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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
d5dd9ab6-aa1a-4dd4-8ae0-7633bea6de41 | adaptive-meta-learner-via-gradient-similarity | 2209.04702 | null | https://arxiv.org/abs/2209.04702v1 | https://arxiv.org/pdf/2209.04702v1.pdf | Adaptive Meta-learner via Gradient Similarity for Few-shot Text Classification | Few-shot text classification aims to classify the text under the few-shot scenario. Most of the previous methods adopt optimization-based meta learning to obtain task distribution. However, due to the neglect of matching between the few amount of samples and complicated models, as well as the distinction between useful and useless task features, these methods suffer from the overfitting issue. To address this issue, we propose a novel Adaptive Meta-learner via Gradient Similarity (AMGS) method to improve the model generalization ability to a new task. Specifically, the proposed AMGS alleviates the overfitting based on two aspects: (i) acquiring the potential semantic representation of samples and improving model generalization through the self-supervised auxiliary task in the inner loop, (ii) leveraging the adaptive meta-learner via gradient similarity to add constraints on the gradient obtained by base-learner in the outer loop. Moreover, we make a systematic analysis of the influence of regularization on the entire framework. Experimental results on several benchmarks demonstrate that the proposed AMGS consistently improves few-shot text classification performance compared with the state-of-the-art optimization-based meta-learning approaches. | ['Xu Wang', 'Dezhong Peng', 'Qiaoyang Luo', 'Honghui Hu', 'Tianyi Lei'] | 2022-09-10 | null | https://aclanthology.org/2022.coling-1.431 | https://aclanthology.org/2022.coling-1.431.pdf | coling-2022-10 | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 1.77641332e-01 -2.93150067e-01 -1.93714887e-01 -3.84537101e-01
-6.69678271e-01 2.74860084e-01 5.83566010e-01 3.06764156e-01
-4.74904448e-01 5.79017162e-01 1.42586350e-01 2.47488752e-01
-3.22021693e-01 -7.77033448e-01 -3.29485387e-01 -8.32617939e-01
6.56940222e-01 2.75240451e-01 2.82349169e-01 -3.34014952e-01
4.31454331e-01 -3.49754423e-01 -1.78751063e+00 1.63886264e-01
1.40556002e+00 1.21257484e+00 4.17096466e-01 1.37981787e-01
-6.43354297e-01 7.53573835e-01 -4.75061536e-01 -4.03763473e-01
4.24751459e-04 -6.66691184e-01 -4.68910903e-01 2.19943792e-01
6.20983653e-02 -1.50760636e-01 -2.13129744e-01 1.25679827e+00
5.97513974e-01 7.50060499e-01 6.89047635e-01 -1.22180963e+00
-6.55662894e-01 5.81014931e-01 -5.73740125e-01 3.07281733e-01
-2.07975395e-02 -1.77127460e-03 9.81170297e-01 -1.11615360e+00
4.23375309e-01 1.10288358e+00 5.86984992e-01 6.48005188e-01
-8.28366518e-01 -6.00764215e-01 3.85674089e-01 5.65353215e-01
-1.24396932e+00 -4.36217546e-01 1.02635360e+00 -3.09312046e-01
6.15296364e-01 5.99848628e-02 4.47899967e-01 9.70214069e-01
-6.07367940e-02 1.01923144e+00 8.59026015e-01 -6.04673207e-01
6.99885130e-01 3.86291146e-01 7.87471652e-01 6.56602085e-01
4.01141420e-02 -2.89908201e-01 -4.46571171e-01 -2.61132926e-01
1.03043228e-01 3.78751308e-01 -2.26933479e-01 -5.19685090e-01
-8.34055364e-01 9.71735954e-01 3.22445571e-01 4.00764138e-01
-1.87523544e-01 -2.57518202e-01 8.53050888e-01 1.53770924e-01
9.21499372e-01 3.38000178e-01 -3.51416647e-01 -1.89804047e-01
-8.88332009e-01 -1.84546243e-02 5.95650673e-01 9.45007801e-01
8.50302279e-01 9.72003341e-02 -5.45389175e-01 1.28362870e+00
1.18928060e-01 -7.16218799e-02 1.20063853e+00 -2.73728848e-01
8.32258761e-01 7.38826275e-01 -8.30251873e-02 -8.55194628e-01
-3.11389238e-01 -7.30307698e-01 -7.81321824e-01 -2.97251701e-01
1.53626308e-01 -2.64185995e-01 -8.38014305e-01 1.52264249e+00
4.07331586e-01 2.71720618e-01 3.82892713e-02 8.10233355e-01
8.77006054e-01 6.92960739e-01 2.44689837e-01 -5.39019644e-01
1.18081927e+00 -1.32515860e+00 -8.83533359e-01 -3.30386937e-01
9.66766477e-01 -3.73285860e-01 1.34072363e+00 -2.50767861e-02
-7.42502868e-01 -6.18374109e-01 -1.19348884e+00 8.59630853e-02
-5.03031313e-01 4.03809324e-02 4.82132554e-01 6.32287502e-01
-2.52482831e-01 7.46502757e-01 -5.28831840e-01 -3.88454169e-01
5.97609699e-01 -6.01268448e-02 6.60972148e-02 -1.58021614e-01
-1.39518833e+00 7.43442774e-01 7.93709517e-01 -1.20655298e-01
-4.89581347e-01 -9.07754779e-01 -8.95191789e-01 4.29553092e-01
5.25854707e-01 -6.90192699e-01 1.03336585e+00 -1.07845628e+00
-1.49679470e+00 3.96376908e-01 -1.84998915e-01 -2.78448850e-01
7.12512016e-01 -1.60080165e-01 -2.25806162e-01 -7.12055042e-02
-2.62116734e-02 1.41919360e-01 9.86795008e-01 -1.01803303e+00
-6.70432985e-01 -6.09260082e-01 -2.43479297e-01 5.03405511e-01
-1.08686841e+00 -3.82296681e-01 -4.15221363e-01 -8.07526946e-01
5.89393452e-02 -5.17141759e-01 -1.45514697e-01 -1.69839188e-01
-1.23918131e-01 -4.14790422e-01 9.76901114e-01 -3.35359722e-01
1.50136209e+00 -2.17535853e+00 2.12479651e-01 -2.02538818e-01
1.69935208e-02 5.22819459e-01 -1.35192588e-01 3.10415298e-01
8.07609856e-02 -2.17859834e-01 -2.70765156e-01 -5.58439493e-01
-8.42696652e-02 -1.56324469e-02 -2.25416705e-01 2.96319842e-01
-5.54611608e-02 8.88868630e-01 -1.08265078e+00 -7.38592923e-01
4.07277882e-01 1.80600524e-01 -3.36979091e-01 2.40241408e-01
-1.77297682e-01 8.20809379e-02 -7.74881065e-01 5.13578534e-01
5.60933173e-01 -2.65077472e-01 -5.71436435e-02 -1.12166457e-01
6.99960217e-02 8.43038037e-03 -1.13334000e+00 1.88940561e+00
-5.47698855e-01 2.23278567e-01 -2.42128029e-01 -1.39356256e+00
9.94007945e-01 1.62922949e-01 4.34019685e-01 -6.43522680e-01
4.44534570e-01 1.03400141e-01 -8.77141729e-02 -6.50388658e-01
4.40792829e-01 -4.64693606e-01 1.75109103e-01 5.30376792e-01
3.70025337e-01 1.12160191e-01 2.94591606e-01 9.61871147e-02
6.92765534e-01 3.26276533e-02 5.21047175e-01 -2.87653327e-01
6.30993485e-01 -1.35987371e-01 6.78405166e-01 8.03056300e-01
-5.16692400e-01 3.88305187e-01 1.72138795e-01 -3.47863108e-01
-8.12552392e-01 -4.41153616e-01 -2.50609040e-01 1.53476334e+00
3.04453731e-01 -4.02811021e-01 -7.51554430e-01 -8.11784744e-01
-2.02407800e-02 1.08159208e+00 -9.35149968e-01 -6.53199196e-01
-1.54847682e-01 -1.23738742e+00 3.00137568e-02 6.08401418e-01
6.33576810e-01 -7.93545008e-01 -4.72195148e-01 3.21475089e-01
-1.83491886e-01 -8.29873085e-01 -5.26829302e-01 2.67164230e-01
-1.07227314e+00 -7.99212456e-01 -9.77166533e-01 -6.72736108e-01
5.84255338e-01 7.65682220e-01 5.15217841e-01 1.06363341e-01
-2.85048574e-01 1.50085613e-01 -7.26523101e-01 -4.53653753e-01
-2.34192625e-01 3.14602375e-01 -4.82147336e-02 3.24979544e-01
6.29508138e-01 -4.54787195e-01 -3.13898444e-01 2.10495278e-01
-8.48205507e-01 8.79638642e-02 3.50908607e-01 1.20982623e+00
2.63915658e-01 2.25971028e-01 9.16526914e-01 -1.04029155e+00
7.92824864e-01 -7.31571317e-01 -1.80261329e-01 4.44059163e-01
-8.94059896e-01 9.54096094e-02 9.26966131e-01 -5.90983212e-01
-1.32219934e+00 -2.58202076e-01 1.65585831e-01 -4.95418221e-01
2.08184570e-01 5.96041501e-01 -2.19882727e-01 1.17516465e-01
7.15809703e-01 5.90341270e-01 6.73804153e-03 -4.33750480e-01
2.34077275e-01 8.60762835e-01 -7.88943321e-02 -4.23920274e-01
5.15665293e-01 4.61878061e-01 -2.10510418e-01 -1.02026010e+00
-1.39294779e+00 -8.01021457e-01 -5.39307356e-01 -1.38326898e-01
6.36756539e-01 -6.93517506e-01 -2.38354385e-01 4.38687414e-01
-8.31648529e-01 -8.25839713e-02 -3.48233998e-01 4.77446258e-01
-5.22112072e-01 6.16357327e-01 -3.52742106e-01 -9.59162235e-01
-6.99639857e-01 -9.48536396e-01 8.28798771e-01 3.20675701e-01
1.12997033e-01 -1.02421272e+00 -1.72182394e-03 3.92831087e-01
3.78494382e-01 -1.70948178e-01 1.13545418e+00 -1.01212728e+00
8.23478699e-02 -3.87511283e-01 -2.89849937e-01 2.60878026e-01
2.21211180e-01 -3.64559054e-01 -1.12911725e+00 -3.04107845e-01
3.87008786e-01 -5.10731578e-01 1.16102099e+00 3.75578433e-01
1.18250966e+00 -8.28339085e-02 -2.31148243e-01 5.79519868e-01
1.35305548e+00 1.51933283e-01 4.69518751e-01 4.89101917e-01
6.91697001e-01 5.87920487e-01 9.49079633e-01 6.99328780e-01
1.77598923e-01 5.36912024e-01 2.27635995e-01 4.19665247e-01
1.36395320e-01 -2.59939492e-01 -1.83543656e-02 1.00775802e+00
3.75548601e-02 3.08125559e-02 -6.55079782e-01 2.57486314e-01
-2.29339886e+00 -1.01261806e+00 2.05465540e-01 2.25878477e+00
6.71118379e-01 2.61753827e-01 4.20190655e-02 2.36374840e-01
1.10361910e+00 4.15617347e-01 -7.75795817e-01 -3.89408544e-02
1.04372777e-01 -1.91861570e-01 7.29801133e-02 9.82055515e-02
-1.20960581e+00 8.73520434e-01 5.14387417e+00 1.37504375e+00
-9.09659326e-01 3.52657706e-01 4.85726207e-01 -1.04073033e-01
-1.66076854e-01 -6.60320446e-02 -9.76379514e-01 6.75414920e-01
4.86349761e-01 -5.68566263e-01 3.13469619e-01 1.06267917e+00
4.73176651e-02 8.83083493e-02 -8.09846818e-01 1.02952015e+00
4.30473894e-01 -1.09136188e+00 2.67956525e-01 -2.75770783e-01
8.00345838e-01 -3.09280097e-01 -1.31225571e-01 8.34647596e-01
-2.87618518e-01 -4.53950495e-01 6.48816407e-01 6.06290460e-01
2.54828095e-01 -8.43330145e-01 8.11193824e-01 8.16731751e-01
-1.05312884e+00 -5.16572833e-01 -8.87874544e-01 -1.65325217e-02
-8.36764425e-02 6.67029858e-01 -4.16052073e-01 6.67386532e-01
2.61968285e-01 8.29754591e-01 -6.43037498e-01 9.71684337e-01
1.51423529e-01 4.24843997e-01 1.57230318e-01 -5.27391374e-01
3.11972737e-01 -3.03406030e-01 6.28401875e-01 9.54726040e-01
1.23848952e-01 7.66641200e-02 2.61493653e-01 8.29350948e-01
-6.77614138e-02 6.24108493e-01 -2.97378808e-01 -4.97613512e-02
4.03436959e-01 1.19665015e+00 -5.81245363e-01 -5.63355803e-01
-5.02560318e-01 8.31406713e-01 6.47163808e-01 2.66508639e-01
-6.45380080e-01 -7.44223416e-01 1.23847902e-01 -1.67894602e-01
3.00366014e-01 1.59802109e-01 -4.86668587e-01 -1.43747795e+00
8.38070586e-02 -6.73354745e-01 6.10230744e-01 -4.68558371e-01
-1.44817495e+00 3.97277027e-01 -6.75824285e-02 -1.34975064e+00
-8.67096186e-02 -2.70071477e-01 -8.36139381e-01 6.59640133e-01
-1.51716018e+00 -1.00694823e+00 -5.01557708e-01 3.96711379e-01
1.14381695e+00 -3.07197511e-01 6.71160638e-01 2.17846781e-01
-9.00525749e-01 7.55375326e-01 4.73692656e-01 -1.69717804e-01
6.83731675e-01 -9.81669247e-01 2.20723040e-02 5.39665639e-01
-2.14918062e-01 5.40331244e-01 6.12472594e-01 -5.71476400e-01
-1.19365561e+00 -1.05331433e+00 5.00689745e-01 -4.37446609e-02
6.13994956e-01 -2.30262518e-01 -1.21630776e+00 3.42428058e-01
-2.14508802e-01 -1.54710514e-02 7.61567473e-01 2.95073032e-01
-3.28770638e-01 -2.18600184e-01 -9.96346533e-01 5.71978271e-01
8.30565870e-01 -3.67761791e-01 -8.80810201e-01 4.13117617e-01
6.81780815e-01 -2.95670647e-02 -6.14249289e-01 2.97121376e-01
5.45340717e-01 -7.65068412e-01 6.58732116e-01 -8.81238401e-01
5.97002447e-01 1.46264702e-01 -1.61822498e-01 -1.55552483e+00
-4.50573981e-01 -1.57090768e-01 -4.27359998e-01 1.35137773e+00
1.44927651e-01 -6.00159287e-01 7.19112098e-01 5.84278703e-01
-1.67859361e-01 -9.31913495e-01 -8.93730640e-01 -9.14906025e-01
1.92739382e-01 -2.40586355e-01 4.60542798e-01 1.19034863e+00
4.46091026e-01 5.97988248e-01 -5.27314246e-01 -4.17212307e-01
7.73942053e-01 2.18230546e-01 6.98159039e-01 -1.40493906e+00
-3.06884646e-01 -4.25977111e-01 -3.18027735e-01 -9.10896838e-01
3.41733634e-01 -9.47274983e-01 8.15020949e-02 -1.40252268e+00
6.00816071e-01 -2.14625686e-01 -4.95956630e-01 1.31647244e-01
-8.62699091e-01 -3.01808715e-01 2.65383303e-01 3.72215956e-01
-8.33854496e-01 1.16860366e+00 1.29598248e+00 -2.46388748e-01
-3.48364860e-01 1.68292880e-01 -7.44322419e-01 7.42464662e-01
7.04093575e-01 -5.57604253e-01 -5.95300555e-01 -1.99886903e-01
-1.23271570e-01 -1.43581301e-01 1.75070576e-02 -9.39215124e-01
3.95464122e-01 -2.16163605e-01 2.61698425e-01 -3.43512088e-01
3.75396937e-01 -6.17623329e-01 -5.64858615e-01 4.68138069e-01
-5.46315074e-01 -5.93313456e-01 -6.52273670e-02 9.26972747e-01
-7.19681457e-02 -8.60424638e-01 9.56519961e-01 -2.37679601e-01
-7.76346564e-01 3.07048440e-01 -1.26162451e-02 3.89758110e-01
1.00862932e+00 -2.52229124e-01 -3.54555964e-01 -2.43053019e-01
-5.02101779e-01 3.54999572e-01 3.45086843e-01 5.04946113e-01
4.55043435e-01 -1.29729033e+00 -5.47860563e-01 1.24014154e-01
5.33559442e-01 -3.22602659e-01 6.39501929e-01 9.09261823e-01
9.77653489e-02 2.45053962e-01 -4.62084599e-02 -4.33183461e-01
-9.24061298e-01 8.60077024e-01 2.51594037e-01 -3.47143918e-01
-7.29432821e-01 5.83837330e-01 2.11873874e-01 -2.89796382e-01
3.78066689e-01 8.38221237e-02 -3.34477216e-01 3.77079159e-01
7.33161092e-01 7.17787325e-01 9.29088295e-02 -3.71100962e-01
-9.00941193e-02 3.87221813e-01 -4.90055025e-01 2.07527056e-01
1.16656125e+00 -2.40518525e-01 2.69059926e-01 8.96084607e-01
1.33819294e+00 -5.72090328e-01 -1.10734057e+00 -5.35023928e-01
-5.35129830e-02 -6.16462231e-01 3.11483741e-01 -4.00689662e-01
-7.44780838e-01 1.07019484e+00 4.65845406e-01 2.22252086e-01
8.00334632e-01 -4.36747313e-01 8.78624082e-01 5.99796593e-01
2.99185216e-01 -1.65382707e+00 2.20424175e-01 5.78043461e-01
4.57381845e-01 -1.44102669e+00 1.50443465e-01 -2.82277793e-01
-8.12136710e-01 1.13737512e+00 6.88546300e-01 -4.59144264e-03
7.22211719e-01 -3.97778809e-01 -1.26849830e-01 3.73254456e-02
-7.75391042e-01 -1.58291355e-01 3.56154442e-01 3.31723899e-01
2.60558546e-01 -2.33079121e-01 -6.35292590e-01 9.17844415e-01
3.35122645e-01 1.09476916e-01 2.49564856e-01 1.08024430e+00
-8.26333702e-01 -6.53202355e-01 -1.57275354e-03 8.08797061e-01
-1.67399004e-01 -1.25995368e-01 -1.45621449e-01 5.52292347e-01
-1.54270127e-01 9.46271241e-01 -1.22525223e-01 -2.04479426e-01
3.80743474e-01 4.68180507e-01 2.87697673e-01 -8.39089394e-01
-4.37104136e-01 -9.39562246e-02 -2.26147711e-01 -1.02169484e-01
-3.09041798e-01 -4.23419893e-01 -9.74803925e-01 5.12995459e-02
-9.48867202e-01 3.79947275e-01 5.12193978e-01 1.37690079e+00
3.27544123e-01 6.29531503e-01 8.36818695e-01 -7.42909610e-01
-1.31251347e+00 -1.36880147e+00 -8.75311375e-01 6.38538599e-01
-6.22610897e-02 -9.10298586e-01 -6.35401070e-01 -3.01874459e-01] | [10.177956581115723, 3.5120670795440674] |
32297d13-2419-42a1-bc25-411d686cb924 | systematic-n-tuple-networks-for-position | 1406.1509 | null | http://arxiv.org/abs/1406.1509v3 | http://arxiv.org/pdf/1406.1509v3.pdf | Systematic N-tuple Networks for Position Evaluation: Exceeding 90% in the Othello League | N-tuple networks have been successfully used as position evaluation functions
for board games such as Othello or Connect Four. The effectiveness of such
networks depends on their architecture, which is determined by the placement of
constituent n-tuples, sequences of board locations, providing input to the
network. The most popular method of placing n-tuples consists in randomly
generating a small number of long, snake-shaped board location sequences. In
comparison, we show that learning n-tuple networks is significantly more
effective if they involve a large number of systematically placed, short,
straight n-tuples. Moreover, we demonstrate that in order to obtain the best
performance and the steepest learning curve for Othello it is enough to use
n-tuples of size just 2, yielding a network consisting of only 288 weights. The
best such network evolved in this study has been evaluated in the online
Othello League, obtaining the performance of nearly 96% --- more than any other
player to date. | ['Wojciech Jaśkowski'] | 2014-06-05 | null | null | null | null | ['board-games'] | ['playing-games'] | [-2.05344558e-01 2.30242804e-01 -3.17650437e-01 -6.27213940e-02
-5.55203140e-01 -8.79270911e-01 4.08967659e-02 5.59319735e-01
-8.02626431e-01 1.14721644e+00 -1.44649059e-01 -3.99059087e-01
-5.97701252e-01 -1.11452591e+00 -8.09053957e-01 -4.48247820e-01
-6.03329122e-01 1.00656378e+00 7.53877103e-01 -8.40883493e-01
3.21130902e-01 4.93414998e-01 -1.78956950e+00 3.28191757e-01
4.80065286e-01 1.19448161e+00 -4.30086285e-01 9.88036275e-01
1.26410052e-01 7.71670640e-01 -8.94888163e-01 -8.48356009e-01
6.13210559e-01 -3.75808567e-01 -6.19069040e-01 -5.11885285e-01
2.21596509e-01 -2.08505645e-01 -1.30355626e-01 7.25047529e-01
5.51486313e-01 4.09015685e-01 5.25124431e-01 -9.73512173e-01
2.01576501e-01 1.21239305e+00 -1.16024613e-01 1.74237281e-01
2.92126298e-01 -9.19960216e-02 1.54118645e+00 -4.07847583e-01
5.67049980e-01 7.93178260e-01 1.11044836e+00 2.52572894e-01
-1.07422149e+00 -3.72582555e-01 -2.73918331e-01 3.48079614e-02
-1.01698971e+00 -1.94768786e-01 4.70956504e-01 1.37319788e-01
7.67687798e-01 6.62138820e-01 1.30373836e+00 4.56696481e-01
2.22443223e-01 6.50083303e-01 6.46538496e-01 -6.10356867e-01
3.78769279e-01 -2.83480287e-01 -1.02954237e-02 5.78051269e-01
5.07663488e-01 1.93357617e-01 -7.11257935e-01 -2.09701046e-01
9.76576924e-01 -2.76637346e-01 3.10409516e-01 -6.84144318e-01
-9.26809967e-01 7.28644907e-01 4.93275791e-01 2.27279320e-01
-2.61721343e-01 4.48049128e-01 5.89179158e-01 4.40672994e-01
1.67224869e-01 1.08525813e+00 -4.49221045e-01 -7.58710086e-01
-7.27461100e-01 7.08816051e-01 1.12018359e+00 6.16767764e-01
6.53450668e-01 -3.88522260e-02 -6.70124814e-02 9.47468102e-01
-3.48473400e-01 -1.19604290e-01 4.07731146e-01 -1.16817951e+00
7.36521661e-01 8.25319827e-01 2.38924712e-01 -9.95635986e-01
-7.88092852e-01 -5.70178568e-01 -4.92065251e-01 4.19425428e-01
9.62403893e-01 -4.48593020e-01 -5.10049462e-01 1.59572065e+00
2.18031272e-01 -3.82425368e-01 -1.65176466e-01 6.37645364e-01
4.35525626e-01 3.24557841e-01 -5.97671866e-01 1.59013897e-01
1.06903601e+00 -7.40271270e-01 -1.34634390e-01 -3.65739316e-02
8.14496219e-01 -3.54080349e-01 7.59723783e-01 8.18949640e-01
-1.63193834e+00 -4.01923090e-01 -1.44950533e+00 3.14654350e-01
-3.50202858e-01 7.82813430e-02 9.38088119e-01 8.36085498e-01
-1.14894974e+00 9.97447848e-01 -4.71299827e-01 -2.20447108e-02
2.17569232e-01 9.46741760e-01 -4.11248147e-01 4.59506273e-01
-1.31026042e+00 5.99274039e-01 8.07355404e-01 2.54470915e-01
-6.26097381e-01 -4.75077480e-01 -6.04724884e-01 2.54744262e-01
6.72556043e-01 -4.39863652e-01 1.53264892e+00 -7.34102130e-01
-1.51610720e+00 5.35826981e-01 5.43157101e-01 -9.84036028e-01
8.65519822e-01 -1.71238020e-01 2.79043149e-02 8.19793046e-02
-2.24633992e-01 7.98012376e-01 3.57714623e-01 -1.03812039e+00
-7.44964123e-01 -2.76867062e-01 5.91452539e-01 5.13009667e-01
-4.62318867e-01 -3.01625609e-01 -3.89991254e-01 -4.95520115e-01
4.39393073e-01 -1.06566966e+00 -5.42562544e-01 -4.30195421e-01
-4.70652759e-01 -2.83542454e-01 -9.17398930e-02 -1.01118892e-01
1.46889734e+00 -1.93018532e+00 -8.53912383e-02 6.78823352e-01
2.74437070e-01 1.44408822e-01 5.88002615e-02 6.83888018e-01
-2.62632757e-01 1.23732813e-01 2.99806356e-01 -1.16486475e-01
7.42581114e-02 2.00769007e-01 2.45875791e-01 2.64472604e-01
-2.59695977e-01 9.70623374e-01 -8.32452834e-01 -3.39419812e-01
-2.31523588e-01 -2.41905600e-01 -8.77498746e-01 -1.29681781e-01
-4.10037428e-01 -3.71644914e-01 -2.59229332e-01 2.35313356e-01
6.27478883e-02 4.41546515e-02 4.47792947e-01 2.07500711e-01
-2.19409615e-02 2.93388814e-01 -1.19055688e+00 1.52090597e+00
5.74034788e-02 5.37318707e-01 -3.46431822e-01 -6.41270161e-01
1.16252601e+00 2.13968635e-01 5.93230367e-01 -7.53835559e-01
2.53226876e-01 4.57636386e-01 4.39843446e-01 -1.51327476e-01
1.14486814e+00 -5.06556816e-02 -5.05500317e-01 3.72176170e-01
-1.56868204e-01 -9.14961547e-02 8.23465347e-01 1.62557960e-01
1.38085413e+00 2.27888003e-02 1.39928639e-01 -5.22028515e-03
1.40445799e-01 1.41937047e-01 6.83989406e-01 9.41189170e-01
6.17372915e-02 7.19084144e-01 1.28642356e+00 -7.41705656e-01
-1.32602787e+00 -8.38242948e-01 2.55008250e-01 1.28437531e+00
-1.94269102e-02 -6.06681168e-01 -9.57908154e-01 -3.53246331e-01
9.37362090e-02 2.78936118e-01 -8.31891298e-01 -5.64874709e-02
-8.56400430e-01 -4.96502161e-01 8.60630989e-01 6.61701322e-01
2.55260140e-01 -1.16876030e+00 -8.15376341e-01 4.68780071e-01
5.17405830e-02 -5.35619617e-01 -1.74645588e-01 8.00786912e-01
-1.15301430e+00 -1.06556952e+00 -4.57169265e-01 -6.05959773e-01
4.73115712e-01 -2.56079316e-01 1.41089821e+00 7.77634308e-02
2.03772746e-02 -2.92395830e-01 -3.88996810e-01 -1.92318156e-01
-1.22988611e-01 5.56278765e-01 -1.97952650e-02 -4.00734454e-01
-1.17800564e-01 -6.00625753e-01 -5.79228401e-01 4.67461020e-01
-8.21757853e-01 -3.34603667e-01 6.04651928e-01 1.05735898e+00
3.07700962e-01 3.21152419e-01 4.18425173e-01 -1.38182807e+00
7.52764344e-01 -6.80394843e-02 -4.61860657e-01 1.23006925e-02
-2.52848566e-01 -4.27395999e-02 7.39751637e-01 -2.55393595e-01
-4.23248291e-01 -1.42177939e-01 -2.12921828e-01 3.52101238e-03
1.02232829e-01 4.86686707e-01 1.29246011e-01 -2.40369514e-01
1.14028037e+00 -3.44687462e-01 1.12300649e-01 -5.49103022e-02
4.77016360e-01 6.25936836e-02 6.68809593e-01 -9.05434608e-01
4.11857158e-01 -1.48784416e-02 1.47614986e-01 -3.60923499e-01
-5.55544734e-01 -2.85361171e-01 -6.01579666e-01 -4.00947005e-01
5.16395807e-01 -3.57940435e-01 -1.55982828e+00 5.18348634e-01
-9.02384222e-01 -3.95395160e-01 -7.18722522e-01 1.51525125e-01
-9.05651987e-01 -2.17074051e-01 -7.36324728e-01 -8.49070132e-01
9.45437476e-02 -9.41596687e-01 3.32017839e-01 1.28593653e-01
-5.08927524e-01 -8.65693390e-01 1.70160577e-01 2.76506633e-01
1.75583884e-01 2.66824454e-01 1.08994138e+00 -1.01535392e+00
-4.62582856e-01 -4.64047462e-01 2.20116273e-01 1.93499178e-01
-3.58188450e-01 -2.11360276e-01 -3.60178739e-01 -1.61742583e-01
-4.10911560e-01 -5.32230079e-01 5.82266629e-01 2.75615752e-01
1.07605338e+00 -2.22816810e-01 3.41701955e-02 4.80298668e-01
1.31037056e+00 4.98123437e-01 7.14288712e-01 7.76974499e-01
3.66171837e-01 7.38499105e-01 5.34823835e-01 4.99376148e-01
2.86112159e-01 6.24701798e-01 7.00687706e-01 1.13253236e-01
1.61575317e-01 -5.19105136e-01 -9.54159200e-02 6.59863889e-01
-3.56808811e-01 -4.19531345e-01 -9.50504601e-01 4.20346588e-01
-2.12749457e+00 -9.35440361e-01 1.01524413e-01 2.42513061e+00
8.36178422e-01 9.10931170e-01 7.13389158e-01 6.74680769e-01
3.74365777e-01 1.87285498e-01 -4.24654007e-01 -5.99034727e-01
-2.23752350e-01 4.02507752e-01 8.99216294e-01 2.61043519e-01
-8.93837571e-01 6.84195280e-01 7.52519655e+00 1.13788736e+00
-5.09206235e-01 -3.60969454e-01 8.65743637e-01 -6.13093555e-01
-4.09906983e-01 -2.76385266e-02 -7.42285490e-01 3.52378935e-01
9.04404402e-01 2.44735450e-01 2.61499196e-01 5.38181424e-01
-3.02442145e-02 -4.56493706e-01 -9.82001007e-01 8.43707442e-01
-1.67659372e-01 -1.69918621e+00 -9.73409936e-02 1.22056551e-01
8.27127874e-01 -4.55350250e-01 3.67182307e-02 3.01539302e-01
5.62397420e-01 -1.26434267e+00 1.00467610e+00 3.43736708e-01
5.32397807e-01 -1.31574333e+00 7.25990117e-01 2.91424185e-01
-1.01996541e+00 -4.28137213e-01 -3.64590049e-01 -3.65137219e-01
-1.30398020e-01 3.82776469e-01 -6.90408587e-01 5.00212967e-01
5.56602836e-01 2.37473831e-01 -5.03062427e-01 1.56804383e+00
-1.49611682e-01 6.44168258e-01 -4.85340357e-01 -5.96387804e-01
4.71992373e-01 -2.16110498e-01 3.31988454e-01 5.64409792e-01
9.80093107e-02 1.36397123e-01 -5.58018982e-02 1.55391112e-01
-1.94747165e-01 9.43554789e-02 -5.12598932e-01 2.78541744e-01
6.22566819e-01 1.05796850e+00 -8.81597400e-01 1.09324828e-01
1.84364513e-01 3.24332833e-01 7.22502589e-01 7.15830624e-02
-6.18325412e-01 -5.92065036e-01 4.79366362e-01 4.54685658e-01
5.29742122e-01 -1.60050333e-01 -6.10521317e-01 -3.06331456e-01
-4.49078716e-02 -1.15113580e+00 3.70504558e-01 -6.24253750e-01
-8.94040465e-01 6.96625471e-01 -1.03610128e-01 -1.27213013e+00
-5.94702005e-01 -7.90409267e-01 -5.01556873e-01 4.06647921e-01
-5.80584347e-01 -5.84759891e-01 2.38795340e-01 2.22669363e-01
5.56362346e-02 -3.99165362e-01 5.68599224e-01 1.71866715e-01
-4.15980816e-01 9.26364660e-01 7.97543377e-02 2.08676800e-01
3.29383433e-01 -1.46746862e+00 3.59353840e-01 3.36633414e-01
3.15550476e-01 4.79658157e-01 7.25065589e-01 -4.12355930e-01
-1.17172599e+00 -2.81415403e-01 6.06594920e-01 -2.77322739e-01
7.05649972e-01 -4.90356833e-01 -4.11732584e-01 4.90854383e-01
-6.32551163e-02 -2.58994371e-01 6.93169534e-01 5.83376527e-01
-1.52840137e-01 -4.54453826e-01 -8.71328831e-01 7.17923582e-01
1.17685938e+00 -1.30919322e-01 -3.04367781e-01 1.42393857e-02
6.47860408e-01 -8.97652328e-01 -8.52908313e-01 3.14191252e-01
7.66917884e-01 -1.63817966e+00 1.00830328e+00 -7.94427335e-01
5.84794939e-01 5.39594442e-02 4.18541580e-02 -1.37460315e+00
-2.82318652e-01 -9.35671151e-01 1.58429474e-01 7.95695066e-01
9.55284536e-01 -7.26625443e-01 1.85646629e+00 3.59284669e-01
-9.91825014e-02 -1.38670254e+00 -9.69207466e-01 -6.66761100e-01
8.57124329e-02 -4.93660629e-01 8.56665075e-01 2.49525756e-01
1.85889527e-01 1.86639771e-01 -3.70870709e-01 -6.30836964e-01
3.12899053e-01 -1.34083033e-01 8.61059427e-01 -1.32676661e+00
-5.60240030e-01 -6.85398281e-01 -7.02340424e-01 -1.21262205e+00
-1.79639861e-01 -6.01580918e-01 1.63499519e-01 -1.05572176e+00
-5.56839593e-02 -1.18146026e+00 -3.01922917e-01 3.36944044e-01
1.02212876e-01 4.33037162e-01 2.39934102e-01 -1.12339363e-01
-7.71613181e-01 1.53325230e-01 1.23020899e+00 3.36567938e-01
-2.76169449e-01 3.02709073e-01 -6.92703485e-01 7.14571238e-01
7.62677133e-01 -3.37615967e-01 -4.70281184e-01 -2.71628410e-01
1.16285920e+00 5.33985138e-01 -1.33513211e-04 -1.11466420e+00
5.80122888e-01 2.18807369e-01 5.31336546e-01 -8.64156425e-01
6.39292300e-01 -5.61906874e-01 2.07136035e-01 3.69803429e-01
-4.21810061e-01 7.30975449e-01 1.31522715e-01 3.56554419e-01
-4.10014331e-01 -7.32288837e-01 1.50393099e-01 -2.63898194e-01
-5.35931945e-01 1.74053445e-01 -2.26093143e-01 2.21614316e-01
1.04293895e+00 -5.91434538e-01 -3.15739602e-01 -5.40716290e-01
-5.74517787e-01 7.66375810e-02 3.25775713e-01 -6.04659021e-02
3.64430755e-01 -1.41184831e+00 -3.31383854e-01 1.92976832e-01
-1.75669670e-01 1.13215722e-01 9.42365080e-02 3.61694336e-01
-1.21066940e+00 3.39969724e-01 -4.89878207e-01 -2.69819438e-01
-1.11452174e+00 -3.41522992e-02 5.76301515e-01 -5.74754715e-01
-3.28525722e-01 1.08996642e+00 -5.99401057e-01 -5.70503533e-01
3.17436248e-01 -1.79664269e-01 -2.12754577e-01 4.21643466e-01
4.33040522e-02 6.66973472e-01 3.84200126e-01 -9.20936912e-02
7.49766175e-03 1.93471789e-01 1.02896839e-01 -4.42843765e-01
1.40354753e+00 6.18017733e-01 -2.47414783e-01 4.16081488e-01
7.32712030e-01 4.94714491e-02 -1.26670682e+00 2.72613794e-01
1.50734290e-01 -4.65403557e-01 -6.37997985e-01 -4.94091064e-01
-1.02459013e+00 3.43608677e-01 -9.55494046e-02 7.24343777e-01
8.40837538e-01 -3.74714196e-01 7.53623247e-01 5.12207329e-01
8.86566937e-01 -9.96307969e-01 4.00128245e-01 7.36700535e-01
4.75587100e-01 -6.04108453e-01 4.44298424e-03 -9.62926224e-02
-3.48677278e-01 1.18546391e+00 6.37435615e-01 -5.50654173e-01
3.72047573e-01 3.71579885e-01 -3.25878672e-02 -1.71809867e-01
-9.60450470e-01 -2.65216324e-02 2.11702317e-01 4.24661338e-01
2.84584284e-01 2.83057123e-01 -4.29218501e-01 5.41499138e-01
-1.06320930e+00 -3.43774259e-01 5.45254946e-01 7.33270288e-01
-6.22622490e-01 -1.51348448e+00 -4.11736101e-01 7.72069395e-01
-5.09365976e-01 -4.89111170e-02 -4.03737336e-01 9.34549153e-01
-3.12431855e-03 7.49329329e-01 2.25487605e-01 -6.06083333e-01
6.42355740e-01 -2.13379428e-01 6.34423256e-01 -5.11257827e-01
-1.38126779e+00 -2.99979508e-01 6.67901218e-01 -6.31065309e-01
1.12915330e-01 -5.96560061e-01 -1.07943738e+00 -7.51722157e-01
-1.77643970e-01 4.72951770e-01 3.74551058e-01 6.76605225e-01
-1.91742286e-01 5.75362623e-01 4.10028785e-01 -7.87407398e-01
-7.72690654e-01 -5.92428684e-01 -8.63328159e-01 2.43308902e-01
8.14511906e-03 -6.03654861e-01 6.22836724e-02 -8.48543167e-01] | [3.4473025798797607, 1.434889793395996] |
c9836cff-4393-4204-9d84-bad432441681 | speed-estimation-evaluation-on-the-kitti | 1907.06989 | null | https://arxiv.org/abs/1907.06989v1 | https://arxiv.org/pdf/1907.06989v1.pdf | Speed estimation evaluation on the KITTI benchmark based on motion and monocular depth information | In this technical report we investigate speed estimation of the ego-vehicle on the KITTI benchmark using state-of-the-art deep neural network based optical flow and single-view depth prediction methods. Using a straightforward intuitive approach and approximating a single scale factor, we evaluate several application schemes of the deep networks and formulate meaningful conclusions such as: combining depth information with optical flow improves speed estimation accuracy as opposed to using optical flow alone; the quality of the deep neural network methods influences speed estimation performance; using the depth and optical flow results from smaller crops of wide images degrades performance. With these observations in mind, we achieve a RMSE of less than 1 m/s for vehicle speed estimation using monocular images as input from recordings of the KITTI benchmark. Limitations and possible future directions are discussed as well. | ['Róbert-Adrian Rill'] | 2019-07-16 | null | null | null | null | ['vehicle-speed-estimation'] | ['computer-vision'] | [-5.41750193e-01 -2.45666265e-01 -3.42722297e-01 -2.63394028e-01
6.34206459e-02 -5.37386596e-01 5.03609955e-01 -5.30000031e-01
-6.32475972e-01 7.52884269e-01 1.54001325e-01 -4.78713036e-01
2.06597522e-02 -4.06571746e-01 -7.43563473e-01 -5.68287313e-01
-3.91151756e-01 9.43296477e-02 5.02944328e-02 -1.17898531e-01
4.34831411e-01 8.85900497e-01 -1.54481494e+00 -2.06276357e-01
4.00540113e-01 1.15000713e+00 -1.99960396e-01 1.37970626e+00
2.58285195e-01 1.15362287e+00 -5.58757186e-01 -3.84191573e-01
7.18107820e-01 1.46090677e-02 -8.76037955e-01 -2.38391429e-01
1.29837132e+00 -1.38858366e+00 -1.08467209e+00 5.88138461e-01
2.90543079e-01 2.24442288e-01 4.28500503e-01 -1.40849233e+00
-1.22253627e-01 5.01552410e-02 -3.43677491e-01 9.58569467e-01
2.18153641e-01 6.47085011e-01 7.03858197e-01 -5.98992348e-01
9.23319519e-01 1.25691009e+00 9.21298563e-01 5.19512057e-01
-8.62351716e-01 -5.37136614e-01 2.20669946e-03 6.65411532e-01
-1.08070624e+00 -6.90943778e-01 5.73628247e-01 -6.16717100e-01
1.22758234e+00 -2.94543713e-01 7.46123612e-01 9.62042332e-01
5.45283496e-01 7.50291705e-01 4.25875336e-01 2.52625018e-01
-7.50686005e-02 -3.03457379e-02 5.43450415e-02 7.89984405e-01
5.11213839e-01 7.70673752e-01 -3.39964747e-01 3.01312625e-01
9.59384382e-01 -3.87523651e-01 -1.43087357e-01 -4.65010464e-01
-1.23453176e+00 8.46972406e-01 5.89636087e-01 -1.95463791e-01
-2.05695406e-01 8.33679199e-01 6.67999148e-01 3.30287039e-01
4.16740149e-01 2.23737121e-01 -5.18018544e-01 -6.09437346e-01
-1.01221669e+00 4.84076709e-01 6.99188352e-01 9.75801229e-01
9.13377106e-01 5.69071531e-01 1.36794209e-01 3.78598571e-02
1.95407465e-01 6.45726800e-01 1.34545997e-01 -1.97127676e+00
7.89111733e-01 -5.40589243e-02 4.66578454e-01 -1.12750506e+00
-6.84492946e-01 -4.24064249e-01 -3.51985455e-01 6.15556300e-01
9.26986217e-01 -6.64369285e-01 -5.53292632e-01 1.49545479e+00
5.08020893e-02 4.75237578e-01 2.00810343e-01 1.27835560e+00
7.92598367e-01 4.66191709e-01 -2.61705518e-01 -3.59297842e-02
7.80984402e-01 -1.09128487e+00 -6.62330270e-01 -3.03417832e-01
9.74757612e-01 -7.38688409e-01 2.66948551e-01 2.25442857e-01
-1.13577425e+00 -8.86258543e-01 -1.19721138e+00 -1.86700791e-01
-2.12391213e-01 2.12376527e-02 7.72660971e-01 5.82983613e-01
-1.29000163e+00 9.26419854e-01 -1.04247665e+00 -3.53633493e-01
3.52093607e-01 2.65593737e-01 -3.69830102e-01 -1.17288612e-01
-1.06280446e+00 1.05560923e+00 -5.85356355e-02 2.99680978e-01
-1.04536951e+00 -1.07314968e+00 -9.64789689e-01 -1.50854185e-01
-7.51763806e-02 -1.05977535e+00 1.38394940e+00 -7.98511863e-01
-1.62447441e+00 4.54518974e-01 -4.24130559e-01 -8.76885116e-01
8.17314804e-01 -6.90500259e-01 -2.93763548e-01 6.11268580e-01
-1.35762125e-01 9.50988233e-01 5.32102346e-01 -1.04471862e+00
-8.46116781e-01 -2.09034577e-01 3.84838969e-01 1.27595782e-01
2.60712743e-01 -3.22523952e-01 -8.57556686e-02 2.85805184e-02
-2.54503936e-01 -9.49430466e-01 -7.90528283e-02 4.22422528e-01
-1.39830858e-01 3.33308995e-01 1.14229345e+00 -5.69299161e-01
8.20159137e-01 -1.70773244e+00 -8.33403841e-02 -2.72074223e-01
3.92757893e-01 2.60054648e-01 -3.27426195e-01 1.38518810e-01
9.92505550e-02 -1.13239042e-01 4.75114137e-01 -3.32510054e-01
-3.95304650e-01 1.02159992e-01 -4.42156456e-02 8.74944031e-01
-5.25658466e-02 1.05571461e+00 -1.10783255e+00 -3.41540098e-01
8.38756800e-01 6.10135853e-01 -7.38332212e-01 1.08790383e-01
2.53399730e-01 4.63011056e-01 -1.45982459e-01 3.16857845e-01
8.97959232e-01 -1.69662107e-02 -3.74375224e-01 -5.02552986e-01
-3.13217938e-01 2.97447175e-01 -9.48145747e-01 1.49554002e+00
-6.16106868e-01 1.64829302e+00 7.00871721e-02 -5.89776754e-01
5.89940846e-01 4.62464280e-02 8.19047332e-01 -6.36130750e-01
2.77698934e-01 -9.77007523e-02 1.44186720e-01 -6.25956893e-01
7.23779261e-01 9.35736746e-02 5.91109753e-01 1.29065007e-01
1.33265376e-01 -4.24968377e-02 3.81752074e-01 1.76621705e-01
1.04962146e+00 3.20614994e-01 -1.85101658e-01 -1.61843613e-01
4.25373793e-01 1.32225677e-01 3.98644567e-01 6.73753917e-01
-1.01766133e+00 4.46668178e-01 4.37579483e-01 -1.05801976e+00
-1.21646023e+00 -1.00326943e+00 -1.47094712e-01 4.22653377e-01
5.32262444e-01 -1.12033740e-01 -5.76497614e-01 -5.91470242e-01
2.66856879e-01 5.20876169e-01 -7.02840030e-01 3.35249119e-02
-8.34961414e-01 -2.16296777e-01 7.34342337e-01 7.34089375e-01
6.13788128e-01 -4.72244740e-01 -8.40478420e-01 2.07529470e-01
-1.47177219e-01 -1.84509766e+00 -2.95989037e-01 -2.82437325e-01
-1.04570079e+00 -1.19722950e+00 -5.81422567e-01 -3.35988700e-01
1.67509779e-01 7.43430078e-01 1.28559136e+00 -1.22554474e-01
5.23397550e-02 3.95779163e-01 -1.57661084e-02 -7.23972451e-03
-3.90808672e-01 3.46158221e-02 1.61435306e-01 -2.78501779e-01
3.58981222e-01 -3.73192728e-01 -1.19738710e+00 3.96821678e-01
-2.80980051e-01 -6.38386086e-02 2.10607931e-01 4.42999601e-01
-1.98734358e-01 -3.01938385e-01 -4.73011145e-03 -3.34448576e-01
-3.04078180e-02 -3.02413732e-01 -9.83900189e-01 -6.29906237e-01
-4.81295168e-01 -4.84578349e-02 6.97594941e-01 -2.16358379e-01
-9.71737146e-01 -9.10603702e-02 -5.02890833e-02 -9.20345902e-01
-8.23391378e-02 -5.69207110e-02 5.73401630e-01 -4.24697310e-01
5.75445950e-01 -8.53352100e-02 3.58600020e-01 8.04636776e-02
5.24487138e-01 1.45795688e-01 6.25886500e-01 1.03780022e-02
7.79675007e-01 1.03341568e+00 3.40829432e-01 -9.66904163e-01
-6.71137393e-01 -5.21024466e-01 -7.51578152e-01 -7.07406163e-01
8.51599693e-01 -1.18871963e+00 -1.23419929e+00 5.72947919e-01
-1.46766722e+00 -4.41514373e-01 4.02374901e-02 9.69856441e-01
-9.42276359e-01 3.80204529e-01 -1.01726639e+00 -4.65256572e-01
-1.26158133e-01 -1.37247312e+00 9.93402481e-01 3.21759969e-01
-4.32502031e-02 -1.45091355e+00 1.05326571e-01 3.67927313e-01
6.47160351e-01 1.72878936e-01 1.44652501e-01 1.98106796e-01
-9.49571848e-01 -7.64547661e-02 -5.04421830e-01 3.66158366e-01
-7.51463547e-02 4.26368713e-01 -1.13995969e+00 -5.34207582e-01
-2.36820713e-01 -1.23066053e-01 8.59647572e-01 1.03540802e+00
7.80806243e-01 -6.64326549e-02 -2.77829856e-01 1.13186872e+00
1.57452822e+00 3.97071205e-02 7.78598070e-01 4.75148469e-01
9.06870544e-01 4.44007695e-01 4.56537634e-01 3.53607595e-01
4.69221503e-01 6.12485707e-01 6.94710314e-01 3.43779698e-02
-3.04195732e-01 -5.37816994e-02 3.73387009e-01 3.70199054e-01
-3.17814887e-01 -1.57522127e-01 -7.28869855e-01 5.77025354e-01
-1.44555247e+00 -1.32941818e+00 -2.72307724e-01 1.97487819e+00
-1.02797851e-01 3.05802256e-01 1.40664324e-01 1.41287027e-02
3.78535926e-01 4.75701898e-01 -4.55833942e-01 -6.22584641e-01
1.02751464e-01 -5.12495995e-01 1.19147599e+00 1.03551936e+00
-1.08180821e+00 9.80757833e-01 7.68389511e+00 1.87391698e-01
-1.54319870e+00 -9.22367051e-02 5.22349954e-01 -3.56711060e-01
4.34161015e-02 -6.34651482e-02 -1.02677619e+00 3.66113693e-01
1.30095208e+00 -6.04132786e-02 4.52982873e-01 7.63579488e-01
7.75798082e-01 -3.57427180e-01 -1.06891513e+00 1.07952082e+00
-1.23055503e-01 -1.64756513e+00 -1.66021675e-01 3.02006215e-01
8.02710831e-01 8.33546400e-01 1.46114826e-03 3.52164470e-02
9.13524851e-02 -8.01609576e-01 6.37365341e-01 4.07246470e-01
6.10674500e-01 -7.23092556e-01 1.00257659e+00 4.75668013e-02
-1.27788520e+00 -1.58553466e-01 -4.90652859e-01 -4.86705929e-01
4.08876359e-01 3.88196379e-01 -6.68433845e-01 4.29227710e-01
6.14435911e-01 1.22451043e+00 -3.69502038e-01 1.14851236e+00
1.40019536e-01 5.10640383e-01 -2.74898142e-01 1.77920818e-01
5.96300602e-01 -6.45256639e-02 8.04866433e-01 1.18931103e+00
1.34630665e-01 -3.58506620e-01 -2.63714015e-01 8.20986927e-01
2.09897503e-01 -5.39798260e-01 -9.37110245e-01 2.84330547e-01
1.98328108e-01 1.20667195e+00 -2.33632401e-01 -3.93512636e-01
-6.68175578e-01 5.72841406e-01 1.13489762e-01 7.73189068e-01
-9.26273942e-01 -2.57210016e-01 1.54532862e+00 1.83185950e-01
3.60930800e-01 -5.80493569e-01 -3.09462249e-01 -1.17633319e+00
-2.01022848e-01 -3.76834683e-02 -1.86097741e-01 -1.00525093e+00
-5.56248426e-01 4.86502558e-01 8.50317702e-02 -1.61669469e+00
-6.06296539e-01 -9.94000077e-01 -5.60561121e-01 6.85803890e-01
-2.04643273e+00 -4.86650288e-01 -7.15228200e-01 1.41105741e-01
5.36343157e-01 -1.18449129e-01 1.45283684e-01 4.89593506e-01
-4.70762074e-01 4.57682282e-01 1.77427009e-01 3.00743192e-01
4.77992773e-01 -9.49424446e-01 9.76108789e-01 9.34053957e-01
-2.90252447e-01 2.72767901e-01 1.02051759e+00 -1.41987145e-01
-1.55025268e+00 -9.32122529e-01 7.34837651e-01 -8.04094493e-01
8.62488329e-01 2.39635840e-01 -4.57839966e-01 7.34348834e-01
3.25892538e-01 5.11043906e-01 -7.01736510e-02 -3.35563689e-01
-1.34152383e-01 -2.45722502e-01 -8.90787303e-01 4.61987197e-01
9.63367701e-01 -3.81666243e-01 -8.85126367e-03 -1.45721054e-02
5.57642937e-01 -6.43134952e-01 -7.78691113e-01 2.40769744e-01
9.10755098e-01 -1.37506938e+00 1.13499844e+00 -5.19878805e-01
4.98491168e-01 -2.51401126e-01 3.53062227e-02 -1.36221969e+00
-1.95585832e-01 -6.91515803e-01 -4.62577730e-01 3.49377781e-01
1.72696635e-01 -5.72894633e-01 1.26826704e+00 2.80698389e-01
-1.95790544e-01 -4.53523427e-01 -9.76433277e-01 -7.92518258e-01
1.94926038e-01 -7.08626986e-01 1.42706603e-01 5.12088180e-01
-3.02765608e-01 2.74738073e-01 -5.11494100e-01 2.58228213e-01
8.50595534e-01 -3.18429172e-01 1.18174016e+00 -9.96239007e-01
9.03841779e-02 -5.27098477e-01 -8.98933709e-01 -1.64778304e+00
3.83495837e-01 -2.75322825e-01 -2.46105880e-01 -1.33107090e+00
-2.92841822e-01 -1.94884054e-02 1.05094485e-01 -2.99600124e-01
8.26461837e-02 2.55672604e-01 3.26194644e-01 -8.50539282e-02
-4.14036989e-01 3.17160398e-01 1.71037292e+00 -1.64357051e-01
7.61863440e-02 8.93938467e-02 -8.48064795e-02 6.43906951e-01
7.13261902e-01 -8.90048221e-02 -4.08664852e-01 -8.54576766e-01
-5.66048920e-02 4.85302061e-01 6.33830547e-01 -1.31582177e+00
1.38670817e-01 -1.15312517e-01 6.30614340e-01 -7.28309691e-01
5.18179595e-01 -7.13337719e-01 -2.61168808e-01 7.42171705e-01
-7.71358311e-02 3.27454299e-01 5.37578702e-01 4.21356529e-01
-1.25126019e-01 1.42662466e-01 8.92173469e-01 -8.27708188e-03
-1.10296404e+00 4.96969432e-01 -6.46033704e-01 9.46256369e-02
6.65614665e-01 -6.36655807e-01 -6.50030553e-01 -8.25320005e-01
-3.62037301e-01 1.62054762e-01 4.86163199e-01 6.34210229e-01
4.58897531e-01 -1.28343976e+00 -6.91716909e-01 1.95316225e-01
-8.94558728e-02 -4.04708117e-01 3.06774765e-01 7.92482138e-01
-1.34040582e+00 9.41914856e-01 -4.88160640e-01 -9.82925236e-01
-8.54661703e-01 4.17340547e-01 9.48467135e-01 1.12955108e-01
-4.45651770e-01 5.99007785e-01 1.49510786e-01 -1.12497598e-01
1.52673230e-01 -6.01547658e-01 -1.79105923e-01 -2.95479894e-01
5.69378316e-01 9.44278121e-01 -4.92714942e-02 -9.83836055e-01
-3.22485477e-01 8.54781628e-01 1.67879134e-01 8.73494819e-02
7.74045467e-01 -6.03366494e-01 5.66058278e-01 1.76610097e-01
1.74952829e+00 -4.27009642e-01 -1.89563513e+00 1.33974239e-01
-4.84323949e-01 -6.15803063e-01 5.85484922e-01 -2.73043364e-01
-1.50058222e+00 1.18037212e+00 8.47345114e-01 -8.28071609e-02
5.66996992e-01 -4.71755415e-01 9.55231488e-01 4.83458430e-01
2.19980448e-01 -9.06852841e-01 -1.20126702e-01 8.23773444e-01
3.58018935e-01 -1.62928116e+00 3.01405713e-02 -2.25785345e-01
-2.78115720e-01 1.44936752e+00 8.92753959e-01 -3.90872657e-01
5.82398236e-01 2.77276605e-01 5.11839569e-01 1.06074721e-01
-9.22813296e-01 -1.37021258e-01 1.10878475e-01 5.77400386e-01
2.77511179e-01 -2.69797891e-01 1.50946483e-01 -6.23829424e-01
-1.66732997e-01 1.77975014e-01 1.00185907e+00 5.96833229e-01
-3.95564973e-01 -3.07023436e-01 -1.79353833e-01 2.08711743e-01
-3.64594609e-01 1.23045094e-01 2.31826261e-01 1.14651656e+00
-2.29773954e-01 9.57049251e-01 5.21755576e-01 -5.48663318e-01
1.98041603e-01 -4.81219590e-01 4.68972981e-01 1.79462522e-01
-1.38073772e-01 -3.36742669e-01 2.68820435e-01 -1.07212532e+00
-6.12672806e-01 -5.32122552e-01 -1.07589877e+00 -1.22309935e+00
2.86275279e-02 -2.03003049e-01 8.11368227e-01 1.05645454e+00
3.23793620e-01 3.39135140e-01 6.79202914e-01 -1.52217853e+00
-2.16353804e-01 -7.84980655e-01 -2.46827379e-01 2.11928621e-01
1.09683776e+00 -6.38122201e-01 -8.57380331e-01 3.59914564e-02] | [8.69487190246582, -1.815500259399414] |
0561f9f4-4ba5-441b-85be-cc5f890d24ca | evolutionary-algorithm-enhanced-neural | 2008.05695 | null | https://arxiv.org/abs/2008.05695v1 | https://arxiv.org/pdf/2008.05695v1.pdf | Evolutionary Algorithm Enhanced Neural Architecture Search for Text-Independent Speaker Verification | State-of-the-art speaker verification models are based on deep learning techniques, which heavily depend on the handdesigned neural architectures from experts or engineers. We borrow the idea of neural architecture search(NAS) for the textindependent speaker verification task. As NAS can learn deep network structures automatically, we introduce the NAS conception into the well-known x-vector network. Furthermore, this paper proposes an evolutionary algorithm enhanced neural architecture search method called Auto-Vector to automatically discover promising networks for the speaker verification task. The experimental results demonstrate our NAS-based model outperforms state-of-the-art speaker verification models. | ['Jing Xiao', 'Jianzong Wang', 'Xiaoyang Qu'] | 2020-08-13 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [ 2.16165688e-02 -7.70670846e-02 -4.49456722e-02 -7.92503238e-01
-5.73721707e-01 -2.08682030e-01 3.12061280e-01 -6.37255490e-01
-2.08880723e-01 2.77588367e-01 1.97875395e-01 -6.14530385e-01
-6.49325997e-02 -2.47937560e-01 -4.64356810e-01 -7.75205374e-01
1.67062759e-01 3.67478371e-01 -3.72833222e-01 -4.98543024e-01
2.25366309e-01 3.92579138e-01 -1.52621424e+00 -8.77874345e-02
8.02677393e-01 1.08566010e+00 -1.28586307e-01 6.37375534e-01
-5.08678555e-01 4.65497702e-01 -7.05535710e-01 -5.25792122e-01
6.08311221e-02 -4.81986076e-01 -5.64536273e-01 -2.76382744e-01
2.34233007e-01 -3.01271193e-02 -4.44369495e-01 1.13312936e+00
7.62777686e-01 4.79934923e-02 5.74803591e-01 -1.42042255e+00
-1.05634260e+00 1.23052061e+00 -2.12014124e-01 2.81522900e-01
-1.08419709e-01 5.34885004e-02 1.05203903e+00 -1.12896466e+00
-4.24833922e-03 1.34072828e+00 7.84196436e-01 1.05978322e+00
-9.04580832e-01 -1.12588036e+00 4.59189355e-01 7.43778408e-01
-1.70229435e+00 -1.10273623e+00 1.56980300e+00 -2.69568145e-01
8.84894550e-01 1.01240605e-01 2.50876516e-01 1.27631187e+00
-2.14426398e-01 1.13661861e+00 5.36216438e-01 -6.48645163e-01
4.50214237e-01 3.33456069e-01 4.86716747e-01 8.35647762e-01
-1.04673952e-01 6.74489141e-01 -8.80609691e-01 -1.33107305e-01
3.79930973e-01 -3.44220519e-01 -4.05342311e-01 -7.21446797e-02
-7.20532954e-01 1.05896783e+00 5.16326427e-01 6.11443460e-01
-4.23219711e-01 -7.70611987e-02 4.77456063e-01 4.61013436e-01
8.03784057e-02 3.54884982e-01 -6.81082129e-01 1.37535438e-01
-1.17656088e+00 -1.38261810e-01 6.04910254e-01 6.49924934e-01
5.10598779e-01 9.72854912e-01 1.79584220e-03 9.04796898e-01
8.65759194e-01 2.11801052e-01 1.05516756e+00 -3.70944172e-01
2.25783914e-01 7.70705044e-01 -5.03073692e-01 -5.38224339e-01
-1.86447367e-01 -1.00300312e+00 -1.07232130e+00 1.00507781e-01
-3.51012141e-01 -3.20490062e-01 -8.57618332e-01 1.84272873e+00
2.31045991e-01 3.99823308e-01 3.89441639e-01 7.96112180e-01
1.15108109e+00 6.29196167e-01 -9.96221751e-02 -2.01982453e-01
1.14873707e+00 -1.17304170e+00 -8.25354397e-01 -2.12160394e-01
2.09376618e-01 -2.52556890e-01 9.46360528e-01 3.27074349e-01
-8.44736338e-01 -1.01336479e+00 -1.36148214e+00 3.87688339e-01
-3.67330283e-01 2.66738504e-01 3.48821253e-01 1.33071339e+00
-1.28990114e+00 3.03229958e-01 -6.14068985e-01 -9.22793895e-02
5.14592052e-01 5.08169174e-01 -1.13991015e-01 3.20458263e-01
-1.27821875e+00 8.63663375e-01 6.20797634e-01 5.40101051e-01
-1.05748320e+00 -5.85522473e-01 -9.43319559e-01 3.95152897e-01
8.41737464e-02 -6.27804816e-01 1.40992188e+00 -1.06258035e+00
-2.30193543e+00 6.03841007e-01 -5.84104598e-01 -5.94528139e-01
-1.30707979e-01 -1.32517237e-03 -8.68940294e-01 -2.55462050e-01
-6.97935283e-01 4.23177809e-01 1.24880469e+00 -1.23403561e+00
-2.92315692e-01 -4.51476753e-01 -5.78905821e-01 -1.75687745e-01
-9.47683096e-01 4.42764372e-01 -7.79506564e-02 -6.41610086e-01
2.66543508e-01 -6.80702984e-01 -1.44216686e-01 -1.86781660e-01
-5.95049322e-01 -6.74803853e-01 1.09620070e+00 -6.95805788e-01
1.43494678e+00 -2.12190151e+00 3.41382056e-01 4.00131047e-01
5.99064641e-02 5.65034389e-01 -4.04026181e-01 2.65574548e-02
-3.95619988e-01 3.41807827e-02 -1.85058162e-01 -6.28033638e-01
2.96088219e-01 -6.29291730e-03 -3.91843557e-01 3.39686871e-01
7.37711936e-02 1.00686836e+00 -1.99617326e-01 -4.24404025e-01
-4.83000837e-03 6.32730007e-01 -3.68755907e-01 3.13490123e-01
-9.06552095e-03 2.60152578e-01 -3.42355341e-01 8.70765865e-01
5.59546471e-01 -1.39808387e-01 1.49687044e-02 -9.54301581e-02
8.88477787e-02 2.91401923e-01 -1.02550566e+00 1.93593049e+00
-3.47607851e-01 6.42105222e-01 3.11636537e-01 -1.00865293e+00
1.38397098e+00 6.61026597e-01 -1.87596262e-01 -1.56969935e-01
6.06041729e-01 2.49013007e-01 2.45832995e-01 -2.03082383e-01
2.07424447e-01 -2.63364911e-01 2.73397416e-01 5.14549196e-01
3.20136845e-01 4.11074340e-01 -4.91069913e-01 -1.96616501e-01
5.66578865e-01 -3.04553241e-01 2.25233272e-01 -1.67788938e-01
1.19006002e+00 -4.48458821e-01 7.74132252e-01 5.64528048e-01
-4.66811776e-01 1.25719860e-01 -1.44988641e-01 -6.45984113e-01
-5.42093933e-01 -7.51768172e-01 -5.07342219e-02 1.33939004e+00
-3.80984753e-01 -1.81206882e-01 -9.07316148e-01 -8.24340224e-01
-2.32589111e-01 9.17605639e-01 -5.47625244e-01 -3.54308486e-01
-6.43656373e-01 -3.54237556e-01 9.86701012e-01 6.83253884e-01
4.89685029e-01 -1.19542253e+00 3.46051008e-02 3.62303704e-01
1.36478662e-01 -5.79901516e-01 -5.53697824e-01 5.99835217e-02
-8.01155150e-01 -5.29813051e-01 -8.01605582e-01 -1.11519742e+00
4.50527221e-01 -1.60952672e-01 8.47291231e-01 3.27824146e-01
2.13245284e-02 4.71966863e-02 -2.28702635e-01 -6.67656243e-01
-7.08450198e-01 4.10949945e-01 6.01005018e-01 4.17593330e-01
7.75650501e-01 -7.26563871e-01 -1.97191894e-01 4.00628328e-01
-3.77740741e-01 -4.43088681e-01 7.38428056e-01 1.13197410e+00
3.30766231e-01 1.32359743e-01 1.06073391e+00 -1.24754556e-01
7.34965801e-01 -7.36286715e-02 -8.78938496e-01 5.45711815e-01
-8.67107213e-01 5.36418319e-01 3.72893333e-01 -5.42949140e-01
-9.28007483e-01 3.80294800e-01 -3.51926208e-01 -1.06931472e+00
-2.77778119e-01 7.70761371e-01 -5.53345740e-01 -2.94480473e-01
7.13547111e-01 8.28654706e-01 7.43816048e-02 -6.76227450e-01
3.25204760e-01 9.20868993e-01 6.05313957e-01 -1.26049325e-01
9.19482768e-01 -1.42629787e-01 -4.69032854e-01 -8.58918965e-01
-4.65873152e-01 -3.33958000e-01 -5.57317615e-01 -1.23925075e-01
7.20086038e-01 -7.80194223e-01 -1.00497913e+00 5.37369490e-01
-1.40434837e+00 1.82516754e-01 1.03362612e-01 4.35654014e-01
-1.44935608e-01 1.99397102e-01 -3.99028242e-01 -1.19821513e+00
-1.04252052e+00 -1.33425438e+00 8.07116449e-01 4.47891712e-01
-1.34341449e-01 -8.02527249e-01 2.02154577e-01 2.93311715e-01
8.80818248e-01 -7.35862195e-01 9.46378529e-01 -1.24234235e+00
-5.07670283e-01 -1.97505057e-01 2.31760502e-01 4.81713563e-01
-1.80753320e-02 -8.84719864e-02 -1.45622730e+00 -3.77952039e-01
3.25263202e-01 -4.81460579e-02 8.04560184e-01 5.19519746e-01
1.34919262e+00 -5.93074203e-01 -4.19010699e-01 7.23480344e-01
1.04355228e+00 3.29617113e-01 3.38900506e-01 1.48229450e-01
6.00396633e-01 6.49776340e-01 -1.85292974e-01 5.30480854e-02
3.39676112e-01 7.27488875e-01 2.20003530e-01 9.58954021e-02
2.26369992e-01 -2.69684196e-01 3.77933651e-01 1.30025268e+00
2.16792971e-01 -2.79197305e-01 -9.78025496e-01 5.64304292e-01
-1.58769166e+00 -1.09533358e+00 5.40246964e-01 1.72518969e+00
5.99493921e-01 -9.70564559e-02 7.39407167e-02 6.36779785e-01
1.01122856e+00 8.81620403e-03 -7.96827853e-01 -3.82997572e-01
-8.76116008e-02 2.63473213e-01 -6.88915253e-02 4.54730123e-01
-9.43996370e-01 9.12747562e-01 6.89435244e+00 7.54972875e-01
-1.34958494e+00 1.22041002e-01 4.58898813e-01 -3.99938673e-02
-1.69042036e-01 -3.49702269e-01 -1.18605387e+00 2.39930362e-01
1.26287556e+00 -2.35708088e-01 4.30122077e-01 1.28807223e+00
-9.26065668e-02 1.19069290e+00 -1.16653883e+00 1.39147067e+00
4.47206289e-01 -1.55707479e+00 1.93732783e-01 7.36903399e-02
4.60544646e-01 1.75274298e-01 2.23801062e-01 6.86383426e-01
3.00519168e-01 -1.24498165e+00 6.09423578e-01 2.48659641e-01
5.06421268e-01 -8.00906837e-01 7.65529215e-01 1.74114197e-01
-1.29304874e+00 -5.89841962e-01 -2.74261057e-01 4.97606725e-01
3.48771900e-01 1.88327909e-01 -9.91418779e-01 3.03541332e-01
5.75728655e-01 3.86158168e-01 -4.88090158e-01 1.10055423e+00
-4.27351207e-01 1.09220731e+00 -1.11826658e-01 -3.22836429e-01
1.57916442e-01 2.27609187e-01 6.31056130e-01 8.24774384e-01
3.72686237e-01 -1.09378062e-01 -2.55404770e-01 1.03882277e+00
-2.91084439e-01 7.25332722e-02 -2.74275333e-01 -7.54440948e-02
8.41275394e-01 9.32956874e-01 9.66755114e-03 -2.75255173e-01
-6.22691140e-02 7.00585723e-01 8.78573582e-02 3.83899510e-01
-5.96140504e-01 -6.41012907e-01 7.27159202e-01 -3.82987887e-01
6.82057023e-01 1.37544081e-01 -5.03990591e-01 -9.35039699e-01
2.52210740e-02 -1.31840956e+00 2.74174690e-01 -6.07160628e-01
-1.54325044e+00 1.23699999e+00 -6.31543398e-01 -1.11707032e+00
-6.09084785e-01 -5.17652452e-01 -1.13577056e+00 8.87459457e-01
-1.78179145e+00 -1.17341053e+00 -5.02274856e-02 7.65262663e-01
7.51134455e-01 -1.29863715e+00 1.12582028e+00 1.85123652e-01
-9.10889030e-01 1.33917189e+00 -1.59322470e-01 5.04673243e-01
2.47881964e-01 -7.51077473e-01 9.18490112e-01 7.57846534e-01
4.83835280e-01 7.79739201e-01 4.53013480e-01 -1.22528590e-01
-1.57935977e+00 -8.45349491e-01 1.06167269e+00 -8.49594325e-02
4.48211253e-01 -2.97494680e-01 -1.10535800e+00 5.88470101e-01
4.34850931e-01 -1.92466214e-01 9.34286118e-01 5.23157477e-01
-6.67832494e-01 -2.33643487e-01 -1.11322665e+00 4.27103847e-01
8.85241568e-01 -7.49792993e-01 -8.96126151e-01 8.72077607e-03
9.87820506e-01 6.56372830e-02 -3.73226494e-01 3.65577489e-01
4.78032887e-01 -5.44457018e-01 1.06015742e+00 -7.93727160e-01
-1.44572362e-01 -3.79911482e-01 -2.73817927e-01 -1.40140927e+00
-4.56975400e-01 -7.12635756e-01 -3.19904119e-01 1.35414684e+00
9.00642335e-01 -7.34605849e-01 1.03276587e+00 4.84419763e-01
-3.67654026e-01 -5.54205000e-01 -1.45504498e+00 -8.73606563e-01
-1.01966180e-01 -3.40610921e-01 1.31041574e+00 1.15625966e+00
-9.45013687e-02 6.82450831e-01 -1.73139334e-01 4.52675223e-01
6.63604438e-01 -5.54154664e-02 5.10387301e-01 -1.58077502e+00
-5.52888691e-01 -9.70712662e-01 -6.22684598e-01 -9.22113776e-01
7.24328876e-01 -8.74761224e-01 1.57735661e-01 -1.02050769e+00
-1.57471791e-01 -2.34724030e-01 -7.73490131e-01 6.83233142e-01
-1.89735994e-01 -3.09470952e-01 -4.91508022e-02 4.89979014e-02
-2.17885360e-01 1.15894043e+00 4.53814089e-01 -5.57250440e-01
-5.45735717e-01 3.09940130e-01 -9.09906983e-01 6.09924555e-01
1.00398338e+00 -4.44936395e-01 -4.41755503e-01 -4.86890554e-01
-3.08756143e-01 4.28472273e-02 3.45809013e-02 -9.07431901e-01
5.91441095e-01 2.46037334e-01 2.64538318e-01 -8.47050726e-01
4.98803973e-01 -7.45139182e-01 -1.61419258e-01 5.67933798e-01
-5.80293477e-01 2.01592594e-01 3.39708358e-01 3.75437617e-01
-4.77291435e-01 -2.26330131e-01 7.35871077e-01 7.07593560e-02
-5.58313251e-01 3.17231357e-01 -2.50704527e-01 -4.71512824e-01
5.17510235e-01 -1.86378270e-01 -2.05945358e-01 -3.98738593e-01
-6.05415583e-01 9.72639024e-02 -2.92274475e-01 7.60051012e-01
9.43454206e-01 -1.45975459e+00 -1.11207306e+00 6.17278039e-01
1.40837729e-01 -5.16619802e-01 3.45000982e-01 2.74886042e-01
9.73473489e-02 6.09841704e-01 2.17796378e-02 -8.00926626e-01
-1.58389652e+00 6.03742480e-01 6.22572064e-01 3.43517289e-02
-9.61302072e-02 1.37953568e+00 5.36360731e-03 -9.40675855e-01
6.68749273e-01 1.44411907e-01 -2.33539522e-01 -3.11425835e-01
8.93567383e-01 7.01346248e-02 2.04735503e-01 -6.57112420e-01
-6.67625427e-01 3.78791273e-01 -3.83007497e-01 -2.85439700e-01
1.37923956e+00 6.23930283e-02 1.01091906e-01 5.13785295e-02
1.15560961e+00 -4.67953116e-01 -6.57502234e-01 -5.95085323e-01
-5.69017939e-02 1.42268926e-01 5.98173678e-01 -9.07950759e-01
-1.37564301e+00 1.05662847e+00 1.01116192e+00 -1.74284235e-01
1.10613227e+00 -2.35633180e-01 7.60254741e-01 8.18257928e-01
1.47222295e-01 -8.97061288e-01 -1.80155501e-01 4.50297534e-01
1.15296328e+00 -1.24092126e+00 -2.42165640e-01 2.65844047e-01
-5.62145531e-01 8.76878619e-01 6.80501580e-01 3.44680488e-01
1.01516008e+00 8.04806650e-02 5.29524386e-01 -5.83434291e-02
-7.66591430e-01 3.19762498e-01 5.18867254e-01 5.89193702e-01
2.19280109e-01 -3.10953371e-02 3.02906066e-01 1.05174601e+00
-5.49464762e-01 -1.77400112e-01 -1.31260991e-01 7.27605939e-01
-4.32187140e-01 -1.18117166e+00 -5.77815771e-01 -9.40406770e-02
-4.37248982e-02 -3.27952027e-01 -6.95038795e-01 2.46136293e-01
-3.78672779e-01 1.10281265e+00 -2.32877448e-01 -8.67488444e-01
1.61674619e-01 6.09340370e-01 1.51626676e-01 -9.18159187e-02
-8.39949429e-01 -2.10583702e-01 -2.89726436e-01 -2.02521637e-01
-2.65601814e-01 -4.33283865e-01 -9.23958004e-01 -1.86889857e-01
-8.29517484e-01 3.80966604e-01 1.44559526e+00 9.44774568e-01
6.62312567e-01 5.01590073e-01 8.06680083e-01 -7.46591687e-01
-8.67300570e-01 -1.25996447e+00 -3.24898094e-01 -3.78018379e-01
5.88359535e-01 -4.64358330e-01 -4.64880884e-01 -9.97372195e-02] | [14.317463874816895, 6.046128749847412] |
7db1c107-a431-4e0a-b49f-2121899c90b8 | lexico-acoustic-neural-based-models-for | 1803.00831 | null | http://arxiv.org/abs/1803.00831v1 | http://arxiv.org/pdf/1803.00831v1.pdf | Lexico-acoustic Neural-based Models for Dialog Act Classification | Recent works have proposed neural models for dialog act classification in
spoken dialogs. However, they have not explored the role and the usefulness of
acoustic information. We propose a neural model that processes both lexical and
acoustic features for classification. Our results on two benchmark datasets
reveal that acoustic features are helpful in improving the overall accuracy.
Finally, a deeper analysis shows that acoustic features are valuable in three
cases: when a dialog act has sufficient data, when lexical information is
limited and when strong lexical cues are not present. | ['Daniel Ortega', 'Ngoc Thang Vu'] | 2018-03-02 | null | null | null | null | ['dialog-act-classification'] | ['natural-language-processing'] | [-2.13278085e-01 5.68871386e-02 -2.17160836e-01 -7.38132775e-01
-4.78670150e-01 -5.06284595e-01 8.92004073e-01 2.42631048e-01
-6.66978002e-01 6.49494112e-01 8.27910483e-01 -2.23582640e-01
2.13558465e-01 -4.87211406e-01 7.88559616e-02 -5.38185894e-01
1.50013402e-01 4.57949400e-01 2.42123455e-01 -7.19442904e-01
1.66707575e-01 7.58236870e-02 -1.35762835e+00 3.89778435e-01
5.10060191e-01 1.03949511e+00 2.43209153e-01 6.56815767e-01
-5.73203087e-01 9.35009241e-01 -9.66784358e-01 -1.16903134e-01
1.49140775e-01 -6.40327930e-01 -1.02345884e+00 -4.26314920e-02
-1.01502016e-01 -4.74553823e-01 -2.38352612e-01 6.28332555e-01
5.93279958e-01 3.90333444e-01 6.79054081e-01 -9.46883619e-01
-3.82294595e-01 9.40810621e-01 1.08885862e-01 3.73480707e-01
7.29888320e-01 2.64319688e-01 1.30075479e+00 -1.00487339e+00
2.30159894e-01 1.38584650e+00 3.68559003e-01 7.20250249e-01
-9.78197634e-01 -5.50424874e-01 3.09422374e-01 3.06239396e-01
-7.97064066e-01 -8.27548504e-01 9.00308073e-01 -3.79475951e-01
1.33929503e+00 2.70798773e-01 4.20249194e-01 1.34547603e+00
-1.36136217e-02 8.99864197e-01 1.07500553e+00 -5.07274449e-01
3.64457160e-01 2.89724439e-01 6.46017611e-01 2.64400691e-01
-8.06176305e-01 7.80219361e-02 -9.19684410e-01 -3.57580900e-01
2.60160953e-01 -2.53613502e-01 -2.03434333e-01 3.77573252e-01
-7.35654354e-01 1.34617829e+00 1.57178044e-01 6.03123248e-01
-4.60380942e-01 -2.88410664e-01 5.01644909e-01 7.03591943e-01
4.09591228e-01 5.34498632e-01 -4.54627842e-01 -6.59198821e-01
-5.73050499e-01 8.02907348e-02 1.11713314e+00 4.70286787e-01
5.62156916e-01 2.06563726e-01 -1.14602201e-01 1.31009746e+00
3.79965901e-01 3.01780790e-01 6.98795080e-01 -8.22774827e-01
3.43018711e-01 6.18554950e-01 3.13860700e-02 -5.50023019e-01
-6.49072766e-01 3.70391011e-01 -5.62199891e-01 -8.17010254e-02
5.30118287e-01 -3.88058096e-01 -5.31636238e-01 1.76960218e+00
7.15250745e-02 -1.31092533e-01 3.27256978e-01 9.12952781e-01
1.39593160e+00 7.18656659e-01 2.99170285e-01 -2.53126085e-01
1.40182304e+00 -7.79465318e-01 -1.26707804e+00 -5.07672489e-01
3.69075179e-01 -9.04322624e-01 1.30932462e+00 2.88135231e-01
-1.09876037e+00 -5.22398353e-01 -9.04675722e-01 1.30446285e-01
-4.50829029e-01 5.22030815e-02 6.68737829e-01 7.86681652e-01
-9.32253838e-01 1.24280356e-01 -5.42205095e-01 -3.68990958e-01
-1.73706964e-01 4.44102705e-01 -1.82457805e-01 4.24071133e-01
-1.67509401e+00 1.10571587e+00 2.71126688e-01 1.21428236e-01
-5.83698094e-01 1.27384327e-02 -7.46186435e-01 1.63581550e-01
2.37526238e-01 -1.15707360e-01 1.39292336e+00 -5.63361645e-01
-2.00245214e+00 4.19629186e-01 -2.00996771e-01 -5.64776838e-01
1.34977609e-01 -2.74885595e-01 -2.54768580e-01 5.71311750e-02
-5.75348735e-01 6.86567128e-01 5.74105859e-01 -1.06377804e+00
-5.73616564e-01 -2.67682999e-01 3.66638511e-01 4.81964260e-01
-6.99190676e-01 1.85043335e-01 -1.04546733e-01 -3.61235350e-01
4.89463508e-02 -8.68155062e-01 -1.55856088e-01 -4.46642697e-01
-3.38191152e-01 -6.89211845e-01 8.06464076e-01 -5.67944229e-01
1.15607619e+00 -2.07297492e+00 1.12403490e-01 2.12998968e-02
-3.30950459e-03 3.07537466e-01 1.51529938e-01 6.67096019e-01
5.15641510e-01 2.52642274e-01 4.20522206e-02 -4.65444475e-01
2.50984758e-01 4.44320917e-01 -4.43678021e-01 -9.31409225e-02
1.66259319e-01 6.99600518e-01 -3.99932772e-01 -3.61791492e-01
5.66980064e-01 2.57116258e-01 -3.93877506e-01 6.60429358e-01
-1.74458444e-01 4.57206160e-01 -4.69558448e-01 2.71595389e-01
2.42956251e-01 7.33699501e-02 2.90159583e-01 9.46357772e-02
-5.72090819e-02 9.07674551e-01 -9.89706695e-01 1.34365952e+00
-7.25543737e-01 6.69574499e-01 4.50541466e-01 -6.49420440e-01
1.02635121e+00 6.10263348e-01 2.32131138e-01 -6.78691566e-01
3.86969894e-01 -2.92596351e-02 3.86491895e-01 -4.98719305e-01
6.09862804e-01 -1.65971711e-01 -1.78276420e-01 6.57088935e-01
1.20522290e-01 -3.21133554e-01 3.07875611e-02 5.44738024e-02
1.01371217e+00 -5.80244839e-01 4.64709759e-01 -3.70196365e-02
6.25078738e-01 -2.50150383e-01 4.26958352e-01 7.91748822e-01
-5.72280943e-01 3.76641542e-01 3.59326839e-01 -2.39883512e-01
-4.09745991e-01 -8.87326360e-01 -1.62380427e-01 1.72827804e+00
1.87602248e-02 -5.55825472e-01 -5.36563694e-01 -6.69394791e-01
-2.98466176e-01 6.73179150e-01 -4.74749535e-01 -1.83024988e-01
-4.91047651e-01 -5.32063603e-01 6.38535142e-01 6.70948923e-01
4.31128889e-01 -1.55524790e+00 -5.16103864e-01 9.96588096e-02
-3.33945900e-01 -1.25299144e+00 -3.12839866e-01 6.32462740e-01
-5.36824405e-01 -7.46401787e-01 -1.21941097e-01 -8.69888544e-01
5.63586727e-02 8.02105814e-02 8.75525057e-01 1.66915074e-01
1.84964702e-01 2.90559620e-01 -6.32879794e-01 -4.99074906e-01
-7.11866617e-01 1.47844866e-01 6.37243241e-02 -1.65552512e-01
7.76027322e-01 -3.71441066e-01 -1.25199124e-01 6.18130505e-01
-4.94662195e-01 -3.67049396e-01 2.76189357e-01 1.16889262e+00
-2.56776631e-01 -3.07514012e-01 1.05655873e+00 -6.94468915e-01
1.36481428e+00 -4.34506446e-01 -8.57211351e-02 -1.38224795e-01
-2.86176562e-01 5.01379296e-02 5.74314058e-01 -3.45477462e-01
-1.25320065e+00 2.36112606e-02 -7.92509317e-01 1.61122471e-01
-6.48097217e-01 6.47054911e-01 -5.33406250e-02 3.17514539e-01
6.31060362e-01 1.83454126e-01 2.39040434e-01 -5.43792129e-01
5.58642074e-02 1.04014301e+00 -7.97749609e-02 -4.81257349e-01
1.13893889e-01 1.13256492e-01 -5.70576489e-01 -1.32909024e+00
-5.22551656e-01 -7.11226642e-01 -8.59995544e-01 -1.64773136e-01
1.07669318e+00 -6.23093009e-01 -8.90419185e-01 5.14490247e-01
-1.11574459e+00 -4.33988899e-01 -1.25067398e-01 6.22289479e-01
-3.44918877e-01 3.82513613e-01 -9.25679386e-01 -1.23934722e+00
-1.68564484e-01 -1.50850916e+00 7.94030309e-01 2.50070721e-01
-6.75002217e-01 -1.14532697e+00 -7.48261511e-02 4.98058438e-01
6.46678150e-01 -5.93000770e-01 8.52298379e-01 -1.41151786e+00
-3.63737419e-02 -3.47959846e-02 2.95952022e-01 3.94977033e-01
4.78245974e-01 -1.61740810e-01 -1.42505336e+00 1.52101651e-01
2.97822237e-01 -8.54270041e-01 7.55311131e-01 4.72566307e-01
6.99295223e-01 -3.42758328e-01 6.32401258e-02 -1.42259836e-01
3.97219032e-01 6.75717831e-01 1.94804728e-01 -1.83462411e-01
1.86756879e-01 1.12645710e+00 6.53762519e-01 5.65023422e-01
4.75745767e-01 9.33269143e-01 2.72025555e-01 -8.30217525e-02
1.59269974e-01 1.80779457e-01 6.60175681e-01 1.16961956e+00
2.55921453e-01 -3.96281958e-01 -9.96328592e-01 3.83028895e-01
-1.89216304e+00 -7.27939963e-01 9.96536389e-03 1.77157712e+00
1.05986476e+00 3.05311024e-01 3.25401574e-01 1.99092880e-01
3.51895124e-01 5.03567874e-01 -2.98736602e-01 -5.87467134e-01
-2.34156609e-01 1.50716737e-01 -2.45492578e-01 9.29811954e-01
-1.31518555e+00 1.20488322e+00 7.27765369e+00 5.37903905e-01
-8.26884568e-01 9.74789821e-03 4.57056373e-01 4.25133039e-05
-1.31681487e-01 -2.54725367e-01 -8.72667730e-01 3.08531880e-01
1.09933150e+00 9.57239196e-02 1.28344968e-01 7.98513174e-01
3.09501886e-01 -3.81844252e-01 -1.15847063e+00 6.66815758e-01
5.69697469e-03 -9.85204935e-01 -3.10022444e-01 3.47597105e-03
-2.75706351e-02 -4.30198014e-02 2.81034615e-02 5.43314517e-01
5.96692562e-01 -1.16497970e+00 2.85177112e-01 1.26093045e-01
-1.10328056e-01 -5.24029851e-01 1.12725139e+00 5.58467627e-01
-1.01661146e+00 -4.63439152e-02 -2.75005102e-01 -4.12918985e-01
3.99138838e-01 -3.14976797e-02 -1.42800152e+00 1.00742236e-01
5.72254479e-01 4.67428505e-01 -4.55252916e-01 3.62501711e-01
-3.25360268e-01 1.06188786e+00 -5.55201948e-01 -4.49580878e-01
3.11723799e-01 -1.70535907e-01 4.56168503e-01 1.30479121e+00
-3.22285920e-01 2.54762411e-01 6.41023993e-01 5.27399421e-01
3.18938464e-01 3.91004026e-01 -7.73236573e-01 4.62915674e-02
6.74419761e-01 9.51889694e-01 -6.39363170e-01 2.18148883e-02
-4.96599376e-01 5.55896163e-01 1.26536682e-01 1.38409093e-01
-3.36044699e-01 1.32909358e-01 7.20932007e-01 -3.86405259e-01
-4.60985675e-02 -4.89947706e-01 -3.02728295e-01 -7.67409205e-01
-1.95259526e-01 -7.97907948e-01 4.28191096e-01 -4.25768524e-01
-1.34861660e+00 7.54944146e-01 8.73158425e-02 -6.82858646e-01
-8.84555578e-01 -7.26228476e-01 -8.20552051e-01 7.84058809e-01
-1.05113816e+00 -9.53544974e-01 -1.44879654e-01 4.50219244e-01
1.08960903e+00 -5.42320788e-01 1.17550492e+00 1.63336173e-01
-3.68267834e-01 4.11894411e-01 -3.34540784e-01 3.93461555e-01
8.41150761e-01 -1.38261056e+00 1.55147702e-01 3.62552673e-01
2.93960541e-01 7.46346354e-01 7.91107953e-01 -3.22963268e-01
-1.06484127e+00 -2.90453434e-01 9.40386772e-01 -5.06117404e-01
5.07625103e-01 -7.29939044e-01 -1.08092165e+00 3.76961023e-01
6.72560096e-01 -5.81087828e-01 1.13778460e+00 7.49944329e-01
-1.85535438e-02 3.13528806e-01 -9.37161446e-01 5.24162471e-01
5.44258595e-01 -7.57835388e-01 -9.15211916e-01 1.33327335e-01
6.97434902e-01 -1.33066714e-01 -8.43947589e-01 1.27647728e-01
5.41334927e-01 -1.01091528e+00 9.00687635e-01 -6.00346446e-01
1.38952300e-01 2.31174320e-01 -3.57716948e-01 -1.50005627e+00
2.61754483e-01 -3.55947882e-01 1.00304157e-01 1.25180447e+00
6.28979385e-01 -5.49708664e-01 6.07242405e-01 7.12432623e-01
-2.72321135e-01 -5.14527082e-01 -1.11010444e+00 -4.79688197e-01
1.80388987e-01 -4.83964294e-01 4.27974671e-01 1.02467120e+00
5.36460221e-01 8.89444530e-01 -6.45398378e-01 -3.34247887e-01
-3.70428264e-01 -3.69947478e-02 6.78897917e-01 -1.30627024e+00
-1.23093203e-01 -6.26965523e-01 -2.59372294e-01 -1.26456320e+00
6.28899693e-01 -5.33369243e-01 3.13524395e-01 -1.40497005e+00
-1.27539352e-01 -2.49227658e-01 -1.40030280e-01 5.72119772e-01
-4.33934420e-01 -1.53212026e-01 3.38540405e-01 1.64152935e-01
-2.80656993e-01 9.40022767e-01 7.57943690e-01 -2.47972339e-01
-5.19183457e-01 3.73638779e-01 -2.83987015e-01 8.54817867e-01
9.56474602e-01 -4.40731347e-02 -4.69253957e-01 3.68393920e-02
-3.06782365e-01 3.02669197e-01 -1.88026845e-01 -7.09085047e-01
3.56700659e-01 -2.88234800e-01 -5.35120219e-02 -6.41003788e-01
1.16243815e+00 -8.16444218e-01 -5.76939225e-01 3.24195325e-01
-8.91266406e-01 2.68536224e-03 2.51744270e-01 3.45147640e-01
-6.22904599e-01 -4.23240960e-01 6.47834420e-01 -2.35537682e-02
-7.44521320e-01 -2.87028462e-01 -1.02980709e+00 1.65694836e-03
5.94774604e-01 -2.58706789e-02 -1.29677832e-01 -1.15051460e+00
-1.00380254e+00 3.06293249e-01 -1.60949975e-01 8.14134181e-01
6.01034641e-01 -1.13832271e+00 -5.71751237e-01 3.14105690e-01
-1.86261293e-02 -5.36555886e-01 2.19518598e-02 6.35448694e-01
2.17275899e-02 5.55680454e-01 -1.93797335e-01 -6.92046583e-01
-1.64752674e+00 -3.10840085e-02 2.75939077e-01 -1.75671637e-01
-2.58545250e-01 9.26863253e-01 2.75789350e-01 -6.94713891e-01
6.85674906e-01 -4.49508995e-01 -7.87002504e-01 5.46159685e-01
3.82439882e-01 1.32058412e-01 8.18599537e-02 -8.65154207e-01
-2.88903743e-01 1.72650501e-01 -1.12160087e-01 -5.23239434e-01
1.12287045e+00 -2.64577270e-01 2.90901303e-01 1.09537601e+00
7.60118663e-01 4.10407111e-02 -8.03734362e-01 -5.52769065e-01
2.29743123e-01 -2.06478000e-01 2.43581086e-02 -7.95252979e-01
-7.93736398e-01 1.28800011e+00 5.21011114e-01 8.23981941e-01
8.66698444e-01 3.48034874e-02 6.28194034e-01 6.96656048e-01
1.17536321e-01 -1.41155767e+00 4.34728712e-01 1.33390236e+00
9.75547731e-01 -1.71517777e+00 -3.31951439e-01 -4.71576691e-01
-1.37636757e+00 1.20576608e+00 8.98284316e-01 2.98840046e-01
1.02891588e+00 4.51146632e-01 5.92555821e-01 -3.32402915e-01
-9.55358088e-01 -4.00724232e-01 2.60038376e-01 4.70524967e-01
7.89294899e-01 -5.03437892e-02 -3.51101816e-01 8.52376163e-01
-6.48467660e-01 -7.57603526e-01 4.57552344e-01 9.19334888e-01
-6.62979662e-01 -1.24640083e+00 -1.66651681e-01 4.31428432e-01
-4.10366982e-01 -1.85760915e-01 -1.20740497e+00 6.63654268e-01
-5.11480987e-01 1.65627098e+00 3.07832006e-02 -6.29257202e-01
1.38812691e-01 7.55228758e-01 -1.49851307e-01 -9.58340824e-01
-1.23141289e+00 3.01874012e-01 5.87710798e-01 -3.83939624e-01
-5.92330277e-01 -4.67847556e-01 -1.27924395e+00 -7.30958059e-02
-5.71340024e-01 4.17140961e-01 6.10174954e-01 1.11276495e+00
7.40618035e-02 3.64302665e-01 7.47252941e-01 -5.24870276e-01
-6.50150061e-01 -1.50366318e+00 -4.82813120e-01 1.95155725e-01
3.09105784e-01 -6.92911983e-01 -5.29434144e-01 -1.32847115e-01] | [12.809038162231445, 7.722132682800293] |
7f4a6ea0-9a55-4251-9f6b-6a01ce2ebf1c | prometheus-a-corpus-of-proverbs-annotated | null | null | https://aclanthology.org/L16-1600 | https://aclanthology.org/L16-1600.pdf | PROMETHEUS: A Corpus of Proverbs Annotated with Metaphors | Proverbs are commonly metaphoric in nature and the mapping across domains is commonly established in proverbs. The abundance of proverbs in terms of metaphors makes them an extremely valuable linguistic resource since they can be utilized as a gold standard for various metaphor related linguistic tasks such as metaphor identification or interpretation. Besides, a collection of proverbs fromvarious languages annotated with metaphors would also be essential for social scientists to explore the cultural differences betweenthose languages. In this paper, we introduce PROMETHEUS, a dataset consisting of English proverbs and their equivalents in Italian.In addition to the word-level metaphor annotations for each proverb, PROMETHEUS contains other types of information such as the metaphoricity degree of the overall proverb, its meaning, the century that it was first recorded in and a pair of subjective questions responded by the annotators. To the best of our knowledge, this is the first multi-lingual and open-domain corpus of proverbs annotated with word-level metaphors. | ['Serra Sinem Tekiro{\\u{g}}lu', 'G{\\"o}zde {\\"O}zbal', 'Carlo Strapparava'] | 2016-05-01 | prometheus-a-corpus-of-proverbs-annotated-1 | https://aclanthology.org/L16-1600 | https://aclanthology.org/L16-1600.pdf | lrec-2016-5 | ['english-proverbs'] | ['natural-language-processing'] | [-5.67576766e-01 -2.32681647e-01 -5.34491003e-01 -1.03157453e-01
-3.69251668e-01 -1.29686522e+00 9.98759985e-01 8.26337337e-01
-3.16845924e-01 7.99892902e-01 4.86348182e-01 -5.40216446e-01
-2.62739122e-01 -8.36051941e-01 -2.85415262e-01 -5.39944649e-01
2.89611399e-01 6.49141371e-01 -1.07444502e-01 -8.63276005e-01
4.60402906e-01 2.06939906e-01 -9.80406642e-01 4.14912760e-01
8.94582987e-01 8.05500269e-01 6.92486346e-01 -2.29350179e-01
-6.47936940e-01 7.11297154e-01 -9.14546907e-01 -5.91968894e-01
-3.71822596e-01 -3.54181916e-01 -1.05361271e+00 -5.10489225e-01
-4.39757183e-02 2.28083625e-01 -2.70907357e-02 9.72806752e-01
1.59690335e-01 -7.81955943e-02 3.90703022e-01 -6.40861809e-01
-8.82992744e-01 9.51338053e-01 -3.26260448e-01 3.04788798e-01
5.69410145e-01 -5.69188237e-01 1.17604995e+00 -7.34609127e-01
1.06319666e+00 1.50870574e+00 3.50437731e-01 3.76465678e-01
-1.32777703e+00 -3.29217613e-01 -3.98647428e-01 4.44897234e-01
-1.02373397e+00 -3.67284156e-02 1.22301197e+00 -6.41585529e-01
6.91974461e-01 4.25510705e-01 7.93197393e-01 1.51930773e+00
-5.31577691e-02 2.40597636e-01 1.57621920e+00 -7.60184646e-01
1.38010442e-01 6.09540701e-01 5.13616623e-03 1.74121514e-01
-4.03271802e-02 -2.66704947e-01 -4.39705342e-01 -3.76360953e-01
6.13592386e-01 -1.82687238e-01 -2.16896772e-01 8.09559599e-02
-1.35140121e+00 1.01701248e+00 4.57396120e-01 1.13780141e+00
-2.83821613e-01 1.07213713e-01 9.01059091e-01 3.29786628e-01
3.82304221e-01 8.53751242e-01 -3.36876601e-01 -4.40153688e-01
-6.10650659e-01 4.50288773e-01 9.74239290e-01 6.62559211e-01
4.10008937e-01 -2.39088431e-01 3.51967305e-01 1.34362900e+00
-2.00540684e-02 4.92135167e-01 7.25739896e-01 -9.62225318e-01
2.03315586e-01 7.37519383e-01 3.43410343e-01 -1.39675903e+00
2.54208520e-02 7.76016563e-02 -3.54433239e-01 -3.61380935e-01
3.62641394e-01 3.66950303e-01 -1.02652721e-01 1.87193203e+00
2.03471318e-01 -6.97846889e-01 -1.18941598e-01 1.13846481e+00
1.16905212e+00 6.05099022e-01 4.83952053e-02 -2.63396174e-01
2.06181717e+00 -2.68745065e-01 -9.08034980e-01 -3.11362267e-01
2.50051022e-01 -1.22927761e+00 1.65481329e+00 2.15163141e-01
-7.51267552e-01 -8.46712440e-02 -8.03246737e-01 -4.71045166e-01
-9.57578480e-01 -1.66194350e-01 6.34397149e-01 2.92169988e-01
-5.63669980e-01 2.82162845e-01 -5.30946970e-01 -8.67471457e-01
2.45051528e-03 -6.71149850e-01 -1.76693574e-01 2.57807642e-01
-1.44500220e+00 1.47890401e+00 8.44815552e-01 -3.21883172e-01
-4.08488184e-01 -4.54912156e-01 -7.02310622e-01 -2.66879201e-01
6.41962230e-01 -2.18731791e-01 1.04692590e+00 -1.02463639e+00
-7.94985771e-01 1.49541533e+00 -7.33470619e-02 -4.78730023e-01
1.38831601e-01 -6.81455582e-02 -6.05434418e-01 7.44688809e-02
6.05995119e-01 5.24438202e-01 3.71503919e-01 -1.00270998e+00
-2.46235609e-01 -1.21996939e-01 6.95364177e-01 -3.74594964e-02
-4.69562113e-01 5.19439280e-01 2.68125683e-02 -1.17542136e+00
1.63783133e-01 -5.37587643e-01 4.03812855e-01 -8.49920064e-02
-2.17081979e-01 -6.24251783e-01 8.17412436e-01 -9.14309323e-01
1.41240466e+00 -2.18026400e+00 3.54292333e-01 -1.91961169e-01
3.23624551e-01 -4.40807253e-01 4.01270747e-01 1.03598750e+00
1.12132184e-01 2.05343410e-01 2.79801413e-02 4.24748749e-01
3.80521059e-01 7.70946801e-01 -8.44843030e-01 2.01447025e-01
-2.10519999e-01 9.97102618e-01 -1.38122988e+00 -8.06522012e-01
2.13922143e-01 3.23469639e-01 -1.27118662e-01 -1.48034170e-01
-3.97392750e-01 4.21706855e-01 -3.87084991e-01 7.44792402e-01
9.41094607e-02 3.94639745e-02 4.54271138e-01 -1.19280554e-01
-4.76224273e-01 9.64404821e-01 -1.14009574e-01 1.84188497e+00
-7.62891293e-01 1.00478780e+00 -8.81631821e-02 -1.02449489e+00
8.76337528e-01 6.20151043e-01 2.51217723e-01 -6.02572620e-01
2.16560751e-01 7.00464129e-01 1.33984581e-01 -3.96742046e-01
6.06593728e-01 -6.82662606e-01 -7.99663246e-01 2.02850834e-01
-1.59684464e-01 -4.95192796e-01 6.72317803e-01 -5.47129214e-02
6.33193672e-01 1.63576812e-01 8.67403686e-01 -1.00615788e+00
3.96964937e-01 3.20611835e-01 5.22499263e-01 1.43976569e-01
2.42737710e-01 5.04806228e-02 5.02968907e-01 -7.44007111e-01
-1.01654875e+00 -1.22174561e+00 -7.96658814e-01 8.04922938e-01
1.43316731e-01 -8.06794941e-01 -5.61052799e-01 8.74013081e-03
-1.29296303e-01 1.11425591e+00 -6.29088879e-01 3.28980535e-01
-4.42372113e-01 -2.23496199e-01 6.53606713e-01 2.18418732e-01
5.19620121e-01 -1.21513057e+00 -1.19553030e+00 4.25033867e-01
-8.88818264e-01 -1.17676282e+00 1.25189275e-01 -7.96503387e-03
-5.10934174e-01 -1.10285962e+00 -2.39440754e-01 -8.49949658e-01
1.07380718e-01 -7.10503384e-02 1.45928943e+00 7.28383884e-02
-1.51109219e-01 -2.86594152e-01 -5.91104746e-01 -6.55414402e-01
-6.32848084e-01 -2.97900081e-01 -8.77074227e-02 -8.38833511e-01
5.65867782e-01 -8.09700906e-01 -9.18955132e-02 2.10168418e-02
-9.09235477e-01 8.14228430e-02 -1.59686074e-01 9.30783093e-01
3.78912359e-01 -1.73013344e-01 1.70438617e-01 -9.03656304e-01
8.81393731e-01 -7.45632112e-01 -5.39989829e-01 2.53688812e-01
-3.21409032e-02 -2.33485430e-01 6.64667308e-01 -4.71346766e-01
-7.53129900e-01 -1.16169357e+00 4.21599299e-02 5.34822531e-02
1.20371588e-01 1.01732862e+00 1.36549518e-01 3.46057445e-01
9.30827439e-01 -1.33256242e-01 -5.19010425e-01 -7.92550027e-01
4.84047681e-01 7.14170694e-01 4.95645791e-01 -1.12425625e+00
4.77831215e-01 3.31638873e-01 -1.62778258e-01 -1.04761231e+00
-9.63967562e-01 -3.27366620e-01 -3.37133735e-01 -1.49326539e-02
8.22595835e-01 -5.35777628e-01 -7.57058620e-01 -2.39136204e-01
-1.63411248e+00 1.27509564e-01 -1.29578292e-01 3.90157327e-02
-5.71688056e-01 1.69778585e-01 -4.97429013e-01 -7.57415175e-01
-2.09753528e-01 -6.76227510e-01 6.73044264e-01 -1.27334461e-01
-1.14834976e+00 -1.50255835e+00 3.09691280e-02 1.55835599e-01
2.32129157e-01 9.19912815e-01 1.54468548e+00 -3.90627027e-01
4.64319997e-02 1.71190947e-01 -8.79591778e-02 -1.31104067e-01
5.93150795e-01 -4.92946029e-01 -6.38355017e-01 5.55736609e-02
5.16950488e-01 -6.00330412e-01 3.64364862e-01 -1.25757024e-01
7.71927357e-01 -6.98051691e-01 -1.75670356e-01 1.58106595e-01
1.73839259e+00 2.30515137e-01 2.42083907e-01 7.26213515e-01
1.50599197e-01 6.79716051e-01 6.82584643e-01 3.08137923e-01
3.66558462e-01 1.00232577e+00 -6.48470670e-02 2.42630661e-01
1.36107534e-01 -2.10594237e-01 3.41816574e-01 9.89663959e-01
-1.38095450e-02 8.46982747e-02 -1.43170810e+00 9.72809136e-01
-1.75327432e+00 -9.54485297e-01 -9.91636654e-04 1.85785544e+00
1.23338115e+00 3.55996042e-02 1.29001841e-01 3.48571055e-02
3.09482306e-01 3.69365424e-01 1.37854859e-01 -8.20000350e-01
-4.21036363e-01 3.90088469e-01 -1.26508668e-01 5.27163506e-01
-8.86390507e-01 1.18617070e+00 5.95184326e+00 9.80212510e-01
-9.30486619e-01 4.93349165e-01 1.21965818e-01 1.03860244e-01
-5.40233612e-01 4.29939002e-01 4.94278632e-02 6.51879847e-01
5.68117678e-01 -3.82678002e-01 9.26772475e-01 6.29303336e-01
2.80535340e-01 -2.92610407e-01 -1.09404838e+00 8.96786213e-01
4.07936163e-02 -1.42258763e+00 -1.62882075e-01 -2.26183623e-01
3.19237888e-01 -1.02934591e-01 -3.42082679e-01 -2.33743981e-01
-5.58023602e-02 -9.20735836e-01 1.09941256e+00 -6.53904211e-03
8.73951256e-01 -5.20634472e-01 5.98555446e-01 3.65254730e-01
-9.54270482e-01 1.24835350e-01 -4.13307071e-01 -2.95479953e-01
2.53753334e-01 2.37922445e-01 -6.82049155e-01 4.88020986e-01
8.99040222e-01 6.73018754e-01 -3.20506275e-01 2.07391009e-01
-7.16019869e-01 6.31253123e-01 -4.26866561e-01 -4.87585157e-01
2.60240585e-01 -5.67780018e-01 1.00553775e+00 1.26899600e+00
2.47147650e-01 -7.51599222e-02 3.28248262e-01 1.45359838e+00
2.08379015e-01 3.36507022e-01 -7.34733760e-01 -8.90061021e-01
8.79967391e-01 1.28710604e+00 -1.01860332e+00 -1.19887508e-01
-2.18605757e-01 8.73713613e-01 3.17485690e-01 1.92340255e-01
-9.39491093e-01 -3.45516056e-01 4.97558832e-01 3.63304466e-01
-3.01538825e-01 -2.30596036e-01 -4.45061564e-01 -9.97573376e-01
6.41324446e-02 -6.64131105e-01 3.03767383e-01 -8.90216589e-01
-1.62517488e+00 4.96839613e-01 4.25629199e-01 -9.35174704e-01
-2.48635232e-01 -7.12826073e-01 -3.20003778e-01 9.51869488e-01
-1.10565937e+00 -1.29356563e+00 7.04296529e-02 1.87344924e-01
4.59302127e-01 2.63787364e-03 1.44302201e+00 1.94474999e-02
2.29646012e-01 3.56720611e-02 1.62921362e-02 2.78415799e-01
7.57597387e-01 -1.34042811e+00 5.52987331e-04 3.05241615e-01
4.08776641e-01 1.14889240e+00 1.33818328e+00 -4.42235917e-01
-9.48172152e-01 -2.36116782e-01 1.36075544e+00 -6.07040286e-01
1.44007397e+00 -5.13181567e-01 -8.40112865e-01 5.62965810e-01
7.13967204e-01 -7.05700755e-01 7.95739412e-01 5.21892846e-01
-5.34710586e-01 7.93026313e-02 -1.06554878e+00 8.85905564e-01
8.80543292e-01 -8.81517470e-01 -1.51882994e+00 6.97784483e-01
5.51432431e-01 -5.96789181e-01 -8.65238249e-01 -1.55175090e-01
3.88328016e-01 -6.38421476e-01 9.33206916e-01 -4.64529663e-01
8.40371549e-01 -1.05654977e-01 -3.67232800e-01 -1.19073451e+00
-2.38235798e-02 -4.77993608e-01 2.02961877e-01 1.36602330e+00
1.64296284e-01 -6.80013895e-01 -4.75626849e-02 2.22608112e-02
4.90349829e-02 -5.37256420e-01 -1.42786872e+00 -6.48904145e-01
3.98296922e-01 -3.08236897e-01 4.70636278e-01 1.44667470e+00
7.43570805e-01 4.91777956e-01 5.52693643e-02 -8.28549623e-01
1.22490399e-01 8.05733621e-01 2.44866982e-01 -1.26155806e+00
-2.90670156e-01 -7.54122555e-01 -4.39387292e-01 -4.77096170e-01
4.55840319e-01 -1.19486380e+00 -2.09980741e-01 -1.27624416e+00
1.92982584e-01 -3.38365585e-01 -1.74762309e-01 4.76841271e-01
2.47951850e-01 3.06675881e-01 2.91432738e-01 2.54270971e-01
-3.13780457e-02 2.49270514e-01 9.87610698e-01 -5.98906055e-02
-4.23393548e-02 -7.53533304e-01 -8.32737505e-01 1.12949300e+00
8.43444347e-01 -5.16970098e-01 -3.04331094e-01 -4.27197635e-01
5.80385089e-01 -4.92567122e-02 6.48505270e-01 -1.11727819e-01
-3.34291548e-01 -7.27480054e-01 -2.57302895e-02 -2.42318317e-01
4.81573284e-01 -8.14155281e-01 3.54913473e-01 3.90505880e-01
6.90708831e-02 3.42540324e-01 4.22744066e-01 2.41029598e-02
-4.40436721e-01 -3.40846330e-01 5.76128185e-01 -4.22386914e-01
-6.43315792e-01 -5.71716130e-01 -4.26854789e-01 5.30529380e-01
8.04177701e-01 -1.68464780e-01 -2.61823177e-01 -1.52495503e-01
-2.12140247e-01 -9.52205434e-02 9.90677655e-01 6.95267081e-01
4.57663000e-01 -1.71920180e+00 -5.42431891e-01 -5.66673696e-01
3.21904063e-01 -3.78235847e-01 -1.46599889e-01 8.77679706e-01
-8.43822777e-01 3.65765959e-01 -4.07062531e-01 -2.12692201e-01
-1.04102743e+00 7.19839036e-01 6.04603700e-02 3.79366800e-03
-6.47565067e-01 4.91442859e-01 3.02481323e-01 -3.48914117e-01
-3.79764706e-01 -1.90436348e-01 -1.56043634e-01 4.48257118e-01
1.57287717e-01 4.75475490e-02 -3.99370968e-01 -9.29040790e-01
-7.71558881e-01 2.12204799e-01 4.04430151e-01 -5.18401027e-01
1.24597096e+00 -5.06517850e-02 -1.11415279e+00 1.36871290e+00
7.78893888e-01 3.91281009e-01 -1.82341598e-02 5.65074831e-02
4.93357569e-01 -6.63202643e-01 -3.48625004e-01 -1.15138423e+00
-4.23365310e-02 7.78067172e-01 -2.33108327e-01 4.42170411e-01
1.04657590e+00 4.50736910e-01 7.90923238e-01 2.71576345e-01
6.58538640e-01 -1.21676803e+00 -1.60812438e-01 6.29698038e-01
1.43056774e+00 -8.08707595e-01 -1.93845749e-01 -3.63306791e-01
-3.89724255e-01 1.00241566e+00 1.91291735e-01 -1.23763479e-01
2.87500948e-01 5.56353107e-02 2.95801222e-01 -4.42735702e-01
-3.51416677e-01 2.70716786e-01 4.35136557e-01 2.32756704e-01
1.07076681e+00 4.79298770e-01 -1.53977513e+00 8.65786552e-01
-1.09015906e+00 -3.79021257e-01 2.34534949e-01 7.78711975e-01
-2.32113987e-01 -1.34822369e+00 -4.16786969e-01 -3.14732678e-02
-4.79128957e-01 -3.47310245e-01 -9.11557496e-01 1.02612615e+00
3.41407746e-01 8.53590369e-01 -8.07130933e-02 9.94056463e-02
-7.66902417e-02 2.84264442e-02 7.45959640e-01 -5.71036994e-01
-6.08859360e-01 1.17160365e-01 4.45574880e-01 -1.49751231e-01
-6.36101604e-01 -4.91015494e-01 -1.38832021e+00 -6.91062868e-01
-1.19858600e-01 5.52545428e-01 4.22544062e-01 1.21857655e+00
-3.28197509e-01 1.67844325e-01 1.72747634e-02 -2.40859777e-01
-1.11389980e-01 -1.12445962e+00 -4.97943759e-01 4.89671081e-01
-1.14998072e-01 -9.39696729e-01 1.30272552e-01 -1.28354520e-01] | [10.635248184204102, 9.206259727478027] |
ca35ef69-8358-4c54-8298-560f0bcf96c6 | crop-zero-shot-cross-lingual-named-entity | 2210.07022 | null | https://arxiv.org/abs/2210.07022v1 | https://arxiv.org/pdf/2210.07022v1.pdf | CROP: Zero-shot Cross-lingual Named Entity Recognition with Multilingual Labeled Sequence Translation | Named entity recognition (NER) suffers from the scarcity of annotated training data, especially for low-resource languages without labeled data. Cross-lingual NER has been proposed to alleviate this issue by transferring knowledge from high-resource languages to low-resource languages via aligned cross-lingual representations or machine translation results. However, the performance of cross-lingual NER methods is severely affected by the unsatisfactory quality of translation or label projection. To address these problems, we propose a Cross-lingual Entity Projection framework (CROP) to enable zero-shot cross-lingual NER with the help of a multilingual labeled sequence translation model. Specifically, the target sequence is first translated into the source language and then tagged by a source NER model. We further adopt a labeled sequence translation model to project the tagged sequence back to the target language and label the target raw sentence. Ultimately, the whole pipeline is integrated into an end-to-end model by the way of self-training. Experimental results on two benchmarks demonstrate that our method substantially outperforms the previous strong baseline by a large margin of +3~7 F1 scores and achieves state-of-the-art performance. | ['Furu Wei', 'Zhoujun Li', 'Hongcheng Guo', 'Dongdong Zhang', 'Li Dong', 'Yuwei Yin', 'Shuming Ma', 'Shaohan Huang', 'Jian Yang'] | 2022-10-13 | null | null | null | null | ['cross-lingual-ner'] | ['natural-language-processing'] | [ 0.22611648 -0.20129961 -0.35886902 -0.52571195 -1.4965664 -0.857012
0.571273 -0.37569842 -0.7370184 1.0856855 0.36788392 -0.43459448
0.6580976 -0.41863054 -0.8395966 -0.38070044 0.5888705 0.40014434
-0.14797713 -0.04680393 -0.2290354 0.06962986 -0.6500076 0.395392
1.1504585 0.34844628 0.34595937 0.09214497 -0.6170468 0.36643407
-0.16611569 -0.94648033 0.34984148 -0.6812152 -0.8652197 -0.1524407
0.15519534 0.22203167 0.12874424 1.2106494 0.5702744 -0.05920695
0.27322397 -0.7710561 -0.90545577 0.67561793 -0.5164077 -0.23484556
0.13793673 0.12450844 0.9318477 -1.5221723 0.9471537 1.1066115
0.7933049 0.71553946 -1.203713 -0.6427837 -0.01054605 -0.22827707
-1.3467723 -0.6131522 0.41152832 -0.29303446 1.0594853 -0.196514
-0.11652644 1.2185639 0.06115042 0.5941951 1.2385709 -0.5619228
-0.01505709 0.3511253 -0.14652355 0.7753826 -0.06864275 -0.05388562
-0.48742267 -0.05620835 0.21926385 -0.11621779 -0.21648711 -0.17845361
-1.6020141 0.61408705 0.34395906 0.47423342 -0.5140934 -0.43702117
0.59085757 0.23881994 0.67705667 0.44382623 -0.92648727 -0.00954883
-0.94505054 -0.53393155 0.86996526 1.0784314 0.98977053 0.07106069
-0.05508181 1.0249485 0.08835345 0.639151 0.592276 -0.3614461
0.94505495 0.53140134 0.13471635 -0.29980248 -0.16005674 -0.43908393
-0.7230678 -0.24626084 0.1275385 -0.48751923 -0.81369734 1.8810254
0.33271047 0.13553198 0.67307526 0.8118415 0.5236574 0.9517106
0.4443572 -0.34549972 1.3602437 -1.4730346 -0.5976119 -0.3850745
1.0515822 -0.98300457 1.2009777 -0.23394775 -0.74091595 -0.53981346
-0.7578816 -0.19350915 -0.40855667 0.7196695 0.03406134 0.51929015
-0.8574844 0.37515104 -0.93921936 -0.61826116 0.07996691 0.12079492
-0.7437007 -0.39072567 -1.2487947 1.1904703 0.6747212 0.2010729
-0.6032669 -0.8005832 -0.9894113 -0.13338244 0.19130576 -0.5476003
0.974361 -0.946441 -1.6033111 0.93911594 -0.28802136 -0.1888676
0.48394316 -0.5355392 -0.6477551 -0.31989986 0.55789477 0.59075797
0.13582954 -0.9336339 -0.6514016 -0.21851763 -0.44856617 0.2940536
-0.37048274 0.4825691 -0.65346706 -0.45446047 -0.25014722 -1.1371405
-0.38446295 -0.50647944 -0.40152323 -0.13791223 0.5004316 -0.97604877
1.0082269 -2.1221216 0.11401432 -0.40433678 -0.51299405 0.545325
-0.47035208 0.66932636 -0.1442139 0.17009957 -0.4773085 -0.564833
-0.0661631 0.00905691 -0.34571138 0.39761743 0.6234346 0.96040964
-1.1025158 -0.5446703 -0.02985821 0.6866601 -0.11161374 0.32778266
-0.01813883 0.77742064 -0.3507137 0.49443054 0.520444 -0.05869485
0.5171355 -0.14133307 -0.5051027 0.80612016 -0.7581587 2.2375927
-0.8201501 0.13890806 -0.23928532 -0.6226195 1.180035 0.6658286
0.16193894 -0.5035799 -0.14591537 0.66171676 -0.2510207 -0.35471165
0.4629806 -0.48076385 -0.43621555 0.5790149 0.35041222 0.36407012
0.16699834 -0.1716561 0.93529755 0.7525454 0.4779063 -0.07165971
