paperID
stringlengths 36
36
| pwc_id
stringlengths 8
47
| arxiv_id
stringlengths 6
16
⌀ | nips_id
float64 | url_abs
stringlengths 18
329
| url_pdf
stringlengths 18
742
| title
stringlengths 8
325
| abstract
stringlengths 1
7.27k
⌀ | authors
stringlengths 2
7.06k
| published
stringlengths 10
10
⌀ | conference
stringlengths 12
47
⌀ | conference_url_abs
stringlengths 16
198
⌀ | conference_url_pdf
stringlengths 27
199
⌀ | proceeding
stringlengths 6
47
⌀ | taskID
stringlengths 7
1.44k
| areaID
stringclasses 688
values | embedding
stringlengths 9.26k
12.5k
| umap_embedding
stringlengths 29
44
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ccd7c7b3-bcb4-4d4a-ab85-4164f819e396 | arrhythmia-classification-using-cgan | 2202.00569 | null | https://arxiv.org/abs/2202.00569v4 | https://arxiv.org/pdf/2202.00569v4.pdf | Arrhythmia Classification using CGAN-augmented ECG Signals | ECG databases are usually highly imbalanced due to the abundance of Normal ECG and scarcity of abnormal cases. As such, deep learning classifiers trained on imbalanced datasets usually perform poorly, especially on minor classes. One solution is to generate realistic synthetic ECG signals using Generative Adversarial Networks (GAN) to augment imbalanced datasets. In this study, we combined conditional GAN with WGAN-GP and developed AC-WGAN-GP in 1D form for the first time to be applied on MIT-BIH Arrhythmia dataset. We investigated the impact of data augmentation on arrhythmia classification. We employed two models for ECG generation: (i) unconditional GAN; Wasserstein GAN with gradient penalty (WGAN-GP) is trained on each class individually; (ii) conditional GAN; one Auxiliary Classifier WGAN-GP (AC-WGAN-GP) model is trained on all classes and then used to generate synthetic beats in all classes. Two scenarios are defined for each case: (a) unscreened; all the generated synthetic beats were used, and (b) screened; only a portion of generated beats are selected and used, based on their Dynamic Time Warping (DTW) distance to a designated template. A state-of-the-art ResNet classifier (EcgResNet34) is trained on each of the augmented datasets and the performance metrics (precision/recall/F1-Score micro- and macro-averaged, confusion matrices, multiclass precision-recall curves) were compared with those of the unaugmented imbalanced case. We also used a simple metric Net Improvement. All the three metrics show consistently that net improvement (total and minor-class), unconditional GAN with raw generated data (not screened) creates the best improvements. | ['John J. Prevost', 'Fatemeh Afghah', 'Edmond Adib'] | 2022-01-26 | null | null | null | null | ['arrhythmia-detection'] | ['medical'] | [ 5.82353830e-01 2.73082405e-01 2.15592161e-01 -8.33806321e-02
-1.14943409e+00 -5.79942107e-01 3.27740967e-01 -7.71575943e-02
-1.50427386e-01 1.16550779e+00 1.31179824e-01 -1.98720500e-01
-8.17873776e-02 -8.12690020e-01 -3.57427895e-01 -9.76655185e-01
-1.85504347e-01 6.51166916e-01 -6.34133667e-02 -6.25727251e-02
-1.62908554e-01 3.46552461e-01 -1.05835795e+00 4.27679211e-01
9.82714415e-01 8.54772091e-01 -5.16993463e-01 9.21125710e-01
4.32618052e-01 7.21173763e-01 -1.12415195e+00 -3.72842818e-01
5.14340401e-01 -1.17906535e+00 -5.05837798e-01 -2.52083898e-01
-8.39527473e-02 -1.33251801e-01 -8.77789780e-02 5.84736526e-01
1.18640625e+00 -3.16037476e-01 7.19476283e-01 -1.32596874e+00
-1.54706165e-01 4.33619857e-01 -5.62357128e-01 3.05695295e-01
2.17614710e-01 5.51636159e-01 3.57381552e-01 -4.71107036e-01
7.01494157e-01 6.05087996e-01 1.00352728e+00 7.46343791e-01
-1.42477274e+00 -7.04180777e-01 -5.66860259e-01 -2.84463644e-01
-1.19085515e+00 2.70636678e-02 8.26703131e-01 -4.45061952e-01
6.64468586e-01 4.55527246e-01 9.41978812e-01 1.22395182e+00
4.89054173e-01 3.22110772e-01 1.40294302e+00 -3.13843042e-01
3.01982492e-01 -1.22920498e-02 -1.25807062e-01 2.41875798e-01
1.90024137e-01 3.03574502e-01 -3.09634596e-01 -3.44163328e-01
5.77431083e-01 -1.36199638e-01 -2.57056445e-01 5.19883148e-02
-1.26169252e+00 7.25676298e-01 1.98421836e-01 4.08647239e-01
-7.50604570e-01 -5.49212433e-02 6.52038753e-01 5.35018444e-01
4.15994257e-01 7.17487097e-01 -4.42565829e-01 -3.66997570e-01
-9.57251728e-01 4.28839415e-01 6.30691588e-01 3.98690343e-01
3.89388770e-01 4.26523060e-01 -6.73401654e-01 8.57054889e-01
-3.42463672e-01 3.71601820e-01 9.04827416e-01 -4.40479606e-01
4.37297821e-01 6.32722855e-01 -1.23528734e-01 -7.52150416e-01
-4.22078818e-01 -8.97197962e-01 -1.24320900e+00 2.19636217e-01
2.85798192e-01 -6.23995543e-01 -1.26537633e+00 1.79945147e+00
9.92214233e-02 1.26055256e-01 3.78527582e-01 6.69366956e-01
7.96698630e-01 5.96474349e-01 1.92939579e-01 -4.29542601e-01
1.00635588e+00 -4.17599857e-01 -6.17443740e-01 4.86236438e-02
7.05640256e-01 -6.90842330e-01 8.74645472e-01 3.76495481e-01
-1.20310152e+00 -7.32030392e-01 -1.18088651e+00 5.99275351e-01
-1.51754364e-01 1.91690594e-01 1.75562665e-01 9.68794584e-01
-9.71635699e-01 7.05728352e-01 -7.76852429e-01 -1.18575022e-01
7.10419416e-01 4.03884858e-01 -1.42686963e-01 8.38698968e-02
-1.37003195e+00 7.84577370e-01 2.09974706e-01 7.20327049e-02
-1.10113919e+00 -8.01580906e-01 -5.69899201e-01 -2.34635994e-01
-3.43824804e-01 -9.14513528e-01 6.28444552e-01 -1.35101068e+00
-1.42551529e+00 8.65905344e-01 4.40782249e-01 -6.29496634e-01
9.50661302e-01 1.31114135e-02 -4.65988547e-01 -6.40946701e-02
2.18965895e-02 5.14254451e-01 7.29288220e-01 -1.25341833e+00
-1.25492811e-02 -4.55316305e-01 -3.55359554e-01 2.37637967e-01
4.80897278e-02 -2.61533916e-01 2.18779176e-01 -1.02743936e+00
4.83983345e-02 -1.09635007e+00 -1.24696918e-01 -5.05294442e-01
-7.08245814e-01 3.34580362e-01 8.40362549e-01 -9.52290058e-01
1.22719049e+00 -2.05127430e+00 1.00982256e-01 4.32390809e-01
1.81003571e-01 6.35223746e-01 -1.39791563e-01 4.14777040e-01
-5.39166868e-01 3.46427649e-01 -6.73156381e-01 -5.17685339e-03
-5.15941978e-01 1.01774931e-01 -1.21773466e-01 3.74619097e-01
2.77331680e-01 9.60626781e-01 -7.97951043e-01 -1.97824791e-01
1.82515591e-01 5.47033370e-01 -4.37604338e-01 4.27693695e-01
2.62097657e-01 8.45056295e-01 -2.75820345e-01 4.98703480e-01
6.15002155e-01 2.16545597e-01 1.29291818e-01 -3.64015102e-01
4.19474393e-01 -1.34631157e-01 -1.07571363e+00 1.44395411e+00
-3.71845663e-01 3.79350275e-01 -6.04443789e-01 -9.60444868e-01
1.19053292e+00 6.29638314e-01 6.01890743e-01 -4.13955897e-01
4.81006540e-02 2.97198743e-01 5.51616609e-01 -3.97231758e-01
-2.22647339e-01 -4.53093261e-01 -9.50531214e-02 4.91600841e-01
1.73637226e-01 -4.02306437e-01 -2.03675851e-02 -1.64201539e-02
1.37398422e+00 2.62261122e-01 2.42694870e-01 -2.03885958e-01
3.08366776e-01 -1.27034113e-01 6.63371146e-01 7.84337699e-01
-1.38319284e-01 1.17913783e+00 8.64752650e-01 -6.88785255e-01
-1.03717375e+00 -1.11387229e+00 4.49103594e-04 2.84875900e-01
-3.41239572e-01 -1.54192716e-01 -9.99256313e-01 -7.61506319e-01
-3.68688732e-01 7.26381838e-01 -7.13990152e-01 -5.67977667e-01
-5.07365644e-01 -1.47705424e+00 1.05595958e+00 5.18595099e-01
5.51279664e-01 -1.40563416e+00 -8.91966581e-01 4.15675968e-01
-9.61100757e-02 -5.66755891e-01 -8.10887963e-02 2.70126611e-01
-9.53521490e-01 -1.11156964e+00 -8.31560254e-01 -4.17485952e-01
5.77686787e-01 -8.06358635e-01 1.24562311e+00 1.02366343e-01
-5.76196432e-01 -5.06888051e-03 -3.74079376e-01 -7.18527973e-01
-7.39321589e-01 -2.40609869e-02 -8.63083825e-02 2.54668891e-01
-6.51274109e-03 -8.66503596e-01 -9.45823252e-01 2.82101393e-01
-7.53929615e-01 3.74800786e-02 6.28352106e-01 1.17349422e+00
6.54042006e-01 -1.84310183e-01 1.16125882e+00 -1.27143884e+00
7.02798128e-01 -3.18233401e-01 -5.08761220e-02 -5.98195158e-02
-6.08672500e-01 -4.57208037e-01 6.54246390e-01 -5.24831653e-01
-6.48941159e-01 -3.51841934e-02 -3.96651626e-01 -3.77541184e-01
-1.76300146e-02 3.25713664e-01 -1.73920676e-01 2.33643264e-01
1.06495368e+00 2.93202996e-01 -2.03431826e-02 4.11421023e-02
-1.27249226e-01 5.53874493e-01 4.63880509e-01 -4.50160563e-01
6.58133209e-01 1.41518936e-01 7.71186203e-02 -5.26141703e-01
-4.28041160e-01 1.76377743e-01 -5.40184319e-01 -1.98394731e-01
1.02937400e+00 -7.24110901e-01 2.33867615e-02 7.93850899e-01
-7.49938428e-01 -5.57157099e-01 -8.72134209e-01 4.66689497e-01
-5.66249192e-01 -1.21299401e-01 -5.64428806e-01 -8.22738409e-01
-9.60705817e-01 -9.93230820e-01 8.75999033e-01 1.48320228e-01
-4.69022751e-01 -7.01290250e-01 3.45838517e-01 2.21595317e-01
5.93624294e-01 1.35670042e+00 9.68209386e-01 -9.33664143e-01
1.77878123e-02 -5.89102030e-01 1.92760155e-01 7.53603518e-01
2.35923916e-01 -1.38543993e-01 -1.06397915e+00 -3.10394078e-01
1.55802131e-01 -2.62126714e-01 3.95057261e-01 5.73488832e-01
1.30849004e+00 -2.17986926e-01 -2.39704877e-01 5.69487214e-01
1.41242254e+00 6.91260815e-01 1.19120741e+00 -9.65850651e-02
6.46575630e-01 1.74778402e-02 2.40676969e-01 1.49976373e-01
-1.73267201e-01 4.61044133e-01 1.48555979e-01 -6.38309121e-01
-2.92343736e-01 -2.12679490e-01 7.94113129e-02 6.55274391e-01
-4.72650617e-01 -4.30543363e-01 -9.66862142e-01 4.78812933e-01
-1.35830927e+00 -9.34482992e-01 -4.24045362e-02 2.43099141e+00
8.84747922e-01 4.28934127e-01 2.85192877e-01 7.21845865e-01
7.06452131e-01 -9.81682837e-02 -5.50737858e-01 -4.31262612e-01
-3.53038728e-01 8.12527657e-01 1.01938687e-01 -2.50114575e-02
-9.02192652e-01 3.10135365e-01 5.50032330e+00 5.98809421e-01
-1.31597674e+00 3.34838301e-01 1.42737806e+00 -2.23616492e-02
-7.14780688e-02 -1.77628681e-01 -7.71609545e-02 6.42961860e-01
1.03170824e+00 2.49948911e-03 1.93589732e-01 4.67307389e-01
4.82394509e-02 6.18668534e-02 -9.63947594e-01 8.74822259e-01
-4.49438915e-02 -1.07467127e+00 2.48914305e-03 -8.81245136e-02
1.03013110e+00 -1.82614177e-01 -1.21661440e-01 4.16254401e-01
4.98471372e-02 -1.19264030e+00 2.76216120e-01 5.84246218e-01
1.48792148e+00 -8.91734362e-01 1.29382050e+00 1.86937258e-01
-6.86666965e-01 9.16403979e-02 7.20363855e-02 2.17828140e-01
1.11713126e-01 8.32115471e-01 -7.53203571e-01 7.76277006e-01
5.15508413e-01 3.58699352e-01 -4.66839403e-01 7.66104996e-01
-1.62365809e-02 1.06102240e+00 -2.14010045e-01 3.41906279e-01
-1.26108184e-01 -1.61317378e-01 6.81365192e-01 1.01042867e+00
4.86834824e-01 9.97936875e-02 -7.48097450e-02 6.99201822e-01
7.86621938e-04 5.59982061e-02 -5.92633188e-01 2.85732359e-01
2.90092647e-01 1.23370790e+00 -8.27439964e-01 -3.72414380e-01
6.65650591e-02 8.76332581e-01 -2.59635419e-01 2.23063573e-01
-9.59908009e-01 -5.56913793e-01 1.43920824e-01 3.76739323e-01
-1.78441808e-01 5.55240989e-01 -6.30560040e-01 -7.33961523e-01
-6.24077655e-02 -1.14062452e+00 6.42882109e-01 -7.67504632e-01
-1.34770834e+00 1.12612736e+00 -1.32149279e-01 -1.48011625e+00
-3.50268006e-01 1.34781469e-02 -9.53234971e-01 1.18242478e+00
-8.33496153e-01 -1.11744404e+00 -5.48708975e-01 3.57599586e-01
4.22960311e-01 -3.49319786e-01 1.19182384e+00 3.80763590e-01
-3.21478814e-01 8.26711357e-01 -3.22373509e-01 3.53172451e-01
4.73789901e-01 -1.25120330e+00 1.85326725e-01 7.23752737e-01
-1.42894998e-01 2.66187191e-01 5.84981203e-01 -6.96637213e-01
-8.50637138e-01 -1.43440092e+00 5.42557478e-01 -4.11268830e-01
-6.10865429e-02 -2.20317096e-01 -7.83256054e-01 4.91164327e-01
7.35835582e-02 1.81790963e-01 6.78564250e-01 -4.00882363e-01
1.72083765e-01 -3.11915398e-01 -1.64776349e+00 4.09634352e-01
7.25295544e-01 -1.47202820e-01 -4.54739243e-01 3.33712846e-01
3.22434932e-01 -7.32260406e-01 -1.17008924e+00 8.36138844e-01
5.57194889e-01 -1.04754996e+00 7.37483323e-01 -5.46391845e-01
5.41347444e-01 -4.06711996e-01 7.69046843e-02 -1.64369369e+00
1.22518703e-01 -6.92900121e-01 2.02151075e-01 1.22359383e+00
5.90694249e-01 -8.92154694e-01 6.44549072e-01 2.21929833e-01
-2.54909664e-01 -1.22467887e+00 -9.36930537e-01 -5.19065976e-01
1.38539672e-01 -6.05605245e-02 6.54489160e-01 9.68415618e-01
-3.47689509e-01 1.78380370e-01 -3.95869195e-01 -2.11905465e-01
3.17137063e-01 -1.46453843e-01 6.51568115e-01 -9.58055556e-01
-3.41952890e-01 -9.79915783e-02 -6.96046650e-01 1.14939690e-01
-3.39304924e-01 -8.68190169e-01 -2.85145402e-01 -1.36747360e+00
6.78713620e-02 -7.09416091e-01 -4.71974522e-01 4.94537175e-01
-3.57121050e-01 8.84375811e-01 9.88208577e-02 8.03921968e-02
2.82895952e-01 1.50805622e-01 1.12252104e+00 -2.48261690e-02
-5.37276387e-01 3.45819890e-01 -6.58306897e-01 3.65752012e-01
9.78328824e-01 -6.37133718e-01 -5.09992361e-01 2.62669295e-01
-5.18658571e-02 4.58048791e-01 2.97025114e-01 -1.55457270e+00
-4.27798003e-01 3.67645085e-01 9.10957575e-01 -4.16489810e-01
1.28747180e-01 -4.10577565e-01 8.84442151e-01 8.27302933e-01
-3.16466928e-01 2.66838700e-01 2.74772793e-01 1.64426088e-01
-2.60205328e-01 6.32762387e-02 9.99047816e-01 -7.37445205e-02
2.34648734e-01 2.67941386e-01 -2.37094149e-01 2.97255129e-01
1.12217522e+00 -3.19782764e-01 -1.34606883e-01 -4.56996411e-01
-1.06356645e+00 -2.49108016e-01 2.16265753e-01 1.21073993e-02
4.45996195e-01 -1.28350246e+00 -1.11443734e+00 3.05127293e-01
-7.59160668e-02 -4.47669160e-03 5.68228960e-01 9.37013447e-01
-7.14518487e-01 -2.33267814e-01 -4.30982590e-01 -7.27044404e-01
-1.07936680e+00 3.39510381e-01 7.64655948e-01 -7.67852843e-01
-5.70997119e-01 6.74495697e-01 1.26995429e-01 -3.70784461e-01
-1.58528700e-01 -1.05690263e-01 -1.30377427e-01 4.35429215e-02
2.01084688e-01 3.76283675e-01 4.56800431e-01 -3.44491810e-01
-2.05009133e-01 5.15559554e-01 3.49200428e-01 6.82113618e-02
1.40160596e+00 4.94370490e-01 -2.37696953e-02 3.60202223e-01
9.32309926e-01 -1.10264398e-01 -1.06344712e+00 4.16579425e-01
-4.91295218e-01 -1.26155019e-01 -2.30913639e-01 -1.28319228e+00
-1.46860027e+00 7.71331191e-01 1.19609380e+00 2.20479473e-01
1.64617729e+00 -5.27343810e-01 5.94071865e-01 -3.93305629e-01
1.73985571e-01 -6.61160707e-01 -9.48641300e-02 -1.47099048e-01
9.25544381e-01 -6.68561578e-01 4.33208309e-02 -1.35099784e-01
-9.90828812e-01 9.11946595e-01 5.17629147e-01 -3.18396389e-01
5.29911399e-01 3.37642074e-01 3.76346737e-01 -1.89166754e-01
-4.40598130e-01 3.90163511e-01 1.70043886e-01 9.04824555e-01
4.64533567e-01 7.94087350e-02 -5.55933654e-01 7.39794850e-01
-4.44547474e-01 2.48403147e-01 4.98147070e-01 8.10818315e-01
5.31525493e-01 -9.63106930e-01 -1.37276188e-01 1.18494415e+00
-7.73686826e-01 -6.07895553e-02 -7.76845515e-02 7.70251930e-01
4.96699035e-01 5.75033724e-01 -2.83907447e-02 -4.85260516e-01
4.28102314e-01 3.14392060e-01 3.00705701e-01 -5.16787767e-01
-1.20951259e+00 -3.17375027e-02 9.63435695e-02 -2.60512948e-01
-2.14413241e-01 -7.76374042e-01 -9.29417253e-01 2.22666800e-01
-2.32645825e-01 1.68382958e-01 5.11050880e-01 5.98552167e-01
4.17029977e-01 9.67271328e-01 7.12715566e-01 -6.15752816e-01
-4.66820210e-01 -1.34284222e+00 -5.76748490e-01 5.99993289e-01
1.37191072e-01 -3.80053639e-01 -5.95106959e-01 2.10777536e-01] | [14.278203010559082, 3.129563093185425] |
cbb8707c-39b4-4a61-bda6-caf98f96f894 | nature-language-reasoning-a-survey | 2303.14725 | null | https://arxiv.org/abs/2303.14725v2 | https://arxiv.org/pdf/2303.14725v2.pdf | Natural Language Reasoning, A Survey | This survey paper proposes a clearer view of natural language reasoning in the field of Natural Language Processing (NLP), both conceptually and practically. Conceptually, we provide a distinct definition for natural language reasoning in NLP, based on both philosophy and NLP scenarios, discuss what types of tasks require reasoning, and introduce a taxonomy of reasoning. Practically, we conduct a comprehensive literature review on natural language reasoning in NLP, mainly covering classical logical reasoning, natural language inference, multi-hop question answering, and commonsense reasoning. The paper also identifies and views backward reasoning, a powerful paradigm for multi-step reasoning, and introduces defeasible reasoning as one of the most important future directions in natural language reasoning research. We focus on single-modality unstructured natural language text, excluding neuro-symbolic techniques and mathematical reasoning. | ['Benyou Wang', 'Prayag Tiwari', 'Hongbo Zhang', 'Fei Yu'] | 2023-03-26 | null | null | null | null | ['multi-hop-question-answering', 'philosophy', 'mathematical-reasoning', 'logical-reasoning'] | ['knowledge-base', 'miscellaneous', 'natural-language-processing', 'reasoning'] | [ 2.05752671e-01 1.09973335e+00 -3.98677528e-01 -5.90374887e-01
-9.02146846e-02 -9.42222297e-01 7.58863628e-01 5.42381525e-01
-4.91200507e-01 9.95268106e-01 5.53930700e-01 -8.72233272e-01
-7.03220248e-01 -1.12038398e+00 -3.89342844e-01 -2.03282475e-01
1.85445994e-01 6.98429763e-01 1.77659437e-01 -6.41484857e-01
5.35554945e-01 6.19558871e-01 -1.01594126e+00 6.34160161e-01
9.54659283e-01 5.77572346e-01 -2.47785270e-01 4.43130493e-01
-8.47596169e-01 2.07374215e+00 -3.47944587e-01 -1.08243775e+00
-2.05624864e-01 -3.64872664e-01 -1.87504900e+00 -4.57219630e-01
-9.09670740e-02 -3.20339948e-01 -1.07279748e-01 1.28815937e+00
6.47174120e-02 3.19304049e-01 4.58787560e-01 -1.46514046e+00
-9.83890533e-01 1.23169231e+00 8.44781771e-02 2.03265116e-01
1.27205443e+00 2.12196317e-02 9.99935389e-01 -4.32527363e-01
7.48838484e-01 1.83051288e+00 5.58612585e-01 8.89237225e-01
-8.20047975e-01 -2.14522243e-01 3.41955066e-01 6.79831922e-01
-9.75749373e-01 -9.08700228e-02 5.66442788e-01 -4.23730999e-01
1.42267847e+00 4.38305914e-01 6.58144534e-01 6.82358265e-01
5.81520021e-01 9.63965416e-01 1.12994313e+00 -8.91566813e-01
5.14867961e-01 1.56295463e-01 7.71704853e-01 8.50929499e-01
3.15460712e-01 -8.94707739e-02 -6.09437466e-01 -4.65053499e-01
4.94482696e-01 -4.94825542e-01 -3.09162550e-02 -8.47798064e-02
-1.20907950e+00 8.20159733e-01 -5.74106956e-03 5.31507552e-01
-4.14946258e-01 8.60713869e-02 6.04001284e-01 3.60276759e-01
-3.08528692e-01 4.87206578e-01 -5.92339814e-01 2.91319098e-02
-6.40702307e-01 5.39554060e-01 1.23708606e+00 6.50543690e-01
7.37433508e-02 -3.24461311e-01 -2.31947094e-01 3.67589951e-01
6.77234650e-01 5.88532150e-01 3.01070839e-01 -1.68752992e+00
4.20825601e-01 5.65587163e-01 1.80225775e-01 -1.06162357e+00
-4.86208439e-01 4.37422872e-01 -5.24630308e-01 1.96290046e-01
4.86270130e-01 -1.00014322e-01 -9.60302912e-03 1.58186650e+00
3.66276383e-01 -6.56951427e-01 7.63686955e-01 6.87435210e-01
1.06462181e+00 4.38107580e-01 4.78808552e-01 -4.97604847e-01
1.98744965e+00 -9.07523692e-01 -1.11739457e+00 -2.16059864e-01
4.31124449e-01 -1.70229971e-01 9.70390856e-01 5.69028556e-01
-1.55683482e+00 2.70747632e-01 -4.62689787e-01 -8.15539241e-01
-7.89895833e-01 -2.84858525e-01 1.02012050e+00 3.19886059e-01
-1.10311782e+00 1.02269486e-01 -3.45172882e-01 -4.99508828e-01
4.47988182e-01 4.67289574e-02 -6.99410588e-02 -3.74866366e-01
-1.63218939e+00 1.38938963e+00 9.06403720e-01 4.95396145e-02
-2.26098448e-01 -5.93820393e-01 -1.02666080e+00 5.27273305e-02
8.52656662e-01 -1.21917212e+00 1.29980206e+00 -4.87044901e-01
-1.51080024e+00 1.41619503e+00 -4.41080660e-01 -9.24767494e-01
5.08584797e-01 -4.10259217e-02 -6.50017321e-01 7.66938925e-01
1.60227582e-01 5.22254407e-01 -3.29121202e-03 -8.51624668e-01
-4.42136049e-01 -4.10108328e-01 8.90057921e-01 2.42770135e-01
4.06323284e-01 5.12877226e-01 1.04104385e-01 -3.02479476e-01
8.62941146e-02 -2.52922624e-01 -2.20617801e-01 6.53912604e-01
-2.71901131e-01 -7.62926280e-01 7.22822994e-02 -5.14983535e-01
9.51604545e-01 -1.69621789e+00 -6.58951476e-02 1.08269371e-01
4.23812449e-01 -1.99351966e-01 1.82853818e-01 6.10253632e-01
-4.53251451e-02 1.21825188e-01 -2.92905599e-01 7.25936115e-01
4.99591112e-01 6.93389595e-01 -1.00325429e+00 -1.19712278e-01
3.33312303e-02 1.37962306e+00 -1.31748724e+00 -1.07265246e+00
2.88528651e-01 9.35800225e-02 -4.52757806e-01 -3.93769860e-01
-5.84436178e-01 1.08262442e-01 -7.85969734e-01 7.21058786e-01
4.76966262e-01 -2.80196816e-01 4.98423487e-01 1.17293475e-02
-4.18251641e-02 4.75250155e-01 -8.56325686e-01 1.43647563e+00
-3.00654024e-01 4.27061200e-01 -1.01204306e-01 -1.00689864e+00
5.05923688e-01 7.27046967e-01 -4.17256430e-02 -6.30236268e-01
6.53203279e-02 3.61955538e-02 -8.74337405e-02 -1.19529486e+00
-4.21044081e-02 -8.64044964e-01 -2.34549269e-02 7.24050701e-01
-1.64035112e-01 -5.89718699e-01 6.23442829e-01 3.53248328e-01
9.01044428e-01 1.83094114e-01 1.28117585e+00 -5.16556442e-01
1.25388741e+00 5.88572383e-01 2.78665006e-01 9.30806935e-01
-4.86641645e-01 -4.26338702e-01 6.92683935e-01 -7.68883944e-01
-2.35139713e-01 -1.42338717e+00 -6.64979219e-02 1.07138526e+00
1.71391845e-01 -1.95829406e-01 -6.15246832e-01 -3.21080178e-01
-8.45577717e-02 1.45874894e+00 -4.07331407e-01 5.34771197e-03
-3.19990128e-01 -5.49899817e-01 9.86962259e-01 4.79496598e-01
7.10232675e-01 -1.70526814e+00 -8.51922035e-01 -6.72783032e-02
-6.32933736e-01 -1.51613712e+00 6.99514091e-01 -1.02912880e-01
-1.10367167e+00 -1.44713557e+00 -1.17857575e-01 -7.32371867e-01
7.13857114e-01 -2.70883381e-01 1.16993082e+00 3.09333861e-01
8.07063505e-02 6.96871579e-01 -2.53514290e-01 -6.10774100e-01
-4.56722230e-01 -9.05068219e-01 1.36524975e-01 -8.57158005e-01
9.63175833e-01 -4.09198046e-01 -2.90756747e-02 -3.72853011e-01
-8.19557965e-01 -6.53252282e-05 1.83451101e-01 4.81387466e-01
2.72458613e-01 6.71017528e-01 2.85965890e-01 -8.15601170e-01
1.30563247e+00 -3.44296783e-01 -2.14035735e-01 8.61395597e-01
-3.85019898e-01 2.56663859e-01 8.33992243e-01 1.00203201e-01
-1.65334630e+00 -8.00140202e-01 -1.51749521e-01 5.41805446e-01
-4.42760736e-01 6.49275064e-01 -4.16513458e-02 5.72199747e-02
6.89979136e-01 4.30543393e-01 -2.06983536e-01 4.43994142e-02
7.07665026e-01 9.18519199e-02 6.71234608e-01 -1.30545700e+00
6.59225702e-01 8.37732792e-01 1.54746696e-01 -5.84360898e-01
-1.24343383e+00 -3.37206502e-03 -4.95412588e-01 -3.07342317e-03
1.00147069e+00 -5.35902441e-01 -1.52302575e+00 -1.91183329e-01
-1.39745438e+00 -4.38923180e-01 -7.69396305e-01 4.89525437e-01
-1.01494980e+00 7.34627604e-01 -9.73565996e-01 -1.28739393e+00
-6.66930676e-01 -5.66446304e-01 6.94358766e-01 1.11069195e-01
-7.71948516e-01 -1.32731843e+00 -6.52468130e-02 9.50128496e-01
-5.62600084e-02 9.18913558e-02 1.66090024e+00 -7.51862347e-01
6.70205150e-03 3.08064908e-01 -3.48867387e-01 -4.21206541e-02
-2.37235531e-01 -1.32940471e-01 -5.86724699e-01 6.68793678e-01
4.53098327e-01 -6.23878896e-01 8.41575637e-02 3.17699790e-01
9.77883339e-01 -5.48116386e-01 -1.24443136e-01 -3.09533864e-01
1.29266727e+00 3.77760530e-01 6.81209445e-01 4.85680610e-01
-1.99371859e-01 1.19011426e+00 7.43086576e-01 1.62195504e-01
8.50613654e-01 -2.59655267e-01 -1.21083945e-01 5.94650090e-01
2.29396373e-01 -1.56591594e-01 6.92540333e-02 4.25354481e-01
-4.87444162e-01 6.33975416e-02 -1.35580552e+00 2.68096298e-01
-1.92897189e+00 -1.44997871e+00 -1.02210604e-01 1.23266339e+00
1.20743656e+00 5.23796305e-02 -2.74061620e-01 3.42074871e-01
4.39990461e-01 -2.51028836e-01 -3.76164109e-01 -9.98050630e-01
-3.10074151e-01 7.38514289e-02 -3.73678386e-01 7.35165536e-01
-5.41797101e-01 1.16294754e+00 8.03025246e+00 5.33018768e-01
-3.65324408e-01 -9.97897685e-02 -1.79304089e-02 1.82814449e-01
-4.74764198e-01 9.69526619e-02 -2.47741237e-01 -1.38990879e-01
7.24016368e-01 -4.20016944e-01 6.93927467e-01 5.08452356e-01
3.24704438e-01 -4.33785200e-01 -1.34323084e+00 8.35366070e-01
1.23040721e-01 -1.52648234e+00 5.85576773e-01 -4.55756813e-01
2.44365931e-01 -3.98096293e-01 -6.10930085e-01 3.35687310e-01
7.21616030e-01 -8.18798482e-01 8.90341163e-01 6.85620368e-01
-9.51569304e-02 -4.02553350e-01 7.25997031e-01 8.00413728e-01
-7.80649006e-01 -4.83796239e-01 -2.86222249e-01 -7.73628712e-01
4.26603526e-01 8.73290062e-01 -2.15013117e-01 7.01163292e-01
7.18974769e-01 2.38991633e-01 -1.09855369e-01 4.98354524e-01
-9.59554076e-01 9.51487385e-03 -3.20891500e-01 -4.03857082e-01
1.00069031e-01 -2.66541213e-01 3.95052493e-01 1.12465417e+00
-4.45672452e-01 8.87642443e-01 -1.98480010e-01 1.21169710e+00
4.07486141e-01 -2.40697339e-02 -4.88381654e-01 -2.15077102e-01
4.21885759e-01 5.64876556e-01 -7.28773832e-01 -7.90303707e-01
-4.39265817e-01 6.31739318e-01 2.67988145e-01 4.86230254e-01
-7.67232776e-01 -1.50565907e-01 2.24796504e-01 -4.41014677e-01
-4.31127340e-01 -1.26466826e-01 -5.48487544e-01 -1.33855045e+00
-1.86483823e-02 -8.26950073e-01 9.45891798e-01 -1.44283402e+00
-1.51772058e+00 1.99990600e-01 5.70695937e-01 -4.83809590e-01
-2.63077199e-01 -1.12348747e+00 -5.18721998e-01 5.11556566e-01
-1.61865664e+00 -9.43886042e-01 1.80885449e-01 6.98242962e-01
2.41434768e-01 1.56914353e-01 1.26379430e+00 -2.02101678e-01
-5.52087314e-02 -1.29967004e-01 -6.79637790e-01 2.90025175e-01
6.99371919e-02 -1.03474557e+00 -3.74751836e-02 6.29203975e-01
-2.79072940e-01 1.24206078e+00 7.56062508e-01 -5.08299649e-01
-1.54709363e+00 -5.04461825e-01 1.44187212e+00 -4.32322681e-01
1.01415813e+00 1.85734451e-01 -5.18316984e-01 1.02692139e+00
2.13674143e-01 -3.53058934e-01 1.08054864e+00 -4.74803858e-02
-4.30690795e-01 4.00457472e-01 -1.85525799e+00 1.19545746e+00
1.07939720e+00 -8.16500127e-01 -1.79608905e+00 6.50234640e-01
8.73600960e-01 -1.64942667e-01 -9.12546277e-01 5.52491322e-02
6.79359794e-01 -6.57520831e-01 1.22118616e+00 -1.08468568e+00
5.30693352e-01 -4.41992223e-01 -2.16372624e-01 -4.01785254e-01
-2.63666630e-01 -3.29095453e-01 -4.03843611e-01 7.62175322e-01
3.26813877e-01 -1.02148116e+00 3.58400643e-01 1.52556503e+00
4.79010075e-01 -6.96622550e-01 -7.63377190e-01 -4.61798906e-01
5.02386928e-01 -8.26395035e-01 4.88060087e-01 1.16280735e+00
1.13221920e+00 4.15747702e-01 3.93902361e-01 -1.69126485e-02
6.48862302e-01 4.23920602e-01 1.14802375e-01 -1.27600002e+00
-1.53438464e-01 -5.75505257e-01 -1.70580700e-01 -7.45034635e-01
6.79449558e-01 -1.07104743e+00 -1.90660492e-01 -2.27336407e+00
2.11050048e-01 4.18172598e-01 2.78971553e-01 9.23905611e-01
4.03509021e-01 -2.71838039e-01 8.05776417e-02 -1.13992982e-01
-4.47799772e-01 8.66427571e-02 1.60561931e+00 -2.39536285e-01
1.81629553e-01 -4.28498715e-01 -1.05966973e+00 1.38786602e+00
8.46412539e-01 -1.79889053e-01 -7.00277030e-01 -4.26230490e-01
1.10728848e+00 2.27831274e-01 6.30735755e-01 -3.75436157e-01
4.90604252e-01 -1.02180541e+00 2.86502931e-02 -4.77412134e-01
9.73137021e-02 -1.04364860e+00 -4.70405430e-01 8.31996441e-01
-7.09717751e-01 -2.43384629e-01 1.67901769e-01 5.00853360e-02
-8.75117555e-02 -7.36766040e-01 4.91784453e-01 -5.74090421e-01
-1.08493125e+00 -5.03246963e-01 -8.30032825e-01 5.47876716e-01
1.05999756e+00 -5.21752872e-02 -5.55432916e-01 -9.00586247e-02
-1.12765944e+00 6.31498933e-01 -2.33228385e-01 -3.96452062e-02
7.30238676e-01 -8.70556712e-01 -3.40400636e-01 -6.61060572e-01
-1.47774443e-01 -1.60386384e-01 2.61591941e-01 1.01190877e+00
-9.81637955e-01 8.37100744e-01 -9.53546464e-02 1.57379642e-01
-9.03175056e-01 9.50438440e-01 4.88627434e-01 -2.73623168e-01
-9.89582062e-01 5.58519959e-01 4.84226421e-02 -7.54657865e-01
1.84034079e-01 -7.44000375e-01 -4.96273100e-01 -3.82778287e-01
9.69537437e-01 2.40567356e-01 -4.48150456e-01 -2.95244336e-01
-9.43855166e-01 5.76094925e-01 3.71260971e-01 -3.29897970e-01
8.13452601e-01 -3.38182986e-01 -1.10445237e+00 7.34514117e-01
3.75394106e-01 -3.97482365e-01 2.18524933e-01 -2.58904159e-01
3.04727554e-01 1.52588978e-01 -3.53959918e-01 -1.25663829e+00
-1.93651289e-01 6.46017671e-01 -3.57012510e-01 1.61176488e-01
9.87841070e-01 3.62566650e-01 5.64434946e-01 1.41501677e+00
7.62044072e-01 -1.41323173e+00 -4.30734366e-01 6.56887174e-01
1.12836170e+00 -9.31865275e-01 4.46656585e-01 -7.68667936e-01
-6.38156950e-01 1.39377904e+00 2.88287312e-01 5.44200279e-02
7.67166913e-01 5.88676870e-01 1.67292971e-02 -6.49565160e-01
-7.35623002e-01 3.45315635e-02 -1.41112596e-01 6.62086427e-01
4.22495782e-01 3.65600698e-02 -7.69780457e-01 7.39063263e-01
-7.67970800e-01 7.44879484e-01 3.82042438e-01 1.37611008e+00
-2.98051327e-01 -7.86594689e-01 -7.44279802e-01 8.14593490e-03
-4.38352257e-01 -2.47669503e-01 -7.40248203e-01 7.38475144e-01
7.54672587e-02 1.30964005e+00 -1.86485320e-01 4.37279254e-01
1.33821547e-01 3.86230707e-01 9.98994410e-01 -3.00293028e-01
-4.82446313e-01 -9.57616866e-01 3.69954616e-01 -5.74418962e-01
-1.07823229e+00 -3.69082958e-01 -2.21787310e+00 -5.08093536e-01
2.45070055e-01 3.45131278e-01 9.71808732e-02 1.74047291e+00
-3.26019347e-01 3.70533466e-01 -6.16789818e-01 1.02673560e-01
-3.92993152e-01 -1.72346041e-01 -3.33585203e-01 4.75040674e-02
1.45624995e-01 -3.06995600e-01 -2.10636631e-01 2.83675104e-01] | [9.180093765258789, 7.12153434753418] |
f1cb92e4-ec87-49c6-9faf-bf1eac56b115 | utility-oriented-underwater-image-quality | 2205.03574 | null | https://arxiv.org/abs/2205.03574v1 | https://arxiv.org/pdf/2205.03574v1.pdf | Utility-Oriented Underwater Image Quality Assessment Based on Transfer Learning | The widespread image applications have greatly promoted the vision-based tasks, in which the Image Quality Assessment (IQA) technique has become an increasingly significant issue. For user enjoyment in multimedia systems, the IQA exploits image fidelity and aesthetics to characterize user experience; while for other tasks such as popular object recognition, there exists a low correlation between utilities and perceptions. In such cases, the fidelity-based and aesthetics-based IQA methods cannot be directly applied. To address this issue, this paper proposes a utility-oriented IQA in object recognition. In particular, we initialize our research in the scenario of underwater fish detection, which is a critical task that has not yet been perfectly addressed. Based on this task, we build an Underwater Image Utility Database (UIUD) and a learning-based Underwater Image Utility Measure (UIUM). Inspired by the top-down design of fidelity-based IQA, we exploit the deep models of object recognition and transfer their features to our UIUM. Experiments validate that the proposed transfer-learning-based UIUM achieves promising performance in the recognition task. We envision our research provides insights to bridge the researches of IQA and computer vision. | ['Patrick Le Callet', 'Ke Gu', 'Tiesong Zhao', 'Honggang Liao', 'Rongfu Lin', 'Weiling Chen'] | 2022-05-07 | null | null | null | null | ['fish-detection'] | ['computer-vision'] | [ 1.20024905e-01 -2.30033368e-01 4.04252827e-01 -4.82484519e-01
-5.79557121e-01 -1.37736484e-01 2.95012444e-01 1.24693848e-01
-4.96136874e-01 2.59193987e-01 2.07833722e-01 3.08228843e-02
-3.66629392e-01 -1.06929207e+00 -6.20164633e-01 -7.45866060e-01
-1.73058376e-01 -3.33022743e-01 1.31899148e-01 -2.45520741e-01
5.37873983e-01 2.59666145e-02 -1.86202896e+00 4.55157757e-02
1.02715552e+00 1.54541850e+00 5.31200230e-01 5.45183420e-01
-1.66863918e-01 5.47660828e-01 -4.21998292e-01 -5.52760482e-01
1.78036958e-01 -5.56485355e-01 -5.47231495e-01 -5.28100394e-02
1.77084729e-01 -6.25656009e-01 -2.98106968e-01 1.30298150e+00
7.10974514e-01 2.46910974e-01 7.40844190e-01 -1.34579539e+00
-1.12324703e+00 3.23313266e-01 -3.08140814e-01 1.80720109e-02
2.53472030e-01 1.15719162e-01 1.18005872e+00 -1.00457621e+00
1.23017021e-01 1.39074528e+00 5.38375795e-01 4.31249887e-01
-5.22751689e-01 -4.44888949e-01 1.29376790e-02 5.73049664e-01
-1.08793306e+00 -1.54015332e-01 6.14695668e-01 -4.13689256e-01
3.03909510e-01 2.52574146e-01 1.03772020e+00 4.99833107e-01
-1.34053864e-02 1.06865895e+00 1.21327293e+00 -2.16677383e-01
5.19134760e-01 2.55067367e-02 -1.46522909e-01 4.56371635e-01
1.91636741e-01 -1.21246334e-02 -5.15109301e-01 2.48233005e-01
8.52522910e-01 1.88044101e-01 -4.89231616e-01 -1.41244635e-01
-7.00044513e-01 8.05790603e-01 7.91681051e-01 -1.18060812e-01
-2.51280874e-01 -2.65630446e-02 2.99900472e-01 4.17699963e-01
2.28728101e-01 3.03366750e-01 -1.05494363e-02 -3.29134285e-01
-5.85134149e-01 -8.21103081e-02 6.83261514e-01 7.19173491e-01
8.48675847e-01 1.68896735e-01 -6.61050081e-02 9.64255929e-01
6.81900561e-01 6.28594279e-01 6.51023448e-01 -1.02146041e+00
-1.70432508e-01 7.08257556e-01 7.43818358e-02 -1.06507361e+00
-2.46463925e-01 -3.91700238e-01 -5.80048263e-01 5.20915449e-01
6.65109605e-02 2.70628422e-01 -6.18302286e-01 1.50350404e+00
6.88078851e-02 1.35630757e-01 4.65422601e-01 1.35866773e+00
1.21903014e+00 8.39701116e-01 9.44223404e-02 -1.67558730e-01
1.30436575e+00 -7.81019032e-01 -5.79689443e-01 1.91419408e-01
2.41481096e-01 -6.01028085e-01 1.48261583e+00 5.08894920e-01
-9.15629208e-01 -5.32332003e-01 -1.27411580e+00 -1.22666854e-04
-2.27055073e-01 -6.75585717e-02 5.65354049e-01 7.77457535e-01
-1.06682551e+00 5.31517565e-01 -5.26412904e-01 -5.25151849e-01
4.33819264e-01 -1.19012287e-02 -3.30777854e-01 -2.45764539e-01
-1.10375440e+00 9.14421082e-01 2.08292559e-01 3.05198222e-01
-1.26303840e+00 -2.63967335e-01 -9.16763425e-01 1.09144993e-01
-6.37716055e-03 -4.37074393e-01 1.30056977e+00 -1.16502583e+00
-1.62349653e+00 5.91370165e-01 3.94240767e-01 -2.24354655e-01
4.59530503e-01 -2.58526832e-01 -2.81442374e-01 1.82022721e-01
-3.73037308e-02 3.41139704e-01 5.59737027e-01 -1.54512739e+00
-8.46031427e-01 -3.20383579e-01 4.76339787e-01 6.12343669e-01
-9.88255799e-01 -2.15346947e-01 -5.56527317e-01 -3.02162826e-01
5.89266084e-02 -2.96319366e-01 1.70676127e-01 7.89004266e-01
1.60223931e-01 -3.74435037e-02 7.66759634e-01 -4.45179790e-01
9.91407454e-01 -2.22390246e+00 1.28299877e-01 -3.24036777e-02
-1.15765691e-01 2.45359316e-01 -3.36767137e-01 2.99745530e-01
5.54387510e-01 6.84897229e-02 -4.85343307e-01 -2.78991759e-01
1.61241472e-01 4.07771111e-01 4.80270013e-02 4.29351121e-01
1.53967291e-01 6.65969968e-01 -1.27600121e+00 -6.42905533e-01
1.73614293e-01 2.55647212e-01 -6.41767502e-01 8.01560819e-01
6.65039429e-03 3.03164333e-01 -5.45956671e-01 1.06646883e+00
8.06509137e-01 1.36392355e-01 -1.66154712e-01 -4.94320095e-01
-3.18428636e-01 -4.30051953e-01 -9.63696182e-01 1.77522445e+00
-8.40574205e-01 6.21205926e-01 1.00669339e-01 -9.52963889e-01
1.06530249e+00 -2.07081214e-02 3.89189601e-01 -1.04837298e+00
9.67001468e-02 3.25932950e-01 -1.42828748e-01 -1.08525550e+00
6.28421366e-01 -3.68108243e-01 1.64317593e-01 1.17385879e-01
1.40165947e-02 -1.12746924e-01 -1.14391848e-01 -4.40572090e-02
8.63009632e-01 3.91564697e-01 3.55678052e-01 -4.30835545e-01
3.80212277e-01 -3.18980038e-01 4.24561709e-01 3.86572093e-01
-4.26666379e-01 7.12335348e-01 1.01309821e-01 -2.65399843e-01
-8.68365228e-01 -1.21791410e+00 -4.26644146e-01 1.02118754e+00
9.65508461e-01 7.10596936e-03 -7.88624227e-01 -2.26402417e-01
-6.76369593e-02 2.35675305e-01 -6.71193898e-01 -3.91040415e-01
1.30351841e-01 -7.10978210e-01 5.37627816e-01 3.76075178e-01
8.38036478e-01 -1.11485517e+00 -8.17180753e-01 5.05665392e-02
-1.26182601e-01 -8.05044293e-01 -2.69740552e-01 -1.60040155e-01
-6.78248107e-01 -8.80722702e-01 -1.05664396e+00 -8.69883716e-01
5.12068212e-01 5.00679314e-01 8.38105798e-01 2.67919272e-01
-7.27719143e-02 5.47616363e-01 -9.18932021e-01 -4.31509793e-01
-8.49788412e-02 -5.41663885e-01 7.20134154e-02 3.11582416e-01
8.03735405e-02 -6.37868345e-01 -1.12003684e+00 3.62261027e-01
-1.34680808e+00 -1.11961409e-01 9.67971683e-01 9.56586599e-01
2.31147021e-01 -1.57481760e-01 7.68386424e-01 -1.37841672e-01
6.36268377e-01 -6.74689591e-01 -3.45070571e-01 2.04746053e-01
-5.29152930e-01 -1.09652646e-01 3.02600175e-01 -3.47023517e-01
-9.70529437e-01 -3.09347481e-01 -4.84860569e-01 -3.14899355e-01
2.98753113e-01 9.89742517e-01 -5.63162506e-01 -3.26971263e-01
3.88512433e-01 4.11891073e-01 1.99276865e-01 -3.99799556e-01
2.00402170e-01 1.11249399e+00 5.39968848e-01 -5.59484065e-01
5.32685161e-01 3.88763696e-01 -2.27954850e-01 -9.99848664e-01
-6.45124316e-01 -5.26679456e-01 -1.93031859e-02 -7.11688161e-01
8.93765509e-01 -1.01081276e+00 -1.00648868e+00 5.98938525e-01
-1.17386806e+00 -1.07539773e-01 -1.02856077e-01 5.16374171e-01
-5.24367750e-01 8.56097937e-01 -4.22122151e-01 -1.29279506e+00
-3.71615022e-01 -1.36603940e+00 1.08740568e+00 6.97741747e-01
5.06896555e-01 -6.84965968e-01 -8.09590369e-02 1.43225372e-01
5.31597674e-01 1.65807188e-01 4.65969086e-01 -1.50981560e-01
-6.41928494e-01 6.30481914e-02 -5.73678315e-01 6.12349510e-01
8.48623961e-02 -1.31795064e-01 -1.10750377e+00 -1.95183948e-01
-1.13448396e-01 -6.76843107e-01 9.65227246e-01 6.44322261e-02
1.06159520e+00 -1.94708481e-01 3.25798869e-01 7.73986757e-01
1.52625418e+00 3.38811576e-01 1.03487039e+00 6.09220445e-01
4.05737877e-01 8.25856090e-01 9.52318013e-01 7.77716577e-01
5.94481170e-01 5.79133868e-01 1.10285199e+00 -1.54163301e-01
1.26591027e-01 -2.99414724e-01 6.22965872e-01 9.16784048e-01
-3.78032118e-01 -3.22213739e-01 -5.31065166e-01 5.67240894e-01
-1.85978961e+00 -7.81954944e-01 8.13632756e-02 2.08627629e+00
6.80287421e-01 -1.74427181e-01 -1.76463470e-01 8.95070359e-02
5.34627676e-01 -1.32704094e-01 -5.40588677e-01 -2.88442552e-01
-1.66447490e-01 -2.41696700e-01 2.06144795e-01 1.92606762e-01
-1.01635063e+00 5.10940611e-01 5.23223209e+00 8.55693698e-01
-9.52465415e-01 1.05613075e-01 4.67810571e-01 4.86207753e-01
-6.77694976e-01 -1.29010901e-01 -2.57535815e-01 4.83766645e-01
4.65207219e-01 -1.51368514e-01 1.62183747e-01 1.10846877e+00
3.23545307e-01 -1.33969421e-02 -9.84647632e-01 1.29006445e+00
9.28501859e-02 -8.20481360e-01 3.39709163e-01 4.42656614e-02
4.43678588e-01 -4.63059902e-01 2.80151755e-01 3.77017587e-01
-4.33828086e-02 -1.03030646e+00 9.81183052e-01 6.00344956e-01
7.56145775e-01 -4.36932892e-01 9.40777004e-01 1.47185072e-01
-1.29711425e+00 -2.14604825e-01 -8.92688751e-01 -3.27777475e-01
8.75237770e-03 4.49037731e-01 -2.93782085e-01 5.80250084e-01
9.82055426e-01 7.19146192e-01 -2.00663567e-01 1.54662383e+00
8.68814066e-02 3.29383969e-01 -6.13962524e-02 -3.50259751e-01
2.52943307e-01 -4.27937120e-01 2.93838233e-01 1.19034362e+00
7.97131002e-01 3.74721259e-01 -8.23179409e-02 8.42138052e-01
-1.98151141e-01 5.69159806e-01 -5.63033044e-01 2.18314484e-01
2.90549338e-01 1.35268092e+00 -4.03832406e-01 8.92855506e-03
-5.42498648e-01 1.05725980e+00 -4.70688343e-02 2.36620731e-03
-6.80904388e-01 -3.51474315e-01 8.23280871e-01 -1.88876480e-01
2.13699546e-02 -1.58862136e-02 -1.58201605e-01 -1.21117353e+00
1.72025505e-02 -5.51365495e-01 1.53573707e-01 -8.84348571e-01
-1.44109380e+00 6.92907155e-01 -1.98592961e-01 -1.95104301e+00
4.00107890e-01 -6.83204770e-01 -7.30591357e-01 5.84877551e-01
-2.02142262e+00 -1.25403261e+00 -8.27743411e-01 3.39558482e-01
7.01237857e-01 2.44446397e-02 7.57544756e-01 5.84831953e-01
-2.64985919e-01 5.97595453e-01 2.71135837e-01 9.93259624e-02
4.68061954e-01 -1.15470707e+00 -1.43532306e-01 7.20968604e-01
-1.97752550e-01 2.85388678e-01 6.95529401e-01 -2.70826995e-01
-1.67183864e+00 -9.20225143e-01 -1.96821205e-02 -2.21238621e-02
5.07645667e-01 1.75677851e-01 -1.00116575e+00 -1.70523394e-02
1.56036407e-01 7.06988573e-03 6.68826938e-01 -4.10706311e-01
-2.22256735e-01 -3.39404285e-01 -1.11947381e+00 4.85474259e-01
1.10526836e+00 -3.00193399e-01 -4.32336420e-01 -5.73548600e-02
6.26049042e-01 -5.30471606e-03 -1.12705719e+00 5.40486634e-01
1.03103089e+00 -1.11232197e+00 9.00182486e-01 -8.62722248e-02
7.34894574e-01 -4.69552517e-01 -5.94864249e-01 -1.26652527e+00
-1.04341954e-01 -2.37704515e-02 2.10926294e-01 1.21341598e+00
1.79327205e-01 -2.86530018e-01 4.99268502e-01 3.49211305e-01
-5.51306367e-01 -6.43491626e-01 -9.55500364e-01 -7.15492904e-01
-4.56803711e-03 -4.25991565e-01 4.99432296e-01 5.92334986e-01
1.47912830e-01 -5.22776842e-02 -6.24810457e-01 3.64043057e-01
6.91665530e-01 2.60879010e-01 6.27724051e-01 -1.35062206e+00
-3.30843300e-01 -6.07741654e-01 -8.79454195e-01 -1.23795521e+00
-3.45384568e-01 -4.66256231e-01 4.85759169e-01 -1.81060004e+00
4.47816074e-01 -2.51308680e-01 -4.86773878e-01 2.33911201e-01
-2.08378375e-01 4.47609514e-01 2.37950355e-01 3.78192037e-01
-8.20423126e-01 1.18376648e+00 1.39675772e+00 -3.51036578e-01
1.82574362e-01 -2.31243983e-01 -7.58263707e-01 7.21257329e-01
5.59810519e-01 -1.84762366e-02 -4.02096391e-01 -5.40953457e-01
1.67867288e-01 2.58681566e-01 2.21889660e-01 -1.14082718e+00
2.20282212e-01 -3.48169684e-01 7.87500814e-02 8.28023851e-02
5.61253965e-01 -8.26072335e-01 -2.44632691e-01 5.87395966e-01
-1.46221891e-01 -1.91231668e-01 -1.67013139e-01 7.68007815e-01
-5.74044585e-01 -3.51020247e-01 9.50189710e-01 -1.92022964e-01
-1.46218801e+00 3.40350211e-01 -1.25251874e-01 -2.45883569e-01
9.00475264e-01 -4.95797426e-01 -2.52607942e-01 -7.26628482e-01
-3.26030612e-01 3.22997898e-01 5.49764514e-01 3.32191855e-01
1.37133861e+00 -1.17940307e+00 -7.98521042e-01 -1.05018549e-01
6.44185364e-01 -9.90922451e-02 4.71643984e-01 6.01060629e-01
-7.76582420e-01 -5.45139909e-01 -5.50496161e-01 -5.66099703e-01
-9.45680499e-01 2.27650598e-01 3.40506613e-01 4.14182842e-01
-2.65425682e-01 7.52451301e-01 6.05175734e-01 -1.43357858e-01
3.50655407e-01 -2.05683753e-01 -6.66599393e-01 -8.92008767e-02
5.94066501e-01 2.77404845e-01 -1.95352584e-01 -6.42972112e-01
-7.92348385e-02 8.08346570e-01 4.89878833e-01 -1.04156002e-01
1.38387656e+00 -4.25663054e-01 -6.14323243e-02 3.59303743e-01
1.20581663e+00 -4.99859303e-01 -1.48720503e+00 -1.16413146e-01
-1.75210565e-01 -6.84975326e-01 1.61830708e-01 -5.93551099e-01
-1.09128726e+00 1.13216019e+00 9.86095965e-01 2.51285344e-01
1.50501716e+00 -9.75708291e-02 8.67523313e-01 4.18317288e-01
5.85981965e-01 -9.86442864e-01 5.40941477e-01 2.69360483e-01
1.02791595e+00 -1.64224041e+00 -1.56499192e-01 -1.35128722e-01
-7.31012106e-01 1.15509188e+00 7.49262273e-01 1.59953117e-01
5.98731577e-01 6.22916073e-02 3.69517326e-01 6.20229132e-02
-2.49891013e-01 -6.40820682e-01 1.76455721e-01 7.87069619e-01
2.37885132e-01 -7.03875208e-03 -4.63916838e-01 9.76290166e-01
1.88082457e-02 -1.12482756e-01 7.03590274e-01 8.78271759e-01
-9.75202322e-01 -7.57950664e-01 -2.96024024e-01 2.12860674e-01
-3.05276096e-01 -8.70510265e-02 1.48279607e-01 3.95917326e-01
1.73798397e-01 9.42650855e-01 -1.73142493e-01 -8.35485458e-01
4.79575753e-01 -5.58598816e-01 2.14427099e-01 -3.10578167e-01
-6.68094009e-02 -2.10997134e-01 -1.33563668e-01 -2.39219099e-01
-8.20444345e-01 -2.89886653e-01 -1.25130808e+00 -9.04389694e-02
-3.53673935e-01 3.10494751e-01 8.96216094e-01 7.29992867e-01
-1.65471002e-01 2.34892458e-01 8.35028052e-01 -1.01115012e+00
-6.37875378e-01 -1.03688014e+00 -8.90610397e-01 5.65431058e-01
1.86887965e-01 -8.53299439e-01 -3.56227487e-01 8.87662545e-02] | [10.702264785766602, -3.5378506183624268] |
d0e87ae7-c19f-4e3e-ae8d-defc87147839 | learning-lightness-from-human-judgement-on | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Narihira_Learning_Lightness_From_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Narihira_Learning_Lightness_From_2015_CVPR_paper.pdf | Learning Lightness From Human Judgement on Relative Reflectance | We develop a new approach to inferring lightness, the perceived reflectance of surfaces, from a single image. Classic methods view this problem from the perspective of intrinsic image decomposition, where an image is separated into reflectance and shading components. Rather than reason about reflectance and shading together, we learn to directly predict lightness differences between pixels. Large-scale training from human judgement data on relative reflectance, and patch representations built using deep networks, provide the foundation for our model. Benchmarked on the Intrinsic Images in the Wild dataset, our local lightness model achieves on-par performance with the state-of-the-art global lightness model, which incorporates multiple shading/reflectance priors and simultaneous reasoning between pairs of pixels in a dense conditional random field formulation. | ['Takuya Narihira', 'Stella X. Yu', 'Michael Maire'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['intrinsic-image-decomposition'] | ['computer-vision'] | [ 1.01311815e+00 8.75925869e-02 3.49117905e-01 -8.66308749e-01
-6.52106941e-01 -3.52612823e-01 4.94080305e-01 -3.58714253e-01
1.08094625e-02 2.95953929e-01 2.11380899e-01 -2.29293182e-02
1.88662112e-01 -1.09645998e+00 -8.60421002e-01 -8.44708145e-01
6.61036551e-01 1.57283887e-01 5.84973358e-02 -3.13585848e-02
2.94619977e-01 4.97366995e-01 -1.75141430e+00 6.30195022e-01
6.65756524e-01 1.16396809e+00 2.64613867e-01 9.98947263e-01
-9.50318649e-02 9.22079146e-01 -8.00374225e-02 -4.08884697e-02
4.69650507e-01 -7.52609000e-02 -6.95100307e-01 3.84232759e-01
1.27406740e+00 -7.96512306e-01 -9.11326930e-02 9.17394161e-01
6.69505671e-02 5.62135428e-02 6.11332417e-01 -7.02470481e-01
-1.00579488e+00 5.15970290e-02 -8.49538863e-01 -3.11695665e-01
4.51626152e-01 4.12948906e-01 1.38807571e+00 -7.36439109e-01
1.40107080e-01 1.34498847e+00 7.95459151e-01 3.25123757e-01
-1.83403182e+00 -9.07955766e-02 3.12786341e-01 -1.92415655e-01
-1.08426201e+00 -5.99547505e-01 6.90238476e-01 -5.66987276e-01
1.15659499e+00 3.13142270e-01 5.04875124e-01 7.10967541e-01
3.11140157e-02 6.50908232e-01 1.50898707e+00 -5.35644412e-01
-2.40501370e-02 -2.42985308e-01 8.76580328e-02 7.39320993e-01
-1.53549467e-04 3.59972417e-01 -6.02623463e-01 -1.38435271e-02
1.07332432e+00 1.08983196e-01 -4.95464146e-01 -1.32052034e-01
-1.01890564e+00 3.83611649e-01 6.02371931e-01 -5.57480276e-01
-4.50024933e-01 4.86437976e-01 -4.02034789e-01 -3.25790383e-02
6.84981883e-01 1.88658044e-01 -6.80403173e-01 3.71533185e-01
-7.98088133e-01 -1.14235226e-02 6.34704590e-01 4.83713210e-01
1.55340588e+00 -2.19414920e-01 1.07354522e-01 7.57574320e-01
6.49551928e-01 1.00840068e+00 -4.40464497e-01 -1.55275893e+00
6.93539158e-02 4.67939019e-01 4.52582955e-01 -9.74404514e-01
-2.49238327e-01 1.26583382e-01 -4.76385117e-01 7.24584818e-01
4.36998487e-01 5.65881841e-02 -1.13715899e+00 1.80433953e+00
1.79866493e-01 1.60830662e-01 -2.08732873e-01 9.41482663e-01
6.16729617e-01 6.29337728e-01 -1.52051717e-01 1.89812824e-01
1.09114456e+00 -8.73034656e-01 -1.91675588e-01 -5.81027925e-01
-1.04842328e-01 -9.11734879e-01 1.17291176e+00 6.68331265e-01
-1.27500582e+00 -6.59056008e-01 -8.30230713e-01 -6.94904685e-01
-1.28403246e-01 2.42113858e-01 8.93063247e-01 5.59451044e-01
-1.40423989e+00 4.97386694e-01 -8.42137694e-01 -2.30087176e-01
3.61843139e-01 1.40956238e-01 1.37808779e-02 -3.96844923e-01
-5.83295286e-01 6.87330186e-01 -4.83880311e-01 3.73671204e-01
-9.67429578e-01 -9.51976836e-01 -7.82286882e-01 -2.02472061e-02
-4.02638130e-03 -9.84678149e-01 1.08423781e+00 -1.43251896e+00
-1.75688386e+00 1.31235790e+00 -5.47451675e-01 1.03315279e-01
1.29840598e-01 -3.84043723e-01 -1.14574455e-01 1.95904955e-01
-1.69973597e-01 8.12164009e-01 9.67194438e-01 -1.90930903e+00
-2.85164446e-01 -4.49893266e-01 4.22948122e-01 3.41623425e-01
3.54348421e-01 -1.39775425e-01 -2.68219978e-01 1.13047741e-01
2.93553740e-01 -5.71086168e-01 -1.25314137e-02 4.80426699e-01
-3.96841705e-01 1.14353582e-01 2.46532917e-01 -6.85592651e-01
4.14609492e-01 -1.99888241e+00 -1.97541947e-03 1.82009175e-01
3.79039764e-01 -2.73235232e-01 -3.71891081e-01 9.89598259e-02
-1.99685305e-01 -2.37749308e-01 -2.97210425e-01 -4.08187002e-01
2.46814024e-02 2.01389670e-01 -5.69977045e-01 6.86249197e-01
3.30929786e-01 7.70419896e-01 -8.23413253e-01 -2.95852795e-02
5.00397861e-01 9.46640909e-01 -5.18201113e-01 3.85906219e-01
-5.40710032e-01 2.83818185e-01 -1.12248212e-01 5.06743550e-01
1.13234246e+00 -3.14165294e-01 1.49831101e-01 -6.12377524e-01
-1.31603524e-01 3.32168162e-01 -9.70388174e-01 1.68764246e+00
-8.01062584e-01 6.99463546e-01 3.88716817e-01 -3.96752328e-01
8.83967996e-01 1.17656505e-02 5.13921142e-01 -6.69909477e-01
-7.82236978e-02 -1.68934748e-01 -4.74736035e-01 -2.70937741e-01
3.95996571e-01 -2.75733531e-01 6.22949541e-01 6.55896842e-01
-2.20691100e-01 -5.84127843e-01 -4.78568166e-01 -6.68240413e-02
8.98303926e-01 8.74326408e-01 -2.02931106e-01 -2.12066367e-01
4.02922422e-01 -4.47639316e-01 3.35488528e-01 5.56708157e-01
8.05759653e-02 1.06011868e+00 1.79617241e-01 -6.87358379e-01
-8.96982133e-01 -1.67633271e+00 -1.19733825e-01 1.23872602e+00
3.24529439e-01 -8.23971108e-02 -5.76819062e-01 1.74617823e-02
2.03432277e-01 7.05048442e-01 -8.56173098e-01 3.46672624e-01
-1.33004248e-01 -7.06883371e-01 6.86807334e-02 4.52829063e-01
5.38855910e-01 -8.41751814e-01 -7.35839605e-01 -3.09552580e-01
-1.42883033e-01 -1.16811907e+00 -2.28055060e-01 4.65259291e-02
-6.40437543e-01 -1.19159925e+00 -2.57196426e-01 -3.79838318e-01
8.14560711e-01 8.18409562e-01 1.81960821e+00 2.89975911e-01
-7.09058225e-01 8.40920210e-01 1.69167772e-01 -2.22381264e-01
-6.33534044e-02 -7.09744692e-01 -4.58426565e-01 2.64086396e-01
4.48826671e-01 -6.22451067e-01 -1.06441832e+00 3.79255354e-01
-8.55640769e-01 4.73315209e-01 3.97027463e-01 5.73752165e-01
8.37244809e-01 -9.85192358e-02 -3.34854007e-01 -1.04889822e+00
-3.54932509e-02 -8.00072029e-02 -6.90800726e-01 4.73442018e-01
-4.93835300e-01 -1.15543038e-01 2.35552073e-01 9.42379460e-02
-1.84792340e+00 1.06637314e-01 1.14791669e-01 -1.57617852e-01
-5.81849754e-01 -1.49419054e-01 -2.76593119e-01 -1.88091427e-01
8.48576248e-01 -4.58998010e-02 -1.14456445e-01 -3.53654534e-01
8.33489954e-01 3.20137173e-01 6.69758677e-01 -8.59640181e-01
6.27546728e-01 1.12533128e+00 1.84871957e-01 -8.57261896e-01
-1.27212286e+00 -4.96942729e-01 -8.65093708e-01 -4.10203747e-02
1.10038471e+00 -1.25380802e+00 -1.00863349e+00 6.91894829e-01
-1.15692306e+00 -1.03202629e+00 -3.81064147e-01 -5.76670282e-03
-8.53070378e-01 3.88031185e-01 -5.48007190e-01 -9.43622112e-01
7.77157024e-02 -8.48150492e-01 1.74188721e+00 2.09191442e-01
2.13115409e-01 -1.25273764e+00 1.46702096e-01 6.59233928e-01
4.57024276e-01 1.80050135e-01 7.15554535e-01 7.84033060e-01
-1.10969400e+00 2.43741527e-01 -9.75223899e-01 6.06383085e-01
3.92849684e-01 3.83682668e-01 -1.65672255e+00 7.14440346e-02
2.13871479e-01 -5.55266201e-01 1.34801280e+00 7.31184602e-01
1.32309592e+00 1.11407503e-01 7.18819648e-02 1.06902575e+00
1.98168385e+00 -5.72196841e-01 9.71076131e-01 -8.27714726e-02
1.06074858e+00 8.63070667e-01 1.16052993e-01 4.19074953e-01
6.80396855e-01 1.91721126e-01 6.17253184e-01 -5.62477946e-01
-5.55825770e-01 -6.16075806e-02 2.58384496e-01 1.88369572e-01
-2.25831166e-01 -1.83333844e-01 -7.48818099e-01 2.68525213e-01
-1.50851154e+00 -1.01392293e+00 -3.45150143e-01 2.16493034e+00
8.88831317e-01 -2.66210556e-01 -3.27497900e-01 -2.92148411e-01
4.55107063e-01 3.65509957e-01 -7.06415117e-01 -3.13843220e-01
-4.98588711e-01 2.21445903e-01 7.20735073e-01 1.26198447e+00
-7.03753769e-01 9.36467528e-01 7.78507519e+00 1.70309484e-01
-1.02239335e+00 -1.29925504e-01 1.03473163e+00 6.61714673e-02
-1.04252720e+00 3.23775679e-01 -6.53636396e-01 -1.49588943e-01
5.97768843e-01 7.77865887e-01 1.24996090e+00 1.86890349e-01
2.65932590e-01 -7.24331677e-01 -1.21700215e+00 9.84450996e-01
3.73369038e-01 -8.91888559e-01 -1.60383433e-01 -9.25759003e-02
9.94665921e-01 5.13886213e-01 3.27861339e-01 -4.82285649e-01
9.20124471e-01 -1.22722208e+00 4.50932473e-01 1.15450621e+00
8.62457871e-01 -1.75272822e-02 5.32276519e-02 -4.35499176e-02
-9.00325119e-01 1.22085802e-01 -6.56086087e-01 -4.01008040e-01
1.17974527e-01 9.20069396e-01 -4.42674249e-01 2.09456891e-01
7.97149003e-01 9.19011593e-01 -5.09059250e-01 3.53251010e-01
-5.93971729e-01 4.34043735e-01 -4.84874606e-01 7.41614640e-01
-3.87353897e-01 -6.54873848e-01 3.33819946e-04 9.11951125e-01
-2.19627172e-01 2.17387453e-01 6.31755069e-02 1.60420454e+00
1.18964165e-01 -3.95390660e-01 -3.88438106e-01 3.14075142e-01
-4.51681092e-02 1.53400528e+00 -4.05396998e-01 -7.85382316e-02
-8.16471338e-01 1.04057348e+00 3.53453785e-01 9.15496826e-01
-5.40566266e-01 1.22330144e-01 9.92894888e-01 1.59514472e-01
1.95821345e-01 -3.47856313e-01 -8.86525869e-01 -1.15362120e+00
-1.59851998e-01 -3.73714626e-01 -2.38568872e-01 -1.65637469e+00
-1.72363162e+00 1.53336182e-01 -2.92687476e-01 -6.95853829e-01
4.05756623e-01 -1.11354291e+00 -6.99943066e-01 1.42466629e+00
-2.22911835e+00 -1.43283534e+00 -8.87935400e-01 7.29402840e-01
1.80961177e-01 5.30345201e-01 9.17790294e-01 -2.96626091e-01
-2.19492406e-01 -3.82961258e-02 1.72172990e-02 -1.08427428e-01
8.18922758e-01 -1.56083977e+00 5.95566332e-01 7.16732204e-01
1.37056261e-01 8.02554250e-01 5.71240664e-01 -3.30528647e-01
-1.58013153e+00 -7.19412923e-01 3.83438259e-01 -8.97874355e-01
5.28525770e-01 -2.71711051e-01 -6.58995688e-01 7.14512885e-01
3.56046379e-01 1.47285998e-01 7.06400573e-01 4.89346176e-01
-9.04802740e-01 -3.79190147e-01 -1.05794907e+00 4.90160316e-01
9.96296346e-01 -1.08856106e+00 -3.09951305e-01 5.54580867e-01
5.38108170e-01 -2.70305216e-01 -9.04668331e-01 1.44755006e-01
7.01698124e-01 -1.58127451e+00 1.27532053e+00 -1.40218526e-01
5.63740313e-01 -3.56247753e-01 -7.50078797e-01 -1.09328496e+00
-5.01402080e-01 -4.68147188e-01 2.91636348e-01 9.68120754e-01
3.30480456e-01 -5.62148452e-01 7.54456639e-01 1.11563802e+00
2.27574483e-02 -2.54812777e-01 -2.84320295e-01 -1.55047998e-01
-1.71948913e-02 -4.48919684e-01 5.26550472e-01 4.79963988e-01
-6.52127922e-01 3.19121480e-01 -2.37068385e-01 4.41644847e-01
1.00311124e+00 5.63919246e-01 8.50395739e-01 -1.41941094e+00
-7.10977972e-01 -4.57999438e-01 3.57315131e-02 -1.50585473e+00
3.16888750e-01 -5.38808584e-01 6.38585150e-01 -1.89676273e+00
4.42757159e-01 -3.99617881e-01 -1.36757851e-01 5.10546744e-01
-3.50992560e-01 4.25617963e-01 -8.82809833e-02 1.36397794e-01
-4.92978036e-01 4.82576877e-01 1.27986228e+00 -2.24829614e-01
-1.13731153e-01 -2.46394932e-01 -8.13767552e-01 8.78656030e-01
5.71177125e-01 2.06553653e-01 -5.08519650e-01 -1.18818450e+00
8.01786900e-01 -1.83124796e-01 6.53418660e-01 -6.91644430e-01
2.93156113e-02 -5.83167553e-01 6.83815241e-01 -3.91289592e-01
6.73190892e-01 -7.58578360e-01 -3.18294279e-02 -2.83937633e-01
-3.51803094e-01 -3.49833935e-01 -3.02355941e-02 6.38102651e-01
1.45650700e-01 3.03033233e-01 7.95904875e-01 -2.36903444e-01
-6.59904897e-01 3.04653466e-01 1.38151988e-01 -3.92047688e-02
2.69633472e-01 -3.84466171e-01 -7.70125508e-01 -2.52104759e-01
-4.98875707e-01 -1.11817732e-01 1.02285385e+00 4.11225930e-02
6.15707874e-01 -8.56821418e-01 -6.52728796e-01 3.88692230e-01
4.51397039e-02 2.50249028e-01 4.54394758e-01 5.80780864e-01
-7.21501410e-01 -9.51641053e-02 -1.16131395e-01 -9.57295477e-01
-8.95350158e-01 3.70310992e-01 7.37493694e-01 3.45775694e-01
-5.09069383e-01 1.10141003e+00 1.06593513e+00 -4.94755834e-01
-1.83525458e-01 -5.75753927e-01 4.35054749e-01 -5.87571383e-01
4.73784387e-01 2.82755971e-01 -1.03802107e-01 -5.55319071e-01
-1.42974496e-01 1.12044621e+00 4.12905663e-01 -1.91208258e-01
1.38375998e+00 -6.79851890e-01 -6.08171463e-01 7.02845693e-01
1.06237888e+00 1.89915910e-01 -2.17501450e+00 -3.20100874e-01
-7.43650496e-01 -9.75464344e-01 5.29409468e-01 -1.15214360e+00
-1.09848571e+00 1.09047830e+00 4.23169464e-01 -6.40838966e-03
1.41246545e+00 -9.43468735e-02 4.50456470e-01 3.43899846e-01
1.60512611e-01 -1.04061210e+00 9.73659828e-02 4.03126121e-01
7.14628398e-01 -1.69457507e+00 3.83528352e-01 -6.70251906e-01
-5.45646489e-01 1.16336060e+00 5.38907707e-01 -3.38919938e-01
9.21609819e-01 4.28897470e-01 4.51169968e-01 -3.24605942e-01
-8.43293130e-01 -2.20280230e-01 4.21464890e-01 9.05407250e-01
7.30170369e-01 1.85444683e-01 6.30917430e-01 -1.22099735e-01
1.37442425e-01 -1.37526020e-01 3.88739765e-01 3.46483767e-01
-6.70344174e-01 -8.46771896e-01 -3.27374816e-01 1.77967533e-01
2.67645437e-02 -4.24371690e-01 -3.34852546e-01 -4.26504128e-02
2.40883172e-01 1.16826010e+00 2.93741673e-01 -9.36516300e-02
1.86657123e-02 -2.07058236e-01 9.72128987e-01 -7.53807664e-01
-1.73736587e-01 1.95976391e-01 -1.21895872e-01 -1.04895401e+00
-8.66187513e-01 -6.90782487e-01 -9.62965012e-01 -2.72228211e-01
-8.97142515e-02 -7.07786500e-01 6.56281471e-01 7.99764693e-01
9.65709388e-02 3.54940206e-01 7.67198622e-01 -1.20855653e+00
-1.79033190e-01 -6.83658421e-01 -9.26291287e-01 6.03611231e-01
6.91513777e-01 -4.10098255e-01 -5.27460694e-01 3.98514897e-01] | [9.861172676086426, -2.9681179523468018] |
6a592448-4e97-4310-b332-6de73646d882 | eegminer-discovering-interpretable-features | 2110.10009 | null | https://arxiv.org/abs/2110.10009v2 | https://arxiv.org/pdf/2110.10009v2.pdf | EEGminer: Discovering Interpretable Features of Brain Activity with Learnable Filters | Patterns of brain activity are associated with different brain processes and can be used to identify different brain states and make behavioral predictions. However, the relevant features are not readily apparent and accessible. To mine informative latent representations from multichannel recordings of ongoing EEG activity, we propose a novel differentiable decoding pipeline consisting of learnable filters and a pre-determined feature extraction module. Specifically, we introduce filters parameterized by generalized Gaussian functions that offer a smooth derivative for stable end-to-end model training and allow for learning interpretable features. For the feature module, we use signal magnitude and functional connectivity estimates. We demonstrate the utility of our model towards emotion recognition from EEG signals on the SEED dataset, as well as on a new EEG dataset of unprecedented size (i.e., 761 subjects), where we identify consistent trends of music perception and related individual differences. The discovered features align with previous neuroscience studies and offer new insights, such as marked differences in the functional connectivity profile between left and right temporal areas during music listening. This agrees with the respective specialisation of the temporal lobes regarding music perception proposed in the literature. | ['Stefanos Zafeiriou', 'Yannis Panagakis', 'Nikolaos Laskaris', 'Dimitrios A. Adamos', 'Stylianos Bakas', 'Siegfried Ludwig'] | 2021-10-19 | null | null | null | null | ['eeg-decoding', 'eeg-decoding'] | ['medical', 'time-series'] | [ 4.35327679e-01 9.85038802e-02 1.10966310e-01 -5.40536284e-01
-6.04051530e-01 -7.35230565e-01 7.12990761e-01 1.58603072e-01
-5.41436136e-01 5.40002048e-01 6.26360774e-01 3.18941295e-01
-7.68380165e-01 -2.59571642e-01 -4.01173294e-01 -5.89479566e-01
-5.94297826e-01 1.58038258e-03 -3.49599302e-01 1.83292821e-01
4.55641478e-01 3.35501581e-01 -1.44350731e+00 5.27415156e-01
7.12475836e-01 1.16394126e+00 2.61401951e-01 2.68475950e-01
4.13999110e-01 1.80705890e-01 -4.36336011e-01 -1.95297942e-01
-8.09506401e-02 -5.02385259e-01 -5.55976927e-01 -2.21873317e-02
3.07740033e-01 1.50028452e-01 -3.68653327e-01 7.89985955e-01
6.89506829e-01 2.26866156e-01 6.53834641e-01 -9.06494260e-01
-3.06593359e-01 7.65353858e-01 -2.63886929e-01 7.32147753e-01
2.63196766e-01 2.03459397e-01 1.23157179e+00 -8.15332294e-01
5.80043435e-01 9.08615887e-01 5.43191314e-01 3.45253438e-01
-1.60716522e+00 -8.02393258e-01 3.10022682e-01 3.18042308e-01
-1.29371667e+00 -8.08624268e-01 6.41567409e-01 -6.50474906e-01
9.36360419e-01 2.40690604e-01 1.10416615e+00 1.52255523e+00
4.04667407e-01 4.37960625e-01 1.44585347e+00 1.15933232e-01
2.00431958e-01 1.42170920e-03 6.70255572e-02 2.20155731e-01
-2.79190484e-02 -1.14168867e-01 -1.13513267e+00 -2.65856683e-01
6.14071906e-01 -6.81559816e-02 -5.14935672e-01 4.71067354e-02
-1.74553692e+00 4.52830970e-01 3.24195355e-01 4.25148398e-01
-8.48485291e-01 2.13845477e-01 4.12280232e-01 1.84702963e-01
4.72816467e-01 6.49999619e-01 -6.77095234e-01 -5.04141748e-01
-1.37587821e+00 -2.12563246e-01 4.23118383e-01 1.74058482e-01
4.88585055e-01 1.56896457e-01 -4.16806966e-01 8.43591273e-01
3.09963614e-01 2.61880457e-02 6.40255928e-01 -7.78582692e-01
1.14775315e-01 3.32157671e-01 -4.19879764e-01 -1.04965270e+00
-7.00823128e-01 -8.04707527e-01 -7.46730804e-01 -3.05704353e-03
3.35186213e-01 -8.95556994e-03 -2.01046675e-01 2.07415557e+00
-2.56891310e-01 4.37428027e-01 -3.70367587e-01 9.93487716e-01
4.05343950e-01 4.30059060e-02 1.17248647e-01 -3.85898560e-01
1.56116879e+00 -2.29303405e-01 -3.81943703e-01 -4.25354540e-01
-3.17542143e-02 -3.25888097e-01 9.85548556e-01 8.26964378e-01
-1.05923975e+00 -5.65087020e-01 -8.98125947e-01 1.74831226e-01
-2.13771135e-01 3.04931104e-01 9.42333221e-01 5.13093174e-01
-9.48752522e-01 9.28003371e-01 -1.00383329e+00 -3.02502483e-01
7.37187803e-01 6.19450450e-01 -4.37536269e-01 5.51127315e-01
-8.77730072e-01 7.45283544e-01 3.67877215e-01 4.99522954e-01
-1.04921734e+00 -8.60470355e-01 -3.71746331e-01 2.13221699e-01
-1.07426755e-01 -6.73029065e-01 6.12094283e-01 -1.13379884e+00
-1.61066949e+00 8.19357455e-01 -1.70229301e-01 -2.52224267e-01
2.85192002e-02 -1.70419857e-01 -4.34370100e-01 3.31962109e-01
-1.57774180e-01 6.23591065e-01 9.47655559e-01 -4.87983733e-01
-5.47985174e-02 -5.74779212e-01 -3.83592755e-01 1.17327169e-01
-4.08957899e-01 2.31994629e-01 1.32073566e-01 -9.00564849e-01
3.48636597e-01 -6.63849294e-01 2.60923505e-02 -9.10768509e-02
-2.05133066e-01 8.86362940e-02 6.56610653e-02 -8.49171698e-01
1.07104659e+00 -2.43535876e+00 4.13186312e-01 5.18282950e-01
5.59193075e-01 -5.92054844e-01 -2.36139670e-01 1.73332095e-01
-5.23199499e-01 -4.86432947e-02 -3.77717674e-01 1.16478734e-01
3.72774750e-01 -2.81091601e-01 -2.62589037e-01 9.09011006e-01
4.35868084e-01 1.19867325e+00 -7.39136100e-01 2.41192147e-01
-1.79882973e-01 5.73744655e-01 -6.26272857e-01 3.47838514e-02
-3.72262262e-02 8.41473997e-01 -1.46648407e-01 4.75080907e-01
2.51106471e-01 -4.55529206e-02 1.92207605e-01 -2.77633190e-01
-2.06771776e-01 9.12489235e-01 -8.78476024e-01 2.12740517e+00
-2.55536556e-01 1.11177933e+00 1.77570894e-01 -1.18720877e+00
8.60484898e-01 5.65899968e-01 4.77814823e-01 -4.80636835e-01
5.18737257e-01 4.85496894e-02 6.56330526e-01 -3.96229208e-01
-1.85552731e-01 -1.35509253e-01 3.17181610e-02 4.99017268e-01
5.72039604e-01 1.43146247e-01 -3.67672183e-02 -1.59556419e-01
1.18834090e+00 7.88447335e-02 1.95285663e-01 -5.76029301e-01
2.68361717e-01 -5.45847237e-01 3.92375529e-01 5.26548326e-01
-5.40400622e-03 4.42770004e-01 8.15252364e-01 3.34852226e-02
-4.08996552e-01 -1.12327731e+00 -3.90443176e-01 1.16981292e+00
-3.29387575e-01 -6.75261438e-01 -5.02335429e-01 -1.22098073e-01
-2.21889898e-01 5.21051347e-01 -8.19238126e-01 -5.47218144e-01
-1.37644306e-01 -8.71991038e-01 4.67403650e-01 3.54157060e-01
7.76018873e-02 -1.20820034e+00 -7.33270347e-01 2.91523308e-01
-2.38829833e-02 -1.01802754e+00 -1.34402424e-01 5.54110408e-01
-9.80108321e-01 -8.91479075e-01 -3.79085541e-01 -4.17988569e-01
4.04109925e-01 -2.32854351e-01 1.03874493e+00 -4.23625618e-01
-5.90686142e-01 6.01567209e-01 -2.23757327e-03 -2.82098949e-01
2.30831161e-01 4.12122980e-02 3.41304541e-01 1.89990252e-01
2.61125475e-01 -1.25528145e+00 -8.94644797e-01 7.28283599e-02
-6.50484741e-01 -8.59206095e-02 8.35759342e-01 7.47238934e-01
4.96677190e-01 -2.57272035e-01 8.83238614e-01 -2.70932853e-01
9.58175242e-01 -6.83125615e-01 -2.47925594e-01 -1.26304096e-02
-5.93621373e-01 -5.87973036e-02 3.31227362e-01 -5.56287348e-01
-7.38845408e-01 -8.61785263e-02 1.38933733e-01 -8.33470374e-02
-5.18913925e-01 6.41883314e-01 -9.56184119e-02 2.33376883e-02
5.67923725e-01 2.27361679e-01 -1.07960649e-01 -4.26895529e-01
2.77744383e-01 5.66686332e-01 7.04178512e-01 -5.71177125e-01
3.24983060e-01 3.76993746e-01 -1.98818341e-01 -8.67022455e-01
-6.22145534e-01 -3.13230664e-01 -1.00397253e+00 -2.24048778e-01
8.83060634e-01 -9.55738902e-01 -9.74094808e-01 1.49163336e-01
-8.54160726e-01 -3.50051135e-01 -3.29553276e-01 9.87297058e-01
-7.64536738e-01 9.23578888e-02 -5.52882671e-01 -7.02562392e-01
-5.00582635e-01 -8.71525466e-01 1.03079534e+00 -8.90542641e-02
-8.11358511e-01 -7.31550336e-01 2.95668602e-01 -6.08672686e-02
2.51496732e-01 1.44219488e-01 9.80033398e-01 -6.05874062e-01
-2.85122603e-01 4.66540530e-02 3.37968804e-02 1.21724434e-01
-2.92285774e-02 -3.08697373e-01 -1.35384190e+00 -1.31726950e-01
9.73478612e-03 -2.82739580e-01 9.99774396e-01 4.72988129e-01
1.26106966e+00 -2.60115433e-02 -1.16181850e-01 7.22031236e-01
8.70631516e-01 -1.52436152e-01 5.43525755e-01 8.51924419e-02
1.36370271e-01 6.49675667e-01 -1.75627947e-01 3.78957301e-01
-1.06448699e-02 5.32756865e-01 5.24647720e-02 2.67541617e-01
1.71889231e-01 -1.43032297e-02 6.92358851e-01 7.66512215e-01
-1.72967985e-01 4.79448736e-01 -5.69410205e-01 3.15230191e-01
-1.60502756e+00 -1.09159946e+00 1.22476116e-01 2.34230256e+00
6.75862730e-01 9.24883932e-02 2.15617627e-01 4.60503362e-02
2.73926198e-01 6.28141835e-02 -7.08417535e-01 -1.78843856e-01
-1.81748182e-01 7.07267880e-01 -1.06991515e-01 -5.04045859e-02
-8.47009957e-01 7.76550174e-01 7.16651249e+00 2.97208518e-01
-1.35776627e+00 2.18195185e-01 3.80279213e-01 -6.68652236e-01
-2.94055551e-01 -1.42184511e-01 -4.60759290e-02 2.56391943e-01
1.30319321e+00 -1.67368174e-01 1.01958311e+00 2.40004748e-01
5.51608682e-01 4.68647182e-02 -1.15461767e+00 1.17838728e+00
5.11669330e-02 -1.03290355e+00 -5.35143375e-01 3.63956168e-02
2.08795398e-01 3.72478396e-01 1.80406809e-01 1.48205355e-01
-4.54339087e-01 -1.20522058e+00 8.26185346e-01 1.13978100e+00
7.83472657e-01 -5.01942515e-01 1.18140243e-01 2.64818579e-01
-9.33551788e-01 -2.83375412e-01 -3.35327893e-01 -2.73618430e-01
-1.43802806e-03 6.41408026e-01 -5.41140497e-01 1.46848589e-01
6.55377388e-01 1.16223717e+00 -7.18376040e-01 1.20138299e+00
-2.83609450e-01 8.11548412e-01 -2.28586599e-01 9.56226066e-02
-1.82562023e-01 -3.02592337e-01 6.65881157e-01 1.17327023e+00
4.92329359e-01 1.58822183e-02 -3.53246897e-01 1.46745038e+00
1.01753771e-01 1.57505348e-01 -2.31482908e-01 -4.87427920e-01
8.27275515e-02 1.76712608e+00 -1.01849139e+00 -7.18586966e-02
-2.38138676e-01 1.15466940e+00 4.28333074e-01 5.26170790e-01
-5.90376556e-01 -2.10531086e-01 7.93577850e-01 -1.86532527e-01
6.40912428e-02 -5.42014062e-01 -2.68885016e-01 -1.41486454e+00
2.82150004e-02 -8.34613681e-01 2.00101703e-01 -7.85576165e-01
-1.36693692e+00 6.38854682e-01 -1.08211443e-01 -8.68021011e-01
-1.57985613e-01 -6.51472688e-01 -6.15497172e-01 1.10587466e+00
-1.12423801e+00 -6.47684753e-01 -2.85518706e-01 6.21832252e-01
2.83010066e-01 -9.21951085e-02 1.14576030e+00 2.84410387e-01
-6.36152744e-01 2.44443372e-01 -5.38394786e-02 -1.91484645e-01
6.93251312e-01 -1.06190014e+00 5.61139062e-02 5.43709695e-01
8.04838240e-01 9.66767728e-01 4.46867466e-01 -3.65679890e-01
-1.28551030e+00 -5.54681301e-01 3.51830006e-01 -5.00849545e-01
9.38085675e-01 -8.99023354e-01 -7.46977806e-01 5.21735072e-01
2.16204062e-01 -1.88838214e-01 1.14833140e+00 5.12213230e-01
-3.20307165e-01 4.59468784e-03 -7.44606853e-01 3.50244254e-01
1.27849746e+00 -8.23416531e-01 -8.17821205e-01 1.58523917e-01
1.58120901e-03 1.20409414e-01 -9.61093366e-01 1.40794799e-01
1.00608516e+00 -8.19516718e-01 8.96581769e-01 -7.02070236e-01
1.16762809e-01 2.34257430e-02 -6.60738796e-02 -1.56861484e+00
-8.03171635e-01 -6.01972461e-01 -1.26348689e-01 1.01118982e+00
5.25384784e-01 -5.72835147e-01 3.66671294e-01 5.97484052e-01
-1.17648728e-01 -6.33963108e-01 -9.98380542e-01 -4.20012385e-01
-2.06946284e-01 -7.11287618e-01 1.68284968e-01 8.79220426e-01
6.10752404e-01 2.72070706e-01 -4.26862910e-02 4.35502194e-02
2.56497324e-01 1.93224877e-01 1.05112083e-01 -1.56802487e+00
-5.34375429e-01 -8.50549579e-01 -6.33536458e-01 -6.78613245e-01
6.11859739e-01 -1.50172746e+00 -1.57665402e-01 -1.35216880e+00
4.23875630e-01 9.47165489e-02 -9.03286815e-01 6.19574249e-01
6.94521144e-02 3.50221068e-01 -1.45338640e-01 2.48101521e-02
-3.33220750e-01 6.20033562e-01 6.80488646e-01 -7.87846837e-03
-2.62351930e-01 -9.86895636e-02 -1.09954548e+00 6.73813164e-01
9.29017425e-01 -4.13275272e-01 -3.41348022e-01 -8.60083625e-02
4.07659918e-01 -2.77860045e-01 5.94111800e-01 -1.06959713e+00
-6.96455389e-02 2.30607584e-01 8.95981371e-01 2.84655541e-02
4.73963201e-01 -6.25696480e-01 5.77716157e-02 3.02496374e-01
-5.42491734e-01 -9.72229168e-02 2.20917821e-01 6.10296905e-01
-2.50280444e-02 1.10531107e-01 3.79011959e-01 2.28529759e-02
-6.07101083e-01 2.30924264e-01 -7.12391198e-01 -1.17863782e-01
6.80479407e-01 -2.65758336e-01 -1.85368508e-01 -3.08184206e-01
-1.07777691e+00 -2.15207070e-01 -2.42833376e-01 5.26106238e-01
5.34622967e-01 -1.18428981e+00 -7.46899307e-01 5.00943959e-01
-3.22310477e-02 -8.56449425e-01 3.11777264e-01 1.55690980e+00
2.67162144e-01 2.91888714e-01 -5.65007269e-01 -5.96984982e-01
-9.19497073e-01 2.33846977e-01 3.67749512e-01 3.97580564e-01
-7.47669518e-01 8.84830058e-01 1.14129938e-01 -9.77657884e-02
8.31606332e-03 -5.00465870e-01 -4.11114782e-01 7.40132868e-01
5.58600008e-01 1.21863060e-01 1.11396618e-01 -5.16263247e-01
-3.91035825e-01 1.28959849e-01 3.14871788e-01 -3.18864733e-01
1.89032066e+00 -4.90844576e-03 -5.44880092e-01 9.81078982e-01
8.82410109e-01 2.47785091e-01 -1.19392478e+00 7.09558204e-02
1.47697642e-01 -3.44443798e-01 1.43727884e-01 -8.85570705e-01
-1.10175323e+00 1.01173043e+00 7.66751111e-01 -2.05327775e-02
1.16790652e+00 1.92612857e-01 2.33380601e-01 2.27620974e-01
3.63754183e-01 -9.97999609e-01 -1.06257655e-01 1.27504945e-01
1.05925345e+00 -7.28036880e-01 2.95194774e-03 1.60222977e-01
-4.42767590e-01 1.22390163e+00 2.59559065e-01 -2.17973098e-01
8.21737945e-01 2.90009584e-02 -2.75362223e-01 -5.08819520e-01
-1.00715566e+00 -1.58783063e-01 8.65871489e-01 2.97042817e-01
8.37148011e-01 3.04117091e-02 -4.83642846e-01 1.33408165e+00
-5.95537305e-01 -1.90229774e-01 1.61530390e-01 2.68963188e-01
-3.52777690e-01 -8.11054349e-01 -1.33655876e-01 9.56546426e-01
-5.48538804e-01 -2.78230876e-01 -4.13902134e-01 3.32698822e-01
3.76116075e-02 8.15359414e-01 1.67706147e-01 -2.97456890e-01
2.37530336e-01 7.42809951e-01 7.02608764e-01 -7.13468075e-01
-6.10746980e-01 2.85557151e-01 1.52243720e-02 -8.70726109e-01
-4.72358674e-01 -1.06926942e+00 -1.06883168e+00 4.00285572e-01
-6.41598627e-02 -3.20755355e-02 5.66846490e-01 9.95031059e-01
5.96968055e-01 7.82056689e-01 2.37897068e-01 -1.03408217e+00
-1.68335155e-01 -1.25960803e+00 -9.15211678e-01 3.35852832e-01
1.48527011e-01 -7.08248317e-01 -3.11675578e-01 1.26598001e-01] | [12.886730194091797, 3.4430789947509766] |
428d797f-d045-4f45-9d53-80a859dfbf88 | learning-jpeg-compression-artifacts-for-image | 2108.12947 | null | https://arxiv.org/abs/2108.12947v2 | https://arxiv.org/pdf/2108.12947v2.pdf | Learning JPEG Compression Artifacts for Image Manipulation Detection and Localization | Detecting and localizing image manipulation are necessary to counter malicious use of image editing techniques. Accordingly, it is essential to distinguish between authentic and tampered regions by analyzing intrinsic statistics in an image. We focus on JPEG compression artifacts left during image acquisition and editing. We propose a convolutional neural network (CNN) that uses discrete cosine transform (DCT) coefficients, where compression artifacts remain, to localize image manipulation. Standard CNNs cannot learn the distribution of DCT coefficients because the convolution throws away the spatial coordinates, which are essential for DCT coefficients. We illustrate how to design and train a neural network that can learn the distribution of DCT coefficients. Furthermore, we introduce Compression Artifact Tracing Network (CAT-Net) that jointly uses image acquisition artifacts and compression artifacts. It significantly outperforms traditional and deep neural network-based methods in detecting and localizing tampered regions. | ['Changick Kim', 'Heung-Kyu Lee', 'In-Jae Yu', 'Seung-Hun Nam', 'Myung-Joon Kwon'] | 2021-08-30 | null | null | null | null | ['image-manipulation-detection'] | ['computer-vision'] | [ 6.04035914e-01 -6.54309392e-01 -2.12680712e-01 -1.33651614e-01
-5.66408098e-01 -5.07124722e-01 3.64326805e-01 -1.01779982e-01
-4.67197925e-01 1.46770775e-01 3.30523471e-03 -5.07873356e-01
3.86158288e-01 -6.97877526e-01 -9.80398178e-01 -6.26242638e-01
-2.96010554e-01 -4.33627933e-01 9.26026553e-02 8.32003132e-02
7.47167468e-01 5.48497856e-01 -8.67932975e-01 8.03368270e-01
6.30789161e-01 1.04449201e+00 1.74141496e-01 1.13145137e+00
1.95092857e-01 1.28786111e+00 -1.08415604e+00 -4.50631753e-02
3.45226169e-01 -5.07844806e-01 -5.37469089e-01 -9.17844251e-02
5.64501584e-01 -1.05820203e+00 -1.11093497e+00 1.60907519e+00
2.13030860e-01 -2.87751704e-01 4.95261163e-01 -1.06760192e+00
-1.27977145e+00 6.34312510e-01 -7.57690072e-01 6.68499410e-01
3.10432762e-02 3.37693542e-01 2.33483806e-01 -6.84893012e-01
3.63851041e-01 8.42795730e-01 1.01022184e+00 3.22429985e-01
-7.14723647e-01 -6.46265328e-01 -3.91401768e-01 5.64507723e-01
-1.41187072e+00 -2.27083012e-01 9.82672274e-01 -8.04961920e-02
7.82457829e-01 2.62673676e-01 3.61027449e-01 1.21732748e+00
5.61228812e-01 8.91177893e-01 7.00340211e-01 -4.75706637e-01
6.12581102e-03 -2.45838493e-01 -2.15613604e-01 6.46515489e-01
9.95470360e-02 1.39276505e-01 -4.40047503e-01 7.05460086e-02
1.18278658e+00 3.12321007e-01 -7.35397041e-01 -2.94418950e-02
-1.29869294e+00 6.89091504e-01 4.31601465e-01 3.74338210e-01
-2.17293054e-01 8.77691269e-01 5.12152612e-01 4.51387912e-01
-5.43004535e-02 3.28906685e-01 -2.68927693e-01 -2.15480626e-01
-1.28270578e+00 -2.56649047e-01 3.18900645e-01 9.93327737e-01
4.89234179e-01 2.93608069e-01 -5.23946211e-02 5.72258949e-01
-2.19633654e-02 5.60335875e-01 6.42952204e-01 -1.13890493e+00
5.46102941e-01 1.95822507e-01 -1.67461365e-01 -1.58509052e+00
1.86743781e-01 -2.22174749e-01 -1.26423454e+00 3.43236864e-01
3.52904707e-01 7.85880387e-02 -9.65402961e-01 1.19188511e+00
-4.02282178e-01 5.04732788e-01 -1.83957741e-01 8.45403314e-01
2.90391624e-01 3.94703478e-01 -2.83845514e-01 2.28642941e-01
1.24800622e+00 -8.98209751e-01 -9.26518142e-01 -1.45868003e-01
6.78384900e-01 -7.01796591e-01 9.19257820e-01 5.11800826e-01
-1.05776978e+00 -7.26146340e-01 -1.53435802e+00 -3.82328808e-01
-6.88051105e-01 3.49275500e-01 2.63609916e-01 8.34111154e-01
-8.95573318e-01 9.14607644e-01 -7.95891285e-01 2.65711010e-01
8.55012596e-01 3.70908946e-01 -3.51579010e-01 -2.10423674e-02
-1.40167892e+00 7.16592789e-01 2.27177978e-01 3.47956359e-01
-1.09015107e+00 -6.06582761e-01 -9.24886346e-01 4.46654916e-01
5.27705625e-03 1.74845442e-01 9.57720280e-01 -1.17131996e+00
-1.10276163e+00 6.88336611e-01 1.87933445e-01 -7.74665534e-01
6.89751804e-01 -1.47922337e-01 -6.95082963e-01 6.99693024e-01
-1.30551919e-01 4.10210758e-01 1.46531343e+00 -1.25342190e+00
-3.72471958e-01 -1.17506124e-01 -2.45220631e-01 -4.50227290e-01
-5.37460625e-01 1.58269569e-01 -6.70464516e-01 -1.05046356e+00
-6.02571480e-02 -4.40165341e-01 1.12602301e-01 2.83156008e-01
-5.73876262e-01 4.89863157e-01 1.58740318e+00 -1.23466444e+00
1.36495245e+00 -2.31708813e+00 -5.01541138e-01 3.40533286e-01
4.56164896e-01 5.93654990e-01 -3.55739236e-01 2.59835124e-01
-3.18959653e-01 3.57581526e-01 -4.06264812e-01 -1.26192778e-01
-2.12961927e-01 -1.61443815e-01 -5.65979958e-01 9.56782579e-01
3.28712851e-01 1.00678897e+00 -7.92853832e-01 -4.65184838e-01
5.37706912e-01 7.77331769e-01 -3.40501249e-01 -2.09243685e-01
-8.99016708e-02 2.42592931e-01 -2.38989130e-01 7.43272245e-01
1.34627414e+00 -2.24169433e-01 3.25229406e-01 -5.66628575e-01
6.17436878e-02 -5.91656975e-02 -6.05037630e-01 1.58820367e+00
-4.88043815e-01 1.43472588e+00 2.23313555e-01 -1.08973503e+00
6.09976411e-01 1.76776841e-01 4.51923877e-01 -9.30469811e-01
5.38713396e-01 5.59942573e-02 -2.53991425e-01 -8.08804870e-01
6.46823823e-01 6.69003427e-01 6.46213489e-03 6.62882030e-01
-1.18304960e-01 1.93838984e-01 -3.34001422e-01 2.79520392e-01
1.25637960e+00 -2.14313239e-01 1.07387297e-01 4.15883586e-03
6.10421360e-01 -4.03427005e-01 2.73161918e-01 1.04344594e+00
-4.91797030e-01 9.81685877e-01 7.41697133e-01 -7.91645169e-01
-1.29410183e+00 -8.28171492e-01 1.22546516e-01 7.08273113e-01
4.42129344e-01 -2.84325451e-01 -9.70489681e-01 -8.64923656e-01
-1.00914679e-01 3.07899356e-01 -8.42868030e-01 -4.23474759e-01
-1.07986641e+00 -1.40485466e-01 1.08395505e+00 6.36116087e-01
1.17553425e+00 -1.05562854e+00 -6.78732991e-01 -3.11377961e-02
-4.18400139e-01 -1.32555175e+00 -8.89637709e-01 1.90311015e-01
-5.71810067e-01 -1.41942596e+00 -5.51492095e-01 -9.31809187e-01
6.72295034e-01 4.32319343e-01 7.76903391e-01 7.04290926e-01
-4.85879362e-01 5.50980382e-02 -2.79253870e-01 -1.10431157e-01
-5.59097230e-01 -1.19877264e-01 -2.31551409e-01 -1.98044330e-02
4.93533403e-01 -5.54628909e-01 -8.40620100e-01 1.21588781e-01
-1.27496123e+00 -3.17167550e-01 5.40845573e-01 7.48726487e-01
1.73178554e-01 5.56259871e-01 -1.38667226e-01 -7.43813038e-01
6.64093554e-01 -2.52947003e-01 -4.82782543e-01 3.80289346e-01
-7.39904791e-02 -9.69010592e-02 1.03142536e+00 -6.35941148e-01
-5.77216566e-01 -2.92615425e-02 -2.72589643e-02 -9.94809866e-01
-9.37639996e-02 1.42453238e-01 5.03145978e-02 -6.13752484e-01
6.77630663e-01 6.60196602e-01 1.58136383e-01 -3.12006533e-01
1.74889341e-01 9.21945214e-01 1.09487009e+00 -2.45473757e-01
9.39058542e-01 7.85563409e-01 -2.32848659e-01 -7.71649301e-01
-1.65696144e-01 -1.59000546e-01 -6.32939041e-01 -9.26108435e-02
8.23403776e-01 -7.21934378e-01 -9.70330298e-01 8.96761537e-01
-1.40951979e+00 -3.52572113e-01 1.07402978e-02 1.75646484e-01
-5.46617389e-01 1.07968938e+00 -9.13022339e-01 -4.80917692e-01
-1.49006203e-01 -1.32880402e+00 9.93274271e-01 -3.38454507e-02
8.78965780e-02 -1.01159346e+00 -4.21851307e-01 3.32146063e-02
8.41398180e-01 3.50995511e-01 8.23383510e-01 -3.02099615e-01
-8.20620358e-01 -5.12042642e-01 -5.68185389e-01 5.83756030e-01
2.98359483e-01 3.54594104e-02 -1.04942572e+00 -2.69290626e-01
3.65834057e-01 -1.59830168e-01 1.09065008e+00 5.15979052e-01
2.14948988e+00 -5.17720819e-01 -1.89699367e-01 1.19647264e+00
1.43562508e+00 2.32127637e-01 1.34611583e+00 5.48241317e-01
7.23388910e-01 9.73899812e-02 8.25022254e-03 3.64235282e-01
-2.48178959e-01 3.08870763e-01 6.58459604e-01 -3.08476210e-01
-2.30192438e-01 -2.96794206e-01 3.22733253e-01 3.99540901e-01
2.15058640e-01 -3.14721793e-01 -5.30371308e-01 4.49133903e-01
-1.36205649e+00 -1.21069050e+00 -8.94191768e-03 1.93863571e+00
7.92530775e-01 2.31459644e-02 -4.31311905e-01 2.65139848e-01
1.03577387e+00 4.73260939e-01 -5.57844341e-01 -2.97535598e-01
-2.00895771e-01 3.42286676e-01 1.05264938e+00 4.46416050e-01
-1.43391156e+00 7.82208025e-01 6.71472645e+00 1.03313196e+00
-1.40202570e+00 -1.78720169e-02 7.51041412e-01 4.05577451e-01
9.89008546e-02 -3.39467406e-01 -1.03492707e-01 9.16642725e-01
7.09414363e-01 5.78721344e-01 7.34858215e-01 7.94059277e-01
3.20537776e-01 -1.92140900e-02 -7.92543471e-01 1.12424791e+00
3.40036422e-01 -1.61804235e+00 6.27010837e-02 2.89099757e-02
5.94133615e-01 -1.21781878e-01 6.10260725e-01 -1.15308337e-01
1.31614164e-01 -1.23768353e+00 6.40398860e-01 9.36024338e-02
1.23723209e+00 -5.13222516e-01 8.99432600e-01 -8.54786369e-04
-1.10488701e+00 -2.48744428e-01 -5.15775740e-01 3.97021413e-01
-1.18671529e-01 2.92133212e-01 -4.86430198e-01 -2.72204690e-02
7.93661296e-01 7.87247419e-01 -5.25301397e-01 7.65599847e-01
-1.95477873e-01 4.94211197e-01 -1.36492671e-02 4.69781339e-01
2.77601629e-01 1.08567150e-02 8.09192806e-02 1.31466055e+00
4.75720227e-01 -3.57086152e-01 -2.91637421e-01 1.14824140e+00
-6.16359234e-01 -5.51437438e-01 -6.72887683e-01 -1.96874514e-01
4.87135500e-01 9.49921727e-01 -8.45373750e-01 -2.98512995e-01
-4.05354381e-01 1.55446768e+00 -1.69840436e-02 5.87231934e-01
-1.03090703e+00 -8.64878356e-01 7.36705840e-01 -1.66034371e-01
6.84771955e-01 -4.12037462e-01 -3.97328883e-01 -1.20754635e+00
5.61560988e-02 -1.14668989e+00 2.00178474e-03 -6.63736761e-01
-9.29740310e-01 2.36597925e-01 -5.39303422e-01 -1.54206872e+00
2.32022017e-01 -8.71466398e-01 -8.46753240e-01 6.48957908e-01
-1.76486015e+00 -1.00742912e+00 -5.83162189e-01 8.48457754e-01
7.12370396e-01 -1.44388586e-01 4.70579416e-01 3.89719099e-01
-3.15762162e-01 9.22378182e-01 3.24727386e-01 1.16004038e+00
6.60923898e-01 -9.89917576e-01 7.21867442e-01 1.10894763e+00
-7.60720149e-02 9.29178178e-01 4.11898166e-01 -7.06181347e-01
-1.45700145e+00 -1.19752884e+00 4.57340717e-01 -1.99908182e-01
4.74227011e-01 -3.40191901e-01 -9.56751704e-01 7.87427127e-01
5.20949006e-01 3.28165263e-01 2.68520623e-01 -7.74945796e-01
-8.25218439e-01 1.17446132e-01 -1.31145084e+00 3.77627730e-01
5.71961462e-01 -1.09087050e+00 1.88465286e-02 3.56011897e-01
4.86721992e-01 -3.49107295e-01 -5.36355019e-01 8.83593559e-02
6.57978356e-01 -1.08681178e+00 1.24574530e+00 -3.04682493e-01
8.33500922e-01 -2.96743304e-01 -7.21657053e-02 -9.82116103e-01
-1.50865704e-01 -6.79718971e-01 -3.59385341e-01 5.90955138e-01
2.42049731e-02 -3.20749015e-01 9.38420355e-01 6.40465990e-02
1.19124152e-01 -3.50113571e-01 -7.10553348e-01 -6.74084306e-01
6.24699853e-02 -4.67252851e-01 7.23851979e-01 1.30624819e+00
-1.55597091e-01 -7.63275325e-01 -8.38136613e-01 5.60930431e-01
7.00034142e-01 -3.56488198e-01 4.43548381e-01 -3.79458517e-01
-1.57974303e-01 -4.14687723e-01 -5.76179802e-01 -1.30244064e+00
-8.37160274e-02 -4.17641342e-01 -2.56763250e-02 -8.61212492e-01
8.84697363e-02 -1.32989302e-01 -2.80067146e-01 2.80628324e-01
3.58386971e-02 6.82463467e-01 1.44868657e-01 2.93964356e-01
-4.08331633e-01 2.43538663e-01 1.09071898e+00 -6.17424488e-01
2.36974970e-01 -2.93710530e-01 -4.51518327e-01 6.79399550e-01
9.90968585e-01 -4.57053423e-01 -1.93557158e-01 -8.87045801e-01
2.86998805e-02 -2.65738722e-02 5.87629735e-01 -1.29746377e+00
4.83532995e-01 7.85818622e-02 9.48544085e-01 -7.55383611e-01
8.57964978e-02 -1.10323560e+00 -2.88542509e-01 7.14790583e-01
-6.81042314e-01 2.24898875e-01 -1.09440319e-01 6.31295502e-01
-3.89877230e-01 -3.33888263e-01 9.71435189e-01 -3.53761762e-01
-5.74933171e-01 2.33925000e-01 -7.53420830e-01 -4.05255049e-01
8.69225681e-01 -5.36847889e-01 -4.93751407e-01 -6.76841259e-01
-4.13848698e-01 -2.82729894e-01 5.59899688e-01 3.37198257e-01
1.12223768e+00 -1.27875268e+00 -2.95541942e-01 7.12899745e-01
-2.27300316e-01 -4.31628585e-01 4.04366553e-01 6.75766408e-01
-1.48229802e+00 2.89754987e-01 -4.07610625e-01 -4.55025643e-01
-1.24585903e+00 8.39484155e-01 5.57588220e-01 -2.21812576e-02
-6.62935853e-01 5.56324184e-01 -1.63429067e-01 3.41066010e-02
3.41296643e-01 -5.56715369e-01 1.02837123e-01 -6.42892957e-01
9.33717847e-01 3.44240457e-01 -3.87679599e-02 -4.43625659e-01
-1.13917358e-01 5.39919496e-01 -2.31130451e-01 2.83456445e-01
9.59846199e-01 -2.69709945e-01 -2.38666385e-01 -4.18416291e-01
1.89669073e+00 -5.91484644e-03 -1.34706569e+00 -1.01040959e-01
-1.45868719e-01 -9.20159757e-01 2.53317416e-01 -5.12683749e-01
-1.64009869e+00 1.26188719e+00 7.81549990e-01 3.44654649e-01
1.24680293e+00 -6.74435198e-01 1.18097651e+00 2.63201296e-01
1.45984560e-01 -1.21876991e+00 2.96432674e-01 4.06995445e-01
6.47338510e-01 -1.21792555e+00 4.11541425e-02 -2.76124060e-01
-2.76570618e-01 1.63152313e+00 4.53650713e-01 -4.32750851e-01
6.73872471e-01 7.17603385e-01 2.28800684e-01 -1.53916776e-01
-1.01834290e-01 4.75354820e-01 -2.37385795e-01 9.31892455e-01
1.43870473e-01 -2.85082906e-01 3.06772500e-01 -9.63787064e-02
3.33861932e-02 -4.25497927e-02 7.06313550e-01 1.10395825e+00
-3.64821225e-01 -8.39360714e-01 -6.02554202e-01 3.11965436e-01
-7.87397981e-01 -2.97549188e-01 -3.89283329e-01 7.38441229e-01
3.24296683e-01 8.02762210e-01 2.20023870e-01 -6.65862501e-01
-3.64901751e-01 -4.26046133e-01 2.70321369e-01 1.34074673e-01
-5.70941150e-01 -2.38608196e-01 -5.57341576e-01 -8.00246954e-01
-1.91710427e-01 -1.15293384e-01 -9.82433200e-01 -6.63132966e-01
-3.50843012e-01 -1.83399171e-01 8.74925554e-01 6.43513978e-01
2.80661136e-01 5.55231810e-01 8.97531807e-01 -8.30754876e-01
-5.57139516e-01 -8.70527685e-01 -6.31534100e-01 4.42544132e-01
9.81169164e-01 -3.04322932e-02 -5.79565942e-01 5.01336277e-01] | [12.303071022033691, 0.9517441391944885] |
1aff20ea-6cf6-481f-ad26-d6ea756028a4 | non-decreasing-quantile-function-network-with-1 | 2105.06696 | null | https://arxiv.org/abs/2105.06696v1 | https://arxiv.org/pdf/2105.06696v1.pdf | Non-decreasing Quantile Function Network with Efficient Exploration for Distributional Reinforcement Learning | Although distributional reinforcement learning (DRL) has been widely examined in the past few years, there are two open questions people are still trying to address. One is how to ensure the validity of the learned quantile function, the other is how to efficiently utilize the distribution information. This paper attempts to provide some new perspectives to encourage the future in-depth studies in these two fields. We first propose a non-decreasing quantile function network (NDQFN) to guarantee the monotonicity of the obtained quantile estimates and then design a general exploration framework called distributional prediction error (DPE) for DRL which utilizes the entire distribution of the quantile function. In this paper, we not only discuss the theoretical necessity of our method but also show the performance gain it achieves in practice by comparing with some competitors on Atari 2600 Games especially in some hard-explored games. | ['Liwen Zhang', 'Qi Kuang', 'Zhoufan Zhu', 'Fan Zhou'] | 2021-05-14 | non-decreasing-quantile-function-network-with | https://openreview.net/forum?id=f_GA2IU9-K- | https://openreview.net/pdf?id=f_GA2IU9-K- | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-4.16535228e-01 3.92912365e-02 -3.67440015e-01 -3.25551331e-01
-1.01162219e+00 -5.55849731e-01 1.23168588e-01 1.61772761e-02
-5.14632940e-01 1.21963036e+00 -1.30429240e-02 -4.85704213e-01
-5.62372029e-01 -9.53033626e-01 -6.01943552e-01 -8.03546846e-01
-4.13605094e-01 4.96481329e-01 2.33933538e-01 -4.31611478e-01
4.18479830e-01 2.97503680e-01 -1.50501978e+00 -2.86190152e-01
1.14577687e+00 1.22614896e+00 1.56932056e-01 3.33166003e-01
-4.18917127e-02 7.33180761e-01 -6.66141093e-01 -4.03689802e-01
3.36551994e-01 -6.04953468e-01 -5.76964915e-01 -3.63013536e-01
-1.13143757e-01 -6.15013301e-01 -7.16189444e-02 1.39416790e+00
7.04256117e-01 3.29140365e-01 5.39494455e-01 -1.45301676e+00
-4.25406039e-01 9.31978703e-01 -1.04167211e+00 4.51743007e-01
4.05524150e-02 -6.66853972e-04 9.88680482e-01 -3.71550173e-01
1.50523320e-01 1.10692132e+00 5.03893375e-01 4.13307160e-01
-5.61400533e-01 -9.22760725e-01 8.03275630e-02 4.63122487e-01
-1.36823559e+00 1.04808167e-01 6.22120440e-01 -6.57456219e-02
4.47700232e-01 -1.23141862e-01 5.08966029e-01 7.16291845e-01
2.61765093e-01 7.39865959e-01 1.28840029e+00 -2.32531309e-01
6.35874033e-01 -7.45442286e-02 2.03010477e-02 3.94162238e-01
2.12204427e-01 4.87112969e-01 -4.64697063e-01 -1.43592894e-01
9.15032685e-01 -4.34516460e-01 -1.64525628e-01 -5.39422631e-01
-5.73999763e-01 1.14568698e+00 1.29334450e-01 -8.27938616e-02
-3.09291542e-01 3.05489719e-01 4.89619046e-01 4.48456198e-01
5.35480976e-01 2.15809062e-01 -3.76832992e-01 -7.52342999e-01
-9.34657454e-01 4.65052396e-01 7.13567138e-01 9.24200296e-01
5.66778839e-01 2.33395427e-01 -1.47748008e-01 6.06145918e-01
2.36584678e-01 5.54165900e-01 3.98211628e-01 -8.98223519e-01
6.63375139e-01 1.09920368e-01 4.21662450e-01 -5.66465974e-01
-2.94668853e-01 -2.63416708e-01 -5.68345428e-01 3.63626480e-01
7.13379145e-01 -6.08497202e-01 -3.87997031e-01 1.94904029e+00
3.95452708e-01 1.94816723e-01 4.03045155e-02 1.03015506e+00
3.91600013e-01 5.72001457e-01 -4.27983254e-02 -1.17815837e-01
9.42554414e-01 -5.48197389e-01 -5.21336854e-01 1.36658773e-01
3.07163417e-01 -4.20291007e-01 1.19249868e+00 6.55787349e-01
-1.14453673e+00 -3.68133247e-01 -1.05180669e+00 2.55428284e-01
-5.60124405e-02 -4.55221161e-02 6.95548236e-01 9.58423197e-01
-7.93871701e-01 6.70757651e-01 -8.60815346e-01 -1.73364133e-02
3.11865449e-01 4.60294724e-01 1.23775557e-01 2.55580872e-01
-1.57827175e+00 6.82593405e-01 5.77735126e-01 -4.05118726e-02
-1.05181229e+00 -4.62233067e-01 -6.85167015e-01 3.84361982e-01
8.06853533e-01 -4.57805255e-03 1.31901896e+00 -4.92258042e-01
-1.88602495e+00 1.78961083e-01 4.68847901e-01 -6.34938776e-01
8.15753520e-01 -3.83082628e-01 -1.48023367e-01 3.95075828e-02
-2.29264703e-03 3.61477375e-01 3.28636467e-01 -6.88428402e-01
-9.85152662e-01 -4.29472178e-01 3.66586506e-01 3.67308557e-01
-3.79842848e-01 -2.95558959e-01 -1.28566936e-01 -6.10950291e-01
-2.45904908e-01 -6.11101985e-01 -3.08560848e-01 -1.99539229e-01
-2.38813460e-01 -6.79471791e-01 2.86071032e-01 -3.20149571e-01
1.43981445e+00 -2.26200628e+00 -1.73549373e-02 3.59570891e-01
-1.18832543e-01 3.27779427e-02 -2.49998234e-02 6.73261881e-01
1.75098315e-01 -1.23494700e-01 -1.43678397e-01 -2.58139474e-03
2.48975843e-01 2.67503709e-01 -8.39101374e-01 6.53099239e-01
-3.25382262e-01 6.10738695e-01 -9.68155444e-01 -3.15464348e-01
-7.47646298e-03 3.26003693e-02 -6.00312769e-01 2.79818714e-01
-3.71459007e-01 1.56398416e-01 -6.19403780e-01 4.34787482e-01
8.78187597e-01 1.62212685e-01 1.30987018e-01 2.22604260e-01
-2.78826118e-01 8.59863311e-02 -1.39363992e+00 1.38491535e+00
-6.68353066e-02 1.72537178e-01 -1.63600653e-01 -1.16589689e+00
1.14081979e+00 -2.01456044e-02 4.86098260e-01 -7.59507716e-01
2.14045674e-01 3.06518942e-01 -3.52754556e-02 -4.77176964e-01
6.57122314e-01 -3.91568482e-01 -2.77975559e-01 5.03593266e-01
-6.25582933e-02 1.60775244e-01 3.22577864e-01 -1.06549352e-01
5.45981228e-01 3.49955857e-01 3.42736423e-01 -6.71021640e-01
3.90962809e-01 -7.94075355e-02 5.69615424e-01 9.53209102e-01
-4.78550136e-01 2.28199005e-01 1.14806449e+00 -3.29030305e-01
-8.89745951e-01 -1.26086414e+00 -1.38673142e-01 1.23826599e+00
5.37320077e-01 -2.31690720e-01 -7.62594104e-01 -8.76782954e-01
1.22633137e-01 8.31812143e-01 -7.30736315e-01 -9.92381275e-02
-2.52477348e-01 -6.20609641e-01 8.32700312e-01 8.05798888e-01
8.04355025e-01 -1.06315875e+00 -7.80435443e-01 1.32367715e-01
6.38282020e-03 -5.91564059e-01 -4.77281183e-01 3.99326295e-01
-7.88143158e-01 -1.07767081e+00 -6.42937660e-01 -3.37443173e-01
4.26309630e-02 2.35175639e-02 9.62207794e-01 -2.93674499e-01
3.74027163e-01 8.80052298e-02 -4.10314590e-01 -5.41967809e-01
1.60153195e-01 2.24401474e-01 -5.54841161e-02 -4.98226613e-01
4.85204041e-01 -6.12729371e-01 -5.14879465e-01 4.58023965e-01
-8.50374281e-01 -5.01404345e-01 4.26239818e-01 8.85273039e-01
6.42423749e-01 3.54772538e-01 1.18574202e+00 -9.22729313e-01
1.05699968e+00 -7.30708301e-01 -1.17035854e+00 2.27814361e-01
-6.05995953e-01 2.37707511e-01 8.58832717e-01 -2.92347252e-01
-1.01870334e+00 -4.27553892e-01 -4.78807926e-01 -1.43367186e-01
1.76628664e-01 7.65513957e-01 -5.92765678e-03 2.88077503e-01
4.87081498e-01 3.10957491e-01 1.11638501e-01 -3.52903813e-01
3.30330223e-01 7.05843449e-01 6.15696192e-01 -1.11472023e+00
5.83848476e-01 2.14617580e-01 -1.46190468e-02 -5.49130738e-01
-1.03683460e+00 -2.95682847e-01 -1.07510164e-01 -1.63052365e-01
5.48689127e-01 -6.39106214e-01 -1.36279488e+00 1.41477823e-01
-5.02798498e-01 -5.67135632e-01 -4.66325283e-01 5.87083459e-01
-1.12254667e+00 4.76114720e-01 -6.04997337e-01 -1.17709315e+00
-4.17998195e-01 -1.02807987e+00 6.85585260e-01 4.29764062e-01
2.90228486e-01 -8.68067563e-01 4.27132368e-01 -8.62487108e-02
3.57320756e-01 1.10774748e-01 9.76742208e-01 -5.58674037e-01
-2.36065716e-01 2.60650426e-01 -3.32433343e-01 3.97255152e-01
-8.43523964e-02 -3.97868395e-01 -5.38086236e-01 -5.68648875e-01
7.01484531e-02 -6.98349714e-01 7.45923936e-01 6.45030260e-01
1.51292241e+00 4.01052088e-02 1.09118760e-01 5.55966198e-01
1.45415854e+00 3.75786811e-01 7.68695891e-01 5.58991432e-01
4.88874763e-02 3.21776956e-01 1.07617080e+00 9.26815987e-01
3.25084686e-01 4.23622757e-01 6.39716387e-01 3.14331383e-01
4.62251931e-01 -5.51780343e-01 3.77962738e-01 4.36279029e-01
-1.77885726e-01 -3.07426244e-01 -6.51609898e-01 3.78343552e-01
-2.00130749e+00 -8.86360168e-01 3.29005271e-01 2.36797428e+00
9.17637825e-01 3.23336571e-01 5.75554311e-01 -4.44895402e-02
7.04693258e-01 -6.33168221e-02 -8.89814258e-01 -5.68666101e-01
1.95503950e-01 4.80488002e-01 5.88891745e-01 4.71049547e-01
-7.98558056e-01 8.57486188e-01 6.70898247e+00 1.34838021e+00
-9.51092720e-01 -1.59475267e-01 6.93952262e-01 9.97295678e-02
-3.20735365e-01 -5.38419448e-02 -8.33745897e-01 5.65083325e-01
7.09553897e-01 -4.23732311e-01 5.57403505e-01 1.35883629e+00
1.44424200e-01 -3.54201615e-01 -6.56540453e-01 7.93471396e-01
-3.62202525e-01 -9.59867179e-01 -2.57757425e-01 2.54433841e-01
5.86743653e-01 -1.12105785e-02 3.11811745e-01 8.63800764e-01
5.18251181e-01 -1.14061654e+00 5.84768236e-01 4.46734726e-01
7.90227592e-01 -1.69997668e+00 9.89075065e-01 6.31809473e-01
-1.01803398e+00 -1.57798856e-01 -1.06511378e+00 -2.69505769e-01
-1.67223781e-01 3.76104146e-01 -4.49113876e-01 5.74457705e-01
7.58745790e-01 3.57921571e-01 -3.15965623e-01 1.36568999e+00
-3.99278075e-01 8.69313955e-01 -3.15239310e-01 -3.48545521e-01
6.77127123e-01 -4.28267628e-01 2.69269854e-01 7.49102771e-01
4.95911628e-01 2.85641546e-03 2.38512710e-01 6.88601315e-01
-2.10437495e-02 1.75979003e-01 -2.80502081e-01 2.34503578e-02
5.43003321e-01 1.01837909e+00 -5.88516951e-01 -3.98583673e-02
-3.00483078e-01 3.87027562e-01 6.35091543e-01 1.60183176e-01
-1.16540158e+00 -5.76805711e-01 5.90692282e-01 -9.26757604e-02
4.71909106e-01 -2.38752633e-01 -9.83596593e-02 -8.71143758e-01
-1.08677894e-01 -7.94319212e-01 7.20615566e-01 -4.39748973e-01
-1.49050725e+00 6.00278914e-01 1.79792866e-01 -1.13941205e+00
-4.59084183e-01 -5.31169653e-01 -5.84805787e-01 7.03987360e-01
-1.64524364e+00 -4.82162029e-01 -4.92242612e-02 5.41235268e-01
4.26523030e-01 -2.44953781e-01 5.28348446e-01 2.12392002e-01
-3.26634496e-01 8.46507967e-01 6.07131541e-01 3.07498649e-02
6.34202182e-01 -1.58492172e+00 6.34174272e-02 3.75043780e-01
-2.75100116e-02 2.76244104e-01 7.41371810e-01 -5.21210790e-01
-1.09303713e+00 -7.24703074e-01 1.73482552e-01 4.55618389e-02
7.43285656e-01 -1.41437069e-01 -7.39169180e-01 4.77233142e-01
3.28264892e-01 -2.27333561e-01 6.27809823e-01 1.06552213e-01
-5.35054877e-02 -1.69612914e-01 -1.14524579e+00 4.91147608e-01
6.39853001e-01 1.20156854e-02 -4.09423709e-01 -1.15795664e-01
4.31783885e-01 -5.69972813e-01 -7.34247983e-01 4.54638869e-01
5.69593370e-01 -1.32058239e+00 6.29992545e-01 -6.72138631e-01
4.29564536e-01 -1.97769344e-01 -2.21834466e-01 -1.42807996e+00
9.48139876e-02 -5.43153167e-01 -7.68886358e-02 1.12976062e+00
3.82631458e-02 -7.06717253e-01 1.10900509e+00 1.92406595e-01
6.98082969e-02 -1.48514676e+00 -1.02428651e+00 -1.09898317e+00
7.52174377e-01 -5.77268481e-01 8.97393882e-01 5.10153234e-01
2.65588164e-01 -1.03620172e-01 -6.82279706e-01 -2.02598162e-02
6.74233198e-01 4.00536388e-01 6.14351332e-01 -8.80081534e-01
-4.65677530e-01 -3.48978490e-01 -2.56145358e-01 -1.45022464e+00
5.25914915e-02 -4.91387755e-01 1.05011195e-01 -1.15192378e+00
1.91002354e-01 -5.41619599e-01 -5.08927226e-01 1.26359582e-01
-1.08800225e-01 -6.03685267e-02 -3.91855799e-02 -8.93420205e-02
-9.19341326e-01 1.03074837e+00 1.29065561e+00 2.70436078e-01
-1.74335659e-01 3.26028436e-01 -9.79024410e-01 5.32725692e-01
1.00118291e+00 -5.05088031e-01 -8.16624999e-01 -1.66326407e-02
3.79922509e-01 5.18146276e-01 1.28378011e-02 -9.00841892e-01
2.19683498e-01 -6.24893546e-01 3.46995652e-01 -8.66672575e-01
-6.98927715e-02 -5.25768042e-01 -4.41803604e-01 2.36022681e-01
-3.70591104e-01 3.69246244e-01 2.66686231e-01 6.71602726e-01
-3.76351655e-01 -5.18729329e-01 8.42051327e-01 8.04676563e-02
-6.54550672e-01 3.78120065e-01 -3.44164819e-01 5.05345821e-01
1.08365262e+00 1.85207091e-02 -2.63966352e-01 -7.62168646e-01
-4.79345351e-01 5.76715291e-01 1.57494053e-01 1.12662045e-02
5.30355096e-01 -1.30741429e+00 -5.43771982e-01 -5.21266013e-02
-9.80195776e-02 -2.37119406e-01 3.20082217e-01 5.63343048e-01
-4.63361084e-01 1.87438816e-01 -3.31244946e-01 -3.07077706e-01
-6.12669170e-01 5.29600978e-01 4.43059713e-01 -5.70200086e-01
-4.63137329e-01 6.18383110e-01 4.73397821e-02 -3.01587582e-01
4.76582646e-01 -7.50044882e-02 -2.14327857e-01 -1.78644270e-01
5.17961800e-01 5.68515003e-01 -2.59640276e-01 -5.89633435e-02
-1.05381846e-01 3.21781963e-01 3.81990932e-02 -1.30750135e-01
1.43271792e+00 -1.33798137e-01 1.38740912e-01 3.03950429e-01
7.14636803e-01 -4.85334657e-02 -1.65504277e+00 1.25332351e-03
1.67039305e-01 -5.17617881e-01 -8.20480585e-02 -7.86753476e-01
-1.21313322e+00 8.73631418e-01 6.01828039e-01 5.00021040e-01
1.26011813e+00 -2.31240287e-01 7.54494429e-01 4.06062603e-01
7.15471148e-01 -1.36397123e+00 -2.02765930e-02 7.13995576e-01
6.31913424e-01 -9.21785057e-01 -5.74641079e-02 -1.33572951e-01
-8.24987292e-01 1.41608417e+00 7.68063307e-01 -5.43716013e-01
6.79555118e-01 3.13427448e-01 -9.78729799e-02 -1.24337249e-01
-5.65948486e-01 -4.45860922e-01 -5.65562360e-02 6.23462260e-01
3.45144629e-01 -3.07386089e-02 -5.80327451e-01 8.27880442e-01
-5.83977938e-01 2.43618079e-02 5.83147168e-01 6.46505415e-01
-7.00685918e-01 -1.27791762e+00 -1.46639556e-01 4.26523596e-01
-4.94837105e-01 6.95769265e-02 1.48635790e-01 1.09385002e+00
-4.27061878e-02 8.36666286e-01 -1.49891004e-01 -2.59788245e-01
2.31305987e-01 -2.85198957e-01 5.47217369e-01 -3.82226408e-01
-2.29833007e-01 9.72554609e-02 -1.09760970e-01 -4.23619062e-01
-7.37890378e-02 -4.24625218e-01 -1.24506998e+00 -5.28778851e-01
-5.16947210e-01 6.83396459e-01 4.29566950e-01 1.00988007e+00
4.64874227e-03 2.49425799e-01 5.80379784e-01 -2.71646500e-01
-1.38747418e+00 -8.22031498e-01 -1.26694381e+00 -8.45429376e-02
-7.20323101e-02 -1.07147300e+00 -2.27655634e-01 -7.71660209e-01] | [4.045472621917725, 2.571512460708618] |
8f99f9cf-3f75-486f-8cfe-a086a14bb378 | profilesr-gan-a-gan-based-super-resolution | 2107.09523 | null | https://arxiv.org/abs/2107.09523v2 | https://arxiv.org/pdf/2107.09523v2.pdf | ProfileSR-GAN: A GAN based Super-Resolution Method for Generating High-Resolution Load Profiles | It is a common practice for utilities to down-sample smart meter measurements from high resolution (e.g. 1-min or 1-sec) to low resolution (e.g. 15-, 30- or 60-min) to lower the data transmission and storage cost. However, down-sampling can remove high-frequency components from time-series load profiles, making them unsuitable for in-depth studies such as quasi-static power flow analysis or non-intrusive load monitoring (NILM). Thus, in this paper, we propose ProfileSR-GAN: a Generative Adversarial Network (GAN) based load profile super-resolution (LPSR) framework for restoring high-frequency components lost through the smoothing effect of the down-sampling process. The LPSR problem is formulated as a Maximum-a-Prior problem. When training the ProfileSR-GAN generator network, to make the generated profiles more realistic, we introduce two new shape-related losses in addition to conventionally used content loss: adversarial loss and feature-matching loss. Moreover, a new set of shape-based evaluation metrics are proposed to evaluate the realisticness of the generated profiles. Simulation results show that ProfileSR-GAN outperforms Mean-Square Loss based methods in all shape-based metrics. The successful application in NILM further demonstrates that ProfileSR-GAN is effective in recovering high-resolution realistic waveforms. | ['Ning Lu', 'Yiyan Li', 'Lidong Song'] | 2021-07-18 | null | null | null | null | ['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'miscellaneous', 'time-series'] | [ 1.72906682e-01 -3.07935953e-01 5.15834056e-03 -2.84185886e-01
-1.17459357e+00 -4.56722498e-01 2.66148597e-01 -8.80383700e-02
2.47360885e-01 1.00569367e+00 2.63337761e-01 -1.66208699e-01
-2.03434899e-01 -1.21212709e+00 -3.02496344e-01 -8.75187337e-01
-2.70007700e-01 6.07476942e-03 -2.07403481e-01 -2.25304976e-01
-2.30906606e-01 6.94468021e-01 -1.09031832e+00 9.77562591e-02
1.27461350e+00 1.00150931e+00 1.05312347e-01 3.60394835e-01
3.73187214e-01 8.57319057e-01 -1.25931513e+00 -3.57228577e-01
2.80615479e-01 -8.56904805e-01 -1.85365751e-01 -2.84155399e-01
-4.12145585e-01 -6.79029822e-01 -4.18403447e-01 1.21729040e+00
9.24117565e-01 2.79605091e-01 6.56463265e-01 -1.63306189e+00
-4.88944650e-01 6.59454465e-01 -6.75770581e-01 3.86089623e-01
5.17108023e-01 2.11329415e-01 6.95630372e-01 -3.52563918e-01
-1.80167288e-01 8.88028145e-01 1.07539690e+00 8.63764733e-02
-1.45316803e+00 -7.14950383e-01 -2.96256393e-01 4.26940560e-01
-1.35513365e+00 -2.71310270e-01 1.32754636e+00 -1.48084611e-01
8.16971481e-01 5.11362374e-01 4.39450473e-01 8.76886666e-01
1.38705699e-02 5.60305655e-01 1.15541375e+00 -1.68042615e-01
2.95987427e-01 -8.84575620e-02 -5.91864526e-01 -1.63055182e-01
1.03272587e-01 7.35451952e-02 -2.15282857e-01 -2.54173428e-01
7.53585875e-01 -2.04436272e-01 -5.85824251e-01 2.76419908e-01
-4.60870057e-01 9.03744936e-01 3.58123422e-01 3.13989103e-01
-5.53548098e-01 -1.05577990e-01 5.21785736e-01 2.84301311e-01
7.42862880e-01 2.07577512e-01 -1.64407462e-01 -4.21850413e-01
-1.25638223e+00 -6.84983358e-02 4.06777054e-01 8.78179729e-01
3.37602526e-01 1.06223607e+00 -3.14379722e-01 9.28141177e-01
1.57500684e-01 8.44786167e-01 6.45981312e-01 -7.50717163e-01
5.89308262e-01 4.88908850e-02 1.91833362e-01 -8.88383448e-01
-4.91526097e-01 -2.71749884e-01 -1.30993760e+00 1.81259587e-01
3.08057457e-01 -3.04157943e-01 -3.65964293e-01 1.82061696e+00
5.51412851e-02 3.14149588e-01 5.09754382e-02 8.20750356e-01
5.19531131e-01 9.84369218e-01 1.92175191e-02 -5.65163612e-01
1.22138155e+00 -5.09132802e-01 -1.03636670e+00 2.85327315e-01
4.91846688e-02 -7.78320849e-01 1.27848649e+00 3.89514357e-01
-1.34594095e+00 -4.88021731e-01 -1.08779323e+00 3.20793122e-01
-3.54046039e-02 -4.86294888e-02 1.07856363e-01 9.17726576e-01
-5.40961981e-01 9.21740413e-01 -6.02951646e-01 1.76907450e-01
5.86523950e-01 -4.06983227e-01 8.13649818e-02 2.31597736e-01
-1.44964099e+00 8.01484406e-01 5.88910393e-02 3.78761888e-02
-7.57293701e-01 -1.31683481e+00 -7.77365267e-01 3.69765550e-01
-1.10261977e-01 -2.88690686e-01 9.55586195e-01 -6.08338714e-01
-1.77256560e+00 1.22860104e-01 1.38142988e-01 -6.94044888e-01
6.19183898e-01 -7.25174770e-02 -9.70823109e-01 4.26965177e-01
-7.67727271e-02 -2.82857031e-01 9.59130764e-01 -1.14839542e+00
-3.14173549e-01 -1.84670120e-01 -1.73348486e-01 -2.48859003e-02
-4.14135568e-02 -1.76346779e-01 5.09692967e-01 -1.23434353e+00
-2.47882068e-01 -1.87702075e-01 1.06565602e-01 -4.38061833e-01
-4.88929242e-01 2.51278520e-01 9.38643515e-01 -1.27603841e+00
1.06667018e+00 -1.97635746e+00 -4.83098269e-01 2.75044858e-01
-2.48813912e-01 5.04053056e-01 -3.80803719e-02 6.21970117e-01
-3.33591074e-01 7.91682303e-02 -4.72000331e-01 -2.87289590e-01
1.93324491e-01 2.07034722e-02 -5.58102250e-01 6.23050809e-01
1.32210389e-01 8.05040658e-01 -8.57317209e-01 9.47021842e-02
6.25291824e-01 8.04508746e-01 -1.08631514e-01 2.60363251e-01
2.22568080e-01 5.09256959e-01 -1.90200046e-01 4.09633636e-01
9.55359817e-01 8.79569948e-02 -2.06727207e-01 -7.65036404e-01
2.19814450e-01 3.60997200e-01 -1.00611985e+00 1.38073146e+00
-9.78151798e-01 5.89148402e-01 8.33715424e-02 -1.21048093e+00
1.15167332e+00 3.72310847e-01 7.61819899e-01 -1.22942984e+00
-2.50170827e-02 -4.65917774e-02 -3.08701605e-01 -3.61332536e-01
3.19398820e-01 -4.44361567e-01 -7.37273619e-02 4.79054838e-01
-1.50733501e-01 -2.80271262e-01 1.64959922e-01 -7.24934340e-02
9.34008479e-01 -3.02185379e-02 2.27698535e-01 -3.27462256e-01
5.70936739e-01 -5.18927872e-01 9.53793108e-01 1.25080511e-01
-1.25443131e-01 5.06403744e-01 3.13727379e-01 -3.88469063e-02
-1.22567689e+00 -1.16906106e+00 -2.85173625e-01 3.77929479e-01
7.60105392e-03 -2.42684230e-01 -7.28675961e-01 -3.82240444e-01
8.24810565e-02 1.50194216e+00 -2.25541100e-01 -4.54955965e-01
-6.12518072e-01 -1.18225646e+00 8.21266413e-01 7.32574165e-01
6.39972329e-01 -9.66296673e-01 -6.52127206e-01 5.37682950e-01
-3.50728691e-01 -1.02435851e+00 -5.38170338e-01 5.34535944e-02
-6.52421832e-01 -8.37216198e-01 -7.53484368e-01 -1.00998633e-01
4.90786195e-01 -1.18642695e-01 1.23249030e+00 -3.91459405e-01
-3.07147950e-01 2.04109296e-01 -5.29928386e-01 -3.59971106e-01
-5.03025651e-01 -2.99323022e-01 -1.13173120e-01 8.63575414e-02
8.73446465e-02 -1.29010653e+00 -6.66679740e-01 2.96413481e-01
-7.23559380e-01 -2.86595911e-01 3.40239942e-01 7.24915147e-01
4.20954019e-01 7.11067796e-01 1.46378136e+00 -4.88311321e-01
9.33133245e-01 -4.58644271e-01 -7.67754197e-01 -5.14084250e-02
-6.47811532e-01 -4.25210327e-01 1.40324891e+00 -5.17152011e-01
-1.13305819e+00 -5.42749465e-01 -4.61006999e-01 -4.61068749e-01
1.15979679e-01 2.32454836e-01 -6.96863770e-01 1.84658453e-01
3.00671726e-01 3.69322658e-01 -1.07030727e-01 -5.44797421e-01
2.99084097e-01 5.49587488e-01 7.82116652e-01 -4.56482410e-01
1.23649514e+00 2.93486565e-01 1.75342798e-01 -9.00008678e-01
-5.81732750e-01 -8.31980035e-02 6.20909296e-02 -6.73490539e-02
5.80093563e-01 -9.60902333e-01 -8.15087378e-01 8.87727737e-01
-7.52784252e-01 -5.97395837e-01 -9.87426221e-01 4.45437521e-01
-7.80179739e-01 4.40672606e-01 -7.77352810e-01 -9.77583170e-01
-8.82898092e-01 -1.06914735e+00 7.76274025e-01 4.25603926e-01
-3.69437486e-02 -1.13592529e+00 -1.07059620e-01 1.33395955e-01
8.81169677e-01 7.76384115e-01 9.24064636e-01 -3.62256527e-01
-1.51456431e-01 -1.51926011e-01 -9.45896506e-02 7.12183475e-01
2.39882067e-01 -4.01168227e-01 -1.10757089e+00 -5.56755602e-01
4.52464312e-01 7.66499620e-03 1.38669461e-01 4.65504497e-01
1.28674841e+00 -5.33859730e-01 1.52759761e-01 8.20284605e-01
1.66454768e+00 4.69234854e-01 1.32788754e+00 4.17475067e-02
2.98698902e-01 6.60140906e-03 2.98374653e-01 8.84009123e-01
2.71071732e-01 5.66959441e-01 2.01717153e-01 -2.24346906e-01
-1.50143474e-01 -3.39853317e-01 2.16110721e-01 8.66816759e-01
9.11243558e-02 -3.79552394e-01 -8.84720758e-02 4.85315502e-01
-1.32351995e+00 -1.21928895e+00 -1.01790823e-01 2.40302253e+00
1.01897931e+00 5.89307062e-02 3.24391156e-01 7.76073635e-01
6.88711226e-01 2.42390558e-01 -7.81952202e-01 -3.81847918e-01
-3.59452546e-01 4.20478672e-01 5.52743495e-01 4.51323181e-01
-7.40780532e-01 -1.64172605e-01 5.15150642e+00 1.24612200e+00
-1.00025415e+00 3.06056708e-01 5.83377659e-01 1.70841604e-01
-6.31455362e-01 -3.99796724e-01 -4.42546040e-01 1.00847304e+00
1.11459208e+00 -7.78271198e-01 6.24928594e-01 6.71837449e-01
5.79779446e-01 -1.19449876e-01 -8.33502471e-01 1.07699096e+00
-3.53210345e-02 -8.63498211e-01 -1.87170446e-01 -2.05612406e-01
6.57991171e-01 -3.95919472e-01 -1.73615336e-01 3.50302577e-01
1.65208161e-01 -1.03229129e+00 5.71683049e-01 7.23598659e-01
1.07526290e+00 -1.06435478e+00 7.73950219e-01 2.25307971e-01
-1.58035588e+00 -1.06293850e-01 -2.40094826e-01 2.91363209e-01
7.85502255e-01 1.29090154e+00 -4.31607962e-01 9.95424509e-01
6.09057963e-01 3.97080421e-01 -4.34840322e-02 1.08351290e+00
-4.01131272e-01 9.41978455e-01 -4.91539568e-01 5.18435776e-01
-2.72255242e-01 -3.50753397e-01 5.73138773e-01 9.62401927e-01
5.20368040e-01 5.57460710e-02 -2.57326523e-03 1.15180874e+00
-1.09094098e-01 -2.59168297e-01 -2.24532068e-01 3.91574889e-01
9.86406744e-01 1.25703943e+00 -2.91113317e-01 -1.86763078e-01
-3.45439434e-01 8.21096420e-01 -4.37100977e-01 4.92405802e-01
-1.10797811e+00 -9.15417790e-01 6.87508941e-01 2.59769827e-01
2.03751266e-01 2.19773576e-01 -2.97944158e-01 -9.26071107e-01
2.94529885e-01 -7.54304528e-01 2.09983259e-01 -8.29049110e-01
-1.73922455e+00 3.38461339e-01 7.18652606e-02 -1.41702938e+00
-5.69004416e-01 3.14394265e-01 -1.19601047e+00 1.39709461e+00
-1.87784362e+00 -8.87762427e-01 -4.76243973e-01 6.09396994e-01
3.91534269e-01 -7.77763575e-02 9.01314735e-01 7.45444179e-01
-4.21183974e-01 7.96018243e-01 2.80567467e-01 1.76042035e-01
1.09953642e-01 -1.28202593e+00 4.01056647e-01 9.36621428e-01
-4.08881634e-01 -1.31812558e-01 8.38648021e-01 -3.58944058e-01
-1.10960567e+00 -1.31768906e+00 2.61052310e-01 2.53777444e-01
4.47907478e-01 -2.81379730e-01 -1.19931173e+00 2.91951329e-01
3.09038550e-01 2.62704752e-02 6.08573079e-01 -8.01705897e-01
-7.27283508e-02 -6.30619824e-01 -1.96636546e+00 1.16027534e-01
3.69109958e-01 -6.45018756e-01 -5.13882816e-01 9.87430960e-02
4.65736568e-01 -1.15335837e-01 -1.55550659e+00 5.57833731e-01
4.85546812e-02 -8.67194533e-01 1.14454317e+00 1.65720694e-02
5.41317351e-02 -4.98034388e-01 -2.65075237e-01 -1.92110479e+00
-1.19975716e-01 -1.21818268e+00 -2.96630144e-01 1.86268044e+00
6.03749417e-02 -9.92329359e-01 4.61501807e-01 2.22076431e-01
-2.04339892e-01 -4.95736778e-01 -1.07509542e+00 -9.99532759e-01
1.33346558e-01 -3.08817297e-01 1.22152388e+00 8.60303998e-01
1.45733193e-01 4.47573476e-02 -2.99076676e-01 4.25240725e-01
1.14079547e+00 9.87659916e-02 4.42903787e-01 -8.81734729e-01
-2.75612265e-01 -4.81189162e-01 -2.88884223e-01 -6.91916168e-01
-3.94067653e-02 -5.83668351e-01 -5.58155254e-02 -1.51900446e+00
-2.79990077e-01 -5.19814789e-01 -1.90129772e-01 1.68593079e-01
-1.64172173e-01 3.18808824e-01 3.19268018e-01 -1.27874598e-01
1.58666775e-01 1.10439515e+00 9.69666302e-01 -1.93349436e-01
-6.42460212e-02 2.85404295e-01 -2.79459298e-01 6.49544358e-01
1.10829341e+00 -2.57687241e-01 -7.52484083e-01 1.03609025e-01
-3.56579162e-02 4.34657365e-01 2.69890577e-01 -1.06744277e+00
-2.54487902e-01 9.49145481e-02 4.05586869e-01 -8.15193594e-01
2.84000099e-01 -7.76840985e-01 4.65467125e-01 3.10041338e-01
1.19662590e-01 1.13148563e-01 3.67638469e-01 1.34722456e-01
-2.12639481e-01 -2.30902016e-01 9.89059925e-01 1.37214065e-01
-1.95826679e-01 3.50955389e-02 -2.19699651e-01 4.75520134e-01
9.44795966e-01 -3.32184113e-03 -4.56734657e-01 -7.66488850e-01
-4.22202110e-01 7.64352828e-02 1.53900817e-01 4.16734777e-02
4.73342031e-01 -1.60239899e+00 -8.56998742e-01 2.77790070e-01
-4.60771561e-01 1.02807924e-01 5.70491374e-01 8.03297162e-01
-1.65098965e-01 -3.06395404e-02 -9.92865637e-02 -2.96245158e-01
-5.79211831e-01 4.16956931e-01 4.55429465e-01 -4.52368706e-01
-1.05205107e+00 7.36173093e-02 -7.20512718e-02 1.59244444e-02
1.01067401e-01 -3.79946440e-01 -4.10472304e-02 1.37371644e-01
7.84019887e-01 8.35813940e-01 3.89394790e-01 -4.80324745e-01
-1.45533364e-02 4.17523950e-01 5.91866016e-01 2.81204373e-01
1.52383530e+00 -2.88128108e-01 1.56633407e-01 3.51172239e-01
1.19358599e+00 6.91597387e-02 -1.39413512e+00 1.04837641e-01
-3.76870632e-01 -4.72179472e-01 1.12058982e-01 -8.29504669e-01
-1.66559505e+00 7.22025096e-01 4.19532150e-01 7.33360410e-01
1.79132652e+00 -5.91920078e-01 1.32333541e+00 -4.17962462e-01
4.76316452e-01 -1.02500057e+00 -1.59865186e-01 -8.27733725e-02
9.45956945e-01 -3.87909830e-01 1.29426956e-01 -1.15443058e-01
-4.23284948e-01 7.99955308e-01 1.25254586e-01 -1.06053188e-01
5.78810751e-01 8.16207528e-01 -1.30170569e-01 3.37253898e-01
-2.85356969e-01 1.74984753e-01 2.68642791e-02 1.07738709e+00
6.88070506e-02 2.54828453e-01 -1.55698434e-01 1.06407475e+00
-4.14868385e-01 -9.33582801e-03 8.18926692e-01 6.47060633e-01
1.18651338e-01 -7.77906597e-01 -5.07143617e-01 6.36383951e-01
-6.83041155e-01 8.52417946e-02 7.41000235e-01 5.45330882e-01
-1.88008323e-01 1.10985291e+00 9.07466188e-02 -6.21526055e-02
8.50836396e-01 -1.36358151e-02 3.65964770e-01 1.47553176e-01
-4.90694702e-01 1.99071348e-01 2.25199796e-02 -5.43211341e-01
-3.11855584e-01 -7.81941712e-01 -1.13041556e+00 -6.29204988e-01
-5.50245285e-01 1.56614795e-01 5.27786791e-01 6.70717478e-01
8.11002851e-02 9.56596375e-01 1.18305278e+00 -7.07135081e-01
-1.00888312e+00 -1.13219368e+00 -1.11843669e+00 8.69586766e-01
2.52696306e-01 -4.30685878e-01 -6.79423928e-01 -9.38682444e-03] | [15.222539901733398, 6.038623332977295] |
016922ae-5035-420f-bae7-6c44487f4a07 | piano-skills-assessment | 2101.04884 | null | https://arxiv.org/abs/2101.04884v2 | https://arxiv.org/pdf/2101.04884v2.pdf | Piano Skills Assessment | Can a computer determine a piano player's skill level? Is it preferable to base this assessment on visual analysis of the player's performance or should we trust our ears over our eyes? Since current CNNs have difficulty processing long video videos, how can shorter clips be sampled to best reflect the players skill level? In this work, we collect and release a first-of-its-kind dataset for multimodal skill assessment focusing on assessing piano player's skill level, answer the asked questions, initiate work in automated evaluation of piano playing skills and provide baselines for future work. Dataset is available from: https://github.com/ParitoshParmar/Piano-Skills-Assessment. | ['Brendan Morris', 'Jaiden Reddy', 'Paritosh Parmar'] | 2021-01-13 | null | null | null | null | ['action-quality-assessment', 'skills-evaluation', 'skills-assessment'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-5.16523719e-02 -1.53410971e-01 -1.28170013e-01 -1.36781812e-01
-1.06901073e+00 -1.03751242e+00 -3.43631580e-02 -1.13489203e-01
-5.85085094e-01 1.06363259e-01 5.41832745e-01 3.51117589e-02
-2.24265441e-01 -4.79622871e-01 -2.02211410e-01 -3.22781622e-01
1.18517026e-01 3.13064128e-01 4.70071375e-01 -3.84109408e-01
4.31765199e-01 2.02371329e-01 -1.73280084e+00 5.64360738e-01
1.99478835e-01 1.01682866e+00 -5.36721013e-02 1.25028038e+00
4.96656895e-01 1.22155142e+00 -5.17931581e-01 -5.00066638e-01
4.79262382e-01 -6.96454108e-01 -9.34487581e-01 -1.58464655e-01
8.27918589e-01 -5.17033339e-01 -6.12533212e-01 9.06640053e-01
7.11683095e-01 3.40351641e-01 1.39329731e-01 -1.15088117e+00
-2.12736905e-01 5.71614861e-01 -3.00851822e-01 8.71066928e-01
5.60755253e-01 6.59573495e-01 9.57536399e-01 -2.50179380e-01
3.77032131e-01 8.60812068e-01 6.79524064e-01 5.18966496e-01
-8.56445909e-01 -1.01230061e+00 -1.41893074e-01 4.50805247e-01
-1.14795256e+00 -6.16032422e-01 7.25585639e-01 -4.38339025e-01
6.26109064e-01 2.78644294e-01 1.14726770e+00 1.12655067e+00
-2.36166269e-01 8.22170675e-01 1.20736158e+00 -1.76929504e-01
-2.95954347e-01 -3.16241980e-01 -1.09891228e-01 5.83475471e-01
-6.72188327e-02 2.13343620e-01 -1.18063927e+00 1.80540029e-02
9.63321328e-01 -4.92085546e-01 -1.30238846e-01 1.64441109e-01
-1.19049954e+00 3.72108161e-01 -9.97052267e-02 2.57208496e-01
-4.22335207e-01 3.99193168e-01 6.59574151e-01 6.59516692e-01
2.34691817e-02 7.80691981e-01 -2.81415701e-01 -1.42281997e+00
-9.98362303e-01 6.09561324e-01 6.77661240e-01 5.23488700e-01
1.15793213e-01 -7.51060024e-02 -1.74188867e-01 6.90230906e-01
-1.48326308e-01 2.18418017e-01 5.38683593e-01 -1.88197112e+00
3.12223047e-01 3.71763706e-01 8.78289789e-02 -9.25288439e-01
-3.62506896e-01 -3.61697301e-02 1.28661975e-01 7.69323707e-01
1.31808734e+00 -3.94087106e-01 -3.73715609e-01 1.68867469e+00
-8.58337879e-02 1.07546479e-01 -2.99764395e-01 1.33318770e+00
1.12566984e+00 1.52588293e-01 1.80299673e-02 1.24304920e-01
1.63990974e+00 -5.57051003e-01 -4.46350813e-01 -2.54812986e-01
4.60550457e-01 -8.21878552e-01 1.39650667e+00 1.13403952e+00
-1.63475287e+00 -9.58729386e-01 -7.98625052e-01 -5.00534428e-03
1.90621465e-01 3.78356278e-01 5.39646983e-01 7.23431468e-01
-9.82587576e-01 8.13522935e-01 -8.90767336e-01 -2.73937851e-01
2.88577408e-01 3.81368279e-01 -6.58063412e-01 4.79210347e-01
-1.14160812e+00 8.47109437e-01 2.58157909e-01 1.16502993e-01
-8.23087990e-01 -5.39860249e-01 -5.58331251e-01 -4.05046381e-02
6.25878990e-01 -2.83387363e-01 1.71510637e+00 -1.40621471e+00
-1.62802303e+00 1.20722055e+00 1.90158978e-01 -1.26639187e-01
4.17362124e-01 -3.62168998e-01 -3.43012005e-01 5.78189194e-01
-2.64912307e-01 6.39898956e-01 4.02763277e-01 -7.40023136e-01
-1.06741858e+00 -3.45865905e-01 6.09034777e-01 3.88156980e-01
-1.58482105e-01 3.30534726e-01 -6.68890119e-01 -4.23438877e-01
-1.06293067e-01 -1.13143647e+00 1.38974831e-01 -2.09558502e-01
-4.27798219e-02 -2.89124250e-01 2.23314628e-01 -1.10531449e+00
1.42706180e+00 -2.19595933e+00 1.09857135e-01 1.22213565e-01
5.41855335e-01 4.84865546e-01 -1.75044239e-01 4.27279115e-01
4.48132642e-02 -1.84219226e-01 4.98729318e-01 -6.40552416e-02
1.13806985e-01 -2.61949867e-01 2.08290413e-01 3.75699133e-01
-1.61719456e-01 8.56693447e-01 -8.25034022e-01 -7.04828739e-01
4.74950951e-03 2.42442191e-01 -4.58266407e-01 8.88030529e-02
2.00936705e-01 4.94470745e-01 -2.73732364e-01 7.87844956e-01
1.38144761e-01 1.59188718e-01 1.42091721e-01 -2.06423715e-01
-1.64217949e-01 2.51936585e-01 -1.14633322e+00 1.63286221e+00
-1.87180713e-01 1.33259451e+00 2.76616484e-01 -6.71665549e-01
6.85766220e-01 3.70767385e-01 5.82812846e-01 -6.46860838e-01
5.75601816e-01 -1.74738064e-01 6.87365413e-01 -8.26726794e-01
7.16655672e-01 -7.01233372e-02 -2.24522665e-01 3.92254621e-01
1.25740230e-01 -1.16041668e-01 5.50777674e-01 2.77329683e-02
1.39014721e+00 3.43518853e-01 -1.90838445e-02 1.50899962e-01
3.35456282e-02 3.42375726e-01 3.79709393e-01 8.46165419e-01
-7.76511252e-01 7.84110487e-01 8.97694409e-01 -2.06217498e-01
-8.59254897e-01 -9.42017734e-01 4.41258430e-01 1.58517039e+00
-1.60292640e-01 -6.32070065e-01 -8.67979527e-01 -3.66805404e-01
-3.81656557e-01 1.02002807e-01 -6.11076176e-01 4.26628031e-02
-4.59586948e-01 2.02773526e-01 1.04158008e+00 8.70098829e-01
3.26927781e-01 -1.50661254e+00 -7.81083763e-01 -5.15834503e-02
-4.89538521e-01 -8.06104183e-01 -4.75337297e-01 -1.90170139e-01
-4.36944425e-01 -1.42896974e+00 -7.66124666e-01 -7.70381749e-01
5.39664254e-02 -9.78081487e-03 1.11740100e+00 -1.20997958e-01
-2.48736143e-02 8.37453544e-01 -5.83955646e-01 -4.40106392e-01
-1.90547436e-01 7.69043788e-02 1.46694511e-01 -5.19353271e-01
7.94855535e-01 -7.51734495e-01 -9.89668250e-01 3.54039133e-01
-5.02617002e-01 -1.18566617e-01 6.65646255e-01 2.95274973e-01
4.95608747e-01 1.39279440e-01 -3.51565145e-02 -5.50629556e-01
9.96278465e-01 -1.31977305e-01 -2.67016470e-01 -1.17378719e-01
1.63517874e-02 -6.60501063e-01 1.13728620e-01 -8.69514406e-01
-7.60999322e-01 4.82951961e-02 -3.81996095e-01 -6.56601727e-01
-3.72298092e-01 3.28529924e-01 2.00786904e-01 4.74738330e-02
8.05030286e-01 -1.51388109e-01 1.02538131e-01 -3.27774674e-01
8.52942020e-02 6.24936581e-01 1.18357348e+00 -6.04538023e-01
5.40266216e-01 5.15797675e-01 -2.10467771e-01 -6.01973236e-01
-9.44226384e-01 -7.33274817e-01 -4.82253820e-01 -1.05312932e+00
7.53893495e-01 -9.91725862e-01 -1.76287806e+00 5.39466619e-01
-6.58798218e-01 -7.63507664e-01 -3.03865582e-01 6.50353730e-01
-7.76549995e-01 3.15871239e-01 -5.38851202e-01 -6.56940520e-01
-3.12109411e-01 -9.01221156e-01 6.96585298e-01 6.03398979e-01
-1.00432575e+00 -6.64650023e-01 3.66507560e-01 9.26810443e-01
1.49313778e-01 1.55713959e-02 6.53955191e-02 -4.19469684e-01
-9.94030293e-03 -5.31485736e-01 1.29229173e-01 6.00821972e-01
-3.05172712e-01 7.27670779e-03 -8.91555250e-01 -1.47902533e-01
-4.50465918e-01 -8.74068916e-01 5.38294494e-01 6.06870532e-01
1.28208303e+00 1.44455299e-01 4.80713665e-01 3.28219593e-01
7.57419348e-01 -1.17415741e-01 8.70341659e-01 3.23596865e-01
4.39009696e-01 7.69728005e-01 8.73542011e-01 3.72507274e-01
4.25637692e-01 7.97008932e-01 6.65415898e-02 2.38902360e-01
-4.01454598e-01 -3.64409119e-01 5.40329278e-01 6.97414100e-01
-9.64499474e-01 -2.07472429e-01 -8.55780840e-01 4.56146657e-01
-1.79570448e+00 -1.33955026e+00 -2.23578304e-01 2.03048778e+00
6.55164540e-01 3.95882279e-01 9.93449688e-01 2.98956603e-01
4.78844881e-01 6.44896831e-03 -3.33106786e-01 -5.23655415e-01
2.12580234e-01 4.63110894e-01 4.15474176e-01 2.92216718e-01
-8.92288506e-01 1.01249313e+00 6.37825394e+00 1.07273650e+00
-1.24892604e+00 1.55194998e-01 3.79643798e-01 -7.45616198e-01
2.85490781e-01 5.19995987e-02 -5.05415559e-01 5.49377918e-01
8.03091168e-01 3.44350487e-02 3.64456087e-01 5.38480759e-01
4.53475952e-01 -4.85083342e-01 -9.18162704e-01 1.23195958e+00
1.06453441e-01 -8.81872296e-01 -7.50444114e-01 1.41551554e-01
1.77044570e-01 3.79244313e-02 1.52419761e-01 4.43901569e-01
4.67864722e-01 -1.17319822e+00 8.46130073e-01 1.06448603e+00
8.90956163e-01 -6.74593329e-01 6.41482174e-01 -2.63317265e-02
-1.07019079e+00 -1.20951772e-01 4.59984206e-02 -6.82618499e-01
-3.00110523e-02 -4.00119394e-01 -3.35453123e-01 -2.33889431e-01
1.02370167e+00 3.40393931e-01 -6.42806470e-01 1.12303150e+00
-1.75869673e-01 1.07207203e+00 -2.81621695e-01 -1.80098101e-01
1.13392442e-01 5.04713431e-02 7.08927214e-01 9.18347836e-01
4.40076590e-01 4.56145227e-01 1.85612589e-01 2.74239361e-01
1.51665121e-01 7.05473423e-02 -2.78614312e-01 -4.68893468e-01
5.00770628e-01 1.36249495e+00 -7.50611126e-01 -1.08807199e-01
-2.43732274e-01 6.80979490e-01 1.99796975e-01 6.30954206e-02
-4.52298224e-01 -4.62701797e-01 8.73292863e-01 5.60879588e-01
3.44971195e-02 -3.04902285e-01 -2.30192885e-01 -7.17718780e-01
6.31234050e-02 -1.16986215e+00 6.54714823e-01 -1.21451485e+00
-8.30738902e-01 2.74070323e-01 -2.47023150e-01 -1.54995763e+00
-3.19630295e-01 -8.40586543e-01 -9.18495119e-01 5.54718971e-01
-6.16738319e-01 -1.02803659e+00 -5.70917130e-01 5.03887415e-01
4.56060618e-01 -2.74708390e-01 5.24560332e-01 2.00339794e-01
-1.52771890e-01 6.78535581e-01 -5.40802717e-01 5.07282257e-01
9.20532823e-01 -1.20916498e+00 -4.38669249e-02 4.15809005e-01
3.01775396e-01 1.26074612e-01 1.10489142e+00 -3.08137894e-01
-1.08074760e+00 5.68523491e-03 4.29269910e-01 -1.08532858e+00
9.47930872e-01 7.73784295e-02 -5.33363044e-01 5.28207481e-01
2.61838377e-01 -4.44576502e-01 9.10883129e-01 5.50542831e-01
-4.42182869e-02 -6.31117895e-02 -6.54831350e-01 5.84483206e-01
1.24935138e+00 -5.52568674e-01 -5.76100886e-01 4.93840836e-02
-1.88657075e-01 -7.44287550e-01 -9.81228113e-01 1.01936705e-01
1.28767228e+00 -1.08123732e+00 8.46024573e-01 -6.31480098e-01
8.39878559e-01 -1.89593270e-01 1.67705521e-01 -1.22023654e+00
-4.52442855e-01 -5.96783400e-01 1.54606640e-01 1.15903544e+00
1.74245358e-01 8.01187679e-02 1.18275702e+00 5.43702185e-01
-9.31229384e-04 -5.37534058e-01 -7.85737574e-01 -4.19276625e-01
1.81937531e-01 -9.25612509e-01 1.01599313e-01 7.11158216e-01
6.62363589e-01 -1.43691733e-01 -5.89450479e-01 -4.66453508e-02
1.40427962e-01 -1.54334798e-01 1.13433444e+00 -1.04336071e+00
-7.08377004e-01 -7.77373672e-01 -9.38736498e-01 -5.69126964e-01
-2.84337997e-02 -5.40746808e-01 -1.69605657e-01 -1.24862850e+00
2.55723178e-01 1.58709049e-01 -3.08319986e-01 6.25191987e-01
1.31573871e-01 1.10067773e+00 5.16050935e-01 2.32080556e-02
-1.09110510e+00 -2.12169081e-01 1.43989778e+00 2.89617777e-01
-1.80361167e-01 2.48111680e-01 -1.00633872e+00 9.01296318e-01
9.16743934e-01 -1.03893824e-01 -1.96695313e-01 -1.67083457e-01
6.54338837e-01 3.48838001e-01 5.46934962e-01 -1.47087944e+00
1.63584694e-01 -2.41235688e-01 2.78375745e-01 -2.05986306e-01
6.98593736e-01 -2.79020786e-01 4.05475423e-02 1.76898062e-01
-5.07385075e-01 1.54014034e-02 3.95634800e-01 -4.06402759e-02
-3.65652621e-01 -2.65709549e-01 4.61591899e-01 -3.45841467e-01
-1.01789677e+00 3.55321132e-02 -9.84103739e-01 1.92869946e-01
6.32788777e-01 -4.38805193e-01 -2.14451969e-01 -1.05119038e+00
-1.17523849e+00 3.11770022e-01 3.32403332e-01 4.45501298e-01
2.71233231e-01 -1.35662806e+00 -7.49463916e-01 -3.86406213e-01
2.55073637e-01 -6.48799956e-01 6.66211963e-01 9.62599814e-01
-8.40606093e-01 -1.53382467e-02 -6.35440111e-01 -1.82716418e-02
-1.85289359e+00 -1.00596160e-01 2.84560889e-01 -2.77706701e-02
-4.13792312e-01 1.00872386e+00 -2.26875126e-01 -1.72451526e-01
2.26218402e-01 4.37025949e-02 -5.56893587e-01 2.31167406e-01
6.30909979e-01 6.15612745e-01 -3.02135706e-01 -8.70645523e-01
-1.21232346e-01 3.40087622e-01 2.48267531e-01 -4.61138368e-01
1.01020122e+00 -2.51817647e-02 4.33161139e-01 5.60408354e-01
5.15973687e-01 3.44456524e-01 -1.36614716e+00 1.31331161e-01
-4.64102834e-01 -7.68785954e-01 1.34394705e-01 -9.19194460e-01
-9.76991355e-01 8.95512521e-01 8.16593349e-01 2.86812812e-01
1.18805909e+00 2.35465184e-01 7.93034196e-01 3.43676120e-01
1.48344308e-01 -1.59750044e+00 3.57156634e-01 3.69723320e-01
7.83241808e-01 -1.24717915e+00 -1.45790458e-01 1.69017389e-02
-1.18775606e+00 1.00797522e+00 8.60395193e-01 -2.73835123e-01
4.68966603e-01 2.43779019e-01 6.00403666e-01 -4.72024918e-01
-6.91651642e-01 -8.61994922e-01 5.59650660e-01 6.27065420e-01
7.26122141e-01 2.18168557e-01 -3.24510634e-01 8.98279309e-01
-1.07950151e+00 3.60373110e-01 4.14668500e-01 7.42317736e-01
-4.04425234e-01 -9.49142158e-01 -3.77369940e-01 6.15634680e-01
-7.05290854e-01 2.83270717e-01 -7.21381605e-01 6.25379026e-01
5.16964436e-01 1.16792476e+00 2.74546951e-01 -7.31816709e-01
6.79293871e-01 1.92133524e-02 9.35470700e-01 -3.74416620e-01
-7.97367156e-01 -4.47761156e-02 5.50633311e-01 -7.09971189e-01
-4.33806509e-01 -1.04684126e+00 -8.23219836e-01 -7.06793129e-01
1.79713100e-01 -1.77122012e-01 3.27651441e-01 9.68269527e-01
6.08103722e-02 3.83352906e-01 -4.26909961e-02 -1.10107040e+00
-1.86068509e-02 -1.22801220e+00 -7.04451621e-01 4.89194006e-01
-1.53474048e-01 -5.04489541e-01 -1.15896299e-01 -3.43345292e-02] | [7.73577356338501, 0.26029351353645325] |
755d74c4-8202-40cb-b71c-98ef5b30ee59 | joint-3d-localization-and-classification-of | 1906.04749 | null | https://arxiv.org/abs/1906.04749v1 | https://arxiv.org/pdf/1906.04749v1.pdf | Joint 3D Localization and Classification of Space Debris using a Multispectral Rotating Point Spread Function | We consider the problem of joint three-dimensional (3D) localization and material classification of unresolved space debris using a multispectral rotating point spread function (RPSF). The use of RPSF allows one to estimate the 3D locations of point sources from their rotated images acquired by a single 2D sensor array, since the amount of rotation of each source image about its x, y location depends on its axial distance z. Using multi-spectral images, with one RPSF per spectral band, we are able not only to localize the 3D positions of the space debris but also classify their material composition. We propose a three-stage method for achieving joint localization and classification. In Stage 1, we adopt an optimization scheme for localization in which the spectral signature of each material is assumed to be uniform, which significantly improves efficiency and yields better localization results than possible with a single spectral band. In Stage 2, we estimate the spectral signature and refine the localization result via an alternating approach. We process classification in the final stage. Both Poisson noise and Gaussian noise models are considered, and the implementation of each is discussed. Numerical tests using multispectral data from NASA show the efficiency of our three-stage approach and illustrate the improvement of point source localization and spectral classification from using multiple bands over a single band. | ['Sudhakar Prasad', 'Grey Ballard', 'Chao Wang', 'Robert Plemmons'] | 2019-06-11 | null | null | null | null | ['material-classification'] | ['computer-vision'] | [ 1.98584393e-01 -6.68636680e-01 1.40647709e-01 3.41974854e-01
-1.08619499e+00 -8.78246307e-01 4.95215297e-01 -5.27823605e-02
-4.18337643e-01 7.17920661e-01 -1.32966086e-01 -7.08716959e-02
-6.22126579e-01 -6.83054686e-01 -5.40000498e-01 -1.00939631e+00
3.98941711e-03 4.72967714e-01 3.33322883e-01 2.37685874e-01
2.16567189e-01 1.15694642e+00 -1.62190235e+00 -3.02041888e-01
6.32814705e-01 1.16065753e+00 5.88788688e-01 8.04253221e-01
9.36379656e-03 1.99209243e-01 -8.49513710e-01 4.08670425e-01
5.96365869e-01 1.82017282e-01 -5.02512872e-01 5.40627539e-01
-1.11822695e-01 2.16490286e-03 -1.74305942e-02 1.16974103e+00
2.80757159e-01 2.63977051e-01 1.07047510e+00 -8.59457672e-01
-1.60541952e-01 -1.44828185e-01 -9.83589828e-01 7.34407306e-02
2.99147218e-01 -1.84920147e-01 5.66520810e-01 -8.97743881e-01
2.56908000e-01 9.15812194e-01 9.94941592e-01 -1.65610567e-01
-1.08941650e+00 -1.36078820e-01 -2.84801990e-01 2.15313192e-02
-1.94853151e+00 -2.06216678e-01 9.76857185e-01 -6.31668866e-01
5.55566728e-01 2.65889555e-01 4.82331663e-01 3.71405810e-01
1.20932586e-01 2.60681868e-01 1.16715991e+00 -6.79960430e-01
4.70066875e-01 -2.29884028e-01 7.88819641e-02 3.64691794e-01
6.73225701e-01 -6.67022318e-02 -2.77665406e-01 -5.68703473e-01
9.10692632e-01 -2.05740035e-02 -4.61077452e-01 -5.57721436e-01
-1.09234071e+00 6.89707577e-01 2.92290688e-01 3.08492959e-01
-7.71688640e-01 5.46666607e-02 -1.90047055e-01 -1.98721141e-01
7.33864129e-01 4.52421218e-01 -1.69894606e-01 2.02669010e-01
-9.36990440e-01 2.50317305e-01 4.60403889e-01 7.03771055e-01
8.98912549e-01 -3.17751952e-02 3.89317900e-01 1.09324622e+00
4.10771698e-01 1.42686260e+00 8.14378783e-02 -1.20525837e+00
1.20771766e-01 2.25436255e-01 6.71387076e-01 -8.91576469e-01
-5.24711728e-01 -5.59998870e-01 -4.79749799e-01 5.72441518e-01
4.61578906e-01 -1.35456935e-01 -8.99794877e-01 1.20099568e+00
3.72113198e-01 2.37194777e-01 1.18087105e-01 9.87249732e-01
-8.24397579e-02 8.12586904e-01 -2.89962709e-01 -3.94414157e-01
1.38407540e+00 -2.45066181e-01 -4.58379209e-01 -4.19766605e-01
2.75983125e-01 -8.64090979e-01 3.42405826e-01 3.76942873e-01
-8.40544105e-01 -2.39552349e-01 -1.00204408e+00 6.07533872e-01
-3.50006729e-01 5.49315453e-01 4.49189156e-01 5.82194805e-01
-8.17361057e-01 1.99102327e-01 -9.49930370e-01 -2.00295523e-01
2.38737434e-01 -4.62549701e-02 -2.76249796e-01 -1.04044322e-02
-5.77797055e-01 8.32319319e-01 2.71058865e-02 1.22567199e-01
-3.14747125e-01 -5.45298278e-01 -6.76284075e-01 -2.35672757e-01
1.93614662e-01 -4.68482286e-01 1.02554965e+00 -5.08706093e-01
-1.05240631e+00 5.71541488e-01 -4.29897696e-01 -1.21810384e-01
1.37432754e-01 -1.70587078e-01 -5.14568448e-01 5.91027737e-01
6.46194816e-01 2.14589876e-03 9.18720484e-01 -1.71021175e+00
-7.45692134e-01 -4.62127388e-01 -4.19695109e-01 2.55171031e-01
1.91928133e-01 -3.21885361e-03 -5.52398860e-02 -4.50043589e-01
7.25965977e-01 -1.03164995e+00 -2.17212394e-01 -7.06065819e-02
-2.75180459e-01 9.12882835e-02 7.09907949e-01 -6.98260188e-01
6.64725244e-01 -2.48299932e+00 -4.23857905e-02 4.81957525e-01
-4.57567833e-02 -1.30520627e-01 1.18388303e-01 1.98772281e-01
-1.51379555e-01 -1.69772774e-01 -7.11358547e-01 -3.34293842e-01
-3.82972747e-01 -9.49475095e-02 -2.70266593e-01 1.06703973e+00
2.63395440e-02 2.94356227e-01 -7.24535286e-01 -5.86628169e-02
3.12999636e-01 3.40623647e-01 -2.47800956e-03 -8.02160054e-02
1.38376057e-01 8.55453759e-02 -5.06373823e-01 8.36449742e-01
1.25892675e+00 -3.01750563e-02 -2.77636945e-01 -4.90796149e-01
-5.39866149e-01 -2.84172565e-01 -1.65332353e+00 1.43117642e+00
-5.05349815e-01 4.55848724e-01 6.69186354e-01 -8.33630085e-01
1.01259005e+00 2.27648422e-01 8.37176085e-01 -1.96218818e-01
-8.94366875e-02 3.29766691e-01 -5.28320074e-01 -3.67581040e-01
7.06512749e-01 -3.83696169e-01 -1.73157007e-01 3.90280724e-01
-1.78052038e-01 -4.65723574e-01 -1.85202733e-01 -2.35825643e-01
1.10223496e+00 -2.32515946e-01 2.08349705e-01 -3.27338547e-01
3.51009727e-01 3.26971740e-01 2.22717524e-01 8.65199506e-01
-5.18511720e-02 8.17157924e-01 -2.11550221e-02 4.50104102e-02
-8.97611439e-01 -1.45708227e+00 -2.19507977e-01 3.55998129e-01
4.37853754e-01 2.24557146e-01 -4.11288381e-01 -3.79854232e-01
2.89927691e-01 7.59625733e-01 -3.03336918e-01 1.42672643e-01
-4.26071525e-01 -1.24489951e+00 2.27447733e-01 2.73719102e-01
4.13581342e-01 -2.52264857e-01 -6.51673257e-01 7.50414953e-02
-2.70183623e-01 -1.05342948e+00 1.30290568e-01 3.47182095e-01
-6.89541638e-01 -1.26431572e+00 -7.55447686e-01 -3.90047044e-01
7.05809653e-01 1.13732827e+00 5.59046149e-01 -4.31105793e-01
-2.84013689e-01 1.00602078e+00 -4.89642888e-01 -3.98935765e-01
-1.47268146e-01 -3.15432906e-01 3.08936179e-01 2.24806771e-01
7.29834437e-02 -4.02776867e-01 -3.42974216e-01 4.75989908e-01
-7.91107774e-01 -4.41322893e-01 6.10280216e-01 4.52048182e-01
6.40465498e-01 7.55118906e-01 3.75656366e-01 -9.46727097e-02
4.14052039e-01 -6.27282083e-01 -7.85333514e-01 1.73805788e-01
-9.35311168e-02 -1.25713930e-01 2.04899669e-01 -1.21499188e-01
-1.17108500e+00 3.41124266e-01 2.25129932e-01 -5.00437915e-01
-5.34311056e-01 4.45433766e-01 6.89103231e-02 -5.03496528e-01
9.28019822e-01 2.03127429e-01 1.64341226e-01 -6.41952932e-01
3.11319977e-01 9.04636562e-01 8.74499679e-01 -5.85091472e-01
1.17062223e+00 7.70615458e-01 2.12871030e-01 -1.41303098e+00
-6.74689651e-01 -1.04680634e+00 -6.95427597e-01 -5.37397444e-01
7.63960958e-01 -1.03837001e+00 -4.15658325e-01 6.42632186e-01
-1.39947772e+00 3.91521938e-02 -3.36635172e-01 9.48575795e-01
-5.90499640e-01 3.99716675e-01 -2.22487794e-03 -1.30237985e+00
1.13736935e-01 -8.81137192e-01 1.49723673e+00 -9.11991969e-02
1.53342441e-01 -9.59140122e-01 5.07545024e-02 1.78196624e-01
2.57899314e-01 8.99512246e-02 5.72686672e-01 -3.93959880e-02
-4.96675819e-01 -5.05265832e-01 -2.64107317e-01 1.89637423e-01
4.87152845e-01 -3.21321160e-01 -8.34867060e-01 -2.36507803e-01
6.21454060e-01 3.31560403e-01 6.47890508e-01 9.07343745e-01
8.37899625e-01 2.92465895e-01 -5.24634421e-01 5.47124505e-01
1.75526941e+00 7.28400052e-02 3.05168241e-01 4.97753501e-01
4.37304497e-01 4.73451614e-01 9.35247540e-01 4.60525930e-01
1.89965963e-02 6.18688643e-01 5.77402771e-01 6.82757422e-02
-2.48765469e-01 4.25997972e-01 1.44191757e-01 2.55314559e-01
-4.88880545e-01 -1.31771088e-01 -9.77463543e-01 6.00362957e-01
-1.60347021e+00 -9.92422044e-01 -4.86142159e-01 2.41230273e+00
7.14485301e-03 -3.60528529e-01 5.42291626e-02 3.32857907e-01
1.11066020e+00 8.46004188e-02 -2.79208481e-01 4.66981322e-01
-3.89690816e-01 2.23868247e-02 1.40542638e+00 7.25247920e-01
-1.25482428e+00 2.34538108e-01 7.10925341e+00 8.04677665e-01
-9.60353792e-01 6.73152208e-02 3.63914436e-03 2.79710032e-02
-1.30577177e-01 -2.11744055e-01 -6.68712616e-01 4.18434411e-01
6.38481200e-01 -1.32923797e-01 7.14351118e-01 7.21803963e-01
2.97461420e-01 -8.23402226e-01 -2.87953049e-01 1.22717512e+00
6.72985539e-02 -1.13050807e+00 -3.04473430e-01 2.23127127e-01
4.54376578e-01 2.22436115e-01 -2.47016639e-01 -4.47643816e-01
7.36192092e-02 -3.79517257e-01 1.16456497e+00 7.89066076e-01
6.78938806e-01 -7.37524271e-01 6.64621890e-01 4.08005536e-01
-1.40839601e+00 -2.96647280e-01 -3.90963316e-01 5.46913818e-02
4.29115683e-01 1.05457163e+00 -9.31879282e-01 9.83783603e-01
7.05938697e-01 3.32743436e-01 -3.87583315e-01 1.45354187e+00
-1.20014481e-01 4.13152367e-01 -7.12568641e-01 1.89408571e-01
-3.76224369e-02 -6.10052347e-01 1.04792786e+00 1.03275836e+00
1.16710818e+00 2.47328371e-01 1.49353758e-01 7.19697654e-01
5.55547774e-01 -3.39650720e-01 -7.04250872e-01 3.73276114e-01
7.80813634e-01 1.23767889e+00 -9.27625239e-01 2.36626714e-02
-2.50605017e-01 6.57904923e-01 -2.92870849e-01 4.77371037e-01
-4.90195781e-01 -3.70882988e-01 6.69797778e-01 3.26757044e-01
3.54055882e-01 -8.73178184e-01 -6.05574012e-01 -6.84275389e-01
5.53328022e-02 -3.25484246e-01 1.24786198e-01 -1.16627657e+00
-1.26385963e+00 2.17145309e-01 3.77353549e-01 -1.59844291e+00
-1.59504294e-01 -8.48029017e-01 -2.96690851e-01 1.26692545e+00
-1.25028503e+00 -1.02226424e+00 -4.36099410e-01 4.44188952e-01
1.85631737e-01 -1.55910596e-01 7.52413929e-01 -4.99827489e-02
-1.02415703e-01 -4.03940767e-01 4.49138761e-01 -2.97142863e-01
4.12529260e-01 -1.20806801e+00 -6.40998185e-02 9.23378348e-01
-2.09538013e-01 1.81793943e-01 8.85726511e-01 -9.17523026e-01
-1.47243381e+00 -1.18435180e+00 2.75204569e-01 -4.40501004e-01
8.22364569e-01 -1.13917828e-01 -5.85585892e-01 3.53016108e-01
-3.29916120e-01 1.23267267e-02 6.19780779e-01 -1.21298142e-01
6.28336146e-02 -2.01878786e-01 -1.32825780e+00 8.60640705e-02
7.12169170e-01 -6.30041182e-01 -6.47257864e-01 5.70314825e-01
2.33488709e-01 -1.96639419e-01 -6.91730618e-01 4.38352197e-01
2.07904652e-01 -7.18812883e-01 1.29708195e+00 3.64885300e-01
-2.50929266e-01 -9.49927211e-01 -5.82391322e-01 -1.59240878e+00
-5.91580749e-01 -5.47671542e-02 2.92288512e-01 9.80899990e-01
2.60223448e-01 -7.62376428e-01 6.72989368e-01 2.27533486e-02
-3.70052010e-01 1.06996447e-01 -1.20951879e+00 -1.20838571e+00
-2.53277719e-01 -6.85388029e-01 3.96147102e-01 5.82612574e-01
-3.63148421e-01 -2.66970042e-02 2.31504701e-02 1.23520160e+00
1.11105323e+00 2.56469935e-01 5.11846900e-01 -1.53634620e+00
-1.95135117e-01 -1.16627768e-01 -4.17737275e-01 -8.44015241e-01
1.69547707e-01 -6.84785008e-01 3.63695353e-01 -1.56857181e+00
-1.05375394e-01 -7.42077947e-01 1.05368763e-01 3.44231762e-02
1.38171434e-01 1.92374676e-01 -2.06340738e-02 5.37033498e-01
-6.16815723e-02 2.81667650e-01 5.38608611e-01 -1.70106471e-01
-1.61319926e-01 2.14365780e-01 -2.37296596e-01 8.46237898e-01
6.46598041e-01 -3.91013294e-01 -9.13861468e-02 -6.34677529e-01
-5.19155525e-02 1.70240879e-01 8.53991508e-01 -1.50001287e+00
3.42950106e-01 -2.31723920e-01 5.42482018e-01 -9.10732985e-01
8.31582606e-01 -1.11059105e+00 5.38276196e-01 2.19662651e-01
4.68300074e-01 -2.90766925e-01 9.53925624e-02 7.56927073e-01
-5.75880446e-02 -7.12668955e-01 9.44252908e-01 -1.00956835e-01
-6.41721547e-01 -4.84812595e-02 -6.28618479e-01 -7.54707098e-01
1.07699430e+00 -1.73923269e-01 -3.96638155e-01 -2.30824679e-01
-7.36684978e-01 -2.73202032e-01 7.41719604e-01 -5.04392199e-02
4.70317334e-01 -1.39516914e+00 -5.56003153e-01 1.85951591e-01
6.49545863e-02 3.52269448e-02 2.41394937e-01 6.95979416e-01
-6.16971552e-01 2.49877691e-01 9.40448344e-02 -8.89919341e-01
-9.37594593e-01 3.97142321e-01 4.92084503e-01 2.16322616e-01
-2.28205010e-01 6.30614996e-01 -1.94247857e-01 -2.40545049e-01
-3.47225517e-01 5.13634905e-02 -9.42942426e-02 2.00992286e-01
5.81783831e-01 7.01959014e-01 2.92381853e-01 -9.64443862e-01
-3.68724048e-01 1.03603041e+00 6.36544228e-01 -3.73749197e-01
1.50574267e+00 -2.36662567e-01 -3.64155293e-01 5.32000840e-01
8.59612763e-01 4.64587867e-01 -1.18446624e+00 -1.10265993e-01
-1.57367229e-01 -7.28091836e-01 3.00950557e-01 -5.90373218e-01
-7.31745899e-01 3.50916773e-01 6.18586481e-01 7.08735049e-01
1.21092391e+00 1.33176178e-01 1.23789005e-01 2.66679794e-01
6.88335299e-01 -9.97396886e-01 -3.93940628e-01 4.55270410e-01
8.44821751e-01 -6.69077098e-01 7.37024248e-02 -8.47199500e-01
-8.32590237e-02 1.24582219e+00 -1.98795460e-02 -1.72487631e-01
7.27334797e-01 3.22160929e-01 -1.21114954e-01 -2.31968746e-01
1.03690520e-01 -3.64571899e-01 1.19630359e-02 8.42140734e-01
-2.21738741e-01 2.75390238e-01 -9.22049768e-03 2.68889606e-01
1.76726356e-02 -2.50909150e-01 5.64769506e-01 1.04051352e+00
-1.01263642e+00 -7.07108617e-01 -1.56495190e+00 4.11576062e-01
-9.96630732e-03 3.51393998e-01 -4.31373417e-02 4.37699407e-01
1.80071428e-01 1.07355940e+00 2.88180530e-01 -1.41746312e-01
4.46364611e-01 2.86770314e-02 4.54328805e-01 -4.09858614e-01
3.58970255e-01 3.39699328e-01 2.25622281e-01 -3.59260410e-01
-6.30309761e-01 -1.24889505e+00 -1.06821239e+00 7.06706122e-02
-4.60877746e-01 4.74322885e-01 1.33245838e+00 7.94583619e-01
1.64681599e-01 2.36699924e-01 1.14472520e+00 -1.37115896e+00
-5.30329466e-01 -7.30795085e-01 -1.30041480e+00 -3.79194975e-01
5.00167370e-01 -1.17416000e+00 -8.98491621e-01 2.02908497e-02] | [10.071433067321777, -2.079148530960083] |
3f5e691d-b0b0-4948-ac85-22efc527b7b7 | semi-supervised-3d-face-reconstruction-with | null | null | https://openreview.net/forum?id=H1lK5kBKvr | https://openreview.net/pdf?id=H1lK5kBKvr | Semi-supervised 3D Face Reconstruction with Nonlinear Disentangled Representations | Recovering 3D geometry shape, albedo and lighting from a single image has wide applications in many areas, which is also a typical ill-posed problem. In order to eliminate the ambiguity, face prior knowledge like linear 3D morphable models (3DMM) learned from limited scan data are often adopted to the reconstruction process. However, methods based on linear parametric models cannot generalize well for facial images in the wild with various ages, ethnicity, expressions, poses, and lightings. Recent methods aim to learn a nonlinear parametric model using convolutional neural networks (CNN) to regress the face shape and texture directly. However, the models were only trained on a dataset that is generated from a linear 3DMM. Moreover, the identity and expression representations are entangled in these models, which hurdles many facial editing applications. In this paper, we train our model with adversarial loss in a semi-supervised manner on hybrid batches of unlabeled and labeled face images to exploit the value of large amounts of unlabeled face images from unconstrained photo collections. A novel center loss is introduced to make sure that different facial images from the same person have the same identity shape and albedo. Besides, our proposed model disentangles identity, expression, pose, and lighting representations, which improves the overall reconstruction performance and facilitates facial editing applications, e.g., expression transfer. Comprehensive experiments demonstrate that our model produces high-quality reconstruction compared to state-of-the-art methods and is robust to various expression, pose, and lighting conditions.
| ['Xiaokang Yang', 'Guangtao Zhai', 'Chao Ma', 'Yudong Guo', 'Juyong Zhang', 'Zhongpai Gao'] | 2019-09-25 | null | null | null | null | ['3d-face-reconstruction', 'facial-editing', 'face-reconstruction'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 4.08606008e-02 -2.27733105e-02 -1.49001330e-01 -6.48539186e-01
-3.10391456e-01 -3.68372947e-01 3.86654764e-01 -8.35349143e-01
-5.93626276e-02 5.21172822e-01 -6.83724461e-03 3.57444495e-01
2.25641280e-01 -6.44883990e-01 -8.47588241e-01 -8.79491270e-01
4.05385822e-01 3.37922722e-01 -4.87980992e-01 -2.31070980e-01
-1.85981825e-01 8.83820295e-01 -1.44793046e+00 -1.65046364e-01
7.99499035e-01 1.10826838e+00 -1.99431956e-01 3.86834033e-02
-1.48141682e-01 4.06272411e-01 -3.63194704e-01 -6.83680892e-01
5.83392203e-01 -5.07918596e-01 -1.14282593e-01 5.14483809e-01
7.70845711e-01 -5.31243026e-01 -5.51251173e-01 1.16315758e+00
4.40099776e-01 -7.65056610e-02 7.18390644e-01 -1.30233335e+00
-9.85737026e-01 -2.37880602e-01 -8.28675508e-01 -6.97541475e-01
2.57146657e-01 1.22092098e-01 4.46954846e-01 -1.08211541e+00
4.68716055e-01 1.49862838e+00 6.61154568e-01 7.74924934e-01
-1.37209547e+00 -1.11650872e+00 4.15859781e-02 -9.21555459e-02
-1.63688815e+00 -6.76790714e-01 1.08488524e+00 -2.84268439e-01
9.10485387e-02 8.31571221e-02 6.60561681e-01 1.17221045e+00
-1.20048009e-01 4.48046029e-01 1.26184416e+00 -2.70840228e-01
-1.42955035e-01 1.97306931e-01 -7.47327626e-01 1.04952335e+00
2.36047432e-03 6.71572378e-03 -6.61615849e-01 5.91285452e-02
1.26224065e+00 1.68188229e-01 -3.21772009e-01 -5.81886232e-01
-8.64448786e-01 6.77675128e-01 3.62714529e-01 -2.32682750e-01
-2.55139202e-01 -9.07840505e-02 -4.84431768e-03 2.60289073e-01
5.62992573e-01 1.11790136e-01 -2.14029416e-01 3.21385980e-01
-7.31629670e-01 2.13237211e-01 6.40949190e-01 1.01029575e+00
1.20542502e+00 3.66779476e-01 1.24302618e-01 1.12690842e+00
4.39990610e-01 1.06409168e+00 1.78741276e-01 -1.09023178e+00
2.16856197e-01 5.86187065e-01 -3.91785577e-02 -1.18702328e+00
7.07906764e-03 -7.34013617e-02 -1.11536169e+00 3.96315515e-01
2.69334197e-01 -5.91092370e-02 -9.68519092e-01 2.01796842e+00
5.90333521e-01 1.42410398e-01 -1.41416624e-01 1.09166193e+00
8.39129567e-01 4.63258892e-01 -1.94315746e-01 -2.90318191e-01
1.09511316e+00 -6.45040274e-01 -7.57299662e-01 -2.50086397e-01
-4.10540812e-02 -9.57599580e-01 9.76052701e-01 1.68731913e-01
-9.80563641e-01 -4.73692268e-01 -9.05114055e-01 -2.83478320e-01
1.49414033e-01 2.41489917e-01 4.87832010e-01 6.40712857e-01
-7.35244632e-01 3.24734837e-01 -6.57039523e-01 -1.62871450e-01
6.03377879e-01 4.11489934e-01 -7.58540630e-01 -4.07550961e-01
-9.38028514e-01 7.03322172e-01 -3.46798778e-01 4.89570290e-01
-8.24195206e-01 -5.99768460e-01 -9.99096453e-01 -2.78014749e-01
2.05671802e-01 -5.82153201e-01 6.72104955e-01 -1.34886563e+00
-2.02860761e+00 1.19221902e+00 -2.27872357e-01 3.49567682e-01
6.50294840e-01 -1.16043374e-01 -3.10074836e-01 -5.02605829e-03
1.22445310e-02 5.77958703e-01 1.30972540e+00 -1.28369570e+00
9.03203636e-02 -7.24626720e-01 -1.16602011e-01 2.92402953e-01
-4.16284502e-01 5.92013039e-02 -7.67370164e-01 -6.03518188e-01
2.37947956e-01 -1.04032958e+00 1.41896918e-01 7.32705832e-01
-2.90093124e-01 3.26364547e-01 7.94183075e-01 -8.46416414e-01
4.36924130e-01 -2.19171667e+00 3.51762593e-01 2.25012824e-01
-6.70723766e-02 1.13291442e-01 -3.28623205e-01 2.35320833e-02
-4.72365208e-02 9.78534226e-04 -2.13374496e-01 -4.84045833e-01
1.70584563e-02 4.14939374e-01 -2.01918051e-01 7.50771642e-01
4.13471580e-01 8.24668765e-01 -5.20208538e-01 -5.58897138e-01
1.23274900e-01 8.89951050e-01 -5.08868933e-01 5.51389098e-01
-1.83590740e-01 1.07871449e+00 -4.46499258e-01 8.90182078e-01
1.01709366e+00 1.30203158e-01 4.02677655e-02 -4.01449680e-01
2.34441161e-01 -3.72830302e-01 -1.02437329e+00 1.91349733e+00
-6.01914704e-01 4.83229756e-01 3.78710032e-01 -9.26659882e-01
1.17856336e+00 3.07414770e-01 5.39262176e-01 -7.67345190e-01
2.25961193e-01 1.56119049e-01 -3.24030995e-01 -5.93864918e-01
-9.55026655e-04 -5.23615003e-01 1.48096427e-01 2.13551313e-01
1.32377133e-01 -4.27336663e-01 -3.67667794e-01 -3.80589753e-01
4.34942424e-01 3.99706542e-01 -3.72928642e-02 9.59041417e-02
4.76022512e-01 -6.35497272e-01 9.23607409e-01 -9.36312750e-02
1.06278397e-01 9.13893938e-01 3.99744958e-01 -4.90915805e-01
-1.11711287e+00 -1.00852823e+00 -1.77903131e-01 7.64132798e-01
2.54979461e-01 9.14408416e-02 -7.17981756e-01 -4.47827935e-01
-5.11696655e-03 1.34152532e-01 -6.18867338e-01 -2.61769384e-01
-5.28237104e-01 -6.22147799e-01 5.97063363e-01 1.72457471e-01
6.87950194e-01 -7.20918894e-01 3.67631987e-02 -2.47551262e-01
-1.75476283e-01 -1.31738007e+00 -7.30788827e-01 -6.57230377e-01
-5.97646415e-01 -9.94100273e-01 -7.70569265e-01 -6.76355839e-01
1.16748405e+00 9.50946957e-02 7.11339653e-01 4.11828235e-03
-2.52881676e-01 2.73463339e-01 -8.25312361e-02 -4.47992295e-01
-2.74486989e-01 -3.87136549e-01 2.87676513e-01 7.91149020e-01
3.52186039e-02 -8.73766720e-01 -6.41777813e-01 5.94234526e-01
-9.16617334e-01 1.83111548e-01 6.13845706e-01 8.05897653e-01
7.30722129e-01 -1.76452085e-01 3.48671108e-01 -8.35646868e-01
8.74752924e-02 -2.64306992e-01 -6.17207408e-01 2.74612218e-01
-5.26370108e-01 -3.77092287e-02 5.64174354e-01 -6.43941343e-01
-1.32750297e+00 2.26721272e-01 -1.40839040e-01 -9.07618105e-01
-7.05073848e-02 5.26439212e-02 -8.45704377e-01 -5.50842583e-01
5.02050579e-01 1.77878723e-01 5.51067054e-01 -3.79831433e-01
3.77308697e-01 4.19182003e-01 5.94133914e-01 -6.94228768e-01
1.23342073e+00 7.29459524e-01 2.90028960e-01 -8.79409432e-01
-8.40087116e-01 5.25825657e-02 -7.12276161e-01 -2.34596297e-01
5.99868894e-01 -1.16645646e+00 -5.29725194e-01 8.89964581e-01
-1.10354698e+00 -1.80635944e-01 1.86178908e-02 3.98065001e-01
-3.99909377e-01 2.58798033e-01 -4.32102114e-01 -6.52206659e-01
-2.81260908e-01 -1.05947649e+00 1.07503068e+00 4.45781708e-01
1.85234874e-01 -7.26427853e-01 -2.64865100e-01 5.45322120e-01
3.61903906e-01 6.19874537e-01 8.08344543e-01 1.98724315e-01
-6.99036300e-01 -2.13485911e-01 -2.18219370e-01 6.18576705e-01
4.50601846e-01 3.98410670e-02 -1.05127037e+00 -2.50510782e-01
-4.96875262e-03 -5.70210040e-01 4.10567790e-01 4.42843400e-02
1.20196950e+00 -4.54989672e-01 1.40503228e-01 1.23259306e+00
1.14138651e+00 -1.98522493e-01 6.33064687e-01 -1.74272105e-01
1.08841968e+00 8.04423630e-01 3.87403846e-01 4.57575947e-01
3.06387544e-01 6.86327636e-01 5.82970500e-01 -2.95899123e-01
-1.23487540e-01 -4.39005166e-01 4.24626201e-01 6.92675054e-01
-3.43198240e-01 1.19681112e-01 -4.33651835e-01 1.25501424e-01
-1.48578310e+00 -7.22284734e-01 3.76380891e-01 2.27997565e+00
1.03483248e+00 -4.31907535e-01 -3.51283520e-01 -3.34865808e-01
7.04473555e-01 1.85363501e-01 -9.22539592e-01 8.54860526e-03
-3.32986921e-01 4.25806433e-01 3.11486959e-01 1.84023649e-01
-7.81736016e-01 9.12563145e-01 5.32351637e+00 7.45419025e-01
-1.54087985e+00 3.67360376e-02 6.87991261e-01 -1.96076646e-01
-4.91368145e-01 -1.09580211e-01 -5.12740374e-01 3.21582556e-01
3.42393279e-01 2.26577446e-02 7.77895331e-01 6.67120397e-01
2.07632571e-01 3.90279979e-01 -9.71607089e-01 1.31354129e+00
4.89575446e-01 -8.60453546e-01 1.45263001e-01 2.29411080e-01
9.84987617e-01 -3.52689505e-01 2.88213402e-01 1.54892638e-01
1.04471261e-03 -1.34911478e+00 6.96883202e-01 7.23267794e-01
1.32427979e+00 -6.61679506e-01 4.00033504e-01 2.30334461e-01
-8.48083675e-01 3.04037511e-01 -3.84790897e-01 1.77086473e-01
8.02271217e-02 4.81919616e-01 -4.08949196e-01 5.36076009e-01
6.24231040e-01 6.55594289e-01 -2.10860714e-01 4.18529510e-01
-5.25385141e-01 3.17662269e-01 -2.85125256e-01 5.10559797e-01
-3.12434942e-01 -7.09410250e-01 3.53588760e-01 5.17726302e-01
5.22728503e-01 1.93621486e-01 8.56147334e-02 1.10582614e+00
-5.45071542e-01 2.24065855e-01 -6.33036017e-01 5.12231607e-03
3.50859195e-01 1.35617125e+00 -1.18004583e-01 2.95702368e-01
-5.54485440e-01 1.09339404e+00 3.53861123e-01 5.33579826e-01
-7.76298583e-01 4.15381528e-02 9.26777840e-01 3.08892488e-01
2.30019800e-02 -2.77896792e-01 -4.09145802e-02 -1.31534648e+00
1.04628906e-01 -9.16894078e-01 -2.35081807e-01 -7.59198546e-01
-1.49100053e+00 4.22777474e-01 -1.73557833e-01 -1.24771869e+00
-1.49529114e-01 -5.77382207e-01 -5.92564166e-01 1.04628956e+00
-1.68302202e+00 -1.60838699e+00 -6.02350354e-01 7.25724697e-01
1.69771612e-01 -2.75870562e-01 8.40998530e-01 3.75333101e-01
-7.43586600e-01 8.21703553e-01 -7.67999142e-02 4.10480291e-01
1.10798895e+00 -7.28249729e-01 -2.39596684e-02 6.61546469e-01
7.84146637e-02 6.05731368e-01 3.14890563e-01 -2.35065907e-01
-1.81790912e+00 -1.30291402e+00 2.64459372e-01 -1.53156653e-01
1.17686994e-01 -3.99158061e-01 -9.81857181e-01 6.67589009e-01
-2.10072070e-01 5.46845973e-01 7.33955681e-01 -5.56089133e-02
-6.23997211e-01 -5.07970095e-01 -1.11855280e+00 5.89662790e-01
1.11158526e+00 -7.35241652e-01 -5.65328859e-02 2.91082710e-01
4.07615066e-01 -7.27502227e-01 -9.70359981e-01 5.39050579e-01
8.81829262e-01 -9.42282736e-01 9.13353264e-01 -4.12064224e-01
4.16571915e-01 -2.58714914e-01 -2.31496930e-01 -1.28981626e+00
8.73819366e-02 -7.11351395e-01 5.26461080e-02 1.41819680e+00
3.50017585e-02 -6.95395768e-01 6.68697953e-01 8.56702924e-01
1.41822889e-01 -7.42317915e-01 -8.07432652e-01 -6.15524471e-01
1.00935385e-01 -7.78655931e-02 8.98671210e-01 1.09117520e+00
-7.76346207e-01 2.92480916e-01 -7.92964578e-01 2.91433066e-01
7.50248075e-01 3.67919892e-01 1.22484469e+00 -1.19770992e+00
2.76769660e-02 -2.15803847e-01 -4.19235080e-01 -9.22953546e-01
7.88631618e-01 -7.96495736e-01 -4.94850837e-02 -9.59737897e-01
1.34478778e-01 -5.12411833e-01 9.47043598e-02 6.25125527e-01
-3.58973891e-02 5.56372702e-01 8.47182646e-02 3.39687318e-01
-5.15263178e-04 1.06639624e+00 1.57009971e+00 -1.69167161e-01
2.34371610e-02 -1.08259633e-01 -6.71488583e-01 7.64830530e-01
5.97446263e-01 -2.73529381e-01 -4.97532845e-01 -7.74836302e-01
1.15678184e-01 4.59138453e-02 4.56433594e-01 -6.53590322e-01
3.40707935e-02 -4.95933563e-01 6.60950661e-01 7.65906796e-02
6.61671042e-01 -8.12896073e-01 4.44243789e-01 -1.03124194e-01
2.23480854e-02 -3.44523847e-01 2.93275639e-02 5.10839164e-01
-3.33344907e-01 1.81778669e-01 9.89636004e-01 -1.58658639e-01
-4.13720518e-01 1.06003499e+00 4.90602553e-01 3.37526575e-02
8.45396578e-01 -2.11635992e-01 6.70375153e-02 -5.99061251e-01
-3.53713930e-01 4.66619842e-02 9.05989349e-01 4.61325288e-01
6.58184171e-01 -1.65346026e+00 -8.31237376e-01 6.51486635e-01
2.15554535e-01 4.58679587e-01 3.30703676e-01 7.15966403e-01
-5.34639060e-01 -2.85727382e-01 -4.43925261e-01 -4.93873119e-01
-1.17704165e+00 2.59208739e-01 4.77356344e-01 2.82518178e-01
-3.01985085e-01 6.08854592e-01 4.42870796e-01 -8.17251325e-01
-1.71227902e-02 2.69425124e-01 1.33944646e-01 -1.83788285e-01
3.64918083e-01 -1.86524950e-02 -1.57752916e-01 -1.10267806e+00
-6.79055750e-02 1.04993367e+00 2.28395447e-01 4.54958677e-02
1.25605905e+00 -1.15864925e-01 -3.62913430e-01 2.17870370e-01
1.39938736e+00 2.79420137e-01 -1.58761787e+00 -4.74989653e-01
-7.94951320e-01 -8.03394914e-01 -3.57638299e-02 -2.93467909e-01
-1.60751152e+00 9.69121873e-01 4.23249722e-01 -6.21856511e-01
1.21400785e+00 -1.68267921e-01 7.76325762e-01 1.51109338e-01
4.00493890e-01 -9.10718799e-01 2.11889878e-01 2.96750575e-01
1.17077339e+00 -1.39288521e+00 1.21175088e-01 -5.43172657e-01
-5.19879580e-01 1.15556335e+00 6.93063319e-01 4.24719453e-02
6.86249077e-01 1.28614185e-02 2.75141567e-01 4.53520864e-02
-8.77777785e-02 1.34538531e-01 2.99944907e-01 6.36443317e-01
2.23545998e-01 6.05648570e-02 2.09983453e-01 4.83709306e-01
-2.62499571e-01 -1.08572468e-01 6.26645684e-02 5.86879134e-01
8.87105539e-02 -1.15693045e+00 -4.71020579e-01 1.11735560e-01
-3.40581834e-01 2.41680428e-01 -2.89023668e-01 7.55601943e-01
2.77700096e-01 4.96424735e-01 -3.84249315e-02 -2.99919635e-01
3.62451077e-01 -3.98196690e-02 8.05574298e-01 -5.12469351e-01
1.87211558e-01 6.14543930e-02 -3.71057481e-01 -5.16798973e-01
-5.16110599e-01 -5.21586359e-01 -1.08939612e+00 -4.42526758e-01
-1.54748201e-01 -3.10399890e-01 7.54436851e-01 7.40709722e-01
3.69370729e-01 2.86086220e-02 1.13798440e+00 -1.03820109e+00
-4.84022528e-01 -8.58729720e-01 -8.77885520e-01 7.46910393e-01
2.60799170e-01 -9.18830693e-01 -2.53550678e-01 3.51525992e-01] | [12.911275863647461, -0.03857969492673874] |
adfe4088-191d-4163-9257-51d3ce1636dc | diffuse-map-guiding-unsupervised-generative | 2205.11951 | null | https://arxiv.org/abs/2205.11951v2 | https://arxiv.org/pdf/2205.11951v2.pdf | Diffuse Map Guiding Unsupervised Generative Adversarial Network for SVBRDF Estimation | Reconstructing materials in the real world has always been a difficult problem in computer graphics. Accurately reconstructing the material in the real world is critical in the field of realistic rendering. Traditionally, materials in computer graphics are mapped by an artist, then mapped onto a geometric model by coordinate transformation, and finally rendered with a rendering engine to get realistic materials. For opaque objects, the industry commonly uses physical-based bidirectional reflectance distribution function (BRDF) rendering models for material modeling. The commonly used physical-based rendering models are Cook-Torrance BRDF, Disney BRDF. In this paper, we use the Cook-Torrance model to reconstruct the materials. The SVBRDF material parameters include Normal, Diffuse, Specular and Roughness. This paper presents a Diffuse map guiding material estimation method based on the Generative Adversarial Network(GAN). This method can predict plausible SVBRDF maps with global features using only a few pictures taken by the mobile phone. The main contributions of this paper are: 1) We preprocess a small number of input pictures to produce a large number of non-repeating pictures for training to reduce over-fitting. 2) We use a novel method to directly obtain the guessed diffuse map with global characteristics, which provides more prior information for the training process. 3) We improve the network architecture of the generator so that it can generate fine details of normal maps and reduce the possibility to generate over-flat normal maps. The method used in this paper can obtain prior knowledge without using dataset training, which greatly reduces the difficulty of material reconstruction and saves a lot of time to generate and calibrate datasets. | ['Hongnan Chen', 'Zhiyao Luo'] | 2022-05-24 | null | null | null | null | ['svbrdf-estimation'] | ['computer-vision'] | [ 5.78115940e-01 -2.99426932e-02 4.81465042e-01 -7.88305625e-02
-4.81150448e-01 -4.66345519e-01 5.39829433e-01 -5.91964483e-01
2.26395026e-01 8.09851706e-01 -1.34954631e-01 -2.79126853e-01
1.35653645e-01 -1.43265045e+00 -9.21310127e-01 -8.21327984e-01
4.62283820e-01 3.03787500e-01 3.43994379e-01 -2.37450778e-01
1.15611650e-01 8.11218739e-01 -1.68746197e+00 4.50742871e-01
1.01845598e+00 1.07724380e+00 4.40305859e-01 8.23571146e-01
-3.54160815e-01 5.51090956e-01 -6.94902658e-01 -5.99404871e-01
5.14076233e-01 -5.73300362e-01 -3.52941751e-01 -5.58650447e-03
2.65257806e-01 -7.40740061e-01 -7.54225105e-02 9.69520390e-01
3.85453254e-01 9.92482826e-02 1.12377965e+00 -9.21962023e-01
-6.13596320e-01 2.64014788e-02 -6.91270590e-01 -6.98634148e-01
3.15900117e-01 8.56506158e-05 3.36806148e-01 -5.27564406e-01
5.21884441e-01 1.38327515e+00 7.36076474e-01 6.88082755e-01
-1.09423828e+00 -7.43190587e-01 -2.01637700e-01 -8.65437833e-05
-1.30663514e+00 -1.35421595e-02 1.21142817e+00 -3.14182073e-01
1.60342038e-01 8.24099243e-01 8.67268085e-01 1.11096931e+00
2.37160787e-01 4.52147156e-01 1.28240252e+00 -5.71144283e-01
1.28177926e-01 2.02237979e-01 -4.43774313e-01 5.87948322e-01
9.89190415e-02 1.02739625e-01 6.09889291e-02 -2.19220296e-01
1.42718232e+00 -6.02014251e-02 -4.92833167e-01 -1.62960872e-01
-8.30574155e-01 6.03168905e-01 4.02479380e-01 -2.20135543e-02
-3.09936017e-01 5.79267442e-02 -8.79667029e-02 1.06133662e-01
6.98449016e-01 2.71690220e-01 -8.14867243e-02 2.38062870e-02
-6.81707561e-01 2.46880084e-01 7.30059922e-01 8.18495631e-01
8.22725892e-01 3.17811817e-01 1.34894639e-01 1.16312480e+00
4.54538733e-01 9.75652695e-01 1.04473196e-01 -1.02907658e+00
4.45106477e-01 2.79112786e-01 4.15416062e-01 -1.23919451e+00
6.78529218e-02 -1.88418347e-02 -1.07307518e+00 7.35929668e-01
2.96131283e-01 -1.30608594e-02 -9.27105427e-01 1.32896101e+00
4.49879438e-01 3.50607872e-01 -1.37170583e-01 8.18672359e-01
7.98448086e-01 1.04991460e+00 -4.16710466e-01 2.87144594e-02
1.01911044e+00 -8.51971328e-01 -5.23561060e-01 2.01139197e-01
-9.07652378e-02 -1.00372183e+00 1.45061338e+00 7.07253337e-01
-1.11086321e+00 -7.29139447e-01 -1.06870461e+00 -1.12809278e-01
-1.63194299e-01 6.34735823e-02 6.39297962e-01 9.12701309e-01
-9.12541866e-01 6.32171690e-01 -5.97990394e-01 2.77295589e-01
1.93251371e-01 1.36570528e-01 -4.46836725e-02 -2.52801239e-01
-1.07619226e+00 4.46888357e-01 -1.09492932e-02 3.10778856e-01
-6.63262010e-01 -7.70656407e-01 -5.63358486e-01 -1.25008181e-01
-2.72571873e-02 -9.93027508e-01 8.77099454e-01 -1.22664607e+00
-2.13838983e+00 5.51486552e-01 6.45730719e-02 2.38444790e-01
8.21639717e-01 7.33476970e-03 -4.13987905e-01 -2.48730183e-02
-4.00592178e-01 4.65508670e-01 1.09381807e+00 -1.92756414e+00
-5.44381253e-02 1.88165620e-01 4.59554084e-02 1.09668016e-01
9.90081429e-02 -2.86658853e-01 -4.02188033e-01 -8.05066824e-01
1.92724779e-01 -5.88312328e-01 -3.24336514e-02 1.84225485e-01
-5.28052151e-01 3.66113156e-01 7.06497133e-01 -1.00495732e+00
6.51249170e-01 -1.89230704e+00 -1.62230760e-01 3.76435488e-01
-7.43696559e-03 1.89077228e-01 -1.59682542e-01 2.73423821e-01
-5.04650846e-02 8.93233269e-02 -3.19788575e-01 -4.26070064e-01
-1.84442624e-01 1.39420545e-02 -5.78488052e-01 1.70414582e-01
-7.50228986e-02 6.37836933e-01 -5.13732076e-01 -1.68860853e-01
3.76335472e-01 9.96038437e-01 -5.33080935e-01 4.39121634e-01
-3.61402750e-01 7.31769562e-01 -2.89212227e-01 4.81344640e-01
1.29428756e+00 1.84198335e-01 -1.73121870e-01 -5.67271650e-01
-8.02552141e-03 -6.10193089e-02 -1.23231888e+00 1.22659862e+00
-9.67824876e-01 4.38877434e-01 2.71434218e-01 -4.85599905e-01
1.34676158e+00 8.58143568e-02 2.57181674e-01 -6.74736500e-01
1.78158179e-01 1.95552006e-01 -3.52794647e-01 -3.45989525e-01
4.12695557e-01 -2.79749483e-01 3.60791147e-01 3.98478746e-01
-7.60337174e-01 -7.83205807e-01 -4.68604594e-01 -2.85762697e-01
6.46059036e-01 4.39713389e-01 -2.69017935e-01 1.01684041e-01
5.63592851e-01 -2.34727591e-01 3.76988441e-01 3.38047743e-01
7.79633880e-01 1.11378658e+00 2.18042776e-01 -5.58269143e-01
-1.31208134e+00 -1.23962164e+00 -1.02876067e-01 3.76918316e-01
3.67079675e-01 1.15658984e-01 -1.01385701e+00 -1.62375465e-01
-1.64017633e-01 8.60989690e-01 -5.41241407e-01 -6.45129234e-02
-7.95664608e-01 -6.48077905e-01 2.68504173e-01 2.55751908e-01
7.68856227e-01 -1.13937426e+00 -1.34910718e-01 -1.26170486e-01
-7.52210245e-02 -7.49986291e-01 -2.00142547e-01 -6.30820990e-01
-8.89990389e-01 -8.20097327e-01 -1.02763855e+00 -5.38563311e-01
9.91549551e-01 2.54094422e-01 1.17356575e+00 7.31029958e-02
-2.45919108e-01 1.70778513e-01 -1.02426492e-01 -4.60979462e-01
-8.01286101e-01 -4.21443731e-01 -4.13555920e-01 3.81153256e-01
-3.48416328e-01 -6.51237547e-01 -7.51175046e-01 5.86553276e-01
-1.01743686e+00 7.95560598e-01 3.61639142e-01 5.12709916e-01
9.02083278e-01 2.70141333e-01 2.00883299e-01 -1.10981584e+00
5.62324822e-01 -2.22657174e-01 -8.08249831e-01 1.51317358e-01
-2.25989655e-01 -2.22924083e-01 9.76684391e-01 -6.70265198e-01
-1.54714978e+00 -3.53572965e-01 -3.40052068e-01 -6.12958670e-01
-1.15511663e-01 -5.25438227e-02 -5.57610393e-01 -2.54326761e-01
5.57125151e-01 1.87548175e-01 -1.23333499e-01 -6.62122130e-01
2.02907786e-01 5.87764204e-01 4.03726667e-01 -7.31945038e-01
9.91420567e-01 6.40453339e-01 2.96386451e-01 -8.91996562e-01
-3.63755733e-01 1.84686482e-01 -2.31130823e-01 -3.92277092e-01
6.21589005e-01 -6.31379485e-01 -7.76256204e-01 7.93925881e-01
-1.31314540e+00 -7.38536835e-01 -3.23998004e-01 2.44670376e-01
-6.37583315e-01 3.75012159e-01 -6.84525788e-01 -8.90631735e-01
-3.66494000e-01 -1.16772330e+00 1.09222817e+00 1.86298773e-01
1.74265146e-01 -1.00067687e+00 1.53017966e-02 4.65958595e-01
6.11641884e-01 5.93258440e-01 9.53863680e-01 6.70041025e-01
-1.00981045e+00 -8.54675099e-02 -4.14657533e-01 5.94201326e-01
4.94911224e-01 3.58230770e-01 -1.10915136e+00 2.06680261e-02
2.62485713e-01 1.73182011e-01 5.31451404e-01 4.20591056e-01
1.67833233e+00 -3.09583545e-01 -1.10088222e-01 9.25311327e-01
1.53571022e+00 3.92818749e-01 1.20633388e+00 2.06720933e-01
1.17003524e+00 6.08270347e-01 5.27051032e-01 2.41773114e-01
2.35803291e-01 7.60475338e-01 4.63634759e-01 -3.70920271e-01
-5.35452127e-01 -3.34750056e-01 2.28230968e-01 8.75927687e-01
-5.82551718e-01 -4.70960528e-01 -5.22886992e-01 -3.70658911e-03
-1.22771311e+00 -7.69549429e-01 -3.16942811e-01 2.30695009e+00
8.80527854e-01 -1.43577293e-01 -1.49276659e-01 1.01858385e-01
7.81427562e-01 -1.75786570e-01 -3.88643473e-01 -5.61609805e-01
-1.05374902e-01 3.54179412e-01 4.36153173e-01 8.33835959e-01
-4.47205752e-01 6.35644019e-01 5.55731344e+00 1.05807590e+00
-1.40473294e+00 -6.63875192e-02 8.26468945e-01 2.65934587e-01
-9.89625871e-01 -3.28830004e-01 -5.69699764e-01 7.49306858e-01
6.58416390e-01 3.07478487e-01 8.49100471e-01 6.17531240e-01
3.51377606e-01 -1.12126179e-01 -6.76227212e-01 1.22203708e+00
1.31567754e-02 -1.13934052e+00 3.07750463e-01 2.47682296e-02
9.89307046e-01 -5.44034541e-01 2.07340464e-01 -1.92442223e-01
1.85817093e-01 -1.13337052e+00 7.40000010e-01 1.07160997e+00
1.22711360e+00 -7.55930603e-01 4.96606648e-01 4.40669775e-01
-9.11327600e-01 4.32188153e-01 -6.13089502e-01 3.51895660e-01
2.54973233e-01 9.89359021e-01 -7.24115372e-01 5.89661658e-01
4.30355519e-01 2.17193022e-01 -1.83860719e-01 1.04195476e+00
-1.59588665e-01 5.91722488e-01 -3.99570256e-01 9.31823775e-02
-3.49795997e-01 -8.44315171e-01 3.07506979e-01 7.85086155e-01
6.63545907e-01 -4.09466475e-02 -1.76050872e-01 1.21316695e+00
1.15204733e-02 2.07206935e-01 -4.79085147e-01 3.30531001e-01
1.43849418e-01 1.12367880e+00 -6.99150383e-01 -1.03247233e-01
-9.78724658e-02 1.07088745e+00 -9.74819511e-02 4.46505636e-01
-9.40373242e-01 -4.25398111e-01 2.17663169e-01 5.82145751e-01
-1.94261856e-02 -1.20578088e-01 -5.22199988e-01 -9.65228975e-01
1.21972397e-01 -7.46206164e-01 -3.83711040e-01 -1.36692309e+00
-1.36323071e+00 6.89478576e-01 -8.14800784e-02 -1.40942442e+00
-4.94417213e-02 -7.47675180e-01 -6.84292614e-01 1.44896388e+00
-1.52643502e+00 -1.28845954e+00 -8.34995985e-01 4.76598591e-01
3.62926483e-01 4.19085026e-02 8.74699235e-01 3.05512130e-01
-1.10348612e-01 3.54168534e-01 2.19539419e-01 -2.09653825e-01
5.65784812e-01 -9.41110492e-01 5.47407627e-01 4.32614803e-01
-2.29353666e-01 4.31155056e-01 6.69796348e-01 -7.33908117e-01
-1.12298071e+00 -1.01026297e+00 8.03842675e-03 -1.63144782e-01
1.15233650e-02 -4.78295028e-01 -9.86745775e-01 2.59077728e-01
-6.94338828e-02 -1.79169998e-01 3.66875231e-01 -4.85382169e-01
-1.60512835e-01 -3.13242555e-01 -1.54642177e+00 7.88894534e-01
8.11635196e-01 -3.26586157e-01 2.27948222e-02 2.12377310e-01
6.33374453e-01 -6.90159082e-01 -7.06452727e-01 3.35908562e-01
6.99958444e-01 -1.27819514e+00 1.13656151e+00 1.26532912e-01
6.59459889e-01 -4.58786696e-01 -6.22831844e-02 -1.43872142e+00
-1.64993465e-01 -6.65797949e-01 -4.43984307e-02 1.29737854e+00
1.82492122e-01 -8.29535842e-01 8.95711482e-01 6.28285408e-01
-1.64019391e-01 -7.03173459e-01 -4.00130063e-01 -5.94535470e-01
-3.40367779e-02 -3.51487905e-01 1.09835672e+00 8.21777880e-01
-9.50093508e-01 -2.09482700e-01 -4.81194496e-01 9.39535871e-02
6.38092160e-01 3.87115926e-01 9.65380549e-01 -1.24011374e+00
-6.06507659e-01 -2.01942146e-01 -1.49291158e-02 -1.20153785e+00
-3.63574363e-03 -4.73505676e-01 7.00931847e-02 -1.49858236e+00
-1.59076735e-01 -1.15635288e+00 2.29419068e-01 -9.05193165e-02
-8.91078170e-03 4.10451502e-01 -1.17994972e-01 1.24823660e-01
4.05108303e-01 5.85060239e-01 1.97984695e+00 -5.54909185e-02
-2.52526551e-01 4.34640199e-01 -5.50022900e-01 7.03066707e-01
9.33507621e-01 -1.61365226e-01 -7.37984717e-01 -5.71020186e-01
3.91447306e-01 2.18265820e-02 4.33878273e-01 -9.04370427e-01
-4.73140836e-01 -3.12661588e-01 6.15044594e-01 -5.29697299e-01
7.58968234e-01 -1.04587042e+00 8.16224277e-01 -7.88799487e-03
1.12651303e-01 -2.14778244e-01 1.85494199e-01 2.92364568e-01
2.69287955e-02 -4.51250166e-01 9.35387373e-01 -1.43241599e-01
-4.18269560e-02 3.01772773e-01 6.46451116e-02 -3.42894852e-01
8.08118284e-01 -4.61289614e-01 -2.91842580e-01 -6.44955456e-01
-1.90467179e-01 -6.62773371e-01 9.07035708e-01 3.63421962e-02
6.19315326e-01 -1.49438095e+00 -4.30280834e-01 4.98941064e-01
-6.10249877e-01 4.27946806e-01 4.97530848e-01 1.49756953e-01
-1.33421123e+00 -5.66164292e-02 -2.27950767e-01 -3.47981155e-01
-1.04047632e+00 4.22165602e-01 3.63547564e-01 -5.15920408e-02
-7.62162924e-01 6.51050150e-01 6.99470818e-01 -4.66943204e-01
-2.72431493e-01 -3.61682326e-01 -3.67561281e-02 -4.33948398e-01
5.37801206e-01 5.63496292e-01 5.31583726e-02 -4.49021548e-01
3.01045597e-01 9.76682425e-01 2.81653196e-01 -1.16117738e-01
1.35491061e+00 8.10240582e-02 -2.76176721e-01 3.79527390e-01
1.08223081e+00 6.17190123e-01 -1.43124604e+00 4.09532666e-01
-9.94315267e-01 -7.94375062e-01 1.56697646e-01 -8.28711033e-01
-1.38773906e+00 1.00179088e+00 3.92764837e-01 3.77193600e-01
1.23156667e+00 -3.85831863e-01 1.03744817e+00 -2.12082371e-01
4.73410994e-01 -7.28962123e-01 -7.53947794e-02 6.39853254e-02
1.19683731e+00 -5.62096953e-01 -6.42914698e-02 -1.07843244e+00
-3.68732452e-01 1.23307765e+00 4.59015846e-01 -2.28134632e-01
6.22138500e-01 4.35984492e-01 1.44290432e-01 6.45657182e-02
-8.89589787e-02 6.23313308e-01 4.35440630e-01 7.50984728e-01
8.77880678e-02 1.39605030e-01 1.98432192e-01 4.95892674e-01
-4.37110126e-01 -1.46928757e-01 5.39048374e-01 4.00182158e-01
-1.44129619e-01 -1.30617261e+00 -7.69056380e-01 3.32765967e-01
-2.41449267e-01 -8.42247680e-02 -1.01635024e-01 3.84542286e-01
2.55331576e-01 5.95286548e-01 9.97146219e-02 -2.34877348e-01
3.45827371e-01 -3.27539712e-01 6.99217200e-01 -3.28130394e-01
-1.52688563e-01 1.23978481e-01 -4.55033332e-02 -2.41199777e-01
-2.80387223e-01 -3.52015972e-01 -9.69941080e-01 -5.29514849e-01
-3.12733531e-01 -1.59551889e-01 9.34474707e-01 4.19697046e-01
1.80392578e-01 5.94587207e-01 8.70664656e-01 -1.23325384e+00
-2.15832479e-02 -7.02390134e-01 -7.30224490e-01 4.74110961e-01
8.36639851e-02 -7.71552205e-01 -4.59091693e-01 1.85662672e-01] | [9.500858306884766, -3.17021107673645] |
ad7ed517-629d-45e6-b239-cec40e7efea9 | evolutionary-framework-for-two-stage | 1903.01885 | null | http://arxiv.org/abs/1903.01885v1 | http://arxiv.org/pdf/1903.01885v1.pdf | Evolutionary framework for two-stage stochastic resource allocation problems | Resource allocation problems are a family of problems in which resources must
be selected to satisfy given demands. This paper focuses on the two-stage
stochastic generalization of resource allocation problems where future demands
are expressed in a finite number of possible scenarios. The goal is to select
cost effective resources to be acquired in the present time (first stage), and
to implement a complete solution for each scenario (second stage), while
minimizing the total expected cost of the choices in both stages.
We propose an evolutionary framework for solving general two-stage stochastic
resource allocation problems. In each iteration of our framework, a local
search algorithm selects resources to be acquired in the first stage. A genetic
metaheuristic then completes the solutions for each scenario and relevant
information is passed onto the next iteration, thereby supporting the
acquisition of promising resources in the following first stage.