0.6295484 0.31535307 1.0806984 0.07097304 -0.18150006 0.46938258
-0.05650897 -0.31780106 -1.1957084 -0.9243161 0.04652191 1.3099065
-0.23343174 -0.28194845 -0.8544077 -1.1876568 -0.52786434 0.92148876
-0.20833889 0.03840663 -0.8201778 -0.7981229 0.97772473 0.4606137
0.47477552 -1.3382517 0.1815912 0.44581255 -0.64675784 -1.5276955
-0.7670096 0.1933091 -0.71994925 -0.47519302 -0.90879387 -1.1461161
0.7030258 0.01480433 0.90077823 -0.5231169 0.34026724 -0.29197577
-0.36509445 0.05947041 -0.55476505 0.5117133 0.3686261 0.16539644
0.64778155 -0.27790496 -0.23494668 0.3278612 -0.5253955 0.11402043
0.95171523 0.9174435 0.8660555 -0.40497687 0.7535524 -1.0872816
0.25928107 -0.5257994 -0.47912034 0.67113614 -0.44298702 0.34537786
1.0652817 -0.38246575 -1.4679278 0.48262307 -0.17649285 -0.2557165
-0.15980016 0.657515 -0.60231197 0.4078819 0.636569 0.47995538
-0.62143475 -0.74696034 0.6233465 0.92266595 0.7602807 -0.48627776
0.7182243 0.03802951 -0.53966856 -0.37799305 -1.1358542 -0.40585256
-1.0848963 0.24256338 0.99243605 -1.2686887 0.19647123 0.07608121
-1.3200018 -0.04639861 0.02724786 0.82899565 -0.28563866 0.21812429
-0.8824685 -0.5160656 -0.7312295 -1.1628672 1.0303468 0.11186209
-0.00720768 -0.8976594 0.48803848 0.3924145 0.25034377 -0.1466613
0.8580488 -1.0576658 -0.24278961 -0.20192349 -0.25573927 0.43426615
0.22100899 -0.45534596 -0.757771 -0.24833523 -0.11064648 -0.41703475
0.5230257 -0.30297592 0.14144605 -0.08057501 -0.31782362 0.634421
1.5581079 -0.2396031 0.4161233 0.3825855 0.9234433 0.67398316
0.65967834 -0.15140569 0.6012924 0.6278989 -0.28364116 -0.19781235
-0.19829056 -0.72039706 0.8219493 1.4876628 0.2281433 -0.14880708
-1.081129 0.70468074 -1.6623272 -0.64381963 -0.06209325 2.2298298
1.1604595 -0.15654548 -0.2337939 -0.6349496 1.1159499 -0.18700415
-0.5188532 -0.26595852 -0.10764321 0.03447792 0.581765 0.24813816
-1.0487466 1.6795145 5.1358347 0.82269454 -1.273377 0.5756857
0.3408567 0.3219611 -0.14127032 0.31741527 -1.2972997 0.47374165
1.3889462 -0.20905648 0.19196032 0.9102555 0.10413634 0.6210877
-0.97170126 0.6448209 0.08970552 -0.93390954 -0.07946147 -0.05849058
0.946919 0.62190163 -0.38539228 0.6903015 0.67261386 -0.56997633
0.3846619 0.33993667 1.3104308 -0.75238353 0.83433366 0.66497797
-1.2997377 0.52784675 -0.62883705 0.38689423 0.5806299 0.5121311
-0.9761968 0.8469554 0.408735 0.67800677 -0.15750895 0.70078295
-0.6036827 0.775431 -0.12761435 0.15117383 0.42766765 -0.38486597
0.19895256 1.6308498 0.6323111 -0.20386516 0.4845355 0.68152404
-0.6338643 0.6831806 -0.5993002 -0.39776024 0.44509152 1.5248747
-0.44139603 -0.4086707 -0.82745284 1.4940898 0.78424937 0.29005632
-0.6101248 -0.4075685 0.47523463 -0.40770367 0.43827918 -0.15325333
-0.08622619 -1.6506456 0.0518642 -0.819446 0.17548785 -0.58683115
-1.5744 1.1291776 -0.77889144 -1.442425 -0.25813302 -0.34776798
-0.36828867 1.2604755 -1.7407387 -1.5995419 0.39143744 0.39239785
0.6941964 -0.30258483 1.2049558 0.59920657 -0.7871699 0.704606
0.28257558 0.5876488 1.198109 -1.0091505 0.8395104 1.2113281
0.19434689 0.71804667 0.14652409 -0.88602406 -1.2246722 -1.6768966
1.7612575 -0.6085742 0.79335237 -0.48589402 -1.0368341 0.8382901
0.3175255 0.22658768 1.058919 0.13557328 -0.59721446 0.19368434
-0.8625655 0.583872 0.97081053 -0.84241045 -0.82598996 0.3123906
0.9543662 -0.18738523 -0.92638016 0.25572625 0.31338534 -0.3075162
0.51590747 -0.76596963 0.3694291 -0.5445583 -0.26935503 -1.4160627
-0.18323846 -0.46081483 0.58423775 1.7780963 0.98691237 -0.38985255
0.38436088 0.29483375 -0.48524198 -0.4207871 -0.7819577 -0.9113747
0.24421838 -0.31426063 0.46673018 1.338045 0.17875195 0.9171336
-0.7717903 0.39637205 0.455638 0.21751894 0.6475999 -0.764895
-0.07491186 0.07469223 0.00631147 -1.0081309 0.62820464 -1.4203688
0.5864809 -1.3860788 0.39888528 -0.4157113 -0.48281008 0.72971123
-0.42066053 0.3496253 0.10349661 0.53978944 -0.71874106 0.665348
0.83768404 0.28531167 -0.15413995 -0.24780267 -0.5027515 0.6008978
0.65341055 -0.9419898 0.14081185 -0.7344076 0.11313813 0.16317074
-0.4076776 -0.67659485 0.16826507 -0.06504511 0.12159757 -0.40548357
-0.0068509 -0.70048964 0.07404584 0.22667699 -0.43509492 0.18337852
-0.03890103 0.4117334 -0.34548187 -0.17069137 0.69028616 -0.11548153
-0.6329253 0.24648291 0.01278083 0.20858365 0.7328239 0.27989867
-0.2523661 0.2522619 -0.52093685 0.14437906 0.6578523 0.5950734
0.00713315 -1.4876499 -1.0467212 0.12641819 0.39329994 -0.48934424
-0.00722445 0.84559405 -0.12313481 0.468146 -0.16761929 -0.4431967
-0.89304173 0.5584484 0.14132267 -0.62478095 -0.46781036 0.52072346
0.20489444 -1.416282 -0.2324272 0.2162226 0.05412625 -0.25547224
0.48201331 0.08314677 0.1637409 -1.1454077 -0.4895662 0.534695
-0.15718195 -0.21823163 1.3983364 -0.40178898 -0.03767084 0.45968944
1.3977374 0.2588089 -1.0755343 -0.7177855 0.36725518 -0.06328233
-0.27963942 -0.9567888 -0.68526596 0.88563365 0.31523916 -0.6895505
0.8634107 -0.02549105 1.0767231 0.60006183 0.5318864 -0.892943
-0.43132052 0.85379595 0.4538469 -1.4619715 -0.4440522 -0.27526385
-1.0042932 0.8592576 0.4739553 0.08352263 0.2889083 0.17004405
0.5690448 0.33158353 -0.6532075 -0.21859793 0.31217185 0.40037104
0.8283135 0.09588879 -0.5089626 0.7674642 -0.16950373 0.13394627
0.11792849 0.7635162 -0.17057303 -1.58226 -0.05612946 -0.20086704
-0.80363554 -0.64335996 -0.3103108 0.44366825 0.13972741 0.8214975
-0.3470723 -0.34720767 0.4276008 0.6737124 -0.06006912 -0.91688794
-0.7637861 0.2852008 0.2280608 -0.30078644 -0.3955526 -0.65744066
-1.2870321 0.05353381 -0.40618825 0.39068514 0.8755493 1.1023122
0.6758746 0.13487963 0.6849721 -0.5078231 -0.54551995 -1.1267571
-0.10829746 0.30533367 -0.2247116 -0.10601399 -0.13092063 0.39645183] | [10.00869369506836, 9.709908485412598] |
3dcce8f3-1757-41b6-872b-7bf6cfff4352 | skeleton-aware-networks-for-deep-motion | 2005.05732 | null | https://arxiv.org/abs/2005.05732v1 | https://arxiv.org/pdf/2005.05732v1.pdf | Skeleton-Aware Networks for Deep Motion Retargeting | We introduce a novel deep learning framework for data-driven motion retargeting between skeletons, which may have different structure, yet corresponding to homeomorphic graphs. Importantly, our approach learns how to retarget without requiring any explicit pairing between the motions in the training set. We leverage the fact that different homeomorphic skeletons may be reduced to a common primal skeleton by a sequence of edge merging operations, which we refer to as skeletal pooling. Thus, our main technical contribution is the introduction of novel differentiable convolution, pooling, and unpooling operators. These operators are skeleton-aware, meaning that they explicitly account for the skeleton's hierarchical structure and joint adjacency, and together they serve to transform the original motion into a collection of deep temporal features associated with the joints of the primal skeleton. In other words, our operators form the building blocks of a new deep motion processing framework that embeds the motion into a common latent space, shared by a collection of homeomorphic skeletons. Thus, retargeting can be achieved simply by encoding to, and decoding from this latent space. Our experiments show the effectiveness of our framework for motion retargeting, as well as motion processing in general, compared to existing approaches. Our approach is also quantitatively evaluated on a synthetic dataset that contains pairs of motions applied to different skeletons. To the best of our knowledge, our method is the first to perform retargeting between skeletons with differently sampled kinematic chains, without any paired examples. | ['Daniel Cohen-Or', 'Olga Sorkine-Hornung', 'Peizhuo Li', 'Kfir Aberman', 'Dani Lischinski', 'Baoquan Chen'] | 2020-05-12 | null | null | null | null | ['motion-retargeting'] | ['computer-vision'] | [ 1.65970087e-01 2.23805279e-01 -3.13486367e-01 6.11421913e-02
-4.32293296e-01 -6.65577650e-01 6.21047378e-01 -3.11499715e-01
-3.12172830e-01 4.43300813e-01 6.36098802e-01 1.19328484e-01
5.93036879e-03 -9.43411291e-01 -8.83622825e-01 -7.93771684e-01
-6.49501383e-02 1.47310853e-01 4.05328482e-01 -1.79736391e-01
-4.87288646e-02 7.00717211e-01 -1.13178456e+00 2.68851221e-03
3.86939108e-01 4.12898242e-01 -4.61955033e-02 8.56224716e-01
2.59051025e-01 5.22926569e-01 -1.57030433e-01 -3.23593408e-01
4.06335443e-01 -2.71848083e-01 -9.81770515e-01 3.61693799e-01
4.20225263e-01 -5.08447528e-01 -8.47747147e-01 7.23409772e-01
3.59578818e-01 4.56423670e-01 6.44692659e-01 -9.64847565e-01
-7.54020631e-01 6.01117611e-01 -7.70478785e-01 -2.67953239e-02
3.22944403e-01 2.50472933e-01 1.16084170e+00 -6.97891355e-01
9.56749082e-01 1.22387159e+00 6.25584781e-01 6.78173423e-01
-1.37897933e+00 -2.81807452e-01 -3.57998582e-03 -9.08912811e-03
-1.17162228e+00 -3.55717987e-01 8.34930301e-01 -5.54395497e-01
8.36652219e-01 1.50136858e-01 9.09303010e-01 1.17978120e+00
3.41274112e-01 7.40585208e-01 2.68534631e-01 -2.50178277e-01
2.17155874e-01 -6.92888796e-01 -3.65709811e-02 1.02419150e+00
6.65133223e-02 -3.13355923e-02 -6.04015470e-01 -2.35206857e-02
1.23485529e+00 1.11785784e-01 -5.28779387e-01 -9.45594430e-01
-1.69119668e+00 8.15807164e-01 4.84422207e-01 1.77692652e-01
-2.16753542e-01 7.47642398e-01 3.02577794e-01 1.74123179e-02
2.04087928e-01 3.19074899e-01 -3.62264782e-01 -5.29237948e-02
-6.83733940e-01 2.89926767e-01 6.27480686e-01 8.00494730e-01
1.04672432e+00 -1.55636340e-01 -3.20768684e-01 3.56369734e-01
2.52905965e-01 1.96484789e-01 7.58037746e-01 -1.21587121e+00
4.54203546e-01 4.76320922e-01 -3.58861648e-02 -1.11613786e+00
-5.32588482e-01 -2.05239132e-01 -8.00928175e-01 1.11053675e-01
4.53421295e-01 -1.68822452e-01 -9.22091305e-01 2.12721586e+00
3.36418062e-01 4.62761462e-01 6.41871393e-02 8.27592373e-01
4.72053200e-01 5.01620829e-01 -6.41623884e-02 2.33851522e-01
1.22132206e+00 -1.21855295e+00 -3.25398892e-01 -4.88962047e-02
8.08984578e-01 -5.65645218e-01 1.13070834e+00 -2.24002466e-01
-1.23733079e+00 -4.94907051e-01 -1.14761198e+00 -4.11446333e-01
-5.89299686e-02 1.36108726e-01 7.76471376e-01 3.06457490e-01
-1.13778043e+00 1.02795184e+00 -1.28969300e+00 -5.63720405e-01
1.96212173e-01 4.00845677e-01 -6.72347844e-01 1.96127757e-01
-1.04413009e+00 4.19144869e-01 2.80794650e-01 3.68877947e-02
-7.74985969e-01 -5.33750176e-01 -1.16886604e+00 -2.34870613e-02
3.25773209e-01 -1.36821270e+00 1.16032577e+00 -6.40082598e-01
-1.73311257e+00 8.68492067e-01 -3.55658345e-02 -3.00455332e-01
6.10410213e-01 -2.58938968e-01 -7.49918260e-03 2.05829963e-01
2.31989279e-01 9.54845726e-01 9.84321237e-01 -7.49902427e-01
-4.93248820e-01 -1.73178807e-01 3.19511592e-01 2.72574425e-01
-3.42376620e-01 -3.56504500e-01 -8.96684825e-01 -1.25019610e+00
2.76936054e-01 -1.15516198e+00 -4.09346551e-01 3.60525787e-01
-6.21549070e-01 7.54137039e-02 8.41444314e-01 -5.45474708e-01
1.11089754e+00 -2.25744176e+00 1.00610626e+00 1.92025229e-01
4.71219867e-01 -2.28869453e-01 -2.57953167e-01 4.74774390e-01
-1.61386296e-01 2.09411711e-01 -6.48589909e-01 -4.39317584e-01
3.47099751e-02 4.98539358e-01 -1.91742897e-01 6.05868340e-01
1.85375720e-01 1.21887136e+00 -1.07478559e+00 -2.26096928e-01
1.22635938e-01 4.46562231e-01 -7.06201553e-01 -5.58707751e-02
-7.49880075e-02 7.40794659e-01 -4.97025937e-01 2.44448960e-01
2.36632407e-01 -4.83964346e-02 3.05367634e-03 -3.77875954e-01
1.83588769e-02 1.53260961e-01 -1.07450712e+00 2.36177754e+00
-2.65616059e-01 4.77021664e-01 -1.90805018e-01 -8.62398386e-01
6.05779588e-01 2.02311948e-01 7.03890085e-01 -4.33618166e-02
-7.30551109e-02 2.12140188e-01 -1.75525293e-01 -5.38413823e-01
4.97812182e-01 -1.60572827e-02 -3.16207737e-01 6.89127684e-01
1.47136301e-01 -1.85718462e-02 2.29296207e-01 1.58168793e-01
1.33426511e+00 4.80235189e-01 2.44061351e-01 -1.76267341e-01
4.53311384e-01 -2.14632630e-01 6.39320791e-01 3.10765594e-01
-7.58968815e-02 7.49996603e-01 5.54594815e-01 -4.31939214e-01
-1.11354840e+00 -1.41291618e+00 3.68654609e-01 7.91046083e-01
1.09179512e-01 -4.14702743e-01 -7.72423148e-01 -7.79293358e-01
-8.54929984e-02 1.29645124e-01 -7.45437264e-01 -5.42213738e-01
-1.10675299e+00 -5.76410115e-01 5.70696354e-01 7.62730956e-01
5.27056634e-01 -8.32913876e-01 -8.03974748e-01 2.03761652e-01
-3.00596245e-02 -1.01702178e+00 -1.03139341e+00 -3.13118729e-03
-1.08363414e+00 -1.09470677e+00 -1.08504617e+00 -8.60621214e-01
6.20765388e-01 2.98724920e-01 6.99024439e-01 3.07806637e-02
-2.38595620e-01 5.95014334e-01 -1.50303185e-01 6.30664945e-01
-3.02131891e-01 3.08118254e-01 -2.89380234e-02 2.77073443e-01
-2.47179687e-01 -1.02834892e+00 -7.51006663e-01 2.53449261e-01
-1.22551048e+00 3.70218396e-01 5.72329044e-01 6.61587596e-01
6.24483645e-01 -3.80880833e-02 7.43009597e-02 -5.48026562e-01
3.42721373e-01 -4.83071595e-01 -2.17258975e-01 1.17043078e-01
2.08033115e-01 5.99764705e-01 3.48268211e-01 -4.13196266e-01
-8.90100062e-01 4.45952177e-01 4.74532209e-02 -4.49619472e-01
1.80401502e-03 4.14000481e-01 -4.35444623e-01 -7.26910830e-02
4.89741921e-01 -4.85779718e-03 -1.37100592e-01 -6.68372393e-01
1.01107383e+00 4.29153349e-03 9.49382007e-01 -6.16748691e-01
1.06899381e+00 9.86461937e-01 4.15118188e-01 -6.63770556e-01
-3.43311727e-01 -2.74455100e-01 -1.19678807e+00 -1.45043079e-02
1.12442517e+00 -5.85918784e-01 -3.45857471e-01 6.08489275e-01
-1.20927203e+00 -5.96712351e-01 -7.18956947e-01 5.49161971e-01
-8.89688611e-01 8.61618161e-01 -9.78509486e-01 3.96164954e-02
-7.84982666e-02 -1.16005421e+00 1.22012663e+00 -6.34701550e-03
-4.29276586e-01 -1.11118877e+00 3.73158246e-01 -1.91528467e-03
-4.72765900e-02 5.25487959e-01 9.89910185e-01 -1.60793945e-01
-6.27774656e-01 -5.31460065e-03 7.72200823e-02 -1.81810912e-02
4.82729942e-01 1.79132000e-01 -5.51967680e-01 -3.43857974e-01
-2.19698668e-01 2.08015032e-02 1.14451051e+00 2.47417495e-01
8.00750613e-01 -3.13183010e-01 -5.65211654e-01 1.01636696e+00
1.02461910e+00 -1.43394798e-01 6.19292557e-01 3.12212855e-01
1.14414191e+00 7.56398737e-01 8.27188343e-02 3.34545553e-01
2.91948974e-01 9.35270667e-01 3.39507341e-01 -1.11470476e-03
-4.18437779e-01 -3.67764920e-01 6.26340747e-01 9.38842952e-01
-2.29410961e-01 -4.34882427e-03 -7.09359586e-01 6.01536512e-01
-2.12186480e+00 -7.17990935e-01 -6.57269955e-02 2.13815475e+00
6.59059167e-01 -1.47365972e-01 3.48176025e-02 1.98575668e-02
7.43824542e-01 3.76843393e-01 -5.93917370e-01 -1.09041348e-01
5.70016215e-03 3.53015602e-01 5.85201383e-01 6.19496226e-01
-1.37462866e+00 1.07263005e+00 5.85789204e+00 7.27084100e-01
-1.02846718e+00 1.30836442e-01 2.71648150e-02 -1.03448354e-01
-4.14602548e-01 1.22251503e-01 -4.79247838e-01 2.40475789e-01
4.26003397e-01 -1.05681419e-01 3.91108453e-01 5.15487909e-01
1.30452618e-01 2.24275768e-01 -1.27722335e+00 6.80988789e-01
-1.23628683e-01 -1.43649912e+00 3.24252516e-01 -4.06653397e-02
8.19499493e-01 -1.91673264e-01 -1.18149500e-02 -1.26873642e-01
3.21553916e-01 -7.30997741e-01 8.77865374e-01 5.65878928e-01
6.13334477e-01 -6.72197104e-01 3.26198131e-01 -3.89485247e-02
-1.58388543e+00 2.08329231e-01 -1.62408561e-01 -2.28190627e-02
4.18612093e-01 2.97184080e-01 -2.97553003e-01 6.99545562e-01
5.41886270e-01 1.14666784e+00 -3.80276084e-01 8.41229856e-01
-5.40812135e-01 1.67485788e-01 -1.90817937e-01 4.64586347e-01
2.01119632e-01 -1.99636638e-01 8.09966087e-01 9.70978498e-01
4.13252145e-01 -2.40397722e-01 -1.98328737e-02 7.97813058e-01
-1.41567469e-01 -6.16969690e-02 -6.74913347e-01 1.00118019e-01
7.38049299e-02 1.21954608e+00 -8.99265885e-01 -2.87451833e-01
-4.54112351e-01 1.46641386e+00 4.18150246e-01 6.36893988e-01
-8.81285489e-01 -3.07427526e-01 1.04234672e+00 9.05475579e-03
4.05397505e-01 -5.51654935e-01 -4.43330631e-02 -1.39355135e+00
1.54785052e-01 -3.15908879e-01 3.17220718e-01 -4.99676436e-01
-1.12075543e+00 2.33411610e-01 1.52760595e-01 -1.21549463e+00
-2.16080740e-01 -5.34529269e-01 -7.02273786e-01 5.52861452e-01
-1.05834913e+00 -1.35182309e+00 -2.35923156e-01 7.06944644e-01
4.30846900e-01 4.70961742e-02 5.56755066e-01 2.36530542e-01
-6.76061809e-01 4.66568410e-01 -1.06639169e-01 2.73251086e-01
6.11149669e-01 -1.15753138e+00 8.85991037e-01 1.03782749e+00
2.89342850e-01 6.22358918e-01 3.71220380e-01 -6.35078430e-01
-1.31760800e+00 -1.32522058e+00 5.10225058e-01 -2.81773180e-01
8.54626894e-01 -2.48716906e-01 -8.39513302e-01 1.11598825e+00
-8.83283615e-02 -2.50935312e-02 5.47141194e-01 -4.42636609e-01
-2.28670165e-01 2.50539690e-01 -7.18725324e-01 1.15482426e+00
1.61087441e+00 -3.91943306e-01 -6.44056678e-01 1.92842022e-01
9.68121469e-01 -4.69199866e-01 -8.84135723e-01 4.34687287e-01
6.42559886e-01 -7.39454091e-01 1.18637061e+00 -7.30328739e-01
4.60922152e-01 -5.42552829e-01 -1.52439699e-01 -1.18629444e+00
-5.16006947e-01 -8.35251570e-01 -3.28469783e-01 9.35722470e-01
1.29669011e-01 -4.94111508e-01 1.00284624e+00 3.84793580e-01
-3.39048803e-01 -7.09782243e-01 -9.56743538e-01 -7.96792984e-01
4.70083095e-02 -3.89723659e-01 6.05674028e-01 9.80286777e-01
-1.11684807e-01 2.19414860e-01 -5.24105787e-01 6.40066788e-02
4.12285745e-01 1.23879872e-02 9.60607588e-01 -7.92387486e-01
-7.53309906e-01 -5.93361735e-01 -7.82950580e-01 -1.49472022e+00
2.80173153e-01 -1.15628231e+00 -9.33488682e-02 -1.38756835e+00
1.34472936e-01 -8.20802972e-02 9.06803384e-02 6.70438647e-01
-4.08949740e-02 4.39872861e-01 1.33553341e-01 4.69990343e-01
-8.08370784e-02 9.02749956e-01 1.51284945e+00 -2.00128704e-01
-4.61174846e-01 -1.01264976e-01 -3.83732885e-01 9.83562410e-01
7.76567996e-01 -3.60941350e-01 -3.53296995e-01 -6.74943626e-01
1.53625131e-01 -2.79625338e-02 6.06578887e-01 -1.05496740e+00
2.34638453e-01 -7.54869804e-02 -3.78464647e-02 -1.86926663e-01
3.66364896e-01 -5.46587586e-01 3.98204923e-01 6.74175620e-01
-1.36832312e-01 5.82826287e-02 1.03366971e-01 8.29091787e-01
-1.45093966e-02 -5.50413020e-02 5.42486608e-01 -3.88686508e-02
-9.14475381e-01 4.60918188e-01 -5.12144268e-01 -1.96066380e-01
1.13790858e+00 -3.19834560e-01 7.53364014e-03 -3.35057735e-01
-1.08604848e+00 -6.27544373e-02 8.61074030e-01 4.62709427e-01
5.36298752e-01 -1.61289680e+00 -4.59785134e-01 1.60693750e-01
-1.41511494e-02 2.19373047e-01 2.01793090e-01 9.36203361e-01
-8.03391099e-01 1.64255112e-01 -4.26709920e-01 -6.00139558e-01
-9.32359099e-01 5.55016816e-01 3.64143461e-01 -3.49409997e-01
-9.72791910e-01 6.73413157e-01 6.61004484e-01 -4.26665664e-01
1.81991479e-03 -6.82199180e-01 7.88837746e-02 -1.46326482e-01
5.68006486e-02 6.02169991e-01 -2.54010558e-01 -8.31308782e-01
-2.74203360e-01 1.18863583e+00 2.73960561e-01 -4.32839245e-01
1.01511526e+00 -6.95547163e-02 -2.92173713e-01 2.85147101e-01
1.33769047e+00 2.59830773e-01 -1.54679537e+00 -1.21747993e-01
1.04817390e-01 -3.46522361e-01 -4.07490343e-01 5.66714182e-02
-1.44318545e+00 7.51953244e-01 1.60821214e-01 -3.34790260e-01
1.04899180e+00 2.47865334e-01 1.07031929e+00 3.12109947e-01
2.24605009e-01 -6.81918919e-01 3.92415971e-01 3.56929332e-01
9.14014816e-01 -5.76114416e-01 -8.50836560e-02 -5.25640190e-01
-1.99701533e-01 1.32235026e+00 4.20559734e-01 -3.35703135e-01
5.70929348e-01 2.20735651e-02 -2.63347387e-01 -1.65926814e-01
-3.35810155e-01 -1.60310939e-01 3.15020323e-01 5.19644558e-01
2.06916451e-01 9.95647348e-03 -4.01005298e-01 3.79966140e-01
-1.59625962e-01 -1.66670997e-02 4.91130799e-01 1.02638471e+00
-2.40220353e-01 -1.46509397e+00 -2.41013274e-01 -2.56241411e-02
-1.94831952e-01 1.99165925e-01 -3.43255371e-01 8.96690726e-01
1.30305901e-01 4.30932343e-01 8.29132944e-02 -5.07683694e-01
2.84860730e-01 -1.57791644e-01 5.32567799e-01 -8.81315112e-01
-1.65647730e-01 1.68152954e-02 -2.24678427e-01 -8.84521663e-01
-5.35759926e-01 -6.31056011e-01 -1.46498966e+00 -1.97396919e-01
2.78018624e-01 -2.13417649e-01 8.45953524e-02 7.94935107e-01
4.19892371e-01 6.12659693e-01 3.31605971e-01 -1.27001154e+00
-2.07205415e-01 -5.85501373e-01 -5.76164961e-01 5.95150292e-01
3.05809408e-01 -8.02075088e-01 -1.96051851e-01 3.51902097e-01] | [7.452986240386963, -0.4098314344882965] |
94b2690c-6c2a-4558-bc82-ddb5b1eeaf7c | object-detection-in-aerial-images-what | 2201.08763 | null | https://arxiv.org/abs/2201.08763v1 | https://arxiv.org/pdf/2201.08763v1.pdf | Object Detection in Aerial Images: What Improves the Accuracy? | Object detection is a challenging and popular computer vision problem. The problem is even more challenging in aerial images due to significant variation in scale and viewpoint in a diverse set of object categories. Recently, deep learning-based object detection approaches have been actively explored for the problem of object detection in aerial images. In this work, we investigate the impact of Faster R-CNN for aerial object detection and explore numerous strategies to improve its performance for aerial images. We conduct extensive experiments on the challenging iSAID dataset. The resulting adapted Faster R-CNN obtains a significant mAP gain of 4.96% over its vanilla baseline counterpart on the iSAID validation set, demonstrating the impact of different strategies investigated in this work. | ['Abdelrahman Mohamed', 'Ikboljon Sobirov', 'Hashmat Shadab Malik'] | 2022-01-21 | null | null | null | null | ['object-detection-in-aerial-images'] | ['computer-vision'] | [ 9.98425260e-02 -5.28658569e-01 1.90786093e-01 -7.02362731e-02
-4.72025275e-01 -7.63333619e-01 4.33885366e-01 -2.00434074e-01
-6.34878039e-01 2.72579521e-01 -3.77488405e-01 -2.46772394e-02
4.80515733e-02 -6.37553275e-01 -5.16565681e-01 -6.04361355e-01
-3.03457946e-01 9.09668058e-02 6.78454876e-01 -3.94817770e-01
-1.17314272e-01 8.44263494e-01 -1.56722975e+00 1.10817328e-01
3.76938999e-01 1.16051161e+00 3.62059891e-01 8.96260500e-01
4.80508715e-01 4.64094073e-01 -8.62335324e-01 -1.42754078e-01
8.96995842e-01 -4.20594402e-03 -7.53594995e-01 3.33925933e-01
9.84886289e-01 -8.25261056e-01 -3.01959574e-01 1.06538904e+00
5.29069841e-01 1.21386461e-01 5.49489439e-01 -1.04061735e+00
-3.33118856e-01 3.48102123e-01 -1.02526331e+00 8.07667792e-01
-1.23377495e-01 2.79451132e-01 9.51128781e-01 -9.11214471e-01
2.20375031e-01 1.13474274e+00 7.12764800e-01 8.63454193e-02
-1.17951643e+00 -6.62734210e-01 1.68681547e-01 9.57903415e-02
-1.83361578e+00 8.25927928e-02 2.64251709e-01 -6.00399077e-01
9.10397410e-01 1.50201738e-01 5.70910096e-01 6.83394790e-01
6.58960342e-02 9.17890966e-01 7.70375371e-01 -2.65805364e-01
-4.95490320e-02 1.75904721e-01 7.39236623e-02 7.77944624e-01
7.32904494e-01 2.27736115e-01 2.17659980e-01 7.99664259e-02
7.56703377e-01 1.72850564e-01 -3.40543568e-01 -3.52131039e-01
-1.04461992e+00 8.60398114e-01 1.19988167e+00 9.50171649e-02
-3.38696927e-01 1.31165728e-01 3.18126500e-01 3.79736722e-02
3.16941649e-01 8.25892508e-01 -6.76339030e-01 5.24845302e-01
-9.29979563e-01 3.60145897e-01 3.13745677e-01 7.53252864e-01
1.82661340e-01 4.77104604e-01 -2.46582568e-01 5.68681657e-01
2.34783947e-01 8.70298982e-01 7.73156285e-02 -5.24716258e-01
3.30756873e-01 7.64777899e-01 4.68588978e-01 -1.15958226e+00
-6.38922334e-01 -8.40462387e-01 -7.31770813e-01 5.44022977e-01
4.67838347e-01 -1.98110029e-01 -9.06955123e-01 1.21106339e+00
5.34617543e-01 -1.06551751e-01 -1.11299410e-01 1.31080115e+00
1.04054856e+00 8.02335024e-01 4.12039123e-02 5.03148019e-01
1.33160675e+00 -7.42498934e-01 -9.43799391e-02 -4.48556393e-01
3.95546108e-01 -7.11847663e-01 6.82996631e-01 4.07886446e-01
-6.12368166e-01 -8.57043922e-01 -1.21681917e+00 1.07738748e-01
-3.50523770e-01 1.14886534e+00 4.86795247e-01 4.43299294e-01
-9.04042661e-01 2.93166429e-01 -6.04940891e-01 -7.03910172e-01
6.64695621e-01 5.22449851e-01 -2.22044662e-01 -6.76770061e-02
-6.37049496e-01 7.91156590e-01 6.00473344e-01 1.84674323e-01
-1.30428302e+00 -4.57010090e-01 -6.14394426e-01 1.29369304e-01
6.26406014e-01 -2.84353882e-01 1.21082520e+00 -7.78752446e-01
-8.85044277e-01 9.63745117e-01 5.11451066e-01 -8.97569120e-01
5.12780190e-01 -5.87861121e-01 -2.15815678e-01 6.20615929e-02
2.63950042e-02 8.94449651e-01 1.06104851e+00 -9.44145083e-01
-9.74764884e-01 -4.72973555e-01 4.71060902e-01 1.84477903e-02
-1.98543474e-01 2.22816542e-01 -2.48769358e-01 -4.94948477e-01
-2.28746444e-01 -1.16697204e+00 -2.41721630e-01 5.38890123e-01
-3.43794316e-01 -1.37458280e-01 1.09050155e+00 -3.93416554e-01
7.15650618e-01 -2.09313631e+00 1.64606407e-01 -4.05454785e-01
7.97365382e-02 9.12987411e-01 -2.46708795e-01 6.04866520e-02
1.18194455e-02 2.00709645e-02 -1.33793443e-01 1.01952985e-01
-4.04640675e-01 -1.77939370e-01 -5.25281966e-01 6.93114936e-01
6.47408605e-01 7.60175228e-01 -5.24153292e-01 -2.77795285e-01
4.80866343e-01 4.33313340e-01 -2.98041731e-01 2.36139223e-01
-1.16216630e-01 9.97599512e-02 -3.77836883e-01 1.10367858e+00
7.83721507e-01 -3.72548103e-01 -1.31926313e-01 -4.76557881e-01
-3.47749144e-01 -2.82107621e-01 -1.07393742e+00 1.10620940e+00
-1.82104692e-01 1.07526267e+00 1.07737005e-01 -7.62600958e-01
8.91219676e-01 -5.13514988e-02 -6.23355769e-02 -2.01601237e-01
5.17648637e-01 -1.62903160e-01 3.49873573e-01 -2.40770578e-01
6.76829159e-01 1.77882403e-01 1.54933676e-01 -1.71745330e-01
1.04264781e-01 -2.23711550e-01 2.28839844e-01 3.58559377e-02
9.63278890e-01 -9.47555061e-03 6.53510332e-01 -4.91550058e-01
4.56929922e-01 4.13844407e-01 1.72963217e-01 8.47200930e-01
-4.02402312e-01 4.92190450e-01 -4.43180166e-02 -9.69341636e-01
-7.21781850e-01 -7.60879636e-01 -4.88527596e-01 9.62633073e-01
4.28017050e-01 -1.08209260e-01 -4.71435547e-01 -8.13783050e-01
1.31050542e-01 3.34518850e-01 -8.02115202e-01 -5.14709242e-02
-3.29243094e-01 -1.06179976e+00 5.57068467e-01 8.56910944e-01
8.40797603e-01 -9.13267314e-01 -1.22768950e+00 5.87893017e-02
8.16744044e-02 -1.44622970e+00 1.07756615e-01 2.46337831e-01
-8.58714283e-01 -1.18957436e+00 -6.62008882e-01 -4.93804604e-01
4.82976556e-01 1.13780332e+00 9.71884370e-01 1.55842707e-01
-1.08934259e+00 3.14527422e-01 -4.19923455e-01 -7.78551340e-01
1.26703277e-01 1.33228540e-01 1.72061309e-01 -1.63216561e-01
2.66813606e-01 2.44581223e-01 -7.62870729e-01 5.52185476e-01
-9.05131400e-01 -4.74948615e-01 7.84582675e-01 7.14747906e-01
3.49485040e-01 1.12681590e-01 -5.40577471e-02 -1.80199087e-01
-1.31980196e-01 -2.67212719e-01 -1.27949059e+00 8.68918970e-02
-1.05652176e-01 -2.73141146e-01 4.46793258e-01 -3.57215255e-01
-6.04303896e-01 3.27769578e-01 2.32681677e-01 -3.86275560e-01
-3.56585413e-01 1.05331719e-01 1.68542489e-01 -5.44863880e-01
8.36616218e-01 -7.38376305e-02 -3.51576418e-01 -2.63194680e-01
8.74966905e-02 4.49827164e-01 3.15984964e-01 -2.38686334e-02
1.16114604e+00 7.96932697e-01 9.44701433e-02 -1.25300872e+00
-1.31570876e+00 -6.67788982e-01 -8.02378535e-01 -2.48973310e-01
1.01619720e+00 -1.44670904e+00 -3.91386688e-01 4.68580931e-01
-1.18718874e+00 -4.04023498e-01 -2.31880248e-01 4.99558359e-01
-8.89446400e-03 9.49924365e-02 -2.91645169e-01 -7.60580301e-01
-3.95540535e-01 -1.17021692e+00 1.32761109e+00 3.71573150e-01
1.68690816e-01 -5.53176761e-01 -8.02099258e-02 1.04549296e-01
2.93084681e-01 4.10688758e-01 1.93768293e-01 -5.50864875e-01
-8.99360359e-01 -5.84207952e-01 -8.08693886e-01 4.80401814e-01
3.02071329e-02 3.19236606e-01 -9.04161692e-01 -5.63793242e-01
-5.04935861e-01 -3.80936414e-01 1.17294979e+00 5.50609946e-01
8.53845060e-01 1.67123020e-01 -4.21166956e-01 7.30647564e-01
1.70497715e+00 1.05147116e-01 2.33653158e-01 4.62225944e-01
6.21055186e-01 3.31133515e-01 1.01860201e+00 5.02583385e-01
-2.18434498e-01 9.27495718e-01 8.86361897e-01 -4.51729476e-01
-6.85382262e-02 2.74586409e-01 2.58516759e-01 -2.60758042e-01
-2.90831327e-01 -4.62040961e-01 -9.68373775e-01 5.47217250e-01
-1.54893875e+00 -7.38576353e-01 -2.79444277e-01 1.93967855e+00
1.52932510e-01 1.13037243e-01 1.92272320e-01 -1.28717988e-03
5.97959757e-01 1.41378045e-01 -4.51894343e-01 2.62086809e-01
-7.31030703e-02 1.09783970e-02 9.19769824e-01 -1.63518906e-01
-1.77380466e+00 1.14101017e+00 6.69369364e+00 4.81467813e-01
-1.20003545e+00 -1.81685477e-01 2.85510957e-01 -8.49117786e-02
1.00989199e+00 -4.62899864e-01 -1.33296549e+00 8.90624598e-02
3.36594373e-01 1.37504190e-01 6.99596852e-02 1.40952861e+00
-6.22347258e-02 -1.77682698e-01 -7.11752832e-01 9.11297321e-01
1.76647231e-01 -1.08169878e+00 7.30093196e-02 -1.28883775e-02
7.00997889e-01 5.27506769e-01 1.66993797e-01 4.49966013e-01
5.07510126e-01 -8.31399083e-01 6.49912953e-01 -2.51059115e-01
6.75281346e-01 -7.10254610e-01 1.05437136e+00 1.02116860e-01
-1.48450804e+00 -5.66279113e-01 -9.97431338e-01 -1.87775686e-01
-1.96725726e-02 1.73007771e-01 -9.73917902e-01 2.33667806e-01
1.32873201e+00 6.94006503e-01 -1.05510473e+00 1.36390305e+00
-1.86381698e-01 6.70424759e-01 -3.46363008e-01 8.79469439e-02
5.38857579e-01 -6.79057017e-02 6.98191822e-01 1.24064934e+00
4.14532572e-01 2.47515991e-01 4.74284470e-01 7.62363791e-01
-8.60383064e-02 -1.94935933e-01 -9.20529544e-01 7.64909163e-02
1.66762508e-02 1.83172667e+00 -9.95221436e-01 -2.63993651e-01
-2.96193391e-01 8.71904016e-01 1.31876737e-01 1.86979279e-01
-1.07492781e+00 -2.83282548e-01 8.23613346e-01 2.30952248e-01
1.02859557e+00 -4.55626249e-01 4.13971603e-01 -8.75251830e-01
-4.14405227e-01 -7.12233663e-01 4.21040952e-01 -7.91702509e-01
-9.22874868e-01 9.00584340e-01 2.10286126e-01 -1.30040824e+00
2.97970712e-01 -1.27679002e+00 -6.50084794e-01 4.19688761e-01
-1.47481453e+00 -1.20470202e+00 -9.34158087e-01 2.97387183e-01
7.66507626e-01 -3.55583668e-01 5.08115232e-01 7.75152370e-02
-8.45707953e-01 2.18041107e-01 2.16492750e-02 3.56832802e-01
4.47044104e-01 -1.15097415e+00 6.41681969e-01 1.28533733e+00
3.62007082e-01 1.59352511e-01 6.07497573e-01 -2.68831402e-01
-1.31499374e+00 -1.55761611e+00 -1.84309006e-01 -5.43956399e-01
4.89817768e-01 -2.10343346e-01 -6.81905925e-01 6.39115989e-01
9.49658453e-02 2.66328603e-01 3.23717922e-01 -2.19528124e-01
-2.84857959e-01 -2.07821414e-01 -1.01729500e+00 4.01100308e-01
6.91546679e-01 6.57123551e-02 -3.36207360e-01 3.90695393e-01
8.17596138e-01 -2.61242181e-01 -4.38867182e-01 8.03964853e-01
2.33595535e-01 -9.29261446e-01 1.46173286e+00 -5.38749278e-01
1.04122557e-01 -7.81037748e-01 -3.68686259e-01 -1.11264586e+00
-5.54234922e-01 -7.29483366e-02 2.20408767e-01 8.59662235e-01
1.08234808e-01 -2.54070997e-01 4.67341989e-01 -1.02440186e-01
7.37694055e-02 -3.25429410e-01 -5.78547478e-01 -1.01030505e+00
-3.43989193e-01 -6.60460442e-02 2.27576345e-01 4.80911463e-01
-8.58831406e-01 3.04036647e-01 -4.89203006e-01 5.76651871e-01
6.43263638e-01 3.79537016e-01 1.09516072e+00 -1.43594301e+00
-2.28002116e-01 -1.67208299e-01 -6.76694036e-01 -1.01460373e+00
-2.61387855e-01 -6.31430924e-01 2.92440921e-01 -1.37912333e+00
7.41717815e-02 -2.45474353e-02 -9.05910134e-02 3.33395481e-01
-2.02607840e-01 8.11689854e-01 3.85199308e-01 8.90218168e-02
-7.72865891e-01 4.24237877e-01 7.78608382e-01 -4.37455505e-01
-3.03073637e-02 7.60298669e-02 -3.25832695e-01 7.04497874e-01
1.15801704e+00 -4.10569578e-01 -1.87689159e-02 -5.54735184e-01
-1.38838500e-01 -3.98423553e-01 7.86957204e-01 -1.45023286e+00
-2.06393644e-01 1.42828926e-01 8.49590898e-01 -8.92752767e-01
4.41969842e-01 -9.34501112e-01 -3.06271672e-01 8.77751589e-01
-3.88905294e-02 1.29535332e-01 8.57742548e-01 6.59842670e-01
-4.52242419e-02 -1.19462416e-01 1.20881951e+00 -1.84502393e-01
-1.05419648e+00 2.49116823e-01 -3.25247765e-01 -8.36339146e-02
1.26969075e+00 -3.63726430e-02 -1.22081056e-01 1.50268152e-01
-2.39174619e-01 2.86879800e-02 3.03986728e-01 5.50995171e-01
5.78872144e-01 -8.84348273e-01 -8.92016172e-01 2.02183709e-01
4.30882096e-01 4.95761037e-02 -6.02495857e-02 8.75769794e-01
-7.52593696e-01 5.33530056e-01 -4.20239031e-01 -8.26911151e-01
-1.69441712e+00 5.40741265e-01 4.38108712e-01 -6.40866458e-02
-6.02152228e-01 1.15397561e+00 5.68289280e-01 -1.72794044e-01
2.09661558e-01 -6.98460519e-01 -2.73790509e-01 7.49887899e-03
8.08949232e-01 5.60689688e-01 -9.74049792e-02 -5.72871864e-01
-4.91331339e-01 6.47584438e-01 -1.88720882e-01 4.52475905e-01
1.23259878e+00 7.60733560e-02 2.42190138e-01 1.20301461e-02
7.42500007e-01 -4.89581466e-01 -1.29769683e+00 -2.85333276e-01
-1.69628307e-01 -6.44704163e-01 3.57053846e-01 -7.79385865e-01
-1.07589853e+00 8.28879774e-01 1.08246410e+00 1.85000941e-01
8.98116529e-01 5.03645949e-02 3.94062847e-01 7.43539751e-01
4.32921827e-01 -6.35693610e-01 2.31947988e-01 4.39661920e-01
1.05446792e+00 -1.77657413e+00 4.69730765e-01 -2.82783151e-01
-5.60659707e-01 1.14594603e+00 8.32964182e-01 -4.50761378e-01
3.42937320e-01 1.64151609e-01 1.40524045e-01 -4.73620236e-01
-2.63613522e-01 -7.32187390e-01 5.41848660e-01 5.60648501e-01
2.01862887e-01 9.33001041e-02 2.18391210e-01 -2.07498460e-03
2.87258804e-01 -2.43695915e-01 4.97612804e-01 7.32006967e-01
-6.21990144e-01 -4.74478483e-01 -5.47362089e-01 3.08512688e-01
-3.93067151e-01 -1.38242105e-02 -6.52677953e-01 1.26368093e+00
3.80810685e-02 8.48748565e-01 4.79046404e-02 -3.22219104e-01
4.17667866e-01 -5.20575643e-01 3.32753539e-01 -6.63643897e-01
-5.78632832e-01 -2.51648158e-01 -3.04053128e-01 -5.65365195e-01
-3.53667200e-01 -6.48418963e-01 -8.42244983e-01 1.17270015e-01
-8.72028649e-01 -3.57477546e-01 6.21983409e-01 5.11465609e-01
8.94390047e-02 6.27739727e-01 3.44591469e-01 -1.18297541e+00
-8.98993611e-01 -7.95669258e-01 -5.73002160e-01 -5.19402549e-02
3.40645254e-01 -9.39423919e-01 -4.55848008e-01 -1.79017693e-01] | [8.64217472076416, -0.8473261594772339] |
154f06e3-9f47-46ab-9e8c-b1759f627e48 | a-graph-based-and-patient-demographics-aware | 2010.10699 | null | https://arxiv.org/abs/2010.10699v2 | https://arxiv.org/pdf/2010.10699v2.pdf | A Weighted Heterogeneous Graph Based Dialogue System | Knowledge based dialogue systems have attracted increasing research interest in diverse applications. However, for disease diagnosis, the widely used knowledge graph is hard to represent the symptom-symptom relations and symptom-disease relations since the edges of traditional knowledge graph are unweighted. Most research on disease diagnosis dialogue systems highly rely on data-driven methods and statistical features, lacking profound comprehension of symptom-disease relations and symptom-symptom relations. To tackle this issue, this work presents a weighted heterogeneous graph based dialogue system for disease diagnosis. Specifically, we build a weighted heterogeneous graph based on symptom co-occurrence and a proposed symptom frequency-inverse disease frequency. Then this work proposes a graph based deep Q-network (Graph-DQN) for dialogue management. By combining Graph Convolutional Network (GCN) with DQN to learn the embeddings of diseases and symptoms from both the structural and attribute information in the weighted heterogeneous graph, Graph-DQN could capture the symptom-disease relations and symptom-symptom relations better. Experimental results show that the proposed dialogue system rivals the state-of-the-art models. More importantly, the proposed dialogue system can complete the task with less dialogue turns and possess a better distinguishing capability on diseases with similar symptoms. | ['Huanhuan Chen', 'LiangWei Chen', 'Xinyan Zhao'] | 2020-10-21 | null | null | null | null | ['dialogue-management'] | ['natural-language-processing'] | [-2.70510077e-01 5.85886002e-01 -4.58462179e-01 -2.96975374e-01
-7.65208080e-02 -7.82371461e-02 2.44252935e-01 7.13240445e-01
-3.07826418e-02 6.48726761e-01 5.83931029e-01 -1.22181736e-01
-6.78573608e-01 -1.22856426e+00 3.36572617e-01 -5.44420183e-01
-3.61433238e-01 7.58292317e-01 2.72397369e-01 -8.04649055e-01
-7.24854097e-02 -9.07655731e-02 -5.17928183e-01 -1.49810389e-01
1.17684460e+00 6.35924816e-01 -3.74144129e-02 9.01410758e-01
-4.77612138e-01 1.07918298e+00 -6.34366035e-01 -6.04979396e-01
-4.07039165e-01 -9.53755260e-01 -1.05891383e+00 7.50705004e-02
-2.90750533e-01 -2.18277097e-01 -8.70317757e-01 9.26867127e-01
7.01439917e-01 -6.12205788e-02 8.05642664e-01 -1.53053522e+00
-1.36801124e+00 5.21810412e-01 -1.75163791e-01 1.30400911e-01
7.65525937e-01 7.70414323e-02 1.28650451e+00 -1.81695566e-01
7.09621668e-01 1.52464187e+00 6.64013147e-01 7.04075038e-01
-7.56706536e-01 -1.58681914e-01 2.07658783e-01 4.84487325e-01
-1.09208488e+00 5.13819814e-01 7.13433385e-01 -2.41673142e-01
1.04285312e+00 3.96781236e-01 1.18295014e+00 9.65199590e-01
3.38100761e-01 4.36105490e-01 5.92633605e-01 3.54894623e-02
-1.50051579e-01 2.30062585e-02 4.78112996e-01 1.41706479e+00
1.58696622e-01 -3.77660543e-01 -2.09930405e-01 -5.35351753e-01
6.46290600e-01 4.22984421e-01 -6.09782040e-01 -3.02390426e-01
-1.11080980e+00 1.20180452e+00 8.21906865e-01 5.18271685e-01
-3.32044274e-01 -2.39959970e-01 7.75480211e-01 4.87805307e-01
7.25071013e-01 3.23331684e-01 -5.16822278e-01 3.54284078e-01
-2.42126912e-01 2.49223039e-01 1.33253145e+00 6.58291578e-01
2.41687939e-01 -9.13477466e-02 -4.78213251e-01 1.03817630e+00
4.54427570e-01 3.46872926e-01 5.88592410e-01 -2.93375552e-01
2.64588088e-01 1.27489352e+00 -5.03106415e-01 -1.64528704e+00
-1.07273197e+00 -2.16444030e-01 -1.42466593e+00 -7.10889578e-01
1.09049745e-01 -4.06953841e-01 -6.07938707e-01 1.69418609e+00
6.25724733e-01 2.55995452e-01 2.97830164e-01 9.12223756e-01
1.51530290e+00 3.82189363e-01 2.02570766e-01 -8.11772719e-02
1.74854624e+00 -1.01286983e+00 -1.08873975e+00 3.20042342e-01
9.87596571e-01 -2.74428487e-01 5.76235831e-01 -1.46562338e-01
-6.18335962e-01 -1.70891240e-01 -6.05422378e-01 -9.89901796e-02
-7.54596889e-01 -3.74657005e-01 9.61291909e-01 3.20918530e-01
-1.12568128e+00 3.82072091e-01 -4.53585804e-01 -8.35711002e-01
2.70791292e-01 3.52419198e-01 -4.51751918e-01 -1.69874012e-01
-2.06897163e+00 1.10175335e+00 4.20686662e-01 1.58558950e-01
-3.28443229e-01 -6.84779108e-01 -1.24971867e+00 6.79885969e-02
3.87093633e-01 -1.33563280e+00 8.34088087e-01 -4.35799479e-01
-1.21982086e+00 7.23663330e-01 3.92954499e-01 -2.73725808e-01
1.81468874e-01 2.89987713e-01 -9.07779157e-01 3.29080462e-01
-1.71300948e-01 2.31523976e-01 3.00067276e-01 -6.56644881e-01
-2.83083171e-01 -3.61626506e-01 2.67130852e-01 5.72445631e-01
-4.52638388e-01 -4.79680836e-01 -3.62900257e-01 -4.93771166e-01
-2.22376123e-01 -7.05963194e-01 -3.20434451e-01 2.26906985e-01
-8.70285630e-01 -7.26843119e-01 6.73021436e-01 -5.93305409e-01
1.37402785e+00 -1.77800548e+00 3.06791276e-01 3.01594269e-02
1.09490669e+00 5.36938787e-01 -2.72108674e-01 1.05768883e+00
1.60558149e-01 -3.18002738e-02 -2.02103749e-01 2.69728214e-01
6.34521618e-02 3.99262041e-01 3.36819381e-01 4.58377361e-01
5.56900740e-01 1.13856077e+00 -1.45011497e+00 -7.40034878e-01
2.17639402e-01 6.80244327e-01 -4.88063931e-01 5.67673206e-01
-2.68738538e-01 1.77667901e-01 -8.78831506e-01 4.86949414e-01
5.33988297e-01 -6.82867467e-01 5.81900477e-01 -3.87347579e-01
6.67385817e-01 7.02832965e-03 -5.15959740e-01 1.53362226e+00
-2.16888323e-01 2.66534716e-01 9.74123850e-02 -1.30730212e+00
9.42758977e-01 5.13925970e-01 8.46210182e-01 -4.83542353e-01
1.24509700e-01 -1.67669654e-01 4.48723465e-01 -1.15906608e+00
1.40669107e-01 4.26413231e-02 -1.48817571e-02 3.09167176e-01
3.00916225e-01 -1.69785872e-01 -2.54842099e-02 6.57543361e-01
1.38718200e+00 -5.19785225e-01 5.23851931e-01 -1.10644795e-01
7.18554854e-01 1.77424207e-01 3.70916218e-01 2.09168613e-01
-4.51069534e-01 1.89152732e-01 9.68616307e-01 -2.99391270e-01
-4.28153694e-01 -1.04467189e+00 -1.53867051e-01 9.07211602e-01
3.47398460e-01 -5.31774104e-01 -5.97898424e-01 -1.06188238e+00
3.71047527e-01 1.16646051e-01 -7.96847582e-01 -7.81940281e-01
-1.29906744e-01 -1.04604185e+00 6.57492399e-01 1.65466636e-01
5.85460007e-01 -1.06906819e+00 2.87233382e-01 2.41584376e-01
-8.30461755e-02 -5.99644363e-01 -7.54783094e-01 -2.33377695e-01
-4.48757231e-01 -1.81340110e+00 -9.84378874e-01 -1.24724925e+00
6.76603377e-01 1.90805852e-01 1.35827100e+00 5.53612411e-01
-6.15695834e-01 7.02595055e-01 -6.52392685e-01 -9.05888751e-02
-3.26557308e-01 1.26091704e-01 -1.44542113e-01 -1.57856047e-01
6.36048794e-01 -3.76025081e-01 -1.07747149e+00 4.48710136e-02
-9.75838542e-01 -9.76042263e-03 5.43167114e-01 1.19880593e+00
1.37442574e-01 -5.93691319e-02 1.02824914e+00 -1.38447165e+00
1.38236821e+00 -1.10232246e+00 2.89008647e-01 5.03092349e-01
-8.58487785e-01 -7.51799271e-02 7.09841073e-01 -3.68001908e-01
-7.12505162e-01 -4.87540781e-01 -2.67320186e-01 7.45136961e-02
2.23322645e-01 8.08713615e-01 1.09872676e-01 3.65115590e-02
3.67931128e-01 2.62979776e-01 1.28949285e-01 -7.49800801e-02
7.36872375e-01 7.93592751e-01 6.94603100e-02 1.20754398e-01
2.21533522e-01 2.45550647e-03 4.98861745e-02 -6.75854445e-01
-8.41656089e-01 -5.93738198e-01 -2.81086266e-01 4.29129153e-02
1.29834497e+00 -7.16363609e-01 -1.02966130e+00 3.95384312e-01
-1.03835142e+00 7.26659819e-02 1.96277052e-02 4.64589387e-01
-1.81036294e-01 3.89508188e-01 -9.26785350e-01 -5.03984153e-01
-7.31946111e-01 -6.98665082e-01 9.28879678e-01 3.01764578e-01
-2.31954500e-01 -1.90269852e+00 5.18765271e-01 1.53335750e-01
6.28290832e-01 4.86447453e-01 1.37280214e+00 -1.08856094e+00
5.75014763e-02 -1.80835754e-01 -6.42975569e-01 -1.22922733e-01
7.59171784e-01 -2.86720932e-01 -3.29937011e-01 -6.08091913e-02
-1.77159220e-01 -4.49812174e-01 8.46882105e-01 3.44349980e-01
7.22791553e-01 -4.06266302e-01 -3.41001749e-01 3.48300040e-01
1.12300885e+00 -1.61640663e-02 9.32286531e-02 -3.34070414e-01
1.14977062e+00 7.94794023e-01 2.56344259e-01 5.35859406e-01
1.28275096e+00 4.41389203e-01 4.20651704e-01 -4.24443007e-01
-1.34412587e-01 -1.46958292e-01 -1.08920351e-01 1.49885297e+00
3.33601147e-01 -5.56827486e-01 -8.36958706e-01 5.62178612e-01
-2.00533509e+00 -5.49277723e-01 -5.19373119e-01 1.35729504e+00
1.07885838e+00 -3.03638190e-01 2.13713512e-01 -2.23396763e-01
7.28350103e-01 3.51443917e-01 -7.12335348e-01 -5.57859361e-01
-1.14280805e-02 -8.22867900e-02 1.14554994e-01 5.15435159e-01
-8.74832153e-01 6.75772607e-01 5.59367275e+00 5.63145280e-01
-7.99512446e-01 -4.39124554e-02 3.44608068e-01 4.17558283e-01
-4.75633055e-01 -5.82528710e-01 -1.87527895e-01 4.28170025e-01
7.53763735e-01 -4.86890048e-01 1.43513888e-01 5.62326968e-01
-9.98698622e-02 1.18260190e-01 -9.54062343e-01 8.91790986e-01
1.77041754e-01 -1.03734624e+00 1.87518075e-01 -1.40355617e-01
4.29622054e-01 -2.27868676e-01 -1.09589577e-01 5.47116518e-01
6.93403125e-01 -1.02205741e+00 -6.15502954e-01 9.66693044e-01
7.61136591e-01 -5.29995024e-01 1.07815492e+00 1.97761226e-02
-1.46770775e+00 1.70447394e-01 -4.32606488e-01 3.19772989e-01
3.14712003e-02 6.78369284e-01 -1.08802211e+00 1.15239561e+00
1.50451824e-01 1.14157951e+00 -4.87788379e-01 7.73091972e-01
-6.15334697e-02 3.80646229e-01 3.41735423e-01 -4.42166239e-01
2.64542341e-01 -4.01059955e-01 1.96164951e-01 1.43026137e+00
5.22093251e-02 3.30694795e-01 6.44910216e-01 6.07061744e-01
-1.43277958e-01 5.35241604e-01 -9.34152603e-01 -5.15792251e-01
2.08152339e-01 1.49015081e+00 -4.51747686e-01 -4.14996654e-01
-5.25844276e-01 1.06895804e+00 4.36440527e-01 9.31090191e-02
-6.16651595e-01 -6.41718566e-01 6.60885930e-01 -3.52032751e-01
-1.78781137e-01 2.67939955e-01 4.88594979e-01 -1.17240477e+00
-5.07955253e-01 -6.52947128e-01 7.56017148e-01 -4.11316484e-01
-2.02560782e+00 8.01873863e-01 -3.82328510e-01 -1.06372046e+00
-2.55051136e-01 -4.97615963e-01 -6.57093108e-01 8.89714777e-01
-1.50513268e+00 -1.22588837e+00 -3.55886489e-01 9.93155003e-01
1.21525012e-01 -2.02936798e-01 1.53449368e+00 1.77677989e-01
-7.26809144e-01 4.40492809e-01 -1.05028048e-01 5.82168698e-01
7.52797306e-01 -1.72599387e+00 1.31205425e-01 -2.63179123e-01
-2.64445812e-01 4.76208746e-01 1.72414333e-01 -7.77552664e-01
-1.47071183e+00 -1.27210462e+00 9.88521934e-01 -2.07356021e-01
9.64606047e-01 -1.24899879e-01 -1.12725711e+00 1.33637652e-01
5.05761921e-01 1.29491791e-01 1.06181848e+00 1.56673297e-01
-3.02367866e-01 1.20974779e-01 -1.36818314e+00 4.46610183e-01
1.04310632e+00 -7.06419289e-01 -6.12510085e-01 9.30412114e-01
1.10685432e+00 -2.43796602e-01 -1.46941423e+00 2.19050065e-01
2.90751487e-01 -6.33799255e-01 8.82059932e-01 -1.09302568e+00
3.89611840e-01 1.92427322e-01 2.68136382e-01 -1.89936447e+00
-5.07644475e-01 -2.55933464e-01 -2.76804388e-01 9.48657215e-01
2.03436479e-01 -7.33813763e-01 4.29394782e-01 2.93625683e-01
1.07305802e-01 -1.30289543e+00 -5.04106104e-01 -2.44349241e-01
-6.65047467e-02 2.57023454e-01 3.71851295e-01 1.57932162e+00
4.94796008e-01 8.04669678e-01 -3.86875480e-01 -5.73082687e-03
1.05377816e-01 7.75967836e-02 2.70264030e-01 -1.48215711e+00
-1.72600016e-01 -4.86649215e-01 -7.40188301e-01 -6.09343767e-01
8.70691687e-02 -1.24705470e+00 -2.75915504e-01 -2.21498060e+00
2.63138860e-01 -1.56702474e-01 -4.33158845e-01 3.49133611e-01
-8.20805609e-01 -2.44578049e-01 -1.71738848e-01 -2.29239047e-01
-9.60251272e-01 7.50258744e-01 1.80384171e+00 -4.87345159e-01
-1.74484015e-01 -2.37138763e-01 -7.37769723e-01 5.50721169e-01
7.57161319e-01 -3.38092715e-01 -8.90948534e-01 -1.10517733e-01
2.78472751e-01 5.88146269e-01 9.82699841e-02 -3.32100481e-01
3.44160616e-01 -1.01507187e-01 -1.02222040e-01 -2.09060416e-01
2.20665351e-01 -7.25439548e-01 -2.69962609e-01 8.34464669e-01
-5.10590911e-01 1.16864234e-01 -2.32119471e-01 1.04139447e+00
-3.40884268e-01 2.01218620e-01 3.79468173e-01 -2.41412252e-01
-4.19570059e-01 6.13879323e-01 -2.66144723e-01 4.95005608e-01
1.02518845e+00 3.52021754e-01 -7.08727181e-01 -5.53609073e-01
-8.66909862e-01 8.52034032e-01 -2.60752320e-01 5.38647652e-01
9.23440576e-01 -1.39323533e+00 -9.89611506e-01 -1.99815616e-01
3.03318113e-01 9.93036199e-03 5.92270374e-01 1.17482483e+00
-7.19957948e-01 4.94719416e-01 -5.51508591e-02 -3.68580937e-01
-1.20040917e+00 6.39488041e-01 4.69931901e-01 -7.91117251e-01
-7.28217125e-01 9.99110878e-01 1.98472098e-01 -1.03057802e+00
2.56427109e-01 -2.15630636e-01 -6.97268724e-01 2.87950695e-01
3.25381935e-01 2.75107712e-01 -2.97870666e-01 -4.80885595e-01
-3.62855613e-01 5.36654472e-01 -2.41840988e-01 4.84662980e-01
1.18007362e+00 -4.77541313e-02 -4.79209334e-01 3.68155330e-01
1.27092624e+00 -3.62305522e-01 -3.10371816e-01 -4.90358740e-01
-1.17502034e-01 -5.01826368e-02 -1.20965876e-01 -8.89662206e-01
-1.32696450e+00 6.39460385e-01 3.78295004e-01 1.07456470e+00
1.00773573e+00 -9.55748707e-02 1.17347729e+00 5.33685148e-01
-9.46075022e-02 -6.23349428e-01 1.98817835e-03 4.14577693e-01
7.58526206e-01 -1.33073854e+00 -6.95436820e-02 -6.09019935e-01
-9.44356263e-01 1.16686237e+00 6.04577839e-01 -2.51358867e-01
1.17907786e+00 -2.53166914e-01 2.24622831e-01 -8.64892185e-01
-9.03960943e-01 -4.35332149e-01 6.02494657e-01 8.97213340e-01
5.92045605e-01 3.13356489e-01 -6.36310935e-01 6.69138432e-01
-6.84019849e-02 -2.37286478e-01 2.27051526e-01 5.10669947e-01
-2.57354528e-01 -1.14829171e+00 2.58462727e-01 8.98166597e-01
-2.57902414e-01 -2.35611618e-01 -9.74673390e-01 7.30694234e-01
5.46870306e-02 1.15388453e+00 -2.11111039e-01 -6.37123644e-01
4.24712628e-01 5.05426573e-03 2.44940996e-01 -8.64071369e-01
-8.87398303e-01 -4.37115222e-01 2.28027314e-01 -4.88400817e-01
-5.17054141e-01 1.45429716e-01 -1.53801990e+00 -4.64795679e-01
-4.88650799e-01 3.15237343e-01 9.37944427e-02 8.04535270e-01
5.19263268e-01 1.13679516e+00 7.00559020e-01 1.18628323e-01
-3.54797661e-01 -1.01975632e+00 -1.01094973e+00 6.31678998e-01
4.77104217e-01 -5.22938013e-01 -4.70651351e-02 -4.98995781e-01] | [7.687839984893799, 6.774497985839844] |
008b2c3b-313b-4c27-be91-0ad854bb1c48 | a-hierarchical-approach-for-joint-multi-view | 1503.01393 | null | http://arxiv.org/abs/1503.01393v1 | http://arxiv.org/pdf/1503.01393v1.pdf | A Hierarchical Approach for Joint Multi-view Object Pose Estimation and Categorization | We propose a joint object pose estimation and categorization approach which
extracts information about object poses and categories from the object parts
and compositions constructed at different layers of a hierarchical object
representation algorithm, namely Learned Hierarchy of Parts (LHOP). In the
proposed approach, we first employ the LHOP to learn hierarchical part
libraries which represent entity parts and compositions across different object
categories and views. Then, we extract statistical and geometric features from
the part realizations of the objects in the images in order to represent the
information about object pose and category at each different layer of the
hierarchy. Unlike the traditional approaches which consider specific layers of
the hierarchies in order to extract information to perform specific tasks, we
combine the information extracted at different layers to solve a joint object
pose estimation and categorization problem using distributed optimization
algorithms. We examine the proposed generative-discriminative learning approach
and the algorithms on two benchmark 2-D multi-view image datasets. The proposed
approach and the algorithms outperform state-of-the-art classification,
regression and feature extraction algorithms. In addition, the experimental
results shed light on the relationship between object categorization, pose
estimation and the part realizations observed at different layers of the
hierarchy. | ['Mete Ozay', 'Krzysztof Walas', 'Ales Leonardis'] | 2015-03-04 | null | null | null | null | ['object-categorization'] | ['computer-vision'] | [ 9.24076047e-03 -8.55828077e-02 -1.29175350e-01 -3.94393712e-01
-7.98730493e-01 -7.35466719e-01 5.93600810e-01 2.79124349e-01
9.40374658e-02 2.51949012e-01 1.09160252e-01 5.69415271e-01
-2.94308364e-01 -6.86925292e-01 -8.52980137e-01 -9.16980505e-01
4.63769995e-02 9.29887950e-01 4.50346291e-01 2.45276734e-01
3.92222226e-01 8.43460500e-01 -2.07351637e+00 5.67562997e-01
4.23261970e-01 1.59053957e+00 3.23915273e-01 3.93009812e-01
-2.08659440e-01 5.08906007e-01 -3.72603238e-01 -2.33521178e-01
4.93980139e-01 6.85158223e-02 -6.59103215e-01 9.04912353e-01
8.62410665e-01 -2.12613240e-01 1.53362960e-01 9.58597839e-01
2.70672590e-01 -1.09089077e-01 1.11093688e+00 -1.35217738e+00
-3.52930278e-01 3.81625921e-01 -9.17850733e-01 -2.84127071e-02
1.90030903e-01 -1.43271178e-01 1.12457740e+00 -1.04074478e+00
6.47561014e-01 1.46855068e+00 3.83282304e-01 3.46500315e-02
-1.27694070e+00 -4.93909329e-01 7.91538000e-01 2.34478816e-01
-1.58926547e+00 -1.21308647e-01 1.00952756e+00 -9.48537648e-01
4.04088527e-01 -4.38673742e-04 6.53516531e-01 4.87785846e-01
1.85216367e-01 1.09777236e+00 1.19256496e+00 6.30423939e-03
1.34850144e-01 4.32786256e-01 4.81501818e-01 7.64150560e-01
5.59086382e-01 -3.52810472e-01 -5.84655225e-01 -2.55341321e-01
4.69007164e-01 2.85656601e-01 -3.05705029e-03 -1.14582777e+00
-9.95033324e-01 6.03612959e-01 6.11587942e-01 -1.93664767e-02
-6.35791004e-01 1.79394763e-02 2.44572535e-01 -2.90981621e-01
3.36602896e-01 1.26856528e-02 -6.08143687e-01 6.98458374e-01
-6.62346482e-01 3.95097315e-01 6.99348271e-01 1.42097163e+00
1.12366223e+00 -2.93860227e-01 -1.92131966e-01 8.26134503e-01
6.97015584e-01 2.43805259e-01 1.15897968e-01 -5.53921282e-01
5.27656853e-01 1.28024340e+00 -2.71191858e-02 -8.94547760e-01
-2.56531507e-01 -6.80851519e-01 -5.60156226e-01 6.87747821e-02
1.02006570e-01 3.67561579e-01 -1.05249834e+00 1.33408129e+00
8.33893657e-01 -2.50235021e-01 -1.34241492e-01 7.75690079e-01
9.51834440e-01 3.86714220e-01 5.86551204e-02 6.00758055e-03
1.81447697e+00 -8.71109247e-01 -2.27927178e-01 -1.88774213e-01
1.40052959e-01 -8.60853672e-01 3.89468223e-01 4.86421704e-01
-8.54936242e-01 -8.31503093e-01 -1.08942413e+00 3.85686681e-02
-5.65819561e-01 6.81483626e-01 6.36609852e-01 2.96220750e-01
-3.33981544e-01 2.54025877e-01 -6.70240164e-01 -7.67783895e-02
7.11224914e-01 6.30930543e-01 -3.74870300e-01 -6.25021309e-02
-3.38555247e-01 5.63563168e-01 6.43398643e-01 2.24594221e-01
-1.38865125e+00 -7.22985446e-01 -8.82464290e-01 -7.28467479e-02
5.78126013e-01 -7.21457839e-01 8.03591251e-01 -7.20432401e-01
-9.69333589e-01 8.62191319e-01 1.48762506e-03 -2.49443464e-02
2.29975715e-01 -4.01158839e-01 2.97041178e-01 9.31294709e-02
5.85952364e-02 6.92146957e-01 1.03550875e+00 -1.97201085e+00
-9.30989623e-01 -9.96908307e-01 2.53615864e-02 3.38429630e-01
7.75773749e-02 -2.88367301e-01 -5.62631428e-01 -2.99590200e-01
7.57405758e-01 -1.06046474e+00 7.22659677e-02 8.13716203e-02
-6.02946699e-01 -4.26805258e-01 1.08185709e+00 -5.86399615e-01
6.53056026e-01 -2.04238963e+00 6.19141877e-01 4.41315062e-02
2.85080820e-01 -3.81258994e-01 2.46462986e-01 3.44762355e-01
3.16187561e-01 -3.19132537e-01 -3.27244848e-02 -3.61862391e-01
7.61262774e-02 2.94537097e-01 -6.81271628e-02 7.49781370e-01
8.85053352e-02 6.05570972e-01 -4.67550695e-01 -7.96791494e-01
2.56872922e-01 3.76477927e-01 -5.89025557e-01 4.97627765e-01
-4.24469590e-01 3.33868086e-01 -7.88462698e-01 9.54731584e-01
9.60466564e-01 -1.07800260e-01 1.51102737e-01 -7.83872962e-01
9.04793218e-02 7.43359933e-03 -1.56819487e+00 1.51986182e+00
-4.12260920e-01 2.43824646e-02 1.22618996e-01 -1.10758054e+00
8.71476114e-01 1.08561516e-01 6.26290262e-01 1.09822266e-01
2.26170182e-01 1.25926077e-01 -1.58684164e-01 -3.92634749e-01
2.73364574e-01 6.44088760e-02 -2.93335348e-01 1.25855207e-01
5.50931096e-01 -3.95644873e-01 1.81249157e-01 1.82213318e-02
4.24193770e-01 4.07845914e-01 4.04683471e-01 -2.10186392e-01
7.81866014e-01 -1.23456299e-01 4.84835923e-01 5.93510330e-01
2.35671237e-01 4.88178313e-01 3.91303748e-01 -3.62612396e-01
-9.39977109e-01 -1.21579969e+00 -4.23847795e-01 1.14035046e+00
4.80583608e-01 -2.49613062e-01 -5.59204221e-01 -9.15826559e-01
4.33580965e-01 2.39930093e-01 -6.88246429e-01 1.52088359e-01
-4.05623525e-01 -6.64986134e-01 -1.51614010e-01 6.95716143e-01
5.97642720e-01 -6.97178662e-01 -5.07962048e-01 -6.62710220e-02
-3.77045497e-02 -1.28438616e+00 -2.63471425e-01 3.52814704e-01
-7.88673818e-01 -1.25481021e+00 -3.03428739e-01 -8.59027267e-01
8.18181276e-01 1.88068002e-01 1.08055246e+00 -3.10341418e-01
-6.11126125e-01 6.75484240e-01 -3.17076504e-01 -5.74943662e-01
2.45801383e-03 1.08236626e-01 -8.25043954e-03 5.80146015e-01
8.56408179e-02 -6.63514435e-01 -6.09524012e-01 5.69722772e-01
-6.83274329e-01 -5.33491708e-02 1.11750555e+00 5.23823738e-01
1.00338435e+00 8.81891698e-02 1.47847310e-01 -8.49171102e-01
-1.91406742e-01 -5.79450727e-01 -8.78578067e-01 3.83266777e-01
-3.64124626e-01 3.73328924e-01 3.15795451e-01 -5.04857302e-01
-9.71942008e-01 4.75482106e-01 5.20725667e-01 -5.87083101e-01
-5.20793259e-01 1.81877036e-02 -7.21663773e-01 4.04210342e-03
2.02418104e-01 3.81656647e-01 -3.02312493e-01 -8.81770313e-01
3.25526059e-01 6.53517485e-01 2.34522387e-01 -7.05065608e-01
8.93138170e-01 5.85240602e-01 2.33477399e-01 -7.97340631e-01
-1.11710775e+00 -8.92164409e-01 -1.21047425e+00 -2.50472277e-01
1.10561466e+00 -1.19043648e+00 -7.89186537e-01 4.07088280e-01
-1.19677556e+00 4.75418299e-01 -6.77650571e-02 3.99120003e-01
-6.97169483e-01 2.67894149e-01 -2.63042837e-01 -9.15082872e-01
-7.52580836e-02 -1.25098574e+00 1.86321342e+00 7.00941756e-02
3.48729014e-01 -5.13705611e-01 -2.17485905e-01 5.62571347e-01
-2.84111142e-01 2.94493020e-01 1.03677130e+00 -7.78391361e-01
-1.22571349e+00 -2.56392926e-01 -2.30182528e-01 3.41747791e-01
1.48773655e-01 -1.37137234e-01 -8.86698246e-01 -2.68823564e-01
1.14368752e-01 -4.43912804e-01 7.08648264e-01 2.59894192e-01
1.32487774e+00 -2.58373260e-01 -5.33156633e-01 5.96260428e-01
1.70502365e+00 2.57388428e-02 1.47885099e-01 -2.99495808e-03
9.57115829e-01 9.86259818e-01 6.38779104e-01 5.89476466e-01
3.56913507e-01 9.36089694e-01 6.46740019e-01 3.57993126e-01
-1.11551210e-01 -2.70555437e-01 1.58678949e-01 6.90801740e-01
-7.10692927e-02 1.04474993e-02 -6.33059144e-01 4.15260524e-01
-1.69190168e+00 -6.03159428e-01 -6.62065819e-02 1.95055842e+00
2.85031766e-01 1.52667165e-02 2.63504237e-01 -4.45691571e-02
8.30412567e-01 1.13929100e-01 -6.08369589e-01 8.50754827e-02
2.37501308e-01 -1.51783019e-01 4.54696536e-01 7.07496563e-03
-1.25978374e+00 6.14222944e-01 5.39186668e+00 7.35184848e-01
-6.87857807e-01 -7.81252608e-03 2.71929771e-01 1.47452101e-01
8.27012807e-02 2.99595427e-02 -1.41025937e+00 1.01049788e-01
5.88834248e-02 2.06061885e-01 2.22862840e-01 1.23171210e+00
-4.28369910e-01 -2.49357805e-01 -1.46962059e+00 9.52617645e-01
4.22430545e-01 -8.25761616e-01 5.13613582e-01 3.02109271e-01
7.91826129e-01 -2.67853081e-01 3.09918299e-02 1.77932635e-01
7.27373064e-02 -5.72751939e-01 1.12285888e+00 5.87482095e-01
2.56006807e-01 -7.05042243e-01 7.99667358e-01 5.70367813e-01
-1.72586071e+00 -5.31061947e-01 -5.51620841e-01 3.58489603e-01
-9.55057889e-02 3.00346702e-01 -7.63783693e-01 8.41158390e-01
7.16303885e-01 5.97874880e-01 -8.54251802e-01 9.45503771e-01
-1.51114851e-01 1.27772659e-01 -2.64230639e-01 3.50625552e-02
2.90590315e-03 -2.99401641e-01 5.19417405e-01 9.10815418e-01
-1.82507690e-02 -7.05990270e-02 6.03363574e-01 9.46993709e-01
-2.89645307e-02 5.45390919e-02 -4.14464384e-01 1.18669130e-01
2.65214145e-01 1.53972018e+00 -8.27733099e-01 -3.99056405e-01
-4.66298044e-01 5.59078336e-01 3.98351759e-01 5.11137657e-02
-8.21882129e-01 -4.34174016e-03 4.50794190e-01 2.80555010e-01
9.06534374e-01 -2.61396676e-01 -2.31481224e-01 -1.07869446e+00
2.28387743e-01 -6.89487994e-01 5.44076025e-01 -5.84596753e-01
-1.44444323e+00 5.36310315e-01 6.85565710e-01 -1.56139481e+00
5.92941903e-02 -9.15983021e-01 -1.05427086e-01 6.22282147e-01
-1.07940018e+00 -1.69918406e+00 -4.61053759e-01 4.12065238e-01
8.84329438e-01 -1.95799068e-01 4.15184945e-01 4.34788875e-02
-3.45479310e-01 1.26421809e-01 -1.21084511e-01 3.40712592e-02
1.64718822e-01 -1.26690698e+00 -3.77481550e-01 3.99283707e-01
4.19899344e-01 4.35571909e-01 3.45557898e-01 -6.16784453e-01
-1.66714716e+00 -1.09825277e+00 1.94483921e-01 -6.24414504e-01
3.35231572e-01 -8.56816649e-01 -5.36672950e-01 6.35591149e-01
-1.24331176e-01 1.38469532e-01 5.31557441e-01 1.21781677e-01
-4.64722723e-01 -3.32324058e-01 -1.00851083e+00 -1.73720568e-02
1.13647854e+00 -4.72598851e-01 -8.35503817e-01 3.91905040e-01
3.58796805e-01 -2.36242652e-01 -9.61358190e-01 6.06470764e-01
8.06190968e-01 -8.73831391e-01 1.07259834e+00 -5.59952199e-01
3.49305540e-01 -4.68264461e-01 -6.70016170e-01 -9.49661911e-01
-4.16938454e-01 2.71876365e-01 -1.89222232e-01 1.32140148e+00
1.16883494e-01 -3.82292777e-01 7.18629360e-01 9.47597474e-02
-2.26961955e-01 -9.76212204e-01 -7.49711514e-01 -6.54323578e-01
-2.08739161e-01 -1.19890660e-01 5.50397277e-01 2.29532078e-01
-7.24189103e-01 5.07503748e-01 5.11380546e-02 7.21197844e-01
1.07687080e+00 1.01304591e+00 9.64113057e-01 -1.52370799e+00
-4.74837840e-01 -2.16236085e-01 -9.28763449e-01 -1.07543480e+00
2.98056215e-01 -1.01067829e+00 1.47693828e-01 -1.57271838e+00
8.17990124e-01 -3.63995641e-01 -1.87897772e-01 1.87400207e-01
-1.30004650e-02 2.79481113e-01 2.79545695e-01 3.37868333e-01
-6.87277496e-01 6.04756057e-01 1.21416867e+00 -4.40875083e-01
-3.85126323e-02 1.10897154e-01 -5.12309253e-01 9.52623427e-01
1.96599975e-01 -5.47580481e-01 -3.14526170e-01 -2.14254498e-01
-2.75205433e-01 -1.43273577e-01 7.12123096e-01 -1.05435336e+00
9.35157910e-02 -2.94980425e-02 7.92254567e-01 -1.36748958e+00
5.56448817e-01 -1.13874292e+00 3.36882293e-01 2.62938321e-01
-1.08849429e-01 -2.49217540e-01 -5.79484273e-03 7.73411691e-01
-1.89641535e-01 -1.87682971e-01 8.76417518e-01 -3.64307731e-01
-5.91370940e-01 5.10218203e-01 3.74613069e-02 -1.35673404e-01
1.27196610e+00 -2.45199904e-01 -8.72085318e-02 1.81410730e-01
-9.61079419e-01 2.13507190e-01 3.07559043e-01 4.67540622e-01
4.89702106e-01 -1.28419161e+00 -5.67562461e-01 3.36181968e-01
4.94009942e-01 2.93415040e-01 2.44297847e-01 7.22141147e-01
-1.54330999e-01 3.86795819e-01 -3.16039920e-01 -1.21405637e+00
-1.28130472e+00 9.27400231e-01 1.98732600e-01 -1.27800256e-01
-2.94490606e-01 8.37663710e-01 8.26043487e-01 -5.01356661e-01
2.39227042e-01 -2.47179732e-01 -4.29333508e-01 4.50002730e-01
-1.17201112e-01 3.54331523e-01 -2.91113462e-02 -9.63356137e-01
-4.67499167e-01 1.15879476e+00 -1.58545837e-01 1.46027476e-01
1.49289620e+00 -1.36579379e-01 -2.59687513e-01 6.87016428e-01
1.30980659e+00 8.90794303e-03 -1.24662995e+00 -3.27232897e-01
-5.50784208e-02 -5.21600246e-01 -1.35782748e-01 -4.52286541e-01
-9.81128633e-01 8.09340000e-01 6.69513702e-01 5.45674823e-02
9.19176579e-01 6.89078212e-01 2.30935887e-01 2.72227079e-01
7.45727718e-01 -9.88288462e-01 2.22112596e-01 1.58072680e-01
1.15686822e+00 -9.66770828e-01 5.13433278e-01 -9.36127543e-01
-5.34474730e-01 1.02725947e+00 8.91368985e-01 -2.03404129e-01
8.06881189e-01 3.58683914e-02 -4.17657912e-01 -4.96100932e-01
-6.27940595e-01 -3.70079428e-01 8.68748128e-01 4.15422469e-01
1.59734234e-01 6.88433573e-02 4.00814451e-02 5.91558933e-01
-8.56827125e-02 -5.17453730e-01 -9.26755071e-02 8.84808302e-01
-5.33754408e-01 -1.06721306e+00 -5.37909687e-01 5.61849773e-01
-2.43820027e-01 4.93671685e-01 -5.76573074e-01 8.75573575e-01
8.05762172e-01 5.05604506e-01 -1.11595374e-02 -2.80317634e-01
5.19924104e-01 2.32081190e-01 8.96586061e-01 -9.17919934e-01
-4.24943149e-01 2.38320187e-01 -1.22939393e-01 -4.49659079e-01
-5.13619781e-01 -8.47657681e-01 -8.92080963e-01 4.55778003e-01
-7.07764447e-01 5.45030981e-02 8.24592710e-01 8.71956706e-01
2.11540535e-01 5.10550857e-01 8.43110383e-01 -1.40907919e+00
-6.73662901e-01 -9.34294283e-01 -8.80013585e-01 4.81737405e-01
2.19597757e-01 -1.33982742e+00 -4.09448504e-01 6.45769536e-02] | [7.535601615905762, -2.7117760181427] |
d74e37c7-dba1-4719-a28b-2a597ba1a52d | few-shot-image-classification-via-contrastive | 2008.09942 | null | https://arxiv.org/abs/2008.09942v1 | https://arxiv.org/pdf/2008.09942v1.pdf | Few-Shot Image Classification via Contrastive Self-Supervised Learning | Most previous few-shot learning algorithms are based on meta-training with fake few-shot tasks as training samples, where large labeled base classes are required. The trained model is also limited by the type of tasks. In this paper we propose a new paradigm of unsupervised few-shot learning to repair the deficiencies. We solve the few-shot tasks in two phases: meta-training a transferable feature extractor via contrastive self-supervised learning and training a classifier using graph aggregation, self-distillation and manifold augmentation. Once meta-trained, the model can be used in any type of tasks with a task-dependent classifier training. Our method achieves state of-the-art performance in a variety of established few-shot tasks on the standard few-shot visual classification datasets, with an 8- 28% increase compared to the available unsupervised few-shot learning methods. | ['Guizhong Liu', 'Jianyi Li'] | 2020-08-23 | null | null | null | null | ['unsupervised-few-shot-learning', 'unsupervised-few-shot-image-classification'] | ['computer-vision', 'computer-vision'] | [ 5.17350256e-01 2.22169593e-01 -5.69874406e-01 -1.75148442e-01
-6.66698456e-01 2.13193834e-01 8.27835917e-01 1.73403502e-01
-2.50655800e-01 6.03520691e-01 5.91383129e-02 1.19922534e-01
5.58469221e-02 -8.58991086e-01 -4.88033444e-01 -4.61135447e-01
1.09243758e-01 5.24532914e-01 6.09530568e-01 -5.35228312e-01
3.26595783e-01 2.25932524e-02 -1.98160160e+00 4.37794507e-01
7.92836785e-01 8.07373881e-01 -3.05238031e-02 7.33963847e-01
-2.57071227e-01 1.17838073e+00 -4.08888161e-01 -2.86749005e-01
1.33299530e-01 -7.86461174e-01 -6.18769407e-01 4.37955618e-01
4.88901109e-01 -2.65646756e-01 -3.71349663e-01 1.04916871e+00
4.56271857e-01 6.08955622e-01 9.39881921e-01 -1.59118843e+00
-6.89402759e-01 2.47921601e-01 -4.81037110e-01 3.13562959e-01
-3.62501740e-02 2.18493506e-01 9.25218284e-01 -1.21020210e+00
9.18688536e-01 8.87210548e-01 7.42404878e-01 8.86205077e-01
-1.15946293e+00 -4.44794923e-01 -3.66466671e-01 5.99163890e-01
-1.10392714e+00 -6.39348805e-01 9.92993593e-01 -6.60025597e-01
1.16397953e+00 -1.67423949e-01 5.33199549e-01 1.29542375e+00
-3.40580754e-02 5.23883581e-01 1.18184805e+00 -8.28840733e-01
7.38664329e-01 1.42620772e-01 5.07548571e-01 1.05706966e+00
3.45280498e-01 1.90663636e-02 -6.93551958e-01 -1.41272545e-01
2.24549443e-01 4.65961933e-01 1.71928823e-01 -7.85675466e-01
-7.21232057e-01 1.21562660e+00 2.97614872e-01 5.62454820e-01
-1.27980456e-01 9.29501727e-02 7.00317383e-01 6.02843583e-01
1.05610633e+00 6.45725250e-01 -2.21699938e-01 1.17145628e-02
-1.05497766e+00 -9.35575292e-02 7.79476583e-01 9.17390704e-01
1.18478286e+00 3.83682936e-01 -4.27169621e-01 9.62166548e-01
-1.36558702e-02 -1.06440280e-02 9.57365394e-01 -5.75754166e-01
1.83527306e-01 5.30570686e-01 -1.50993004e-01 -5.22997320e-01
-2.06140444e-01 -1.44603446e-01 -7.11799562e-01 4.33951259e-01
2.35451877e-01 -8.41775760e-02 -1.38939857e+00 1.27758598e+00
2.66827494e-01 7.55030513e-01 1.13408500e-02 5.20728409e-01
9.84634578e-01 4.50348824e-01 1.06034473e-01 -4.76959646e-01
1.29766273e+00 -1.49544406e+00 -9.37830389e-01 -4.80354279e-01
1.17704785e+00 -3.44794452e-01 1.30830884e+00 1.89567339e-02
-4.32084441e-01 -5.88550925e-01 -1.45548213e+00 -3.71596143e-02
-9.42670286e-01 -3.66109759e-01 5.24066925e-01 8.40112269e-01
-5.95463932e-01 1.13519573e+00 -4.97648329e-01 -5.60667098e-01
8.53648603e-01 5.65374568e-02 -4.50884610e-01 -3.86845708e-01
-1.06354785e+00 1.04643202e+00 5.67954957e-01 -7.10009038e-01
-1.09660387e+00 -7.59272397e-01 -1.20942044e+00 8.17471892e-02
7.40843296e-01 -5.53331077e-01 1.01280200e+00 -6.19424760e-01
-1.41106486e+00 1.05582809e+00 2.25393862e-01 -5.46541929e-01
4.23479646e-01 -2.26318054e-02 -3.44823062e-01 9.79646668e-02
2.46664926e-01 2.48668477e-01 1.61931491e+00 -1.11928809e+00
-3.11071485e-01 -4.91807520e-01 -2.32409492e-01 8.82479101e-02
-7.09613383e-01 -1.67385474e-01 -1.11546321e-02 -5.65731108e-01
-3.09662342e-01 -7.17641652e-01 -1.69680566e-01 -5.08407280e-02
-1.12798594e-01 -2.10100800e-01 1.38232863e+00 -1.25773609e-01
8.41435611e-01 -2.09520411e+00 8.79649445e-02 -4.65358317e-01
5.85290611e-01 7.26572394e-01 -3.55526298e-01 5.11821806e-01
-1.10413671e-01 -4.76899184e-02 -3.90313864e-01 -8.03400755e-01
-1.24557614e-01 2.92728752e-01 -1.86455026e-01 6.40645742e-01
3.04487437e-01 1.14732897e+00 -1.25327063e+00 -6.87760651e-01
6.95506454e-01 1.18049428e-01 -2.02250645e-01 3.69625926e-01
-1.27512887e-01 -7.65186176e-03 -9.34211910e-02 8.24490666e-01
3.36951584e-01 -3.24099422e-01 -1.29246786e-01 1.22275621e-01
2.05565333e-01 -2.65438408e-01 -8.46949637e-01 1.99823129e+00
-3.63950163e-01 7.58834422e-01 -5.26795864e-01 -1.42813826e+00
8.09744954e-01 2.80715972e-01 3.94289881e-01 -3.53115469e-01
4.05575454e-01 1.12848759e-01 -1.43856362e-01 -6.03621721e-01
2.55201042e-01 -7.09536314e-01 -5.66251995e-03 6.47047400e-01
1.02376401e+00 -3.41300279e-01 3.63687515e-01 2.50381172e-01
1.36735737e+00 8.87974165e-03 7.97267973e-01 -2.63334382e-02
1.53612392e-03 2.08468556e-01 2.00254366e-01 9.19310987e-01
-5.45776546e-01 6.72687829e-01 2.59377390e-01 -4.97422129e-01
-1.23167765e+00 -7.32942283e-01 1.53224552e-02 1.60157287e+00
-2.61919834e-02 -6.38214231e-01 -6.52259767e-01 -1.06946468e+00
-7.39644617e-02 9.97713149e-01 -1.05067074e+00 -7.56619751e-01
-5.84173985e-02 -7.03172326e-01 2.48014718e-01 2.29486704e-01
3.89161527e-01 -1.22628736e+00 -7.09499657e-01 7.98664764e-02
4.05106515e-01 -1.13574779e+00 -1.53369993e-01 4.71400768e-01
-1.00413764e+00 -1.27349317e+00 -7.66120374e-01 -7.87764072e-01
6.10172272e-01 8.01862478e-01 6.79012120e-01 9.18562934e-02
-7.74155796e-01 2.54813522e-01 -7.82622755e-01 -5.54712117e-01
-3.88208389e-01 1.90970004e-02 1.14071891e-01 1.56705454e-01
4.75706846e-01 -7.19587862e-01 -1.43089592e-01 -9.34149548e-02
-7.98908412e-01 -8.66454318e-02 4.62668389e-01 1.42466795e+00
2.89818972e-01 -1.58662528e-01 7.91789651e-01 -1.48629642e+00
5.45567691e-01 -6.68789744e-01 -9.55647528e-02 2.76246995e-01
-9.76583481e-01 7.68829882e-02 6.60729825e-01 -6.58124506e-01
-1.01708972e+00 1.61593497e-01 4.55446303e-01 -1.00128508e+00
-2.04859689e-01 2.77391613e-01 2.14560881e-01 -4.36301172e-01
1.31603515e+00 1.81671873e-01 2.24984288e-01 -4.05113339e-01
8.61579776e-01 7.11884797e-01 8.86167660e-02 7.17671663e-02
9.24222291e-01 5.23631513e-01 8.86582956e-02 -1.27986205e+00
-1.29891169e+00 -9.56216455e-01 -1.17389512e+00 -2.44592637e-01
7.85719573e-01 -7.33755946e-01 2.27471158e-01 4.61049050e-01
-7.62053609e-01 -4.39081252e-01 -9.27022696e-01 3.29116493e-01
-8.85470688e-01 4.38859224e-01 -4.80044395e-01 -7.60415554e-01
-3.79705518e-01 -6.29728079e-01 9.78013277e-01 -7.94624537e-03
4.01977971e-02 -1.05691218e+00 3.90503258e-01 4.07373577e-01
2.31814206e-01 2.77529538e-01 7.94026494e-01 -1.02953362e+00
-6.79847319e-03 -5.39226413e-01 -1.69163808e-01 2.75554985e-01
2.86334306e-01 -4.82743174e-01 -1.38280702e+00 -3.76713753e-01
1.31669536e-01 -1.03883636e+00 1.31314349e+00 9.29363221e-02
7.61714578e-01 -1.61828414e-01 -2.37548187e-01 5.96248925e-01
1.54697239e+00 -2.12945908e-01 5.42171180e-01 9.57099870e-02
8.68534207e-01 4.68994230e-01 7.64737785e-01 4.41078454e-01
-1.68929815e-01 3.37649226e-01 3.28053892e-01 2.20767841e-01
-4.48169380e-01 -3.16171676e-01 2.27471571e-02 7.41753757e-01
-1.15210190e-01 1.53487444e-01 -8.89198363e-01 6.69205368e-01
-2.29792500e+00 -1.39123464e+00 1.69622779e-01 1.96979403e+00
5.29171705e-01 2.05725446e-01 1.59987330e-01 3.56017023e-01
7.69204140e-01 6.26520932e-01 -5.34711361e-01 -1.52706742e-01
1.60072640e-01 3.15211922e-01 3.11659575e-01 7.66163841e-02
-1.34575856e+00 1.40511775e+00 6.23083210e+00 9.92075741e-01
-8.23292911e-01 8.22586298e-01 3.23711932e-01 -1.35246217e-01
3.28041345e-01 2.36043066e-01 -4.40060824e-01 1.29624262e-01
9.11836922e-01 -2.80621439e-01 3.53185326e-01 1.22648597e+00
-3.28357249e-01 -6.29970133e-02 -9.77522552e-01 1.14476871e+00
7.05634713e-01 -1.57662761e+00 -9.46886390e-02 -1.06409945e-01
1.12834609e+00 8.60887319e-02 -1.42725006e-01 8.99361908e-01
1.39812082e-01 -8.30933988e-01 1.53914377e-01 2.66507953e-01
1.02057230e+00 -4.82032895e-01 5.70008099e-01 7.43208051e-01
-9.97169733e-01 -3.20907176e-01 -6.81669772e-01 -4.27037597e-01
-1.39501812e-02 3.20852280e-01 -8.89890611e-01 3.23790312e-01
3.18253249e-01 9.62276161e-01 -7.33543277e-01 1.02073848e+00
-1.57145858e-01 5.59961915e-01 2.67086148e-01 -1.29218966e-01
1.48450494e-01 2.22662073e-02 5.94228506e-01 9.34478402e-01
9.97455567e-02 1.19924001e-01 3.39203805e-01 5.53118467e-01
-2.11657152e-01 2.54077584e-01 -1.27651572e+00 -3.43417853e-01
2.17947736e-01 1.41681588e+00 -8.39609742e-01 -8.60476673e-01
-6.84125125e-01 1.19304121e+00 7.38450229e-01 6.13611639e-02
-4.81570572e-01 -7.67721236e-01 1.44053265e-01 2.40680709e-01
3.34749639e-01 1.59847806e-03 -1.71137974e-01 -1.54931927e+00
-4.26568091e-01 -4.70774263e-01 5.66626012e-01 -7.36844182e-01
-1.50992894e+00 2.61365563e-01 1.02384560e-01 -1.40490639e+00
-4.07697022e-01 -5.83609819e-01 -9.98915255e-01 2.05896020e-01
-1.57087958e+00 -1.45314670e+00 -5.00911057e-01 6.55912817e-01
9.39236581e-01 -6.92716777e-01 1.11148298e+00 2.98422519e-02
-4.09218222e-01 3.44721347e-01 -7.00790435e-04 3.43528166e-02
9.45307732e-01 -1.20670974e+00 2.68059522e-01 7.53226101e-01
4.45812106e-01 1.66692868e-01 7.34110415e-01 -8.21227968e-01
-1.36408865e+00 -1.23392951e+00 6.28922522e-01 -3.04640412e-01
9.84297276e-01 -5.68956733e-01 -9.61791575e-01 6.68989658e-01
5.92469946e-02 8.31116378e-01 8.89994502e-01 1.81337878e-01
-6.90001905e-01 1.90055549e-01 -1.20788383e+00 4.20865864e-01
1.21954310e+00 -8.77812386e-01 -1.07605064e+00 8.35808158e-01
8.20169270e-01 1.22927651e-01 -4.39539433e-01 6.84299767e-02
1.65501639e-01 -8.11717451e-01 7.47322440e-01 -1.28109419e+00
4.60628688e-01 1.86283067e-01 -2.01084837e-01 -1.73350108e+00
-4.31764573e-01 -6.54139161e-01 -6.49021089e-01 7.71743178e-01
9.75374803e-02 -3.67635548e-01 1.02929103e+00 2.21511230e-01
-2.87608773e-01 -3.96537364e-01 -9.56562519e-01 -1.06702161e+00
-3.22470039e-01 -2.21450210e-01 -8.85083601e-02 1.42205489e+00
4.58785594e-01 9.62135792e-01 -9.89807487e-01 -5.56401432e-01
1.15053463e+00 3.99951153e-02 8.13306272e-01 -1.64546692e+00
-2.13461041e-01 -1.45843057e-02 -7.25538790e-01 4.37229648e-02
5.25646210e-01 -1.22751379e+00 8.80529638e-03 -1.48644269e+00
5.09772420e-01 2.15947986e-01 -3.13318521e-01 7.98827350e-01
-1.39850006e-01 4.28928494e-01 2.48353973e-01 3.80248427e-01
-9.93927360e-01 8.25301170e-01 1.07224035e+00 -3.40372652e-01
-2.34755471e-01 -8.49688426e-02 -3.07695597e-01 8.00978184e-01
6.72875106e-01 -7.60441363e-01 -6.28614187e-01 3.17795753e-01
-4.13149536e-01 -7.28169903e-02 3.45914453e-01 -1.19183242e+00
2.45287210e-01 -1.97478816e-01 2.17332974e-01 -1.40512437e-01
6.86192811e-01 -5.14824152e-01 -5.57254970e-01 7.13669360e-01
-7.70657212e-02 -5.98036885e-01 -1.86456352e-01 9.70605791e-01
-9.23573598e-02 -6.48511648e-01 1.08647406e+00 -4.16495204e-01
-1.27436364e+00 4.65116024e-01 -1.99371979e-01 4.02032763e-01
1.49414742e+00 -3.52309763e-01 -6.44927204e-01 -2.18658909e-01
-1.02977276e+00 -2.91021734e-01 4.00377333e-01 4.23484236e-01
8.79969418e-01 -1.45895994e+00 -4.07010645e-01 1.09862745e-01
8.26976657e-01 -5.70656717e-01 4.62877840e-01 6.09137535e-01
-4.27355245e-02 -2.05190200e-02 -7.01856375e-01 -2.95122206e-01
-1.18525922e+00 1.10094655e+00 -3.80951837e-02 -1.60258755e-01
-9.80206788e-01 5.04246891e-01 -3.34498167e-01 -1.93787560e-01
-6.22908622e-02 4.01257396e-01 -3.31228703e-01 5.51758766e-01
6.01927459e-01 6.58276856e-01 -6.85138395e-03 -4.97962177e-01
-3.53097506e-02 2.29013383e-01 -1.30120322e-01 1.71266403e-02
1.59578872e+00 2.77150124e-01 2.85775006e-01 1.10994434e+00
1.37491345e+00 -6.78967416e-01 -1.11013031e+00 -4.93380427e-01
7.42989257e-02 -6.43548608e-01 2.68419415e-01 -2.50862569e-01
-5.93953848e-01 1.15039372e+00 6.21192396e-01 2.21951276e-01
6.19286060e-01 1.11979246e-01 5.75055182e-01 7.04517066e-01
4.98631954e-01 -1.44136190e+00 7.42312014e-01 5.62327564e-01
5.33468246e-01 -1.81735456e+00 1.66252300e-01 -3.85607600e-01
-7.92921960e-01 9.21723604e-01 6.54299378e-01 -5.41094422e-01
1.04636419e+00 -5.37990555e-02 -2.67491221e-01 -5.31329513e-01
-8.53947163e-01 -6.32872641e-01 4.46356237e-01 8.51485550e-01
6.21876866e-02 -8.92234817e-02 -1.84102923e-01 4.83214617e-01
3.32513452e-01 1.68231726e-01 6.48366809e-01 1.25127029e+00
-9.47992086e-01 -7.75241077e-01 1.77619472e-01 8.46982539e-01
2.72856951e-02 -1.48820847e-01 -4.65613723e-01 7.41466582e-01
-7.12532178e-02 9.12360311e-01 -1.18360527e-01 -5.54429352e-01
1.67635292e-01 6.29188180e-01 5.33842385e-01 -1.48676407e+00
-2.38727331e-01 -2.16996476e-01 9.15958062e-02 -4.68019545e-01
-3.30259472e-01 -3.08899909e-01 -9.06037629e-01 4.14652973e-02
-6.31388962e-01 -1.57687798e-01 4.26264584e-01 1.23528290e+00
2.98344284e-01 4.63601321e-01 5.06349921e-01 -1.11940455e+00
-9.41464067e-01 -1.30925512e+00 -9.57817972e-01 7.41207004e-01
2.27093995e-01 -1.11867404e+00 -5.85280776e-01 -3.29871327e-02] | [9.979411125183105, 2.9641973972320557] |
90b7672f-7978-4ce4-ae63-ff896028c916 | predicting-indian-stock-market-using-the | 1911.06193 | null | https://arxiv.org/abs/1911.06193v1 | https://arxiv.org/pdf/1911.06193v1.pdf | Predicting Indian stock market using the psycho-linguistic features of financial news | Financial forecasting using news articles is an emerging field. In this paper, we proposed hybrid intelligent models for stock market prediction using the psycholinguistic variables (LIWC and TAALES) extracted from news articles as predictor variables. For prediction purpose, we employed various intelligent techniques such as Multilayer Perceptron (MLP), Group Method of Data Handling (GMDH), General Regression Neural Network (GRNN), Random Forest (RF), Quantile Regression Random Forest (QRRF), Classification and regression tree (CART) and Support Vector Regression (SVR). We experimented on the data of 12 companies stocks, which are listed in the Bombay Stock Exchange (BSE). We employed chi-squared and maximum relevance and minimum redundancy (MRMR) feature selection techniques on the psycho-linguistic features obtained from the new articles etc. After extensive experimentation, using the Diebold-Mariano test, we conclude that GMDH and GRNN are statistically the best techniques in that order with respect to the MAPE and NRMSE values. | ['Vadlamani Ravi', 'Rishabh Miglani', 'B. Shravan Kumar'] | 2019-11-07 | null | null | null | null | ['stock-market-prediction'] | ['time-series'] | [-5.80189347e-01 -1.09675832e-01 -2.56011367e-01 -6.12295389e-01
8.14479813e-02 -3.49152833e-01 6.57777190e-01 2.60044068e-01
-6.28467500e-01 1.12269139e+00 4.79577452e-01 -6.32454515e-01
-4.62227672e-01 -1.04344440e+00 -1.52814671e-01 -3.65044296e-01
-3.02093118e-01 3.72606188e-01 -3.54243629e-02 -5.74707925e-01
1.31971455e+00 6.72180176e-01 -1.91265213e+00 2.39163756e-01
7.62758732e-01 1.16524255e+00 -9.17217601e-03 5.35144925e-01
-2.88994968e-01 1.59896564e+00 -4.03802484e-01 -5.37917554e-01
2.12041557e-01 -2.12778866e-01 -5.59884787e-01 -4.09570247e-01
-4.28225458e-01 -3.27506959e-02 2.22641572e-01 7.07665741e-01
2.23560989e-01 4.40159887e-01 1.08065140e+00 -9.66029406e-01
-9.34883475e-01 9.01747048e-01 -6.83127642e-01 6.16597354e-01
2.48448685e-01 -6.79954469e-01 8.44625473e-01 -1.09394729e+00
5.28977394e-01 1.14319921e+00 5.79720497e-01 9.21704173e-02
-8.85576427e-01 -9.08293962e-01 -1.89540029e-01 4.97396916e-01
-9.31415498e-01 -1.97145134e-01 6.15246773e-01 -6.79641128e-01
1.18190885e+00 2.66362250e-01 5.30064225e-01 4.73978460e-01
7.40749657e-01 2.16205493e-01 1.64896417e+00 -7.85047412e-01
3.43723893e-01 9.32534277e-01 8.59066069e-01 4.12717700e-01
4.53946590e-01 3.42696279e-01 -7.36174643e-01 -2.83165991e-01
4.95741487e-01 -9.91014019e-02 2.87342906e-01 5.73882759e-01
-1.03027797e+00 1.39084244e+00 2.27622259e-02 6.78713620e-01
-9.00390029e-01 -5.54978967e-01 2.40703717e-01 8.03250551e-01
7.17845976e-01 5.97255111e-01 -1.06688237e+00 1.33605018e-01
-8.37351799e-01 3.66254568e-01 1.27524400e+00 4.98865694e-01
3.84633631e-01 2.27228910e-01 3.40594471e-01 5.02660096e-01
4.58581388e-01 5.25768042e-01 1.19691396e+00 -2.71523625e-01
1.85388923e-01 4.91381556e-01 1.28436863e-01 -1.45976269e+00
-7.95928180e-01 -4.53947753e-01 -5.85435390e-01 5.28517962e-01
1.39273345e-01 -6.23542905e-01 -2.62504697e-01 9.81437147e-01
4.92394939e-02 -2.32869193e-01 4.59296525e-01 4.57747459e-01
1.00412333e+00 7.78881788e-01 2.23638102e-01 -7.56074309e-01
1.26127434e+00 -5.72082937e-01 -8.16702545e-01 2.53713220e-01
4.15241778e-01 -6.91645741e-01 3.11334461e-01 9.45416689e-01
-8.48209083e-01 -6.97780311e-01 -8.96550775e-01 4.47172701e-01
-7.34462023e-01 5.86595796e-02 7.14168727e-01 7.78095186e-01
-7.46942997e-01 7.67209291e-01 -3.38103145e-01 5.70357293e-02
-3.77803773e-01 4.75469381e-01 -2.72941649e-01 7.89529920e-01
-1.10682106e+00 1.38496649e+00 5.79917848e-01 -5.76239526e-02
2.59243608e-01 -7.58912340e-02 -3.49067956e-01 -1.37732029e-01
-5.94758540e-02 -2.08676890e-01 7.31466532e-01 -1.14198387e+00
-1.58185983e+00 4.73859370e-01 2.28435770e-01 -7.63354480e-01
8.94115791e-02 -9.87433568e-02 -8.12774301e-01 -1.52501717e-01
-3.24669182e-01 1.54280603e-01 6.08864129e-01 -6.00903690e-01
-8.45833898e-01 -6.15224063e-01 -5.84324360e-01 -4.99737784e-02
-5.05520329e-02 6.72994912e-01 1.07829165e+00 -9.78474081e-01
4.66813624e-01 -5.38305223e-01 -1.97135597e-01 -1.21409440e+00
-1.87730893e-01 -3.50834876e-01 5.34709468e-02 -1.27959824e+00
1.20011330e+00 -1.89327514e+00 -6.50354326e-02 7.92256117e-01
-2.65274942e-01 -3.00633132e-01 6.21752143e-01 4.75613683e-01
-3.13406438e-01 2.08358243e-01 3.26930225e-01 3.67709070e-01
-5.84468357e-02 -5.99851459e-02 -3.95098299e-01 1.83231652e-01
-2.26546437e-01 4.02525693e-01 -1.35793924e-01 -6.33853495e-01
4.66881879e-02 -2.86137150e-03 -1.74633205e-01 -4.95869443e-02
5.12665696e-02 1.17023602e-01 -5.22326291e-01 6.89278424e-01
4.86268789e-01 3.50463986e-01 -1.81977570e-01 -7.95113146e-02
-6.66467130e-01 1.97915062e-01 -1.23857999e+00 3.83133739e-01
-1.56841248e-01 4.75878894e-01 -6.24722064e-01 -8.99736702e-01
1.87920165e+00 3.92581105e-01 1.67168513e-01 -6.92168474e-01
4.07651335e-01 4.95567769e-01 2.62149535e-02 -7.16367066e-01
5.84318936e-01 -4.95722026e-01 1.68762624e-01 3.73730898e-01
-1.69127621e-02 6.52281106e-01 1.78261817e-01 -5.76888502e-01
4.31309432e-01 9.53322947e-02 9.33788061e-01 -4.64998513e-01
7.58934259e-01 1.48840368e-01 4.30247933e-01 3.71699840e-01
6.38637170e-02 4.53535430e-02 7.43599594e-01 -6.03613079e-01
-7.52372980e-01 -4.20968294e-01 -3.51089299e-01 1.13610423e+00
-5.31719148e-01 3.31646562e-01 -4.28827405e-01 -3.83293331e-01
1.21605679e-01 1.54869783e+00 -4.94569033e-01 3.92836362e-01
-9.88301784e-02 -1.24449956e+00 1.92231790e-03 1.35845631e-01
2.65538275e-01 -1.55228209e+00 -7.33016253e-01 3.95041972e-01
5.74394763e-01 -5.31546235e-01 6.98154032e-01 4.44059461e-01
-1.20651054e+00 -6.86887264e-01 -4.26186591e-01 -5.16061485e-01
1.70969516e-01 -4.25544262e-01 9.95781302e-01 -1.37097016e-01
3.00923944e-01 -3.03251058e-01 -8.24571908e-01 -8.57068062e-01
-3.25063974e-01 2.37436667e-02 7.32594579e-02 -5.83852828e-02
7.50854075e-01 -7.51539409e-01 3.31290215e-02 5.62232323e-02
-2.97355115e-01 -1.68125063e-01 7.58216798e-01 6.55788124e-01
1.61083221e-01 4.60852414e-01 1.15540266e+00 -1.09436214e+00
8.55112195e-01 -8.83881927e-01 -8.89424086e-01 1.22150533e-01
-1.25183368e+00 1.17833912e-01 3.20943475e-01 -3.09537441e-01
-1.15898991e+00 -3.57524544e-01 -1.39984816e-01 4.67959970e-01
-3.60549420e-01 1.19209409e+00 4.65697855e-01 -1.95822939e-01
7.49406874e-01 2.07189247e-01 -5.60919084e-02 -6.75514340e-01
-2.19533235e-01 9.26076710e-01 -3.80574614e-02 4.63697240e-02
3.84224802e-01 -3.27633768e-01 8.67406428e-02 -7.18558550e-01
-5.91758668e-01 9.23970640e-02 -6.20415807e-01 -3.33345711e-01
8.93966079e-01 -5.20280898e-01 -1.03145432e+00 3.06781948e-01
-6.70532882e-01 5.22081673e-01 2.73385465e-01 1.28570354e+00
-2.60108590e-01 -3.07099581e-01 -6.46078646e-01 -1.69211257e+00
-5.15720427e-01 -4.78958517e-01 -1.24969244e-01 4.05087501e-01
-2.75334179e-01 -8.32691729e-01 1.05105333e-01 2.73062617e-01
3.75706404e-01 6.45079374e-01 1.18422866e+00 -1.73574102e+00
7.11703002e-02 -4.28179651e-01 8.23980123e-02 4.29694265e-01
-3.30050737e-01 3.01850885e-01 -6.35538757e-01 4.61483449e-01
5.23997486e-01 -5.97481765e-02 6.93315923e-01 5.30412138e-01
2.73457080e-01 -4.74132597e-01 2.98410952e-01 1.35791764e-01
1.53059936e+00 9.89053071e-01 4.15570945e-01 1.01542878e+00
1.88580900e-01 8.94393504e-01 7.52494574e-01 7.58821130e-01
4.71957684e-01 1.45743915e-03 -2.27113053e-01 4.73616898e-01
9.41758156e-01 2.05569968e-01 6.34076416e-01 9.19505894e-01
-5.86046696e-01 1.02813795e-01 -8.96500647e-01 -2.05458671e-01
-1.72208643e+00 -1.20121932e+00 -4.32067275e-01 2.03999662e+00
3.86034042e-01 6.08815551e-01 5.01725852e-01 6.28821015e-01
5.38266659e-01 -2.61611670e-01 -3.20666544e-02 -8.90529692e-01
-3.51605088e-01 1.47263557e-01 8.18231583e-01 2.59088039e-01
-1.02153826e+00 4.94976729e-01 5.33246374e+00 4.15366441e-01
-9.06003833e-01 -1.16160385e-01 7.64791369e-01 2.55887657e-01
-1.04922265e-01 -1.86141029e-01 -1.00136542e+00 5.01548231e-01
1.41785455e+00 -3.56849462e-01 3.81928205e-01 1.03544497e+00
1.27557144e-01 -4.36777979e-01 -3.89776349e-01 5.25392890e-01
5.39637171e-02 -1.11499166e+00 -2.33365849e-01 4.56435010e-02
4.25688922e-01 -8.72435421e-02 -1.43844038e-01 3.76150668e-01
9.46545899e-02 -8.00359428e-01 8.57512236e-01 1.23264718e+00
-3.31435204e-01 -1.30710745e+00 1.33608401e+00 6.93377376e-01
-6.04049146e-01 -2.75834441e-01 -5.68403065e-01 -4.66210872e-01
-2.36995175e-01 6.38149679e-01 -6.00765526e-01 6.74232364e-01
4.62677598e-01 4.53901142e-01 -6.48122072e-01 7.45059550e-01
3.27277631e-01 7.56377161e-01 -1.02471888e-01 -8.37317884e-01
7.97012672e-02 -4.89363611e-01 2.49047011e-01 8.51574242e-01
4.19800311e-01 3.51805359e-01 -4.00757641e-01 4.42001998e-01
7.16481686e-01 9.42422390e-01 -2.07997784e-01 9.39553380e-02
1.52998939e-01 8.51839781e-01 -1.11233103e+00 -3.50435376e-01
-5.54683745e-01 8.53559375e-02 -8.52079093e-02 2.76019163e-02
-3.39225084e-01 -6.12675130e-01 -8.15180987e-02 -1.50449648e-01
3.18567693e-01 -1.78685747e-02 -8.24772716e-01 -9.64910150e-01
-3.72602910e-01 -7.43923903e-01 4.35236275e-01 -7.02914298e-01
-1.59514463e+00 1.02326071e+00 2.29138911e-01 -9.98867631e-01
-6.21433020e-01 -9.24419105e-01 -2.53465503e-01 1.22900736e+00
-1.06836069e+00 -5.60492218e-01 3.77633691e-01 3.86450708e-01
2.58632064e-01 -1.14165723e+00 9.37517822e-01 5.09908982e-02
-4.20843214e-01 -5.08171096e-02 2.99892098e-01 7.68230259e-02
2.47189805e-01 -1.29708910e+00 1.20883055e-01 3.59696090e-01
7.65696913e-02 2.54030079e-01 9.94330943e-01 -1.04921722e+00
-7.10626960e-01 -2.72285759e-01 1.46723485e+00 -6.72828127e-03
7.91589200e-01 2.59317964e-01 -8.14038694e-01 5.02861261e-01
2.07820058e-01 -7.66314387e-01 9.12385345e-01 1.03719041e-01
3.02977469e-02 -1.22481681e-01 -1.34013987e+00 2.73417115e-01
-1.34342328e-01 8.99542421e-02 -1.09729421e+00 2.42761016e-01
4.78440702e-01 2.28024364e-01 -1.19690561e+00 1.13174133e-01
8.01382780e-01 -1.56059086e+00 6.94793403e-01 -6.12407625e-01
4.28422868e-01 -1.86035018e-02 -1.42534688e-01 -9.48447466e-01
-5.01891673e-01 -5.69387600e-02 2.28390232e-01 1.36932039e+00
1.05999792e+00 -1.06585813e+00 5.10313332e-01 9.03608322e-01
4.32234257e-01 -7.72513986e-01 -6.37843251e-01 -3.96737367e-01
-1.17407091e-01 -2.55660385e-01 6.14621282e-01 1.14005852e+00
1.48069620e-01 4.07211244e-01 -4.86769646e-01 -2.17619717e-01
4.13410068e-01 6.66573271e-02 2.21057460e-01 -1.67557299e+00
-3.55342865e-01 -4.49316114e-01 -5.36713898e-01 1.58250913e-01
1.78577974e-01 -4.21628028e-01 -5.50849915e-01 -1.13242292e+00
-3.77076685e-01 -1.56576917e-01 -6.04749680e-01 -4.09982447e-03
1.82911500e-01 -3.10673088e-01 2.61949211e-01 3.08009207e-01
2.78753668e-01 1.59941941e-01 7.88071215e-01 5.90372741e-01
-5.44851303e-01 7.01738238e-01 -6.82771325e-01 9.28678513e-01
9.94822621e-01 -7.47733533e-01 -1.53449669e-01 4.80111808e-01
6.80054605e-01 6.01465046e-01 -2.00538889e-01 -7.54592657e-01
1.95146248e-01 -4.08037454e-01 8.72181773e-01 -9.55786884e-01
-6.31892160e-02 -6.45385385e-01 4.14237589e-01 7.46889591e-01
-6.04969919e-01 8.19347978e-01 -1.90348282e-01 1.21270485e-01
-4.01074797e-01 -8.93534482e-01 3.52659523e-01 -2.26016670e-01
-5.72953165e-01 -4.52562749e-01 -7.27898180e-01 -5.45043528e-01
1.05326295e+00 -1.98088989e-01 -1.11755043e-01 -3.29356521e-01
-7.30085373e-01 -2.12550953e-01 -2.45118648e-01 7.00977743e-02
5.58732808e-01 -8.60618353e-01 -9.91708755e-01 1.98752671e-01
-4.23579216e-01 -9.01717782e-01 9.49052395e-04 9.35216367e-01
-9.71814036e-01 4.39664871e-01 -7.42281616e-01 5.64854205e-01
-1.11202741e+00 7.00134993e-01 2.25269664e-02 -1.17604755e-01
-2.16062054e-01 7.85539925e-01 -6.86967671e-01 -8.36510286e-02
1.03092514e-01 -2.67175883e-01 -1.58872771e+00 7.43083775e-01
5.11189997e-01 9.12765086e-01 5.05031422e-02 -8.05535138e-01
-1.76461294e-01 4.87617999e-01 -7.68908300e-03 -5.96166372e-01
1.63805795e+00 7.62390867e-02 -4.48985189e-01 9.29033697e-01
6.46836579e-01 1.41368017e-01 -1.39436737e-01 -9.66726895e-03
9.63419139e-01 -1.76503122e-01 3.60449851e-01 -9.30986047e-01
-6.45640790e-01 1.12810954e-01 3.77842158e-01 9.48620915e-01
1.20264244e+00 -6.21022880e-01 2.05582276e-01 7.62998164e-01
1.92349836e-01 -1.36399961e+00 -9.94202077e-01 3.75702411e-01
9.31983829e-01 -1.04157019e+00 3.71099353e-01 -7.43179321e-02
-1.25991535e+00 1.68912208e+00 8.62443373e-02 -7.25499094e-01
1.56436515e+00 1.98774397e-01 2.19082329e-02 6.96196631e-02
-1.17654240e+00 -3.29459645e-02 3.07962388e-01 1.32685184e-01
8.93396795e-01 1.84960902e-01 -9.84731495e-01 1.25972342e+00
-8.84184361e-01 2.10583210e-01 6.16187096e-01 9.32437420e-01
-8.22670341e-01 -7.26484656e-01 -7.33909667e-01 9.81045902e-01
-7.88787842e-01 -2.58718610e-01 -3.72060210e-01 1.02678967e+00
1.59820125e-01 1.08317637e+00 1.08877420e-01 -7.79082358e-01
3.26779544e-01 2.57100314e-01 -4.85863201e-02 -2.24098861e-01
-1.26843584e+00 2.65614629e-01 5.29630363e-01 1.59511477e-01
-6.36328101e-01 -1.06792223e+00 -1.14653933e+00 -5.23548067e-01
-4.78275687e-01 5.27146459e-01 1.17023647e+00 1.17777395e+00
-2.94106305e-01 2.73262203e-01 1.00870824e+00 -6.32645905e-01
-6.21839106e-01 -1.32629800e+00 -1.15635097e+00 -3.28488499e-01
5.25598526e-02 -7.06187487e-01 -5.30451119e-01 1.21462740e-01] | [4.54866361618042, 4.1901774406433105] |
cbee047c-2352-436e-aa66-9b24550513d4 | conformalized-unconditional-quantile | 2304.01426 | null | https://arxiv.org/abs/2304.01426v1 | https://arxiv.org/pdf/2304.01426v1.pdf | Conformalized Unconditional Quantile Regression | We develop a predictive inference procedure that combines conformal prediction (CP) with unconditional quantile regression (QR) -- a commonly used tool in econometrics that involves regressing the recentered influence function (RIF) of the quantile functional over input covariates. Unlike the more widely-known conditional QR, unconditional QR explicitly captures the impact of changes in covariate distribution on the quantiles of the marginal distribution of outcomes. Leveraging this property, our procedure issues adaptive predictive intervals with localized frequentist coverage guarantees. It operates by fitting a machine learning model for the RIFs using training data, and then applying the CP procedure for any test covariate with respect to a ``hypothetical'' covariate distribution localized around the new instance. Experiments show that our procedure is adaptive to heteroscedasticity, provides transparent coverage guarantees that are relevant to the test instance at hand, and performs competitively with existing methods in terms of efficiency. | ['David Sontag', 'Zeshan Hussain', 'Ahmed M. Alaa'] | 2023-04-04 | null | null | null | null | ['econometrics'] | ['miscellaneous'] | [ 3.00981909e-01 3.46008688e-01 -7.13480473e-01 -5.65568805e-01
-1.27223384e+00 -4.36057240e-01 5.36381841e-01 4.57922071e-01
2.80560441e-02 1.07890737e+00 3.91822338e-01 -7.07750201e-01
-5.59036493e-01 -1.01708484e+00 -1.07089031e+00 -7.27508605e-01
-3.31436992e-01 6.52935684e-01 -6.81037130e-03 3.06986362e-01
2.63592392e-01 3.60892028e-01 -1.21386182e+00 -3.26173492e-02
1.04772890e+00 7.84304798e-01 -4.64839190e-01 3.18458319e-01
3.95570576e-01 5.71420670e-01 -4.02413398e-01 -6.85667157e-01
4.71622236e-02 -3.86625350e-01 -4.28353071e-01 -3.52530986e-01
-2.70516910e-02 -2.25621164e-02 9.97895300e-02 6.65134132e-01
2.30312031e-02 4.68282551e-02 1.43430424e+00 -1.08618450e+00
-5.33800483e-01 8.31738889e-01 -1.04293227e+00 1.25812113e-01
2.50029922e-01 -2.77733952e-01 1.06121755e+00 -7.32878983e-01
2.72898734e-01 1.04595852e+00 1.08253872e+00 -1.21131644e-01
-1.77599967e+00 -5.64820468e-01 -1.52073771e-01 -4.28625762e-01
-1.35802126e+00 -3.27073395e-01 3.27691376e-01 -6.63635075e-01
4.23211902e-01 2.19346330e-01 2.52459615e-01 8.27356100e-01
7.71292686e-01 4.09726471e-01 9.52806652e-01 -4.70918864e-01
5.78518033e-01 2.58567333e-01 5.00833243e-02 2.33771577e-01
3.03344250e-01 4.22557294e-01 -3.31750691e-01 -7.99950123e-01
7.13407993e-01 2.09129062e-02 -1.23436920e-01 -5.18366754e-01
-8.44938934e-01 1.14253223e+00 -2.28434373e-02 -2.33510807e-01
-5.13068199e-01 2.17311114e-01 2.54131466e-01 1.39821231e-01
1.02478242e+00 -5.67268929e-04 -7.50489771e-01 1.58853624e-02
-1.08624661e+00 4.52379674e-01 7.36191750e-01 7.95065463e-01
6.63224995e-01 -2.18373939e-01 -6.68833077e-01 6.04159892e-01
1.37553573e-01 7.19748318e-01 -8.13692808e-02 -7.51929998e-01
3.95208567e-01 3.13859731e-01 3.19723189e-01 -6.41824126e-01
-2.29527846e-01 -6.62588477e-01 -7.69064188e-01 -6.59878924e-02
6.23248816e-01 -2.70555764e-01 -5.88431656e-01 1.80332112e+00
5.43019176e-01 1.67932093e-01 -3.39476913e-01 2.10415959e-01
-4.06023711e-02 5.81433892e-01 4.15981054e-01 -5.20921648e-01
1.01031733e+00 -1.17518850e-01 -3.54621261e-01 1.46943405e-01
6.61816359e-01 -2.40393072e-01 8.35831761e-01 1.73266083e-01
-1.11206043e+00 -3.07446510e-01 -6.25888824e-01 2.21921951e-01
-5.81094855e-03 -1.52191356e-01 5.37768960e-01 7.66930521e-01
-7.17672288e-01 7.57065773e-01 -7.67058671e-01 1.48095891e-01
3.77137333e-01 2.41478592e-01 -1.92609206e-01 1.32603280e-03
-9.40037072e-01 5.15161812e-01 -8.14619362e-02 -3.85590076e-01
-6.20826960e-01 -1.51023591e+00 -8.53904963e-01 4.67729360e-01
3.17124754e-01 -6.40592933e-01 1.20827830e+00 -8.38597596e-01
-1.24225533e+00 4.80116874e-01 -4.04562175e-01 -4.84035641e-01
6.60868526e-01 -1.91300407e-01 -1.68490097e-01 -2.63351768e-01
4.67821985e-01 -2.21992508e-02 6.89901829e-01 -8.22589695e-01
-5.69322884e-01 -4.91696090e-01 -4.92876559e-01 -3.24778378e-01
2.53619909e-01 -9.87695754e-02 -1.26394644e-01 -7.54980445e-01
-3.14025267e-04 -4.91849542e-01 -3.79199296e-01 -4.74246949e-01
-4.82518524e-01 -3.26635718e-01 4.79765376e-03 -6.94658577e-01
1.50333405e+00 -2.20940995e+00 -1.01382658e-01 9.24919188e-01
-3.90570372e-01 -6.66674435e-01 3.52239043e-01 5.01121104e-01
-2.31448144e-01 -3.89283970e-02 -5.90338707e-01 -1.48850054e-01
6.58377782e-02 -1.07895173e-01 -7.57544935e-01 8.87597203e-01
2.29875341e-01 7.94637799e-01 -5.24282038e-01 -2.60099053e-01
-8.88067037e-02 1.48749381e-01 -8.12212050e-01 1.03786282e-01
-2.87065536e-01 4.19406831e-01 -1.69586837e-01 5.44362247e-01
6.12744331e-01 -1.35938376e-01 2.67301410e-01 4.35747206e-01
-4.32426557e-02 1.66967124e-01 -1.03573251e+00 9.27046895e-01
-1.38584554e-01 1.27109438e-01 -4.19712454e-01 -1.22376502e+00
9.82378542e-01 1.29001766e-01 4.18879598e-01 -3.21727425e-01
-1.76839277e-01 1.16146386e-01 -5.80649972e-01 -1.53926283e-01
8.96667913e-02 -3.29746038e-01 -4.17270511e-01 2.37207994e-01
-5.31887859e-02 2.47155622e-01 -2.14429021e-01 -2.58613005e-02
1.01374424e+00 4.63256449e-01 8.37553084e-01 -7.15491295e-01
1.80856124e-01 -1.45838812e-01 7.28528678e-01 8.92734408e-01
4.82317835e-01 3.12373370e-01 1.16096425e+00 -1.76555514e-01
-1.09623933e+00 -1.52225721e+00 -8.80144477e-01 1.18920612e+00
-4.20053393e-01 1.13096774e-01 -5.16063750e-01 -5.39195716e-01
6.89662516e-01 1.03621995e+00 -1.17685544e+00 2.33690664e-02
-1.02063440e-01 -8.16723228e-01 1.78127095e-01 6.70035064e-01
-8.25488344e-02 -6.36168480e-01 -3.13572794e-01 2.64088303e-01
3.08878064e-01 -3.30426663e-01 -3.56974214e-01 3.11264396e-01
-9.43823636e-01 -1.11475050e+00 -6.86438143e-01 -3.23443890e-01
4.97283608e-01 -4.22957361e-01 1.28592908e+00 -4.77382123e-01
4.83122021e-02 3.14134210e-01 4.20389473e-02 -4.26231474e-01
-2.84220070e-01 2.32335031e-02 -4.04921651e-01 1.51235238e-01
2.08015442e-01 -6.69478953e-01 -5.07330358e-01 2.19304547e-01
-7.53305435e-01 -2.96031058e-01 4.08673942e-01 7.33223915e-01
7.66080081e-01 -7.17966482e-02 1.21018624e+00 -1.50367188e+00
1.91517740e-01 -1.09447873e+00 -1.06034791e+00 4.61885780e-01
-7.50230491e-01 2.78380159e-02 2.34949529e-01 -2.81679392e-01
-1.23169303e+00 -2.96343446e-01 3.04271370e-01 5.46173863e-02
-7.54294842e-02 9.72686112e-01 -1.72289237e-01 5.33665001e-01
6.68561280e-01 -2.63120085e-01 -1.55308545e-01 -4.07396257e-01
1.96305126e-01 4.34113741e-01 9.29690361e-01 -9.41166580e-01
6.33004606e-01 4.96162951e-01 5.57812214e-01 -1.90449968e-01
-9.31103110e-01 -3.82149726e-01 -5.30405939e-01 1.34660989e-01
5.44852912e-01 -8.26194704e-01 -7.62865603e-01 -6.49016276e-02
-4.93409455e-01 -4.72208947e-01 -5.38946807e-01 5.92321932e-01
-6.73216105e-01 -8.47653002e-02 -2.20949948e-01 -1.24855483e+00
-9.86866206e-02 -6.07270956e-01 1.05779314e+00 1.09978840e-01
-2.44981840e-01 -1.37456286e+00 4.34742808e-01 -1.35948345e-01
2.04803452e-01 6.56654000e-01 1.46943319e+00 -6.89364612e-01
-1.47470281e-01 -5.93498886e-01 -1.49019003e-01 3.73853510e-03
-1.23732783e-01 2.85117596e-01 -8.17814589e-01 -1.96030974e-01
-1.69155031e-01 1.58637628e-01 6.87237322e-01 1.15726101e+00
1.39019978e+00 -3.46533149e-01 -6.19557559e-01 4.25919682e-01
1.37477314e+00 -3.87412198e-02 6.10209107e-01 1.93760484e-01
-9.75892916e-02 6.29433751e-01 6.88468575e-01 9.13184464e-01
4.09594834e-01 4.93103266e-01 1.04849273e-02 5.18037053e-03
6.73398614e-01 -5.43439269e-01 2.54203588e-01 -1.41843051e-01
8.81812796e-02 8.02165717e-02 -8.18650544e-01 6.72558188e-01
-1.86387277e+00 -1.04281318e+00 -1.53955221e-01 2.85278559e+00
1.09707129e+00 1.61949307e-01 4.57214326e-01 -6.14889041e-02
7.25238264e-01 -3.61435860e-01 -5.16584694e-01 -5.97605169e-01
1.27188221e-01 5.66822171e-01 6.91474438e-01 5.61950564e-01
-9.13222194e-01 1.68697268e-01 6.79575825e+00 7.63126254e-01
-4.79385197e-01 -1.00736231e-01 1.13965261e+00 1.19197167e-01
-7.74336278e-01 2.57658899e-01 -7.52331972e-01 4.86908764e-01
1.30438447e+00 -4.74700511e-01 4.56164777e-02 8.21461558e-01
3.23048979e-01 -4.90847051e-01 -1.37204254e+00 1.69377644e-02
-3.35339099e-01 -1.17824054e+00 -2.61312962e-01 3.41546446e-01
1.02393723e+00 -3.08140278e-01 4.18596894e-01 5.81661940e-01
5.50350547e-01 -1.21246719e+00 5.76406240e-01 1.03163910e+00
1.06594527e+00 -1.31171620e+00 8.50339532e-01 4.94613767e-01
-7.58467555e-01 -5.89208007e-02 -4.08375859e-01 8.15990120e-02
-1.71011254e-01 1.05591655e+00 -7.03167737e-01 5.62969744e-01
5.48600614e-01 3.68519098e-01 -3.04502875e-01 1.09733474e+00
-5.92482090e-02 1.21993697e+00 -1.70118570e-01 4.65589792e-01
-2.37566113e-01 -3.75248343e-01 1.41386300e-01 1.14213777e+00
5.34359396e-01 1.14001788e-01 -8.98052529e-02 9.27599847e-01
-9.74554103e-03 2.58482575e-01 -6.35084212e-01 4.84638095e-01
6.46418095e-01 7.13089228e-01 -2.82644391e-01 -2.22349599e-01
-4.59873259e-01 1.94808543e-01 2.23933801e-01 5.33530295e-01
-8.78190517e-01 -1.31454453e-01 2.65601784e-01 2.88990259e-01
5.09991646e-01 3.79099637e-01 -6.89832985e-01 -6.53046310e-01
-1.69394031e-01 -5.01887023e-01 8.47196281e-01 -5.53346574e-01
-1.55415809e+00 -1.98996216e-01 4.19012666e-01 -8.86370838e-01
-6.31423891e-01 -2.00453848e-01 -9.10236180e-01 1.20176077e+00
-1.37293530e+00 -8.01092803e-01 3.25262249e-01 4.86516297e-01
1.55216502e-02 1.51887193e-01 7.54137099e-01 -3.64121079e-01
-3.55847895e-01 7.70908058e-01 5.55747747e-01 -3.12123299e-01
6.98772967e-01 -1.39327002e+00 1.41927123e-01 5.20514131e-01
-2.39179790e-01 5.41165769e-01 7.72861481e-01 -9.43267465e-01
-8.31540525e-01 -1.06750965e+00 1.04657876e+00 -5.11093557e-01
7.95940936e-01 -4.00620289e-02 -9.58474159e-01 1.08844483e+00
-2.49554604e-01 -1.58411548e-01 1.08814323e+00 6.93668664e-01
-3.39652270e-01 -3.17990303e-01 -1.35252285e+00 2.85891145e-01
5.25775135e-01 -3.08390051e-01 -3.74340326e-01 2.09344670e-01
5.41351795e-01 -2.01284647e-01 -1.30606532e+00 8.43205094e-01
6.91611946e-01 -9.46376741e-01 6.87437773e-01 -8.19582343e-01
6.31854236e-01 8.40877071e-02 -2.48854324e-01 -1.05849600e+00
-4.70034361e-01 -5.70048213e-01 7.42877126e-02 1.36160874e+00
6.62206650e-01 -9.08316433e-01 5.29421389e-01 8.81337106e-01
3.09135430e-02 -9.37252343e-01 -1.18696237e+00 -5.70641518e-01
5.95214784e-01 -5.06304324e-01 9.95518744e-01 7.62765467e-01
1.48020491e-01 -9.45030972e-02 -2.71003723e-01 5.12624979e-01
7.89516866e-01 4.57723558e-01 9.18592572e-01 -1.52141929e+00
-6.99916184e-01 -2.15527147e-01 -5.68732284e-02 -5.28524339e-01
2.14003608e-01 -7.37581134e-01 -9.59057063e-02 -9.65191483e-01
5.13418436e-01 -4.67443138e-01 -2.92331457e-01 3.66721988e-01
-4.00016665e-01 -1.00847520e-01 -5.98164558e-01 -1.59898140e-02
-9.82657596e-02 3.86190832e-01 7.69931078e-01 3.16315502e-01
-2.91166514e-01 7.13757873e-01 -7.96681285e-01 6.20591581e-01
7.10852504e-01 -5.17426074e-01 -2.84419239e-01 2.99336582e-01
2.94247687e-01 7.56389499e-01 4.64318305e-01 -5.20511031e-01
4.55334187e-02 -4.66294855e-01 6.47406042e-01 -7.98812211e-01
-3.15656036e-01 -5.28087020e-01 3.94607574e-01 1.79498583e-01
-5.97779989e-01 1.39153406e-01 1.74274296e-01 8.20078969e-01
4.67776880e-02 -1.12172596e-01 5.62880099e-01 4.87475663e-01
3.14035654e-01 1.83968395e-01 -2.87708670e-01 1.74208179e-01
9.17043924e-01 9.25324336e-02 -2.79128343e-01 -4.65946019e-01
-6.53764009e-01 3.14337164e-01 2.80819327e-01 -2.80192584e-01
2.02812359e-01 -1.22435129e+00 -9.86487031e-01 3.05616796e-01
9.25304145e-02 -9.32304934e-02 9.44637284e-02 1.03583360e+00
2.02156276e-01 4.48455215e-01 3.32902730e-01 -5.77479720e-01
-6.52305484e-01 6.15297914e-01 1.61659271e-01 -2.89362103e-01
-5.25109231e-01 5.51946223e-01 6.32936537e-01 -3.04745376e-01
1.27748176e-01 -2.59235412e-01 4.42181490e-02 -1.16423927e-01
2.93439746e-01 6.51525617e-01 -1.46401420e-01 -1.73546329e-01
-1.36056006e-01 2.41317496e-01 2.83238173e-01 -2.48604745e-01
1.37359321e+00 -1.03058852e-02 -1.58807784e-01 8.17578197e-01
9.36995089e-01 4.40797448e-01 -1.49230671e+00 -3.83626878e-01
4.34891254e-01 -4.47499722e-01 -2.74106443e-01 -9.47199345e-01
-6.02801383e-01 3.53943288e-01 1.38243996e-02 2.04428509e-01
9.18324649e-01 1.99085250e-01 -7.35370293e-02 -1.88177839e-01
2.12819859e-01 -8.40840638e-01 -5.33755898e-01 1.41957086e-02
7.46009052e-01 -8.08868170e-01 2.09806502e-01 -2.26324514e-01
-4.56185400e-01 8.18825901e-01 -7.05183297e-02 -3.20688874e-01
9.42365706e-01 3.37713450e-01 -3.67474467e-01 5.00535418e-04
-8.88810635e-01 3.05263847e-01 7.54076898e-01 5.53312898e-01
3.36265892e-01 2.35462636e-01 -1.73678264e-01 8.80739152e-01
-3.11016947e-01 2.34434544e-03 2.55176216e-01 3.95450115e-01
-2.98568845e-01 -7.87775457e-01 -3.08769822e-01 8.98994088e-01
-7.04518795e-01 -7.42745101e-02 2.15592399e-01 9.98629689e-01
3.34380604e-02 7.79414415e-01 3.21575791e-01 2.49499112e-01
2.14388266e-01 3.83591950e-01 1.15805894e-01 -5.65255284e-01
-2.05397993e-01 4.02091503e-01 -1.40858293e-01 -4.15774614e-01
-1.02852002e-01 -1.40419888e+00 -7.70797491e-01 -5.24192274e-01
-4.14353013e-01 4.98382539e-01 3.26625705e-01 1.13507104e+00
7.42533430e-02 1.36099190e-01 1.02827239e+00 -2.88925916e-01
-8.53443980e-01 -8.86561632e-01 -9.91465449e-01 7.19837612e-03
2.93406755e-01 -6.77357733e-01 -4.71962303e-01 -5.31691574e-02] | [7.56197452545166, 4.382304668426514] |
a4f1e29a-6b8d-48b3-9fce-d43a053e304a | a-principal-agent-framework-for-optimal | 2302.12167 | null | https://arxiv.org/abs/2302.12167v1 | https://arxiv.org/pdf/2302.12167v1.pdf | A Principal-Agent Framework for Optimal Incentives in Renewable Investments | We investigate the optimal regulation of energy production reflecting the long-term goals of the Paris Climate Agreement. We analyze the optimal regulatory incentives to foster the development of non-emissive electricity generation when the demand for power is served either by a monopoly or by two competing agents. The regulator wishes to encourage green investments to limit carbon emissions, while simultaneously reducing intermittency of the total energy production. We find that the regulation of a competitive market is more efficient than the one of the monopoly as measured with the certainty equivalent of the Principal's value function. This higher efficiency is achieved thanks to a higher degree of freedom of the incentive mechanisms which involves cross-subsidies between firms. A numerical study quantifies the impact of the designed second-best contract in both market structures compared to the business-as-usual scenario. In addition, we expand the monopolistic and competitive setup to a more general class of tractable Principal-Multi-Agent incentives problems when both the drift and the volatility of a multi-dimensional diffusion process can be controlled by the Agents. We follow the resolution methodology of Cvitani\'c et al. (2018) in an extended linear quadratic setting with exponential utilities and a multi-dimensional state process of Ornstein-Uhlenbeck type. We provide closed-form expression of the second-best contracts. In particular, we show that they are in rebate form involving time-dependent prices of each state-variable. | ['Nizar Touzi', 'Annika Kemper', 'René Aïd'] | 2023-02-23 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [-2.40654096e-01 3.46019059e-01 -9.89633724e-02 3.11519474e-01
-3.88938695e-01 -1.11168075e+00 6.00985289e-01 6.84087910e-03
-3.56473476e-01 1.00703907e+00 3.65807712e-02 -3.63427192e-01
-6.44960046e-01 -9.73465383e-01 -3.96416485e-01 -1.47841191e+00
8.70722905e-02 4.10795778e-01 -4.98313785e-01 -9.60232317e-02
-6.07572831e-02 4.03495908e-01 -9.36223567e-01 -5.96071422e-01
1.09258389e+00 8.94110501e-01 2.67787337e-01 5.37730932e-01
3.34702641e-01 2.76571721e-01 -2.99011767e-01 1.42867370e-02
9.10355866e-01 -1.18655607e-01 -1.26897216e-01 -1.43906012e-01
-7.46618390e-01 -2.67743349e-01 -2.82302387e-02 1.21228552e+00
5.01981735e-01 1.39117092e-01 8.52252007e-01 -1.45026195e+00
-6.55747116e-01 6.78483963e-01 -6.54193640e-01 1.07309066e-01
-3.51925522e-01 3.14127237e-01 1.03979516e+00 1.20823562e-01
2.66008794e-01 1.06527042e+00 -5.22139715e-03 6.14882350e-01
-1.50293660e+00 -4.84591544e-01 9.33895484e-02 -3.66282254e-01
-9.90264893e-01 -2.16295291e-03 3.68833691e-01 -7.69552231e-01
6.90305054e-01 4.32081878e-01 6.60325468e-01 7.01069236e-01
4.26735640e-01 2.72576064e-01 1.26981056e+00 -1.51756316e-01
5.79635084e-01 4.96861003e-02 -2.65002668e-01 -3.02329063e-01
7.30929971e-01 4.68235075e-01 2.58187324e-01 -3.79027396e-01
4.76389885e-01 -1.33417830e-01 -2.67766565e-01 -5.04178405e-01
-7.83025861e-01 7.95935929e-01 3.02012563e-01 5.08435309e-01
-1.03632224e+00 3.14402372e-01 9.56224278e-02 3.33820730e-01
3.39154571e-01 4.64655519e-01 -4.47795928e-01 4.38002273e-02
-5.84204376e-01 1.71782330e-01 7.76048362e-01 7.17742682e-01
9.77508053e-02 9.47540998e-02 -4.83954072e-01 -5.19130379e-02
5.35259485e-01 1.07390070e+00 1.33475706e-01 -1.24117196e+00
3.86593282e-01 8.77615139e-02 9.11994636e-01 -3.02245677e-01
-2.83953279e-01 -7.15321660e-01 -9.70800936e-01 3.35089475e-01
6.49903953e-01 -6.57683671e-01 -2.21513167e-01 2.07515311e+00
1.86515465e-01 -4.31175888e-01 1.63200110e-01 6.19036734e-01
-6.52608573e-01 9.97297525e-01 -1.10426500e-01 -9.42593515e-01
1.35639203e+00 -4.93070513e-01 -8.52360368e-01 3.73546809e-01
4.23022687e-01 -2.04550236e-01 3.55420887e-01 1.72284007e-01
-1.17484498e+00 2.54054189e-01 -5.47579944e-01 6.13030612e-01
-3.41132134e-01 -1.82003915e-01 -9.83833969e-02 5.63748062e-01
-9.52784717e-01 5.86994827e-01 -6.45727575e-01 5.21000624e-02
2.04576120e-01 1.23652548e-01 2.27099583e-01 6.18387222e-01
-1.22679961e+00 8.95818949e-01 -6.23273617e-03 3.13056886e-01
-8.48897934e-01 -9.57310796e-01 -2.54004478e-01 5.91256440e-01
5.49828589e-01 -7.07760870e-01 1.33097363e+00 -8.04906249e-01
-1.37726986e+00 2.23627120e-01 4.66623813e-01 -5.64980626e-01
1.10115778e+00 2.46220782e-01 8.63959938e-02 5.48711680e-02
4.23436195e-01 -3.65123339e-02 3.20879996e-01 -1.01844895e+00
-6.36402190e-01 -5.51062107e-01 4.45008911e-02 3.78623039e-01
-1.14596434e-01 -2.12284952e-01 7.25158393e-01 -5.15638530e-01
-6.72602534e-01 -1.21773231e+00 -3.39956403e-01 -4.59300637e-01
-4.07339513e-01 -3.84297669e-01 5.50320745e-01 -5.64367294e-01
8.38843584e-01 -1.98858953e+00 1.81056008e-01 1.86341748e-01
-3.35224360e-01 -1.60631061e-01 1.22150764e-01 6.60936415e-01
-8.25974941e-02 4.63591069e-01 -5.07514715e-01 -1.86885074e-01
6.83539212e-01 3.55781615e-01 -2.14151964e-01 7.06391752e-01
-2.51843128e-02 6.12444937e-01 -9.72501874e-01 3.99472415e-01
-5.89296445e-02 4.68443818e-02 -9.38729271e-02 -1.24650277e-01
-2.88570285e-01 4.91242856e-01 -7.07684278e-01 9.76795778e-02
6.43184483e-01 -2.84085069e-02 3.06673318e-01 4.83783990e-01
-7.16883004e-01 -1.66166797e-01 -1.41865385e+00 1.06369936e+00
-5.53965688e-01 2.90014833e-01 7.22112238e-01 -1.11076105e+00
3.81182432e-01 4.54214752e-01 7.54248738e-01 -4.64301229e-01
-4.66434732e-02 4.86250997e-01 2.94777662e-01 -3.52610201e-01
1.52944833e-01 -6.83275163e-01 2.11772881e-02 6.64207518e-01
-6.08110785e-01 -3.76041293e-01 4.44070727e-01 9.40913558e-02
8.75364423e-01 -1.76979959e-01 2.39497453e-01 -1.08921862e+00
3.03005219e-01 -4.39512998e-01 6.93716526e-01 4.91791666e-01
-2.14697003e-01 -2.85276383e-01 1.09308839e+00 5.96820891e-01
-1.06034267e+00 -8.81064832e-01 -9.27063301e-02 4.29435670e-01
-2.11723953e-01 4.44419801e-01 -5.61154723e-01 -3.43584448e-01
6.15190446e-01 1.35637283e+00 -6.99667513e-01 1.40484646e-01
-7.14781284e-02 -9.66400445e-01 -3.25116329e-02 1.01986557e-01
3.89279783e-01 -5.68492472e-01 -1.06401527e+00 1.33796528e-01
2.19190359e-01 -5.68856120e-01 -6.86040461e-01 4.52304423e-01
-3.97034347e-01 -9.15577829e-01 -9.56636608e-01 6.85916394e-02
6.93735540e-01 -8.95914659e-02 6.15326464e-01 -6.84058309e-01
2.51727521e-01 6.86888158e-01 3.17001194e-01 -7.37120807e-01
-4.72735524e-01 -7.04975575e-02 1.75611228e-01 1.48068100e-01
-1.93134204e-01 -4.11236465e-01 -8.62011433e-01 1.04872800e-01
-1.12502420e+00 -3.00375789e-01 1.88391715e-01 5.76846123e-01
4.28234428e-01 4.97510523e-01 1.02876782e+00 -2.84246057e-01
9.22063231e-01 -6.27513111e-01 -1.53550386e+00 2.81933576e-01
-1.08658803e+00 3.26912135e-01 6.03272438e-01 -2.24219799e-01
-1.38706577e+00 1.66865736e-01 5.97187281e-01 5.32770678e-02
2.00440869e-01 2.43299723e-01 -3.41691524e-01 4.02947575e-01
-1.33744031e-01 -3.15606482e-02 5.67482412e-02 -5.44698656e-01
5.77677369e-01 6.70741618e-01 2.92124152e-01 -8.75219882e-01
1.13314521e+00 4.89830852e-01 4.73647326e-01 -4.14061129e-01
-9.34435129e-02 -3.40097964e-01 -2.93705940e-01 -1.97677404e-01
9.70060050e-01 -9.89405394e-01 -1.30959857e+00 4.77670014e-01
-1.18079996e+00 -4.03715640e-01 -9.89871919e-01 5.00508606e-01
-7.86733210e-01 9.24688503e-02 -3.49041969e-01 -1.72772288e+00
-2.19741479e-01 -9.10753012e-01 5.55066288e-01 5.01408696e-01
6.42256856e-01 -9.44359362e-01 5.36578178e-01 1.14820719e-01
7.60210216e-01 5.08837879e-01 7.60838747e-01 -2.15495691e-01
-7.50601351e-01 1.84632048e-01 2.67181605e-01 4.37502235e-01
3.00713122e-01 1.17100857e-01 -5.58148623e-01 -4.54254299e-01
3.72582167e-01 2.56409794e-01 4.97685462e-01 7.20966935e-01
3.15048039e-01 -7.39541650e-01 -1.52762994e-01 9.55366157e-03
1.70955360e+00 7.62652397e-01 2.99402982e-01 4.66053277e-01
3.40913571e-02 9.78216946e-01 4.03851748e-01 6.55603409e-01
3.60919267e-01 5.48128426e-01 7.02057719e-01 2.51539320e-01
8.95673573e-01 3.40675563e-02 4.09752041e-01 2.05079198e-01
-1.57084942e-01 -2.53878027e-01 -6.42613173e-01 7.59154439e-01
-1.91050577e+00 -1.19773483e+00 -2.55729318e-01 2.49566078e+00
6.25747263e-01 -1.56345263e-01 6.08446002e-01 -1.77361637e-01
7.48751819e-01 -2.05086872e-01 -7.15365708e-01 -6.43922687e-01
-4.74282056e-01 -3.33602339e-01 1.18794632e+00 4.41700041e-01
-4.55142826e-01 -1.39625445e-01 5.46638536e+00 4.36204553e-01
-7.64409423e-01 2.41240457e-01 9.26536858e-01 -3.47623676e-01
-5.47317207e-01 1.09225415e-01 -2.85353869e-01 8.38101089e-01
1.07585311e+00 -9.96492863e-01 5.44454873e-01 3.80182266e-01
1.09282732e+00 -3.44837666e-01 -7.98548996e-01 2.82404214e-01
-9.28525448e-01 -9.91601825e-01 -1.09665895e+00 9.23006356e-01
1.08804119e+00 1.88731194e-01 1.37706161e-01 -9.85073745e-02
5.05127072e-01 -6.25888169e-01 8.25567603e-01 7.10490286e-01
2.76855201e-01 -8.23378265e-01 6.03562295e-01 6.15999818e-01
-1.28692400e+00 -6.88010871e-01 8.53862986e-02 -1.43066153e-01
4.73447412e-01 7.32834399e-01 2.21161414e-02 7.63773084e-01
4.48483139e-01 3.45095880e-02 2.37359300e-01 8.02379549e-01
-1.31100118e-01 4.14054930e-01 -6.70848310e-01 -1.59364149e-01
3.02565545e-01 -1.11050200e+00 7.04714417e-01 7.49829292e-01
4.89531368e-01 9.33673009e-02 -2.23957896e-01 1.28077829e+00
2.40233894e-02 6.28725961e-02 -6.69781804e-01 -3.05527925e-01
3.59366953e-01 1.28231752e+00 -7.17598498e-01 5.48583381e-02
-3.08225960e-01 4.51930434e-01 -4.36245620e-01 4.56216782e-01
-6.47141814e-01 -2.63815463e-01 7.22910762e-01 2.51099560e-03
4.56875622e-01 -1.63062736e-01 -2.24686459e-01 -9.84839678e-01
1.37669757e-01 -2.66707420e-01 4.75223750e-01 -5.07273138e-01
-1.28193581e+00 -2.33547419e-01 1.03304863e-01 -5.78508198e-01
-4.91122633e-01 -2.12521568e-01 -8.12609196e-01 1.39235139e+00
-1.84679151e+00 -4.19635177e-01 6.53985023e-01 1.43462136e-01
2.44965255e-02 2.00695485e-01 3.38654310e-01 5.14292046e-02
-8.74415040e-01 -3.98312211e-01 1.26473534e+00 -2.84624308e-01
3.85739584e-03 -1.82246172e+00 -2.26538386e-02 9.08697128e-01
-4.14893508e-01 3.74085099e-01 8.35810006e-01 -6.42903447e-01
-1.50914049e+00 -7.74678051e-01 4.42048073e-01 9.86648351e-02
1.23215961e+00 -3.19880694e-01 -5.41197956e-01 3.29354525e-01
7.88441658e-01 -3.16770911e-01 2.07967281e-01 -6.05414271e-01
2.17899263e-01 -3.12459648e-01 -1.47342610e+00 1.79799184e-01
4.32219684e-01 -2.43464321e-01 -6.39222413e-02 3.62877369e-01
4.28974032e-01 4.36303169e-01 -1.03310239e+00 1.61945075e-01
3.47155184e-01 -3.51612091e-01 4.04609740e-01 -4.49378818e-01
-7.96691924e-02 -3.41178745e-01 -1.28537014e-01 -1.79033399e+00
-3.03993553e-01 -1.29655111e+00 2.52369512e-02 1.38112962e+00
2.85167456e-01 -1.09132266e+00 1.75542876e-01 7.17657685e-01
4.73391294e-01 -4.24666584e-01 -1.55698776e+00 -1.30174208e+00
9.63009894e-01 2.38315299e-01 5.63799381e-01 7.20555604e-01
1.79260522e-01 1.96194246e-01 -1.23589218e-01 2.46204868e-01
9.74013329e-01 7.01237693e-02 6.69699088e-02 -1.06350243e+00
-4.08088148e-01 -7.04499960e-01 3.93621862e-01 -5.24282575e-01
1.59754798e-01 -6.94693804e-01 8.74544978e-02 -1.60781944e+00
1.40651748e-01 -2.36054584e-01 -4.49551612e-01 4.41575013e-02
3.06558102e-01 -5.33682942e-01 6.71913624e-01 1.17006063e-01
9.40627232e-02 8.93112183e-01 1.08382511e+00 -3.55432212e-01
-1.83083922e-01 4.64408517e-01 -8.47151458e-01 1.71743155e-01
9.02912617e-01 -4.76353526e-01 -5.62375963e-01 -1.68962292e-02
4.80323106e-01 3.78683507e-01 2.29754388e-01 -3.36991251e-01
8.94832909e-02 -8.11500430e-01 -5.01467347e-01 -4.65699613e-01
-2.64161319e-01 -9.58810031e-01 8.06042373e-01 9.35467422e-01
-3.04634184e-01 1.29313678e-01 -1.77920237e-01 8.83063316e-01
4.53202903e-01 -3.49389523e-01 9.11548376e-01 -9.75254457e-04
4.83055115e-01 1.32861599e-01 -7.21859574e-01 -5.24827056e-02
1.39096606e+00 5.21831214e-01 -6.67203248e-01 -5.33597171e-01
-5.07098496e-01 8.33681285e-01 4.83446568e-01 1.24488123e-01
-3.84799659e-01 -1.08587146e+00 -7.98232079e-01 -4.68179256e-01
-4.94864792e-01 -2.72072792e-01 1.84883967e-01 9.32792366e-01
2.13450655e-01 5.44710636e-01 -3.02752722e-02 -2.82997131e-01
-5.77894151e-01 7.03456283e-01 1.00794673e+00 -4.95193630e-01
-1.33691564e-01 -7.12726638e-02 3.40338409e-01 -3.90250306e-03
-1.69193819e-01 -4.49092746e-01 2.10312888e-01 6.02454722e-01
1.98661253e-01 7.89198697e-01 -2.41768256e-01 -3.86933208e-01
-3.42797995e-01 1.11261174e-01 5.75761676e-01 -5.51170766e-01
1.49383152e+00 -3.79299700e-01 -2.46461764e-01 5.88125229e-01
7.34373093e-01 -8.11679885e-02 -1.47508824e+00 2.06160396e-01
1.79007053e-01 9.39945430e-02 8.62232894e-02 -9.97458041e-01
-1.27913320e+00 4.35493678e-01 5.62634885e-01 1.17637932e+00
8.80886972e-01 -3.09948087e-01 -8.60118866e-03 1.17674656e-01
8.42848122e-02 -1.48404515e+00 -4.77840483e-01 -8.63850042e-02
9.65858698e-01 -5.90697944e-01 -1.83409646e-01 6.98853582e-02
-4.46765423e-01 6.98275208e-01 5.49512319e-02 -1.07297964e-01
7.92991400e-01 4.36021149e-01 -5.35937548e-01 2.96039253e-01
-7.78509080e-01 -2.48968303e-01 -1.46544054e-01 4.38905329e-01
-3.47547941e-02 7.77949274e-01 -9.10265982e-01 5.48189759e-01
3.07681978e-01 -1.59534067e-01 1.07309651e+00 7.54065394e-01
-8.15072507e-02 -8.74917209e-01 -4.56619054e-01 2.79502273e-01
-4.96684760e-01 1.63852423e-01 -7.45777860e-02 6.50081158e-01
1.96935497e-02 9.64700580e-01 2.04744816e-01 8.05830002e-01
6.48735762e-01 4.44745570e-02 1.36448875e-01 -2.95360476e-01
-4.15005773e-01 4.45855230e-01 -3.38004321e-01 5.66508919e-02
-5.15533149e-01 -1.06225479e+00 -9.78320420e-01 -2.72836804e-01
-5.58169782e-01 6.23945177e-01 9.07145381e-01 1.01178908e+00
2.26323947e-01 3.36521924e-01 1.19600356e+00 -4.17766720e-01
-1.48847592e+00 -7.68556535e-01 -1.25620735e+00 -4.87370901e-02
3.60400617e-01 -2.09926724e-01 -9.63794470e-01 -4.71243024e-01] | [5.102737903594971, 3.2528750896453857] |
c7517ce7-855f-429a-917c-6319835217a6 | a-semi-supervised-approach-for-a-better | 2210.11899 | null | https://arxiv.org/abs/2210.11899v2 | https://arxiv.org/pdf/2210.11899v2.pdf | A Semi-supervised Approach for a Better Translation of Sentiment in Dialectical Arabic UGT | In the online world, Machine Translation (MT) systems are extensively used to translate User-Generated Text (UGT) such as reviews, tweets, and social media posts, where the main message is often the author's positive or negative attitude towards the topic of the text. However, MT systems still lack accuracy in some low-resource languages and sometimes make critical translation errors that completely flip the sentiment polarity of the target word or phrase and hence delivers a wrong affect message. This is particularly noticeable in texts that do not follow common lexico-grammatical standards such as the dialectical Arabic (DA) used on online platforms. In this research, we aim to improve the translation of sentiment in UGT written in the dialectical versions of the Arabic language to English. Given the scarcity of gold-standard parallel data for DA-EN in the UGT domain, we introduce a semi-supervised approach that exploits both monolingual and parallel data for training an NMT system initialised by a cross-lingual language model trained with supervised and unsupervised modeling objectives. We assess the accuracy of sentiment translation by our proposed system through a numerical 'sentiment-closeness' measure as well as human evaluation. We will show that our semi-supervised MT system can significantly help with correcting sentiment errors detected in the online translation of dialectical Arabic UGT. | ['Ashraf Tantawy', 'Emad Mohamed', 'Constantin Orasan', 'Hadeel Saadany'] | 2022-10-21 | null | null | null | null | ['nmt'] | ['computer-code'] | [ 1.62420034e-01 1.84886694e-01 -2.12483257e-01 -4.88734454e-01
-8.07627797e-01 -9.70319569e-01 7.47211993e-01 3.73130381e-01
-4.23666209e-01 9.01080906e-01 1.98833108e-01 -5.26811659e-01
5.34910440e-01 -7.48640776e-01 -6.39799774e-01 -3.20301622e-01
6.65212512e-01 8.65862429e-01 -4.95813787e-01 -1.12138546e+00
2.51524150e-01 1.60589106e-02 -1.01426101e+00 6.42061472e-01
1.34634089e+00 6.34632349e-01 -3.50495009e-03 2.42846161e-01
-3.86838526e-01 8.68884742e-01 -5.93136370e-01 -1.17614853e+00
2.09206849e-01 -7.61295438e-01 -1.01131415e+00 4.64948565e-02
1.32458761e-01 2.25725159e-01 6.52971208e-01 1.32097602e+00
3.86194140e-01 -4.77092266e-01 7.08673120e-01 -1.03806913e+00
-6.58901095e-01 8.64895999e-01 -3.75254065e-01 -2.17987284e-01
5.75669885e-01 -1.76911071e-01 1.14372897e+00 -1.22927141e+00
1.03911805e+00 1.24330449e+00 6.04320645e-01 4.03571963e-01
-1.03794944e+00 -3.48815411e-01 -2.72512972e-01 8.40175077e-02
-9.66586292e-01 -3.79734606e-01 4.40291345e-01 -3.95138830e-01
9.66943264e-01 2.26105183e-01 5.30595839e-01 1.20035887e+00
6.48556054e-01 5.75838804e-01 1.60011053e+00 -9.35228169e-01
-8.77254829e-02 7.28085220e-01 -3.59262437e-01 4.24884975e-01
2.83891987e-02 -4.03996140e-01 -7.41683364e-01 1.53676301e-01
-1.60616100e-01 -5.60736358e-01 3.52071226e-02 3.43878716e-01
-1.34059501e+00 1.08013523e+00 2.13901866e-02 6.29495263e-01
-4.64488983e-01 -5.75581253e-01 6.58447623e-01 8.86576951e-01
9.41291153e-01 5.44687092e-01 -5.91824770e-01 -3.13842326e-01
-8.66281629e-01 9.21335444e-02 1.09366798e+00 8.89985502e-01
8.64943266e-01 -1.49889275e-01 2.90419340e-01 1.07937276e+00
4.78029042e-01 1.03419363e+00 8.11821043e-01 -3.18113625e-01
6.44167423e-01 8.38291168e-01 2.42770404e-01 -1.20932627e+00
-1.83229595e-01 -3.23899180e-01 -4.60047394e-01 -2.18527824e-01
3.52745622e-01 -4.61343378e-01 -4.27001297e-01 1.55397737e+00
3.85497481e-01 -1.16202450e+00 6.93696797e-01 9.60311592e-01
6.22268975e-01 8.90020609e-01 -2.01343909e-01 -5.57526588e-01
1.21038365e+00 -9.46368337e-01 -1.00614989e+00 -5.30940771e-01
1.08115685e+00 -1.70662880e+00 1.23042929e+00 1.54091373e-01
-1.07676852e+00 -1.20802090e-01 -1.05013430e+00 -1.10474840e-01
-7.50192761e-01 1.59778163e-01 -3.62625122e-02 7.70307600e-01
-8.85446846e-01 4.17600691e-01 -3.76780331e-01 -6.28528595e-01
-1.12658471e-01 2.39608541e-01 -3.71111006e-01 1.22955954e-02
-1.57721448e+00 1.63827002e+00 5.47175519e-02 1.63939536e-01
-1.81573197e-01 -2.18963623e-01 -7.63004899e-01 -5.96823931e-01
-1.65946353e-02 -1.19070075e-01 1.30495334e+00 -1.83989370e+00
-1.79823041e+00 1.33393776e+00 -2.70600766e-01 -3.18977624e-01
6.62696302e-01 -1.86674282e-01 -6.00355685e-01 -2.47911632e-01
5.83507359e-01 3.50673258e-01 8.28638077e-01 -1.00476980e+00
-7.30901957e-01 -4.18502718e-01 -1.49210185e-01 4.32397902e-01
-4.27405536e-01 6.21599853e-01 1.06537223e-01 -5.97671568e-01
-1.76785085e-02 -1.19862497e+00 4.42420226e-03 -7.82512307e-01
-2.51197040e-01 9.81583726e-03 6.96277380e-01 -9.77790415e-01
1.18780410e+00 -1.62645137e+00 3.67437482e-01 3.17830116e-01
-3.73888940e-01 1.57905579e-01 1.48636639e-01 9.45947409e-01
1.11344762e-01 1.50071785e-01 -3.30151647e-01 -2.66138494e-01
1.01717971e-01 4.38199043e-01 -3.98164630e-01 3.76586229e-01
1.77274048e-01 9.80335057e-01 -1.04846787e+00 -5.45953393e-01
-2.33483091e-01 2.19928876e-01 -1.05453424e-01 6.02916405e-02
-2.11167395e-01 5.87386549e-01 -9.89287943e-02 5.40586531e-01
3.61532271e-01 9.55798179e-02 4.77496386e-01 -8.26732442e-02
-4.96907771e-01 8.41365576e-01 -4.57752943e-01 1.36365831e+00
-8.28993678e-01 6.78331912e-01 -2.47221768e-01 -6.40281737e-01
1.09477925e+00 4.02169526e-01 2.50096112e-01 -9.20297384e-01
4.45593774e-01 1.00037539e+00 1.26386762e-01 -2.64089763e-01
8.20204139e-01 -5.12760580e-01 -3.03622603e-01 7.86348462e-01
-4.55852374e-02 -5.40498734e-01 4.67150450e-01 4.79349717e-02
3.90825063e-01 8.29904228e-02 3.11729491e-01 -4.60156500e-01
7.16954768e-01 6.24206901e-01 3.00861627e-01 -5.86789511e-02
1.20016515e-01 5.41447878e-01 5.54211140e-01 -1.52254030e-01
-1.12905514e+00 -3.72659951e-01 -1.46041036e-01 1.05473244e+00
-3.31786215e-01 -6.00682199e-01 -1.15209889e+00 -8.32571507e-01
-4.60777730e-01 9.52844620e-01 -5.93378425e-01 1.00018628e-01
-5.28208375e-01 -1.08216059e+00 4.81348962e-01 -1.40676498e-01
2.85999000e-01 -9.49216068e-01 -1.27587169e-01 4.41483945e-01
-8.23463142e-01 -1.33138466e+00 -4.17611063e-01 8.78612176e-02
-5.90374231e-01 -8.23638380e-01 -4.82973337e-01 -8.83922815e-01
8.95938337e-01 -2.14887217e-01 1.08180058e+00 -2.26469964e-01
6.86095536e-01 2.84445495e-03 -7.93641210e-01 -6.74265742e-01
-1.38077378e+00 3.57001603e-01 2.11255640e-01 3.46765161e-01
6.35211766e-01 -1.23481683e-01 -4.22463976e-02 4.29286361e-01
-8.93760324e-01 2.76564866e-01 4.00925994e-01 7.16125786e-01
1.95016980e-01 -4.94840175e-01 6.14385486e-01 -1.23110449e+00
9.40494895e-01 -5.01580596e-01 -2.44967520e-01 1.91284463e-01
-1.00032651e+00 -7.48298168e-02 9.18378174e-01 -2.03516513e-01
-9.23944592e-01 -3.62627476e-01 -4.72398490e-01 3.13905567e-01
3.98496687e-01 1.10245299e+00 2.03485973e-02 -3.50061692e-02
9.76305723e-01 1.48004636e-01 1.16937041e-01 -6.74912930e-02
1.58683032e-01 1.19211876e+00 2.19106823e-01 -4.90817308e-01
7.84918606e-01 1.67480186e-01 -3.27135444e-01 -4.39761788e-01
-9.60520864e-01 -2.00733051e-01 -7.64227629e-01 -4.61936295e-01
6.05654597e-01 -8.59856665e-01 2.88207848e-02 6.05072677e-01
-1.27714074e+00 -1.82923764e-01 3.37592922e-02 2.09918201e-01
-4.56498623e-01 1.44562975e-01 -8.26502204e-01 -3.66967440e-01
-8.10693622e-01 -1.27320075e+00 9.40760434e-01 -1.19288683e-01
-7.73481846e-01 -1.25127876e+00 3.90707463e-01 4.70974058e-01
3.95291179e-01 2.15498865e-01 1.09339643e+00 -6.62049770e-01
4.14195538e-01 -2.81389266e-01 1.48963574e-02 5.79910278e-01
4.89243507e-01 1.73674792e-01 -5.00684798e-01 -1.29686609e-01
1.47409052e-01 -4.34329301e-01 -1.36934966e-01 -3.30484599e-01
-3.24586689e-01 -6.75077319e-01 3.50962669e-01 -1.21287443e-01
1.31056082e+00 -1.40853867e-01 5.23869038e-01 7.57133365e-01
5.66659153e-01 1.03304589e+00 1.09501231e+00 6.19760901e-02
6.62547112e-01 6.44766271e-01 3.05478908e-02 -1.78609751e-02
2.54565716e-01 -2.91765392e-01 1.17831886e+00 1.66384518e+00
5.34972176e-02 -9.13009942e-02 -1.05785215e+00 6.77400529e-01
-1.80064678e+00 -4.38891560e-01 -6.86669350e-01 1.93868423e+00
1.37983739e+00 2.09716074e-02 -1.43286601e-01 1.58917606e-01
5.79206169e-01 -1.39287278e-01 1.14808187e-01 -1.18054473e+00
-5.02503932e-01 1.38593659e-01 5.11277795e-01 7.35931635e-01
-5.94942272e-01 1.41129267e+00 5.35206079e+00 7.43213475e-01
-1.54544961e+00 5.08268118e-01 4.85325366e-01 2.88016915e-01
-4.06244457e-01 1.18884826e-02 -6.88197732e-01 3.33638370e-01
1.29126334e+00 -3.24983269e-01 3.34315389e-01 5.67759693e-01
5.95016181e-01 2.01515183e-02 -9.14878726e-01 7.29804218e-01
3.84122968e-01 -9.85936999e-01 1.55389994e-01 -6.26394674e-02
1.18481517e+00 1.66397020e-01 1.56730995e-01 9.41549763e-02
7.86741078e-02 -8.16106379e-01 1.11675441e+00 -1.53974891e-01
7.92054355e-01 -8.27128589e-01 1.09404898e+00 3.03343385e-01
-4.81766820e-01 4.85589176e-01 -2.59009540e-01 -2.14447141e-01
5.86762577e-02 5.76445282e-01 -1.14892936e+00 6.21623993e-01
5.18152654e-01 8.63423824e-01 -6.11254930e-01 -1.12265989e-01
-6.30664945e-01 6.91418409e-01 -2.27964535e-01 -3.57613295e-01
6.18427634e-01 -7.26905286e-01 6.37880325e-01 1.37159443e+00
3.89133424e-01 -3.60289574e-01 -1.94269687e-01 4.76606309e-01
-1.72266543e-01 1.18124366e+00 -7.39973307e-01 -3.77214491e-01
-2.21170306e-01 1.35042942e+00 -7.45544255e-01 -4.32123721e-01
-4.12774146e-01 1.15872705e+00 2.21171021e-01 -6.57798583e-03
-5.09968817e-01 1.69154555e-02 3.60933870e-01 1.29656523e-01
-6.67244121e-02 -6.47144243e-02 -5.09953439e-01 -1.34698617e+00
2.23927885e-01 -1.64850032e+00 -3.42136286e-02 -7.67258048e-01
-1.27828789e+00 1.18183911e+00 -4.81223285e-01 -1.48856974e+00
-6.25335157e-01 -7.84765005e-01 -2.25179821e-01 9.01032269e-01
-1.42367613e+00 -1.32470548e+00 3.77358526e-01 3.73434603e-01
4.70563114e-01 -2.10783094e-01 1.03518665e+00 3.83546293e-01
-3.28090072e-01 4.89945859e-01 2.43944570e-01 8.63310918e-02
1.29178107e+00 -1.21902728e+00 2.78698355e-01 9.45615411e-01
7.95940310e-02 6.37658834e-01 1.17945325e+00 -8.61461878e-01
-1.28229797e+00 -1.08520508e+00 1.84248745e+00 -5.77555597e-01
1.02565300e+00 -3.23915362e-01 -5.33297718e-01 6.41258895e-01
7.59079278e-01 -7.47969508e-01 8.80591810e-01 -7.03183711e-02
-1.71306670e-01 6.74943775e-02 -1.17564368e+00 9.46074665e-01
4.03275967e-01 -5.38716376e-01 -6.13576472e-01 7.50635028e-01
3.55223089e-01 -5.18472254e-01 -9.52509046e-01 2.19027385e-01
4.62566018e-01 -5.74688852e-01 1.58147022e-01 -6.10215127e-01
9.28947330e-01 -2.72374868e-01 -3.36201906e-01 -1.79039681e+00
4.59102541e-01 -7.36842513e-01 5.74362159e-01 1.18872833e+00
1.11164212e+00 -7.19630361e-01 2.80493140e-01 3.77485633e-01
-3.18362303e-02 -5.29785395e-01 -7.70142138e-01 -2.22400784e-01
4.05685753e-01 -4.64444816e-01 3.21931064e-01 1.28357673e+00
5.20261645e-01 7.94003963e-01 -4.29914922e-01 -1.51677176e-01
1.63276657e-01 2.09322602e-01 7.64349163e-01 -8.27757776e-01
2.06684351e-01 -2.82645077e-01 -1.57710746e-01 -5.04352868e-01
2.00791046e-01 -1.13965034e+00 5.21217585e-02 -1.27233791e+00
-2.84114957e-01 -2.10321426e-01 4.34833318e-01 4.73325849e-01
2.34400798e-02 6.88867331e-01 2.15384662e-02 4.15718943e-01
-2.17677817e-01 5.25295556e-01 1.16907012e+00 -4.97912206e-02
-1.83632150e-01 -2.90782124e-01 -6.66543424e-01 7.20427096e-01
7.89814711e-01 -7.90006042e-01 -1.56621505e-02 -3.31280589e-01
1.13274610e+00 -2.96098650e-01 -3.04607660e-01 -3.44879538e-01
-7.06719235e-02 -2.03545630e-01 -5.26627079e-02 -2.87292361e-01
-2.27293208e-01 -8.23390186e-01 -1.90251656e-02 4.17775393e-01
-2.18303174e-01 6.90292478e-01 7.11018220e-02 -2.69257993e-01
-5.02158284e-01 -3.95335525e-01 7.69441843e-01 -1.53032377e-01
-6.42869696e-02 -1.52048603e-01 -8.98133576e-01 3.66684198e-02
6.39715016e-01 2.66905818e-02 -2.38780200e-01 -5.99016488e-01
-3.45424235e-01 3.73902693e-02 8.43558073e-01 6.84580743e-01
3.11823934e-01 -1.28831255e+00 -1.20080054e+00 9.05616488e-03
3.04931670e-01 -5.34715354e-01 -4.09354180e-01 1.33359766e+00
-8.27645719e-01 3.11291099e-01 -3.89152050e-01 -3.73888314e-01
-1.04678178e+00 5.76315895e-02 3.95818472e-01 -3.93034577e-01
1.02399122e-02 1.47167072e-01 -4.63878632e-01 -1.10281134e+00
-6.48978710e-01 -3.19011621e-02 -2.93560356e-01 4.77438956e-01
2.48045579e-01 3.13290358e-02 5.88543415e-01 -1.61398828e+00
-9.53834057e-02 3.45620990e-01 -1.61464453e-01 -6.51284516e-01
1.05391479e+00 -5.53951025e-01 -9.10615385e-01 7.64407575e-01
1.19285178e+00 5.99979162e-01 -7.73471817e-02 -1.14342414e-01
1.87222108e-01 -2.00561896e-01 -2.22556457e-01 -9.27640676e-01
-6.03186548e-01 5.25340021e-01 2.63334602e-01 1.36207655e-01
8.21984053e-01 -3.97364557e-01 1.02201879e+00 4.02685642e-01
4.48486477e-01 -1.58980000e+00 -3.82679462e-01 9.50747192e-01
1.01848042e+00 -1.45620322e+00 -3.36534739e-01 -3.65195304e-01
-9.46854830e-01 1.27169597e+00 3.27083141e-01 2.03035861e-01
3.52020711e-01 4.46656952e-03 1.00483119e+00 -8.84593204e-02
-5.62047660e-01 -5.70962578e-02 1.85805544e-01 2.17803627e-01
8.91074836e-01 6.99436292e-02 -1.06077182e+00 4.43866700e-01
-9.86836672e-01 -3.66014093e-01 8.72910738e-01 8.80882919e-01
-1.74008198e-02 -1.53342366e+00 -4.95164484e-01 1.34244800e-01
-9.69846606e-01 -3.84073347e-01 -8.86364162e-01 4.76591885e-01
-1.06218249e-01 1.31279576e+00 -3.96991342e-01 -5.26874185e-01
2.05972001e-01 3.30382049e-01 2.68437117e-01 -5.98740041e-01
-1.14547551e+00 1.21897683e-01 5.41124284e-01 -1.84192702e-01
-8.20991278e-01 -7.70847559e-01 -1.18708920e+00 -5.59109032e-01
-1.94356695e-01 5.88314116e-01 1.10082400e+00 1.49561560e+00
-4.99143358e-03 -1.08517759e-01 9.32692826e-01 -3.72974724e-01
-2.49345109e-01 -1.38748431e+00 -1.00120164e-01 4.99905646e-01
-5.15402071e-02 -5.99668585e-02 -3.07925910e-01 3.09833765e-01] | [11.508479118347168, 10.312270164489746] |
5b9d9bdb-ce88-49ad-ae8a-93b7a2cb4ccb | neural-language-modeling-by-jointly-learning | 1711.02013 | null | http://arxiv.org/abs/1711.02013v2 | http://arxiv.org/pdf/1711.02013v2.pdf | Neural Language Modeling by Jointly Learning Syntax and Lexicon | We propose a neural language model capable of unsupervised syntactic
structure induction. The model leverages the structure information to form
better semantic representations and better language modeling. Standard
recurrent neural networks are limited by their structure and fail to
efficiently use syntactic information. On the other hand, tree-structured
recursive networks usually require additional structural supervision at the
cost of human expert annotation. In this paper, We propose a novel neural
language model, called the Parsing-Reading-Predict Networks (PRPN), that can
simultaneously induce the syntactic structure from unannotated sentences and
leverage the inferred structure to learn a better language model. In our model,
the gradient can be directly back-propagated from the language model loss into
the neural parsing network. Experiments show that the proposed model can
discover the underlying syntactic structure and achieve state-of-the-art
performance on word/character-level language model tasks. | ['Chin-wei Huang', 'Zhouhan Lin', 'Yikang Shen', 'Aaron Courville'] | 2017-11-02 | neural-language-modeling-by-jointly-learning-1 | https://openreview.net/forum?id=rkgOLb-0W | https://openreview.net/pdf?id=rkgOLb-0W | iclr-2018-1 | ['constituency-grammar-induction'] | ['natural-language-processing'] | [ 3.22044641e-01 7.22539246e-01 -4.91476715e-01 -8.24900985e-01
-7.18616545e-01 -4.03558105e-01 1.40822053e-01 -6.54418394e-02
-4.21504229e-01 4.91132498e-01 4.74912435e-01 -8.03654909e-01
3.55762869e-01 -9.10691619e-01 -9.04200137e-01 -3.00262660e-01
1.00249425e-01 3.38217020e-01 -7.98386708e-02 1.28670409e-02
2.98810434e-02 -1.86823290e-02 -9.83105183e-01 6.06160820e-01
9.83675480e-01 7.36565173e-01 5.77221513e-01 4.57292467e-01
-7.72646308e-01 1.47732902e+00 -1.43852755e-01 -2.28686914e-01
-1.88200902e-02 -5.73597789e-01 -9.93126988e-01 -1.71632648e-01
-3.12347654e-02 -2.83519506e-01 -2.13383928e-01 1.07199407e+00
7.93905556e-02 -1.15208730e-01 2.31333494e-01 -4.45701212e-01
-7.19796240e-01 1.32056236e+00 -4.27387118e-01 9.41610932e-02
-6.93662390e-02 -1.62244916e-01 1.66946638e+00 -9.04376030e-01
5.83503067e-01 1.66745293e+00 4.62106556e-01 7.19693482e-01
-1.26513755e+00 -6.26021922e-01 8.37458730e-01 -7.53246918e-02
-6.55641735e-01 -3.77603620e-01 1.19118321e+00 -5.86208738e-02
1.23265314e+00 -3.29719514e-01 1.55262321e-01 1.29073215e+00
4.61725183e-02 1.31961441e+00 9.22921717e-01 -5.91787934e-01
9.50763524e-02 -1.63544893e-01 8.09140146e-01 1.16809690e+00
1.69387966e-01 -1.26778081e-01 -4.35458601e-01 2.82727573e-02
5.39483547e-01 -7.84230009e-02 -2.17895862e-02 -1.44935086e-01
-5.61683178e-01 1.08237338e+00 4.99506235e-01 2.94471115e-01
-2.06292659e-01 2.91152537e-01 4.86150116e-01 2.29558095e-01
6.69218481e-01 5.55161834e-01 -8.85469258e-01 -3.38385217e-02
-6.64383650e-01 -2.49665499e-01 6.79379582e-01 5.32233775e-01
8.18630219e-01 1.68063745e-01 1.02237165e-01 1.07577682e+00
4.64503497e-01 1.66360408e-01 7.09741950e-01 -7.50812948e-01
7.83677101e-01 7.62536108e-01 -5.58062077e-01 -5.63460112e-01
-5.35163045e-01 -3.99641871e-01 -7.20164180e-01 -3.70598257e-01
2.54981190e-01 -3.77470881e-01 -9.26966846e-01 1.98846948e+00
-1.17337525e-01 1.96743682e-01 2.81807810e-01 5.02558827e-01
7.85525918e-01 8.64881456e-01 3.63510340e-01 -6.49321601e-02
1.28739107e+00 -1.22990310e+00 -5.40521324e-01 -7.45867372e-01
1.14063334e+00 -1.92457989e-01 1.22675335e+00 6.91413879e-02
-9.97309566e-01 -4.64177877e-01 -7.71117806e-01 -2.57842124e-01
-8.85610208e-02 2.85215080e-01 6.76569462e-01 2.37140030e-01
-9.11594450e-01 5.18180609e-01 -1.08426559e+00 1.38418227e-02
3.89296979e-01 2.40839794e-01 -6.40677847e-03 -5.34759685e-02
-1.37789118e+00 6.44508481e-01 6.86176360e-01 1.45873979e-01
-6.75976336e-01 -5.54647028e-01 -1.18375146e+00 2.55063653e-01
4.87689078e-01 -7.55929708e-01 1.51052976e+00 -1.24220252e+00
-1.81458068e+00 8.05087268e-01 -5.96894503e-01 -7.57851720e-01
-9.08734053e-02 -3.25199842e-01 1.09105259e-01 -4.38244604e-02
-5.15257455e-02 6.90726936e-01 5.75926006e-01 -9.97589648e-01
-4.53492641e-01 -3.63945305e-01 7.59206414e-02 1.87631752e-02
-1.79417998e-01 1.22664712e-01 -2.35845700e-01 -7.46968746e-01
2.37198710e-01 -7.67755210e-01 -5.12888491e-01 -5.83888829e-01
-7.68406391e-01 -6.43148780e-01 5.93956053e-01 -8.16442370e-01
1.18534148e+00 -1.91495419e+00 2.84244895e-01 -2.10365597e-02
1.78564742e-01 2.50757307e-01 -5.10744572e-01 8.39572623e-02
-2.49071773e-02 3.53060186e-01 -3.19900751e-01 -5.37078917e-01
-1.71827376e-01 5.40464103e-01 -5.29759824e-01 -9.75571200e-02
4.96155679e-01 1.32113731e+00 -7.75559962e-01 -1.76879674e-01
-2.11488500e-01 2.73534536e-01 -7.44369447e-01 3.89509857e-01
-6.72238946e-01 3.69452864e-01 -6.59208119e-01 4.09665823e-01
2.33191729e-01 -4.65474278e-01 5.90206265e-01 1.38059407e-01
1.81411356e-01 1.23356342e+00 -4.64394867e-01 1.63492596e+00
-8.93339515e-01 2.84630358e-01 3.74537259e-02 -1.39990699e+00
9.97444987e-01 4.92229015e-02 -1.15964092e-01 -7.50747919e-01
1.71110481e-01 1.65513501e-01 3.01669359e-01 -2.91935563e-01
5.05869314e-02 -1.91759244e-01 -1.86947405e-01 6.18696570e-01
1.53228849e-01 4.59877491e-01 -1.26594096e-01 6.40961304e-02
1.12508917e+00 1.54718161e-01 1.99075520e-01 -2.24442601e-01
5.06850660e-01 -3.75318438e-01 9.26555395e-01 8.00154269e-01
2.22911745e-01 5.01668677e-02 7.94846117e-01 -5.06863654e-01
-8.80286098e-01 -8.24944198e-01 1.25421762e-01 1.71055365e+00
-2.72082537e-01 -4.54884678e-01 -8.64870429e-01 -1.07409585e+00
-2.29310930e-01 9.65228438e-01 -3.79040122e-01 -2.18826473e-01
-1.05740929e+00 -6.95925713e-01 7.05287337e-01 8.67955208e-01
4.30310577e-01 -1.49635231e+00 -2.19686836e-01 4.51540083e-01
-1.31947383e-01 -1.20992517e+00 -2.65809178e-01 4.57715303e-01
-1.32865143e+00 -7.63385177e-01 -1.88438818e-01 -1.21415687e+00
7.91239679e-01 -2.03942209e-01 1.28526533e+00 2.41007358e-01
1.97798908e-01 -1.81174889e-01 -3.29567194e-01 -2.07093924e-01
-6.44929171e-01 5.92161596e-01 -1.98566929e-01 8.73323239e-04
3.96574467e-01 -6.03348017e-01 -8.66797790e-02 -1.41462982e-01
-6.70597672e-01 3.65855068e-01 8.41099262e-01 1.11557090e+00
6.17466927e-01 -1.74944058e-01 8.34643066e-01 -1.56838071e+00
6.21237636e-01 -3.42218012e-01 -6.89873636e-01 3.01553458e-01
-6.21441901e-01 8.61991227e-01 8.85981321e-01 -2.53747791e-01
-1.41481590e+00 2.16762170e-01 -4.14565682e-01 -1.36507839e-01
-1.29210904e-01 9.14592266e-01 -3.20592403e-01 4.62934196e-01
1.92402065e-01 3.47612977e-01 -1.88878551e-01 -9.89045560e-01
5.57351470e-01 3.15664470e-01 4.04053330e-01 -8.05082440e-01
5.81536770e-01 8.14630240e-02 -3.77360851e-01 -6.25249803e-01
-1.63558185e+00 -1.73788011e-01 -7.13553488e-01 5.42075634e-01
8.92005205e-01 -8.80786300e-01 -2.97642112e-01 3.40455323e-01
-1.51098776e+00 -5.19732714e-01 5.88777736e-02 2.37592369e-01
-2.74063140e-01 3.83102775e-01 -1.21914756e+00 -8.32212090e-01
-6.20368659e-01 -9.42051113e-01 9.50485647e-01 -2.35480312e-02
-8.64883810e-02 -1.20184100e+00 -4.38915081e-02 3.56868058e-01
2.53934145e-01 -3.81491750e-01 1.59553349e+00 -9.43301618e-01
-7.17584729e-01 7.58059993e-02 -2.87338436e-01 4.99909639e-01
3.98381948e-02 -4.03607041e-01 -1.02200460e+00 -4.46563661e-02
1.20748274e-01 -5.29087603e-01 1.31506515e+00 4.49935794e-01
1.45262563e+00 -6.28075182e-01 -1.68609485e-01 6.08187854e-01
1.07912731e+00 -8.91411025e-03 4.51694041e-01 1.68486238e-01
9.03144956e-01 7.70664215e-01 6.33782595e-02 -6.55806288e-02
4.11287844e-01 2.43085146e-01 1.68680891e-01 -5.21348752e-02
4.55844887e-02 -8.91449153e-01 6.92775726e-01 1.42110038e+00
3.35078210e-01 -2.71252804e-02 -1.00532258e+00 3.15886497e-01
-1.78620732e+00 -5.20974219e-01 2.67627716e-01 1.60264599e+00
1.04567289e+00 5.84120989e-01 -2.98608273e-01 -2.97724277e-01
5.57670474e-01 3.67815703e-01 -6.32650197e-01 -6.61793172e-01
-8.89809877e-02 3.66043687e-01 3.95226657e-01 8.74636054e-01
-8.88336539e-01 1.59324348e+00 6.06880903e+00 5.55940628e-01
-1.05329025e+00 2.11697165e-02 8.37506652e-01 2.34348059e-01
-5.22695959e-01 2.12495476e-01 -1.24196339e+00 2.45708793e-01
1.23589921e+00 4.55403067e-02 3.28846961e-01 1.06447506e+00
1.97736278e-01 2.64182091e-01 -1.14534080e+00 5.86191177e-01
-1.22473173e-01 -1.40219259e+00 4.06293541e-01 4.68882695e-02
3.63732278e-01 4.00084406e-01 -2.30623305e-01 7.04542458e-01
7.18369067e-01 -1.14850271e+00 3.47841352e-01 3.07527840e-01
4.29994732e-01 -6.96997404e-01 7.90976584e-01 8.21071029e-01
-9.76149023e-01 -1.73270509e-01 -4.32448834e-01 -4.02103096e-01
2.83169925e-01 6.55114412e-01 -7.21176147e-01 1.83755279e-01
2.37519607e-01 7.85933137e-01 -5.36345482e-01 2.25339562e-01
-8.53239775e-01 1.25624192e+00 -2.14549825e-01 -3.50561142e-01
5.50919294e-01 -1.20697185e-01 2.41254687e-01 1.24998677e+00
-1.31892398e-01 -6.43649399e-02 5.95311761e-01 1.02900279e+00
-6.17077470e-01 2.15195432e-01 -5.56064785e-01 -2.85959393e-01
4.11425561e-01 9.36378062e-01 -5.38641095e-01 -3.43466461e-01
-5.75026929e-01 7.51418889e-01 1.06881332e+00 3.90321553e-01
-4.12883610e-01 -2.30543941e-01 5.31003654e-01 -1.24096908e-01
4.41088825e-01 -2.28566960e-01 -4.77470696e-01 -1.42609787e+00
-1.56921204e-02 -7.94630170e-01 3.56335849e-01 -4.00968641e-01
-1.41484272e+00 6.81844234e-01 -3.33002150e-01 -2.97992408e-01
-5.00204027e-01 -1.00648344e+00 -7.05652595e-01 8.28237772e-01
-1.78055620e+00 -1.06334293e+00 5.58784425e-01 8.86085853e-02
9.42311406e-01 -2.79324234e-01 9.72714245e-01 -1.78932905e-01
-8.72248948e-01 6.88843191e-01 -1.00333296e-01 8.06127489e-01
8.75918642e-02 -1.16315985e+00 6.90151036e-01 8.40347350e-01
4.60016459e-01 7.75010645e-01 1.70679584e-01 -5.04025996e-01
-1.23470211e+00 -1.29622769e+00 1.30423176e+00 -3.32366556e-01
8.16406071e-01 -7.05966413e-01 -1.15457880e+00 9.29629505e-01
3.91005687e-02 -1.26712427e-01 7.45066762e-01 6.92657411e-01
-8.14519703e-01 1.35075748e-01 -4.66775447e-01 4.45176065e-01
1.05672443e+00 -8.73028576e-01 -1.03266156e+00 5.98600134e-02
1.20696223e+00 2.17187628e-02 -2.50358433e-01 4.43426877e-01
2.36152828e-01 -5.10385394e-01 6.76948309e-01 -1.15694237e+00
7.41622984e-01 1.53815821e-01 -6.33159187e-03 -1.14130795e+00
-3.42508048e-01 -4.05245245e-01 -2.20973521e-01 1.10993350e+00
1.02432168e+00 -5.42298496e-01 8.94649208e-01 4.42397654e-01
-8.07237327e-02 -1.01729929e+00 -7.95553267e-01 -7.60970116e-01
4.03356165e-01 -7.31233060e-01 2.12417603e-01 8.06432545e-01
8.76960531e-02 1.25532317e+00 -2.78962255e-01 1.17329709e-01
4.35614735e-01 2.43735239e-01 3.57861668e-01 -1.28461862e+00
-7.18905747e-01 -4.66437548e-01 -3.43658868e-03 -1.67713058e+00
1.09714246e+00 -1.26060402e+00 2.57879555e-01 -1.45864785e+00
3.29150558e-01 -3.41704577e-01 -5.01722336e-01 7.88528860e-01
-3.57928306e-01 -4.54992950e-01 1.85197368e-01 1.80844124e-02
-7.57608950e-01 6.35807395e-01 9.00966763e-01 -5.13314549e-03
-2.19459727e-01 -4.72792005e-03 -8.81034970e-01 1.19403374e+00
9.16298568e-01 -6.49746418e-01 -4.19058233e-01 -8.41266930e-01
3.15291196e-01 9.51555297e-02 -1.61312725e-02 -3.74588430e-01
1.29233366e-02 8.71063024e-02 1.41235478e-02 -4.11728263e-01
-1.99078601e-02 -3.85393649e-01 -7.36425221e-01 4.09521848e-01
-8.46160948e-01 3.13378987e-05 1.32016921e-02 7.26374924e-01
-1.85597718e-01 -5.20334601e-01 6.71982825e-01 -4.08773869e-01
-5.38998365e-01 1.10677667e-01 -5.15158117e-01 2.86948323e-01
1.41839042e-01 4.05168623e-01 -1.45751163e-01 -3.45794320e-01
-6.64307654e-01 2.27479026e-01 -3.16936225e-02 5.97442448e-01
4.47171956e-01 -9.79234397e-01 -5.67748189e-01 1.77382424e-01
-2.32641771e-01 2.44726315e-01 -1.79104120e-01 3.03216159e-01
-1.00711584e-01 5.84948897e-01 4.42990720e-01 -3.92717481e-01
-9.91588950e-01 5.30422509e-01 3.22037011e-01 -9.97121334e-01
-7.59558976e-01 9.84676182e-01 6.36937857e-01 -8.56786370e-01
4.41496760e-01 -7.64075756e-01 -1.70543194e-01 -2.89167851e-01
4.87859964e-01 -2.63285160e-01 -2.29191124e-01 -3.21122110e-01
-1.33569419e-01 3.64352286e-01 -5.02200544e-01 1.29599497e-01
1.45806611e+00 5.81097156e-02 -3.63859177e-01 5.97921491e-01
1.31276393e+00 -8.20402354e-02 -1.20521677e+00 -7.52386212e-01
8.91413033e-01 3.44785452e-01 8.59358385e-02 -7.40776002e-01
-9.65217888e-01 1.17729199e+00 -3.92834693e-02 -2.32414007e-01
8.91833127e-01 3.17482769e-01 1.22124600e+00 1.02468324e+00
1.92032337e-01 -1.01515114e+00 1.04864247e-01 1.25326395e+00
4.40511853e-01 -1.10302842e+00 -5.46776891e-01 -6.24580026e-01
-4.31028545e-01 1.08391249e+00 6.40997350e-01 -3.44835281e-01
6.95512056e-01 6.13599002e-01 8.25802237e-02 -4.57091480e-02
-1.13691294e+00 -2.02117547e-01 1.19437158e-01 1.50067568e-01
6.97225809e-01 1.87667124e-02 -1.10588893e-01 1.25164473e+00
-1.92141518e-01 -1.72977462e-01 1.02430336e-01 7.22105384e-01
-7.56268144e-01 -1.44577909e+00 1.50476232e-01 3.35002869e-01
-5.77847481e-01 -5.70395350e-01 -5.76469541e-01 2.30694503e-01
-3.73577833e-01 6.65904760e-01 3.94552872e-02 -1.90752968e-01
2.38544479e-01 7.05400527e-01 1.25796705e-01 -1.01935124e+00
-3.05934966e-01 7.96244144e-02 4.00667965e-01 -5.47607124e-01
-2.72563368e-01 -2.47319579e-01 -1.61422503e+00 1.76942065e-01
-1.90458402e-01 2.12820664e-01 4.14160252e-01 1.24384367e+00
4.77584749e-01 5.08866429e-01 5.54634273e-01 -4.54873472e-01
-6.93836689e-01 -1.01870751e+00 -1.85739547e-01 2.33342990e-01
1.82817623e-01 -2.42440432e-01 -1.77327305e-01 5.36223799e-02] | [10.601807594299316, 9.271988868713379] |
7f7e7b2f-a693-4a01-9dff-f38804cfafed | deep-multimodal-feature-analysis-for-action | 1603.07120 | null | http://arxiv.org/abs/1603.07120v2 | http://arxiv.org/pdf/1603.07120v2.pdf | Deep Multimodal Feature Analysis for Action Recognition in RGB+D Videos | Single modality action recognition on RGB or depth sequences has been
extensively explored recently. It is generally accepted that each of these two
modalities has different strengths and limitations for the task of action
recognition. Therefore, analysis of the RGB+D videos can help us to better
study the complementary properties of these two types of modalities and achieve
higher levels of performance. In this paper, we propose a new deep autoencoder
based shared-specific feature factorization network to separate input
multimodal signals into a hierarchy of components. Further, based on the
structure of the features, a structured sparsity learning machine is proposed
which utilizes mixed norms to apply regularization within components and group
selection between them for better classification performance. Our experimental
results show the effectiveness of our cross-modality feature analysis framework
by achieving state-of-the-art accuracy for action classification on five
challenging benchmark datasets. | ['Tian-Tsong Ng', 'Gang Wang', 'Amir Shahroudy', 'Yihong Gong'] | 2016-03-23 | null | null | null | null | ['multimodal-activity-recognition'] | ['computer-vision'] | [ 4.03087556e-01 -5.17758787e-01 -3.69949847e-01 -5.00005245e-01
-6.71100438e-01 -2.03067511e-01 5.18255651e-01 -2.38460451e-01
-2.85769224e-01 4.39495474e-01 5.85026503e-01 1.95527390e-01
-2.58744419e-01 -4.51793730e-01 -3.16048592e-01 -9.61425483e-01
8.44953135e-02 -1.84896633e-01 1.94725057e-03 -9.79015231e-03
1.02577806e-01 4.06808794e-01 -1.83495319e+00 7.61497200e-01
7.27425396e-01 1.40412807e+00 -4.58290912e-02 3.85701567e-01
1.08746693e-01 1.18968916e+00 -1.21060885e-01 -1.14706725e-01
5.16409457e-01 -5.88851392e-01 -6.76755667e-01 5.00253439e-01
2.52943099e-01 -5.40624559e-01 -6.40200853e-01 8.57883096e-01
4.29590702e-01 2.77973711e-01 4.12386596e-01 -1.30264235e+00
-3.94880176e-01 3.62681538e-01 -6.02834463e-01 1.12178206e-01
6.81461155e-01 2.23219961e-01 1.06362343e+00 -7.37633407e-01
2.31304139e-01 1.00992775e+00 4.57678109e-01 4.37763691e-01
-1.06472731e+00 -5.99440157e-01 1.94382802e-01 5.46880782e-01
-1.23914230e+00 -3.88560086e-01 9.19809401e-01 -4.77294505e-01
9.60083544e-01 1.10187516e-01 7.24869430e-01 1.19092560e+00
1.27586931e-01 1.15849459e+00 1.30481088e+00 -2.44049519e-01
2.05266982e-01 -2.56568640e-01 -1.59459218e-01 6.39143646e-01
-1.62050039e-01 -2.37859085e-01 -9.91204500e-01 5.18916026e-02
6.94320858e-01 3.87615323e-01 -4.60323393e-01 -6.21966064e-01
-1.35297322e+00 8.20103049e-01 3.75764757e-01 4.65841889e-01
-5.34817994e-01 -3.56746726e-02 5.11033177e-01 4.65619843e-04
1.58014014e-01 4.63394187e-02 -4.32724863e-01 -4.82771009e-01
-5.47761381e-01 -7.73081407e-02 3.55312228e-01 3.53326440e-01
6.35150731e-01 -8.01928341e-03 -7.22677782e-02 8.83087456e-01
4.30272400e-01 4.29526061e-01 7.50705302e-01 -9.26006258e-01
4.92085665e-01 1.03481662e+00 -1.62154436e-01 -1.07049108e+00
-5.81191480e-01 -1.47753596e-01 -1.00110173e+00 2.43441045e-01
4.53747362e-01 1.84462130e-01 -7.79148281e-01 1.56518865e+00
3.00946593e-01 4.04880732e-01 2.70290971e-01 1.25991917e+00
7.66340554e-01 4.08499628e-01 6.12493716e-02 9.76325944e-02
1.19315970e+00 -7.97320843e-01 -6.75720572e-01 -2.37250747e-03
6.46785736e-01 -4.09515560e-01 8.83903027e-01 4.35956061e-01
-7.25411415e-01 -6.80471718e-01 -1.01209223e+00 -1.04232773e-01
-1.76964894e-01 4.00108933e-01 1.00617409e+00 4.88117069e-01
-4.57955360e-01 5.81763804e-01 -1.23035324e+00 -2.73874998e-01
6.70272946e-01 2.82545775e-01 -7.84995377e-01 -3.55607569e-01
-1.00560296e+00 5.17193019e-01 2.39062250e-01 1.15501709e-01
-7.99242198e-01 -5.25147319e-01 -9.70066130e-01 -1.57153115e-01
3.00165564e-01 -4.80446637e-01 7.24579513e-01 -1.11702621e+00
-1.69053185e+00 5.84794819e-01 1.29281044e-01 -3.17117631e-01
2.02245519e-01 -4.22353297e-01 -4.02054191e-01 4.36193794e-01
-2.73169093e-02 4.34036404e-01 8.99953187e-01 -8.59354794e-01
-7.70717978e-01 -7.85777092e-01 2.14058131e-01 3.40247393e-01
-7.03079104e-01 -9.15045142e-02 -3.65969360e-01 -6.78984225e-01
4.36406255e-01 -9.15634692e-01 -3.30584571e-02 6.07121438e-02
-1.19611368e-01 -2.07032755e-01 6.74349785e-01 -5.87042391e-01
1.04697764e+00 -2.40179110e+00 7.17735767e-01 1.49644360e-01
1.18734747e-01 3.63838524e-02 -1.07152998e-01 3.70659918e-01
-8.16776901e-02 -4.34155196e-01 -1.44361570e-01 -3.07469606e-01
-4.95474674e-02 4.30921733e-01 8.36191815e-04 6.78917527e-01
2.08095342e-01 7.11645246e-01 -6.40903771e-01 -1.90063432e-01
5.01192570e-01 7.34519303e-01 -4.30577219e-01 3.32004517e-01
3.25877160e-01 7.94085264e-01 -6.01916015e-01 9.18368578e-01
4.31236178e-01 -2.71082878e-01 2.29417875e-01 -5.84783852e-01
1.91989556e-01 6.51822090e-02 -1.27021289e+00 2.19188380e+00
-2.96705306e-01 3.87881726e-01 -1.46301478e-01 -1.18120706e+00
5.96657395e-01 2.41080463e-01 1.02799654e+00 -7.14967608e-01
4.29649293e-01 8.97856876e-02 -5.16719374e-05 -7.09639847e-01
1.48927346e-01 -1.38474837e-01 -7.17453063e-02 4.44714606e-01
3.22837532e-01 3.65882248e-01 1.19414382e-01 -1.71045102e-02
1.14747787e+00 3.29826951e-01 1.99592680e-01 1.71013892e-01
8.09979141e-01 -3.51661086e-01 6.58038557e-01 4.38899100e-01
-4.47648644e-01 6.55085206e-01 2.42319599e-01 -4.91671920e-01
-3.29404652e-01 -7.06137478e-01 -5.85892536e-02 1.05903494e+00
1.72610670e-01 -3.66476029e-01 -4.21638638e-01 -8.04173112e-01
1.18052825e-01 1.51198998e-01 -9.14979279e-01 -2.27313653e-01
-3.81581277e-01 -6.16687059e-01 4.83631194e-01 9.61252868e-01
7.12295353e-01 -9.53165352e-01 -8.57817233e-01 -1.63359255e-01
-2.89587677e-01 -1.45013082e+00 5.34815267e-02 2.29923680e-01
-9.48890626e-01 -1.39165592e+00 -7.70136297e-01 -3.90728652e-01
5.19752085e-01 5.33988118e-01 5.18951714e-01 -2.21855789e-01
-7.13396072e-02 8.86611104e-01 -8.84196103e-01 9.09740403e-02
1.24822445e-01 -2.15398669e-01 1.44023180e-01 7.39390075e-01
4.27002311e-01 -5.69257975e-01 -6.69307113e-01 2.04191461e-01
-1.12324095e+00 -1.07945167e-01 7.05588937e-01 8.09006631e-01
5.19252777e-01 6.33749440e-02 1.32456556e-01 -1.32197842e-01
2.28532955e-01 -4.13946807e-01 -8.63062143e-02 1.20885976e-01
-7.04055652e-02 1.58076603e-02 4.93822485e-01 -3.25948060e-01
-1.04892206e+00 3.37465703e-01 1.02035977e-01 -6.02345586e-01
-3.48310560e-01 7.41593838e-01 -4.10978645e-01 -2.60454625e-01
2.31305838e-01 4.67174292e-01 1.91182062e-01 -5.19548476e-01
2.55083740e-01 5.89904189e-01 2.81477928e-01 -2.89347947e-01
4.31528360e-01 7.00039089e-01 8.47713947e-02 -6.50699496e-01
-8.27492595e-01 -5.46682358e-01 -7.60073960e-01 -4.77988064e-01
1.09390521e+00 -1.16323352e+00 -7.12135792e-01 9.00283992e-01
-5.06758809e-01 -1.17655352e-01 -2.34100312e-01 8.18329036e-01
-5.70676982e-01 6.91125393e-01 -4.73881930e-01 -6.20266795e-01
3.86797674e-02 -1.25602281e+00 1.22943580e+00 2.97294021e-01
6.23529255e-02 -7.16088951e-01 1.76238894e-01 8.30591977e-01
1.71353340e-01 2.59055763e-01 4.55516845e-01 -5.57190478e-01
-3.95449668e-01 -3.41197938e-01 -8.37940499e-02 7.20191300e-01
3.85099500e-01 -3.97774696e-01 -8.73109341e-01 -3.12785685e-01
3.78490947e-02 -7.11434364e-01 1.15035319e+00 2.20477149e-01
1.21580005e+00 1.21413641e-01 -1.12344593e-01 6.48028135e-01
1.34563208e+00 3.54608186e-02 7.14047015e-01 3.61805737e-01
9.39204097e-01 4.40098494e-01 5.87831318e-01 8.80571187e-01
3.35819811e-01 7.09646523e-01 5.02605617e-01 1.88860580e-01
1.25375390e-01 4.82514128e-02 5.39745808e-01 5.82242370e-01
-3.81455541e-01 1.05352765e-02 -6.66675806e-01 3.90938699e-01
-2.10157108e+00 -8.96462500e-01 1.89443290e-01 2.01899099e+00
5.07688642e-01 -2.01709002e-01 3.45835954e-01 5.62635064e-01
3.15706879e-01 4.21891659e-01 -4.35828924e-01 1.00993432e-01
-2.56439924e-01 2.11302564e-01 1.98431954e-01 -1.23329662e-01
-1.38633597e+00 4.89946365e-01 5.41157103e+00 6.22251093e-01
-1.16583991e+00 -6.43821955e-02 3.36756259e-01 -1.97440788e-01
1.85596436e-01 -1.78037152e-01 -4.13768262e-01 4.53411937e-01
5.61537683e-01 3.82035106e-01 3.60438317e-01 6.62868381e-01
9.98794660e-02 -3.09018165e-01 -1.01610601e+00 1.31275809e+00
3.36948425e-01 -9.97886777e-01 2.64025871e-02 -5.21995798e-02
7.05001831e-01 -9.74754244e-02 1.15557969e-01 1.84968069e-01
-9.12946612e-02 -9.74355102e-01 3.71750444e-01 7.22228229e-01
2.81234503e-01 -7.30978847e-01 8.07173967e-01 1.52545765e-01
-1.10117865e+00 -4.75173414e-01 -1.27523243e-01 -3.88429999e-01
1.82577848e-01 3.62504929e-01 5.71896136e-02 8.50419283e-01
9.65986013e-01 1.30349243e+00 -5.58883309e-01 7.74917781e-01
-1.45442355e-02 3.53390366e-01 -2.41845042e-01 3.30896318e-01
1.39452398e-01 -1.63403422e-01 2.50321150e-01 8.15744877e-01
2.28013262e-01 4.37554449e-01 3.77283037e-01 2.56774396e-01
2.41247073e-01 1.48766607e-01 -3.88202369e-01 -3.13311934e-01
-1.70304716e-01 1.26457417e+00 -5.71733356e-01 -1.12361796e-01
-9.11238611e-01 1.11502039e+00 1.67656437e-01 1.63068399e-01
-8.90687108e-01 -6.34951591e-02 9.31974709e-01 -2.96421021e-01
7.55865216e-01 -3.66701782e-01 -1.53905347e-01 -1.62812543e+00
8.01023990e-02 -1.13509631e+00 5.95473349e-01 -5.71941197e-01
-1.24411976e+00 3.40694129e-01 -1.91978455e-01 -1.68990815e+00
3.14460322e-02 -8.90394211e-01 -2.20909789e-01 4.25699621e-01
-1.31764305e+00 -1.32573795e+00 -6.10808492e-01 1.04278731e+00
3.02115053e-01 -4.49309140e-01 8.86505485e-01 3.14725965e-01
-7.79929101e-01 4.42688137e-01 9.90708396e-02 3.45731050e-01
5.60430944e-01 -8.65994990e-01 -6.31651103e-01 8.49696398e-01
4.03661698e-01 3.24347764e-01 1.70911759e-01 -3.75968993e-01
-1.80671895e+00 -7.94014096e-01 3.13512743e-01 -1.99266642e-01
5.96323431e-01 7.76683865e-03 -6.77615821e-01 5.01262486e-01
8.26477781e-02 2.22164676e-01 1.27425385e+00 6.90718517e-02
-4.27631348e-01 -3.36308330e-01 -8.78433585e-01 1.77208170e-01
9.25156295e-01 -6.94580197e-01 -4.48012948e-01 1.78948030e-01
1.11773431e-01 -2.59789318e-01 -1.15853930e+00 5.94376504e-01
8.68519425e-01 -1.39408696e+00 8.26723695e-01 -8.01349163e-01
7.77117372e-01 -4.13577467e-01 -7.10300803e-01 -1.03216207e+00
-2.46679008e-01 -2.22074781e-02 -4.32510018e-01 1.02144063e+00
-1.13400497e-01 -4.58151191e-01 7.71266341e-01 4.97865587e-01
9.84580591e-02 -9.99178171e-01 -1.02667832e+00 -5.94781160e-01
-3.37113827e-01 -7.35246897e-01 3.18978429e-01 8.82245779e-01
1.32457659e-01 1.74592257e-01 -6.40691161e-01 9.16951224e-02
4.82534379e-01 3.68661344e-01 8.24479342e-01 -9.54174876e-01
-4.44686174e-01 -2.25625440e-01 -9.79413807e-01 -1.09535992e+00
2.68198192e-01 -7.44583666e-01 -1.65786624e-01 -1.36233795e+00
3.04895371e-01 -3.27348746e-02 -8.49879265e-01 7.20654964e-01
-8.72579813e-02 5.45251131e-01 3.12809139e-01 2.27675393e-01
-9.37985420e-01 1.04346645e+00 1.03950179e+00 -1.77114844e-01
-1.09406419e-01 -1.30535603e-01 -6.11494839e-01 6.47209048e-01
5.48783302e-01 -1.91476971e-01 -3.61555099e-01 -4.51108277e-01
-1.60919912e-02 -5.70920715e-03 4.50092882e-01 -1.33918655e+00
3.99199175e-03 -1.96509883e-01 6.13152444e-01 -4.19056743e-01
5.50127447e-01 -1.06041193e+00 -9.94987935e-02 3.11898261e-01
-3.36086035e-01 -4.74512905e-01 5.97417429e-02 7.69968808e-01
-6.19662702e-01 2.45047078e-01 5.23709059e-01 -1.59193464e-02
-9.85424697e-01 1.73740089e-01 -2.42725581e-01 -2.26424262e-01
1.07671416e+00 -3.76742929e-01 1.53081313e-01 -4.54890132e-01
-7.56206810e-01 9.03661400e-02 2.79265493e-01 6.65041924e-01
7.90970147e-01 -1.68265474e+00 -4.51992780e-01 3.46720099e-01
3.75256151e-01 -4.67822969e-01 7.28455424e-01 1.21758389e+00
-6.58900663e-02 3.48065943e-01 -7.13353813e-01 -8.52781594e-01
-1.25071621e+00 3.14921528e-01 2.83012271e-01 -2.62047648e-01
-5.39362907e-01 7.43702710e-01 1.23556919e-01 -1.89496741e-01
2.82649964e-01 -2.30441749e-01 -5.12965679e-01 2.05836505e-01
6.74676955e-01 4.17124718e-01 -9.09983814e-02 -1.10366726e+00
-6.36264563e-01 7.26852238e-01 1.06204554e-01 4.36539017e-02
1.58004022e+00 -9.52379927e-02 5.86616760e-03 5.67016602e-01
1.27411640e+00 -3.31768841e-01 -1.55392396e+00 -2.48786151e-01
-3.13371301e-01 -8.88314128e-01 9.72888842e-02 -5.31845868e-01
-1.43784797e+00 7.73392558e-01 8.07029963e-01 6.21957751e-03
1.66539764e+00 -1.12112753e-01 8.19956839e-01 2.18475685e-01
2.93042064e-01 -1.07107472e+00 3.83466989e-01 3.77662301e-01
8.45498741e-01 -1.52553296e+00 2.35125884e-01 -2.30155200e-01
-9.09691274e-01 1.24724472e+00 5.05214810e-01 -2.12154999e-01
6.55023575e-01 -1.08057328e-01 5.35843708e-02 -1.48347169e-01
-3.65522385e-01 -4.43004936e-01 5.59328973e-01 5.52580953e-01
3.23692083e-01 -3.64352949e-02 -2.50049025e-01 7.96012998e-01
3.53897989e-01 1.67930305e-01 9.04890075e-02 1.22983634e+00
-1.31106898e-01 -1.16537356e+00 -1.57823831e-01 3.12599689e-01
-4.84236419e-01 3.09839368e-01 -3.78555477e-01 5.98707557e-01
1.87495410e-01 9.42065537e-01 -1.98684022e-01 -7.77980566e-01
2.67755598e-01 1.83382139e-01 7.57853150e-01 -2.91866362e-01
-3.96473765e-01 2.76438519e-02 -1.23613819e-01 -1.15484822e+00
-1.18818092e+00 -8.64917040e-01 -1.03878379e+00 2.17390612e-01
-1.99455932e-01 -3.43399197e-01 4.98517901e-01 1.25469208e+00
5.50803602e-01 4.46330428e-01 6.57700956e-01 -9.79905427e-01
-5.50495327e-01 -8.40363264e-01 -7.65161812e-01 7.66909420e-01
2.08289236e-01 -1.00750899e+00 -3.28635871e-01 6.97764456e-02] | [7.978432655334473, 0.43726420402526855] |
49cd9d1b-fdc3-4389-b59d-035a856a476b | deep-active-learning-with-structured-neural | 2306.02808 | null | https://arxiv.org/abs/2306.02808v1 | https://arxiv.org/pdf/2306.02808v1.pdf | Deep Active Learning with Structured Neural Depth Search | Previous work optimizes traditional active learning (AL) processes with incremental neural network architecture search (Active-iNAS) based on data complexity change, which improves the accuracy and learning efficiency. However, Active-iNAS trains several models and selects the model with the best generalization performance for querying the subsequent samples after each active learning cycle. The independent training processes lead to an insufferable computational budget, which is significantly inefficient and limits search flexibility and final performance. To address this issue, we propose a novel active strategy with the method called structured variational inference (SVI) or structured neural depth search (SNDS) whereby we could use the gradient descent method in neural network depth search during AL processes. At the same time, we theoretically demonstrate that the current VI-based methods based on the mean-field assumption could lead to poor performance. We apply our strategy using three querying techniques and three datasets and show that our strategy outperforms current methods. | ['Jianwei Zhang', 'Xieyi Ping', 'Xiaoyun Zhang'] | 2023-06-05 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 3.18060040e-01 2.71845460e-01 -5.10947406e-01 -4.37106043e-01
-9.99755025e-01 -3.75533223e-01 6.11588180e-01 -1.08641326e-01
-7.20526159e-01 9.39733922e-01 -3.76205206e-01 -2.95354277e-01
-4.78138119e-01 -8.22905362e-01 -8.10262859e-01 -9.34444666e-01
2.17559919e-01 7.71928549e-01 6.42252386e-01 4.78825152e-01
5.05954802e-01 5.59894145e-01 -1.67690587e+00 6.95996732e-02
1.17732501e+00 1.09354424e+00 3.43550265e-01 4.79389668e-01
-6.73725128e-01 9.63797748e-01 -5.52216649e-01 -3.83117110e-01
2.82274485e-01 -3.25266302e-01 -7.18434811e-01 -4.96385604e-01
4.17107403e-01 -3.03370208e-01 9.32761803e-02 8.26739192e-01
5.30340672e-01 5.68243742e-01 4.05075788e-01 -1.17818654e+00
-3.72679412e-01 6.24050915e-01 -2.99537331e-01 2.00823098e-01
-2.69916534e-01 1.62722096e-02 8.56177688e-01 -1.15444815e+00
5.24950862e-01 1.12758934e+00 7.67871022e-01 1.07527626e+00
-1.35178852e+00 -7.31688142e-01 3.68482143e-01 3.61056089e-01
-1.25939393e+00 -8.39021087e-01 9.38997924e-01 -7.61284605e-02
9.88867879e-01 3.45356941e-01 8.64300132e-01 1.11580813e+00
-1.54280484e-01 1.29873168e+00 6.90296531e-01 -6.71333492e-01
9.52020407e-01 1.65785015e-01 3.18126321e-01 1.02913463e+00
3.88872981e-01 1.41363591e-01 -1.07886410e+00 -5.56231320e-01
7.66202509e-01 1.16487471e-02 -1.05409874e-02 -6.20382309e-01
-6.12036407e-01 1.01556897e+00 8.42324495e-02 -7.20905736e-02
-3.32636863e-01 2.99991548e-01 1.29056685e-02 2.32482955e-01
6.95147872e-01 5.14060736e-01 -4.73169476e-01 -1.56183854e-01
-1.48294854e+00 2.89910436e-01 8.09901655e-01 5.82204401e-01
9.00393844e-01 2.31584221e-01 -1.36008754e-01 9.91969943e-01
9.08793032e-01 7.38107860e-02 1.67437449e-01 -1.51781774e+00
3.31518799e-01 7.95939267e-01 3.46481875e-02 -3.71148199e-01
3.25220041e-02 -5.08472681e-01 -4.74921048e-01 5.72372556e-01
3.15481633e-01 -2.17682242e-01 -8.62956226e-01 1.45028770e+00
4.22913730e-01 1.48070842e-01 -1.26198739e-01 3.54636043e-01
8.09687436e-01 6.56030416e-01 3.75552997e-02 -7.74714410e-01
4.18032616e-01 -1.14170957e+00 -7.52219081e-01 -2.08681405e-01
4.68469888e-01 -2.44806707e-01 9.27922547e-01 6.64158523e-01
-1.76115704e+00 -4.54320669e-01 -1.06092095e+00 1.98304608e-01
-2.56169826e-01 -2.26641297e-01 8.41619253e-01 6.91806018e-01
-1.33268905e+00 5.98473310e-01 -1.45964742e+00 3.48963626e-02
7.55035102e-01 6.26416624e-01 2.50642776e-01 5.28888226e-01
-9.08228338e-01 4.59341556e-01 3.83090556e-01 2.00514033e-01
-1.01388240e+00 -8.65618050e-01 -4.64691520e-01 1.28361985e-01
6.76868916e-01 -6.65873587e-01 1.27931666e+00 -1.04305696e+00
-1.74754894e+00 3.40899765e-01 -5.86404860e-01 -8.04766774e-01
4.90690559e-01 -1.92518339e-01 1.19304650e-01 1.99789777e-02
-5.03523350e-01 8.55695188e-01 8.13358188e-01 -1.33623064e+00
-4.41885084e-01 -5.02107561e-01 4.37118895e-02 2.70555288e-01
-4.98867393e-01 -2.05630913e-01 -6.48478985e-01 -2.43103713e-01
3.94149989e-01 -6.85911298e-01 -2.80665725e-01 2.66916096e-01
-1.90416008e-01 -5.55589974e-01 7.54238367e-01 -2.29900721e-02
1.48048604e+00 -1.95468402e+00 1.43546984e-01 2.77794361e-01
2.99542814e-01 3.21780533e-01 1.41621962e-01 1.69486642e-01
4.79043275e-01 2.31621429e-01 -2.73431182e-01 -8.19476008e-01
-1.19059317e-01 4.83060956e-01 -3.43594134e-01 1.06627401e-02
1.85480490e-02 9.41332698e-01 -6.70589328e-01 -9.20118690e-01
-1.98386461e-01 2.01409698e-01 -7.37748086e-01 2.62828112e-01
-3.79302889e-01 2.53750026e-01 -4.40961242e-01 6.46183372e-01
5.11408269e-01 -5.41980922e-01 8.67156088e-02 1.68685559e-02
-1.82746440e-01 2.61067957e-01 -1.20954537e+00 1.65986013e+00
-2.46396139e-01 6.06411099e-01 1.29082233e-01 -1.15828001e+00
8.68882179e-01 1.21162936e-01 4.69288588e-01 -7.27570415e-01
-4.64184880e-01 3.00789267e-01 -1.89206272e-01 -1.86375648e-01
3.47080708e-01 5.19314408e-01 7.16149688e-01 3.63122374e-01
9.31364298e-02 3.80933523e-01 1.12738207e-01 1.98103249e-01
8.86067331e-01 4.44143385e-01 -8.91507324e-03 -2.33583689e-01
2.68461496e-01 -3.03651635e-02 7.29241014e-01 1.50386465e+00
-2.80053198e-01 3.54817986e-01 2.59149820e-01 -2.51398772e-01
-7.20183849e-01 -1.36520696e+00 -2.04524979e-01 1.30672312e+00
-7.83587471e-02 -2.92130977e-01 -7.45159566e-01 -8.03360224e-01
-2.66461015e-01 8.72017205e-01 -6.38863087e-01 -1.84200257e-02
-6.03509426e-01 -8.22862387e-01 5.87421119e-01 4.85112637e-01
6.33638561e-01 -1.08070993e+00 -7.93369412e-01 1.16802134e-01
3.25991809e-01 -1.78950235e-01 2.16972530e-01 5.52494586e-01
-1.45479238e+00 -6.80616915e-01 -6.87977433e-01 -4.37809974e-01
5.86791575e-01 -3.31788629e-01 1.10138452e+00 1.99329127e-02
-3.85329016e-02 1.42372504e-01 -1.44074187e-01 -5.74286938e-01
-1.89453125e-01 4.31559205e-01 -1.53767809e-01 -2.45608106e-01
3.45362663e-01 -4.46203530e-01 -8.24129641e-01 -1.58459604e-01
-6.32488370e-01 1.11916840e-01 4.27820086e-01 8.89708757e-01
8.14507246e-01 -1.35237038e-01 5.50104976e-01 -1.16272783e+00
6.13788247e-01 -3.70180815e-01 -9.71454442e-01 5.74981391e-01
-1.48635817e+00 3.02453518e-01 2.79764473e-01 -4.90438521e-01
-1.52932727e+00 1.07312948e-01 -1.30956367e-01 -5.12485981e-01
1.48447454e-01 5.02874792e-01 1.48298159e-01 -1.10835582e-01
8.81358504e-01 3.36967885e-01 -8.85994732e-02 -5.80950081e-01
-1.17775671e-01 3.03793579e-01 3.99711281e-02 -3.99137706e-01
4.17638063e-01 4.74423051e-01 4.30360213e-02 -5.04208744e-01
-9.62837577e-01 -1.73854873e-01 -4.65557367e-01 -2.53348321e-01
4.97564495e-01 -5.47778547e-01 -6.91846907e-01 5.36943793e-01
-1.08623004e+00 -6.50015056e-01 -4.94534135e-01 6.00616872e-01
-4.70677853e-01 1.79433867e-01 -6.22801781e-01 -1.46881437e+00
-4.87452835e-01 -1.18489432e+00 7.11379170e-01 2.63733268e-01
-2.09202394e-01 -1.22981822e+00 3.12980294e-01 3.49107862e-01
5.33221304e-01 -3.19520801e-01 8.63533854e-01 -9.01462793e-01
-9.80678558e-01 3.49433243e-01 1.63692117e-01 9.83643904e-02
-3.06013733e-01 2.85184592e-01 -1.08463335e+00 -2.11986154e-01
1.82582922e-02 -5.23351610e-01 1.06076384e+00 6.04206026e-01
1.48345721e+00 -2.67749995e-01 -2.75354683e-01 7.51606762e-01
1.46963418e+00 6.63795888e-01 6.83592141e-01 2.42198601e-01
4.72104847e-01 4.01041627e-01 3.82818401e-01 2.25352615e-01
-2.18029410e-01 4.32795644e-01 3.08304995e-01 -8.08246434e-02
2.03921303e-01 -1.11991704e-01 4.49655026e-01 8.32327724e-01
-2.24288940e-01 -4.17047411e-01 -1.11769438e+00 3.02739233e-01
-2.07458973e+00 -1.11614442e+00 1.35905415e-01 2.30571151e+00
1.16564584e+00 4.86605287e-01 -2.64159203e-01 1.08379044e-01
2.66723990e-01 1.93706632e-01 -1.07998919e+00 -4.60508734e-01
5.99560998e-02 4.08279091e-01 4.27289486e-01 6.26947522e-01
-6.36250257e-01 6.60037339e-01 7.03629303e+00 1.11691272e+00
-9.23241436e-01 3.62474501e-01 5.45175970e-01 -6.05349243e-01
-3.62027556e-01 6.46464601e-02 -1.27452838e+00 3.10503572e-01
1.00611234e+00 2.98634171e-01 3.97453427e-01 9.28177059e-01
1.01986155e-01 -2.96581656e-01 -1.08054996e+00 9.58762169e-01
-3.80641259e-02 -1.79291511e+00 2.26641968e-01 1.62358359e-02
8.43887150e-01 1.33006811e-01 1.32378796e-02 3.01513195e-01
3.37426335e-01 -8.20541441e-01 6.03223205e-01 9.75625992e-01
4.25728679e-01 -7.15078056e-01 2.27322891e-01 7.57605255e-01
-8.75666559e-01 -1.57754928e-01 -2.68645823e-01 1.89490601e-01
3.03818583e-02 4.03991818e-01 -5.86549819e-01 1.05319835e-01
6.66019320e-01 3.14102709e-01 -5.81878006e-01 1.12180459e+00
2.49986842e-01 1.27420151e+00 -4.38598782e-01 -2.86061406e-01
2.18600541e-01 -2.74412274e-01 6.85488105e-01 7.24474251e-01
1.03649586e-01 -4.92129624e-01 -5.07124178e-02 1.24758744e+00
8.64619911e-02 -1.44068658e-01 -2.42086500e-01 -8.66204500e-02
9.67681706e-01 7.40150094e-01 -7.37639666e-01 -3.89085591e-01
-3.06984544e-01 8.10292065e-01 5.97918689e-01 6.32812202e-01
-5.38952529e-01 -1.22933239e-01 -7.14139491e-02 3.60322185e-02
1.52158916e-01 -6.41188249e-02 -4.64983493e-01 -9.23875690e-01
-1.69968143e-01 -4.86298829e-01 3.62999737e-01 -5.74525356e-01
-1.11408532e+00 1.90462351e-01 3.24571222e-01 -5.97256839e-01
-5.84371209e-01 -2.46403888e-01 -5.88809371e-01 8.62486959e-01
-1.14312422e+00 -6.67067230e-01 -8.12096894e-02 5.15885532e-01
9.03107226e-01 -4.17095691e-01 7.04807520e-01 2.27929011e-01
-7.06723392e-01 8.06642592e-01 4.87680435e-01 -2.13723123e-01
3.11367273e-01 -1.17046905e+00 3.44900668e-01 6.47699893e-01
3.21190089e-01 1.01599789e+00 3.72644991e-01 -6.57580733e-01
-1.30203032e+00 -8.55733097e-01 7.15990663e-01 -2.32338920e-01
3.30364674e-01 -4.76294190e-01 -1.10569644e+00 5.16839743e-01
1.82472088e-03 -3.06499541e-01 7.05263436e-01 3.56709480e-01
4.24023382e-02 -2.05019102e-01 -9.57672656e-01 6.26430750e-01
1.20817804e+00 -4.96274263e-01 -4.85990912e-01 6.79363087e-02
5.27317762e-01 -5.86503446e-02 -5.55829942e-01 5.57297289e-01
6.47097290e-01 -1.14092875e+00 9.81850445e-01 -4.73126441e-01
-1.24217784e-02 1.48943812e-01 3.63596939e-02 -7.28758872e-01
-1.35220990e-01 -4.93625551e-01 -8.46616149e-01 1.10295808e+00
8.92336726e-01 -8.87064993e-01 1.41967165e+00 6.80915475e-01
5.89296259e-02 -1.42272365e+00 -9.14116502e-01 -6.09010160e-01
7.70230666e-02 -4.95888978e-01 3.26684028e-01 7.04135776e-01
-6.26610219e-01 1.76682040e-01 1.33782690e-02 -2.70961791e-01
8.84107411e-01 -9.12684724e-02 2.85028607e-01 -1.73807847e+00
-5.84691286e-01 -2.38010347e-01 2.43840650e-01 -1.28771985e+00
1.09249860e-01 -4.56757367e-01 3.62208039e-02 -1.51656771e+00
2.92908132e-01 -7.91310966e-01 -5.29504895e-01 4.00040656e-01
-7.76278377e-02 4.76059727e-02 -8.66817608e-02 5.90239346e-01
-7.91872680e-01 4.86941397e-01 9.35609519e-01 -3.26780677e-02
-4.88767505e-01 1.23810604e-01 -1.41357809e-01 7.74308443e-01
7.08390474e-01 -4.67481911e-01 -1.10114229e+00 -3.72185349e-01
7.83405542e-01 8.06645975e-02 3.79479796e-01 -8.87142181e-01
8.06187034e-01 -2.53403068e-01 5.62440038e-01 -1.04799831e+00
6.91720307e-01 -5.80691397e-01 2.32099220e-01 5.76160431e-01
-9.20487523e-01 -1.69291288e-01 -1.71850741e-01 6.24303401e-01
-6.17110543e-02 -8.42262208e-01 6.75582647e-01 -2.97146499e-01
-5.76179385e-01 4.10412669e-01 -3.96989435e-01 4.18708706e-03
6.36702895e-01 -7.61991084e-01 -9.57661346e-02 -7.17800260e-02
-7.16402113e-01 5.93726560e-02 3.38270605e-01 -5.91107234e-02
6.15254581e-01 -1.01942921e+00 -3.05870891e-01 3.19969207e-01
-1.87442645e-01 4.90829349e-01 8.88948962e-02 7.80681014e-01
-4.39032733e-01 3.92215073e-01 3.49078298e-01 -7.34665334e-01
-1.17085385e+00 1.30214617e-01 5.12583435e-01 -3.30092430e-01
-1.60038620e-01 1.14848912e+00 -6.35295883e-02 -4.14890140e-01
7.61069298e-01 1.74888611e-01 -2.89288282e-01 2.30588198e-01
1.87019423e-01 7.60745883e-01 1.80911511e-01 1.41356051e-01
-1.33774132e-01 1.49575084e-01 -2.36456499e-01 -3.77530187e-01
1.06821239e+00 -1.19283244e-01 1.61725551e-01 1.08005619e+00
1.01680553e+00 3.91981639e-02 -1.57775187e+00 -3.26896012e-01
8.51953924e-02 -1.44151196e-01 6.02804661e-01 -7.25647151e-01
-1.04370201e+00 7.69290924e-01 9.78404045e-01 2.31738806e-01
8.85347068e-01 -1.21473305e-01 4.95494366e-01 9.26947296e-01
3.43846589e-01 -1.52038801e+00 -2.36226954e-02 3.52460146e-01
3.21762502e-01 -1.17290092e+00 6.08922876e-02 -1.14790499e-01
-2.78259814e-01 9.39383447e-01 6.85957015e-01 1.45235643e-01
6.87943220e-01 3.26300710e-01 -4.75963950e-02 -2.78902829e-01
-1.34860468e+00 2.22092003e-01 2.64971524e-01 2.20402434e-01
2.58951873e-01 -4.85106200e-01 -2.45752200e-01 8.32905844e-02
2.76117980e-01 2.20571786e-01 -2.35658497e-01 1.16330945e+00
-5.92512965e-01 -1.08244991e+00 -9.10393074e-02 6.95648491e-01
-3.44636142e-01 -1.27041325e-01 -4.56897885e-01 5.78372002e-01
7.21824467e-02 7.63916314e-01 5.25884748e-01 3.01104970e-02
-2.77750611e-01 4.95522887e-01 6.30758226e-01 -4.31312114e-01
-6.22622371e-01 -5.47294170e-02 1.46105932e-03 -6.35870218e-01
-7.30856776e-01 -6.94097638e-01 -1.19094205e+00 -1.26901507e-01
-6.96582198e-01 3.66109937e-01 6.31904662e-01 9.12725389e-01
4.37248766e-01 1.47380278e-01 3.80005598e-01 -3.21022630e-01
-7.34013319e-01 -7.89444923e-01 -2.82771856e-01 -1.71023473e-01
1.26191005e-01 -7.21195936e-01 -5.77914059e-01 -1.07968025e-01] | [8.996773719787598, 3.895050048828125] |
47ac765d-ce4a-452d-9615-d0f93ae02c58 | thesis-multiple-kernel-learning-for-object | 1604.03247 | null | http://arxiv.org/abs/1604.03247v1 | http://arxiv.org/pdf/1604.03247v1.pdf | Thesis: Multiple Kernel Learning for Object Categorization | Object Categorization is a challenging problem, especially when the images
have clutter background, occlusions or different lighting conditions. In the
past, many descriptors have been proposed which aid object categorization even
in such adverse conditions. Each descriptor has its own merits and de-merits.