Experimentation on numerous instances of the two-stage stochastic Steiner tree
problem suggests that our evolutionary framework is powerful enough to address
large instances of a wide variety of two-stage stochastic resource allocation
problems. | ['Fábio L. Usberti', 'Evandro C. Bracht', 'Mário C. San Felice', 'Pedro H. D. B. Hokama'] | 2018-11-29 | null | null | null | null | ['steiner-tree-problem'] | ['graphs'] | [ 6.79621816e-01 -1.12111308e-01 -5.52117467e-01 -2.01118171e-01
-3.21092784e-01 -4.06805217e-01 9.82011296e-03 -2.16775224e-01
-2.46493578e-01 1.19672227e+00 -1.44021779e-01 -3.04416746e-01
-9.62026536e-01 -9.43457723e-01 7.59128332e-02 -7.51278222e-01
-1.07068844e-01 1.11482990e+00 3.06311280e-01 -3.75255108e-01
5.70649147e-01 9.18836951e-01 -1.52491581e+00 -1.45193845e-01
7.76469707e-01 1.06109250e+00 8.25375736e-01 1.96727172e-01
-5.11747718e-01 1.03212856e-01 -6.59705579e-01 -4.51579802e-02
5.43177307e-01 -4.52848315e-01 -1.17851138e+00 6.74079001e-01
-8.39104950e-01 7.90993348e-02 3.81242633e-01 8.89355958e-01
5.26999414e-01 3.96620721e-01 4.58794624e-01 -1.49557495e+00
1.04065560e-01 8.09675097e-01 -7.78913736e-01 3.51100892e-01
4.01953816e-01 -6.29158989e-02 8.63492787e-01 -8.17159295e-01
5.95188200e-01 9.52580690e-01 4.79825810e-02 7.79596746e-01
-1.20327127e+00 -1.84361592e-01 3.20173174e-01 9.40138325e-02
-1.40967631e+00 -4.77216005e-01 8.43188584e-01 1.56311184e-01
1.05905676e+00 5.62257171e-01 7.85790265e-01 1.98906690e-01
-2.05006897e-01 6.99335873e-01 6.29867852e-01 -6.48656487e-01
7.03926563e-01 1.07628942e-01 -4.67366546e-01 1.63069859e-01
3.98119897e-01 -4.83752713e-02 -3.47604841e-01 -3.16023439e-01
4.95750338e-01 -1.15560561e-01 -3.81265096e-02 -6.31933331e-01
-8.08255196e-01 8.29215229e-01 4.72280104e-03 4.93600219e-01
-8.45790625e-01 1.80208739e-02 2.13746607e-01 3.99082184e-01
3.03804815e-01 7.83111572e-01 -7.35436261e-01 7.71439299e-02
-1.18113196e+00 1.74089298e-01 7.97688663e-01 1.15039313e+00
5.00272632e-01 1.43864006e-01 -6.35753665e-03 7.88069904e-01
1.03840999e-01 1.80012479e-01 -8.16574469e-02 -9.09604073e-01
7.07007766e-01 2.39093199e-01 5.78224838e-01 -6.48014367e-01
-4.05054182e-01 -5.54802477e-01 -2.95493364e-01 2.09996421e-02
2.63974220e-01 -4.18041795e-01 -5.53291798e-01 1.58894384e+00
4.56919760e-01 -2.25598618e-01 -8.65281224e-02 7.60112703e-01
-4.22285870e-02 7.84589946e-01 -1.93247706e-01 -1.13627696e+00
8.08478236e-01 -1.05620110e+00 -3.89188886e-01 -1.15887642e-01
2.40331143e-01 -7.59507120e-01 4.14062053e-01 2.41673484e-01
-1.66708195e+00 3.50283720e-02 -7.38381624e-01 8.93610716e-01
-8.32545906e-02 -2.05633387e-01 6.55148089e-01 7.63702631e-01
-1.25892115e+00 4.08964157e-01 -2.26770848e-01 -4.05654907e-01
1.75449789e-01 6.61522985e-01 3.60826671e-01 -1.94193453e-01
-1.03375244e+00 8.13879490e-01 4.03616846e-01 3.38263273e-01
-6.76395655e-01 -5.63320696e-01 -3.71288449e-01 4.64907616e-01
1.05507958e+00 -8.24885488e-01 1.04215837e+00 -9.36671853e-01
-1.24901175e+00 6.56172514e-01 -2.31529385e-01 -6.40249625e-02
5.69176376e-01 6.62493587e-01 -3.13893795e-01 1.22239694e-01
3.35687310e-01 3.61471802e-01 6.88765049e-01 -1.11538601e+00
-1.13467824e+00 -1.19385213e-01 2.59202719e-01 6.09754920e-01
-2.25007117e-01 4.38037723e-01 -5.49153328e-01 -3.78541052e-01
1.53043434e-01 -9.28642571e-01 -1.00897729e+00 -4.08760369e-01
-3.10722798e-01 -2.21893594e-01 3.38337570e-01 8.70674327e-02
1.36949849e+00 -1.65068698e+00 7.30577886e-01 6.32082224e-01
-3.87927175e-01 -9.85258445e-02 -2.68888354e-01 6.03077292e-01
-1.27325892e-01 3.74151707e-01 -1.75775439e-01 -1.97281569e-01
-1.43218234e-01 2.16557339e-01 -5.69125488e-02 5.58610186e-02
1.13822393e-01 3.69560063e-01 -7.43611038e-01 -3.12227696e-01
-4.51365449e-02 -5.02333164e-01 -4.99644846e-01 1.36700626e-02
-3.59759450e-01 1.30583763e-01 -9.42656100e-01 8.73731434e-01
5.20783544e-01 -5.62145114e-02 3.90760273e-01 4.05905187e-01
-2.57190943e-01 -4.56850678e-02 -1.35141945e+00 1.43110669e+00
-8.75913143e-01 6.02725893e-03 2.91605592e-01 -1.15951955e+00
7.20452607e-01 1.49320319e-01 1.13144851e+00 -4.53319281e-01
1.23907380e-01 5.41177750e-01 -1.01686545e-01 -3.38814408e-01
6.07702374e-01 -2.58577198e-01 -3.58871788e-01 7.61290252e-01
-5.25206268e-01 -4.72043544e-01 5.51655054e-01 4.11855280e-02
8.35088313e-01 -2.21477598e-02 3.54855061e-01 -4.21110481e-01
7.57370591e-01 2.68831819e-01 9.05334771e-01 6.60023689e-01
-2.23895505e-01 1.51269853e-01 3.23057324e-01 -5.60397625e-01
-1.00851631e+00 -6.47919655e-01 2.13059485e-01 8.57100964e-01
3.95845741e-01 3.77493382e-01 -2.78675944e-01 -6.65784359e-01
-2.37619970e-02 9.02810812e-01 -2.86234558e-01 2.80206025e-01
-5.64365029e-01 -7.12530255e-01 -5.10870457e-01 2.61371788e-02
2.27419764e-01 -1.13122559e+00 -1.11413515e+00 5.50863922e-01
-7.23368675e-02 -7.39799321e-01 -4.66753721e-01 4.51502770e-01
-8.62326205e-01 -9.99862075e-01 -8.17953050e-01 -6.53174579e-01
9.81703281e-01 5.62253177e-01 9.97963846e-01 3.35482806e-01
-5.79432487e-01 2.88161635e-01 -3.70262742e-01 -1.93649441e-01
9.36242491e-02 2.21539065e-01 1.83759019e-01 -2.27265745e-01
2.52109952e-03 -4.28757906e-01 -4.16324437e-01 6.05715513e-01
-7.15430140e-01 -1.77591056e-01 3.54060680e-01 8.02308559e-01
7.35357106e-01 1.03338337e+00 9.07919168e-01 -7.17971206e-01
5.52845299e-01 -8.71020913e-01 -8.23322833e-01 6.54927731e-01
-5.41592479e-01 -1.24210812e-01 7.11188495e-01 -3.02849650e-01
-1.27610004e+00 -2.82541141e-02 2.47028083e-01 -7.87654370e-02
2.60209620e-01 5.92461467e-01 -5.12580693e-01 -5.46870986e-03
-1.27680264e-02 8.65496993e-02 -4.58817154e-01 -2.46985018e-01
-1.58187456e-03 3.41604233e-01 -1.79414637e-02 -8.49420130e-01
6.74032331e-01 -1.97438914e-02 4.25829411e-01 -2.64764935e-01
-5.81235886e-01 -3.40302080e-01 -4.91915762e-01 -3.26210648e-01
1.51373997e-01 -1.72498330e-01 -5.50352216e-01 2.20028356e-01
-7.25899041e-01 -2.01086402e-01 -8.17109525e-01 -6.12546243e-02
-9.40260768e-01 -4.28578211e-03 3.01502943e-01 -1.25074160e+00
-1.06508538e-01 -1.26121700e+00 5.81051350e-01 5.37510872e-01
-1.00615859e-01 -9.38878834e-01 -3.67791116e-01 1.50162861e-01
5.73631942e-01 2.49569476e-01 1.05663931e+00 -4.96361196e-01
-6.74162447e-01 9.99671519e-02 3.42865825e-01 -4.33070451e-01
2.59367377e-01 4.51096445e-02 1.72942892e-01 -4.92178857e-01
2.43744418e-01 -1.10883772e-01 3.00568998e-01 6.12289906e-01
1.15779734e+00 -2.17015386e-01 -6.78929806e-01 3.15988421e-01
1.83623922e+00 9.12471414e-01 3.83402497e-01 5.70036471e-01
-2.10582241e-01 1.02593338e+00 1.28978968e+00 7.95721292e-01
8.42392743e-02 8.74348342e-01 5.93065202e-01 2.09519297e-01
5.08065343e-01 3.08518678e-01 -1.91670209e-01 1.71965018e-01
3.61263193e-02 -7.71038890e-01 -9.09547985e-01 9.80050087e-01
-1.84931052e+00 -1.07537031e+00 2.89120406e-01 2.14250827e+00
4.67286587e-01 2.22591057e-01 4.07172889e-01 3.56808096e-01
1.11103010e+00 8.08001235e-02 -8.25520694e-01 -9.65886354e-01
9.90241542e-02 1.11809991e-01 4.72330093e-01 1.97933197e-01
-5.29075563e-01 7.64112592e-01 6.93730974e+00 9.19177413e-01
-7.33368099e-01 -3.23845088e-01 1.01642740e+00 -6.60977840e-01
-8.22779953e-01 1.38859704e-01 -6.89253271e-01 6.24787211e-01
8.51382732e-01 -8.75310242e-01 8.35780621e-01 6.83539808e-01
3.99065107e-01 -5.24934888e-01 -9.08356905e-01 5.28560340e-01
-2.83459455e-01 -1.38056946e+00 -1.75010964e-01 1.17633037e-01
1.16327596e+00 -5.44045091e-01 2.62260795e-01 -1.42818853e-01
4.12769943e-01 -5.97208679e-01 6.88804984e-01 2.90132105e-01
9.68876421e-01 -1.47627473e+00 4.16077077e-01 6.71363175e-01
-1.17071867e+00 -6.73325181e-01 -3.49089146e-01 1.73299044e-01
8.82821679e-01 5.67141950e-01 -6.32509589e-01 8.88836384e-01
5.99435866e-01 -1.08169347e-01 4.31541324e-01 1.39431727e+00
-2.14829873e-02 -1.93072245e-01 -4.60454822e-01 -2.98606247e-01
4.52977657e-01 -3.19479972e-01 8.23408604e-01 5.02745926e-01
6.52511954e-01 5.91312408e-01 3.10504675e-01 5.46624124e-01
2.19096303e-01 2.03564778e-01 -4.42579001e-01 9.05686915e-02
8.66606832e-01 9.75229919e-01 -1.24812925e+00 -2.85553653e-02
-9.90022793e-02 7.18876183e-01 -4.57522161e-02 1.90059319e-01
-6.18638754e-01 -5.17421663e-01 3.67302626e-01 -7.40019679e-02
3.94150019e-01 -6.13776743e-02 -5.03730536e-01 -7.14279950e-01
-2.66359597e-01 -4.89827096e-01 5.75797677e-01 -5.40898740e-01
-8.61082613e-01 5.60728550e-01 2.83048064e-01 -9.50796664e-01
-5.21586776e-01 -6.27041310e-02 -8.26232612e-01 1.09477687e+00
-1.60359597e+00 -4.80863184e-01 2.77139902e-01 5.98642230e-01
1.26239133e+00 -3.94193232e-01 5.73309541e-01 -9.03595053e-03
-7.07162678e-01 9.12278295e-02 1.24327488e-01 -1.06889105e+00
-1.93726823e-01 -9.46066439e-01 -7.89656937e-02 1.02084124e+00
-5.32495618e-01 4.18661684e-01 8.76314998e-01 -7.45245159e-01
-1.14113963e+00 -7.42319286e-01 9.95900333e-01 4.14234310e-01
5.49149156e-01 2.68217564e-01 -1.87619086e-02 6.61782250e-02
-1.67236090e-01 -4.58179563e-01 7.94338226e-01 -2.32203841e-01
6.42012954e-01 -4.51201275e-02 -1.82928073e+00 6.70990288e-01
1.26052988e+00 2.77081817e-01 -5.73682897e-02 4.22382087e-01
5.15790224e-01 -2.86431126e-02 -5.62328100e-01 4.98120368e-01
2.98601717e-01 -5.36869168e-01 7.61670232e-01 -5.99190950e-01
2.81651020e-01 -6.41808435e-02 -1.92391396e-01 -1.46965146e+00
-4.60385352e-01 -1.08664930e+00 2.12296531e-01 1.05775237e+00
7.43276238e-01 -5.32047689e-01 1.00727332e+00 1.12525320e+00
1.23723187e-02 -1.11112618e+00 -1.08361006e+00 -8.21422398e-01
-2.53363818e-01 -1.35004325e-02 1.15146792e+00 6.36296213e-01
-1.75555423e-01 -1.94294438e-01 -5.53648293e-01 -1.79156840e-01
7.20941603e-01 7.80849278e-01 2.17550084e-01 -9.39837754e-01
-4.09274369e-01 -4.58857208e-01 2.51590014e-01 -9.27081823e-01
9.93220210e-02 -5.34633458e-01 1.10908531e-01 -1.69643462e+00
2.08447248e-01 -1.19745219e+00 -2.81361014e-01 3.56885940e-01
4.44666110e-02 -1.61089286e-01 3.78812313e-01 3.06229919e-01
-5.30160010e-01 3.51166666e-01 1.11406589e+00 1.10913582e-01
-5.07956982e-01 7.98201084e-01 -8.44357371e-01 5.39302349e-01
9.03523386e-01 -5.35542607e-01 -7.95612335e-01 -3.96184444e-01
2.49127701e-01 9.53466833e-01 -7.91426778e-01 -3.75650138e-01
3.63148272e-01 -1.18430793e+00 1.02917150e-01 -8.49860668e-01
2.80449033e-01 -1.47201610e+00 6.56897366e-01 6.03331566e-01
-1.79947436e-01 1.97611883e-01 -4.01359759e-02 5.09313107e-01
1.95143491e-01 -8.72125447e-01 5.32913566e-01 -1.40868366e-01
-8.81482065e-01 4.72221613e-01 -4.92537022e-01 -3.28314990e-01
1.80805397e+00 -6.35642707e-01 1.35206416e-01 -2.22599834e-01
-1.12853706e+00 9.17174995e-01 5.16846061e-01 1.65194824e-01
5.42746544e-01 -1.03117001e+00 -4.54167157e-01 -3.60107452e-01
-3.94683450e-01 -8.67683589e-02 2.48095125e-01 5.12869060e-01
-3.81617010e-01 5.73595226e-01 -5.23480654e-01 -1.24456413e-01
-1.12608504e+00 8.99766684e-01 1.57783344e-01 -6.55016422e-01
1.35254532e-01 1.10797751e+00 -4.37339932e-01 2.12985396e-01
-3.66362371e-02 6.07938290e-01 -3.76074076e-01 3.64100873e-01
2.21094131e-01 6.53997064e-01 -1.44394413e-01 -5.08175373e-01
-7.12767482e-01 5.24342120e-01 1.04460619e-01 -4.36470032e-01
1.73556554e+00 -6.58473432e-01 -3.38049352e-01 -1.84253842e-01
6.62476182e-01 -3.11454609e-02 -8.58859181e-01 -3.05144638e-01
2.91209728e-01 -1.00364757e+00 -2.44464129e-01 -8.33440721e-01
-1.57493353e+00 7.71018490e-02 -7.79281333e-02 5.07696092e-01
2.09269500e+00 -2.69920677e-01 5.51453948e-01 7.10532814e-02
1.27105868e+00 -1.18834651e+00 -1.34637997e-01 1.93605766e-01
7.78377295e-01 -4.35253769e-01 1.70741603e-01 -7.24111974e-01
-7.55725920e-01 1.20653963e+00 4.28629607e-01 -4.26204270e-03
3.89313281e-01 2.84745693e-01 -7.34944165e-01 -4.28751223e-02
-1.08536625e+00 -3.55035186e-01 -2.35756442e-01 4.36311692e-01
-2.96905607e-01 2.72888760e-03 -9.59832132e-01 1.55060560e-01
1.05030268e-01 -2.43554831e-01 6.96678758e-01 1.19612920e+00
-6.52796209e-01 -1.46192157e+00 -3.57950896e-01 4.79641646e-01
-4.70591187e-01 1.19310781e-01 -2.59393215e-01 4.32010651e-01
1.94912255e-01 1.33773029e+00 7.71887973e-02 4.97020744e-02
3.39608550e-01 -2.80917794e-01 5.01255870e-01 -9.38030720e-01
-5.92231631e-01 3.25980246e-01 5.74079812e-01 -5.27299464e-01
-5.31875789e-01 -1.02986944e+00 -9.69125032e-01 -2.06647187e-01
-6.62350953e-01 7.15695977e-01 8.10011148e-01 6.49581909e-01
9.82716158e-02 6.30581439e-01 1.42676830e+00 -7.53598809e-01
-5.41461051e-01 -2.15538163e-02 -8.00261259e-01 -4.52499300e-01
-3.16324115e-01 -8.34525883e-01 -8.46282765e-02 -4.69055295e-01] | [5.398467540740967, 3.2222445011138916] |
dda04f99-ad93-4b60-b10b-3a340ce9f3a8 | unleashing-the-power-of-user-reviews | 2306.15541 | null | https://arxiv.org/abs/2306.15541v1 | https://arxiv.org/pdf/2306.15541v1.pdf | Unleashing the Power of User Reviews: Exploring Airline Choices at Catania Airport, Italy | This study aims to investigate the possible relationship between the mechanisms of social influence and the choice of airline, through the use of new tools, with the aim of understanding whether they can contribute to a better understanding of the factors influencing the decisions of consumers in the aviation sector. We have chosen to extract user reviews from well-known platforms: Trustpilot, Google, and Twitter. By combining web scraping techniques, we have been able to collect a comprehensive dataset comprising a wide range of user opinions, feedback, and ratings. We then refined the BERT model to focus on insightful sentiment in the context of airline reviews. Through our analysis, we observed an intriguing trend of average negative sentiment scores across various airlines, giving us deeper insight into the dynamics between airlines and helping us identify key partnerships, popular routes, and airlines that play a central role in the aeronautical ecosystem of Catania airport during the specified period. Our investigation led us to find that, despite an airline having received prestigious awards as a low-cost leader in Europe for two consecutive years 2021 and 2022, the "Catanese" user tends to suffer the dominant position of other companies. Understanding the impact of positive reviews and leveraging sentiment analysis can help airlines improve their reputation, attract more customers, and ultimately gain a competitive edge in the marketplace. | ['Antonio Picone', 'Vincenzo Miracula'] | 2023-06-27 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [-4.48999465e-01 -8.39499198e-03 -3.08399856e-01 -7.06013963e-02
-2.86253184e-01 -7.30186641e-01 5.75674534e-01 7.66330004e-01
-5.37525773e-01 2.32502267e-01 4.43645507e-01 -6.03742838e-01
-2.54968762e-01 -9.26349759e-01 -3.54467630e-01 -3.88244092e-01
-2.31995359e-02 -1.83970705e-01 -1.56266406e-01 -7.48195529e-01
5.50260007e-01 5.30116558e-01 -1.13036752e+00 -6.08155429e-02
4.17984456e-01 9.55403984e-01 -2.01938033e-01 1.49776503e-01
3.69485110e-01 5.43396533e-01 -4.82417256e-01 -7.99674273e-01
4.81063902e-01 -8.38891417e-02 -1.78976312e-01 -9.19726714e-02
-1.28794596e-01 6.60997257e-02 1.91729143e-01 7.61498451e-01
2.25700736e-01 3.09245866e-02 1.03575237e-01 -9.70636129e-01
-1.51797488e-01 4.07571852e-01 -2.79208362e-01 2.17467576e-01
2.86408693e-01 1.66170001e-01 1.60243189e+00 -6.96720123e-01
6.35081351e-01 5.91507494e-01 4.62640107e-01 -2.84234732e-01
-7.88634300e-01 -6.35681570e-01 4.41557944e-01 -2.58282367e-02
-9.87268686e-01 -1.29385799e-01 5.35453796e-01 -6.55317545e-01
4.57354814e-01 3.80131751e-01 1.12152541e+00 6.59108639e-01
4.81718481e-01 3.21498722e-01 1.10299838e+00 -1.69224888e-01
1.86727941e-01 7.56706059e-01 5.09604365e-02 2.08627522e-01
4.81591851e-01 -8.44567567e-02 -5.88119686e-01 -2.52338946e-01
1.94628716e-01 2.40431830e-01 -3.38267922e-01 -1.70476079e-01
-1.10596347e+00 9.99199808e-01 5.81635714e-01 5.71905792e-01
-6.62282765e-01 -3.73329312e-01 3.10522139e-01 4.53841984e-01
5.39922297e-01 8.62669349e-01 -4.99319583e-01 -4.87583160e-01
-4.51723874e-01 6.90165833e-02 9.58268702e-01 3.22551169e-02
5.44345140e-01 -4.45305526e-01 4.79412436e-01 4.16166455e-01
4.08011645e-01 4.19105798e-01 1.71878666e-01 -5.01540780e-01
4.28697318e-01 8.25873435e-01 3.58039141e-01 -1.49343336e+00
-2.12234288e-01 -1.07301474e+00 -1.11335933e-01 1.70493573e-01
2.27803603e-01 -4.68005031e-01 8.06727856e-02 9.65476155e-01
4.09502722e-02 -6.71019793e-01 -3.64984989e-01 9.82760966e-01
8.09657499e-02 3.41898650e-01 -1.61440909e-01 -5.48674092e-02
1.00701952e+00 -6.43893659e-01 -3.16868752e-01 -4.38869223e-02
6.44021392e-01 -7.94724166e-01 9.21360612e-01 7.26217747e-01
-7.48307765e-01 -3.79027754e-01 -1.20648551e+00 6.68481171e-01
-6.49721384e-01 -1.89045101e-01 6.86172843e-01 7.68472195e-01
-6.07992947e-01 6.71052992e-01 -4.16046202e-01 -2.96322882e-01
1.48745760e-01 2.28471979e-01 -3.23894083e-01 1.41700711e-02
-1.13845706e+00 9.38996255e-01 -5.77022791e-01 1.66577607e-01
-2.93804795e-01 -4.66461569e-01 -3.95591259e-01 -6.38919547e-02
4.23298270e-01 -3.16888958e-01 8.30911219e-01 -1.15168238e+00
-9.38732147e-01 2.82091171e-01 2.49582842e-01 -2.98131168e-01
2.50018686e-01 -7.56397471e-02 -7.20943213e-01 -2.54040778e-01
7.66880587e-02 -1.41755059e-01 2.67989904e-01 -8.96122873e-01
-8.44306469e-01 -5.51444769e-01 4.67850387e-01 5.74587211e-02
-5.13642371e-01 4.35030274e-02 6.34127185e-02 -2.80363709e-01
-1.32336617e-01 -1.02274454e+00 -2.95110613e-01 -6.55006707e-01
5.10153286e-02 6.28607348e-02 1.59216195e-01 -3.01455200e-01
1.32682848e+00 -2.03542733e+00 -9.77150127e-02 6.06049359e-01
1.67525858e-01 1.66185558e-01 2.66286701e-01 9.63107765e-01
1.09484963e-01 5.53127825e-01 4.65448141e-01 5.98955676e-02
-7.17168078e-02 -3.26152623e-01 -1.91333130e-01 4.65957642e-01
7.79986084e-02 5.53197980e-01 -9.68788326e-01 3.59582901e-01
-7.04774959e-03 2.86978483e-01 -3.85155350e-01 -2.21541718e-01
7.19198957e-02 3.22308153e-01 -5.66202343e-01 7.75854528e-01
2.62542903e-01 -2.68704951e-01 2.28452653e-01 9.82633419e-03
-7.47271180e-01 5.79099357e-01 -6.20940864e-01 7.71381259e-01
-6.34447038e-01 9.58047152e-01 2.16235250e-01 -2.61917651e-01
9.75514114e-01 -4.43060026e-02 5.58699906e-01 -7.30826676e-01
4.41008598e-01 2.36185834e-01 2.51466602e-01 -2.40856096e-01
7.54274786e-01 -1.45709842e-01 -1.66740254e-01 5.91860771e-01
-6.20339274e-01 1.05539948e-01 2.05563858e-01 2.16043800e-01
8.60075116e-01 -4.28345770e-01 4.04820591e-02 -7.18969926e-02
3.44685644e-01 8.42859149e-02 2.22514510e-01 1.56531215e-01
-1.95714831e-01 9.10955295e-02 7.11531878e-01 -2.83873975e-01
-5.27850747e-01 -4.34832662e-01 6.12479262e-02 7.86642373e-01
4.90948185e-02 -6.08310938e-01 -1.43095762e-01 -5.99311471e-01
2.05404848e-01 7.13905573e-01 -6.30356491e-01 3.10581326e-02
5.98479733e-02 -7.48324692e-01 -3.10333788e-01 -1.79291293e-01
4.24825937e-01 -3.49120080e-01 -4.97231603e-01 -7.99043402e-02
-8.69673677e-03 -8.76311064e-01 -2.40113795e-01 8.43820162e-03
-5.38765669e-01 -1.12597549e+00 -3.79674852e-01 8.13362971e-02
6.28849387e-01 4.80959624e-01 9.33049738e-01 5.06650172e-02
1.46984056e-01 4.13706541e-01 -4.82779920e-01 -6.36397302e-01
-2.65167981e-01 4.48567152e-01 -1.22037726e-02 2.41398945e-01
3.37290794e-01 -3.07814032e-01 -6.98403597e-01 6.39867127e-01
-6.21709585e-01 -4.81633872e-01 4.54685926e-01 2.43129171e-02
-5.49633652e-02 3.60562086e-01 8.45331192e-01 -9.60690498e-01
9.10998762e-01 -9.48751390e-01 -5.39557159e-01 -3.42708498e-01
-1.23372757e+00 -6.03686154e-01 3.46203953e-01 3.36196512e-01
-6.89875901e-01 -5.99521935e-01 1.37455417e-02 3.48109454e-01
3.35129738e-01 1.00554812e+00 2.63145357e-01 -2.09095523e-01
3.89616847e-01 -4.08022344e-01 2.30215624e-01 -2.31723696e-01
1.49385810e-01 7.17345059e-01 -3.30341339e-01 2.79450685e-01
7.14283526e-01 5.95318496e-01 -3.78738612e-01 -6.57703698e-01
-7.36358285e-01 -7.14894176e-01 -2.42501706e-01 -7.74445117e-01
5.10158837e-01 -9.72307324e-01 -8.76288652e-01 9.82469395e-02
-4.86469746e-01 1.52649777e-02 -1.41871780e-01 5.93706787e-01
3.01877230e-01 5.29311411e-02 -3.65358263e-01 -7.26406097e-01
8.24880134e-03 -8.75865936e-01 9.91376024e-03 3.09070170e-01
-6.67049468e-01 -1.07961893e+00 2.46209487e-01 8.79715323e-01
7.47644663e-01 2.54417837e-01 6.49932444e-01 -9.16560173e-01
-7.82424212e-01 -6.89924777e-01 6.57668710e-03 6.87005877e-01
3.80445033e-01 3.63175035e-01 -5.25844395e-01 -2.71613240e-01
-1.20404735e-02 2.09588423e-01 3.32920611e-01 -9.09090266e-02
1.52586311e-01 -2.64747739e-02 -2.65611380e-01 -3.12422603e-01
1.29143381e+00 1.69652671e-01 2.78060406e-01 8.09820056e-01
1.22661740e-01 9.35554564e-01 1.20731723e+00 6.64264441e-01
5.79814255e-01 3.82080227e-01 7.68486261e-01 -1.02707006e-01
5.23780584e-01 -2.38791242e-01 5.82883477e-01 1.10312188e+00
-7.52992332e-02 -5.62515669e-02 -7.75917470e-01 6.70002162e-01
-1.49929488e+00 -4.26192790e-01 -1.45639822e-01 2.35937810e+00
2.13761888e-02 6.78321421e-01 3.87861729e-01 1.22955609e-02
4.08758074e-01 2.66055614e-01 -2.09305942e-01 -4.75881308e-01
4.67061810e-02 -1.60025030e-01 1.01119196e+00 2.40471184e-01
-2.98650086e-01 2.48822600e-01 6.32552433e+00 2.15105608e-01
-1.33178377e+00 -1.49989545e-01 7.79622972e-01 -4.56413180e-01
-7.11545885e-01 2.21093699e-01 -7.19147444e-01 2.63754666e-01
1.16380250e+00 -4.54938471e-01 2.94602424e-01 5.12429833e-01
7.98104525e-01 -4.94609743e-01 -4.71547812e-01 1.96700931e-01
1.04380831e-01 -1.33170688e+00 -4.31756139e-01 5.48589945e-01
7.77485073e-01 4.50453013e-01 2.75596470e-01 -3.06079127e-02
2.26667747e-02 -5.49905837e-01 5.19947708e-01 5.86486816e-01
-2.03479503e-04 -9.22360957e-01 9.91022766e-01 2.80420661e-01
-5.55435956e-01 -4.64745909e-01 1.62266418e-01 -7.03349769e-01
3.17490906e-01 9.39226151e-01 -9.37525094e-01 5.66278756e-01
5.45270681e-01 7.82142997e-01 -3.16502243e-01 6.43484175e-01
-4.19917814e-02 5.12876451e-01 1.84779149e-02 -4.38511938e-01
3.82961899e-01 -5.10200381e-01 4.91871983e-01 4.60818648e-01
1.37423486e-01 -2.17490822e-01 -3.48061502e-01 5.37285328e-01
1.01911812e-03 4.60365444e-01 -5.86100399e-01 -7.49140263e-01
1.69228181e-01 1.55972743e+00 -7.98331261e-01 1.22183144e-01
-5.32803535e-01 2.73271263e-01 -2.98886031e-01 1.67273000e-01
-6.39680803e-01 -3.69119912e-01 8.41078639e-01 8.00662100e-01
2.41298467e-01 -6.10835135e-01 -3.17116141e-01 -7.44316876e-01
-1.71749488e-01 -9.21704471e-01 -1.60149366e-01 -6.01535738e-01
-8.20766091e-01 5.44350326e-01 -5.21168768e-01 -1.02303612e+00
1.10935405e-01 -4.11322683e-01 -5.77596486e-01 8.70174706e-01
-1.65851009e+00 -6.16233885e-01 3.08183469e-02 5.90464063e-02
1.73891470e-01 -2.32179165e-01 3.72795850e-01 3.91018033e-01
-5.85824788e-01 -1.44387960e-01 7.71318302e-02 -1.96712568e-01
6.01553917e-01 -6.60367429e-01 4.22350824e-01 4.02081400e-01
8.10360536e-02 9.42671359e-01 5.73442042e-01 -7.98993170e-01
-1.24387312e+00 -4.40357924e-01 1.11920369e+00 -5.99558890e-01
1.28057361e+00 -1.60192907e-01 -1.79377809e-01 3.52719367e-01
3.75965714e-01 -7.76966155e-01 1.16513860e+00 5.28939843e-01
1.45593554e-01 -3.65795255e-01 -7.20272243e-01 5.86366415e-01
3.45010430e-01 -5.47868431e-01 -1.28116403e-02 3.28706056e-01
5.39359868e-01 2.49401167e-01 -1.10851622e+00 5.02364058e-03
7.13024437e-01 -1.34510791e+00 2.83915728e-01 -2.74764687e-01
3.62230569e-01 -1.54140025e-01 4.12894972e-02 -1.45452046e+00
-2.25951374e-01 -7.15477109e-01 8.48733068e-01 1.11471403e+00
9.21925366e-01 -1.09723186e+00 7.89204597e-01 1.00293648e+00
1.53962970e-01 -1.06689537e+00 -4.59879935e-01 -2.89105803e-01
-3.78022283e-01 -4.75582749e-01 4.67104256e-01 7.34968722e-01
2.38325119e-01 2.17389241e-01 2.91728899e-02 -4.93010096e-02
3.61300856e-02 2.55034864e-01 8.41848016e-01 -1.09013331e+00
-1.57663390e-01 -2.68196791e-01 -1.59062818e-01 -6.98114812e-01
-2.88071960e-01 -7.15658486e-01 -6.74606264e-01 -1.17534471e+00
-1.86205432e-01 -7.77051449e-01 -5.05008221e-01 -2.24236175e-02
3.77302080e-01 2.48163164e-01 3.87197882e-01 2.95499980e-01
-3.22907209e-01 1.28625944e-01 1.21567678e+00 1.21634692e-01
-1.18261926e-01 5.99110544e-01 -1.43403757e+00 6.22147083e-01
7.48082340e-01 -3.30287337e-01 -1.42521918e-01 5.98746277e-02
1.36752737e+00 -1.55454278e-01 9.70908478e-02 -5.86759031e-01
2.23629758e-01 -1.63332775e-01 -1.06890507e-01 -3.43320131e-01
1.81621566e-01 -1.11482191e+00 3.46545935e-01 3.15762311e-01
-3.23310226e-01 4.63302791e-01 -4.99284528e-02 3.81236613e-01
-5.01473010e-01 -2.01794893e-01 -1.14234742e-02 -8.95757824e-02
-1.21364467e-01 -2.74677221e-02 -7.91020989e-01 -3.97260189e-01
1.00732267e+00 -1.98284134e-01 -1.96793407e-01 -8.13465118e-01
-6.49836659e-01 1.57400861e-01 7.89293885e-01 6.82792664e-01
2.15826500e-02 -8.27092707e-01 -4.08266187e-01 1.78864822e-01
-6.62055518e-03 -7.68566668e-01 8.64452571e-02 1.04064238e+00
-2.20610365e-01 5.54925084e-01 -9.50936750e-02 -8.64071399e-02
-1.08885920e+00 1.43695055e-02 1.27175525e-01 -2.45409027e-01
2.48188168e-01 6.44790351e-01 -3.97088170e-01 -6.23885132e-02
-1.68213904e-01 -2.85051942e-01 -3.54056418e-01 7.80420184e-01
1.85763776e-01 6.74841046e-01 3.89451057e-01 -5.22608697e-01
-2.03901604e-01 2.11498916e-01 -1.21716194e-01 -1.13821790e-01
1.32464528e+00 -3.91674548e-01 -1.67689770e-01 8.22927117e-01
9.56324041e-01 7.85154641e-01 -7.30486929e-01 3.41828376e-01
-3.25850323e-02 -8.10503662e-01 2.59580374e-01 -9.14873421e-01
-1.15128231e+00 5.28609395e-01 1.72051772e-01 8.47135007e-01
6.96892679e-01 -2.90088356e-01 7.94295907e-01 -6.62776977e-02
4.27536786e-01 -9.71463740e-01 -2.33196858e-02 2.30840206e-01
6.80661321e-01 -9.78588223e-01 2.54980326e-01 -3.33045781e-01
-7.87911534e-01 8.51857126e-01 4.22502458e-02 7.16441870e-02
8.60581458e-01 -3.11277449e-01 2.08555028e-01 -3.29029322e-01
-8.95479500e-01 -8.62607826e-03 -1.80821232e-02 -2.24533919e-02
5.38976967e-01 2.54722923e-01 -4.81484354e-01 7.00506508e-01
-1.75253853e-01 3.85935903e-02 8.75618696e-01 5.71231246e-01
-4.53400165e-01 -1.46852338e+00 -6.56101704e-02 4.70302969e-01
-5.79634428e-01 -9.00658742e-02 -6.52040243e-01 8.91341984e-01
1.82522103e-01 1.01700258e+00 -1.47857308e-01 -5.87344766e-01
6.38593554e-01 -2.82714814e-01 -1.32325456e-01 -5.60747862e-01
-1.25698519e+00 -1.19487748e-01 6.69133067e-01 -2.86835283e-01
-4.09049451e-01 -1.00301957e+00 -7.05831230e-01 -4.93154138e-01
-6.45035505e-01 6.07491672e-01 1.32783163e+00 6.80833280e-01
6.15041435e-01 3.70718390e-01 1.12411535e+00 -2.89348572e-01
-3.23535413e-01 -6.77629173e-01 -8.04514349e-01 3.90250981e-02
8.17005783e-02 -3.15288156e-01 -7.77798116e-01 -6.27801776e-01] | [10.686920166015625, 6.8262410163879395] |
2d2a7aa3-7d7a-4517-8c83-b8f3e07d6bdc | bayesian-persuasion-in-sequential-trials | 2110.09594 | null | https://arxiv.org/abs/2110.09594v3 | https://arxiv.org/pdf/2110.09594v3.pdf | Bayesian Persuasion in Sequential Trials | We consider a Bayesian persuasion or information design problem where the sender tries to persuade the receiver to take a particular action via a sequence of signals. This we model by considering multi-phase trials with different experiments conducted based on the outcomes of prior experiments. In contrast to most of the literature, we consider the problem with constraints on signals imposed on the sender. This we achieve by fixing some of the experiments in an exogenous manner; these are called determined experiments. This modeling helps us understand real-world situations where this occurs: e.g., multi-phase drug trials where the FDA determines some of the experiments, funding of a startup by a venture capital firm, start-up acquisition by big firms where late-stage assessments are determined by the potential acquirer, multi-round job interviews where the candidates signal initially by presenting their qualifications but the rest of the screening procedures are determined by the interviewer. The non-determined experiments (signals) in the multi-phase trial are to be chosen by the sender in order to persuade the receiver best. With a binary state of the world, we start by deriving the optimal signaling policy in the only non-trivial configuration of a two-phase trial with binary-outcome experiments. We then generalize to multi-phase trials with binary-outcome experiments where the determined experiments can be placed at any chosen node in the trial tree. Here we present a dynamic programming algorithm to derive the optimal signaling policy that uses the two-phase trial solution's structural insights. We also contrast the optimal signaling policy structure with classical Bayesian persuasion strategies to highlight the impact of the signaling constraints on the sender. | ['Grant Schoenebeck', 'Vijay G. Subramanian', 'Shih-Tang Su'] | 2021-10-18 | null | null | null | null | ['persuasion-strategies'] | ['computer-vision'] | [ 7.54972458e-01 5.60713887e-01 -4.84785795e-01 -3.06759536e-01
-8.31122518e-01 -7.08229721e-01 5.78012109e-01 3.62400264e-01
-9.38134909e-01 8.93926442e-01 1.19599812e-01 -9.51202691e-01
-7.14060605e-01 -5.94838917e-01 -7.01633036e-01 -8.43855500e-01
2.81033576e-01 8.95196736e-01 -1.83405817e-01 3.08682323e-02
5.20075381e-01 3.05424273e-01 -7.89909244e-01 8.92196819e-02
3.01806748e-01 4.55811381e-01 1.87228382e-01 6.32316887e-01
2.68027782e-01 3.90331417e-01 -6.57435596e-01 -3.28269690e-01
4.20089811e-01 -5.00681698e-01 -5.45729518e-01 1.89198375e-01
-5.51204979e-01 -4.72670138e-01 3.20879549e-01 8.81950021e-01
7.26938009e-01 -1.35574952e-01 6.31304741e-01 -1.14664972e+00
-1.33537009e-01 1.00254965e+00 -7.03071117e-01 -9.58174653e-03
5.39351285e-01 3.08843434e-01 9.31557178e-01 -7.02236444e-02
5.52582443e-01 1.48086345e+00 3.64928730e-02 3.22546452e-01
-1.59438503e+00 -5.69587409e-01 1.81641817e-01 5.57694733e-02
-1.07717180e+00 -2.31068760e-01 6.66149199e-01 -6.88297570e-01
1.35756275e-02 2.29741827e-01 4.06285852e-01 1.28144419e+00
3.53430480e-01 2.29659706e-01 1.78518236e+00 -5.94817221e-01
8.41331601e-01 2.22879961e-01 3.73651117e-01 3.66725065e-02
6.70504093e-01 4.73522246e-01 -2.59772092e-01 -5.13756990e-01
3.68947238e-01 -1.84239358e-01 -2.57834911e-01 -3.03075641e-01
-8.24127316e-01 1.01943707e+00 -6.28857175e-03 2.98409760e-01
-8.91392350e-01 3.57276201e-02 -1.56235725e-01 4.35322762e-01
-3.84107828e-01 6.42971098e-01 -5.24780393e-01 3.24281082e-02
-6.77334845e-01 3.35180253e-01 7.42783606e-01 2.89279968e-01
6.33716464e-01 -5.35528362e-01 -5.07257104e-01 9.45835561e-02
6.16588354e-01 8.38170424e-02 -1.24706931e-01 -8.56743157e-01
4.39588696e-01 9.82423574e-02 7.70915270e-01 -4.51437622e-01
-3.35980117e-01 -4.83122408e-01 -2.17503697e-01 2.29472145e-01
7.54203439e-01 -7.84699798e-01 -8.67435515e-01 1.78441644e+00
3.93985152e-01 -3.99651192e-02 9.85396951e-02 8.91017139e-01
1.88336343e-01 5.84132373e-01 3.83605599e-01 -7.58921266e-01
1.59364450e+00 7.18005896e-02 -6.89862192e-01 -2.53064245e-01
4.82904196e-01 -6.49403572e-01 6.41530812e-01 5.57605505e-01
-8.29742968e-01 2.53534578e-02 -1.02787495e+00 6.69141173e-01
4.44812961e-02 -7.56633803e-02 3.03376108e-01 9.68024552e-01
-5.72505891e-01 3.95537049e-01 -4.66698259e-01 -9.52994078e-02
3.28466207e-01 5.45622885e-01 5.73601089e-02 -3.02405655e-01
-1.41013956e+00 7.63284445e-01 -6.96754688e-03 2.90207863e-01
-1.22072911e+00 -5.91848910e-01 -4.46172148e-01 1.79235414e-01
9.96863306e-01 -7.12205648e-01 1.34281182e+00 -8.92088115e-01
-1.45024610e+00 5.09624898e-01 1.16437271e-01 -2.39031091e-01
7.31298029e-01 3.46600533e-01 3.53434272e-02 -1.68183669e-01
3.68084788e-01 4.58332330e-01 4.71471578e-01 -1.22750664e+00
-4.96891081e-01 -5.44113636e-01 4.65156466e-01 1.60048261e-01
5.14725626e-01 3.64702165e-01 8.72471556e-02 -6.86976686e-02
-3.11371893e-01 -1.27179623e+00 -6.40148401e-01 -6.65796757e-01
-7.00113893e-01 -2.20885873e-02 4.74753827e-01 -9.66031328e-02
8.39878798e-01 -1.92457819e+00 -4.53801975e-02 3.53944719e-01
-1.40157729e-01 -2.17439443e-01 -8.68776962e-02 5.37143290e-01
-1.01705149e-01 3.07020426e-01 -2.49147695e-02 6.55853450e-02
1.78460404e-01 1.01133056e-01 4.41249423e-02 5.61408937e-01
4.72775288e-03 4.28054243e-01 -7.30818748e-01 -1.89695537e-01
-3.47566485e-01 -1.69303101e-02 -4.63498682e-01 1.11067578e-01
-6.64843991e-02 7.82279670e-01 -1.01595533e+00 2.87543893e-01
5.42505085e-01 -2.05167338e-01 8.34096909e-01 3.36195648e-01
-1.94502637e-01 3.53861809e-01 -1.44264221e+00 9.45946753e-01
1.39891617e-02 1.18378974e-01 4.47089583e-01 -1.14509010e+00
4.69309419e-01 4.43138719e-01 3.46646935e-01 -2.35406965e-01
5.07040322e-01 6.59141690e-02 6.56162322e-01 -2.32858911e-01
1.98606834e-01 -6.39086843e-01 -3.73486519e-01 5.99787533e-01
-5.01040280e-01 -4.69273776e-02 5.05514704e-02 3.06246787e-01
1.28025770e+00 -3.08408946e-01 3.24019730e-01 -3.09643567e-01
1.89638481e-01 1.16524838e-01 1.00907469e+00 1.25667727e+00
-1.10933810e-01 1.36687189e-01 1.21599329e+00 3.01721901e-01
-6.70440316e-01 -4.70175445e-01 -9.98434052e-02 6.39866889e-01
-9.63800028e-02 3.45126912e-02 -6.59099579e-01 -6.26349270e-01
4.54109833e-02 8.95057857e-01 -7.54752398e-01 4.12298515e-02
-8.39345977e-02 -8.75245631e-01 -9.55396593e-02 -4.20747995e-02
3.40722203e-01 -5.23420691e-01 -9.00294483e-01 3.84524941e-01
1.31402284e-01 -7.42529452e-01 -3.10417950e-01 5.67315578e-01
-4.71045524e-01 -1.11238372e+00 -6.10842168e-01 -1.03578985e-01
7.01280713e-01 -2.08528131e-01 3.30987662e-01 -4.21566844e-01
1.08175404e-01 5.04272640e-01 -1.08353764e-01 -7.83628047e-01
-6.03762865e-01 -3.96054924e-01 3.16797756e-03 1.86113074e-01
3.12588304e-01 -4.60633151e-02 -6.35770500e-01 4.53551084e-01
-9.08557415e-01 -4.16965723e-01 7.09515870e-01 7.98214018e-01
3.96711230e-01 3.91825765e-01 7.59319961e-01 -1.02710378e+00
8.27902555e-01 -6.47126079e-01 -1.09758914e+00 3.19653749e-01
-5.22077143e-01 1.31015867e-01 3.79012339e-02 -8.55759561e-01
-1.10843301e+00 2.25245878e-01 3.81619453e-01 2.64173746e-01
-2.47213587e-01 7.28296757e-01 -6.13084793e-01 5.26795201e-02
5.69452584e-01 -4.57809091e-01 9.53361113e-03 -3.71043980e-01
1.53748959e-03 7.41735697e-01 -1.48747295e-01 -9.99709368e-01
5.02589405e-01 7.34681860e-02 4.17914510e-01 -4.23253477e-01
-5.97645521e-01 1.44512638e-01 -8.76978189e-02 -2.28970200e-01
7.95589209e-01 -7.80707598e-01 -1.26010311e+00 1.33250535e-01
-1.06812108e+00 -4.29679781e-01 -1.12039953e-01 9.45019543e-01
-4.25328821e-01 8.10268819e-02 -2.13495344e-01 -1.19174933e+00
6.09070957e-01 -1.74679327e+00 6.31821871e-01 4.38385785e-01
-2.35325277e-01 -5.66276312e-01 -7.90979192e-02 6.26706302e-01
1.48852557e-01 4.54850718e-02 1.04679239e+00 -1.03272605e+00
-8.72724056e-01 2.21836083e-02 4.07007515e-01 -2.11886123e-01
-6.16592094e-02 -3.06243479e-01 -4.27138925e-01 -1.47220418e-01
2.09695175e-01 -9.71307904e-02 2.63034195e-01 1.04166734e+00
4.97323275e-01 -2.80747920e-01 -5.69428325e-01 -3.06838483e-01
1.04964030e+00 1.07696116e+00 4.87693518e-01 5.26217043e-01
-1.81287512e-01 1.10761404e+00 7.92955995e-01 5.07518888e-01
1.66763797e-01 9.01197135e-01 1.77321404e-01 9.56958830e-02
5.87043643e-01 -1.88914075e-01 3.67924154e-01 -4.30255622e-01
3.51993144e-01 -2.60274917e-01 -6.34589434e-01 3.96320581e-01
-1.80151141e+00 -5.64521909e-01 -1.72077507e-01 2.52218795e+00
9.66675341e-01 3.71212870e-01 3.35478365e-01 2.10465714e-01
9.98862326e-01 -3.15854877e-01 -4.43701267e-01 -5.36270440e-01
1.80216432e-01 2.62185242e-02 8.05778503e-01 6.78799272e-01
-4.76939499e-01 3.68767828e-01 6.42357588e+00 5.48283875e-01
-7.23507583e-01 1.71314061e-01 1.18404460e+00 -8.37314874e-03
-4.57414061e-01 8.65978360e-01 -1.09179139e+00 4.66315567e-01
1.08780932e+00 -2.35045463e-01 1.41189545e-01 2.14054987e-01
7.15773821e-01 -7.41505265e-01 -1.46018279e+00 3.84054214e-01
-6.20045245e-01 -9.81881499e-01 -6.06761038e-01 5.22141814e-01
4.73708004e-01 -7.01195240e-01 3.00940983e-02 8.96731298e-03
1.01550972e+00 -7.71416903e-01 6.86768591e-01 2.14423329e-01
4.81879741e-01 -5.12449861e-01 6.02565527e-01 6.51427090e-01
-3.66995990e-01 -5.70064247e-01 6.15059622e-02 -4.38418746e-01
4.32932943e-01 8.30279231e-01 -1.14873290e+00 5.66515803e-01
2.92126328e-01 -3.37668210e-01 9.67133045e-02 8.46574187e-01
-5.25094807e-01 8.79564226e-01 -2.65666306e-01 -3.72729242e-01
3.11516464e-01 -4.26201522e-01 5.75463533e-01 6.22992337e-01
3.01000535e-01 5.34431159e-01 8.46354663e-02 1.07015836e+00
2.33347520e-01 -1.29506484e-01 -5.82266092e-01 -1.65107623e-01
5.00228047e-01 9.53879118e-01 -7.46413887e-01 -1.41569212e-01
-2.02816669e-02 -1.10624775e-01 -4.42926198e-01 7.10232198e-01
-4.20184404e-01 -1.89406514e-01 3.47624391e-01 3.69707078e-01
2.22054258e-01 1.16415210e-01 -2.10134581e-01 -4.03399080e-01
-2.75525123e-01 -1.05532205e+00 4.68308449e-01 -7.12896049e-01
-8.08828354e-01 -2.34367847e-01 6.57280624e-01 -4.98419791e-01
-2.72565633e-01 -4.05629069e-01 -4.42549229e-01 1.11149108e+00
-1.33174992e+00 -3.39730650e-01 5.73585868e-01 1.84506983e-01
1.81774482e-01 1.26042068e-01 8.02843347e-02 -2.33058538e-02
-6.36134744e-01 1.81792781e-01 -1.74306661e-01 -9.15022120e-02
4.83631998e-01 -9.62282479e-01 -4.68933284e-01 7.14507937e-01
-3.68644327e-01 8.38805437e-01 9.49680209e-01 -9.78041708e-01
-1.41163051e+00 -3.81669790e-01 8.07813644e-01 -1.40964225e-01
6.60391629e-01 -3.32216918e-01 -3.80962461e-01 6.31554484e-01
1.36970475e-01 -7.00187147e-01 4.72818136e-01 2.01894298e-01
2.72283703e-01 -1.05710626e-01 -1.20437717e+00 8.11853349e-01
3.56629431e-01 -2.09928341e-02 -6.11359715e-01 3.25811207e-01
3.43788892e-01 -6.39627650e-02 -5.25863171e-01 2.40621537e-01
2.62328267e-01 -3.56243491e-01 5.55106997e-01 -7.53713310e-01
9.57837701e-02 -4.43028331e-01 -2.55317464e-02 -1.36536527e+00
-3.79005343e-01 -1.14777625e+00 8.93257439e-01 1.14372146e+00
7.42626905e-01 -8.73156607e-01 7.08397031e-01 1.00600898e+00
4.12581831e-01 -4.28300232e-01 -1.15318811e+00 -4.66844499e-01
8.09296146e-02 -1.45302247e-02 6.00169241e-01 6.14542902e-01
5.89515045e-02 6.56986713e-01 -3.02295506e-01 2.95642674e-01
7.98013508e-01 -1.22298207e-02 6.33797586e-01 -1.26491034e+00
-7.94202328e-01 -1.24777138e-01 9.74621847e-02 -1.07278752e+00
3.17173302e-02 -3.20307732e-01 3.23099285e-01 -1.23285222e+00
4.91560340e-01 -7.09595501e-01 -1.51338324e-01 4.94833648e-01
-1.57623962e-01 -7.65044630e-01 1.02088630e-01 -1.88474223e-01
-1.81637555e-02 -6.49004728e-02 1.05615234e+00 -2.92930841e-01
-3.38814020e-01 4.59601790e-01 -1.42638171e+00 3.31908345e-01
3.92766714e-01 -9.18314755e-01 -4.51104730e-01 1.50529280e-01
4.87733275e-01 9.36831355e-01 3.19198370e-01 -2.43058261e-02
2.87589312e-01 -6.19338930e-01 7.67679140e-02 -2.05118611e-01
8.45125616e-02 -8.47997904e-01 8.65634501e-01 7.19227076e-01
-6.69044614e-01 -1.97121650e-01 -1.46284059e-01 8.24276268e-01
3.04873049e-01 -7.94590414e-01 6.89365149e-01 -1.42760664e-01
2.85771161e-01 -3.10235620e-01 -9.79067326e-01 -3.87068450e-01
1.14142144e+00 -8.03550631e-02 -5.81358671e-01 -6.52648091e-01
-1.03023720e+00 4.57350433e-01 2.19537139e-01 -1.48140982e-01
1.56921640e-01 -7.55920529e-01 -7.53705144e-01 -4.40878540e-01
-3.55319738e-01 -2.04169810e-01 6.83418736e-02 1.11953759e+00
2.65579075e-01 3.60864252e-01 2.49439664e-02 -4.59616870e-01
-1.29390049e+00 4.85221803e-01 1.75539106e-01 -5.11164486e-01
9.22305062e-02 4.70680565e-01 4.62607771e-01 1.77753136e-01
4.36905846e-02 1.60234496e-02 -2.65084326e-01 1.58354729e-01
2.09479451e-01 1.32052094e-01 -9.72990021e-02 -2.42797211e-01
-4.02176768e-01 2.92529583e-01 -2.37496942e-01 -8.37416947e-01
1.22979677e+00 2.99714729e-02 1.04205556e-01 3.79879296e-01
5.89824617e-01 1.52073205e-01 -1.16979194e+00 -1.21139415e-01
3.65968086e-02 -4.54976737e-01 2.32664138e-01 -1.15272021e+00
-6.45815313e-01 3.50499779e-01 3.30900341e-01 7.75991455e-02
8.82509232e-01 -3.04510389e-02 -2.07901642e-01 2.61363208e-01
5.54165363e-01 -1.00983679e+00 -1.58555917e-02 -7.77313411e-02
5.74771821e-01 -8.78376245e-01 2.03176618e-01 -3.03417444e-01
-7.51386225e-01 4.17527288e-01 6.66474272e-03 4.10089612e-01
6.08107865e-01 1.83511451e-01 -1.15300141e-01 -2.73061514e-01
-8.17963183e-01 4.80648130e-02 -9.10597518e-02 4.08194125e-01
2.14822143e-02 2.99014896e-01 -1.20395911e+00 7.66047120e-01
2.64760613e-01 9.77442861e-02 1.10686696e+00 1.15117550e+00
-2.41242096e-01 -1.79695475e+00 -9.37059820e-01 3.87874156e-01
-7.71389425e-01 7.87173808e-02 -7.61517704e-01 7.59428442e-01
-1.41037963e-02 1.47080338e+00 -3.03539127e-01 8.78688321e-02
4.59377289e-01 1.10293940e-01 2.56336480e-01 -6.48261487e-01
-6.82002425e-01 6.34313226e-01 5.00756323e-01 -5.34248874e-02
-4.73056018e-01 -1.10756099e+00 -9.08701539e-01 5.96390851e-02
-4.79367733e-01 7.48548329e-01 9.10540462e-01 1.02340460e+00
4.31468040e-01 5.06324112e-01 9.25623357e-01 -5.22927463e-01
-9.56509769e-01 -9.20528531e-01 -8.39648604e-01 -8.98818597e-02
2.31960341e-01 -8.69923055e-01 -5.14399767e-01 -2.90963560e-01] | [7.912792682647705, 5.2510905265808105] |
57d4f134-265c-46a0-ba15-f2822a5743e3 | scale-invariant-adversarial-attack-for | 2201.12527 | null | https://arxiv.org/abs/2201.12527v1 | https://arxiv.org/pdf/2201.12527v1.pdf | Scale-Invariant Adversarial Attack for Evaluating and Enhancing Adversarial Defenses | Efficient and effective attacks are crucial for reliable evaluation of defenses, and also for developing robust models. Projected Gradient Descent (PGD) attack has been demonstrated to be one of the most successful adversarial attacks. However, the effect of the standard PGD attack can be easily weakened by rescaling the logits, while the original decision of every input will not be changed. To mitigate this issue, in this paper, we propose Scale-Invariant Adversarial Attack (SI-PGD), which utilizes the angle between the features in the penultimate layer and the weights in the softmax layer to guide the generation of adversaries. The cosine angle matrix is used to learn angularly discriminative representation and will not be changed with the rescaling of logits, thus making SI-PGD attack to be stable and effective. We evaluate our attack against multiple defenses and show improved performance when compared with existing attacks. Further, we propose Scale-Invariant (SI) adversarial defense mechanism based on the cosine angle matrix, which can be embedded into the popular adversarial defenses. The experimental results show the defense method with our SI mechanism achieves state-of-the-art performance among multi-step and single-step defenses. | ['Daoqiang Zhang', 'Zhongnian Li', 'Tao Zhang', 'Mengting Xu'] | 2022-01-29 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [ 1.83148175e-01 -3.88579905e-01 3.71335521e-02 -2.89435893e-01
-5.27046680e-01 -1.20689917e+00 7.94098318e-01 -3.79202098e-01
-4.95101035e-01 4.64483410e-01 1.87479798e-02 -5.69095910e-01
-1.19408913e-01 -8.12042832e-01 -5.89659870e-01 -9.01543260e-01
-1.89281836e-01 -3.00499558e-01 4.36422229e-01 -6.50782108e-01
2.74311781e-01 8.30799341e-01 -3.57976973e-01 9.33046490e-02
6.15499854e-01 9.16706324e-01 -2.50990778e-01 5.29655337e-01
2.47064486e-01 1.89044520e-01 -9.62056696e-01 -6.49203837e-01
9.29775178e-01 -2.00833544e-01 -2.62986630e-01 -7.75915444e-01
2.90230274e-01 -5.50009549e-01 -9.34525073e-01 1.33517385e+00
6.75202727e-01 2.04690434e-02 5.24727046e-01 -1.43426013e+00
-6.41547620e-01 4.25342888e-01 -7.52770185e-01 1.82687730e-01
7.98334181e-02 3.49358581e-02 6.02180123e-01 -5.90960026e-01
2.04291180e-01 1.56048095e+00 4.17162389e-01 8.70109499e-01
-9.22765195e-01 -1.27900171e+00 4.93232220e-01 2.14942962e-01
-1.25785458e+00 -1.11596093e-01 1.14450717e+00 7.72583038e-02
3.98189455e-01 5.19471586e-01 3.21993604e-02 1.43364537e+00
4.72231925e-01 5.20163953e-01 1.02357364e+00 -2.88786083e-01
2.56321102e-01 8.05114284e-02 -2.68132001e-01 7.31954157e-01
4.12714064e-01 5.62718332e-01 -1.08706973e-01 -5.33666849e-01
8.01732779e-01 1.89954117e-01 -2.17875898e-01 -3.15298289e-01
-6.14478946e-01 1.18777800e+00 7.73530662e-01 1.49868964e-03
-1.11800462e-01 1.12410195e-01 2.86582142e-01 4.64961201e-01
1.66325867e-01 4.95825917e-01 -3.84993583e-01 2.39376217e-01
-5.57288766e-01 3.10977101e-01 5.38061440e-01 4.99500185e-01
1.33248404e-01 5.11933208e-01 -2.00912446e-01 6.01957500e-01
4.28005785e-01 6.94402516e-01 3.59941363e-01 -5.58588505e-01
7.82650054e-01 4.38270599e-01 -2.00559884e-01 -1.40901828e+00
-2.01515600e-01 -4.72212881e-01 -9.53062356e-01 7.91317999e-01
2.82251805e-01 -7.42250383e-01 -8.63054037e-01 2.14532018e+00
3.91883910e-01 1.62762582e-01 6.90885931e-02 6.64853036e-01
2.49961257e-01 5.69917738e-01 -1.11848354e-01 3.25042605e-02
8.60726595e-01 -6.43718779e-01 -3.06416661e-01 -3.92834246e-01
7.22867623e-02 -8.21780384e-01 8.15440416e-01 2.05203041e-01
-7.00621068e-01 -4.00725514e-01 -1.49367559e+00 5.55299163e-01
-4.58110631e-01 -2.26287514e-01 5.44779778e-01 1.14121056e+00
-4.12124574e-01 5.63557923e-01 -7.99057424e-01 2.09515840e-01
3.55369270e-01 5.18323720e-01 -5.09149432e-01 7.07721934e-02
-1.57485843e+00 8.04921746e-01 2.02099353e-01 8.29085112e-02
-1.00796306e+00 -3.62128317e-01 -6.83745921e-01 -1.44045493e-02
2.33711645e-01 -2.81712890e-01 6.71009958e-01 -3.79478216e-01
-1.54754674e+00 2.14442730e-01 3.19980323e-01 -6.06998026e-01
5.11239350e-01 -1.19175777e-01 -5.82171679e-01 1.07356146e-01
-3.92904043e-01 2.30000585e-01 1.16631877e+00 -1.10818291e+00
-3.37720603e-01 -4.49789762e-01 5.73808789e-01 1.73310116e-01
-8.83021772e-01 1.42000675e-01 -2.46359020e-01 -1.11975515e+00
6.56457469e-02 -1.21420646e+00 -3.28387618e-01 2.78213322e-02
-5.91263473e-01 2.34639987e-01 1.40011311e+00 -6.29263282e-01
1.40271699e+00 -2.39073157e+00 1.08751498e-01 5.00618160e-01
2.43483987e-02 7.78708160e-01 -3.51977229e-01 4.75659221e-01
-2.48027146e-01 1.70283645e-01 -1.72385424e-01 3.98558415e-02
1.21696167e-01 9.34361368e-02 -8.67885768e-01 5.32501817e-01
7.31904656e-02 5.44754922e-01 -6.77387118e-01 1.60041042e-02
8.07461366e-02 4.71068114e-01 -5.08220375e-01 3.28588188e-01
3.51295292e-01 3.15643340e-01 -7.01041162e-01 4.65045631e-01
1.04810894e+00 5.56379855e-01 2.52177566e-02 -3.12124074e-01
3.48327607e-01 7.01827481e-02 -1.36688936e+00 1.06682467e+00
-4.14811790e-01 2.73983598e-01 1.09503396e-01 -9.33542490e-01
1.04067159e+00 2.33819082e-01 2.35872969e-01 -5.20820320e-01
6.47547916e-02 -7.21734622e-03 1.84675843e-01 1.13455951e-01
-2.37607583e-02 1.16037233e-02 -4.82633859e-01 3.08945686e-01
-2.99011648e-01 -2.24533781e-01 -1.55775592e-01 3.49212646e-01
1.10580432e+00 -1.75605208e-01 8.50325525e-02 -1.82921533e-02
8.58547986e-01 -7.08191037e-01 7.93601573e-01 8.11441541e-01
-3.00767094e-01 3.36615264e-01 6.62522376e-01 -4.45037007e-01
-6.74102187e-01 -1.31430745e+00 2.80567762e-02 8.68520737e-01
3.60018104e-01 -3.69353950e-01 -7.81229436e-01 -1.39494812e+00
1.67761728e-01 5.57145178e-01 -5.88254094e-01 -7.04530239e-01
-7.65114605e-01 -8.76937628e-01 1.12101293e+00 6.22227728e-01
7.97127664e-01 -6.01745009e-01 3.92881501e-03 -1.36555163e-02
3.40516239e-01 -9.43915009e-01 -6.39234483e-01 1.55420214e-01
-3.66090685e-01 -1.06890798e+00 -4.02415663e-01 -5.21932662e-01
8.89843822e-01 3.88552755e-01 2.01513037e-01 -1.01439707e-01
-3.98052931e-02 1.84335783e-02 -3.40591699e-01 -5.27455211e-01
-3.05567443e-01 -2.03664586e-01 5.16860187e-01 5.83467521e-02
8.31170566e-03 -6.88587129e-01 -5.95465541e-01 5.48237741e-01
-1.19903231e+00 -4.81211275e-01 6.23110831e-01 9.68889475e-01
1.23385362e-01 3.36651862e-01 5.80941617e-01 -7.81430066e-01
1.13393569e+00 -2.90637553e-01 -5.90949535e-01 2.00046092e-01
-4.92601424e-01 1.34672478e-01 1.29784763e+00 -7.64608562e-01
-8.84046078e-01 -2.69190580e-01 -4.43011433e-01 -4.80388701e-01
2.66149402e-01 3.29561047e-02 -4.29683536e-01 -8.59748721e-01
7.38648772e-01 6.79284781e-02 -4.43053037e-01 -3.59039605e-01
4.46639746e-01 5.38333833e-01 5.50580084e-01 -6.65311992e-01
1.76006460e+00 3.90383184e-01 3.76006544e-01 -4.02464092e-01
-5.56552470e-01 9.29065421e-02 -2.90956020e-01 2.43430421e-01
4.63610470e-01 -4.92567211e-01 -5.88940024e-01 6.75264060e-01
-9.27083910e-01 8.05902034e-02 2.13309839e-01 4.56789672e-01
8.72356743e-02 6.08897567e-01 -6.56368554e-01 -6.85311854e-01
-5.61433136e-01 -1.20397627e+00 5.12820780e-01 4.32318777e-01
1.75524950e-01 -9.53669965e-01 5.71316946e-03 -4.29006293e-02
6.61161363e-01 7.57506609e-01 9.13649201e-01 -9.35095191e-01
-7.49041811e-02 -8.28056037e-01 7.60331517e-03 8.42471659e-01
2.68341154e-01 -1.01983167e-01 -6.50838435e-01 -8.19699585e-01
2.75786817e-01 -2.81352192e-01 6.81974947e-01 -1.64328411e-01
1.23715138e+00 -6.84441566e-01 -2.37851262e-01 9.68055308e-01
1.17699981e+00 4.86801863e-01 4.93314594e-01 4.53267246e-01
7.31162250e-01 2.52043307e-01 5.17455637e-01 1.81720138e-01
-1.75408214e-01 8.54685247e-01 7.00474858e-01 -1.50765643e-01
2.10441291e-01 -4.08912987e-01 6.79054677e-01 4.28507209e-01
9.83931944e-02 -5.35957590e-02 -4.25780773e-01 -1.56742841e-01
-1.38871276e+00 -9.87991869e-01 3.06270152e-01 2.34266329e+00
8.19714069e-01 7.81206727e-01 -2.10900217e-01 2.36600667e-01
5.16199350e-01 5.87064087e-01 -6.36534214e-01 -7.84627616e-01
-6.04643449e-02 4.31513518e-01 7.00507760e-01 6.34045541e-01
-1.36875105e+00 9.59236681e-01 6.08293200e+00 1.23113608e+00
-1.31447458e+00 -1.06604263e-01 2.46463731e-01 -1.84516788e-01
-7.90520012e-02 9.48047191e-02 -8.79989028e-01 6.79388583e-01
5.41611969e-01 -3.68297130e-01 6.67246401e-01 8.95675600e-01
-1.14489347e-01 5.94244480e-01 -6.68671608e-01 5.23257017e-01
-8.28011800e-03 -8.40075314e-01 1.95598155e-01 1.02127828e-01
6.99292064e-01 -4.00760978e-01 6.35310173e-01 4.53015298e-01
5.73673010e-01 -7.90196598e-01 1.68435693e-01 -8.47698301e-02
7.49608219e-01 -1.14617765e+00 6.70449317e-01 2.40828753e-01
-1.16286325e+00 -3.42012107e-01 -4.70276743e-01 3.06785911e-01
-7.80003741e-02 1.59031749e-01 -6.85755372e-01 5.70635438e-01
3.30217689e-01 -5.23198508e-02 -5.12417674e-01 6.21050537e-01
-7.86704600e-01 6.75970018e-01 -3.53305191e-01 7.74107873e-02
5.62293231e-01 -2.93980818e-02 9.02893662e-01 1.06808710e+00
7.87761733e-02 -1.37811571e-01 2.47205019e-01 2.52844363e-01
-1.94979906e-01 -2.25159779e-01 -5.93740642e-01 8.83831754e-02
8.66587222e-01 1.27703893e+00 -1.47607014e-01 1.00453652e-01
-1.18238412e-01 1.05117130e+00 7.09310099e-02 3.00558031e-01
-1.19559407e+00 -1.03877950e+00 9.47415769e-01 -2.04958558e-01
1.23411171e-01 -4.37664151e-01 -1.01714116e-03 -9.77436185e-01
1.27603989e-02 -1.26303256e+00 4.61862624e-01 -2.05270499e-02
-1.38506007e+00 7.69369543e-01 7.36935204e-03 -1.27856553e+00
-3.50108415e-01 -7.48437583e-01 -1.16164601e+00 9.92725492e-01
-1.19918954e+00 -1.24705136e+00 8.06397758e-03 9.68797505e-01
3.12890410e-01 -6.24415398e-01 1.05568659e+00 2.60271698e-01
-7.95894086e-01 1.31074870e+00 2.13367820e-01 4.25331324e-01
8.74511480e-01 -1.07757914e+00 8.18646550e-01 1.32005000e+00
1.41093791e-01 1.15479112e+00 6.57708943e-01 -5.75676322e-01
-1.30453253e+00 -1.04763210e+00 1.45247262e-02 -2.21177012e-01
7.09424675e-01 -5.42632937e-01 -6.89698696e-01 4.05941278e-01
-1.11454971e-01 5.81009090e-02 7.22935855e-01 -2.83805221e-01
-8.44749033e-01 -3.43406230e-01 -1.45495868e+00 9.10744488e-01
8.19964409e-01 -5.39772689e-01 -4.61319625e-01 3.29651475e-01
7.98856080e-01 -4.67202485e-01 -4.99413490e-01 6.72469258e-01
6.66800976e-01 -7.49738932e-01 1.40684819e+00 -7.48068035e-01
1.33543164e-02 -3.57430935e-01 -3.23472589e-01 -1.45360196e+00
-5.36241889e-01 -7.88634300e-01 -3.96890104e-01 1.11041284e+00
2.17312694e-01 -9.67075944e-01 6.37324691e-01 3.27185035e-01
1.77026764e-01 -1.00055444e+00 -1.03082323e+00 -1.06682754e+00
3.36389720e-01 -7.48098269e-02 6.12798452e-01 8.45311224e-01
-2.64903247e-01 -6.01992421e-02 -7.56020963e-01 6.41652286e-01
6.82585001e-01 -3.34809572e-01 8.87850046e-01 -6.68562353e-01
-6.31474555e-01 -4.37355876e-01 -5.71020365e-01 -9.03292596e-01
1.43466191e-02 -6.79599345e-01 -5.14068127e-01 -8.69485438e-01
-3.54545116e-01 -3.47327590e-01 -6.59960270e-01 5.47649622e-01
-4.50305343e-01 3.83094877e-01 3.33403945e-01 -1.96423735e-02
-2.45771166e-02 4.23074126e-01 9.99175787e-01 -2.07378581e-01
-1.42093733e-01 3.05750132e-01 -9.74703968e-01 1.01685834e+00
9.39909279e-01 -6.18492842e-01 -4.91164297e-01 -3.04697126e-01
-1.96508840e-01 -3.55287343e-01 5.07685505e-02 -9.23405349e-01
2.43037879e-01 -4.29937392e-01 7.64899015e-01 -1.13383397e-01
4.49517548e-01 -7.12173164e-01 -2.77051777e-01 8.57958317e-01
-2.53146499e-01 2.18445554e-01 2.78904557e-01 5.58584273e-01
8.49419832e-03 -2.56706268e-01 9.82452810e-01 2.21130505e-01
-2.14487225e-01 5.32565594e-01 1.67785585e-01 -1.06917985e-01
1.18191099e+00 1.17600247e-01 -5.02533078e-01 -3.44408929e-01
-1.46671280e-01 5.54305315e-02 1.64084226e-01 6.65664732e-01
7.42662311e-01 -1.44557095e+00 -7.20603228e-01 5.22587538e-01
-3.26811671e-01 -5.21987617e-01 1.34165734e-01 1.97747096e-01
-4.06234890e-01 3.19516599e-01 -3.43008518e-01 2.38907427e-01
-1.46369886e+00 7.14141965e-01 1.35885373e-01 -4.91616428e-01
-2.62108415e-01 9.88344908e-01 2.55439013e-01 -3.47489238e-01
4.09449220e-01 4.10982013e-01 -3.09898853e-01 -4.52291459e-01
7.06369996e-01 2.61998951e-01 -2.15911061e-01 -3.78734648e-01
-4.82094526e-01 5.01176238e-01 -3.64032686e-01 -7.52106085e-02
1.21134615e+00 2.99673676e-01 8.76851752e-02 -3.92538399e-01
1.23427987e+00 7.65383005e-01 -1.31904221e+00 -1.03005163e-01
-5.98764598e-01 -8.15575957e-01 -1.66602120e-01 -8.19792569e-01
-1.16214776e+00 7.74177432e-01 6.60446167e-01 4.46704298e-01
1.20971096e+00 -7.97170460e-01 1.02542937e+00 4.23056126e-01
3.85648757e-01 -7.25202322e-01 2.38743812e-01 3.32042903e-01
9.76646543e-01 -8.50275159e-01 1.10198945e-01 -2.41263494e-01
-4.96041238e-01 1.02177465e+00 7.88976967e-01 -5.58584750e-01
5.79384625e-01 4.29021209e-01 1.34844529e-02 3.48765403e-01
-3.38409454e-01 4.94422585e-01 4.76605535e-01 6.93595648e-01
-1.33513734e-01 -1.86145902e-02 -5.74350238e-01 6.85149550e-01
-2.35479549e-01 -9.85446632e-01 8.63811970e-02 9.75991249e-01
-2.47181892e-01 -1.69085622e+00 -5.65118551e-01 9.19538084e-03
-8.01999867e-01 3.21769603e-02 -6.85742736e-01 5.56237102e-01
1.29568338e-01 9.97133434e-01 -6.19223416e-01 -9.99972522e-01
4.43039417e-01 -3.61515462e-01 3.36996704e-01 -2.57446647e-01
-5.01664340e-01 -3.25464904e-01 -2.43869796e-01 -5.71793020e-01
4.02754575e-01 -1.08600490e-01 -9.45765495e-01 -4.89355445e-01
-4.23553944e-01 2.23060578e-01 8.75268519e-01 7.05793977e-01
1.65313363e-01 4.87265736e-01 1.61040854e+00 -6.54697001e-01
-1.41482913e+00 -7.46624708e-01 -2.36769319e-01 4.75707322e-01
1.42014012e-01 -6.68327570e-01 -7.17322469e-01 -5.43232918e-01] | [5.558252334594727, 7.914463520050049] |
b60f468d-ddd9-4e53-b384-bb0f51cab2f1 | the-challenges-of-htr-model-training-feedback | 2212.11146 | null | https://arxiv.org/abs/2212.11146v3 | https://arxiv.org/pdf/2212.11146v3.pdf | The Challenges of HTR Model Training: Feedback from the Project Donner le gout de l'archive a l'ere numerique | The arrival of handwriting recognition technologies offers new possibilities for research in heritage studies. However, it is now necessary to reflect on the experiences and the practices developed by research teams. Our use of the Transkribus platform since 2018 has led us to search for the most significant ways to improve the performance of our handwritten text recognition (HTR) models which are made to transcribe French handwriting dating from the 17th century. This article therefore reports on the impacts of creating transcribing protocols, using the language model at full scale and determining the best way to use base models in order to help increase the performance of HTR models. Combining all of these elements can indeed increase the performance of a single model by more than 20% (reaching a Character Error Rate below 5%). This article also discusses some challenges regarding the collaborative nature of HTR platforms such as Transkribus and the way researchers can share their data generated in the process of creating or training handwritten text recognition models. | ['Deslandres Dominique', 'Gohier Maxime', 'Verret Farah', 'Couture Beatrice'] | 2022-12-13 | null | null | null | null | ['handwriting-recognition'] | ['computer-vision'] | [ 8.81064832e-02 -6.04729168e-02 4.69099171e-02 -2.97032654e-01
-7.79006481e-01 -8.55535567e-01 6.24288261e-01 -1.60721093e-01
-5.69280267e-01 3.70705068e-01 5.64587831e-01 -5.86327493e-01
-3.32736969e-03 -6.00650072e-01 -4.87791061e-01 -2.93287188e-01
5.59535921e-01 5.99282146e-01 -9.96867567e-02 -3.99430990e-01
7.74887264e-01 8.08801711e-01 -1.39126527e+00 5.04331231e-01
5.44172823e-01 1.81062281e-01 2.35349447e-01 9.18200850e-01
-2.50603765e-01 8.33931267e-01 -7.47656226e-01 -5.23180783e-01
1.57774299e-01 -1.55934945e-01 -9.37739611e-01 -1.07814997e-01
3.91206205e-01 -7.41197944e-01 -2.41858184e-01 3.42841417e-01
6.70528114e-01 -5.04867174e-02 2.52140403e-01 -4.49924618e-01
-8.41941059e-01 9.55400527e-01 -1.11352652e-01 9.03603733e-02
5.88401020e-01 2.04121545e-01 5.49925864e-01 -8.24254692e-01
9.16142523e-01 8.42468739e-01 9.06904578e-01 4.12364244e-01
-8.67518902e-01 -5.60601473e-01 -2.22620741e-01 -1.00134924e-01
-1.27879477e+00 -8.41528177e-01 2.56374657e-01 -5.16448438e-01
1.29395664e+00 3.39582831e-01 8.43958378e-01 1.14880228e+00
-1.03902556e-01 5.42849362e-01 1.16408360e+00 -9.19987023e-01
-2.12538108e-01 4.67632972e-02 2.05917209e-01 5.37905812e-01
1.15532629e-01 -2.43323132e-01 -8.50582421e-01 3.02111059e-02
8.33503723e-01 -5.16532183e-01 -1.69999957e-01 6.34918034e-01
-1.33259058e+00 2.88268685e-01 -9.54213738e-02 8.73500228e-01
2.44624615e-02 2.65595559e-02 3.60898435e-01 2.11351007e-01
2.43729070e-01 7.94160783e-01 -2.31033131e-01 -1.04209149e+00
-1.22280192e+00 -9.93467495e-02 8.72970402e-01 8.66303325e-01
5.03034234e-01 -1.67848587e-01 7.95966685e-02 1.13736761e+00
5.47701061e-01 4.49748725e-01 4.98675853e-01 -7.34744251e-01
4.75193113e-01 5.94582498e-01 1.58954766e-02 -8.14260721e-01
-3.20542566e-02 -7.23395590e-03 -1.13448620e-01 -2.95704380e-02
7.77849674e-01 -1.25673011e-01 -1.21795011e+00 5.50626159e-01
-1.50741756e-01 -3.05755705e-01 -2.00605884e-01 9.04218376e-01
5.98238766e-01 5.76052964e-01 -9.64967310e-02 2.93440580e-01
8.04593027e-01 -5.82072854e-01 -5.12017608e-01 -9.55025479e-02
8.92066658e-01 -1.30370057e+00 1.17587674e+00 5.94507456e-01
-8.04302514e-01 -2.94216543e-01 -1.15308750e+00 -3.74600440e-01
-2.55507648e-01 4.69735265e-01 5.82584381e-01 1.19506836e+00
-1.03063512e+00 6.45459890e-01 -1.06645298e+00 -6.53933406e-01
1.69342577e-01 3.15517187e-01 -5.37769616e-01 -3.36807489e-01
-5.48905432e-01 1.17110324e+00 1.04145981e-01 6.15682602e-01
-6.46506071e-01 -6.12211704e-01 -1.93935737e-01 -4.91549194e-01
2.60858219e-02 1.14657000e-01 1.07108176e+00 -6.73683167e-01
-1.59305084e+00 9.58097398e-01 -2.40288880e-02 8.75837505e-02
1.00854659e+00 -3.39615524e-01 -3.53099912e-01 -1.17091045e-01
-1.91534951e-01 1.54165044e-01 2.35769883e-01 -9.47966456e-01
-5.71295857e-01 -3.71266246e-01 -4.18126971e-01 -2.35341694e-02
-3.92035246e-01 4.33661729e-01 -4.85115141e-01 -4.74252343e-01
9.10819992e-02 -1.04040682e+00 3.41685265e-01 -3.23691964e-01
-6.14930084e-03 2.06100658e-01 7.17721879e-01 -1.47215378e+00
9.95863974e-01 -2.15644526e+00 -9.94649678e-02 2.99681485e-01
-2.44430587e-01 4.57103997e-01 -1.41106471e-01 9.76813912e-01
5.40175080e-01 5.40512383e-01 6.10985085e-02 -1.23748712e-01
-8.92653689e-02 3.65773022e-01 -4.48401958e-01 2.86918133e-01
-1.22645684e-01 6.30203068e-01 -6.76640987e-01 -1.26654282e-01
5.20361997e-02 7.55365193e-01 3.90771925e-02 -9.77023542e-02
9.53578353e-02 3.27155530e-01 -1.67132422e-01 8.96338284e-01
3.88529420e-01 2.56166577e-01 5.65375090e-01 2.30565414e-01
-7.45955050e-01 3.49161178e-01 -1.01015902e+00 1.32019746e+00
-5.70559382e-01 1.12552023e+00 -4.67526615e-02 -8.79608914e-02
1.41141999e+00 2.40991056e-01 1.19315185e-01 -4.43975687e-01
-2.52246261e-02 8.74541163e-01 8.07952508e-02 -6.19270504e-01
9.90942538e-01 -1.64580755e-02 4.24293786e-01 7.38974214e-01
1.09633300e-02 -7.05167130e-02 9.23016965e-02 -7.90495425e-02
9.16610479e-01 4.17792231e-01 -4.35584784e-01 -1.51458859e-01
1.13567874e-01 2.52144426e-01 9.10209119e-02 7.62407601e-01
1.23015873e-01 7.26121962e-01 1.60037652e-02 -4.36067045e-01
-1.47016001e+00 -3.37554067e-01 -4.16420847e-02 1.03959727e+00
-5.26664793e-01 -3.41425806e-01 -4.72967923e-01 -2.83142209e-01
-4.98818643e-02 6.45822823e-01 -5.05699635e-01 1.88932002e-01
-6.07358515e-01 -5.93360782e-01 1.31794703e+00 5.72526395e-01
6.87158346e-01 -7.43479490e-01 -5.69654107e-01 3.24009776e-01
-2.37497211e-01 -9.36388016e-01 -1.23787783e-01 -1.10641785e-01
-8.13855588e-01 -6.52268350e-01 -1.08672297e+00 -5.37412405e-01
5.36296487e-01 -4.79326537e-03 4.37733829e-01 4.16212618e-01
-3.62528235e-01 5.09360552e-01 -8.53213668e-01 -2.45591894e-01
-7.62285590e-01 5.16006291e-01 -2.97125578e-01 -2.39833072e-01
3.39096308e-01 -1.78853422e-01 1.23626582e-01 3.17854494e-01
-8.47912848e-01 1.35758415e-01 6.48604572e-01 5.11263967e-01
-5.82527183e-03 -4.34530705e-01 3.39850783e-01 -6.83157206e-01
7.62832344e-01 7.82052502e-02 -2.45952412e-01 9.30921853e-01
-5.99442601e-01 -2.60449857e-01 1.36069283e-01 -4.50595498e-01
-1.17080426e+00 -6.41942248e-02 -1.02807149e-01 1.82476580e-01
-3.93211134e-02 7.66250432e-01 3.47908020e-01 -4.84797210e-01
6.72591984e-01 3.20522875e-01 2.24517435e-02 -5.84330857e-01
8.13706964e-02 1.38319540e+00 3.40546876e-01 -8.44300091e-01
2.46174887e-01 2.27326512e-01 -4.83421832e-01 -1.44078290e+00
1.09190539e-01 -2.01045603e-01 -9.16010141e-01 -7.41112947e-01
3.42624843e-01 -6.76422119e-01 -5.23334444e-01 1.01647425e+00
-8.35553348e-01 -8.86673748e-01 1.29874110e-01 5.72077096e-01
-1.49687812e-01 4.28123087e-01 -7.38720119e-01 -1.07340741e+00
-1.71525687e-01 -9.93821084e-01 1.01785505e+00 2.58152694e-01
-3.38858634e-01 -7.09364295e-01 1.99665964e-01 9.03104186e-01
5.63441336e-01 9.24090156e-04 7.00040460e-01 -3.24559718e-01
-4.54405516e-01 -4.15642291e-01 -9.48257670e-02 1.90946773e-01
1.37540698e-01 7.95766532e-01 -9.11511779e-01 -1.83211848e-01
-3.95705491e-01 -1.91803917e-01 4.47275698e-01 -3.31997424e-01
4.04714882e-01 -6.68316558e-02 -7.63165504e-02 2.25853533e-01
1.32798886e+00 3.41573715e-01 1.06928563e+00 7.23852873e-01
7.12301850e-01 5.13811052e-01 3.26822370e-01 1.94659308e-01
3.37768793e-01 4.47382122e-01 -3.72631609e-01 3.65848780e-01
-1.94534764e-01 -3.75620186e-01 4.91838366e-01 1.15336239e+00
-7.69514382e-01 4.17853966e-02 -1.81504214e+00 5.59606493e-01
-1.60309267e+00 -7.01928556e-01 -4.25907969e-01 2.09804845e+00
8.33914101e-01 -2.51485676e-01 3.64583619e-02 1.35799736e-01
5.07411659e-01 -2.73449391e-01 -3.97298224e-02 -6.74925923e-01
-2.32071579e-01 1.33477062e-01 6.82586908e-01 5.66849649e-01
-4.49773431e-01 1.15055585e+00 7.43061876e+00 2.82998025e-01
-1.52253199e+00 -2.98304379e-01 3.80263597e-01 -1.02506997e-03
-2.95136005e-01 1.97426111e-01 -9.60580349e-01 3.36722404e-01
1.05493164e+00 -6.63872734e-02 7.84914494e-01 3.70746851e-01
2.47089058e-01 -3.96748275e-01 -7.44520545e-01 7.14501023e-01
2.94404477e-01 -1.47191072e+00 3.58325280e-02 2.18613923e-01
4.90737379e-01 2.06581309e-01 -2.49902025e-01 -8.08559582e-02
4.16764677e-01 -1.24521875e+00 8.63650382e-01 9.94982123e-01
8.25902402e-01 -4.69496191e-01 6.81933463e-01 1.85606420e-01
-5.68950415e-01 7.60515705e-02 -2.47672230e-01 -2.31621176e-01
4.49690260e-02 2.71780550e-01 -1.41930842e+00 5.17081499e-01
6.29141748e-01 6.30519509e-01 -8.63762200e-01 9.92191374e-01
-8.48863125e-02 9.36947763e-01 -5.82741201e-01 -2.69381315e-01
1.19079553e-01 -1.94429621e-01 6.53369352e-02 1.39575720e+00
5.31250596e-01 1.44429535e-01 -3.68975997e-01 3.63981694e-01
8.68559480e-02 1.50409579e-01 -4.05282378e-01 -6.46065831e-01
5.19413769e-01 9.55523133e-01 -1.02797925e+00 3.24972831e-02
-1.45862356e-01 9.64752972e-01 1.34591982e-01 1.00917749e-01
-2.68378794e-01 -4.81384397e-01 5.66955432e-02 3.72729957e-01
7.41398633e-02 -8.88819337e-01 -5.52343667e-01 -1.02391303e+00
1.27531961e-01 -1.31202173e+00 -1.11119688e-01 -9.31445181e-01
-8.33675861e-01 2.66130596e-01 -3.71736616e-01 -4.70957160e-01
5.74446358e-02 -6.91116095e-01 -1.45146310e-01 1.01582754e+00
-8.73178720e-01 -1.59102905e+00 -3.24084669e-01 4.74283509e-02
2.23739892e-01 -1.48181304e-01 8.41035604e-01 3.49446565e-01
-4.64223951e-01 4.86044616e-01 3.96292418e-01 4.67187047e-01
7.90553868e-01 -7.24017501e-01 4.75749195e-01 7.51469731e-01
3.02485466e-01 9.64520335e-01 3.49722564e-01 -1.06178617e+00
-1.75286317e+00 -3.19149524e-01 1.03539348e+00 -7.25641310e-01
6.03876472e-01 -3.59822810e-01 -6.94016099e-01 9.79496598e-01
-3.70699391e-02 -8.35381031e-01 8.44396889e-01 2.72060543e-01
-2.33637810e-01 1.06880121e-01 -9.71851826e-01 7.41194367e-01
8.93497467e-01 -9.11281347e-01 -5.16052306e-01 8.97955224e-02
-1.40465915e-01 -4.51228499e-01 -1.18963778e+00 2.68173702e-02
1.26530492e+00 -6.58660054e-01 2.44768977e-01 -2.99154133e-01
6.35442913e-01 -2.95195013e-01 -1.68059498e-01 -9.86945331e-01
-3.59112248e-02 -7.50777125e-01 4.70180899e-01 1.68764639e+00
6.30671978e-01 -7.07128704e-01 7.58022010e-01 1.19139624e+00
-6.07409589e-02 -2.73760706e-01 -9.29543376e-01 -5.22835970e-01
1.71960711e-01 -5.65826952e-01 5.35406709e-01 1.04749537e+00
2.42918894e-01 -1.86262891e-01 -3.43554527e-01 -1.01913579e-01
8.91824812e-02 -4.56472695e-01 9.98559594e-01 -8.73078942e-01
-1.15547821e-01 -2.61164933e-01 -5.64527094e-01 -7.45447993e-01
-4.60659564e-01 -8.47849786e-01 -2.55776793e-02 -1.68768275e+00
-4.55931798e-02 -5.80070555e-01 5.05341232e-01 8.18783820e-01
3.74608755e-01 3.46375495e-01 6.84096396e-01 7.23834097e-01
2.09826216e-01 -1.01739027e-01 9.62554932e-01 5.74535318e-02
-2.24826261e-01 -4.97263998e-01 -6.38664842e-01 2.27466017e-01
7.36794889e-01 -1.35957941e-01 5.65372668e-02 -1.01904261e+00
5.36628127e-01 -3.30836177e-01 7.44844601e-02 -1.02087927e+00
5.60591459e-01 -9.59743280e-03 5.09595752e-01 -2.38917261e-01
8.28931257e-02 -5.87005317e-01 6.66860342e-01 2.07495645e-01
-4.10447389e-01 -6.15012459e-03 4.36592877e-01 -1.03683211e-01
5.88984601e-02 -5.16414523e-01 2.69592732e-01 -2.44035020e-01
-5.62346399e-01 -6.01827681e-01 -8.68086874e-01 -5.12698412e-01
7.92769790e-01 -5.90706885e-01 -6.25582635e-01 -2.45082989e-01
-6.85213625e-01 5.26800146e-03 7.42297232e-01 4.83366162e-01
3.62593889e-01 -6.55151367e-01 -6.59287572e-01 1.50630251e-01
-1.02735884e-01 -3.80824447e-01 1.09392121e-01 6.46787405e-01
-1.40208590e+00 4.27567750e-01 -6.13156676e-01 -1.99826121e-01
-1.38663304e+00 -3.37021858e-01 3.97090912e-01 -6.00552745e-02
-6.78672016e-01 5.47768652e-01 -1.15872598e+00 -4.65329587e-01
3.64858061e-02 -1.79962032e-02 -4.13355790e-02 1.59969822e-01
5.52715123e-01 8.35592210e-01 4.25424576e-01 -5.85345089e-01
-2.86534846e-01 5.04142284e-01 -3.58744442e-01 -8.51903617e-01
1.65468931e+00 1.90864146e-01 -3.31841320e-01 7.18782723e-01
6.90065682e-01 3.66336972e-01 -9.65573132e-01 4.98019874e-01
2.10207865e-01 -7.02364802e-01 -4.49762866e-02 -1.38237977e+00
-6.03964031e-01 6.09460533e-01 4.28670317e-01 -3.16221491e-02
6.45476103e-01 -3.46503019e-01 4.96045828e-01 6.61312461e-01
5.37211597e-01 -1.52124202e+00 -3.87316674e-01 7.37247288e-01
9.92567301e-01 -5.94425082e-01 1.30657539e-01 -1.38218001e-01
-5.82477927e-01 1.80728519e+00 1.60639808e-01 1.70738980e-01
3.87864202e-01 5.00669539e-01 6.05507612e-01 5.93525171e-03
-3.12004447e-01 2.09923446e-01 -1.81795076e-01 5.93019485e-01
8.49244058e-01 8.39301646e-02 -7.14128137e-01 1.24073774e-01
-3.77621114e-01 5.73042750e-01 1.01713085e+00 1.22334445e+00
-1.97169080e-01 -1.55806088e+00 -6.44530475e-01 4.92619246e-01
-4.45377946e-01 1.06445663e-01 -1.03591931e+00 6.87978029e-01
-1.47787482e-01 9.40205991e-01 -2.12366834e-01 -6.30210161e-01
2.57933378e-01 4.01591480e-01 5.51824927e-01 -4.38229680e-01
-1.05670094e+00 -1.73526421e-01 5.43464184e-01 9.51144695e-02
-3.44866395e-01 -8.85508955e-01 -8.60189378e-01 -7.08535433e-01
-3.31672639e-01 -4.38372232e-02 1.20803821e+00 9.29459631e-01
3.73979479e-01 2.11295933e-02 -3.39883030e-03 -7.83726335e-01
-3.98823351e-01 -1.04045963e+00 -4.40464109e-01 -2.38224417e-01
-1.39163077e-01 -2.00874154e-02 -3.04096080e-02 2.86396891e-01] | [11.838233947753906, 2.5776703357696533] |
6ba718b7-df97-4427-b461-bb371b436660 | from-unsupervised-to-few-shot-graph-anomaly | 2202.05525 | null | https://arxiv.org/abs/2202.05525v1 | https://arxiv.org/pdf/2202.05525v1.pdf | From Unsupervised to Few-shot Graph Anomaly Detection: A Multi-scale Contrastive Learning Approach | Anomaly detection from graph data is an important data mining task in many applications such as social networks, finance, and e-commerce. Existing efforts in graph anomaly detection typically only consider the information in a single scale (view), thus inevitably limiting their capability in capturing anomalous patterns in complex graph data. To address this limitation, we propose a novel framework, graph ANomaly dEtection framework with Multi-scale cONtrastive lEarning (ANEMONE in short). By using a graph neural network as a backbone to encode the information from multiple graph scales (views), we learn better representation for nodes in a graph. In maximizing the agreements between instances at both the patch and context levels concurrently, we estimate the anomaly score of each node with a statistical anomaly estimator according to the degree of agreement from multiple perspectives. To further exploit a handful of ground-truth anomalies (few-shot anomalies) that may be collected in real-life applications, we further propose an extended algorithm, ANEMONE-FS, to integrate valuable information in our method. We conduct extensive experiments under purely unsupervised settings and few-shot anomaly detection settings, and we demonstrate that the proposed method ANEMONE and its variant ANEMONE-FS consistently outperform state-of-the-art algorithms on six benchmark datasets. | ['Yi-Ping Phoebe Chen', 'Shirui Pan', 'Khoa T. Phan', 'Lianhua Chi', 'Yixin Liu', 'Ming Jin', 'Yu Zheng'] | 2022-02-11 | null | null | null | null | ['graph-anomaly-detection'] | ['graphs'] | [ 1.14858367e-01 -5.43899238e-02 -2.94846799e-02 -3.93089920e-01
-2.39020914e-01 -3.03240627e-01 3.86457741e-01 7.41466939e-01
9.69241709e-02 1.91370174e-01 -5.76447807e-02 -1.85707286e-01
-2.56625891e-01 -1.10723984e+00 -5.09382367e-01 -5.54554462e-01
-4.32264626e-01 3.02537590e-01 3.87038767e-01 -2.76990950e-01
2.51935840e-01 5.51690757e-01 -1.19842696e+00 4.90395054e-02
1.16765881e+00 1.05944896e+00 -7.02526987e-01 4.17721152e-01
-3.49866271e-01 7.82553792e-01 -4.98751581e-01 -5.64000607e-01
4.50032920e-01 -3.58932704e-01 -5.11232615e-01 4.71410304e-01
5.65538883e-01 -1.77003026e-01 -4.30038422e-01 1.43767834e+00
9.12665501e-02 1.94603652e-01 5.63068569e-01 -1.70948958e+00
-4.78426158e-01 3.29121947e-01 -1.09457934e+00 7.09138870e-01
2.03389958e-01 1.98317945e-01 1.37166059e+00 -6.84373081e-01
3.68785709e-01 1.01917410e+00 6.37615025e-01 1.47935361e-01
-1.19744134e+00 -4.09147263e-01 7.91330040e-01 2.41308391e-01
-1.08964968e+00 -1.55744419e-01 1.21544504e+00 -3.35586965e-01
7.29880512e-01 2.82562196e-01 6.19500577e-01 9.09371436e-01
2.51242250e-01 7.19174027e-01 6.23341858e-01 -5.06219082e-02
3.47368896e-01 -4.23862725e-01 3.34816217e-01 9.79986250e-01
6.69867039e-01 -1.88279897e-01 -5.30411899e-01 -5.95050275e-01
4.72857893e-01 6.23249114e-01 7.76763633e-02 -6.37805164e-01
-8.26121628e-01 9.48339522e-01 6.00937426e-01 1.66488722e-01
-6.00852191e-01 -1.48237228e-01 7.16872752e-01 8.50024283e-01
8.54249597e-01 3.88534069e-01 -2.35932022e-01 1.76672846e-01
-4.83225852e-01 1.07423685e-01 7.55775094e-01 6.84180021e-01
8.05148363e-01 3.67485613e-01 3.38633992e-02 8.02128911e-01
3.71432245e-01 -1.84791838e-03 2.61559933e-01 -3.45162004e-01
4.15181071e-01 1.25599313e+00 -4.94411141e-01 -1.41389477e+00
-3.28308612e-01 -6.29082263e-01 -1.20524693e+00 -9.34888236e-03
2.90429682e-01 6.74181953e-02 -9.41834688e-01 1.46179295e+00
6.33871257e-01 6.20440304e-01 -2.79933542e-01 5.98300219e-01
5.88058472e-01 2.94906318e-01 -4.98352461e-02 -3.66237402e-01
9.14955854e-01 -8.92024875e-01 -5.67789853e-01 -3.55869412e-01
9.24202859e-01 -2.41480708e-01 9.99394417e-01 3.56014371e-01
-4.32708770e-01 -1.11951254e-01 -1.05626059e+00 6.05882227e-01
-5.02974331e-01 -6.24492228e-01 6.64598227e-01 3.22536260e-01
-6.88110292e-01 8.94352973e-01 -9.30446088e-01 -4.72103804e-01
6.48373604e-01 -2.48603038e-02 -5.16572714e-01 -1.35888964e-01
-9.13252652e-01 3.51692915e-01 2.87691265e-01 -3.90931815e-02
-7.09734619e-01 -4.32209611e-01 -1.08121765e+00 1.29813522e-01
1.05537224e+00 -3.19628894e-01 6.89760447e-01 -8.89743626e-01
-6.89026833e-01 7.40091205e-01 1.09342724e-01 -4.70778316e-01
2.28399739e-01 -1.14943832e-01 -1.08715570e+00 4.23048884e-02
2.39688247e-01 -3.11181456e-01 1.00313103e+00 -1.09481502e+00
-5.49680650e-01 -8.22775900e-01 -1.24489710e-01 -1.90712065e-02
-5.88891625e-01 -1.31284207e-01 -1.86285168e-01 -9.15391386e-01
5.21505773e-01 -6.11317694e-01 -4.20919418e-01 -6.38690516e-02
-6.40898824e-01 -3.26025039e-01 1.15113688e+00 -5.01827776e-01
1.54401565e+00 -2.17735863e+00 -1.94659736e-02 8.16380501e-01
8.98479223e-01 2.08253086e-01 -2.72745997e-01 4.78229910e-01
-1.83709249e-01 6.80793673e-02 -6.28534913e-01 -1.13203906e-01
-1.94281846e-01 5.43286562e-01 -1.75501987e-01 6.06292903e-01
3.93457711e-01 6.86940372e-01 -1.13654113e+00 -3.52880985e-01
2.21905543e-06 -2.51301527e-01 -5.74588060e-01 3.71770829e-01
-1.10300004e-01 3.02858263e-01 -6.55997574e-01 1.20165002e+00
5.34421086e-01 -5.11952460e-01 5.38664795e-02 2.04891309e-01
4.28558171e-01 -3.03114682e-01 -1.21034586e+00 1.54909647e+00
8.75839498e-03 1.78083941e-01 -1.49198487e-01 -1.51660037e+00
1.04647648e+00 3.71533483e-02 7.93055534e-01 -5.93997061e-01
-3.11846614e-01 2.75928944e-01 2.57794261e-01 -3.55251133e-01
2.19027579e-01 2.49176443e-01 -1.57741923e-02 6.38503850e-01
1.98415443e-01 4.84109491e-01 4.07957673e-01 6.67287469e-01
1.85059106e+00 -4.09804761e-01 6.25240922e-01 -2.30722930e-02
6.15510702e-01 -3.31335485e-01 7.06159890e-01 9.69545662e-01
-5.50910890e-01 4.12600070e-01 1.12832379e+00 -9.23269093e-01
-8.40355575e-01 -1.10899925e+00 2.40230992e-01 1.19712901e+00
2.33390518e-02 -8.06656778e-01 -3.01057637e-01 -1.48614395e+00
3.05515975e-01 4.18137819e-01 -5.64659476e-01 -5.10622501e-01
-4.61987913e-01 -8.39349091e-01 2.03075930e-01 3.34979296e-01
2.66270906e-01 -1.20828950e+00 1.37846068e-01 1.98630378e-01
5.62272780e-02 -1.15599072e+00 -4.50531006e-01 -4.41855118e-02
-1.07692850e+00 -1.35675037e+00 9.11696553e-02 -3.22990477e-01
8.52202654e-01 3.57596666e-01 1.38968718e+00 4.85346258e-01
-4.40906614e-01 4.40203905e-01 -5.21966338e-01 -2.79021263e-01
-3.87471586e-01 3.20938043e-02 2.07446814e-01 5.39026439e-01
5.95404625e-01 -1.05168676e+00 -5.57529032e-01 1.57607585e-01
-1.06304264e+00 -6.46670699e-01 6.14433169e-01 7.92931736e-01
6.29751980e-01 3.28097492e-02 8.71016979e-01 -1.35827816e+00
8.36089730e-01 -1.06311202e+00 -5.11622667e-01 1.45952031e-01
-1.11637032e+00 1.33569790e-02 7.71017075e-01 -8.81262347e-02
-4.35276449e-01 -3.90460193e-01 3.50871533e-02 -7.24341393e-01
-2.50769049e-01 7.37578154e-01 -5.46547733e-02 -9.89526361e-02
7.39577770e-01 2.11481169e-01 1.41172558e-01 -3.36065382e-01
1.66449159e-01 1.69251248e-01 5.23909390e-01 -4.36420441e-01
9.96840417e-01 3.06224197e-01 3.29126805e-01 -7.68626809e-01
-8.74387920e-01 -8.54180098e-01 -6.19717360e-01 -2.75355786e-01
2.87570626e-01 -6.96679473e-01 -2.65328258e-01 4.57672536e-01
-6.23472512e-01 2.68010408e-01 -3.52774888e-01 -8.15357640e-02
-6.72139227e-02 7.87582695e-01 -3.52843165e-01 -7.09171414e-01
-4.12614286e-01 -6.14740372e-01 9.43450570e-01 -2.65056472e-02
-1.15044750e-02 -1.16241169e+00 3.11073244e-01 -6.46893531e-02
2.35844135e-01 7.92952240e-01 9.00621235e-01 -1.50349581e+00
-4.66530174e-01 -6.11334026e-01 -2.75707752e-01 2.13215008e-01
4.62479293e-01 -4.14485447e-02 -7.06795692e-01 -5.63122392e-01
-1.88437879e-01 -1.01194561e-01 9.14515555e-01 2.40532774e-02
1.63328314e+00 -5.01920223e-01 -2.72042602e-01 5.15733004e-01
1.41508043e+00 -2.32199609e-01 3.95023316e-01 1.35096014e-01
1.25109053e+00 3.91691267e-01 4.76800472e-01 6.48826838e-01
1.93471074e-01 3.47605556e-01 8.69363368e-01 6.53573021e-04
4.27437276e-01 -2.76010543e-01 3.06279629e-01 9.11537766e-01
-1.69309005e-01 -2.37233520e-01 -9.60963488e-01 5.54129541e-01
-2.23073483e+00 -9.18554068e-01 -2.00737379e-02 2.13147068e+00
6.85360357e-02 4.60151076e-01 5.06994247e-01 1.03303872e-01
8.63940179e-01 6.45002186e-01 -8.25789750e-01 -3.51539224e-01
1.21363126e-01 -5.72126284e-02 1.31091729e-01 4.85091358e-02
-1.14494169e+00 6.27630770e-01 5.51792288e+00 6.14641726e-01
-8.16070080e-01 -1.56346723e-01 6.20761693e-01 1.21063441e-01
-3.40529203e-01 1.06336765e-01 -7.34531209e-02 4.65969622e-01
8.69819343e-01 -2.29782730e-01 4.54625100e-01 1.02621067e+00
-1.66435331e-01 2.74108678e-01 -1.11449981e+00 7.92503297e-01
1.38967991e-01 -1.02536583e+00 2.72288144e-01 3.08201760e-01
6.62900686e-01 2.06958905e-01 -1.76531270e-01 3.84602249e-01
2.67441869e-01 -7.75954664e-01 -1.52070582e-01 4.82994050e-01
3.54241878e-01 -7.39858985e-01 7.39373446e-01 2.87029266e-01
-1.40642011e+00 -2.35923290e-01 -1.95751294e-01 2.99468841e-02
-1.16291329e-01 9.73365605e-01 -6.69146895e-01 9.21867967e-01
6.98804677e-01 1.20750165e+00 -9.40792680e-01 8.69883776e-01
2.17671432e-02 7.20234513e-01 -2.48344392e-01 3.58216971e-01
2.93321699e-01 -5.18984914e-01 1.07385969e+00 7.42652655e-01
2.72554696e-01 -8.45328346e-02 5.67121744e-01 6.75994635e-01
-2.86401540e-01 2.72556126e-01 -1.13610411e+00 -2.24737942e-01
4.24481332e-01 1.54196692e+00 -8.64893973e-01 -1.95238009e-01
-8.29819500e-01 8.79155159e-01 6.02833331e-01 2.45223701e-01
-4.72896636e-01 -4.25010562e-01 7.00507998e-01 1.30212232e-01
4.51589860e-02 4.78662811e-02 3.03565077e-02 -1.51406276e+00
3.58162075e-01 -1.09429049e+00 1.14958107e+00 -7.97304288e-02
-2.00176501e+00 5.34465611e-01 -2.39413798e-01 -1.55397069e+00
-3.05074722e-01 -6.25777781e-01 -1.31117344e+00 2.65083283e-01
-1.20389962e+00 -1.12621808e+00 -5.59442878e-01 7.88764238e-01
4.26923215e-01 -6.36139512e-01 7.24111140e-01 1.86544746e-01
-9.07569885e-01 6.78036928e-01 -1.07886247e-01 3.83316994e-01
6.83673203e-01 -1.54838276e+00 8.61772835e-01 1.34633565e+00
5.86236835e-01 3.08042884e-01 5.29895842e-01 -8.58436644e-01
-1.18267763e+00 -1.25417697e+00 1.74105123e-01 -3.32410723e-01
9.96343851e-01 -1.71486020e-01 -1.47960961e+00 8.19043517e-01
-2.62883246e-01 9.02204454e-01 6.77657306e-01 5.46742380e-01
-4.86001641e-01 -3.14174086e-01 -1.24983466e+00 5.38558662e-01
1.40841281e+00 -4.72212434e-01 -2.60513693e-01 2.38893300e-01
7.03006923e-01 -1.52960733e-01 -8.50967050e-01 6.26819909e-01
1.04716286e-01 -1.17487741e+00 6.91176236e-01 -1.09318602e+00
2.95477390e-01 -1.58566639e-01 -9.69270915e-02 -1.62564611e+00
-2.80089468e-01 -5.36950946e-01 -9.17656898e-01 1.15130544e+00
1.73405215e-01 -1.08927715e+00 8.21134090e-01 2.69139498e-01
-1.91569671e-01 -8.51009011e-01 -8.22124422e-01 -6.47681117e-01
-4.91526455e-01 -3.05058241e-01 7.85301030e-01 1.27481759e+00
-8.13721865e-02 1.06553428e-01 -4.14025515e-01 4.80015755e-01
9.27045524e-01 1.45343170e-01 8.88820469e-01 -1.84352291e+00
-1.45016149e-01 -3.26432407e-01 -1.00637901e+00 -3.44150007e-01
1.26672596e-01 -1.04140794e+00 -3.88152480e-01 -1.12247574e+00
1.59714729e-01 -1.98357031e-01 -8.67781222e-01 3.82864684e-01
-5.50097406e-01 3.29249427e-02 -3.45208615e-01 2.43612185e-01
-1.00103664e+00 5.73636651e-01 8.92311096e-01 -1.17850877e-01
-4.95336130e-02 1.79893762e-01 -8.11604559e-01 9.99382854e-01
6.34023130e-01 -4.82540131e-01 -3.52437556e-01 1.37367368e-01
2.90842026e-01 -1.15816258e-01 3.03589374e-01 -9.77796018e-01
1.71861082e-01 -1.35891929e-01 3.31406713e-01 -3.97261947e-01
-3.55030924e-01 -8.34184170e-01 -2.46217728e-01 3.31419468e-01
-1.81408212e-01 4.75279868e-01 -1.31242767e-01 1.29124570e+00
-4.63735431e-01 8.12952667e-02 5.28200924e-01 -1.10912710e-01
-9.92849171e-01 1.14266145e+00 1.59956247e-01 3.34104121e-01
9.70484197e-01 5.03852777e-02 -4.81546015e-01 -4.31532651e-01
-7.66448796e-01 5.15010774e-01 3.65682930e-01 5.70057094e-01
8.06902468e-01 -1.69924641e+00 -6.57951772e-01 5.95543623e-01
8.19288135e-01 1.20268531e-01 3.61429363e-01 8.16326618e-01
-3.05141300e-01 -3.79107505e-01 -2.89436281e-01 -7.41992652e-01
-1.03513062e+00 6.19198918e-01 2.31038451e-01 -6.94794059e-01
-8.97841871e-01 3.71920288e-01 8.13734829e-02 -5.08783937e-01
1.92728601e-02 1.15336038e-01 -2.53750175e-01 -6.78017884e-02
3.14556956e-01 4.63299036e-01 1.90118462e-01 -3.90208572e-01
-2.59193778e-01 1.74328968e-01 -4.74275172e-01 5.06159425e-01
1.32196712e+00 -7.61538222e-02 -3.94169360e-01 5.58873713e-01
8.13431323e-01 5.68910316e-02 -9.38918233e-01 -7.17677414e-01
3.89669150e-01 -7.95803070e-01 -2.77744144e-01 -2.47961670e-01
-1.31430089e+00 4.74859148e-01 3.93592834e-01 8.40079725e-01
1.19910550e+00 1.60534345e-02 7.58219719e-01 5.89783609e-01
5.87281920e-02 -1.06657314e+00 5.78440726e-01 2.23408729e-01
7.27053761e-01 -1.71825612e+00 1.22567907e-01 -4.08210486e-01
-7.00816631e-01 9.99556720e-01 1.16501069e+00 -3.73558819e-01
7.19214439e-01 -2.41322890e-01 -1.29182324e-01 -7.86338031e-01
-6.97646677e-01 -1.00482635e-01 5.15643656e-01 5.02968848e-01
4.66706865e-02 7.60459155e-02 6.68275952e-02 3.73177707e-01
3.40068489e-01 -7.05258489e-01 5.34120381e-01 8.73389959e-01
-5.05336404e-01 -8.76600087e-01 -3.13778631e-02 1.25323606e+00
-5.55710733e-01 1.58941627e-01 -5.07978439e-01 5.88436246e-01
-3.10608864e-01 7.37174451e-01 2.13404536e-01 -4.69024420e-01
4.05326724e-01 1.34735316e-01 -6.46337867e-02 -6.49229646e-01
-2.03099474e-01 -1.90961152e-01 2.46241456e-03 -9.37819541e-01
-1.08583309e-01 -4.20825958e-01 -8.69407892e-01 -4.28716838e-01
-2.78968960e-01 1.93597283e-02 7.18102157e-02 1.06821907e+00
5.55861652e-01 7.82056987e-01 1.05098808e+00 -2.61654794e-01
-5.85959733e-01 -9.82875347e-01 -1.03812706e+00 8.69815946e-01
4.82649088e-01 -6.27914131e-01 -6.63303018e-01 -6.24578714e-01] | [6.618313312530518, 5.759117603302002] |
2c79fd3b-87ec-4f55-9912-fd543194e776 | image-provenance-analysis-at-scale | 1801.06510 | null | http://arxiv.org/abs/1801.06510v2 | http://arxiv.org/pdf/1801.06510v2.pdf | Image Provenance Analysis at Scale | Prior art has shown it is possible to estimate, through image processing and
computer vision techniques, the types and parameters of transformations that
have been applied to the content of individual images to obtain new images.
Given a large corpus of images and a query image, an interesting further step
is to retrieve the set of original images whose content is present in the query
image, as well as the detailed sequences of transformations that yield the
query image given the original images. This is a problem that recently has
received the name of image provenance analysis. In these times of public media
manipulation ( e.g., fake news and meme sharing), obtaining the history of
image transformations is relevant for fact checking and authorship
verification, among many other applications. This article presents an
end-to-end processing pipeline for image provenance analysis, which works at
real-world scale. It employs a cutting-edge image filtering solution that is
custom-tailored for the problem at hand, as well as novel techniques for
obtaining the provenance graph that expresses how the images, as nodes, are
ancestrally connected. A comprehensive set of experiments for each stage of the
pipeline is provided, comparing the proposed solution with state-of-the-art
results, employing previously published datasets. In addition, this work
introduces a new dataset of real-world provenance cases from the social media
site Reddit, along with baseline results. | ['Michael Parowski', 'Walter J. Scheirer', 'Kevin W. Bowyer', 'Joel Brogan', 'Daniel Moreira', 'Anderson Rocha', 'Allan Pinto', 'Patrick J. Flynn', 'Aparna Bharati'] | 2018-01-19 | null | null | null | null | ['authorship-verification'] | ['natural-language-processing'] | [ 6.52554214e-01 -5.36229461e-02 1.43359080e-01 -2.08300039e-01
-5.42555928e-01 -8.95952225e-01 9.50085342e-01 6.91469550e-01
-6.00423634e-01 3.33388776e-01 4.49254662e-02 -1.03342846e-01
1.40786514e-01 -7.93492377e-01 -1.00263453e+00 -5.34362853e-01
-1.59022987e-01 2.97011018e-01 5.88268936e-01 -1.01905540e-01
5.58975101e-01 4.53404218e-01 -1.75231123e+00 5.49215436e-01
4.05598313e-01 8.09796154e-01 -1.10147893e-01 8.79252791e-01
-7.36309821e-03 9.09066260e-01 -6.20759130e-01 -1.09092057e+00
2.43266508e-01 -3.59467775e-01 -1.06928062e+00 1.57689214e-01
9.09524620e-01 -3.49533945e-01 -3.36123914e-01 1.19102156e+00
1.17038257e-01 -1.76807344e-01 4.70261127e-01 -1.40710592e+00
-7.10520089e-01 6.56360686e-01 -6.42493844e-01 3.78038049e-01
3.57986450e-01 2.99274921e-01 6.89257264e-01 -6.92039251e-01
1.41610324e+00 9.35491025e-01 5.58386445e-01 4.55025546e-02
-1.00114930e+00 -3.08561057e-01 -4.04017061e-01 4.65402722e-01
-1.29783988e+00 -5.36597490e-01 4.13675278e-01 -7.07327724e-01
5.45251548e-01 2.48536140e-01 6.69842482e-01 7.97143042e-01
1.14171609e-01 5.67939878e-01 1.16625106e+00 -5.25553763e-01
2.38129571e-01 4.39755976e-01 1.43166542e-01 6.85918689e-01
2.78281063e-01 -7.07368478e-02 -8.37522149e-01 -3.37650061e-01
1.83412194e-01 -2.14078635e-01 -3.29025030e-01 -2.91587979e-01
-1.27260470e+00 3.80460531e-01 4.98833954e-01 1.11164689e-01
-5.77566445e-01 3.38332355e-01 4.33995634e-01 1.48412377e-01
5.07904410e-01 1.92878276e-01 -1.03818893e-01 4.92156669e-02
-1.38150883e+00 3.66307795e-01 6.54947579e-01 9.05510962e-01
1.03001821e+00 -4.95099336e-01 -1.79594189e-01 9.17718485e-02
1.58787787e-01 3.25174600e-01 -6.54686540e-02 -8.33357573e-01
3.42299819e-01 7.02666461e-01 1.45738810e-01 -1.30941010e+00
2.78278202e-01 -1.01484850e-01 -5.40449440e-01 3.86427730e-01
6.92805529e-01 5.27621090e-01 -8.81124318e-01 1.28624284e+00
7.79657483e-01 4.13231820e-01 -9.18770358e-02 8.29440534e-01
5.13963342e-01 4.64259595e-01 -4.89192791e-02 -2.70951260e-02
1.80318213e+00 -6.15075648e-01 -5.52569091e-01 -5.18642180e-02
2.49619320e-01 -9.03007925e-01 8.06502521e-01 3.82981449e-01
-9.84911382e-01 -2.70069152e-01 -1.04735851e+00 -2.44614303e-01
-6.23398244e-01 -1.55403614e-01 2.64780492e-01 7.89863229e-01
-1.02282739e+00 7.66325831e-01 -4.60440546e-01 -7.32209325e-01
7.76652277e-01 2.89145811e-03 -4.89276141e-01 -1.44861504e-01
-1.01811182e+00 6.32140875e-01 5.94765484e-01 -6.52410910e-02
-8.98704588e-01 -8.98266435e-01 -4.46914554e-01 -2.85061836e-01
6.61513805e-01 -5.73265016e-01 1.01530957e+00 -1.16454315e+00
-8.44392478e-01 1.46968567e+00 -9.12958458e-02 -6.67320549e-01
8.55571270e-01 -5.34595363e-02 -4.04267907e-01 5.55145264e-01
2.10070267e-01 4.97982651e-01 1.29566503e+00 -1.38296020e+00
-9.86048400e-01 -5.13417900e-01 -4.12419885e-02 -3.30171287e-01
-3.39834750e-01 4.01709408e-01 -9.03792560e-01 -6.05480134e-01
-2.19632477e-01 -1.01134181e+00 2.00954005e-01 4.58729088e-01
-3.23734194e-01 2.19573557e-01 7.67902315e-01 -1.17985284e+00
1.14420283e+00 -2.34375882e+00 -3.32592130e-02 4.29886490e-01
3.59537333e-01 2.56489456e-01 1.60007644e-02 6.59241736e-01
1.51692824e-02 5.42032480e-01 -6.64771318e-01 -3.74218076e-01
-2.89125085e-01 6.99924976e-02 -5.63595772e-01 8.08386445e-01
1.91264421e-01 8.10301781e-01 -1.13953125e+00 -7.37529635e-01
-4.24154140e-02 2.68651187e-01 -1.21206202e-01 1.47557169e-01
-3.86687815e-01 4.53183770e-01 5.68988780e-03 6.27700150e-01
8.29671741e-01 -1.75302878e-01 3.18772509e-03 -3.69734854e-01
-1.51696682e-01 -2.66674221e-01 -1.23324144e+00 1.63173807e+00
1.39397783e-02 8.50793958e-01 -4.88874428e-02 -3.48582655e-01
5.17457724e-01 2.62255907e-01 3.54335964e-01 -5.51130533e-01
1.83130875e-02 1.63501963e-01 -5.60217679e-01 -7.49478042e-01
1.10002673e+00 2.86739051e-01 -3.78408171e-02 8.94717455e-01
-4.99112532e-02 -2.03948412e-02 4.30505544e-01 7.28114307e-01
1.11605549e+00 4.83362049e-01 2.78049648e-01 9.52287465e-02
6.25371039e-01 5.76503575e-01 2.65202433e-01 6.81693316e-01
-1.90309286e-01 7.52528548e-01 5.85517287e-01 -5.91786325e-01
-1.43321896e+00 -7.32975900e-01 1.41100451e-01 6.20953560e-01
3.84871989e-01 -5.79040289e-01 -9.03068900e-01 -5.78805745e-01
-8.64666607e-03 8.15717101e-01 -9.09756541e-01 -7.96887800e-02
-5.76914966e-01 -4.14943814e-01 8.79553854e-01 -2.01912254e-01
5.48775017e-01 -9.66503322e-01 -9.57519889e-01 -4.83082347e-02
-3.62771690e-01 -1.24412858e+00 -3.20066154e-01 -5.85703433e-01
-5.46435118e-01 -1.43297565e+00 -2.91868567e-01 -2.75713533e-01
8.16479027e-01 2.48684287e-01 1.15763748e+00 5.21448970e-01
-6.42178714e-01 4.26059365e-01 -5.10083497e-01 -2.34100074e-01
-8.94490659e-01 -1.99247897e-01 -4.56222534e-01 6.68832898e-01
-2.58937720e-02 -3.65549535e-01 -5.23911119e-01 5.46239540e-02
-1.67421854e+00 -2.85079628e-02 2.88752615e-01 4.95531708e-01
7.16396391e-01 2.59913087e-01 -1.91066816e-01 -1.25502002e+00
3.96361440e-01 -5.70182800e-01 -6.51466310e-01 7.43701220e-01
-6.50687039e-01 2.69840807e-02 3.76013428e-01 -7.42023066e-02
-1.20577383e+00 1.84277967e-01 4.54815984e-01 -4.55953360e-01
-1.22861214e-01 5.46349823e-01 3.48020233e-02 9.48346108e-02
7.95442522e-01 1.66415676e-01 5.72789162e-02 -3.59150499e-01
8.46933305e-01 6.05586290e-01 1.07508981e+00 -2.05836669e-01
9.71707344e-01 1.09004271e+00 2.82215122e-02 -6.01761639e-01
-4.23546046e-01 -4.60607946e-01 -1.02750862e+00 -5.83398640e-01
6.69069886e-01 -6.30688310e-01 -4.36315596e-01 7.81066239e-01
-1.29715168e+00 -1.67410020e-02 -4.19812083e-01 -1.63425714e-01
-3.69880676e-01 6.84118330e-01 -3.06757659e-01 -6.43661559e-01
-4.13191110e-01 -9.99609888e-01 9.25053895e-01 4.47424017e-02
-6.87114373e-02 -6.38503790e-01 1.48403561e-02 5.58540940e-01
2.09660381e-01 5.96021593e-01 8.54283512e-01 -2.97547370e-01
-1.01806760e+00 -3.78713250e-01 -2.45678142e-01 1.50018498e-01
-2.43767500e-01 4.93770242e-01 -9.90345895e-01 -1.44074440e-01
-2.36538485e-01 1.10673465e-01 7.54924834e-01 -1.54741719e-01
5.77666700e-01 -4.55650508e-01 -2.42636755e-01 7.10976362e-01
1.67066026e+00 -3.23329568e-01 1.06627584e+00 6.35671794e-01
6.08625948e-01 8.12074840e-01 7.21423626e-01 5.02823651e-01
3.90151322e-01 6.34693444e-01 7.04374850e-01 2.40187123e-01
-4.50315803e-01 -3.12042743e-01 2.38706663e-01 2.00238287e-01
2.96923928e-02 -2.99815744e-01 -9.01498914e-01 6.93092406e-01
-1.98073006e+00 -1.13681006e+00 -6.99727476e-01 2.37846494e+00
5.39400637e-01 -1.89692929e-01 6.08061180e-02 1.07088629e-02
9.59161103e-01 1.40154347e-01 -5.53079069e-01 -1.29235983e-01
-2.87083864e-01 3.12592424e-02 7.95838237e-01 1.49363235e-01
-1.02782238e+00 9.42986012e-01 5.82214975e+00 7.32422829e-01
-9.65154171e-01 3.90578181e-01 4.46408093e-01 1.35992039e-02
-1.91234365e-01 6.03782833e-01 -4.18905258e-01 5.03657937e-01
9.49316680e-01 -2.13875517e-01 6.11293077e-01 4.57961708e-01
9.87374410e-02 -6.62780643e-01 -9.42912161e-01 7.67213285e-01
3.10379028e-01 -1.49802971e+00 3.15347388e-02 8.37725624e-02
4.83356506e-01 -1.45854250e-01 -1.06150843e-01 -5.42328060e-01
1.92956746e-01 -5.47969699e-01 1.49179840e+00 7.73897290e-01
8.30048144e-01 -4.52425659e-01 6.35383427e-01 9.03465524e-02
-7.91258097e-01 1.24311656e-01 -1.79339349e-01 2.86177903e-01
3.12350601e-01 9.02962983e-01 -1.11868238e+00 8.07620168e-01
1.01984143e+00 8.63399565e-01 -1.31976807e+00 1.03077340e+00
-6.42605007e-01 5.75602055e-01 -2.27768034e-01 3.77901763e-01
-2.77388960e-01 -1.41847417e-01 7.31075525e-01 1.15078938e+00
3.54776680e-01 -1.76737264e-01 -4.80431110e-01 1.02209544e+00
-3.16698372e-01 -7.38388747e-02 -4.14473176e-01 -2.10516602e-01
5.63128233e-01 1.49694908e+00 -1.11560071e+00 -5.55431426e-01
9.06846747e-02 1.40369594e+00 2.17763260e-01 1.22160546e-01
-8.31384480e-01 -1.66181713e-01 4.84748244e-01 4.81630683e-01
3.77755255e-01 -1.83102399e-01 -3.61609198e-02 -1.17272353e+00
3.18947166e-01 -9.22899842e-01 6.31505609e-01 -1.11029756e+00
-9.41954911e-01 5.75470209e-01 1.15703516e-01 -1.11455429e+00
-6.37262538e-02 -8.75234231e-02 -3.90858740e-01 6.29956782e-01
-1.45482361e+00 -1.48280406e+00 -5.11837959e-01 6.50200903e-01
-5.83944470e-02 7.26397783e-02 4.97585386e-01 3.55386138e-01
-4.90758061e-01 3.86535525e-01 1.37706339e-01 1.56172559e-01
9.12539959e-01 -9.19512272e-01 6.68796659e-01 1.72566020e+00
3.54713440e-01 5.14523387e-01 8.31271410e-01 -9.56780076e-01
-1.62619352e+00 -1.13411140e+00 7.76878059e-01 -5.84177554e-01
8.66042912e-01 -3.19281608e-01 -9.06525612e-01 4.87116307e-01
4.84530717e-01 1.25503138e-01 3.04752022e-01 -7.32796192e-01
-6.81355000e-01 -1.38886273e-01 -1.24081373e+00 4.63776022e-01
1.10566437e+00 -7.00890124e-01 -2.95056969e-01 2.46686682e-01
4.99005258e-01 -3.59113336e-01 -8.40823293e-01 -1.34541512e-01
5.23702562e-01 -1.17515528e+00 9.20961678e-01 -3.11273962e-01
6.34725809e-01 -8.43011141e-01 1.09088413e-01 -8.96610022e-01
8.70813280e-02 -8.36134136e-01 -1.01436079e-02 1.68686295e+00
4.18623798e-02 -3.92888457e-01 4.99196947e-01 7.54852474e-01
3.06996346e-01 1.86047852e-02 -8.43290925e-01 -4.38757747e-01
-5.57220280e-01 -4.26202118e-01 8.24364245e-01 8.90322983e-01
-5.47560811e-01 -9.82220694e-02 -4.68190640e-01 3.56650531e-01
9.14539814e-01 1.13439821e-01 1.09876955e+00 -1.00437081e+00
-5.76590523e-02 -2.27225438e-01 -6.95743680e-01 -1.85158163e-01
-2.34802246e-01 -9.27535653e-01 -8.36958811e-02 -1.41824961e+00
2.47678965e-01 -4.73608881e-01 2.30375275e-01 3.82319510e-01
-2.52749503e-01 4.93249893e-01 3.89774621e-01 8.07888329e-01
-4.27605510e-01 -1.36196300e-01 7.34646380e-01 -2.70456165e-01
1.76260561e-01 -3.94387126e-01 -3.85589451e-01 4.04170692e-01
5.67250252e-01 -1.04721415e+00 -4.01011817e-02 -4.10062104e-01
7.86809921e-01 -2.78679997e-01 7.45122671e-01 -9.16387379e-01
6.01606965e-01 -3.54730226e-02 6.36044815e-02 -6.34925127e-01
1.07065491e-01 -8.70559692e-01 9.01551306e-01 4.82370377e-01
-3.16344380e-01 1.92012087e-01 -9.89811346e-02 8.88083041e-01
-1.87842935e-01 -5.26320517e-01 6.17058635e-01 -3.72423410e-01
-1.20499337e+00 2.75597483e-01 -1.98694766e-01 -4.73770499e-02
1.32101238e+00 -5.30568421e-01 -6.41541243e-01 -1.65760592e-01
-3.40711653e-01 -9.94236022e-02 1.15973139e+00 4.28478867e-01
5.68910778e-01 -9.21992898e-01 -9.82237041e-01 6.35218620e-02
4.86086607e-01 -1.99419484e-01 2.84271747e-01 8.00903857e-01
-9.52708840e-01 -4.30843711e-01 -3.73781204e-01 -5.25435328e-01
-1.63806176e+00 8.82313490e-01 4.47030226e-03 -1.79915339e-01
-7.14399278e-01 4.47692513e-01 -5.07732689e-01 1.52677028e-02
-1.98825255e-01 9.13101956e-02 -1.39702568e-02 4.46480006e-01
8.49822581e-01 4.35708463e-01 3.07525188e-01 -1.11062288e+00
-3.83204013e-01 3.60414505e-01 2.23144829e-01 -3.47175747e-01
1.45480478e+00 -3.11241955e-01 -8.21579337e-01 5.46093509e-02
1.20384336e+00 8.47243369e-02 -1.01764119e+00 -3.55712503e-01
2.71761119e-01 -1.03461635e+00 -8.42960738e-03 -7.10628629e-01
-1.34818995e+00 7.53495336e-01 6.03327632e-01 2.40449473e-01
1.13451970e+00 8.78078938e-02 6.64530933e-01 -6.75157085e-02
4.28572088e-01 -9.26742792e-01 -1.16393395e-01 9.76824984e-02
7.31886744e-01 -9.99357820e-01 4.73906755e-01 -5.17523825e-01
-5.13140619e-01 1.17509592e+00 -4.23138402e-02 2.00994536e-01
5.33895910e-01 6.24238513e-02 -5.04057892e-02 -4.51124877e-01
-3.83080184e-01 -7.26590157e-02 5.17036915e-02 5.20091891e-01
-2.37228483e-01 -9.96902138e-02 -2.91424524e-02 -1.33351028e-01
-2.54784734e-03 1.99701741e-01 9.70307469e-01 1.10371351e+00
-1.24112819e-03 -1.18354988e+00 -7.32001185e-01 3.02567393e-01
-5.35374284e-01 -1.06468171e-01 -6.65047467e-01 4.95018780e-01
3.59362870e-01 8.82671416e-01 -1.93551052e-02 -3.20572883e-01
3.40147167e-01 -1.29318282e-01 3.78888696e-01 -3.31050515e-01
-1.10134757e+00 -5.05870223e-01 1.00821979e-01 -5.73322535e-01
-8.97389174e-01 -9.33206022e-01 -1.06337726e+00 -4.52567577e-01
-3.15163553e-01 -2.68389434e-01 1.06679797e+00 5.93664289e-01
6.61952972e-01 8.95571895e-03 3.26424271e-01 -7.22426891e-01
-5.67196449e-03 -4.73021299e-01 -4.07549798e-01 1.12827027e+00
1.00207284e-01 -1.45383984e-01 -2.43729383e-01 9.13301706e-01] | [12.366605758666992, 1.0184673070907593] |
adb7b837-d019-4872-b003-df6920a9de57 | humor-detection-in-english-hindi-code-mixed | 1806.05513 | null | http://arxiv.org/abs/1806.05513v1 | http://arxiv.org/pdf/1806.05513v1.pdf | Humor Detection in English-Hindi Code-Mixed Social Media Content : Corpus and Baseline System | The tremendous amount of user generated data through social networking sites
led to the gaining popularity of automatic text classification in the field of
computational linguistics over the past decade. Within this domain, one problem
that has drawn the attention of many researchers is automatic humor detection
in texts. In depth semantic understanding of the text is required to detect
humor which makes the problem difficult to automate. With increase in the
number of social media users, many multilingual speakers often interchange
between languages while posting on social media which is called code-mixing. It
introduces some challenges in the field of linguistic analysis of social media
content (Barman et al., 2014), like spelling variations and non-grammatical
structures in a sentence. Past researches include detecting puns in texts (Kao
et al., 2016) and humor in one-lines (Mihalcea et al., 2010) in a single
language, but with the tremendous amount of code-mixed data available online,
there is a need to develop techniques which detects humor in code-mixed tweets.
In this paper, we analyze the task of humor detection in texts and describe a
freely available corpus containing English-Hindi code-mixed tweets annotated
with humorous(H) or non-humorous(N) tags. We also tagged the words in the
tweets with Language tags (English/Hindi/Others). Moreover, we describe the
experiments carried out on the corpus and provide a baseline classification
system which distinguishes between humorous and non-humorous texts. | ['Manish Shrivastava', 'Ankush Khandelwal', 'Syed S. Akhtar', 'Sahil Swami'] | 2018-06-14 | humor-detection-in-english-hindi-code-mixed-1 | https://aclanthology.org/L18-1193 | https://aclanthology.org/L18-1193.pdf | lrec-2018-5 | ['humor-detection'] | ['natural-language-processing'] | [-6.21427000e-01 -1.53074130e-01 1.20087698e-01 -7.01092631e-02
-2.10853472e-01 -5.34594655e-01 5.31304479e-01 5.09077191e-01
-1.74269423e-01 5.82100213e-01 6.42759204e-01 -3.21024060e-01
4.73946571e-01 -7.05574453e-01 4.51833084e-02 -2.87115097e-01
1.16659477e-01 -8.61721337e-02 2.06380367e-01 -7.20899343e-01
7.62089670e-01 -9.75676849e-02 -1.39748836e+00 6.08389676e-01
1.06985188e+00 2.73787230e-01 4.33030635e-01 5.91103494e-01
-6.29979789e-01 1.66750264e+00 -4.57589090e-01 -7.55338788e-01
-1.35058552e-01 -7.42557168e-01 -1.07378531e+00 1.02704823e-01
3.38081457e-02 1.89125508e-01 -3.22044045e-01 1.51166081e+00
4.43824589e-01 -7.18174502e-02 4.09691185e-01 -9.94353354e-01
-6.18175447e-01 1.11620545e+00 -6.83247149e-01 2.13091418e-01
7.42600203e-01 -7.74645135e-02 1.06250298e+00 -1.01473141e+00
6.13221407e-01 1.22976100e+00 6.54907346e-01 3.73385638e-01
-8.34392190e-01 -6.91512525e-01 -8.48125279e-01 6.21325254e-01
-1.33122802e+00 -8.05559680e-02 1.03279388e+00 -9.81350720e-01
7.58425057e-01 2.52223492e-01 4.20962900e-01 8.10575962e-01
2.43165091e-01 8.76555204e-01 1.23918152e+00 -7.01152027e-01
-1.99553538e-02 4.69345838e-01 2.78492540e-01 6.68702066e-01
-1.63675517e-01 -7.31958210e-01 -5.37346959e-01 -2.42548957e-01
-4.88793664e-02 -5.08473739e-02 -2.71912545e-01 3.30082834e-01
-6.49251580e-01 1.32161665e+00 2.38245040e-01 8.85538101e-01
-4.46733981e-02 -6.37600303e-01 1.05726838e+00 3.93574655e-01
3.57533991e-01 4.05941159e-01 1.54283777e-01 -3.65208328e-01
-1.23621726e+00 4.27861542e-01 9.69618022e-01 1.06562197e+00
7.00870931e-01 -1.90910295e-01 2.93583065e-01 1.26356375e+00
2.27653086e-01 3.31598788e-01 1.06635356e+00 -3.34428042e-01
5.62330484e-01 1.08531189e+00 -1.17297992e-01 -1.69724917e+00
-6.26080990e-01 -2.29422718e-01 -7.00817347e-01 -1.25896364e-01
3.79273415e-01 -7.50945881e-02 -4.61383909e-02 1.20760572e+00
8.38894844e-02 -5.71209848e-01 -1.60444230e-01 8.64896834e-01
1.14700556e+00 7.56062448e-01 -1.11508854e-01 -3.55381846e-01
1.35902238e+00 -1.06259739e+00 -7.95638621e-01 -1.11841664e-01
8.13399851e-01 -1.37872171e+00 1.41178930e+00 2.87715077e-01
-8.66410315e-01 -2.48456895e-01 -8.30598593e-01 -3.22722822e-01
-4.55619842e-01 -2.34973475e-01 1.38793528e-01 7.17651069e-01
-4.50850815e-01 3.70797604e-01 -1.88560560e-01 -5.96093178e-01
1.75546363e-01 -2.34827638e-01 -1.62981749e-01 1.45924821e-01
-1.36903799e+00 9.37884867e-01 3.88624907e-01 -6.09235168e-01
-3.47371340e-01 -2.22019404e-01 -7.15310693e-01 -2.29919732e-01
7.53216892e-02 2.61301458e-01 1.15211785e+00 -1.54579389e+00
-1.07484150e+00 1.43886495e+00 -1.97939714e-03 -3.69873464e-01
5.08716106e-01 3.59894074e-02 -7.77229548e-01 -1.10952936e-01
3.71561974e-01 -6.13901392e-02 6.87088668e-01 -8.02033067e-01
-5.37572145e-01 -2.42776483e-01 6.24714158e-02 -1.38571128e-01
-8.00483942e-01 8.22389543e-01 2.45017767e-01 -6.24129891e-01
-7.42927343e-02 -9.04922843e-01 3.05069566e-01 -7.75146425e-01
-5.52768350e-01 -4.12760407e-01 7.96918571e-01 -9.68124270e-01
1.97206676e+00 -2.24585891e+00 -2.99875308e-02 -1.84652299e-01
4.74379867e-01 2.00355992e-01 2.08446681e-01 9.13971722e-01
4.17324640e-02 1.83672994e-01 -3.33952665e-01 -1.43482521e-01
-1.11439414e-01 7.75598362e-02 -3.83539319e-01 5.86650193e-01
-3.95783216e-01 6.23907745e-01 -1.10118890e+00 -9.83531415e-01
-1.17782332e-01 -2.45065205e-02 -5.94522059e-01 2.69086182e-01
9.79177505e-02 3.24121296e-01 -1.84034884e-01 5.65053940e-01
5.75719595e-01 -6.68001845e-02 -9.30833817e-03 3.11740458e-01
-7.40529597e-01 6.25021577e-01 -7.39886940e-01 1.10889804e+00
-5.17763615e-01 1.10983074e+00 -4.51923274e-02 -5.42595267e-01
9.69926655e-01 3.19728166e-01 2.54059315e-01 -4.51102316e-01
5.16525805e-01 4.74763960e-01 2.40734279e-01 -9.15682375e-01
9.31945384e-01 -4.91692454e-01 -3.21138293e-01 3.62775296e-01
-1.57271042e-01 -6.87923729e-02 6.30142331e-01 3.63495678e-01
1.02804101e+00 -4.10516977e-01 8.54058862e-01 -6.05500817e-01
1.10356975e+00 1.89304113e-01 3.41712087e-01 3.00062299e-01
-3.97028923e-01 5.53077459e-01 5.97017467e-01 -3.22262436e-01
-1.22703731e+00 -3.38405907e-01 -4.96531934e-01 1.26581752e+00
-3.16280931e-01 -7.16230452e-01 -7.76659787e-01 -2.90829688e-01
-4.00124006e-02 6.86000645e-01 -2.97724396e-01 3.02178681e-01
-4.81721550e-01 -6.80608690e-01 6.74701631e-01 -2.09932193e-01
6.41121686e-01 -1.33881187e+00 -5.94619632e-01 3.77591550e-01
-5.92111707e-01 -9.86362636e-01 -3.95776391e-01 3.39606106e-02
-3.89748484e-01 -1.06161809e+00 -3.66362274e-01 -8.73555660e-01
3.98943365e-01 4.28665012e-01 1.09700203e+00 4.30310547e-01
-2.16015592e-01 -2.09969088e-01 -1.06122923e+00 -4.09886509e-01
-9.39478338e-01 1.62563652e-01 -1.79431170e-01 -1.84779972e-01
8.07361245e-01 -3.38545084e-01 -1.05431555e-02 1.28913656e-01
-9.11480367e-01 1.72073409e-01 -4.32612486e-02 6.73298001e-01
-6.19967937e-01 5.44490889e-02 4.71569061e-01 -9.97503757e-01
9.53182697e-01 -9.97714639e-01 -1.44051969e-01 -3.51194978e-01
-1.80719823e-01 -1.95625365e-01 1.17439771e+00 -2.64989674e-01
-7.55888522e-01 -3.28396410e-01 -2.08592117e-01 2.95336246e-01
2.16044374e-02 8.54551852e-01 1.02668740e-01 -1.93355093e-03
1.06719184e+00 8.60146731e-02 -1.41194165e-01 -4.09162968e-01
1.52903711e-02 1.53639758e+00 2.97840267e-01 -1.93552166e-01
8.09526086e-01 1.84636280e-01 -4.65606004e-01 -1.18737829e+00
-9.86128449e-01 -1.06097353e+00 -6.38041914e-01 -4.67940748e-01
8.01228404e-01 -7.77830958e-01 -4.88563269e-01 6.90645874e-01
-1.35801756e+00 6.99146762e-02 4.05088454e-01 1.32029623e-01
-2.46994406e-01 8.72612774e-01 -9.25680578e-01 -8.69574964e-01
-3.64560992e-01 -7.81187773e-01 9.50409025e-02 7.36760050e-02
-8.11033905e-01 -9.53262687e-01 3.32909316e-01 7.29550242e-01
2.37536520e-01 6.28461465e-02 1.07062232e+00 -6.70830727e-01
1.59358621e-01 -1.52667895e-01 -2.01428816e-01 2.90133864e-01
-7.28674559e-03 6.41902769e-03 -7.50384510e-01 6.78850412e-02
4.82136339e-01 -6.03211939e-01 4.78218257e-01 -3.28105152e-01
7.06162632e-01 -8.11366141e-01 3.40344578e-01 -4.42553014e-02
1.30069363e+00 -1.83634385e-02 7.47630000e-01 7.12380767e-01
6.95958972e-01 6.95503592e-01 3.17989886e-01 1.02481103e+00
3.76328290e-01 3.45358491e-01 3.23104858e-01 5.77033162e-01
1.23444080e-01 -2.85495192e-01 6.42649293e-01 1.53712666e+00
8.90393741e-03 -1.09865423e-03 -1.21748221e+00 6.63117826e-01
-1.73883903e+00 -1.42144501e+00 -1.05843318e+00 1.99090433e+00
1.31694007e+00 -8.99012759e-02 4.75394011e-01 6.03760540e-01
7.94287801e-01 2.44718745e-01 2.54020393e-01 -7.98832059e-01
-2.00619265e-01 -1.53642029e-01 1.69101521e-01 7.28341937e-01
-8.61037195e-01 8.42588246e-01 4.56653023e+00 8.23590517e-01
-1.17848802e+00 4.43673819e-01 1.07915960e-01 2.64609098e-01
-2.41793126e-01 -1.29855881e-02 -6.11045480e-01 9.36129928e-01
8.47212315e-01 -4.18816924e-01 6.24670804e-01 1.05534172e+00
4.26895082e-01 -2.47374490e-01 -5.35705149e-01 1.24727440e+00
6.29126012e-01 -7.78423429e-01 -4.06468332e-01 -2.36082569e-01
9.97857094e-01 1.16970710e-01 -4.28278893e-01 4.70848113e-01
-4.73605283e-02 -9.41852927e-01 9.62534845e-01 -5.65608181e-02
1.30824447e-01 -6.52801633e-01 9.23022330e-01 9.09416199e-01
-8.79331708e-01 -2.84447491e-01 -6.02455914e-01 -7.30657876e-01
1.88762881e-02 9.17947233e-01 -7.67953575e-01 -9.58490297e-02
5.22024810e-01 7.85609305e-01 -8.63840997e-01 1.06342435e+00
-3.44024897e-01 7.25947559e-01 9.81739089e-02 -6.91358566e-01
2.18803927e-01 -2.00687215e-01 6.44837201e-01 1.76511788e+00
1.63083062e-01 -1.32736355e-01 1.44720778e-01 9.48371530e-01
-1.34899348e-01 9.04180884e-01 -4.92380708e-01 -2.82797039e-01
1.23136297e-01 1.52749968e+00 -7.10191727e-01 -3.74965817e-02
-6.53982341e-01 9.45462525e-01 4.47341830e-01 -2.78255731e-01
-7.45126009e-01 -4.61420029e-01 2.26363353e-02 4.54559088e-01
-3.71381938e-01 -3.76621366e-01 -4.64193910e-01 -1.18973517e+00
-1.79356620e-01 -9.90903378e-01 3.13254535e-01 -5.63739717e-01
-1.50997734e+00 4.66685653e-01 -3.36441368e-01 -1.30094802e+00
2.00910005e-03 -5.26464820e-01 -6.06617987e-01 7.57910311e-01
-1.03328097e+00 -9.54887569e-01 -4.62361634e-01 4.93619561e-01
7.58367360e-01 -2.65243739e-01 4.47743565e-01 7.25178421e-01
-2.70293802e-01 2.57039934e-01 4.04115543e-02 5.42756379e-01
7.14150965e-01 -1.16150022e+00 -5.35794459e-02 7.20893204e-01
-1.77735552e-01 5.94475448e-01 1.17507493e+00 -7.63655186e-01
-1.02657115e+00 -7.35784709e-01 1.89000559e+00 -4.30520892e-01
1.39380395e+00 -2.34442040e-01 -9.59886074e-01 1.90861911e-01
6.45634353e-01 -5.16810775e-01 8.31896305e-01 -2.58677956e-02
-4.48616356e-01 4.75549191e-01 -1.07360756e+00 5.58056891e-01
5.47233284e-01 -8.19093406e-01 -7.56017089e-01 4.86726135e-01
6.26257733e-02 -7.22264126e-02 -3.58090639e-01 -2.86674142e-01
1.25369489e-01 -1.31799173e+00 6.98560625e-02 -4.18213874e-01
1.25708556e+00 -3.10010970e-01 -6.89352974e-02 -1.07263792e+00
-4.01354581e-01 -5.13849318e-01 4.52112228e-01 1.45277321e+00
1.42416731e-01 -6.37141988e-02 3.35056722e-01 3.99111181e-01
-1.65827945e-01 -9.51221064e-02 -6.51271880e-01 -4.13165659e-01
4.72401321e-01 -3.80647570e-01 -3.05848252e-02 1.36661541e+00
1.23139703e+00 6.22793734e-01 -6.51089787e-01 -5.35527289e-01
3.44949156e-01 2.74421591e-02 8.14580381e-01 -1.06689489e+00
-6.92768395e-02 -6.97165012e-01 -6.65232897e-01 -6.99596941e-01
3.39330792e-01 -1.34216702e+00 -6.54113516e-02 -1.11629236e+00
8.83756459e-01 1.14314213e-01 4.40808862e-01 3.81716430e-01
1.25393262e-02 3.01863998e-01 3.99435222e-01 5.16899407e-01
-5.23778379e-01 3.37723523e-01 9.64011431e-01 -9.40872729e-02
-2.29209855e-01 -3.77304316e-01 -4.22039062e-01 9.19936717e-01
8.13259423e-01 -5.90753138e-01 6.19358420e-02 2.55133957e-01
1.01806450e+00 -1.40245393e-01 1.36789441e-01 -1.03323948e+00
2.03031763e-01 -2.69215196e-01 -4.09337074e-01 -4.29383993e-01
-5.47909975e-01 -6.34000301e-01 -2.21020669e-01 4.54942465e-01
-2.12853208e-01 1.16220728e-01 -1.45167232e-01 -1.10481322e-01
-3.77286673e-01 -1.05774641e+00 1.29667878e+00 -3.42118442e-01
-4.36036825e-01 -4.28916067e-01 -9.18506861e-01 3.84085298e-01
9.05400634e-01 -3.83646153e-02 -3.76200527e-01 -4.38078642e-01
-1.49340823e-01 -7.82739222e-02 7.09247410e-01 5.07056057e-01
2.18394995e-01 -1.16524351e+00 -9.44142401e-01 -1.90890223e-01
3.48537743e-01 -6.40905440e-01 1.81704879e-01 1.04380047e+00
-9.66142178e-01 1.43915936e-01 -3.49740267e-01 -8.42904225e-02
-1.51045644e+00 6.30304694e-01 1.10816367e-01 -3.11204325e-03
-3.35791975e-01 3.08077514e-01 -2.97150284e-01 -4.05902803e-01
-2.38689587e-01 3.88004720e-01 -5.93026042e-01 5.74422717e-01
7.48382211e-01 7.31505692e-01 -9.80425179e-02 -1.26380980e+00
-1.07874855e-01 2.53608376e-02 2.51173042e-02 1.25047520e-01
1.03504694e+00 -2.83616692e-01 -9.85162199e-01 1.00038719e+00
1.45516717e+00 5.72884500e-01 -6.05225526e-02 -4.35724147e-02
4.39428777e-01 -6.61456645e-01 -7.88975284e-02 -3.64320338e-01
-5.56396842e-01 8.73509169e-01 7.29600340e-02 7.98109412e-01
6.08301103e-01 -5.94629496e-02 9.90909696e-01 1.02097370e-01
4.90947157e-01 -1.55782151e+00 1.21611297e-01 1.26466942e+00
9.41315949e-01 -1.30367911e+00 -9.02463049e-02 -2.79784322e-01
-7.18284011e-01 1.32506382e+00 3.47872674e-01 3.31603810e-02
6.32681787e-01 1.97622225e-01 1.46214992e-01 -1.86029181e-01
-1.90764934e-01 -2.30030373e-01 1.24635227e-01 -9.82369334e-02
1.31980801e+00 1.79599151e-01 -1.40002370e+00 7.59930253e-01
-8.50807130e-01 -1.83035523e-01 1.17343390e+00 7.05131710e-01
-1.07044971e+00 -9.64830816e-01 -6.75165772e-01 3.86974335e-01
-8.52224290e-01 -2.89812535e-01 -6.90751672e-01 3.95016700e-01
2.08688185e-01 1.38289618e+00 -3.84156406e-01 -5.92338026e-01
-2.29054838e-01 2.65379876e-01 -4.16176813e-03 -8.04157495e-01
-1.13196826e+00 -2.26453528e-01 1.49112865e-01 1.40451998e-01
-4.90849048e-01 -8.44451010e-01 -1.51315475e+00 -9.06947911e-01
-3.44381720e-01 4.02041972e-01 4.47547257e-01 9.93895888e-01
-2.19941854e-01 -1.61412448e-01 8.36402833e-01 -3.86534840e-01
-3.41120332e-01 -1.27748990e+00 -8.01369011e-01 9.18597043e-01
3.52463610e-02 -3.68217260e-01 -7.96191692e-01 2.22703576e-01] | [8.955350875854492, 10.8219633102417] |
a53cb31b-a7ab-4bee-90cb-28cdd182604b | ndjir-neural-direct-and-joint-inverse | 2302.00675 | null | https://arxiv.org/abs/2302.00675v1 | https://arxiv.org/pdf/2302.00675v1.pdf | NDJIR: Neural Direct and Joint Inverse Rendering for Geometry, Lights, and Materials of Real Object | The goal of inverse rendering is to decompose geometry, lights, and materials given pose multi-view images. To achieve this goal, we propose neural direct and joint inverse rendering, NDJIR. Different from prior works which relies on some approximations of the rendering equation, NDJIR directly addresses the integrals in the rendering equation and jointly decomposes geometry: signed distance function, lights: environment and implicit lights, materials: base color, roughness, specular reflectance using the powerful and flexible volume rendering framework, voxel grid feature, and Bayesian prior. Our method directly uses the physically-based rendering, so we can seamlessly export an extracted mesh with materials to DCC tools and show material conversion examples. We perform intensive experiments to show that our proposed method can decompose semantically well for real object in photogrammetric setting and what factors contribute towards accurate inverse rendering. | ['Takuya Narihira', 'Kazuki Yoshiyama'] | 2023-02-02 | null | null | null | null | ['inverse-rendering'] | ['computer-vision'] | [ 4.10735458e-01 -6.64689243e-02 7.60807991e-01 -3.51138711e-01
-5.26219785e-01 -4.74482119e-01 7.94573665e-01 -5.22958815e-01
2.51078457e-01 5.93421519e-01 1.22839585e-02 -1.41752616e-01
-3.14651608e-01 -1.27598965e+00 -5.82016528e-01 -4.56716985e-01
3.27703327e-01 9.46598291e-01 3.36161584e-01 -2.06947416e-01
3.08341801e-01 8.46493185e-01 -1.71014512e+00 1.05162352e-01
9.82406080e-01 7.48511314e-01 1.07020941e-02 7.97680795e-01
-3.21495712e-01 2.67602295e-01 -9.88540575e-02 -2.43450403e-01
4.74242449e-01 -1.64144695e-01 -6.54036045e-01 5.25606945e-02
5.11200726e-01 -7.44466007e-01 1.45169765e-01 9.41672564e-01
4.00524497e-01 2.70050347e-01 1.12712801e+00 -9.73081529e-01
-6.99053943e-01 -2.54298627e-01 -9.24875498e-01 -6.37165606e-01
6.84004426e-01 -7.46364817e-02 5.36275923e-01 -9.74532485e-01
5.21143615e-01 1.69221020e+00 7.02832580e-01 3.38886082e-01
-1.35928786e+00 -6.40525579e-01 4.02541310e-02 -3.34455729e-01
-1.29897761e+00 -7.03870654e-02 1.04957473e+00 -4.37215328e-01
5.52450418e-01 7.00021446e-01 9.17787433e-01 5.74943662e-01
1.34003326e-01 1.33160297e-02 1.63867188e+00 -3.43445212e-01
2.56734550e-01 6.75133243e-02 1.04195274e-01 6.84371710e-01
-1.01239689e-01 4.39696640e-01 -3.02164495e-01 -7.20076263e-01
1.25512421e+00 4.85783182e-02 -4.64099139e-01 -2.58899540e-01
-9.97173846e-01 3.93761694e-01 1.10522203e-01 -5.74717522e-01
-1.93987250e-01 2.46017605e-01 -2.37680748e-01 -1.05623193e-01
7.86932826e-01 -9.94296372e-02 -3.87409866e-01 1.21743672e-01
-4.75810558e-01 6.89941868e-02 8.44299912e-01 1.08514750e+00
1.05135024e+00 -9.85672348e-04 2.03898981e-01 8.87729406e-01
9.29403245e-01 1.13176191e+00 -5.39165556e-01 -1.52260435e+00
1.46133164e-02 4.08957541e-01 2.12939262e-01 -9.28584635e-01
-1.03783362e-01 -5.83287627e-02 -7.30145276e-01 8.37770164e-01
-1.27682239e-01 2.69757777e-01 -9.94686186e-01 1.16897357e+00
8.94565165e-01 3.86783332e-01 -2.30877236e-01 7.93311477e-01
9.18939650e-01 5.41250408e-01 -1.38646170e-01 -3.22347060e-02
1.30346966e+00 -3.61539632e-01 -6.25123322e-01 2.89094537e-01
-7.04874471e-02 -1.24334705e+00 1.11471570e+00 5.95872462e-01
-1.17366278e+00 -3.20194304e-01 -9.01110649e-01 -3.98445398e-01
1.15369350e-01 1.16246879e-01 7.44635284e-01 6.04516745e-01
-9.21110272e-01 7.69991696e-01 -7.04748034e-01 1.14094682e-01
7.36985207e-02 1.51633799e-01 2.68557239e-02 -2.66187280e-01
-6.22142076e-01 5.30707598e-01 -3.74199808e-01 1.96863949e-01
-5.74187636e-01 -9.89011347e-01 -5.53234577e-01 -2.35382125e-01
6.44565970e-02 -1.22359192e+00 6.80389524e-01 -5.16257644e-01
-2.13942218e+00 6.74189389e-01 -1.95752606e-02 6.75210297e-01
5.38480461e-01 -3.01618785e-01 -2.22439747e-02 -2.43798108e-03
-1.50189057e-01 7.38852471e-02 7.68799365e-01 -2.13138747e+00
-1.19304061e-01 -5.48410296e-01 1.69298232e-01 3.04934055e-01
3.08536649e-01 -3.69540930e-01 -4.25749898e-01 -3.69361073e-01
5.90198755e-01 -6.77857876e-01 -2.26515308e-01 4.73808020e-01
-4.33486670e-01 1.30156964e-01 9.03262854e-01 -9.55050826e-01
3.41096580e-01 -2.12789226e+00 -2.63488404e-02 4.73578781e-01
4.39922623e-02 -2.60722637e-01 -8.63814261e-03 2.54178166e-01
-2.61951014e-02 7.28752762e-02 -3.38583291e-01 -4.73721534e-01
1.36835784e-01 3.23875964e-01 -5.23551106e-01 7.22865283e-01
-3.12186092e-01 3.30666840e-01 -7.00805604e-01 -5.30210912e-01
6.59335017e-01 1.22798073e+00 -7.56357849e-01 2.62506753e-01
-2.43669868e-01 6.45121396e-01 -5.36459565e-01 6.26084507e-01
1.57257009e+00 1.48844764e-01 -1.44198656e-01 -6.93664610e-01
-1.63653269e-01 4.10063751e-02 -1.48781645e+00 1.66806221e+00
-8.54991496e-01 -2.07319986e-02 6.76582575e-01 -1.55628398e-01
1.00080645e+00 2.32209057e-01 3.57062966e-01 -4.52002734e-01
-7.59781227e-02 9.86241773e-02 -5.36922514e-01 -9.49012712e-02
4.64763999e-01 -4.54700828e-01 6.84812069e-01 5.09635985e-01
-5.23494124e-01 -1.14309990e+00 -8.30303192e-01 1.79342762e-01
6.95058167e-01 1.06838036e+00 -5.96105568e-02 -2.71201879e-01
4.50561434e-01 -1.07740693e-01 4.39297795e-01 1.73347071e-01
7.36382365e-01 9.48250711e-01 -1.91548262e-02 -2.27131933e-01
-7.65943468e-01 -1.61353672e+00 -2.92622894e-01 6.34504974e-01
4.24701065e-01 -6.91693425e-02 -4.26348537e-01 7.84315343e-04
5.37317134e-02 9.68734562e-01 -3.37505877e-01 3.01393479e-01
-7.08521545e-01 -7.56113827e-01 -1.23759300e-01 1.36198297e-01
4.38128799e-01 -4.07915592e-01 -3.69291484e-01 -1.59578651e-01
1.32903913e-02 -8.47485363e-01 -2.47331932e-01 -5.07732451e-01
-1.14936030e+00 -1.18607640e+00 -4.11421925e-01 -8.78996402e-02
7.88748264e-01 2.87533373e-01 1.51608634e+00 3.78827035e-01
-5.19731820e-01 1.06155789e+00 -4.17490378e-02 -2.14234799e-01
-3.96576375e-01 -7.91934371e-01 -4.94804204e-01 2.63200756e-02
-5.11941314e-01 -1.24898779e+00 -8.44274282e-01 3.92380297e-01
-8.47303092e-01 7.33308375e-01 3.18206064e-02 3.19853365e-01
1.01949263e+00 -1.96743876e-01 -3.28199476e-01 -9.20011938e-01
2.30082706e-01 -1.80936709e-01 -8.92493486e-01 2.25256369e-01
-4.94252324e-01 -1.87425837e-01 3.37926894e-01 -1.62672341e-01
-1.64853859e+00 -3.89158316e-02 -2.96085715e-01 -3.36562634e-01
-1.82626247e-01 -1.44164264e-01 -3.64133030e-01 -2.67160594e-01
3.00783068e-01 -2.26956904e-02 -2.03516677e-01 -8.52215767e-01
5.11487544e-01 3.25595558e-01 3.48954618e-01 -1.07521880e+00
1.05659473e+00 1.30795538e+00 5.51498652e-01 -8.55150104e-01
-3.70609850e-01 -9.06196833e-02 -4.78389025e-01 -4.07372952e-01
7.20157981e-01 -6.81297779e-01 -1.05814004e+00 1.66267931e-01
-1.40249479e+00 -2.54248738e-01 -3.69926661e-01 4.76004899e-01
-7.31101930e-01 6.00850463e-01 -5.06682396e-01 -1.21111870e+00
-1.47470459e-01 -1.15249360e+00 1.57437241e+00 -1.57813057e-01
2.17551187e-01 -1.03224576e+00 1.97220862e-01 2.82669097e-01
2.68814176e-01 6.81857347e-01 9.38915312e-01 6.34823442e-01
-1.24577844e+00 2.16259956e-01 -4.45888788e-01 3.25695395e-01
3.30963850e-01 4.71332103e-01 -1.19935453e+00 1.55240819e-01
2.75949121e-01 1.33576021e-01 4.49690700e-01 3.49968731e-01
1.11382043e+00 4.73605208e-02 -2.80697942e-01 1.21634817e+00
1.90984666e+00 -7.17235059e-02 9.54726279e-01 -9.16282758e-02
9.83990788e-01 6.07853115e-01 5.70044279e-01 6.78460300e-01
6.11149371e-01 6.74552739e-01 8.26976836e-01 -2.66759574e-01
-5.86772978e-01 -5.34924231e-02 4.35714349e-02 1.10083687e+00
-9.10691202e-01 4.09100205e-02 -6.83994055e-01 -7.50763342e-02
-1.40857422e+00 -5.53968012e-01 -8.79405558e-01 2.36738157e+00
5.50434053e-01 -5.69208741e-01 -6.49720848e-01 -2.92695791e-01
3.63599509e-01 -1.72979802e-01 -3.62409115e-01 -2.62616485e-01
2.29967777e-02 4.06393021e-01 5.50917506e-01 9.65600014e-01
-3.10140222e-01 4.95442897e-01 6.60807133e+00 7.89470434e-01
-7.23478436e-01 3.99735481e-01 2.84826279e-01 2.20750168e-01
-1.33595014e+00 4.26043451e-01 -4.93280202e-01 7.25434795e-02
4.46009696e-01 3.75928998e-01 9.44327533e-01 5.65916419e-01
3.05379808e-01 -4.38878447e-01 -9.20822322e-01 1.10063422e+00
-1.59912631e-02 -1.20041227e+00 2.60641634e-01 2.85877675e-01
6.12573385e-01 -1.29891291e-01 -4.53551151e-02 -1.56072989e-01
8.73912871e-01 -8.35514843e-01 7.70465136e-01 9.46534336e-01
9.51493204e-01 -5.06267786e-01 8.34933221e-02 4.10401285e-01
-1.17791450e+00 5.97934484e-01 -2.04813555e-01 -4.32984494e-02
6.39742255e-01 9.81742144e-01 -2.98241735e-01 9.68184650e-01
6.80454910e-01 3.63841116e-01 -8.98198783e-02 5.50834298e-01
-4.94141802e-02 3.45716506e-01 -7.25146711e-01 6.64528489e-01
-5.38638353e-01 -1.06406486e+00 5.81315219e-01 5.96040428e-01
5.23165941e-01 4.48507190e-01 2.57841557e-01 1.45127392e+00
4.76983994e-01 5.53042628e-02 -3.65259856e-01 6.52619362e-01
1.72819942e-01 1.28507578e+00 -8.47454309e-01 -2.49183431e-01
-1.29219159e-01 9.45511460e-01 -4.54312079e-02 7.86170065e-01
-9.44807708e-01 9.15111825e-02 4.53710258e-01 3.88927370e-01
-2.50833899e-01 -4.13055152e-01 -7.15967238e-01 -9.27579641e-01
4.55503948e-02 -3.47545475e-01 -2.43734032e-01 -1.28111303e+00
-1.36576343e+00 3.65381330e-01 2.31984779e-01 -1.27920318e+00
1.95159733e-01 -7.33195305e-01 -3.59930784e-01 1.32435632e+00
-1.77264905e+00 -1.47405541e+00 -4.86135304e-01 6.67853057e-01
2.83323258e-01 5.24831533e-01 8.91845345e-01 9.31166038e-02
9.78667364e-02 -2.66022831e-01 1.47639990e-01 -3.65481585e-01
5.29911935e-01 -1.05441022e+00 2.28575483e-01 3.14052343e-01
-3.09574336e-01 6.50766075e-01 7.71391392e-01 -9.30099845e-01
-1.78878069e+00 -7.46500134e-01 -2.39197537e-01 -5.06271303e-01
2.46013533e-02 -1.84906647e-01 -6.77652061e-01 4.99547303e-01
7.68882036e-03 3.26925665e-01 4.20852840e-01 -1.24234580e-01
-2.88660407e-01 1.21060394e-01 -1.52027285e+00 4.07578409e-01
1.37970412e+00 -4.40317899e-01 -2.51348048e-01 4.10566360e-01
5.38550854e-01 -8.39867353e-01 -1.11003292e+00 6.33545339e-01
6.53958261e-01 -1.49691308e+00 1.49019814e+00 2.29407772e-01
2.66381562e-01 -6.04765415e-01 -5.12632906e-01 -1.22574604e+00
-2.56381452e-01 -6.25591993e-01 8.51803571e-02 1.38595772e+00
-2.24205643e-01 -1.04400337e+00 3.48847896e-01 8.62687409e-01
-4.45995122e-01 -4.28820044e-01 -8.00698400e-01 -5.90510964e-01
-2.18557864e-01 -6.07259810e-01 9.63317692e-01 8.07214499e-01
-9.96085584e-01 -1.57043695e-01 -1.12605602e-01 6.57289565e-01
1.23305416e+00 5.42342782e-01 8.16292226e-01 -1.67813694e+00
-5.60986996e-01 7.74151608e-02 2.91445941e-01 -1.00013292e+00
1.99885778e-02 -6.36865199e-01 1.15661465e-01 -1.87542033e+00
2.00448081e-01 -1.22538185e+00 5.91017365e-01 -8.13058913e-02
1.96204513e-01 2.96363086e-01 -1.34565949e-01 1.43378943e-01
1.99380428e-01 7.50386596e-01 1.74589682e+00 9.33655351e-02
-4.58033346e-02 -1.02327392e-01 -2.51865268e-01 1.25671637e+00
2.72960633e-01 -3.57814759e-01 -5.31841040e-01 -7.51901746e-01
5.66739261e-01 7.43534341e-02 6.88286483e-01 -5.03008187e-01
-3.34789157e-01 -5.18267691e-01 2.43070319e-01 -8.75749469e-01
9.96199131e-01 -1.31144834e+00 9.74356115e-01 -2.26039952e-03
3.53061199e-01 -2.04770505e-01 2.99186092e-02 5.58552921e-01
4.05491412e-01 -1.11356050e-01 7.27414131e-01 -3.67440879e-01
-5.97891361e-02 3.83079439e-01 2.53711402e-01 -1.33678392e-01
5.07456660e-01 -5.12274623e-01 -1.65734112e-01 -2.18918845e-01
-5.68014801e-01 -2.82863140e-01 1.10826552e+00 -1.89325958e-01
8.16087782e-01 -1.46952200e+00 -5.54280400e-01 2.48634651e-01
-4.83387291e-01 4.32199448e-01 3.58156025e-01 7.26050615e-01
-9.22634184e-01 -4.48426366e-01 1.59002244e-01 -7.75137246e-01
-1.18160605e+00 6.09448291e-02 2.74222493e-01 1.71268076e-01
-9.80192125e-01 6.30082309e-01 6.42527759e-01 -9.52057779e-01
-3.22038293e-01 -5.09509623e-01 1.71699762e-01 -4.54561353e-01
2.93810546e-01 7.68410265e-01 2.76608653e-02 -6.77253544e-01
-2.71444708e-01 1.28238177e+00 7.86745310e-01 -4.45010126e-01
1.54492795e+00 -3.64962637e-01 -6.42754555e-01 3.97496074e-01
9.03203487e-01 5.55309832e-01 -1.25235355e+00 2.34344348e-01
-8.91212642e-01 -8.75855386e-01 2.97435015e-01 -6.42422616e-01
-1.01380408e+00 8.90297294e-01 5.33928156e-01 -6.84960932e-02
9.89959359e-01 -1.71114206e-01 6.19866490e-01 -2.02023327e-01
8.62261415e-01 -8.82273853e-01 -3.58892620e-01 4.37299937e-01
1.19710326e+00 -7.01692641e-01 7.13300169e-01 -1.37949562e+00
1.44654252e-02 1.14166880e+00 3.78774643e-01 -4.17325616e-01
1.14353693e+00 6.69150710e-01 -1.13788247e-01 -4.14354116e-01
-3.06478500e-01 2.12376356e-01 4.14265841e-01 7.93446541e-01
3.83451819e-01 1.15256861e-01 1.77899554e-01 1.04689375e-01
-1.79338872e-01 2.14707019e-04 3.66530180e-01 7.50174582e-01
-8.50958526e-02 -1.01340723e+00 -9.47732449e-01 2.59564847e-01
1.81791082e-01 -4.73181605e-02 -7.14100664e-03 6.74681664e-01
2.28556365e-01 4.25641388e-01 1.33042052e-01 -7.77231082e-02
3.85875225e-01 -2.80103773e-01 7.41345763e-01 -6.51461661e-01
-1.54728979e-01 4.54603314e-01 -4.96072322e-02 -7.91901350e-01
-6.47108316e-01 -5.36965013e-01 -1.34375834e+00 -1.73386708e-01
-5.26924551e-01 -2.34411255e-01 9.99344409e-01 6.79895341e-01
1.98123440e-01 5.59151590e-01 5.11907518e-01 -1.37703443e+00
-2.67622262e-01 -5.37709117e-01 -7.77747512e-01 3.24690819e-01
1.76863477e-01 -1.04379535e+00 -6.55068219e-01 1.07041247e-01] | [9.663744926452637, -3.1272757053375244] |
60bd774c-df37-4680-a860-cb25b4d343ba | adaptive-multi-teacher-knowledge-distillation | 2306.06634 | null | https://arxiv.org/abs/2306.06634v1 | https://arxiv.org/pdf/2306.06634v1.pdf | Adaptive Multi-Teacher Knowledge Distillation with Meta-Learning | Multi-Teacher knowledge distillation provides students with additional supervision from multiple pre-trained teachers with diverse information sources. Most existing methods explore different weighting strategies to obtain a powerful ensemble teacher, while ignoring the student with poor learning ability may not benefit from such specialized integrated knowledge. To address this problem, we propose Adaptive Multi-teacher Knowledge Distillation with Meta-Learning (MMKD) to supervise student with appropriate knowledge from a tailored ensemble teacher. With the help of a meta-weight network, the diverse yet compatible teacher knowledge in the output layer and intermediate layers is jointly leveraged to enhance the student performance. Extensive experiments on multiple benchmark datasets validate the effectiveness and flexibility of our methods. Code is available: https://github.com/Rorozhl/MMKD. | ['Can Wang', 'Defang Chen', 'Hailin Zhang'] | 2023-06-11 | null | null | null | null | ['meta-learning'] | ['methodology'] | [-1.59604460e-01 1.30485326e-01 -5.25774002e-01 -4.86405462e-01
-4.50689852e-01 -4.96193081e-01 2.54956782e-01 7.30289072e-02
-3.80525351e-01 9.19801056e-01 1.29492015e-01 -2.26508245e-01
-4.62257475e-01 -9.12717342e-01 -4.43649113e-01 -9.26714420e-01
7.08899617e-01 2.11514354e-01 2.32294336e-01 -2.89033204e-01
-8.90367031e-02 2.31717840e-01 -1.42423773e+00 1.10171326e-01
1.69487715e+00 6.16640568e-01 2.97632694e-01 4.52026844e-01
-2.08906800e-01 8.54001224e-01 -5.48624337e-01 -4.60909545e-01
1.89596117e-02 -2.84518212e-01 -6.55006945e-01 -1.52683005e-01
5.28215110e-01 -3.87806147e-01 -2.43460611e-01 9.34897125e-01
7.75516987e-01 4.55192477e-01 4.10478681e-01 -1.06911278e+00
-7.39804924e-01 1.15412605e+00 -7.00448096e-01 3.43076050e-01
-1.66231930e-01 1.86745301e-01 7.21285701e-01 -9.77762818e-01
-5.39155230e-02 8.72776210e-01 2.98728943e-01 6.37600183e-01
-1.08945179e+00 -1.00039053e+00 4.42711025e-01 3.19420278e-01
-1.27869105e+00 -1.98615909e-01 9.63643074e-01 -2.25309476e-01
3.71713161e-01 6.59694057e-03 6.12944722e-01 9.58088100e-01
-4.42707419e-01 1.12880039e+00 1.12128115e+00 -3.12736928e-01
-3.10683250e-01 5.98393440e-01 4.41861928e-01 8.14375341e-01
2.83845425e-01 -1.26681849e-01 -5.53621352e-01 -8.47715791e-03
3.88700515e-01 3.44440579e-01 -4.07673150e-01 -2.31641516e-01
-8.85799825e-01 5.86389542e-01 3.35524231e-01 3.54070812e-01
-3.20927620e-01 -3.18073705e-02 6.47580251e-02 6.04770184e-01
5.30775726e-01 4.10691530e-01 -8.19217443e-01 -6.36332557e-02
-7.17609465e-01 -8.14914629e-02 5.44189572e-01 8.69670868e-01
9.71783638e-01 3.62951010e-01 -3.52627903e-01 1.07750285e+00
3.03577721e-01 4.38301384e-01 7.52337933e-01 -7.91039884e-01
3.82115245e-01 1.04324174e+00 -3.62491667e-01 -4.42216724e-01
1.30441844e-01 -9.91267979e-01 -8.00310969e-01 2.70815998e-01
1.62434742e-01 -6.18218541e-01 -9.18551624e-01 1.50302029e+00
7.76401758e-01 1.01117241e+00 2.57289022e-01 6.11320853e-01
1.23314083e+00 5.35753310e-01 2.00996548e-01 -8.49959552e-02
1.13151503e+00 -1.21318769e+00 -3.42039466e-01 5.25951274e-02
5.09131849e-01 -5.69677711e-01 7.02052593e-01 5.71346700e-01
-1.08272159e+00 -7.44467437e-01 -8.84045541e-01 1.50198132e-01
-1.64971933e-01 2.01847479e-01 3.48858654e-01 3.72930408e-01
-7.47734070e-01 3.48674715e-01 -4.11578536e-01 4.70912755e-01
6.59201443e-01 5.13466716e-01 -5.72477765e-02 2.89492924e-02
-1.12284470e+00 7.30721891e-01 5.83528221e-01 -2.05781594e-01
-1.23339915e+00 -1.21560729e+00 -3.68209571e-01 3.29523683e-01
6.68246448e-01 -8.23202014e-01 1.57194734e+00 -1.08347857e+00
-1.88506877e+00 2.54078150e-01 2.30469644e-01 1.39055047e-02
4.15124536e-01 -4.57727075e-01 -1.55730799e-01 4.50354256e-02
-3.61874491e-01 4.32191879e-01 8.05431962e-01 -1.16349494e+00
-9.85457897e-01 -3.75752032e-01 5.43568470e-02 6.75764263e-01
-8.51161420e-01 -2.99642771e-01 -2.76069313e-01 -4.80705202e-01
-1.93885610e-01 -5.28922021e-01 -3.27439874e-01 -6.39635384e-01
-1.40514702e-01 -7.76160181e-01 9.05270755e-01 -2.19711468e-01
1.50960720e+00 -1.85971332e+00 1.98418066e-01 3.14527005e-01
6.21551394e-01 7.95482874e-01 -3.83310229e-01 1.11000158e-01
-5.97663671e-02 -1.48524465e-02 1.73160627e-01 -4.12298627e-02
-2.46306419e-01 1.79399058e-01 -2.19611034e-01 -5.03236391e-02
5.47833368e-02 7.38484740e-01 -1.28076756e+00 -4.08363402e-01
2.02319667e-01 5.91350675e-01 -3.66126060e-01 7.01645672e-01
-2.02244699e-01 6.29353523e-01 -1.11367011e+00 6.24893665e-01
3.59216422e-01 -3.73428851e-01 7.68570751e-02 2.97834650e-02
2.74097305e-02 2.74161786e-01 -1.21473670e+00 1.51995647e+00
-6.38744414e-01 1.23401262e-01 2.71919191e-01 -1.16557717e+00
9.43210304e-01 5.27742088e-01 3.39300662e-01 -3.84094179e-01
1.69784904e-01 6.77971542e-02 1.89989299e-01 -5.30457675e-01
4.64669555e-01 2.04772279e-01 4.87041891e-01 6.08218491e-01
4.32386786e-01 3.85086797e-02 -9.13238674e-02 3.29568475e-01
7.05090225e-01 1.77239254e-01 1.44618191e-02 -1.01483010e-01
5.53230166e-01 -3.12430233e-01 7.71547616e-01 5.84616303e-01
-1.99178830e-01 7.26656914e-02 -1.11242719e-01 -2.36925393e-01
-5.15378118e-01 -8.56163442e-01 6.49422333e-02 1.88151336e+00
-4.55059606e-04 -2.54314840e-01 -5.16431630e-01 -1.04144347e+00
2.11640447e-01 6.79765046e-01 -4.43062246e-01 -3.05574566e-01
-6.66728973e-01 -6.75253093e-01 3.89496744e-01 6.14664018e-01
4.98416692e-01 -8.72224629e-01 -3.93888175e-01 2.80197054e-01
7.71543458e-02 -5.51126063e-01 -3.15109998e-01 2.99750239e-01
-9.57669914e-01 -1.05002606e+00 -7.22963274e-01 -6.67706370e-01
8.64960670e-01 5.22549272e-01 1.18702245e+00 5.08313835e-01
6.90527707e-02 6.32725298e-01 -2.98123509e-01 -5.49634099e-01
-2.14096680e-01 3.70634675e-01 2.20653370e-01 -1.11450963e-01
5.12408972e-01 -8.81817341e-01 -4.55493242e-01 2.07509294e-01
-7.79285908e-01 3.89404327e-01 4.78316039e-01 9.46795166e-01
5.01869321e-01 3.15044194e-01 1.01507688e+00 -1.03054953e+00
7.21611083e-01 -8.52920413e-01 -5.66310287e-01 5.57403922e-01
-9.73463714e-01 1.62758678e-01 6.94973052e-01 -8.30410600e-01
-1.54935014e+00 -1.74535319e-01 -2.18640305e-02 -5.81938446e-01
-1.80766180e-01 5.09685397e-01 -1.72196224e-01 -1.15157358e-01
4.88124669e-01 2.35629722e-01 -2.68569767e-01 -6.34569943e-01
5.83802581e-01 6.30373240e-01 3.19375128e-01 -1.17604911e+00
8.53870451e-01 -6.57086670e-02 -4.51853395e-01 -2.94225812e-01
-1.24158525e+00 -2.35918656e-01 -6.65557504e-01 -1.87340781e-01
2.80923039e-01 -1.24417996e+00 -6.78947330e-01 5.11273980e-01
-5.92342257e-01 -6.01717293e-01 -3.84308428e-01 5.86559296e-01
7.21884593e-02 -1.32677168e-01 -3.47166210e-01 -6.78344131e-01
-4.91023660e-01 -1.21586668e+00 1.10971428e-01 1.07594192e+00
3.58341515e-01 -1.28263032e+00 1.39238790e-01 6.59461439e-01
5.48840940e-01 -2.57785439e-01 5.66011369e-01 -1.10376930e+00
-5.14166117e-01 2.27140605e-01 9.66075212e-02 4.34037447e-01
2.98202097e-01 2.65789777e-01 -1.15035510e+00 -2.01473683e-01
-2.77188659e-01 -8.24423075e-01 1.19310904e+00 2.15099931e-01
1.40420783e+00 -3.58980507e-01 -2.84700751e-01 6.01672590e-01
1.23510361e+00 -8.09867829e-02 -1.89256892e-01 2.40984499e-01
1.00188720e+00 4.14931148e-01 2.22338468e-01 3.48415375e-01
8.32336247e-01 2.85672337e-01 3.09566379e-01 1.51251644e-01
-2.55053490e-01 -3.20580095e-01 4.70619142e-01 1.29080796e+00
-3.27353090e-01 -1.40703201e-01 -9.52575505e-01 6.68829501e-01
-1.80115879e+00 -8.64633679e-01 1.64903224e-01 1.82336724e+00
1.52291167e+00 -1.58106625e-01 9.90547836e-02 -2.47150674e-01
5.07247746e-01 2.31414940e-02 -7.21803308e-01 -1.68680981e-01
1.13605864e-01 3.05736005e-01 2.15459809e-01 4.29207951e-01
-7.79504836e-01 9.22700405e-01 4.99830675e+00 1.21316135e+00
-1.02679396e+00 2.48429865e-01 6.75815761e-01 -2.13901922e-01
-6.42498136e-01 -1.41886026e-01 -1.28956211e+00 3.55828881e-01
1.05243027e+00 -3.69131714e-01 2.42260575e-01 8.42163444e-01
2.75830012e-02 -2.70121358e-02 -5.61471760e-01 4.23180729e-01
-1.65338829e-01 -1.03893244e+00 -3.14334594e-02 -2.11395219e-01
1.27012432e+00 1.70023024e-01 3.14671427e-01 7.38935471e-01
1.15759599e+00 -8.41654360e-01 1.01288119e-02 7.05170691e-01
3.97422135e-01 -9.02392268e-01 4.13272887e-01 6.16597831e-01
-1.05856061e+00 -3.70663881e-01 -3.82076710e-01 6.68582097e-02
-3.59972417e-01 7.09320664e-01 -9.96645033e-01 7.72160590e-01
5.06121516e-01 6.90773904e-01 -6.85764372e-01 9.48778927e-01
-8.46700549e-01 1.28661299e+00 -1.95239365e-01 7.10834339e-02
1.14576355e-01 -2.03015536e-01 2.82714605e-01 1.01275063e+00
3.78051132e-01 6.07679844e-01 5.24664581e-01 5.46676099e-01
-4.49192017e-01 2.08873507e-02 -2.11462274e-01 1.09500587e-02
8.85289907e-01 1.64804494e+00 -8.63983706e-02 -6.41458452e-01
-5.74104071e-01 4.79191571e-01 8.13669980e-01 5.05091906e-01
-4.85586703e-01 -1.64260998e-01 5.46709061e-01 -2.30869338e-01
2.14300781e-01 1.18993133e-01 -2.45300040e-01 -1.24296820e+00
-3.32891524e-01 -8.97651732e-01 7.40411401e-01 -5.94879329e-01
-1.24980593e+00 3.37169915e-01 6.34257197e-02 -1.00358880e+00
2.89387796e-02 -2.75312901e-01 -1.09710574e+00 9.89472389e-01
-1.88486290e+00 -1.09794915e+00 -4.80508983e-01 6.16367817e-01
4.12413597e-01 -6.21848702e-01 5.48459232e-01 9.40992311e-02
-1.09951568e+00 8.05318296e-01 1.80182517e-01 1.37043148e-01
8.12219083e-01 -1.54537797e+00 -2.99271345e-01 4.42883849e-01
8.52126330e-02 5.14277756e-01 3.17417741e-01 -4.59041059e-01
-1.35043919e+00 -1.17251062e+00 3.93028140e-01 -5.56957006e-01
5.96589923e-01 4.64049965e-01 -1.30094159e+00 5.97604871e-01
7.26440310e-01 -4.15371023e-02 1.21476185e+00 3.76633853e-01
-4.93696779e-01 -4.53514248e-01 -8.95795166e-01 3.57817262e-01
4.82237548e-01 -1.57281935e-01 -8.90749335e-01 1.01455286e-01
7.53574848e-01 -4.74393755e-01 -1.24780655e+00 5.96626878e-01
4.46393847e-01 -7.40793228e-01 8.75892758e-01 -7.76996493e-01
4.25737649e-01 -2.40518630e-01 4.71041143e-01 -1.87904894e+00
-2.94043362e-01 -3.16614002e-01 -6.12980843e-01 1.44644237e+00
3.05572093e-01 -5.13216794e-01 7.07225323e-01 6.40294075e-01
-3.20116460e-01 -1.32757115e+00 -3.66126120e-01 -4.44229990e-01
3.15054417e-01 -4.01147977e-02 9.43153501e-01 1.45423770e+00
-5.44941351e-02 4.57330257e-01 -1.57575175e-01 1.87808111e-01
5.07192016e-01 3.52664351e-01 8.34477782e-01 -1.36971211e+00
-2.82840639e-01 -6.68534636e-01 2.45985374e-01 -9.75680888e-01
3.78868818e-01 -1.04113543e+00 -4.04347092e-01 -1.35069489e+00
3.98909390e-01 -8.02016914e-01 -1.00195062e+00 9.90704834e-01
-8.89140069e-01 -1.27195790e-01 -2.46316679e-02 -2.06683502e-01
-8.88438940e-01 7.89735079e-01 1.62463260e+00 -3.03989220e-02
-3.51057678e-01 2.10483104e-01 -1.19360995e+00 8.77426803e-01
1.20148325e+00 -6.36153400e-01 -8.33786368e-01 -7.09633887e-01
3.18060741e-02 -2.00335383e-01 5.16969264e-02 -6.68101549e-01
6.04366302e-01 -5.97943485e-01 6.26194358e-01 -2.86201239e-01
1.03414096e-01 -8.29715490e-01 -2.44324341e-01 3.16007972e-01
-4.41517562e-01 -1.50224105e-01 2.34152734e-01 2.83850074e-01
-9.10586491e-02 -4.39296693e-01 7.56909370e-01 -1.77203983e-01
-6.46827579e-01 5.18494487e-01 1.10762686e-01 1.79220602e-01
9.45723653e-01 1.33915752e-01 -6.07785225e-01 -1.57507896e-01
-5.43452501e-01 9.47233796e-01 1.32847484e-02 3.96646976e-01
6.93722248e-01 -1.22776866e+00 -1.04282391e+00 2.63302952e-01
-1.43107325e-01 4.60811168e-01 4.31951493e-01 8.35486472e-01
1.87197179e-01 -2.56316923e-02 -1.05745502e-01 -2.31020525e-01
-1.35059214e+00 1.52319327e-01 3.90615940e-01 -3.98792446e-01
-2.18951821e-01 1.29945898e+00 -2.55439593e-03 -8.31179500e-01
3.50166976e-01 -8.84146020e-02 -5.26297390e-01 2.09816784e-01
8.31711411e-01 5.32037318e-01 -1.96069345e-01 -2.43126526e-01
9.36494321e-02 2.28320524e-01 -4.60841596e-01 2.20509648e-01
1.45589590e+00 2.88322233e-02 2.53668457e-01 2.27021992e-01
5.89105964e-01 2.15936080e-01 -1.41243100e+00 -7.73478508e-01
-1.61683992e-01 -1.31584838e-01 3.02105546e-01 -1.12430799e+00
-1.48627520e+00 8.88107002e-01 2.90938526e-01 -1.20229296e-01
1.41568303e+00 -3.29608679e-01 5.53392112e-01 5.69544017e-01
-9.15501490e-02 -1.01348555e+00 1.94870025e-01 5.75510740e-01
4.17203099e-01 -1.40282941e+00 1.79939419e-01 -4.52662222e-02
-5.90675771e-01 1.11852658e+00 1.50745714e+00 1.22402377e-01
7.87922978e-01 1.87074006e-01 2.27416873e-01 2.45563057e-03
-1.24148524e+00 -8.67391229e-02 6.20359898e-01 8.49551111e-02
6.92052960e-01 2.73325533e-01 -6.27307817e-02 1.08609080e+00
3.97467092e-02 -8.12899172e-02 5.10994315e-01 9.03436124e-01
-9.55575049e-01 -1.33599627e+00 -1.16018139e-01 5.57736754e-01
-3.28516871e-01 -2.50383317e-01 -2.63491511e-01 2.93563426e-01
3.96544814e-01 9.11990166e-01 -3.66814256e-01 -5.61718762e-01
1.08257338e-01 2.71128386e-01 4.93645012e-01 -8.38510633e-01
-1.24770713e+00 -7.89341331e-02 -2.62746245e-01 -1.03914827e-01
-6.31574392e-01 -9.15107876e-02 -1.28227985e+00 -1.54347166e-01
-6.11068368e-01 7.00680494e-01 1.55996561e-01 8.63699019e-01
4.28897023e-01 9.66799140e-01 7.77148902e-01 -2.40605086e-01
-7.72738934e-01 -1.06312442e+00 -4.04392213e-01 -2.20613740e-02
3.24912697e-01 -6.37668312e-01 -3.32973301e-01 -6.12760447e-02] | [9.516715049743652, 3.379037618637085] |
5717e949-f007-489d-bc5c-a85de20a6040 | towards-complex-artificial-life | 1805.06366 | null | http://arxiv.org/abs/1805.06366v1 | http://arxiv.org/pdf/1805.06366v1.pdf | Towards Complex Artificial Life | An object-oriented combinator chemistry was used to construct an artificial
organism with a system architecture possessing characteristics necessary for
organisms to evolve into more complex forms. This architecture supports
modularity by providing a mechanism for the construction of executable modules
called $methods$ that can be duplicated and specialized to increase complexity.
At the same time, its support for concurrency provides the flexibility in
execution order necessary for redundancy, degeneracy and parallelism to
mitigate increased replication costs. The organism is a moving,
self-replicating, spatially distributed assembly of elemental combinators
called a $roving \: pile.$ The pile hosts an asynchronous message passing
computation implemented by parallel subprocesses encoded by genes distributed
through out the pile like the plasmids of a bacterial cell. | ['Lance R. Williams'] | 2018-05-16 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [-1.63805291e-01 1.14573650e-01 3.79852355e-01 2.75638044e-01
5.72842896e-01 -6.45123839e-01 6.70759737e-01 9.81724113e-02
-2.24926963e-01 5.69865644e-01 -4.43098575e-01 -3.14573824e-01
-3.39339375e-01 -1.19349253e+00 -3.62582356e-01 -8.64686728e-01
-7.31978655e-01 3.21643353e-01 5.90012014e-01 -2.90846407e-01
3.25633198e-01 5.07503748e-01 -1.93343484e+00 2.94852108e-01
5.82187414e-01 4.31667119e-01 9.25726712e-01 8.21565688e-01
-3.09173584e-01 9.06314611e-01 -5.55047154e-01 -4.22009975e-02
1.65374756e-01 -6.16614223e-01 -3.74598771e-01 3.53542805e-01
-7.97398269e-01 2.11871415e-01 -5.58771640e-02 4.50487435e-01
-9.76171903e-03 -1.48382828e-01 6.37459457e-01 -1.34886038e+00
-2.13822052e-01 7.55124912e-02 -2.52366811e-02 -4.30820525e-01
4.37297970e-01 4.61805165e-01 6.00601077e-01 -8.01434994e-01
8.56443048e-01 1.12434411e+00 6.66467428e-01 4.59661514e-01
-1.40920413e+00 7.70831928e-02 -4.35281634e-01 -6.38943791e-01
-1.64621866e+00 -5.07812917e-01 -2.06712075e-02 -5.20074069e-01
1.56138444e+00 8.17727923e-01 1.23083234e+00 2.94205546e-01
9.91899550e-01 4.16506752e-02 8.14911008e-01 -3.87890399e-01
6.81428313e-01 -5.77487946e-02 -4.49185789e-01 1.05906796e+00
8.99699807e-01 -2.67865121e-01 -5.05548954e-01 -7.56753504e-01
1.10566342e+00 1.13546019e-02 -7.26858228e-02 -5.72680533e-01
-1.28241467e+00 3.23956668e-01 -6.81766123e-02 4.94963825e-01
-3.50454003e-01 6.21535718e-01 4.26829576e-01 4.69108164e-01
-1.95492446e-01 6.85971856e-01 -6.58949494e-01 -2.51049429e-01
8.13655555e-02 3.48454624e-01 1.39884818e+00 8.64344478e-01
6.69559717e-01 2.58512974e-01 5.63209832e-01 4.52936053e-01
5.55679023e-01 2.12904841e-01 5.75688601e-01 -1.13488042e+00
-2.58358330e-01 1.28082597e+00 1.12147644e-01 -9.34032261e-01
-4.94526476e-01 -3.97015572e-01 -7.25062907e-01 5.57687104e-01
1.65170029e-01 1.36716843e-01 -2.33963430e-01 1.41271019e+00
3.58217269e-01 -5.05284548e-01 8.95363167e-02 2.38383144e-01
4.68254872e-02 8.66903305e-01 1.26860842e-01 -1.62587523e-01
1.44717205e+00 -5.66341639e-01 -3.09789211e-01 3.06610852e-01
7.97759712e-01 -6.86179042e-01 6.71841979e-01 3.24696749e-01
-1.42705834e+00 -6.00017719e-02 -1.08917356e+00 5.38393855e-01
-4.94145215e-01 -2.21515372e-01 1.09675801e+00 7.81379700e-01
-1.38484919e+00 5.14752567e-01 -9.20105636e-01 -5.54803848e-01
-6.33815378e-02 3.85919333e-01 -3.83914858e-01 4.85434443e-01
-3.18969399e-01 8.73051703e-01 4.92112875e-01 -2.37106770e-01
-1.14939630e+00 -1.81776777e-01 -5.03142118e-01 1.88059792e-01
-2.03597490e-02 -1.43475640e+00 8.31506431e-01 -1.01592553e+00
-1.63805377e+00 8.50726485e-01 5.10355318e-03 -1.08241081e-01
1.52431980e-01 6.65403008e-01 1.10633783e-01 2.03314573e-02
8.38627517e-02 3.10711682e-01 4.95374918e-01 -1.30591881e+00
-6.69336736e-01 -2.51665235e-01 -1.48492068e-01 1.59422129e-01
6.10484481e-02 1.77498311e-01 -7.87591003e-03 -6.89848423e-01
3.92069936e-01 -7.47560322e-01 -4.22737241e-01 -8.23766142e-02
6.05708398e-02 -2.46277690e-01 6.12049639e-01 -2.78988808e-01
7.90261149e-01 -2.21697593e+00 3.93340528e-01 1.95126072e-01
3.87234062e-01 -2.65256733e-01 1.10118471e-01 1.28162694e+00
2.26335734e-01 2.90106118e-01 -2.54868388e-01 3.67512405e-01
5.39316088e-02 2.49761716e-01 3.65966678e-01 3.29678476e-01
2.78397799e-01 4.84114349e-01 -7.25085020e-01 -5.96582144e-02
-5.48816919e-01 -2.23814975e-02 -6.35349214e-01 2.08750650e-01
-5.29118717e-01 1.50108606e-01 -3.42376351e-01 7.72338331e-01
2.25810483e-01 -3.25427324e-01 6.84865296e-01 7.58172750e-01
-8.33228707e-01 1.99887499e-01 -1.27946556e+00 1.30917466e+00
-1.95459992e-01 2.84848303e-01 7.45023131e-01 -7.87040532e-01
8.15627396e-01 5.01732409e-01 4.95643765e-01 -8.28986615e-02
1.13726214e-01 5.01285315e-01 3.23198795e-01 -4.10279393e-01
5.02496302e-01 -1.04655527e-01 -1.56852797e-01 7.14343011e-01
-2.41168275e-01 -6.17265165e-01 3.08983505e-01 2.94969290e-01
1.61946630e+00 1.96788415e-01 5.38818479e-01 -8.26189697e-01
3.65679175e-01 3.77666533e-01 8.04814696e-01 7.44044125e-01
1.48410305e-01 -1.92433760e-01 4.59644586e-01 -5.10636866e-01
-1.58949912e+00 -1.26583743e+00 -6.13638610e-02 1.01662481e+00
7.87239671e-02 -8.33600581e-01 -7.03507125e-01 3.55713844e-01
-4.26223874e-02 1.67323560e-01 -5.36708832e-02 1.89526081e-01
-4.24639493e-01 -9.39247012e-01 5.82036614e-01 -1.89762145e-01
4.85075444e-01 -9.18089867e-01 -1.22603655e+00 6.83263242e-01
3.03933710e-01 -2.12146461e-01 2.10253179e-01 4.00111377e-01
-9.01063144e-01 -7.68919945e-01 -1.99973062e-01 -9.22900558e-01
8.25104713e-01 3.82144034e-01 9.53431726e-01 7.35402465e-01
-8.06294978e-01 6.01072729e-01 -3.60591598e-02 -5.00099838e-01
-7.51063347e-01 -2.70595551e-01 2.33749077e-02 -3.98093641e-01
-6.20699465e-01 -8.86886835e-01 -4.40034539e-01 3.50642234e-01
-1.26217508e+00 3.00920486e-01 2.62606114e-01 9.18773830e-01
1.12648465e-01 5.03135085e-01 4.30632144e-01 -3.94749790e-02
5.17381251e-01 -4.90900129e-01 -5.99292219e-01 1.15419574e-01
-1.84427425e-01 -1.30725831e-01 6.20182574e-01 1.03423875e-02
-1.02796471e+00 3.98831964e-02 2.17809245e-01 9.03319478e-01
8.98909755e-03 3.50729644e-01 -3.58119786e-01 -3.06188196e-01
4.21166450e-01 6.52758121e-01 4.81191128e-01 -1.83884159e-01
1.22461133e-01 5.99371314e-01 1.62342191e-02 -8.16696882e-01
3.39159310e-01 5.53651571e-01 2.48435810e-01 -1.26680529e+00
9.23108637e-01 -2.21589338e-02 -2.12010011e-01 -2.37579286e-01
5.15739679e-01 -5.86390793e-01 -9.89372253e-01 8.72307241e-01
-1.20895886e+00 -3.49257439e-01 -5.34760833e-01 -4.59418222e-02
-8.50606322e-01 1.69916645e-01 -7.37337708e-01 -9.08829689e-01
-3.05028036e-02 -8.56144547e-01 5.24332464e-01 1.18797854e-01
-4.39093083e-01 -7.23220706e-01 4.46048677e-01 -1.62917569e-01
8.01181138e-01 6.44584447e-02 1.17459166e+00 -2.53698081e-01
-1.18595183e+00 1.98590271e-02 2.03418061e-01 -2.99697936e-01
1.04182139e-01 3.33236456e-01 -3.20177585e-01 -1.60388380e-01
1.34699672e-01 1.13624021e-01 2.03169480e-01 -2.76862770e-01
4.99885559e-01 -4.73331183e-01 -6.29616499e-01 3.00003171e-01
1.41986716e+00 7.18181431e-01 9.04019892e-01 6.21574819e-01
-6.12139292e-02 6.69141591e-01 -1.35480016e-01 6.82212532e-01
3.80067706e-01 2.83545136e-01 3.44088018e-01 3.71639013e-01
1.97891131e-01 3.93937588e-01 2.66603917e-01 1.00773370e+00
-3.84423196e-01 -2.21297175e-01 -1.18635499e+00 4.86653894e-01
-1.70723355e+00 -1.08988702e+00 -2.77395278e-01 2.22332716e+00
9.00771379e-01 -2.26195395e-01 9.65165198e-02 1.58814833e-01
5.41866243e-01 -5.62517762e-01 5.98870292e-02 -7.55309522e-01
-1.82222649e-01 9.08139274e-02 3.34433317e-01 8.37377012e-02
-6.18719101e-01 6.02986157e-01 7.60821533e+00 4.95574623e-01
-7.99726248e-01 9.21268202e-03 3.81684899e-01 1.64116144e-01
-4.08316374e-01 3.99977177e-01 -4.55011755e-01 4.02635694e-01
9.75838661e-01 -3.84245366e-01 6.47389114e-01 5.40781081e-01
3.66095275e-01 -6.29482925e-01 -6.03689313e-01 4.70151812e-01
-3.59786004e-01 -1.76980174e+00 -4.72714603e-02 3.25102925e-01
6.48795366e-01 -1.15897246e-01 -5.03844261e-01 -2.02941000e-01
7.34116137e-01 -7.55381942e-01 8.12727392e-01 6.43360913e-01
2.12276101e-01 -4.15273905e-01 2.01854169e-01 7.58077741e-01
-1.19505703e+00 -3.23735327e-01 -1.22853242e-01 -7.85924911e-01
1.28269091e-01 2.71027595e-01 -7.86335230e-01 4.45672244e-01
5.65369368e-01 -2.66971916e-01 -2.16887459e-01 1.22943842e+00
3.68183345e-01 3.99920195e-02 -2.93885440e-01 -4.65386480e-01
1.47640379e-03 -5.38013518e-01 9.06901002e-01 1.16805530e+00
4.27104115e-01 2.23190576e-01 -5.98466694e-02 6.93199217e-01
3.71592283e-01 1.18636690e-01 -1.03415811e+00 -2.14563817e-01
4.44150627e-01 1.14948237e+00 -1.13574386e+00 -4.96051788e-01
-1.56107187e-01 6.60602450e-01 3.25696617e-02 4.49907556e-02
-3.52968723e-01 -6.65337265e-01 7.00216234e-01 4.98895049e-01
3.19021314e-01 -1.10525262e+00 -4.21646535e-01 -8.76791120e-01
-3.13451856e-01 -8.70309830e-01 -6.89839870e-02 -6.63827360e-01
-1.00873852e+00 1.54267788e-01 -3.78381521e-01 -5.92130184e-01
-1.96624383e-01 -4.89678890e-01 -6.37670517e-01 8.03411663e-01
-2.02272043e-01 -7.28592277e-01 9.95323062e-03 5.04645221e-02
2.61166066e-01 -4.15822208e-01 1.15528679e+00 -4.20605898e-01
-4.94025409e-01 -2.90042460e-01 5.37476659e-01 -4.14038211e-01
-1.24141984e-01 -8.71978641e-01 1.98555052e-01 6.06617391e-01
-5.28059840e-01 1.17387617e+00 4.91983652e-01 -9.13878739e-01
-2.18611503e+00 -7.74753749e-01 9.81639802e-01 -1.81470990e-01
6.50106668e-01 -6.12361968e-01 -4.99908328e-01 4.31140065e-01
4.30386275e-01 -6.89249694e-01 6.80262864e-01 -4.00145173e-01
5.51115051e-02 -1.21858910e-01 -1.37802708e+00 1.01319170e+00
1.09975362e+00 -9.27402526e-02 -3.88736308e-01 8.76304135e-02
6.09919906e-01 5.72705790e-02 -8.94753456e-01 -8.98817405e-02
6.95379853e-01 -8.87751102e-01 9.20334995e-01 -1.28283024e-01
1.39934540e-01 -6.97784066e-01 -7.46397823e-02 -9.33624029e-01
-6.04146779e-01 -1.10786116e+00 5.42078018e-01 1.00458574e+00
1.96443155e-01 -1.30659258e+00 3.57970357e-01 6.75106227e-01
-2.22513750e-01 -2.92717725e-01 -1.09824586e+00 -9.97974753e-01
-3.75053525e-01 1.12256430e-01 6.46037638e-01 9.00679708e-01
6.74993753e-01 6.15450442e-02 3.68321300e-01 -1.84431404e-01
5.52521884e-01 -2.07222909e-01 8.14855993e-01 -9.85126972e-01
-6.14927709e-01 -6.46728933e-01 -7.58054018e-01 -6.46003306e-01
-5.05838752e-01 -9.41012979e-01 6.77273273e-02 -1.22358501e+00
2.91039914e-01 -8.72120142e-01 4.12187755e-01 2.79503763e-01
5.23611724e-01 -1.55430391e-01 3.20046663e-01 4.86027241e-01
-3.91835213e-01 2.68726379e-01 9.50535059e-01 2.14734539e-01
-2.59792000e-01 -5.10617375e-01 -4.65619028e-01 8.43778193e-01
9.95945036e-01 -5.77560723e-01 -2.01420039e-01 -2.06951305e-01
6.88620389e-01 3.74259412e-01 4.08796877e-01 -1.02045071e+00
2.82886356e-01 -4.62234080e-01 5.03961034e-02 -1.66538402e-01
2.90762454e-01 -7.94120669e-01 1.15094125e+00 1.20961452e+00
2.29980260e-01 6.12852514e-01 9.30258259e-02 4.75334853e-01
2.41789937e-01 -4.39771622e-01 4.05337542e-01 -7.95201480e-01
-2.50595093e-01 -2.65634775e-01 -1.60131645e+00 -5.08566558e-01
1.61717522e+00 -3.40566188e-01 -3.64756823e-01 1.13643043e-01
-5.02156734e-01 -2.33940333e-01 1.30794203e+00 -1.61488593e-01
3.94923896e-01 -8.32197607e-01 -3.42683822e-01 2.42257729e-01
-2.29438081e-01 -2.39558965e-01 -3.28151315e-01 5.07409990e-01
-1.50117254e+00 4.71797884e-01 -5.76025963e-01 -6.18934274e-01
-9.94029343e-01 2.27945089e-01 4.26056564e-01 7.77994171e-02
-4.35432792e-01 6.83865786e-01 1.63099766e-01 -4.19637114e-01
-3.69519144e-01 -1.71779707e-01 3.94496530e-01 -4.81585979e-01
5.37298858e-01 5.36139071e-01 -6.38486519e-02 -1.84364513e-01
-3.86177212e-01 -6.04036078e-02 6.99715078e-01 -4.07296121e-01
1.37815571e+00 -8.38439986e-02 -1.34499264e+00 5.28855741e-01
3.50916445e-01 8.18253867e-03 -6.51451349e-01 6.76301301e-01
1.50441512e-01 -5.08913577e-01 -7.35919356e-01 -4.06843483e-01
-1.26342326e-01 -8.16468596e-02 -4.20722067e-02 8.36095333e-01
7.32856154e-01 -1.86496209e-02 1.36817440e-01 6.30214512e-01
8.80829394e-01 -9.66363907e-01 1.09275006e-01 5.90492785e-01
7.61802912e-01 1.36548271e-02 -1.09435230e-01 -5.25264800e-01
-1.39925584e-01 1.06825209e+00 2.20099226e-01 -2.32941955e-01
4.77230281e-01 8.45115185e-01 -6.00708961e-01 -4.49648559e-01
-1.39774239e+00 2.47198492e-01 -7.70414889e-01 8.07673812e-01
5.61572075e-01 2.49597520e-01 -1.08611846e+00 4.00376678e-01
1.55813903e-01 -3.18125561e-02 6.46553278e-01 1.72982204e+00
-1.09399951e+00 -1.45923555e+00 -8.09285700e-01 2.11519033e-01
-1.01752050e-01 6.92568272e-02 -2.94240415e-01 7.76443601e-01
6.03706181e-01 7.68229365e-01 4.57356483e-01 6.40232936e-02
-1.03816934e-01 2.97982663e-01 5.94779730e-01 -5.58678389e-01
-1.00597513e+00 -4.84873578e-02 4.94789451e-01 -2.91196793e-01
4.26852740e-02 -1.04740405e+00 -1.45602369e+00 -5.84896743e-01
1.13048919e-01 3.08993846e-01 1.03819180e+00 4.74260420e-01
6.52800441e-01 4.80220169e-01 6.43901050e-01 -6.64142132e-01
-3.30845475e-01 -3.97484809e-01 -8.04068387e-01 -1.54939517e-01
-4.64046627e-01 -2.57943243e-01 -9.77814570e-02 4.22138929e-01] | [5.614199638366699, 4.185136795043945] |
a0a6ac45-f3f6-4956-8de2-d48f28fdc897 | basn-learning-steganography-with-binary | 1907.04362 | null | https://arxiv.org/abs/1907.04362v1 | https://arxiv.org/pdf/1907.04362v1.pdf | BASN -- Learning Steganography with Binary Attention Mechanism | Secret information sharing through image carrier has aroused much research attention in recent years with images' growing domination on the Internet and mobile applications. However, with the booming trend of convolutional neural networks, image steganography is facing a more significant challenge from neural-network-automated tasks. To improve the security of image steganography and minimize task result distortion, models must maintain the feature maps generated by task-specific networks being irrelative to any hidden information embedded in the carrier. This paper introduces a binary attention mechanism into image steganography to help alleviate the security issue, and in the meanwhile, increase embedding payload capacity. The experimental results show that our method has the advantage of high payload capacity with little feature map distortion and still resist detection by state-of-the-art image steganalysis algorithms. | ['Yang Yang'] | 2019-07-09 | null | null | null | null | ['steganalysis', 'image-steganography'] | ['computer-vision', 'computer-vision'] | [ 8.32077205e-01 2.48137355e-01 -1.66563615e-01 1.22823484e-01
2.49443009e-01 -7.91766271e-02 2.86998063e-01 -6.73239291e-01
-3.30044955e-01 3.12227398e-01 -8.95808712e-02 -4.87919182e-01
1.37514725e-01 -8.72116923e-01 -4.61129606e-01 -8.87607515e-01
-1.91205531e-01 -4.77020502e-01 3.82751733e-01 -4.20975238e-01
5.19702435e-01 6.62262663e-02 -1.34242618e+00 1.54356882e-01
8.12302291e-01 1.14420021e+00 4.66265798e-01 3.32605630e-01
6.16647378e-02 9.47009444e-01 -4.82504010e-01 -4.64063078e-01
4.24954832e-01 -8.00391853e-01 -3.87364864e-01 2.62209594e-01
-1.93873674e-01 -2.36404881e-01 -1.03158951e+00 1.76509559e+00
4.16399866e-01 -5.05179107e-01 4.03997190e-02 -1.35009491e+00
-9.66335416e-01 6.58701360e-01 -6.45175040e-01 1.70893624e-01
-2.20429182e-01 5.17446101e-01 4.08604711e-01 -4.19980466e-01
5.30456066e-01 1.14787054e+00 4.60233212e-01 6.65011585e-01
-7.37742245e-01 -1.14535296e+00 -2.33097121e-01 6.10673666e-01
-1.35468996e+00 -4.63518351e-01 9.80632484e-01 6.76195994e-02
6.55108511e-01 3.87975305e-01 9.26795125e-01 5.90914905e-01
5.06525874e-01 4.47056413e-01 9.16498780e-01 -4.02454734e-01
-3.05182129e-01 2.47320175e-01 -7.23984659e-01 9.80447531e-01
6.53231919e-01 1.63055912e-01 -6.02412857e-02 3.45479250e-01
9.93200183e-01 2.44770095e-01 -6.59701884e-01 -3.76746446e-01
-1.27559400e+00 8.26969743e-01 5.76650798e-01 5.19439697e-01
-1.77732646e-01 4.48732346e-01 1.52649790e-01 5.01919329e-01
2.56559730e-01 2.91489214e-01 -9.02038887e-02 2.73570836e-01
-5.84041715e-01 -1.95645139e-01 6.64005280e-01 8.07164550e-01
6.56003773e-01 2.96211898e-01 5.01266956e-01 2.59241551e-01
5.23137748e-01 6.22855246e-01 7.59517670e-01 -6.33668602e-01
5.83563209e-01 6.77242875e-01 -4.57548290e-01 -1.76711810e+00
1.80861130e-01 -6.07547998e-01 -1.26503897e+00 2.02749133e-01
1.12893127e-01 4.57378328e-02 -8.43183935e-01 1.36802673e+00
-1.51882218e-02 1.40424699e-01 1.89060435e-01 7.19735384e-01
4.76648360e-01 7.81507373e-01 -2.60658532e-01 -2.59086788e-01
1.42659879e+00 -7.63102770e-01 -8.92869174e-01 -5.35926700e-01
6.19850695e-01 -7.84272611e-01 3.72456759e-01 3.75390053e-02
-8.19542050e-01 -4.99037236e-01 -1.57432520e+00 3.95360827e-01
-1.54761761e-01 -4.80645508e-01 5.16331494e-01 1.18260574e+00
-7.98077881e-01 2.07430482e-01 -4.44994628e-01 -2.36865375e-02
7.45175183e-01 6.88894093e-01 -4.00271863e-01 -2.39271238e-01
-1.56679964e+00 7.39283621e-01 8.98101807e-01 1.53860331e-01
-6.82649732e-01 -5.77049963e-02 -9.89227414e-01 2.12147355e-01
3.56994987e-01 -1.23478763e-01 6.03847980e-01 -1.33756411e+00
-1.11675572e+00 7.17487395e-01 3.77243787e-01 -5.52489877e-01
5.53904057e-01 5.53701401e-01 -7.66438246e-01 2.70962507e-01
-2.58578211e-01 7.69222081e-01 1.18775654e+00 -1.02879190e+00
-6.64638996e-01 -9.43757519e-02 -4.47739661e-01 -4.68189083e-02
-7.63528228e-01 -6.30096421e-02 -4.71268803e-01 -5.59867799e-01
4.06505734e-01 -1.16436708e+00 -3.32446009e-01 1.78016439e-01
-3.71660084e-01 2.99913257e-01 1.56677008e+00 -7.18658149e-01
1.24099457e+00 -2.38562775e+00 -1.43269464e-01 3.99502039e-01
3.35823983e-01 8.95972669e-01 -1.04517065e-01 3.19091290e-01
-1.70158610e-01 5.04186749e-01 -4.15217787e-01 3.13364327e-01
-5.30314326e-01 1.22144714e-01 -2.07444727e-01 8.20875943e-01
-8.58043581e-02 1.12303150e+00 -7.95708537e-01 -6.59369051e-01
2.65758544e-01 5.35888731e-01 -3.62169832e-01 -1.11755833e-01
1.33731514e-01 3.31004530e-01 -5.37695825e-01 3.87613952e-01
8.07787895e-01 -5.93571126e-01 3.72637033e-01 -5.43633364e-02
5.41717745e-02 -2.39914224e-01 -7.75472045e-01 1.11981440e+00
-8.43733922e-02 9.99865592e-01 -1.20425753e-01 -1.06163919e+00
9.16715860e-01 3.42255205e-01 3.87463212e-01 -7.93265641e-01
5.60228169e-01 3.96075249e-01 6.12926364e-01 -7.78552234e-01
2.99532086e-01 1.24158949e-01 6.03106804e-02 4.55686271e-01
-3.90248686e-01 1.48794845e-01 -3.29185098e-01 -4.10342626e-02
8.37068141e-01 -2.96945363e-01 3.01368445e-01 -6.02658875e-02
7.25989521e-01 -1.28004834e-01 4.39816564e-01 4.35886800e-01
-1.73255578e-01 2.51645386e-01 2.47030228e-01 -3.25515062e-01
-1.38258970e+00 -1.07689515e-01 1.99059084e-01 2.92103976e-01
6.58968568e-01 1.13658488e-01 -9.20505583e-01 -5.96795797e-01
-2.97554463e-01 2.38252893e-01 -3.92012209e-01 -6.21626735e-01
-7.89750636e-01 -4.59341615e-01 9.04520214e-01 -2.89257228e-01
1.62099600e+00 -1.32716274e+00 -7.17901468e-01 2.58133858e-01
-2.71237314e-01 -1.19235063e+00 -5.70632219e-01 -4.14426744e-01
-6.96819901e-01 -1.12289715e+00 -1.04717112e+00 -1.31415176e+00
1.01029968e+00 7.53993213e-01 3.26109558e-01 7.37121463e-01
-1.92040682e-01 -2.51431853e-01 -4.86324549e-01 -4.11236405e-01
-8.96722376e-01 1.06168181e-01 -4.00367379e-01 3.05706501e-01
1.99022770e-01 -4.42686200e-01 -8.18165421e-01 5.76278031e-01
-1.34405887e+00 4.51945305e-01 9.71345305e-01 7.57115304e-01
-1.31135099e-02 7.74476588e-01 3.17020565e-01 -6.39203787e-01
2.58332789e-01 -4.55085307e-01 -5.94864368e-01 1.28400087e-01
-9.71359551e-01 -7.02115744e-02 4.28418368e-01 -5.61204791e-01
-7.02789366e-01 -2.70353258e-01 1.32121280e-01 -3.31490129e-01
4.19139266e-01 4.36083794e-01 -4.43503141e-01 -8.16555619e-01
3.34725559e-01 8.91278565e-01 8.37042868e-01 3.46135460e-02
-1.63701534e-01 8.43151391e-01 4.37175304e-01 6.65597320e-01
1.19330037e+00 5.86122632e-01 1.82826266e-01 -7.32905149e-01
-1.33886725e-01 5.82119823e-02 8.85343328e-02 -2.58307815e-01
8.04331243e-01 -6.72828138e-01 -8.18850040e-01 1.03547001e+00
-1.14955842e+00 2.65000075e-01 3.19584101e-01 2.33135596e-01
-1.59837514e-01 8.30666423e-01 -3.52303863e-01 -5.70684731e-01
-3.32111299e-01 -1.31026137e+00 8.55449289e-02 1.18670188e-01
4.26914901e-01 -9.39596236e-01 -5.18371105e-01 2.44278505e-01
6.56457305e-01 3.11839968e-01 7.98640311e-01 -4.32738066e-01
-1.14497006e+00 -5.71909070e-01 -5.36059022e-01 6.18238151e-01
2.70993322e-01 -7.43968189e-01 -5.83607435e-01 -4.76423562e-01
4.18872654e-01 1.53976664e-01 9.86581266e-01 6.94806455e-03
1.27495492e+00 -9.18105900e-01 -4.00729626e-01 8.22802067e-01
1.54162741e+00 5.63572168e-01 1.35597050e+00 5.96068263e-01
7.55669236e-01 7.22522795e-01 2.46847361e-01 9.22725871e-02
9.88830626e-02 2.20343962e-01 9.45126295e-01 -2.06137806e-01
-1.54464468e-01 -3.85889858e-01 2.03659564e-01 8.07927489e-01
-3.37429978e-02 -7.12132156e-01 -5.89713931e-01 4.88212138e-01
-1.67821598e+00 -1.13879633e+00 5.18347733e-02 1.78573871e+00
6.66869462e-01 2.45054930e-01 -6.23288572e-01 4.93205607e-01
1.33836067e+00 5.94299436e-01 -4.96722400e-01 4.97974493e-02
-2.95586377e-01 -3.92514229e-01 1.13525486e+00 8.31786096e-02
-8.62918973e-01 8.60577345e-01 5.45834732e+00 1.04331589e+00
-1.29680538e+00 3.09697147e-02 7.04994798e-01 6.31018400e-01
-2.87128657e-01 6.68451786e-02 -3.87479156e-01 9.09599781e-01
5.52246034e-01 -8.77015889e-02 5.82109749e-01 6.54619694e-01
-1.78837553e-01 1.26653239e-01 -3.72855574e-01 1.16301465e+00
3.30647498e-01 -1.45365834e+00 5.70606962e-02 7.63171673e-01
8.01159799e-01 -4.03476387e-01 5.76422811e-01 -3.09780985e-01
-2.05810323e-01 -1.01931059e+00 5.00025749e-01 2.78042313e-02
1.23304141e+00 -7.40805686e-01 9.24744487e-01 4.32207316e-01
-9.18620110e-01 -2.26986989e-01 -4.69723165e-01 2.03664764e-03
1.08668238e-01 2.74347067e-01 -8.03702116e-01 2.21168339e-01
3.24980259e-01 6.88348293e-01 -2.81662911e-01 9.06081438e-01
-2.87508368e-01 5.29155016e-01 8.90040770e-02 -2.31067270e-01
4.34047371e-01 7.13934302e-02 6.86257303e-01 7.20919073e-01
6.64265335e-01 4.01493087e-02 -1.97033763e-01 5.22908390e-01
-1.52157858e-01 -1.63548797e-01 -8.94467294e-01 -3.06935281e-01
4.44034487e-01 8.68843615e-01 -9.25221264e-01 -3.04622918e-01
-1.48317590e-01 1.03045487e+00 -5.86919010e-01 5.07475846e-02
-7.44133174e-01 -8.03981006e-01 3.30179602e-01 2.34845430e-01
5.87568700e-01 -2.46440664e-01 -1.08030640e-01 -9.28226352e-01
-1.85494706e-01 -9.85971868e-01 -2.74762034e-01 -4.83639270e-01
-5.29016674e-01 4.59957451e-01 -3.56301516e-01 -1.49755013e+00
6.54002698e-03 -3.41058522e-01 -4.46634769e-01 6.35576785e-01
-1.69220817e+00 -1.11955357e+00 -2.53422409e-01 6.32215321e-01
3.77483428e-01 -5.91125369e-01 5.48684895e-01 1.67118296e-01
-2.89994240e-01 5.63726842e-01 1.51258543e-01 4.32484835e-01
1.86256632e-01 -1.74787074e-01 7.19547212e-01 1.01817203e+00
-3.10338229e-01 2.97051132e-01 7.63315916e-01 -8.04056287e-01
-1.46594882e+00 -1.09613776e+00 9.27606583e-01 4.35025513e-01
4.80143666e-01 -2.17563331e-01 -7.98088193e-01 4.55349058e-01
3.41655165e-01 -1.80655658e-01 2.18803406e-01 -1.03004324e+00
-3.19770634e-01 5.42481430e-02 -1.39100182e+00 5.89101732e-01
1.08438087e+00 -3.26717436e-01 -3.83430459e-02 1.66266318e-02
8.78005445e-01 -2.24874079e-01 -2.42693931e-01 3.58971804e-01
6.24331713e-01 -9.20388997e-01 8.39155912e-01 1.14091158e-01
4.61383641e-01 -2.40873724e-01 1.39341176e-01 -9.48964119e-01
-3.12988073e-01 -8.62198889e-01 2.35644709e-02 7.45986223e-01
2.08006978e-01 -1.03859091e+00 1.00810814e+00 9.06776264e-02
1.66394606e-01 -3.63814712e-01 -9.03105199e-01 -6.28838718e-01
-5.40144861e-01 -1.65883582e-02 6.84738874e-01 1.02720678e+00
-7.88528398e-02 -2.13856071e-01 -1.00359654e+00 3.35351713e-02
8.28541934e-01 -4.40594316e-01 5.01965821e-01 -1.05560541e+00
1.40431672e-02 -3.70085329e-01 -8.51274908e-01 -1.04480517e+00
-3.35787460e-02 -8.30896676e-01 -8.98988247e-02 -1.06987512e+00
9.84479263e-02 -4.73458081e-01 -1.26868680e-01 4.15407389e-01
8.33122432e-02 8.37637842e-01 2.64952660e-01 5.47555566e-01
-4.81909424e-01 3.28855038e-01 1.80033600e+00 -3.45576376e-01
1.90569907e-01 -2.12681487e-01 -8.73785079e-01 5.91564119e-01
9.73393381e-01 -7.70469368e-01 -5.23130178e-01 -3.56839687e-01
4.20148104e-01 1.34887397e-01 4.06357229e-01 -1.22578371e+00
3.84117573e-01 -3.66561189e-02 3.28935146e-01 -1.14704827e-02
1.20945573e-01 -1.19337845e+00 5.07977664e-01 1.42182744e+00
-1.32740393e-01 -2.57051051e-01 -2.57832110e-01 6.68932498e-01
-2.24766552e-01 -2.19434530e-01 9.46253598e-01 -4.34844613e-01
-1.00929880e+00 3.30678433e-01 -5.65053105e-01 -4.84895647e-01
1.21988082e+00 -8.62079501e-01 -2.49014199e-01 -5.56741655e-01
-5.63174672e-02 1.91558674e-02 5.44294775e-01 5.39135575e-01
1.10802543e+00 -1.08935845e+00 -6.40616417e-01 6.70215011e-01
-6.78727478e-02 -3.63287777e-01 4.02495503e-01 4.90259379e-01
-8.43167424e-01 4.85847324e-01 -4.41108525e-01 -3.44765961e-01
-1.40152323e+00 5.31643808e-01 2.08647445e-01 -2.49170333e-01
-5.51982522e-01 6.83968604e-01 1.57158837e-01 2.29089513e-01
-4.02291752e-02 3.36020082e-01 -2.82173723e-01 -4.64532197e-01
7.47452259e-01 2.82348961e-01 -3.93420130e-01 -8.29702735e-01
2.96396948e-02 3.76465887e-01 -3.09793323e-01 7.80209899e-02
9.66589332e-01 -4.21045184e-01 -2.89556980e-01 -4.28774625e-01
1.42690456e+00 -3.60258669e-01 -1.03470349e+00 -2.29793757e-01
-2.52746940e-01 -8.75453174e-01 3.56605679e-01 -4.96121287e-01
-1.53833950e+00 6.63772106e-01 7.45770752e-01 6.26001537e-01
1.02236533e+00 -4.06141818e-01 1.32340395e+00 2.96683997e-01
5.97982287e-01 -5.85666180e-01 2.48767897e-01 2.74036050e-01
4.91706938e-01 -1.16104627e+00 -9.44672674e-02 -5.72571218e-01
-4.68499273e-01 8.95313144e-01 3.87185991e-01 -2.06907749e-01
6.94723487e-01 -5.94348833e-02 -1.06204607e-01 -1.52613088e-01
-2.61757582e-01 1.11036487e-01 3.03074392e-03 7.18121290e-01
-2.42172524e-01 -3.47736061e-01 -2.82237619e-01 -1.76002413e-01
-3.74002792e-02 5.43394871e-03 5.57068706e-01 9.71922457e-01
-9.36315775e-01 -1.19875419e+00 -5.12528777e-01 2.39466980e-01
-9.43146646e-01 -2.88179100e-01 1.04714580e-01 6.99352741e-01
2.64292210e-01 9.21945393e-01 -3.56171280e-02 -7.95155644e-01
-2.62011200e-01 -4.15551096e-01 2.07883164e-01 -1.38647661e-01
-2.97848046e-01 -3.31604406e-02 -3.10055196e-01 -1.31862089e-01
-4.84429449e-01 -1.49396494e-01 -9.82377410e-01 -6.97035372e-01
-7.40509152e-01 1.42350987e-01 1.01167452e+00 8.22553039e-01
3.90663326e-01 5.02343118e-01 1.03796220e+00 -5.18009663e-01
-3.70992482e-01 -8.27258468e-01 -5.62150061e-01 2.19642803e-01
4.53590393e-01 6.12719078e-03 -4.86453056e-01 1.03734717e-01] | [4.294663429260254, 8.062056541442871] |
4924c6b0-6689-49d5-894f-2ee729ee742b | named-entity-recognition-only-from-word | 1909.00164 | null | https://arxiv.org/abs/1909.00164v2 | https://arxiv.org/pdf/1909.00164v2.pdf | Named Entity Recognition Only from Word Embeddings | Deep neural network models have helped named entity (NE) recognition achieve amazing performance without handcrafting features. However, existing systems require large amounts of human annotated training data. Efforts have been made to replace human annotations with external knowledge (e.g., NE dictionary, part-of-speech tags), while it is another challenge to obtain such effective resources. In this work, we propose a fully unsupervised NE recognition model which only needs to take informative clues from pre-trained word embeddings. We first apply Gaussian Hidden Markov Model and Deep Autoencoding Gaussian Mixture Model on word embeddings for entity span detection and type prediction, and then further design an instance selector based on reinforcement learning to distinguish positive sentences from noisy sentences and refine these coarse-grained annotations through neural networks. Extensive experiments on CoNLL benchmark datasets demonstrate that our proposed light NE recognition model achieves remarkable performance without using any annotated lexicon or corpus. | ['Junlang Zhan', 'Ying Luo', 'Hai Zhao'] | 2019-08-31 | null | https://aclanthology.org/2020.emnlp-main.723 | https://aclanthology.org/2020.emnlp-main.723.pdf | emnlp-2020-11 | ['type-prediction'] | ['computer-code'] | [ 2.44760718e-02 -1.31500736e-02 -2.75180489e-01 -6.87499106e-01
-7.48749614e-01 -5.65184295e-01 3.66949618e-01 7.64279254e-03
-9.38728809e-01 8.10179532e-01 4.19132471e-01 -1.98536173e-01
2.32625976e-01 -9.09695268e-01 -4.20655638e-01 -3.70919019e-01
2.19335333e-01 5.37701428e-01 8.33334997e-02 -8.35350007e-02
1.20024338e-01 4.32597935e-01 -1.05930078e+00 7.32718185e-02
1.06303239e+00 9.16276097e-01 3.92548591e-01 6.02900565e-01
-6.87266290e-01 7.63155580e-01 -7.37904608e-01 -5.41323006e-01
6.24972731e-02 -1.36551976e-01 -8.16555500e-01 -2.94230301e-02
-1.45104080e-01 -2.63152778e-01 -5.06319046e-01 1.08307433e+00
7.71252036e-01 3.27498019e-01 7.46464133e-01 -8.26204717e-01
-1.04741514e+00 8.17351043e-01 -1.55572847e-01 3.69194329e-01
-6.89426214e-02 1.64542779e-01 1.13788736e+00 -1.17622256e+00
6.59030914e-01 7.53949881e-01 5.81597686e-01 6.96510673e-01
-1.01996601e+00 -7.75971770e-01 1.87978912e-02 3.18976074e-01
-1.63854611e+00 -6.67909920e-01 7.32344449e-01 -2.39564806e-01
1.31855941e+00 1.06549986e-01 1.99971661e-01 1.17408955e+00
-2.97744989e-01 1.03912759e+00 6.49232566e-01 -3.61634284e-01
1.08390681e-01 3.60334784e-01 2.96767622e-01 7.46945262e-01
2.21505448e-01 -4.03747469e-01 -1.02837801e-01 -1.96254358e-01
5.84706306e-01 7.39305019e-02 -2.61432320e-01 -1.50691643e-01
-1.07203770e+00 7.42507517e-01 2.00674310e-01 6.07052565e-01
-5.41836143e-01 -1.59314528e-01 5.38948834e-01 9.37794596e-02
3.77427220e-01 6.30160689e-01 -1.09027159e+00 -2.16144845e-01
-9.71508622e-01 -4.33091402e-01 9.30070221e-01 1.06164384e+00
7.81117260e-01 1.47945896e-01 -4.43938375e-01 1.14753795e+00
3.06222588e-01 5.61897337e-01 7.22813666e-01 -3.26249093e-01
4.94739443e-01 6.49324894e-01 6.91913292e-02 -1.05028248e+00
-3.26539218e-01 -4.38881576e-01 -1.07594860e+00 -6.50043905e-01
1.58399157e-02 -4.23099637e-01 -9.91694331e-01 1.53184640e+00
1.45166308e-01 3.02983940e-01 2.66808391e-01 6.00155771e-01
1.01413631e+00 6.56677902e-01 5.32726586e-01 -2.29535857e-03
1.38997138e+00 -9.75543976e-01 -9.04293835e-01 -1.61352590e-01
8.55523825e-01 -6.85421646e-01 8.73784363e-01 1.17917350e-02
-6.56984270e-01 -5.17161489e-01 -8.28949749e-01 -2.38630459e-01
-6.61402106e-01 6.71543181e-01 5.04428506e-01 8.73769343e-01
-5.33884823e-01 3.46079946e-01 -9.53229129e-01 -1.65288389e-01
5.61654210e-01 5.86787522e-01 -3.06705803e-01 5.76495845e-03
-1.39633858e+00 7.25109756e-01 7.95899451e-01 3.40218216e-01
-8.27848375e-01 -4.43281919e-01 -1.11293828e+00 4.60923314e-01
4.08126295e-01 -6.11434698e-01 1.13850796e+00 -7.53975570e-01
-1.60398698e+00 7.53031194e-01 -3.23808104e-01 -3.64617378e-01
-1.98517442e-02 -2.23561406e-01 -9.78872538e-01 -1.23766009e-02
5.11659542e-03 6.66899145e-01 6.29481673e-01 -9.12807286e-01
-6.22737885e-01 -9.08801854e-02 -1.99280858e-01 -1.64255247e-01
-7.25394845e-01 2.14744896e-01 -2.45389715e-01 -6.97253585e-01
-2.34575182e-01 -6.60364747e-01 -4.58798021e-01 -5.11726379e-01
-6.27469122e-01 -6.99505746e-01 6.36192262e-01 -8.17741036e-01
1.55671155e+00 -1.98423469e+00 -3.55359942e-01 2.26338342e-01
3.51631582e-01 6.72190309e-01 -2.86743134e-01 1.65432826e-01
1.31516516e-01 3.04879010e-01 -1.41973644e-01 -2.41450354e-01
4.02666271e-01 1.78732440e-01 -3.08362395e-01 8.60618800e-02
5.44798970e-01 1.04443622e+00 -9.53765869e-01 -6.62274003e-01
-1.35278171e-02 4.47663009e-01 -5.39126635e-01 4.11181152e-01
-1.35001987e-01 1.65061444e-01 -7.13022411e-01 4.50074911e-01
6.38387203e-01 -4.61843312e-01 2.12026164e-01 -3.42502445e-01
7.65508488e-02 6.07784688e-01 -1.33606386e+00 1.34040058e+00
-6.76775753e-01 4.20319885e-01 -2.93861300e-01 -1.10949039e+00
1.14510417e+00 4.93730783e-01 8.84284973e-02 -4.04830009e-01
2.91365325e-01 2.72398770e-01 -8.19036067e-02 -7.62397349e-01
7.48164654e-01 -1.33369982e-01 -2.89802432e-01 3.39369446e-01
4.21801031e-01 5.17937601e-01 2.37875149e-01 -7.54918456e-02
1.21522784e+00 -2.96676338e-01 3.87245089e-01 1.53359100e-01
7.71030605e-01 2.40352340e-02 1.08883166e+00 5.48109055e-01
-2.85627395e-01 2.71985352e-01 2.06419200e-01 -3.98395717e-01
-1.14608729e+00 -6.60514653e-01 -1.39828846e-01 1.42522824e+00
-1.25653076e-03 -4.87204969e-01 -6.07101083e-01 -1.19155252e+00
-4.75027740e-01 8.30290854e-01 -2.81357676e-01 -2.22555146e-01
-6.52890682e-01 -7.94753671e-01 9.70408082e-01 8.94252837e-01
7.27338195e-01 -1.39897990e+00 -5.89133091e-02 3.93270940e-01
-1.91571355e-01 -1.19717038e+00 -7.03904331e-01 3.28142047e-01
-5.49384475e-01 -8.35260630e-01 -6.94489062e-01 -1.06374967e+00
7.23537266e-01 -1.38587564e-01 1.03360796e+00 4.61118706e-02
-1.41745552e-01 1.56616583e-01 -4.25214201e-01 -1.09174810e-01
-1.82648867e-01 5.65218866e-01 1.56663954e-01 -8.03903490e-02
1.06085324e+00 -2.58149862e-01 -5.19099057e-01 2.96146154e-01
-9.17123139e-01 -3.25040698e-01 9.98730779e-01 1.20020711e+00
5.30982196e-01 1.71417549e-01 8.88833940e-01 -1.03397298e+00
6.26447082e-01 -5.62498331e-01 -4.01848406e-01 5.49300015e-01
-5.29471099e-01 3.89222592e-01 7.67192006e-01 -4.80850846e-01
-1.49069285e+00 5.83767071e-02 -5.56467533e-01 -2.19779432e-01
-5.95817506e-01 5.68667412e-01 -6.23273611e-01 3.79904211e-01
3.79093945e-01 4.85012174e-01 -7.22044826e-01 -6.95442617e-01
4.82065380e-01 1.23250759e+00 4.25452858e-01 -5.30391634e-01
6.07253194e-01 -6.35674456e-03 -6.78100109e-01 -8.91856670e-01
-1.13827384e+00 -5.99949896e-01 -8.72428477e-01 5.88454306e-01
1.00553513e+00 -8.73599648e-01 -4.83289719e-01 4.45985571e-02
-1.27280712e+00 -1.39856055e-01 -7.52896592e-02 7.23574340e-01
-1.83771759e-01 2.54311770e-01 -6.96840286e-01 -5.89409351e-01
-5.00487566e-01 -8.58739495e-01 8.00928473e-01 4.12035912e-01
-2.04266056e-01 -8.66931438e-01 3.02959114e-01 3.66356552e-01
4.69337970e-01 -5.13246179e-01 7.77499914e-01 -1.48390841e+00
-4.85213995e-01 -3.39566439e-01 -4.30391192e-01 5.24690270e-01
1.00131772e-01 -1.48146674e-01 -9.54884887e-01 7.89540932e-02
-2.68384933e-01 -2.92129278e-01 9.44850028e-01 -2.49011777e-02
1.29525745e+00 -4.15540725e-01 -5.08601546e-01 6.90688908e-01
1.23472369e+00 2.87609875e-01 4.76193547e-01 2.74686456e-01
9.07671630e-01 2.56599456e-01 3.29357535e-01 4.64892775e-01
4.81163949e-01 2.15185449e-01 -1.24713309e-01 4.54169922e-02
9.34043378e-02 -4.45999324e-01 1.23292468e-01 1.18526876e+00
1.46719307e-01 -4.86751884e-01 -9.29783046e-01 7.01857388e-01
-1.38902569e+00 -9.78139460e-01 3.43428016e-01 1.75477099e+00
1.08968508e+00 1.27293870e-01 -2.86506683e-01 -2.09261686e-01
8.52045655e-01 1.50526777e-01 -4.69545782e-01 -1.23645552e-01
-1.21290579e-01 4.05706167e-01 5.39732754e-01 6.90381974e-02
-1.44941640e+00 1.31134892e+00 5.45296717e+00 1.19753551e+00
-1.01383412e+00 1.83648795e-01 5.65982342e-01 4.06677574e-01
-3.84934872e-01 -2.93569058e-01 -1.36042917e+00 4.57585752e-01
1.00624025e+00 2.23841257e-02 6.06527962e-02 1.12162912e+00
-1.79175958e-01 5.02661884e-01 -8.16023946e-01 9.29647624e-01
2.39872690e-02 -1.26487422e+00 -6.34261146e-02 -9.71882418e-02
8.42620075e-01 2.77801275e-01 -2.83303350e-01 9.37888741e-01
7.08586633e-01 -1.04657543e+00 5.02433255e-02 7.02516556e-01
7.49758244e-01 -8.22089911e-01 1.01410580e+00 4.97352183e-01
-1.25523543e+00 3.72841135e-02 -7.29060650e-01 3.47134501e-01
1.33165956e-01 5.59139431e-01 -1.06490648e+00 2.48801559e-01
3.99889171e-01 4.90963697e-01 -3.41364443e-01 1.18088460e+00
-5.33110142e-01 9.34911728e-01 -3.84908646e-01 -3.47866088e-01
5.41188419e-02 6.64457977e-02 2.92269915e-01 1.71434557e+00
2.63283432e-01 3.47375691e-01 3.21267307e-01 9.52256262e-01
-6.17518067e-01 3.40520620e-01 -2.45133519e-01 -6.85156107e-01
7.40205824e-01 1.52688849e+00 -7.63364375e-01 -6.57121301e-01
-4.95708138e-01 1.16610658e+00 6.56379879e-01 3.79827678e-01
-6.29206479e-01 -9.96341050e-01 6.67671323e-01 -3.86962235e-01
8.15191865e-01 -2.78492719e-01 -5.39165772e-02 -1.60276020e+00
-3.39084089e-01 -5.56374669e-01 4.66111213e-01 -3.60889971e-01
-1.61303163e+00 8.53537619e-01 -5.65141737e-01 -1.01728165e+00
-1.86340794e-01 -5.59519947e-01 -6.18155599e-01 7.87619770e-01
-1.60099506e+00 -9.72721279e-01 2.66573634e-02 3.36049289e-01
5.31368434e-01 -4.23667341e-01 1.06716180e+00 6.59496307e-01
-9.86608505e-01 8.85054052e-01 2.65235931e-01 1.01378870e+00
8.46076071e-01 -1.27249253e+00 3.87524188e-01 6.02700830e-01
4.19478118e-01 8.85895014e-01 3.68928164e-01 -7.30402231e-01
-1.06702709e+00 -1.17726815e+00 1.40946412e+00 -4.84833896e-01
5.67237794e-01 -3.13287646e-01 -9.48108196e-01 8.81276309e-01
1.02667645e-01 2.36081123e-01 1.15049100e+00 2.67281890e-01
-3.50663155e-01 9.73993465e-02 -1.06216776e+00 3.84972453e-01
9.99409735e-01 -7.50041306e-01 -9.82079446e-01 1.71135709e-01
8.86455297e-01 -2.02833295e-01 -9.59833920e-01 3.27545255e-01
2.40694359e-01 -2.08053052e-01 9.19091344e-01 -9.42589283e-01
1.74381480e-01 -3.71962935e-01 -8.45703483e-02 -1.30954909e+00
-3.00347239e-01 -2.13633910e-01 -1.02262922e-01 1.74016273e+00
7.09013879e-01 -3.75417620e-01 8.72484803e-01 7.61788249e-01
-2.10013852e-01 -6.07428372e-01 -4.45721328e-01 -7.27703333e-01
-1.90474346e-01 -4.81240153e-01 8.49487126e-01 1.42929673e+00
-9.63319317e-02 7.25535333e-01 -2.99906015e-01 4.07198280e-01
1.79890439e-01 -1.59401018e-02 6.12415373e-01 -1.01064968e+00
-1.55462489e-01 -3.14685434e-01 -2.64096975e-01 -1.48150861e+00
5.17813385e-01 -1.05932057e+00 2.69653708e-01 -1.50680315e+00
2.87686288e-01 -6.85282290e-01 -6.39595211e-01 6.22441709e-01
-4.61351424e-01 1.18134543e-01 -1.78527370e-01 -5.11538051e-02
-1.23602629e+00 1.01420033e+00 7.68612385e-01 -2.49776587e-01
-2.09255382e-01 -2.10763827e-01 -4.87784833e-01 8.26745212e-01
8.16024125e-01 -7.11892843e-01 -3.31987031e-02 -6.35441065e-01
3.62297148e-01 -1.76125899e-01 -3.14947255e-02 -6.98221803e-01
4.84497994e-01 -5.38576357e-02 6.88347042e-01 -6.45742834e-01
5.51569462e-02 -7.32824087e-01 -3.40210557e-01 8.15013889e-03
-4.74572718e-01 -2.67660797e-01 -5.60896173e-02 6.45917118e-01
-3.34999830e-01 -5.90549469e-01 3.52166474e-01 -1.07684091e-01
-1.14331245e+00 5.12191117e-01 -3.61302704e-01 3.39948922e-01
6.67453885e-01 6.76550344e-02 -1.92858338e-01 -2.24695671e-02
-9.13339674e-01 3.53140444e-01 -1.79641843e-02 3.36249381e-01
4.72570390e-01 -1.47816205e+00 -4.88244355e-01 3.03208530e-01
2.81782717e-01 -1.55775428e-01 4.07877892e-01 4.92245138e-01
-1.83859780e-01 6.30658329e-01 -3.95565853e-02 -7.98171461e-02
-1.10031557e+00 6.58328116e-01 1.13988258e-01 -6.20373607e-01
-2.49329656e-01 1.02692187e+00 1.33530917e-02 -1.03973842e+00
1.78462356e-01 -2.57069498e-01 -5.98452508e-01 -5.45543730e-02
6.21576250e-01 2.25438759e-01 -8.92974213e-02 -5.42833149e-01
-2.77738810e-01 2.56046206e-01 -2.50914395e-01 2.78054148e-01
1.48099613e+00 -2.72095054e-01 1.45878449e-01 8.40105303e-03
1.30546212e+00 1.94707304e-01 -7.66884983e-01 -5.89457393e-01
5.30878067e-01 -1.46525264e-01 1.33514196e-01 -6.14730239e-01
-9.21729445e-01 1.06937003e+00 3.35140139e-01 3.14545408e-02
9.00475860e-01 5.57746645e-03 1.44541943e+00 1.01144814e+00
1.22029863e-01 -1.19844556e+00 -1.47529885e-01 8.05434346e-01
2.04297215e-01 -1.53658032e+00 -4.83709663e-01 -1.78738877e-01
-6.36118293e-01 1.01261151e+00 6.70500576e-01 -1.15525536e-01
8.85803699e-01 2.12942481e-01 4.13631089e-02 -7.09863007e-02
-6.37294948e-01 -4.21088964e-01 4.86429036e-01 4.97978032e-01
4.98341262e-01 1.08791545e-01 -3.62067252e-01 1.44261777e+00
-4.84269708e-02 5.47493733e-02 1.08700968e-01 6.11641943e-01
-7.11570323e-01 -1.32033634e+00 4.13352363e-02 6.93989038e-01
-6.33387625e-01 -4.91523981e-01 -2.79644206e-02 5.05788982e-01
2.49425635e-01 6.11438930e-01 7.24754035e-02 -3.54180455e-01
2.52471805e-01 3.67849261e-01 5.82326017e-02 -9.61645126e-01
-4.58194166e-01 -6.36543706e-03 3.26632619e-01 -9.14504007e-02
-1.77082971e-01 -3.42135251e-01 -1.36349857e+00 -2.96584349e-02
-6.05967999e-01 4.27862585e-01 4.74013865e-01 1.14915287e+00
5.30336857e-01 3.92718494e-01 4.96992618e-01 -3.85907561e-01
-5.52582204e-01 -1.12162471e+00 -5.08930266e-01 2.34277397e-01
1.43905412e-02 -4.44742322e-01 -1.32213324e-01 1.26612514e-01] | [9.670619010925293, 9.427018165588379] |
3a524748-47af-4af5-9b39-a6058ad94236 | synthetic-ct-generation-from-mri-using-3d | 2305.19467 | null | https://arxiv.org/abs/2305.19467v1 | https://arxiv.org/pdf/2305.19467v1.pdf | Synthetic CT Generation from MRI using 3D Transformer-based Denoising Diffusion Model | Magnetic resonance imaging (MRI)-based synthetic computed tomography (sCT) simplifies radiation therapy treatment planning by eliminating the need for CT simulation and error-prone image registration, ultimately reducing patient radiation dose and setup uncertainty. We propose an MRI-to-CT transformer-based denoising diffusion probabilistic model (MC-DDPM) to transform MRI into high-quality sCT to facilitate radiation treatment planning. MC-DDPM implements diffusion processes with a shifted-window transformer network to generate sCT from MRI. The proposed model consists of two processes: a forward process which adds Gaussian noise to real CT scans, and a reverse process in which a shifted-window transformer V-net (Swin-Vnet) denoises the noisy CT scans conditioned on the MRI from the same patient to produce noise-free CT scans. With an optimally trained Swin-Vnet, the reverse diffusion process was used to generate sCT scans matching MRI anatomy. We evaluated the proposed method by generating sCT from MRI on a brain dataset and a prostate dataset. Qualitative evaluation was performed using the mean absolute error (MAE) of Hounsfield unit (HU), peak signal to noise ratio (PSNR), multi-scale Structure Similarity index (MS-SSIM) and normalized cross correlation (NCC) indexes between ground truth CTs and sCTs. MC-DDPM generated brain sCTs with state-of-the-art quantitative results with MAE 43.317 HU, PSNR 27.046 dB, SSIM 0.965, and NCC 0.983. For the prostate dataset, MC-DDPM achieved MAE 59.953 HU, PSNR 26.920 dB, SSIM 0.849, and NCC 0.948. In conclusion, we have developed and validated a novel approach for generating CT images from routine MRIs using a transformer-based DDPM. This model effectively captures the complex relationship between CT and MRI images, allowing for robust and high-quality synthetic CT (sCT) images to be generated in minutes. | ['Xiaofeng Yang', 'Hui Mao', 'David S. Yu', 'Pretesh Patel', 'Justin Roper', 'Junbo Peng', 'Chih-Wei Chang', 'Yuheng Li', 'Richard L. J. Qiu', 'Tonghe Wang', 'Jacob Wynne', 'Elham Abouei', 'Shaoyan Pan'] | 2023-05-31 | null | null | null | null | ['image-registration', 'ms-ssim', 'anatomy'] | ['computer-vision', 'computer-vision', 'miscellaneous'] | [ 4.44251597e-01 1.36550069e-01 3.48496139e-01 -3.15524071e-01
-1.13160014e+00 -3.54437590e-01 6.63223386e-01 1.00853242e-01
-6.55314267e-01 6.25757992e-01 4.14447874e-01 -3.06757241e-01
-2.69468457e-01 -9.37987685e-01 -3.51040035e-01 -9.71065938e-01
-2.62534767e-01 8.00394654e-01 5.55684745e-01 2.10090727e-01
-5.77169731e-02 6.64254725e-01 -5.69053471e-01 3.00235450e-01
8.37885201e-01 1.08169699e+00 6.32110476e-01 6.85952365e-01
1.06051184e-01 8.26003969e-01 -4.42949623e-01 -3.99127603e-01
3.44814181e-01 -5.71807504e-01 -7.47505665e-01 -1.32054701e-01
-3.92680429e-02 -2.68244684e-01 -3.61859411e-01 9.60825145e-01
8.61669183e-01 1.78667773e-02 8.89793575e-01 -7.84576654e-01
-4.60275382e-01 6.56013727e-01 -7.54871130e-01 3.51437628e-01
6.04159534e-02 5.55316150e-01 -1.08530857e-01 -6.62879288e-01
9.07554805e-01 8.18941057e-01 8.26234281e-01 6.35163844e-01
-1.38176966e+00 -5.43805838e-01 -9.35692012e-01 3.72439027e-02
-1.06271720e+00 1.53503597e-01 3.65873992e-01 -4.98125345e-01
7.29116738e-01 5.35978734e-01 8.67470145e-01 8.20654273e-01
1.08071744e+00 2.37272382e-01 1.64606667e+00 -1.66678548e-01
4.13853496e-01 -4.67653275e-01 -4.63953137e-01 5.90428412e-01
1.27540722e-01 2.46935025e-01 4.30565029e-02 -1.21613137e-01
1.06892192e+00 -1.70442954e-01 -4.79503632e-01 -1.91759780e-01
-1.45243108e+00 6.57425165e-01 7.89188564e-01 7.33425200e-01
-8.18223417e-01 1.30106598e-01 5.17125607e-01 -1.43007681e-01
1.72698796e-01 1.10851184e-01 3.49967241e-01 7.15890229e-02
-1.20663536e+00 -5.48187606e-02 4.09526348e-01 5.91482341e-01
-1.10789672e-01 1.07522145e-01 -5.51747680e-01 6.32846534e-01
1.89269871e-01 7.69672275e-01 1.05395257e+00 -1.03755355e+00
1.84872113e-02 1.03390910e-01 -4.80426401e-01 -5.39454401e-01
-2.93172032e-01 -6.61001682e-01 -1.37696612e+00 2.65724033e-01
2.22339064e-01 9.15144384e-02 -1.38820255e+00 1.39729738e+00
3.67963165e-01 5.72721660e-02 -1.22675829e-01 1.09512532e+00
7.44975388e-01 5.08185387e-01 2.79279858e-01 -4.05935496e-01
1.34261918e+00 -6.61065519e-01 -6.87518239e-01 2.55432352e-03
4.63477522e-01 -9.31466043e-01 8.74263167e-01 2.29139924e-01
-1.58733189e+00 -8.79470482e-02 -1.01047814e+00 4.31093007e-01
3.92591596e-01 -2.42039084e-01 1.77815855e-01 7.05705225e-01
-1.19123065e+00 7.52354324e-01 -1.21346867e+00 3.55822816e-02
5.86024582e-01 4.39744711e-01 -4.32039708e-01 -4.94073182e-01
-8.72650743e-01 1.15130222e+00 1.52135208e-01 8.78568739e-02
-1.32030129e+00 -1.07282150e+00 -6.78669035e-01 -3.00293595e-01
8.62005726e-02 -8.87235105e-01 1.32446957e+00 -5.54838300e-01
-1.45500135e+00 6.79480791e-01 1.16276570e-01 -6.22262120e-01
9.93520498e-01 7.07230508e-01 -3.54729503e-01 5.63704133e-01
3.81984413e-01 6.42990649e-01 4.68835086e-01 -1.29185379e+00
1.17196836e-01 -4.86340046e-01 -8.41640532e-01 1.83717921e-01
4.61708754e-01 1.39602408e-01 -2.84194827e-01 -8.16922307e-01
4.98930454e-01 -9.45024371e-01 -4.98070627e-01 2.28313401e-01
-2.27297336e-01 5.53055346e-01 4.90789503e-01 -1.04028141e+00
7.85237610e-01 -1.65515196e+00 -9.07820910e-02 5.23954153e-01
3.57193798e-01 1.31634414e-01 -2.81832553e-02 -3.87706369e-01
-3.43087703e-01 9.73283779e-03 -9.51543689e-01 -5.35781831e-02
-5.43676555e-01 3.20208818e-01 3.29165131e-01 6.56397581e-01
-4.25329119e-01 1.16551614e+00 -1.05478454e+00 -6.32452846e-01
3.47149998e-01 7.91876137e-01 -1.04863837e-01 9.87509489e-02
4.25082207e-01 8.93371582e-01 -2.45238379e-01 4.92513180e-01
9.67143893e-01 9.84583274e-02 1.85560331e-01 -5.54273784e-01
-4.03650030e-02 -2.89016604e-01 -8.05718243e-01 1.65386355e+00
-5.98863780e-01 3.60025167e-01 2.81194061e-01 -4.94458824e-01
7.00279593e-01 5.83519280e-01 9.11350012e-01 -1.18261611e+00
3.84549528e-01 5.24224758e-01 3.38152558e-01 -3.45151484e-01
-8.08291361e-02 -9.46392179e-01 2.03515485e-01 4.82758880e-01
2.74889451e-02 -8.67212713e-01 -3.57680470e-02 2.18280777e-01
1.31204271e+00 -2.61718690e-01 -1.33704692e-01 -3.46275568e-01
5.52412391e-01 1.82883404e-02 2.97046900e-01 5.66343963e-01
-3.28284204e-01 9.50356007e-01 1.00583293e-01 -1.46523252e-01
-1.33408797e+00 -1.52396655e+00 -2.26189002e-01 -1.00546815e-01
-1.49515629e-01 1.34664565e-01 -1.06522107e+00 -4.87057507e-01
-5.57050526e-01 9.16327834e-01 -5.55676937e-01 -8.78415033e-02
-7.36064255e-01 -1.02278197e+00 6.46022558e-01 4.30997699e-01
5.96576571e-01 -1.04804623e+00 -6.25778019e-01 4.46080297e-01
-5.25827646e-01 -9.44843411e-01 -6.99033141e-01 6.38995245e-02
-1.26281106e+00 -6.86549664e-01 -1.34217811e+00 -4.90224272e-01
7.50444472e-01 -8.21535140e-02 1.01371074e+00 -6.54760897e-02
-5.60716093e-01 1.46599859e-01 4.22868393e-02 -5.01755858e-03
-9.58554804e-01 -6.36856318e-01 -6.59000427e-02 -4.00161177e-01
-4.55948681e-01 -6.93416238e-01 -9.88658667e-01 4.55632806e-01
-1.42602897e+00 1.39445841e-01 7.87560582e-01 9.91461158e-01
1.07240868e+00 8.65875557e-02 -1.27274198e-02 -4.42683369e-01
7.29902446e-01 -1.81049868e-01 -4.88237053e-01 2.84368336e-01
-7.57626832e-01 4.58818153e-02 3.99320602e-01 -4.49022055e-01
-1.22712398e+00 -3.54472362e-03 -3.63256246e-01 -4.06138450e-01
1.38963580e-01 3.28959405e-01 2.56594032e-01 -5.29705822e-01
8.36693585e-01 6.05549693e-01 2.30894998e-01 3.10843550e-02
2.48651244e-02 2.84449339e-01 9.47536886e-01 -3.85852277e-01
6.10757768e-01 6.27396941e-01 3.48403096e-01 -4.50334251e-01
-1.18883900e-01 -1.05760321e-01 -5.85339010e-01 -5.19656420e-01
1.12242258e+00 -4.45638627e-01 -4.91648346e-01 5.94885826e-01
-8.55337143e-01 -3.68461221e-01 -5.36370933e-01 8.66767526e-01
-6.02682650e-01 5.99408925e-01 -9.75353241e-01 -2.40219906e-01
-9.71271276e-01 -1.82710254e+00 6.84625447e-01 9.92562175e-02
-2.11300224e-01 -8.06311488e-01 1.27219949e-02 4.54630703e-01
9.30878699e-01 6.72301710e-01 8.58061612e-01 -1.60440907e-01
-4.55668688e-01 -1.12631269e-01 -2.07177833e-01 5.15494049e-01
-2.26427857e-02 -5.26388884e-01 -5.97226739e-01 -1.36977330e-01
6.71362877e-01 4.36413027e-02 4.25451815e-01 1.01961768e+00
1.09752429e+00 6.62354827e-02 -3.58832040e-04 8.01054537e-01
1.59923160e+00 5.45725048e-01 9.61172938e-01 3.07112396e-01
4.08755064e-01 1.66211262e-01 2.55591840e-01 1.06289826e-01
1.28450757e-02 4.52309072e-01 4.64056134e-01 -2.75088578e-01
-6.29509985e-01 -3.30107920e-02 -2.38500461e-02 7.50239611e-01
-3.21799040e-01 3.02644242e-02 -1.20390320e+00 5.27536929e-01
-1.03937542e+00 -7.88562894e-01 -5.24366319e-01 2.20089960e+00
8.07394981e-01 1.44529417e-01 -2.64905274e-01 5.61167784e-02
7.80224621e-01 -2.75517851e-01 -3.37991774e-01 -1.40791267e-01
-6.35923743e-02 7.90831566e-01 9.21539962e-01 6.29727364e-01
-5.57922423e-01 1.74850091e-01 5.13217354e+00 1.08848524e+00
-1.28060102e+00 7.57838547e-01 8.79832685e-01 5.45218103e-02
-4.03045624e-01 -2.52662927e-01 2.48066530e-01 4.73148644e-01
1.02897668e+00 -3.27262998e-01 2.36057684e-01 4.47143257e-01
4.08377558e-01 -6.20712698e-01 -5.38693726e-01 8.38626266e-01
-1.09974355e-01 -1.40307951e+00 -7.10009858e-02 1.83979541e-01
6.58260763e-01 2.04459816e-01 8.32324848e-02 -1.66929394e-01
1.97400823e-01 -1.23054612e+00 5.73378026e-01 6.25464618e-01
1.18493414e+00 -6.14309430e-01 9.16599989e-01 1.90510571e-01
-9.10684228e-01 4.42117423e-01 -1.30974174e-01 9.40298080e-01
5.13818860e-01 8.75261068e-01 -1.26736104e+00 9.27307844e-01
6.00407958e-01 -8.40366334e-02 -2.39900112e-01 1.29696357e+00
-9.54185575e-02 3.67796957e-01 -4.05861348e-01 4.12788838e-01
3.68929178e-01 -3.12263131e-01 5.82561016e-01 1.12353945e+00
6.56392455e-01 4.74358410e-01 -3.50552589e-01 7.23154068e-01
2.99834814e-02 -1.14484124e-01 -3.56909521e-02 5.85677981e-01
1.56981066e-01 1.44526398e+00 -1.31078196e+00 -4.61448640e-01
2.59632796e-01 1.04465115e+00 -5.18491209e-01 -4.69133034e-02
-9.03950751e-01 1.33858427e-01 -2.25301310e-01 4.16130483e-01
-2.21849963e-01 -1.21427134e-01 -5.36726236e-01 -7.21294582e-01
-1.30407810e-01 -6.80820048e-01 2.62718916e-01 -1.05975235e+00
-1.25035655e+00 1.16957128e+00 1.97130159e-01 -1.14505935e+00
-1.34263530e-01 -1.49690479e-01 -7.45585918e-01 1.17543972e+00
-1.11237466e+00 -9.14970279e-01 -4.66408283e-01 5.98622799e-01
3.42841268e-01 3.53202164e-01 6.73454762e-01 4.25991267e-01
1.42646488e-02 2.51630396e-01 1.41921982e-01 -1.64137259e-02
4.62478340e-01 -1.31455767e+00 1.58679634e-01 6.23786330e-01
-5.99102139e-01 2.40179077e-01 5.76413631e-01 -1.10655928e+00
-1.10074151e+00 -1.09237134e+00 5.49704909e-01 -5.98910684e-03
5.00911355e-01 1.58401713e-01 -7.38495886e-01 2.62820244e-01
1.56813323e-01 3.42807353e-01 3.52604717e-01 -1.17895305e+00
3.94532949e-01 1.54636335e-02 -2.06315136e+00 4.29794043e-01
5.13140261e-01 -1.44922942e-01 -4.14782852e-01 3.49201649e-01
3.38571608e-01 -8.35504591e-01 -1.49764037e+00 5.52615285e-01
5.18709064e-01 -8.46931159e-01 1.20145464e+00 4.24219698e-01
4.17713344e-01 -2.04038888e-01 1.20499715e-01 -1.44160092e+00
-2.34924465e-01 -2.30039537e-01 5.06052971e-01 5.99424183e-01
2.61926204e-01 -3.80935878e-01 7.01078534e-01 7.64277160e-01
-3.96132469e-01 -7.75374174e-01 -1.46299744e+00 -7.15988994e-01
2.61516750e-01 -6.02469444e-01 2.90385246e-01 8.50064933e-01
-5.57163775e-01 -3.13729554e-01 2.60514859e-02 6.34872494e-03
1.09712279e+00 -5.52148342e-01 -2.07002833e-01 -7.61042535e-01
-2.34992296e-01 -3.22589070e-01 -2.95821100e-01 -2.97780097e-01
-3.85701805e-01 -1.19272470e+00 8.85794777e-03 -1.82640803e+00
3.84737462e-01 -4.65021700e-01 8.73090001e-04 1.37316212e-01
1.90181360e-01 5.01866400e-01 1.38483077e-01 2.86940396e-01
3.90084058e-01 4.37104642e-01 2.11183023e+00 -2.46512190e-01
1.40318736e-01 -7.36081675e-02 -4.29960415e-02 6.09930515e-01
6.36988461e-01 -8.41147244e-01 -3.44777197e-01 -2.29988635e-01
-5.24412692e-01 7.65701115e-01 5.17145574e-01 -1.30388975e+00
3.73729885e-01 1.19739823e-01 9.78351116e-01 -6.19932115e-01
2.92797118e-01 -8.24821115e-01 7.70824432e-01 1.28749382e+00
-3.78297307e-02 2.83676714e-01 1.24883883e-01 1.31795108e-01
-1.03530221e-01 -3.13594013e-01 1.19082308e+00 -4.34930593e-01
-6.76792935e-02 3.55746120e-01 -5.05023658e-01 -2.52547801e-01
1.07916403e+00 -2.59244025e-01 6.37175888e-02 -3.58665228e-01
-1.13182664e+00 -1.79530740e-01 2.44391829e-01 -2.62167633e-01
9.91426110e-01 -1.34071529e+00 -8.13613176e-01 7.25132152e-02
-4.16512072e-01 7.44418129e-02 6.91754401e-01 1.40837801e+00
-1.03947067e+00 1.60045832e-01 -4.28172380e-01 -8.22899640e-01
-1.21795332e+00 2.07420260e-01 7.94144869e-01 -8.55234027e-01
-7.42510498e-01 8.02906215e-01 2.55217846e-03 -5.10885596e-01
-3.19131166e-01 -5.15755832e-01 3.25624824e-01 -4.59032416e-01
3.58517408e-01 2.29528785e-01 5.67407131e-01 -7.89447308e-01
-2.90040284e-01 4.85352397e-01 1.36870191e-01 -6.62854135e-01
1.32982516e+00 1.90088304e-03 -1.25412166e-01 -2.82617062e-01
9.98853147e-01 -2.26891592e-01 -9.60362911e-01 -1.04349598e-01
-2.13985294e-01 -4.25249696e-01 6.02211952e-01 -1.31947052e+00
-1.26860821e+00 6.25532925e-01 1.29790795e+00 -3.98303151e-01
1.45521653e+00 -1.02941468e-01 1.08866465e+00 -5.08332670e-01
4.43121344e-01 -6.53190911e-01 -7.94678554e-02 1.56934515e-01
9.67825353e-01 -8.07871044e-01 1.19555667e-01 -5.27941942e-01
-9.65187252e-01 1.10722828e+00 2.98899174e-01 -8.32568668e-03
4.42310780e-01 7.26171315e-01 2.81248074e-02 -1.88285753e-01
-1.60458028e-01 3.56146365e-01 2.64705032e-01 6.17172956e-01
4.70588744e-01 1.67504922e-01 -5.15912652e-01 2.87414074e-01
-1.73995450e-01 2.65899599e-01 5.90902150e-01 1.04226816e+00
-1.60633266e-01 -1.06255960e+00 -6.68926418e-01 4.26565945e-01
-5.87549686e-01 -2.28993386e-01 1.37840375e-01 6.10374331e-01
-5.65077644e-03 4.17144120e-01 -2.51515061e-01 -1.54408827e-01
3.62281233e-01 -3.61094385e-01 8.58673096e-01 -3.06623369e-01
-1.07116592e+00 3.90323728e-01 -1.85275733e-01 -5.18790960e-01
-2.92205244e-01 -5.01287639e-01 -1.52225959e+00 -3.87247175e-01
7.22802579e-02 8.32988769e-02 1.17276025e+00 8.21240187e-01
-2.04514965e-01 8.99315953e-01 4.87913042e-01 -7.72997737e-01
-3.61912310e-01 -8.86422753e-01 -5.97681701e-01 2.05278292e-01
-8.42354372e-02 -3.04572403e-01 -1.06551133e-01 -1.82840690e-01] | [13.626974105834961, -2.499552011489868] |
9842ae22-17c6-4d78-a7fc-9ed031e344d2 | boxcars-improving-fine-grained-recognition-of | 1703.00686 | null | http://arxiv.org/abs/1703.00686v3 | http://arxiv.org/pdf/1703.00686v3.pdf | BoxCars: Improving Fine-Grained Recognition of Vehicles using 3-D Bounding Boxes in Traffic Surveillance | In this paper, we focus on fine-grained recognition of vehicles mainly in
traffic surveillance applications. We propose an approach that is orthogonal to
recent advancements in fine-grained recognition (automatic part discovery and
bilinear pooling). In addition, in contrast to other methods focused on
fine-grained recognition of vehicles, we do not limit ourselves to a
frontal/rear viewpoint, but allow the vehicles to be seen from any viewpoint.