Some descriptors are invariant to transformations while the others are more
discriminative. Past research has shown that, employing multiple descriptors
rather than any single descriptor leads to better recognition. The problem of
learning the optimal combination of the available descriptors for a particular
classification task is studied. Multiple Kernel Learning (MKL) framework has
been developed for learning an optimal combination of descriptors for object
categorization. Existing MKL formulations often employ block l-1 norm
regularization which is equivalent to selecting a single kernel from a library
of kernels. Since essentially a single descriptor is selected, the existing
formulations maybe sub- optimal for object categorization. A MKL formulation
based on block l-infinity norm regularization has been developed, which chooses
an optimal combination of kernels as opposed to selecting a single kernel. A
Composite Multiple Kernel Learning(CKL) formulation based on mixed l-infinity
and l-1 norm regularization has been developed. These formulations end in
Second Order Cone Programs(SOCP). Other efficient alter- native algorithms for
these formulation have been implemented. Empirical results on benchmark
datasets show significant improvement using these new MKL formulations. | ['Dinesh Govindaraj'] | 2016-04-12 | null | null | null | null | ['object-categorization'] | ['computer-vision'] | [ 4.36957069e-02 -6.96485937e-01 -4.28442150e-01 -5.17388999e-01
-8.52265298e-01 -4.43164289e-01 5.55420935e-01 3.21197897e-01
-4.52668697e-01 4.54936892e-01 -2.24597961e-01 1.76084116e-01
-6.04957283e-01 -5.42148829e-01 -1.87816203e-01 -9.80533421e-01
-1.25780240e-01 1.36106089e-01 3.23679149e-01 -4.43032719e-02
6.04383111e-01 8.17996502e-01 -1.72026384e+00 2.36658826e-01
7.94218838e-01 1.08056772e+00 2.21218109e-01 5.64946115e-01
-2.32226014e-01 4.93698925e-01 -2.77262092e-01 -8.90486538e-02
5.41364908e-01 -1.04507558e-01 -5.35646439e-01 1.77977487e-01
6.80907309e-01 1.37141198e-01 1.56500220e-01 1.14261079e+00
3.92359346e-01 4.26164448e-01 1.03467882e+00 -1.51430285e+00
-7.32011795e-01 -9.58148241e-02 -6.46677017e-01 2.41521910e-01
3.46362263e-01 -2.81120420e-01 8.42245579e-01 -1.08587801e+00
1.29428446e-01 1.47162950e+00 7.82826543e-01 8.86813998e-02
-1.38877928e+00 -5.21950006e-01 1.99839845e-01 4.85219330e-01
-1.73883510e+00 -1.41106695e-01 9.69904125e-01 -5.01764238e-01
8.12835217e-01 3.82564604e-01 2.29995012e-01 5.01605809e-01
5.39116859e-01 7.64068902e-01 1.50210595e+00 -6.56583488e-01
2.65645273e-02 5.51063418e-01 7.94187486e-01 6.23851836e-01
3.46462220e-01 1.60948578e-02 -2.87553281e-01 -5.25046110e-01
3.31448883e-01 1.04466528e-01 6.61233962e-02 -5.78528166e-01
-8.98585141e-01 9.75726783e-01 1.88666314e-01 2.00273976e-01
-3.03135425e-01 -1.53507903e-01 6.56121910e-01 3.25955302e-01
2.51323104e-01 3.04973334e-01 -2.24283203e-01 3.82903099e-01
-7.40375638e-01 3.64290208e-01 6.59131229e-01 8.18847537e-01
9.52853084e-01 4.96441498e-02 -2.72191107e-01 1.27920759e+00
4.41782266e-01 3.79820079e-01 4.00597155e-01 -5.90638816e-01
2.38443807e-01 8.39651465e-01 3.96875478e-02 -1.29297912e+00
-3.34960848e-01 -1.95018962e-01 -7.63687253e-01 5.96719146e-01
2.37025529e-01 2.86293685e-01 -7.84731507e-01 1.41364503e+00
3.97428662e-01 2.77687013e-01 2.01555654e-01 8.01122129e-01
6.48935199e-01 6.58400059e-01 3.29197019e-01 -1.51612401e-01
1.20406997e+00 -8.25006783e-01 -5.93348682e-01 9.95222330e-02
5.49282610e-01 -1.17757082e+00 8.49359810e-01 5.99323392e-01
-5.60971797e-01 -8.30655694e-01 -1.06414604e+00 1.17727168e-01
-9.21825945e-01 3.97304863e-01 6.50110006e-01 6.70953095e-01
-7.92410016e-01 3.01106602e-01 -4.52022702e-01 -3.43044400e-01
2.60348283e-02 7.35489905e-01 -5.70024073e-01 -9.92440432e-02
-8.52484882e-01 1.09304917e+00 6.19539142e-01 1.08242683e-01
-3.61638576e-01 -4.20496732e-01 -7.31582284e-01 -2.79991776e-01
2.08622515e-01 -5.34448922e-02 7.46211469e-01 -1.08046353e+00
-1.24094176e+00 7.72761583e-01 -1.00960232e-01 -4.90329340e-02
2.79899985e-01 -2.55240630e-02 -4.98437673e-01 1.59684539e-01
2.19930410e-01 4.66787040e-01 1.06085277e+00 -1.26417863e+00
-6.16574705e-01 -3.49102616e-01 3.67563702e-02 2.94966072e-01
-4.43940759e-01 2.24670500e-01 -1.26282528e-01 -7.27911830e-01
2.42646620e-01 -1.10015821e+00 -1.40677646e-01 1.52526855e-01
-7.31185749e-02 -6.47731245e-01 1.34359610e+00 -4.44152385e-01
1.12882376e+00 -2.11572409e+00 1.59388259e-01 4.32941884e-01
-2.60449439e-01 6.32895350e-01 -1.03745423e-01 4.18894470e-01
-2.23748475e-01 -1.76698253e-01 -1.16025478e-01 -1.27033547e-01
-1.12848625e-01 3.19589287e-01 -1.93690718e-03 7.08312988e-01
2.04802707e-01 2.94066072e-01 -5.32967746e-01 -6.69567168e-01
4.69558895e-01 5.48382461e-01 -2.73407727e-01 7.81074315e-02
3.63426954e-01 -5.26861772e-02 -6.64131999e-01 8.74908328e-01
1.09532702e+00 1.04567595e-01 -3.33301842e-01 -5.10400414e-01
-2.40366161e-01 -5.42503238e-01 -1.90242779e+00 1.24353004e+00
-4.97528076e-01 4.14827347e-01 -1.46807190e-02 -1.37477279e+00
1.21201921e+00 2.59500146e-01 6.71658874e-01 -2.91991681e-01
2.83940695e-02 3.03087234e-01 -1.38811573e-01 -6.26766980e-01
1.89504072e-01 -2.78810989e-02 7.88419768e-02 5.04496954e-02
5.73929958e-02 8.09926689e-02 3.67747337e-01 -9.18871984e-02
8.80500317e-01 -7.98396990e-02 6.48522735e-01 -5.26929855e-01
1.11294031e+00 9.72626731e-02 5.55953622e-01 7.39501476e-01
-3.73464733e-01 4.12008822e-01 9.69235003e-02 -6.96851134e-01
-7.89597332e-01 -1.02287841e+00 -5.87445021e-01 9.81672406e-01
3.02621245e-01 -1.86739638e-01 -1.60486907e-01 -5.00261426e-01
2.13204652e-01 4.48682249e-01 -4.68830079e-01 -2.37094179e-01
-4.59080905e-01 -9.78424847e-01 4.45965916e-01 4.94628966e-01
7.82505035e-01 -8.51346135e-01 -4.53994364e-01 2.88914144e-01
2.42879570e-01 -9.97555315e-01 -3.63222599e-01 3.81906927e-01
-8.02491426e-01 -1.27072513e+00 -5.22212744e-01 -8.15761745e-01
9.16550219e-01 6.62821531e-01 5.63609064e-01 -1.08242370e-01
-6.50890231e-01 7.16139555e-01 -6.05545938e-01 -5.05086124e-01
-1.05837561e-01 -1.70368537e-01 2.69720018e-01 3.53273809e-01
6.35358572e-01 -1.28406405e-01 -3.09388369e-01 2.98174471e-01
-1.06619179e+00 -4.80206341e-01 7.25734353e-01 1.06903374e+00
6.76642358e-01 9.88490582e-02 5.98299980e-01 -6.17899120e-01
8.31975758e-01 -4.57379341e-01 -6.77823901e-01 5.93656719e-01
-6.27364218e-01 5.71261756e-02 8.75173032e-01 -7.27326691e-01
-1.00538242e+00 2.31885225e-01 2.67248541e-01 -5.84923208e-01
-3.14184189e-01 3.08287740e-01 -2.91065145e-02 -6.07895255e-01
6.17853403e-01 3.15350950e-01 6.92624878e-03 -4.29568321e-01
1.22557588e-01 8.14284146e-01 5.41309640e-02 -7.60579765e-01
7.05381393e-01 3.22478175e-01 1.48271486e-01 -1.12235475e+00
-7.65920341e-01 -1.14821506e+00 -1.01319754e+00 -3.06224793e-01
8.26549649e-01 -6.58296347e-01 -5.51357448e-01 5.03571630e-01
-8.91277850e-01 4.14794475e-01 6.22566380e-02 8.32183599e-01
-4.94149834e-01 4.57244396e-01 -7.88045079e-02 -9.15082455e-01
-2.91699797e-01 -1.41488016e+00 8.30865324e-01 4.70327646e-01
1.67501289e-02 -1.04081082e+00 1.16157047e-02 2.61969566e-01
3.50318491e-01 2.01434672e-01 9.96659338e-01 -8.13299775e-01
-2.27625668e-01 -6.11326218e-01 -4.29946721e-01 7.24160075e-01
5.17669261e-01 2.17746217e-02 -7.22471118e-01 -5.64407647e-01
-5.48479781e-02 -4.36001629e-01 7.34355271e-01 3.55427444e-01
1.04769397e+00 -2.07736984e-01 -4.08499449e-01 6.83981776e-01
1.81777883e+00 5.22843122e-01 2.87190735e-01 5.69094956e-01
6.45931542e-01 5.10088027e-01 8.97397101e-01 2.64712334e-01
-1.65916905e-02 7.43301809e-01 1.24952726e-01 2.36507878e-03
3.67132295e-03 4.20875371e-01 2.52068341e-01 6.69477224e-01
-2.12392900e-02 2.50168324e-01 -8.34937692e-01 3.12633723e-01
-1.81246543e+00 -9.04429615e-01 -2.68913716e-01 2.18020344e+00
4.26404893e-01 -7.46162534e-02 4.55334187e-02 2.83691287e-01
7.08467185e-01 -1.61216445e-02 -6.22382641e-01 -7.81448662e-01
-1.69812202e-01 1.62204981e-01 6.02163136e-01 5.18517077e-01
-1.46092486e+00 6.99766040e-01 6.35525608e+00 1.17087197e+00
-1.27001095e+00 -2.28172205e-02 1.94575340e-01 4.55150008e-01
2.17255995e-01 1.82366893e-01 -1.21069634e+00 4.10099953e-01
4.95469958e-01 -2.57380188e-01 5.31882420e-02 1.21329689e+00
1.38000473e-01 -4.29859310e-01 -7.77269483e-01 1.28844559e+00
3.80276144e-01 -8.86562407e-01 1.53333098e-01 -7.20857680e-02
5.28759480e-01 -4.26575840e-01 4.26137410e-02 3.22078168e-01
-3.44320796e-02 -7.50600636e-01 3.80304635e-01 5.70040762e-01
4.60633218e-01 -8.59771729e-01 8.97834241e-01 2.69444704e-01
-1.52320766e+00 -3.11464638e-01 -7.26002634e-01 1.40165001e-01
-4.02145654e-01 3.11858654e-01 -6.78254128e-01 6.18499517e-01
6.42972350e-01 7.19625533e-01 -8.89923573e-01 1.38150704e+00
3.94537628e-01 1.99864313e-01 -3.98556441e-01 -4.45878431e-02
3.62733513e-01 -6.23198032e-01 7.16759801e-01 1.51148844e+00
1.39286131e-01 1.22722968e-01 7.62199223e-01 4.38390672e-01
5.48988938e-01 6.70529306e-01 -8.67189646e-01 3.65308404e-01
1.85714096e-01 1.43105280e+00 -8.94933820e-01 -2.24713981e-01
-6.85005188e-01 9.25040841e-01 3.61926824e-01 2.89938718e-01
-6.00754440e-01 -5.39099574e-01 6.70391142e-01 -2.41787940e-01
1.57221332e-01 -3.29903275e-01 2.89206188e-02 -9.16896820e-01
1.38399228e-01 -7.66597331e-01 6.66314185e-01 -3.64178121e-01
-1.64069927e+00 4.87105995e-01 5.97423375e-01 -1.61026764e+00
2.60870010e-01 -1.01230788e+00 -5.70975006e-01 8.94454837e-01
-1.54907811e+00 -1.24521399e+00 -2.67398030e-01 7.44533598e-01
8.28060150e-01 -5.72363377e-01 8.99202049e-01 3.41953814e-01
-5.89020908e-01 4.76982385e-01 3.65511477e-01 -5.49776852e-02
8.62809241e-01 -1.21105969e+00 -7.49925315e-01 6.61063671e-01
-8.99665877e-02 6.67715073e-01 5.99504590e-01 -3.36566120e-01
-1.52286077e+00 -9.78526950e-01 5.46113729e-01 -3.66839953e-02
4.37776417e-01 8.05176497e-02 -1.05025327e+00 3.55919480e-01
-9.12148803e-02 3.96582037e-01 9.94102895e-01 -1.13989726e-01
-4.05578941e-01 -5.38107455e-01 -1.20626652e+00 2.43196279e-01
3.09098542e-01 -4.37440157e-01 -6.03295326e-01 4.54758108e-01
-1.25399664e-01 -3.85718159e-02 -9.15195823e-01 4.87611324e-01
6.38997197e-01 -9.07023728e-01 1.22658324e+00 -4.98260826e-01
-3.94567132e-01 -6.96854591e-01 -4.25073773e-01 -1.14149761e+00
-5.62169194e-01 5.10757007e-02 8.70764852e-02 1.23720109e+00
1.25428319e-01 -5.97265720e-01 3.89221311e-01 6.00949943e-01
2.00362485e-02 -9.15842891e-01 -8.49048078e-01 -1.27099514e+00
-7.17979446e-02 -2.40809768e-01 3.85446590e-03 9.25178945e-01
-4.59740818e-01 5.70639148e-02 -3.46536487e-01 2.96169430e-01
9.39767838e-01 1.24471888e-01 5.58933496e-01 -1.36944175e+00
6.96794242e-02 -4.35553610e-01 -8.62671554e-01 -4.84728098e-01
4.56479400e-01 -1.02036691e+00 -2.15979204e-01 -1.38708591e+00
3.22843075e-01 -8.45122755e-01 -6.15380406e-01 4.80108768e-01
-2.35140115e-01 9.61027071e-02 2.69120246e-01 3.75522971e-01
-3.11564416e-01 2.97888070e-01 9.08175647e-01 -3.95583928e-01
-3.53026390e-01 1.92938253e-01 -3.57712984e-01 7.59460747e-01
9.27233100e-01 -5.96623003e-01 -5.10947227e-01 1.81930717e-02
-3.08744133e-01 -2.70680875e-01 3.93573463e-01 -1.13541675e+00
3.02491188e-01 -4.10088718e-01 5.36090851e-01 -6.87093735e-01
5.81452012e-01 -1.05442643e+00 4.63175885e-02 2.59592950e-01
-3.86848748e-01 2.31444523e-01 2.12381095e-01 6.31057978e-01
-5.39573133e-01 -6.47961080e-01 1.28976440e+00 -1.91766530e-01
-1.24062347e+00 2.04709962e-01 -2.34633923e-01 -3.53314787e-01
1.51449430e+00 -8.06097329e-01 2.80016959e-01 1.40064880e-01
-6.91217184e-01 -3.52683268e-03 7.83848390e-03 5.67251861e-01
7.93184161e-01 -1.58505630e+00 -5.90401292e-01 1.52237341e-01
3.12291175e-01 -6.73843622e-01 -5.90547398e-02 7.75631011e-01
-5.68227232e-01 5.52743554e-01 -5.62078536e-01 -6.59650564e-01
-1.66325653e+00 5.55377245e-01 2.45274469e-01 -1.59563497e-01
-3.75691921e-01 7.36885726e-01 1.51362540e-02 -3.26287210e-01
2.07640797e-01 -1.52348012e-01 -5.50505102e-01 3.60824883e-01
4.12523746e-01 6.56714678e-01 8.93872827e-02 -1.05817890e+00
-5.24294734e-01 1.10332584e+00 -2.43232280e-01 3.00372154e-01
1.22940385e+00 3.43478029e-03 -2.97979742e-01 5.74181736e-01
1.69537532e+00 -2.03945950e-01 -8.00299287e-01 -2.62349665e-01
2.33038157e-01 -7.99789965e-01 7.83256367e-02 -3.99481088e-01
-7.77675986e-01 6.19402170e-01 9.56880689e-01 1.60773292e-01
1.28504479e+00 -4.23652619e-01 1.76129207e-01 5.87968528e-01
3.73403847e-01 -1.38604116e+00 2.72720277e-01 4.29510117e-01
9.97299075e-01 -1.57625842e+00 2.45278418e-01 -4.51128334e-01
-6.15291238e-01 1.55088258e+00 8.53418112e-01 -4.55687910e-01
1.09653234e+00 2.42385492e-02 5.30752838e-02 -1.44901332e-02
-4.37548637e-01 -2.92483687e-01 5.56390226e-01 5.70538998e-01
3.92059982e-01 1.13825463e-01 -6.31003916e-01 4.63008210e-02
4.00498956e-01 -2.96392918e-01 1.69217065e-01 1.03488517e+00
-5.49021602e-01 -1.40402782e+00 -8.64984393e-01 7.26389468e-01
-2.88760960e-01 3.57686996e-01 -1.67839140e-01 8.25509191e-01
4.35403436e-01 9.78960931e-01 -3.92685294e-01 -1.89880624e-01
3.61880124e-01 2.14032993e-01 4.06865001e-01 -5.76301277e-01
-4.32497710e-01 -1.31144011e-02 -4.02507216e-01 -3.12480122e-01
-7.55191624e-01 -7.80604005e-01 -1.01890731e+00 2.26392914e-02
-5.98807096e-01 1.53959140e-01 6.67098165e-01 8.94261479e-01
2.29290035e-02 8.01960304e-02 6.79478049e-01 -8.20347428e-01
-7.42207825e-01 -6.78500056e-01 -8.10095191e-01 3.84908944e-01
2.66665161e-01 -1.17636657e+00 -4.85556483e-01 1.62015751e-01] | [8.130358695983887, 4.005428791046143] |
8ce8f11a-b22e-4581-9a7d-1c43035fda50 | evaluating-bert-and-parsbert-for-analyzing | 2305.02426 | null | https://arxiv.org/abs/2305.02426v1 | https://arxiv.org/pdf/2305.02426v1.pdf | evaluating bert and parsbert for analyzing persian advertisement data | This paper discusses the impact of the Internet on modern trading and the importance of data generated from these transactions for organizations to improve their marketing efforts. The paper uses the example of Divar, an online marketplace for buying and selling products and services in Iran, and presents a competition to predict the percentage of a car sales ad that would be published on the Divar website. Since the dataset provides a rich source of Persian text data, the authors use the Hazm library, a Python library designed for processing Persian text, and two state-of-the-art language models, mBERT and ParsBERT, to analyze it. The paper's primary objective is to compare the performance of mBERT and ParsBERT on the Divar dataset. The authors provide some background on data mining, Persian language, and the two language models, examine the dataset's composition and statistical features, and provide details on their fine-tuning and training configurations for both approaches. They present the results of their analysis and highlight the strengths and weaknesses of the two language models when applied to Persian text data. The paper offers valuable insights into the challenges and opportunities of working with low-resource languages such as Persian and the potential of advanced language models like BERT for analyzing such data. The paper also explains the data mining process, including steps such as data cleaning and normalization techniques. Finally, the paper discusses the types of machine learning problems, such as supervised, unsupervised, and reinforcement learning, and the pattern evaluation techniques, such as confusion matrix. Overall, the paper provides an informative overview of the use of language models and data mining techniques for analyzing text data in low-resource languages, using the example of the Divar dataset. | ['Pegah Ahadian', 'Ali Mehrban'] | 2023-05-03 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [-6.61763608e-01 -1.82595000e-01 -5.20076692e-01 -4.75412667e-01
-4.96579558e-01 -7.35365331e-01 5.68239868e-01 4.32298392e-01
-9.23802555e-01 4.69954818e-01 6.35899752e-02 -7.85530925e-01
-2.85646558e-01 -8.82092774e-01 -2.72869468e-01 -2.43839443e-01
-1.48994476e-01 1.30766201e+00 -3.47445905e-01 -5.12265384e-01
7.00349331e-01 6.11934483e-01 -1.41714776e+00 2.70484388e-01
5.52238703e-01 7.37547398e-01 2.44323283e-01 2.74327099e-01
-9.92768764e-01 1.12923229e+00 -7.44928718e-01 -8.74433398e-01
6.17001414e-01 1.99632943e-01 -9.13120329e-01 -2.31360011e-02
-2.93850321e-02 -2.09044050e-02 -2.19436556e-01 7.50072241e-01
1.08522780e-01 -7.99430087e-02 5.03770530e-01 -1.39037526e+00
-6.08800888e-01 1.36095524e+00 -6.56060100e-01 3.10915262e-01
2.78789438e-02 -2.99077898e-01 1.04809928e+00 -7.70691991e-01
7.57431626e-01 1.52690613e+00 5.92799306e-01 -9.86264870e-02
-9.03065920e-01 -1.05428004e+00 -1.49217807e-02 1.09497987e-01
-1.37209022e+00 -9.00280923e-02 5.55948496e-01 -6.75682127e-01
1.27198994e+00 5.53254485e-02 5.58586061e-01 6.77262783e-01
5.46442091e-01 9.01939631e-01 1.50011921e+00 -5.67953467e-01
-4.70098894e-04 5.93185961e-01 6.53487325e-01 2.85147130e-01
4.17389721e-01 -1.55864254e-01 -6.44454658e-01 -3.78139287e-01
3.04790527e-01 -1.56597406e-01 8.48642766e-01 -5.62473871e-02
-8.24665606e-01 1.46237993e+00 -2.56460905e-01 5.67847729e-01
-3.66084933e-01 -6.32537425e-01 4.80300635e-01 4.16506022e-01
4.27983910e-01 6.70052946e-01 -1.07929790e+00 -4.57466036e-01
-1.02852464e+00 3.81151855e-01 1.36751628e+00 1.06185305e+00
7.83805728e-01 1.36186123e-01 4.67034847e-01 1.21046233e+00
6.32846475e-01 7.05409110e-01 7.62506545e-01 -6.76722825e-01
6.28224254e-01 8.46674919e-01 -2.22470015e-01 -9.39898372e-01
-5.10732353e-01 -1.27560943e-01 -4.16847199e-01 -1.76200077e-01
4.56055969e-01 -2.58702040e-01 -8.36203456e-01 9.92382526e-01
-1.65081009e-01 -9.81089890e-01 3.60127717e-01 3.77518535e-01
6.45404875e-01 7.34425962e-01 2.76965052e-01 -3.51675034e-01
1.51123524e+00 -7.67741024e-01 -8.59311521e-01 -5.97074687e-01
7.32733190e-01 -1.34636009e+00 7.61204600e-01 7.58857131e-01
-9.27871466e-01 -3.12091112e-01 -6.19993448e-01 -1.00712441e-01
-9.40462172e-01 4.96931337e-02 8.73714924e-01 7.40226269e-01
-3.23985219e-01 3.10703307e-01 -7.74350584e-01 -7.46832967e-01
7.09926710e-02 4.64660555e-01 -7.44353458e-02 2.69311547e-01
-1.14408386e+00 1.10427606e+00 6.94453418e-01 -1.37572706e-01
-4.53819335e-02 -2.55937606e-01 -7.48617887e-01 -3.84089023e-01
3.60395432e-01 2.36647099e-01 1.24326015e+00 -9.97345150e-01
-1.12399650e+00 1.08249128e+00 -3.13488953e-02 -7.55570769e-01
3.48161668e-01 -9.22406390e-02 -9.05607700e-01 -5.03538132e-01
4.30468351e-01 3.29773724e-01 3.75185937e-01 -8.72507095e-01
-1.19450068e+00 -5.85195720e-01 -4.26606357e-01 -2.36768201e-01
-2.86155101e-02 6.87586188e-01 -5.53830624e-01 -8.88298273e-01
1.22570815e-02 -8.69424343e-01 -1.82777509e-01 -9.82446134e-01
-1.57687426e-01 -4.03459579e-01 7.85024405e-01 -1.20578289e+00
1.59509313e+00 -2.19193673e+00 -5.43910801e-01 7.31929898e-01
-7.80386552e-02 1.02020971e-01 -1.07778870e-01 1.00780010e+00
-6.67127892e-02 4.16659921e-01 6.91211298e-02 2.13714793e-01
1.13029778e-01 6.62646711e-01 -3.71335149e-01 1.21955320e-01
1.46630462e-02 8.08971345e-01 -4.20235604e-01 -3.91700447e-01
2.46849105e-01 1.59509957e-01 -1.67363778e-01 -1.88990355e-01
3.89344767e-02 -1.82063490e-01 -1.62021786e-01 8.22220802e-01
7.93123126e-01 2.95136809e-01 6.34027123e-01 1.26028314e-01
-6.86318219e-01 5.98614335e-01 -1.32366836e+00 9.35424984e-01
-3.85593742e-01 6.14207387e-01 3.13243926e-01 -9.02422011e-01
1.35248792e+00 -3.23635906e-01 5.59983373e-01 -1.13359475e+00
1.44324943e-01 5.74939609e-01 3.17796439e-01 -4.53045785e-01
8.30596387e-01 -8.60049352e-02 -3.82699966e-01 8.68044138e-01
-1.58592816e-02 6.95482492e-02 9.17807519e-01 1.58460215e-01
5.59053719e-01 -1.10060200e-01 6.26212299e-01 -4.07034725e-01
6.14503086e-01 4.71858203e-01 6.87136352e-01 4.92409319e-01
9.25928578e-02 -1.66930571e-01 5.43866098e-01 -6.18409097e-01
-1.02879202e+00 -6.63128972e-01 -3.67376924e-01 1.38517475e+00
-3.66003871e-01 -7.91286230e-01 -3.36488396e-01 -6.75180078e-01
3.73235852e-01 1.14432156e+00 -2.86695302e-01 3.51145595e-01
-5.30635178e-01 -1.12624872e+00 6.22575879e-01 2.57053405e-01
5.47024131e-01 -1.12991655e+00 -3.67594600e-01 3.58528763e-01
-1.50722370e-01 -8.52462232e-01 -4.65372987e-02 6.37996137e-01
-8.56554687e-01 -1.08571720e+00 1.35092080e-01 -1.04101765e+00
1.70883626e-01 5.17555047e-03 1.11242020e+00 -4.37619388e-01
-1.44670144e-01 2.54769444e-01 -3.37488085e-01 -1.03607070e+00
-8.58522356e-01 2.60845304e-01 -4.78632050e-03 -3.99465173e-01
1.47466421e+00 -8.68012533e-02 3.50605577e-01 3.49889308e-01
-9.23893750e-01 -2.74682909e-01 1.04340911e+00 7.04341650e-01
3.42447400e-01 7.26594865e-01 3.76912385e-01 -1.35660136e+00
8.35532367e-01 -5.82005560e-01 -8.14929068e-01 1.87534139e-01
-1.29227483e+00 1.99198171e-01 2.16873959e-01 -1.67080373e-01
-9.78761971e-01 -7.19746649e-02 -7.79658854e-02 1.92907482e-01
-1.19755615e-03 9.92706537e-01 1.27197132e-01 3.14242154e-01
2.94196457e-01 1.97468176e-01 2.25303009e-01 -8.75527203e-01
1.54270858e-01 1.35599947e+00 -1.71966106e-02 -3.57238233e-01
6.80918872e-01 2.32031107e-01 -6.82959259e-01 -1.19918489e+00
-2.63350397e-01 -8.79410207e-01 -8.84510696e-01 2.28107780e-01
3.76529932e-01 -8.25983822e-01 -5.43260932e-01 5.89179397e-01
-7.45261848e-01 -1.07732542e-01 -2.63394594e-01 5.92198730e-01
-2.88474053e-01 1.41761094e-01 -8.66410613e-01 -1.15604198e+00
-4.08994019e-01 -9.31882799e-01 5.95037103e-01 -1.65732980e-01
-6.91075146e-01 -1.06201184e+00 5.61702473e-04 9.54786181e-01
1.51949629e-01 -5.26344538e-01 1.28487706e+00 -1.62183273e+00
-2.30259709e-02 -1.47744030e-01 -6.79020882e-02 5.16507745e-01
1.71495706e-01 1.69914871e-01 -5.23609281e-01 -1.92404419e-01
5.14737070e-02 -1.96221724e-01 3.27247590e-01 9.59576294e-02
6.51404858e-01 -5.98274171e-01 6.25351667e-02 -9.09523014e-03
1.18793118e+00 6.92280054e-01 3.58858615e-01 8.70733738e-01
3.28452438e-01 1.25068641e+00 1.03469157e+00 4.76369858e-01
5.71021497e-01 3.09905529e-01 -3.04592103e-01 2.05404073e-01
3.21012378e-01 -3.90089720e-01 5.66740870e-01 1.33913791e+00
2.02073425e-01 2.95878977e-01 -1.26623678e+00 4.02203321e-01
-1.68804812e+00 -6.84579551e-01 -3.75911564e-01 1.98716712e+00
7.21073747e-01 1.29736006e-01 5.19945502e-01 1.48962915e-01
4.38003749e-01 -2.27390751e-01 -2.17340723e-01 -1.06597567e+00
-4.26967293e-01 3.97501916e-01 1.07606184e+00 4.62479621e-01
-1.11949801e+00 1.13185215e+00 7.14807177e+00 7.50975668e-01
-1.09813046e+00 -1.30960569e-01 4.24778968e-01 8.84473771e-02
-7.09606260e-02 2.37528402e-02 -9.60426748e-01 4.39113200e-01
1.39212334e+00 -4.05629784e-01 6.99512303e-01 8.91028404e-01
4.29647982e-01 -3.01306129e-01 -6.76844895e-01 1.13538218e+00
3.16214442e-01 -1.10333741e+00 3.22928250e-01 5.39521337e-01
3.65154326e-01 4.64737922e-01 -2.39874288e-01 5.78134358e-01
5.43200970e-01 -8.19700480e-01 1.03700650e+00 2.51033604e-02
4.96058650e-02 -1.06315708e+00 1.08298528e+00 4.04967278e-01
-7.94928730e-01 -4.63703096e-01 -3.42216313e-01 -2.08848014e-01
-2.33626530e-01 4.45101380e-01 -9.55801725e-01 6.46401703e-01
1.00809443e+00 7.99167514e-01 -7.74362922e-01 3.88020277e-01
3.10149878e-01 7.32012749e-01 -3.20258468e-01 -1.70294553e-01
2.63755620e-01 -8.61777604e-01 3.28059077e-01 1.42973006e+00
1.59283325e-01 -3.18496168e-01 1.83216617e-01 6.28540635e-01
2.46113524e-01 8.17596793e-01 -5.55466771e-01 -8.38273704e-01
5.51620662e-01 1.18107247e+00 -8.20221305e-01 -8.25705677e-02
-8.84383559e-01 3.31796408e-01 -3.42557840e-02 1.21006906e-01
-2.16252238e-01 -4.08725500e-01 5.04979670e-01 3.39479089e-01
1.59904107e-01 -3.10086757e-01 -6.84216738e-01 -8.42138410e-01
-1.16616979e-01 -1.55787444e+00 8.83513510e-01 -4.35813814e-01
-1.55374479e+00 3.27572733e-01 6.01916797e-02 -8.92909825e-01
-2.85801440e-01 -1.04938161e+00 2.08373427e-01 8.98454607e-01
-1.13091063e+00 -1.16180587e+00 4.99556273e-01 3.35306019e-01
6.82344675e-01 -9.19472694e-01 6.85093760e-01 3.79762828e-01
-4.05131876e-01 2.28269324e-01 3.72081608e-01 6.04249418e-01
8.49814892e-01 -1.07535565e+00 2.84179300e-01 5.03185630e-01
3.97642195e-01 9.90424037e-01 4.05072421e-01 -1.10847867e+00
-1.68237424e+00 -8.31057608e-01 1.51127708e+00 -5.79761922e-01
1.21265304e+00 -5.50133348e-01 -5.57774127e-01 9.73200619e-01
3.73474836e-01 -1.20884836e+00 1.09083831e+00 3.48234773e-01
-2.31298715e-01 -3.06897134e-01 -1.06199396e+00 4.00413632e-01
3.32728952e-01 -3.22596043e-01 -1.06735432e+00 1.47543818e-01
-1.29854279e-02 4.66351695e-02 -8.29515219e-01 -1.75200626e-01
6.61462665e-01 -6.71650648e-01 4.91462409e-01 -7.94682324e-01
1.06660657e-01 -1.78817241e-03 -2.25820661e-01 -9.26590025e-01
-3.69179249e-01 -5.07750392e-01 4.95146126e-01 1.53715563e+00
8.48193288e-01 -9.89320815e-01 5.58292866e-01 4.88223821e-01
4.40377563e-01 -2.50195771e-01 -7.00844049e-01 -7.31198668e-01
4.34125870e-01 -8.78884435e-01 4.58568066e-01 1.12720799e+00
2.08680518e-02 5.73114097e-01 -1.03690304e-01 -3.68856788e-01
3.01465154e-01 1.44523725e-01 9.05074835e-01 -1.41124249e+00
-5.32721495e-03 -3.94509554e-01 -3.55153471e-01 -5.50037682e-01
1.17957890e-01 -1.08675861e+00 -4.58778352e-01 -1.29686737e+00
1.37598231e-01 -5.39412737e-01 5.44615835e-02 5.67843258e-01
5.59001327e-01 -3.27699669e-02 4.78052229e-01 6.60556138e-01
-7.82966241e-02 -2.35023811e-01 5.71647227e-01 -3.71707201e-01
-5.11977911e-01 -4.91371714e-02 -1.20738530e+00 7.82731235e-01
7.88848519e-01 -4.10679936e-01 -1.69254616e-01 -1.70993522e-01
5.97677529e-01 -5.15259802e-01 -3.04498851e-01 -1.63875192e-01
1.17394544e-01 -3.98944974e-01 3.87522787e-01 -1.03515744e+00
-7.98531249e-02 -1.09339893e+00 2.89288253e-01 4.17842269e-01
-3.08526218e-01 6.13819063e-01 4.13015902e-01 4.08682674e-02
-4.45038587e-01 -2.35408723e-01 5.51006615e-01 -1.56309813e-01
-9.97159064e-01 -2.04869092e-01 -1.10611832e+00 1.99176773e-01
1.04383719e+00 -1.42461926e-01 -1.40513077e-01 -2.80785531e-01
-6.55623496e-01 4.91553426e-01 1.12142883e-01 8.95017266e-01
8.51209909e-02 -1.01150441e+00 -8.06073666e-01 6.60197198e-01
1.52010828e-01 -5.50924599e-01 -3.02368313e-01 7.86813974e-01
-8.99207115e-01 8.37604702e-01 -3.05507720e-01 -1.55153468e-01
-1.23292935e+00 6.20568633e-01 -2.65743602e-02 -4.85609382e-01
-1.62586421e-01 6.18607886e-02 -2.70818681e-01 -9.44035232e-01
9.17529836e-02 -2.35843793e-01 -4.65608925e-01 6.52638435e-01
3.97150457e-01 6.17719173e-01 2.70315677e-01 -9.35233951e-01
-4.67470288e-01 2.64544308e-01 -3.61607909e-01 -3.27538431e-01
1.46957982e+00 -3.73367041e-01 -8.67049396e-01 9.85346794e-01
8.21699977e-01 4.82181787e-01 -1.28462061e-01 -1.57611847e-01
7.31106579e-01 -1.64120182e-01 -1.59216464e-01 -1.23990059e+00
-9.39890444e-01 4.84886646e-01 4.45207536e-01 3.78648221e-01
8.51527035e-01 8.60861838e-02 5.88801503e-01 5.41716516e-01
3.93093407e-01 -1.84294140e+00 -7.31180787e-01 9.65119183e-01
7.22806454e-01 -1.05079138e+00 1.85439475e-02 -3.75199080e-01
-9.25387681e-01 1.34981859e+00 2.39329189e-01 2.49224052e-01
1.08543205e+00 6.48454905e-01 7.58656740e-01 -8.08816627e-02
-5.41774511e-01 -7.11796200e-03 -1.97192281e-03 4.78285074e-01
7.02591002e-01 3.08059782e-01 -9.84169841e-01 1.04928875e+00
-1.04885650e+00 -8.95191804e-02 5.27207255e-01 1.28630698e+00
-1.46794796e-01 -1.70988917e+00 -7.15418577e-01 9.81060326e-01
-8.67887795e-01 -1.69329137e-01 -1.11395109e+00 1.36506498e+00
1.13148890e-01 1.08531988e+00 3.21031153e-01 -4.42230970e-01
3.89820784e-01 4.56817448e-01 -1.64652348e-01 -7.00166047e-01
-8.96760166e-01 5.01995385e-01 4.81403649e-01 -1.76407874e-01
-3.57025176e-01 -1.09247303e+00 -1.32016480e+00 -8.20224822e-01
-4.07832190e-02 5.86832285e-01 9.02295589e-01 8.41858864e-01
1.20631315e-01 -9.36486796e-02 4.66435105e-01 -2.64902785e-02
-4.93964404e-01 -1.05242360e+00 -1.14512467e+00 2.78283745e-01
-2.15090141e-01 -3.02154899e-01 -3.01676571e-01 2.55396552e-02] | [10.280202865600586, 9.911498069763184] |
8f82e0fd-f9db-4a33-bf6f-3cb2dcac5b02 | first-image-then-video-a-two-stage-network | 2001.00346 | null | https://arxiv.org/abs/2001.00346v2 | https://arxiv.org/pdf/2001.00346v2.pdf | First image then video: A two-stage network for spatiotemporal video denoising | Video denoising is to remove noise from noise-corrupted data, thus recovering true signals via spatiotemporal processing. Existing approaches for spatiotemporal video denoising tend to suffer from motion blur artifacts, that is, the boundary of a moving object tends to appear blurry especially when the object undergoes a fast motion, causing optical flow calculation to break down. In this paper, we address this challenge by designing a first-image-then-video two-stage denoising neural network, consisting of an image denoising module for spatially reducing intra-frame noise followed by a regular spatiotemporal video denoising module. The intuition is simple yet powerful and effective: the first stage of image denoising effectively reduces the noise level and, therefore, allows the second stage of spatiotemporal denoising for better modeling and learning everywhere, including along the moving object boundaries. This two-stage network, when trained in an end-to-end fashion, yields the state-of-the-art performances on the video denoising benchmark Vimeo90K dataset in terms of both denoising quality and computation. It also enables an unsupervised approach that achieves comparable performance to existing supervised approaches. | ['S. Kevin Zhou', 'Zhiwei Cheng', 'Ce Wang'] | 2020-01-02 | null | null | null | null | ['video-denoising'] | ['computer-vision'] | [ 2.37106293e-01 -4.63419706e-01 2.60086685e-01 -1.38333827e-01
-4.85089064e-01 -2.82991439e-01 3.35374922e-01 -1.97761729e-01
-5.32840967e-01 4.14681822e-01 3.40040654e-01 -3.88650340e-03
6.11938573e-02 -4.47136670e-01 -6.74870014e-01 -1.12587738e+00
-5.36507517e-02 -4.85023528e-01 2.75120616e-01 -1.02119118e-01
9.74405557e-02 1.05756231e-01 -1.34856951e+00 2.80404478e-01
8.14973056e-01 1.16459799e+00 2.97270775e-01 8.33184779e-01
3.13883722e-01 1.42886007e+00 -4.32120204e-01 -2.67135262e-01
2.80691385e-01 -5.15848398e-01 -4.79229003e-01 1.53921217e-01
5.22560060e-01 -8.19597781e-01 -8.60590756e-01 1.28897345e+00
3.73673648e-01 4.94254053e-01 3.09384108e-01 -8.09720278e-01
-6.55150890e-01 2.10051447e-01 -6.76031888e-01 5.25428474e-01
9.00115594e-02 3.57154757e-01 5.38371384e-01 -7.41402030e-01
5.71107328e-01 1.32266223e+00 7.85119772e-01 6.97285712e-01
-1.21498060e+00 -2.54977316e-01 1.43912598e-01 3.99214178e-01
-1.14113796e+00 -6.38483107e-01 9.52323675e-01 -4.14635330e-01
5.00406682e-01 7.58424625e-02 5.26953399e-01 1.11346185e+00
3.89490575e-01 7.48657525e-01 6.61589146e-01 -3.84530649e-02
3.53235096e-01 -3.83703232e-01 4.90125678e-02 3.07093412e-01
1.70235947e-01 2.39528522e-01 -6.04372323e-01 2.25090653e-01
8.32916141e-01 2.22712263e-01 -6.52238727e-01 -2.11839363e-01
-1.02016592e+00 3.92755300e-01 4.19568717e-01 4.05595273e-01
-6.72233045e-01 4.37046021e-01 7.05848992e-01 4.15824413e-01
7.47360408e-01 -1.52116701e-01 -1.89091235e-01 -2.59519309e-01
-1.42734468e+00 1.33644581e-01 5.32352388e-01 5.74634910e-01
4.71860707e-01 1.66333526e-01 -2.95498669e-01 6.71646595e-01
2.92497307e-01 3.97173427e-02 3.61872017e-01 -1.46407497e+00
4.70116675e-01 -1.91533826e-02 2.57493705e-01 -1.18248296e+00
-4.72032577e-02 -2.81422943e-01 -1.54228461e+00 4.20578718e-01
5.82365155e-01 -1.23841576e-01 -1.04112804e+00 1.71708369e+00
1.93893805e-01 7.75120676e-01 1.42047167e-01 1.35334361e+00
7.73154974e-01 9.16827679e-01 1.70535162e-01 -5.34386456e-01
1.13193977e+00 -1.13736713e+00 -1.20474541e+00 -1.71910554e-01
2.71891087e-01 -6.46898091e-01 6.35907292e-01 6.78973436e-01
-1.36537540e+00 -8.98282707e-01 -1.01497602e+00 -4.09322649e-01
1.23557569e-02 -7.69155771e-02 2.92189240e-01 2.96647698e-01
-1.24527597e+00 8.30029130e-01 -1.15424109e+00 -1.37106374e-01
6.40060484e-01 1.34207541e-04 -3.85217577e-01 -4.31022406e-01
-1.08351302e+00 6.15596950e-01 1.47304103e-01 7.67651618e-01
-1.19690967e+00 -8.66347492e-01 -1.01202381e+00 1.24100700e-01
3.58330280e-01 -8.87512326e-01 1.00475490e+00 -1.21617496e+00
-1.35398626e+00 4.86211091e-01 -5.23845971e-01 -5.69097161e-01
9.10824180e-01 -5.11752427e-01 -3.13529372e-01 4.31157440e-01
1.04253413e-02 4.12463427e-01 1.46919739e+00 -1.36161733e+00
-5.64612687e-01 -3.35968852e-01 -2.47586906e-01 1.09456316e-01
-1.38659090e-01 -1.76846534e-01 -8.49557221e-01 -1.06170011e+00
7.47454986e-02 -4.06930834e-01 -3.09086442e-01 1.23795159e-01
-6.91758990e-02 1.05214886e-01 1.13477945e+00 -1.14492166e+00
1.44309318e+00 -2.41915178e+00 3.45642567e-01 -1.48492560e-01
4.58038688e-01 4.80365694e-01 -2.07334936e-01 -4.07972373e-02
-3.30658704e-01 2.74568330e-03 -3.93685549e-01 -6.70861423e-01
-5.43054461e-01 3.51294614e-02 -1.44251227e-01 6.95361912e-01
1.52362466e-01 7.37252116e-01 -1.05767238e+00 -1.53548151e-01
5.44243395e-01 8.26452613e-01 -4.71780151e-01 3.47808629e-01
2.35486567e-01 7.78925896e-01 -2.29251221e-01 3.86054546e-01
1.09610975e+00 7.56853372e-02 -7.47361183e-02 -5.07157862e-01
-1.57731503e-01 -1.43067464e-01 -1.30868614e+00 1.85902786e+00
-3.45688462e-01 9.19893444e-01 6.79983675e-01 -1.02745020e+00
4.39804643e-01 4.19697076e-01 6.10561311e-01 -7.38736570e-01
1.18458018e-01 -5.33544691e-03 -2.79781282e-01 -8.60885918e-01
3.22714537e-01 -1.07281834e-01 4.75124180e-01 -7.83089623e-02
-4.18611057e-02 1.29175708e-01 3.72067332e-01 2.63947934e-01
1.31254804e+00 3.06355417e-01 -1.13982461e-01 -2.62109518e-01
7.88731933e-01 -4.06057715e-01 7.84141421e-01 7.67299652e-01
-4.96577471e-01 9.94344056e-01 5.10117710e-01 -5.24873674e-01
-9.58860099e-01 -9.72932041e-01 2.31576249e-01 7.89181828e-01
5.02240598e-01 -2.70521760e-01 -1.06745851e+00 -3.53130907e-01
-1.82751045e-01 3.87328893e-01 -6.17312670e-01 -3.22583407e-01
-8.06076229e-01 -5.29912829e-01 2.46216580e-01 4.76030529e-01
6.85477436e-01 -9.53505695e-01 -5.70372581e-01 4.29743350e-01
-5.37883282e-01 -1.29957652e+00 -7.96977162e-01 7.11488351e-02
-1.15345073e+00 -9.94662941e-01 -9.47925091e-01 -8.57894301e-01
5.03855169e-01 6.55594468e-01 1.11697054e+00 4.46938962e-01
-9.77312475e-02 1.29794151e-01 -3.00499648e-01 2.04355985e-01
-2.90119529e-01 -4.89622742e-01 -9.77819636e-02 3.19209963e-01
1.55930504e-01 -6.40949011e-01 -1.03777206e+00 1.89628705e-01
-1.43599415e+00 -1.85421795e-01 3.79682690e-01 9.30740654e-01
4.32804435e-01 4.44460154e-01 2.26990223e-01 -4.84955728e-01
4.75209117e-01 -3.13049793e-01 -5.94374657e-01 -1.08838622e-02
-1.44645631e-01 -2.63706893e-01 8.31949770e-01 -3.27448875e-01
-1.31695783e+00 6.86332285e-02 -2.54526645e-01 -7.70151079e-01
-1.59123540e-01 3.04971933e-01 -1.79259732e-01 1.05025023e-01
4.48063701e-01 3.97211820e-01 2.26649940e-01 -7.65710771e-01
3.21800262e-01 4.95223403e-01 9.52218592e-01 -3.51885222e-02
7.85115242e-01 8.33315313e-01 -5.01331799e-02 -1.24610484e+00
-7.37689316e-01 -8.07828188e-01 -5.65188110e-01 -3.53198737e-01
1.06666100e+00 -1.33636332e+00 -7.87911952e-01 1.08438265e+00
-1.27766573e+00 -4.72611576e-01 -1.27071291e-01 3.96592081e-01
-3.93600464e-01 8.12652707e-01 -1.12802029e+00 -5.27705729e-01
-2.89635360e-01 -1.18733263e+00 9.60216701e-01 2.91726410e-01
1.81141511e-01 -1.12114239e+00 -2.07518429e-01 3.14430863e-01
5.25010645e-01 7.11781010e-02 4.66490895e-01 2.46951595e-01
-6.76691413e-01 9.68814865e-02 -2.87669867e-01 1.03595817e+00
1.57161906e-01 -1.12134583e-01 -9.42960739e-01 -4.22948271e-01
6.48925006e-01 9.82194617e-02 1.41009796e+00 9.43147123e-01
1.29308593e+00 -7.87054375e-02 3.95196714e-02 1.01621771e+00
1.64710903e+00 4.67132358e-03 1.18916225e+00 3.07911992e-01
9.28397357e-01 4.68195796e-01 4.63263303e-01 2.19061598e-01
1.32737860e-01 5.00773311e-01 5.40073037e-01 -2.79167235e-01
-3.13939065e-01 1.76944733e-01 4.95718628e-01 6.04766488e-01
-1.31168544e-01 -3.36777031e-01 -4.84635055e-01 6.73979759e-01
-1.92181945e+00 -1.05936027e+00 -3.88939768e-01 2.02535725e+00
7.68303454e-01 -1.66094508e-02 -1.73482686e-01 1.36028007e-01
6.36063218e-01 5.35558105e-01 -3.56341541e-01 -3.55268680e-02
-1.74648166e-01 -4.65052165e-02 3.96814793e-01 6.45123243e-01
-1.31393611e+00 7.99029410e-01 5.53141451e+00 7.54740596e-01
-1.13383532e+00 2.08567291e-01 9.20216858e-01 -5.69132492e-02
2.77900219e-01 -2.61401534e-01 -3.26835930e-01 8.06218207e-01
7.02611864e-01 3.12411219e-01 6.45324707e-01 4.21496838e-01
1.02475142e+00 -4.29307550e-01 -9.81425583e-01 1.35011685e+00
-6.40129521e-02 -1.32208943e+00 -7.13461041e-02 -3.19953322e-01
7.24873543e-01 -1.62242696e-01 -1.77676156e-01 -7.29440227e-02
-3.01831126e-01 -9.09939051e-01 8.60080659e-01 8.76837850e-01
3.68726254e-01 -6.15873337e-01 9.41804230e-01 3.38558078e-01
-1.15080154e+00 -1.64507598e-01 -3.34762216e-01 -1.00404695e-01
5.50702691e-01 1.00573766e+00 4.39361364e-01 5.93752086e-01
1.19482124e+00 1.18566310e+00 -3.16148221e-01 1.12687016e+00
-3.13677311e-01 6.59748137e-01 1.11209035e-01 7.02302873e-01
3.63346726e-01 -5.98500848e-01 7.43409991e-01 1.31575465e+00
2.32889101e-01 1.86424732e-01 -1.75216794e-01 6.00181937e-01
-1.36700362e-01 -4.65864152e-01 -4.01070654e-01 4.34120536e-01
-3.39597911e-02 1.21992898e+00 -5.44706106e-01 -5.35753429e-01
-5.61275125e-01 1.32496154e+00 -8.25891346e-02 8.75059843e-01
-8.06821942e-01 -2.00543091e-01 9.31713164e-01 6.04280680e-02
5.94549954e-01 -2.76474714e-01 -3.94773036e-01 -1.30835259e+00
2.65054971e-01 -9.31737065e-01 2.05330268e-01 -7.87391067e-01
-1.19675386e+00 4.45340544e-01 -4.52104867e-01 -1.22756684e+00
7.08385631e-02 -4.93865132e-01 -7.61549473e-01 7.90101767e-01
-1.62078273e+00 -7.45937705e-01 -7.36663103e-01 6.13543630e-01
7.95132041e-01 3.83047789e-01 7.96689540e-02 7.17485428e-01
-7.19030678e-01 1.76661149e-01 3.26269537e-01 2.42153466e-01
8.56612623e-01 -1.10299683e+00 1.81288257e-01 1.51644838e+00
-3.81161779e-01 5.28102219e-01 8.75727475e-01 -6.04775846e-01
-1.49591601e+00 -1.33061385e+00 5.37445188e-01 -2.00076938e-01
5.30284643e-01 -2.58850902e-01 -1.26696944e+00 4.18733001e-01
3.94594222e-01 2.88358867e-01 -8.17490220e-02 -5.00156105e-01
1.17988111e-02 -3.89814407e-01 -9.85641420e-01 5.78919351e-01
9.24811959e-01 -3.76091868e-01 -4.26600426e-01 1.90838648e-03
5.06667614e-01 -3.57156128e-01 -8.11407149e-01 2.11550549e-01
2.86482513e-01 -1.19164205e+00 1.06703019e+00 -1.90679863e-01
6.56102896e-01 -6.67285323e-01 1.29635155e-01 -1.30900550e+00
-4.27122444e-01 -1.06651473e+00 -4.76261675e-01 1.15700889e+00
-1.34085312e-01 -1.48837194e-01 6.53474331e-01 3.30590159e-01
-8.89949203e-02 -4.73253012e-01 -8.33076715e-01 -5.61408043e-01
-2.29973018e-01 -5.28349459e-01 -1.81162491e-01 7.61212528e-01
-6.11291409e-01 1.87541708e-01 -7.01207221e-01 2.48286679e-01
1.04379916e+00 -5.80250323e-01 4.77450997e-01 -7.25517571e-01
9.82054994e-02 -3.45202684e-01 -3.36779982e-01 -1.63716662e+00
-1.46040052e-01 -1.31143585e-01 4.85477865e-01 -1.48961437e+00
4.80061397e-02 2.20909104e-01 -2.18418822e-01 -4.58662622e-02
-4.66941655e-01 4.36427742e-01 6.78236485e-02 1.63378790e-01
-7.01275051e-01 5.88246346e-01 1.39133358e+00 -1.28813267e-01
-2.88357794e-01 -1.39135391e-01 -5.88845670e-01 8.08661819e-01
2.95872658e-01 -3.28019291e-01 -2.68690944e-01 -7.83088565e-01
-2.21535206e-01 1.47443423e-02 7.67769217e-01 -9.50138807e-01
4.60887641e-01 2.10309133e-01 5.68515241e-01 -5.10864735e-01
1.79229155e-01 -1.00621367e+00 1.52101917e-02 4.88169044e-01
-5.22715151e-02 -1.53956071e-01 9.26877707e-02 8.83247256e-01
-7.36550748e-01 -1.11187756e-01 1.22253847e+00 -1.71134651e-01
-1.05274713e+00 2.26702452e-01 -6.33127570e-01 2.76161861e-02
8.42834353e-01 -3.11779112e-01 -6.88858926e-02 -6.18903816e-01
-8.93698990e-01 1.69515654e-01 4.09789294e-01 3.79605889e-01
6.91165864e-01 -1.04115391e+00 -7.46424377e-01 2.24254265e-01
-4.88856524e-01 2.42547125e-01 7.80357420e-01 1.19589591e+00
-8.52326334e-01 -6.39695376e-02 3.77083570e-02 -8.45194817e-01
-1.11361790e+00 6.59099460e-01 5.65080404e-01 -9.87552777e-02
-9.25554514e-01 9.35507476e-01 2.81904072e-01 3.31764311e-01
4.92101669e-01 -5.71612954e-01 -1.85278177e-01 1.25857323e-01
9.21408653e-01 6.36558414e-01 9.29757059e-02 -6.59100175e-01
-2.54836619e-01 7.07098365e-01 9.56779271e-02 2.19287887e-01
1.45980632e+00 -5.53414643e-01 -3.36574435e-01 6.90758079e-02
1.37156391e+00 -3.19555640e-01 -1.90882504e+00 -9.62420106e-02
-2.08633944e-01 -6.27852023e-01 4.78840917e-01 -4.52277213e-01
-1.48689234e+00 8.11095357e-01 8.28923702e-01 2.94318259e-01
1.64033008e+00 -4.71706122e-01 9.97177064e-01 -1.98227968e-02
-2.07744837e-01 -1.13173819e+00 1.65134028e-01 5.23258448e-01
7.36329913e-01 -1.32971263e+00 -7.41466656e-02 -4.67643976e-01
-4.48812068e-01 1.15484178e+00 2.86801815e-01 -2.71301955e-01
7.57660747e-01 2.71396875e-01 2.07417801e-01 -1.11736758e-02
-6.21498883e-01 4.73004766e-03 3.17921400e-01 4.52339768e-01
2.88788974e-01 -5.58240473e-01 -1.91292912e-01 3.35918367e-01
5.12290597e-01 2.60945380e-01 4.47885513e-01 7.96477556e-01
-2.26483062e-01 -5.91463685e-01 -4.25460458e-01 1.14058763e-01
-7.60899305e-01 -2.08743140e-01 1.56189337e-01 5.71835399e-01
2.50770539e-01 1.27821779e+00 3.44656222e-02 -8.23771767e-03
4.66234446e-01 -4.16375667e-01 2.64738530e-01 -3.56305763e-02
-5.44847310e-01 4.43593770e-01 -2.87314326e-01 -1.03845501e+00
-6.72718287e-01 -5.43630421e-01 -9.10703242e-01 -4.54734087e-01
4.21779975e-02 -3.00928559e-02 4.13570970e-01 9.29505646e-01
2.44330227e-01 8.35596561e-01 5.84126472e-01 -1.21478701e+00
-3.45750779e-01 -1.01008713e+00 -5.24159193e-01 8.82830262e-01
9.35229361e-01 -3.02584887e-01 -7.24568665e-01 6.07036054e-01] | [11.386120796203613, -2.156062126159668] |
6775c7d8-1fdb-49a2-85af-5fc80841b74e | supervised-learning-of-universal-sentence | 1705.02364 | null | http://arxiv.org/abs/1705.02364v5 | http://arxiv.org/pdf/1705.02364v5.pdf | Supervised Learning of Universal Sentence Representations from Natural Language Inference Data | Many modern NLP systems rely on word embeddings, previously trained in an
unsupervised manner on large corpora, as base features. Efforts to obtain
embeddings for larger chunks of text, such as sentences, have however not been
so successful. Several attempts at learning unsupervised representations of
sentences have not reached satisfactory enough performance to be widely
adopted. In this paper, we show how universal sentence representations trained
using the supervised data of the Stanford Natural Language Inference datasets
can consistently outperform unsupervised methods like SkipThought vectors on a
wide range of transfer tasks. Much like how computer vision uses ImageNet to
obtain features, which can then be transferred to other tasks, our work tends
to indicate the suitability of natural language inference for transfer learning
to other NLP tasks. Our encoder is publicly available. | ['Holger Schwenk', 'Loic Barrault', 'Douwe Kiela', 'Antoine Bordes', 'Alexis Conneau'] | 2017-05-05 | supervised-learning-of-universal-sentence-1 | https://aclanthology.org/D17-1070 | https://aclanthology.org/D17-1070.pdf | emnlp-2017-9 | ['cross-lingual-natural-language-inference'] | ['natural-language-processing'] | [ 1.28453627e-01 4.20614004e-01 -3.69823456e-01 -8.16029191e-01
-7.34168291e-01 -2.66854495e-01 9.47968543e-01 3.83034647e-01
-8.23206604e-01 1.13036835e+00 5.28639615e-01 -4.53209966e-01
1.85173675e-01 -9.05345619e-01 -6.50447190e-01 -3.36849630e-01
3.82969575e-03 5.30211508e-01 2.22946018e-01 -2.37980917e-01
3.89771581e-01 2.42078543e-01 -1.25983238e+00 1.25148252e-01
3.73602927e-01 3.22039396e-01 2.20956817e-01 6.74050152e-01
-5.80545664e-01 9.31069791e-01 -4.73295629e-01 -4.40671742e-01
-1.38165087e-01 -3.74613881e-01 -1.12472427e+00 -8.55431855e-02
4.93968189e-01 -4.91445690e-01 -5.63963890e-01 8.72001708e-01
3.16756457e-01 2.05593258e-01 9.56689954e-01 -8.70724618e-01
-1.19703519e+00 6.48375690e-01 1.24204857e-03 4.23844427e-01
3.06801796e-01 -1.58778325e-01 1.39788842e+00 -9.47704732e-01
7.66694725e-01 1.25757504e+00 5.60760260e-01 5.67620397e-01
-1.44125199e+00 -2.75857925e-01 -3.50501865e-01 1.44433394e-01
-1.13714993e+00 -3.85047019e-01 6.04754508e-01 -3.66487354e-01
1.54685712e+00 -2.28622586e-01 4.11021948e-01 1.11565733e+00
3.35344613e-01 7.73350060e-01 9.39997852e-01 -7.33950853e-01
-4.90396284e-02 4.56173062e-01 2.51080692e-01 8.18245411e-01
4.22650993e-01 3.47762294e-02 -3.75952065e-01 1.06742471e-01
7.94495225e-01 -4.90016378e-02 -7.09314570e-02 -2.39860982e-01
-1.28997958e+00 1.39632034e+00 7.38631487e-01 8.49717379e-01
-1.86916471e-01 4.30241644e-01 5.82199812e-01 5.65480649e-01
7.44491756e-01 7.16604054e-01 -6.56674623e-01 -3.46410543e-01
-7.25833356e-01 1.03357323e-01 8.09874296e-01 8.84404778e-01
1.09780085e+00 -5.21531375e-03 1.51565358e-01 9.41677690e-01
1.87356398e-01 1.37707666e-01 7.24179506e-01 -6.90910697e-01
5.11909246e-01 4.17103320e-01 -9.49875936e-02 -1.03617990e+00
-1.51674837e-01 3.38210613e-01 -4.80875999e-01 -1.63438022e-02
3.21694165e-01 -3.66912305e-01 -8.38339627e-01 1.49094021e+00
-2.30746701e-01 -9.87379625e-02 4.88994718e-01 5.12607872e-01
8.19215715e-01 9.01934266e-01 2.26130918e-01 2.58933045e-02
1.07133043e+00 -7.39398241e-01 -5.90360165e-01 -4.16173935e-01
9.60260928e-01 -7.05046415e-01 8.44252467e-01 -1.07946189e-03
-8.76846015e-01 -6.56270146e-01 -1.07492769e+00 -5.37309468e-01
-8.64620328e-01 -2.53117681e-01 9.76740241e-01 4.56855536e-01
-1.15261865e+00 6.88099205e-01 -7.22411335e-01 -8.14078629e-01
7.77764142e-01 3.67487371e-01 -6.42600536e-01 -2.27357954e-01
-1.31977797e+00 1.46268463e+00 7.65024900e-01 -2.63179958e-01
-6.28544748e-01 -4.19587165e-01 -1.30765224e+00 1.22573577e-01
-7.16682971e-02 -5.86433768e-01 1.04258215e+00 -1.10764885e+00
-1.24156451e+00 1.07420909e+00 -6.71172142e-02 -7.56196380e-01
-2.57894754e-01 -1.07969798e-01 -2.57325053e-01 3.13907474e-01
8.77655968e-02 1.15508342e+00 8.87149215e-01 -7.44255304e-01
-2.22301960e-01 -1.34031892e-01 3.64082277e-01 6.53488711e-02
-8.12017977e-01 1.41477466e-01 1.19658515e-01 -4.05992985e-01
-4.52749759e-01 -6.56658173e-01 -2.94253618e-01 4.07280140e-02
-2.56956629e-02 -7.55027294e-01 4.95793819e-01 -4.13223803e-01
8.02631974e-01 -2.13658142e+00 2.26220801e-01 -3.25729579e-01
3.12951282e-02 3.63519371e-01 -2.25903675e-01 7.91228175e-01
-5.38482182e-02 3.00928384e-01 -2.08616108e-01 -1.28587201e-01
1.07436337e-01 6.48209453e-01 -3.23607564e-01 3.63796085e-01
6.88380480e-01 1.10524237e+00 -1.13211858e+00 -8.26439142e-01
3.29383075e-01 3.86068970e-01 -6.34191155e-01 8.11266378e-02
-1.57982394e-01 -1.21096544e-01 -6.13740683e-01 5.49783632e-02
3.93668078e-02 -2.95807034e-01 1.25290543e-01 2.45594800e-01
-2.08994020e-02 7.34564185e-01 -3.60043108e-01 1.91094947e+00
-6.94548965e-01 1.22082806e+00 -3.57620627e-01 -1.49647510e+00
9.97792959e-01 6.08600199e-01 2.19142109e-01 -2.74821639e-01
2.11359739e-01 8.49652514e-02 2.73799181e-01 -7.03988254e-01
6.30463660e-01 -7.90691793e-01 -2.39172548e-01 4.59556937e-01
9.08384442e-01 -5.83016217e-01 2.87211388e-01 3.99507225e-01
1.12967861e+00 -5.45494556e-02 4.33543235e-01 -3.23765516e-01
3.82049084e-01 2.73464084e-01 2.98868008e-02 5.10989785e-01
-1.29149064e-01 5.29165149e-01 4.08081859e-01 -4.61743414e-01
-1.29377699e+00 -1.15023935e+00 -5.58297455e-01 1.12404943e+00
-3.32335413e-01 -5.99841177e-01 -3.97173017e-01 -7.33379006e-01
2.37833664e-01 8.78885329e-01 -6.38589025e-01 -9.13643762e-02
-1.89404503e-01 -3.22668344e-01 5.01848757e-01 8.70446801e-01
9.08971857e-03 -1.56041086e+00 -2.75976956e-01 3.86788607e-01
5.18096030e-01 -9.73699749e-01 -6.07556365e-02 4.28191543e-01
-9.85678971e-01 -8.23999584e-01 -6.23100698e-01 -1.27951908e+00
8.38446915e-01 8.45403448e-02 1.17062414e+00 2.02729419e-01
-5.42690098e-01 6.92768574e-01 -5.02216101e-01 -5.39802194e-01
-4.74199831e-01 1.18567921e-01 8.91034156e-02 -4.62268502e-01
1.10678589e+00 -5.91169775e-01 -3.92847173e-02 -2.98390955e-01
-1.04622912e+00 -3.97357643e-01 5.65268695e-01 1.10578430e+00
2.18228344e-03 -3.38763088e-01 9.58784699e-01 -1.07307959e+00
8.05053353e-01 -5.36516547e-01 -6.59066364e-02 -5.62354624e-02
-3.69358301e-01 1.55609950e-01 6.55739963e-01 -2.15225324e-01
-8.66170466e-01 -1.93446532e-01 -4.23429757e-01 -1.70768172e-01
-4.88126367e-01 7.51335502e-01 3.39563966e-01 2.33357802e-01
5.85379899e-01 3.32899094e-01 1.53032824e-01 -2.49077916e-01
5.24906814e-01 9.45443153e-01 8.12267363e-02 -5.58616340e-01
8.47259521e-01 1.56164601e-01 -4.21990484e-01 -1.37880421e+00
-1.01344347e+00 -4.76096511e-01 -9.59341109e-01 3.07031572e-01
1.29820085e+00 -8.56476903e-01 -6.60440326e-03 -1.94726646e-01
-1.27122629e+00 -2.11986691e-01 -5.85808635e-01 7.86886156e-01
-7.39239573e-01 3.41234237e-01 -6.41166925e-01 -4.66246456e-01
-8.39610472e-02 -5.81034243e-01 8.43161583e-01 1.71235487e-01
-5.08986712e-01 -1.66143990e+00 3.55536729e-01 1.52477816e-01
5.25054932e-01 -3.41363668e-01 8.82260919e-01 -1.01641202e+00
-2.77566135e-01 -4.45373774e-01 -2.29257673e-01 8.27967942e-01
4.73772049e-01 -3.46470848e-02 -8.82947147e-01 -2.78304338e-01
-2.24503964e-01 -1.02962446e+00 1.07969773e+00 6.86821761e-03
9.51597750e-01 -6.99063316e-02 -3.21113348e-01 2.83503324e-01
1.48423016e+00 -2.17458457e-01 8.69633079e-01 2.53730327e-01
4.07992929e-01 3.80629689e-01 3.36983413e-01 2.94246897e-02
2.34138831e-01 1.31177843e-01 -9.59270373e-02 1.06032953e-01
6.47094324e-02 -3.42842251e-01 4.31729525e-01 1.23195076e+00
-7.36729950e-02 -9.79139134e-02 -8.46196294e-01 1.06217647e+00
-1.37126112e+00 -1.19260836e+00 3.87996763e-01 1.65658855e+00
1.17961395e+00 3.49933386e-01 -4.85253423e-01 -1.23858929e-01
3.40962291e-01 5.54347575e-01 -6.80406913e-02 -1.03624129e+00
2.33151734e-01 7.37723351e-01 2.22634211e-01 4.11128819e-01
-1.11985159e+00 1.04422855e+00 6.92947054e+00 2.86592335e-01
-8.88607025e-01 2.02082589e-01 2.56765068e-01 2.68890619e-01
-4.81181711e-01 1.34941250e-01 -7.62979865e-01 3.02442044e-01
1.32398164e+00 -4.70709205e-01 4.12566438e-02 7.88735867e-01
-2.15851679e-01 1.34909704e-01 -1.43898058e+00 7.22966433e-01
2.66734838e-01 -1.48061717e+00 1.85872898e-01 -6.98127300e-02
7.57700026e-01 2.97736794e-01 -3.05976331e-01 7.91948318e-01
6.19466603e-01 -1.31910372e+00 -1.44896567e-01 4.26899225e-01
6.41384840e-01 -5.21461725e-01 8.07574630e-01 4.03391033e-01
-6.97058856e-01 1.67798057e-01 -1.14955235e+00 -4.59130943e-01
1.83648676e-01 2.52435952e-01 -1.06042469e+00 2.50351548e-01
3.21738958e-01 1.15281570e+00 -5.23304820e-01 6.10875249e-01
-5.20929635e-01 7.62101233e-01 -2.22842142e-01 -5.20748973e-01
6.23416305e-01 -3.81729528e-02 1.32334962e-01 1.38630843e+00
-5.97746000e-02 -2.84866579e-02 1.85283516e-02 6.53582633e-01
-3.95613074e-01 3.11261505e-01 -1.34986174e+00 -6.64310753e-01
1.58862785e-01 1.02503824e+00 -2.73336828e-01 -6.15289330e-01
-1.00591958e+00 8.36724877e-01 8.14106941e-01 1.51998237e-01
-5.58770239e-01 -7.52528369e-01 6.34213030e-01 -1.42859697e-01
7.40968406e-01 -5.28072655e-01 1.09249838e-01 -1.38697767e+00
-1.95266724e-01 -3.70814413e-01 4.60834429e-02 -8.65900576e-01
-1.76730800e+00 3.95435810e-01 7.44898394e-02 -8.97139192e-01
-5.36777556e-01 -9.38643813e-01 -9.88417506e-01 7.21509039e-01
-1.51366246e+00 -8.11819911e-01 3.35150480e-01 4.96358335e-01
8.24417591e-01 -4.17208999e-01 1.37778592e+00 -4.86426540e-02
-1.65131673e-01 3.55537772e-01 2.13930696e-01 6.52497530e-01
7.55237937e-01 -1.32381451e+00 2.35498145e-01 4.21786934e-01
7.83692181e-01 8.81927073e-01 6.11559391e-01 -1.75923005e-01
-1.39196444e+00 -8.93846273e-01 1.32448459e+00 -6.45123959e-01
1.08474731e+00 -3.63582432e-01 -9.33421791e-01 1.22513616e+00
7.39011526e-01 8.20260122e-02 8.31544638e-01 4.28642809e-01
-3.95413399e-01 2.74432246e-02 -7.80142665e-01 3.51076573e-01
6.27055526e-01 -9.16241586e-01 -1.55068338e+00 5.47467172e-01
6.48666501e-01 2.55729049e-01 -1.14761889e+00 -3.53531875e-02
2.28805169e-01 -4.90546048e-01 1.02894878e+00 -1.28038287e+00
1.02716589e+00 3.37532759e-01 -2.07244128e-01 -1.45338786e+00
-2.43812606e-01 -4.84120063e-02 2.40988567e-01 1.24675298e+00
5.34973919e-01 -7.74132133e-01 5.93021452e-01 2.49970630e-01
-8.28631371e-02 -6.09074533e-01 -9.29318666e-01 -7.19848454e-01
7.18598545e-01 -2.19266161e-01 6.36617094e-02 1.12684536e+00
4.36185688e-01 9.30208206e-01 -2.37112548e-02 -2.08004445e-01
4.91486341e-01 1.45612448e-01 8.09899211e-01 -1.30050445e+00
-4.47961614e-02 -2.82572389e-01 -9.27830517e-01 -1.13923335e+00
6.78496242e-01 -1.32716680e+00 2.73439705e-01 -1.79324460e+00
2.64715225e-01 -2.64268756e-01 -3.41279000e-01 4.72710192e-01
-1.04199618e-01 2.99156368e-01 1.61069259e-01 -9.78695750e-02
-4.83311713e-01 7.12539136e-01 1.18364835e+00 -4.07327801e-01
3.59557599e-01 -4.48103011e-01 -6.07252598e-01 8.36580873e-01
9.91516232e-01 -5.56335926e-01 -4.09194350e-01 -7.21012294e-01
-2.61460971e-02 -1.73741177e-01 1.13138311e-01 -7.83446312e-01
-8.37409496e-03 -4.43588756e-02 5.51948607e-01 -2.70714849e-01
4.73199487e-01 -7.68218637e-01 -7.71951556e-01 1.49664238e-01
-6.21886313e-01 -1.16599560e-01 2.28019416e-01 5.72448015e-01
-5.57597458e-01 -8.63286555e-01 5.22945166e-01 -3.59876037e-01
-7.73051739e-01 2.04390064e-01 -6.88323200e-01 2.40549281e-01
9.67999816e-01 -1.05684564e-01 -1.46984726e-01 -3.52642924e-01
-8.52317274e-01 1.10858992e-01 5.61253667e-01 4.33720797e-01
8.51391435e-01 -1.24397349e+00 -8.64013672e-01 8.80904198e-02
1.60215110e-01 -2.18273461e-01 -2.24798262e-01 6.04335785e-01
-5.41206837e-01 7.58412898e-01 -3.55709583e-01 -4.36720401e-01
-8.42078686e-01 8.40859890e-01 -1.15094304e-01 -1.71080962e-01
-7.54662812e-01 8.52603316e-01 -4.00595553e-02 -4.55145657e-01
-9.93521512e-02 -2.63391793e-01 -1.74910218e-01 6.08090870e-02
5.61394513e-01 -3.82022321e-01 -3.36934507e-01 -5.39536119e-01
-1.92708284e-01 2.37500086e-01 -4.21250850e-01 1.30164906e-01
1.93856692e+00 1.40802175e-01 -1.22772790e-01 8.04659784e-01
1.79146862e+00 -2.27459773e-01 -8.90758216e-01 -1.97539657e-01
8.72238502e-02 -3.68082047e-01 -7.50707462e-02 -1.10288233e-01
-6.68785989e-01 1.32623947e+00 -5.38252965e-02 4.03875411e-01
5.50891280e-01 2.97835052e-01 6.56695485e-01 9.09129798e-01
5.15194833e-01 -9.58175361e-01 9.67568308e-02 8.05639923e-01
7.73270130e-01 -1.42789435e+00 2.92124242e-01 1.10116899e-01
-4.13652658e-01 1.47258902e+00 3.00505400e-01 -7.58524299e-01
9.21528041e-01 9.08037350e-02 -1.67984009e-01 -2.25057513e-01
-1.01246953e+00 -2.42504120e-01 -4.61463910e-03 6.28218532e-01
8.06221783e-01 -1.18738882e-01 -3.76868516e-01 1.35657489e-01
-1.18814781e-01 1.53739274e-01 7.45920539e-01 1.19939351e+00
-8.17697048e-01 -1.24966872e+00 6.05544411e-02 8.50866199e-01
-4.44281489e-01 -3.40453953e-01 -3.39003384e-01 1.00732422e+00
-1.33985788e-01 5.16223192e-01 4.62133676e-01 -8.96802929e-04
-1.01029322e-01 5.70984960e-01 8.62502933e-01 -1.40588260e+00
-3.37387323e-01 -5.30231893e-01 4.02265370e-01 -2.09806383e-01
-7.84355998e-01 -5.37639439e-01 -1.20950353e+00 -1.74121454e-01
-2.04445913e-01 3.03869009e-01 5.26421726e-01 1.08242536e+00
-1.24444030e-01 3.07889104e-01 3.89767587e-01 -9.83968496e-01
-7.72731483e-01 -1.45371807e+00 -6.70990586e-01 5.86697102e-01
1.72986120e-01 -6.27837956e-01 -3.96132201e-01 2.85158157e-01] | [10.711503028869629, 8.774624824523926] |
0e9b9cbd-a748-430f-91ea-ed57a6832dee | more-recent-advances-in-hyper-graph | 2205.13202 | null | https://arxiv.org/abs/2205.13202v3 | https://arxiv.org/pdf/2205.13202v3.pdf | More Recent Advances in (Hyper)Graph Partitioning | In recent years, significant advances have been made in the design and evaluation of balanced (hyper)graph partitioning algorithms. We survey trends of the last decade in practical algorithms for balanced (hyper)graph partitioning together with future research directions. Our work serves as an update to a previous survey on the topic. In particular, the survey extends the previous survey by also covering hypergraph partitioning and streaming algorithms, and has an additional focus on parallel algorithms. | ['Lars Gottesbüren', 'Dorothea Wagner', 'Daniel Seemaier', 'Christian Schulz', 'Sebastian Schlag', 'Peter Sanders', 'Henning Meyerhenke', 'Tobias Heuer', 'Marcelo Fonseca Faraj', 'Karen D. Devine', 'Ümit V. Çatalyürek'] | 2022-05-26 | null | null | null | null | ['hypergraph-partitioning', 'graph-partitioning'] | ['graphs', 'graphs'] | [ 3.42476726e-01 5.31426311e-01 -6.42404020e-01 -2.65520483e-01
-3.33175242e-01 -6.82524383e-01 1.75735489e-01 5.64717114e-01
2.08522797e-01 7.08933234e-01 1.60804182e-01 -4.12261546e-01
-3.81038249e-01 -1.07939160e+00 -2.16614306e-01 -5.53446054e-01
-7.44011343e-01 1.07596910e+00 8.66702199e-01 -1.38592139e-01
2.36486837e-01 5.77120483e-01 -1.45075512e+00 3.22045624e-01
4.41145480e-01 3.12538922e-01 -4.06764388e-01 1.12523139e+00
-2.59001702e-01 7.01657414e-01 -6.96707070e-01 -5.10815918e-01
1.27200872e-01 -6.74029768e-01 -1.42760587e+00 5.25115788e-01
2.33371481e-01 8.49160329e-02 -7.93440759e-01 8.13746810e-01
5.26484191e-01 3.21698412e-02 1.67457938e-01 -1.74519682e+00
-4.14687246e-02 1.24276805e+00 -1.21745837e+00 4.78581071e-01
5.76180160e-01 -4.10200953e-01 9.83374178e-01 -4.17643905e-01
6.59626186e-01 1.22458458e+00 6.33167922e-01 5.31788729e-02
-1.21424675e+00 -3.53794903e-01 5.29249430e-01 3.29749137e-01
-1.92780578e+00 -2.51103252e-01 4.82688814e-01 -3.74503359e-02
1.36710167e+00 7.68542826e-01 1.42204773e+00 -3.24773908e-01
2.49845032e-02 9.04021442e-01 6.72967553e-01 -6.75363243e-01
2.61884600e-01 -1.42289475e-01 5.72499394e-01 3.93066019e-01
1.09042072e+00 -2.04213157e-01 -2.99323648e-01 -5.98720491e-01
4.69639629e-01 -6.72250450e-01 -9.87274721e-02 -1.12783360e+00
-1.05630481e+00 8.78400326e-01 6.90825060e-02 3.47195983e-01
-1.14783555e-01 3.82897258e-01 8.21934700e-01 4.74441946e-01
4.65123057e-01 1.68867797e-01 -1.03642657e-01 -1.07698748e-02
-1.27741027e+00 5.71885705e-01 1.46565485e+00 1.57418013e+00
5.51625252e-01 3.05882897e-02 -2.51739882e-02 4.75631297e-01
1.15811951e-01 3.16950977e-01 -3.97644550e-01 -8.09351087e-01
3.96801412e-01 4.91763264e-01 -2.83395082e-01 -1.05681145e+00
-5.30848265e-01 -2.26520747e-01 -6.94131017e-01 -3.71982366e-01
-1.11063071e-01 1.07377522e-01 -7.10762024e-01 9.72446978e-01
8.10255289e-01 -2.22358868e-01 -2.97280818e-01 3.61715436e-01
1.02402604e+00 7.11153090e-01 -2.21427277e-01 -7.78933823e-01
1.12460780e+00 -1.02882326e+00 -8.96409094e-01 3.07251632e-01
5.14520466e-01 -8.12728882e-01 3.17704976e-01 4.50677335e-01
-1.64780462e+00 -3.55939604e-02 -1.04092216e+00 3.39415699e-01
-3.46411854e-01 -9.96911108e-01 1.05001748e+00 1.48976707e+00
-1.62028229e+00 -1.65923573e-02 -7.92784274e-01 -1.13845956e+00
9.96511877e-02 7.69179940e-01 1.69155881e-01 -2.54914254e-01
-9.60995972e-01 4.41305637e-01 7.13771820e-01 -4.65943575e-01
-3.73969615e-01 -8.29816580e-01 -7.68838167e-01 1.27378404e-01
7.11098671e-01 -8.88487816e-01 1.27622938e+00 -2.96961546e-01
-1.13556659e+00 9.26463664e-01 9.96627361e-02 -4.13069993e-01
6.53324351e-02 5.93132019e-01 -5.28965235e-01 3.94455969e-01
-3.87979716e-01 5.36248326e-01 -1.63702518e-01 -1.22382307e+00
-8.11740458e-01 -3.61571461e-01 3.16150755e-01 5.39828718e-01
-4.54306662e-01 5.53841233e-01 -7.86114156e-01 -5.07904053e-01
1.71781331e-01 -9.63291526e-01 -4.84948128e-01 -7.59118676e-01
-6.50000691e-01 -4.13969666e-01 4.35566634e-01 3.02753836e-01
2.11854696e+00 -1.75052011e+00 2.57524014e-01 5.62118411e-01
6.26737416e-01 -6.02677539e-02 1.43591136e-01 1.39550829e+00
-6.76542148e-02 3.44927460e-01 5.02415188e-02 1.95374876e-01
-1.05717499e-02 9.09706950e-02 -6.99774623e-02 7.93520808e-01
-9.57813084e-01 5.47116876e-01 -8.07654738e-01 -7.14737356e-01
2.87487566e-01 -9.42303762e-02 -5.94894052e-01 -3.98659885e-01
-6.90285712e-02 -4.02469814e-01 -9.39873829e-02 5.71488798e-01
9.78842676e-01 -3.23528975e-01 6.83505237e-01 3.75551105e-01
-1.85914382e-01 2.49319419e-01 -1.65080345e+00 1.16600406e+00
4.33067739e-01 2.75550574e-01 3.51175874e-01 -1.02649796e+00
5.21948099e-01 5.89230597e-01 1.07968354e+00 -2.16527328e-01
1.44593744e-02 1.97466344e-01 -6.30061235e-03 -1.32367998e-01
7.98126876e-01 -3.39590153e-03 -2.50407845e-01 8.66592050e-01
-3.29023689e-01 -6.27673268e-01 1.11330807e+00 8.49644721e-01
1.24171460e+00 -7.10699797e-01 7.29349613e-01 -7.03698516e-01
6.18269622e-01 5.74232280e-01 2.53618713e-02 7.88736463e-01
-3.25087368e-01 3.19973171e-01 6.44461989e-01 -6.65151715e-01
-9.73976135e-01 -7.36141443e-01 -9.61524546e-02 1.05326557e+00
7.54950941e-01 -1.12586617e+00 -8.99012506e-01 -4.94450986e-01
3.94528098e-02 1.58572167e-01 -6.41792834e-01 3.48929346e-01
-7.65023828e-01 -1.12783289e+00 4.20885086e-01 7.49730229e-01
-6.15721606e-02 -7.46592879e-01 -3.04817379e-01 1.09813966e-01
-1.68273628e-01 -8.69052351e-01 -4.73558754e-01 -3.08682546e-02
-1.13954473e+00 -1.36238527e+00 -3.39846849e-01 -1.04554141e+00
6.12759411e-01 1.21707582e+00 1.47014415e+00 2.85644650e-01
-2.90862352e-01 5.84697545e-01 -5.94297409e-01 -3.39160353e-01
-3.61263044e-02 6.52665555e-01 -2.78192639e-01 -5.99617362e-01
6.58719301e-01 -3.54914248e-01 -5.08625567e-01 4.86612022e-01
-9.46499944e-01 2.75708884e-02 -5.02047129e-03 3.42079848e-01
7.95977294e-01 7.64941156e-01 3.70917767e-01 -1.57727456e+00
5.35293519e-01 -7.27550745e-01 -6.67502403e-01 1.50324941e-01
-7.97041416e-01 -5.12654126e-01 2.11199552e-01 1.35016837e-03
-8.97603154e-01 -3.06805074e-01 -3.70750725e-02 2.42789567e-01
1.55018747e-01 5.50384283e-01 -2.20071655e-02 -3.71612459e-01
3.93780947e-01 -3.74525100e-01 -3.49765152e-01 4.67378870e-02
4.74603176e-01 3.84951949e-01 -6.97032735e-02 -3.27385157e-01
6.65073156e-01 8.48411322e-01 1.70749292e-01 -9.70014691e-01
-1.07165128e-01 -9.95961189e-01 -5.48340619e-01 -2.54645407e-01
1.23535253e-01 -4.45238829e-01 -5.15267909e-01 2.39903346e-01
-6.35671377e-01 -4.15555120e-01 -4.76732552e-01 3.23209852e-01
-6.19915962e-01 5.08732915e-01 -8.04693699e-01 -5.45667946e-01
-1.96222991e-01 -8.41017663e-01 7.65568197e-01 1.04626156e-01
-6.10082209e-01 -1.20419109e+00 6.78233564e-01 1.52732134e-01
3.18882406e-01 1.39118224e-01 9.23065782e-01 -6.29106581e-01
-7.31585383e-01 -3.70855451e-01 -2.16981739e-01 -7.27834761e-01
-3.54923427e-01 2.97335953e-01 -4.89639193e-01 -8.39666247e-01
-2.97885627e-01 -5.46701625e-02 6.40512764e-01 8.50397229e-01
9.69835043e-01 -3.60339917e-02 -1.10757947e+00 4.69031870e-01
1.74331427e+00 4.54306126e-01 6.35105073e-01 1.89895719e-01
4.02226567e-01 7.58785665e-01 5.88794172e-01 4.32026505e-01
6.91444218e-01 4.28393126e-01 1.34826943e-01 -3.42759341e-01
-3.44993323e-01 1.83039665e-01 -3.40173304e-01 1.12460423e+00
4.04395089e-02 -9.92660344e-01 -8.86583626e-01 9.86607373e-01
-1.83399332e+00 -1.00924230e+00 -6.87313855e-01 2.03964257e+00
3.68895203e-01 -1.20594101e-02 9.34588194e-01 7.67062545e-01
1.27252972e+00 4.48060095e-01 -3.56615663e-01 -7.72120833e-01
1.24305703e-01 4.84071262e-02 8.28937471e-01 6.44416571e-01
-9.43279922e-01 7.69663692e-01 8.93657398e+00 9.23551261e-01
-4.85086739e-01 -1.34990707e-01 4.22843426e-01 -2.68994451e-01
-8.20412874e-01 2.02297851e-01 -5.96731067e-01 -2.03894705e-01
1.03351748e+00 -9.87539530e-01 3.62096071e-01 5.47123313e-01
-1.28357381e-01 -9.48559865e-02 -8.09451520e-01 6.42310858e-01
3.11827272e-01 -1.24837494e+00 -1.14616878e-01 5.23998082e-01
1.05409479e+00 -7.43528828e-02 -2.07758695e-01 1.96646482e-01
7.27586448e-01 -6.86260283e-01 2.23275289e-01 -4.45918947e-01
7.65999675e-01 -1.13442576e+00 8.33300769e-01 -5.46634346e-02
-1.94397449e+00 -2.90074106e-02 -4.14567053e-01 -3.11994493e-01
4.15107042e-01 5.38654864e-01 -6.44248307e-01 9.47925508e-01
9.08961058e-01 3.85803491e-01 -2.32161265e-02 1.59878135e+00
2.93640256e-01 6.34330988e-01 -4.57396626e-01 -6.51728734e-02
-7.15411827e-02 -8.92107114e-02 6.41699195e-01 1.25518024e+00
-2.81429619e-01 5.89640200e-01 8.08174670e-01 -2.43263505e-03
-1.07535720e-02 6.73574924e-01 -7.12038994e-01 6.38470873e-02
7.61339307e-01 1.04191208e+00 -1.67041183e+00 -6.39912963e-01
-8.09251070e-01 1.09418586e-01 1.95439905e-02 3.00616533e-01
-7.19488382e-01 -8.79225135e-01 1.62667885e-01 4.38253403e-01
4.35413361e-01 -1.19993642e-01 -6.96772635e-01 -6.00951314e-01
-5.35616994e-01 -9.65611339e-01 1.26801908e+00 -8.51902515e-02
-8.19903255e-01 7.11953402e-01 6.87970161e-01 -5.92728198e-01
-9.16109979e-03 9.98660624e-02 -1.63843691e-01 2.16217145e-01
-9.72437739e-01 -6.74173415e-01 -2.71723479e-01 3.16808850e-01
3.84618968e-01 2.36875176e-01 5.37669659e-01 3.35203528e-01
-3.12506646e-01 3.36129040e-01 -9.30479988e-02 -4.43624288e-01
5.52366257e-01 -1.23648071e+00 7.23365843e-01 9.12814021e-01
-5.33281714e-02 5.39705634e-01 7.56717682e-01 -7.96307564e-01
-1.60117126e+00 -7.47216046e-01 8.10464621e-01 -3.74699272e-02
6.06725931e-01 -3.14525366e-01 -5.78556895e-01 9.39493477e-01
7.17122555e-01 -2.83838898e-01 9.74560201e-01 3.87068689e-01
2.13992685e-01 -1.77456841e-01 -1.21943426e+00 4.32199270e-01
1.27513480e+00 1.29524678e-01 2.62111239e-02 2.83708751e-01
8.15870047e-01 -5.45036077e-01 -1.05014396e+00 4.06210661e-01
6.67504549e-01 -8.80518258e-01 8.26005399e-01 -3.39745045e-01
-6.13613725e-01 -2.81809926e-01 1.45554870e-01 -7.75235176e-01
-7.35739827e-01 -1.07030737e+00 -2.43472457e-01 8.44015062e-01
1.13479257e-01 -6.00051701e-01 1.41854572e+00 1.77880377e-01
-1.02308378e-01 -9.68259335e-01 -3.85997355e-01 -6.12726748e-01
4.59437370e-02 -1.12562656e-01 9.98499572e-01 7.35988557e-01
7.88102388e-01 2.58725584e-01 4.60388418e-03 -5.62613234e-02
7.57958472e-01 8.42023194e-01 1.07415843e+00 -1.27468181e+00
6.47045746e-02 -8.42556536e-01 -5.99268317e-01 -1.21384609e+00
-2.55333483e-01 -9.24248815e-01 -3.38987231e-01 -2.09760475e+00
6.74000025e-01 -2.89816797e-01 -5.49979173e-02 1.23714849e-01
1.48292035e-01 7.83235312e-01 -1.67265274e-02 -3.85800004e-02
-1.11290824e+00 -9.52646658e-02 1.11020339e+00 -7.90256169e-03
-3.80465627e-01 3.15915048e-01 -9.21582222e-01 3.82452995e-01
9.90221739e-01 -3.89600724e-01 -1.00933075e+00 -2.30995417e-01
5.49807131e-01 3.70350093e-01 -7.68576384e-01 -6.15844786e-01
6.05962574e-01 -3.81085217e-01 -3.19794893e-01 -1.34082961e+00
-1.93851218e-01 -7.50278533e-01 4.69869733e-01 7.49352574e-01
-1.34799019e-01 7.28754401e-01 1.75609916e-01 6.52414262e-01
-3.53405923e-01 1.92990135e-02 7.49700904e-01 -2.73927450e-01
-5.24951577e-01 6.05867922e-01 -7.31622159e-01 3.67238194e-01
1.64227092e+00 -8.56442153e-01 -3.49834442e-01 -4.32541668e-01
-8.13907683e-01 7.44984031e-01 8.42604578e-01 -1.40328869e-01
4.12145942e-01 -8.42704952e-01 -7.25676298e-01 -5.90030998e-02
1.71287850e-01 -1.42672315e-01 2.96971738e-01 8.67901921e-01
-1.09722567e+00 8.38863313e-01 4.97999713e-02 -4.81545419e-01
-1.79691565e+00 1.27727330e+00 6.98180646e-02 -3.13646197e-01
-5.84366024e-01 7.08220065e-01 1.94582298e-01 7.31683196e-03
3.79263401e-01 1.52175128e-01 -9.29752737e-02 -2.35727727e-01
5.07113338e-01 1.23796427e+00 1.61337942e-01 -3.32099229e-01
-6.26192629e-01 3.42374831e-01 -1.32581949e-01 -2.16189455e-02
1.12332225e+00 -7.03969538e-01 -8.62315357e-01 5.22464693e-01
7.13448763e-01 3.23849559e-01 -2.52410099e-02 -4.37959507e-02
-7.43005276e-02 -6.03765726e-01 -1.03257649e-01 -1.34409770e-01
-1.08399570e+00 3.00789058e-01 -3.24924380e-01 1.41497040e+00
1.21390045e+00 3.81120801e-01 8.24143291e-01 -6.94809183e-02
7.73507178e-01 -9.62132692e-01 -2.33331367e-01 2.62618244e-01
3.12078953e-01 -3.84762019e-01 6.35649979e-01 -1.44223595e+00
-3.09963167e-01 9.08086836e-01 5.17652750e-01 -6.20492026e-02
1.08783937e+00 7.66146660e-01 -3.88623267e-01 -8.23373437e-01
-9.69316602e-01 -5.97626492e-02 -1.73436373e-01 7.10505784e-01
5.37450433e-01 2.91997313e-01 -7.71596968e-01 -2.42834240e-01
-4.09050673e-01 -3.19021702e-01 9.07020152e-01 1.06634641e+00
-6.79817557e-01 -1.46361530e+00 -4.83483195e-01 4.22728479e-01
-3.39396447e-01 1.14251934e-02 -6.32954895e-01 1.12511516e+00
-3.44616890e-01 1.24304354e+00 3.08416635e-01 -1.26708880e-01
3.27484816e-01 -5.63784719e-01 8.70242238e-01 -8.63743186e-01
-5.73096275e-01 2.05950499e-01 5.75138986e-01 -4.09207672e-01
-4.31629866e-01 -7.84426987e-01 -9.62588906e-01 -1.33434653e+00
-6.41860187e-01 7.28323340e-01 1.44171312e-01 7.19420984e-02
2.09635608e-02 7.62878656e-01 4.60010350e-01 -4.53991443e-01
-2.21060123e-03 -5.00495553e-01 -1.07394326e+00 -2.62039661e-01
-1.33449942e-01 -4.63937879e-01 -6.31799400e-02 -2.73323148e-01] | [6.991598129272461, 5.210996627807617] |
f26ab1c6-7cbd-46e9-b975-cdd8d17eb869 | the-webnlg-challenge-generating-text-from-rdf | null | null | https://aclanthology.org/W17-3518 | https://aclanthology.org/W17-3518.pdf | The WebNLG Challenge: Generating Text from RDF Data | The WebNLG challenge consists in mapping sets of RDF triples to text. It provides a common benchmark on which to train, evaluate and compare {``}microplanners{''}, i.e. generation systems that verbalise a given content by making a range of complex interacting choices including referring expression generation, aggregation, lexicalisation, surface realisation and sentence segmentation. In this paper, we introduce the microplanning task, describe data preparation, introduce our evaluation methodology, analyse participant results and provide a brief description of the participating systems. | ['Laura Perez-Beltrachini', 'Claire Gardent', 'Shashi Narayan', 'Anastasia Shimorina'] | 2017-09-01 | null | null | null | ws-2017-9 | ['referring-expression-generation'] | ['computer-vision'] | [ 4.48561937e-01 8.23121607e-01 2.72818983e-01 -8.28496993e-01
-1.10126019e+00 -9.38265026e-01 1.28690839e+00 6.26068234e-01
-5.15792012e-01 1.19360673e+00 7.71343291e-01 -6.81283399e-02
-2.16426358e-01 -8.85988951e-01 -6.93067729e-01 1.96849480e-02
2.88306117e-01 1.01561999e+00 1.53927743e-01 -6.19418502e-01
5.20874448e-02 -1.33053273e-01 -1.77195013e+00 1.06496727e+00
9.33787644e-01 9.54854131e-01 -2.39067703e-01 6.53187573e-01
-6.48454428e-01 1.10635972e+00 -1.05401015e+00 -9.42966521e-01
-4.96104002e-01 -4.82530832e-01 -1.53918266e+00 -4.09599811e-01
4.61460292e-01 2.61854082e-01 3.71668786e-01 8.88558924e-01
8.54735494e-01 5.62268257e-01 5.31945705e-01 -1.25222397e+00
-6.24316573e-01 1.60545170e+00 4.63705361e-01 2.29139589e-02
1.49395776e+00 -1.24684088e-01 1.12829673e+00 -6.94583297e-01
1.30757153e+00 1.53776526e+00 4.78740335e-01 1.15223682e+00
-1.06894612e+00 -2.40599841e-01 -5.82640730e-02 8.45157653e-02
-1.22101450e+00 -8.23679209e-01 2.30960980e-01 -1.39167294e-01
1.67938495e+00 8.44451487e-01 4.26886141e-01 1.46766484e+00
-2.89496154e-01 8.18026781e-01 1.05812728e+00 -7.38291383e-01
4.24882248e-02 7.43510351e-02 2.39264935e-01 3.94553125e-01
5.33301011e-03 -3.96502644e-01 -9.97741044e-01 -1.62556946e-01
3.04302201e-02 -1.16825461e+00 -3.45687181e-01 1.98839203e-01
-1.38690186e+00 4.29107577e-01 1.66804478e-01 9.79486927e-02
-4.55973983e-01 2.37506241e-01 7.83391118e-01 2.15811372e-01
2.34237596e-01 7.32731700e-01 -3.90008897e-01 -3.47565889e-01
-5.08759201e-01 1.04677999e+00 1.29793668e+00 1.44086409e+00
3.41508567e-01 -4.10187989e-01 -7.04775393e-01 6.67775989e-01
4.77623850e-01 4.13744897e-01 5.09975910e-01 -8.84293735e-01
1.02571082e+00 7.02702820e-01 2.02761218e-01 -4.05286402e-01
-6.52178764e-01 4.93119448e-01 -3.39772314e-01 -3.08438271e-01
3.56849164e-01 -5.40913880e-01 -6.80299997e-01 1.49911439e+00
3.98376048e-01 -4.60906029e-01 6.24144316e-01 6.22705221e-01
1.91413641e+00 5.31650424e-01 6.84270799e-01 -2.65308827e-01
1.52038789e+00 -4.11805809e-01 -1.10380661e+00 -1.07970491e-01
9.57378268e-01 -6.39248312e-01 1.06789362e+00 3.86892967e-02
-1.62063694e+00 -3.14814866e-01 -7.05271006e-01 -5.92247963e-01
-8.06188464e-01 -3.39308530e-01 6.14997208e-01 4.86296535e-01
-1.12264466e+00 5.91099024e-01 -5.61515868e-01 -6.12761021e-01
1.00394361e-01 3.34049135e-01 -5.21719515e-01 1.22071318e-01
-1.76001990e+00 1.04862511e+00 1.23995280e+00 -4.55658175e-02
-2.17372268e-01 -8.40967357e-01 -1.11117721e+00 -4.27790225e-01
3.51326793e-01 -1.04469585e+00 1.57371557e+00 -7.15110064e-01
-1.37305367e+00 1.31157231e+00 -1.07089706e-01 -7.07555592e-01
5.62947929e-01 -1.43706933e-01 -7.45719194e-01 -2.36217663e-01
2.32609004e-01 1.02820873e+00 6.58372343e-02 -9.29146528e-01
-7.48734474e-01 -1.80987537e-01 3.13952148e-01 7.14619935e-01
6.56174779e-01 5.98831475e-01 -6.68570623e-02 -4.31659728e-01
-2.15646416e-01 -4.60906744e-01 1.48123041e-01 -9.50235844e-01
-1.04597902e+00 -8.46446633e-01 3.19652915e-01 -4.83103991e-01
1.35758603e+00 -1.56251669e+00 2.49955162e-01 2.04636291e-01
1.73648685e-01 -2.24533137e-02 2.27492675e-01 1.00430870e+00
-1.98595688e-01 7.52059102e-01 -1.22543290e-01 -1.64087892e-01
7.19867349e-01 2.37089008e-01 -1.66988537e-01 -2.99855202e-01
1.84411973e-01 1.33609414e+00 -9.67644751e-01 -8.03674936e-01
1.12938225e-01 3.67069960e-01 -1.71510145e-01 2.19493136e-01
-8.82622778e-01 2.40351260e-01 -5.04952371e-01 4.17537183e-01
5.54505084e-03 8.04228038e-02 2.67345697e-01 -4.33551699e-01
-2.04030737e-01 9.38874245e-01 -1.14243996e+00 1.91662192e+00
-3.53416294e-01 3.24339122e-01 -3.41083825e-01 -1.63755715e-01
7.96370447e-01 6.49147749e-01 -7.75404871e-02 -7.57076323e-01
1.07224271e-01 3.75675470e-01 -4.64504302e-01 -8.33459914e-01
7.85574973e-01 -2.34080523e-01 -8.04759741e-01 5.16444027e-01
1.60290733e-01 -4.05606151e-01 7.39653111e-01 2.76416391e-01
1.18220687e+00 6.41780376e-01 6.25073671e-01 -3.20997000e-01
6.28906429e-01 -1.15892105e-02 -1.42150754e-02 5.28718710e-01
3.04892212e-01 2.74732322e-01 6.48030162e-01 -4.94502157e-01
-7.94777274e-01 -1.00907290e+00 1.67283248e-02 1.22573161e+00
-6.63639009e-02 -8.80701661e-01 -1.07617462e+00 -8.09954643e-01
-3.88671011e-01 1.51038432e+00 -5.13889372e-01 3.18572938e-01
-6.58985078e-01 -6.34638250e-01 9.80875611e-01 3.63767385e-01
5.90574980e-01 -1.78088915e+00 -9.20213938e-01 2.49008939e-01
-8.39549541e-01 -1.31050181e+00 1.02343440e-01 1.82408572e-03
-2.12798715e-01 -1.01732624e+00 1.18238650e-01 -6.40009940e-01
2.66185373e-01 -8.93417895e-01 1.95627451e+00 1.00908481e-01
-2.36401819e-02 2.10586056e-01 -7.10390210e-01 -6.75257921e-01
-6.15976393e-01 3.01347971e-01 -3.24514419e-01 -4.76825446e-01
5.64805806e-01 -1.42168805e-01 -9.75685269e-02 -2.15408549e-01
-1.11087143e+00 3.42534691e-01 -2.84248292e-02 3.49337220e-01
6.60435796e-01 -3.76703024e-01 4.13074464e-01 -1.62688422e+00
1.31493998e+00 -3.69889796e-01 -5.07138260e-02 7.89919317e-01
-2.81444609e-01 2.98031390e-01 3.29049170e-01 4.24205959e-01
-1.36128032e+00 -7.74294585e-02 -5.48510134e-01 6.23346150e-01
-5.63832939e-01 7.50357926e-01 -7.91461587e-01 5.59753478e-01
1.20612955e+00 -2.61041820e-01 -5.47274351e-01 -1.80009916e-01
1.08784604e+00 5.65617263e-01 6.77209496e-01 -1.07537568e+00
3.35738569e-01 -7.80602843e-02 -2.55892962e-01 -5.76712132e-01
-9.15054739e-01 -4.23222631e-02 -7.54841089e-01 -2.44899958e-01
1.15803158e+00 -6.01129830e-01 -7.81509042e-01 -8.74029398e-02
-1.57758808e+00 -5.36371112e-01 -8.91970634e-01 -9.12927985e-02
-8.26521099e-01 -1.61296010e-01 -2.63912976e-01 -5.89344442e-01
-7.59929836e-01 -6.27495825e-01 1.38018203e+00 2.62911528e-01
-1.02871621e+00 -1.24740613e+00 -2.22960836e-03 3.57598335e-01
1.15201317e-01 9.41229582e-01 9.08270597e-01 -1.05490875e+00
3.03867441e-02 3.33183371e-02 -1.67443469e-01 -2.06356928e-01
-2.13440806e-01 2.14006469e-01 -7.40534127e-01 2.26142228e-01
-6.31913841e-01 -8.47850442e-01 2.56487876e-01 -7.05132708e-02
8.61759543e-01 -7.47923911e-01 -3.72157574e-01 3.93973231e-01
1.28061700e+00 1.00236177e-01 8.67943704e-01 3.94953310e-01
6.11286402e-01 1.00729036e+00 4.45522398e-01 2.97287285e-01
8.76643956e-01 6.57888234e-01 2.83504993e-01 5.38981915e-01
-2.29859784e-01 -4.33721632e-01 1.61741257e-01 4.41654980e-01
-1.98401287e-01 -9.34836388e-01 -1.19628894e+00 3.49719048e-01
-1.91562629e+00 -1.08726275e+00 -3.75067741e-01 1.75518525e+00
1.21850610e+00 -9.78780240e-02 1.21883310e-01 -2.30985861e-02
7.74120748e-01 1.93516016e-01 -5.42515703e-02 -8.40141356e-01
-4.35515344e-01 6.44970775e-01 4.73343097e-02 6.73117042e-01
-9.07283902e-01 1.21538889e+00 6.50508785e+00 5.84881246e-01
-4.87223506e-01 1.20941428e-02 3.01893800e-01 -6.06974028e-02
-7.29394555e-01 -2.64141649e-01 -1.04147005e+00 2.59272575e-01
1.50921142e+00 -6.53656840e-01 4.06443715e-01 1.41904041e-01
-6.48866072e-02 5.56570888e-02 -1.53916419e+00 6.99786544e-01
1.83231622e-01 -1.49419320e+00 4.68744606e-01 -6.99892998e-01
4.78240311e-01 -7.61461332e-02 -5.36873341e-01 4.29451525e-01
5.99152982e-01 -1.42856443e+00 1.07377994e+00 8.57828259e-01
8.93875241e-01 -5.40306270e-01 7.29718924e-01 5.39306775e-02
-9.98129904e-01 4.46349770e-01 2.58737892e-01 5.67673892e-02
6.12639844e-01 5.61617836e-02 -6.62044168e-01 1.02768672e+00
5.88960230e-01 2.44632185e-01 -6.99260056e-01 5.68493605e-01
-7.70916164e-01 1.89013973e-01 -3.89181793e-01 -3.97253782e-01
6.58391416e-02 2.01160595e-01 4.26446557e-01 1.63913643e+00
1.59782082e-01 4.30837244e-01 -7.44251087e-02 8.08411002e-01
-4.01670635e-01 4.60953504e-01 -5.01307607e-01 -6.66580945e-02
7.40944326e-01 1.11334002e+00 -5.36862493e-01 -3.90362233e-01
6.83572963e-02 7.15405762e-01 3.79781187e-01 2.73436457e-01
-6.53414965e-01 -8.16379488e-01 1.07696615e-01 2.07326606e-01
-1.68618232e-01 2.10978955e-01 5.91900796e-02 -8.60691428e-01
1.77491769e-01 -9.36001778e-01 7.49725878e-01 -1.20558393e+00
-1.27380908e+00 9.87119675e-01 5.16471803e-01 -4.83154625e-01
-8.75676513e-01 -4.35792565e-01 -5.32318175e-01 8.92152429e-01
-1.00172281e+00 -1.31217408e+00 -4.47057962e-01 5.92271805e-01
2.51950860e-01 2.41219357e-01 1.18027842e+00 6.62058964e-02
-4.13907856e-01 3.03930581e-01 -7.69771278e-01 1.35783717e-01
5.10119319e-01 -1.71444678e+00 7.13054180e-01 5.19652307e-01
1.44948408e-01 6.74138427e-01 1.04071426e+00 -8.34822237e-01
-1.06816697e+00 -1.09968102e+00 1.79568815e+00 -8.32468092e-01
7.29935646e-01 -4.06594753e-01 -7.60491431e-01 1.04536140e+00
6.79473460e-01 -9.71958339e-02 9.57516134e-01 1.83770403e-01
-1.81064695e-01 1.54079869e-01 -1.22704661e+00 5.13739884e-01
1.35552275e+00 -4.87491757e-01 -1.03171349e+00 7.17864037e-01
8.43343973e-01 -1.19967651e+00 -1.21349883e+00 3.01483959e-01
2.86660582e-01 -8.77389967e-01 6.70357764e-01 -1.15790737e+00
5.00142813e-01 -2.01894104e-01 -2.32035518e-01 -1.39760375e+00
2.80488193e-01 -1.06555855e+00 -1.10196479e-01 1.84574771e+00
1.08211648e+00 -3.03710967e-01 4.53692883e-01 1.24342525e+00
-3.89323473e-01 -4.17217612e-01 -1.00849211e+00 -3.78415346e-01
-1.60840969e-03 -8.11768234e-01 1.17023110e+00 8.44025373e-01
6.48281693e-01 7.55808890e-01 3.08485061e-01 -1.84537143e-01
4.54543710e-01 -2.85039693e-01 3.87302577e-01 -1.08707917e+00
1.13165155e-01 -3.99539769e-01 -2.22933307e-01 -3.80772412e-01
2.94906020e-01 -1.33105791e+00 1.29076600e-01 -2.22192717e+00
-4.51130629e-01 -2.94849992e-01 2.91643620e-01 5.91992557e-01
-7.06709623e-02 5.41362315e-02 1.47978589e-01 -1.35870516e-01
-1.02255595e+00 2.10225910e-01 9.37251687e-01 2.28481472e-01
-8.29752684e-02 -2.64580518e-01 -8.06922555e-01 5.22833824e-01
6.20642602e-01 -7.88144842e-02 -4.63821560e-01 -6.94769979e-01
1.16178524e+00 7.61642167e-03 3.24360043e-01 -7.26994991e-01
3.33788246e-01 -1.06628910e-01 2.80611455e-01 -3.40435207e-01
6.80331662e-02 -2.83950597e-01 4.50653106e-01 -2.42452323e-01
-8.94130528e-01 5.89188635e-01 3.23147744e-01 -8.94127414e-02
-1.66342780e-01 -4.91218746e-01 1.60887867e-01 -3.98613721e-01
-8.47914338e-01 -1.60168231e-01 -1.70446366e-01 6.92040443e-01
9.70928013e-01 1.88182145e-02 -6.60487473e-01 -1.46936774e-01
-1.03960907e+00 5.43742418e-01 1.33024693e-01 3.69330823e-01
5.77103794e-01 -1.58289719e+00 -9.63145971e-01 -3.15843433e-01
2.86543012e-01 3.89507324e-01 1.78499538e-02 3.68958384e-01
-8.27955425e-01 4.83744472e-01 -2.36253247e-01 -2.91538499e-02
-1.14865649e+00 7.26821795e-02 4.85605478e-01 -3.31837505e-01
-4.24893022e-01 1.04067504e+00 -5.39790869e-01 -5.96902609e-01
-5.45009859e-02 -3.32889646e-01 -6.63212478e-01 4.01363343e-01
6.75952315e-01 3.76693875e-01 3.71350169e-01 -8.95557523e-01
-5.90972900e-01 2.26837769e-02 2.10412756e-01 -4.98458892e-01
9.53349411e-01 -1.49497464e-01 -4.22909319e-01 6.72382414e-01
6.61441386e-01 -2.48336181e-01 -5.28625250e-01 -7.18717882e-03
6.31631851e-01 2.00314194e-01 -3.68052661e-01 -1.33249581e+00
-3.76286358e-01 3.49824995e-01 -9.21178311e-02 7.13966727e-01
8.09987903e-01 4.21395630e-01 6.65852666e-01 5.90263546e-01
2.40992218e-01 -1.33960819e+00 -6.36360765e-01 6.93866372e-01
1.31310141e+00 -7.71971524e-01 2.42161788e-02 -5.73459804e-01
-8.42497349e-01 1.22525501e+00 6.48126066e-01 3.29095483e-01
1.47101089e-01 2.39132404e-01 3.07509918e-02 -6.60136104e-01
-1.02444685e+00 -4.71508384e-01 4.09816206e-01 9.66935754e-01
1.10431921e+00 1.56768531e-01 -5.18467546e-01 7.95162916e-01
-1.03388584e+00 -5.78660704e-02 5.70423186e-01 8.62173855e-01
-1.61813483e-01 -1.41898525e+00 -6.04749098e-02 5.55101097e-01
-4.59297091e-01 -2.65974075e-01 -1.04660654e+00 1.11934686e+00
2.77802318e-01 1.02322447e+00 1.16159037e-01 -1.47852704e-01
1.23998189e+00 5.41083574e-01 7.82341301e-01 -1.24502182e+00
-1.24317431e+00 -6.53986812e-01 1.23855221e+00 -5.99249065e-01
-7.75590897e-01 -1.15757608e+00 -1.86062968e+00 -3.27592969e-01
9.22655836e-02 5.04836857e-01 5.83979785e-01 1.30069602e+00
1.78854585e-01 5.61574817e-01 -3.63599837e-01 -4.39506114e-01
-1.58900581e-02 -9.49682474e-01 -1.85921654e-01 6.71698809e-01
-3.57456446e-01 -1.38765603e-01 2.87913159e-02 3.23783070e-01] | [11.184332847595215, 9.059237480163574] |
60a2b8c8-e227-491a-8744-e0a51a709763 | efficient-human-pose-estimation-with | 2012.03316 | null | https://arxiv.org/abs/2012.03316v1 | https://arxiv.org/pdf/2012.03316v1.pdf | Efficient Human Pose Estimation with Depthwise Separable Convolution and Person Centroid Guided Joint Grouping | In this paper, we propose efficient and effective methods for 2D human pose estimation. A new ResBlock is proposed based on depthwise separable convolution and is utilized instead of the original one in Hourglass network. It can be further enhanced by replacing the vanilla depthwise convolution with a mixed depthwise convolution. Based on it, we propose a bottom-up multi-person pose estimation method. A rooted tree is used to represent human pose by introducing person centroid as the root which connects to all body joints directly or hierarchically. Two branches of sub-networks are used to predict the centroids, body joints and their offsets to their parent nodes. Joints are grouped by tracing along their offsets to the closest centroids. Experimental results on the MPII human dataset and the LSP dataset show that both our single-person and multi-person pose estimation methods can achieve competitive accuracies with low computational costs. | ['Hong Wu', 'Jie Ou'] | 2020-12-06 | null | null | null | null | ['2d-human-pose-estimation'] | ['computer-vision'] | [-2.89597869e-01 2.60433167e-01 1.43573999e-01 -4.25903499e-01
-1.86677530e-01 -3.50571685e-02 3.04071069e-01 -2.21238315e-01
-5.60988188e-01 6.44748032e-01 4.12795991e-01 5.30374527e-01
9.57675800e-02 -7.85982549e-01 -5.52347779e-01 -4.02066290e-01
-3.92355621e-01 8.15215647e-01 3.79154712e-01 -2.60394841e-01
-2.99197346e-01 3.54965657e-01 -1.35682726e+00 -4.20647860e-02
5.73183060e-01 6.90168381e-01 -2.04443961e-01 5.87955773e-01
3.34100693e-01 3.39284122e-01 -6.32479727e-01 -5.73795915e-01
2.80093819e-01 -7.66871050e-02 -9.49808657e-01 3.35816512e-05
6.58029377e-01 -7.98171520e-01 -5.45429468e-01 5.16854167e-01
8.46815646e-01 3.51958483e-01 4.09181118e-01 -1.32733786e+00
-3.76470126e-02 5.09492874e-01 -8.78022134e-01 -2.18109921e-01
6.71229780e-01 -2.84523249e-01 6.47710204e-01 -8.86059284e-01
4.53515619e-01 1.72697902e+00 1.26319325e+00 5.93488812e-01
-7.37428784e-01 -5.30767500e-01 2.67925441e-01 1.61586463e-01
-1.56296182e+00 -4.80498411e-02 5.14774084e-01 -1.72916442e-01
7.81529427e-01 9.28020924e-02 1.03487492e+00 9.24841762e-01
-3.50437574e-02 9.51242805e-01 6.84615493e-01 -3.24109316e-01
-3.71307760e-01 -5.12629390e-01 1.52692139e-01 1.21371388e+00
2.13151172e-01 -1.51046768e-01 -6.24231935e-01 -2.70914048e-01
1.09389591e+00 -1.77203212e-02 -4.81000170e-02 -4.14310157e-01
-1.45005620e+00 6.32677913e-01 8.12410772e-01 -8.10867697e-02
-2.57888973e-01 5.68777204e-01 3.61436248e-01 -3.43119115e-01
3.90148193e-01 -4.89634015e-02 -4.65255052e-01 1.34400904e-01
-9.96029675e-01 8.10917020e-01 8.03810716e-01 1.04133987e+00
5.15365660e-01 -4.12306488e-01 -4.36376005e-01 9.64163065e-01
3.54114860e-01 1.21830940e-01 3.60561877e-01 -8.69067192e-01
4.18877691e-01 6.72153950e-01 3.36764663e-01 -1.11467683e+00
-1.00013244e+00 -5.47573864e-01 -9.58261847e-01 -1.90891683e-01
6.05954528e-01 -3.20832312e-01 -1.11326313e+00 1.55133092e+00
7.88477778e-01 3.47408503e-01 -5.55460513e-01 1.05735850e+00
1.07251275e+00 4.12157655e-01 8.16514641e-02 3.79126817e-01
1.89966571e+00 -1.46692884e+00 -4.98712152e-01 -1.35890171e-01
3.52466345e-01 -4.90064174e-01 6.09624445e-01 2.48685151e-01
-1.21650362e+00 -9.51409638e-01 -9.93146181e-01 -4.57235485e-01
-2.72744477e-01 6.20228887e-01 8.27952564e-01 6.60637140e-01
-9.07630265e-01 8.27483118e-01 -1.08803153e+00 -4.09779698e-01
1.51897997e-01 4.90980536e-01 -4.05774087e-01 2.81422377e-01
-1.36532664e+00 8.76213253e-01 3.90785664e-01 6.81842089e-01
-6.49787724e-01 -3.14456046e-01 -1.02849388e+00 -1.21206954e-01
3.14050406e-01 -1.37300527e+00 1.16345799e+00 -1.91034883e-01
-1.65855563e+00 6.96375728e-01 -1.77588701e-01 -4.62733686e-01
1.02298403e+00 -9.12075818e-01 -2.41724993e-04 2.24625811e-01
1.62664682e-01 1.08113480e+00 7.34090269e-01 -9.96866584e-01
-7.27706313e-01 -6.11864686e-01 -2.11945698e-02 5.60892701e-01
-2.25830764e-01 -8.19381699e-02 -1.09880114e+00 -6.77377582e-01
5.38397133e-01 -1.26395261e+00 -4.79552060e-01 9.82474461e-02
-9.58111107e-01 -4.92509037e-01 6.79232478e-01 -1.11053693e+00
1.20841932e+00 -1.56479418e+00 4.18508500e-01 4.15996104e-01
4.73310202e-01 -1.53690040e-01 1.63627788e-01 2.84192711e-01
9.37041640e-02 -3.73503715e-01 -4.28969823e-02 -9.02262926e-01
1.43813372e-01 1.46252766e-01 3.58272433e-01 6.59474909e-01
-2.55856544e-01 9.09894526e-01 -7.15006828e-01 -7.27778316e-01
2.23083615e-01 7.87901163e-01 -5.00343859e-01 6.47935495e-02
2.87057072e-01 3.85833651e-01 -1.86594456e-01 6.21708870e-01
7.13219464e-01 -6.61400631e-02 6.03049211e-02 -5.35386324e-01
8.43738578e-03 1.06613338e-01 -1.61943352e+00 1.83209193e+00
-8.63283351e-02 2.62932908e-02 -2.04884354e-03 -4.83280212e-01
8.31685185e-01 2.73207158e-01 4.45589244e-01 1.92799538e-01
1.89007178e-01 -1.21714234e-01 -2.48590380e-01 -1.59175694e-01
7.12531984e-01 1.12777606e-01 -4.39682931e-01 5.11372685e-02
4.10191864e-02 2.94965863e-01 3.03989619e-01 -2.01234743e-02
6.02673531e-01 6.19058132e-01 2.08911225e-01 -1.66276813e-01
6.48091435e-01 -2.94676900e-01 6.77340150e-01 5.53858638e-01
-3.56702268e-01 7.71186769e-01 3.13585252e-01 -7.41053820e-01
-8.41571748e-01 -1.06144845e+00 1.43872097e-01 1.26755142e+00
1.87562928e-01 -6.43615246e-01 -1.13653445e+00 -5.42400301e-01
5.71698621e-02 -1.75300971e-01 -6.98525012e-01 1.22648500e-01
-1.09566486e+00 -6.08696640e-01 8.96270394e-01 1.03350413e+00
1.12137544e+00 -8.23975503e-01 -6.89568460e-01 2.31798381e-01
-6.72202289e-01 -1.25597847e+00 -5.30934513e-01 -2.08096907e-01
-7.22863436e-01 -9.97986674e-01 -1.39098716e+00 -1.02020288e+00
6.56755447e-01 -1.35935619e-01 9.42053914e-01 1.26548082e-01
-2.44251192e-01 6.08057491e-02 -2.20954558e-03 -9.90471914e-02
3.00340295e-01 4.40558672e-01 4.16535109e-01 -2.47339204e-01
3.74312550e-02 -4.31301147e-01 -1.05749798e+00 4.42939997e-01
-3.04595307e-02 2.34715849e-01 3.77892971e-01 5.97674608e-01
4.10155237e-01 -1.04375020e-01 2.10371420e-01 -5.26418805e-01
5.32800674e-01 3.24511677e-02 -2.06136584e-01 1.53590992e-01
9.79676619e-02 -1.69766128e-01 2.94607043e-01 -1.67738423e-01
-1.03722894e+00 5.06863177e-01 -2.73665071e-01 -1.90194771e-01
-3.01908374e-01 1.52532175e-01 -9.21089575e-02 -8.88719484e-02
3.08046967e-01 -1.46822691e-01 -6.03843220e-02 -7.14276075e-01
5.92149913e-01 3.63310456e-01 1.05718791e+00 -8.12257886e-01
8.10165405e-01 5.29788613e-01 9.18157622e-02 -7.56397605e-01
-6.85791075e-01 -4.78069186e-01 -1.29711366e+00 -3.46681058e-01
1.14987302e+00 -1.14803791e+00 -1.05727792e+00 9.89515543e-01
-1.26956987e+00 -1.16780147e-01 1.48603901e-01 5.07215917e-01
-4.62331921e-01 6.63418293e-01 -1.14767385e+00 -6.68279052e-01
-6.05185747e-01 -8.04359376e-01 1.37119138e+00 2.51872867e-01
-5.06727576e-01 -6.45034432e-01 -6.19362108e-02 2.84747779e-01
-3.77181955e-02 5.30020297e-01 5.17266095e-01 -2.40050644e-01
-1.34112895e-01 -4.53783631e-01 -1.41194701e-01 1.08728083e-02
-2.54044324e-01 -2.49040946e-01 -5.59475839e-01 -4.18509692e-01
-5.82916319e-01 -1.41756728e-01 8.07425380e-01 5.11529446e-01
1.15352905e+00 -5.74495494e-02 -7.88874507e-01 7.30030835e-01
9.62125480e-01 -3.70146185e-01 5.67641258e-01 3.07291687e-01
1.08321202e+00 7.33228743e-01 2.25133061e-01 5.08271217e-01
7.78556585e-01 8.84562075e-01 1.76316813e-01 -1.88004762e-01
-1.40587091e-01 -4.45033073e-01 7.64603689e-02 5.78933895e-01
-8.28301966e-01 7.17036799e-02 -7.74349153e-01 3.53869587e-01
-1.97166610e+00 -6.98075116e-01 -4.39138144e-01 2.25169277e+00
5.83901346e-01 1.09794624e-01 8.27791572e-01 1.60219386e-01
9.84470725e-01 5.66239506e-02 -1.66432068e-01 1.44954890e-01
2.80077994e-01 1.77119955e-01 4.52582479e-01 3.14878702e-01
-1.55938053e+00 9.93892789e-01 6.74736691e+00 5.40519774e-01
-3.91778886e-01 -5.61630400e-03 2.43403196e-01 -1.27358422e-01
5.76642513e-01 -3.22118610e-01 -1.29538941e+00 2.62196124e-01
3.42227042e-01 3.39292139e-01 7.55055845e-02 8.31753314e-01
7.22473189e-02 -6.97023273e-02 -1.08599257e+00 9.90030289e-01
-8.36768653e-03 -7.21876025e-01 -1.50699273e-01 -1.13898225e-01
5.16007721e-01 -4.99274194e-01 -2.58374929e-01 2.97007769e-01
1.96692243e-01 -8.86644661e-01 7.52416432e-01 5.34220874e-01
4.85411763e-01 -1.13132620e+00 6.84466779e-01 4.04171616e-01
-1.81167400e+00 9.77712125e-02 -3.04670483e-01 -3.47569406e-01
4.99765784e-01 2.82568872e-01 -5.05928457e-01 6.21492088e-01
9.85163689e-01 4.91348475e-01 -5.57316184e-01 1.13128495e+00
-6.01442873e-01 2.01703105e-02 -6.23406529e-01 2.15244144e-02
6.99199215e-02 7.84998108e-03 4.28493500e-01 1.24859202e+00
2.50422508e-01 1.41239792e-01 5.37385702e-01 4.32938933e-01
1.16792575e-01 5.06443251e-03 -6.49096072e-02 6.91868603e-01
4.84141350e-01 1.45427191e+00 -9.42444324e-01 -5.88676929e-01
-8.96981210e-02 1.42995143e+00 2.74774492e-01 2.30208173e-01
-1.05112910e+00 -5.59533238e-01 5.56320012e-01 2.61400104e-01
1.57135531e-01 -5.13614416e-01 -1.05936239e-02 -8.30301285e-01
-9.54734907e-02 -4.97757137e-01 5.69952428e-01 -6.37493789e-01
-1.13820434e+00 4.13402975e-01 3.81590694e-01 -1.11205554e+00
-2.25580171e-01 -5.39519489e-01 -4.36441392e-01 8.72231424e-01
-8.52609336e-01 -1.43211889e+00 -5.37143052e-01 8.75237942e-01
3.25773686e-01 2.86113948e-01 8.03692400e-01 3.00743192e-01
-6.97120667e-01 8.75754476e-01 -6.09901488e-01 6.54766381e-01
5.20164430e-01 -1.21949530e+00 7.18563437e-01 6.14621818e-01
-2.65536487e-01 1.02679312e+00 5.86807787e-01 -9.88783777e-01
-7.69590080e-01 -9.54474032e-01 1.12661386e+00 -2.46833339e-01
1.94846526e-01 -3.79405349e-01 -4.68859315e-01 9.67782855e-01
7.66730011e-02 -1.46475211e-01 4.77124095e-01 4.08577710e-01
-1.22805707e-01 1.92450032e-01 -1.13892400e+00 8.27151060e-01
1.48255336e+00 -1.50635438e-02 -7.02613235e-01 3.69029671e-01
5.14834404e-01 -9.20708895e-01 -9.50407922e-01 5.05678296e-01
9.74165618e-01 -7.72805333e-01 1.54806566e+00 -4.18545067e-01
3.42609793e-01 -4.95313197e-01 2.42273957e-01 -1.07069075e+00
-6.08623803e-01 -4.63820904e-01 -3.95265251e-01 8.98532927e-01
-1.02815650e-01 -3.46380353e-01 1.14667416e+00 4.51306522e-01
-1.37135997e-01 -9.23614681e-01 -9.17339802e-01 -6.23352528e-01
-1.62974089e-01 -1.24532007e-01 5.91709912e-01 4.66006309e-01
-1.34923562e-01 3.55867833e-01 -9.61915553e-01 3.26899379e-01
9.91527319e-01 -8.09328109e-02 1.10669768e+00 -1.28905594e+00
-4.28655475e-01 -2.20896631e-01 -4.43365574e-01 -1.47422421e+00
-1.09354980e-01 -5.75626850e-01 2.00615942e-01 -1.73308694e+00
2.55067229e-01 -1.09311767e-01 1.77148089e-01 5.64683855e-01
-3.37604672e-01 5.02164006e-01 1.64349288e-01 3.13241445e-02
-4.39409703e-01 4.10856128e-01 1.32219458e+00 1.18821241e-01
-2.44211257e-01 2.86488473e-01 -1.59327626e-01 1.37087679e+00
6.29436553e-01 -1.06072269e-01 -6.53255656e-02 -2.92206734e-01
-1.16426639e-01 9.84474868e-02 7.60803699e-01 -1.58587074e+00
2.84016579e-01 4.53552246e-01 1.02384281e+00 -1.33465874e+00
7.41575718e-01 -3.83389026e-01 1.75917536e-01 8.49442959e-01
-1.14935592e-01 2.77226955e-01 -1.55768007e-01 4.31976080e-01
1.35874316e-01 7.69122690e-02 5.83350241e-01 -5.58689594e-01
-6.61838174e-01 5.23164928e-01 -1.26956075e-01 -3.93750787e-01
8.10299695e-01 -3.31316382e-01 1.08777665e-01 -3.34412128e-01
-1.47218192e+00 4.69861716e-01 2.64104512e-02 4.07968402e-01
6.44781232e-01 -1.68221414e+00 -6.12040937e-01 6.71108626e-03
-2.50347883e-01 2.26085246e-01 3.39735627e-01 8.42117965e-01
-7.08817124e-01 4.30758625e-01 -2.61076927e-01 -5.08291364e-01
-1.49074686e+00 -3.79822627e-02 5.66514492e-01 -4.08992678e-01
-8.36901009e-01 1.30850172e+00 -3.41470563e-03 -6.61694169e-01
5.06831050e-01 -3.43164533e-01 -2.86569238e-01 6.05995245e-02
4.56049830e-01 9.21166718e-01 -3.01462293e-01 -9.23433065e-01
-4.90484565e-01 8.29478264e-01 8.21229592e-02 -1.00469388e-01
1.11108363e+00 -1.65365905e-01 -2.14939073e-01 1.98942915e-01
9.97031510e-01 -1.75094828e-01 -1.20914972e+00 -7.73079023e-02
-2.41880462e-01 -3.63111764e-01 -4.64516521e-01 -5.27244627e-01
-1.06362343e+00 7.47955978e-01 5.19414783e-01 -1.67725801e-01
7.74515867e-01 -4.77753542e-02 1.26479721e+00 2.39786804e-01
6.34135127e-01 -1.28756297e+00 2.03866392e-01 5.12985706e-01
8.34681094e-01 -6.27006292e-01 2.02697530e-01 -8.29609156e-01
-2.80982733e-01 1.16754198e+00 1.03786695e+00 -3.91560733e-01
5.03365159e-01 1.14812955e-01 -2.21013516e-01 -1.75882950e-01
3.22919637e-02 -2.90629655e-01 4.54043120e-01 5.90516031e-01
6.02825940e-01 1.69721484e-01 -5.52385449e-01 7.54193246e-01
-6.90967321e-01 3.66943628e-02 1.11721918e-01 8.24304461e-01
-4.47239220e-01 -9.92152095e-01 -7.71604240e-01 1.92130253e-01
-3.08879822e-01 1.38235256e-01 -2.54997134e-01 1.08308113e+00
4.41956311e-01 5.82465351e-01 3.99733372e-02 -5.36653519e-01
6.00715697e-01 2.11608797e-01 7.60011971e-01 -5.21681845e-01
-6.26161098e-01 2.47463524e-01 2.36800283e-01 -6.56290591e-01
-3.52690130e-01 -7.45988786e-01 -1.48048985e+00 -4.42154676e-01
-2.03484312e-01 -1.88642561e-01 3.08222800e-01 8.71900976e-01
4.60986793e-02 5.17966151e-01 1.98652316e-03 -1.46831751e+00
-4.00442868e-01 -1.09665751e+00 -4.59043890e-01 2.37185717e-01
-3.43514234e-02 -1.02885509e+00 5.16965985e-02 -4.95452769e-02] | [7.127902507781982, -0.7593551278114319] |
b23bb495-a4b5-4624-825b-ae341505d063 | towards-learning-rubik-s-cube-with-n-tuple | 2301.12167 | null | https://arxiv.org/abs/2301.12167v1 | https://arxiv.org/pdf/2301.12167v1.pdf | Towards Learning Rubik's Cube with N-tuple-based Reinforcement Learning | This work describes in detail how to learn and solve the Rubik's cube game (or puzzle) in the General Board Game (GBG) learning and playing framework. We cover the cube sizes 2x2x2 and 3x3x3. We describe in detail the cube's state representation, how to transform it with twists, whole-cube rotations and color transformations and explain the use of symmetries in Rubik's cube. Next, we discuss different n-tuple representations for the cube, how we train the agents by reinforcement learning and how we improve the trained agents during evaluation by MCTS wrapping. We present results for agents that learn Rubik's cube from scratch, with and without MCTS wrapping, with and without symmetries and show that both, MCTS wrapping and symmetries, increase computational costs, but lead at the same time to much better results. We can solve the 2x2x2 cube completely, and the 3x3x3 cube in the majority of the cases for scrambled cubes up to p = 15 (QTM). We cannot yet reliably solve 3x3x3 cubes with more than 15 scrambling twists. Although our computational costs are higher with MCTS wrapping and with symmetries than without, they are still considerably lower than in the approaches of McAleer et al. (2018, 2019) and Agostinelli et al. (2019) who provide the best Rubik's cube learning agents so far. | ['Wolfgang Konen'] | 2023-01-28 | null | null | null | null | ['rubik-s-cube'] | ['graphs'] | [-2.81992763e-01 3.40865433e-01 1.65206611e-01 3.64303082e-01
-9.52715278e-01 -1.10396206e+00 6.49003267e-01 -3.35116625e-01
-3.63643706e-01 1.24642754e+00 1.54137611e-01 -4.59289700e-01
-4.03959692e-01 -9.31165516e-01 -9.91050661e-01 -9.58966851e-01
-5.44417083e-01 1.03909791e+00 3.53380144e-01 -4.97973889e-01
1.59399420e-01 1.94701836e-01 -1.56822801e+00 2.36668319e-01
4.06018525e-01 7.55472660e-01 -1.14422567e-01 1.14353645e+00
3.78313899e-01 1.03470671e+00 -6.83073580e-01 -2.71062821e-01
7.71791816e-01 -4.38570827e-01 -1.14107645e+00 -1.41569292e-02
9.90353599e-02 -4.28409696e-01 -2.62378395e-01 5.88346720e-01
3.16881001e-01 1.58514798e-01 4.73871052e-01 -1.43677378e+00
-3.20538878e-01 7.41056919e-01 -5.90368450e-01 -1.91945285e-01
3.20307612e-01 5.16889632e-01 1.02665007e+00 -1.39396697e-01
7.09329367e-01 1.20362377e+00 4.13662612e-01 7.07399607e-01
-7.55018353e-01 -8.03560257e-01 2.07604781e-01 4.62789416e-01
-1.22044456e+00 -3.15341689e-02 1.38877451e-01 -1.97169870e-01
1.44181156e+00 3.23043078e-01 1.02329063e+00 1.01062989e+00
1.57954335e-01 5.40197432e-01 1.30193293e+00 -2.51492053e-01
6.02410376e-01 -7.59509623e-01 -2.24990621e-01 7.10541785e-01
5.49483895e-01 5.53140044e-01 -3.42583239e-01 -8.94878283e-02
9.12990868e-01 -3.34094912e-01 6.93023428e-02 -6.64296865e-01
-1.23064923e+00 9.57686484e-01 5.71470082e-01 -5.73294982e-02
-2.67657042e-01 8.36808443e-01 3.93128216e-01 3.91662717e-01
-6.31000921e-02 7.59206235e-01 -6.41549587e-01 -5.68272769e-01
-5.55316746e-01 1.02473950e+00 1.03745723e+00 1.10307717e+00
5.36538124e-01 1.79822668e-01 2.94900358e-01 3.21283638e-02
-1.65902689e-01 4.07279432e-01 1.41529694e-01 -1.45647657e+00
7.59445071e-01 2.46768340e-01 3.31917197e-01 -2.88176715e-01
-7.68829942e-01 -1.65729955e-01 -5.78002036e-01 1.00679016e+00
6.71078026e-01 -6.15501285e-01 -7.22306311e-01 1.79277384e+00
3.67221206e-01 7.33117163e-02 4.70408082e-01 9.91079390e-01
5.22601128e-01 5.90142846e-01 -4.18293566e-01 4.69949245e-02
1.65583432e+00 -1.35467100e+00 -1.42796308e-01 -1.62955076e-01
9.45811033e-01 -3.71658385e-01 6.99456573e-01 8.57216716e-01
-1.39080179e+00 -1.61534157e-02 -1.18144727e+00 1.59308329e-01
-2.26679444e-01 -4.27565187e-01 1.04637110e+00 7.49773204e-01
-1.25898206e+00 6.37888193e-01 -1.05519998e+00 -9.68028158e-02
2.93218762e-01 8.01207483e-01 -6.02424383e-01 -3.64420682e-01
-9.38953996e-01 1.33882737e+00 5.68810523e-01 -1.88643426e-01
-1.06361032e+00 -3.54190618e-01 -8.87808442e-01 6.54528588e-02
9.70369399e-01 -9.13791955e-01 1.51642978e+00 -5.70800066e-01
-1.88133609e+00 3.70760918e-01 3.31586570e-01 -5.74687481e-01
5.41063249e-01 -8.05500671e-02 3.98547649e-01 -5.59917018e-02
-1.45434543e-01 8.49338651e-01 4.20848429e-01 -1.12148416e+00
-4.54914570e-01 -3.18052560e-01 7.69199193e-01 4.18081701e-01
6.37418985e-01 -4.59997654e-02 -1.00164250e-01 -4.23217833e-01
-2.73644663e-02 -1.35140574e+00 -2.74499625e-01 -5.26591539e-01
-4.05041993e-01 -1.07730865e-01 3.60705465e-01 -4.41260755e-01
4.38207865e-01 -1.56681120e+00 6.22634768e-01 1.75271742e-02
3.37217867e-01 9.43985060e-02 -6.18611217e-01 7.80416012e-01
-9.01464522e-02 4.38506231e-02 -2.18262210e-01 -3.39522183e-01
2.46750563e-01 5.93118787e-01 -5.15671261e-03 4.27992284e-01
1.35356113e-01 1.31345558e+00 -8.78800035e-01 4.35572788e-02
-4.59355302e-02 2.50623286e-01 -1.07895529e+00 -7.25863443e-04
-4.22051013e-01 5.09005725e-01 -2.01523617e-01 4.59042072e-01
5.53304434e-01 9.03027356e-02 3.64392281e-01 5.07854521e-01
-2.63242483e-01 4.65117514e-01 -1.32355762e+00 1.54475474e+00
-3.00927103e-01 2.93172836e-01 2.48866126e-01 -7.58661091e-01
4.34170574e-01 3.89721632e-01 1.74773738e-01 -8.34244072e-01
1.00382738e-01 2.10331991e-01 4.50105727e-01 -2.99631864e-01
4.48681653e-01 -3.00009370e-01 -3.50366473e-01 7.42041707e-01
-1.47455186e-01 -7.21992314e-01 6.12647653e-01 -1.80190685e-03
1.43886232e+00 5.93257129e-01 4.03714120e-01 -1.47291392e-01
1.59184173e-01 1.01902559e-01 4.49895561e-01 7.89720058e-01
2.54358023e-01 7.49328017e-01 1.07708776e+00 -6.56966150e-01
-1.23590469e+00 -9.03508425e-01 5.93948781e-01 9.95652437e-01
-2.22965386e-02 -7.38470256e-01 -8.04609060e-01 -4.44779932e-01
-1.57206744e-01 6.55129135e-01 -9.71504092e-01 -7.30224997e-02
-9.65737700e-01 -5.27168751e-01 5.66078067e-01 6.49283230e-01
4.91310209e-01 -1.24245751e+00 -1.05326021e+00 1.69725820e-01
2.22449657e-02 -7.46642053e-01 -8.27030912e-02 5.06980538e-01
-7.23647356e-01 -1.49944878e+00 -4.10740197e-01 -4.56544876e-01
4.20901090e-01 1.95630074e-01 1.20445871e+00 1.57024384e-01
-1.84296519e-01 3.12754661e-01 -6.13178432e-01 -2.06795111e-01
-1.14470057e-01 2.17423171e-01 -5.30692004e-02 -1.04853356e+00
-2.48000458e-01 -7.19656169e-01 -4.11392748e-01 4.84621003e-02
-6.79139376e-01 2.76430994e-01 4.31752264e-01 8.23888361e-01
1.39980733e-01 1.38698488e-01 -1.06391340e-01 -7.11495697e-01
4.11219478e-01 -2.12502435e-01 -1.14939868e+00 -2.04995602e-01
-9.62188095e-02 4.02838111e-01 6.73880816e-01 -2.51377076e-01
-5.00790179e-01 -1.16247497e-01 1.77185833e-01 -1.44619331e-01
1.18008949e-01 1.06194526e-01 1.04528978e-01 -2.42040038e-01
6.04462743e-01 -6.39805943e-02 -1.48153514e-01 -2.86451340e-01
6.00680649e-01 -9.87244472e-02 2.72137731e-01 -1.10125196e+00
1.00003004e+00 3.16490769e-01 1.99727833e-01 -3.30790654e-02
-4.19652134e-01 2.85340548e-01 -4.46693361e-01 2.74770886e-01
8.88035715e-01 -7.61922419e-01 -1.69969738e+00 4.08806056e-01
-1.14350462e+00 -8.85903060e-01 -4.50530708e-01 4.39106911e-01
-1.14248085e+00 9.78847221e-02 -6.61242187e-01 -8.31376135e-01
-1.36976972e-01 -1.28449607e+00 1.07134652e+00 1.09177525e-03
-1.57535106e-01 -4.51940715e-01 3.53437573e-01 4.43460912e-01
4.15745348e-01 4.45614666e-01 1.01220441e+00 -3.60019714e-01
-9.49932754e-01 2.16598764e-01 -2.19223201e-02 -8.53150338e-02
-6.18026108e-02 -3.09012741e-01 -4.79565322e-01 -6.81568384e-01
-2.45576039e-01 -6.62379503e-01 6.37179315e-01 1.56213582e-01
7.40288496e-01 -6.98838949e-01 1.50952965e-01 5.93617439e-01
1.20672524e+00 4.12284642e-01 8.43603134e-01 9.21349525e-01
5.39087772e-01 5.04688263e-01 3.63121808e-01 5.05527020e-01
7.58134902e-01 6.10015333e-01 1.21560752e+00 3.79364610e-01
1.13136932e-01 1.03297107e-01 5.36448538e-01 4.06213224e-01
-6.87388539e-01 -2.55161613e-01 -8.24235380e-01 2.90131539e-01
-1.89091921e+00 -8.34010780e-01 1.96864516e-01 1.94515169e+00
6.93674564e-01 1.71204597e-01 3.86303604e-01 3.11461627e-01
1.91354543e-01 1.08808041e-01 -4.98290956e-01 -7.99107313e-01
1.05278954e-01 8.01233411e-01 7.01908171e-01 8.26154351e-01
-7.38820136e-01 1.13725245e+00 6.42167854e+00 4.73430514e-01
-7.83509016e-01 -4.05582078e-02 2.18437195e-01 -6.74861968e-01
3.21698859e-02 3.07362616e-01 -4.34720159e-01 1.84764683e-01
7.50076771e-01 2.31353760e-01 1.37995207e+00 5.65329850e-01
-1.71211436e-01 -3.28508556e-01 -1.18692827e+00 8.77072155e-01
5.05485013e-03 -1.27984250e+00 -2.94479072e-01 2.41458774e-01
8.98353040e-01 1.40976310e-01 1.01976827e-01 7.62499511e-01
9.79999185e-01 -1.50915134e+00 9.41306531e-01 -1.44706860e-01
4.30114657e-01 -9.47079062e-01 6.36348963e-01 5.21813333e-01
-9.94110703e-01 -2.11735979e-01 -2.54111737e-01 -6.85398161e-01
-1.36349007e-01 -2.23454773e-01 -6.53710127e-01 5.97339332e-01
5.55528402e-01 1.07817531e-01 -2.09701285e-01 8.14109683e-01
-6.54487848e-01 2.99622178e-01 -5.15784442e-01 -1.08967490e-01
7.52258778e-01 -5.56577444e-01 2.93190181e-01 5.30078828e-01
4.48002785e-01 5.37113845e-01 -9.46621224e-02 6.10625029e-01
7.74495536e-03 -4.47545856e-01 -2.29653165e-01 1.19687974e-01
3.31063092e-01 9.20599759e-01 -7.31184542e-01 -2.99281269e-01
-1.39455467e-01 6.26448333e-01 4.04742539e-01 2.84672886e-01
-8.18655789e-01 -1.10214822e-01 8.51300061e-01 6.61793575e-02
8.51680815e-01 -5.73399127e-01 -5.01354858e-02 -1.09045017e+00
2.19990108e-02 -1.39278877e+00 2.48450294e-01 -9.42679524e-01
-7.56080866e-01 6.13723636e-01 2.64672607e-01 -1.01091146e+00
-7.91416585e-01 -9.51617360e-01 -6.33942664e-01 4.36285257e-01
-1.42958951e+00 -1.27243042e+00 -3.18550020e-02 6.50227726e-01
1.80171400e-01 -8.82233381e-02 1.01007628e+00 -3.85923475e-01
-1.71401948e-01 1.87613904e-01 2.42384851e-01 -2.88538896e-02
9.64964032e-02 -1.29458642e+00 8.36436093e-01 3.65627229e-01
3.61420736e-02 4.43336040e-01 7.32216597e-01 -3.45156729e-01
-1.77559304e+00 -7.08700538e-01 1.95058733e-01 -4.86997694e-01
5.64483225e-01 -6.25435591e-01 -5.71856022e-01 1.06932700e+00
3.88748527e-01 -3.31678875e-02 1.82092980e-01 1.00949332e-01
-8.86781335e-01 2.51043946e-01 -1.05428529e+00 7.77660131e-01
1.24778473e+00 -1.09388709e-01 -5.95465064e-01 4.88097548e-01
5.90474069e-01 -9.49605048e-01 -6.44807458e-01 1.42606601e-01
5.24256527e-01 -1.27221501e+00 8.14550459e-01 -7.66631782e-01
4.01683867e-01 -4.17741239e-01 -1.63048252e-01 -1.43441224e+00
-3.11122388e-01 -7.64216065e-01 -2.34438889e-02 3.56257468e-01
1.07992798e-01 -9.14392948e-01 1.08873522e+00 1.62716597e-01
-9.01299936e-04 -7.75261581e-01 -1.37223864e+00 -8.86509657e-01
8.69907796e-01 -3.75233620e-01 9.16983545e-01 5.67863405e-01
2.75101990e-01 5.72795011e-02 -4.44798619e-01 -2.02573510e-03
3.91370654e-01 2.92507082e-01 1.07976937e+00 -7.63932467e-01
-8.03207397e-01 -5.12828290e-01 -4.03321505e-01 -7.33643889e-01
3.11293509e-02 -7.47957885e-01 -2.40076572e-01 -1.50005984e+00
2.86423981e-01 -3.06267798e-01 1.84953868e-01 7.64553726e-01
1.70919955e-01 3.01600814e-01 8.73422563e-01 -1.59689337e-01
-8.46719742e-01 4.85966891e-01 1.49544621e+00 7.74711883e-03
1.27208114e-01 -2.63085485e-01 -8.70065391e-01 4.82145935e-01
9.32523191e-01 -4.34652209e-01 -3.37027550e-01 -6.99301481e-01
6.96278453e-01 2.46086463e-01 3.40522349e-01 -1.13777983e+00
1.48849472e-01 -2.99647123e-01 1.99870452e-01 -3.46019953e-01
7.26250708e-01 -6.10810697e-01 3.70723933e-01 8.95722806e-01
3.59697640e-02 4.18205321e-01 3.69091094e-01 9.23296288e-02
1.94132730e-01 -2.16272414e-01 5.15680611e-01 -4.79494929e-01
-1.27580449e-01 8.79675001e-02 -5.91872454e-01 1.34904623e-01
1.12866735e+00 -2.18463585e-01 -5.51415503e-01 -7.16325998e-01
-7.64208794e-01 5.00334680e-01 5.89158177e-01 3.07192475e-01
2.25808814e-01 -1.28275526e+00 -7.91819096e-01 1.68005586e-01
-3.22986007e-01 4.83144045e-01 1.51199415e-01 5.76148570e-01
-8.35487247e-01 5.44891536e-01 -6.28326297e-01 -7.21791834e-02
-1.17545426e+00 6.52074873e-01 3.67785990e-01 -7.65498161e-01
-6.06138289e-01 7.61045516e-01 3.13725978e-01 -5.96064448e-01
-4.29483205e-02 -7.21108854e-01 6.72728866e-02 -2.81412184e-01
3.03337812e-01 4.25015122e-01 8.15826058e-02 -3.29207540e-01
-3.72898310e-01 9.01259482e-01 -2.26659030e-02 -2.75398999e-01
1.51259172e+00 5.32308042e-01 -4.78116870e-02 -1.80576593e-01
7.72006631e-01 -9.33677033e-02 -1.34305489e+00 1.99704170e-01
-2.88049757e-01 -1.53184414e-01 -6.67374730e-01 -7.60765731e-01
-1.08976722e+00 7.54863441e-01 -2.55785752e-02 7.49580041e-02
7.39241064e-01 4.31002602e-02 5.97037554e-01 7.24963009e-01
9.61962700e-01 -6.80836260e-01 1.65289491e-01 9.61346507e-01
9.96889591e-01 -7.28234172e-01 1.93320557e-01 -1.63624555e-01
-7.30798423e-01 1.06347001e+00 5.69960535e-01 -6.57870412e-01
-1.26955852e-01 5.90062499e-01 -2.16127306e-01 -8.85589197e-02
-1.23643446e+00 -3.87393624e-01 -3.70071948e-01 9.60767686e-01
-2.28504404e-01 2.48565182e-01 -5.82347251e-02 6.52730167e-01
-1.09546924e+00 -3.01884919e-01 7.99593270e-01 1.23717558e+00
-4.01457995e-01 -1.40613043e+00 -8.35076332e-01 9.58153978e-02
-6.59762695e-02 3.81733552e-02 -4.52342361e-01 1.26411676e+00
2.90514439e-01 7.88796902e-01 1.46435738e-01 -3.13451052e-01
3.25484723e-01 -1.04737408e-01 1.17570245e+00 -5.56886435e-01
-8.15945804e-01 -2.10592255e-01 1.78315371e-01 -7.76672781e-01
-1.78508610e-01 -6.44019008e-01 -1.47818053e+00 -8.34340334e-01
-8.25568661e-02 2.68497199e-01 4.92816448e-01 9.43189263e-01
2.21890211e-01 5.51936090e-01 2.16152117e-01 -1.30861807e+00
-6.16358340e-01 -7.50317991e-01 -5.23706734e-01 -2.23184284e-02
3.26738566e-01 -7.63110876e-01 -2.00843394e-01 -4.46921170e-01] | [3.7012155055999756, 1.5364125967025757] |
48fa85eb-03ed-4f65-a539-f19de4811d7c | monitoring-and-detection-of-low-current-high | 2210.16981 | null | https://arxiv.org/abs/2210.16981v1 | https://arxiv.org/pdf/2210.16981v1.pdf | Monitoring and Detection of Low-current High-Impedance Faults in Distribution Networks | Faults in electricity distribution networks have the potential to ignite fires, cause electrocution, and damage the system itself. High current Low Impedance Faults (LIF) are typically detected and mitigated via over-current, distance, directional relays, fuses, etc. In contrast, while High Impedance Faults (HIF) are equally hazardous, they are much more challenging to detect due to the fault current being much lower than load currents and their time-varying and nonlinear behaviour. Moreover, New Zealand distribution networks are extensive and largely unmonitored beyond the substation, and suitable HIF detection schemes are still an ongoing research challenge. To date, we have built a physical test facility for power system fault analysis and developing and evaluating our sensing and fault detection system. We have simulated LIF and HIF with different fault surface materials and load-switching events. From the data collected, we have characterized the unique fault behaviour for both LIF and HIF in 400V networks and trained a Deep Learning classifier to recognize the type of fault present from its unique signature. We have developed an outdoor pole mountable sensing system and have installed this in Wellington Electricity's network for ongoing data collection and evaluation. This paper will describe the test facility and our experience developing and implementing the sensing system. The widest range of HIF phenomena observed was in the fault experiments involving the tree branch. For brevity, therefore, this paper reports on the results of just these tree-branch experiments. HIF faults on other surface materials will be reported elsewhere. Finally, we will detail the pole-mountable sensing system installed in Wellington Electricity's network and the outcomes thus far. | ['Joseph Bailey', 'Ramesh Rayudu', 'Fiona J. Stevens McFadden', 'Anwarul Islam Sifat'] | 2022-10-30 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [ 1.42595127e-01 -1.90859184e-01 4.23898935e-01 -1.07080735e-01
-5.02774417e-01 -8.17408502e-01 5.28456986e-01 -1.78473815e-01
4.64078307e-01 8.20294082e-01 5.10994568e-02 -6.22491062e-01
-8.02508652e-01 -8.58354330e-01 -2.68487662e-01 -8.12159121e-01
-5.90960741e-01 4.12400514e-01 2.01735660e-01 -3.35862011e-01
2.80030191e-01 1.14516485e+00 -1.37633646e+00 6.36528656e-02
7.30907738e-01 7.80507922e-01 6.29040822e-02 4.78531539e-01
9.13858950e-01 8.34785640e-01 -1.35494554e+00 9.13895190e-01
2.02971905e-01 -2.71680355e-01 -9.84379113e-01 -2.91036796e-02
-1.54386356e-01 -7.74755538e-01 -4.71188843e-01 3.95930409e-01
6.82053983e-01 1.02055863e-01 9.26582515e-01 -1.50549591e+00
6.94464296e-02 4.41496044e-01 -2.05044627e-01 7.52558231e-01
6.30082905e-01 2.08346650e-01 4.57220674e-01 -4.56678450e-01
9.32088047e-02 6.61343098e-01 8.90736461e-01 -6.99177384e-02
-8.55393171e-01 -5.14912188e-01 -5.25031328e-01 1.42300740e-01
-1.26597774e+00 -3.55296075e-01 7.80283511e-01 -5.04118621e-01
1.91135967e+00 3.04771513e-01 7.07956612e-01 6.85049772e-01
5.00903606e-01 1.71539918e-01 9.55225110e-01 -2.86844194e-01
4.00889426e-01 -2.72536427e-01 -3.26843828e-01 2.83828676e-01
2.16193736e-01 3.96783054e-02 2.16459528e-01 -1.16379701e-01
7.01271892e-01 -2.85587221e-01 -9.61117268e-01 5.23331463e-01
-2.90662050e-01 4.43399340e-01 3.80370408e-01 1.01870096e+00
-2.93085277e-01 5.05778193e-02 1.90998957e-01 4.78958458e-01
4.18706000e-01 6.16397321e-01 -5.29150724e-01 -5.53174078e-01
-1.00595641e+00 -3.79646905e-02 1.00331044e+00 2.75096297e-01
1.64808944e-01 4.97681588e-01 3.98803025e-01 8.81671250e-01
1.70941874e-01 4.68728751e-01 2.85153419e-01 -4.90668148e-01
2.34685436e-01 2.79451787e-01 2.19586402e-01 -7.45675325e-01
-7.29050875e-01 -6.71991855e-02 -4.50430274e-01 4.87302512e-01
-9.17820260e-02 -9.25121605e-01 -9.60316718e-01 9.78878140e-01
-1.89551696e-01 1.24453627e-01 1.24990325e-02 8.18270624e-01
3.52902651e-01 8.43338370e-01 -4.43324476e-01 1.00009032e-01
8.69574845e-01 -2.76763085e-02 -5.66899538e-01 -6.93620816e-02
8.14430714e-01 -4.80131000e-01 4.85523880e-01 3.62661093e-01
-8.62642407e-01 6.72518313e-02 -1.53720665e+00 6.91755354e-01
-3.22410315e-01 -6.90851957e-02 2.65070230e-01 8.37701857e-01
-1.21602428e+00 9.17684317e-01 -8.41837227e-01 -8.20463777e-01
3.14163685e-01 1.87985137e-01 -2.22940311e-01 4.60107289e-02
-1.70730627e+00 1.48387742e+00 -2.52250135e-01 5.98179042e-01
-1.13718379e+00 -6.07508898e-01 -6.27132952e-01 5.30710965e-02
-9.00083780e-02 2.19546601e-01 1.14643240e+00 -1.98560342e-01
-1.16106486e+00 1.06579691e-01 4.36816812e-01 -3.11557412e-01
2.66105235e-01 2.39093918e-02 -1.07990599e+00 5.87932885e-01
1.69445068e-01 -2.85330415e-01 2.48578742e-01 -1.05448973e+00
-5.62196672e-01 1.45368233e-01 -1.39837395e-02 5.48201650e-02
-2.68410016e-02 5.29685058e-03 8.86897087e-01 -2.38919601e-01
1.34886518e-01 -4.67632473e-01 3.01796615e-01 -5.78877449e-01
-3.50027621e-01 -4.43314821e-01 1.57788229e+00 -7.54321992e-01
6.70180738e-01 -2.02712035e+00 -3.54700089e-01 6.53843880e-01
-3.84143353e-01 1.54514104e-01 4.78317618e-01 1.15681231e+00
-3.21612954e-01 2.22388774e-01 -6.20970726e-01 7.22745717e-01
-3.94809172e-02 5.59000909e-01 -3.25930446e-01 9.97528732e-01
4.51281071e-01 3.55281651e-01 -7.18072355e-01 5.35243690e-01
5.83188951e-01 2.92033643e-01 2.15176165e-01 3.56453210e-02
4.25947517e-01 -1.83952171e-02 -1.24642313e-01 8.33559990e-01
6.79587960e-01 4.43968236e-01 -1.15678191e-01 -2.47949064e-01
-2.83519864e-01 4.12464619e-01 -1.10846353e+00 1.04192412e+00
-4.24946666e-01 8.58076692e-01 4.21670914e-01 -1.51398098e+00
1.18894827e+00 8.33546877e-01 7.54079401e-01 -6.60452008e-01
-4.79467101e-02 5.60845256e-01 -1.49284065e-01 -1.06902337e+00
5.54963909e-02 -4.05189157e-01 6.05029054e-02 4.90103990e-01
-1.60788625e-01 -6.31814420e-01 -7.62435645e-02 7.60276243e-02
1.73632073e+00 -2.09531665e-01 -6.15108967e-01 -9.21901226e-01
1.07123494e-01 -5.59422746e-03 4.41050828e-01 3.35722059e-01
-9.14565325e-02 4.33967680e-01 5.47550201e-01 -2.07200125e-01
-6.77532494e-01 -9.26452756e-01 -8.53572309e-01 1.66803040e-02
4.39561337e-01 2.39541590e-01 -4.63917524e-01 -4.34535980e-01
3.60349149e-01 9.48096514e-01 -2.11636871e-01 -5.05795360e-01
-2.47223318e-01 -7.72853315e-01 1.06528842e+00 7.86018848e-01
6.34999514e-01 -1.13265240e+00 -7.89957941e-01 4.36969489e-01
5.89551814e-02 -5.39733589e-01 4.03905451e-01 7.02286482e-01
-4.81009781e-01 -1.28166723e+00 -4.01870251e-01 -7.35477328e-01
4.63174671e-01 -6.68208823e-02 8.45354140e-01 3.10888708e-01
-6.96933806e-01 2.89182365e-01 -3.08059037e-01 -3.38589638e-01
-6.24369830e-02 -2.95293659e-01 7.70764649e-02 -6.08474255e-01
1.09380245e-01 -7.52934456e-01 -3.55523348e-01 4.81066167e-01
-9.40412819e-01 -7.78598785e-01 2.72064567e-01 4.73738283e-01
-5.05524874e-01 1.16522312e+00 1.21712589e+00 -5.75550199e-01
7.69838214e-01 -9.43532169e-01 -1.84118301e-01 -1.11521095e-01
-4.41943109e-01 -5.59603572e-01 3.18205774e-01 2.15407208e-01
-1.12859786e+00 -4.35201496e-01 -4.19482499e-01 1.36290073e-01
-8.21640193e-01 5.67638278e-01 -3.06368470e-01 -3.13725591e-01
6.87515914e-01 -6.11154497e-01 -9.98568758e-02 -3.22527498e-01
-5.25147140e-01 1.11378574e+00 5.70344090e-01 -4.56411123e-01
9.76506889e-01 4.17457223e-01 -2.38944646e-02 -1.29608345e+00
2.48624869e-02 -1.61389306e-01 -2.15672374e-01 -6.25302553e-01
5.04960716e-01 -8.01028669e-01 -2.54288405e-01 9.72539783e-01
-6.45792305e-01 -7.56042600e-01 -1.85315445e-01 4.65773314e-01
7.48052970e-02 1.69354901e-02 -9.52140450e-01 -7.89103329e-01
-9.59137902e-02 -1.06091702e+00 7.73443937e-01 1.74073249e-01
-3.36553812e-01 -1.34858847e+00 -6.58012629e-02 -1.07762497e-02
8.34050238e-01 7.41780519e-01 9.60828722e-01 -1.90473109e-01
-5.60839251e-02 -3.74250978e-01 1.06735140e-01 5.36466837e-01
5.15770912e-01 2.40302041e-01 -1.13827372e+00 -6.23201251e-01
3.20734411e-01 -1.67096719e-01 4.42678362e-01 1.60610318e-01
5.75759590e-01 1.70448169e-01 -6.31647825e-01 3.13384831e-01
1.67893541e+00 7.98599899e-01 7.28720605e-01 5.40680707e-01
2.90955663e-01 3.59799951e-01 4.10477430e-01 4.27950054e-01
-1.81318134e-01 1.21246219e-01 4.78398383e-01 -5.08219361e-01
3.27587157e-01 2.81955838e-01 3.27706963e-01 1.51359379e-01
1.64576452e-02 -6.10430419e-01 -1.07310283e+00 8.09274912e-01
-1.08832145e+00 -8.65263224e-01 -6.08003736e-01 1.51401699e+00
2.71814346e-01 2.71522701e-01 -3.37352782e-01 1.02757192e+00
7.46986270e-01 -1.68613672e-01 -3.07518691e-01 -7.58580863e-01
-2.46761143e-01 3.28475386e-01 6.43034756e-01 5.29793739e-01
-7.61722147e-01 -7.22905174e-02 6.45221901e+00 3.15568656e-01
-1.21305954e+00 -2.08152637e-01 1.86204001e-01 -1.19441055e-01
-4.41053540e-01 7.70736411e-02 -1.25167789e-02 3.16189170e-01
1.01547265e+00 7.47643933e-02 2.43730724e-01 9.11153704e-02
2.16374516e-01 -6.80800915e-01 -9.74739730e-01 2.76466101e-01
-1.71448454e-01 -7.52111852e-01 -5.89145243e-01 -3.97712551e-02
5.21260440e-01 3.63430321e-01 -7.53750861e-01 -1.06505498e-01
9.08980221e-02 -9.33993101e-01 6.03416085e-01 3.64591748e-01
8.80324185e-01 -9.97813165e-01 1.03387105e+00 6.77130520e-02
-1.05061436e+00 -5.60262024e-01 -1.10631824e-01 -5.26956320e-01
6.01079881e-01 1.14042592e+00 -6.75514221e-01 8.48648608e-01
1.15592360e+00 6.44006729e-01 -2.19838023e-01 8.02877665e-01
-3.24330658e-01 1.07044458e+00 -7.92076290e-01 2.45997593e-01
3.83489370e-01 -1.16225548e-01 4.01079029e-01 5.84893942e-01
5.02557814e-01 1.22905008e-01 -1.13657124e-01 5.63376725e-01
4.22378808e-01 -7.47027695e-01 -9.73156810e-01 1.69346660e-01
8.08249712e-01 1.14686692e+00 -8.28697026e-01 -1.12174094e-01
-2.86408454e-01 5.46499848e-01 -5.39036751e-01 5.14404118e-01
-7.67983198e-01 -1.11240184e+00 7.24830985e-01 3.34407002e-01
-1.94903880e-01 2.39470646e-01 -2.23103866e-01 -4.42273170e-01
1.23501174e-01 -1.94496766e-01 1.15521409e-01 -1.03626120e+00
-1.41887987e+00 2.84624904e-01 3.16123724e-01 -6.84715152e-01
-1.95670009e-01 -3.54354262e-01 -1.26169395e+00 1.12792110e+00
-1.22294164e+00 -3.99261296e-01 -2.78000176e-01 4.47075993e-01
4.04534817e-01 -4.63358462e-02 6.42548442e-01 5.38407087e-01
-7.75229692e-01 6.76554218e-02 4.05575216e-01 6.95050061e-02
1.34207055e-01 -1.15210688e+00 1.93799779e-01 8.74864638e-01
-5.27837276e-01 -1.99371412e-01 4.94587541e-01 -9.83406246e-01
-1.52223110e+00 -8.22022974e-01 4.54503208e-01 6.19916990e-02
7.04782724e-01 -3.01615983e-01 -1.10683453e+00 7.07701087e-01
6.92570090e-01 -3.87613386e-01 1.40511259e-01 -3.85229707e-01
4.81560022e-01 -2.04442181e-02 -1.87776279e+00 9.42374095e-02
6.15549624e-01 -8.81675005e-01 -6.15080655e-01 6.45381749e-01
-1.31988660e-01 -1.60069585e-01 -1.21602869e+00 7.90259123e-01
1.69102535e-01 -8.52480233e-01 4.54708099e-01 3.51239175e-01
7.75429234e-02 -3.46134126e-01 2.49003008e-01 -1.64653313e+00
-4.60596561e-01 -5.33492327e-01 5.29015124e-01 1.64947486e+00
4.34032202e-01 -1.27962601e+00 3.22784662e-01 3.47021878e-01
-6.81381524e-01 -7.84696281e-01 -1.00953436e+00 -5.30819416e-01
1.69503260e-02 -3.06283146e-01 8.74422669e-01 1.05369461e+00
5.90766191e-01 -1.69685513e-01 3.99584979e-01 8.67225826e-01
2.47903824e-01 -2.10783094e-01 -2.37705097e-01 -1.20765758e+00
1.51160538e-01 -1.36862636e-01 -6.38032913e-01 8.31122845e-02
-7.07736239e-02 -6.55073881e-01 4.66037512e-01 -2.14638567e+00
-7.78193474e-01 -6.28784776e-01 1.51060922e-02 8.22975814e-01
3.44485760e-01 1.53078601e-01 -5.29889286e-01 1.07824944e-01
6.85538411e-01 1.85497865e-01 5.19930184e-01 -2.63361305e-01
2.74367541e-01 -3.25535566e-01 1.34992786e-03 5.19216597e-01
1.34065044e+00 -1.13843292e-01 -5.69032133e-01 -4.87821341e-01
7.16517167e-03 2.85556883e-01 2.81531632e-01 -1.56468463e+00
7.11584538e-02 1.71879679e-01 6.77663684e-01 -4.56483513e-01
-1.13158226e-01 -1.07374132e+00 6.08432055e-01 6.85858190e-01
2.51743317e-01 1.32976860e-01 4.48254377e-01 -1.92770306e-02
4.69263224e-03 -4.58846688e-01 6.24516666e-01 5.18658571e-02
-7.47351348e-01 -2.79912204e-01 -1.35523903e+00 -2.68465519e-01
1.44117260e+00 -1.26966879e-01 -7.10484743e-01 -2.59680271e-01
-3.21758866e-01 3.56194347e-01 5.93477249e-01 3.21531087e-01
5.88644147e-01 -1.01703668e+00 -5.14826119e-01 7.27089226e-01
-4.06630248e-01 -1.04501538e-01 2.02325195e-01 7.36683726e-01
-8.41723561e-01 9.15249214e-02 -3.18095416e-01 -6.06968522e-01
-6.39330864e-01 -1.84144303e-01 9.11876857e-01 6.00316048e-01
-9.83912766e-01 7.01313972e-01 -5.86799085e-01 -1.31010935e-01
-1.96486220e-01 -3.26116085e-01 -2.22218648e-01 1.32473364e-01
1.46466553e-01 8.15705717e-01 9.70685661e-01 -3.70738238e-01
-4.62683141e-01 3.95337909e-01 5.78548551e-01 7.11628236e-03
1.41530168e+00 9.55727175e-02 -2.33743906e-01 4.69230175e-01
9.76232409e-01 -5.93010664e-01 -1.20642364e+00 7.64997184e-01
-1.40950501e-01 -2.07046837e-01 3.01286697e-01 -1.05851114e+00
-1.35475349e+00 4.85578448e-01 4.22986984e-01 1.08374524e+00
1.36308610e+00 1.60521048e-03 5.12581408e-01 1.06684662e-01
4.68634337e-01 -1.12847495e+00 -5.52766681e-01 2.21722528e-01
9.33060408e-01 -5.22742905e-02 -1.16772316e-01 -1.12935297e-01
1.03011116e-01 1.15202129e+00 4.76309240e-01 -3.38175178e-01
7.45107651e-01 1.19040942e+00 9.58329290e-02 -6.66944444e-01
-6.53812706e-01 4.18842405e-01 -6.67962074e-01 8.67626488e-01
3.40426683e-01 1.10711962e-01 1.92438774e-02 -1.54121086e-01
-1.79364048e-02 -2.19048765e-02 9.72166836e-01 1.63620925e+00
-5.49786270e-01 -6.23705924e-01 -8.14599574e-01 1.14922988e+00
-4.51323211e-01 4.29297239e-01 -2.24417567e-01 7.24346995e-01
1.96067728e-02 1.53036642e+00 2.66892791e-01 -3.06197375e-01
7.83006907e-01 3.59735303e-02 3.36736023e-01 -2.26444796e-01
-6.44753873e-01 -2.03844681e-01 5.21411359e-01 -4.52881455e-01
-1.19023770e-02 -8.80306542e-01 -1.72361863e+00 -2.84429282e-01
-8.03610265e-01 5.64625740e-01 9.13547397e-01 1.25095809e+00
-6.94660991e-02 1.05230856e+00 1.01100516e+00 -1.16714835e+00
-4.74088252e-01 -1.36723924e+00 -1.55289352e+00 6.96916059e-02
2.17618808e-01 -9.41928685e-01 -1.11864114e+00 -6.50171041e-01] | [6.331514358520508, 2.4845423698425293] |
ebf3b142-656e-472d-9e46-f75162986716 | train-diagnose-and-fix-interpretable-approach | 1711.08502 | null | http://arxiv.org/abs/1711.08502v1 | http://arxiv.org/pdf/1711.08502v1.pdf | Train, Diagnose and Fix: Interpretable Approach for Fine-grained Action Recognition | Despite the growing discriminative capabilities of modern deep learning
methods for recognition tasks, the inner workings of the state-of-art models
still remain mostly black-boxes. In this paper, we propose a systematic
interpretation of model parameters and hidden representations of Residual
Temporal Convolutional Networks (Res-TCN) for action recognition in time-series
data. We also propose a Feature Map Decoder as part of the interpretation
analysis, which outputs a representation of model's hidden variables in the
same domain as the input. Such analysis empowers us to expose model's
characteristic learning patterns in an interpretable way. For example, through
the diagnosis analysis, we discovered that our model has learned to achieve
view-point invariance by implicitly learning to perform rotational
normalization of the input to a more discriminative view. Based on the findings
from the model interpretation analysis, we propose a targeted refinement
technique, which can generalize to various other recognition models. The
proposed work introduces a three-stage paradigm for model learning: training,
interpretable diagnosis and targeted refinement. We validate our approach on
skeleton based 3D human action recognition benchmark of NTU RGB+D. We show that
the proposed workflow is an effective model learning strategy and the resulting
Multi-stream Residual Temporal Convolutional Network (MS-Res-TCN) achieves the
state-of-the-art performance on NTU RGB+D. | ['Jingxuan Hou', 'Austin Reiter', 'Tae Soo Kim'] | 2017-11-22 | null | null | null | null | ['3d-human-action-recognition', 'fine-grained-action-recognition'] | ['computer-vision', 'computer-vision'] | [ 7.57867038e-01 2.61056125e-01 -3.87741446e-01 -4.18780237e-01
-3.30348641e-01 -1.61158994e-01 6.04868233e-01 -4.04440433e-01
4.60693426e-02 1.25263080e-01 2.66932845e-01 -1.30907193e-01
-4.16167200e-01 -3.31713319e-01 -7.07050025e-01 -7.91040659e-01
1.40488455e-02 6.84153318e-01 2.09512755e-01 -8.75963867e-02
7.89888427e-02 8.25385630e-01 -1.67991567e+00 7.38642335e-01
9.29320082e-02 1.38992214e+00 -2.85205185e-01 7.13420153e-01
1.48237616e-01 1.16797996e+00 -4.29184675e-01 6.83772862e-02
3.26047003e-01 -4.36651319e-01 -1.07642066e+00 4.94006097e-01
5.41664541e-01 -4.04771358e-01 -4.99168038e-01 4.47068393e-01
2.40765363e-01 2.95319334e-02 7.15784311e-01 -1.32438445e+00
-5.15101254e-01 1.97515160e-01 -2.49168590e-01 2.27505192e-01
2.15677559e-01 3.69438708e-01 7.90459394e-01 -7.83060849e-01
7.16836572e-01 1.28273571e+00 5.85163534e-01 7.81853437e-01
-1.18954015e+00 -2.59000868e-01 3.26396286e-01 6.03390396e-01
-8.57288957e-01 -3.62759292e-01 6.70704126e-01 -4.37177241e-01
1.32603729e+00 3.43797833e-01 1.07658029e+00 1.41362321e+00
4.06148821e-01 9.60089922e-01 1.00117862e+00 -4.29167300e-01
2.62107372e-01 -5.23775518e-01 2.26231411e-01 7.72154272e-01
-2.26169541e-01 2.83753783e-01 -8.78860712e-01 3.92052174e-01
1.02867186e+00 2.53802568e-01 -1.39841273e-01 -6.92386389e-01
-1.20080316e+00 4.43432152e-01 3.60623747e-01 1.60182461e-01
-3.94771427e-01 5.91148734e-01 4.12956893e-01 1.82794005e-01
3.67813885e-01 6.57809004e-02 -7.59004354e-01 -1.49171531e-01
-6.03764772e-01 -1.83809269e-02 3.40490818e-01 7.60191381e-01
5.14228880e-01 2.51331955e-01 -3.31149310e-01 4.06895399e-01
2.90822506e-01 4.01164860e-01 7.08551347e-01 -9.42926526e-01
2.55023360e-01 1.06729198e+00 -3.12610775e-01 -6.42883062e-01
-6.30823493e-01 -5.82636833e-01 -1.00729847e+00 6.02636635e-01
3.01285237e-01 4.73651320e-01 -1.33764255e+00 1.41830003e+00
2.18729645e-01 3.83777618e-01 1.75166979e-01 9.34859812e-01
5.37417233e-01 6.38202429e-02 -8.67560655e-02 1.56421840e-01
1.22341049e+00 -1.02037859e+00 -5.99702775e-01 -7.44013786e-02
8.35037291e-01 -2.45439127e-01 8.20969999e-01 6.02948189e-01
-7.60744333e-01 -8.44064832e-01 -1.05775964e+00 -1.54807925e-01
-2.72642463e-01 3.74285758e-01 7.54249334e-01 3.34037095e-01
-9.03664649e-01 8.95962000e-01 -1.38406897e+00 -5.29827595e-01
7.11994112e-01 6.31059110e-01 -7.95868039e-01 1.55958086e-01
-7.55842686e-01 9.78490233e-01 3.56246114e-01 4.19991791e-01
-1.18967748e+00 -4.63590652e-01 -7.93848932e-01 -1.70341477e-01
3.52006227e-01 -9.60767150e-01 1.27442098e+00 -1.31258595e+00
-1.60576499e+00 8.98593783e-01 -1.65459156e-01 -7.27601767e-01
5.97175241e-01 -2.73351759e-01 -2.83652574e-01 3.05126101e-01
-2.83603251e-01 5.08471727e-01 1.17110801e+00 -1.01765835e+00
-4.98001158e-01 -6.22647345e-01 7.77156204e-02 -4.44726422e-02
-1.78477436e-01 -2.04410955e-01 -3.75509351e-01 -6.15966737e-01
5.16772807e-01 -1.03288543e+00 -2.16759324e-01 2.66195387e-01
-3.55134547e-01 -2.54047266e-03 1.10850537e+00 -6.50746465e-01
1.05519891e+00 -1.92875254e+00 7.08261371e-01 2.25375786e-01
3.43303859e-01 3.18836927e-01 7.19954446e-03 -2.50483002e-03
-7.24444270e-01 -1.56558543e-01 -1.98563635e-01 -6.87839150e-01
-4.65310887e-02 7.35552907e-01 -2.61353225e-01 6.05030835e-01
3.32739413e-01 1.01824415e+00 -5.96399844e-01 -2.39983022e-01
6.75153136e-01 5.36787629e-01 -2.58089870e-01 1.64012581e-01
-2.09660709e-01 8.22257340e-01 -4.38000381e-01 7.34498739e-01
1.84784755e-01 -2.96686709e-01 7.02762082e-02 -4.42198306e-01
8.98302346e-02 -7.25533068e-02 -1.01834559e+00 2.00372934e+00
-2.83726573e-01 6.32628977e-01 -5.79173028e-01 -1.28054762e+00
8.91669452e-01 3.58954698e-01 7.10283041e-01 -8.21203232e-01
2.05976367e-01 3.66560183e-02 -6.57635555e-02 -7.44842350e-01
7.44331479e-02 -7.70054460e-02 2.31355861e-01 5.39192200e-01
2.50475377e-01 3.41531992e-01 -3.40290278e-01 -1.20282061e-01
1.35094345e+00 8.02628875e-01 3.67113441e-01 1.86453998e-01
6.35085881e-01 -8.52718726e-02 4.17965502e-01 5.51362216e-01
-2.13029012e-01 7.18115628e-01 5.26139677e-01 -1.07123792e+00
-1.02876103e+00 -8.57258260e-01 7.77863562e-02 7.64673948e-01
-2.80043423e-01 -3.58554780e-01 -7.08521366e-01 -9.42331195e-01
-8.66446570e-02 4.89436001e-01 -1.23583710e+00 -4.55417842e-01
-7.65713871e-01 -1.85429499e-01 6.04748309e-01 1.05182922e+00
5.08810580e-01 -1.18701792e+00 -1.14203870e+00 -3.14067863e-02
1.82166725e-01 -1.18460965e+00 2.21549198e-02 3.71378303e-01
-1.30413830e+00 -1.31187105e+00 -3.94130707e-01 -3.68356228e-01
7.84322977e-01 -5.88300526e-02 7.70946562e-01 1.81415126e-01
-2.97775894e-01 8.41417670e-01 -4.79727268e-01 -2.99765855e-01
-4.39497650e-01 -1.34274438e-01 1.57505721e-01 3.92426312e-01
1.54183492e-01 -6.66695297e-01 -5.56710243e-01 3.73078465e-01
-1.10999167e+00 4.78778094e-01 6.72465444e-01 8.56339812e-01
8.53608966e-01 -2.08406851e-01 1.11705311e-01 -5.80837786e-01
-1.66739851e-01 -5.93767129e-02 -1.94191128e-01 4.29819465e-01
-5.13826072e-01 4.87559885e-01 6.28894150e-01 -5.61390877e-01
-8.47595870e-01 5.00653267e-01 7.95571432e-02 -1.19751155e+00
-3.91049564e-01 4.61631805e-01 -2.10239395e-01 -1.34914564e-02
5.92933416e-01 3.24099034e-01 2.16432005e-01 -7.03848362e-01
1.64476380e-01 2.27967069e-01 6.35775626e-01 -3.63436908e-01
6.71936810e-01 7.99370527e-01 5.23970604e-01 -5.71818233e-01
-8.42640281e-01 -3.62800211e-01 -1.35936177e+00 -5.31929553e-01
1.09844589e+00 -5.97848654e-01 -6.53724134e-01 7.08083510e-01
-1.11219943e+00 -5.29844403e-01 -6.56256616e-01 3.79406482e-01
-9.69646513e-01 4.09522682e-01 -1.20812662e-01 -6.17098987e-01
-1.94294035e-01 -1.21553707e+00 1.38702214e+00 -1.09596193e-01
-2.76003838e-01 -1.08867586e+00 1.59020185e-01 5.23547173e-01
1.89329103e-01 6.04847133e-01 9.76867139e-01 -7.35810220e-01
-6.81956410e-01 -1.14368975e-01 1.14491865e-01 5.95916808e-01
8.87236968e-02 -1.89462855e-01 -1.15849221e+00 -1.44268470e-02
9.25756395e-02 -2.17667282e-01 9.77405906e-01 3.42112899e-01
1.51973450e+00 -1.04863063e-01 -1.55047998e-01 8.76497388e-01
1.08643556e+00 1.84166487e-02 9.33941960e-01 5.24824739e-01
9.11104500e-01 3.73381913e-01 5.45899093e-01 4.62103844e-01
-2.19427738e-02 1.02974474e+00 7.49147117e-01 -2.08501890e-01
-3.61727148e-01 -1.00575387e-01 5.41030169e-01 4.51436698e-01
-7.45161712e-01 1.64109230e-01 -9.23978984e-01 2.15641901e-01
-2.26270151e+00 -1.00395203e+00 -1.18758194e-01 2.05512357e+00
1.85530782e-01 2.52249986e-01 8.54974389e-02 5.71580231e-01
3.06234509e-01 1.03534840e-01 -7.43630826e-01 -4.11753446e-01
-8.92110839e-02 6.40550435e-01 3.67312044e-01 1.23086728e-01
-1.11949456e+00 7.83717573e-01 5.90906143e+00 5.70739210e-01
-1.20884657e+00 5.76817282e-02 4.68634069e-01 6.28537908e-02
7.31198862e-02 7.38459155e-02 -4.58264917e-01 -1.35665610e-01
1.02259886e+00 2.24015236e-01 1.14024244e-01 8.70289207e-01
4.50328469e-01 1.08621933e-01 -1.60440075e+00 8.95599186e-01
2.20784649e-01 -1.32331729e+00 4.05404836e-01 1.42889187e-01
3.04542005e-01 -2.83602089e-01 6.29187673e-02 1.95639938e-01
-2.94912875e-01 -1.38260686e+00 8.39590192e-01 9.15192544e-01
7.28429079e-01 -3.61378372e-01 6.04241133e-01 1.99470118e-01
-1.17833126e+00 -2.70948797e-01 -1.21047720e-01 -2.64856189e-01
-3.86484452e-02 -7.75100067e-02 -9.30668116e-01 8.28792572e-01
5.97887278e-01 1.32888889e+00 -8.77130568e-01 5.68244219e-01
-3.09141248e-01 4.67316955e-01 -2.93224398e-02 3.89079064e-01
1.56093612e-01 1.53934538e-01 4.68353033e-01 9.37315822e-01
1.82141438e-01 -9.14159603e-03 -3.34217846e-02 7.32567430e-01
3.23706627e-01 -3.82538855e-01 -6.87675416e-01 1.00122154e-01
-4.74538624e-01 9.82288420e-01 -6.43822908e-01 -3.42592031e-01
-1.68168247e-01 1.17148328e+00 2.48564884e-01 3.13167393e-01
-9.60861683e-01 5.12926459e-01 5.20654142e-01 4.92806211e-02
4.87039804e-01 -1.96122050e-01 -2.37123713e-01 -1.22906625e+00
1.17836066e-01 -1.04141355e+00 5.04270554e-01 -9.69426990e-01
-7.13768840e-01 4.72551852e-01 3.37082505e-01 -1.58081686e+00
-4.04119432e-01 -1.09619856e+00 -3.54019225e-01 5.45719743e-01
-1.16976047e+00 -1.74021161e+00 -4.21542019e-01 9.52679098e-01
7.86269963e-01 -3.16417545e-01 1.10654080e+00 -3.02276928e-02
-5.75355053e-01 4.12438571e-01 -2.48806998e-01 1.43110424e-01
1.95949435e-01 -1.21979654e+00 3.16511720e-01 8.13775599e-01
3.77240568e-01 4.47752386e-01 4.87913460e-01 -3.96570504e-01
-1.61172116e+00 -1.00963593e+00 4.25490886e-01 -7.72209048e-01
6.15494311e-01 -3.38440463e-02 -7.09206223e-01 1.08007479e+00
-1.40924051e-01 1.96548909e-01 6.08434498e-01 -5.20201661e-02
-4.17142540e-01 -1.84582755e-01 -7.84972787e-01 3.95617574e-01
1.30333805e+00 -5.15475571e-01 -7.86985874e-01 2.92169034e-01
4.30848002e-01 -6.18235767e-01 -8.97301137e-01 6.35057211e-01
1.03169179e+00 -1.10470855e+00 1.06501949e+00 -1.17682517e+00
6.10609472e-01 -3.02190453e-01 -2.04135790e-01 -9.98489976e-01
-2.80802697e-01 -3.54825556e-01 -6.78518355e-01 5.93155801e-01
7.27533773e-02 -4.10321057e-01 8.49968016e-01 5.20404637e-01
-4.86838639e-01 -9.61966395e-01 -1.26649570e+00 -6.71195447e-01
-3.46507102e-01 -8.47212374e-01 3.58496040e-01 5.90846479e-01
-4.22014982e-01 1.76137358e-01 -4.72449005e-01 1.95491850e-01
5.63420713e-01 -9.77992639e-02 8.49887609e-01 -1.04392636e+00
-5.47146976e-01 -3.03219527e-01 -1.10873401e+00 -9.72001433e-01
2.25785211e-01 -8.43043983e-01 -3.07425767e-01 -1.46817112e+00
3.27423513e-02 6.84874877e-02 -5.77688038e-01 9.36169207e-01
3.80576789e-01 1.61451221e-01 1.93442956e-01 3.03259701e-01
-6.40757740e-01 5.07339537e-01 1.36712039e+00 -3.20780843e-01
2.36750524e-02 1.99167743e-01 -4.62160558e-02 8.10030520e-01
5.77033401e-01 -3.05153221e-01 -4.81671572e-01 -4.74755138e-01
-1.35509610e-01 -5.16229868e-03 1.06537414e+00 -1.23567927e+00
1.43931568e-01 -4.73382547e-02 7.18893290e-01 -6.68178976e-01
4.90889221e-01 -1.13188624e+00 5.38421571e-01 6.99620605e-01
-4.18680757e-01 6.78238794e-02 2.31838465e-01 5.59094965e-01
-6.66986555e-02 9.64834839e-02 5.77205896e-01 -8.25287104e-02
-1.05492353e+00 3.63208503e-01 -2.17243984e-01 -4.90036070e-01
9.41284835e-01 -7.47656286e-01 -1.57142505e-01 -1.69976100e-01
-1.19309103e+00 -1.95234939e-01 2.27972776e-01 4.56922203e-01
9.34428155e-01 -1.34276402e+00 -3.43613029e-01 5.08333504e-01
2.86948264e-01 -1.27648525e-02 5.04964530e-01 1.22221434e+00
-4.58003432e-01 4.86156255e-01 -5.62969148e-01 -1.07654130e+00
-1.45821309e+00 3.97489220e-01 7.44597733e-01 -4.51448441e-01
-1.07630265e+00 3.85953248e-01 2.03987241e-01 -2.61356562e-01
4.28064615e-01 -7.98575521e-01 -3.32803309e-01 -2.42593840e-01
3.53318870e-01 2.93395609e-01 1.79775760e-01 -9.20338392e-01
-3.66393834e-01 7.05446661e-01 6.70509040e-02 5.59709929e-02
1.51915038e+00 2.00672224e-01 1.47106320e-01 5.48327565e-01
1.13526034e+00 -8.49847019e-01 -1.44205332e+00 -5.66880517e-02
-2.65783407e-02 -3.80552024e-01 -4.54303883e-02 -7.56948054e-01
-1.23876297e+00 1.03712928e+00 1.00521755e+00 -1.38953432e-01
1.36923993e+00 5.39240502e-02 4.26928014e-01 4.39455956e-01
3.33471656e-01 -9.89624262e-01 5.18842638e-01 5.16821802e-01
1.15651882e+00 -1.06453395e+00 2.18961701e-01 -1.47813410e-02
-5.25987566e-01 1.55176055e+00 5.05163729e-01 6.06826805e-02
5.18296123e-01 -1.37615174e-01 -1.68793164e-02 -5.68048835e-01
-7.89560080e-01 -2.05090091e-01 5.86802721e-01 7.33125925e-01
1.85446162e-02 -6.14989139e-02 2.63878852e-01 5.64345241e-01
7.04578906e-02 1.69310659e-01 2.22539857e-01 9.32813287e-01
-5.25016077e-02 -1.13466418e+00 -2.29400501e-01 2.31680602e-01
-1.38262630e-01 3.04044336e-01 -5.26851237e-01 1.06990266e+00
1.95996284e-01 4.36723471e-01 -2.35878304e-01 -7.11050868e-01
6.54770553e-01 3.04501534e-01 7.49804616e-01 -4.73919541e-01
-6.20309114e-01 -8.75303987e-03 5.28017767e-02 -1.05314064e+00
-9.55624759e-01 -8.06832910e-01 -1.33078742e+00 1.66551560e-01
6.82918131e-02 -6.43805563e-01 5.12743831e-01 1.44721556e+00
2.84190565e-01 8.78083169e-01 3.04441899e-01 -9.79417741e-01
-5.41667998e-01 -9.00789261e-01 -3.69376689e-01 5.70222080e-01
4.39369708e-01 -1.05067468e+00 -2.36084208e-01 4.72536534e-01] | [7.895671844482422, 0.41903677582740784] |
e26d3fdf-ca13-48cb-96c9-ae1863c2cde7 | training-multimedia-event-extraction-with | 2306.08966 | null | https://arxiv.org/abs/2306.08966v1 | https://arxiv.org/pdf/2306.08966v1.pdf | Training Multimedia Event Extraction With Generated Images and Captions | Contemporary news reporting increasingly features multimedia content, motivating research on multimedia event extraction. However, the task lacks annotated multimodal training data and artificially generated training data suffer from the distribution shift from the real-world data. In this paper, we propose Cross-modality Augmented Multimedia Event Learning (CAMEL), which successfully utilizes artificially generated multimodal training data and achieves state-of-the-art performance. Conditioned on unimodal training data, we generate multimodal training data using off-the-shelf image generators like Stable Diffusion and image captioners like BLIP. In order to learn robust features that are effective across domains, we devise an iterative and gradual annealing training strategy. Substantial experiments show that CAMEL surpasses state-of-the-art (SOTA) baselines on the M2E2 benchmark. On multimedia events in particular, we outperform the prior SOTA by 4.2\% F1 on event mention identification and by 9.8\% F1 on argument identification, which demonstrates that CAMEL learns synergistic representations from the two modalities. | ['Boyang Li', 'Yidan Sun', 'Xu Guo', 'Yunxin Li', 'Zilin Du'] | 2023-06-15 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [ 6.14277959e-01 2.62556911e-01 -2.03008771e-01 -1.47013962e-01
-1.70456803e+00 -6.85310960e-01 1.27289259e+00 2.64670938e-01
-8.08456361e-01 8.12404394e-01 4.92758840e-01 5.40920272e-02
2.24492729e-01 -6.32403851e-01 -1.36006224e+00 -3.47088635e-01
3.57421041e-02 4.38230932e-01 1.02060981e-01 -4.02547091e-01
4.76830937e-02 -3.20687652e-01 -1.75236964e+00 1.09548604e+00
6.61638856e-01 8.90903056e-01 1.83364227e-01 7.41705537e-01
-2.46238559e-01 1.01698911e+00 -6.29880846e-01 -8.84006679e-01
-3.01705092e-01 -6.73484504e-01 -7.65981853e-01 1.43198267e-01
3.92855614e-01 -1.78056285e-01 -5.68507433e-01 7.79603481e-01
6.79201782e-01 2.32098222e-01 8.30399752e-01 -1.27936864e+00
-9.52946603e-01 1.04901326e+00 -5.94603121e-01 3.66076589e-01
6.07294440e-01 6.16061427e-02 1.03012896e+00 -1.26129222e+00
1.11404037e+00 1.18842053e+00 5.22785246e-01 6.38880134e-01
-1.04308987e+00 -4.35105085e-01 2.46954069e-01 3.29341292e-01
-1.06738567e+00 -4.85162824e-01 7.21462309e-01 -1.77826986e-01
8.60024333e-01 2.61062473e-01 1.21735699e-01 2.01723623e+00
-3.18650216e-01 1.36477375e+00 8.98877919e-01 -5.93138278e-01
2.53321640e-02 2.17145100e-01 -1.15014791e-01 5.70484102e-01
-3.61112244e-02 -1.89120606e-01 -1.14088500e+00 -4.23884466e-02
3.11565787e-01 -4.39577669e-01 -2.76313901e-01 1.40560970e-01
-1.66778624e+00 8.59930336e-01 1.05517544e-01 1.75745055e-01
-5.32151222e-01 5.73537238e-02 5.60860157e-01 1.34459227e-01
5.06708324e-01 3.41963083e-01 -2.11291611e-01 -3.26777250e-01
-8.27793360e-01 2.62313247e-01 6.09380186e-01 8.35986912e-01
3.35390955e-01 -7.74798393e-02 -2.71976531e-01 1.07656252e+00
1.00475684e-01 7.16035008e-01 6.94930255e-01 -7.29294062e-01
9.65324283e-01 6.48421824e-01 9.88366380e-02 -8.22632432e-01
-4.02247161e-01 -1.61029264e-01 -6.53670669e-01 -3.38386089e-01
5.11475980e-01 -3.88893902e-01 -9.68384802e-01 2.02255607e+00
1.48229167e-01 2.79670417e-01 4.83983248e-01 7.96385288e-01
1.37776566e+00 1.09630442e+00 3.14390928e-01 -7.64597505e-02
1.52485037e+00 -9.04891551e-01 -8.17181945e-01 -6.18069172e-01
4.46586430e-01 -7.63537467e-01 1.14274800e+00 2.60361522e-01
-1.35220718e+00 -4.33686018e-01 -1.00790811e+00 7.83220306e-02
-5.08206964e-01 1.53764024e-01 3.82580012e-01 3.83311272e-01
-4.48912233e-01 8.91879052e-02 -6.77911937e-01 -3.79412979e-01
3.58345598e-01 -2.49920543e-02 -6.41940951e-01 -1.47215128e-01
-1.49665153e+00 8.48817170e-01 7.32744515e-01 -2.76985645e-01
-1.03370869e+00 -7.91030943e-01 -1.17320466e+00 -1.46061435e-01
5.63065171e-01 -6.67645574e-01 1.42108142e+00 -9.93547678e-01
-1.31795919e+00 1.17306900e+00 3.38083971e-03 -6.40848815e-01
5.46388865e-01 -4.31271493e-01 -7.30345845e-01 6.28500223e-01
1.24056041e-01 1.23215067e+00 9.12981629e-01 -1.49046385e+00
-7.56954551e-01 1.28658742e-01 -1.37806684e-01 2.83358246e-01
-5.57072937e-01 4.23257202e-02 -5.43645799e-01 -8.82590950e-01
-3.59595507e-01 -9.63380039e-01 4.66546342e-02 -5.00476360e-01
-5.32673776e-01 -1.29073709e-01 5.36785722e-01 -7.84743369e-01
1.08274674e+00 -2.25000310e+00 2.88363963e-01 8.86260252e-03
-1.14429668e-01 -9.63425711e-02 -4.79665577e-01 6.14386559e-01
-1.53723154e-02 2.99296044e-02 -2.54526347e-01 -6.28036499e-01
4.00261253e-01 4.26238850e-02 -3.94496083e-01 1.36229530e-01
5.40280461e-01 1.02739739e+00 -9.63557303e-01 -6.37982190e-01
-3.48122008e-02 5.16864538e-01 -4.55724508e-01 2.55434930e-01
-4.88264143e-01 4.54131216e-01 -1.55624330e-01 5.86787641e-01
2.38172099e-01 -6.66315734e-01 2.97525227e-01 -2.50745863e-01
6.40706047e-02 1.47615701e-01 -1.08351123e+00 2.07378578e+00
-4.51617390e-01 7.49244034e-01 -3.72822404e-01 -9.24464881e-01
5.58078408e-01 5.80360532e-01 4.27963972e-01 -9.30078983e-01
2.18476579e-01 1.46204665e-01 -4.47253406e-01 -6.23499095e-01
7.94280231e-01 1.51404649e-01 -6.32859707e-01 4.68311280e-01
4.66182619e-01 2.71369874e-01 4.52822626e-01 5.98802865e-01
9.17541385e-01 7.47557878e-02 6.02569580e-02 2.43586540e-01
2.67906070e-01 1.54593602e-01 1.96213901e-01 9.64512587e-01
1.32708460e-01 7.74147451e-01 4.06398118e-01 -1.07859708e-01
-9.58531559e-01 -1.11988115e+00 3.32037620e-02 1.41134286e+00
3.29647094e-01 -4.11855131e-01 -7.65851378e-01 -8.87477875e-01
-2.58370727e-01 9.37664747e-01 -7.12517738e-01 -8.97221342e-02
-5.41095018e-01 -1.09830260e+00 8.05928349e-01 5.69881439e-01
6.50285065e-01 -1.18286562e+00 -5.11175513e-01 3.35521489e-01
-8.51096094e-01 -1.53149533e+00 -3.89794022e-01 -1.01765744e-01
-2.62619257e-01 -9.35406208e-01 -8.87211025e-01 -8.65139365e-01
4.54602897e-01 -8.42456799e-03 1.44836056e+00 -4.90998119e-01
-1.45215169e-01 7.72315323e-01 -6.83930576e-01 -3.74632090e-01
-5.05715966e-01 2.91647851e-01 -3.43890876e-01 2.19668537e-01
2.23324940e-01 -2.35466212e-01 -6.39702678e-01 1.70239538e-01
-1.25714326e+00 2.80463785e-01 7.39757895e-01 1.03287315e+00
5.97462595e-01 -1.58810228e-01 1.00829542e+00 -8.35579097e-01
3.82767290e-01 -9.17743742e-01 -6.59365654e-02 2.95468360e-01
-1.16632566e-01 -1.44243578e-03 1.87509581e-01 -8.26048076e-01
-1.52558506e+00 5.90576641e-02 7.43145794e-02 -1.64632544e-01
-3.17986429e-01 8.43510568e-01 -1.53978735e-01 5.36013305e-01
8.96602452e-01 2.45596230e-01 -2.89392561e-01 -2.98686296e-01
8.38128746e-01 6.61635756e-01 1.04496980e+00 -6.84832633e-01
4.97087747e-01 5.23087025e-01 -5.58526218e-01 -6.56717300e-01
-1.01927221e+00 -1.43496960e-01 -1.13137968e-01 -4.86980557e-01
1.11198115e+00 -1.21166170e+00 -4.18674558e-01 3.17111641e-01
-1.07015884e+00 -2.08632112e-01 -1.19367398e-01 4.60226297e-01
-5.73316872e-01 7.44878277e-02 -7.31334329e-01 -6.30313873e-01
-1.53261945e-01 -9.30534124e-01 1.22466099e+00 1.64318711e-01
-2.97479838e-01 -7.90630817e-01 8.74847472e-02 6.55217707e-01
4.43884656e-02 5.57351410e-01 6.69219375e-01 -8.85414720e-01
-4.75830793e-01 -2.59791076e-01 -1.47100970e-01 1.33982617e-02
-3.03881824e-01 -3.19143593e-01 -1.07334650e+00 -4.23794538e-02
-5.12720168e-01 -8.10034454e-01 1.02473915e+00 1.17226779e-01
8.76708686e-01 -2.58405209e-01 -3.30445975e-01 1.29060805e-01
1.04772067e+00 -2.16189567e-02 6.83989048e-01 6.65326118e-01
4.24545586e-01 7.25311160e-01 5.25180876e-01 5.97643733e-01
6.55843318e-01 5.22905231e-01 3.61675233e-01 -2.39532385e-02
-2.53055036e-01 -5.34324586e-01 6.54516816e-01 6.85024023e-01
2.09439829e-01 -6.85445130e-01 -8.72158885e-01 7.69960105e-01
-2.14356279e+00 -1.25474226e+00 -2.10545352e-03 1.80562603e+00
1.15100920e+00 2.18627259e-01 2.16026962e-01 -9.29281302e-03
8.82781208e-01 1.37820765e-01 -3.54805529e-01 1.11020654e-01
-5.88507295e-01 -3.18379514e-02 2.58267134e-01 5.98415509e-02
-1.49229944e+00 7.48427987e-01 5.74691963e+00 9.51668084e-01
-8.23653400e-01 2.86108136e-01 7.32092977e-01 -1.94047585e-01
-3.74907464e-01 -4.47496206e-01 -7.49890149e-01 5.56279182e-01
1.31145597e+00 -1.14962555e-01 1.19989693e-01 6.04548335e-01
-3.13191861e-01 1.29155237e-02 -1.07094061e+00 1.20851409e+00
4.62121993e-01 -1.54924798e+00 2.48491973e-01 -1.61513641e-01
9.62354779e-01 -2.86217686e-02 3.41073066e-01 6.01244986e-01
1.80790171e-01 -8.15698862e-01 9.08549547e-01 4.16046023e-01
6.85658455e-01 -8.05013835e-01 6.61499262e-01 2.04445258e-01
-9.40909088e-01 5.86937554e-02 2.35254124e-01 4.37848330e-01
6.59232616e-01 3.81839246e-01 -7.07223058e-01 6.24185085e-01
5.60802281e-01 6.62082970e-01 -6.80509329e-01 8.01099062e-01
-1.31882101e-01 6.76765025e-01 -3.53940189e-01 7.47768506e-02
3.95142555e-01 4.27272379e-01 6.00683033e-01 1.64750576e+00
3.21649164e-01 -6.26952946e-02 2.47210607e-01 5.03682077e-01
-6.99751198e-01 1.93141639e-01 -5.65807343e-01 -3.83907378e-01
3.85640740e-01 1.11745548e+00 -6.60398662e-01 -5.13685763e-01
-6.71190560e-01 1.22327173e+00 1.92239687e-01 3.48220885e-01
-1.27491200e+00 -2.31205180e-01 -1.58129856e-02 -3.63798708e-01
4.65844482e-01 9.92743373e-02 9.07976031e-02 -1.26531672e+00
-1.05021447e-01 -1.06483841e+00 8.57390463e-01 -7.80283630e-01
-1.69687331e+00 8.08286011e-01 1.89574242e-01 -1.28805602e+00
-4.76122916e-01 -5.07317185e-01 -2.60450065e-01 1.83278158e-01
-1.43253744e+00 -1.38435340e+00 -1.40614346e-01 6.94545984e-01
8.58681679e-01 -2.86307484e-01 7.20348120e-01 6.99150920e-01
-6.18028879e-01 7.43983865e-01 -5.67790978e-02 2.65564382e-01
9.83072698e-01 -1.11641073e+00 2.55852520e-01 7.78327405e-01
4.96767372e-01 7.64591694e-02 5.97266614e-01 -6.49080753e-01
-1.39982641e+00 -9.97839570e-01 7.63566256e-01 -5.21068156e-01
8.59570861e-01 -2.26419717e-01 -8.14248860e-01 8.04284692e-01
8.09696734e-01 -2.11802959e-01 9.89158571e-01 -8.61353949e-02
-5.76200247e-01 3.92553449e-01 -9.54589367e-01 7.70529270e-01
9.83410120e-01 -6.73593342e-01 -8.61147165e-01 4.95237559e-01
6.47599876e-01 -4.98713732e-01 -8.70741069e-01 4.91189539e-01
2.63478488e-01 -5.03959179e-01 1.20846832e+00 -9.32602048e-01
8.57894123e-01 -3.83165665e-02 -4.99642551e-01 -1.15967762e+00
7.17325807e-02 -6.96869195e-01 -4.24062103e-01 1.52140570e+00
9.32888567e-01 -1.59589201e-01 4.92263287e-01 4.23451334e-01
-2.40444332e-01 -1.93552062e-01 -9.20940042e-01 -6.17472231e-01
-9.67498273e-02 -6.39712572e-01 2.37578377e-01 1.17816675e+00
2.56816357e-01 6.32451236e-01 -4.58712697e-01 3.45717877e-01
5.94025433e-01 -4.19883709e-03 4.86739814e-01 -6.94609880e-01
-3.58323991e-01 -3.99549514e-01 -8.28845873e-02 -8.06828797e-01
4.37209845e-01 -8.60353529e-01 6.69947267e-02 -1.39149308e+00
4.12694395e-01 1.24382466e-01 -4.20095176e-01 4.16598052e-01
-4.96688843e-01 5.56250215e-01 2.80951887e-01 2.51381267e-02
-1.13935590e+00 7.10606217e-01 8.96673679e-01 -3.25666875e-01
-2.49357354e-02 -5.56228340e-01 -6.32115543e-01 6.69650495e-01
6.44033015e-01 -3.95825684e-01 -4.59881634e-01 -4.66879308e-01
4.97406900e-01 3.60306390e-02 4.64891523e-01 -7.68325388e-01
8.73426050e-02 7.14445710e-02 3.52343112e-01 -5.79356015e-01
6.48784578e-01 -4.47904706e-01 5.78107089e-02 -1.13250859e-01
-7.15239763e-01 1.52961180e-01 4.60349858e-01 7.99359560e-01
-3.87340963e-01 -1.97800964e-01 3.10498744e-01 -7.04052150e-02
-1.01429033e+00 -1.20730147e-01 -5.69988608e-01 5.98848641e-01
9.56649125e-01 2.59566665e-01 -9.50571835e-01 -6.19663775e-01
-7.97486484e-01 2.04491720e-01 1.36226229e-02 6.71476364e-01
7.05968320e-01 -1.52656841e+00 -1.09438646e+00 -2.81827301e-01
4.08690333e-01 -3.47599179e-01 4.37079489e-01 6.65040314e-01
-1.33267716e-01 1.05887696e-01 2.87592076e-02 -5.85559666e-01
-1.07492447e+00 4.47493643e-01 -1.32492498e-01 -1.61177203e-01
-5.40649295e-01 8.68160963e-01 -1.78025831e-02 -2.96125442e-01
2.75991321e-01 1.36215225e-01 -2.91013449e-01 4.34672654e-01
7.80889869e-01 2.18335256e-01 -7.09993392e-02 -5.84667444e-01
-2.78080165e-01 1.10669747e-01 -1.40543371e-01 -6.82550490e-01
1.30467188e+00 -1.98361620e-01 4.63443995e-01 4.95902300e-01
1.16505384e+00 -7.81368613e-02 -1.26475060e+00 -2.74165720e-01
8.57220516e-02 -1.05221078e-01 -1.07213005e-03 -1.07054031e+00
-7.82950759e-01 5.11556983e-01 5.12033284e-01 3.36138844e-01
9.64493454e-01 3.37662369e-01 1.11442256e+00 4.63022351e-01
-3.95280309e-02 -1.19248104e+00 4.92331773e-01 3.54277462e-01
9.13992643e-01 -1.57713878e+00 -3.16918731e-01 -2.63963342e-01
-1.28456581e+00 7.56148398e-01 5.12967408e-01 2.17246234e-01
2.37015441e-01 8.68397728e-02 5.88530786e-02 -1.95797905e-01
-8.88452351e-01 -3.48170608e-01 5.96095741e-01 4.33898956e-01
3.67308021e-01 -2.55368173e-01 -6.14053681e-02 9.62849438e-01
-3.27411704e-02 -2.26148337e-01 2.82316566e-01 1.09877741e+00
-2.25146204e-01 -9.79148269e-01 -4.00156200e-01 1.59800574e-01
-6.89558685e-01 -1.41131729e-01 -3.12795371e-01 1.07551026e+00
-2.62647364e-02 1.00528109e+00 8.61772224e-02 -1.26092896e-01
5.01202881e-01 4.31874961e-01 4.70507383e-01 -2.11688623e-01
-5.31680644e-01 7.32414573e-02 6.62304103e-01 -3.33648264e-01
-6.96892321e-01 -8.45359027e-01 -1.29101729e+00 -8.33974183e-02
-8.19546878e-02 4.96399812e-02 6.52569175e-01 9.41944063e-01
5.78412414e-01 6.24434292e-01 2.94048518e-01 -9.97082829e-01
-1.26452118e-01 -9.31357443e-01 -2.38680597e-02 9.43274796e-01
9.71397012e-02 -6.32329702e-01 -2.34058470e-01 6.25181258e-01] | [10.71348762512207, 1.0775126218795776] |
446c3db3-e450-424e-a85f-c8ddd883f789 | marionette-few-shot-face-reenactment | 1911.08139 | null | https://arxiv.org/abs/1911.08139v1 | https://arxiv.org/pdf/1911.08139v1.pdf | MarioNETte: Few-shot Face Reenactment Preserving Identity of Unseen Targets | When there is a mismatch between the target identity and the driver identity, face reenactment suffers severe degradation in the quality of the result, especially in a few-shot setting. The identity preservation problem, where the model loses the detailed information of the target leading to a defective output, is the most common failure mode. The problem has several potential sources such as the identity of the driver leaking due to the identity mismatch, or dealing with unseen large poses. To overcome such problems, we introduce components that address the mentioned problem: image attention block, target feature alignment, and landmark transformer. Through attending and warping the relevant features, the proposed architecture, called MarioNETte, produces high-quality reenactments of unseen identities in a few-shot setting. In addition, the landmark transformer dramatically alleviates the identity preservation problem by isolating the expression geometry through landmark disentanglement. Comprehensive experiments are performed to verify that the proposed framework can generate highly realistic faces, outperforming all other baselines, even under a significant mismatch of facial characteristics between the target and the driver. | ['Dongyoung Kim', 'Beomsu Kim', 'Martin Kersner', 'Sungjoo Ha', 'Seokjun Seo'] | 2019-11-19 | null | null | null | null | ['face-reenactment'] | ['computer-vision'] | [ 2.67251432e-01 4.59936559e-02 1.42588153e-01 -3.62542540e-01
-7.66352057e-01 -6.08202457e-01 6.11849546e-01 -5.11070788e-01
2.52761319e-02 4.94685173e-01 2.35626325e-01 3.44572097e-01
1.37555242e-01 -3.52020115e-01 -6.89555466e-01 -8.04853499e-01
5.24705946e-01 1.64082870e-01 -5.98287024e-02 -3.70961070e-01
1.40959904e-01 4.39533383e-01 -1.93681300e+00 2.58455992e-01
8.39265406e-01 1.10185492e+00 -1.18618265e-01 6.73074052e-02
-9.11381692e-02 5.87468445e-01 -7.67171502e-01 -8.82264256e-01
5.62430799e-01 -3.35348576e-01 -4.72384959e-01 2.38820434e-01
9.34097767e-01 -4.18063253e-01 -2.33634397e-01 1.30653226e+00
6.40771687e-01 4.21008607e-03 3.82124931e-01 -1.73760176e+00
-5.38824499e-01 9.31813717e-02 -8.92690778e-01 -2.19631523e-01
4.19241607e-01 1.01428762e-01 7.71776140e-01 -1.35886371e+00
6.30402625e-01 1.62441456e+00 5.98903477e-01 7.34569490e-01
-1.31354332e+00 -1.25591600e+00 1.09633990e-01 2.69758493e-01
-1.70421374e+00 -1.03713846e+00 9.39108551e-01 -3.78002793e-01
2.46814743e-01 2.07084879e-01 2.72205085e-01 1.26194715e+00
3.52682583e-02 4.59253848e-01 8.87200058e-01 -2.24261642e-01
-1.60433993e-01 4.09792542e-01 -1.43936396e-01 5.42329609e-01
5.33372015e-02 2.68059283e-01 -6.33535206e-01 -1.15127496e-01
4.99442458e-01 -9.26046222e-02 -4.84111458e-01 -3.81821156e-01
-8.23454559e-01 3.71625036e-01 4.65868890e-01 3.38663533e-02
-1.12795107e-01 -2.00246066e-01 3.08977008e-01 2.11930126e-01
3.74562025e-01 2.79473692e-01 -1.58652849e-02 1.49740666e-01
-7.46067584e-01 2.33809575e-01 3.70383382e-01 1.22055173e+00
9.69924271e-01 1.63412467e-01 -3.79306644e-01 8.03903759e-01
7.62494579e-02 3.46200228e-01 3.65988433e-01 -7.81572700e-01
3.82070631e-01 6.19770408e-01 1.80282727e-01 -1.24136293e+00
-9.03569069e-03 -3.67155552e-01 -8.03961098e-01 4.52471823e-01
3.47835243e-01 -1.56880379e-01 -9.35935557e-01 2.32725191e+00
4.72468048e-01 4.99935120e-01 1.26854956e-01 1.00676572e+00
1.07608283e+00 4.04352039e-01 -7.09408335e-03 -2.14620978e-01
1.49031365e+00 -9.09249842e-01 -9.31564808e-01 -4.51386422e-01
1.49600148e-01 -9.83691990e-01 8.72632563e-01 -4.93249558e-02
-8.74508083e-01 -8.42822611e-01 -1.06720901e+00 -1.21415667e-01
-2.98462640e-02 1.02770440e-01 2.30946928e-01 5.50796926e-01
-1.02090919e+00 4.70152795e-01 -1.64172351e-01 -3.72179687e-01
4.74240661e-01 4.70843256e-01 -8.55383933e-01 -2.27584526e-01
-1.14340127e+00 8.29971135e-01 1.03958815e-01 4.44537520e-01
-7.91777313e-01 -9.51743245e-01 -1.01124859e+00 2.09108237e-02
3.54620308e-01 -5.44738591e-01 1.01787794e+00 -1.51208413e+00
-1.34493864e+00 1.10651708e+00 -4.04851228e-01 1.31364867e-01
7.38772631e-01 -2.67691314e-01 -5.66688657e-01 -1.77476406e-01
1.19426459e-01 7.24632382e-01 1.31310165e+00 -1.51325727e+00
-5.09779990e-01 -6.07092202e-01 -1.82697624e-01 2.93548733e-01
-1.83183163e-01 2.30677370e-02 -5.02092957e-01 -5.82406163e-01
-4.06133123e-02 -8.92064810e-01 3.10752839e-01 1.59484044e-01
-5.10311544e-01 4.00809124e-02 1.11109793e+00 -5.41874886e-01
7.69033253e-01 -2.57683754e+00 1.08557798e-01 -1.19975619e-01
1.81240290e-01 3.39684457e-01 -4.75318134e-01 2.46272042e-01
-5.79375148e-01 3.28430608e-02 7.63084963e-02 -6.32858336e-01
-3.09506416e-01 -1.44452683e-03 -4.95494694e-01 3.85723948e-01
5.04617214e-01 9.01619792e-01 -7.01803446e-01 -4.25572813e-01
-1.01951599e-01 5.95712066e-01 -2.91639686e-01 4.48186547e-01
2.61261225e-01 5.05833864e-01 -1.00258216e-01 5.40805280e-01
1.18126118e+00 3.20117563e-01 -1.98465705e-01 -6.66650593e-01
9.76290852e-02 -3.10154676e-01 -1.14900637e+00 1.55062330e+00
-1.76984176e-01 5.93287468e-01 3.11960906e-01 -4.05501842e-01
1.08295619e+00 3.80029589e-01 2.31661528e-01 -7.08001077e-01
2.20881000e-01 8.77608582e-02 2.06729537e-03 -4.62572724e-01
5.16919911e-01 -4.73267078e-01 5.08332103e-02 2.53885835e-01
8.45136195e-02 2.75545835e-01 -2.76853621e-01 1.11290254e-01
6.96799517e-01 4.84281220e-02 7.90832937e-03 -2.27514982e-01
5.32489419e-01 -3.42685103e-01 1.07073021e+00 2.73468882e-01
-5.06134689e-01 8.48776937e-01 5.58298767e-01 -2.59914160e-01
-9.68029082e-01 -1.10993445e+00 6.72526285e-02 8.80957544e-01
5.81037283e-01 -3.46100539e-01 -9.27055418e-01 -4.90484774e-01
-5.24498615e-03 5.89141071e-01 -7.56572485e-01 -7.62708843e-01
-4.12551939e-01 -5.38503587e-01 6.02201045e-01 2.99683630e-01
6.69429958e-01 -8.27453494e-01 -2.24231154e-01 -1.87037855e-01
-3.76493096e-01 -1.32585227e+00 -8.13740075e-01 -5.68769932e-01
-3.07382286e-01 -1.10776627e+00 -5.60422838e-01 -9.45157170e-01
9.88566875e-01 6.04850352e-01 7.28136420e-01 2.61857659e-01
-2.73848832e-01 3.99724208e-02 -1.25739083e-01 -3.55872631e-01
-4.79433686e-01 -4.20397609e-01 1.92296103e-01 7.96099186e-01
2.17925608e-01 -4.62485135e-01 -5.60532153e-01 4.97008145e-01
-6.72758937e-01 2.48316482e-01 4.01987046e-01 9.53707337e-01
2.73006797e-01 2.08620429e-01 7.28368104e-01 -7.79386938e-01
4.22665596e-01 -2.33500078e-01 -3.81671637e-01 1.45458817e-01
-3.65100741e-01 -8.32838789e-02 7.65409887e-01 -5.84515512e-01
-1.41093493e+00 1.88769564e-01 2.94475351e-02 -7.14071274e-01
-9.18195546e-02 -2.66505599e-01 -8.14176679e-01 -2.04306886e-01
5.60238302e-01 1.36015415e-01 4.18974251e-01 -5.04395366e-01
2.67640680e-01 5.97499311e-01 8.56098413e-01 -4.16426301e-01
1.19946015e+00 6.51329935e-01 -1.38839528e-01 -5.88211060e-01
-6.78807497e-01 -2.65781842e-02 -6.80987597e-01 -3.14209253e-01
4.84445542e-01 -1.08276260e+00 -7.02069998e-01 9.01991904e-01
-1.36307871e+00 3.29866111e-01 -1.58418477e-01 1.45384381e-02
-3.98824662e-01 2.12207407e-01 -2.09156066e-01 -6.62541151e-01
-3.02230209e-01 -1.29504180e+00 1.09500670e+00 5.71420014e-01
-1.30543455e-01 -3.89016271e-01 -2.17076153e-01 3.15767318e-01
3.93319160e-01 4.09841567e-01 1.06264722e+00 -5.07981539e-01
-4.28738356e-01 -2.76654303e-01 -3.83367509e-01 3.03769052e-01
4.93981570e-01 1.84950173e-01 -1.46038651e+00 -5.17307639e-01
1.63352802e-01 -2.50897706e-01 5.35538971e-01 -1.90241456e-01
6.16754532e-01 -2.30597198e-01 -3.81907701e-01 6.80017054e-01
1.07071304e+00 1.48993999e-01 7.89014041e-01 -7.34766126e-02
7.95027316e-01 1.01704335e+00 6.84668005e-01 2.10617304e-01
1.55514136e-01 8.60777259e-01 5.47816515e-01 -4.41034496e-01
-2.95202225e-01 -4.94504571e-01 3.53979498e-01 3.24910998e-01
3.18009168e-01 -1.59408301e-02 -4.96446222e-01 5.09689271e-01
-1.75178051e+00 -9.88967955e-01 2.52193838e-01 2.37737775e+00
6.52607799e-01 -1.07064746e-01 -1.77613217e-02 2.28322241e-02
1.19342434e+00 1.09395206e-01 -7.71848857e-01 -1.79749116e-01
-2.33112261e-01 -2.10284725e-01 -3.13573293e-02 4.77020025e-01
-8.21283042e-01 1.17798781e+00 5.98598433e+00 7.62983382e-01
-1.27033341e+00 8.99733044e-03 6.53519332e-01 -1.92991346e-01
-1.10391177e-01 2.77922954e-02 -7.27059841e-01 4.72115815e-01
2.70467043e-01 -4.81960118e-01 3.44857872e-01 7.90870011e-01
1.76604778e-01 1.54017046e-01 -1.20016110e+00 1.28962111e+00
5.29765427e-01 -8.70672405e-01 1.75204575e-01 -7.22617730e-02
4.12626952e-01 -5.41751623e-01 3.14116091e-01 2.02154830e-01
-1.23573087e-01 -1.10342181e+00 9.95714009e-01 3.90056461e-01
1.11540449e+00 -9.96797681e-01 7.02099919e-01 1.61004990e-01
-1.18007851e+00 -1.07358254e-01 -3.17864656e-01 9.64085087e-02
9.82414037e-02 1.72233745e-01 -7.24982202e-01 5.79316258e-01
5.67333341e-01 6.26592278e-01 -6.69092357e-01 8.40248704e-01
-2.10890502e-01 -7.64058307e-02 -8.65625218e-02 6.47216678e-01
-2.60964572e-01 -1.06892213e-01 8.81134868e-01 6.63893640e-01
3.51809502e-01 5.32872118e-02 -9.88014266e-02 1.14243937e+00
-2.59834051e-01 3.76298688e-02 -7.29708672e-01 3.03241819e-01
6.42261744e-01 1.42700803e+00 -1.22622453e-01 -1.42245805e-02
-2.59265542e-01 1.22907889e+00 3.59465927e-01 5.10537684e-01
-8.57997596e-01 -4.38813716e-01 1.29033160e+00 1.68158665e-01
1.16006685e-02 3.62614870e-01 -7.27798194e-02 -1.07093263e+00
3.77339780e-01 -1.08358467e+00 1.41304046e-01 -8.86334598e-01
-1.15350723e+00 9.16596830e-01 -2.27124676e-01 -1.29665744e+00
-1.06204368e-01 -2.10747346e-01 -8.60800445e-01 9.54957843e-01
-1.40690029e+00 -1.40983200e+00 -6.49859786e-01 7.11927831e-01
4.84462857e-01 -2.34873816e-01 6.54242873e-01 5.45244455e-01
-9.58808303e-01 1.19201195e+00 -3.71815711e-01 4.03205454e-02
1.10135937e+00 -6.26678765e-01 3.06664973e-01 9.15586889e-01
-1.18992075e-01 5.26661456e-01 7.58481860e-01 -4.63545173e-01
-1.30900216e+00 -1.21896732e+00 6.98827744e-01 -3.75447154e-01
2.18554288e-01 -7.07685232e-01 -1.11437929e+00 5.96683800e-01
4.29008603e-02 7.10119903e-02 4.45664823e-01 -2.33049601e-01
-6.44209385e-01 -4.66546655e-01 -1.24412012e+00 7.68150032e-01
1.06809556e+00 -6.05942667e-01 -3.97198826e-01 -9.48065072e-02
6.81776285e-01 -4.01383191e-01 -5.55975854e-01 5.07307470e-01
7.57783353e-01 -1.14931357e+00 8.13238323e-01 -6.47401810e-01
3.02248240e-01 -4.51678485e-01 1.02794506e-02 -1.34782302e+00
-3.51518154e-01 -9.58276868e-01 2.72116214e-01 1.93538046e+00
1.79922536e-01 -5.21305144e-01 6.16112947e-01 8.97690177e-01
1.30593687e-01 -4.86966193e-01 -1.08482945e+00 -6.92471743e-01
-5.16711026e-02 1.20057464e-01 1.07351506e+00 9.19575930e-01
-2.41088286e-01 5.68495929e-01 -7.63626695e-01 1.96554616e-01
6.54542744e-01 1.53517410e-01 1.05949080e+00 -1.04155362e+00
1.58920497e-01 -3.55403244e-01 -6.16096139e-01 -7.48469472e-01
5.53566039e-01 -7.19075799e-01 2.38714322e-01 -8.33248854e-01
3.89804423e-01 -2.70391941e-01 -1.13080747e-01 5.15203834e-01
-3.54099125e-01 2.99067616e-01 3.65151584e-01 2.03134015e-01
-6.90416470e-02 7.96395183e-01 1.35659230e+00 -3.07582822e-02
-3.29631045e-02 -1.94852293e-01 -1.00624907e+00 7.18794584e-01
5.02165496e-01 -4.99565154e-01 -4.07181561e-01 -4.91873354e-01
-2.04610541e-01 -8.97131637e-02 5.15458584e-01 -8.08109462e-01
2.85076976e-01 -7.53147304e-02 3.78519237e-01 -2.72407204e-01
7.09863722e-01 -9.51314569e-01 4.99170065e-01 9.12076533e-02
-7.93234110e-02 1.72299683e-01 4.94603723e-01 4.97989506e-01
-3.30099672e-01 9.14337263e-02 1.28935754e+00 2.28517383e-01
-6.69137597e-01 3.99407059e-01 1.34504989e-01 7.68290535e-02
1.19243646e+00 -5.58927178e-01 -4.86368299e-01 -5.12381971e-01
-4.37682092e-01 2.27465913e-01 7.87520885e-01 9.23503339e-01
5.59963524e-01 -1.62110722e+00 -9.03595388e-01 7.97713757e-01
3.11484158e-01 -2.17085868e-01 5.51422775e-01 7.27748573e-01
3.36601250e-02 -9.18738544e-02 -4.77486044e-01 -4.72916722e-01
-1.54028893e+00 5.96062839e-01 5.94376504e-01 3.54418635e-01
-5.99288166e-01 9.37759459e-01 9.15176392e-01 -1.12014361e-01
1.58568203e-01 3.33725929e-01 -6.06214814e-02 1.69679582e-01
7.67304540e-01 1.38813242e-01 -3.75296175e-02 -1.46624255e+00
-4.56296235e-01 8.94147515e-01 -3.16478372e-01 1.22730121e-01
8.65444183e-01 -3.80835414e-01 -8.99043381e-02 1.07271068e-01
1.26552224e+00 1.61832199e-02 -1.42384541e+00 -3.69433343e-01
-3.86454999e-01 -9.11311269e-01 -1.49953455e-01 -6.26412988e-01
-1.30144405e+00 8.83823812e-01 6.09643757e-01 -3.54606003e-01
1.21468103e+00 -2.34710276e-01 9.04819071e-01 -2.39257336e-01
2.69679904e-01 -7.79658437e-01 1.58590928e-01 1.75008222e-01
1.17022133e+00 -1.35767782e+00 -3.00055385e-01 -6.91607058e-01
-8.20473433e-01 8.07779074e-01 1.01120734e+00 1.03386350e-01
4.21074688e-01 8.67271274e-02 3.55196565e-01 -7.12218508e-02
-6.66588485e-01 4.38655466e-02 3.64850432e-01 7.03201175e-01
-7.39301145e-02 -2.00520515e-01 3.80238682e-01 7.63986826e-01
-4.60267246e-01 -4.13503617e-01 4.26969558e-01 5.66544235e-01
-1.20487005e-01 -9.55609083e-01 -5.48492610e-01 -5.64209977e-03
-3.61821949e-01 7.45652691e-02 -8.15240860e-01 7.59174347e-01
2.86139727e-01 9.80999589e-01 2.08984718e-01 -6.51150227e-01
6.81424737e-01 9.36688334e-02 3.82857114e-01 -4.97724771e-01
-5.16504109e-01 9.85336974e-02 -6.69935197e-02 -6.47745013e-01
-1.45753874e-02 -3.98699284e-01 -1.05074358e+00 -5.73020816e-01
-5.82344472e-01 -8.86542350e-02 2.36464292e-01 7.91986287e-01
8.19179535e-01 2.70491123e-01 9.23181713e-01 -7.27174222e-01
-5.20804703e-01 -7.59958029e-01 -7.25226760e-01 9.45281446e-01
5.64572275e-01 -9.98591006e-01 -4.80272949e-01 -7.92474449e-02] | [12.770840644836426, 0.036588504910469055] |