Our approach is based on 3-D bounding boxes built around the vehicles. The
bounding box can be automatically constructed from traffic surveillance data.
For scenarios where it is not possible to use precise construction, we propose
a method for an estimation of the 3-D bounding box. The 3-D bounding box is
used to normalize the image viewpoint by "unpacking" the image into a plane. We
also propose to randomly alter the color of the image and add a rectangle with
random noise to a random position in the image during the training of
convolutional neural networks (CNNs). We have collected a large fine-grained
vehicle data set BoxCars116k, with 116k images of vehicles from various
viewpoints taken by numerous surveillance cameras. We performed a number of
experiments, which show that our proposed method significantly improves CNN
classification accuracy (the accuracy is increased by up to 12% points and the
error is reduced by up to 50% compared with CNNs without the proposed
modifications). We also show that our method outperforms the state-of-the-art
methods for fine-grained recognition. | ['Jakub Špaňhel', 'Adam Herout', 'Jakub Sochor'] | 2017-03-02 | null | null | null | null | ['vehicle-pose-estimation'] | ['computer-vision'] | [ 1.36260465e-01 -1.57147467e-01 9.00097415e-02 -4.60140586e-01
-5.78782439e-01 -6.62793279e-01 8.57612729e-01 -2.81996876e-01
-5.04021108e-01 5.03206909e-01 -1.94923267e-01 -2.54177719e-01
4.03328799e-02 -9.77553248e-01 -1.27070415e+00 -8.40379059e-01
2.16581136e-01 4.90825772e-01 6.01068974e-01 7.16464594e-03
1.87327817e-01 1.22634864e+00 -1.89508533e+00 3.59009415e-01
3.24140757e-01 1.33194757e+00 -1.57298997e-01 7.14942098e-01
2.74095479e-02 4.80062366e-01 -6.59447372e-01 -5.28090775e-01
6.59243524e-01 2.18015537e-01 -5.06939888e-01 5.85788667e-01
1.08259165e+00 -5.04816294e-01 -2.26620063e-01 1.04538476e+00
-1.10114954e-01 6.95636943e-02 7.42241800e-01 -1.28717291e+00
-2.33773798e-01 -2.42109954e-01 -4.86972094e-01 1.11105457e-01
-1.02135405e-01 -6.77766278e-02 3.51485401e-01 -8.62354219e-01
4.74474907e-01 1.30286348e+00 7.14750350e-01 5.84457517e-01
-9.23358917e-01 -8.13470721e-01 1.90196157e-01 3.15550804e-01
-1.74830973e+00 -5.23302019e-01 6.32500708e-01 -6.24256849e-01
7.92190850e-01 3.32332015e-01 5.41773140e-01 6.69106543e-01
2.05168083e-01 1.50312394e-01 1.03125870e+00 -3.61388564e-01
1.38040558e-01 1.20647490e-01 3.97967100e-01 9.13169503e-01
5.90285063e-01 2.49322265e-01 9.48013291e-02 -9.61285979e-02
8.67976010e-01 4.95186150e-01 1.62512451e-01 -6.11731470e-01
-1.15378463e+00 8.11640203e-01 3.42298061e-01 2.05238089e-01
-3.50711048e-01 2.08044022e-01 1.84541464e-01 4.44542468e-02
3.97274941e-01 -5.08533306e-02 -2.93311745e-01 1.40520409e-01
-9.63595271e-01 4.30976123e-01 7.53717422e-01 8.75008941e-01
1.17456365e+00 -6.49200529e-02 -3.35092366e-01 6.48112953e-01
1.47899732e-01 6.51323676e-01 3.71757895e-02 -9.92984891e-01
4.90833312e-01 6.74129367e-01 2.24720374e-01 -9.70536530e-01
-3.29861611e-01 -1.27658501e-01 -1.00561666e+00 6.24726713e-01
7.09844649e-01 -3.47357579e-02 -1.13184679e+00 1.30706871e+00
4.53770012e-01 2.27432892e-01 -2.27389574e-01 8.55434895e-01
8.19253922e-01 5.47969520e-01 1.94738749e-02 1.59148827e-01
1.58013380e+00 -1.00939178e+00 -2.07305416e-01 9.75497365e-02
2.32938707e-01 -6.19645715e-01 4.49969530e-01 2.73943782e-01
-6.53029859e-01 -7.87339568e-01 -1.03129554e+00 2.56373733e-01
-7.79584110e-01 2.11655691e-01 3.07985574e-01 9.24392521e-01
-1.08585989e+00 3.50799501e-01 -5.71520925e-01 -3.51384044e-01
5.62798917e-01 3.62989366e-01 -8.58035386e-01 -1.39448151e-01
-7.20793903e-01 9.13006186e-01 1.28763095e-01 2.09092706e-01
-1.09050894e+00 -4.02923286e-01 -9.01531518e-01 1.55373380e-01
3.59149009e-01 -6.48674548e-01 9.93592680e-01 -8.84733140e-01
-1.23757565e+00 9.34007227e-01 -2.36475199e-01 -5.58480859e-01
4.10593837e-01 9.12353024e-02 -2.08063141e-01 2.62765028e-02
-1.36279956e-01 6.68422520e-01 1.08882010e+00 -1.18770742e+00
-6.46255612e-01 -3.57132345e-01 4.59575236e-01 -2.82497644e-01
-6.38111457e-02 2.11395144e-01 -5.40770888e-01 -6.41112685e-01
-1.08455993e-01 -1.13899863e+00 -2.30475351e-01 2.24638119e-01
-9.98595506e-02 -1.62056491e-01 1.12959659e+00 -3.15024883e-01
6.70997322e-01 -2.02325082e+00 -3.97046685e-01 1.48252800e-01
3.70236099e-01 5.55387437e-01 -2.11159587e-01 -1.11830905e-02
-7.97426403e-02 1.38713524e-01 1.20456330e-03 -3.54669005e-01
7.82231838e-02 3.30830246e-01 -7.05968887e-02 7.42995620e-01
2.27011800e-01 8.55597317e-01 -3.11656833e-01 -4.31988925e-01
6.50736034e-01 5.88159204e-01 -4.12097812e-01 2.10131958e-01
6.38494715e-02 -7.97137469e-02 -4.83294278e-01 6.01702094e-01
1.17652929e+00 -1.83455758e-02 -2.72436619e-01 -4.16755736e-01
-2.54491687e-01 -1.28985092e-01 -1.24013221e+00 7.47794151e-01
-3.64323527e-01 6.86658084e-01 8.33826289e-02 -1.07396209e+00
9.19484496e-01 2.77099758e-01 2.81569898e-01 -6.35009468e-01
2.47786164e-01 -1.64811552e-01 -3.03736567e-01 -2.04625502e-01
5.33069313e-01 -5.98168783e-02 -1.45307571e-01 2.06900373e-01
-1.36285886e-01 3.54200043e-02 2.10973993e-01 -2.92758793e-01
8.53310049e-01 -2.85782784e-01 3.54212493e-01 -2.83477157e-01
7.62025058e-01 -5.13541624e-02 4.44708407e-01 7.08493829e-01
-3.06278050e-01 6.68866158e-01 2.99964935e-01 -1.00275362e+00
-1.10136497e+00 -8.49943280e-01 -2.22739339e-01 1.04676151e+00
1.51725575e-01 -6.19944669e-02 -1.26035869e+00 -6.91932261e-01
7.08217174e-02 1.20400533e-01 -1.01828325e+00 3.10579300e-01
-8.57412696e-01 -5.39216399e-01 5.86897910e-01 5.21706283e-01
7.81072438e-01 -8.75932634e-01 -5.12014031e-01 -1.33900002e-01
4.33243066e-02 -1.46843219e+00 -3.88078094e-01 -1.80785246e-02
-7.14925170e-01 -1.41045284e+00 -6.51355028e-01 -6.42435730e-01
9.83081460e-01 6.68981850e-01 9.72097456e-01 -1.49598410e-02
-1.79940775e-01 2.08976403e-01 -2.79968619e-01 -2.86350459e-01
-2.20309690e-01 -1.65014431e-01 -4.10826355e-02 4.65335637e-01
4.38238472e-01 -8.16454664e-02 -4.23243850e-01 6.65311575e-01
-8.48613739e-01 -6.12134300e-02 6.17629766e-01 6.81114018e-01
6.42578363e-01 1.90596417e-01 8.07285830e-02 -7.22618222e-01
6.57381788e-02 -1.74305975e-01 -1.22379041e+00 1.49545565e-01
-9.19863135e-02 1.11704180e-02 6.50945723e-01 -3.40222836e-01
-7.51295447e-01 1.80119231e-01 -2.12020829e-01 -8.00937831e-01
-8.01711619e-01 -3.71315658e-01 -2.69385785e-01 -7.26518393e-01
3.13427091e-01 -7.27197826e-02 -2.86401957e-01 -4.89703268e-01
3.22668314e-01 6.40031099e-01 3.20067763e-01 -3.51611406e-01
9.65508819e-01 5.74827552e-01 2.00661093e-01 -9.30432916e-01
-3.55243742e-01 -4.24403816e-01 -8.23747575e-01 -2.19834700e-01
1.12493467e+00 -7.86641955e-01 -1.08082914e+00 6.21043444e-01
-1.38346267e+00 -9.47548226e-02 -6.57449588e-02 3.03977817e-01
-3.29689056e-01 2.51198024e-01 -2.73856789e-01 -6.92939699e-01
-2.64749467e-01 -1.28353453e+00 1.32659829e+00 1.71105310e-01
2.63621032e-01 -6.31102502e-01 -2.37826467e-01 5.42415202e-01
5.62445641e-01 3.36381763e-01 6.35849595e-01 -3.81090373e-01
-1.06869900e+00 -3.67883861e-01 -5.31871259e-01 3.91941637e-01
-2.84609217e-02 1.11470982e-01 -1.07431173e+00 -6.11327030e-02
-2.86925822e-01 9.80411246e-02 1.00714195e+00 4.13248867e-01
1.24111271e+00 -4.15315837e-01 -5.42031109e-01 6.68433726e-01
1.26396954e+00 2.84897685e-01 7.49264657e-01 4.11939412e-01
7.69389033e-01 6.35069549e-01 4.42346156e-01 2.82336205e-01
4.08636332e-01 1.03135931e+00 6.05427742e-01 -1.63880810e-01
-3.27157170e-01 8.78710374e-02 1.07226476e-01 -3.98675539e-02
-3.41094464e-01 -2.31871918e-01 -6.49782419e-01 5.92909992e-01
-1.57546175e+00 -1.13357675e+00 2.95547489e-02 2.11286139e+00
1.20397814e-01 1.35620788e-01 1.62937909e-01 5.38576804e-02
8.57746959e-01 2.41876096e-01 -6.67454228e-02 -5.84738255e-01
7.62936100e-02 7.22262561e-02 9.65414524e-01 5.57715356e-01
-1.51374793e+00 9.67352092e-01 6.14992762e+00 9.21897829e-01
-1.26403081e+00 1.24867171e-01 6.55360818e-01 1.25399306e-01
3.98119576e-02 -3.72587204e-01 -1.23937333e+00 2.80168444e-01
7.95493960e-01 4.90336448e-01 4.19774890e-01 9.80288386e-01
1.25883132e-01 2.19968855e-02 -9.88284230e-01 1.01063931e+00
1.72480851e-01 -1.48464739e+00 1.34062365e-01 1.39470220e-01
8.12868416e-01 1.32461801e-01 -1.56166077e-01 9.42215919e-02
1.03417896e-01 -1.08051360e+00 8.34743381e-01 5.73399961e-01
7.85930753e-01 -6.75736785e-01 7.81544447e-01 4.19893712e-01
-1.46533692e+00 1.64265707e-01 -4.69599754e-01 -4.22598831e-02
-1.65756688e-01 4.76754516e-01 -9.55736518e-01 1.97426870e-01
8.93950999e-01 1.78583086e-01 -6.52315915e-01 9.37862754e-01
2.23070055e-01 3.81298542e-01 -2.60979623e-01 -4.40918207e-02
2.85356671e-01 -1.79644395e-02 3.23670954e-01 1.43217933e+00
2.30665803e-01 1.29454464e-01 7.36605749e-02 5.34731150e-01
-1.67373732e-01 -2.96804160e-01 -7.62909651e-01 3.50714922e-01
2.15578243e-01 1.37614369e+00 -6.96602881e-01 -6.32114232e-01
-6.59530938e-01 6.36715412e-01 1.67427912e-01 3.10598642e-01
-9.21158016e-01 -4.21938211e-01 9.92019773e-01 3.55912745e-01
1.01060152e+00 -2.09839180e-01 9.80126485e-03 -9.74861979e-01
2.05323860e-01 -7.84553528e-01 4.51108813e-02 -6.53278112e-01
-1.06149495e+00 1.04101992e+00 1.90060869e-01 -1.23412883e+00
-3.64413589e-01 -9.63835180e-01 -2.96532065e-01 7.74234295e-01
-1.61519241e+00 -1.24631357e+00 -4.78316486e-01 7.79798269e-01
4.41073418e-01 -2.27740835e-02 7.75741279e-01 4.39192742e-01
-3.29926223e-01 6.11906528e-01 -1.67348385e-02 4.17076945e-01
3.29223007e-01 -7.37782240e-01 6.77130401e-01 8.46926510e-01
-7.58165568e-02 4.69559997e-01 4.95839864e-01 -3.44600677e-01
-1.23596442e+00 -1.49889600e+00 1.01977563e+00 -5.87893784e-01
3.44317913e-01 -7.33956277e-01 -6.01603925e-01 6.84755862e-01
-1.08111668e-02 6.89399242e-01 2.47257456e-01 -2.79748946e-01
-6.65403843e-01 -6.50398135e-01 -1.51821744e+00 2.84100235e-01
8.14606547e-01 -4.31519300e-01 -5.81315160e-01 9.37616080e-02
4.39240694e-01 -2.53102064e-01 -6.16477609e-01 5.47392249e-01
9.30339813e-01 -9.96297717e-01 1.08972645e+00 -5.07189035e-01
-2.02478588e-01 -6.88991606e-01 -3.95798951e-01 -9.72006500e-01
-4.55599099e-01 -1.45396516e-01 1.27632767e-01 9.18192029e-01
2.83217616e-02 -8.35392594e-01 9.03406918e-01 5.12005389e-01
-1.72941595e-01 -5.59216380e-01 -1.16421175e+00 -6.64068580e-01
-1.00761771e-01 -6.67396188e-01 8.21954846e-01 3.01988810e-01
-7.09870577e-01 -1.44084230e-01 -3.10888350e-01 5.03962755e-01
6.69605911e-01 3.89151871e-01 1.06649852e+00 -1.33425689e+00
4.56753857e-02 -4.27664578e-01 -8.55115950e-01 -1.01173317e+00
1.67496905e-01 -1.81026265e-01 8.83742198e-02 -1.25312996e+00
3.79397392e-01 -3.48676264e-01 -2.02773944e-01 4.76767868e-01
2.96116501e-01 9.70685005e-01 3.44204068e-01 -1.26162454e-01
-7.67574310e-01 1.27240404e-01 1.21009839e+00 -2.78993011e-01
4.03914481e-01 1.82913035e-01 -5.18335521e-01 8.28326821e-01
6.27413929e-01 -3.44080120e-01 1.38243675e-01 -3.41958672e-01
-3.80855531e-01 -1.21901967e-02 8.70647967e-01 -1.00439489e+00
1.93025857e-01 -1.96308553e-01 5.37741601e-01 -8.56363893e-01
5.82339287e-01 -1.24407935e+00 4.21922535e-01 2.13472620e-01
2.28823736e-01 2.28498243e-02 3.88782263e-01 1.92069292e-01
-2.34673679e-01 -7.33532235e-02 8.86387765e-01 -1.68124691e-01
-7.33824611e-01 4.41930652e-01 -5.14149904e-01 -4.33135837e-01
1.24139488e+00 -5.56114614e-01 -4.36636895e-01 2.39557261e-03
-4.27821606e-01 -3.13863009e-01 5.95583856e-01 4.50472862e-01
4.38377976e-01 -1.49190414e+00 -6.18895471e-01 6.20362103e-01
2.28500113e-01 -2.92940855e-01 4.25862670e-01 5.29693305e-01
-6.93684876e-01 1.12674975e+00 -3.17767024e-01 -5.35343170e-01
-1.48296213e+00 8.52692425e-01 4.60124493e-01 -1.72706291e-01
-3.11984181e-01 4.81000930e-01 6.72266424e-01 -3.37465286e-01
1.52965441e-01 -7.01583624e-01 -4.00469393e-01 -1.92724496e-01
7.05094934e-01 4.32382196e-01 4.70410943e-01 -1.21109343e+00
-6.16659105e-01 1.12518287e+00 -3.92018035e-02 3.11889589e-01
1.27861464e+00 4.00473177e-02 7.15119541e-02 -2.21072435e-01
1.25379646e+00 1.14078254e-01 -1.43213260e+00 1.61960088e-02
-5.02356172e-01 -6.50221229e-01 -1.24572523e-01 -4.24587131e-01
-1.23236442e+00 8.67559016e-01 7.17370093e-01 3.15038711e-01
1.14180827e+00 1.26563817e-01 6.30379260e-01 5.20284176e-01
4.98285264e-01 -5.34274101e-01 -3.71688694e-01 6.88056052e-01
7.40697920e-01 -1.34084260e+00 -1.76873282e-01 -5.15129507e-01
-1.65516943e-01 1.28066540e+00 4.63416219e-01 -3.73975307e-01
6.76533937e-01 4.03706402e-01 1.04901083e-01 -2.18590349e-01
-5.36454499e-01 -2.89152235e-01 6.17712379e-01 6.56304598e-01
3.96027714e-02 2.36726865e-01 4.46636081e-02 3.65298122e-01
-9.68634337e-03 -4.61218357e-02 2.79781371e-01 4.89096671e-01
-6.32253349e-01 -9.16367650e-01 -9.16732073e-01 1.77383721e-01
-3.84448379e-01 8.79542828e-02 -1.33089662e-01 8.56917858e-01
7.20425129e-01 9.93839681e-01 3.27771544e-01 -3.19455177e-01
4.82036322e-01 -2.41215840e-01 5.16979694e-01 -3.11169595e-01
-3.82519662e-01 -1.78135097e-01 1.56662583e-01 -7.17662334e-01
-3.78125817e-01 -6.29909277e-01 -7.97906816e-01 -5.18555343e-01
-9.19319168e-02 1.51895743e-03 7.74355948e-01 1.01501799e+00
1.87167138e-01 3.40087861e-01 5.96733630e-01 -1.23670042e+00
-3.71856913e-02 -7.58077800e-01 -3.86477798e-01 1.34788081e-01
8.24075699e-01 -9.50267375e-01 -2.81768441e-01 2.57990479e-01] | [8.221046447753906, -0.8083351254463196] |
6309bdc2-698d-4232-814a-9ca48dd2d23d | improving-the-modality-representation-with | 2210.15824 | null | https://arxiv.org/abs/2210.15824v3 | https://arxiv.org/pdf/2210.15824v3.pdf | Improving the Modality Representation with Multi-View Contrastive Learning for Multimodal Sentiment Analysis | Modality representation learning is an important problem for multimodal sentiment analysis (MSA), since the highly distinguishable representations can contribute to improving the analysis effect. Previous works of MSA have usually focused on multimodal fusion strategies, and the deep study of modal representation learning was given less attention. Recently, contrastive learning has been confirmed effective at endowing the learned representation with stronger discriminate ability. Inspired by this, we explore the improvement approaches of modality representation with contrastive learning in this study. To this end, we devise a three-stages framework with multi-view contrastive learning to refine representations for the specific objectives. At the first stage, for the improvement of unimodal representations, we employ the supervised contrastive learning to pull samples within the same class together while the other samples are pushed apart. At the second stage, a self-supervised contrastive learning is designed for the improvement of the distilled unimodal representations after cross-modal interaction. At last, we leverage again the supervised contrastive learning to enhance the fused multimodal representation. After all the contrast trainings, we next achieve the classification task based on frozen representations. We conduct experiments on three open datasets, and results show the advance of our model. | ['Limin Sun', 'Hongsong Zhu', 'Yimo Ren', 'Jie Liu', 'Hong Li', 'Xin Zheng', 'Peipei Liu'] | 2022-10-28 | null | null | null | null | ['multimodal-sentiment-analysis', 'multimodal-sentiment-analysis'] | ['computer-vision', 'natural-language-processing'] | [ 3.92260045e-01 -1.67463616e-01 -9.08003822e-02 -3.26010555e-01
-9.21117783e-01 -2.81351417e-01 7.79835880e-01 1.72394022e-01
-1.60049453e-01 4.51797068e-01 5.39734662e-01 9.33775082e-02
-1.06557868e-01 -6.11910105e-01 -5.51878989e-01 -1.18393004e+00
4.14194882e-01 1.53910384e-01 -8.24407861e-02 -6.31220877e-01
3.51728946e-01 1.72304604e-02 -1.56972992e+00 7.42551804e-01
9.90126669e-01 1.02276087e+00 -6.85586482e-02 2.52253383e-01
-4.28758204e-01 8.37245166e-01 -1.98892564e-01 -4.99640077e-01
-1.86852232e-01 -6.10298395e-01 -7.47019410e-01 6.64214715e-02
2.13534627e-02 5.26730604e-02 -7.10827634e-02 1.02273583e+00
6.17822886e-01 3.35382283e-01 7.82746017e-01 -1.25511146e+00
-5.63666701e-01 9.83266592e-01 -1.05722558e+00 -3.95793058e-02
4.58279729e-01 8.24778080e-02 1.06449580e+00 -1.20312178e+00
2.35039622e-01 1.45227575e+00 2.27714255e-01 5.17650187e-01
-9.57177520e-01 -6.38763130e-01 4.69925582e-01 2.38923714e-01
-1.00531840e+00 -4.75590855e-01 1.36825716e+00 -2.53121287e-01
1.44832045e-01 1.67443588e-01 4.13828671e-01 1.17082870e+00
-2.42155164e-01 1.12611520e+00 1.13954425e+00 -4.37562466e-01
-1.78246245e-01 3.17255169e-01 3.95138055e-01 5.60140848e-01
-2.27824375e-01 -1.53978586e-01 -4.55196053e-01 1.58189133e-01
2.16018841e-01 4.39799130e-01 -3.10074359e-01 -2.95725822e-01
-1.21485317e+00 8.96893084e-01 7.28733063e-01 6.17531478e-01
-3.03016186e-01 -1.94060490e-01 5.14308512e-01 2.83110976e-01
4.37812030e-01 2.30557367e-01 -2.38152668e-01 1.25074714e-01
-6.57738388e-01 -8.42493400e-02 1.20835930e-01 3.36806029e-01
9.76167023e-01 -1.09522454e-01 -2.88984090e-01 1.07767081e+00
5.63397348e-01 3.80311012e-01 7.20884442e-01 -4.32457805e-01
7.18305588e-01 1.08371198e+00 -4.05301154e-01 -1.04109013e+00
-5.07887483e-01 -3.78396928e-01 -1.16117084e+00 7.57509544e-02
1.37960300e-01 -2.43317217e-01 -8.04103434e-01 1.72126043e+00
2.75692552e-01 3.74294370e-02 4.32897061e-01 9.12959456e-01
1.09890473e+00 8.44982326e-01 1.35680392e-01 -3.01176935e-01
1.54293168e+00 -1.16589379e+00 -7.56783485e-01 6.16704226e-02
6.07429445e-01 -7.77018547e-01 8.95646036e-01 3.52744937e-01
-1.06746924e+00 -7.82574356e-01 -1.14847708e+00 -1.39174135e-02
-5.76617658e-01 1.19663835e-01 4.82181281e-01 3.47320557e-01
-5.30739903e-01 3.06652546e-01 -5.23536921e-01 -8.85463133e-02
4.84215349e-01 1.72538087e-01 -5.36929309e-01 -1.58099934e-01
-1.40813601e+00 7.48656511e-01 5.95597863e-01 5.33295929e-01
-5.66730499e-01 -4.30056781e-01 -1.00108469e+00 5.37837520e-02
3.06440741e-01 -5.91552317e-01 7.80679345e-01 -1.43054855e+00
-1.48557341e+00 6.54351652e-01 -3.02745663e-02 2.88804471e-02
1.94800436e-01 -1.76027883e-02 -5.53093374e-01 1.85752362e-01
-1.17335394e-01 7.28231609e-01 1.05002618e+00 -1.80616772e+00
-7.04698563e-01 -5.86284101e-01 3.44248354e-01 5.65392137e-01
-6.23297393e-01 -1.62831470e-01 -2.62030184e-01 -5.16874850e-01
2.57001907e-01 -6.68283105e-01 -3.12251672e-02 -5.43048561e-01
-3.39028418e-01 -2.13388816e-01 7.02480078e-01 -6.46714449e-01
1.30037034e+00 -2.39332032e+00 8.55037451e-01 2.82858968e-01
2.30992183e-01 1.07859291e-01 -3.32007736e-01 3.15419823e-01
-3.22221935e-01 6.54644221e-02 -4.53155458e-01 -5.88145256e-01
-1.33695896e-03 -1.24009036e-01 -2.59366006e-01 9.80200917e-02
5.35151541e-01 8.58668029e-01 -7.62620926e-01 -6.17955565e-01
1.91265136e-01 4.04830605e-01 -5.02571464e-01 2.86435127e-01
8.57063308e-02 6.78743899e-01 -4.05159831e-01 7.26858139e-01
1.08196902e+00 -1.17803387e-01 1.92928776e-01 -7.43174851e-01
1.93947889e-02 -2.37212881e-01 -1.16886449e+00 1.88843775e+00
-5.21307707e-01 1.87920198e-01 8.30585882e-02 -1.38623559e+00
8.75851452e-01 9.43106785e-03 4.03889626e-01 -8.35891604e-01
3.76080006e-01 1.72300786e-01 -2.21961811e-02 -6.19115114e-01
6.48535609e-01 -4.36574370e-01 -2.06496418e-01 3.33296090e-01
1.60687238e-01 7.72016346e-02 -3.25384550e-02 2.35784873e-01
2.68779129e-01 3.54014069e-01 -1.18364440e-02 1.99452966e-01
9.97317076e-01 -2.92594850e-01 1.64433390e-01 1.70763955e-01
-4.65094447e-02 5.47002375e-01 5.56045532e-01 -1.08239517e-01
-4.69040692e-01 -9.56940472e-01 -6.45615608e-02 1.42717385e+00
5.39619565e-01 -2.05737203e-01 -4.71141309e-01 -8.85853171e-01
-2.81692028e-01 3.87325853e-01 -9.59519327e-01 -5.86001098e-01
-4.24028248e-01 -9.51293051e-01 1.97701797e-01 6.66779697e-01
7.36664951e-01 -1.00159538e+00 -1.52153075e-01 -1.31653354e-01
-4.24293011e-01 -5.20913482e-01 -1.50454342e-01 2.67155647e-01
-7.59244144e-01 -8.70239198e-01 -1.05174315e+00 -1.04124141e+00
7.15096474e-01 4.18577641e-01 6.32122278e-01 7.58292675e-02
4.06847268e-01 4.21969920e-01 -6.02239311e-01 -2.05520809e-01
-1.77525610e-01 2.70646185e-01 -2.69955516e-01 6.13726914e-01
1.80685297e-01 -4.87943143e-01 -5.58870196e-01 1.31840602e-01
-1.17938983e+00 1.10186622e-01 6.89686239e-01 1.14963257e+00
4.30117399e-01 -1.09581165e-01 7.10051954e-01 -6.96704745e-01
6.64118409e-01 -7.45081544e-01 1.55454571e-03 2.66586900e-01
-1.64084658e-01 1.86074346e-01 5.01507342e-01 -5.65487802e-01
-1.46174550e+00 -6.38849661e-02 -2.06161126e-01 -4.94353980e-01
4.96168621e-03 9.79108036e-01 -4.01174694e-01 1.67879343e-01
3.49364966e-01 2.80298650e-01 3.41153413e-01 -2.24219486e-01
6.55571938e-01 7.22727954e-01 2.54974931e-01 -5.10639131e-01
6.03023291e-01 3.97172123e-01 -1.73465163e-01 -4.45358336e-01
-7.70791173e-01 -2.80294299e-01 -4.75455284e-01 -2.83898503e-01
9.36877012e-01 -8.57885897e-01 -7.25928605e-01 5.11941314e-01
-8.56125593e-01 1.91372260e-01 -5.41857406e-02 4.66201901e-01
-2.21070796e-01 4.80238348e-01 -2.65240103e-01 -7.74141192e-01
-1.67574018e-01 -1.38666272e+00 1.23951876e+00 7.08712876e-01
2.04834014e-01 -1.05049217e+00 2.51216382e-01 5.62093616e-01
3.11887920e-01 1.36252388e-01 9.28221881e-01 -7.50189185e-01
-4.75514419e-02 -3.35307270e-02 -2.44884163e-01 3.36264580e-01
1.73939720e-01 1.77429663e-03 -1.26639545e+00 -2.61810392e-01
-8.04991275e-02 -5.94190598e-01 1.38656890e+00 1.56784117e-01
1.06068146e+00 1.25409484e-01 -3.33070308e-01 3.89945775e-01
1.03256869e+00 1.17067464e-01 7.23130763e-01 4.15810019e-01
7.80015647e-01 8.31016541e-01 8.00439179e-01 2.59033740e-01
5.11324883e-01 3.93635303e-01 6.11302674e-01 -2.66159713e-01
1.77277401e-01 -1.65237904e-01 3.83981824e-01 1.05938005e+00
-2.62754023e-01 1.47114589e-03 -5.57731986e-01 2.79367894e-01
-1.99028707e+00 -1.11879599e+00 1.46010175e-01 1.90028560e+00
5.38972139e-01 5.01182191e-02 2.32415676e-01 2.67690271e-01
8.56379628e-01 4.43831414e-01 -3.22010875e-01 -1.32168517e-01
-3.03886980e-01 -1.64655149e-01 -4.51335877e-01 2.47188851e-01
-1.29053450e+00 7.37092614e-01 4.86145782e+00 8.25845480e-01
-1.27888477e+00 -1.14571281e-01 7.21315742e-01 9.68592241e-02
-7.93806612e-01 -2.36973520e-02 -4.39472169e-01 4.61393476e-01
4.78213906e-01 2.46062428e-01 4.01542038e-01 3.77921939e-01
-2.42614031e-01 -2.03796220e-03 -8.61860037e-01 1.28924525e+00
3.74963164e-01 -1.06042075e+00 4.40403938e-01 -1.80770591e-01
7.10122943e-01 -5.25823832e-01 3.81452709e-01 7.99416006e-01
-2.53719985e-01 -8.43418658e-01 4.59401399e-01 9.67019022e-01
3.14274490e-01 -1.09711993e+00 1.09231842e+00 2.29847625e-01
-1.35806727e+00 -2.80861795e-01 -2.83846229e-01 2.02440053e-01
1.27866924e-01 2.22109973e-01 -7.88876787e-02 1.27194607e+00
4.34286475e-01 9.82549310e-01 -7.70538449e-01 6.03659928e-01
2.62110587e-02 1.61164701e-01 2.03456491e-01 3.05075776e-02
2.26640299e-01 -4.35315341e-01 3.15664411e-01 1.02444410e+00
1.57356128e-01 3.54462638e-02 9.22822952e-02 5.86694598e-01
-2.13449374e-01 3.65192741e-01 -4.82280999e-01 -5.79615049e-02
1.33415699e-01 1.59136057e+00 -4.49380040e-01 -3.79048765e-01
-4.53326046e-01 8.37027371e-01 4.86939132e-01 2.63643831e-01
-9.94097650e-01 -3.94617409e-01 9.96003821e-02 -4.14292872e-01
3.16417873e-01 1.95003375e-01 -2.51591504e-01 -1.38178122e+00
-1.17191397e-01 -1.07706940e+00 7.38021791e-01 -7.41536140e-01
-1.56944835e+00 6.09553337e-01 -2.04653237e-02 -1.59340084e+00
-1.83271244e-02 -5.47510743e-01 -8.12162101e-01 9.19852197e-01
-1.74310076e+00 -1.58019185e+00 -3.55512291e-01 7.85269856e-01
4.20064360e-01 -3.26714128e-01 6.01736963e-01 3.33823085e-01
-7.61768281e-01 7.12385297e-01 -1.51331723e-01 -4.56905477e-02
8.86792779e-01 -9.58383024e-01 -7.08022058e-01 5.76499164e-01
-1.08267300e-01 7.33027756e-01 3.07406098e-01 -3.22656572e-01
-1.39862001e+00 -8.50602090e-01 2.79494673e-01 -1.85328677e-01
5.72733581e-01 6.38948306e-02 -1.10896027e+00 4.32750583e-01
6.50781631e-01 -1.92690521e-01 9.55932200e-01 4.02249128e-01
-4.01415408e-01 -3.30664128e-01 -8.58462274e-01 5.82550406e-01
5.13089538e-01 -5.92089832e-01 -8.59222770e-01 -1.16346940e-01
6.47272885e-01 -1.76352501e-01 -9.34122205e-01 6.43156052e-01
6.66339159e-01 -7.71042585e-01 7.86700189e-01 -7.72647262e-01
9.33469653e-01 -4.47201550e-01 -2.94056714e-01 -1.55556476e+00
-2.96296895e-01 -7.30507076e-02 4.62618656e-02 1.76020539e+00
5.33581436e-01 -5.38371921e-01 2.29010612e-01 9.94765311e-02
-2.46561185e-01 -7.44744122e-01 -6.20545089e-01 -9.33385920e-03
2.44400993e-01 -1.19469471e-01 6.09787881e-01 1.16126680e+00
4.54184800e-01 7.67389417e-01 -4.66042906e-01 9.70617756e-02
2.49920323e-01 4.92396057e-01 7.59365380e-01 -1.06962144e+00
-1.78496048e-01 -7.26626039e-01 -1.51331842e-01 -9.44193602e-01
2.01047078e-01 -1.01852930e+00 -1.22033276e-01 -1.44045639e+00
6.30978823e-01 -5.00219576e-02 -7.10655868e-01 4.16470587e-01
-7.39390612e-01 1.65721580e-01 3.55908453e-01 1.09739736e-01
-8.41053188e-01 1.05288088e+00 1.46213186e+00 -6.18558466e-01
-2.51285732e-01 -1.26747191e-01 -1.18214631e+00 5.83078384e-01
5.59970617e-01 3.33651118e-02 -4.86964166e-01 -2.41904348e-01
2.45250359e-01 9.69363190e-03 2.04472229e-01 -7.07100689e-01
3.26371640e-02 3.69117269e-03 5.31591833e-01 -7.09501505e-01
4.39049900e-01 -9.06603575e-01 -2.83462703e-01 2.98821896e-01
-4.84643877e-01 -1.00600481e-01 2.72797585e-01 4.47758079e-01
-6.89213991e-01 -1.89000219e-01 7.86486745e-01 1.17584161e-01
-7.30612695e-01 1.23849034e-01 -8.45095739e-02 -8.53752270e-02
8.61086547e-01 -2.26328261e-02 -4.98212367e-01 -2.28606433e-01
-8.94688070e-01 5.15120983e-01 5.58518842e-02 4.87309456e-01
6.94885194e-01 -1.69661820e+00 -6.83612287e-01 1.06710747e-01
3.76789898e-01 -1.79277837e-01 7.20544815e-01 1.15592122e+00
1.29081517e-01 -1.67476892e-01 -4.39888000e-01 -7.62256861e-01
-1.06909525e+00 8.67461503e-01 2.83387512e-01 -3.28453451e-01
-2.52530500e-02 5.40680170e-01 3.70648414e-01 -6.02288187e-01
7.47452956e-03 -2.12023184e-02 -1.07773721e+00 8.17065835e-01
5.87352455e-01 3.31435412e-01 -9.93344784e-02 -7.87427545e-01
-3.83738995e-01 7.70905912e-01 -3.80754352e-01 -1.55150026e-01
1.27651179e+00 -2.85642952e-01 -2.13232175e-01 7.19348907e-01
1.34315336e+00 4.38355058e-02 -9.12304044e-01 -1.18766777e-01
-3.82318795e-01 -9.51071605e-02 -9.96399596e-02 -6.00002289e-01
-1.22709239e+00 1.21596992e+00 7.62832701e-01 2.83600241e-01
1.51248443e+00 -8.66509881e-03 5.11460483e-01 1.29277527e-01
-1.13091342e-01 -9.06751096e-01 5.04441142e-01 4.35093611e-01
8.89912903e-01 -1.55405164e+00 -2.05437347e-01 -2.30417594e-01
-1.05741990e+00 1.21587217e+00 7.05155075e-01 5.87850399e-02
4.74670023e-01 -1.56811967e-01 5.42754419e-02 -2.18313724e-01
-3.99033487e-01 -3.60557556e-01 4.46074486e-01 3.68353486e-01
4.62320715e-01 -6.21584132e-02 -3.48717868e-01 1.00317764e+00
2.77636945e-01 -3.01747590e-01 4.03760262e-02 8.82644057e-01
-3.73841316e-01 -9.66379881e-01 -4.87380207e-01 1.97565332e-01
-1.15188006e-02 -2.55906098e-02 -3.49123299e-01 5.37307382e-01
2.07245037e-01 1.04375231e+00 -6.26265556e-02 -7.43502319e-01
4.44168419e-01 1.78073093e-01 4.65833247e-01 -2.19401762e-01
-6.74564004e-01 3.55568022e-01 -3.95466574e-02 -2.04696342e-01
-9.15281057e-01 -6.34390891e-01 -1.30170548e+00 -1.70648135e-02
-3.47687215e-01 2.40413800e-01 3.84779304e-01 1.10097301e+00
1.91054448e-01 8.64176929e-01 1.15248299e+00 -1.11991096e+00
-3.75104338e-01 -1.03524268e+00 -3.57634962e-01 7.48203993e-01
3.48649651e-01 -8.85748267e-01 -4.74536270e-01 -1.40115947e-01] | [13.077596664428711, 5.007893085479736] |
af6fb4fd-d802-4813-b4f6-25eec9582ba8 | benchmarking-the-impact-of-noise-on-deep | 2303.13915 | null | https://arxiv.org/abs/2303.13915v1 | https://arxiv.org/pdf/2303.13915v1.pdf | Benchmarking the Impact of Noise on Deep Learning-based Classification of Atrial Fibrillation in 12-Lead ECG | Electrocardiography analysis is widely used in various clinical applications and Deep Learning models for classification tasks are currently in the focus of research. Due to their data-driven character, they bear the potential to handle signal noise efficiently, but its influence on the accuracy of these methods is still unclear. Therefore, we benchmark the influence of four types of noise on the accuracy of a Deep Learning-based method for atrial fibrillation detection in 12-lead electrocardiograms. We use a subset of a publicly available dataset (PTBXL) and use the metadata provided by human experts regarding noise for assigning a signal quality to each electrocardiogram. Furthermore, we compute a quantitative signal-to-noise ratio for each electrocardiogram. We analyze the accuracy of the Deep Learning model with respect to both metrics and observe that the method can robustly identify atrial fibrillation, even in cases signals are labelled by human experts as being noisy on multiple leads. False positive and false negative rates are slightly worse for data being labelled as noisy. Interestingly, data annotated as showing baseline drift noise results in an accuracy very similar to data without. We conclude that the issue of processing noisy electrocardiography data can be addressed successfully by Deep Learning methods that might not need preprocessing as many conventional methods do. | ['Nicolai Spicher', 'Dagmar Krefting', 'Henning Dathe', 'Ennio Idrobo-Avila', 'Philip Gemke', 'Theresa Bender'] | 2023-03-24 | null | null | null | null | ['atrial-fibrillation-detection'] | ['medical'] | [ 1.23014741e-01 -2.59214312e-01 3.20276380e-01 -5.71968675e-01
-1.11719787e+00 -7.03737080e-01 2.03969866e-01 4.96541977e-01
-7.49907017e-01 7.01615393e-01 4.19663489e-02 -3.95374358e-01
-3.64068866e-01 -5.90856493e-01 -4.30906534e-01 -8.80646765e-01
-3.09574515e-01 3.82808775e-01 -3.75875831e-01 1.29393294e-01
-2.92836763e-02 4.97568429e-01 -8.75998914e-01 4.62362587e-01
5.33200622e-01 1.15615201e+00 -4.87005919e-01 8.17439795e-01
3.08618307e-01 5.00452638e-01 -1.21264708e+00 -1.65783450e-01
3.18616003e-01 -6.39003754e-01 -5.83552182e-01 -1.78427055e-01
1.78962097e-01 -2.15391070e-01 1.63477119e-02 7.97480881e-01
1.34054160e+00 -4.86243635e-01 5.95591068e-01 -6.16459131e-01
1.61694065e-01 7.35723197e-01 -1.99039310e-01 8.47009480e-01
2.35364243e-01 4.88286167e-01 5.41171014e-01 -6.76683068e-01
4.90716368e-01 5.97977519e-01 1.26913822e+00 9.93843675e-02
-1.64375210e+00 -4.25562769e-01 -3.04747373e-01 2.30185408e-02
-1.42280805e+00 -4.35895085e-01 7.28316247e-01 -6.75969899e-01
6.77156389e-01 2.61881381e-01 7.08450496e-01 1.19217217e+00
4.67306942e-01 2.74373710e-01 1.11761820e+00 -2.96298414e-01
4.39799696e-01 -1.06104203e-01 3.21659833e-01 8.42190310e-02
3.33303630e-01 1.93274900e-01 -2.83749610e-01 -4.21180010e-01
5.33380330e-01 -1.78959906e-01 -4.36005563e-01 -1.01703905e-01
-1.44911146e+00 5.82250416e-01 1.78365231e-01 6.36835277e-01
-6.73383415e-01 1.90594673e-01 7.82158017e-01 6.78667963e-01
4.07071084e-01 7.49947965e-01 -6.37150109e-01 -4.22296852e-01
-1.13889217e+00 4.55850601e-01 7.50670254e-01 2.56202608e-01
2.14490727e-01 1.41580209e-01 -4.64756817e-01 6.00559354e-01
-1.54584184e-01 3.68238956e-01 5.22117317e-01 -9.23720300e-01
1.83424845e-01 3.14533740e-01 1.60468981e-01 -8.02556157e-01
-1.01136851e+00 -9.22124326e-01 -1.19191027e+00 2.68920541e-01
6.64406538e-01 -5.19490659e-01 -8.84097219e-01 1.36446869e+00
-1.58528492e-01 -1.09460108e-01 -6.97377622e-02 1.09388340e+00
7.64981449e-01 1.04393058e-01 1.80129409e-01 -4.21427459e-01
1.05463505e+00 1.14448123e-01 -9.85872805e-01 4.07755896e-02
7.34154761e-01 -6.22463882e-01 8.34825039e-01 9.26789343e-01
-9.89179015e-01 -5.76867521e-01 -9.35055614e-01 3.36220324e-01
-2.40109507e-02 8.27454776e-02 2.05174729e-01 1.05889213e+00
-1.03445816e+00 9.89569604e-01 -9.75880265e-01 -1.11484760e-02
7.74601102e-01 4.77617472e-01 -2.25257948e-01 3.55414093e-01
-1.34136462e+00 8.90738726e-01 7.33857900e-02 3.21491539e-01
-8.65854919e-01 -6.93196058e-01 -5.98987877e-01 3.43810255e-03
-1.23012606e-02 -5.27199149e-01 1.08811116e+00 -9.71631587e-01
-9.97302949e-01 9.39127266e-01 2.55138814e-01 -8.03759694e-01
1.03557885e+00 -3.16776693e-01 -5.01699746e-01 -1.05091311e-01
-3.93459052e-02 -2.73441710e-02 9.67130303e-01 -1.02681446e+00
-3.44739616e-01 -5.00265718e-01 -2.10273653e-01 -3.19891900e-01
2.00469628e-01 -1.27860516e-01 6.45119175e-02 -9.72081065e-01
2.42462799e-01 -7.15146005e-01 -3.85784805e-01 -2.66354203e-01
-1.92031741e-01 1.25810564e-01 3.50322366e-01 -6.88204527e-01
1.45724630e+00 -2.21220684e+00 -1.26150936e-01 6.34477913e-01
5.15736461e-01 5.14823854e-01 3.38367671e-02 2.17740729e-01
-2.36803308e-01 4.87780541e-01 -5.46199858e-01 -1.97747331e-02
-3.55555415e-01 1.51137665e-01 -2.96351220e-02 7.54656374e-01
2.46257439e-01 8.26694965e-01 -6.93244696e-01 -5.35194501e-02
2.60711104e-01 5.47938645e-01 -1.79132029e-01 3.52767520e-02
3.14904779e-01 6.49745345e-01 4.83100815e-03 3.27605486e-01
5.28358161e-01 -7.72058405e-03 3.69295806e-01 -3.75643522e-01
2.44280636e-01 3.45108032e-01 -1.22487569e+00 1.45871186e+00
-3.42942983e-01 7.13261783e-01 -1.12055272e-01 -1.14733636e+00
9.79987264e-01 6.33112848e-01 6.12218499e-01 -5.62325001e-01
1.44079134e-01 2.04912618e-01 5.85196435e-01 -5.74454010e-01
-2.29712829e-01 -4.52890277e-01 2.01531202e-01 5.28990328e-01
-1.02959327e-01 4.72070798e-02 -1.63270503e-01 -1.39125302e-01
1.43580937e+00 -1.47328258e-01 2.49740824e-01 -5.06955802e-01
2.54941493e-01 -2.85604179e-01 7.44135559e-01 1.33409798e+00
-3.82678300e-01 1.03856575e+00 6.93988442e-01 -9.80327249e-01
-8.20769250e-01 -9.71328676e-01 -5.81322253e-01 2.94775873e-01
-3.44196558e-01 -4.07357126e-01 -7.07970083e-01 -7.60352790e-01
2.22511869e-02 2.63004988e-01 -6.73051059e-01 -2.15633780e-01
-6.83333576e-01 -1.32329500e+00 9.83685553e-01 7.94971228e-01
3.61340074e-03 -1.24733281e+00 -1.19744134e+00 5.66210151e-01
-2.39599645e-01 -9.15646970e-01 6.37478009e-02 7.01576948e-01
-1.05902421e+00 -1.25779653e+00 -7.59318531e-01 -2.77608901e-01
2.14093983e-01 -5.90813935e-01 1.48714149e+00 3.27459902e-01
-5.73626757e-01 2.90325969e-01 -3.80349338e-01 -6.80642247e-01
-5.65896332e-01 -7.84936845e-02 1.63570195e-01 6.08152710e-02
5.27113259e-01 -4.74556535e-01 -7.79319644e-01 8.19525272e-02
-8.32989097e-01 -6.27032042e-01 4.12792027e-01 7.73119390e-01
4.36440766e-01 -2.76660677e-02 1.10820806e+00 -1.12178183e+00
1.08289993e+00 -2.85496116e-01 -3.04690272e-01 -3.08342844e-01
-6.36244118e-01 -1.18216731e-01 5.69736362e-01 -1.80871516e-01
-3.64868790e-01 2.28846341e-01 -4.24574077e-01 -1.45101890e-01
-5.18229902e-01 6.12472892e-01 -1.59378629e-02 1.67738885e-01
1.08503568e+00 -1.62264287e-01 4.86832559e-02 -4.13731843e-01
-1.59916267e-01 6.98707759e-01 3.61146957e-01 -2.14792788e-01
3.76030654e-01 4.52931732e-01 5.27416728e-02 -7.59201765e-01
-5.59039474e-01 -4.30688560e-01 -6.64982557e-01 -1.88513339e-01
7.39448488e-01 -9.04533327e-01 -4.67965662e-01 5.35148919e-01
-1.13675320e+00 -2.55230248e-01 -5.10655940e-01 4.65467274e-01
-3.93238485e-01 4.66076136e-01 -3.66438210e-01 -9.20372069e-01
-5.89637876e-01 -1.09188402e+00 9.13867116e-01 -5.86949646e-01
-7.32847273e-01 -8.48933578e-01 5.42927682e-02 -4.44747545e-02
4.59911168e-01 8.26822817e-01 9.26561296e-01 -8.59726369e-01
1.57066926e-01 -4.09024864e-01 1.47531033e-01 7.57941604e-01
2.47930944e-01 -3.05926055e-01 -1.35224104e+00 -2.54797339e-01
5.16646445e-01 1.06822252e-01 9.32577372e-01 8.47979248e-01
1.21292114e+00 3.28318290e-02 -4.79749404e-02 4.57001179e-01
1.19001889e+00 1.35762662e-01 7.76164293e-01 3.25517923e-01
4.00861651e-01 3.19467038e-01 1.50577798e-01 4.91542459e-01
-2.74880558e-01 4.90972787e-01 8.58296454e-02 -5.06701887e-01
1.79435477e-01 5.03121436e-01 -5.63280797e-03 1.81301028e-01
-2.65145004e-01 -2.77522001e-02 -1.23023880e+00 5.24480343e-01
-1.75564873e+00 -7.65156269e-01 -4.54764277e-01 2.36055732e+00
6.69323802e-01 4.04748887e-01 3.46992731e-01 1.03562558e+00
4.40335721e-01 -1.65303305e-01 -4.43779409e-01 -4.12743688e-01
-2.70251215e-01 3.47207159e-01 3.61225635e-01 1.84159428e-01
-1.16723478e+00 4.17351909e-02 7.04885197e+00 1.24315135e-01
-1.15721023e+00 -1.48222242e-02 1.05115116e+00 -3.13091092e-02
1.11029856e-01 -5.17012417e-01 -1.20588906e-01 5.33369899e-01
1.22714925e+00 2.31623337e-01 1.09229684e-02 4.48331892e-01
7.94334710e-01 -1.45460129e-01 -1.41353834e+00 1.31446207e+00
-1.25752717e-01 -1.35615122e+00 -2.56251454e-01 -1.17509015e-01
3.48527193e-01 -1.07876882e-01 -7.91292787e-02 1.04823649e-01
-4.95722413e-01 -1.18175530e+00 5.90669990e-01 7.27793932e-01
7.40684092e-01 -7.40925729e-01 1.45019734e+00 2.22349644e-01
-5.18821657e-01 -1.32422060e-01 -2.30516437e-02 -1.76496476e-01
7.00026676e-02 1.28874040e+00 -6.91117942e-01 3.55597883e-01
1.02689147e+00 6.11281753e-01 -5.66579819e-01 1.12807631e+00
1.03171699e-01 1.04134202e+00 -3.82488370e-01 4.92174685e-01
-1.79193005e-01 2.25504726e-01 5.65879762e-01 1.40378273e+00
3.75353098e-02 1.05931543e-01 3.88488322e-02 8.12478721e-01
2.02458221e-02 -2.72343848e-02 -5.08120656e-01 2.37616017e-01
1.29580185e-01 8.73619676e-01 -8.41235101e-01 -4.53547567e-01
-1.60609707e-01 6.12457454e-01 -2.45554179e-01 3.19238961e-01
-5.13214469e-01 -3.55258256e-01 3.60839605e-01 6.23435378e-01
-1.06671505e-01 1.81218982e-01 -1.00205731e+00 -7.66901255e-01
3.93664122e-01 -1.26127005e+00 4.95259255e-01 -3.38556498e-01
-1.27121115e+00 8.17925334e-01 -3.80495161e-01 -1.29276729e+00
-3.16588670e-01 -4.70338672e-01 -5.02996564e-01 1.00292873e+00
-1.12277770e+00 -3.93590242e-01 -2.35263020e-01 2.96740294e-01
3.24109584e-01 -7.35488832e-02 9.37366724e-01 5.60829818e-01
-1.05666623e-01 5.14419734e-01 -1.96565911e-01 4.68287140e-01
7.50373244e-01 -1.53605735e+00 6.11555338e-01 8.14572215e-01
3.62910658e-01 5.86525977e-01 7.88687825e-01 -4.76119459e-01
-8.45590770e-01 -1.06058335e+00 7.89318860e-01 -8.05048406e-01
7.55865797e-02 -9.58231241e-02 -1.11990559e+00 1.87681079e-01
-1.86488122e-01 3.06198895e-01 7.47072399e-01 2.91007251e-01
9.93838757e-02 -2.60566324e-01 -1.14519346e+00 3.51552442e-02
6.61479890e-01 -4.08999890e-01 -6.06440187e-01 2.54188567e-01
-1.71223938e-01 -4.05539215e-01 -9.83040512e-01 7.20523715e-01
6.06652200e-01 -1.18375468e+00 7.80984759e-01 -5.72515786e-01
-5.53604923e-02 -1.77585140e-01 3.66737157e-01 -1.40427113e+00
-3.12225461e-01 -6.47364736e-01 9.72728282e-02 7.82251716e-01
5.05899966e-01 -4.57955271e-01 6.09101951e-01 3.29769760e-01
1.37841972e-02 -7.22953916e-01 -9.77903187e-01 -6.44286275e-01
8.26588273e-02 -7.82143891e-01 3.33005488e-01 8.29854906e-01
-2.99892038e-01 1.75451308e-01 -2.47213975e-01 1.15174785e-01
4.48360115e-01 -2.10637406e-01 4.35050130e-01 -1.55073225e+00
-1.37190089e-01 -2.91010529e-01 -6.80157542e-01 -9.80023071e-02
-2.51398295e-01 -8.22084546e-01 -4.13816795e-03 -1.37171960e+00
-2.70411462e-01 -5.94531775e-01 -7.42392957e-01 3.36965472e-01
-2.60902166e-01 5.63567877e-01 -7.86406621e-02 1.60646051e-01
-2.68879116e-01 -1.71634346e-01 5.67007363e-01 -6.99739009e-02
-4.42213863e-01 3.42391729e-01 -4.52294409e-01 8.64307761e-01
8.28070939e-01 -7.27068365e-01 -1.44477151e-02 -3.41890365e-01
3.12466025e-01 8.98892954e-02 4.71014023e-01 -1.30443358e+00
-2.54004568e-01 6.24389291e-01 9.07926977e-01 -2.36603588e-01
-1.07086442e-01 -8.74886572e-01 2.38098815e-01 6.97751820e-01
-5.13804078e-01 3.85887951e-01 2.97384322e-01 4.58661199e-01
-2.23468319e-01 -1.03178779e-02 8.69055331e-01 -1.76157460e-01
-9.30459648e-02 -3.98793370e-02 -8.13628137e-01 3.25156540e-01
6.15998328e-01 -1.55350909e-01 3.01766425e-01 -3.86981964e-01
-1.26328278e+00 -1.95272282e-01 -7.24041089e-02 2.65087754e-01
5.92108905e-01 -9.40318882e-01 -9.99240935e-01 4.29811895e-01
1.27433717e-01 -1.76243320e-01 1.62620947e-01 1.16969419e+00
-5.83337247e-01 1.43989712e-01 -1.64688334e-01 -1.09127319e+00
-1.17358851e+00 2.57679611e-01 8.14657271e-01 -1.11458234e-01
-1.00969267e+00 5.31024456e-01 -3.44686866e-01 -2.46672006e-03
5.72331071e-01 -7.70555556e-01 -1.23019405e-01 2.42333427e-01
6.08495355e-01 4.00769770e-01 8.67377341e-01 -3.12936664e-01
-6.08925521e-01 3.83171618e-01 1.00868978e-01 5.86546818e-03
1.23889625e+00 1.67581216e-01 8.87248218e-02 5.64400852e-01
7.74955094e-01 -3.09263468e-01 -7.12101102e-01 1.52501792e-01
2.26271510e-01 -1.94598198e-01 2.73316592e-01 -1.24801135e+00
-1.14670181e+00 1.22912300e+00 1.43181765e+00 3.74707073e-01
1.16448808e+00 -5.53750575e-01 3.84807140e-01 2.95004874e-01
5.59215769e-02 -1.08830917e+00 -3.07920635e-01 3.62099968e-02
6.45390153e-01 -1.35142827e+00 -1.24068174e-03 1.45824030e-01
-6.61410630e-01 1.07789338e+00 -1.82423536e-02 -8.62642080e-02
8.58357728e-01 5.97724974e-01 8.16336930e-01 -2.24625334e-01
-3.83855999e-01 1.09404147e-01 4.99466397e-02 9.10926998e-01
7.18671024e-01 1.62124470e-01 -5.53930819e-01 8.90933394e-01
1.17267296e-01 3.23590130e-01 6.15211606e-01 8.72234941e-01
-1.02479212e-01 -1.03732777e+00 -4.08180356e-01 8.74256849e-01
-1.09936571e+00 -5.73767051e-02 -3.70085508e-01 4.36125159e-01
2.88276464e-01 1.03592908e+00 -1.18174784e-01 -1.85574383e-01
4.90278095e-01 1.97928980e-01 2.39805534e-01 -6.18586123e-01
-1.25665879e+00 2.55970120e-01 3.19796540e-02 -4.96815085e-01
-3.41717631e-01 -5.55270314e-01 -9.96831536e-01 3.78126830e-01
-1.53032556e-01 1.42657101e-01 5.66320837e-01 9.43908632e-01
3.47192258e-01 8.09608221e-01 2.82605767e-01 -5.73657990e-01
-7.10839033e-01 -9.75941718e-01 -5.03541648e-01 6.94749057e-01
8.29839170e-01 -2.21251830e-01 -5.58152795e-01 2.68534690e-01] | [14.318840980529785, 3.2930819988250732] |
06826626-8872-4ee9-b6f6-a049b77056b8 | musiac-an-extensible-generative-framework-for | 2202.05528 | null | https://arxiv.org/abs/2202.05528v1 | https://arxiv.org/pdf/2202.05528v1.pdf | MusIAC: An extensible generative framework for Music Infilling Applications with multi-level Control | We present a novel music generation framework for music infilling, with a user friendly interface. Infilling refers to the task of generating musical sections given the surrounding multi-track music. The proposed transformer-based framework is extensible for new control tokens as the added music control tokens such as tonal tension per bar and track polyphony level in this work. We explore the effects of including several musically meaningful control tokens, and evaluate the results using objective metrics related to pitch and rhythm. Our results demonstrate that adding additional control tokens helps to generate music with stronger stylistic similarities to the original music. It also provides the user with more control to change properties like the music texture and tonal tension in each bar compared to previous research which only provided control for track density. We present the model in a Google Colab notebook to enable interactive generation. | ['Dorien Herremans', 'Thor Magnusson', 'Chris Kiefer', 'Ivor Simpson', 'Rui Guo'] | 2022-02-11 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 3.88596058e-02 -7.68597648e-02 5.57615645e-02 2.93530762e-01
-6.15801096e-01 -9.11123037e-01 5.49743652e-01 7.20865801e-02
6.16830774e-02 5.76151729e-01 4.64154810e-01 8.94456804e-02
-4.61101145e-01 -8.29527557e-01 -4.16008711e-01 -4.27082062e-01
-9.04323906e-02 1.46480650e-01 2.95836091e-01 -5.46416640e-01
5.72966933e-01 3.50132823e-01 -1.90461266e+00 3.90788138e-01
7.78346956e-01 6.56680465e-01 3.43722522e-01 8.88623714e-01
-9.72534120e-02 5.92024207e-01 -1.03709126e+00 -2.17163533e-01
4.25602168e-01 -5.00312924e-01 -1.98890105e-01 -2.74441004e-01
6.17025018e-01 -7.34685734e-03 3.42189252e-01 7.09581435e-01
8.12303066e-01 2.63641089e-01 5.19965172e-01 -1.04637289e+00
-3.87846738e-01 1.30048668e+00 -4.41384465e-01 -1.92315474e-01
6.34388685e-01 -5.34454398e-02 1.11535859e+00 -5.09981036e-01
5.61980128e-01 1.12667978e+00 8.85412991e-01 2.89235234e-01
-1.44604766e+00 -8.73192787e-01 -2.73136288e-01 -3.65166701e-02
-1.33304822e+00 -1.79140180e-01 8.74946356e-01 -3.95724058e-01
5.90094090e-01 7.80013084e-01 1.16500533e+00 6.12834990e-01
1.46597266e-01 5.21397650e-01 8.49769592e-01 -7.96737134e-01
9.02748555e-02 -1.18685156e-01 -4.51007545e-01 1.18355826e-01
-1.81268319e-01 2.51183063e-01 -8.65896344e-01 -8.25934708e-02
1.26689076e+00 -6.84650242e-01 -6.83464259e-02 -1.67184606e-01
-1.45195055e+00 4.88006383e-01 1.73855931e-01 5.11264265e-01
-1.55936018e-01 5.06090641e-01 4.58323807e-01 1.88186318e-01
9.07954276e-02 1.21393085e+00 -1.95753381e-01 -6.71674311e-01
-1.35857546e+00 9.48015273e-01 7.74089992e-01 1.02889478e+00
4.53658760e-01 4.68774140e-01 -7.81318128e-01 9.13240671e-01
1.11186475e-01 4.82684344e-01 4.38356221e-01 -1.27197194e+00
1.30387440e-01 4.04680222e-01 2.32448980e-01 -9.10622895e-01
-2.62172729e-01 -5.27872741e-01 -5.17968655e-01 6.76333785e-01
4.51666087e-01 -1.48458276e-02 -5.27259588e-01 1.60990477e+00
1.58239722e-01 1.43136665e-01 -3.99637610e-01 7.49207616e-01
5.98869205e-01 4.93402183e-01 -2.12900877e-01 -1.15072675e-01
1.43222928e+00 -8.91997695e-01 -1.21666908e+00 4.13159221e-01
1.16485320e-01 -1.54991615e+00 1.82269037e+00 7.41670489e-01
-1.54745519e+00 -9.17933345e-01 -1.11424720e+00 -8.54114536e-03
3.10139768e-02 4.36142325e-01 5.50401747e-01 8.90511632e-01
-9.28627968e-01 1.13867509e+00 -4.17751580e-01 -1.27379177e-02
-2.37695023e-01 2.17590958e-01 2.54865289e-01 7.88559318e-01
-1.04312408e+00 5.51042795e-01 2.70846516e-01 -3.51654023e-01
-3.67740512e-01 -1.16690874e+00 -6.25426471e-01 1.59250751e-01
8.57171044e-02 -6.77086711e-01 1.57779038e+00 -6.75124168e-01
-2.08587432e+00 3.20927113e-01 3.45354199e-01 -1.80487275e-01
5.46006918e-01 -4.75068927e-01 -2.70153582e-01 -8.38936195e-02
1.21400066e-01 7.06738830e-01 7.17149854e-01 -1.06859398e+00
-5.43344438e-01 4.26867306e-02 1.88811347e-01 3.06297481e-01
-3.44278097e-01 -3.13527405e-01 -4.22428846e-01 -1.56866956e+00
-1.23807177e-01 -1.00226808e+00 3.83729301e-03 -2.07009032e-01
-6.01523101e-01 2.23922372e-01 7.27871716e-01 -5.14016449e-01
1.79955876e+00 -2.15751886e+00 1.15094803e-01 2.83616930e-01
-3.68414730e-01 -9.41525698e-02 -3.05604711e-02 5.62460303e-01
-1.44693116e-02 1.24484241e-01 2.36052141e-01 -1.90406889e-01
2.16582015e-01 -1.84676528e-01 -2.07482666e-01 -3.32451791e-01
-2.87204832e-01 4.89454150e-01 -7.28577256e-01 -2.39685819e-01
2.62747288e-01 4.86083806e-01 -9.69028056e-01 -8.77782777e-02
-3.18171591e-01 4.25142825e-01 -1.14053018e-01 5.41355312e-01
4.69962448e-01 4.36869323e-01 -1.04481000e-02 -4.45525765e-01
-6.69177413e-01 4.04175073e-01 -1.71534693e+00 2.07568312e+00
-5.24309754e-01 4.95893449e-01 -1.34333849e-01 4.56817895e-02
1.36699212e+00 4.04705495e-01 3.81886601e-01 -4.08854097e-01
-6.04950041e-02 1.72388569e-01 1.08131051e-01 -2.84794837e-01
1.09429979e+00 -2.39332065e-01 -8.99177790e-02 5.09524703e-01
-2.83131033e-01 -7.90717304e-01 5.96413672e-01 1.29465042e-02
7.13707328e-01 6.21320844e-01 2.57166058e-01 -5.37797987e-01
1.80816278e-01 -1.49963409e-01 2.06505835e-01 5.81636727e-01
4.42769557e-01 6.92044437e-01 2.66753167e-01 -5.06783798e-02
-1.39966774e+00 -1.28557575e+00 2.25614235e-02 1.20672071e+00
-2.33128518e-02 -1.11208081e+00 -7.44304121e-01 2.67537594e-01
-8.67155790e-02 6.90743983e-01 -5.06120384e-01 1.06060803e-01
-5.28789759e-01 -2.11578459e-01 8.06353748e-01 3.47122818e-01
3.28142464e-01 -1.25302601e+00 -6.39714777e-01 3.53555977e-01
-2.38080174e-01 -5.06840169e-01 -1.02946925e+00 -9.90894586e-02
-1.02614772e+00 -6.50790036e-01 -7.49448180e-01 -5.76663733e-01
-3.23717110e-02 -2.20443234e-01 1.12292171e+00 -1.32329941e-01
-5.90892732e-01 1.14660263e-01 -5.78267276e-01 -7.07833111e-01
-3.97997200e-01 4.22206789e-01 -4.09473106e-02 -3.84171516e-01
-3.76494020e-01 -1.06973290e+00 -7.88501561e-01 3.07904720e-01
-9.78385746e-01 4.27792400e-01 7.94388652e-02 3.59193981e-01
6.29998922e-01 6.28179731e-03 6.20220125e-01 -5.35928309e-01
1.22820210e+00 2.95533866e-01 -3.63549799e-01 -1.24459401e-01
-3.92795205e-01 1.94419220e-01 5.18915117e-01 -8.60976994e-01
-8.34521472e-01 -8.17742273e-02 -8.54907036e-02 -2.23685846e-01
2.40254074e-01 3.87668788e-01 2.45966744e-02 1.55710936e-01
9.88930464e-01 -1.35503367e-01 -2.24347174e-01 -6.34427249e-01
5.86151063e-01 2.88351327e-01 7.84433424e-01 -9.40212190e-01
8.25156450e-01 3.89167406e-02 -4.27823290e-02 -4.79444057e-01
-3.06700706e-01 -1.14147067e-01 -7.02444375e-01 -5.64119577e-01
5.08863568e-01 -5.43643534e-01 -9.41044629e-01 1.55459508e-01
-8.71721864e-01 -4.17256683e-01 -9.23301756e-01 5.85340619e-01
-9.67363238e-01 4.52647358e-02 -5.21702111e-01 -8.07542920e-01
-5.14738500e-01 -8.82503927e-01 1.10701609e+00 3.82051080e-01
-8.74941289e-01 -6.27215803e-01 5.28601885e-01 -7.92205036e-02
4.97226536e-01 5.64444721e-01 8.55660915e-01 3.10533971e-01
-3.99145424e-01 1.34146772e-02 4.14337337e-01 -1.74471717e-02
4.60207433e-01 3.90422106e-01 -8.43616545e-01 -6.09676577e-02
-5.84824979e-01 1.12033613e-01 3.95044804e-01 5.06781101e-01
8.03277671e-01 -3.15747827e-01 9.16371793e-02 4.70966429e-01
1.24263477e+00 2.71897703e-01 9.53236222e-01 6.65102720e-01
5.40875137e-01 3.73644233e-01 7.60126889e-01 9.69236076e-01
-1.88173205e-01 1.36944950e+00 4.02769707e-02 -1.13504276e-01
-5.27919590e-01 -5.93461514e-01 3.79210800e-01 8.76180589e-01
-6.82454109e-01 -4.74963970e-02 -3.82087916e-01 3.50700170e-01
-1.64472270e+00 -1.31784606e+00 -2.08401918e-01 2.27536821e+00
1.10675418e+00 8.51687565e-02 6.86410785e-01 7.62017548e-01
6.88021719e-01 -1.89513624e-01 -2.32098997e-01 -7.66698062e-01
-2.48992722e-02 7.47326076e-01 2.45137900e-01 4.62460220e-01
-6.28392458e-01 8.92203391e-01 7.44748974e+00 1.05358994e+00
-1.14014065e+00 -2.99907714e-01 1.03998836e-02 -6.58882916e-01
-5.80216050e-01 -4.98849433e-03 -5.44722021e-01 3.14573914e-01
4.01016235e-01 -5.40456057e-01 6.74982548e-01 5.73900163e-01
8.54491115e-01 3.48681882e-02 -9.22333717e-01 8.43872309e-01
-2.32382715e-01 -1.50876665e+00 3.66550237e-01 -1.07625099e-02
6.72735810e-01 -8.13275695e-01 3.07582587e-01 1.49044454e-01
-3.39792483e-02 -9.15835261e-01 1.31018758e+00 7.34225094e-01
1.09654593e+00 -9.83264506e-01 1.02138594e-01 -1.65215448e-01
-1.57554948e+00 5.66512868e-02 1.34692743e-01 -3.46074343e-01
3.32032293e-01 4.27224547e-01 -1.00366640e+00 4.42298532e-01
6.32589936e-01 3.85606319e-01 -5.59799016e-01 1.40072680e+00
-1.24009855e-01 4.46927667e-01 -3.65779638e-01 8.65358021e-03
-2.35283300e-01 -2.13369414e-01 7.01248527e-01 1.23929989e+00
1.09700692e+00 -2.65757233e-01 -4.98862602e-02 1.06107473e+00
4.30028886e-01 5.90087533e-01 -1.50246605e-01 8.48318413e-02
7.14409411e-01 1.25190115e+00 -7.19647586e-01 -1.49800643e-01
3.48600954e-01 8.40310216e-01 -2.32472524e-01 3.93997692e-02
-7.94653416e-01 -6.08154237e-01 4.06746566e-01 6.71471596e-01
1.92751199e-01 -2.62460262e-01 -5.46067715e-01 -6.46371365e-01
-1.06402919e-01 -1.04624355e+00 4.01338488e-02 -1.20098448e+00
-8.13366532e-01 5.21711648e-01 -2.90709734e-03 -1.52851915e+00
-4.75529879e-01 -2.11184695e-01 -8.67128074e-01 9.16497409e-01
-5.94224274e-01 -1.18388188e+00 -2.94983745e-01 4.37297940e-01
5.08690238e-01 -1.64942205e-01 1.01977170e+00 4.86572713e-01
5.60091026e-02 8.08596849e-01 -1.26258612e-01 -4.91855353e-01
9.93380010e-01 -1.45849383e+00 3.34009916e-01 3.21192086e-01
3.66037190e-01 6.46344960e-01 1.16765451e+00 -5.96122980e-01
-8.79675210e-01 -5.75172842e-01 5.74511230e-01 -1.08566217e-01
3.76591355e-01 -3.34201723e-01 -3.73901367e-01 1.11892104e-01
2.36603633e-01 -8.91409874e-01 8.62726629e-01 3.50363612e-01
-1.31165445e-01 -1.89428687e-01 -8.78463566e-01 9.25490737e-01
9.11140621e-01 -3.60917866e-01 -3.79026890e-01 -2.46581540e-01
5.92912912e-01 -5.01548767e-01 -1.10534048e+00 3.90028030e-01
1.19915056e+00 -1.13876188e+00 8.88587594e-01 2.06330284e-01
3.60435128e-01 -8.27135384e-01 3.53998691e-02 -1.49412143e+00
-6.08996987e-01 -1.17914844e+00 2.31655478e-01 1.38490236e+00
4.74881470e-01 2.59777188e-01 6.80366933e-01 -1.48870545e-02
-3.75906020e-01 -2.06205651e-01 -5.42853236e-01 -8.11361194e-01
-2.39287332e-01 -4.40186560e-01 8.29335392e-01 7.25126803e-01
4.26529676e-01 3.18361640e-01 -5.83503664e-01 -1.81670517e-01
2.48561487e-01 1.27780676e-01 1.05642116e+00 -1.19959867e+00
-6.62660301e-01 -6.60606742e-01 -1.70722157e-01 -6.26864970e-01
-7.40180194e-01 -7.86743701e-01 -2.37361059e-01 -1.38966107e+00
-4.82813716e-02 -6.70620620e-01 -1.53584480e-01 4.10485297e-01
-3.41508240e-02 7.13431716e-01 8.02679360e-01 1.03893735e-01
7.39590302e-02 3.86550903e-01 1.72787297e+00 -1.48718655e-01
-8.70502412e-01 2.26672798e-01 -6.75743639e-01 6.06732011e-01
8.04032445e-01 -3.92299920e-01 -5.77160299e-01 -9.15066674e-02
6.03532732e-01 2.98172049e-02 1.79873295e-02 -1.46089566e+00
-2.04297118e-02 -2.69588828e-03 2.26039842e-01 -6.66268110e-01
3.11817884e-01 -3.51460129e-01 8.11561763e-01 4.00910050e-01
-5.47106087e-01 3.78197074e-01 7.40024924e-01 -9.63563658e-03
-1.17133997e-01 -2.58099645e-01 5.64193785e-01 1.18689872e-01
-6.05260506e-02 -2.63537437e-01 -3.91195953e-01 -3.62389565e-01
7.39939392e-01 -5.55926502e-01 1.11081526e-01 -5.81794143e-01
-1.10864699e+00 -4.65949655e-01 3.73511285e-01 5.96272111e-01
2.24548385e-01 -1.82953572e+00 -6.25857532e-01 1.44521311e-01
1.30829856e-01 -4.59967643e-01 2.58088976e-01 3.38275075e-01
-8.11987579e-01 2.05713268e-02 -5.72901845e-01 -4.11526918e-01
-1.61793005e+00 2.95006871e-01 9.01095718e-02 -2.53902227e-01
-6.15831137e-01 5.41617036e-01 -1.88972037e-02 -2.74109364e-01
3.50846767e-01 -6.72196925e-01 -3.41953784e-01 2.54888266e-01
5.82301676e-01 5.55617332e-01 -3.26220058e-02 -3.43720540e-02
1.41835183e-01 8.31788063e-01 4.75882024e-01 -6.71189725e-01
9.89244401e-01 1.86450891e-02 1.35786682e-01 7.55157828e-01
2.15939239e-01 8.96551967e-01 -1.07419348e+00 4.57760841e-01
-2.57155091e-01 -6.05560541e-01 -1.95147619e-01 -1.07985902e+00
-6.80436075e-01 3.62137407e-01 7.17519164e-01 3.40951800e-01
1.08833921e+00 -4.16117013e-01 5.48112214e-01 -3.48624513e-02
2.78940409e-01 -1.21051919e+00 3.99730057e-01 4.38651770e-01
1.13755965e+00 -5.02246805e-02 -1.28564727e-03 -3.84870142e-01
-4.96287137e-01 1.14408588e+00 4.41935956e-01 -6.57386854e-02
4.04627919e-01 7.74083495e-01 2.41587818e-01 7.73061812e-02
-4.18798536e-01 -7.71182775e-02 4.62682337e-01 5.84140599e-01
1.11657763e+00 1.86157852e-01 -6.34587288e-01 5.85140944e-01
-1.30426729e+00 1.91305697e-01 5.24777174e-01 5.20448208e-01
-4.63387787e-01 -1.62109661e+00 -7.84093916e-01 1.38681322e-01
-4.33382839e-01 -4.59249109e-01 -2.31096208e-01 7.58580208e-01
6.71816826e-01 8.03673148e-01 2.21018374e-01 -5.78270078e-01
6.92789495e-01 -1.77727342e-01 8.63451779e-01 -6.00178838e-01
-1.16823852e+00 7.83102810e-01 5.57524934e-02 -2.42173374e-01
-3.30362022e-01 -4.46761906e-01 -1.17254388e+00 -5.03478050e-01
-4.60813254e-01 3.25242579e-01 6.77170217e-01 2.85057634e-01
2.67873675e-01 1.18549180e+00 4.58555669e-01 -1.15374589e+00
9.11796167e-02 -1.33594561e+00 -9.16925907e-01 3.47586662e-01
-2.64456421e-01 -5.81332386e-01 -6.98682293e-02 4.11157817e-01] | [15.99519157409668, 5.433119297027588] |
c163f1ef-b5e9-4fb7-8a7d-d7191ee914bf | self-supervised-real-time-video-stabilization | 2111.05980 | null | https://arxiv.org/abs/2111.05980v1 | https://arxiv.org/pdf/2111.05980v1.pdf | Self-Supervised Real-time Video Stabilization | Videos are a popular media form, where online video streaming has recently gathered much popularity. In this work, we propose a novel method of real-time video stabilization - transforming a shaky video to a stabilized video as if it were stabilized via gimbals in real-time. Our framework is trainable in a self-supervised manner, which does not require data captured with special hardware setups (i.e., two cameras on a stereo rig or additional motion sensors). Our framework consists of a transformation estimator between given frames for global stability adjustments, followed by scene parallax reduction module via spatially smoothed optical flow for further stability. Then, a margin inpainting module fills in the missing margin regions created during stabilization to reduce the amount of post-cropping. These sequential steps reduce distortion and margin cropping to a minimum while enhancing stability. Hence, our approach outperforms state-of-the-art real-time video stabilization methods as well as offline methods that require camera trajectory optimization. Our method procedure takes approximately 24.3 ms yielding 41 fps regardless of resolution (e.g., 480p or 1080p). | ['In So Kweon', 'Jaesik Park', 'Jinsoo Choi'] | 2021-11-10 | null | null | null | null | ['video-stabilization'] | ['computer-vision'] | [ 2.93658286e-01 -1.29028201e-01 -1.13707945e-01 -5.52664734e-02
-6.47831023e-01 -7.88340092e-01 2.77533740e-01 2.24407226e-01
-4.92374927e-01 5.55819929e-01 4.32823263e-02 -1.79541856e-01
2.99222887e-01 -3.97888243e-01 -1.17775011e+00 -7.66700327e-01
7.71353394e-02 -2.21423075e-01 6.84072495e-01 2.11885702e-02
3.93113226e-01 3.84094894e-01 -1.51117909e+00 -1.24664374e-01
7.13488996e-01 9.89953458e-01 2.19242692e-01 7.68141210e-01
4.12673891e-01 1.08094788e+00 -1.58388227e-01 -1.94483981e-01
4.78948206e-01 -3.40100914e-01 -5.38360059e-01 6.49889529e-01
8.16719532e-01 -7.11256146e-01 -4.13259655e-01 1.05032277e+00
3.17491382e-01 2.84466982e-01 2.89695188e-02 -1.03163588e+00
1.10542886e-01 1.03312664e-01 -9.63814139e-01 1.55437246e-01
5.63293576e-01 3.08906645e-01 5.02775371e-01 -7.44662285e-01
6.13950789e-01 1.00232351e+00 4.52690423e-01 2.96746254e-01
-1.30330801e+00 -4.11384434e-01 -1.75705571e-02 2.13348255e-01
-1.25224030e+00 -8.54547977e-01 7.51141191e-01 -5.02345622e-01
3.72948647e-01 2.56461531e-01 7.42308915e-01 5.93088090e-01
1.96781799e-01 4.70440835e-01 6.34134293e-01 -4.66834664e-01
2.62546897e-01 -1.22923531e-01 -3.30530226e-01 6.98047042e-01
2.66710315e-02 -2.42154136e-01 -7.60731936e-01 -5.13654947e-02
1.28528392e+00 2.83494331e-02 -6.00102007e-01 -5.80366790e-01
-1.53148079e+00 4.27732855e-01 5.89366741e-02 -1.76168203e-01
-3.48275661e-01 1.53080642e-01 5.45743644e-01 3.69127065e-01
5.60997128e-01 8.38151053e-02 -3.24175149e-01 -3.48342806e-01
-1.30730116e+00 4.76675816e-02 3.60299855e-01 9.89000261e-01
9.79354560e-01 1.31109148e-01 1.22153953e-01 4.16776657e-01
2.05836877e-01 3.03364903e-01 3.89982283e-01 -1.48839962e+00
6.50327802e-01 2.78808594e-01 4.47653025e-01 -1.26784301e+00
-5.76962121e-02 1.42376602e-01 -6.99137568e-01 2.78339893e-01
6.35027885e-01 -2.43143842e-01 -3.14148962e-01 1.55016041e+00
6.94178224e-01 7.07629681e-01 -1.91239357e-01 1.02648807e+00
1.96039915e-01 8.59683692e-01 -6.09104574e-01 -7.28359759e-01
1.15577853e+00 -1.22431803e+00 -7.75687397e-01 -1.09867834e-01
4.21727121e-01 -1.00405955e+00 9.90350723e-01 5.71612298e-01
-1.34510219e+00 -4.77647126e-01 -1.00848508e+00 -1.43800542e-01
5.16954780e-01 6.30185008e-02 -4.68918793e-02 3.88449013e-01
-1.11616695e+00 6.66581690e-01 -1.22952461e+00 -3.02725792e-01
-2.03815158e-02 3.64933461e-01 -5.34071982e-01 2.00857744e-01
-7.25933015e-01 6.07196867e-01 1.75437585e-01 6.96063973e-03
-8.79012764e-01 -6.84629500e-01 -1.00694990e+00 1.08844033e-02
7.31411278e-01 -3.92555773e-01 1.21803045e+00 -1.42983317e+00
-2.03201699e+00 7.63105929e-01 -4.04404879e-01 -4.77751613e-01
7.02315986e-01 -5.78574002e-01 -9.10142530e-03 4.90325272e-01
-9.05584469e-02 4.22157228e-01 1.33162701e+00 -9.67026532e-01
-5.04405379e-01 -4.02712040e-02 2.27808580e-01 4.31117117e-01
-4.65762168e-01 3.57157648e-01 -8.41634929e-01 -6.57468617e-01
1.54657811e-01 -1.03960407e+00 -1.08160347e-01 4.00923580e-01
-1.26582667e-01 4.77261186e-01 1.02811980e+00 -7.22795725e-01
1.32535589e+00 -2.17104030e+00 1.97310254e-01 -5.34797646e-02
7.40338266e-02 3.01396221e-01 1.90669283e-01 3.18297535e-01
-1.43122956e-01 -4.52979505e-01 -9.26497355e-02 -5.49685359e-01
-5.29735923e-01 -1.00319058e-01 -3.31890315e-01 9.23566699e-01
-1.01699941e-01 2.68029004e-01 -8.69113147e-01 -6.08320355e-01
6.54764354e-01 5.06864548e-01 -7.43158579e-01 3.44505340e-01
3.02892774e-02 8.54746044e-01 -2.62935180e-02 4.98010665e-01
7.18732595e-01 -8.13887119e-02 1.24018885e-01 -2.85361171e-01
-7.18481719e-01 1.33368403e-01 -1.46738982e+00 2.00368595e+00
-3.56947422e-01 8.28967690e-01 4.32303429e-01 -9.20448482e-01
6.12043440e-01 3.51697445e-01 7.33075857e-01 -8.07714015e-02
3.49882662e-01 1.08925477e-01 -4.96316433e-01 -5.03028572e-01
5.90468884e-01 2.80791461e-01 4.00728732e-01 2.88743824e-01
-3.17417592e-01 -1.22328237e-01 3.41251820e-01 1.91425398e-01
9.37260985e-01 3.58726293e-01 3.89229894e-01 -3.21139038e-01
8.69481325e-01 -2.42195293e-01 8.17175090e-01 1.49067923e-01
-3.15611154e-01 1.01115477e+00 4.13265377e-01 -3.94446969e-01
-1.13815200e+00 -6.43812180e-01 2.11771280e-01 7.69428194e-01
5.86976290e-01 -7.61370599e-01 -1.10078657e+00 -6.55621812e-02
-3.82597357e-01 4.82833534e-02 -2.07651660e-01 -6.46220753e-04
-9.21536565e-01 -1.26257092e-01 9.99271497e-02 2.15119451e-01
4.48431045e-01 -7.41243362e-01 -1.04897237e+00 2.48878971e-01
-3.95173162e-01 -1.21234035e+00 -1.07530558e+00 -2.51860797e-01
-1.16156375e+00 -1.07200837e+00 -7.13842809e-01 -8.05788517e-01
7.94206858e-01 8.17821980e-01 7.37375915e-01 -8.96020830e-02
1.04019642e-01 2.25265026e-01 -2.08708838e-01 2.33188286e-01
-1.75265506e-01 -2.43777752e-01 3.36839408e-01 5.22590160e-01
-2.62349457e-01 -5.20596743e-01 -8.20804298e-01 5.07082224e-01
-1.03378510e+00 4.15066719e-01 -1.03639383e-02 6.77905917e-01
6.85366333e-01 1.46924844e-02 -3.46909106e-01 -2.70585418e-01
3.78740802e-02 -5.49121499e-02 -1.06079197e+00 1.62836965e-02
-1.34499416e-01 -2.23408118e-01 8.57177973e-01 -5.45462310e-01
-9.74864066e-01 4.05662179e-01 4.31725323e-01 -9.35605407e-01
1.55137375e-01 2.03622252e-01 -6.62916303e-02 -2.29401395e-01
5.76203942e-01 2.00429112e-01 3.35367233e-01 -2.41753474e-01
1.79994240e-01 4.63460118e-01 8.44296217e-01 -3.62057596e-01
9.55201626e-01 7.55892932e-01 -9.73856226e-02 -8.97425115e-01
-6.00616395e-01 -5.44717610e-01 -7.62757480e-01 -5.19457757e-01
8.05138588e-01 -1.21340227e+00 -8.57122660e-01 8.58151555e-01
-1.11494434e+00 -2.62473762e-01 -8.58441815e-02 6.04467452e-01
-6.80100143e-01 7.94589758e-01 -7.47173369e-01 -4.80967164e-01
-2.35823929e-01 -1.20135736e+00 1.02113235e+00 3.85203868e-01
2.39457190e-02 -7.03307509e-01 1.32629216e-01 3.58345807e-01
2.70811468e-01 2.99960941e-01 -1.44747511e-01 4.42391813e-01
-8.44836712e-01 6.62800157e-03 4.96329973e-03 4.40502137e-01
4.27330166e-01 3.78599018e-01 -6.47082031e-01 -7.34962404e-01
2.41875052e-01 -2.18874127e-01 3.78246903e-01 5.73496401e-01
1.06183219e+00 -5.02281308e-01 -7.01061962e-03 9.46186483e-01
1.42143774e+00 1.94661051e-01 5.67796946e-01 6.32869720e-01
6.75734341e-01 4.19332832e-01 8.85303557e-01 6.91306353e-01
2.40621269e-01 8.66658628e-01 6.00954890e-01 -7.69169703e-02
1.63753197e-01 -8.44598338e-02 7.55356789e-01 8.19949031e-01
-2.47805521e-01 -1.65384188e-01 -7.05555141e-01 4.65665936e-01
-2.06138372e+00 -9.80215430e-01 -1.12041317e-01 2.68610120e+00
9.18773234e-01 -6.55564368e-02 5.16514964e-02 2.23796740e-01
1.12522960e+00 4.79162008e-01 -5.48677146e-01 -1.01736180e-01
-5.93666434e-02 -2.80986339e-01 6.24117196e-01 7.92080522e-01
-1.21657145e+00 8.86745393e-01 4.99500704e+00 4.60540086e-01
-1.48093998e+00 -9.55342501e-02 5.86103559e-01 -3.23795408e-01
1.63672090e-01 2.81918257e-01 -4.19456333e-01 7.14600325e-01
8.10844481e-01 -1.93769470e-01 5.79145253e-01 8.06844532e-01
7.81080246e-01 -3.82093042e-01 -9.01315033e-01 1.35306454e+00
2.42815301e-01 -1.44094026e+00 -4.22363490e-01 -1.29450291e-01
6.72392726e-01 -2.18345657e-01 -2.24456355e-01 -4.89657968e-01
-2.63983607e-01 -3.36609662e-01 1.10609949e+00 3.16101372e-01
7.61895061e-01 -6.26285613e-01 2.78352708e-01 3.62428099e-01
-1.27926672e+00 1.24217682e-01 -3.41254145e-01 -1.80391103e-01
6.11402571e-01 6.70579731e-01 -1.62989348e-01 3.91458809e-01
7.25254893e-01 9.54139948e-01 -3.01802874e-01 1.01859665e+00
-2.00959459e-01 5.88289559e-01 -5.17111719e-01 6.97955728e-01
-1.60187297e-02 -6.36280477e-01 5.94737649e-01 8.45536113e-01
4.74234968e-01 2.53785729e-01 2.01403543e-01 6.16912693e-02
9.69704539e-02 5.03782853e-02 -2.55644709e-01 4.46804196e-01
3.26845169e-01 1.15942907e+00 -6.78651512e-01 -4.77236748e-01
-6.24077141e-01 1.29351890e+00 -2.13924646e-02 1.74924672e-01
-1.19522703e+00 -2.32989892e-01 6.44819140e-01 5.30104399e-01
2.91313887e-01 -4.28229988e-01 7.75212571e-02 -1.77585816e+00
3.24620247e-01 -8.83649468e-01 7.00064227e-02 -8.99425805e-01
-5.58778346e-01 4.40237224e-01 -2.02313662e-01 -1.83745360e+00
-1.13474384e-01 -4.56836373e-02 -6.53533518e-01 4.15366977e-01
-1.50296402e+00 -6.53653204e-01 -6.10208929e-01 9.60415602e-01
7.47961879e-01 1.48343533e-01 2.10313529e-01 4.73033875e-01
-7.66911268e-01 3.03609163e-01 4.09898758e-01 -1.62817627e-01
1.30957210e+00 -8.36021483e-01 1.69862017e-01 1.36980104e+00
-2.31581196e-01 5.21642208e-01 8.99877429e-01 -3.76118720e-01
-1.52625084e+00 -9.48389292e-01 8.06116939e-01 7.04040867e-05
7.49043107e-01 -2.20927492e-01 -8.23170424e-01 6.86401665e-01
3.28219175e-01 3.97233248e-01 1.73634976e-01 -7.11564064e-01
-7.37132803e-02 -5.92515826e-01 -8.17663431e-01 5.13110816e-01
7.05465972e-01 -3.64513159e-01 -2.22662404e-01 2.19221294e-01
5.69923460e-01 -9.37237561e-01 -6.33552790e-01 9.80689898e-02
5.19282758e-01 -1.16290152e+00 8.15874994e-01 2.84086075e-02
6.14599943e-01 -8.60794783e-01 1.33562371e-01 -9.19605196e-01
-1.70633405e-01 -1.48716354e+00 -3.49186391e-01 1.15502453e+00
-1.69646084e-01 -3.97530347e-01 8.20046365e-01 6.17200315e-01
-1.10478006e-01 -4.34951931e-01 -8.75530720e-01 -6.23526752e-01
-6.18393540e-01 -1.14741459e-01 6.51145875e-02 9.09268856e-01
1.49537310e-01 2.29046624e-02 -7.85368502e-01 3.26674432e-01
7.47748137e-01 1.20970532e-01 1.10227001e+00 -6.82334006e-01
-1.68686613e-01 1.41824841e-01 -4.87801999e-01 -1.48329496e+00
-3.80449183e-02 -8.00782591e-02 3.81798646e-03 -8.43522370e-01
6.03594817e-03 -3.41424495e-02 -2.46795844e-02 2.75645435e-01
-1.93702981e-01 2.79382080e-01 3.27181816e-01 5.04881084e-01
-5.24254441e-01 4.83388931e-01 1.03792620e+00 2.11005285e-01
-4.42943335e-01 -8.06739852e-02 -2.22364455e-01 9.02033269e-01
7.15556085e-01 -2.72803634e-01 -5.97600043e-01 -7.49263763e-01
8.67282674e-02 4.69712943e-01 2.71872938e-01 -1.20087254e+00
4.18878019e-01 -3.12154740e-01 3.47687788e-02 -4.46141183e-01
3.60534191e-01 -9.04629290e-01 2.24325776e-01 3.29780668e-01
-1.07409261e-01 4.48777020e-01 6.49330616e-02 5.03958046e-01
-3.74865919e-01 1.27436575e-02 1.13081694e+00 1.07157119e-01
-3.56799722e-01 3.34873110e-01 -5.10470927e-01 -1.36428639e-01
1.32355046e+00 -3.87116671e-01 -7.75358602e-02 -5.78793108e-01
-4.48343098e-01 9.94060189e-02 1.05534613e+00 2.67579794e-01
4.02414739e-01 -1.07485461e+00 -4.15861785e-01 1.25559270e-01
-4.08114463e-01 3.27661008e-01 3.26417476e-01 1.04674792e+00
-1.20011544e+00 1.52387917e-01 -1.69046715e-01 -9.28010881e-01
-1.63793635e+00 5.68688571e-01 9.92818028e-02 4.13771830e-02
-7.52642274e-01 7.20203340e-01 1.61331877e-01 1.60589457e-01
2.64022470e-01 -3.75901520e-01 3.25424634e-02 7.84226358e-02
7.15639949e-01 5.22263050e-01 4.11290266e-02 -7.88793206e-01
-3.89316529e-01 8.71133983e-01 8.54557678e-02 -2.37393960e-01
1.19291639e+00 -6.41826272e-01 -1.98143229e-01 2.98060358e-01
1.03998172e+00 2.14845151e-01 -2.04611421e+00 -1.38204098e-01
-3.55902344e-01 -8.84546638e-01 4.14714403e-02 1.55390754e-01
-1.26139116e+00 5.10240376e-01 4.49090779e-01 -9.21411216e-02
1.43485272e+00 -4.10732538e-01 9.92926717e-01 1.73334211e-01
2.39293709e-01 -1.12336898e+00 1.75926909e-01 4.43534702e-01
6.06470764e-01 -1.27666414e+00 3.13379705e-01 -5.77471673e-01
-4.42610979e-01 1.27902675e+00 5.41224837e-01 -3.35080832e-01
4.51342046e-01 2.38891706e-01 1.29193261e-01 3.54257137e-01
-7.51003265e-01 2.86772698e-01 -7.82499388e-02 1.17973216e-01
3.02712828e-01 -4.51796263e-01 -2.78394043e-01 -1.26123250e-01
3.07609662e-02 4.06620316e-02 1.01088548e+00 1.05590069e+00
-4.42008615e-01 -8.14634919e-01 -6.33863211e-01 -1.77480131e-01
-4.13722605e-01 -1.27060227e-02 1.84394896e-01 3.30188274e-01
-1.54581383e-01 9.93442237e-01 1.18856832e-01 -1.42033905e-01
2.67567635e-01 -4.34981257e-01 3.80226612e-01 -3.31966966e-01
-4.85585153e-01 5.12443721e-01 -2.15324268e-01 -1.05062044e+00
-8.16666126e-01 -8.98100615e-01 -1.08786917e+00 -4.62338269e-01
-4.10506845e-01 7.35064130e-03 3.65155190e-01 6.31466210e-01
4.11370546e-01 -7.95242842e-03 1.03099477e+00 -1.33332646e+00
-1.94193229e-01 -5.96642613e-01 -2.93996394e-01 3.37660283e-01
7.07641840e-01 -3.53547871e-01 -5.20168006e-01 9.28361833e-01] | [10.62294864654541, -1.3856219053268433] |
1f636fa6-18c1-4b84-9ec5-b96cfd4a20ac | hoiclip-efficient-knowledge-transfer-for-hoi | 2303.15786 | null | https://arxiv.org/abs/2303.15786v2 | https://arxiv.org/pdf/2303.15786v2.pdf | HOICLIP: Efficient Knowledge Transfer for HOI Detection with Vision-Language Models | Human-Object Interaction (HOI) detection aims to localize human-object pairs and recognize their interactions. Recently, Contrastive Language-Image Pre-training (CLIP) has shown great potential in providing interaction prior for HOI detectors via knowledge distillation. However, such approaches often rely on large-scale training data and suffer from inferior performance under few/zero-shot scenarios. In this paper, we propose a novel HOI detection framework that efficiently extracts prior knowledge from CLIP and achieves better generalization. In detail, we first introduce a novel interaction decoder to extract informative regions in the visual feature map of CLIP via a cross-attention mechanism, which is then fused with the detection backbone by a knowledge integration block for more accurate human-object pair detection. In addition, prior knowledge in CLIP text encoder is leveraged to generate a classifier by embedding HOI descriptions. To distinguish fine-grained interactions, we build a verb classifier from training data via visual semantic arithmetic and a lightweight verb representation adapter. Furthermore, we propose a training-free enhancement to exploit global HOI predictions from CLIP. Extensive experiments demonstrate that our method outperforms the state of the art by a large margin on various settings, e.g. +4.04 mAP on HICO-Det. The source code is available in https://github.com/Artanic30/HOICLIP. | ['Xuming He', 'Yongfei Liu', 'Longtian Qiu', 'Shan Ning'] | 2023-03-28 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Ning_HOICLIP_Efficient_Knowledge_Transfer_for_HOI_Detection_With_Vision-Language_Models_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Ning_HOICLIP_Efficient_Knowledge_Transfer_for_HOI_Detection_With_Vision-Language_Models_CVPR_2023_paper.pdf | cvpr-2023-1 | ['human-object-interaction-detection'] | ['computer-vision'] | [ 2.67327160e-01 1.75209213e-02 -3.62428516e-01 -2.21050501e-01
-9.80693221e-01 -3.87159646e-01 5.72681010e-01 -1.17813930e-01
-2.63186961e-01 3.67847770e-01 4.52498525e-01 1.87138841e-01
2.07046241e-01 -4.47872818e-01 -1.11468947e+00 -4.52425808e-01
9.82118547e-02 2.48912305e-01 3.08460534e-01 -2.40186155e-02
-7.50626028e-02 4.07280922e-02 -1.52520394e+00 6.87557995e-01
7.78593004e-01 1.15231013e+00 4.42197680e-01 6.01239681e-01
3.44614893e-01 9.36756432e-01 -4.80215490e-01 -3.26415867e-01
2.75601566e-01 -5.54643810e-01 -5.72554171e-01 1.43018559e-01
3.87674898e-01 -5.90390503e-01 -6.45762324e-01 8.09579074e-01
6.70756698e-01 -4.50142473e-02 6.11874938e-01 -1.47491407e+00
-6.59499526e-01 4.84347224e-01 -7.13056326e-01 1.97423771e-01
6.36038423e-01 6.19749606e-01 1.01476848e+00 -1.47317016e+00
6.29324734e-01 1.33351004e+00 4.35271114e-01 3.40210289e-01
-9.93700266e-01 -8.08411658e-01 1.13998726e-02 5.24529755e-01
-1.73116469e+00 -4.66739327e-01 7.10468113e-01 -4.51916397e-01
1.23634660e+00 1.12110123e-01 8.01699817e-01 1.31794345e+00
-2.08604082e-01 1.57206726e+00 6.05305791e-01 -5.04298210e-01
-2.30756894e-01 2.33922705e-01 4.67763357e-02 8.97176981e-01
1.48179278e-01 1.77668199e-01 -9.79552031e-01 1.97791353e-01
7.20229387e-01 2.55101621e-02 -4.82435644e-01 -4.69790876e-01
-1.33245850e+00 7.12575853e-01 7.62778759e-01 8.12456533e-02
-3.15711796e-01 1.32190585e-01 4.78963166e-01 -2.06821755e-01
4.35950235e-02 8.35015997e-02 -1.07091188e-01 8.47167075e-02
-6.38941705e-01 2.12007552e-01 5.45318723e-01 1.41527879e+00
5.68620861e-01 -3.97167861e-01 -8.71575415e-01 7.63002157e-01
2.67533153e-01 6.30505443e-01 2.59994209e-01 -7.83272207e-01
8.56586099e-01 7.22692668e-01 4.51104678e-02 -8.56054604e-01
-2.40894064e-01 -4.22101200e-01 -6.07373416e-01 4.17729169e-02
2.82156497e-01 1.04753621e-01 -7.45222628e-01 1.57268155e+00
4.23440099e-01 7.71490782e-02 -1.17104918e-01 1.20411706e+00
8.63758981e-01 5.56346059e-01 1.24624297e-01 3.22701722e-01
1.66028142e+00 -1.30136275e+00 -5.60538590e-01 -4.20485407e-01
8.53374004e-01 -4.67313588e-01 1.14150703e+00 2.63223082e-01
-1.10018969e+00 -7.66695142e-01 -1.07144260e+00 -4.33578312e-01
-2.41338253e-01 6.14387214e-01 5.59343100e-01 3.14339936e-01
-6.72138274e-01 2.14982145e-02 -7.37445652e-01 -5.08386195e-01
7.20649540e-01 2.84281552e-01 -2.70869136e-01 -9.93083566e-02
-1.16442716e+00 6.07097328e-01 7.69150078e-01 7.60229528e-02
-1.04022932e+00 -7.20190108e-01 -1.15863132e+00 2.50351757e-01
8.31390977e-01 -8.17575455e-01 1.15249085e+00 -9.57173645e-01
-1.27938199e+00 8.45375419e-01 -1.72984466e-01 -5.52375674e-01
5.56665957e-01 -4.47712809e-01 -1.61254883e-01 4.72072482e-01
1.71903566e-01 1.05740535e+00 8.63815904e-01 -1.13829064e+00
-9.55517173e-01 -2.12049529e-01 1.36742204e-01 5.01355410e-01
-3.72592062e-01 2.73837298e-01 -1.13978207e+00 -6.27064884e-01
-3.59811157e-01 -9.91668582e-01 3.56134176e-01 2.22008511e-01
-6.36421144e-01 -3.19413811e-01 7.82597721e-01 -9.09683049e-01
1.19326723e+00 -2.34901881e+00 1.05231449e-01 9.88547653e-02
4.05034155e-01 3.27441096e-01 -1.71343029e-01 2.44371831e-01
9.10790637e-02 -2.62867033e-01 2.98710410e-02 -3.89212281e-01
1.61541685e-01 -1.90903887e-01 -9.14866254e-02 4.22524482e-01
3.96561682e-01 1.47243762e+00 -7.98416495e-01 -7.34054506e-01
3.89910966e-01 6.87690318e-01 -8.28639269e-01 3.72868091e-01
-1.05406195e-01 4.94145483e-01 -2.72395939e-01 9.18448389e-01
4.26394999e-01 -5.81546307e-01 1.64558040e-03 -5.27189016e-01
6.10686056e-02 4.05935831e-02 -9.65701759e-01 1.78371418e+00
-3.14831138e-01 7.64645696e-01 3.94261628e-02 -1.12681842e+00
4.45087254e-01 2.03223467e-01 2.16958538e-01 -7.73797810e-01
3.58161569e-01 -4.44082133e-02 -2.59753875e-03 -4.87901807e-01
2.17879340e-01 4.96298075e-01 -2.33698770e-01 -1.13635749e-01
2.93255836e-01 4.41707760e-01 2.04030260e-01 4.07995135e-01
1.11670983e+00 1.45939320e-01 3.97917777e-01 1.56092599e-01
5.74245632e-01 -7.43923560e-02 4.98886704e-01 9.21030581e-01
-4.27966893e-01 6.92418277e-01 1.86360508e-01 6.95273504e-02
-8.55714917e-01 -1.06365657e+00 5.14936745e-02 1.25536788e+00
4.65124875e-01 -6.10912263e-01 -7.29872048e-01 -8.26713681e-01
-2.54949927e-02 5.26994646e-01 -5.69635868e-01 -2.04601705e-01
-5.70087969e-01 -3.30189526e-01 2.86787629e-01 9.84041810e-01
8.24453712e-01 -1.15281892e+00 -7.46628821e-01 -8.18235353e-02
-5.54713011e-01 -1.42912257e+00 -7.62023270e-01 5.76154180e-02
-1.58655539e-01 -9.58536327e-01 -7.80976117e-01 -9.27106977e-01
4.93552536e-01 6.70888066e-01 9.15351272e-01 -7.24710003e-02
-8.35198104e-01 6.29475892e-01 -4.93099868e-01 -3.23401064e-01
4.68749041e-03 2.15729736e-02 -1.62349090e-01 4.16950434e-02
5.11775315e-01 -1.31605163e-01 -9.64479923e-01 3.82542640e-01
-4.96998549e-01 4.89455312e-01 9.73488867e-01 9.37975585e-01
5.02469182e-01 -2.00172171e-01 7.59809315e-02 -5.03840089e-01
2.09516231e-02 -4.81859326e-01 -3.91184360e-01 3.57106715e-01
-1.21979015e-02 -6.81959987e-02 3.71173918e-01 -4.11764562e-01
-1.39978838e+00 4.01970267e-01 2.76263833e-01 -6.71048820e-01
-1.89292163e-01 2.35965088e-01 -4.70433265e-01 -5.94958849e-02
5.44842303e-01 2.42748842e-01 -3.56750309e-01 -2.55373597e-01
4.99985456e-01 8.31810057e-01 8.55266273e-01 -3.68573755e-01
7.81472802e-01 5.47881663e-01 -4.39088583e-01 -7.21733809e-01
-1.09672546e+00 -8.25002253e-01 -6.32270515e-01 -1.90981358e-01
1.13041198e+00 -1.42034543e+00 -8.59784663e-01 2.72643983e-01
-1.17896187e+00 -4.99003261e-01 -9.66155156e-03 5.55224955e-01
-6.43594086e-01 3.08058470e-01 -4.60018575e-01 -6.76469028e-01
-4.03600782e-01 -9.42532659e-01 1.44507468e+00 2.73413509e-01
-2.04406962e-01 -4.17580813e-01 -3.78142864e-01 8.24388504e-01
-1.72499102e-02 2.01865174e-02 4.02559847e-01 -4.69961733e-01
-9.42212701e-01 -1.83365658e-01 -7.17780352e-01 1.94175526e-01
-3.21604699e-01 -5.06383538e-01 -9.13040102e-01 -2.60665238e-01
-3.22863996e-01 -6.14949405e-01 9.10864592e-01 2.02505052e-01
1.23159099e+00 -2.46918544e-01 -7.60020733e-01 7.28311539e-01
1.20252621e+00 3.86134945e-02 5.95660031e-01 6.67681769e-02
1.16085351e+00 5.09326339e-01 8.37932467e-01 6.80107117e-01
5.39574087e-01 1.11928725e+00 1.15637131e-01 -9.56650525e-02
-6.88127041e-01 -3.85594130e-01 6.19865298e-01 3.01006138e-01
2.20299102e-02 -3.09196830e-01 -1.09958088e+00 6.76093042e-01
-2.10367894e+00 -1.07345724e+00 -7.81271199e-04 2.05211759e+00
6.92103267e-01 1.47616401e-01 2.94687301e-01 -3.42229344e-02
7.52685785e-01 -2.49828488e-01 -5.13634741e-01 3.13275903e-01
-6.55725971e-02 -1.76146440e-02 5.77498555e-01 3.32616597e-01
-1.35298169e+00 1.12340009e+00 4.72487211e+00 9.54287291e-01
-7.00779617e-01 4.34062719e-01 2.10992321e-01 -3.39664340e-01
3.63633245e-01 -1.27243862e-01 -1.20822704e+00 5.86801410e-01
6.23958886e-01 9.12028830e-04 3.25622052e-01 8.68299603e-01
1.02645211e-01 -2.79861629e-01 -1.17925906e+00 1.48299515e+00
5.28289258e-01 -1.16829848e+00 -1.18405603e-01 -4.97723520e-02
5.09483933e-01 7.13704973e-02 -1.90538019e-02 6.15428388e-01
5.41856186e-03 -7.88355231e-01 8.55223656e-01 4.33486521e-01
8.14550459e-01 -6.00523353e-01 5.17444611e-01 3.46879154e-01
-1.55890810e+00 -3.80043328e-01 1.10123111e-02 2.05728412e-02
1.13699898e-01 2.40283847e-01 -9.18375432e-01 4.89515960e-01
9.33005691e-01 7.58742929e-01 -6.91638529e-01 1.13006973e+00
-4.21103954e-01 5.43012381e-01 -3.70254636e-01 1.34411275e-01
-4.13421221e-04 2.76942223e-01 4.66145128e-01 1.49388206e+00
1.08826436e-01 2.31710270e-01 4.76545423e-01 9.74282086e-01
-1.43857330e-01 6.76497668e-02 -5.44070125e-01 -8.06096569e-03
4.49623019e-01 1.12739551e+00 -5.75560153e-01 -4.97073859e-01
-8.93662035e-01 1.65397286e+00 5.37452638e-01 4.01657939e-01
-1.34165776e+00 -6.92286909e-01 3.67899090e-01 1.08225383e-01
6.68826163e-01 -9.46941152e-02 3.50517556e-02 -1.31019986e+00
2.09088594e-01 -7.07921863e-01 4.35255468e-01 -9.02530193e-01
-1.02925360e+00 1.55656934e-01 1.78107023e-01 -1.28314996e+00
-2.06337273e-01 -7.29426622e-01 -4.64065343e-01 5.84726214e-01
-1.32638419e+00 -1.64517772e+00 -7.87795305e-01 6.73718512e-01
7.34665215e-01 -5.62485196e-02 4.38395113e-01 4.49032009e-01
-6.91893220e-01 9.89986539e-01 -2.09005550e-01 4.78506833e-01
7.20979989e-01 -1.01889265e+00 1.47953436e-01 8.00017357e-01
1.93931103e-01 2.83095837e-01 4.49402958e-01 -6.48768187e-01
-1.44268262e+00 -1.24880552e+00 6.30502224e-01 -6.27576828e-01
3.90568882e-01 -8.50067198e-01 -7.94398129e-01 8.56692076e-01
1.08072571e-01 3.64506811e-01 4.98988867e-01 -7.42415935e-02
-4.38663453e-01 1.54940009e-01 -6.99054062e-01 6.33925736e-01
1.57015812e+00 -7.54845202e-01 -3.88445318e-01 4.76820052e-01
5.16833425e-01 -4.29014951e-01 -6.23433650e-01 4.39852953e-01
6.84520960e-01 -8.42258036e-01 1.20602143e+00 -6.70849383e-02
2.37148985e-01 -3.77235770e-01 -1.34876341e-01 -8.03486824e-01
-4.92781430e-01 -4.70603377e-01 -5.84503412e-01 1.14407206e+00
7.20197037e-02 -1.16089933e-01 7.45882630e-01 3.04866463e-01
-9.19521824e-02 -6.64904952e-01 -5.31457841e-01 -1.10489571e+00
-5.31576157e-01 -5.04576564e-01 7.47949928e-02 4.88748580e-01
2.78073013e-01 6.77251220e-01 -5.06071329e-01 3.35806340e-01
7.48184323e-01 2.18433887e-02 1.14913845e+00 -8.07131410e-01
-7.09599555e-01 -2.82678366e-01 -4.77240354e-01 -1.32843494e+00
2.72832483e-01 -1.05905759e+00 2.60006279e-01 -1.41799104e+00
7.97637463e-01 8.45282078e-02 -1.52761787e-01 6.41192555e-01
-3.36246282e-01 4.89808828e-01 4.51960623e-01 3.84514809e-01
-1.14125049e+00 7.43416846e-01 9.40386236e-01 -2.19735041e-01
-2.75648922e-01 -4.69450384e-01 -5.12869477e-01 7.34529316e-01
7.19959021e-01 -2.13952035e-01 -3.76930654e-01 -3.29792142e-01
-1.13373511e-01 -8.14704448e-02 9.62188482e-01 -1.40260363e+00
2.66767651e-01 2.84757137e-01 6.38245285e-01 -7.68906176e-01
3.62602741e-01 -5.85429430e-01 -2.15346709e-01 2.75084794e-01
-3.30813050e-01 -6.19872153e-01 1.07827358e-01 7.09555328e-01
-1.55872658e-01 7.85173196e-03 5.96350431e-01 1.00510985e-01
-1.12284899e+00 1.47119224e-01 -2.53508170e-03 1.46196783e-01
1.24180758e+00 -1.91201955e-01 -1.74900800e-01 -2.92157263e-01
-5.01979411e-01 5.40513277e-01 2.25642249e-01 5.51595390e-01
6.37055457e-01 -1.39184773e+00 -6.62763357e-01 2.06446752e-01
7.16941237e-01 -1.85416251e-01 4.80141103e-01 1.15794635e+00
-4.94691581e-02 6.13251090e-01 -9.13477391e-02 -6.68311834e-01
-1.45933568e+00 7.51615703e-01 1.03434928e-01 2.73504052e-02
-8.90588641e-01 9.15725112e-01 8.19229603e-01 6.24618679e-02
6.36708796e-01 -2.58042485e-01 1.52184203e-01 -1.90495849e-01
7.74714887e-01 2.22942337e-01 -3.28968495e-01 -8.24000776e-01
-4.70769405e-01 3.52085888e-01 -1.54389486e-01 -8.47800002e-02
7.47336984e-01 -2.71919876e-01 5.03803730e-01 1.13885283e-01
1.41095507e+00 -2.93351471e-01 -1.56336272e+00 -4.64609444e-01
-2.49105260e-01 -5.77737331e-01 -3.14381644e-02 -9.39791203e-01
-7.26423740e-01 1.01755643e+00 7.24831402e-01 -3.97301197e-01
1.11556685e+00 6.25133991e-01 8.73938203e-01 4.35480952e-01
1.98684901e-01 -1.08127964e+00 5.29547036e-01 3.88136625e-01
1.13926959e+00 -1.66411543e+00 -3.10890168e-01 -7.93720067e-01
-9.21254814e-01 7.13388205e-01 9.60638463e-01 9.94757637e-02
3.00181717e-01 1.49059847e-01 -4.54091817e-01 -2.49500185e-01
-4.90151584e-01 -6.94338322e-01 4.69059974e-01 6.74096525e-01
2.94014275e-01 -3.99866216e-02 -1.38256710e-03 9.34483469e-01
1.34700686e-01 1.74498782e-01 -1.14823930e-01 8.85358274e-01
-4.32196796e-01 -5.78471303e-01 -3.21400076e-01 2.55084068e-01
-5.43199889e-02 -2.21951142e-01 -2.47188285e-01 7.85523593e-01
3.31895769e-01 8.09995770e-01 2.15794332e-02 -3.75834823e-01
3.62450719e-01 9.20268372e-02 6.42965794e-01 -6.69688046e-01
-3.34995449e-01 1.02179103e-01 2.28600483e-02 -1.01288152e+00
-1.46223590e-01 -5.84043443e-01 -1.41116834e+00 1.79844573e-01
-3.88566375e-01 -3.94903958e-01 2.09260866e-01 8.56957436e-01
7.05769062e-01 5.30333281e-01 2.19381452e-01 -1.23007560e+00
-3.33371490e-01 -7.86933362e-01 -3.53557944e-01 6.73600256e-01
5.01782075e-02 -1.00687063e+00 -1.56642050e-01 2.24659175e-01] | [9.628889083862305, 1.4010441303253174] |
20f2c611-28f3-462a-a352-4fa55d339a52 | efficient-video-segmentation-models-with-per | 2202.12427 | null | https://arxiv.org/abs/2202.12427v1 | https://arxiv.org/pdf/2202.12427v1.pdf | Efficient Video Segmentation Models with Per-frame Inference | Most existing real-time deep models trained with each frame independently may produce inconsistent results across the temporal axis when tested on a video sequence. A few methods take the correlations in the video sequence into account,e.g., by propagating the results to the neighboring frames using optical flow or extracting frame representations using multi-frame information, which may lead to inaccurate results or unbalanced latency. In this work, we focus on improving the temporal consistency without introducing computation overhead in inference. To this end, we perform inference at each frame. Temporal consistency is achieved by learning from video frames with extra constraints during the training phase. introduced for inference. We propose several techniques to learn from the video sequence, including a temporal consistency loss and online/offline knowledge distillation methods. On the task of semantic video segmentation, weighing among accuracy, temporal smoothness, and efficiency, our proposed method outperforms keyframe-based methods and a few baseline methods that are trained with each frame independently, on datasets including Cityscapes, Camvid, and 300VW-Mask. We further apply our training method to video instance segmentation on YouTubeVISand develop an application of portrait matting in video sequences, by segmenting temporally consistent instance-level trimaps across frames. Experiments show superior qualitative and quantitative results. Code is available at: https://git.io/vidseg. | ['Jingdong Wang', 'Changqian Yu', 'Chunhua Shen', 'Yifan Liu'] | 2022-02-24 | null | null | null | null | ['image-matting', 'video-instance-segmentation'] | ['computer-vision', 'computer-vision'] | [-3.15047875e-02 -3.11115414e-01 -4.73150611e-01 -5.02162755e-01
-6.75502956e-01 -5.19910812e-01 3.26664388e-01 -3.71291906e-01
-5.24017274e-01 7.52577722e-01 -1.73674718e-01 -1.18155047e-01
8.68497938e-02 -6.29932404e-01 -1.01684153e+00 -6.27221286e-01
-1.19216785e-01 7.22923055e-02 6.77843750e-01 1.74111351e-01
5.55742644e-02 3.29675913e-01 -1.33893073e+00 4.12451595e-01
8.00839782e-01 1.19420207e+00 2.07090467e-01 7.63329804e-01
-1.56359822e-01 1.12070918e+00 -4.78877515e-01 -5.19941807e-01
3.23941290e-01 -3.70678276e-01 -9.30600524e-01 4.26783353e-01
7.09970355e-01 -8.73053133e-01 -5.23770392e-01 1.00651670e+00
1.19418852e-01 3.14862430e-01 1.40532762e-01 -1.48546350e+00
-1.60596207e-01 4.27720279e-01 -7.12829053e-01 5.26377738e-01
2.93005973e-01 3.08840483e-01 7.43668139e-01 -7.13881910e-01
8.24640393e-01 1.13725793e+00 7.21315026e-01 5.26279449e-01
-9.53886211e-01 -7.34272301e-01 7.16855586e-01 7.38269508e-01
-1.41605914e+00 -6.82105362e-01 5.42333126e-01 -4.20997798e-01
7.62335360e-01 1.98737770e-01 7.62916863e-01 9.61118400e-01
-3.96409668e-02 1.11949944e+00 6.33888483e-01 8.70723724e-02
1.63021639e-01 -4.94613767e-01 -5.13260104e-02 8.02486598e-01
-2.26508006e-01 -1.50646996e-02 -6.89207911e-01 1.46627948e-01
8.34942639e-01 9.76901799e-02 -5.25946081e-01 8.68622288e-02
-1.21381605e+00 4.85351682e-01 2.50537694e-01 9.96703953e-02
-2.91297048e-01 4.24757153e-01 5.45323968e-01 1.83892474e-01
7.49770999e-01 -3.85422260e-01 -5.34506381e-01 -2.65821129e-01
-1.56433523e+00 3.10999632e-01 6.37716293e-01 1.04770100e+00
8.14151943e-01 1.14123628e-01 -3.19039106e-01 4.91915792e-01
2.34442353e-01 3.22855055e-01 3.38030368e-01 -1.50331700e+00
4.62791175e-01 4.07154635e-02 9.71598998e-02 -9.92263794e-01
-1.18246995e-01 2.59095520e-01 -6.44873202e-01 -6.16564453e-02
4.91742700e-01 -2.18019366e-01 -1.07010520e+00 1.50354087e+00
6.50852978e-01 1.09268904e+00 -1.68519005e-01 1.16017401e+00
7.96344042e-01 8.21839929e-01 3.86558883e-02 -3.35115075e-01
1.03419566e+00 -1.29668188e+00 -8.49331617e-01 4.47775461e-02
5.42779744e-01 -6.89187050e-01 7.30447948e-01 4.20993984e-01
-1.33897698e+00 -7.14597881e-01 -7.44549096e-01 -2.15357825e-01
-1.10238798e-01 4.89686197e-03 4.41830009e-01 1.89550325e-01
-1.04859889e+00 8.32384169e-01 -1.30466461e+00 -1.02294259e-01
5.39961159e-01 1.98351562e-01 -1.29115820e-01 -7.63133764e-02
-1.15994179e+00 2.92347103e-01 4.45101082e-01 4.00234580e-01
-1.05471885e+00 -8.17661643e-01 -9.36236382e-01 -1.20028846e-01
6.01712644e-01 -5.89539230e-01 1.36868393e+00 -1.41624951e+00
-1.65224147e+00 4.41240966e-01 -4.23472673e-01 -7.27027416e-01
8.69070709e-01 -5.58424234e-01 -2.31154054e-01 5.02470911e-01
4.43680175e-02 9.05489564e-01 9.54224527e-01 -8.92400205e-01
-9.07440603e-01 -6.30266219e-02 3.71336907e-01 8.33924636e-02
-4.22779545e-02 -1.16319753e-01 -1.22722316e+00 -7.20469296e-01
-1.14187211e-01 -9.67694163e-01 -4.25671563e-02 2.97390282e-01
-3.78145576e-01 -2.01881640e-02 1.07351494e+00 -1.13764703e+00
1.30683112e+00 -2.03984404e+00 1.78565994e-01 2.94278227e-02
2.11324580e-02 2.60195941e-01 -8.70388523e-02 -1.23018555e-01
2.32161000e-01 1.08724110e-01 -2.69867361e-01 -5.22424519e-01
-2.65233099e-01 3.89630318e-01 -2.57286936e-01 5.41451693e-01
1.82530537e-01 9.25870597e-01 -9.21547651e-01 -8.96149814e-01
3.56940031e-01 6.01617038e-01 -7.58770347e-01 1.43468484e-01
-5.54459989e-01 5.37556410e-01 -3.28626990e-01 6.75967872e-01
7.59873331e-01 -3.49611640e-01 1.48253381e-01 -4.18922544e-01
-6.47722855e-02 1.13129415e-01 -1.11656523e+00 2.09686327e+00
-3.56868118e-01 8.56680810e-01 5.10880090e-02 -1.09467912e+00
2.51811862e-01 3.90408218e-01 7.80047059e-01 -6.47273600e-01
-1.65544953e-02 -1.43784001e-01 -3.67593288e-01 -7.17937231e-01
4.39040720e-01 3.85373503e-01 4.02521849e-01 1.52224258e-01
1.16420597e-01 1.94167197e-01 6.12071097e-01 2.37698779e-01
7.97537029e-01 6.81678832e-01 -1.25706047e-01 1.46146435e-02
4.63364691e-01 -2.06643902e-03 8.81999671e-01 4.89529729e-01
-2.49581888e-01 7.52707183e-01 4.54078943e-01 -5.80710411e-01
-9.17163134e-01 -8.68556619e-01 3.15767713e-02 8.38209212e-01
4.85761285e-01 -5.90415239e-01 -8.97153318e-01 -7.71425188e-01
-2.17771888e-01 3.26798856e-01 -4.72917527e-01 2.26925477e-01
-8.15458655e-01 -4.56478208e-01 4.21226919e-01 6.08578205e-01
9.38408077e-01 -6.36858523e-01 -6.53460264e-01 2.20998317e-01
-6.48467839e-01 -1.70935380e+00 -7.67523885e-01 -6.05279028e-01
-9.39589024e-01 -1.12339878e+00 -7.97425866e-01 -5.09724140e-01
4.14410025e-01 3.00682724e-01 1.14898789e+00 3.24805140e-01
-1.29124880e-01 3.23125482e-01 -3.46902966e-01 1.80485547e-01
-2.63059745e-03 -3.06242257e-02 -2.56471157e-01 6.43518567e-02
3.57846692e-02 -3.42313498e-01 -8.95089030e-01 5.06961167e-01
-1.00843191e+00 4.81685728e-01 9.93177369e-02 6.32210732e-01
8.39419186e-01 -2.84192134e-02 9.61768553e-02 -6.92147434e-01
-7.29031041e-02 -4.73716408e-01 -7.33896196e-01 3.16125542e-01
-9.45485607e-02 -1.29601821e-01 5.58970511e-01 -4.90270376e-01
-1.14862812e+00 2.66926456e-02 -1.79096490e-01 -1.00639522e+00
-7.16881230e-02 3.12476367e-01 -1.60059184e-02 1.53892547e-01
2.09501907e-01 1.94352075e-01 -4.08275314e-02 -1.73144817e-01
3.30790192e-01 2.31056929e-01 6.52702034e-01 -6.30613685e-01
5.41363478e-01 8.39638233e-01 -2.82932669e-01 -6.93930984e-01
-9.18686569e-01 -3.90820444e-01 -6.25981629e-01 -4.64041412e-01
1.03407967e+00 -1.08881497e+00 -7.06664681e-01 5.06669283e-01
-1.32681012e+00 -7.52484977e-01 1.71826722e-03 5.02801061e-01
-5.24741769e-01 5.49931824e-01 -7.97883093e-01 -4.36907560e-01
-2.42319584e-01 -1.21162665e+00 1.19427836e+00 2.88140148e-01
-1.17851064e-01 -1.14722848e+00 -1.95623964e-01 2.94983625e-01
6.67102635e-02 3.44522446e-01 1.90323859e-01 -1.51072115e-01
-9.49791789e-01 2.25503445e-01 -2.82538205e-01 3.79379153e-01
1.13745150e-03 4.66086835e-01 -8.09556603e-01 -1.73419073e-01
-1.69339657e-01 -7.17986748e-02 9.37693298e-01 7.52501130e-01
1.62235844e+00 -4.60133940e-01 -3.26322377e-01 9.87231791e-01
1.34770536e+00 2.28959292e-01 6.46029890e-01 2.96978682e-01
8.81259739e-01 3.57993007e-01 8.29731703e-01 5.01777053e-01
4.96279538e-01 6.90211535e-01 2.76813686e-01 5.36873341e-02
-2.05585957e-01 1.26330322e-02 4.60862517e-01 5.99849820e-01
-3.78491461e-01 -4.58139241e-01 -6.95460260e-01 5.10896325e-01
-2.23788285e+00 -1.19315851e+00 4.82530408e-02 1.97108269e+00
8.23888719e-01 1.96596265e-01 1.87673435e-01 -4.76371646e-02
7.93684423e-01 2.71023005e-01 -6.38207853e-01 7.79604912e-02
8.41286220e-03 -1.13554066e-02 4.38903868e-01 7.16780663e-01
-1.16112435e+00 1.09129345e+00 5.35616684e+00 8.29885125e-01
-1.25629473e+00 2.23186240e-01 1.02455699e+00 -4.89732742e-01
-1.35563329e-01 6.47764578e-02 -6.80707991e-01 7.55922616e-01
7.88419425e-01 5.16796261e-02 4.62620914e-01 5.76937675e-01
5.82647502e-01 -2.76247829e-01 -1.11601388e+00 1.09043872e+00
-1.17460545e-02 -1.62614453e+00 -6.01380989e-02 -2.82269567e-01
8.53997529e-01 -5.29174414e-03 -1.34665132e-01 6.67472109e-02
7.91155919e-02 -7.19318628e-01 9.49060798e-01 5.38105428e-01
6.13719046e-01 -6.79119170e-01 5.58966517e-01 2.42140498e-02
-1.36456013e+00 3.29639584e-01 -6.62672296e-02 7.99961314e-02
4.54735309e-01 5.82782865e-01 -4.19557959e-01 6.70334101e-01
1.10126197e+00 1.24970818e+00 -3.42733264e-01 1.00963449e+00
-1.59753367e-01 6.85873628e-01 -3.89355332e-01 4.45753783e-01
3.79038781e-01 -3.18334013e-01 2.83957541e-01 1.19394493e+00
3.40662420e-01 2.54494786e-01 3.40120673e-01 7.28376567e-01
-4.29571122e-02 -3.63386363e-01 -1.82620406e-01 1.78208560e-01
3.54685456e-01 1.16700947e+00 -9.06094491e-01 -7.68576205e-01
-6.90262854e-01 1.28811371e+00 1.56520624e-02 7.13597775e-01
-1.55306590e+00 -2.36347457e-03 8.24931085e-01 8.74240920e-02
4.59127456e-01 -4.53731507e-01 -6.96264207e-02 -1.38767064e+00
1.69033855e-01 -5.97486019e-01 5.07226765e-01 -7.47373283e-01
-8.48068655e-01 5.31738043e-01 2.78215051e-01 -1.18454063e+00
-3.27671736e-01 -2.47931644e-01 -6.01417065e-01 3.74609083e-01
-1.71293771e+00 -8.58949721e-01 -5.17358780e-01 8.47949266e-01
1.09657872e+00 3.06070387e-01 3.57804932e-02 5.58394074e-01
-8.19051862e-01 4.18652773e-01 -5.49566150e-02 3.97421330e-01
6.11781776e-01 -8.49618137e-01 5.33063531e-01 1.15379012e+00
2.45602340e-01 1.11402042e-01 5.49861193e-01 -5.78128874e-01
-1.17836821e+00 -1.29293776e+00 4.97641981e-01 -1.31645218e-01
6.36852562e-01 -8.63115862e-02 -1.01631367e+00 7.36962378e-01
2.32949287e-01 3.68634820e-01 2.89195716e-01 -3.70792955e-01
-7.60789216e-02 -1.00871101e-01 -9.00493026e-01 6.52700424e-01
1.32289863e+00 -3.37675720e-01 -9.00572687e-02 4.13811207e-01
8.42715383e-01 -8.23570728e-01 -8.71867895e-01 3.37173492e-01
5.55182576e-01 -9.99755859e-01 1.00935817e+00 -4.25722778e-01
6.22406423e-01 -4.68435317e-01 7.42441788e-02 -9.08547997e-01
5.86478971e-02 -8.18235815e-01 -4.64469969e-01 1.22856128e+00
1.51444906e-02 -2.14473143e-01 9.57793832e-01 9.19448018e-01
-7.38198981e-02 -7.10528851e-01 -9.05871809e-01 -7.73499310e-01
-3.05413783e-01 -6.20913327e-01 4.80685443e-01 7.93877602e-01
-4.52237964e-01 -1.57052562e-01 -4.48388457e-01 2.67605036e-01
5.61656117e-01 2.31161192e-01 7.42449999e-01 -7.12953389e-01
-2.75277168e-01 -3.19242269e-01 -3.11012745e-01 -1.34545982e+00
4.05120969e-01 -4.79217201e-01 6.45181313e-02 -1.31880164e+00
-5.67564778e-02 -9.20032859e-02 -3.29345278e-02 5.31390131e-01
-2.48470992e-01 3.62017930e-01 3.64902437e-01 3.23060304e-01
-9.46229577e-01 3.78386945e-01 1.31260383e+00 -2.59432971e-01
-9.94680077e-02 -2.25066394e-01 1.13469139e-01 9.11805868e-01
7.75551438e-01 -2.99070299e-01 -4.94728267e-01 -7.87731588e-01
-6.97770044e-02 2.65660584e-01 7.02507079e-01 -9.89847422e-01
3.28856409e-01 -3.47327113e-01 3.83059353e-01 -4.92166191e-01
4.28880006e-01 -7.42012441e-01 3.16776305e-01 4.35870379e-01
-1.25853226e-01 2.38106191e-01 2.72186071e-01 5.76583087e-01
-4.61529911e-01 -2.83515751e-02 7.26731122e-01 -8.38043466e-02
-1.31243849e+00 7.56970882e-01 -1.47344083e-01 1.91182047e-01
1.06332338e+00 -2.85211146e-01 -2.65095055e-01 -3.74226183e-01
-7.87075698e-01 4.38926816e-01 4.13996428e-01 4.14304286e-01
6.68691516e-01 -1.25858629e+00 -5.59422791e-01 2.03608461e-02
-4.21603650e-01 3.41824204e-01 5.98510206e-01 1.14229763e+00
-7.05082595e-01 1.49195597e-01 -1.29099870e-02 -9.84253287e-01
-1.24022031e+00 4.26035553e-01 3.57837319e-01 4.31025624e-02
-6.76790297e-01 8.11866462e-01 2.45845959e-01 2.46838197e-01
3.12572062e-01 -6.59854591e-01 8.78375024e-02 4.24497128e-02
5.83281696e-01 3.94159526e-01 -2.57508215e-02 -5.98874927e-01
-3.63392681e-01 6.77991331e-01 -9.06567648e-02 -1.02608368e-01
9.79467154e-01 -4.06128109e-01 1.12544917e-01 2.32128367e-01
1.38592923e+00 -4.18736786e-01 -1.84336293e+00 -1.40322849e-01
-1.60960928e-01 -8.03246498e-01 7.02428073e-02 -2.75094718e-01
-1.63367605e+00 6.28345609e-01 4.88293648e-01 -6.24188641e-03
1.26253045e+00 -2.27101207e-01 1.25147223e+00 7.09667280e-02
2.28701219e-01 -1.24470627e+00 -8.14499613e-03 3.65609139e-01
4.11679834e-01 -1.38142049e+00 3.76769155e-02 -4.79264408e-01
-5.77526808e-01 1.17824054e+00 6.49481773e-01 -9.41341072e-02
6.50580406e-01 3.42438787e-01 -1.03338793e-01 1.35580674e-01
-8.51515651e-01 -1.71863645e-01 2.30104432e-01 2.34191269e-01
3.78676385e-01 -3.05579513e-01 -2.84703523e-01 1.15456194e-01
1.70595646e-01 4.20508265e-01 3.43366146e-01 8.16302955e-01
-2.85360012e-02 -7.64381409e-01 -2.09049702e-01 2.07832530e-01
-5.47555268e-01 -6.23590089e-02 1.26890913e-01 8.56000304e-01
1.48198381e-01 8.64715636e-01 3.90795887e-01 -2.50030756e-01
-1.70948971e-02 -1.32141083e-01 4.34815317e-01 -6.37947246e-02
-2.52541363e-01 3.03069800e-01 5.19390963e-02 -1.07397223e+00
-9.94331181e-01 -7.21566677e-01 -1.40382218e+00 -6.18175328e-01
-1.71378478e-01 2.72430424e-02 3.31892312e-01 1.00062191e+00
4.17575419e-01 5.24417996e-01 3.73173326e-01 -1.15091491e+00
1.23678729e-01 -4.46429968e-01 -7.03817755e-02 5.67904592e-01
4.00379121e-01 -4.43126708e-01 -2.58426785e-01 7.43831635e-01] | [9.177655220031738, -0.10053399205207825] |
d709ca85-8a7e-4db6-8c5e-c705e62d11ee | stock-movement-prediction-from-tweets-and | null | null | https://aclanthology.org/P18-1183 | https://aclanthology.org/P18-1183.pdf | Stock Movement Prediction from Tweets and Historical Prices | Stock movement prediction is a challenging problem: the market is highly stochastic, and we make temporally-dependent predictions from chaotic data. We treat these three complexities and present a novel deep generative model jointly exploiting text and price signals for this task. Unlike the case with discriminative or topic modeling, our model introduces recurrent, continuous latent variables for a better treatment of stochasticity, and uses neural variational inference to address the intractable posterior inference. We also provide a hybrid objective with temporal auxiliary to flexibly capture predictive dependencies. We demonstrate the state-of-the-art performance of our proposed model on a new stock movement prediction dataset which we collected. | ['Yumo Xu', 'Shay B. Cohen'] | 2018-07-01 | null | null | null | acl-2018-7 | ['stock-trend-prediction', 'stock-market-prediction'] | ['time-series', 'time-series'] | [-4.09530073e-01 -2.92220414e-01 -3.53632450e-01 -2.18646646e-01
-9.18234169e-01 -5.65650403e-01 1.07425117e+00 -5.80676675e-01
-9.41641331e-02 7.77521551e-01 4.57376182e-01 -1.07348591e-01
-1.78118601e-01 -8.70782733e-01 -6.94978476e-01 -7.00541735e-01
-3.46320182e-01 9.84467506e-01 1.92802250e-01 -1.86028957e-01
2.47257322e-01 6.45409897e-02 -1.04968929e+00 -1.66719943e-01
6.80639029e-01 1.15363669e+00 1.45176888e-01 6.46379769e-01
-1.21853672e-01 1.16191983e+00 -4.46014136e-01 -5.42346001e-01
5.06458700e-01 -2.71455526e-01 -2.54959404e-01 1.65894553e-01
-3.66197042e-02 -6.27094388e-01 -5.86762488e-01 8.40074301e-01
1.18156999e-01 1.29847080e-01 8.65853131e-01 -1.34552097e+00
-9.70990777e-01 8.66197109e-01 -5.15800118e-01 5.92284620e-01
-3.58541340e-01 2.23588794e-01 1.41933405e+00 -6.22484505e-01
4.75487530e-01 1.08318830e+00 5.92698812e-01 1.27321213e-01
-1.50983620e+00 -8.78654242e-01 7.40621507e-01 3.29479203e-02
-1.04440522e+00 -1.90462068e-01 9.25372243e-01 -6.27222419e-01
1.00733578e+00 -2.22270399e-01 9.10129309e-01 1.63265204e+00
6.32806301e-01 1.12220633e+00 9.81689572e-01 3.56589675e-01
3.14061284e-01 -2.40383029e-01 4.25133072e-02 1.29069969e-01
-9.12631676e-02 2.52638429e-01 -7.86873817e-01 -3.65950644e-01
9.01528120e-01 8.27172816e-01 1.40474290e-01 -2.42475234e-02
-1.01021957e+00 1.21462715e+00 -1.08413704e-01 -6.61294982e-02
-7.40487099e-01 6.77484155e-01 3.13018635e-03 3.04080307e-01
1.03673410e+00 2.50179395e-02 -7.41173327e-01 -5.65140843e-01
-1.62904489e+00 6.43408775e-01 1.16782558e+00 1.01782084e+00
3.99963617e-01 5.33083797e-01 -2.07727104e-01 3.39403242e-01
7.13162303e-01 7.70741880e-01 6.83438778e-01 -6.75703704e-01
3.84176195e-01 -9.96082425e-02 4.56070065e-01 -5.17750263e-01
-3.27480763e-01 -7.60756612e-01 -7.09586263e-01 -6.96453005e-02
1.73787922e-01 -4.08468038e-01 -1.05920577e+00 1.66509068e+00
-1.98119462e-01 7.25766063e-01 3.85148935e-02 5.16859651e-01
2.31149957e-01 1.10287642e+00 7.61624426e-02 -3.80735219e-01
8.97758245e-01 -9.86597776e-01 -1.01601183e+00 -2.51912773e-01
-2.50035711e-02 -5.45304537e-01 6.27176166e-01 1.91930681e-01
-1.17363429e+00 -1.75227135e-01 -7.79412627e-01 1.35130614e-01
-2.91523516e-01 -1.28075749e-01 7.40323186e-01 1.53345078e-01
-1.03632557e+00 7.57684469e-01 -1.59425783e+00 2.98969448e-01
2.23520398e-01 1.96722642e-01 5.21073580e-01 5.66697955e-01
-1.37291825e+00 7.91033447e-01 -8.52959137e-03 2.35919073e-01
-9.68622923e-01 -8.24745417e-01 -5.57909012e-01 1.59898520e-01
2.53747910e-01 -6.61764860e-01 1.39675665e+00 -3.30175370e-01
-1.87317348e+00 2.02839643e-01 -3.32606316e-01 -1.07715225e+00
9.05333757e-01 -5.21350682e-01 -3.43086541e-01 -2.60855079e-01
1.54478788e-01 2.66780674e-01 1.02185059e+00 -7.61192322e-01
-4.44509506e-01 -8.59350339e-02 -4.72257465e-01 -6.22311607e-02
6.69807643e-02 -1.67732775e-01 -3.45373154e-01 -1.32915914e+00
3.29218656e-02 -1.13726175e+00 -4.65335906e-01 -5.12101293e-01
-3.30397129e-01 -3.11955988e-01 7.56973684e-01 -8.45400572e-01
1.32009709e+00 -1.73942435e+00 9.85164195e-02 3.27907205e-02
1.07913859e-01 -3.89866352e-01 1.03096873e-01 6.86276734e-01
3.48968565e-01 1.44211948e-01 -2.57790148e-01 -1.05059659e+00
5.89506567e-01 4.45750147e-01 -1.32391703e+00 2.59543747e-01
3.13943893e-01 1.33708906e+00 -5.39246202e-01 1.21730305e-02
-2.83943061e-02 4.51557875e-01 -5.14711976e-01 7.74297565e-02
-7.13618517e-01 3.98895144e-01 -6.27199531e-01 5.59040606e-01
5.02134085e-01 -6.39050364e-01 -5.06251082e-02 5.01085043e-01
-1.12999514e-01 4.65260655e-01 -1.04293251e+00 1.29851115e+00
-3.48390490e-01 6.64674759e-01 -5.37754476e-01 -6.34447694e-01
9.03841257e-01 2.90706187e-01 6.38133764e-01 -5.28530717e-01
-8.73070657e-02 2.01468710e-02 -3.29924166e-01 -3.39894444e-02
6.91218257e-01 -4.27336752e-01 -2.12670535e-01 7.15714991e-01
3.51735950e-02 3.18920203e-02 -2.61196420e-02 9.77926776e-02
7.92187333e-01 4.56647158e-01 -1.42427146e-01 -2.65180469e-01
-3.90936255e-01 -2.06139028e-01 8.12280715e-01 9.72791314e-01
-4.69566621e-02 4.06108677e-01 6.77633762e-01 -4.82532173e-01
-8.55456233e-01 -1.34843612e+00 -8.10580626e-02 9.63319719e-01
1.21854236e-02 -3.75121653e-01 3.35782319e-02 -2.97458589e-01
3.37813199e-01 9.55367208e-01 -7.97057390e-01 2.28838637e-01
-4.68335479e-01 -1.22835302e+00 2.75358241e-02 7.41822779e-01
3.18096817e-01 -8.88115287e-01 -3.69764686e-01 6.57553911e-01
4.25850935e-02 -9.73256528e-01 -5.43329835e-01 3.60710412e-01
-9.50759113e-01 -4.38538849e-01 -9.22517657e-01 -3.49507183e-01
-6.46287799e-02 -4.06971157e-01 1.37083840e+00 -5.11949658e-01
1.99784279e-01 3.34220946e-01 2.64991969e-02 -5.92993617e-01
-9.38305259e-02 2.12879747e-01 -8.78534615e-02 3.78985591e-02
2.90506452e-01 -9.62350488e-01 -7.26705432e-01 1.60104394e-01
-8.83044362e-01 -8.32076445e-02 4.18190897e-01 9.27511454e-01
7.59813011e-01 1.68511868e-02 5.25359154e-01 -7.38676548e-01
7.12938666e-01 -8.39089870e-01 -1.20215023e+00 9.18088704e-02
-7.44920969e-01 4.24673140e-01 2.34055340e-01 -8.22303414e-01
-9.89619970e-01 -1.69115245e-01 -6.29515722e-02 -7.53999352e-01
2.68405229e-01 8.29016209e-01 4.88698632e-01 6.42906249e-01
-3.30085844e-01 6.74681187e-01 -1.70638025e-01 -5.31426311e-01
3.55730176e-01 -3.46357822e-02 4.03641820e-01 -2.77191043e-01
9.22435522e-01 6.58478796e-01 -4.54675779e-02 -4.86539006e-01
-8.38920116e-01 -5.24806045e-02 -5.95195830e-01 7.63389766e-02
8.15315366e-01 -1.35554230e+00 -6.59180284e-01 6.28596067e-01
-1.02335501e+00 -7.82199919e-01 -4.13483620e-01 6.46926403e-01
-7.36616790e-01 -5.51806614e-02 -1.13700342e+00 -1.27563107e+00
-2.58145541e-01 -9.62234616e-01 1.14907455e+00 -8.53580236e-02
-2.08504111e-01 -1.52692378e+00 6.36577010e-01 -1.46948338e-01
7.10262239e-01 2.28636950e-01 4.99880373e-01 -8.67457807e-01
-1.14246178e+00 -2.32904717e-01 1.59513548e-01 -1.02091432e-02
-8.28084648e-02 1.82042733e-01 -6.29793465e-01 -9.50990617e-02
5.94740584e-02 -5.93618751e-02 1.34528017e+00 9.26588714e-01
6.34469032e-01 -4.79599059e-01 -1.98918581e-01 7.07792163e-01
1.23643744e+00 1.28963441e-01 3.26359928e-01 3.72794837e-01
3.79608482e-01 2.85615474e-01 3.73412669e-01 8.75713944e-01
7.89657950e-01 4.62874621e-01 2.63297975e-01 3.13405037e-01
6.38442934e-01 -6.22917116e-01 6.52866364e-01 8.23332191e-01
3.94197516e-02 -4.65480030e-01 -8.37603450e-01 5.25881052e-01
-2.23468900e+00 -1.32517374e+00 1.94092244e-01 1.54976892e+00
7.31764793e-01 3.64862323e-01 3.86760831e-01 -5.54759800e-01
2.85119534e-01 6.11078620e-01 -8.25423598e-01 2.70212978e-01
-2.21711695e-01 8.27250779e-02 8.61104727e-01 5.00954390e-01
-1.31424117e+00 1.16135180e+00 8.01351547e+00 6.90136015e-01
-1.14122629e+00 7.05516115e-02 5.92400908e-01 -5.62354803e-01
-5.84902763e-01 8.13420936e-02 -1.18494940e+00 9.86637354e-01
1.15957904e+00 -3.33496809e-01 2.68018186e-01 7.47050166e-01
3.18115830e-01 2.12549746e-01 -1.00469434e+00 8.10730994e-01
-2.54765451e-01 -1.62013781e+00 9.23901498e-02 5.16277671e-01
9.28440332e-01 5.32920957e-01 6.92715049e-01 4.40624416e-01
1.00446630e+00 -8.29416037e-01 1.16320276e+00 1.11371291e+00
-2.04514340e-02 -7.30133057e-01 3.96330416e-01 4.51533347e-01
-1.14870214e+00 5.01165390e-02 -1.86849147e-01 -2.59760678e-01
8.03825855e-01 4.70766783e-01 -2.64168531e-01 2.63708364e-03
6.04177237e-01 1.27266335e+00 -2.20447898e-01 6.91769719e-01
-2.58587718e-01 9.45247710e-01 -5.24623036e-01 -3.08015980e-02
5.73749185e-01 -5.25934815e-01 6.44653082e-01 9.20714021e-01
5.56707084e-01 -1.24774039e-01 3.01760018e-01 1.38030052e+00
5.47786243e-04 -4.61754501e-01 -4.24963415e-01 -2.74000287e-01
2.70562798e-01 6.08853936e-01 -7.77781129e-01 -3.87501687e-01
-6.16327345e-01 9.95422363e-01 1.28711745e-01 7.07949162e-01
-1.09932244e+00 2.06459954e-01 8.97355437e-01 -1.44313946e-01
9.41340685e-01 -6.48902416e-01 -1.36172146e-01 -1.72088027e+00
-8.37620646e-02 -6.37241378e-02 3.64033163e-01 -5.78975439e-01
-1.79530668e+00 4.34621692e-01 7.60896653e-02 -1.23949432e+00
-9.83231187e-01 -3.83554757e-01 -1.01554585e+00 1.06958735e+00
-1.73503804e+00 -1.07494879e+00 5.36094606e-01 5.28537393e-01
8.37056875e-01 -2.99561113e-01 2.90295362e-01 -6.35048226e-02
-5.68941176e-01 -3.64566781e-02 5.25861204e-01 1.63516119e-01
3.12895000e-01 -1.53793526e+00 1.05388427e+00 9.12461996e-01
4.78726089e-01 5.05692005e-01 9.24390793e-01 -1.00727987e+00
-1.20948577e+00 -1.17478442e+00 7.21900225e-01 -7.59023428e-01
1.52781248e+00 -4.42138433e-01 -8.45851541e-01 1.37869370e+00
2.42965907e-01 -2.88942605e-01 7.65951037e-01 1.09517723e-01
-1.41677365e-01 3.01149189e-01 -4.96376514e-01 4.88511980e-01
7.14652300e-01 -4.87478077e-01 -8.16900134e-01 2.74839818e-01
9.46349263e-01 -1.45361751e-01 -9.07882214e-01 -6.97423071e-02
5.61985612e-01 -5.93757331e-01 8.50652635e-01 -6.46657467e-01
3.64059061e-01 2.31887773e-02 -4.84051108e-02 -1.34615767e+00
-4.53979015e-01 -1.13839614e+00 -6.89757466e-01 1.04793549e+00
7.97810555e-01 -8.77329350e-01 9.31628287e-01 8.79417181e-01
2.30369955e-01 -5.94740212e-01 -8.46970916e-01 -9.34631884e-01
3.58289450e-01 -5.76423168e-01 7.20073462e-01 7.88904905e-01
-3.04312259e-01 1.96444780e-01 -9.55081701e-01 2.65739083e-01
6.66621447e-01 6.85544789e-01 5.42187512e-01 -1.25197756e+00
-6.62343621e-01 -6.03395045e-01 -1.09008566e-01 -1.67768526e+00
3.46242636e-01 -4.40848291e-01 -3.70638408e-02 -1.23901320e+00
1.12509415e-01 7.14187175e-02 -3.37977469e-01 3.35771702e-02
-2.04096381e-02 -4.04440574e-02 3.40071261e-01 7.26060331e-01
-6.70391083e-01 1.17487538e+00 9.34600115e-01 -1.01048544e-01
-3.36881250e-01 4.96486992e-01 -3.79987448e-01 5.09708405e-01
5.87990165e-01 -5.94930530e-01 -3.53811264e-01 -2.43756399e-01
4.15630162e-01 3.33431244e-01 4.63250756e-01 -3.00795227e-01
4.11392182e-01 -3.21412385e-01 4.80274618e-01 -1.40446067e+00
5.74658513e-01 -5.23100197e-01 2.98862040e-01 4.07480538e-01
-3.71909112e-01 4.54811454e-01 7.65779689e-02 1.10175657e+00
-4.71628428e-01 4.20261860e-01 3.03039670e-01 -1.00685805e-01
-4.75579083e-01 8.69358540e-01 -5.81283450e-01 1.90065086e-01
7.53752351e-01 2.66188473e-01 -1.04546994e-01 -9.13877666e-01
-1.19451451e+00 6.70024395e-01 6.95721507e-02 4.13822442e-01
4.76249337e-01 -1.32002425e+00 -7.69197226e-01 1.07505977e-01
-3.35624665e-01 -1.99063927e-01 2.28863120e-01 8.68025541e-01
-1.43623263e-01 4.20916975e-01 2.21634835e-01 -5.15978396e-01
-3.20883095e-01 5.98120809e-01 2.13508993e-01 -7.44399965e-01
-8.16927195e-01 6.55857205e-01 2.23077476e-01 8.28873962e-02
1.29282743e-01 -7.72112370e-01 -3.55080329e-02 3.39414299e-01
2.26513579e-01 2.50978023e-01 -5.14180481e-01 -3.62421989e-01
-6.72054524e-03 3.21280062e-01 -1.31017029e-01 -5.74141145e-01
1.78356338e+00 -3.21476430e-01 2.39394739e-01 1.10420728e+00
9.51344848e-01 -3.60168606e-01 -2.02250075e+00 -4.54298675e-01
2.68329620e-01 -2.53600199e-02 1.99640080e-01 -4.69382286e-01
-1.01867127e+00 7.08393574e-01 1.22933768e-01 6.30302429e-01
6.40947461e-01 1.12871647e-01 9.84144270e-01 1.56792298e-01
2.92524341e-02 -1.13263988e+00 -1.32691205e-01 8.38499427e-01
7.82394409e-01 -1.31230521e+00 -1.61183849e-01 1.67031735e-01
-9.22077656e-01 9.28538620e-01 -2.73241699e-02 -6.56504631e-01
1.41404271e+00 4.67260331e-01 -1.07467048e-01 -2.18616918e-01
-1.42536926e+00 -2.05394104e-01 3.24006975e-01 9.54799727e-02
9.07631069e-02 2.56501697e-02 1.64686173e-01 1.01170492e+00
-5.76268196e-01 -4.12375443e-02 3.27692866e-01 9.39407587e-01
-1.52902618e-01 -8.51379633e-01 4.22438420e-02 5.03193676e-01
-7.29271173e-01 -4.07598674e-01 1.84284970e-02 8.93349409e-01
-5.87915599e-01 4.97829825e-01 6.46303177e-01 -1.22213811e-01
-9.62703601e-02 2.76997864e-01 -7.84402490e-02 -3.46144855e-01
-2.37112805e-01 8.32117796e-01 -4.85355079e-01 -3.89326930e-01
-3.07632327e-01 -1.14909077e+00 -7.19931602e-01 -3.90728623e-01
5.70169948e-02 3.84300277e-02 3.74572664e-01 1.04153407e+00
2.84900665e-01 5.48783958e-01 7.49106050e-01 -1.05883384e+00
-9.72317696e-01 -9.94022310e-01 -1.01079214e+00 1.33510441e-01
7.44643807e-01 -6.13939583e-01 -5.43663502e-01 1.19532987e-01] | [6.8355793952941895, 3.4699034690856934] |
f4fdb41b-f5af-4655-aacc-243d38a8ea0e | chart-rcnn-efficient-line-chart-data | 2211.14362 | null | https://arxiv.org/abs/2211.14362v1 | https://arxiv.org/pdf/2211.14362v1.pdf | Chart-RCNN: Efficient Line Chart Data Extraction from Camera Images | Line Chart Data Extraction is a natural extension of Optical Character Recognition where the objective is to recover the underlying numerical information a chart image represents. Some recent works such as ChartOCR approach this problem using multi-stage networks combining OCR models with object detection frameworks. However, most of the existing datasets and models are based on "clean" images such as screenshots that drastically differ from camera photos. In addition, creating domain-specific new datasets requires extensive labeling which can be time-consuming. Our main contributions are as follows: we propose a synthetic data generation framework and a one-stage model that outputs text labels, mark coordinates, and perspective estimation simultaneously. We collected two datasets consisting of real camera photos for evaluation. Results show that our model trained only on synthetic data can be applied to real photos without any fine-tuning and is feasible for real-world application. | ['Haoshuai Zhou', 'Linkai Li', 'Congxi Lu', 'Shufan Li'] | 2022-11-25 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 6.18046522e-01 -1.95154354e-01 -7.10165501e-02 -5.32850146e-01
-6.82091117e-01 -1.00230801e+00 5.72989225e-01 7.07362145e-02
-1.81947559e-01 4.86277282e-01 -5.76876216e-02 -2.85648137e-01
3.19411218e-01 -7.71609962e-01 -1.02866876e+00 -1.17356412e-01
4.83066499e-01 2.89439917e-01 2.93718010e-01 -1.11110054e-01
7.85164952e-01 7.49111474e-01 -1.36335385e+00 7.28087068e-01
8.24366868e-01 7.98236549e-01 5.93203753e-02 1.03621221e+00
-2.78402060e-01 9.50296462e-01 -8.26955557e-01 -5.52457035e-01
4.00171757e-01 -2.35546872e-01 -4.78543699e-01 6.82302356e-01
8.57923567e-01 -7.68701792e-01 -2.78049409e-01 1.20714808e+00
3.29807371e-01 -1.41468719e-01 6.51498914e-01 -1.24514425e+00
-1.26040280e+00 3.60710293e-01 -5.50736248e-01 -3.68203640e-01
5.34395456e-01 1.27299279e-01 7.14303136e-01 -1.03606200e+00
6.18365347e-01 9.27544713e-01 6.69508755e-01 5.31279027e-01
-1.29203749e+00 -3.19297016e-01 -2.03404706e-02 7.59450579e-03
-1.00254750e+00 -4.87414449e-01 9.71717894e-01 -5.00343561e-01
5.80021203e-01 4.31538403e-01 5.16131580e-01 1.04603970e+00
-3.39109302e-01 1.18888211e+00 1.14321554e+00 -8.09342146e-01
3.11113715e-01 3.27744693e-01 1.00851201e-01 6.94606781e-01
3.87979925e-01 -4.09314752e-01 -3.33366930e-01 4.03413773e-01
1.08477807e+00 1.27935708e-02 -2.73153961e-01 -7.22942650e-01
-1.29870903e+00 4.19213384e-01 5.38655184e-02 -5.45735918e-02
8.07821155e-02 4.95546833e-02 2.42176339e-01 2.43042260e-01
1.94174312e-02 5.61215043e-01 -1.98667988e-01 -2.04707295e-01
-1.25740361e+00 3.39053184e-01 6.59412920e-01 1.63134611e+00
8.42242897e-01 -3.28073204e-02 -1.47292763e-01 7.75306046e-01
5.95297441e-02 5.22580028e-01 3.78844261e-01 -8.14282417e-01
9.13617134e-01 1.00994754e+00 2.00927839e-01 -1.04725301e+00
-3.16410482e-01 5.88300526e-02 -5.69140375e-01 3.53315502e-01
6.36373103e-01 -2.94734746e-01 -1.23135030e+00 9.88553047e-01
-1.45320818e-01 -3.52720797e-01 -7.96819776e-02 7.66222298e-01
6.31363690e-01 5.28874934e-01 -5.27881205e-01 3.44881117e-01
1.23584354e+00 -1.30583382e+00 -7.12531209e-01 -3.53238642e-01
6.19809151e-01 -1.07762492e+00 1.48304129e+00 5.34288943e-01
-9.32599604e-01 -5.54710627e-01 -1.38300169e+00 -5.48622131e-01
-6.12765551e-01 9.91689086e-01 5.46934128e-01 9.14919198e-01
-9.87628341e-01 3.86797190e-01 -5.97604692e-01 -4.80369568e-01
4.62245226e-01 5.13703004e-02 -4.22066569e-01 -1.20113909e-01
-4.46886420e-01 5.06983757e-01 5.55919409e-01 2.54189074e-01
-5.05083442e-01 -4.10546273e-01 -8.24412346e-01 1.06571494e-02
6.92889869e-01 -1.25490338e-01 1.37230647e+00 -9.90201533e-01
-1.75860214e+00 6.43102586e-01 7.70569593e-02 -1.07539348e-01
9.02495980e-01 -4.14147764e-01 -3.59977037e-01 3.24811518e-01
-2.18786672e-02 6.02886677e-01 8.31035137e-01 -1.30322182e+00
-4.80753779e-01 -2.30299175e-01 1.25145301e-01 5.46781980e-02
-4.58005190e-01 1.44534176e-02 -9.23815131e-01 -7.83664823e-01
-9.37935058e-03 -9.57292855e-01 1.42904334e-02 4.38661247e-01
-9.23361182e-01 3.24970752e-01 9.08900321e-01 -7.00768709e-01
1.19485128e+00 -1.83359146e+00 -3.04812878e-01 8.07883665e-02
-1.67530984e-01 4.96204048e-01 -2.18696028e-01 5.85007608e-01
-4.81625535e-02 2.43763208e-01 -3.62460256e-01 -2.41298437e-01
8.36901646e-03 -2.13761970e-01 -4.95079607e-01 2.01925382e-01
4.60249186e-01 7.71729469e-01 -5.65590084e-01 -5.99525392e-01
4.74728495e-01 1.46444216e-01 -4.90057200e-01 2.24715620e-01
-4.83697593e-01 -4.87637110e-02 -1.03193171e-01 9.49977160e-01
9.37429547e-01 -2.85786748e-01 1.37370542e-01 -1.79240808e-01
-2.81178445e-01 -1.77355513e-01 -1.50796890e+00 1.79969573e+00
-1.13324538e-01 1.09980154e+00 -4.13963288e-01 -5.73048294e-01
1.11417389e+00 -6.07566796e-02 -1.09056877e-02 -5.97224653e-01
8.25538635e-02 -3.16870064e-02 -5.03571153e-01 -6.52397275e-01
1.14244211e+00 5.60621619e-01 -1.38632348e-02 4.14898455e-01
-2.28271708e-01 -3.71537745e-01 6.70837462e-01 1.68448970e-01
8.88419032e-01 6.58521891e-01 1.40682802e-01 2.79875547e-01
6.30696297e-01 4.89138216e-01 3.57420444e-01 7.18831301e-01
7.61236846e-02 1.21764147e+00 8.47348094e-01 -4.64863539e-01
-1.56942558e+00 -7.27005064e-01 1.90408543e-01 6.85221672e-01
1.59812272e-01 -5.20905495e-01 -9.70972300e-01 -6.72755480e-01
-2.74363846e-01 7.71745801e-01 -4.67837960e-01 4.78274494e-01
-6.87269568e-01 -5.71537673e-01 6.45663440e-01 8.84586036e-01
6.37106538e-01 -9.75381672e-01 -5.39495349e-01 -1.76973343e-01
2.46540591e-01 -1.44265306e+00 -6.17683351e-01 -8.54173973e-02
-8.76058638e-01 -1.09286761e+00 -8.94851625e-01 -9.53641891e-01
1.17153895e+00 3.54323089e-01 6.75930619e-01 -6.70886263e-02
-1.20510839e-01 1.57054394e-01 -3.37325573e-01 -5.72848320e-01
-3.79331410e-01 1.15352891e-01 -5.02707481e-01 1.44207090e-01
3.11164230e-01 -1.14593901e-01 -4.54222679e-01 2.16228336e-01
-1.28537607e+00 5.66131651e-01 7.28565931e-01 5.24115324e-01
4.82342005e-01 -2.43356317e-01 1.73332289e-01 -1.18212593e+00
7.49294102e-01 2.98036665e-01 -1.01578867e+00 7.13065863e-01
-4.45178628e-01 2.99420327e-01 8.89110446e-01 -3.59437615e-01
-1.19613349e+00 6.04757130e-01 3.21941435e-01 -2.03624681e-01
-4.48094904e-01 1.59154892e-01 -3.09097171e-01 8.69861469e-02
5.95460534e-01 2.66342759e-01 -2.18006641e-01 -6.08428717e-01
6.19967401e-01 8.75642955e-01 8.14767182e-01 -4.22547817e-01
1.03956068e+00 5.80509543e-01 -1.46402821e-01 -8.63780320e-01
-5.66058815e-01 -3.08380336e-01 -1.12041938e+00 -1.68067798e-01
7.49687850e-01 -7.52914190e-01 -5.97343147e-01 7.44619071e-01
-1.31754935e+00 -3.04818213e-01 -1.40530244e-01 3.18950951e-01
-5.43468952e-01 7.21623659e-01 -3.94730300e-01 -6.58886015e-01
-1.35244265e-01 -1.16574204e+00 1.14782178e+00 4.92159814e-01
1.61825120e-01 -7.11654484e-01 3.58676095e-03 3.81127656e-01
9.21698660e-02 4.53197181e-01 9.26153898e-01 -3.93301487e-01
-1.00196755e+00 -6.17612183e-01 -7.00426042e-01 4.59786147e-01
7.94130042e-02 6.43638313e-01 -9.57811177e-01 1.43769398e-01
-4.89459544e-01 -5.92855155e-01 5.53172112e-01 -8.12637880e-02
1.27834713e+00 -2.60716587e-01 -1.50266644e-02 6.69868529e-01
1.62827981e+00 3.54385495e-01 8.28179479e-01 5.41570544e-01
1.16044831e+00 5.05679905e-01 3.57496947e-01 1.31343648e-01
4.38508958e-01 4.30197746e-01 8.05723220e-02 -1.76637188e-01
-1.16485976e-01 -6.92245066e-01 2.93980151e-01 6.34876370e-01
9.91814956e-02 -5.01238406e-01 -1.05393386e+00 5.77255547e-01
-1.87709177e+00 -8.39304447e-01 -2.48623267e-01 2.12463617e+00
6.61568940e-01 2.38658994e-01 5.64578250e-02 2.02508718e-01
9.84256744e-01 1.17975593e-01 -5.02627194e-01 -5.42945981e-01
-2.80422896e-01 -1.64547309e-01 7.33493388e-01 4.97003272e-02
-1.13324702e+00 8.88984978e-01 6.31387711e+00 4.73457426e-01
-1.09876251e+00 -6.51388228e-01 5.42298317e-01 1.41342267e-01
-1.68280855e-01 2.37078309e-01 -6.80493832e-01 2.60754257e-01
6.01221919e-01 2.36520935e-02 2.91351795e-01 1.04229558e+00
2.11129300e-02 -4.97951880e-02 -1.42310238e+00 1.23187101e+00
4.15252596e-01 -1.53287411e+00 3.99528950e-01 -1.03243843e-01
9.22106683e-01 -3.03553671e-01 -3.63187157e-02 6.62911087e-02
2.60059595e-01 -9.19334710e-01 9.51955140e-01 6.34644389e-01
1.14483380e+00 -4.37511563e-01 3.12022924e-01 2.60023147e-01
-9.06768441e-01 1.38219045e-02 -3.56062979e-01 9.72815696e-03
1.10473417e-01 1.97529718e-01 -1.07722223e+00 3.35141510e-01
3.12110424e-01 7.31899261e-01 -1.36777294e+00 1.22539306e+00
-5.26885629e-01 3.32477361e-01 -7.38701224e-02 -3.48528147e-01
9.72180739e-02 -2.04728544e-01 1.28487408e-01 1.33716583e+00
3.52542311e-01 -3.74669015e-01 -6.84940964e-02 1.01623583e+00
-3.75060588e-01 2.00811163e-01 -6.60958052e-01 -3.00931007e-01
2.63311118e-01 1.36896646e+00 -1.04839134e+00 -3.56740803e-01
-6.19118810e-01 9.94061470e-01 2.82649577e-01 3.99619669e-01
-7.95332909e-01 -9.18442667e-01 -1.49921879e-01 6.59020171e-02
3.35034430e-01 -3.25495750e-01 -7.51533210e-01 -1.56883669e+00
3.76780003e-01 -9.47767138e-01 -1.14093848e-01 -1.10259247e+00
-7.85420597e-01 5.26549101e-01 -2.06669927e-01 -1.60880506e+00
-2.02748105e-01 -1.07303226e+00 -4.81600553e-01 4.80920315e-01
-1.55783010e+00 -1.18745136e+00 -6.73201144e-01 2.35794067e-01
8.65882754e-01 -9.52280313e-02 6.00953639e-01 2.49593765e-01
-8.19794476e-01 6.94157481e-01 3.95691276e-01 9.06609952e-01
9.26993251e-01 -1.60432303e+00 9.35278118e-01 1.21365929e+00
3.09284627e-01 4.85729277e-01 3.98805290e-01 -5.46312630e-01
-1.42182422e+00 -1.05565512e+00 4.94224846e-01 -5.78329623e-01
4.78310466e-01 -8.17087948e-01 -7.56842017e-01 7.33940244e-01
1.36086747e-01 1.35390665e-02 3.82584274e-01 -4.82260048e-01
-3.16870719e-01 -9.22801197e-02 -8.67887914e-01 9.18137729e-01
7.67190278e-01 -4.90828454e-01 -4.62434858e-01 2.94226676e-01
4.11654711e-01 -6.14434004e-01 -4.78618771e-01 8.63739923e-02
7.92291462e-01 -1.13790822e+00 5.99490523e-01 -4.12049919e-01
7.84188151e-01 -5.89655161e-01 -7.27054328e-02 -9.25122142e-01
2.37651452e-01 -7.21212327e-01 1.64919406e-01 1.36520970e+00
6.45840228e-01 -7.10340962e-02 9.47832227e-01 1.00824273e+00
-3.00208782e-03 -3.11992884e-01 -1.78680763e-01 -5.45431674e-01
-3.04110765e-01 -4.07695234e-01 7.72252202e-01 8.15886736e-01
-1.67290494e-01 3.64195615e-01 -5.14685869e-01 7.45749334e-03
4.52489853e-01 2.81514883e-01 1.33004320e+00 -1.04762423e+00
3.61214764e-03 -2.36027345e-01 -4.65422243e-01 -1.12932396e+00
-3.46487522e-01 -4.93784219e-01 6.44563138e-02 -1.79166818e+00
-1.01937773e-03 -1.62552312e-01 2.68574148e-01 2.98295826e-01
2.61538923e-02 3.37607533e-01 4.77786481e-01 1.86292812e-01
-5.17476797e-01 1.40945092e-01 1.25728774e+00 -1.54773831e-01
-3.64645243e-01 -1.04077846e-01 -5.64287305e-01 1.00781786e+00
7.44599342e-01 -2.39307299e-01 -4.83588070e-01 -7.78035343e-01
4.64106709e-01 8.94163102e-02 2.34818548e-01 -1.22480726e+00
4.17997122e-01 -3.07069868e-01 8.40671480e-01 -1.00562954e+00
2.29294635e-02 -7.33879387e-01 -4.21681315e-01 -9.03231800e-02
-5.82232475e-01 3.22534323e-01 6.18983693e-02 5.48782587e-01
-1.13800332e-01 -6.13813758e-01 5.42630374e-01 -1.68173194e-01
-7.85467684e-01 -1.03977509e-01 -1.10189199e-01 -6.82023913e-02
1.08743954e+00 -5.79585195e-01 -7.43606031e-01 -3.34354579e-01
-1.25767946e-01 -3.24117169e-02 8.45699310e-01 4.96330142e-01
7.01618016e-01 -1.23482120e+00 -4.29711491e-01 2.24873617e-01
4.82661307e-01 3.51769418e-01 -1.38930947e-01 1.80375308e-01
-1.38772285e+00 5.98436296e-01 -4.43976521e-01 -4.99940842e-01
-1.05083656e+00 7.40578413e-01 1.87759325e-02 -9.95414332e-03
-5.95189929e-01 2.48926520e-01 -8.45289305e-02 -4.40927118e-01
2.36605525e-01 -8.08403134e-01 -2.31454507e-01 -1.07216472e-02
6.20244026e-01 4.83705968e-01 2.81440653e-02 -3.33727419e-01
-2.21307650e-02 7.66828895e-01 -4.03431356e-01 -3.46260428e-01
1.24557590e+00 -1.10953553e-02 8.93101916e-02 3.65897030e-01
1.06383789e+00 1.84706420e-01 -1.55823171e+00 -1.62932783e-01
3.43659133e-01 -4.72793847e-01 -3.42223406e-01 -8.26255202e-01
-8.02520692e-01 1.00575793e+00 3.66508394e-01 6.47300035e-02
1.09062505e+00 -6.79331303e-01 6.57726347e-01 9.06431019e-01
1.36939511e-02 -1.70929480e+00 3.07012051e-01 2.79357135e-01
8.31096530e-01 -1.36052513e+00 2.41324574e-01 -3.56234580e-01
-8.29358339e-01 1.78709459e+00 7.91139603e-01 -2.40632623e-01
2.41304785e-02 2.99902707e-01 2.23181114e-01 3.29589583e-02
-4.02647823e-01 -1.08658699e-02 4.08127427e-01 5.95437586e-01
3.60241711e-01 -1.67005554e-01 -9.93895829e-02 2.30705068e-01
-2.08567873e-01 2.01781273e-01 1.33949673e+00 1.19962037e+00
-1.76219031e-01 -1.18289983e+00 -5.37805796e-01 3.43675226e-01
-2.62587935e-01 -1.94619354e-02 -8.09655964e-01 9.52440083e-01
-2.26821423e-01 7.36653090e-01 3.45054343e-02 -3.08652997e-01
4.46374774e-01 9.92955863e-02 2.93485641e-01 -6.95456862e-01
-1.32000387e-01 -1.33772064e-02 -1.58410650e-02 -1.98450550e-01
-4.39102858e-01 -5.37607312e-01 -1.11565506e+00 1.12042623e-02
-3.52468073e-01 -2.73355722e-01 9.30854321e-01 5.67722678e-01
3.14734250e-01 4.05215979e-01 4.64595944e-01 -6.14673793e-01
-3.71157825e-01 -8.91612530e-01 -3.61399770e-01 4.36826646e-01
5.26462555e-01 -8.66558477e-02 9.34880823e-02 6.80300772e-01] | [11.621419906616211, 2.255239725112915] |
e7df1829-68cf-4e5b-a384-502c8f12643c | discohead-audio-and-video-driven-talking-head | 2303.07697 | null | https://arxiv.org/abs/2303.07697v1 | https://arxiv.org/pdf/2303.07697v1.pdf | DisCoHead: Audio-and-Video-Driven Talking Head Generation by Disentangled Control of Head Pose and Facial Expressions | For realistic talking head generation, creating natural head motion while maintaining accurate lip synchronization is essential. To fulfill this challenging task, we propose DisCoHead, a novel method to disentangle and control head pose and facial expressions without supervision. DisCoHead uses a single geometric transformation as a bottleneck to isolate and extract head motion from a head-driving video. Either an affine or a thin-plate spline transformation can be used and both work well as geometric bottlenecks. We enhance the efficiency of DisCoHead by integrating a dense motion estimator and the encoder of a generator which are originally separate modules. Taking a step further, we also propose a neural mix approach where dense motion is estimated and applied implicitly by the encoder. After applying the disentangled head motion to a source identity, DisCoHead controls the mouth region according to speech audio, and it blinks eyes and moves eyebrows following a separate driving video of the eye region, via the weight modulation of convolutional neural networks. The experiments using multiple datasets show that DisCoHead successfully generates realistic audio-and-video-driven talking heads and outperforms state-of-the-art methods. Project page: https://deepbrainai-research.github.io/discohead/ | ['Gyeongsu Chae', 'Sungwoo Park', 'SeungHyun Lee', 'Sunwon Hong', 'Geumbyeol Hwang'] | 2023-03-14 | null | null | null | null | ['talking-head-generation'] | ['computer-vision'] | [-2.73815274e-01 3.93427461e-01 -8.43001753e-02 -2.57993698e-01
-7.87135839e-01 -4.14088845e-01 5.35542786e-01 -9.19107735e-01
-2.51445740e-01 4.74598438e-01 5.25220811e-01 1.90123767e-01
6.01856172e-01 -1.09617554e-01 -8.12332511e-01 -9.37223315e-01
2.44025171e-01 1.67680338e-01 1.79263707e-02 -3.46070863e-02
3.40879261e-02 3.07224095e-01 -1.80126333e+00 9.93899629e-02
4.88528430e-01 8.31000924e-01 1.89834267e-01 9.66070056e-01
3.48363668e-01 8.34698081e-01 -6.16927326e-01 -2.49174133e-01
1.86536133e-01 -5.24002671e-01 -3.96920919e-01 4.36604731e-02
5.63028753e-01 -5.88943779e-01 -5.68566561e-01 8.01900744e-01
1.32476604e+00 -5.19991741e-02 3.36432010e-01 -1.55854774e+00
-2.76421636e-01 3.80720913e-01 -6.49695575e-01 -1.30524427e-01
6.25984609e-01 6.07554257e-01 5.52050292e-01 -1.07428360e+00
5.59336483e-01 1.42832863e+00 5.92031300e-01 1.08310640e+00
-1.22105014e+00 -1.05493164e+00 -1.02851987e-01 2.44142890e-01
-1.64760673e+00 -1.57746315e+00 9.65966225e-01 -3.14211339e-01
5.70108712e-01 2.60282755e-01 7.24793911e-01 1.44598508e+00
-3.33633721e-02 1.05612457e+00 6.18875325e-01 -2.66919613e-01
-1.06121041e-03 2.62266472e-02 -3.55269939e-01 5.26269138e-01
-2.62693167e-01 8.93209428e-02 -9.17335808e-01 -2.28860937e-02
5.39382398e-01 -5.15960932e-01 -9.25166249e-01 -2.89650559e-01
-1.02567220e+00 6.54179811e-01 6.81253746e-02 1.33602405e-02
-1.72157139e-01 1.82411328e-01 3.27733904e-01 -1.03923097e-01
3.20863605e-01 -5.13241217e-02 -4.86471169e-02 -1.64991707e-01
-1.27437425e+00 4.05116469e-01 9.15606081e-01 9.80551898e-01
4.90356952e-01 1.94273964e-01 -2.58264452e-01 7.64771461e-01
5.09704709e-01 5.87415993e-01 6.93411410e-01 -1.17019951e+00
2.87094563e-01 -7.42010958e-03 -2.67276317e-02 -7.23909616e-01
-4.74853545e-01 -5.33560067e-02 -7.37655580e-01 3.08282435e-01
4.04325455e-01 -5.35173059e-01 -6.94692969e-01 2.21614337e+00
7.56290793e-01 4.91090596e-01 -1.63291693e-01 1.16212881e+00
7.42410123e-01 6.15661323e-01 -2.71488577e-01 -4.11482036e-01
1.40540087e+00 -1.01069355e+00 -1.12747955e+00 -2.40243509e-01
4.91147965e-01 -6.95804000e-01 9.36545312e-01 4.19055313e-01
-1.53420234e+00 -5.02546370e-01 -9.37405705e-01 -3.58148128e-01
1.74482822e-01 3.58752787e-01 2.20535249e-01 4.92759675e-01
-1.36529112e+00 1.83853030e-01 -8.95030677e-01 3.84980813e-02
1.92384914e-01 5.51579118e-01 -5.06969512e-01 2.91928440e-01
-1.15330112e+00 7.33428717e-01 -2.13034153e-01 3.03386062e-01
-7.75507450e-01 -6.60555959e-01 -1.26929021e+00 -1.47418723e-01
5.68217859e-02 -9.74518597e-01 1.54661703e+00 -9.46651518e-01
-2.24365878e+00 9.04005766e-01 -6.58216596e-01 -3.35029721e-01
7.96903312e-01 -4.20570940e-01 -2.33619481e-01 2.79844046e-01
-1.96541790e-02 1.11906791e+00 1.42062998e+00 -1.04495335e+00
-2.98447073e-01 -3.64559710e-01 -4.34510171e-01 2.64428288e-01
-1.97693393e-01 1.61766335e-01 -6.90252900e-01 -5.85121214e-01
-2.68971771e-01 -1.17998075e+00 4.39520538e-01 1.72178656e-01
-6.44718885e-01 -5.58020957e-02 1.06789649e+00 -8.63227963e-01
1.36657166e+00 -2.29607368e+00 3.73174757e-01 -2.59640157e-01
3.57768744e-01 1.95587277e-01 3.57030667e-02 -8.11678246e-02
-1.86884522e-01 -2.91776508e-01 -8.23175982e-02 -1.09833872e+00
4.25326563e-02 -2.54560977e-01 -1.43036529e-01 9.18312907e-01
6.43325821e-02 8.78979087e-01 -5.40694654e-01 -5.49877942e-01
1.00561574e-01 1.01829410e+00 -8.34114850e-01 3.37154448e-01
6.13151416e-02 6.09802783e-01 1.83483765e-01 4.22401309e-01
9.42976892e-01 1.65359437e-01 2.50649936e-02 -2.97568589e-01
-1.42788455e-01 4.78707761e-01 -1.18552816e+00 1.82232881e+00
-5.38319767e-01 1.00048518e+00 6.89391375e-01 -4.91791189e-01
6.19350076e-01 7.09257841e-01 2.38895148e-01 -4.20318902e-01
4.66484398e-01 8.71426985e-02 -8.74116570e-02 -7.61661232e-01
1.38254359e-01 -2.82077510e-02 1.85461611e-01 3.77233654e-01
1.31912395e-01 -2.55823821e-01 -3.22465181e-01 -5.08876517e-02
7.40431607e-01 2.47336119e-01 1.00141592e-01 -7.39292949e-02
7.39477456e-01 -7.15591848e-01 5.54962993e-01 3.78484055e-02
-3.68677408e-01 1.03398967e+00 4.61914897e-01 1.30644128e-01
-7.82667339e-01 -8.43549669e-01 1.10573895e-01 1.03379607e+00
-1.21818185e-01 -3.40401947e-01 -1.40466309e+00 -2.47865409e-01
-1.40693396e-01 6.75838530e-01 -6.59146845e-01 -2.08124444e-01
-7.79205680e-01 -4.80220973e-01 8.77522290e-01 2.46517777e-01
4.43730086e-01 -9.38819289e-01 -5.24400532e-01 -4.80494872e-02
-5.06700516e-01 -1.19871306e+00 -1.21617377e+00 -2.60440767e-01
-1.08897254e-01 -6.42752469e-01 -1.02110457e+00 -7.70781875e-01
4.97004598e-01 2.22192705e-01 5.90034902e-01 -3.38677049e-01
-1.95163235e-01 -7.77651593e-02 2.08274335e-01 -3.30402076e-01
-3.52970779e-01 1.17434211e-01 3.39753330e-01 5.49146771e-01
1.30635202e-01 -8.54064465e-01 -9.32238519e-01 3.45348507e-01
-5.46449184e-01 3.55812967e-01 1.97022915e-01 6.78174853e-01
-4.52900901e-02 -5.51986516e-01 4.17046994e-01 -1.50283158e-01
5.55579245e-01 -4.18655336e-01 -6.35569513e-01 -3.41770113e-01
-5.12706377e-02 2.67243031e-02 3.87382239e-01 -8.17823231e-01
-1.03748858e+00 3.78399551e-01 -3.62480611e-01 -7.59552360e-01
-4.15370315e-02 -3.80467355e-01 -7.74820268e-01 1.78097710e-01
4.92234826e-01 2.06939623e-01 5.06182730e-01 -2.85808474e-01
4.28935856e-01 1.03410673e+00 9.24626172e-01 -1.76496003e-02
5.55714190e-01 7.61633456e-01 -3.81220132e-01 -1.10175729e+00
-3.37478578e-01 -3.60539407e-01 -5.11659980e-01 -2.87706554e-01
1.05424547e+00 -1.11627579e+00 -1.25736296e+00 9.50259209e-01
-1.45121682e+00 -5.97386420e-01 8.84143487e-02 4.34085697e-01
-6.73767447e-01 1.35069817e-01 -5.40727794e-01 -7.90647984e-01
-4.89767283e-01 -1.40037024e+00 1.38936675e+00 1.52627498e-01
-5.93227863e-01 -6.27566755e-01 8.83893147e-02 4.22986001e-01
2.87832111e-01 2.45788489e-02 1.10018127e-01 -2.20059589e-01
-4.53311235e-01 -1.13644242e-01 2.50442386e-01 2.21604615e-01
1.23057671e-01 1.55889720e-01 -1.62638378e+00 -2.66018778e-01
2.37896591e-01 -1.31632671e-01 5.80410361e-01 7.41759360e-01
7.92346060e-01 -7.26614177e-01 -2.48882562e-01 1.00176835e+00
5.34232438e-01 -4.85734716e-02 7.29494214e-01 2.00997815e-02
7.58446395e-01 9.26315129e-01 -2.58068834e-02 5.59729695e-01
7.54364967e-01 1.10031307e+00 3.13522041e-01 -2.15026326e-02
-5.35227239e-01 -2.72282600e-01 9.55048919e-01 9.08787847e-01
2.56509453e-01 -1.50941342e-01 -6.12785280e-01 5.10796130e-01
-1.54787588e+00 -1.05255806e+00 -4.75821039e-03 2.17520833e+00
1.04702866e+00 -5.71013652e-02 4.61018592e-01 2.03131735e-01
9.17152703e-01 2.09648430e-01 -7.22959399e-01 -1.36758566e-01
-7.22573400e-02 -2.50320509e-02 2.14571372e-01 9.76244867e-01
-8.34293425e-01 1.02052259e+00 5.17758560e+00 6.22166157e-01
-1.67887557e+00 2.47489184e-01 2.83391863e-01 -6.59444392e-01
-5.45868911e-02 -3.13702494e-01 -1.04453778e+00 6.49098337e-01
9.69558537e-01 -1.90383736e-02 3.90083611e-01 6.60079658e-01
7.84670234e-01 6.71413988e-02 -1.22465539e+00 1.22182620e+00
3.74586999e-01 -1.07866454e+00 -4.26002145e-01 2.53754914e-01
1.69014186e-01 -6.72288835e-02 1.86004460e-01 -2.36142911e-02
-1.00412585e-01 -9.89488721e-01 1.14734769e+00 4.17160749e-01
9.59204912e-01 -5.05425453e-01 2.12492272e-01 4.19905186e-01
-1.12311995e+00 9.58991274e-02 1.64650843e-01 1.16700657e-01
4.71564382e-01 2.54661947e-01 -1.02227056e+00 -6.71520755e-02
5.65333247e-01 3.71980071e-01 -1.48863927e-01 7.91582406e-01
-4.82046157e-01 4.47840244e-01 -3.57798278e-01 3.57056320e-01
-2.98643619e-01 2.57085055e-01 8.85201991e-01 1.24005997e+00
1.73537299e-01 -2.11756885e-01 -5.19178689e-01 8.73574734e-01
-1.83311045e-01 -1.48352563e-01 -6.97411537e-01 4.89481777e-01
5.07584810e-01 1.21262729e+00 2.18203589e-02 -1.27771199e-01
-2.08623156e-01 1.19428778e+00 4.63778973e-02 3.79433632e-01
-1.09740329e+00 -4.39718753e-01 1.16088665e+00 4.26301122e-01
1.03283495e-01 -1.81056112e-01 9.13479105e-02 -1.33203137e+00
1.32632121e-01 -8.98403227e-01 -2.50304699e-01 -1.07925010e+00
-5.09142339e-01 6.83138013e-01 -1.53073713e-01 -1.21097124e+00
-7.60501146e-01 -2.47542500e-01 -7.37416923e-01 9.14928496e-01
-1.47984672e+00 -1.12270570e+00 -4.54238325e-01 9.24994290e-01
6.22099876e-01 1.74402148e-01 5.00004590e-01 4.62119013e-01
-9.23016131e-01 1.09733176e+00 -4.68854249e-01 1.90956414e-01
1.01373386e+00 -8.42355549e-01 6.55924797e-01 6.93598270e-01
-3.45056444e-01 4.95567888e-01 8.57903004e-01 -2.23404586e-01
-1.47100663e+00 -8.74117494e-01 8.57995927e-01 -3.15686584e-01
4.33099061e-01 -1.00392473e+00 -8.75010371e-01 6.26457810e-01
5.02177536e-01 7.63391480e-02 4.84473377e-01 -7.63241231e-01
-1.13210358e-01 -2.17549965e-01 -9.35279667e-01 7.83514142e-01
1.03804290e+00 -6.39505565e-01 -4.59279150e-01 -1.66698713e-02
7.98164070e-01 -5.11265874e-01 -4.05783385e-01 1.59881890e-01
9.40442979e-01 -1.19237113e+00 8.00304115e-01 -5.55749461e-02
1.05778545e-01 -2.91581631e-01 1.68109328e-01 -1.26763797e+00
2.29511768e-01 -1.62163711e+00 -4.16615456e-01 1.46184301e+00
2.31341675e-01 -7.31947899e-01 7.81899095e-01 5.30722916e-01
-9.54415128e-02 -5.48489571e-01 -9.87207651e-01 -3.84053439e-01
-2.75323950e-02 -4.95935827e-01 5.64859450e-01 7.07591295e-01
1.32742554e-01 4.49731082e-01 -5.80399811e-01 1.87259182e-01
4.88543510e-01 -4.39026713e-01 1.27605438e+00 -8.12466860e-01
-1.42713830e-01 -4.34560478e-01 -2.31632739e-01 -1.45543623e+00
5.68718731e-01 -4.58573192e-01 1.76522195e-01 -9.04609442e-01
-3.19285810e-01 1.93535686e-01 4.13162440e-01 3.04657400e-01
-2.55559888e-02 1.32288203e-01 2.93628901e-01 1.67121321e-01
3.78743634e-02 7.61844993e-01 1.23421156e+00 5.49948215e-02
-5.21550298e-01 1.88161537e-01 -6.34707212e-01 9.83637214e-01
6.74425364e-01 -2.40797684e-01 -4.85802203e-01 -4.97241884e-01
-1.53633535e-01 3.96230012e-01 5.32148361e-01 -9.28345740e-01
5.68647087e-01 3.64025384e-01 1.72817305e-01 -2.25403875e-01
8.48700643e-01 -4.17597890e-01 6.25025108e-02 2.81124085e-01
-1.71368420e-01 1.59383621e-02 3.10313493e-01 2.40263287e-02
-1.15493201e-01 2.59871870e-01 1.08287549e+00 2.14892551e-01
-1.36178061e-01 2.62109518e-01 -4.57432717e-01 1.27934784e-01
7.82541513e-01 -2.65523523e-01 -1.46290496e-01 -9.16746616e-01
-7.52911568e-01 1.61498427e-01 3.91618431e-01 4.92662907e-01
6.26884818e-01 -1.24807560e+00 -6.77815974e-01 6.83143795e-01
-2.78372496e-01 1.64223969e-01 1.92960188e-01 1.18322647e+00
-3.08311015e-01 2.30467021e-01 7.78533518e-02 -8.36141765e-01
-1.33076513e+00 3.57582718e-01 6.43317401e-01 4.60035145e-01
-5.12108564e-01 9.57642674e-01 5.67394435e-01 -1.87335864e-01
4.80974674e-01 -2.63931274e-01 -1.26629144e-01 3.41946423e-01
6.95535719e-01 3.19656909e-01 -1.74605809e-02 -1.07610226e+00
-2.94835478e-01 6.70547068e-01 2.22676292e-01 -5.49360693e-01
1.04485166e+00 -5.37349582e-01 1.35883734e-01 2.82443672e-01
1.64654469e+00 3.96828741e-01 -1.58851039e+00 2.34194040e-01
-5.18403232e-01 -1.98547453e-01 7.04724863e-02 -2.48867288e-01
-1.13499510e+00 1.16570711e+00 4.79867816e-01 -1.31241113e-01
1.21962011e+00 -7.39596691e-03 9.40266907e-01 -2.46003807e-01
-4.79022153e-02 -8.74457359e-01 3.11261625e-03 2.29779050e-01
1.21745563e+00 -1.01077724e+00 -5.09056330e-01 -3.79309416e-01
-7.88847804e-01 8.85126710e-01 6.30180001e-01 2.00180858e-01
7.64875054e-01 6.21439636e-01 1.53518721e-01 2.31224805e-01
-8.30343783e-01 -7.36567527e-02 1.85963348e-01 5.83797276e-01
4.15741742e-01 -1.84088349e-01 1.59958377e-01 5.80751777e-01
-7.04144657e-01 1.54753432e-01 3.57906848e-01 6.09532833e-01
-1.13980994e-01 -6.64485097e-01 -5.97933650e-01 -3.94983053e-01
-4.91600633e-01 -1.34367868e-01 -2.63386130e-01 6.81345820e-01
1.19137347e-01 9.64811921e-01 1.17415369e-01 -3.65539134e-01
2.68932492e-01 2.94739425e-01 3.20282400e-01 -2.42701098e-01
-2.97139555e-01 5.42589366e-01 -5.83116636e-02 -7.45890677e-01
-1.72455370e-01 -7.26564646e-01 -1.26180756e+00 -4.45097446e-01
-2.97907770e-01 -8.07344243e-02 7.66823232e-01 7.07198083e-01
6.81137323e-01 3.11127275e-01 8.59240949e-01 -1.62282908e+00
-3.94725591e-01 -1.13590348e+00 -2.66581386e-01 1.85371459e-01
1.02043402e+00 -7.04566598e-01 -8.37792575e-01 2.75766581e-01] | [13.19989013671875, -0.4321385622024536] |
c110a81e-1661-4369-ad80-a813b0b70a0c | a-unified-survey-on-anomaly-novelty-open-set | 2110.14051 | null | https://arxiv.org/abs/2110.14051v5 | https://arxiv.org/pdf/2110.14051v5.pdf | A Unified Survey on Anomaly, Novelty, Open-Set, and Out-of-Distribution Detection: Solutions and Future Challenges | Machine learning models often encounter samples that are diverged from the training distribution. Failure to recognize an out-of-distribution (OOD) sample, and consequently assign that sample to an in-class label significantly compromises the reliability of a model. The problem has gained significant attention due to its importance for safety deploying models in open-world settings. Detecting OOD samples is challenging due to the intractability of modeling all possible unknown distributions. To date, several research domains tackle the problem of detecting unfamiliar samples, including anomaly detection, novelty detection, one-class learning, open set recognition, and out-of-distribution detection. Despite having similar and shared concepts, out-of-distribution, open-set, and anomaly detection have been investigated independently. Accordingly, these research avenues have not cross-pollinated, creating research barriers. While some surveys intend to provide an overview of these approaches, they seem to only focus on a specific domain without examining the relationship between different domains. This survey aims to provide a cross-domain and comprehensive review of numerous eminent works in respective areas while identifying their commonalities. Researchers can benefit from the overview of research advances in different fields and develop future methodology synergistically. Furthermore, to the best of our knowledge, while there are surveys in anomaly detection or one-class learning, there is no comprehensive or up-to-date survey on out-of-distribution detection, which our survey covers extensively. Finally, having a unified cross-domain perspective, we discuss and shed light on future lines of research, intending to bring these fields closer together. | ['Mohammad Sabokrou', 'Mohammad Hossein Rohban', 'Yixuan Li', 'Dan Hendrycks', 'Hossein Mirzaei', 'Mohammadreza Salehi'] | 2021-10-26 | null | null | null | null | ['open-set-learning'] | ['miscellaneous'] | [ 1.01689599e-01 -7.45884180e-02 -3.99073184e-01 -4.33519810e-01
-5.79093814e-01 -8.02904069e-01 4.55093682e-01 4.88245219e-01
8.97605810e-03 6.15212977e-01 -4.62823451e-01 -4.88807142e-01
-4.51735735e-01 -6.49442613e-01 -4.17064041e-01 -6.90395892e-01
-2.55133808e-01 6.40818477e-01 1.29274428e-01 1.45015180e-01
3.96194667e-01 5.69903076e-01 -2.12478018e+00 3.31924856e-01
8.39956641e-01 9.82049644e-01 -4.09882158e-01 3.86190534e-01
-5.00749409e-01 3.85192931e-01 -9.17329967e-01 -3.55667621e-01
2.09196821e-01 -3.34347486e-01 -6.88237846e-01 2.40756720e-01
6.60509706e-01 -2.04621583e-01 -9.37387496e-02 1.12743509e+00
4.74878937e-01 -6.57310933e-02 1.12595952e+00 -1.86160529e+00
-8.21511924e-01 1.59708217e-01 -6.32589638e-01 6.63238168e-01
3.59250367e-01 -2.72300523e-02 1.06907201e+00 -7.16091692e-01
3.83187920e-01 9.02796686e-01 6.62916362e-01 5.61401606e-01
-1.17236972e+00 -7.40698397e-01 5.66852033e-01 2.88894892e-01
-1.25905776e+00 -1.20094985e-01 7.05533803e-01 -6.27880514e-01
9.63570416e-01 4.25956994e-01 4.75369394e-01 1.49623775e+00
2.06394792e-02 1.03094554e+00 1.04988694e+00 -4.23901290e-01
4.48918402e-01 3.69213820e-01 2.83080429e-01 1.16648741e-01
6.36713862e-01 2.18186513e-01 -3.84010613e-01 -4.92862463e-01
2.70043015e-01 2.44047657e-01 8.66851658e-02 -5.35668433e-01
-5.41090131e-01 8.27827930e-01 -1.54420689e-01 6.18952155e-01
-1.48880601e-01 -6.07376158e-01 5.41773260e-01 6.91270590e-01
5.31602085e-01 4.70501512e-01 -5.38608968e-01 -4.02533174e-01
-9.49164987e-01 3.45454872e-01 1.14040792e+00 9.13461626e-01
6.74305439e-01 1.29936382e-01 1.14720002e-01 9.61450338e-01
2.01972365e-01 4.06061798e-01 6.71841741e-01 -3.49453032e-01
1.02346152e-01 6.10612512e-01 -1.91644400e-01 -9.80979145e-01
-3.86205047e-01 -3.44785511e-01 -7.43321240e-01 2.50756890e-01
7.73172081e-01 -7.36741275e-02 -6.80125058e-01 1.43074632e+00
4.75062788e-01 2.20354155e-01 -4.71386202e-02 5.40790498e-01
5.30401528e-01 2.14289352e-01 -1.01786613e-01 -7.42494687e-02
1.09783626e+00 -6.64569676e-01 -6.65706456e-01 -3.96800667e-01
7.06144869e-01 -8.28405261e-01 8.90560567e-01 7.46224046e-01
-5.44115782e-01 -3.00150126e-01 -9.60525155e-01 4.25392985e-01
-7.87067235e-01 -4.54282165e-01 6.18808448e-01 9.11578834e-01
-5.60947299e-01 5.47418714e-01 -6.53812766e-01 -7.19243765e-01
7.04146504e-01 2.19376869e-02 -3.84813756e-01 -1.05404511e-01
-9.96527433e-01 8.90343428e-01 1.82857424e-01 -2.88677454e-01
-7.16259122e-01 -8.25399399e-01 -7.31928289e-01 -3.50870550e-01
3.97198975e-01 -2.99920052e-01 1.24122941e+00 -9.68046963e-01
-8.65655124e-01 1.07740545e+00 -1.81069925e-01 -3.34345967e-01
4.01073456e-01 -1.63865358e-01 -1.07796073e+00 -2.74340928e-01
1.07792526e-01 3.29976529e-02 8.68605077e-01 -1.13076854e+00
-9.11102653e-01 -5.90487897e-01 -3.11791360e-01 -3.39076109e-02
-5.23351908e-01 1.54922649e-01 1.61225662e-01 -8.17664385e-01
1.14058807e-01 -5.82364380e-01 -3.31263654e-02 -1.01121910e-01
-4.07871693e-01 -5.99778116e-01 1.18372524e+00 -7.63387904e-02
1.75946844e+00 -2.38429117e+00 -5.22680044e-01 2.58887857e-01
3.79132926e-01 2.46246144e-01 1.70032363e-02 7.10877597e-01
-4.33357686e-01 1.12797596e-01 -3.97629857e-01 -7.38934353e-02
1.81627139e-01 4.47680414e-01 -6.05457008e-01 7.01174200e-01
3.76969725e-01 4.21155244e-01 -9.69166875e-01 -8.57121274e-02
2.90061027e-01 7.70081067e-03 -3.33678812e-01 1.07254751e-01
-1.35297790e-01 2.10116833e-01 -3.41723502e-01 1.09806204e+00
8.64986539e-01 2.28090044e-02 -1.17144123e-01 4.06868190e-01
-1.21652428e-02 7.78434202e-02 -1.52136016e+00 9.19933319e-01
-3.82078849e-02 6.87190235e-01 -1.77726105e-01 -1.43609142e+00
1.17185140e+00 2.05915183e-01 5.54914534e-01 -4.51206923e-01
-1.02145478e-01 7.69865036e-01 3.13564420e-01 -5.27835369e-01
4.06951040e-01 -1.66429356e-01 -4.47456352e-02 4.82913136e-01
6.93201199e-02 1.28838927e-01 2.15441093e-01 -2.29185194e-01
1.16396022e+00 -1.96693510e-01 5.88485062e-01 -2.21832953e-02
4.57000673e-01 -1.56921312e-01 1.89453736e-01 1.09287322e+00
-6.60016596e-01 6.16289675e-01 7.52735138e-01 -5.40758908e-01
-8.66302967e-01 -1.21383691e+00 -7.71185517e-01 9.94789541e-01
-2.27085426e-01 -2.71903843e-01 -4.90921438e-01 -1.14109051e+00
4.24100459e-01 8.68354917e-01 -5.73590517e-01 -1.80581346e-01
-1.57448471e-01 -8.45591247e-01 6.81336045e-01 4.28976864e-01
-6.95135742e-02 -9.40008879e-01 -3.25014204e-01 -9.74162761e-03
-5.64195625e-02 -8.15610230e-01 7.42554665e-02 3.94433945e-01
-1.08179283e+00 -1.47792661e+00 -6.17337704e-01 -7.19096065e-01
5.46899974e-01 1.10773042e-01 1.23394740e+00 -4.69508693e-02
-4.03815776e-01 5.78448772e-01 -4.27693665e-01 -8.32982004e-01
-4.03620332e-01 1.04661591e-01 3.86177182e-01 -3.46247964e-02
1.36677802e+00 -5.89001536e-01 -1.53252587e-01 6.88507140e-01
-1.03825510e+00 -1.10495377e+00 4.63871330e-01 6.83916330e-01
4.30084139e-01 3.20071608e-01 9.60505366e-01 -1.00502622e+00
7.41691947e-01 -1.13152468e+00 -2.89662927e-01 -6.67582378e-02
-8.47759783e-01 -3.50226611e-01 5.67158759e-01 -5.85031450e-01
-6.51632309e-01 -4.75599259e-01 -1.59213424e-01 -4.92829889e-01
-1.02315259e+00 1.94738835e-01 -8.69991407e-02 2.01946154e-01
8.99928629e-01 2.07877785e-01 1.64290711e-01 -6.33528590e-01
5.63514270e-02 1.11655009e+00 1.50456980e-01 -5.02787888e-01
8.52642596e-01 4.43792373e-01 -3.59812409e-01 -1.14182973e+00
-9.18679714e-01 -9.17036414e-01 -6.46846414e-01 -1.48754576e-02
2.29267493e-01 -5.95120192e-01 -5.45853414e-02 8.44342887e-01
-5.61341703e-01 -7.01669529e-02 -6.13943279e-01 3.35373074e-01
-3.72489750e-01 6.69006705e-01 -1.02506541e-01 -1.13454580e+00
1.51046604e-01 -8.41648161e-01 7.10855603e-01 2.08440423e-01
-8.28090429e-01 -1.26528549e+00 1.15579069e-01 1.25820637e-01
3.26876134e-01 2.08305389e-01 7.52034485e-01 -1.68643534e+00
-6.93804547e-02 -5.94715953e-01 8.13151821e-02 5.78268409e-01
4.48466539e-01 2.02083394e-01 -1.31798792e+00 -2.84367055e-01
4.79637794e-02 -1.90804005e-01 5.42042911e-01 3.00960958e-01
1.34793746e+00 3.28557775e-03 -5.43402910e-01 2.19663590e-01
9.98508513e-01 2.42994308e-01 5.38650930e-01 6.51620507e-01
2.15786353e-01 7.43148446e-01 8.11106086e-01 6.49397075e-01
1.42700702e-01 3.62613350e-01 4.14717913e-01 9.66913477e-02
1.17245398e-03 -5.14114313e-02 3.14681143e-01 1.33948997e-01
3.19902897e-01 -2.93606162e-01 -9.65477884e-01 7.61901140e-01
-1.49762821e+00 -1.20802224e+00 -4.77858871e-01 2.48840261e+00
5.47901332e-01 2.86591083e-01 5.12408435e-01 4.78873461e-01
8.48974824e-01 -1.15158828e-02 -7.18099594e-01 -5.31722367e-01
-1.91486210e-01 1.52427986e-01 1.15390599e-01 1.35178864e-01
-1.35155821e+00 4.55193013e-01 6.95217276e+00 9.36914980e-01
-9.89652812e-01 -3.00119221e-01 6.51041865e-01 1.00886915e-02
-1.71539798e-01 -1.26639709e-01 -1.03195250e+00 5.66987991e-01
8.11453700e-01 9.98517424e-02 6.57804534e-02 1.23477709e+00
-1.68171644e-01 -3.04321855e-01 -1.45760643e+00 1.07382512e+00
3.40193748e-01 -6.56626046e-01 -1.64122418e-01 2.90323108e-01
6.72540903e-01 -6.85356278e-03 1.94662809e-01 6.11945927e-01
-8.81124586e-02 -9.42354858e-01 3.67084563e-01 3.36396009e-01
4.04277354e-01 -7.88247764e-01 8.85741174e-01 5.05600393e-01
-7.49534786e-01 -2.11882979e-01 -2.94813931e-01 -4.28827345e-01
-1.60428286e-01 9.72148597e-01 -6.75297737e-01 4.78057235e-01
9.78500724e-01 7.41752803e-01 -5.64034998e-01 1.36742318e+00
8.96188095e-02 8.74002397e-01 -4.14791048e-01 7.42265657e-02
1.41937226e-01 -3.34650055e-02 7.40199447e-01 9.90503192e-01
2.56321639e-01 -5.03754020e-01 3.15006167e-01 8.10996234e-01
4.08131272e-01 1.45816999e-02 -9.82885599e-01 -3.06859881e-01
5.03174603e-01 9.45105314e-01 -5.64451277e-01 -1.23507328e-01
-8.89858544e-01 7.35580802e-01 1.33357450e-01 2.05617473e-01
-5.88024437e-01 -6.43950582e-01 1.06171167e+00 2.74916053e-01
1.62649810e-01 1.95303395e-01 -5.80858588e-01 -1.01609302e+00
1.74191728e-01 -1.16200376e+00 9.38048184e-01 6.03352534e-03
-2.35049891e+00 2.38947347e-01 2.20247954e-01 -1.41034448e+00
-2.64834851e-01 -8.22083116e-01 -5.70186675e-01 7.67258883e-01
-1.48330700e+00 -5.00203013e-01 -1.55682608e-01 3.67889911e-01
5.37683785e-01 -4.47150737e-01 1.03879213e+00 3.71741444e-01
-6.91255271e-01 7.72451282e-01 3.56581211e-01 1.36873588e-01
1.07578993e+00 -1.40928185e+00 2.86375463e-01 5.65079331e-01
8.02397728e-02 5.73308527e-01 5.38255930e-01 -6.93976939e-01
-7.81618178e-01 -9.94592607e-01 1.01143181e+00 -9.92229283e-01
8.38218868e-01 -1.33362651e-01 -1.29485309e+00 7.55478978e-01
-4.01645228e-02 2.22138658e-01 1.31008351e+00 5.19412160e-01
-4.08103079e-01 9.91161261e-03 -1.37266183e+00 3.20019275e-01
8.23615551e-01 -2.62500018e-01 -8.37429583e-01 1.45089045e-01
-1.79689005e-01 -3.96676302e-01 -7.32329845e-01 2.89810598e-01
4.79756892e-01 -1.18288171e+00 8.82667661e-01 -8.02241087e-01
1.87084302e-01 -1.09470412e-01 -2.22306475e-01 -1.15794635e+00
-1.88929662e-01 -3.77402574e-01 -4.97355551e-01 1.48094642e+00
2.90568024e-01 -1.11797178e+00 7.48003483e-01 4.82526243e-01
-1.85766354e-01 -9.50832129e-01 -8.19997489e-01 -1.16083241e+00
3.50484759e-01 -7.48227537e-01 5.25604367e-01 1.15921640e+00
1.73943534e-01 4.89789359e-02 -1.44610211e-01 2.59034902e-01
5.12363017e-01 4.29656729e-02 9.99599218e-01 -1.67587423e+00
6.50818646e-03 -7.21838653e-01 -6.76612616e-01 -9.40147579e-01
1.32456824e-01 -9.19844091e-01 -2.98013985e-01 -1.13281369e+00
7.81254098e-02 -4.92984265e-01 -3.94493550e-01 3.57663393e-01
-5.39556704e-02 2.24849969e-01 -2.89178163e-01 1.90332115e-01
-5.37774980e-01 1.94491312e-01 8.63374591e-01 6.88673556e-02
-1.83308840e-01 4.63105947e-01 -9.99783814e-01 1.03189826e+00
9.64996338e-01 -5.94914556e-01 -4.95316535e-01 1.00351952e-01
-3.23539488e-02 -6.17552698e-01 3.49459946e-01 -1.05682373e+00
-1.31160259e-01 -2.68003315e-01 5.41008174e-01 -8.08399677e-01
-7.70994648e-02 -9.54431653e-01 -3.02420557e-01 2.84417152e-01
5.12668863e-02 1.76523209e-01 4.63725567e-01 8.01146865e-01
-3.28379720e-01 -6.39541209e-01 7.11583734e-01 -8.92445967e-02
-8.31097186e-01 2.73040414e-01 -7.78880954e-01 5.81682503e-01
1.48916423e+00 -6.38330996e-01 -8.93153846e-02 -2.86190122e-01
-7.13145912e-01 1.84287578e-01 4.26878661e-01 8.15465868e-01
3.43930870e-01 -1.11584973e+00 -3.67696851e-01 6.16396904e-01
6.22594297e-01 5.46865091e-02 4.62761372e-01 7.74016023e-01
-6.44080639e-02 2.81927466e-01 5.00084506e-03 -8.34026694e-01
-1.08233380e+00 6.59986258e-01 3.67932200e-01 -2.36105904e-01
-4.30876642e-01 5.98132133e-01 1.38053000e-01 -8.25748026e-01
5.81923962e-01 -1.60668969e-01 -1.17763840e-01 2.92116225e-01
6.53822124e-01 7.01170683e-01 2.69046485e-01 -2.46574610e-01
-5.74035525e-01 1.82279527e-01 -3.13882500e-01 4.67133641e-01
1.04981637e+00 -4.31312695e-02 -8.45043212e-02 1.00867629e+00
9.17391300e-01 3.27868871e-02 -7.65090525e-01 -1.85386986e-01
1.79075643e-01 -7.33310401e-01 -4.34545100e-01 -8.53980839e-01
-5.72073162e-01 7.79324174e-01 6.17584407e-01 6.42078340e-01
9.93014097e-01 1.94524020e-01 4.44865048e-01 2.41257563e-01
6.37126043e-02 -1.17332554e+00 3.98234129e-01 7.22211778e-01
6.59601152e-01 -1.40124035e+00 -1.87947199e-01 -2.66182244e-01
-5.10353327e-01 1.19984424e+00 1.10883546e+00 1.52181266e-02
9.11998570e-01 2.74683535e-01 5.51188327e-02 -1.14415713e-01
-4.17592347e-01 -1.37181446e-01 2.43874714e-01 1.13455224e+00
4.98328537e-01 -5.41532189e-02 1.93100851e-02 5.89406013e-01
-2.50493050e-01 -1.57504067e-01 4.26921278e-01 9.02978659e-01
-4.10682410e-01 -1.35296071e+00 -6.86228931e-01 9.83995199e-01
-4.51045930e-01 1.20917350e-01 -5.14022589e-01 9.46202517e-01
2.49596730e-01 9.50888216e-01 4.68131751e-01 -1.25556216e-01
5.55857658e-01 5.35119116e-01 3.10906351e-01 -6.73848629e-01
-3.06261569e-01 -2.79301226e-01 1.18823070e-02 -2.03093633e-01
-1.18753891e-02 -1.04038334e+00 -7.72540510e-01 -4.19009805e-01
-3.71950477e-01 1.56165538e-02 2.79654145e-01 9.08600330e-01
3.40230852e-01 3.35116208e-01 4.32382226e-01 -3.20601195e-01
-8.93541157e-01 -9.87759709e-01 -1.02564895e+00 5.64700723e-01
5.08472323e-01 -9.79227841e-01 -7.34429538e-01 -4.83614385e-01] | [7.718326091766357, 2.583534002304077] |
1e4bfd28-af0d-4ad6-9c68-c1e744c4cb02 | rstgen-imbuing-fine-grained-interpretable | 2205.12590 | null | https://arxiv.org/abs/2205.12590v1 | https://arxiv.org/pdf/2205.12590v1.pdf | RSTGen: Imbuing Fine-Grained Interpretable Control into Long-FormText Generators | In this paper, we study the task of improving the cohesion and coherence of long-form text generated by language models. To this end, we propose RSTGen, a framework that utilises Rhetorical Structure Theory (RST), a classical language theory, to control the discourse structure, semantics and topics of generated text. Firstly, we demonstrate our model's ability to control structural discourse and semantic features of generated text in open generation evaluation. Then we experiment on the two challenging long-form text tasks of argument generation and story generation. Evaluation using automated metrics and a metric with high correlation to human evaluation, shows that our model performs competitively against existing models, while offering significantly more controls over generated text than alternative methods. | ['Yulan He', 'Ritabrata Dutta', 'Rilwan A. Adewoyin'] | 2022-05-25 | null | https://aclanthology.org/2022.naacl-main.133 | https://aclanthology.org/2022.naacl-main.133.pdf | naacl-2022-7 | ['story-generation'] | ['natural-language-processing'] | [ 4.61014628e-01 1.28591311e+00 -1.59483507e-01 9.48563814e-02
-7.97735929e-01 -7.32018769e-01 1.59733796e+00 3.93027574e-01
2.10931078e-02 1.14783549e+00 1.32417929e+00 -4.49860722e-01
-7.64827384e-03 -8.17096710e-01 -3.88284773e-01 2.22558845e-02
8.35950300e-02 8.68231654e-01 2.07187623e-01 -8.88585031e-01
7.07807541e-01 -2.78571546e-01 -1.55195427e+00 1.02997553e+00
1.10447061e+00 3.34699541e-01 8.62098709e-02 8.58512998e-01
-4.18584883e-01 1.67808068e+00 -1.17350841e+00 -4.10880774e-01
-4.91478235e-01 -8.18829656e-01 -1.60476983e+00 -9.92893353e-02
1.46664873e-01 1.47237390e-01 7.51991048e-02 4.52337295e-01
5.26822865e-01 1.69994403e-03 8.78628671e-01 -8.61259222e-01
-7.02045202e-01 1.53096223e+00 1.64133999e-02 2.12203354e-01
9.92798150e-01 -2.12066714e-03 1.27806866e+00 -6.97297454e-01
1.46853077e+00 1.78661692e+00 5.08944988e-01 6.06445551e-01
-1.53152406e+00 -1.82252787e-02 2.16247831e-02 -8.07607174e-02
-8.01104426e-01 -5.89472651e-01 6.83325171e-01 -6.83184385e-01
1.20006895e+00 4.79898840e-01 6.45621121e-01 1.28562212e+00
1.22831240e-01 7.92481422e-01 9.97019529e-01 -9.34815705e-01
7.34661147e-02 8.22068453e-02 1.13794014e-01 3.11330259e-01
1.26404809e-02 3.04706190e-02 -6.97592795e-01 -3.46159071e-01
2.15098798e-01 -1.21875215e+00 -2.76446879e-01 5.79623319e-02
-1.54186940e+00 1.28558052e+00 1.52854532e-01 6.01871431e-01
-2.22014382e-01 8.74357596e-02 5.98446012e-01 1.23867646e-01
7.90903986e-01 1.24030519e+00 -4.92337160e-02 -3.12067628e-01
-7.94243932e-01 1.18946445e+00 1.32075715e+00 8.38238120e-01
6.74426481e-02 -1.35761201e-01 -1.07065094e+00 8.25872540e-01
2.87771761e-01 3.47938538e-01 6.83452845e-01 -1.05003750e+00
7.40087748e-01 6.98447168e-01 2.73197532e-01 -1.15133977e+00
-2.78198302e-01 -1.15059033e-01 -3.13144416e-01 -9.50763151e-02
2.86721259e-01 -3.90518039e-01 -3.17683131e-01 1.64621615e+00
1.22840449e-01 -4.49594200e-01 3.61243784e-01 4.83633637e-01
1.05002546e+00 8.61127853e-01 1.91443503e-01 -4.92052794e-01
1.36074507e+00 -9.73545551e-01 -1.05920434e+00 -2.36557066e-01
9.65989411e-01 -1.13726759e+00 1.17428505e+00 2.02624332e-02
-1.63295650e+00 -3.00595582e-01 -1.02079773e+00 -2.25897431e-01
-1.44722149e-01 9.49433595e-02 2.83230454e-01 2.42332458e-01
-9.47849095e-01 5.57043493e-01 -2.77349710e-01 -1.62195832e-01
2.09328458e-01 -2.04088628e-01 2.13921398e-01 5.07223010e-01
-1.49262226e+00 1.02626979e+00 7.90625691e-01 -6.67596400e-01
-4.04462934e-01 -6.75469577e-01 -8.96607518e-01 -6.29888996e-02
5.53702831e-01 -9.81856465e-01 2.00002718e+00 -6.23755813e-01
-1.63676727e+00 8.90885055e-01 -4.53150943e-02 -7.53731012e-01
5.62249720e-01 -3.99266243e-01 1.03949700e-02 1.20982863e-01
3.43972594e-01 6.95927024e-01 3.97007942e-01 -1.24081564e+00
-4.68551844e-01 1.83850408e-01 2.68485814e-01 5.24317801e-01
-3.49422731e-02 1.80761173e-01 4.19004470e-01 -9.48946178e-01
-4.57590282e-01 -6.15027308e-01 -2.26965129e-01 -7.10951686e-01
-7.73587286e-01 -7.63708413e-01 5.18418431e-01 -5.53467631e-01
1.65171504e+00 -1.56067193e+00 3.47891003e-01 3.50116976e-02
2.80727655e-01 1.42957449e-01 -3.95527370e-02 9.47268963e-01
1.71675310e-01 5.55755854e-01 -7.92938396e-02 -1.38201445e-01
2.71192431e-01 6.08984418e-02 -6.39374614e-01 -4.70752925e-01
3.08289826e-01 1.19328690e+00 -9.44462478e-01 -7.72571087e-01
-3.66545394e-02 1.60443500e-01 -6.48177981e-01 3.41775596e-01
-1.08123147e+00 1.78487629e-01 -5.42487025e-01 -2.04704806e-01
-1.86514854e-01 -4.80385780e-01 4.50765997e-01 2.71855980e-01
-2.68997729e-01 1.04356360e+00 -8.19609761e-01 1.55039930e+00
-3.00508589e-01 7.65897036e-01 -2.69816160e-01 -4.35041457e-01
8.73200059e-01 5.67155004e-01 -1.77983895e-01 -6.19715214e-01
1.93455443e-01 9.63687599e-02 1.39783561e-01 -5.79506040e-01
9.66867208e-01 6.06061518e-03 -2.84216672e-01 1.04964900e+00
-1.54079825e-01 -6.03209376e-01 8.28195035e-01 7.45391190e-01
8.88388216e-01 1.21563077e-01 5.11472702e-01 -5.06475270e-01
5.01557112e-01 4.07282114e-01 -2.62777925e-01 8.29325497e-01
4.91087675e-01 4.48893130e-01 8.47294211e-01 -4.99353819e-02
-1.17968166e+00 -6.89526260e-01 6.34270087e-02 1.27521598e+00
-2.32725903e-01 -1.01474297e+00 -1.20115352e+00 -5.37593067e-01
-2.31931135e-01 1.43790638e+00 -6.94101036e-01 1.97210640e-01
-9.61930811e-01 -4.63621229e-01 7.30045557e-01 2.51445204e-01
8.01326036e-02 -1.51389754e+00 -9.74910378e-01 4.45567548e-01
-8.53426456e-01 -9.75977540e-01 -2.21643656e-01 -3.51612866e-01
-4.89765167e-01 -1.22036457e+00 -3.52166444e-01 -5.88933468e-01
1.61895826e-01 -7.61721879e-02 1.68246329e+00 3.18399757e-01
-3.49884555e-02 2.27393210e-01 -7.30346918e-01 -7.00186610e-01
-1.24487531e+00 5.50893307e-01 -6.55090749e-01 -6.25001371e-01
-1.67946771e-01 -2.76589006e-01 -1.53719589e-01 -1.21712826e-01
-1.11332178e+00 6.87475860e-01 9.31294337e-02 9.04430985e-01
-3.06659136e-02 -5.25829673e-01 7.55432010e-01 -1.16751206e+00
1.60483241e+00 -3.48309934e-01 -8.13210607e-02 3.71648133e-01
-7.13474929e-01 3.00782770e-01 3.57413888e-01 -1.20229878e-01
-1.43994856e+00 -7.33471394e-01 -1.31736466e-04 5.77496588e-01
1.57672495e-01 8.23646247e-01 1.43559560e-01 6.68586075e-01
1.26716757e+00 -2.20067114e-01 1.81602448e-01 -1.99851543e-01
8.39476407e-01 4.57000315e-01 4.67942744e-01 -9.19420421e-01
5.87481558e-01 1.03121549e-01 -3.15368682e-01 -6.15642905e-01
-1.11184537e+00 2.34069359e-02 -3.66517454e-01 -1.92804307e-01
8.22385192e-01 -8.21230352e-01 -1.54798269e-01 -1.45151839e-01
-1.69615519e+00 -6.38574719e-01 -7.15301156e-01 -1.87994968e-02
-8.54837477e-01 1.76761031e-01 -6.63988173e-01 -8.09243202e-01
-6.51934981e-01 -6.43766820e-01 1.18804789e+00 3.40647846e-02
-1.19153094e+00 -1.36986756e+00 4.94908214e-01 4.14941192e-01
4.59825844e-01 7.06852019e-01 1.05941355e+00 -8.31359565e-01
-2.43718341e-01 3.02600861e-01 -2.01136321e-02 -2.42498577e-01
-2.50443459e-01 8.09928030e-02 -7.43415534e-01 1.68033063e-01
-2.22852111e-01 -7.21572042e-01 6.15685642e-01 1.10245876e-01
4.93236691e-01 -8.81171465e-01 -3.65490794e-01 -7.81982318e-02
7.98467755e-01 -1.11692652e-01 7.37359643e-01 6.18882179e-01
3.40825558e-01 1.06075251e+00 8.27334225e-01 6.26143694e-01
5.97638190e-01 7.03550041e-01 4.21202369e-03 7.72537217e-02
-5.09288967e-01 -7.56518543e-01 2.57396698e-01 7.49272287e-01
-3.13632824e-02 -7.06838012e-01 -1.00390840e+00 5.12628436e-01
-2.08957839e+00 -1.33241999e+00 -5.99673986e-01 1.33051610e+00
1.21866763e+00 2.36141145e-01 1.00273684e-01 9.47571844e-02
4.44414735e-01 5.01854479e-01 3.05552155e-01 -6.32209659e-01
-5.07062256e-01 3.00647885e-01 -2.85146773e-01 8.23002875e-01
-7.08084404e-01 1.02272809e+00 7.41264772e+00 9.00757492e-01
-4.05393839e-01 5.01041226e-02 6.05791569e-01 -7.23787174e-02
-8.26602876e-01 1.25066236e-01 -6.43130422e-01 2.58382261e-01
9.58824694e-01 -7.94446647e-01 1.53272659e-01 5.71319163e-01
4.36130971e-01 1.43639848e-03 -1.24694383e+00 2.85481036e-01
3.34681720e-01 -2.06630158e+00 3.88961077e-01 -3.66063453e-02
9.78887498e-01 -4.91632432e-01 -3.60406578e-01 3.80388588e-01
7.99142838e-01 -1.28179574e+00 1.15744150e+00 3.55021685e-01
5.24317503e-01 -4.37394053e-01 7.08291411e-01 5.78594565e-01
-7.11095750e-01 3.05289403e-02 1.34218141e-01 -5.30451655e-01
5.30509531e-01 2.95954078e-01 -1.25057542e+00 4.76661801e-01
9.38482732e-02 4.50691432e-01 -5.84417284e-01 2.40661696e-01
-7.15574682e-01 6.09157383e-01 1.17117114e-01 -3.66075814e-01
2.35568821e-01 3.10527056e-01 8.41787875e-01 1.50541091e+00
1.36814803e-01 3.03110868e-01 4.13468540e-01 1.01588070e+00
-6.63832426e-02 4.45621699e-01 -7.88163900e-01 -1.28435373e-01
6.44351959e-01 8.81233335e-01 -4.22171175e-01 -7.13106573e-01
3.85575742e-01 4.34392124e-01 4.11210209e-01 5.26878312e-02
-5.95815241e-01 -1.10269129e-01 1.70905627e-02 3.86374176e-01
6.79340810e-02 9.93405879e-02 -5.39863467e-01 -8.93380463e-01
-9.02109370e-02 -1.28591001e+00 3.68909180e-01 -9.56251144e-01
-1.13274229e+00 8.16991985e-01 3.85494858e-01 -7.07075655e-01
-1.10836911e+00 -2.09794194e-01 -9.13501263e-01 7.46797025e-01
-1.06896436e+00 -1.19538403e+00 8.42973311e-03 1.00937322e-01
8.48287582e-01 -2.23295599e-01 9.33356404e-01 -5.14628470e-01
-4.92875613e-02 2.08109036e-01 -2.58237958e-01 -1.14446655e-01
5.28057754e-01 -1.41591299e+00 7.33147502e-01 4.74247545e-01
1.33763745e-01 4.89913404e-01 1.09635544e+00 -8.15328121e-01
-6.92129612e-01 -7.91813791e-01 1.60530615e+00 -8.65829229e-01
9.04202104e-01 -2.29152098e-01 -7.29110003e-01 3.79404813e-01
1.01267850e+00 -9.88005936e-01 7.44749904e-01 2.12447241e-01
-2.10196123e-01 8.68166804e-01 -7.99117208e-01 7.30840921e-01
1.16140282e+00 -2.91425198e-01 -1.50866258e+00 6.65988326e-01
1.06492865e+00 -6.59958005e-01 -8.11052024e-01 1.78893447e-01
2.66502053e-01 -9.00887668e-01 8.36818635e-01 -7.00699806e-01
1.27124417e+00 -1.12832054e-01 7.53008500e-02 -1.53413534e+00
-2.32595682e-01 -1.10112250e+00 -3.25196773e-01 1.40246367e+00
7.41367877e-01 -2.52207667e-01 3.20726514e-01 4.21054155e-01
-3.63766044e-01 -5.14031231e-01 -6.61546648e-01 -3.73936445e-01
2.68631428e-01 -1.65915862e-01 6.29926503e-01 8.17859411e-01
6.14446223e-01 1.24106956e+00 -2.32737400e-02 -6.31581664e-01
1.86286747e-01 7.03362525e-02 9.58459735e-01 -1.35180199e+00
-3.16941202e-01 -7.81528234e-01 2.95150578e-01 -8.21348906e-01
2.90653378e-01 -1.02020979e+00 1.46699641e-02 -1.91766274e+00
7.83694014e-02 -1.42625332e-01 6.32227123e-01 1.96562901e-01
-2.85465121e-01 -6.84501976e-02 5.66583991e-01 3.62690330e-01
-6.18583143e-01 6.53853714e-01 1.41580510e+00 -6.68393523e-02
-5.28149247e-01 -3.85472238e-01 -1.06354344e+00 7.01537728e-01
8.18642557e-01 -2.81593442e-01 -6.56518042e-01 -3.99662435e-01
5.07056236e-01 6.57838956e-02 2.97338724e-01 -6.13410950e-01
-1.09589487e-01 -1.63299561e-01 1.00010941e-02 -4.77443546e-01
7.09200650e-03 1.99681059e-01 -1.23554662e-01 3.24690610e-01
-1.24902046e+00 2.57196933e-01 1.45305917e-01 1.66034117e-01
-3.35198760e-01 -4.32390213e-01 2.92643726e-01 -1.72223806e-01
-5.99444024e-02 -5.35637975e-01 -6.37302995e-01 7.88942873e-01
8.28669727e-01 -7.63685554e-02 -8.31998467e-01 -5.07732391e-01
-5.34025550e-01 2.98174649e-01 3.43315810e-01 6.31240010e-01
3.86487007e-01 -1.31704664e+00 -1.27109039e+00 -2.90202349e-01
1.33305117e-01 1.03914835e-01 -2.72112668e-01 4.57931161e-01
-6.02614284e-01 7.45125949e-01 2.70664036e-01 -4.66583312e-01
-1.23975015e+00 2.18757495e-01 -9.70806777e-02 -1.07369268e+00
-6.43337429e-01 5.45337021e-01 -7.49593228e-03 -3.29440773e-01
-1.76781639e-02 -2.30961889e-01 -7.56128907e-01 2.72362590e-01
7.32390285e-01 3.23384374e-01 -2.74178803e-01 -6.58898711e-01
3.05072278e-01 -4.64200489e-02 -1.22023821e-01 -7.21298456e-01
1.05687952e+00 -1.58308357e-01 -3.23686868e-01 5.26290536e-01
5.57160974e-01 1.68951213e-01 -7.64851391e-01 -6.42774627e-02
5.99877536e-01 -1.29468128e-01 -2.35734031e-01 -1.12490201e+00
1.80821717e-01 5.06855607e-01 -9.16251913e-02 9.03118372e-01
3.58116478e-01 3.64792943e-01 6.57138348e-01 4.30241644e-01
-2.21179239e-02 -1.37520552e+00 4.78502035e-01 1.06413567e+00
1.45507658e+00 -7.13832200e-01 -5.52209318e-02 -5.24446130e-01
-9.27695811e-01 1.13403499e+00 5.63600123e-01 9.37649906e-02
9.31966379e-02 2.19650656e-01 1.65892601e-01 -5.74642360e-01
-1.39653862e+00 5.68360500e-02 3.97852242e-01 6.17663264e-01
1.06152785e+00 9.99973789e-02 -8.16850245e-01 3.57942969e-01
-1.10794985e+00 -1.31671771e-01 6.89362943e-01 7.87128806e-01
-7.24780858e-01 -1.17363977e+00 -5.99983752e-01 2.81618059e-01
-3.67735326e-01 -1.62024453e-01 -1.09000266e+00 8.63489032e-01
3.66903245e-02 1.29997444e+00 -2.95771547e-02 2.10300740e-02
3.71788561e-01 2.30964184e-01 3.69089574e-01 -1.03186095e+00
-1.13746691e+00 1.97301000e-01 1.12687683e+00 -2.33262837e-01
-5.91306627e-01 -7.53508806e-01 -1.28929126e+00 -2.84191012e-01
-2.47269467e-01 6.03239536e-01 2.97993183e-01 1.00157118e+00
3.28292280e-01 8.01092803e-01 4.28949505e-01 -6.64865136e-01
-7.84559488e-01 -1.39689136e+00 1.40105993e-01 7.34230995e-01
-2.74439249e-02 -4.55942035e-01 -1.65236413e-01 3.27735752e-01] | [11.548778533935547, 9.062111854553223] |
307a851a-c5e9-411f-ba11-e961679b4d30 | layout-based-causal-inference-for-object | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhang_Layout-Based_Causal_Inference_for_Object_Navigation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhang_Layout-Based_Causal_Inference_for_Object_Navigation_CVPR_2023_paper.pdf | Layout-Based Causal Inference for Object Navigation | Previous works for ObjectNav task attempt to learn the association (e.g. relation graph) between the visual inputs and the goal during training. Such association contains the prior knowledge of navigating in training environments, which is denoted as the experience. The experience performs a positive effect on helping the agent infer the likely location of the goal when the layout gap between the unseen environments of the test and the prior knowledge obtained in training is minor. However, when the layout gap is significant, the experience exerts a negative effect on navigation. Motivated by keeping the positive effect and removing the negative effect of the experience, we propose the layout-based soft Total Direct Effect (L-sTDE) framework based on the causal inference to adjust the prediction of the navigation policy. In particular, we propose to calculate the layout gap which is defined as the KL divergence between the posterior and the prior distribution of the object layout. Then the sTDE is proposed to appropriately control the effect of the experience based on the layout gap. Experimental results on AI2THOR, RoboTHOR, and Habitat demonstrate the effectiveness of our method. | ['Shuqiang Jiang', 'Xinyao Yu', 'Yubing Bai', 'Weijie Li', 'Xinhang Song', 'Sixian Zhang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['causal-inference', 'causal-inference'] | ['knowledge-base', 'miscellaneous'] | [-1.28994184e-02 2.38233015e-01 1.79122925e-01 -4.10371065e-01
1.74793124e-01 -3.41603577e-01 5.14223754e-01 1.49522960e-01
-6.74043953e-01 6.96599483e-01 2.28409141e-01 -1.44363225e-01
-4.40362543e-01 -8.56169403e-01 -1.23889887e+00 -9.08519685e-01
-1.41625211e-01 1.32452995e-01 4.61206555e-01 -1.55385330e-01
4.90370363e-01 8.03793147e-02 -1.36574626e+00 -3.01182598e-01
1.19227397e+00 7.58237541e-01 9.74547863e-01 2.40570202e-01
2.15648860e-02 7.02907324e-01 -5.12647152e-01 1.96353957e-01
2.38507926e-01 -5.46834111e-01 -3.18452328e-01 2.95983609e-02
-8.72828513e-02 -2.79682487e-01 -2.85286456e-01 1.29937041e+00
4.84389007e-01 4.47440624e-01 6.71739459e-01 -1.35104012e+00
-6.35422945e-01 6.76750004e-01 -6.26610994e-01 2.81651795e-01
1.99745432e-01 3.87048125e-01 6.38313890e-01 -5.75279117e-01
6.85586035e-01 1.31482077e+00 1.78190112e-01 -7.84990657e-03
-1.12211764e+00 -5.57082951e-01 9.77141917e-01 3.36992413e-01
-1.39585674e+00 3.76942940e-02 7.25857437e-01 -5.03710270e-01
5.40328503e-01 -1.95365146e-01 8.81318510e-01 8.00465643e-01
4.17082846e-01 6.41917050e-01 9.27961290e-01 -4.28116351e-01
4.23262179e-01 2.78724045e-01 -2.02568233e-01 8.75818074e-01
1.76264897e-01 2.48034060e-01 -5.29268682e-01 2.26327941e-01
8.22013795e-01 -1.72811210e-01 -4.42857176e-01 -6.85363233e-01
-9.08666551e-01 4.62496847e-01 8.38253796e-01 1.09564386e-01
-5.18301964e-01 1.78013086e-01 3.37108262e-02 1.86446771e-01
5.26775420e-02 5.24537385e-01 -4.16232675e-01 6.43379092e-02
-6.91917771e-03 2.28035837e-01 3.71409625e-01 9.03604031e-01
8.85686576e-01 -1.25805795e-01 -2.78821647e-01 6.45122290e-01
6.30603135e-01 3.81618112e-01 1.23455107e-01 -6.30114794e-01
5.98011613e-01 7.42905617e-01 4.32365209e-01 -1.22819459e+00
-2.80894905e-01 -5.29101670e-01 -2.05853820e-01 3.97646159e-01
5.37183821e-01 -5.36712110e-01 -1.14853954e+00 2.22960329e+00
5.65919638e-01 3.42506796e-01 -6.21011518e-02 1.21868956e+00
4.87082571e-01 5.96095622e-01 3.35506469e-01 -1.48271650e-01
9.24994230e-01 -7.60765910e-01 -8.74929547e-01 -8.06896329e-01
5.98479271e-01 -3.97659034e-01 1.18642974e+00 1.25968352e-01
-6.34384155e-01 -4.31950420e-01 -1.10388243e+00 2.31802478e-01
-2.05323264e-01 2.08386153e-01 4.80587721e-01 4.96497899e-02
-7.10489750e-01 4.06563610e-01 -7.92911887e-01 -5.00921547e-01
1.78870618e-01 1.38317078e-01 -1.43588588e-01 -1.62757054e-01
-1.11997950e+00 1.01179135e+00 7.59300351e-01 3.48252922e-01
-1.04411161e+00 -4.22563970e-01 -6.78231716e-01 1.00026913e-01
9.05616701e-01 -5.65977931e-01 7.80539870e-01 -1.07752204e+00
-1.30002213e+00 1.74012676e-01 2.01459691e-01 -1.94867492e-01
4.60491091e-01 -4.04717177e-01 2.01482445e-01 -2.55251080e-01
3.81867439e-02 9.38975632e-01 6.85874462e-01 -1.55128670e+00
-8.25395346e-01 -4.32281494e-01 2.29948685e-01 8.90324771e-01
-2.19105348e-01 -6.75756872e-01 -9.12120223e-01 -2.38540918e-01
3.00651819e-01 -9.72004652e-01 -4.61753696e-01 5.15246764e-02
-3.05280745e-01 -1.85375974e-01 6.60560131e-01 -5.11197209e-01
1.05730820e+00 -2.39452934e+00 4.90655601e-01 4.65228558e-01
-9.92278531e-02 -2.13629678e-01 -2.35367343e-01 2.79827625e-01
2.95570642e-01 -1.31826207e-01 1.84304814e-03 1.75549999e-01
-1.83431000e-01 3.28493863e-01 -2.29137465e-01 3.04941475e-01
1.10104576e-01 5.83875775e-01 -1.09770834e+00 -3.91286820e-01
1.70828909e-01 2.62099445e-01 -6.97070003e-01 5.63084483e-01
-4.18573260e-01 5.89731395e-01 -9.06598866e-01 -7.24555030e-02
5.86468697e-01 -1.10464226e-02 4.91419107e-01 -1.37346694e-02
-4.44598436e-01 1.53891280e-01 -1.29155803e+00 1.52500927e+00
-2.77732491e-01 2.17623949e-01 -6.80911988e-02 -6.05023742e-01
9.41768765e-01 -2.26032585e-01 9.69300047e-03 -7.34566927e-01
1.17548451e-01 -1.22959599e-01 3.76656204e-01 -7.98865259e-01
2.54592508e-01 1.73565909e-01 2.25926235e-01 -1.07070468e-02
-8.54359567e-02 -3.06781586e-02 -2.41270028e-02 2.78162986e-01
8.76446903e-01 5.19868016e-01 2.92365283e-01 -2.30372190e-01
1.66645318e-01 -1.63635924e-01 7.05851495e-01 9.06084239e-01
-1.35116875e-01 7.53610283e-02 8.47557485e-01 9.99263674e-03
-7.08975852e-01 -9.45284128e-01 1.59754902e-01 1.07443845e+00
8.20532799e-01 -1.84065059e-01 -5.94569445e-01 -6.83771193e-01
-1.62005827e-01 1.04545796e+00 -8.84871483e-01 -6.33881092e-01
-4.46331471e-01 -6.41133189e-01 -1.13460250e-01 7.01174974e-01
6.84003830e-01 -1.12664247e+00 -6.12507999e-01 -5.68172969e-02
-5.15512675e-02 -6.65484369e-01 -3.88026595e-01 4.47815001e-01
-6.63213909e-01 -9.09496546e-01 -1.88406765e-01 -7.88825452e-01
1.16562760e+00 6.26094267e-02 5.33541560e-01 2.89349526e-01
7.10808188e-02 3.25180650e-01 -2.21532837e-01 -3.36434960e-01
2.64632106e-01 -3.48817706e-01 -1.87553510e-01 -2.33336061e-01
1.26799541e-02 -4.85519350e-01 -5.94969273e-01 4.27152425e-01
-5.63834369e-01 2.13798076e-01 6.43572509e-01 8.05040359e-01
5.95768571e-01 4.67604339e-01 4.83563989e-01 -4.98889208e-01
5.10827959e-01 -6.82049453e-01 -7.72027075e-01 2.35815004e-01
-7.10110068e-01 3.44156116e-01 2.53770202e-01 -6.74474895e-01
-1.40000343e+00 -1.93550135e-03 1.85766324e-01 -2.45558277e-01
-7.08225593e-02 8.74160886e-01 -5.30957937e-01 2.86140740e-01
4.37211454e-01 1.48059260e-02 -5.11748493e-01 -2.10402399e-01
2.35650480e-01 -2.64553186e-02 1.96269959e-01 -6.11293733e-01
5.97457767e-01 2.21211269e-01 -9.40565988e-02 -4.46332783e-01
-6.85504973e-01 -3.62359211e-02 -4.56661016e-01 -4.90737706e-01
1.09190059e+00 -6.93456888e-01 -6.12603009e-01 2.90706575e-01
-8.46064389e-01 -6.90326035e-01 2.38098223e-02 7.43974566e-01
-4.25575048e-01 -9.10550822e-03 -1.69987548e-02 -9.74917352e-01
5.01740336e-01 -1.15189159e+00 4.70458150e-01 4.96876895e-01
1.33913919e-01 -9.44986999e-01 -5.99659160e-02 -1.88765511e-01
-4.40955535e-03 9.87665355e-02 1.17000091e+00 -3.56268555e-01
-9.17733848e-01 1.68701857e-01 -3.10924500e-01 -4.94558290e-02
-6.09403057e-03 -2.60486007e-01 -6.34147644e-01 -1.35705441e-01
-1.14257626e-01 -1.06506452e-01 8.47336531e-01 6.73607826e-01
9.55989003e-01 -3.24357241e-01 -5.41535854e-01 3.45471293e-01
1.41671729e+00 7.59343505e-01 7.66548514e-01 5.54469943e-01
7.60191143e-01 6.31454587e-01 1.20278001e+00 4.76353288e-01
2.75695562e-01 6.67373955e-01 7.27147162e-01 1.69088468e-01
8.94451737e-02 -7.17395127e-01 2.21613288e-01 2.01610088e-01
-2.12464929e-02 -4.98640239e-01 -8.05033386e-01 4.52521741e-01
-2.16130090e+00 -5.70336640e-01 -2.61816978e-02 2.45520782e+00
6.61482453e-01 5.44330776e-01 -4.67083395e-01 -4.40901071e-01
6.89243674e-01 1.08559959e-01 -8.03430617e-01 -7.64868855e-02
1.84433550e-01 -6.55744970e-01 4.01384771e-01 8.15405071e-01
-7.13117778e-01 1.09671724e+00 5.48972082e+00 7.34392107e-01
-8.47740769e-01 -3.27249348e-01 2.68561453e-01 6.77909255e-02
-3.89704585e-01 1.84328765e-01 -7.77896702e-01 3.66547257e-01
2.23830998e-01 -3.26987752e-03 5.60586512e-01 8.79867733e-01
3.66083652e-01 -8.00649285e-01 -1.02795124e+00 3.13586920e-01
-1.26936272e-01 -4.92067903e-01 -2.30443850e-01 6.63989484e-02
5.27283430e-01 -2.63521045e-01 1.99102357e-01 5.33278704e-01
4.14612621e-01 -5.50609767e-01 9.41556096e-01 8.29014599e-01
2.09227294e-01 -7.43668318e-01 7.39165366e-01 7.35449851e-01
-9.92767394e-01 -3.03184092e-01 -3.73161316e-01 -1.85171872e-01
6.03948534e-03 3.12254876e-01 -1.16353393e+00 2.06529155e-01
6.96012497e-01 5.58715224e-01 -5.75165391e-01 1.11075342e+00
-8.56933534e-01 3.60722870e-01 -2.30251625e-01 -1.75114363e-01
1.78035155e-01 -5.43683350e-01 7.66460061e-01 5.69193959e-01
3.30335438e-01 1.76453739e-01 5.25665164e-01 9.89831269e-01
2.79584765e-01 -4.64434288e-02 -7.41074562e-01 -6.36649951e-02
6.95802033e-01 8.16935420e-01 -8.49330962e-01 -1.51102409e-01
-2.40642130e-02 7.58875072e-01 6.58719242e-01 7.33963311e-01
-8.13273966e-01 -2.38838717e-01 3.28520209e-01 -1.24990031e-01
5.74633777e-01 -8.81478712e-02 -4.25266236e-01 -5.22475004e-01
8.85145292e-02 -4.50951904e-01 1.31694824e-01 -1.03893256e+00
-8.23813200e-01 3.06799054e-01 1.90085337e-01 -1.03978264e+00
4.17514630e-02 -3.29301298e-01 -7.46533096e-01 9.63340580e-01
-1.07092750e+00 -6.76071227e-01 -2.58113682e-01 1.01230808e-01
2.29978994e-01 1.84636638e-02 4.17972833e-01 -3.43483016e-02
-7.67150581e-01 1.54064000e-01 -8.91937315e-02 -2.78965414e-01
7.04710901e-01 -1.20763373e+00 -1.55028328e-01 8.68474901e-01
-4.51198071e-01 8.13039839e-01 1.19768953e+00 -1.24032319e+00
-1.17307246e+00 -7.64104068e-01 5.19399464e-01 -2.42503330e-01
4.51261342e-01 -2.68300921e-01 -9.61964011e-01 7.79779136e-01
1.02113904e-02 -3.67187351e-01 3.36064488e-01 3.04900467e-01
-2.05512196e-01 -1.36650470e-03 -9.34944928e-01 9.79724765e-01
1.16391540e+00 -2.16705818e-02 -6.43649459e-01 -4.03394876e-03
8.37711751e-01 -4.48126912e-01 -3.86551142e-01 3.57088804e-01
4.59004343e-01 -6.90359294e-01 6.90178275e-01 -4.61535245e-01
4.90799338e-01 -6.03287160e-01 -7.02396929e-02 -1.80416512e+00
-5.91526091e-01 2.02267811e-01 1.29725367e-01 1.19102407e+00
4.73032713e-01 -3.86811197e-01 3.62140059e-01 6.08513653e-01
-2.91852385e-01 -7.47479737e-01 -5.86307585e-01 -3.96661967e-01
-4.21585500e-01 -1.26679927e-01 4.65425998e-01 7.44109690e-01
-5.86939380e-02 4.39143449e-01 -3.69434983e-01 7.85658300e-01
3.99196446e-01 -8.75006169e-02 7.12165773e-01 -9.45508957e-01
-3.14523041e-01 -5.14259972e-02 -1.49014518e-01 -1.26792812e+00
1.09499022e-01 -4.54049766e-01 6.24876678e-01 -1.84081697e+00
2.82432526e-01 -5.49273372e-01 -3.83431554e-01 5.43137610e-01
-5.58435678e-01 -5.84200799e-01 6.55261055e-02 9.13913827e-03
-7.11853802e-01 8.27883184e-01 1.66751754e+00 5.52314483e-02
-7.43604064e-01 -1.83822453e-01 -5.99430144e-01 9.19846475e-01
6.00608349e-01 -5.21360934e-01 -7.56223679e-01 -5.88631213e-01
5.23171186e-01 3.21271271e-02 2.47067124e-01 -7.07911372e-01
2.41922870e-01 -4.63010967e-01 5.61998904e-01 -5.66871285e-01
2.58654475e-01 -1.08686054e+00 -1.02050669e-01 5.41731298e-01
-6.16906404e-01 -2.73916990e-01 2.83926725e-01 1.15240550e+00
1.82649687e-01 -3.23739707e-01 4.55164522e-01 7.62226209e-02
-9.89145756e-01 9.95776355e-02 -3.95384222e-01 5.11164106e-02
8.50230396e-01 -1.05825610e-01 -3.32133114e-01 -3.59864205e-01
-6.82038188e-01 8.35736573e-01 3.48199785e-01 5.86991966e-01
7.21176863e-01 -1.26621008e+00 -2.78174669e-01 1.86703071e-01
1.83375493e-01 -9.56617743e-02 3.83524239e-01 7.56102324e-01
3.89617383e-02 1.90902889e-01 -4.05721188e-01 -4.64194566e-01
-1.10243082e+00 5.56944370e-01 3.30743402e-01 -5.92887290e-02
-5.01028061e-01 1.00042832e+00 8.86213839e-01 -2.46411175e-01
5.02565444e-01 -1.44400135e-01 -6.67679727e-01 -2.00268611e-01
3.92018437e-01 1.68261379e-01 -3.67473006e-01 -8.92157033e-02
-2.55811393e-01 2.03184471e-01 -1.25181571e-01 -3.12620342e-01
1.25463724e+00 -5.00031769e-01 -1.34369163e-02 7.04731762e-01
5.95245540e-01 -1.03616133e-01 -1.98644412e+00 -1.05307393e-01
-1.61368623e-01 -7.60846913e-01 3.68124187e-01 -1.06636095e+00
-8.63004088e-01 5.80943882e-01 8.11672986e-01 -3.92549112e-02
9.86472309e-01 4.71958742e-02 -1.42206009e-02 5.13744712e-01
4.49232280e-01 -1.17805958e+00 3.48070264e-01 5.83311260e-01
1.04814732e+00 -1.13871014e+00 -6.33567572e-02 -5.04409492e-01
-8.45322728e-01 5.54812908e-01 1.27464223e+00 -4.27328274e-02
5.53628922e-01 9.39565450e-02 -1.04680039e-01 -2.72420257e-01
-7.73190081e-01 -2.82295048e-01 2.52282470e-01 6.16671741e-01
1.09717324e-01 -5.24331443e-02 -3.10192704e-01 5.64695060e-01
5.68428263e-02 -1.85534790e-01 1.69503674e-01 1.02902055e+00
-8.16650569e-01 -6.29872978e-01 -3.59301805e-01 1.25125617e-01
2.12739944e-01 5.58701158e-02 -4.37894195e-01 7.49487340e-01
5.28448582e-01 7.15676367e-01 1.98252752e-01 -4.43319201e-01
4.24978882e-01 -3.28235924e-02 4.36132550e-01 -5.52663863e-01
-1.43551141e-01 2.56254554e-01 1.14604518e-01 -4.92821842e-01
-9.84794199e-02 -5.31395078e-01 -1.46673441e+00 2.22103357e-01
-6.29620552e-01 1.39380038e-01 5.07914007e-01 1.08656704e+00
1.30756438e-01 7.97288656e-01 4.41402346e-01 -5.93615830e-01
-2.53522813e-01 -1.01148880e+00 -6.60089850e-01 2.29659811e-01
7.43385926e-02 -1.09440911e+00 -4.90483314e-01 -1.91784099e-01] | [4.497059345245361, 0.5689058899879456] |
70f0bd37-cbd9-4b75-9e70-695dc6b38f88 | sar-image-despeckling-based-on-nonlocal | 1611.07559 | null | http://arxiv.org/abs/1611.07559v1 | http://arxiv.org/pdf/1611.07559v1.pdf | Sar image despeckling based on nonlocal similarity sparse decomposition | This letter presents a method of synthetic aperture radar (SAR) image
despeckling aimed to preserve the detail information while suppressing speckle
noise. This method combines the nonlocal self-similarity partition and a
proposed modified sparse decomposition. The nonlocal partition method groups a
series of structure-similarity data sets. Each data set has a good sparsity for
learning an over-complete dictionary in sparse representation. In the sparse
decomposition, we propose a novel method to identify principal atoms from
over-complete dictionary to form a principal dictionary. Despeckling is
performed on each data set over the principal dictionary with principal atoms.
Experimental results demonstrate that the proposed method can achieve high
performances in terms of both speckle noise reduction and structure details
preservation. | ['Cheng-Wei Sang', 'Quisong Xia', 'Hong Sun'] | 2016-11-22 | null | null | null | null | ['sar-image-despeckling'] | ['computer-vision'] | [ 5.47652125e-01 -6.93964124e-01 2.04190388e-01 -2.39900693e-01
-8.27762008e-01 -2.86234111e-01 3.88741612e-01 -2.94380546e-01
-9.41965953e-02 4.31962490e-01 6.82842553e-01 2.76711702e-01
-5.09957135e-01 -7.62275815e-01 -2.08268300e-01 -1.13712931e+00
7.82159567e-02 2.14976341e-01 -3.88314319e-03 -1.78301066e-01
2.86942899e-01 6.70369148e-01 -1.48079658e+00 4.30352688e-01
8.97407413e-01 8.35680068e-01 5.39202392e-01 2.52159446e-01
4.50117737e-02 7.22024858e-01 -3.52025896e-01 3.62654805e-01
7.43149221e-01 -6.65342510e-01 -1.99152514e-01 5.60344934e-01
5.52223444e-01 3.44248712e-02 -3.31636429e-01 1.51912391e+00
5.57167649e-01 2.37998173e-01 5.65047383e-01 -3.68009657e-01
-4.83518064e-01 3.56203496e-01 -8.99591684e-01 6.17324531e-01
7.81331956e-02 -2.19124585e-01 6.21917427e-01 -1.13310683e+00
5.68923056e-01 1.19998407e+00 6.60088301e-01 -1.32061755e-02
-1.15379870e+00 -4.69667077e-01 -2.70121753e-01 2.25519583e-01
-1.59023499e+00 -7.49053657e-01 1.13167334e+00 -4.33593512e-01
3.87959480e-01 5.86770058e-01 6.37472987e-01 1.90276682e-01
3.49754304e-01 3.50748539e-01 1.55504882e+00 -4.23930466e-01
1.23810083e-01 -3.83567184e-01 5.16041517e-01 5.71403623e-01
6.38163447e-01 2.62464970e-01 -4.55533653e-01 -4.77938533e-01
7.87272453e-01 3.38387638e-01 -4.74800467e-01 -4.23016191e-01
-1.34952259e+00 8.68301511e-01 2.14829609e-01 7.37089574e-01
-6.69698596e-01 -4.05588567e-01 5.49440198e-02 4.89495605e-01
1.48101538e-01 2.23852709e-01 1.09587908e-01 6.17574811e-01
-1.25032449e+00 4.23846617e-02 5.78257859e-01 6.27734005e-01
1.07302761e+00 6.50314987e-01 -1.21138968e-01 8.18375647e-01
3.00064445e-01 1.20538807e+00 6.35708570e-01 -9.01219070e-01
1.01686396e-01 4.01567250e-01 -2.28246916e-02 -1.62253571e+00
-1.93392754e-01 -7.97834635e-01 -1.46955025e+00 1.91608131e-01
-1.30689457e-01 7.12129548e-02 -8.19815397e-01 1.10481584e+00
1.36904106e-01 4.19233173e-01 5.98145306e-01 1.21847510e+00
1.05704904e+00 1.02076209e+00 -3.56714338e-01 -7.93338776e-01
1.32983983e+00 -6.37807310e-01 -8.98652196e-01 -2.34460726e-01
-2.94036996e-02 -1.15904403e+00 2.90093303e-01 4.97147232e-01
-9.44640875e-01 -7.34974802e-01 -1.14452875e+00 4.12396967e-01
5.86236537e-01 3.72499675e-01 3.16553622e-01 3.08555126e-01
-6.35917783e-01 3.10410917e-01 -6.39746070e-01 -1.10009432e-01
1.18495546e-01 8.84944499e-02 -4.21764016e-01 -4.73420024e-01
-8.43426049e-01 6.13668144e-01 2.98949450e-01 1.47721320e-01
-9.10171926e-01 -4.72620487e-01 -8.69999766e-01 6.89050257e-02
-7.08424821e-02 -5.99658012e-01 4.05813962e-01 -8.63515735e-01
-9.32163894e-01 8.18475664e-01 -3.89356464e-01 -3.75317574e-01
-3.09574038e-01 8.30591470e-02 -8.26814532e-01 3.91444802e-01
2.66935825e-01 -2.66672879e-01 1.45838249e+00 -1.48541379e+00
-4.29676533e-01 -5.28341830e-01 -7.44323492e-01 3.73417258e-01
-1.41877279e-01 -1.50228813e-01 5.95940351e-02 -1.07622290e+00
1.01547325e+00 -5.33290565e-01 -5.68894804e-01 -5.95435858e-01
-1.55180395e-02 6.12773478e-01 1.15849066e+00 -7.80542254e-01
1.26532125e+00 -2.29995704e+00 2.71749914e-01 5.65874815e-01
1.78963393e-01 2.85560548e-01 -3.11664581e-01 5.30449092e-01
-3.39245290e-01 -6.05900049e-01 -6.18287504e-01 3.10038507e-01
-7.46592224e-01 8.62163976e-02 -4.71929014e-01 6.47421658e-01
-3.51775289e-01 3.41137618e-01 -7.09182918e-01 -3.15455556e-01
1.71499044e-01 4.92172867e-01 -2.42879763e-01 1.11866660e-01
3.48273963e-01 6.66540146e-01 -6.53446615e-01 8.62913072e-01
1.24239182e+00 1.85406934e-02 2.10146606e-01 -7.98216343e-01
-3.15064758e-01 -4.13433969e-01 -1.62767112e+00 1.46528351e+00
-2.13591143e-01 1.72328442e-01 6.37061715e-01 -1.32683837e+00
1.60941637e+00 9.20963809e-02 7.03366280e-01 -5.67178488e-01
9.10198018e-02 3.30209523e-01 -1.29478589e-01 -5.29994786e-01
2.21052915e-01 -7.01595962e-01 2.61502624e-01 3.15676510e-01
-2.31848344e-01 -1.31853372e-01 2.40445957e-02 -6.12080246e-02
1.01172614e+00 -6.66823626e-01 6.83006525e-01 -8.85988116e-01
1.09817767e+00 3.22396815e-01 8.95576239e-01 5.98368108e-01
1.35057271e-01 5.94071567e-01 -3.83918822e-01 -7.54730582e-01
-1.02225149e+00 -9.21209633e-01 -2.12067708e-01 4.86272812e-01
3.26919943e-01 -1.16007857e-01 -3.37256968e-01 -1.30280048e-01
-1.26868427e-01 2.80525923e-01 -1.53796509e-01 -5.26157022e-02
-8.35899472e-01 -9.53665972e-01 4.87623960e-02 -1.67504519e-01
9.37363923e-01 -6.98354125e-01 -3.23971242e-01 1.58256128e-01
-1.74343482e-01 -7.69522905e-01 -4.89016324e-01 -8.67344514e-02
-1.25470161e+00 -1.03370488e+00 -6.87113822e-01 -1.36351717e+00
9.44549799e-01 1.17292154e+00 7.28895485e-01 -2.73945481e-02
-2.34298065e-01 2.34191492e-01 -5.40846825e-01 8.02422166e-02
-3.84558558e-01 -6.05567694e-01 2.21514672e-01 6.05573535e-01
4.29791868e-01 -9.48178411e-01 -4.36095119e-01 2.46748179e-01
-9.94148016e-01 -2.68156171e-01 1.00920928e+00 1.00217807e+00
1.11349511e+00 7.70520568e-01 1.78532407e-01 -7.40918100e-01
5.10762572e-01 -3.61770779e-01 -6.99288607e-01 4.85961661e-02
-3.43723208e-01 -1.33757656e-02 6.83933914e-01 -3.35214846e-02
-1.18189251e+00 3.85615855e-01 1.87921301e-01 -5.63571632e-01
-5.85295912e-03 8.34299743e-01 -1.31801724e-01 -6.12930715e-01
7.32854843e-01 1.06820846e+00 4.95623112e-01 -8.13813031e-01
-3.26742083e-02 4.92493004e-01 7.60566473e-01 -3.66610557e-01
1.17847013e+00 7.96027601e-01 2.79155016e-01 -1.27931297e+00
-9.32231486e-01 -9.40518498e-01 -5.93913019e-01 6.41607493e-02
4.30958956e-01 -1.28422856e+00 -1.46320745e-01 2.91586787e-01
-7.58099198e-01 5.43696642e-01 -4.81701523e-01 8.18389416e-01
-4.33017373e-01 1.03591096e+00 -2.22321033e-01 -5.68768919e-01
-5.32723069e-01 -8.14019501e-01 6.56280756e-01 4.75631543e-02
3.47279489e-01 -7.01775253e-01 2.57819384e-01 4.84700620e-01
3.93332630e-01 2.10695684e-01 6.56191409e-01 -4.11603034e-01
-5.68876922e-01 -1.40620023e-01 -1.15884982e-01 5.49898326e-01
3.20438355e-01 -8.61226857e-01 -3.52203190e-01 -8.76671135e-01
1.13471901e+00 5.23330532e-02 1.17603827e+00 5.93666434e-01
6.49366379e-01 -4.84111995e-01 -2.91386992e-01 9.16554213e-01
1.88636124e+00 3.77441078e-01 7.17303753e-01 2.88288593e-01
5.58527231e-01 3.46513420e-01 7.34202564e-01 7.14859486e-01
-3.30229402e-01 2.68664002e-01 8.09184015e-02 -8.83559138e-02
-2.48021767e-01 1.58368513e-01 1.90963924e-01 1.33363056e+00
-3.23648714e-02 2.36343175e-01 -8.09780657e-01 7.11668015e-01
-1.63692820e+00 -1.47294331e+00 -4.49299812e-01 1.93364382e+00
5.80921888e-01 -3.17525059e-01 -3.63324076e-01 2.42746800e-01
9.49135900e-01 6.93042219e-01 -2.28772104e-01 1.41123399e-01
-7.44295537e-01 1.87079832e-01 5.41655302e-01 9.23865318e-01
-1.05880594e+00 6.44678116e-01 6.55058718e+00 1.01690972e+00
-9.45597410e-01 3.43267731e-02 3.08067668e-02 4.97653753e-01
-4.90313947e-01 2.23750427e-01 -7.76292324e-01 3.17599654e-01
2.83814907e-01 -2.81712681e-01 3.50368083e-01 6.08148336e-01
3.91244352e-01 -1.95816040e-01 -2.26483181e-01 1.40866649e+00
5.16658545e-01 -1.58004248e+00 7.50130892e-01 -2.03270882e-01
1.05747640e+00 -2.40114853e-01 -4.96200658e-02 -3.76594394e-01
1.05164640e-01 -7.78956175e-01 1.58112288e-01 9.60701406e-01
6.18946016e-01 -6.89535856e-01 6.99391723e-01 5.40584862e-01
-1.29967153e+00 -2.02708900e-01 -8.24767232e-01 -1.59421846e-01
1.50930882e-01 1.29770660e+00 -2.19755545e-01 7.32027173e-01
3.41536582e-01 1.15372026e+00 -1.12456135e-01 1.10756350e+00
1.36619091e-01 5.61181903e-01 -2.07865700e-01 5.22877395e-01
2.56798089e-01 -9.06220734e-01 1.27670336e+00 9.89326239e-01
7.48739123e-01 9.41474676e-01 6.31340325e-01 4.70958710e-01
5.30631185e-01 1.83938801e-01 -6.59808040e-01 1.08989313e-01
5.46327472e-01 1.12392676e+00 -4.92816031e-01 -2.20287263e-01
-3.66813660e-01 6.89070404e-01 -5.45155585e-01 2.38733292e-01
1.36472499e-02 -2.14060172e-01 5.13116419e-01 2.49602050e-01
6.34116292e-01 -4.36039954e-01 -4.47025418e-01 -1.28085852e+00
-1.58010721e-01 -1.30575800e+00 4.64747310e-01 -5.75260401e-01
-1.43537712e+00 7.69802332e-01 -1.01008810e-01 -1.76736259e+00
3.11930209e-01 -2.06640854e-01 -5.46828032e-01 9.87911522e-01
-1.60217237e+00 -1.10568202e+00 -4.58864748e-01 1.10616171e+00
5.14965773e-01 -9.93741870e-01 7.72706687e-01 7.06257075e-02
-5.65155596e-02 -1.51176587e-01 6.90669477e-01 -8.29852670e-02
3.78763199e-01 -5.26087046e-01 -3.40238094e-01 1.13929200e+00
-3.81705910e-02 7.33587503e-01 9.20348287e-01 -1.00891542e+00
-1.45336509e+00 -9.89106715e-01 8.37997019e-01 3.34883809e-01
3.53174448e-01 3.18815947e-01 -7.99110651e-01 4.51330036e-01
1.59176528e-01 3.40313017e-02 8.40489626e-01 -2.92582095e-01
-2.39720270e-01 -5.22841990e-01 -1.21655035e+00 1.89991668e-02
6.06743813e-01 -2.09779084e-01 -1.15653598e+00 4.65245515e-01
1.19015589e-01 -1.52358234e-01 -8.48238647e-01 6.51792407e-01
1.66197211e-01 -9.55044866e-01 1.19737339e+00 -4.55523692e-02
7.73854554e-02 -9.39710736e-01 -6.40677512e-01 -1.18926609e+00
-1.20476866e+00 -7.31879115e-01 1.36740461e-01 8.83171976e-01
-2.13741228e-01 -3.60602647e-01 8.28582764e-01 -4.87922341e-01
-5.28384894e-02 -2.39696026e-01 -8.51996005e-01 -9.54291046e-01
-4.69145417e-01 2.89728642e-01 4.42901969e-01 1.15895641e+00
-4.78480935e-01 4.84741002e-01 -7.96248376e-01 6.24218285e-01
1.27690494e+00 7.98786283e-01 3.76589149e-01 -1.52355492e+00
-1.37803972e-01 1.00626826e-01 -5.10490716e-01 -1.08228660e+00
6.20174557e-02 -9.21366572e-01 -1.53149724e-01 -1.53596377e+00
4.87220138e-01 -4.29205090e-01 -2.33323410e-01 8.16918984e-02
1.28282279e-01 3.66422921e-01 1.34767458e-01 1.07167006e+00
-2.87014276e-01 6.10562682e-01 1.37110984e+00 -3.70732635e-01
-2.13652030e-01 1.95507612e-02 -7.25024045e-01 7.05311239e-01
6.76551282e-01 -5.22208452e-01 -2.80316591e-01 -5.44220209e-01
-2.58202523e-01 4.06698406e-01 1.93681136e-01 -1.46706772e+00
3.83091658e-01 -3.10672164e-01 2.73023725e-01 -8.04782033e-01
1.56245559e-01 -1.11852729e+00 6.23058736e-01 7.16207206e-01
1.11131109e-01 -2.31078893e-01 -9.10433829e-02 7.69082427e-01
-7.75774419e-01 -3.60387236e-01 1.20512390e+00 -3.54137480e-01
-1.04734719e+00 1.33033410e-01 -4.27513480e-01 -1.67318583e-01
8.37949991e-01 -5.26114762e-01 -1.84660964e-02 -2.50522971e-01
-1.14798987e+00 -1.42600030e-01 3.23879898e-01 -1.80490062e-01
1.00583923e+00 -1.56118298e+00 -1.23739696e+00 8.49143445e-01
-1.07196914e-02 -3.93360078e-01 7.47313619e-01 6.62056386e-01
-7.15054810e-01 2.58254737e-01 -7.21106827e-01 -6.13231301e-01
-1.43640661e+00 5.31197786e-01 2.32692137e-01 -3.36464554e-01
-8.39153588e-01 6.33861959e-01 2.23290235e-01 -1.51621206e-02
-3.00849319e-01 2.89474487e-01 -5.31548083e-01 2.62507461e-02
8.59876215e-01 3.84061188e-01 -4.64030690e-02 -1.16270447e+00
-2.56635815e-01 1.25843108e+00 1.05624169e-01 5.60026914e-02
1.64396369e+00 -3.10137033e-01 -8.43795240e-01 -3.76630994e-03
1.10195017e+00 5.35523653e-01 -7.91999280e-01 -6.84035420e-01
-3.55560899e-01 -8.03846538e-01 4.26134735e-01 -5.47282338e-01
-1.19681799e+00 4.05919969e-01 6.72171175e-01 2.15302464e-02
1.63064766e+00 -3.81731182e-01 9.58371401e-01 6.50123358e-01
3.59544396e-01 -6.73493922e-01 -8.91622007e-02 6.63068831e-01
1.12438798e+00 -1.03372645e+00 5.17722487e-01 -6.46218538e-01
-4.66313094e-01 1.07811117e+00 7.96332303e-03 -7.74540722e-01
1.00873709e+00 2.45457008e-01 7.88482726e-02 -5.80390513e-01
-2.39858359e-01 -3.34409654e-01 3.00753862e-01 7.81215370e-01
-7.06151053e-02 -7.75256157e-02 -8.21775973e-01 6.32032931e-01
-7.37386122e-02 -2.20446631e-01 4.18724805e-01 8.32182646e-01
-1.22240591e+00 -9.63329732e-01 -1.11741042e+00 4.54200417e-01
-1.50647819e-01 -1.85514763e-01 9.59210936e-03 2.10253701e-01
1.03322819e-01 9.22287226e-01 -9.34047401e-02 -4.63583678e-01
3.95332485e-01 -3.22742790e-01 2.27733195e-01 -5.95751941e-01
-2.46964991e-01 6.22755766e-01 -1.69767797e-01 -3.03417087e-01
-7.65979052e-01 -7.64019251e-01 -9.11798060e-01 -1.28218085e-01
6.07702211e-02 5.99938393e-01 9.83864218e-02 5.76456070e-01
2.45393783e-01 1.03394523e-01 9.86588061e-01 -6.11028671e-01
-5.83148658e-01 -6.18382454e-01 -1.31791961e+00 3.75189185e-01
4.93583679e-01 -3.50765705e-01 -6.09688878e-01 4.26984519e-01] | [10.442651748657227, -2.004106044769287] |
48040ebf-da3b-411f-b6b2-c5c940512dde | galois-boosting-deep-reinforcement-learning | 2205.13728 | null | https://arxiv.org/abs/2205.13728v1 | https://arxiv.org/pdf/2205.13728v1.pdf | GALOIS: Boosting Deep Reinforcement Learning via Generalizable Logic Synthesis | Despite achieving superior performance in human-level control problems, unlike humans, deep reinforcement learning (DRL) lacks high-order intelligence (e.g., logic deduction and reuse), thus it behaves ineffectively than humans regarding learning and generalization in complex problems. Previous works attempt to directly synthesize a white-box logic program as the DRL policy, manifesting logic-driven behaviors. However, most synthesis methods are built on imperative or declarative programming, and each has a distinct limitation, respectively. The former ignores the cause-effect logic during synthesis, resulting in low generalizability across tasks. The latter is strictly proof-based, thus failing to synthesize programs with complex hierarchical logic. In this paper, we combine the above two paradigms together and propose a novel Generalizable Logic Synthesis (GALOIS) framework to synthesize hierarchical and strict cause-effect logic programs. GALOIS leverages the program sketch and defines a new sketch-based hybrid program language for guiding the synthesis. Based on that, GALOIS proposes a sketch-based program synthesis method to automatically generate white-box programs with generalizable and interpretable cause-effect logic. Extensive evaluations on various decision-making tasks with complex logic demonstrate the superiority of GALOIS over mainstream baselines regarding the asymptotic performance, generalizability, and great knowledge reusability across different environments. | ['Yang Liu', 'Jianye Hao', 'Yi Li', 'Yan Zheng', 'Hao Zhang', 'Tianpei Yang', 'Zhiming Li', 'Yushi Cao'] | 2022-05-27 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [-7.14714304e-02 2.99537599e-01 -5.60268879e-01 -2.88875937e-01
-3.31660032e-01 -7.14806795e-01 5.40850639e-01 3.57301198e-02
1.76209718e-01 7.08486378e-01 2.60366332e-02 -9.15585637e-01
-2.74319470e-01 -1.22525918e+00 -1.10826683e+00 -1.10692978e-01
6.78578541e-02 1.68456078e-01 3.34974885e-01 -4.45187598e-01
5.61021194e-02 1.44097626e-01 -1.48086929e+00 5.27251244e-01
1.44515312e+00 7.39262819e-01 -1.14796728e-01 4.84501123e-01
2.51186341e-02 1.51426923e+00 -2.82781571e-01 -4.55778301e-01
1.65779173e-01 -3.64566952e-01 -7.53283679e-01 -3.82716894e-01
7.19704330e-01 -6.97384953e-01 -3.54054302e-01 1.11068261e+00
4.35702922e-03 -1.18864119e-01 3.12532902e-01 -1.56059694e+00
-1.01607895e+00 1.14159012e+00 3.23914066e-02 -4.51920956e-01
4.63827282e-01 8.20921302e-01 1.55955708e+00 -4.19551790e-01
4.22937840e-01 1.73145902e+00 7.39731312e-01 8.28936517e-01
-1.59743762e+00 -5.58848023e-01 3.94807965e-01 3.22292261e-02
-1.06131470e+00 -2.12630317e-01 6.44953012e-01 -5.42155921e-01
1.15672874e+00 1.45654753e-01 8.36926281e-01 1.06121051e+00
4.46681827e-01 1.10363925e+00 1.44950187e+00 -3.77251685e-01
5.63705206e-01 1.07887395e-01 3.31249952e-01 1.21345520e+00
6.23023629e-01 4.92761642e-01 -5.90352952e-01 -1.15226395e-01
7.08533168e-01 -1.45264655e-01 -6.89506158e-02 -5.63666582e-01
-1.16959345e+00 8.05946350e-01 4.36955512e-01 -9.95773077e-02
-3.81074026e-02 7.30064154e-01 6.31132126e-01 5.69448948e-01
-3.23501348e-01 9.85040009e-01 -3.49430174e-01 1.63935423e-02
-6.18229985e-01 9.24612701e-01 1.14136517e+00 1.11695230e+00
6.95559740e-01 3.82713377e-01 -6.57747269e-01 2.51038045e-01
4.33273762e-01 7.23789334e-01 4.24995124e-02 -1.23734498e+00
3.36840570e-01 1.03457141e+00 -1.09265447e-01 -1.06137538e+00
-2.33632997e-01 -2.66946375e-01 -5.41474164e-01 4.53235269e-01
2.04186618e-01 -2.59801179e-01 -5.18695235e-01 2.00448489e+00
-1.14563882e-01 -3.62606227e-01 2.86954910e-01 6.34372592e-01
7.09772050e-01 6.24766827e-01 1.47487596e-01 1.93717584e-01
1.39098656e+00 -1.02399552e+00 -6.89101279e-01 -5.41812301e-01
5.04816532e-01 2.97373980e-01 1.57074261e+00 6.56443000e-01
-1.11652625e+00 -3.91688228e-01 -1.32756341e+00 -9.64825004e-02
-2.42005423e-01 -6.55523241e-02 1.24601746e+00 6.66355073e-01
-1.20805490e+00 4.16174889e-01 -5.87568402e-01 6.80263266e-02
5.62832177e-01 2.56032109e-01 8.58817622e-02 7.64413625e-02
-1.28394830e+00 7.54339159e-01 7.91782916e-01 -2.11647302e-01
-1.24326968e+00 -1.08553338e+00 -1.13446343e+00 2.67573297e-01
8.02524269e-01 -9.62227345e-01 1.73852170e+00 -1.00270927e+00
-1.79603422e+00 4.15830791e-01 2.55819738e-01 -7.42966950e-01
5.87733030e-01 -1.59600541e-01 -7.12411925e-02 -6.44580126e-02
-1.97831199e-01 8.23866069e-01 8.15724969e-01 -1.27892900e+00
-6.17918611e-01 4.45253663e-02 9.12399530e-01 -7.27705583e-02
-3.88276167e-02 -5.12296855e-01 -7.60618299e-02 -5.82946122e-01
-4.34545547e-01 -8.83238614e-01 7.47076236e-03 -2.83789132e-02
-2.59775251e-01 -5.19196570e-01 5.74258804e-01 -2.72450328e-01
1.47209740e+00 -2.09128523e+00 1.63635835e-01 2.00259313e-01
5.35206795e-01 1.11813463e-01 -6.67982250e-02 3.42979252e-01
2.34112293e-01 2.29981244e-01 -1.85749203e-01 4.32054251e-01
7.91404068e-01 3.37584972e-01 -7.89796948e-01 -1.18237257e-01
4.32139099e-01 1.49084318e+00 -1.31951404e+00 -4.58245635e-01
3.66180055e-02 -3.08920503e-01 -1.33311141e+00 1.61596492e-01
-1.21278429e+00 -1.09945998e-01 -4.76279765e-01 7.73140430e-01
2.33124092e-01 -1.93509907e-01 6.16976917e-01 -7.46557191e-02
1.70517325e-01 4.02221859e-01 -9.75064158e-01 1.47254586e+00
-6.41941965e-01 5.39394021e-01 -1.03638090e-01 -5.56264758e-01
7.47466385e-01 1.88645020e-01 -1.58849090e-01 -6.65864050e-01
-2.13494971e-01 2.56658018e-01 2.36474633e-01 -5.25978923e-01
2.43425742e-01 -1.76773474e-01 -3.40159178e-01 4.69220281e-01
-4.32971977e-02 -7.59531975e-01 5.03684998e-01 1.93744510e-01
1.24379349e+00 4.95947987e-01 4.41826880e-01 -4.64835852e-01
5.05693972e-01 1.80645391e-01 6.69073820e-01 1.21575499e+00
-1.51654541e-01 -2.00547248e-01 1.18367648e+00 -5.92617393e-01
-7.56494522e-01 -1.13697076e+00 2.33605415e-01 1.32153440e+00
-3.56418006e-02 -7.84010231e-01 -7.89695084e-01 -8.36454034e-01
3.68522972e-01 1.08495688e+00 -3.69701624e-01 -5.42778432e-01
-6.11581564e-01 -1.34202540e-01 1.03827977e+00 5.78100204e-01
7.84298778e-01 -1.33903861e+00 -8.88383985e-01 1.07266784e-01
8.54782313e-02 -8.34253550e-01 -1.08739786e-01 1.10003084e-01
-8.42234313e-01 -9.97434676e-01 3.06849211e-01 -5.93212306e-01
4.71113622e-01 -2.10555241e-01 1.36500621e+00 1.43296421e-01
8.61084387e-02 5.28791070e-01 -4.81752008e-02 -4.87149537e-01
-7.44584739e-01 -1.41077697e-01 3.69001739e-02 -4.77059335e-01
2.49183774e-01 -4.33616400e-01 -1.65368021e-01 -1.45711005e-01
-1.01362658e+00 3.35655034e-01 8.02623332e-01 9.06585991e-01
1.69996843e-01 3.24648172e-01 6.66743815e-01 -9.74098325e-01
1.06302249e+00 2.38647265e-03 -1.01142836e+00 5.05780220e-01
-9.56908822e-01 4.71806824e-01 1.09359539e+00 -3.52103859e-01
-1.01898575e+00 -1.63100377e-01 2.48960212e-01 -2.37665907e-01
-8.93270597e-03 5.62117755e-01 -2.72683620e-01 -5.38740158e-02
8.99855077e-01 3.71792018e-01 4.21506353e-02 1.58266723e-01
6.26124978e-01 1.78558424e-01 5.09738266e-01 -1.49488616e+00
8.41075897e-01 -4.15064208e-02 1.01653039e-02 -1.96730152e-01
-7.66840696e-01 5.83581448e-01 -4.25569154e-02 6.98352531e-02
5.60562670e-01 -8.67454708e-01 -1.37571371e+00 2.86337048e-01
-9.89115477e-01 -9.87668872e-01 -4.16053623e-01 -4.98581640e-02
-8.42912376e-01 5.52024841e-02 -7.53885746e-01 -6.86713934e-01
-3.23850721e-01 -1.53413761e+00 8.85145307e-01 -1.64092761e-02
-4.54007387e-01 -9.40655470e-01 -7.46031106e-02 3.62962671e-03
4.02018547e-01 1.89522326e-01 1.79814351e+00 -2.89547503e-01
-8.51455986e-01 1.38646349e-01 -1.97358966e-01 5.65183043e-01
-1.46030754e-01 1.38986275e-01 -7.35026538e-01 -4.98865210e-02
-3.60083580e-01 -7.39662468e-01 4.92064983e-01 1.32677510e-01
1.27929366e+00 -8.84052336e-01 1.05927482e-01 5.00270605e-01
1.29410100e+00 2.10996404e-01 4.85823572e-01 3.31126511e-01
5.55653155e-01 4.01273012e-01 3.58455092e-01 3.14789742e-01
7.36348867e-01 3.93558621e-01 3.27135146e-01 3.13818574e-01
-5.48620820e-02 -8.36457133e-01 8.67689431e-01 2.15563506e-01
2.25070909e-01 2.34940141e-01 -1.18435609e+00 2.84886837e-01
-2.08800340e+00 -1.02676773e+00 1.41640648e-01 1.73359108e+00
1.53452945e+00 3.41372520e-01 6.34441078e-02 -7.16912597e-02
8.98571312e-02 6.33191466e-02 -7.11835325e-01 -6.26990139e-01
9.14557204e-02 2.71528125e-01 2.75250763e-01 6.11311197e-01
-8.29800665e-01 1.27154315e+00 6.87329721e+00 8.56170237e-01
-1.18090069e+00 -2.89160043e-01 1.50962025e-01 1.45736575e-01
-7.16131330e-01 1.82591245e-01 -7.12329805e-01 7.89259449e-02
6.36195183e-01 -3.86266321e-01 1.00161791e+00 1.12221265e+00
6.94984803e-04 2.34870650e-02 -1.75069535e+00 6.90888226e-01
-3.40130150e-01 -1.59799421e+00 5.35883307e-01 -3.33675712e-01
7.53442049e-01 -5.17424166e-01 1.92505047e-01 1.00386536e+00
7.53463268e-01 -1.20520747e+00 1.27747202e+00 4.11205232e-01
6.36508763e-01 -6.09337151e-01 2.37892717e-01 3.84359777e-01
-8.47008288e-01 -4.84610945e-01 1.80176727e-03 -4.70069140e-01
-5.26918709e-01 3.28576744e-01 -6.27428651e-01 3.74115407e-01
4.22822416e-01 6.86963618e-01 -8.00564349e-01 3.30095321e-01
-8.63931775e-01 7.00249851e-01 -1.23350490e-02 -3.05254936e-01
3.46159935e-01 -6.78481311e-02 3.49891216e-01 1.30657637e+00
-2.24931687e-01 -5.44588119e-02 3.88087511e-01 1.70602500e+00
-1.93398781e-02 -4.14944082e-01 -8.28700781e-01 -3.46878648e-01
4.45544958e-01 1.04739344e+00 -2.52850592e-01 -5.34353435e-01
-4.26465631e-01 2.76009649e-01 5.01486480e-01 5.50422013e-01
-9.54822063e-01 -4.50134844e-01 4.78871822e-01 -1.40518267e-02
4.85168993e-02 -2.50370830e-01 -7.52179980e-01 -1.14958644e+00
9.50220972e-02 -1.77709627e+00 2.85184652e-01 -6.43771052e-01
-1.20104408e+00 1.43167436e-01 3.86202425e-01 -7.02794552e-01
-2.34560341e-01 -9.54813480e-01 -5.03411472e-01 4.06855226e-01
-1.30248570e+00 -1.13565290e+00 -1.33814707e-01 4.19583559e-01
2.79659092e-01 -3.04194927e-01 6.68531418e-01 -1.71910182e-01
-5.02510190e-01 7.08359420e-01 -5.59563160e-01 7.99339339e-02
3.70073706e-01 -1.61885726e+00 1.26840159e-01 7.80887485e-01
-4.73306388e-01 1.19622254e+00 5.15040636e-01 -6.35427952e-01
-2.26011109e+00 -1.22619426e+00 5.02143979e-01 -5.60774803e-01
9.42480624e-01 -5.72674513e-01 -7.01937914e-01 9.89135981e-01
2.21026346e-01 -3.08273941e-01 2.10562393e-01 1.28448159e-01
-8.91267002e-01 -5.09972274e-01 -1.04396582e+00 1.22444952e+00
1.12091601e+00 -7.69797742e-01 -1.04447222e+00 2.57546276e-01
1.09673166e+00 -3.15140009e-01 -7.77259052e-01 5.52842498e-01
5.84663153e-01 -1.03251433e+00 8.96541238e-01 -6.53233111e-01
9.36795771e-01 -6.23275578e-01 -1.72460154e-01 -1.02238083e+00
-5.22059619e-01 -8.26444864e-01 -5.88499248e-01 1.03472579e+00
2.79911131e-01 -7.23649621e-01 3.85682136e-01 6.97863638e-01
-1.73062772e-01 -8.59957099e-01 -2.93069571e-01 -1.02605987e+00
3.40075105e-01 -6.24029338e-01 8.17079246e-01 6.32565796e-01
5.10433555e-01 6.40461147e-01 4.68078405e-02 1.23626355e-03
4.39880699e-01 4.65488404e-01 8.73846948e-01 -9.47897673e-01
-7.23279536e-01 -9.00547087e-01 1.39899030e-01 -9.73815918e-01
7.77602494e-01 -1.30666625e+00 2.94676095e-01 -1.39447176e+00
2.31503230e-02 -4.91401374e-01 3.46940830e-02 1.10905242e+00
-1.09658092e-01 -4.76119250e-01 3.07576299e-01 -8.26204941e-02
-7.92560875e-01 4.69065666e-01 1.27463472e+00 -4.76352841e-01
-3.05900455e-01 -3.44169706e-01 -9.98514652e-01 7.97149062e-01
6.80620551e-01 2.92021432e-03 -8.24234307e-01 -4.88481551e-01
7.98381627e-01 2.65572150e-03 6.89229906e-01 -1.10584641e+00
2.05056772e-01 -7.70554125e-01 -1.17991775e-01 -1.49854757e-02
-3.99543881e-01 -7.78783083e-01 -9.90780666e-02 8.62758696e-01
-6.78280473e-01 7.02884421e-02 4.03522611e-01 3.48886758e-01
-1.03624992e-01 -1.29524142e-01 6.83949828e-01 -1.97254375e-01
-9.92915809e-01 4.90942933e-02 -4.60177124e-01 2.31925026e-01
1.04732680e+00 1.70849070e-01 -6.00463748e-01 -1.84441209e-01
-8.85399207e-02 4.48790371e-01 5.10984778e-01 3.51996064e-01
5.64271927e-01 -1.34175527e+00 -4.99745995e-01 2.43546963e-01
3.87957573e-01 8.81463513e-02 -2.24865124e-01 7.52912045e-01
-8.20987284e-01 7.04847872e-01 -2.88709611e-01 -4.17266935e-01
-6.02546453e-01 8.38424385e-01 5.77424943e-01 -5.56473017e-01
-5.60687959e-01 3.64736021e-01 4.99273509e-01 -8.80571604e-01
2.87712127e-01 -1.02519858e+00 2.40828648e-01 -4.64013666e-01
4.52539772e-01 1.33806393e-01 -2.08034143e-01 5.54398835e-01
-3.65970224e-01 1.50367916e-01 1.96743682e-01 -8.54834542e-02
1.09490693e+00 5.49010992e-01 -5.57950139e-01 3.79143894e-01
3.80366832e-01 -8.50970298e-02 -1.27198339e+00 -2.38403440e-01
2.71783412e-01 -1.29767910e-01 -4.09641676e-02 -1.16689467e+00
-5.34864724e-01 7.05414593e-01 -1.81464925e-01 1.33715495e-01
9.93601322e-01 -3.46298844e-01 3.91307443e-01 1.11684835e+00
6.28185928e-01 -9.92276430e-01 1.66203260e-01 1.06159627e+00
9.49658334e-01 -8.94786775e-01 -4.61084619e-02 -1.72071829e-01
-6.90955997e-01 1.18666029e+00 1.06922734e+00 -2.39823580e-01
1.33358790e-02 5.29295385e-01 -4.59521323e-01 -1.10261612e-01
-1.24148703e+00 3.28356251e-02 3.27455282e-01 5.92118442e-01
4.26437020e-01 1.69640332e-01 -1.85167819e-01 7.69758284e-01
-3.75682265e-01 5.42240322e-01 3.24891210e-01 1.16016030e+00
-3.33327681e-01 -9.12290275e-01 -2.51577020e-01 2.81632751e-01
-1.11082315e-01 -2.27569699e-01 -3.48886132e-01 8.70478630e-01
1.25517696e-01 8.04043531e-01 -3.66431952e-01 -3.24539721e-01
4.03466225e-01 1.77039325e-01 8.10748637e-01 -6.80616021e-01
-8.03004265e-01 -5.35070181e-01 1.16258129e-01 -8.30099523e-01
9.06033143e-02 -1.42989859e-01 -1.43565619e+00 -6.06369674e-01
4.50579673e-01 -5.13402075e-02 6.83775172e-02 8.13230515e-01
3.57714802e-01 7.08273947e-01 9.94929969e-02 -2.70741850e-01
-1.13455474e+00 -3.14018011e-01 -2.70329863e-01 3.80501121e-01
5.15236259e-01 -4.22000140e-01 4.35878411e-02 7.62122348e-02] | [9.142034530639648, 7.198415756225586] |
6997cc53-f544-42e1-be6c-976032574202 | pv2tea-patching-visual-modality-to-textual | 2306.01016 | null | https://arxiv.org/abs/2306.01016v1 | https://arxiv.org/pdf/2306.01016v1.pdf | PV2TEA: Patching Visual Modality to Textual-Established Information Extraction | Information extraction, e.g., attribute value extraction, has been extensively studied and formulated based only on text. However, many attributes can benefit from image-based extraction, like color, shape, pattern, among others. The visual modality has long been underutilized, mainly due to multimodal annotation difficulty. In this paper, we aim to patch the visual modality to the textual-established attribute information extractor. The cross-modality integration faces several unique challenges: (C1) images and textual descriptions are loosely paired intra-sample and inter-samples; (C2) images usually contain rich backgrounds that can mislead the prediction; (C3) weakly supervised labels from textual-established extractors are biased for multimodal training. We present PV2TEA, an encoder-decoder architecture equipped with three bias reduction schemes: (S1) Augmented label-smoothed contrast to improve the cross-modality alignment for loosely-paired image and text; (S2) Attention-pruning that adaptively distinguishes the visual foreground; (S3) Two-level neighborhood regularization that mitigates the label textual bias via reliability estimation. Empirical results on real-world e-Commerce datasets demonstrate up to 11.74% absolute (20.97% relatively) F1 increase over unimodal baselines. | ['Xian Li', 'Carl Yang', 'Jingbo Shang', 'Chenwei Zhang', 'Nasser Zalmout', 'Rongmei Lin', 'Hejie Cui'] | 2023-06-01 | null | null | null | null | ['attribute-value-extraction'] | ['natural-language-processing'] | [ 6.47711992e-01 4.97001968e-02 -3.79063308e-01 -6.46998107e-01
-1.32146323e+00 -7.28032887e-01 5.99733889e-01 2.07297832e-01
-4.17250752e-01 6.57599509e-01 1.69308439e-01 1.96166113e-02
1.86246842e-01 -3.79423469e-01 -7.77927816e-01 -8.83281291e-01
2.81847626e-01 3.66986901e-01 -3.82237136e-02 7.39805549e-02
1.22518949e-01 1.40674561e-01 -1.64078176e+00 7.33838737e-01
9.17636991e-01 1.63864088e+00 1.17661707e-01 4.48418081e-01
-3.07405978e-01 7.01757550e-01 -4.07789797e-01 -8.18199515e-01
2.46639065e-02 -4.15630311e-01 -7.13818491e-01 5.18254638e-01
8.02385092e-01 -3.41370046e-01 1.03881508e-01 1.22375846e+00
4.95221406e-01 -1.92626774e-01 8.12155008e-01 -1.35059762e+00
-8.51199448e-01 5.35694301e-01 -9.12513793e-01 -2.29066685e-01
9.21603143e-02 3.29797745e-01 1.17321157e+00 -1.28235030e+00
6.67704761e-01 1.23471677e+00 3.73287052e-01 4.28893089e-01
-1.47794986e+00 -5.08738816e-01 2.75898039e-01 3.17627311e-01
-1.62053430e+00 -4.98630106e-01 8.78032029e-01 -3.96032810e-01
6.67799890e-01 2.99998999e-01 4.34673429e-01 1.50122678e+00
-1.15510955e-01 1.20620215e+00 1.22472501e+00 -2.55341798e-01
-5.50913513e-02 5.66397905e-01 -1.00540183e-01 6.44578457e-01
2.22711563e-02 -1.42922148e-01 -7.59227574e-01 2.54416913e-01
4.87025857e-01 -2.88400441e-01 -1.78187266e-01 -3.17523807e-01
-1.30301869e+00 5.98479450e-01 2.59317011e-01 -5.64739527e-03
-2.76115090e-01 -5.54653332e-02 4.91724849e-01 1.40198069e-02
3.64250034e-01 1.94077536e-01 -5.04946411e-01 -3.78671056e-03
-8.05146277e-01 -1.84137195e-01 3.65250409e-01 1.34204495e+00
6.19301319e-01 1.41062036e-01 -4.20334280e-01 1.07421279e+00
4.34277296e-01 8.91573191e-01 1.81950063e-01 -7.70640969e-01
8.37348223e-01 6.30664766e-01 -9.42336470e-02 -1.01945603e+00
-3.53314847e-01 -4.03430194e-01 -1.07169890e+00 -7.35596716e-02
5.28479099e-01 6.36488423e-02 -9.33855355e-01 1.65629995e+00
1.98294744e-01 -4.73493278e-01 -4.60300557e-02 1.16815150e+00
1.15969622e+00 5.48677444e-01 4.97047186e-01 -1.83181003e-01
1.72432482e+00 -9.57744420e-01 -8.85014296e-01 -4.16521817e-01
2.29762763e-01 -8.70051086e-01 1.21539021e+00 4.49477404e-01
-1.10425866e+00 -5.80730379e-01 -1.00180542e+00 -4.27378386e-01
-4.06540960e-01 6.26777291e-01 3.54944438e-01 4.80385363e-01
-6.80027843e-01 7.66614750e-02 -3.61672342e-01 -1.56369343e-01
7.20679224e-01 3.72309327e-01 -5.37017703e-01 -1.02628604e-01
-1.04485142e+00 7.01651812e-01 4.81113434e-01 2.07730368e-01
-7.75468349e-01 -4.78273869e-01 -1.04355562e+00 -3.23981699e-03
8.36128354e-01 -5.28870821e-01 8.27415526e-01 -1.49479103e+00
-1.17926753e+00 1.20944643e+00 -2.07684532e-01 1.70193166e-01
4.51040715e-01 -1.13799587e-01 -5.11944771e-01 2.91032821e-01
1.58587888e-01 1.05748379e+00 1.14519000e+00 -1.70352554e+00
-7.41149068e-01 -4.73748356e-01 -3.02450985e-01 4.30746913e-01
-5.82308829e-01 -2.31510222e-01 -8.87414396e-01 -9.28521156e-01
1.30154043e-01 -8.63886237e-01 2.91625977e-01 1.57654256e-01
-7.63481379e-01 -6.37384281e-02 6.50500476e-01 -7.95759201e-01
1.01880515e+00 -2.25978422e+00 1.90020099e-01 7.23870322e-02
1.37695462e-01 3.00291032e-02 -3.84137660e-01 -1.00257648e-02
5.56536727e-02 1.51093647e-01 -1.95580959e-01 -4.29994822e-01
1.16444968e-01 4.49833423e-02 -1.73169345e-01 3.31418961e-01
6.36901557e-01 1.04632854e+00 -6.04082406e-01 -1.04595256e+00
1.43210098e-01 5.94038010e-01 -2.48800099e-01 1.98181942e-01
-1.39612481e-01 5.32060623e-01 -2.67924666e-01 1.20668709e+00
7.31120169e-01 -4.15024340e-01 1.88953429e-01 -9.30954218e-01
1.37121901e-01 -6.42197505e-02 -1.05962777e+00 1.55405414e+00
-2.29247212e-01 6.06395721e-01 2.61274964e-01 -7.12323427e-01
7.94592500e-01 1.43081054e-01 4.55528378e-01 -8.90775502e-01
3.41549039e-01 6.17705509e-02 -3.21829408e-01 -7.31056094e-01
5.91718495e-01 -7.80343711e-02 -1.58282325e-01 8.81602019e-02
2.20697060e-01 8.80537778e-02 1.58398271e-01 1.58779845e-01
4.09256488e-01 2.92784423e-01 1.51294887e-01 -8.80413502e-02
5.91118097e-01 -1.51704147e-01 6.71261132e-01 4.93840843e-01
-3.72287065e-01 8.57989967e-01 5.90634346e-01 -1.02758057e-01
-1.00486052e+00 -9.77935791e-01 -3.41966063e-01 1.37820816e+00
4.72549349e-01 -2.56657571e-01 -4.98071581e-01 -9.40648615e-01
-6.68382049e-02 6.30691588e-01 -5.84070206e-01 -1.09385885e-01
-2.58069247e-01 -7.36419916e-01 4.91665989e-01 8.00514519e-01
5.18580973e-01 -8.87838721e-01 -2.49916494e-01 -9.22653824e-02
-5.90612650e-01 -1.42803347e+00 -7.42064714e-01 4.02125180e-01
-5.88419259e-01 -8.07970703e-01 -7.53630757e-01 -7.11489141e-01
8.14134002e-01 8.22569206e-02 1.28133631e+00 -6.80977702e-02
3.07080727e-02 5.42211533e-01 -3.53561610e-01 -2.42233455e-01
-1.67042196e-01 4.50122207e-02 -2.59190589e-01 4.47969019e-01
6.00227237e-01 7.08467066e-02 -6.21420264e-01 4.63568926e-01
-7.74106443e-01 3.96029115e-01 9.37679052e-01 1.10892045e+00
8.95304382e-01 -2.71843910e-01 4.02321696e-01 -9.10736203e-01
1.34553656e-01 -2.74017572e-01 -4.52367276e-01 5.24435699e-01
-6.94747865e-01 3.29782031e-02 3.67496729e-01 -6.02932990e-01
-1.33963037e+00 4.02756721e-01 8.84905681e-02 -3.04370522e-01
-2.82909214e-01 3.47511977e-01 -7.23179519e-01 1.38992161e-01
2.02471837e-01 2.63810009e-01 -1.13891780e-01 -1.98441073e-01
4.83939588e-01 7.94901431e-01 6.68312192e-01 -6.42472029e-01
5.43122172e-01 3.17079097e-01 -9.67312828e-02 -7.57322133e-01
-9.22045410e-01 -2.76450604e-01 -8.08635414e-01 -4.43101466e-01
1.09217060e+00 -1.03715658e+00 -7.47626662e-01 3.60868484e-01
-1.11034191e+00 -1.14730276e-01 3.38698179e-03 3.73981625e-01
-3.20960581e-01 3.41557533e-01 -6.19046986e-01 -9.04108584e-01
-3.39404404e-01 -1.22966337e+00 1.43130708e+00 2.88196981e-01
-7.33311698e-02 -6.64373338e-01 -6.34283006e-01 9.76623833e-01
2.28594065e-01 7.42603391e-02 9.77160633e-01 -5.81315100e-01
-5.95618665e-01 3.18113863e-02 -7.32384443e-01 4.67779875e-01
-5.68883717e-02 4.01423015e-02 -1.33919144e+00 -1.21946730e-01
-3.25071961e-01 -6.37632728e-01 8.57487857e-01 2.31362268e-01
1.29878521e+00 -2.58611828e-01 -1.98916599e-01 5.93501091e-01
1.15479922e+00 1.41634479e-01 5.59226751e-01 1.66451320e-01
1.19554746e+00 9.23153698e-01 7.88368225e-01 3.81619900e-01
5.55764318e-01 7.60189772e-01 5.19399107e-01 -3.61348152e-01
-3.79931450e-01 -1.90912873e-01 4.38991159e-01 7.65957177e-01
1.79158058e-02 -2.81193852e-01 -7.00126529e-01 4.13188517e-01
-1.80556095e+00 -7.14511573e-01 -3.11199367e-01 2.03108048e+00
1.16407120e+00 4.38739620e-02 1.42773256e-01 -9.44212973e-02
8.62001181e-01 -2.17789859e-02 -7.08738267e-01 2.37789564e-02
-5.53837836e-01 -3.79495591e-01 5.67929685e-01 1.22907259e-01
-1.47037137e+00 7.51342416e-01 5.05659771e+00 1.03071010e+00
-9.88555431e-01 2.81471033e-02 1.09466875e+00 -1.57887831e-01
-3.41571808e-01 -2.96212703e-01 -9.72302139e-01 5.96169829e-01
2.91465312e-01 3.97598952e-01 2.08535597e-01 6.94178998e-01
-2.27448106e-01 -1.02297015e-01 -1.19183648e+00 1.28313935e+00
4.85087812e-01 -7.50804245e-01 2.27671057e-01 3.25684473e-02
6.83005691e-01 -4.65441167e-01 4.19465750e-01 3.89591426e-01
-1.90591633e-01 -1.04365253e+00 1.16984141e+00 3.16797614e-01
1.36130905e+00 -7.73419261e-01 7.70089030e-01 1.33823566e-02
-1.26902092e+00 -6.70551062e-02 -3.78194898e-02 6.99506938e-01
1.68709800e-01 5.54382384e-01 -5.11963308e-01 5.02799749e-01
7.90412843e-01 7.31478751e-01 -8.94007385e-01 4.67586070e-01
-3.82244349e-01 5.10028303e-01 -1.08716704e-01 7.15292394e-02
1.08768471e-01 -2.75646865e-01 3.98414731e-01 1.43309629e+00
7.24123865e-02 -4.33111414e-02 7.32372925e-02 9.65905309e-01
-1.90593228e-01 1.37516633e-01 -2.56732672e-01 -1.88880041e-01
3.94493699e-01 1.48977852e+00 -8.17033648e-01 -2.77633250e-01
-6.78010821e-01 1.03625858e+00 1.13772698e-01 5.19227207e-01
-9.57782865e-01 -2.23408952e-01 3.00301582e-01 9.87089588e-04
5.14040411e-01 1.14308037e-01 -6.82777405e-01 -1.17815852e+00
9.62402672e-02 -1.10036814e+00 5.05127430e-01 -9.21785951e-01
-1.58789420e+00 6.32341921e-01 -2.27783397e-01 -1.36107385e+00
8.68012682e-02 -7.64331520e-01 1.88656673e-01 7.21425474e-01
-1.35798931e+00 -1.64702642e+00 -4.00641799e-01 5.63356400e-01
6.52679682e-01 -1.98665172e-01 6.26769423e-01 6.53255820e-01
-7.07956970e-01 9.38937664e-01 -1.76303074e-01 3.59284908e-01
1.09968698e+00 -1.28121161e+00 -3.27517539e-01 7.69129276e-01
1.05601482e-01 3.84535044e-01 4.34772581e-01 -6.00478768e-01
-1.61304796e+00 -1.11151314e+00 8.34023654e-01 -5.34352422e-01
4.22531158e-01 -5.35237968e-01 -9.64365542e-01 4.07564849e-01
3.36827338e-01 -1.17531987e-02 7.51361072e-01 1.55966997e-01
-5.96946716e-01 -3.62336755e-01 -1.04757082e+00 6.35825157e-01
6.21681094e-01 -6.89747870e-01 -1.71866953e-01 8.09561089e-02
3.43423396e-01 -3.40572178e-01 -9.38700318e-01 4.82139170e-01
6.12011135e-01 -7.10256338e-01 1.01216686e+00 -3.18739504e-01
7.20262706e-01 -3.79545659e-01 -3.88363630e-01 -7.93937027e-01
-2.06865191e-01 -2.14909375e-01 -1.97652966e-01 1.74964631e+00
5.70894539e-01 -6.80942535e-02 4.69539642e-01 8.50908279e-01
1.33298606e-01 -7.20364690e-01 -7.76643455e-01 -5.69003761e-01
-2.59798378e-01 -4.31200981e-01 2.91979104e-01 1.00496328e+00
-1.33116618e-01 8.33999753e-01 -7.08359182e-01 1.24170212e-02
6.92819715e-01 1.52927727e-01 5.17427087e-01 -9.38616693e-01
-1.27515122e-01 -6.05669022e-01 -1.07464686e-01 -1.11234343e+00
-4.90521640e-03 -8.92547011e-01 2.60861427e-01 -1.22055650e+00
6.62873805e-01 -2.53575593e-01 -2.59491682e-01 6.38642728e-01
-3.32950830e-01 5.27081430e-01 1.82929561e-01 2.39139330e-03
-9.19482052e-01 6.93260789e-01 1.29620969e+00 -4.47706312e-01
1.60124227e-01 -2.95390695e-01 -7.93599308e-01 6.55653775e-01
4.62045997e-01 -3.53948861e-01 -3.51299405e-01 -4.07691061e-01
4.42766637e-01 5.17647602e-02 3.83509636e-01 -4.56664652e-01
3.14222127e-02 -2.63701618e-01 8.15694332e-01 -7.88041949e-01
4.24952745e-01 -1.08865571e+00 -2.51978844e-01 -4.06816043e-02
-5.42220891e-01 -1.37473881e-01 1.35458380e-01 7.05220580e-01
-2.98887521e-01 -2.07391024e-01 8.20013881e-01 2.75267720e-01
-7.70005703e-01 1.39663085e-01 -2.00743854e-01 1.75476834e-01
7.93417335e-01 -1.73647046e-01 -5.19882441e-01 -4.12192196e-01
-5.12667954e-01 3.57541025e-01 3.66669357e-01 4.97677207e-01
6.13574684e-01 -1.55724025e+00 -7.32484519e-01 1.00516520e-01
5.21391809e-01 -1.18889615e-01 3.28878492e-01 1.22408700e+00
-1.00452565e-02 3.67439277e-02 -1.16626710e-01 -8.68171751e-01
-1.53065848e+00 6.62342131e-01 -6.14465214e-02 1.00110121e-01
-4.13884372e-01 8.14657688e-01 3.29214275e-01 -9.25102606e-02
4.78110492e-01 -3.05421837e-02 -2.60417163e-01 5.84492326e-01
4.08873171e-01 2.52148062e-01 -3.43366228e-02 -1.12527454e+00
-4.67708766e-01 5.78985691e-01 -1.15237884e-01 -1.21982880e-01
1.04425895e+00 -5.06435335e-01 -2.38002073e-02 5.52869380e-01
1.23375130e+00 -2.07294837e-01 -1.42834890e+00 -3.92418474e-01
-6.38859570e-02 -4.42416549e-01 3.22687700e-02 -1.19840848e+00
-1.27311265e+00 1.03990591e+00 7.75459051e-01 7.45027512e-02
1.32932663e+00 5.06003536e-02 7.13479757e-01 1.25990197e-01
-8.83794948e-03 -1.55779958e+00 2.62748629e-01 3.37902606e-01
8.21688533e-01 -1.73269200e+00 6.65387213e-02 -5.64055860e-01
-1.34808993e+00 1.03262711e+00 7.38379240e-01 6.61256909e-01
2.38850310e-01 2.87807941e-01 1.97508797e-01 -1.54653013e-01
-6.90504074e-01 -3.73713046e-01 9.61116612e-01 5.64765692e-01
5.91165543e-01 -6.11094758e-02 2.89264116e-02 1.11678886e+00
3.21056187e-01 -4.32318807e-01 -3.78966779e-02 5.73679745e-01
-1.12195656e-01 -7.48553693e-01 -5.26803195e-01 4.64028388e-01
-4.54018056e-01 -1.65980190e-01 -5.25451422e-01 6.76382422e-01
4.02860403e-01 1.03890729e+00 4.27853279e-02 -3.09814751e-01
1.79514140e-01 1.15049161e-01 4.85050201e-01 -2.13923991e-01
-5.26692569e-01 5.98481596e-01 2.07252741e-01 -4.94688421e-01
-4.73225534e-01 -6.46769941e-01 -1.03799832e+00 3.81455794e-02
-3.61831546e-01 -3.35129052e-01 7.49902248e-01 8.06302011e-01
3.37916583e-01 4.72474426e-01 3.36616874e-01 -7.51650870e-01
-3.60118598e-01 -9.09888446e-01 -5.80656767e-01 7.93222606e-01
2.46088669e-01 -7.28850424e-01 -3.04261148e-01 5.22565663e-01] | [10.77523422241211, 1.3511593341827393] |
80b4b623-4f73-41c7-9ed0-2a813cba315f | synopses-of-movie-narratives-a-video-language-1 | 2203.05711 | null | https://arxiv.org/abs/2203.05711v4 | https://arxiv.org/pdf/2203.05711v4.pdf | Synopses of Movie Narratives: a Video-Language Dataset for Story Understanding | Despite recent advances of AI, story understanding remains an open and under-investigated problem. We collect, preprocess, and publicly release a video-language story dataset, Synopses of Movie Narratives (SyMoN), containing 5,193 video summaries of popular movies and TV series with a total length of 869 hours. SyMoN captures naturalistic storytelling videos made by human creators and intended for a human audience. As a prototypical and naturalistic story dataset, SyMoN features high coverage of multimodal story events and abundant mental-state descriptions. Its use of storytelling techniques cause cross-domain semantic gaps that provide appropriate challenges to existing models. We establish benchmarks on video-text retrieval and zero-shot alignment on movie summary videos, which showcase the importance of in-domain data and long-term memory in story understanding. With SyMoN, we hope to lay the groundwork for progress in multimodal story understanding. | ['Yangfeng Ji', 'Boyang Li', 'Qin Chao', 'Yidan Sun'] | 2022-03-11 | null | null | null | null | ['video-text-retrieval'] | ['computer-vision'] | [ 3.61415476e-01 -2.84411430e-01 -6.25104368e-01 -2.51361549e-01
-1.04674006e+00 -9.31906044e-01 1.07229698e+00 3.02013248e-01
-9.70924273e-02 6.05060995e-01 1.19459832e+00 3.28618407e-01
-1.08722508e-01 -3.64591271e-01 -7.07125604e-01 -1.46287799e-01
-1.25204548e-01 3.31510216e-01 1.78438902e-01 -3.12320381e-01
4.85750824e-01 -2.54708886e-01 -1.67228067e+00 1.27530968e+00
-4.67062779e-02 7.13944137e-01 4.32048470e-01 1.12618327e+00
-3.42766605e-02 1.69001019e+00 -3.99544716e-01 -7.07684338e-01
-2.46511027e-01 -5.10405302e-01 -1.00581324e+00 4.34542984e-01
6.61184311e-01 -4.58171606e-01 -1.14806259e+00 5.96448541e-01
3.69374543e-01 4.01337266e-01 6.28465950e-01 -1.52036345e+00
-5.90996981e-01 9.81566131e-01 -5.42509444e-02 5.53945482e-01
1.10540318e+00 9.69787240e-02 1.22308576e+00 -7.25853205e-01
1.48173344e+00 1.04228604e+00 6.61386490e-01 5.31883538e-01
-8.79194021e-01 -2.27525190e-01 -9.46068168e-02 6.88413799e-01
-1.16457856e+00 -8.34259570e-01 6.51048958e-01 -7.06500828e-01
1.04092157e+00 2.32767835e-01 8.96153867e-01 1.74736023e+00
-9.83119011e-04 1.36929476e+00 2.91771770e-01 4.58393693e-02
-1.18833920e-02 -3.20892662e-01 1.02386968e-02 2.84102976e-01
-4.52533573e-01 -5.49630642e-01 -1.32100201e+00 -8.80743638e-02
8.23719203e-01 -7.59309009e-02 -3.72771293e-01 -4.06386048e-01
-1.88028657e+00 6.35327458e-01 -2.54564553e-01 3.37478310e-01
-3.83310705e-01 1.55034646e-01 1.02383745e+00 4.04476076e-01
2.99761593e-01 5.69971561e-01 7.55030289e-02 -1.30820596e+00
-9.53997314e-01 8.39100182e-01 8.18406582e-01 1.23255765e+00
-8.78909752e-02 -2.19225124e-01 -2.73417979e-01 9.42790031e-01
-4.08991963e-01 3.87867808e-01 3.55839282e-01 -1.35614336e+00
9.04214859e-01 4.14536387e-01 1.44315049e-01 -1.41942644e+00
-3.46989632e-01 3.09727818e-01 -4.17463541e-01 -8.71938050e-01
3.87952298e-01 7.37634078e-02 -2.63393134e-01 1.65730202e+00
-1.51616737e-01 3.13096374e-01 2.34097913e-01 8.94758463e-01
1.24581742e+00 1.05344307e+00 -1.42698035e-01 -2.73089141e-01
1.31379211e+00 -9.28117573e-01 -8.73118818e-01 -4.18269128e-01
6.58832133e-01 -7.68751442e-01 1.00923669e+00 4.28357214e-01
-1.47215557e+00 -2.93744177e-01 -9.22659397e-01 -3.92343014e-01
-1.33292362e-01 -1.92231357e-01 5.57834923e-01 -1.42071724e-01
-6.13180816e-01 5.20110130e-01 -5.46882212e-01 -8.81843269e-01
4.59786534e-01 -3.79630327e-01 -7.07609177e-01 -4.99715328e-01
-1.03503120e+00 6.26176536e-01 5.73901713e-01 -5.21213293e-01
-1.22622871e+00 -8.49190772e-01 -1.04238904e+00 -3.05193752e-01
6.21952713e-01 -3.74194473e-01 1.56131470e+00 -7.94679821e-01
-1.04092693e+00 1.02004135e+00 8.37704912e-03 -5.28311789e-01
2.43476436e-01 -3.89367104e-01 -6.46573961e-01 8.31898212e-01
2.43185595e-01 8.65244389e-01 6.14970207e-01 -6.44172132e-01
-4.79024202e-01 1.01576507e-01 3.40038627e-01 4.11557168e-01
-4.90956396e-01 5.02457619e-01 -6.39096379e-01 -9.57287073e-01
-1.93667755e-01 -7.42384136e-01 2.94034094e-01 -2.87390202e-01
-3.57269526e-01 1.35613993e-01 7.21588790e-01 -8.62613678e-01
1.44015157e+00 -2.23017788e+00 5.53621709e-01 -5.93926966e-01
2.45928004e-01 -3.17975730e-01 -3.28537911e-01 1.25241017e+00
-1.25490710e-01 -6.31071329e-02 -8.70338455e-02 -3.69806916e-01
6.89211339e-02 -1.25499824e-02 -6.36580348e-01 3.38398069e-01
3.94574553e-02 9.64650869e-01 -1.10553670e+00 -7.07680523e-01
1.17976926e-01 2.27920607e-01 -5.82225561e-01 2.25402877e-01
-5.83303034e-01 2.21157625e-01 -3.42670560e-01 5.96581399e-01
-1.37614891e-01 -3.87532711e-01 -2.24297997e-02 -1.46854401e-01
-8.87052342e-02 4.33664560e-01 -7.55544126e-01 2.54165101e+00
1.06231414e-01 1.41285098e+00 -3.61065984e-01 -5.69598138e-01
4.87065315e-01 5.86427391e-01 7.18721449e-01 -4.43554401e-01
2.51823038e-01 -3.63110900e-01 -5.52154720e-01 -1.10541081e+00
1.07195342e+00 -9.31808129e-02 -7.20136404e-01 6.21798813e-01
3.03549767e-01 -3.56545061e-01 6.93050861e-01 7.79033661e-01
1.35054874e+00 1.33482859e-01 4.08154011e-01 2.33447313e-01
-4.24911082e-02 7.56331682e-01 4.60000336e-02 7.23766744e-01
-7.27690980e-02 1.09814227e+00 8.12965393e-01 -5.02571046e-01
-1.42977965e+00 -8.71383011e-01 1.98068619e-01 1.28253305e+00
2.75460239e-02 -1.12201405e+00 -7.19881117e-01 -6.17340617e-02
-2.76433647e-01 8.11584413e-01 -4.25813556e-01 -1.68612674e-02
-6.15719855e-01 -3.37093592e-01 8.13971102e-01 5.22559583e-01
3.57374310e-01 -1.01023626e+00 -4.96172518e-01 4.05766726e-01
-8.87624025e-01 -1.68791628e+00 -6.84011161e-01 -5.58941245e-01
-3.58731538e-01 -1.28093159e+00 -7.47187138e-01 -5.88375926e-01
2.00178679e-02 4.92040396e-01 1.48386157e+00 -2.99982220e-01
-2.25744292e-01 8.29434216e-01 -7.32303798e-01 8.25709626e-02
-5.46124160e-01 5.88737801e-03 1.07324339e-01 2.37993747e-02
2.63914406e-01 -6.00391090e-01 -8.00076351e-02 3.11343580e-01
-8.46688211e-01 6.22104228e-01 -4.61925082e-02 5.90928972e-01
2.51025230e-01 -2.23202541e-01 4.22664493e-01 -4.38784331e-01
5.31590521e-01 -1.01504731e+00 2.78081354e-02 1.40581086e-01
3.17029834e-01 -5.45280814e-01 4.02215093e-01 -5.50519049e-01
-9.80095387e-01 -2.41314292e-01 2.34892651e-01 -6.41509831e-01
-1.59484804e-01 5.73518455e-01 1.45574406e-01 6.84183121e-01
5.05732477e-01 3.54708910e-01 -2.89768159e-01 -3.82618278e-01
4.24828678e-01 4.46263433e-01 1.25611031e+00 -5.30208170e-01
3.16960365e-01 5.25953770e-01 -3.92277718e-01 -1.04647219e+00
-8.65731657e-01 -6.21427476e-01 -5.06061077e-01 -7.68996537e-01
1.11400735e+00 -1.21434295e+00 -7.59906471e-01 3.80771399e-01
-1.35958707e+00 -2.08238319e-01 -2.35986188e-01 3.78771931e-01
-1.08063734e+00 2.55863637e-01 -8.24397504e-01 -4.14575785e-01
9.68766510e-02 -5.95262825e-01 1.02331948e+00 -1.27434656e-01
-9.60323274e-01 -7.99912632e-01 1.60386980e-01 8.93866539e-01
-8.37191343e-02 4.24084157e-01 6.99463665e-01 -6.63067162e-01
-5.41332781e-01 -3.81804436e-01 -1.15788661e-01 -3.72522146e-01
-4.46651369e-01 1.06909715e-01 -7.28467405e-01 -8.58052373e-02
-2.42210746e-01 -9.69125450e-01 7.69712269e-01 3.09440821e-01
8.34078550e-01 -2.81277418e-01 -1.71141326e-01 2.09273264e-01
1.04469478e+00 9.17241052e-02 6.64226830e-01 3.04113477e-01
7.20763743e-01 7.22348273e-01 7.61000872e-01 9.86602604e-01
7.42312133e-01 6.30789697e-01 2.68598020e-01 5.09785473e-01
-1.41775236e-01 -7.95940161e-01 6.35596752e-01 1.20484483e+00
-1.08545743e-01 -7.88397729e-01 -1.05132103e+00 9.35138047e-01
-2.13570094e+00 -1.80805874e+00 -7.84047469e-02 1.66481173e+00
6.15296423e-01 1.59793913e-01 3.76732856e-01 -1.04356229e-01
7.12409735e-01 7.42600918e-01 -3.55615377e-01 -1.22523405e-01
-5.98820508e-01 -7.25402653e-01 1.39912471e-01 1.52248219e-01
-1.20379341e+00 7.73495972e-01 6.84465170e+00 1.03104508e+00
-4.12788898e-01 1.35396242e-01 2.44795606e-01 -8.77275527e-01
-1.85885355e-01 -1.18241228e-01 -4.30608362e-01 3.64446521e-01
1.05650270e+00 -4.95260388e-01 5.06894767e-01 7.11153686e-01
2.56127268e-01 -2.94782847e-01 -1.49476576e+00 1.35427701e+00
7.92363226e-01 -1.99018157e+00 9.13600549e-02 -4.02775675e-01
8.27199280e-01 -2.50299852e-02 -3.20278287e-01 2.72516489e-01
-3.12855281e-02 -9.24081028e-01 1.07104039e+00 6.03928089e-01
6.93914115e-01 -6.37400627e-01 4.26915914e-01 4.19920772e-01
-1.29670548e+00 -1.12277947e-01 -9.18308794e-02 -4.22750831e-01
7.27070510e-01 -1.70862507e-02 -5.82447231e-01 2.41594359e-01
5.20104408e-01 1.67043781e+00 -4.53362942e-01 7.27924824e-01
2.70089298e-01 3.78856272e-01 1.49653303e-02 -8.42646975e-03
3.83399576e-01 6.44484982e-02 9.82355118e-01 1.39529073e+00
1.04482986e-01 4.93178278e-01 1.98422715e-01 4.94469166e-01
-2.49646366e-01 -4.87486757e-02 -1.11399508e+00 -9.53118384e-01
4.13422823e-01 8.25501442e-01 -7.02737868e-01 -5.78906596e-01
-4.34140086e-01 1.07938528e+00 3.91522236e-02 5.98686077e-02
-1.00553823e+00 7.06075355e-02 7.03803122e-01 1.82165846e-01
5.09155095e-02 -2.94501454e-01 1.97252166e-02 -1.42963469e+00
-1.03514992e-01 -1.01243103e+00 4.03215855e-01 -1.38590825e+00
-1.16252041e+00 5.33716619e-01 5.98237574e-01 -1.28262556e+00
-6.15645945e-01 -8.95205364e-02 -4.02557611e-01 -2.90832072e-01
-4.88725185e-01 -1.10497200e+00 -2.71123946e-01 6.90619469e-01
1.46706104e+00 -3.68445486e-01 5.54277718e-01 3.84127170e-01
-4.40478235e-01 3.15072611e-02 9.92034078e-02 4.95760255e-02
6.95108891e-01 -7.29374349e-01 5.81582308e-01 6.62051380e-01
4.25791413e-01 1.58282906e-01 1.05952024e+00 -7.90130436e-01
-1.72629261e+00 -7.39703655e-01 1.01604605e+00 -1.00680983e+00
1.18115938e+00 -5.86149573e-01 -5.90357602e-01 8.75075042e-01
5.43293655e-01 -6.74870670e-01 7.77732670e-01 -2.62464732e-01
-2.87059933e-01 5.09444535e-01 -5.86508930e-01 8.04253876e-01
1.38452351e+00 -9.15470839e-01 -1.07340693e+00 7.19361126e-01
7.76816189e-01 -5.47942162e-01 -1.03742826e+00 -4.42216806e-02
8.51375639e-01 -8.14809561e-01 1.00419044e+00 -8.87250662e-01
1.47682440e+00 2.57429123e-01 -4.63119417e-01 -8.79179955e-01
1.20510221e-01 -9.61622000e-01 5.08309491e-02 1.44780886e+00
1.42835572e-01 5.34110844e-01 6.21132433e-01 7.02116251e-01
-2.32944444e-01 -4.01353270e-01 -6.04555428e-01 -5.48606873e-01
-4.15777475e-01 -1.09136820e+00 3.62068892e-01 1.17939889e+00
7.30603755e-01 4.91721004e-01 -8.74736011e-01 -4.61032838e-01
3.16449702e-01 5.21760322e-02 9.44948614e-01 -8.21142972e-01
-2.05836058e-01 -2.54775375e-01 -4.51487660e-01 -1.22656643e+00
2.83531159e-01 -7.19711006e-01 -6.48633018e-02 -1.45958865e+00
7.49912500e-01 4.92017984e-01 2.64195472e-01 2.66579151e-01
4.46347952e-01 4.68640566e-01 4.83323932e-01 3.72355372e-01
-1.35586083e+00 5.41543663e-01 1.09853828e+00 -1.49763465e-01
1.33297071e-01 -5.63325047e-01 -2.85900086e-01 9.95686531e-01
3.88793558e-01 -4.68908697e-01 -5.61292112e-01 -5.47898352e-01
6.37502849e-01 7.77454078e-01 5.57073891e-01 -9.15652752e-01
3.63183737e-01 -3.33714455e-01 -2.01213546e-02 -9.76018906e-01
1.02956259e+00 -4.15259689e-01 5.01445711e-01 -1.53122768e-01
-7.97795475e-01 3.60498101e-01 2.58530110e-01 6.29240572e-01
-5.35733402e-01 -9.06990934e-03 1.33861274e-01 -1.23467036e-01
-1.16350234e+00 1.04565106e-01 -7.68037081e-01 4.90314484e-01
1.03929317e+00 -5.09334624e-01 -6.54446423e-01 -9.90104258e-01
-6.26772821e-01 2.60130882e-01 6.28138900e-01 9.26043093e-01
8.58453035e-01 -1.51939452e+00 -1.01421404e+00 -3.04846555e-01
5.69622338e-01 -5.83226383e-01 6.60821259e-01 5.77982903e-01
-3.94482225e-01 4.05858964e-01 -4.18463856e-01 -3.61634284e-01
-1.48309004e+00 4.17350173e-01 -4.81769115e-01 1.80160165e-01
-8.43908012e-01 8.30194235e-01 2.31701974e-02 1.73459232e-01
1.34367377e-01 -1.12887071e-02 -3.50916386e-01 4.56644297e-01
9.66614485e-01 4.85534698e-01 -5.31257331e-01 -9.65241611e-01
-2.54987687e-01 3.39915991e-01 5.77218086e-02 -3.05635899e-01
1.50767326e+00 -4.94861960e-01 1.07738584e-01 1.12365639e+00
1.22518408e+00 -4.18290585e-01 -1.20879090e+00 -1.27606228e-01
1.34216949e-01 -4.74899411e-01 -2.07799062e-01 -4.81140971e-01
-5.70973814e-01 6.76614583e-01 -2.71007568e-01 9.18589011e-02
8.99463713e-01 3.58435303e-01 1.27443421e+00 6.28462255e-01
2.75974721e-01 -1.05022442e+00 3.92893344e-01 7.05186427e-01
1.19069660e+00 -1.30040276e+00 6.28434420e-02 -1.15263306e-01
-1.22403753e+00 1.09131777e+00 2.41168916e-01 -3.86715122e-02
9.68851224e-02 1.22047454e-01 -4.52223063e-01 -4.36358571e-01
-1.37857866e+00 1.42349586e-01 2.36469612e-01 2.59889811e-01
1.89481497e-01 -6.96090907e-02 -5.56464307e-03 1.02223349e+00
-3.70330095e-01 1.87010959e-01 8.18995535e-01 8.19743812e-01
-3.18599820e-01 -4.88619208e-01 -3.47262770e-01 1.34308577e-01
-3.53235692e-01 -1.38952568e-01 -6.42409146e-01 8.36371422e-01
-2.18579933e-01 1.06045806e+00 3.05272788e-01 -5.25925756e-01
2.04860106e-01 -1.86747331e-02 6.48766458e-01 -5.02203524e-01
-5.53072155e-01 -1.47171751e-01 8.57948959e-01 -7.62880325e-01
-5.17224133e-01 -1.09554482e+00 -9.55624104e-01 -7.20800102e-01
4.72449154e-01 -1.41729295e-01 4.46594477e-01 1.13943851e+00
3.18543881e-01 2.90648073e-01 1.29644200e-01 -1.01263583e+00
2.59025186e-01 -8.74891937e-01 -4.74020630e-01 7.44700134e-01
1.79960370e-01 -5.70227444e-01 -2.67540943e-02 7.19816864e-01] | [10.502279281616211, 0.8036318421363831] |
1ed08b4b-f559-4b45-8a79-3dbdc8a013f3 | 2305-14704 | 2305.14704 | null | https://arxiv.org/abs/2305.14704v2 | https://arxiv.org/pdf/2305.14704v2.pdf | An Evaluation on Practical Batch Bayesian Sampling Algorithms for Online Adaptive Traffic Experimentation | To speed up online testing, adaptive traffic experimentation through multi-armed bandit algorithms is rising as an essential complementary alternative to the fixed horizon A/B testing. Based on recent research on best arm identification and statistical inference with adaptively collected data, this paper derives and evaluates four Bayesian batch bandit algorithms (NB-TS, WB-TS, NB-TTTS, WB-TTTS), which are combinations of two ways of weighting batches (Naive Batch and Weighted Batch) and two Bayesian sampling strategies (Thompson Sampling and Top-Two Thompson Sampling) to adaptively determine traffic allocation. These derived Bayesian sampling algorithms are practically based on summary batch statistics of a reward metric for pilot experiments, where one of the combination WB-TTTS in this paper seems to be newly discussed. The comprehensive evaluation on the four Bayesian sampling algorithms covers trustworthiness, sensitivity and regret of a testing methodology. Moreover, the evaluation includes 4 real-world eBay experiments and 40 reproducible synthetic experiments to reveal the learnings, which covers both stationary and non-stationary situations. Our evaluation reveals that, (a) There exist false positives inflation with equivalent best arms, while seldom discussed in literatures; (b) To control false positives, connections between convergence of posterior optimal probabilities and neutral posterior reshaping are discovered; (c) WB-TTTS shows competitive recall, higher precision, and robustness against non-stationary trend; (d) NB-TS outperforms on minimizing regret trials except on precision and robustness; (e) WB-TTTS is a promising alternative if regret of A/B Testing is affordable, otherwise NB-TS is still a powerful choice with regret consideration for pilot experiments. | ['Ted Yuan', 'Zezhong Zhang'] | 2023-05-24 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [-1.95068028e-02 -2.11981654e-01 -7.90600598e-01 -3.45608145e-01
-8.45927298e-01 -4.78174001e-01 3.40191871e-01 -4.31913555e-01
-3.64374965e-01 1.39676809e+00 -2.70864725e-01 -1.00812936e+00
-9.81504440e-01 -5.62463760e-01 -7.90776432e-01 -8.68102014e-01
-3.03878009e-01 9.23328757e-01 5.41399777e-01 1.55392522e-02
3.25339556e-01 5.01721025e-01 -1.49244213e+00 3.96362059e-02
7.10825205e-01 1.56268311e+00 -2.37154856e-01 6.14919245e-01
5.81577681e-02 4.58589733e-01 -6.77800417e-01 -1.00245821e+00
6.40923619e-01 -4.28545237e-01 -5.01296341e-01 -9.08097029e-02
-1.60021141e-01 -4.77274895e-01 1.18302643e-01 9.54572916e-01
4.88009363e-01 8.30193236e-03 4.53063041e-01 -1.40409017e+00
-1.96577519e-01 9.23187137e-01 -9.34458256e-01 5.69724441e-01
1.36978561e-02 4.92872447e-01 9.48005259e-01 -7.43712932e-02
-1.26887560e-01 1.28948534e+00 6.29474878e-01 1.45479500e-01
-9.81228292e-01 -8.37106466e-01 3.59930366e-01 5.85431993e-01
-1.22028828e+00 -2.48771057e-01 5.10091662e-01 -1.67311624e-01
4.66649741e-01 8.09813917e-01 1.04183328e+00 1.23652017e+00
2.47042075e-01 8.33165348e-01 1.48228574e+00 -5.34163952e-01
5.74632347e-01 5.08697212e-01 2.96209753e-01 1.93870470e-01
7.37435758e-01 7.05788434e-01 -3.19604069e-01 -4.15398747e-01
6.20132506e-01 -3.18448156e-01 8.37774575e-02 -2.03625619e-01
-7.98671603e-01 8.37774217e-01 -2.99918056e-01 -1.20778143e-01
-6.10949636e-01 1.14441462e-01 3.94105464e-01 3.81381691e-01
3.72434050e-01 -1.03264354e-01 -6.09581113e-01 -4.50188398e-01
-9.10424829e-01 5.30016541e-01 8.01055610e-01 1.01352978e+00
1.73544362e-01 4.01916057e-01 -9.66787219e-01 6.66856229e-01
3.16464663e-01 8.80331635e-01 3.41605932e-01 -7.74716675e-01
4.73614037e-01 -4.41128202e-02 7.47964263e-01 -3.48839194e-01
-2.93406576e-01 -8.22068870e-01 -4.09735054e-01 3.96128111e-02
6.94314182e-01 -4.08987373e-01 -5.72865248e-01 1.44633937e+00
4.65009242e-01 2.59797722e-01 -2.74565816e-01 7.43059397e-01
2.61650141e-02 4.18923825e-01 -1.16961852e-01 -9.19595301e-01
1.37340260e+00 -5.76591909e-01 -7.84063816e-01 1.59629837e-01
8.27550739e-02 -5.25380075e-01 8.75205874e-01 8.99872899e-01
-1.21850026e+00 -3.17703009e-01 -9.04645979e-01 1.31787789e+00
-1.37212202e-01 -2.03628585e-01 7.41726160e-01 1.82188845e+00
-5.84797323e-01 2.84806401e-01 -4.05689150e-01 -3.19172367e-02
4.87467200e-01 3.15370470e-01 4.96138215e-01 1.39010102e-01
-1.24068320e+00 7.87253499e-01 3.00932825e-01 2.29558170e-01
-1.02301824e+00 -5.35080731e-01 -3.51405367e-02 -9.61870551e-02
8.42743814e-01 -4.20089394e-01 1.32444251e+00 -9.69240010e-01
-2.08430433e+00 1.32240593e-01 -1.27422124e-01 -8.88605356e-01
9.63658869e-01 -2.55696308e-02 -7.25165784e-01 -1.27514720e-01
-1.09993108e-01 7.67322108e-02 9.58985388e-01 -1.05348516e+00
-9.97665346e-01 -2.32169330e-01 -4.41286527e-02 -1.82272017e-01
-8.72446522e-02 1.95735648e-01 5.79982065e-02 -5.47502637e-01
-2.20503807e-01 -7.91507363e-01 -9.59115773e-02 -8.31394672e-01
-5.68024695e-01 -4.08945918e-01 5.78408599e-01 -3.31913650e-01
1.46465778e+00 -1.48884153e+00 -6.41839206e-01 6.47224724e-01
-6.32725596e-01 2.31754526e-01 1.60441488e-01 2.21609354e-01
1.71359554e-02 1.05964355e-01 1.41511530e-01 4.86363918e-01
3.47083479e-01 1.31857708e-01 -5.93564332e-01 4.50013399e-01
1.83624048e-02 6.71528995e-01 -5.38541913e-01 -3.99042755e-01
1.94043741e-01 -3.62532228e-01 -4.59426701e-01 -4.70029823e-02
-3.23084414e-01 2.44558424e-01 -5.85078895e-01 9.98726904e-01
6.85799241e-01 -4.59201494e-03 8.89151469e-02 -8.76159519e-02
-9.07362401e-02 -1.01424359e-01 -1.43201876e+00 3.82074416e-01
-1.95703164e-01 1.08550921e-01 -2.27653161e-01 -1.29856038e+00
6.98262870e-01 2.47764200e-01 3.32194120e-01 -8.53141129e-01
3.65019053e-01 1.10532232e-01 9.14592296e-02 -6.07241094e-01
1.41208425e-01 -2.15859741e-01 1.66208953e-01 5.95865011e-01
-3.17718744e-01 1.09364510e-01 3.31887603e-01 -3.14696759e-01
8.11996400e-01 1.00195564e-01 1.22398421e-01 -3.52770180e-01
4.02522027e-01 -3.06440860e-01 6.65542543e-01 1.32175314e+00
-4.11493927e-01 -1.98931068e-01 7.26168573e-01 -3.89221013e-01
-6.86866581e-01 -1.03156126e+00 -5.05888581e-01 1.29873788e+00
1.60614267e-01 5.51618636e-01 -4.72164065e-01 -6.55222237e-01
2.29520380e-01 1.16454065e+00 -5.35965264e-01 2.80819740e-02
-1.95845321e-01 -1.27245545e+00 4.06980366e-01 2.21844703e-01
5.81558287e-01 -6.94439411e-01 -8.33158195e-01 2.60138243e-01
1.38930958e-02 -9.44199681e-01 1.41606089e-02 3.05487096e-01
-7.97203124e-01 -9.50725675e-01 -5.97787857e-01 3.00915748e-01
8.02205577e-02 2.14694574e-01 9.92887676e-01 -2.84363419e-01
1.96964340e-03 4.49467093e-01 -5.04567862e-01 -1.04970551e+00
-1.98571965e-01 -3.85459870e-01 7.46516213e-02 2.64041543e-01
4.15166318e-01 -4.06733125e-01 -6.29057050e-01 1.01941276e+00
-4.84489053e-01 -4.61313456e-01 8.39622140e-01 9.11044061e-01
3.72166008e-01 2.25857973e-01 8.97836804e-01 -8.09972048e-01
6.96371853e-01 -5.38383842e-01 -1.22157991e+00 5.46279013e-01
-8.97148669e-01 -1.69205070e-01 9.37787220e-02 -7.81741917e-01
-1.25264823e+00 -5.89758635e-01 9.16565880e-02 -2.77610600e-01
9.01690722e-02 2.77473122e-01 1.32349670e-01 8.49201381e-02
5.73635340e-01 6.54327199e-02 -8.00887719e-02 -1.58586547e-01
3.22473198e-02 9.29158926e-01 -3.81575041e-02 -1.11086369e+00
5.69673657e-01 3.15884918e-01 7.45411068e-02 -4.78200614e-01
-9.66142535e-01 -1.85582377e-02 -3.12746502e-02 -6.00750566e-01
3.18769783e-01 -5.14857709e-01 -1.23987603e+00 1.77644100e-02
-5.84336579e-01 -2.94585854e-01 -3.78976375e-01 8.93127978e-01
-7.73273408e-01 1.92506567e-01 -1.66708499e-01 -1.80891657e+00
-1.24885261e-01 -1.19083011e+00 5.15089869e-01 2.62099922e-01
-1.32279925e-03 -6.12994373e-01 -1.82697028e-01 6.40234172e-01
4.00882542e-01 -1.63042061e-02 8.63331735e-01 -9.64014590e-01
-5.21317601e-01 -4.26380157e-01 -9.89149734e-02 3.21700215e-01
-3.51929843e-01 2.75735468e-01 -7.02067912e-01 -4.17691916e-01
-7.96865951e-03 -1.74611032e-01 2.33484030e-01 9.80005443e-01
1.30597293e+00 -6.56865537e-01 -1.17482156e-01 1.08746909e-01
1.03111935e+00 8.66265476e-01 6.87326252e-01 6.35436118e-01
-4.14979428e-01 5.34168661e-01 1.02184010e+00 8.70050788e-01
-1.29846498e-01 6.15022659e-01 6.58929825e-01 3.84254485e-01
3.74535918e-01 1.06682338e-01 4.59263235e-01 2.78003663e-01
-2.86043823e-01 -3.35174501e-01 -5.67591429e-01 3.16380620e-01
-1.69013393e+00 -1.19233155e+00 -1.54234156e-01 2.85280991e+00
7.12700307e-01 9.40462649e-01 8.96363497e-01 2.67179400e-01
9.29417789e-01 -2.36911550e-01 -8.07480097e-01 -5.08963406e-01
9.84875020e-03 3.37799668e-01 9.94669378e-01 6.92156926e-02
-7.40073144e-01 3.69071841e-01 6.86157513e+00 1.43952322e+00
-8.18465412e-01 3.28075975e-01 1.08334899e+00 -4.63958561e-01
-8.25319216e-02 1.16220169e-01 -1.16558552e+00 6.78933144e-01
1.39324594e+00 -1.32415771e-01 3.94729465e-01 8.38082910e-01
4.06915337e-01 -4.39578027e-01 -7.85954237e-01 6.76269054e-01
-4.49772537e-01 -1.16142046e+00 5.04840072e-03 1.14637323e-01
6.91381991e-01 -1.38829172e-01 1.56940311e-01 6.65471971e-01
8.36600900e-01 -7.10740805e-01 1.08286667e+00 5.96591890e-01
3.75607193e-01 -1.07953918e+00 1.17052495e+00 3.32599849e-01
-3.69191587e-01 -6.09363496e-01 -4.32255715e-01 -1.22633921e-02
9.24259722e-02 8.65731061e-01 -5.59289157e-01 8.57444823e-01
8.89275670e-01 -1.21872745e-01 -2.55817562e-01 1.35587382e+00
-4.21100017e-03 1.15413070e+00 -4.87265706e-01 -6.29577339e-01
2.63804048e-01 -3.54893476e-01 7.34231770e-01 1.00654793e+00
5.17444193e-01 -1.24490246e-01 -2.85545867e-02 7.48776317e-01
5.11269748e-01 -1.27145857e-01 -1.28739490e-03 2.23042041e-01
8.58473122e-01 1.10701799e+00 -7.85130739e-01 -1.62079811e-01
-1.77059546e-01 -1.15538687e-01 -4.42961156e-01 4.83209491e-01
-1.44289100e+00 -1.03619635e-01 2.60929853e-01 1.31846368e-01
5.20123184e-01 4.58295256e-01 -4.67886209e-01 -3.60128999e-01
-1.82405755e-01 -9.58114922e-01 6.62900090e-01 -7.25971162e-01
-1.26069236e+00 2.61179239e-01 8.04277956e-01 -1.23777735e+00
-2.14404702e-01 -6.16522551e-01 -4.76558030e-01 5.04075646e-01
-1.21563447e+00 -7.32766092e-01 1.79376602e-01 5.34636855e-01
5.45642018e-01 -4.86848027e-01 1.73808634e-01 5.80628477e-02
-9.83413398e-01 9.09063995e-01 2.69024700e-01 -4.45718795e-01
2.40554988e-01 -7.10694432e-01 -1.34105325e-01 6.26323342e-01
-1.60995007e-01 3.48709822e-01 1.09768939e+00 -5.82408786e-01
-1.23635328e+00 -7.31629491e-01 -1.02086902e-01 -2.64862657e-01
8.16792667e-01 5.00563681e-02 -3.65363985e-01 3.98355156e-01
-5.47014773e-02 -2.94003040e-01 5.99353313e-01 2.75442690e-01
-1.22921079e-01 -7.28209019e-01 -1.37489164e+00 5.97036779e-01
6.30335867e-01 2.13946074e-01 -1.75480947e-01 5.83545387e-01
4.62482840e-01 -1.12173960e-01 -8.36875975e-01 6.04036152e-01
9.29042876e-01 -1.34664166e+00 6.90892994e-01 -6.57350481e-01
-3.63117903e-01 1.66356955e-02 -2.32477620e-01 -8.55693400e-01
-5.42646609e-02 -1.07989872e+00 -1.11735694e-01 1.29288399e+00
5.48497617e-01 -1.05923510e+00 8.08089793e-01 4.79960144e-01
9.37013403e-02 -8.46038520e-01 -1.63280487e+00 -1.30149686e+00
-1.10926740e-01 -8.52662504e-01 8.52203846e-01 4.05166030e-01
-2.20416576e-01 -1.16251647e-01 -5.87250471e-01 -1.62293185e-02
8.20350468e-01 5.82683831e-02 8.50227118e-01 -1.11748731e+00
-6.87268674e-01 -6.20005131e-01 -7.89308995e-02 -9.33182359e-01
-2.11628377e-01 -9.73082781e-02 -1.83831111e-01 -7.50527322e-01
2.63260663e-01 -3.78284693e-01 -7.02121913e-01 3.00141215e-01
1.87434793e-01 -7.60992840e-02 -3.02658260e-01 -3.20135243e-02
-6.16845489e-01 3.16428244e-01 1.05240202e+00 -3.65534611e-02
-6.97573945e-02 1.02614737e+00 -7.30558515e-01 4.58764195e-01
7.59229302e-01 -6.10921443e-01 -4.73474771e-01 3.49128872e-01
4.34104532e-01 5.89739025e-01 3.08974296e-01 -7.07587183e-01
-1.59149598e-02 -6.97078764e-01 2.82869250e-01 -9.48233783e-01
6.78620040e-02 -8.90944302e-01 3.57419580e-01 6.49832129e-01
-4.33772691e-02 -1.58792406e-01 2.02019632e-01 8.93428445e-01
2.80120760e-01 -6.25551641e-01 6.83824480e-01 -3.26475129e-02
-2.25681037e-01 6.86497474e-03 -4.94005531e-01 -1.42795190e-01
1.30787158e+00 -7.11259604e-01 -3.39170635e-01 -6.03979647e-01
-4.14546013e-01 3.28179985e-01 -4.64865237e-01 1.59699529e-01
1.14362769e-01 -1.15737975e+00 -6.49120986e-01 7.82674327e-02
-2.77366102e-01 -5.38769901e-01 3.46146405e-01 1.36560059e+00
-8.67645368e-02 6.07454956e-01 -3.92689258e-02 -7.93935776e-01
-9.34694827e-01 4.98204529e-01 2.14218110e-01 -4.12346870e-01
-6.25943914e-02 8.16746771e-01 -3.33250225e-01 1.17188141e-01
5.61289132e-01 -1.05804719e-01 6.83898926e-02 7.30476156e-02
3.41674179e-01 9.37943101e-01 2.62841702e-01 -1.69332195e-02
-1.15991645e-01 2.50358909e-01 -2.35159174e-02 -2.47620106e-01
1.10039592e+00 -1.50514215e-01 2.54058894e-02 5.32273114e-01
1.91557854e-01 -1.08984850e-01 -1.27794683e+00 -2.39638073e-04
2.68276989e-01 -6.71386659e-01 1.43914551e-01 -1.12986565e+00
-7.87673831e-01 2.89725035e-01 8.55032027e-01 8.00055206e-01
1.03210425e+00 -3.52275938e-01 3.79564255e-01 3.22666407e-01
6.86417878e-01 -1.26169407e+00 -2.80909836e-01 1.69155195e-01
6.54199421e-01 -9.72960413e-01 3.54998767e-01 -3.70385759e-02
-6.52864456e-01 8.96306574e-01 2.67774224e-01 2.33542010e-01
7.20608413e-01 1.04331955e-01 -4.42445755e-01 1.84370011e-01
-9.97125328e-01 -2.72958457e-01 2.09275022e-01 4.77382153e-01
-1.01056159e-01 1.45527631e-01 -6.87823474e-01 9.56131041e-01
-2.43321180e-01 1.90663040e-02 2.24024802e-01 6.45179331e-01
-4.89802361e-01 -8.12229812e-01 -7.27785945e-01 8.64712596e-01
-6.72087669e-01 3.40356499e-01 -1.72722731e-02 1.25571620e+00
-1.33231208e-01 1.26192665e+00 -2.88535580e-02 -3.92257363e-01
4.34950382e-01 4.11032774e-02 5.36086738e-01 2.21378550e-01
-4.89627212e-01 3.81012231e-01 3.32597226e-01 -4.03513342e-01
-4.47673500e-01 -8.28682065e-01 -3.69278789e-01 -4.59940791e-01
-9.47992027e-01 6.33373439e-01 7.80261397e-01 1.02673972e+00
-6.00142106e-02 4.73193765e-01 8.48738134e-01 -6.48892641e-01
-1.35942459e+00 -1.14311814e+00 -9.08897758e-01 -1.34721756e-01
-7.42687061e-02 -1.03956532e+00 -5.90375125e-01 -5.64907253e-01] | [4.530822277069092, 3.2678725719451904] |
917c071d-8dce-439b-a574-1a8f3c07ff1e | drotrack-high-speed-drone-based-object | 2005.00828 | null | https://arxiv.org/abs/2005.00828v1 | https://arxiv.org/pdf/2005.00828v1.pdf | DroTrack: High-speed Drone-based Object Tracking Under Uncertainty | We present DroTrack, a high-speed visual single-object tracking framework for drone-captured video sequences. Most of the existing object tracking methods are designed to tackle well-known challenges, such as occlusion and cluttered backgrounds. The complex motion of drones, i.e., multiple degrees of freedom in three-dimensional space, causes high uncertainty. The uncertainty problem leads to inaccurate location predictions and fuzziness in scale estimations. DroTrack solves such issues by discovering the dependency between object representation and motion geometry. We implement an effective object segmentation based on Fuzzy C Means (FCM). We incorporate the spatial information into the membership function to cluster the most discriminative segments. We then enhance the object segmentation by using a pre-trained Convolution Neural Network (CNN) model. DroTrack also leverages the geometrical angular motion to estimate a reliable object scale. We discuss the experimental results and performance evaluation using two datasets of 51,462 drone-captured frames. The combination of the FCM segmentation and the angular scaling increased DroTrack precision by up to $9\%$ and decreased the centre location error by $162$ pixels on average. DroTrack outperforms all the high-speed trackers and achieves comparable results in comparison to deep learning trackers. DroTrack offers high frame rates up to 1000 frame per second (fps) with the best location precision, more than a set of state-of-the-art real-time trackers. | ['Flora Salim', 'Ali Hamdi', 'Du Yong Kim'] | 2020-05-02 | null | null | null | null | ['drone-based-object-tracking'] | ['computer-vision'] | [-4.08146948e-01 -6.33727193e-01 -1.06029108e-01 -1.67921945e-01
-4.94324803e-01 -8.86547625e-01 3.32933992e-01 -2.48869091e-01
-6.85964942e-01 4.82122183e-01 -3.94466072e-01 1.42831802e-01
1.26984492e-01 -5.13317227e-01 -8.97778749e-01 -7.31639206e-01
-2.91033477e-01 3.30034673e-01 7.79919744e-01 8.80400613e-02
3.36331464e-02 6.42501771e-01 -1.71667206e+00 -2.54963130e-01
8.09418440e-01 1.56688070e+00 2.92306766e-02 9.09307241e-01
3.34104270e-01 2.67300725e-01 -8.77952158e-01 -3.72601718e-01
6.28655612e-01 2.47239664e-01 -1.56758670e-02 -8.51143301e-02
1.32210910e+00 -5.36344051e-01 -4.26414371e-01 1.18129492e+00
3.39194953e-01 2.99041063e-01 3.23950410e-01 -1.59302306e+00
-3.08537513e-01 -9.90981460e-02 -6.35150135e-01 7.12744415e-01
4.56398092e-02 4.96083498e-01 3.83902311e-01 -7.43650794e-01
2.50883311e-01 1.18561113e+00 1.17740893e+00 4.33688641e-01
-8.36760521e-01 -1.11483014e+00 1.88221082e-01 -1.19484648e-01
-1.50188029e+00 -2.34034240e-01 3.81554514e-01 -7.58389235e-01
7.70213187e-01 1.04370236e-01 8.54161859e-01 7.12949574e-01
3.02814245e-01 5.25218904e-01 5.94064713e-01 2.61582404e-01
-7.41559565e-02 -2.02651978e-01 -1.19048476e-01 9.54827785e-01
8.12768638e-01 7.40393341e-01 -4.12348241e-01 1.15026549e-01
1.01549125e+00 1.28199413e-01 -4.07150030e-01 -5.07265687e-01
-1.51460028e+00 5.25505066e-01 8.47168505e-01 -1.34929642e-01
1.15645610e-01 6.89308882e-01 3.06015730e-01 -2.42345467e-01
3.09542358e-01 4.74439293e-01 -4.58027869e-01 -2.39358932e-01
-1.15520287e+00 3.83701056e-01 3.54690254e-01 1.36460936e+00
5.17463028e-01 4.29462820e-01 -3.01333249e-01 7.12008178e-02
6.26389205e-01 1.14731181e+00 7.22055435e-02 -1.01369464e+00
4.92246836e-01 4.82661933e-01 5.96926868e-01 -1.18780792e+00
-6.83175147e-01 -5.49045444e-01 -4.49863970e-01 4.31126028e-01
6.45259559e-01 -4.70220536e-01 -9.79102790e-01 1.48107517e+00
6.79061174e-01 5.54361403e-01 -3.95388097e-01 1.57765472e+00
9.49482322e-01 5.36037087e-01 2.00170264e-01 5.03480583e-02
1.27916110e+00 -9.73985374e-01 -7.75990069e-01 -2.01356545e-01
3.78577113e-01 -7.59034872e-01 4.31276798e-01 2.83996999e-01
-8.14202964e-01 -9.29281771e-01 -1.11882412e+00 2.14587256e-01
-3.41291577e-01 6.25309706e-01 5.87643623e-01 8.91328990e-01
-8.22538137e-01 3.39379042e-01 -1.20284069e+00 -1.44541599e-02
6.10716343e-01 6.40377879e-01 -1.18725458e-02 3.23841423e-01
-7.56969631e-01 6.93278611e-01 4.46652949e-01 1.41365647e-01
-8.33372772e-01 -1.10350680e+00 -1.01219296e+00 -1.89893201e-01
6.47411644e-01 -6.30905390e-01 1.15707755e+00 -6.41666651e-01
-1.26721501e+00 5.03495991e-01 1.34090513e-01 -6.57278061e-01
6.47000372e-01 -7.19589293e-01 -6.29485905e-01 1.45233795e-01
1.19911246e-01 1.02045345e+00 9.63276505e-01 -1.05355608e+00
-1.20123613e+00 -2.14644819e-01 -1.29605383e-01 -3.02966568e-03
-7.59273469e-02 -1.15279607e-01 -8.31249595e-01 -8.38161647e-01
-9.05876607e-02 -1.15558147e+00 1.25658453e-01 6.07116818e-01
3.68521363e-02 -1.11689165e-01 1.24777102e+00 -4.37417507e-01
1.31028354e+00 -1.96462548e+00 -1.62856251e-01 -2.60082632e-01
3.56686026e-01 6.59919143e-01 2.55824655e-01 -3.68605494e-01
4.26533520e-01 -2.36915454e-01 1.35147020e-01 -3.56393367e-01
5.22551462e-02 -1.19360998e-01 -2.82718003e-01 9.04245973e-01
1.45600557e-01 1.02562487e+00 -9.96348321e-01 -6.40266240e-01
6.65123165e-01 7.59556055e-01 -4.81308550e-01 8.88620988e-02
-2.96083093e-01 3.62715751e-01 -2.85289288e-01 9.71412599e-01
9.81524646e-01 -1.87594309e-01 -3.85527581e-01 -5.07879019e-01
-5.53796947e-01 -6.42363876e-02 -1.21746743e+00 1.60690546e+00
1.14101253e-01 1.03426945e+00 -1.10817922e-03 -3.48263711e-01
9.20741379e-01 5.90593629e-02 6.95040762e-01 -3.65749836e-01
4.12927806e-01 5.95435314e-02 -2.50577271e-01 -3.32831264e-01
7.79594481e-01 1.81586593e-01 -8.25212374e-02 -1.57737315e-01
1.27369657e-01 -1.36876479e-02 2.82196313e-01 -1.21262453e-01
6.09037757e-01 4.62947458e-01 -1.57441959e-01 -3.29344034e-01
2.27873251e-01 2.87262857e-01 9.10595000e-01 6.17317736e-01
-6.81770265e-01 5.87309361e-01 -2.73736924e-01 -7.78916299e-01
-6.44818068e-01 -1.03590953e+00 -1.41870752e-01 8.53380442e-01
8.88394654e-01 -3.15491796e-01 -6.39583111e-01 -6.18060112e-01
3.11031103e-01 1.50734097e-01 -4.96872932e-01 -6.23342991e-02
-7.41328418e-01 -5.74399173e-01 6.37164950e-01 8.66626382e-01
9.32163954e-01 -5.47343552e-01 -1.22944832e+00 1.60371289e-01
-5.66919632e-02 -1.50962353e+00 -7.82557726e-01 -2.06892222e-01
-7.76612103e-01 -1.18928254e+00 -4.92567569e-01 -4.22693402e-01
4.47958469e-01 5.76753914e-01 9.43392038e-01 8.74087512e-02
-4.63334978e-01 3.34886104e-01 -9.76822823e-02 -4.24155116e-01
2.99353957e-01 -2.41237968e-01 4.33175981e-01 -2.04812691e-01
4.79483038e-01 2.37895131e-01 -9.40593779e-01 6.73804522e-01
-5.15895307e-01 -1.87387452e-01 3.77536684e-01 4.05032754e-01
5.46484470e-01 2.13607341e-01 1.58784445e-02 1.22709209e-02
-2.22466081e-01 -9.57726687e-02 -1.54189312e+00 2.84708254e-02
-2.77741402e-01 -2.45365784e-01 3.52068990e-01 -8.24279785e-01
-8.73909771e-01 4.01112080e-01 3.19936216e-01 -1.01558793e+00
-1.25607178e-01 -2.13209406e-01 1.11198314e-02 -7.08056152e-01
5.66127896e-01 -1.43247291e-01 -2.11966723e-01 -2.14454308e-01
2.26729050e-01 2.50963688e-01 9.50351298e-01 -3.16339731e-01
1.10054481e+00 7.24929571e-01 -5.19899465e-02 -7.30389178e-01
-7.98770607e-01 -6.79869831e-01 -5.62745154e-01 -7.40140021e-01
1.13460755e+00 -1.28766525e+00 -1.11382937e+00 4.41988140e-01
-1.11013281e+00 -3.12392682e-01 -1.80108026e-02 8.68757129e-01
-3.05894971e-01 2.24852249e-01 -3.78711611e-01 -7.64361441e-01
-3.34210515e-01 -1.27432442e+00 1.39339209e+00 8.05382669e-01
2.22833768e-01 -7.55868554e-01 -6.16523214e-02 2.56231368e-01
3.17087352e-01 5.90327799e-01 -2.92121857e-01 -1.38168290e-01
-1.25448298e+00 -1.84420794e-01 -4.09309775e-01 -1.13352090e-01
3.94659340e-02 1.97114021e-01 -7.81109750e-01 -6.19005442e-01
-3.45942140e-01 2.61329096e-02 9.03014839e-01 7.64722705e-01
7.83840477e-01 -2.93106921e-02 -8.03406835e-01 1.14301050e+00
1.37251079e+00 3.24158221e-01 2.19256356e-01 2.58165091e-01
9.89044309e-01 -1.52467057e-01 1.17374623e+00 4.97784466e-01
3.68897587e-01 9.79169846e-01 6.48972452e-01 2.00706087e-02
-1.77860245e-01 -1.17299020e-01 3.32652211e-01 1.42874405e-01
-2.57454515e-01 -2.79886067e-01 -8.09588909e-01 4.52557743e-01
-1.75367451e+00 -8.54838014e-01 -2.29175627e-01 2.30723524e+00
4.71790850e-01 2.23974034e-01 3.28495741e-01 -3.36810976e-01
9.25704896e-01 3.91387157e-02 -5.52377999e-01 3.24704170e-01
9.66163538e-03 -2.22476557e-01 1.19400549e+00 3.56868744e-01
-1.81783783e+00 1.12439859e+00 5.72606611e+00 6.47295058e-01
-1.35370493e+00 -1.30135596e-01 2.30622012e-02 -4.21829909e-01
6.10589802e-01 -1.99242577e-01 -1.59727085e+00 8.01246405e-01
6.56409562e-01 2.01741323e-01 2.47085154e-01 1.07245207e+00
4.02250141e-03 -1.39841765e-01 -6.74662113e-01 1.19900668e+00
2.48431757e-01 -1.54456365e+00 -4.07185495e-01 7.16406032e-02
8.63219917e-01 3.08034688e-01 1.51820555e-01 2.30877995e-01
2.74551690e-01 -8.04761469e-01 1.02042365e+00 5.89323938e-01
8.50811541e-01 -7.61874855e-01 6.01966262e-01 2.15590313e-01
-1.95728242e+00 -1.80552185e-01 -4.46960598e-01 1.28325820e-01
2.42311954e-01 2.33906612e-01 -6.14953578e-01 2.60300964e-01
1.26607239e+00 8.08781922e-01 -7.02695072e-01 1.68492138e+00
1.89989023e-02 2.74536401e-01 -7.90286601e-01 6.18937016e-02
2.80325115e-01 4.97064739e-02 7.95783937e-01 1.22578108e+00
3.58499616e-01 1.78048313e-01 4.66455132e-01 9.02244031e-01
5.80027252e-02 -6.54517412e-01 -2.60419905e-01 2.44541615e-02
5.71990430e-01 1.39526260e+00 -9.36943114e-01 -4.81996447e-01
-3.05058002e-01 5.49690366e-01 -1.03878111e-01 1.94710955e-01
-1.52744925e+00 -3.06529850e-01 9.91055131e-01 1.24733582e-01
8.14072251e-01 -6.11537516e-01 1.19441235e-02 -1.12414074e+00
-6.52204901e-02 -3.78582627e-01 2.81253725e-01 -6.35391593e-01
-9.30709779e-01 6.49188519e-01 5.32400757e-02 -1.66625786e+00
1.03083789e-01 -7.87862897e-01 -2.91855842e-01 4.72876191e-01
-1.56417513e+00 -1.19022238e+00 -8.50625813e-01 3.88554424e-01
5.35404503e-01 1.86172742e-02 2.07384452e-01 6.59576416e-01
-6.56276584e-01 6.12971127e-01 -1.98416159e-01 4.30794537e-01
8.69910479e-01 -1.18637013e+00 5.15644968e-01 1.11511481e+00
3.47776823e-02 3.93756986e-01 6.05391741e-01 -9.70106006e-01
-1.55658913e+00 -1.53739405e+00 3.46637338e-01 -8.28150511e-01
5.69383621e-01 -3.03916246e-01 -6.39243543e-01 7.31501460e-01
-1.80181131e-01 8.23600292e-01 3.04386348e-01 -6.17368102e-01
-2.14342959e-02 -1.45490646e-01 -9.89484131e-01 2.09906414e-01
1.14678764e+00 7.39704957e-03 -4.40797746e-01 -7.40664592e-03
8.73546720e-01 -1.03259051e+00 -9.60116625e-01 6.28197193e-01
6.20898843e-01 -9.28269327e-01 1.22747290e+00 -1.75118372e-01
-4.49729115e-01 -1.09223378e+00 6.41558468e-02 -9.22835231e-01
-3.94493163e-01 -7.52082229e-01 -5.07985115e-01 9.16017056e-01
-2.05874026e-01 -2.84136951e-01 8.87725830e-01 6.51859105e-01
-3.13658178e-01 -5.01628280e-01 -1.01990879e+00 -9.91621315e-01
-4.94209737e-01 -4.31729645e-01 6.37856603e-01 5.26237667e-01
-6.08831227e-01 -2.35089704e-01 -3.26319903e-01 6.72783256e-01
1.03397977e+00 3.22760463e-01 9.89929199e-01 -1.36800015e+00
1.21484362e-01 -3.62217426e-01 -7.57153332e-01 -1.51268232e+00
-1.84744209e-01 -2.66372383e-01 2.72420496e-01 -9.87802267e-01
-4.78680372e-01 -5.98188519e-01 -1.22204795e-03 1.73943445e-01
-2.16589049e-01 7.09399402e-01 4.15669024e-01 4.08393741e-02
-1.08530903e+00 4.29733366e-01 1.09849250e+00 -1.03990883e-01
-2.26977199e-01 1.73887774e-01 -1.39873907e-01 8.29418004e-01
6.01758122e-01 -4.38128084e-01 7.35562528e-03 -5.51639676e-01
-4.06303927e-02 -5.54693043e-02 7.71886289e-01 -1.68650186e+00
5.99997103e-01 -3.53794433e-02 9.12403166e-01 -1.10709679e+00
5.36328673e-01 -1.10145140e+00 1.33407339e-01 5.13197422e-01
2.16217071e-01 1.09889559e-01 7.69326806e-01 6.21660113e-01
-2.73917448e-02 3.44997942e-01 8.29409957e-01 2.22361550e-01
-9.74301934e-01 7.00740516e-01 -1.39029354e-01 2.56693631e-01
1.14330685e+00 -6.16210938e-01 -4.97542977e-01 1.40421271e-01
-2.23250210e-01 4.27988291e-01 5.82655609e-01 7.66376376e-01
5.39929092e-01 -1.47303009e+00 -1.49937883e-01 3.09926331e-01
5.06518967e-02 3.10223788e-01 9.21541899e-02 1.03209591e+00
-6.44635558e-01 7.28702724e-01 -2.80807346e-01 -1.13527393e+00
-1.23728418e+00 5.32029152e-01 5.73939979e-01 3.07255685e-01
-6.20881796e-01 9.33251023e-01 9.29030180e-02 8.62813517e-02
3.93705428e-01 -9.00779784e-01 -2.23821178e-02 -4.41680141e-02
6.37504637e-01 5.97453773e-01 -2.02923506e-01 -1.01127017e+00
-5.95988870e-01 1.19956505e+00 3.15149546e-01 3.18887860e-01
8.25925112e-01 -1.12763330e-01 5.33119857e-01 -3.35549228e-02
8.58193815e-01 -2.42632300e-01 -2.26932549e+00 -2.61454494e-03
-2.44389653e-01 -8.31407726e-01 2.87729412e-01 -5.88051379e-01
-1.46650994e+00 6.39200628e-01 9.81322944e-01 -7.86746740e-02
7.27338910e-01 -2.44251937e-01 9.30409133e-01 2.68695712e-01
4.35046375e-01 -1.02204406e+00 1.31123722e-01 3.69274348e-01
3.30339491e-01 -1.69395280e+00 2.49397516e-01 -4.02061492e-01
-5.28236210e-01 1.12113142e+00 1.08966100e+00 -2.70982951e-01
4.76619303e-01 5.28651953e-01 1.52218208e-01 -2.79781133e-01
-2.21890718e-01 -2.45842993e-01 8.19209635e-01 6.89992964e-01
-1.59301925e-02 2.47314200e-02 2.29701072e-01 4.51742083e-01
-2.04220876e-01 1.02096032e-02 1.59536097e-02 8.97117794e-01
-6.73680305e-01 -2.51159877e-01 -8.37532163e-01 1.30380347e-01
-2.67663062e-01 3.03643674e-01 -7.77687225e-03 9.74877179e-01
6.11258030e-01 8.58339727e-01 5.43904960e-01 -3.68350327e-01
9.95793268e-02 -5.13643205e-01 4.30333853e-01 -1.40359834e-01
-6.32695794e-01 2.16007724e-01 -3.02361339e-01 -8.61450672e-01
-6.02370620e-01 -5.10243535e-01 -1.52385259e+00 -3.15795600e-01
-7.10117340e-01 -1.82524681e-01 7.86350310e-01 7.56386340e-01
3.59896034e-01 5.00976980e-01 3.79014499e-02 -1.30349743e+00
-9.60581899e-02 -5.55488169e-01 -2.05781415e-01 4.22278792e-02
6.80542827e-01 -1.34432530e+00 -2.93765634e-01 2.75450885e-01] | [6.470572471618652, -2.1865131855010986] |
b27a655d-6836-4c28-a71a-8022d3fdb48a | sentiment-analysis-for-emotional-speech | null | null | https://aclanthology.org/2020.coling-main.440 | https://aclanthology.org/2020.coling-main.440.pdf | Sentiment Analysis for Emotional Speech Synthesis in a News Dialogue System | As smart speakers and conversational robots become ubiquitous, the demand for expressive speech synthesis has increased. In this paper, to control the emotional parameters of the speech synthesis according to certain dialogue contents, we construct a news dataset with emotion labels ({``}positive,{''} {``}negative,{''} or {``}neutral{''}) annotated for each sentence. We then propose a method to identify emotion labels using a model combining BERT and BiLSTM-CRF, and evaluate its effectiveness using the constructed dataset. The results showed that the classification model performance can be efficiently improved by preferentially annotating news articles with low confidence in the human-in-the-loop machine learning framework. | ['Tetsunori Kobayashi', 'Yoichi Matsuyama', 'Ryota Ando', 'Hiroaki Takatsu'] | 2020-12-01 | null | null | null | coling-2020-8 | ['emotional-speech-synthesis', 'expressive-speech-synthesis'] | ['speech', 'speech'] | [-1.65324569e-01 7.55803406e-01 -1.65880978e-01 -7.22152710e-01
-5.05836189e-01 -4.24407661e-01 5.60972154e-01 -2.04562426e-01
-3.31006348e-01 1.02406764e+00 4.25556332e-01 -8.25323686e-02
4.98199373e-01 -4.50298667e-01 -3.70864719e-01 -6.15858257e-01
1.88464269e-01 5.53976119e-01 -1.08821794e-01 -3.40795457e-01
4.61180955e-02 1.19776934e-01 -1.55491304e+00 6.39168322e-01
5.54013252e-01 1.27798235e+00 3.10148537e-01 6.73656166e-01
-3.37725341e-01 1.48752797e+00 -1.02261972e+00 -6.54013038e-01
-3.51196766e-01 -5.54800868e-01 -1.02359581e+00 -2.79703457e-02
-6.50125742e-01 -1.94813415e-01 2.30270326e-02 1.19679272e+00
5.42443395e-01 4.40687835e-01 6.07595980e-01 -1.33634377e+00
-3.42634678e-01 9.42659974e-01 -8.12975094e-02 -2.29764163e-01
5.12037456e-01 -6.44150004e-02 1.03893793e+00 -6.80762947e-01
7.18144417e-01 1.26096845e+00 3.40939224e-01 9.22092021e-01
-8.65849853e-01 -5.34840882e-01 4.77164149e-01 1.55375168e-01
-9.86588359e-01 -6.91919208e-01 1.09313774e+00 -3.81035000e-01
1.12996781e+00 2.51378536e-01 5.38383186e-01 1.61564898e+00
-3.58373523e-02 1.09764218e+00 1.09279037e+00 -5.78402519e-01
2.77755678e-01 6.25288367e-01 -1.32043377e-01 5.18250048e-01
-8.58606875e-01 -1.85222328e-01 -3.81482780e-01 -1.13951758e-01
2.32189640e-01 -5.39069057e-01 -5.78316301e-02 1.80268571e-01
-1.16190195e+00 9.69304800e-01 -5.98947741e-02 4.49538141e-01
-6.50027096e-01 2.56064236e-02 8.79207313e-01 4.25963372e-01
8.81942630e-01 4.28496271e-01 -7.43139029e-01 -6.21206760e-01
-9.01127756e-02 5.76235093e-02 1.19241011e+00 1.09173238e+00
4.25769359e-01 4.51987842e-03 -3.87529522e-01 1.47481000e+00
3.80033404e-01 3.86215866e-01 5.33253253e-01 -1.31454730e+00
1.72756448e-01 2.59033561e-01 5.55790424e-01 -6.48358941e-01
-7.92564690e-01 -1.14691809e-01 -8.02278697e-01 -2.32628688e-01
6.04473948e-02 -7.04340160e-01 -4.03522849e-01 1.96197593e+00
3.94111723e-01 -1.96223095e-01 6.65188789e-01 7.27101624e-01
9.67279375e-01 1.10134017e+00 4.13740069e-01 -6.93394125e-01
1.61289179e+00 -1.27562153e+00 -1.30468190e+00 -2.59920806e-01
7.38495827e-01 -8.07299435e-01 1.21618736e+00 4.87543762e-01
-7.42382586e-01 -3.71214598e-01 -6.33723080e-01 2.47066896e-02
-1.25412017e-01 5.26793480e-01 7.07717359e-01 3.52757543e-01
-9.26806092e-01 1.15732580e-01 -5.94587684e-01 -1.54031217e-01
-2.03478768e-01 1.97563812e-01 -1.15486689e-01 5.66020966e-01
-1.70717740e+00 1.13081503e+00 4.39723551e-01 2.00687379e-01
-5.67418993e-01 1.46357223e-01 -7.95161784e-01 -3.29603888e-02
1.89699292e-01 -2.49791279e-01 1.91255617e+00 -1.36428034e+00
-2.37792468e+00 8.64836514e-01 -1.41109169e-01 -2.28293464e-01
2.55923480e-01 -1.77877575e-01 -6.61264718e-01 9.58120748e-02
-5.40069155e-02 8.24628174e-01 5.54135799e-01 -1.43045020e+00
-8.09265077e-01 -5.68931624e-02 1.16019107e-01 3.67713660e-01
-1.72134191e-02 6.11070752e-01 -1.18508548e-01 -5.82326114e-01
-1.32338241e-01 -1.07559097e+00 -1.36768669e-01 -2.94580162e-01
-4.79456812e-01 -7.43337452e-01 4.50205415e-01 -7.75775313e-01
9.27309513e-01 -2.09348226e+00 7.74537548e-02 -1.53069735e-01
-2.36248165e-01 1.84329823e-01 2.33235791e-01 3.98286194e-01
2.69086748e-01 -2.50063032e-01 8.97566006e-02 -3.91083419e-01
3.81819814e-01 4.74767834e-01 -2.73549139e-01 -2.98498943e-03
4.34945971e-02 4.14366663e-01 -1.01834297e+00 -6.34246528e-01
2.02396601e-01 4.92875278e-01 -4.25344557e-01 7.62561083e-01
-6.15545869e-01 6.84797108e-01 -7.83013940e-01 2.09664717e-01
1.93397738e-02 2.05381569e-02 3.76707375e-01 -6.73348755e-02
-2.23980263e-01 5.95717132e-01 -7.56691813e-01 1.34944904e+00
-6.14875734e-01 4.90896583e-01 4.17978704e-01 -9.68908370e-01
1.37087715e+00 9.42831516e-01 3.62716198e-01 -6.53195024e-01
5.57995319e-01 2.18500551e-02 1.88072268e-02 -1.01198447e+00
4.09341455e-01 -5.11982918e-01 -5.86186826e-01 2.53347635e-01
-1.82686578e-02 -3.75461280e-01 -2.37458304e-01 -2.57408649e-01
8.47950161e-01 1.12552375e-01 2.35598996e-01 8.06630962e-03
6.26402497e-01 -2.54792333e-01 6.28616452e-01 2.41194367e-01
-4.64527130e-01 -8.12135190e-02 8.27868462e-01 -1.54241562e-01
-7.79823542e-01 -4.91377950e-01 1.30233616e-01 1.73977768e+00
-4.47695144e-02 1.28243417e-02 -9.87298191e-01 -4.45081860e-01
-7.15432525e-01 1.29521716e+00 -4.57998216e-01 -7.31004402e-02
-2.78855205e-01 -6.12596989e-01 6.64732635e-01 3.68169308e-01
6.20399117e-01 -1.75947225e+00 -5.88905156e-01 3.67368609e-01
-9.00809228e-01 -1.19135761e+00 -9.54861194e-02 5.50163150e-01
-1.59478143e-01 -2.92729616e-01 -3.68349344e-01 -9.88687277e-01
3.32930177e-01 -6.18287563e-01 1.06371963e+00 -3.35036844e-01
4.71250445e-01 8.47361684e-02 -8.43240678e-01 -5.13845503e-01
-8.40456247e-01 3.69103695e-03 4.65477221e-02 1.31262317e-01
2.17004359e-01 -4.51656193e-01 -4.26797509e-01 3.05338591e-01
-4.16851670e-01 4.10080642e-01 3.96422952e-01 7.73101628e-01
2.24229708e-01 -3.16129833e-01 1.03919446e+00 -9.18609619e-01
9.50017571e-01 -2.97845423e-01 4.33960855e-02 1.89084098e-01
-1.90121651e-01 2.33095400e-02 6.63237929e-01 -6.49130940e-01
-1.59381986e+00 1.35584593e-01 -8.43540192e-01 -1.81669533e-01
-4.77825731e-01 4.96026009e-01 -5.31016707e-01 5.83557725e-01
3.16700906e-01 -3.66861932e-02 -2.65256882e-01 -9.23904479e-02
5.11950076e-01 1.30822504e+00 6.14267588e-01 -6.56656146e-01
-9.42436904e-02 -3.77458632e-02 -6.20250344e-01 -5.88044763e-01
-9.77068126e-01 -4.36937287e-02 -3.28359693e-01 -7.01563060e-01
1.09048319e+00 -9.88643944e-01 -1.18614900e+00 3.88415039e-01
-1.55086005e+00 -5.18559158e-01 -3.63000073e-02 7.63597429e-01
-7.02829957e-01 1.41004443e-01 -1.09145069e+00 -1.32913291e+00
-5.59503913e-01 -1.04592335e+00 9.23724592e-01 1.30533367e-01
-7.73833930e-01 -7.26553202e-01 4.54183929e-02 4.80165750e-01
1.72885641e-01 -1.49668939e-02 8.86627495e-01 -1.03040552e+00
3.69502366e-01 -2.15848386e-01 -2.22281460e-02 5.81984878e-01
-1.74826067e-02 1.29129097e-01 -1.25035441e+00 3.11091125e-01
3.04562181e-01 -8.02718222e-01 -5.21861529e-03 3.22188251e-02
9.51233506e-01 -8.24289262e-01 -6.86771348e-02 -9.77028161e-02
4.36180770e-01 8.37629497e-01 5.72841346e-01 5.30293807e-02
3.44621599e-01 1.05778110e+00 9.56766725e-01 7.08223164e-01
6.46398962e-01 5.85475028e-01 1.54510602e-01 7.20290244e-02
4.73841280e-01 -2.71466911e-01 6.69521749e-01 1.23432946e+00
8.69190469e-02 -6.10959053e-01 -4.92763937e-01 2.85415918e-01
-2.00790071e+00 -9.31370378e-01 8.91798288e-02 1.44599807e+00
1.29085577e+00 9.38974619e-02 -9.45033357e-02 -1.14637606e-01
9.83242035e-01 2.33141467e-01 -3.68102401e-01 -9.32328284e-01
-3.71103324e-02 -3.30329567e-01 -1.23688027e-01 6.28855467e-01
-1.00991249e+00 1.29229617e+00 5.67923403e+00 7.50486732e-01
-1.24906123e+00 1.53178647e-01 8.03891599e-01 2.36764491e-01
-2.38616899e-01 -2.11839512e-01 -4.82379854e-01 6.08718932e-01
1.10039628e+00 1.49639264e-01 3.27538967e-01 1.27881587e+00
4.38530058e-01 -1.06807612e-01 -9.82110381e-01 8.02045524e-01
1.05014935e-01 -7.46267259e-01 -3.68257225e-01 -5.17558932e-01
4.07323897e-01 -1.34328067e-01 -2.48417154e-01 7.77353048e-01
6.55632615e-01 -5.92694223e-01 9.44410682e-01 6.05978489e-01
5.06138980e-01 -6.42546058e-01 7.91304231e-01 7.06044674e-01
-6.68493211e-01 1.27954915e-01 -3.97398844e-02 -2.28874341e-01
3.55617106e-01 4.11850095e-01 -1.03039515e+00 3.09357941e-01
5.96403182e-01 9.24099013e-02 1.59039348e-01 1.64743796e-01
-6.80619955e-01 6.72316670e-01 -6.32886738e-02 -8.33925962e-01
8.47157687e-02 -2.99726218e-01 2.76082724e-01 1.34972441e+00
2.52002805e-01 4.59555447e-01 3.45927268e-01 3.52459133e-01
-2.59747028e-01 3.36014301e-01 -2.75395542e-01 -1.48215145e-01
4.23934221e-01 1.31449175e+00 -9.10860419e-01 -5.68343163e-01
-1.94288179e-01 1.12159693e+00 3.43430758e-01 2.82071441e-01
-1.02956772e+00 -4.32494134e-01 3.19289029e-01 -5.77136874e-01
-6.91317394e-02 2.34212667e-01 1.08233243e-02 -8.75187755e-01
-1.19785048e-01 -7.67636538e-01 5.02252541e-02 -1.34529316e+00
-1.20406437e+00 1.07786238e+00 -2.04341471e-01 -8.70616376e-01
-6.32197022e-01 -5.35740912e-01 -3.88670474e-01 3.88405710e-01
-9.38075483e-01 -1.04521203e+00 3.26232836e-02 3.19474518e-01
5.67633748e-01 -1.89200342e-02 1.17026937e+00 3.26859802e-01
-4.67079729e-01 2.16843084e-01 -2.93945428e-02 1.93367839e-01
7.12425888e-01 -9.38352406e-01 -2.84272343e-01 7.72395507e-02
-3.22535425e-01 1.43531337e-01 1.25723720e+00 -3.82967770e-01
-8.35808396e-01 -8.38181674e-01 1.27933431e+00 -5.90220187e-03
6.95967734e-01 -6.18219137e-01 -7.21322656e-01 7.30289638e-01
5.55352569e-01 -4.48443502e-01 7.73457885e-01 2.20395654e-01
-2.06639543e-02 7.48823807e-02 -1.36721361e+00 6.67101443e-01
6.39309943e-01 -6.74995124e-01 -6.00129485e-01 5.27530134e-01
1.03733432e+00 -4.89720106e-01 -1.00365186e+00 5.31351089e-01
5.63131690e-01 -8.14217925e-01 4.33376402e-01 -4.32107925e-01
4.02127922e-01 6.39354289e-02 -2.06583977e-01 -1.43032014e+00
-6.78870175e-03 -6.05844498e-01 2.78161585e-01 1.43432438e+00
7.13837504e-01 -5.34620523e-01 4.16547596e-01 8.36716473e-01
-6.17498994e-01 -5.43761492e-01 -1.11964238e+00 -1.34631127e-01
-3.13941091e-01 -4.98257816e-01 3.62595320e-01 1.01730847e+00
7.43673384e-01 9.79695082e-01 -7.52491057e-01 -3.43995839e-02
-3.46706182e-01 3.43333697e-03 5.05318046e-01 -1.15187705e+00
-3.08147132e-01 -3.45499605e-01 4.45641167e-02 -1.11605632e+00
7.32599080e-01 -4.39612508e-01 7.86636531e-01 -1.44486630e+00
-3.97862867e-02 -4.15095210e-01 -1.16626970e-01 6.41966760e-01
-9.37257335e-03 -9.19225663e-02 -1.89195424e-01 -1.35243639e-01
-8.22568059e-01 1.01318872e+00 1.14666426e+00 -2.07535680e-02
-6.13052137e-02 1.92387000e-01 -4.02426243e-01 9.50110257e-01
7.47542322e-01 -1.91625208e-01 -3.64644349e-01 -1.28128966e-02
5.13076484e-01 7.32351780e-01 -1.55335858e-01 -4.35198814e-01
9.24691651e-03 -2.00244188e-01 -2.03767400e-02 -3.84553641e-01
6.98144853e-01 -5.94687164e-01 8.20275545e-02 1.31363913e-01
-8.26835692e-01 -2.06997365e-01 6.49032965e-02 1.18481599e-01
-3.72920215e-01 -4.11710441e-01 7.73036957e-01 6.00535423e-03
-4.41740394e-01 -3.06660891e-01 -8.51448298e-01 -1.45570442e-01
9.41155732e-01 4.73874062e-01 -1.75500125e-01 -8.97876740e-01
-9.67034817e-01 3.39044660e-01 -6.12854846e-02 6.21497095e-01
1.80558532e-01 -1.18560565e+00 -3.85624200e-01 -2.24359304e-01
7.57684037e-02 2.15466563e-02 3.40703487e-01 4.98202026e-01
-2.27075964e-01 3.23490173e-01 -1.54926971e-01 -2.03061953e-01
-1.15164828e+00 4.90114272e-01 2.39864320e-01 -2.67658383e-01
-2.25605324e-01 1.00228512e+00 1.02965057e-01 -7.56875217e-01
6.49507284e-01 -3.17760110e-01 -5.04100978e-01 2.45283201e-01
-2.20692251e-02 1.77177101e-01 -1.67525247e-01 -7.93894291e-01
-9.12459940e-02 -1.48765430e-01 2.85654426e-01 -5.35991669e-01
1.03745246e+00 -4.43381995e-01 -3.25319976e-01 1.04366338e+00
1.04645395e+00 -1.40701756e-01 -1.17061508e+00 -1.12735955e-02
1.16303742e-01 2.63045073e-01 -1.81283846e-01 -1.13820887e+00
-6.27911806e-01 6.34728611e-01 2.93291599e-01 4.69816566e-01
9.61145997e-01 1.97238967e-01 7.58358002e-01 6.39983714e-01
3.87051851e-01 -1.74196362e+00 -2.13456769e-02 7.47862279e-01
9.44577277e-01 -1.15336895e+00 -5.12785196e-01 -4.27116305e-01
-1.41295910e+00 7.75338948e-01 6.02886498e-01 2.28307724e-01
4.51543868e-01 3.92728984e-01 4.89404887e-01 -1.28789455e-01
-1.10852301e+00 1.49027959e-01 -2.10454971e-01 3.10204804e-01
7.18535006e-01 3.62068594e-01 -4.94045675e-01 1.14555919e+00
-4.79746491e-01 -8.65676999e-02 2.37537801e-01 6.16821527e-01
-6.29360795e-01 -8.66797686e-01 -8.05592313e-02 2.03080580e-01
-5.25952160e-01 2.72931188e-01 -5.64347208e-01 3.96696746e-01
1.44818589e-01 1.42708111e+00 1.18675590e-01 -4.04101878e-01
3.90079290e-01 5.72502315e-01 7.37178549e-02 -4.22351241e-01
-6.88874424e-01 4.64772880e-01 9.33126807e-01 -1.45383656e-01
-6.96350336e-01 -4.89785373e-01 -1.73379862e+00 1.50257245e-01
-3.26667041e-01 6.65652037e-01 6.25854015e-01 1.21785712e+00
1.36403590e-01 5.33349037e-01 1.02805531e+00 -7.91098237e-01
-5.00511050e-01 -1.51078963e+00 -4.44993973e-01 3.19280922e-01
-1.20989360e-01 -4.91211861e-01 -4.39340740e-01 1.15888081e-01] | [13.01169490814209, 6.1897478103637695] |
3daaa7b3-0688-447b-8b65-a3ece4fd511c | leveraging-relational-information-for-1 | 2205.10056 | null | https://arxiv.org/abs/2205.10056v1 | https://arxiv.org/pdf/2205.10056v1.pdf | Leveraging Relational Information for Learning Weakly Disentangled Representations | Disentanglement is a difficult property to enforce in neural representations. This might be due, in part, to a formalization of the disentanglement problem that focuses too heavily on separating relevant factors of variation of the data in single isolated dimensions of the neural representation. We argue that such a definition might be too restrictive and not necessarily beneficial in terms of downstream tasks. In this work, we present an alternative view over learning (weakly) disentangled representations, which leverages concepts from relational learning. We identify the regions of the latent space that correspond to specific instances of generative factors, and we learn the relationships among these regions in order to perform controlled changes to the latent codes. We also introduce a compound generative model that implements such a weak disentanglement approach. Our experiments shows that the learned representations can separate the relevant factors of variation in the data, while preserving the information needed for effectively generating high quality data samples. | ['Davide Bacciu', 'Andrea Valenti'] | 2022-05-20 | leveraging-relational-information-for | https://openreview.net/forum?id=TNmJgFmz2k | https://openreview.net/pdf?id=TNmJgFmz2k | null | ['relational-reasoning'] | ['natural-language-processing'] | [ 3.60340774e-01 5.41365564e-01 -3.33316028e-01 -2.12223679e-01
-6.48603082e-01 -9.29797947e-01 9.96390343e-01 -1.87519968e-01
-4.14799675e-02 6.80127740e-01 7.35755622e-01 -1.95278749e-01
-7.00483203e-01 -6.66515231e-01 -7.23466814e-01 -8.43734086e-01
6.66996017e-02 4.56508785e-01 -2.20000699e-01 -2.04284802e-01
1.28791958e-01 6.12361372e-01 -1.61781144e+00 3.97399902e-01
6.19756937e-01 2.70497978e-01 -1.62067682e-01 5.66825926e-01
7.75921568e-02 3.91734660e-01 -5.49340487e-01 -3.11003417e-01
3.67776364e-01 -6.26928151e-01 -3.33554596e-01 1.85765477e-03
4.24413383e-01 -1.38357341e-01 -3.06841403e-01 8.53930175e-01
8.38123783e-02 -1.29086077e-01 1.17199707e+00 -1.29695141e+00
-6.62875950e-01 7.52563953e-01 -5.52781284e-01 3.92570823e-01
-1.02547295e-01 -8.92138407e-02 1.54837370e+00 -6.33195817e-01
3.93773437e-01 1.21623564e+00 2.69720584e-01 7.27137148e-01
-1.99999642e+00 -6.54196858e-01 3.68700624e-01 -4.41797376e-01
-1.19693398e+00 -7.45707870e-01 7.81503439e-01 -8.73451591e-01
7.15289593e-01 4.60928023e-01 6.74837291e-01 1.38832664e+00
1.69253156e-01 6.48478508e-01 9.02273536e-01 -3.72197717e-01
2.89038926e-01 6.03998490e-02 3.17812622e-01 5.76945662e-01
9.82905149e-01 3.66082966e-01 -6.69212580e-01 -3.39384764e-01
9.16182935e-01 1.07760496e-01 -4.41684872e-01 -1.14038634e+00
-1.15852344e+00 1.13202929e+00 2.48876899e-01 2.98809141e-01
3.54378372e-02 1.64076209e-01 2.01802209e-01 2.50117809e-01
3.81908238e-01 8.67814422e-01 -5.34395099e-01 5.97159788e-02
-7.44053006e-01 4.14240956e-01 7.65905738e-01 7.22433448e-01
8.99674594e-01 6.90575540e-02 -1.72698691e-01 5.16980350e-01
5.43234229e-01 9.47921425e-02 2.21722722e-01 -7.48706639e-01
4.93018508e-01 6.16207659e-01 -1.06954262e-01 -8.23617756e-01
-1.52941763e-01 -4.57639426e-01 -6.42079413e-01 3.36624533e-01
5.84481299e-01 -1.52400017e-01 -8.47590029e-01 2.15933561e+00
-1.90509915e-01 1.26980349e-01 4.35360409e-02 6.66673422e-01
1.46870419e-01 2.86628544e-01 -1.57225430e-01 -6.32583052e-02
1.24072552e+00 -5.06218255e-01 -5.61327159e-01 -4.25753146e-01
4.27144438e-01 -1.71852574e-01 1.01951838e+00 3.03177029e-01
-9.82615471e-01 -2.09533751e-01 -1.34879005e+00 -9.04642642e-02
-1.96845740e-01 1.08314462e-01 9.34054077e-01 7.61835277e-01
-7.47887015e-01 5.96698940e-01 -1.03189230e+00 -9.55123827e-02
5.06538033e-01 3.34158629e-01 -3.87014329e-01 1.13575205e-01
-9.48222339e-01 6.84971869e-01 2.55534649e-01 -1.27592951e-01
-9.97087896e-01 -8.33925784e-01 -9.01574731e-01 3.73887569e-01
3.95417660e-01 -8.48793089e-01 7.56469786e-01 -7.88090765e-01
-1.01627505e+00 8.62184465e-01 -6.05802573e-02 -1.89478844e-01
1.92796677e-01 -3.38873208e-01 -7.94369876e-02 -1.34671822e-01
-9.92162377e-02 3.38494807e-01 1.03676569e+00 -1.50646460e+00
-6.66256174e-02 -7.42689013e-01 9.01633278e-02 9.07236636e-02
-1.36043906e-01 -2.12793872e-01 -1.16322555e-01 -7.24938929e-01
2.45402217e-01 -1.11058939e+00 -1.52051851e-01 5.22056222e-03
-5.12149930e-01 4.98349965e-02 3.29143733e-01 -2.37943739e-01
9.65360463e-01 -2.16712117e+00 6.78963304e-01 2.57507652e-01
7.36274898e-01 -2.94529013e-02 -2.47658104e-01 4.02302206e-01
-4.03078437e-01 4.21347231e-01 -1.17787801e-01 -4.47656065e-01
7.23080784e-02 4.61240262e-01 -8.49750340e-01 5.05074322e-01
5.46071827e-01 8.67718875e-01 -7.45629191e-01 7.31397271e-02
-2.82929629e-01 5.30858636e-01 -8.52486074e-01 1.42533258e-01
-3.76460969e-01 4.61241156e-01 -5.65068662e-01 1.04973517e-01
3.94443989e-01 -4.29634228e-02 3.71407241e-01 -1.14316076e-01
1.95431948e-01 6.93631113e-01 -1.03361356e+00 1.79616904e+00
-1.27398774e-01 6.95783436e-01 -1.22478418e-01 -1.11660159e+00
7.86575794e-01 2.89209306e-01 3.42820436e-01 -7.86091536e-02
2.74552442e-02 -3.86141501e-02 2.76997656e-01 -4.03323859e-01
3.04284215e-01 -2.70995498e-01 -8.60546827e-02 7.91266382e-01
3.18020463e-01 -2.60979794e-02 1.08713722e-02 2.08966553e-01
1.00108099e+00 3.35206181e-01 3.94014269e-01 -5.44232547e-01
-1.67434633e-01 -3.28346133e-01 6.59216166e-01 7.69173861e-01
4.10455884e-03 7.12799549e-01 1.15251541e+00 -1.27312407e-01
-1.14171469e+00 -1.45912659e+00 -7.80738741e-02 8.28448892e-01
-3.63688469e-01 -6.54529512e-01 -4.44707245e-01 -6.40554905e-01
-3.79646979e-02 6.39325798e-01 -9.00645077e-01 -4.83400464e-01
-4.10269409e-01 -1.05486429e+00 7.08640218e-01 4.96120423e-01
-3.76546621e-01 -4.39164370e-01 -7.48425186e-01 -3.70892406e-01
2.99085855e-01 -6.41374111e-01 -7.78347347e-03 7.04892695e-01
-1.01402509e+00 -1.06329215e+00 -3.62433702e-01 -1.46896273e-01
8.71386647e-01 3.31628472e-01 1.18139112e+00 -1.54701084e-01
-5.55498078e-02 9.86674950e-02 -5.15903309e-02 -1.92788497e-01
-3.38588893e-01 2.18135729e-01 -1.42152265e-01 -1.28523543e-01
5.85513353e-01 -9.93720710e-01 -3.08681816e-01 4.86152396e-02
-1.11430120e+00 1.84813991e-01 6.87356949e-01 8.39805305e-01
2.30153441e-01 -9.95343700e-02 3.78289729e-01 -1.07048929e+00
8.05016518e-01 -6.91563427e-01 -4.62514311e-01 2.71925926e-01
-5.96167564e-01 8.05679262e-01 3.14242661e-01 -5.55882394e-01
-6.64142668e-01 -1.10507049e-02 3.27843130e-01 -5.19351006e-01
-2.16002494e-01 3.93605798e-01 -5.69736302e-01 5.48685908e-01
8.93306255e-01 -1.17740100e-02 1.06119327e-01 -5.49237549e-01
7.91780829e-01 1.97464794e-01 -3.51897888e-02 -8.52294087e-01
9.46251094e-01 4.60702509e-01 -1.47307003e-02 -4.34271991e-01
-8.86992216e-01 6.62855133e-02 -7.51205504e-01 3.83082300e-01
7.88668692e-01 -1.00768864e+00 -3.58411521e-01 -3.10671896e-01
-9.08904076e-01 -1.03197463e-01 -5.69273651e-01 4.58627522e-01
-7.67280400e-01 -1.32172137e-01 -3.80990565e-01 -7.05924034e-01
4.43805754e-01 -1.20249152e+00 1.07413912e+00 -5.35150170e-02
-5.15830278e-01 -7.89369702e-01 4.85433072e-01 1.63646430e-01
4.69280221e-02 3.05556804e-01 1.35490847e+00 -9.22525883e-01
-7.35903561e-01 -5.78576140e-02 1.02148689e-02 4.11910117e-02
3.17397416e-01 8.64438638e-02 -1.23088765e+00 -2.44788289e-01
1.60350397e-01 -3.25578272e-01 1.20669162e+00 1.50783479e-01
8.13563287e-01 -4.52111691e-01 -3.70168120e-01 6.43338084e-01
1.22639430e+00 -9.03212950e-02 4.91542459e-01 -1.07978389e-01
8.14377725e-01 8.45886648e-01 -2.42567867e-01 1.98651150e-01
1.35518879e-01 7.32171416e-01 9.93971229e-02 4.22860570e-02
4.97559346e-02 -5.69943488e-01 3.49177808e-01 5.81506312e-01
-2.92464532e-02 -3.02620586e-02 -6.30907476e-01 4.67461795e-01
-1.84577298e+00 -1.04203081e+00 2.12840289e-01 2.10575223e+00
9.42054868e-01 1.82244197e-01 2.27125213e-01 2.98347831e-01
4.18392688e-01 4.58289504e-01 -6.75453246e-01 -4.04253185e-01
1.51568279e-02 1.40794367e-01 2.85917908e-01 5.83234191e-01
-6.45864189e-01 5.45922399e-01 6.93795395e+00 4.20119584e-01
-7.69480169e-01 -1.18791528e-01 4.55828071e-01 -3.49022120e-01
-1.06262207e+00 2.45276511e-01 -8.21110070e-01 1.77901015e-01
8.51927757e-01 -1.45829856e-01 6.88954532e-01 5.33615947e-01
-1.64017093e-03 2.53845453e-01 -1.82939935e+00 6.44961059e-01
1.44422263e-01 -1.25103438e+00 3.50579232e-01 4.91994619e-01
6.08452499e-01 -1.95659995e-01 2.73316205e-01 1.58036262e-01
6.51626766e-01 -1.35869026e+00 5.08490741e-01 6.33461714e-01
3.71985853e-01 -7.31053293e-01 -1.03722867e-02 4.94920850e-01
-6.20960951e-01 2.04934552e-02 -3.82783324e-01 -1.56039909e-01
-2.97850072e-01 7.57166147e-01 -6.46815479e-01 3.21401089e-01
1.69731587e-01 6.92050040e-01 -6.25913501e-01 4.17952031e-01
-4.97411907e-01 4.49988484e-01 5.40199280e-02 2.56705552e-01
-2.58775532e-01 -3.02870005e-01 6.00177407e-01 9.24318850e-01
1.74993590e-01 -1.94052890e-01 -3.79597247e-01 1.61158001e+00
4.20523360e-02 -4.62666333e-01 -1.16227913e+00 -3.77083987e-01
3.40829730e-01 1.03380358e+00 -5.97761512e-01 -1.48453638e-02
-4.68673229e-01 6.28165662e-01 6.45655394e-01 6.13644004e-01
-4.91336554e-01 1.68602437e-01 1.16415346e+00 -7.63662010e-02
4.26070035e-01 -5.70055485e-01 -5.62951863e-01 -1.76508570e+00
-1.12389870e-01 -9.79163527e-01 1.73049018e-01 -5.49255371e-01
-1.37969470e+00 4.44605172e-01 2.16297165e-01 -1.01133609e+00
-5.40620983e-01 -6.37334228e-01 -3.89392436e-01 1.08131695e+00
-1.24493492e+00 -9.12562430e-01 1.93421021e-01 3.87765706e-01
1.93193257e-01 -1.82205319e-01 1.06091869e+00 -1.63625330e-01
-7.14911997e-01 4.97571468e-01 6.96341097e-02 1.41372904e-02
3.72337371e-01 -1.30772841e+00 3.75710309e-01 9.51747298e-01
8.24580610e-01 1.33511817e+00 8.04451108e-01 -4.06679213e-01
-1.31245041e+00 -6.58322513e-01 6.91352010e-01 -8.66591513e-01
7.00915515e-01 -1.02865744e+00 -8.81784737e-01 1.02576125e+00
1.47468597e-01 -1.55226424e-01 1.25766420e+00 5.53062499e-01
-1.01053131e+00 1.18453249e-01 -7.59909570e-01 8.77809882e-01
1.13103116e+00 -8.73196244e-01 -8.58942926e-01 -1.60611123e-01
9.11185026e-01 1.38799310e-01 -6.60085022e-01 2.50518441e-01
7.06661463e-01 -1.13120615e+00 1.03566504e+00 -1.08191466e+00
6.25984311e-01 -2.11240396e-01 -4.21144873e-01 -1.53799713e+00
-4.55894887e-01 -5.15555382e-01 -2.60923773e-01 1.27570367e+00
5.31781971e-01 -4.75989521e-01 8.73599827e-01 8.88744295e-01
2.33236000e-01 -7.52610385e-01 -6.98945642e-01 -5.41154563e-01
4.84582841e-01 -2.68048763e-01 7.30546713e-01 1.00127780e+00
6.76303655e-02 6.96260571e-01 -2.21491098e-01 8.51113424e-02
4.34689462e-01 2.91197568e-01 6.56943321e-01 -1.36702955e+00
-8.26115191e-01 -5.91662705e-01 -4.28824335e-01 -1.01443636e+00
3.82158279e-01 -1.02921402e+00 -1.74770087e-01 -1.11875081e+00
3.91310275e-01 -4.23586279e-01 -4.08394516e-01 4.11473662e-01
-1.17333993e-01 -2.26746826e-03 1.92544177e-01 4.28369015e-01
4.27581593e-02 6.07466698e-01 9.69313502e-01 2.14058552e-02
-5.25029749e-02 -8.14806148e-02 -1.51087368e+00 5.57627618e-01
7.46881247e-01 -5.83352864e-01 -9.10291076e-01 -6.19464099e-01
6.45880878e-01 -1.35967046e-01 4.20821428e-01 -4.51921493e-01
-2.55478323e-01 -3.69471401e-01 4.13786232e-01 -4.53189276e-02
4.18069214e-01 -7.62715995e-01 1.77949131e-01 2.23641515e-01
-9.17577505e-01 -8.15985128e-02 2.65404861e-02 6.72010064e-01
1.25888914e-01 -1.66505963e-01 5.18041849e-01 -1.72049984e-01
1.17025683e-02 4.83284937e-03 -2.37109512e-01 1.20724998e-01
8.15350413e-01 -1.31151974e-01 -5.04509032e-01 -1.37404323e-01
-7.04344392e-01 -3.08326840e-01 4.86887604e-01 5.19997537e-01
4.48244691e-01 -1.23803413e+00 -5.50770521e-01 6.14135444e-01
3.20288599e-01 1.29046477e-02 -3.05294544e-01 4.47903663e-01
1.56410277e-01 4.61072296e-01 -2.76190609e-01 -3.05252492e-01
-8.31560075e-01 6.46009147e-01 3.39389384e-01 -2.04408154e-01
-5.44818223e-01 5.66153944e-01 7.26121664e-01 -2.11180672e-01
1.59764830e-02 -5.13437331e-01 -1.91511571e-01 2.25164652e-01
3.95120323e-01 2.62103304e-02 -7.16990307e-02 -4.12693232e-01
-2.53989279e-01 2.70946532e-01 -1.73656166e-01 -3.94874454e-01
1.43016708e+00 -3.51661816e-02 1.28316309e-03 8.51567507e-01
1.12290668e+00 4.48593587e-01 -1.60465670e+00 -7.25558251e-02
8.15661252e-02 -5.31370699e-01 -1.45258904e-01 -6.31555259e-01
-1.03542793e+00 9.92893994e-01 2.01062888e-01 3.53132188e-01
8.07526112e-01 2.77028710e-01 3.37127969e-02 1.67538553e-01
1.44489124e-01 -4.06038046e-01 4.07817289e-02 1.65403530e-01
8.42714965e-01 -6.66723192e-01 1.12983242e-01 -2.84574807e-01
-5.52949190e-01 9.05157983e-01 4.74410921e-01 -6.52111650e-01
5.99238694e-01 5.07674813e-01 -3.08993697e-01 -4.00163829e-01
-1.12230217e+00 -8.81066695e-02 5.07400513e-01 7.12974012e-01
5.75842738e-01 8.49962011e-02 -1.83609113e-01 7.54635096e-01
-2.82555044e-01 -3.39825630e-01 5.72726607e-01 6.64479077e-01
-2.83550441e-01 -1.19359088e+00 -1.36518940e-01 3.99969161e-01
-3.13314050e-01 -1.15710750e-01 -5.86327434e-01 6.72534466e-01
1.45061389e-01 6.99982166e-01 1.93602666e-01 -2.13913918e-01
1.40350863e-01 2.27797791e-01 6.85841084e-01 -9.34000969e-01
-1.13236949e-01 1.16707452e-01 -7.97306374e-02 -4.68628407e-01
-3.24323803e-01 -6.60587907e-01 -8.71769488e-01 1.09135240e-01
-1.75050765e-01 8.65331963e-02 4.08655435e-01 1.09089673e+00
5.47954977e-01 5.29045284e-01 4.64314729e-01 -7.44669735e-01
-6.38777971e-01 -7.26438999e-01 -6.66743875e-01 3.10175627e-01
7.14301944e-01 -7.60427177e-01 -6.01241529e-01 7.11897984e-02] | [9.25931453704834, 4.870980739593506] |
b16381c9-ce5c-46ce-ba81-1bb93f855b04 | text-to-speech-synthesis-based-on-latent | 2212.08329 | null | https://arxiv.org/abs/2212.08329v1 | https://arxiv.org/pdf/2212.08329v1.pdf | Text-to-speech synthesis based on latent variable conversion using diffusion probabilistic model and variational autoencoder | Text-to-speech synthesis (TTS) is a task to convert texts into speech. Two of the factors that have been driving TTS are the advancements of probabilistic models and latent representation learning. We propose a TTS method based on latent variable conversion using a diffusion probabilistic model and the variational autoencoder (VAE). In our TTS method, we use a waveform model based on VAE, a diffusion model that predicts the distribution of latent variables in the waveform model from texts, and an alignment model that learns alignments between the text and speech latent sequences. Our method integrates diffusion with VAE by modeling both mean and variance parameters with diffusion, where the target distribution is determined by approximation from VAE. This latent variable conversion framework potentially enables us to flexibly incorporate various latent feature extractors. Our experiments show that our method is robust to linguistic labels with poor orthography and alignment errors. | ['Tomoki Toda', 'Yusuke Yasuda'] | 2022-12-16 | null | null | null | null | ['text-to-speech-synthesis'] | ['speech'] | [ 4.38556522e-02 2.75842190e-01 -3.61097276e-01 -4.20427382e-01
-9.25899029e-01 -6.66421771e-01 9.77339208e-01 -6.74640715e-01
9.56203565e-02 5.01303732e-01 8.32387209e-01 -2.75720328e-01
2.13221446e-01 -7.79411793e-01 -5.94691813e-01 -7.77432978e-01
5.98404408e-01 8.37038517e-01 3.82214375e-02 -1.10705167e-01
1.07364245e-02 2.69598931e-01 -1.19528532e+00 2.46817559e-01
7.64220953e-01 6.33573592e-01 4.01658595e-01 8.40250850e-01
-7.11016178e-01 6.67224884e-01 -7.72851706e-01 -3.92935187e-01
9.04145911e-02 -6.55330181e-01 -4.30152804e-01 -5.09514883e-02
-1.61840066e-01 -2.66941279e-01 -4.19999331e-01 1.01391423e+00
3.21433038e-01 2.72270679e-01 1.39384437e+00 -1.15199375e+00
-1.20838380e+00 1.11022222e+00 -3.03468376e-01 -9.11282822e-02
9.62378830e-02 -4.59670201e-02 1.14267945e+00 -8.20145905e-01
6.31718338e-01 1.70725524e+00 5.21147132e-01 5.85716367e-01
-1.23696864e+00 -6.33675635e-01 -5.99362254e-02 1.41478226e-01
-1.37246478e+00 -7.59659410e-01 9.28291380e-01 -8.32945526e-01
1.06113565e+00 -2.56166141e-02 4.48828548e-01 1.51955628e+00
4.17063773e-01 9.51357782e-01 8.51341367e-01 -8.60474110e-01
3.55946630e-01 2.01654017e-01 -2.84440160e-01 4.71514970e-01
-5.67868054e-01 1.84933349e-01 -7.40751088e-01 -2.43776232e-01
8.77701819e-01 -2.50109643e-01 1.04503736e-01 -2.13357378e-02
-8.81767094e-01 1.15008414e+00 -4.97718066e-01 2.18293190e-01
-4.76226479e-01 2.84646571e-01 2.53693610e-02 9.39322487e-02
5.86148143e-01 6.04815185e-02 -3.86823088e-01 -3.95810604e-01
-1.27523506e+00 -6.12577274e-02 7.40003526e-01 1.21527755e+00
4.49320853e-01 6.92173302e-01 -4.06344324e-01 1.08524549e+00
9.03140366e-01 1.04316640e+00 9.16006625e-01 -1.13786256e+00
4.58492786e-01 1.11268200e-02 4.78416532e-02 -5.75355232e-01
2.15826958e-01 -6.84572607e-02 -4.08533841e-01 -2.99370219e-03
7.02064112e-02 -1.96652427e-01 -1.12195182e+00 1.89584637e+00
1.06344953e-01 1.26044527e-01 1.87776819e-01 3.51960659e-01
4.58678037e-01 1.21804774e+00 -2.06518129e-01 -5.58629274e-01
1.04322922e+00 -1.16830778e+00 -1.39801800e+00 -1.68541849e-01
2.62560636e-01 -1.11317968e+00 1.10587502e+00 2.19842866e-01
-1.14636302e+00 -6.46290064e-01 -1.08268344e+00 -2.12669998e-01
-1.60188973e-01 1.57586053e-01 -1.33166304e-02 7.02998996e-01
-1.11989760e+00 6.17494762e-01 -1.10021389e+00 -3.14054713e-02
-2.81379193e-01 2.23141804e-01 2.01934457e-01 3.62659186e-01
-1.34928834e+00 8.42353761e-01 1.60400290e-03 -1.81579590e-01
-8.40018451e-01 -6.94978386e-02 -7.11690187e-01 1.85754314e-01
-4.95093353e-02 -7.87228227e-01 1.43617153e+00 -9.06543136e-01
-2.46240568e+00 4.82573695e-02 -5.66351593e-01 -2.76574671e-01
2.55248368e-01 -1.88412845e-01 -5.08177638e-01 -3.75223830e-02
-1.58282340e-01 6.26523256e-01 1.46532476e+00 -9.53354359e-01
-3.17832947e-01 -4.74980026e-02 -8.03617179e-01 2.04506993e-01
-4.06219393e-01 1.07257888e-01 -5.13040364e-01 -9.58720744e-01
1.62940592e-01 -1.08972490e+00 1.96534976e-01 -2.53518462e-01
-2.60080576e-01 -5.17633140e-01 7.24906325e-01 -9.89617348e-01
1.29540873e+00 -2.24357486e+00 5.86573184e-01 8.31656158e-02
1.23357475e-02 -1.01162061e-01 -2.05959305e-01 6.46276712e-01
2.03914464e-01 6.14956096e-02 -1.43253759e-01 -9.11133051e-01
4.50463504e-01 5.57373106e-01 -1.01191461e+00 1.24908850e-01
-6.61659390e-02 8.86773765e-01 -6.27638161e-01 -5.23397088e-01
1.34891063e-01 8.01696301e-01 -3.97275686e-01 5.55118084e-01
-5.63148081e-01 3.61017913e-01 -1.51705563e-01 1.79669783e-01
2.44819641e-01 5.23475334e-02 1.73407272e-01 2.60099359e-02
-3.06771755e-01 6.46762609e-01 -1.04473281e+00 1.64195585e+00
-3.15680742e-01 8.98644030e-01 -2.10717604e-01 -3.74166965e-01
1.19439542e+00 6.35694504e-01 1.81963533e-01 -1.77072480e-01
2.24406511e-01 1.53390869e-01 -1.74194619e-01 -4.55886751e-01
7.60215938e-01 -3.77351552e-01 5.96931204e-02 6.48766458e-01
2.84772128e-01 -6.17134690e-01 -1.33055344e-01 1.31868482e-01
7.69441187e-01 4.37551349e-01 6.37864098e-02 3.63026783e-02
4.29259520e-03 -3.30615491e-01 6.51215851e-01 5.66590071e-01
7.77270347e-02 7.23148465e-01 4.00492400e-01 6.88346550e-02
-1.23692381e+00 -1.36057436e+00 2.38989666e-01 9.45987344e-01
-3.63078684e-01 -5.47824442e-01 -8.34187090e-01 -2.28196919e-01
-2.46312976e-01 1.40357530e+00 -3.40427488e-01 -1.24601685e-01
-3.86961639e-01 -2.67987728e-01 6.31553710e-01 5.94859362e-01
-2.00175390e-01 -9.66921568e-01 3.56812805e-01 4.64088768e-01
-4.15215969e-01 -9.60505664e-01 -7.48754859e-01 1.00596845e-01
-7.30670989e-01 -7.21103027e-02 -7.01032817e-01 -6.25370681e-01
1.85572386e-01 -2.06095934e-01 5.69393098e-01 -6.86387241e-01
3.42990786e-01 3.31382513e-01 -3.53954077e-01 -4.12304014e-01
-1.18291426e+00 6.14287369e-02 5.20963371e-01 1.14442063e-02
3.27789187e-01 -7.56474853e-01 -1.86108157e-01 1.29265055e-01
-7.72001028e-01 1.14318475e-01 4.28875983e-01 6.07960761e-01
6.83638275e-01 -3.10699102e-02 1.56357065e-01 -6.46331966e-01
6.92974329e-01 -4.81563479e-01 -7.97942281e-01 2.99327940e-01
-7.34080315e-01 5.00711858e-01 5.54105103e-01 -9.18271959e-01
-1.33608949e+00 1.19171301e-02 -3.84971380e-01 -7.45942473e-01
8.92648250e-02 4.54353303e-01 -1.15635380e-01 7.11030841e-01
4.68755454e-01 4.02203202e-01 2.67984439e-02 -5.77092290e-01
9.10267770e-01 1.10377347e+00 4.47969109e-01 -5.54220438e-01
8.99192214e-01 2.14259312e-01 -5.50848246e-01 -9.56520021e-01
-5.11923790e-01 -1.04867825e-02 -7.39174485e-01 -1.15942113e-01
1.05124593e+00 -8.48917067e-01 -9.57575589e-02 4.39753115e-01
-1.49645352e+00 -4.79456216e-01 -5.38594544e-01 9.08520997e-01
-7.82572329e-01 4.51798379e-01 -9.56435323e-01 -9.36930835e-01
-1.20285593e-01 -1.36926591e+00 1.04381967e+00 3.70765217e-02
-5.82959175e-01 -1.09806252e+00 5.20410359e-01 3.15845996e-01
6.34442925e-01 -3.14446658e-01 1.20507812e+00 -4.74720806e-01
-5.80460072e-01 1.70002133e-01 3.91531706e-01 5.59684813e-01
3.15996617e-01 6.26310110e-01 -1.04654479e+00 -1.83467921e-02
2.45880872e-01 2.37815399e-02 4.65455949e-01 8.10786009e-01
5.55292189e-01 -4.39240992e-01 -1.74893916e-01 7.21068859e-01
9.13843930e-01 4.77148980e-01 5.21513104e-01 -1.75855771e-01
8.06265831e-01 4.77580249e-01 7.21995905e-02 4.13823634e-01
5.13962388e-01 6.61110461e-01 -2.20471352e-01 2.60155171e-01
-4.64443892e-01 -5.72957575e-01 9.09851670e-01 2.19147968e+00
2.78611511e-01 -5.71776867e-01 -9.24596250e-01 4.30589825e-01
-1.69606435e+00 -9.13757324e-01 2.97739878e-02 1.81026542e+00
1.00569320e+00 1.01869665e-01 -2.22532019e-01 -1.67247921e-01
8.68593633e-01 2.98039526e-01 -4.61937666e-01 -6.31068468e-01
2.78104506e-02 8.19841549e-02 2.86647797e-01 8.66518080e-01
-5.21174610e-01 1.42964613e+00 7.11306953e+00 1.08375537e+00
-1.02551615e+00 2.78798044e-01 -2.75546312e-02 4.34473194e-02
-9.11436975e-01 4.85354401e-02 -1.06109810e+00 4.80158180e-01
1.44047153e+00 -9.48681906e-02 6.80572689e-01 7.46104240e-01
3.86202008e-01 5.80164254e-01 -9.81934011e-01 6.92541003e-01
1.68435231e-01 -1.18572748e+00 2.90185183e-01 1.14387758e-01
9.27438378e-01 1.66433617e-01 4.66229916e-01 4.08325613e-01
8.24277461e-01 -9.40101445e-01 1.18970466e+00 6.74182653e-01
1.03103995e+00 -3.32264394e-01 1.83486924e-01 3.98924142e-01
-1.11344492e+00 9.84005406e-02 -3.16759437e-01 1.57999828e-01
3.82095426e-01 3.40382665e-01 -9.99473453e-01 -7.00366721e-02
3.25264484e-02 3.66724223e-01 1.85657144e-01 4.16595906e-01
-7.22573519e-01 1.23392808e+00 -2.75193125e-01 -6.88612536e-02
2.14105826e-02 -4.45690095e-01 7.70519853e-01 1.12137711e+00
6.47186995e-01 -1.15702741e-01 -4.72101010e-02 1.04682076e+00
6.63900152e-02 1.89972669e-01 -3.67166102e-01 -5.85405648e-01
1.03864133e+00 7.59994268e-01 -6.14236891e-01 -2.98893690e-01
-3.53356153e-01 1.14443743e+00 1.08774833e-01 6.35615647e-01
-8.32134247e-01 -2.80641675e-01 4.61050570e-01 -2.48965457e-01
6.17003739e-01 -7.80038178e-01 -2.16940880e-01 -1.19328856e+00
-3.54107916e-01 -8.14301372e-01 -1.89547166e-01 -1.06539965e+00
-1.41356039e+00 9.21025515e-01 8.82094279e-02 -8.74486983e-01
-8.18240285e-01 -3.59960228e-01 -3.45479608e-01 1.15675962e+00
-1.30082071e+00 -1.16014373e+00 3.91544729e-01 4.34062243e-01
9.98895943e-01 -5.76789498e-01 9.03734505e-01 2.16025263e-02
-4.51446623e-01 4.77942586e-01 7.08634496e-01 1.48938298e-02
8.12356293e-01 -1.24362087e+00 8.61724854e-01 7.81657100e-01
5.61124802e-01 7.77398050e-01 9.31670487e-01 -9.16465700e-01
-9.91346180e-01 -8.28826249e-01 1.10702145e+00 -6.99713647e-01
7.90327847e-01 -5.24983943e-01 -8.72626007e-01 9.26270664e-01
3.84205312e-01 -5.84924877e-01 8.04401398e-01 9.06055346e-02
-3.42543811e-01 2.59505719e-01 -4.47836041e-01 7.41594970e-01
6.08393252e-01 -1.13430655e+00 -1.01688051e+00 1.32754833e-01
1.16260862e+00 -4.44012731e-01 -6.96569920e-01 -1.60589576e-01
6.47698462e-01 -3.06615919e-01 5.75711906e-01 -1.70745641e-01
3.42982680e-01 -2.38520160e-01 -4.84991252e-01 -1.59656954e+00
-5.30314684e-01 -9.33490813e-01 -3.85222435e-01 1.58447361e+00
5.69527805e-01 -3.71872604e-01 4.29572016e-01 6.44372880e-01
-2.61794001e-01 -2.05237374e-01 -1.00061071e+00 -8.35383594e-01
2.46675804e-01 -6.01920128e-01 7.80888259e-01 9.49369133e-01
-3.03634793e-01 5.52932680e-01 -5.58375716e-01 3.29521835e-01
3.77287507e-01 -1.41430005e-01 4.65748966e-01 -9.93530989e-01
-7.86407292e-01 -4.24992502e-01 1.57123387e-01 -1.46412086e+00
4.55451071e-01 -8.50409269e-01 3.85146141e-01 -1.47661150e+00
-9.17367712e-02 -1.29878357e-01 -6.44429773e-02 2.55282521e-01
1.07422195e-01 -3.36197793e-01 4.37774621e-02 5.52947938e-01
4.31364067e-02 1.17065883e+00 8.21080863e-01 -4.00969051e-02
-6.53620422e-01 -1.23052998e-02 -1.33850902e-01 8.95999372e-01
7.22461164e-01 -1.00090480e+00 -6.06852770e-01 -7.41674781e-01
1.43940359e-01 4.08719629e-01 -4.65422183e-01 -5.87405384e-01
4.86736566e-01 -4.47490305e-01 1.35050640e-01 -7.08188117e-01
7.55122185e-01 -4.29573953e-01 3.28216583e-01 -3.85965928e-02
-5.84022999e-01 1.37390792e-01 -1.64116144e-01 5.37312746e-01
-9.51245576e-02 -2.55099088e-01 5.88589072e-01 2.51394749e-01
-2.02943996e-01 1.74216375e-01 -1.16265357e+00 -9.76613462e-02
4.34545696e-01 -5.03933690e-02 -9.34729725e-02 -6.88718915e-01
-6.94330931e-01 -2.19817698e-01 1.91509515e-01 6.58371329e-01
5.52218437e-01 -1.51318753e+00 -6.49644196e-01 4.16567832e-01
-2.44549707e-01 -3.66727352e-01 -1.22046478e-01 2.47576550e-01
-1.07689649e-01 3.58276546e-01 2.20986739e-01 -5.07631719e-01
-1.18736851e+00 2.99112439e-01 6.44356832e-02 -2.79790573e-02
-4.03933913e-01 1.01507056e+00 5.69578595e-02 -5.55180967e-01
3.65926147e-01 -3.26564074e-01 -2.85320729e-02 2.21605971e-02
3.04789066e-01 3.46696347e-01 -5.15948832e-01 -9.14074063e-01
-2.79229451e-02 4.99753147e-01 7.92827010e-02 -1.12051761e+00
1.24221468e+00 -5.95685363e-01 1.97215974e-01 1.03002131e+00
1.03528965e+00 5.96782744e-01 -1.47437942e+00 -3.84804845e-01
-1.16218939e-01 -1.41091302e-01 3.95390809e-01 -6.61743283e-01
-6.68557107e-01 1.28069544e+00 4.02297080e-01 2.52451412e-02
4.63545114e-01 -2.29379814e-02 1.06581688e+00 7.54494146e-02
-1.89758372e-02 -1.51989663e+00 1.66641042e-01 8.49475622e-01
7.45263100e-01 -6.60100818e-01 -3.86970073e-01 -9.34795663e-02
-9.44623291e-01 1.27401090e+00 5.05316034e-02 2.25168541e-01
7.51117587e-01 5.44709086e-01 4.66899574e-01 9.79406014e-02
-9.40210938e-01 9.68243256e-02 5.11191547e-01 6.17692828e-01
4.50976074e-01 2.52696365e-01 3.44850421e-02 7.43861854e-01
-5.71746349e-01 -2.11611584e-01 4.85311061e-01 3.90093654e-01
-5.10146916e-01 -1.51014733e+00 -4.78036702e-01 -8.15175008e-03
-2.66594708e-01 -3.98540199e-01 -2.35829204e-01 1.42095640e-01
-8.52681231e-03 9.43895638e-01 1.20791897e-01 -4.83324051e-01
-1.56800762e-01 7.27025628e-01 3.30491871e-01 -6.56709790e-01
-7.56987706e-02 7.00633109e-01 -1.94236368e-01 -9.76969749e-02
4.34559658e-02 -8.21625412e-01 -1.31956899e+00 -2.11330697e-01
-6.14095986e-01 3.10476750e-01 1.09761405e+00 1.18511438e+00
2.47921377e-01 6.02086723e-01 5.96871614e-01 -5.08194447e-01
-8.55499446e-01 -1.04291070e+00 -8.18286479e-01 -1.32274806e-01
2.34893009e-01 -5.50655603e-01 -5.96544087e-01 3.96914303e-01] | [15.00573444366455, 6.563294410705566] |
ffc70f5a-ee0f-4032-ac31-627d92b854fe | direct-robot-configuration-space-construction | 2303.05653 | null | https://arxiv.org/abs/2303.05653v1 | https://arxiv.org/pdf/2303.05653v1.pdf | Direct Robot Configuration Space Construction using Convolutional Encoder-Decoders | Intelligent robots must be able to perform safe and efficient motion planning in their environments. Central to modern motion planning is the configuration space. Configuration spaces define the set of configurations of a robot that result in collisions with obstacles in the workspace, C-clsn, and the set of configurations that do not, C-free. Modern approaches to motion planning first compute the configuration space and then perform motion planning using the calculated configuration space. Real-time motion planning requires accurate and efficient construction of configuration spaces. We are the first to apply a convolutional encoder-decoder framework for calculating highly accurate approximations to configuration spaces. Our model achieves an average 97.5% F1-score for predicting C-free and C-clsn for 2-D robotic workspaces with a dual-arm robot. Our method limits undetected collisions to less than 2.5% on robotic workspaces that involve translation, rotation, and removal of obstacles. Our model learns highly transferable features between robotic workspaces, requiring little to no fine-tuning to adapt to new transformations of obstacles in the workspace. | ['Hod Lipson', 'Riya Gupta', 'Carl Gross', 'Christopher Benka'] | 2023-03-10 | null | null | null | null | ['motion-planning'] | ['robots'] | [-1.90994248e-01 1.99802220e-01 4.55438606e-02 -4.38529514e-02
-5.75299561e-01 -7.03133941e-01 4.79370683e-01 -7.32441545e-02
-6.33248746e-01 5.12035131e-01 1.41614974e-01 -5.83348930e-01
-2.76059300e-01 -6.10461056e-01 -9.08292174e-01 -2.96214491e-01
-4.27929997e-01 8.42431247e-01 2.42768109e-01 -5.62162817e-01
3.94911319e-01 8.76090646e-01 -1.23897207e+00 -1.81222067e-03
2.74575204e-01 4.45771754e-01 7.51769602e-01 9.30437863e-01
1.57200456e-01 2.90336728e-01 -5.61677277e-01 2.63095140e-01
6.17112517e-01 -4.43540551e-02 -1.14176738e+00 -1.70590445e-01
-4.07434970e-01 -1.96880355e-01 -6.66553736e-01 5.74755192e-01
3.32810640e-01 4.87465441e-01 7.91073442e-01 -1.29284430e+00
-2.77995974e-01 4.72879201e-01 -1.83021545e-01 -2.78702468e-01
4.68397915e-01 6.32368505e-01 3.22981745e-01 -7.48414218e-01
8.84265304e-01 1.28727281e+00 7.73595691e-01 5.37879884e-01
-9.22972798e-01 -4.97922599e-02 -7.79886171e-02 -3.73383868e-03
-1.50996125e+00 -2.86107957e-01 3.28229547e-01 -5.52732050e-01
1.67299843e+00 -1.11272998e-01 5.66001356e-01 7.95627296e-01
1.20372617e+00 1.86549455e-01 4.72624898e-02 -3.83469790e-01
3.65303546e-01 -8.01334620e-01 -5.10576010e-01 7.73789525e-01
2.56377578e-01 1.23884119e-01 1.39076397e-01 1.91913590e-01
1.10413885e+00 -1.02511182e-01 1.47345886e-01 -8.48513424e-01
-1.70368421e+00 4.28761214e-01 6.76362336e-01 -1.06523642e-02
-1.72086731e-01 5.86456358e-01 3.27925235e-01 1.42739555e-02
-5.87551177e-01 1.05811357e+00 -5.13170779e-01 -3.19350690e-01
-9.01708379e-02 5.17731726e-01 7.58533478e-01 1.79170835e+00
5.75857878e-01 -1.07354544e-01 3.48121405e-01 3.02289933e-01
7.33664706e-02 4.11393225e-01 3.33441675e-01 -1.58368540e+00
8.28040063e-01 5.01729846e-01 5.84754586e-01 -6.55295789e-01
-1.04809284e+00 3.46721336e-02 -4.55367833e-01 7.22409904e-01
3.82138163e-01 -3.74664187e-01 -9.03177440e-01 1.32891953e+00
2.44739890e-01 -7.27327406e-01 2.69546896e-01 9.21296000e-01
-3.14207107e-04 4.87659872e-01 -1.84025317e-01 3.70053023e-01
8.51576209e-01 -1.01514292e+00 -3.88555288e-01 -5.98567247e-01
9.98271883e-01 -9.50550556e-01 9.11694705e-01 4.41151768e-01
-1.05412984e+00 -5.08008897e-01 -1.40991628e+00 -2.31397316e-01
8.37766286e-03 2.52625108e-01 6.04820728e-01 -1.13089560e-02
-9.01456356e-01 9.53415215e-01 -1.29123890e+00 -2.55323321e-01
1.24931157e-01 7.57827520e-01 -7.99444795e-01 -9.90126878e-02
-6.44049466e-01 1.61073744e+00 5.57020247e-01 2.56399333e-01
-1.07383823e+00 1.00859581e-02 -1.03304517e+00 -3.01595002e-01
3.38737369e-01 -7.28945851e-01 1.55248678e+00 -3.07760775e-01
-1.46354413e+00 2.31769443e-01 1.03363410e-01 -1.40421435e-01
4.34378415e-01 -3.95318091e-01 -1.03869068e-03 -1.68847367e-01
2.01290190e-01 6.99208617e-01 3.82490158e-01 -1.15071297e+00
-5.18616974e-01 -1.49633750e-01 1.32412866e-01 4.18716997e-01
5.17360389e-01 -5.02816856e-01 -5.91476679e-01 -6.74504638e-02
6.07948840e-01 -1.55773079e+00 -8.04681003e-01 1.48542643e-01
-5.13718367e-01 -1.00821943e-03 5.37676275e-01 -3.36737424e-01
4.17648882e-01 -1.78009093e+00 5.86836517e-01 1.19782567e-01
-1.59457192e-01 -8.07913095e-02 -3.08708966e-01 5.36866426e-01
2.29483381e-01 -1.40135065e-01 4.83555198e-02 -7.19407946e-02
4.09086235e-02 5.29255748e-01 -1.85810387e-01 6.21967077e-01
1.75777137e-01 8.57541978e-01 -1.01015425e+00 -1.22340128e-01
4.10122067e-01 1.79758310e-01 -7.09604204e-01 1.28908694e-01
-2.67356664e-01 5.14340997e-01 -5.07714093e-01 3.58847439e-01
2.77389884e-01 2.92299360e-01 4.13662642e-01 1.14977755e-01
-5.18434942e-01 4.33817387e-01 -1.02742207e+00 2.26372099e+00
-6.15134299e-01 5.66920280e-01 9.03186351e-02 -4.42328990e-01
1.03213751e+00 -1.54817462e-01 5.77989340e-01 -3.96235704e-01
3.56642485e-01 3.31036896e-01 1.53693661e-01 -5.13643026e-01
1.04866397e+00 9.71665531e-02 -7.65389025e-01 3.56308550e-01
-1.46548539e-01 -6.65714502e-01 -4.11767736e-02 -1.92684114e-01
1.51588058e+00 6.14165246e-01 2.99635351e-01 -1.55880913e-01
5.46990149e-02 6.45635426e-01 3.88583392e-01 4.93142694e-01
-1.08288921e-01 6.54442132e-01 3.34854752e-01 -6.09282613e-01
-1.63910615e+00 -1.15987551e+00 4.09308463e-01 7.04605937e-01
7.24724472e-01 -3.01787913e-01 -6.67471230e-01 -4.01355326e-02
9.36867148e-02 6.78619266e-01 -3.35600555e-01 -4.26471978e-01
-1.15990090e+00 -2.67372979e-03 4.39865112e-01 7.21907079e-01
8.71921033e-02 -1.02674747e+00 -1.21200395e+00 3.83589953e-01
-2.77808636e-01 -8.99445593e-01 -5.52793443e-01 5.90775192e-01
-8.98239493e-01 -1.42096150e+00 -2.33768165e-01 -1.09560239e+00
9.60730255e-01 3.00364643e-01 6.01460695e-01 -1.98905930e-01
-6.51060581e-01 2.36378789e-01 -3.29076260e-01 -1.19401895e-01
-6.19634688e-01 1.13157228e-01 3.84362906e-01 -1.14274728e+00
-1.91691101e-01 -2.74411201e-01 -4.45787907e-01 6.44930720e-01
-2.66634881e-01 1.14263259e-01 9.83430207e-01 5.55395186e-01
6.91718936e-01 7.49235824e-02 -6.34619966e-02 -2.74461005e-02
4.96740937e-01 -5.06879926e-01 -4.94474590e-01 2.14325404e-03
-3.04201961e-01 5.78826010e-01 6.38660073e-01 -5.08049190e-01
-9.19642985e-01 8.59364986e-01 4.52159680e-02 -5.08264005e-01
-1.89488947e-01 4.46291983e-01 -2.04499692e-01 6.13814853e-02
8.50488126e-01 -3.07717621e-01 2.38628648e-02 -3.34624052e-01
6.81833804e-01 4.34534311e-01 1.09015691e+00 -6.76791131e-01
6.04122460e-01 3.14557433e-01 4.12386447e-01 -5.28388798e-01
5.79493903e-02 -2.72978127e-01 -1.32992530e+00 -1.84400678e-01
8.11745584e-01 -5.83167672e-01 -8.58917773e-01 1.01285204e-01
-1.45971572e+00 -7.68486619e-01 -2.85294294e-01 5.30559123e-01
-1.32777977e+00 2.17412665e-01 -2.44616106e-01 -7.15817094e-01
1.51210092e-02 -1.50505972e+00 1.04967892e+00 -2.31454283e-01
-6.22265875e-01 -2.92905033e-01 3.03014740e-02 -1.11479066e-01
1.70733169e-01 5.66386223e-01 9.89715040e-01 2.98490245e-02
-6.05999351e-01 -3.31282735e-01 4.38384384e-01 -4.17258084e-01
2.30226398e-01 -2.15548083e-01 -1.83710024e-01 -2.63208598e-01
-4.37252462e-01 -1.19106419e-01 3.42615008e-01 3.04170519e-01
8.56467128e-01 -2.94300646e-01 -8.10108244e-01 4.51095372e-01
1.20450997e+00 4.88748252e-01 8.69080424e-01 6.80943489e-01
4.08213168e-01 5.79662323e-01 1.06684196e+00 4.56373602e-01
3.19941431e-01 8.67332280e-01 9.15629327e-01 6.42304599e-01
-1.16829708e-01 -2.76626945e-01 5.11377692e-01 4.96563613e-01
-1.40193880e-01 -1.13608226e-01 -1.28682613e+00 6.83269799e-01
-2.03162789e+00 -6.51873589e-01 -2.39501461e-01 2.07113910e+00
6.83980227e-01 3.77563626e-01 -2.14077964e-01 -1.11693598e-01
5.16612053e-01 -4.02365685e-01 -6.77267849e-01 -6.64917052e-01
4.59161222e-01 -9.32743400e-02 8.77648294e-01 7.20262945e-01
-1.27610278e+00 9.67350125e-01 6.80181980e+00 1.98103949e-01
-6.95530832e-01 -2.77750939e-01 -1.87811330e-01 -5.20288467e-01
7.15770051e-02 2.22255886e-02 -6.06459260e-01 1.04849689e-01
9.53693330e-01 -1.00784535e-02 6.44851387e-01 1.48723161e+00
5.49746305e-02 -3.31915379e-01 -1.31826174e+00 7.11486280e-01
-2.89346308e-01 -1.21642148e+00 -4.29851890e-01 -1.49436845e-02
6.04219913e-01 1.92485809e-01 -6.80377185e-02 3.66846174e-01
7.05880642e-01 -1.20982194e+00 1.27769041e+00 4.92107242e-01
9.88414586e-01 -9.12461877e-01 6.60776019e-01 5.60274899e-01
-1.24797428e+00 -3.81796271e-01 -7.35085487e-01 -2.31053710e-01
4.30704117e-01 -1.13296546e-01 -1.34525263e+00 4.30702150e-01
3.46727937e-01 2.13106781e-01 -3.37159373e-02 9.98854756e-01
-1.08705118e-01 -5.25779665e-01 -2.22424179e-01 -1.82475388e-01
2.33012006e-01 1.50890023e-01 5.65693974e-01 8.08560252e-01
7.26373255e-01 -3.37337120e-03 2.16062278e-01 6.79378390e-01
4.47616011e-01 -6.46499693e-01 -9.09515381e-01 2.98332423e-01
6.79536164e-01 8.17254424e-01 -6.85467482e-01 4.31108594e-01
1.36839718e-01 1.03268516e+00 2.93682098e-01 1.33388430e-01
-7.78077304e-01 -8.30475688e-01 1.05960715e+00 8.19043070e-03
2.57843345e-01 -1.37322497e+00 -4.16943997e-01 -4.25602645e-01
-4.74220887e-02 -3.09906423e-01 -3.09560120e-01 -9.73777235e-01
-5.78695774e-01 5.02375960e-01 1.23153389e-01 -1.59431505e+00
-7.25911617e-01 -1.09435356e+00 -2.62742490e-01 7.14688361e-01
-8.76898706e-01 -8.36294591e-01 -3.46215785e-01 3.47395599e-01
7.98119724e-01 -1.82680607e-01 1.08111274e+00 -1.81044623e-01
-6.22749850e-02 2.33676478e-01 5.40168248e-02 2.63489224e-02
4.25408006e-01 -8.82003486e-01 9.26948845e-01 5.78332663e-01
-4.62874949e-01 7.94311047e-01 8.72693300e-01 -8.67042065e-01
-2.15221715e+00 -1.39264524e+00 5.91936827e-01 -8.51043403e-01
2.58772135e-01 -1.68878511e-01 -3.19739968e-01 1.09829009e+00
-4.29672033e-01 -2.22678959e-01 4.60321382e-02 -1.11415520e-01
-2.37939715e-01 5.03662825e-01 -8.74657154e-01 1.12831366e+00
1.53121114e+00 -4.58560549e-02 -7.49547422e-01 2.56145984e-01
7.23515511e-01 -1.05200303e+00 -7.68780351e-01 2.05660522e-01
7.53351748e-01 -3.51665735e-01 1.14081383e+00 -6.59454107e-01
4.29418147e-01 -5.10816574e-01 -1.94110692e-01 -1.33847845e+00
-6.07817352e-01 -8.19306731e-01 1.38032334e-02 1.55986875e-01
4.71439481e-01 -6.62233233e-02 7.10280061e-01 7.36044884e-01
-9.91528153e-01 -7.14063525e-01 -1.14175761e+00 -1.14788604e+00
2.75026768e-01 -6.54934227e-01 6.52570605e-01 3.68710667e-01
5.71339726e-01 -1.93076566e-01 -5.28373150e-03 2.91690022e-01
2.28530988e-01 4.03717364e-04 1.13443041e+00 -8.64872932e-01
3.77863348e-02 -4.40245569e-01 -5.01516402e-01 -1.11111605e+00
2.44515806e-01 -8.18503380e-01 9.49143171e-01 -1.83444071e+00
-2.63123572e-01 -8.24259341e-01 4.23569232e-01 6.92090392e-01
5.64699948e-01 -2.99738795e-01 8.94814581e-02 4.68719214e-01
-7.18943536e-01 6.15739524e-01 1.30167544e+00 -1.08931057e-01
-3.40593815e-01 -2.19730988e-01 -2.20145226e-01 8.08851063e-01
1.16227651e+00 -3.40684801e-01 -2.60031849e-01 -9.14597452e-01
3.12504560e-01 1.40765801e-01 1.86743960e-02 -1.30432296e+00
3.67587745e-01 -6.73061132e-01 5.57021618e-01 -5.62920928e-01
5.99032640e-01 -8.37101936e-01 4.34744835e-01 9.17260826e-01
-3.66432369e-01 4.48140144e-01 5.15983641e-01 5.86883903e-01
5.02625644e-01 -4.02294874e-01 2.98242599e-01 -2.63217777e-01
-1.08434522e+00 -1.48554191e-01 -1.03055644e+00 -2.91736931e-01
1.41891384e+00 -3.66051823e-01 -4.19505537e-01 -1.81914344e-01
-7.96625078e-01 3.38123173e-01 7.48640239e-01 8.34495008e-01
8.84668708e-01 -1.34313118e+00 -1.18974380e-01 1.37569800e-01
-2.68848985e-02 5.62813401e-01 -8.85809883e-02 4.07505155e-01
-1.18120861e+00 7.48130143e-01 -6.97383046e-01 -3.75223130e-01
-9.35113311e-01 4.82187569e-01 3.73679936e-01 7.11496323e-02
-7.99033225e-01 6.91544354e-01 -2.98691720e-01 -9.19852138e-01
6.01176172e-03 -6.40220284e-01 1.74970612e-01 -8.59836817e-01
1.14642575e-01 5.63788235e-01 4.80347462e-02 -8.03620219e-01
-5.53120255e-01 5.26486993e-01 4.13419336e-01 -3.02383691e-01
1.06518960e+00 -5.38179502e-02 1.49975926e-01 1.98946938e-01
9.77001131e-01 -4.16212946e-01 -1.63939023e+00 4.95738953e-01
1.75165877e-01 -3.86894464e-01 -3.92425537e-01 -6.65919542e-01
-3.91555429e-01 6.48703992e-01 2.45506808e-01 -5.03420174e-01
4.11086410e-01 3.00375242e-02 6.88916266e-01 1.14575899e+00
1.21131122e+00 -1.23832989e+00 3.78792912e-01 1.28072119e+00
1.26672673e+00 -7.19557226e-01 8.46686661e-02 -2.57167161e-01
-7.04700947e-01 1.33132064e+00 6.28718913e-01 -1.89344093e-01
2.51096576e-01 5.42166173e-01 -1.60175413e-01 3.83626707e-02
-6.48593903e-01 4.72106859e-02 8.14841315e-02 1.00428343e+00
4.42574313e-03 1.97046831e-01 1.65710524e-01 6.32056057e-01
-6.43521369e-01 -2.96846390e-01 7.30982602e-01 1.39056695e+00
-8.60012054e-01 -9.46262419e-01 -4.29336965e-01 9.11575034e-02
3.02226067e-01 5.69631398e-01 -3.59094024e-01 1.08140838e+00
8.17099363e-02 9.90764856e-01 4.16373253e-01 -9.23468113e-01
5.79634428e-01 2.20533460e-03 9.08710718e-01 -8.38324904e-01
-2.13519186e-01 -1.99101165e-01 3.10384661e-01 -1.05437505e+00
3.80625933e-01 -7.27226555e-01 -2.22309422e+00 -2.81238794e-01
-1.66202918e-01 -2.90170878e-01 1.11269593e+00 8.02317798e-01
6.51794910e-01 6.90641224e-01 1.85276598e-01 -1.59176779e+00
-7.37637579e-01 -6.59876525e-01 -1.65189192e-01 1.36559755e-02
3.45510751e-01 -8.87027860e-01 5.99649921e-02 1.38325952e-02] | [4.726498603820801, 0.9186777472496033] |
516cd210-6a41-460a-9326-3677934227d4 | bayesian-analysis-of-dynamic-linear-topic | 1511.03947 | null | http://arxiv.org/abs/1511.03947v1 | http://arxiv.org/pdf/1511.03947v1.pdf | Bayesian Analysis of Dynamic Linear Topic Models | In dynamic topic modeling, the proportional contribution of a topic to a
document depends on the temporal dynamics of that topic's overall prevalence in
the corpus. We extend the Dynamic Topic Model of Blei and Lafferty (2006) by
explicitly modeling document level topic proportions with covariates and
dynamic structure that includes polynomial trends and periodicity. A Markov
Chain Monte Carlo (MCMC) algorithm that utilizes Polya-Gamma data augmentation
is developed for posterior inference. Conditional independencies in the model
and sampling are made explicit, and our MCMC algorithm is parallelized where
possible to allow for inference in large corpora. To address computational
bottlenecks associated with Polya-Gamma sampling, we appeal to the Central
Limit Theorem to develop a Gaussian approximation to the Polya-Gamma random
variable. This approximation is fast and reliable for parameter values relevant
in the text mining domain. Our model and inference algorithm are validated with
multiple simulation examples, and we consider the application of modeling
trends in PubMed abstracts. We demonstrate that sharing information across
documents is critical for accurately estimating document-specific topic
proportions. We also show that explicitly modeling polynomial and periodic
behavior improves our ability to predict topic prevalence at future time
points. | ['Brian Howard', 'Surya T. Tokdar', 'David L. Banks', 'Chris Glynn'] | 2015-11-12 | null | null | null | null | ['dynamic-topic-modeling'] | ['natural-language-processing'] | [ 6.49982691e-02 6.18920103e-02 -5.91910124e-01 -2.67620683e-01
-1.06513965e+00 -5.62026501e-01 8.66194725e-01 5.26183426e-01
-3.43005359e-01 9.78476465e-01 3.70912045e-01 -6.89225852e-01
-2.89543778e-01 -7.73365021e-01 -7.77763128e-01 -5.36184371e-01
-3.73438776e-01 9.90749955e-01 2.41474852e-01 5.06655991e-01
1.98649675e-01 1.61648765e-01 -1.15881681e+00 -1.40371351e-02
8.91107738e-01 9.21097919e-02 3.03574324e-01 9.07539546e-01
-3.01220894e-01 1.99286133e-01 -7.32758820e-01 -2.43455201e-01
-2.97716409e-01 -8.69294181e-02 -6.50113463e-01 -9.20186117e-02
-1.83829218e-01 -1.62936136e-01 1.92411467e-01 7.37328827e-01
4.07595001e-02 1.74109921e-01 1.34258044e+00 -1.35973406e+00
-6.00840151e-02 6.39649391e-01 -1.04111123e+00 4.00344342e-01
1.24295294e-01 -2.48999178e-01 8.36622477e-01 -4.53285247e-01
7.44380474e-01 1.37456322e+00 7.42723048e-01 -1.48146329e-02
-1.49864519e+00 -5.71945786e-01 5.51850975e-01 -3.08179557e-01
-1.40830076e+00 1.21578008e-01 4.18971866e-01 -7.72320151e-01
8.02405715e-01 1.08068012e-01 8.28465223e-01 1.14321995e+00
8.02321970e-01 8.02020729e-01 7.37975240e-01 -3.87897372e-01
5.93229115e-01 6.81840554e-02 5.70280433e-01 1.71577111e-01
4.51908797e-01 -4.89792705e-01 -3.47790062e-01 -1.03151894e+00
5.33790410e-01 2.67012030e-01 -1.78242242e-03 -1.51470795e-01
-8.46768796e-01 1.11872840e+00 -4.59172815e-01 3.94028351e-02
-5.38234472e-01 3.42033237e-01 3.17036092e-01 -9.98285487e-02
1.00474823e+00 -5.21583576e-03 -5.39917469e-01 -5.04922569e-01
-1.32461798e+00 7.31390655e-01 1.14740014e+00 1.08059669e+00
3.37623298e-01 -4.54638451e-01 -2.35623896e-01 6.63715303e-01
4.79050905e-01 6.02569580e-01 2.96238571e-01 -8.59213710e-01
2.00551346e-01 8.81092772e-02 5.19882262e-01 -5.60834706e-01
-2.88782507e-01 -3.81114483e-01 -4.76353198e-01 -4.52895075e-01
6.93223238e-01 -2.10233837e-01 -8.79742980e-01 1.99183202e+00
4.42786396e-01 1.08002007e-01 -2.85273820e-01 7.23226368e-02
-1.63715735e-01 1.06693852e+00 6.65316522e-01 -8.17951739e-01
1.74340248e+00 -2.17549279e-01 -1.01007378e+00 2.60753751e-01
7.30162382e-01 -7.86133349e-01 8.81758749e-01 4.38938022e-01
-1.08776200e+00 1.46647289e-01 -4.32977587e-01 2.51937620e-02
-1.56222239e-01 -2.84497797e-01 7.69145429e-01 6.59485281e-01
-8.17678750e-01 2.82386422e-01 -1.45583379e+00 -4.78763908e-01
3.10404360e-01 6.65599480e-02 4.26534027e-01 -1.05508104e-01
-1.07577324e+00 4.86378342e-01 1.43862262e-01 -5.01786113e-01
-7.05027699e-01 -1.28593373e+00 -3.02925855e-01 3.70746851e-01
1.05873317e-01 -9.41056788e-01 1.45317852e+00 -2.90820181e-01
-1.01356149e+00 2.39944056e-01 -7.41084754e-01 -5.67791462e-01
5.43446243e-01 -1.37402723e-02 -9.22161490e-02 9.53458920e-02
2.86885619e-01 3.42649043e-01 5.39652407e-01 -9.74381745e-01
-6.71604514e-01 -3.58112961e-01 -4.69879627e-01 2.44666323e-01
-3.26154411e-01 2.31167618e-02 -7.46050239e-01 -6.96041167e-01
-8.32791403e-02 -8.46989274e-01 -3.14913154e-01 -2.91455090e-01
-2.19329432e-01 -5.06239235e-01 5.95905960e-01 -6.78937733e-01
1.44961560e+00 -1.75560117e+00 -2.98551500e-01 4.59360182e-01
9.28544477e-02 -6.23320282e-01 3.64345282e-01 6.65708005e-01
3.31948996e-01 1.02192611e-01 -3.99940908e-01 -4.42744076e-01
-7.50633925e-02 1.28473669e-01 -4.12726432e-01 6.79794550e-01
-1.36389107e-01 4.14978594e-01 -6.84490561e-01 -5.15631676e-01
-3.57877135e-01 4.83007997e-01 -7.72073030e-01 -3.02063972e-01
-5.92824042e-01 -4.19129580e-02 -4.55384672e-01 3.22942585e-01
6.83484077e-01 -7.45035231e-01 5.44780970e-01 3.52924049e-01
-2.42361546e-01 3.76440138e-01 -1.13545465e+00 1.18004775e+00
-3.77046525e-01 4.38472092e-01 3.70544903e-02 -6.65321767e-01
4.90132749e-01 3.36680502e-01 5.87186754e-01 2.40599111e-01
-1.53395668e-01 2.32047755e-02 -1.94980830e-01 6.18584156e-02
7.05839753e-01 -2.51373678e-01 -2.14190315e-02 9.59394395e-01
-2.15698823e-01 -1.17626645e-01 3.37739199e-01 5.55379629e-01
1.03742647e+00 -1.42012358e-01 3.03432256e-01 -7.51924038e-01
-3.12333196e-01 2.57434875e-01 4.48785663e-01 1.10423660e+00
2.40934432e-01 1.90046906e-01 9.79411721e-01 -3.40277404e-02
-1.25959957e+00 -1.21053362e+00 -8.29056382e-01 9.93855953e-01
-3.56540293e-01 -4.66896921e-01 -6.61691248e-01 -1.05039380e-01
4.22928929e-02 9.78407204e-01 -7.64704406e-01 2.17607886e-01
-1.46825656e-01 -1.53262448e+00 7.09718019e-02 3.60556096e-01
-1.89576164e-01 -4.25409377e-01 -5.61948597e-01 6.14018798e-01
-2.84244835e-01 -6.37447000e-01 -4.96082038e-01 9.60031375e-02
-1.17548335e+00 -9.62615788e-01 -1.12836528e+00 -1.70579597e-01
5.77502906e-01 5.16298376e-02 9.74719107e-01 -4.46433157e-01
-2.35296473e-01 6.59200072e-01 4.65279855e-02 -7.64461279e-01
-5.23796916e-01 1.53537601e-01 -1.09767996e-01 -5.25321484e-01
4.70789820e-01 -4.93549734e-01 -5.95351279e-01 5.65281101e-02
-9.44902420e-01 -4.83000502e-02 7.40431920e-02 7.26246834e-01
5.34028769e-01 1.21091053e-01 3.62952560e-01 -1.20870864e+00
8.78778100e-01 -1.03243458e+00 -8.03663850e-01 2.73062319e-01
-8.30834508e-01 -1.02513574e-01 -1.91221476e-01 -7.60993361e-01
-1.32555032e+00 -4.36620533e-01 3.49916518e-01 -1.64538380e-02
8.50689858e-02 9.47881162e-01 1.79258376e-01 7.76764572e-01
3.45947206e-01 2.60683126e-03 8.81825611e-02 -5.12377679e-01
1.74691156e-01 5.57260692e-01 2.21073776e-01 -7.34801233e-01
1.32070258e-01 7.37604678e-01 6.75932914e-02 -1.01976097e+00
-6.79142714e-01 -7.12651432e-01 -1.23934798e-01 1.79260820e-02
6.11822009e-01 -1.24158621e+00 -8.89169097e-01 3.53566527e-01
-1.16842699e+00 -4.43077236e-01 -2.16532737e-01 8.94612432e-01
-6.19385064e-01 2.28398502e-01 -6.69495583e-01 -1.28707564e+00
-1.23633347e-01 -9.70029354e-01 1.11027944e+00 8.77935998e-03
-5.85610926e-01 -1.51933241e+00 5.15500903e-01 -7.81405717e-02
1.59745201e-01 -1.20647037e-02 1.21454632e+00 -6.28091037e-01
-4.53664422e-01 -2.67533422e-01 -8.47855732e-02 -6.01985395e-01
2.29095295e-03 4.57550138e-01 -6.32817745e-01 -2.81195015e-01
-2.71093864e-02 4.10400987e-01 7.82776833e-01 1.22936475e+00
9.96108592e-01 -4.47443396e-01 -9.90059555e-01 1.75974697e-01
1.24474061e+00 1.00530230e-01 2.53153384e-01 2.49067873e-01
1.64167419e-01 5.99668086e-01 4.94677573e-01 9.02359426e-01
6.76493406e-01 5.15683651e-01 -3.57110530e-01 1.92166492e-01
4.16461974e-01 -3.54982734e-01 1.26671091e-01 6.35686457e-01
2.75381476e-01 -6.27734244e-01 -1.19446945e+00 9.82736945e-01
-1.75371122e+00 -9.72124994e-01 -4.71209526e-01 2.22195768e+00
1.25702274e+00 3.14684123e-01 5.96456051e-01 -3.49363804e-01
8.90730202e-01 -3.73195678e-01 -5.70351303e-01 -2.37291396e-01
2.03518607e-02 -5.66120893e-02 7.01950550e-01 7.44867086e-01
-7.09350705e-01 6.53221369e-01 7.52744293e+00 9.11505282e-01
-3.55348140e-01 2.82365113e-01 9.19683516e-01 -1.42632931e-01
-6.88885093e-01 1.58961624e-01 -1.19524944e+00 7.33650267e-01
1.54614878e+00 -6.75763607e-01 -1.67617872e-01 6.19664073e-01
3.79222691e-01 -6.95042908e-01 -8.25130224e-01 2.22366035e-01
-3.39614987e-01 -1.18326414e+00 -3.19681276e-04 5.73570192e-01
9.63436723e-01 -9.20118093e-02 1.58184424e-01 6.97492063e-02
1.18830907e+00 -4.14706409e-01 3.04099828e-01 6.79605842e-01
5.36742210e-01 -9.48349476e-01 4.41756278e-01 4.37206060e-01
-7.01900661e-01 9.38930288e-02 -3.21730912e-01 1.75448671e-01
5.34166157e-01 1.18084824e+00 -1.14789701e+00 1.01168372e-01
5.25320828e-01 3.50249171e-01 -2.47477423e-02 1.08588600e+00
2.51733691e-01 1.17356038e+00 -9.51131880e-01 -1.61697000e-01
2.26908538e-04 -1.24004133e-01 5.39268792e-01 1.25254333e+00
4.92619544e-01 -1.77591713e-03 -6.43837750e-02 7.52440631e-01
1.64160028e-01 1.50027245e-01 -2.91854352e-01 -4.75480817e-02
8.11068773e-01 7.37704813e-01 -1.17264867e+00 -5.56205332e-01
-3.87639850e-01 3.78193915e-01 -8.10395777e-02 6.82027340e-01
-4.97712612e-01 -1.60025150e-01 5.29921174e-01 2.56981552e-01
4.09315944e-01 -4.11089063e-01 -5.31490982e-01 -9.48063970e-01
-3.15768212e-01 -4.98000234e-01 6.73840404e-01 -4.24289525e-01
-1.26893497e+00 -2.49046702e-02 7.44257629e-01 -5.91287434e-01
-5.68583012e-01 -2.00501621e-01 -6.92295611e-01 1.12369025e+00
-1.13027418e+00 -8.77692401e-01 4.47565347e-01 3.85302931e-01
4.50997084e-01 2.75295109e-01 6.02153361e-01 -2.25668594e-01
-3.29640478e-01 3.76692295e-01 6.55256927e-01 -6.57134473e-01
5.16301632e-01 -1.23917627e+00 5.75813651e-01 5.25490701e-01
-3.08638662e-01 1.16579199e+00 1.17748272e+00 -1.35758841e+00
-1.12273884e+00 -8.23508561e-01 8.68774652e-01 -5.08630276e-01
9.80022728e-01 -4.69891846e-01 -1.04533923e+00 8.88217032e-01
-1.66404903e-01 -8.38895857e-01 1.05319715e+00 7.90314615e-01
-6.89828470e-02 4.39186335e-01 -9.39816594e-01 6.38229847e-01
3.57432455e-01 -2.10006386e-01 -3.00522357e-01 7.46690750e-01
7.51791060e-01 -5.47563359e-02 -1.10714400e+00 2.67053172e-02
6.24478996e-01 -1.51027173e-01 6.56959474e-01 -6.28568590e-01
2.79978961e-01 8.39468651e-03 5.48141375e-02 -1.07265079e+00
-1.95523888e-01 -7.61099041e-01 -3.36094201e-01 1.14304173e+00
5.27277827e-01 -4.08094227e-01 8.67688894e-01 9.64684010e-01
2.57256627e-01 -3.76597196e-01 -9.93693650e-01 -6.76650405e-01
5.49625516e-01 -4.72066730e-01 4.59764153e-01 7.34549582e-01
2.66644448e-01 -1.91136934e-02 -1.14113241e-01 1.75533071e-01
8.07436526e-01 2.10478306e-01 6.40814960e-01 -1.46299839e+00
-3.94601464e-01 -1.69586539e-01 1.80276275e-01 -1.05855751e+00
-4.52322811e-02 -4.42570657e-01 -7.28065223e-02 -1.35177159e+00
8.56303096e-01 -2.99892575e-01 2.58975983e-01 1.94806159e-02
-3.56034517e-01 -3.08606923e-01 -3.32240462e-01 5.05806267e-01
-3.56321424e-01 4.49468672e-01 8.98724496e-01 9.47328955e-02
-4.37057376e-01 3.95174384e-01 -5.26338816e-01 6.61449850e-01
5.77823877e-01 -7.17629850e-01 -5.07690609e-01 8.09923708e-02
3.86967957e-01 3.76721859e-01 6.44057468e-02 -2.99494445e-01
4.21649426e-01 -2.27516472e-01 2.84792632e-01 -1.14098883e+00
2.84438461e-01 -4.36977297e-01 4.25974548e-01 4.45668906e-01
-5.71535051e-01 1.20263018e-01 3.70975286e-01 1.18353271e+00
2.00039208e-01 -2.34411269e-01 4.69278544e-01 -1.40494918e-02
3.93470794e-01 2.03280658e-01 -1.10244370e+00 7.42774829e-02
8.05013955e-01 2.11722001e-01 -2.32811391e-01 -7.55326867e-01
-8.58774006e-01 4.61931586e-01 3.37111056e-01 -1.38839588e-01
6.94307610e-02 -8.74480009e-01 -5.61718762e-01 -2.84266621e-01
-1.49863213e-01 -6.78839087e-02 4.50605750e-01 8.03314090e-01
-1.91366911e-01 5.09774029e-01 4.24234271e-01 -7.70107985e-01
-1.19045150e+00 4.34825271e-01 -2.22770274e-01 -6.56054556e-01
-3.68738323e-01 5.16126692e-01 2.98064679e-01 -4.73718159e-02
2.53917724e-01 -2.21023619e-01 1.83992326e-01 1.96771502e-01
6.62552714e-01 4.83348787e-01 -3.28762621e-01 1.54615596e-01
-1.38453022e-01 3.62814888e-02 -6.59740627e-01 -8.57546270e-01
1.46331036e+00 -3.58710885e-01 -2.34823022e-02 9.70273018e-01
9.52540874e-01 -1.86079413e-01 -1.34613502e+00 -2.38246769e-01
8.34451541e-02 1.67368557e-02 1.18838862e-01 -6.29228175e-01
-2.42293730e-01 6.61651790e-01 1.74963027e-01 3.74078304e-01
5.76095998e-01 1.21095635e-01 3.52830142e-01 3.42873521e-02
1.22307718e-01 -9.87840235e-01 -3.58540326e-01 3.48516971e-01
4.19005424e-01 -6.70264721e-01 5.02170265e-01 -4.01107073e-01
-2.55393118e-01 8.58954787e-01 6.62652180e-02 3.12815487e-01
1.34016716e+00 4.80368823e-01 -5.03110886e-01 -3.55534792e-01
-1.15653312e+00 3.90253842e-01 1.02895074e-01 2.64716268e-01
4.68532056e-01 1.59855276e-01 -7.49770403e-01 4.68892127e-01
-2.15670004e-01 4.89341728e-02 7.61945963e-01 9.67287362e-01
-3.95320863e-01 -9.98638451e-01 -4.63432908e-01 7.75363684e-01
-9.24314022e-01 -2.94668376e-01 2.32838184e-01 8.72475982e-01
-5.99450827e-01 7.09451020e-01 7.79589176e-01 5.67594528e-01
-3.40736777e-01 3.99111629e-01 2.15645581e-01 -6.09062135e-01
-5.12870066e-02 9.28226411e-01 -6.55071437e-02 5.18363789e-02
-1.92916304e-01 -1.47943866e+00 -8.65512908e-01 -4.72460628e-01
-4.60918009e-01 4.94066775e-01 1.04757595e+00 7.73885667e-01
3.08739543e-01 4.69180167e-01 3.08780938e-01 -2.55304754e-01
-4.45338905e-01 -1.13271701e+00 -8.42241824e-01 -1.16671257e-01
2.92723682e-02 -6.51231766e-01 -4.90528286e-01 3.03332448e-01] | [10.269942283630371, 6.903753280639648] |
e6a82332-517b-422c-95f4-bebc18cc0e2c | deep-clustering-with-a-constraint-for | 2303.03036 | null | https://arxiv.org/abs/2303.03036v1 | https://arxiv.org/pdf/2303.03036v1.pdf | Deep Clustering with a Constraint for Topological Invariance based on Symmetric InfoNCE | We consider the scenario of deep clustering, in which the available prior knowledge is limited. In this scenario, few existing state-of-the-art deep clustering methods can perform well for both non-complex topology and complex topology datasets. To address the problem, we propose a constraint utilizing symmetric InfoNCE, which helps an objective of deep clustering method in the scenario train the model so as to be efficient for not only non-complex topology but also complex topology datasets. Additionally, we provide several theoretical explanations of the reason why the constraint can enhances performance of deep clustering methods. To confirm the effectiveness of the proposed constraint, we introduce a deep clustering method named MIST, which is a combination of an existing deep clustering method and our constraint. Our numerical experiments via MIST demonstrate that the constraint is effective. In addition, MIST outperforms other state-of-the-art deep clustering methods for most of the commonly used ten benchmark datasets. | ['Takafumi Kanamori', 'Yusaku Hino', 'Kaito Goto', 'Hiroki Waida', 'Yuichiro Wada', 'Yuhui Zhang'] | 2023-03-06 | null | null | null | null | ['deep-clustering', 'deep-clustering'] | ['miscellaneous', 'natural-language-processing'] | [-7.72479951e-01 -5.16275585e-01 2.04082385e-01 -2.67196625e-01
-1.04565300e-01 -4.88390267e-01 2.27751344e-01 -1.98903337e-01
-3.21215838e-01 5.33143520e-01 -9.29328054e-03 -1.68699339e-01
-4.47872311e-01 -7.73866236e-01 -5.29105902e-01 -1.06796229e+00
-2.81704813e-01 8.95872474e-01 2.52648592e-01 4.59075458e-02
2.18870595e-01 3.94443601e-01 -1.10687327e+00 -6.94376975e-02
1.04833484e+00 8.43985736e-01 3.66466492e-01 1.23711899e-01
-1.19353697e-01 2.19393402e-01 -7.77274251e-01 -9.76362154e-02
3.91810775e-01 -1.32607907e-01 -7.75015116e-01 1.41979009e-01
-6.16125204e-02 -1.17115773e-01 -4.92579222e-01 8.63229990e-01
7.37983108e-01 2.71289408e-01 7.98164904e-01 -1.37146389e+00
-4.94057119e-01 8.87207627e-01 -9.70950365e-01 1.19762786e-01
-2.42248863e-01 1.23321088e-02 5.70912182e-01 -7.17153609e-01
4.10281658e-01 1.14645517e+00 8.33902121e-01 1.92425206e-01
-8.73923540e-01 -8.77718687e-01 3.52913827e-01 2.81057239e-01
-1.86230862e+00 -8.04074928e-02 1.05932426e+00 -3.03707957e-01
3.76829743e-01 -2.61676639e-01 4.99829262e-01 9.28030074e-01
-1.38696551e-01 7.26266444e-01 9.11594152e-01 1.15225606e-01
4.65563148e-01 -2.52523363e-01 4.37259562e-02 5.81347704e-01
3.80790353e-01 -1.80572554e-01 -2.68324632e-02 1.40779510e-01
8.29872668e-01 9.31673795e-02 -8.04440677e-02 -5.48617780e-01
-1.30256379e+00 8.51155519e-01 7.59410858e-01 5.62690794e-01
-1.39460891e-01 3.44434887e-01 3.33258510e-01 -7.90914446e-02
2.10526630e-01 3.76642406e-01 -4.33865428e-01 -8.61146301e-02
-1.15024984e+00 -6.36032671e-02 6.82731509e-01 9.12991226e-01
5.71068347e-01 2.42226869e-02 2.56686211e-01 7.29997516e-01
3.66959602e-01 8.62182602e-02 2.38596499e-01 -9.55535412e-01
4.23525959e-01 8.61674011e-01 -2.77188212e-01 -1.12618577e+00
-7.24505723e-01 -7.98168659e-01 -1.57773793e+00 -2.98096567e-01
3.94668609e-01 -4.26142514e-01 -8.76513660e-01 1.86476350e+00
4.88616198e-01 5.69522440e-01 -8.75021070e-02 1.04213417e+00
7.29101896e-01 6.39368415e-01 -4.88789529e-01 -1.82329878e-01
9.14191902e-01 -8.86198401e-01 -7.22961366e-01 3.43115330e-01
5.96594512e-01 -6.80899620e-01 1.04349422e+00 3.40838253e-01
-5.93940377e-01 -5.18745124e-01 -1.00848532e+00 1.36449575e-01
-6.34572983e-01 2.86754757e-01 8.74187946e-01 4.22266841e-01
-1.16151428e+00 5.00390589e-01 -1.11860502e+00 -5.86619496e-01
3.83949637e-01 5.96663415e-01 -1.64230749e-01 -1.47726625e-01
-1.02877629e+00 3.03521067e-01 5.44035196e-01 3.11413884e-01
-8.64008188e-01 -4.64337200e-01 -4.22610879e-01 2.80403823e-01
5.39793134e-01 -5.71825802e-01 5.84571838e-01 -5.96941888e-01
-1.43452036e+00 4.16000485e-01 9.38209668e-02 -9.45986360e-02
6.28019273e-01 -2.06021238e-02 -1.57106921e-01 2.30480209e-01
-1.22190371e-01 6.51104212e-01 2.39607781e-01 -1.57593131e+00
-2.01072738e-01 -4.29483801e-01 3.33719477e-02 1.15758710e-01
-4.91335034e-01 -2.85524487e-01 -9.67835844e-01 -6.12567842e-01
5.53661585e-02 -8.93653810e-01 -4.55760926e-01 -1.86235607e-02
-9.41646039e-01 -4.04962093e-01 1.33602011e+00 -8.45719129e-02
1.34213579e+00 -2.07759929e+00 2.31206819e-01 5.35142124e-01
3.32997113e-01 2.34123960e-01 -4.60030548e-02 4.89521205e-01
2.13407382e-01 6.02396607e-01 -3.55481654e-01 -4.56980973e-01
1.76768526e-01 3.98217946e-01 2.18620911e-01 5.84319234e-01
-5.34703396e-02 6.66459262e-01 -6.36511147e-01 -7.82181859e-01
4.52060193e-01 6.13538384e-01 -5.21675646e-01 1.80858567e-01
3.74058262e-02 5.60907722e-01 -5.67148447e-01 5.23146927e-01
9.81237710e-01 -4.47524935e-01 4.27920163e-01 -3.59307855e-01
-1.41777143e-01 -2.66122162e-01 -1.43703425e+00 1.95314360e+00
-1.11265272e-01 5.84426045e-01 3.05587351e-01 -1.42740357e+00
8.68745506e-01 3.83507833e-02 7.55252600e-01 -3.57622862e-01
3.66924673e-01 -3.96099016e-02 3.54018301e-01 -2.68854797e-01
1.92302525e-01 1.56325743e-01 9.44960713e-02 5.90009212e-01
-2.42822915e-01 3.19459230e-01 3.57738733e-01 5.02269566e-01
1.06904972e+00 -2.88696498e-01 -3.21866006e-01 -7.78117478e-01
2.70937562e-01 -2.48615652e-01 8.22771013e-01 7.62028515e-01
-2.97848821e-01 6.77528858e-01 5.50977468e-01 -3.52986634e-01
-9.34897006e-01 -8.75484705e-01 -1.50508180e-01 6.67304933e-01
6.12129450e-01 -3.68385643e-01 -1.03689015e+00 -6.74990892e-01
-7.68487155e-02 9.59631149e-03 -7.12741792e-01 -1.47122117e-02
-6.31461322e-01 -1.07811928e+00 7.62531281e-01 7.15948343e-01
1.17656326e+00 -6.34336174e-01 -3.87266517e-01 1.06610693e-01
-3.27966303e-01 -1.23198688e+00 -3.40773731e-01 8.65906626e-02
-7.31078804e-01 -1.39746451e+00 -5.90772688e-01 -1.01924860e+00
6.67805731e-01 3.87082517e-01 1.09403849e+00 6.93237603e-01
-1.64540038e-01 -7.56442696e-02 -4.28071767e-01 1.94209427e-01
-3.50966342e-02 4.83667582e-01 2.41591319e-01 -2.10863352e-01
3.51426303e-01 -8.38543594e-01 -8.83296847e-01 7.06235170e-01
-9.39829588e-01 -2.12474614e-02 6.45686567e-01 5.56018889e-01
4.95406747e-01 8.55084181e-01 6.02077127e-01 -5.17654657e-01
6.46110833e-01 -5.80749810e-01 -4.72694188e-01 1.98107839e-01
-3.46819133e-01 3.04395761e-02 1.02251983e+00 -6.19064629e-01
-7.80628622e-01 1.03221349e-01 2.38459427e-02 -6.66667938e-01
-1.71103209e-01 7.26379395e-01 -4.54859853e-01 1.37381569e-01
1.75522625e-01 -2.13463604e-03 -1.60450518e-01 -7.02837586e-01
2.31451556e-01 5.97614050e-01 5.21051347e-01 -9.34608817e-01
8.49346578e-01 7.60607541e-01 2.49075532e-01 -6.29008472e-01
-3.77276510e-01 -4.24811453e-01 -1.10697544e+00 -7.90301934e-02
9.69580591e-01 -6.75313592e-01 -9.43732381e-01 6.40942693e-01
-9.66080666e-01 -4.98541921e-01 4.62860197e-01 3.19244355e-01
-2.52445579e-01 5.46785712e-01 -7.27138817e-01 -4.89132643e-01
-1.18693836e-01 -1.45861232e+00 8.04126441e-01 7.68314376e-02
3.29146862e-01 -1.36613321e+00 -2.34259337e-01 7.17021450e-02
3.52359623e-01 6.72932565e-01 8.47360194e-01 -6.48112655e-01
-5.49710095e-01 2.65760481e-01 -5.31895757e-01 9.94825512e-02
1.31825745e-01 5.56300044e-01 -4.51270133e-01 -4.79075760e-01
-2.36496508e-01 -2.36173898e-01 8.85227442e-01 4.71633255e-01
1.82160699e+00 -1.09859137e-03 -6.46928728e-01 8.97566259e-01
1.61310112e+00 3.20488334e-01 6.98351204e-01 1.17195331e-01
9.83361244e-01 3.23755771e-01 3.14580739e-01 4.63949770e-01
5.90695381e-01 6.94388986e-01 5.61249912e-01 -4.44585025e-01
7.16758296e-02 6.30898476e-02 -2.28000209e-01 1.29690909e+00
1.25744447e-01 -5.02326906e-01 -1.34599829e+00 6.64697289e-01
-2.05645323e+00 -6.67019010e-01 -1.99291125e-01 1.61280632e+00
5.91733694e-01 9.34791192e-02 1.07689738e-01 2.74099022e-01
1.06968045e+00 -1.35947555e-01 -7.68367171e-01 1.32163450e-01
-6.16506673e-02 -1.75123587e-01 2.02118322e-01 1.48295134e-01
-1.20734763e+00 9.25147891e-01 6.22386789e+00 9.62511837e-01
-1.06577146e+00 -1.68075189e-01 6.80044949e-01 1.48116186e-01
-2.63104513e-02 -1.60224438e-01 -3.40574563e-01 8.12739909e-01
5.15140057e-01 5.34717552e-02 3.79029512e-01 6.54200613e-01
4.13147539e-01 -5.78326173e-02 -1.27598131e+00 9.84850705e-01
-1.74440891e-01 -1.48233259e+00 -7.85839483e-02 3.57345730e-01
7.34748125e-01 4.80995215e-02 7.88941234e-02 3.17921281e-01
6.15781188e-01 -1.16267717e+00 2.35947639e-01 1.33111343e-01
7.34809279e-01 -1.02795720e+00 1.06456745e+00 3.95408273e-01
-1.46099150e+00 -1.18562981e-01 -5.04461169e-01 3.34207341e-02
3.97234596e-02 9.32645559e-01 -4.35416311e-01 8.35415125e-01
9.18100715e-01 9.49397504e-01 -5.65254807e-01 1.21347415e+00
7.10074455e-02 7.53452957e-01 -8.40211034e-01 1.50029778e-01
5.60133457e-01 -2.51710624e-01 1.95885271e-01 1.25456667e+00
3.72795105e-01 1.77352831e-01 5.10174334e-01 1.16026807e+00
-3.23492318e-01 -2.58540720e-01 -5.35207689e-01 1.20425627e-01
1.02327383e+00 1.19327211e+00 -1.27169430e+00 -3.00798714e-01
-3.09629571e-02 6.45377874e-01 3.25486362e-01 5.56959748e-01
-1.01780474e+00 -4.75287199e-01 5.71545601e-01 -1.08721904e-01
3.72944295e-01 -5.52870750e-01 -4.92608130e-01 -1.10834801e+00
-3.59835811e-02 -6.31846845e-01 3.27177703e-01 -5.91579676e-01
-1.38309944e+00 3.62781554e-01 -3.09127513e-02 -1.02692473e+00
3.03188026e-01 -4.38385457e-01 -1.09357500e+00 1.89307928e-01
-1.28829753e+00 -1.06144369e+00 -6.41118646e-01 9.32871938e-01
3.86294633e-01 -1.10995978e-01 3.58096838e-01 7.93002427e-01
-1.16689384e+00 5.90530396e-01 4.79237288e-01 5.08385956e-01
5.62790751e-01 -1.23676872e+00 1.70902431e-01 9.28916991e-01
-1.60281375e-01 8.60707581e-01 3.88720602e-01 -4.61588889e-01
-1.28978610e+00 -1.18314266e+00 -1.49171025e-01 -9.71651226e-02
4.50280815e-01 -6.44900620e-01 -8.52235675e-01 2.61285603e-01
1.54620558e-01 -4.34370525e-03 5.87931931e-01 4.52608615e-02
-1.65056735e-01 -1.96243063e-01 -1.16038132e+00 3.93995196e-01
1.21226871e+00 7.14797759e-03 -2.31765479e-01 3.16673964e-01
8.81511092e-01 -2.27343410e-01 -9.54841793e-01 5.87431431e-01
2.42550239e-01 -1.12470794e+00 9.17488277e-01 -2.13678733e-01
4.60522175e-01 -6.73649192e-01 -1.82325274e-01 -1.45031250e+00
-3.75425458e-01 -4.83350039e-01 -1.10491626e-01 1.35545945e+00
1.99658852e-02 -3.08630228e-01 8.75563323e-01 1.42879009e-01
-3.11265230e-01 -8.74959826e-01 -1.02055335e+00 -9.22138214e-01
5.94809294e-01 -7.95985162e-02 1.04745984e+00 1.52229965e+00
-3.40392381e-01 1.29598260e-01 -1.00008860e-01 1.38689414e-01
6.54199302e-01 8.52193907e-02 8.53530288e-01 -1.61277926e+00
1.04443088e-01 -4.83843505e-01 -3.39103609e-01 -1.19628966e+00
3.03791225e-01 -6.20186210e-01 -2.63542861e-01 -1.87395370e+00
3.74983102e-01 -8.34236681e-01 -3.00840795e-01 3.42742562e-01
-2.08810717e-02 1.92377623e-02 1.84515432e-01 3.03117305e-01
-8.80405545e-01 8.00852418e-01 1.32413423e+00 -1.16989031e-01
-3.27361524e-02 -3.22584927e-01 -7.79269755e-01 6.97654903e-01
9.93731320e-01 -3.09293628e-01 -6.03271842e-01 -8.05158496e-01
7.06102178e-02 -3.42319757e-01 1.32943988e-01 -1.28884256e+00
4.65878159e-01 -1.79568484e-01 4.43671405e-01 -1.03478432e+00
1.52686477e-01 -1.11848283e+00 5.81470057e-02 3.94603133e-01
2.93261223e-02 3.24843466e-01 9.40777361e-02 5.98263800e-01
-3.27276111e-01 5.07629454e-01 8.96221459e-01 -3.91080379e-02
-5.90414345e-01 5.75550199e-01 -2.20892251e-01 2.56092578e-01
1.12588441e+00 -3.16695035e-01 -5.77739298e-01 -4.03630048e-01
-5.21092653e-01 9.12996352e-01 6.37922585e-01 2.19790250e-01
5.15446961e-01 -1.42545259e+00 -5.00666559e-01 -1.17313221e-01
-3.71223807e-01 5.41710079e-01 1.09915257e-01 1.02679300e+00
-7.68487453e-01 1.98043033e-01 -1.31750837e-01 -8.26728642e-01
-8.31834435e-01 8.10389578e-01 5.29501200e-01 -1.75050169e-01
-4.85306680e-01 6.16278529e-01 2.82696784e-01 -7.42719412e-01
6.61300719e-01 -3.26687157e-01 8.95848405e-03 -2.34096348e-01
1.68980926e-01 5.48952341e-01 -7.57838190e-02 -4.03383583e-01
-5.88579297e-01 7.68271387e-01 5.27335797e-03 2.45978460e-01
1.37999117e+00 -3.06258619e-01 -2.65447110e-01 1.14345588e-01
1.25905252e+00 -4.14276272e-01 -1.19660616e+00 -3.14864404e-02
-1.76597103e-01 -2.21102580e-01 3.31894457e-02 -7.58111000e-01
-1.82130730e+00 1.35303593e+00 5.72039247e-01 1.66539520e-01
1.08989108e+00 -1.00984871e-01 9.26651895e-01 5.31527817e-01
3.02562028e-01 -1.37442076e+00 4.27229166e-01 5.91158450e-01
5.31426549e-01 -1.13742638e+00 1.67658910e-01 -4.16175216e-01
-4.18642312e-01 1.08392990e+00 1.07845485e+00 -4.26799506e-02
1.01988411e+00 3.91450733e-01 8.82860348e-02 -4.95305121e-01
-6.34037435e-01 -1.31739855e-01 -1.32249564e-01 7.57082343e-01
3.22579682e-01 9.99646634e-02 -1.44954443e-01 4.67923254e-01
-2.48148099e-01 -3.19589019e-01 5.60094655e-01 5.29725194e-01
-2.62315392e-01 -8.43363702e-01 -2.75058091e-01 2.96007097e-01
-4.16706502e-01 1.03983775e-01 -8.25247109e-01 1.11442935e+00
2.58792967e-01 1.22934592e+00 3.56513739e-01 -6.51272058e-01
-1.80127695e-01 -5.60631275e-01 -7.50599653e-02 -1.34595945e-01
-4.58218008e-01 2.13904187e-01 -2.99869448e-01 -5.09302199e-01
-5.24301052e-01 -1.01447940e-01 -1.61102724e+00 -8.20006847e-01
-4.72338766e-01 3.29892784e-01 6.25603437e-01 9.07712221e-01
5.94884157e-01 4.93500173e-01 6.64417505e-01 -8.51471186e-01
-1.35206640e-01 -8.94369125e-01 -6.75896287e-01 4.87015575e-01
-1.33758225e-02 -1.14564407e+00 -4.34281051e-01 -3.93859655e-01] | [9.076977729797363, 3.3687026500701904] |
33d26b30-7848-4bbc-877d-fcf2366a54e7 | viewnet-a-novel-projection-based-backbone | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chen_ViewNet_A_Novel_Projection-Based_Backbone_With_View_Pooling_for_Few-Shot_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chen_ViewNet_A_Novel_Projection-Based_Backbone_With_View_Pooling_for_Few-Shot_CVPR_2023_paper.pdf | ViewNet: A Novel Projection-Based Backbone With View Pooling for Few-Shot Point Cloud Classification | Although different approaches have been proposed for 3D point cloud-related tasks, few-shot learning (FSL) of 3D point clouds still remains under-explored. In FSL, unlike traditional supervised learning, the classes of training and test data do not overlap, and a model needs to recognize unseen classes from only a few samples. Existing FSL methods for 3D point clouds employ point-based models as their backbone. Yet, based on our extensive experiments and analysis, we first show that using a point-based backbone is not the most suitable FSL approach, since (i) a large number of points' features are discarded by the max pooling operation used in 3D point-based backbones, decreasing the ability of representing shape information; (ii)point-based backbones are sensitive to occlusion. To address these issues, we propose employing a projection- and 2D Convolutional Neural Network-based backbone, referred to as the ViewNet, for FSL from 3D point clouds. Our approach first projects a 3D point cloud onto six different views to alleviate the issue of missing points. Also, to generate more descriptive and distinguishing features, we propose View Pooling, which combines different projected plane combinations into five groups and performs max-pooling on each of them. The experiments performed on the ModelNet40, ScanObjectNN and ModelNet40-C datasets, with cross validation, show that our method consistently outperforms the state-of-the-art baselines. Moreover, compared to traditional image classification backbones, such as ResNet, the proposed ViewNet can extract more distinguishing features from multiple views of a point cloud. We also show that ViewNet can be used as a backbone with different FSL heads and provides improved performance compared to traditionally used backbones. | ['Senem Velipasalar', 'Minmin Yang', 'Jiajing Chen'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['few-shot-point-cloud-classification', 'point-cloud-classification'] | ['computer-vision', 'computer-vision'] | [-1.41509771e-01 -1.40165329e-01 -1.76866397e-01 -4.30983484e-01
-5.45645058e-01 -5.39210081e-01 6.98980093e-01 -5.92663251e-02
-3.03677190e-03 1.05066232e-01 -9.66536626e-02 -4.52687517e-02
5.58199994e-02 -1.09846008e+00 -9.08173978e-01 -5.80699921e-01
5.18415980e-02 3.68120581e-01 6.87462389e-01 -1.63404882e-01
4.00669008e-01 1.06213057e+00 -1.99931884e+00 3.50530088e-01
6.00401461e-01 1.16273749e+00 1.91027731e-01 1.43252071e-02
-5.98902822e-01 2.41669357e-01 -4.52085614e-01 -1.60581172e-01
6.57566249e-01 1.92835823e-01 -5.43736696e-01 1.55416831e-01
8.76901925e-01 -3.97239596e-01 -3.49703848e-01 9.72949564e-01
5.56522429e-01 1.53018296e-01 6.48893416e-01 -1.52224970e+00
-6.15318596e-01 -1.27580732e-01 -6.39964581e-01 -5.28813936e-02
3.35285604e-01 1.70128763e-01 9.18003023e-01 -1.43832397e+00
8.29245329e-01 1.43852234e+00 7.66344011e-01 6.18365884e-01
-1.13089466e+00 -8.87152791e-01 3.10828358e-01 7.26295561e-02
-1.26330936e+00 -1.41578421e-01 1.10233867e+00 -3.86979967e-01
1.26463020e+00 1.11150019e-01 9.51685786e-01 1.20774519e+00
1.53804764e-01 8.93756866e-01 1.01689482e+00 -6.64848611e-02
3.32421750e-01 -1.20981745e-02 1.59361899e-01 6.46728277e-01
5.75613640e-02 2.03762189e-01 -4.56502020e-01 -3.44263554e-01
9.37740088e-01 5.12191594e-01 -4.12936091e-01 -1.10984647e+00
-1.10235226e+00 8.40935290e-01 8.09894323e-01 1.02762701e-02
-1.54837340e-01 -7.96662942e-02 2.87278473e-01 2.36454710e-01
7.67224729e-01 1.97474539e-01 -3.31123084e-01 3.07958663e-01
-8.87292325e-01 2.93877065e-01 6.12935662e-01 1.30420458e+00
9.37808812e-01 -2.27048486e-01 1.32401556e-01 7.61725605e-01
2.60876209e-01 5.38704991e-01 1.82355627e-01 -7.43095398e-01
5.93074918e-01 1.01495314e+00 -1.61168709e-01 -8.69201303e-01
-3.18310857e-01 -3.11411917e-01 -8.19383144e-01 7.76774585e-01
-8.65890160e-02 3.66796523e-01 -1.42589784e+00 1.32163024e+00
3.05289745e-01 4.25063252e-01 -1.37678469e-02 8.60751748e-01
1.32258546e+00 6.14055693e-01 -2.83177197e-01 2.39370212e-01
1.08293664e+00 -6.84214294e-01 -7.06393644e-02 -9.10878107e-02
3.62957269e-01 -5.49209058e-01 1.02272403e+00 1.41630590e-01
-9.92028713e-01 -7.15178490e-01 -1.24414659e+00 -1.30871728e-01
-6.43118978e-01 -2.89460510e-01 5.95104694e-01 3.00078988e-01
-9.91165340e-01 7.28145242e-01 -9.45631921e-01 -5.64275801e-01
8.22835624e-01 3.01167339e-01 -7.00227857e-01 -3.40932786e-01
-7.19673574e-01 7.72777319e-01 2.58912206e-01 -1.73156723e-01
-7.88930357e-01 -9.85473812e-01 -1.00031471e+00 2.09532678e-01
2.33897269e-01 -9.14826393e-01 8.03520203e-01 -2.70435244e-01
-1.24730086e+00 1.09570837e+00 -1.89991146e-01 -9.59380269e-02
3.63854527e-01 -1.65205136e-01 -1.10925198e-01 2.18947053e-01
2.19949082e-01 9.57063913e-01 7.83315420e-01 -1.67482471e+00
-5.91714680e-01 -5.35094321e-01 3.10094416e-01 2.77613819e-01
7.13553354e-02 -3.16097230e-01 -6.68102145e-01 -3.21761459e-01
8.36677372e-01 -1.00296891e+00 -2.36752823e-01 3.02355409e-01
-3.46599698e-01 -4.45039839e-01 1.13309646e+00 1.91150792e-02
5.00598848e-01 -2.30033350e+00 -5.11847697e-02 1.68459252e-01
2.95984536e-01 2.49481797e-01 -1.25066668e-01 4.16563511e-01
-2.53928751e-01 2.18007684e-01 -1.57300249e-01 -4.50138867e-01
-1.39516860e-01 2.97947317e-01 -3.74358207e-01 4.51278359e-01
3.77297461e-01 8.88010561e-01 -7.52652824e-01 -2.79932261e-01
7.40502417e-01 4.32288855e-01 -5.42227089e-01 4.92185168e-03
-1.42078966e-01 2.74699092e-01 -3.09550315e-01 8.83906126e-01
1.14488208e+00 -3.68697703e-01 -5.20758808e-01 -2.24621281e-01
-4.03775647e-02 1.18333846e-01 -9.95913267e-01 2.06703401e+00
-4.06264663e-01 2.76937693e-01 -3.90310317e-01 -8.21821928e-01
1.21935380e+00 3.09656709e-01 6.67087018e-01 -2.85014123e-01
-5.13375551e-02 9.80962440e-02 -3.84506047e-01 -3.47998261e-01
7.71873668e-02 -1.47901967e-01 6.76328549e-03 1.27523333e-01
4.51461792e-01 -4.76578742e-01 -2.96794981e-01 6.22657686e-03
1.13799524e+00 3.53380471e-01 1.56403095e-01 7.18221664e-02
4.38377559e-01 -8.07536766e-02 6.30084395e-01 6.75193131e-01
-1.88744903e-01 1.22781098e+00 2.82895267e-01 -7.03220725e-01
-9.77530241e-01 -1.37654245e+00 -2.66707569e-01 4.57804054e-01
5.59435427e-01 -2.91385204e-01 -1.65996745e-01 -1.08045447e+00
2.68525451e-01 7.09013700e-01 -4.25034940e-01 -1.00470513e-01
-4.67471302e-01 -3.16010118e-01 6.36425018e-02 5.38116574e-01
5.55026531e-01 -8.69455993e-01 -5.75290322e-01 -3.88829336e-02
1.43221006e-01 -1.28777945e+00 7.04496820e-03 2.22345561e-01
-1.28342843e+00 -1.18064499e+00 -6.96315825e-01 -7.56200254e-01
7.71986902e-01 8.98325682e-01 1.20825434e+00 -1.56686515e-01
1.41628876e-01 5.60730815e-01 -4.42909420e-01 -7.57171810e-01
7.72756413e-02 9.33422819e-02 -7.68975541e-02 -2.09046170e-01
8.86280835e-01 -1.05644858e+00 -6.08034015e-01 3.48788679e-01
-7.07755327e-01 2.52908673e-02 6.16326034e-01 7.35221624e-01
8.15864384e-01 -2.18330577e-01 2.40248129e-01 -7.94761240e-01
1.24031626e-01 -3.86639714e-01 -5.39993405e-01 9.78839844e-02
-2.83215493e-01 -2.08246320e-01 3.50899726e-01 -2.22207338e-01
-8.19171667e-01 1.58707336e-01 -1.67158589e-01 -1.40783060e+00
-5.59094012e-01 1.06919959e-01 -2.66509086e-01 -4.00922805e-01
6.38783038e-01 2.20289916e-01 1.28175870e-01 -5.57317972e-01
3.48310441e-01 4.01149035e-01 1.37201875e-01 -1.98816642e-01
1.06891477e+00 9.28935409e-01 1.28035143e-01 -9.24816728e-01
-7.49508321e-01 -7.91491508e-01 -9.17447925e-01 -1.43773079e-01
9.22436357e-01 -1.01614165e+00 -5.21061301e-01 2.80094624e-01
-1.44523931e+00 3.41998786e-01 -3.71914685e-01 5.57708144e-01
-6.93768203e-01 2.66465008e-01 -3.39336544e-01 -5.24565518e-01
-2.67905563e-01 -1.23433256e+00 1.48411250e+00 1.41159445e-01
9.40064192e-02 -6.01069033e-01 -1.99766085e-02 6.20144680e-02
3.39928567e-02 4.88308966e-01 1.00409245e+00 -7.62177885e-01
-9.86709833e-01 -4.15531069e-01 -2.37477452e-01 4.36960638e-01
2.04088669e-02 -1.70799106e-01 -1.27012432e+00 -4.19917434e-01
2.37394497e-01 -2.49966636e-01 9.33175325e-01 3.52309763e-01
1.31805885e+00 3.20458770e-01 -5.87240934e-01 9.74025786e-01
1.56269634e+00 1.26077443e-01 6.98237658e-01 1.91441596e-01
7.90342987e-01 5.24596214e-01 3.60190332e-01 1.69415966e-01
2.09331781e-01 5.15959620e-01 8.48753273e-01 -2.07957581e-01
-6.12966530e-03 -3.30677330e-01 3.99704874e-02 8.60439241e-01
-3.36508274e-01 -4.63111177e-02 -8.70670855e-01 3.84674370e-01
-1.70918036e+00 -9.73369837e-01 6.36691749e-02 2.28615427e+00
-1.82885602e-02 3.86191726e-01 -1.81625679e-01 -3.70434448e-02
6.08539522e-01 3.82218510e-01 -6.63293600e-01 -2.19956003e-02
-5.95376231e-02 4.85312223e-01 5.52954793e-01 -6.77291006e-02
-1.21803463e+00 8.34042847e-01 5.83865356e+00 6.37216330e-01
-1.26679385e+00 5.41139171e-02 8.73577744e-02 -3.03324342e-01
-2.49477550e-01 1.15009733e-01 -8.94355953e-01 2.99113601e-01
2.06655994e-01 1.47442684e-01 -7.17343464e-02 1.21378624e+00
-1.45429343e-01 2.45215654e-01 -1.30913937e+00 1.31329930e+00
2.13987559e-01 -1.47507763e+00 3.32293987e-01 1.64633840e-01
7.07984686e-01 6.33025110e-01 -1.93868190e-01 4.90046680e-01
5.90744764e-02 -7.98148453e-01 5.43758750e-01 4.53031838e-01
5.90465546e-01 -7.05707312e-01 7.49973714e-01 4.23055351e-01
-1.26744616e+00 8.39628503e-02 -7.47384727e-01 1.85785592e-01
1.01627402e-01 4.96083736e-01 -6.46758020e-01 9.53827262e-01
1.04455936e+00 9.98203516e-01 -4.96212512e-01 1.26479733e+00
-9.72190127e-02 1.02543861e-01 -4.50804830e-01 1.48172766e-01
3.74602348e-01 -1.38200119e-01 8.22076559e-01 6.65134370e-01
5.26710749e-01 3.92840691e-02 2.70743936e-01 1.20501578e+00
5.61350062e-02 -1.68266654e-01 -1.31360781e+00 4.27054673e-01
5.15369296e-01 1.08729541e+00 -7.72044182e-01 -2.69864976e-01
-8.64533484e-01 7.63362110e-01 3.74582767e-01 4.27222371e-01
-5.24758339e-01 -4.16076541e-01 7.84114778e-01 1.33106247e-01
6.10665619e-01 -3.33763808e-01 -1.78627372e-01 -1.33510435e+00
1.98799551e-01 -4.47555125e-01 9.24038440e-02 -9.56093848e-01
-1.69322097e+00 5.93650162e-01 2.96527117e-01 -2.01430559e+00
1.02997683e-01 -6.84353828e-01 -7.50720441e-01 9.09799457e-01
-1.54532635e+00 -1.33219433e+00 -4.91758615e-01 5.40236175e-01
7.54487932e-01 -1.48599759e-01 7.56693006e-01 7.43624792e-02
-1.75116062e-02 1.47251353e-01 -2.62702644e-01 -1.19846612e-02
4.87078339e-01 -1.17214835e+00 8.22961926e-01 4.09077018e-01
2.41552696e-01 5.38183928e-01 2.05933183e-01 -6.46303713e-01
-1.31891501e+00 -1.12150943e+00 5.75719833e-01 -6.85048997e-01
5.00403866e-02 -6.10643744e-01 -1.29472697e+00 6.02965474e-01
-1.06841065e-01 4.01667267e-01 6.38357639e-01 1.20408401e-01
-3.37972373e-01 1.97908115e-02 -1.07825398e+00 4.17847365e-01
1.50784254e+00 -5.39992452e-01 -9.14680123e-01 2.23475561e-01
7.57586241e-01 -4.92284864e-01 -7.89026558e-01 8.10029864e-01
3.26586127e-01 -1.32336175e+00 1.34338033e+00 -5.57588875e-01
3.93963128e-01 -2.84096956e-01 -2.55415887e-01 -1.30593598e+00
-5.57764590e-01 6.46198466e-02 -8.94651487e-02 9.30512130e-01
8.43331739e-02 -5.93429506e-01 1.18010116e+00 2.13418663e-01
-4.28403050e-01 -1.02541935e+00 -1.06113648e+00 -1.01278043e+00
2.25771740e-02 -5.56910813e-01 8.15081477e-01 9.57542419e-01
-5.33640504e-01 4.05796826e-01 2.13233642e-02 4.51220751e-01
7.30033875e-01 3.82257342e-01 1.11218369e+00 -1.70237970e+00
1.53339118e-01 -4.58785981e-01 -9.15889144e-01 -9.80190694e-01
1.88639909e-01 -1.18855834e+00 -2.03694403e-01 -1.70604861e+00
-5.52632399e-02 -5.32532334e-01 -2.78335571e-01 3.03472131e-01
8.79237056e-02 1.38704076e-01 4.58509117e-01 4.33663219e-01
-3.11053723e-01 7.36080229e-01 1.36076224e+00 -2.11123735e-01
-2.66100228e-01 2.94250190e-01 -1.36145055e-01 1.00119054e+00
4.56018329e-01 -3.33238810e-01 -5.41106522e-01 -3.11967134e-01
-6.20944910e-02 -8.47144946e-02 6.01543963e-01 -1.26162338e+00
2.18085185e-01 2.58859154e-02 6.56600296e-01 -1.40748072e+00
7.26520300e-01 -1.10594928e+00 7.86787197e-02 1.67715088e-01
2.48506457e-01 -1.15658954e-01 1.62635669e-01 7.44428873e-01
-3.05827349e-01 -2.29213923e-01 5.85104823e-01 -5.87075293e-01
-8.71095300e-01 7.92953730e-01 2.03011364e-01 -3.07543337e-01
1.10689414e+00 -6.87252402e-01 -1.83516487e-01 -7.67929479e-02
-6.90922976e-01 2.76877135e-01 7.57685006e-01 7.60423779e-01
9.35423255e-01 -1.57817757e+00 -3.37740749e-01 6.31056786e-01
5.34052730e-01 5.97491920e-01 3.53562534e-01 4.69953448e-01
-4.81868714e-01 2.81781644e-01 -3.08569878e-01 -1.28823471e+00
-1.06957304e+00 7.20130920e-01 1.93109557e-01 9.99118388e-02
-1.16175246e+00 7.98992574e-01 5.84064126e-01 -8.84709656e-01
2.62655884e-01 -4.82015401e-01 -2.24651471e-01 -7.16669410e-02
1.62260592e-01 1.28418609e-01 3.04441750e-01 -5.08430481e-01
-3.83518636e-01 1.00428259e+00 -5.22250794e-02 1.88371360e-01
1.59847176e+00 1.63023859e-01 1.44666135e-01 6.66644812e-01
1.35372329e+00 -3.53306293e-01 -1.28302979e+00 -4.02156025e-01
-1.79190129e-01 -6.79683805e-01 8.64188299e-02 -2.32530683e-01
-1.01125729e+00 1.16568482e+00 7.05805480e-01 2.52202958e-01
8.41436923e-01 1.91649571e-01 6.81008458e-01 3.02064747e-01
6.88227296e-01 -6.42862856e-01 -1.36463791e-01 5.30663013e-01
9.76452112e-01 -1.45945215e+00 -3.52494381e-02 -7.50030100e-01
-2.87726223e-01 1.18668818e+00 8.48997355e-01 -4.05388862e-01
9.19689178e-01 -8.08919966e-02 -1.07825987e-01 -6.82901323e-01
-6.81428730e-01 -2.18236268e-01 3.89895380e-01 7.63438940e-01
-7.27771968e-02 -2.90394753e-01 2.76854187e-01 2.28035614e-01
-2.42903903e-01 5.01093976e-02 1.08927302e-01 1.18870556e+00
-3.93980354e-01 -8.95024955e-01 -3.35319787e-01 5.80803990e-01
2.43846583e-03 1.41542345e-01 -2.70303428e-01 1.07938707e+00
2.66657293e-01 4.42148685e-01 4.34813350e-01 -5.65231085e-01
7.76139915e-01 2.01004624e-01 5.12116134e-01 -9.22416866e-01
-2.74451822e-01 -9.51977223e-02 -4.34500724e-01 -8.10320258e-01
-6.33306384e-01 -4.09824193e-01 -8.97947609e-01 -1.60120353e-01
-4.49631006e-01 -3.41132849e-01 6.21969104e-01 7.56724119e-01
6.21345162e-01 3.14959079e-01 7.77606547e-01 -1.51883745e+00
-4.60882723e-01 -7.62535691e-01 -5.02608836e-01 6.19749904e-01
2.58803397e-01 -1.12711227e+00 -5.04979312e-01 -4.15019870e-01] | [8.07878589630127, -3.3366682529449463] |
e4f0d702-2004-4622-a76d-8d452536d02f | computing-education-in-the-era-of-generative | 2306.02608 | null | https://arxiv.org/abs/2306.02608v1 | https://arxiv.org/pdf/2306.02608v1.pdf | Computing Education in the Era of Generative AI | The computing education community has a rich history of pedagogical innovation designed to support students in introductory courses, and to support teachers in facilitating student learning. Very recent advances in artificial intelligence have resulted in code generation models that can produce source code from natural language problem descriptions -- with impressive accuracy in many cases. The wide availability of these models and their ease of use has raised concerns about potential impacts on many aspects of society, including the future of computing education. In this paper, we discuss the challenges and opportunities such models present to computing educators, with a focus on introductory programming classrooms. We summarize the results of two recent articles, the first evaluating the performance of code generation models on typical introductory-level programming problems, and the second exploring the quality and novelty of learning resources generated by these models. We consider likely impacts of such models upon pedagogical practice in the context of the most recent advances at the time of writing. | ['Sami Sarsa', 'Eddie Antonio Santos', 'Brent N. Reeves', 'Andrew Luxton-Reilly', 'Juho Leinonen', 'Arto Hellas', 'James Finnie-Ansley', 'Brett A. Becker', 'James Prather', 'Paul Denny'] | 2023-06-05 | null | null | null | null | ['code-generation'] | ['computer-code'] | [ 4.44944166e-02 3.61607790e-01 1.94672607e-02 -3.73362184e-01
-4.87452328e-01 -8.10075462e-01 3.02412063e-01 8.06499183e-01
-2.62928724e-01 3.27468634e-01 1.78015143e-01 -9.33339179e-01
-3.18543106e-01 -8.71221483e-01 -7.05985308e-01 -1.75230056e-01
5.70985153e-02 8.76594707e-02 2.19296739e-01 -3.54355991e-01
7.80448556e-01 6.94284976e-01 -2.08286428e+00 2.42281601e-01
1.28616071e+00 1.29804835e-01 2.41295218e-01 5.27991235e-01
-6.41871989e-01 1.25501454e+00 -7.71749675e-01 -4.13565129e-01
-2.37602606e-01 -3.42458427e-01 -8.88969064e-01 -1.28704071e-01
4.51921731e-01 -1.69432253e-01 -4.25112993e-02 9.12610412e-01
2.44650587e-01 3.07441384e-01 1.47207111e-01 -1.04985893e+00
-5.22455454e-01 4.69914496e-01 -6.35947511e-02 2.08979666e-01
7.62201369e-01 8.16623718e-02 7.49275565e-01 -4.78801459e-01
6.83506966e-01 6.43714845e-01 7.33373702e-01 5.76542437e-01
-9.79886591e-01 -4.13052946e-01 1.36934280e-01 5.71736880e-03
-1.10060203e+00 -1.53627485e-01 4.36056823e-01 -9.18103635e-01
9.67307508e-01 1.47659704e-01 1.13752854e+00 2.33606309e-01
2.84337997e-01 5.90031624e-01 8.68281960e-01 -8.94173682e-01
2.09240422e-01 7.34062493e-01 3.62913191e-01 8.29960644e-01
4.25203502e-01 -2.33629569e-01 -3.45753431e-01 -4.90931831e-02
5.33167660e-01 -2.61885047e-01 -2.08844811e-01 -3.27330053e-01
-9.91892159e-01 8.09846461e-01 1.85013339e-01 6.31094694e-01
1.83091715e-01 7.88301677e-02 1.97119161e-01 6.87894076e-02
-3.46230902e-02 8.35409641e-01 -2.39664540e-01 -6.18192792e-01
-7.95282900e-01 5.57492971e-01 1.13542569e+00 1.15290225e+00
3.30333561e-01 2.26572171e-01 3.94044459e-01 6.49140596e-01
5.01032412e-01 -1.03954062e-01 6.27280235e-01 -9.28969502e-01
1.23509288e-01 7.56658018e-01 -3.77828509e-01 -9.31340933e-01
8.61514583e-02 -4.40297365e-01 5.07849976e-02 4.64543521e-01
3.46184015e-01 -8.68890509e-02 -4.31134403e-01 1.18199360e+00
2.34057188e-01 -1.00335687e-01 -5.63883819e-02 3.08744431e-01
1.16000116e+00 7.34653175e-01 3.67712140e-01 2.24084571e-01
1.18460536e+00 -9.34491277e-01 -3.53380263e-01 6.28979057e-02
9.32928145e-01 -9.04940844e-01 9.32350874e-01 5.96232295e-01
-1.36087620e+00 -6.27759695e-01 -8.80806029e-01 -1.30158275e-01
-5.26944041e-01 -2.06199855e-01 6.29878283e-01 1.25930405e+00
-1.04645026e+00 4.69686240e-01 -5.96537828e-01 -2.32358843e-01
7.50056505e-02 6.81233257e-02 2.42942981e-02 -1.66685686e-01
-6.32327199e-01 8.94341707e-01 2.28048891e-01 -5.63455403e-01
-3.96400243e-01 -1.23474002e+00 -7.43302286e-01 5.03546476e-01
1.24390565e-01 -3.70513260e-01 1.70145571e+00 -7.41845489e-01
-1.41077077e+00 8.65069091e-01 1.80101439e-01 1.14427254e-01
2.38860130e-01 8.58961493e-02 -5.72814420e-02 3.66642401e-02
8.34795982e-02 6.10302091e-01 -2.07640365e-01 -9.76958632e-01
-8.22734475e-01 8.00312161e-02 4.03010398e-01 3.66062433e-01
-4.75288481e-01 1.41441301e-01 -1.63861006e-01 -4.58241820e-01
2.45309234e-01 -7.18392611e-01 -3.83768171e-01 1.49985030e-01
5.36463082e-01 -4.48428303e-01 3.36670309e-01 -2.02855632e-01
1.04358411e+00 -1.87749958e+00 -3.30763936e-01 1.75316527e-01
1.07235134e-01 1.54572949e-01 7.73733929e-02 6.81065083e-01
-2.90595323e-01 3.91267955e-01 1.13405965e-01 2.19028935e-01
1.77559227e-01 -1.58982262e-01 -4.16738957e-01 -2.38587379e-01
-4.58639376e-02 3.38766932e-01 -1.10118747e+00 -3.69798034e-01
3.35581332e-01 4.49222505e-01 -8.60230684e-01 3.69104773e-01
-2.55575508e-01 1.85447738e-01 -3.74851167e-01 1.91437349e-01
1.46191731e-01 7.41361931e-04 1.77949920e-01 8.10012639e-01
-7.70135105e-01 7.12460637e-01 -1.05184937e+00 1.65645564e+00
-8.15105855e-01 9.51931119e-01 -1.08825676e-02 -8.22873533e-01
8.57977092e-01 5.13222218e-01 3.57807010e-01 -2.58767515e-01
-2.15809762e-01 2.40921170e-01 3.37698460e-01 -7.14351714e-01
5.25781274e-01 -3.40003148e-02 2.45016456e-01 8.81955028e-01
1.50910378e-01 -1.05205297e+00 5.15478492e-01 4.87567216e-01
7.66749501e-01 5.26627779e-01 -9.21668205e-03 -5.09189129e-01
3.76934230e-01 3.18330288e-01 1.53939128e-01 6.92638934e-01
-1.14569962e-01 3.68559718e-01 4.32026595e-01 -5.06916165e-01
-8.37782741e-01 -6.09675407e-01 -1.99232638e-01 1.09624791e+00
-4.74731117e-01 -5.57029366e-01 -8.93616080e-01 -2.62583315e-01
-2.37547427e-01 1.02512717e+00 3.92344370e-02 1.52167216e-01
-4.11668956e-01 -4.10235137e-01 2.58514136e-01 2.56273836e-01
9.55289602e-02 -1.21732318e+00 -1.00863242e+00 3.30272615e-01
3.17579657e-02 -1.04143047e+00 1.55242041e-01 5.37922569e-02
-9.06689644e-01 -6.71909332e-01 -2.73679674e-01 -1.36717260e+00
9.09935355e-01 2.85013795e-01 1.22358549e+00 8.88508141e-01
-5.80099285e-01 1.01488292e+00 -3.07843179e-01 -7.41556704e-01
-7.73480833e-01 -4.57780026e-02 -4.02318954e-01 -1.02283752e+00
4.85135257e-01 -4.78398681e-01 5.29247038e-02 -5.55165708e-01
-1.09525454e+00 3.81618410e-01 2.54573792e-01 3.23942959e-01
-9.49190035e-02 4.06426340e-01 5.06178498e-01 -9.74119246e-01
7.22157955e-01 -4.60674584e-01 -7.78684199e-01 3.19758028e-01
-6.77130103e-01 -1.49434552e-01 5.98492742e-01 -8.72378424e-02
-1.09786355e+00 -2.27536187e-01 -4.63273883e-01 4.50501323e-01
-5.41466236e-01 9.21556175e-01 1.84454575e-01 -7.18418658e-01
7.62193441e-01 1.40918776e-01 -2.30282247e-01 -1.93576381e-01
-9.35011506e-02 4.57675993e-01 1.63593888e-01 -1.35508120e+00
6.49998963e-01 -4.00556982e-01 -1.49977403e-02 -1.17114592e+00
-5.05321980e-01 -2.90539086e-01 -4.61488098e-01 -4.49121356e-01
2.73564428e-01 -8.31738591e-01 -6.23293400e-01 1.07704327e-01
-1.01655817e+00 -3.21206927e-01 -3.71197730e-01 5.61795235e-01
-2.23426357e-01 1.27085134e-01 -4.78793174e-01 -5.54399848e-01
2.57357955e-01 -1.57026851e+00 9.66836289e-02 6.99510396e-01
-2.72384644e-01 -1.30218768e+00 1.22523502e-01 7.31148303e-01
4.06236589e-01 2.88674552e-02 1.24005282e+00 -5.52926183e-01
-6.77451611e-01 -1.49582043e-01 2.59409964e-01 6.05625734e-02
-2.90437758e-01 5.36079347e-01 -8.73974741e-01 1.92109928e-01
-3.09681818e-02 -2.39393055e-01 -3.56852561e-02 -1.29417386e-02
1.14966023e+00 -1.88168809e-01 -1.14246391e-01 2.28031561e-01
1.60693157e+00 3.08977127e-01 2.42879465e-01 4.54766154e-01
2.40097135e-01 8.65949094e-01 2.87416339e-01 7.64283687e-02
3.93187970e-01 1.37773231e-01 -8.07348192e-02 5.25516629e-01
-1.54649757e-03 -2.45109349e-01 1.83580533e-01 1.43389177e+00
-3.89201306e-02 2.98136741e-01 -1.56680954e+00 9.98610795e-01
-1.34094906e+00 -8.72834086e-01 -2.99382597e-01 1.80313814e+00
8.93628657e-01 3.57005775e-01 -4.70634848e-02 -8.57586712e-02
3.44509810e-01 -2.92590588e-01 2.58296967e-01 -9.78551745e-01
6.78859115e-01 7.92171597e-01 -2.79905587e-01 6.12205684e-01
-3.00263882e-01 5.72183073e-01 6.80204153e+00 5.01937270e-01
-9.20597911e-01 -3.26983541e-01 4.30200040e-01 9.72301885e-02
-7.52134085e-01 1.87585920e-01 -6.81455672e-01 2.34544307e-01
1.21998847e+00 -7.62731194e-01 3.67667019e-01 1.15908694e+00
9.60887522e-02 -2.32921407e-01 -9.57372427e-01 2.46250451e-01
5.07931784e-02 -1.70369768e+00 -1.28867283e-01 -2.16400623e-01
1.16729069e+00 -2.89889216e-01 -5.95283136e-02 5.29009461e-01
6.19367361e-01 -9.00022268e-01 7.47754693e-01 -1.16710728e-02
1.77625686e-01 -8.41968894e-01 3.50631803e-01 4.41212565e-01
-9.05659020e-01 -1.12337939e-01 -7.05599636e-02 -6.93582594e-01
-5.54864407e-01 2.43802428e-01 -9.99166965e-01 2.21340135e-01
5.54957628e-01 4.20419686e-02 -5.01162350e-01 1.45093822e+00
-2.82692343e-01 5.64427197e-01 -4.89965752e-02 -7.12848246e-01
2.52196580e-01 -1.41700998e-01 1.28754422e-01 1.33787775e+00
3.76642704e-01 7.45007932e-01 2.84767747e-01 1.06742251e+00
2.95252025e-01 2.91755915e-01 -5.24452984e-01 -2.57239968e-01
7.76193082e-01 1.29446149e+00 -8.26676846e-01 -8.03161934e-02
-7.82116473e-01 1.63055122e-01 2.08577141e-01 2.34386042e-01
-3.94801110e-01 -9.19354439e-01 4.77618694e-01 5.28048098e-01
-1.49448648e-01 -4.63349640e-01 -4.89768893e-01 -9.06570017e-01
-1.87227845e-01 -1.19683385e+00 -1.03869952e-01 -8.39465857e-01
-7.23171353e-01 2.30081737e-01 9.93826613e-02 -7.34550118e-01
-2.93162502e-02 -7.20157802e-01 -8.71119440e-01 8.69663417e-01
-1.17753005e+00 -6.16438687e-01 -3.29558313e-01 -1.68842539e-01
4.52470362e-01 -2.22667649e-01 9.11818206e-01 4.43402259e-03
-3.40116352e-01 3.68825525e-01 2.18083203e-01 1.93905067e-02
3.13134938e-01 -1.23686111e+00 3.81912053e-01 7.42305636e-01
1.50357530e-01 9.84665990e-01 7.03221321e-01 -2.85659552e-01
-1.44418716e+00 -5.93076229e-01 1.25639808e+00 -3.91258627e-01
5.80244243e-01 1.43663064e-01 -7.83358037e-01 6.44590914e-01
2.27934286e-01 -3.62525761e-01 1.05389369e+00 6.72124326e-02
-4.41916622e-02 2.16850400e-01 -1.14398515e+00 6.20198905e-01
4.52211380e-01 -4.48706627e-01 -8.43197942e-01 3.65925074e-01
4.46206629e-01 -6.50864422e-01 -1.05540156e+00 -1.26102954e-01
5.99396944e-01 -9.95120525e-01 7.30618834e-01 -4.08930629e-01
8.92512321e-01 5.51315174e-02 3.88614923e-01 -9.29485023e-01
-1.88109353e-01 -4.86720592e-01 5.68355381e-01 1.04676712e+00
2.96402350e-02 -3.16413671e-01 9.16160524e-01 1.29714203e+00
-5.27451098e-01 -6.46184564e-01 -2.89191693e-01 -2.84242809e-01
5.10504067e-01 -4.21561331e-01 6.19562030e-01 1.25153208e+00
7.21480727e-01 -9.72633287e-02 8.14617574e-01 -1.36683568e-01
4.84172791e-01 1.87981442e-01 7.18680441e-01 -1.34733129e+00
-2.35955328e-01 -7.88640618e-01 -4.72285628e-01 -7.62113869e-01
9.06870663e-02 -9.54687715e-01 8.02976415e-02 -1.43087947e+00
8.75553265e-02 -6.13502920e-01 2.92251289e-01 4.34816211e-01
-1.27822548e-01 -1.07109599e-01 2.67855942e-01 -2.58117706e-01
-2.18714684e-01 -7.99679607e-02 1.12897742e+00 3.59962076e-01
-1.66858241e-01 -2.88488977e-02 -8.48729789e-01 1.01405132e+00
6.76292896e-01 -3.37009490e-01 -4.68826950e-01 -5.68452179e-01
4.36849177e-01 5.05038723e-02 -9.06539895e-03 -1.42708302e+00
5.61827183e-01 -6.54298604e-01 2.30597243e-01 6.04373477e-02
-3.16445321e-01 -8.32570732e-01 -9.92383808e-02 4.70193624e-01
-5.72161376e-01 3.32246512e-01 6.78863883e-01 -2.24232644e-01
-2.00730637e-01 -1.22146440e+00 7.20125973e-01 -5.03185153e-01
-6.26429379e-01 -3.84206682e-01 -1.06248868e+00 7.97115266e-02
1.22916150e+00 -2.09610447e-01 -3.05693448e-01 -3.70012671e-01
-5.14492989e-01 -3.05117052e-02 4.36797082e-01 3.40638667e-01
4.09551769e-01 -8.96988332e-01 -3.10777247e-01 3.36170256e-01
-1.41362637e-01 9.49777514e-02 4.67332937e-02 1.51666448e-01
-1.17627001e+00 7.88749516e-01 -6.62175655e-01 -2.40394697e-01
-1.27737582e+00 2.13472337e-01 1.27964556e-01 -6.57544658e-02
-3.02614629e-01 9.54460323e-01 -1.32119581e-01 -8.31706882e-01
4.89480913e-01 -2.08713636e-01 -1.81171462e-01 -3.54645848e-01
6.04797840e-01 2.97201633e-01 1.08351901e-01 -1.34768128e-01
1.70800105e-01 2.46218368e-01 -5.15519120e-02 -1.76120356e-01
1.38708568e+00 2.76459664e-01 -1.58169925e-01 4.14074004e-01
6.74990475e-01 2.65954524e-01 -4.24713731e-01 2.58219205e-02
2.04863578e-01 -3.46311927e-01 7.15912580e-02 -8.43433559e-01
-6.47528231e-01 1.19342804e+00 2.61267543e-01 2.15685755e-01
7.16815889e-01 -5.79895496e-01 1.04079805e-01 3.57690364e-01
2.80067235e-01 -8.69521320e-01 2.06881791e-01 6.67208970e-01
4.53023016e-01 -8.51288080e-01 2.67911822e-01 -4.75271255e-01
-2.26900071e-01 1.69470263e+00 1.10618484e+00 2.51222670e-01
5.85301280e-01 3.07959974e-01 -7.57862926e-02 -1.37966171e-01
-8.62886429e-01 2.96217412e-01 8.46631303e-02 5.03460884e-01
1.42073119e+00 -3.53375405e-01 -4.44053441e-01 3.05566669e-01
-3.03455323e-01 2.75690407e-01 1.30199790e+00 1.54873776e+00
-8.61164808e-01 -1.36658466e+00 -6.72523677e-01 6.11701496e-02
-4.47378784e-01 -2.67941266e-01 -3.87754649e-01 6.79203987e-01
1.66293830e-01 8.46642911e-01 3.97147313e-02 8.06065872e-02
9.06248167e-02 3.92440826e-01 7.76881039e-01 -1.23025692e+00
-1.14999938e+00 -4.90581006e-01 -1.45478487e-01 2.72128195e-01
-6.19121753e-02 -5.88793874e-01 -1.31490600e+00 -6.95826888e-01
-3.77384156e-01 7.22686291e-01 9.78317857e-01 6.85026765e-01
2.29871109e-01 6.48827136e-01 -2.12833583e-02 -5.86442292e-01
-6.52677178e-01 -4.30677980e-01 -1.96015567e-01 -1.64427355e-01
-4.15334329e-02 -3.90368439e-02 -1.04706831e-01 4.05145168e-01] | [9.831528663635254, 7.3534393310546875] |
e26d04b2-693f-4ea0-8eac-39c4bf3c3001 | hp-gan-probabilistic-3d-human-motion | 1711.09561 | null | http://arxiv.org/abs/1711.09561v1 | http://arxiv.org/pdf/1711.09561v1.pdf | HP-GAN: Probabilistic 3D human motion prediction via GAN | Predicting and understanding human motion dynamics has many applications,
such as motion synthesis, augmented reality, security, and autonomous vehicles.