6c0c9a30-fade-44fc-9ba7-93de04f62fb3 | error-mitigated-quantum-approximate | 2303.14877 | null | https://arxiv.org/abs/2303.14877v1 | https://arxiv.org/pdf/2303.14877v1.pdf | Error-mitigated Quantum Approximate Optimization via Learning-based Adaptive Optimization | Combinatorial optimization problems are ubiquitous and computationally hard to solve in general. Quantum computing is envisioned as a powerful tool offering potential computational advantages for solving some of these problems. Quantum approximate optimization algorithm (QAOA), one of the most representative quantum-classical hybrid algorithms, is designed to solve certain combinatorial optimization problems by transforming a discrete optimization problem into a classical optimization problem over a continuous circuit parameter domain. QAOA objective landscape over the parameter variables is notorious for pervasive local minima and barren plateaus, and its viability in training significantly relies on the efficacy of the classical optimization algorithm. To enhance the performance of QAOA, we design double adaptive-region Bayesian optimization (DARBO), an adaptive classical optimizer for QAOA. Our experimental results demonstrate that the algorithm greatly outperforms conventional gradient-based and gradient-free optimizers in terms of speed, accuracy, and stability. We also address the issues of measurement efficiency and the suppression of quantum noise by successfully conducting the full optimization loop on the superconducting quantum processor. This work helps to unlock the full power of QAOA and paves the way toward achieving quantum advantage in practical classical tasks. | ['Shengyu Zhang', 'Shi-Xin Zhang', 'Yu-Qin Chen', 'Lixue Cheng'] | 2023-03-27 | null | null | null | null | ['combinatorial-optimization'] | ['methodology'] | [ 1.18428677e-01 -3.39723527e-01 -9.57609117e-02 -2.66446769e-01
-1.14513838e+00 -2.79348671e-01 2.02712685e-01 8.20765793e-02
-7.00782597e-01 1.09907889e+00 -3.89531732e-01 -3.82852465e-01
-2.68552750e-01 -7.84124851e-01 -5.12282372e-01 -1.12627304e+00
1.13319322e-01 6.55129969e-01 4.02536727e-02 -4.73339051e-01
6.59462273e-01 5.20071328e-01 -1.15718007e+00 -2.87036777e-01
1.26403105e+00 1.06944966e+00 -1.10913850e-02 5.06068051e-01
5.12424335e-02 2.16926277e-01 -4.26003665e-01 -4.72815514e-01
2.08244264e-01 -6.69942617e-01 -6.75026894e-01 -7.07090855e-01
1.17354341e-01 -2.47454178e-02 -5.66261768e-01 1.48075771e+00
8.39012682e-01 2.73478657e-01 4.00026739e-01 -7.74619520e-01
-3.37398857e-01 5.31974792e-01 -1.63807526e-01 3.07206810e-01
1.82982072e-01 4.15607929e-01 1.25690699e+00 -6.77478790e-01
4.17467594e-01 8.64999354e-01 4.64888930e-01 1.99824348e-01
-1.35073864e+00 -4.09504145e-01 -6.95220411e-01 6.38541162e-01
-1.79663742e+00 -4.26553041e-01 4.58595663e-01 4.33776490e-02
1.28354323e+00 2.26663992e-01 7.14884877e-01 3.86017710e-01
7.96347499e-01 4.88075554e-01 1.29995048e+00 -8.05216849e-01
7.24666715e-01 -2.10590865e-02 3.14854205e-01 8.34176123e-01
1.18521176e-01 6.18560612e-01 -9.59739566e-01 -4.62095588e-01
1.12966917e-01 -4.57522988e-01 -2.14554131e-01 -2.44017705e-01
-9.13839340e-01 9.29096937e-01 5.21915019e-01 7.96505287e-02
-2.70998836e-01 4.33678061e-01 1.25277778e-02 1.99213073e-01
-9.09334887e-03 9.54942286e-01 9.00428072e-02 -4.74385321e-01
-9.93578970e-01 6.29434824e-01 8.10723722e-01 6.37336791e-01
1.00993502e+00 8.22522119e-02 -2.08761841e-01 3.87325346e-01
2.88044870e-01 7.48366058e-01 1.29030511e-01 -9.85376835e-01
1.63175501e-02 2.63659537e-01 3.50612283e-01 -3.71033520e-01
-5.69179654e-01 -6.42271578e-01 -4.21269149e-01 1.37203798e-01
4.29648727e-01 9.05923396e-02 -6.65572524e-01 1.38306248e+00
4.18652236e-01 -3.14938247e-01 -6.79826215e-02 1.01038742e+00
3.50975662e-01 8.43783796e-01 -2.90919781e-01 -3.81482750e-01
1.16294575e+00 -8.14416468e-01 -9.01330113e-01 -1.83508083e-01
5.75593829e-01 -7.16613770e-01 7.95498610e-01 5.08603334e-01
-9.94391084e-01 -2.05619529e-01 -1.62497389e+00 -1.17593162e-01
-2.04348639e-01 -2.04316273e-01 9.67998803e-01 1.26587307e+00
-6.70310795e-01 1.02013648e+00 -8.99801254e-01 2.85751611e-01
2.93055981e-01 8.12453687e-01 1.24857642e-01 -2.88067281e-01
-1.19106805e+00 1.18149447e+00 5.53770602e-01 2.48436093e-01
-4.96332973e-01 -4.48468834e-01 -4.55892771e-01 6.70432076e-02
4.45407897e-01 -4.62520987e-01 1.11712730e+00 -1.19262747e-01
-1.96929479e+00 3.96629661e-01 -2.90069073e-01 -5.48068404e-01
-7.94985965e-02 1.05211236e-01 -3.76258612e-01 1.25783488e-01
4.41602655e-02 5.71851358e-02 6.12476826e-01 -2.96250343e-01
-1.20724089e-01 -5.21880031e-01 -2.62403548e-01 1.04604825e-01
1.10127255e-01 2.71313591e-03 -1.38845265e-01 5.97290695e-02
4.08176512e-01 -9.71349418e-01 -5.20318151e-01 -5.61925530e-01
-2.30178624e-01 -1.67186096e-01 1.49828151e-01 -2.32733592e-01
1.43917954e+00 -1.79556715e+00 3.02933007e-01 5.63341558e-01
6.29035905e-02 2.29413345e-01 3.61882210e-01 4.97013450e-01
3.93904120e-01 -2.16752782e-01 -4.54897463e-01 9.61104482e-02
2.96282858e-01 1.29325524e-01 -5.06822541e-02 9.22303081e-01
-9.43725556e-02 1.10786045e+00 -1.02022743e+00 -2.97633886e-01
1.22025728e-01 1.35405194e-02 -8.24725688e-01 1.33990264e-02
-2.91308105e-01 5.30794144e-01 -6.86266661e-01 8.77042949e-01
7.32089996e-01 -2.98640341e-01 2.88322955e-01 -2.83946127e-01
-3.88292730e-01 4.99572873e-01 -1.21931112e+00 1.75729406e+00
-6.93936348e-02 3.95511985e-01 1.78805932e-01 -1.04002929e+00
6.26490593e-01 -1.15550518e-01 3.66913140e-01 -1.12592661e+00
3.24770004e-01 6.75037861e-01 3.46877456e-01 -3.13542217e-01
1.00306797e+00 -5.75164139e-01 -3.32476914e-01 5.57820261e-01
3.94458413e-01 -7.99781024e-01 2.31128708e-01 1.88638821e-01
9.48787928e-01 1.33808091e-01 4.53024864e-01 -6.13766074e-01
4.11783934e-01 2.42327631e-01 5.83918452e-01 1.12123883e+00
-4.40537095e-01 2.97676384e-01 2.49095768e-01 -4.10595119e-01
-1.11250448e+00 -1.17438555e+00 -7.45862961e-01 7.58824229e-01
5.97766280e-01 -7.30865598e-01 -5.89985490e-01 -8.31784308e-02
-1.46542907e-01 8.29096019e-01 5.40580601e-02 -2.49733254e-01
-5.36605418e-01 -1.67377162e+00 4.30444688e-01 -1.32325232e-01
5.12703717e-01 -6.09616220e-01 -2.33484253e-01 5.49444795e-01
-5.17559536e-02 -1.18902004e+00 -1.30450040e-01 4.49192613e-01
-7.01929331e-01 -7.25163221e-01 -1.65830433e-01 -7.21189603e-02
3.05261165e-01 -1.05158821e-01 8.63982201e-01 -8.83537903e-02
-5.14633536e-01 3.73698175e-02 -1.15256719e-01 -1.70483708e-01
-6.24373019e-01 -2.32766420e-02 3.88727665e-01 -2.18173787e-01
4.60076541e-01 -5.76450706e-01 -5.50039291e-01 1.73428386e-01
-5.26871502e-01 -3.81671071e-01 5.69275677e-01 1.21143663e+00
6.44792438e-01 -5.39832290e-05 3.34398687e-01 -6.58832908e-01
4.99425322e-01 -8.13892707e-02 -1.10201895e+00 2.26448491e-01
-9.70724940e-01 6.79152966e-01 5.79297364e-01 1.68515027e-01
-7.64397979e-01 -4.15507816e-02 -4.27085191e-01 2.68687099e-01
4.71647352e-01 6.17394865e-01 1.34385660e-01 -8.98205996e-01
1.12323904e+00 2.77885079e-01 -1.33560404e-01 -1.24812856e-01
3.52211356e-01 6.67531848e-01 5.55178285e-01 -9.60066259e-01
6.25666976e-01 3.05904776e-01 5.25529027e-01 -1.01367652e+00
-9.01742101e-01 -4.97636676e-01 -3.31842154e-01 -1.02739841e-01
7.77923167e-01 -6.27354145e-01 -9.05197203e-01 2.05564663e-01
-8.49953651e-01 1.16698511e-01 -8.62811059e-02 7.63116956e-01
-4.22736108e-01 6.94345355e-01 -4.40990359e-01 -1.00357735e+00
-3.40435505e-01 -1.62867427e+00 7.18118489e-01 6.81870759e-01
1.59579977e-01 -6.61350369e-01 1.26806453e-01 5.30594945e-01
5.69750249e-01 -3.46427858e-02 8.27082634e-01 -2.34729886e-01
-1.05170739e+00 -3.55641574e-01 -4.89478670e-02 1.25429183e-01
-6.88709199e-01 -2.45298579e-01 -9.26735103e-01 -3.64812285e-01
1.74175814e-01 -6.30868495e-01 7.87670553e-01 3.76886010e-01
9.12860513e-01 3.28888834e-01 -2.59149998e-01 7.79912233e-01
1.48947680e+00 5.99461608e-02 5.87311506e-01 2.44906515e-01
2.63906598e-01 -1.59285009e-01 5.70750952e-01 3.57087880e-01
8.32641572e-02 9.72601175e-01 3.89181346e-01 5.64083755e-01
3.46927226e-01 2.92283837e-02 3.43626350e-01 1.00396097e+00
-1.44400463e-01 1.86899751e-01 -7.42673337e-01 -4.83312719e-02
-1.64408648e+00 -8.35733116e-01 -2.25785837e-01 2.29190445e+00
8.68041396e-01 2.23166779e-01 -2.38098606e-01 -1.92093506e-01
4.38168913e-01 7.43420273e-02 -5.48360825e-01 -7.59310782e-01
-9.67954174e-02 1.01669490e+00 7.52630472e-01 6.26038671e-01
-7.99896359e-01 9.93444026e-01 6.80349541e+00 1.34387600e+00
-8.61478090e-01 2.66879261e-01 1.37860551e-01 -8.32270309e-02
-2.10220039e-01 4.05610859e-01 -9.49844897e-01 4.28899199e-01
1.28306592e+00 -1.15183465e-01 1.08088577e+00 6.79033637e-01
-6.62987530e-02 -6.61542356e-01 -9.69266713e-01 1.14048934e+00
-2.14583457e-01 -1.68651402e+00 -4.29200798e-01 7.38799199e-02
8.71764481e-01 3.97487849e-01 5.76932244e-02 5.37460983e-01
-1.07880183e-01 -1.21508563e+00 7.00360656e-01 3.45813096e-01
6.63855135e-01 -7.33860850e-01 7.67210305e-01 4.10387963e-01
-5.99250376e-01 3.61621417e-02 -5.69361448e-01 -2.70872533e-01
3.54063898e-01 6.71967328e-01 -5.46141565e-01 6.33834898e-01
4.34838682e-01 2.86339298e-02 -3.21266413e-01 1.28749430e+00
-2.01067671e-01 6.05245352e-01 -8.00258875e-01 -5.56670368e-01
4.53144878e-01 -9.71321940e-01 6.88734829e-01 8.57966423e-01
2.18040794e-01 2.72880733e-01 -8.13041553e-02 1.30157626e+00
-2.72974130e-02 -3.95818837e-02 -1.40329614e-01 -4.62284356e-01
6.95036530e-01 1.12283814e+00 -6.11252010e-01 1.23158740e-02
-8.04209560e-02 5.64420402e-01 1.94166392e-01 2.18575016e-01
-8.44973862e-01 -5.82686424e-01 3.03332895e-01 -4.29240853e-01
1.22935519e-01 -5.50981522e-01 -5.75351655e-01 -1.32231545e+00
-1.16187587e-01 -8.24582577e-01 7.37527758e-02 -2.17169195e-01
-8.48164856e-01 2.46113405e-01 -2.21356645e-01 -7.74574161e-01
-8.25587362e-02 -8.83287013e-01 -5.51884949e-01 9.59297597e-01
-1.33331764e+00 -5.24303675e-01 9.43579748e-02 2.42372110e-01
-1.76958159e-01 -6.35210052e-02 8.33117306e-01 2.23324537e-01
-5.25968730e-01 6.12714767e-01 7.72040665e-01 -4.23679113e-01
3.34621906e-01 -1.23125982e+00 1.06507748e-01 8.70036364e-01
2.15946332e-01 7.19552994e-01 1.04748893e+00 -2.35846192e-01
-2.22036409e+00 -1.38408646e-01 5.07815778e-01 -3.05815667e-01
7.95634687e-01 -3.50156605e-01 -5.69793701e-01 1.23562880e-01
-1.63669631e-01 -9.15671065e-02 6.85139775e-01 3.49458247e-01
8.60902742e-02 -8.73322263e-02 -1.06346059e+00 3.69379669e-01
5.18333554e-01 -9.17074323e-01 -3.90362024e-01 7.23955929e-01
3.03945750e-01 -8.23247075e-01 -9.22323763e-01 3.85990947e-01
5.80838978e-01 -1.16224420e+00 1.04827285e+00 -2.14955807e-01
-2.82229017e-02 -4.23239619e-01 -4.14969534e-01 -1.09034836e+00
3.08980905e-02 -1.24776208e+00 -1.89455003e-01 3.66027355e-01
4.40855205e-01 -7.94486165e-01 7.85190225e-01 5.68527877e-01
-4.34265226e-01 -8.98656666e-01 -1.53220093e+00 -7.71998763e-01
1.58721745e-01 -6.48348868e-01 4.11625743e-01 3.92699480e-01
2.00185031e-01 4.05124009e-01 -2.48784840e-01 2.83881038e-01
9.59614277e-01 4.03547794e-01 5.08430541e-01 -5.05090475e-01
-8.20223212e-01 -4.68996316e-01 -4.00407672e-01 -1.12921464e+00
-2.00689360e-01 -1.08289361e+00 1.35319814e-01 -8.00418854e-01
1.11692026e-01 -5.69847286e-01 -1.64293036e-01 -2.46348426e-01
-2.66505361e-01 3.86281133e-01 -5.87290153e-02 6.15753233e-02
-7.41315842e-01 8.88025641e-01 1.04310834e+00 -7.89845884e-02
-2.26615548e-01 1.24644153e-01 -2.37166226e-01 3.18064094e-01
4.65253592e-01 -6.35511041e-01 2.78280020e-01 4.33293059e-02
7.03313708e-01 1.70303017e-01 6.51591718e-02 -1.05819666e+00
4.86024261e-01 -1.22021459e-01 -3.54952402e-02 -5.62772155e-01
6.37885630e-01 -5.14656976e-02 -6.73349202e-02 5.49576521e-01
2.01314360e-01 -2.42195874e-01 -5.62163927e-02 4.47383016e-01
-2.04290986e-01 -7.69486189e-01 1.06271577e+00 -7.67495707e-02
-6.40248597e-01 1.72377378e-01 -2.85322815e-01 1.39292404e-01
6.55233383e-01 4.55122143e-02 -3.49130988e-01 -8.67531926e-04
-5.83080947e-01 3.09570521e-01 3.63501787e-01 -4.54760730e-01
4.03154701e-01 -1.08338404e+00 -3.58329415e-01 1.39209270e-01
-8.18594173e-02 -2.88552016e-01 2.26123258e-01 1.24732637e+00
-1.11815262e+00 6.13810241e-01 4.85368781e-02 -7.64853001e-01
-6.14167690e-01 1.98176026e-01 5.89310706e-01 -3.58579576e-01
-2.95588493e-01 1.00482512e+00 -6.49038076e-01 -4.91492987e-01
-1.22492373e-01 -1.93875171e-02 5.19202948e-01 -5.06259620e-01
3.83693039e-01 6.11576021e-01 3.22426081e-01 -3.62366855e-01
-2.51776725e-01 3.21577817e-01 6.56158403e-02 -1.57093987e-01
1.23598802e+00 8.95335525e-02 -6.16680920e-01 5.25655374e-02
1.00175393e+00 6.77766129e-02 -7.08242953e-01 -3.35806221e-01
6.51166663e-02 -2.17545837e-01 6.15157127e-01 -6.98473096e-01
-1.36599094e-01 9.63174880e-01 6.01861596e-01 8.75951946e-02
6.84998214e-01 -1.94569916e-01 8.83613408e-01 8.62847865e-01
7.98771560e-01 -1.43999302e+00 -1.95699155e-01 6.25286758e-01
9.73857343e-02 -1.25681448e+00 7.07352221e-01 -9.27533060e-02
-1.21072166e-01 1.42818809e+00 1.82571203e-01 -1.11168399e-01
4.32815939e-01 -1.74200218e-02 -4.32083726e-01 -4.56611931e-01
-2.73210555e-01 -6.62383437e-02 3.61663610e-01 3.31643247e-03
5.80098107e-02 3.25508535e-01 -5.67961335e-01 3.94457102e-01
-4.15427566e-01 -3.26455623e-01 4.76075351e-01 8.82247388e-01
-7.42273092e-01 -1.40120804e+00 -5.00768244e-01 3.82292271e-01
-5.30545413e-01 -3.29842448e-01 9.88270789e-02 5.58381200e-01
-6.48329109e-02 8.18906069e-01 -5.03874362e-01 -3.02527755e-01
-1.67345852e-01 1.59183502e-01 1.07508790e+00 -3.86348784e-01
-4.85088289e-01 -7.59054869e-02 6.15577586e-02 -7.52687573e-01
-1.43270001e-01 -4.53131735e-01 -1.36965740e+00 -4.91014093e-01
-8.33682239e-01 6.82973623e-01 9.20376599e-01 1.38089001e+00
1.32109851e-01 2.09853575e-01 5.91505468e-01 -8.89530778e-01
-1.35153854e+00 -5.83845913e-01 -6.62140965e-01 -4.28860992e-01
2.15602070e-01 -7.21518159e-01 -2.52940863e-01 -9.62403119e-01] | [5.61993932723999, 4.892773151397705] |
1f624a87-e00e-493e-a65f-acb651986c83 | value-function-factorisation-with-hypergraph | 2112.06771 | null | https://arxiv.org/abs/2112.06771v2 | https://arxiv.org/pdf/2112.06771v2.pdf | Cooperative Multi-Agent Reinforcement Learning with Hypergraph Convolution | Recent years have witnessed the great success of multi-agent systems (MAS). Value decomposition, which decomposes joint action values into individual action values, has been an important work in MAS. However, many value decomposition methods ignore the coordination among different agents, leading to the notorious "lazy agents" problem. To enhance the coordination in MAS, this paper proposes HyperGraph CoNvolution MIX (HGCN-MIX), a method that incorporates hypergraph convolution with value decomposition. HGCN-MIX models agents as well as their relationships as a hypergraph, where agents are nodes and hyperedges among nodes indicate that the corresponding agents can coordinate to achieve larger rewards. Then, it trains a hypergraph that can capture the collaborative relationships among agents. Leveraging the learned hypergraph to consider how other agents' observations and actions affect their decisions, the agents in a MAS can better coordinate. We evaluate HGCN-MIX in the StarCraft II multi-agent challenge benchmark. The experimental results demonstrate that HGCN-MIX can train joint policies that outperform or achieve a similar level of performance as the current state-of-the-art techniques. We also observe that HGCN-MIX has an even more significant improvement of performance in the scenarios with a large amount of agents. Besides, we conduct additional analysis to emphasize that when the hypergraph learns more relationships, HGCN-MIX can train stronger joint policies. | ['Yu Liu', 'Xinwen Hou', 'Guoliang Fan', 'Bin Zhang', 'Chen Gong', 'Yunpeng Bai'] | 2021-12-09 | null | null | null | null | ['smac'] | ['playing-games'] | [-6.05302632e-01 1.39702931e-02 -2.77056277e-01 -1.64531861e-02
-8.26957822e-02 -4.83822405e-01 7.75769949e-01 1.06437206e-01
-5.88629484e-01 8.93605113e-01 6.43312693e-01 5.55364415e-02
-3.20944726e-01 -1.15114832e+00 -5.35110056e-01 -1.15129411e+00
-3.95565093e-01 8.03701997e-01 2.33685136e-01 -5.50512791e-01
-1.44100398e-01 3.14313807e-02 -9.19319928e-01 9.64951143e-02
6.78754449e-01 5.12632906e-01 -3.33837606e-02 6.09802425e-01
1.25512406e-01 1.45360136e+00 -9.23394859e-01 -3.56597602e-01
3.40945810e-01 -2.41139635e-01 -6.16071999e-01 -1.76538453e-02
-3.77434522e-01 -5.62819064e-01 -8.13521624e-01 1.09139287e+00
3.93565714e-01 4.64721560e-01 4.24646765e-01 -1.85091388e+00
-9.25101280e-01 1.57997835e+00 -7.91137993e-01 1.47783220e-01
-3.46699618e-02 8.01093102e-01 1.16815209e+00 7.82482475e-02
2.81017274e-01 1.66730535e+00 2.45032191e-01 5.10831058e-01
-1.12205732e+00 -6.20057821e-01 6.82003498e-01 2.41183162e-01
-7.10691869e-01 3.30133021e-01 4.71127510e-01 -1.66192129e-01
1.17694080e+00 1.04444167e-02 7.65349030e-01 9.48529184e-01
1.26160933e-02 1.08491182e+00 9.76733804e-01 1.80672318e-01
2.01164693e-01 -2.46568277e-01 8.49358141e-02 5.10738611e-01
2.72971988e-01 2.45150983e-01 -1.88494071e-01 -3.26582491e-01
1.09673584e+00 2.93880135e-01 -1.08371422e-01 -2.46428534e-01
-1.79557490e+00 8.58648062e-01 7.68227041e-01 7.35707879e-02
-8.61091018e-01 8.78194511e-01 5.09798586e-01 5.50663471e-01
1.25061065e-01 5.40171325e-01 -2.52872258e-01 -2.33726561e-01
4.80202883e-02 5.60395777e-01 9.93093371e-01 8.70550275e-01
6.58958256e-01 2.31904745e-01 -5.34357190e-01 4.41469103e-01
3.11822712e-01 5.13525069e-01 1.99877396e-01 -1.27216196e+00
5.87997139e-01 8.51275563e-01 2.51235485e-01 -8.68939817e-01
-5.53469181e-01 -6.51432872e-01 -9.52715456e-01 4.03091997e-01
2.72685289e-01 -7.83293307e-01 -6.19261384e-01 1.95679271e+00
3.48025531e-01 4.03837621e-01 5.60658932e-01 1.08415163e+00
7.42906868e-01 7.57823467e-01 9.66998637e-02 -1.27494201e-01
1.16619039e+00 -1.56905138e+00 -6.86680973e-01 -3.40508409e-02
3.59030664e-01 -1.66818276e-01 4.71938342e-01 1.39155239e-01
-1.04293084e+00 -1.54101625e-01 -7.86511898e-01 7.43673265e-01
-7.63721466e-02 -4.92689312e-01 1.07205951e+00 1.53016942e-02
-9.19480443e-01 4.70790595e-01 -1.02156532e+00 6.83517307e-02
3.44425082e-01 6.31838202e-01 -4.07388732e-02 1.59492359e-01
-1.31800044e+00 8.25270295e-01 4.57702100e-01 -1.53403595e-01
-1.64719033e+00 -5.44152856e-01 -8.17443967e-01 4.90135610e-01
1.12007368e+00 -7.78999329e-01 1.53200352e+00 -7.72802234e-01
-1.50806320e+00 -9.03414190e-02 5.95049977e-01 -6.47885501e-01
4.24688786e-01 1.03870623e-01 -9.78997126e-02 -9.98505112e-03
8.83178562e-02 4.40894753e-01 4.56876904e-01 -1.30927134e+00
-1.04810953e+00 4.99611162e-02 9.10625398e-01 5.85645497e-01
-2.99560934e-01 -2.42652521e-01 -3.40357989e-01 -5.34875870e-01
-5.46793759e-01 -1.02375066e+00 -6.84908450e-01 -6.34713948e-01
-2.22398594e-01 -7.42575645e-01 3.72549087e-01 3.59725440e-03
1.05870104e+00 -1.84538615e+00 7.03141928e-01 2.01227456e-01
8.91133964e-01 2.41282552e-01 -5.43997169e-01 8.17508399e-01
4.40377235e-01 -2.92970501e-02 6.43048510e-02 -1.27060771e-01
4.87969518e-01 5.69746435e-01 5.30504584e-02 3.36643666e-01
-2.32315436e-01 1.01522434e+00 -1.28503227e+00 -1.11191012e-01
1.68771863e-01 3.49712998e-01 -5.55136263e-01 3.57633173e-01
-6.45721018e-01 3.52721483e-01 -8.37828875e-01 3.71839046e-01
6.04857564e-01 -4.09395695e-01 3.76527637e-01 2.63925850e-01
-2.33169738e-02 -1.57233894e-01 -1.22523475e+00 1.19645321e+00
3.01512191e-04 3.20306093e-01 2.18831241e-01 -8.01376224e-01
6.46463752e-01 3.05997342e-01 9.06029046e-01 -6.16623521e-01
4.04297084e-01 -2.03367338e-01 4.84430134e-01 -2.32110783e-01
4.19743150e-01 2.31798500e-01 -1.78402722e-01 7.57290363e-01
-7.65677691e-02 2.10721895e-01 6.31365240e-01 6.81558073e-01
1.28519046e+00 -1.10712804e-01 7.75322318e-02 -2.27077100e-02
5.22998989e-01 1.78119745e-02 9.56681132e-01 9.86557245e-01
-4.46953624e-01 -1.74160436e-01 1.09190500e+00 -5.42895973e-01
-7.44783998e-01 -9.23277497e-01 7.29244947e-01 1.32740915e+00
5.39568067e-01 -4.36760187e-01 -7.31145084e-01 -8.24265897e-01
4.60057706e-01 3.88385922e-01 -8.32952917e-01 -1.38818890e-01
-5.29670477e-01 -1.08802450e+00 2.32762218e-01 7.08182275e-01
8.85351717e-01 -1.43656039e+00 -4.94955838e-01 4.07349765e-01
2.33378001e-02 -9.10294652e-01 -5.67822874e-01 -1.23703718e-01
-3.64322662e-01 -1.29147768e+00 -7.23092556e-01 -5.89242220e-01
6.63216650e-01 4.66778606e-01 1.11173022e+00 3.06863993e-01
4.58309025e-01 7.41907835e-01 -6.87521696e-01 -3.80101830e-01
-4.13079441e-01 7.61955827e-02 2.58121341e-01 1.23058055e-02
4.06614602e-01 -6.43705368e-01 -7.02118397e-01 4.70737845e-01
-6.24716759e-01 -2.17099544e-02 8.11656356e-01 9.10229325e-01
1.97589695e-01 3.26668978e-01 7.28682816e-01 -8.74005854e-01
1.15076697e+00 -7.76839137e-01 -7.12053537e-01 2.59168804e-01
-6.64663851e-01 1.46530867e-01 8.04389656e-01 -6.87063396e-01
-9.12434638e-01 -4.45657551e-01 4.04600918e-01 -3.72523278e-01
2.41334081e-01 7.16494679e-01 2.35115308e-02 1.70475960e-01
2.18175292e-01 1.99684948e-01 3.57453704e-01 -1.28515437e-01
3.62602919e-01 2.33051598e-01 4.07136291e-01 -7.88438261e-01
7.45706081e-01 3.17673206e-01 5.08072674e-02 -1.08951502e-01
-4.48740959e-01 -2.28688583e-01 1.93294168e-01 -2.84718156e-01
8.44992161e-01 -9.80596304e-01 -1.72836542e+00 6.99277341e-01
-1.03754365e+00 -6.90043569e-01 -1.20761439e-01 7.47514009e-01
-4.58854735e-01 1.02288604e-01 -9.74877715e-01 -6.03704870e-01
-1.69303581e-01 -1.41458464e+00 1.37361228e-01 9.11903381e-01
5.27429938e-01 -1.12707043e+00 2.82557458e-01 1.68895110e-01
5.90585530e-01 1.44398645e-01 6.39892578e-01 -7.99453139e-01
-9.61882591e-01 1.60623610e-01 -1.46576270e-01 1.59160972e-01
7.97823444e-02 -1.80925235e-01 -3.57658178e-01 -6.73267484e-01
-5.84707737e-01 -2.77925670e-01 7.50389576e-01 5.01136720e-01
5.85171044e-01 -5.28057575e-01 -3.63975644e-01 3.36832523e-01
1.19138169e+00 4.14080352e-01 3.86340499e-01 6.86006248e-01
7.62132287e-01 3.75177532e-01 5.08398056e-01 7.82423556e-01
1.13442326e+00 3.16402704e-01 1.24082625e+00 -1.80899397e-01
1.79633483e-01 -2.30141040e-02 5.79034328e-01 6.09933972e-01
-6.21337891e-01 -3.80544752e-01 -5.77794552e-01 4.19286698e-01
-2.70929098e+00 -1.12114048e+00 -1.12702467e-01 1.76808143e+00
6.28639340e-01 4.96317297e-02 6.39501333e-01 -6.06594205e-01
7.75810122e-01 3.64381701e-01 -9.92016196e-01 -1.38867915e-01
-1.71121866e-01 -5.28101087e-01 6.26850307e-01 4.40808207e-01
-9.44595218e-01 9.60777819e-01 5.92836285e+00 5.13893127e-01
-4.64092910e-01 1.19275548e-01 3.22917730e-01 -3.23055118e-01
-2.50577360e-01 4.85476926e-02 -5.71796119e-01 5.89096189e-01
2.95385420e-01 -5.46892107e-01 8.76039684e-01 8.62949848e-01
9.12487507e-02 6.94505945e-02 -8.65594447e-01 7.19015181e-01
-2.00561017e-01 -1.35407329e+00 -3.03746350e-02 2.96800792e-01
1.15331697e+00 2.40798041e-01 -3.74813080e-02 8.37553024e-01
1.68322086e+00 -9.31596696e-01 4.12474573e-01 4.72633302e-01
-1.70877114e-01 -1.05819333e+00 1.07560921e+00 4.49566960e-01
-1.40406859e+00 -4.15725440e-01 -4.93831247e-01 -3.82806093e-01
8.88267308e-02 -1.53632224e-01 -5.92771590e-01 6.97585166e-01
4.88749713e-01 8.20360839e-01 -2.58200586e-01 1.00558949e+00
-5.06233096e-01 3.46616328e-01 6.74778968e-02 -4.04378116e-01
8.71888697e-01 -3.93084079e-01 6.91474378e-01 7.83835053e-01
-2.44556725e-01 3.10501158e-01 8.42238843e-01 7.88881302e-01
-2.31214762e-01 -2.69042373e-01 -1.08629018e-01 -3.94924819e-01
6.31610930e-01 1.52345324e+00 -4.93543595e-01 -5.38889349e-01
-6.93342745e-01 5.43400943e-01 6.07109308e-01 5.91908634e-01
-1.00907362e+00 -1.84950575e-01 1.12905765e+00 -4.03233379e-01
1.37183100e-01 -3.10798734e-01 2.12062776e-01 -1.03162861e+00
-3.31433147e-01 -1.19143462e+00 6.13022149e-01 -4.52991813e-01
-1.47631764e+00 6.69854403e-01 -3.49125592e-03 -1.16212797e+00
-3.65216225e-01 -4.18247372e-01 -8.70341957e-01 5.81177592e-01
-1.51018953e+00 -1.27176547e+00 -2.80902982e-01 6.73533559e-01
2.92011917e-01 -7.50095546e-01 4.75963652e-01 -5.16607538e-02
-7.80067980e-01 3.34318995e-01 1.37892857e-01 4.57710922e-01
3.28381538e-01 -1.48802829e+00 1.79897919e-01 6.13138199e-01
-2.26549581e-01 4.51628506e-01 6.35367692e-01 -6.46574736e-01
-1.49999416e+00 -7.50497401e-01 1.57976313e-03 -2.89477974e-01
8.97897184e-01 6.24728352e-02 -6.67661428e-01 9.36659932e-01
8.53004038e-01 -1.96381524e-01 4.62684363e-01 2.01819584e-01
-1.70686483e-01 -8.26297477e-02 -8.07417631e-01 7.55395293e-01
8.42781723e-01 1.00257322e-01 -6.36110008e-01 2.03015164e-01
1.16099036e+00 -3.47218245e-01 -9.49064016e-01 7.75154531e-02
2.12599277e-01 -8.34663212e-01 7.97990978e-01 -9.10465539e-01
4.77818221e-01 -4.09498364e-01 1.96733177e-02 -2.08220625e+00
-7.07603991e-01 -8.04654300e-01 -2.39684656e-01 1.12615454e+00
2.55962968e-01 -9.39618170e-01 6.98969841e-01 7.72282004e-01
-8.77135620e-02 -6.80594087e-01 -4.69777554e-01 -7.14744925e-01
-1.21759735e-02 2.64433593e-01 1.23086739e+00 1.13178658e+00
2.59578139e-01 3.11879158e-01 -6.13616288e-01 3.46170515e-01
8.14106941e-01 9.71695483e-02 1.05837500e+00 -1.00777495e+00
-8.22746992e-01 -8.17115426e-01 -1.42681494e-01 -9.60367739e-01
2.92828411e-01 -8.28591466e-01 -2.95461535e-01 -2.06910086e+00
5.56076884e-01 -5.62700808e-01 -6.66397154e-01 7.50900090e-01
-3.35933775e-01 -3.20403576e-01 7.88927317e-01 2.54193306e-01
-1.10876012e+00 7.19562232e-01 1.75237191e+00 -4.97856081e-01
-3.98671955e-01 4.14597802e-03 -9.13422823e-01 5.29782355e-01
8.62310171e-01 -4.43774641e-01 -5.61402500e-01 -6.60603404e-01
3.32856178e-01 2.97945440e-01 2.89520025e-01 -6.90399110e-01
5.55221856e-01 -8.01239967e-01 -3.03733256e-02 -2.73962945e-01
7.21056238e-02 -7.76423573e-01 1.68629065e-01 7.10699737e-01
-3.94324929e-01 4.11310345e-01 -1.08159564e-01 6.35769606e-01
-1.65346846e-01 -1.28938898e-01 2.59100467e-01 -4.85070229e-01
-7.03854322e-01 6.48410499e-01 -4.10367429e-01 -6.44162819e-02
1.29315531e+00 3.12137187e-01 -8.66303444e-01 -6.95062876e-01
-6.02591097e-01 1.45528960e+00 1.77011058e-01 3.52032959e-01
5.51176190e-01 -1.40774190e+00 -1.01716828e+00 -5.35880476e-02
-2.49100104e-01 2.50430793e-01 4.80910927e-01 8.31017613e-01
-2.60388017e-01 4.31608642e-03 -5.58991849e-01 -2.67467022e-01
-1.08518815e+00 7.00272262e-01 3.96741658e-01 -6.45226777e-01
-5.77398479e-01 6.21709228e-01 3.46925288e-01 -3.83051634e-01
4.06661987e-01 -5.14090769e-02 -5.33855975e-01 6.38658330e-02
6.28695488e-01 5.91926396e-01 -8.88249874e-01 -3.38580102e-01
-1.59987539e-01 -8.90604313e-03 -4.86375183e-01 -2.37650797e-01
1.75111878e+00 2.98656374e-02 -2.91770190e-01 -1.23251475e-01
3.55143368e-01 -3.33187521e-01 -1.86241925e+00 -5.34325838e-01
-3.96204263e-01 -5.54357350e-01 2.94032637e-02 -8.47544074e-01
-1.56469345e+00 3.91401708e-01 -2.16964036e-02 5.83444595e-01
8.43281269e-01 2.00125314e-02 6.22317791e-01 5.00816107e-01
5.54988027e-01 -1.25160539e+00 5.02242804e-01 7.14475453e-01
8.53212833e-01 -1.21472013e+00 3.13224085e-03 -5.83376409e-03
-1.33546877e+00 8.32421720e-01 1.16410339e+00 -3.97295296e-01
2.95728505e-01 2.25176916e-01 2.42241286e-02 -2.87381083e-01
-1.12980866e+00 -5.61914027e-01 -3.44259292e-01 8.46302807e-01
-2.00156361e-01 5.15502393e-01 -1.17059439e-01 8.18306446e-01
1.90988645e-01 -2.49041975e-01 9.66863692e-01 8.22567523e-01
-3.62597823e-01 -1.04500175e+00 -2.40152061e-01 3.46501797e-01
-1.02306932e-01 4.61475216e-02 -2.78308600e-01 7.38497555e-01
-1.73960462e-01 1.09615529e+00 1.09812655e-01 -5.42738736e-01
3.87350380e-01 -7.25816727e-01 2.66663104e-01 -3.83366793e-01
-1.09955740e+00 1.49970338e-01 -2.72407494e-02 -5.49705565e-01
-5.31465888e-01 -3.14310551e-01 -1.77350843e+00 -9.01540816e-01
9.08218846e-02 3.11495602e-01 8.74036625e-02 5.79470217e-01
2.52587169e-01 1.11171412e+00 7.84579575e-01 -6.18512928e-01
-9.72602010e-01 -1.06018734e+00 -6.61035776e-01 4.66131389e-01
2.87655801e-01 -1.00205195e+00 -2.64848948e-01 -6.27918541e-01] | [3.7426846027374268, 1.928741216659546] |
0df72ea1-3ff7-4484-9f46-e9c55324828e | semi-supervised-hypothesis-transfer-for | 2107.06735 | null | https://arxiv.org/abs/2107.06735v1 | https://arxiv.org/pdf/2107.06735v1.pdf | Semi-Supervised Hypothesis Transfer for Source-Free Domain Adaptation | Domain Adaptation has been widely used to deal with the distribution shift in vision, language, multimedia etc. Most domain adaptation methods learn domain-invariant features with data from both domains available. However, such a strategy might be infeasible in practice when source data are unavailable due to data-privacy concerns. To address this issue, we propose a novel adaptation method via hypothesis transfer without accessing source data at adaptation stage. In order to fully use the limited target data, a semi-supervised mutual enhancement method is proposed, in which entropy minimization and augmented label propagation are used iteratively to perform inter-domain and intra-domain alignments. Compared with state-of-the-art methods, the experimental results on three public datasets demonstrate that our method gets up to 19.9% improvements on semi-supervised adaptation tasks. | ['Xifeng Yan', 'Sheng Zhou', 'Zhen Zhang', 'Jun Wen', 'Lixian Lu', 'Jiajun Bu', 'Ning Ma'] | 2021-07-14 | null | null | null | null | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 4.68096524e-01 -9.36041623e-02 -2.47216910e-01 -6.52077317e-01
-7.99469471e-01 -4.41177845e-01 6.34364784e-01 5.27953170e-02
-7.50272930e-01 1.15951419e+00 5.10495305e-02 6.33556843e-02
2.53080517e-01 -4.20045853e-01 -4.93158281e-01 -6.56214535e-01
3.93725932e-01 3.59164178e-01 3.84343863e-01 -4.88459505e-02
1.12734519e-01 6.47671297e-02 -1.16108203e+00 1.02949016e-01
1.18545711e+00 7.93549538e-01 2.39622459e-01 3.19298394e-02
-3.65340084e-01 2.43743464e-01 -2.88588285e-01 -4.78583336e-01
3.77152383e-01 -4.25174296e-01 -9.09128726e-01 3.77750963e-01
2.18527898e-01 -3.05931807e-01 -9.57048610e-02 1.27453864e+00
7.23749578e-01 2.59422779e-01 5.96965551e-01 -1.22791457e+00
-6.90032482e-01 1.23746507e-01 -8.54862392e-01 6.90637082e-02
3.02366585e-01 -2.02873901e-01 5.25288939e-01 -9.36514139e-01
7.46773064e-01 7.98064590e-01 4.00688589e-01 7.88495958e-01
-1.10193598e+00 -1.05289316e+00 1.94666669e-01 3.43007773e-01
-1.44807124e+00 -4.77317095e-01 9.14008975e-01 -2.19085410e-01
4.88684505e-01 -2.13119552e-01 1.73714116e-01 1.29880428e+00
-2.56891727e-01 7.51943648e-01 1.37799048e+00 -6.45781636e-01
1.29309639e-01 8.80109727e-01 -2.03645170e-01 3.46827865e-01
4.31694128e-02 -3.31939459e-02 -6.08726799e-01 -3.46693516e-01
4.89669979e-01 -5.40355854e-02 -2.54799634e-01 -6.94333076e-01
-1.28537536e+00 7.64006972e-01 2.46995509e-01 1.39646769e-01
-2.34373763e-01 -8.10413599e-01 5.78769445e-01 4.22385752e-01
7.68047810e-01 6.18707240e-02 -8.09205413e-01 1.60685852e-01
-7.44026482e-01 1.85584854e-02 7.09768474e-01 1.27549469e+00
9.44374621e-01 -3.74150187e-01 2.88048148e-01 1.15898943e+00
2.77535588e-01 5.91575563e-01 8.50881159e-01 -4.98873681e-01
6.99359000e-01 5.46600938e-01 1.47801071e-01 -5.32572091e-01
-2.43927330e-01 -2.43568510e-01 -1.03483784e+00 -7.29721636e-02
3.96756858e-01 -2.78778940e-01 -9.23291326e-01 1.67614841e+00
7.32556045e-01 2.88828492e-01 2.98205495e-01 9.03540969e-01
6.71410561e-01 5.09013176e-01 3.44930530e-01 -4.30604249e-01
1.22398674e+00 -1.01752067e+00 -8.62889826e-01 -3.52448255e-01
3.21425915e-01 -9.07592595e-01 9.63368952e-01 2.87510753e-01
-6.04264915e-01 -6.61235631e-01 -9.86049592e-01 1.27762958e-01
-4.86852616e-01 1.93793513e-02 1.82229459e-01 5.24107337e-01
-4.92178500e-01 1.42087311e-01 -6.77256823e-01 -7.02211678e-01
6.71533644e-01 3.83821487e-01 -8.38233352e-01 -2.39744902e-01
-1.37180841e+00 8.99047673e-01 7.17655361e-01 -4.16089535e-01
-5.21553159e-01 -5.16784370e-01 -6.81366146e-01 -3.27933639e-01
4.81186002e-01 -6.31030619e-01 1.11722493e+00 -1.39625204e+00
-1.67935157e+00 1.09013903e+00 -2.08475664e-01 -3.51731092e-01
5.77735782e-01 -2.94312209e-01 -6.96527779e-01 -5.96918017e-02
2.62093339e-02 5.90858996e-01 9.35215056e-01 -1.04481256e+00
-9.28444445e-01 -5.32030106e-01 -3.42173547e-01 5.80873787e-01
-7.61820197e-01 7.54810050e-02 -6.10477984e-01 -5.99876344e-01
7.99628869e-02 -1.03551340e+00 -2.97937632e-01 5.13080992e-02
-1.52087063e-01 -6.19953647e-02 9.95016456e-01 -9.15307105e-01
9.84179974e-01 -2.12362194e+00 4.70483415e-02 1.68059871e-01
-1.28308564e-01 6.56496108e-01 -7.35923462e-03 2.37717107e-01
-1.87241714e-02 -1.85683638e-01 -7.17130303e-01 -4.09756809e-01
-2.24315092e-01 8.08564499e-02 -1.16517298e-01 4.11065668e-01
1.46077797e-01 2.97151566e-01 -7.71903157e-01 -8.47321033e-01
1.77570432e-01 5.24583519e-01 -3.13974857e-01 5.22681832e-01
1.17958672e-02 8.61282229e-01 -7.04136372e-01 6.11420691e-01
1.06966102e+00 -2.21382350e-01 2.08480984e-01 4.92313616e-02
1.96815193e-01 -8.90109465e-02 -1.21176350e+00 2.16453457e+00
-2.71078974e-01 2.52799332e-01 -2.55406722e-02 -1.20518935e+00
1.03296149e+00 5.06999433e-01 3.93455803e-01 -5.81534147e-01
-1.38814554e-01 2.64993191e-01 -3.27543080e-01 -5.05997002e-01
2.51591593e-01 -2.85000890e-01 -3.19067538e-02 1.05107874e-01
2.50145257e-01 2.06164643e-01 -2.46149853e-01 -4.24803197e-02
7.69215584e-01 3.57447743e-01 7.21002340e-01 -2.09018216e-02
9.68753576e-01 1.21354192e-01 9.45103168e-01 2.73562104e-01
-6.04749382e-01 6.77098215e-01 -1.71233833e-01 -2.44173512e-01
-8.75323474e-01 -8.85987937e-01 -7.66547322e-02 1.11296189e+00
1.70373946e-01 -4.06791680e-02 -7.67206252e-01 -1.28881860e+00
-2.17104793e-01 4.77438778e-01 -3.63029569e-01 -2.49255642e-01
-2.94341624e-01 -6.83458507e-01 2.63351142e-01 2.61288136e-01
1.12466455e+00 -8.43706608e-01 -2.67918617e-01 2.82449722e-01
-4.88812566e-01 -1.31243050e+00 -5.81555784e-01 -5.19441813e-02
-8.57172608e-01 -7.89701164e-01 -9.88510966e-01 -9.32462156e-01
9.39371824e-01 2.27131888e-01 7.22209692e-01 -4.56495672e-01
6.63833274e-03 2.29789928e-01 -5.78601897e-01 -4.50044632e-01
-3.33280921e-01 4.45981711e-01 1.88828588e-01 3.14999223e-01
8.75517786e-01 -5.77850938e-01 -5.52648246e-01 4.79640454e-01
-9.25763607e-01 -8.62867311e-02 8.09505463e-01 1.11148262e+00
7.37427831e-01 2.45424230e-02 8.06460381e-01 -1.31305027e+00
3.71865422e-01 -6.68403089e-01 -5.70698440e-01 4.53508556e-01
-9.29592431e-01 -5.06863222e-02 6.59431756e-01 -6.82275057e-01
-1.66177344e+00 5.55667639e-01 1.46342114e-01 -3.09548914e-01
-5.38183868e-01 4.45185393e-01 -6.41576707e-01 -1.94198519e-01
6.95398510e-01 4.78476852e-01 -1.43193908e-03 -5.50999641e-01
1.75726071e-01 1.17134237e+00 4.34767336e-01 -3.07763755e-01
9.28168833e-01 3.83418709e-01 -2.89411098e-01 -5.66199720e-01
-7.00924397e-01 -7.09165394e-01 -9.36463177e-01 2.30632335e-01
7.74945259e-01 -1.29631102e+00 7.62271285e-02 4.35472846e-01
-9.45425212e-01 1.05011985e-01 -4.42102663e-02 8.14805567e-01
-5.44008911e-01 6.07785046e-01 4.23079059e-02 -4.41615760e-01
-3.35833490e-01 -9.65655148e-01 6.41439795e-01 3.82636607e-01
1.48109600e-01 -9.86022353e-01 2.57682323e-01 5.58094263e-01
4.03569281e-01 1.09164082e-01 4.08663452e-01 -1.19938743e+00
-2.44775876e-01 -1.84702128e-01 -2.29359075e-01 5.83828211e-01
4.61415648e-01 -6.01206064e-01 -9.85798597e-01 -3.71096164e-01
4.52849790e-02 -3.83423835e-01 5.37436366e-01 -4.68070619e-03
1.05794132e+00 -2.40720764e-01 -5.24278283e-01 5.71829796e-01
1.39395678e+00 1.71029195e-01 4.32245016e-01 4.34801370e-01
6.47146821e-01 7.01234162e-01 1.17660749e+00 6.60892010e-01
5.82496226e-01 7.67798424e-01 -4.30574529e-02 -4.78309616e-02
-3.40634771e-02 -3.26631546e-01 3.96230638e-01 6.26847148e-01
1.33266523e-01 -1.37188554e-01 -8.79905641e-01 7.19714224e-01
-1.83977604e+00 -6.48877919e-01 2.18713611e-01 2.57798481e+00
1.19020772e+00 -7.56807476e-02 1.52009785e-01 -3.58492702e-01
9.26572025e-01 -1.42463833e-01 -8.96808684e-01 -1.45507921e-02
-8.41092765e-02 -9.53142121e-02 7.16945589e-01 2.67727733e-01
-1.49518895e+00 1.07665896e+00 5.71219730e+00 9.06076193e-01
-1.05204558e+00 5.24985671e-01 3.58330250e-01 1.82451680e-01
1.89926848e-02 -1.64672285e-02 -7.74310112e-01 6.08843744e-01
8.42214823e-01 -2.01938033e-01 1.90608844e-01 8.46073627e-01
-1.67051524e-01 5.04632555e-02 -8.38060200e-01 1.03140128e+00
1.52041554e-01 -6.82512641e-01 -1.88619748e-01 1.40265927e-01
7.73197174e-01 2.75310669e-02 -5.79675473e-02 2.67913014e-01
2.09046781e-01 -4.90287721e-01 4.37939502e-02 2.76990771e-01
8.61829400e-01 -8.54304910e-01 7.26534367e-01 5.37865281e-01
-8.78757358e-01 1.45838410e-01 -5.59481502e-01 3.33201140e-01
1.00370586e-01 5.86449206e-01 -9.98437345e-01 7.38148868e-01
7.71354556e-01 7.44916618e-01 -4.12986845e-01 9.12492335e-01
-1.49074852e-01 6.62826419e-01 -3.53840142e-01 3.71891946e-01
-1.60670310e-01 -1.65600911e-01 5.69453537e-01 1.20008004e+00
3.05904120e-01 1.07652202e-01 2.97279775e-01 3.82107705e-01
-1.17941424e-01 4.65602100e-01 -7.90380418e-01 1.22483224e-01
5.97971022e-01 1.15532196e+00 -3.24262321e-01 -3.63082439e-01
-7.66098738e-01 1.57612133e+00 3.09715331e-01 4.16932195e-01
-8.06231022e-01 -5.00273883e-01 5.46430290e-01 -1.00564805e-03
3.36468816e-01 -9.52168107e-02 -7.10083470e-02 -1.40687704e+00
1.74610689e-01 -1.01576209e+00 7.26923585e-01 -4.73203659e-01
-1.57279456e+00 5.64918756e-01 1.02138564e-01 -1.74667799e+00
-3.03291380e-01 -1.64704174e-01 -1.74010798e-01 8.22062254e-01
-1.94860125e+00 -1.40472806e+00 -1.75866991e-01 1.06733406e+00
5.37836194e-01 -6.41462386e-01 1.05382884e+00 4.89612311e-01
-3.87195170e-01 1.04220414e+00 4.03343260e-01 -9.86694079e-03
1.52775657e+00 -1.08690894e+00 4.79020774e-02 8.24707985e-01
-1.70120485e-02 4.52008456e-01 5.78970671e-01 -5.89793801e-01
-9.38714445e-01 -1.11031997e+00 9.59325194e-01 -1.74893901e-01
3.00669283e-01 -2.15204254e-01 -1.23194528e+00 6.79310918e-01
4.70426559e-01 2.72312760e-01 1.15963256e+00 -2.55303197e-02
-4.91164029e-01 -3.33388627e-01 -1.63234770e+00 2.26920992e-01
1.04411125e+00 -3.79153579e-01 -7.21873760e-01 4.76852328e-01
6.00059628e-01 -4.52899724e-01 -1.02258062e+00 4.57995027e-01
3.25693011e-01 -5.67898273e-01 8.90165150e-01 -5.54908514e-01
7.89322183e-02 -5.20806193e-01 -2.38273814e-01 -1.43423223e+00
-1.99821502e-01 -5.80680132e-01 6.01574630e-02 1.69116569e+00
3.55816871e-01 -8.74288261e-01 6.67932272e-01 8.21410537e-01
3.24893534e-01 -7.33248889e-02 -1.01986814e+00 -8.19053054e-01
-3.06319036e-02 1.04348995e-01 6.83232009e-01 1.34900737e+00
3.47382352e-02 3.50507349e-01 -5.70790112e-01 3.63084435e-01
7.70009756e-01 5.44294566e-02 8.38741302e-01 -1.17776072e+00
-7.22955465e-02 1.61339357e-01 -3.04141372e-01 -9.08910334e-01
4.17113930e-01 -7.03713357e-01 5.82021438e-02 -1.13359118e+00
3.19268376e-01 -5.27242541e-01 -7.26370931e-01 4.68826115e-01
-3.14547479e-01 1.43028334e-01 -2.25109980e-02 3.72098446e-01
-7.13697970e-01 6.87495530e-01 1.02381361e+00 3.32276896e-02
-1.98788360e-01 9.89943370e-02 -7.79342532e-01 7.02309310e-01
1.04153347e+00 -6.88389301e-01 -5.98051667e-01 -3.79801303e-01
-3.62833530e-01 -2.47006431e-01 -3.54785211e-02 -1.01011801e+00
3.08609754e-01 -3.85475546e-01 4.34048921e-01 -4.27790284e-01
9.69977304e-02 -1.29405093e+00 6.76326603e-02 1.40995607e-01
-3.46239805e-01 -3.21735054e-01 1.56621218e-01 8.66230845e-01
-5.81349730e-01 -8.55934694e-02 8.75347853e-01 -3.44982147e-02
-1.02589166e+00 4.47244912e-01 2.92119868e-02 -1.49298841e-02
1.22710896e+00 -5.11791036e-02 -4.93951887e-02 -2.94309735e-01
-6.76713765e-01 2.38444552e-01 5.90167761e-01 4.88429010e-01
5.59253395e-01 -1.60506165e+00 -8.84905875e-01 3.20715994e-01
6.14628315e-01 1.93326604e-02 4.15494084e-01 5.50470233e-01
5.29378168e-02 1.14276648e-01 -4.33019996e-01 -4.24120039e-01
-1.42003715e+00 6.63395405e-01 1.95497703e-02 -4.06519294e-01
-2.21082702e-01 7.92416215e-01 2.42102161e-01 -7.78122187e-01
1.05433211e-01 3.03804040e-01 -2.02327326e-01 -8.89403454e-04
6.15702271e-01 1.57729343e-01 -9.64810252e-02 -6.84628308e-01
-5.76813579e-01 6.63311601e-01 -4.75618720e-01 -2.37213165e-01
1.07441974e+00 -7.13279188e-01 2.00874135e-01 2.27406323e-01
1.26515543e+00 -1.44128501e-01 -1.44976389e+00 -1.06411529e+00
-1.87105257e-02 -5.87956905e-01 -1.46857142e-01 -9.89095449e-01
-9.73139405e-01 9.04329002e-01 1.01720417e+00 -3.17385614e-01
1.58707309e+00 -2.18130827e-01 1.01356280e+00 3.66669804e-01
3.89002353e-01 -1.54028893e+00 -2.78907835e-01 2.49989286e-01
6.97622538e-01 -1.95769358e+00 1.90642819e-01 -4.43338484e-01
-1.12397122e+00 7.60873854e-01 9.79651809e-01 2.73989886e-01
7.73347855e-01 -3.26003671e-01 2.23158032e-01 4.38142866e-01
-4.50420797e-01 -1.88405037e-01 4.18253124e-01 9.45192516e-01
2.34781012e-01 -2.47332025e-02 -4.70227152e-01 5.33181131e-01
3.39299798e-01 2.89006948e-01 2.09867865e-01 1.01973176e+00
-2.77447790e-01 -1.68167317e+00 -3.03557605e-01 1.19532168e-01
-5.55969834e-01 1.00130863e-01 -3.05233628e-01 6.40620232e-01
2.70681113e-01 9.65376556e-01 -3.05867523e-01 -3.37775528e-01
2.55393535e-01 3.02653849e-01 2.33648002e-01 -5.85680306e-01
-2.50619024e-01 8.46283808e-02 -9.68229324e-02 -3.42812687e-01
-8.51327062e-01 -9.35800791e-01 -1.18050635e+00 9.57567170e-02
-3.57931942e-01 4.54026684e-02 6.63262665e-01 8.36955190e-01
6.44310057e-01 1.22226784e-02 7.80326784e-01 -2.53581882e-01
-5.79639494e-01 -9.95995104e-01 -3.86044651e-01 5.76200664e-01
2.93421537e-01 -5.75803995e-01 -1.15097523e-01 5.16899824e-01] | [10.352641105651855, 3.053542137145996] |
7df42f46-1ed2-4aa0-9b8e-7cebc6ee28f9 | smoothed-bernstein-online-aggregation-for-day | 2107.06268 | null | https://arxiv.org/abs/2107.06268v1 | https://arxiv.org/pdf/2107.06268v1.pdf | Smoothed Bernstein Online Aggregation for Day-Ahead Electricity Demand Forecasting | We present a winning method of the IEEE DataPort Competition on Day-Ahead Electricity Demand Forecasting: Post-COVID Paradigm. The day-ahead load forecasting approach is based on online forecast combination of multiple point prediction models. It contains four steps: i) data cleaning and preprocessing, ii) a holiday adjustment procedure, iii) training of individual forecasting models, iv) forecast combination by smoothed Bernstein Online Aggregation (BOA). The approach is flexible and can quickly adopt to new energy system situations as they occurred during and after COVID-19 shutdowns. The pool of individual prediction models ranges from rather simple time series models to sophisticated models like generalized additive models (GAMs) and high-dimensional linear models estimated by lasso. They incorporate autoregressive, calendar and weather effects efficiently. All steps contain novel concepts that contribute to the excellent forecasting performance of the proposed method. This holds particularly for the holiday adjustment procedure and the fully adaptive smoothed BOA approach. | ['Florian Ziel'] | 2021-07-13 | null | null | null | null | ['novel-concepts'] | ['reasoning'] | [-1.60779372e-01 -2.41693303e-01 -6.43107668e-02 -7.80670404e-01
-7.40462959e-01 -5.67273259e-01 8.86013627e-01 1.44983470e-01
2.50627816e-01 9.12314653e-01 3.15417200e-01 -5.79233587e-01
-6.08282387e-01 -6.96951866e-01 -3.16195011e-01 -9.04093266e-01
-7.28267193e-01 7.96371937e-01 -5.23464978e-01 -2.63968259e-01
6.63630739e-02 5.77774227e-01 -1.38598597e+00 -6.78392947e-02
1.23200190e+00 1.18804526e+00 7.02706873e-02 5.54077268e-01
-1.28533438e-01 4.96120214e-01 -2.44200766e-01 -8.42272043e-02
6.93543553e-01 3.19554470e-02 -7.37507790e-02 -2.72380352e-01
-4.26890194e-01 -4.19353485e-01 3.82447302e-01 5.51727831e-01
6.12829924e-01 4.16265279e-01 7.06107914e-01 -1.73936725e+00
-5.91672063e-01 9.44598317e-01 -5.93635380e-01 1.33314699e-01
3.14492822e-01 -3.25413235e-02 7.25449920e-01 -1.14444602e+00
-2.56720763e-02 9.11703587e-01 9.47838604e-01 -1.82329059e-01
-1.58182728e+00 -5.49451768e-01 2.56796986e-01 3.43376040e-01
-1.43579626e+00 -2.89194763e-01 7.32315063e-01 -7.58058131e-01
1.42654610e+00 7.72637486e-01 6.24122620e-01 5.18325627e-01
2.59568721e-01 5.60947597e-01 1.05316138e+00 -3.75112027e-01
5.48048198e-01 1.00214079e-01 5.76552927e-01 -3.23452055e-01
-1.34802073e-01 3.59383613e-01 -5.53340241e-02 -6.55070424e-01
3.67292687e-02 3.64182293e-01 -2.14226302e-02 3.67249139e-02
-9.60891008e-01 7.90362298e-01 -6.03924952e-02 1.42637044e-01
-1.03529096e+00 -3.19190234e-01 5.45477092e-01 3.35132599e-01
1.02262282e+00 3.01687699e-02 -9.55305159e-01 -6.84985593e-02
-1.58444989e+00 4.15240049e-01 9.06235993e-01 1.01344192e+00
6.48017704e-01 7.20480323e-01 -2.29816824e-01 6.48568511e-01
2.84186244e-01 7.51043022e-01 4.68178600e-01 -4.71888751e-01
2.38761976e-01 2.64110863e-02 5.38471699e-01 -8.48325789e-01
-9.19976234e-01 -3.48595053e-01 -1.18812644e+00 1.86135188e-01
1.05060205e-01 -6.13445461e-01 -4.66659784e-01 1.41453075e+00
3.28659773e-01 4.50880826e-01 -3.31951603e-02 3.37819189e-01
2.86399841e-01 1.13442576e+00 1.01418965e-01 -1.03980064e+00
1.02641737e+00 -8.92644584e-01 -1.06229532e+00 3.79447550e-01
7.49957502e-01 -7.59004712e-01 3.00956100e-01 6.49053991e-01
-1.11537921e+00 -5.32306135e-01 -6.43432438e-01 2.60900706e-01
-8.02875817e-01 8.67656618e-02 4.72994745e-01 7.23935962e-01
-1.09393406e+00 4.97400552e-01 -7.03558028e-01 1.01008832e-01
-1.54500470e-01 2.89449543e-01 2.26661801e-01 4.01551217e-01
-1.26735771e+00 9.85688806e-01 3.42486024e-01 4.72470969e-01
-4.01454687e-01 -1.21994317e+00 -6.65349364e-01 2.61245370e-01
-3.33244950e-02 -5.74646056e-01 9.01828527e-01 -9.20820296e-01
-1.61539781e+00 5.11449389e-02 -5.11828899e-01 -9.07791078e-01
4.50238675e-01 -1.38349146e-01 -1.06135511e+00 -9.06265557e-01
-1.95059091e-01 -2.36363709e-01 8.72679174e-01 -8.61667037e-01
-5.27978241e-01 -4.72370416e-01 -8.05060744e-01 3.44744156e-04
2.29081228e-01 1.14931718e-01 6.31964386e-01 -9.84543800e-01
-1.54276073e-01 -7.26317585e-01 -4.46051925e-01 -9.59778666e-01
-4.21423048e-01 -6.02712929e-01 9.22062397e-01 -1.33014941e+00
1.66166472e+00 -1.86709738e+00 -3.31500471e-01 6.96584821e-01
-4.86813515e-01 -8.33201706e-02 2.88242280e-01 9.53711867e-01
-6.42362773e-01 -2.45983470e-02 -1.07716084e-01 -5.75618684e-01
5.61156452e-01 2.35808626e-01 -9.79243517e-01 4.20290589e-01
-2.03422621e-01 9.26936507e-01 -5.27360559e-01 3.18188131e-01
4.32732403e-01 2.38342717e-01 -1.60233304e-01 8.51115435e-02
-1.39005929e-01 3.35048527e-01 -4.84788939e-02 4.94585693e-01
1.38508284e+00 1.14918649e-01 -3.14455964e-02 -8.22385997e-02
-8.01526427e-01 1.99125055e-02 -1.42212129e+00 1.04204619e+00
-4.35931116e-01 1.63221255e-01 -6.89582750e-02 -1.19957221e+00
1.12695682e+00 4.80639935e-01 8.90211523e-01 -2.97416985e-01
-3.51758242e-01 2.30155677e-01 -5.28962314e-01 -3.26310456e-01
4.99950558e-01 -4.78426144e-02 3.34874690e-02 5.68220794e-01
-2.49901444e-01 -1.23974122e-01 2.23620817e-01 -2.04647526e-01
2.96273649e-01 1.12613901e-01 4.84871387e-01 -7.90160298e-01
4.93460953e-01 -2.40686223e-01 8.61381054e-01 5.93784869e-01
2.26053864e-01 1.67618945e-01 2.24059403e-01 -8.92988563e-01
-1.03355694e+00 -8.50983739e-01 -4.17415917e-01 1.09042168e+00
-6.51532650e-01 -1.73632681e-01 -2.09293187e-01 -1.42371044e-01
3.74537766e-01 1.64756000e+00 -5.44545293e-01 4.78262872e-01
-1.75475270e-01 -1.26541615e+00 -1.46307066e-01 5.44919074e-01
-4.02133800e-02 -7.36701488e-01 -5.85769638e-02 5.74198723e-01
1.56655177e-01 -6.18056059e-01 -4.50396895e-01 3.35206777e-01
-7.14970112e-01 -3.26321751e-01 -8.05253625e-01 -2.25604072e-01
3.04450512e-01 -1.79597259e-01 1.32947350e+00 -4.96206492e-01
1.56204686e-01 2.49749631e-01 -1.94729373e-01 -9.26739931e-01
-4.91068475e-02 -3.07270944e-01 5.02014816e-01 1.86415300e-01
4.39651519e-01 -9.28394377e-01 -5.48098505e-01 3.76966327e-01
-3.85434479e-01 3.09305582e-02 7.28050396e-02 8.61245751e-01
7.11552382e-01 1.95608407e-01 1.22502840e+00 -7.62123883e-01
6.78020656e-01 -9.17956293e-01 -1.18087888e+00 4.78006333e-01
-1.22065294e+00 -4.71224338e-01 8.79153788e-01 8.06190073e-02
-1.20347345e+00 1.67495370e-01 -1.07459843e-01 -7.01793134e-02
-5.26799671e-02 7.84029722e-01 2.65576810e-01 2.38774061e-01
2.94325620e-01 4.74544525e-01 -1.86911955e-01 -6.86731994e-01
3.65921259e-01 5.83276391e-01 2.54341274e-01 -1.30139932e-01
8.15245807e-01 2.03743801e-01 -3.04874200e-02 -7.06262708e-01
-3.53432000e-01 -4.46555495e-01 -6.71637714e-01 -1.11325674e-01
7.71273732e-01 -1.12979746e+00 -7.55111754e-01 7.07681775e-01
-8.97529125e-01 -3.92887205e-01 -5.37005723e-01 4.91046637e-01
-4.76615816e-01 1.25523165e-01 -3.30951989e-01 -1.57593942e+00
-7.34286964e-01 -7.85235047e-01 6.32168651e-01 9.69634131e-02
-5.50922714e-02 -1.41904199e+00 2.30531961e-01 -1.22236095e-01
8.77489805e-01 5.16845763e-01 8.14047515e-01 -1.00795102e+00
1.91993155e-02 -5.04805565e-01 1.15529671e-01 3.67479205e-01
7.75947347e-02 2.68194646e-01 -9.25217152e-01 -4.52910900e-01
7.00659398e-03 2.23844662e-01 2.53883839e-01 9.30780470e-01
1.22088075e+00 -6.42050326e-01 -2.50043720e-01 4.21097904e-01
1.37564611e+00 4.03498203e-01 5.01318216e-01 -6.93312585e-02
2.41231337e-01 2.69090414e-01 4.05647635e-01 1.09269249e+00
7.25876331e-01 6.89438403e-01 1.75242633e-01 -1.03347331e-01
7.11727977e-01 1.17764898e-01 4.23459023e-01 1.18379283e+00
-1.89269558e-01 -1.30429253e-01 -9.71788704e-01 5.29842436e-01
-2.06807351e+00 -1.23837388e+00 -6.75180256e-01 2.35124493e+00
2.64914483e-01 -1.38372913e-01 3.82309973e-01 1.20692052e-01
3.39256048e-01 -5.82848261e-05 -5.42240739e-01 -1.12568069e+00
-3.87435019e-01 -8.34053010e-02 5.69981575e-01 5.96042454e-01
-9.60675478e-01 -9.00304839e-02 7.20318604e+00 8.41018200e-01
-6.60400867e-01 6.89336807e-02 9.71071064e-01 2.72528052e-01
-4.63847011e-01 -1.97942704e-02 -8.33607316e-01 9.46782351e-01
1.40563464e+00 -7.24387705e-01 7.23159969e-01 9.83171046e-01
1.07282710e+00 -1.65766075e-01 -8.85306358e-01 7.22151160e-01
-1.43314943e-01 -1.18985856e+00 -3.28657120e-01 1.25253946e-01
1.42204881e+00 4.31646556e-01 -1.51855841e-01 5.92375636e-01
4.01890337e-01 -8.18259597e-01 3.29716206e-01 1.34426761e+00
2.57331312e-01 -8.89337480e-01 8.22259843e-01 7.76496291e-01
-1.20966387e+00 -3.77765387e-01 -2.06992924e-01 -3.00594091e-01
7.17795491e-01 1.37241673e+00 -5.79682410e-01 1.06740320e+00
6.26833022e-01 7.09343195e-01 7.97876343e-02 9.72300291e-01
2.60350764e-01 7.35014439e-01 -8.87488365e-01 3.21294934e-01
-1.59761384e-02 -6.65875256e-01 5.11151731e-01 9.34641004e-01
8.94127369e-01 1.60606310e-01 3.17757130e-01 5.30586302e-01
5.83678365e-01 2.76407361e-01 -3.95323962e-01 4.91589606e-01
3.42540622e-01 1.26098919e+00 -1.01825707e-01 -5.29937804e-01
-5.41982830e-01 4.69228148e-01 -4.23974395e-01 5.97454250e-01
-8.16605091e-01 1.57899335e-02 4.44787174e-01 -5.50033376e-02
4.52431619e-01 -3.24747890e-01 -7.40760028e-01 -1.09483624e+00
-1.89268500e-01 -5.14012396e-01 6.28600836e-01 -8.57388854e-01
-1.78864646e+00 3.71984392e-01 1.87030405e-01 -1.29871523e+00
-8.70897114e-01 -1.79304972e-01 -1.35931742e+00 1.42971396e+00
-1.45399320e+00 -9.91752625e-01 8.96851495e-02 5.90376318e-01
6.11897767e-01 -3.16335857e-01 1.32633317e+00 3.83987844e-01
-5.94296515e-01 2.56066054e-01 1.15340829e+00 -6.69403851e-01
1.74828336e-01 -1.53588557e+00 4.76828128e-01 8.41272533e-01
-3.67673606e-01 2.13909954e-01 9.66395915e-01 -6.15722656e-01
-1.12817180e+00 -9.63671029e-01 1.27554440e+00 -2.32374340e-01
7.65620947e-01 -2.93984026e-01 -9.31367755e-01 8.73209476e-01
5.51436424e-01 -2.08009198e-01 9.88227904e-01 3.42500091e-01
3.33303571e-01 -5.39444983e-01 -1.28263485e+00 -4.04111594e-02
6.55355826e-02 5.64385811e-03 -2.83428848e-01 8.31426799e-01
1.52790919e-01 -1.18774422e-01 -1.41504157e+00 3.30313534e-01
4.85444397e-01 -7.31652081e-01 7.78047144e-01 -4.07249361e-01
-3.00833642e-01 -2.79551864e-01 -9.01146010e-02 -1.67135274e+00
-4.92771506e-01 -1.31228936e+00 -4.00972635e-01 1.40684700e+00
6.98040962e-01 -1.19023013e+00 4.08280380e-02 1.33463395e+00
-1.51272371e-01 -7.17479289e-01 -1.10214341e+00 -8.28246117e-01
1.18704610e-01 -7.08296239e-01 1.37656438e+00 1.18563795e+00
2.51506031e-01 -3.54437262e-01 -8.12501609e-01 1.84733972e-01
7.83592761e-01 5.21271288e-01 7.77791917e-01 -1.23975408e+00
-2.11495697e-01 -3.57157499e-01 5.90311661e-02 -7.15938389e-01
-1.12110421e-01 -6.90556824e-01 -3.90643537e-01 -1.10583770e+00
-3.26785773e-01 -4.98116165e-01 -4.38094497e-01 4.50898707e-01
2.03312665e-01 -2.93004245e-01 1.26938820e-01 2.12341800e-01
-9.54336207e-03 6.12474382e-01 4.20789838e-01 2.09782287e-01
-4.40515518e-01 7.04355061e-01 -1.84787631e-01 6.18729472e-01
8.64626765e-01 -1.27379119e-01 -3.87264103e-01 -1.18053362e-01
4.05703396e-01 4.01253223e-01 7.87005723e-02 -5.50089836e-01
2.48798802e-01 -3.24997038e-01 4.11614329e-01 -1.21492124e+00
5.65808564e-02 -1.05971611e+00 9.61832166e-01 2.74803013e-01
8.29149857e-02 6.33799493e-01 2.33049348e-01 6.26651824e-01
-2.07230672e-02 2.20773399e-01 2.30964467e-01 3.28372985e-01
-6.13014817e-01 2.73334026e-01 -4.75943148e-01 -5.12818217e-01
1.27534008e+00 -2.13259682e-01 -9.77172852e-02 -6.36982441e-01
-1.51422632e+00 8.61214459e-01 -2.49069318e-01 3.26097637e-01
-1.88341434e-03 -1.46610451e+00 -1.05846763e+00 5.23370087e-01
-4.74214584e-01 -5.20463526e-01 6.50940776e-01 1.30816507e+00
2.52979528e-02 5.28885424e-01 1.94606349e-01 -4.99440134e-01
-6.16122961e-01 6.98204994e-01 3.20190102e-01 -6.01278424e-01
-4.34140295e-01 3.27525169e-01 -1.27060682e-01 -4.26912218e-01
-3.34461853e-02 -6.12703681e-01 -2.70695895e-01 5.22809148e-01
6.77922785e-01 9.17968929e-01 3.87972593e-01 -6.43337071e-01
-4.03569520e-01 3.17503065e-01 3.53557229e-01 3.04034412e-01
1.77465475e+00 -6.65799975e-01 -3.87070000e-01 9.34972644e-01
8.71826410e-01 -2.69780874e-01 -9.79901135e-01 2.15219357e-03
1.91864833e-01 -1.38765290e-01 1.95736200e-01 -1.23320067e+00
-7.97996819e-01 3.85216922e-01 6.95836961e-01 1.18087924e+00
1.39315736e+00 -7.76166499e-01 7.20543385e-01 -1.04812220e-01
2.03961000e-01 -1.45326638e+00 -1.45479536e+00 5.30089259e-01
1.22171330e+00 -1.09870720e+00 1.28585011e-01 6.79043233e-02
-5.70289373e-01 1.10427058e+00 -3.45318735e-01 -1.42595366e-01
1.63380933e+00 4.11273688e-01 -1.52720794e-01 3.29824954e-01
-1.08451092e+00 1.92791060e-01 4.67295885e-01 4.20852035e-01
1.92524210e-01 6.26531005e-01 -7.26249635e-01 1.26216078e+00
-3.25545132e-01 2.00039104e-01 2.94643670e-01 3.30078542e-01
7.10016415e-02 -6.43965960e-01 -4.56315368e-01 9.33803320e-01
-3.13169211e-01 -2.26431191e-01 5.55028617e-01 4.10557896e-01
3.51034030e-02 1.03223658e+00 2.45511264e-01 1.84797198e-02
3.59583497e-01 5.45311332e-01 -5.59218824e-01 2.07663402e-02
-6.83129609e-01 3.06577325e-01 3.73261496e-02 -4.19816554e-01
-2.17133909e-01 -1.30089951e+00 -7.34317243e-01 -6.06303930e-01
-6.07102871e-01 3.57692987e-01 1.01115704e+00 9.07354534e-01
5.43854237e-01 2.49571040e-01 1.34346974e+00 -1.29039776e+00
-7.84161150e-01 -1.01510859e+00 -1.09539545e+00 7.10016340e-02
3.37537766e-01 -4.04944986e-01 -6.30246282e-01 5.17244004e-02] | [6.273411273956299, 2.946685552597046] |
29c82d5d-4d1a-49de-838f-b89ca216d327 | improved-generator-objectives-for-gans | 1612.02780 | null | http://arxiv.org/abs/1612.02780v1 | http://arxiv.org/pdf/1612.02780v1.pdf | Improved generator objectives for GANs | We present a framework to understand GAN training as alternating density
ratio estimation and approximate divergence minimization. This provides an
interpretation for the mismatched GAN generator and discriminator objectives
often used in practice, and explains the problem of poor sample diversity. We
also derive a family of generator objectives that target arbitrary
$f$-divergences without minimizing a lower bound, and use them to train
generative image models that target either improved sample quality or greater
sample diversity. | ['Anelia Angelova', 'Jascha Sohl-Dickstein', 'Alexander A. Alemi', 'Ben Poole'] | 2016-12-08 | null | null | null | null | ['density-ratio-estimation'] | ['methodology'] | [ 3.55898380e-01 5.20116091e-01 -4.17551070e-01 -5.15960336e-01
-1.19380963e+00 -5.01033127e-01 7.44270384e-01 -6.99436128e-01
-3.48651074e-02 1.35659814e+00 1.23464182e-01 -1.33003071e-01
1.78259641e-01 -8.34625065e-01 -5.51389575e-01 -8.23279977e-01
4.35440779e-01 7.86710382e-01 -4.00761932e-01 1.45174563e-01
1.56851858e-01 2.46026784e-01 -1.50770581e+00 -2.12112635e-01
1.31089687e+00 9.33196604e-01 8.57355352e-03 1.06854594e+00
-3.41986679e-02 8.19098234e-01 -1.14785552e+00 -5.10884702e-01
3.48779380e-01 -1.53564346e+00 -6.12428010e-01 2.48774335e-01
3.75382572e-01 -3.60686570e-01 -3.42388935e-02 1.07855952e+00
5.88776171e-01 1.34184256e-01 1.33587015e+00 -1.36500871e+00
-1.04839718e+00 2.89166182e-01 -3.98923904e-01 2.86757737e-01
6.09862767e-02 2.81277210e-01 8.50095272e-01 -4.01152492e-01
5.49240708e-01 1.00023150e+00 3.44885141e-01 1.06744266e+00
-1.38871610e+00 -5.80776632e-01 -1.78247765e-01 -4.08391267e-01
-1.26020825e+00 -6.22355402e-01 6.21184647e-01 -5.19498587e-01
8.24274004e-01 3.19092333e-01 5.77367127e-01 1.26929402e+00
1.22263551e-01 8.36922646e-01 8.84531975e-01 -3.09388727e-01
4.48531181e-01 3.99040222e-01 -4.21131432e-01 6.35013819e-01
2.71922946e-01 2.92872488e-01 -2.48701647e-01 -1.03746392e-01
1.21153569e+00 -5.22438109e-01 -4.28101271e-01 -1.83488309e-01
-6.59488559e-01 1.33578813e+00 1.78314462e-01 2.21720397e-01
-1.78160027e-01 5.28298676e-01 -8.62377323e-03 4.20104951e-01
5.82847536e-01 6.09860241e-01 8.45909864e-02 -3.44323993e-01
-1.08958113e+00 4.29070950e-01 8.18578422e-01 1.11356616e+00
8.58389437e-01 7.77096212e-01 -2.17335448e-01 9.22573507e-01
3.57329398e-01 7.89493918e-01 4.56116557e-01 -1.16981196e+00
2.23198906e-01 -7.97967166e-02 1.61139801e-01 -3.72043103e-01
2.52021879e-01 -6.31836176e-01 -7.20442712e-01 4.04874384e-01
3.54774803e-01 -1.00198120e-01 -1.20083249e+00 2.21094084e+00
-3.85774970e-02 8.60276297e-02 -6.46023154e-02 6.33969009e-01
2.72996783e-01 7.29644358e-01 -1.73288077e-01 -4.39155847e-01
4.86186892e-01 -8.32209706e-01 -7.57392704e-01 -3.74174416e-01
5.98333836e-01 -5.64431012e-01 1.27995574e+00 1.88009605e-01
-1.66287339e+00 -4.09289300e-01 -1.18215978e+00 1.84162050e-01
2.40569115e-02 -4.92728613e-02 4.40282315e-01 1.14894819e+00
-1.37041116e+00 6.89754128e-01 -7.91741610e-01 -1.46715976e-02
6.85251653e-01 2.64852226e-01 2.52766877e-01 1.47365704e-01
-5.76288283e-01 9.54042971e-01 2.63724327e-01 -1.29187882e-01
-1.20225310e+00 -7.28113234e-01 -7.59897888e-01 -1.18472576e-01
-1.86235666e-01 -1.10551322e+00 1.35671568e+00 -1.31649387e+00
-1.94773972e+00 9.00701404e-01 -9.77984667e-02 -4.64925557e-01
6.41009569e-01 -1.40021741e-01 -2.56458759e-01 -2.30776101e-01
6.12614937e-02 7.80218363e-01 1.07199287e+00 -1.33287323e+00
-3.43793064e-01 -7.78887197e-02 -3.64830434e-01 4.61832017e-01
5.62946796e-02 -4.51770186e-01 9.25353393e-02 -8.93829525e-01
-4.65707600e-01 -6.21432304e-01 -1.95959046e-01 -2.81868398e-01
-4.25910354e-01 1.34833634e-01 6.21183753e-01 -6.07235014e-01
1.03661919e+00 -1.70574307e+00 6.13677680e-01 2.56240994e-01
1.86017647e-01 5.95135912e-02 -3.45272064e-01 5.39393425e-02
2.58699238e-01 3.65252852e-01 -6.55646265e-01 -4.03453410e-01
4.63034883e-02 3.88351351e-01 -4.25563097e-01 3.41021359e-01
4.17271554e-01 1.08299780e+00 -8.69069695e-01 -2.26260498e-01
6.84460923e-02 5.55126131e-01 -7.44470060e-01 5.74152827e-01
-4.59000468e-01 5.63851774e-01 -1.14379525e-01 5.64175010e-01
5.93067527e-01 -1.31670713e-01 -4.02232744e-02 1.78347141e-01
2.68743336e-01 2.88143575e-01 -7.69126534e-01 1.48361850e+00
-4.48348582e-01 8.20237577e-01 9.42940786e-02 -1.08924770e+00
9.10827637e-01 5.57943657e-02 4.45805378e-02 -4.30768996e-01
3.47531050e-01 5.67728996e-01 -9.54370499e-02 5.29961698e-02
3.57264549e-01 -7.53126860e-01 1.48549378e-01 5.83164275e-01
4.21255201e-01 -7.90813386e-01 9.43127722e-02 -1.18825406e-01
7.70812154e-01 8.23780894e-02 1.31334051e-01 -5.74736595e-01
2.86159426e-01 -3.52992952e-01 3.05127591e-01 9.04492497e-01
1.69913143e-01 8.73672843e-01 6.21737361e-01 3.88267376e-02
-1.46059060e+00 -1.52633190e+00 -7.73448199e-02 6.59103751e-01
-7.77875036e-02 5.73163368e-02 -7.39903271e-01 -8.26912344e-01
-1.68122098e-01 1.15094972e+00 -8.43517482e-01 -4.14003849e-01
-5.59497774e-01 -8.90629888e-01 5.87729573e-01 7.20835388e-01
3.85036439e-01 -8.61704350e-01 -4.11232531e-01 -1.06598407e-01
1.35646004e-03 -3.01889598e-01 -7.12120235e-01 2.80181587e-01
-1.04372835e+00 -7.04802513e-01 -1.09680617e+00 -7.35521019e-01
6.86688423e-01 -4.66591597e-01 1.67663097e+00 -9.81116518e-02
-2.09533006e-01 4.71384376e-01 5.52940704e-02 -3.50574166e-01
-9.19918120e-01 1.62723497e-01 -1.98173925e-01 -5.34086883e-01
-7.72875026e-02 -8.09315562e-01 -5.14740825e-01 2.80943751e-01
-8.27654123e-01 -2.43850946e-01 4.32639092e-01 1.25233698e+00
7.17579544e-01 -2.67992437e-01 9.26696002e-01 -6.87640309e-01
9.61110413e-01 -5.30842304e-01 -7.65257359e-01 1.83966070e-01
-9.03467655e-01 5.40579498e-01 6.06934607e-01 -4.27495927e-01
-1.07509160e+00 -5.54421246e-01 -4.47577447e-01 -5.28887868e-01
4.59172055e-02 5.57540134e-02 -3.74170125e-01 1.25573173e-01
8.51558089e-01 4.26406682e-01 2.79429555e-01 -1.74586609e-01
5.89693308e-01 4.17173862e-01 5.39467454e-01 -6.14343882e-01
6.52484357e-01 2.36908853e-01 -1.63252473e-01 -6.47838831e-01
-7.28148818e-01 2.90815383e-01 -2.53328141e-02 -5.42409457e-02
9.18153703e-01 -6.72323465e-01 -1.16834514e-01 3.16156864e-01
-7.35903442e-01 -8.52217019e-01 -1.25354970e+00 3.24623555e-01
-1.19417214e+00 -1.49658665e-01 -4.46932614e-01 -9.31354940e-01
-3.67486030e-01 -1.06488287e+00 9.63953495e-01 3.83295655e-01
-2.01310471e-01 -1.30406499e+00 4.04208660e-01 -2.39697188e-01
6.90568388e-01 3.02130520e-01 9.04746056e-01 -3.89244378e-01
-4.67865407e-01 1.90005034e-01 4.48131040e-02 7.20930338e-01
2.76158750e-01 -1.13362044e-01 -8.68538737e-01 -4.21826035e-01
2.59380341e-01 -5.24032950e-01 7.64133513e-01 7.89414823e-01
1.06804252e+00 -5.33772588e-01 -1.41279668e-01 9.65706229e-01
1.49399078e+00 5.21876454e-01 8.49070549e-01 -2.36315623e-01
5.02404630e-01 6.33996725e-02 -2.76210196e-02 2.94859201e-01
-3.14867012e-02 3.90313804e-01 2.08187416e-01 4.36689705e-02
-2.27984220e-01 -4.51478243e-01 3.31515133e-01 8.23052466e-01
2.85084266e-02 -7.39437699e-01 -4.56174046e-01 6.57409251e-01
-1.54448247e+00 -1.09103990e+00 4.33488309e-01 2.40700817e+00
8.08290422e-01 1.66944955e-02 5.48958838e-01 -1.55880541e-01
7.11696565e-01 1.81121618e-01 -8.59096110e-01 -5.97198665e-01
-2.40615204e-01 6.69096708e-01 4.56938922e-01 9.34957027e-01
-4.57215130e-01 6.54199302e-01 8.95796299e+00 9.44186568e-01
-8.28234553e-01 5.40618747e-02 1.14141333e+00 -3.29791009e-01
-1.08310807e+00 -2.25415140e-01 -5.57854652e-01 5.75405777e-01
9.87132967e-01 -3.58978868e-01 7.14169621e-01 8.11572075e-01
-4.03437436e-01 -6.37345875e-05 -1.40761042e+00 1.13628197e+00
2.64503181e-01 -1.43006289e+00 7.70451054e-02 3.56700301e-01
1.05963886e+00 -1.94376588e-01 5.19699454e-01 2.70988435e-01
6.94622397e-01 -1.56214428e+00 7.23469853e-01 3.82798344e-01
1.07840121e+00 -1.07218027e+00 4.85593289e-01 2.76482165e-01
-7.62522876e-01 9.96851400e-02 -3.59643728e-01 1.61423698e-01
4.89424229e-01 5.28855026e-01 -7.66473889e-01 1.89278066e-01
1.06057718e-01 5.05729616e-01 -1.26482546e-01 5.92428684e-01
-3.14701796e-01 4.49944854e-01 -4.24828082e-01 -2.72650588e-02
5.27366363e-02 -7.23304212e-01 6.23979688e-01 1.00270152e+00
8.54664028e-01 -2.18148246e-01 -3.12876016e-01 1.53519654e+00
-1.98402002e-01 -2.25347176e-01 -7.93783545e-01 -2.41696596e-01
5.13402343e-01 6.71268046e-01 -5.54233730e-01 -4.76102024e-01
1.56508759e-01 1.31119025e+00 3.26730698e-01 4.19022083e-01
-1.01453781e+00 -2.65493393e-01 9.41983819e-01 1.21636346e-01
3.25702667e-01 -1.16237618e-01 -4.75791007e-01 -1.26050377e+00
-8.87918994e-02 -8.19877684e-01 2.39030749e-01 -5.35393119e-01
-1.51830721e+00 6.26357079e-01 2.00623527e-01 -9.04994369e-01
-9.92835045e-01 -3.94942224e-01 -9.16184008e-01 1.19960511e+00
-1.12063789e+00 -6.13984585e-01 -3.71058621e-02 2.80623019e-01
4.90236998e-01 -1.65782109e-01 7.10463703e-01 1.41392246e-01
-3.28808516e-01 1.04643297e+00 2.00545967e-01 -3.50214332e-01
1.76999956e-01 -1.55047452e+00 4.10175353e-01 7.67408431e-01
6.37153611e-02 2.00669900e-01 8.83810759e-01 -4.46600169e-01
-8.57480526e-01 -9.03408825e-01 2.96688497e-01 -5.79540491e-01
1.48268089e-01 -2.13476032e-01 -5.23263812e-01 6.97358131e-01
4.88328010e-01 -4.31090087e-01 7.53730595e-01 -1.53236687e-01
-2.12804392e-01 2.94066072e-01 -1.57833934e+00 4.84771043e-01
1.16638672e+00 -4.54488575e-01 -2.32259840e-01 2.93358937e-02
4.04347092e-01 -4.13283199e-01 -6.97777748e-01 2.83296764e-01
4.51754540e-01 -1.14059365e+00 7.81437874e-01 -4.33260858e-01
3.84011745e-01 -5.93283959e-02 -2.70271599e-01 -1.83713102e+00
-1.50354877e-01 -6.67713821e-01 -2.91706085e-01 1.16542065e+00
4.54985648e-01 -7.39543676e-01 9.30816114e-01 3.74803513e-01
-1.42775148e-01 -6.95250630e-01 -9.82952178e-01 -1.09099817e+00
6.30071282e-01 -1.52324736e-01 5.95958471e-01 6.55749202e-01
-1.80433020e-01 3.24802786e-01 -4.70798314e-01 -4.20076042e-01
6.97301686e-01 -1.07845955e-01 5.50596714e-01 -8.20219398e-01
-5.98176360e-01 -9.66472864e-01 -3.63902539e-01 -1.45178342e+00
1.75428882e-01 -7.94788003e-01 2.78438389e-01 -1.33104420e+00
-5.08197993e-02 -4.51666862e-01 9.18254778e-02 -1.58916861e-01
-1.67959407e-01 4.25299585e-01 -4.76780273e-02 -4.24664617e-02
-5.57116866e-02 9.18560803e-01 1.17379093e+00 -1.79735087e-02
-3.61683905e-01 -4.09216918e-02 -1.03301597e+00 3.98860604e-01
8.13339412e-01 -3.31218839e-01 -9.39959466e-01 -6.34761453e-01
3.41179073e-01 -3.39131132e-02 1.49144307e-01 -1.02191222e+00
-5.40560424e-01 -3.08044314e-01 4.68593180e-01 -1.48126245e-01
3.18622053e-01 -3.15973818e-01 4.76242691e-01 3.35740834e-01
-4.86158669e-01 9.04554792e-04 -8.99126083e-02 3.82626951e-01
-1.35008410e-01 -3.77550513e-01 1.23440194e+00 -1.20493822e-01
2.17341170e-01 2.56437600e-01 -3.52246106e-01 5.79051256e-01
8.20889592e-01 -2.95663029e-01 -2.69641966e-01 -7.90713012e-01
-6.37200832e-01 -3.54704633e-02 8.19355667e-01 4.39481176e-02
5.70123136e-01 -1.74649084e+00 -8.14000607e-01 4.49882686e-01
-2.26964951e-01 1.50906667e-01 -1.11254059e-01 3.74491274e-01
-4.93105829e-01 -7.96810240e-02 -2.20110208e-01 -4.90951866e-01
-5.96786320e-01 1.39368653e-01 8.27295065e-01 -2.27168500e-01
-2.64135242e-01 1.31432700e+00 3.00817966e-01 -2.10719764e-01
-1.03691176e-01 -1.16965070e-01 4.18911159e-01 -2.75124788e-01
4.50610727e-01 4.98964787e-01 -3.08377534e-01 -3.99342358e-01
-9.55341980e-02 3.38402659e-01 2.37339616e-01 -6.62164330e-01
8.53765368e-01 -1.12521239e-01 3.02491605e-01 4.56906974e-01
1.25923097e+00 -2.12493882e-01 -1.50440204e+00 2.78424710e-01
-5.21373212e-01 -7.91637659e-01 -1.85303792e-01 -7.62520254e-01
-1.11293042e+00 7.94917107e-01 7.26947844e-01 3.23358148e-01
1.23585987e+00 2.70399302e-01 6.31823599e-01 -6.98392242e-02
-2.46951710e-02 -1.05907035e+00 2.55774379e-01 3.52413625e-01
8.82257581e-01 -9.82610822e-01 -2.36315489e-01 -7.03024492e-02
-6.29027963e-01 6.52958989e-01 8.32371354e-01 -4.84590411e-01
4.06610489e-01 6.72460079e-01 -2.86457092e-02 3.25935930e-02
-7.17674613e-01 -1.10746913e-01 2.23469630e-01 1.11979246e+00
6.58286631e-01 -6.43848330e-02 -4.67741817e-01 1.68342650e-01
-5.45948505e-01 1.35833425e-02 2.27892816e-01 7.38514900e-01
-4.60292041e-01 -1.22444022e+00 -9.15677845e-03 7.49500275e-01
-2.76270181e-01 -6.47925586e-02 -3.15722525e-01 5.16862750e-01
-2.65471041e-02 6.30951881e-01 4.85648304e-01 -3.68961722e-01
-1.89376414e-01 3.05906594e-01 1.04795885e+00 -4.16213393e-01
-2.94057816e-01 1.72608212e-01 -9.30643082e-02 -2.00349897e-01
-3.68423462e-01 -3.99124742e-01 -6.53038263e-01 -3.85766178e-01
-6.83131099e-01 4.30717140e-01 3.93698812e-01 8.08367074e-01
1.55946165e-01 4.17840749e-01 6.97405696e-01 -7.50501573e-01
-6.73304200e-01 -1.00797069e+00 -5.92321754e-01 2.20809937e-01
2.85170019e-01 -3.95886958e-01 -7.47477233e-01 9.13564265e-02] | [11.646082878112793, -0.06982655078172684] |
d7c12dad-8e14-4f9a-8ca7-07db33675a7e | incremental-few-shot-learning-via-vector | null | null | https://openreview.net/forum?id=3SV-ZePhnZM | https://openreview.net/pdf?id=3SV-ZePhnZM | Incremental few-shot learning via vector quantization in deep embedded space | The capability of incrementally learning new tasks without forgetting old ones is a challenging problem due to catastrophic forgetting. This challenge becomes greater when novel tasks contain very few labelled training samples. Currently, most methods are dedicated to class-incremental learning and rely on sufficient training data to learn additional weights for newly added classes. Those methods cannot be easily extended to incremental regression tasks and could suffer from severe overfitting when learning few-shot novel tasks. In this study, we propose a nonparametric method in deep embedded space to tackle incremental few-shot learning problems. The knowledge about the learned tasks are compressed into a small number of quantized reference vectors. The proposed method learns new tasks sequentially by adding more reference vectors to the model using few-shot samples in each novel task. For classification problems, we employ the nearest neighbor scheme to make classification on sparsely available data and incorporate intra-class variation, less forgetting regularization and calibration of reference vectors to mitigate catastrophic forgetting. In addition, the proposed learning vector quantization (LVQ) in deep embedded space can be customized as a kernel smoother to handle incremental few-shot regression tasks. Experimental results demonstrate that the proposed method outperforms other state-of-the-art methods in incremental learning. | ['Chi-Guhn Lee', 'Kuilin Chen'] | 2021-01-01 | null | null | null | iclr-2021-1 | ['few-shot-class-incremental-learning'] | ['methodology'] | [ 4.77930427e-01 -7.47569129e-02 -2.61586905e-01 -4.87364173e-01
-6.64669991e-01 5.83742261e-02 3.41868341e-01 2.47144431e-01
-6.72955036e-01 1.09133649e+00 3.69470119e-02 4.28796440e-01
-2.53211230e-01 -6.67126715e-01 -6.45011604e-01 -7.20371783e-01
7.49567971e-02 3.16458404e-01 6.68892205e-01 -8.86205360e-02
2.03404546e-01 8.09756443e-02 -2.01477003e+00 2.48108283e-01
9.73777056e-01 9.28793132e-01 7.03149736e-01 5.13534307e-01
-2.43354395e-01 7.97066987e-01 -5.15742302e-01 -1.29906118e-01
2.47779474e-01 -1.89128771e-01 -1.98358268e-01 2.16520373e-02
3.19343597e-01 -5.08819103e-01 -3.47889394e-01 7.89790213e-01
7.87032545e-01 7.73493230e-01 7.51573384e-01 -1.27067804e+00
-8.38892758e-01 3.65823507e-01 -3.93662244e-01 4.73544002e-01
-9.22579467e-02 -5.85503019e-02 3.80794376e-01 -1.54730821e+00
5.09782672e-01 9.09168959e-01 1.02033341e+00 7.98123717e-01
-9.67576563e-01 -7.16494083e-01 3.12284470e-01 7.54233897e-01
-1.38754404e+00 -7.06376314e-01 8.79214406e-01 -3.03915113e-01
9.36858773e-01 -1.65528223e-01 4.87876922e-01 1.13586795e+00
2.38901824e-01 6.64805651e-01 7.85964370e-01 -3.98038119e-01
8.77419472e-01 5.09641707e-01 1.69882268e-01 4.48626846e-01
4.07690197e-01 -2.08304822e-01 -7.34387875e-01 -1.44954950e-01
2.44124129e-01 1.11121154e+00 -1.95528224e-01 -6.61612511e-01
-9.21308756e-01 9.74432945e-01 2.70893246e-01 2.56370664e-01
-3.77755135e-01 -9.08763111e-02 8.07675898e-01 5.45514882e-01
7.85725594e-01 -6.72904402e-02 -6.95347011e-01 -1.29133597e-01
-1.07436657e+00 8.76401067e-02 4.21917260e-01 1.07167161e+00
1.06538224e+00 4.33669358e-01 -4.01738644e-01 1.14572561e+00
-1.48304164e-01 2.22104520e-01 1.41867077e+00 -7.93581963e-01
3.29493493e-01 3.59909981e-01 -5.12440577e-02 -8.49107862e-01
-1.02238618e-01 -5.62104762e-01 -8.49519193e-01 9.76033658e-02
-2.29143068e-01 -1.26160294e-01 -8.85694504e-01 1.27508128e+00
4.90522534e-01 5.79708815e-01 9.88319144e-02 3.88914555e-01
5.39364934e-01 7.59023547e-01 2.06807274e-02 -7.98831582e-01
8.44407141e-01 -1.11664832e+00 -8.77378285e-01 -3.34597975e-01
5.50612986e-01 -3.61884236e-01 1.27132487e+00 3.83148789e-01
-6.04759753e-01 -1.01978230e+00 -1.24688578e+00 -2.64640488e-02
-5.78786790e-01 -8.42667297e-02 2.61735737e-01 4.66655016e-01
-6.74872518e-01 8.00407946e-01 -8.07109654e-01 -2.54056633e-01
6.66856706e-01 1.20655790e-01 -1.99791729e-01 -5.25520980e-01
-1.18015933e+00 7.13684082e-01 7.15967774e-01 -2.68059254e-01
-1.08131766e+00 -9.10589695e-01 -9.93801057e-01 6.71024099e-02
6.16930723e-01 -4.55706567e-01 1.31073916e+00 -7.53538430e-01
-1.51566482e+00 2.07926735e-01 -3.90364677e-01 -6.20551705e-01
3.33804429e-01 -4.35513973e-01 -3.49475324e-01 -8.33476260e-02
1.37892440e-01 3.12491119e-01 1.58462203e+00 -9.06912804e-01
-7.20589995e-01 -5.67644894e-01 -4.21715796e-01 4.84403193e-01
-1.13165832e+00 -6.96781337e-01 6.32581562e-02 -9.72358108e-01
-1.61422595e-01 -5.49524128e-01 -1.08648449e-01 1.97411418e-01
4.97139573e-01 -3.49956065e-01 1.39396930e+00 -5.00486970e-01
1.32502437e+00 -2.28693342e+00 5.69185019e-02 -5.81627846e-01
1.60388455e-01 4.75636989e-01 2.87804361e-02 2.44921774e-01
1.71311468e-01 -5.81888080e-01 -3.77169907e-01 -5.49327135e-01
-1.86564237e-01 4.10656959e-01 -4.81956005e-01 2.25638628e-01
1.66770872e-02 7.72500277e-01 -1.12877846e+00 -3.44920158e-01
2.44299024e-01 6.42504334e-01 -3.41358572e-01 2.11613134e-01
3.05589046e-02 2.31110044e-02 -1.48081973e-01 5.59640586e-01
5.92209995e-01 -2.21011415e-02 -2.73645401e-01 3.79783101e-02
1.97762117e-01 -3.37372005e-01 -1.21582305e+00 2.00844312e+00
-6.75117731e-01 3.64922762e-01 -3.98115844e-01 -1.27600467e+00
1.15658200e+00 1.80565193e-01 1.60540178e-01 -4.08846080e-01
-1.91809908e-01 1.11456156e-01 -4.52855587e-01 -5.00960946e-01
4.77948606e-01 -5.48865199e-01 1.06422547e-02 3.30430180e-01
5.57787180e-01 -1.09115586e-01 9.34915394e-02 1.00009993e-01
1.02825201e+00 -1.76538248e-03 6.37739480e-01 3.15758526e-01
3.62635523e-01 -2.69073665e-01 9.33384478e-01 7.84042060e-01
-5.58772683e-01 4.41691607e-01 -3.36278886e-01 -7.22468734e-01
-1.01800954e+00 -1.06775296e+00 -5.92294894e-02 1.63238204e+00
-2.32051447e-01 -3.55827540e-01 -3.78657132e-01 -9.46690679e-01
2.94154853e-01 9.74867046e-01 -7.25846648e-01 -8.27074468e-01
-3.06873918e-01 -7.06050217e-01 -9.58654359e-02 6.16575599e-01
3.43411267e-01 -1.01467705e+00 -8.59092176e-01 5.38900495e-01
1.53787553e-01 -7.36531019e-01 -4.76709306e-01 5.38652480e-01
-1.36963952e+00 -8.66469741e-01 -1.19418252e+00 -9.65157926e-01
6.02105260e-01 7.29826510e-01 5.28698385e-01 -5.38853526e-01
-5.45704484e-01 5.35953760e-01 -6.36887670e-01 -5.47337592e-01
-1.15952734e-02 4.29634079e-02 6.33869350e-01 1.95912272e-01
6.14207566e-01 -5.91989756e-01 -5.19507647e-01 3.08884364e-02
-8.92928421e-01 -3.90663236e-01 6.12520516e-01 1.26294208e+00
8.22188675e-01 3.69930446e-01 1.00713205e+00 -1.17726707e+00
7.08227694e-01 -7.83934951e-01 3.96657847e-02 2.74927944e-01
-8.89238834e-01 1.02717042e-01 1.14766991e+00 -9.87893760e-01
-1.33542860e+00 1.21569701e-01 4.43979591e-01 -8.49665463e-01
6.97196797e-02 2.72739381e-01 2.89682984e-01 -2.40473915e-02
7.66340435e-01 6.69350445e-01 -9.35206413e-02 -6.26583934e-01
6.33074939e-01 7.96305120e-01 3.09935778e-01 -2.27909714e-01
6.47540510e-01 4.28789616e-01 -3.31433684e-01 -8.80124211e-01
-1.02033496e+00 -5.69504857e-01 -8.56119215e-01 -1.35581985e-01
3.11726570e-01 -1.09581757e+00 1.22343823e-01 4.30339009e-01
-6.73028827e-01 -3.09409559e-01 -1.03795028e+00 4.88406688e-01
-6.85559154e-01 4.51804847e-01 -2.71159083e-01 -9.03241754e-01
-3.88448358e-01 -6.28609836e-01 8.47065568e-01 2.08872780e-01
1.78847220e-02 -8.37397516e-01 2.64993340e-01 5.65748215e-02
7.40226507e-01 -7.52890706e-02 6.50842786e-01 -7.09927022e-01
1.78283811e-01 -3.28833997e-01 4.68925536e-02 6.42866373e-01
5.65107822e-01 -7.65377998e-01 -1.04748511e+00 -7.88417161e-01
6.10021651e-01 -6.94686949e-01 1.25688517e+00 1.43621117e-01
1.20617652e+00 -4.76192683e-01 -2.48800099e-01 4.78362739e-01
1.31779420e+00 2.09875405e-01 2.20786735e-01 4.71719466e-02
6.40978634e-01 3.68457764e-01 9.53911245e-01 9.25602198e-01
2.63256401e-01 2.86895007e-01 5.01443818e-02 7.15707481e-01
-3.48125160e-01 -3.03810716e-01 4.74631369e-01 1.34547853e+00
1.63351297e-01 4.84131247e-01 -6.73912406e-01 6.90454006e-01
-2.01636887e+00 -1.05725050e+00 6.53869033e-01 2.34090185e+00
1.06321335e+00 2.48838410e-01 -2.74952888e-01 4.74617243e-01
7.80355752e-01 1.68353587e-01 -1.14894533e+00 -3.76499593e-02
2.62519300e-01 3.25285226e-01 1.68880507e-01 2.33732596e-01
-1.03868139e+00 8.07409346e-01 5.28601551e+00 1.11667907e+00
-1.04485393e+00 7.75148034e-01 5.06967485e-01 -4.72858906e-01
1.96818888e-01 -2.27461189e-01 -1.14509642e+00 7.36190379e-01
1.09900212e+00 -4.41024005e-01 3.06363940e-01 1.30559528e+00
-8.26682746e-02 -5.25359251e-02 -9.65800405e-01 1.31633949e+00
6.27288520e-01 -9.86451924e-01 1.59995601e-01 -6.94068968e-01
9.96077299e-01 -2.13311195e-01 2.96658456e-01 1.10814285e+00
-2.16929186e-02 -4.94278133e-01 2.72948653e-01 9.13057804e-01
9.80388880e-01 -7.34012604e-01 6.21845126e-01 6.76685035e-01
-1.12945116e+00 -7.39252865e-01 -1.03418386e+00 -3.06077212e-01
-1.29633591e-01 7.56664217e-01 -1.03897822e+00 7.92402476e-02
8.95239055e-01 1.10046709e+00 -7.68346846e-01 9.81969774e-01
1.19263150e-01 4.31431741e-01 4.08253521e-02 -4.05823886e-02
-2.21715629e-01 3.10851425e-01 2.88718075e-01 8.02094400e-01
6.95201695e-01 3.82155143e-02 1.24518439e-01 2.39676818e-01
-4.35505174e-02 1.77091643e-01 -9.58471894e-01 1.87452778e-01
7.13214338e-01 9.27480340e-01 -4.12641823e-01 -6.56931818e-01
-5.50402999e-01 1.62349987e+00 6.74768090e-01 3.46814543e-01
-4.25860196e-01 -8.30280602e-01 3.23180705e-01 2.02408224e-01
7.75669098e-01 -1.85718089e-01 1.56891271e-01 -1.28079462e+00
1.87606633e-01 -4.37905550e-01 5.19888282e-01 -7.46959746e-01
-1.57021332e+00 3.20847064e-01 -4.20514308e-02 -1.43904519e+00
-3.56410861e-01 -2.66434044e-01 -4.75763053e-01 3.39493483e-01
-1.72100890e+00 -8.15758467e-01 -4.81304765e-01 8.12548220e-01
1.40863609e+00 -7.16484308e-01 9.08308029e-01 3.25601071e-01
-2.56128788e-01 6.86881185e-01 7.29087055e-01 -5.18669069e-01
1.17046905e+00 -9.68593955e-01 2.87249163e-02 4.53845203e-01
-3.82436290e-02 4.88036811e-01 4.56644595e-01 -8.00625086e-01
-1.27131045e+00 -1.55541444e+00 6.94525480e-01 -3.49526107e-01
4.13967788e-01 -5.28445005e-01 -1.24181199e+00 6.04542494e-01
-2.75217891e-01 5.56385756e-01 7.83183694e-01 -1.96233585e-01
-3.36789936e-01 -6.23416543e-01 -1.28222859e+00 2.08742082e-01
8.89077067e-01 -5.84277451e-01 -9.21056509e-01 3.30610305e-01
1.18039048e+00 9.06120464e-02 -6.70144379e-01 3.51273835e-01
4.00952429e-01 -6.27903640e-01 8.57252538e-01 -5.25262475e-01
-1.39228880e-01 -1.71991855e-01 -2.92659879e-01 -1.63978875e+00
-5.47551572e-01 -2.17870429e-01 -8.56961250e-01 9.75795150e-01
-5.29054478e-02 -5.28326035e-01 8.37494135e-01 3.19564015e-01
-1.38696760e-01 -7.37358391e-01 -1.11147463e+00 -1.04038584e+00
-1.95051551e-01 -2.54466385e-01 2.44345218e-01 1.02950168e+00
-7.89247304e-02 4.42570627e-01 -8.13882709e-01 -3.57744962e-01
8.76697004e-01 -2.84446806e-01 6.64738953e-01 -1.51202500e+00
-3.40559810e-01 1.89866111e-01 -5.59359848e-01 -6.36188567e-01
4.87972051e-02 -8.47430110e-01 5.20243049e-02 -1.23445487e+00
3.06522310e-01 -2.36601576e-01 -7.95611024e-01 4.96043473e-01
-4.32739079e-01 2.22195670e-01 8.88149068e-02 3.79818439e-01
-1.06814969e+00 1.23718882e+00 8.33856761e-01 -3.16984832e-01
-3.98890316e-01 1.08088486e-01 -5.14036477e-01 6.13745570e-01
7.31358826e-01 -7.62445092e-01 -8.57147694e-01 -2.68838108e-01
-4.11810949e-02 -2.20680609e-01 -6.96803629e-02 -1.44515336e+00
5.74824929e-01 5.56433909e-02 7.57774949e-01 -4.87102419e-01
4.75191563e-01 -9.50235367e-01 -1.95225045e-01 5.89810550e-01
-2.51450419e-01 -2.86874086e-01 -1.59107540e-02 1.23994589e+00
-1.07399963e-01 -6.10689580e-01 7.46869504e-01 -3.47564995e-01
-1.15398693e+00 4.45699155e-01 -9.27698389e-02 1.35103211e-01
1.44636810e+00 -3.92165840e-01 1.16032146e-01 -2.92462390e-02
-1.07834291e+00 7.81356543e-02 2.35878497e-01 6.56837761e-01
1.27068949e+00 -1.50421154e+00 -5.23298383e-01 3.92036498e-01
4.58408058e-01 7.12620690e-02 6.98835909e-01 3.14210027e-01
2.26335656e-02 1.41475514e-01 -3.70385319e-01 -3.25600624e-01
-9.40673709e-01 1.01809585e+00 -2.94184715e-01 -3.71760316e-02
-6.48754299e-01 9.35115278e-01 -1.41414687e-01 -4.64185447e-01
6.27109885e-01 -1.16153255e-01 -2.16899931e-01 5.06187677e-01
9.58370447e-01 7.62741446e-01 1.85432985e-01 -2.03554526e-01
-5.96333221e-02 4.48359489e-01 -6.68722272e-01 2.76156515e-01
1.62489831e+00 -3.72002423e-01 3.21483523e-01 1.40555882e+00
1.23543584e+00 -5.42510867e-01 -1.64931619e+00 -8.90153885e-01
-1.99486032e-01 -5.66209197e-01 4.42038029e-02 -4.96474057e-01
-6.58639491e-01 1.09543991e+00 1.12616301e+00 -2.44773865e-01
9.22842562e-01 -3.81316960e-01 1.09496808e+00 8.32264006e-01
5.69571555e-01 -1.61264300e+00 5.56764543e-01 5.09051323e-01
7.57271826e-01 -1.46147907e+00 1.86077565e-01 2.05357417e-01
-6.31205499e-01 1.13085365e+00 7.59376407e-01 -2.10810572e-01
1.02702832e+00 -1.40296787e-01 -3.92062217e-01 3.60259354e-01
-9.86999869e-01 1.27432197e-01 6.53252453e-02 9.08132315e-01
5.79048172e-02 -1.14642724e-01 -1.58440471e-01 8.92572284e-01
1.62000284e-01 4.08297479e-01 4.18581635e-01 1.29047334e+00
-1.06872904e+00 -7.82637477e-01 -1.82710782e-01 8.68960619e-01
-3.29839252e-02 -1.38101146e-01 2.03584120e-01 2.07898512e-01
4.05598372e-01 5.09873807e-01 6.00137971e-02 -2.86508262e-01
3.63419622e-01 5.99405408e-01 4.61302727e-01 -1.12475288e+00
-2.90958174e-02 -4.12332535e-01 -6.86544538e-01 -2.65489340e-01
-7.15522617e-02 -6.44115508e-01 -9.27340567e-01 1.64018407e-01
-3.02240819e-01 1.51511371e-01 4.80072379e-01 6.46040320e-01
4.94035780e-01 6.95923448e-01 7.44588673e-01 -9.73143756e-01
-1.04856384e+00 -1.16638482e+00 -8.23654234e-01 3.43653083e-01
4.18170065e-01 -1.04255426e+00 -5.06416678e-01 3.42651188e-01] | [9.910062789916992, 3.2912869453430176] |
cefa1939-ef27-4ed0-8adb-3a15cb555370 | embedding-individual-table-columns-for | 1811.00633 | null | http://arxiv.org/abs/1811.00633v1 | http://arxiv.org/pdf/1811.00633v1.pdf | Embedding Individual Table Columns for Resilient SQL Chatbots | Most of the world's data is stored in relational databases. Accessing these
requires specialized knowledge of the Structured Query Language (SQL), putting
them out of the reach of many people. A recent research thread in Natural
Language Processing (NLP) aims to alleviate this problem by automatically
translating natural language questions into SQL queries. While the proposed
solutions are a great start, they lack robustness and do not easily generalize:
the methods require high quality descriptions of the database table columns,
and the most widely used training dataset, WikiSQL, is heavily biased towards
using those descriptions as part of the questions.