Due to the recent success of generative adversarial networks (GAN), there has
been much interest in probabilistic estimation and synthetic data generation
using deep neural network architectures and learning algorithms.
We propose a novel sequence-to-sequence model for probabilistic human motion
prediction, trained with a modified version of improved Wasserstein generative
adversarial networks (WGAN-GP), in which we use a custom loss function designed
for human motion prediction. Our model, which we call HP-GAN, learns a
probability density function of future human poses conditioned on previous
poses. It predicts multiple sequences of possible future human poses, each from
the same input sequence but a different vector z drawn from a random
distribution. Furthermore, to quantify the quality of the non-deterministic
predictions, we simultaneously train a motion-quality-assessment model that
learns the probability that a given skeleton sequence is a real human motion.
We test our algorithm on two of the largest skeleton datasets: NTURGB-D and
Human3.6M. We train our model on both single and multiple action types. Its
predictive power for long-term motion estimation is demonstrated by generating
multiple plausible futures of more than 30 frames from just 10 frames of input.
We show that most sequences generated from the same input have more than 50\%
probabilities of being judged as a real human sequence. We will release all the
code used in this paper to Github. | ['Zicheng Liu', 'John Kender', 'Emad Barsoum'] | 2017-11-27 | null | null | null | null | ['human-pose-forecasting'] | ['computer-vision'] | [ 3.76055658e-01 4.33298379e-01 3.64586376e-02 -1.30166799e-01
-9.05182779e-01 -3.46548319e-01 7.92051852e-01 -8.91605854e-01
-2.41899729e-01 1.10613871e+00 4.99381542e-01 2.72633974e-02
4.51642305e-01 -8.59603047e-01 -1.11781561e+00 -7.46025503e-01
6.94933254e-03 5.93290091e-01 4.36788797e-01 -1.67839006e-01
-3.55751365e-01 4.24569666e-01 -1.28196275e+00 1.87764570e-01
3.48488897e-01 6.25054300e-01 4.44612652e-02 1.27467287e+00
5.71909547e-01 1.02284682e+00 -7.85491049e-01 -6.28734767e-01
2.95298934e-01 -7.35993445e-01 -6.68022692e-01 -3.82627845e-02
-7.74211250e-03 -6.55282021e-01 -7.95175254e-01 7.33244956e-01
5.91061532e-01 2.41274670e-01 8.79962444e-01 -1.59784091e+00
-2.93521851e-01 3.02213013e-01 -3.09055328e-01 -2.73054332e-01
7.21050322e-01 7.77091205e-01 5.55652142e-01 -5.38004756e-01
9.65021431e-01 1.32760561e+00 7.90560007e-01 1.15237319e+00
-1.16353965e+00 -5.22821248e-01 -4.08710808e-01 2.27114215e-01
-1.22141457e+00 -2.15918615e-01 6.94115639e-01 -5.39673328e-01
6.76298499e-01 4.68869768e-02 8.94747496e-01 2.07078528e+00
6.26701772e-01 9.64073479e-01 5.59456646e-01 -1.54762194e-01
2.34895602e-01 -5.58978260e-01 -7.32280433e-01 6.80393398e-01
-1.03995860e-01 4.80093122e-01 -4.84451801e-01 -1.24880895e-01
9.91523921e-01 -1.25514910e-01 -2.71696746e-01 -4.35459018e-01
-1.57118726e+00 8.45498621e-01 1.87992454e-01 -2.15322331e-01
-5.64556658e-01 9.24242556e-01 1.57571778e-01 -2.94087559e-01
3.52906957e-02 7.50910640e-02 -1.14736527e-01 -5.08214176e-01
-8.18645120e-01 9.49672401e-01 6.26038194e-01 9.24549162e-01
4.24377054e-01 4.22301799e-01 -3.67262393e-01 3.96528512e-01
2.77137727e-01 9.74447787e-01 6.57791793e-01 -1.40751982e+00
3.51043463e-01 -2.35341206e-01 4.24436867e-01 -9.15931165e-01
-2.94947445e-01 9.35074612e-02 -9.64308560e-01 4.65041459e-01
4.67998266e-01 -5.07578075e-01 -1.12406039e+00 2.04577947e+00
2.48707563e-01 5.48648536e-01 1.60390198e-01 8.30377817e-01
4.89108920e-01 8.94730449e-01 1.45911053e-01 1.64078191e-01
8.56177151e-01 -8.57918262e-01 -4.95519519e-01 -1.52357206e-01
2.22733602e-01 -5.48521161e-01 8.03448915e-01 2.18566790e-01
-1.17804074e+00 -7.44895279e-01 -8.90371263e-01 6.17997609e-02
1.62241578e-01 -1.87351331e-01 2.49221534e-01 5.98495424e-01
-9.33532119e-01 8.90505254e-01 -1.19834554e+00 -1.00963235e-01
3.32827508e-01 2.51060985e-02 -3.76853228e-01 -1.01859644e-01
-1.33902764e+00 7.19723761e-01 4.26626831e-01 -9.72269997e-02
-1.32629454e+00 -2.76975006e-01 -9.93823171e-01 -2.77557611e-01
3.67677137e-02 -1.37318397e+00 1.30629265e+00 -8.33283246e-01
-1.75834620e+00 4.70687568e-01 -1.22807503e-01 -7.80209780e-01
9.92085159e-01 -3.99035037e-01 -3.54446143e-01 1.36549383e-01
2.43156403e-01 1.22554696e+00 1.03263116e+00 -1.06020081e+00
-4.06442702e-01 8.80374238e-02 -3.06283355e-01 6.67539909e-02
4.75321680e-01 -3.62557322e-01 -3.90674442e-01 -1.13273525e+00
-3.76474947e-01 -1.34686077e+00 -4.68111038e-01 3.56843136e-02
-6.42729580e-01 2.08508626e-01 6.35302305e-01 -8.44894588e-01
7.50862241e-01 -1.74884057e+00 5.09616554e-01 1.02856522e-02
-1.56266704e-01 1.15483165e-01 -2.57885635e-01 3.06201279e-01
2.48319767e-02 1.28307462e-01 -5.40928185e-01 -3.57920825e-01
1.28552541e-01 5.02651751e-01 -4.44278091e-01 2.61924297e-01
3.23955327e-01 1.31664801e+00 -1.11323535e+00 -3.52200270e-01
3.54083210e-01 6.26358390e-01 -5.84122479e-01 4.45926666e-01
-5.33887446e-01 8.63089740e-01 -1.61735475e-01 3.23215574e-01
2.79119641e-01 -4.57953885e-02 -6.24639280e-02 5.86286187e-02
6.06552720e-01 -1.82264462e-01 -1.00871301e+00 1.81749690e+00
-1.20225884e-01 5.68850338e-01 -7.04019904e-01 -4.31789696e-01
6.26635015e-01 5.34602940e-01 5.88616610e-01 -2.35851347e-01
-4.08893228e-02 -8.92860144e-02 -9.73715261e-02 -4.65828806e-01
6.53124690e-01 -3.08820218e-01 -3.09297800e-01 4.62833941e-01
6.00998364e-02 -3.74861658e-01 -6.36895224e-02 1.58063427e-01
1.37067211e+00 8.02585661e-01 1.71320319e-01 5.11823535e-01
1.01952448e-01 -1.02341948e-02 6.97628081e-01 7.55921185e-01
-2.72645742e-01 1.22092044e+00 4.39153165e-01 -4.12771314e-01
-1.60451162e+00 -1.57591808e+00 5.08353531e-01 4.64250654e-01
2.01561842e-02 -1.96060583e-01 -7.31877744e-01 -6.46111369e-01
-2.26043999e-01 9.71061349e-01 -6.78971112e-01 -4.68593717e-01
-8.99522245e-01 -5.61163008e-01 9.36163008e-01 7.28481472e-01
3.04919571e-01 -1.53878534e+00 -8.22697997e-01 3.81406188e-01
-4.95751590e-01 -1.12968051e+00 -5.28493524e-01 -5.09820759e-01
-5.96898139e-01 -8.83873582e-01 -1.20399606e+00 -3.12720090e-01
2.19044998e-01 -3.98659170e-01 1.25401950e+00 -3.19385231e-01
-1.79393813e-01 4.32694823e-01 -3.00825417e-01 -2.03112289e-01
-1.03106260e+00 -2.28812292e-01 1.83066115e-01 -4.24434356e-02
-1.46551952e-01 -4.62562233e-01 -7.64014602e-01 3.09202403e-01
-9.19928372e-01 4.22418475e-01 4.38145965e-01 1.01493514e+00
7.04169393e-01 -2.31178090e-01 4.49945956e-01 -5.57226837e-01
2.24157214e-01 -7.82036424e-01 -1.35474056e-01 -2.40430143e-02
-2.12151129e-02 1.74362242e-01 5.06975591e-01 -6.42683506e-01
-1.10759652e+00 3.85532558e-01 -5.59898376e-01 -7.32304931e-01
-3.35985839e-01 6.67386577e-02 -2.55013049e-01 4.34858590e-01
8.01556706e-01 4.08296317e-01 5.26283076e-03 6.49536848e-02
7.46540129e-01 -1.80109907e-02 1.04044044e+00 -5.47665417e-01
9.84902978e-01 5.57743907e-01 2.03953773e-01 -6.97903991e-01
-2.57437885e-01 1.65796742e-01 -3.39180678e-01 -4.48888421e-01
1.26070225e+00 -8.65460694e-01 -5.50572455e-01 8.38684678e-01
-1.44146597e+00 -7.18731463e-01 -5.12601554e-01 5.15220344e-01
-1.26438701e+00 5.40334344e-01 -5.98618746e-01 -8.36494625e-01
-1.13569856e-01 -1.23146617e+00 1.22373450e+00 1.47226024e-02
-6.86536789e-01 -8.14436495e-01 3.95564198e-01 1.58846289e-01
3.29187177e-02 9.57123876e-01 5.16505599e-01 -1.41807660e-01
-7.83515632e-01 -2.99078375e-01 4.05753374e-01 3.32768679e-01
-9.30631757e-02 1.29707918e-01 -6.19170845e-01 -5.26063144e-02
-2.66819447e-01 -4.11713511e-01 7.01284111e-01 5.92048526e-01
9.09438014e-01 -4.35187906e-01 -3.11367154e-01 6.74120784e-01
1.00487554e+00 3.57004076e-01 1.23772538e+00 9.06152725e-02
9.96802986e-01 4.09030765e-01 4.93040204e-01 4.67702210e-01
2.20630258e-01 8.30710948e-01 4.68025982e-01 3.25155020e-01
-2.25501642e-01 -8.37382257e-01 6.14606321e-01 2.93167293e-01
-2.75965154e-01 -6.53838456e-01 -9.17858899e-01 6.16038561e-01
-1.87072790e+00 -1.36019802e+00 8.30606669e-02 2.09857464e+00
6.54248655e-01 2.89929450e-01 4.00384545e-01 6.74876198e-02
5.78745544e-01 2.09010959e-01 -6.40165746e-01 -2.88930982e-02
-1.36086255e-01 3.26534659e-01 3.57209265e-01 4.50724870e-01
-1.02986515e+00 9.26537097e-01 6.36235094e+00 7.98304021e-01
-8.41370642e-01 -6.33707121e-02 7.49076307e-01 -9.95214731e-02
-3.98589343e-01 -2.19982505e-01 -5.39548337e-01 6.76878273e-01
1.08477604e+00 -1.46720693e-01 1.20457046e-01 8.41526091e-01
3.12109172e-01 3.74590531e-02 -1.01165104e+00 8.83158684e-01
-8.84339660e-02 -1.41690910e+00 3.35657537e-01 2.26484034e-02
9.23659444e-01 -1.98267642e-02 8.43953639e-02 2.12712675e-01
7.42001772e-01 -1.28332460e+00 1.09653008e+00 9.87481773e-01
8.77286494e-01 -8.60679269e-01 6.69463992e-01 5.48822820e-01
-9.66438293e-01 3.42058450e-01 -3.47857662e-02 1.79111287e-01
8.61352563e-01 1.74623311e-01 -8.67912591e-01 4.76218104e-01
3.62218082e-01 6.50882304e-01 -1.53848633e-01 7.46280253e-01
-4.98944581e-01 6.04832709e-01 -2.48996451e-01 1.82899311e-01
-1.09260110e-02 1.38994455e-01 7.68236220e-01 8.85491014e-01
6.31523967e-01 2.56087035e-02 -5.63674048e-02 8.38789344e-01
1.46916494e-01 -5.02745509e-01 -8.20122004e-01 5.96003048e-02
3.12295794e-01 6.36293292e-01 -3.57102454e-01 -3.82536054e-01
3.08829881e-02 1.35264456e+00 -1.09505065e-01 4.08462286e-01
-1.28694844e+00 -1.91240944e-02 7.52830327e-01 -3.91077623e-02
3.67708147e-01 -3.60833436e-01 -1.12507656e-01 -1.12258232e+00
-1.16719455e-01 -8.30104470e-01 2.58271415e-02 -1.11837626e+00
-1.07463372e+00 7.51502991e-01 1.21908434e-01 -1.46709967e+00
-1.34287751e+00 -3.28519851e-01 -6.41273081e-01 8.74297500e-01
-6.47989094e-01 -1.12229264e+00 -1.75226554e-01 5.15550196e-01
5.24071038e-01 -2.78991908e-01 7.87802815e-01 -5.17367981e-02
-1.87085286e-01 5.28694928e-01 -2.21864998e-01 2.90428013e-01
4.29199368e-01 -1.10976660e+00 1.20749485e+00 9.53902245e-01
2.20252320e-01 -5.65561280e-02 1.03436506e+00 -9.60223496e-01
-9.63640630e-01 -1.36522138e+00 6.11265004e-01 -8.02563787e-01
3.94623041e-01 5.68927564e-02 -6.82363391e-01 8.91654611e-01
-2.04694849e-02 -3.94233391e-02 2.36839876e-01 -9.12920237e-01
7.85919055e-02 4.28240865e-01 -1.22518897e+00 9.33487475e-01
1.24351001e+00 -2.57834375e-01 -4.63510782e-01 1.72176868e-01
8.71234655e-01 -6.81453407e-01 -8.41656983e-01 5.96305788e-01
7.35269010e-01 -1.09791172e+00 1.24825382e+00 -6.39074802e-01
9.56040800e-01 -3.07618141e-01 -2.18664855e-01 -1.37145102e+00
-1.84203580e-01 -6.46531582e-01 -4.72083420e-01 7.00222313e-01
2.03279689e-01 -2.22643688e-01 1.24833906e+00 6.24051034e-01
2.94093676e-02 -7.65268266e-01 -9.72207069e-01 -8.34330559e-01
9.80814695e-02 -7.27826416e-01 6.16105795e-01 4.11010355e-01
-6.07137561e-01 1.17061593e-01 -1.04486179e+00 4.46100123e-02
7.21783459e-01 -3.46934974e-01 1.33385277e+00 -5.23368537e-01
-7.50298798e-01 -1.67299509e-01 -7.36810982e-01 -1.25155294e+00
2.33838439e-01 -4.12915230e-01 4.17725205e-01 -1.40320790e+00
-7.20026791e-02 8.73790234e-02 3.98228705e-01 1.55393749e-01
-2.68764049e-01 3.63201439e-01 3.30741912e-01 2.45887607e-01
-2.17261642e-01 1.03882968e+00 1.36245537e+00 -6.81963861e-02
1.12786427e-01 2.69611567e-01 9.44449753e-02 8.68982613e-01
5.80254138e-01 -4.72475946e-01 -4.56145227e-01 -2.25357935e-01
-7.38278627e-02 7.08831608e-01 8.37502658e-01 -1.46139681e+00
-2.00770214e-01 -3.16227138e-01 6.20536029e-01 -6.66903615e-01
7.83557475e-01 -5.00058055e-01 9.41253543e-01 7.94759691e-01
-2.11174309e-01 1.47762537e-01 -1.91170238e-02 9.43381608e-01
-5.12969494e-03 5.26948385e-02 7.01884508e-01 -2.20821172e-01
-8.43488634e-01 5.39056122e-01 -4.93643045e-01 1.67780951e-01
1.04989886e+00 -2.08757401e-01 -1.97720267e-02 -9.50946331e-01
-1.00425100e+00 -1.61053166e-01 6.53707564e-01 4.24283594e-01
8.11435997e-01 -1.70887446e+00 -8.78629684e-01 -8.61903504e-02
-6.11500368e-02 1.42549038e-01 3.70286047e-01 1.44563600e-01
-8.42014492e-01 1.46430314e-01 -3.67627025e-01 -6.56258821e-01
-8.18005264e-01 4.14368868e-01 3.63815486e-01 -3.97805363e-01
-6.95853949e-01 7.20508516e-01 1.66419163e-01 -2.53511429e-01
-1.41365290e-01 -7.72447437e-02 1.89961776e-01 -7.75296152e-01
4.18293327e-01 5.22139907e-01 -6.59281254e-01 -1.05253482e+00
-1.36687413e-01 3.02406788e-01 5.00610471e-01 -7.19792962e-01
9.69321728e-01 9.59743559e-02 5.41621447e-01 4.15973037e-01
9.89114523e-01 -1.63284630e-01 -1.93891096e+00 1.63669840e-01
-2.72336721e-01 -2.94811726e-01 -7.45111406e-01 -6.58407867e-01
-9.94253695e-01 6.42191410e-01 5.24205565e-01 -2.51536131e-01
7.60423660e-01 -1.41600534e-01 1.28392518e+00 3.05196363e-02
7.10033596e-01 -5.69092870e-01 3.73706967e-01 4.08738375e-01
1.03595412e+00 -1.05694640e+00 -2.71497190e-01 -8.62247422e-02
-9.89596605e-01 8.61368120e-01 5.17138660e-01 -3.49459350e-01
4.11122203e-01 1.49275109e-01 1.79459639e-02 2.92078316e-01
-6.37850285e-01 4.08640392e-02 3.29082817e-01 1.06912887e+00
1.26211479e-01 3.28680068e-01 3.64549533e-02 4.63679880e-01
-5.34835100e-01 3.88095587e-01 6.41106963e-01 6.46732330e-01
-4.24889103e-03 -1.05720150e+00 -4.77835476e-01 7.39253312e-02
-3.36756319e-01 1.95982039e-01 -3.04297525e-02 7.01897979e-01
1.14814311e-01 4.86363232e-01 -1.09326169e-01 -7.59951711e-01
7.74434358e-02 1.24628060e-01 5.46289563e-01 -3.58342707e-01
1.04189329e-01 -2.48930886e-01 1.27205715e-01 -7.74288297e-01
-3.67547393e-01 -8.57763648e-01 -1.23996198e+00 -4.73784089e-01
4.06190902e-01 -2.35282928e-01 3.31115484e-01 8.07790399e-01
2.53512532e-01 5.82215309e-01 2.72323966e-01 -1.43876207e+00
-5.07201493e-01 -8.98318529e-01 -3.21643203e-01 7.95379937e-01
1.88515618e-01 -6.66134119e-01 -5.96094243e-02 5.22951484e-01] | [7.299492835998535, -0.10407783091068268] |
1c201a88-23b4-49a7-a545-99808cfe0b97 | vvc-extension-scheme-for-object-detection | 2305.18782 | null | https://arxiv.org/abs/2305.18782v1 | https://arxiv.org/pdf/2305.18782v1.pdf | VVC Extension Scheme for Object Detection Using Contrast Reduction | In recent years, video analysis using Artificial Intelligence (AI) has been widely used, due to the remarkable development of image recognition technology using deep learning. In 2019, the Moving Picture Experts Group (MPEG) has started standardization of Video Coding for Machines (VCM) as a video coding technology for image recognition. In the framework of VCM, both higher image recognition accuracy and video compression performance are required. In this paper, we propose an extention scheme of video coding for object detection using Versatile Video Coding (VVC). Unlike video for human vision, video used for object detection does not require a large image size or high contrast. Since downsampling of the image can reduce the amount of information to be transmitted. Due to the decrease in image contrast, entropy of the image becomes smaller. Therefore, in our proposed scheme, the original image is reduced in size and contrast, then coded with VVC encoder to achieve high compression performance. Then, the output image from the VVC decoder is restored to its original image size using the bicubic method. Experimental results show that the proposed video coding scheme achieves better coding performance than regular VVC in terms of object detection accuracy. | ['Hiroshi Watanabe', 'Kein Yamada', 'Taiju Watanabe', 'Takahiro Shindo'] | 2023-05-30 | null | null | null | null | ['video-compression'] | ['computer-vision'] | [ 5.22523701e-01 -3.41147900e-01 -1.35391429e-01 1.33619770e-01
3.96336615e-02 -2.34849602e-02 2.26972550e-01 -1.80919051e-01
-5.51371813e-01 4.23582464e-01 -1.98460668e-02 -1.06752686e-01
3.87575179e-01 -8.72394979e-01 -4.82218742e-01 -7.06114650e-01
1.14803314e-01 -4.24760818e-01 5.48826456e-01 2.72596121e-01
4.92299169e-01 3.79544884e-01 -1.66605926e+00 4.70067799e-01
5.38627446e-01 1.37926483e+00 7.66917646e-01 6.33605003e-01
-4.06573504e-01 1.26463521e+00 -5.41158080e-01 -1.27551675e-01
3.39901924e-01 -6.71886384e-01 -6.95070207e-01 4.91834372e-01
-8.36539343e-02 -7.01358378e-01 -6.55290902e-01 1.33849740e+00
6.83331043e-02 5.74817806e-02 6.26896024e-01 -1.17016828e+00
-5.85816026e-01 1.21515378e-01 -6.59960747e-01 3.48389745e-01
2.15400890e-01 5.30432863e-03 3.91242594e-01 -7.58532763e-01
5.01732111e-01 1.18425262e+00 2.03224912e-01 5.01873016e-01
-8.03935409e-01 -5.71231723e-01 -3.06082934e-01 9.32202816e-01
-1.39623022e+00 -1.76342919e-01 7.46269047e-01 -2.65834451e-01
8.21022093e-01 3.25256646e-01 9.32500422e-01 4.15652484e-01
4.15904075e-01 7.24432647e-01 6.51527584e-01 -6.77026451e-01
2.67682135e-01 -1.41309025e-02 -2.88292944e-01 6.26220047e-01
3.84773344e-01 -9.53693166e-02 -2.85133310e-02 3.95119935e-01
9.06021059e-01 4.00922894e-01 -4.72917169e-01 -1.21453740e-01
-1.08452141e+00 7.46889889e-01 3.77354652e-01 5.53476751e-01
-4.47243184e-01 1.86545700e-01 5.74119687e-01 2.39186078e-01
-8.75514820e-02 -1.72841981e-01 1.12221628e-01 -3.22465926e-01
-9.06795382e-01 -3.36403400e-01 5.70526898e-01 8.09054315e-01
4.14459974e-01 4.01026547e-01 2.68032581e-01 8.48365664e-01
5.03332555e-01 5.88347673e-01 9.15942192e-01 -1.23130667e+00
4.65120584e-01 7.10505962e-01 -2.29081303e-01 -1.38152313e+00
-8.28733370e-02 -4.41634189e-03 -1.21484911e+00 3.53542268e-01
2.32819431e-02 1.00213520e-01 -7.99431682e-01 1.00970745e+00
-1.47738084e-01 1.02545016e-01 4.60555911e-01 9.01283503e-01
7.38946140e-01 1.38621414e+00 -6.65762499e-02 -5.44171870e-01
1.27520335e+00 -8.03127646e-01 -9.57678258e-01 -5.24199866e-02
4.73289996e-01 -8.41827154e-01 4.26331609e-01 5.60194612e-01
-9.68999684e-01 -8.33597124e-01 -1.27674258e+00 9.27930027e-02
-4.15151380e-03 3.90000418e-02 2.54000664e-01 5.85700035e-01
-7.06450403e-01 1.71235442e-01 -6.11782908e-01 -3.58736992e-01
4.15916413e-01 2.77579606e-01 -3.96087527e-01 -3.64886731e-01
-9.34688032e-01 7.27720022e-01 9.51038182e-01 -7.79654458e-02
-5.26461601e-01 -1.50113061e-01 -8.20176125e-01 2.58750707e-01
3.09587363e-02 -1.07715428e-01 8.90411973e-01 -1.40108657e+00
-1.29859293e+00 6.37070537e-01 -2.01737955e-02 -7.54540563e-01
2.58760720e-01 2.10601643e-01 -5.45150340e-01 7.26523638e-01
-1.95662871e-01 8.61862957e-01 1.03050160e+00 -8.77587855e-01
-9.00218785e-01 -7.04071820e-02 -1.00542717e-01 8.89222994e-02
-3.42385203e-01 7.08044320e-02 -7.14232147e-01 -4.85273719e-01
2.82955557e-01 -6.81012988e-01 1.13325141e-01 2.28952095e-01
2.32023492e-01 -7.53152417e-03 1.47216642e+00 -6.93221271e-01
1.33021104e+00 -2.50386214e+00 4.04748656e-02 7.06180977e-03
6.69194013e-02 7.07471550e-01 -3.30924802e-02 2.25393400e-01
1.43001825e-01 -9.15757790e-02 -3.18992078e-01 3.70277196e-01
-8.56158674e-01 3.67172360e-01 1.03610910e-01 3.37188542e-01
-5.75448908e-02 5.19254625e-01 -5.25783002e-01 -9.04989660e-01
6.78688347e-01 4.43182409e-01 -7.37043440e-01 5.07759824e-02
2.47948900e-01 1.67402193e-01 -3.11739475e-01 6.35728002e-01
1.01184964e+00 -2.36280233e-01 2.28904337e-01 -2.01552033e-01
-1.29735708e-01 -5.65870106e-01 -1.18101001e+00 1.04160738e+00
-3.81807536e-01 1.10849190e+00 8.36738497e-02 -1.44658422e+00
9.46149409e-01 4.73089874e-01 6.10987186e-01 -9.30594027e-01
3.23456198e-01 1.67154193e-01 1.50020078e-01 -8.92077565e-01
3.02020282e-01 1.17540449e-01 4.96007532e-01 -6.46671802e-02
-6.07365742e-02 -6.06860220e-02 3.76127362e-01 9.56349820e-02
6.58891976e-01 -2.67434806e-01 5.31746149e-01 1.17851235e-01
1.12941194e+00 -5.37534505e-02 6.65470302e-01 3.47230732e-01
-2.73336887e-01 4.72111404e-01 7.67763928e-02 -5.39483249e-01
-1.32796001e+00 -6.21430337e-01 -2.35468715e-01 2.40356371e-01
5.39223135e-01 -2.67917901e-01 -8.83486688e-01 -3.44215669e-02
-2.92762756e-01 4.11296129e-01 -4.56136558e-03 -2.32783973e-01
-6.79275811e-01 -3.97813290e-01 2.39354908e-01 1.87157065e-01
1.32738900e+00 -1.38899469e+00 -1.00201523e+00 3.49586010e-01
-3.11721593e-01 -1.31520534e+00 -2.51322687e-01 -4.61935490e-01
-1.01373661e+00 -1.13552701e+00 -9.01127160e-01 -1.33266270e+00
5.66449404e-01 8.63210440e-01 3.63131732e-01 4.67068672e-01
-4.35868323e-01 3.22701544e-01 -6.41526341e-01 -2.17328668e-01
-6.50941908e-01 -5.45645475e-01 -1.56173244e-01 2.18389139e-01
3.54334950e-01 -2.43419111e-01 -7.99813867e-01 1.25964791e-01
-1.36491680e+00 3.70616943e-01 6.76213861e-01 6.72059953e-01
2.26547942e-01 5.36434650e-01 2.46698618e-01 -2.38590002e-01
1.15758307e-01 -1.34004101e-01 -7.35553443e-01 4.49014418e-02
-4.58020002e-01 -3.20724040e-01 7.68872499e-01 -2.04328001e-01
-9.69557881e-01 -4.50134166e-02 -8.20679292e-02 -4.30266768e-01
1.31317943e-01 3.59183967e-01 -2.43913531e-02 -3.00550193e-01
1.56530485e-01 8.43255103e-01 5.63431799e-01 -1.47953942e-01
-8.77030939e-02 1.16778994e+00 5.24798036e-01 4.10272837e-01
3.90062809e-01 3.60120177e-01 9.10169706e-02 -1.23532140e+00
3.68211158e-02 -3.43158811e-01 -3.12757283e-01 -6.65889978e-01
1.27835488e+00 -9.13489759e-01 -8.13331783e-01 7.43362665e-01
-1.36273658e+00 3.32784414e-01 2.36438677e-01 1.05499649e+00
-3.89555722e-01 8.76614213e-01 -6.84332132e-01 -7.95200825e-01
-4.01754856e-01 -1.35275304e+00 4.87192452e-01 3.52003545e-01
2.75172472e-01 -6.69562280e-01 -6.11664355e-01 3.37588519e-01
4.46471661e-01 -7.91958421e-02 8.94111693e-01 1.25536487e-01
-8.66199434e-01 -3.86446714e-01 -4.96147066e-01 8.13964009e-01
1.44173399e-01 3.26262452e-02 -4.55118090e-01 -1.28385410e-01
3.92944455e-01 9.97362658e-02 8.72646034e-01 4.87441570e-01
1.53090537e+00 -3.93751830e-01 -1.95527852e-01 6.80547953e-01
1.78155673e+00 1.15478849e+00 1.21286488e+00 6.17393076e-01
5.02487838e-01 2.15184584e-01 4.96722162e-01 3.55837524e-01
-3.58288772e-02 5.61062872e-01 6.07442677e-01 -5.48900776e-02
-2.57050514e-01 4.56008725e-02 4.35215265e-01 1.05513299e+00
-3.09985012e-01 -3.27113897e-01 -5.71591556e-01 6.17329776e-02
-1.55215943e+00 -1.39028049e+00 -1.94611058e-01 2.18625736e+00
4.11670536e-01 4.14667800e-02 -4.25088614e-01 5.59392810e-01
9.57382143e-01 5.46541177e-02 -2.58141875e-01 -5.98116934e-01
-7.35915452e-02 -3.39598268e-01 5.78934848e-01 3.06216508e-01
-9.71677661e-01 5.46545684e-01 5.74353790e+00 1.03596151e+00
-1.40217447e+00 -1.14345774e-01 5.27669013e-01 1.53721064e-01
1.57382056e-01 -1.99469015e-01 -4.08244967e-01 9.65400517e-01
6.90663874e-01 -2.23161072e-01 4.72372025e-01 9.47261989e-01
1.82377473e-01 -3.66013139e-01 -5.23901463e-01 1.59779799e+00
3.26830804e-01 -1.39790297e+00 3.43632787e-01 3.11363172e-02
4.73276496e-01 -4.30913448e-01 -1.18317150e-01 1.72730610e-01
-7.01850235e-01 -5.96087813e-01 5.39634407e-01 1.87518969e-01
7.90648282e-01 -6.95155382e-01 9.00809586e-01 2.65154302e-01
-1.31803560e+00 -5.08405745e-01 -8.67971361e-01 -1.59327909e-01
1.37472406e-01 3.45281899e-01 -4.59646672e-01 2.27684155e-01
7.39260554e-01 7.35515773e-01 -3.39415312e-01 1.11813724e+00
2.98409045e-01 4.86048996e-01 -4.22566831e-02 -3.58573161e-02
3.68469357e-01 -4.97417301e-01 5.67413867e-01 9.21514928e-01
5.50010324e-01 4.26046044e-01 5.32728955e-02 2.60164857e-01
-1.09726943e-01 2.74186522e-01 -7.39987373e-01 -2.77511086e-02
3.77130866e-01 8.52118433e-01 -9.12556767e-01 -7.25212455e-01
-8.61719608e-01 1.15107632e+00 -3.04052949e-01 1.88886017e-01
-7.42429733e-01 -3.96200627e-01 -3.01362691e-03 7.39347190e-02
5.60386062e-01 -1.74523741e-01 1.34192169e-01 -1.10749590e+00
-2.89743301e-02 -7.70384312e-01 1.75262034e-01 -9.24898207e-01
-4.03318971e-01 5.00662804e-01 -1.51669368e-01 -1.81757176e+00
-6.91746920e-02 -6.06853664e-01 -4.54290211e-01 4.69581634e-01
-1.17921460e+00 -5.82533717e-01 -3.76732618e-01 4.68041778e-01
9.82190132e-01 -4.82981920e-01 5.30537486e-01 4.76137996e-01
-4.07434374e-01 8.96156281e-02 5.35128355e-01 2.56499320e-01
1.97378069e-01 -5.65707922e-01 -2.06638753e-01 9.81685638e-01
-1.11280821e-01 1.55302867e-01 3.91435713e-01 -4.69401032e-01
-1.55253255e+00 -9.73945439e-01 6.05763018e-01 5.13696730e-01
7.49270245e-02 9.22464132e-02 -8.39582026e-01 2.80371875e-01
3.58001113e-01 -6.00266792e-02 4.04055715e-01 -9.41292584e-01
9.15800631e-02 -3.68810803e-01 -1.28990841e+00 4.23729241e-01
3.37192357e-01 -2.53039092e-01 -4.39133108e-01 1.55489549e-01
6.43400252e-01 1.64106805e-02 -6.92625701e-01 3.36090684e-01
7.54442811e-01 -1.00310457e+00 8.82192373e-01 7.10233673e-02
6.19278252e-01 -5.21476805e-01 -4.23822314e-01 -6.38400435e-01
-4.36526746e-01 2.77877301e-02 8.05408061e-02 8.81995618e-01
-4.73706722e-02 -3.84892613e-01 6.42644346e-01 2.34162301e-01
8.79463181e-02 -5.22963166e-01 -1.01282489e+00 -7.56760478e-01
-4.84770209e-01 -3.00384909e-01 1.41188487e-01 5.71052730e-01
5.19623570e-02 2.19955556e-02 -6.68716371e-01 -1.84044372e-02
5.99148393e-01 2.63637640e-02 3.68424505e-01 -9.66554165e-01
-2.27454394e-01 -3.73354912e-01 -1.17071474e+00 -1.04248738e+00
-3.56428564e-01 -7.17218518e-01 -3.12330961e-01 -1.61438870e+00
4.11103040e-01 2.64097482e-01 -1.70730948e-01 -1.20170027e-01
7.15779439e-02 5.94368517e-01 7.59038210e-01 6.13385141e-01
-4.87762183e-01 4.25158143e-01 1.28910315e+00 -3.81621003e-01
-5.98081090e-02 -2.20126674e-01 -4.13465649e-02 7.03379631e-01
8.07954133e-01 -1.39721051e-01 -3.40751171e-01 -4.01645035e-01
-4.02215987e-01 5.09344697e-01 1.65997744e-01 -1.34932935e+00
1.54188469e-01 -6.00977466e-02 7.62629151e-01 -5.88034809e-01
3.06063920e-01 -1.22255266e+00 2.95811564e-01 1.18961656e+00
-9.20867473e-02 -1.23165578e-01 8.01559836e-02 4.91178691e-01
-6.31745100e-01 -6.43030405e-01 1.03657007e+00 -2.04047292e-01
-1.34527206e+00 5.75624034e-02 -1.00011432e+00 -5.93413293e-01
1.38868916e+00 -7.99507439e-01 -2.75182594e-02 -4.57249880e-01
-2.94025093e-01 -1.28807157e-01 4.01073337e-01 4.18486029e-01
1.06603336e+00 -1.43301153e+00 -4.43857282e-01 4.43478048e-01
-4.70275469e-02 -3.26546133e-01 5.69999754e-01 6.57405436e-01
-1.34162974e+00 5.18123627e-01 -4.86922890e-01 -7.17578351e-01
-1.60512340e+00 9.10163760e-01 4.85984683e-02 1.97485819e-01
-9.09607112e-01 3.73355001e-01 3.12281191e-01 6.17063642e-01
1.95245042e-01 -1.78852290e-01 -5.01845241e-01 -2.52302080e-01
9.74898994e-01 4.49108928e-01 -2.60338098e-01 -7.34695256e-01
-1.84600249e-01 7.33933568e-01 -5.02669588e-02 7.84202814e-02
9.71732080e-01 -3.45701844e-01 -2.66319752e-01 -6.91488758e-03
1.64339340e+00 -3.42330545e-01 -9.62624431e-01 1.82474494e-01
-3.46597701e-01 -6.90053642e-01 3.49599421e-01 -1.64623246e-01
-1.36121321e+00 8.20575416e-01 9.66049314e-01 2.78593242e-01
1.49314439e+00 -4.18902755e-01 8.71113181e-01 2.90359706e-01
4.99489427e-01 -1.11693454e+00 3.78275327e-02 2.19679296e-01
6.84609413e-01 -1.24441969e+00 8.39645043e-02 -4.17585671e-01
-6.63390815e-01 1.49620247e+00 4.77908850e-01 -1.19312584e-01
6.02924049e-01 1.26163706e-01 -1.23369910e-01 1.55839875e-01
-4.38725531e-01 7.37617835e-02 -5.80416247e-02 5.93244374e-01
2.16263697e-01 -2.20751464e-01 -6.93118215e-01 -1.25258103e-01
3.10434759e-01 1.71456590e-01 8.09011281e-01 9.07991946e-01
-1.07421315e+00 -8.42779517e-01 -6.88482940e-01 4.35567826e-01
-6.52604938e-01 4.88425270e-02 2.24160999e-01 5.92342615e-01
2.32918382e-01 1.12443173e+00 3.84526283e-01 -3.05185646e-01
-6.85891807e-02 -3.07430506e-01 5.20795166e-01 -1.14445291e-01
2.35639974e-01 1.84084140e-02 -4.75962460e-01 -2.81087011e-01
-7.32502341e-01 -2.23715603e-01 -1.48417628e+00 -2.84001231e-01
-1.07075535e-01 2.47004837e-01 8.82325113e-01 7.27448702e-01
1.58798508e-02 3.79870683e-01 7.31123149e-01 -6.23350024e-01
-8.29722546e-03 -8.17459822e-01 -6.96804225e-01 5.17515361e-01
1.58553436e-01 -4.23782200e-01 -2.69041210e-01 4.90303993e-01] | [11.175915718078613, -1.5825532674789429] |
8cfbac87-70b1-45b4-aafd-56b2264e8fc3 | powerplanningdl-reliability-aware-framework | 2005.01386 | null | https://arxiv.org/abs/2005.01386v2 | https://arxiv.org/pdf/2005.01386v2.pdf | PowerPlanningDL: Reliability-Aware Framework for On-Chip Power Grid Design using Deep Learning | With the increase in the complexity of chip designs, VLSI physical design has become a time-consuming task, which is an iterative design process. Power planning is that part of the floorplanning in VLSI physical design where power grid networks are designed in order to provide adequate power to all the underlying functional blocks. Power planning also requires multiple iterative steps to create the power grid network while satisfying the allowed worst-case IR drop and Electromigration (EM) margin. For the first time, this paper introduces Deep learning (DL)-based framework to approximately predict the initial design of the power grid network, considering different reliability constraints. The proposed framework reduces many iterative design steps and speeds up the total design cycle. Neural Network-based multi-target regression technique is used to create the DL model. Feature extraction is done, and the training dataset is generated from the floorplans of some of the power grid designs extracted from the IBM processor. The DL model is trained using the generated dataset. The proposed DL-based framework is validated using a new set of power grid specifications (obtained by perturbing the designs used in the training phase). The results show that the predicted power grid design is closer to the original design with minimal prediction error (~2%). The proposed DL-based approach also improves the design cycle time with a speedup of ~6X for standard power grid benchmarks. | ['Sukanta Dey', 'Sukumar Nandi', 'Gaurav Trivedi'] | 2020-05-04 | null | null | null | null | ['multi-target-regression'] | ['miscellaneous'] | [-9.16796401e-02 -1.04212416e-02 -3.44022423e-01 -1.44536451e-01
-4.53529507e-01 -3.40372562e-01 3.02909344e-01 2.47569263e-01
1.47387415e-01 9.36793685e-01 -1.01078607e-01 -4.79498237e-01
-4.53460544e-01 -9.26720262e-01 -3.17689985e-01 -7.59642005e-01
-1.36620045e-01 5.21883309e-01 -9.44636390e-03 -1.91381592e-02
5.10314167e-01 8.80196095e-01 -1.02410150e+00 4.07625549e-02
6.50952280e-01 1.12277198e+00 1.44256294e-01 4.59122986e-01
2.79639453e-01 5.93772352e-01 -8.56835067e-01 4.32328373e-01
2.48840854e-01 -4.27206397e-01 -3.71424109e-01 7.58674070e-02
2.86573246e-02 -2.46583238e-01 -2.78259099e-01 1.23248887e+00
6.50043488e-01 -1.30028024e-01 6.57324255e-01 -1.36732006e+00
-7.72748590e-02 8.02594066e-01 -7.98817277e-01 1.44048899e-01
-1.37075871e-01 3.57394442e-02 6.80847824e-01 -3.46029758e-01
1.03193313e-01 8.08115005e-01 3.81757885e-01 5.98818138e-02
-1.27197385e+00 -1.00574911e+00 -3.98499638e-01 2.93251306e-01
-1.63264871e+00 -1.93201572e-01 1.17798042e+00 -1.96488231e-01
1.57094550e+00 -8.40548649e-02 5.72777510e-01 4.84624356e-01
1.26950061e+00 2.06885681e-01 1.03238022e+00 -4.81833160e-01
8.04851413e-01 -8.35473314e-02 1.80045336e-01 4.18245196e-01
6.83874667e-01 2.84880221e-01 -1.19607337e-01 -1.87442094e-01
6.07477963e-01 -3.43913525e-01 -5.89338876e-02 -2.26371661e-01
-6.22124732e-01 5.46176076e-01 5.77751756e-01 7.13186681e-01
-2.97119796e-01 5.14778078e-01 5.05334854e-01 1.48715109e-01
-1.50326803e-01 7.45790422e-01 -6.73067451e-01 -9.06755850e-02
-1.27083802e+00 3.44063431e-01 8.68843377e-01 1.03371572e+00
6.42394602e-01 7.91733980e-01 1.47917628e-01 1.02422833e-01
3.75486314e-01 3.67267281e-01 3.77620906e-01 -4.62451696e-01
1.22550137e-01 7.63852358e-01 -1.60830375e-02 -1.02540779e+00
-9.12341118e-01 -9.69473302e-01 -1.29219699e+00 4.76208150e-01
1.20060019e-01 -5.35213232e-01 -8.95583451e-01 1.11762702e+00
-4.24638577e-02 -4.84607667e-02 -9.00912285e-02 3.22261572e-01
2.15827778e-01 1.15529144e+00 -1.42058864e-01 -5.23123682e-01
1.02762103e+00 -7.26747811e-01 -9.72537279e-01 -2.40697220e-01
6.17652714e-01 -6.00864887e-01 3.99516761e-01 8.94224048e-01
-5.91975987e-01 -6.52354717e-01 -1.99138153e+00 5.21612644e-01
-2.25191399e-01 3.30731958e-01 3.61035347e-01 8.37975383e-01
-9.15118277e-01 8.54669809e-01 -7.73417652e-01 -1.61653876e-01
4.57653910e-01 8.68454993e-01 1.46821544e-01 -1.29382946e-02
-8.82822812e-01 8.81258070e-01 5.08807838e-01 1.86766773e-01
-1.14756846e+00 -1.10269415e+00 -6.78125858e-01 3.83316070e-01
2.73958385e-01 -6.53082877e-02 9.45529401e-01 -7.49179959e-01
-1.54050028e+00 -2.48501346e-01 3.82285893e-01 -4.10172820e-01
-1.28297761e-01 1.27961054e-01 -5.79845607e-01 -5.10120988e-01
-1.60907790e-01 1.83370233e-01 6.43455863e-01 -1.24776053e+00
-6.43600821e-01 -2.06762537e-01 -5.74773610e-01 -2.20208362e-01
-2.37656623e-01 -3.27070504e-01 8.50635022e-03 -2.25372404e-01
-2.33240470e-01 -6.82218492e-01 -6.15013599e-01 -6.19868517e-01
-5.75465441e-01 -2.26268664e-01 9.60087180e-01 -6.97925508e-01
1.41789079e+00 -1.72233498e+00 -1.22213185e-01 6.50570989e-01
-2.55950719e-01 -1.38577381e-02 2.42398903e-01 5.20601213e-01
-1.78272024e-01 -1.33698806e-01 9.61083099e-02 6.72799423e-02
4.88182157e-02 1.00628793e-01 5.41548505e-02 6.13473773e-01
1.53126851e-01 5.98461986e-01 -4.94088590e-01 6.57060519e-02
3.88977468e-01 2.04943448e-01 -4.45085973e-01 5.96647337e-02
-3.76383007e-01 3.98833454e-01 -3.77574027e-01 4.23894882e-01
7.47516870e-01 -1.53976521e-02 6.84216261e-01 -6.90696776e-01
-1.40146345e-01 -4.97393794e-02 -1.37718260e+00 1.41081715e+00
-7.48262167e-01 6.50123775e-01 -3.75157744e-01 -1.13885736e+00
1.31292832e+00 2.24675909e-01 4.61112916e-01 -9.77532446e-01
6.59036875e-01 2.29210854e-01 5.65779865e-01 1.23710759e-01
2.26942629e-01 1.53595880e-01 -5.86813748e-01 3.96899104e-01
1.19817868e-01 -3.04617405e-01 1.67595372e-01 -3.58651906e-01
1.36739933e+00 -8.58349726e-03 6.98238492e-01 -1.10565031e+00
6.13829017e-01 2.37080529e-01 8.85448337e-01 9.79318097e-02
1.85202822e-01 4.51957881e-02 8.97502303e-01 -6.07475460e-01
-1.50222588e+00 -8.72440994e-01 -2.64913112e-01 -1.04287535e-01
-2.01733157e-01 -2.38060787e-01 -6.43975854e-01 -6.17334843e-01
-2.99609691e-01 1.10538757e+00 -3.33190262e-01 -3.14297080e-01
-6.98843777e-01 -1.05147922e+00 2.41053745e-01 5.80198109e-01
5.13455153e-01 -9.09268379e-01 -7.46863306e-01 4.20581907e-01
9.48261738e-01 -8.97160709e-01 -1.52037814e-01 1.02829897e+00
-6.54654741e-01 -8.37334037e-01 4.70063537e-02 -9.59763587e-01
1.01448309e+00 -7.41802335e-01 1.09172046e+00 2.28481814e-01
-6.34254813e-01 -5.68121731e-01 -4.60252501e-02 -1.52827039e-01
-5.57078660e-01 3.38550240e-01 2.13394403e-01 -6.66797936e-01
2.67132044e-01 -6.79528236e-01 -2.43900880e-01 1.29697710e-01
-5.53127825e-01 -6.44170493e-02 7.72367001e-01 8.99195611e-01
4.49609339e-01 1.59212577e+00 1.04349101e+00 -3.24384332e-01
4.26992089e-01 -3.47678781e-01 -1.40768123e+00 -1.33014441e-01
-7.91408837e-01 3.96253347e-01 1.36868107e+00 -4.03494090e-01
-8.36610198e-01 2.95097560e-01 -6.79801255e-02 -2.08468392e-01
1.00174904e-01 3.72637600e-01 -1.13635015e+00 -2.58119315e-01
4.43403751e-01 -3.51052284e-02 -4.21167046e-01 -3.56726021e-01
-1.28749803e-01 5.55885613e-01 2.51180857e-01 -3.17594141e-01
1.13244879e+00 -1.56196177e-01 6.81822121e-01 -6.61654115e-01
-2.53491998e-01 3.63029689e-01 -7.06963181e-01 -1.63634300e-01
4.80004609e-01 -5.50146401e-01 -7.10953414e-01 2.07487673e-01
-8.97967398e-01 -4.42447335e-01 1.46354083e-02 1.18957050e-01
-3.06037247e-01 -1.02800980e-01 -1.95574477e-01 -6.98918164e-01
-6.69131160e-01 -1.41853321e+00 5.26032031e-01 6.15044415e-01
-5.77875257e-01 -1.03832150e+00 -2.65076697e-01 -3.02876472e-01
3.65420401e-01 4.50903535e-01 1.60127139e+00 -7.22830355e-01
-6.90720201e-01 -2.15406865e-01 -3.34855095e-02 4.76044357e-01
5.54531753e-01 -1.37141545e-03 -4.72523540e-01 -5.37227631e-01
2.50042677e-01 1.71586275e-01 -1.26973152e-01 6.63370609e-01
7.57174909e-01 -1.32040381e-01 -5.11381388e-01 3.69859993e-01
2.27723265e+00 9.83149886e-01 6.30829930e-01 1.95205122e-01
5.52638829e-01 -1.20387018e-01 5.37243962e-01 6.50657773e-01
-1.59082457e-01 4.85056967e-01 4.78541195e-01 1.97881064e-03
8.73888135e-02 -8.80885348e-02 8.02088678e-02 5.96461952e-01
9.89401817e-01 -3.38377297e-01 -1.11275542e+00 6.64945364e-01
-1.45054185e+00 -2.10112587e-01 8.73145163e-02 1.99384248e+00
7.54235685e-01 7.40220010e-01 -3.52991790e-01 6.89939320e-01
3.02470297e-01 -2.92506218e-01 -7.46626735e-01 -7.64245927e-01
5.28415799e-01 8.05074334e-01 8.37103844e-01 6.09378815e-01
-7.99075902e-01 5.92940271e-01 5.35758543e+00 1.02471972e+00
-1.26215982e+00 -3.24082971e-01 7.87648797e-01 5.23334183e-02
1.49741694e-01 -1.83861516e-02 -1.10456586e+00 3.64397615e-01
1.21649218e+00 -6.46185815e-01 2.98502326e-01 1.01940131e+00
4.93689448e-01 -5.45134187e-01 -1.51673114e+00 8.30552876e-01
-3.91234644e-03 -1.40196335e+00 -3.51222426e-01 3.71828794e-01
1.16085064e+00 -5.44866323e-01 -2.27677688e-01 2.52579272e-01
3.24789912e-01 -1.12195575e+00 3.17493796e-01 2.55640805e-01
5.89239717e-01 -1.54713166e+00 1.15109527e+00 2.63913423e-01
-1.11880374e+00 -2.85064310e-01 -1.68530419e-01 -1.90046072e-01
5.28234653e-02 9.42519903e-01 -1.07525480e+00 7.65554428e-01
6.32670760e-01 5.40085852e-01 -5.29487312e-01 9.16968703e-01
1.52189404e-01 3.84620428e-01 -4.63314146e-01 -5.11588991e-01
8.08244348e-02 1.40512828e-02 7.28826523e-02 6.33459747e-01
4.33410138e-01 -3.88900906e-01 2.71457583e-01 9.92592692e-01
-9.32315588e-02 -1.88786328e-01 -3.94183964e-01 1.33572415e-01
6.66618168e-01 1.60673988e+00 -9.89064038e-01 5.57546839e-02
-1.56705618e-01 4.52785105e-01 -2.48507142e-01 -9.46335793e-02
-9.09696221e-01 -6.09208286e-01 5.70938051e-01 3.23659182e-01
2.50476867e-01 -4.42942560e-01 -8.73364747e-01 -7.36683980e-02
-3.15136701e-01 -9.07471001e-01 -2.73771822e-01 -4.77520078e-01
-9.28888142e-01 6.63469136e-01 1.29027545e-01 -1.02784431e+00
-4.93673414e-01 -5.92956424e-01 -7.87052631e-01 1.06122220e+00
-1.20445526e+00 -7.45407879e-01 -2.03492776e-01 -2.30376258e-01
6.90771103e-01 -5.53932905e-01 6.51448727e-01 4.95911628e-01
-7.02349484e-01 6.28908694e-01 1.62044525e-01 -1.41721874e-01
1.82481185e-01 -1.17531598e+00 3.94009829e-01 9.94980216e-01
-4.44581985e-01 1.08440369e-01 1.05051160e+00 -1.03277695e+00
-1.79588723e+00 -1.28924608e+00 5.16165674e-01 2.26519302e-01
8.15768480e-01 -4.65449482e-01 -5.46437025e-01 3.35107774e-01
6.76936626e-01 -2.06630602e-01 4.31618005e-01 -4.79895502e-01
7.16108501e-01 -5.91569543e-01 -1.33952725e+00 4.83699322e-01
2.18013704e-01 2.92856902e-01 -6.33057281e-02 8.32952932e-03
2.80657738e-01 -1.42581001e-01 -1.08237183e+00 7.97994733e-01
1.77020639e-01 -5.00942111e-01 2.72239685e-01 1.97393388e-01
7.85300285e-02 -7.52050817e-01 -1.68623313e-01 -1.41466916e+00
-5.31487346e-01 -4.27949220e-01 -1.11529134e-01 1.26876438e+00
4.19970989e-01 -1.12915915e-02 1.11249411e+00 1.22237526e-01
-2.16098607e-01 -1.00557506e+00 -1.12348592e+00 -7.12072670e-01
7.42239058e-02 -2.00987384e-01 6.43011689e-01 5.65942585e-01
-1.32841289e-01 6.04315996e-01 -1.52087376e-01 4.39433098e-01
7.19905376e-01 -3.14701945e-01 4.58752781e-01 -1.08289421e+00
-4.18874741e-01 -9.81164500e-02 -5.44875920e-01 -4.97960925e-01
2.45328560e-01 -5.32768309e-01 -5.44156805e-02 -1.45683384e+00
-2.72832662e-01 -4.60486323e-01 -1.00980803e-01 5.03469467e-01
3.47036958e-01 -2.54662275e-01 -1.81005955e-01 -5.62780559e-01
1.77812397e-01 5.26649952e-01 9.36066210e-01 -2.31936559e-01
-2.89384156e-01 -8.18194747e-02 -5.08890033e-01 4.94613707e-01
1.03152859e+00 -3.29528630e-01 -6.54185295e-01 4.28924337e-02
3.75336558e-01 1.62601650e-01 -3.61811459e-01 -1.68388140e+00
3.64378035e-01 3.74362469e-02 8.69826198e-01 -1.09194553e+00
-1.33575186e-01 -1.44775450e+00 5.86978793e-01 9.33954358e-01
3.89632076e-01 3.48959625e-01 8.89476180e-01 1.02501929e-01
1.13348179e-01 -3.48295897e-01 1.04420745e+00 3.26575011e-01
-5.84888637e-01 1.22742131e-02 -6.78978801e-01 -6.48192763e-01
1.17792606e+00 -6.00416735e-02 1.04575619e-01 2.08543375e-01
-4.45308745e-01 3.74330759e-01 1.79931313e-01 4.17930514e-01
2.95294136e-01 -1.51986551e+00 -3.16188186e-01 5.14223397e-01
-4.57969725e-01 -1.80433951e-02 1.87495708e-01 2.19922289e-01
-7.23619699e-01 5.76684713e-01 -4.85241234e-01 -5.12323499e-01
-9.09326553e-01 5.43034792e-01 7.55618334e-01 -8.33806753e-01
-4.43660647e-01 1.59676269e-01 -3.76756012e-01 9.10400674e-02
-3.23864110e-02 -6.03015840e-01 -1.32255018e-01 -2.70153940e-01
3.92708868e-01 2.99858421e-01 4.24559593e-01 -3.03728264e-02
-3.96048397e-01 5.07663012e-01 -4.64858897e-02 8.70604068e-02
1.30305266e+00 3.76395404e-01 -3.06556374e-01 1.92198247e-01
1.24621654e+00 -3.15633655e-01 -9.37904418e-01 5.70402890e-02
4.37936693e-01 9.23565403e-02 8.32649589e-01 -1.00675654e+00
-1.26723981e+00 3.70132715e-01 1.07823741e+00 -1.86560556e-01
1.57551432e+00 -6.13378465e-01 5.96326530e-01 2.72625417e-01
6.45854592e-01 -1.24746966e+00 -2.68981345e-02 3.64082187e-01
5.61577439e-01 -5.88506818e-01 4.27111059e-01 -1.68401212e-01
1.36654511e-01 1.32458949e+00 1.05583000e+00 -4.33786958e-01
1.05934680e+00 1.33923519e+00 -5.12416244e-01 2.25853585e-02
-8.05563092e-01 6.52323425e-01 8.51290897e-02 5.98835766e-01
2.03245610e-01 -2.08417282e-01 -2.15063155e-01 8.79759431e-01
-2.04571217e-01 1.50157288e-02 6.17370307e-01 9.99327958e-01
-4.41560805e-01 -1.34648454e+00 -4.07292098e-01 7.13977158e-01
-2.49773502e-01 6.10015020e-02 1.56947136e-01 8.53588998e-01
2.62961209e-01 7.84465790e-01 3.40606987e-01 -4.13019389e-01
3.11500221e-01 -2.96773940e-01 4.65134442e-01 -4.98809099e-01
-6.91983700e-01 1.83167338e-01 1.59985662e-01 -2.89987147e-01
4.84136790e-01 -3.02967966e-01 -1.50147235e+00 -1.98124066e-01
-3.25032622e-01 2.14498803e-01 7.53309488e-01 8.87890577e-01
3.70811298e-02 1.42663479e+00 1.07309771e+00 -9.65660274e-01
-1.62509888e-01 -9.68818188e-01 -7.68816054e-01 -6.19558215e-01
3.05267107e-02 -6.30848408e-01 -2.54859805e-01 -1.70798317e-01] | [5.933703899383545, 3.346686363220215] |
6a4d6284-30d6-4d99-b9d8-3d873afa87df | learning-affinity-via-spatial-propagation-1 | null | null | https://arxiv.org/pdf/1710.01020.pdf | https://arxiv.org/pdf/1710.01020.pdf | Learning Affinity via Spatial Propagation Network | In this paper, we propose spatial propagation networks for learning the affinity matrix for vision tasks. We show that by constructing a row/column linear propagation
model, the spatially varying transformation matrix exactly constitutes an affinity
matrix that models dense, global pairwise relationships of an image. Specifically,
we develop a three-way connection for the linear propagation model, which (a)
formulates a sparse transformation matrix, where all elements can be the output
from a deep CNN, but (b) results in a dense affinity matrix that effectively models
any task-specific pairwise similarity matrix. Instead of designing the similarity
kernels according to image features of two points, we can directly output all the
similarities in a purely data-driven manner. The spatial propagation network is a
generic framework that can be applied to many affinity-related tasks, including
but not limited to image matting, segmentation and colorization, to name a few.
Essentially, the model can learn semantically-aware affinity values for high-level
vision tasks due to the powerful learning capability of the deep neural network classifier. We validate the framework on the task of refinement for image segmentation
boundaries. Experiments on the HELEN face parsing and PASCAL VOC-2012
semantic segmentation tasks show that the spatial propagation network provides a
general, effective and efficient solution for generating high-quality segmentation
results. | ['Jan Kautz', 'Ming-Hsuan Yang', 'Guangyu Zhong', 'Jinwei Gu', 'Shalini De Mello', 'Sifei Liu'] | 2017-10-03 | null | null | null | null | ['face-parsing'] | ['computer-vision'] | [ 2.27516755e-01 1.41534552e-01 -3.59814465e-02 -6.23511672e-01
-5.14614284e-01 -5.47494650e-01 3.27408999e-01 -2.11136397e-02
-5.11394382e-01 5.24868332e-02 -1.39568165e-01 -1.29080787e-01
-2.79190361e-01 -9.18365359e-01 -1.09682965e+00 -8.49065900e-01
1.78656161e-01 6.82326555e-01 3.66713196e-01 -2.64037937e-01
1.73346892e-01 6.83771014e-01 -1.41109300e+00 2.43227988e-01
8.77797246e-01 9.44619536e-01 1.90604821e-01 6.24033332e-01
-3.52113843e-01 6.08836114e-01 -2.00327843e-01 -4.75681603e-01
3.45540464e-01 -2.97322392e-01 -1.09846807e+00 1.79880738e-01
7.54472315e-01 -1.06990196e-01 -1.73781350e-01 1.18048978e+00
-1.36407418e-02 -2.19852291e-02 6.64788485e-01 -1.16181028e+00
-9.07917738e-01 7.60600209e-01 -7.14014530e-01 -1.41370490e-01
-3.74795087e-02 -5.21419048e-02 1.19420993e+00 -1.14238477e+00
6.75527573e-01 1.16676831e+00 7.17935920e-01 4.83673960e-01
-1.44528568e+00 -3.19988221e-01 2.92145580e-01 2.95475572e-01
-1.36485505e+00 -2.26379093e-02 8.11747730e-01 -5.77836931e-01
6.10233665e-01 2.84375012e-01 8.24721277e-01 6.66318834e-01
-1.79003682e-02 1.03596961e+00 8.73268545e-01 -3.47727686e-01
1.37863189e-01 -6.87754527e-02 5.39008796e-01 9.35203612e-01
-1.14962541e-01 -2.08678558e-01 -3.68455559e-01 9.25062373e-02
9.11073506e-01 -4.03672270e-02 -1.71481848e-01 -8.42032969e-01
-1.12272382e+00 9.04638529e-01 1.01310742e+00 3.95251781e-01
-2.93093801e-01 2.94620246e-01 7.31027452e-03 -5.68503654e-03
3.59087735e-01 4.94229972e-01 -4.25279677e-01 3.60652328e-01
-9.50431824e-01 2.18505323e-01 7.00290382e-01 8.35446477e-01
1.19891346e+00 -1.65045559e-01 -2.48953700e-01 9.85307038e-01
3.15762192e-01 3.78988475e-01 2.55420487e-02 -1.26480722e+00
2.15359822e-01 5.58590710e-01 -1.60224363e-01 -1.11574268e+00
-5.44604957e-01 -4.80457962e-01 -8.77164185e-01 2.41931394e-01
6.42521620e-01 2.26899423e-02 -1.17803597e+00 2.05580473e+00
4.36002940e-01 2.46475011e-01 -1.46447778e-01 9.95566785e-01
7.56019711e-01 6.43927336e-01 -8.70907903e-02 2.56539524e-01
1.27487671e+00 -1.13294768e+00 -4.21628386e-01 -3.22923660e-01
4.15355146e-01 -6.28164053e-01 1.09732139e+00 2.50898302e-01
-1.28729522e+00 -5.48220277e-01 -9.63754356e-01 -4.50526834e-01
-3.87403429e-01 -1.54172387e-02 8.12450588e-01 2.23054186e-01
-1.43246424e+00 5.83266139e-01 -8.92015636e-01 -4.46248472e-01
4.41250473e-01 4.99449641e-01 -3.29185188e-01 -1.11783892e-01
-9.04765785e-01 6.42319560e-01 2.48097256e-01 3.17433119e-01
-5.44878781e-01 -8.21914196e-01 -9.55936491e-01 1.58894360e-01
1.55570924e-01 -1.04867578e+00 8.28102529e-01 -1.19935441e+00
-1.42336059e+00 1.05494404e+00 -1.56082809e-01 -3.66009027e-01
4.14113045e-01 -1.14029817e-01 2.36429468e-01 1.99914113e-01
2.26824492e-01 1.27035797e+00 9.20803547e-01 -1.41679335e+00
-5.81873655e-01 -4.79976714e-01 1.48711205e-01 2.20020741e-01
-2.88682610e-01 -1.63656980e-01 -9.83782947e-01 -6.06352270e-01
4.02125150e-01 -9.63364065e-01 -5.15919924e-01 1.72859162e-01
-5.43803096e-01 -8.99080932e-02 6.12178802e-01 -4.73566175e-01
7.30862319e-01 -2.12928319e+00 7.39904642e-01 5.77843547e-01
2.13281766e-01 2.84670554e-02 -3.37294549e-01 1.06067456e-01
-1.44074531e-02 4.42230515e-02 -8.86691272e-01 -4.01701301e-01
4.07227501e-02 4.44315195e-01 -6.02747612e-02 3.37619215e-01
5.09169400e-01 1.19043577e+00 -7.93532252e-01 -3.52503419e-01
2.71912277e-01 7.33816445e-01 -7.69822896e-01 1.96340710e-01
-2.31944367e-01 4.73280877e-01 -2.28493169e-01 2.91592598e-01
7.56186604e-01 -4.32760268e-01 -8.00133348e-02 -5.85856438e-01
1.31931733e-02 -2.77595013e-01 -1.07145822e+00 2.12373257e+00
-3.53634834e-01 4.41410661e-01 3.18750441e-01 -1.00305223e+00
5.52529156e-01 -2.89166331e-01 5.99225342e-01 -5.19006491e-01
6.34129792e-02 -1.57118794e-02 -5.12287915e-02 -2.25707829e-01
3.53153199e-01 1.57528594e-01 3.33124362e-02 2.94498980e-01
3.11095953e-01 -2.21321523e-01 2.78051555e-01 4.34483767e-01
9.05046463e-01 1.41663447e-01 -2.23743930e-01 -4.42002654e-01
6.08950973e-01 6.52389154e-02 5.32044530e-01 7.49634624e-01
7.88156688e-02 8.85357976e-01 5.16144156e-01 -4.23696518e-01
-9.53644395e-01 -1.24030864e+00 -2.44720519e-01 1.18972707e+00
3.91618580e-01 -3.30436826e-01 -1.12740207e+00 -5.39489388e-01
-8.14535469e-02 3.05971444e-01 -8.59570444e-01 -2.97805872e-02
-6.78936422e-01 -7.55980253e-01 2.15424135e-01 6.28391623e-01
5.36304951e-01 -1.00155783e+00 -1.42951593e-01 -5.97804375e-02
-8.04283991e-02 -1.21278882e+00 -6.75052404e-01 1.85400784e-01
-6.02835238e-01 -1.05560529e+00 -6.62299812e-01 -1.13213980e+00
9.40486312e-01 2.09350526e-01 1.15222514e+00 2.75299937e-01
-3.57235193e-01 5.92172027e-01 -9.71259028e-02 6.88780546e-02
-1.66331530e-01 7.48032555e-02 -3.72087479e-01 3.10737610e-01
2.10117787e-01 -5.39834738e-01 -6.32602036e-01 2.07425460e-01
-8.94060135e-01 3.14564466e-01 5.98398924e-01 9.57053244e-01
9.18547988e-01 -2.88043320e-01 5.71271852e-02 -1.20959830e+00
2.92374343e-01 -2.18211785e-01 -6.99560761e-01 3.67539793e-01
-3.26546818e-01 2.74851441e-01 4.19936597e-01 -1.27010912e-01
-8.77104461e-01 4.79227453e-01 -5.06102145e-01 -3.46074045e-01
-1.81479827e-01 4.08896208e-01 -3.01523387e-01 -3.37041706e-01
4.36326057e-01 8.15115273e-02 7.33234512e-04 -3.26271206e-01
1.02300465e+00 5.23338467e-02 8.57548118e-01 -8.20529938e-01
9.84481692e-01 6.66769683e-01 1.31160185e-01 -8.31867695e-01
-8.72494876e-01 -4.12397414e-01 -1.04722488e+00 -2.76978742e-02
1.45894682e+00 -6.80461884e-01 -6.40912056e-01 6.66794598e-01
-1.22140491e+00 -5.68430007e-01 -4.07572031e-01 2.06404567e-01
-5.88965356e-01 2.83979446e-01 -9.92942393e-01 -2.35671118e-01
-1.03284106e-01 -1.38601470e+00 1.17775941e+00 2.12520748e-01
1.23198912e-01 -1.22372687e+00 -9.98743474e-02 3.48393232e-01
1.98133856e-01 8.14864039e-02 1.12803090e+00 -2.17669740e-01
-8.20444226e-01 2.34640896e-01 -5.03986835e-01 4.74207878e-01
-2.13677257e-01 2.44827509e-01 -9.72122729e-01 3.56299267e-03
-1.56354204e-01 -1.54653966e-01 1.16758418e+00 6.45946622e-01
1.48092461e+00 -1.13799348e-01 -1.61152557e-01 1.13414967e+00
1.25801444e+00 -2.06649706e-01 4.82243210e-01 9.92089361e-02
1.45250678e+00 8.09579730e-01 3.19577336e-01 5.59724346e-02
5.33644438e-01 6.90595925e-01 6.34847999e-01 -6.89919531e-01
-1.51151136e-01 -1.47772431e-01 -5.20554818e-02 8.20567131e-01
7.34898895e-02 2.57805258e-01 -7.93052495e-01 4.21465099e-01
-2.11125207e+00 -7.19522536e-01 -4.68046814e-01 1.99007368e+00
7.84682393e-01 -1.00068189e-01 -4.02463228e-02 -1.99472815e-01
6.14486516e-01 1.49842259e-02 -6.07463360e-01 -3.47551942e-01
-2.01876223e-01 4.98278916e-01 6.36914432e-01 8.44269514e-01
-1.19463503e+00 1.33787072e+00 6.38089228e+00 6.83211803e-01
-9.48209584e-01 6.31183609e-02 7.79524267e-01 2.00166211e-01
-5.75392187e-01 -8.95903334e-02 -6.04871511e-01 1.26824558e-01
3.66204828e-01 2.94475257e-01 3.95774066e-01 8.03679824e-01
-1.78268582e-01 1.95276827e-01 -1.27987111e+00 1.01503980e+00
-4.28065248e-02 -1.52346563e+00 2.69855767e-01 -3.41859390e-03
7.76386738e-01 1.98069960e-02 4.02256668e-01 3.39259878e-02
4.04777557e-01 -1.08233595e+00 8.61938298e-01 3.54755253e-01
3.81215990e-01 -7.70811915e-01 4.68145192e-01 2.15400800e-01
-1.02708638e+00 9.06289443e-02 -5.32643855e-01 3.06311965e-01
9.01283100e-02 6.94492042e-01 -3.94277245e-01 2.94582874e-01
9.02352393e-01 9.11172986e-01 -5.98070800e-01 8.73004138e-01
-3.97993565e-01 4.16704267e-01 -3.37525576e-01 3.68208468e-01
5.04673541e-01 -6.50237918e-01 1.88732579e-01 1.27575374e+00
1.09342650e-01 -9.71635431e-03 2.95784716e-02 1.16360426e+00
-2.33147070e-01 1.56235904e-01 -4.09837902e-01 3.95420760e-01
1.72399104e-01 1.62591326e+00 -9.33755398e-01 -1.39233038e-01
-3.99680853e-01 1.23921120e+00 7.45650649e-01 5.78682959e-01
-8.48439217e-01 -7.23753348e-02 7.97622979e-01 -1.47402138e-01
4.47670519e-01 -3.62721682e-01 -5.52709222e-01 -9.96553838e-01
-2.18579993e-01 -6.38048470e-01 1.58765256e-01 -7.38482296e-01
-1.42384958e+00 5.61718941e-01 1.15104122e-02 -6.00470603e-01
2.70493515e-03 -8.83297145e-01 -6.11837626e-01 8.06422710e-01
-1.38217604e+00 -1.39331853e+00 -3.23814362e-01 8.29139829e-01
3.41923594e-01 1.71810538e-01 6.19948506e-01 2.77104914e-01
-6.97860003e-01 4.16161716e-01 -1.86009943e-01 3.59561622e-01
3.97624999e-01 -1.67697716e+00 4.75251764e-01 9.74754572e-01
6.28872454e-01 8.06861341e-01 4.88980949e-01 -1.93639949e-01
-1.29411232e+00 -1.15015137e+00 5.76669037e-01 -4.25880939e-01
6.66785359e-01 -5.98785281e-01 -1.15646493e+00 6.89660251e-01
2.25424260e-01 2.03129873e-01 5.26880264e-01 1.79555148e-01
-5.49109936e-01 -1.57699376e-01 -8.24617565e-01 6.35446429e-01
1.22757101e+00 -5.38757920e-01 -3.38697582e-01 4.79839087e-01
7.22845197e-01 -5.57995260e-01 -8.05507243e-01 2.37063199e-01
4.16781932e-01 -1.11275482e+00 1.25792396e+00 -5.38020849e-01
4.09504473e-01 -4.36915547e-01 -8.79067704e-02 -1.15168583e+00
-6.13006175e-01 -1.97080120e-01 2.03352422e-01 1.12010670e+00
6.31002784e-01 -4.38367069e-01 8.57739091e-01 7.80249655e-01
-2.39316896e-01 -8.28629375e-01 -6.61065400e-01 -2.64614195e-01
3.33220810e-01 -5.63542843e-01 6.32625341e-01 9.00551856e-01
-5.71277797e-01 4.80488777e-01 -1.81017146e-01 3.33319604e-01
7.73948789e-01 2.12428778e-01 6.23905540e-01 -1.07990324e+00
-4.86905873e-01 -6.18577182e-01 -3.42279315e-01 -1.43690133e+00
4.25801218e-01 -1.11664963e+00 1.60649732e-01 -1.59361100e+00
3.25344205e-01 -6.38360739e-01 -2.57454425e-01 5.13040066e-01
-2.07351208e-01 5.46962976e-01 2.42334157e-01 1.92793369e-01
-5.72447658e-01 3.86122257e-01 1.32609689e+00 -4.25252348e-01
-1.06832735e-01 -2.32809380e-01 -7.49040306e-01 8.93126190e-01
5.93803763e-01 -2.04034120e-01 -4.95035529e-01 -8.55598152e-01
3.75540704e-01 -3.69372010e-01 5.32619178e-01 -7.83980072e-01
4.20677423e-01 -1.32099837e-01 3.43803495e-01 -3.80223215e-01
3.99162471e-01 -8.68910134e-01 4.66224086e-03 1.60836607e-01
-4.23047006e-01 -8.32422674e-02 -4.64750975e-02 3.62288088e-01
-3.49985778e-01 -1.45267785e-01 9.20143545e-01 -1.04858458e-01
-8.25041652e-01 7.08195865e-01 -5.89646250e-02 3.71105969e-02
8.96209598e-01 -2.97224373e-01 -1.34266779e-01 -1.42376676e-01
-9.10850465e-01 2.74399340e-01 5.60509980e-01 2.65269101e-01
6.22660041e-01 -1.31529593e+00 -5.54564178e-01 4.22507495e-01
6.91736564e-02 3.52933109e-01 2.15273812e-01 8.83723021e-01
-6.52896464e-01 3.58710401e-02 -2.98403442e-01 -1.10603166e+00
-1.13913405e+00 4.37309653e-01 4.98129338e-01 -5.86389378e-02
-6.29112065e-01 1.35307539e+00 8.25096488e-01 -4.74049002e-01
1.29511401e-01 -4.17164505e-01 -2.70439386e-01 -1.76991761e-01
3.33335817e-01 -5.94618693e-02 4.96256202e-02 -9.30278778e-01
-3.90038401e-01 1.10207939e+00 -8.84612426e-02 -1.82748750e-01
1.27571702e+00 -3.73232588e-02 -5.65820098e-01 2.60254443e-01
1.41604316e+00 -2.08221629e-01 -1.46442223e+00 -2.33716160e-01
-4.48238617e-03 -3.15625370e-01 4.08977605e-02 -4.83429909e-01
-1.48482513e+00 9.96535182e-01 4.71988082e-01 1.00619584e-01
1.02515018e+00 2.68370003e-01 6.75962210e-01 2.71885484e-01
1.46105245e-01 -9.26350832e-01 8.22064951e-02 6.13388300e-01
7.93025136e-01 -1.11612737e+00 -3.72397840e-01 -6.22375011e-01
-5.55994153e-01 9.92272675e-01 5.82243502e-01 -2.58452445e-01
8.24332535e-01 3.17492813e-01 1.75552204e-01 -2.99016297e-01
-3.17404956e-01 -4.34858203e-01 5.60499847e-01 6.65941417e-01
4.00557846e-01 6.84101507e-02 -2.02444419e-01 3.28135729e-01
-1.71122044e-01 -4.67286229e-01 1.70074761e-01 5.04696071e-01
-4.01015311e-01 -1.36245966e+00 -2.51597226e-01 2.89094359e-01
-7.05523640e-02 -1.30717248e-01 -5.44625759e-01 6.28294110e-01
3.01318139e-01 6.56781137e-01 3.07938308e-01 -3.20569456e-01
3.27844620e-01 -8.58455077e-02 7.14166999e-01 -6.30131721e-01
-5.62561631e-01 -2.46842969e-02 -3.40193182e-01 -8.65360856e-01
-5.18620610e-01 -4.76560950e-01 -1.54037917e+00 -2.51320213e-01
2.56282799e-02 1.17433153e-01 7.29060888e-01 9.50411499e-01
2.17641339e-01 5.28534412e-01 4.26841110e-01 -9.16499376e-01
-7.86885098e-02 -5.53602755e-01 -5.71409583e-01 7.94617891e-01
1.66309670e-01 -5.34177780e-01 -1.77737907e-01 1.54443294e-01] | [9.655721664428711, 0.5238730311393738] |
f57ad705-c41c-4f7b-9828-953590a3dc85 | multi-stage-distillation-framework-for-cross-2 | 2209.05869 | null | https://arxiv.org/abs/2209.05869v1 | https://arxiv.org/pdf/2209.05869v1.pdf | Multi-stage Distillation Framework for Cross-Lingual Semantic Similarity Matching | Previous studies have proved that cross-lingual knowledge distillation can significantly improve the performance of pre-trained models for cross-lingual similarity matching tasks. However, the student model needs to be large in this operation. Otherwise, its performance will drop sharply, thus making it impractical to be deployed to memory-limited devices. To address this issue, we delve into cross-lingual knowledge distillation and propose a multi-stage distillation framework for constructing a small-size but high-performance cross-lingual model. In our framework, contrastive learning, bottleneck, and parameter recurrent strategies are combined to prevent performance from being compromised during the compression process. The experimental results demonstrate that our method can compress the size of XLM-R and MiniLM by more than 50\%, while the performance is only reduced by about 1%. | ['Xuefeng Yang', 'Qi Ju', 'Zhe Zhao', 'Yuejian Fang', 'Weijie Liu', 'Kunbo Ding'] | 2022-09-13 | multi-stage-distillation-framework-for-cross-1 | https://aclanthology.org/2022.findings-naacl.167 | https://aclanthology.org/2022.findings-naacl.167.pdf | findings-naacl-2022-7 | ['xlm-r'] | ['natural-language-processing'] | [-3.67969126e-02 -2.32942745e-01 -5.55676639e-01 -4.37422574e-01
-1.06314170e+00 -3.87741268e-01 3.47296864e-01 2.21674219e-01
-6.50042057e-01 4.57289428e-01 -1.96144562e-02 -7.52005100e-01
1.03225209e-01 -5.46472490e-01 -7.14726985e-01 -3.22817773e-01
3.51462156e-01 3.58079046e-01 2.96039701e-01 -7.24694505e-02
8.11981782e-02 6.54230267e-02 -1.56876624e+00 2.15005562e-01
1.27396238e+00 7.58234441e-01 5.73331237e-01 2.59988934e-01
-4.82265830e-01 5.69859445e-01 -3.89153868e-01 -6.52135432e-01
1.34234995e-01 -2.63723403e-01 -7.60182500e-01 -4.84687686e-01
4.77642119e-01 -3.18389505e-01 -5.08709788e-01 1.15748680e+00
5.76156080e-01 -3.19469385e-02 4.00051326e-01 -9.30903673e-01
-4.56265867e-01 1.06095076e+00 -6.30831599e-01 1.52508676e-01
4.69778553e-02 -6.62915334e-02 1.00833082e+00 -7.68092453e-01
2.13966101e-01 1.23711038e+00 6.80763841e-01 2.97078282e-01
-9.72983062e-01 -1.07248902e+00 1.12562068e-01 2.64123172e-01
-1.49009609e+00 -7.02012837e-01 5.48993170e-01 2.32079014e-01
1.17027950e+00 2.05889568e-01 3.49365324e-01 6.39964938e-01
-1.73596859e-01 1.03847313e+00 9.65584517e-01 -5.81436872e-01
-2.55922496e-01 1.95932925e-01 -1.12273470e-01 7.41402864e-01
4.74029809e-01 -2.46323138e-01 -5.01624703e-01 -1.90400288e-01
4.20894742e-01 -1.99929804e-01 -9.23933163e-02 -1.63912565e-01
-8.60479057e-01 7.31672287e-01 3.37927550e-01 5.44949710e-01
1.20382965e-01 -2.50641964e-02 7.63990283e-01 4.35998172e-01
4.67766196e-01 5.34483552e-01 -5.76589108e-01 -4.38771278e-01
-1.25488949e+00 -3.92388925e-02 6.65472746e-01 1.08314264e+00
7.09252894e-01 -1.23909779e-01 1.48726478e-01 1.14451468e+00
2.37945050e-01 6.76806927e-01 8.73883784e-01 -7.21995592e-01
8.03986669e-01 6.95921659e-01 -4.26116288e-01 -7.09734738e-01
-3.58801857e-02 -3.54022086e-01 -8.25611651e-01 -5.39858282e-01
1.14350565e-01 7.27692991e-02 -7.61061847e-01 1.62299824e+00
1.66545719e-01 2.82403052e-01 1.25300288e-01 4.98936683e-01
6.94587052e-01 6.35719836e-01 1.28338277e-01 -3.54462802e-01
1.42228925e+00 -1.11065793e+00 -6.98764324e-01 -4.22511607e-01
1.31202948e+00 -1.20529020e+00 1.33733082e+00 4.63083535e-02
-1.46787059e+00 -7.54370272e-01 -1.22888041e+00 -2.30620831e-01
-2.20546097e-01 7.79349655e-02 6.59800410e-01 5.75972199e-01
-7.48378754e-01 5.30961990e-01 -8.22903156e-01 -1.98071197e-01
1.38543308e-01 2.54653186e-01 -2.63893288e-02 -4.62036610e-01
-1.30115831e+00 9.59818959e-01 5.81107378e-01 -2.16096759e-01
-1.93085834e-01 -7.45839477e-01 -7.46734619e-01 1.56830341e-01
2.20784888e-01 -4.47058886e-01 1.18031037e+00 -3.23126167e-01
-1.25049293e+00 7.65255809e-01 -3.78557056e-01 -4.54017937e-01
2.98524499e-01 -4.57608908e-01 -4.18516934e-01 -1.93575740e-01
-1.96524501e-01 4.57123756e-01 4.27232504e-01 -9.20307636e-01
-6.87538743e-01 -2.14391649e-01 -1.75765168e-03 4.71978992e-01
-8.58848274e-01 9.45714489e-02 -1.16372406e+00 -7.40629554e-01
1.56690702e-01 -9.33873475e-01 -1.29111364e-01 -4.18820709e-01
-1.25621408e-01 -3.58407646e-01 7.99252331e-01 -6.05154872e-01
1.69627011e+00 -2.09361434e+00 -2.11287707e-01 1.04362249e-01
-1.00454427e-01 8.59435797e-01 -1.69607103e-01 1.86743557e-01
2.87926853e-01 7.42969960e-02 -3.38164978e-02 -6.00977778e-01
-2.99331993e-02 4.21661168e-01 -2.58776933e-01 2.31300920e-01
-3.26116949e-01 9.16343451e-01 -7.52667487e-01 -9.38513100e-01
1.03910476e-01 3.94025892e-01 -6.56309843e-01 4.42456722e-01
-1.55430555e-01 -1.79323047e-01 -3.63852769e-01 4.97882158e-01
7.03085423e-01 -1.92482740e-01 5.49920976e-01 -3.58980924e-01
1.11904472e-01 8.70127141e-01 -9.22179997e-01 2.06143737e+00
-8.84107471e-01 3.29777688e-01 -2.41260707e-01 -9.96613264e-01
9.91395414e-01 2.76397824e-01 4.24175769e-01 -1.09078705e+00
1.13143101e-01 5.55270314e-01 -1.05919644e-01 -1.96992129e-01
7.25718379e-01 1.78692788e-02 -2.50253886e-01 7.80192018e-01
-3.11126202e-01 -7.62956068e-02 1.37373433e-01 9.43594351e-02
7.67977118e-01 7.08967671e-02 1.03727095e-02 -1.30414471e-01
6.46812379e-01 -1.56939551e-01 6.20572448e-01 5.53669930e-01
-2.18860116e-02 -1.28406007e-02 -1.39098421e-01 -9.34381858e-02
-1.08730805e+00 -7.01819062e-01 -1.79866984e-01 1.30362952e+00
1.45994082e-01 -7.13344932e-01 -9.86548543e-01 -5.41628480e-01
1.23027325e-01 5.76722145e-01 1.95482790e-01 -7.49940693e-01
-1.10753262e+00 -8.47445250e-01 9.78078604e-01 4.99931365e-01
7.42251158e-01 -6.71041846e-01 -2.52214730e-01 1.89431190e-01
-3.30206513e-01 -1.34065950e+00 -7.09509790e-01 1.07117176e-01
-1.18429673e+00 -7.96314001e-01 -5.38199544e-01 -1.07438564e+00
4.96407211e-01 5.28308213e-01 1.10705554e+00 1.65006727e-01
-7.11431652e-02 -1.99774817e-01 -2.61449248e-01 1.63147822e-02
-5.88668764e-01 6.53565645e-01 2.76308984e-01 -6.13418341e-01
7.95696139e-01 -7.35834539e-01 -5.86028397e-01 3.88369083e-01
-7.09969580e-01 1.19370140e-01 8.08892906e-01 6.69529140e-01
5.92623115e-01 1.33930027e-01 5.33981264e-01 -6.92504168e-01
7.09862590e-01 -1.34265319e-01 -6.11713946e-01 6.05113328e-01
-1.11734724e+00 4.98371750e-01 8.31267536e-01 -6.61039710e-01
-9.05825913e-01 -2.56272554e-01 -4.67337035e-02 -5.44554949e-01
3.45135272e-01 6.32805407e-01 -1.34447515e-01 -1.01379685e-01
3.14847469e-01 3.91442001e-01 -1.10108428e-01 -9.00851369e-01
5.97267210e-01 9.61554766e-01 6.76546454e-01 -7.83292770e-01
7.53223538e-01 -2.90623754e-02 -4.58862692e-01 -3.85042280e-01
-8.85062754e-01 -5.55790842e-01 -4.30758238e-01 3.25969160e-01
3.18955928e-01 -1.23835170e+00 -6.08561993e-01 2.26058304e-01
-8.66679013e-01 -1.81839898e-01 4.98577170e-02 6.44803643e-01
-3.65792096e-01 5.87354064e-01 -8.54992867e-01 -5.07041335e-01
-7.23211467e-01 -1.14171576e+00 7.09644735e-01 8.50398913e-02
-8.78121480e-02 -8.79323006e-01 2.67216787e-02 7.40173340e-01
5.91669023e-01 -7.10669279e-01 1.06508303e+00 -5.89321971e-01
-5.70068955e-01 -1.61992595e-01 -3.33323240e-01 3.29928905e-01
1.38993347e-02 -4.21044409e-01 -8.07393074e-01 -6.26477897e-01
-1.18845426e-01 -6.60268009e-01 7.84940898e-01 -1.76410213e-01
1.43170404e+00 -1.99245885e-01 -5.32132983e-01 9.11960304e-01
1.19157302e+00 1.66548118e-01 3.72804046e-01 2.95527160e-01
7.16735661e-01 2.10657850e-01 6.96176946e-01 1.63903683e-01
6.66446686e-01 9.99713123e-01 -1.59389660e-01 -1.15578704e-01
-3.73308539e-01 -6.15864754e-01 4.95608687e-01 1.86380184e+00
2.15414345e-01 5.74284000e-03 -7.54396021e-01 6.40692115e-01
-1.70705509e+00 -5.58515191e-01 2.58710355e-01 2.39855528e+00
1.43477952e+00 2.81949162e-01 -1.16948828e-01 3.19747999e-02
4.89299327e-01 1.36752933e-01 -6.74783468e-01 -6.06683075e-01
3.85597092e-03 2.47326881e-01 6.42994881e-01 4.34539974e-01
-7.69616902e-01 1.30510497e+00 6.38970375e+00 1.29065216e+00
-1.18838882e+00 1.49478644e-01 4.99032766e-01 -1.78211108e-01
-4.47200239e-01 -1.37927057e-02 -1.25208056e+00 5.24182379e-01
1.34978628e+00 -4.72540796e-01 2.97612667e-01 9.30876017e-01
-1.09662451e-01 9.74061936e-02 -9.20454264e-01 1.06619191e+00
7.04652220e-02 -9.55091059e-01 8.23074281e-02 4.72688526e-02
6.12378359e-01 2.48647898e-01 3.35867554e-02 6.67168081e-01
4.15232807e-01 -8.27431142e-01 1.91220701e-01 -6.81337714e-02
9.95839775e-01 -1.04528487e+00 5.52691758e-01 4.71041054e-01
-1.45222533e+00 2.69100517e-01 -7.21909165e-01 3.59443575e-01
1.71707243e-01 5.63795865e-01 -8.00228417e-01 4.89252359e-01
5.42696893e-01 4.71330494e-01 -5.31559348e-01 7.05752015e-01
6.69215247e-02 6.52660847e-01 -5.17202675e-01 1.09466486e-01
2.78226495e-01 2.39030067e-02 -1.04264885e-01 1.32181096e+00
4.55501616e-01 -1.74051747e-01 1.36016354e-01 4.40598577e-01
-3.75316173e-01 3.07771236e-01 -4.63305563e-01 -5.68329962e-03
8.51273775e-01 8.51990938e-01 1.56993810e-02 -5.44692039e-01
-6.25735402e-01 1.06305265e+00 5.76227605e-01 4.30771261e-02
-1.03831065e+00 -4.70011950e-01 7.83714950e-01 -2.43346132e-02
3.05605412e-01 -2.73910910e-01 -2.09467560e-01 -1.42047429e+00
2.77812928e-01 -1.02415228e+00 5.15678823e-01 -3.92927855e-01
-9.82014358e-01 5.74178278e-01 -4.11433093e-02 -1.01911163e+00
-4.59393650e-01 -8.97785276e-02 -1.93657905e-01 9.69531059e-01
-1.93477380e+00 -1.02255297e+00 1.52087975e-02 5.98780572e-01
3.82822305e-01 -1.89359277e-01 9.46073353e-01 8.13409328e-01
-7.00060904e-01 1.51842630e+00 4.53620702e-01 5.83570637e-02
9.22484815e-01 -7.12786019e-01 6.24279201e-01 7.71531820e-01
1.91726834e-01 9.31925893e-01 2.84041047e-01 -4.60250080e-01
-1.58852470e+00 -1.06922078e+00 1.41126704e+00 -1.16733033e-02
6.41439974e-01 -3.33401978e-01 -1.27081859e+00 3.78878534e-01
5.09590656e-02 -1.36174150e-02 7.42429554e-01 4.13069457e-01
-8.04746330e-01 -3.35691571e-01 -8.28558266e-01 7.48867631e-01
1.11071706e+00 -9.21243727e-01 -6.95211649e-01 1.87854439e-01
9.37259495e-01 -3.72175813e-01 -1.31317866e+00 5.47514737e-01
5.53567410e-01 -5.84188521e-01 1.18195331e+00 -4.19379711e-01
1.97058916e-01 -3.76614816e-02 -1.57451123e-01 -9.72981036e-01
-4.03725430e-02 -6.17473483e-01 -4.94257122e-01 1.39019513e+00
4.31803375e-01 -4.79013234e-01 7.99572051e-01 4.66985434e-01
-5.80090173e-02 -8.21484268e-01 -8.50259483e-01 -1.03608966e+00
2.84910500e-01 -4.53896672e-01 7.38619387e-01 1.13802791e+00
3.90910298e-01 4.87280786e-01 -4.90990341e-01 -1.03378057e-01
4.78171468e-01 4.82401937e-01 8.39912295e-01 -8.20704520e-01
-3.72735500e-01 -3.77693921e-01 9.49182808e-02 -1.73478115e+00
2.91536868e-01 -1.07928205e+00 -4.00947854e-02 -1.01355052e+00
3.71395797e-01 -1.13937986e+00 -5.57537556e-01 4.99940008e-01
-4.60295707e-01 4.17336710e-02 1.76182285e-01 5.28460741e-01
-6.28908813e-01 5.81762433e-01 1.19390881e+00 6.86956868e-02
-5.97250313e-02 -2.63765216e-01 -7.03908265e-01 4.84603852e-01
8.63759398e-01 -4.86278504e-01 -6.32570803e-01 -6.73862457e-01
2.73488183e-03 -2.26019938e-02 -4.65016127e-01 -8.86665940e-01
4.08620089e-01 8.68129823e-03 -1.33271098e-01 -5.01418173e-01
4.17949975e-01 -7.57632196e-01 -2.13536993e-01 6.19877040e-01
-3.09760481e-01 2.70651579e-01 3.65016162e-01 1.36305258e-01
-4.24043745e-01 -1.44364998e-01 7.57434964e-01 -7.17538446e-02
-4.93015975e-01 2.94712305e-01 1.70109138e-01 2.15333730e-01
5.64152062e-01 1.02589451e-01 -3.08748484e-01 -1.37176335e-01
-1.08410604e-01 2.86057472e-01 5.73408365e-01 6.12799108e-01
3.19613785e-01 -1.54898608e+00 -5.08372843e-01 3.62852931e-01
2.61709034e-01 7.26824626e-02 6.75683990e-02 6.95788264e-01
-2.58226305e-01 8.73853385e-01 2.04360917e-01 -4.08483714e-01
-1.53803194e+00 6.01988852e-01 3.76472063e-02 -7.45825350e-01
-5.60980380e-01 7.39562571e-01 -1.03278197e-01 -5.81468344e-01
4.08124655e-01 -3.62820715e-01 7.77619258e-02 -2.59172350e-01
5.58794916e-01 1.25876352e-01 9.60086063e-02 -4.80120063e-01
-2.26664275e-01 6.66394234e-01 -5.26870787e-01 2.45828733e-01
1.08740962e+00 -3.26744944e-01 -5.45697138e-02 2.06277296e-01
1.54817140e+00 -1.08522654e-01 -5.68573415e-01 -9.20777559e-01
1.40111391e-02 -3.26087117e-01 2.80383289e-01 -4.43897307e-01
-1.01101410e+00 8.75530660e-01 4.35355484e-01 -2.19962686e-01
1.25615120e+00 -5.71669675e-02 1.68240893e+00 6.34332538e-01
3.99076879e-01 -1.29306316e+00 -1.19315118e-01 6.09082758e-01
1.25848025e-01 -1.20106971e+00 1.80344209e-01 -4.13259715e-01
-4.04278129e-01 7.54719257e-01 7.50938177e-01 3.40546817e-01
4.88551795e-01 3.73386830e-01 1.05201434e-02 4.00550365e-02
-8.33523989e-01 -1.62646353e-01 4.41110134e-01 6.77558333e-02
7.42792904e-01 7.92251751e-02 -3.32296580e-01 2.40671620e-01
-3.82895023e-01 -8.59103575e-02 -1.77326113e-01 1.04861355e+00
-5.95591784e-01 -1.53216600e+00 8.25193431e-03 2.95698851e-01
-6.09541655e-01 -4.87872064e-01 2.23877668e-01 5.43156743e-01
3.62774618e-02 7.64783144e-01 1.24413967e-01 -6.35130703e-01
2.83279508e-01 2.06586346e-01 4.32314098e-01 -2.40732595e-01
-5.41029751e-01 1.06993474e-01 1.64319530e-01 -5.07764220e-01
-2.80009717e-01 -3.63820702e-01 -1.07589173e+00 -7.55022645e-01
-5.38179278e-01 4.14361060e-01 6.54031694e-01 8.91320646e-01
5.02133012e-01 2.02756301e-01 6.11764848e-01 -5.22414185e-02
-7.20760882e-01 -9.79684353e-01 -6.54688850e-02 2.98715353e-01
-2.04599395e-01 -3.56458634e-01 -2.81278074e-01 -1.03673980e-01] | [13.631701469421387, 7.037595272064209] |
df844211-4dbf-4ceb-836d-98de495c2b00 | game-theoretic-algorithms-for-conditional | 2208.09551 | null | https://arxiv.org/abs/2208.09551v1 | https://arxiv.org/pdf/2208.09551v1.pdf | Game-Theoretic Algorithms for Conditional Moment Matching | A variety of problems in econometrics and machine learning, including instrumental variable regression and Bellman residual minimization, can be formulated as satisfying a set of conditional moment restrictions (CMR). We derive a general, game-theoretic strategy for satisfying CMR that scales to nonlinear problems, is amenable to gradient-based optimization, and is able to account for finite sample uncertainty. We recover the approaches of Dikkala et al. and Dai et al. as special cases of our general framework before detailing various extensions and how to efficiently solve the game defined by CMR. | ['Zhiwei Steven Wu', 'J. Andrew Bagnell', 'Sanjiban Choudhury', 'Gokul Swamy'] | 2022-08-19 | null | null | null | null | ['econometrics'] | ['miscellaneous'] | [-4.47512530e-02 2.99114048e-01 -5.19118845e-01 -2.67742306e-01
-1.27955842e+00 -7.70463407e-01 5.89451969e-01 -2.52488226e-01
-3.49397212e-01 1.01249886e+00 -1.35316700e-01 -8.92406166e-01
-9.03560162e-01 -2.87569433e-01 -4.40413445e-01 -7.06188381e-01
-2.87040442e-01 5.97047567e-01 -3.55742246e-01 -4.68432941e-02
4.28649664e-01 5.20325124e-01 -7.48247147e-01 -6.49156988e-01
9.24304068e-01 1.17280066e+00 2.99996193e-02 7.21894205e-01
3.15014124e-01 9.36386943e-01 -1.79224744e-01 -5.91261446e-01
4.19368923e-01 -2.77082831e-01 -7.54921496e-01 1.78997383e-01
-2.17198417e-01 -1.93192273e-01 -2.45077372e-01 1.21013141e+00
5.14344335e-01 4.61295366e-01 1.07236803e+00 -1.48447394e+00
-7.37183154e-01 6.97863042e-01 -8.57164919e-01 1.83523968e-01
1.16837367e-01 -1.15053551e-02 1.10671377e+00 -6.79054081e-01
4.35054541e-01 1.37697768e+00 6.34940207e-01 3.67843479e-01
-1.34942055e+00 -3.29543322e-01 1.33056298e-01 -2.82757998e-01
-1.30156791e+00 -6.22376978e-01 6.42936647e-01 -5.37762105e-01
6.53738618e-01 3.52049619e-01 9.92066339e-02 6.00848675e-01
1.50962994e-01 9.84159887e-01 1.13747644e+00 -5.04201829e-01
4.09152985e-01 -5.59981801e-02 2.44935509e-02 5.18701553e-01
1.56406015e-01 1.98044956e-01 -4.95190024e-02 -5.00103295e-01
9.81940746e-01 -1.23532437e-01 5.20485230e-02 -6.42375052e-01
-7.78510153e-01 1.52881181e+00 -2.80218273e-01 -5.14215454e-02
-4.11794543e-01 2.42023215e-01 1.52848437e-01 3.44085157e-01
7.31347799e-01 5.51162243e-01 -6.87406242e-01 -1.36133626e-01
-8.86224508e-01 6.40562236e-01 9.69461381e-01 1.00388324e+00
5.71838677e-01 3.33473057e-01 -1.11385055e-01 6.98063195e-01
4.54203904e-01 6.85483515e-01 1.63083479e-01 -1.62744141e+00
3.91683906e-01 -2.38920763e-01 5.09511709e-01 -8.23207498e-01
-6.59247041e-01 -2.80222476e-01 -5.86863518e-01 5.10843471e-02
6.38848543e-01 -7.23717928e-01 -5.13500154e-01 1.85108161e+00
1.34580895e-01 1.69378877e-01 9.97120664e-02 6.47682190e-01
1.56605110e-01 5.26160181e-01 -2.27619201e-01 -8.15050304e-01
8.96281302e-01 -5.02671480e-01 -8.12113464e-01 -3.52564454e-01
6.14805400e-01 -5.84632754e-01 6.85628414e-01 4.21392202e-01
-1.34258425e+00 1.62052587e-01 -5.97643375e-01 -6.77600428e-02
1.51959315e-01 -1.67450622e-01 1.28877795e+00 7.72101223e-01
-1.17968857e+00 5.72615623e-01 -6.52055323e-01 1.13948341e-02
1.55012503e-01 5.87976038e-01 -9.93737280e-02 2.40685567e-01
-9.49433923e-01 7.25170195e-01 5.54551370e-02 1.63918555e-01
-6.79156005e-01 -5.87567747e-01 -1.17263293e+00 7.44091049e-02
7.59960115e-01 -4.98019367e-01 1.76311767e+00 -7.71644115e-01
-1.94385648e+00 7.73715615e-01 -1.15989439e-01 -5.12250841e-01
6.26880467e-01 5.51443398e-02 -1.27452224e-01 -2.18014531e-02
3.77070725e-01 1.53047279e-01 6.49505079e-01 -7.58340836e-01
-4.61166948e-01 -3.84216607e-01 2.58460701e-01 1.33432344e-01
3.73785257e-01 6.35283947e-01 -2.95763195e-01 -6.61840022e-01
1.76173419e-01 -1.09693241e+00 -8.35758686e-01 -6.38037562e-01
-4.23256457e-01 -4.09749858e-02 3.35678604e-04 -5.58078349e-01
1.12155867e+00 -1.78624141e+00 3.72736663e-01 4.33700085e-01
-1.58985928e-02 -2.94971913e-01 -5.74857928e-02 4.31135595e-01
-1.68031365e-01 2.84045070e-01 -6.69459403e-01 -5.88909447e-01
4.75438625e-01 1.53455898e-01 -2.36017719e-01 8.99569750e-01
1.75748855e-01 1.19757318e+00 -7.95300603e-01 -2.56828338e-01
1.21960782e-01 -2.56310999e-01 -6.67260408e-01 -2.47099176e-02
-3.37408483e-02 4.72890198e-01 -5.91983020e-01 7.76534855e-01
6.56643808e-01 -4.08870801e-02 3.19714814e-01 5.91771483e-01
-3.13210458e-01 1.56843916e-01 -1.52827239e+00 1.44113445e+00
-2.88676172e-01 5.47399640e-01 6.37619734e-01 -1.54739058e+00
5.30489385e-01 3.28317225e-01 5.90768814e-01 -2.88459450e-01
3.08415979e-01 3.18713903e-01 -2.86431462e-01 -5.01216769e-01
4.16157603e-01 -6.08453155e-01 -6.27232850e-01 4.35453236e-01
1.85922801e-01 -4.30838764e-01 1.94294199e-01 1.53111503e-01
7.03036427e-01 -1.60206899e-01 8.34741533e-01 -6.97661281e-01
2.35887572e-01 -3.49186510e-01 7.00877190e-01 1.28779793e+00
-2.21101031e-01 3.72344643e-01 1.09806609e+00 4.47440669e-02
-7.86998510e-01 -8.92417431e-01 -2.89950222e-01 1.12986207e+00
-4.06078100e-01 -3.63972075e-02 -5.18070877e-01 -2.46062189e-01
2.39649490e-01 5.54815710e-01 -5.86217105e-01 2.23395929e-01
-1.50569379e-01 -1.06175804e+00 4.31487292e-01 2.88315654e-01
-3.16344649e-02 -5.37855685e-01 -1.21300250e-01 2.57139087e-01
-6.99036717e-02 -9.62909639e-01 -5.16906381e-01 4.16765749e-01
-8.16669703e-01 -9.99069691e-01 -6.11295283e-01 -3.90999109e-01
2.18185753e-01 7.67252967e-02 1.06413209e+00 -3.86992246e-01
6.06095828e-02 6.18005037e-01 8.79487768e-03 -6.67398512e-01
-2.12497368e-01 -2.11845949e-01 2.57975489e-01 -2.86967624e-02
7.14193210e-02 -4.81057495e-01 -1.25150606e-01 1.65117867e-02
-8.72470379e-01 -6.08250618e-01 2.70164430e-01 8.98944378e-01
5.02897441e-01 -1.21894136e-01 6.38815999e-01 -1.10701334e+00
9.01239991e-01 -8.55523527e-01 -1.27915537e+00 6.92983568e-02
-8.07525575e-01 2.65771061e-01 1.86725914e-01 -2.52710968e-01
-1.07030141e+00 -4.81705926e-02 4.54243682e-02 -1.36418909e-01
1.37800947e-01 1.07819188e+00 -1.67499036e-01 -3.57315540e-01
3.17033470e-01 -2.96398610e-01 1.84725616e-02 -5.22566557e-01
5.27108908e-01 5.38249373e-01 3.56286138e-01 -7.24470854e-01
6.90170348e-01 3.66352320e-01 3.85183901e-01 -5.10381758e-01
-6.56525135e-01 -4.35259402e-01 -4.37922776e-01 8.27299654e-02
5.49478710e-01 -7.74701476e-01 -9.68105733e-01 3.24988902e-01
-8.87872994e-01 -5.35264790e-01 -2.90799230e-01 8.15977931e-01
-1.09873879e+00 3.41590971e-01 -6.69006348e-01 -1.57749116e+00
2.75400132e-01 -1.04661226e+00 8.79967868e-01 1.62510946e-01
1.18582703e-01 -1.45127749e+00 2.96958923e-01 6.35285676e-02
1.50036797e-01 3.33536059e-01 6.14975095e-01 -3.74460608e-01
-2.84003973e-01 -3.32831651e-01 6.50710687e-02 1.40149951e-01
-2.22519800e-01 1.35484144e-01 -5.80716312e-01 -2.22678304e-01
3.94618154e-01 -2.03853965e-01 8.49753201e-01 1.31548178e+00
9.45959389e-01 -5.86336970e-01 7.12029217e-03 8.73972058e-01
1.27152038e+00 2.50124663e-01 3.28124970e-01 3.34348977e-01
2.07511723e-01 6.54144704e-01 6.47198021e-01 7.10787117e-01
4.71141905e-01 4.94358569e-01 3.30643237e-01 7.76062608e-02
1.06983960e+00 5.21693677e-02 2.31780261e-01 5.57635903e-01
-1.62426218e-01 -1.76535308e-01 -7.15909421e-01 6.13366246e-01
-2.42063522e+00 -1.16824114e+00 -1.65726379e-01 2.19303179e+00
5.87590039e-01 -1.75302655e-01 5.31536818e-01 -1.90354571e-01
5.75686991e-01 2.32147038e-01 -4.65240628e-01 -7.43812382e-01
-3.07141691e-01 1.75793335e-01 1.23911381e+00 9.29387748e-01
-1.14112318e+00 9.32120681e-01 8.65270710e+00 8.95529687e-01
-5.34469068e-01 2.40090042e-01 7.33139038e-01 -9.19478014e-02
-4.61861998e-01 3.00134897e-01 -3.72496545e-01 1.00244209e-01
9.09170091e-01 -3.58066142e-01 7.71293104e-01 7.78706968e-01
5.93094826e-01 -4.23831373e-01 -7.13765383e-01 9.01189387e-01
-3.15218002e-01 -9.44425881e-01 -7.31250346e-01 3.54048222e-01
1.08863604e+00 -6.28169859e-03 4.23393309e-01 3.78308386e-01
7.20385075e-01 -1.16173339e+00 6.47931516e-01 3.04711699e-01
5.80739498e-01 -9.94741142e-01 6.20268047e-01 5.01042008e-01
-6.80606008e-01 -3.59801918e-01 -5.12642920e-01 -6.86627924e-01
2.61320829e-01 6.86476767e-01 -1.78322881e-01 5.80248296e-01
1.70479462e-01 3.99092138e-01 3.48497517e-02 7.88671017e-01
-1.67326316e-01 6.39133573e-01 -3.05792540e-01 2.23603919e-01
5.55828452e-01 -7.79363871e-01 5.61097145e-01 1.06846273e+00
2.45873779e-01 4.64915127e-01 -2.18105037e-02 8.96945596e-01
-1.35311792e-02 1.24043651e-01 -5.49455464e-01 -1.49270967e-01
3.64161342e-01 1.02230918e+00 -5.99190891e-01 -1.63760558e-02
-5.94183683e-01 6.77121580e-01 2.88545549e-01 7.50147343e-01
-7.19481468e-01 -2.45961145e-01 7.45555639e-01 -2.60126114e-01
2.45502263e-01 -4.23123062e-01 -3.11151415e-01 -1.44554651e+00
-1.18266888e-01 -8.95543575e-01 5.75963199e-01 -2.30356395e-01
-1.00454140e+00 -2.11196944e-01 3.74509275e-01 -5.44250906e-01
-8.96972001e-01 -8.31049979e-01 -6.03452981e-01 1.06498992e+00
-1.43389463e+00 -6.83770776e-01 7.73734093e-01 7.32204199e-01
2.31402621e-01 -1.17619269e-01 4.80209678e-01 9.98601690e-02
-7.48823702e-01 3.77783060e-01 8.01244020e-01 -1.03572406e-01
2.30509326e-01 -1.70239556e+00 2.00651795e-01 1.13811827e+00
1.38914455e-02 5.50314069e-01 9.93348539e-01 -4.52448696e-01
-1.46264672e+00 -6.91475093e-01 7.83419132e-01 -3.00843358e-01
1.05872905e+00 -1.55506685e-01 -3.15628529e-01 1.32461941e+00
-2.93357354e-02 -2.03047812e-01 5.70157468e-01 5.27616978e-01
-3.91884856e-02 1.79197699e-01 -1.17655611e+00 5.35044491e-01
7.10713208e-01 -5.50536573e-01 -2.86932290e-01 4.16197389e-01
3.01525533e-01 -5.64220548e-01 -7.37003028e-01 2.83447206e-01
4.96012807e-01 -7.75919259e-01 8.79468799e-01 -1.13145351e+00
2.05444358e-02 2.27854654e-01 -1.50557637e-01 -1.26249003e+00
-4.47444618e-01 -1.36862588e+00 -6.11685365e-02 1.04306030e+00
3.69877338e-01 -8.32149088e-01 2.96377212e-01 1.02593732e+00
9.12747085e-02 -5.70002556e-01 -1.24921918e+00 -1.02761781e+00
6.23739600e-01 -8.56791854e-01 4.17135715e-01 9.82853413e-01
3.54924828e-01 2.26722565e-02 -9.35109258e-01 -3.30863297e-02
7.84540117e-01 6.71330169e-02 6.03751421e-01 -1.14679492e+00
-6.37740433e-01 -7.35513210e-01 -2.08262324e-01 -1.25234246e+00
5.53295195e-01 -7.69245803e-01 3.56390290e-02 -1.15569174e+00
2.32710376e-01 -3.58429134e-01 -3.21760893e-01 1.25077935e-02
-1.41914502e-01 7.15655908e-02 2.37656966e-01 -3.95343229e-02
-6.07226491e-01 2.61642277e-01 8.32585275e-01 1.34817719e-01
-2.57072598e-01 5.49638867e-01 -1.20691311e+00 8.56609046e-01
8.06987941e-01 -5.97600877e-01 -2.19352901e-01 -2.65544891e-01
5.46500087e-01 7.12327123e-01 2.41998151e-01 -1.05127044e-01
7.01163709e-02 -8.24645579e-01 -1.40999436e-01 -1.83839336e-01
1.22179627e-01 -2.88877904e-01 9.10212770e-02 8.31890106e-02
-4.01164770e-01 1.21804133e-01 2.37089545e-02 3.40334386e-01
-4.72290404e-02 -6.51627541e-01 7.20842063e-01 -2.18160868e-01
-3.21202904e-01 3.65692407e-01 -6.94352806e-01 5.45388579e-01
6.55115604e-01 2.24846467e-01 1.11058831e-01 -1.07584250e+00
-7.23801196e-01 4.78933126e-01 1.93346649e-01 -1.05086220e-02
1.75029173e-01 -1.16771412e+00 -8.00476074e-01 -2.50885069e-01
-4.62239146e-01 -4.00813997e-01 7.67999217e-02 1.31372929e+00
-1.89357623e-01 9.31892693e-01 3.01086694e-01 6.41925111e-02
-6.05963647e-01 7.09390700e-01 4.60900545e-01 -4.37434971e-01
3.00759543e-03 8.81764650e-01 2.74941266e-01 -5.56377888e-01
-2.02100322e-01 -3.12400788e-01 9.94942784e-02 -6.02919199e-02
3.39612842e-01 6.53615355e-01 -3.61606419e-01 -6.75999880e-01
-3.54227424e-01 2.80447125e-01 4.78554368e-01 -7.15553999e-01
1.30017924e+00 -5.25035024e-01 -2.28940085e-01 6.71647966e-01
1.17238867e+00 3.51332128e-01 -1.21249878e+00 -2.98684508e-01
3.57086122e-01 -2.87738979e-01 2.38331974e-01 -4.69127089e-01
-9.96932030e-01 5.10928452e-01 -1.35718480e-01 3.07476461e-01
1.05900538e+00 2.28924435e-02 4.48851027e-02 4.14921761e-01
2.20539674e-01 -1.17806160e+00 -6.11886024e-01 5.29510200e-01
6.74942672e-01 -1.38045692e+00 9.03498381e-03 -2.05715895e-01
-5.99190772e-01 8.99508297e-01 -2.30898052e-01 -2.36580297e-01
8.19725394e-01 3.92818421e-01 -1.71534076e-01 -5.26028425e-02
-5.35945952e-01 -4.98945713e-01 3.97324055e-01 7.07359791e-01
2.14546487e-01 3.39839607e-01 -8.65589261e-01 1.01845264e+00
-6.18450902e-02 -1.70278162e-01 5.76730788e-01 8.02716076e-01
-2.13921010e-01 -1.05654573e+00 -4.32921022e-01 6.12906396e-01
-9.49924290e-01 -2.49458477e-01 -2.78223336e-01 9.16317105e-01
-3.82356703e-01 1.19845009e+00 -8.43735114e-02 2.03153357e-01
5.43723069e-02 -4.63692918e-02 4.20089543e-01 -4.09935981e-01
-7.45974705e-02 5.93962789e-01 -5.20413443e-02 -4.96843815e-01
-6.41115904e-01 -1.32080722e+00 -6.79369628e-01 -6.57584012e-01
-6.30789876e-01 3.95747483e-01 4.08103794e-01 1.29250991e+00
-1.69956818e-01 1.07497275e-01 8.50884080e-01 -7.10559726e-01
-1.05378115e+00 -8.29260826e-01 -1.44348490e+00 -2.45550141e-01
5.42421579e-01 -6.84213281e-01 -6.39141679e-01 -4.54135239e-01] | [6.525765895843506, 4.097433090209961] |
36b4bc3b-bfab-4b78-a659-6e9fc740ab02 | differentially-private-distributed-data | 1910.12832 | null | https://arxiv.org/abs/1910.12832v2 | https://arxiv.org/pdf/1910.12832v2.pdf | Differentially Private Distributed Data Summarization under Covariate Shift | We envision AI marketplaces to be platforms where consumers, with very less data for a target task, can obtain a relevant model by accessing many private data sources with vast number of data samples. One of the key challenges is to construct a training dataset that matches a target task without compromising on privacy of the data sources. To this end, we consider the following distributed data summarizataion problem. Given K private source datasets denoted by $[D_i]_{i\in [K]}$ and a small target validation set $D_v$, which may involve a considerable covariate shift with respect to the sources, compute a summary dataset $D_s\subseteq \bigcup_{i\in [K]} D_i$ such that its statistical distance from the validation dataset $D_v$ is minimized. We use the popular Maximum Mean Discrepancy as the measure of statistical distance. The non-private problem has received considerable attention in prior art, for example in prototype selection (Kim et al., NIPS 2016). Our work is the first to obtain strong differential privacy guarantees while ensuring the quality guarantees of the non-private version. We study this problem in a Parsimonious Curator Privacy Model, where a trusted curator coordinates the summarization process while minimizing the amount of private information accessed. Our central result is a novel protocol that (a) ensures the curator accesses at most $O(K^{\frac{1}{3}}|D_s| + |D_v|)$ points (b) has formal privacy guarantees on the leakage of information between the data owners and (c) closely matches the best known non-private greedy algorithm. Our protocol uses two hash functions, one inspired by the Rahimi-Recht random features method and the second leverages state of the art differential privacy mechanisms. We introduce a novel "noiseless" differentially private auctioning protocol for winner notification and demonstrate the efficacy of our protocol using real-world datasets. | ['Venkata Sitaramagiridharganesh Ganapavarapu', 'Roman Vaculin', 'Karthikeyan Shanmugam', 'Kanthi Sarpatwar', 'Ashish Jagmohan'] | 2019-10-28 | differentially-private-distributed-data-1 | http://papers.nips.cc/paper/9589-differentially-private-distributed-data-summarization-under-covariate-shift | http://papers.nips.cc/paper/9589-differentially-private-distributed-data-summarization-under-covariate-shift.pdf | neurips-2019-12 | ['data-summarization'] | ['miscellaneous'] | [ 2.48018280e-01 4.49264646e-02 -3.05887163e-01 -4.92395818e-01
-1.46386516e+00 -1.17706263e+00 1.54851168e-01 4.25726950e-01
-4.53043312e-01 8.39914560e-01 -1.04187123e-01 -1.35544553e-01
-3.78545046e-01 -9.09028590e-01 -9.50733066e-01 -1.14281940e+00
-3.05921197e-01 4.66774344e-01 -7.11854696e-02 -1.24001674e-01
2.69524395e-01 9.34234262e-02 -1.40474772e+00 1.60602957e-01
7.41231263e-01 1.48517048e+00 -4.00599748e-01 3.72544169e-01
3.51892442e-01 1.71863779e-01 -6.90838516e-01 -8.29005718e-01
1.17753613e+00 -3.27609211e-01 -7.37905920e-01 -4.30694729e-01
1.62244469e-01 -2.75676399e-01 -1.68283373e-01 1.34835756e+00
6.30069852e-01 -1.95868298e-01 4.40784663e-01 -1.94673502e+00
-6.41142905e-01 1.08829284e+00 -8.95610034e-01 -1.38679385e-01
1.47084504e-01 2.11263910e-01 1.10261667e+00 -1.04016900e-01
7.68777251e-01 7.20778823e-01 4.90876764e-01 4.07302827e-01
-1.51528072e+00 -1.30995154e+00 -8.11035410e-02 -1.30578279e-01
-1.70782781e+00 -4.07523125e-01 7.32844114e-01 -1.43788695e-01
3.41689646e-01 9.07885075e-01 3.06518286e-01 8.16654861e-01
3.38471904e-02 8.40018570e-01 1.19259763e+00 4.81571630e-02
7.23991752e-01 6.56575978e-01 1.92538053e-01 1.97783872e-01
7.30270326e-01 8.89878199e-02 -8.87474597e-01 -1.18631113e+00
2.83621028e-02 1.15161262e-01 -4.10903305e-01 -6.44787014e-01
-9.40040171e-01 9.85644042e-01 1.18663751e-01 -1.76260710e-01
-3.35413694e-01 2.26531565e-01 5.10768354e-01 8.19667935e-01
2.90456682e-01 1.43229350e-01 -8.50477159e-01 1.66635111e-01
-8.00769210e-01 7.82445669e-01 9.74262595e-01 1.41028082e+00
8.50645065e-01 -5.80819845e-01 6.38369843e-02 9.75853205e-02
1.49113402e-01 6.41808987e-01 2.97965854e-01 -9.75581586e-01
6.71014190e-01 6.15897894e-01 2.95787007e-01 -9.08540428e-01
2.44050682e-01 -1.07446402e-01 -1.10134029e+00 1.00284800e-01
4.80874985e-01 -3.82055521e-01 -1.10295095e-01 2.13667655e+00
6.65135026e-01 -3.33308429e-01 4.30678517e-01 8.08986664e-01
3.28318387e-01 5.01535773e-01 -7.24819079e-02 -4.63403165e-01
1.30354142e+00 -1.52931154e-01 -2.60126144e-01 3.30566853e-01
6.90849245e-01 -3.31315905e-01 7.30470657e-01 5.11049449e-01
-1.13488460e+00 2.76956409e-01 -9.63652551e-01 -1.26369461e-01
-4.62279290e-01 -3.42449248e-01 7.70870328e-01 1.05051446e+00
-8.49749744e-01 4.48159933e-01 -3.40501875e-01 -7.54270107e-02
9.12656486e-01 8.50169003e-01 -6.95994496e-01 1.34892374e-01
-1.26622558e+00 -3.38449255e-02 1.93003207e-01 -5.10284483e-01
-8.39408100e-01 -9.12340999e-01 -5.34003794e-01 -6.87002987e-02
4.62785780e-01 -8.59099329e-01 8.58482242e-01 -9.20179427e-01
-1.02137756e+00 1.10886729e+00 3.23031366e-01 -8.79400432e-01
8.20769548e-01 3.57427835e-01 -1.13272248e-02 -7.76162818e-02
1.97186425e-01 2.17935905e-01 7.02753246e-01 -1.27326763e+00
-1.01673663e+00 -1.16264522e+00 -2.57767797e-01 8.01819488e-02
-2.27012157e-01 1.50703806e-02 9.64983329e-02 -5.57754755e-01
-9.75290760e-02 -9.30324256e-01 -3.28548014e-01 1.99879080e-01
-7.66051829e-01 -1.04191052e-02 9.69882369e-01 -3.86730701e-01
1.05863345e+00 -2.36049581e+00 -4.80932966e-02 7.44836271e-01
4.37761307e-01 -5.27756813e-04 2.51537234e-01 4.28334922e-01
4.20725435e-01 3.63112420e-01 -6.07461333e-01 -4.11147863e-01
3.74070972e-01 -1.03843845e-01 -5.85235834e-01 1.04870200e+00
-5.26995718e-01 6.47358477e-01 -5.03641844e-01 -1.44915342e-01
-4.16906297e-01 2.46713355e-01 -4.68652576e-01 1.13005355e-01
-2.02962890e-01 1.78452209e-01 -7.12401927e-01 5.65613389e-01
1.23407149e+00 -4.70886789e-02 1.89198837e-01 5.57578802e-02
6.08085878e-02 -9.09173191e-02 -1.73819780e+00 1.69867074e+00
3.32651258e-01 -1.42230183e-01 7.01159179e-01 -8.45187664e-01
8.44301999e-01 1.80136561e-01 6.32463217e-01 -3.57025594e-01
3.41944039e-01 4.44689631e-01 -3.97465646e-01 7.05055967e-02
3.82500768e-01 7.44446814e-02 -8.35690677e-01 9.50486779e-01
-3.06784749e-01 5.24285110e-03 -6.39536023e-01 3.31049293e-01
1.22771764e+00 -4.88725901e-01 4.51733708e-01 -4.92055565e-01
2.75521874e-01 -6.53370917e-02 8.79899204e-01 9.27220225e-01
-3.63100052e-01 4.45738494e-01 6.97610915e-01 -2.22023189e-01
-8.67297947e-01 -8.09553206e-01 3.83827873e-02 9.78269815e-01
4.47253197e-01 -2.64115632e-01 -7.83892810e-01 -8.07759166e-01
7.29270160e-01 5.39995253e-01 -8.45030844e-01 -2.05163926e-01
3.58084813e-02 -5.97968936e-01 7.96313643e-01 -3.44784651e-03
5.97100914e-01 -5.62733591e-01 -8.85246038e-01 -1.93071559e-01
6.16017953e-02 -3.95182759e-01 -8.80276799e-01 3.96961004e-01
-5.84184229e-01 -9.80887890e-01 -4.73177642e-01 -3.41801673e-01
6.71499074e-01 1.99131802e-01 4.47931916e-01 -3.32086056e-01
-2.61414468e-01 2.43953139e-01 -1.15343079e-01 -8.85463536e-01
-7.47406185e-02 1.50515139e-01 1.17627822e-01 3.29226315e-01
7.06164658e-01 -6.83028877e-01 -8.68549168e-01 1.77528530e-01
-1.12865353e+00 -4.55852926e-01 3.19246560e-01 4.66832340e-01
9.33749139e-01 4.96251620e-02 5.52012444e-01 -1.28816807e+00
6.62819266e-01 -7.84585714e-01 -8.70820045e-01 3.09718013e-01
-9.33516443e-01 3.18173356e-02 7.69944608e-01 -3.72415364e-01
-5.52402794e-01 2.77675837e-01 5.04111528e-01 -3.37158769e-01
2.67306715e-02 -1.03551596e-02 -6.61391854e-01 -1.75274119e-01
5.79424620e-01 4.04322684e-01 3.22043151e-01 -3.83069843e-01
4.81618404e-01 9.74592984e-01 4.21253234e-01 -6.30388498e-01
7.31974423e-01 7.09464848e-01 1.53186962e-01 -2.40320861e-01
-9.33744945e-03 -2.49648318e-01 -1.78872034e-01 6.89250708e-01
2.45954245e-01 -8.70880425e-01 -1.58506930e+00 5.41865349e-01
-7.01793969e-01 6.52014613e-02 -8.72797728e-01 8.95207375e-02
-6.15478635e-01 3.76672417e-01 -9.48860049e-02 -9.89657342e-01
-8.76590312e-01 -1.09580088e+00 9.97122407e-01 1.33084252e-01
-1.02174513e-01 -3.62863511e-01 -5.14749624e-02 3.10005039e-01
4.42055017e-01 5.62228620e-01 7.55836666e-01 -1.28630269e+00
-8.71563137e-01 -5.31003833e-01 3.88952345e-02 1.92803040e-01
-7.17865378e-02 -4.48272854e-01 -9.96876001e-01 -4.43983525e-01
3.57631564e-01 -3.82311106e-01 6.07497871e-01 2.82312870e-01
1.34742975e+00 -1.03584826e+00 -3.50615025e-01 8.25988054e-01
1.45788252e+00 1.53548613e-01 3.90599340e-01 1.19985573e-01
1.47877902e-01 5.13864100e-01 5.49322546e-01 1.13722742e+00
4.86956239e-01 4.54523206e-01 2.76173294e-01 2.05032095e-01
7.58561492e-01 -2.72871017e-01 2.15339810e-01 -9.18786079e-02
4.58029866e-01 -2.65415072e-01 -3.35550368e-01 6.64022326e-01
-1.96987689e+00 -7.91919649e-01 7.43136033e-02 2.80901241e+00
1.23502481e+00 -3.63616735e-01 4.12162155e-01 4.33674976e-02
5.49373806e-01 -4.96405624e-02 -9.16355252e-01 -4.78791028e-01
-3.89102817e-01 1.92728862e-01 1.31934488e+00 7.09247515e-02
-7.76150823e-01 3.92392576e-01 4.31082106e+00 8.75326991e-01
-8.77824187e-01 1.92864731e-01 1.07335806e+00 -4.78815913e-01
-6.71764851e-01 3.02821934e-01 -6.59872472e-01 8.57148170e-01
9.29922998e-01 -1.01727438e+00 5.12128472e-01 1.07148528e+00
-2.65211731e-01 -3.36603224e-01 -1.47375846e+00 1.13378096e+00
-1.63935587e-01 -1.30601656e+00 -1.74983442e-01 5.19933879e-01
5.80486774e-01 -1.96190417e-01 5.96720576e-01 -1.03335276e-01
7.37291038e-01 -8.37625682e-01 8.00894558e-01 2.10800573e-01
1.02136064e+00 -1.14578378e+00 4.26734000e-01 5.43110371e-01
-8.24538767e-01 -2.62861639e-01 -3.93414050e-01 4.66370553e-01
-1.09395407e-01 5.73459148e-01 -4.22615558e-01 7.77618170e-01
9.86503541e-01 -6.26612529e-02 -1.80891842e-01 6.59146070e-01
4.11642462e-01 3.17439288e-01 -7.99281836e-01 8.84691067e-03
-1.42333895e-01 -2.70890802e-01 6.66694701e-01 7.72680819e-01
3.27150047e-01 4.93932426e-01 1.45719439e-01 8.09826612e-01
-7.93944359e-01 3.97738755e-01 -8.86471093e-01 2.39730909e-01
9.53891933e-01 9.40605104e-01 -1.62266687e-01 -9.08154622e-02
1.00433350e-01 9.58605409e-01 -4.65311073e-02 6.25516176e-02
-5.10734439e-01 -5.26176870e-01 1.04553354e+00 5.65793999e-02
2.68719047e-01 2.80907422e-01 -5.07226646e-01 -8.99235129e-01
2.65950799e-01 -9.89433527e-01 9.30256248e-01 1.20710030e-01
-1.63011062e+00 8.26072693e-02 -8.61400142e-02 -1.05069232e+00
1.42317146e-01 1.82230026e-01 -2.76282698e-01 1.01850259e+00
-1.25265813e+00 -1.11828470e+00 2.48846084e-01 1.10580468e+00
-3.77086520e-01 -2.64019996e-01 9.59076166e-01 -7.51572195e-03
-2.45914608e-01 1.33822322e+00 5.85347593e-01 -1.53790936e-01
7.54472435e-01 -1.04977369e+00 1.59309492e-01 5.16858816e-01
-2.65747279e-01 7.69246340e-01 6.64148569e-01 -6.69283152e-01
-2.06553888e+00 -1.12930596e+00 7.80166566e-01 -3.22679371e-01
2.56947547e-01 -7.07055271e-01 -5.85097849e-01 7.99821734e-01
-6.00369088e-02 3.69822025e-01 1.14164495e+00 -4.25425917e-01
-6.15428507e-01 -7.04846501e-01 -2.31738734e+00 1.43477663e-01
8.64416957e-01 -4.83483583e-01 5.12740090e-02 2.40551844e-01
7.17253447e-01 -2.29696870e-01 -8.89182210e-01 4.71690372e-02
5.66771746e-01 -8.86765838e-01 6.39848113e-01 -5.06123483e-01
-2.86964923e-01 -2.72215396e-01 -6.26460552e-01 -8.06530654e-01
2.16140419e-01 -1.45400548e+00 -3.59505676e-02 1.56021428e+00
4.59206969e-01 -9.72701371e-01 1.08654511e+00 1.39665389e+00
8.33329380e-01 -4.09825981e-01 -1.47134030e+00 -5.78478217e-01
2.90532172e-01 -1.57545045e-01 1.11419761e+00 1.01309061e+00
2.90779937e-02 -3.59740078e-01 -4.96234387e-01 2.77337015e-01
1.25096428e+00 3.23136628e-01 1.15822077e+00 -1.26505029e+00
-2.19379038e-01 4.94881123e-02 -2.96941698e-01 -4.53928769e-01
-4.47452348e-03 -1.03740442e+00 -2.33056128e-01 -6.89699471e-01
5.01815259e-01 -9.70878959e-01 -3.86420280e-01 5.46717286e-01
2.08501294e-01 -1.74316555e-01 3.08670606e-02 2.30677292e-01
-5.58451235e-01 4.11277145e-01 4.98315871e-01 -7.67438784e-02
-3.84029001e-01 3.03646892e-01 -1.48336935e+00 2.34868247e-02
6.37298882e-01 -8.61750662e-01 -5.22201002e-01 -5.79741597e-03
2.70556897e-01 1.70003176e-01 3.85410756e-01 -3.53514284e-01
6.26028299e-01 -3.90423536e-01 2.17944104e-02 -4.36235905e-01
-6.09720014e-02 -1.22873902e+00 6.97463095e-01 3.19352388e-01
-7.11046457e-01 -4.75551859e-02 -2.77754277e-01 9.09296155e-01
2.47347593e-01 1.99254200e-01 7.61061490e-01 -1.17726721e-01
3.43081094e-02 7.11955190e-01 2.58572161e-01 5.99050112e-02
1.73188329e+00 -8.78208354e-02 -4.95422661e-01 -3.02944154e-01
-1.20392948e-01 5.03462911e-01 9.14814174e-01 -4.80787084e-03
3.21932673e-01 -9.90524948e-01 -7.68589318e-01 2.94822663e-01
2.33227640e-01 4.43073303e-01 2.87973911e-01 7.56043077e-01
1.33823946e-01 8.88558701e-02 -6.08577169e-02 -2.67885029e-01
-1.38780928e+00 8.03754032e-01 1.98431108e-02 3.12603381e-03
-3.58555555e-01 1.04457605e+00 -2.56717242e-02 -2.62517303e-01
4.18018430e-01 -1.26080319e-01 6.17399991e-01 1.31128505e-01
6.24456584e-01 5.52492321e-01 -7.45310122e-03 -5.66536844e-01
-4.44245994e-01 -7.33596692e-03 -3.64652544e-01 -2.93648273e-01
1.52404642e+00 -3.71609330e-01 -2.60558456e-01 2.66535790e-03
1.60875094e+00 1.74542665e-01 -1.00019240e+00 -3.82209808e-01
-2.45715261e-01 -8.86179686e-01 -2.10278794e-01 -6.45394385e-01
-1.29212344e+00 3.53903055e-01 8.59077692e-01 2.84147769e-01
1.26760316e+00 4.01087143e-02 1.08565474e+00 3.42185274e-02
9.61107135e-01 -1.10750866e+00 -7.76999354e-01 -4.12153155e-01
4.88368452e-01 -1.09927011e+00 1.59231871e-01 -7.45852888e-02
-9.90818143e-01 4.90416974e-01 9.30585042e-02 -1.46626951e-02
7.86529481e-01 4.79542017e-01 -2.21327588e-01 -1.61028504e-01
-8.14089835e-01 3.87997508e-01 -3.87210816e-01 6.08111382e-01
-4.75758642e-01 3.65423381e-01 -4.98867989e-01 1.32497358e+00
-2.50344425e-01 1.71903819e-01 3.88381153e-01 1.24384034e+00
-7.44654089e-02 -1.42890465e+00 -2.93326050e-01 6.11658633e-01
-9.37624454e-01 1.50688663e-01 -6.31624520e-01 4.26218212e-01
2.13651016e-01 7.76239634e-01 -2.82109827e-01 -3.17280799e-01
2.16102704e-01 -9.39134788e-03 -1.29510194e-01 -2.08005264e-01
-1.26571763e+00 -2.32658103e-01 -3.72417659e-01 -6.33157492e-01
-1.23871185e-01 -9.77427363e-01 -1.17683768e+00 -7.67506719e-01
-3.60002697e-01 4.62141037e-01 8.00678730e-01 2.81579971e-01
9.07777488e-01 -6.44993305e-01 1.20951200e+00 -5.24083860e-02
-1.15131128e+00 -1.93695471e-01 -1.33744419e+00 5.85475445e-01
4.46172655e-01 -4.59512360e-02 -6.01216257e-01 -3.25971171e-02] | [5.872924327850342, 6.698272705078125] |
d4fcb9db-b9e4-4a6a-9f1c-cd510368e80f | actor-director-critic-a-novel-deep | 2301.03887 | null | https://arxiv.org/abs/2301.03887v1 | https://arxiv.org/pdf/2301.03887v1.pdf | Actor-Director-Critic: A Novel Deep Reinforcement Learning Framework | In this paper, we propose actor-director-critic, a new framework for deep reinforcement learning. Compared with the actor-critic framework, the director role is added, and action classification and action evaluation are applied simultaneously to improve the decision-making performance of the agent. Firstly, the actions of the agent are divided into high quality actions and low quality actions according to the rewards returned from the environment. Then, the director network is trained to have the ability to discriminate high and low quality actions and guide the actor network to reduce the repetitive exploration of low quality actions in the early stage of training. In addition, we propose an improved double estimator method to better solve the problem of overestimation in the field of reinforcement learning. For the two critic networks used, we design two target critic networks for each critic network instead of one. In this way, the target value of each critic network can be calculated by taking the average of the outputs of the two target critic networks, which is more stable and accurate than using only one target critic network to obtain the target value. In order to verify the performance of the actor-director-critic framework and the improved double estimator method, we applied them to the TD3 algorithm to improve the TD3 algorithm. Then, we carried out experiments in multiple environments in MuJoCo and compared the experimental data before and after the algorithm improvement. The final experimental results show that the improved algorithm can achieve faster convergence speed and higher total return. | ['Yuanlin Zhang', 'Yonghong Song', 'Zongwei Liu'] | 2023-01-10 | null | null | null | null | ['action-classification'] | ['computer-vision'] | [-2.30556130e-01 -1.02153920e-01 -2.19987676e-01 -2.40936037e-02
-2.24808753e-01 -4.85805385e-02 3.32203478e-01 2.49857139e-02
-9.28997457e-01 8.72157693e-01 -5.43109290e-02 9.11689177e-02
-1.30533114e-01 -8.98378968e-01 -4.66588378e-01 -9.98233497e-01
9.46771502e-02 3.50845397e-01 4.92189676e-01 -7.38527626e-02
3.12129885e-01 3.12938750e-01 -1.35146654e+00 -1.22812323e-01
9.91810381e-01 1.04827893e+00 3.90249282e-01 3.84290069e-01
-1.87505372e-02 1.02914417e+00 -8.99928927e-01 3.24340671e-01
3.00493628e-01 -8.26665878e-01 -3.22633535e-01 8.69704783e-02
-5.32046080e-01 -6.44432724e-01 -8.93796831e-02 1.10871935e+00
7.13308215e-01 4.37873572e-01 3.06633770e-01 -1.06837404e+00
-1.29928872e-01 6.98717892e-01 -5.78455806e-01 1.61003470e-01
7.23886043e-02 3.66467953e-01 6.63577259e-01 -3.28917116e-01
2.25547835e-01 1.57328999e+00 8.88954997e-02 6.28526509e-01
-7.42803454e-01 -7.43346512e-01 6.82817280e-01 2.81380326e-01
-8.40826929e-01 -1.93570659e-01 7.48583138e-01 -6.65882379e-02
5.87227643e-01 -1.46545351e-01 1.05840755e+00 6.83622897e-01
2.29700908e-01 1.08493960e+00 9.91377413e-01 -3.00779909e-01
6.37308717e-01 2.08418950e-01 -2.44984508e-01 6.76711857e-01
1.84027925e-01 4.89706963e-01 -7.02062100e-02 1.21711627e-01
9.42579925e-01 1.31026059e-01 -5.71759790e-02 -3.49259108e-01
-1.07466805e+00 7.89012671e-01 6.26691937e-01 3.46640229e-01
-8.09176743e-01 3.16585839e-01 4.61888939e-01 3.01537216e-01
2.60378778e-01 4.59710419e-01 -2.42554441e-01 -2.44654074e-01
-4.07020092e-01 2.38300636e-01 4.13529843e-01 1.77164525e-01
5.92722058e-01 5.63758254e-01 -3.39839637e-01 8.86292934e-01
4.67005312e-01 5.65877378e-01 7.55020261e-01 -1.10860360e+00
4.55624133e-01 9.14696634e-01 2.78874338e-01 -9.70725775e-01
-2.35230118e-01 -5.98198295e-01 -6.41907156e-01 8.12739134e-01
3.16423178e-01 -3.80317181e-01 -6.32919908e-01 1.52427554e+00
5.36775470e-01 2.24507097e-02 2.63074338e-01 1.01846790e+00
2.39927143e-01 7.36186028e-01 -4.02977802e-02 -5.60624659e-01
7.99615145e-01 -1.19750655e+00 -8.64510596e-01 -1.57312870e-01
5.64106405e-01 -2.34061152e-01 8.32925618e-01 6.76758051e-01
-9.23488379e-01 -6.96725368e-01 -1.06185973e+00 8.31233323e-01
-4.86900285e-02 3.60926092e-01 4.55231130e-01 1.79309875e-01
-4.08344865e-01 7.69876003e-01 -8.72538984e-01 1.79849714e-01
3.22463185e-01 1.84504732e-01 1.87398165e-01 2.05981135e-01
-1.28014886e+00 9.53063369e-01 8.13348353e-01 1.67467341e-01
-1.25699854e+00 -3.72173544e-03 -4.63265240e-01 2.04725042e-01
6.94234133e-01 -9.38711911e-02 1.39422369e+00 -1.46744871e+00
-1.81424928e+00 -1.71745062e-01 2.38824487e-01 -4.03399467e-01
6.46466434e-01 -3.65754329e-02 -1.88132569e-01 -3.59005504e-03
-4.35939580e-02 3.80592048e-01 8.01983237e-01 -1.07920229e+00
-1.04617631e+00 -2.65898824e-01 3.05474013e-01 6.64582670e-01
-3.36198688e-01 -3.06250840e-01 -2.20943898e-01 -2.29537547e-01
-1.06834993e-01 -6.85208678e-01 -4.53276455e-01 -1.14360496e-01
1.99342564e-01 -3.93833399e-01 6.68230116e-01 -2.90209949e-01
1.33720076e+00 -2.20257640e+00 1.49567172e-01 3.20230693e-01
8.25631768e-02 6.01686656e-01 -2.54301876e-01 8.54100138e-02
1.57856479e-01 -2.36405671e-01 -1.73230488e-02 2.32546642e-01
-2.33151808e-01 3.73694807e-01 1.79598778e-01 2.16375500e-01
-3.26841623e-02 3.59964579e-01 -1.17974377e+00 -4.25448477e-01
2.97381252e-01 -2.20495462e-02 -5.00834346e-01 4.47715610e-01
-3.96024942e-01 4.57685292e-01 -1.00878930e+00 3.71522188e-01
4.51112598e-01 1.38350815e-01 2.49836639e-01 3.23343128e-01
-1.92713022e-01 5.40324226e-02 -1.66480458e+00 9.60354447e-01
-5.82947493e-01 9.65637937e-02 1.75570786e-01 -1.10781252e+00
1.30533481e+00 3.35997432e-01 5.40800631e-01 -1.10892928e+00
3.98804307e-01 2.56633610e-01 2.11838514e-01 -5.82936704e-01
6.32209107e-02 1.62663773e-01 2.87675530e-01 4.68170941e-01
-2.58472621e-01 9.45117474e-02 3.83824110e-01 -6.75410479e-02
8.94634843e-01 3.44184697e-01 2.65697181e-01 1.40349254e-01
7.65893936e-01 -1.13730304e-01 9.55252111e-01 6.24677896e-01
-3.11735719e-01 -3.28241259e-01 6.54463708e-01 -5.82995057e-01
-8.05905223e-01 -5.27257860e-01 3.70702326e-01 1.00857317e+00
3.96299154e-01 8.30004141e-02 -4.83782411e-01 -1.00779283e+00
9.51814950e-02 6.24006331e-01 -4.61340338e-01 -4.76542145e-01
-6.07899427e-01 -7.79473722e-01 2.71438301e-01 5.59797108e-01
9.44023609e-01 -1.43840671e+00 -9.21502888e-01 5.36318183e-01
2.00571463e-01 -3.20136517e-01 -1.44880310e-01 1.22421861e-01
-8.29829574e-01 -1.25749719e+00 -7.03332007e-01 -6.20980501e-01
7.16192424e-01 5.50788045e-02 4.24824625e-01 2.33305275e-01
3.49622279e-01 -2.77134702e-02 -4.49919194e-01 -3.76843244e-01
-5.13445497e-01 -2.18161866e-01 1.72858804e-01 -8.61514211e-02
-5.53167686e-02 -3.42398703e-01 -4.29251522e-01 3.69602352e-01
-7.19201863e-01 -1.72891110e-01 8.53332460e-01 1.01486135e+00
2.60491878e-01 2.52995253e-01 8.48671675e-01 -5.25879860e-01
1.02180600e+00 -2.60068417e-01 -1.04794705e+00 1.31392375e-01
-9.74814475e-01 4.50137496e-01 1.04025972e+00 -8.61460388e-01
-1.13320255e+00 -8.52557570e-02 -1.29190952e-01 -4.40687388e-01
2.05585331e-01 3.77067864e-01 -8.56779963e-02 1.45531237e-01
4.52652693e-01 2.70080835e-01 3.70439947e-01 -3.20078045e-01
-5.06058596e-02 6.10084593e-01 4.89408430e-03 -3.88510764e-01
3.78061414e-01 -1.17095739e-01 -1.09993555e-01 -4.61412296e-02
-5.77695251e-01 -1.82059221e-02 -9.41132233e-02 -7.26122618e-01
6.45163000e-01 -6.98134959e-01 -1.07839406e+00 5.71622133e-01
-1.05677295e+00 -4.26809847e-01 -4.27366376e-01 8.41922343e-01
-3.77090931e-01 2.17983529e-01 -3.72727156e-01 -1.18947101e+00
-2.50293344e-01 -1.31290507e+00 3.14597309e-01 7.24169970e-01
4.48805302e-01 -9.27667499e-01 2.74974883e-01 -2.15782657e-01
3.56348991e-01 7.08209351e-02 6.36775196e-01 -6.78195834e-01
-2.90800959e-01 -5.20560443e-02 2.30237350e-01 6.58900797e-01
-1.05369180e-01 6.85443655e-02 -5.61989188e-01 -3.78353626e-01
1.25221506e-01 -4.07231212e-01 7.55302787e-01 2.12859407e-01
8.87429774e-01 -4.02708501e-01 -1.80614710e-01 2.19957620e-01
1.33753896e+00 1.00629556e+00 4.73702878e-01 8.59222233e-01
2.06967086e-01 2.19378024e-01 1.14329851e+00 6.79424644e-01
-4.79941741e-02 5.29973149e-01 7.99746692e-01 -4.19535227e-02
1.92727536e-01 -2.19699442e-01 7.47814178e-01 6.08557105e-01
-8.56076553e-02 -9.54641029e-02 -3.62391114e-01 3.65206569e-01
-2.21490884e+00 -1.06561244e+00 1.35207221e-01 2.26050878e+00
6.39504433e-01 3.93894941e-01 2.70025671e-01 1.50039718e-01
7.51934052e-01 2.20886335e-01 -9.33255792e-01 -5.76871395e-01
2.61120677e-01 -3.15018594e-01 2.51301229e-01 3.98217648e-01
-6.71240449e-01 6.79584801e-01 5.83256388e+00 9.34359193e-01
-1.23206604e+00 -1.08217478e-01 4.69898045e-01 -8.56438056e-02
-1.82761922e-02 3.89914624e-02 -7.05709040e-01 6.97714448e-01
5.60580730e-01 -1.24882936e-01 8.62620473e-01 1.25924969e+00
4.60978299e-01 -6.07214570e-01 -7.15958476e-01 6.16693139e-01
-3.13020825e-01 -8.05521011e-01 -1.90844223e-01 -1.46798193e-01
6.40851259e-01 -2.61971384e-01 -2.89353937e-01 6.43028975e-01
5.17133534e-01 -4.74934489e-01 6.68858409e-01 5.73152423e-01
1.07879661e-01 -1.05765450e+00 1.07415259e+00 8.88778150e-01
-1.05162930e+00 -6.08881891e-01 -6.70822263e-01 -2.35206887e-01
-2.49604121e-01 4.40898985e-01 -7.11584687e-01 6.12987459e-01
4.17602181e-01 6.29125118e-01 -3.00406516e-01 1.06804359e+00
-5.21545708e-01 2.99418002e-01 -1.81065857e-01 -5.67433178e-01
4.76540744e-01 -3.36663485e-01 4.30864125e-01 5.12573302e-01
1.77692339e-01 -7.70280370e-03 4.16289151e-01 6.60967112e-01
2.62945920e-01 1.09417297e-01 -3.03624839e-01 6.48347661e-02
6.42528117e-01 1.25148046e+00 -4.55386609e-01 -6.46078587e-01
3.96642201e-02 5.29801428e-01 4.75154430e-01 2.40919903e-01
-9.33789194e-01 -5.22709250e-01 1.41296789e-01 -3.09459150e-01
2.05402404e-01 -3.06258146e-02 3.23383547e-02 -8.22290659e-01
-7.37385228e-02 -9.04008925e-01 2.15841964e-01 -5.61437607e-01
-8.54106486e-01 5.37752211e-01 -1.36583447e-01 -1.31535518e+00
-4.28306371e-01 -3.87316704e-01 -8.30918312e-01 7.31548905e-01
-1.36424327e+00 -2.69418240e-01 -1.99079156e-01 4.47687566e-01
5.66401720e-01 -4.64101791e-01 5.83414495e-01 1.15975775e-01
-7.67608106e-01 2.66352803e-01 4.60289776e-01 3.45101720e-03
4.05857593e-01 -1.02081501e+00 -4.18211371e-01 6.54088080e-01
-5.59066296e-01 4.18309011e-02 3.67629111e-01 -6.30312860e-01
-1.01558769e+00 -9.10514176e-01 2.08562419e-01 4.02370572e-01
3.31801534e-01 1.74911141e-01 -6.92297637e-01 2.13291198e-01
-8.65163188e-03 -2.80668139e-01 -9.57056805e-02 -1.33761317e-01
3.23004782e-01 -5.73830009e-01 -1.15976143e+00 5.04315615e-01
4.75776106e-01 2.20181569e-01 -4.94426697e-01 7.84994513e-02
5.43015301e-01 -3.24404448e-01 -6.14360869e-01 2.37018093e-01
4.88238692e-01 -1.01457000e+00 4.95680004e-01 -3.40093464e-01
2.31784403e-01 -4.22998726e-01 3.59319389e-01 -1.66922355e+00
-3.70647043e-01 -9.01164785e-02 -6.17692322e-02 1.00209498e+00
1.82211414e-01 -8.31649899e-01 6.11066282e-01 1.16707161e-01
-4.47057225e-02 -1.08110380e+00 -8.71653259e-01 -8.88841808e-01
-2.86824048e-01 6.86259344e-02 6.32352829e-01 5.30495524e-01
-1.00652121e-01 3.33639652e-01 -5.01757503e-01 -1.76311865e-01
3.32171619e-01 3.84250805e-02 7.14847505e-01 -9.46538389e-01
-4.79072005e-01 -4.54332262e-01 -5.76520786e-02 -1.13643956e+00
1.13865258e-02 -4.91171271e-01 2.49257818e-01 -1.66390288e+00
4.92146425e-02 -5.48352778e-01 -5.40550411e-01 6.42177880e-01
-3.98552924e-01 -4.41102624e-01 3.55094999e-01 2.31321290e-01
-8.24832261e-01 1.03619969e+00 1.74970770e+00 -1.45779952e-01
-5.56632876e-01 4.16344464e-01 -3.70548606e-01 7.65436649e-01
9.66864109e-01 -6.73951328e-01 -4.10322994e-01 -3.75651032e-01
-1.04371943e-01 3.59081388e-01 3.60868424e-02 -1.17798734e+00
5.80360442e-02 -4.97658819e-01 6.26469195e-01 -2.68780112e-01
1.25484005e-01 -9.64188337e-01 -2.31003657e-01 1.02696669e+00
-4.89683658e-01 1.11054063e-01 -6.57391697e-02 5.73306620e-01
-1.32627845e-01 -4.57273096e-01 9.19666409e-01 -3.85040611e-01
-5.63571036e-01 4.94940616e-02 -6.08890653e-01 -1.34534240e-01
1.27885973e+00 -3.90614122e-02 -1.95270851e-01 -4.12179708e-01
-5.97724736e-01 7.86207676e-01 1.23929873e-01 2.34556705e-01
7.77603447e-01 -1.44681692e+00 -4.53919590e-01 1.68798506e-01
-3.79487753e-01 -8.60942379e-02 -2.64589000e-03 7.76620567e-01
-2.69828945e-01 2.26475764e-02 -5.04285216e-01 -2.87466824e-01
-1.01702332e+00 5.42901158e-01 7.09163666e-01 -6.44746542e-01
-2.53814191e-01 3.35313559e-01 -1.18330963e-01 -4.29086655e-01
5.27565181e-01 -1.07321836e-01 -6.43501043e-01 -3.19024064e-02
4.72763985e-01 6.45050228e-01 -3.01220745e-01 7.01602176e-02
-2.32825398e-01 2.99605608e-01 -1.09967748e-02 -1.90427244e-01
1.37313902e+00 1.47901937e-01 -1.95642915e-02 3.71003360e-01
7.36560047e-01 -1.53655767e-01 -1.45071614e+00 -5.89345172e-02
-2.91150719e-01 -4.53195214e-01 1.21489793e-01 -9.29333448e-01
-1.15665483e+00 7.07114160e-01 8.28879118e-01 1.87866390e-01
1.33327162e+00 -5.59581220e-01 4.93799627e-01 4.58290577e-01
2.78004795e-01 -1.48500609e+00 4.79699284e-01 5.28695345e-01
6.71818435e-01 -9.29765642e-01 1.09243676e-01 4.43275928e-01
-9.87687707e-01 1.12822258e+00 1.16852105e+00 -3.46569777e-01
1.13843530e-01 -4.15693074e-02 3.00332084e-02 1.01364963e-01
-9.19504046e-01 -1.63945466e-01 -1.89781189e-01 2.52698511e-01
-1.17329709e-01 -1.02440149e-01 -6.87495112e-01 3.69339973e-01
5.28648973e-01 9.02878493e-02 4.90898281e-01 9.26529229e-01
-1.02733684e+00 -1.20826459e+00 -4.14102763e-01 3.01297575e-01
-1.00844249e-01 4.20753479e-01 -6.84683099e-02 7.06514001e-01
2.75572896e-01 9.38340366e-01 -2.08874810e-02 -4.76457864e-01
4.29460913e-01 -2.47241423e-01 2.30316341e-01 -3.08616489e-01
-6.54040873e-01 3.14552218e-01 -7.55172521e-02 -6.36510551e-01
-3.67051423e-01 -3.28286350e-01 -1.63677740e+00 -1.36999875e-01
-5.69977582e-01 6.48645043e-01 5.98629177e-01 8.17532659e-01
8.87215957e-02 9.63543117e-01 1.23322260e+00 -6.68248892e-01
-1.13392699e+00 -1.01799250e+00 -5.83570898e-01 8.01189989e-02
2.03723058e-01 -7.96335936e-01 -4.30200994e-01 -5.68497479e-01] | [4.036343574523926, 2.065160036087036] |
31555700-8920-4f63-9937-186e1f75211b | flightbert-a-non-autoregressive-multi-horizon | 2305.01658 | null | https://arxiv.org/abs/2305.01658v1 | https://arxiv.org/pdf/2305.01658v1.pdf | FlightBERT++: A Non-autoregressive Multi-Horizon Flight Trajectory Prediction Framework | Flight Trajectory Prediction (FTP) is an essential task in Air Traffic Control (ATC), which can assist air traffic controllers to manage airspace more safely and efficiently. Existing approaches generally perform multi-horizon FTP tasks in an autoregressive manner, which is prone to suffer from error accumulation and low-efficiency problems. In this paper, a novel framework, called FlightBERT++, is proposed to i) forecast multi-horizon flight trajectories directly in a non-autoregressive way, and ii) improved the limitation of the binary encoding (BE) representation in the FlightBERT framework. Specifically, the proposed framework is implemented by a generalized Encoder-Decoder architecture, in which the encoder learns the temporal-spatial patterns from historical observations and the decoder predicts the flight status for the future time steps. Compared to conventional architecture, an extra horizon-aware contexts generator (HACG) is dedicatedly designed to consider the prior horizon information that enables us to perform multi-horizon non-autoregressive prediction. Additionally, a differential prediction strategy is designed by well considering both the stationarity of the differential sequence and the high-bits errors of the BE representation. Moreover, the Bit-wise Weighted Binary Cross Entropy loss function is proposed to optimize the proposed framework that can further constrain the high-bits errors of the predictions. Finally, the proposed framework is validated on a real-world flight trajectory dataset. The experimental results show that the proposed framework outperformed the competitive baselines. | ['Yi Lin', 'Jianwei Zhang', 'Zheng Zhang', 'Dongyue Guo'] | 2023-05-02 | null | null | null | null | ['trajectory-prediction'] | ['computer-vision'] | [ 4.62187111e-01 -2.56379485e-01 -3.42727631e-01 -2.44433627e-01
-4.74181622e-01 -3.55012864e-01 4.09060925e-01 -9.82753113e-02
-2.02145785e-01 6.12036109e-01 2.31066197e-01 -4.77873355e-01
-3.50632221e-01 -8.89626324e-01 -5.55458546e-01 -6.19007111e-01
-2.25938261e-01 -2.79128477e-02 2.23432377e-01 -2.24868819e-01
-1.76451844e-03 2.13839754e-01 -1.59604836e+00 2.99845129e-01
9.18655932e-01 1.58504486e+00 1.88722387e-01 5.22635102e-01
3.45388204e-01 8.78141642e-01 -3.39156061e-01 -3.11701298e-01
5.82132399e-01 -4.38427418e-01 -2.02526122e-01 -8.12910721e-02
-1.51039377e-01 -4.83033210e-01 -6.27095938e-01 8.54974687e-01
3.15339804e-01 5.66911757e-01 4.28503275e-01 -1.08554196e+00
-1.27448095e-02 1.53529286e-01 4.25646268e-03 3.46510738e-01
-1.69149805e-02 4.80196595e-01 8.51652205e-01 -5.36488831e-01
9.94880199e-02 1.08025837e+00 6.36309564e-01 1.33885309e-01
-5.79227030e-01 -7.59275913e-01 4.57135260e-01 3.62438023e-01
-1.52986443e+00 -1.75333396e-01 6.25785410e-01 -7.10497618e-01
8.79056573e-01 4.24662948e-01 8.08904350e-01 8.09018970e-01
7.02329338e-01 5.87137818e-01 7.49364197e-01 -8.70676991e-03
1.69711068e-01 2.55729482e-02 -2.38326818e-01 6.43630028e-01
-2.64005810e-01 8.69052172e-01 -1.79912880e-01 1.01561241e-01
4.46565688e-01 3.70607942e-01 -4.48585242e-01 -5.87653881e-03
-1.11566889e+00 7.33233094e-01 4.89514679e-01 -6.83246180e-02
-6.41698658e-01 -5.50038666e-02 4.86092538e-01 8.95055160e-02
4.33878124e-01 1.73208326e-01 -3.19501817e-01 -3.84956270e-01
-1.01650560e+00 3.38962346e-01 5.05638719e-01 1.27094948e+00
3.36087793e-01 3.59539419e-01 -9.24219370e-01 3.03872138e-01
2.41004795e-01 3.23246181e-01 7.39949167e-01 -2.71391809e-01
8.27174902e-01 3.78922850e-01 1.35012999e-01 -1.00847423e+00
-2.47768655e-01 -7.63608217e-01 -9.57419574e-01 -7.90799707e-02
-3.00614983e-01 -4.87572193e-01 -7.39829361e-01 1.41962588e+00
2.57178724e-01 5.18642843e-01 -9.45663371e-04 9.99350190e-01
1.33695707e-01 1.02139950e+00 -1.31880954e-01 -4.18186903e-01
1.13224328e+00 -1.22751915e+00 -9.96272206e-01 -1.35892257e-01
4.55710799e-01 -6.53866410e-01 6.46248519e-01 2.34133556e-01
-8.25550079e-01 -8.83521855e-01 -1.26967347e+00 2.96380699e-01
-2.24254325e-01 6.52285159e-01 2.82217771e-01 6.77978098e-01
-4.57819402e-01 5.70420802e-01 -7.40319073e-01 3.62981021e-01
2.27399915e-01 2.07954362e-01 2.73800522e-01 3.94367985e-02
-1.61552978e+00 5.51308334e-01 1.01829708e+00 4.15018559e-01
-9.68848169e-01 -6.86922371e-01 -9.73268449e-01 2.95541972e-01
6.13393545e-01 -5.10800540e-01 1.18657541e+00 -5.97372472e-01
-1.67188334e+00 -1.51326686e-01 -1.50162473e-01 -9.92670894e-01
4.70026910e-01 -2.88104683e-01 -1.02270138e+00 -1.32704392e-01
-4.81754422e-01 3.71599466e-01 1.03288186e+00 -3.64015281e-01
-1.00476766e+00 -6.17779829e-02 2.58030128e-02 3.78525138e-01
-2.02735856e-01 -2.91284412e-01 -2.61807114e-01 -1.12463009e+00
-3.43484312e-01 -1.07780302e+00 -4.13863301e-01 -3.84868175e-01
-2.63296694e-01 2.64037609e-01 7.98718512e-01 -8.89612317e-01
2.11880302e+00 -2.28304338e+00 8.36250633e-02 1.81491300e-01
-3.31688702e-01 6.30163491e-01 2.03566447e-01 6.28634989e-01
-1.24396086e-01 -2.43455589e-01 -2.98955202e-01 -1.48358373e-02
5.16655445e-02 -3.01111722e-03 -9.54158843e-01 1.29228234e-01
2.09255069e-01 6.72340214e-01 -7.91218400e-01 -1.55039623e-01
5.05981326e-01 3.49066645e-01 -6.07017994e-01 4.82298106e-01
-4.36882496e-01 5.19976437e-01 -3.19964200e-01 4.40190434e-01
5.41920602e-01 5.90200089e-02 1.59544945e-02 -7.46013597e-02
-3.95662367e-01 4.51473325e-01 -8.86052966e-01 1.36157644e+00
-6.54436052e-01 3.49358916e-01 -4.50672060e-01 -5.19105315e-01
1.04941416e+00 5.19126415e-01 3.19850236e-01 -4.21432376e-01
1.30702421e-01 1.44291753e-02 -4.08409489e-03 -2.53642172e-01
6.79683030e-01 -3.34252506e-01 8.03956017e-02 -2.42724001e-01
-2.73347467e-01 -1.52320834e-02 -2.14962140e-02 -3.73927087e-01
7.77544916e-01 1.95261894e-03 3.81935894e-01 2.59274542e-01
9.99473929e-01 -1.19401060e-01 1.01628232e+00 1.32022157e-01
-7.46866837e-02 3.11821699e-01 2.75517136e-01 -4.62907642e-01
-6.62533283e-01 -7.12163925e-01 -8.00531581e-02 3.55665624e-01
1.24156147e-01 -6.93242192e-01 -4.19504434e-01 -7.17530370e-01
-1.98325872e-01 1.15706813e+00 -2.37995118e-01 -5.46824455e-01
-2.80659318e-01 -5.65407693e-01 3.65783781e-01 4.29882020e-01
7.80999720e-01 -6.81488991e-01 -7.29363918e-01 4.26958263e-01
-8.85426924e-02 -1.22039747e+00 -7.18757093e-01 -1.37422428e-01
-6.78605855e-01 -7.58251905e-01 -5.68793952e-01 -3.50817919e-01
3.11265320e-01 9.90148485e-02 3.59587848e-01 -2.69240141e-01
-1.27881140e-01 4.32607867e-02 -4.82348174e-01 -3.57367486e-01
-2.01968774e-01 -7.71123124e-03 -1.29749009e-03 4.43846256e-01
1.36818588e-01 -2.94929177e-01 -6.28339171e-01 4.05942857e-01
-1.00679362e+00 1.78701207e-01 7.17171907e-01 1.06321669e+00
8.74272585e-01 7.03332007e-01 5.26799917e-01 -5.85694790e-01
5.48907101e-01 -5.86479783e-01 -1.00114512e+00 3.14191759e-01
-8.64583850e-01 -1.55868068e-01 1.13701725e+00 -1.69624895e-01
-1.07807887e+00 1.75808016e-02 -4.37939018e-01 -7.05966353e-01
1.27158687e-01 5.51468730e-01 -1.63302228e-01 -6.19954895e-03
5.63627621e-03 6.12245321e-01 -3.15333813e-01 -2.91047841e-01
8.70570093e-02 6.69115126e-01 3.09606224e-01 -7.90458769e-02
7.69530833e-01 4.75528724e-02 2.64913648e-01 -4.64806646e-01
-7.29918718e-01 -3.77547592e-01 -5.47943115e-01 -2.77919471e-01
7.04176486e-01 -1.22535908e+00 -6.13623798e-01 1.72180265e-01
-9.92125630e-01 -2.00592354e-02 -3.31819862e-01 9.59995210e-01
-7.54008412e-01 4.32122767e-01 -4.07497644e-01 -1.12819612e+00
-2.90331900e-01 -1.24181235e+00 9.23585415e-01 -9.15835276e-02
1.30483791e-01 -8.60751688e-01 -5.85346818e-02 1.10682964e-01
3.28285575e-01 2.92236656e-01 1.00907564e+00 -5.36156893e-01
-8.26439500e-01 -4.33684409e-01 8.11330304e-02 6.54108822e-01
4.42075655e-02 -4.24601197e-01 -7.36639857e-01 -4.61731672e-01
1.74355507e-01 1.25517845e-01 8.49984169e-01 6.24197759e-02
1.59019506e+00 -7.13934422e-01 -2.58699745e-01 7.77221084e-01
1.28994298e+00 7.17107117e-01 6.89775944e-01 -1.06044911e-01
3.25588971e-01 4.59165603e-01 1.36892867e+00 9.46738005e-01
1.95523724e-01 9.24834788e-01 4.81081635e-01 3.84390295e-01
1.45456910e-01 -6.81786060e-01 6.65111542e-01 1.07368374e+00
1.19702332e-01 -5.64102888e-01 -5.36096156e-01 3.09889048e-01
-1.81898093e+00 -9.66626227e-01 1.86181411e-01 2.58143210e+00
5.27537525e-01 2.08230704e-01 -1.99292243e-01 1.23015217e-01
4.25416321e-01 5.32613039e-01 -5.67676663e-01 -3.13182920e-01
2.01688573e-01 6.58083707e-02 6.70117497e-01 2.59004563e-01
-1.31897926e+00 6.23763740e-01 5.38583946e+00 1.31284010e+00
-1.03700125e+00 -5.69690131e-02 4.65471447e-01 -2.41867378e-01
8.50765500e-04 -1.14725813e-01 -1.15921104e+00 9.92210746e-01
1.31360972e+00 -4.31411088e-01 6.21449351e-01 7.90465832e-01
2.00117603e-01 3.95870745e-01 -7.99614847e-01 9.44087207e-01
-7.05692247e-02 -1.10419548e+00 1.02190748e-01 5.05651440e-03
5.39125383e-01 -3.70848000e-01 4.59231436e-01 6.81922495e-01
-1.29256889e-01 -8.39558065e-01 6.49062455e-01 8.30295980e-01
8.85483801e-01 -1.05106521e+00 8.09864461e-01 6.43473208e-01
-1.66261590e+00 -6.80360913e-01 -4.55089450e-01 -3.63326184e-02
6.15141869e-01 6.08226240e-01 -1.07184875e+00 1.12671852e+00
2.76812792e-01 9.04896021e-01 -2.98332781e-01 1.20223963e+00
-1.75689593e-01 6.60314679e-01 -3.43022235e-02 1.68200359e-01
5.18021405e-01 -3.46466690e-01 8.02948475e-01 9.84580398e-01
8.43095422e-01 2.70110995e-01 3.62395346e-01 6.14593148e-01
3.29629689e-01 1.08422395e-02 -6.60293996e-01 -3.46928418e-01
4.15109485e-01 8.95703554e-01 -8.02044868e-02 -1.48661017e-01
-3.49790961e-01 8.53307545e-01 -1.96086749e-01 3.06906104e-01
-1.20220363e+00 -4.66796607e-01 7.16349721e-01 6.65971339e-02
6.73690677e-01 -2.94547170e-01 2.99172670e-01 -9.50940192e-01
2.00358108e-01 -7.23414719e-01 4.17498648e-01 -6.49870694e-01
-8.38835835e-01 8.47748816e-01 -5.93979098e-02 -2.04497886e+00
-6.04160845e-01 -3.74822319e-01 -3.84607136e-01 1.00307584e+00
-1.86620891e+00 -9.04921055e-01 -5.77068403e-02 4.22336310e-01
7.15998232e-01 -3.49299431e-01 5.61651170e-01 4.62117821e-01
-8.56937408e-01 5.88671029e-01 7.59786218e-02 -3.43736917e-01
3.60817671e-01 -7.09666014e-01 3.50075662e-01 9.85123932e-01
-2.94324905e-01 3.49108130e-01 3.36618841e-01 -8.82319629e-01
-1.19698834e+00 -1.87903917e+00 7.33090222e-01 1.35885850e-01
4.28427726e-01 -4.85633351e-02 -7.38541007e-01 7.23857045e-01
-4.56342064e-02 -2.73121923e-01 6.74060047e-01 -4.50448394e-01
2.34191269e-01 -2.08435416e-01 -8.15878034e-01 5.61117291e-01
7.56524444e-01 -3.42228442e-01 -3.11767817e-01 5.79422534e-01
1.17332745e+00 -7.54305363e-01 -8.87275517e-01 7.60812342e-01
2.63867319e-01 -8.78265202e-01 9.66441751e-01 -6.52982593e-01
3.49894881e-01 -5.68476796e-01 -1.67626023e-01 -1.30587327e+00
-3.68732005e-01 -7.46676385e-01 -4.37757313e-01 1.03004742e+00
1.43723652e-01 -6.07192695e-01 3.42072487e-01 3.95740241e-01
-5.63805401e-01 -1.20151782e+00 -1.03910673e+00 -1.08514905e+00
-4.86142218e-01 -4.57416862e-01 1.07580280e+00 3.24863404e-01
-1.96037412e-01 5.41411750e-02 -8.54939401e-01 5.13570905e-01
-3.89693305e-02 1.13219284e-01 3.98528159e-01 -7.53765285e-01
-3.69951665e-01 -1.32916644e-01 -4.97567892e-01 -1.59051263e+00
2.01335236e-01 -9.20911491e-01 1.51482776e-01 -8.44073832e-01
-5.96157968e-01 -2.27066278e-01 -4.62461680e-01 -3.93060818e-02
-9.25975963e-02 -5.70342839e-01 1.52071118e-01 -6.98296577e-02
-4.03628856e-01 1.31896245e+00 1.10631192e+00 8.23000539e-03
-1.73406631e-01 6.19530320e-01 -5.44451028e-02 4.94457155e-01
6.59100056e-01 -2.10885122e-01 -8.40284526e-01 1.53658697e-02
-4.88008782e-02 6.02562547e-01 2.99743980e-01 -1.48380101e+00
3.72222185e-01 -1.20944999e-01 2.33170182e-01 -1.08097398e+00
7.87432194e-01 -1.26404178e+00 2.72964984e-01 5.46258926e-01
-4.18256432e-01 3.46196741e-01 3.31180483e-01 1.15302432e+00
-6.28611326e-01 -1.28730819e-01 4.64413434e-01 2.15257794e-01
-7.25726187e-01 7.28504598e-01 -4.45289344e-01 -2.11359397e-01
1.38802004e+00 9.34900790e-02 7.54786981e-03 -2.59912282e-01
-3.49438816e-01 5.46979427e-01 8.14892501e-02 5.53825855e-01
8.28178823e-01 -1.50385559e+00 -3.15228820e-01 6.21963203e-01
1.29716292e-01 -1.26067698e-01 8.19571376e-01 8.99378061e-01
-2.76262671e-01 1.15946043e+00 4.14164066e-02 -1.95749745e-01
-9.08218086e-01 9.82848704e-01 2.64837027e-01 -6.28354311e-01
-5.93305230e-01 4.86941010e-01 2.15208694e-01 -1.85142934e-01
1.68183625e-01 -6.17852688e-01 -3.38717401e-01 -1.34168983e-01
7.34603345e-01 4.25193369e-01 2.15279460e-01 -6.25192404e-01
-9.54014622e-03 2.47817114e-01 8.30746721e-03 -7.03498721e-02
6.52933538e-01 -1.64092913e-01 3.02755594e-01 4.34448719e-01
9.53533173e-01 -2.71686584e-01 -1.42283499e+00 -1.71821088e-01
-2.77498156e-01 -6.71401262e-01 2.63233572e-01 -7.32032835e-01
-9.60038364e-01 1.21571743e+00 5.86454272e-01 -2.21726313e-01
1.55951369e+00 -1.13914609e+00 1.40310752e+00 1.44299984e-01
4.22893912e-01 -8.09153438e-01 -4.89995360e-01 5.89543402e-01
7.57438481e-01 -7.83667922e-01 -1.68832943e-01 -5.20578444e-01
-1.00432897e+00 1.18097663e+00 4.42921788e-01 2.33371586e-01
7.39089787e-01 -2.14260414e-01 -5.38256764e-01 2.13370323e-01
-1.26833034e+00 -2.12732822e-01 7.26934195e-01 3.25784504e-01
1.28159538e-01 2.17621908e-01 -3.57423276e-01 1.10982430e+00
-2.92093515e-01 3.95941913e-01 1.48063794e-01 6.58161402e-01
-2.93186992e-01 -8.45324159e-01 -1.76436588e-01 4.81113553e-01
-1.33796677e-01 -2.62324184e-01 1.38446972e-01 4.35951501e-01
2.49851465e-01 9.30329382e-01 7.78729841e-02 -9.70830083e-01
4.26402718e-01 1.42182559e-01 -1.39915362e-01 -3.84491861e-01
-7.28255510e-01 -4.47613634e-02 8.27709213e-03 -5.84317148e-01
1.14090845e-01 -5.39411485e-01 -9.63055074e-01 -3.24935503e-02
-1.30358800e-01 2.50369281e-01 3.75000983e-01 7.11706102e-01
6.40343070e-01 1.07088172e+00 1.03874671e+00 -3.52718323e-01
-8.69624674e-01 -6.78954363e-01 -6.29045784e-01 -1.82381898e-01
4.33935016e-01 -7.39578485e-01 -2.09405035e-01 -1.08513474e-01] | [6.935773849487305, 2.7537682056427] |
20549a74-dd86-471c-ac9b-a38ff5c187c0 | policy-learning-for-active-target-tracking | 2212.01498 | null | https://arxiv.org/abs/2212.01498v2 | https://arxiv.org/pdf/2212.01498v2.pdf | Policy Learning for Active Target Tracking over Continuous SE(3) Trajectories | This paper proposes a novel model-based policy gradient algorithm for tracking dynamic targets using a mobile robot, equipped with an onboard sensor with limited field of view. The task is to obtain a continuous control policy for the mobile robot to collect sensor measurements that reduce uncertainty in the target states, measured by the target distribution entropy. We design a neural network control policy with the robot $SE(3)$ pose and the mean vector and information matrix of the joint target distribution as inputs and attention layers to handle variable numbers of targets. We also derive the gradient of the target entropy with respect to the network parameters explicitly, allowing efficient model-based policy gradient optimization. | ['Nikolay Atanasov', 'Arash Asgharivaskasi', 'Shumon Koga', 'Pengzhi Yang'] | 2022-12-03 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-2.37404592e-02 4.87139374e-01 -6.33888006e-01 -2.33788729e-01
-5.48060834e-01 -3.36814702e-01 3.26755941e-01 -2.94704467e-01
-9.79909778e-01 7.56266892e-01 -9.99903604e-02 -3.41504246e-01
-3.39637816e-01 -4.53603059e-01 -9.73106325e-01 -7.81476617e-01
-9.75607932e-02 4.45355505e-01 -1.28406703e-01 -5.23744933e-02
2.54489839e-01 5.06358922e-01 -1.07280421e+00 -6.86955512e-01
5.89729965e-01 1.38473856e+00 7.02049077e-01 6.89063549e-01
5.95424652e-01 6.72990203e-01 -3.91046315e-01 2.50932336e-01
6.04561985e-01 -9.33310315e-02 -3.11315328e-01 1.61425505e-05
2.39637345e-02 -2.00029850e-01 -4.57434714e-01 1.72184551e+00
4.64253962e-01 5.40846050e-01 8.89822483e-01 -9.04335797e-01
-2.94849247e-01 4.21732008e-01 -1.87560841e-01 2.73065776e-01
-2.66902655e-01 4.06537175e-01 6.45522594e-01 -4.23880845e-01
5.28936028e-01 1.53948307e+00 4.78012532e-01 8.29740047e-01
-7.66106844e-01 -2.63754368e-01 5.51231205e-01 3.35883141e-01
-1.16558671e+00 -2.92863131e-01 5.37332475e-01 -4.37216014e-01
1.08454514e+00 -1.42705470e-01 6.21299565e-01 1.00277388e+00
6.21584356e-01 5.35081089e-01 5.53850114e-01 -1.17307059e-01
6.99834287e-01 1.73407048e-01 3.35248001e-02 8.23185503e-01
4.65336651e-01 7.30969667e-01 -4.21771705e-02 3.06663755e-02
5.61048925e-01 5.47942631e-02 -2.51619339e-01 -8.44928145e-01
-8.69572639e-01 1.05412209e+00 9.02608514e-01 -6.17949478e-02
-8.57259631e-01 5.69572687e-01 1.14202127e-01 4.81471002e-01
1.62202507e-01 8.39238346e-01 -6.60110414e-01 -1.01164363e-01
4.55795228e-02 1.11918353e-01 8.55491042e-01 1.09355879e+00
6.20792329e-01 5.65494835e-01 -4.91718240e-02 3.08997661e-01
7.39455998e-01 1.39440167e+00 5.58966160e-01 -1.22180688e+00
7.21425593e-01 4.41201597e-01 8.46822977e-01 -9.41542625e-01
-6.88891411e-01 -4.53178316e-01 -6.54338419e-01 6.33942246e-01
2.16620520e-01 -8.55614543e-01 -9.17275071e-01 1.98558104e+00
3.95239800e-01 -3.25175852e-01 4.55935240e-01 1.03381789e+00
-2.59891391e-01 6.24568999e-01 -1.92261979e-01 -4.68452573e-01
6.30696177e-01 -7.74889767e-01 -7.44132817e-01 -8.13193679e-01
5.25476456e-01 1.38846431e-02 9.63046908e-01 5.84302433e-02
-8.54971170e-01 -1.46511331e-01 -1.23510563e+00 5.89670599e-01
-4.25748825e-01 1.76897556e-01 -4.07110639e-02 2.83802480e-01
-9.80378091e-01 6.32053912e-01 -1.24834979e+00 -2.96603769e-01
2.10452408e-01 7.26347268e-01 8.81682485e-02 1.85080230e-01
-1.00404572e+00 1.85399246e+00 9.76328433e-01 2.99808472e-01
-1.29785466e+00 -2.34613240e-01 -1.02815640e+00 2.73521934e-02
6.67975068e-01 -5.48815131e-01 1.39915895e+00 -7.21044779e-01
-2.17295480e+00 -3.25711742e-02 1.25880167e-01 -5.93484223e-01
3.13679576e-01 -4.55649942e-01 2.17569217e-01 -1.38344839e-01
-1.21458627e-01 6.19524181e-01 1.08494961e+00 -9.64353383e-01
-8.84751141e-01 -7.13005722e-01 -2.77647555e-01 8.30943763e-01
-3.54174495e-01 -5.79384565e-01 1.94788035e-02 2.07972527e-01
-1.81971684e-01 -1.04274595e+00 -6.40170336e-01 -1.55403569e-01
-3.00202221e-01 -4.50778231e-02 9.26388860e-01 -6.31933331e-01
7.00165451e-01 -1.91954398e+00 5.57845473e-01 2.61591464e-01
-7.44965598e-02 1.30098268e-01 -3.43926728e-01 -2.37883300e-01
3.75681877e-01 -4.81395006e-01 -2.54325569e-02 -8.12470913e-02
3.32035050e-02 2.30291560e-01 -2.40539134e-01 7.68327415e-01
-7.52543584e-02 8.24141860e-01 -1.05094969e+00 7.02020824e-02
2.41999403e-01 3.13386694e-02 -2.86522061e-01 2.56572276e-01
-5.35747528e-01 3.99495304e-01 -8.58219087e-01 2.34357029e-01
2.73838520e-01 -1.04366802e-01 1.44705400e-01 2.01778457e-01
-1.56942904e-01 -9.86067019e-03 -8.98407578e-01 1.11274290e+00
-4.11186337e-01 4.42681611e-01 7.75970697e-01 -1.01185215e+00
1.10793352e+00 -1.15741104e-01 4.53196585e-01 -4.60947722e-01
6.77281678e-01 5.29249907e-02 -3.86756100e-02 -2.30987355e-01
4.36779112e-01 -3.82739166e-03 -3.05243582e-01 -4.29896079e-02
-9.51633751e-02 -2.81864375e-01 -2.22481340e-01 -3.27308506e-01
9.69927788e-01 -2.81579196e-01 3.76196653e-01 -2.93417096e-01
2.28222489e-01 1.37093022e-01 3.55797738e-01 1.02070093e+00
-2.07230374e-01 -3.02038193e-01 2.14427635e-01 -1.67876080e-01
-1.01378989e+00 -8.32485259e-01 4.13635790e-01 8.71171653e-01
2.86389291e-01 3.78605872e-01 -5.42759240e-01 -7.08971620e-01
2.59544522e-01 9.21218276e-01 -6.24031723e-01 -7.57233143e-01
-6.77318335e-01 -5.50017118e-01 -1.07926726e-01 4.45933014e-01
5.04055679e-01 -9.77303863e-01 -1.24957490e+00 1.74015313e-01
1.61535248e-01 -6.89053655e-01 -5.62814355e-01 6.16626024e-01
-8.61479104e-01 -1.27041352e+00 -3.31967562e-01 -5.74111342e-01
7.44954824e-01 -2.21885830e-01 3.11040610e-01 -8.03210020e-01
2.59177566e-01 8.28511417e-01 2.50865996e-01 -7.25228310e-01
-2.97143221e-01 3.09348762e-01 4.86880869e-01 -2.58234352e-01
-5.25534945e-03 -4.63028289e-02 -2.40182579e-01 1.18508361e-01
-3.14712226e-01 -4.54763949e-01 7.00761259e-01 1.07855129e+00
4.48413640e-01 -9.16085765e-02 2.63179302e-01 -9.69862118e-02
1.09067380e+00 -4.94758606e-01 -1.63646638e+00 1.18448831e-01
-9.32902634e-01 3.69837105e-01 6.42343700e-01 -8.61625016e-01
-8.93988729e-01 4.01651382e-01 3.40157032e-01 -7.63088703e-01
4.00974810e-01 5.26846886e-01 -2.51182735e-01 -9.07370597e-02
7.00845182e-01 1.65164322e-01 6.17508411e-01 -2.19463646e-01
4.83171731e-01 4.42368895e-01 4.02915776e-01 -1.95512846e-01
3.79332513e-01 1.27578661e-01 2.75315911e-01 -5.56701124e-01
-8.08539927e-01 -3.37362438e-01 -5.26234448e-01 -2.65164018e-01
7.44530559e-01 -7.85058022e-01 -1.26674521e+00 3.62270296e-01
-1.09518945e+00 -8.36102486e-01 -6.29742265e-01 8.98053527e-01
-1.18705904e+00 -1.24000475e-01 -2.90337890e-01 -1.23796570e+00
-4.71055776e-01 -9.60739553e-01 7.27090418e-01 3.07303578e-01
2.24345312e-01 -9.44910944e-01 2.80199777e-02 -7.85161555e-01
5.51944077e-01 8.90127346e-02 5.88166654e-01 -5.10728478e-01
-4.52863425e-01 -3.86987150e-01 5.25382273e-02 4.30110008e-01
6.75857887e-02 -8.02213967e-01 -2.95900732e-01 -6.54016614e-01
7.06759572e-01 -1.88372508e-01 8.29461634e-01 1.04232585e+00
4.65895772e-01 -8.02827537e-01 -7.57376134e-01 4.93742794e-01
1.24888551e+00 7.66898513e-01 -1.32582396e-01 4.98813838e-01
6.34809375e-01 3.41461092e-01 8.79998446e-01 5.28315127e-01
1.47158995e-01 4.52389836e-01 9.15417552e-01 7.21111894e-01
4.67584610e-01 -2.80694216e-01 7.30037808e-01 4.37154859e-01
5.85320592e-01 -2.45802596e-01 -7.15405703e-01 4.89301473e-01
-2.27013040e+00 -6.67635798e-01 6.73494279e-01 2.36070061e+00
4.01371360e-01 5.01899943e-02 -1.01550883e-02 -6.26292884e-01
8.72586846e-01 -5.15130796e-02 -1.52438307e+00 -4.42566872e-02
3.60459000e-01 -7.79619038e-01 1.33356643e+00 1.16106844e+00
-1.07089949e+00 9.77440417e-01 7.30458450e+00 4.58532631e-01
-1.32959771e+00 -1.46795407e-01 4.36838776e-01 -2.43165612e-01
2.75579810e-01 -4.94320512e-01 -1.34904730e+00 4.74951148e-01
1.17491949e+00 -2.75387079e-01 8.13671350e-01 1.34025419e+00
2.21752927e-01 -1.13014773e-01 -9.29602861e-01 9.35808420e-01
-1.81848690e-01 -9.16362226e-01 -4.64743942e-01 2.06027240e-01
4.35006440e-01 6.29160166e-01 5.58124781e-01 5.41670680e-01
8.34755480e-01 -6.10516250e-01 7.84925282e-01 6.64629996e-01
5.24737060e-01 -6.78219557e-01 6.67286694e-01 7.75047839e-01
-7.32378483e-01 -9.39480245e-01 -7.66748548e-01 -1.46326855e-01
1.91249564e-01 2.09969252e-01 -1.35988045e+00 -1.57216758e-01
4.81988609e-01 7.45865464e-01 4.53344807e-02 9.84159410e-01
-6.08766079e-02 5.46120778e-02 -8.35418820e-01 -9.56730783e-01
4.07692581e-01 -3.30958396e-01 1.13853312e+00 6.37841046e-01
5.95369816e-01 -9.95056555e-02 4.85492498e-01 7.40020514e-01
2.76113540e-01 -4.16331351e-01 -9.75595593e-01 -2.20003817e-02
5.00011027e-01 8.09130013e-01 -4.22181189e-01 -2.39061162e-01
2.51242548e-01 4.99761492e-01 6.42500699e-01 4.92685080e-01
-3.75605941e-01 -3.52480501e-01 8.53133500e-01 -4.73611951e-01
6.09416008e-01 -4.40041333e-01 -1.26420766e-01 -9.14910674e-01
-5.33617334e-03 -3.62605244e-01 2.90915579e-01 -3.67750049e-01
-1.06651390e+00 4.90922630e-01 1.76963657e-01 -9.50100839e-01
-9.34892952e-01 -1.14223230e+00 -1.74929529e-01 7.87195504e-01
-1.25017715e+00 -3.01915437e-01 3.60512704e-01 6.33323491e-01
3.69486570e-01 -4.44841892e-01 5.17101288e-01 -4.58448082e-01
-5.69014728e-01 2.76770562e-01 6.65749669e-01 -1.96121395e-01
1.25189096e-01 -1.20246899e+00 7.42893592e-02 7.11324155e-01
-7.08148062e-01 3.32494467e-01 9.77955878e-01 -9.59263325e-01
-1.88074374e+00 -1.33161223e+00 -2.00779028e-02 -4.92550671e-01
7.73036540e-01 -1.78840116e-01 -1.72881633e-01 9.36460733e-01
-1.92559630e-01 -6.54443502e-02 -3.98132831e-01 -1.91912055e-01
1.49063513e-01 -1.44621789e-01 -1.36753809e+00 6.68670058e-01
6.57603443e-01 -5.79331676e-03 -3.24959576e-01 3.17630410e-01
1.01043034e+00 -7.28292882e-01 -4.26240057e-01 4.89100873e-01
2.96317279e-01 1.21764615e-02 6.51021421e-01 -8.12963009e-01
-5.33906758e-01 -1.14158846e-01 -3.17492306e-01 -1.95871294e+00
-4.48233902e-01 -6.05724454e-01 -4.69235748e-01 1.69389024e-01
5.56253970e-01 -9.07311797e-01 1.09790933e+00 6.26431227e-01
-2.55845845e-01 -6.96003139e-01 -1.39365125e+00 -9.81349647e-01
-4.92037600e-03 -1.53275684e-01 1.11284545e-02 8.78287777e-02
1.86950922e-01 6.02939069e-01 -5.82601190e-01 4.45252180e-01
5.70677996e-01 -3.36280018e-01 3.38953674e-01 -9.27813828e-01
-4.63037528e-02 -5.58612704e-01 -1.07177332e-01 -1.43178761e+00
5.09525299e-01 -5.02645493e-01 8.84346366e-01 -1.36909759e+00
-2.39248201e-01 -9.18959528e-02 -1.41338691e-01 1.75771162e-01
1.46563968e-03 -8.79517853e-01 3.99792939e-01 -9.76115242e-02
-8.74564350e-01 1.05456793e+00 1.10713315e+00 -3.54269087e-01
-4.73765016e-01 6.19862497e-01 -5.58473051e-01 7.79565930e-01
8.73082221e-01 -5.64486504e-01 -5.35937548e-01 -3.51869732e-01
-4.51187342e-02 2.58429289e-01 -9.01631713e-02 -9.23294842e-01
3.23473930e-01 -5.24064064e-01 4.35624361e-01 -5.97123444e-01
7.22260535e-01 -1.20128715e+00 -3.69567752e-01 1.23384726e+00
-5.61081052e-01 2.07996935e-01 1.40737087e-01 1.04457033e+00
2.10934728e-01 -4.31164801e-01 1.05577767e+00 -1.78030714e-01
-6.81948900e-01 3.41485977e-01 -6.80993795e-01 -1.76075608e-01
9.23805535e-01 3.80946904e-01 -1.80895671e-01 -7.45968223e-01
-7.80622900e-01 6.87449515e-01 1.30389303e-01 3.79249960e-01
7.23935902e-01 -1.22678065e+00 -2.51640260e-01 7.61997402e-02
-4.35220242e-01 -1.30048051e-01 -2.29977563e-01 3.75785410e-01
9.14914533e-02 8.34929883e-01 -1.45985767e-01 -5.36215901e-01
-7.10468292e-01 9.74629462e-01 9.60616529e-01 -3.55528682e-01
-2.08651245e-01 8.24231029e-01 -9.48577076e-02 -8.27388465e-01
5.82212031e-01 -5.81815004e-01 -1.64099678e-01 -2.71377325e-01
6.05389237e-01 5.22968948e-01 -3.23289335e-01 -2.48066261e-01
-1.54522896e-01 3.17230046e-01 1.40505970e-01 -4.26776171e-01
1.04109001e+00 -4.03734952e-01 2.12080359e-01 4.96456176e-01
1.17587209e+00 -8.19783866e-01 -2.08224797e+00 -2.35313222e-01
2.42213249e-01 -1.29958779e-01 1.25140205e-01 -8.34705770e-01
-8.46135855e-01 4.08510983e-01 1.13562894e+00 -1.47131801e-01
5.97155333e-01 -1.27715230e-01 1.76158220e-01 1.39023364e+00
4.36109036e-01 -1.24499190e+00 1.66978359e-01 1.36780214e+00
8.51902485e-01 -1.17928660e+00 -3.51506412e-01 1.72710031e-01
-6.26545310e-01 8.34743321e-01 1.01204157e+00 -3.74051720e-01
9.25061703e-01 2.81334162e-01 7.31952935e-02 -2.20736698e-03
-8.17175925e-01 -1.39614850e-01 1.96368337e-01 9.48220432e-01
-5.46794415e-01 1.29687175e-01 1.09442614e-01 1.14051975e-01
-1.18004039e-01 -3.63001525e-01 1.72629252e-01 8.93150508e-01
-1.26655889e+00 -3.78202438e-01 -5.68573773e-01 5.30942738e-01
-5.07802591e-02 1.54147804e-01 -9.52214599e-02 6.37010276e-01
-5.97669303e-01 9.25937414e-01 9.81954560e-02 -3.38940948e-01
4.44884002e-01 -1.25339508e-01 4.08238173e-01 -5.02577126e-01
1.72036573e-01 -1.41359657e-01 -1.78544596e-01 -5.57815790e-01
2.42252350e-01 -6.29433692e-01 -1.15801358e+00 2.38709152e-01
-4.54905421e-01 2.51831234e-01 1.10323763e+00 9.44321632e-01
5.48871040e-01 2.90974140e-01 6.68679714e-01 -1.18525660e+00
-1.68901694e+00 -1.36156857e+00 -5.36094248e-01 -2.98512548e-01
6.36266589e-01 -9.15437520e-01 -4.05800432e-01 -4.76579756e-01] | [4.658564567565918, 2.18918776512146] |
caeb04ec-5f1a-460a-a8be-189f139843df | questions-for-flat-minima-optimization-of | 2202.00661 | null | https://arxiv.org/abs/2202.00661v5 | https://arxiv.org/pdf/2202.00661v5.pdf | When Do Flat Minima Optimizers Work? | Recently, flat-minima optimizers, which seek to find parameters in low-loss neighborhoods, have been shown to improve a neural network's generalization performance over stochastic and adaptive gradient-based optimizers. Two methods have received significant attention due to their scalability: 1. Stochastic Weight Averaging (SWA), and 2. Sharpness-Aware Minimization (SAM). However, there has been limited investigation into their properties and no systematic benchmarking of them across different domains. We fill this gap here by comparing the loss surfaces of the models trained with each method and through broad benchmarking across computer vision, natural language processing, and graph representation learning tasks. We discover several surprising findings from these results, which we hope will help researchers further improve deep learning optimizers, and practitioners identify the right optimizer for their problem. | ['Matt J. Kusner', 'Ricardo Silva', 'Linqing Liu', 'Jean Kaddour'] | 2022-02-01 | null | null | null | null | ['self-supervised-image-classification'] | ['computer-vision'] | [ 2.54378021e-02 -1.08951055e-01 -3.86685878e-01 -6.95288897e-01
-8.54160786e-01 -3.92455488e-01 3.32047135e-01 3.04634362e-01
-6.64186478e-01 6.21902108e-01 4.27192330e-01 -3.28686446e-01
-3.75813484e-01 -6.28288507e-01 -7.06865311e-01 -5.83874702e-01
-1.41390994e-01 3.01134944e-01 5.63114956e-02 -1.11132771e-01
5.14006078e-01 4.35776860e-01 -1.14592636e+00 1.43246591e-01
9.80499268e-01 1.10181034e+00 1.10592470e-01 2.82907009e-01
-2.35501692e-01 6.16072834e-01 -1.63960204e-01 -6.94208860e-01
2.43327439e-01 -4.34843749e-01 -7.54992723e-01 -3.00969630e-01
8.85187447e-01 2.00697437e-01 -4.15620297e-01 1.15587389e+00
4.51840013e-01 5.08583426e-01 5.76271474e-01 -8.66588712e-01
-9.12303746e-01 4.92384374e-01 -5.60048342e-01 5.58660567e-01
-1.31923422e-01 2.46189818e-01 1.68327272e+00 -9.54656661e-01
6.54163003e-01 1.31219161e+00 1.11788309e+00 6.23729765e-01
-1.37679446e+00 -3.33962917e-01 4.39876497e-01 3.85687411e-01
-1.29266620e+00 -6.59363270e-01 7.98913419e-01 -1.58343092e-01
1.29685342e+00 1.10311531e-01 5.35610676e-01 7.91469753e-01
3.09407353e-01 1.05606484e+00 7.67773926e-01 -2.31850505e-01
1.66539893e-01 5.35175875e-02 5.77738285e-02 1.20270622e+00
2.30990127e-01 -1.61561191e-01 -7.63012648e-01 -1.78871322e-02
3.33325833e-01 -1.76583782e-01 -3.52378756e-01 -4.65109557e-01
-7.64268816e-01 1.07086003e+00 7.03906715e-01 1.78737238e-01
-2.28817537e-01 3.46135616e-01 3.59020412e-01 3.83641273e-01
8.22904646e-01 1.05984437e+00 -4.37602729e-01 -9.75233987e-02
-1.01626587e+00 3.99249017e-01 7.25678504e-01 5.05770028e-01
8.29793394e-01 1.99720219e-01 -4.63315919e-02 1.17433810e+00
4.93465483e-01 1.97476968e-01 3.78435731e-01 -7.82350302e-01
7.55916893e-01 6.98399723e-01 -4.03057277e-01 -1.38898897e+00
-4.94127750e-01 -6.77627861e-01 -5.45560956e-01 1.25585303e-01
3.88097376e-01 -2.53860205e-01 -7.10430980e-01 1.86278367e+00
1.11417891e-02 2.60902178e-02 -3.21017295e-01 9.95368302e-01
7.39020944e-01 3.75114709e-01 1.33103520e-01 2.20345989e-01
7.37319827e-01 -1.22763503e+00 -3.91248047e-01 -5.50192297e-01
8.35951447e-01 -6.95595682e-01 1.28877223e+00 2.60581970e-01
-1.18569505e+00 -2.16915220e-01 -1.28350496e+00 -2.24702373e-01
-4.07403052e-01 -2.25666445e-02 8.65937293e-01 6.32944286e-01
-1.29135239e+00 1.13568354e+00 -9.86457944e-01 -3.43721151e-01
8.00752223e-01 3.25572610e-01 -1.45004302e-01 -1.61916897e-01
-9.65810537e-01 1.08097827e+00 -3.27031016e-02 7.16181397e-02
-7.62910247e-01 -8.30995679e-01 -8.23070049e-01 2.04252750e-02
2.83756852e-01 -8.63995254e-01 8.28583896e-01 -9.61096227e-01
-1.23821831e+00 9.93104398e-01 -3.80616963e-01 -7.35449374e-01
2.21083611e-01 -2.68922210e-01 5.84269688e-02 -2.08834976e-01
-2.46368274e-01 4.31262970e-01 7.38704503e-01 -9.20178115e-01
-4.15789992e-01 -6.06098354e-01 -1.35185897e-01 4.30149078e-01
-6.08722866e-01 3.45772952e-02 -3.49491388e-01 -6.74171567e-01
1.62850544e-01 -7.31856942e-01 -3.71015906e-01 6.59768656e-02
-5.10627031e-01 -3.72649580e-01 6.17559850e-01 -6.39758587e-01
1.30854714e+00 -1.88763130e+00 3.78395349e-01 2.00223967e-01
4.10117507e-01 4.22933906e-01 -4.33783591e-01 2.34573767e-01
1.02427632e-01 3.70960563e-01 -3.68806273e-01 -7.93588698e-01
3.41068432e-02 1.20977029e-01 -2.33660027e-01 6.28797472e-01
3.61930847e-01 1.21042442e+00 -8.09127808e-01 -3.51611078e-01
-2.77105533e-02 4.86784965e-01 -7.27737010e-01 -9.68992412e-02
-2.01254770e-01 -1.02104895e-01 -5.49015760e-01 5.99762917e-01
4.10306424e-01 -4.05478597e-01 1.30012557e-02 -2.44177982e-01
1.44520476e-01 6.90086484e-01 -9.00085568e-01 1.57224596e+00
-4.30295438e-01 1.00276184e+00 1.76057681e-01 -1.35580409e+00
8.79534304e-01 -1.23763040e-01 4.59289789e-01 -6.90817773e-01
-7.10329637e-02 1.31073952e-01 -1.23060353e-01 -4.17197764e-01
3.78491759e-01 -1.47658646e-01 5.41355431e-01 4.41791534e-01
9.24450457e-02 4.81181592e-03 1.15641288e-01 1.18338332e-01
1.24683905e+00 -7.64162913e-02 -5.66220395e-02 -4.31184202e-01
2.20475018e-01 4.91086021e-03 4.72136229e-01 9.91052747e-01
-3.42707247e-01 7.90410876e-01 2.89172560e-01 -6.32757485e-01
-8.31284046e-01 -1.12717235e+00 3.01039480e-02 1.26887977e+00
-1.50744831e-02 -3.94293845e-01 -6.20974958e-01 -8.01770627e-01
1.52101457e-01 6.65265381e-01 -4.43591326e-01 -4.20872837e-01
-8.05142820e-01 -1.21140563e+00 4.60533977e-01 4.42586333e-01
4.50941086e-01 -8.84226739e-01 -1.63376480e-01 2.70991385e-01
2.83553023e-02 -9.07065332e-01 -7.94484675e-01 1.02262191e-01
-1.32978594e+00 -9.35479999e-01 -5.43793857e-01 -8.94865572e-01
5.53780675e-01 1.89195961e-01 1.48814833e+00 3.73472065e-01
-3.72778744e-01 4.83164519e-01 -6.43792152e-02 -2.92147160e-01
-1.93840209e-02 4.33083266e-01 1.07523397e-01 2.22357810e-02
4.25663412e-01 -6.21642351e-01 -8.33360076e-01 2.90129304e-01
-5.37203372e-01 -3.68123114e-01 4.14404333e-01 6.43205702e-01
6.92341745e-01 -1.73944265e-01 5.17655611e-01 -8.82025540e-01
1.19631124e+00 -4.66147453e-01 -4.22421247e-01 4.72085893e-01
-1.28220427e+00 4.88296449e-01 6.58242881e-01 -1.65399179e-01
-7.48975813e-01 -3.12932372e-01 -1.31636202e-01 -2.36229762e-01
2.73939461e-01 6.98226810e-01 2.81258225e-01 -6.41218603e-01
9.63488400e-01 1.02785341e-01 7.92522728e-03 -4.93586183e-01
2.51429886e-01 1.66868001e-01 1.69585526e-01 -7.03109741e-01
5.92693269e-01 3.54986459e-01 3.51519585e-02 -7.89870679e-01
-1.26488721e+00 -4.55009282e-01 -1.35964841e-01 2.70907767e-02
7.62319863e-01 -6.59285307e-01 -3.72518629e-01 3.91586542e-01
-9.66024160e-01 -4.17736202e-01 -1.98407307e-01 4.81376588e-01
-4.71049607e-01 2.85753459e-01 -6.20376587e-01 -5.28040707e-01
-5.22357941e-01 -1.24845791e+00 6.98456287e-01 4.50956613e-01
-1.63164094e-01 -1.61002743e+00 1.11315794e-01 2.84470707e-01
8.54278564e-01 1.05580583e-01 1.03384221e+00 -6.30233705e-01
-5.73274910e-01 -1.78508878e-01 -3.78108591e-01 5.28512061e-01
6.06387504e-04 3.34244519e-02 -8.03688705e-01 -3.36942136e-01
3.18577215e-02 -4.66646254e-01 1.36751914e+00 8.00913155e-01
1.36882758e+00 -3.82710725e-01 -4.31709945e-01 1.06419027e+00
1.55956042e+00 -2.71766335e-01 3.59023571e-01 4.01638120e-01
7.69233167e-01 3.57341766e-01 6.69545904e-02 -5.73070757e-02
4.09133166e-01 7.48907626e-01 3.96638304e-01 -2.53311135e-02
-3.91768515e-01 -3.12666029e-01 2.81075537e-01 7.07079113e-01
-1.48189172e-01 -1.81963339e-01 -9.61357832e-01 5.60404360e-01
-1.93685579e+00 -8.49367917e-01 1.69879958e-01 2.06263876e+00
7.98224032e-01 1.78552657e-01 -2.92523429e-02 -2.98105597e-01
5.15675902e-01 5.96712351e-01 -9.00907457e-01 -5.58301687e-01
-3.08575660e-01 4.06422079e-01 6.62237704e-01 6.83964908e-01
-1.09553516e+00 1.09502017e+00 7.08038855e+00 9.68871593e-01
-1.06555641e+00 -1.51533097e-01 8.97301257e-01 -3.53107810e-01
-4.24297690e-01 -5.23386300e-02 -9.76327658e-01 1.37280256e-01
8.16902876e-01 -1.53276369e-01 9.59835768e-01 9.05878007e-01
2.72666693e-01 2.73533136e-01 -1.16413033e+00 1.06009901e+00
8.20968449e-02 -1.73387039e+00 -4.31222245e-02 -4.68581766e-02
9.55694079e-01 7.13715613e-01 3.49321008e-01 3.05515856e-01
4.11368012e-01 -1.37153399e+00 2.65369356e-01 6.24370456e-01
1.67718410e-01 -5.44597030e-01 3.72031271e-01 9.34736580e-02
-1.04611290e+00 -1.17724128e-01 -5.90222359e-01 1.31421685e-02
1.80361748e-01 7.87115872e-01 -7.54351735e-01 -5.63215874e-02
6.16277218e-01 9.40641165e-01 -7.45312870e-01 1.31960154e+00
-1.40333831e-01 1.01308537e+00 -3.68637234e-01 -4.30945873e-01
5.17360389e-01 -4.57770944e-01 7.89491236e-01 1.17274725e+00
2.76903231e-02 -3.30848038e-01 -3.80926009e-04 1.12070954e+00
-4.35321569e-01 3.08370948e-01 -4.94929850e-01 -2.98959553e-01
1.97570190e-01 9.93603706e-01 -6.36732876e-01 1.82053164e-01
-3.49943072e-01 6.66696012e-01 9.68491554e-01 4.85311329e-01
-5.03892362e-01 -5.60287833e-01 9.85901535e-01 1.08108602e-01
1.56406477e-01 -5.35241842e-01 -9.06923234e-01 -1.08818281e+00
1.50655210e-01 -8.53246748e-01 5.20166397e-01 -3.51693749e-01
-1.56077635e+00 4.83012974e-01 -2.40306437e-01 -3.04923654e-01
4.92240004e-02 -8.84329259e-01 -7.32029974e-01 8.01782012e-01
-1.50412953e+00 -6.38380706e-01 1.29312471e-01 3.83480817e-01
5.53337693e-01 -3.89739931e-01 5.25704801e-01 2.43749157e-01
-7.07057297e-01 9.37368095e-01 3.04335535e-01 5.39083704e-02
3.62825125e-01 -1.29495621e+00 6.07401788e-01 6.78405881e-01
6.61451221e-01 6.97672427e-01 6.09056413e-01 -3.50550592e-01
-1.53683031e+00 -9.20004487e-01 9.74879920e-01 -5.62349617e-01
6.88835919e-01 -1.09639831e-01 -1.16366720e+00 5.57745099e-01
1.58546850e-01 2.92387232e-02 4.90997583e-01 6.29164815e-01
-4.17076766e-01 -2.26726696e-01 -1.03057909e+00 7.07504809e-01
1.29592836e+00 -4.73772734e-01 -3.68880540e-01 6.09213471e-01
4.36196029e-01 -2.44682178e-01 -7.23316193e-01 3.22962493e-01
4.32556391e-01 -1.09631658e+00 1.11944056e+00 -1.07946014e+00
4.86631244e-01 2.11423695e-01 -1.38666898e-01 -1.58863127e+00
-2.07743108e-01 -5.53290367e-01 -3.46030109e-03 8.06403756e-01
8.48580062e-01 -8.67036283e-01 1.12591708e+00 8.19562197e-01
-2.74585217e-01 -1.48389757e+00 -8.37729692e-01 -7.72193432e-01
3.28751832e-01 -4.68081772e-01 3.39587897e-01 7.85099566e-01
-2.13233709e-01 3.44036043e-01 -1.20488495e-01 -2.07472622e-01
7.47977734e-01 4.16485546e-03 4.09136415e-01 -1.02679074e+00
-3.08805317e-01 -1.14099181e+00 -4.26958174e-01 -1.13122690e+00
3.07447970e-01 -1.29433370e+00 -1.06674872e-01 -1.56579065e+00
1.31733119e-01 -5.50683022e-01 -4.40550685e-01 4.89723355e-01
-3.90571862e-01 1.68865666e-01 -9.26827714e-02 1.52482584e-01
-8.03781390e-01 8.49069715e-01 1.13275623e+00 -2.99945474e-01
-2.50790507e-01 1.82601541e-01 -9.65884149e-01 6.90688789e-01
9.68703091e-01 -6.28594756e-01 -4.09473777e-01 -8.64524841e-01
5.86755514e-01 -6.68642640e-01 6.94421232e-02 -9.48857963e-01
2.05820546e-01 -2.34419525e-01 2.22692952e-01 1.28764391e-01
1.90262288e-01 -1.84853584e-01 -5.55697978e-01 3.08092147e-01
-6.21026337e-01 2.48094186e-01 1.61072478e-01 5.60500383e-01
-1.15458302e-01 -4.40938354e-01 9.31183994e-01 -2.03425199e-01
-6.27043962e-01 6.77158713e-01 1.51704818e-01 7.27489531e-01
4.43706095e-01 -7.62259439e-02 -1.44492149e-01 -5.28200448e-01
-5.20151258e-01 2.89171398e-01 2.81890720e-01 3.39472830e-01
7.56534457e-01 -1.31590569e+00 -8.44836533e-01 -2.42287181e-02
-1.40383750e-01 -1.31280452e-01 -1.20588951e-01 9.01802242e-01
-4.64205801e-01 3.05105150e-01 2.17741475e-01 -3.89070749e-01
-1.03037953e+00 1.94528744e-01 6.30501986e-01 -4.25281435e-01
-5.29138982e-01 1.40182614e+00 -1.19175375e-01 -6.03182554e-01
4.31715339e-01 -1.26833513e-01 7.81005099e-02 -2.54880667e-01
3.65086406e-01 4.58807081e-01 2.90262103e-01 -3.89573544e-01
-4.54567611e-01 7.64981806e-01 -2.45501220e-01 2.02156305e-01
1.47963440e+00 -2.77678594e-02 -5.50229549e-02 2.59367824e-01
1.61999547e+00 -1.52956039e-01 -1.32571149e+00 -4.92453128e-01
3.12221974e-01 -4.37910765e-01 3.36707562e-01 -4.15652752e-01
-1.33864748e+00 9.00673568e-01 4.21145171e-01 2.91303575e-01
9.44977164e-01 9.08265822e-03 7.89457023e-01 7.05449522e-01
7.61050805e-02 -1.28416061e+00 1.47510841e-01 6.71952248e-01
8.71769965e-01 -1.32613099e+00 2.83142060e-01 1.43718347e-01
-5.33849418e-01 1.02722168e+00 3.50734860e-01 -3.45697135e-01
8.39341879e-01 2.63057780e-02 2.86365431e-02 -4.66871351e-01
-8.77566993e-01 -1.23547621e-01 6.30515397e-01 5.48177779e-01
6.36550605e-01 -1.10715553e-01 -2.47193128e-01 1.63391933e-01
-2.72520244e-01 -3.33827615e-01 -9.56905261e-02 4.10727978e-01
-5.75740337e-01 -9.87019598e-01 1.40622973e-01 7.68524230e-01
-5.04019856e-01 -3.21171641e-01 -5.46609104e-01 4.60180849e-01
-4.48669672e-01 9.28984463e-01 -1.49206832e-01 -4.10546839e-01
3.08067292e-01 1.78388998e-01 6.65358007e-01 -4.85657483e-01
-4.86419767e-01 -7.13394821e-01 6.88983351e-02 -8.59292567e-01
-1.03417084e-01 -7.38169491e-01 -9.65944588e-01 -2.34261379e-01
-3.00963670e-01 5.33667989e-02 8.95983934e-01 1.09701073e+00
5.05487978e-01 1.99167952e-01 4.96586770e-01 -6.68469727e-01
-9.39349830e-01 -6.92969799e-01 -3.31496954e-01 3.83405447e-01
2.73952127e-01 -4.94781971e-01 -6.06662214e-01 -4.31752682e-01] | [8.25529956817627, 3.523998498916626] |
07e67162-9d77-4280-aec8-b63c774f3e97 | efficientad-accurate-visual-anomaly-detection | 2303.14535 | null | https://arxiv.org/abs/2303.14535v1 | https://arxiv.org/pdf/2303.14535v1.pdf | EfficientAD: Accurate Visual Anomaly Detection at Millisecond-Level Latencies | Detecting anomalies in images is an important task, especially in real-time computer vision applications. In this work, we focus on computational efficiency and propose a lightweight feature extractor that processes an image in less than a millisecond on a modern GPU. We then use a student-teacher approach to detect anomalous features. We train a student network to predict the extracted features of normal, i.e., anomaly-free training images. The detection of anomalies at test time is enabled by the student failing to predict their features. We propose a training loss that hinders the student from imitating the teacher feature extractor beyond the normal images. It allows us to drastically reduce the computational cost of the student-teacher model, while improving the detection of anomalous features. We furthermore address the detection of challenging logical anomalies that involve invalid combinations of normal local features, for example, a wrong ordering of objects. We detect these anomalies by efficiently incorporating an autoencoder that analyzes images globally. We evaluate our method, called EfficientAD, on 32 datasets from three industrial anomaly detection dataset collections. EfficientAD sets new standards for both the detection and the localization of anomalies. At a latency of two milliseconds and a throughput of six hundred images per second, it enables a fast handling of anomalies. Together with its low error rate, this makes it an economical solution for real-world applications and a fruitful basis for future research. | ['Rebecca König', 'Lars Heckler', 'Kilian Batzner'] | 2023-03-25 | null | null | null | null | ['semi-supervised-anomaly-detection'] | ['computer-vision'] | [ 2.13263869e-01 -4.00319338e-01 5.06830513e-01 -2.94768244e-01
-3.51682127e-01 -3.69789153e-01 4.74149704e-01 6.75174057e-01
-4.55187351e-01 1.04110986e-01 -8.34078610e-01 -5.03641665e-01
1.20083578e-01 -9.41326082e-01 -8.02513063e-01 -8.83469641e-01
-1.79040447e-01 3.00736099e-01 6.29790962e-01 1.43814400e-01
4.64864552e-01 1.00617182e+00 -2.25137615e+00 3.15142602e-01
7.73336232e-01 1.26606429e+00 -8.94482285e-02 9.90066886e-01
-9.03076306e-03 7.31290996e-01 -9.54776406e-01 -2.50895005e-02
2.63849646e-01 -1.92847311e-01 -5.00421584e-01 4.00371820e-01
8.52426231e-01 -6.28722370e-01 -5.35436086e-02 9.90365326e-01
3.24494451e-01 4.46319245e-02 3.46535861e-01 -1.61043024e+00
-1.17480353e-01 -1.18702583e-01 -7.43310034e-01 7.40336001e-01
2.36643896e-01 3.02805036e-01 6.34525955e-01 -6.16959095e-01
1.50463089e-01 7.51489580e-01 4.76067007e-01 3.27244669e-01
-1.20971870e+00 -4.38766569e-01 2.64359057e-01 7.28110135e-01
-1.06505442e+00 -1.88903481e-01 6.35218382e-01 -4.93770599e-01
1.11482239e+00 4.77738947e-01 5.53322911e-01 8.42299223e-01
2.72672176e-01 7.99216151e-01 6.28795028e-01 -4.51831073e-01
3.95957857e-01 -5.83218746e-02 3.77287388e-01 7.66535580e-01
2.99543738e-01 -5.69600202e-02 -4.03028667e-01 -2.34150931e-01
5.81178367e-01 4.72867370e-01 1.82414521e-02 -2.50012904e-01
-1.00377178e+00 6.62994266e-01 1.16829187e-01 1.40353382e-01
-6.70120776e-01 -9.95618179e-02 6.77314222e-01 7.63463259e-01
3.22752833e-01 3.24263453e-01 -5.47455609e-01 -1.90282926e-01
-7.12836325e-01 1.24538243e-01 6.54904842e-01 4.94703025e-01
7.84587622e-01 2.35401168e-01 1.21181376e-01 4.27929014e-01
-2.44327039e-01 3.53833944e-01 5.68835497e-01 -7.31996059e-01
-9.61521789e-02 9.59699392e-01 -1.54143363e-01 -1.10779238e+00
-2.64999539e-01 -2.45169476e-01 -7.90832818e-01 6.58129275e-01
5.47113717e-01 1.33190364e-01 -8.85565937e-01 1.13380206e+00
7.19672382e-01 6.24488056e-01 -1.06804937e-01 6.47235572e-01
2.29261458e-01 5.93540013e-01 -1.88205481e-01 -1.29614562e-01
1.45034349e+00 -7.00107098e-01 -2.47490749e-01 -8.88704658e-02
8.79184365e-01 -8.33825231e-01 1.06610513e+00 1.06029117e+00
-8.09043467e-01 -6.14584565e-01 -8.79185557e-01 2.84215748e-01
-4.16624337e-01 -2.22681761e-02 5.49592793e-01 3.73753011e-01
-8.66668642e-01 8.31353128e-01 -1.18357420e+00 -2.33894616e-01
3.11217785e-01 5.68900883e-01 -4.06746924e-01 7.69729838e-02
-3.80868882e-01 4.79780793e-01 4.23455715e-01 -1.18454374e-01
-6.38344884e-01 -7.09411204e-01 -8.33441854e-01 2.12029234e-01
4.04108882e-01 -9.85787362e-02 1.26972234e+00 -1.23665047e+00
-1.14439464e+00 7.80099571e-01 -8.07158351e-02 -5.75168669e-01
3.63534838e-01 -3.99471849e-01 -5.21611631e-01 3.53356481e-01
-2.10553817e-02 1.68594554e-01 1.33734548e+00 -6.56557858e-01
-1.07594037e+00 -4.78239983e-01 -4.07512575e-01 -3.17826599e-01
-5.50208807e-01 -2.37636324e-02 -2.04752386e-01 -4.27724212e-01
1.74113736e-01 -8.98985088e-01 -1.91313118e-01 -2.90570371e-02
-3.49050075e-01 -5.72971582e-01 1.40348125e+00 -2.89657146e-01
8.23473752e-01 -2.46965003e+00 -6.17717743e-01 6.35426819e-01
2.95204729e-01 5.05147040e-01 -1.13314293e-01 5.94238080e-02
-3.06356668e-01 -4.60698187e-01 -1.04848862e-01 -1.57808512e-01
-4.45931971e-01 4.39665169e-01 -6.28284693e-01 5.98503947e-01
5.88151395e-01 3.56907040e-01 -7.42970407e-01 -2.77907580e-01
4.49291140e-01 1.82956353e-01 -6.61911726e-01 5.35854161e-01
5.23165315e-02 3.95400733e-01 -4.18552786e-01 5.63688040e-01
6.48094416e-01 -1.66477814e-01 -3.10084164e-01 2.18334675e-01
-1.85885698e-01 8.24489966e-02 -1.34280610e+00 9.40113962e-01
-4.11468118e-01 6.47493541e-01 -2.23899052e-01 -1.26208103e+00
9.51009214e-01 1.32599592e-01 5.45028269e-01 -8.17583323e-01
-1.39272228e-01 3.11281711e-01 1.11947909e-01 -4.81845528e-01
2.80687869e-01 5.57029247e-01 2.33435124e-01 6.39834583e-01
6.21519461e-02 2.60923713e-01 3.79916102e-01 1.46619514e-01
1.53103256e+00 -2.99617887e-01 1.80234089e-01 -1.29056571e-04
6.78463519e-01 -9.22569185e-02 4.90102589e-01 8.94739091e-01
-1.00662932e-01 3.63436133e-01 7.10856259e-01 -1.02750647e+00
-9.85018671e-01 -1.10172951e+00 4.66561839e-02 1.25701213e+00
-2.25265808e-02 -3.57425392e-01 -5.38369715e-01 -9.61207688e-01
1.26713350e-01 5.34718275e-01 -3.31507444e-01 -3.67748678e-01
-8.38994145e-01 -4.42534626e-01 2.34899819e-01 6.02306902e-01
2.79634058e-01 -1.28470790e+00 -1.15306664e+00 1.42457664e-01
2.85158575e-01 -1.16083789e+00 -1.02840081e-01 4.39595461e-01
-9.98359025e-01 -1.30974436e+00 -9.39222127e-02 -6.89777553e-01
1.04891205e+00 2.52474040e-01 1.14071751e+00 6.73721075e-01
-7.98000097e-01 4.51846957e-01 -2.21685842e-01 -5.95090330e-01
-3.87925267e-01 -3.32007498e-01 2.64069796e-01 1.58386126e-01
8.00347924e-01 -7.12531328e-01 -5.98665833e-01 2.80621976e-01
-1.01006651e+00 -4.67390835e-01 4.90819812e-01 8.52006674e-01
7.59918272e-01 3.68037075e-01 3.26736271e-01 -7.31784642e-01
2.87510306e-01 -4.31896508e-01 -1.12632501e+00 -1.38278276e-01
-6.51959717e-01 7.57244825e-02 1.09446335e+00 -5.94833732e-01
-6.04077458e-01 3.16053092e-01 -2.27124095e-01 -5.52389264e-01
-6.41566277e-01 -1.28426284e-01 3.30525041e-01 -1.44807994e-01
7.14660227e-01 4.53184664e-01 3.59919332e-02 -3.93839031e-01
-2.90288538e-01 4.22714919e-01 8.11281383e-01 -5.57103992e-01
7.56522417e-01 4.17532772e-01 9.37107280e-02 -1.16677916e+00
-5.50395191e-01 -8.11645985e-01 -5.36081791e-01 -7.63682574e-02
4.34951723e-01 -7.25480497e-01 -1.08822489e+00 6.44832015e-01
-1.03801036e+00 -1.43751755e-01 -4.33849633e-01 3.08573008e-01
-2.96142250e-01 5.16834080e-01 -5.74968517e-01 -6.76728606e-01
-3.31244856e-01 -1.10047376e+00 1.05738711e+00 2.44127378e-01
-2.06141219e-01 -6.90094233e-01 -1.55011073e-01 -1.49470285e-01
2.71670163e-01 2.95789510e-01 7.12973595e-01 -1.20297146e+00
-6.22404695e-01 -4.91115361e-01 -1.33432195e-01 5.33851981e-01
-1.81160830e-02 3.54141384e-01 -1.03515863e+00 -3.67988169e-01
2.01694727e-01 -1.81369543e-01 8.71827245e-01 -1.68583691e-02
1.78621936e+00 -3.05810958e-01 -1.21222608e-01 4.27199721e-01
1.09372330e+00 2.47956902e-01 4.78915840e-01 5.57221353e-01
6.45433843e-01 4.31960583e-01 6.18573904e-01 6.98798358e-01
-4.44339663e-02 4.20267075e-01 6.70592427e-01 -9.05360952e-02
4.48614508e-01 3.10951799e-01 3.87875319e-01 6.04821801e-01
7.12823635e-03 -6.31342903e-02 -1.05073225e+00 6.53240860e-01
-1.71298563e+00 -8.17684412e-01 -4.14283961e-01 2.28885627e+00
4.63361651e-01 2.39424706e-01 1.94880471e-01 6.94991589e-01
5.09364963e-01 -2.25804254e-01 -5.88585019e-01 -8.65524411e-01
3.80653888e-01 5.12720048e-01 1.44114912e-01 1.10395476e-01
-1.33219731e+00 6.33119226e-01 5.14774656e+00 4.64433998e-01
-1.34694529e+00 -3.13447237e-01 5.54685354e-01 8.70295276e-04
2.54148543e-01 -3.56596053e-01 -6.94979548e-01 5.88659585e-01
1.04435253e+00 2.07912192e-01 1.14873976e-01 1.43141520e+00
-1.28771096e-01 -3.49760890e-01 -1.29341245e+00 1.00933409e+00
7.94775039e-02 -9.63797152e-01 -7.45668039e-02 9.98582393e-02
5.35061419e-01 1.94266089e-03 8.83289725e-02 2.07151100e-01
4.92620245e-02 -6.86197877e-01 2.29031116e-01 1.40956521e-01
2.91442603e-01 -1.14703548e+00 9.03920352e-01 4.82931376e-01
-8.01609993e-01 -2.72808373e-01 -3.74860078e-01 -2.37538353e-01
-5.01920640e-01 9.99556661e-01 -1.04960263e+00 5.71848489e-02
9.75068867e-01 3.38415861e-01 -7.74495065e-01 9.51804161e-01
-1.31629810e-01 6.49246335e-01 -6.36269689e-01 2.10770831e-01
2.86384225e-01 -3.39753344e-03 5.07530749e-01 1.11334467e+00
5.06892979e-01 -6.69100210e-02 4.06869859e-01 3.92262638e-01
2.16581851e-01 6.58177286e-02 -7.26909757e-01 2.60127336e-01
2.69040078e-01 1.29501140e+00 -9.37568009e-01 -3.81698728e-01
-4.11124289e-01 1.29249203e+00 2.16641486e-01 -2.89034694e-02
-5.22431195e-01 -6.05345786e-01 8.12947333e-01 -6.80308864e-02
5.52499115e-01 -1.65960997e-01 -2.11479202e-01 -1.07341146e+00
4.07867163e-01 -1.02380884e+00 5.97539604e-01 -1.50622085e-01
-1.15197408e+00 5.55545330e-01 -4.32387799e-01 -1.21880221e+00
-7.33576655e-01 -8.19940627e-01 -1.11058486e+00 4.07349408e-01
-1.36933315e+00 -5.46686649e-01 -5.38162053e-01 8.92134070e-01
4.55300152e-01 -2.03663811e-01 8.58605623e-01 1.58774897e-01
-7.15676308e-01 5.92259288e-01 -1.08004138e-01 1.40683129e-01
6.27521515e-01 -1.48119640e+00 7.62607813e-01 1.20023894e+00
4.69650745e-01 3.49706501e-01 7.97316194e-01 -4.02280599e-01
-1.15441585e+00 -1.13630152e+00 6.68949068e-01 -3.14618647e-01
6.20091856e-01 -1.22634009e-01 -1.40017235e+00 6.12439334e-01
-1.56302959e-01 5.50723970e-01 5.64262986e-01 2.15498656e-02
-3.92766923e-01 -3.03075075e-01 -1.13576686e+00 4.32969064e-01
4.91591007e-01 -3.49049658e-01 -3.65013570e-01 3.35335881e-01
2.75457025e-01 -3.26044172e-01 -5.20770788e-01 2.23580435e-01
2.59876370e-01 -1.26338089e+00 8.11216831e-01 -5.10630429e-01
2.92180419e-01 -2.98414826e-01 2.74185956e-01 -1.18853545e+00
-1.44266486e-02 -4.32695121e-01 -6.08915925e-01 8.81411076e-01
-1.41222760e-01 -8.03929567e-01 8.77446055e-01 3.48306000e-01
-7.11790174e-02 -8.11397374e-01 -8.72621715e-01 -8.43222857e-01
-5.83665192e-01 -6.57834053e-01 4.62644190e-01 8.67609322e-01
-4.31748539e-01 -1.87963694e-01 -4.11406346e-03 6.36613309e-01
6.38770461e-01 2.68049806e-01 9.26414490e-01 -1.49265802e+00
-2.38587543e-01 -2.82010496e-01 -1.12274849e+00 -6.85089827e-01
5.55862971e-02 -3.20173711e-01 -4.58317213e-02 -5.90690196e-01
-1.78703129e-01 -1.94287091e-01 -4.61007327e-01 6.73125982e-01
-2.07831502e-01 4.99489337e-01 -1.44951969e-01 -5.09561524e-02
-5.80511272e-01 9.65267345e-02 6.45184875e-01 1.28973886e-01
-3.28867406e-01 3.64729464e-01 -1.73808962e-01 1.17037439e+00
9.08596277e-01 -5.80511689e-01 -1.66733846e-01 -2.37996191e-01
8.28035474e-02 -4.04664636e-01 6.65408969e-01 -1.30056751e+00
3.46430600e-01 9.62272193e-03 7.51891315e-01 -7.19048321e-01
-1.22530863e-03 -1.06208646e+00 -6.08789325e-01 6.40496075e-01
1.24583638e-03 5.78970790e-01 3.30565721e-01 4.91864741e-01
-2.83014327e-01 -2.08655626e-01 7.82034218e-01 6.22485206e-02
-9.72278535e-01 2.31528312e-01 -4.26704228e-01 -2.42742270e-01
1.42997313e+00 -2.76113540e-01 -1.93218976e-01 -1.91516191e-01
-5.37611604e-01 2.75338590e-01 2.99060106e-01 2.70597517e-01
8.22417796e-01 -1.02860963e+00 -5.27786076e-01 9.47617590e-01
2.30577603e-01 2.39725351e-01 3.04247558e-01 7.92177260e-01
-7.35199451e-01 -9.96438414e-02 -3.06010246e-01 -1.17318344e+00
-1.77110898e+00 6.26420200e-01 1.19896233e-01 -1.83528975e-01
-1.00769305e+00 6.89991117e-01 1.01117313e-01 -6.08964562e-02
3.41401935e-01 -2.07784355e-01 -5.89587949e-02 -6.27089217e-02
1.03474867e+00 4.62812662e-01 5.30916035e-01 -1.74307480e-01
-3.40221375e-01 2.53746957e-01 -4.94057477e-01 5.29862285e-01
1.36471212e+00 2.42411822e-01 -2.36477673e-01 3.73615235e-01
9.80226994e-01 -2.92979274e-02 -1.12175596e+00 -2.40028575e-01
3.96609038e-01 -6.78839326e-01 -1.08565856e-02 -4.31983888e-01
-1.06845248e+00 9.49775457e-01 9.06580925e-01 6.46629393e-01
1.55687928e+00 -2.21560657e-01 8.92462373e-01 7.91507423e-01
-1.82684660e-02 -9.86065507e-01 3.55129629e-01 6.71826005e-01
4.25505012e-01 -1.43110371e+00 -1.11925066e-01 -1.33644238e-01
-4.52812344e-01 1.52805293e+00 1.01252139e+00 -4.17626888e-01
3.13521981e-01 4.61612135e-01 2.14226156e-01 -1.54951945e-01
-9.06552434e-01 -8.72221589e-02 2.54379719e-01 4.45965886e-01
-5.19305058e-02 -1.01241685e-01 2.89047897e-01 9.85984318e-03
-1.04628958e-01 -4.33991939e-01 6.27729535e-01 9.89982307e-01
-5.70499599e-01 -8.90306771e-01 -5.24482548e-01 6.57652557e-01
-6.51908517e-01 1.32126123e-01 -1.22143157e-01 5.58724523e-01
2.50247538e-01 8.04588675e-01 8.81004095e-01 -2.79296577e-01
3.07942986e-01 1.83829114e-01 1.11948371e-01 -4.54051077e-01
-5.92540085e-01 -2.48513386e-01 -3.82648289e-01 -1.04711509e+00
1.25656560e-01 -5.25449693e-01 -1.22594285e+00 -4.34424430e-01
-1.69290915e-01 8.86486396e-02 7.90563047e-01 9.66715395e-01
5.08976579e-01 6.25724852e-01 1.02759099e+00 -6.41615927e-01
-6.94172144e-01 -4.75866854e-01 -4.03555930e-01 5.54674923e-01
7.68406570e-01 -3.25825274e-01 -4.45083022e-01 -4.27456461e-02] | [7.663540363311768, 2.210428237915039] |
e247d3ff-e010-4575-8c9b-23364df64463 | understanding-and-mitigating-multi-sided | 2111.05564 | null | https://arxiv.org/abs/2111.05564v1 | https://arxiv.org/pdf/2111.05564v1.pdf | Understanding and Mitigating Multi-Sided Exposure Bias in Recommender Systems | Fairness is a critical system-level objective in recommender systems that has been the subject of extensive recent research. It is especially important in multi-sided recommendation platforms where it may be crucial to optimize utilities not just for the end user, but also for other actors such as item sellers or producers who desire a fair representation of their items. Existing solutions do not properly address various aspects of multi-sided fairness in recommendations as they may either solely have one-sided view (i.e. improving the fairness only for one side), or do not appropriately measure the fairness for each actor involved in the system. In this thesis, I aim at first investigating the impact of unfair recommendations on the system and how these unfair recommendations can negatively affect major actors in the system. Then, I seek to propose solutions to tackle the unfairness of recommendations. I propose a rating transformation technique that works as a pre-processing step before building the recommendation model to alleviate the inherent popularity bias in the input data and consequently to mitigate the exposure unfairness for items and suppliers in the recommendation lists. Also, as another solution, I propose a general graph-based solution that works as a post-processing approach after recommendation generation for mitigating the multi-sided exposure bias in the recommendation results. For evaluation, I introduce several metrics for measuring the exposure fairness for items and suppliers, and show that these metrics better capture the fairness properties in the recommendation results. I perform extensive experiments to evaluate the effectiveness of the proposed solutions. The experiments on different publicly-available datasets and comparison with various baselines confirm the superiority of the proposed solutions in improving the exposure fairness for items and suppliers. | ['Masoud Mansoury'] | 2021-11-10 | null | null | null | null | ['exposure-fairness'] | ['adversarial'] | [-1.99102640e-01 -3.63270119e-02 -3.06230485e-01 -4.51426804e-01
-9.00419354e-02 -6.98373079e-01 3.24479610e-01 3.03949416e-01
-3.11850518e-01 5.19668102e-01 4.58650082e-01 -4.69984740e-01
-6.69411480e-01 -9.88434196e-01 -2.89352477e-01 -5.33957243e-01
1.66832462e-01 1.76554680e-01 -8.96959379e-02 -5.46108425e-01
5.76722622e-01 4.59310383e-01 -1.38134789e+00 1.75338969e-01
1.09979129e+00 8.92719209e-01 -2.36996338e-01 3.50078374e-01
-5.34387305e-03 5.91658592e-01 -5.98280430e-01 -9.62006867e-01
7.04330146e-01 -4.48792487e-01 -3.70620728e-01 -1.48896411e-01
2.42085576e-01 -3.58601183e-01 -9.44104139e-03 1.13290691e+00
5.63784301e-01 4.13374811e-01 7.70113051e-01 -1.55815685e+00
-6.78814888e-01 8.75353098e-01 -8.01529586e-01 2.17019260e-01
2.38652468e-01 -2.10423276e-01 1.45673800e+00 -6.22543573e-01
2.33901158e-01 1.08651340e+00 3.17112774e-01 1.54112116e-01
-1.03354037e+00 -6.11556172e-01 3.61259967e-01 1.11405484e-01
-1.21229494e+00 -2.38834828e-01 4.91958022e-01 -2.91426808e-01
5.32325387e-01 8.10434937e-01 4.28645998e-01 4.37362164e-01
2.00838923e-01 3.99278462e-01 8.82461965e-01 -2.03696772e-01
1.79564148e-01 5.79098642e-01 4.90480065e-01 -8.14012159e-03
7.72575498e-01 2.02611446e-01 -2.34317541e-01 -2.94892550e-01
3.24891001e-01 2.44377747e-01 -1.52610838e-01 -2.19137713e-01
-5.90241551e-01 9.95989799e-01 5.91099024e-01 1.50791243e-01
-6.06818020e-01 -2.81708807e-01 2.64891773e-01 5.28873861e-01
6.82554007e-01 5.91446340e-01 -1.95668593e-01 1.73240140e-01
-9.95614350e-01 2.51285613e-01 8.26771379e-01 7.53384650e-01
4.08310354e-01 -6.46994710e-02 -6.18665278e-01 6.15468502e-01
5.29112756e-01 1.92143992e-01 1.08713187e-01 -7.41857529e-01
5.34077883e-01 7.46389508e-01 4.20751005e-01 -1.49145091e+00
-1.88617393e-01 -8.29893827e-01 -7.68217087e-01 3.23565036e-01
4.56860274e-01 -2.16989666e-01 -3.23924035e-01 1.62349892e+00
3.48592788e-01 -2.45065138e-01 -2.40745470e-01 1.31190443e+00
6.64090872e-01 6.80805683e-01 2.10398957e-01 -4.43387985e-01
1.07073116e+00 -1.15241957e+00 -7.35911608e-01 3.42545271e-01
3.00997853e-01 -9.92426634e-01 9.32375610e-01 3.42168927e-01
-1.18524337e+00 -3.55148435e-01 -8.04161727e-01 3.07827145e-01
-3.35151106e-01 -6.21414883e-03 4.01086271e-01 1.07091689e+00
-6.45037711e-01 7.04170763e-01 9.09342915e-02 -1.02390505e-01
1.02717370e-01 5.90341151e-01 6.01955764e-02 1.34711161e-01
-1.44696105e+00 1.03327382e+00 -1.59116700e-01 7.86575824e-02
-2.86213309e-01 -8.93526912e-01 -2.18262985e-01 5.98426819e-01
5.67252517e-01 -7.47630775e-01 8.59816492e-01 -1.16796422e+00
-1.22873580e+00 1.01664402e-01 4.06391352e-01 -1.46214128e-01
7.86564052e-01 -5.98578192e-02 -5.92055619e-01 -5.28991699e-01
-1.91245943e-01 -7.30056054e-05 6.54278219e-01 -1.47332740e+00
-9.91112947e-01 -4.84402359e-01 5.92189074e-01 6.07625544e-01
-5.64393163e-01 1.05848432e-01 -6.51417971e-02 -8.07158053e-01
-4.02655274e-01 -9.08063829e-01 -3.58530432e-01 -4.69718009e-01
-3.89795542e-01 -1.53905347e-01 4.01520014e-01 -4.97887403e-01
1.71241570e+00 -2.05801272e+00 -1.36439741e-01 8.01926315e-01
1.32519960e-01 4.83834475e-01 -1.46012694e-01 5.04214942e-01
5.43073080e-02 3.37164044e-01 4.11179632e-01 -7.46068880e-02
3.60517651e-01 9.48096346e-03 -1.31676629e-01 5.37208855e-01
-4.07478005e-01 5.04988611e-01 -7.07065046e-01 1.34973479e-02
2.57257104e-01 3.52305889e-01 -7.45374143e-01 2.96590537e-01
1.49578422e-01 2.04684287e-01 -3.10965598e-01 2.34053478e-01
9.04290080e-01 2.94544995e-02 2.02100217e-01 -3.64534646e-01
-2.05409855e-01 2.96274334e-01 -1.52028143e+00 9.00932848e-01
-4.91470426e-01 -2.03341722e-01 1.18076503e-01 -7.90755630e-01
7.94426739e-01 2.44998917e-01 5.48899889e-01 -7.59547830e-01
3.98239374e-01 9.29837078e-02 2.76717782e-01 -4.51366231e-02
1.01721883e+00 -1.93603858e-01 -5.00445552e-02 7.87265956e-01
-4.26505327e-01 4.18194860e-01 2.09169060e-01 4.74724382e-01
5.41756809e-01 -3.91230643e-01 3.49631757e-01 -4.45637107e-01
6.98351920e-01 -4.32750225e-01 4.33852643e-01 6.48303032e-01
-1.94011740e-02 2.66983747e-01 5.23825109e-01 -2.90805221e-01
-7.44022608e-01 -5.91349542e-01 2.35710800e-01 1.36476099e+00
5.26617527e-01 -3.22674662e-01 -4.66585338e-01 -8.50841880e-01
3.19793522e-01 1.09377170e+00 -5.42052567e-01 -3.80082816e-01
1.67967081e-01 -6.57841742e-01 1.24230087e-01 4.01942693e-02
6.07389286e-02 -5.52848876e-01 -3.44499707e-01 1.21105105e-01
-1.21719517e-01 -6.26139998e-01 -8.17289710e-01 -2.98627347e-01
-5.92044115e-01 -9.28008020e-01 -6.13420188e-01 7.38529712e-02
7.75032043e-01 9.25861299e-01 9.72263157e-01 4.18375760e-01
1.73083529e-01 1.29065171e-01 -6.01330638e-01 -3.87974173e-01
-2.27634922e-01 -4.32335138e-02 1.75746381e-01 3.00903976e-01
4.48711962e-02 -3.23242188e-01 -9.08490181e-01 9.12878633e-01
-8.99377406e-01 -3.67039859e-01 3.27622086e-01 4.81879115e-01
2.13213861e-01 4.08648431e-01 8.88377309e-01 -1.28883576e+00
1.27850366e+00 -9.19749618e-01 -5.48476517e-01 4.68101382e-01
-1.42115939e+00 -2.05808505e-01 9.68758881e-01 -1.09222636e-01
-1.24037397e+00 -4.23148602e-01 -4.24440242e-02 -6.79619312e-02
1.67025149e-01 4.13392574e-01 -2.51309574e-01 -1.79686755e-01
4.24003184e-01 -5.62518358e-01 -1.04421258e-01 -4.62054789e-01
7.27105141e-01 8.03155839e-01 -1.00533299e-01 -2.77248889e-01
6.91003859e-01 1.53931022e-01 -6.05413457e-03 -1.31236717e-01
-7.80240476e-01 -6.27054930e-01 -1.41129475e-02 -3.89941424e-01
2.72034049e-01 -6.38975441e-01 -8.99871945e-01 -1.23399161e-01
-7.20898926e-01 2.93218285e-01 -3.88945997e-01 4.57182467e-01
-2.49384083e-02 4.05494213e-01 -2.93427020e-01 -9.12063360e-01
-7.78465390e-01 -1.19555640e+00 2.53908277e-01 5.75372815e-01
-2.25429475e-01 -9.72948015e-01 1.30944140e-02 5.75762808e-01
7.72047997e-01 -2.46331319e-01 8.82245421e-01 -9.36824262e-01
-2.17509866e-01 -3.20178926e-01 -3.40742946e-01 3.52622509e-01
1.10090658e-01 1.60810292e-01 -4.97545451e-01 -4.78461653e-01
-2.59924322e-01 3.01709533e-01 3.88633698e-01 3.77881825e-01
7.38308132e-01 -6.44731939e-01 1.13736793e-01 1.15194485e-01
1.43312633e+00 -1.40331872e-02 6.00455284e-01 1.43322408e-01
4.87447500e-01 8.78070772e-01 1.06600904e+00 8.34527075e-01
5.32897592e-01 8.88011515e-01 7.57567286e-01 -3.10902804e-01
1.93449706e-02 -7.37611875e-02 2.66171157e-01 7.53854275e-01
-4.50801045e-01 -7.63164103e-01 -5.45936152e-02 2.63293386e-01
-2.05622506e+00 -9.82854187e-01 -5.30777037e-01 2.63812542e+00
1.59846276e-01 -2.05169275e-01 5.31341791e-01 2.36278504e-01
8.74122858e-01 -1.64258942e-01 -2.75360197e-01 -9.89688873e-01
8.60827044e-02 -1.62716106e-01 8.12938750e-01 4.88867700e-01
-6.24471247e-01 5.43946564e-01 5.95300293e+00 7.04214692e-01
-8.77155960e-01 -6.21983688e-03 6.31612897e-01 -3.38386297e-01
-7.36199915e-01 1.21389553e-01 -5.36227703e-01 6.04316175e-01
7.03695297e-01 -7.60748684e-01 6.33272886e-01 6.85997486e-01
6.63444459e-01 -1.08673550e-01 -7.71850407e-01 6.39632344e-01
-5.59311695e-02 -8.69853854e-01 2.93027926e-02 4.11953300e-01
9.07934546e-01 -4.70907688e-01 2.52117902e-01 8.19824934e-02
3.56579334e-01 -7.75165498e-01 7.07910538e-01 5.50110698e-01
1.35565132e-01 -1.06726134e+00 1.08298826e+00 3.37391704e-01
-8.87706995e-01 -2.33287558e-01 -5.71708798e-01 -2.88474619e-01
3.46101940e-01 8.19094539e-01 -3.42123151e-01 1.02158761e+00
3.65179539e-01 3.47365439e-01 -3.46696645e-01 1.24312949e+00
-1.69942051e-01 3.91443461e-01 -5.33826351e-02 -4.58962880e-02
1.93723664e-02 -6.50486171e-01 4.49957967e-01 9.20238256e-01
5.45763671e-01 2.82303751e-01 -1.64142072e-01 6.59907460e-01
-2.97503233e-01 7.18931258e-01 -3.68173152e-01 1.83289111e-01
4.33853656e-01 1.72955668e+00 -5.39631605e-01 -1.31373107e-01
-5.07062435e-01 4.81393456e-01 1.09716775e-02 2.07788616e-01
-1.00440502e+00 -1.33210436e-01 7.54477680e-01 2.71597743e-01
8.14586282e-02 3.06061029e-01 -4.43249583e-01 -9.38148737e-01
-3.22327018e-01 -1.09348643e+00 6.42301679e-01 -2.32900202e-01
-1.51910508e+00 2.34000817e-01 -2.66730189e-01 -1.32735908e+00
1.82007238e-01 -5.05727082e-02 -7.11506784e-01 9.42498267e-01
-1.57731307e+00 -8.46096933e-01 -1.94735840e-01 7.46000767e-01
9.90031809e-02 -1.96445093e-01 4.62295085e-01 8.29298079e-01
-6.43858433e-01 8.91751051e-01 1.87065557e-01 -7.03178883e-01
9.45200861e-01 -1.03715158e+00 -1.91912919e-01 1.00231278e+00
9.87652987e-02 8.82959902e-01 8.63706112e-01 -4.61238325e-01
-1.29676628e+00 -9.06131625e-01 8.53647292e-01 -9.34520289e-02
3.63903135e-01 6.02659509e-02 -5.75211823e-01 1.23418182e-01
2.52442390e-01 -3.66901428e-01 1.16210139e+00 3.12442362e-01
-1.49737597e-01 -3.99665534e-01 -1.36791372e+00 4.84188139e-01
6.89815640e-01 -6.24246076e-02 -2.64418185e-01 4.19733077e-02
2.98004299e-01 2.14929860e-02 -9.33169067e-01 2.24715650e-01
8.36986482e-01 -1.25736618e+00 8.50004196e-01 -7.41249979e-01
2.87120461e-01 -4.89169896e-01 -4.61794771e-02 -1.58628738e+00
-7.71836758e-01 -4.56214517e-01 2.52331477e-02 1.49673128e+00
6.15735888e-01 -5.83129525e-01 5.31951070e-01 1.20551097e+00
1.86143771e-01 -4.70486820e-01 -4.98014450e-01 -6.71508789e-01
-1.48067325e-01 -3.72301675e-02 8.87408853e-01 9.12716269e-01
1.14539757e-01 4.08613622e-01 -9.02950108e-01 9.80970114e-02
5.26379228e-01 4.56420273e-01 8.20963264e-01 -9.72697616e-01
-1.95555553e-01 -6.17263496e-01 -2.08809823e-02 -6.30495787e-01
-2.02936575e-01 -8.12975347e-01 -3.25971127e-01 -1.65184748e+00
2.67514110e-01 -7.74692774e-01 -8.79755378e-01 2.00658873e-01
-3.47735047e-01 2.39463776e-01 5.39748907e-01 2.43375510e-01
-5.81663966e-01 3.08448046e-01 1.26657689e+00 2.19016463e-01
-2.03415975e-01 6.26740038e-01 -1.53191757e+00 3.57802331e-01
7.00368047e-01 -6.17470384e-01 -6.82607472e-01 -2.00109467e-01
7.46444404e-01 -5.07982112e-02 -6.62110969e-02 -4.04688269e-01
2.56297350e-01 -4.94810641e-01 -2.09779531e-01 -2.41770640e-01
-1.89262360e-01 -1.43096948e+00 5.42153537e-01 2.89008945e-01
-4.83744532e-01 1.02279186e-01 -3.50913435e-01 5.71436107e-01
-2.62800436e-02 -3.38857830e-01 6.89734757e-01 1.02743559e-01
-1.65146828e-01 3.36384863e-01 -1.77534938e-01 -3.46714944e-01
1.24732554e+00 4.25253548e-02 -5.44019639e-01 -8.26425254e-01
-4.08109128e-01 4.09403533e-01 4.97087747e-01 5.70393920e-01
1.72603086e-01 -1.23641992e+00 -7.14501917e-01 -4.01558936e-01
3.93907242e-02 -8.72968078e-01 3.92638385e-01 9.39165354e-01
-1.65723398e-01 2.23455369e-01 -4.11479056e-01 3.33428979e-01
-1.40169251e+00 6.86332524e-01 8.99864361e-02 -5.05228817e-01
1.82090774e-01 6.67015851e-01 8.97793025e-02 -3.01881731e-01
1.84439465e-01 9.52714607e-02 -6.81564808e-01 3.47415954e-01
3.10312331e-01 9.32181001e-01 8.60669091e-02 -8.57681692e-01
-3.51512164e-01 2.40112394e-01 8.99999961e-02 2.06781790e-01
1.34590364e+00 -6.62502110e-01 -1.74268275e-01 -8.67265090e-02
6.84693456e-01 5.77379286e-01 -6.58575416e-01 -4.09295000e-02
-2.04561278e-01 -1.00084794e+00 2.74173707e-01 -9.97141063e-01
-1.49651027e+00 5.59519053e-01 3.49988431e-01 7.39580750e-01
1.22545862e+00 -6.05434775e-01 6.65265799e-01 -2.23890990e-01
2.75154054e-01 -1.25760996e+00 -2.98820734e-01 3.76730002e-02
7.06676900e-01 -1.03151500e+00 4.71878380e-01 -5.38131654e-01
-9.73628759e-01 9.54793394e-01 4.92617548e-01 -5.69545627e-02
7.01772630e-01 -1.23935357e-01 1.09337099e-01 -2.62878346e-03
-6.16854548e-01 -8.72112364e-02 6.07479274e-01 8.04191753e-02
6.88313663e-01 4.33742791e-01 -1.25032365e+00 1.13715410e+00
1.13015641e-02 -1.03712574e-01 6.74530387e-01 4.62871522e-01
-3.01687896e-01 -1.49916303e+00 -1.32167131e-01 7.35980749e-01
-6.62248075e-01 1.20363459e-02 -4.41354275e-01 3.98357540e-01
2.63718218e-01 1.58555114e+00 -4.18204844e-01 -6.87652588e-01
8.11515629e-01 -4.95752484e-01 -6.63743094e-02 -3.69913876e-01
-1.23900115e+00 1.14457026e-01 3.14069897e-01 -5.16278923e-01
-4.39970016e-01 -4.37349677e-01 -8.20585072e-01 -9.10104156e-01
-8.83834183e-01 4.72926110e-01 8.21280003e-01 7.71428347e-01
5.42411983e-01 5.81094980e-01 1.27966821e+00 -2.93641180e-01
-8.50286186e-01 -6.03072286e-01 -9.75059986e-01 7.99321115e-01
-2.05676273e-01 -6.48308694e-01 -4.60285544e-01 -5.47019362e-01] | [9.695348739624023, 5.644174575805664] |
5e8f746f-4d8e-47b2-8373-be3fd9625045 | eeny-meeny-miny-moe-how-to-choose-data-for | 2210.14465 | null | https://arxiv.org/abs/2210.14465v1 | https://arxiv.org/pdf/2210.14465v1.pdf | Eeny, meeny, miny, moe. How to choose data for morphological inflection | Data scarcity is a widespread problem in numerous natural language processing (NLP) tasks for low-resource languages. Within morphology, the labour-intensive work of tagging/glossing data is a serious bottleneck for both NLP and language documentation. Active learning (AL) aims to reduce the cost of data annotation by selecting data that is most informative for improving the model. In this paper, we explore four sampling strategies for the task of morphological inflection using a Transformer model: a pair of oracle experiments where data is chosen based on whether the model already can or cannot inflect the test forms correctly, as well as strategies based on high/low model confidence, entropy, as well as random selection. We investigate the robustness of each strategy across 30 typologically diverse languages. We also perform a more in-depth case study of Nat\"ugu. Our results show a clear benefit to selecting data based on model confidence and entropy. Unsurprisingly, the oracle experiment, where only incorrectly handled forms are chosen for further training, which is presented as a proxy for linguist/language consultant feedback, shows the most improvement. This is followed closely by choosing low-confidence and high-entropy predictions. We also show that despite the conventional wisdom of larger data sets yielding better accuracy, introducing more instances of high-confidence or low-entropy forms, or forms that the model can already inflect correctly, can reduce model performance. | ['Mans Hulden', 'Saliha Muradoglu'] | 2022-10-26 | null | null | null | null | ['morphological-inflection'] | ['natural-language-processing'] | [ 2.77299523e-01 2.73735195e-01 -2.34907046e-01 -3.62442017e-01
-1.23752713e+00 -8.94263923e-01 4.53322887e-01 6.66930795e-01
-8.54193747e-01 7.58501291e-01 4.22415942e-01 -7.37495780e-01
-1.56191081e-01 -5.18320382e-01 -5.25505126e-01 -6.38754189e-01
2.23863780e-01 8.07857096e-01 1.11274784e-02 -5.26802801e-02
2.54914701e-01 2.40028381e-01 -1.17072952e+00 2.13703945e-01
1.22316968e+00 3.94916415e-01 3.38226169e-01 1.78566754e-01
-2.77668864e-01 5.70616782e-01 -6.24994993e-01 -9.35646772e-01
2.89135426e-01 -1.52057678e-01 -8.82166862e-01 -1.24571338e-01
3.67706299e-01 1.90120786e-01 5.06217837e-01 9.55779910e-01
5.85069537e-01 -7.36266077e-02 6.28623962e-01 -6.52187228e-01
-5.05698502e-01 1.00705576e+00 -2.09213361e-01 2.03864992e-01
4.06940818e-01 3.36076915e-01 1.30312777e+00 -8.26538444e-01
7.45424271e-01 1.27114511e+00 6.90288663e-01 4.31786060e-01
-1.50436783e+00 -4.19173688e-01 2.08579198e-01 -6.66632652e-02
-1.25579858e+00 -8.35405231e-01 5.09676456e-01 -4.79190499e-01
1.11696732e+00 3.49936098e-01 3.45113337e-01 6.60908759e-01
-7.37218857e-02 7.14275956e-01 1.13712251e+00 -9.31066632e-01
2.78626800e-01 5.39189100e-01 1.30881429e-01 3.72700244e-01
5.51237762e-01 4.73275706e-02 -5.09951830e-01 -5.55544794e-01
2.06138641e-01 -6.96191847e-01 -2.76094645e-01 -6.19602464e-02
-8.61786008e-01 7.50537634e-01 -1.17544442e-01 4.03010279e-01
-2.99302489e-01 -4.43856120e-01 2.04297975e-01 2.67177731e-01
5.98196149e-01 1.08866739e+00 -1.01743770e+00 -2.43181363e-01
-9.41414714e-01 6.36633262e-02 1.11807930e+00 6.91481471e-01
7.06823587e-01 9.34720854e-04 4.52467315e-02 1.30303729e+00
3.73830646e-01 2.41761580e-01 4.85099375e-01 -7.49856114e-01
6.76178098e-01 7.70093501e-01 1.78670466e-01 -5.12913704e-01
-3.25860143e-01 -2.96748579e-01 -1.73475236e-01 7.78905675e-02
7.67500877e-01 -2.20620111e-01 -8.29148591e-01 1.80306113e+00
1.65198326e-01 -6.78984404e-01 1.03078783e-02 5.59680700e-01
3.75309378e-01 4.84285980e-01 4.14680123e-01 -7.80762851e-01
1.28499651e+00 -3.75087261e-01 -6.75043106e-01 -5.89094162e-01
9.55061078e-01 -1.01071441e+00 1.38593340e+00 5.23188591e-01
-1.09472609e+00 -2.04191312e-01 -7.37860620e-01 -1.17246658e-01
-3.68254930e-01 -7.94751793e-02 6.24083936e-01 6.91901088e-01
-9.59426224e-01 6.13505542e-01 -8.26475441e-01 -2.93212563e-01
1.92929372e-01 4.34879780e-01 -3.14548612e-01 6.34560287e-02
-1.03819466e+00 1.26317394e+00 5.58766663e-01 1.62603632e-02
-7.09458962e-02 -5.82335413e-01 -9.45232630e-01 -4.00730185e-02
4.35391068e-01 -1.97032496e-01 1.17144322e+00 -1.17515671e+00
-1.10357285e+00 9.95171428e-01 -1.10977083e-01 -4.75097865e-01
3.60983372e-01 -3.35782081e-01 -3.19791257e-01 -5.40494382e-01
8.35280269e-02 5.56449175e-01 3.61989975e-01 -1.31195152e+00
-4.75648463e-01 -3.61692399e-01 -9.26229879e-02 3.97533417e-01
-2.72569180e-01 4.57712054e-01 -2.09688872e-01 -5.13405561e-01
7.16311932e-02 -9.14723277e-01 -1.64757490e-01 -5.76105952e-01
-2.48808831e-01 -6.22762978e-01 1.22302718e-01 -9.62927699e-01
1.67012107e+00 -2.13665104e+00 -5.46691380e-02 2.05020234e-01
-1.29526675e-01 1.88998520e-01 2.29317725e-01 3.49840224e-01
1.86170209e-02 6.13927901e-01 -4.61964935e-01 -2.55579472e-01
7.07086474e-02 4.00141507e-01 -1.92213356e-01 1.42839566e-01
4.81651843e-01 5.60553789e-01 -8.73028636e-01 -6.84702754e-01
-6.08559102e-02 1.36722356e-01 -6.05680943e-01 -3.62096019e-02
-2.28122026e-01 1.62756413e-01 1.32125139e-01 7.11587906e-01
2.89271951e-01 -2.60449555e-02 5.13282478e-01 1.85783923e-01
-4.94680732e-01 9.13746119e-01 -1.09581232e+00 1.12447596e+00
-6.25702322e-01 3.44046026e-01 1.92474239e-02 -4.71940249e-01
9.13445771e-01 3.08409810e-01 5.03363870e-02 -5.41097403e-01
-2.47187063e-01 7.23210931e-01 6.95652366e-01 -4.13863331e-01
5.37742853e-01 -2.65563190e-01 -1.96472868e-01 3.52439612e-01
-1.25781912e-02 -3.96456182e-01 5.09933352e-01 4.24196683e-02
1.03046596e+00 2.61542678e-01 6.63307726e-01 -5.11004984e-01
2.42966235e-01 2.22850561e-01 9.56899941e-01 7.73031771e-01
-1.46516487e-01 4.28615689e-01 5.25825620e-01 -1.20291285e-01
-1.07220840e+00 -8.17799568e-01 -4.67182368e-01 1.29704857e+00
-3.80601406e-01 -5.49581110e-01 -4.48299766e-01 -8.05868030e-01
-1.28638878e-01 1.43363082e+00 -3.02494496e-01 9.34295263e-03
-8.15571606e-01 -1.06201065e+00 2.97480851e-01 3.23788315e-01
-8.67723376e-02 -1.32865143e+00 -3.62248480e-01 2.27027521e-01
-2.04080060e-01 -6.23701334e-01 -2.69776672e-01 7.58987606e-01
-7.64842749e-01 -7.84146965e-01 -2.02985644e-01 -7.64976203e-01
5.76856732e-01 -4.21357721e-01 1.47188592e+00 3.20908904e-01
2.77556360e-01 3.35534550e-02 -5.25783181e-01 -6.94488943e-01
-7.88139641e-01 2.38453567e-01 -3.93775702e-02 -3.97628725e-01
6.31329060e-01 -1.62153721e-01 -1.00763328e-01 3.54453996e-02
-7.51276612e-01 -1.89679176e-01 6.50085270e-01 8.96968961e-01
6.11500084e-01 -1.46688446e-01 5.00656843e-01 -1.32303762e+00
5.83817482e-01 -4.92875844e-01 -4.91349071e-01 5.31692147e-01
-8.89101446e-01 2.15697661e-01 6.63788021e-01 -3.25701088e-01
-1.12288439e+00 2.45533630e-01 -3.26994985e-01 3.92363697e-01
-1.16137013e-01 6.62216902e-01 -3.92317027e-01 3.24339479e-01
8.83300483e-01 -2.07398832e-01 -2.75271058e-01 -7.64540553e-01
-8.44338238e-02 7.85365522e-01 1.48889989e-01 -7.26886034e-01
5.93750894e-01 -1.75123826e-01 -6.03258252e-01 -7.89169967e-01
-1.06084919e+00 -4.61620331e-01 -7.71364093e-01 1.32006720e-01
4.31343675e-01 -4.91398275e-01 -1.02588654e-01 8.84396508e-02
-1.00928712e+00 -5.55339098e-01 -3.76685500e-01 5.39084315e-01
-4.18774068e-01 2.84947246e-01 -4.16834801e-01 -1.01765192e+00
-2.77908444e-01 -1.21362936e+00 9.66716111e-01 3.45867127e-02
-7.82617152e-01 -1.17864525e+00 1.30256236e-01 2.81003028e-01
2.34539509e-01 -7.01621398e-02 1.48012865e+00 -1.28656518e+00
-1.55344129e-01 -2.74515413e-02 3.52567762e-01 3.20370555e-01
1.69887841e-01 1.88761443e-01 -8.59901428e-01 -2.50394225e-01
4.54998687e-02 -4.13428366e-01 6.74211204e-01 1.51479408e-01
7.61245489e-01 -4.78862911e-01 -4.60357852e-02 2.47211725e-01
1.44665694e+00 3.48158538e-01 4.73327577e-01 4.49505687e-01
3.90658498e-01 8.58317077e-01 6.12131536e-01 5.77362105e-02
3.90143126e-01 6.70841813e-01 -5.29829264e-02 1.62587110e-02
6.46045953e-02 -2.37648144e-01 5.11785865e-01 1.07680678e+00
1.37409747e-01 -3.80472749e-01 -1.50662196e+00 5.82766414e-01
-1.47799230e+00 -7.06348181e-01 -8.07015076e-02 2.67065620e+00
1.44618464e+00 5.38459241e-01 1.21200293e-01 4.62885112e-01
5.66036999e-01 -1.15238585e-01 -2.40536958e-01 -7.65474856e-01
-2.81427860e-01 2.50483334e-01 4.35211778e-01 8.53660822e-01
-9.68088150e-01 1.08134508e+00 5.90982533e+00 8.95425260e-01
-9.27463412e-01 -5.69469482e-02 6.90192342e-01 7.18450695e-02
-5.21041214e-01 3.19610596e-01 -9.33116674e-01 5.42693913e-01
1.05983818e+00 -1.71320438e-01 3.19635987e-01 7.19751596e-01
1.59373105e-01 -4.14786071e-01 -1.14856505e+00 4.14483368e-01
-1.00408472e-01 -8.40696335e-01 -3.86879481e-02 1.17847733e-01
4.35976803e-01 2.38491781e-02 -2.86401838e-01 4.19576287e-01
6.52051508e-01 -7.61327803e-01 9.57819223e-01 2.07986355e-01
8.33604097e-01 -5.86603224e-01 9.18916285e-01 6.70503497e-01
-6.45909131e-01 -1.41839728e-01 -3.67523313e-01 -1.05023451e-01
1.48730576e-01 5.92659712e-01 -1.16155672e+00 2.37227753e-01
4.96191442e-01 1.52772695e-01 -7.43536413e-01 1.05873716e+00
-3.40397418e-01 1.25664282e+00 -5.51288188e-01 -9.56371352e-02
1.65890902e-02 -1.39643297e-01 5.50931513e-01 1.41555035e+00
-3.61459120e-03 6.94191232e-02 1.66343644e-01 5.12850225e-01
1.44735575e-01 5.72359204e-01 -6.18187308e-01 -1.01472452e-01
7.61746347e-01 1.08268499e+00 -8.33488166e-01 -3.00663263e-01
-3.53240073e-01 4.74310458e-01 6.17020845e-01 5.92030808e-02
-2.02327073e-01 -2.01214924e-01 6.58279508e-02 2.35322744e-01
-3.53367850e-02 -1.92089051e-01 -6.72435999e-01 -9.26884532e-01
8.51760339e-03 -9.83891606e-01 5.58925688e-01 -4.87397313e-01
-1.34110200e+00 5.48745990e-01 -5.18703498e-02 -8.64854097e-01
-4.57572818e-01 -5.86696982e-01 -2.73937553e-01 9.58003998e-01
-1.11049378e+00 -7.13912129e-01 2.79216081e-01 -2.06550620e-02
5.44700027e-01 1.13771237e-01 9.02371824e-01 2.56286263e-01
-4.94730413e-01 6.52150035e-01 -3.04275677e-02 2.00460941e-01
7.94434071e-01 -1.66355157e+00 3.74683529e-01 7.94920266e-01
4.62661415e-01 9.05724823e-01 7.26090193e-01 -9.84222174e-01
-9.71842229e-01 -6.90888107e-01 1.62913561e+00 -7.07106352e-01
5.76460481e-01 -4.20689791e-01 -1.17864668e+00 6.30724907e-01
3.65841910e-02 -4.89549845e-01 9.34905887e-01 5.75767994e-01
-1.52573526e-01 9.11304578e-02 -1.18411660e+00 6.77540421e-01
1.01126921e+00 -2.57707804e-01 -8.90590250e-01 3.68470252e-01
5.96555054e-01 -1.71900928e-01 -7.19520628e-01 4.87078667e-01
3.34755033e-01 -7.01417983e-01 3.34793448e-01 -5.39778292e-01
2.16949806e-01 2.56359056e-02 -1.07904218e-01 -1.46327698e+00
-4.69214797e-01 -6.17238760e-01 3.80175471e-01 1.54739141e+00
1.19318068e+00 -4.39181030e-01 5.35770655e-01 8.80711436e-01
-2.41333961e-01 -8.28789473e-01 -8.59104156e-01 -6.71524942e-01
3.40802580e-01 -5.89804888e-01 2.85581648e-01 1.04919875e+00
9.41215605e-02 4.53921318e-01 1.35559142e-01 -1.14718169e-01
2.68814564e-01 -1.54529095e-01 3.61550331e-01 -1.27605903e+00
-2.77514309e-01 -4.13456917e-01 -8.55391696e-02 -5.60399890e-01
3.21247429e-02 -9.81195569e-01 3.92767161e-01 -1.43331957e+00
1.06039554e-01 -8.00397336e-01 -1.10226013e-01 6.44921482e-01
-4.52516407e-01 -3.49714160e-02 2.28085220e-01 3.24331224e-01
-3.67660791e-01 -1.40488252e-01 6.95389211e-01 2.77261019e-01
-5.91297030e-01 6.17239326e-02 -7.78788209e-01 9.48734999e-01
7.35012174e-01 -6.56816483e-01 -1.60226017e-01 -6.43961906e-01
7.21335411e-01 -7.45174959e-02 -1.11637510e-01 -4.58588064e-01
4.84420434e-02 -2.57065177e-01 2.70933002e-01 -3.62883955e-01
8.35019499e-02 -7.05621600e-01 5.70353270e-02 3.02221507e-01
-5.06058156e-01 3.45198750e-01 2.25726396e-01 -3.43000633e-03
6.74947426e-02 -6.22638583e-01 6.83860421e-01 -4.15983766e-01
-4.16163564e-01 -1.14804059e-01 -2.23785236e-01 4.33405131e-01
3.87061357e-01 -2.93661922e-01 -1.14743643e-01 -3.39784101e-02
-7.35589504e-01 -1.55347819e-02 6.62251711e-01 2.71087706e-01
1.25469893e-01 -8.60354841e-01 -8.88563216e-01 1.74277008e-01
1.21048346e-01 -5.45020401e-03 -4.11037087e-01 8.63085091e-01
-3.64658624e-01 3.42212230e-01 2.89676934e-01 -3.41068983e-01
-1.10202456e+00 2.41211817e-01 2.68101543e-01 -4.81096268e-01
-8.73246342e-02 1.02196836e+00 -8.82626772e-02 -6.05805874e-01
1.54786870e-01 -1.96688354e-01 -2.45925695e-01 3.37108195e-01
1.09161004e-01 2.66598046e-01 5.18912435e-01 -5.77551603e-01
-4.04714853e-01 9.46604535e-02 -3.23036045e-01 -3.74821514e-01
1.34715486e+00 -1.23897344e-02 -1.54224932e-01 8.94853592e-01
5.53744555e-01 6.56107664e-01 -1.00162578e+00 -2.17235535e-01
8.89463723e-01 -3.14349383e-01 -6.63056970e-02 -1.29071605e+00
-4.64007169e-01 5.97436070e-01 2.07777888e-01 3.92952472e-01
7.18906999e-01 7.68714468e-04 2.79605955e-01 3.18224400e-01
4.96891558e-01 -1.41569948e+00 -3.96118522e-01 5.73946655e-01
9.01369750e-01 -1.30441284e+00 1.35310292e-01 -2.42329329e-01
-7.92576551e-01 7.59652615e-01 7.28521347e-01 2.50735909e-01
3.60747546e-01 4.32727545e-01 2.16410264e-01 -8.83186385e-02
-1.03574848e+00 -2.46221334e-01 1.94938540e-01 3.13716739e-01
8.62860262e-01 2.22451374e-01 -7.60258436e-01 6.75273657e-01
-6.61715031e-01 -4.63249087e-01 3.35443139e-01 8.33248198e-01
-4.16609526e-01 -1.23507094e+00 -3.87574732e-01 7.89136291e-01
-6.56288207e-01 -6.31118536e-01 -8.93747091e-01 9.09841299e-01
4.14385587e-01 9.12240028e-01 2.87309974e-01 -8.95711407e-03
1.90786704e-01 6.02410316e-01 3.72238308e-01 -1.14115012e+00
-9.42399979e-01 1.44491285e-01 6.16889775e-01 3.08613703e-02
-1.40704945e-01 -1.08297586e+00 -1.13719702e+00 -1.23641305e-01
-8.62275541e-01 3.78183097e-01 6.00969315e-01 1.04778898e+00
5.55274934e-02 -2.20774963e-01 3.06805968e-01 -1.91498205e-01
-9.46955800e-01 -1.27382350e+00 -2.37873077e-01 4.76293296e-01
-1.14647131e-02 -5.01400232e-01 -7.38738716e-01 3.21443342e-02] | [10.758749961853027, 9.466996192932129] |
b012529a-52b4-499a-971f-1a117cef7e99 | robust-counterfactual-inferences-using | 1808.07569 | null | http://arxiv.org/abs/1808.07569v1 | http://arxiv.org/pdf/1808.07569v1.pdf | Robust Counterfactual Inferences using Feature Learning and their Applications | In a wide variety of applications, including personalization, we want to
measure the difference in outcome due to an intervention and thus have to deal
with counterfactual inference. The feedback from a customer in any of these
situations is only 'bandit feedback' - that is, a partial feedback based on
whether we chose to intervene or not. Typically randomized experiments are
carried out to understand whether an intervention is overall better than no
intervention. Here we present a feature learning algorithm to learn from a
randomized experiment where the intervention in consideration is most effective
and where it is least effective rather than only focusing on the overall
impact, thus adding a context to our learning mechanism and extract more
information. From the randomized experiment, we learn the feature
representations which divide the population into subpopulations where we
observe statistically significant difference in average customer feedback
between those who were subjected to the intervention and those who were not,
with a level of significance l, where l is a configurable parameter in our
model. We use this information to derive the value of the intervention in
consideration for each instance in the population. With experiments, we show
that using this additional learning, in future interventions, the context for
each instance could be leveraged to decide whether to intervene or not. | ['Abhimanyu Mitra', 'Sushant Kumar', 'Kannan Achan'] | 2018-08-22 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [ 4.62089151e-01 2.09004238e-01 -7.76381016e-01 -3.76206249e-01
-5.90437829e-01 -4.89441901e-01 5.57173133e-01 5.85798025e-01
-7.78246522e-01 9.18771565e-01 5.67247093e-01 -7.52532244e-01
-4.39769298e-01 -8.93754363e-01 -9.30403829e-01 -8.57185841e-01
-2.49109487e-03 5.80751777e-01 -1.78370833e-01 -4.77806143e-02
4.82927173e-01 2.15084642e-01 -1.44399762e+00 4.29827094e-01
6.07086778e-01 8.38110924e-01 -2.02525146e-02 5.91512322e-01
1.09599486e-01 4.68023419e-01 -5.97556949e-01 -1.60500541e-01
3.40445727e-01 -7.34669268e-01 -7.09583342e-01 1.90417439e-01
1.55864879e-01 -4.42083061e-01 1.66355297e-01 6.93968594e-01
4.35743034e-01 2.21543640e-01 9.21317816e-01 -1.03228474e+00
-2.15009525e-01 1.10671961e+00 -5.96014500e-01 2.82440424e-01
5.47793090e-01 4.43502337e-01 1.30080307e+00 -3.13514173e-02
5.22136509e-01 1.30190885e+00 3.51977766e-01 3.35929036e-01
-1.81766200e+00 -6.96262240e-01 5.13354242e-01 1.00356765e-01
-6.63089275e-01 -4.17755395e-01 8.41747165e-01 -4.90728080e-01
4.94809330e-01 2.92780757e-01 8.46127212e-01 9.26125407e-01
1.72427952e-01 6.14769578e-01 1.25269854e+00 -6.29439652e-01
6.87305868e-01 3.58302712e-01 -9.36140865e-03 2.65809149e-01
5.34211159e-01 5.51753998e-01 -2.15596706e-01 -3.98867369e-01
3.87668580e-01 2.33922780e-01 -3.28646630e-01 -4.00726676e-01
-8.51260245e-01 1.21584773e+00 3.26837689e-01 1.03009358e-01
-8.02759945e-01 6.14665449e-02 4.24450725e-01 4.37416524e-01
2.40675718e-01 7.79931307e-01 -7.45840013e-01 -1.96409076e-02
-7.16255307e-01 5.29761910e-01 7.53411949e-01 1.49988949e-01
1.03964412e+00 -4.62089092e-01 -3.39140266e-01 5.38384974e-01
-1.13561183e-01 1.79669321e-01 5.19513428e-01 -8.87841463e-01
4.61556196e-01 4.98615324e-01 5.34416735e-01 -6.32388055e-01
-2.85141438e-01 -1.12599088e-02 -2.55341083e-01 2.28560433e-01
5.48874438e-01 -8.15369606e-01 -8.10434163e-01 2.01300788e+00
2.95380920e-01 2.57724643e-01 -2.04935104e-01 7.80401945e-01
-4.83562984e-02 4.52865422e-01 1.70942560e-01 -6.57363832e-01
1.03660882e+00 -1.21813551e-01 -4.85462278e-01 -3.16374272e-01
9.53833222e-01 -4.15467471e-01 1.11226428e+00 3.74874830e-01
-7.09451795e-01 -1.75019830e-01 -9.92274106e-01 7.61677563e-01
-7.55701587e-02 -3.93720627e-01 7.68184423e-01 6.29800022e-01
-6.63921595e-01 8.69028866e-01 -3.82968873e-01 -2.77027547e-01
3.67221892e-01 5.35054743e-01 -1.61434650e-01 1.11464068e-01
-1.23859429e+00 7.23877311e-01 4.49468046e-01 -3.99013400e-01
-5.06649315e-01 -8.53885651e-01 -6.00775659e-01 4.74352986e-01
8.10215950e-01 -8.24635208e-01 1.26239347e+00 -1.30863643e+00
-1.22728169e+00 5.10374129e-01 -1.23704635e-01 -8.51585805e-01
5.51853299e-01 2.60036975e-01 -6.35762289e-02 -2.52946705e-01
2.11768404e-01 3.74882996e-01 7.83460259e-01 -1.08855462e+00
-1.15074885e+00 -6.80714309e-01 4.32918519e-01 1.90484673e-01
4.51646782e-02 -5.66563345e-02 1.85548529e-01 -3.12793225e-01
-3.79067719e-01 -1.17414117e+00 -4.08161730e-01 -4.64062393e-01
-3.38859320e-01 -2.77545005e-01 5.02929211e-01 -2.60953248e-01
1.19278336e+00 -1.91608906e+00 -1.68977484e-01 3.03509891e-01
-3.19988765e-02 -1.84228480e-01 7.27495402e-02 4.88057882e-01
-1.20721564e-01 3.85734528e-01 -1.37257367e-01 1.89556509e-01
-4.22468372e-02 -1.82370581e-02 -3.23885053e-01 5.24049938e-01
1.22103073e-01 3.22155923e-01 -6.95184231e-01 -2.01075912e-01
1.91329092e-01 -1.51378840e-01 -1.04203749e+00 3.03089470e-01
-3.13806325e-01 4.84988898e-01 -7.16138363e-01 -2.05024164e-02
4.16283965e-01 2.07241885e-02 4.08748031e-01 1.45382732e-01
-7.92912096e-02 5.48507214e-01 -1.36686194e+00 9.19245005e-01
-7.59325981e-01 4.68602508e-01 -4.88913953e-02 -1.40015602e+00
5.18776715e-01 1.54882669e-01 5.15462697e-01 -4.57400143e-01
3.23931098e-01 5.22384886e-03 3.38190764e-01 -4.77516413e-01
7.65121207e-02 -6.54297829e-01 -2.18920976e-01 7.08530664e-01
-3.13797921e-01 2.08269805e-01 1.07744291e-01 -1.11454636e-01
1.14462614e+00 -3.34836364e-01 6.60562754e-01 -2.72545427e-01
2.46848360e-01 -2.12122247e-01 8.51898432e-01 1.13659430e+00
-1.26811698e-01 1.34646162e-01 8.91066968e-01 -2.76884049e-01
-8.14978063e-01 -8.00931871e-01 -5.37163466e-02 1.14591384e+00
-1.55045748e-01 1.85113102e-01 -4.64439124e-01 -8.68869007e-01
4.83714342e-01 1.12791550e+00 -1.01432407e+00 -4.49433237e-01
-3.54627430e-01 -6.33652151e-01 -3.41677427e-01 3.14089447e-01
3.29438955e-01 -1.16922009e+00 -8.26399148e-01 1.25987723e-01
1.36362240e-01 -2.70547569e-01 -6.29063427e-01 3.53189617e-01
-8.72794032e-01 -1.25782311e+00 -2.29600444e-01 -2.31883436e-01
5.20061553e-01 -2.63422839e-02 9.12043214e-01 -2.37706497e-01
1.14100173e-01 2.27355048e-01 -2.54284501e-01 -5.83144307e-01
-2.55013078e-01 -1.52078226e-01 7.37015009e-02 2.04342604e-01
3.74616146e-01 -4.94710028e-01 -8.61922622e-01 8.60024542e-02
-6.98799253e-01 -3.76516938e-01 6.08253717e-01 1.00556326e+00
2.98261661e-02 4.15995628e-01 8.68143797e-01 -1.62016857e+00
8.11436296e-01 -6.63363695e-01 -5.80856144e-01 1.69376925e-01
-7.41047859e-01 4.21946585e-01 7.27413595e-01 -7.83956587e-01
-1.07826018e+00 -1.82999611e-01 1.48478150e-01 -6.30703568e-02
-2.31797308e-01 7.27150321e-01 -4.62696701e-01 6.61413372e-01
6.91501379e-01 -2.99983114e-01 -2.65871678e-02 -3.96338493e-01
3.15508306e-01 7.40411997e-01 5.54991923e-02 -8.27566445e-01
1.22191347e-01 1.68244332e-01 -1.81002572e-01 -5.04166603e-01
-7.24059820e-01 -2.17313230e-01 -2.19364852e-01 3.03650033e-02
4.30129439e-01 -4.83945638e-01 -1.18485785e+00 -1.83178335e-01
-5.59688568e-01 -6.06588125e-01 -4.95920807e-01 6.96059287e-01
-7.71781266e-01 -2.90730000e-01 6.64466321e-02 -1.04161382e+00
1.75904781e-01 -1.17329240e+00 4.79748100e-01 2.74406940e-01
-3.95137846e-01 -1.18424475e+00 2.06032284e-02 2.24508807e-01
1.71062946e-01 1.76908702e-01 1.25035191e+00 -1.07127726e+00
-3.15141618e-01 -3.99067730e-01 2.13476896e-01 1.02125205e-01
5.70735455e-01 -2.15198323e-01 -6.50869370e-01 -5.60988367e-01
1.16746575e-01 -2.90023126e-02 8.09394658e-01 9.13646996e-01
1.10251987e+00 -9.70351756e-01 -5.17987907e-01 1.06142879e-01
1.30974126e+00 5.72565079e-01 3.78078967e-01 2.97863394e-01
2.29146436e-01 8.03486168e-01 7.59707987e-01 7.59133816e-01
1.72382325e-01 8.13066721e-01 3.30749929e-01 2.30956301e-01
5.11272490e-01 -3.95126462e-01 3.02711725e-01 -3.70125622e-01
2.71073222e-01 -9.11280364e-02 -4.26543504e-01 5.56125700e-01
-1.79698288e+00 -1.14315045e+00 5.31707168e-01 3.03243113e+00
8.59550536e-01 4.68956947e-01 6.53795123e-01 7.95630813e-02
8.70665193e-01 2.45018378e-02 -8.43318999e-01 -8.41432273e-01
3.66832495e-01 5.39542176e-04 6.93062901e-01 6.10168099e-01
-8.98471713e-01 3.31283689e-01 6.15993643e+00 5.25541604e-01
-1.31427610e+00 -4.68981206e-01 9.92228091e-01 -2.17309460e-01
-4.96968836e-01 3.83779824e-01 -7.18469083e-01 6.51427507e-01
1.36584628e+00 -5.67063570e-01 5.36770761e-01 5.01162469e-01
6.27954066e-01 -4.82096404e-01 -1.58213055e+00 2.63055623e-01
-4.14987475e-01 -1.15799868e+00 -4.80245501e-02 4.12248671e-01
5.92537642e-01 -5.55211663e-01 1.05923191e-01 5.25747359e-01
7.30814397e-01 -7.64772117e-01 5.77678978e-01 3.79412025e-01
4.26009566e-01 -1.04663157e+00 8.89987707e-01 7.36364782e-01
-4.43000644e-01 -6.72613502e-01 -6.69301003e-02 -5.02313614e-01
-2.05491588e-01 6.64464116e-01 -1.38452709e+00 1.00633681e-01
3.18740785e-01 1.98398620e-01 -3.36427093e-02 8.62561822e-01
-6.21230491e-02 8.67423058e-01 -1.69946596e-01 -2.72566229e-01
-4.17689160e-02 -2.03745425e-01 3.41130137e-01 9.23713386e-01
3.05880338e-01 6.79267496e-02 3.40280652e-01 5.69654107e-01
-3.98132801e-02 2.13934138e-01 -8.02256107e-01 5.45396749e-03
5.20218074e-01 6.86784744e-01 -5.57853103e-01 -3.87076020e-01
-3.89484107e-01 4.33008164e-01 1.64516345e-01 3.55899662e-01
-4.14465100e-01 -2.38496408e-01 7.16496348e-01 4.01794255e-01
4.56952095e-01 6.58409417e-01 -2.27036729e-01 -7.37032950e-01
-3.01161975e-01 -8.24012578e-01 6.89764917e-01 -3.33526522e-01
-1.32699251e+00 -3.20778042e-01 3.23683411e-01 -9.17916119e-01
-5.73066950e-01 -3.66820455e-01 -8.11762810e-01 1.01081073e+00
-8.98589849e-01 -4.25731540e-01 5.78668535e-01 6.21127784e-02
3.67760867e-01 1.94151610e-01 4.90492791e-01 -1.56575754e-01
-4.94020849e-01 5.21561980e-01 3.84872928e-02 -1.27828419e-01
6.91468537e-01 -1.34788811e+00 -4.92540836e-01 3.19281161e-01
-2.82519728e-01 8.30226779e-01 1.17286873e+00 -6.56186998e-01
-1.12649596e+00 -7.30340064e-01 6.94455445e-01 -1.64042458e-01
6.01918399e-01 -1.69611067e-01 -6.92651570e-01 9.43214476e-01
1.64410770e-02 -3.28207910e-01 7.60399520e-01 6.79045260e-01
-1.02379739e-01 -3.55772197e-01 -1.56567526e+00 8.12853992e-01
6.08415782e-01 -2.17938080e-01 -6.78642333e-01 -6.32931367e-02
7.22234070e-01 2.80698668e-02 -5.66658854e-01 2.68147856e-01
6.80290341e-01 -1.07336414e+00 6.24873221e-01 -1.02892721e+00
4.48405057e-01 3.45779769e-02 -1.16986364e-01 -1.78258669e+00
-3.54635686e-01 -3.69491369e-01 2.91583151e-01 1.23976421e+00
5.64691901e-01 -8.71927917e-01 8.91956806e-01 8.26116502e-01
4.55634892e-01 -9.70390558e-01 -8.38431299e-01 -3.73012722e-01
2.11103484e-01 -1.26285717e-01 8.76025200e-01 7.68164337e-01
8.72063711e-02 5.28530717e-01 -1.89518586e-01 1.16761155e-01
3.25482935e-01 5.36294818e-01 7.92924464e-01 -1.04322195e+00
-6.09682679e-01 -5.30399382e-01 -1.63279057e-01 -7.02431917e-01
1.40384719e-01 -5.73136151e-01 7.30300322e-02 -1.22326565e+00
5.73872745e-01 -3.72257203e-01 -4.93138611e-01 3.99396658e-01
-5.10827780e-01 -5.44502795e-01 1.55796900e-01 -2.22292647e-01
-5.53839877e-02 1.91568404e-01 9.47612703e-01 -2.52962232e-01
-7.44715989e-01 6.06905580e-01 -1.27116144e+00 4.82479930e-01
7.07511187e-01 -5.05033255e-01 -4.44963843e-01 3.42917591e-01
4.41369526e-02 5.70701241e-01 1.51112482e-01 -3.31947565e-01
-1.70329332e-01 -7.92396367e-01 3.07381570e-01 -5.94982430e-02
-1.57769695e-01 -8.68978500e-01 1.95488334e-01 6.71231806e-01
-9.21551049e-01 -3.42142642e-01 8.38147700e-02 7.23907590e-01
2.33593971e-01 -3.49392295e-01 6.40075922e-01 -4.36074845e-02
-1.57465488e-01 8.28462169e-02 -3.38623106e-01 -4.19225767e-02
7.60833621e-01 1.26265496e-01 1.36572942e-02 -7.79012561e-01
-7.70692945e-01 2.99866617e-01 5.42521179e-01 1.65790096e-01
1.53296858e-01 -1.16161919e+00 -7.49428868e-01 8.85031968e-02
5.46669438e-02 -3.64872158e-01 2.99459845e-01 6.04493439e-01
2.80643225e-01 3.98346573e-01 1.68440044e-02 -1.97179303e-01
-9.87728298e-01 9.32158828e-01 1.22486942e-01 -5.14985144e-01
-3.15348655e-01 3.24570388e-01 5.06367505e-01 -1.60694093e-01
-1.57773420e-02 -2.49720559e-01 -3.34530681e-01 4.23845857e-01
4.40392017e-01 2.52757370e-01 7.84806162e-02 -2.17393994e-01
-1.64527252e-01 8.81216899e-02 -3.06914717e-01 -2.75864899e-01
1.35830498e+00 -1.65096417e-01 2.32631058e-01 8.48573148e-01
1.21745455e+00 8.80453736e-02 -1.43223691e+00 -1.90881014e-01
1.71308983e-02 -8.40550005e-01 4.97348756e-02 -9.34654832e-01
-1.02646053e+00 5.00373781e-01 6.89340353e-01 5.40055275e-01
1.24435556e+00 -4.38982844e-02 4.30712029e-02 2.51364619e-01
3.88408005e-01 -1.06850386e+00 -3.15523207e-01 -8.37269351e-02
7.39064276e-01 -1.33391309e+00 1.84018195e-01 2.27086246e-01
-6.42965317e-01 6.97968841e-01 2.18810305e-01 -3.65487695e-01
6.64864123e-01 -1.19586021e-01 -2.06560418e-01 5.22939069e-03
-1.05395067e+00 -1.34212434e-01 2.93232724e-02 3.36527884e-01
3.37948769e-01 5.32467723e-01 -6.26780510e-01 5.14816642e-01
-3.41814011e-01 5.39680235e-02 6.33757710e-01 8.02118659e-01
-4.19858903e-01 -1.12482548e+00 -4.55567330e-01 1.20488620e+00
-4.09843832e-01 2.43320793e-01 -3.55327338e-01 8.50268424e-01
1.48622692e-01 9.85690475e-01 3.47367436e-01 -2.58288473e-01
6.43511832e-01 1.54659644e-01 4.05066192e-01 -6.97303236e-01
-3.96965504e-01 1.23531856e-01 2.93291926e-01 -2.86881447e-01
-1.39101803e-01 -1.14222682e+00 -1.10170412e+00 -4.03557867e-01
-4.66219932e-01 3.23061973e-01 5.20831645e-01 1.17380440e+00
1.68687358e-01 4.18234885e-01 1.20933735e+00 -8.92993689e-01
-7.94913650e-01 -9.24411297e-01 -7.36728013e-01 4.70386058e-01
6.48183405e-01 -8.67817640e-01 -8.01390290e-01 -2.88118988e-01] | [8.367703437805176, 5.411685466766357] |
ac4c7daf-4d6f-423a-b499-7383cd3555e4 | gatortron-a-large-clinical-language-model-to | 2203.03540 | null | https://arxiv.org/abs/2203.03540v3 | https://arxiv.org/pdf/2203.03540v3.pdf | GatorTron: A Large Clinical Language Model to Unlock Patient Information from Unstructured Electronic Health Records | There is an increasing interest in developing artificial intelligence (AI) systems to process and interpret electronic health records (EHRs). Natural language processing (NLP) powered by pretrained language models is the key technology for medical AI systems utilizing clinical narratives. However, there are few clinical language models, the largest of which trained in the clinical domain is comparatively small at 110 million parameters (compared with billions of parameters in the general domain). It is not clear how large clinical language models with billions of parameters can help medical AI systems utilize unstructured EHRs. In this study, we develop from scratch a large clinical language model - GatorTron - using >90 billion words of text (including >82 billion words of de-identified clinical text) and systematically evaluate it on 5 clinical NLP tasks including clinical concept extraction, medical relation extraction, semantic textual similarity, natural language inference (NLI), and medical question answering (MQA). We examine how (1) scaling up the number of parameters and (2) scaling up the size of the training data could benefit these NLP tasks. GatorTron models scale up the clinical language model from 110 million to 8.9 billion parameters and improve 5 clinical NLP tasks (e.g., 9.6% and 9.5% improvement in accuracy for NLI and MQA), which can be applied to medical AI systems to improve healthcare delivery. The GatorTron models are publicly available at: https://catalog.ngc.nvidia.com/orgs/nvidia/teams/clara/models/gatortron_og. | ['Tanja Magoc', 'Ying Zhang', 'Mona G Flores', 'Cheryl Martin', 'Colin Compas', 'Christopher Parisien', 'Kaleb E Smith', 'Hoo Chang Shin', 'Nima PourNejatian', 'Aokun Chen', 'Yonghui Wu', 'Jiang Bian', 'Elizabeth A Shenkman', 'William R Hogan', 'Duane A Mitchell', 'Gloria Lipori', 'Christopher A Harle', 'Xi Yang'] | 2022-02-02 | null | null | null | null | ['medical-relation-extraction', 'clinical-concept-extraction'] | ['medical', 'medical'] | [ 8.63025114e-02 3.38367134e-01 -1.92757100e-01 -3.77910882e-01
-9.69210327e-01 -3.91119212e-01 1.02175698e-01 7.26878643e-01
-4.97756809e-01 6.46594167e-01 5.91991365e-01 -8.51814032e-01
-1.44322127e-01 -7.06717372e-01 -3.04904878e-01 -2.23903820e-01
-1.77480057e-01 1.06330609e+00 -4.40679044e-01 -3.09761427e-02
-2.52104849e-01 3.76446307e-01 -5.71695328e-01 9.65039790e-01
1.07029724e+00 8.28444421e-01 -1.33992761e-01 1.15796876e+00
-2.45276853e-01 1.11930370e+00 -5.58711112e-01 -3.14390153e-01
-1.02330223e-01 -2.31252477e-01 -1.15446341e+00 -4.05324101e-01
1.46719776e-02 -3.24404061e-01 -1.16550818e-01 7.66367614e-01
7.15079486e-01 -5.10302365e-01 5.48745036e-01 -9.68257070e-01
-7.90889502e-01 7.60836124e-01 -1.80972204e-01 3.84069353e-01
4.62391108e-01 3.69332135e-01 8.47918987e-01 -5.94987452e-01
7.39974856e-01 1.17848432e+00 8.22433114e-01 8.00695181e-01
-7.55425751e-01 -6.76006734e-01 -2.53416210e-01 -7.66755193e-02
-1.33128715e+00 -3.31143409e-01 6.45437613e-02 -4.85190660e-01
1.56963074e+00 3.97823364e-01 5.65570831e-01 1.04278123e+00
8.78422618e-01 8.16093683e-01 6.66847587e-01 -2.03035027e-01
1.74228549e-01 1.02584206e-01 5.34699440e-01 8.96123528e-01
4.79885310e-01 -2.64194965e-01 -2.79084742e-01 -8.61524940e-01
4.49982345e-01 2.61185735e-01 -2.06760447e-02 5.05318403e-01
-1.40501177e+00 1.09300113e+00 3.03996176e-01 3.82266939e-01
-5.32157362e-01 6.97159097e-02 8.13241601e-01 1.54897988e-01
4.20122564e-01 1.06799126e+00 -7.81493187e-01 -1.84652939e-01
-5.98707557e-01 2.01946095e-01 1.17511082e+00 9.69736099e-01
-3.28641385e-02 -9.75338891e-02 -1.07353866e-01 7.17522800e-01
1.66342631e-01 7.14041352e-01 7.77949274e-01 -7.41485059e-01
6.57745838e-01 6.56017542e-01 -2.81718671e-01 -8.05492043e-01
-8.95337820e-01 -1.35574162e-01 -1.15677142e+00 -7.41788208e-01
3.73210721e-02 -5.78818858e-01 -1.04671180e+00 1.32254469e+00
1.75472483e-01 5.15952669e-02 4.76296246e-01 3.77412379e-01
1.32690454e+00 6.34012282e-01 6.85722172e-01 -2.71639917e-02
2.11863112e+00 -7.75557637e-01 -8.69003177e-01 -4.21220124e-01
1.32411027e+00 -8.47708404e-01 8.28088760e-01 2.73315877e-01
-1.16055763e+00 -4.99377698e-02 -6.78498566e-01 -2.72777528e-01
-4.95371968e-01 -1.18461184e-01 1.01029992e+00 3.85193646e-01
-9.95281994e-01 1.34904116e-01 -1.16518104e+00 -4.67023224e-01
7.48579264e-01 5.06097436e-01 -2.61102080e-01 -2.47225299e-01
-1.51002860e+00 9.10619020e-01 3.14858317e-01 -1.49391010e-01
-4.72193867e-01 -1.20635641e+00 -8.55656922e-01 2.97081135e-02
1.55736491e-01 -1.54511845e+00 1.22319865e+00 -4.51381117e-01
-1.01565123e+00 9.62559164e-01 -2.16653436e-01 -7.95790255e-01
2.72174310e-02 -2.87036568e-01 -5.95102847e-01 4.14988011e-01
4.73799855e-02 7.00676680e-01 6.28777966e-02 -5.11932373e-01
-4.14731026e-01 -3.84851217e-01 -2.93216854e-01 1.46000495e-03
-4.00500804e-01 2.34257787e-01 -1.20083436e-01 -6.40836775e-01
-3.81004572e-01 -8.51126373e-01 -6.57083690e-01 -6.19213507e-02
-5.83709717e-01 -1.48341745e-01 9.24589038e-02 -8.46744955e-01
1.27826321e+00 -1.88816357e+00 -4.00131464e-01 9.05314684e-02
7.24427879e-01 4.66213614e-01 -2.99962580e-01 4.25061733e-01
-2.03931462e-02 5.01518369e-01 -1.95536599e-01 2.58782264e-02
-4.17482138e-01 1.87939689e-01 -2.87935644e-01 1.01020029e-02
4.28343296e-01 1.42128718e+00 -9.76080120e-01 -8.04726243e-01
-1.11694373e-01 6.36050463e-01 -9.05740082e-01 2.77822375e-01
-2.56898880e-01 1.79536581e-01 -9.79251742e-01 8.58891129e-01
3.18385720e-01 -9.83148634e-01 2.86663920e-01 -1.98334575e-01
4.93389517e-01 5.66235363e-01 -4.70454544e-01 1.48529458e+00
-4.05388653e-01 1.72469065e-01 -1.16319485e-01 -6.58243239e-01
5.36901891e-01 6.32441401e-01 1.03794301e+00 -4.14766312e-01
1.95516363e-01 2.81513799e-02 2.70821393e-01 -1.06279874e+00
1.12245895e-01 -3.30981761e-01 -2.50323534e-01 4.81239051e-01
-1.66869015e-01 -2.33499147e-02 -2.11585388e-02 4.29650187e-01
1.60171270e+00 -7.38283455e-01 8.49309266e-01 -2.37962708e-01
3.26465666e-01 5.16495287e-01 5.10974109e-01 5.96744895e-01
-1.14992641e-01 2.52113014e-01 3.56270432e-01 -7.13866830e-01
-8.64570558e-01 -1.07661223e+00 -4.47151661e-01 7.56777227e-01
-4.51950401e-01 -7.79357374e-01 -6.59769237e-01 -5.15639067e-01
1.78540528e-01 7.12126017e-01 -2.98480183e-01 -2.02139676e-01
-6.62475049e-01 -1.35293686e+00 1.09961498e+00 8.23008060e-01
7.99291953e-02 -1.27412057e+00 -6.92093372e-01 3.98543328e-01
-2.76829422e-01 -1.42515135e+00 -6.19153678e-01 -1.04186870e-01
-1.06893158e+00 -1.02008772e+00 -2.88054407e-01 -6.37394369e-01
6.28235221e-01 -4.10333991e-01 1.40956998e+00 3.23703475e-02
-8.78242612e-01 3.12319428e-01 -1.74056739e-01 -8.66220355e-01
-7.78490961e-01 2.05923557e-01 6.54541701e-02 -6.53964043e-01
9.20436442e-01 -1.91855177e-01 -7.40674078e-01 -3.54301870e-01
-1.01415503e+00 2.47280553e-01 7.19463885e-01 8.76883864e-01
5.70588112e-01 -5.29823065e-01 9.13358569e-01 -1.52415347e+00
8.95754337e-01 -8.73744071e-01 2.03222185e-02 3.20796281e-01
-7.29384661e-01 6.12373948e-02 7.39110887e-01 -4.81770843e-01
-6.00999475e-01 -2.36182928e-01 -5.89937270e-01 -8.74699950e-02
-2.38384411e-01 8.90498519e-01 3.41443181e-01 4.48798895e-01
7.72591114e-01 -1.76438183e-01 1.60247803e-01 -3.39752972e-01
3.51010084e-01 9.60713148e-01 1.53435513e-01 -3.27708304e-01
1.87962890e-01 4.84655768e-01 -2.33659104e-01 -6.12881362e-01
-8.06296408e-01 -5.38045526e-01 -1.30778432e-01 5.80929577e-01
1.34782732e+00 -9.59548593e-01 -1.08322060e+00 -7.83077180e-02
-1.16227460e+00 -1.65881261e-01 -3.05825859e-01 6.00904346e-01
-2.95099437e-01 2.58263916e-01 -1.36150730e+00 -4.20011014e-01
-1.32495511e+00 -1.00124729e+00 1.15305209e+00 -1.44521058e-01
-1.01965797e+00 -1.20511723e+00 1.12843879e-01 6.33585155e-01
4.39229608e-01 2.12210804e-01 1.55734682e+00 -1.18067002e+00
-9.39696953e-02 -2.40824431e-01 -2.47525111e-01 1.84219014e-02
2.99653918e-01 -3.52113426e-01 -7.45578587e-01 -1.56452894e-01
1.68321371e-01 -3.42970699e-01 4.79081303e-01 5.24141073e-01
1.38030088e+00 -6.51563823e-01 -6.52596653e-01 7.51279950e-01
1.18283057e+00 3.89196873e-01 3.89400870e-01 -1.41446099e-01
8.17516506e-01 3.37581307e-01 3.13074142e-01 4.12947536e-01
6.34620249e-01 1.24581099e-01 -2.53866345e-01 -3.25834483e-01
5.08078597e-02 1.15542682e-02 1.84390470e-01 1.36811090e+00
1.35459006e-01 -1.75837666e-01 -1.65915024e+00 5.55823863e-01
-1.56702292e+00 -4.06531453e-01 6.70319796e-02 1.56706822e+00
1.38021910e+00 -8.77059996e-02 -2.88003594e-01 -5.13064504e-01
1.39100030e-01 -2.96287447e-01 -7.51987875e-01 -7.38082647e-01
2.10957274e-01 4.87462550e-01 3.45753670e-01 5.80797315e-01
-8.47050726e-01 8.59821677e-01 6.18319654e+00 6.73325896e-01
-1.00415885e+00 1.51369736e-01 1.06374991e+00 -3.93427610e-01
-2.05417380e-01 -4.82355028e-01 -9.55076277e-01 3.05804968e-01
1.60080159e+00 -2.94311374e-01 -3.63483503e-02 7.31344402e-01
3.19763333e-01 2.53185302e-01 -1.44775701e+00 1.01917398e+00
1.27318397e-01 -1.76690865e+00 5.42741835e-01 2.17934296e-01
6.28752112e-01 2.95739979e-01 -7.41716400e-02 2.28470132e-01
4.94041681e-01 -1.54319298e+00 -1.63703725e-01 5.57343066e-01
1.01447833e+00 -4.62714493e-01 1.12622237e+00 1.75417274e-01
-8.82041335e-01 -8.47729668e-03 -1.32222548e-01 3.17806184e-01
3.34614038e-01 6.86665833e-01 -1.39779460e+00 4.25141633e-01
6.94151878e-01 6.86395466e-01 -4.60927367e-01 4.43075269e-01
4.20584440e-01 7.67993152e-01 -2.69188970e-01 -1.18231609e-01
1.70615628e-01 1.17581278e-01 2.47344226e-01 1.53984201e+00
9.34635848e-02 6.89571023e-01 3.44829351e-01 7.40604222e-01
-2.80230075e-01 3.16782475e-01 -5.35221338e-01 -4.51136827e-01
3.87511224e-01 1.11267817e+00 -3.79721820e-01 -8.05660248e-01
-3.31004620e-01 5.74073672e-01 7.80814886e-02 3.01358670e-01
-7.72858381e-01 -2.41463035e-01 6.91641986e-01 1.67148128e-01
-3.28899652e-01 2.46794224e-01 -6.72520041e-01 -1.12903845e+00
-2.91225284e-01 -1.40085936e+00 8.91441107e-01 -6.26135230e-01
-1.89035344e+00 9.56613600e-01 -1.77340120e-01 -8.80181253e-01
-6.34256423e-01 -6.92830801e-01 -1.04122892e-01 8.58673275e-01
-1.36366963e+00 -1.10334718e+00 -1.17386580e-01 4.68524396e-01
4.00192589e-01 -2.10108131e-01 1.38742578e+00 4.55836952e-01
-3.89559984e-01 6.51374698e-01 -1.40673012e-01 4.30622369e-01
7.59248853e-01 -1.01532400e+00 4.97065723e-01 1.04071042e-02
-2.72675633e-01 1.14576662e+00 2.24458307e-01 -7.80890465e-01
-1.47371590e+00 -1.35385108e+00 1.46724236e+00 -8.45415533e-01
7.98847437e-01 -2.03200713e-01 -1.07609642e+00 8.99891794e-01
1.15388371e-01 -6.06280938e-02 1.38784385e+00 -4.77090217e-02
-1.96581706e-01 5.26798554e-02 -1.27739143e+00 6.50057316e-01
7.94612885e-01 -6.04273677e-01 -7.20733166e-01 7.07416117e-01
1.12667942e+00 -2.83324689e-01 -1.56549168e+00 6.77775145e-01
4.05377746e-01 6.15491420e-02 1.15974641e+00 -1.27372885e+00
6.92283511e-01 9.20027718e-02 1.60746992e-01 -8.66421700e-01
-8.84671435e-02 -4.78876978e-01 -1.87684670e-01 5.47866702e-01
7.64750063e-01 -9.87016439e-01 5.51637053e-01 8.72636199e-01
5.28412908e-02 -1.48203909e+00 -4.97835368e-01 -2.44894505e-01
3.84906560e-01 -3.20763737e-01 5.16651034e-01 1.27592611e+00
4.63562250e-01 5.88971615e-01 1.86562344e-01 4.39714044e-02
1.38939247e-01 -2.92443931e-02 1.47416890e-01 -9.78630841e-01
-4.06530946e-01 -3.12460095e-01 -1.87588423e-01 -6.27618611e-01
-7.20232027e-03 -1.19786704e+00 -2.55465806e-01 -1.84915161e+00
5.62827766e-01 -5.64229906e-01 -1.68880701e-01 8.07020605e-01
-4.00227666e-01 -5.13629057e-02 -1.33286538e-02 1.31698787e-01
-4.45204854e-01 9.57789198e-02 9.49137568e-01 -2.54591733e-01
-2.53462404e-01 -2.59807467e-01 -9.09056902e-01 7.77172685e-01
9.19869006e-01 -5.42951524e-01 -2.27996126e-01 -5.59024572e-01
3.35278422e-01 2.39413932e-01 1.10791914e-01 -5.75342536e-01
3.43295395e-01 -1.05299711e-01 3.38868141e-01 -2.12015629e-01
1.32402286e-01 -4.79579866e-01 -1.24431424e-01 9.71624374e-01
-6.47834063e-01 4.44379508e-01 5.58009386e-01 1.86847150e-01
-1.56102046e-01 1.41608745e-01 5.00135601e-01 -3.76166463e-01
-1.84737355e-01 3.69597763e-01 -8.00938487e-01 4.50456321e-01
7.52133667e-01 2.03625888e-01 -4.79221702e-01 -5.50817959e-02
-7.20991910e-01 5.62222242e-01 -1.96990341e-01 3.00777465e-01
8.00429821e-01 -9.38067973e-01 -9.89436090e-01 -2.54374482e-02
2.91756868e-01 1.25860199e-01 3.47124934e-01 8.88858795e-01
-1.00380135e+00 9.17244375e-01 1.57049343e-01 -4.73169595e-01
-1.42957544e+00 7.12609410e-01 2.83361256e-01 -8.64745259e-01
-7.29302883e-01 6.86461449e-01 3.52277964e-01 -4.60865080e-01
5.01922034e-02 -8.22004080e-01 -1.73943534e-01 -2.81212121e-01
8.04668427e-01 -6.99228048e-02 1.67851914e-02 -8.88048187e-02
-6.68244123e-01 3.04706007e-01 -4.37958866e-01 2.40755901e-01
1.47384918e+00 3.63271534e-01 -4.02461469e-01 2.63870329e-01
1.23904514e+00 -1.13644175e-01 -1.66396216e-01 -8.86742696e-02
2.67530441e-01 3.26084495e-01 -2.47760981e-01 -1.07737350e+00
-6.91660464e-01 9.34727907e-01 3.83994371e-01 -2.57438093e-01
1.02569973e+00 1.84226558e-01 1.33945572e+00 5.98211706e-01
2.08281279e-02 -7.17101097e-01 -6.87067062e-02 4.26605225e-01
6.30482733e-01 -1.14310229e+00 1.34997830e-01 -5.44032812e-01
-9.42253530e-01 8.81335258e-01 4.70259428e-01 1.09274246e-01
9.43863273e-01 8.49980235e-01 3.36089164e-01 -7.01724231e-01
-1.16730332e+00 2.33608752e-01 2.65745759e-01 2.35841811e-01
9.04559672e-01 4.31640208e-01 -2.36509502e-01 9.43740547e-01
-3.34543407e-01 3.55030537e-01 2.88287789e-01 8.55568230e-01
4.19631600e-02 -9.99723136e-01 -1.45486102e-01 1.00334728e+00
-9.03174222e-01 -8.86259496e-01 -2.75152087e-01 4.67904150e-01
3.90432812e-02 9.18819189e-01 8.49638656e-02 -1.84984341e-01
1.64954334e-01 3.37252438e-01 1.16958926e-02 -1.09755671e+00
-1.10794866e+00 -1.45010948e-01 4.90038633e-01 -6.63006604e-01
-8.37253705e-02 -3.07050437e-01 -1.82818270e+00 -2.89140135e-01
1.01856507e-01 1.38326958e-01 3.80763233e-01 7.81557202e-01
9.73467946e-01 7.41974890e-01 -1.88472584e-01 3.09138268e-01
-5.86051047e-01 -9.47427988e-01 -8.60887840e-02 3.31108153e-01
2.67620981e-01 2.99185395e-01 -5.93460537e-02 3.97909164e-01] | [8.59883975982666, 8.548994064331055] |
28fa7c83-e8eb-453e-af28-f30b502828f8 | adversarial-synthesis-learning-enables | 1712.07695 | null | http://arxiv.org/abs/1712.07695v1 | http://arxiv.org/pdf/1712.07695v1.pdf | Adversarial Synthesis Learning Enables Segmentation Without Target Modality Ground Truth | A lack of generalizability is one key limitation of deep learning based
segmentation. Typically, one manually labels new training images when
segmenting organs in different imaging modalities or segmenting abnormal organs
from distinct disease cohorts. The manual efforts can be alleviated if one is
able to reuse manual labels from one modality (e.g., MRI) to train a
segmentation network for a new modality (e.g., CT). Previously, two stage
methods have been proposed to use cycle generative adversarial networks
(CycleGAN) to synthesize training images for a target modality. Then, these
efforts trained a segmentation network independently using synthetic images.
However, these two independent stages did not use the complementary information
between synthesis and segmentation. Herein, we proposed a novel end-to-end
synthesis and segmentation network (EssNet) to achieve the unpaired MRI to CT
image synthesis and CT splenomegaly segmentation simultaneously without using
manual labels on CT. The end-to-end EssNet achieved significantly higher median
Dice similarity coefficient (0.9188) than the two stages strategy (0.8801), and
even higher than canonical multi-atlas segmentation (0.9125) and ResNet method
(0.9107), which used the CT manual labels. | ['Bennett A. Landman', 'Richard G. Abramson', 'Albert Assad', 'Shunxing Bao', 'Zhoubing Xu', 'Yuankai Huo'] | 2017-12-20 | null | null | null | null | ['splenomegaly-segmentation-on-multi-modal-mri'] | ['medical'] | [ 3.98474574e-01 4.95575517e-01 1.41119942e-01 -3.96200716e-01
-9.53363597e-01 -8.74881268e-01 3.95226032e-01 -2.13598326e-01
-4.44166541e-01 8.36987257e-01 -1.36460379e-01 -3.75589520e-01
3.66284639e-01 -7.15127110e-01 -7.04293370e-01 -7.75329113e-01
2.08341330e-01 7.61530876e-01 2.49802470e-01 1.86493456e-01
-1.23734355e-01 5.77629864e-01 -5.44203579e-01 1.11588286e-02
1.14288914e+00 6.99901819e-01 1.67656094e-01 7.45471954e-01
-1.22474656e-01 3.66102785e-01 -6.70810401e-01 -4.52016473e-01
5.63824773e-01 -1.01244521e+00 -8.06966305e-01 1.45941123e-01
3.16444039e-01 -5.40101707e-01 -2.17996031e-01 1.04193819e+00
6.27300024e-01 -1.31874055e-01 7.30894625e-01 -1.10621417e+00
-6.13760710e-01 8.07015181e-01 -7.56887078e-01 -8.15024227e-03
-8.82077813e-02 4.90359545e-01 3.07504326e-01 -4.29340273e-01
7.54910648e-01 8.85584891e-01 7.59844780e-01 7.51267612e-01
-1.27194512e+00 -9.07270253e-01 -3.04292887e-01 -5.12777030e-01
-1.26536274e+00 6.19203560e-02 5.48052013e-01 -5.14643967e-01
1.81655526e-01 2.69327402e-01 7.80203938e-01 9.47518647e-01
4.03325617e-01 4.82607365e-01 1.41367304e+00 -5.47855757e-02
3.61810476e-02 -1.09296553e-01 -3.34685504e-01 7.34282970e-01
2.30120242e-01 2.47750636e-02 2.45804146e-01 -1.75690111e-02
1.25086212e+00 9.24495421e-03 -9.74012688e-02 -1.80727586e-01
-1.53965282e+00 8.92571151e-01 6.84872210e-01 4.47815746e-01
-3.91173184e-01 8.48612189e-02 5.21618366e-01 9.51627344e-02
1.20655976e-01 5.62414169e-01 -7.26062208e-02 4.33272928e-01
-1.32922304e+00 -1.93210647e-01 5.28164566e-01 6.69942677e-01
2.96562850e-01 3.88860375e-01 -2.59190828e-01 5.91976702e-01
2.30139121e-01 5.09422183e-01 8.82853985e-01 -9.66630161e-01
2.87315659e-02 3.65428388e-01 -3.19401443e-01 -6.78553879e-01
-6.94060743e-01 -6.68664813e-01 -1.26491368e+00 8.43833387e-02
5.58039129e-01 -3.09684068e-01 -1.54640496e+00 1.93177104e+00
4.91200030e-01 2.19898283e-01 -5.38895354e-02 1.11357152e+00
1.01895475e+00 2.50539780e-01 3.53481978e-01 -7.76226632e-03
1.18423343e+00 -1.14521050e+00 -4.55533147e-01 2.71231178e-02
6.36365473e-01 -8.09506357e-01 1.00869107e+00 1.81214452e-01
-1.22602630e+00 -3.64577562e-01 -1.15186346e+00 1.79793447e-01
-1.96376428e-01 -4.36922126e-02 5.16420782e-01 8.56061757e-01
-1.16947305e+00 5.02567351e-01 -1.08010197e+00 -3.80562037e-01
4.83922839e-01 6.19859755e-01 -2.78170556e-01 -1.83353014e-02
-1.19702244e+00 7.96913266e-01 4.30253148e-01 -1.44491270e-01
-1.22927833e+00 -7.85153449e-01 -6.06699646e-01 -2.30122343e-01
2.17710137e-01 -9.53271210e-01 9.60501134e-01 -1.23332810e+00
-1.62308943e+00 1.12511039e+00 4.39759463e-01 -4.14382517e-01
8.31317008e-01 3.53897333e-01 -1.80259973e-01 5.61183453e-01
2.49318510e-01 1.16322684e+00 7.23305404e-01 -1.33988738e+00
-9.90281329e-02 -2.53074080e-01 -1.43393159e-01 1.26074135e-01
3.06685984e-01 4.82170135e-02 -3.15448642e-01 -9.55850303e-01
3.18789631e-01 -1.24897408e+00 -4.13440019e-01 2.14511290e-01
-5.35557508e-01 3.90797198e-01 4.32570010e-01 -1.08547556e+00
5.02683461e-01 -1.77506733e+00 4.40457091e-02 3.66806865e-01
5.27410448e-01 2.72522986e-01 -1.41440973e-01 -1.84396252e-01
-2.32546598e-01 3.76712143e-01 -7.96882927e-01 -1.67563170e-01
-3.41716051e-01 1.96250156e-01 1.88961744e-01 6.89742863e-01
-2.47375444e-02 1.12107825e+00 -1.00067437e+00 -9.29158807e-01
2.43555948e-01 3.40075344e-01 -5.66187739e-01 2.65145391e-01
2.01429307e-01 1.33537054e+00 -2.92885751e-01 7.27555692e-01
8.30407381e-01 -3.25124443e-01 3.48184824e-01 -3.25401306e-01
3.47352833e-01 -3.57238531e-01 -7.34096825e-01 1.95333290e+00
-4.28119451e-01 1.49103850e-01 5.67151234e-03 -8.91449928e-01
7.50472128e-01 5.03754199e-01 8.96758556e-01 -5.20442545e-01
3.41091692e-01 4.31252629e-01 4.24955070e-01 -1.34641945e-01
-1.30600065e-01 -6.48462355e-01 -1.92152739e-01 4.93119359e-01
2.32810125e-01 -5.58516324e-01 2.04405859e-02 2.31684409e-02
8.51565838e-01 -4.48692450e-03 2.93773636e-02 -2.23345459e-01
5.65064192e-01 1.17039330e-01 5.88467419e-01 5.86412728e-01
-4.80623990e-01 1.11713541e+00 4.70272511e-01 -1.98834211e-01
-1.30976605e+00 -1.37273157e+00 -1.42528206e-01 6.80640042e-01
2.98573938e-03 2.31972396e-01 -9.40691590e-01 -1.03783262e+00
-3.32295597e-01 4.84660923e-01 -6.90492749e-01 -1.99307814e-01
-7.59034455e-01 -8.87441814e-01 1.14888608e+00 4.34752554e-01
6.78424835e-01 -8.46621871e-01 -3.98865402e-01 1.95848763e-01
-1.63636059e-01 -9.93237376e-01 -7.43462443e-01 -1.18842028e-01
-1.05316818e+00 -9.66171086e-01 -1.40531313e+00 -7.58921444e-01
9.93276477e-01 -3.10683727e-01 9.68670845e-01 6.11068457e-02
-1.98350713e-01 9.68346447e-02 -1.17231868e-01 1.95003711e-02
-7.49736905e-01 2.52786934e-01 -2.20856536e-02 -1.44668356e-01
-3.78427774e-01 -5.78287721e-01 -9.66511369e-01 4.45366234e-01
-1.31179917e+00 3.02425414e-01 7.73501813e-01 1.04460371e+00
9.88076448e-01 -4.00451899e-01 6.13241255e-01 -1.01786351e+00
4.12776291e-01 -4.70630139e-01 -4.73054707e-01 4.44000334e-01
-6.16664410e-01 -1.76359609e-01 8.36530387e-01 -5.95904589e-01
-9.54160154e-01 2.20308781e-01 -2.70092160e-01 -5.67343473e-01
-6.42268360e-02 3.42770338e-01 2.02491224e-01 -3.55381280e-01
3.94590467e-01 3.01438570e-01 2.10177034e-01 -1.38525963e-01
2.96023875e-01 2.73360908e-01 6.61307931e-01 -4.36076581e-01
6.23357177e-01 3.64521146e-01 2.26468578e-01 -7.38313794e-02
-4.44561332e-01 1.44607246e-01 -8.74387801e-01 -1.61734626e-01
1.24702132e+00 -7.05122590e-01 -3.16351682e-01 5.60330212e-01
-7.82176793e-01 -4.45894748e-01 -2.91246265e-01 6.84599936e-01
-4.69613314e-01 4.66610163e-01 -7.91159511e-01 2.24338681e-03
-7.93113649e-01 -1.63123930e+00 7.60133743e-01 5.05890250e-01
-2.29096130e-01 -1.12054324e+00 7.68780634e-02 3.49091679e-01
5.84057271e-01 9.49379861e-01 6.55646503e-01 -9.53227341e-01
-3.27923685e-01 -2.14253202e-01 -3.78170848e-01 3.52779180e-01
2.05755129e-01 -9.64621753e-02 -5.88478208e-01 -4.66203332e-01
1.83678254e-01 -2.92250514e-01 4.07408267e-01 5.65937757e-01
1.15736163e+00 -1.48555590e-02 -1.28217369e-01 9.73726511e-01
1.42423081e+00 5.36373854e-01 5.30752361e-01 2.06714366e-02
8.55921507e-01 3.06113511e-01 2.91840166e-01 3.23170871e-02
1.73486203e-01 3.49915445e-01 1.61051258e-01 -6.83894455e-01
-4.74861294e-01 -2.49069363e-01 -4.38331254e-02 8.51819277e-01
1.73323199e-01 -1.86422810e-01 -1.15717566e+00 4.64327097e-01
-1.17825961e+00 -2.65152514e-01 1.06032565e-02 1.97693503e+00
9.32746530e-01 3.07618566e-02 9.70743150e-02 -4.50535059e-01
9.38680649e-01 -1.60091445e-01 -8.61026943e-01 -3.18980664e-01
1.97037216e-03 4.67339784e-01 7.48361588e-01 3.19693118e-01
-8.22576642e-01 8.59986186e-01 6.41233206e+00 7.29910135e-01
-1.51028788e+00 5.28132379e-01 9.84458208e-01 5.75290881e-02
-2.14897051e-01 -4.40364070e-02 -1.65152088e-01 6.51847064e-01
1.00276971e+00 1.02868211e-02 2.70694107e-01 4.25619066e-01
-5.60183004e-02 -1.76510915e-01 -8.17927659e-01 6.50712788e-01
4.11268556e-03 -1.07378590e+00 -1.55190472e-02 -3.22532207e-02
9.15784419e-01 4.87285592e-02 1.32384226e-01 2.31566951e-01
4.25195843e-01 -1.20960140e+00 2.74584293e-01 4.63417262e-01
1.31226718e+00 -5.82643867e-01 8.15685868e-01 5.66101633e-03
-8.39303553e-01 6.97585404e-01 -7.61099532e-02 6.90184474e-01
3.47380430e-01 4.29635435e-01 -1.18526399e+00 6.21774316e-01
3.06878120e-01 9.72117111e-02 -5.90541303e-01 9.26273286e-01
-1.88602567e-01 5.89581490e-01 -4.11727756e-01 4.79573667e-01
4.56525832e-01 -3.36053103e-01 4.77313250e-01 1.05247235e+00
2.95329988e-01 5.50359040e-02 1.08287834e-01 1.04255593e+00
-2.67289430e-01 1.53543502e-01 -2.45995164e-01 -7.09808916e-02
1.29496351e-01 1.46101677e+00 -1.38056719e+00 -5.94697833e-01
-2.61240184e-01 1.13045907e+00 -3.11698943e-01 8.64071473e-02
-1.28733957e+00 -3.79490435e-01 -1.35334909e-01 5.56636192e-02
-1.86910667e-02 -1.43485283e-02 -5.89704871e-01 -1.13017988e+00
-5.09932101e-01 -8.24647248e-01 4.71178532e-01 -7.89814055e-01
-1.18577230e+00 7.38589048e-01 -7.56234601e-02 -1.24124432e+00
-1.59480572e-01 -2.08841041e-01 -7.29329824e-01 1.08245802e+00
-1.04504812e+00 -1.40464139e+00 -3.03235024e-01 5.82322776e-01
2.83886105e-01 3.77163477e-02 6.12280250e-01 4.30607498e-01
-5.00593126e-01 8.90697241e-01 1.38715506e-01 4.47064430e-01
9.39019263e-01 -1.26196969e+00 3.86790633e-02 8.66043150e-01
-4.74517882e-01 5.53620458e-01 3.83018881e-01 -8.84347081e-01
-9.55449641e-01 -1.12583220e+00 3.13132614e-01 -5.09801209e-02
4.12047476e-01 6.23822808e-02 -7.64281929e-01 8.43387306e-01
4.23663080e-01 2.10522100e-01 8.02644849e-01 -6.92365229e-01
1.64484918e-01 3.91856171e-02 -1.89902997e+00 6.46267653e-01
6.62067771e-01 -1.74168319e-01 -4.37118828e-01 3.19969982e-01
8.44834447e-01 -8.77301216e-01 -1.44658029e+00 5.32062590e-01
6.47823393e-01 -5.38442850e-01 9.90377903e-01 -3.84868890e-01
5.07090509e-01 -3.05562973e-01 1.61534667e-01 -1.34770977e+00
7.60724917e-02 -4.67723608e-01 3.70711565e-01 1.14871562e+00
3.75585705e-01 -7.93154299e-01 7.70480752e-01 6.86784983e-01
-3.24789315e-01 -6.98479950e-01 -8.88181329e-01 -6.32443309e-01
5.69463670e-01 1.41367495e-01 7.01201797e-01 1.21506488e+00
-4.96163845e-01 2.58441940e-02 -2.40290061e-01 -7.15301558e-02
6.86810970e-01 1.07548617e-01 4.83187556e-01 -8.11434567e-01
-2.25697219e-01 -4.62028593e-01 -1.75649434e-01 -6.14153981e-01
-4.82821018e-02 -1.31734514e+00 -1.13305420e-01 -1.36393762e+00
2.85959303e-01 -6.33460224e-01 -4.37877476e-01 4.95968819e-01
-7.75892287e-02 6.58679008e-01 3.19053680e-01 3.63969296e-01
-1.48633569e-01 2.95879930e-01 2.14742708e+00 -2.65112296e-02
-9.41918269e-02 -1.45165488e-01 -6.91101909e-01 6.08206451e-01
1.06736791e+00 -7.03490615e-01 -3.41164619e-01 -2.14623213e-01
-2.29084492e-01 4.53193307e-01 3.29265386e-01 -9.28943813e-01
9.52202231e-02 6.15595281e-02 8.15426111e-01 -4.58395243e-01
-1.35448918e-01 -6.97473764e-01 6.25192463e-01 9.90209460e-01
-2.22081020e-01 6.10824861e-02 1.89373851e-01 2.73789000e-02
-1.51996091e-01 -2.66081512e-01 1.19277906e+00 -5.59391499e-01
-1.00726604e-01 4.37163472e-01 -2.78055459e-01 1.92789167e-01
1.13445485e+00 -1.45926431e-01 -1.95193570e-02 -2.16489092e-01
-1.09228635e+00 1.66389033e-01 5.44891596e-01 -6.39400482e-02
5.35803974e-01 -1.27694857e+00 -7.46844649e-01 1.30151242e-01
-4.88963246e-01 2.41392791e-01 3.37279588e-01 1.43414617e+00
-1.02623188e+00 2.81517357e-01 -7.01157033e-01 -8.23373199e-01
-8.33161473e-01 4.47705656e-01 6.71881139e-01 -7.59074032e-01
-4.20617312e-01 6.84741735e-01 4.70143080e-01 -8.40827107e-01
-3.37900370e-01 -3.13603252e-01 2.24398762e-01 -8.98085609e-02
-1.06093995e-01 2.15731323e-01 5.18198162e-02 -6.32507145e-01
-3.10979009e-01 6.17523193e-01 -7.42129534e-02 -2.44990930e-01
1.12378812e+00 -1.10740840e-01 -9.33873057e-02 3.51793803e-02
1.19452655e+00 -1.67301908e-01 -1.10502303e+00 -3.71348113e-02
-5.22678077e-01 -2.97871053e-01 4.21336628e-02 -1.12221014e+00
-1.66153967e+00 6.54601216e-01 9.48005080e-01 -4.01253477e-02
1.29487741e+00 -2.03583524e-01 1.25701594e+00 -4.03540671e-01
2.67505616e-01 -6.07562542e-01 -1.34433806e-01 1.11777112e-01
7.86637187e-01 -1.18254304e+00 -7.69565329e-02 -2.83629030e-01
-9.36439216e-01 9.72812593e-01 6.88614905e-01 -3.00761759e-01
4.15706515e-01 1.80473238e-01 1.85220644e-01 -8.90941396e-02
5.57177551e-02 4.40653525e-02 4.17785019e-01 5.81774712e-01
5.33146560e-01 2.17755467e-01 -6.64165020e-01 5.59664547e-01
-1.81677207e-01 6.89136162e-02 4.29919988e-01 7.96304107e-01
1.48084551e-01 -1.17833698e+00 -4.71722513e-01 6.06806219e-01
-9.15865302e-01 -3.37260105e-02 -2.89440900e-01 9.68631923e-01
2.31228337e-01 3.41817737e-01 -6.81188405e-02 -2.26718545e-01
-5.11645302e-02 -1.76014621e-02 6.09895110e-01 -4.27375108e-01
-9.13728654e-01 2.83022732e-01 -4.32492107e-01 -4.09354329e-01
-4.62567836e-01 -5.03083050e-01 -1.45878482e+00 -3.07750285e-01
-1.30563408e-01 -2.42959604e-01 5.58593512e-01 6.92805827e-01
3.16207968e-02 7.84067392e-01 6.80496037e-01 -6.46223962e-01
-1.59918621e-01 -9.25585508e-01 -5.58691502e-01 4.47743148e-01
-3.92521061e-02 -4.40501273e-01 -2.36409515e-01 2.00641781e-01] | [14.285225868225098, -2.2306711673736572] |
a7fcb22f-eeac-4729-afa0-2f09f4be0273 | guir-mup-2022-towards-generating-topic-aware | null | null | https://aclanthology.org/2022.sdp-1.34 | https://aclanthology.org/2022.sdp-1.34.pdf | GUIR @ MuP 2022: Towards Generating Topic-aware Multi-perspective Summaries for Scientific Documents | This paper presents our approach for the MuP 2022 shared task —-Multi-Perspective Scientific Document Summarization, where the objective is to enable summarization models to explore methods for generating multi-perspective summaries for scientific papers. We explore two orthogonal ways to cope with this task. The first approach involves incorporating a neural topic model (i.e., NTM) into the state-of-the-art abstractive summarizer (LED); the second approach involves adding a two-step summarizer that extracts the salient sentences from the document and then writes abstractive summaries from those sentences. Our latter model outperformed our other submissions on the official test set. Specifically, among 10 participants (including organizers’ baseline) who made their results public with 163 total runs. Our best system ranks first in Rouge-1 (F), and second in Rouge-1 (R), Rouge-2 (F) and Average Rouge (F) scores. | ['Nazli Goharian', 'Sajad Sotudeh'] | null | null | null | null | sdp-coling-2022-10 | ['scientific-article-summarization', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.86002026e-02 5.38892448e-01 -1.46553874e-01 -1.39883935e-01
-1.68196285e+00 -7.11305916e-01 8.99196327e-01 4.80504990e-01
-2.59879112e-01 1.14794385e+00 1.10411608e+00 -2.13011354e-01
7.03996941e-02 -3.69512171e-01 -8.08185399e-01 -3.95329356e-01
1.15119226e-01 5.86859524e-01 1.98947433e-02 -1.52978212e-01
1.05183268e+00 1.85026705e-01 -1.09948790e+00 5.96277237e-01
1.27962184e+00 3.47686261e-01 3.27659696e-01 1.27831876e+00
-2.76797533e-01 6.40870988e-01 -1.25017452e+00 -3.88533145e-01
-3.58687103e-01 -5.81923664e-01 -1.01971805e+00 -4.32061762e-01
8.28571081e-01 -5.65318502e-02 -2.17225879e-01 8.06858063e-01
7.27126896e-01 4.63760078e-01 9.46117699e-01 -8.01137626e-01
-7.32910216e-01 1.13285208e+00 -7.90487528e-01 4.24625039e-01
5.26728570e-01 8.07033572e-03 1.26911008e+00 -9.64012623e-01
9.47633684e-01 1.42324615e+00 3.66394281e-01 6.21036768e-01
-1.06218219e+00 -2.93985546e-01 6.14893883e-02 -1.46212563e-01
-7.15207517e-01 -6.74439490e-01 5.31164467e-01 -1.13025427e-01
1.25654423e+00 6.05073929e-01 3.64186317e-01 1.21650958e+00
5.88215470e-01 1.06903040e+00 6.03982985e-01 -2.67305762e-01
3.48024368e-01 -5.08786775e-02 7.84104168e-01 1.92255825e-01
7.16078699e-01 -7.65052259e-01 -8.96548092e-01 -3.47237229e-01
7.46464878e-02 -5.21285713e-01 -3.28091383e-01 4.51066434e-01
-1.58540058e+00 6.09706402e-01 1.57393873e-01 3.10535043e-01
-7.50044703e-01 8.44607726e-02 7.47296810e-01 9.67767462e-02
9.60585654e-01 1.32123399e+00 -2.50390917e-01 -2.50329524e-01
-1.66310441e+00 9.01891649e-01 1.10708463e+00 8.81144464e-01
1.05375744e-01 8.99345949e-02 -1.14720714e+00 8.10393155e-01
-1.27827451e-01 3.85910183e-01 6.21124625e-01 -1.00793231e+00
7.34575808e-01 1.30353630e-01 3.26984495e-01 -8.79226685e-01
-3.27015460e-01 -6.22967839e-01 -7.98884988e-01 -4.49327856e-01
-1.42080158e-01 -4.40249801e-01 -8.04703534e-01 1.36677456e+00
-3.43218625e-01 -1.01757839e-01 4.58998322e-01 4.94212508e-01
1.74778926e+00 1.32530546e+00 5.31627089e-02 -5.34743786e-01
1.39654469e+00 -1.47746384e+00 -8.30191433e-01 -1.44633606e-01
3.85921180e-01 -7.65361428e-01 9.50327456e-01 4.88872141e-01
-1.68830216e+00 -4.09259886e-01 -1.28797328e+00 -3.64207774e-01
-2.68343478e-01 4.62471604e-01 2.19246984e-01 7.12269619e-02
-1.59407330e+00 9.40722942e-01 -5.62643290e-01 -5.54620266e-01
2.09135801e-01 -8.75040442e-02 -1.28841490e-01 2.88888961e-01
-9.26642776e-01 8.63595128e-01 4.25011098e-01 -3.68475735e-01
-9.09211576e-01 -9.91952538e-01 -5.63919365e-01 5.60505629e-01
2.56779552e-01 -9.65958118e-01 1.52971280e+00 9.38938558e-02
-1.49496853e+00 7.01955080e-01 -5.00089765e-01 -6.28235340e-01
4.10982788e-01 -6.37985885e-01 1.72257666e-02 3.54615718e-01
4.04898822e-01 6.09426558e-01 2.98030287e-01 -1.31989288e+00
-4.83842552e-01 -8.06831568e-02 -3.05427819e-01 4.78192449e-01
-2.61721522e-01 1.43342167e-01 -1.89054161e-01 -5.56755126e-01
-2.91667640e-01 -5.11637449e-01 -2.13909075e-01 -9.73908544e-01
-1.14623678e+00 -8.22374880e-01 5.91216564e-01 -9.70536888e-01
1.33503401e+00 -1.52845037e+00 4.00578767e-01 -4.43169326e-01
3.15680951e-01 1.82390437e-01 -3.69614184e-01 8.33462358e-01
7.17913061e-02 5.48547924e-01 -5.99527806e-02 -8.61169875e-01
3.16441129e-03 -5.53467393e-01 -8.61343861e-01 -1.54947296e-01
8.41782317e-02 9.44557786e-01 -1.10757864e+00 -4.87705052e-01
-3.37712467e-01 -4.33731824e-02 -3.26498032e-01 1.82579473e-01
-5.51840663e-01 2.44504884e-01 -6.50237739e-01 2.49833092e-01
4.28879499e-01 -3.14290792e-01 -2.31154859e-01 8.19325447e-02
-5.10376155e-01 7.76437044e-01 -5.57226717e-01 1.77851331e+00
-1.89852327e-01 9.03417647e-01 -1.73099697e-01 -6.36220276e-01
9.99329984e-01 4.58346099e-01 1.46581590e-01 -9.81057659e-02
-2.93183029e-01 3.11255276e-01 -4.93157387e-01 -1.09259322e-01
1.45170534e+00 3.94656569e-01 -3.02943408e-01 7.32834041e-01
2.51641721e-01 -5.10957658e-01 8.39269817e-01 1.10291684e+00
1.21377110e+00 1.38976695e-02 4.65275735e-01 -7.26802349e-01
4.17319477e-01 3.73274833e-02 1.24858655e-01 1.40209901e+00
1.07846931e-01 9.75670159e-01 1.06070006e+00 -2.67973721e-01
-1.06208491e+00 -8.71995211e-01 3.52527231e-01 9.66182530e-01
-2.90889829e-01 -8.84113312e-01 -8.24891984e-01 -6.38524175e-01
-2.32974872e-01 1.62165952e+00 -4.65711504e-01 -1.54528156e-01
-5.55270910e-01 -6.65395379e-01 6.49838090e-01 2.75021434e-01
3.59319448e-01 -1.41114223e+00 -5.07266760e-01 1.07687294e-01
-6.71597898e-01 -6.58259809e-01 -6.75237298e-01 -4.92871553e-02
-8.80906343e-01 -4.04279172e-01 -1.20111430e+00 -4.09328461e-01
1.69266984e-01 3.20337325e-01 1.37581491e+00 -3.08786362e-01
1.98986471e-01 1.39344424e-01 -1.54149339e-01 -8.51282358e-01
-5.05118012e-01 7.18032777e-01 -1.26317382e-01 -6.75265968e-01
4.18950543e-02 -2.96348572e-01 -3.47130239e-01 -5.71777463e-01
-6.62954271e-01 2.45221332e-01 6.78725660e-01 6.00102365e-01
4.10557330e-01 -6.06614053e-01 1.10045826e+00 -8.60879958e-01
1.53063595e+00 -3.81035239e-01 -5.84032238e-02 5.64034939e-01
-5.34080684e-01 4.16187197e-02 6.96957111e-01 -5.01562133e-02
-1.17536831e+00 -7.13406444e-01 5.18301055e-02 3.31832170e-02
1.53804481e-01 8.67136776e-01 8.27859938e-02 6.67777836e-01
7.97623515e-01 5.52248478e-01 -3.26433480e-01 -5.65191507e-01
5.05645573e-01 6.71393871e-01 7.54912555e-01 -6.00236595e-01
3.15791488e-01 -2.26507843e-01 -2.71835506e-01 -1.12049544e+00
-1.29423404e+00 -4.37348098e-01 -1.90140501e-01 -8.79662260e-02
7.72025466e-01 -8.87476683e-01 -4.29587394e-01 3.32842708e-01
-1.82049072e+00 1.07549340e-01 -4.45145994e-01 3.11433345e-01
-5.62076867e-01 4.64794397e-01 -7.97954559e-01 -6.34121120e-01
-1.23748600e+00 -9.43865180e-01 1.21177828e+00 8.38942945e-01
-7.35117674e-01 -6.83320403e-01 3.98136139e-01 3.07121634e-01
5.02124310e-01 2.91464776e-01 8.17391515e-01 -1.20145214e+00
-2.56031215e-01 -9.83508006e-02 -2.72090465e-01 5.34111783e-02
-1.38510317e-01 2.39169598e-01 -8.44446898e-01 -2.18434528e-01
-9.42955390e-02 -4.20610011e-01 1.49715054e+00 9.62002099e-01
1.11896443e+00 -6.38211131e-01 -3.76653314e-01 8.02190080e-02
8.58908772e-01 -4.30456735e-02 6.60075128e-01 2.78973430e-01
5.44343948e-01 4.95471269e-01 2.65080690e-01 3.43701005e-01
4.75905806e-01 3.42776000e-01 -1.86117664e-01 1.59176946e-01
-2.83003360e-01 -4.01354223e-01 4.54922706e-01 1.25232887e+00
-5.55383302e-02 -9.25756991e-01 -7.43831933e-01 7.66679585e-01
-1.99629700e+00 -1.13180220e+00 -2.68783361e-01 1.99264312e+00
9.96055782e-01 2.36128747e-01 6.33856878e-02 -5.11400640e-01
5.83923995e-01 7.89774835e-01 -4.22723651e-01 -8.83024573e-01
-2.66321748e-01 6.25694245e-02 4.88973670e-02 4.47821349e-01
-9.33021069e-01 9.56701815e-01 6.67094231e+00 9.37982917e-01
-6.79018736e-01 -2.01908156e-01 7.81075716e-01 -3.11707169e-01
-5.50550759e-01 -8.79418254e-02 -1.02879024e+00 4.94527102e-01
1.51882875e+00 -1.04236460e+00 -1.65387481e-01 6.68368697e-01
3.58403236e-01 -2.96045661e-01 -1.06886530e+00 4.49979484e-01
4.65950698e-01 -1.87016284e+00 6.77521467e-01 -1.28201142e-01
1.05557859e+00 -1.09045627e-03 -1.43257201e-01 6.40655577e-01
5.22309661e-01 -8.81322622e-01 5.81466436e-01 9.60265040e-01
5.06301582e-01 -7.63947427e-01 8.50815594e-01 4.15437132e-01
-4.70027149e-01 5.20272434e-01 -5.56547880e-01 1.62080303e-01
3.29322100e-01 9.40168858e-01 -8.36883247e-01 1.02816284e+00
4.07621950e-01 9.29641902e-01 -6.70053720e-01 1.18363643e+00
-2.19009280e-01 7.95901060e-01 5.49173132e-02 -5.57296276e-01
3.13329756e-01 -4.80106100e-02 1.35115361e+00 1.64551699e+00
4.62571293e-01 -7.57667348e-02 2.06415891e-03 9.93030310e-01
-7.10200548e-01 8.22315216e-02 -5.26321888e-01 -4.84057456e-01
4.34543312e-01 1.38409150e+00 -7.53694713e-01 -9.24921393e-01
3.72216046e-01 8.74787450e-01 2.16319636e-01 4.02628213e-01
-4.06618088e-01 -9.04443741e-01 1.67095885e-02 -4.57411051e-01
9.07139406e-02 -1.36307646e-02 -7.22616255e-01 -1.24422276e+00
-9.72317345e-03 -7.34021842e-01 2.59776503e-01 -1.14909983e+00
-1.05483449e+00 8.39438379e-01 4.02199551e-02 -7.17616975e-01
-3.03204924e-01 6.66680783e-02 -1.19484735e+00 9.92747247e-01
-1.22440100e+00 -9.05955493e-01 9.11565870e-03 -4.71217334e-01
1.14955401e+00 -3.90299439e-01 7.53664732e-01 -4.09653097e-01
-5.54253399e-01 3.38465273e-01 3.55253041e-01 -4.00933832e-01
1.16738617e+00 -1.60134625e+00 8.51452887e-01 9.09590483e-01
-1.85480639e-02 9.13864315e-01 1.26057220e+00 -9.70064461e-01
-1.06648839e+00 -9.73402023e-01 1.58616316e+00 -5.98120213e-01
5.93377888e-01 -1.87942162e-02 -9.24343884e-01 5.89384973e-01
9.37238932e-01 -9.30352151e-01 4.77303684e-01 2.11350441e-01
1.49239469e-02 2.45654017e-01 -7.79726088e-01 7.94558346e-01
4.91965979e-01 -1.35277677e-02 -1.02593565e+00 6.75584078e-01
1.27962732e+00 -5.86918235e-01 -5.98515570e-01 1.83834936e-02
4.19241905e-01 -5.16517878e-01 8.13918352e-01 -8.53003561e-01
1.48372102e+00 -3.42399701e-02 2.38275185e-01 -1.90814579e+00
-3.90102416e-01 -8.48218083e-01 -3.90045345e-01 1.45655894e+00
6.52257681e-01 -3.46799463e-01 4.12364125e-01 2.53204912e-01
-7.31642008e-01 -6.45796061e-01 -6.14334583e-01 -6.06814265e-01
4.61181760e-01 4.44406480e-01 2.91870952e-01 4.44177866e-01
9.58463177e-02 8.51730168e-01 -4.06403899e-01 -4.65743929e-01
5.36104083e-01 1.57928824e-01 7.49132693e-01 -1.23455465e+00
1.38100935e-02 -9.14921880e-01 2.78929442e-01 -9.89443898e-01
1.48369879e-01 -8.76703024e-01 2.89713085e-01 -2.33933020e+00
6.99053824e-01 4.53634501e-01 -2.23254636e-01 1.85825974e-01
-7.64125049e-01 -2.86202252e-01 2.94931263e-01 4.23269719e-01
-1.05729461e+00 7.44277060e-01 1.11477482e+00 -2.75952965e-01
-3.35638434e-01 3.62351276e-02 -1.48656702e+00 3.01324129e-01
6.93728030e-01 -4.30830449e-01 -1.72176436e-01 -3.18001956e-01
5.01632579e-02 2.73669124e-01 1.45790409e-02 -9.58673239e-01
5.05259633e-01 -2.78648268e-03 2.80535728e-01 -1.28026187e+00
-7.85399452e-02 5.13592660e-01 -2.66839266e-01 2.29701191e-01
-1.10273647e+00 2.15438381e-01 3.15184593e-01 4.19805616e-01
-7.40359575e-02 -3.58811110e-01 3.88029486e-01 -3.28556359e-01
-9.27450284e-02 -1.97597519e-01 -4.47385967e-01 2.25327134e-01
4.67596442e-01 2.89033115e-01 -1.09412301e+00 -7.22300887e-01
-3.35956126e-01 4.75840151e-01 2.23182335e-01 4.47058499e-01
5.38887382e-01 -8.53803098e-01 -1.47556365e+00 -5.88039398e-01
-1.01646021e-01 7.88306911e-03 3.42469364e-01 6.23839855e-01
-4.50231105e-01 9.70140994e-01 2.21181363e-02 -2.67384678e-01
-1.02213848e+00 1.02040090e-01 -2.77202159e-01 -1.00051367e+00
-7.41123557e-01 9.01256263e-01 8.70758593e-02 -2.44443938e-01
1.24324732e-01 -9.80498716e-02 -5.47344148e-01 3.18875015e-01
1.15046728e+00 7.46469736e-01 2.65313804e-01 -1.37912974e-01
-1.17530815e-01 7.24551603e-02 -7.21444368e-01 -4.36827749e-01
1.50581956e+00 8.14394578e-02 -3.15190196e-01 7.42775202e-01
9.62913811e-01 1.16325401e-01 -5.87003767e-01 8.03997293e-02
2.09354892e-01 2.40794301e-01 9.40664411e-02 -1.20769703e+00
-3.28943372e-01 6.88448012e-01 -4.34309453e-01 4.07291263e-01
4.67197716e-01 -9.23222378e-02 8.47520232e-01 7.18409121e-01
-1.97199970e-01 -1.06959820e+00 2.01468691e-01 8.27109694e-01
1.46494818e+00 -8.85740817e-01 5.29128432e-01 8.24609175e-02
-9.04621124e-01 1.20446789e+00 4.28549260e-01 -1.40719876e-01
-1.37426600e-01 -3.22071671e-01 -3.33358079e-01 -4.15712059e-01
-1.27429152e+00 6.08763516e-01 6.34468257e-01 1.47178456e-01
7.78967381e-01 9.01251063e-02 -8.99068832e-01 8.55165124e-01
-5.45865059e-01 -6.21718019e-02 1.20081758e+00 6.41098797e-01
-7.20446527e-01 -4.92678374e-01 -5.87735176e-02 8.40117335e-01
-6.02423549e-01 -2.22344860e-01 -8.40538561e-01 3.72059852e-01
-9.12875235e-01 1.13261795e+00 -7.24280477e-02 -8.33197609e-02
4.16532665e-01 1.86990455e-01 -1.86816081e-02 -9.00276005e-01
-8.51881385e-01 6.90895021e-02 3.19113284e-01 -2.25825876e-01
-3.77285555e-02 -7.50897646e-01 -9.44827139e-01 -2.99824953e-01
6.91251382e-02 7.60445833e-01 7.96022773e-01 6.23022020e-01
7.02243865e-01 1.04055929e+00 4.78928387e-01 -1.14129603e+00
-8.38129342e-01 -1.47637236e+00 -4.17168647e-01 -1.00656845e-01
4.45276797e-01 -4.95589115e-02 -3.79152000e-01 6.49685562e-02] | [12.556046485900879, 9.614302635192871] |
79747bee-b5e5-4cec-8d98-76598f0dbffb | specific-investments-under-negotiated | 2303.14515 | null | https://arxiv.org/abs/2303.14515v1 | https://arxiv.org/pdf/2303.14515v1.pdf | Specific investments under negotiated transfer pricing: effects of different surplus sharing parameters on managerial performance: An agent-based simulation with fuzzy Q-learning agents | This paper focuses on a decentralized profit-center firm that uses negotiated transfer pricing as an instrument to coordinate the production process. Moreover, the firm's headquarters gives its divisions full authority over operating decisions and it is assumed that each division can additionally make an upfront investment decision that enhances the value of internal trade. On early works, the paper expands the number of divisions by one downstream division and relaxes basic assumptions, such as the assumption of common knowledge of rationality. Based on an agent-based simulation, it is examined whether cognitively bounded individuals modeled by fuzzy Q-learning achieve the same results as fully rational utility maximizers. In addition, the paper investigates different constellations of bargaining power to see whether a deviation from the recommended optimal bargaining power leads to a higher managerial performance. The simulation results show that fuzzy Q-learning agents perform at least as well or better than fully individual rational utility maximizers. The study also indicates that, in scenarios with different marginal costs of divisions, a deviation from the recommended optimal distribution ratio of the bargaining power of divisions can lead to higher investment levels and, thus, to an increase in the headquarters' profit. | ['Christian Mitsch'] | 2023-03-25 | null | null | null | null | ['q-learning'] | ['methodology'] | [-4.58814859e-01 6.47650659e-01 -4.73300129e-01 1.58472165e-01
-1.36850134e-01 -6.17589414e-01 9.50234011e-02 2.72200089e-02
-5.73132575e-01 1.05737317e+00 -1.41943544e-01 -3.22554350e-01
-6.57284617e-01 -9.54688013e-01 -1.42283395e-01 -8.63223255e-01
2.77389407e-01 7.63669372e-01 -2.82497913e-01 -3.00460875e-01
5.76342285e-01 3.87000203e-01 -1.12096584e+00 -2.09639058e-01
8.53834569e-01 8.33707631e-01 2.49353439e-01 1.77444413e-01
7.50493854e-02 7.11458623e-01 -7.92437255e-01 -4.74597722e-01
7.99214959e-01 -1.62059203e-01 -5.69282353e-01 3.65220100e-01
-5.95329225e-01 -6.20388865e-01 2.60611087e-01 1.04588413e+00
3.89716655e-01 2.97191292e-01 7.63853788e-01 -1.44485271e+00
-7.03002393e-01 1.03386331e+00 -4.85015094e-01 1.03110850e-01
2.99771689e-02 1.89538851e-01 1.02566874e+00 -3.93296257e-02
4.62821931e-01 1.13415992e+00 -6.44370262e-03 1.25437185e-01
-1.13690054e+00 -6.16401434e-01 1.16051950e-01 -1.44496888e-01
-1.34579265e+00 3.69885266e-02 4.50120032e-01 -3.51421624e-01
8.40426147e-01 -1.68258771e-01 6.13741875e-01 -3.30695584e-02
6.21485293e-01 4.73658517e-02 1.43401122e+00 -7.47382820e-01
6.79174960e-01 2.73116022e-01 -4.00914937e-01 -1.32472605e-01
8.87377799e-01 2.07757801e-01 2.01828361e-01 2.80235615e-02
9.53214407e-01 -1.96050122e-01 -1.29353717e-01 -3.14423889e-01
-8.13036263e-01 7.44963884e-01 2.36505404e-01 6.01434469e-01
-1.03535962e+00 2.36608759e-02 -3.61413322e-02 7.41472483e-01
-1.02418773e-01 7.88931966e-01 -7.78470412e-02 -2.33880337e-02
-4.23285604e-01 1.59485385e-01 9.88119304e-01 5.37237763e-01
6.99522197e-01 1.45867422e-01 -8.73473957e-02 5.18192649e-02
2.57276177e-01 3.65364522e-01 2.05433533e-01 -1.81698847e+00
3.16780448e-01 6.13623738e-01 8.48126650e-01 -8.69768262e-01
-1.73406214e-01 -8.88789475e-01 -1.64087892e-01 6.43846929e-01
5.10740101e-01 -5.82661629e-01 -1.12704724e-01 1.44363391e+00
-8.38580206e-02 -7.62016058e-01 4.25796241e-01 1.01217735e+00
-3.24097127e-01 5.75565577e-01 -1.11944713e-01 -7.46360779e-01
1.18944180e+00 -6.31678939e-01 -7.99011111e-01 2.82313854e-01
2.21726999e-01 -5.39759874e-01 4.36527640e-01 5.69616437e-01
-1.26856935e+00 -2.28195056e-01 -6.79093003e-01 6.66952372e-01
-9.06759501e-02 -2.80630022e-01 6.31132841e-01 8.28682005e-01
-9.80166018e-01 5.75590432e-01 -5.61743677e-01 -1.02370478e-01
5.51244728e-02 5.22493243e-01 9.34226736e-02 2.60952920e-01
-1.10431099e+00 1.09358776e+00 4.16294724e-01 4.90079559e-02
-5.59411466e-01 -3.34413439e-01 -3.08039963e-01 5.97062826e-01
7.86026418e-01 -7.28728354e-01 1.50921178e+00 -1.45260465e+00
-1.41111732e+00 2.77170897e-01 4.22736734e-01 -4.93743271e-01
6.45387769e-01 5.37439287e-01 2.82902747e-01 2.62947619e-01
3.63540262e-01 3.39607716e-01 1.09631665e-01 -1.48793387e+00
-1.01321924e+00 -5.87436020e-01 8.24359059e-01 7.26884663e-01
-2.07062215e-02 -1.23015769e-01 5.33557296e-01 -2.20753342e-01
-1.00431085e-01 -9.09054458e-01 -5.15383005e-01 -8.49014938e-01
-4.29880470e-02 -2.26119265e-01 3.78264510e-03 -1.76121779e-02
7.54175067e-01 -1.78533125e+00 -2.31236249e-01 5.81228554e-01
1.67561788e-02 -4.18795377e-01 2.55491048e-01 8.00544858e-01
2.17201531e-01 2.53410518e-01 3.26999068e-01 3.02512228e-01
5.90631127e-01 4.10405725e-01 2.58202821e-01 2.54294783e-01
-3.47944677e-01 5.27196288e-01 -6.29160285e-01 2.64058840e-02
-1.70049056e-01 -4.19704586e-01 -2.66352385e-01 -3.88599724e-01
1.63318574e-01 5.51487617e-02 -5.58940053e-01 5.88650048e-01
5.33986449e-01 -1.01653099e-01 7.98723519e-01 5.42166650e-01
-6.10979378e-01 -8.63277465e-02 -1.27342641e+00 7.02589452e-01
-4.26258773e-01 4.08863604e-01 5.84791541e-01 -1.04181540e+00
8.19681883e-01 4.81994212e-01 4.96752083e-01 -7.09047675e-01
4.73724693e-01 2.48645186e-01 6.63926065e-01 1.99967101e-02
7.49397099e-01 -6.08038783e-01 4.97836769e-02 9.23414826e-01
-3.90219271e-01 7.48013407e-02 2.44326800e-01 -4.04090136e-02
6.74465239e-01 -1.54559121e-01 4.88837749e-01 -5.25543928e-01
1.53848514e-01 1.13827817e-01 9.04712796e-01 6.70489490e-01
-3.08751136e-01 -2.78127044e-01 8.94913316e-01 3.00868243e-01
-7.39679575e-01 -6.95858717e-01 1.88089266e-01 1.00252545e+00
5.49807847e-01 4.87346321e-01 -6.72328770e-01 -1.79591581e-01
3.54854792e-01 1.01043797e+00 -3.85944635e-01 1.44651935e-01
1.97728232e-01 -5.39260745e-01 6.45513386e-02 4.64139640e-01
7.15266347e-01 -7.59463310e-01 -1.12954509e+00 9.39519331e-02
-3.04008927e-02 -5.72569788e-01 -1.63662419e-01 1.35112181e-01
-6.12457514e-01 -9.01087761e-01 -7.09058881e-01 -3.05599600e-01
8.36292267e-01 5.49809754e-01 5.54375589e-01 -1.01404741e-01
3.57474297e-01 4.34834749e-01 -3.90683383e-01 -8.22629273e-01
-5.39905071e-01 -5.19381016e-02 6.74480870e-02 -3.43742043e-01
3.41999322e-01 -2.71940708e-01 -5.45388460e-01 3.63427639e-01
-6.14571631e-01 -3.14799786e-01 7.54972041e-01 4.27351356e-01
1.28074676e-01 9.71955538e-01 1.34725678e+00 -4.85043019e-01
1.09086692e+00 -3.78672898e-01 -8.77941966e-01 2.18749046e-01
-1.15077209e+00 -2.23562308e-02 5.08151829e-01 1.04665151e-02
-1.70704937e+00 -5.90553939e-01 7.73288190e-01 1.49129294e-02
-1.20937623e-01 6.45053566e-01 -3.24894041e-01 1.51296064e-01
1.83554053e-01 -1.25655875e-01 4.57654893e-01 -3.34709547e-02
-1.06580198e-01 7.33429670e-01 5.19842841e-02 -5.59101045e-01
7.71507382e-01 2.30164573e-01 8.88323635e-02 -2.72486091e-01
-6.96956413e-03 -2.08568573e-01 -3.47232878e-01 -4.70284671e-01
6.33530676e-01 -9.28209245e-01 -1.40235019e+00 6.51743039e-02
-5.92594087e-01 -3.37410033e-01 -3.15070391e-01 9.11332369e-01
-8.77940059e-01 2.19235182e-01 -3.79232049e-01 -1.14876211e+00
1.13620371e-01 -1.01988149e+00 -1.63707510e-01 6.58653796e-01
-4.70614582e-02 -9.01671529e-01 -3.97115290e-01 9.21854079e-01
5.18245280e-01 9.98623222e-02 8.25143993e-01 -5.84364951e-01
-9.04623151e-01 1.49620086e-01 1.89379454e-01 2.68673539e-01
3.99111211e-01 5.39703555e-02 -3.17893147e-01 -3.15715522e-01
-5.42125246e-03 -8.79489258e-02 3.72558199e-02 6.34127200e-01
-5.65204676e-03 -4.31414455e-01 -8.24535266e-02 -2.51413107e-01
1.46691072e+00 1.09492958e+00 4.79380071e-01 9.32807922e-01
-4.19362217e-01 1.19381905e+00 1.30051923e+00 8.09423685e-01
6.68892145e-01 5.14745533e-01 5.95932484e-01 2.61327535e-01
6.97308421e-01 7.20574036e-02 2.82628715e-01 1.47975564e-01
-8.87526155e-01 -1.25934510e-02 -5.51114380e-01 5.71783543e-01
-1.80818796e+00 -1.31172895e+00 3.46188515e-01 2.28030562e+00
4.44931716e-01 2.16804177e-01 4.72990572e-01 1.36174276e-01
1.03093207e+00 -4.70040262e-01 -4.38548058e-01 -7.96466529e-01
3.04200109e-02 -3.13590288e-01 8.91139984e-01 4.75896358e-01
-5.97393475e-02 4.71552819e-01 5.70227861e+00 3.49983990e-01
-7.08420396e-01 -3.77225950e-02 7.46832013e-01 -2.84379423e-01
-5.05727589e-01 4.75548953e-01 -3.96450132e-01 3.93594056e-01
6.56607270e-01 -1.12497342e+00 4.74553198e-01 5.49417794e-01
8.07376444e-01 -6.72725439e-01 -6.19887769e-01 1.30730867e-01
-5.49604714e-01 -1.12882674e+00 -5.18895447e-01 8.44909251e-01
1.02066267e+00 -6.70657456e-01 4.70591336e-02 2.63928860e-01
7.56281316e-01 -7.84833670e-01 1.01719582e+00 4.63244289e-01
3.16579267e-02 -1.49778354e+00 1.35840821e+00 7.41001606e-01
-8.24728727e-01 -7.61337876e-01 -4.35895234e-01 -8.62625241e-01
1.50846183e-01 -3.24724726e-02 -1.13428986e+00 8.79736245e-01
3.63891572e-01 -5.71026683e-01 1.75044715e-01 7.81727374e-01
-8.79243612e-02 2.72478461e-01 1.25773028e-01 -1.41044170e-01
3.97347122e-01 -7.93612957e-01 1.16680928e-01 3.70430589e-01
3.68823856e-01 4.50922996e-01 2.74557889e-01 1.02729642e+00
7.47409761e-02 3.56072962e-01 -4.12059456e-01 -3.09368391e-02
1.04053366e+00 9.37748253e-01 -1.08677232e+00 -8.62215757e-02
-1.75826445e-01 2.11888641e-01 -3.69946569e-01 4.21900392e-01
-4.25936460e-01 -5.77588379e-01 4.39643890e-01 3.33248675e-01
2.33682856e-01 1.23825043e-01 -3.44741762e-01 -4.90018994e-01
-3.13771099e-01 -6.02060497e-01 1.18187480e-01 -5.92401803e-01
-8.66125047e-01 -2.09760532e-01 1.36048719e-01 -7.84387827e-01
-4.95084256e-01 -3.43635887e-01 -6.86481237e-01 1.09700525e+00
-1.43510079e+00 -5.85597336e-01 8.89870748e-02 1.97651591e-02
6.98507056e-02 -2.84060329e-01 2.70382673e-01 -3.46031398e-01
-4.02499855e-01 1.62076235e-01 3.97611707e-01 2.22181026e-02
1.99111238e-01 -1.17752743e+00 -8.49439025e-01 5.04364252e-01
-6.77910507e-01 6.88029706e-01 6.05760634e-01 -6.29907906e-01
-1.12141061e+00 -5.23674488e-01 5.72346032e-01 4.01945740e-01
5.22357643e-01 5.04954815e-01 -3.79975200e-01 6.36974096e-01
8.91548038e-01 -9.40878272e-01 7.94147015e-01 -1.12920627e-01
6.31628156e-01 -5.23730814e-01 -1.41745126e+00 5.37553668e-01
5.24242103e-01 -1.23989256e-02 -7.90155053e-01 -1.95428982e-01
2.59545624e-01 2.93910325e-01 -1.10752010e+00 -1.05358511e-01
3.70127976e-01 -1.05118012e+00 4.81120229e-01 -2.33002156e-01
1.03076383e-01 -3.26963961e-01 -5.68731092e-02 -1.38948047e+00
-6.69587076e-01 -6.35494947e-01 7.36210823e-01 1.06605089e+00
3.48514467e-01 -1.19282842e+00 7.63828874e-01 8.03950727e-01
1.85387209e-02 -4.61341470e-01 -1.11611080e+00 -8.25011790e-01
4.88144785e-01 3.77531737e-01 5.82117915e-01 9.32742238e-01
5.27305841e-01 1.10789528e-02 3.70557547e-01 1.08969763e-01
8.14298868e-01 4.65811193e-01 6.80051506e-01 -1.30432594e+00
-3.93979996e-01 -6.31669104e-01 -1.59717664e-01 -5.78894496e-01
1.76044539e-01 -5.21755457e-01 -1.56896263e-01 -1.88859427e+00
3.44696231e-02 -5.94120681e-01 -4.11555111e-01 3.36711437e-01
4.70069587e-01 -3.58414859e-01 6.16198242e-01 3.00183326e-01
-1.55743912e-01 9.47552398e-02 1.52009547e+00 1.56606600e-01
-4.92411345e-01 4.56138492e-01 -1.50221300e+00 5.91285467e-01
9.94844973e-01 -5.58928065e-02 -7.60151982e-01 8.47648233e-02
4.83182877e-01 9.19542849e-01 9.03166947e-04 -2.88291097e-01
2.61671573e-01 -9.64526176e-01 -1.36416899e-02 -2.38482237e-01
-7.67575428e-02 -9.56464291e-01 5.22180200e-01 7.80843318e-01
-2.91207671e-01 1.63211182e-01 -3.40287715e-01 2.56662220e-01
-1.75907388e-01 -8.05814683e-01 4.41920727e-01 -5.01235545e-01
-1.32165909e-01 -5.56607366e-01 -1.10697770e+00 -5.56790471e-01
1.52626491e+00 -6.97763324e-01 -4.45984155e-01 -8.80346954e-01
-6.76900685e-01 4.16687459e-01 8.59994411e-01 -1.45507812e-01
-8.78120121e-03 -7.96595156e-01 -6.60360932e-01 -6.64750338e-01
-3.90024662e-01 -2.61590093e-01 1.31356955e-01 1.00915897e+00
-4.70373690e-01 7.56939769e-01 -6.83643520e-01 1.23759300e-01
-9.13585365e-01 3.44718933e-01 5.91757774e-01 -1.44993961e-01
5.14300652e-02 1.58340037e-01 1.38869807e-01 1.46551738e-02
-2.09349856e-01 -2.02149972e-01 -1.89655796e-01 5.19045293e-01
4.67975140e-02 8.25367153e-01 -1.92965969e-01 -5.06841362e-01
-2.02106107e-02 1.62571505e-01 1.39533430e-01 -5.06984532e-01
1.06890202e+00 -4.66682523e-01 -1.00450389e-01 1.84786022e-01
-7.33935833e-03 1.96784168e-01 -1.35864902e+00 7.88756236e-02
3.05479858e-02 -8.34742069e-01 2.75143664e-02 -8.14827561e-01
-9.37336683e-01 2.28479490e-01 -5.96279763e-02 6.27178192e-01
1.13994563e+00 -3.27975392e-01 -3.30737859e-01 4.70343262e-01
9.91230905e-01 -1.62837791e+00 1.46477099e-03 3.30130309e-02
4.55263108e-01 -7.10450947e-01 -8.03699270e-02 -3.19924176e-01
-1.07376301e+00 7.95103192e-01 4.83456373e-01 7.45049678e-03
3.47349912e-01 1.22975998e-01 1.16712503e-01 2.71026880e-01
-9.16921735e-01 -2.07153633e-01 -5.31412303e-01 7.77402818e-01
1.56071290e-01 5.85902929e-01 -1.02951145e+00 8.18599761e-01
-3.74986559e-01 1.94355205e-01 1.58696139e+00 1.11731732e+00
-8.26693833e-01 -1.01278007e+00 -7.76086330e-01 2.60683000e-01
-7.07445860e-01 3.45078677e-01 -2.98112333e-01 1.28330696e+00
4.47327197e-01 1.11005569e+00 5.86865425e-01 4.32189137e-01
3.95024568e-01 -8.51246119e-02 4.16891247e-01 -5.96941590e-01
-4.37467694e-01 4.51698393e-01 1.22872656e-02 3.14079016e-01
-5.95314622e-01 -8.82026672e-01 -1.78104854e+00 -5.96138418e-01
-4.94591981e-01 7.39916146e-01 4.09282565e-01 8.19469154e-01
1.43849507e-01 3.71107489e-01 9.27489400e-01 -2.33145222e-01
-1.18627977e+00 -7.94068217e-01 -1.32669783e+00 -1.22423753e-01
-3.15693736e-01 -7.00587153e-01 -4.15729195e-01 -2.92649388e-01] | [4.260954856872559, 2.9026646614074707] |
30ae32c7-7bc9-4f5a-81c2-b38f53afa5d7 | amr-to-text-generation-with-graph-transformer | null | null | https://aclanthology.org/2020.tacl-1.2 | https://aclanthology.org/2020.tacl-1.2.pdf | AMR-To-Text Generation with Graph Transformer | Abstract meaning representation (AMR)-to-text generation is the challenging task of generating natural language texts from AMR graphs, where nodes represent concepts and edges denote relations. The current state-of-the-art methods use graph-to-sequence models; however, they still cannot significantly outperform the previous sequence-to-sequence models or statistical approaches. In this paper, we propose a novel graph-to-sequence model (Graph Transformer) to address this task. The model directly encodes the AMR graphs and learns the node representations. A pairwise interaction function is used for computing the semantic relations between the concepts. Moreover, attention mechanisms are used for aggregating the information from the incoming and outgoing neighbors, which help the model to capture the semantic information effectively. Our model outperforms the state-of-the-art neural approach by 1.5 BLEU points on LDC2015E86 and 4.8 BLEU points on LDC2017T10 and achieves new state-of-the-art performances. | ['Xiaojun Wan', 'Tianming Wang', 'Hanqi Jin'] | 2020-01-01 | null | null | null | tacl-2020-1 | ['graph-to-sequence'] | ['natural-language-processing'] | [ 6.94048345e-01 6.45857811e-01 -1.10181637e-01 -2.28670269e-01
-7.30250299e-01 -4.07041818e-01 1.00577652e+00 3.28408092e-01
-4.41347361e-02 8.92364919e-01 7.21103251e-01 -4.76421952e-01
2.78713644e-01 -1.07443428e+00 -7.75609553e-01 -2.96506733e-01
1.20698340e-01 6.77467048e-01 3.71950641e-02 -6.26514375e-01
9.61926356e-02 -2.84193635e-01 -1.24863005e+00 6.08134806e-01
1.09844589e+00 5.65292358e-01 3.14800531e-01 7.87629068e-01
-7.75219083e-01 9.30014253e-01 -7.39994884e-01 -6.78101003e-01
-3.70125562e-01 -1.09655952e+00 -8.34334612e-01 -1.88137591e-01
1.79543681e-02 1.75988674e-01 -6.11761808e-01 1.04731679e+00
5.30448973e-01 2.16929898e-01 8.80196214e-01 -1.18128347e+00
-1.30872238e+00 1.23788428e+00 -4.51391280e-01 -1.85235962e-01
8.16264391e-01 -8.66535604e-02 1.50250709e+00 -8.08604360e-01
6.86595321e-01 1.54781365e+00 1.83143079e-01 9.72795367e-01
-9.57993865e-01 -5.12214601e-01 6.36094213e-01 2.54638553e-01
-1.14726174e+00 1.74444709e-02 7.08496690e-01 -7.65132159e-02
1.33780622e+00 1.96239322e-01 7.34628618e-01 1.36434639e+00
-3.01308520e-02 9.75924253e-01 5.20646632e-01 -5.36851585e-01
2.33222693e-01 -5.42604804e-01 9.82524604e-02 7.41877675e-01
4.26909000e-01 -4.26580101e-01 -5.69992721e-01 -1.14097506e-01
5.69764435e-01 -1.13495797e-01 -3.47093076e-01 2.29648665e-01
-1.27472866e+00 9.92413104e-01 6.27835035e-01 3.86115819e-01
-4.11342084e-01 5.39659262e-01 2.53599435e-01 1.84394330e-01
4.88385260e-01 4.87423152e-01 -1.37391239e-01 -3.03978212e-02
-5.52480102e-01 3.41374993e-01 7.43264735e-01 1.44451988e+00
3.31506073e-01 1.29820719e-01 -6.99154437e-01 6.59090042e-01
4.87604201e-01 3.81171227e-01 5.63379943e-01 -3.37152988e-01
8.77599776e-01 9.00246620e-01 -1.63710907e-01 -9.73214686e-01
-1.72053337e-01 -3.16843539e-01 -1.19696581e+00 -6.73524559e-01
-1.02296825e-02 -2.88520277e-01 -1.12586844e+00 1.66219795e+00
-7.19432756e-02 4.23251808e-01 3.40084404e-01 6.45195067e-01
1.27940750e+00 1.04756916e+00 1.78398430e-01 4.40543108e-02
1.46546626e+00 -1.18906069e+00 -9.33618426e-01 -5.19521654e-01
7.53247559e-01 -4.39015687e-01 9.87187386e-01 -6.60642013e-02
-1.08461106e+00 -5.33630848e-01 -9.30564582e-01 -7.20997108e-03
-3.20270687e-01 -2.36234572e-02 4.70770687e-01 1.06123172e-01
-1.21485817e+00 4.01952416e-01 -4.45015550e-01 -2.66223639e-01
1.50404304e-01 -2.89020278e-02 -1.78980619e-01 -4.63967294e-01
-1.77837050e+00 6.78330064e-01 6.11304522e-01 1.06079392e-01
-7.40877509e-01 -4.85576659e-01 -1.26817262e+00 2.16480523e-01
3.25043261e-01 -1.07588816e+00 1.07369244e+00 -7.65766144e-01
-1.41928339e+00 4.92258400e-01 -3.31552446e-01 -6.37304068e-01
2.14941248e-01 -2.38598093e-01 -4.18731332e-01 1.32596582e-01
6.28216891e-04 9.15319383e-01 5.07993877e-01 -1.11444783e+00
-3.65286708e-01 2.31894981e-02 4.09863554e-02 4.09637570e-01
-9.92723256e-02 -1.63251474e-01 -5.99818051e-01 -9.71520960e-01
-1.16101332e-01 -9.04019713e-01 -4.68086898e-01 -5.58586240e-01
-6.47377014e-01 -6.75404549e-01 5.27631760e-01 -7.14638054e-01
1.38645935e+00 -1.74503243e+00 3.82962376e-01 -3.03962510e-02
2.44153097e-01 2.04455137e-01 -7.37457037e-01 9.03444350e-01
-1.20441556e-01 3.97970647e-01 -3.37590158e-01 -3.39164138e-01
1.63778365e-01 1.33515537e-01 -2.76518553e-01 -1.73230112e-01
5.75391352e-01 1.45556498e+00 -1.25423551e+00 -2.79716790e-01
-1.19218536e-01 6.32139266e-01 -3.11857432e-01 3.78119677e-01
-7.10824609e-01 2.99215049e-01 -6.59254730e-01 8.78019184e-02
2.39732608e-01 -5.76241374e-01 6.33523047e-01 1.70340717e-01
6.81235552e-01 6.79672837e-01 -5.91041207e-01 1.98451507e+00
-5.40865004e-01 4.43776488e-01 -5.17596185e-01 -1.05650783e+00
1.30272222e+00 4.77536798e-01 -2.14683283e-02 -5.82427084e-01
-5.30937836e-02 4.46812175e-02 3.86686139e-02 -1.07560746e-01
7.23316491e-01 4.12652716e-02 -3.05391580e-01 4.81996745e-01
7.01331720e-02 -2.85294384e-01 3.48381191e-01 7.09417462e-01
1.28843689e+00 1.38908774e-01 4.75173652e-01 -9.91611853e-02
6.14427209e-01 -3.70907456e-01 2.22677484e-01 5.68819761e-01
3.96682471e-01 7.03310847e-01 7.32722342e-01 -2.23351881e-01
-8.02561700e-01 -8.69270325e-01 7.95254052e-01 8.76384377e-01
-6.95628449e-02 -8.05432379e-01 -8.75432432e-01 -1.01975811e+00
-2.24252015e-01 1.38439763e+00 -5.96953988e-01 -4.34496731e-01
-5.22930443e-01 -4.64631289e-01 5.86344719e-01 7.06841111e-01
2.43935555e-01 -1.27878571e+00 1.64682403e-01 4.45933938e-01
-6.73968196e-01 -1.30304646e+00 -6.56359136e-01 -4.82320637e-01
-6.30777657e-01 -7.09302187e-01 -9.35414553e-01 -7.69190431e-01
8.74709189e-01 1.30201533e-01 1.65527606e+00 3.10223222e-01
-3.27616464e-03 1.22670852e-01 -9.59522963e-01 -2.85976052e-01
-6.18467629e-01 3.38345677e-01 -4.43864673e-01 1.89341716e-02
3.33134949e-01 -4.80758548e-01 -4.80349034e-01 -8.26830119e-02
-8.20313692e-01 5.18450141e-01 6.80645347e-01 8.14190030e-01
6.24252737e-01 -1.40308127e-01 9.07895386e-01 -1.02831089e+00
9.66971278e-01 -5.49324274e-01 -1.67296305e-01 4.68095988e-01
-5.16189873e-01 4.51338083e-01 7.95845687e-01 -2.18647256e-01
-1.06803429e+00 -9.30732116e-02 -8.72628540e-02 -1.22725852e-01
5.98755255e-02 6.99917555e-01 -3.54270607e-01 7.27842331e-01
3.80088478e-01 5.00146270e-01 -3.36642295e-01 -1.34623468e-01
1.01240456e+00 5.92807174e-01 4.89830613e-01 -6.34049416e-01
5.43597519e-01 -3.70203666e-02 -2.53007803e-02 -6.13716006e-01
-1.06960893e+00 -2.58132309e-01 -3.63690078e-01 -1.20380111e-02
1.14489985e+00 -1.09371769e+00 -3.07293773e-01 2.32146367e-01
-1.60864818e+00 -3.81392449e-01 -1.31097361e-01 2.14030087e-01
-4.79689479e-01 2.82310814e-01 -6.68676436e-01 -8.34509611e-01
-8.45773935e-01 -7.36147344e-01 1.29439628e+00 1.96898863e-01
-2.84218669e-01 -1.21095800e+00 -1.51198208e-01 3.15003455e-01
3.12081486e-01 3.46824348e-01 1.16816401e+00 -7.78455436e-01
-5.29239893e-01 -2.35848323e-01 -3.23350102e-01 1.29787372e-02
2.12454394e-01 -2.31579214e-01 -5.16346037e-01 -1.82468966e-01
-6.95511818e-01 -2.26966128e-01 1.07280207e+00 1.38541654e-01
1.23766184e+00 -5.88252604e-01 -3.82593304e-01 -1.30268245e-03
1.13063121e+00 3.29477638e-02 1.02098393e+00 -3.36103320e-01
1.10974693e+00 6.61781728e-01 3.06527644e-01 1.63989440e-01
8.43722880e-01 5.58335006e-01 4.42246854e-01 9.03558545e-03
-4.67598140e-01 -9.56488550e-01 5.34551144e-01 1.19452262e+00
-1.12390863e-02 -9.97407556e-01 -7.79929161e-01 5.09911299e-01
-2.15416741e+00 -8.12959075e-01 -3.73421460e-01 1.81331873e+00
8.85865450e-01 2.12551102e-01 -2.57565439e-01 -1.46249458e-01
9.63069320e-01 4.54300374e-01 -2.32226267e-01 -3.69468629e-01
-2.34591793e-02 2.46925771e-01 9.39269438e-02 5.13685524e-01
-6.63802743e-01 1.23043132e+00 5.50313759e+00 8.91890883e-01
-4.82695729e-01 -2.29819238e-01 5.88197827e-01 2.71468133e-01
-9.85621333e-01 -6.07156828e-02 -7.33171582e-01 3.90604168e-01
1.05404031e+00 -3.61170024e-01 4.11619395e-01 6.42139852e-01
-5.01593091e-02 4.31072980e-01 -1.11365676e+00 8.58848989e-01
5.29408753e-01 -1.47687531e+00 7.86596298e-01 -9.75034386e-02
8.99138689e-01 -1.95769846e-01 -4.18059021e-01 4.85104024e-01
8.59663367e-01 -1.42237413e+00 4.75970596e-01 7.72741318e-01
9.04230177e-01 -8.32674325e-01 9.29448187e-01 2.12000057e-01
-1.62507820e+00 2.53997445e-01 -3.88146043e-01 -4.98819277e-02
4.70675111e-01 6.28415108e-01 -8.14963758e-01 1.07074833e+00
6.79357499e-02 1.01984501e+00 -4.96295452e-01 3.44062746e-01
-9.75474417e-01 6.68212891e-01 2.67188519e-01 -6.44317210e-01
4.49970990e-01 -8.75492021e-02 3.51481885e-01 1.41936350e+00
5.75128734e-01 2.40687765e-02 2.06807554e-01 1.01114678e+00
-6.06614053e-01 9.70114544e-02 -7.48649180e-01 -4.86326396e-01
5.33401906e-01 1.17544365e+00 -5.47840059e-01 -5.91522932e-01
-4.27969933e-01 1.26821160e+00 4.98527378e-01 3.80789310e-01
-5.58385015e-01 -6.26820207e-01 4.09037650e-01 -1.46623710e-02
3.39649171e-01 -1.48107439e-01 8.07785466e-02 -1.10982692e+00
-5.77757694e-02 -9.17101204e-01 4.89310563e-01 -9.88285184e-01
-1.51148045e+00 7.42544293e-01 -4.05417718e-02 -9.14812505e-01
-6.46401167e-01 -4.77618277e-01 -9.10479605e-01 9.45001364e-01
-1.38253856e+00 -1.19946468e+00 -2.47300789e-01 1.98651046e-01
7.26773143e-01 -3.96045864e-01 1.04751384e+00 -1.92204729e-01
-1.86520979e-01 5.95832527e-01 -2.10448742e-01 4.41844821e-01
2.17736065e-01 -1.42615914e+00 1.47697210e+00 9.26254988e-01
4.54534471e-01 5.69518983e-01 4.47071671e-01 -7.95358121e-01
-1.29561293e+00 -1.42154860e+00 1.27210152e+00 -2.87647188e-01
6.64900124e-01 -6.08641863e-01 -8.30380082e-01 5.85759401e-01
5.37519991e-01 -1.00174040e-01 5.89476049e-01 -5.14638387e-02
-5.47132075e-01 3.52548748e-01 -5.48422575e-01 9.48673844e-01
1.51833725e+00 -5.03610015e-01 -6.09441340e-01 2.91062236e-01
1.50211728e+00 -2.40935326e-01 -5.96701503e-01 9.41989794e-02
1.24292769e-01 -3.18180442e-01 7.40759015e-01 -8.78777146e-01
8.61049056e-01 -2.88428098e-01 -1.16259106e-01 -1.79017699e+00
-2.46831089e-01 -9.81287003e-01 -4.45370883e-01 1.28103518e+00
8.75431061e-01 -4.41323429e-01 6.31689966e-01 1.06734321e-01
-2.38826588e-01 -7.78737903e-01 -6.07472181e-01 -6.31315351e-01
1.39817312e-01 -3.06708038e-01 8.57205987e-01 7.34534621e-01
1.98753655e-01 1.17877519e+00 -3.23679030e-01 -3.17681819e-01
4.95909840e-01 3.26625556e-02 4.79948401e-01 -1.13183606e+00
-3.66861314e-01 -4.18700874e-01 -2.33251080e-01 -1.22758889e+00
6.23034596e-01 -1.34542465e+00 1.07226059e-01 -2.49519181e+00
1.53971642e-01 1.53288189e-02 -8.72456506e-02 5.10610938e-01
-8.87121797e-01 -2.30206400e-01 4.64804173e-01 -3.10781419e-01
-7.83362508e-01 1.02878129e+00 1.44847810e+00 -3.25866640e-01
1.09953605e-01 -2.46916160e-01 -1.05446076e+00 2.64464647e-01
9.39991891e-01 -3.76912773e-01 -8.48519921e-01 -6.15624189e-01
5.28281212e-01 1.13280065e-01 1.06120348e-01 -5.82345724e-01
1.16274450e-02 -2.23735645e-02 6.05887733e-02 -5.27619779e-01
1.68161243e-01 -2.49589115e-01 -7.93688148e-02 4.59845275e-01
-8.41130793e-01 3.17243874e-01 -1.44160450e-01 9.27009463e-01
-1.90713823e-01 -1.11910827e-01 2.21117124e-01 -3.19949150e-01
-4.57460344e-01 3.44762534e-01 -3.82709712e-01 5.36460876e-01
6.56242251e-01 4.07240361e-01 -5.06570518e-01 -1.05749023e+00
-3.80483449e-01 4.72443223e-01 -7.79300481e-02 8.16555560e-01
9.55484509e-01 -1.64894187e+00 -1.34462953e+00 -1.46372885e-01
4.02978837e-01 2.01244280e-01 1.14482693e-01 2.26525888e-01
-1.71678618e-01 4.15785998e-01 2.44949162e-01 -1.28614053e-01
-1.09893596e+00 4.98588771e-01 2.42624190e-02 -7.12278962e-01
-7.07420588e-01 8.48603904e-01 3.10266167e-01 -5.69270551e-01
3.84742790e-03 -1.81605428e-01 -4.68669951e-01 -1.38665691e-01
7.77217448e-01 1.20909721e-01 -1.75220415e-01 -5.38701952e-01
-2.00058252e-01 3.03628623e-01 -1.75140351e-01 -1.44406170e-01
1.06998026e+00 7.63147622e-02 -1.34241968e-01 1.62136495e-01
1.05060172e+00 -3.73127997e-01 -7.66201019e-01 -3.32070708e-01
6.17674626e-02 2.49741673e-02 -3.27297807e-01 -7.91850328e-01
-9.30749476e-01 1.03716981e+00 -2.21024618e-01 1.83155864e-01
7.99856246e-01 1.60221323e-01 1.07550502e+00 4.54023689e-01
1.55436754e-01 -8.84111106e-01 3.99838805e-01 8.16256464e-01
1.23955834e+00 -7.91655302e-01 -2.53153294e-01 -6.66897833e-01
-1.05097723e+00 1.01118994e+00 4.61679667e-01 -7.63243735e-02
2.46775076e-01 5.39389066e-02 -2.54778713e-01 -4.01407667e-02
-1.15057766e+00 -3.91608238e-01 5.98752797e-01 9.00724351e-01
8.30691576e-01 3.63229692e-01 -4.71921116e-01 7.59346604e-01
-3.27894837e-01 -2.58452147e-01 4.46207970e-01 4.76301700e-01
-3.09150726e-01 -1.30137122e+00 2.58016944e-01 4.11765218e-01
-2.60925442e-01 -6.11374140e-01 -8.47534418e-01 5.03572583e-01
-4.02405113e-01 1.25033462e+00 1.09690540e-01 -4.78650451e-01
2.93954790e-01 1.88000262e-01 2.72538006e-01 -8.25491369e-01
-3.75694066e-01 -1.46826893e-01 5.95206797e-01 -3.41747731e-01
-2.08901301e-01 -3.17193002e-01 -1.68605566e+00 -1.77023485e-01
-1.80016115e-01 3.66940022e-01 4.07952189e-01 7.80887306e-01
6.17169023e-01 1.05111647e+00 4.15409029e-01 -4.98178303e-01
-3.84094238e-01 -1.24952614e+00 -3.23890418e-01 5.33413351e-01
-1.35228425e-01 -2.07593530e-01 -5.45017347e-02 -4.26583318e-03] | [10.356513977050781, 8.365818977355957] |
026acdd7-1580-4960-a05d-72e0c25f0bec | comprehensive-evaluation-of-no-reference-1 | 2011.07950 | null | https://arxiv.org/abs/2011.07950v1 | https://arxiv.org/pdf/2011.07950v1.pdf | Comprehensive evaluation of no-reference image quality assessment algorithms on authentic distortions | Objective image quality assessment deals with the prediction of digital images' perceptual quality. No-reference image quality assessment predicts the quality of a given input image without any knowledge or information about its pristine (distortion free) counterpart. Machine learning algorithms are heavily used in no-reference image quality assessment because it is very complicated to model the human visual system's quality perception. Moreover, no-reference image quality assessment algorithms are evaluated on publicly available benchmark databases. These databases contain images with their corresponding quality scores. In this study, we evaluate several machine learning based NR-IQA methods and one opinion unaware method on databases consisting of authentic distortions. Specifically, LIVE In the Wild and KonIQ-10k databases were applied to evaluate the state-of-the-art. For machine learning based methods, appx. 80% were used for training and the remaining 20% were used for testing. Furthermore, average PLCC, SROCC, and KROCC values were reported over 100 random train-test splits. The statistics of PLCC, SROCC, and KROCC values were also published using boxplots. Our evaluation results may be helpful to obtain a clear understanding about the status of state-of-the-art no-reference image quality assessment methods. | ['Domonkos Varga'] | 2020-10-26 | null | null | null | null | ['no-reference-image-quality-assessment'] | ['computer-vision'] | [ 2.29968965e-01 -3.59933168e-01 -8.83348286e-02 -3.71059775e-01
-1.17200077e+00 -3.97817343e-01 3.71176124e-01 3.56915176e-01
-3.55709881e-01 6.80777133e-01 -3.10197920e-02 -1.04208015e-01
-8.13607201e-02 -5.83495796e-01 -5.13006330e-01 -9.06826496e-01
-1.47691593e-01 -2.41674930e-01 2.16944814e-01 -1.11694232e-01
4.75364864e-01 2.98763543e-01 -1.68148947e+00 4.01680887e-01
1.02146959e+00 1.52160871e+00 1.28153890e-01 1.06336713e+00
4.37097818e-01 8.48207891e-01 -1.07795250e+00 -8.47524107e-01
4.78848845e-01 -6.02129936e-01 -4.52610999e-01 1.00142695e-01
7.03864753e-01 -3.58061552e-01 -3.18925411e-01 1.29506850e+00
7.46399701e-01 -1.65768832e-01 7.54778683e-01 -1.32728887e+00
-8.19975972e-01 -5.08395582e-02 -3.14637482e-01 5.31800926e-01
4.84538704e-01 8.00153792e-01 7.88900971e-01 -6.28503621e-01
3.60478073e-01 8.80043149e-01 2.78924853e-01 1.43238723e-01
-1.19986880e+00 -4.54636395e-01 -4.96568888e-01 8.41811776e-01
-1.36876953e+00 -4.62319732e-01 4.96613026e-01 -5.01511633e-01
6.58210218e-01 3.63688499e-01 4.75497156e-01 7.09672570e-01
7.68623173e-01 2.95161754e-01 1.56236923e+00 -5.85748494e-01
3.52222025e-01 2.52777755e-01 -9.04642418e-02 4.20932382e-01
2.35246405e-01 6.03615701e-01 -4.92069483e-01 2.10582823e-01
5.36017001e-01 -6.20043874e-01 -4.84339893e-01 -1.51879281e-01
-1.25806487e+00 4.53473866e-01 4.97799248e-01 2.63719708e-01
-2.74449915e-01 -1.93002522e-01 2.85312623e-01 7.86205649e-01
2.40147799e-01 4.82772648e-01 -2.88686842e-01 -1.66468307e-01
-9.71592724e-01 -1.42061278e-01 3.73476475e-01 6.47103906e-01
7.63953567e-01 1.20301239e-01 -3.28864187e-01 9.40801203e-01
-5.77664338e-02 8.68974805e-01 6.05044782e-01 -1.41128588e+00
3.66186202e-01 2.53209978e-01 2.58248031e-01 -1.23104036e+00
1.13729186e-01 -4.91932184e-01 -1.00910890e+00 1.00365841e+00
5.02105772e-01 3.41027260e-01 -8.16812277e-01 1.09966576e+00
-2.28533477e-01 -1.28212333e-01 1.80073842e-01 1.11436808e+00
7.83082247e-01 6.44469619e-01 -5.95651828e-02 -4.80951548e-01
1.07881916e+00 -7.29847312e-01 -7.07224905e-01 2.42162868e-01
1.19111717e-01 -1.06476891e+00 1.36612082e+00 8.75503600e-01
-1.25165081e+00 -1.28893924e+00 -1.36468613e+00 2.11386636e-01
-2.33322680e-01 3.35135281e-01 -4.21252288e-02 1.05275571e+00
-1.19431531e+00 6.96198344e-01 -3.36237401e-01 5.47828376e-02
2.93438107e-01 1.07433461e-01 -5.41590869e-01 -2.37073824e-01
-9.72790360e-01 9.84269023e-01 2.37281650e-01 -1.01080097e-01
-1.00417817e+00 -6.35141313e-01 -6.12386048e-01 -2.60300368e-01
1.63222492e-01 -3.95321935e-01 9.95361507e-01 -1.32546747e+00
-1.53787744e+00 1.05291915e+00 8.68341476e-02 -2.95184284e-01
5.19659698e-01 4.11921553e-02 -9.62064922e-01 3.80972296e-01
-1.09003223e-01 3.95149529e-01 1.01196754e+00 -1.49693668e+00
-6.50837958e-01 -1.18536063e-01 1.06992543e-01 1.19866937e-01
1.94633201e-01 -1.99718513e-02 -5.92072129e-01 -6.74407482e-01
6.92283139e-02 -4.33174849e-01 2.46498019e-01 2.45568827e-01
-2.06543624e-01 2.41380736e-01 4.09516841e-01 -8.73059630e-01
1.22399890e+00 -2.06146002e+00 -4.06087250e-01 1.39798895e-01
1.34390414e-01 3.93138140e-01 -6.01814032e-01 8.28738287e-02
-1.07552379e-01 2.41598114e-01 -8.12284276e-02 1.33399665e-01
-7.11136311e-02 -1.26351133e-01 1.13649189e-01 6.30969405e-01
2.08862260e-01 5.50026119e-01 -6.71704233e-01 -7.02794790e-01
6.25900984e-01 4.10522938e-01 -2.11127743e-01 4.82761651e-01
2.28896037e-01 3.40781629e-01 3.03206086e-01 7.55078673e-01
7.56616056e-01 -2.01470475e-03 -2.73908347e-01 -7.67087519e-01
8.97004604e-02 -1.90819904e-01 -1.16658807e+00 1.39103794e+00
-5.51007867e-01 8.88984263e-01 -3.11946005e-01 -5.20013750e-01
8.78220260e-01 3.19244444e-01 1.39213845e-01 -1.55841780e+00
-4.15477641e-02 2.71475643e-01 3.05409878e-01 -5.52838504e-01
3.94347608e-01 -3.37864384e-02 3.21782291e-01 1.36532575e-01
2.01640263e-01 -4.44378942e-01 4.18904692e-01 -4.36299816e-02
8.70083213e-01 -2.33253598e-01 4.20552403e-01 2.31701997e-03
8.17102015e-01 -4.35293972e-01 4.30273861e-01 7.51534581e-01
-6.30573630e-01 1.06753814e+00 3.97326618e-01 -2.84558147e-01
-1.34144604e+00 -1.57031250e+00 -2.94343114e-01 5.99300325e-01
3.53246599e-01 -1.47869676e-01 -7.19056129e-01 -4.32739973e-01
-4.44979370e-01 6.39937282e-01 -4.58637685e-01 -1.21423334e-01
-2.17990637e-01 -6.11173511e-01 2.38992602e-01 1.04352556e-01
7.31211603e-01 -1.10945487e+00 -5.01959860e-01 -1.35106131e-01
-3.11006367e-01 -1.12331879e+00 -3.72557640e-01 -4.37278748e-01
-6.70139849e-01 -1.50616682e+00 -8.45200837e-01 -2.92804688e-01
5.00655949e-01 1.09160997e-01 1.54585922e+00 2.23870531e-01
-3.13105434e-01 1.94196895e-01 -4.90769386e-01 -7.77142495e-02
-8.18265796e-01 -4.73962128e-01 -1.44086555e-01 9.15492401e-02
-3.93885113e-02 -3.46819043e-01 -1.00020111e+00 8.09319258e-01
-1.04186893e+00 -2.55128622e-01 7.72617579e-01 6.95655644e-01
9.81762290e-01 5.26681185e-01 2.63793945e-01 -1.74310282e-01
5.57479620e-01 5.48528321e-02 -7.31976390e-01 4.59310949e-01
-1.04750121e+00 -2.09997728e-01 5.53725183e-01 -2.84602046e-01
-9.61269319e-01 -6.70103431e-01 -4.11990248e-02 -3.17775637e-01
-2.86610186e-01 1.94906622e-01 -4.86513197e-01 -1.42484501e-01
9.39410031e-01 1.79205135e-01 -1.93787009e-01 -2.52304077e-01
1.97265178e-01 7.51255512e-01 1.04716539e+00 -2.62285411e-01
9.20876086e-01 1.23912603e-01 -4.70805056e-02 -7.21296191e-01
-4.30759430e-01 -3.78928602e-01 -4.65587199e-01 -5.85038304e-01
7.07720041e-01 -7.91422665e-01 -5.83265245e-01 8.34153891e-01
-7.70286858e-01 -3.05335850e-01 -4.12471443e-01 6.28261805e-01
-6.90923154e-01 4.97011125e-01 -3.64775479e-01 -6.48887336e-01
-3.79480004e-01 -1.30177152e+00 7.94524908e-01 3.53977829e-01
1.85940098e-02 -7.72814870e-01 1.21106461e-01 4.70599085e-01
6.45237863e-01 1.07846908e-01 8.63121450e-01 -1.15267634e-01
-4.00745302e-01 -2.87349224e-01 -4.93781000e-01 1.02305913e+00
3.13162804e-01 2.58032586e-02 -1.05966163e+00 -4.53279406e-01
-1.19526666e-02 -2.76664734e-01 3.78179669e-01 4.41210508e-01
1.20574355e+00 -4.34162855e-01 3.86848152e-01 5.22792101e-01
1.84489608e+00 3.26439768e-01 1.32887924e+00 3.57309848e-01
1.35864019e-01 4.00537074e-01 8.60562265e-01 1.56345576e-01
-1.32428393e-01 9.45251524e-01 4.26376522e-01 -2.33918741e-01
-5.08946955e-01 4.27005403e-02 3.59438330e-01 7.20491230e-01
-2.12060317e-01 -4.67437416e-01 -7.38319218e-01 2.22038075e-01
-9.75161076e-01 -1.16826403e+00 -2.93361768e-02 2.42342353e+00
8.30214143e-01 2.51455516e-01 -1.36267336e-03 7.74846673e-01
6.93019748e-01 1.14788130e-01 -5.15862823e-01 -2.74977177e-01
-5.70817649e-01 2.46444494e-01 3.03592235e-01 2.66730726e-01
-1.07010829e+00 3.22669297e-01 6.62069893e+00 1.14708149e+00
-1.27280569e+00 1.61192954e-01 1.06622064e+00 1.00770835e-02
1.20405696e-01 -3.11765522e-01 -1.25797793e-01 6.67271316e-01
1.23177719e+00 -1.70839861e-01 4.49553877e-01 4.94090587e-01
5.35432935e-01 -7.62942255e-01 -8.23122859e-01 1.51672876e+00
2.20466346e-01 -1.01345146e+00 1.85492821e-02 -1.02898300e-01
8.23386252e-01 -1.18551701e-01 3.86944562e-01 -1.38315335e-01
-1.68939665e-01 -1.20546865e+00 8.18186820e-01 8.79624367e-01
1.35769010e+00 -5.97227931e-01 1.23482645e+00 -4.11725491e-02
-7.88978636e-01 7.86037892e-02 -5.39816856e-01 2.65435755e-01
1.32518010e-02 8.60531747e-01 -2.27403417e-01 8.35402727e-01
8.20305645e-01 3.79050255e-01 -1.23014796e+00 1.45512831e+00
-2.81316519e-01 5.64667404e-01 2.14775071e-01 5.67892253e-01
-3.49083185e-01 -2.03179389e-01 3.36054623e-01 1.09861505e+00
5.24041295e-01 9.02784467e-02 -3.72495264e-01 6.81618392e-01
-2.86478139e-02 2.49911472e-01 -1.62856594e-01 1.46696463e-01
6.61572516e-02 1.04335022e+00 -5.61319292e-01 -2.54583210e-01
-4.24312681e-01 9.89934981e-01 -4.70289588e-01 4.20974910e-01
-7.42695630e-01 -5.13836324e-01 4.73758757e-01 1.54308766e-01
-1.81710813e-03 9.15677249e-02 -1.50583372e-01 -1.08795047e+00
1.54835194e-01 -1.33421147e+00 2.02626169e-01 -1.35188329e+00
-1.41744375e+00 8.64760637e-01 -4.11643013e-02 -1.76239347e+00
-1.37473971e-01 -6.91318333e-01 -4.83320564e-01 1.04886460e+00
-1.58301485e+00 -6.45916283e-01 -7.26455152e-01 4.98890162e-01
4.21418726e-01 -4.28167641e-01 6.01234853e-01 2.84874201e-01
-1.30738541e-01 7.59887397e-01 1.69468015e-01 2.45879337e-01
8.80179048e-01 -1.19986629e+00 4.37109880e-02 9.81305122e-01
1.62188068e-01 2.23064609e-02 8.90109837e-01 -2.81396657e-01
-9.64460075e-01 -9.27687705e-01 3.90602261e-01 -3.64980519e-01
1.98013052e-01 4.37971681e-01 -9.71441865e-01 -1.98505491e-01
3.95534158e-01 3.96290213e-01 5.61405599e-01 -5.46981275e-01
-3.54611546e-01 -6.22249842e-01 -1.38630497e+00 2.13355333e-01
5.39086998e-01 -6.54012382e-01 -3.33149433e-01 -2.59782732e-01
3.84928912e-01 -2.04999849e-01 -1.12647450e+00 4.88981932e-01
6.77435875e-01 -1.72311616e+00 1.03127980e+00 1.06471248e-01
5.32808781e-01 -6.52087450e-01 -5.54056227e-01 -1.40622580e+00
-5.41313365e-02 -9.95755419e-02 1.87105998e-01 1.00790107e+00
4.18440938e-01 -3.17692548e-01 4.32508528e-01 1.71479955e-01
4.18165550e-02 -4.58580256e-01 -9.06913757e-01 -1.04413855e+00
1.79967508e-02 -6.31851673e-01 5.23583770e-01 5.39646685e-01
-4.55407858e-01 -1.45227686e-01 -2.22921342e-01 1.60365433e-01
9.11457598e-01 -2.01065481e-01 8.08499336e-01 -8.64955127e-01
-3.72502595e-01 -4.29469168e-01 -9.05821323e-01 -3.18272918e-01
-3.34457636e-01 -4.55905885e-01 -1.18650764e-01 -1.33103693e+00
2.51606703e-01 -1.61952555e-01 -6.02069616e-01 2.57007615e-03
-1.60146728e-01 7.19946384e-01 3.43263388e-01 3.81352454e-01
-6.52611971e-01 3.94835472e-01 1.40123320e+00 -5.29938221e-01
-8.23803842e-02 6.31308332e-02 -2.96845555e-01 4.44436044e-01
9.67721045e-01 -3.70985061e-01 -4.98592883e-01 -2.26706460e-01
8.45192522e-02 1.59135059e-01 5.54315984e-01 -1.61624575e+00
-6.04583845e-02 -1.99731261e-01 6.77583516e-01 -5.15707672e-01
9.24283415e-02 -7.64629066e-01 4.30439055e-01 3.42476130e-01
-2.91822970e-01 2.25997686e-01 1.01023940e-02 2.59781837e-01
-6.58151746e-01 -2.39852950e-01 1.25773978e+00 3.13551305e-03
-7.82849252e-01 1.66267812e-01 -1.37790898e-02 -1.77412555e-02
8.35724175e-01 -5.34155130e-01 -4.54317331e-01 -6.08142614e-01
-6.96228385e-01 -4.88952756e-01 8.74494553e-01 2.28000537e-01
1.06378031e+00 -1.30663264e+00 -8.87826025e-01 2.07115024e-01
5.06017447e-01 -7.53914058e-01 4.86470789e-01 6.95151389e-01
-6.93751752e-01 2.82480121e-01 -6.41719997e-01 -6.59099042e-01
-1.41662407e+00 6.44560814e-01 6.37352407e-01 -1.15835503e-01
-1.23036213e-01 2.64045119e-01 -6.18340187e-02 9.66595933e-02
2.07739562e-01 -2.33209237e-01 -2.18384981e-01 -2.48351112e-01
8.61180961e-01 6.61839724e-01 3.96712184e-01 -8.00573647e-01
-3.14165279e-02 7.49466956e-01 4.70706314e-01 -3.69787008e-01
8.21747303e-01 -3.98214012e-01 9.86509491e-04 4.08933252e-01
1.29518604e+00 -4.99090403e-02 -1.05890155e+00 -1.02467664e-01
-2.05240458e-01 -9.30597365e-01 2.74110675e-01 -1.41847098e+00
-1.27426124e+00 9.30841386e-01 1.55238247e+00 1.73260346e-01
1.84936333e+00 -2.90785015e-01 2.19403028e-01 -1.03077833e-02
5.28505683e-01 -1.04203594e+00 3.26573491e-01 -1.01520307e-01
1.16383612e+00 -1.46038449e+00 1.28369942e-01 -1.60348609e-01
-6.58613622e-01 1.11877143e+00 4.41856146e-01 2.84493744e-01
4.85566169e-01 -8.78483877e-02 5.46831429e-01 2.13688582e-01
-5.96948326e-01 -9.39683989e-02 6.68485940e-01 1.09543693e+00
3.16213667e-01 -1.19562559e-01 -1.31784305e-01 2.52650082e-01
-2.42311522e-01 1.30960524e-01 6.98931634e-01 3.37541163e-01
-2.83481389e-01 -9.72974360e-01 -6.45129681e-01 5.15834630e-01
-6.99237347e-01 8.76595080e-03 1.52682081e-01 6.16468191e-01
3.91336471e-01 1.48789048e+00 -3.87728997e-02 -5.97024500e-01
5.72297931e-01 -4.32262748e-01 5.01338780e-01 -5.77563196e-02
-3.93902391e-01 -1.47389174e-01 -2.37677023e-01 -8.47107053e-01
-5.89687109e-01 -3.87522012e-01 -6.89586878e-01 -4.89720196e-01
-1.51495516e-01 6.72407374e-02 8.03667784e-01 6.13478541e-01
-6.25085160e-02 4.43800360e-01 7.97710359e-01 -5.68636417e-01
-2.28975207e-01 -9.25777912e-01 -5.84832668e-01 6.27040088e-01
4.77018595e-01 -3.43073219e-01 -6.16978288e-01 4.79475290e-01] | [11.773789405822754, -1.904897689819336] |
ca8ae175-279d-4de5-96ac-79bcc06d4716 | lscp-locally-selective-combination-in | 1812.01528 | null | http://arxiv.org/abs/1812.01528v2 | http://arxiv.org/pdf/1812.01528v2.pdf | LSCP: Locally Selective Combination in Parallel Outlier Ensembles | In unsupervised outlier ensembles, the absence of ground truth makes the
combination of base outlier detectors a challenging task. Specifically,
existing parallel outlier ensembles lack a reliable way of selecting competent
base detectors, affecting accuracy and stability, during model combination. In
this paper, we propose a framework---called Locally Selective Combination in
Parallel Outlier Ensembles (LSCP)---which addresses the issue by defining a
local region around a test instance using the consensus of its nearest
neighbors in randomly selected feature subspaces. The top-performing base
detectors in this local region are selected and combined as the model's final
output. Four variants of the LSCP framework are compared with seven widely used
parallel frameworks. Experimental results demonstrate that one of these
variants, LSCP_AOM, consistently outperforms baselines on the majority of
twenty real-world datasets. | ['Zheng Li', 'Maciej K. Hryniewicki', 'Zain Nasrullah', 'Yue Zhao'] | 2018-12-04 | null | null | null | null | ['outlier-ensembles'] | ['methodology'] | [-3.13261032e-01 -7.51821220e-01 1.17451914e-01 -6.42284676e-02
-1.18229377e+00 -3.56854916e-01 5.95455647e-01 4.85569119e-01
-4.03409451e-01 5.46117544e-01 3.60719077e-02 8.88446420e-02
-2.23473072e-01 -4.69433755e-01 -5.68100989e-01 -8.45146835e-01
-1.92703068e-01 5.45975566e-01 4.02571350e-01 1.83386635e-02
4.57603395e-01 4.43553627e-01 -1.59613252e+00 4.43762034e-01
1.32353890e+00 7.01387107e-01 -4.34821844e-01 3.01355839e-01
2.51867503e-01 5.08833349e-01 -5.87632418e-01 -5.82783930e-02
6.63192391e-01 -5.44956803e-01 -7.00943917e-02 -2.14253187e-01
7.37904906e-01 -1.79884046e-01 -9.42755938e-02 9.00269389e-01
5.56385458e-01 3.17139894e-01 7.17475593e-01 -1.41578913e+00
2.85875779e-02 5.77054322e-01 -4.10625696e-01 5.23739755e-01
4.51373816e-01 4.11758155e-01 1.00061095e+00 -1.50326335e+00
5.31250358e-01 9.43930447e-01 1.04809356e+00 1.47611007e-01
-1.54109371e+00 -7.76129246e-01 2.52879024e-01 2.88395714e-02
-1.85256135e+00 -4.30699110e-01 5.90625465e-01 -4.35475975e-01
1.15133297e+00 4.24836755e-01 3.98451835e-01 1.28576112e+00
4.90390986e-01 5.09692073e-01 1.06238544e+00 -1.53076753e-01