In this work, we propose solutions to both problems: we entirely eliminate
the need for column descriptions, by relying solely on their contents, and we
augment the WikiSQL dataset by paraphrasing column names to reduce bias. We
show that the accuracy of existing methods drops when trained on our augmented,
column-agnostic dataset, and that our own method reaches state of the art
accuracy, while relying on column contents only. | ['Andreea Hossmann', 'Ignacio Aguado', 'Michael Baeriswyl', 'Claudiu Musat', 'Bojan Petrovski'] | 2018-11-01 | embedding-individual-table-columns-for-1 | https://aclanthology.org/W18-5710 | https://aclanthology.org/W18-5710.pdf | ws-2018-10 | ['sql-chatbots'] | ['computer-code'] | [-1.73457399e-01 3.48797709e-01 -3.62982750e-01 -5.38231313e-01
-8.74380231e-01 -1.00178885e+00 5.83387673e-01 7.38006294e-01
-4.66220289e-01 7.56080687e-01 3.47257972e-01 -4.96408045e-01
-1.25433221e-01 -1.24213398e+00 -8.72463107e-01 7.12556541e-02
4.73270625e-01 8.56238782e-01 6.03326619e-01 -5.37989259e-01
3.00872386e-01 5.06049275e-01 -1.57312155e+00 6.28395081e-01
9.81006265e-01 1.12278605e+00 -1.29890576e-01 2.45355546e-01
-1.02735984e+00 1.20963514e+00 -4.66971606e-01 -6.69853091e-01
2.66764969e-01 -1.47260979e-01 -1.16606951e+00 -2.29104415e-01
4.11036074e-01 -4.74417731e-02 -1.21006571e-01 1.01594341e+00
1.26435861e-01 -2.29516625e-01 1.86266914e-01 -1.17720020e+00
-4.94357854e-01 7.37120986e-01 3.37935723e-02 -3.07452559e-01
6.20131016e-01 -1.97304279e-01 1.27007735e+00 -8.70459676e-01
8.21895480e-01 9.01656985e-01 5.83081067e-01 2.48287678e-01
-1.19356513e+00 -1.64385125e-01 -9.44416523e-02 1.15158871e-01
-1.48567498e+00 -5.22288918e-01 4.95699197e-01 -3.32556576e-01
1.12448585e+00 6.00441575e-01 1.48480579e-01 4.78477299e-01
-2.59342641e-01 5.74675262e-01 1.03988039e+00 -4.64558840e-01
3.14331353e-01 9.11557317e-01 2.79431343e-01 4.77990448e-01
4.86628890e-01 -4.16824758e-01 -3.95536244e-01 -2.62305319e-01
3.75605524e-01 9.67521816e-02 -8.79253373e-02 -7.63819873e-01
-1.11835551e+00 7.32306242e-01 1.52518451e-01 4.34723854e-01
-2.53157169e-01 -4.50723976e-01 4.98184770e-01 4.49826658e-01
8.13390408e-03 9.80391800e-01 -7.90788531e-01 -1.55762792e-01
-8.71855795e-01 4.77852374e-01 1.54797482e+00 1.51425111e+00
1.12892830e+00 -6.37902796e-01 -1.65341385e-02 7.16692150e-01
-7.84211755e-02 3.65753174e-01 4.46609855e-01 -9.06307578e-01
7.46891201e-01 1.27869284e+00 3.12502086e-01 -1.09356964e+00
-3.65014672e-01 -5.73907457e-02 -4.58991379e-01 -3.70626628e-01
6.34326279e-01 1.86125383e-01 -3.60485256e-01 1.35251975e+00
2.01077357e-01 -9.20538127e-01 2.55209744e-01 5.63294470e-01
7.18883455e-01 6.41982615e-01 -1.68261260e-01 -1.50157049e-01
1.34161878e+00 -6.50194407e-01 -6.70931101e-01 -3.67071837e-01
7.74886668e-01 -6.22745633e-01 1.48947763e+00 4.39714760e-01
-9.56303358e-01 -2.99675882e-01 -8.79145086e-01 -4.56747890e-01
-1.01536286e+00 2.32830066e-02 5.94608843e-01 7.62957036e-01
-9.20191407e-01 2.59362549e-01 -5.73930383e-01 -6.30795240e-01
3.19029167e-02 2.10592955e-01 -6.27220988e-01 -2.24130377e-01
-1.09599221e+00 9.03181136e-01 4.94436771e-01 -1.60419211e-01
-7.99674988e-02 -6.94381416e-01 -8.57998133e-01 1.86191663e-01
1.28287888e+00 -4.41925794e-01 1.16541147e+00 -2.07538471e-01
-1.15525877e+00 8.83342087e-01 -3.40728909e-01 -4.56690520e-01
4.90213871e-01 -2.40505219e-01 -2.51452893e-01 7.97432214e-02
1.31098896e-01 2.51312435e-01 1.81572288e-01 -1.23885059e+00
-4.62324351e-01 -5.63022196e-01 3.76320839e-01 -2.55367100e-01
-4.57790017e-01 1.16198458e-01 -8.63501310e-01 -4.07174647e-01
2.44192839e-01 -7.81733632e-01 -2.84373611e-01 5.48530258e-02
-6.40203536e-01 -2.41306856e-01 3.71968716e-01 -5.65935373e-01
1.67543495e+00 -1.91356611e+00 -1.58411741e-01 4.35933679e-01
2.33477548e-01 1.59623012e-01 1.59049511e-01 1.03235757e+00
1.36780977e-01 4.94169742e-01 -3.93070638e-01 1.27521440e-01
3.24574739e-01 4.69006211e-01 -7.22499371e-01 -1.18107095e-01
3.80254328e-01 8.10657382e-01 -6.07070982e-01 -6.43755436e-01
-1.14059396e-01 -9.05904844e-02 -8.86727929e-01 2.61871994e-01
-7.73555875e-01 -1.75137416e-01 -3.67060035e-01 5.56165218e-01
6.30793869e-01 -2.22415850e-01 4.65598553e-01 -2.43444979e-01
-2.20602274e-01 6.63851976e-01 -1.43801415e+00 1.49195111e+00
-4.69751537e-01 8.85986984e-02 -5.38041666e-02 -1.01067173e+00
1.06890297e+00 4.21523303e-02 5.40745020e-01 -7.58914411e-01
-4.67539668e-01 4.51863855e-01 -3.23984623e-01 -8.03869069e-01
4.61390018e-01 2.24572979e-02 -4.50483054e-01 4.41797465e-01
-9.81374234e-02 -4.40264404e-01 6.69877648e-01 3.51822317e-01
1.12113488e+00 -1.21834137e-01 5.44385135e-01 -1.82758078e-01
8.12711716e-01 5.05494595e-01 4.03725266e-01 8.97772193e-01
3.06924701e-01 5.24835348e-01 9.90951538e-01 -4.25028414e-01
-1.02870131e+00 -1.07063389e+00 -7.10451826e-02 9.95683491e-01
-1.90806329e-01 -9.59617555e-01 -6.42612636e-01 -7.85788000e-01
2.44009733e-01 7.81725705e-01 -4.13995415e-01 3.67323458e-02
-6.05302691e-01 -3.60984623e-01 4.97954607e-01 5.40793478e-01
3.47087741e-01 -9.06850934e-01 -5.17382085e-01 4.07340854e-01
-3.15892488e-01 -1.46531069e+00 -4.81868759e-02 1.07631981e-01
-6.51054621e-01 -1.29575706e+00 -1.62258893e-01 -3.86924386e-01
3.74462336e-01 -1.44043028e-01 1.56779599e+00 8.33328888e-02
-1.78682640e-01 3.08686972e-01 -3.84773672e-01 -6.19704664e-01
-4.50100660e-01 3.32388043e-01 -4.48375285e-01 -1.71172246e-01
7.12952971e-01 -3.01876873e-01 -1.26793891e-01 3.43169391e-01
-1.32406020e+00 -1.54428810e-01 6.09452128e-01 4.95437473e-01
5.89299977e-01 -9.16143656e-02 4.98347342e-01 -1.68428326e+00
5.10987461e-01 -3.90782833e-01 -7.91581929e-01 5.64881444e-01
-9.61578488e-01 5.69262445e-01 9.74616885e-01 1.02185473e-01
-7.88057506e-01 -2.06608865e-02 -3.13256413e-01 1.47004828e-01
-2.73595870e-01 8.43830287e-01 -5.43987155e-01 2.66746879e-01
6.41052485e-01 1.66460171e-01 2.93344371e-02 -7.25136936e-01
4.21794415e-01 4.63915110e-01 5.06214797e-01 -7.43458748e-01
9.08110380e-01 4.30715024e-01 -1.45555764e-01 -6.65214598e-01
-9.54389334e-01 -5.62971354e-01 -6.67931437e-01 3.84272814e-01
4.82522070e-01 -5.43414831e-01 -7.47661531e-01 -1.48958102e-01
-1.03071606e+00 9.00034457e-02 -5.92882752e-01 1.24748118e-01
-5.60690999e-01 2.57381797e-01 -3.50981414e-01 -5.62366247e-01
-6.17647246e-02 -9.99034226e-01 8.61841083e-01 -2.26776570e-01
-2.63043851e-01 -5.69215119e-01 8.46301541e-02 3.97099465e-01
6.77977920e-01 1.90171041e-02 1.51388788e+00 -1.08839250e+00
-7.15456903e-01 -4.54001427e-01 -4.89475101e-01 2.77734190e-01
1.48503870e-01 -1.50076300e-01 -5.96765101e-01 1.53426260e-01
5.42827621e-02 -3.11999708e-01 5.05159855e-01 -4.55316037e-01
1.34388924e+00 -6.13002837e-01 -3.44664082e-02 2.44752094e-01
1.79273999e+00 -1.08370017e-02 6.80868447e-01 4.38943148e-01
4.88106728e-01 1.10196602e+00 5.32327831e-01 5.26267588e-01
8.57660174e-01 5.62678576e-01 3.20843071e-01 1.64855719e-01
2.68721789e-01 -4.74156439e-01 -3.19387279e-02 6.50313139e-01
3.30324382e-01 -4.10734676e-02 -1.28960323e+00 5.35636306e-01
-1.77567923e+00 -6.87655091e-01 -1.07756533e-01 2.34560466e+00
1.15457928e+00 1.74141183e-01 6.47825375e-02 2.18362972e-01
1.24874078e-01 -8.63799900e-02 -3.82659882e-01 -3.02512914e-01
-2.18483612e-01 3.75284284e-01 5.02901077e-01 2.92142391e-01
-8.50705862e-01 8.58594000e-01 6.24086380e+00 2.57848173e-01
-9.24047947e-01 -2.59160519e-01 8.13026428e-02 1.30934015e-01
-5.70774317e-01 2.08849460e-01 -9.13868666e-01 2.17370600e-01
1.06023586e+00 -3.87984127e-01 4.18777466e-01 9.12360728e-01
-1.04094796e-01 -6.09612800e-02 -1.44702327e+00 8.21620762e-01
1.65421426e-01 -1.44959104e+00 3.00343037e-01 -1.10591576e-01
1.34695649e-01 -2.78868914e-01 -2.59745061e-01 5.47289371e-01
9.46395025e-02 -1.02926636e+00 6.23581350e-01 6.93223238e-01
3.93919796e-01 -5.76065600e-01 9.31784034e-01 4.42765921e-01
-8.83781075e-01 -1.18561104e-01 -4.40980405e-01 5.89333028e-02
-2.80069172e-01 5.65448523e-01 -8.31507564e-01 6.78595245e-01
7.89258659e-01 1.86176986e-01 -1.14408529e+00 7.64701724e-01
6.35895431e-02 3.67387891e-01 -4.49014693e-01 -2.65547693e-01
5.44274449e-02 -1.51447967e-01 3.15652043e-02 1.11343479e+00
9.68709067e-02 -2.49183755e-02 2.75183380e-01 9.36011314e-01
-2.29753509e-01 5.12594759e-01 -7.32707798e-01 -4.02008653e-01
4.25248712e-01 8.59592199e-01 -3.27075094e-01 -5.75068295e-01
-8.74058485e-01 4.59681243e-01 4.63805348e-01 3.10872644e-01
-4.48674977e-01 -7.13902473e-01 3.86448026e-01 4.79436040e-01
2.82374144e-01 -1.01853341e-01 -4.14073318e-01 -1.29814470e+00
7.69925594e-01 -1.27885973e+00 5.10718048e-01 -6.72804058e-01
-1.28344285e+00 5.68234146e-01 1.75902769e-01 -9.38896477e-01
-6.39669716e-01 -6.21566892e-01 1.08519278e-01 6.90248728e-01
-1.50357008e+00 -7.36823738e-01 -2.22294405e-01 6.58072472e-01
1.32766485e-01 7.75174573e-02 9.54937994e-01 6.17022514e-01
-2.39283189e-01 5.58600366e-01 -1.20354474e-01 3.99735153e-01
9.80651498e-01 -1.35873830e+00 2.31324941e-01 6.46501422e-01
2.22617969e-01 1.22061348e+00 7.37703621e-01 -2.92911768e-01
-1.99642158e+00 -9.77648854e-01 1.45579469e+00 -6.64137900e-01
8.20849240e-01 -6.25529170e-01 -1.33127844e+00 6.84811056e-01
-3.17523070e-02 9.98112112e-02 7.31266081e-01 1.75151899e-01
-8.03182483e-01 -4.40921456e-01 -1.15775406e+00 4.65302676e-01
7.21820414e-01 -8.77056301e-01 -7.44778812e-01 4.79785770e-01
8.59226406e-01 -1.07693240e-01 -9.89436328e-01 3.68379503e-01
3.94743234e-01 -1.01185036e+00 8.69364619e-01 -8.18589866e-01
3.60572815e-01 -5.03840387e-01 -3.84767711e-01 -6.65933430e-01
3.85961056e-01 -3.05180877e-01 -7.66427517e-02 1.22731900e+00
8.05650711e-01 -8.19500387e-01 9.30636644e-01 1.11426353e+00
9.53238159e-02 -7.17368364e-01 -5.83940446e-01 -6.69257760e-01
1.69223715e-02 -4.40098703e-01 8.19099426e-01 9.02099907e-01
2.49451220e-01 3.06390047e-01 8.06661844e-02 9.91349816e-02
1.64447561e-01 4.93550062e-01 1.14169252e+00 -1.23806262e+00
-2.58696944e-01 -2.40346447e-01 -2.04212338e-01 -9.04065669e-01
-1.76277012e-01 -8.03853333e-01 -1.24540746e-01 -1.65581954e+00
-2.07384266e-02 -4.95401859e-01 1.73599645e-01 7.05616176e-01
1.52176082e-01 -1.76801771e-01 3.70791912e-01 1.08457170e-01
-5.98940730e-01 8.39294344e-02 6.46283627e-01 -1.90682083e-01
-1.15497522e-01 -3.57962921e-02 -1.01528358e+00 4.84763592e-01
4.99888003e-01 -5.93932688e-01 -3.54990631e-01 -4.05277967e-01
8.30408335e-01 1.26062915e-01 1.58825919e-01 -1.05623901e+00
6.83781385e-01 -2.52485365e-01 -8.65981802e-02 -6.58049703e-01
-4.11160523e-03 -1.15658641e+00 5.19168638e-02 1.91855565e-01
-5.25821149e-01 3.57685804e-01 8.07673931e-02 2.35081479e-01
-6.55814588e-01 -3.55982155e-01 4.50394034e-01 -3.82128328e-01
-5.43106556e-01 1.98568270e-01 -1.42568737e-01 5.37799835e-01
7.12695062e-01 1.43760607e-01 -3.19686770e-01 -1.97559074e-01
-4.89545882e-01 3.45950902e-01 6.68760657e-01 3.81606400e-01
3.69269550e-01 -8.76584530e-01 -3.44319612e-01 2.80802578e-01
5.54662704e-01 9.89361331e-02 -2.19995186e-01 5.70671320e-01
-7.53564477e-01 8.83568704e-01 5.53918481e-02 -2.93858975e-01
-6.34454429e-01 9.63620484e-01 1.97177436e-02 -5.81496656e-01
-2.76129425e-01 4.90132749e-01 -1.54128924e-01 -9.77812231e-01
3.09676647e-01 -7.31020093e-01 -2.08533704e-01 2.13754714e-01
4.28475171e-01 2.39741919e-03 5.81588030e-01 -1.04334116e-01
-4.52612817e-01 3.30885828e-01 -2.06137374e-01 -5.59138320e-02
1.35465765e+00 -3.10490988e-02 -5.69689870e-01 6.10352218e-01
1.19199657e+00 5.19591987e-01 -3.12389344e-01 -5.30236423e-01
7.11838186e-01 -4.06129152e-01 -6.19956851e-01 -8.44972193e-01
-6.49146795e-01 8.68923843e-01 -9.80284214e-02 4.92097050e-01
1.09752631e+00 1.46526411e-01 7.87920177e-01 1.07630515e+00
6.17920041e-01 -9.43251908e-01 -2.89053410e-01 6.04693651e-01
8.37578773e-01 -1.28834474e+00 -2.14470085e-02 -6.43999755e-01
-4.94369358e-01 1.20719230e+00 5.57658970e-01 3.76241133e-02
6.05798304e-01 3.68216574e-01 1.71318084e-01 -2.91285276e-01
-8.48958671e-01 -3.25224489e-01 6.62794039e-02 3.96926254e-01
5.03553987e-01 -3.85614038e-01 -4.34903413e-01 7.87048876e-01
-4.73314822e-01 -1.31536908e-02 5.87508798e-01 1.33258176e+00
-2.63133526e-01 -1.44560063e+00 -2.61347950e-01 6.61279023e-01
-5.07206976e-01 -2.48071820e-01 -6.38984740e-01 9.03576255e-01
-1.50811583e-01 8.65917742e-01 -2.70065311e-02 -1.27140716e-01
8.98825705e-01 3.55557114e-01 1.50796786e-01 -9.26764727e-01
-6.73770249e-01 -6.10913754e-01 1.33714169e-01 -7.65188873e-01
-1.59083113e-01 -4.49517190e-01 -1.26354623e+00 -1.63514689e-01
8.39354321e-02 6.20775878e-01 6.35293841e-01 8.04455638e-01
4.78431910e-01 -4.77018161e-03 3.84619862e-01 1.61263958e-01
-8.70859385e-01 -6.00703955e-01 -2.97507524e-01 5.14422655e-01
2.69090265e-01 -2.60638863e-01 -1.37847379e-01 9.15434584e-02] | [9.79900074005127, 7.887511730194092] |
99c74226-7a40-4681-922c-473e1cbb8719 | learning-to-rank-microphones-for-distant | 2104.02819 | null | https://arxiv.org/abs/2104.02819v2 | https://arxiv.org/pdf/2104.02819v2.pdf | Learning to Rank Microphones for Distant Speech Recognition | Fully exploiting ad-hoc microphone networks for distant speech recognition is still an open issue. Empirical evidence shows that being able to select the best microphone leads to significant improvements in recognition without any additional effort on front-end processing. Current channel selection techniques either rely on signal, decoder or posterior-based features. Signal-based features are inexpensive to compute but do not always correlate with recognition performance. Instead decoder and posterior-based features exhibit better correlation but require substantial computational resources. In this work, we tackle the channel selection problem by proposing MicRank, a learning to rank framework where a neural network is trained to rank the available channels using directly the recognition performance on the training set. The proposed approach is agnostic with respect to the array geometry and type of recognition back-end. We investigate different learning to rank strategies using a synthetic dataset developed on purpose and the CHiME-6 data. Results show that the proposed approach is able to considerably improve over previous selection techniques, reaching comparable and in some instances better performance than oracle signal-based measures. | ['Stefano Squartini', 'Marco Matassoni', 'Alessio Brutti', 'Samuele Cornell'] | 2021-04-06 | null | null | null | null | ['distant-speech-recognition'] | ['speech'] | [ 3.78487110e-01 -2.79151559e-01 1.25342369e-01 -5.91979086e-01
-1.72485149e+00 -6.34022117e-01 7.68867135e-01 3.72003704e-01
-6.12574160e-01 5.83039999e-01 2.91097313e-01 -2.31172919e-01
-5.10952473e-01 -3.44584376e-01 -5.31225443e-01 -8.17955434e-01
-2.02711388e-01 4.56817299e-01 1.76754028e-01 2.10631024e-02
4.16335076e-01 5.85584402e-01 -1.78695202e+00 3.24001729e-01
4.30598289e-01 1.40768862e+00 2.87615776e-01 1.02587569e+00
1.30009860e-01 4.61275131e-01 -7.31840014e-01 -4.36598016e-03
3.06858629e-01 -4.31864321e-01 -2.92283237e-01 -6.31193146e-02
4.58690822e-01 1.49268597e-01 4.87357378e-02 7.85214663e-01
1.20377827e+00 -4.29564118e-02 6.45761609e-01 -6.64381981e-01
4.84385878e-01 7.39083111e-01 2.03302011e-01 2.14573458e-01
6.84738219e-01 -1.88977078e-01 1.29246438e+00 -1.25044823e+00
7.27979317e-02 9.05384243e-01 8.00467968e-01 4.64289896e-02
-1.25392771e+00 -5.14176130e-01 -1.29319251e-01 2.76822984e-01
-1.53156090e+00 -1.13664925e+00 7.13297665e-01 -4.22994375e-01
1.00337684e+00 4.58967716e-01 2.61554807e-01 9.96110439e-01
-4.64345396e-01 6.57745242e-01 1.43803573e+00 -6.62435591e-01
4.09160733e-01 3.83020878e-01 -5.44997267e-02 4.04242694e-01
-2.55237240e-03 2.09268481e-01 -1.00832617e+00 -2.37939700e-01
2.08463788e-01 -5.73994637e-01 -4.07463878e-01 -3.80187184e-01
-1.17412007e+00 5.36039114e-01 -4.74112853e-02 5.49620152e-01
-4.75309551e-01 -3.06306090e-02 2.53659666e-01 6.56693876e-01
2.93500960e-01 6.11546695e-01 -7.06536710e-01 -4.63464677e-01
-1.30468547e+00 -1.07229918e-01 1.07813287e+00 4.79371518e-01
6.10137165e-01 1.76617742e-01 -1.79377645e-01 1.16892564e+00
4.50916857e-01 4.92661834e-01 4.54720378e-01 -6.28236771e-01
5.96714795e-01 6.38870224e-02 -4.55158725e-02 -8.86051655e-01
-5.31010091e-01 -9.05348778e-01 -4.76134747e-01 1.71951234e-01
4.67253685e-01 -2.04109818e-01 -5.82726359e-01 1.53656566e+00
1.08124027e-02 2.42167100e-01 9.32345390e-02 7.72867441e-01
5.09353995e-01 3.91138822e-01 -4.04395282e-01 -3.79269600e-01
8.30407083e-01 -6.52685821e-01 -3.69199395e-01 1.97470728e-02
4.26438183e-01 -1.08287525e+00 7.22894311e-01 9.96629357e-01
-1.01606894e+00 -4.50952590e-01 -1.28187609e+00 7.32606173e-01
-2.89407909e-01 5.29879093e-01 1.40604004e-01 1.29527855e+00
-1.18040526e+00 6.87553883e-01 -4.51153815e-01 -1.89116403e-01
-6.14857227e-02 7.34791458e-01 -2.83475310e-01 8.77602771e-02
-9.31241751e-01 5.75564563e-01 2.32619941e-01 1.59673646e-01
-9.90073323e-01 -3.13632935e-01 -4.02139544e-01 -3.07367812e-03
1.95698097e-01 -3.53297472e-01 1.20384037e+00 -9.54389870e-01
-2.10374761e+00 5.43237507e-01 -1.67339951e-01 -5.88041544e-01
5.93086660e-01 -1.21599197e-01 -6.39013827e-01 5.23319654e-02
-3.24025959e-01 -1.74914449e-02 1.20716405e+00 -1.12880421e+00
-6.89216852e-01 -2.33833492e-01 -3.60963821e-01 1.88024342e-01
-4.69263971e-01 -3.48652788e-02 -4.86279100e-01 -6.43423855e-01
4.20207381e-01 -7.92906046e-01 -1.84177369e-01 -4.61314738e-01
-4.84352648e-01 8.00184980e-02 2.15569869e-01 -6.04249299e-01
1.23459017e+00 -2.05218816e+00 8.37262794e-02 7.92396963e-01
-1.97730616e-01 4.09019947e-01 -7.24656954e-02 4.65080649e-01
-9.04560536e-02 -3.35593045e-01 -1.03187814e-01 -3.34865183e-01
1.48198724e-01 -1.13888800e-01 1.46840751e-01 7.35637665e-01
-1.01527840e-01 5.49042709e-02 -5.14714003e-01 -3.39753777e-01
5.01289248e-01 5.20642221e-01 -5.49490273e-01 3.48490953e-01
1.77453980e-01 4.50356871e-01 -2.77848274e-01 5.85957229e-01
4.04795438e-01 1.41332716e-01 2.23942548e-01 -1.03087267e-02
-1.01759389e-01 6.23280823e-01 -1.60128057e+00 1.57770860e+00
-9.00346696e-01 8.14924121e-01 3.36248010e-01 -1.27111375e+00
1.12426031e+00 6.25345528e-01 2.18406200e-01 -6.17415965e-01
1.09777130e-01 8.87835503e-01 1.77276671e-01 -3.16023648e-01
5.83204255e-02 -1.29329795e-02 -3.15327011e-02 7.16390461e-02
1.34495050e-01 1.13390619e-02 -1.58402577e-01 -2.94949025e-01
1.24108624e+00 -3.31586748e-01 3.74885201e-01 -1.03045881e-01
1.07368994e+00 -5.29475570e-01 1.53919056e-01 1.02866912e+00
1.08540662e-01 7.50172317e-01 2.47332558e-01 3.75579953e-01
-6.41279280e-01 -9.79524314e-01 -2.05896959e-01 1.18458188e+00
-2.73582339e-01 -2.85748661e-01 -6.75910234e-01 -3.98467541e-01
-1.88830584e-01 4.45308566e-01 -1.96152166e-01 2.63969630e-01
-4.28484827e-01 -5.47056377e-01 6.35568500e-01 1.79798126e-01
4.92823422e-02 -6.53370380e-01 -2.73826897e-01 4.17769283e-01
1.47462726e-01 -9.95390534e-01 2.68638600e-02 7.03846157e-01
-9.33979213e-01 -6.60479307e-01 -9.28695977e-01 -5.52064717e-01
4.18641418e-01 -7.12318942e-02 9.06796396e-01 -1.87920004e-01
7.44958743e-02 5.42692959e-01 -5.37659883e-01 -9.64279547e-02
-3.71350735e-01 3.08407038e-01 3.37435342e-02 4.15509522e-01
3.44373882e-02 -6.61376178e-01 -5.01534998e-01 4.75157082e-01
-3.13925236e-01 -7.02185988e-01 9.44499314e-01 1.05344713e+00
4.41742390e-01 7.20250085e-02 5.49365819e-01 -5.69988787e-01
7.10625827e-01 -1.94032118e-01 -6.23290420e-01 5.35269454e-02
-8.74632359e-01 3.37242067e-01 6.51278794e-01 -1.56763390e-01
-6.95389450e-01 5.32590628e-01 -6.60727918e-01 -1.53735518e-01
-2.39537418e-01 5.57586908e-01 -6.02235794e-01 -4.55434382e-01
6.62624657e-01 2.81661391e-01 -3.38627607e-01 -8.55887234e-01
5.13416827e-02 1.01388550e+00 3.94726306e-01 -4.61544633e-01
6.14844799e-01 8.78937393e-02 -8.26569945e-02 -1.03493011e+00
-3.25053811e-01 -9.67729568e-01 -4.70903426e-01 -1.79254070e-01
2.81604886e-01 -9.55711186e-01 -5.65493286e-01 4.09754038e-01
-8.63566101e-01 9.17320251e-02 7.41278157e-02 8.74247551e-01
-4.83981907e-01 1.99539334e-01 -2.37670615e-02 -1.10834527e+00
-9.44303125e-02 -1.16863048e+00 1.11445045e+00 -2.04652444e-01
-6.57846704e-02 -8.15110624e-01 -6.44484833e-02 3.82833809e-01
5.96395433e-01 -1.55754998e-01 4.65694666e-01 -1.17372108e+00
-4.99756068e-01 -5.91636777e-01 1.51582971e-01 5.88024497e-01
-2.74038553e-01 -3.80819798e-01 -1.45562661e+00 -3.45332593e-01
-7.81394765e-02 2.99862325e-02 9.98141646e-01 2.65195012e-01
1.02200902e+00 -8.97725746e-02 -1.27119184e-01 4.19175029e-01
1.36526299e+00 1.45430699e-01 4.49736834e-01 2.85059839e-01
3.02773833e-01 6.09441340e-01 4.92504030e-01 7.13053286e-01
-1.70845926e-01 1.08428288e+00 1.46895558e-01 1.06766768e-01
-1.51610315e-01 -7.33675808e-02 6.43023431e-01 1.11716473e+00
1.96399853e-01 -3.03455263e-01 -9.67916071e-01 4.35739189e-01
-1.30196452e+00 -8.13610852e-01 -1.35080934e-01 2.50551605e+00
4.87015575e-01 3.62114638e-01 2.25410312e-01 8.72134686e-01
5.52709341e-01 -5.17642833e-02 -2.29587227e-01 -2.07290083e-01
-3.39509070e-01 4.70939904e-01 6.79320693e-01 7.69579291e-01
-9.80627716e-01 3.52155924e-01 5.97459793e+00 1.08633935e+00
-1.36039364e+00 2.94016302e-01 4.78458107e-01 -9.56587940e-02
-3.35763581e-02 -7.57837966e-02 -6.86612070e-01 2.49663234e-01
1.22316694e+00 2.31459290e-01 4.27433938e-01 6.23124719e-01
1.84977919e-01 -8.74547064e-02 -1.29679060e+00 1.37798309e+00
3.04704756e-01 -9.30452347e-01 -4.85405147e-01 -2.81681749e-03
3.69600147e-01 2.21433535e-01 2.60770887e-01 5.70614748e-02
-3.70480835e-01 -1.08754170e+00 8.71564507e-01 5.65620542e-01
6.38669908e-01 -5.99113226e-01 9.07397211e-01 2.61062652e-01
-1.19909239e+00 -2.73267359e-01 5.96037880e-02 2.84107178e-02
-4.15634625e-02 8.20384860e-01 -1.22836375e+00 4.12457764e-01
5.10266662e-01 3.10159892e-01 -5.32441020e-01 1.49206865e+00
-6.03932589e-02 9.85696137e-01 -6.18175447e-01 -3.15821677e-01
1.58277363e-01 1.31070212e-01 8.63934398e-01 1.54354703e+00
7.17969775e-01 -3.91147822e-01 -1.02161013e-01 4.97061312e-02
2.57519990e-01 4.52131450e-01 -4.58790451e-01 1.35167152e-01
4.52572882e-01 9.44114447e-01 -6.49083018e-01 -1.04426168e-01
-2.76117742e-01 7.42394984e-01 -8.54040980e-02 1.71114400e-01
-4.20483410e-01 -4.82077569e-01 3.43083709e-01 7.27062002e-02
6.97264969e-01 -2.40686283e-01 -2.37163126e-01 -7.55507529e-01
2.09664702e-01 -1.13740051e+00 6.71751276e-02 -2.67683148e-01
-8.82206500e-01 8.14748347e-01 -3.59521091e-01 -1.42356300e+00
-5.23225367e-01 -8.78144741e-01 -2.01424181e-01 8.45406711e-01
-1.26950812e+00 -6.22602701e-01 5.26895784e-02 5.36386609e-01
5.37291765e-01 -6.52996004e-01 9.30475712e-01 7.25041509e-01
-2.44454712e-01 9.56215441e-01 6.94279134e-01 -1.42129526e-01
6.41809225e-01 -1.27252662e+00 -1.57437012e-01 7.49893904e-01
6.28882945e-01 4.45637763e-01 1.03837562e+00 -1.86681207e-02
-1.50250852e+00 -6.04307175e-01 1.03953600e+00 -3.63566786e-01
5.22260904e-01 -5.15160084e-01 -4.81885642e-01 -2.16443226e-01
5.75857237e-02 -8.63822475e-02 8.98010373e-01 4.77741808e-01
-2.56567359e-01 -5.75272858e-01 -7.50921667e-01 4.94155567e-03
7.54736781e-01 -8.02381575e-01 -3.07009488e-01 2.56410807e-01
-1.99502744e-02 -1.27026916e-01 -7.68293738e-01 2.81435937e-01
7.64763772e-01 -1.26754248e+00 9.76393580e-01 2.25929588e-01
-2.72250026e-01 -3.58863711e-01 -6.48192763e-01 -1.27421904e+00
1.66681319e-01 -8.82215142e-01 2.28090003e-01 1.33888447e+00
1.12317133e+00 -7.10765362e-01 9.72727060e-01 5.63864373e-02
-5.99133447e-02 -6.32939994e-01 -1.31980467e+00 -9.33717370e-01
-2.99291909e-01 -9.19882536e-01 2.99193591e-01 5.27267396e-01
-1.62249193e-01 4.61270869e-01 -4.77819353e-01 3.03877950e-01
4.79930669e-01 -1.46976441e-01 7.20279813e-01 -1.30145824e+00
-9.39739704e-01 -6.52851164e-01 -7.15821624e-01 -1.14541376e+00
-7.32142404e-02 -8.07210386e-01 3.08900982e-01 -1.05072558e+00
-3.93522710e-01 -6.62902832e-01 -6.67055726e-01 2.63087824e-02
2.34868869e-01 1.60958901e-01 3.63909714e-02 1.72882900e-02
-7.06741273e-01 2.57911295e-01 3.82713050e-01 -1.48301095e-01
-1.57800481e-01 6.16263807e-01 -3.39646190e-01 5.02706587e-01
7.71667540e-01 -5.21087050e-01 -2.39750981e-01 -1.19369172e-01
2.32871875e-01 4.00839955e-01 1.31927788e-01 -1.61631083e+00
4.76927906e-01 6.57029450e-01 2.52063811e-01 -4.48631406e-01
6.72203481e-01 -1.02508414e+00 1.53086752e-01 3.12969267e-01
-6.31641567e-01 -1.32462084e-01 -7.69682452e-02 7.56012082e-01
-5.85762203e-01 -4.25893545e-01 6.28196895e-01 2.55597681e-01
-4.90630507e-01 -2.07768634e-01 -5.20479798e-01 -3.43595952e-01
5.54166317e-01 -3.77033144e-01 3.98146302e-01 -7.07064033e-01
-9.69161868e-01 -3.27192158e-01 2.75622141e-02 1.20997407e-01
5.19830644e-01 -9.12017226e-01 -7.25926876e-01 2.31747568e-01
1.16643295e-01 -7.24233985e-01 -1.38786241e-01 9.05519903e-01
-2.68733442e-01 7.03043520e-01 2.26598457e-01 -8.23364854e-01
-1.63577342e+00 1.13626853e-01 3.97968471e-01 -1.23926856e-01
-3.60389077e-03 1.15389395e+00 -4.91970092e-01 -3.85021418e-01
7.80502558e-01 -1.88362092e-01 -3.70439142e-01 3.59370083e-01
1.82674021e-01 4.15624708e-01 6.73284054e-01 -7.55669236e-01
-5.96928596e-01 7.14551568e-01 3.36704820e-01 -7.23951876e-01
1.18380070e+00 -1.59573153e-01 2.71293342e-01 5.43313622e-01
1.53103781e+00 4.67220575e-01 -7.85582542e-01 -2.69872397e-01
4.46095228e-01 -5.13808966e-01 2.84469068e-01 -9.68011796e-01
-8.99490595e-01 1.06171298e+00 1.26511514e+00 2.54230708e-01
1.40092695e+00 -1.62315175e-01 2.87263215e-01 6.66861594e-01
5.15624404e-01 -1.27261889e+00 -2.48485692e-02 3.79067153e-01
7.84510255e-01 -1.25461638e+00 -6.23904951e-02 -2.97549248e-01
-2.33805895e-01 1.12654746e+00 -1.92499477e-02 1.43819332e-01
9.20678854e-01 2.41950110e-01 2.76709974e-01 9.46877748e-02
-6.03784680e-01 -3.76138568e-01 3.96266639e-01 5.76230645e-01
6.33179903e-01 2.31448159e-01 -3.62604767e-01 4.50738639e-01
-3.33166569e-01 -5.26737154e-01 -3.01581547e-02 6.13238394e-01
-5.58499932e-01 -1.59255636e+00 -7.58371115e-01 6.25977874e-01
-5.51291585e-01 -2.16117322e-01 -5.47659755e-01 3.12557787e-01
7.45844841e-02 1.44175935e+00 -4.26559657e-01 -6.71766937e-01
4.55528706e-01 4.47852761e-02 3.15106034e-01 -4.37191308e-01
-8.33852291e-01 4.60194707e-01 6.02293491e-01 -4.81128126e-01
-3.87073725e-01 -1.14588022e+00 -6.47955835e-01 2.68502206e-01
-7.75116861e-01 4.01722074e-01 1.17903447e+00 7.95910537e-01
2.57985055e-01 4.72478926e-01 9.56575871e-01 -9.48244572e-01
-6.98287010e-01 -8.14199030e-01 -5.95205307e-01 -1.98445208e-02
5.12268364e-01 -6.05707109e-01 -5.42023122e-01 -2.66363710e-01] | [14.957423210144043, 5.6774115562438965] |
b7089136-571a-4564-a19f-68ee10310a10 | vssa-net-vertical-spatial-sequence-attention | 1905.01583 | null | https://arxiv.org/abs/1905.01583v1 | https://arxiv.org/pdf/1905.01583v1.pdf | VSSA-NET: Vertical Spatial Sequence Attention Network for Traffic Sign Detection | Although traffic sign detection has been studied for years and great progress has been made with the rise of deep learning technique, there are still many problems remaining to be addressed. For complicated real-world traffic scenes, there are two main challenges. Firstly, traffic signs are usually small size objects, which makes it more difficult to detect than large ones; Secondly, it is hard to distinguish false targets which resemble real traffic signs in complex street scenes without context information. To handle these problems, we propose a novel end-to-end deep learning method for traffic sign detection in complex environments. Our contributions are as follows: 1) We propose a multi-resolution feature fusion network architecture which exploits densely connected deconvolution layers with skip connections, and can learn more effective features for the small size object; 2) We frame the traffic sign detection as a spatial sequence classification and regression task, and propose a vertical spatial sequence attention (VSSA) module to gain more context information for better detection performance. To comprehensively evaluate the proposed method, we do experiments on several traffic sign datasets as well as the general object detection dataset and the results have shown the effectiveness of our proposed method. | ['Qi. Wang', 'IEEE', 'Senior Member', 'Student Member', 'Zhitong Xiong', 'Yuan Yuan'] | 2019-05-05 | null | null | null | null | ['traffic-sign-detection'] | ['computer-vision'] | [ 4.09622937e-01 -8.32299173e-01 9.30321440e-02 -3.88942599e-01
-2.82709867e-01 -2.30892980e-03 4.79796410e-01 -8.21936011e-01
-4.57027823e-01 6.01088941e-01 7.87639022e-02 -3.43550026e-01
-4.69041131e-02 -4.31156546e-01 -5.12550294e-01 -8.87180984e-01
2.19584033e-01 -1.88543685e-02 1.02800250e+00 -1.82105035e-01
5.19194365e-01 4.39313024e-01 -1.95488703e+00 3.81709784e-01
1.14267194e+00 1.18259883e+00 4.91747409e-01 5.77628374e-01
-2.22936496e-01 1.18481708e+00 -5.53371787e-01 -2.37280518e-01
3.29683721e-01 -5.20247638e-01 -3.27675998e-01 1.56434551e-01
8.68843615e-01 -5.85459769e-01 -6.11785233e-01 1.09444273e+00
6.68770730e-01 2.21141756e-01 4.81476396e-01 -1.35136497e+00
-7.02403486e-01 3.41808870e-02 -6.71839774e-01 5.64480305e-01
-2.29204834e-01 7.35768139e-01 7.51723409e-01 -8.01608622e-01
3.60197932e-01 1.35622239e+00 5.03857732e-01 5.15405059e-01
-6.31856322e-01 -1.01757133e+00 3.36672157e-01 8.59954059e-01
-9.82120156e-01 -4.91443247e-01 7.91695178e-01 -3.95385206e-01
6.06571317e-01 8.23262036e-02 5.63588917e-01 9.48993146e-01
-1.54172927e-01 1.30552757e+00 1.29538906e+00 -7.83682317e-02
-2.52350062e-01 -1.84557155e-01 3.85956198e-01 6.98774815e-01
4.82038170e-01 3.65589708e-01 2.78172866e-02 3.44649285e-01
7.75623739e-01 1.50272533e-01 -3.08427572e-01 -3.32018584e-01
-1.31397402e+00 5.02426863e-01 6.93717718e-01 4.69924718e-01
-4.47615445e-01 4.91378248e-01 5.05473018e-01 2.86127795e-02
-9.74632949e-02 -2.72295326e-01 -2.62172461e-01 -2.62633055e-01
-6.43622756e-01 3.04365873e-01 2.32667282e-01 7.20832169e-01
3.62746626e-01 3.15213829e-01 -3.07874441e-01 9.44036603e-01
4.05413955e-01 9.32360411e-01 4.45537537e-01 -5.73423207e-01
7.43739009e-01 3.39418292e-01 1.03250161e-01 -1.11342311e+00
-3.87171209e-01 -4.40305710e-01 -7.00732708e-01 5.12572527e-01
6.92375958e-01 -4.17560041e-02 -1.10473001e+00 1.51328063e+00
1.82749584e-01 7.62158453e-01 -1.73233211e-01 1.49162543e+00
1.01243258e+00 5.32625318e-01 2.98630744e-01 2.36425728e-01
1.50340819e+00 -1.20419240e+00 -7.05402136e-01 -3.06704104e-01
4.26815122e-01 -7.48368204e-01 8.95782590e-01 1.62305504e-01
-7.97052443e-01 -8.83446157e-01 -9.55855131e-01 -2.03794211e-01
-4.23515111e-01 4.46850032e-01 7.45334089e-01 6.60749257e-01
-6.45845592e-01 6.00689873e-02 -5.81103206e-01 -3.89272094e-01
8.01899672e-01 9.02194679e-02 -3.79002132e-02 -4.29743409e-01
-1.09901989e+00 9.76920843e-01 3.60825866e-01 6.93327367e-01
-6.00926638e-01 -2.34396994e-01 -6.06039405e-01 5.82471713e-02
5.18416643e-01 -6.15760803e-01 1.09297168e+00 -1.00956297e+00
-1.26917446e+00 6.22269988e-01 -3.48728448e-01 -2.05173910e-01
6.71386480e-01 -5.55488244e-02 -7.69598007e-01 -1.04413554e-01
2.99833361e-02 4.04261470e-01 8.02646995e-01 -1.18469322e+00
-1.07557881e+00 -8.32122788e-02 -1.75205827e-01 -2.86105629e-02
2.77500618e-02 5.49450576e-01 -4.22226667e-01 -6.56108975e-01
-3.66643183e-02 -6.71428263e-01 -9.54775512e-02 2.86844850e-01
-1.74586520e-01 -3.52800667e-01 1.23029709e+00 -7.13216305e-01
1.08794224e+00 -2.06289697e+00 -3.71277839e-01 -4.79967743e-02
2.64347166e-01 1.09288239e+00 -4.16559994e-01 -1.34281456e-01
-4.78217453e-02 -2.75095761e-01 -3.02649289e-01 9.84194875e-02
1.00600310e-02 2.00753585e-01 -3.62552315e-01 2.16975018e-01
3.33577842e-01 1.16127872e+00 -9.17348385e-01 -5.45258880e-01
5.31094491e-01 3.90585214e-01 -9.17548835e-02 -1.46414593e-01
7.13995248e-02 4.34134811e-01 -6.50965750e-01 8.80860925e-01
9.89085078e-01 -1.93179220e-01 -4.99873340e-01 -2.99844027e-01
-3.50163549e-01 4.67405580e-02 -1.33843458e+00 1.03011060e+00
-1.39096871e-01 9.77566957e-01 -2.07082946e-02 -1.11436832e+00
8.40462506e-01 -7.78061524e-02 2.25631893e-01 -1.14052248e+00
2.69923627e-01 4.89291519e-01 4.10010278e-01 -1.01488185e+00
2.21332759e-01 -2.21709028e-01 3.43383729e-01 3.52850616e-01
-3.83316159e-01 3.35018247e-01 4.56804931e-01 -7.90146962e-02
1.04028165e+00 -1.61140648e-04 -1.09947309e-01 3.64729583e-01
9.70752954e-01 -1.58887506e-01 8.86291265e-01 6.11041903e-01
-7.81213045e-01 4.70230430e-01 3.17727834e-01 -6.23564780e-01
-8.07745039e-01 -9.27335680e-01 -2.56660953e-02 9.20636475e-01
5.26943803e-01 2.04055637e-01 -1.31353319e-01 -8.45594406e-01
1.03710666e-01 4.23597336e-01 -3.98778707e-01 -8.09235848e-04
-1.17736530e+00 -6.92641020e-01 6.16541982e-01 9.04585361e-01
1.18533695e+00 -1.35616517e+00 -6.06680751e-01 2.15022974e-02
-1.82250619e-01 -1.32203329e+00 -6.29716814e-01 -3.48753542e-01
-5.74544311e-01 -1.44353199e+00 -9.89298642e-01 -1.08422267e+00
5.59049785e-01 8.81391108e-01 5.59285522e-01 2.46306285e-01
-5.29656112e-01 -3.49995680e-02 -3.56185079e-01 -4.04998958e-01
-2.01297123e-02 -4.11034316e-01 -2.75511652e-01 4.28187966e-01
5.55114388e-01 -1.64755121e-01 -7.70093441e-01 7.19421268e-01
-7.05564320e-01 4.53109741e-02 1.03366113e+00 9.36819553e-01
6.37212545e-02 -1.14975959e-01 5.93934655e-01 -3.25808823e-01
6.01334035e-01 3.56224598e-03 -7.95940876e-01 2.96358436e-01
-1.04607403e-01 -5.59702329e-02 4.13412392e-01 -4.08421904e-01
-1.21961820e+00 -9.22398120e-02 -1.62700161e-01 -2.79142171e-01
-3.10931623e-01 7.45036826e-02 -1.73077390e-01 -3.82445067e-01
2.69543260e-01 5.28440058e-01 4.64196987e-02 -4.03712094e-01
3.29247385e-01 8.30568731e-01 4.97055769e-01 -2.35412210e-01
8.06365252e-01 5.31594574e-01 4.61751521e-02 -8.55749547e-01
-6.02585912e-01 -5.23920774e-01 -5.24699330e-01 -4.97756094e-01
8.24816525e-01 -5.78335464e-01 -1.13097799e+00 8.65863442e-01
-1.03268969e+00 -2.05590755e-01 4.48749624e-02 6.02852404e-01
-3.52052122e-01 8.37518215e-01 -3.03864926e-01 -8.11371744e-01
-1.94979995e-01 -1.19321203e+00 9.73508358e-01 4.34046447e-01
3.90256822e-01 -6.55932724e-01 -1.76315442e-01 5.81903398e-01
8.17502618e-01 2.27807797e-02 6.34798765e-01 -2.74362564e-01
-1.19449735e+00 -2.69428849e-01 -1.05032718e+00 5.25497079e-01
-7.86006898e-02 -8.83771107e-02 -7.83517599e-01 1.27727225e-01
-3.71344745e-01 -3.30125421e-01 1.41938829e+00 4.50427324e-01
9.84229565e-01 1.17575772e-01 -4.99178261e-01 6.22406602e-01
9.94053721e-01 4.11078215e-01 7.44386792e-01 4.07891959e-01
7.33809412e-01 3.82064611e-01 6.77799582e-01 9.68090817e-03
3.28997433e-01 7.10315704e-01 2.15222746e-01 -2.28495866e-01
-7.27368295e-01 -1.70116886e-01 2.99310297e-01 6.63386106e-01
-2.57703364e-01 9.96048097e-03 -6.17133439e-01 5.00711083e-01
-2.03193736e+00 -1.50191998e+00 -5.07433534e-01 1.85574496e+00
6.58961162e-02 7.62373731e-02 3.45004767e-01 2.75242388e-01
8.66797388e-01 2.40246147e-01 -6.82161450e-01 -1.42796489e-03
-4.27557617e-01 -1.27688989e-01 4.87717867e-01 -1.04331234e-02
-1.25694561e+00 1.08769190e+00 5.78653669e+00 8.23129833e-01
-1.55665278e+00 1.40890768e-02 1.50242448e-01 1.59751832e-01
2.36420006e-01 -1.46823078e-01 -7.30246425e-01 9.01759684e-01
2.42577210e-01 2.15372548e-01 1.68108359e-01 7.84621596e-01
4.02969182e-01 3.46384458e-02 -5.18661082e-01 1.33592272e+00
1.73416346e-01 -1.06992638e+00 -1.45315230e-01 -1.15929574e-01
5.91822982e-01 2.86759675e-01 2.40581647e-01 5.69016576e-01
2.42796153e-01 -7.86710739e-01 5.51831663e-01 6.55560434e-01
5.07462263e-01 -3.42717916e-01 8.08629632e-01 3.82597953e-01
-1.43148220e+00 -4.89531845e-01 -5.14782548e-01 -4.11351211e-02
6.30726695e-01 3.91153336e-01 -4.72956389e-01 3.89360100e-01
4.04483080e-01 1.01647031e+00 -6.68429673e-01 2.01199126e+00
-4.43386823e-01 5.92597425e-01 -2.74230123e-01 -3.75346750e-01
3.69431466e-01 -2.61835873e-01 4.30005640e-01 1.28309131e+00
2.12618560e-01 -3.00338697e-02 8.72549415e-02 9.32450175e-01
1.89543068e-01 -4.31325436e-02 -3.51892769e-01 6.12081140e-02
8.03500563e-02 1.10384858e+00 -7.30031490e-01 -3.86500627e-01
-6.73948050e-01 9.35537875e-01 5.73748313e-02 6.69047236e-01
-1.12319183e+00 -7.49119163e-01 7.38012493e-01 -6.81414530e-02
8.38202059e-01 -1.92235500e-01 -2.00672135e-01 -1.39021671e+00
3.06147009e-01 -7.34146535e-01 3.39946538e-01 -1.00756693e+00
-1.20908248e+00 2.87429988e-01 -4.96134669e-01 -1.51332200e+00
2.74407238e-01 -1.02580011e+00 -8.36101115e-01 7.03890324e-01
-1.97541440e+00 -1.24290097e+00 -6.43566728e-01 4.93273407e-01
5.33510745e-01 -1.52609989e-01 1.52722970e-02 8.21578443e-01
-7.27265537e-01 5.26434064e-01 2.18113244e-01 5.46509802e-01
6.43707395e-01 -6.11861110e-01 3.90342027e-01 1.06692994e+00
-2.80258805e-02 3.42888176e-01 1.55148447e-01 -6.08588457e-01
-1.14860380e+00 -9.67278779e-01 8.78651202e-01 -3.32272142e-01
6.19423270e-01 -7.62477219e-02 -7.70272076e-01 4.07708079e-01
-2.28615537e-01 3.62827182e-01 1.67565141e-02 -2.55993962e-01
-3.81228417e-01 -3.82806838e-01 -8.35148692e-01 5.17413259e-01
1.36083555e+00 -2.10203543e-01 -7.60098040e-01 4.93837968e-02
1.24193773e-01 -1.16465010e-01 2.17519924e-02 5.28291762e-01
8.01014304e-01 -7.75071383e-01 1.05737782e+00 -6.65390909e-01
2.12923009e-02 -9.35672760e-01 9.42411572e-02 -9.11817431e-01
-3.34114015e-01 -1.77735120e-01 -1.58528648e-02 1.02820182e+00
2.04413645e-02 -8.24902892e-01 7.23196745e-01 4.13492471e-01
-2.75057077e-01 -6.13562226e-01 -9.84145224e-01 -9.56968784e-01
-2.39087313e-01 -5.15517354e-01 4.74392802e-01 5.98089576e-01
-4.16129678e-01 2.59246647e-01 -6.96214318e-01 -9.84924585e-02
6.52398646e-01 3.99184883e-01 1.05250633e+00 -1.28020930e+00
1.95270516e-02 -9.92469907e-01 -9.64702606e-01 -1.66054249e+00
-8.26264322e-02 -6.42640471e-01 2.09446490e-01 -1.68364692e+00
1.57605693e-01 -4.11936134e-01 -4.10583466e-01 3.06564659e-01
-4.69012618e-01 1.12073518e-01 4.12901849e-01 9.62347463e-02
-9.73668933e-01 7.71376312e-01 1.64366698e+00 -2.77283072e-01
2.63056874e-01 2.09260598e-01 -5.23122907e-01 6.17034912e-01
3.80631864e-01 -2.60897763e-02 -4.91269007e-02 -4.89463180e-01
-2.60318220e-01 -2.75517374e-01 6.71971798e-01 -1.09009933e+00
3.81411970e-01 -7.08648935e-02 4.64739978e-01 -1.05331600e+00
1.71172768e-01 -8.21540415e-01 -5.94746053e-01 6.63172781e-01
-3.11650913e-02 -4.28072363e-01 5.58056086e-02 6.98445559e-01
-3.46062422e-01 2.59181135e-03 8.68032396e-01 1.83484212e-01
-1.45560336e+00 2.91420490e-01 -3.68932039e-01 2.50275414e-02
1.03704655e+00 -5.51035821e-01 -5.11306882e-01 -1.99227750e-01
-2.82111079e-01 5.40060282e-01 2.84827984e-05 7.20321774e-01
8.36021543e-01 -1.46954453e+00 -7.68671632e-01 4.32055384e-01
2.47388154e-01 -2.77171820e-01 5.30027866e-01 1.10516274e+00
-6.53501511e-01 7.44206369e-01 -3.52802604e-01 -6.08304918e-01
-1.41618586e+00 6.38847232e-01 3.96471679e-01 2.50651330e-01
-7.87487447e-01 8.05462420e-01 2.63436407e-01 -1.50045961e-01
3.72607797e-01 -7.23720789e-01 -2.95003623e-01 -2.19057351e-01
8.98511827e-01 5.39203227e-01 -3.17623377e-01 -8.73528540e-01
-3.11995417e-01 1.07132471e+00 -8.79930556e-02 3.87872756e-01
1.15049458e+00 -1.41735598e-01 2.95743376e-01 4.83339727e-02
9.67042267e-01 -3.92538130e-01 -1.36082292e+00 -4.93885964e-01
-1.95376016e-02 -9.62052703e-01 -8.09259415e-02 -8.61840725e-01
-1.30714846e+00 1.20579243e+00 8.23315918e-01 -1.62984639e-01
1.03864872e+00 -2.60029733e-01 1.23666728e+00 5.53499401e-01
1.35156110e-01 -1.05034125e+00 9.06491801e-02 7.18514502e-01
7.53252625e-01 -1.60132444e+00 -3.46882254e-01 -3.62466693e-01
-7.94621170e-01 1.17194152e+00 7.12217271e-01 -1.91514939e-01
3.88411164e-01 -1.05633929e-01 2.27037042e-01 -1.30197868e-01
-3.93944621e-01 -1.02820945e+00 3.74873579e-01 6.80005312e-01
1.76211864e-01 -2.15903580e-01 -3.98606062e-01 4.60655540e-01
3.98107767e-01 4.43053484e-01 1.39903858e-01 7.54587710e-01
-7.99389541e-01 -7.97919691e-01 -4.89751607e-01 4.10256833e-01
-3.11599504e-02 8.01061243e-02 -2.96690434e-01 8.01603079e-01
3.86496842e-01 8.90707791e-01 -1.30783305e-01 -2.43725717e-01
6.06416643e-01 1.27210483e-01 3.97192210e-01 -5.75051978e-02
-1.34429872e-01 -1.11334488e-01 -3.02874367e-03 -4.91200954e-01
-3.65616113e-01 -8.09809387e-01 -1.12707567e+00 -1.28725648e-01
-5.44191718e-01 -3.11351985e-01 6.03000522e-01 1.08955014e+00
2.78535187e-01 6.78835392e-01 3.48057985e-01 -7.51990199e-01
-5.00913858e-01 -8.82804632e-01 -4.00674045e-01 7.49316752e-01
4.63601738e-01 -1.14063025e+00 -1.70637459e-01 -1.65953338e-01] | [7.988999366760254, -0.8294830322265625] |
27d8d630-ae3b-4767-b07f-d7aa192dd215 | efficient-reuse-of-structured-and | null | null | https://aclanthology.org/L14-1235 | https://aclanthology.org/L14-1235.pdf | Efficient Reuse of Structured and Unstructured Resources for Ontology Population | We study the problem of ontology population for a domain ontology and present solutions based on semi-automatic techniques. A domain ontology for an organization, often consists of classes whose instances are either specific to, or independent of the organization. E.g. in an academic domain ontology, classes like Professor, Department could be organization (university) specific, while Conference, Programming languages are organization independent. This distinction allows us to leverage data sources both―within the organization and those in the Internet ― to extract entities and populate an ontology. We propose techniques that build on those for open domain IE. Together with user input, we show through comprehensive evaluation, how these semi-automatic techniques achieve high precision. We experimented with the academic domain and built an ontology comprising of over 220 classes. Intranet documents from five universities formed our organization specific corpora and we used open domain knowledge bases like Wikipedia, Linked Open Data, and web pages from the Internet as the organization independent data sources. The populated ontology that we built for one of the universities comprised of over 75,000 instances. We adhere to the semantic web standards and tools and make the resources available in the OWL format. These could be useful for applications such as information extraction, text annotation, and information retrieval. | ['Ganesh Ramakrishnan', 'Chetana Gavankar', 'Ashish Kulkarni'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['text-annotation'] | ['natural-language-processing'] | [-3.48432779e-01 6.96883202e-01 -3.64777595e-01 -1.29262000e-01
-3.43637913e-01 -1.07376981e+00 6.05065584e-01 6.26040041e-01
-5.05241454e-01 1.04219365e+00 1.52242467e-01 -2.71953464e-01
-6.62708461e-01 -1.30864298e+00 -4.14872736e-01 1.29967496e-01
2.03239784e-01 8.83683205e-01 6.76511884e-01 -5.81538320e-01
1.99286759e-01 3.59053105e-01 -1.91727149e+00 3.58720571e-01
9.20156538e-01 1.03552926e+00 6.78221583e-02 1.04969725e-01
-1.04273629e+00 8.11967850e-01 -6.85361564e-01 -5.34837842e-01
1.28540829e-01 3.83682191e-01 -1.47706437e+00 1.00282133e-01
1.53335541e-01 2.74637461e-01 -1.34999290e-01 1.19352376e+00
1.38647676e-01 6.40074722e-03 4.50391948e-01 -1.48896062e+00
-7.27764070e-01 7.66494930e-01 1.81184918e-01 -7.82499090e-02
6.55906379e-01 -5.81805229e-01 1.04440391e+00 -4.61131662e-01
1.51610446e+00 9.74028826e-01 3.26228380e-01 2.93699086e-01
-6.89022243e-01 -4.83254671e-01 -1.58939615e-01 1.54348731e-01
-1.42669666e+00 -3.76981556e-01 3.69582832e-01 -6.17533743e-01
8.00641179e-01 2.68001735e-01 3.05540532e-01 6.60612345e-01
-8.06885436e-02 3.93887945e-02 8.38147640e-01 -1.03581893e+00
2.03736916e-01 9.96075749e-01 8.80560338e-01 7.40267396e-01
9.57091153e-01 -6.40042722e-01 -2.92842358e-01 -3.75031441e-01
5.52993059e-01 -2.20654279e-01 -9.56067070e-02 -5.90614974e-01
-9.26361740e-01 4.93651092e-01 1.36698456e-02 8.51873636e-01
-1.83453739e-01 -4.51021105e-01 5.13193309e-01 4.39704657e-01
4.04651687e-02 6.45700336e-01 -8.03467155e-01 1.98893338e-01
-2.52223432e-01 3.49402785e-01 1.57941890e+00 1.66277432e+00
1.12294531e+00 -6.61775649e-01 5.65949857e-01 9.60625648e-01
4.73539442e-01 -2.85394229e-02 5.91512263e-01 -1.31427145e+00
4.75048959e-01 1.31500435e+00 4.73618954e-01 -7.30506063e-01
-2.44859099e-01 -9.31884721e-02 2.44095344e-02 -9.59317014e-02
5.28371572e-01 4.70897630e-02 -6.22717857e-01 1.17585325e+00
4.44449872e-01 -4.06450063e-01 5.79803824e-01 2.99189895e-01
1.34266281e+00 1.42496988e-01 2.48170763e-01 -1.31257951e-01
2.04104352e+00 -7.30592906e-01 -1.22931445e+00 -1.16021419e-02
8.43708932e-01 -7.22752929e-01 5.18646955e-01 1.68751776e-01
-8.61097097e-01 -2.67110914e-01 -1.06481647e+00 -2.41173327e-01
-1.57608306e+00 -3.04859370e-01 5.83784640e-01 5.01799941e-01
-7.53309190e-01 5.28513074e-01 -3.38063359e-01 -1.04523849e+00
2.48147026e-01 2.12790161e-01 -7.70369291e-01 3.48513722e-02
-1.48530614e+00 9.49941039e-01 1.12145483e+00 -8.69657457e-01
-2.67244458e-01 -3.58242810e-01 -8.56486261e-01 1.29163995e-01
8.39738607e-01 -6.70622170e-01 1.05739939e+00 -7.13346541e-01
-7.78330147e-01 1.48797357e+00 1.85921490e-01 -3.87039781e-01
-9.98457819e-02 1.61848515e-01 -1.10174549e+00 2.10212275e-01
5.77335417e-01 9.02345479e-02 -1.86357927e-02 -1.39795327e+00
-1.24673128e+00 -7.00064898e-01 6.61687672e-01 -2.75378913e-01
-9.27274346e-01 4.03498262e-01 -4.43738341e-01 -1.73968002e-01
1.15388021e-01 -7.11459935e-01 -6.09337501e-02 -4.58680183e-01
-1.76376343e-01 -4.48050201e-01 6.86765254e-01 -6.31426275e-01
1.70023990e+00 -1.74035490e+00 -1.11922897e-01 4.32631135e-01
4.02749240e-01 -1.39962882e-01 2.31306702e-01 7.95766652e-01
-9.29228067e-02 6.42337978e-01 -3.49225812e-02 6.53220594e-01
4.36447561e-01 4.84843582e-01 -9.72928256e-02 -2.53716230e-01
-3.08900714e-01 1.92383111e-01 -8.72185767e-01 -8.91824484e-01
-2.61013210e-01 -6.23350069e-02 -4.89284486e-01 -9.59463418e-02
-3.63437086e-01 -1.05668865e-01 -8.90756726e-01 8.36289942e-01
2.09243715e-01 -3.01339030e-01 8.14992726e-01 -1.60648435e-01
-5.02406180e-01 2.36093849e-01 -1.43135691e+00 1.79179394e+00
-6.59485817e-01 2.72914797e-01 3.11285913e-01 -9.18396354e-01
1.06332338e+00 8.65079224e-01 7.16402531e-01 -2.21187934e-01
6.23379983e-02 7.25131452e-01 -2.94681162e-01 -8.26799750e-01
5.48855960e-01 2.86866099e-01 -1.86118618e-01 2.40564227e-01
5.89419484e-01 9.77193378e-03 1.08889925e+00 1.94705173e-01
1.12184811e+00 3.75696093e-01 9.04574215e-01 -4.99667227e-01
8.04093182e-01 4.72661465e-01 5.55049241e-01 5.58111548e-01
8.92904960e-03 -1.84187084e-01 5.81655681e-01 -5.68454027e-01
-1.05389845e+00 -7.82716155e-01 -7.47085333e-01 1.10216558e+00
1.32680964e-02 -9.53257024e-01 -7.50372648e-01 -7.77047575e-01
-3.21047120e-02 5.64181745e-01 -1.38447374e-01 4.07343268e-01
-2.09504664e-01 -2.33596057e-01 3.02436739e-01 -3.70309800e-02
3.76262963e-01 -9.73951101e-01 -2.78619051e-01 5.32458484e-01
-3.16457123e-01 -1.71360886e+00 3.67882669e-01 1.71973899e-01
-5.15259147e-01 -1.49260354e+00 8.82505476e-02 -9.16585386e-01
4.47990924e-01 -3.36306244e-01 1.46930230e+00 1.02196865e-01
-1.88868165e-01 7.92489648e-01 -5.70541680e-01 -8.91816258e-01
-3.98388565e-01 3.74172509e-01 1.37481064e-01 -3.20937246e-01
9.63814497e-01 -6.29442155e-01 1.10559091e-01 3.88192803e-01
-1.26601315e+00 -6.16788685e-01 5.32301702e-02 3.88779998e-01
3.40253562e-01 5.78429401e-01 7.40916371e-01 -1.36564171e+00
5.51805973e-01 -8.86038959e-01 -5.75628877e-01 5.18917561e-01
-6.77980125e-01 5.03944233e-02 2.81595558e-01 5.75896800e-02
-9.98111486e-01 -1.71468541e-01 1.12940684e-01 8.97357911e-02
-4.93550330e-01 8.39900315e-01 -6.28831148e-01 -9.04751867e-02
7.39651024e-01 -3.83737385e-01 -3.30815256e-01 -7.35912204e-01
4.87127006e-02 1.18953156e+00 3.56740505e-01 -1.28228390e+00
7.16743767e-01 3.96871448e-01 -1.51040182e-01 -8.47427726e-01
-9.93421316e-01 -7.12587237e-01 -9.15735483e-01 -4.17947024e-02
9.34229851e-01 -9.51414049e-01 -6.00606978e-01 -3.69407684e-01
-1.09630752e+00 2.06151888e-01 -4.91605103e-01 2.24256843e-01
-5.36873817e-01 2.06219822e-01 -2.73585171e-01 -4.64484423e-01
2.54231263e-02 -7.46542811e-01 5.42620301e-01 2.92299807e-01
-2.07870245e-01 -1.29310465e+00 -2.53163767e-03 6.34918690e-01
1.26091912e-01 2.50746340e-01 1.17035997e+00 -1.46652377e+00
-3.24970394e-01 -4.61104810e-01 -6.81411922e-02 1.27724335e-01
2.50026882e-01 1.83077920e-02 -7.38205254e-01 7.34961331e-02
-5.11526167e-01 -3.88897434e-02 -6.27603903e-02 -5.13208151e-01
1.15021944e+00 -4.19529229e-01 -6.62359536e-01 2.73012687e-02
1.80833316e+00 2.49569774e-01 7.39997149e-01 1.34286773e+00
3.79559636e-01 1.03230023e+00 8.16620946e-01 6.10539317e-01
4.85829502e-01 8.10663462e-01 6.50546625e-02 5.04986644e-01
2.51033872e-01 1.82319153e-02 -3.41412812e-01 8.60489607e-01
-6.16139770e-01 -1.66710448e-02 -1.37812030e+00 6.84233427e-01
-1.89366651e+00 -1.07698607e+00 -1.33027524e-01 2.16769695e+00
9.96022284e-01 1.94996089e-01 -9.52437446e-02 -2.49760086e-03
6.46279037e-01 -4.02716279e-01 6.19118512e-02 -2.18591794e-01
-9.04010311e-02 4.43291277e-01 6.30206108e-01 2.65215456e-01
-1.02118266e+00 8.38447750e-01 5.80805683e+00 5.87725401e-01
-1.24986358e-01 2.27130547e-01 -2.68911809e-01 3.90660018e-01
-2.86792964e-01 5.10408759e-01 -1.33474135e+00 2.38905042e-01
1.26944125e+00 -8.51703465e-01 2.93804884e-01 9.82822657e-01
-2.44347543e-01 2.32174590e-01 -9.46067214e-01 5.24580657e-01
-2.64708877e-01 -1.58208656e+00 9.71718580e-02 5.30355990e-01
6.67043865e-01 -2.19600484e-01 -8.08203816e-01 5.72097421e-01
7.96462715e-01 -5.10380626e-01 6.28316462e-01 5.95342755e-01
5.17538846e-01 -6.11983240e-01 8.13827097e-01 3.00257742e-01
-1.24155092e+00 -3.36181104e-01 -4.22756106e-01 2.06660420e-01
-2.78993100e-01 4.11852926e-01 -5.15903533e-01 1.17567933e+00
9.35045481e-01 5.72403967e-01 -3.72144967e-01 7.31195569e-01
1.21793579e-02 -2.37094443e-02 -1.00940280e-01 2.81035691e-01
6.25611767e-02 -3.88415158e-01 5.13177276e-01 1.09961867e+00
3.24449033e-01 2.14444086e-01 4.95048493e-01 4.77394730e-01
-2.79848307e-01 5.14056027e-01 -9.60613966e-01 -2.52025872e-01
9.29685175e-01 1.38234520e+00 -4.21846658e-01 -6.34384930e-01
-8.17332625e-01 2.93498307e-01 2.66145557e-01 4.39164847e-01
-5.31383514e-01 -1.24332786e+00 8.26066375e-01 4.03319091e-01
5.44840358e-02 -1.43400118e-01 2.93269724e-01 -1.21197593e+00
6.49658144e-02 -9.27421689e-01 9.39175904e-01 -6.95444524e-01
-1.18167293e+00 5.70862174e-01 2.70647466e-01 -1.33013046e+00
-2.21860021e-01 -1.05928636e+00 1.45321757e-01 5.89056551e-01
-1.62866545e+00 -1.02885771e+00 -2.96431094e-01 5.37109971e-01
1.61574319e-01 -4.06218827e-01 1.19417739e+00 7.74528861e-01
-4.46485311e-01 -1.36867389e-01 1.47828951e-01 4.48131800e-01
9.14918840e-01 -1.31920803e+00 -4.59459037e-01 2.79011726e-01
-2.57184170e-02 1.13183594e+00 4.47640270e-01 -5.68347752e-01
-1.09577096e+00 -8.85075271e-01 1.34671617e+00 -7.15445995e-01
1.26094449e+00 1.02910334e-02 -8.34229648e-01 1.18887603e+00
4.64135736e-01 9.81869921e-02 1.15072882e+00 2.58292437e-01
-3.39017540e-01 -1.98437959e-01 -1.32445896e+00 3.33658338e-01
1.36152828e+00 -5.11254549e-01 -1.06014538e+00 7.24496663e-01
6.60955548e-01 -2.32937112e-01 -1.89492404e+00 1.07531697e-01
4.40054476e-01 -6.17282450e-01 1.04847753e+00 -1.19375682e+00
2.53291965e-01 -3.20182502e-01 -4.42607611e-01 -8.91629696e-01
-2.17850566e-01 -2.60125965e-01 -1.35221392e-01 1.65625548e+00
4.36714023e-01 -9.09008861e-01 2.57194489e-01 9.87650454e-01
-1.79279923e-01 1.07774958e-01 -7.84378052e-01 -1.11509550e+00
-2.24235952e-02 -1.00134134e-01 1.01025343e+00 1.39103305e+00
6.74202800e-01 3.98039877e-01 5.10959089e-01 3.20126891e-01
6.83751225e-01 1.77618206e-01 7.24958837e-01 -2.16134858e+00
1.87936276e-01 -1.34555683e-01 -5.85426986e-01 -2.17976525e-01
4.70669806e-01 -1.06790745e+00 -6.76955044e-01 -1.98156071e+00
-1.68738589e-01 -8.45817029e-01 -2.39533052e-01 7.50384808e-01
5.30812800e-01 -1.38267815e-01 -1.79163978e-01 3.95930916e-01
-5.63309014e-01 -2.57807821e-01 1.00554001e+00 -2.96056420e-02
-4.87285182e-02 -6.05931997e-01 -9.15321112e-01 1.01490188e+00
5.41886747e-01 -5.92828631e-01 6.28226101e-02 -1.08197719e-01
3.33639532e-01 -1.37595937e-01 -7.40371794e-02 -1.14499009e+00
3.66746902e-01 -3.91095608e-01 -2.11233512e-01 1.78365365e-01
-4.25633602e-02 -1.50536060e+00 2.77966201e-01 4.72909622e-02
-1.05427116e-01 -5.66161752e-01 7.26578832e-02 2.29405746e-01
-5.67363620e-01 -1.02969503e+00 5.05135477e-01 -7.27559745e-01
-1.10946703e+00 8.60433504e-02 -2.50818342e-01 2.07913429e-01
1.20324147e+00 -2.61733651e-01 -3.34134728e-01 1.97997838e-01
-1.19485867e+00 1.84149832e-01 6.26782775e-01 4.44978654e-01
-1.73685387e-01 -1.28987110e+00 -3.26083541e-01 -1.91898629e-01
6.84400082e-01 -3.31855685e-01 -4.44805205e-01 2.57626474e-01
-5.01406848e-01 7.94395268e-01 -4.77610171e-01 -4.97673824e-03
-9.59118426e-01 6.10870361e-01 2.17152745e-01 -4.17561650e-01
-2.54005760e-01 -1.26593322e-01 -1.87235072e-01 -8.98380458e-01
1.56903341e-01 -4.38165367e-02 -9.14962113e-01 4.59178984e-01
5.00491619e-01 2.69308567e-01 3.46606553e-01 -4.68056858e-01
-2.99554825e-01 5.62478840e-01 1.32638082e-01 -2.84069721e-02
1.42061341e+00 -3.21175784e-01 -6.05258644e-01 4.69792694e-01
6.68083310e-01 3.19462031e-01 -5.77430576e-02 -4.11664188e-01
7.64932513e-01 -3.79128128e-01 -1.58577934e-01 -6.54567063e-01
-6.79731965e-01 4.69872728e-02 2.33333513e-01 9.92997706e-01
7.01576412e-01 1.34581879e-01 3.78451020e-01 7.60026932e-01
9.51541007e-01 -1.46054065e+00 -5.16347945e-01 6.40387356e-01
5.62427580e-01 -1.00402260e+00 3.37342150e-03 -8.98306251e-01
-1.84379131e-01 1.56901670e+00 6.15994453e-01 3.63329917e-01
8.99486423e-01 2.56188691e-01 1.16995960e-01 -7.30496168e-01
-6.64751351e-01 -5.78880131e-01 9.83551890e-02 6.02230251e-01
7.05152094e-01 -2.29247600e-01 -7.75338352e-01 1.00353146e+00
-4.44285236e-02 4.51273620e-01 7.17122972e-01 1.50530744e+00
-8.43506038e-01 -1.60102201e+00 -5.22008955e-01 4.21333939e-01
-7.02784538e-01 1.52097657e-01 -3.16008359e-01 1.00643933e+00
6.86005175e-01 9.55049455e-01 2.21284658e-01 1.48117155e-01
6.54800355e-01 4.78626341e-01 3.34654301e-01 -9.77831066e-01
-4.11619842e-01 -4.72346097e-01 8.57197106e-01 -2.24941507e-01
-6.96205258e-01 -4.37390208e-01 -1.67627752e+00 -8.57556425e-03
-3.18190902e-01 7.38140225e-01 8.34450543e-01 8.18747282e-01
3.96713346e-01 4.93237793e-01 2.48416215e-01 -2.89692637e-02
-2.49817502e-02 -6.55147374e-01 -8.56935322e-01 5.82140684e-01
-2.94166028e-01 -8.82894456e-01 -2.85909146e-01 5.60474992e-01] | [9.204201698303223, 8.094673156738281] |
597f4ab7-525c-49c0-8640-0ad7d70ec5db | from-arabic-sentiment-analysis-to-sarcasm | null | null | https://aclanthology.org/2020.osact-1.5 | https://aclanthology.org/2020.osact-1.5.pdf | From Arabic Sentiment Analysis to Sarcasm Detection: The ArSarcasm Dataset | Sarcasm is one of the main challenges for sentiment analysis systems. Its complexity comes from the expression of opinion using implicit indirect phrasing. In this paper, we present ArSarcasm, an Arabic sarcasm detection dataset, which was created through the reannotation of available Arabic sentiment analysis datasets. The dataset contains 10,547 tweets, 16{\%} of which are sarcastic. In addition to sarcasm the data was annotated for sentiment and dialects. Our analysis shows the highly subjective nature of these tasks, which is demonstrated by the shift in sentiment labels based on annotators{'} biases. Experiments show the degradation of state-of-the-art sentiment analysers when faced with sarcastic content. Finally, we train a deep learning model for sarcasm detection using BiLSTM. The model achieves an F1 score of 0.46, which shows the challenging nature of the task, and should act as a basic baseline for future research on our dataset. | ['Walid Magdy', 'Ibrahim Abu Farha'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['arabic-sentiment-analysis'] | ['natural-language-processing'] | [-1.37308747e-01 4.22546148e-01 1.24201020e-02 -6.25477433e-01
-6.85382962e-01 -6.36164248e-01 5.25637150e-01 2.20966429e-01
-5.41315138e-01 4.02019262e-01 7.61314213e-01 -5.98776676e-02
8.59121501e-01 -3.14345151e-01 -2.17185929e-01 -6.69138551e-01
5.88517368e-01 3.17958266e-01 -4.24379855e-02 -1.07597065e+00
3.38762611e-01 -1.23956949e-01 -1.05088580e+00 1.03389013e+00
3.31396759e-01 9.09054399e-01 -3.55091512e-01 7.58543074e-01
1.16564855e-01 1.61871457e+00 -1.05435455e+00 -1.20306838e+00
-3.17105591e-01 -6.26874089e-01 -9.07698274e-01 8.39410722e-03
2.91592389e-01 -2.12315664e-01 2.00048700e-01 1.04229271e+00
6.46248460e-01 -1.83900118e-01 3.56510192e-01 -9.33531463e-01
-7.80852675e-01 1.03939033e+00 -6.44987047e-01 4.08672215e-03
2.11348101e-01 -9.35701951e-02 1.33857262e+00 -1.16252518e+00
4.52013135e-01 1.15472043e+00 1.11337852e+00 7.07097650e-01
-8.06771457e-01 -3.74528468e-01 -1.42004639e-01 -3.49757820e-02
-6.54377222e-01 -4.57608759e-01 8.31225812e-01 -6.11432076e-01
9.05008137e-01 1.11324571e-01 5.20719767e-01 1.25241673e+00
-1.03443183e-01 1.12636113e+00 1.20122826e+00 -5.67876637e-01
4.35638651e-02 4.76704389e-01 4.50000852e-01 5.16778648e-01
-1.38843656e-01 -6.46009147e-01 -6.96851850e-01 2.48471904e-03
-1.25060186e-01 -4.78517145e-01 -1.94111664e-04 1.67591665e-02
-9.11804676e-01 1.23500347e+00 5.16915202e-01 1.15219764e-01
3.62690985e-02 2.61051282e-02 9.78270590e-01 3.95340502e-01
6.93850100e-01 6.27654254e-01 -6.57969475e-01 -4.25010026e-01
-7.56806195e-01 1.63665414e-01 9.67538893e-01 4.89751756e-01
1.78294271e-01 1.30944967e-01 1.32919833e-01 1.21242034e+00
2.83639789e-01 7.64217973e-01 7.92331278e-01 -7.21704364e-01
3.54619175e-01 7.09814191e-01 2.02638283e-01 -1.28855669e+00
-7.92233288e-01 -3.08325410e-01 -5.17349958e-01 7.40460074e-03
5.25922954e-01 -3.82939667e-01 -4.56725985e-01 1.76155341e+00
8.38224590e-03 -7.39293098e-01 3.10726196e-01 1.03012133e+00
1.17004263e+00 4.39879388e-01 1.17808972e-02 -7.58733302e-02
1.53594422e+00 -1.33169246e+00 -7.68640518e-01 -4.92171317e-01
1.21190012e+00 -1.14512086e+00 1.53794205e+00 6.18136406e-01
-1.07862079e+00 -3.22312713e-01 -1.05652964e+00 -2.75245398e-01
-4.65372443e-01 5.20193577e-01 2.91651458e-01 7.69256651e-01
-8.21368098e-01 1.68578088e-01 -5.81968069e-01 -3.48053157e-01
2.46981174e-01 1.38280153e-01 -4.17159759e-02 5.18895566e-01
-1.14792740e+00 1.16614318e+00 -2.01851904e-01 2.47761801e-01
-2.13928759e-01 -3.42148632e-01 -8.57085586e-01 -2.30815023e-01
-3.41114774e-02 7.33419657e-02 1.77742553e+00 -1.66685295e+00
-1.44846082e+00 1.55357599e+00 -7.03115389e-02 -3.90922040e-01
3.52306217e-01 -5.54029405e-01 -5.80008507e-01 1.61750302e-01
3.38549465e-01 4.01068687e-01 7.28098989e-01 -9.85952854e-01
-4.40493524e-01 -2.48819903e-01 7.62403309e-02 2.11980343e-01
-6.61386847e-01 6.70083642e-01 2.70020850e-02 -7.09674299e-01
-2.58642077e-01 -1.20846534e+00 5.78815266e-02 -7.25960135e-01
-4.76753622e-01 -2.14272574e-01 7.06177056e-01 -7.41411686e-01
1.37667584e+00 -2.21412635e+00 1.03476606e-02 -2.84133017e-01
1.92945242e-01 2.78448403e-01 -3.72774387e-03 5.90416789e-01
-1.23359434e-01 7.33906478e-02 -2.26740703e-01 -6.58174217e-01
5.36603369e-02 -1.41443342e-01 -6.29982591e-01 4.13329154e-01
2.77222514e-01 8.93827200e-01 -9.00456548e-01 -1.86439395e-01
-1.86780810e-01 2.48304069e-01 -3.90048295e-01 1.26890808e-01
-1.66304663e-01 1.18242599e-01 -8.19349363e-02 5.25178969e-01
4.99717295e-01 -3.76677394e-01 3.30722064e-01 -3.88562083e-01
-9.83831510e-02 6.91053987e-01 -3.00933450e-01 1.44728696e+00
-3.09281856e-01 7.34812379e-01 7.99738839e-02 -9.42054331e-01
9.93707538e-01 2.68148005e-01 1.47555888e-01 -6.64566755e-01
6.20726705e-01 5.01807809e-01 5.63907139e-02 -5.01503229e-01
8.08328450e-01 -4.26072687e-01 -7.34866023e-01 8.04109216e-01
-2.03416035e-01 -5.36979198e-01 3.54435742e-01 5.38395882e-01
7.85636246e-01 -1.52980760e-01 2.22980037e-01 -5.30785620e-01
6.17856562e-01 4.22701418e-01 2.31399044e-01 4.82496560e-01
-3.50349844e-01 7.73698270e-01 9.03571069e-01 -7.01172054e-01
-1.01596379e+00 -7.41020620e-01 -4.27690148e-03 1.42593396e+00
-2.09008217e-01 -5.90276003e-01 -8.64198208e-01 -9.12970841e-01
-4.52838153e-01 5.52134454e-01 -9.06948090e-01 -2.23386809e-01
-5.77512324e-01 -1.20141745e+00 6.45286679e-01 5.83855391e-01
3.15597624e-01 -1.46371996e+00 -7.28778183e-01 7.00855181e-02
-5.49818039e-01 -1.21594977e+00 -2.31823727e-01 2.60818928e-01
-3.55117559e-01 -1.03862238e+00 -4.15808827e-01 -9.78472710e-01
4.03323531e-01 1.09614618e-01 1.71316910e+00 2.11012229e-01
2.08836347e-01 -1.10054210e-01 -9.02303696e-01 -7.66720593e-01
-8.34021628e-01 1.90027937e-01 -9.79406089e-02 -1.83087140e-01
1.08815360e+00 5.50914416e-03 -5.33041775e-01 1.72283038e-01
-6.36215031e-01 2.94611473e-02 1.08270705e-01 1.04139626e+00
8.50923657e-02 -6.17580831e-01 1.00741434e+00 -1.29475379e+00
8.55154037e-01 -4.54607099e-01 1.09828329e-02 -2.73004919e-01
-3.12558055e-01 -4.93468732e-01 7.98913717e-01 -8.00959319e-02
-9.37570572e-01 -1.24185607e-01 -4.08808827e-01 5.09397745e-01
1.27051383e-01 7.26450384e-01 3.93636584e-01 6.14023209e-01
9.49017227e-01 -2.85733581e-01 2.37051889e-01 -3.95654202e-01
3.96192968e-01 1.28003705e+00 6.81601584e-01 -1.83651760e-01
-2.99142785e-02 6.45574570e-01 -6.02113783e-01 -7.78085351e-01
-2.01468396e+00 -4.16615397e-01 -4.30979908e-01 -1.99916109e-01
9.04979408e-01 -1.29809940e+00 -7.12494791e-01 9.32020485e-01
-1.05072200e+00 -4.20460224e-01 -2.66857415e-01 2.09387783e-02
-3.99697810e-01 4.41880763e-01 -1.27055442e+00 -7.36515820e-01
-7.13481367e-01 -1.01450467e+00 9.38857675e-01 -1.35006234e-02
-1.11595702e+00 -9.74209249e-01 3.61749858e-01 9.26747859e-01
2.00081870e-01 1.83778137e-01 6.00178301e-01 -8.29914033e-01
7.66877770e-01 -3.65970641e-01 -1.94182888e-01 7.55256832e-01
-8.05484802e-02 -1.10856825e-02 -1.25407863e+00 -3.28532994e-01
2.63107657e-01 -1.30421174e+00 9.11112428e-01 1.46075651e-01
4.93157685e-01 -2.52556413e-01 4.85017359e-01 -7.40571320e-02
9.15082693e-01 -2.75859267e-01 5.88990510e-01 6.68877542e-01
6.61773980e-01 9.88538444e-01 7.22400606e-01 6.37325823e-01
6.86472118e-01 3.91524732e-01 5.86219668e-01 -3.50784391e-01
-6.11916147e-02 -7.89026618e-02 8.56664300e-01 1.35719216e+00
4.25723404e-01 -1.83636099e-01 -8.02875996e-01 6.85852051e-01
-1.95551932e+00 -9.53701913e-01 -8.75592589e-01 1.59260964e+00
1.10038137e+00 1.70546725e-01 4.38661814e-01 3.61300528e-01
4.02015269e-01 2.40262032e-01 -1.36477500e-01 -1.24244225e+00
-6.23474061e-01 1.78127736e-01 4.55513820e-02 5.15485287e-01
-1.27508509e+00 1.29785335e+00 6.35505533e+00 3.88412237e-01
-1.17849469e+00 4.96649295e-02 7.58630633e-01 -1.69781327e-01
2.07615253e-02 -3.75413507e-01 -6.06451750e-01 6.25565052e-01
8.40867996e-01 4.82726604e-01 -1.88526154e-01 9.68725026e-01
3.45008224e-01 -1.97328582e-01 -5.75571060e-01 6.30404711e-01
6.33101404e-01 -9.35045838e-01 -3.19453448e-01 -5.94855964e-01
9.38010335e-01 4.14773822e-01 2.13279784e-01 4.34715390e-01
7.05810308e-01 -1.04841411e+00 1.10263658e+00 -1.31732062e-01
4.67539877e-01 -7.54923761e-01 1.22943985e+00 5.63952141e-02
-4.68107969e-01 -5.15205301e-02 -2.15746403e-01 -6.41813934e-01
3.38137448e-01 7.11937547e-01 -6.43779993e-01 -2.97311902e-01
1.09872556e+00 1.19141376e+00 -7.61098325e-01 1.21705465e-01
-6.26997769e-01 9.69001889e-01 -1.45832509e-01 -4.47924972e-01
3.30388665e-01 -9.39982533e-02 1.00011058e-01 1.68757296e+00
-1.19889766e-01 -2.82191247e-01 8.98358226e-03 4.36221480e-01
-1.32515699e-01 2.89303333e-01 -2.13124394e-01 -2.17576772e-01
-1.14121087e-01 1.67351663e+00 -7.47945368e-01 -3.63716096e-01
-4.78308290e-01 9.54967737e-01 4.79446024e-01 -9.27677378e-03
-7.63690054e-01 -2.80449361e-01 3.57611448e-01 -2.19977573e-01
2.45614454e-01 1.69172078e-01 -6.97311580e-01 -1.32424033e+00
-4.56371456e-02 -1.39838219e+00 4.19535279e-01 -1.03883266e+00
-1.58339763e+00 8.61379504e-01 -7.75022447e-01 -9.24422443e-01
-1.76521793e-01 -8.55763018e-01 -4.99566704e-01 6.04838848e-01
-1.22030973e+00 -1.51734042e+00 -2.57392347e-01 2.71625072e-01
5.93351305e-01 -1.84273928e-01 9.71313655e-01 1.04079567e-01
-5.17519236e-01 5.94270825e-01 -1.81681037e-01 4.43150848e-01
1.13053656e+00 -1.36458552e+00 3.51859391e-01 6.83868408e-01
-1.92141179e-02 2.94770598e-01 1.02811241e+00 -2.41374016e-01
-6.77156866e-01 -6.09179080e-01 1.54288149e+00 -1.09276915e+00
1.28950512e+00 -3.55504513e-01 -6.88544929e-01 5.52329421e-01
5.12083054e-01 -5.56830823e-01 1.18654573e+00 3.30461472e-01
-7.34824300e-01 2.42596075e-01 -8.59286368e-01 6.50941849e-01
4.33144152e-01 -5.50521076e-01 -7.65423894e-01 4.24256206e-01
5.08071542e-01 -2.61275262e-01 -5.49272716e-01 4.23176169e-01
5.99774241e-01 -1.26165044e+00 4.73602295e-01 -8.17856431e-01
1.49376059e+00 -3.40362303e-02 -1.62720412e-01 -1.40030646e+00
3.74245122e-02 -2.77490199e-01 7.95497000e-02 1.19679821e+00
7.50447154e-01 -1.60271935e-02 9.18242157e-01 3.16353738e-01
-3.10887694e-01 -7.17855632e-01 -2.66167641e-01 -2.76651662e-02
4.53547478e-01 -4.07182008e-01 1.05945945e-01 1.20010197e+00
8.09232473e-01 1.06383252e+00 -5.83759189e-01 -4.53177214e-01
8.61985087e-02 3.72493774e-01 8.86869371e-01 -7.54061878e-01
-3.64752471e-01 -5.79162717e-01 5.71106141e-03 -9.78926837e-01
1.54918954e-01 -6.95718765e-01 5.09266146e-02 -1.22224677e+00
2.67786473e-01 -1.72734067e-01 -1.34965390e-01 5.79142869e-01
-1.01273723e-01 1.05470729e+00 2.29710296e-01 2.06676349e-01
-8.24348092e-01 3.71160299e-01 9.50659215e-01 -8.14379081e-02
-7.30573386e-02 -1.72284573e-01 -1.17781878e+00 1.17381144e+00
1.18742085e+00 -3.28137964e-01 -4.27430123e-02 -5.35948157e-01
1.21148515e+00 -4.35652226e-01 3.75886038e-02 -4.51913029e-01
-2.82523006e-01 1.86308056e-01 -4.85005789e-02 -6.65832818e-01
5.98466218e-01 -3.40699613e-01 -7.20600605e-01 5.27552187e-01
-5.95808387e-01 4.72620487e-01 6.63539246e-02 5.83904721e-02
-4.46261376e-01 -3.16156656e-01 9.96696591e-01 -1.53207496e-01
-4.07105833e-01 -6.03191078e-01 -7.50503480e-01 6.19911671e-01
3.94118190e-01 2.06954360e-01 -7.45266557e-01 -5.67334294e-01
-5.62981606e-01 1.14551105e-01 5.22901654e-01 4.78605300e-01
2.38624379e-01 -1.05170631e+00 -1.02498674e+00 -1.56476080e-01
4.87075180e-01 -2.81416804e-01 -2.57081036e-02 1.15387225e+00
-7.17785716e-01 -1.07785709e-01 -9.03020129e-02 -3.66150588e-01
-1.47619236e+00 2.34739751e-01 2.05035612e-01 -3.50127190e-01
-2.33725697e-01 1.05854857e+00 -5.53216413e-02 -6.25127316e-01
-2.38103181e-01 5.10999113e-02 -7.18331635e-01 4.64603990e-01
6.07892632e-01 1.09617934e-01 1.84426844e-01 -1.13208342e+00
-2.49625623e-01 2.39834234e-01 -2.85006374e-01 -1.60080090e-01
1.25471926e+00 -2.57784307e-01 -5.33614695e-01 6.79409087e-01
1.12518120e+00 4.55193222e-01 -6.13172650e-01 4.40993719e-02
7.66797215e-02 -3.94396149e-02 -3.43808591e-01 -1.21202850e+00
-8.97189975e-01 8.33576620e-01 2.64299363e-02 5.37022591e-01
8.33965540e-01 2.02661231e-02 9.91369069e-01 4.09324914e-01
-2.19166517e-01 -1.62167931e+00 4.84287441e-01 1.11518550e+00
8.74487400e-01 -1.59969521e+00 -4.01655212e-02 -2.04163447e-01
-1.25181377e+00 1.11625791e+00 7.26905107e-01 -3.87502253e-01
5.45844376e-01 3.36222887e-01 1.01403308e+00 -5.39686918e-01
-5.67720354e-01 1.77147627e-01 -1.77595913e-01 2.50679642e-01
1.04016232e+00 1.31297514e-01 -6.67127132e-01 1.11923504e+00
-1.03238630e+00 -3.81586611e-01 1.09078836e+00 6.67730868e-01
-3.34785044e-01 -8.63506675e-01 -9.61731821e-02 1.89172521e-01
-9.08683896e-01 -2.92081863e-01 -9.81253862e-01 3.86750460e-01
-2.58456200e-01 1.38651764e+00 -1.21670574e-01 -4.85690027e-01
2.89736122e-01 -1.61658078e-01 -1.66477654e-02 -6.64977133e-01
-1.28317547e+00 -1.06978804e-01 5.38556576e-01 -2.31342658e-01
-7.42421985e-01 -5.69105506e-01 -1.33587217e+00 -3.21513027e-01
-2.55214661e-01 2.88638771e-01 6.17758393e-01 8.77775848e-01
7.20109642e-02 2.20172629e-01 6.07185721e-01 -4.98384744e-01
-7.33859479e-01 -1.22429085e+00 -4.82291371e-01 1.12648141e+00
2.27868125e-01 -1.58582181e-01 -6.78701758e-01 3.68091762e-01] | [9.142120361328125, 10.541312217712402] |
506a44b9-77ee-48a9-ac5c-d30e539c778c | multiscale-autoencoder-with-structural | 2208.04945 | null | https://arxiv.org/abs/2208.04945v1 | https://arxiv.org/pdf/2208.04945v1.pdf | Multiscale Autoencoder with Structural-Functional Attention Network for Alzheimer's Disease Prediction | The application of machine learning algorithms to the diagnosis and analysis of Alzheimer's disease (AD) from multimodal neuroimaging data is a current research hotspot. It remains a formidable challenge to learn brain region information and discover disease mechanisms from various magnetic resonance images (MRI). In this paper, we propose a simple but highly efficient end-to-end model, a multiscale autoencoder with structural-functional attention network (MASAN) to extract disease-related representations using T1-weighted Imaging (T1WI) and functional MRI (fMRI). Based on the attention mechanism, our model effectively learns the fused features of brain structure and function and finally is trained for the classification of Alzheimer's disease. Compared with the fully convolutional network, the proposed method has further improvement in both accuracy and precision, leading by 3% to 5%. By visualizing the extracted embedding, the empirical results show that there are higher weights on putative AD-related brain regions (such as the hippocampus, amygdala, etc.), and these regions are much more informative in anatomical studies. Conversely, the cerebellum, parietal lobe, thalamus, brain stem, and ventral diencephalon have little predictive contribution. | ['Qiankun Zuo', 'Changhong Jing', 'Yongcheng Zong'] | 2022-08-09 | null | null | null | null | ['disease-prediction'] | ['medical'] | [-7.21353069e-02 1.99741811e-01 1.85456336e-01 -4.43352073e-01
-4.60262448e-01 1.92483477e-02 4.60955799e-01 7.11944029e-02
-5.03921211e-01 6.61245942e-01 5.29233277e-01 3.80602665e-02
-3.18644732e-01 -5.68160892e-01 -2.67073601e-01 -6.69894338e-01
-6.41235113e-01 3.00732255e-01 3.74062010e-03 1.65604651e-02
6.04272075e-02 4.58647698e-01 -1.26955247e+00 3.65923077e-01
7.70476997e-01 1.28421056e+00 2.82088339e-01 2.71955639e-01
1.72985107e-01 7.55502284e-01 -1.02228589e-01 -2.98506200e-01
-1.75822049e-01 -3.06757241e-01 -8.59328628e-01 1.23183597e-02
7.20259547e-02 -5.11021137e-01 -2.52362043e-01 1.24060631e+00
7.85291314e-01 -2.79344171e-01 7.65124381e-01 -8.13787997e-01
-6.42310381e-01 3.16187561e-01 -6.35380268e-01 7.27828205e-01
-1.74288630e-01 1.04791641e-01 1.01668966e+00 -1.13476288e+00
5.61824203e-01 1.26497507e+00 4.81504947e-01 3.37133437e-01
-9.33897853e-01 -4.95951325e-01 1.01999015e-01 7.72594333e-01
-1.07565236e+00 -1.05994955e-01 6.66993320e-01 -4.92019325e-01
6.53188169e-01 -1.50492210e-02 8.19299579e-01 1.19389510e+00
6.95997477e-01 5.73290586e-01 1.28761935e+00 -4.03448381e-03
2.89144069e-01 -5.17798699e-02 4.15241033e-01 7.24066615e-01
1.75091833e-01 -6.51014736e-03 1.68917581e-01 -4.53698099e-01
7.70771801e-01 5.75508118e-01 -4.01590914e-01 -2.53729343e-01
-1.41823232e+00 9.90056872e-01 9.87403154e-01 5.80984294e-01
-9.44632411e-01 -7.48000517e-02 4.80652928e-01 2.69458108e-02
5.67023158e-01 2.39784773e-02 -5.95659971e-01 4.66607451e-01
-8.09109330e-01 5.21483049e-02 1.39199510e-01 5.26702665e-02
4.70339805e-01 -5.58423288e-02 -1.65684633e-02 1.01356351e+00
6.24022782e-01 2.66224623e-01 1.02855778e+00 -7.88567305e-01
1.19302124e-01 8.33971441e-01 -4.76272523e-01 -7.86468327e-01
-8.06651115e-01 -5.92351973e-01 -1.22278762e+00 2.76278466e-01
-6.36509014e-03 -2.63794303e-01 -6.13210380e-01 1.52745652e+00
4.04112458e-01 -1.26016811e-01 -1.86151400e-01 1.28926206e+00
6.23966932e-01 1.97814882e-01 4.72744584e-01 -9.76555422e-02
1.88114941e+00 -7.40356445e-01 -6.37797892e-01 -5.10382950e-01
6.05229318e-01 -1.06382214e-01 6.74593210e-01 -4.60664593e-02
-8.08798790e-01 -3.69841456e-01 -8.22702944e-01 -1.44209102e-01
-3.84723186e-01 3.86085540e-01 7.20967710e-01 8.69370848e-02
-1.00242460e+00 4.83722448e-01 -9.20604050e-01 -4.63466942e-01
8.85058045e-01 2.49763995e-01 -8.29185724e-01 -2.50140011e-01
-1.17933047e+00 1.06403995e+00 3.18512470e-01 1.46912530e-01
-8.26968610e-01 -7.94899702e-01 -6.62028909e-01 1.61947533e-01
-1.61260664e-02 -8.65732729e-01 7.63420880e-01 -1.00601161e+00
-1.11577129e+00 7.92619288e-01 4.70687002e-02 -4.42996949e-01
-6.60819709e-02 -2.08778322e-01 -4.96106297e-01 6.43150151e-01
1.10118218e-01 9.19270992e-01 6.13485813e-01 -7.52649605e-01
-3.03623140e-01 -1.27408862e+00 -2.15399817e-01 -9.92920995e-02
-4.22270507e-01 2.87951767e-01 5.12097836e-01 -7.09742367e-01
2.55844921e-01 -7.15464592e-01 -4.73648190e-01 3.47629249e-01
-8.47931057e-02 -2.71864146e-01 7.97917783e-01 -1.30555868e+00
9.03293192e-01 -2.04856753e+00 3.48216891e-01 4.82749240e-03
8.13284993e-01 -1.23593114e-01 8.55386779e-02 -1.52484581e-01
-4.71432894e-01 1.51649445e-01 -4.35743153e-01 3.01950008e-01
-1.31687135e-01 1.52035733e-03 3.37790996e-01 6.29402459e-01
4.37614202e-01 1.12075710e+00 -5.66664636e-01 -4.58172500e-01
2.64685694e-02 7.03594148e-01 -6.45006299e-01 -4.94914781e-03
8.92698690e-02 5.80928683e-01 -5.98516107e-01 7.14264572e-01
3.73465180e-01 -4.87292647e-01 2.16301337e-01 -4.01297748e-01
1.00173227e-01 7.02798460e-03 -4.80631739e-01 1.43034124e+00
-7.04763234e-02 3.66642326e-01 3.73721570e-01 -1.35331690e+00
7.29101777e-01 4.80072051e-01 5.75149715e-01 -9.18136477e-01
4.07590508e-01 -4.58105952e-02 6.54043674e-01 -9.09374893e-01
-6.05265439e-01 -1.75830483e-01 3.26208621e-01 5.31049550e-01
2.92655200e-01 9.27838802e-01 -1.19453095e-01 -7.15382323e-02
1.15620589e+00 -2.23755643e-01 5.16636550e-01 -4.81954277e-01
6.77137613e-01 -4.26576227e-01 5.82705975e-01 -1.58408731e-02
-6.03439033e-01 3.07776779e-01 5.32734871e-01 -6.42958701e-01
-1.07857788e+00 -8.98321390e-01 -3.43279541e-01 8.91232967e-01
-5.58802187e-01 -1.14204757e-01 -8.04121792e-01 -6.86062336e-01
-5.08048683e-02 3.69625717e-01 -8.64366114e-01 -4.05679256e-01
-5.69330692e-01 -1.19269764e+00 2.45422021e-01 9.06383991e-01
8.36669683e-01 -1.03941643e+00 -6.50625944e-01 -6.26356341e-03
-6.08499125e-02 -8.70606542e-01 -1.69310763e-01 -4.84727696e-02
-1.51609027e+00 -1.22887135e+00 -1.20803714e+00 -9.04657483e-01
7.50821590e-01 -6.63129389e-02 8.60309303e-01 -1.88601062e-01
-6.67350471e-01 3.82300973e-01 -1.47013202e-01 -1.27526388e-01
1.19740777e-02 -1.83433563e-01 -1.58567354e-01 3.25369537e-01
3.37458163e-01 -8.91814172e-01 -1.15087175e+00 1.66467145e-01
-7.05656052e-01 -2.94213086e-01 1.04075253e+00 7.43586957e-01
6.13600433e-01 -1.04065411e-01 7.80602217e-01 -4.73253876e-01
6.55591607e-01 -9.15169358e-01 -2.37135645e-02 1.09097123e-01
-6.04452848e-01 -3.21819098e-03 3.35608572e-01 -2.26609185e-01
-8.86614919e-01 -5.42195626e-02 -1.62826866e-01 -2.79904574e-01
-4.87820804e-01 5.87891996e-01 -3.20832640e-01 3.61728817e-02
5.22510707e-01 2.25210100e-01 3.64011526e-01 -7.72579372e-01
5.50575964e-02 5.41431665e-01 2.80555099e-01 -1.48575559e-01
-2.93879360e-02 4.48688418e-01 -9.97686014e-02 -6.50702715e-01
-6.11553729e-01 -1.59161597e-01 -7.96150267e-01 -1.39759287e-01
1.25414205e+00 -8.98919880e-01 -5.41321278e-01 3.07677947e-02
-9.68573809e-01 4.37115841e-02 1.84138462e-01 8.48505676e-01
-1.89988747e-01 3.81159812e-01 -8.12551618e-01 -3.94977778e-01
-7.96626925e-01 -1.10375106e+00 1.01328993e+00 -1.08427286e-01
-1.43093273e-01 -1.11672020e+00 6.86200336e-02 4.47731763e-01
5.94474971e-01 2.76024491e-01 1.66140985e+00 -6.63593650e-01
-2.74856389e-01 -7.13369697e-02 -5.31191766e-01 4.19208348e-01
-4.46481667e-02 -5.84726989e-01 -7.27186859e-01 -8.35999250e-02
4.43704456e-01 -1.04011312e-01 8.94079030e-01 6.92397952e-01
1.12268949e+00 -4.64527339e-01 -2.41259679e-01 1.60115659e-01
1.16210663e+00 2.19078124e-01 7.44910240e-01 1.83320478e-01
4.71825987e-01 7.27286994e-01 -9.37225893e-02 2.85904229e-01
5.74118197e-01 3.56062293e-01 5.90996861e-01 -8.28533769e-02
-1.05949491e-01 5.34766316e-01 5.23860574e-01 8.91291320e-01
-3.66952121e-01 4.66300666e-01 -9.03862834e-01 5.24819613e-01
-1.56719077e+00 -8.51862609e-01 -7.51258656e-02 1.94317997e+00
5.76929271e-01 -2.47088641e-01 1.32713705e-01 -2.97995582e-02
7.97122538e-01 8.16763937e-02 -6.22376680e-01 4.20729704e-02
-3.99691723e-02 2.81423897e-01 -5.18281246e-03 4.84534986e-02
-1.02142298e+00 3.63139868e-01 6.14807606e+00 8.43117535e-02
-1.06327760e+00 5.67726076e-01 8.34862590e-01 -6.73089102e-02
4.69547063e-02 -3.35583270e-01 -2.67768323e-01 5.66898823e-01
1.16887522e+00 1.84914500e-01 3.88133496e-01 8.47412765e-01
2.16175392e-01 3.79243940e-02 -9.55186009e-01 7.66222596e-01
4.11210619e-02 -8.76359582e-01 5.41881621e-02 3.34152102e-01
1.62714332e-01 1.85496494e-01 -6.17851764e-02 2.54425585e-01
-2.51337767e-01 -1.04566503e+00 1.59435809e-01 9.24642622e-01
3.94504398e-01 -6.61713779e-01 9.74060476e-01 1.34050101e-01
-8.01773906e-01 -4.05327886e-01 -4.38329995e-01 2.52198987e-02
-8.08424652e-02 8.35314035e-01 -5.63805640e-01 1.35747448e-01
9.11021531e-01 7.69440830e-01 -7.50150323e-01 9.46420133e-01
-1.85179844e-01 6.32155597e-01 1.39081851e-01 1.44222006e-01
2.42159054e-01 -1.96474731e-01 2.81336129e-01 9.73894298e-01
3.91749948e-01 3.71763498e-01 1.52878597e-01 1.22112000e+00
-2.22681779e-02 2.50166535e-01 -4.42961693e-01 1.55568612e-03
-1.47854432e-01 1.57543218e+00 -5.42693973e-01 -3.34191442e-01
-4.67477769e-01 8.82579446e-01 2.86379606e-01 3.69622558e-01
-7.39193320e-01 -8.86919573e-02 4.90770966e-01 1.74536750e-01
3.26690227e-01 -1.01490147e-01 -1.01965360e-01 -9.95840013e-01
-2.02046596e-02 -7.69510508e-01 4.08650130e-01 -8.32653284e-01
-1.50361812e+00 7.34004319e-01 -2.22854048e-01 -7.16878593e-01
-1.79900397e-02 -8.42506647e-01 -6.21428668e-01 7.83029497e-01
-1.41019058e+00 -9.13817406e-01 -2.52014875e-01 7.88774073e-01
2.45091125e-01 -2.99704880e-01 1.07905817e+00 6.53645515e-01
-8.20778370e-01 -2.57793684e-02 -8.91577974e-02 3.92589808e-01
3.87977898e-01 -1.18394506e+00 -3.36603403e-01 2.54229724e-01
-2.15227932e-01 6.17359936e-01 1.85242519e-02 -6.19773626e-01
-9.31487262e-01 -1.31076574e+00 7.25181282e-01 8.67613067e-04
7.25697696e-01 -8.69455189e-02 -9.94719744e-01 7.00345874e-01
2.63113141e-01 2.68649459e-01 8.80025804e-01 3.49877402e-02
-3.02616775e-01 -1.36714116e-01 -1.22109532e+00 4.99862880e-01
7.46834874e-01 -5.34215271e-01 -8.22919726e-01 5.17385304e-01
3.99905771e-01 4.02372241e-01 -1.37708735e+00 4.24990445e-01
5.02025068e-01 -7.52944171e-01 1.11439955e+00 -8.92995059e-01
7.39753008e-01 5.11074625e-02 -1.52826369e-01 -1.30295813e+00
-8.86298299e-01 3.75509024e-01 -2.00102687e-01 8.29995394e-01
2.14447439e-01 -7.17832983e-01 3.26965362e-01 4.12920266e-01
-5.57632744e-02 -9.38929915e-01 -1.09599328e+00 -3.19404751e-01
2.17846587e-01 -7.07292482e-02 1.81696787e-01 8.91021252e-01
-7.60875270e-02 6.33814454e-01 -5.61155640e-02 1.92654908e-01
6.86496317e-01 1.50278201e-02 -1.68462396e-01 -1.65460336e+00
9.28731114e-02 -4.54120040e-01 -7.12969065e-01 -1.34073183e-01
2.13745177e-01 -1.32006180e+00 -6.17704868e-01 -1.49159122e+00
6.17352843e-01 2.76834428e-01 -9.12671566e-01 6.04173839e-01
-5.31874336e-02 8.11535046e-02 9.23596881e-03 8.66552666e-02
-4.17662561e-01 7.57715881e-01 1.34926784e+00 -3.05738658e-01
1.56396821e-01 -3.03007066e-01 -8.38348210e-01 9.62560773e-01
9.42868948e-01 -4.37391669e-01 -1.18006565e-01 -2.50176162e-01
-3.20387691e-01 -1.57159138e-02 9.89047587e-01 -9.83905852e-01
-2.13291552e-02 3.91239017e-01 1.13398433e+00 -1.35300472e-01
1.36302412e-01 -9.26645696e-01 -1.78940386e-01 6.71282172e-01
-2.66379356e-01 2.20185310e-01 -6.39524981e-02 4.32538182e-01
-2.05007061e-01 -1.40520195e-02 8.69931042e-01 -5.61400652e-01
-4.85203832e-01 4.95078474e-01 -6.30523443e-01 -2.41070360e-01
9.28564250e-01 9.95250046e-02 -1.33249924e-01 1.20022977e-02
-1.13813353e+00 3.20498049e-02 -3.45124692e-01 2.24306256e-01
8.04667711e-01 -1.54900038e+00 -8.33299875e-01 9.72134918e-02
-1.64223641e-01 -5.81253469e-01 5.59406757e-01 1.45570517e+00
-1.09963320e-01 5.39660096e-01 -8.00585210e-01 -5.60045063e-01
-1.08641088e+00 4.98889923e-01 4.53812331e-01 -3.51340503e-01
-8.84634614e-01 3.98990691e-01 4.14647430e-01 -3.00981313e-01
1.39286295e-01 -5.09410679e-01 -6.56837761e-01 2.40640536e-01
7.77612627e-01 5.26541471e-01 6.12742528e-02 -6.34728849e-01
-5.03464639e-01 3.13222826e-01 -4.20929700e-01 2.14268610e-01
1.80024290e+00 9.58254114e-02 -5.56803167e-01 1.13867320e-01
1.31147981e+00 -5.02579987e-01 -9.24192667e-01 -1.35470495e-01
2.43406817e-01 1.69733763e-01 6.37245536e-01 -1.08509195e+00
-1.62481189e+00 1.27331483e+00 1.34722090e+00 -1.00923479e-01
1.03713238e+00 2.67478764e-01 9.08661783e-01 2.64812440e-01
2.66041130e-01 -8.28133345e-01 -3.13769467e-02 2.21914262e-01
1.28770304e+00 -1.14781678e+00 1.29401414e-02 2.03579795e-02
-6.20904148e-01 1.25021100e+00 5.00919521e-01 -2.51920313e-01
9.54224646e-01 -8.92417580e-02 -2.60250688e-01 -6.25087976e-01
-6.18375361e-01 -8.00300762e-02 4.40222681e-01 3.27188343e-01
4.49880749e-01 6.26773760e-02 -4.14636791e-01 1.26588523e+00
2.09610745e-01 -3.48743796e-02 -1.52582809e-01 6.63471758e-01
-7.14093089e-01 -7.01598227e-01 -4.26156968e-01 8.31241250e-01
-7.00884342e-01 -1.73608199e-01 -4.72544938e-01 6.48201823e-01
3.03112268e-01 3.20000589e-01 1.05948329e-01 -3.86643670e-02
3.39649208e-02 5.50935924e-01 2.71158248e-01 -3.75582665e-01
-4.47196424e-01 1.73575982e-01 -1.99835002e-01 -7.16149807e-01
-5.87483048e-01 -6.54048085e-01 -1.37777936e+00 1.70803174e-01
3.15363593e-02 -2.76846647e-01 5.38447022e-01 9.39778984e-01
8.35518479e-01 8.14597487e-01 4.00822610e-01 -7.66422093e-01
-2.40578607e-01 -1.31633413e+00 -5.88127971e-01 1.72520593e-01
1.21102482e-01 -8.50088835e-01 -1.38156623e-01 5.67211583e-02] | [14.182737350463867, -1.7141987085342407] |
2963886d-0e0b-44ed-84d3-0930062e2c97 | two-heads-are-better-than-one-enhancing | 2201.10113 | null | https://arxiv.org/abs/2201.10113v7 | https://arxiv.org/pdf/2201.10113v7.pdf | Multimodal data matters: language model pre-training over structured and unstructured electronic health records | As two important textual modalities in electronic health records (EHR), both structured data (clinical codes) and unstructured data (clinical narratives) have recently been increasingly applied to the healthcare domain. Most existing EHR-oriented studies, however, either focus on a particular modality or integrate data from different modalities in a straightforward manner, which usually treats structured and unstructured data as two independent sources of information about patient admission and ignore the intrinsic interactions between them. In fact, the two modalities are documented during the same encounter where structured data inform the documentation of unstructured data and vice versa. In this paper, we proposed a Medical Multimodal Pre-trained Language Model, named MedM-PLM, to learn enhanced EHR representations over structured and unstructured data and explore the interaction of two modalities. In MedM-PLM, two Transformer-based neural network components are firstly adopted to learn representative characteristics from each modality. A cross-modal module is then introduced to model their interactions. We pre-trained MedM-PLM on the MIMIC-III dataset and verified the effectiveness of the model on three downstream clinical tasks, i.e., medication recommendation, 30-day readmission prediction and ICD coding. Extensive experiments demonstrate the power of MedM-PLM compared with state-of-the-art methods. Further analyses and visualizations show the robustness of our model, which could potentially provide more comprehensive interpretations for clinical decision-making. | ['Buzhou Tang', 'Yang Xiang', 'Hui Xu', 'Hui Wang', 'Ge Li', 'Yongshuai Hou', 'Xiaolong Wang', 'Sicen Liu'] | 2022-01-25 | null | null | null | null | ['readmission-prediction'] | ['medical'] | [ 0.3899944 0.20975412 -0.39946374 -0.53926325 -0.9633922 -0.24009174
0.46034044 0.70193124 -0.11143761 0.5588394 0.89730275 -0.5262905
-0.33218718 -0.70285743 -0.47631425 -0.4222189 -0.05638197 0.599654
-0.516466 0.01564781 -0.21164551 -0.08389643 -0.9469486 1.0792164
0.9174899 1.0601672 0.05524606 0.22241588 -0.2988661 1.3576745
-0.11515354 -0.42796844 -0.14262812 -0.46416092 -0.6899569 0.04064392
-0.28539932 -0.28647718 -0.5776359 0.7458635 0.61698586 -0.46554083
0.82039696 -0.7210632 -0.95790803 0.924079 -0.39084852 -0.17093265
0.6940615 0.06691646 0.6287771 -0.7529381 0.6721399 1.0052829
0.87120533 0.47809288 -1.0654811 -0.53676134 0.08934562 0.08958965
-1.1742378 -0.26057693 0.56213915 -0.77085924 0.8518772 0.2038638
0.6076556 1.330098 0.48187214 0.9344992 0.960991 -0.15657005
-0.07238117 0.2324453 0.31197566 0.5225307 0.27092814 0.1095577
-0.2639459 -0.45146403 0.43426272 0.74781334 -0.42063856 -0.13120407
-1.5838165 0.65147716 0.29889452 0.32996342 -0.73103726 -0.44277844
0.7768684 0.06422375 0.19146003 0.13175918 -0.52199835 -0.10106786
-0.7630095 0.03009326 0.8410205 0.965266 0.06145249 -0.29466793
-0.52329487 0.82501525 0.5260569 0.38124663 0.7476695 -0.39890012
0.8509696 1.0474205 -0.06736735 -1.0488715 -0.6883975 -0.25422502
-1.4686333 -0.73449963 0.04487404 -0.28031573 -0.9100744 1.5160121
0.02069077 0.10361861 0.45955434 0.6471427 1.463321 0.56287014
0.45952535 -0.28967774 1.6349208 -0.665224 -1.1389445 0.11716373
1.0395151 -0.4938757 0.58817816 0.24394748 -1.0789002 -0.44473276
-0.71417975 -0.09384031 -0.23475026 0.32087427 0.29094595 0.06851598
-0.5356143 0.3614998 -0.9104364 -0.22356431 0.56469053 0.12365209
-0.4340379 -0.350307 -1.4668046 0.6884853 0.361635 0.21621081
-0.6011944 -0.8414357 -1.0348368 0.23729847 0.24091776 -0.99488205
1.2097094 -0.4435496 -1.0500567 0.84712654 -0.2457474 -0.46692592
0.45055157 -0.17211692 -0.7722068 0.02802327 -0.10275647 0.2899434
0.21548365 -1.2235898 -0.46596304 -0.4462165 -0.21386163 0.03978119
-0.35589093 -0.1270108 -0.46840033 -0.8394173 -0.11833646 -0.5857056
-0.3278711 -0.32893026 -0.592326 -0.05749303 0.37723002 -0.9886183
1.7778734 -2.1855087 0.23993434 0.06134264 0.55998063 0.32557967
-0.09273627 0.8303498 -0.20501377 0.0440632 -0.3921896 -0.3881979
-0.13836898 0.3147526 -0.43341032 0.1902405 0.21209438 1.3084198
-0.79775804 -0.7195246 0.05751996 0.76287293 -0.52403075 0.4590633
0.10350206 0.7396235 -0.51678795 0.75799966 0.44706616 -0.9039389
0.6130428 -0.55943495 0.23293446 0.48195115 -0.6892128 1.6584787
-0.5519023 0.01224325 -0.3290371 -1.0228261 0.61870515 0.8417645
0.91171885 -0.80859417 0.2544123 0.00754441 0.02908205 -1.116699
0.09118346 -0.22321853 -0.12435418 0.33234093 -0.17374991 0.5544911
-0.2758613 0.13911599 1.1610599 0.01203698 0.8588864 0.17267847
0.509748 -0.1181885 0.62275636 0.60766226 0.0915013 0.667846
0.3042625 -0.4671762 -0.6753587 -0.8303431 -0.4920278 0.55005634
0.09689348 -0.72439593 -0.32386872 -0.731168 0.03497757 0.5333024
-0.73900455 -0.2430641 -0.4304323 -0.881159 0.65091026 0.84660524
0.23211327 -1.0621719 -0.6595122 0.40901184 -0.47630087 -1.216454
-0.40361992 0.16338666 -1.0446227 -1.1593051 -0.61436963 -0.6150851
0.59574157 -0.17013328 1.1380131 0.06029811 -0.14470115 0.4127616
-0.50465024 -0.27890915 -0.51598746 -0.05340878 -0.23907594 0.21755512
0.61530846 -0.37980035 -0.6525637 0.0447218 -1.1718736 0.41216028
0.75369555 1.1036097 0.62025845 -0.4470664 0.6280319 -1.3825423
0.71684957 -1.0235432 0.0443752 0.48969945 -0.6030137 0.00594989
0.76611865 -0.43440726 -1.0635118 0.108601 -0.33087057 -0.41270667
-0.3731967 1.3677473 -0.3368369 0.70570475 0.27877814 0.38802093
-0.20734835 -0.7001846 0.13347211 1.145159 0.49813694 -0.34282163
0.34116307 0.30223542 -0.28979734 -0.27632925 -0.8696859 -0.47593865
-0.4502944 0.21548705 0.8811997 -1.0935043 -0.8713333 0.30292222
-1.0677958 0.05459959 -0.29141223 0.6179999 -0.2733563 0.4887836
-0.91975385 -0.6747658 -0.42443165 -1.1179632 1.2755234 -0.12370382
-0.4753522 -1.2659163 0.07986521 0.5239471 0.17017107 0.26730517
1.3648008 -1.0633612 -0.17009267 -0.3393612 -0.3576714 -0.00761097
0.47548583 -0.45256352 -0.8819983 -0.14927213 0.12552115 -0.3388754
0.7326072 0.35190678 1.4443045 -0.569469 -0.6473472 0.76497334
1.1409079 0.54638207 0.6596757 0.0612953 0.8765482 0.5238369
0.36959663 0.6344107 0.7957289 0.55201507 0.19964233 -0.3883834
0.34214965 -0.5186201 0.14502403 1.3351127 0.03594308 -0.19837122
-1.1531702 0.36121547 -2.1613948 -0.7011895 0.02993796 1.9131669
1.1056914 -0.18178909 -0.06830194 0.01448328 0.3860947 -0.18103583
-0.63439614 0.0253061 0.01607169 -0.0404718 0.08770754 0.0588779
-1.1670517 0.11309203 5.728084 0.46071258 -1.0355595 0.17753989
0.68969053 -0.02905602 -0.48128754 -0.43583223 -0.55087495 0.70477164
1.0518591 0.03739355 0.01097493 0.63522935 0.21929425 0.43469298
-1.5451388 1.1849655 -0.08562355 -1.5490413 0.278706 0.23600876
0.49198404 -0.05281075 -0.04324317 0.46881434 0.17844275 -1.302094
0.39196184 0.92030627 0.9887323 -0.19860855 1.1817104 0.46241114
-1.2565029 -0.2163665 0.0915065 0.21303664 0.47873172 0.5627573
-0.61457825 1.1772985 0.5710827 1.4619286 -0.37902966 0.87191695
0.2238083 0.69369924 0.1366276 0.491301 0.02690974 -0.08441316
0.11246932 1.5161808 0.47035003 0.43123838 0.4218941 0.72950906
-0.14241792 0.26350173 -0.76620907 -0.37273946 0.2738083 1.0346246
-0.18606703 -0.59714717 -0.8839396 0.5017882 0.13020563 0.38820785
-0.7941115 0.05985432 0.33173224 0.22569463 0.16768824 0.27833048
-0.46456742 -1.66378 -0.10628968 -1.1942891 0.74389905 -0.61847246
-1.7735437 0.8538227 -0.03155975 -1.7137731 -0.43749723 -0.43053454
-0.12897673 0.8170034 -1.4187722 -1.2016885 -0.19185983 0.6937696
0.27202782 -0.33055267 1.1813635 0.73455936 -0.56245375 0.71144956
0.19211459 0.55721235 0.6648698 -0.8103234 -0.0911482 0.24545005
-0.09822594 1.0756572 0.10226202 -0.776107 -1.4185791 -1.3033998
1.031507 -0.49774662 0.32931155 -0.14504032 -1.2954892 0.8415135
0.24756287 -0.14539267 1.291963 0.03279428 -0.36118847 0.11933008
-0.9174731 0.44026676 0.88562584 -0.5994182 -0.9320939 0.08976228
0.7314711 -0.3951752 -1.3947366 0.81957835 0.6741893 -0.580302
1.0133609 -1.1611034 0.9853064 -0.14379133 -0.31995746 -1.1280359
-0.18823463 -0.15538472 -0.47230095 0.9589906 0.39102706 -0.6241082
0.1387918 0.4798674 -0.04986933 -1.3065003 -0.5846297 -0.11053173
-0.13030033 -0.44853792 0.7580827 1.3374949 0.4810032 0.395159
-0.5846987 0.06879284 0.12064936 0.32082197 0.28216013 -1.259346
-0.43452784 -0.1839852 -0.18583374 -0.9987006 -0.06254213 -1.1650089
-0.14799203 -1.8577375 0.70677865 -0.29509038 -0.6899727 0.8167519
-0.29390708 -0.05416071 0.06933242 0.42172575 -0.43286663 0.5268544
1.2208383 -0.46662566 -0.30699843 -0.18621726 -0.873979 0.72622085
0.4242337 -0.48889235 -0.39497066 -0.35285425 0.31943825 0.70251083
0.22474638 -0.5711267 0.10886311 -0.09079753 0.39960256 -0.58528346
0.10006566 -1.1041297 0.35280776 0.44858646 -0.7907414 0.03315254
0.23954245 0.600071 -0.59200066 0.13845083 0.27803186 -0.06327738
-0.15351625 0.33299148 -0.40341696 0.02247043 0.7543426 0.03198308
-0.2701635 -0.36260435 -1.0012532 0.32871422 -0.01667056 0.49866837
0.8710945 -1.4636693 -0.8510829 0.25170124 0.46232268 -0.01428429
0.5824874 1.3410292 -0.08177906 0.6039898 0.0297078 -0.73778933
-1.1778578 0.9406866 0.22290862 -0.81601125 -0.9623427 0.10703383
0.5540589 -0.6192596 0.50354385 -0.6716985 -0.60010964 0.12198976
0.62581044 -0.16048104 -0.01803364 -0.58428824 -0.39599994 0.48129782
-0.23996371 0.27281174 1.5457102 -0.1482545 -0.10398407 0.6933345
1.152402 -0.2914085 -0.5830097 -0.54438853 0.02684738 -0.04530168
-0.17130958 -1.1174402 -1.1183797 1.0719651 0.49508613 -0.01520473
1.3783665 0.02765267 0.9154323 0.28739536 0.04610842 -0.4861272
-0.29292464 0.31339335 0.7569851 -1.3040135 -0.20125288 -0.31329173
-1.0747277 1.0485823 0.28386128 0.5159793 0.9103938 0.4236803
0.27546582 -0.22439134 -0.91541976 0.26215324 0.6146392 0.35855865
0.89787126 0.09461699 -0.29580313 1.4609698 0.3054669 0.7534509
0.14505495 1.0749896 0.38534632 -1.0910308 -0.33026248 0.74205977
-0.4333373 -0.37693226 -0.13437249 0.49830082 0.24835739 0.8020679
-0.22029272 -0.69943064 0.5092019 0.3760157 0.08948018 -0.6226941
-0.86990327 0.31186384 -0.05808 -0.44492212 -0.63195884 -0.62048715
-1.3666245 0.07072604 0.16786477 0.03751421 0.09457739 1.0667624
0.7772511 0.994848 0.2318943 -0.43311203 -0.52432054 -0.99092096
-0.3478208 0.6332509 0.520515 -0.42747855 0.22720364 0.41327646] | [7.8908257484436035, 6.397800445556641] |
d6a7d0bb-96fb-47a6-bdcb-a5082afdfa0a | differentially-private-submodular | null | null | https://icml.cc/Conferences/2017/Schedule?showEvent=637 | http://proceedings.mlr.press/v70/mitrovic17a/mitrovic17a.pdf | Differentially Private Submodular Maximization: Data Summarization in Disguise |
Many data summarization applications are captured by the general framework of submodular maximization. As a consequence, a wide range of efficient approximation algorithms have been developed. However, when such applications involve sensitive data about individuals, their privacy concerns are not automatically addressed. To remedy this problem, we propose a general and systematic study of differentially private submodular maximization. We present privacy-preserving algorithms for both monotone and non-monotone submodular maximization under cardinality, matroid, and p-extendible system constraints, with guarantees that are competitive with optimal. Along the way, we analyze a new algorithm for non-monotone submodular maximization, which is the first (even non-privately) to achieve a constant approximation ratio while running in linear time. We additionally provide two concrete experiments to validate the efficacy of these algorithms.