6.54849410e-01 -1.62123397e-01 3.70527744e-01 3.10381085e-01
1.09899008e+00 2.39650846e-01 -8.52679431e-01 -1.03554130e+00
1.71979219e-01 2.67224967e-01 -1.72752306e-01 -5.62549055e-01
-1.14116681e+00 5.70262551e-01 3.81171703e-01 3.24845374e-01
-5.56666732e-01 -8.66202265e-02 4.88482445e-01 3.91236901e-01
5.03095865e-01 8.08430314e-01 -3.11574250e-01 -7.17156939e-03
-1.32835090e+00 6.55958533e-01 7.52866626e-01 6.39665902e-01
6.43058121e-01 1.80934288e-03 -3.02119732e-01 6.04095101e-01
4.66597408e-01 3.22094113e-01 3.79780650e-01 -4.18303311e-01
6.95260942e-01 9.52525616e-01 2.14711532e-01 -8.57634008e-01
-5.15101612e-01 -6.81160271e-01 -7.23748565e-01 1.90481305e-01
2.46518299e-01 -2.21290197e-02 -9.86839414e-01 1.19620085e+00
3.66193950e-01 8.75403404e-01 3.90343592e-02 6.72868550e-01
4.93900955e-01 2.97724158e-01 1.47018358e-01 -2.44410023e-01
5.01633108e-01 -6.98866546e-01 -1.58570439e-01 -1.32699504e-01
6.40106738e-01 -5.58039129e-01 6.97174251e-01 5.25699556e-01
-5.47166467e-01 -3.32116216e-01 -1.10738289e+00 6.19576097e-01
-3.71287376e-01 5.71463369e-02 1.36276022e-01 3.75315189e-01
-8.98281336e-01 7.75345206e-01 -1.00367796e+00 -4.30227220e-01
1.93003207e-01 5.04708827e-01 -5.15495300e-01 -7.71698654e-02
-7.72587776e-01 7.29218960e-01 5.70964277e-01 1.14985213e-01
-9.14967000e-01 -5.63978076e-01 -6.49636269e-01 -2.26845846e-01
3.53542536e-01 -8.45235288e-01 9.34774041e-01 -5.49458742e-01
-9.29024875e-01 4.68492150e-01 -2.08378360e-01 -5.82119226e-01
8.95427227e-01 -7.36515701e-01 -7.09131420e-01 -3.69921148e-01
3.59484971e-01 -4.47091199e-02 8.97401989e-01 -1.51637721e+00
-7.51905978e-01 -4.06475008e-01 -7.14383066e-01 2.31387779e-01
5.06060719e-02 6.31188899e-02 -4.07406926e-01 -6.43904924e-01
6.59039736e-01 -9.77513194e-01 -6.93081498e-01 -6.92982554e-01
-7.74121702e-01 -2.25760162e-01 1.09942460e+00 -7.93958977e-02
1.81073475e+00 -2.23382854e+00 1.85407884e-02 9.78452742e-01
3.21460694e-01 2.73963422e-01 -2.22179383e-01 7.60985434e-01
-1.88591108e-01 7.36759156e-02 -2.03431174e-01 -5.31597674e-01
-1.07965536e-01 1.72717661e-01 -6.32900238e-01 6.68341696e-01
2.10213855e-01 4.47442889e-01 -1.16583693e+00 -2.26056933e-01
3.51471484e-01 9.65977553e-03 -6.87125504e-01 1.52108580e-01
1.23643816e-01 5.40232956e-01 -4.60576355e-01 9.32763040e-01
5.75472534e-01 2.79050171e-01 -1.50576249e-01 3.44164968e-01
5.37684932e-02 2.16179237e-01 -1.65887892e+00 1.18309915e+00
3.37585099e-02 9.16574076e-02 -5.01810849e-01 -3.95418465e-01
1.07697415e+00 1.39938846e-01 7.15475321e-01 -2.59024471e-01
-2.76793897e-01 8.38809431e-01 1.91434532e-01 -9.82650220e-02
5.45225561e-01 3.52739573e-01 -2.76440978e-01 3.20734501e-01
4.00036909e-02 1.08318314e-01 6.09659433e-01 1.80376709e-01
1.63668716e+00 6.36575744e-02 5.01575589e-01 -2.18114480e-01
3.53437424e-01 -1.16680473e-01 1.16349888e+00 1.40947664e+00
-4.80564088e-01 1.03920233e+00 4.70339209e-01 -6.49129808e-01
-9.28968966e-01 -1.59660912e+00 -3.26879591e-01 6.76097095e-01
1.49603030e-02 -1.01232064e+00 -4.98144865e-01 -1.01259768e+00
4.89170641e-01 8.84450972e-01 -4.46456730e-01 -3.50519568e-01
-5.70299089e-01 -9.73634064e-01 5.26625752e-01 4.06133592e-01
2.32591212e-01 -9.95331526e-01 -4.35240686e-01 3.40470105e-01
-9.59178209e-02 -7.72055268e-01 -1.49074048e-01 3.80989403e-01
-1.19913745e+00 -1.24096072e+00 -4.78967503e-02 -6.27309754e-02
7.19509006e-01 2.51725465e-01 1.41012204e+00 1.96512252e-01
4.80594486e-03 6.10669255e-02 -4.56922323e-01 -3.75639349e-01
-2.76605994e-01 3.50849070e-02 6.71377242e-01 8.57105926e-02
6.78128958e-01 -6.34099007e-01 -3.49087417e-01 2.73235440e-01
-7.12096274e-01 -7.84141421e-01 5.95339358e-01 9.29126322e-01
7.83498228e-01 -2.88764909e-02 5.16497910e-01 -1.07957256e+00
6.82531416e-01 -7.78262019e-01 -3.77926767e-01 8.79078284e-02
-5.99998236e-01 -1.59681857e-01 7.76785493e-01 -1.40319049e-01
-3.88907164e-01 1.50015756e-01 2.75372773e-01 -6.70600235e-01
-3.16983014e-01 4.40579623e-01 -1.36825636e-01 1.59535438e-01
1.14770710e+00 1.01434916e-01 -4.85025913e-01 -4.37735319e-01
-6.71059415e-02 3.53665531e-01 5.36561310e-01 -6.67221785e-01
1.10651398e+00 2.46405780e-01 -2.71358788e-01 -6.60817564e-01
-5.38774252e-01 -1.12751377e+00 -8.17218363e-01 -8.75515044e-02
1.63670152e-01 -1.06336844e+00 2.68955171e-01 5.81410050e-01
-8.47821236e-01 5.52632399e-02 -4.05770212e-01 5.02351940e-01
-3.93336952e-01 2.88584083e-01 -2.98908114e-01 -8.25977445e-01
-2.22204655e-01 -1.26221585e+00 1.17784810e+00 1.56812608e-01
-6.20512247e-01 -5.74049652e-01 6.04925394e-01 -4.42125052e-02
1.86706707e-01 6.95469797e-01 3.60614896e-01 -1.38004982e+00
-4.86246884e-01 -7.09674418e-01 2.60110557e-01 2.26254359e-01
-5.38989305e-02 4.69178528e-01 -1.01591754e+00 -5.21252096e-01
-2.25112870e-01 1.47310659e-01 1.12947738e+00 1.61219344e-01
9.02142286e-01 2.35156640e-02 -7.53574491e-01 5.98738730e-01
1.55517471e+00 -2.16843799e-01 5.48787594e-01 4.74922240e-01
6.39307201e-01 -2.34473959e-01 6.93448842e-01 7.12208092e-01
-6.25034282e-03 6.04513347e-01 3.33045006e-01 1.79254442e-01
3.35264295e-01 -2.55288750e-01 8.33005250e-01 8.56281936e-01
-2.40881234e-01 -1.00989841e-01 -1.37894607e+00 5.27577043e-01
-2.22966290e+00 -1.16567075e+00 -2.16792762e-01 2.64877152e+00
3.37321699e-01 3.26826692e-01 1.86781108e-01 8.80357176e-02
5.01117587e-01 5.84669337e-02 -5.49596131e-01 -1.05046526e-01
-2.11978987e-01 4.59433533e-02 3.66861045e-01 2.16626555e-01
-1.49541974e+00 6.34712815e-01 6.49638748e+00 4.98729646e-01
-7.18020201e-01 3.14111076e-02 2.83235520e-01 -3.57296914e-01
7.93414861e-02 1.65055379e-01 -9.35494959e-01 7.45482028e-01
1.20311701e+00 -1.37561873e-01 -2.14608777e-02 1.00236106e+00
3.36889952e-01 -3.23174000e-01 -1.37463915e+00 9.03960109e-01
1.35838792e-01 -8.66543174e-01 1.17360555e-01 6.12561591e-02
1.25812840e+00 5.74027359e-01 3.36989947e-02 5.78380823e-01
6.48506343e-01 -7.56829619e-01 5.63839912e-01 8.22740912e-01
2.18576625e-01 -9.11566436e-01 9.74624336e-01 3.11406583e-01
-1.11010110e+00 -5.64808130e-01 -3.15789044e-01 2.10181892e-01
-1.88342676e-01 9.63085175e-01 -7.47810066e-01 7.86942244e-01
1.05817854e+00 7.02231586e-01 -1.00615656e+00 1.70849812e+00
-8.81286040e-02 7.54419565e-01 -8.17398071e-01 4.45511222e-01
2.10653886e-01 -2.57937580e-01 1.05138195e+00 1.27357507e+00
5.58321357e-01 -3.71761769e-01 7.44167209e-01 5.98400831e-01
2.15928592e-02 2.09719405e-01 -1.08636534e+00 4.88978893e-01
7.87603319e-01 9.50981617e-01 -7.64718413e-01 -2.62314081e-01
-2.43918136e-01 7.86300957e-01 4.35799152e-01 4.44364160e-01
-8.04785430e-01 -1.72184303e-01 8.37557316e-01 3.28362733e-02
9.35141817e-02 -9.70196500e-02 -4.78838027e-01 -1.48098993e+00
3.35364372e-01 -1.40399027e+00 7.57894158e-01 -1.74538553e-01
-1.68517661e+00 5.46986878e-01 -1.39556199e-01 -1.85911787e+00
-1.84194922e-01 -4.02074039e-01 -1.20368111e+00 5.76315582e-01
-7.59937346e-01 -9.12688017e-01 -4.12570059e-01 4.15764362e-01
3.42964441e-01 -3.32582623e-01 7.78015971e-01 1.87137470e-01
-1.09323835e+00 7.37005293e-01 6.15878522e-01 -2.12563574e-02
1.11337721e+00 -1.35490191e+00 5.29937506e-01 1.40756679e+00
5.18457949e-01 8.41423690e-01 7.87519157e-01 -9.57251072e-01
-1.00914431e+00 -1.50836241e+00 6.78023219e-01 -1.00814676e+00
4.97004896e-01 -1.50066748e-01 -1.11296320e+00 6.89020634e-01
-3.27406198e-01 3.03732634e-01 7.70763636e-01 3.25889796e-01
-4.97723877e-01 -3.17063004e-01 -1.14477396e+00 6.95746839e-01
9.71283674e-01 -3.95279169e-01 -6.19980395e-01 3.28151822e-01
2.00417694e-02 -5.25375903e-01 -6.94545329e-01 6.88439369e-01
2.68761870e-02 -1.38756156e+00 9.24652994e-01 -5.72971880e-01
-3.61407623e-02 -8.51115048e-01 -2.83785015e-01 -1.51642942e+00
-3.00891489e-01 -5.26708245e-01 -3.27161103e-01 1.05478179e+00
4.72162038e-01 -9.77109611e-01 4.10246432e-01 3.94891620e-01
-3.22941452e-01 -7.09227324e-01 -1.19840360e+00 -9.66614425e-01
-1.15014382e-01 -5.25953174e-01 6.00698352e-01 8.33121121e-01
-5.29695675e-02 -1.11365244e-01 -5.02456464e-02 6.03420079e-01
7.77777970e-01 -2.07502037e-01 1.22928560e+00 -1.48501885e+00
-6.10441752e-02 -4.11464900e-01 -7.94366896e-01 -1.90184206e-01
-1.06187731e-01 -7.77208447e-01 1.57798633e-01 -1.13697946e+00
1.58069745e-01 -5.87328315e-01 -1.05783737e+00 3.13438505e-01
-6.32068753e-01 3.03029120e-01 5.02954684e-02 6.95402563e-01
-1.13483357e+00 3.98468465e-01 2.36409351e-01 2.28031129e-01
-5.48592269e-01 1.26257176e-02 -3.52980644e-01 8.90532076e-01
7.42934108e-01 -7.40845084e-01 -2.01160833e-02 5.75933754e-02
-1.00772917e-01 -6.03196084e-01 2.99013346e-01 -1.84243941e+00
3.34484100e-01 -1.15677200e-01 6.60636663e-01 -8.22831929e-01
-1.23970464e-01 -7.85877705e-01 3.01309198e-01 5.41524947e-01
6.07971251e-02 6.45551324e-01 9.50711444e-02 8.00482452e-01
-3.69024038e-01 7.29459301e-02 7.06907511e-01 1.95279121e-01
-7.76498735e-01 2.07858533e-01 -1.37109786e-01 1.28089011e-01
1.40923035e+00 -3.00043970e-01 -2.13400543e-01 9.44795460e-02
-6.06579244e-01 4.14948463e-01 9.04667675e-01 5.60617805e-01
7.01661646e-01 -1.32422423e+00 -9.19650972e-01 4.98480886e-01
5.85422933e-01 6.36742264e-02 3.74989808e-02 1.07743931e+00
-4.82668132e-01 1.54003277e-01 9.29210261e-02 -9.16648984e-01
-1.04646444e+00 2.99570978e-01 5.38350224e-01 -6.25566661e-01
-7.87828207e-01 5.72335362e-01 -2.60574073e-01 -6.06751978e-01
3.22348997e-02 -3.17083359e-01 2.31312990e-01 -1.62672698e-01
4.20513719e-01 5.99225879e-01 4.43270206e-01 -8.13352287e-01
-5.39895952e-01 2.47473761e-01 -2.43523359e-01 2.69270718e-01
1.21042132e+00 3.14601928e-01 -3.16101789e-01 7.40262270e-01
5.78323066e-01 2.50343263e-01 -9.79521632e-01 -1.86077103e-01
5.47936022e-01 -5.90054631e-01 -2.92433053e-01 -5.52021801e-01
-5.84254682e-01 1.75686434e-01 5.07376671e-01 -6.13370761e-02
1.00974977e+00 -4.52308327e-01 3.78840387e-01 5.50988436e-01
5.26080906e-01 -1.24570012e+00 -5.64972311e-02 6.62096202e-01
7.78763890e-01 -1.23409235e+00 2.62514234e-01 -4.04506661e-02
-5.02128720e-01 8.35357666e-01 1.13714063e+00 -6.74942851e-01
5.87873459e-01 1.15361333e-01 2.01232538e-01 -1.64002523e-01
-9.79326606e-01 -1.68610495e-02 4.11829472e-01 3.17865789e-01
1.87084213e-01 5.39840981e-02 -2.19720677e-01 4.00920391e-01
-1.61203556e-02 -2.79651374e-01 3.56890112e-01 9.13571775e-01
-4.15227503e-01 -1.04607224e+00 -7.60769129e-01 1.00484896e+00
-2.85724878e-01 1.12095840e-01 -5.60132146e-01 7.26548254e-01
4.89377558e-01 9.22276139e-01 1.42474666e-01 -5.83233953e-01
6.28306866e-01 5.08678615e-01 -1.19014017e-01 -8.26344311e-01
-9.94106174e-01 -1.18348233e-01 -1.36564717e-01 -8.81673038e-01
1.68579876e-01 -1.01388741e+00 -9.33991730e-01 -1.58711076e-01
-6.24146163e-01 2.50983328e-01 9.92962420e-02 8.00142586e-01
5.87330341e-01 2.45790496e-01 6.41545475e-01 -9.67470109e-01
-8.92897367e-01 -9.27866399e-01 -5.69894373e-01 7.75391996e-01
3.05041164e-01 -7.84487605e-01 -6.72175407e-01 -5.40224612e-01] | [7.586382865905762, 2.691741466522217] |
744229a7-f64f-41c7-a797-30cce148808d | neural-program-repair-systems-challenges-and | 2202.10868 | null | https://arxiv.org/abs/2202.10868v2 | https://arxiv.org/pdf/2202.10868v2.pdf | Neural Program Repair: Systems, Challenges and Solutions | Automated Program Repair (APR) aims to automatically fix bugs in the source code. Recently, as advances in Deep Learning (DL) field, there is a rise of Neural Program Repair (NPR) studies, which formulate APR as a translation task from buggy code to correct code and adopt neural networks based on encoder-decoder architecture. Compared with other APR techniques, NPR approaches have a great advantage in applicability because they do not need any specification (i.e., a test suite). Although NPR has been a hot research direction, there isn't any overview on this field yet. In order to help interested readers understand architectures, challenges and corresponding solutions of existing NPR systems, we conduct a literature review on latest studies in this paper. We begin with introducing the background knowledge on this field. Next, to be understandable, we decompose the NPR procedure into a series of modules and explicate various design choices on each module. Furthermore, we identify several challenges and discuss the effect of existing solutions. Finally, we conclude and provide some promising directions for future research. | ['Bin Luo', 'Jidong Ge', 'Chuanyi Li', 'Wenkang Zhong'] | 2022-02-22 | null | null | null | null | ['program-repair', 'program-repair'] | ['computer-code', 'reasoning'] | [-0.05321573 0.191861 -0.5353088 -0.35074255 -0.7286398 -0.33887127
-0.09774721 0.15763718 0.15543945 0.53853405 0.07489485 -0.61450356
0.19576462 -0.7736008 -1.1116349 -0.13514072 -0.00898371 -0.37688515
0.11187957 -0.27539384 0.48292577 -0.05001323 -1.4239157 0.39636245
0.9998652 0.6039018 0.32473534 0.49847057 -0.19941562 1.1177436
-0.9461697 -0.6624966 -0.18642072 -0.31897286 -0.93268263 -0.56903726
0.22858682 -0.34993985 -0.5104187 1.5393279 0.3151166 -0.3285179
0.10373767 -1.2051448 -1.3095971 0.90449244 -0.41183943 0.23434599
0.56733793 0.19658034 1.1510241 -0.7876232 0.38677844 0.9904155
0.9708603 0.6515229 -0.87042207 -0.3861637 0.2740777 0.28500777
-1.1827017 -0.1836105 0.91607225 -0.5369667 1.5939256 0.09508027
0.461187 1.1891474 0.741326 0.80561566 0.5017949 -0.45349094
0.15879907 -0.13995612 0.66320294 1.0671355 0.4084342 0.2456449
-0.27607572 -0.1496597 0.39916286 -0.13673542 -0.41946363 -0.05296808
-0.865606 0.9109188 0.5673662 0.25920954 -0.10593642 0.47199437
0.70558405 0.4296135 0.10980594 0.68690854 -0.343552 -0.2380043
-0.7460688 0.28164712 0.73713416 1.0843965 0.46980175 0.6099621
-0.1738547 0.813734 0.36228096 0.11305877 0.6525923 -0.5582268
0.6531863 0.81344604 -0.17388411 -1.3416226 -0.24866064 -0.2912286
-0.778361 -0.10478557 -0.18112554 -0.15571693 -0.5692241 1.6360213
-0.34613982 0.08364254 -0.04988154 0.76710546 1.2048115 0.8570772
-0.29223704 0.16230616 1.2001277 -1.2854073 -0.74911046 -0.38222995
0.72775304 -0.48351502 1.2571547 0.37979865 -1.0391034 -0.59940255
-1.270167 -0.29029617 -0.32224968 0.55720466 0.7158672 0.70625895
-1.2564424 0.56460696 -0.9616497 -0.3071554 0.19922471 0.32162517
-0.19641936 -0.08943658 -1.1108061 0.9538779 0.44070676 0.36421737
-1.2457325 -0.5500032 -1.1261328 0.20340261 0.3095956 -0.54149306
1.7946309 -0.8620941 -1.4853235 0.67484224 -0.25101987 -0.6418015
-0.21802008 -0.22217995 -0.45766363 -0.35942298 -0.20751755 0.42669323
0.4105074 -0.932251 -0.4817159 0.11991496 0.7427592 -0.4381832
-0.36114663 0.3631611 -0.34625202 -0.74088925 -0.23726965 -0.79714364
-0.14519466 -0.43276647 -0.5423849 -0.3162753 0.4168542 -0.97950804
2.0263903 -2.240975 0.30334264 -0.3277097 0.47636336 0.40159273
-0.31103572 0.6600674 -0.42873332 0.2923023 -0.27863303 -0.09562572
0.26647392 0.10196088 -0.67245805 0.32133985 0.48929068 1.1502236
-0.8335673 -0.08056559 -0.2626081 0.16822122 -0.8592177 0.3472522
-0.66136515 -0.12669562 -0.30756527 1.031439 0.45491984 -0.1671289
-0.1122776 0.03938538 -0.3480876 0.7227411 -0.68394774 1.5681152
-0.46239042 0.7739956 -0.04694 -1.2292392 1.165965 0.2762416
-0.14855023 -0.5876455 0.04432822 0.39673164 0.129169 -0.9561644
0.76378244 0.37954536 -0.4318419 0.3240674 0.04908695 0.3189121
0.16323777 -0.11248837 1.6344385 0.10436822 0.507461 0.01396786
0.60289603 0.20225891 0.77219087 0.7385992 -0.1549082 0.44650832
0.9547394 -0.7429786 -0.93417144 -0.64446884 0.271097 0.68529123
-0.15485983 -0.7795771 -0.9548326 -0.81464505 -0.2540602 0.6255162
-0.572478 -0.47631183 -0.80164 -0.82823044 0.89918065 0.6204358
0.5169865 -1.2527423 -0.53322184 -0.00846785 -0.3327901 -0.43460223
-0.2676339 0.21625336 -0.90476096 -0.96782434 -0.4283637 -1.2704934
0.7414655 0.3232677 1.3220584 0.7016686 -0.17797321 0.11136922
-0.7128669 -0.09305968 -0.6745148 0.3046966 -0.14768825 -0.6910503
0.50932056 -0.32082134 -0.02887959 -0.10801737 -0.83646405 -0.10392258
0.7427682 0.91510355 0.34644577 0.33739582 0.6857389 -0.782783
1.1119727 -0.71295714 -0.89225686 0.45414993 -0.51842356 0.06178696
0.87369806 -0.12246682 -0.9049828 -0.14865877 -0.5818116 -0.22733538
-0.00894115 1.1752652 -0.17984176 -0.2848694 0.7385229 0.19048649
-0.3897769 -0.4781481 0.01305835 0.8186959 0.4541172 -0.6536489
0.6464767 -0.34811866 -0.58417773 -0.29207888 -0.7759329 0.11022112
-0.2702723 0.00903609 0.58659905 -0.7362173 -0.64777356 0.34778252
-1.6736298 -0.24948838 0.10902081 0.01028145 -0.34311745 0.48577687
-0.89532566 -0.42727283 -0.4132703 -1.6607813 0.74956876 0.25687256
-0.37254354 -0.88639903 0.17777137 0.16155833 0.44985554 -0.01336614
1.339512 -0.33332798 -0.89566875 -0.35634017 -0.13079365 0.4784962
0.04508226 0.16557297 -0.8103085 -0.2780086 0.2729249 -0.19489847
0.7239167 0.24851944 1.4560419 -0.6089582 -0.36463308 0.5449681
1.6411096 0.42843315 0.81640804 0.442031 0.6887562 0.4479092
0.5106383 0.07407711 0.5625551 0.5424867 0.6952115 0.24767978
-0.15380523 -0.44595018 0.9271685 1.1348699 0.25131217 -0.29351363
-1.203597 0.42630002 -1.7831646 -0.6691084 -0.31373462 1.9929354
0.8568394 0.05605876 -0.12537113 0.08766621 0.7158609 0.0212555
-0.36286983 -0.7331991 0.19351637 0.15423484 0.11933587 0.44476622
-0.95297116 0.9763221 6.886136 0.5130306 -1.1458635 0.15860493
0.31289914 0.2226099 -0.3507953 0.15078756 -0.8109204 0.5227659
1.1293811 -0.16912192 0.62577635 1.2504653 0.03684419 -0.04073936
-1.3572037 0.86754566 0.2437699 -1.4909078 -0.2790012 -0.27689573
0.57275987 0.03067705 0.11524366 0.8268561 0.2354482 -1.0719919
0.64693105 0.41249302 0.5396983 -0.7339924 1.0520248 0.35849908
-0.91422284 -0.5421182 -0.6217718 -0.39483607 -0.08853017 0.74700814
-0.3764833 0.5886781 0.92172444 0.9186274 -0.85527664 1.3345399
-0.60845345 0.71943414 0.38206086 -0.28938788 0.01977823 0.27105755
0.3549309 1.2107617 0.4704428 -0.2753893 -0.00983923 1.4376457
-0.2235916 -0.32892123 -0.734202 -0.29514444 0.4172186 0.83061975
-0.37251517 -0.04068417 -0.7372531 0.7590059 0.521491 0.19242401
-1.1250331 -0.7393766 0.4337199 -0.17844024 0.04697481 -0.26134372
-0.45914972 -1.2425601 0.35069904 -1.3471898 -0.04634919 -0.82728714
-0.94751936 0.63631064 -0.09672394 -1.2087748 -0.06665944 -0.6290779
-0.71477336 0.6781761 -1.4482038 -0.7775789 -0.14665475 -0.30146873
0.5319248 -0.27168414 0.9713025 0.5205582 -1.1223799 0.86913294
-0.11697479 0.28623095 0.33338448 -0.9421191 0.84074736 1.3403183
-0.2565535 1.1873734 0.58101875 -0.83311766 -1.8784618 -1.2860785
1.103652 -0.4155845 0.6559326 -0.37499845 -1.1129593 0.94005066
0.46304598 -0.24005418 0.4772818 0.1451068 -0.44148695 -0.12353303
-0.8991551 0.65450066 0.7659718 -0.54555845 -0.6869761 0.14730518
1.1794891 -0.7794508 -0.6406816 0.37655985 0.05903256 -1.0670973
0.55384094 -0.30756974 0.9169884 -0.58499074 -0.11326847 -1.2049822
-0.41346994 -0.51270264 -0.39610454 1.2700498 0.34452885 -0.4713845
0.5902393 0.41046447 -0.924918 -1.0015849 -0.64316595 -0.7238976
0.01796892 -0.52486616 0.66016483 0.79558086 0.4294881 0.23982225
-0.49345052 0.36076605 -0.13023295 0.18755212 0.5337839 -0.9098685
-0.55470514 -0.46829733 -0.15173416 -0.9801457 0.50471574 -1.1576706
0.30695835 -1.5086066 0.25106296 -0.07580555 -0.03545702 0.9727402
0.05049247 -0.16691066 -0.14525275 0.01551419 -0.511789 0.48399425
0.7225203 -0.40596035 -0.1691954 -0.06363284 -1.0298718 0.6138745
0.9918874 -0.67705 -0.3621858 -1.1191992 0.857297 0.20940647
0.39304456 -1.0405004 0.24550577 0.03066354 -0.2262465 -0.3987475
-0.28692126 -0.48556828 0.02729417 0.5924082 -0.38794446 0.6987338
0.37284875 0.35978708 -0.46888366 -0.90269667 0.67201567 -0.17143512
-0.88748777 0.01496055 -0.6797752 -0.3742612 0.92523205 0.09365834
-0.5484356 -0.05150444 -0.02859754 0.19510855 0.43586832 0.738918
0.7959159 -1.1560875 -0.23752865 0.1922295 0.29031602 0.01627672
0.30906796 0.61944485 -0.6729337 0.81698525 -0.07480188 -0.17640811
-1.0113621 0.82938945 0.48377824 -0.12809739 -0.6019519 0.7800945
-0.10530933 -0.6042999 0.5786599 -0.69254804 -0.32099804 -0.49988964
0.6877849 0.05625316 0.41252968 -0.09725304 -0.27831832 0.22231042
-0.17326355 0.5967302 1.3105283 0.30271867 -0.6596564 0.29500458
1.0143336 -0.14071907 -0.7033378 0.0712574 0.2367699 -0.1567619
0.0125481 -0.86470467 -1.208127 1.0146716 0.36200476 0.15971488
1.2483362 -0.18944505 0.76829034 0.57773274 0.50036126 -0.76012516
0.24892096 1.0449256 0.9743097 -0.9301011 -0.41977337 -0.20477441
-0.14535059 1.32122 0.98321986 -0.26813716 0.3745103 0.5369301
-0.37113294 -0.0918272 -0.9037528 0.28467873 -0.04766983 0.6848823
0.8793912 -0.27790445 -0.3454047 1.1533862 -0.44193482 0.28266668
0.92468214 0.9297876 -0.501228 -1.3681108 -0.31855518 0.43045017
-0.4245963 -0.30460754 -0.27489108 0.5642734 0.10813608 0.7954948
-0.31882307 -0.8488234 0.4116615 -0.06607208 0.34417102 -1.1066027
-0.64930326 -0.53532356 -0.00980501 -0.54927665 0.12916836 -0.26965186
-1.0309812 -0.3922892 -0.3059064 0.00962749 0.4402792 0.6989608
0.47415566 0.958419 0.2385101 -0.44608223 -0.63674647 -0.6855782
-0.08789552 -0.5627288 0.55497026 -0.42469448 -0.10216448 0.0114427 ] | [7.597423076629639, 7.752069473266602] |
5069a673-e4c1-4484-81af-1029dde7378f | open-set-recognition-with-gradient-based | 2206.08229 | null | https://arxiv.org/abs/2206.08229v1 | https://arxiv.org/pdf/2206.08229v1.pdf | Open-Set Recognition with Gradient-Based Representations | Neural networks for image classification tasks assume that any given image during inference belongs to one of the training classes. This closed-set assumption is challenged in real-world applications where models may encounter inputs of unknown classes. Open-set recognition aims to solve this problem by rejecting unknown classes while classifying known classes correctly. In this paper, we propose to utilize gradient-based representations obtained from a known classifier to train an unknown detector with instances of known classes only. Gradients correspond to the amount of model updates required to properly represent a given sample, which we exploit to understand the model's capability to characterize inputs with its learned features. Our approach can be utilized with any classifier trained in a supervised manner on known classes without the need to model the distribution of unknown samples explicitly. We show that our gradient-based approach outperforms state-of-the-art methods by up to 11.6% in open-set classification. | ['Ghassan AlRegib', 'Jinsol Lee'] | 2022-06-16 | null | null | null | null | ['open-set-learning'] | ['miscellaneous'] | [ 6.83935702e-01 2.42557779e-01 -4.81039345e-01 -6.79565847e-01
-6.53683186e-01 -6.93617642e-01 4.07603234e-01 2.15483367e-01
-4.90967542e-01 7.01049805e-01 -5.96046686e-01 -3.26785803e-01
-1.52533501e-01 -8.66410613e-01 -1.12437427e+00 -5.65782309e-01
4.22157384e-02 8.98230076e-01 1.47180527e-01 2.54637897e-01
1.72845900e-01 7.94820249e-01 -2.00314212e+00 7.00129688e-01
6.18472338e-01 1.40938425e+00 6.53044730e-02 8.86054695e-01
-1.57171503e-01 6.27959907e-01 -5.29796481e-01 -1.38658300e-01
4.68373060e-01 -2.71991104e-01 -6.51734054e-01 4.06206906e-01
9.43890631e-01 -3.87684375e-01 -2.06977308e-01 1.22016323e+00
6.86835451e-03 -3.06233335e-02 1.16968489e+00 -1.22318494e+00
-6.28946602e-01 2.74155200e-01 -9.68107656e-02 3.09870571e-01
-1.68397359e-03 1.02936246e-01 8.61613512e-01 -8.27545583e-01
5.18345416e-01 9.43210244e-01 5.08674979e-01 7.40982890e-01
-1.45083046e+00 -5.53072870e-01 5.04694581e-01 4.74706382e-01
-1.25545764e+00 -4.68149394e-01 6.18969798e-01 -5.30324638e-01
7.40028977e-01 2.07854688e-01 4.43798423e-01 1.06968212e+00
-7.65379295e-02 9.02686059e-01 1.18893039e+00 -5.01851320e-01
4.83823210e-01 7.08595872e-01 8.40987563e-01 8.23238134e-01
5.02543330e-01 3.50581884e-01 -1.48926139e-01 -1.09488182e-01
4.52382773e-01 1.36688605e-01 -2.46530145e-01 -5.28469622e-01
-7.34623909e-01 8.25568438e-01 5.72754622e-01 1.79924130e-01
-3.31987113e-01 -6.37385473e-02 2.19839320e-01 4.09689486e-01
6.77530408e-01 4.67169762e-01 -7.10719645e-01 3.01148444e-01
-8.66351128e-01 -7.82750174e-02 1.11012173e+00 8.62960994e-01
9.53760684e-01 -9.55757871e-03 4.53372486e-02 8.29311550e-01
1.65861621e-01 5.56741595e-01 5.30863881e-01 -6.07851982e-01
6.50414452e-02 6.35638237e-01 -9.27138627e-02 -5.63930035e-01
-2.23329455e-01 -7.88870811e-01 -7.28761911e-01 3.13706577e-01
8.89841259e-01 4.09144908e-02 -1.19402146e+00 1.52574623e+00
1.79351598e-01 4.47574019e-01 3.14502090e-01 6.09371245e-01
4.84133095e-01 5.80332637e-01 -4.98824298e-01 -1.45379260e-01
9.19834554e-01 -5.39406836e-01 -2.08294198e-01 -4.86070961e-01
5.10678351e-01 -2.20008478e-01 8.72963428e-01 5.91204524e-01
-4.75782156e-01 -5.09412885e-01 -1.28470767e+00 4.17522877e-01
-6.44808948e-01 4.91876125e-01 4.00993526e-01 8.04949284e-01
-6.38028860e-01 7.87841797e-01 -7.61072874e-01 -1.72375455e-01
9.92846608e-01 5.68088531e-01 -2.81964362e-01 -3.52167159e-01
-5.35430849e-01 8.70144546e-01 3.79301161e-01 1.57331824e-01
-1.18692780e+00 -5.14760911e-01 -8.40103388e-01 1.90220103e-01
5.46807289e-01 -3.05492789e-01 1.16684747e+00 -1.29313266e+00
-1.00505233e+00 9.20706809e-01 -3.57109845e-01 -8.44324827e-01
5.62354684e-01 -3.25359441e-02 -1.49889126e-01 1.40864819e-01
-1.82205752e-01 4.07675415e-01 1.29082489e+00 -1.35288680e+00
-5.43557405e-01 -5.74011087e-01 4.06240299e-02 -1.27221853e-01
-4.66085374e-01 -3.98605853e-01 -6.84604142e-03 -1.36614777e-02
3.87422144e-01 -9.28231895e-01 -1.54910132e-01 4.13088739e-01
-4.71537411e-01 -9.96323898e-02 9.18421209e-01 -1.40645713e-01
7.97940791e-01 -1.89814639e+00 -2.35809684e-01 3.57428223e-01
2.75541157e-01 4.31752384e-01 -5.69159128e-02 -6.31015152e-02
-3.32826078e-02 -4.69735228e-02 -4.45447057e-01 -2.51258790e-01
-1.63704678e-01 6.45284593e-01 -6.08050346e-01 6.73444152e-01
5.16926169e-01 3.95118505e-01 -8.05027962e-01 -2.41782397e-01
3.05008739e-01 2.74200618e-01 -3.68801385e-01 2.34210491e-01
-3.64421070e-01 1.23343758e-01 -2.54787296e-01 5.68760216e-01
6.29905403e-01 -4.23376054e-01 1.82345986e-01 5.75924739e-02
4.52484488e-01 1.94535814e-02 -1.48854959e+00 9.02772248e-01
-5.98758101e-01 7.53815830e-01 -2.86417276e-01 -1.47489595e+00
9.07574952e-01 -7.97759928e-03 -9.12530571e-02 -1.11360952e-01
2.46030137e-01 3.90314668e-01 3.48016411e-01 -2.76201844e-01
-8.38222504e-02 3.42938825e-02 4.03859735e-01 3.01568985e-01
4.80241776e-01 1.23489708e-01 2.51207411e-01 -2.70284656e-02
1.01831317e+00 -6.12448603e-02 2.36055732e-01 -2.19622672e-01
4.35261756e-01 -1.32376477e-01 4.50476289e-01 1.33538949e+00
-1.28178254e-01 5.62835515e-01 2.98843026e-01 -6.91912770e-01
-8.25849354e-01 -1.14498997e+00 -5.82949102e-01 7.09834576e-01
-9.05122682e-02 2.27852121e-01 -6.26325130e-01 -1.10541523e+00
9.79984775e-02 5.93271017e-01 -9.08582509e-01 -2.30141833e-01
-1.13811634e-01 -6.21156871e-01 2.76111960e-01 5.02821624e-01
1.91081479e-01 -7.82585025e-01 -7.63763428e-01 2.53825379e-03
2.28666723e-01 -9.06884789e-01 -4.15534638e-02 6.51395619e-01
-9.39501464e-01 -1.64512837e+00 -5.16122162e-01 -7.31230795e-01
1.27101970e+00 -2.67239958e-02 1.00139725e+00 5.79258986e-02
-4.61510986e-01 3.74794364e-01 -1.25838369e-01 -7.37247586e-01
-5.97907841e-01 -3.77597585e-02 2.79295474e-01 4.54938143e-01
5.80059469e-01 -3.27163398e-01 -2.24116221e-01 3.51446867e-01
-8.36184621e-01 -2.72816271e-01 5.20024419e-01 1.18839037e+00
6.54141903e-01 6.85979426e-02 7.24988878e-01 -1.22282708e+00
1.92040399e-01 -4.89828855e-01 -7.80556500e-01 4.26093876e-01
-7.04752803e-01 2.61398196e-01 1.06186736e+00 -1.00089633e+00
-6.95285678e-01 3.85794222e-01 1.55509636e-01 -6.84615195e-01
-5.22777259e-01 3.62603545e-01 -9.42574814e-02 -1.95749640e-01
9.71071720e-01 3.62145036e-01 5.57951331e-02 -3.43504757e-01
8.13086107e-02 8.15278769e-01 4.12732869e-01 -5.11739373e-01
6.63714111e-01 5.57510197e-01 1.20702669e-01 -8.58577907e-01
-1.44297969e+00 -5.77898204e-01 -8.51646602e-01 -2.86935449e-01
2.59125620e-01 -6.44890845e-01 -5.61206281e-01 4.41341668e-01
-9.82149959e-01 -2.29423523e-01 -3.98093730e-01 5.07948279e-01
-4.72258568e-01 7.69674182e-02 -4.08678830e-01 -1.05521739e+00
-6.02341816e-02 -1.03173923e+00 7.69873857e-01 2.44205117e-01
5.73365774e-04 -9.38341856e-01 -3.45551074e-01 9.68975723e-02
1.73193812e-01 -1.49240077e-01 7.17529774e-01 -1.48814106e+00
-4.51993287e-01 -6.76660597e-01 -9.21578109e-02 8.55963111e-01
1.64101079e-01 -4.15726490e-02 -1.34348214e+00 -2.36508936e-01
9.95209366e-02 -7.01938093e-01 1.16633987e+00 3.48050803e-01
1.42191875e+00 -4.91405576e-01 -3.12165886e-01 4.40453440e-01
1.38255167e+00 2.14720652e-01 3.69955182e-01 -1.09517965e-02
3.78338873e-01 5.65050304e-01 4.36107039e-01 3.63934010e-01
-1.76601425e-01 1.09297737e-01 4.31248844e-01 1.59159854e-01
2.36010864e-01 -1.87753707e-01 4.62662475e-03 3.29968594e-02
5.08142054e-01 -2.92300671e-01 -8.42709243e-01 5.68975568e-01
-1.74218440e+00 -9.95617449e-01 1.74094699e-02 2.46050882e+00
7.58410096e-01 6.38647199e-01 -2.47876838e-01 4.92737591e-01
6.67418301e-01 -2.95318425e-01 -9.77419972e-01 -1.61476925e-01
-7.60019571e-02 2.41585135e-01 5.33443809e-01 4.57941175e-01
-1.23400402e+00 5.30357897e-01 6.04601717e+00 6.53782368e-01
-1.26207185e+00 -2.55831331e-01 9.36731279e-01 2.49470562e-01
1.22278757e-01 6.22542836e-02 -1.18241489e+00 2.76668310e-01
1.03895962e+00 -9.91792884e-04 3.03409249e-01 1.12979949e+00
-3.13953042e-01 -2.33403146e-01 -1.80436397e+00 7.92881906e-01
3.38380337e-01 -1.38053322e+00 1.21455848e-01 1.12326153e-01
5.94502151e-01 4.77854759e-02 1.95812911e-01 5.48451006e-01
6.30921796e-02 -8.92013490e-01 5.01469076e-01 5.22000790e-01
5.93558788e-01 -5.42101622e-01 7.26927221e-01 9.15993512e-01
-6.09182835e-01 -2.83288568e-01 -4.44160223e-01 -3.71210694e-01
-5.54770887e-01 5.33711255e-01 -1.34000075e+00 5.12536988e-02
2.70265251e-01 6.95245385e-01 -5.92218876e-01 1.14850795e+00
-4.33729887e-02 1.03759038e+00 -7.38189638e-01 -1.02136128e-01
2.18138322e-01 -5.13829291e-02 2.77837485e-01 8.03843915e-01
-7.66533753e-03 -1.47458032e-01 4.82591778e-01 9.28845704e-01
-1.52586013e-01 -2.37318292e-01 -8.34401369e-01 -1.27871171e-01
3.30608606e-01 8.39131773e-01 -9.08872008e-01 -7.15805888e-01
-3.12839746e-01 9.17989016e-01 5.04094243e-01 2.97602117e-01
-4.60926116e-01 -4.01385427e-01 3.85413915e-01 6.37797564e-02
3.36398184e-01 -4.19520698e-02 -2.07401186e-01 -1.25200689e+00
2.36475348e-01 -5.60501873e-01 4.38754469e-01 -3.61254930e-01
-1.55887365e+00 6.20838881e-01 2.31209025e-02 -1.30096507e+00
-4.83658731e-01 -1.23706043e+00 -4.79021400e-01 6.67468011e-01
-1.66819572e+00 -8.98006618e-01 -8.86376575e-02 3.32878649e-01
6.38444066e-01 -1.19318888e-01 1.00303042e+00 -1.58313643e-02
-4.77120757e-01 5.52500606e-01 4.63486135e-01 4.47722316e-01
4.18365479e-01 -1.34195030e+00 3.87857519e-02 8.17886591e-01
6.53930426e-01 4.33859676e-01 7.21428394e-01 -3.36581945e-01
-1.20429540e+00 -1.14002144e+00 6.12412989e-01 -4.65678185e-01
5.07756531e-01 -4.89902794e-01 -1.15313601e+00 8.13269019e-01
-5.20441115e-01 9.25269961e-01 6.50091290e-01 1.51085883e-01
-7.05049992e-01 -2.75534749e-01 -1.26278543e+00 1.60487279e-01
6.36962414e-01 -6.01542950e-01 -6.31065309e-01 5.59635580e-01
1.94638684e-01 -3.70962024e-01 -5.24605691e-01 2.85355151e-01
4.86120433e-01 -6.48063362e-01 7.96400428e-01 -1.00773585e+00
3.05465430e-01 -1.72607079e-01 -3.15716356e-01 -1.30679274e+00
1.95109665e-01 5.47307208e-02 -3.50804538e-01 7.52833366e-01
6.23668373e-01 -8.33163559e-01 1.07472157e+00 5.47521174e-01
8.24333131e-02 -9.07022595e-01 -9.45856214e-01 -8.80409122e-01
3.71673852e-02 -6.93998754e-01 -1.74028929e-02 8.37113857e-01
-3.07241678e-01 3.46903294e-01 -1.29581422e-01 5.11595666e-01
7.72425294e-01 3.06513339e-01 4.88371730e-01 -1.54787433e+00
-5.22207618e-01 -1.45108968e-01 -8.09571981e-01 -9.76470470e-01
5.42514384e-01 -8.51921439e-01 7.75339082e-02 -1.22342157e+00
2.05971137e-01 -7.15084672e-01 -3.81540865e-01 4.57400709e-01
-2.67004557e-02 2.49984369e-01 4.13202271e-02 9.79029089e-02
-5.08655906e-01 3.14381897e-01 7.11187243e-01 -4.37284231e-01
-2.18375959e-02 5.88957489e-01 -4.78271395e-01 8.90438080e-01
6.58244789e-01 -5.56545854e-01 -5.34452915e-01 -1.25859261e-01
-1.67762205e-01 -1.24420457e-01 5.77775955e-01 -1.21830654e+00
2.28213951e-01 -9.95564461e-02 8.52274001e-01 -3.37836117e-01
5.52564025e-01 -1.02911460e+00 -3.71655196e-01 5.95553637e-01
-6.61684513e-01 -7.10974038e-01 2.13682801e-01 8.41591895e-01
-9.92539451e-02 -6.72659814e-01 9.83009636e-01 -9.78631303e-02
-6.84225500e-01 3.22715342e-01 -4.87443507e-01 4.49936166e-02
1.06434011e+00 -3.05116147e-01 -1.98109657e-01 -2.88843364e-01
-1.02013326e+00 1.74956486e-01 1.90901741e-01 1.57058761e-01
9.40820992e-01 -7.52969086e-01 -5.24567127e-01 4.49064642e-01
4.45238143e-01 -4.24553156e-02 -7.58068040e-02 2.83652902e-01
-1.49702415e-01 1.02731198e-01 4.95210588e-02 -9.77019668e-01
-1.34287846e+00 6.68364406e-01 6.57735288e-01 -1.43319085e-01
-2.57994473e-01 7.78634906e-01 8.88222754e-02 -7.62848318e-01
4.89237696e-01 -4.79732603e-01 -5.94987422e-02 -8.85063559e-02
7.00817049e-01 3.19071785e-02 2.17321783e-01 -3.82675797e-01
-2.02442482e-01 1.17076844e-01 -4.56940800e-01 2.32625142e-01
1.13971555e+00 3.74005705e-01 4.34081137e-01 9.54578936e-01
1.47510684e+00 -4.75853503e-01 -1.36029565e+00 -6.09016716e-01
4.53236187e-03 -4.76747960e-01 2.58022677e-02 -9.57937717e-01
-6.08296990e-01 1.01457667e+00 9.56477642e-01 -7.89681226e-02
8.29199076e-01 2.86772139e-02 1.53257430e-01 1.05763984e+00
5.27962863e-01 -8.57018530e-01 -1.40762717e-01 5.87245643e-01
6.10011399e-01 -1.68180537e+00 -2.03428239e-01 -4.45301503e-01
-2.01597229e-01 1.35528481e+00 7.52955019e-01 -3.87157261e-01
9.06998277e-01 9.85433832e-02 -1.01373293e-01 1.50074050e-01
-1.01477599e+00 -2.05696225e-01 5.30658126e-01 5.78466535e-01
4.20930795e-02 7.05489516e-02 3.19415927e-01 3.24493766e-01
2.40560174e-01 -7.22517669e-02 6.99616790e-01 1.13318181e+00
-6.20106339e-01 -8.33991706e-01 -4.32715595e-01 1.24321747e+00
-2.64884412e-01 -5.11656590e-02 -4.69060779e-01 6.03058875e-01
8.45678821e-02 7.88060784e-01 3.00526649e-01 -2.56289870e-01
1.00104369e-01 1.81921303e-01 6.01157784e-01 -9.04050231e-01
-1.99865118e-01 -3.66117150e-01 6.07234472e-03 -2.33833209e-01
-2.18414247e-01 -6.21928334e-01 -1.01764894e+00 3.46803904e-01
-6.42375648e-01 2.43731327e-02 7.36599803e-01 1.04682958e+00
6.76753595e-02 3.95267755e-01 6.71888590e-01 -8.76338303e-01
-1.23775971e+00 -9.04141724e-01 -5.29722750e-01 2.72400945e-01
6.94131315e-01 -6.74859881e-01 -6.74036086e-01 9.94280651e-02] | [9.544711112976074, 3.0263195037841797] |
606ddf0c-ba63-4770-9857-bee53c8beb35 | crystal-transformer-self-learning-neural | 2204.11953 | null | https://arxiv.org/abs/2204.11953v1 | https://arxiv.org/pdf/2204.11953v1.pdf | Crystal Transformer: Self-learning neural language model for Generative and Tinkering Design of Materials | Self-supervised neural language models have recently achieved unprecedented success, from natural language processing to learning the languages of biological sequences and organic molecules. These models have demonstrated superior performance in the generation, structure classification, and functional predictions for proteins and molecules with learned representations. However, most of the masking-based pre-trained language models are not designed for generative design, and their black-box nature makes it difficult to interpret their design logic. Here we propose BLMM Crystal Transformer, a neural network based probabilistic generative model for generative and tinkering design of inorganic materials. Our model is built on the blank filling language model for text generation and has demonstrated unique advantages in learning the "materials grammars" together with high-quality generation, interpretability, and data efficiency. It can generate chemically valid materials compositions with as high as 89.7\% charge neutrality and 84.8\% balanced electronegativity, which are more than 4 and 8 times higher compared to a pseudo random sampling baseline. The probabilistic generation process of BLMM allows it to recommend tinkering operations based on learned materials chemistry and makes it useful for materials doping. Combined with the TCSP crysal structure prediction algorithm, We have applied our model to discover a set of new materials as validated using DFT calculations. Our work thus brings the unsupervised transformer language models based generative artificial intelligence to inorganic materials. A user-friendly web app has been developed for computational materials doping and can be accessed freely at \url{www.materialsatlas.org/blmtinker}. | ['Jianjun Hu', 'Fanglin Chen', 'Edirisuriya M. D. Siriwardane', 'Stanislav Stefanov', 'Yuqi Song', 'Qinyang Li', 'Lai Wei'] | 2022-04-25 | null | null | null | null | ['self-learning'] | ['natural-language-processing'] | [ 5.15445411e-01 2.33826712e-01 -6.89911544e-02 -2.36502439e-01
-6.15465641e-01 -4.70497131e-01 5.83073199e-01 1.11522801e-01
1.15686506e-01 1.32673275e+00 1.59164533e-01 -6.14455342e-01
1.94051534e-01 -1.23529112e+00 -1.04387343e+00 -1.11725104e+00
3.63116831e-01 9.80538070e-01 -8.51172507e-02 -2.72588581e-01
3.90080124e-01 3.44761461e-01 -1.71611214e+00 8.33580732e-01
1.11450899e+00 7.29021132e-01 8.40805829e-01 3.25816751e-01
-4.44880813e-01 6.60105526e-01 -4.00449187e-01 -2.04058230e-01
-2.31833130e-01 -5.50930977e-01 -6.32771313e-01 -4.95833129e-01
-1.67822614e-01 6.99213594e-02 1.43551064e-04 6.74899817e-01
7.20576406e-01 -1.48344904e-01 1.23846006e+00 -5.97865641e-01
-1.30776441e+00 1.07954419e+00 -7.53828324e-03 -4.01834190e-01
4.42764193e-01 4.25000101e-01 1.14494121e+00 -1.04551256e+00
7.64705658e-01 1.07835412e+00 4.61972058e-01 1.03567135e+00
-1.39826345e+00 -9.87328708e-01 -2.89153844e-01 3.59499070e-04
-1.13445854e+00 -3.66118491e-01 6.46400094e-01 -6.01989150e-01
1.57301259e+00 1.89050764e-01 6.89726591e-01 1.38835728e+00
6.34015560e-01 5.93208194e-01 1.11971629e+00 -4.91344035e-01
4.35943663e-01 1.93533137e-01 -3.37206453e-01 6.58180654e-01
4.48908061e-01 2.05355123e-01 -7.72603273e-01 -1.47469029e-01
3.90680045e-01 -7.12005571e-02 1.02536678e-01 -1.68956295e-01
-8.63146663e-01 9.68607247e-01 3.80751431e-01 3.09063941e-01
-4.00535703e-01 1.80097818e-01 -3.43310758e-02 -3.57979357e-01
2.93668389e-01 9.31392133e-01 -4.52057511e-01 1.49995059e-01
-6.44341052e-01 4.45358872e-01 6.85555220e-01 9.57248390e-01
7.40258634e-01 4.88957554e-01 -9.96773914e-02 7.45483518e-01
4.44169283e-01 6.33680582e-01 4.24540162e-01 -5.16852498e-01
-1.34433468e-03 5.74953735e-01 -8.36345181e-02 -3.78372252e-01
-3.53721261e-01 -3.08692902e-01 -8.78123641e-01 8.41010287e-02
-1.47895843e-01 -7.79721653e-03 -1.05400622e+00 1.63227534e+00
4.48295139e-02 -6.12202227e-01 3.72174382e-01 3.74778122e-01
1.23928404e+00 1.15940845e+00 3.09473038e-01 -2.47435361e-01
1.15746140e+00 -3.84123147e-01 -4.85500753e-01 1.20359875e-01
4.83414680e-01 -6.09984159e-01 9.81539428e-01 4.81406331e-01
-1.10621428e+00 -4.24485624e-01 -1.13737476e+00 -3.13587068e-03
-6.44216597e-01 8.94891396e-02 1.19860697e+00 9.05934393e-01
-9.38238025e-01 8.14489722e-01 -8.38125885e-01 -9.61430222e-02
6.24192059e-01 6.32796466e-01 -2.61260029e-02 1.39985263e-01
-1.13657606e+00 4.91363078e-01 8.54103148e-01 -2.48536482e-01
-9.49477255e-01 -8.15034032e-01 -5.47050118e-01 -1.61162466e-01
-3.23326699e-02 -9.83176351e-01 1.11275649e+00 -5.07365584e-01
-1.76252210e+00 7.04762042e-01 -3.16521674e-01 -4.64063972e-01
-3.70274894e-02 1.65020332e-01 -2.58379072e-01 -1.94703713e-01
1.63292691e-01 1.03025615e+00 3.42439115e-01 -1.25380516e+00
-9.82944518e-02 -5.96493930e-02 -4.80496556e-01 -1.25449970e-01
-2.68357009e-01 -1.92047060e-01 2.79894441e-01 -7.88432896e-01
4.08109874e-02 -7.05400050e-01 -3.42178702e-01 -4.22963828e-01
-5.51987350e-01 -2.59437114e-01 4.27149713e-01 -3.98141235e-01
9.08698082e-01 -1.39103317e+00 -9.25842375e-02 3.34866583e-01
1.09435767e-01 2.47525573e-01 1.75670125e-02 8.28607559e-01
-2.56359190e-01 3.45923841e-01 -3.15913856e-01 3.54699671e-01
1.15164742e-02 -7.33142272e-02 -3.75664711e-01 -1.25823095e-01
3.58622879e-01 1.28406751e+00 -8.34789038e-01 1.50362268e-01
7.14755431e-02 5.98160207e-01 -9.22973216e-01 1.71445817e-01
-1.01382875e+00 6.00764930e-01 -4.03196990e-01 7.29068220e-01
6.17417216e-01 -4.69220161e-01 5.66949069e-01 -1.90010026e-01
-2.95654714e-01 7.66041994e-01 -5.86726189e-01 1.51822662e+00
-7.31669739e-02 1.13924138e-01 -7.01315165e-01 -7.26811588e-01
1.33569872e+00 1.54305413e-01 4.97200608e-01 -8.91100109e-01
-6.59687072e-02 2.88172185e-01 1.70452863e-01 -3.96978706e-01
4.43626255e-01 -5.63442886e-01 1.15346849e-01 5.20382106e-01
1.10716343e-01 -4.10772681e-01 3.04176480e-01 6.59513399e-02
8.40896606e-01 5.99291980e-01 1.47184134e-01 -4.50497150e-01
2.71563381e-01 9.92663354e-02 3.48141789e-01 7.12646484e-01
7.37921655e-01 3.40973139e-01 1.54970393e-01 -4.35198337e-01
-1.35443985e+00 -1.24282908e+00 -9.96631756e-02 1.00721550e+00
-2.70646960e-01 -7.57500529e-01 -7.62215614e-01 1.13538086e-01
-2.09588185e-01 1.03502095e+00 -2.49485165e-01 -4.57663059e-01
-3.17067623e-01 -1.34155393e+00 3.95308942e-01 5.35260856e-01
2.42222443e-01 -1.44804490e+00 4.52352539e-02 4.61413711e-01
9.65326875e-02 -6.07368171e-01 -7.51779974e-02 7.16424465e-01
-7.26459622e-01 -7.05577970e-01 -3.94908696e-01 -8.50742877e-01
7.35946774e-01 -3.05129081e-01 9.47777748e-01 -2.32920736e-01
-4.52493489e-01 -1.40020952e-01 -1.22225126e-02 -7.31811106e-01
-1.20520103e+00 2.49752715e-01 3.39876324e-01 -5.40476322e-01
3.21417838e-01 -8.17614377e-01 -5.83222806e-01 -4.51733060e-02
-9.87175226e-01 6.09480321e-01 7.70868361e-01 9.10887778e-01
9.13061917e-01 1.47947490e-01 8.63636136e-01 -1.19511008e+00
6.18348718e-01 -2.57940233e-01 -5.35338938e-01 2.96322823e-01
-1.11520529e+00 6.50375783e-01 7.91431963e-01 -2.14833751e-01
-1.14326918e+00 1.71115860e-01 -6.17207706e-01 2.27101326e-01
-1.99474722e-01 3.64059150e-01 -6.74282432e-01 1.49817213e-01
8.13365936e-01 6.13054097e-01 3.42164487e-02 -2.31976792e-01
4.13642198e-01 5.37868202e-01 2.33401984e-01 -1.05948246e+00
5.60153127e-01 9.37091187e-02 -5.87716140e-02 -9.08034027e-01
-3.97779882e-01 1.97066590e-01 -3.17875236e-01 1.53346136e-01
8.97556245e-01 -7.90681779e-01 -1.08384740e+00 4.56620932e-01
-9.47228909e-01 -2.54084915e-01 -2.24715307e-01 1.55917585e-01
-6.84573770e-01 1.97810948e-01 -6.61878645e-01 -9.13589358e-01
-8.05698931e-01 -1.18594277e+00 8.59064996e-01 1.39120653e-01
-5.15348494e-01 -7.73611426e-01 -6.52955696e-02 3.50375772e-01
4.72976983e-01 2.53289580e-01 1.60203731e+00 -5.29093862e-01
-8.33746016e-01 2.71392077e-01 1.67077780e-01 1.79155186e-01
4.36405897e-01 -5.92276305e-02 -9.23751533e-01 6.49348833e-03
-2.38640174e-01 -3.79865915e-01 9.90279257e-01 3.35341781e-01
1.33298111e+00 -4.84566718e-01 -5.38925648e-01 4.64035243e-01
1.22603881e+00 8.42290401e-01 9.45300221e-01 8.59242231e-02
8.23025703e-01 3.17673266e-01 6.17136061e-02 4.21164930e-01
-2.99864420e-04 3.29435647e-01 1.86112970e-01 1.05430655e-01
3.42427045e-02 -7.28100598e-01 5.96480370e-01 8.43970835e-01
-3.08517456e-01 -5.09921849e-01 -9.96076405e-01 -1.48427160e-02
-1.34169638e+00 -1.05505741e+00 -1.07899047e-01 2.02485633e+00
1.27776349e+00 3.22153896e-01 2.88328901e-02 -6.03728145e-02
5.22478044e-01 -2.57852703e-01 -8.69779348e-01 -5.79402089e-01
-4.85381246e-01 9.49594498e-01 4.00230944e-01 1.80389225e-01
-6.13511086e-01 1.20318484e+00 6.67434692e+00 1.05312181e+00
-1.21306038e+00 -1.76591396e-01 8.57441068e-01 1.11212358e-01
-1.05525923e+00 8.92327949e-02 -1.17464745e+00 4.45829719e-01
1.05179811e+00 6.20950647e-02 4.47599739e-01 5.60863554e-01
2.63974994e-01 1.94502369e-01 -1.09457874e+00 9.22805071e-01
-2.23644525e-01 -2.29032850e+00 7.14873195e-01 1.54689878e-01
8.42092395e-01 -1.01317331e-01 1.28996328e-01 8.42400491e-02
4.11602944e-01 -1.40796506e+00 7.73100972e-01 6.16085470e-01
1.02614474e+00 -6.86770260e-01 1.17059819e-01 1.88737333e-01
-8.19328547e-01 8.44091028e-02 -4.03655678e-01 6.08140193e-02
2.57524271e-02 6.25977278e-01 -1.28790164e+00 3.83625299e-01
4.68709052e-01 4.77021337e-01 -6.98841810e-02 4.03639466e-01
-1.46595448e-01 6.63529634e-01 -1.69248268e-01 -6.46123946e-01
-1.10425457e-01 -4.74598795e-01 1.90476149e-01 9.74482059e-01
5.27139008e-01 9.01208222e-02 -4.96716648e-02 1.49938869e+00
-2.37375349e-01 1.39967009e-01 -6.15106285e-01 -7.33165443e-01
2.77691543e-01 7.66041517e-01 -7.92227745e-01 -2.81834453e-01
1.04427323e-01 6.53533161e-01 -5.88678867e-02 2.37781897e-01
-7.47927368e-01 -1.42382130e-01 4.39138591e-01 5.07072806e-01
4.78550762e-01 -1.46416739e-01 -3.51940453e-01 -7.21576214e-01
-2.85214961e-01 -9.14600670e-01 -1.36171699e-01 -1.01186538e+00
-1.36989594e+00 4.75317419e-01 -1.96431443e-01 -7.58094132e-01
-2.39234090e-01 -1.07756627e+00 -4.16366220e-01 7.56710708e-01
-9.07672346e-01 -1.21995246e+00 2.90884763e-01 -5.74938804e-02
2.01462910e-01 -4.72187519e-01 1.18353820e+00 -5.40802740e-02
-3.95586640e-01 5.44033885e-01 5.28441370e-01 -4.12773639e-01
3.26391369e-01 -1.10357523e+00 4.85620916e-01 2.06308171e-01
2.54049823e-02 8.75761390e-01 8.09105158e-01 -1.01115668e+00
-1.55633557e+00 -1.19478261e+00 7.53263593e-01 -4.86634374e-01
3.47753167e-01 -8.56131554e-01 -7.92022228e-01 1.42250180e-01
2.30165094e-01 -8.72136176e-01 1.14516509e+00 -2.46498004e-01
-2.32528880e-01 4.00780290e-02 -1.06734407e+00 6.43789768e-01
1.20294166e+00 -3.28940272e-01 -2.10967228e-01 8.34488928e-01
9.31204915e-01 -3.97406727e-01 -9.24111247e-01 5.25383949e-01
5.35409033e-01 -7.80768037e-01 1.04471540e+00 -6.80125892e-01
6.56074405e-01 -2.78977633e-01 -2.03119904e-01 -7.55661190e-01
-4.13300931e-01 -7.97327340e-01 5.07064350e-03 1.17982137e+00
9.70835805e-01 -7.54302800e-01 9.20699179e-01 4.76964653e-01
-3.71639729e-01 -9.57418740e-01 -5.17581403e-01 -5.43672800e-01
2.77481288e-01 -4.12269175e-01 8.49887848e-01 5.31664073e-01
-1.70931220e-01 6.67107344e-01 -2.67903179e-01 -2.44088009e-01
4.16500807e-01 2.41504103e-01 3.75952840e-01 -1.13261366e+00
-5.46999395e-01 -4.01393086e-01 -6.33917451e-02 -8.29463840e-01
1.50767356e-01 -1.50527501e+00 -1.35640070e-01 -1.62643194e+00
2.81519353e-01 -5.78915775e-01 -1.81631178e-01 6.02594018e-01
3.68931442e-01 9.66002420e-03 -3.80968809e-01 1.43441841e-01
-1.88626930e-01 8.44236314e-01 1.10659611e+00 -4.13826138e-01
-3.32921773e-01 -3.07988506e-02 -1.22115183e+00 2.24782288e-01
1.04011333e+00 -4.93842244e-01 -4.47978556e-01 4.53516394e-02
7.07618117e-01 -3.85014206e-01 -1.36567196e-02 -1.08514261e+00
-1.96827188e-01 -2.11926371e-01 6.49363995e-01 -6.62645578e-01
3.65746677e-01 -2.15117887e-01 8.02407563e-01 6.68217838e-01
-3.41276735e-01 -2.68722087e-01 2.08692878e-01 3.34095389e-01
3.77063870e-01 -1.96141049e-01 5.05795717e-01 -5.27879953e-01
-3.20116013e-01 3.84247333e-01 -8.22391927e-01 -2.71491021e-01
7.03652740e-01 -4.33954298e-01 -2.48100013e-01 -8.34995359e-02
-5.71268439e-01 -2.81187534e-01 6.11915052e-01 2.64717966e-01
7.46532500e-01 -1.14869380e+00 -3.29903632e-01 3.58766824e-01
1.60534173e-01 6.42120019e-02 1.23849660e-01 2.51339078e-01
-7.00747311e-01 7.30023265e-01 -2.69735843e-01 -5.43108761e-01
-7.25070238e-01 5.67835927e-01 2.06061989e-01 -7.53019154e-02
-2.67459750e-01 6.43643558e-01 4.00547594e-01 -4.90591079e-01
-2.48398080e-01 -4.25992072e-01 8.15786887e-03 -3.68263304e-01
2.59383917e-01 -7.81378746e-02 2.87780404e-01 -2.36861497e-01
-1.45896241e-01 2.90291011e-01 -2.78189063e-01 2.10815325e-01
1.81212056e+00 5.65993547e-01 -3.73619288e-01 3.87044400e-01
8.26454222e-01 1.36210825e-02 -8.91574681e-01 9.74004269e-02
-1.21530909e-02 3.91435266e-01 -4.67249423e-01 -9.05284286e-01
-5.60052693e-01 6.30404174e-01 5.60806155e-01 -3.02897722e-01
7.97270477e-01 2.20276549e-01 8.50068688e-01 7.03017473e-01
4.24264342e-01 -9.91210043e-01 1.35930017e-01 4.99361485e-01
7.80406535e-01 -8.87889624e-01 5.52868880e-02 -3.90775681e-01
-4.38697726e-01 1.06851673e+00 5.51581681e-01 2.67022759e-01
2.95187056e-01 4.69960451e-01 -3.20648879e-01 -4.40355957e-01
-9.90389109e-01 9.48216990e-02 6.44153953e-02 9.32547808e-01
8.78437638e-01 2.39240244e-01 -2.05624253e-01 8.36846828e-01
-5.14360964e-01 -7.56116817e-03 2.14282468e-01 8.45793307e-01
-5.25507331e-01 -1.73012066e+00 -1.59386709e-01 5.77928007e-01
-2.13329285e-01 -5.16272247e-01 -6.12272203e-01 1.71507537e-01
2.65016258e-01 5.38310170e-01 -2.51109302e-01 -4.74454433e-01
-1.09387845e-01 3.58108819e-01 6.52524471e-01 -7.23656595e-01
-4.42011148e-01 -4.84103151e-02 1.06351793e-01 -3.84254917e-03
-4.07751024e-01 -5.03932595e-01 -1.70831048e+00 -2.85769850e-01
-3.57532442e-01 4.62932378e-01 6.51618779e-01 7.87751794e-01
7.43710995e-01 5.70304453e-01 2.58162141e-01 -7.76860297e-01
-5.66492938e-02 -7.07155108e-01 -3.41672748e-01 1.76666733e-02
-4.62203473e-01 -5.60737669e-01 2.09756047e-01 1.15759797e-01] | [5.127840042114258, 5.408814907073975] |
8541c411-88fc-4fc0-87be-453c4bda3496 | clear-the-fog-combat-value-assessment-in | 1811.12627 | null | http://arxiv.org/abs/1811.12627v2 | http://arxiv.org/pdf/1811.12627v2.pdf | Clear the Fog: Combat Value Assessment in Incomplete Information Games with Convolutional Encoder-Decoders | StarCraft, one of the most popular real-time strategy games, is a compelling
environment for artificial intelligence research for both micro-level unit
control and macro-level strategic decision making. In this study, we address an
eminent problem concerning macro-level decision making, known as the
'fog-of-war', which rises naturally from the fact that information regarding
the opponent's state is always provided in the incomplete form. For intelligent
agents to play like human players, it is obvious that making accurate
predictions of the opponent's status under incomplete information will increase
its chance of winning. To reflect this fact, we propose a convolutional
encoder-decoder architecture that predicts potential counts and locations of
the opponent's units based on only partially visible and noisy information. To
evaluate the performance of our proposed method, we train an additional
classifier on the encoder-decoder output to predict the game outcome (win or
lose). Finally, we designed an agent incorporating the proposed method and
conducted simulation games against rule-based agents to demonstrate both
effectiveness and practicality. All experiments were conducted on actual game
replay data acquired from professional players. | ['Changhyeon Bae', 'Hyungu Kahng', 'Young Joon Park', 'Yoon Sang Cho', 'Junseung Lee', 'Iljoo Yoon', 'Hyunjin Choi', 'Hyunjae Lee', 'Gonie Ahn', 'Yonghyun Jeong', 'Seoung Bum Kim', 'Hyungrok Do', 'Uk Jo', 'Hankyu Lee', 'Daehun Jun'] | 2018-11-30 | null | null | null | null | ['real-time-strategy-games'] | ['playing-games'] | [ 3.22207771e-02 6.68475851e-02 -3.14809307e-02 -1.37316748e-01
-3.11200827e-01 -4.57349956e-01 5.30661345e-01 -4.17945050e-02
-8.88483405e-01 8.86781931e-01 8.68018866e-02 -4.05971408e-01
2.42566857e-02 -1.01888704e+00 -4.89674777e-01 -5.79787791e-01
-1.63198307e-01 3.38282347e-01 4.30889398e-01 -8.60633194e-01
5.09867728e-01 -1.56871546e-02 -1.52288926e+00 3.46289605e-01
5.20632863e-01 1.19453883e+00 3.34217072e-01 8.14248204e-01
3.86735678e-01 1.58620608e+00 -1.05151010e+00 -7.56825745e-01
5.63218236e-01 -4.76689935e-01 -4.02462870e-01 -3.37971121e-01
-6.31222546e-01 -4.88339871e-01 -5.70131242e-01 1.18693697e+00
5.72394073e-01 8.11919272e-02 3.39348584e-01 -1.09951758e+00
5.74913155e-03 7.94856250e-01 -3.44531685e-01 5.79915583e-01
2.81732649e-01 6.28340721e-01 1.10863507e+00 -9.97043550e-02
4.89484489e-01 9.54936504e-01 2.54129559e-01 5.08974671e-01
-5.34513354e-01 -6.45986319e-01 2.04408497e-01 5.13277113e-01
-1.24036264e+00 -3.25747788e-01 8.29600930e-01 -1.77859575e-01
8.60085785e-01 1.22046228e-02 9.32360888e-01 1.08480048e+00
7.08129525e-01 9.53370988e-01 1.15278995e+00 -1.91954225e-01
6.14613652e-01 -1.38430580e-01 -3.61451209e-01 5.87961972e-01
1.07483417e-01 6.64763629e-01 -6.13161325e-01 6.78002983e-02
8.38421881e-01 -2.47917980e-01 5.16730770e-02 5.13836928e-02
-9.53919649e-01 8.01937878e-01 4.16446745e-01 1.34358913e-01
-6.42279088e-01 1.84897467e-01 4.23917800e-01 4.49723423e-01
-6.90463558e-02 6.02854431e-01 -7.93886259e-02 -8.62750828e-01
-5.60179353e-01 4.24266696e-01 8.87921274e-01 4.33296472e-01
3.00232202e-01 3.25400770e-01 -1.19091973e-01 2.64943063e-01
2.15697616e-01 4.92444783e-02 7.81376243e-01 -8.62717807e-01
5.60978711e-01 7.02124655e-01 3.15036893e-01 -1.17258120e+00
-4.36544776e-01 -8.68694663e-01 -5.60800970e-01 5.67925155e-01
4.04670179e-01 -5.40323019e-01 -4.43821996e-01 1.83082843e+00
-4.40726392e-02 4.11722302e-01 3.98216039e-01 1.23701024e+00
4.32407290e-01 5.47401130e-01 -2.45489225e-01 -8.83094743e-02
1.14766347e+00 -5.85822880e-01 -6.34977043e-01 -5.66111803e-01
3.12428504e-01 -2.52447933e-01 5.40230095e-01 5.79522252e-01
-1.06391251e+00 -2.94794112e-01 -1.25121450e+00 5.27526915e-01
-8.08253363e-02 -1.23224318e-01 8.56058538e-01 4.58755016e-01
-7.12006330e-01 5.19737184e-01 -7.92147577e-01 4.08603922e-02
2.13551089e-01 5.19276321e-01 -2.00903878e-01 4.48150963e-01
-1.63305295e+00 1.11227238e+00 5.37652731e-01 3.25837463e-01
-1.25691378e+00 8.77135247e-02 -7.18965411e-01 2.70200193e-01
8.33844900e-01 -2.62800992e-01 1.55302882e+00 -8.90635848e-01
-1.79113400e+00 5.91189861e-01 3.91354889e-01 -9.26672459e-01
5.76902688e-01 2.70085305e-01 -3.05544078e-01 -3.02542448e-01
2.96856444e-02 1.72596991e-01 5.87100267e-01 -7.63068080e-01
-1.09249353e+00 -5.45780122e-01 7.10513055e-01 7.95206189e-01
1.64084136e-01 -2.04526648e-01 -7.91362301e-02 -3.86799932e-01
-1.51091889e-01 -5.92315435e-01 -4.60377812e-01 -7.02980280e-01
-3.04931760e-01 -1.46905854e-01 2.07754031e-01 -4.15694386e-01
1.35774183e+00 -2.04414296e+00 2.18366832e-02 1.00640796e-01
2.87888378e-01 3.08349997e-01 1.10711493e-01 4.89125043e-01
3.28768671e-01 -3.63275141e-01 1.97283849e-01 1.36679605e-01
4.67639528e-02 2.58493543e-01 -1.54901996e-01 3.03467274e-01
-8.25253949e-02 9.09680843e-01 -7.84064591e-01 -1.07698321e-01
1.55696064e-01 -2.20847711e-01 -6.92691028e-01 3.51309806e-01
-1.38416454e-01 3.08043897e-01 -7.07204103e-01 5.02065063e-01
9.64789540e-02 8.35570768e-02 4.03236061e-01 2.51353085e-01
-2.52770662e-01 5.50175011e-01 -1.20068097e+00 1.49654937e+00
-1.52799651e-01 5.28288841e-01 2.06989661e-01 -1.27861285e+00
9.12212193e-01 2.88540006e-01 2.92914212e-01 -1.11176157e+00
8.26853514e-01 4.68976274e-02 7.39494562e-01 -2.01050714e-01
6.44604504e-01 -2.78710991e-01 -7.00020492e-01 4.63074118e-01
-1.00951999e-01 -1.41732313e-03 2.12638274e-01 7.24262968e-02
1.43821096e+00 -1.74900085e-01 6.16610229e-01 -5.13279950e-03
3.06177050e-01 1.23115756e-01 8.09917629e-01 1.11238503e+00
-6.89414203e-01 4.81178984e-02 9.27011132e-01 -6.82149470e-01
-7.08773673e-01 -7.89599478e-01 5.05293489e-01 9.50482488e-01
5.17710209e-01 -3.83476138e-01 -7.29237735e-01 -3.03394586e-01
-3.05616200e-01 7.40377367e-01 -7.05803573e-01 -5.89181423e-01
-3.95022005e-01 -7.25442708e-01 6.29815280e-01 2.28462636e-01
8.52290928e-01 -1.23328841e+00 -1.30135739e+00 5.94182312e-01
-2.06604570e-01 -9.80603278e-01 -1.39824674e-02 3.34190607e-01
-2.34815538e-01 -1.25709355e+00 -6.17187805e-02 -4.67052549e-01
9.94743779e-02 -1.03881480e-02 8.77027333e-01 5.44651644e-03
-3.08895353e-02 -2.84722149e-01 -3.26711565e-01 -5.48759103e-01
-4.44471449e-01 6.07160926e-02 2.42108732e-01 1.61212236e-02
4.11190510e-01 -4.60975707e-01 -5.85586131e-01 2.69227505e-01
-6.65723026e-01 2.96415687e-01 7.01233864e-01 9.66485143e-01
-7.10189417e-02 6.13819242e-01 5.11584342e-01 -6.97910786e-01
1.29213405e+00 -2.93182313e-01 -8.12507629e-01 -1.14970142e-03
-1.84712052e-01 5.14354110e-02 7.42111206e-01 -2.49514565e-01
-1.00759304e+00 -2.35299304e-01 -2.43843511e-01 -4.03278992e-02
6.08357452e-02 7.11249232e-01 -1.50497764e-01 1.96961522e-01
6.08232796e-01 4.73753452e-01 5.40250465e-02 -4.65283468e-02
5.28743900e-02 8.97141814e-01 7.61250615e-01 -3.37518662e-01
3.80861580e-01 2.22255841e-01 -2.24494979e-01 -2.77443141e-01
-5.07167459e-01 -9.81835350e-02 -1.31066412e-01 -5.39888501e-01
5.06612718e-01 -1.00966036e+00 -1.49347830e+00 7.37770677e-01
-9.56362069e-01 -2.49379411e-01 -1.79809704e-01 4.10784334e-01
-7.11968005e-01 -3.18793096e-02 -5.74665010e-01 -1.06222999e+00
7.96163280e-04 -1.35571826e+00 4.88729239e-01 4.92727101e-01
5.35394996e-02 -7.28951156e-01 9.68521684e-02 3.95257592e-01
3.52936029e-01 1.75973922e-02 4.37109292e-01 -7.99450934e-01
-4.61544245e-01 -5.24374604e-01 2.12262422e-01 7.06192292e-03
-3.99070866e-02 -5.79569757e-01 -6.76261365e-01 -1.37979880e-01
3.84510159e-01 -2.45540053e-01 4.64614421e-01 3.00827354e-01
6.27291501e-01 -3.53722185e-01 5.72590977e-02 2.06877455e-01
1.19337475e+00 6.98783457e-01 7.13332355e-01 5.30804634e-01
7.70362839e-02 3.18676591e-01 6.98349059e-01 1.01200509e+00
5.87915659e-01 8.76001835e-01 1.00046992e+00 3.43889147e-01
3.87429655e-01 -4.92182910e-01 6.75370276e-01 6.05233014e-01
-3.33736092e-01 -4.66094851e-01 -5.58402002e-01 2.20096856e-01
-1.95573449e+00 -1.34457099e+00 4.63818133e-01 2.19454312e+00
6.40515268e-01 7.56378710e-01 1.59537464e-01 2.58969963e-01
7.11204946e-01 2.28998885e-01 -6.22358203e-01 -5.63311100e-01
-1.37006864e-01 9.53890532e-02 6.12745047e-01 4.37860101e-01
-9.53182161e-01 1.27250576e+00 5.97475910e+00 1.04571092e+00
-1.04114044e+00 -9.88431973e-04 6.51005983e-01 -3.48792821e-01
1.85903683e-01 -1.73381165e-01 -6.09577537e-01 6.97659016e-01
8.32158744e-01 -3.75376105e-01 6.96357846e-01 7.36971378e-01
4.25209701e-01 -5.35657585e-01 -7.02343345e-01 9.32448983e-01
-9.67223793e-02 -1.42807090e+00 -3.91238600e-01 1.63848907e-01
4.65278238e-01 -6.46039620e-02 1.18139565e-01 6.50533020e-01
1.01127219e+00 -1.04218435e+00 9.83568847e-01 3.09584826e-01
3.55934024e-01 -9.04882610e-01 1.16551733e+00 9.33036804e-01
-9.15988922e-01 -5.37597120e-01 -4.29632455e-01 -1.02516592e+00
9.80068073e-02 -6.43084571e-02 -9.64603484e-01 3.57271254e-01
2.43970901e-01 4.38228816e-01 -2.47786909e-01 8.88938129e-01
-4.33734506e-01 5.06352186e-01 -3.84435296e-01 -5.21060407e-01
5.87714553e-01 -1.10123552e-01 4.80067343e-01 4.87444013e-01
1.75131604e-01 5.93858421e-01 1.15274385e-01 5.83976865e-01
1.90472960e-01 -4.01121318e-01 -6.36112392e-01 -4.29726653e-02
4.51199591e-01 8.64171207e-01 -6.76061511e-01 -1.85926538e-02
-6.81613684e-02 7.64186502e-01 2.63722628e-01 7.10031688e-02
-9.31549072e-01 -1.15512848e-01 7.66604662e-01 -5.96353337e-02
4.84526567e-02 -1.15495451e-01 -2.88201243e-01 -1.16921413e+00
-1.59875870e-01 -1.04095793e+00 2.65918911e-01 -5.47104120e-01
-8.37487996e-01 7.04065204e-01 -5.10233343e-01 -1.26829576e+00
-7.36230016e-01 -6.62047923e-01 -8.42535615e-01 5.07744312e-01
-1.07375383e+00 -6.05250180e-01 1.04940154e-01 5.44661701e-01
4.46904659e-01 -6.83026373e-01 5.29474974e-01 2.75091752e-02
-6.49328530e-01 5.44083595e-01 -1.30147427e-01 5.18425882e-01
4.21882123e-02 -9.21704888e-01 2.15925172e-01 8.91023993e-01
8.74414667e-02 7.59564154e-03 9.39105630e-01 -5.51482379e-01
-1.56600201e+00 -4.34826374e-01 3.81755233e-01 -2.17704475e-01
7.23632395e-01 -3.04562628e-01 -2.21289188e-01 4.13783640e-01
1.84374135e-02 -4.33895767e-01 3.94497246e-01 -1.73202142e-01
2.58719712e-01 -9.19471495e-03 -9.88012195e-01 9.39999759e-01
8.66033614e-01 -2.97175914e-01 -8.14617813e-01 -1.82840765e-01
2.32209727e-01 -5.59825659e-01 -2.79801756e-01 1.48944959e-01
4.13085997e-01 -1.38998258e+00 5.64412057e-01 -8.19882393e-01
5.34034014e-01 -2.67314941e-01 -3.41543972e-01 -1.58696139e+00
-2.21453622e-01 -6.89690888e-01 1.23098925e-01 4.88013357e-01
3.02699417e-01 -4.13315862e-01 1.21583474e+00 5.20977259e-01
3.54179367e-03 -7.85295963e-01 -1.41736150e+00 -3.71479243e-01
-2.01058373e-01 -6.63939118e-01 5.84852040e-01 5.32206178e-01
6.92189336e-01 4.54332322e-01 -6.99007213e-01 1.20992362e-01
2.67202377e-01 3.13848769e-03 8.09184253e-01 -8.41063380e-01
-6.00172698e-01 -6.06612086e-01 -9.72094178e-01 -1.16157377e+00
6.35641664e-02 -5.01248479e-01 1.82273909e-01 -1.22559011e+00
-1.87787469e-02 -2.27066368e-01 -4.75790858e-01 3.12461853e-01
7.83153810e-03 1.38361707e-01 3.93358111e-01 8.62841159e-02
-8.79261613e-01 5.57456791e-01 1.28426611e+00 -5.65356910e-02
-3.50023285e-02 4.56955701e-01 -9.06813622e-01 7.50262201e-01
7.80356228e-01 -9.30395126e-02 -3.41099858e-01 -2.06458941e-01
5.17975271e-01 7.10677385e-01 1.19237080e-01 -1.22607374e+00
6.52911246e-01 -4.28851962e-01 -6.62651733e-02 -7.89755434e-02
7.03726649e-01 -7.16002762e-01 -6.21531568e-02 7.76747167e-01
-4.08294976e-01 1.62117109e-01 -7.10998848e-02 5.10936379e-01
-4.08509046e-01 -2.00783819e-01 4.88414377e-01 -3.56968939e-01
-1.04390156e+00 1.33365363e-01 -8.81559551e-01 -4.92160991e-02
1.27259982e+00 -3.26496184e-01 -2.71869361e-01 -9.40411866e-01
-5.40495336e-01 4.26348716e-01 1.64521873e-01 3.45857650e-01
6.12410545e-01 -1.01162219e+00 -8.12327087e-01 2.97329068e-01
-2.00662628e-01 -3.62832516e-01 2.51488328e-01 4.71999407e-01
-6.28565490e-01 5.00820637e-01 -5.34105122e-01 -1.50926886e-02
-8.53754103e-01 2.92093307e-01 6.98634148e-01 -6.63697302e-01
-2.92405576e-01 8.23944509e-01 1.81383997e-01 -2.76678294e-01
1.22191250e-01 -1.92294419e-02 -4.48897153e-01 -1.38333261e-01
5.45034051e-01 4.88469638e-02 -2.18388699e-02 -5.86916864e-01
-2.62358308e-01 -3.04431170e-01 -5.03219813e-02 -4.44684625e-01
1.13343513e+00 1.02788098e-01 3.17321837e-01 3.32380652e-01
3.44055265e-01 -2.63798147e-01 -1.50518000e+00 -2.23534063e-01
-1.13248609e-01 -5.45514166e-01 1.94108397e-01 -9.24832523e-01
-1.04726589e+00 8.29518616e-01 3.03701371e-01 2.75181115e-01
1.01230299e+00 -4.37336057e-01 5.07524490e-01 3.93498331e-01
9.47712123e-01 -1.26797175e+00 -3.37834060e-02 7.91748226e-01
3.70888621e-01 -1.19176733e+00 -3.26591760e-01 1.33064672e-01
-1.04986095e+00 7.28785992e-01 8.78666043e-01 -2.82918483e-01
4.69535291e-01 5.52098811e-01 1.20122269e-01 -1.67944729e-01
-1.39693272e+00 -2.33132944e-01 -3.42534810e-01 5.44374168e-01
-1.19049929e-01 3.73337388e-01 -2.59903580e-01 1.14012718e+00
-7.15960205e-01 1.32258549e-01 9.27055180e-01 9.49481606e-01
-6.60843432e-01 -8.60582471e-01 -2.32703477e-01 6.02830589e-01
-6.54845655e-01 1.57699157e-02 -3.91483575e-01 5.78374743e-01
1.83127850e-01 1.11448348e+00 2.65271157e-01 -8.40294302e-01
3.45163256e-01 -3.67156804e-01 2.54809946e-01 -4.67683047e-01
-7.73553252e-01 -2.45355502e-01 1.11722678e-01 -5.62431574e-01
-1.35969864e-02 -4.28321719e-01 -1.19358242e+00 -6.54321849e-01
-1.95343062e-01 3.88114095e-01 4.68408763e-01 1.23993003e+00
7.89196268e-02 6.79591060e-01 7.16804087e-01 -6.17441893e-01
-9.59993064e-01 -8.37183475e-01 -8.33792031e-01 1.51815280e-01
-2.42230538e-02 -8.59891951e-01 -1.66870922e-01 -6.11453414e-01] | [3.5622565746307373, 1.555107831954956] |
9bb74dce-917f-4257-8629-87028db0f45d | user-satisfaction-modeling-with-domain | null | null | https://aclanthology.org/2022.sigdial-1.59 | https://aclanthology.org/2022.sigdial-1.59.pdf | User Satisfaction Modeling with Domain Adaptation in Task-oriented Dialogue Systems | User Satisfaction Estimation (USE) is crucial in helping measure the quality of a task-oriented dialogue system. However, the complex nature of implicit responses poses challenges in detecting user satisfaction, and most datasets are limited in size or not available to the public due to user privacy policies. Unlike task-oriented dialogue, large-scale annotated chitchat with emotion labels is publicly available. Therefore, we present a novel user satisfaction model with domain adaptation (USMDA) to utilize this chitchat. We adopt a dialogue Transformer encoder to capture contextual features from the dialogue. And we reduce domain discrepancy to learn dialogue-related invariant features. Moreover, USMDA jointly learns satisfaction signals in the chitchat context with user satisfaction estimation, and user actions in task-oriented dialogue with dialogue action recognition. Experimental results on two benchmarks show that our proposed framework for the USE task outperforms existing unsupervised domain adaptation methods. To the best of our knowledge, this is the first work to study user satisfaction estimation with unsupervised domain adaptation from chitchat to task-oriented dialogue. | ['Georg Groh', 'Bernhard Pflugfelder', 'Mingyang Ma', 'Yan Pan'] | null | null | null | null | sigdial-acl-2022-9 | ['task-oriented-dialogue-systems'] | ['natural-language-processing'] | [ 2.56180108e-01 2.54357487e-01 -2.55232513e-01 -1.00490093e+00
-9.71724868e-01 -4.90180433e-01 5.24610579e-01 7.53184855e-02
-5.20057797e-01 9.75404203e-01 6.06322944e-01 1.62674591e-01
2.92365462e-01 -4.41993028e-01 2.27986854e-02 -3.19557875e-01
3.78352642e-01 7.45337427e-01 7.00854063e-02 -8.19876790e-01
3.82826179e-01 -2.33268812e-01 -9.00394201e-01 4.55242038e-01
1.22237873e+00 1.18709147e+00 2.91869957e-02 5.80857515e-01
-2.25361332e-01 9.02571619e-01 -6.53097093e-01 -6.59960091e-01
-1.36734173e-01 -7.21996486e-01 -1.37274325e+00 5.61606586e-01
1.06973266e-02 -5.41960120e-01 -2.32484713e-01 8.49005342e-01
6.16039813e-01 4.65060443e-01 5.78837097e-01 -1.40630031e+00
-4.28918988e-01 2.96459645e-01 -2.24477495e-03 -1.74211010e-01
8.70589972e-01 -9.44208354e-03 1.17982459e+00 -9.82276022e-01
5.55386782e-01 1.22852361e+00 4.11542356e-01 8.10947478e-01
-8.92551780e-01 -3.44337940e-01 -2.58193873e-02 1.18208766e-01
-8.53308856e-01 -6.06964111e-01 1.01442695e+00 -3.14149082e-01
8.80868673e-01 2.47548580e-01 1.98813945e-01 1.26326263e+00
-2.34594926e-01 1.13439393e+00 1.36542034e+00 -3.97359967e-01
3.23306978e-01 6.68709993e-01 3.77298892e-01 5.84626257e-01
-7.22872019e-01 -7.43786931e-01 -6.78725839e-01 -4.62664962e-01
4.69535083e-01 -4.49584313e-02 -2.27415338e-01 -4.28681076e-01
-7.64762700e-01 1.06681907e+00 -1.10205866e-01 2.69509673e-01
-2.97101647e-01 -6.89619303e-01 1.04434896e+00 9.18345571e-01
7.59558797e-01 5.39876282e-01 -9.75839674e-01 -6.94361866e-01
-9.98931378e-02 1.96725011e-01 1.49994063e+00 1.15819323e+00
7.50034392e-01 -1.38992310e-01 -2.66613603e-01 1.54452848e+00
-1.29028156e-01 4.14646357e-01 6.70313001e-01 -1.17510152e+00
4.58875716e-01 8.91004264e-01 2.15868995e-01 -6.80200636e-01
-4.47198510e-01 3.69853601e-02 -8.01973462e-01 -5.50659835e-01
2.74047762e-01 -6.17319882e-01 6.32902905e-02 1.57323492e+00
1.02785051e-01 -2.10346952e-01 4.47110444e-01 9.79835927e-01
1.12588418e+00 5.23211598e-01 2.99074072e-02 -5.77010989e-01
1.45031023e+00 -1.01878178e+00 -1.08122694e+00 -2.05331072e-01
9.52181935e-01 -5.45616210e-01 1.28504956e+00 3.81374568e-01
-9.99018550e-01 -4.11960334e-01 -4.52446580e-01 -1.54423833e-01
-1.78759262e-01 2.14482788e-02 5.96120238e-01 5.37927508e-01
-5.81432402e-01 -1.85890663e-02 -2.71831632e-01 -5.12630105e-01
7.30623975e-02 3.85824144e-01 -3.40689123e-01 -8.25293511e-02
-1.60027254e+00 1.05749238e+00 -3.53391692e-02 -1.71924144e-01
-3.23518276e-01 -8.32979232e-02 -1.20292044e+00 -1.76631555e-01
5.62621057e-01 -1.96944043e-01 1.93391371e+00 -1.37562251e+00
-2.05713797e+00 8.27827930e-01 -4.95766282e-01 -3.48131865e-01
2.57424980e-01 -2.82203913e-01 -2.96324581e-01 -1.74447805e-01
9.59064513e-02 2.09838882e-01 4.22731847e-01 -9.09044266e-01
-6.67368650e-01 -4.00434852e-01 2.88676381e-01 6.75519049e-01
-6.98691905e-01 2.22404048e-01 -4.04471517e-01 1.63449475e-03
-1.81070611e-01 -7.67401338e-01 -2.14382529e-01 -5.89254439e-01
-1.85205843e-02 -7.90684462e-01 8.03002179e-01 -6.30038977e-01
1.16180766e+00 -1.92446733e+00 2.19691694e-01 -1.28900051e-01
1.68688059e-01 3.14721048e-01 -5.92952929e-02 6.22495830e-01
4.15558994e-01 -4.13091391e-01 -3.91285628e-01 -4.62099522e-01
2.35914439e-01 4.14864123e-01 7.83226278e-04 2.08284333e-01
2.64197201e-01 7.69246519e-01 -1.07909024e+00 -5.44503570e-01
1.73552543e-01 -8.03341195e-02 -7.99870014e-01 1.01370811e+00
-2.38723159e-01 5.45748413e-01 -8.12998295e-01 4.28411484e-01
5.55653036e-01 -4.78739217e-02 4.72551525e-01 -1.12301685e-01
8.36637337e-03 5.70514381e-01 -5.89166522e-01 1.84828949e+00
-8.57056677e-01 4.53967541e-01 1.46969289e-01 -1.35818708e+00
1.31523621e+00 7.24749446e-01 6.30805790e-01 -9.97666240e-01
2.78884917e-01 1.46048114e-01 -1.43692181e-01 -8.00755918e-01
7.87318110e-01 -9.28573683e-02 -7.37842083e-01 4.68680680e-01
4.32552725e-01 -3.60986203e-01 -1.15508568e-02 2.41590872e-01
1.02984750e+00 -1.01426281e-01 6.10042274e-01 7.83929750e-02
9.23480272e-01 -5.45247411e-03 7.87978470e-01 3.88303161e-01
-6.90285683e-01 2.81172812e-01 7.75861502e-01 -1.13671996e-01
-6.20728195e-01 -4.44250435e-01 -7.55430833e-02 1.77346611e+00
-4.50200886e-02 -4.31633741e-01 -8.10917318e-01 -1.08295286e+00
-4.25491989e-01 6.32707238e-01 -3.43684733e-01 -2.24693939e-01
-4.13400501e-01 -3.23110312e-01 2.96424180e-01 3.21991920e-01
8.78450334e-01 -1.12697005e+00 -2.17978194e-01 3.29182476e-01
-7.96834886e-01 -1.43147457e+00 -6.88258231e-01 1.18398316e-01
-5.15060306e-01 -1.02031040e+00 -6.25040948e-01 -1.10541546e+00
4.11893159e-01 9.94882658e-02 1.27212512e+00 -3.17483753e-01
3.32782000e-01 6.66649759e-01 -7.13139772e-01 -1.90826043e-01
-3.25481713e-01 1.05352104e-01 1.27202153e-01 1.28609031e-01
9.35916901e-01 -1.15696706e-01 -3.38266879e-01 5.85702598e-01
-5.39903402e-01 -6.83205575e-02 2.17041269e-01 1.24330163e+00
-6.96868449e-02 -1.24962606e-01 1.05333841e+00 -1.37062442e+00
1.26659584e+00 -6.02591813e-01 8.10717344e-02 6.76880851e-02
-5.07767558e-01 -2.46536538e-01 6.79885030e-01 -2.00516075e-01
-1.52697444e+00 1.67969130e-02 -3.82617980e-01 1.35614738e-01
-4.33864564e-01 4.13420916e-01 -3.91126692e-01 2.67239273e-01
5.57478368e-01 2.61272371e-01 7.59854214e-03 -3.58068436e-01
1.81103453e-01 1.36963117e+00 3.79166335e-01 -6.62450969e-01
1.73138693e-01 -1.27283350e-01 -6.85701907e-01 -9.40838575e-01
-1.00381327e+00 -1.06662881e+00 -8.35660696e-01 -2.12903410e-01
7.49083459e-01 -8.68687868e-01 -7.72941470e-01 3.72853458e-01
-1.12036681e+00 -2.86179096e-01 -8.46824422e-03 2.16849029e-01
-8.51731777e-01 5.60859382e-01 -7.11118162e-01 -1.22351921e+00
-3.32891315e-01 -1.00432026e+00 9.52937424e-01 1.60199374e-01
-5.68918347e-01 -1.23479939e+00 1.22024536e-01 8.55960608e-01
2.60370016e-01 -1.11353181e-01 3.95404011e-01 -1.19681120e+00
3.37635666e-01 -2.83504516e-01 -2.02669606e-01 6.88515723e-01
3.23064983e-01 -8.03940833e-01 -9.25663173e-01 -1.16156332e-01
4.23449159e-01 -1.17748368e+00 2.73748487e-01 4.79503833e-02
8.40908766e-01 -6.19731188e-01 2.14660570e-01 -2.42405191e-01
6.98874652e-01 2.59271979e-01 4.88364458e-01 1.34641510e-02
3.46273929e-01 1.02669930e+00 1.37155843e+00 9.32077110e-01
7.93216467e-01 8.04298639e-01 -7.33714998e-02 -2.70214915e-01
4.97128755e-01 1.68255977e-02 5.77486336e-01 1.12978816e+00
1.22074500e-01 -1.99007601e-01 -6.87041163e-01 4.32737112e-01
-2.09922433e+00 -6.87363744e-01 -1.55748963e-01 1.84540367e+00
1.33028722e+00 8.96111596e-03 3.10969144e-01 -8.20882842e-02
5.58810592e-01 3.88159379e-02 -4.52646196e-01 -8.23343813e-01
1.05153777e-01 1.25312015e-01 7.13128448e-02 5.83314419e-01
-1.06479084e+00 1.05475748e+00 4.99875259e+00 4.54840362e-01
-6.37710214e-01 4.05861557e-01 5.21057546e-01 4.51102614e-01
-6.53528646e-02 -1.57450825e-01 -3.61439049e-01 2.65553862e-01
8.48804176e-01 -2.20533088e-01 2.17121411e-02 1.02408767e+00
4.03747082e-01 1.49206817e-02 -1.05328298e+00 8.77834320e-01
3.29582334e-01 -3.71603519e-01 -7.13012815e-01 -2.27180004e-01
5.72146475e-01 -3.76226395e-01 -6.39414415e-02 9.66641128e-01
4.07527715e-01 -6.15621209e-01 -2.91908205e-01 4.64739412e-01
6.97654068e-01 -8.60046268e-01 1.18309152e+00 5.51572502e-01
-7.48702228e-01 1.42658696e-01 -3.97610009e-01 -1.86107725e-01
7.08059669e-02 1.84146374e-01 -1.33022571e+00 1.81846604e-01
2.80990779e-01 8.76650035e-01 -1.62367150e-01 3.66512865e-01
2.16422930e-01 6.15641892e-01 2.64064699e-01 -2.51267701e-01
2.52620190e-01 -2.93335825e-01 2.42481589e-01 1.45723236e+00
-8.07581469e-02 4.67858732e-01 6.32414639e-01 4.20195103e-01
-2.51151353e-01 4.60527748e-01 -7.54268944e-01 -6.25010282e-02
3.32483321e-01 1.33516669e+00 1.92058593e-01 -2.89337307e-01
-8.35292459e-01 1.58085072e+00 2.55633622e-01 1.78768992e-01
-4.61411446e-01 -3.54839385e-01 6.30939186e-01 -2.67079145e-01
-2.99323678e-01 -6.59227744e-02 -8.84769261e-02 -1.28994083e+00
-2.13917926e-01 -1.15857959e+00 4.42479640e-01 -4.77698803e-01
-1.50171268e+00 4.02441978e-01 -4.03358221e-01 -1.23665273e+00
-5.05839586e-01 -4.09037083e-01 -4.45950121e-01 7.48675942e-01
-1.40300965e+00 -9.09012675e-01 -3.00510347e-01 8.77651453e-01
1.34112859e+00 -3.85933340e-01 1.29090071e+00 4.28093225e-02
-5.75238466e-01 6.61593318e-01 6.26270846e-02 2.92168885e-01
1.41370285e+00 -1.45306802e+00 -5.00355475e-02 -2.70421356e-02
-6.27290666e-01 4.29010242e-01 6.78580463e-01 -4.02013570e-01
-1.45164549e+00 -8.65132511e-01 1.24410939e+00 -5.00170887e-01
5.35744250e-01 -3.84376675e-01 -1.12255418e+00 5.04174113e-01
6.13083720e-01 -4.79708999e-01 1.18487096e+00 6.29843473e-01
-1.48166865e-01 6.87187165e-02 -1.30381024e+00 3.66614968e-01
6.46958709e-01 -6.60642266e-01 -8.08121502e-01 4.97336149e-01
6.22068048e-01 -4.26850736e-01 -1.25912404e+00 3.51922214e-01
2.67047197e-01 -8.86669993e-01 4.88170922e-01 -6.70046508e-01
5.02857387e-01 4.85611826e-01 -2.43046507e-01 -1.48698807e+00
-6.06596805e-02 -5.68130434e-01 9.18011069e-02 1.40110409e+00
2.05943093e-01 -4.61190343e-01 7.20755219e-01 1.22453129e+00
-2.83329844e-01 -4.11626250e-01 -5.88408589e-01 -2.66199529e-01
4.06758264e-02 -1.78565249e-01 2.71947354e-01 1.36135507e+00
9.36333597e-01 9.88642216e-01 -8.46128643e-01 -2.90761501e-01
2.19155535e-01 -4.51250374e-02 1.03914201e+00 -1.24765003e+00
-2.15208247e-01 1.33439787e-02 4.15676832e-02 -1.40648472e+00
5.74642539e-01 -3.67657572e-01 3.09325963e-01 -1.25846601e+00
1.91802964e-01 -3.06694448e-01 -1.12992458e-01 3.11240107e-01
-2.43022680e-01 -1.44847363e-01 -2.82436401e-01 -1.36950472e-02
-1.28357124e+00 8.60006690e-01 1.18850088e+00 -1.15758441e-01
-3.62104595e-01 2.34458044e-01 -7.21068621e-01 7.54808009e-01
1.08119118e+00 -6.92212656e-02 -4.92116153e-01 -2.95833293e-02
-2.44216532e-01 7.67984450e-01 -2.20505029e-01 -5.02597690e-01
1.44928053e-01 -5.34424663e-01 -1.44263849e-01 -2.45447367e-01
5.92443645e-01 -7.84226835e-01 -8.64504337e-01 6.63610101e-02
-8.31808865e-01 -4.07855392e-01 -1.94808900e-01 2.87510514e-01
-4.87310648e-01 -2.76444256e-01 6.93861246e-01 -2.27168381e-01
-8.85362566e-01 1.44884676e-01 -7.37586737e-01 3.68599266e-01
6.24497652e-01 1.19012438e-01 -1.10444725e-01 -1.05697572e+00
-4.98670906e-01 5.54836810e-01 1.46096408e-01 5.17088056e-01
7.38567173e-01 -1.25364375e+00 -7.96908379e-01 -3.96840982e-02
5.81125438e-01 -1.42639786e-01 2.29491889e-01 7.11000979e-01
3.18298906e-01 5.77793479e-01 -1.09564193e-01 -4.11152154e-01
-1.54536307e+00 1.20523967e-01 2.02212587e-01 -3.07974309e-01
-1.19144723e-01 6.57532752e-01 2.84263909e-01 -1.13365424e+00
4.12985921e-01 1.93066150e-01 -5.70924938e-01 1.46951243e-01
2.95852423e-01 2.66663820e-01 -3.07794977e-02 -7.81864226e-01
-7.94373900e-02 -7.67646916e-03 -1.95976898e-01 -1.77459106e-01
1.14005280e+00 -6.21708632e-01 1.61097571e-02 6.18383288e-01
1.31114352e+00 -2.22741619e-01 -9.53840613e-01 -1.00881076e+00
4.15445976e-02 -4.71052587e-01 -1.76186755e-01 -8.09739649e-01
-6.23405337e-01 8.28255832e-01 3.11924934e-01 3.51702839e-01
1.16852522e+00 -1.84229061e-01 1.16346359e+00 9.79945779e-01
3.55285406e-01 -1.77417207e+00 5.68678737e-01 1.08908653e+00
9.60909128e-01 -1.75244498e+00 -5.24156690e-01 -3.96333069e-01
-1.67413735e+00 8.71427417e-01 1.19927502e+00 2.05198690e-01
3.07548702e-01 -1.23698801e-01 3.28483552e-01 -3.12702656e-02
-1.04582202e+00 -3.89645994e-01 7.52426982e-02 6.48199260e-01
9.27157700e-01 -3.24081592e-02 -6.84019327e-01 9.35227931e-01
-1.19985089e-01 -1.90463867e-02 4.92424846e-01 9.82442498e-01
-6.43326938e-01 -1.47482586e+00 -3.61473076e-02 4.29831356e-01
-3.33643824e-01 -2.72655673e-02 -9.04357493e-01 2.39786848e-01
-2.91240305e-01 1.52252400e+00 -2.06478551e-01 -4.13691700e-01
8.14879060e-01 5.45206547e-01 6.66569024e-02 -8.46565187e-01
-7.83820808e-01 7.78924674e-02 6.75264359e-01 -3.06381196e-01
-4.45346981e-01 -5.62505126e-01 -1.21404815e+00 -1.34178624e-01
-3.42393965e-01 6.78226113e-01 4.27743047e-01 9.32619214e-01
1.23189777e-01 2.35689908e-01 1.36212373e+00 -3.00074667e-01
-5.89518309e-01 -1.23846030e+00 -5.75026274e-01 8.71680558e-01
1.65280595e-01 -4.85844642e-01 -7.88105726e-02 -5.87233640e-02] | [12.784561157226562, 7.934021472930908] |
81f60e72-24bf-4e75-8b76-de5bb1794d11 | modeling-and-recognition-of-smart-grid-faults | 1407.7008 | null | http://arxiv.org/abs/1407.7008v2 | http://arxiv.org/pdf/1407.7008v2.pdf | Modeling and Recognition of Smart Grid Faults by a Combined Approach of Dissimilarity Learning and One-Class Classification | Detecting faults in electrical power grids is of paramount importance, either
from the electricity operator and consumer viewpoints. Modern electric power
grids (smart grids) are equipped with smart sensors that allow to gather
real-time information regarding the physical status of all the component
elements belonging to the whole infrastructure (e.g., cables and related
insulation, transformers, breakers and so on). In real-world smart grid
systems, usually, additional information that are related to the operational
status of the grid itself are collected such as meteorological information.
Designing a suitable recognition (discrimination) model of faults in a
real-world smart grid system is hence a challenging task. This follows from the
heterogeneity of the information that actually determine a typical fault
condition. The second point is that, for synthesizing a recognition model, in
practice only the conditions of observed faults are usually meaningful.
Therefore, a suitable recognition model should be synthesized by making use of
the observed fault conditions only. In this paper, we deal with the problem of
modeling and recognizing faults in a real-world smart grid system, which
supplies the entire city of Rome, Italy. Recognition of faults is addressed by
following a combined approach of multiple dissimilarity measures customization
and one-class classification techniques. We provide here an in-depth study
related to the available data and to the models synthesized by the proposed
one-class classifier. We offer also a comprehensive analysis of the fault
recognition results by exploiting a fuzzy set based reliability decision rule. | ['Enrico De Santis', 'Alireza Sadeghian', 'Antonello Rizzi', 'Lorenzo Livi'] | 2014-07-25 | null | null | null | null | ['one-class-classifier'] | ['methodology'] | [ 1.45465480e-02 -3.38967413e-01 2.71808922e-01 -1.48775578e-01
-3.28712195e-01 -5.66706777e-01 4.67728645e-01 6.67927861e-01
-4.01173756e-02 8.89311790e-01 -5.13175726e-01 -3.39435071e-01
-6.67253733e-01 -9.78768885e-01 -1.01028932e-02 -1.09202087e+00
-1.97079271e-01 8.10030341e-01 -5.60286529e-02 -2.22368360e-01
3.14579546e-01 1.03762507e+00 -1.80503607e+00 1.07635930e-02
1.18757808e+00 1.35941017e+00 3.59140664e-01 3.25145274e-01
2.37476736e-01 3.09434295e-01 -1.16982174e+00 5.06944597e-01
-9.73337982e-03 -4.79739964e-01 -4.95120108e-01 5.93667686e-01
-3.30187589e-01 -1.34142399e-01 3.17439944e-01 1.22041941e+00
2.19771504e-01 2.30195135e-01 8.83575618e-01 -1.30884719e+00
1.60677463e-01 4.86719757e-01 -2.46742237e-02 3.53317350e-01
3.86693954e-01 -1.92308389e-02 8.99758160e-01 -4.74640518e-01
1.90110564e-01 4.89161134e-01 -8.82315636e-02 -2.34256670e-01
-1.04454851e+00 -4.84275166e-03 -8.13440979e-03 8.71511042e-01
-1.40225101e+00 -6.52162880e-02 8.63155186e-01 -5.78831494e-01
7.90502727e-01 4.79872823e-01 6.20369196e-01 5.23112059e-01
2.43081227e-01 1.06763259e-01 1.30113971e+00 -3.83569330e-01
6.64096594e-01 2.56429285e-01 1.68876410e-01 5.16457111e-02
4.92261678e-01 -1.40972510e-01 2.83939421e-01 3.00196167e-02
1.73037454e-01 -1.29861042e-01 -7.09114671e-01 -1.76068485e-01
-4.86120582e-01 6.01949453e-01 1.12161338e-01 1.33706224e+00
-4.69309270e-01 -7.56133318e-01 3.83760363e-01 5.15988231e-01
2.92177528e-01 4.08317059e-01 -3.89095545e-01 -1.26773104e-01
-9.46381450e-01 -2.56758004e-01 8.74087870e-01 4.79162663e-01
7.45258689e-01 4.59937990e-01 5.30821741e-01 3.13839644e-01
-9.86955613e-02 3.24773580e-01 6.98940933e-01 -1.17147200e-01
7.78570920e-02 6.46906018e-01 1.90896377e-01 -9.84651625e-01
-5.13860047e-01 -7.90196240e-01 -9.00198579e-01 5.54840565e-01
4.22012419e-01 -1.21640280e-01 -1.89547181e-01 1.10685861e+00
2.01618746e-01 -6.97554797e-02 1.24957614e-01 6.99944854e-01
2.33371317e-01 5.10288060e-01 -3.27186435e-01 -6.87518537e-01
1.32569504e+00 -1.84105471e-01 -7.51475036e-01 5.08550107e-01
6.15624070e-01 -5.63826442e-01 5.47147334e-01 1.14890850e+00
-6.81069791e-01 -5.33084095e-01 -1.30501616e+00 7.77596891e-01
-5.72350144e-01 4.04752105e-01 -1.15633376e-01 5.45816660e-01
-6.12661302e-01 8.97713184e-01 -5.26348710e-01 -3.06755722e-01
-3.04691136e-01 1.62028745e-01 -5.22381365e-01 1.43130571e-01
-1.11043513e+00 1.38677907e+00 5.87391853e-01 2.94124812e-01
-6.14086747e-01 -2.29397476e-01 -5.19723415e-01 4.04560030e-01
5.24098396e-01 1.09985374e-01 6.66147947e-01 -9.44530904e-01
-1.14048338e+00 1.89602852e-01 2.03474373e-01 -4.85392213e-01
5.46650589e-01 3.86313766e-01 -1.00179887e+00 4.73254770e-01
-8.42386261e-02 -7.14173734e-01 6.89689875e-01 -1.20409548e+00
-7.73449719e-01 -2.28181586e-01 -1.62605152e-01 -2.09080055e-01
-2.08861262e-01 -3.61400604e-01 5.98019123e-01 -3.35594177e-01
-6.88943788e-02 -4.50627953e-01 -6.01724312e-02 -6.87427163e-01
-3.69518131e-01 -3.91529202e-01 1.03334403e+00 -8.90955269e-01
1.10197973e+00 -2.06998396e+00 2.12198481e-01 8.56309831e-01
-1.50720090e-01 2.75960624e-01 4.09060925e-01 6.86616242e-01
-4.48705137e-01 -4.22297627e-01 -4.23519999e-01 2.51658499e-01
4.84174453e-02 4.13971931e-01 -1.24233313e-01 7.17577696e-01
8.72006938e-02 5.06037511e-02 -6.44489527e-01 -2.10847184e-01
7.71651804e-01 3.15800577e-01 6.73650131e-02 1.83703437e-01
-1.06150500e-01 6.38454914e-01 -4.61200953e-01 3.29962760e-01
5.30268133e-01 8.49828422e-02 3.44843745e-01 -4.30644989e-01
-1.93466648e-01 -5.39181791e-02 -1.70034909e+00 9.25469458e-01
-7.46000648e-01 7.14731365e-02 1.35769146e-02 -1.82989001e+00
1.28689539e+00 8.16430449e-01 8.28233361e-01 -5.98496854e-01
3.69711995e-01 5.16992092e-01 1.82130069e-01 -4.01208222e-01
9.51145068e-02 4.26806025e-02 1.74829498e-01 3.41053545e-01
1.27723515e-01 -1.41522586e-01 8.17652345e-01 -4.45091158e-01
8.69569063e-01 -2.76034355e-01 6.67052984e-01 -7.43187547e-01
9.99143839e-01 -2.65349984e-01 4.42526311e-01 -7.24759549e-02
1.76922530e-01 5.00261895e-02 5.86341500e-01 -3.35322112e-01
-7.20476151e-01 -6.71966732e-01 -5.54123104e-01 -1.73178777e-01
-1.61533743e-01 1.90242138e-02 -6.29613996e-01 -6.65089190e-01
-8.56339000e-03 9.04459119e-01 -3.48004341e-01 -1.50907502e-01
-2.72374064e-01 -9.78503764e-01 -6.11811392e-02 3.29796821e-02
2.23297790e-01 -8.68418276e-01 -9.76225019e-01 5.17018318e-01
-3.55825461e-02 -9.30412650e-01 3.74662131e-01 5.45736313e-01
-8.60450804e-01 -1.50254607e+00 -2.01367110e-01 -4.17778224e-01
8.83363247e-01 -2.23747924e-01 1.11820388e+00 3.79297704e-01
-4.69287962e-01 -1.47219658e-01 -7.05331624e-01 1.61353856e-01
-6.98581815e-01 -2.54188240e-01 1.31648645e-01 2.53630638e-01
2.55864784e-02 -8.91335011e-01 -7.28623495e-02 5.26181519e-01
-1.08609486e+00 -6.02871120e-01 4.39735472e-01 7.07858026e-01
1.73106983e-01 1.27061903e+00 8.33673835e-01 -5.57492912e-01
4.60087895e-01 -5.39521456e-01 -1.08361793e+00 4.04696405e-01
-7.67265618e-01 -2.42075518e-01 1.26195371e+00 -1.18037686e-01
-8.11026692e-01 -1.91346586e-01 -3.43999654e-01 9.34922546e-02
-8.52140367e-01 5.23015320e-01 -8.79263103e-01 -1.20012917e-01
2.86315560e-01 2.12252080e-01 -2.84250617e-01 -8.19495142e-01
-1.61511555e-01 7.87510991e-01 4.14205551e-01 -4.82389182e-01
8.40873122e-01 6.49521649e-02 3.57284755e-01 -1.03036916e+00
-3.05112340e-02 -3.74690413e-01 -7.35778809e-01 -3.61010909e-01
2.85466701e-01 -3.73753190e-01 -8.07847679e-01 4.96489435e-01
-8.29059541e-01 -1.61627680e-03 -7.33835757e-01 5.14418721e-01
-3.51797998e-01 6.91175878e-01 -4.40588295e-01 -8.51955414e-01
-2.04392765e-02 -1.42574990e+00 5.00387371e-01 2.90392917e-02
-1.28450513e-01 -1.20199549e+00 -2.52929211e-01 -1.01507485e-01
1.87435150e-01 3.91428918e-01 1.22456825e+00 -1.00326180e+00
-1.39066324e-01 -3.76499683e-01 2.77888954e-01 8.84180725e-01
7.26277947e-01 1.35124549e-01 -5.33946157e-01 -5.08949578e-01
5.58826685e-01 3.27363044e-01 2.74650156e-01 1.37993485e-01
8.36329460e-01 -9.53925550e-02 -1.72017235e-02 1.07911192e-01
1.87013173e+00 6.17843807e-01 4.98022884e-01 2.41975591e-01
1.46889597e-01 6.12099826e-01 7.92912304e-01 9.33104932e-01
3.19049461e-04 8.81011426e-01 4.51154023e-01 -5.57729751e-02
4.06465918e-01 5.10999203e-01 1.06505156e-01 7.60164142e-01
-6.89613968e-02 -3.76188844e-01 -4.84951407e-01 4.75073010e-01
-1.40739202e+00 -8.71508002e-01 -2.45624572e-01 2.30339289e+00
3.73992413e-01 3.05509359e-01 1.96559075e-02 1.41143513e+00
7.65147567e-01 -3.77512187e-01 -2.23937750e-01 -3.04280549e-01
-4.65445340e-01 2.12428585e-01 4.89908680e-02 5.17588377e-01
-8.39184046e-01 -2.00364664e-01 4.29391718e+00 9.53539610e-01
-1.25240552e+00 -1.45611867e-01 4.05438334e-01 3.65590453e-01
-2.99407970e-02 -7.23609105e-02 -3.68108004e-01 8.00409079e-01
8.59689057e-01 -2.66492218e-01 2.69043058e-01 7.20196486e-01
4.13507789e-01 -7.34058499e-01 -1.06159055e+00 6.43271327e-01
1.27928272e-01 -5.68021417e-01 -6.29865155e-02 1.19367108e-01
5.81865489e-01 -4.92333591e-01 -5.21313608e-01 -4.79903758e-01
-1.76127404e-01 -5.95121622e-01 6.25454485e-01 5.70922852e-01
4.27204251e-01 -9.11019385e-01 1.11914349e+00 3.11911106e-01
-1.20648062e+00 -3.43306988e-01 1.06547751e-01 2.57387012e-02
5.45803249e-01 1.37290752e+00 -5.51685870e-01 1.37578893e+00
6.00830615e-01 6.46386981e-01 -3.93532395e-01 1.11169577e+00
-4.08412129e-01 5.12690663e-01 -3.59318435e-01 2.41566271e-01
-4.61062491e-02 -5.31540036e-01 4.96720195e-01 7.19018519e-01
6.65792525e-01 -1.46220759e-01 1.43187419e-01 5.50568402e-01
5.95326722e-01 2.91573554e-01 -4.13223892e-01 3.20067972e-01
3.11291486e-01 1.46305919e+00 -1.04486859e+00 -6.38505161e-01
-2.72744864e-01 4.73695934e-01 -3.44640106e-01 2.06103608e-01
-2.97295451e-01 -5.55457056e-01 5.36513269e-01 3.62416692e-02
3.15216541e-01 -1.68850169e-01 -1.48643225e-01 -1.03967488e+00
1.98114693e-01 -8.50096822e-01 2.92975128e-01 -5.20118892e-01
-1.39058578e+00 9.29550886e-01 9.15272608e-02 -1.52721584e+00
-7.45956004e-01 -6.42574370e-01 -6.86466217e-01 1.10105228e+00
-1.42954338e+00 -5.62752366e-01 -1.29369289e-01 7.35167444e-01
2.84689397e-01 -1.24477699e-01 8.78129840e-01 4.88481373e-01
-6.63795233e-01 6.32224604e-02 3.52872670e-01 -7.97882453e-02
7.74688646e-02 -1.33793485e+00 -2.08780736e-01 1.15486062e+00
1.54733017e-01 -1.08649127e-01 1.01601493e+00 -4.53307599e-01
-1.00989211e+00 -4.68972385e-01 9.69969571e-01 4.03150648e-01
7.57710218e-01 1.13497779e-01 -1.02986634e+00 1.84227362e-01
2.51867741e-01 -3.78315039e-02 4.37895477e-01 -3.11714262e-01
2.60723561e-01 -3.87272775e-01 -1.49600661e+00 -1.75406206e-02
3.92247811e-02 -4.18730527e-01 -6.75643504e-01 2.18954280e-01
-4.31454748e-01 1.94655120e-01 -1.40209961e+00 4.42014992e-01
1.85741894e-02 -1.06384242e+00 4.93643522e-01 6.81067631e-02
-2.26150662e-01 -6.54937148e-01 9.88487080e-02 -1.81665456e+00
-7.30647743e-02 -4.85155992e-02 1.39451250e-01 1.13596582e+00
-7.88894948e-03 -1.14288723e+00 2.87039787e-01 2.91093420e-02
-1.40248433e-01 -5.20763695e-01 -1.24437523e+00 -7.60131657e-01
-3.30103487e-01 -1.36780396e-01 9.04302478e-01 1.01384687e+00
4.95402187e-01 -6.21883161e-02 1.37444600e-01 5.03722131e-01
6.40160203e-01 4.23796713e-01 9.52661633e-02 -1.57087994e+00
-4.26888824e-01 -4.79284167e-01 -7.85499334e-01 -1.21586584e-01
1.40251547e-01 -4.15506363e-01 -8.28741565e-02 -1.39395142e+00
-6.47189081e-01 -5.24956822e-01 -4.42896903e-01 1.99218512e-01
2.22312972e-01 2.51660705e-01 1.11799203e-01 5.34156077e-02
1.82280004e-01 2.50592053e-01 6.95124388e-01 -6.26524165e-02
3.46139312e-01 2.48437166e-01 5.51101156e-02 5.41568756e-01
9.56310451e-01 8.71530399e-02 -3.77257794e-01 9.78770107e-02
-3.05328015e-02 4.44881827e-01 9.17837694e-02 -1.40013409e+00
1.32595420e-01 -1.11570004e-02 2.64724731e-01 -5.33794284e-01
-8.46630894e-03 -1.47930157e+00 6.57530725e-01 7.39204049e-01
3.92076105e-01 2.25811988e-01 -1.68435290e-01 -3.66373248e-02
-5.60626388e-01 -7.10522890e-01 6.72585964e-01 4.12584096e-02
-6.32259130e-01 -2.03789324e-01 -7.60364234e-01 -4.80069637e-01
1.17392123e+00 -8.12533349e-02 -2.24291801e-01 -1.88010052e-01
-1.02046776e+00 -1.58342142e-02 4.61769938e-01 1.34605899e-01
2.74543285e-01 -9.35339153e-01 -5.95755160e-01 5.38654566e-01
5.70704788e-02 -3.08899641e-01 3.91833037e-01 9.25495386e-01
-3.24788451e-01 4.83810335e-01 -4.39257085e-01 -3.40331942e-01
-1.11814272e+00 6.97862506e-01 3.33973825e-01 -3.35720330e-01
-3.48558486e-01 -9.38159451e-02 -4.69609261e-01 4.59359258e-01
-1.74712732e-01 -6.23868525e-01 -8.16563308e-01 5.99030972e-01
3.77010107e-01 6.90615594e-01 1.02469659e+00 -8.76775861e-01
-2.63957113e-01 5.58317244e-01 5.99225044e-01 2.68762559e-01
1.21276915e+00 -6.90038130e-02 -3.70726138e-01 5.29323578e-01
8.68302524e-01 -1.79403946e-01 -8.60009730e-01 4.83003370e-02
1.97190478e-01 -2.72473663e-01 -1.31636835e-03 -8.41179907e-01
-1.32114601e+00 7.11496234e-01 4.64119673e-01 1.06151688e+00
1.58309269e+00 -3.51935148e-01 2.71626949e-01 9.98648927e-02
8.73684824e-01 -1.00988913e+00 -6.30380750e-01 -1.37696132e-01
6.27236843e-01 -6.28895700e-01 4.03663032e-02 -4.30107206e-01
-1.30151406e-01 1.37958229e+00 -2.64722351e-02 -1.64990664e-01
8.76003325e-01 5.30010641e-01 -7.99689889e-02 1.26340166e-01
-5.16454756e-01 -2.44336545e-01 5.88765889e-02 5.11811197e-01
8.50547999e-02 4.09451038e-01 -8.12069237e-01 4.72600132e-01
-5.72029278e-02 -9.25030187e-02 6.67080641e-01 7.92874575e-01
-3.73283446e-01 -1.47935927e+00 -7.19967723e-01 6.02616370e-01
-3.23784351e-01 2.79146791e-01 2.97012389e-01 5.15965462e-01
3.95250320e-01 1.45003164e+00 6.62912503e-02 -2.09008351e-01
5.89982212e-01 -9.13874432e-02 4.29938763e-01 -2.26991594e-01
-5.12862504e-01 -1.17984593e-01 -2.58052889e-02 -8.67606997e-02
-3.42080772e-01 -9.23534989e-01 -1.15987813e+00 -1.38838097e-01
-5.58646381e-01 1.05095792e+00 1.02463770e+00 1.28670096e+00
-3.04651529e-01 7.29676127e-01 1.23372602e+00 -9.10189450e-01
-6.43875122e-01 -1.08269835e+00 -1.51932907e+00 4.73741502e-01
1.80009753e-01 -8.23237956e-01 -7.42654920e-01 -1.05959967e-01] | [6.465163707733154, 2.4062306880950928] |
c2d81f18-db3c-4b10-bf7f-7f29a0f364bb | monocular-real-time-full-body-capture-with | 2012.06087 | null | https://arxiv.org/abs/2012.06087v2 | https://arxiv.org/pdf/2012.06087v2.pdf | Monocular Real-time Full Body Capture with Inter-part Correlations | We present the first method for real-time full body capture that estimates shape and motion of body and hands together with a dynamic 3D face model from a single color image. Our approach uses a new neural network architecture that exploits correlations between body and hands at high computational efficiency. Unlike previous works, our approach is jointly trained on multiple datasets focusing on hand, body or face separately, without requiring data where all the parts are annotated at the same time, which is much more difficult to create at sufficient variety. The possibility of such multi-dataset training enables superior generalization ability. In contrast to earlier monocular full body methods, our approach captures more expressive 3D face geometry and color by estimating the shape, expression, albedo and illumination parameters of a statistical face model. Our method achieves competitive accuracy on public benchmarks, while being significantly faster and providing more complete face reconstructions. | ['Feng Xu', 'Christian Theobalt', 'Ayush Tewari', 'Ikhsanul Habibie', 'Marc Habermann', 'Yuxiao Zhou'] | 2020-12-11 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zhou_Monocular_Real-Time_Full_Body_Capture_With_Inter-Part_Correlations_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zhou_Monocular_Real-Time_Full_Body_Capture_With_Inter-Part_Correlations_CVPR_2021_paper.pdf | cvpr-2021-1 | ['face-model'] | ['computer-vision'] | [-7.79646933e-02 -1.40222525e-02 7.19006360e-02 -4.23767358e-01
-4.47510719e-01 -5.88790894e-01 4.07139778e-01 -8.98428977e-01
-6.77003190e-02 5.22758722e-01 7.89606050e-02 2.66429067e-01
4.01079863e-01 -4.89370435e-01 -6.59319162e-01 -7.07407415e-01
2.61827037e-02 7.74863303e-01 -2.90819436e-01 -6.30686283e-02
-4.64347243e-01 9.39118981e-01 -1.63763666e+00 -9.61959437e-02
1.52545348e-01 1.18853390e+00 -4.87575352e-01 8.23497534e-01
2.31244385e-01 3.38387191e-01 -2.81178206e-01 -7.58102596e-01
6.76559746e-01 -2.82521814e-01 -4.68422204e-01 4.11508828e-01
1.33366668e+00 -7.04837024e-01 -2.63248354e-01 6.00507796e-01
9.40418243e-01 -3.28391939e-02 5.05972683e-01 -1.23837900e+00
-2.92519450e-01 -1.15460657e-01 -9.05314088e-01 -5.29677331e-01
6.30150735e-01 2.76852965e-01 5.23862660e-01 -9.67478335e-01
6.62304461e-01 1.54459107e+00 1.00397503e+00 9.32381570e-01
-1.36557877e+00 -1.00120580e+00 9.22380760e-02 -4.31921124e-01
-1.51195669e+00 -9.89281714e-01 8.70124638e-01 -2.82271892e-01
6.49430513e-01 5.82783855e-02 1.08055580e+00 1.20219469e+00
-2.81553477e-01 5.42862713e-01 9.51599121e-01 -4.12466556e-01
-2.63061404e-01 -2.18033880e-01 -4.58866119e-01 1.21555936e+00
6.95097446e-03 4.52976562e-02 -7.09384859e-01 -1.56087726e-01
1.15620840e+00 -3.18124406e-02 -2.08989710e-01 -8.63312125e-01
-8.40256929e-01 4.15632159e-01 8.21711197e-02 -8.12793151e-02
-1.96415201e-01 5.36074579e-01 1.11781619e-01 6.15348443e-02
5.54077387e-01 -2.74109006e-01 -5.51218212e-01 -2.81532691e-03
-1.19739962e+00 3.21679860e-01 1.03431141e+00 9.21296179e-01
6.88243449e-01 3.34957629e-01 1.05959915e-01 6.92154288e-01
4.44152653e-01 1.11477816e+00 8.75447765e-02 -1.36392057e+00
5.97840548e-02 5.26143312e-01 7.64876083e-02 -7.72790372e-01
-5.99580646e-01 -2.48318583e-01 -7.34970987e-01 6.91661298e-01
8.16716254e-01 -2.94251084e-01 -1.04512429e+00 1.90357101e+00
9.27973688e-01 1.09195448e-01 -2.26842299e-01 9.57545698e-01
1.00677633e+00 7.71297440e-02 -1.64498940e-01 -6.81833327e-02
1.28759170e+00 -8.51281226e-01 -4.67379391e-01 -3.13876748e-01
1.43336896e-02 -7.13936806e-01 7.36242354e-01 5.18106043e-01
-1.38109410e+00 -4.78884667e-01 -6.72736347e-01 -2.86159694e-01
-9.28038657e-02 3.78066152e-01 9.44562316e-01 1.05680549e+00
-1.51722074e+00 4.81459260e-01 -8.49439263e-01 -4.00705695e-01
3.71906966e-01 8.16955984e-01 -8.03180337e-01 -5.71621656e-02
-5.75976610e-01 6.61067367e-01 -2.61895180e-01 2.52899170e-01
-7.70293474e-01 -6.59191191e-01 -1.14928651e+00 -1.82389170e-01
2.17256024e-01 -7.63962805e-01 1.18159020e+00 -1.35453653e+00
-2.07450414e+00 1.34878504e+00 -2.17661038e-01 1.15386009e-01
8.43017519e-01 -2.98083395e-01 -1.45729899e-01 8.01182836e-02
-4.57347095e-01 9.64478135e-01 1.13281131e+00 -1.38007760e+00
5.27371727e-02 -7.91769385e-01 -9.55703035e-02 3.09616357e-01
-2.21535116e-01 1.06263310e-01 -1.15264285e+00 -3.02356809e-01
1.71263758e-02 -1.05143881e+00 1.27093717e-01 8.23838592e-01
-2.26755321e-01 1.99915230e-01 7.76737034e-01 -8.78727198e-01
5.81023157e-01 -2.05663490e+00 1.85500190e-01 1.07738905e-01
2.40735307e-01 1.71429679e-01 -2.02025875e-01 -4.76231128e-02
3.70634943e-02 -3.07449967e-01 -1.21843427e-01 -9.49762702e-01
9.85795707e-02 1.94900990e-01 1.39720812e-01 8.81063819e-01
-3.37109901e-02 9.31579947e-01 -5.26770055e-01 -6.98523819e-01
1.16724797e-01 9.73247707e-01 -6.59342408e-01 4.55784351e-02
-1.37318792e-02 5.31094432e-01 4.53883223e-02 1.13223910e+00
9.92175519e-01 -1.41633213e-01 4.71918076e-01 -3.14838350e-01
3.31731290e-01 -3.51453453e-01 -1.32675242e+00 2.00411844e+00
-5.80087841e-01 5.18316448e-01 6.29190981e-01 -5.09588659e-01
7.17018068e-01 5.58477402e-01 8.00827920e-01 -4.74260718e-01
2.34627783e-01 -2.86620259e-02 -2.89582223e-01 -2.68084943e-01
2.76508108e-02 -4.08849269e-01 1.66140765e-01 6.03877068e-01
1.40780434e-01 -2.90018380e-01 -2.97709644e-01 -1.84862435e-01
5.30437529e-01 8.73356402e-01 5.13889901e-02 -2.38944311e-02
3.03495705e-01 -5.17956913e-01 5.96827686e-01 1.70389876e-01
-2.26239204e-01 8.93126488e-01 1.61037222e-01 -6.15254998e-01
-8.04215670e-01 -1.14746964e+00 6.46071928e-03 1.14530253e+00
-1.28219342e-02 -1.56246737e-01 -7.35768855e-01 -5.29493034e-01
3.96221191e-01 -1.54160842e-01 -9.35249686e-01 2.50822574e-01
-6.83161139e-01 -5.52790880e-01 7.00435877e-01 7.03603327e-01
4.11184996e-01 -8.44550192e-01 -6.53392971e-01 -2.27673963e-01
-1.45259835e-02 -1.15512657e+00 -6.33160770e-01 -1.86572850e-01
-8.08314264e-01 -1.06475115e+00 -9.00408566e-01 -5.48677206e-01
6.71275735e-01 -1.35619462e-01 1.31647754e+00 2.61116475e-01
-6.67793155e-01 6.53859377e-01 1.83508217e-01 -3.65187049e-01
2.04916974e-03 -3.76416415e-01 2.94244856e-01 1.24976344e-01
2.22769063e-02 -7.10600138e-01 -6.87691510e-01 2.53438860e-01
-4.25460994e-01 5.61152473e-02 2.58448571e-01 6.49907053e-01
3.34225953e-01 -5.19465864e-01 -3.56800146e-02 -5.14095664e-01
-1.61312521e-01 -2.14349814e-02 -6.03808641e-01 2.64381319e-01
-2.20161781e-01 -1.99258804e-01 3.72377187e-01 -5.55173099e-01
-1.27792990e+00 5.70158780e-01 -8.01523924e-02 -7.24562466e-01
-2.62911767e-01 -3.00103366e-01 -3.15957725e-01 -5.03519833e-01
5.39866865e-01 5.53964116e-02 4.32874382e-01 -6.09763265e-01
4.29840595e-01 1.92425877e-01 7.84761071e-01 -6.69745862e-01
9.11936164e-01 1.01158023e+00 3.75072181e-01 -7.50930727e-01
-6.33614779e-01 -3.39669883e-01 -1.26703060e+00 -4.52239722e-01
5.48710227e-01 -1.07728601e+00 -1.02557027e+00 7.92507708e-01
-1.01653636e+00 -6.39848888e-01 -1.19194694e-01 3.03099424e-01
-4.92828280e-01 3.52300227e-01 -6.37247562e-01 -1.02549887e+00
-5.43758214e-01 -7.56615281e-01 1.71992838e+00 7.67490342e-02
-1.92865551e-01 -9.81866479e-01 1.77350268e-01 4.18394297e-01
2.92380452e-01 7.45314837e-01 3.15855742e-01 1.03269145e-01
-2.84183949e-01 -3.84269416e-01 -9.60333645e-02 3.95135134e-02
1.64663061e-01 3.00203413e-01 -1.37325180e+00 -5.52396476e-01
-4.45042998e-01 -7.13256240e-01 5.27285457e-01 4.06121075e-01
9.63442564e-01 -2.48670250e-01 -2.52789110e-01 1.22591221e+00
1.36982834e+00 -2.07101494e-01 3.65210980e-01 -3.58367473e-01
9.50500488e-01 6.92742348e-01 4.61723050e-03 6.25933051e-01
3.73391926e-01 8.20701241e-01 4.37155068e-01 -4.12500739e-01
-4.89850700e-01 -2.09504757e-02 4.93108660e-01 3.01333398e-01
-5.33928275e-01 6.31178841e-02 -8.56785715e-01 2.76702911e-01
-1.27319860e+00 -8.03851187e-01 2.11722896e-01 2.19979763e+00
8.06041896e-01 -5.91557384e-01 5.38863301e-01 -1.53435051e-01
3.44164491e-01 -1.28966242e-01 -6.73638582e-01 -9.64151323e-02
-1.81469902e-01 6.24779642e-01 2.35655442e-01 3.56662750e-01
-9.89253402e-01 1.00764596e+00 7.29021311e+00 3.36036295e-01
-1.27803767e+00 7.79638514e-02 5.02333105e-01 -8.72047961e-01
-1.00669026e-01 -3.70274723e-01 -6.73440516e-01 8.30860250e-03
4.82058167e-01 5.19763529e-01 6.90607786e-01 6.47934258e-01
-1.80078939e-01 -8.90176371e-02 -1.12215197e+00 1.38255107e+00
5.12687862e-01 -8.23698461e-01 -1.70028314e-01 2.44451180e-01
6.57917440e-01 -1.56699330e-01 9.78068337e-02 -2.85458751e-02
4.75908875e-01 -1.20424521e+00 9.23980474e-01 6.37878358e-01
1.40305340e+00 -5.94568789e-01 2.26249307e-01 9.21538770e-02
-1.07710195e+00 2.99513284e-02 -5.98917864e-02 3.71322148e-02
-8.82288516e-02 2.13192716e-01 -3.51961076e-01 3.72859836e-01
8.49241436e-01 4.54120159e-01 -4.74957734e-01 6.31649315e-01
-1.86929777e-02 2.61551797e-01 -6.76538408e-01 4.11809564e-01
-3.81678551e-01 -7.39284083e-02 2.38391116e-01 1.11391497e+00
2.88906932e-01 7.42614269e-02 1.93198636e-01 5.96717715e-01
-3.41524214e-01 1.74255759e-01 -5.76553166e-01 2.21287742e-01
6.17118850e-02 1.57353902e+00 -6.63333595e-01 -1.63524300e-01
-4.07958627e-01 1.19140708e+00 3.33037376e-01 2.71028489e-01
-7.08399832e-01 3.16636831e-01 7.73928046e-01 1.75248101e-01
2.34493032e-01 -2.45789960e-01 -3.40403207e-02 -1.25233483e+00
1.36401467e-02 -7.99266040e-01 2.79689580e-01 -8.71104181e-01
-1.15200615e+00 4.08018947e-01 -1.89078152e-01 -7.62988508e-01
-4.03599560e-01 -8.78240049e-01 -3.74589324e-01 8.57898533e-01
-1.30243492e+00 -1.81215751e+00 -6.55473888e-01 9.10160244e-01
9.54532102e-02 7.98811316e-02 1.00245237e+00 3.55185509e-01
-6.54188991e-01 9.04159307e-01 -1.42742664e-01 3.56515557e-01
1.00783122e+00 -1.18940675e+00 3.68739039e-01 4.18360740e-01
1.86602101e-01 5.14893293e-01 2.99298733e-01 -4.51920629e-01
-1.87990010e+00 -6.63142562e-01 3.45221549e-01 -7.86894202e-01
1.84993986e-02 -6.87892020e-01 -3.58040690e-01 6.69546127e-01
3.93534750e-02 4.66917723e-01 6.88744068e-01 3.31419259e-01
-5.91987550e-01 -3.70432973e-01 -1.30419922e+00 4.33944315e-01
1.33960271e+00 -5.09368181e-01 6.18942864e-02 1.36539370e-01
4.03217934e-02 -8.51354957e-01 -7.60705590e-01 2.66308606e-01
1.34324062e+00 -1.09262466e+00 1.20499372e+00 -3.74846429e-01
1.37778014e-01 -1.22380622e-01 -1.94501176e-01 -9.19588506e-01
-1.00568935e-01 -6.56207085e-01 -4.57503825e-01 1.06399465e+00
-6.15164042e-02 -2.71436036e-01 1.12271810e+00 8.46775055e-01
4.20822650e-01 -6.20195150e-01 -8.85626376e-01 -6.27424061e-01
8.05180073e-02 -3.52912396e-01 8.65787685e-01 8.58241796e-01
-5.29931486e-01 -1.38016164e-01 -7.24673748e-01 5.50584570e-02
8.51330817e-01 5.09492636e-01 1.28505337e+00 -1.52231872e+00
-2.71845043e-01 -2.81070352e-01 -2.87982643e-01 -6.71124518e-01
4.73071754e-01 -7.97300518e-01 2.29361765e-02 -1.14616859e+00
3.94541860e-01 -2.29691893e-01 2.15823904e-01 9.07604218e-01
1.36484236e-01 1.05126989e+00 8.79770368e-02 -7.51891807e-02
-3.27126414e-01 4.12430108e-01 1.27306890e+00 -9.14091524e-03
-5.61851375e-02 -2.30958372e-01 -4.32332337e-01 1.04168046e+00
3.11401367e-01 4.75193486e-02 -1.69408426e-01 -5.93956113e-01
1.08938307e-01 2.32889310e-01 7.87494123e-01 -9.78582025e-01
1.95303187e-02 -8.45089704e-02 1.21448064e+00 -2.92431623e-01
8.86560261e-01 -9.13346469e-01 5.83709121e-01 2.14375108e-01
2.63096303e-01 -1.21376105e-01 4.07999814e-01 3.08557630e-01
2.59764969e-01 3.43539983e-01 9.41891611e-01 -4.13923711e-01
-4.07364160e-01 7.66538620e-01 1.31606326e-01 -6.60835803e-02
7.20395982e-01 -3.61541390e-01 3.10910016e-01 -6.43759727e-01
-7.92046309e-01 -1.42999627e-02 9.03514147e-01 2.10249603e-01
4.23086673e-01 -1.35903811e+00 -6.67117059e-01 4.95294064e-01
-1.59622371e-01 -2.39641070e-02 3.06625664e-01 6.49144650e-01
-6.38821006e-01 3.80747207e-02 -2.95110971e-01 -6.85667455e-01
-1.73626900e+00 1.87897727e-01 8.36562753e-01 6.29924014e-02
-6.69111311e-01 9.46388006e-01 1.75253212e-01 -7.45613813e-01
3.29524487e-01 2.14024127e-01 1.75897941e-01 1.14159286e-02
3.48876297e-01 2.06492573e-01 -8.85752887e-02 -1.14086044e+00
-4.48294282e-01 1.20068729e+00 5.76002479e-01 -2.69607276e-01
1.36565423e+00 -1.61196485e-01 -1.84355557e-01 1.25117257e-01
1.07273149e+00 3.12938660e-01 -1.72105587e+00 -2.04854950e-01
-6.65885985e-01 -6.02210224e-01 7.08596259e-02 -9.50955391e-01
-1.58737516e+00 1.03112018e+00 7.01834857e-01 -5.88746428e-01
1.22626674e+00 1.01924082e-03 6.54756486e-01 3.18107009e-01
4.47149783e-01 -9.26253319e-01 1.33048519e-01 2.53313363e-01
9.07850266e-01 -1.36771476e+00 3.12710494e-01 -3.36480498e-01
-3.15965056e-01 1.23827779e+00 8.14708531e-01 2.92900562e-01
6.57823205e-01 5.79395235e-01 3.08189511e-01 -3.27211767e-01
-2.86673367e-01 -2.26729691e-01 6.24332607e-01 6.17823362e-01
4.88257498e-01 -4.20680121e-02 5.63379705e-01 1.79057896e-01
-3.62060815e-01 7.56218238e-03 -5.13955466e-02 8.25307667e-01
2.56632984e-01 -1.11046970e+00 -6.21141672e-01 1.23608872e-01
-5.58956206e-01 1.38297096e-01 -5.20174444e-01 9.40519154e-01
2.96683431e-01 5.12730658e-01 1.67707101e-01 -6.02071062e-02
5.09313382e-02 4.90331292e-01 1.23822498e+00 -4.58111584e-01
-4.88262832e-01 3.08718950e-01 -2.61050947e-02 -8.37016821e-01
-7.07607627e-01 -6.95853412e-01 -9.61438894e-01 -4.21739757e-01
-2.16080800e-01 -4.97735292e-01 7.45650351e-01 8.32487822e-01
2.73592830e-01 -1.95863675e-02 4.97495830e-01 -1.46690679e+00
-1.28114820e-01 -7.82997489e-01 -8.30300570e-01 4.80372876e-01
5.81366122e-01 -8.13338041e-01 -1.06545515e-01 3.29939544e-01] | [13.105039596557617, -0.06154513359069824] |
d7370e65-6cae-462e-893a-53207ccef749 | celebv-hq-a-large-scale-video-facial | 2207.12393 | null | https://arxiv.org/abs/2207.12393v1 | https://arxiv.org/pdf/2207.12393v1.pdf | CelebV-HQ: A Large-Scale Video Facial Attributes Dataset | Large-scale datasets have played indispensable roles in the recent success of face generation/editing and significantly facilitated the advances of emerging research fields. However, the academic community still lacks a video dataset with diverse facial attribute annotations, which is crucial for the research on face-related videos. In this work, we propose a large-scale, high-quality, and diverse video dataset with rich facial attribute annotations, named the High-Quality Celebrity Video Dataset (CelebV-HQ). CelebV-HQ contains 35,666 video clips with the resolution of 512x512 at least, involving 15,653 identities. All clips are labeled manually with 83 facial attributes, covering appearance, action, and emotion. We conduct a comprehensive analysis in terms of age, ethnicity, brightness stability, motion smoothness, head pose diversity, and data quality to demonstrate the diversity and temporal coherence of CelebV-HQ. Besides, its versatility and potential are validated on two representative tasks, i.e., unconditional video generation and video facial attribute editing. Furthermore, we envision the future potential of CelebV-HQ, as well as the new opportunities and challenges it would bring to related research directions. Data, code, and models are publicly available. Project page: https://celebv-hq.github.io. | ['Chen Change Loy', 'Ziwei Liu', 'Li Zhang', 'Siwei Tang', 'Liming Jiang', 'Wentao Zhu', 'Wayne Wu', 'Hao Zhu'] | 2022-07-25 | null | null | null | null | ['video-generation', 'unconditional-video-generation'] | ['computer-vision', 'computer-vision'] | [-1.97889790e-01 -4.37164724e-01 -1.50202483e-01 -5.93456924e-01
-5.13628006e-01 -1.84843823e-01 4.00305122e-01 -4.56357002e-01
-4.02293392e-02 7.05101252e-01 3.98026884e-01 4.40405756e-01
1.75975990e-02 -4.84173566e-01 -5.19554377e-01 -8.99953008e-01
-9.15562883e-02 -1.25036687e-01 -3.00706685e-01 -1.81345433e-01
-1.28502250e-01 3.16235334e-01 -1.84366894e+00 1.42314538e-01
6.77111030e-01 1.41348839e+00 -2.64553756e-01 2.37657040e-01
4.27169323e-01 5.76143980e-01 -2.35961333e-01 -9.72968698e-01
2.02080518e-01 -3.01192075e-01 -4.22633439e-01 2.84141660e-01
9.23156142e-01 -5.24731159e-01 -5.81330061e-01 9.59778070e-01
7.68625855e-01 6.98026419e-02 3.15351397e-01 -1.80685234e+00
-9.15131629e-01 -6.41517416e-02 -6.25940025e-01 1.41882570e-02
6.07045770e-01 3.56770873e-01 7.35108435e-01 -1.14111507e+00
9.11569595e-01 1.22824228e+00 6.77724779e-01 8.42247784e-01
-9.32016075e-01 -1.20338666e+00 9.68930349e-02 5.72086155e-01
-1.86449492e+00 -1.02067101e+00 6.59819186e-01 -5.37003100e-01
1.35541365e-01 2.95390755e-01 9.40747142e-01 1.39156628e+00
-2.43039370e-01 7.09729552e-01 9.85855699e-01 1.34254605e-01
-1.31294027e-01 -1.23599209e-01 -5.15830040e-01 6.59510016e-01
-5.45849912e-02 -3.48947421e-02 -9.09823537e-01 -3.80799398e-02
8.13165486e-01 -1.91264264e-02 -3.69256496e-01 -1.54317737e-01
-1.26932919e+00 6.26635790e-01 6.22004196e-02 -3.10710929e-02
-3.17861080e-01 -9.37375426e-02 6.13073885e-01 2.48923749e-01
6.78997576e-01 -6.48533404e-02 -2.00947523e-01 -2.63855785e-01
-7.91291118e-01 3.65824938e-01 2.74951190e-01 1.42332804e+00
4.64555949e-01 2.78517693e-01 -3.99808407e-01 9.93843675e-01
8.69911835e-02 6.82676196e-01 1.84878737e-01 -1.27911520e+00
1.19701982e-01 2.81409740e-01 -6.27731979e-02 -1.31531560e+00
-2.45389745e-01 1.72650158e-01 -1.07126749e+00 -1.98778838e-01
2.12208703e-01 -1.33647308e-01 -4.49865401e-01 2.04946375e+00
6.51845992e-01 4.74478424e-01 -3.34681600e-01 1.04894185e+00
1.30491292e+00 4.80973482e-01 1.98222771e-01 -4.94524837e-01
1.41382003e+00 -6.76302254e-01 -8.31474304e-01 3.74564886e-01
1.00575246e-01 -9.16013777e-01 1.03012717e+00 3.75997216e-01
-1.07464004e+00 -6.45288706e-01 -4.52687889e-01 -1.18225086e-02
1.35683879e-01 5.10252833e-01 8.38592350e-01 6.19989872e-01
-1.25906038e+00 1.83209836e-01 -3.64507526e-01 -3.88327122e-01
8.90584469e-01 2.67638981e-01 -8.39341581e-01 -2.15742826e-01
-1.19051218e+00 2.86373466e-01 -3.99544789e-03 1.22301333e-01
-9.24524188e-01 -7.30742753e-01 -8.08053970e-01 -2.93584079e-01
4.54103440e-01 -5.50118625e-01 1.09233642e+00 -1.21584511e+00
-1.37221658e+00 1.03939271e+00 -2.75853306e-01 1.05064951e-01
6.07927322e-01 -1.79118544e-01 -7.70158768e-01 3.20014685e-01
1.58118293e-01 7.95882523e-01 1.02515256e+00 -8.40176582e-01
-5.45164287e-01 -5.42395890e-01 -9.99596044e-02 5.98297119e-02
-8.95998836e-01 2.94370681e-01 -9.82142746e-01 -9.68145370e-01
-3.24630409e-01 -1.02755475e+00 2.47103736e-01 4.27683830e-01
-1.28702000e-01 -2.38126695e-01 6.31937385e-01 -8.56983244e-01
1.37073541e+00 -2.51169825e+00 7.52362311e-02 -9.59800556e-02
2.62885094e-01 2.13103130e-01 -3.09184849e-01 8.84679332e-02
-4.03808840e-02 6.69360831e-02 1.48336247e-01 -3.38115573e-01
-6.35293499e-02 -1.29015580e-01 8.82012919e-02 6.20670140e-01
1.11767031e-01 7.53296554e-01 -7.25657880e-01 -8.46131742e-01
-2.18778406e-03 8.28706980e-01 -7.16061294e-01 1.70010030e-01
1.16455063e-01 6.95158958e-01 -6.25859022e-01 1.12187266e+00
8.38254511e-01 5.50824627e-02 -2.13447273e-01 -5.39198339e-01
3.60007510e-02 -5.16464174e-01 -8.78363371e-01 1.76067340e+00
-1.16833188e-01 7.72674203e-01 3.39455843e-01 -4.28358823e-01
8.99632692e-01 5.01445234e-01 9.32795405e-01 -8.57026756e-01
1.55158088e-01 1.15303047e-01 -3.90849382e-01 -8.42547953e-01
5.34653425e-01 -3.24750952e-02 9.70327780e-02 3.69635448e-02
6.69729710e-02 3.12883139e-01 4.28992212e-01 2.28950530e-01
5.35980999e-01 2.65577286e-01 1.41217768e-01 -3.69808562e-02
5.62618315e-01 -4.90165412e-01 9.34666693e-01 -3.88730168e-02
-5.52178144e-01 7.24013686e-01 4.56733584e-01 -4.43681002e-01
-1.13346350e+00 -8.84920835e-01 -4.36278671e-01 1.20583332e+00
8.01782608e-02 -6.64815903e-01 -8.59197021e-01 -2.70370513e-01
1.99616943e-02 1.67042166e-02 -8.26319456e-01 -1.25000626e-01
-3.09749484e-01 -6.75346136e-01 6.21528745e-01 4.55547541e-01
7.87284851e-01 -1.09018064e+00 1.16613530e-01 -1.57508865e-01
-6.49124384e-01 -1.42675877e+00 -9.01089370e-01 -1.25032938e+00
-5.29440939e-01 -1.09679163e+00 -8.85031521e-01 -6.19908452e-01
6.07825398e-01 2.94109851e-01 1.09584570e+00 1.23627067e-01
-3.99485052e-01 5.59734404e-01 -4.84552979e-01 -1.49870858e-01
1.72110140e-01 -2.40794048e-01 4.40005571e-01 6.04302585e-01
3.01592350e-01 -5.35385609e-01 -8.52577865e-01 6.43707395e-01
-6.98967397e-01 1.32729977e-01 4.80546415e-01 6.96030676e-01
6.99972451e-01 -1.59401029e-01 6.53125823e-01 -5.55160522e-01
3.69425058e-01 -5.77240050e-01 -2.43690997e-01 1.07369900e-01
-2.72049010e-01 -6.99010491e-01 3.94028395e-01 -5.03688872e-01
-1.22758961e+00 -9.19244345e-03 -3.00751597e-01 -6.16429985e-01
-2.39362463e-01 1.01994708e-01 -5.94976008e-01 -1.08582996e-01
2.41442382e-01 2.27571651e-01 1.64726958e-01 -3.01702529e-01
1.35882542e-01 7.12311149e-01 7.01328337e-01 -7.41122901e-01
5.65731704e-01 5.02192914e-01 -1.26129866e-01 -1.02014184e+00
-6.37154877e-01 -2.10671082e-01 -4.53279793e-01 -6.21736169e-01
8.95238757e-01 -1.27716851e+00 -9.90473509e-01 9.31132436e-01
-8.65918458e-01 -1.75356474e-02 -2.11296324e-03 4.07332510e-01
-4.80147690e-01 1.91181019e-01 -6.05780423e-01 -5.54824054e-01
-2.83471525e-01 -1.02694297e+00 1.08391786e+00 4.48934197e-01
-4.68558893e-02 -5.58480382e-01 -2.67851502e-01 6.94028735e-01
4.20483977e-01 4.33176339e-01 3.73210311e-01 -1.59058064e-01
-4.92699414e-01 -4.31938730e-02 -3.68702173e-01 2.80764937e-01
1.02195859e-01 4.51305896e-01 -1.01108992e+00 -4.05637920e-01
-3.99031281e-01 -5.17321706e-01 5.02287686e-01 4.44433391e-01
1.45299971e+00 -4.14805710e-01 8.65210127e-03 8.93136919e-01
9.31036174e-01 1.04422383e-01 7.73653030e-01 1.13460712e-01
8.57553720e-01 6.20645761e-01 9.29738760e-01 8.65338027e-01
4.51450795e-01 9.28705812e-01 3.17978412e-01 -7.79081061e-02
-8.74433666e-02 -2.02754304e-01 4.53746468e-01 8.53330433e-01
-7.46678114e-01 8.16981494e-02 -5.36026835e-01 4.24422026e-01
-1.50500190e+00 -1.28302932e+00 -6.79564057e-03 2.13808966e+00
8.64580274e-01 -5.32226264e-01 4.23979133e-01 -1.53181210e-01
9.31774557e-01 2.33471587e-01 -5.69663405e-01 1.12323225e-01
-3.27817440e-01 -1.92500532e-01 -1.17057160e-01 -8.06211606e-02
-1.21996140e+00 8.03234398e-01 5.31791687e+00 1.08429801e+00
-1.06769514e+00 1.62983760e-01 1.04888546e+00 -4.64494973e-01
-8.74761194e-02 -4.84351426e-01 -8.08596194e-01 7.97012925e-01
6.45021498e-01 -4.00585681e-01 4.28070992e-01 9.27469075e-01
4.07056153e-01 2.68776357e-01 -8.86347055e-01 1.45830500e+00
5.06493628e-01 -1.20052588e+00 1.66070119e-01 3.41514051e-02
7.67689466e-01 -3.88450295e-01 3.20127845e-01 1.84765786e-01
-3.30096871e-01 -1.03930390e+00 7.13509083e-01 5.76345026e-01
1.51414418e+00 -7.94832587e-01 5.73346555e-01 -2.74320304e-01
-1.52674818e+00 1.78326759e-02 -3.34159136e-01 2.57737309e-01
1.08506843e-01 2.91955471e-01 -8.85149650e-03 5.40251791e-01
1.08223367e+00 1.25080168e+00 -6.82117939e-01 9.54351425e-01
6.35723397e-02 3.53520602e-01 -5.90242632e-02 2.94468164e-01
-2.48914912e-01 -3.31051201e-01 2.94458419e-01 9.36458051e-01
5.96231461e-01 4.90216672e-01 1.07789598e-02 3.15365940e-01
-4.55031365e-01 4.62273657e-01 -4.25408036e-01 -1.70942962e-01
7.52410591e-01 1.49476242e+00 -2.58864701e-01 -1.99285179e-01
-6.63826346e-01 8.28006327e-01 -9.77748632e-02 3.61798882e-01
-1.15705824e+00 4.71029722e-04 1.16200161e+00 1.83167234e-01
4.35580648e-02 9.19244736e-02 2.82140762e-01 -1.32471108e+00
1.34897768e-01 -1.21983635e+00 3.88242811e-01 -9.53218699e-01
-1.39017248e+00 8.27909529e-01 -9.91727561e-02 -1.48762190e+00
1.72631621e-01 -3.65172267e-01 -2.94049472e-01 4.26834285e-01
-1.18352151e+00 -1.36817527e+00 -9.51544344e-01 1.16058731e+00
5.71689785e-01 -4.37518209e-01 7.32036769e-01 9.31200981e-01
-9.13132966e-01 9.57565546e-01 -8.61990303e-02 3.26687813e-01
1.18181992e+00 -4.31804091e-01 8.07285905e-02 4.57431555e-01
-7.97529221e-02 3.90654802e-01 3.99064928e-01 -3.81801993e-01
-1.52903962e+00 -1.25726771e+00 5.77912748e-01 -2.62049913e-01
5.11336684e-01 -2.95431077e-01 -7.27329314e-01 5.91734588e-01
3.17082033e-02 3.30695868e-01 8.39458644e-01 -5.72882853e-02
-2.89790154e-01 -4.56604332e-01 -1.03612745e+00 5.91294467e-01
1.30723500e+00 -4.76393819e-01 2.23517820e-01 2.28973135e-01
4.38136607e-01 -4.39634234e-01 -1.20725310e+00 4.75965947e-01
9.69270408e-01 -1.11431289e+00 9.68860447e-01 -4.53454882e-01
5.32888949e-01 -1.78031567e-02 -2.10856989e-01 -9.94315982e-01
-2.77123332e-01 -7.65978515e-01 4.61427718e-02 1.73640037e+00
-5.98938800e-02 -3.79563063e-01 7.33806431e-01 7.44548738e-01
-1.84221659e-02 -9.05474842e-01 -8.99226189e-01 -5.03173530e-01
-3.09452772e-01 -2.67288744e-01 9.30014789e-01 1.07790625e+00
-4.57008034e-01 -1.75896659e-02 -1.00441134e+00 -2.28827953e-01
5.43710172e-01 6.03932925e-02 1.02398300e+00 -1.17693317e+00
2.16869548e-01 -3.77072215e-01 -6.78783238e-01 -7.52084792e-01
3.09527993e-01 -4.48966354e-01 -4.62899566e-01 -1.00086153e+00
4.76049513e-01 -3.88514936e-01 -6.90311864e-02 5.49569666e-01
-1.75307304e-01 7.38067746e-01 1.42438799e-01 2.88936257e-01
-8.49255502e-01 9.34239864e-01 1.55444467e+00 3.25376876e-02
2.02989236e-01 -7.02747479e-02 -6.78870618e-01 7.60217369e-01
7.60209084e-01 -2.34448686e-02 -3.72561753e-01 -2.96808720e-01
5.92995510e-02 7.95157254e-02 4.08718437e-01 -7.64827073e-01
-3.61219645e-02 -4.03064698e-01 6.22724056e-01 -4.43804227e-02
6.45631313e-01 -6.09223425e-01 4.25549060e-01 -2.99701327e-03
-1.26605660e-01 2.74841309e-01 1.03191093e-01 4.02912349e-01
-5.02754748e-01 4.44230825e-01 7.69722760e-01 1.72430068e-01
-1.17924798e+00 1.17360544e+00 1.33052289e-01 2.67163962e-01
1.34981978e+00 -2.37785488e-01 -1.70303151e-01 -6.04167700e-01
-7.37227261e-01 1.74474239e-01 7.14863360e-01 7.29885578e-01
7.85000563e-01 -1.90887487e+00 -1.11468470e+00 4.76155937e-01
3.85986090e-01 -3.88905823e-01 8.06411743e-01 1.13668692e+00
-2.87954628e-01 -2.06168909e-02 -7.59824097e-01 -5.15359223e-01
-1.74180973e+00 3.49186927e-01 -1.40033187e-02 5.79786301e-01
-4.89939094e-01 8.91040146e-01 2.40293190e-01 -5.65684885e-02
1.74940646e-01 5.05071998e-01 -3.11432123e-01 3.40820283e-01
8.12599063e-01 5.04718423e-01 -2.25600779e-01 -1.32903600e+00
-3.89205605e-01 7.67645776e-01 1.34307235e-01 3.03994715e-01
1.29927516e+00 -3.99982125e-01 -1.61457315e-01 3.01271062e-02
1.11305165e+00 1.07107393e-01 -1.38745546e+00 -8.49478245e-02
-5.34452379e-01 -1.05401421e+00 -3.28261793e-01 -2.86868602e-01
-1.73881829e+00 6.88107193e-01 6.22638047e-01 -3.18336934e-01
1.56836724e+00 -7.02140555e-02 8.40459049e-01 -1.56388983e-01
5.72428584e-01 -9.81025338e-01 2.06483006e-01 1.30423054e-01
1.04498434e+00 -1.42036462e+00 1.25507325e-01 -5.07320702e-01
-1.00968099e+00 7.86294639e-01 9.23165083e-01 3.89547765e-01
5.17785668e-01 -5.11846803e-02 4.87918779e-02 3.34132574e-02
-6.36210203e-01 9.57735553e-02 2.80314505e-01 6.74399555e-01
6.85355067e-01 -3.50958481e-02 -2.49548674e-01 6.99350953e-01
-3.68896008e-01 2.44317979e-01 9.71235186e-02 4.33887213e-01
2.91159227e-02 -8.71087193e-01 -2.54903823e-01 4.40910220e-01
-5.64962149e-01 -4.08116914e-02 -1.84408560e-01 7.99929321e-01
3.65150720e-01 8.31030488e-01 -5.26580922e-02 -6.00820541e-01
2.34434605e-01 -2.91833222e-01 4.24402595e-01 -9.57296416e-02
-1.22239187e-01 6.92424476e-02 1.77287266e-01 -9.10385966e-01
-6.22100413e-01 -7.61600912e-01 -7.71614969e-01 -9.51819479e-01
5.44749200e-02 9.08391178e-02 3.71607274e-01 3.64079773e-01
6.57315314e-01 1.33808568e-01 8.21826100e-01 -9.39460754e-01
-8.56777579e-02 -8.67775142e-01 -6.67434037e-01 8.00063908e-01
4.74444591e-02 -8.87310624e-01 -1.57059029e-01 4.53677684e-01] | [12.935416221618652, 0.17907829582691193] |
d561b245-7242-4300-bfed-ac9a9e19f025 | global-norm-aware-pooling-for-pose-robust | 1808.00435 | null | http://arxiv.org/abs/1808.00435v1 | http://arxiv.org/pdf/1808.00435v1.pdf | Global Norm-Aware Pooling for Pose-Robust Face Recognition at Low False Positive Rate | In this paper, we propose a novel Global Norm-Aware Pooling (GNAP) block,
which reweights local features in a convolutional neural network (CNN)
adaptively according to their L2 norms and outputs a global feature vector with
a global average pooling layer. Our GNAP block is designed to give dynamic
weights to local features in different spatial positions without losing spatial
symmetry. We use a GNAP block in a face feature embedding CNN to produce
discriminative face feature vectors for pose-robust face recognition. The GNAP
block is of very cheap computational cost, but it is very powerful for
frontal-profile face recognition. Under the CFP frontal-profile protocol, the
GNAP block can not only reduce EER dramatically but also boost TPR@FPR=0.1%
(TPR i.e. True Positive Rate, FPR i.e. False Positive Rate) substantially. Our
experiments show that the GNAP block greatly promotes pose-robust face
recognition over the base model especially at low false positive rate. | ['Zhen Han', 'Xiang Gao', 'Jia Guo', 'Yang Liu', 'Sheng Chen'] | 2018-08-01 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [ 4.16923203e-02 -2.51266688e-01 -1.38943508e-01 -5.74438870e-01
-5.13089538e-01 -3.32203507e-01 4.09248054e-01 -7.18904138e-01
-4.09428149e-01 2.96081662e-01 1.55212536e-01 1.36565328e-01
3.76811773e-02 -7.63426900e-01 -8.08939815e-01 -9.73080039e-01
-1.99523523e-01 -4.18542027e-01 7.05515966e-02 9.74779353e-02
4.66667935e-02 1.25862420e+00 -1.59845591e+00 4.76437151e-01
1.15369879e-01 1.46187794e+00 -1.12186195e-02 4.82023746e-01
1.77037552e-01 2.62599409e-01 -5.43775856e-01 -6.40732050e-01
6.07024729e-01 -1.00012027e-01 -5.30050814e-01 -4.57364507e-02
7.87219524e-01 -2.53054529e-01 -5.13862669e-01 1.06498003e+00
9.09269631e-01 9.36582237e-02 3.96432728e-01 -1.31881154e+00
-5.56350827e-01 9.38500743e-03 -6.78312242e-01 2.64169037e-01
9.75218415e-02 1.58769578e-01 6.75892532e-01 -1.57564056e+00
3.59776437e-01 1.38444185e+00 7.63924658e-01 7.99503505e-01
-1.20700240e+00 -9.56528068e-01 -2.27901200e-03 2.94907521e-02
-1.82612216e+00 -7.31401980e-01 7.12285161e-01 -1.94323007e-02
9.19262111e-01 3.35347503e-01 3.62876713e-01 8.46938789e-01
4.13302571e-01 4.47946817e-01 6.25870168e-01 -5.27645685e-02
-4.25793380e-02 -2.13388577e-01 -3.33298981e-01 9.14782882e-01
-1.06568197e-02 -2.06578299e-02 -9.37387943e-01 -8.67968053e-02
1.07067370e+00 3.01673532e-01 -4.46285218e-01 -1.52165532e-01
-8.56204093e-01 6.42099857e-01 8.60220492e-01 2.15684056e-01
-3.90372097e-01 2.72769898e-01 2.72995323e-01 4.10713643e-01
1.50480136e-01 3.35285872e-01 -6.18118346e-01 2.14134738e-01
-1.00744629e+00 1.39598757e-01 5.54593861e-01 8.69254291e-01
7.48769104e-01 1.00856908e-01 -4.82044607e-01 1.24163747e+00
4.63255644e-01 7.19605625e-01 2.61729777e-01 -9.10595596e-01
2.45561019e-01 5.21501899e-01 -2.28730202e-01 -1.42299747e+00
-4.64654982e-01 -2.85544872e-01 -8.54325593e-01 1.44128844e-01
1.97764650e-01 -1.34455264e-01 -1.01256001e+00 1.87108243e+00
1.62197948e-01 9.20786485e-02 -2.58678883e-01 7.04675198e-01
9.82195556e-01 6.44300401e-01 6.34758361e-03 -2.03621373e-01
1.48293591e+00 -8.32697570e-01 -4.93020236e-01 -2.06452146e-01
2.38262117e-01 -8.60432267e-01 7.87378490e-01 1.32285297e-01
-9.15686727e-01 -5.73552489e-01 -1.08126771e+00 2.31895726e-02
-3.59048814e-01 5.36440492e-01 3.60890388e-01 9.64905679e-01
-1.30131316e+00 6.42928839e-01 -7.13501871e-01 8.91274028e-03
8.68435979e-01 8.18417728e-01 -9.32788789e-01 -4.39591676e-01
-7.95227766e-01 4.73094106e-01 3.68877091e-02 5.89717984e-01
-8.58798265e-01 -6.84878170e-01 -9.00721729e-01 2.15058476e-01
2.50188351e-01 -6.32842183e-02 7.79358804e-01 -8.33096445e-01
-1.67621386e+00 6.24909580e-01 -5.58514595e-01 1.34392276e-01
2.48254597e-01 1.59123670e-02 -5.94622970e-01 1.34936646e-01
-1.34305552e-01 9.62619483e-01 1.21554708e+00 -8.38225424e-01
-2.11588204e-01 -6.15074098e-01 -2.82624871e-01 -1.49115279e-01
-4.93059605e-01 3.99186432e-01 -6.71436667e-01 -6.67417526e-01
3.18952024e-01 -8.11344028e-01 1.11613490e-01 4.67193812e-01
-4.36343215e-02 -3.03859264e-01 1.11543524e+00 -5.04489422e-01
1.03602087e+00 -2.38119411e+00 -1.94237977e-01 4.18247432e-01
8.31682533e-02 5.27232885e-01 -6.02427304e-01 -5.49999415e-04
-3.58083010e-01 1.27836823e-01 6.87544607e-03 -1.63060144e-01
-2.60261208e-01 9.96939465e-02 -6.66579232e-02 5.49172878e-01
6.01056218e-01 1.08967924e+00 -5.52615047e-01 -7.49657601e-02
-1.51729047e-01 1.05871856e+00 -8.78016949e-01 -4.06546071e-02
4.16873515e-01 -1.03950545e-01 -9.70038846e-02 9.69580173e-01
1.23731267e+00 6.94682971e-02 1.15513772e-01 -5.85262775e-01
7.77433366e-02 -1.37690425e-01 -1.11275101e+00 1.25836563e+00
-2.71232426e-01 6.93273127e-01 1.81013674e-01 -6.27095103e-01
1.10466170e+00 3.29117417e-01 1.21286623e-01 -7.40395069e-01
1.97346836e-01 3.16824108e-01 -1.67734966e-01 -2.05295578e-01
1.86477572e-01 -5.94788976e-02 1.48862645e-01 -1.48545712e-01
4.92449343e-01 4.37085927e-01 -2.15508133e-01 -4.15015817e-01
8.69713306e-01 -1.94104031e-01 6.36118129e-02 -6.78660214e-01
7.43048012e-01 -9.74991500e-01 8.57442796e-01 4.02831078e-01
-4.43058968e-01 9.41693366e-01 5.87604642e-01 -7.25126088e-01
-6.62941456e-01 -9.55999255e-01 -3.75574678e-01 1.09635794e+00
-2.10867882e-01 -4.11852568e-01 -6.25844002e-01 -9.54710603e-01
6.02913834e-02 -2.86270320e-01 -7.77119458e-01 -3.27409238e-01
-6.94593191e-01 -7.98376203e-01 6.53433383e-01 8.30052972e-01
1.00955534e+00 -1.02181864e+00 -2.46053368e-01 -5.11840060e-02
1.73353538e-01 -9.31903720e-01 -9.35625732e-01 7.41705522e-02
-5.85332632e-01 -8.90956581e-01 -9.94332790e-01 -9.76404607e-01
9.66675222e-01 2.61303842e-01 5.62576234e-01 -1.22418262e-01
-4.71583337e-01 1.77298691e-02 -1.55350670e-01 -2.25076094e-01
3.42611641e-01 -2.27502644e-01 2.27592394e-01 4.00222003e-01
5.45265079e-01 -3.38629991e-01 -8.71616364e-01 7.81599998e-01
-7.69996881e-01 -5.69060564e-01 3.44309032e-01 1.03382456e+00
5.09425938e-01 -9.13018882e-02 5.41873932e-01 -3.30633640e-01
3.45405608e-01 -9.86496825e-03 -4.94326025e-01 1.19298384e-01
-1.51550546e-01 -1.06770560e-01 5.54592490e-01 -4.05580074e-01
-6.95869088e-01 1.03086680e-01 -3.27986628e-01 -5.19328594e-01
2.09238321e-01 1.21713378e-01 -5.90372264e-01 -7.32387066e-01
6.46975100e-01 2.44261116e-01 1.20418407e-01 -4.58246291e-01
-1.11583741e-02 5.86292803e-01 3.83550406e-01 -2.25106329e-01
7.06573367e-01 3.37993234e-01 2.82621384e-01 -1.00370812e+00
-4.12261307e-01 -2.41382286e-01 -5.48670530e-01 -8.12925920e-02
5.03122807e-01 -1.00808358e+00 -1.02288282e+00 6.46892548e-01
-9.52163160e-01 -3.10784131e-02 3.04327272e-02 2.61102647e-01
-7.01329261e-02 9.62341949e-03 -4.59548593e-01 -3.96786243e-01
-3.64347219e-01 -1.30290604e+00 9.70850170e-01 5.31728208e-01
1.29460648e-01 -5.20055413e-01 -5.63383520e-01 1.67335942e-02
7.61199236e-01 1.28158793e-01 4.82204556e-01 -2.91156858e-01
-2.53668964e-01 -5.24969339e-01 -6.21451139e-01 7.72717893e-01
2.52122313e-01 -5.76604391e-03 -1.32763970e+00 -6.47825897e-01
-1.16680816e-01 -1.85273767e-01 9.00353849e-01 2.85775572e-01
1.40158153e+00 -5.94853103e-01 -2.01662928e-01 7.89930999e-01
1.35718358e+00 2.21803084e-01 9.66394186e-01 -1.09276921e-01
6.53409958e-01 3.82461250e-01 9.34814066e-02 2.77508676e-01
-1.93657070e-01 8.14004660e-01 1.40344962e-01 9.39748213e-02
-4.35055792e-01 -1.07855275e-01 5.79842985e-01 3.05692554e-01
-1.00588031e-01 3.53289917e-02 -6.61319137e-01 3.07286859e-01
-1.25555396e+00 -8.61358285e-01 5.72117388e-01 2.18749547e+00
7.16907799e-01 -4.13606912e-01 -3.82224172e-02 -3.56053561e-02
6.82713032e-01 3.26110870e-01 -2.00375125e-01 -4.13792670e-01
-1.08543232e-01 4.96892095e-01 5.85525870e-01 3.75296503e-01
-1.08752191e+00 7.21804142e-01 6.23208570e+00 9.67167258e-01
-1.54530311e+00 1.13696806e-01 7.86122024e-01 -3.84401947e-01
1.92897424e-01 -6.88205898e-01 -1.04787695e+00 4.39203441e-01
8.67924869e-01 1.21001966e-01 4.26234037e-01 9.95862305e-01
-8.34941864e-02 2.83476561e-01 -9.93129551e-01 1.35810721e+00
3.51146966e-01 -1.16121197e+00 3.48516047e-01 1.26817539e-01
6.50361955e-01 -5.01841344e-02 4.25025970e-01 1.61675334e-01
-3.39866370e-01 -1.51016092e+00 4.17629510e-01 2.73899049e-01
1.17503595e+00 -1.20163786e+00 8.78996193e-01 -3.40714008e-01
-1.40129447e+00 -2.03082338e-01 -8.05282354e-01 3.36528748e-01
-4.37204182e-01 4.44149077e-01 -4.73483771e-01 1.96320564e-01
8.96370947e-01 3.39344054e-01 -4.91518408e-01 9.91818249e-01
-1.41397133e-01 2.29752764e-01 -5.46799183e-01 9.92083177e-02
2.44983509e-01 3.47564846e-01 2.64061034e-01 1.30010426e+00
3.57212931e-01 8.99757221e-02 -1.65433347e-01 7.33530760e-01
-7.14089572e-01 1.29642084e-01 -4.86713231e-01 5.28307408e-02
4.76032436e-01 1.44970000e+00 -6.44409895e-01 1.05160587e-01
-2.17531249e-01 1.11259568e+00 1.20820291e-01 3.48744780e-01
-5.64729750e-01 -8.29731345e-01 1.01471651e+00 6.06696717e-02
6.20186269e-01 4.12043035e-02 5.70784733e-02 -9.55076158e-01
2.86073029e-01 -8.15790415e-01 -2.58759707e-02 -2.16148347e-01
-1.00978446e+00 7.64226496e-01 -3.47497761e-01 -1.01490986e+00
1.63024753e-01 -1.09484637e+00 -5.35241604e-01 1.00789595e+00
-1.36953855e+00 -1.07396090e+00 -4.04329211e-01 7.93647826e-01
2.08784431e-01 -2.29718089e-01 8.46604347e-01 5.15257776e-01
-7.92511344e-01 1.45937216e+00 -8.16027354e-03 4.72119540e-01
6.59821212e-01 -5.37990153e-01 3.73780906e-01 7.45754123e-01
-9.37922448e-02 1.06503987e+00 1.81442291e-01 -2.51123220e-01
-1.52950609e+00 -1.41719508e+00 9.90188062e-01 -2.26168171e-01
-3.03434525e-02 -4.71233428e-01 -7.71284103e-01 4.50242311e-01
-3.74901116e-01 8.22746038e-01 6.92208827e-01 -4.02673036e-02
-9.27469611e-01 -5.46885729e-01 -1.71682489e+00 5.93173802e-01
9.80561256e-01 -6.95009530e-01 -2.40236819e-02 1.65033624e-01
2.70978630e-01 -1.64963260e-01 -8.80225062e-01 6.18134856e-01
1.00269115e+00 -7.21000433e-01 1.13891542e+00 -2.39473209e-01
1.68892354e-01 -4.29652780e-01 -4.59173650e-01 -9.34953153e-01
-6.56808138e-01 -5.39648771e-01 9.66414288e-02 1.11295414e+00
5.07659137e-01 -8.22968781e-01 8.36046875e-01 5.16995430e-01
1.93062067e-01 -8.51765037e-01 -1.43694246e+00 -1.05172336e+00
-9.65846032e-02 -2.70281553e-01 5.87919533e-01 5.88808835e-01
-1.32277951e-01 -2.13371292e-01 -2.66765326e-01 3.11273903e-01
4.28461194e-01 -3.59546244e-01 3.44494700e-01 -9.51999187e-01
9.04289559e-02 -4.10365343e-01 -9.86267984e-01 -9.13671315e-01
-6.81528226e-02 -9.05000091e-01 1.14019655e-01 -7.33330011e-01
3.16341907e-01 1.95647534e-02 -5.13993680e-01 8.91455770e-01
2.97448523e-02 1.14812160e+00 2.03779489e-01 -1.29258543e-01
-1.77377596e-01 5.19155145e-01 1.11759293e+00 -1.42495811e-01
-1.60179168e-01 -2.29461923e-01 -6.94712460e-01 5.37469864e-01
8.33371937e-01 -3.61779958e-01 -1.10353790e-01 -4.16433960e-01
-2.96769738e-01 -5.17654061e-01 3.93092036e-01 -1.13244557e+00
1.46876305e-01 4.01640646e-02 1.23211110e+00 -8.25517625e-02
4.93383706e-01 -6.42053425e-01 -6.10830896e-02 4.99173850e-01
6.67773634e-02 -4.18172292e-02 6.06728196e-01 2.85485834e-01
-2.46801555e-01 9.83032361e-02 1.07092464e+00 1.79641441e-01
-4.52901930e-01 6.55191004e-01 -3.53509299e-02 -4.95634407e-01
8.96668971e-01 -4.29127872e-01 -1.06283329e-01 -5.86730987e-02
-5.14163196e-01 -4.04532664e-02 1.19430654e-01 5.48966646e-01
9.26173627e-01 -1.76110566e+00 -5.50617695e-01 9.60102439e-01
2.08781913e-01 -4.44122463e-01 3.50151002e-01 6.31550491e-01
-5.13536096e-01 5.04712224e-01 -4.50564265e-01 -3.96154612e-01
-1.58952987e+00 2.29865149e-01 5.42153418e-01 2.73501754e-01
-3.21062803e-01 1.54382968e+00 2.59150624e-01 -3.22480977e-01
2.62214631e-01 7.16677383e-02 -1.64720397e-02 1.35838538e-01
1.19540632e+00 2.60026753e-01 3.39799821e-01 -8.58816028e-01
-6.89457417e-01 8.30457091e-01 -4.40706640e-01 9.94310081e-02
1.35920012e+00 2.82898962e-01 -2.81739026e-01 -2.34661087e-01
1.87672222e+00 -3.88476346e-03 -1.38713837e+00 -1.18793443e-01
-3.82572234e-01 -8.56193006e-01 3.80138069e-01 -7.18317091e-01
-1.62323534e+00 8.47622693e-01 8.78198504e-01 -4.40742612e-01
1.24705338e+00 -9.09553021e-02 3.55125844e-01 3.54429543e-01
2.92141050e-01 -8.34527969e-01 7.34363422e-02 5.49749553e-01
1.33314490e+00 -9.33362126e-01 -1.29299179e-01 -4.13124442e-01
-2.79269636e-01 1.35548937e+00 6.48321331e-01 -9.56611335e-02
9.52505887e-01 3.34673166e-01 -1.01811774e-01 -2.63469577e-01
-3.52096707e-01 1.79826453e-01 5.94980061e-01 5.46334505e-01
4.32798564e-01 4.67382744e-02 -4.05595601e-02 5.76952755e-01
1.09320953e-02 6.49665892e-02 -5.60328476e-02 8.75944555e-01
-5.15719593e-01 -8.59061718e-01 -2.51079798e-01 4.14993137e-01
-6.07440829e-01 -3.02939601e-02 -2.21885011e-01 4.61023092e-01
2.08685875e-01 7.15962827e-01 2.43259117e-01 -7.41065502e-01
4.96148884e-01 2.32832819e-01 5.91508150e-01 -3.21055174e-01
-5.40866256e-01 -9.19436291e-02 -3.88409704e-01 -1.06683278e+00
1.35334237e-02 -4.10307556e-01 -8.42714310e-01 -6.25794172e-01
-1.78530186e-01 -2.82255560e-01 6.16068065e-01 4.78206217e-01
5.76467216e-01 2.30011284e-01 9.33862507e-01 -1.01120174e+00
-2.86214441e-01 -9.33695018e-01 -6.45124555e-01 -5.53018861e-02
4.06498611e-01 -6.34406805e-01 -3.43704045e-01 -2.44063333e-01] | [13.25134563446045, 0.6959513425827026] |
81f960ba-a365-45ad-b6c5-3f75d3d80ba8 | generative-modeling-for-small-data-object | 1910.07169 | null | https://arxiv.org/abs/1910.07169v1 | https://arxiv.org/pdf/1910.07169v1.pdf | Generative Modeling for Small-Data Object Detection | This paper explores object detection in the small data regime, where only a limited number of annotated bounding boxes are available due to data rarity and annotation expense. This is a common challenge today with machine learning being applied to many new tasks where obtaining training data is more challenging, e.g. in medical images with rare diseases that doctors sometimes only see once in their life-time. In this work we explore this problem from a generative modeling perspective by learning to generate new images with associated bounding boxes, and using these for training an object detector. We show that simply training previously proposed generative models does not yield satisfactory performance due to them optimizing for image realism rather than object detection accuracy. To this end we develop a new model with a novel unrolling mechanism that jointly optimizes the generative model and a detector such that the generated images improve the performance of the detector. We show this method outperforms the state of the art on two challenging datasets, disease detection and small data pedestrian detection, improving the average precision on NIH Chest X-ray by a relative 20% and localization accuracy by a relative 50%. | ['Li-Jia Li', 'Jia Deng', 'Tomas Pfister', 'Michael Muelly', 'Lanlan Liu'] | 2019-10-16 | generative-modeling-for-small-data-object-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Liu_Generative_Modeling_for_Small-Data_Object_Detection_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Liu_Generative_Modeling_for_Small-Data_Object_Detection_ICCV_2019_paper.pdf | iccv-2019-10 | ['small-data'] | ['computer-vision'] | [ 4.38489646e-01 3.53559226e-01 1.47564799e-01 -2.33205333e-01
-9.53389406e-01 -4.23439026e-01 4.56875831e-01 4.48852241e-01
-8.33291829e-01 6.80110574e-01 -1.18730284e-01 -2.75419623e-01
2.35197857e-01 -6.81333780e-01 -7.62201607e-01 -9.01295841e-01
1.29024103e-01 8.59597862e-01 5.37349701e-01 5.24266176e-02
-1.48922443e-01 5.29108107e-01 -1.49445486e+00 2.80033439e-01
6.87383771e-01 7.35318661e-01 3.31389129e-01 1.08780885e+00
4.11035687e-01 5.17875075e-01 -6.47137463e-01 -3.23021352e-01
3.37630421e-01 -3.88805568e-01 -5.43827713e-01 5.26018620e-01
4.72300112e-01 -4.25153792e-01 2.51693204e-02 9.23445582e-01
7.66413391e-01 -5.38189001e-02 8.04397285e-01 -9.68728721e-01
-3.43334705e-01 2.79961795e-01 -8.24658513e-01 2.99658209e-01
4.22187857e-02 1.32485196e-01 7.46973693e-01 -7.29112685e-01
6.33274376e-01 1.10679483e+00 5.01611888e-01 8.18976343e-01
-1.41419554e+00 -3.18513662e-01 5.31649799e-04 -8.09775367e-02
-1.44135177e+00 -2.05693796e-01 4.02596414e-01 -5.66798091e-01
6.42771006e-01 3.51539016e-01 4.22378063e-01 8.40733409e-01
-8.58306419e-03 6.72945321e-01 1.05767620e+00 -5.67280054e-01
2.40769461e-01 5.38614213e-01 -1.07758135e-01 8.86774659e-01
4.65727925e-01 6.84546381e-02 -1.86922178e-02 -2.83829626e-02
8.71174991e-01 1.04725517e-01 -8.80135894e-02 -6.06608927e-01
-1.19024897e+00 9.22679186e-01 6.96910679e-01 3.57070029e-01
-4.21542227e-01 5.04787713e-02 1.59644157e-01 -1.80477500e-01
4.77554798e-01 3.78782451e-01 -1.22585595e-01 2.48484805e-01
-9.56226110e-01 3.24097931e-01 6.16934001e-01 7.35004306e-01
4.20764059e-01 -2.26041988e-01 -3.18322808e-01 6.55827284e-01
1.25687465e-01 4.61352170e-01 2.64768332e-01 -4.15891021e-01
2.86495745e-01 6.66889608e-01 1.98462456e-01 -7.10103214e-01
-6.18778646e-01 -7.39604771e-01 -8.44900727e-01 3.44034225e-01
7.83996880e-01 -1.71628550e-01 -1.16119587e+00 1.51129305e+00
6.66602194e-01 -1.31110117e-01 -1.90014750e-01 1.04120827e+00
5.50615132e-01 4.10311669e-01 2.19390750e-01 -1.17125966e-01
1.63726723e+00 -8.82263064e-01 -3.31679642e-01 -2.77215123e-01
8.47896159e-01 -7.61239111e-01 1.00015199e+00 4.45412308e-01
-1.05795205e+00 -5.10759115e-01 -9.86559212e-01 -9.33817923e-02
-2.33953223e-01 5.63848019e-01 3.15638840e-01 9.24788117e-01
-8.03045750e-01 4.36870486e-01 -1.01354349e+00 -6.15241051e-01
8.15948486e-01 4.86167073e-01 -1.82681948e-01 -3.25868540e-02
-5.10090292e-01 9.72433329e-01 3.81156564e-01 -3.05316895e-01
-8.21718097e-01 -7.08479762e-01 -5.73217452e-01 -2.52920482e-02
6.05883360e-01 -9.08719063e-01 1.16974449e+00 -6.35271132e-01
-9.94365513e-01 1.04318261e+00 2.54887819e-01 -7.00704515e-01
1.05048192e+00 -2.40072742e-01 6.15694150e-02 5.39296195e-02
6.02031238e-02 9.25695360e-01 8.46512794e-01 -1.22838247e+00
-7.70128429e-01 -5.62836111e-01 2.97519714e-02 1.16510257e-01
-2.92381734e-01 -8.21657777e-02 -5.22549987e-01 -5.30560076e-01
-5.26147224e-02 -1.01925123e+00 -6.20878756e-01 1.90215349e-01
-5.42213380e-01 -6.53756559e-02 8.24591994e-01 -4.54674810e-01
1.02205086e+00 -1.82804251e+00 2.09738791e-01 9.11964029e-02
4.53931928e-01 3.62025917e-01 1.24631658e-01 -8.76550227e-02
1.13212325e-01 -1.35847390e-01 -4.22470450e-01 -5.52257001e-01
-2.10203812e-01 3.02170247e-01 -9.84077901e-02 6.12871349e-01
4.16989088e-01 7.59602427e-01 -8.28393281e-01 -7.84364164e-01
3.49721342e-01 6.25054598e-01 -7.66577363e-01 1.79581210e-01
-1.85405418e-01 6.92686319e-01 -3.32170904e-01 5.40356219e-01
5.76725662e-01 -5.54095745e-01 6.43619746e-02 -8.63363221e-02
3.61236930e-02 -1.91448137e-01 -1.27550089e+00 1.34864128e+00
-4.33945745e-01 4.13030952e-01 -1.44102737e-01 -9.80557561e-01
5.79292238e-01 7.13778809e-02 4.91498321e-01 -3.49689573e-01
3.44035655e-01 2.01934054e-01 3.28172266e-01 -4.79177564e-01
2.16957986e-01 -3.93985152e-01 6.95004463e-02 3.80380332e-01
-1.67827263e-01 -1.49426565e-01 3.74937564e-01 1.35705322e-01
1.17946064e+00 -1.80839315e-01 5.54399669e-01 -1.67669103e-01
4.41477537e-01 1.32058844e-01 2.12406561e-01 9.44584906e-01
-4.00043763e-02 9.48539376e-01 4.06909138e-01 -3.83631796e-01
-1.43353772e+00 -1.02812445e+00 -3.94435912e-01 1.04313624e+00
-1.47049397e-01 -1.07829362e-01 -1.02246737e+00 -9.58886206e-01
-9.27158222e-02 3.97443801e-01 -9.04867768e-01 -4.37053479e-02
-6.80054963e-01 -1.18758869e+00 3.75508338e-01 7.49739408e-01
3.46437395e-01 -8.34133685e-01 -1.18531835e+00 1.80561706e-01
8.88571069e-02 -1.27262759e+00 -3.24042380e-01 2.16866240e-01
-8.32443237e-01 -1.07442367e+00 -1.13837564e+00 -5.81630349e-01
1.08311784e+00 4.90890704e-02 1.07005727e+00 2.80890882e-01
-1.08316231e+00 8.50974098e-02 -2.88081199e-01 -6.00235820e-01
-6.07910454e-01 1.60455063e-01 -9.90088880e-02 -9.91231352e-02
1.15801863e-01 -6.96309134e-02 -7.71572053e-01 3.30688685e-01
-1.06130016e+00 1.96716100e-01 8.62589121e-01 9.96540487e-01
6.52374446e-01 -1.79259852e-01 3.09020996e-01 -1.20093846e+00
1.93962306e-01 -2.91981190e-01 -8.03460181e-01 1.69618696e-01
-4.07435715e-01 5.28254434e-02 3.32794875e-01 -5.84161043e-01
-7.95307219e-01 5.46364427e-01 -1.87734187e-01 -3.28491271e-01
-2.37516597e-01 -1.51589334e-01 4.56018113e-02 -6.88901022e-02
1.00684524e+00 6.40335400e-03 -1.21467307e-01 -4.69350964e-01
3.16660136e-01 4.99132305e-01 5.61926842e-01 -1.68088391e-01
7.37572610e-01 7.63240635e-01 3.15749496e-01 -6.98898733e-01
-9.34830189e-01 -6.88736081e-01 -9.14329350e-01 -1.37417123e-01
1.04201138e+00 -6.30027175e-01 -4.90901113e-01 -5.58631904e-02
-1.10127592e+00 -1.14040978e-01 -4.75511461e-01 4.70520377e-01
-5.99878132e-01 1.89883053e-01 -3.26251745e-01 -8.06012154e-01
-3.30218136e-01 -1.28437698e+00 1.49108315e+00 1.35509700e-01
-5.15087061e-02 -7.93634593e-01 -4.01934329e-03 2.49028012e-01
2.41355419e-01 4.27127391e-01 7.73554444e-01 -5.57367623e-01
-7.25057006e-01 -4.04935122e-01 -5.18427372e-01 2.58516967e-01
7.43989348e-02 -2.86430836e-01 -1.01009536e+00 -3.66926789e-01
-1.02025144e-01 -6.50473088e-02 8.89070809e-01 4.94770020e-01
1.26656258e+00 -9.61082280e-02 -6.21692598e-01 2.31892422e-01
1.37608552e+00 1.04063503e-01 4.77436453e-01 1.11295752e-01
6.87117279e-01 6.74602091e-01 5.79645813e-01 5.57704508e-01
-2.54408810e-02 8.68575335e-01 4.20578182e-01 -4.61702168e-01
-3.32655370e-01 1.24321021e-01 -2.40612552e-01 -2.07287520e-02
-1.72219902e-01 -2.26508796e-01 -1.17599404e+00 7.23121643e-01
-1.89897335e+00 -6.76667333e-01 -1.87520236e-01 2.37573099e+00
7.64713168e-01 1.87804520e-01 4.62126315e-01 2.23758280e-01
6.51678383e-01 -4.24353838e-01 -4.24964011e-01 1.91346765e-01
2.00723469e-01 3.11071873e-01 5.63565493e-01 4.46424216e-01
-1.37325585e+00 5.39877176e-01 5.81582546e+00 5.98210096e-01
-9.38246071e-01 4.10865933e-01 8.96658421e-01 -1.84530005e-01
3.86819661e-01 -2.50088900e-01 -9.99450266e-01 3.62019897e-01
6.03677690e-01 1.39921367e-01 -1.43635213e-01 9.99252677e-01
1.16233863e-01 -4.80642110e-01 -1.25162494e+00 1.03335297e+00
1.90289199e-01 -1.21009147e+00 1.63308643e-02 2.53626138e-01
5.96948385e-01 -2.21970528e-01 2.70345777e-01 -2.46179067e-02
9.12920833e-02 -1.00053060e+00 5.67990184e-01 2.63314694e-01
4.97118026e-01 -6.06384218e-01 7.67732084e-01 6.03349268e-01
-8.60967040e-01 -3.02629210e-02 -3.28972280e-01 3.39473456e-01
2.45054543e-01 5.95491648e-01 -1.36404335e+00 1.24400683e-01
4.84521240e-01 2.22272202e-01 -8.34656775e-01 1.39463246e+00
-8.58941227e-02 3.27900440e-01 -4.49380487e-01 -7.92163759e-02
2.13454306e-01 2.14816049e-01 6.14595294e-01 1.34712565e+00
3.04588675e-01 1.42285556e-01 2.65976816e-01 9.01926398e-01
-8.29780847e-02 2.25109652e-01 -4.49069053e-01 4.00134444e-01
1.03789549e-02 1.41688061e+00 -1.21113789e+00 -4.87061501e-01
-1.80650353e-01 9.08296049e-01 1.77721888e-01 -9.99650210e-02
-9.97024655e-01 -3.58357392e-02 1.22109853e-01 6.47940218e-01
3.86486441e-01 -7.27765188e-02 -2.87873358e-01 -9.30691719e-01
-1.75336339e-02 -5.45055211e-01 4.23519045e-01 -5.48907518e-01
-1.02417588e+00 7.56116450e-01 1.84999034e-01 -1.21241164e+00
-3.89265448e-01 -7.70995378e-01 -1.46371543e-01 6.36287868e-01
-1.23171759e+00 -1.16525459e+00 -4.53895032e-01 3.18631709e-01
5.93444407e-01 8.86726528e-02 6.16288066e-01 6.24472141e-01
-5.24463952e-01 5.16234875e-01 5.59498603e-03 7.81910866e-02
5.88344514e-01 -1.49598134e+00 3.01650494e-01 8.33890736e-01
3.92228335e-01 2.90790528e-01 9.41014946e-01 -4.88983572e-01
-9.43366230e-01 -1.17977214e+00 6.64060831e-01 -8.03876400e-01
3.47684264e-01 -5.75022161e-01 -7.57086277e-01 3.79271775e-01
-1.94137499e-01 4.94447976e-01 4.44252223e-01 -1.57013610e-01
-5.53882867e-02 9.96096209e-02 -1.33460379e+00 3.67926866e-01
8.87997985e-01 4.75594494e-03 -1.60575837e-01 6.57707155e-01
3.07880610e-01 -6.66879475e-01 -4.65678275e-01 3.55962336e-01
3.53591114e-01 -8.29041958e-01 1.03815866e+00 -5.90874553e-01
1.87851399e-01 -3.74433905e-01 -4.24280502e-02 -9.96241808e-01
8.61194432e-02 -2.22613662e-01 1.52884955e-02 7.30461478e-01
3.73734623e-01 -3.38088930e-01 9.63266671e-01 3.40074867e-01
1.93951532e-01 -8.12409282e-01 -9.59499300e-01 -7.06463456e-01
-1.12767763e-01 -2.32644960e-01 1.39470726e-01 3.90450686e-01
-4.29583520e-01 2.90334880e-01 -2.96369731e-01 1.90769911e-01
7.30993807e-01 -3.22315916e-02 9.10902977e-01 -1.12169123e+00
-5.32309651e-01 -3.15817744e-01 -6.26812935e-01 -8.26070428e-01
-5.38513958e-01 -6.57233298e-01 1.45320877e-01 -1.30136681e+00
5.81538796e-01 -5.14379323e-01 1.46861479e-01 3.34251791e-01
-3.63155782e-01 6.83822691e-01 2.12745845e-01 1.37040794e-01
-6.13831401e-01 2.38719434e-02 1.21910846e+00 -2.05843691e-02
-1.35433555e-01 2.90377438e-01 -5.13775587e-01 8.40083063e-01
5.07339299e-01 -7.03869998e-01 -1.84754550e-01 -2.57757783e-01
-3.44951376e-02 -1.71492547e-01 8.42739522e-01 -1.22460008e+00
1.07381582e-01 3.10425520e-01 5.97196639e-01 -5.05536079e-01
3.22990060e-01 -7.99499929e-01 -7.07643628e-02 7.76381195e-01
-1.91090673e-01 -9.75486338e-02 1.09011590e-01 7.32199609e-01
1.10220127e-01 -2.72382110e-01 1.23317766e+00 -1.52022630e-01
-4.14183617e-01 1.57082602e-01 -2.62064673e-02 -6.00658134e-02
1.37180698e+00 -1.67861387e-01 -1.16372913e-01 -1.31667376e-01
-1.09201288e+00 -1.17654711e-01 3.05769145e-01 2.74151117e-01
4.53687131e-01 -1.07266760e+00 -8.10154378e-01 1.93412155e-01
1.74138114e-01 1.96087569e-01 1.70708671e-01 1.13098407e+00
-5.55360079e-01 3.50552589e-01 -4.79913084e-03 -1.06574416e+00
-1.71007121e+00 6.97230518e-01 3.48293751e-01 -5.92137039e-01
-8.11134815e-01 8.78628314e-01 6.72254503e-01 3.22051235e-02
2.75566727e-01 -5.27165949e-01 -6.94260225e-02 -2.28165071e-02
5.40594578e-01 3.28300983e-01 3.10931861e-01 -6.23986483e-01
-2.52350569e-01 5.02977073e-01 -4.56531763e-01 5.73082715e-02
1.13270938e+00 5.27040325e-02 4.13727343e-01 1.22843049e-01
9.87350762e-01 -2.19563216e-01 -1.23714685e+00 -2.24567279e-01
-7.94109479e-02 -6.67876005e-01 7.48521322e-03 -7.16268837e-01
-8.50489318e-01 1.05568302e+00 1.06380546e+00 2.29906380e-01
1.07725585e+00 5.22523403e-01 5.11948645e-01 3.13166440e-01
3.20442587e-01 -8.83959234e-01 3.13880652e-01 -2.48190500e-02
6.76134288e-01 -1.52090096e+00 2.30737418e-01 -5.33091903e-01
-6.42717779e-01 8.48802328e-01 4.60663676e-01 -2.51281917e-01
4.53329146e-01 4.26095903e-01 -1.75079077e-01 -1.99042723e-01
-5.02756715e-01 -5.21016896e-01 5.79427481e-01 5.26527762e-01
3.55920374e-01 2.12124661e-01 -3.13330770e-01 3.49520296e-01
-5.14915735e-02 -1.66319638e-01 3.85991782e-01 9.92065012e-01
-6.67598844e-01 -1.23744953e+00 -6.66004837e-01 6.09303951e-01
-6.94905758e-01 1.13192402e-01 -1.86267614e-01 9.68973339e-01
3.62201303e-01 6.48674071e-01 1.65508986e-01 1.57247320e-01
3.86741549e-01 -2.43900299e-01 6.52847767e-01 -9.97538209e-01
-4.59322751e-01 2.61309028e-01 -1.44441783e-01 -3.29218894e-01
-4.71691728e-01 -7.71395862e-01 -1.15709829e+00 1.01406492e-01
-5.65211475e-01 -2.21042767e-01 7.01070130e-01 8.64445508e-01
5.54261245e-02 7.17748463e-01 2.67412335e-01 -8.20229471e-01
-6.46251619e-01 -7.77129591e-01 -3.99490863e-01 3.39723796e-01
4.48827147e-01 -7.64633298e-01 3.88480723e-02 3.62970293e-01] | [15.010220527648926, -2.423823356628418] |
52964014-da4f-4834-a19b-1d211f943472 | visual-relationship-detection-with-language | 1608.00187 | null | http://arxiv.org/abs/1608.00187v1 | http://arxiv.org/pdf/1608.00187v1.pdf | Visual Relationship Detection with Language Priors | Visual relationships capture a wide variety of interactions between pairs of
objects in images (e.g. "man riding bicycle" and "man pushing bicycle").
Consequently, the set of possible relationships is extremely large and it is
difficult to obtain sufficient training examples for all possible
relationships. Because of this limitation, previous work on visual relationship
detection has concentrated on predicting only a handful of relationships.
Though most relationships are infrequent, their objects (e.g. "man" and
"bicycle") and predicates (e.g. "riding" and "pushing") independently occur
more frequently. We propose a model that uses this insight to train visual
models for objects and predicates individually and later combines them together
to predict multiple relationships per image. We improve on prior work by
leveraging language priors from semantic word embeddings to finetune the
likelihood of a predicted relationship. Our model can scale to predict
thousands of types of relationships from a few examples. Additionally, we
localize the objects in the predicted relationships as bounding boxes in the
image. We further demonstrate that understanding relationships can improve
content based image retrieval. | ['Li Fei-Fei', 'Michael Bernstein', 'Cewu Lu', 'Ranjay Krishna'] | 2016-07-31 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [-7.13536702e-03 -1.27408892e-01 -4.56321865e-01 -6.81303859e-01
-2.57250309e-01 -6.86032712e-01 9.10989404e-01 3.13487560e-01
-1.82523802e-01 4.12393421e-01 2.09093675e-01 -3.64008874e-01
-1.09171867e-01 -7.75397241e-01 -9.90331829e-01 -2.22081915e-01
-2.89964527e-01 5.77285945e-01 5.02316833e-01 -2.43860990e-01
5.91907501e-02 4.97812867e-01 -1.72412086e+00 5.85298240e-01
1.75081253e-01 7.17901468e-01 5.43789506e-01 4.93191451e-01
-1.39900625e-01 7.92472005e-01 -5.76623440e-01 -4.35827821e-01
6.45059422e-02 -1.35700643e-01 -8.78707051e-01 2.96749204e-01
7.56718278e-01 -5.91096699e-01 -5.75297236e-01 8.70613039e-01
-1.41111791e-01 3.78914207e-01 6.67111278e-01 -1.69991362e+00
-1.08427536e+00 3.33057165e-01 -8.44567955e-01 4.97207433e-01
6.01330996e-01 -5.78606538e-02 1.51362789e+00 -9.82062340e-01
8.23713183e-01 1.53497922e+00 1.75190270e-01 2.76016712e-01
-1.24573696e+00 -8.57017457e-01 6.36960149e-01 5.38198292e-01
-1.41175103e+00 -2.13106826e-01 4.99712169e-01 -4.81542796e-01
1.25840557e+00 2.71325290e-01 8.51236284e-01 8.51513624e-01
2.45391950e-02 9.27229702e-01 5.88524997e-01 -4.60246861e-01
-3.09714735e-01 4.02039528e-01 2.53174514e-01 6.45802617e-01
2.88694799e-01 -1.93986028e-01 -6.48797095e-01 -5.54675385e-02
8.01275313e-01 2.03217849e-01 5.39102331e-02 -6.76083088e-01
-1.28245676e+00 6.93777561e-01 6.79979861e-01 8.22381750e-02
-3.37995328e-02 4.48520720e-01 -3.64287607e-02 -2.34151483e-02
2.60278285e-01 3.68338197e-01 -2.01824531e-01 1.99716270e-01
-4.47069466e-01 4.82421547e-01 6.75075591e-01 1.43294609e+00
1.05308807e+00 -6.47186935e-01 7.65108243e-02 9.64533031e-01
3.37100893e-01 2.24684700e-01 -3.86796556e-02 -7.58465886e-01
6.10375583e-01 5.41456521e-01 4.56308007e-01 -1.35622251e+00
-3.21578532e-01 1.00617178e-01 -1.44756466e-01 -1.34581223e-01
3.05002183e-01 3.76244187e-01 -9.46175814e-01 1.61066961e+00
3.09764624e-01 2.03084320e-01 -3.03033143e-01 1.01643872e+00
9.48571563e-01 8.01716924e-01 5.30074596e-01 2.81421453e-01
1.75960767e+00 -1.16129804e+00 -5.96586227e-01 -6.56325400e-01
5.07319391e-01 -9.23486352e-01 1.30020547e+00 -9.35780033e-02
-8.96952331e-01 -5.35629153e-01 -7.18177378e-01 -4.35477734e-01
-8.29537570e-01 -6.97697997e-02 1.00209570e+00 1.45627424e-01
-9.66772914e-01 2.85877794e-01 -7.92019129e-01 -7.09602833e-01
5.33350587e-01 3.96166384e-01 -5.06298482e-01 -3.23704213e-01
-1.04294956e+00 1.11723065e+00 3.48165125e-01 -1.46474093e-01
-6.55639112e-01 -4.56321925e-01 -9.95142043e-01 5.66286482e-02
4.01143700e-01 -7.49506176e-01 1.10114551e+00 -8.07674170e-01
-4.27728564e-01 1.19065690e+00 -5.87576687e-01 -3.36741418e-01
2.89411284e-02 -3.90139490e-01 -5.31186998e-01 3.39742094e-01
3.94850582e-01 1.34690666e+00 6.64278090e-01 -1.54489756e+00
-7.43902981e-01 -1.32519811e-01 5.33648133e-01 5.07078648e-01
-4.18654203e-01 5.85130155e-01 -8.32908571e-01 -5.20677328e-01
1.88762590e-01 -8.95330667e-01 1.46709278e-01 4.70652610e-01
-4.05174494e-01 -5.05092740e-01 1.21443999e+00 -3.28440636e-01
1.01655209e+00 -2.19346499e+00 -2.65837520e-01 9.00728330e-02
2.67640769e-01 -4.38436195e-02 -2.26929292e-01 5.66737056e-01
-3.08177590e-01 3.05717647e-01 3.90332252e-01 -3.38233650e-01
-1.32731527e-01 7.77438700e-01 -4.40748900e-01 2.32067838e-01
3.35610390e-01 9.92387474e-01 -1.17308331e+00 -7.25181639e-01
2.46502668e-01 5.72153151e-01 -5.28077483e-01 1.77578449e-01
-3.27714771e-01 2.00324491e-01 -3.52829993e-01 6.12856865e-01
5.89676738e-01 -6.48663640e-01 2.01325968e-01 -4.50576782e-01
-5.94851598e-02 5.23227632e-01 -8.87166381e-01 1.04601550e+00
-4.19062108e-01 8.92759383e-01 -4.45952922e-01 -8.93295407e-01
7.84078956e-01 8.18522125e-02 3.53117645e-01 -4.86095726e-01
-1.62961707e-01 -3.03068250e-01 1.24273300e-02 -8.08697224e-01
7.23594487e-01 -1.15546688e-01 5.29874563e-02 4.49997365e-01
-2.14890838e-01 -2.13493198e-01 3.71069372e-01 5.35846829e-01
8.13356042e-01 2.75663823e-01 4.51408803e-01 9.09090936e-02
8.50037932e-02 2.81963587e-01 2.84378588e-01 8.07253540e-01
-2.78955311e-01 5.24680197e-01 6.97024047e-01 -4.79476601e-01
-1.10622740e+00 -1.25795209e+00 7.37103522e-02 1.26559258e+00
9.13258493e-01 -6.96645558e-01 1.61711901e-01 -5.68112075e-01
2.37373620e-01 7.34639823e-01 -5.94825745e-01 5.94768785e-02
-7.10077822e-01 -3.05366933e-01 -5.66825196e-02 8.51895213e-01
1.70866132e-01 -1.06138778e+00 -5.13931870e-01 -1.17775992e-01
-1.45971656e-01 -1.50972497e+00 -4.39318836e-01 -8.92068148e-02
-4.20306116e-01 -1.10388124e+00 -1.21933170e-01 -1.11388731e+00
9.07561302e-01 7.96334386e-01 1.48496318e+00 2.83952832e-01
-3.86924207e-01 6.33341312e-01 -3.61860633e-01 -4.17654455e-01
-1.03005737e-01 -6.10616028e-01 -1.78341269e-02 -1.97914764e-01
6.08222783e-01 -3.76272023e-01 -7.40041733e-01 4.50062245e-01
-5.85881710e-01 2.17519477e-01 4.04483765e-01 7.31143355e-01
6.44927979e-01 2.32206583e-01 1.68648139e-01 -8.47046018e-01
2.14023620e-01 -8.37893069e-01 -3.18509191e-01 4.79201615e-01
-1.44915521e-01 -3.23585927e-01 4.09984410e-01 -8.07490528e-01
-9.72562373e-01 8.94598365e-02 4.38562423e-01 -6.01692677e-01
-3.06541651e-01 2.69886851e-01 3.78602222e-02 3.10697705e-01
3.81738842e-01 -6.59175366e-02 -3.27405304e-01 -2.02785745e-01
6.83524191e-01 2.54535675e-01 4.94378120e-01 -4.72247124e-01
7.61734068e-01 7.43431807e-01 -4.07471955e-02 -7.24927068e-01
-9.11509991e-01 -8.38422477e-01 -4.94777679e-01 -2.56178349e-01
9.51625526e-01 -1.10268879e+00 -7.06299245e-01 -1.01991668e-01
-1.32220328e+00 -1.09541200e-01 1.24001779e-01 5.48148990e-01
-3.71652782e-01 3.77189100e-01 -5.12177289e-01 -6.60678685e-01
3.43527198e-01 -9.17800784e-01 1.13302910e+00 2.23013803e-01
-6.12039089e-01 -9.58956599e-01 -4.90787268e-01 4.68478829e-01
-7.49358460e-02 -1.74007758e-01 1.11899018e+00 -4.89582777e-01
-1.05266476e+00 -1.33719042e-01 -6.45142019e-01 -8.25829804e-02
3.94207805e-01 2.23279044e-01 -4.15652782e-01 -1.53498966e-02
-5.07033885e-01 -3.11880648e-01 6.98153317e-01 1.41240031e-01
1.34371877e+00 -3.52932960e-01 -9.23301518e-01 1.29745081e-01
1.16847301e+00 2.95288354e-01 5.60554922e-01 2.58918703e-01
8.24649572e-01 7.33385921e-01 9.25335646e-01 2.77859181e-01
7.73817003e-01 8.56539369e-01 4.65473354e-01 -5.53537123e-02
-2.59350717e-01 -3.63190114e-01 5.28871566e-02 -2.88786590e-02
1.82390735e-01 -4.67401534e-01 -1.04328454e+00 8.78735542e-01
-1.83366287e+00 -1.04121506e+00 -1.79302990e-01 1.81922734e+00
7.25427926e-01 2.73194145e-02 -2.20605009e-03 -4.57241625e-01
8.31533670e-01 2.72999555e-01 -3.96024913e-01 -2.73841321e-01
2.29462944e-02 -4.39139605e-02 2.47189730e-01 3.71219307e-01
-1.25305009e+00 1.28011334e+00 6.80845451e+00 4.39833254e-01
-7.39675879e-01 -1.74709678e-01 4.99718964e-01 -1.03795141e-01
-4.53912079e-01 4.09155190e-01 -1.10975742e+00 1.81494087e-01
2.76731759e-01 1.13108195e-01 3.38878185e-01 8.15398633e-01
9.00704861e-02 -4.55876708e-01 -1.34014785e+00 1.10978222e+00
2.58692235e-01 -1.23905957e+00 4.70478475e-01 -1.87249482e-01
5.62472999e-01 -3.15177232e-01 1.14688516e-01 1.37014329e-01
4.18506891e-01 -1.18238902e+00 8.32845211e-01 2.88121819e-01
5.07747889e-01 -3.79633278e-01 8.74619633e-02 2.26500720e-01
-1.28928936e+00 -8.88014734e-02 -3.76438797e-01 -2.84898281e-01
3.09970260e-01 2.09824160e-01 -9.84614849e-01 3.57171781e-02
9.02804852e-01 9.64127183e-01 -5.53610861e-01 8.25313210e-01
-4.76839155e-01 1.58543274e-01 -4.25919771e-01 -1.95596740e-01
1.83077365e-01 3.24340258e-03 4.35514212e-01 1.02585614e+00
1.35103375e-01 2.53305018e-01 2.37574682e-01 9.80636299e-01
-1.87222138e-02 -7.17605203e-02 -8.69422257e-01 -1.42246783e-01
7.62894750e-01 1.12452579e+00 -9.28196847e-01 -4.94272202e-01
-8.97760868e-01 7.95445383e-01 5.59959471e-01 4.72421288e-01
-9.76859987e-01 -6.96657151e-02 9.23274338e-01 3.81373644e-01
4.98804927e-01 -5.05768716e-01 -8.30907375e-02 -9.81613040e-01
1.46606177e-01 -4.58657473e-01 5.34735143e-01 -1.35622323e+00
-1.38256109e+00 2.18615636e-01 4.94961351e-01 -1.21681762e+00
-1.11210734e-01 -6.91356540e-01 -5.11460304e-01 7.46345401e-01
-1.41694748e+00 -1.38078070e+00 -3.33762765e-01 5.17572343e-01
5.27011216e-01 2.78200865e-01 6.87184274e-01 2.43532434e-01
-3.03878307e-01 3.39156389e-01 -5.61898589e-01 1.31478474e-01
7.36463368e-01 -1.02426398e+00 4.70285296e-01 6.85204864e-01
5.19519091e-01 1.00540149e+00 7.48780847e-01 -7.34021068e-01
-1.01673353e+00 -8.89822245e-01 1.32420743e+00 -7.04333842e-01
1.06717849e+00 -4.52587157e-01 -9.34955299e-01 1.22442794e+00
1.49256095e-01 2.78714567e-01 5.88713527e-01 3.34388256e-01
-8.78966868e-01 6.28449842e-02 -7.00857759e-01 1.01931238e+00
1.24838829e+00 -7.21830666e-01 -7.06656098e-01 6.75623775e-01
6.07208967e-01 -4.00738806e-01 -4.92682189e-01 1.87746048e-01
5.86721182e-01 -8.86144698e-01 1.51883531e+00 -8.91893923e-01
6.80584669e-01 -3.83335948e-01 -6.20429777e-02 -9.49623168e-01
-2.20903769e-01 -8.89331028e-02 -3.21211219e-01 1.06020617e+00
3.87601048e-01 -3.87943536e-01 6.23852551e-01 9.64974582e-01
7.34437034e-02 -6.98589206e-01 -5.85320354e-01 -9.03956413e-01
-2.75477141e-01 -2.76175708e-01 4.89840478e-01 1.02470517e+00
1.74336970e-01 2.42745623e-01 -4.70644146e-01 2.96318620e-01
3.22929740e-01 3.33874851e-01 7.50666499e-01 -7.32467949e-01
-4.56546426e-01 -3.24783474e-01 -5.89032054e-01 -1.43995357e+00
1.86243787e-01 -6.67665303e-01 -7.40379766e-02 -1.75017715e+00
5.58282971e-01 -7.68903196e-01 -7.42267519e-02 7.85072803e-01
-3.74152482e-01 3.58193129e-01 2.24894211e-01 3.19928795e-01
-7.54513860e-01 8.76849294e-02 1.31317735e+00 -2.00168848e-01
3.29304710e-02 -2.22941473e-01 -6.26163304e-01 7.42502153e-01
6.13309741e-01 -3.35294604e-01 -6.67132378e-01 -6.98965847e-01
4.55327630e-01 -1.73285827e-02 8.47573340e-01 -4.84408408e-01
2.07112774e-01 -3.98284316e-01 4.39450681e-01 -7.25149214e-01
8.47095609e-01 -7.75271535e-01 7.39440694e-02 2.55477950e-02
-4.08843756e-01 9.83814150e-03 1.14923395e-01 8.13750803e-01
-3.21314424e-01 -7.24170208e-02 4.25941885e-01 -1.54310346e-01
-1.23220146e+00 2.44085863e-01 -2.55942255e-01 -1.91865236e-01
1.33422792e+00 -3.99662316e-01 -5.34614921e-01 -6.45379126e-01
-9.99020040e-01 4.90587473e-01 3.09029669e-01 8.48719895e-01
7.22421169e-01 -1.25299668e+00 -2.19755217e-01 -2.77175941e-02
6.06194079e-01 -4.62112390e-03 1.71679273e-01 4.90072787e-01
-3.63629490e-01 2.55484313e-01 -1.13395758e-01 -4.90183771e-01
-1.66280019e+00 8.29184651e-01 1.21644363e-01 8.21739361e-02
-7.53718436e-01 1.13424599e+00 5.82985640e-01 3.49998772e-02
3.13576967e-01 -2.10930869e-01 -3.99290264e-01 1.96110420e-02
4.22004879e-01 -2.09282830e-01 -5.93811393e-01 -9.01528776e-01
-6.37873113e-01 5.73630631e-01 -2.97603101e-01 1.78274065e-01
1.16253805e+00 -2.39473954e-01 -3.70133251e-01 4.24238294e-01
1.37847173e+00 -2.39305332e-01 -1.16748905e+00 -2.11406678e-01
-7.30305463e-02 -9.37305033e-01 -4.46705818e-01 -4.74181116e-01
-6.48061693e-01 8.22998703e-01 1.29691422e-01 1.06626913e-01
8.83800983e-01 8.26054454e-01 5.88026226e-01 3.24426830e-01
3.52956802e-01 -7.59984910e-01 5.24419725e-01 2.66115993e-01
8.73588979e-01 -1.38469112e+00 3.39082897e-01 -1.05868948e+00
-7.34523952e-01 9.61835742e-01 8.70917618e-01 -7.77208209e-02
7.37404943e-01 6.71181679e-02 -1.16963819e-01 -4.96225983e-01
-9.00142193e-01 -2.97908336e-01 4.85813469e-01 5.97749531e-01
4.37440753e-01 2.37079889e-01 -1.82263568e-01 6.18210919e-02
-4.26428020e-02 -3.38268846e-01 3.88765514e-01 8.87310982e-01
-3.65253299e-01 -1.13889658e+00 -2.95648187e-01 4.83543694e-01
-1.54084936e-01 -2.40142331e-01 -3.60042840e-01 9.89390016e-01
3.03975523e-01 9.47186708e-01 5.88942409e-01 -2.01553106e-01
1.41329899e-01 -2.16878980e-01 5.48381567e-01 -9.40405965e-01
-1.46732911e-01 -1.64350480e-01 2.30449408e-01 -5.14790773e-01
-6.54537797e-01 -5.20627260e-01 -1.39158964e+00 -1.65569320e-01
-1.86241359e-01 -3.14868599e-01 2.66476959e-01 8.39964986e-01
1.53924003e-01 1.35883763e-01 1.68328509e-01 -5.33891261e-01
-2.24057119e-03 -4.69208509e-01 -4.81113851e-01 9.96994972e-01
1.77029818e-01 -1.00342488e+00 -8.34456384e-02 3.11380953e-01] | [10.2926664352417, 1.592879056930542] |
6104635e-8262-445d-9d06-7c2f4ff9b438 | algorithmic-trading-in-a-microstructural | 1705.01446 | null | https://arxiv.org/abs/1705.01446v3 | https://arxiv.org/pdf/1705.01446v3.pdf | Algorithmic trading in a microstructural limit order book model | We propose a microstructural modeling framework for studying optimal market making policies in a FIFO (first in first out) limit order book (LOB). In this context, the limit orders, market orders, and cancel orders arrivals in the LOB are modeled as Cox point processes with intensities that only depend on the state of the LOB. These are high-dimensional models which are realistic from a micro-structure point of view and have been recently developed in the literature. In this context, we consider a market maker who stands ready to buy and sell stock on a regular and continuous basis at a publicly quoted price, and identifies the strategies that maximize her P\&L penalized by her inventory. We apply the theory of Markov Decision Processes and dynamic programming method to characterize analytically the solutions to our optimal market making problem. The second part of the paper deals with the numerical aspect of the high-dimensional trading problem. We use a control randomization method combined with quantization method to compute the optimal strategies. Several computational tests are performed on simulated data to illustrate the efficiency of the computed optimal strategy. In particular, we simulated an order book with constant/ symmet-ric/ asymmetrical/ state dependent intensities, and compared the computed optimal strategy with naive strategies. Some codes are available on https://github.com/comeh. | ['Huyên Pham', 'Côme Huré', 'Frédéric Abergel'] | 2017-05-03 | null | null | null | null | ['algorithmic-trading'] | ['time-series'] | [-4.49363589e-01 -2.81315178e-01 -2.41716087e-01 2.38834117e-02
-4.06683236e-01 -8.35421324e-01 5.98949790e-01 1.44392192e-01
-4.72972393e-01 7.58774281e-01 -5.99460416e-02 -3.27951640e-01
-5.72940886e-01 -9.16322887e-01 -6.55476511e-01 -6.94780946e-01
-3.44720602e-01 1.31251812e+00 7.78825358e-02 -3.12503567e-03
7.03391254e-01 6.99662983e-01 -1.16852605e+00 -1.25005931e-01
5.39208531e-01 1.30264354e+00 4.45648059e-02 5.29632628e-01
-3.21405292e-01 9.62513626e-01 -3.67892891e-01 -4.88114029e-01
8.09539914e-01 -3.76141071e-01 -3.46799642e-01 2.02924326e-01
-6.21746063e-01 -5.20260155e-01 -7.95319080e-02 1.02788568e+00
2.00637549e-01 3.29469629e-02 1.07360959e+00 -1.17028964e+00
-3.82170767e-01 7.57671177e-01 -6.64205968e-01 7.04660773e-01
-2.54191160e-01 1.13566421e-01 1.21396303e+00 -4.80527610e-01
4.22431171e-01 1.08790398e+00 -4.54014651e-02 2.00528339e-01
-1.42342472e+00 -4.70317602e-01 -4.48395796e-02 7.33847916e-02
-1.00004423e+00 -1.40197620e-01 6.15007520e-01 -5.69808722e-01
4.39458877e-01 2.06026420e-01 9.00790274e-01 5.83334804e-01
1.00549138e+00 8.05755436e-01 1.53420770e+00 -2.01176882e-01
7.20490098e-01 6.26734421e-02 3.76266122e-01 -6.05813153e-02
4.25074905e-01 3.70909721e-01 -1.81284592e-01 -4.15049016e-01
9.99352038e-01 3.06061536e-01 3.53497207e-01 -3.91594738e-01
-9.19323325e-01 9.63824987e-01 -9.95662957e-02 2.27523088e-01
-8.50231886e-01 6.57027289e-02 -1.88370302e-01 5.51345050e-01
4.31939721e-01 2.66840488e-01 -1.51068792e-01 -9.12274420e-03
-9.67562437e-01 5.73897719e-01 1.36862385e+00 8.84010136e-01
2.44135857e-01 -2.87184536e-01 -4.23884660e-01 1.86540753e-01
3.11896592e-01 7.92221248e-01 -7.85613507e-02 -1.18731225e+00
4.16995615e-01 -1.76979721e-01 6.60254359e-01 -5.77660382e-01
-3.08698326e-01 -5.81933141e-01 -8.02750885e-01 1.09451927e-01
6.30654514e-01 -2.76056468e-01 -2.69816756e-01 1.25518656e+00
-1.02488399e-01 -1.74223427e-02 -2.01632571e-03 6.53135478e-01
-5.87583423e-01 9.76119399e-01 -4.56308931e-01 -8.81800950e-01
1.40582573e+00 -5.05893707e-01 -9.47629333e-01 4.87962484e-01
-4.22323542e-03 -9.58144367e-01 4.28720742e-01 6.64287925e-01
-1.44197047e+00 1.70497835e-01 -4.98999357e-01 6.20247662e-01
9.07364413e-02 -2.26768538e-01 9.66240913e-02 1.59410238e-01
-1.05092514e+00 7.24564195e-01 -9.25370753e-01 2.56950203e-02
9.02110562e-02 2.24047065e-01 6.32402837e-01 4.38498586e-01
-1.05267048e+00 6.42794907e-01 6.14806591e-03 1.51282102e-01
-1.07362545e+00 -4.79107797e-01 2.20573425e-01 3.54651362e-01
9.28951502e-01 -5.47587693e-01 1.70476258e+00 -6.41920507e-01
-1.44180667e+00 2.85791397e-01 2.39340767e-01 -8.14893901e-01
9.53985095e-01 -3.40626240e-02 -8.22962299e-02 1.43781811e-01
1.92715153e-01 -1.38187751e-01 7.43927479e-01 -9.88937974e-01
-7.44168043e-01 -4.08501953e-01 -1.07780069e-01 2.73199588e-01
3.39604884e-01 1.92340225e-01 6.56867996e-02 -8.17608058e-01
-8.07452872e-02 -9.78758872e-01 -3.73340517e-01 -6.99390709e-01
-4.40314710e-01 -1.80869997e-02 -4.28427868e-02 -4.51076716e-01
1.02738595e+00 -1.92983687e+00 -8.39913171e-03 5.34815371e-01
-2.47991625e-02 -5.62370837e-01 4.99715924e-01 1.12682223e+00
3.71976167e-01 3.00881535e-01 3.03366948e-02 -7.88475126e-02
4.39436495e-01 1.89647421e-01 -5.16821563e-01 5.32558441e-01
-4.61305648e-01 5.60944498e-01 -3.43925536e-01 -3.72939825e-01
-1.10916547e-01 -3.04363251e-01 -5.06533444e-01 1.19721420e-01
-3.02995503e-01 2.17143029e-01 -7.03524053e-01 5.45331776e-01
6.82694256e-01 -1.97062492e-01 3.52659196e-01 1.87637210e-01
-6.85942471e-01 -1.12172164e-01 -1.47796857e+00 5.73551953e-01
-2.33295247e-01 2.82893777e-02 4.35302407e-01 -8.05505216e-01
5.40804088e-01 1.87812865e-01 5.36409974e-01 -5.87204099e-01
4.54580069e-01 4.16541994e-01 1.24439299e-01 -6.44043088e-02
4.83440816e-01 -7.24570155e-01 -2.27895543e-01 9.76943851e-01
-3.89462382e-01 8.03428292e-02 5.38438082e-01 1.55550495e-01
8.82606864e-01 -5.15029848e-01 1.51536360e-01 -8.92653525e-01
9.55651179e-02 -5.53206615e-02 5.93225241e-01 9.52607691e-01
-2.70277634e-02 -1.40083674e-02 1.24143016e+00 -9.49726347e-03
-1.23057520e+00 -1.29108655e+00 -1.53828487e-01 5.49719214e-01
3.32475632e-01 3.04540128e-01 -7.26168394e-01 2.43439171e-02
4.17713761e-01 9.62827742e-01 -4.00344998e-01 5.50348818e-01
-3.38362664e-01 -9.31316733e-01 -1.92030713e-01 1.27999216e-01
4.90690023e-01 -1.00656176e+00 -6.83385968e-01 4.67919946e-01
2.76907206e-01 -6.88207865e-01 -5.48659027e-01 1.70633182e-01
-8.81908655e-01 -9.88105118e-01 -9.61961508e-01 -3.20966899e-01
5.11494994e-01 -2.97761321e-01 7.73569763e-01 -4.82358903e-01
1.82910949e-01 5.31595647e-01 5.95333129e-02 -3.47600609e-01
-4.15116996e-01 2.35169102e-02 2.67389208e-01 6.08090997e-01
-5.44087263e-03 -1.75557956e-01 -8.10282767e-01 4.51634973e-01
-1.14737582e+00 -1.83554620e-01 6.69305563e-01 5.65573454e-01
9.32702720e-01 4.45479900e-01 3.41290355e-01 -6.29097342e-01
1.01096284e+00 -4.54317540e-01 -1.44533265e+00 2.91667104e-01
-7.93888688e-01 4.28085685e-01 4.30612683e-01 -1.42196223e-01
-1.10951710e+00 -3.68844390e-01 4.40886110e-01 -3.43072206e-01
1.69195816e-01 2.96414584e-01 -4.91683371e-02 5.07194340e-01
-4.87482846e-01 2.54942358e-01 3.16234171e-01 -9.33456421e-01
1.13240823e-01 4.22203809e-01 1.36967227e-01 -6.03352368e-01
6.26402795e-01 6.11091316e-01 3.05474132e-01 -5.32460034e-01
-2.32710987e-01 -3.09043042e-02 -3.39250445e-01 -1.81262538e-01
7.40542352e-01 -2.99367309e-01 -1.35508597e+00 7.55240619e-01
-8.69260609e-01 -5.19944906e-01 -6.08463824e-01 6.82685316e-01
-1.04118204e+00 -5.07945307e-02 -1.18782139e+00 -1.32505512e+00
1.11265406e-01 -9.56971765e-01 5.11618793e-01 7.50750229e-02
2.45392278e-01 -9.74709332e-01 2.85977989e-01 1.15802653e-01
4.47211951e-01 -4.17854935e-02 1.04230130e+00 -1.04918957e+00
-1.24727881e+00 3.99296992e-02 2.29960591e-01 3.52640778e-01
-9.81506556e-02 -3.89271080e-02 -4.71577756e-02 -2.14226723e-01
6.75598919e-01 5.61274290e-01 4.70003068e-01 7.97698200e-01
5.73935509e-01 -7.05371261e-01 -3.29619855e-01 6.81992397e-02
1.52521801e+00 7.21871912e-01 1.23595320e-01 3.68086517e-01
-1.72915041e-01 8.48426044e-01 8.99537742e-01 9.79637980e-01
2.43513525e-01 6.07957602e-01 2.56230980e-01 5.41177154e-01
8.15836251e-01 -2.79720873e-01 2.16567203e-01 7.35593498e-01
5.30659296e-02 -4.34481502e-01 -8.24094713e-01 3.35783511e-01
-1.73933303e+00 -1.13928771e+00 7.94670135e-02 2.30652595e+00
4.95240390e-01 3.16542983e-01 3.91038656e-01 -1.81575686e-01
9.11783755e-01 -1.23782448e-01 -5.44533432e-01 -3.83450925e-01
-1.32120430e-01 -9.76564642e-03 1.14916754e+00 6.23421311e-01
-5.56379735e-01 4.87097949e-01 6.02934122e+00 9.29064393e-01
-8.25289547e-01 1.47340417e-01 9.57315981e-01 -6.00499153e-01
-3.31601113e-01 1.92965567e-01 -1.05762649e+00 9.82254028e-01
1.08781660e+00 -6.74901187e-01 6.48553252e-01 4.22823638e-01
6.09149933e-01 -3.64197791e-01 -9.99607384e-01 8.28093886e-01
-5.69623828e-01 -1.41202259e+00 -2.67796487e-01 8.84136081e-01
6.17823720e-01 -2.70121872e-01 4.47387815e-01 -1.51825070e-01
4.13217068e-01 -4.94184494e-01 1.00744474e+00 1.12085700e+00
5.85225485e-02 -1.00674534e+00 8.15810382e-01 6.25965774e-01
-9.06672478e-01 -2.92509824e-01 2.40548421e-02 -1.95557818e-01
7.89525986e-01 5.94927490e-01 -3.80716324e-01 4.03185487e-01
4.14640099e-01 1.52224258e-01 4.73768450e-02 9.45356667e-01
1.38925061e-01 5.32967567e-01 -7.02449381e-01 -4.23076302e-01
2.64196008e-01 -8.22174072e-01 5.55381894e-01 3.43845010e-01
5.93371391e-01 2.71290958e-01 -5.67502752e-02 1.10263693e+00
3.00469548e-02 2.06726447e-01 -2.45864555e-01 -2.27743268e-01
3.01896811e-01 8.13958287e-01 -1.36731231e+00 -3.56159091e-01
-1.67355686e-01 7.16201782e-01 -4.77666140e-01 4.01921570e-01
-6.46700025e-01 -1.10214494e-01 3.76447290e-01 5.63837171e-01
4.77166265e-01 -2.94770777e-01 -2.19517559e-01 -1.09759641e+00
-1.09827733e-02 -6.19468272e-01 3.65071774e-01 -1.34868383e-01
-1.51431072e+00 1.43168166e-01 4.17156965e-01 -9.76694345e-01
-4.73792106e-01 -2.94333041e-01 -3.95007700e-01 7.37284482e-01
-1.16395080e+00 -8.90583247e-02 7.81899869e-01 4.25633252e-01
3.87200952e-01 -1.59488708e-01 -1.86558366e-01 7.49404877e-02
-4.73946214e-01 -1.98653921e-01 1.04266417e+00 -1.61006734e-01
-5.58350459e-02 -1.36588001e+00 1.81653932e-01 5.48404098e-01
-1.49507383e-02 4.81624991e-01 8.48527491e-01 -1.02482522e+00
-1.42406440e+00 -4.50773358e-01 7.75412858e-01 -2.56683588e-01
1.12814474e+00 -3.53710651e-01 -6.69592500e-01 5.85314989e-01
4.59019363e-01 -5.17847061e-01 3.41353565e-01 -5.35476208e-01
6.15144372e-01 -3.84059876e-01 -1.08857810e+00 4.06831384e-01
4.91812915e-01 -6.46843091e-02 -4.96263206e-01 3.03586394e-01
1.54302731e-01 1.14722237e-01 -8.28612030e-01 -6.24447577e-02
4.11610365e-01 -1.07054031e+00 4.54495192e-01 -1.24989592e-01
-1.39310837e-01 6.01862371e-02 -1.44526377e-01 -1.05446994e+00
-1.49823755e-01 -1.06994784e+00 1.45066485e-01 1.28131616e+00
3.17291439e-01 -1.14043629e+00 5.46022236e-01 5.19014299e-01
6.24412477e-01 -6.85602069e-01 -1.40324807e+00 -1.17066169e+00
4.33929145e-01 -6.34497106e-02 7.85612822e-01 1.63627908e-01
-2.44832084e-01 4.74538878e-02 -3.08653474e-01 -3.93739864e-02
1.24941468e+00 5.19712031e-01 1.58794373e-01 -1.13697004e+00
-5.32482266e-01 -6.12367988e-01 2.67253965e-01 -1.07874656e+00
-3.90223302e-02 -6.29900753e-01 -7.86269009e-02 -1.12867582e+00
2.57116050e-01 -5.03649235e-01 -3.58230889e-01 -4.29706037e-01
7.87668288e-01 -4.76138741e-01 7.49841213e-01 8.40601087e-01
-5.06779611e-01 5.77947080e-01 1.07214534e+00 1.56587988e-01
-7.91569576e-02 6.14991009e-01 -4.22665060e-01 3.56864244e-01
9.32298899e-01 -5.16264379e-01 -1.43940195e-01 1.11594439e-01
3.64904404e-01 7.23628938e-01 2.26502180e-01 -5.24132609e-01
2.88609803e-01 -4.53825623e-01 -2.57409304e-01 -9.70506966e-01
1.32963359e-01 -9.11911845e-01 6.32343531e-01 8.40039909e-01
-6.05264306e-01 5.18513799e-01 -4.03639138e-01 8.06258261e-01
-1.54429927e-01 -4.58828688e-01 6.00483119e-01 -2.02421919e-01
-4.05447744e-02 3.38004053e-01 -7.47280061e-01 1.16293140e-01
1.37071538e+00 4.27366674e-01 8.12939089e-03 -7.09762871e-01
-9.31086957e-01 4.66748804e-01 5.07772148e-01 -5.25218993e-02
1.84350014e-01 -1.03726268e+00 -4.54345137e-01 -3.44469249e-02
-7.49184310e-01 -5.72935462e-01 1.81918904e-01 1.05978286e+00
-6.81748033e-01 6.53174579e-01 -1.84099153e-01 -3.70009452e-01
-6.72129929e-01 9.30060804e-01 4.39460963e-01 -6.28664076e-01
-1.48941875e-01 1.05274692e-01 2.59371907e-01 2.14636058e-01
-1.38773471e-01 -6.18173182e-01 -3.38062532e-02 5.65354109e-01
3.21313560e-01 8.78517866e-01 -3.00691694e-01 -3.42863858e-01
-9.75271463e-02 5.32760501e-01 -1.02124505e-01 -9.13565218e-01
1.17279875e+00 -3.98976028e-01 -3.06850165e-01 8.10829103e-01
6.77663982e-01 5.30135557e-02 -1.29636836e+00 -9.29829702e-02
3.93805385e-01 -3.75097126e-01 -1.90430209e-01 -3.70935649e-01
-1.06926179e+00 7.06693351e-01 3.43529075e-01 1.08351767e+00
6.95352495e-01 1.74762458e-01 5.35688758e-01 1.39242189e-03
6.09650433e-01 -1.28374267e+00 -1.73883602e-01 2.93089002e-01
7.65159369e-01 -4.85123426e-01 -3.71119469e-01 9.79532371e-04
-7.60782242e-01 8.03834975e-01 -2.25908518e-01 -5.52490115e-01
1.18848562e+00 6.03235006e-01 -2.22235277e-01 -7.20017403e-02
-1.03091276e+00 -1.51313841e-01 -4.11225408e-01 -1.72153845e-01
-1.03046529e-01 4.07214046e-01 -7.76381850e-01 7.38606155e-01
-1.41235128e-01 2.14339942e-01 8.88541937e-01 1.09234142e+00
-4.09588873e-01 -1.12278354e+00 -5.48002779e-01 6.64283216e-01
-9.26484466e-01 -7.44495913e-02 -1.34082690e-01 7.50842214e-01
-5.07809877e-01 6.92803800e-01 5.67609012e-01 2.40135506e-01
3.91584039e-01 6.85334057e-02 2.65784770e-01 -3.37745279e-01
-5.39680243e-01 7.72283435e-01 -3.78819376e-01 -2.39295125e-01
-3.72422002e-02 -1.18176341e+00 -1.04361832e+00 -8.18505287e-01
-3.58317979e-02 6.17194831e-01 4.63860661e-01 7.07046747e-01
2.62154013e-01 2.02451721e-01 8.47017467e-01 -5.46748102e-01
-1.59620333e+00 -6.31458342e-01 -1.63142729e+00 -1.26868993e-01
2.37013310e-01 -7.20990539e-01 -7.04023242e-01 -3.68251950e-01] | [4.849298477172852, 3.956382989883423] |
f8e7d834-1ed3-47cf-bb81-25a649b0c857 | chemical-detection-and-indexing-in-pubmed | null | null | https://biocreative.bioinformatics.udel.edu/media/store/files/2021/TRACK2_pos_03_BC7_submission_136.pdf | https://biocreative.bioinformatics.udel.edu/media/store/files/2021/TRACK2_pos_03_BC7_submission_136.pdf | Chemical detection and indexing in PubMed full text articles using deep learning and rule-based methods | Identifying chemicals in biomedical scientific literature is a crucial task for drug development research. The BioCreative NLM-Chem challenge promoted the development of automatic systems that can identify chemicals in full-text articles and decide which chemical concepts are relevant to be indexed. This work describes the participation of the BIT.UA team from the University of Aveiro, where we propose a three-stage automatic pipeline that individually tackles (i) chemical mention detection, (ii) entity normalization and (iii) indexing. We adopted a deep learning solution based on a biomedical BERT variant for chemical identification. For normalization we used a rule-based approach and a hybrid version that explores a dense retrieval mechanism. Similarly, for indexing we also followed two distinct approaches: a rule-based, and a TF-IDF based method. Our best official results are consistently above the official median and benchmark in the three subtasks, with respectively 0.8454, 0.8136, and 0.4664 F1-scores. | ['Sérgio Matos', 'João Rafael Almeida', 'João Figueira Silva', 'Rui Antunes', 'Tiago Almeida'] | 2021-11-08 | null | null | null | biocreative-vii-challenge-evaluation-workshop | ['chemical-indexing'] | ['natural-language-processing'] | [ 2.87756294e-01 2.03003377e-01 -1.90168217e-01 -6.22632615e-02
-9.40467358e-01 -8.60179722e-01 1.00218081e+00 1.12622988e+00
-7.78217673e-01 1.07700217e+00 2.86921620e-01 -3.51853251e-01
-4.21455503e-01 -7.60455668e-01 -7.88142145e-01 -8.88639867e-01
2.25402117e-01 7.58791983e-01 -1.09161705e-01 1.06928855e-01
3.38642359e-01 8.94146383e-01 -1.32435501e+00 4.28563178e-01
7.50895679e-01 8.26668680e-01 -5.72276711e-02 4.45042551e-01
-2.42971376e-01 6.97352171e-01 -5.10024965e-01 -3.45550776e-01
7.69007951e-02 -1.75358921e-01 -9.82834816e-01 -5.79437852e-01
1.78028837e-01 2.64754474e-01 -1.21343844e-02 9.29311037e-01
7.49666095e-01 4.39995676e-02 9.07937706e-01 -6.76832914e-01
-2.67199099e-01 6.86136067e-01 -3.45392168e-01 -1.59869883e-02
5.92399538e-01 -8.25939775e-02 8.69193196e-01 -8.81676853e-01
1.02959740e+00 9.88234878e-01 6.93623602e-01 4.63077277e-01
-1.20000041e+00 -7.45348930e-01 -1.98805183e-01 2.11612329e-01
-1.50689197e+00 -6.01284742e-01 1.98353864e-02 -7.62330234e-01
1.26694167e+00 3.29471469e-01 2.03277498e-01 9.68467057e-01
3.75462711e-01 2.27819115e-01 1.03871500e+00 -5.28346419e-01
6.12237215e-01 2.24026069e-01 1.62027314e-01 4.42073166e-01
6.77994967e-01 -2.23425671e-01 -4.91070509e-01 -3.05518210e-01
4.25467081e-02 -4.63969596e-02 -1.15077056e-01 1.82276696e-01
-1.15274823e+00 7.89453268e-01 1.54688075e-01 9.80030656e-01
-9.00071263e-01 -2.32380971e-01 5.27808011e-01 2.74546072e-02
5.23423314e-01 1.11355090e+00 -8.21884215e-01 3.65540087e-01
-1.23896039e+00 5.22799373e-01 1.12875664e+00 7.05083549e-01
3.92608047e-01 -7.17132866e-01 -6.73565030e-01 6.22688293e-01
3.56671095e-01 1.53057963e-01 5.10161638e-01 -5.45373619e-01
1.49696797e-01 6.32293880e-01 4.26925570e-02 -8.06708395e-01
-7.10790336e-01 -4.65462416e-01 -5.84112823e-01 -3.60350341e-01
3.20942611e-01 -1.87885925e-01 -1.13458180e+00 1.41688895e+00
3.68021458e-01 8.85232016e-02 1.23477772e-01 4.84486580e-01
1.23225725e+00 4.07242656e-01 8.76031578e-01 -2.55407989e-01
1.67504489e+00 -5.79798043e-01 -9.89529669e-01 4.28650588e-01
6.90747261e-01 -9.79184449e-01 2.29931742e-01 5.77251375e-01
-9.92348850e-01 -1.63064510e-01 -1.16410410e+00 -2.40361065e-01
-1.29243481e+00 2.52255678e-01 4.81097370e-01 5.64735711e-01
-1.05441916e+00 8.49524319e-01 -5.22747993e-01 -4.61815357e-01
4.60417241e-01 6.85361505e-01 -6.85536623e-01 -1.11971959e-01
-1.33252513e+00 1.11398995e+00 6.23290241e-01 -1.66091278e-01
-8.04423690e-01 -1.01562381e+00 -4.59293783e-01 1.16713062e-01
1.77818999e-01 -7.17774630e-01 9.19430435e-01 -1.82656869e-01
-1.13573718e+00 1.14094079e+00 -5.82026578e-02 -6.11580253e-01
4.55007315e-01 -1.54480696e-01 -3.97462785e-01 2.20902786e-01
5.17097533e-01 6.09582067e-01 1.24620907e-01 -8.86512101e-01
-6.71302855e-01 -5.64659953e-01 -1.44263610e-01 1.23938704e-02
-4.23454881e-01 1.45983815e-01 -4.38240856e-01 -4.82668996e-01
-3.46238554e-01 -8.32580507e-01 -3.73241544e-01 -4.91245925e-01
-6.56713724e-01 -4.42076325e-01 6.74841031e-02 -8.21205199e-01
1.28525043e+00 -1.61237800e+00 1.81663364e-01 3.47354710e-01
4.25311863e-01 3.75868231e-01 -1.75023917e-02 6.20379448e-01
-3.97892326e-01 4.45696235e-01 -4.62318882e-02 -9.66324955e-02
1.85806490e-02 -2.78676629e-01 -4.45109159e-02 4.19048965e-01
2.94326752e-01 7.10429072e-01 -1.06693769e+00 -5.18610001e-01
-9.09293070e-02 6.33804560e-01 -2.33373627e-01 -1.18536558e-02
-2.89354116e-01 1.90177202e-01 -5.86561143e-01 7.66149819e-01
5.19725144e-01 -1.28896758e-01 4.38487589e-01 -5.51899850e-01
-5.64930379e-01 4.10495043e-01 -9.36197877e-01 1.61104405e+00
-1.07041478e-01 1.36901051e-01 5.60454233e-03 -8.68909955e-01
7.13903606e-01 6.98335826e-01 9.67657268e-01 -5.87833226e-01
1.89711258e-01 5.22392213e-01 1.73438732e-02 -4.12295043e-01
3.50499600e-01 -1.30201846e-01 2.26953372e-01 -5.69875464e-02
5.45646846e-01 3.52542698e-01 7.33080268e-01 1.09454520e-01
1.50337958e+00 3.42716902e-01 6.27563179e-01 -7.27219582e-01
9.79897738e-01 1.97969481e-01 4.28543627e-01 5.41601956e-01
1.18507132e-01 2.11987823e-01 4.94515508e-01 -3.58717501e-01
-7.49110162e-01 -3.84300828e-01 -5.06089866e-01 9.73649204e-01
-4.06523436e-01 -5.07769108e-01 -1.02207458e+00 -6.02413237e-01
4.38767597e-02 5.92999935e-01 -9.00494337e-01 -1.72838792e-01
-3.09071958e-01 -1.10846472e+00 8.55388403e-01 8.65267683e-03
4.56655510e-02 -1.03486669e+00 -1.23208374e-01 4.88712043e-01
-1.49747264e-02 -9.88316774e-01 -8.68212432e-02 7.73267090e-01
-4.62704748e-01 -1.18123698e+00 -8.82082641e-01 -4.23649967e-01
2.91138530e-01 -3.82344246e-01 1.01281500e+00 -2.72403628e-01
-5.05772293e-01 -6.87089860e-02 -2.45394126e-01 -9.73158956e-01
-4.83152181e-01 4.86825854e-01 -8.13823938e-02 -1.95230782e-01
7.69499242e-01 -1.10341750e-01 -7.09483385e-01 -2.35825971e-01
-9.90603924e-01 -4.73858207e-01 1.03364575e+00 4.11153346e-01
8.64952624e-01 -2.21975341e-01 7.25631773e-01 -1.15275753e+00
6.70309365e-01 -7.27439046e-01 -5.56737781e-01 4.67610359e-01
-1.03182280e+00 2.65156776e-01 4.19952095e-01 -2.27592945e-01
-7.45282292e-01 6.43998623e-01 -5.64827561e-01 3.20393592e-02
-5.20771384e-01 6.61597371e-01 -2.77204454e-01 -6.86399043e-02
7.53232002e-01 -1.67998001e-01 -3.36147398e-01 -6.39645815e-01
3.57765287e-01 6.84183836e-01 2.40694836e-01 -5.30269802e-01
4.32501525e-01 1.08852483e-01 3.37691724e-01 -6.60690308e-01
-7.17246771e-01 -9.42484021e-01 -5.61663449e-01 1.79174989e-01
1.31451416e+00 -9.81057525e-01 -8.36379528e-01 1.37315020e-01
-1.29356575e+00 1.14847183e-01 -1.67520031e-01 4.20083880e-01
-7.90885910e-02 2.14908123e-01 -5.81630886e-01 -2.95047641e-01
-8.61735523e-01 -1.03406489e+00 1.20165372e+00 8.45364034e-02
-4.99048859e-01 -8.55831087e-01 5.94284058e-01 3.09941530e-01
2.19356313e-01 5.75792611e-01 1.06487632e+00 -1.50245082e+00
1.97661556e-02 -2.79406726e-01 -1.99073181e-01 -7.32135773e-02
2.06675649e-01 -8.22442397e-02 -1.20407927e+00 -3.22478637e-02
-2.33113274e-01 9.56334025e-02 9.63395596e-01 3.25872391e-01
8.65071714e-01 -2.70035625e-01 -7.73045659e-01 2.77603835e-01
1.45853901e+00 5.03694952e-01 6.74083292e-01 6.46243155e-01
6.42880857e-01 5.91360390e-01 4.42112297e-01 2.34933838e-01
8.93465281e-02 5.27088761e-01 1.01134442e-01 -2.53404260e-01
-1.85427368e-02 3.39676030e-02 -2.55173743e-02 2.69264698e-01
-1.82101995e-01 -2.17038304e-01 -1.12685502e+00 6.04014397e-01
-1.64815617e+00 -7.59465754e-01 -3.28580350e-01 2.01609230e+00
1.30583596e+00 -8.18923563e-02 -1.38420695e-02 1.09228052e-01
5.26225269e-01 -5.41134119e-01 -4.04018819e-01 -3.58392030e-01
-9.90310535e-02 8.34318876e-01 8.30641806e-01 1.65200770e-01
-1.47242105e+00 8.65733624e-01 5.76830101e+00 9.46887016e-01
-9.65866864e-01 1.16326027e-02 6.43782079e-01 2.06690416e-01
7.76480511e-02 -4.28408474e-01 -1.04759169e+00 3.31443489e-01
1.45875657e+00 -2.09647357e-01 3.17454413e-02 4.54787076e-01
1.06009245e-01 4.55626026e-02 -1.25237536e+00 5.47313750e-01
-1.54379860e-01 -1.73762226e+00 3.31800401e-01 3.83306473e-01
6.10627651e-01 2.59829104e-01 -2.69232810e-01 1.49475217e-01
1.43405214e-01 -1.19716787e+00 5.26757896e-01 7.20315635e-01
7.14221299e-01 -5.48860669e-01 8.78690302e-01 5.76717546e-03
-9.64046597e-01 1.51156723e-01 -1.24068424e-01 6.40540779e-01
-4.09196228e-01 8.85956109e-01 -8.43541801e-01 1.04840958e+00
6.67947173e-01 7.42929518e-01 -5.19863188e-01 1.15590286e+00
-4.93717864e-02 3.07162762e-01 -8.77046678e-03 4.03187657e-03
2.44573593e-01 1.35273606e-01 3.28257471e-01 1.89013898e+00
1.59630731e-01 -3.17125092e-03 -1.65282972e-02 7.53810525e-01
-3.78332525e-01 6.48189306e-01 -4.17605937e-01 -3.83420736e-01
5.07421672e-01 1.56175840e+00 -9.81931865e-01 -4.15679038e-01
-8.87539834e-02 6.59312487e-01 -8.79805833e-02 1.17434286e-01
-5.49245834e-01 -6.81611896e-01 4.79821980e-01 -4.96026613e-02
2.65481651e-01 4.21352625e-01 -1.34544909e-01 -7.87884533e-01
-5.40634155e-01 -9.99907076e-01 5.99663019e-01 -2.71682054e-01
-1.22825503e+00 6.44661725e-01 -1.09376937e-01 -8.51131499e-01
-1.67504683e-01 -8.19192290e-01 -3.14950906e-02 1.03420639e+00
-1.45577645e+00 -9.60666597e-01 1.36435568e-01 3.13317001e-01
6.07918166e-02 -1.53696224e-01 1.27942705e+00 8.36421490e-01
-6.21623039e-01 3.11721832e-01 2.53349036e-01 -3.74823436e-02
9.37045097e-01 -1.45952702e+00 6.84492290e-02 3.25273424e-01
-6.42055348e-02 1.06526649e+00 7.49835312e-01 -6.60731614e-01
-1.31512451e+00 -1.39651072e+00 1.47656238e+00 -3.75829995e-01
6.37331963e-01 -1.90001115e-01 -8.43418598e-01 2.30716854e-01
3.31408143e-01 -4.89548713e-01 1.10825193e+00 -6.78864643e-02
-2.50045329e-01 5.10986596e-02 -1.34920382e+00 1.95384651e-01
5.74051261e-01 -3.47760499e-01 -4.56241578e-01 7.77357697e-01
5.54192603e-01 -2.20924780e-01 -1.45934939e+00 2.97720999e-01
5.89547515e-01 -2.15114832e-01 1.01493287e+00 -7.40365148e-01
2.76109129e-01 -4.37059015e-01 1.07208878e-01 -9.27811265e-01
-5.35399854e-01 -4.48299944e-01 4.00188416e-02 1.32659400e+00
7.31352806e-01 -4.46458519e-01 3.20944190e-01 4.50549215e-01
-5.08355685e-02 -6.94461942e-01 -8.02809060e-01 -5.02547681e-01
3.53475660e-01 2.06138834e-01 5.85818887e-01 1.20087624e+00
-2.84236167e-02 7.25502968e-01 2.67313451e-01 -1.13446787e-01
3.14611882e-01 -2.51709402e-01 1.38944715e-01 -1.62272406e+00
8.71293023e-02 -7.62745321e-01 -2.76494354e-01 -1.99938826e-02
1.50175542e-01 -1.16927600e+00 2.26152912e-02 -1.81043935e+00
4.38774407e-01 5.93800582e-02 -7.80484021e-01 7.92674303e-01
-4.31404971e-02 1.29627764e-01 -3.26396376e-01 1.76635578e-01
-5.76933086e-01 -2.04031676e-01 4.43139404e-01 -2.84095109e-01
-9.27814841e-02 -4.53154057e-01 -1.02480507e+00 3.60507816e-01
7.59401023e-01 -7.96711683e-01 1.17063135e-01 3.58311944e-02
4.40980434e-01 -4.07585144e-01 -1.02472551e-01 -9.17931557e-01
3.56507927e-01 1.06180813e-02 5.38058400e-01 -4.77324754e-01
-1.54254749e-01 -6.25455976e-01 4.83902812e-01 8.31158876e-01
-6.00078225e-01 -1.80295736e-01 5.07999003e-01 4.21694338e-01
-5.87966070e-02 -3.21749181e-01 6.25919938e-01 -2.94680715e-01
-2.23907918e-01 2.16294840e-01 -5.41334033e-01 -3.72711360e-01
8.64121258e-01 1.82361975e-01 -3.95411164e-01 3.11863393e-01
-8.55916858e-01 8.64275768e-02 2.13568043e-02 1.53978914e-01
1.03239797e-01 -9.64696527e-01 -8.75267506e-01 -4.35993850e-01
2.79432803e-01 -3.93969327e-01 -2.25743473e-01 9.11179543e-01
-5.77254832e-01 9.94831324e-01 -3.10653523e-02 -2.12189913e-01
-1.27081680e+00 6.19424641e-01 3.90594095e-01 -8.07296276e-01
-6.91341236e-02 6.45833552e-01 3.70977260e-02 -3.37658226e-01
3.11202854e-01 -2.58990347e-01 -8.25347126e-01 6.68537021e-01
6.51141703e-01 4.17324156e-01 9.37022924e-01 -6.76059246e-01
-6.96381390e-01 4.68362629e-01 -2.66530067e-01 9.89238918e-02
1.58991539e+00 5.18824875e-01 -5.55496812e-01 1.48093477e-01
1.25368178e+00 -7.39534721e-02 -3.10303450e-01 8.55912864e-02
7.60977924e-01 4.73461181e-01 3.35666686e-01 -1.46557403e+00
-7.64062941e-01 5.27350843e-01 6.45844400e-01 1.32576481e-01
9.29149985e-01 -1.40203699e-01 3.85715514e-01 6.50383294e-01
-1.41277969e-01 -1.07775378e+00 -6.81896865e-01 4.72374350e-01
9.03564632e-01 -7.98349679e-01 4.72712755e-01 -3.24722975e-01
-2.34107319e-02 1.11573160e+00 -4.22256105e-02 2.69120365e-01
6.14758193e-01 2.88406670e-01 -1.04857482e-01 -5.90261400e-01
-7.63247132e-01 -3.47710520e-01 6.98525310e-01 2.13927954e-01
1.08581066e+00 -2.75729708e-02 -1.05369604e+00 7.32719660e-01
2.94594288e-01 1.99743718e-01 8.84553492e-02 8.55898082e-01
-1.33206353e-01 -1.47923291e+00 -2.52209544e-01 4.63118136e-01
-1.26360273e+00 -4.65553582e-01 -1.03318322e+00 7.30082870e-01
4.21298444e-01 1.03096414e+00 -5.22622943e-01 -2.68658072e-01
6.27925754e-01 3.99348378e-01 3.98240089e-01 -7.26830065e-01
-1.29976583e+00 2.82183737e-01 1.85757667e-01 -5.70464194e-01
-6.71826959e-01 -7.38017797e-01 -1.35978687e+00 8.84919912e-02
-1.90596849e-01 5.65842807e-01 1.01508093e+00 8.34514201e-01
6.40542388e-01 8.07493210e-01 1.57106161e-01 -5.50083876e-01
-1.27289578e-01 -1.04540384e+00 -2.71511316e-01 1.53203353e-01
1.11175440e-01 -4.29205179e-01 -5.06616943e-02 2.20289171e-01] | [8.509193420410156, 8.733352661132812] |
098072cd-02df-4afe-97c5-b6adb7751526 | identifying-trades-using-technical-analysis | 2304.09936 | null | https://arxiv.org/abs/2304.09936v1 | https://arxiv.org/pdf/2304.09936v1.pdf | Identifying Trades Using Technical Analysis and ML/DL Models | The importance of predicting stock market prices cannot be overstated. It is a pivotal task for investors and financial institutions as it enables them to make informed investment decisions, manage risks, and ensure the stability of the financial system. Accurate stock market predictions can help investors maximize their returns and minimize their losses, while financial institutions can use this information to develop effective risk management policies. However, stock market prediction is a challenging task due to the complex nature of the stock market and the multitude of factors that can affect stock prices. As a result, advanced technologies such as deep learning are being increasingly utilized to analyze vast amounts of data and provide valuable insights into the behavior of the stock market. While deep learning has shown promise in accurately predicting stock prices, there is still much research to be done in this area. | ['Prof. Pramila M. Chawan', 'Nirmit Deliwala', 'Meet Parekh', 'Mann Doshi', 'Aayush Shah'] | 2023-04-12 | null | null | null | null | ['stock-market-prediction'] | ['time-series'] | [-7.60105491e-01 -5.27899027e-01 -6.00241601e-01 -5.11510558e-02
-1.53273612e-01 -5.85039556e-01 2.97234356e-01 3.17455590e-01
-3.55242133e-01 5.62963605e-01 2.64960587e-01 -7.37041891e-01
1.25068560e-01 -1.07105744e+00 -4.51239124e-02 -4.87188339e-01
-4.04226296e-02 1.90858662e-01 7.16790408e-02 -2.34697998e-01
6.71135962e-01 8.02677333e-01 -9.39341486e-01 -3.64127904e-01
5.63237607e-01 1.46143997e+00 -1.12325348e-01 1.04426794e-01
-5.06998122e-01 1.15235472e+00 -2.25951418e-01 -5.53812623e-01
6.33705258e-01 -1.81610256e-01 -1.04866773e-01 -2.13439912e-01
-4.41781759e-01 -7.68632650e-01 -1.49205908e-01 1.14037812e+00
1.29050184e-02 -7.01596364e-02 3.81969392e-01 -7.97165573e-01
-4.62820292e-01 4.85140920e-01 -5.92949092e-01 7.13269114e-01
-2.84591585e-01 9.68752056e-02 1.49281490e+00 -5.07923007e-01
-7.20253512e-02 6.08105004e-01 1.53267428e-01 1.04872659e-02
-9.38810945e-01 -1.03779519e+00 2.45324984e-01 1.27123594e-01
-7.59814501e-01 -9.95843336e-02 7.82166839e-01 -7.13580310e-01
7.38162756e-01 -6.06326154e-03 9.28192496e-01 3.54695708e-01
6.94267333e-01 5.37379503e-01 8.54259133e-01 -3.13672945e-02
2.24907592e-01 1.44875899e-01 -2.76329797e-02 -5.43836178e-03
5.87270021e-01 2.77998894e-01 -3.02949607e-01 1.71207618e-02
8.28665614e-01 4.01682585e-01 -2.12502591e-02 1.14051802e-02
-7.99098969e-01 1.05671263e+00 4.60345596e-01 4.69911605e-01
-7.24647462e-01 7.43355229e-02 1.46864146e-01 5.76094925e-01
4.03529197e-01 6.46529794e-01 -3.67458403e-01 -4.18675959e-01
-8.63490403e-01 7.03543186e-01 7.98426092e-01 -1.17308609e-01
2.83029556e-01 5.48228025e-01 4.56919760e-01 3.42485994e-01
2.52771854e-01 5.65768898e-01 5.60255647e-01 -8.24968815e-01
5.99470556e-01 7.75903046e-01 4.92749035e-01 -1.20511532e+00
-2.18349949e-01 -7.03658700e-01 -6.06626749e-01 7.40170717e-01
4.80271667e-01 -1.47682846e-01 -3.37060899e-01 1.25270617e+00
-1.60471588e-01 -9.90686268e-02 -6.99119791e-02 7.64937699e-01
-2.62706012e-01 9.15683925e-01 -1.43238544e-01 -1.01830989e-01
9.36638117e-01 -2.50943631e-01 -5.13562858e-01 -4.05589432e-01
2.68729180e-01 -6.77954793e-01 4.09274876e-01 2.89306045e-01
-9.65145528e-01 -2.10785624e-02 -1.06836748e+00 5.20655930e-01
-2.56408453e-01 -4.98769760e-01 7.09507406e-01 2.79144436e-01
-4.73975271e-01 7.99494386e-01 -9.51074243e-01 7.37996817e-01
4.82108116e-01 3.28141868e-01 1.45198509e-01 3.05406809e-01
-1.34584546e+00 1.24843109e+00 4.78598654e-01 2.17380121e-01
-3.28635097e-01 -6.75329447e-01 -5.01212239e-01 5.07059395e-01
4.81401145e-01 -6.10602833e-02 1.46431088e+00 -6.95962131e-01
-9.58992362e-01 7.57551715e-02 3.66974205e-01 -1.03382337e+00
5.36545396e-01 -1.68513730e-01 -2.19967857e-01 -2.06432864e-01
-1.07357927e-01 -2.11995140e-01 4.95128363e-01 -4.87899363e-01
-1.03868961e+00 -5.54141283e-01 -8.97462964e-02 7.64738172e-02
-4.27210271e-01 2.74960458e-01 4.74366784e-01 -1.01138949e+00
2.84638584e-01 -7.34468341e-01 -3.31088185e-01 -3.26829344e-01
-7.33660087e-02 8.39709491e-02 4.80295599e-01 -9.51616347e-01
1.25641382e+00 -1.70243394e+00 -4.08165067e-01 3.49117786e-01
2.00186834e-01 3.40117842e-01 5.02212346e-01 5.69611609e-01
-1.45062909e-01 3.68745029e-01 -8.74606967e-02 1.48059845e-01
4.69576009e-02 -1.40594974e-01 -7.57814467e-01 2.34596774e-01
3.86856496e-01 9.12387073e-01 -3.87921959e-01 3.09608191e-01
3.95024359e-01 2.86760721e-02 -1.19878665e-01 2.09682181e-01
-3.69303912e-01 1.22101465e-02 -8.30327213e-01 6.65765047e-01
3.33222002e-01 -3.05202097e-01 -4.68867980e-02 4.70708251e-01
-4.98152941e-01 5.63122034e-01 -1.14653873e+00 2.60552347e-01
-2.74172664e-01 7.27612972e-01 -6.48150817e-02 -1.21990561e+00
9.77478206e-01 2.93060690e-01 5.69056511e-01 -1.10893869e+00
2.89660573e-01 4.69659418e-01 3.15800667e-01 -4.20837291e-02
4.80038792e-01 -8.54971945e-01 6.32925779e-02 8.99275720e-01
-8.38050604e-01 1.90789476e-01 1.67624414e-01 -1.64126381e-01
8.21445644e-01 -4.36508417e-01 4.72176582e-01 3.37603651e-02
7.93704093e-02 -1.20281786e-01 8.49000752e-01 2.88844872e-02
-5.10439314e-02 2.22401116e-02 8.66670012e-01 -8.25237215e-01
-9.53510463e-01 -7.50924468e-01 8.03700183e-03 3.71020526e-01
-4.11750853e-01 3.97404462e-01 -1.30909503e-01 -2.67025769e-01
7.45065928e-01 7.03173041e-01 -4.57818866e-01 7.46382549e-02
-3.34913373e-01 -8.88711214e-01 -2.51880676e-01 8.21561933e-01
5.51422119e-01 -9.51304376e-01 -9.39396441e-01 5.72626770e-01
3.34238350e-01 -8.78397942e-01 -8.22753534e-02 3.35199609e-02
-8.64142001e-01 -1.25485945e+00 -7.84509897e-01 -4.17781137e-02
2.48667836e-01 1.82252377e-01 9.73444879e-01 -5.01905009e-02
8.83919224e-02 -1.97710991e-01 -1.96206681e-02 -9.22876954e-01
-3.77894670e-01 5.55001386e-02 -9.14412588e-02 3.16002071e-02
3.27924967e-01 -5.49644053e-01 -6.06773973e-01 9.34672728e-02
-7.50729740e-01 -1.76567242e-01 4.27769691e-01 6.42453015e-01
2.29340807e-01 7.91393280e-01 1.10390651e+00 -5.42522788e-01
7.50295997e-01 -5.98006725e-01 -1.59897840e+00 6.61729127e-02
-1.02440119e+00 8.45576227e-02 4.85386997e-01 5.49417324e-02
-1.15279675e+00 -4.30341452e-01 -6.49511218e-02 -7.96736479e-02
3.90555799e-01 1.17310119e+00 1.00311175e-01 1.63124472e-01
-2.74790794e-01 6.45552948e-02 1.17915183e-01 -6.34277701e-01
-3.20252806e-01 2.02047125e-01 2.88382053e-01 -2.98517179e-02
9.95853424e-01 2.13044301e-01 1.48817196e-01 -4.55220431e-01
-7.70708025e-01 -2.25101352e-01 -3.86013031e-01 -1.16521209e-01
6.68537676e-01 -9.05047417e-01 -5.31998277e-01 6.35890722e-01
-6.62822545e-01 -3.34772617e-02 -1.09646274e-02 5.49473047e-01
7.82172680e-02 -8.65180790e-02 -7.17959404e-01 -1.18974936e+00
-1.93406269e-01 -1.00737631e+00 -1.22161202e-01 4.52445358e-01
-3.02911311e-01 -1.25544429e+00 1.55382715e-02 4.66151267e-01
6.34151459e-01 4.47345316e-01 9.70241785e-01 -8.05580378e-01
-1.15871131e+00 -9.90884364e-01 -1.57443479e-01 5.89646101e-01
3.86834264e-01 1.26894370e-01 -4.02285844e-01 3.22100408e-02
3.97523671e-01 -6.12032637e-02 8.92541945e-01 5.42779684e-01
5.51727295e-01 -3.78555566e-01 2.33291745e-01 1.93234652e-01
1.24583924e+00 7.84418166e-01 4.17649835e-01 8.66932392e-01
3.05666566e-01 6.91467643e-01 5.94761431e-01 7.81397641e-01
5.08347571e-01 2.21976429e-01 5.84267259e-01 3.19874883e-01
8.31749678e-01 -1.85357153e-01 1.43157199e-01 5.86006403e-01
3.31755728e-02 2.00469598e-01 -1.17550623e+00 2.89827108e-01
-1.64911842e+00 -1.39064717e+00 2.98003167e-01 2.17565513e+00
4.90669876e-01 7.70860255e-01 3.55060786e-01 5.25576174e-01
3.88199508e-01 3.07217956e-01 -8.30385208e-01 -2.19331637e-01
-2.29977891e-01 -9.15299505e-02 6.60394490e-01 1.50216833e-01
-8.61435235e-01 5.99074125e-01 5.85456705e+00 1.82794452e-01
-1.47412586e+00 -7.41230488e-01 1.17118430e+00 2.94785481e-03
-5.16899705e-01 -1.08300559e-01 -8.88527930e-01 6.80201411e-01
7.02680767e-01 -7.88743317e-01 2.77996421e-01 9.19552326e-01
5.46916604e-01 -3.41338128e-01 -5.36680281e-01 4.90245461e-01
-6.44890010e-01 -1.67585421e+00 -2.83277571e-01 5.57692349e-01
5.58789551e-01 -1.48421571e-01 4.50394779e-01 1.70294531e-02
3.90758574e-01 -1.05368853e+00 6.98751807e-01 5.89598835e-01
-1.70658588e-01 -1.19557583e+00 9.29016292e-01 5.22817791e-01
-1.24073315e+00 -5.72625756e-01 -3.98884922e-01 -7.81630993e-01
1.97567433e-01 7.99089134e-01 -4.37102973e-01 1.91032924e-02
4.85679120e-01 6.14260912e-01 -1.35191873e-01 1.04414392e+00
-1.82959750e-01 6.15282476e-01 -1.90281779e-01 -8.43854845e-02
4.07287687e-01 -6.61477149e-01 6.03141263e-02 1.69059768e-01
4.33253825e-01 2.40183949e-01 -1.11224644e-01 8.69993806e-01
-2.06252962e-01 6.17158413e-02 -4.39359665e-01 -7.90815175e-01
4.21231985e-01 8.06084514e-01 -8.12866390e-01 -6.78688381e-03
-7.46726096e-01 1.19610049e-01 3.60520445e-02 5.67843504e-02
-3.15524876e-01 -1.78633798e-02 1.10676777e+00 3.82723689e-01
2.04789326e-01 -5.96831262e-01 -6.37184381e-01 -1.09744728e+00
2.36445025e-01 -8.02902102e-01 4.29896861e-01 -1.14085510e-01
-1.00725889e+00 -8.87074694e-02 -3.20288062e-01 -1.00550437e+00
-6.52748704e-01 -6.12559080e-01 -9.70651209e-01 1.00832903e+00
-2.00317717e+00 -2.43866444e-01 3.88321728e-01 -8.92829448e-02
2.75820285e-01 -4.81333256e-01 3.33793133e-01 -2.56609052e-01
-8.05527210e-01 -1.39348000e-01 6.23746097e-01 5.56404471e-01
-3.79121751e-02 -1.14070702e+00 7.04088390e-01 9.43774760e-01
2.77659029e-01 3.14372808e-01 2.33171597e-01 -8.64458919e-01
-1.26530623e+00 -6.37279034e-01 7.87889063e-01 -1.12126566e-01
1.24164581e+00 1.86841428e-01 -1.13441765e+00 5.74760675e-01
-8.68441537e-02 -3.06081027e-01 8.23675811e-01 -9.45406109e-02
-2.17651457e-01 -4.62071389e-01 -6.49089813e-01 5.56955159e-01
-2.66696751e-01 -2.89553612e-01 -4.91960227e-01 -3.00955266e-01
3.28385115e-01 3.91809940e-02 -8.32820714e-01 8.78682658e-02
6.88434124e-01 -1.16405487e+00 7.85928369e-01 -3.78261238e-01
3.53104621e-01 1.97475925e-01 6.83744177e-02 -1.31329298e+00
-2.60916501e-01 -3.86508226e-01 -7.13163316e-02 1.11179435e+00
5.25268137e-01 -1.01713896e+00 1.06550479e+00 1.53941035e+00
4.33611572e-01 -9.23547029e-01 -8.77627730e-01 -6.77393198e-01
4.09250021e-01 -6.25645697e-01 1.06591856e+00 8.18586051e-01
-2.47382969e-02 -2.08167434e-01 -4.09252286e-01 -5.66679910e-02
7.57121623e-01 6.21372342e-01 4.22162741e-01 -1.35500300e+00
-6.12271130e-02 -1.04836702e+00 -1.82611570e-01 -4.38344061e-01
2.72503555e-01 -3.92789871e-01 -7.04871237e-01 -1.37503529e+00
-1.48353502e-01 -3.92275512e-01 -8.47378373e-01 2.25403905e-01
-1.37139663e-01 -1.56025380e-01 5.12770832e-01 4.45109874e-01
2.25367337e-01 5.30888259e-01 9.68504727e-01 2.24622339e-02
-9.79682580e-02 6.95995152e-01 -9.04262364e-01 8.65916193e-01
1.38661051e+00 -1.40163213e-01 -8.79344419e-02 -1.66764021e-01
7.09761918e-01 3.78630459e-01 1.63541704e-01 -6.39248788e-01
2.02977727e-03 -8.15690875e-01 5.81393600e-01 -8.12475145e-01
1.07493855e-01 -7.71009445e-01 1.15877137e-01 7.31213093e-01
-2.09772617e-01 6.65240288e-01 1.00591674e-01 2.96977013e-01
-6.43592000e-01 -1.86142445e-01 6.48555696e-01 -3.29528511e-01
-6.56340241e-01 3.87376696e-01 -6.15904748e-01 -5.69152739e-03
1.15081263e+00 9.25235301e-02 -7.42074102e-03 -7.57569194e-01
-2.00934604e-01 5.65447152e-01 5.49650609e-01 4.34353948e-01
5.28846920e-01 -1.16703594e+00 -6.47542000e-01 1.37147605e-01
-4.18659955e-01 -2.24584892e-01 8.41022953e-02 2.50387192e-01
-6.47159815e-01 6.84824109e-01 -1.63108572e-01 2.95446724e-01
-6.00972176e-01 2.89647728e-01 5.15050590e-01 -4.29058582e-01
-5.75632513e-01 6.65562153e-01 3.14749405e-02 5.95149755e-01
1.52523726e-01 -3.48435581e-01 -2.71349162e-01 6.39040947e-01
1.08283699e+00 5.25595129e-01 -5.80222383e-02 -4.46522027e-01
-1.12670682e-01 2.58387208e-01 -2.54918009e-01 4.62694047e-03
1.68944824e+00 1.22992262e-01 -7.20416522e-03 5.35434484e-01
7.66130507e-01 -1.45843208e-01 -1.46208358e+00 -9.92608890e-02
7.36049592e-01 -6.81425512e-01 4.22880560e-01 -4.66282904e-01
-1.54391646e+00 1.07802343e+00 5.80217950e-02 4.39105958e-01
8.28898370e-01 -4.61502314e-01 1.18884969e+00 2.55118042e-01
1.32444903e-01 -1.23522747e+00 1.57272130e-01 1.70590222e-01
6.54272079e-01 -1.34599829e+00 7.75033683e-02 3.83684278e-01
-6.79790139e-01 1.09209907e+00 1.27408057e-01 -2.20359713e-01
1.06178927e+00 4.67024863e-01 4.41971958e-01 2.06589267e-01
-6.79875374e-01 3.63012627e-02 1.38890386e-01 -4.05854918e-02
3.57927173e-01 8.68120119e-02 -3.37475613e-02 7.55995274e-01
-3.47069144e-01 -1.00695580e-01 6.33203566e-01 9.72189963e-01
-7.49706566e-01 -1.21428442e+00 -4.00523156e-01 9.89964843e-01
-1.05240214e+00 -3.18572074e-02 -1.76008940e-01 8.43666434e-01
-6.40373826e-01 5.65805733e-01 1.81820512e-01 -1.86857074e-01
2.70560026e-01 1.52633950e-01 -1.87939987e-01 -4.79547054e-01
-4.02183592e-01 7.11433366e-02 -1.85105085e-01 -1.88740179e-01
1.90374061e-01 -1.10477614e+00 -1.24406338e+00 -6.38437331e-01
-1.75119609e-01 1.33587599e-01 6.30056560e-01 1.01228225e+00
2.97352690e-02 1.85016260e-01 1.07287133e+00 -5.01140118e-01
-1.26944661e+00 -3.16007257e-01 -1.09271979e+00 -1.54849673e-02
3.89337420e-01 -6.96798980e-01 -4.08045053e-01 -4.67814654e-01] | [4.481984615325928, 4.196460247039795] |
b3bccf0c-17d7-4403-b8ff-207005039fc3 | multi-crossre-a-multi-lingual-multi-domain | 2305.10985 | null | https://arxiv.org/abs/2305.10985v1 | https://arxiv.org/pdf/2305.10985v1.pdf | Multi-CrossRE A Multi-Lingual Multi-Domain Dataset for Relation Extraction | Most research in Relation Extraction (RE) involves the English language, mainly due to the lack of multi-lingual resources. We propose Multi-CrossRE, the broadest multi-lingual dataset for RE, including 26 languages in addition to English, and covering six text domains. Multi-CrossRE is a machine translated version of CrossRE (Bassignana and Plank, 2022), with a sub-portion including more than 200 sentences in seven diverse languages checked by native speakers. We run a baseline model over the 26 new datasets and--as sanity check--over the 26 back-translations to English. Results on the back-translated data are consistent with the ones on the original English CrossRE, indicating high quality of the translation and the resulting dataset. | ['Barbara Plank', 'Rob van der Goot', 'Sampo Pyysalo', 'Filip Ginter', 'Elisa Bassignana'] | 2023-05-18 | null | null | null | null | ['relation-extraction'] | ['natural-language-processing'] | [-2.99167901e-01 2.60549992e-01 -7.06481338e-01 -1.35745555e-01
-1.30130816e+00 -8.90844166e-01 6.72710180e-01 -7.50142187e-02
-5.76904595e-01 1.45654416e+00 5.50284505e-01 -4.20709431e-01
1.53192624e-01 -6.80549562e-01 -7.01208055e-01 6.46210238e-02
1.81168765e-01 7.62767494e-01 7.79716447e-02 -5.54593623e-01
-3.96001458e-01 1.10249355e-01 -6.02676570e-01 5.72948158e-01
8.48773718e-01 3.52918267e-01 -5.54524437e-02 3.21971834e-01
1.60397366e-02 7.06148446e-01 -4.04587716e-01 -1.28154504e+00
1.66908428e-01 -4.73676950e-01 -1.18416154e+00 -1.78082466e-01
3.38236898e-01 1.25080898e-01 -2.60877311e-01 8.80357265e-01
4.87526178e-01 -3.10794890e-01 4.42804456e-01 -8.42207253e-01
-1.11051035e+00 1.25749934e+00 -4.60844040e-01 3.49854678e-01
7.53959537e-01 -6.39786899e-01 1.25428760e+00 -1.33523095e+00
1.30862296e+00 1.14485466e+00 8.69630456e-01 4.55152601e-01
-1.05419600e+00 -6.84918702e-01 -3.16033959e-01 1.03291653e-01
-1.70618331e+00 -8.19601774e-01 1.92212537e-01 -1.65342435e-01
1.82831538e+00 2.23042756e-01 2.03122750e-01 1.25558722e+00
3.94076526e-01 3.20659339e-01 1.25802088e+00 -9.40340221e-01
-3.86756927e-01 4.65428054e-01 -1.61147550e-01 4.80896801e-01
4.80315477e-01 -1.28775209e-01 -7.47569621e-01 -3.51199433e-02
3.01324576e-01 -7.75153399e-01 -1.60328686e-01 2.08562985e-01
-1.38265383e+00 4.28850770e-01 -1.81035578e-01 6.24629200e-01
-2.64492959e-01 -5.84732890e-01 6.53378189e-01 7.27743804e-01
9.37366247e-01 5.69182873e-01 -1.33818829e+00 -2.34507918e-01
-7.02562928e-01 5.01047894e-02 1.11926186e+00 1.52927577e+00
5.85382700e-01 -3.03972572e-01 3.62341888e-02 1.07624543e+00
7.47894719e-02 6.48392320e-01 6.96609974e-01 -3.80136400e-01
1.43997407e+00 7.18104362e-01 -1.00605980e-01 -8.11666250e-01
-3.12419292e-02 -3.83769244e-01 -8.84428561e-01 -5.77333331e-01
-2.30371002e-02 -4.55976248e-01 -3.88316065e-01 1.51781511e+00
5.22133224e-02 -5.77722728e-01 7.67146826e-01 1.86795548e-01
1.19123256e+00 4.42844123e-01 4.87524830e-02 -3.81914943e-01
1.40685868e+00 -1.10164499e+00 -8.74321401e-01 -6.06316030e-01
8.65550280e-01 -1.05584550e+00 9.78683352e-01 2.51340568e-01
-8.94375384e-01 -3.55838716e-01 -9.44077194e-01 -3.96030664e-01
-8.00144434e-01 6.88596368e-01 4.23213214e-01 4.21565205e-01
-8.85488808e-01 3.48818511e-01 -4.37753737e-01 -6.56081021e-01
-1.08227879e-01 6.47633299e-02 -1.07130408e+00 -4.39300090e-02
-1.75613725e+00 1.59131920e+00 8.36484730e-01 -6.30315468e-02
-2.39324659e-01 -4.21891689e-01 -9.72043574e-01 -4.95231390e-01
5.58552504e-01 -2.84146518e-01 1.08955479e+00 -4.11205590e-01
-1.31752598e+00 1.26317811e+00 -4.50399339e-01 -3.79131615e-01
5.65655351e-01 -6.06635273e-01 -1.03084421e+00 -3.82332474e-01
4.39079434e-01 1.34235904e-01 8.91982391e-02 -9.40790594e-01
-6.01493895e-01 -1.25529170e-01 -2.10536361e-01 1.81821093e-01
-1.39573827e-01 6.82143569e-01 -6.07415855e-01 -8.56463373e-01
-2.13259280e-01 -9.74065840e-01 -9.68868285e-02 -1.01003838e+00
-7.85432339e-01 -3.35816324e-01 4.56730008e-01 -1.25329721e+00
1.68212485e+00 -1.83056867e+00 5.43437935e-02 -2.18153179e-01
-1.15186438e-01 2.84960061e-01 -2.71263659e-01 8.51018369e-01
-5.01421452e-01 5.43073356e-01 -1.62658900e-01 -4.63952035e-01
-2.04425022e-01 2.91856796e-01 -1.10487081e-01 1.92889243e-01
4.03709233e-01 1.39603853e+00 -9.03369427e-01 -7.00847983e-01
-4.72922809e-02 2.71767199e-01 -2.87820101e-02 -8.41199234e-02
9.54723135e-02 2.61759281e-01 -3.64941061e-01 6.47892654e-01
3.31773996e-01 1.45450011e-01 4.62369800e-01 -1.02112077e-01
-2.68629417e-02 1.02495587e+00 -9.23057795e-01 1.58742309e+00
-8.24181736e-01 6.19887412e-01 -4.44413036e-01 -2.11307690e-01
1.05277312e+00 8.32770169e-01 3.76743078e-01 -6.39598846e-01
-2.57753278e-03 8.11223149e-01 2.17143297e-02 -5.05894423e-01
6.89492106e-01 -4.88287285e-02 -3.91567826e-01 3.21792215e-01
2.26638392e-01 -2.61141449e-01 7.71048367e-01 3.37724909e-02
9.20566678e-01 6.23596132e-01 1.00645959e+00 -2.61187822e-01
7.24845469e-01 4.22802806e-01 7.22503304e-01 4.70169149e-02
3.14723551e-01 4.86320406e-01 2.40144268e-01 -9.84291136e-02
-1.02661276e+00 -5.94892085e-01 -4.06986684e-01 7.24000335e-01
-3.57733518e-01 -9.68087554e-01 -6.77644551e-01 -1.00036013e+00
-2.46979535e-01 9.45264697e-01 -3.86052698e-01 2.18875289e-01
-8.64544749e-01 -9.88908410e-01 9.29441154e-01 2.51304150e-01
6.13827705e-01 -1.19956946e+00 1.05965631e-02 5.29394805e-01
-7.59664714e-01 -1.90718472e+00 -3.52498621e-01 3.09873819e-01
-4.71794754e-01 -9.81849909e-01 -3.17654192e-01 -9.95207667e-01
1.08682118e-01 -4.28173512e-01 1.86317706e+00 -5.17743826e-01
1.36815339e-01 -3.35068315e-01 -7.09451199e-01 -1.96735919e-01
-8.81281078e-01 5.70284665e-01 1.15317963e-01 -6.29395664e-01
1.04309082e+00 -3.35147679e-01 5.18313885e-01 1.67953730e-01
-2.90619135e-01 -6.32903650e-02 6.29292309e-01 5.69966912e-01
6.17270231e-01 7.56595209e-02 7.09258676e-01 -1.33143508e+00
7.47218668e-01 -5.11371195e-01 -2.44007662e-01 7.65971243e-01
-6.53803229e-01 4.75477055e-02 6.61906719e-01 -3.14465910e-01
-8.39966297e-01 -3.16672355e-01 -1.54430836e-01 3.47598881e-01
7.49318898e-02 8.11225772e-01 -4.94485587e-01 2.61522651e-01
6.84943318e-01 -6.61276728e-02 -8.02119911e-01 -7.12843716e-01
5.23154557e-01 1.08457947e+00 7.39545524e-01 -7.86648750e-01
6.86949492e-01 -4.15270597e-01 -1.70386568e-01 -5.50285161e-01
-9.64004636e-01 -1.63110033e-01 -1.16210914e+00 4.42239761e-01
8.26396465e-01 -1.10446203e+00 1.57243878e-01 3.36073458e-01
-1.40286767e+00 -2.18819723e-01 -2.87924379e-01 5.66780627e-01
-3.64032388e-02 1.19090796e-01 -9.33402121e-01 -2.97381908e-01
-5.43166459e-01 -8.79342914e-01 1.03830647e+00 -3.47696185e-01
-5.72337508e-01 -1.27733648e+00 4.64740396e-01 1.95894703e-01
1.37881096e-03 2.11392194e-01 9.85249639e-01 -7.66196191e-01
2.31219694e-01 -1.51893437e-01 -1.73066020e-01 4.05491561e-01
6.35486543e-01 1.42107964e-01 -6.33525908e-01 -1.02412149e-01
-2.92046219e-01 -5.61974764e-01 4.19754595e-01 -2.33222336e-01
3.29906434e-01 -2.78630376e-01 -2.54644394e-01 2.26688653e-01
1.50088894e+00 5.94803803e-02 6.76811576e-01 6.75978601e-01
7.42206573e-01 6.02760673e-01 7.00842917e-01 -7.57583752e-02
8.86228561e-01 7.51967788e-01 -3.39155763e-01 -3.37919891e-01
-2.97273368e-01 -5.57114840e-01 6.99045241e-01 1.79606545e+00
-2.39785388e-01 -4.32613790e-01 -9.83665824e-01 8.01892638e-01
-1.51595938e+00 -5.41742682e-01 -4.51175630e-01 1.87533355e+00
1.67640078e+00 5.51049039e-02 -6.21894971e-02 -5.05344458e-02
5.42814612e-01 -1.30245790e-01 1.97132695e-02 -6.60462499e-01
-6.57925904e-01 6.19801044e-01 6.66446447e-01 7.29075193e-01
-1.11226928e+00 1.54060102e+00 6.79535246e+00 9.34729815e-01
-7.83362031e-01 3.04141700e-01 4.14402217e-01 4.18230921e-01
-1.83098376e-01 2.44683266e-01 -1.34639227e+00 1.04154117e-01
1.28550053e+00 -4.68270838e-01 2.91296601e-01 4.74888265e-01
-5.53756021e-02 4.15466949e-02 -1.39941287e+00 7.29190230e-01
3.29204977e-01 -9.84502435e-01 -9.39213410e-02 -1.27829965e-02
8.98586631e-01 4.17289406e-01 -5.16108871e-01 4.75831509e-01
7.83433855e-01 -1.03493690e+00 7.41357505e-01 1.72755525e-01
1.49118876e+00 -6.39906108e-01 1.02451658e+00 3.40985000e-01
-1.39121640e+00 6.52842879e-01 -3.06461751e-01 2.21556410e-01
2.83838779e-01 6.21949852e-01 -6.73465669e-01 1.13268423e+00
7.68219471e-01 1.14464414e+00 -1.00120938e+00 9.74988863e-02
-7.26306081e-01 5.53353488e-01 -2.76870251e-01 1.70596346e-01
-2.13122815e-01 -3.40776801e-01 4.86984730e-01 1.77000844e+00
2.84947067e-01 -3.10508132e-01 7.61723220e-02 2.88827717e-01
-3.55008572e-01 4.78209972e-01 -6.50521398e-01 -1.88113466e-01
5.69845259e-01 1.21088183e+00 -1.64725825e-01 -5.07450402e-01
-7.30169833e-01 1.05269051e+00 5.90853691e-01 2.95480222e-01
-4.68973517e-01 -6.05284691e-01 1.64483249e-01 -1.93608984e-01
7.59262145e-02 -3.42693299e-01 -2.30704755e-01 -1.53327179e+00
3.57805878e-01 -1.37869573e+00 2.59945273e-01 -5.53691924e-01
-1.55518055e+00 1.36203361e+00 4.31438014e-02 -1.13903332e+00
-5.38503587e-01 -5.58203161e-01 3.01199585e-01 1.18701243e+00
-1.62114370e+00 -1.49380815e+00 4.47769016e-01 4.54575837e-01
6.13266885e-01 -4.90831673e-01 1.26840019e+00 8.38358760e-01
-6.70490086e-01 8.84245276e-01 -8.37177634e-02 5.97287953e-01
1.17144310e+00 -1.18051565e+00 8.95850718e-01 1.12692392e+00
5.95654011e-01 8.10243547e-01 2.79799402e-01 -1.01228476e+00
-9.86235321e-01 -1.28463113e+00 2.31575632e+00 -1.02046251e+00
1.12808108e+00 -5.24231076e-01 -8.44955921e-01 1.18970048e+00
6.50059998e-01 -1.35632366e-01 5.34665167e-01 2.04983830e-01
-2.39865303e-01 1.66832134e-01 -1.07971287e+00 5.51904380e-01
1.16281164e+00 -8.15494061e-01 -1.02716923e+00 3.06240022e-01
6.05261862e-01 -6.19366705e-01 -1.54199624e+00 5.21338761e-01
2.61798888e-01 -2.77678251e-01 6.22051239e-01 -7.33135283e-01
6.67585552e-01 -1.11045539e-01 -4.14472759e-01 -1.42303097e+00
-1.81843311e-01 -6.41378045e-01 -1.83622986e-02 1.87716389e+00
1.23692322e+00 -8.14612150e-01 1.64833739e-01 1.61013544e-01
4.75620478e-02 -6.72011733e-01 -9.42671180e-01 -9.46287632e-01
4.56422240e-01 -4.59402889e-01 7.40796864e-01 1.19291735e+00
2.91665215e-02 1.01078486e+00 -4.22481120e-01 -9.47901085e-02
1.56461865e-01 -1.26258537e-01 7.07145572e-01 -9.54120278e-01
-2.54051536e-01 1.28628835e-01 9.83205810e-02 -7.41921663e-01
4.49459434e-01 -1.28933144e+00 -5.68107842e-03 -1.70658863e+00
3.13230753e-01 -5.24788320e-01 6.98053166e-02 1.04404199e+00
-2.16856614e-01 5.50830126e-01 -2.34545678e-01 2.12831989e-01
-1.77867532e-01 4.21952844e-01 1.12086701e+00 2.22189277e-01
-1.66444987e-01 -2.79191345e-01 -6.59534276e-01 6.44526720e-01
6.12343669e-01 -1.00984931e+00 -9.12454128e-02 -6.57212555e-01
6.69900000e-01 -2.73440152e-01 -3.21659327e-01 -5.33252239e-01
-1.47683620e-01 -2.06200942e-01 9.22652185e-02 -4.94442880e-01
-1.28780484e-01 -7.11821079e-01 1.94244117e-01 -8.17217585e-03
-1.69647619e-01 7.80685365e-01 3.00570935e-01 -1.19865805e-01
-6.21701300e-01 -2.61004478e-01 3.98868799e-01 -2.62563169e-01
-6.23076975e-01 5.62244561e-03 -2.26901814e-01 8.02531421e-01
6.73239470e-01 8.87356251e-02 -4.86004472e-01 4.52236384e-02
-4.80160981e-01 -4.70356382e-02 2.70192862e-01 7.08869338e-01
-8.36617686e-03 -1.38180637e+00 -1.35791373e+00 3.54556218e-02
2.05656186e-01 -1.04836181e-01 -7.33983517e-01 6.02728188e-01
-4.42880839e-01 5.99976063e-01 1.18169390e-01 1.14183113e-01
-1.19638658e+00 2.17109084e-01 2.89342016e-01 -7.83505440e-01
-6.46940112e-01 6.13112450e-01 -7.10058033e-01 -9.20758605e-01
-3.41549814e-01 -4.25332308e-01 -3.38459283e-01 -3.54382992e-02
1.41572371e-01 3.24265003e-01 5.72933674e-01 -9.90128517e-01
-7.92573869e-01 7.09824383e-01 7.20969588e-02 -4.43124294e-01
1.43320560e+00 -3.65553707e-01 -7.19716012e-01 6.54956460e-01
9.92717981e-01 1.00730228e+00 -2.70641148e-01 -5.82539797e-01
7.06474543e-01 -6.12706654e-02 -2.85639554e-01 -1.18922651e+00
-8.60690415e-01 3.75553936e-01 -1.10703044e-01 -2.18296409e-01
8.74764860e-01 1.30508885e-01 9.28292215e-01 4.79171425e-01
6.82222307e-01 -1.03790438e+00 -6.64496720e-01 1.07357252e+00
9.47397470e-01 -1.30463982e+00 3.72671485e-01 -7.72163332e-01
-8.35012376e-01 9.24430609e-01 4.87916470e-01 1.22935221e-01
7.23909020e-01 6.25376582e-01 4.61301208e-01 6.59789890e-02
-7.86653936e-01 -2.30345987e-02 5.53920984e-01 5.89149833e-01
1.03007960e+00 2.40697220e-01 -7.48245478e-01 8.81444514e-01
-8.78408968e-01 1.46828100e-01 2.93282509e-01 8.02942693e-01
5.47740340e-01 -1.80524588e+00 1.94021508e-01 1.46461487e-01
-9.26095843e-01 -6.77285433e-01 -9.24044430e-01 1.14258480e+00
4.41714942e-01 1.30696869e+00 -4.12460476e-01 -4.28088009e-01
5.69703758e-01 5.08828387e-02 5.20821214e-01 -1.14715278e+00
-8.57747376e-01 -3.92369628e-02 1.09222472e+00 -1.74562544e-01
-6.10609829e-01 -6.40903592e-01 -8.94875765e-01 -2.97023803e-01
-4.25273776e-01 3.68352711e-01 3.76196653e-01 1.16298771e+00
1.24947101e-01 3.70041996e-01 4.06636834e-01 7.51194134e-02
-1.51920512e-01 -1.47899902e+00 -3.39770854e-01 9.41774473e-02
-2.11594235e-02 -3.74896497e-01 -2.40328804e-01 3.10163826e-01] | [10.550929069519043, 9.441484451293945] |
ac93416c-847f-4f19-b9de-46ab14a35145 | multi-level-contrast-network-for-wearables | 2208.07547 | null | https://arxiv.org/abs/2208.07547v1 | https://arxiv.org/pdf/2208.07547v1.pdf | Multi-level Contrast Network for Wearables-based Joint Activity Segmentation and Recognition | Human activity recognition (HAR) with wearables is promising research that can be widely adopted in many smart healthcare applications. In recent years, the deep learning-based HAR models have achieved impressive recognition performance. However, most HAR algorithms are susceptible to the multi-class windows problem that is essential yet rarely exploited. In this paper, we propose to relieve this challenging problem by introducing the segmentation technology into HAR, yielding joint activity segmentation and recognition. Especially, we introduce the Multi-Stage Temporal Convolutional Network (MS-TCN) architecture for sample-level activity prediction to joint segment and recognize the activity sequence. Furthermore, to enhance the robustness of HAR against the inter-class similarity and intra-class heterogeneity, a multi-level contrastive loss, containing the sample-level and segment-level contrast, has been proposed to learn a well-structured embedding space for better activity segmentation and recognition performance. Finally, with comprehensive experiments, we verify the effectiveness of the proposed method on two public HAR datasets, achieving significant improvements in the various evaluation metrics. | ['Robert C. Qiu', 'Wenxian Yu', 'Ling Pei', 'Lei Chu', 'Songpengcheng Xia'] | 2022-08-16 | null | null | null | null | ['activity-prediction', 'activity-prediction'] | ['computer-vision', 'time-series'] | [ 3.15922976e-01 -4.73241746e-01 -3.26599479e-01 -2.35644639e-01
-7.73304880e-01 -9.16230772e-03 2.38583490e-01 1.13940522e-01
-3.69803369e-01 6.64830387e-01 4.96368855e-01 3.16303849e-01
-1.09935813e-01 -5.21020651e-01 -3.65919918e-01 -9.63494003e-01
-1.31651342e-01 -4.37697709e-01 2.36074135e-01 3.59119445e-01
-1.52730510e-01 1.36072800e-01 -1.27108645e+00 1.19883686e-01
9.56502497e-01 1.24443126e+00 -1.81857228e-01 2.74598867e-01
1.15060531e-01 7.73259342e-01 -5.41189432e-01 -9.88708511e-02
5.28587624e-02 -7.93259203e-01 -3.75263929e-01 1.38430417e-01
1.00874733e-02 -6.40344247e-02 -3.31120223e-01 7.55217314e-01
7.56020010e-01 1.32711247e-01 4.03025210e-01 -9.28824842e-01
-2.15198621e-01 2.82092750e-01 -6.57245874e-01 3.30630481e-01
3.93649220e-01 3.29435587e-01 7.15539694e-01 -7.69583881e-01
1.76895455e-01 8.08315158e-01 8.10683489e-01 1.30647957e-01
-1.14342666e+00 -6.18579388e-01 -4.78539802e-02 5.17809749e-01
-1.57174003e+00 -1.11999959e-01 1.07618904e+00 -3.75415146e-01
6.87496662e-01 2.71461487e-01 1.12055051e+00 1.38411343e+00
3.57363373e-01 1.00902331e+00 1.05434275e+00 -9.28314924e-02
1.89342931e-01 -2.18044966e-01 4.48604226e-02 6.80294931e-01
2.67161310e-01 -1.80962518e-01 -4.86330509e-01 -1.55308768e-02
8.21502030e-01 3.31398726e-01 -3.54105204e-01 -3.80013466e-01
-1.44313514e+00 5.42190790e-01 5.61064482e-01 5.74999809e-01
-5.74702203e-01 -2.29314622e-03 6.58748448e-01 -4.54277813e-01
3.01750422e-01 1.34671092e-01 -8.78235027e-02 -3.48780304e-01
-1.13947570e+00 -1.57197297e-01 5.06378293e-01 4.20408845e-01
2.73142278e-01 2.48002201e-01 -7.39624739e-01 9.95345771e-01
3.83731186e-01 2.82377869e-01 7.12875426e-01 -6.08893275e-01
3.59551340e-01 7.37339735e-01 -7.66156688e-02 -1.17977381e+00
-6.75622463e-01 -9.51716006e-01 -1.27886009e+00 -4.86653090e-01
2.47378051e-01 -5.15033193e-02 -7.08642125e-01 1.60860538e+00
4.71695483e-01 8.18665564e-01 -7.19706640e-02 9.19155300e-01
6.57409668e-01 6.82289183e-01 5.20009279e-01 -6.67773902e-01
1.65280604e+00 -1.16736245e+00 -1.05855322e+00 -1.58330247e-01
5.84235966e-01 -4.34794605e-01 9.33957040e-01 3.81113023e-01
-8.52027595e-01 -6.83554530e-01 -1.41349816e+00 8.74343663e-02
-2.60470398e-02 2.99897045e-01 5.34816146e-01 6.66176915e-01
-2.50563860e-01 2.87676156e-01 -1.09034181e+00 -3.21063876e-01
8.07878196e-01 1.72162294e-01 -1.52697086e-01 -1.80661753e-02
-1.37743068e+00 4.75769311e-01 4.54917282e-01 5.15629292e-01
-8.89853895e-01 -5.24696112e-01 -8.01356852e-01 8.23685452e-02
1.99602067e-01 -4.37755853e-01 8.33754480e-01 -7.32819915e-01
-1.26962197e+00 4.64891195e-01 -8.17368366e-03 -5.28248131e-01
5.22012055e-01 -2.82081991e-01 -8.14004004e-01 1.79915771e-01
-4.66258675e-02 6.93752542e-02 6.30141437e-01 -7.37723947e-01
-3.28201741e-01 -5.53645432e-01 -4.85098034e-01 2.66414344e-01
-5.81463814e-01 -3.11730653e-01 -5.07509530e-01 -1.01369560e+00
1.73530392e-02 -7.05212891e-01 -4.17615362e-02 -2.07028165e-01
-3.09212774e-01 -1.54191241e-01 5.91466188e-01 -9.35453653e-01
1.65277565e+00 -2.41096783e+00 -9.17904377e-02 1.71529293e-01
1.30117804e-01 4.73991603e-01 1.18996471e-01 2.25108266e-01
2.01182291e-01 -2.33430192e-01 -4.76946443e-01 -2.07087710e-01
-1.17761418e-01 1.05368957e-01 2.43478060e-01 6.81688726e-01
9.08359587e-02 1.00689209e+00 -7.41677046e-01 -6.75959647e-01
2.63646781e-01 7.52725899e-01 -3.10311913e-01 2.52288997e-01
1.41726524e-01 6.28678560e-01 -6.02296352e-01 6.45556211e-01
4.12984729e-01 -3.94505531e-01 2.56417632e-01 -5.39483905e-01
7.96295181e-02 -3.47055458e-02 -1.24268293e+00 1.81400990e+00
-1.35701746e-01 1.79372013e-01 -3.61889780e-01 -1.29517305e+00
9.37703013e-01 3.60662550e-01 1.03062797e+00 -9.48876083e-01
1.76112816e-01 2.48003006e-01 -4.96690460e-02 -6.02905810e-01
4.19275239e-02 4.38490920e-02 -1.52887434e-01 9.54939649e-02
-1.00750424e-01 7.72558033e-01 -7.95030519e-02 -3.27398181e-01
1.11975658e+00 1.91500053e-01 4.69295621e-01 -4.49342690e-02
7.07068682e-01 -4.48040456e-01 1.04796898e+00 4.05324370e-01
-7.04592466e-01 5.17411053e-01 2.01129302e-01 -3.90174925e-01
-6.96795523e-01 -1.04846537e+00 -2.21500963e-01 7.40535796e-01
4.39571768e-01 -1.91340238e-01 -8.66013765e-01 -7.29867280e-01
-3.12995136e-01 2.26993635e-01 -5.68431616e-01 -4.68925804e-01
-7.28431821e-01 -1.23359966e+00 8.59006107e-01 8.05222750e-01
1.13242376e+00 -1.17458165e+00 -7.00932920e-01 6.33220851e-01
-4.85284567e-01 -1.13131130e+00 -6.54789686e-01 5.68667464e-02
-7.96571493e-01 -1.07437515e+00 -9.23295856e-01 -9.24273551e-01
2.10470766e-01 1.18263707e-01 5.42395771e-01 4.78733256e-02
-5.84304273e-01 1.83664709e-01 -3.96315366e-01 -1.34965509e-01
3.45921546e-01 2.09868625e-01 -1.55650869e-01 5.43989360e-01
4.79751080e-01 -6.99244380e-01 -1.18782330e+00 3.85087162e-01
-9.33481216e-01 -6.76456047e-03 8.24266732e-01 8.99796247e-01
7.87983358e-01 -5.25188185e-02 7.68266499e-01 -4.07271206e-01
5.17951369e-01 -4.25216317e-01 5.86018013e-03 3.55811864e-01
-7.12572098e-01 -2.16619596e-01 6.24903798e-01 -5.09212673e-01
-1.05260849e+00 9.83174443e-02 -2.29623139e-01 -1.55294359e-01
-1.33655388e-02 5.35253227e-01 -4.18340117e-01 1.64764717e-01
4.12492871e-01 6.09338403e-01 -1.74836516e-01 -4.88273323e-01
2.55396403e-02 5.91324270e-01 5.29465973e-01 -2.80690610e-01
3.62569928e-01 4.06645179e-01 -8.43927562e-02 -8.53479862e-01
-8.30010355e-01 -5.98024070e-01 -3.40660691e-01 -1.44151583e-01
1.32839477e+00 -1.12190437e+00 -4.88341242e-01 7.80019701e-01
-8.68230164e-01 -1.45206541e-01 -1.89513937e-01 6.97595775e-01
-3.95970523e-01 5.86398482e-01 -5.39292693e-01 -8.08994234e-01
-6.65570378e-01 -1.08825076e+00 9.66044247e-01 6.73104107e-01
-3.99399623e-02 -7.94548988e-01 2.20258042e-01 6.20578945e-01
2.57956803e-01 6.14920259e-01 6.04837298e-01 -7.62919009e-01
-4.86960858e-01 -1.50636762e-01 -1.10845923e-01 4.31295246e-01
2.51222134e-01 -4.09640282e-01 -8.44683766e-01 -2.16435537e-01
1.18151782e-02 -1.32633358e-01 7.09066212e-01 3.73839617e-01
1.35675693e+00 -2.12668240e-01 -4.15114135e-01 6.65935993e-01
1.15125358e+00 3.30505371e-01 9.37920511e-01 3.08873713e-01
9.13237154e-01 1.49040088e-01 7.62098610e-01 6.66930556e-01
4.38228726e-01 8.24851871e-01 -1.56057226e-02 -3.37314457e-01
-1.39461115e-01 -1.78910390e-01 4.54480648e-01 1.07772541e+00
-1.10284753e-01 8.05348009e-02 -5.86144924e-01 4.55769956e-01
-1.93696320e+00 -1.02037132e+00 -6.27133716e-03 2.07443142e+00
7.95324862e-01 8.89912024e-02 3.44341427e-01 3.17001134e-01
5.36848783e-01 5.71230769e-01 -6.72310650e-01 4.31398749e-02
-2.19049558e-01 1.66515902e-01 2.60623425e-01 -3.47428694e-02
-1.40927875e+00 3.69795114e-01 5.20071840e+00 1.06727421e+00
-1.09162772e+00 1.97563305e-01 5.65333247e-01 5.23263291e-02
8.31000507e-03 -4.29381371e-01 -5.50241649e-01 8.67016435e-01
7.77823448e-01 1.87418342e-01 2.62593299e-01 7.18414009e-01
4.06831950e-01 -6.43194914e-02 -7.48095393e-01 1.26113844e+00
9.13580582e-02 -9.71041083e-01 -2.61430919e-01 2.95359157e-02
5.56521475e-01 -3.79341483e-01 -1.81924641e-01 5.04462123e-01
-5.77327549e-01 -8.54500890e-01 2.90010303e-01 6.43885612e-01
5.24208784e-01 -7.65868366e-01 8.95206869e-01 1.95779085e-01
-1.63666987e+00 -2.17160165e-01 1.06484279e-01 2.87564099e-01
1.73396736e-01 6.29913032e-01 -4.18290824e-01 8.32438827e-01
7.24507809e-01 9.86874819e-01 -6.09288216e-01 1.31925511e+00
-2.73446105e-02 6.79293156e-01 -1.01669475e-01 8.40176549e-03
6.78396299e-02 -2.51603723e-01 2.82112449e-01 1.39639556e+00
2.79867947e-01 7.00655058e-02 4.05783057e-01 4.31371450e-01
-3.37544568e-02 3.49815428e-01 -1.38516888e-01 -2.70904064e-01
2.87266523e-01 1.14772010e+00 -6.71331227e-01 -1.79965422e-01
-4.53005970e-01 1.07625473e+00 1.26981074e-02 3.14200044e-01
-1.16846573e+00 -5.66135585e-01 5.01007140e-01 1.47311911e-01
2.99685895e-01 -2.15124980e-01 -3.15994263e-01 -1.24268723e+00
1.74457312e-01 -9.47387815e-01 6.32593513e-01 -2.50847191e-01
-1.02286851e+00 1.77175939e-01 -2.19329864e-01 -1.41839015e+00
2.32730463e-01 -1.03476480e-01 -6.34775341e-01 4.12002623e-01
-1.29549658e+00 -1.29148328e+00 -5.39041817e-01 7.21092939e-01
5.52520394e-01 1.64947644e-01 6.41794682e-01 8.87673080e-01
-1.12635720e+00 9.15316761e-01 1.71962231e-02 3.95572215e-01
5.66575348e-01 -8.36716890e-01 -1.66920602e-01 9.64942873e-01
2.08693370e-02 6.14922225e-01 3.75180930e-01 -5.11999786e-01
-1.27458119e+00 -1.15098298e+00 4.51436728e-01 -2.62618773e-02
3.35354358e-01 -3.04011226e-01 -1.07949841e+00 4.68829632e-01
-2.60982681e-02 2.70139307e-01 9.16840672e-01 -1.94860876e-01
-1.16523124e-01 -6.67518735e-01 -1.13580000e+00 4.56893623e-01
1.10589015e+00 -3.94546002e-01 -5.10930538e-01 3.04659549e-02
4.27213788e-01 -1.63609743e-01 -1.29050267e+00 7.76705980e-01
9.00176585e-01 -8.28248382e-01 9.85239506e-01 -2.86167800e-01
2.54031047e-02 -5.40051758e-01 -2.14456413e-02 -8.90897036e-01
-2.61605442e-01 -3.65933806e-01 -5.40949345e-01 1.19348288e+00
-1.23374619e-01 -4.66838479e-01 7.38515973e-01 1.24322861e-01
-1.74992964e-01 -9.87035692e-01 -1.17878079e+00 -7.48937607e-01
-5.05449474e-01 -2.05766901e-01 4.19341087e-01 8.28843772e-01
8.97035655e-03 2.42689937e-01 -7.98347414e-01 5.73559627e-02
6.63776636e-01 2.72021741e-02 5.79130888e-01 -1.06427455e+00
-2.39317670e-01 -2.53316760e-01 -5.08517206e-01 -1.01442897e+00
-3.67061585e-01 -6.69648290e-01 2.10130051e-01 -1.57102287e+00
3.66222888e-01 -4.84441184e-02 -9.53763545e-01 3.00658554e-01
-2.73699224e-01 3.76913279e-01 -1.08401380e-01 3.16615373e-01
-9.95483637e-01 9.55184996e-01 1.29250491e+00 -1.61678880e-01
-3.51928920e-01 -1.71395421e-01 -4.71676350e-01 4.90255952e-01
8.02453756e-01 -3.45861971e-01 -3.92529368e-01 4.72698212e-02
-2.80825675e-01 9.08210650e-02 3.42145383e-01 -1.48471725e+00
1.11686334e-01 -8.27730000e-02 7.04294026e-01 -5.15386403e-01
4.18597162e-01 -7.95744598e-01 3.08207929e-01 7.43227303e-01
-2.13746250e-01 -3.34525049e-01 -4.88515310e-02 9.09394741e-01
-2.37682775e-01 3.81918490e-01 9.72172916e-01 1.76837683e-01
-5.87119460e-01 5.44530749e-01 -2.26889044e-01 6.63222373e-02
1.23963141e+00 -3.92416358e-01 -1.15120314e-01 6.96174949e-02
-6.15855575e-01 2.61982709e-01 -7.33855926e-03 4.82989937e-01
5.75011313e-01 -1.66540110e+00 -3.18397880e-01 2.37031445e-01
4.40302610e-01 -1.95560202e-01 8.15233588e-01 1.50964904e+00
-3.54701966e-01 2.38142461e-01 -1.99743778e-01 -5.12768269e-01
-1.06589401e+00 3.72147918e-01 3.67513895e-01 -5.53696871e-01
-7.12490916e-01 4.12683964e-01 3.07545979e-02 5.14807627e-02
4.09282357e-01 -2.49537662e-01 -3.59188914e-01 1.06776789e-01
6.41730189e-01 5.66988587e-01 -9.50230584e-02 -5.96474349e-01
-6.61354959e-01 5.19778788e-01 1.13386884e-01 3.16110730e-01
1.14294708e+00 -1.18509457e-01 1.73037767e-01 4.93948370e-01
1.23410964e+00 -2.11910799e-01 -1.47549784e+00 -2.09475204e-01
1.35928378e-01 -1.62972018e-01 -4.81055398e-03 -7.27509439e-01
-9.61761594e-01 9.48576152e-01 1.13286448e+00 4.53954488e-02
1.23554718e+00 -4.09702599e-01 1.42379010e+00 -3.06763612e-02
1.51358336e-01 -1.23913443e+00 4.80911404e-01 3.50449122e-02
4.51688886e-01 -1.04621243e+00 5.53766340e-02 -1.23444565e-01
-6.42489195e-01 8.06918502e-01 5.30126750e-01 -9.34758559e-02
6.04048133e-01 -1.69447020e-01 -6.30390272e-02 -1.41170278e-01
-8.90770704e-02 4.52298224e-02 3.86975765e-01 4.86820579e-01
3.36205631e-01 1.81889802e-01 -8.13718259e-01 9.95999813e-01
4.28862721e-01 3.26323301e-01 8.25192258e-02 8.96163642e-01
-4.88680094e-01 -9.07289922e-01 -1.59384742e-01 4.98827457e-01
-6.83017671e-01 1.95100784e-01 1.05652057e-01 4.68017250e-01
2.96002060e-01 7.62993157e-01 -3.37565750e-01 -5.19102275e-01
4.84499753e-01 1.31919384e-01 3.32415998e-01 -3.20414156e-01
-5.27101219e-01 3.50001574e-01 -6.77017421e-02 -6.75903618e-01
-6.44062221e-01 -6.36550784e-01 -1.29317391e+00 1.51570141e-01
-6.33137599e-02 1.35434762e-01 3.18921685e-01 9.94852185e-01
4.17468131e-01 6.92501247e-01 6.38963819e-01 -4.41159248e-01
-4.54710543e-01 -8.70273054e-01 -7.85328686e-01 4.93377030e-01
2.50067532e-01 -7.69407809e-01 -1.71464145e-01 1.02301054e-02] | [7.736607074737549, 0.850323498249054] |
bffd2a61-0d88-4a06-9f35-2aea90f47671 | text-to-audio-grounding-based-novel-metric | 2210.06354 | null | https://arxiv.org/abs/2210.06354v1 | https://arxiv.org/pdf/2210.06354v1.pdf | Text-to-Audio Grounding Based Novel Metric for Evaluating Audio Caption Similarity | Automatic Audio Captioning (AAC) refers to the task of translating an audio sample into a natural language (NL) text that describes the audio events, source of the events and their relationships. Unlike NL text generation tasks, which rely on metrics like BLEU, ROUGE, METEOR based on lexical semantics for evaluation, the AAC evaluation metric requires an ability to map NL text (phrases) that correspond to similar sounds in addition lexical semantics. Current metrics used for evaluation of AAC tasks lack an understanding of the perceived properties of sound represented by text. In this paper, wepropose a novel metric based on Text-to-Audio Grounding (TAG), which is, useful for evaluating cross modal tasks like AAC. Experiments on publicly available AAC data-set shows our evaluation metric to perform better compared to existing metrics used in NL text and image captioning literature. | ['Sunil Kumar Kopparapu', 'Rupayan Chakraborty', 'Swapnil Bhosale'] | 2022-10-03 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 6.58478916e-01 2.41576493e-01 1.32845163e-01 -1.84139639e-01
-1.23745441e+00 -6.26273453e-01 9.29897666e-01 5.99997461e-01
-1.65183157e-01 9.79864836e-01 1.13460243e+00 4.42126133e-02
1.13879731e-02 -3.99660796e-01 -5.54017067e-01 -1.75507545e-01
3.89184840e-02 2.53354818e-01 1.68621495e-01 -1.78873558e-02
1.40419036e-01 -1.76711455e-01 -1.75973082e+00 8.49312842e-01
4.90147412e-01 1.09189939e+00 3.30404580e-01 1.11872840e+00
-4.95281249e-01 8.48051786e-01 -9.13093209e-01 -1.06458120e-01
-3.60819459e-01 -9.24556494e-01 -1.00217998e+00 -3.42847914e-01
3.95558089e-01 2.18101636e-01 1.53240755e-01 7.90862679e-01
5.44062674e-01 4.46404293e-02 9.77650404e-01 -1.51207042e+00
-4.15089875e-01 9.88011003e-01 2.48913869e-01 -1.17434969e-03
1.24675167e+00 -3.53637666e-01 1.24931669e+00 -7.09639847e-01
5.91082335e-01 1.26330233e+00 7.62432337e-01 5.39130747e-01
-1.10802376e+00 -4.90015417e-01 -2.82573104e-01 1.16662286e-01
-1.23724711e+00 -4.20898706e-01 5.93061507e-01 -7.01749086e-01
6.56838834e-01 4.77377385e-01 5.02465189e-01 1.39884913e+00
-7.59832980e-03 3.93450469e-01 9.37804937e-01 -8.07552695e-01
3.26082677e-01 1.29173324e-01 -1.49882808e-01 2.33754031e-02
-2.38331228e-01 -1.67813107e-01 -1.05495429e+00 -2.59079099e-01
3.05679530e-01 -8.67509663e-01 -3.48380953e-01 1.09032542e-01
-1.60961902e+00 6.63669407e-01 -3.90722975e-02 4.29910630e-01
-4.69530612e-01 4.24304962e-01 8.24914157e-01 2.46901900e-01
1.67662814e-01 5.82470715e-01 -1.78356633e-01 -6.35365009e-01
-1.04881167e+00 3.61310452e-01 8.80826592e-01 9.32997942e-01
3.02778631e-01 -2.58973520e-02 -6.46315336e-01 7.26655602e-01
5.59609175e-01 6.61358833e-01 8.79883707e-01 -8.18957865e-01
3.29269290e-01 1.44995943e-01 3.15523267e-01 -6.53119862e-01
-9.63686183e-02 -6.66451007e-02 -5.40775716e-01 -1.94164574e-01
6.02311119e-02 -1.22125082e-01 -8.57715786e-01 1.69677663e+00
-1.43340573e-01 3.29861224e-01 2.94810027e-01 6.83099151e-01
1.18918121e+00 1.06402898e+00 4.48951721e-01 -3.30266744e-01
1.58256376e+00 -6.23096764e-01 -9.63612974e-01 8.23295563e-02
2.31736526e-01 -1.35729337e+00 1.58578253e+00 3.51360977e-01
-9.39352989e-01 -8.28791022e-01 -1.10018468e+00 3.19420062e-02
-4.65745240e-01 -9.67742410e-03 2.36991793e-01 7.23687470e-01
-1.19813049e+00 2.11300895e-01 -1.24661401e-01 -5.98523140e-01
-1.43874973e-01 -1.08901188e-01 -2.38490999e-01 5.70641398e-01
-1.32081974e+00 4.37756985e-01 6.50076270e-01 -4.87521082e-01
-1.17338610e+00 -6.03977144e-01 -8.56526792e-01 -5.75012416e-02
-1.49863958e-01 -7.45864153e-01 1.70536196e+00 -1.18576097e+00
-1.73440933e+00 7.65874684e-01 -1.83501184e-01 -7.77164996e-01
2.86444426e-01 -2.80527771e-01 -6.41061187e-01 4.10895348e-01
2.77395517e-01 1.18419027e+00 8.99818659e-01 -1.29053199e+00
-6.98991954e-01 2.97723353e-01 -8.29501450e-02 3.77565771e-01
-2.19194651e-01 2.41102844e-01 -1.05517417e-01 -9.92095232e-01
-1.84698284e-01 -7.28181303e-01 4.13034618e-01 -1.91221699e-01
-5.79314590e-01 -1.29243314e-01 5.48241854e-01 -5.79899609e-01
1.36003435e+00 -2.37094569e+00 -1.41929567e-01 -2.37033912e-03
-3.20712596e-01 -1.08417504e-01 -1.99262440e-01 7.11317241e-01
-1.46203950e-01 3.14317405e-01 -1.46494240e-01 -2.87941366e-01
3.53669912e-01 1.06787467e-02 -8.88787031e-01 -2.70751953e-01
2.41815690e-02 4.13088709e-01 -1.06416166e+00 -7.67995477e-01
1.66727111e-01 5.37886798e-01 -3.09972614e-01 3.43592584e-01
-5.90776682e-01 4.42644030e-01 -4.01408762e-01 1.79104283e-01
-4.37402874e-02 2.32435405e-01 -3.48260880e-01 -3.04601222e-01
1.65974151e-03 5.42200506e-01 -1.05148852e+00 1.91424835e+00
-8.20385754e-01 1.08272755e+00 -4.05747920e-01 -2.91513503e-01
8.39367628e-01 1.12135923e+00 3.77073795e-01 -2.69484401e-01
7.10772946e-02 3.58782321e-01 -4.22384739e-01 -4.63950843e-01
5.38210571e-01 -8.53307918e-02 -3.49102110e-01 4.46668416e-01
1.60567209e-01 -6.19947672e-01 3.07337165e-01 1.06252700e-01
7.58119047e-01 3.09311420e-01 4.75594789e-01 -1.30927548e-01
7.64881492e-01 3.51812877e-02 -3.39011788e-01 8.70098293e-01
2.91763786e-02 1.00713360e+00 2.61075020e-01 1.05199181e-02
-1.18361652e+00 -1.28452456e+00 -1.39338821e-01 1.13050640e+00
-1.67457759e-01 -8.08077395e-01 -8.94506514e-01 -2.03485236e-01
-5.52603662e-01 8.72380018e-01 -5.11131883e-01 5.30662872e-02
-1.64608896e-01 -2.71968633e-01 1.07448697e+00 3.30146581e-01
1.93059817e-01 -1.32742107e+00 -5.09298563e-01 4.32267219e-01
-7.35802650e-01 -1.28192115e+00 -5.90344667e-01 -1.45095482e-01
-4.74320620e-01 -6.46369159e-01 -8.22386324e-01 -9.27693963e-01
-1.08974939e-02 -3.58830541e-01 1.29972839e+00 -4.73650426e-01
-2.14274392e-01 5.87486982e-01 -9.63554382e-01 -1.00036573e+00
-9.63927209e-01 -1.19156033e-01 -3.16976011e-02 -6.22965535e-03
2.21561477e-01 -5.75404346e-01 -2.78955251e-01 7.87103847e-02
-9.82165933e-01 3.30754787e-01 3.85304660e-01 3.72165561e-01
6.90926492e-01 -3.14958632e-01 8.39569628e-01 -2.56176502e-01
1.12808323e+00 -2.23003522e-01 1.25879422e-01 1.16700083e-01
-2.42564335e-01 1.33781567e-01 5.62042177e-01 -5.45243442e-01
-8.17797720e-01 -4.47110496e-02 -1.43328041e-01 -3.04061174e-02
-6.48355126e-01 3.90068799e-01 -1.16395861e-01 3.92784804e-01
9.20331776e-01 2.66901582e-01 -3.61837298e-01 -3.13649863e-01
4.21780676e-01 1.07043350e+00 7.61707664e-01 -6.78430200e-01
4.12001431e-01 2.88127333e-01 5.55346571e-02 -9.37608719e-01
-8.68501067e-01 -5.11102438e-01 -2.25560248e-01 -4.30331409e-01
1.27295196e+00 -8.95206094e-01 -3.53229076e-01 2.22084820e-02
-1.40971375e+00 1.87206001e-03 -6.31109476e-01 8.97625506e-01
-1.14157760e+00 2.00696766e-01 -2.58447677e-01 -8.89653623e-01
-6.03943586e-01 -7.53995180e-01 1.41461265e+00 -7.30001181e-02
-8.07088912e-01 -7.71567106e-01 4.12226409e-01 2.03170389e-01
3.36697698e-01 3.14953059e-01 1.03779459e+00 -7.13443935e-01
7.54695535e-02 -9.49120596e-02 -1.23863824e-01 4.17143732e-01
2.02923343e-01 -5.57294451e-02 -1.33466673e+00 3.71152610e-01
-4.30384338e-01 -4.31271464e-01 5.77742219e-01 4.09835935e-01
9.22368288e-01 -3.24029088e-01 2.70965938e-02 4.79610142e-04
1.30404007e+00 3.35137814e-01 6.68020666e-01 2.04170078e-01
3.29413772e-01 5.01827300e-01 6.86972618e-01 4.10847843e-01
2.34322205e-01 7.74388313e-01 2.55507290e-01 8.23424384e-02
-6.39723241e-01 -6.16735518e-01 6.63181961e-01 8.30810189e-01
1.60798237e-01 -6.76295817e-01 -9.64267790e-01 5.16629100e-01
-1.69069052e+00 -8.21243823e-01 -1.64676949e-01 2.17596459e+00
1.11255157e+00 1.76031575e-01 2.01656535e-01 4.82890725e-01
7.34162331e-01 -7.99380988e-02 2.64406741e-01 -6.09836936e-01
-1.33014023e-01 2.11450383e-01 -1.35376275e-01 6.53800249e-01
-1.05818117e+00 6.79024994e-01 6.77776146e+00 8.40102136e-01
-8.53373706e-01 2.31853545e-01 1.85999289e-01 1.95320264e-01
-2.90646225e-01 -1.66323438e-01 -4.19277877e-01 3.68711978e-01
1.44767344e+00 -4.45240527e-01 2.76922345e-01 4.52008694e-01
6.46136224e-01 1.45438179e-01 -1.45204091e+00 1.15987396e+00
3.42076600e-01 -1.00433588e+00 7.49180377e-01 -4.12717462e-01
5.31010568e-01 -2.44506896e-01 4.08311039e-02 1.65325403e-02
6.27902970e-02 -1.03841734e+00 1.22779715e+00 5.47823966e-01
9.58965302e-01 -4.56084251e-01 8.78404260e-01 -3.49148482e-01
-1.28774071e+00 3.55327576e-01 5.07350825e-02 2.77327579e-02
3.41209501e-01 1.57682836e-01 -1.40165818e+00 6.21381342e-01
5.21024883e-01 3.11435580e-01 -4.06184971e-01 1.24790919e+00
-2.53297538e-01 7.52665877e-01 -1.40235454e-01 -3.05382133e-01
3.19905043e-01 3.56672764e-01 8.51681948e-01 1.68323934e+00
7.52453744e-01 -5.12705207e-01 8.87108147e-02 5.52620471e-01
8.20539668e-02 7.67411470e-01 -4.82891917e-01 -4.82724458e-01
5.18196642e-01 7.06021965e-01 -6.27963603e-01 -2.85310864e-01
-6.56921938e-02 9.94785488e-01 -7.54135311e-01 2.08412394e-01
-9.15353417e-01 -5.89334488e-01 2.89035827e-01 1.79296389e-01
-2.04291493e-01 -1.91420410e-02 -1.92667022e-02 -5.27145982e-01
-7.62680173e-03 -7.51801252e-01 2.98591167e-01 -1.57994640e+00
-1.11880350e+00 1.02876484e+00 2.56134331e-01 -1.61816561e+00
-6.97618902e-01 -4.48628634e-01 -5.37004411e-01 7.11797476e-01
-1.24805117e+00 -1.28281546e+00 -4.54516679e-01 6.29423499e-01
8.46472561e-01 -2.22836539e-01 1.29591846e+00 2.33701810e-01
3.46283466e-01 1.69120163e-01 -3.39089185e-01 -1.47328913e-01
1.01218987e+00 -1.36580086e+00 1.04273774e-01 4.13692445e-01
5.82081854e-01 5.60371988e-02 1.32021737e+00 -5.32243431e-01
-5.20880282e-01 -1.06757247e+00 1.24239194e+00 -4.34445888e-01
8.62755537e-01 -2.23017707e-01 -3.91761661e-01 3.96841437e-01
6.03170395e-01 -6.44045413e-01 1.06203616e+00 -2.35695273e-01
-2.84525365e-01 2.86234729e-02 -7.74036765e-01 5.28148592e-01
7.42492318e-01 -9.11884904e-01 -9.34772670e-01 5.38282931e-01
1.08627355e+00 7.93869421e-03 -6.93305373e-01 1.29585490e-01
5.56268871e-01 -5.98982751e-01 8.91010821e-01 -5.22489905e-01
4.41772759e-01 -5.13690293e-01 -4.40035641e-01 -1.24733841e+00
2.32034072e-01 -9.89045382e-01 1.67997748e-01 1.36401105e+00
6.10034764e-01 7.10402802e-02 3.66344899e-01 -2.31802598e-01
-3.50539416e-01 1.42563924e-01 -9.91928279e-01 -7.39966333e-01
-3.72789413e-01 -8.58335316e-01 6.69135332e-01 6.63169742e-01
8.10504183e-02 5.41085958e-01 -2.84349024e-01 8.40300992e-02
2.92447388e-01 -2.58496016e-01 5.87217569e-01 -1.49121511e+00
-5.13434894e-02 -4.97296065e-01 -8.24009895e-01 -2.97115594e-01
9.64551196e-02 -8.03333879e-01 3.43352139e-01 -1.68648803e+00
-7.27569638e-03 7.26170698e-03 -2.49859750e-01 3.15592021e-01
2.42279470e-01 5.89763165e-01 2.53612965e-01 8.34689811e-02
-6.15301669e-01 4.53913182e-01 9.00335312e-01 -1.70128196e-01
-2.53588229e-01 -2.10819617e-01 -1.82991281e-01 7.29436159e-01
7.33220875e-01 -5.96744478e-01 -5.27582884e-01 -2.02295244e-01
4.04559642e-01 1.45461019e-02 5.22141874e-01 -1.55114472e+00
6.44502714e-02 -4.54128068e-03 -2.40125656e-02 -3.68978500e-01
5.30192852e-01 -7.72881746e-01 3.33040923e-01 3.00475150e-01
-7.39487469e-01 3.56854796e-02 1.79318994e-01 4.66489583e-01
-7.79487252e-01 -6.27116680e-01 4.48745281e-01 3.00223846e-02
-5.56480467e-01 -2.33205840e-01 -7.22991645e-01 2.26360545e-01
7.39122927e-01 -2.47858778e-01 -3.61589044e-02 -1.03885508e+00
-1.14350581e+00 -3.62725526e-01 1.12579368e-01 6.54686689e-01
5.83050013e-01 -1.48995161e+00 -1.07012475e+00 -2.32151166e-01
4.61569667e-01 -6.03722513e-01 -9.68015194e-02 4.07001048e-01
-6.97536409e-01 7.88432837e-01 -2.29886919e-01 -5.04097223e-01
-1.35743058e+00 1.24094009e-01 8.30108896e-02 -3.79890064e-03
-3.23678136e-01 6.69668257e-01 7.41235539e-02 1.42052561e-01
5.50926268e-01 -6.14896357e-01 -5.07690430e-01 2.51851946e-01
7.33411729e-01 7.08039626e-02 2.66063586e-03 -7.45047212e-01
-2.33307794e-01 4.14739251e-01 5.59128463e-01 -9.48240638e-01
7.95822144e-01 -2.08757415e-01 9.85308960e-02 8.86319399e-01
9.68483150e-01 2.32010797e-01 -4.88069534e-01 9.08620805e-02
1.91613674e-01 -7.12442324e-02 1.05634779e-01 -9.01067197e-01
-3.44372988e-02 7.97498703e-01 1.05199075e+00 5.02359211e-01
1.04383183e+00 6.68703243e-02 6.05087519e-01 3.72354954e-01
1.56760037e-01 -1.13390589e+00 4.14831936e-01 5.01183391e-01
1.21130753e+00 -8.03634226e-01 -4.60460216e-01 -3.83871913e-01
-7.22255290e-01 1.31251478e+00 1.49903432e-01 4.05550867e-01
5.93552530e-01 1.66137770e-01 2.13290513e-01 1.42487049e-01
-5.59515893e-01 -3.90525371e-01 5.90619147e-01 8.58415544e-01
6.86196566e-01 1.47434145e-01 -2.56007612e-01 6.36861444e-01
-8.62128854e-01 8.97360370e-02 5.96650720e-01 7.12759733e-01
-6.05133593e-01 -1.05078745e+00 -4.37090099e-01 1.14172265e-01
-6.64009988e-01 -3.05146247e-01 -6.99444592e-01 4.36036885e-01
1.93178400e-01 1.23223007e+00 2.01516151e-01 -2.62619346e-01
2.58184165e-01 3.69122952e-01 3.16072673e-01 -9.28070247e-01
-5.42133272e-01 9.85412300e-02 4.17684287e-01 -1.54592246e-01
-7.95024455e-01 -5.33844054e-01 -1.28657806e+00 4.73944277e-01
7.51185557e-03 5.89069605e-01 9.35268760e-01 8.49691570e-01
1.53044134e-01 7.27151394e-01 2.42615610e-01 -7.00890183e-01
-1.46026192e-02 -1.25733042e+00 -3.37763399e-01 6.32873952e-01
3.67543668e-01 -2.44998783e-01 -3.12858641e-01 8.40275705e-01] | [15.336762428283691, 4.831185817718506] |
7d463332-cb81-4d1b-b8f9-36d42d3d9993 | factual-a-benchmark-for-faithful-and | 2305.17497 | null | https://arxiv.org/abs/2305.17497v2 | https://arxiv.org/pdf/2305.17497v2.pdf | FACTUAL: A Benchmark for Faithful and Consistent Textual Scene Graph Parsing | Textual scene graph parsing has become increasingly important in various vision-language applications, including image caption evaluation and image retrieval. However, existing scene graph parsers that convert image captions into scene graphs often suffer from two types of errors. First, the generated scene graphs fail to capture the true semantics of the captions or the corresponding images, resulting in a lack of faithfulness. Second, the generated scene graphs have high inconsistency, with the same semantics represented by different annotations. To address these challenges, we propose a novel dataset, which involves re-annotating the captions in Visual Genome (VG) using a new intermediate representation called FACTUAL-MR. FACTUAL-MR can be directly converted into faithful and consistent scene graph annotations. Our experimental results clearly demonstrate that the parser trained on our dataset outperforms existing approaches in terms of faithfulness and consistency. This improvement leads to a significant performance boost in both image caption evaluation and zero-shot image retrieval tasks. Furthermore, we introduce a novel metric for measuring scene graph similarity, which, when combined with the improved scene graph parser, achieves state-of-the-art (SOTA) results on multiple benchmark datasets for the aforementioned tasks. The code and dataset are available at https://github.com/zhuang-li/FACTUAL . | ['Terry Yue Zhuo', 'Quan Hung Tran', 'Donghong Ji', 'Fei Li', 'Gholamreza Haffari', 'Lizhen Qu', 'Yuyang Chai', 'Zhuang Li'] | 2023-05-27 | null | null | null | null | ['image-captioning', 'graph-similarity'] | ['computer-vision', 'graphs'] | [ 5.37250876e-01 2.10404575e-01 -1.22292139e-01 -5.64519763e-01
-1.03709209e+00 -7.20468700e-01 5.60749829e-01 3.30741554e-01
6.12187723e-04 4.17471170e-01 2.73014635e-01 -1.00054115e-01
3.62690955e-01 -7.51752496e-01 -1.12065530e+00 -4.40358996e-01
5.56471825e-01 3.97866338e-01 4.61322874e-01 -6.62939772e-02
4.71336618e-02 -1.56294703e-01 -1.46084130e+00 4.70676184e-01
8.59824657e-01 7.85735786e-01 4.90866572e-01 5.41102648e-01
-4.13672805e-01 1.05396831e+00 -3.86004329e-01 -9.09924209e-01
-3.81451212e-02 -6.34360909e-01 -8.46687019e-01 1.60312518e-01
8.84927332e-01 -3.07096034e-01 -4.58752453e-01 1.46096766e+00
2.21681356e-01 -1.35327667e-01 2.58955479e-01 -1.51308393e+00
-1.15346336e+00 5.70234835e-01 -5.55973113e-01 -8.15246776e-02
4.83610511e-01 1.38872415e-01 1.16265142e+00 -7.24146366e-01
8.41956258e-01 1.30634630e+00 3.26673657e-01 5.64261198e-01
-9.83154058e-01 -3.01485598e-01 8.09015185e-02 3.56308222e-01
-1.29544830e+00 -3.27090532e-01 7.33712971e-01 -4.94312286e-01
6.99024081e-01 3.20775479e-01 3.70527178e-01 9.22999263e-01
-8.86472389e-02 6.02429092e-01 7.87603140e-01 -3.20403129e-01
9.93408263e-02 -8.54866207e-02 1.82851315e-01 9.33091283e-01
3.94175529e-01 -3.68638754e-01 -2.84325033e-01 8.77688229e-02
5.29036224e-01 -9.62893590e-02 -4.17214304e-01 -5.75514555e-01
-1.29863870e+00 7.73304343e-01 7.45104253e-01 1.48695409e-01
-1.34405077e-01 4.45490688e-01 6.10483944e-01 -2.20536068e-01
2.49814719e-01 2.93980151e-01 1.15654826e-01 1.19906448e-01
-4.14377034e-01 2.13824529e-02 4.26120251e-01 1.28398633e+00
6.65608227e-01 -1.32979408e-01 -5.00362754e-01 7.35900283e-01
3.61035854e-01 7.46718645e-01 3.03570628e-01 -7.98996747e-01
6.40697598e-01 6.70586944e-01 -2.31514290e-01 -1.37008703e+00
6.54983297e-02 -1.49391234e-01 -7.62462258e-01 -4.86122280e-01
2.18774095e-01 5.44120073e-01 -1.11063826e+00 1.92397940e+00
3.76443803e-01 4.59674835e-01 3.12523305e-01 1.05749416e+00
1.53779757e+00 8.18398833e-01 4.37990040e-01 4.84317504e-02
1.68804586e+00 -1.34878957e+00 -8.44326377e-01 -6.85243487e-01
5.33598781e-01 -8.05505991e-01 1.40809548e+00 -1.86997950e-01
-8.95491540e-01 -6.51187778e-01 -9.88607466e-01 -5.12229800e-01
-2.53894508e-01 2.81156730e-02 5.58611095e-01 3.07958186e-01
-1.04610729e+00 6.53226152e-02 -5.80527067e-01 -5.07484376e-01
5.43064415e-01 -2.21470714e-01 -5.17309189e-01 -5.36208153e-01
-1.04285467e+00 5.39127767e-01 6.09588206e-01 3.49088758e-02
-8.71647894e-01 -4.89571095e-01 -1.30654788e+00 7.83490911e-02
4.04388458e-01 -7.75112748e-01 1.18411505e+00 -9.24680650e-01
-8.92575264e-01 1.28457522e+00 -2.14352757e-01 -3.25961739e-01
3.01166773e-01 7.04867998e-04 -2.71664500e-01 3.41849327e-01
4.89549786e-01 8.35395813e-01 4.61408973e-01 -1.42079532e+00
-2.59417266e-01 -2.40397856e-01 2.00200394e-01 2.52857506e-01
-7.81875625e-02 -8.02613571e-02 -1.16628039e+00 -4.93509531e-01
-6.50730915e-03 -8.84247601e-01 -9.48912352e-02 -3.39901112e-02
-7.44999707e-01 -1.06518775e-01 6.21553779e-01 -6.47938609e-01
9.47115302e-01 -2.22336364e+00 1.00514226e-01 -2.40188196e-01
2.27701783e-01 4.09208119e-01 -4.74023104e-01 3.71894181e-01
-6.65947869e-02 1.89494208e-01 -5.96906662e-01 -3.59730750e-01
-1.53866455e-01 3.68355751e-01 -5.69426835e-01 2.03396246e-01
2.79175580e-01 1.38080263e+00 -1.27806091e+00 -7.80245125e-01
3.94256473e-01 4.22305018e-01 -2.23157167e-01 5.33500552e-01
-4.41183776e-01 3.56708467e-01 -5.35913467e-01 5.10482669e-01
7.22020864e-01 -7.58323550e-01 2.68457264e-01 -4.77288693e-01
3.69526565e-01 -2.79513393e-02 -6.87278569e-01 1.95017195e+00
-4.30074669e-02 6.63979411e-01 -6.07446313e-01 -8.85139167e-01
8.34094226e-01 6.40425365e-04 1.42552629e-01 -9.12798584e-01
8.55090842e-03 1.20928578e-01 -4.99833465e-01 -6.35256052e-01
5.30122876e-01 1.90153390e-01 -3.51165771e-01 4.77283150e-02
4.11821641e-02 -3.74817550e-01 4.57126498e-01 7.24567056e-01
8.74996006e-01 2.18510330e-01 2.97497541e-01 6.61592782e-02
5.15399158e-01 3.25486839e-01 5.21937013e-01 6.10463679e-01
-2.04922989e-01 9.64058757e-01 7.42632926e-01 -2.67807424e-01
-1.18344700e+00 -1.18112099e+00 2.43505150e-01 8.02881956e-01
7.23571360e-01 -5.98690987e-01 -9.59703743e-01 -6.13268852e-01
-3.66300076e-01 7.98630059e-01 -4.80634868e-01 -3.30135614e-01
-3.58325124e-01 -4.16425198e-01 6.31832361e-01 3.98940802e-01
7.26351142e-01 -1.07253397e+00 -4.24996614e-01 -1.20487653e-01
-6.09066010e-01 -1.79442084e+00 -7.24487901e-01 -4.89812106e-01
-4.33228850e-01 -1.18396962e+00 -5.37245333e-01 -1.07708967e+00
8.96628439e-01 6.69169128e-01 1.37533700e+00 2.50436336e-01
-3.17541748e-01 4.84410077e-01 -5.59660435e-01 -2.05200419e-01
-6.83803916e-01 -3.78796577e-01 -4.72572982e-01 -2.37873085e-02
1.60198316e-01 8.95190332e-03 -4.97562468e-01 9.61133912e-02
-1.23936653e+00 6.82104528e-01 4.10026848e-01 6.69583201e-01
1.09636509e+00 -3.47364604e-01 2.83726662e-01 -1.22214437e+00
3.31136137e-01 -4.33315903e-01 -7.68534482e-01 7.51758218e-01
-2.76844919e-01 1.05809487e-01 5.87998867e-01 -2.94955689e-02
-1.02343249e+00 2.39871308e-01 2.01800950e-02 -5.47061563e-01
-5.54009341e-02 5.09098947e-01 -2.14830190e-01 7.71934092e-02
4.49568391e-01 3.15123737e-01 -1.58835486e-01 -7.12537616e-02
7.80481696e-01 5.49781799e-01 1.05232036e+00 -2.86670893e-01
7.11532652e-01 4.02523547e-01 -2.36326493e-02 -5.58399022e-01
-1.22237301e+00 -5.54509640e-01 -2.76067615e-01 -2.54083067e-01
1.25427127e+00 -1.15868890e+00 -2.10265398e-01 5.01173079e-01
-1.46857560e+00 -1.90314442e-01 -1.17370345e-01 1.51816219e-01
-6.06405318e-01 7.76877582e-01 -4.05287951e-01 -4.11817908e-01
-5.49360752e-01 -1.34812236e+00 1.51146519e+00 4.36246783e-01
2.35137478e-01 -7.72975028e-01 -2.96751373e-02 6.69518173e-01
8.08179900e-02 5.58214366e-01 1.12425411e+00 -3.45633209e-01
-7.61715531e-01 -6.37539029e-02 -8.00778627e-01 7.34561384e-02
9.86153111e-02 -7.54154772e-02 -8.79645526e-01 -2.22574458e-01
-4.40728277e-01 -5.15983343e-01 6.50681496e-01 1.46767572e-01
1.10541117e+00 -1.65648729e-01 -1.94820642e-01 6.01549745e-01
1.75616872e+00 -9.00785625e-02 8.72855425e-01 2.01543003e-01
1.20575964e+00 5.47081828e-01 7.16952205e-01 3.59575637e-02
6.41538799e-01 8.80272627e-01 6.21335387e-01 -2.80363262e-01
-7.03408241e-01 -7.70648837e-01 2.13100597e-01 8.98938239e-01
3.53856117e-01 -6.78969562e-01 -9.64868903e-01 6.92762911e-01
-2.20868635e+00 -7.79692471e-01 -4.74996179e-01 2.11698318e+00
5.94541967e-01 -1.57765806e-01 -2.55694211e-01 -4.82111901e-01
1.09040093e+00 2.97994554e-01 -5.88157296e-01 -1.09039135e-01
-3.53100508e-01 -3.25960398e-01 5.15241027e-01 4.06599432e-01
-1.01559079e+00 1.31592417e+00 5.30808640e+00 6.20018542e-01
-1.02857745e+00 2.56950289e-01 7.31610000e-01 4.03600037e-01
-4.74377632e-01 1.83172166e-01 -4.58697140e-01 4.90547895e-01
6.21327579e-01 -4.17198330e-01 3.11299652e-01 9.58316326e-01
-1.20559767e-01 1.11963011e-01 -9.86384332e-01 1.32272696e+00
5.44731200e-01 -1.46880651e+00 6.06466234e-01 -2.76026607e-01
7.13377178e-01 1.53990045e-01 -1.39292374e-01 8.92760605e-02
1.95298985e-01 -8.46142173e-01 8.55114102e-01 3.38303000e-01
9.86318767e-01 -2.72832572e-01 8.91475141e-01 -1.39029622e-01
-1.35013545e+00 4.23419684e-01 -5.42141974e-01 3.44895571e-01
4.53076363e-01 4.41316426e-01 -8.69115591e-01 8.45635533e-01
6.50737286e-01 8.45308781e-01 -9.32755828e-01 9.37596381e-01
-5.76579452e-01 3.82491738e-01 1.22947320e-01 1.27366623e-02
1.91147208e-01 -1.22886114e-01 4.28476930e-01 1.04063451e+00
2.27635190e-01 3.26311849e-02 1.68957815e-01 1.00102878e+00
-3.63217443e-01 2.51235753e-01 -8.73006999e-01 -3.95563871e-01
4.43799645e-01 1.24138534e+00 -9.02262628e-01 -5.20048738e-01
-4.75256324e-01 1.21792924e+00 4.80373025e-01 3.71789217e-01
-1.13323486e+00 -2.76616424e-01 3.82665694e-01 -1.12730287e-01
1.20172165e-01 -4.43912745e-02 7.13782012e-02 -1.35357201e+00
2.86338717e-01 -7.23783314e-01 5.23004055e-01 -1.32896173e+00
-1.28127468e+00 8.39649200e-01 8.76352265e-02 -1.13169992e+00
-1.29913062e-01 -4.79830176e-01 -4.27245229e-01 2.75013387e-01
-1.55205369e+00 -1.42969120e+00 -9.09019887e-01 4.64654505e-01
6.09976113e-01 1.25022203e-01 6.83108449e-01 3.58479261e-01
-5.64354181e-01 5.17982483e-01 -2.30492398e-01 2.42208585e-01
5.95032871e-01 -1.09620988e+00 7.27020800e-01 1.22048092e+00
5.35732687e-01 2.24720448e-01 6.80731416e-01 -6.12351000e-01
-1.51329994e+00 -1.60104752e+00 8.15944850e-01 -4.30696547e-01
4.87644941e-01 -4.64774370e-01 -1.00994205e+00 6.44204736e-01
2.29468778e-01 1.84783474e-01 4.38062757e-01 -5.03713489e-01
-6.26900375e-01 1.69772029e-01 -9.02918756e-01 6.36498153e-01
1.25467134e+00 -6.07046843e-01 -3.83458734e-01 7.27067649e-01
1.22832155e+00 -6.29000545e-01 -6.00558817e-01 3.95746887e-01
1.32068336e-01 -8.36824119e-01 9.50454414e-01 -5.51357329e-01
6.90969110e-01 -6.02060199e-01 -3.88050109e-01 -9.08918738e-01
-2.29403779e-01 -2.21868411e-01 2.64087379e-01 1.44744396e+00
2.01907665e-01 -4.34890270e-01 5.27323186e-01 5.22083163e-01
-2.67085344e-01 -3.76941174e-01 -7.18229353e-01 -7.55557656e-01
-4.20166433e-01 -2.73433894e-01 6.51746511e-01 9.94402826e-01
-2.79724479e-01 6.66110814e-01 -3.97312015e-01 2.35469639e-01
8.14072728e-01 4.17198300e-01 8.45733643e-01 -8.10048103e-01
-2.11717248e-01 -2.13246837e-01 -7.53926694e-01 -8.84006083e-01
1.97262168e-01 -1.14124501e+00 4.00169104e-01 -2.04682612e+00
7.25415528e-01 -2.75335491e-01 -6.32709116e-02 6.17053807e-01
-4.90194768e-01 5.87075889e-01 3.59945536e-01 2.75166214e-01
-1.09818363e+00 5.70009410e-01 1.23232019e+00 -2.89417177e-01
1.48921728e-01 -7.34919071e-01 -7.37184763e-01 5.86895764e-01
6.38763547e-01 -4.44030404e-01 -6.29766643e-01 -9.38988805e-01
2.58525431e-01 -1.51301753e-02 6.43262267e-01 -8.36722970e-01
8.61972421e-02 -1.98417798e-01 -1.18351184e-01 -2.34841883e-01
2.05629990e-01 -4.29091722e-01 3.94812822e-01 3.99118870e-01
-2.77431935e-01 8.38154331e-02 2.68628925e-01 7.95447767e-01
-3.90165746e-01 -2.11982623e-01 6.93562448e-01 -2.76649576e-02
-1.20278418e+00 3.68580967e-01 2.79897273e-01 3.27738971e-01
1.17187405e+00 5.24953939e-02 -8.91808331e-01 -4.59741592e-01
-2.38934949e-01 3.14965785e-01 7.77597964e-01 7.56909132e-01
7.65450895e-01 -1.40740144e+00 -8.20341349e-01 -1.12525530e-01
8.22805047e-01 1.64843380e-01 4.30325329e-01 4.72217321e-01
-8.40415299e-01 4.38423127e-01 -2.01910853e-01 -8.47561121e-01
-1.51681244e+00 6.67366743e-01 1.10735528e-01 -8.82716626e-02
-5.78683317e-01 7.56245375e-01 7.24790990e-01 -2.37513512e-01
-1.17533123e-02 -2.63160050e-01 -6.33476898e-02 -4.50024277e-01
4.05886382e-01 -2.13218242e-01 -1.68327630e-01 -1.10408545e+00
-4.29384768e-01 7.87698686e-01 5.21660075e-02 2.81250536e-01
9.96227026e-01 -1.83720022e-01 -2.76359409e-01 1.01035833e-01
1.20068312e+00 -2.24192590e-01 -9.56940830e-01 -1.96588352e-01
-8.57871547e-02 -6.20132208e-01 -8.01889822e-02 -4.75343227e-01
-1.08295071e+00 7.94964969e-01 5.74747741e-01 9.80580375e-02
1.13331044e+00 5.17945468e-01 9.23989952e-01 2.85396963e-01
3.01931769e-01 -5.38958013e-01 3.21204394e-01 2.83693492e-01
8.86326492e-01 -1.47103131e+00 -1.11433499e-01 -1.06476665e+00
-9.65109169e-01 6.79791629e-01 7.08853304e-01 1.54331461e-01
-1.11879013e-01 -2.55588949e-01 9.89496931e-02 -3.84398431e-01
-4.79698330e-01 -4.82537329e-01 4.28622991e-01 6.08788669e-01
3.66475552e-01 1.12562418e-01 -1.94067731e-01 3.66597384e-01
-8.87340829e-02 -1.96357578e-01 5.62648714e-01 6.76241219e-01
-2.89928854e-01 -9.35144842e-01 -2.00355356e-03 1.85231149e-01
-2.08665431e-01 -3.18698019e-01 -5.96001565e-01 5.08918226e-01
-1.41119033e-01 1.00361359e+00 1.50522783e-01 -2.21554443e-01
3.17078054e-01 -2.85234243e-01 4.57473010e-01 -7.88634479e-01
6.83504045e-02 -2.09077597e-01 1.24058597e-01 -7.17256486e-01
-4.54602689e-01 -1.94697157e-01 -1.52445221e+00 3.83574627e-02
-1.09834775e-01 1.27585083e-01 5.32035351e-01 6.99089229e-01
5.79225183e-01 5.71141183e-01 3.76023084e-01 -2.56069005e-01
4.39598523e-02 -6.19209766e-01 -8.50168541e-02 1.08943868e+00
-5.51287420e-02 -4.51022327e-01 -7.62213916e-02 4.86332864e-01] | [10.499361991882324, 1.5132032632827759] |
dd7401de-56ae-4eae-b4bf-1aeefb548903 | deep-rgb-d-saliency-detection-with-depth | 2103.11832 | null | https://arxiv.org/abs/2103.11832v1 | https://arxiv.org/pdf/2103.11832v1.pdf | Deep RGB-D Saliency Detection with Depth-Sensitive Attention and Automatic Multi-Modal Fusion | RGB-D salient object detection (SOD) is usually formulated as a problem of classification or regression over two modalities, i.e., RGB and depth. Hence, effective RGBD feature modeling and multi-modal feature fusion both play a vital role in RGB-D SOD. In this paper, we propose a depth-sensitive RGB feature modeling scheme using the depth-wise geometric prior of salient objects. In principle, the feature modeling scheme is carried out in a depth-sensitive attention module, which leads to the RGB feature enhancement as well as the background distraction reduction by capturing the depth geometry prior. Moreover, to perform effective multi-modal feature fusion, we further present an automatic architecture search approach for RGB-D SOD, which does well in finding out a feasible architecture from our specially designed multi-modal multi-scale search space. Extensive experiments on seven standard benchmarks demonstrate the effectiveness of the proposed approach against the state-of-the-art. | ['Xi Li', 'Songyuan Li', 'Huanyu Wang', 'Wenhu Zhang', 'Peng Sun'] | 2021-03-22 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Sun_Deep_RGB-D_Saliency_Detection_With_Depth-Sensitive_Attention_and_Automatic_Multi-Modal_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Sun_Deep_RGB-D_Saliency_Detection_With_Depth-Sensitive_Attention_and_Automatic_Multi-Modal_CVPR_2021_paper.pdf | cvpr-2021-1 | ['rgb-d-salient-object-detection'] | ['computer-vision'] | [ 1.53497830e-01 3.71346064e-02 8.88197124e-02 -1.38836980e-01
-1.04332149e+00 -1.12687133e-01 2.97842711e-01 1.80030428e-02
-3.17463964e-01 3.27842563e-01 3.24393898e-01 6.57324269e-02
-2.30338916e-01 -7.59119034e-01 -4.95945364e-01 -9.95889187e-01
5.31977892e-01 -2.79532629e-03 6.03887856e-01 -4.60848302e-01
3.42749447e-01 7.39129543e-01 -1.88499320e+00 7.86888227e-02
7.37027109e-01 1.53965414e+00 5.89906037e-01 4.34677273e-01
-2.71560669e-01 7.61956215e-01 -2.75068611e-01 -2.30380014e-01
4.40100431e-01 -1.96192563e-01 -7.08113194e-01 4.81507331e-01
1.95147917e-01 -2.86069036e-01 -3.92750174e-01 1.00533640e+00
6.60159767e-01 2.29764774e-01 2.91338682e-01 -1.22557104e+00
-3.29956651e-01 -5.74252121e-02 -9.68197823e-01 4.32583690e-01
2.94987321e-01 1.72911406e-01 1.01461995e+00 -1.02478158e+00
3.05922598e-01 1.15054798e+00 2.82786459e-01 3.30063999e-01
-7.53959000e-01 -2.50239909e-01 1.64541006e-01 3.38064641e-01
-1.18872929e+00 -1.20659389e-01 1.66456807e+00 -7.84247741e-02
7.32978404e-01 5.92462599e-01 9.24555600e-01 4.55155224e-01
7.59049878e-02 1.13463771e+00 1.20182526e+00 -5.45498013e-01
1.86525539e-01 1.69099808e-01 -1.00503169e-01 9.26225960e-01
1.12546861e-01 -4.08010818e-02 -7.05708563e-01 -5.64991171e-03
9.65548694e-01 3.28652024e-01 -3.38101774e-01 -6.41328514e-01
-9.70874250e-01 6.97962940e-01 8.83526623e-01 2.31169179e-01
-5.88444173e-01 2.17948675e-01 6.50967732e-02 -3.49543661e-01
5.44324338e-01 2.53574044e-01 -4.76237386e-01 1.98699519e-01
-7.18454182e-01 -6.03589453e-02 -1.20210527e-02 8.06230783e-01
9.55789864e-01 -1.71612322e-01 -2.27690578e-01 4.91478652e-01
6.09602213e-01 5.02862453e-01 4.92194176e-01 -9.62763309e-01
4.68680978e-01 1.15572369e+00 1.35197759e-01 -1.14215195e+00
-6.63816988e-01 -2.76414424e-01 -6.37758434e-01 1.94259331e-01
2.81730980e-01 2.86280960e-01 -7.07430542e-01 1.46375811e+00
8.72206390e-01 -4.08672504e-02 -5.28195985e-02 1.20485389e+00
9.87728834e-01 3.51278067e-01 -7.18897507e-02 -2.43751124e-01
1.55083978e+00 -9.01752472e-01 -5.88684499e-01 -4.01440144e-01
3.25034410e-01 -6.41806066e-01 1.14843237e+00 2.21220121e-01
-1.23454809e+00 -6.23271823e-01 -9.26852465e-01 -5.40565491e-01
-3.95237565e-01 3.05533081e-01 1.03417838e+00 6.67654991e-01
-9.45210934e-01 1.89315632e-01 -9.29134846e-01 -2.38125086e-01
4.69318956e-01 4.01404977e-01 -2.70172358e-01 -1.01963021e-01
-9.63006735e-01 7.44643509e-01 2.76345015e-01 3.82640690e-01
-5.45846760e-01 -4.30049986e-01 -7.85409272e-01 -1.21577024e-01
4.99023646e-01 -8.17412853e-01 8.54629695e-01 -6.36396646e-01
-1.29907036e+00 9.45554316e-01 -2.67096728e-01 1.18324004e-01
2.66739964e-01 -2.25677177e-01 1.07584007e-01 5.73433161e-01
-3.46040092e-02 4.95901227e-01 8.53351831e-01 -1.45619833e+00
-9.67401683e-01 -8.63050163e-01 3.94771487e-01 6.22218788e-01
-6.91990554e-01 -7.41376653e-02 -7.41045296e-01 -3.73471826e-01
6.50281608e-01 -6.58005059e-01 -3.51125479e-01 2.70895034e-01
-4.36924726e-01 -2.49134481e-01 6.62850082e-01 -7.78391063e-01
9.65714753e-01 -2.23583770e+00 3.42944622e-01 7.43324161e-02
2.67672807e-01 -2.03226164e-01 3.46245259e-01 -2.30318323e-01
7.22555816e-02 -1.86504081e-01 -9.17204916e-02 -5.89307308e-01
-1.10784017e-01 2.02681613e-03 -5.40114641e-02 7.32803345e-01
3.52764457e-01 9.26592290e-01 -7.74460793e-01 -7.45030582e-01
5.18985569e-01 5.08438349e-01 -4.75176483e-01 1.38615668e-01
-6.96826801e-02 3.63429308e-01 -1.00140369e+00 1.23375964e+00
7.44081199e-01 -2.33180001e-01 -4.55590963e-01 -5.84126711e-01
-2.40764096e-01 -8.58586803e-02 -1.09871519e+00 2.11742878e+00
-4.39881921e-01 2.99025625e-01 -2.53529307e-02 -7.30918348e-01
9.60511923e-01 -1.36388376e-01 7.57590950e-01 -9.53561962e-01
4.23422813e-01 1.63351312e-01 -5.80056846e-01 -3.79542500e-01
7.30488598e-01 4.71416973e-02 -1.59290865e-01 2.35067233e-01
-3.17969412e-01 -2.84443051e-01 -3.17917436e-01 -4.69730683e-02
8.94602776e-01 7.52878934e-02 2.83157289e-01 -1.93138998e-02
7.51291573e-01 1.20292366e-01 3.73968124e-01 4.40974146e-01
-3.57859761e-01 5.91175616e-01 7.64058977e-02 -3.22925299e-01
-6.63101971e-01 -8.02331030e-01 1.10817410e-01 8.34251106e-01
8.66139650e-01 -8.79809931e-02 -4.35635656e-01 -6.23134553e-01
-2.12778851e-01 3.17178369e-01 -7.49687195e-01 -3.80119950e-01
-4.43163604e-01 -7.89769471e-01 -5.94643168e-02 6.48576617e-01
8.05777490e-01 -8.76700699e-01 -1.42058659e+00 1.75968766e-01
-2.96845943e-01 -1.05615222e+00 -2.21487999e-01 6.38768017e-01
-1.00961912e+00 -1.01208270e+00 -7.92111874e-01 -7.17746675e-01
4.87648338e-01 8.68859112e-01 8.20583045e-01 9.80025828e-02
-6.20009542e-01 5.98801374e-01 -5.26779234e-01 -2.82568932e-01
3.25596094e-01 -7.61248991e-02 -2.77071029e-01 4.46586451e-03
1.63428009e-01 -3.41141105e-01 -1.02486312e+00 2.68714339e-01
-9.33285713e-01 2.30370998e-01 9.78207171e-01 4.86336350e-01
9.59855080e-01 3.07217866e-01 -4.86054504e-03 -1.03298903e-01
-3.33868153e-03 -7.61062056e-02 -5.19856215e-01 3.57256532e-01
-1.12955213e-01 -1.24283805e-02 6.63761720e-02 -1.49660453e-01
-1.23962176e+00 5.42241096e-01 -4.59801517e-02 -6.60797417e-01
-9.19116810e-02 8.48970115e-02 -5.85550606e-01 -3.79583597e-01
3.15037042e-01 5.88326991e-01 -2.76131362e-01 -4.62975711e-01
3.99487823e-01 5.73310971e-01 2.08458140e-01 -3.77114534e-01
8.28225195e-01 9.70532894e-01 2.80368388e-01 -7.79865324e-01
-8.59145641e-01 -7.72707701e-01 -6.27849400e-01 -2.87740976e-01
9.64253426e-01 -1.04163027e+00 -6.80459321e-01 4.82564688e-01
-1.08144379e+00 -8.10444355e-04 -3.41332346e-01 2.46425524e-01
-6.29449248e-01 2.65990496e-01 -3.49364668e-01 -1.07971561e+00
-4.05687243e-01 -1.32653749e+00 1.77451861e+00 4.96733814e-01
3.86857420e-01 -6.61020517e-01 -2.54084200e-01 6.16680145e-01
1.38545677e-01 2.89029390e-01 7.74755180e-01 -8.50281641e-02
-9.67219651e-01 -1.40354276e-01 -5.31502962e-01 -5.31789847e-02
1.11056842e-01 -3.67869705e-01 -1.13236785e+00 6.27539828e-02
2.38084450e-01 -2.12839127e-01 8.18347156e-01 5.13033092e-01
1.27980220e+00 1.45037174e-01 -3.21298987e-01 7.69671142e-01
1.75526488e+00 1.40119540e-02 5.86508870e-01 5.35654962e-01
9.72539127e-01 4.46748108e-01 1.19919562e+00 8.01476777e-01
5.96790135e-01 8.50416720e-01 1.01542699e+00 -5.43272138e-01
-1.33801296e-01 9.44607481e-02 4.94934916e-02 3.48415732e-01
-1.05714269e-01 -1.39655396e-02 -7.47000635e-01 4.24844265e-01
-1.70916522e+00 -6.30542040e-01 -2.45937034e-01 1.96460629e+00
6.36104107e-01 9.75285769e-02 2.96594471e-01 4.05103534e-01
5.97351730e-01 1.80422142e-02 -6.00948036e-01 1.09455734e-01
-3.95990938e-01 7.02451244e-02 5.82935631e-01 8.66064429e-02
-1.18974602e+00 6.45353734e-01 4.57434940e+00 9.98689473e-01
-1.10044050e+00 1.61446929e-01 7.16841519e-01 -1.93925261e-01
-3.81475359e-01 -2.21258923e-01 -9.38104808e-01 2.46808663e-01
1.67524144e-01 1.52903602e-01 8.85491371e-02 9.89502251e-01
1.28077075e-01 -5.16609490e-01 -6.76395237e-01 1.34280920e+00
9.61125121e-02 -1.05948436e+00 -1.79780707e-01 3.08410048e-01
5.79355180e-01 -2.22320601e-01 4.87963296e-02 -1.86594110e-02
-2.35107139e-01 -3.94238561e-01 1.02533400e+00 5.81411600e-01
3.34230691e-01 -9.29504812e-01 5.34105361e-01 2.25855649e-01
-1.44800842e+00 -3.48983198e-01 -4.78974819e-01 3.31060201e-01
8.05960223e-02 6.73675358e-01 -1.29603967e-01 7.85196126e-01
1.01957703e+00 6.84690297e-01 -8.89664173e-01 1.11682951e+00
-1.19058095e-01 -5.17797209e-02 -4.45787102e-01 -3.85108031e-02
1.12727188e-01 1.52071893e-01 3.85261685e-01 7.67051220e-01
3.18864852e-01 3.42005163e-01 5.71445711e-02 6.78705513e-01
2.75237262e-01 1.47649184e-01 -1.85696438e-01 3.75546724e-01
1.65776655e-01 1.49275482e+00 -1.09052372e+00 6.92012981e-02
-4.80336428e-01 1.20623243e+00 2.70107418e-01 1.04549967e-01
-1.02721536e+00 -8.45311210e-02 5.33174217e-01 -2.86083925e-03
5.42289913e-01 -6.87157512e-02 -5.07282019e-01 -9.62756693e-01
1.51671186e-01 -4.97934937e-01 3.40356946e-01 -1.13443518e+00
-1.01164746e+00 5.82426727e-01 -2.47659728e-01 -1.45904791e+00
2.76552081e-01 -6.09369814e-01 -3.28221411e-01 8.45024049e-01
-2.04432535e+00 -1.45009267e+00 -7.65371382e-01 1.05451465e+00
5.26795030e-01 2.64113307e-01 3.94101709e-01 2.00081274e-01
-5.05806863e-01 3.05017442e-01 -2.39621639e-01 -3.03418845e-01
2.48586223e-01 -1.09043956e+00 -2.44846523e-01 8.57979774e-01
-1.29045561e-01 2.65385479e-01 4.42653805e-01 -4.36665297e-01
-1.81704581e+00 -9.35714066e-01 2.57498860e-01 -2.55599618e-01
4.23157305e-01 -1.88375399e-01 -4.90679413e-01 1.00775599e-01
-2.56763071e-01 3.13411981e-01 4.57737356e-01 -4.16530460e-01
1.02913640e-01 -3.05806369e-01 -1.12355757e+00 3.91036779e-01
1.09461510e+00 -5.50380290e-01 -4.96846676e-01 2.35616609e-01
8.70335400e-01 -5.08319795e-01 -7.51447976e-01 4.09328550e-01
4.22133237e-01 -1.16464841e+00 1.48520029e+00 -1.06311418e-01
5.51895618e-01 -3.96749139e-01 -5.77280045e-01 -8.30908358e-01
-1.83543846e-01 -1.95590377e-01 -4.25730616e-01 1.30680227e+00
-1.44623503e-01 -1.88233182e-01 9.52632606e-01 7.17830658e-01
-2.49344900e-01 -9.62835133e-01 -1.08365119e+00 -3.63744706e-01
-5.85605800e-01 -5.52654862e-01 5.18077612e-01 5.17743409e-01
-3.06455880e-01 -1.06561065e-01 -1.29395097e-01 4.40760523e-01
7.32487440e-01 6.06888950e-01 5.68544388e-01 -1.16357327e+00
-3.07552308e-01 -5.23860157e-01 -6.65727794e-01 -1.07155669e+00
-2.09047869e-01 -4.25962448e-01 1.11945920e-01 -1.53454435e+00
2.33148292e-01 -5.06056309e-01 -4.77219582e-01 4.04406279e-01
-4.86679882e-01 5.06536663e-01 7.67583251e-02 6.66324645e-02
-7.50129998e-01 9.49219525e-01 1.43058288e+00 -1.25035658e-01
-3.55520487e-01 -1.10855587e-01 -9.57037449e-01 6.80596769e-01
6.94804013e-01 -2.74508476e-01 -3.33492130e-01 -3.04299206e-01
2.20650867e-01 2.07588837e-01 6.90822184e-01 -1.03126550e+00
2.25211576e-01 -2.68548787e-01 5.86017430e-01 -9.25511539e-01
8.94173920e-01 -1.10938311e+00 -3.62057686e-01 3.15181971e-01
1.35044158e-01 -2.11526945e-01 2.26181135e-01 6.24477327e-01
-2.21153736e-01 1.47863388e-01 7.78693497e-01 -3.03621650e-01
-1.24097002e+00 1.97688162e-01 1.67914391e-01 -1.94006726e-01
1.10429370e+00 -5.74532449e-01 -1.92030236e-01 -5.38590690e-03
-5.40937722e-01 4.51864116e-02 4.95238066e-01 3.01755458e-01
9.67690408e-01 -1.25873637e+00 -1.95462540e-01 2.48818442e-01
3.41109991e-01 2.49639034e-01 4.93844599e-01 1.07859099e+00
-3.39869410e-01 3.86225730e-01 -1.84846714e-01 -6.93054616e-01
-1.19288719e+00 6.33584857e-01 2.81186640e-01 -2.39455938e-01
-3.70771080e-01 1.16527998e+00 2.81707525e-01 2.51470450e-02
2.73096770e-01 -4.51048493e-01 -1.59801915e-01 1.26850352e-01
4.40813988e-01 2.62519866e-01 2.05115721e-01 -8.23891938e-01
-7.61725903e-01 9.78265345e-01 3.45705241e-01 1.13627113e-01
1.43441355e+00 -7.10866928e-01 3.77611592e-02 2.15011403e-01
1.11655617e+00 -3.86871286e-02 -1.37075555e+00 -1.97746396e-01
-2.53840536e-01 -6.87255561e-01 6.27180815e-01 -4.30986434e-01
-1.31472147e+00 9.11239147e-01 1.00047505e+00 6.88549802e-02
1.68461609e+00 3.10816139e-01 7.77206242e-01 1.24672025e-01
5.60389578e-01 -8.75021636e-01 4.67684031e-01 -3.77246775e-02
8.99008811e-01 -1.56589520e+00 2.62112707e-01 -6.25323176e-01
-5.24158001e-01 9.13231373e-01 7.34077871e-01 -1.69760063e-02
5.83255112e-01 -3.71985584e-02 -2.20434144e-01 -3.51043463e-01
-3.02130908e-01 -7.44806945e-01 4.95521396e-01 4.03779745e-01
3.63957919e-02 -3.05260271e-01 1.15705281e-01 7.10716426e-01
1.15253352e-01 -2.04359949e-01 1.00478537e-01 1.03132105e+00
-6.04219735e-01 -7.07443535e-01 -6.95722640e-01 1.33011043e-01
-7.90886059e-02 -3.79666686e-02 -3.19865733e-01 8.64663184e-01
3.45573932e-01 9.61311817e-01 -1.64763987e-01 -4.23768729e-01
3.40090483e-01 -1.75860479e-01 6.91680431e-01 -3.02699596e-01
-4.52255517e-01 3.37350965e-01 -3.48404408e-01 -8.73982608e-01
-7.29916036e-01 -7.10203886e-01 -1.25234878e+00 1.07122302e-01
-5.71216822e-01 -4.07538176e-01 8.80166113e-01 1.01698411e+00
1.32861093e-01 6.29894137e-01 8.40991020e-01 -1.20733249e+00
-9.02425423e-02 -5.93723774e-01 -7.14345753e-01 2.46183634e-01
3.98488671e-01 -9.71685648e-01 -3.41335267e-01 -2.92741448e-01] | [9.69139575958252, -0.8251104950904846] |
437d4d72-0c9c-4948-b27a-0c951c03fe0a | improving-dialogue-act-classification-for | 1806.00522 | null | http://arxiv.org/abs/1806.00522v1 | http://arxiv.org/pdf/1806.00522v1.pdf | Improving Dialogue Act Classification for Spontaneous Arabic Speech and Instant Messages at Utterance Level | The ability to model and automatically detect dialogue act is an important
step toward understanding spontaneous speech and Instant Messages. However, it
has been difficult to infer a dialogue act from a surface utterance because it
highly depends on the context of the utterance and speaker linguistic
knowledge; especially in Arabic dialects. This paper proposes a statistical
dialogue analysis model to recognize utterance's dialogue acts using a
multi-classes hierarchical structure. The model can automatically acquire
probabilistic discourse knowledge from a dialogue corpus were collected and
annotated manually from multi-genre Egyptian call-centers. Extensive
experiments were conducted using Support Vector Machines classifier to evaluate
the system performance. The results attained in the term of average F-measure
scores of 0.912; showed that the proposed approach has moderately improved
F-measure by approximately 20%. | ['AbdelRahim Elmadany', 'Sherif Abdou', 'Mervat Gheith'] | 2018-05-30 | improving-dialogue-act-classification-for-2 | https://aclanthology.org/L18-1020 | https://aclanthology.org/L18-1020.pdf | lrec-2018-5 | ['dialogue-act-classification'] | ['natural-language-processing'] | [ 5.61866723e-03 6.56082511e-01 1.73814744e-01 -7.19714761e-01
-6.45638943e-01 -6.89126432e-01 9.79371250e-01 3.42974931e-01
-1.72634438e-01 1.00347841e+00 6.22039914e-01 -2.96590924e-01
-1.90567616e-02 -5.28072774e-01 2.57889122e-01 -4.77988422e-01
4.86362390e-02 8.19706976e-01 1.67586073e-01 -6.71723068e-01
7.74011850e-01 2.38776967e-01 -1.10525751e+00 6.15188420e-01
8.17907810e-01 5.36882579e-01 1.71685189e-01 1.20473123e+00
-4.69878584e-01 1.48545969e+00 -1.15575218e+00 -1.06693484e-01
-3.17902863e-01 -7.55674422e-01 -1.57476807e+00 5.86095810e-01
-3.06788504e-01 -3.91813070e-01 7.23836794e-02 7.63515055e-01
8.95084217e-02 2.34405428e-01 1.02957511e+00 -8.58768106e-01
-1.17119849e-01 8.08894098e-01 -8.02157074e-02 3.08532417e-01
9.11090016e-01 -1.96631446e-01 9.38702285e-01 -5.38850784e-01
2.98927754e-01 1.60822129e+00 2.40661532e-01 5.88494599e-01
-1.02092171e+00 -2.55177617e-01 -3.67421657e-01 -5.78919128e-02
-1.00563169e+00 -4.23700452e-01 7.00612485e-01 -5.73313892e-01
1.12807977e+00 3.36064368e-01 1.18522100e-01 6.38093591e-01
7.05860332e-02 8.04263294e-01 1.45934474e+00 -9.05544221e-01
-6.33419603e-02 7.22908974e-01 8.03381503e-01 7.18924344e-01
-6.03942215e-01 -6.04106545e-01 -4.44598824e-01 -3.52550179e-01
4.99381304e-01 -7.97628164e-01 7.57348680e-05 6.12686813e-01
-8.75036716e-01 1.19111454e+00 -3.75941455e-01 9.03203428e-01
-4.57078665e-01 -9.05617237e-01 6.14733279e-01 4.60256845e-01
3.38831425e-01 5.24520993e-01 -4.98970360e-01 -8.20729315e-01
-4.37472522e-01 5.98992258e-02 1.43118334e+00 6.63366854e-01
2.80301929e-01 9.34238881e-02 4.65921825e-03 1.23212290e+00
3.89791489e-01 2.57099599e-01 5.56851089e-01 -7.42015183e-01
3.80940944e-01 8.64565492e-01 1.71843290e-01 -1.05404139e+00
-5.39884269e-01 4.32947099e-01 -4.41903502e-01 -2.20492497e-01
7.54935324e-01 -5.19938767e-01 -3.24304700e-01 1.21237075e+00
4.58073825e-01 -6.31567657e-01 6.15763545e-01 5.38882315e-01
1.05999017e+00 1.16240597e+00 1.27892971e-01 -5.68455458e-01
1.44888175e+00 -7.67156005e-01 -1.17377806e+00 1.25401929e-01
5.55255175e-01 -1.09829497e+00 1.01059997e+00 6.15984559e-01
-9.29331839e-01 -5.72980702e-01 -8.45795274e-01 2.51863778e-01
-5.20965606e-02 3.88411999e-01 4.22177583e-01 9.28061545e-01
-6.43744826e-01 -1.17686503e-01 -3.82988304e-01 -2.92227149e-01
-2.35757276e-01 3.36653411e-01 -2.80900091e-01 4.37945873e-01
-1.25609565e+00 1.08363616e+00 6.42937779e-01 -2.58161575e-02
-5.03993452e-01 2.29479611e-01 -7.66730309e-01 -1.44313529e-01
1.14101447e-01 2.81339973e-01 1.51984334e+00 -1.01964140e+00
-1.99198294e+00 7.22330511e-01 -3.03405493e-01 -3.75712395e-01
2.74366081e-01 -7.00181425e-02 -4.73935246e-01 3.46755296e-01
-1.20220423e-01 -3.58252674e-02 3.99395674e-01 -9.88632798e-01
-8.87290776e-01 -3.91134322e-01 2.16061756e-01 4.80552852e-01
-1.19696923e-01 4.79884088e-01 1.71790719e-01 -1.88976318e-01
1.22717410e-01 -8.10345531e-01 2.98098266e-01 -9.99097586e-01
-2.91640669e-01 -7.32643485e-01 8.32249522e-01 -1.06173146e+00
1.32756639e+00 -1.78102672e+00 -7.69347548e-02 1.08230382e-01
-4.89760302e-02 3.09190810e-01 5.71186900e-01 6.64772809e-01
3.25150996e-01 -1.62465781e-01 -1.05794087e-01 2.59733260e-01
1.72435716e-02 2.34093294e-01 -1.74792573e-01 6.05966523e-02
1.12430669e-01 6.29544333e-02 -7.53123045e-01 -7.29416788e-01
4.58454460e-01 2.79088974e-01 -1.23060457e-01 6.98321164e-01
-1.81680679e-01 5.38341641e-01 -6.31344497e-01 2.30977550e-01
1.82494491e-01 1.43246993e-01 6.00261927e-01 2.44844053e-02
-1.71129882e-01 5.56594372e-01 -1.04789305e+00 1.05256331e+00
-5.00166059e-01 9.78518546e-01 2.62821257e-01 -1.21321905e+00
1.21275353e+00 9.45511460e-01 7.87839293e-02 -1.54161587e-01
5.01099288e-01 8.55321661e-02 3.58915687e-01 -9.19570029e-01
7.69081891e-01 -3.58268052e-01 -3.62270832e-01 5.10994434e-01
1.68995216e-01 -1.68328404e-01 2.54747033e-01 1.61769763e-01
5.43465734e-01 -3.86846185e-01 7.24702299e-01 -2.96566069e-01
1.33228600e+00 2.10048437e-01 1.70704886e-01 4.35374498e-01
-2.89321452e-01 -6.05130605e-02 7.78989911e-01 -2.06042230e-01
-5.58245301e-01 -6.61726952e-01 -3.05221707e-01 1.35760295e+00
-4.15334284e-01 -2.00031046e-02 -1.02653873e+00 -8.65221798e-01
-4.70066398e-01 1.20138693e+00 -2.24508837e-01 3.71255726e-01
-6.21207178e-01 -7.86949039e-01 8.25105965e-01 -6.17548078e-02
7.43875444e-01 -1.08700275e+00 -2.33904392e-01 4.92150038e-01
-4.84195262e-01 -1.18317676e+00 -6.98194057e-02 -1.46554753e-01
-4.64345604e-01 -1.24779618e+00 -3.27464581e-01 -1.08302748e+00
4.78093028e-01 -1.92827687e-01 6.80693924e-01 -6.70957565e-03
-4.96327989e-02 2.70900667e-01 -6.93964243e-01 -5.76604486e-01
-1.32623303e+00 -2.83879470e-02 -4.14799377e-02 1.11429334e-01
8.98215473e-01 4.76755463e-02 5.92464916e-02 4.29219693e-01
-5.79403579e-01 -1.10856622e-01 2.35224158e-01 9.71280038e-01
-4.41167384e-01 4.27515984e-01 1.00094783e+00 -1.00650334e+00
1.28410077e+00 -3.76276970e-01 -3.71065110e-01 3.74655992e-01
-2.70074815e-01 1.11041591e-01 4.93108630e-01 -1.51223242e-02
-1.67112446e+00 -1.64310038e-01 -3.90867501e-01 8.13650906e-01
-6.51748478e-01 3.50411773e-01 -6.75370693e-02 2.50481665e-01
5.51543355e-01 3.86160821e-01 3.52368414e-01 -2.35805973e-01
3.75864543e-02 1.67390406e+00 2.60508567e-01 -6.76305473e-01
-1.01867281e-02 -1.96974456e-01 -5.13426483e-01 -1.61795604e+00
-6.96499705e-01 -7.46385098e-01 -8.41606259e-01 -6.67947531e-01
1.04395723e+00 -5.91610610e-01 -1.01936603e+00 6.47282958e-01
-1.16284370e+00 -1.09781832e-01 5.78954756e-01 6.51792526e-01
-3.68270189e-01 5.34140587e-01 -8.30273569e-01 -1.45317709e+00
-2.54325539e-01 -1.01841342e+00 4.65474725e-01 3.72983992e-01
-8.75020087e-01 -1.33626568e+00 -8.90446901e-02 9.09428477e-01
2.01542839e-01 7.82813877e-02 1.07210970e+00 -1.45486474e+00
2.66415954e-01 -2.58830398e-01 7.65355527e-02 5.88060677e-01
6.20010734e-01 1.77679658e-01 -1.06805229e+00 1.76259920e-01
5.41547537e-01 -7.67116189e-01 -1.20589219e-01 1.14841096e-01
3.59147042e-01 -6.39116824e-01 1.55432388e-01 -7.66857028e-01
8.69975209e-01 8.03590000e-01 2.48597398e-01 6.14409372e-02
1.64881498e-01 1.20920241e+00 9.66747224e-01 5.52446783e-01
3.62578034e-01 3.67990613e-01 -3.45801711e-01 4.66327190e-01
2.99205303e-01 1.48441538e-01 4.14428025e-01 1.38148594e+00
2.25914046e-02 -1.21456474e-01 -1.16925168e+00 3.86745512e-01
-1.63667274e+00 -1.05819166e+00 -2.88751543e-01 1.68458116e+00
1.31794047e+00 2.96401441e-01 3.96784216e-01 5.83077610e-01
8.08844626e-01 6.78810403e-02 1.80397153e-01 -8.87032032e-01
7.90354684e-02 -4.22500372e-02 -2.20181569e-01 1.27445757e+00
-1.03057730e+00 8.81090283e-01 5.90241146e+00 6.70256793e-01
-8.70078266e-01 -5.07376939e-02 6.69448197e-01 4.54717427e-01
2.37584904e-01 -3.05405676e-01 -7.83313632e-01 2.86505699e-01
1.26980007e+00 -3.41596186e-01 1.64483398e-01 6.38424993e-01
3.51880580e-01 -5.22195816e-01 -8.24866652e-01 5.44883132e-01
3.57016981e-01 -9.29827929e-01 1.08648062e-01 -1.00977719e-01
4.26585466e-01 -5.22541940e-01 -5.85674345e-01 4.99731481e-01
4.56376463e-01 -9.45892990e-01 1.56203404e-01 3.11745524e-01
1.35193095e-01 -8.93137693e-01 9.54965532e-01 9.47982967e-01
-6.88558817e-01 1.58971250e-01 -1.13477454e-01 -3.46294224e-01
1.05696447e-01 -9.63079631e-02 -1.89562154e+00 1.76788732e-01
5.06863818e-02 -3.04623879e-02 -1.42789796e-01 3.01371574e-01
-2.45667204e-01 1.10415065e+00 -1.76827282e-01 -7.77834833e-01
4.36494768e-01 -3.15463930e-01 5.67530930e-01 1.46375000e+00
-2.19245628e-01 6.50495410e-01 2.52150536e-01 1.32199511e-01
4.25943255e-01 6.16616666e-01 -5.21921277e-01 -1.27486318e-01
4.83515114e-01 9.34915364e-01 -7.37153769e-01 -4.52180117e-01
-3.56565654e-01 5.96635342e-01 -5.16294576e-02 -9.31536481e-02
-5.64399064e-01 -5.70250809e-01 -5.44393901e-03 -2.53232121e-01
-2.16247633e-01 -2.09483951e-01 -1.89593613e-01 -6.87596321e-01
-3.27547759e-01 -1.14330018e+00 3.78017604e-01 -3.46057802e-01
-1.09108412e+00 9.08619165e-01 1.64182380e-01 -7.37162590e-01
-7.61962831e-01 -7.06956863e-01 -4.58995730e-01 9.73311245e-01
-7.18362153e-01 -8.44398797e-01 -3.22801918e-02 4.55365330e-01
1.31049109e+00 -6.80505514e-01 1.41489899e+00 -1.17083468e-01
-4.13926512e-01 2.98192173e-01 -1.87206361e-02 6.88188732e-01
5.04580736e-01 -1.51185727e+00 -4.18877900e-01 3.28373492e-01
1.20425075e-02 5.37036538e-01 9.75234210e-01 -4.59118903e-01
-8.53016317e-01 -3.24364781e-01 1.24476278e+00 -3.97948414e-01
6.66971028e-01 1.27338115e-02 -9.43504035e-01 4.92014259e-01
7.39778519e-01 -1.08007038e+00 1.21440029e+00 2.77336299e-01
1.75490171e-01 2.05980524e-01 -1.34059834e+00 2.42436394e-01
-2.25868031e-01 -6.79154634e-01 -1.51513827e+00 4.15896118e-01
1.70617044e-01 -2.96426892e-01 -1.12799501e+00 -3.57809477e-02
3.79133105e-01 -8.94457400e-01 4.36999440e-01 -6.86795533e-01
1.45565689e-01 -6.33432567e-02 -2.43328825e-01 -9.98134196e-01
2.93899328e-01 -6.26344919e-01 2.32376263e-01 1.52485168e+00
7.29173183e-01 -4.95923966e-01 5.27142584e-01 9.38147843e-01
1.97071105e-01 -2.24098638e-01 -7.14879632e-01 -1.66665569e-01
-5.28038964e-02 -2.45817184e-01 1.85079545e-01 1.10607135e+00
7.38414645e-01 1.12397063e+00 -2.19999656e-01 -2.09289268e-04
3.05735618e-01 -2.77349651e-01 8.34226787e-01 -1.17793262e+00
-1.63703427e-01 -3.38348508e-01 -2.91724920e-01 -1.05771279e+00
2.96998799e-01 -2.28629291e-01 2.50285387e-01 -1.26317775e+00
-1.81030899e-01 -1.65849626e-01 3.22623521e-01 1.72543123e-01
-2.59513378e-01 -3.95895571e-01 -1.03739947e-01 2.04838067e-01
-2.56607920e-01 4.75001842e-01 8.17169309e-01 -5.07383831e-02
-4.45686042e-01 6.06040657e-01 -2.86355853e-01 9.78871226e-01
1.06153166e+00 -2.75656730e-01 -7.23092258e-01 7.73696154e-02
-4.70623165e-01 7.39520133e-01 -4.46657389e-01 -5.13736367e-01
8.79843235e-02 -3.31146181e-01 -1.19740196e-01 -6.39411747e-01
4.40671116e-01 -5.35669863e-01 -4.23993379e-01 3.94218147e-01
-8.19066405e-01 -9.96756777e-02 1.26846001e-01 3.35934252e-01
-6.39067471e-01 -7.43740916e-01 7.81141520e-01 -2.46579200e-01
-5.34483731e-01 -7.14815199e-01 -1.13176560e+00 1.47362677e-02
1.24029624e+00 -1.31039187e-01 -1.20377928e-01 -6.69254422e-01
-8.89169812e-01 9.46666673e-02 -1.12731859e-01 3.46601874e-01
4.04678583e-01 -7.85146713e-01 -8.90650511e-01 2.35438952e-03
-1.85732633e-01 -2.96183467e-01 5.00335693e-02 4.73230809e-01
-9.43494499e-01 8.81213248e-01 -3.15942407e-01 -4.93322253e-01
-1.92891872e+00 -2.91898757e-01 7.92761073e-02 -2.43921936e-01
-1.02878571e-01 7.74056256e-01 -2.77056485e-01 -6.02827609e-01
3.46102834e-01 1.01164214e-01 -1.09380054e+00 2.18912631e-01
6.99632287e-01 4.23388571e-01 -1.83557749e-01 -1.17638075e+00
-7.04316273e-02 -2.94542983e-02 -2.84019232e-01 -7.06852853e-01
9.75693345e-01 -4.26249981e-01 -3.97874862e-01 9.28573489e-01
9.59925473e-01 1.76250279e-01 -5.85585177e-01 -3.24392438e-01
4.13738936e-01 -3.03373367e-01 -1.61757886e-01 -9.19814944e-01
-9.06232670e-02 6.98292255e-01 1.16615765e-01 9.85434830e-01
5.90910256e-01 -1.54071391e-01 4.54553336e-01 7.62448668e-01
2.74449974e-01 -1.59698200e+00 -6.64867926e-03 9.54874456e-01
9.25660312e-01 -1.49074543e+00 -2.00686261e-01 -5.24161637e-01
-1.19991779e+00 1.52089357e+00 6.78484619e-01 2.54648030e-01
8.08419883e-01 -6.99427873e-02 5.50029457e-01 -2.68179357e-01
-7.13280380e-01 2.34675884e-01 1.72228321e-01 3.91894847e-01
1.09284377e+00 2.41876483e-01 -7.38870084e-01 5.17216384e-01
-4.64688987e-01 -4.22960252e-01 8.15167367e-01 9.67807114e-01
-9.04414475e-01 -1.05333602e+00 -5.89312792e-01 3.72024208e-01
-8.37270975e-01 1.49396226e-01 -8.87552977e-01 7.70409942e-01
-5.43543160e-01 1.68439341e+00 -1.54049352e-01 -3.05534124e-01
1.29853159e-01 5.39372444e-01 1.10100158e-01 -8.15235376e-01
-8.30153465e-01 6.64525405e-02 9.96452987e-01 3.17249924e-01
-8.14712167e-01 -7.13225245e-01 -1.35338056e+00 -2.63074994e-01
-5.07619202e-01 1.10161054e+00 6.58777058e-01 1.35359335e+00
-4.20185804e-01 2.87792653e-01 1.03766298e+00 -1.89916775e-01
-7.09811568e-01 -1.71988630e+00 -4.24911201e-01 2.61323333e-01
1.16590850e-01 -2.80817986e-01 -3.39317381e-01 3.91376048e-01] | [12.808451652526855, 7.897951126098633] |
776fddf9-a53b-4d35-9e76-573d3a869739 | multi-level-context-ultra-aggregation-for | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Nie_Multi-Level_Context_Ultra-Aggregation_for_Stereo_Matching_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Nie_Multi-Level_Context_Ultra-Aggregation_for_Stereo_Matching_CVPR_2019_paper.pdf | Multi-Level Context Ultra-Aggregation for Stereo Matching | Exploiting multi-level context information to cost volume can improve the performance of learning-based stereo matching methods. In recent years, 3-D Convolution Neural Networks (3-D CNNs) show the advantages in regularizing cost volume but are limited by unary features learning in matching cost computation. However, existing methods only use features from plain convolution layers or a simple aggregation of multi-level features to calculate cost volume, which is insufficient because stereo matching requires discriminative features to identify corresponding pixels in rectified stereo image pairs. In this paper, we propose a unary features descriptor using multi-level context ultra-aggregation (MCUA), which encapsulates all convolutional features into a more discriminative representation by intra- and inter-level features combination. Specifically, a child module that takes low-resolution images as input captures larger context information; the larger context information from each layer is densely connected to the main branch of the network. MCUA makes good usage of multi-level features with richer context and performs the image-to-image prediction holistically. We introduce our MCUA scheme for cost volume calculation and test it on PSM-Net. We also evaluate our method on Scene Flow and KITTI 2012/2015 stereo datasets. Experimental results show that our method outperforms state-of-the-art methods by a notable margin and effectively improves the accuracy of stereo matching.
| [' Yongtian Wang', ' Yue Liu', ' Deng-Ping Fan', ' Zhengfa Liang', ' Yun Liu', ' Ming-Ming Cheng', 'Guang-Yu Nie'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['stereo-matching'] | ['computer-vision'] | [ 1.08089916e-01 -6.71106100e-01 -2.17219695e-01 -6.07946157e-01
-3.76005977e-01 -5.37758023e-02 5.82104325e-01 -2.54532024e-02
-5.94753861e-01 5.08784413e-01 3.85832250e-01 6.78685531e-02
-4.19948483e-03 -1.24612820e+00 -7.67639697e-01 -5.42322040e-01
4.44655456e-02 8.09648037e-02 5.19379020e-01 -3.06124657e-01
3.47539067e-01 8.66260529e-01 -1.99521101e+00 5.86630642e-01
8.81924450e-01 1.63053060e+00 4.02216911e-01 3.37228626e-01
-3.42492133e-01 9.37311709e-01 -8.16923380e-02 -9.39775482e-02
7.67718136e-01 -5.07184714e-02 -7.75409877e-01 -2.07703486e-01
1.18645179e+00 -7.16054857e-01 -8.77226174e-01 9.15543258e-01
6.06265187e-01 1.64532304e-01 4.13529426e-01 -1.10504138e+00
-2.46916339e-01 4.32145409e-02 -5.69192111e-01 5.44151068e-01
2.32446149e-01 3.48752528e-01 8.96477401e-01 -9.36548889e-01
5.20005584e-01 1.41936815e+00 8.53776932e-01 2.61061072e-01
-8.98707628e-01 -8.30926955e-01 -1.46508012e-02 4.54820514e-01
-1.32159710e+00 -2.12361515e-01 9.36038256e-01 -4.67490166e-01
1.27790201e+00 1.22660235e-01 9.08069074e-01 3.71729970e-01
1.14539489e-01 6.73422873e-01 1.27011132e+00 -7.27704912e-02
-2.77952194e-01 -3.49285126e-01 3.96737009e-02 9.35980856e-01
3.21358144e-02 6.03125215e-01 -6.10544264e-01 1.15644388e-01
1.20732057e+00 4.24568564e-01 -3.65389347e-01 -5.31335652e-01
-1.17483723e+00 7.36932158e-01 1.07890654e+00 2.14805201e-01
-2.53199935e-01 3.14429671e-01 4.20438111e-01 2.03337327e-01
2.83964306e-01 -7.88092613e-03 -5.35574496e-01 -1.04342699e-02
-9.03604627e-01 2.34640956e-01 3.00651819e-01 9.03929651e-01
1.46537220e+00 -1.86654851e-01 -3.25613648e-01 9.67873812e-01
-1.55017167e-01 4.08874184e-01 3.90749991e-01 -1.13543379e+00
7.46012747e-01 1.04481280e+00 -3.39837462e-01 -9.19263959e-01
-4.20050085e-01 -2.24863976e-01 -1.07801080e+00 3.55559051e-01
3.39770883e-01 2.58162200e-01 -7.95164347e-01 1.39409161e+00
2.45079070e-01 4.17814225e-01 -2.20956177e-01 1.26415718e+00
1.22429800e+00 6.08698666e-01 -5.21796802e-03 1.38255954e-01
1.10411906e+00 -1.06616771e+00 -1.24306440e-01 2.75927149e-02
5.36044002e-01 -8.28192770e-01 9.00007188e-01 -9.16505083e-02
-1.02309453e+00 -1.00272155e+00 -9.79310215e-01 -4.17872041e-01
-4.98328984e-01 1.50012346e-02 9.20675218e-01 2.47648984e-01
-9.70599294e-01 9.72232223e-01 -3.61769736e-01 -7.38609731e-02
7.98365474e-01 5.16895592e-01 -6.02160931e-01 -3.21630180e-01
-1.20738506e+00 8.07959735e-01 1.27409205e-01 -4.42267023e-02
-6.71668589e-01 -9.51742768e-01 -1.23665690e+00 1.38828084e-01
3.78273726e-02 -9.64155793e-01 8.76123309e-01 -8.43474209e-01
-1.20040071e+00 9.76089239e-01 -1.79459110e-01 -2.29318738e-01
4.89595890e-01 -7.21430704e-02 -4.76225987e-02 2.87018389e-01
1.99026108e-01 1.10590637e+00 6.51828945e-01 -8.60544264e-01
-1.09562278e+00 -2.91928887e-01 3.53813618e-01 1.98723838e-01
-2.24127650e-01 -2.04639405e-01 -5.12176692e-01 -5.98799050e-01
1.68730333e-01 -4.66533840e-01 -3.25909615e-01 3.64259452e-01
1.28467456e-01 -3.15550715e-01 8.83771122e-01 -3.34384769e-01
1.07514155e+00 -1.93727314e+00 -1.77518830e-01 1.73197761e-01
2.63876110e-01 3.90995681e-01 -1.71436250e-01 3.20107900e-02
-1.10458441e-01 -3.04697335e-01 -1.43459871e-01 -1.88612565e-01
-1.77160844e-01 2.71269172e-01 -1.11020759e-01 4.22911376e-01
3.20063680e-01 1.00180364e+00 -8.84989381e-01 -8.94660115e-01
1.00654304e+00 5.61143816e-01 -8.55816841e-01 2.84776986e-01
3.99023801e-01 3.07600230e-01 -2.54841596e-01 8.08249831e-01
1.12644780e+00 -3.28630321e-02 -3.60099047e-01 -7.38716960e-01
-3.50297004e-01 3.56767476e-01 -1.15164018e+00 1.96763813e+00
-8.96263063e-01 8.18259835e-01 -3.26839179e-01 -1.09795964e+00
8.86910796e-01 -2.13264555e-01 7.10387766e-01 -1.08680463e+00
1.96654961e-01 4.04212356e-01 -2.13282824e-01 -3.24622661e-01
3.70372266e-01 1.35311231e-01 -1.95710007e-02 -6.61244690e-02
1.85442299e-01 -1.25492185e-01 7.07573891e-02 -7.69774541e-02
8.68625581e-01 1.65469661e-01 2.75863826e-01 -3.24377239e-01
1.07252729e+00 -2.34482139e-01 8.55988860e-01 5.92449367e-01
-4.05917376e-01 7.83665180e-01 7.32904896e-02 -1.04290843e+00
-1.06309962e+00 -9.74068224e-01 -4.83269304e-01 6.81471884e-01
5.64481854e-01 -3.46991539e-01 -4.73426163e-01 -6.29996121e-01
3.88902843e-01 -2.00162411e-01 -7.16422319e-01 7.87571445e-02
-9.56682742e-01 -3.43548775e-01 2.57394612e-01 9.25738513e-01
1.22149301e+00 -1.00658357e+00 -7.99835861e-01 3.64831209e-01
-1.56854987e-01 -1.25012207e+00 -7.48851836e-01 -1.42562645e-03
-1.05734062e+00 -1.20949233e+00 -6.72012687e-01 -8.88064921e-01
6.48794055e-01 4.86752063e-01 1.16349256e+00 3.53250951e-01
-6.70089066e-01 4.56494913e-02 8.25784579e-02 7.87994340e-02
2.25191429e-01 -1.44301951e-01 -2.18779504e-01 -1.27557009e-01
4.13961887e-01 -9.00012374e-01 -1.05747914e+00 4.51098114e-01
-7.85340369e-01 2.88886458e-01 5.75183690e-01 9.61748421e-01
6.25059485e-01 -3.54020894e-01 -7.12522939e-02 -3.98666114e-01
-1.27313519e-02 1.04433283e-01 -7.56266057e-01 7.14432681e-03
-3.85894358e-01 1.26389071e-01 6.19278967e-01 -1.92970753e-01
-9.81098533e-01 2.64430285e-01 -2.21243590e-01 -7.15940654e-01
-1.61283940e-01 -1.67112108e-02 -1.86746761e-01 -6.18545115e-01
3.39972466e-01 3.04594666e-01 -1.84711710e-01 -3.57994229e-01
1.90411717e-01 4.96131122e-01 6.05924487e-01 -6.38234556e-01
7.32508004e-01 5.89967489e-01 4.58240658e-01 -5.21536648e-01
-7.39905179e-01 -6.50907815e-01 -7.99510777e-01 -4.11668628e-01
6.05934501e-01 -1.16660237e+00 -9.20566201e-01 7.07038522e-01
-1.15788221e+00 -7.52497166e-02 -3.48634303e-01 4.74910676e-01
-6.08712971e-01 3.86011928e-01 -7.65421689e-01 -1.53686032e-01
-4.54473197e-01 -1.41562331e+00 1.24033654e+00 4.58151400e-01
3.33813697e-01 -7.77614295e-01 -6.68734089e-02 3.50691736e-01
5.47271848e-01 2.16754034e-01 6.51542306e-01 1.35418162e-01
-9.38525438e-01 2.96800826e-02 -8.94142210e-01 4.93805379e-01
8.83051306e-02 -2.46015713e-01 -1.18676293e+00 -2.24790424e-02
-4.61404979e-01 -2.61859924e-01 1.37533855e+00 6.12033010e-01
1.65278292e+00 1.67529937e-02 -2.60430962e-01 1.25232887e+00
1.81961346e+00 -1.50377661e-01 8.44555974e-01 5.78603506e-01
9.78974700e-01 5.70073247e-01 5.34235775e-01 3.66292775e-01
5.56459427e-01 7.50502944e-01 4.85790521e-01 -3.74449134e-01
-5.52385688e-01 -2.66765714e-01 1.02679797e-01 5.98630846e-01
-4.38106567e-01 5.13164520e-01 -6.63605750e-01 4.41247016e-01
-1.80728841e+00 -1.08474803e+00 -1.58702016e-01 2.15927219e+00
6.66357517e-01 7.79711753e-02 -1.26493707e-01 7.83898160e-02
6.64373219e-01 3.14774245e-01 -2.76487887e-01 -3.33457053e-01
-4.49594975e-01 5.49282491e-01 8.06862473e-01 4.74455118e-01
-1.37550962e+00 8.87404442e-01 5.13994217e+00 1.18393183e+00
-1.11249554e+00 -8.43918696e-02 6.69120908e-01 4.30716835e-02
-1.09821916e-01 8.97973925e-02 -7.65287220e-01 4.24342126e-01
1.39314279e-01 1.08153120e-01 2.32290134e-01 9.60396171e-01
2.71480549e-02 -2.28967860e-01 -1.18905663e+00 1.48372567e+00
-1.21553659e-01 -1.77940607e+00 2.74516314e-01 5.82498610e-02
9.28543091e-01 2.55325198e-01 -2.64757484e-01 1.95328563e-01
-1.49658203e-01 -9.84885812e-01 4.78230685e-01 4.33685184e-01
9.28747892e-01 -8.55246603e-01 9.51574326e-01 -1.43422429e-02
-1.89114249e+00 -1.40951931e-01 -7.41241038e-01 -3.93355161e-01
2.19321903e-02 6.77557707e-01 -8.32145661e-02 7.24713504e-01
9.70600069e-01 1.20608282e+00 -5.23115337e-01 1.26427162e+00
9.84843522e-02 -1.16084002e-01 -3.06349099e-01 3.19107413e-01
5.01629829e-01 -1.91958934e-01 2.02005237e-01 1.41316688e+00
2.42803276e-01 1.78512320e-01 2.20470056e-01 7.14447796e-01
-2.11613268e-01 1.39375433e-01 -6.64908826e-01 8.21345747e-01
2.05715656e-01 1.31377780e+00 -3.80138993e-01 -6.52976453e-01
-6.54222369e-01 8.76634181e-01 4.21441495e-01 -3.96367982e-02
-6.40521705e-01 -4.63698626e-01 1.10281992e+00 2.01278791e-01
4.12449449e-01 -6.54191151e-02 -3.54715645e-01 -1.22105098e+00
5.06931283e-02 -3.08690488e-01 4.16422188e-01 -7.01403856e-01
-1.32023954e+00 5.00803351e-01 -1.52523622e-01 -1.77893460e+00
1.17885526e-02 -6.97265804e-01 -6.95378780e-01 1.05752456e+00
-2.19039297e+00 -1.15836871e+00 -8.70071530e-01 9.60146904e-01
5.21335304e-01 -7.26466626e-02 5.62076509e-01 6.31169319e-01
-3.60992223e-01 5.18356085e-01 -3.21970642e-01 3.74641269e-01
5.74966311e-01 -9.70884323e-01 2.35838130e-01 5.94534397e-01
-2.46226817e-01 2.41973206e-01 -1.96022056e-02 -2.95197546e-01
-1.04022288e+00 -1.21331668e+00 1.06220436e+00 -9.92120579e-02
2.97674447e-01 -3.89081165e-02 -6.45363271e-01 1.45483702e-01
1.51192248e-02 6.74727261e-01 2.99627542e-01 -3.70137542e-01
-5.09141982e-01 -6.37018979e-01 -1.23619080e+00 3.06525588e-01
1.62344909e+00 -5.65202773e-01 -6.03958368e-01 -8.22354034e-02
5.73367715e-01 -5.35296023e-01 -1.00309205e+00 7.97779620e-01
6.74874723e-01 -1.52575839e+00 1.33343577e+00 -2.35160157e-01
8.26626122e-01 -4.07715589e-01 -3.75678748e-01 -8.19066167e-01
-3.08303356e-01 -1.51565269e-01 1.82134047e-01 9.83709812e-01
-2.07091242e-01 -5.55550754e-01 7.85495162e-01 3.77477348e-01
-3.66438746e-01 -9.54311430e-01 -1.09025788e+00 -7.21074998e-01
-1.93483382e-02 -5.08615553e-01 8.17130864e-01 9.38819349e-01
-2.19166368e-01 -6.20510280e-02 -2.65199751e-01 -1.05122119e-01
7.78862357e-01 6.69954717e-01 8.27606320e-01 -1.17328763e+00
4.64939363e-02 -7.83975780e-01 -9.55885708e-01 -1.36690080e+00
1.18475154e-01 -9.07663167e-01 -2.32791513e-01 -1.28594172e+00
4.18133348e-01 -6.62828863e-01 -3.27607155e-01 4.01731342e-01
-2.32441723e-01 5.59235573e-01 3.03998530e-01 2.09417462e-01
-3.52048725e-01 6.28416836e-01 1.58094728e+00 -3.33188266e-01
-8.37590545e-02 -2.65075594e-01 -1.30779520e-01 7.28805959e-01
6.34439290e-01 -1.22134693e-01 -6.58706129e-02 -4.67066675e-01
-2.29571298e-01 -5.91150997e-03 6.86960578e-01 -1.37655818e+00
1.91561967e-01 -1.56173959e-01 7.97477067e-01 -8.48192692e-01
2.99372703e-01 -9.78974164e-01 -4.25350340e-03 5.60043752e-01
-5.66084087e-02 -2.26557832e-02 1.73291087e-01 7.80668855e-02
-6.22833908e-01 1.04124114e-01 8.93803537e-01 -2.94743806e-01
-1.33431566e+00 9.27814007e-01 2.08530515e-01 -5.34133017e-02
6.91337585e-01 -4.78233844e-01 -2.61668712e-01 -2.04554610e-02
-2.16856331e-01 2.09583476e-01 5.62826157e-01 4.16542560e-01
8.14274311e-01 -1.77191412e+00 -5.28570175e-01 3.76332492e-01
3.35053861e-01 3.12758625e-01 5.25832176e-01 6.81807995e-01
-8.57125342e-01 5.43031096e-01 -7.23816335e-01 -1.02337921e+00
-1.18130672e+00 3.33435059e-01 5.88442743e-01 -2.67193615e-01
-6.86896145e-01 8.26951981e-01 5.32476306e-01 -3.71924520e-01
2.05121607e-01 -5.13567984e-01 -3.22794527e-01 -6.88442662e-02
4.19338614e-01 3.46903801e-01 7.32293651e-02 -8.76043916e-01
-4.41679835e-01 1.32280815e+00 2.15069711e-01 5.29280126e-01
1.39316237e+00 -8.25212374e-02 -1.67287380e-01 -1.24602109e-01
1.73693538e+00 -3.80424052e-01 -1.57343519e+00 -6.30208969e-01
-4.98387218e-01 -1.01045811e+00 3.37330103e-01 -3.25631231e-01
-1.74493325e+00 1.18198466e+00 7.82905042e-01 -4.99121130e-01
1.46313143e+00 -1.17017306e-01 1.18341565e+00 1.14062689e-01
5.37943661e-01 -1.06697416e+00 3.80123369e-02 5.06633699e-01
7.49377608e-01 -1.53095627e+00 5.32510318e-02 -6.02789938e-01
-2.44615227e-01 1.42918360e+00 1.04531026e+00 -3.99872035e-01
8.19990098e-01 1.69664815e-01 -8.78105238e-02 -1.11593552e-01
-3.55910301e-01 -5.73056221e-01 5.42568803e-01 4.73274022e-01
2.20732778e-01 -4.40159813e-02 -2.47322947e-01 2.37914652e-01
-1.34637907e-01 4.80882414e-02 8.97215009e-02 8.73189390e-01
-4.35864121e-01 -1.12451506e+00 -2.10088715e-01 5.20715237e-01
-6.68267906e-02 -3.68392766e-01 1.64914981e-01 7.78828025e-01
6.75631881e-01 5.08720636e-01 4.87540156e-01 -6.25380158e-01
5.79479694e-01 -3.83905292e-01 6.90221012e-01 -2.10475057e-01
-6.80336893e-01 -1.50869131e-01 -8.66991356e-02 -1.18635809e+00
-8.15312028e-01 -5.00235915e-01 -1.13519239e+00 -6.32915080e-01
-1.28169907e-02 -4.13529873e-01 3.93456787e-01 7.58794487e-01
2.20611632e-01 2.90321797e-01 9.59076822e-01 -1.37455654e+00
-9.15939212e-02 -7.40073860e-01 -3.10941875e-01 6.80055141e-01
5.28043091e-01 -8.74919176e-01 -2.13905111e-01 -9.06292498e-02] | [8.89608383178711, -2.2205333709716797] |
ac471f66-38f9-4f54-bb57-eaf26a77c8dc | spherical-convolutional-neural-network-for-3d | 1805.07872 | null | http://arxiv.org/abs/1805.07872v2 | http://arxiv.org/pdf/1805.07872v2.pdf | Spherical Convolutional Neural Network for 3D Point Clouds | We propose a neural network for 3D point cloud processing that exploits
`spherical' convolution kernels and octree partitioning of space. The proposed
metric-based spherical kernels systematically quantize point neighborhoods to
identify local geometric structures in data, while maintaining the properties
of translation-invariance and asymmetry. The network architecture itself is
guided by octree data structuring that takes full advantage of the sparse
nature of irregular point clouds. We specify spherical kernels with the help of
neurons in each layer that in turn are associated with spatial locations. We
exploit this association to avert dynamic kernel generation during network
training, that enables efficient learning with high resolution point clouds. We
demonstrate the utility of the spherical convolutional neural network for 3D
object classification on standard benchmark datasets. | ['Ajmal Mian', 'Huan Lei', 'Naveed Akhtar'] | 2018-05-21 | null | null | null | null | ['3d-object-classification'] | ['computer-vision'] | [-3.21035951e-01 -1.67488337e-01 2.74889544e-02 -3.63353819e-01
-6.06434569e-02 -6.86670601e-01 5.67740738e-01 3.04324865e-01
-3.39855701e-01 7.07942108e-03 -1.38701499e-01 -3.68159175e-01
-4.42351371e-01 -1.16812909e+00 -9.04931724e-01 -4.17809427e-01
-7.94892490e-01 5.90338349e-01 4.59685862e-01 -5.58738895e-02
4.10322547e-01 1.65892947e+00 -1.59109330e+00 4.31498736e-01
8.28712940e-01 1.33826637e+00 3.00186928e-02 5.16417861e-01
-3.42317194e-01 4.61182058e-01 -1.76949248e-01 8.27868730e-02
4.88994092e-01 7.28142381e-01 -5.00198364e-01 -1.52522281e-01
5.93764603e-01 -1.58373326e-01 -4.17353332e-01 7.65517890e-01
2.52139837e-01 2.35983968e-01 8.82563591e-01 -1.06937313e+00
-1.00247681e+00 1.16679214e-01 -2.40576863e-01 4.15654331e-01
-2.16529801e-01 1.12272874e-01 8.32220256e-01 -1.18294919e+00
3.19240957e-01 9.78169560e-01 9.42205846e-01 1.50836051e-01
-1.07911789e+00 -3.87725204e-01 -1.34182693e-02 -2.24896953e-01
-1.67749834e+00 -2.69969553e-01 9.67778623e-01 -6.15801394e-01
1.36899137e+00 3.59377801e-01 7.57801712e-01 4.29820716e-01
3.33424538e-01 4.23136145e-01 5.43751001e-01 -1.65339008e-01
4.10526931e-01 -2.21372440e-01 4.48740572e-02 7.73232460e-01
3.01175416e-01 2.62555242e-01 -3.17965657e-01 -3.50453436e-01
1.47416520e+00 2.97844917e-01 -1.68808132e-01 -1.16577184e+00
-1.32528651e+00 7.94101536e-01 1.02030027e+00 1.53068691e-01
-5.89200258e-01 2.51811087e-01 2.64332831e-01 2.17007875e-01
7.22708642e-01 6.00132406e-01 -4.11044657e-01 1.73627689e-01
-9.03397441e-01 2.28702545e-01 4.51117486e-01 9.76532578e-01
1.01739776e+00 9.38914195e-02 -4.86263856e-02 7.26653039e-01
1.76488832e-01 4.11379158e-01 1.95600525e-01 -8.64301980e-01
3.14792722e-01 1.13239944e+00 -1.68670416e-02 -1.21208537e+00
-6.01886451e-01 -4.91228163e-01 -1.04910922e+00 6.97225332e-01
-1.00280508e-01 3.85338455e-01 -9.17559147e-01 1.17810392e+00
5.34246922e-01 5.58534622e-01 -1.30599737e-01 7.15908468e-01
8.66024792e-01 3.35832179e-01 -3.48153830e-01 6.43558919e-01
9.88239527e-01 -3.98377806e-01 2.80469626e-01 3.14523548e-01
6.38618529e-01 -3.60976636e-01 9.31461334e-01 -5.73284067e-02
-1.10022759e+00 -5.23133576e-01 -1.08482933e+00 -1.95198461e-01
-6.10256970e-01 1.09723717e-01 8.28429341e-01 4.62633967e-01
-1.50289297e+00 8.16328168e-01 -1.00817847e+00 -4.36326973e-02
8.33102584e-01 6.86846972e-01 -3.58395934e-01 3.42730224e-01
-6.06926858e-01 4.86152351e-01 3.40583295e-01 1.06458433e-01
-4.71074164e-01 -1.12472010e+00 -9.89672184e-01 2.66962320e-01
-2.51915872e-01 -8.40644419e-01 8.92125547e-01 -2.63654113e-01
-1.19076061e+00 1.10329676e+00 -7.85016418e-02 -4.35810536e-01
1.39234543e-01 -1.97393522e-02 -1.01362646e-01 2.75843620e-01
5.52292652e-02 6.17632926e-01 1.03942025e+00 -1.06468296e+00
-5.21628678e-01 -7.54970014e-01 1.34703830e-01 3.45387340e-01
-1.87692448e-01 -1.10857502e-01 -3.10321152e-01 -5.98943830e-01
5.85446239e-01 -7.87411153e-01 -3.42678368e-01 1.64810747e-01
-3.22557360e-01 -5.03414810e-01 8.84124160e-01 2.54932612e-01
7.57577300e-01 -2.29342413e+00 -1.06790304e-01 7.63042212e-01
7.93592751e-01 -8.88255015e-02 -8.45467150e-02 2.80805323e-02
-3.81393343e-01 7.40546510e-02 9.06703100e-02 -1.71373770e-01
-1.44195557e-02 -9.68602765e-03 -4.21280682e-01 7.12859929e-01
4.20802057e-01 1.19857931e+00 -6.53124154e-01 -1.61374569e-01
5.14936507e-01 7.19217598e-01 -7.16556907e-01 -1.40166674e-02
-1.86766118e-01 -2.87155006e-02 -7.22787559e-01 7.75195241e-01
9.83280718e-01 -4.70822901e-01 -6.77629113e-01 -3.62844467e-01
-3.22148889e-01 4.19147283e-01 -1.12161064e+00 1.63877439e+00
-3.02586108e-01 3.38943422e-01 -1.69005967e-03 -7.28956223e-01
1.10231352e+00 1.37642457e-03 8.32327366e-01 -2.31522828e-01
3.63642946e-02 9.45002064e-02 -2.66312420e-01 1.23176768e-01
5.07725537e-01 3.52128625e-01 2.02971354e-01 5.25148511e-01
-5.34753874e-02 -4.13289011e-01 -3.08901042e-01 1.92001816e-02
1.09789610e+00 -1.54665470e-01 -8.56218785e-02 -5.58058321e-01
3.33366126e-01 5.86868147e-04 -2.73425598e-02 7.35316277e-01
6.04897253e-02 6.85021639e-01 2.17971250e-01 -1.12815964e+00
-1.08963895e+00 -1.54076529e+00 -3.57373506e-01 1.03032374e+00
1.23237632e-01 -2.12113276e-01 -3.12597156e-01 -4.83546764e-01
5.28650701e-01 3.09603333e-01 -8.42437744e-01 -1.20420754e-01
-6.47600532e-01 -3.37516695e-01 5.22174358e-01 6.31456673e-01
2.01856270e-01 -1.02025652e+00 -9.31063890e-01 -3.01971491e-02
6.49098217e-01 -1.03380024e+00 -3.90633553e-01 5.42092741e-01
-1.24400067e+00 -1.06635284e+00 -3.21642339e-01 -7.76696801e-01
8.20696175e-01 5.38229823e-01 1.31489551e+00 3.32319252e-02
-1.07513152e-01 4.27413523e-01 -2.29809925e-01 -4.29105729e-01
7.30763003e-03 3.37087512e-01 2.50510812e-01 -1.71370655e-01
6.17439628e-01 -1.17092133e+00 -5.66544235e-01 8.73415470e-02
-8.49825084e-01 -1.21704467e-01 5.71552694e-01 5.22683203e-01
6.82943642e-01 9.19624940e-02 -2.61501014e-01 -5.97922146e-01
5.89385629e-01 -3.24762583e-01 -8.90256941e-01 -1.23876818e-02
-5.14549538e-02 1.83476046e-01 3.68833184e-01 -1.94838926e-01
-2.95984030e-01 1.18491411e-01 1.24708615e-01 -9.98371780e-01
-4.13989991e-01 2.10654482e-01 1.90886796e-01 -7.37332940e-01
9.02697086e-01 3.75553429e-01 -6.58135787e-02 -2.98916757e-01
5.97667694e-01 3.59619290e-01 3.85313481e-01 -8.52702260e-01
1.09546649e+00 8.71802628e-01 1.51492134e-01 -9.92299616e-01
-3.53192747e-01 -4.57254380e-01 -1.19283938e+00 1.21447608e-01
8.30606997e-01 -9.29353476e-01 -1.08034635e+00 2.32999608e-01
-1.40898120e+00 -1.47914872e-01 -6.73418760e-01 3.33380550e-01
-7.71960080e-01 -1.77989807e-02 -5.52785754e-01 -5.11208475e-01
-3.94829005e-01 -1.03699338e+00 1.33840406e+00 -9.50153768e-02
5.89484116e-03 -8.32273185e-01 8.14291090e-02 -4.64151919e-01
4.39403921e-01 2.52652228e-01 1.15652490e+00 -5.61440229e-01
-1.04716289e+00 -1.74222738e-01 -5.62461853e-01 8.41537938e-02
2.32360274e-01 -1.22470282e-01 -7.96627939e-01 -2.24944994e-01
-5.85062914e-02 5.36770783e-02 7.84308076e-01 5.75770259e-01
1.65469444e+00 -5.34426160e-02 -3.32475871e-01 1.27069271e+00
1.37595141e+00 -1.71550408e-01 3.53935331e-01 3.45757514e-01
9.34551656e-01 1.97668239e-01 -1.37743950e-01 4.25106347e-01
3.06592464e-01 2.82459795e-01 8.20528924e-01 -2.83480525e-01
3.41139346e-01 -1.08003337e-02 -3.15312952e-01 4.84396577e-01
-4.98107612e-01 3.27275634e-01 -1.11412883e+00 6.09354854e-01
-1.55195415e+00 -7.77416766e-01 6.97553754e-02 2.03794813e+00
4.00446951e-01 2.53963232e-01 -4.79164459e-02 -2.36334562e-01
3.52109104e-01 3.33440810e-01 -4.76080239e-01 -3.96139234e-01
-1.30344272e-01 6.31010473e-01 8.18278432e-01 2.46602267e-01
-1.41537273e+00 9.29679215e-01 6.35171032e+00 5.24868011e-01
-1.25116777e+00 -3.08337837e-01 3.80324304e-01 -1.48464203e-01
-2.78016537e-01 -4.95130271e-01 -5.94751596e-01 2.02974737e-01
4.46568191e-01 2.46278737e-02 3.43065113e-01 1.09660017e+00
1.27594233e-01 5.27741313e-01 -1.11358607e+00 1.00043464e+00
-2.78729230e-01 -1.88781035e+00 3.79753619e-01 8.91302824e-02
5.86403012e-01 8.50653887e-01 3.61863792e-01 -1.20423943e-01
5.71588516e-01 -1.33076406e+00 6.22666001e-01 2.95831978e-01
9.34409618e-01 -1.06424427e+00 2.38687471e-01 2.43682191e-01
-1.43319631e+00 -2.32572611e-02 -8.89764905e-01 9.25247371e-02
-3.34028572e-01 7.09028602e-01 -1.10734248e+00 4.16803479e-01
1.01813352e+00 8.59570861e-01 -4.94054049e-01 1.17296040e+00
1.80639714e-01 1.83826461e-02 -7.89545238e-01 3.88176814e-02
4.68682736e-01 -4.38510984e-01 6.19173229e-01 1.09053087e+00
4.10153955e-01 6.75826967e-02 -3.83460894e-02 1.25526917e+00
-2.34496947e-02 -9.15265232e-02 -1.14334893e+00 2.27584004e-01
7.20890880e-01 1.00192904e+00 -8.13912570e-01 -2.12188605e-02
-2.91046351e-01 7.86015570e-01 7.39158869e-01 4.32355702e-01
-3.71595114e-01 -4.02452528e-01 1.33206785e+00 3.51157576e-01
7.23391831e-01 -7.92704821e-01 -7.19848633e-01 -9.43486333e-01
9.67330933e-02 -3.31361443e-01 -6.98611215e-02 -6.63844109e-01
-1.21831429e+00 7.81732261e-01 -2.61886090e-01 -1.30035281e+00
7.85175711e-02 -9.18378055e-01 -8.17722738e-01 9.74589646e-01
-1.61179733e+00 -1.33219314e+00 -2.68849015e-01 8.51183295e-01
1.30659752e-02 -1.77391201e-01 8.80083382e-01 -2.12714472e-03
1.00462027e-01 3.71598035e-01 1.56476587e-01 3.35269153e-01
3.66694182e-02 -1.40429688e+00 9.78991032e-01 3.62057835e-01
2.77765453e-01 1.05088139e+00 2.37519406e-02 -5.69322526e-01
-1.35977852e+00 -1.42424583e+00 5.88053644e-01 -7.87241220e-01
6.35831535e-01 -6.01175487e-01 -1.02146471e+00 6.31897569e-01
-4.09467846e-01 3.82884830e-01 5.71875930e-01 2.59523809e-01
-6.40256345e-01 -2.99464464e-02 -1.08863151e+00 4.87111926e-01
1.25146127e+00 -7.98511863e-01 -4.83577341e-01 3.54198635e-01
9.78035867e-01 -7.08377957e-01 -8.59480619e-01 7.67252624e-01
2.04440340e-01 -1.05928278e+00 1.54910636e+00 -8.11684072e-01
2.17867792e-01 -4.59804863e-01 -1.39502794e-01 -1.20532191e+00
-7.50319064e-01 -2.29621261e-01 -6.97662309e-02 3.27464163e-01
3.82230908e-01 -5.32825053e-01 1.32487392e+00 3.28409076e-01
-4.06770796e-01 -9.23751235e-01 -1.22222137e+00 -6.92333639e-01
1.28069460e-01 -7.47943461e-01 1.04939592e+00 1.08867300e+00
-4.39525902e-01 -1.09118789e-01 4.18502510e-01 7.75907218e-01
7.16852725e-01 2.93140888e-01 6.27277374e-01 -1.65190411e+00
2.02195510e-01 -7.69651711e-01 -9.44546223e-01 -1.16481078e+00
1.23572171e-01 -1.14324284e+00 -3.39924246e-01 -1.10740912e+00
-3.53406012e-01 -9.74542499e-01 -4.55141068e-01 3.78298640e-01
2.16172099e-01 4.56813216e-01 -1.62875041e-01 3.03875834e-01
-5.05954325e-01 6.27499998e-01 1.00624216e+00 -3.23502682e-02
-4.72504646e-01 1.20227531e-01 -3.07965428e-01 6.51745975e-01
7.27887452e-01 -8.77252594e-02 -3.88482541e-01 -9.41588759e-01
2.31237248e-01 -5.61565638e-01 7.11541653e-01 -1.17093110e+00
4.96953636e-01 -8.48823190e-02 7.50079572e-01 -1.07084656e+00
5.06899893e-01 -1.11081111e+00 -3.49930264e-02 7.58046806e-02
-3.29560675e-02 3.01154375e-01 3.85341048e-01 3.72941822e-01
5.29566966e-02 4.25646693e-01 7.40423620e-01 -1.19748175e-01
-5.50526083e-01 9.10943091e-01 6.33953884e-02 -3.86032075e-01
8.91746879e-01 -6.97212577e-01 1.80696324e-01 1.94309801e-01
-6.71361685e-01 1.73448831e-01 8.67059171e-01 2.95502007e-01
1.08513677e+00 -1.58320010e+00 -5.36436617e-01 8.05712044e-01
1.71356991e-01 6.75662935e-01 1.31037712e-01 4.99679565e-01
-9.26231027e-01 4.59826052e-01 -2.50271142e-01 -1.17448294e+00
-8.78144324e-01 5.53383112e-01 5.83037913e-01 5.85944392e-02
-8.82204950e-01 1.03268397e+00 3.47299367e-01 -8.55596542e-01
9.72730145e-02 -8.83463144e-01 -8.34498778e-02 -5.47655880e-01
2.38187805e-01 7.28706196e-02 4.19499487e-01 -4.67207193e-01
-3.91466200e-01 6.79269433e-01 3.76135670e-02 3.78326446e-01
1.59956944e+00 2.69355923e-01 -4.19197112e-01 2.89982498e-01
1.21777654e+00 -9.16173384e-02 -1.03481185e+00 -3.64336640e-01
6.84038401e-02 -5.84732652e-01 1.46973684e-01 -1.82069331e-01
-8.82574975e-01 7.48398066e-01 4.87130046e-01 3.17556620e-01
8.11659336e-01 2.52806067e-01 3.97953480e-01 5.94968915e-01
3.04997861e-01 -5.85637331e-01 -2.46446863e-01 9.97669160e-01
8.33935976e-01 -1.04937804e+00 -2.22772196e-01 -4.59766090e-01
1.37252584e-02 1.34902442e+00 4.39025253e-01 -6.53532743e-01
1.13160050e+00 3.65992963e-01 -1.32509843e-01 -7.88156807e-01
-4.72051919e-01 -8.25549141e-02 6.04261577e-01 8.97588849e-01
1.18699290e-01 1.49628237e-01 6.08833790e-01 2.18835175e-01
-5.74318051e-01 -1.40027910e-01 -1.39985457e-01 7.47587681e-01
-6.77591324e-01 -5.73710442e-01 -4.20425206e-01 6.53057992e-01
9.60327834e-02 -1.87445059e-01 -2.61248678e-01 6.78422987e-01
1.28079414e-01 7.76987970e-02 8.08943868e-01 -3.22143465e-01
4.66383845e-01 -1.59133226e-01 4.18395966e-01 -7.32545614e-01
-2.51299024e-01 -2.91373461e-01 -5.96160412e-01 -8.00002933e-01
-2.21779644e-01 -3.43079120e-01 -1.26504791e+00 -5.11204004e-01
1.10076733e-01 2.56184816e-01 6.56528592e-01 5.54859042e-01
8.38232219e-01 3.64934474e-01 7.23761499e-01 -1.28403652e+00
-3.71421516e-01 -5.67724645e-01 -6.43727422e-01 1.35949522e-01
5.30182064e-01 -6.77103996e-01 -1.13726169e-01 -3.06945890e-01] | [7.944940090179443, -3.6841373443603516] |
6edabdae-ed99-4332-864c-d2951f05477d | hybrid-classical-quantum-deep-learning-models | 2108.01125 | null | https://arxiv.org/abs/2108.01125v1 | https://arxiv.org/pdf/2108.01125v1.pdf | Hybrid Classical-Quantum Deep Learning Models for Autonomous Vehicle Traffic Image Classification Under Adversarial Attack | Image classification must work for autonomous vehicles (AV) operating on public roads, and actions performed based on image misclassification can have serious consequences. Traffic sign images can be misclassified by an adversarial attack on machine learning models used by AVs for traffic sign recognition. To make classification models resilient against adversarial attacks, we used a hybrid deep-learning model with both the quantum and classical layers. Our goal is to study the hybrid deep-learning architecture for classical-quantum transfer learning models to support the current era of intermediate-scale quantum technology. We have evaluated the impacts of various white box adversarial attacks on these hybrid models. The classical part of hybrid models includes a convolution network from the pre-trained Resnet18 model, which extracts informative features from a high dimensional LISA traffic sign image dataset. The output from the classical processor is processed further through the quantum layer, which is composed of various quantum gates and provides support to various quantum mechanical features like entanglement and superposition. We have tested multiple combinations of quantum circuits to provide better classification accuracy with decreasing training data and found better resiliency for our hybrid classical-quantum deep learning model during attacks compared to the classical-only machine learning models. | ['Mashrur Chowdhury', 'Dimitra Michalaka', 'Judith Mwakalonge', 'Gurcan Comert', 'Frank Ngeni', 'Zadid Khan', 'Fahim Ahmed', 'Sakib Mahmud Khan', 'Reek Majumder'] | 2021-08-02 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 3.52526754e-01 2.97322929e-01 1.17381059e-01 -2.02756017e-01
-7.71158516e-01 -6.02405787e-01 6.55407429e-01 -6.33756340e-01
-6.10842645e-01 4.57010239e-01 -6.03636801e-01 -1.05228019e+00
3.45110357e-01 -1.24035740e+00 -9.70522523e-01 -9.07754898e-01
1.80528332e-02 2.49412477e-01 7.50425577e-01 -6.96230412e-01
4.74483520e-01 8.22634876e-01 -1.53332853e+00 6.56229556e-01
6.15589261e-01 9.17351425e-01 -6.61016643e-01 1.30578899e+00
1.40836552e-01 1.11715102e+00 -7.14569747e-01 -7.46858895e-01
7.47319162e-01 -4.99418199e-01 -6.77062333e-01 -7.30880022e-01
8.58039558e-01 -4.31149840e-01 -1.30531156e+00 1.48667109e+00
3.72246206e-01 -1.67264089e-01 6.59282982e-01 -1.87782538e+00
-6.10144496e-01 1.61135718e-01 5.82644999e-01 2.91892141e-01
-1.36465609e-01 1.09901714e+00 6.91617846e-01 -5.16166277e-02
7.39790082e-01 1.22237873e+00 7.93616474e-01 9.08582866e-01
-1.10669732e+00 -1.26010656e+00 -1.06722486e+00 1.03916347e+00
-1.03209782e+00 -5.36341071e-01 2.24553213e-01 -7.29760751e-02
1.42288184e+00 -5.75262047e-02 2.50932604e-01 9.30270612e-01
9.97896254e-01 3.49923670e-01 1.36118186e+00 -1.63434878e-01
2.86937147e-01 2.25640163e-01 2.13596761e-01 1.12287772e+00
1.89359844e-01 1.05372608e+00 -2.20522299e-01 1.15004519e-03
1.13742910e-01 -3.40979666e-01 4.31024581e-01 7.41326287e-02
-6.02565885e-01 8.68504584e-01 9.56533074e-01 -9.21150595e-02
-1.03325270e-01 9.51401055e-01 5.92782319e-01 1.02122235e+00
-4.74512607e-01 5.08802116e-01 -2.97883123e-01 -8.38039741e-02
-4.75841492e-01 1.91740498e-01 8.42673361e-01 6.98127687e-01
1.15212715e+00 1.57347307e-01 -2.41952851e-01 -3.57660621e-01
2.36898184e-01 1.42245054e+00 1.83797181e-01 -9.78907168e-01
3.06127518e-01 2.41409495e-01 -2.73527175e-01 -6.25025630e-01
-4.71239835e-01 7.72790685e-02 -5.60960948e-01 9.26182926e-01
5.05766273e-01 -1.92183688e-01 -1.47418714e+00 1.42737496e+00
-1.41082913e-01 5.08810103e-01 7.28298843e-01 6.57523572e-01
7.67496169e-01 5.69309711e-01 2.70625412e-01 6.38372421e-01
1.10297894e+00 -6.27209306e-01 -5.15880048e-01 1.33308917e-01
1.36870682e+00 -5.82598507e-01 4.07539159e-01 -8.91841725e-02
-6.09596670e-01 -5.35572708e-01 -1.55948508e+00 -1.31469265e-01
-1.13749039e+00 -4.80945379e-01 4.63874251e-01 1.50205803e+00
-1.03879333e+00 1.00338531e+00 -7.80696511e-01 -1.41503602e-01
7.39131212e-01 8.07999134e-01 -5.08247316e-01 -1.13838933e-01
-1.85207832e+00 1.51676607e+00 1.96823403e-01 9.36621204e-02
-9.56592500e-01 -1.42843261e-01 -6.58493042e-01 -8.79658163e-02
-1.64517075e-01 -6.29147708e-01 1.17008626e+00 -8.56162786e-01
-1.64024365e+00 7.81376362e-01 -1.50419967e-02 -1.05728018e+00
3.67750943e-01 6.28196359e-01 -7.40430415e-01 3.32621723e-01
-2.37553671e-01 6.31340623e-01 1.08964276e+00 -6.90178573e-01
-5.28588295e-01 -2.18163729e-01 2.17024490e-01 -4.41746324e-01
3.01834732e-01 -6.09191619e-02 5.69528639e-01 3.66751939e-01
-6.37838468e-02 -1.51096916e+00 -1.92469612e-01 -2.21981928e-01
-1.75265580e-01 6.48701414e-02 1.16406250e+00 -1.74732193e-01
5.11492074e-01 -2.08928847e+00 -5.96808672e-01 5.76318324e-01
-3.64165567e-02 9.31286335e-01 -4.60963309e-01 4.50361550e-01
-1.10870980e-01 2.67623693e-01 2.55733300e-02 1.28896222e-01
2.64966100e-01 5.42223692e-01 -6.34395063e-01 7.22068191e-01
5.96569657e-01 1.54326797e+00 -8.25774491e-01 -2.63641775e-01
2.54934996e-01 1.53737113e-01 -4.40362126e-01 -2.35678107e-01
7.18488246e-02 2.85758018e-01 -4.52878326e-01 4.70128179e-01
9.00118351e-01 1.23142086e-01 -2.86486328e-01 -1.28300563e-01
1.47526562e-01 5.01018047e-01 -7.59046733e-01 8.63130450e-01
-3.19145828e-01 1.16332603e+00 -3.38172406e-01 -7.94617951e-01
4.76830751e-01 1.73120871e-01 -1.31839901e-01 -1.33461893e+00
5.39805770e-01 5.27178168e-01 6.76239312e-01 -5.96536100e-01
3.80111158e-01 -3.93300593e-01 -2.58446842e-01 4.06744421e-01
1.80786729e-01 -5.21879792e-01 4.22690921e-02 2.53236920e-01
1.65692604e+00 -1.59075618e-01 -5.30897081e-01 2.14405492e-01
7.04802036e-01 3.47821116e-01 2.49371260e-01 1.20809007e+00
-9.72248375e-01 1.09185264e-01 6.70269847e-01 -6.03956163e-01
-1.29642677e+00 -1.24653792e+00 -2.96421796e-01 7.10535347e-01
4.41397458e-01 -2.60857251e-02 -5.08288920e-01 -7.90666878e-01
8.19168240e-02 9.61922944e-01 -5.06637990e-01 -1.11885166e+00
-6.23521507e-01 -7.20084012e-01 1.63775253e+00 3.32782716e-01
1.00874996e+00 -1.12247670e+00 -5.59384704e-01 -3.31627168e-02
4.69765812e-01 -1.32758760e+00 1.92684352e-01 4.54193771e-01
-4.22635406e-01 -1.39737415e+00 2.93164074e-01 -5.54114401e-01
5.21602333e-01 -8.85547325e-02 3.85287911e-01 1.53377101e-01
-1.94971576e-01 9.60825011e-03 -5.03171504e-01 -4.69535917e-01
-1.63925457e+00 -1.26229107e-01 1.16842516e-01 2.71891756e-03
7.43750632e-01 -6.24313653e-02 -3.18920523e-01 1.98286220e-01
-1.12453449e+00 -3.50534499e-01 7.60184526e-01 9.52846646e-01
-1.70855865e-01 1.27718464e-01 1.29799113e-01 -6.74288571e-01
2.25186557e-01 -5.92264533e-02 -7.79738784e-01 1.07545905e-01
-4.47048813e-01 6.19556367e-01 8.77392769e-01 -1.52758822e-01
-6.18954122e-01 -2.47272407e-03 -3.88160348e-01 -2.63336629e-01
-7.04326853e-02 -8.51411521e-02 1.11639768e-01 -1.10947430e+00
1.06270707e+00 2.04886943e-01 3.07130486e-01 6.86915219e-01
5.23941040e-01 9.02403176e-01 3.69570762e-01 1.99362278e-01
1.60752928e+00 6.06230319e-01 9.14352715e-01 -8.49279225e-01
-3.25144738e-01 7.61560276e-02 -6.37793183e-01 -2.30976030e-01
1.04399371e+00 -3.93701196e-01 -1.09346080e+00 1.01337266e+00
-1.19138396e+00 -4.97926861e-01 -2.10815325e-01 4.89607304e-01
-6.14667833e-01 2.83375114e-01 -8.92227113e-01 -6.33928001e-01
-6.41746596e-02 -1.44223666e+00 7.36499310e-01 3.58944416e-01
4.34639007e-01 -6.86062932e-01 -5.10222018e-02 4.81827110e-01
7.65349269e-01 -1.87091921e-02 9.37749147e-01 -7.57390022e-01
-1.07470155e+00 -7.14997113e-01 -4.21271801e-01 6.91675127e-01
-5.37091136e-01 4.71405163e-02 -1.39623082e+00 -1.50199071e-01
-3.83817047e-01 -6.28105938e-01 1.16769755e+00 -2.11960554e-01
5.63347518e-01 -4.43858206e-02 -7.62595013e-02 7.90972173e-01
1.13257563e+00 2.72093385e-01 1.31066418e+00 4.23795760e-01
5.02716184e-01 -4.52158004e-02 3.04434121e-01 -4.45146322e-01
1.84533343e-01 3.80548567e-01 6.69711113e-01 1.05216287e-01
-2.59468347e-01 -7.08476231e-02 8.71850133e-01 2.05285311e-01
1.55863941e-01 1.71837732e-01 -9.23616409e-01 -1.75496563e-01
-1.45297790e+00 -1.57529235e+00 -4.44810539e-01 1.99650466e+00
3.81572932e-01 5.80615282e-01 -3.27177495e-01 1.17341511e-01
3.81090850e-01 -7.30442107e-02 -5.76694965e-01 -1.21170080e+00
-2.77663052e-01 7.60443211e-01 1.60055721e+00 4.21877474e-01
-1.36306262e+00 1.41283333e+00 6.21433640e+00 7.79291153e-01
-1.39070523e+00 2.78072953e-01 1.11883141e-01 4.09120977e-01
7.91087896e-02 3.11133832e-01 -5.93210280e-01 2.62180597e-01
1.96613610e+00 1.26342356e-01 5.70983231e-01 3.81630808e-01
-1.72187120e-01 2.04571653e-02 -8.54164422e-01 5.32928705e-01
-1.41270667e-01 -1.45260417e+00 -8.47895741e-02 1.33735448e-01
8.48302066e-01 8.04446638e-01 3.05505067e-01 8.83503616e-01
4.30894196e-01 -1.10729098e+00 5.18486500e-01 5.18863082e-01
9.11015153e-01 -5.95164716e-01 1.17439616e+00 9.77304205e-02
-6.52627885e-01 -4.49129283e-01 -3.47249776e-01 -3.57950665e-02
-7.40200877e-02 -5.43404937e-01 -9.46567118e-01 3.32387626e-01
2.81927198e-01 1.84765413e-01 -7.71473765e-01 7.51833260e-01
-4.36261266e-01 7.46834040e-01 -3.63192618e-01 -4.11503762e-01
7.90011227e-01 -3.59588489e-02 5.28644562e-01 9.66051638e-01
-1.31080464e-01 -4.62273546e-02 -2.99517244e-01 9.35357392e-01
1.98021457e-02 -5.30451000e-01 -9.77848470e-01 -2.32669279e-01
1.14901513e-01 8.17821622e-01 -5.69501698e-01 -6.39945090e-01
-2.63998896e-01 1.12722540e+00 -2.88181782e-01 3.43075246e-01
-1.10770941e+00 -7.27301776e-01 6.75512016e-01 -2.35602483e-01
2.75460094e-01 -3.03202957e-01 -2.65912414e-01 -1.03007376e+00
-5.06203175e-01 -6.20814264e-01 -4.04073037e-02 -6.78593934e-01
-1.02390873e+00 3.70041072e-01 -4.84230518e-01 -1.22789693e+00
-1.23486593e-01 -1.19911742e+00 -7.72975862e-01 1.01229489e+00
-1.68393314e+00 -1.16570842e+00 1.54375643e-01 6.20046675e-01
-4.03690577e-01 -5.12195706e-01 1.02446628e+00 1.70627043e-01
-3.85038525e-01 9.37679708e-01 2.94728637e-01 4.89957005e-01
4.95541155e-01 -9.95128155e-01 8.58634472e-01 1.05075276e+00
-1.49081517e-02 2.95220345e-01 6.44001961e-01 -5.43198824e-01
-1.84436667e+00 -1.16148305e+00 8.74301314e-01 -7.78797567e-01
1.07473564e+00 -7.34950528e-02 -4.87678707e-01 5.78736246e-01
-1.12217247e-01 2.96465218e-01 4.13988352e-01 -5.65117657e-01
-8.20830047e-01 -2.02749640e-01 -1.39006543e+00 6.19991720e-01
5.76632798e-01 -1.41668046e+00 -6.74800456e-01 3.18896711e-01
3.63669544e-01 -2.89636374e-01 -3.31014901e-01 9.03669670e-02
6.63800240e-01 -7.32254028e-01 6.59166157e-01 -1.14202964e+00
-3.25830095e-03 -5.18902361e-01 -1.09437540e-01 -1.07145476e+00
4.47209477e-02 -7.36799955e-01 3.01292807e-01 8.79755169e-02
6.03291214e-01 -1.28013825e+00 1.04025197e+00 7.91914821e-01
-2.52978534e-01 1.50039166e-01 -1.51894772e+00 -9.95497406e-01
4.26822871e-01 -8.48789871e-01 3.11574340e-01 4.49464411e-01
-1.05835944e-01 2.03202963e-01 -4.66286987e-02 5.64491332e-01
7.12818146e-01 -2.72784382e-01 1.00996494e+00 -6.27718091e-01
1.18734509e-01 -6.76718533e-01 -1.57968593e+00 -4.63853002e-01
3.61914217e-01 -1.44202352e+00 4.10261601e-02 -8.01060319e-01
-2.81544656e-01 -3.26587588e-01 -4.96574908e-01 4.77675200e-01
1.56619236e-01 6.78639114e-01 3.52153689e-01 7.72856921e-02
-4.04474527e-01 3.33716452e-01 1.22415400e+00 -6.17168367e-01
1.87669218e-01 1.03138730e-01 6.99420646e-02 3.66312802e-01
7.59419441e-01 -8.70840728e-01 8.56248960e-02 3.23272981e-02
2.01853335e-01 -2.51618415e-01 8.82440448e-01 -1.41807163e+00
6.27623737e-01 1.96973130e-01 -4.20136191e-02 -2.08603173e-01
2.40273193e-01 -8.24912190e-01 -2.65293241e-01 1.24199474e+00
-9.57267284e-02 -4.78346825e-01 2.12653428e-01 3.33511293e-01
-2.05084812e-02 -2.47103617e-01 1.05344915e+00 2.25689858e-01
-1.00335979e+00 2.66741991e-01 -7.14257717e-01 -4.64414597e-01
1.08321679e+00 -4.36674207e-01 -8.55881095e-01 -1.27501518e-01
-6.55843019e-01 1.49604917e-01 4.88137275e-01 3.23210716e-01
4.71556783e-01 -1.10108030e+00 -4.96482104e-01 6.07018054e-01
2.01234698e-01 -9.28458512e-01 2.58937299e-01 8.29530478e-01
-1.24964011e+00 7.64775574e-01 -6.86953366e-01 -3.61035556e-01
-9.87715840e-01 5.37351549e-01 9.86940265e-01 9.90619138e-02
-2.48345852e-01 5.60769081e-01 -5.58447659e-01 -5.48892319e-01
-1.36868462e-01 -2.91872829e-01 2.20516801e-01 -3.91828209e-01
3.09413433e-01 4.52743292e-01 2.75424987e-01 -1.07897878e+00
-3.85713100e-01 3.71354282e-01 1.49880871e-01 -1.94306299e-01
7.05522895e-01 3.85946065e-01 7.93271661e-02 3.59796882e-02
1.61523736e+00 -6.32982790e-01 -8.14498723e-01 -4.25174758e-02
-1.74228340e-01 -6.45568371e-02 2.09771082e-01 -7.19780862e-01
-6.96896195e-01 1.27095175e+00 1.10531795e+00 4.18169558e-01
6.85960889e-01 -4.70828921e-01 1.10093939e+00 1.22104633e+00
5.39300203e-01 -1.12592757e+00 -3.34657162e-01 8.96049201e-01
9.97247472e-02 -1.41072857e+00 -4.66087550e-01 1.43871352e-01
-3.32798690e-01 1.46226311e+00 3.19248378e-01 -5.51244974e-01
7.60861158e-01 3.90277505e-02 4.24604386e-01 -3.04061890e-01
-6.08760595e-01 -5.70839286e-01 8.49879440e-03 5.94621778e-01
-4.10595983e-01 2.41131186e-01 9.15263966e-02 -2.73515344e-01
-3.33688647e-01 -2.60921419e-02 8.17370892e-01 1.00883985e+00
-5.03985345e-01 -1.18577433e+00 -4.27682579e-01 4.33535576e-01
-1.64234236e-01 -1.50440574e-01 -4.01947528e-01 6.55468047e-01
4.56955403e-01 1.04362917e+00 -3.86834666e-02 -1.01436317e+00
3.62440825e-01 4.42357183e-01 5.30002117e-01 -1.38386846e-01
-6.41316533e-01 -1.10767233e+00 1.89351499e-01 -8.76307309e-01
-7.62441829e-02 -6.69203460e-01 -1.69162929e+00 -7.87098587e-01
-4.51546818e-01 -2.37243488e-01 9.84851360e-01 1.14453733e+00
2.44421586e-01 4.25053716e-01 7.84555674e-01 -7.28547156e-01
-1.16082489e+00 -7.81824291e-01 -3.72010618e-01 5.19882917e-01
6.09035492e-01 -3.42479527e-01 -5.36995053e-01 -1.78652272e-01] | [5.628922462463379, 5.051033020019531] |
65a96c84-e0f4-4c28-9c5a-3e7b3651c970 | a-comprehensive-empirical-analysis-on-cross | 2106.12797 | null | https://arxiv.org/abs/2106.12797v1 | https://arxiv.org/pdf/2106.12797v1.pdf | A comprehensive empirical analysis on cross-domain semantic enrichment for detection of depressive language | We analyze the process of creating word embedding feature representations designed for a learning task when annotated data is scarce, for example, in depressive language detection from Tweets. We start with a rich word embedding pre-trained from a large general dataset, which is then augmented with embeddings learned from a much smaller and more specific domain dataset through a simple non-linear mapping mechanism. We also experimented with several other more sophisticated methods of such mapping including, several auto-encoder based and custom loss-function based methods that learn embedding representations through gradually learning to be close to the words of similar semantics and distant to dissimilar semantics. Our strengthened representations better capture the semantics of the depression domain, as it combines the semantics learned from the specific domain coupled with word coverage from the general language. We also present a comparative performance analyses of our word embedding representations with a simple bag-of-words model, well known sentiment and psycholinguistic lexicons, and a general pre-trained word embedding. When used as feature representations for several different machine learning methods, including deep learning models in a depressive Tweets identification task, we show that our augmented word embedding representations achieve a significantly better F1 score than the others, specially when applied to a high quality dataset. Also, we present several data ablation tests which confirm the efficacy of our augmentation techniques. | ['Osmar Zaiane', 'Randy Goebel', 'Nawshad Farruque'] | 2021-06-24 | null | null | null | null | ['data-ablation'] | ['computer-vision'] | [ 3.73815969e-02 4.75711197e-01 -2.25454494e-01 -5.74930429e-01
-7.92598128e-01 -1.71756238e-01 7.36230314e-01 7.94455886e-01
-8.35182667e-01 5.80122650e-01 8.01586986e-01 7.55296424e-02
1.12122901e-01 -1.03632176e+00 -3.56024176e-01 -4.76782739e-01
-8.02930072e-02 7.20157623e-01 -8.42376798e-02 -8.67239714e-01
-1.07712433e-01 2.39880621e-01 -1.40896511e+00 2.07335517e-01
5.00386298e-01 5.16041577e-01 -9.91735458e-02 4.19604152e-01
-4.45205063e-01 3.47136468e-01 -5.27181983e-01 -5.27823329e-01
-1.15550399e-01 -2.33002037e-01 -7.60564804e-01 -1.98695600e-01
5.38754724e-02 -6.17675483e-02 -4.38598752e-01 8.43042910e-01
7.94953585e-01 2.39748824e-02 8.49499404e-01 -9.70193923e-01
-1.16629899e+00 6.85538888e-01 -3.55363041e-01 1.85636163e-01
3.89043659e-01 -6.96772709e-02 1.37462962e+00 -9.28530931e-01
6.47631764e-01 1.36392784e+00 9.02112961e-01 7.49822319e-01
-1.42175364e+00 -6.14366710e-01 1.36734158e-01 -3.15917097e-02
-1.21818233e+00 -1.25858471e-01 6.66079402e-01 -4.61621135e-01
1.21537519e+00 -7.62592405e-02 5.17732799e-01 1.51584208e+00
7.76619688e-02 6.28416657e-01 6.02773011e-01 -5.81016541e-01
-6.47932962e-02 7.82997370e-01 5.14539003e-01 7.18609333e-01
3.89283687e-01 -1.16558030e-01 -4.44590390e-01 -4.88875479e-01
7.65706226e-02 2.43577495e-01 -4.69999462e-02 -3.53595644e-01
-1.01025391e+00 1.52541196e+00 4.01096135e-01 6.85975194e-01
-3.11767042e-01 1.22825436e-01 8.25714529e-01 3.58879924e-01
9.65545058e-01 5.25388658e-01 -6.15117550e-01 1.67284012e-01
-5.81363261e-01 2.39611298e-01 6.63913190e-01 5.60478091e-01
1.00859571e+00 -9.03498977e-02 -3.71799797e-01 1.03420746e+00
2.19913065e-01 3.62888396e-01 1.32494092e+00 -2.71828379e-02
2.50830710e-01 8.02044988e-01 -1.61334962e-01 -1.06905639e+00
-7.17126787e-01 -2.82901943e-01 -6.51017606e-01 -1.86371610e-01
-2.73515675e-02 -2.07153812e-01 -6.65681541e-01 2.20859218e+00
9.79654714e-02 -1.06157251e-02 3.61075014e-01 5.34037888e-01
1.02739644e+00 5.56015968e-01 2.67965287e-01 1.46918103e-01
1.81176984e+00 -7.70817339e-01 -7.63698220e-01 -5.32355487e-01
1.33980227e+00 -3.49432141e-01 1.37150586e+00 -3.08477227e-02
-6.68415964e-01 -5.17196894e-01 -1.05900419e+00 -3.68724018e-01
-1.08985639e+00 -1.23524494e-01 7.25341082e-01 6.83049440e-01
-1.12676680e+00 4.25205708e-01 -3.93791288e-01 -7.43839443e-01
3.06043267e-01 3.53227943e-01 -7.27019012e-01 -9.96198654e-02
-1.78113556e+00 1.05921710e+00 4.92652804e-01 -6.36199594e-01
-5.69664299e-01 -8.09068680e-01 -1.41318572e+00 1.70993507e-01
-2.42700502e-01 -6.27289772e-01 7.47659326e-01 -1.00195408e+00
-1.00926459e+00 1.53697658e+00 1.60183273e-02 -5.36025584e-01
-1.41145319e-01 -2.22373873e-01 -5.17069161e-01 -1.33676469e-01
3.26932758e-01 6.81401253e-01 7.30767131e-01 -8.67554784e-01
-1.17899612e-01 -3.40040654e-01 1.40148088e-01 1.54469525e-02
-1.17989171e+00 -2.96363775e-02 -5.09246485e-03 -8.69198501e-01
-4.71815377e-01 -6.81933165e-01 -2.80417860e-01 6.97848573e-02
-2.13727921e-01 -3.90541852e-01 5.10954618e-01 -5.49942672e-01
1.17486429e+00 -2.37635803e+00 1.17737390e-01 1.48015961e-01
5.42132080e-01 2.55692333e-01 -6.40344381e-01 8.32015693e-01
-5.79165161e-01 2.22023711e-01 -2.49615341e-01 -5.00568807e-01
1.61920443e-01 3.73983860e-01 -1.99098840e-01 5.10223746e-01
5.96462250e-01 9.44199741e-01 -1.18150795e+00 -2.31776834e-01
-3.19129527e-02 6.12853050e-01 -8.64736080e-01 1.89256966e-01
1.19694196e-01 -1.77383631e-01 -3.45496505e-01 1.83449343e-01
3.42801273e-01 -1.97024629e-01 1.48284882e-01 -2.05520466e-02
3.88497710e-01 3.54158461e-01 -7.48951495e-01 1.80055928e+00
-9.34272587e-01 6.93466127e-01 -3.66257429e-01 -1.33214080e+00
1.15692496e+00 2.89315373e-01 4.99095082e-01 -6.42628968e-01
1.87963605e-01 3.69866304e-02 2.68247537e-02 -5.67729115e-01
6.16831064e-01 -5.55641353e-01 -4.01940316e-01 7.78738737e-01
6.05533957e-01 2.89410278e-02 9.72051453e-03 4.19132948e-01
1.28985560e+00 -4.57451433e-01 6.44114196e-01 -3.83633703e-01
5.69466233e-01 -2.06159666e-01 1.56995893e-01 4.87826675e-01
-9.25836191e-02 6.54213548e-01 7.23486006e-01 -4.94165808e-01
-9.80857849e-01 -8.45244646e-01 -4.81439114e-01 1.38596332e+00
-2.30523422e-01 -9.31386828e-01 -3.73770058e-01 -7.44818687e-01
3.01885217e-01 7.57884920e-01 -1.19342530e+00 -9.52362359e-01
-1.08602270e-01 -1.11413205e+00 6.35642052e-01 6.10188425e-01
-1.66242003e-01 -1.21157038e+00 -1.81575358e-01 2.47950971e-01
8.83958191e-02 -1.00174701e+00 -2.23199040e-01 5.35091162e-01
-5.30085087e-01 -8.00097823e-01 -6.81294501e-01 -9.66589987e-01
5.39887130e-01 -1.99987143e-01 1.46994615e+00 2.11035565e-01
-4.04945433e-01 4.68391627e-01 -6.64416432e-01 -6.36976898e-01
-4.90355521e-01 2.06629470e-01 3.70815963e-01 6.92862421e-02
1.06878734e+00 -4.23654765e-01 -4.07931536e-01 -3.30148906e-01
-1.20946980e+00 -3.78413260e-01 3.66092294e-01 1.18997729e+00
1.69283807e-01 -3.50757688e-01 8.20749164e-01 -1.14878392e+00
1.17460322e+00 -9.78301883e-01 2.25874782e-01 -1.64151877e-01
-5.15523612e-01 3.33296776e-01 4.30658489e-01 -6.44700408e-01
-2.90979117e-01 -1.69180438e-01 -5.98329604e-01 -1.01375654e-01
1.53198969e-02 4.87041235e-01 9.52626243e-02 3.49563897e-01
9.89446282e-01 2.77445257e-01 2.83176452e-01 -4.31524873e-01
6.91810787e-01 8.61783981e-01 3.59523371e-02 -3.37006539e-01
7.75539637e-01 4.16158140e-01 -5.15221179e-01 -9.51680839e-01
-8.25782716e-01 -5.77270508e-01 -5.32567084e-01 5.51043749e-01
9.96916473e-01 -9.77519095e-01 -2.24792883e-01 -9.89301130e-02
-1.18800128e+00 -2.38799043e-02 -7.67065108e-01 4.83252883e-01
-3.82846773e-01 2.38803566e-01 -5.31958938e-01 -4.60073382e-01
-4.31692421e-01 -8.16119432e-01 1.28480816e+00 -2.16416985e-01
-7.90088236e-01 -1.55861855e+00 7.39404619e-01 -2.48201698e-01
3.72149259e-01 1.96040154e-01 1.34457040e+00 -1.48578060e+00
7.76174724e-01 -6.19241476e-01 -3.10268462e-01 6.55282795e-01
3.54146183e-01 -3.83200735e-01 -1.11516333e+00 -3.05968374e-01
-2.24240288e-01 -6.59723103e-01 1.07555437e+00 8.89257342e-02
1.03711772e+00 -2.33339131e-01 -3.77932698e-01 4.84449089e-01
1.56235182e+00 -4.00291771e-01 4.92567003e-01 4.76895601e-01
5.28482556e-01 7.49691784e-01 1.81492150e-01 5.99148035e-01
3.21807981e-01 5.19828022e-01 2.50266403e-01 -4.70936239e-01
1.67524531e-01 -1.66602224e-01 4.69946057e-01 8.44404578e-01
4.53711271e-01 -2.09575266e-01 -8.69735718e-01 9.59210992e-01
-1.59351420e+00 -6.72023654e-01 1.63777053e-01 1.77011836e+00
1.00973940e+00 1.17548637e-01 1.95239201e-01 2.71442473e-01
4.04286951e-01 2.70037055e-01 -2.66372412e-01 -1.02445889e+00
-1.67464629e-01 8.11013401e-01 2.50232756e-01 4.19986576e-01
-1.02826142e+00 1.10044384e+00 6.69369173e+00 7.03747511e-01
-9.24850225e-01 4.70937133e-01 3.01062286e-01 -2.43960191e-02
-7.32852399e-01 -4.35318977e-01 -6.38515115e-01 3.72621775e-01
1.16504514e+00 -3.79171729e-01 -1.48491353e-01 8.63184929e-01
6.41386211e-02 4.40203130e-01 -1.39375794e+00 9.84901786e-01
3.24954003e-01 -1.17066514e+00 2.94911712e-01 5.69017008e-02
4.64982241e-01 1.54347658e-01 7.91076943e-02 7.27164149e-01
2.99608558e-01 -1.27867496e+00 2.77434260e-01 3.57661486e-01
8.19856584e-01 -7.61881709e-01 1.08910370e+00 1.74914986e-01
-7.74591088e-01 -5.75712360e-02 -7.07014620e-01 -1.72423735e-01
-1.31869987e-01 7.43838251e-01 -6.74482286e-01 3.03053409e-01
4.64371204e-01 9.96316314e-01 -6.11018836e-01 2.54774779e-01
-1.63909510e-01 3.25075924e-01 -3.26661244e-02 -2.79400915e-01
5.26194334e-01 1.37780765e-02 3.14312607e-01 1.63233817e+00
1.52866304e-01 -2.11255088e-01 -1.58067673e-01 9.18369532e-01
-3.37859601e-01 6.15178704e-01 -1.18151617e+00 -4.07907218e-01
7.18112513e-02 1.30968332e+00 -2.15360135e-01 -4.74996209e-01
-6.60578847e-01 1.07885993e+00 5.68182528e-01 1.27254963e-01
-6.91622376e-01 -6.50186777e-01 1.35793281e+00 1.53370067e-01
9.57883373e-02 5.97953089e-02 7.45749846e-02 -1.26390350e+00
-2.81808585e-01 -6.13172114e-01 4.02805060e-01 -6.27569795e-01
-1.78131223e+00 7.64309704e-01 -1.06812268e-01 -9.99390900e-01
-4.16635036e-01 -8.88372123e-01 -6.76113248e-01 1.05256116e+00
-1.64459205e+00 -9.84508097e-01 6.17946163e-02 6.06538236e-01
4.58957970e-01 -5.56094587e-01 1.45752859e+00 3.79678071e-01
-3.49019617e-01 8.12899768e-01 1.32354677e-01 3.14802974e-01
8.64060462e-01 -1.27475619e+00 3.50372881e-01 2.47276574e-01
3.09445560e-01 7.23340988e-01 4.99084115e-01 -3.17864746e-01
-8.42893064e-01 -1.14974868e+00 1.28964126e+00 -6.85490251e-01
9.57024336e-01 -8.25367510e-01 -1.06476414e+00 6.98130667e-01
2.21400201e-01 2.22964045e-02 1.22172523e+00 5.10556579e-01
-3.77381951e-01 5.05800406e-03 -1.11512172e+00 4.56038743e-01
8.28707516e-01 -8.02019656e-01 -1.02918756e+00 7.42628396e-01
1.11964285e+00 2.96642929e-01 -6.93270802e-01 3.28150280e-02
3.63154024e-01 -5.37123382e-01 1.20805120e+00 -1.34802067e+00
6.91237986e-01 4.20295656e-01 -3.70757580e-01 -1.62724459e+00
-5.51651895e-01 -9.47180688e-02 1.65365025e-01 1.27561510e+00
4.60399479e-01 -7.94589639e-01 4.98815209e-01 1.54351264e-01
1.66170850e-01 -8.09901357e-01 -7.03757167e-01 -6.94328308e-01
5.09167790e-01 -4.25177097e-01 4.50834662e-01 1.28658843e+00
1.98164448e-01 7.21848488e-01 -2.88787663e-01 -1.78415015e-01
1.54221088e-01 -2.88305372e-01 4.23349708e-01 -1.32307804e+00
3.51383798e-02 -3.22061419e-01 -9.41085517e-01 -4.25434202e-01
7.17806280e-01 -1.53279305e+00 -3.69883925e-01 -1.47925138e+00
3.12787145e-01 -2.00670645e-01 -6.03810966e-01 5.99905968e-01
-1.71417847e-01 4.85028267e-01 -1.91721812e-01 -2.83623457e-01
-1.41274020e-01 9.44406152e-01 7.97026455e-01 -3.94665688e-01
-1.21432863e-01 -4.68450934e-01 -1.23748827e+00 7.52622724e-01
7.23798513e-01 -9.40197289e-01 -4.94785905e-01 -3.52250278e-01
5.49353182e-01 -6.55412912e-01 1.44487798e-01 -5.98234475e-01
-3.79185617e-01 2.82471329e-01 2.66246378e-01 3.11133987e-03
4.01245415e-01 -6.99967384e-01 -6.70159161e-01 6.31138682e-01
-5.33495784e-01 1.42437980e-01 3.74605298e-01 5.79191089e-01
-2.49448687e-01 -4.81015623e-01 7.21910000e-01 -5.08019999e-02
-7.78730035e-01 1.85152084e-01 -4.90768135e-01 3.92072439e-01
8.51585031e-01 -9.38285738e-02 8.85994062e-02 -4.24058676e-01
-1.02079976e+00 1.01386653e-02 1.71047121e-01 8.01894724e-01
6.53041005e-01 -1.61605072e+00 -9.49550033e-01 4.68480676e-01
7.36556709e-01 -6.32151067e-01 -1.92202672e-01 5.58459044e-01
-3.49766642e-01 3.11150312e-01 -2.07681239e-01 -4.48089600e-01
-9.25089419e-01 7.30154335e-01 1.84917614e-01 -4.00264382e-01
-7.01813102e-01 7.69578934e-01 5.37994564e-01 -9.70637739e-01
-4.52980446e-03 -4.77943987e-01 -5.31910837e-01 4.82241303e-01
6.72966003e-01 -7.85028115e-02 1.51315868e-01 -6.52140737e-01
-6.43480718e-01 5.46424866e-01 -2.54220199e-02 2.75517013e-02
1.56290078e+00 1.95618588e-02 -1.14871599e-01 5.63750923e-01
1.73904145e+00 -4.20109890e-02 -1.94409773e-01 -4.84814495e-01
6.21783473e-02 -1.41380757e-01 -7.14137731e-03 -2.96039373e-01
-8.83383632e-01 9.72668350e-01 7.40877092e-01 3.18744987e-01
7.60982633e-01 2.50602871e-01 7.01857448e-01 3.64054620e-01
-7.40371123e-02 -9.89489794e-01 4.34633702e-01 6.11567378e-01
8.97936821e-01 -1.41106153e+00 -2.17707515e-01 1.26276016e-01
-7.15857804e-01 1.08937824e+00 2.91422844e-01 -4.10681129e-01
9.86471534e-01 1.78236812e-01 4.87705134e-03 -4.00275975e-01
-7.18363702e-01 -6.01827383e-01 2.82232493e-01 8.19639862e-01
5.92938185e-01 -1.28900081e-01 -5.57897627e-01 1.05998731e+00
-2.95081198e-01 -4.10342723e-01 4.40988749e-01 4.77127910e-01
-3.43522012e-01 -1.31249154e+00 8.28670934e-02 5.74321747e-01
-4.18038428e-01 -4.28534210e-01 -5.20767391e-01 1.00398052e+00
2.44952992e-01 6.04304433e-01 4.26833868e-01 -4.79344338e-01
3.55254829e-01 4.95920807e-01 7.80048314e-03 -1.29032910e+00
-6.59757674e-01 -4.66206700e-01 6.37391061e-02 -4.93970633e-01
-2.62057841e-01 -3.21141362e-01 -1.22475767e+00 -1.62160277e-01
-1.31147683e-01 2.76815239e-02 5.75434029e-01 8.95922899e-01
2.52228439e-01 6.00077331e-01 5.06161749e-01 -6.34473562e-01
-4.24247622e-01 -1.10770845e+00 -9.08720672e-01 9.46767390e-01
3.93675953e-01 -5.64822257e-01 -3.68346274e-01 -3.11521024e-01] | [10.458312034606934, 8.711012840270996] |
cfa7723f-59bc-42f4-8f68-670adc72868d | scops-self-supervised-co-part-segmentation | 1905.01298 | null | https://arxiv.org/abs/1905.01298v1 | https://arxiv.org/pdf/1905.01298v1.pdf | SCOPS: Self-Supervised Co-Part Segmentation | Parts provide a good intermediate representation of objects that is robust with respect to the camera, pose and appearance variations. Existing works on part segmentation is dominated by supervised approaches that rely on large amounts of manual annotations and can not generalize to unseen object categories. We propose a self-supervised deep learning approach for part segmentation, where we devise several loss functions that aids in predicting part segments that are geometrically concentrated, robust to object variations and are also semantically consistent across different object instances. Extensive experiments on different types of image collections demonstrate that our approach can produce part segments that adhere to object boundaries and also more semantically consistent across object instances compared to existing self-supervised techniques. | ['Ming-Hsuan Yang', 'Varun Jampani', 'Wei-Chih Hung', 'Sifei Liu', 'Jan Kautz', 'Pavlo Molchanov'] | 2019-05-03 | scops-self-supervised-co-part-segmentation-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Hung_SCOPS_Self-Supervised_Co-Part_Segmentation_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Hung_SCOPS_Self-Supervised_Co-Part_Segmentation_CVPR_2019_paper.pdf | cvpr-2019-6 | ['unsupervised-facial-landmark-detection'] | ['computer-vision'] | [-7.27536099e-04 2.83239502e-02 -3.22389275e-01 -6.76652372e-01
-5.70609689e-01 -8.08356941e-01 2.76466578e-01 6.26502186e-02
1.08934671e-01 3.79810363e-01 -1.03916824e-01 4.66592640e-01
5.51495627e-02 -5.54337621e-01 -1.07455635e+00 -1.86135098e-01
3.20484750e-02 8.48036647e-01 9.49805439e-01 3.04459292e-03
-9.78996530e-02 8.76994908e-01 -1.54414499e+00 -6.74865916e-02
6.10714257e-01 1.09075582e+00 -3.31706703e-02 3.94823134e-01
-4.70342785e-01 4.99711901e-01 -7.25403726e-01 -3.73437345e-01
7.38547504e-01 -2.52277195e-01 -1.18161809e+00 9.68216419e-01
8.29063177e-01 7.22085731e-03 -3.11845373e-02 1.13103926e+00
3.03278901e-02 3.76279764e-02 8.63326609e-01 -1.37348223e+00
-5.88711977e-01 7.18734980e-01 -6.42547250e-01 -1.76219478e-01
2.17374027e-01 -2.92737521e-02 6.85118258e-01 -4.40567225e-01
8.67739558e-01 1.32491410e+00 1.07703817e+00 5.70258379e-01
-1.36819839e+00 -3.45073819e-01 5.74007750e-01 -2.62312084e-01
-1.25849617e+00 -2.19791010e-01 1.03136802e+00 -4.21090662e-01
4.50704992e-01 7.94705078e-02 7.47135103e-01 7.72953033e-01
4.46540266e-02 1.31064546e+00 7.85988033e-01 -1.96114853e-01
3.92577261e-01 -2.08462263e-03 4.35002953e-01 6.00328803e-01
5.02773643e-01 -4.74749327e-01 3.12989168e-02 4.69773747e-02
8.56110096e-01 2.84240752e-01 -3.25536460e-01 -1.09652913e+00
-1.14006817e+00 4.62976784e-01 7.14848816e-01 3.13741237e-01
-3.22160453e-01 3.44516069e-01 3.58751625e-01 -8.07599574e-02
3.80844921e-01 2.86063582e-01 -9.69888926e-01 3.83269489e-01
-1.00001192e+00 2.34501302e-01 9.67324197e-01 1.69169641e+00
9.88484144e-01 -1.26214311e-01 -1.39552385e-01 1.03569090e+00
5.06096303e-01 4.21393037e-01 5.27053356e-01 -9.97029305e-01
3.12627107e-02 1.17725158e+00 -7.00219199e-02 -3.82445276e-01
-3.82648975e-01 -1.72031313e-01 -5.57025731e-01 2.08625451e-01
1.32741392e-01 -1.39542657e-03 -1.61393857e+00 1.48183775e+00
5.95290005e-01 -2.34622240e-01 -2.43480802e-01 5.94059408e-01
1.03926015e+00 1.13938637e-01 1.98467627e-01 1.69207484e-01
1.23846364e+00 -1.24090409e+00 -6.11209691e-01 -4.29426163e-01
8.09677467e-02 -9.30000961e-01 6.80997729e-01 1.47526771e-01
-1.10812628e+00 -8.42425704e-01 -1.01948452e+00 -5.91650419e-02
-5.24984837e-01 5.99328801e-02 7.51680732e-01 6.21486008e-01
-8.43430400e-01 7.50013769e-01 -9.01912153e-01 -4.57049578e-01
1.05791354e+00 5.89140832e-01 -6.11360788e-01 4.07595858e-02
-1.95053875e-01 5.74498117e-01 8.88259530e-01 -4.95681092e-02
-8.61142159e-01 -4.46186453e-01 -1.04121840e+00 -2.98028708e-01
3.13784212e-01 -8.48314285e-01 1.46060050e+00 -1.65273678e+00
-1.24572599e+00 9.64557230e-01 5.93928844e-02 -4.10887361e-01
6.89415157e-01 -3.45838964e-01 -8.05573687e-02 2.58232057e-01
2.40681514e-01 1.26567829e+00 1.06927693e+00 -1.77090561e+00
-5.24317682e-01 -3.98337245e-01 -6.28327951e-02 1.02681004e-01
1.89339563e-01 -2.33590394e-01 -7.63844728e-01 -6.35857463e-01
6.24666154e-01 -9.75477219e-01 -4.46278661e-01 3.46983999e-01
-7.98393488e-01 -2.71813750e-01 1.33487606e+00 -1.13068342e-01
4.31017756e-01 -1.71359193e+00 -5.61663695e-02 -8.22878536e-03
-4.97357957e-02 3.56675625e-01 -3.77068669e-02 1.23638265e-01
-8.96437000e-03 1.83820248e-01 -4.77310419e-01 -4.35530543e-01
-6.16474152e-02 3.27788621e-01 4.80756834e-02 5.88269651e-01
2.61034817e-01 1.12412429e+00 -5.54818690e-01 -8.62335980e-01
3.24077785e-01 1.74293384e-01 -2.47236565e-01 4.21046853e-01
-6.64324284e-01 2.13038623e-01 -4.52150673e-01 1.13700461e+00
8.82885814e-01 -3.26431453e-01 -3.38219196e-01 -2.38361597e-01
2.50140339e-01 -1.56931862e-01 -1.22161329e+00 1.81054688e+00
8.03471431e-02 3.14848155e-01 1.15072541e-01 -8.57179880e-01
8.92921507e-01 1.21698335e-01 6.18745208e-01 4.37571332e-02
4.40066844e-01 -4.79575852e-03 -1.84152484e-01 -3.14231277e-01
2.62783855e-01 2.42882725e-02 6.85035288e-02 3.58330756e-01
3.73927206e-01 -6.06401443e-01 2.48408452e-01 1.65677734e-03
7.21572995e-01 4.67053056e-01 2.58598894e-01 -3.55341315e-01
3.21046472e-01 4.18021023e-01 6.81088328e-01 6.11572742e-01
-4.15290326e-01 1.11594653e+00 3.81930657e-02 -4.96551812e-01
-1.01941848e+00 -1.05626130e+00 -3.59785140e-01 7.05276489e-01
7.90768206e-01 6.74959421e-02 -1.13724387e+00 -1.10292900e+00
1.69114694e-01 1.92734852e-01 -5.77009618e-01 4.24859673e-02
-4.50349331e-01 -4.20408785e-01 2.25903556e-01 1.09779692e+00
8.38554382e-01 -1.20580125e+00 -4.32658285e-01 1.18691131e-01
9.36445072e-02 -1.35687292e+00 -5.80434322e-01 2.55400926e-01
-1.24140000e+00 -1.46834671e+00 -8.09217453e-01 -1.20608974e+00
1.11512446e+00 4.43263680e-01 1.39230347e+00 1.02029033e-01
-5.15789688e-01 7.04707205e-01 -4.74285185e-01 -7.21556425e-01
-4.10206735e-01 -1.13261631e-03 -2.58142442e-01 -4.71255369e-02
1.95200369e-01 -3.26803237e-01 -6.26244187e-01 5.26143134e-01
-1.02732706e+00 -2.03786150e-01 7.83157706e-01 4.49655026e-01
1.10776711e+00 5.54513335e-02 1.49272323e-01 -1.09472275e+00
-5.01669422e-02 -3.00660998e-01 -5.75822115e-01 4.14092332e-01
-2.54062444e-01 7.30693340e-03 3.62508774e-01 -6.36176825e-01
-1.01018536e+00 7.58342445e-01 7.83099532e-02 -6.94200516e-01
-7.80529201e-01 -4.35437381e-01 -5.07022917e-01 -2.24663973e-01
4.45460945e-01 1.35618642e-01 -9.17365775e-02 -6.28981948e-01
5.31087041e-01 3.50376636e-01 4.83341664e-01 -5.17463326e-01
1.08623350e+00 8.88629675e-01 -1.04812704e-01 -7.27379143e-01
-9.00266111e-01 -8.56903791e-01 -1.32637846e+00 -7.20538199e-02
1.02671432e+00 -9.05776978e-01 -2.18404844e-01 7.47022808e-01
-1.02993906e+00 -3.01684737e-01 -5.82176089e-01 2.26572037e-01
-7.04515696e-01 5.61436415e-01 -5.21036565e-01 -4.92925376e-01
-5.06853580e-01 -9.51937973e-01 1.46306217e+00 5.09118855e-01
-1.78636357e-01 -9.33123589e-01 -8.29964206e-02 3.80699128e-01
-2.96963789e-02 4.68751580e-01 5.50099790e-01 -8.64087641e-01
-9.10399377e-01 -4.40190732e-01 -2.70205624e-02 5.06448567e-01
2.98226714e-01 3.30773145e-01 -8.58198166e-01 -1.36776507e-01
-1.53815761e-01 -4.33363438e-01 7.09448874e-01 5.82021713e-01
1.38947141e+00 -3.67503911e-01 -6.89535141e-01 6.28505170e-01
1.43622136e+00 2.40262840e-02 5.57681441e-01 2.37702727e-01
9.62007880e-01 5.28596222e-01 5.96259058e-01 -3.28048803e-02
1.83450148e-01 4.38036710e-01 5.23801088e-01 -4.34013247e-01
-2.67407864e-01 -3.87424290e-01 -7.93272555e-02 2.81638235e-01
3.01692963e-01 -2.13637963e-01 -5.98487973e-01 8.74992311e-01
-1.76558828e+00 -6.99913561e-01 -4.73011494e-01 1.82584739e+00
6.27849162e-01 2.88684726e-01 4.18579549e-01 4.09333371e-02
9.41288531e-01 -1.30897567e-01 -7.71093428e-01 -1.60083845e-01
-1.30616903e-01 2.00560123e-01 7.58718729e-01 -1.18488014e-01
-1.57245779e+00 1.22096515e+00 7.68625927e+00 5.41791320e-01
-6.71027124e-01 -1.09174728e-01 5.81080616e-01 3.93798441e-01
-8.25602263e-02 2.49806643e-02 -7.59509385e-01 1.98277771e-01
6.65547177e-02 2.89846569e-01 -1.91000640e-01 1.46231735e+00
-3.27288777e-01 -1.56351298e-01 -1.30609512e+00 7.71610618e-01
1.54492512e-01 -1.12409675e+00 5.48025519e-02 -2.78086692e-01
1.34190071e+00 -1.65909901e-02 -3.91798586e-01 7.72806108e-02
5.33747196e-01 -8.39529157e-01 1.04655480e+00 3.01026314e-01
4.01498377e-02 -4.82383430e-01 7.55137682e-01 2.91655362e-01
-1.34478986e+00 1.89063117e-01 -5.22170246e-01 3.75050813e-01
6.70401379e-02 2.10455284e-01 -8.46866131e-01 3.44809115e-01
9.37061131e-01 6.30171180e-01 -9.08009529e-01 1.50468540e+00
-2.07728967e-01 3.54691744e-01 -3.22217524e-01 7.00036809e-02
1.26740500e-01 -2.84790285e-02 3.33027184e-01 1.06657529e+00
-2.87654430e-01 -5.82883060e-02 6.27453566e-01 1.04674983e+00
-2.69215196e-01 1.71048284e-01 -5.35345376e-01 6.28886670e-02
3.46753865e-01 1.26919127e+00 -1.43462408e+00 -4.62589920e-01
-2.43510634e-01 1.20166171e+00 2.06238449e-01 2.60475278e-01
-6.76126301e-01 -2.04565510e-01 5.65714598e-01 1.64802760e-01
8.77949119e-01 -1.99993588e-02 -4.59409922e-01 -9.26958084e-01
1.89809307e-01 -4.80097294e-01 3.23028773e-01 -7.13730097e-01
-1.59744239e+00 6.19150877e-01 2.62146860e-01 -1.47306728e+00
8.27855244e-02 -6.23315692e-01 -7.71944404e-01 1.13640770e-01
-1.27831221e+00 -1.62840700e+00 -3.12405497e-01 5.91327488e-01
8.68939996e-01 3.51838656e-02 5.70059121e-01 -1.68973297e-01
-5.24412811e-01 3.73856962e-01 -1.65898278e-01 5.09540617e-01
2.90562600e-01 -1.51127815e+00 4.05493438e-01 9.00206029e-01
4.27644074e-01 6.18593156e-01 5.01724899e-01 -6.70875788e-01
-1.23945081e+00 -1.44216025e+00 1.87635198e-01 -5.44550955e-01
1.67969242e-01 -2.93206096e-01 -8.80962014e-01 9.58210647e-01
9.60041061e-02 4.73678708e-01 5.18638968e-01 -1.05800547e-01
-3.94529194e-01 -6.16841950e-02 -1.49657834e+00 3.37005138e-01
1.11551595e+00 -8.87573734e-02 -8.76519024e-01 5.58211029e-01
5.95637977e-01 -5.74461281e-01 -8.34914923e-01 7.18653083e-01
3.03518653e-01 -9.96851444e-01 9.69245553e-01 -6.67693734e-01
1.53801348e-02 -5.74994683e-01 1.00042261e-01 -7.85760641e-01
-3.48767132e-01 -4.57158267e-01 7.08623305e-02 1.49088299e+00
1.90454036e-01 -1.72731757e-01 9.68675554e-01 8.11038375e-01
-2.86704957e-01 -5.85869968e-01 -4.80319798e-01 -1.03583860e+00
-2.11349074e-02 -2.11320713e-01 6.45235062e-01 5.91823339e-01
-5.81170857e-01 9.67836529e-02 2.16362894e-01 2.24538818e-01
6.87776446e-01 6.33976042e-01 1.07593429e+00 -1.46611154e+00
-7.86710123e-04 -6.13650382e-01 -9.35684443e-01 -9.87772942e-01
3.36641401e-01 -6.09812558e-01 6.10976577e-01 -1.92402720e+00
3.80009383e-01 -4.83949572e-01 1.25404224e-01 7.09455788e-01
-1.52080223e-01 5.97823560e-01 -7.58658275e-02 3.60139221e-01
-9.83567178e-01 4.14801896e-01 1.26624489e+00 -4.41514701e-01
-1.81478277e-01 3.41809750e-01 -5.70421934e-01 1.08442736e+00
6.23105049e-01 -4.96838987e-01 -2.92272091e-01 -3.46640766e-01
-8.13380241e-01 -6.08076692e-01 1.97114080e-01 -1.17272103e+00
-1.01132490e-01 -8.07796493e-02 8.85340333e-01 -9.83486354e-01
2.54838586e-01 -1.16333103e+00 3.04491699e-01 3.90667915e-01
-2.36622617e-02 -2.38873065e-01 3.44340652e-01 7.05968499e-01
-1.69375569e-01 -4.61412102e-01 1.14853108e+00 -5.70147514e-01
-8.55667055e-01 6.51576579e-01 -8.73053074e-03 2.41798088e-01
1.40681660e+00 -7.13570297e-01 1.80268958e-01 4.27266676e-03
-6.03799760e-01 2.26918012e-01 9.45743859e-01 5.31085014e-01
3.52877468e-01 -1.27782083e+00 -3.49295855e-01 2.94447131e-02
2.77589172e-01 5.56584120e-01 5.60045242e-03 2.88193673e-01
-7.89346755e-01 2.38183975e-01 -2.42420703e-01 -1.01057589e+00
-1.32550931e+00 7.26306260e-01 4.86994058e-01 3.11685175e-01
-8.48326206e-01 1.14872849e+00 1.60730064e-01 -5.39162636e-01
4.31849509e-01 -6.05381966e-01 8.57252255e-02 -1.52568251e-01
1.11588025e-02 2.81705707e-01 -3.44188660e-02 -1.03952444e+00
-4.54152524e-01 1.06483006e+00 -1.84211776e-01 3.52956355e-01
1.11267269e+00 4.82761450e-02 8.18132311e-02 3.26179266e-01
1.24535525e+00 -2.01261401e-01 -1.64850986e+00 -2.04865277e-01
1.73044488e-01 -4.51955199e-01 -3.17587942e-01 -5.86468399e-01
-1.36114252e+00 4.96365666e-01 6.60656154e-01 1.95437491e-01
9.85485494e-01 3.98741543e-01 8.51931870e-01 2.83178896e-01
4.39712584e-01 -1.20608842e+00 2.14710608e-01 2.26143271e-01
7.50885785e-01 -1.47869170e+00 2.60113895e-01 -7.82125354e-01
-7.53068030e-01 1.10162950e+00 8.65741313e-01 -4.44479764e-01
7.35745192e-01 1.89773127e-01 3.95198524e-01 -3.15989107e-01
-8.25646967e-02 -6.15409076e-01 6.21492743e-01 9.21653748e-01
2.38156319e-01 3.75348702e-02 -4.38342243e-02 2.38972709e-01
-5.67780547e-02 -3.89171958e-01 1.58772141e-01 1.17928255e+00
-4.41329569e-01 -1.21400464e+00 -4.48175222e-01 2.62590528e-01
-4.22333002e-01 6.52375042e-01 -9.68578041e-01 1.08025968e+00
3.90365154e-01 6.69880152e-01 8.22554380e-02 1.03022359e-01
4.42585170e-01 1.51829928e-01 6.23473644e-01 -9.31235969e-01
-5.93560576e-01 3.17239493e-01 -3.49546015e-01 -5.14724493e-01
-7.95216024e-01 -7.03392506e-01 -1.42465115e+00 3.89968812e-01
-6.66825116e-01 -1.89029351e-01 5.57948291e-01 9.52609360e-01
1.21757381e-01 3.49931121e-01 5.82978308e-01 -1.13183498e+00
-2.17294723e-01 -9.20502007e-01 -6.05070949e-01 8.90310824e-01
9.75618809e-02 -5.59355676e-01 -9.44118798e-02 3.91932189e-01] | [9.304349899291992, 0.5434898138046265] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.