| ['Andreas Krause', 'Amin Karbasi', 'Marko Mitrovic', 'Mark Bun'] | 2017-08-01 | null | null | null | icml-2017-8 | ['data-summarization'] | ['miscellaneous'] | [ 3.89050186e-01 4.06738698e-01 -4.11302179e-01 -4.68394011e-01
-7.22989559e-01 -9.36690629e-01 -3.91953513e-02 4.81523067e-01
-6.83762878e-02 9.92627561e-01 1.56194478e-01 2.52719969e-01
-4.56141531e-01 -1.05185556e+00 -7.08733022e-01 -7.80937433e-01
-7.79312178e-02 5.53427815e-01 -1.27783835e-01 -1.22077428e-01
1.69756815e-01 3.25020164e-01 -1.32342744e+00 1.01759113e-01
9.46289420e-01 8.12072933e-01 -3.94683093e-01 3.53326887e-01
2.72051066e-01 2.01092869e-01 -3.81258607e-01 -8.34102988e-01
7.66335785e-01 -3.15177590e-01 -8.44141603e-01 6.24426126e-01
5.38004875e-01 -6.21289790e-01 -5.28811038e-01 1.26898062e+00
4.94133323e-01 -1.05837882e-01 2.67736286e-01 -1.62116468e+00
-5.26593328e-01 1.06384301e+00 -8.52801204e-01 -2.26965323e-01
4.29151326e-01 -2.07183197e-01 1.37828743e+00 -8.78086537e-02
6.56682849e-01 1.11560321e+00 3.28027040e-01 5.22320151e-01
-1.18487132e+00 -3.21015000e-01 1.08343534e-01 -2.12428480e-01
-1.34031725e+00 -2.53930151e-01 6.33690417e-01 1.21858612e-01
5.40885091e-01 1.11752379e+00 6.67120755e-01 2.90477782e-01
-1.57617867e-01 1.11292160e+00 7.73170352e-01 -6.35174960e-02
3.77250046e-01 5.10682225e-01 3.55468452e-01 1.58322945e-01
1.14043391e+00 -3.79324347e-01 -4.24061298e-01 -8.30258608e-01
1.39131829e-01 3.02382588e-01 -6.54420137e-01 -8.91363740e-01
-7.45299578e-01 6.91967070e-01 -1.38768023e-02 2.78123133e-02
-2.78884113e-01 2.83695281e-01 2.98053056e-01 5.51872313e-01
4.59621966e-01 3.41392070e-01 -4.42673475e-01 1.44857511e-01
-1.03821671e+00 8.17773521e-01 1.20459175e+00 1.35843575e+00
5.37927151e-01 -4.84849572e-01 -2.28287458e-01 4.14006978e-01
2.52600797e-02 4.17183340e-01 -1.27290443e-01 -1.18458092e+00
6.54479146e-01 7.67860472e-01 2.19536930e-01 -1.13658786e+00
-2.20180213e-01 -1.70726433e-01 -9.36652064e-01 -4.50521827e-01
3.36803734e-01 -3.07907552e-01 -9.34457555e-02 1.87823856e+00
6.89304769e-01 -4.20850456e-01 1.73798174e-01 6.82181180e-01
3.99358869e-01 6.57033563e-01 -4.38219845e-01 -9.28476453e-01
1.23830307e+00 -4.96836692e-01 -8.44872117e-01 1.42921776e-01
8.63971472e-01 -1.18896671e-01 3.46259534e-01 5.09165823e-01
-1.52900648e+00 5.08122265e-01 -9.51033473e-01 -2.24861607e-01
-5.94899692e-02 -5.78628659e-01 8.26215208e-01 1.16568089e+00
-8.70445192e-01 3.00393671e-01 -4.37010795e-01 -4.69915390e-01
7.10754693e-01 4.73654091e-01 -3.73785675e-01 -1.41064808e-01
-8.38350952e-01 2.63024747e-01 4.59086984e-01 -4.06037748e-01
-5.85265815e-01 -7.59425282e-01 -6.81210279e-01 3.57819438e-01
7.32784569e-01 -9.39547777e-01 1.02234185e+00 -6.02416635e-01
-8.05297554e-01 9.89651799e-01 -1.05951667e-01 -7.42662787e-01
8.47533643e-01 7.86475465e-02 1.61787748e-01 2.80858070e-01
-1.74435481e-01 3.41974854e-01 2.90171206e-01 -1.24345005e+00
-8.69004846e-01 -1.13140047e+00 5.12056172e-01 4.03209805e-01
-1.02223074e+00 -1.50136231e-02 -1.43657044e-01 -3.92304987e-01
1.42822534e-01 -7.41507590e-01 -4.97714460e-01 8.75623748e-02
-7.56053984e-01 -1.22405283e-01 5.80011487e-01 -4.08357918e-01
1.34002030e+00 -2.20469475e+00 2.63076276e-01 3.78956199e-01
5.11043727e-01 -1.82512030e-01 1.06891401e-01 7.61568427e-01
3.46616089e-01 2.94015199e-01 -7.28935122e-01 -4.28718805e-01
2.74920940e-01 1.82287872e-01 -4.04302299e-01 8.80241156e-01
-5.25130689e-01 8.49340379e-01 -5.62348187e-01 -2.93326199e-01
-4.38270926e-01 -1.66174352e-01 -7.48352110e-01 -8.44568163e-02
-4.46930766e-01 -1.18585743e-01 -6.26928091e-01 8.74837160e-01
1.39342737e+00 -1.10631540e-01 7.39894629e-01 3.57178986e-01
1.45062685e-01 -2.70471483e-01 -1.46887696e+00 1.42249084e+00
1.45774588e-01 4.20444272e-02 6.25505209e-01 -1.04859006e+00
6.25409961e-01 2.31320485e-01 8.52453172e-01 1.89136527e-02
1.58413768e-01 2.83537686e-01 -5.90193450e-01 -3.98586601e-01
8.65522027e-01 -1.06558725e-01 -3.58551592e-01 8.97247434e-01
-5.18825948e-01 1.31765693e-01 2.20099553e-01 4.37424719e-01
9.65605617e-01 -6.51829362e-01 5.30925274e-01 -4.27622527e-01
3.64760697e-01 -6.96942136e-02 7.00871944e-01 7.63133645e-01
-6.50100634e-02 6.94571078e-01 9.08056200e-01 -3.26428235e-01
-7.18245447e-01 -7.33115733e-01 -2.58940190e-01 8.19911182e-01
3.93627644e-01 -6.05891407e-01 -1.05033708e+00 -5.88429093e-01
5.80015838e-01 5.87353349e-01 -5.11421442e-01 6.26007933e-03
-1.81895450e-01 -1.03469598e+00 2.87118614e-01 2.16709778e-01
4.56403643e-01 -3.54562074e-01 -3.23187649e-01 1.06507428e-02
-6.98107183e-02 -1.04063737e+00 -7.27020144e-01 -2.67103791e-01
-1.09885883e+00 -9.50134814e-01 -6.70535803e-01 -5.76378524e-01
9.01497364e-01 7.09815323e-01 5.73819935e-01 1.01123922e-01
4.70079342e-03 6.42506838e-01 1.51521862e-02 -4.50204551e-01
2.96087749e-02 1.12925515e-01 -1.75621323e-02 2.44913518e-01
4.50915277e-01 -7.48585522e-01 -4.15825665e-01 1.67861991e-02
-1.30369413e+00 -3.66988868e-01 2.04983145e-01 3.14394504e-01
8.20058942e-01 2.92262614e-01 5.10038197e-01 -1.33435559e+00
7.77520537e-01 -7.68782020e-01 -9.22790408e-01 4.24683809e-01
-6.16913915e-01 -8.96128714e-02 4.46536183e-01 -2.57348530e-02
-8.28361392e-01 2.39333242e-01 2.64900416e-01 8.28144774e-02
3.26871015e-02 3.34107280e-01 -8.90210092e-01 -6.51256442e-02
3.08883846e-01 5.24590194e-01 2.10162014e-01 -4.29538488e-01
4.31937844e-01 8.93603384e-01 3.00149977e-01 -4.29307044e-01
6.54348016e-01 1.02730334e+00 4.36036617e-01 -7.26863861e-01
-6.77301943e-01 -4.70609367e-01 -2.21525699e-01 2.58400708e-01
4.10342425e-01 -7.72029459e-01 -8.97187173e-01 4.82885361e-01
-9.43515718e-01 4.74568248e-01 -6.37989819e-01 -6.20827405e-03
-7.49789357e-01 9.94133115e-01 -2.73565203e-01 -1.06636655e+00
-5.37369132e-01 -7.50979841e-01 6.47898853e-01 6.32048771e-02
-4.18694243e-02 -7.25603640e-01 7.44814947e-02 7.29290009e-01
1.95643321e-01 4.80842441e-01 5.49776912e-01 -8.83765280e-01
-1.13315785e+00 -4.12505001e-01 -2.04116646e-02 1.67692211e-02
8.53057131e-02 -4.57451910e-01 -7.59406686e-01 -6.89188957e-01
1.93293035e-01 -1.66243777e-01 8.35160851e-01 4.72452044e-01
1.45521772e+00 -1.00852239e+00 -4.23561424e-01 8.63414645e-01
1.43174732e+00 -3.88177574e-01 5.69173396e-01 8.45935568e-02
3.11193079e-01 8.55901659e-01 6.27165675e-01 1.18974221e+00
7.12486625e-01 4.53730702e-01 5.80560088e-01 4.74980652e-01
7.98986733e-01 -2.80297667e-01 1.46524653e-01 1.76501811e-01
8.80095884e-02 -6.95450962e-01 -3.09629906e-02 8.38842273e-01
-2.07350612e+00 -9.59174454e-01 -3.65650237e-01 2.53776860e+00
8.18129778e-01 -4.27285850e-01 5.84829867e-01 2.27104947e-01
7.40951598e-01 1.95283875e-01 -6.76166117e-01 -4.84431654e-01
-5.00255466e-01 -3.71925861e-01 9.10922825e-01 2.27110758e-01
-8.59333515e-01 4.05854911e-01 6.40864325e+00 7.21814275e-01
-2.82649010e-01 3.77570242e-02 7.44948268e-01 -5.68540275e-01
-1.12177241e+00 7.65833855e-02 -7.73330569e-01 3.27446669e-01
5.44480383e-01 -1.01393080e+00 4.41078424e-01 1.01178491e+00
-1.00443922e-01 -2.56082565e-02 -1.29148817e+00 1.16630673e+00
-7.15525541e-03 -1.35823071e+00 1.13135375e-01 7.55422473e-01
1.11889517e+00 -7.20926166e-01 8.40124860e-02 -3.15173805e-01
-7.87135214e-02 -7.20854521e-01 2.51529276e-01 -1.21131632e-02
4.27272767e-01 -1.12831247e+00 5.20600259e-01 5.91095030e-01
-6.37482941e-01 -3.76082152e-01 -7.03319907e-01 4.60346863e-02
4.33102816e-01 9.01022017e-01 -1.29820868e-01 7.51789868e-01
5.44172227e-01 3.49206179e-01 1.85125005e-02 1.14343441e+00
1.93719506e-01 2.11022109e-01 -7.77838588e-01 -8.07588175e-02
-1.77329108e-01 -2.31061444e-01 9.70298290e-01 8.85716617e-01
4.07137752e-01 3.95439655e-01 -7.29983672e-02 7.17694938e-01
-5.83860576e-01 4.17562157e-01 -7.99399674e-01 -1.04182966e-01
5.83749950e-01 1.22017121e+00 -2.82146305e-01 1.47519922e-02
-3.15430820e-01 8.93294632e-01 3.08622308e-02 -6.91478550e-02
-6.42695069e-01 -1.81641102e-01 9.88369882e-01 5.23296952e-01
2.21900597e-01 7.89684653e-02 -6.81894243e-01 -1.13300455e+00
4.46756691e-01 -7.69615829e-01 1.04481065e+00 1.07709624e-01
-1.29168260e+00 -1.54921085e-01 2.92526186e-01 -7.95501888e-01
6.20298497e-02 -1.27430856e-01 -3.54945719e-01 1.98078424e-01
-1.23121893e+00 -8.87657106e-01 -1.36172295e-01 5.99254131e-01
-2.37699732e-01 1.03646852e-01 4.27199215e-01 1.68821171e-01
-3.67051601e-01 1.06326175e+00 3.67931128e-01 -5.54898560e-01
3.01366031e-01 -1.18920600e+00 -1.13774076e-01 8.67440522e-01
-2.04507634e-01 5.16920626e-01 7.74497330e-01 -5.00504911e-01
-1.98631787e+00 -1.04547024e+00 1.09327364e+00 -2.90587753e-01
2.24308908e-01 -3.88126850e-01 -7.56281257e-01 8.46082211e-01
1.24405093e-01 -1.29497156e-01 8.99658918e-01 -8.16084892e-02
-3.87260504e-02 -5.97217560e-01 -2.02669621e+00 3.73156965e-01
1.36261463e+00 8.77064466e-02 -2.42780581e-01 6.87271357e-01
8.93043399e-01 -3.29400718e-01 -9.03797090e-01 3.15302640e-01
2.85769522e-01 -8.24980021e-01 8.39114904e-01 -6.64015830e-01
1.15370631e-01 -1.10910952e-01 -4.59054738e-01 -6.85754836e-01
-1.24328852e-01 -1.23625052e+00 -6.19744241e-01 1.34687066e+00
2.00679854e-01 -8.96588743e-01 1.15326607e+00 1.13598287e+00
4.41349059e-01 -7.78702974e-01 -9.79906559e-01 -8.00674081e-01
8.97955745e-02 -2.90042292e-02 1.02853835e+00 9.61297333e-01
4.09371555e-01 -1.85674056e-01 -7.81923950e-01 3.07838887e-01
1.04927981e+00 6.57548487e-01 9.79525626e-01 -1.09153962e+00
-4.69113618e-01 -1.32497296e-01 -4.97250259e-01 -1.37404478e+00
-7.72287622e-02 -9.16483700e-01 -4.62686419e-01 -1.36212397e+00
1.04469514e+00 -1.46046504e-01 -7.27004046e-03 3.52758616e-01
6.84621856e-02 -1.46576479e-01 9.40320790e-02 -7.02882558e-02
-8.33638191e-01 5.06620944e-01 1.07308161e+00 -4.99354564e-02
-2.29011029e-01 5.05017519e-01 -1.48581994e+00 2.70360231e-01
9.95495617e-01 -5.12551427e-01 -5.19034028e-01 -2.54760355e-01
4.24072355e-01 2.09790349e-01 8.92712250e-02 -4.41150904e-01
2.64621496e-01 -5.61383247e-01 -3.78240049e-01 -7.98570275e-01
1.27525792e-01 -9.81846273e-01 4.61338788e-01 4.50784653e-01
-3.84896845e-01 -2.82090753e-01 -2.59748191e-01 7.05905497e-01
1.47019746e-02 -3.88963580e-01 7.40369856e-01 -7.53889382e-02
-8.27600881e-02 8.06706250e-01 -1.45138934e-01 2.67963737e-01
1.66314399e+00 -3.19399983e-01 -3.63553792e-01 -7.97988534e-01
-2.74435788e-01 7.85124183e-01 7.87823617e-01 -1.68199599e-01
6.11162961e-01 -9.89941180e-01 -9.95169818e-01 -7.15962499e-02
1.75643653e-01 2.33366549e-01 6.33099794e-01 8.35069656e-01
-3.11303973e-01 4.48459566e-01 1.67242922e-02 -1.67331144e-01
-1.41926181e+00 8.80209744e-01 9.43903029e-02 -2.79586285e-01
-3.91236722e-01 6.81687474e-01 2.38137111e-01 -3.95415217e-01
3.03032070e-01 8.93607512e-02 2.99745858e-01 4.30301391e-02
6.09988213e-01 7.78252542e-01 -2.56597191e-01 -3.73919606e-01
-2.76600838e-01 1.12817727e-01 -2.79836744e-01 -2.62932125e-02
1.55871367e+00 -4.19656992e-01 -6.61174357e-01 -2.00439930e-01
1.03178585e+00 3.16686094e-01 -7.71435320e-01 -1.75301120e-01
-2.47391298e-01 -7.88012505e-01 -2.88507581e-01 -1.77612633e-01
-1.33261561e+00 3.28035146e-01 4.47828099e-02 6.38904154e-01
1.43055129e+00 1.53371960e-01 1.13136339e+00 4.81817096e-01
6.50965869e-01 -1.08725500e+00 -6.42397404e-01 -2.38532633e-01
7.46183574e-01 -8.49927604e-01 4.39688265e-01 -8.54944646e-01
-4.36162263e-01 8.12445998e-01 4.76105273e-01 1.46117896e-01
5.70080519e-01 3.30722094e-01 -6.49892867e-01 -9.07465070e-02
-8.29695284e-01 2.10162684e-01 -2.03167722e-01 4.43523020e-01
-2.09511355e-01 2.92335927e-01 -9.65413511e-01 1.04575264e+00
-7.54767060e-02 3.00898701e-02 8.59847009e-01 1.00802958e+00
-5.94779730e-01 -1.24230194e+00 -3.52832824e-01 4.78955925e-01
-8.02311778e-01 2.64769405e-01 -7.43594944e-01 4.45761323e-01
-2.13947505e-01 9.80760694e-01 -1.71236709e-01 -7.64155686e-02
6.33413717e-02 -5.64380944e-01 4.96365219e-01 -2.53516376e-01
-5.28819263e-01 -2.31706172e-01 2.30551381e-02 -3.52376282e-01
-1.44767448e-01 -8.56227875e-01 -1.01675391e+00 -8.27254176e-01
-3.37807178e-01 1.75324708e-01 1.60530463e-01 4.12439674e-01
4.14173424e-01 -4.46359098e-01 9.05043185e-01 1.16040051e-01
-1.26708639e+00 -3.80310893e-01 -1.22013879e+00 4.07706350e-01
8.56270194e-02 1.30394369e-01 -4.23913836e-01 -4.08031702e-01] | [6.554113864898682, 4.970730304718018] |
63b66670-7b4d-416f-a554-77ef772a5a99 | development-and-validation-of-a-natural | 2303.13451 | null | https://arxiv.org/abs/2303.13451v1 | https://arxiv.org/pdf/2303.13451v1.pdf | Development and validation of a natural language processing algorithm to pseudonymize documents in the context of a clinical data warehouse | The objective of this study is to address the critical issue of de-identification of clinical reports in order to allow access to data for research purposes, while ensuring patient privacy. The study highlights the difficulties faced in sharing tools and resources in this domain and presents the experience of the Greater Paris University Hospitals (AP-HP) in implementing a systematic pseudonymization of text documents from its Clinical Data Warehouse. We annotated a corpus of clinical documents according to 12 types of identifying entities, and built a hybrid system, merging the results of a deep learning model as well as manual rules. Our results show an overall performance of 0.99 of F1-score. We discuss implementation choices and present experiments to better understand the effort involved in such a task, including dataset size, document types, language models, or rule addition. We share guidelines and code under a 3-Clause BSD license. | ['Romain Bey', 'Martin Hilka', 'Alexandre Mouchet', 'Basile Dura', 'Alice Calliger', 'Perceval Wajsbürt', 'Xavier Tannier'] | 2023-03-23 | null | null | null | null | ['de-identification'] | ['natural-language-processing'] | [-4.50549535e-02 3.90617073e-01 -7.34185576e-02 -4.76185292e-01
-6.41068161e-01 -5.13805687e-01 2.58228213e-01 1.06608522e+00
-7.83772826e-01 8.82616758e-01 2.29094774e-01 -7.30665147e-01
-4.56763655e-01 -6.77798986e-01 -3.35994363e-01 -3.50910217e-01
4.24770117e-02 8.63825083e-01 -3.24275225e-01 2.47395471e-01
1.19747654e-01 7.06928551e-01 -8.22622657e-01 6.88212216e-01
6.14689410e-01 6.97977066e-01 -3.28959301e-02 6.00227654e-01
-4.02950525e-01 9.33764577e-01 -7.44575143e-01 -7.62152314e-01
2.76418686e-01 9.92185473e-02 -1.08325636e+00 -1.18100010e-01
9.59686860e-02 -3.85461390e-01 -2.63575017e-02 8.28390360e-01
6.02193594e-01 -1.80328578e-01 5.06483257e-01 -1.11248827e+00
-5.04625618e-01 6.62805974e-01 8.25554058e-02 7.11730421e-02
4.50871378e-01 -6.11267723e-02 7.13590324e-01 -5.55271327e-01
1.13021100e+00 3.93001527e-01 7.63000131e-01 4.42261130e-01
-7.38229990e-01 -5.97867250e-01 -4.42960829e-01 -5.03940694e-02
-1.54529727e+00 -5.54904044e-01 8.92013162e-02 -6.79414809e-01
1.29014361e+00 2.82268524e-01 4.43185478e-01 8.16004574e-01
3.47427398e-01 3.73613417e-01 6.10331297e-01 -6.71606660e-01
3.28598887e-01 7.39400506e-01 2.69409806e-01 7.30884075e-01
7.99054384e-01 -5.59557974e-01 -5.39119653e-02 -6.36944830e-01
5.76577902e-01 1.21131446e-02 -4.47566137e-02 -3.01908672e-01
-9.01264548e-01 8.17937672e-01 -6.16596490e-02 7.69158542e-01
-5.51293254e-01 -6.08353853e-01 6.67998731e-01 7.96622038e-02
1.41123571e-02 9.40110624e-01 -8.77890050e-01 -1.07370101e-01
-9.77293789e-01 1.46625549e-01 1.26437593e+00 1.24048555e+00
1.20242946e-01 -6.14299655e-01 -1.03796311e-01 7.72137523e-01
1.24636292e-01 -1.37780458e-01 5.92485368e-01 -8.30776274e-01
4.74573046e-01 7.25362420e-01 3.07976365e-01 -9.22687888e-01
-7.72846580e-01 -3.81571352e-01 -7.09415436e-01 -3.27137023e-01
4.69277233e-01 -4.78111058e-01 -6.14832640e-01 1.20073855e+00
5.09267822e-02 -5.73202729e-01 1.96965888e-01 2.59548903e-01
9.49507058e-01 1.26576811e-01 4.90752399e-01 -4.07208860e-01
1.58637083e+00 -3.40976596e-01 -1.19346941e+00 2.45417029e-01
1.22307372e+00 -8.58446836e-01 2.72640675e-01 4.08510804e-01
-1.13223231e+00 -1.50381997e-01 -8.03441226e-01 -1.68802246e-01
-8.86897743e-01 3.96857053e-01 5.74409604e-01 7.51520395e-01
-8.69017601e-01 4.66067433e-01 -9.14862931e-01 -8.15231264e-01
6.06931031e-01 5.24789035e-01 -6.88558578e-01 1.81111559e-01
-1.16965008e+00 1.08400786e+00 7.72432208e-01 -1.02099948e-01
3.45245935e-02 -7.93204546e-01 -7.58318365e-01 7.85837993e-02
2.58855909e-01 -6.95290506e-01 1.12913895e+00 -2.17439577e-01
-7.28605151e-01 1.21035600e+00 4.85327430e-02 -4.97931153e-01
6.85923934e-01 -3.76125164e-02 -8.63045812e-01 1.42821610e-01
1.89866602e-01 3.59458953e-01 -1.61090612e-01 -7.71744192e-01
-9.61252213e-01 -3.28307241e-01 -1.88952014e-01 -2.47149587e-01
-4.14241254e-01 3.24570388e-01 -3.49644989e-01 -4.29990411e-01
-4.17314023e-01 -6.96398497e-01 -3.59090030e-01 -1.13878816e-01
-4.21429545e-01 -8.12094212e-02 2.80223966e-01 -1.15598083e+00
1.64544082e+00 -1.94492209e+00 -5.47554970e-01 3.02941918e-01
6.33117199e-01 5.32640576e-01 3.30674410e-01 6.85997427e-01
-2.24652514e-01 5.06696165e-01 -8.34805891e-02 -1.32656828e-01
-8.40055868e-02 2.42305677e-02 2.84081362e-02 1.95402086e-01
1.60866633e-01 7.43321180e-01 -7.84318149e-01 -7.78388143e-01
-1.52163464e-03 4.14061219e-01 -3.88663173e-01 1.03277817e-01
1.09094024e-01 7.61253983e-02 -5.09165049e-01 6.52546108e-01
4.98083353e-01 -4.31144923e-01 7.26779103e-01 -3.35139751e-01
-1.27842501e-01 3.86665732e-01 -1.17174280e+00 1.31379211e+00
-1.97936177e-01 4.43693280e-01 1.44765273e-01 -5.20830989e-01
8.73281837e-01 6.31190717e-01 9.83936250e-01 -3.45532477e-01
3.51480573e-01 3.71471465e-01 4.66932170e-02 -9.14332211e-01
6.31786227e-01 2.15128079e-01 1.04472339e-01 3.62492681e-01
1.12197854e-01 2.42884710e-01 3.71399224e-01 1.79963306e-01
1.24105036e+00 -2.80165106e-01 9.87218022e-01 -1.30176455e-01
4.79170471e-01 2.66574681e-01 5.17550766e-01 6.54501796e-01
-2.82095402e-01 3.95315021e-01 6.16591573e-01 -6.70915246e-01
-1.01782322e+00 -3.87020260e-01 -5.34210622e-01 3.95616829e-01
-6.31138921e-01 -7.48947144e-01 -6.33028924e-01 -8.94964874e-01
1.85679406e-01 1.08390582e+00 -7.20620513e-01 1.17579430e-01
-4.67913926e-01 -8.01692426e-01 8.39691162e-01 4.58292097e-01
-2.40488723e-02 -1.08076513e+00 -1.11076188e+00 4.79923695e-01
2.43852288e-02 -1.28926349e+00 -2.17272490e-01 4.29731816e-01
-4.38162148e-01 -1.42281461e+00 -4.36577260e-01 -5.25706947e-01
6.97196066e-01 -6.12089694e-01 9.95975077e-01 -1.44635448e-02
-6.60421252e-01 2.57442653e-01 -2.53815711e-01 -7.45838344e-01
-7.84394145e-01 2.12091804e-01 -2.95236349e-01 -5.66279292e-01
1.24987924e+00 1.46128625e-01 -3.08268368e-01 -7.46856704e-02
-1.22635007e+00 -3.88492852e-01 6.71912253e-01 6.55754447e-01
2.78718740e-01 6.16379119e-02 3.83438021e-01 -1.27974558e+00
1.05559897e+00 -5.09252369e-01 -4.52954769e-01 5.59544086e-01
-9.94517207e-01 -1.58630520e-01 3.74892503e-01 1.51262254e-01
-7.93521583e-01 2.11193785e-01 -2.99189717e-01 -1.69214346e-02
-5.15432119e-01 6.31174386e-01 -9.50731114e-02 2.62505084e-01
4.64916468e-01 -2.43997663e-01 2.36041620e-01 -4.19987440e-01
9.27013606e-02 1.05357420e+00 1.60786852e-01 -1.79444835e-01
9.00661722e-02 1.33576691e-01 -2.41770759e-01 -5.61317205e-01
-3.77736658e-01 -6.73879623e-01 -8.85844827e-01 3.09747249e-01
8.90338898e-01 -6.76822662e-01 -7.53234208e-01 -5.17068990e-02
-1.20862937e+00 1.67061493e-01 -3.67783070e-01 6.48856103e-01
-2.80729830e-01 2.00232774e-01 -7.27485538e-01 -5.04511535e-01
-5.57139277e-01 -1.09549499e+00 7.58289278e-01 -6.49595857e-02
-8.63639772e-01 -1.12196743e+00 7.10903332e-02 2.54847616e-01
3.55678707e-01 2.26074725e-01 1.11348140e+00 -1.58588016e+00
3.83347124e-02 -4.71314937e-01 -2.17285410e-01 1.20123230e-01
4.31796193e-01 1.08863525e-01 -7.81554222e-01 -1.04784638e-01
7.32514486e-02 6.81108683e-02 3.24265689e-01 2.83558011e-01
1.11400414e+00 -5.03815055e-01 -5.67416072e-01 3.35713178e-01
1.38864756e+00 7.08915174e-01 4.64468151e-01 6.38757229e-01
4.75838810e-01 8.51028502e-01 5.48020244e-01 6.74889266e-01
1.93164244e-01 3.53560477e-01 -1.21182509e-01 -1.94601994e-02
3.51387590e-01 8.06063134e-03 -4.05430079e-01 4.96257126e-01
1.64477915e-01 -2.07572803e-01 -1.28505361e+00 5.99258661e-01
-1.59818971e+00 -5.15715301e-01 -5.30510172e-02 2.13554025e+00
1.01345956e+00 1.52441813e-02 2.00209226e-02 -5.22843264e-02
5.34117699e-01 -4.96131927e-01 -2.15962708e-01 -5.95564961e-01
1.53300390e-01 1.75437331e-01 8.78632545e-01 3.24313343e-01
-1.10673070e+00 4.43010807e-01 7.09426546e+00 3.84443820e-01
-1.01853430e+00 -6.82130605e-02 4.73328322e-01 -1.82036281e-01
2.94436049e-02 -3.89325947e-01 -9.57261682e-01 4.15573806e-01
1.20488501e+00 -4.46515292e-01 -1.44561961e-01 7.81798720e-01
1.15662195e-01 1.16881810e-01 -1.32103539e+00 6.86685860e-01
1.13364182e-01 -1.77113771e+00 -1.02064177e-01 4.99475151e-01
2.28648484e-01 8.66886135e-03 -2.37288550e-01 4.81448583e-02
4.17686164e-01 -9.78272021e-01 3.47366035e-01 4.78646815e-01
7.48350441e-01 -5.16381860e-01 1.26292562e+00 3.67686749e-02
-4.71612334e-01 -2.20956087e-01 -6.15718886e-02 3.11310410e-01
-5.71705177e-02 5.43477476e-01 -1.62321329e+00 7.76706457e-01
6.92068040e-01 4.22048569e-01 -6.24081314e-01 1.12140751e+00
3.41951519e-01 1.03230691e-02 -1.46719620e-01 -3.27412002e-02
8.19757134e-02 1.40662715e-01 1.45467088e-01 1.69024289e+00
2.61261940e-01 3.12003605e-02 -1.01129927e-01 6.90109730e-01
-1.32835701e-01 4.60184783e-01 -5.46454310e-01 -5.08145273e-01
6.82758510e-01 1.29259586e+00 -7.75746286e-01 -5.41608095e-01
-4.09409553e-01 3.59437317e-01 7.54519254e-02 6.60808478e-03
-5.53587079e-01 -6.89167500e-01 4.02166694e-01 2.91529179e-01
3.03461224e-01 6.90885261e-02 -3.91835809e-01 -8.56201410e-01
9.22434703e-02 -1.15508962e+00 8.70421410e-01 -6.27173126e-01
-1.04296207e+00 7.51321137e-01 3.44983535e-03 -1.15396130e+00
-3.88805360e-01 -7.68978953e-01 2.06383035e-01 8.87405574e-01
-1.06411994e+00 -1.02052665e+00 -8.04369375e-02 1.38608739e-01
-6.90220967e-02 -3.83473963e-01 1.31336451e+00 7.61077344e-01
-5.74384570e-01 6.90579236e-01 2.93832153e-01 5.56343496e-01
9.95629907e-01 -1.10469973e+00 1.33299669e-02 4.54973996e-01
-1.68796241e-01 1.33703125e+00 5.71231365e-01 -8.44884157e-01
-1.05064857e+00 -1.10011399e+00 1.61394715e+00 -8.04311216e-01
7.13342011e-01 -1.08273141e-01 -9.11642253e-01 8.22338402e-01
4.09865111e-01 -3.80037695e-01 1.43487501e+00 -3.22362520e-02
-7.94655606e-02 1.34359822e-01 -1.64206088e+00 2.57380515e-01
5.53788602e-01 -4.02292162e-01 -7.80376673e-01 5.04469693e-01
4.19449270e-01 -4.38536108e-01 -1.37557936e+00 2.85568208e-01
5.31091511e-01 -5.79347730e-01 6.38706565e-01 -9.48663652e-01
4.26786721e-01 -1.24843217e-01 1.55911567e-02 -8.43146443e-01
-3.36902827e-01 -4.78203326e-01 7.34007955e-02 1.13710046e+00
5.84974229e-01 -7.37921715e-01 5.91022074e-01 1.22697103e+00
9.06705260e-02 -6.88414395e-01 -6.72833920e-01 -3.11858803e-01
2.70009544e-02 -2.77118295e-01 8.57867718e-01 1.36442399e+00
2.92695761e-01 -8.96782875e-02 1.87210336e-01 9.64992791e-02
1.47690222e-01 -3.34693223e-01 4.42471564e-01 -1.20444489e+00
8.21999554e-03 -5.27483821e-01 -3.05892169e-01 -5.92829799e-03
-2.37064168e-01 -8.61538947e-01 -4.40555513e-01 -1.74572003e+00
1.92210421e-01 -5.64381957e-01 -4.58229363e-01 7.84503520e-01
5.86143807e-02 -4.02215630e-01 3.72622572e-02 3.38250786e-01
-4.46161449e-01 -4.41018164e-01 6.19638503e-01 4.95394170e-02
-3.36036950e-01 -2.66902000e-01 -1.13682485e+00 6.01837873e-01
6.73558831e-01 -7.43018806e-01 -7.29708374e-02 -4.22714949e-01
1.86175212e-01 -1.70782626e-01 8.91105458e-03 -5.96515834e-01
6.35931790e-01 -3.65350805e-02 6.43832862e-01 -7.25756586e-01
-9.63631943e-02 -1.21733189e+00 4.53302652e-01 6.04850352e-01
-7.32966661e-01 1.26968220e-01 4.69187021e-01 5.91843128e-02
-2.16199219e-01 -5.31479716e-01 3.88074994e-01 -3.64826620e-01
-3.41319650e-01 -9.77125391e-03 -4.54684556e-01 -1.61863461e-01
1.19076407e+00 -1.87951386e-01 -4.28121567e-01 6.10156022e-02
-9.50372815e-01 3.50760967e-01 5.37033141e-01 3.48946095e-01
2.34311253e-01 -8.55945289e-01 -5.55846691e-01 3.34501535e-01
3.50881636e-01 -1.60084233e-01 5.50172888e-02 7.96424925e-01
-9.47028458e-01 9.66884494e-01 -3.98697108e-01 -1.37049764e-01
-1.44747806e+00 5.75549901e-01 2.75321692e-01 -6.46723926e-01
-5.87178826e-01 4.49287236e-01 -3.79264653e-01 -4.85713631e-01
4.06818718e-01 -3.29691231e-01 -5.20422637e-01 3.35003555e-01
6.72703385e-01 3.25706601e-01 7.42470503e-01 -2.81084359e-01
-6.23206556e-01 1.33823603e-02 -5.59682012e-01 6.05743453e-02
1.47608304e+00 1.37799114e-01 -2.51072973e-01 1.16718397e-01
1.25997055e+00 2.00713351e-01 -2.34477699e-01 3.78117971e-02
4.08148408e-01 -1.15124449e-01 -1.74678802e-01 -1.15240371e+00
-8.51454914e-01 5.24071693e-01 4.62443441e-01 4.34555382e-01
9.34388816e-01 -1.67601869e-01 4.76866156e-01 5.95965505e-01
4.80426475e-02 -1.25598967e+00 -8.01612318e-01 3.21292341e-01
6.10658526e-01 -1.01855338e+00 3.06007534e-01 -3.25583786e-01
-7.84821570e-01 1.26694703e+00 2.43366599e-01 5.48152685e-01
8.46565008e-01 5.50317049e-01 2.80051738e-01 -4.06176984e-01
-6.22632921e-01 2.42425248e-01 -1.30573809e-01 5.93741417e-01
8.41965139e-01 1.08725131e-01 -7.00004935e-01 7.29118109e-01
-1.63366109e-01 7.58742511e-01 5.20674109e-01 1.40504599e+00
1.39391243e-01 -1.32782352e+00 -2.01116338e-01 8.57486546e-01
-1.05256271e+00 -1.64399326e-01 -6.71375215e-01 9.00825441e-01
3.75842333e-01 7.48568714e-01 7.15366825e-02 -2.09526956e-01
4.28671718e-01 3.30178231e-01 1.40917912e-01 -6.96953475e-01
-9.50930655e-01 1.09896109e-01 5.57318866e-01 -2.00496018e-01
-1.76905587e-01 -8.31699371e-01 -1.11536372e+00 1.35853486e-02
-3.16104293e-02 5.27965069e-01 8.66898000e-01 7.51070678e-01
8.39840829e-01 7.24401534e-01 -1.95481963e-02 3.18989724e-01
-4.39676851e-01 -7.75748312e-01 -4.77055132e-01 4.37055230e-01
1.79451734e-01 -9.54470932e-02 -4.84490097e-02 3.33417237e-01] | [8.422920227050781, 8.62608528137207] |
b3ea809d-d657-401a-953b-5d8ab76dc4d9 | vibration-fault-detection-in-wind-turbines | 2206.12452 | null | https://arxiv.org/abs/2206.12452v1 | https://arxiv.org/pdf/2206.12452v1.pdf | Vibration fault detection in wind turbines based on normal behaviour models without feature engineering | Most wind turbines are remotely monitored 24/7 to allow for an early detection of operation problems and developing damage. We present a new fault detection method for vibration-monitored drivetrains that does not require any feature engineering. Our method relies on a simple model architecture to enable a straightforward implementation in practice. We propose to apply convolutional autoencoders for identifying and extracting the most relevant features from the half spectrum in an automated manner, saving time and effort. Thereby, a spectral model of the normal vibration response is learnt for the monitored component from past measurements. We demonstrate that the model can successfully distinguish damaged from healthy components and detect a damaged generator bearing and damaged gearbox parts from their vibration responses. Using measurements from commercial wind turbines and a test rig, we show that vibration-based fault detection in wind turbine drivetrains can be performed without the usual upfront definition of spectral features. Another advantage of the presented method is that the entire half spectrum is monitored instead of the usual focus on monitoring individual frequencies and harmonics. | ['Angela Meyer', 'Bernhard Brodbeck', 'Dimitrios Anagnostos', 'Stefan Jonas'] | 2022-06-24 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [ 1.58874858e-02 -1.51232988e-01 5.35148978e-01 7.10461661e-02
-2.70624906e-01 -4.48851138e-01 1.22221395e-01 6.14097863e-02
6.64223731e-02 2.92661190e-01 -3.33799630e-01 -1.40802125e-02
-4.57072943e-01 -7.73491383e-01 -4.78615880e-01 -8.27542126e-01
-5.27558565e-01 5.53884581e-02 1.20059043e-01 -3.29382658e-01
-1.49054170e-01 7.70389140e-01 -2.20547032e+00 -4.15810011e-02
4.81459498e-01 9.50786173e-01 4.36156332e-01 8.50947082e-01
4.93566751e-01 2.60568112e-01 -1.38120902e+00 4.28971291e-01
1.57230616e-01 -5.04441917e-01 -7.13721275e-01 4.98944402e-01
1.71742007e-01 -7.22889721e-01 -1.29413366e-01 8.29180300e-01
6.86084151e-01 1.69583648e-01 6.91096306e-01 -1.12027621e+00
2.70552672e-02 3.40432882e-01 -1.39948592e-01 3.55631471e-01
1.26106873e-01 2.25597061e-02 7.82590389e-01 -7.47741699e-01
2.16852576e-01 4.90341783e-01 8.04812908e-01 2.72859126e-01
-1.15737355e+00 -2.49074608e-01 -2.95589954e-01 3.81617218e-01
-1.19892359e+00 -1.37591511e-01 1.04509580e+00 -7.43886411e-01
1.21470952e+00 1.68238297e-01 7.06823349e-01 7.66743004e-01
3.55305523e-01 3.24223846e-01 5.90341926e-01 -5.57209313e-01
3.89577359e-01 -2.54997164e-01 -4.77710832e-03 6.43252730e-01
1.97623149e-01 4.09929812e-01 -3.18012267e-01 8.39059576e-02
7.16051042e-01 -5.66620380e-02 -8.35949242e-01 -1.54751375e-01
-9.07183468e-01 6.39387012e-01 1.79782063e-01 7.14382887e-01
-5.41288137e-01 2.23292619e-01 5.89681506e-01 7.66421676e-01
4.39139456e-01 5.19480109e-01 -6.59201503e-01 -2.36056447e-01
-8.84948432e-01 -8.44313949e-02 6.88661814e-01 2.67191768e-01
6.07157171e-01 7.61745691e-01 5.92246234e-01 5.54104209e-01
-3.75336483e-02 4.07570750e-01 9.04795945e-01 -6.78258836e-01
-2.19217956e-01 1.77252010e-01 2.57675320e-01 -8.10895026e-01
-6.13790154e-01 -4.80069906e-01 -6.00319088e-01 7.97267437e-01
2.41093757e-03 -6.31214738e-01 -5.60020924e-01 1.15478742e+00
3.72180551e-01 2.26206303e-01 -8.65110978e-02 9.12060678e-01
2.72305846e-01 3.87334973e-01 -7.08752871e-01 -2.51297563e-01
1.25182807e+00 -3.49258393e-01 -8.91482174e-01 2.71672965e-03
6.62863374e-01 -6.30972683e-01 7.82242715e-01 8.02886844e-01
-6.86739564e-01 -9.26080525e-01 -1.72102821e+00 3.99285287e-01
-5.43625116e-01 5.18781543e-01 1.16778627e-01 5.59329689e-01
-9.66265857e-01 1.36593926e+00 -1.10626030e+00 -1.63421154e-01
-2.57854730e-01 2.76985049e-01 -6.08471394e-01 6.85860813e-01
-1.10162652e+00 1.09209991e+00 4.90939260e-01 4.40058380e-01
-9.82295334e-01 -6.70919776e-01 -8.15909088e-01 3.38776827e-01
3.17402542e-01 -2.78420508e-01 1.53017950e+00 -7.15290368e-01
-1.64081371e+00 5.32380603e-02 2.76816368e-01 -5.73005438e-01
1.68482855e-01 -6.93809688e-01 -6.60775959e-01 7.12119639e-01
-3.34015548e-01 -3.54175627e-01 1.67740214e+00 -1.02408397e+00
-6.52204692e-01 3.38593572e-02 -2.72399276e-01 -2.81694651e-01
-6.90663695e-01 -2.85870314e-01 7.33402908e-01 -6.77476704e-01
7.30956113e-03 -5.20626605e-01 2.57944018e-01 -3.13946456e-01
-1.47388175e-01 -1.83345392e-01 1.27465701e+00 -7.67497897e-01
9.88937736e-01 -2.14125085e+00 7.75305554e-02 1.29427582e-01
2.43408531e-01 4.07360852e-01 1.17867202e-01 9.16225314e-01
-5.31891584e-01 -3.02269399e-01 -2.59939525e-02 4.21337783e-02
4.76680286e-02 2.11966693e-01 -4.28089559e-01 6.21585667e-01
9.63783622e-01 1.84324950e-01 -7.30764031e-01 3.78394812e-01
5.87330937e-01 5.58434248e-01 -1.19294740e-01 4.55582112e-01
8.52471441e-02 1.53828844e-01 -1.12511829e-01 2.52602488e-01
3.33106279e-01 3.99874687e-01 -1.57418251e-01 -3.91359240e-01
-1.71927929e-01 1.75506741e-01 -1.32538068e+00 1.18296158e+00
-7.93179333e-01 7.82733440e-01 3.67998272e-01 -1.42393351e+00
9.83761609e-01 8.27536821e-01 6.56196654e-01 -1.38907164e-01
1.76233351e-01 2.98718095e-01 -1.10618677e-02 -8.67415428e-01
1.87289432e-01 -2.15470120e-01 1.24281108e-01 4.42580223e-01
6.76164210e-01 -3.97735298e-01 1.47127599e-01 -5.32111764e-01
1.12111068e+00 1.29876807e-01 -9.85983666e-03 -3.69305491e-01
5.30333757e-01 -1.33867621e-01 3.89902145e-01 1.63592592e-01
1.00915000e-01 3.99271667e-01 3.46798420e-01 -4.50681627e-01
-7.06983745e-01 -8.15748811e-01 -1.02917120e-01 6.31834328e-01
-3.17221731e-01 -3.87046963e-01 -8.04081261e-01 -5.65799236e-01
1.60274833e-01 4.32185262e-01 -6.04888856e-01 -6.38190746e-01
-6.87812686e-01 -4.15689766e-01 3.61165106e-01 7.80420601e-01
-6.10460974e-02 -7.57396042e-01 -1.20579422e+00 4.01631266e-01
2.49567211e-01 -6.73022151e-01 -7.23671690e-02 8.62219751e-01
-9.66349781e-01 -1.30869627e+00 -4.24188107e-01 -9.05634463e-01
4.68038350e-01 2.34441772e-01 9.47962761e-01 2.15864509e-01
-8.15358102e-01 4.72171396e-01 -4.44092244e-01 -5.38017511e-01
-5.14598429e-01 -3.55070412e-01 5.31329632e-01 -1.20979346e-01
-2.87913997e-02 -7.80221343e-01 -4.13127512e-01 3.11071932e-01
-7.98638344e-01 -7.63287663e-01 4.56671178e-01 1.28417587e+00
1.99340895e-01 1.20880616e+00 8.38650882e-01 -2.12109715e-01
5.45515954e-01 -3.87311310e-01 -7.27355063e-01 -3.77670795e-01
-6.92788482e-01 -1.13146476e-01 1.08690619e+00 -3.50241214e-01
-7.39055037e-01 -7.26646557e-03 -2.29534954e-01 -7.17346191e-01
-5.98377347e-01 7.70059705e-01 1.05663557e-02 4.85204123e-02
5.67232013e-01 9.44565143e-03 4.40771639e-01 -9.33977664e-01
4.74639274e-02 7.14157522e-01 8.66347134e-01 -1.46383122e-01
1.13001275e+00 2.32607290e-01 -3.67658064e-02 -1.31215882e+00
-8.48425552e-02 -4.93802965e-01 -7.93149352e-01 -4.32440549e-01
3.93016577e-01 -8.78501952e-01 -8.22814524e-01 6.18556082e-01
-9.43569601e-01 -2.09029421e-01 -6.10837638e-01 7.27811933e-01
-5.23311079e-01 6.26405597e-01 -8.82843018e-01 -8.72715652e-01
-4.62926567e-01 -8.57868612e-01 1.25423753e+00 -7.21154585e-02
-1.20652921e-01 -9.43897367e-01 -6.14310578e-02 -1.70522586e-01
4.26054657e-01 5.40644109e-01 8.70158553e-01 -4.37545359e-01
1.38901964e-01 -7.14748979e-01 5.83162308e-01 1.03798735e+00
4.77604777e-01 4.16932613e-01 -1.22462118e+00 -6.10768735e-01
6.20231807e-01 -1.45555854e-01 5.48798800e-01 1.60125002e-01
6.59296751e-01 5.91267347e-02 -8.40137601e-02 1.14075243e-01
1.39786375e+00 9.69823822e-02 1.60425454e-01 3.64660984e-03
4.65103954e-01 7.32647419e-01 7.28104532e-01 5.55807948e-01
-4.76119667e-01 4.71078187e-01 8.01426232e-01 -1.31410509e-01
1.54089019e-01 2.29367569e-01 8.46474826e-01 1.08770776e+00
-1.57065421e-01 5.15163243e-02 -5.21227002e-01 8.45921636e-01
-1.31995010e+00 -8.09598744e-01 -2.69693613e-01 2.19870734e+00
4.25635576e-01 1.39462888e-01 1.59800082e-01 1.14552355e+00
7.49031425e-01 -2.06002921e-01 -2.30050310e-01 -4.85271692e-01
2.85304755e-01 7.38455892e-01 1.26923006e-02 3.19103569e-01
-1.24321544e+00 -8.11062753e-03 6.56754160e+00 5.28448582e-01
-1.20149362e+00 -1.96482539e-01 -3.00758541e-01 -1.52778357e-01
3.77770334e-01 -3.68963391e-01 -1.85593188e-01 2.07657814e-01
1.25755954e+00 -1.08686343e-01 1.90942183e-01 9.52101827e-01
2.94617772e-01 5.26920473e-03 -1.08589506e+00 5.79979658e-01
-1.20314233e-01 -8.31884861e-01 -5.27896345e-01 7.11974427e-02
2.15309128e-01 -1.43048242e-01 -3.54514360e-01 7.40779936e-02
-3.88025910e-01 -8.26477051e-01 7.27849126e-01 4.36303198e-01
4.88128811e-01 -1.06981063e+00 1.22487795e+00 3.47996116e-01
-1.27435684e+00 -4.44103271e-01 -3.44355524e-01 -3.51600975e-01
1.05804943e-01 1.03952169e+00 -1.16183448e+00 9.94249284e-01
8.84125769e-01 6.41902566e-01 -1.49379060e-01 7.63188481e-01
-2.10362747e-01 8.88341308e-01 -2.81852305e-01 3.01721036e-01
-1.09188765e-01 6.89703822e-02 5.95537305e-01 9.05112445e-01
7.82750070e-01 -5.98096728e-01 1.36647254e-01 6.41992331e-01
4.86697584e-01 -2.48524830e-01 -7.14607894e-01 -2.05785766e-01
3.56417179e-01 1.41840494e+00 -5.18742025e-01 2.22273748e-02
-2.23513976e-01 7.53009081e-01 -2.22652793e-01 4.56957333e-02
-5.53737342e-01 -1.04290831e+00 8.65670383e-01 1.23948343e-02
8.49725962e-01 -2.69045562e-01 3.79876852e-01 -7.29840398e-01
3.12529862e-01 -6.42872214e-01 1.61493197e-01 -7.44485617e-01
-1.22750556e+00 6.49359703e-01 4.50468622e-02 -1.64948177e+00
-8.55861247e-01 -9.96842861e-01 -1.08467841e+00 9.56906259e-01
-1.41747224e+00 -6.19358718e-01 -2.93856859e-01 3.77394885e-01
5.50525188e-01 -6.19111173e-02 1.17521763e+00 1.58718392e-01
-7.54102886e-01 9.55089629e-02 1.75857157e-01 1.63743109e-01
3.76569718e-01 -1.78919709e+00 3.23505163e-01 1.05544353e+00
6.01637661e-02 2.76555926e-01 1.00549030e+00 -4.02988613e-01
-1.42111075e+00 -8.66172075e-01 4.40942079e-01 -8.39380398e-02
9.94102836e-01 -1.50297567e-01 -1.05767572e+00 2.77192831e-01
5.01818597e-01 1.12423953e-02 5.94607413e-01 -3.54445040e-01
2.68387914e-01 -1.37840763e-01 -9.10850048e-01 -5.46055771e-02
5.39254904e-01 -6.98553443e-01 -9.30264056e-01 3.07437181e-01
4.87317950e-01 -2.72975653e-01 -1.46211398e+00 3.31219077e-01
1.91250876e-01 -8.12312722e-01 7.78310180e-01 -2.56317049e-01
2.45804429e-01 -7.17533171e-01 4.85572845e-01 -1.91719699e+00
-4.50896502e-01 -9.32805300e-01 -6.31971419e-01 1.06812370e+00
-1.45676315e-01 -7.52718210e-01 2.56083727e-01 -3.35385770e-01
-6.14192009e-01 -4.84830618e-01 -1.06551349e+00 -9.99692976e-01
-9.04287547e-02 -2.89531678e-01 5.75301766e-01 7.92830110e-01
2.66494691e-01 1.45466626e-01 -1.10378191e-01 8.64159584e-01
1.90819830e-01 2.81985104e-01 3.73679936e-01 -1.54252124e+00
-6.01326466e-01 -1.84419945e-01 -8.29840004e-01 -2.55949110e-01
5.83318025e-02 -3.18665445e-01 4.54117090e-01 -1.00849652e+00
-9.41289604e-01 3.98570061e-01 -4.13369149e-01 4.62137014e-01
-4.22083028e-02 1.97184935e-01 -3.63779873e-01 -2.04890475e-01
5.34389913e-01 4.99247491e-01 7.15871096e-01 -1.64312497e-01
1.36258796e-01 1.18356943e-01 -2.12312937e-02 7.13143826e-01
8.61813724e-01 -2.78510630e-01 -5.94141543e-01 -3.06403399e-01
-6.21250384e-02 -1.16017694e-02 5.59827924e-01 -1.42967176e+00
-1.40720725e-01 4.53002006e-01 4.71338063e-01 -5.79655051e-01
3.40723284e-02 -1.03353572e+00 2.00776964e-01 8.31609726e-01
2.87569195e-01 3.09369922e-01 3.96981955e-01 6.91289783e-01
-5.33159494e-01 -5.92829049e-01 3.79851282e-01 1.93342477e-01
-7.02001810e-01 -3.23281348e-01 -9.78239238e-01 -7.11123407e-01
8.42317283e-01 -1.18038677e-01 -1.49916768e-01 -1.52304888e-01
-9.09860730e-01 -3.07761952e-02 2.77910382e-01 3.75310898e-01
6.25685453e-01 -1.00561059e+00 -6.87181354e-01 7.26991534e-01
-5.16393073e-02 -6.37169778e-02 4.45908308e-01 8.00270617e-01
-6.75108314e-01 1.00374475e-01 -2.46167555e-01 -6.57341063e-01
-1.11389446e+00 6.07275486e-01 6.12432122e-01 1.92369819e-01
-9.37748313e-01 6.22862220e-01 -4.18165028e-01 8.29278827e-02
-4.45426293e-02 -7.59106278e-01 -4.57364947e-01 2.96096146e-01
5.98910928e-01 5.06507993e-01 9.50986147e-01 -5.06540418e-01
-2.24300399e-01 5.07036805e-01 3.44321996e-01 2.10954130e-01
1.40948808e+00 -1.81872658e-02 3.87685224e-02 5.81945658e-01
9.73101199e-01 -9.58625600e-02 -1.32874846e+00 3.93196523e-01
-2.14097183e-02 -1.83084220e-01 5.31278908e-01 -5.46106935e-01
-1.29791975e+00 9.74741995e-01 8.07490051e-01 1.05904269e+00
1.50071907e+00 -5.99452853e-01 7.72765517e-01 4.40495759e-01
3.99314851e-01 -1.25830376e+00 -5.13603259e-03 1.89515203e-01
9.38352406e-01 -6.78244770e-01 -2.21101999e-01 -1.44182116e-01
-1.05724253e-01 1.81465256e+00 2.63458192e-01 -4.78241116e-01
8.23605597e-01 5.62115371e-01 2.26291344e-01 -3.54845226e-01
-8.47711325e-01 -4.29678649e-01 2.21981034e-01 1.03593194e+00
3.57484400e-01 -6.17016926e-02 4.87759300e-02 5.51077366e-01
-2.61245400e-01 -2.98019290e-01 7.96648324e-01 1.25816977e+00
-5.77686369e-01 -1.03431141e+00 -6.35214925e-01 4.37714785e-01
-4.83203799e-01 3.41834575e-01 -1.24484748e-01 8.30810845e-01
2.73969531e-01 1.24963379e+00 2.01141909e-01 -7.58831561e-01
6.13207877e-01 4.45191026e-01 3.51898670e-01 -6.00279212e-01
-6.14318132e-01 3.17637742e-01 1.73648328e-01 -5.27516901e-01
-2.12186486e-01 -6.60118580e-01 -1.16781437e+00 2.43043408e-01
-1.03205264e+00 4.02998090e-01 7.79320121e-01 9.32509422e-01
4.25627857e-01 1.20050240e+00 1.14729774e+00 -1.24764407e+00
-9.77086663e-01 -1.24035442e+00 -1.16697133e+00 2.30557740e-01
8.71628940e-01 -1.02462626e+00 -7.77882338e-01 2.50355899e-01] | [6.731602191925049, 2.359441041946411] |
0d0838ac-d870-4856-938d-0b1a9423f79d | exploring-difference-in-public-perceptions-on | 1907.03167 | null | https://arxiv.org/abs/1907.03167v1 | https://arxiv.org/pdf/1907.03167v1.pdf | Exploring difference in public perceptions on HPV vaccine between gender groups from Twitter using deep learning | In this study, we proposed a convolutional neural network model for gender prediction using English Twitter text as input. Ensemble of proposed model achieved an accuracy at 0.8237 on gender prediction and compared favorably with the state-of-the-art performance in a recent author profiling task. We further leveraged the trained models to predict the gender labels from an HPV vaccine related corpus and identified gender difference in public perceptions regarding HPV vaccine. The findings are largely consistent with previous survey-based studies. | ['Qiang Wei', 'Cui Tao', 'Jingcheng Du', 'Yong Chen', 'Chongliang Luo'] | 2019-07-06 | null | null | null | null | ['gender-prediction'] | ['computer-vision'] | [-2.76511461e-01 6.50575876e-01 -6.07882738e-01 -7.95397043e-01
-2.24316061e-01 -4.83491510e-01 8.96615446e-01 8.98712337e-01
-7.61522651e-01 8.56234312e-01 4.10732716e-01 -7.24548936e-01
2.46711731e-01 -1.06953859e+00 -6.22864246e-01 -3.77425075e-01
5.74926473e-03 6.82959378e-01 -6.47836626e-01 -1.65824592e-01
7.94553399e-01 2.87765618e-02 -1.42455876e+00 1.09679095e-01
6.72969997e-01 7.76023149e-01 -4.98153508e-01 8.84179473e-01
-3.30351144e-01 6.91254556e-01 -4.83838499e-01 -7.83845842e-01
-2.67737895e-01 1.26739755e-01 -7.65670717e-01 -6.99571550e-01
5.84226429e-01 -3.31713945e-01 -2.98564404e-01 7.30241299e-01
5.30608952e-01 -4.31101084e-01 1.08867311e+00 -1.24213612e+00
-8.61542523e-01 7.16530740e-01 -9.11208212e-01 3.21698576e-01
3.90028566e-01 -7.22687244e-01 6.75457478e-01 -6.01200342e-01
8.75214696e-01 1.35789275e+00 1.16627669e+00 7.14471281e-01
-5.59581578e-01 -1.32259786e+00 -1.55737475e-01 -2.33771384e-01
-1.13001895e+00 -1.97250754e-01 4.57366496e-01 -7.29301274e-01
8.12961757e-01 1.93022206e-01 3.71969461e-01 1.36094058e+00
6.22627497e-01 5.48937023e-01 1.56122863e+00 -1.81855530e-01
-4.50054854e-01 5.05548418e-01 4.41540033e-01 1.03330743e+00
3.71200114e-01 -1.29289851e-01 -5.65974295e-01 -3.42427969e-01
4.45842631e-02 -1.94005311e-01 8.18300784e-01 4.49577630e-01
-7.47175932e-01 1.31504393e+00 2.17410624e-01 3.80638354e-02
-2.35898927e-01 -1.97112024e-01 1.07276201e+00 -2.94919442e-02
1.06157672e+00 1.91364780e-01 -6.68379068e-01 -1.15783498e-01
-9.74815428e-01 6.15340590e-01 9.64447200e-01 3.59064013e-01
3.21679115e-01 -1.50036111e-01 -4.26834017e-01 5.56395471e-01
4.98289794e-01 6.56497061e-01 3.45610261e-01 -4.58111644e-01
3.74919116e-01 6.41219616e-01 5.13239279e-02 -1.22349310e+00
-7.61169910e-01 -5.37380874e-01 -8.09241056e-01 -3.93569916e-01
5.21302938e-01 -4.37440962e-01 -1.09008646e+00 1.63046360e+00
1.07594833e-01 -2.56855845e-01 -1.35165989e-01 2.29086310e-01
1.71825802e+00 3.50533754e-01 7.71160364e-01 9.18173045e-02
1.73911083e+00 -5.32024205e-01 -8.84470046e-01 2.31791157e-02
5.11153281e-01 -8.43257844e-01 2.99045205e-01 1.20675318e-01
-6.76622927e-01 -3.28182489e-01 -8.61531138e-01 1.17124058e-01
-1.06296122e+00 2.48147264e-01 1.14562058e+00 1.42291880e+00
-1.01480031e+00 4.28377301e-01 -5.82832634e-01 -1.02735746e+00
6.23935819e-01 8.08431149e-01 -2.56028712e-01 3.63402247e-01
-1.13876712e+00 6.32017791e-01 1.89569995e-01 -4.94220061e-03
-5.19341707e-01 -6.72347546e-01 -9.21621621e-01 -5.24298847e-01
-5.22690535e-01 -7.06820607e-01 1.42708969e+00 -5.06158531e-01
-1.01751733e+00 1.54143620e+00 -6.43919230e-01 -3.44470114e-01
4.30318296e-01 8.79803672e-02 -5.53754449e-01 -4.33288872e-01
3.36162299e-01 1.03651643e+00 1.59269199e-01 -8.43716860e-01
-8.21751714e-01 -7.10809946e-01 -3.60543817e-01 -7.46816248e-02
-4.93067086e-01 4.56936181e-01 8.95176679e-02 -2.28290170e-01
-6.17000312e-02 -9.09111917e-01 1.32112624e-02 -1.00145161e+00
-4.26231772e-01 -8.13823700e-01 7.25584865e-01 -1.16959810e+00
1.13569248e+00 -1.36176240e+00 -4.87652481e-01 2.24105343e-01
5.64376533e-01 -1.30317673e-01 3.60852838e-01 6.28039956e-01
9.52403471e-02 3.58876109e-01 6.48240805e-01 -5.78273475e-01
1.25804305e-01 -2.80089099e-02 -3.07084415e-02 6.41870737e-01
-2.30788887e-01 8.28279197e-01 -7.26030052e-01 -7.50017524e-01
-1.32164836e-01 4.19214159e-01 -4.26862478e-01 2.09291577e-01
3.80026996e-02 5.94528913e-01 -4.68910515e-01 1.08205676e+00
1.19667912e+00 1.48806959e-01 3.62047613e-01 -2.33685270e-01
-2.78721839e-01 -1.39637798e-01 -3.62191349e-01 9.19626594e-01
-1.99319184e-01 8.71470809e-01 -1.49488360e-01 -7.70351171e-01
1.28408360e+00 6.57766312e-02 4.34220344e-01 -6.01540446e-01
9.24255550e-01 4.09122616e-01 1.61372527e-01 -8.53948653e-01
8.42303753e-01 2.32406557e-01 -6.42232656e-01 2.64332384e-01
2.82041192e-01 5.87264955e-01 1.90086424e-01 -3.20403054e-02
3.88660133e-01 9.98517200e-02 1.30492970e-01 -6.31231129e-01
6.97048485e-01 -1.49701580e-01 3.81392449e-01 8.71325731e-01
-7.23038554e-01 1.41989037e-01 9.07741547e-01 -8.73701334e-01
-9.47694063e-01 -6.06164992e-01 -3.29044431e-01 1.87230682e+00
-4.99976426e-01 -1.89399377e-01 -1.14796627e+00 -7.89860904e-01
2.47386977e-01 2.06916794e-01 -1.62637663e+00 6.49815083e-01
-5.20589054e-01 -1.03992796e+00 1.18332219e+00 5.64693391e-01
6.28062963e-01 -7.58768559e-01 -2.20129624e-01 -1.01745531e-01
8.74507725e-02 -1.02193761e+00 -1.38726503e-01 1.24520719e-01
-5.82412720e-01 -1.26636410e+00 -7.19202936e-01 -1.13191032e+00
6.00591719e-01 -6.81889057e-01 1.28409505e+00 -1.27366588e-01
1.40736625e-02 -1.99137121e-01 1.44951835e-01 -1.20742953e+00
-3.10873002e-01 9.29923654e-01 -1.01958681e-02 -4.52668071e-01
1.03405929e+00 -2.42377326e-01 -8.58489633e-01 -3.17002028e-01
-2.95686632e-01 -7.85496980e-02 2.64592618e-01 6.87492371e-01
-1.68261141e-01 -4.15586472e-01 9.87685382e-01 -1.53430903e+00
9.54559147e-01 -9.95147526e-01 -1.12972543e-01 -4.08227772e-01
-1.02189493e+00 -1.89464152e-01 2.52478302e-01 -1.15428030e-01
-1.01878583e+00 -3.47865939e-01 -5.54346323e-01 5.46609879e-01
-1.32534862e-01 7.25649118e-01 5.58461249e-01 1.05324611e-01
2.41435349e-01 -7.60658681e-02 5.86139113e-02 -3.96964639e-01
-2.11286470e-01 1.27408612e+00 4.37513709e-01 -5.55317402e-01
1.28726736e-01 -1.52792688e-03 9.24126059e-02 -3.12833667e-01
-8.74325693e-01 -2.88537413e-01 -4.42500412e-01 -2.66530395e-01
1.11336684e+00 -9.42613840e-01 -1.37769437e+00 9.65514600e-01
-8.31379831e-01 2.25866854e-01 1.18914366e+00 -4.59032543e-02
3.33079812e-03 -3.52467567e-01 -9.95531261e-01 -1.11639893e+00
-9.29261148e-01 -6.94405973e-01 1.09697068e+00 7.26646662e-01
-6.79830492e-01 -1.39903903e+00 2.81714141e-01 5.46125829e-01
7.39810348e-01 7.94718385e-01 9.62795615e-01 -1.22590530e+00
5.23776233e-01 -3.59067500e-01 -6.03467584e-01 -6.62916243e-01
-7.62365162e-02 2.05778077e-01 -1.22186971e+00 -2.05103144e-01
-6.81109786e-01 -3.88601601e-01 9.94326413e-01 5.62942505e-01
1.54248226e+00 -3.25336009e-01 -7.25480497e-01 4.23198164e-01
1.39189458e+00 -1.49940243e-02 8.70162994e-02 4.71773833e-01
6.79662406e-01 6.06331229e-01 4.35213238e-01 6.38973176e-01
1.31366038e+00 3.46226469e-02 1.66751206e-01 -2.56872296e-01
4.49862391e-01 -5.31241417e-01 -5.76960385e-01 5.24603724e-01
-4.27000821e-01 2.56901258e-03 -1.52906215e+00 5.11526525e-01
-1.69497025e+00 -7.17775702e-01 -8.68295506e-02 1.39062881e+00
7.53955126e-01 1.91725165e-01 6.34982407e-01 -2.85740942e-01
6.50113344e-01 2.98966944e-01 -3.34944695e-01 -1.44434190e+00
1.63946301e-01 4.74898428e-01 1.04612935e+00 2.66852677e-01
-1.74395704e+00 8.54951262e-01 7.47932005e+00 4.58814442e-01
-1.25351143e+00 2.29455769e-01 1.31682324e+00 3.72847140e-01
9.96555537e-02 -8.10699463e-01 -9.50745523e-01 2.78698206e-01
1.40198243e+00 -2.85893530e-01 1.61965899e-02 5.25125861e-01
-1.05543844e-01 -4.81936075e-02 -8.31346333e-01 8.75629127e-01
1.90418497e-01 -1.31328475e+00 -5.33061206e-01 3.93974543e-01
9.10427272e-01 2.27623060e-01 4.65100110e-01 7.38884568e-01
1.94330677e-01 -1.78572774e+00 1.83338314e-01 3.99392992e-01
7.86100626e-01 -1.29707015e+00 1.48182225e+00 2.52008319e-01
-8.32725525e-01 -3.83956134e-01 9.96706039e-02 -4.43498909e-01
-5.34258604e-01 -5.89643754e-02 -1.22766626e+00 2.18673512e-01
8.59818220e-01 5.25260866e-01 -9.43832874e-01 4.73760486e-01
4.49322343e-01 6.84243798e-01 1.37819737e-01 -4.48238671e-01
3.12879533e-01 1.86781585e-01 -2.49167264e-01 1.70902371e+00
2.40314975e-01 -3.72908056e-01 -1.63554728e-01 3.13363522e-01
-1.28202736e-01 5.68549335e-01 -7.98094928e-01 -4.93045032e-01
5.86685956e-01 1.52990758e+00 -1.07933795e+00 -3.03153008e-01
-3.79230261e-01 1.75088689e-01 7.36023486e-02 1.87595829e-01
-7.81054378e-01 -3.31915677e-01 3.75452280e-01 6.38257787e-02
-9.76208895e-02 2.08729669e-01 -9.48095381e-01 -3.69109511e-01
-7.79026389e-01 -8.10191810e-01 7.13238478e-01 -1.37739122e-01
-1.14858913e+00 3.87735009e-01 7.81485066e-03 -6.23097539e-01
-4.38671559e-01 -7.57842124e-01 -7.26361513e-01 9.98933852e-01
-9.85856473e-01 -2.11969280e+00 -2.68969119e-01 -2.94313058e-02
2.55531877e-01 -7.37049580e-01 1.32706618e+00 4.93414909e-01
-6.93103194e-01 1.20200241e+00 2.48940811e-02 4.59107399e-01
8.35327446e-01 -1.25516915e+00 2.81625092e-01 -3.49452347e-02
-8.98652434e-01 9.64453220e-01 1.05779421e+00 -9.11467791e-01
-1.35460782e+00 -9.61896539e-01 1.50703061e+00 -6.45999730e-01
5.48355997e-01 -5.95873713e-01 2.35100021e-03 8.32500279e-01
9.82600033e-01 -5.97721696e-01 1.12200236e+00 6.08852983e-01
-4.91915584e-01 1.23643935e-01 -1.76429260e+00 2.24937707e-01
5.95995843e-01 -4.09857631e-01 3.59425554e-03 2.08616182e-01
2.33788416e-01 -8.90497386e-01 -1.07589710e+00 4.96907324e-01
1.35815585e+00 -7.77633846e-01 6.96322620e-01 -8.53587925e-01
1.07179427e+00 5.87470114e-01 -1.66322306e-01 -8.12648058e-01
-8.66829753e-02 6.29618987e-02 -1.98628962e-01 1.48061049e+00
3.88719559e-01 -6.31444752e-01 1.25609517e+00 5.69022536e-01
3.13052118e-01 -1.23237169e+00 -5.20643115e-01 1.60486102e-01
7.71445692e-01 1.94178969e-01 8.28981340e-01 1.13377988e+00
-1.48654521e-01 2.08224565e-01 -2.11627230e-01 -3.17186058e-01
7.38039255e-01 -1.60528451e-01 6.55552506e-01 -1.43699610e+00
4.13172007e-01 -6.32055640e-01 -2.97226459e-01 1.94937527e-01
8.41285229e-01 -7.70920396e-01 -3.96089852e-01 -1.40497828e+00
1.04943609e+00 -1.26929328e-01 -4.35457230e-01 2.04554200e-01
-1.30131200e-01 6.97998524e-01 -3.50523382e-01 -2.65118092e-01
-3.34414124e-01 -1.56422332e-01 8.91737759e-01 -3.30921710e-01
1.96993485e-01 2.45427743e-01 -1.40679979e+00 8.08932662e-01
1.08088923e+00 -4.11210507e-01 -9.21548903e-02 -2.94500906e-02
6.10116899e-01 -7.45108351e-02 -1.84444577e-01 -3.65283370e-01
-5.44725209e-02 -1.39815003e-01 1.12582731e+00 -1.20574117e+00
-3.36658061e-02 -1.51491945e-03 -1.02678277e-01 6.22939527e-01
-4.34857190e-01 3.95525277e-01 3.88652891e-01 2.33801797e-01
6.68683276e-02 1.03552513e-01 1.62165940e-01 6.44314438e-02
-2.17596129e-01 2.42326856e-01 -5.96028626e-01 -2.64465749e-01
5.06577015e-01 -3.22571397e-01 -7.57750630e-01 -2.90424794e-01
-4.51027364e-01 1.83675423e-01 2.93495476e-01 7.05263853e-01
5.78083806e-02 -1.29879689e+00 -8.56059849e-01 1.83598533e-01
8.23253468e-02 -8.91464710e-01 1.00756034e-01 7.14093447e-01
-6.09541535e-01 9.50047016e-01 -5.70371866e-01 -5.09844363e-01
-1.46320498e+00 2.08035987e-02 2.97134280e-01 -3.33591551e-01
2.85436541e-01 8.89798760e-01 -2.39293709e-01 -1.24920452e+00
4.07107919e-01 -4.62114103e-02 -1.10008419e+00 5.80253303e-01
9.99718606e-02 5.41465521e-01 9.19073969e-02 -7.45170116e-01
-7.49957740e-01 3.18111449e-01 -2.04153791e-01 -3.32026780e-02
1.51032805e+00 6.10294044e-02 -4.29009616e-01 3.78619492e-01
1.33615506e+00 2.40775391e-01 -3.31471413e-01 3.19545895e-01
-1.50481928e-02 -2.51536608e-01 -1.72577624e-03 -9.57706392e-01
-1.02607453e+00 3.73142093e-01 1.12227213e+00 1.67772144e-01
6.48207545e-01 -2.27936178e-01 7.51869023e-01 1.72479957e-01
-9.83358175e-02 -1.30820036e+00 -2.58854657e-01 8.23191524e-01
2.50193477e-01 -1.75694859e+00 5.41689023e-02 -3.01552657e-02
-3.13748837e-01 1.12580013e+00 8.76710653e-01 -1.44194722e-01
1.02885664e+00 1.83534205e-01 3.75477016e-01 -4.33356941e-01
-4.83399421e-01 3.45817953e-01 3.86050940e-01 7.21113026e-01
1.52926803e+00 6.38437092e-01 -8.41248930e-01 1.02074337e+00
-8.38196814e-01 2.95854330e-01 4.28612590e-01 7.16131568e-01
-1.76131472e-01 -1.07198703e+00 -3.21752936e-01 9.32409346e-01
-1.23900652e+00 -1.09804864e-03 -5.93498766e-01 7.53365338e-01
4.66507763e-01 9.49336529e-01 3.74399871e-01 -6.22752249e-01
-1.90047249e-01 2.34994143e-01 6.76397681e-01 -1.65781438e-01
-9.48885322e-01 -3.12477469e-01 7.74967968e-01 -2.80319676e-02
-5.11454761e-01 -7.58544922e-01 -8.18883002e-01 -8.82678926e-01
3.48366857e-01 -6.78136274e-02 1.12579215e+00 8.46374214e-01
1.24132805e-01 4.62628186e-01 5.34022033e-01 -5.42479455e-01
-3.12834904e-02 -1.45706272e+00 -5.23137212e-01 4.00742799e-01
3.13100219e-01 -5.38373530e-01 1.08103782e-01 -2.95763046e-01] | [9.449271202087402, 10.317344665527344] |
b64618c5-13a4-4f49-b4e5-156329b2ed0c | an-fea-surrogate-model-with-boundary-oriented | 2108.13509 | null | https://arxiv.org/abs/2108.13509v1 | https://arxiv.org/pdf/2108.13509v1.pdf | An FEA surrogate model with Boundary Oriented Graph Embedding approach | In this work, we present a Boundary Oriented Graph Embedding (BOGE) approach for the Graph Neural Network (GNN) to serve as a general surrogate model for regressing physical fields and solving boundary value problems. Providing shortcuts for both boundary elements and local neighbor elements, the BOGE approach can embed structured mesh elements into the graph and performs an efficient regression on large-scale triangular-mesh-based FEA results, which cannot be realized by other machine-learning-based surrogate methods. Focusing on the cantilever beam problem, our BOGE approach cannot only fit the distribution of stress fields but also regresses the topological optimization results, which show its potential of realizing abstract decision-making design process. The BOGE approach with 3-layer DeepGCN model \textcolor{blue}{achieves the regression with MSE of 0.011706 (2.41\% MAPE) for stress field prediction and 0.002735 MSE (with 1.58\% elements having error larger than 0.01) for topological optimization.} The overall concept of the BOGE approach paves the way for a general and efficient deep-learning-based FEA simulator that will benefit both industry and design-related areas. | ['Vaneet Aggarwal', 'Martin Byung-Guk Jun', 'Zhengyang Kang', 'Dheeraj Peddireddy', 'Fengfeng Zhou', 'Xingyu Fu'] | 2021-08-30 | null | null | null | null | ['cantilever-beam'] | ['miscellaneous'] | [-2.78189480e-01 4.34221298e-01 7.01195225e-02 -5.74854650e-02
-4.11598861e-01 1.39267206e-01 2.57478338e-02 -1.59134567e-01
1.06331319e-01 9.24567819e-01 -3.53863537e-01 -4.02634919e-01
-3.10999781e-01 -1.27519965e+00 -9.27290499e-01 -8.16408455e-01
-2.90224910e-01 5.53757906e-01 1.56183228e-01 -6.09933257e-01
5.33192232e-02 6.89829469e-01 -1.13917267e+00 2.03823280e-02
6.59137309e-01 1.31294262e+00 -1.80961668e-01 2.79982716e-01
1.83490813e-01 4.67544734e-01 -3.11487883e-01 -1.79287776e-01
3.27632353e-02 -2.46386603e-01 -5.95371008e-01 -5.94718158e-01
5.01984477e-01 -1.39562264e-01 -4.44102526e-01 8.47266793e-01
6.14493906e-01 2.13144720e-01 1.07975852e+00 -9.87490892e-01
-8.94063354e-01 4.38376695e-01 -6.19952738e-01 -1.90874174e-01
-1.48386747e-01 2.20315814e-01 9.39786375e-01 -1.12783992e+00
5.78369737e-01 1.01585412e+00 1.14407933e+00 7.04962730e-01
-1.39271009e+00 -6.11857772e-01 -1.28076941e-01 -1.94112156e-02
-1.40207231e+00 8.90009627e-02 1.33533347e+00 -6.34707808e-01
9.02251303e-01 4.64550674e-01 7.08597541e-01 9.08247411e-01
1.03182006e+00 1.60977483e-01 1.04263604e+00 -1.42394826e-01
2.46562898e-01 -2.74812847e-01 1.26297519e-01 9.45402682e-01
3.76468271e-01 3.37721616e-01 -2.00905919e-01 5.31763136e-02
1.10728741e+00 -2.78779119e-01 -1.94422513e-01 -2.56214648e-01
-7.71046460e-01 7.44071662e-01 9.46049333e-01 2.25631133e-01
-3.37950557e-01 9.38741088e-01 2.15283722e-01 2.22063497e-01
6.80995107e-01 6.54317737e-01 -1.03142805e-01 3.97125989e-01
-1.14722705e+00 4.52264488e-01 5.79756320e-01 3.47270012e-01
7.24307120e-01 9.64383543e-01 -1.57924131e-01 7.43442595e-01
6.55807376e-01 8.13705564e-01 -2.23444000e-01 -9.18870866e-01
1.24910489e-01 7.15886712e-01 -2.59388164e-02 -1.67479813e+00
-8.64222467e-01 -5.82324386e-01 -1.28680420e+00 8.48112941e-01
2.98664540e-01 -3.50739986e-01 -9.10853744e-01 1.58910275e+00
1.87025532e-01 3.86472903e-02 -4.16200876e-01 9.93290246e-01
1.02716100e+00 7.90302336e-01 -1.83207303e-01 1.60730332e-01
8.05434108e-01 -7.14414299e-01 -5.96722424e-01 3.44537608e-02
7.19903171e-01 -3.22312981e-01 1.19434011e+00 3.06114137e-01
-1.16633451e+00 -6.08553708e-01 -1.47175288e+00 1.31753802e-01
-3.72279167e-01 7.58750886e-02 5.18154562e-01 4.55405354e-01
-1.36923599e+00 1.03293860e+00 -8.58075082e-01 2.66731083e-01
4.61287260e-01 5.95867932e-01 4.34499606e-03 9.58896428e-02
-1.17568851e+00 8.29861581e-01 -1.18642271e-01 7.01830447e-01
-9.08904374e-01 -9.94590402e-01 -6.89005375e-01 1.20394744e-01
1.41214669e-01 -7.51696765e-01 4.27152455e-01 -2.86995679e-01
-1.56872404e+00 6.74758792e-01 4.42107946e-01 -3.60982627e-01
4.71398383e-01 4.91988771e-02 -3.31376225e-01 -1.98255077e-01
-1.77823648e-01 5.12637675e-01 6.90882802e-01 -1.42884123e+00
4.20775503e-01 -2.16009736e-01 -6.49985522e-02 -3.94501835e-01
-2.74705201e-01 -6.07543290e-01 1.20529443e-01 -9.62747395e-01
2.16105521e-01 -9.15974200e-01 -5.09095728e-01 2.10424796e-01
-4.58506465e-01 -6.11983165e-02 9.73723292e-01 -9.36853468e-01
1.39262748e+00 -1.71541631e+00 1.82930604e-01 5.42904437e-01
5.88053942e-01 1.19499207e-01 1.06407098e-01 6.72952950e-01
-1.53187573e-01 2.39238128e-01 -5.41796088e-01 5.63642718e-02
1.06780171e-01 -1.26737520e-01 -2.96351127e-02 6.58459604e-01
1.66436464e-01 1.36488163e+00 -4.86469299e-01 -9.53630805e-02
2.72106320e-01 6.07602358e-01 -8.82986426e-01 -1.95974424e-01
-2.78028518e-01 4.73924190e-01 -5.52960277e-01 5.06995499e-01
8.73921990e-01 -2.38508627e-01 1.12186022e-01 -5.74124753e-01
-2.01321170e-02 -3.30059558e-01 -1.16988063e+00 1.51520753e+00
-6.56297386e-01 7.30171978e-01 4.63971376e-01 -1.25099397e+00
1.40283012e+00 4.51552086e-02 7.80861080e-01 -9.30773616e-01
3.57539564e-01 3.04128379e-01 -7.45128319e-02 -2.41467401e-01
7.27707744e-02 -3.45198750e-01 -3.66371095e-01 2.73127586e-01
-8.07425827e-02 -4.22257394e-01 -2.73984820e-01 -2.07999021e-01
1.03514910e+00 1.52253658e-01 -6.25823021e-01 -1.05476940e+00
5.64579844e-01 -5.97217642e-02 3.49479496e-01 4.63815600e-01
1.81654483e-01 4.93865699e-01 5.90494752e-01 -8.82015824e-01
-1.24564791e+00 -8.78677487e-01 -2.62836754e-01 4.33876634e-01
2.57084608e-01 -1.79173917e-01 -8.31303954e-01 -2.69197792e-01
2.96004951e-01 5.71182728e-01 -7.40868270e-01 -5.51063240e-01
-8.56564760e-01 -7.69948006e-01 4.02209997e-01 7.30475962e-01
3.85293901e-01 -1.04545462e+00 -2.33377054e-01 2.88500011e-01
3.23471874e-01 -5.65887272e-01 -1.86080970e-02 7.90498629e-02
-8.50515068e-01 -7.34981298e-01 -5.95766723e-01 -7.98317850e-01
4.80256200e-01 -6.97170079e-01 1.14343166e+00 5.19275010e-01
-2.83832043e-01 1.97249860e-01 5.08981645e-02 4.11236510e-02
-4.59135503e-01 -4.70407605e-02 7.90314451e-02 -1.57172397e-01
-4.13437456e-01 -8.62850487e-01 -7.41705298e-01 4.22223747e-01
-5.90488851e-01 8.38399008e-02 1.69074327e-01 1.02601218e+00
7.73060679e-01 9.25669372e-02 8.48164976e-01 -5.90494096e-01
7.58706510e-01 -3.26986521e-01 -7.37263560e-01 1.12449467e-01
-8.96213114e-01 -4.95070266e-03 9.44546402e-01 -5.77700250e-02
-7.13458598e-01 -4.98650879e-01 -4.40292299e-01 -6.62362754e-01
2.97668606e-01 6.56609237e-01 -1.43502206e-01 -5.92144072e-01
5.55966437e-01 -2.45326355e-01 7.35402061e-03 -5.06400168e-01
1.67351410e-01 2.16508061e-01 3.64981443e-01 -8.00491333e-01
7.19246387e-01 2.93781430e-01 7.63408184e-01 -8.29272032e-01
-2.00990960e-01 3.04993510e-01 -3.24926436e-01 -6.92455530e-01
8.47054958e-01 -3.90912712e-01 -1.06175590e+00 5.84365606e-01
-8.93927872e-01 -7.87471890e-01 -3.33474070e-01 2.18233675e-01
-5.77851772e-01 1.79010957e-01 -7.72131205e-01 -6.32916391e-01
-5.28869450e-01 -1.15362620e+00 8.97063375e-01 -9.25955642e-03
-1.96828783e-01 -1.28280151e+00 -1.45381272e-01 3.54279727e-01
6.07400835e-01 9.25562561e-01 1.28898335e+00 -3.17247622e-02
-5.00132680e-01 -3.18754077e-01 -1.63598508e-01 4.94437933e-01
-3.24556530e-01 8.65248591e-02 -8.32737088e-01 -5.27625442e-01
1.73411280e-01 -1.66320741e-01 7.22615957e-01 7.49138057e-01
1.30827975e+00 -1.67676300e-01 -3.45037609e-01 8.04891229e-01
1.79001200e+00 1.15714267e-01 7.16854632e-01 -3.27463411e-02
1.13336778e+00 2.65998632e-01 2.60639042e-01 3.25081766e-01
-1.06202766e-01 6.87332988e-01 7.49177575e-01 -5.05337298e-01
-3.52925956e-01 -1.55404359e-01 5.63519411e-02 1.02097905e+00
-2.47838631e-01 -3.14905584e-01 -1.31426156e+00 3.14917088e-01
-1.65485990e+00 -4.61482435e-01 -6.49724245e-01 1.83460414e+00
4.27982569e-01 3.81909013e-01 -4.01789874e-01 1.58808231e-01
6.17638528e-01 2.43554011e-01 -6.51877642e-01 -8.72367322e-01
-4.48822863e-02 7.19088614e-01 5.34314215e-01 5.88595331e-01
-8.55313063e-01 8.44303310e-01 6.13833809e+00 9.84789133e-01
-1.54874837e+00 4.22927365e-03 8.35422873e-01 2.57939667e-01
-6.13529682e-01 -2.49342576e-01 -4.69507128e-01 5.32333016e-01
1.07226193e+00 9.24824029e-02 3.85913491e-01 6.57505095e-01
3.09712112e-01 4.20520790e-02 -9.92110491e-01 8.46258283e-01
-3.19717467e-01 -1.87576509e+00 4.00955277e-03 1.71411216e-01
8.39246571e-01 -1.84897363e-01 2.11954802e-01 2.89804697e-01
7.99556747e-02 -1.43485820e+00 8.38657677e-01 7.60678530e-01
1.15415275e+00 -6.11723006e-01 7.14816749e-01 1.33744314e-01
-1.16744161e+00 4.55898531e-02 -3.25625867e-01 -1.99010894e-01
3.42164308e-01 8.00253510e-01 -5.41609228e-01 6.22019291e-01
9.58294511e-01 5.46367943e-01 -1.86275601e-01 5.88353217e-01
1.61298305e-01 8.06509733e-01 -4.18658793e-01 -2.14729249e-01
2.48342499e-01 -6.85819030e-01 7.03981578e-01 8.48069191e-01
1.80077434e-01 -6.03233464e-03 4.95553203e-02 1.59248042e+00
-1.79606661e-01 6.12902157e-02 -5.28312624e-01 2.82058045e-02
1.38717994e-01 9.59510624e-01 -9.62367058e-01 1.28532588e-01
-1.68388747e-02 3.57701838e-01 1.14951536e-01 5.27478814e-01
-1.23581350e+00 -4.88481343e-01 4.23126936e-01 7.87995338e-01
3.94878387e-01 -3.96811873e-01 -8.21501493e-01 -6.15273654e-01
1.47044197e-01 -4.57988590e-01 -1.74310878e-01 -8.76396775e-01
-1.34744442e+00 5.16578913e-01 -9.80627388e-02 -8.67703140e-01
1.60151675e-01 -1.05521464e+00 -8.20501864e-01 9.08446312e-01
-1.10990846e+00 -1.17125094e+00 -1.83364585e-01 2.05813870e-01
-8.82746205e-02 1.63968522e-02 6.47774518e-01 4.86477673e-01
-6.79731309e-01 6.52722776e-01 3.23437870e-01 1.62733674e-01
1.07114740e-01 -9.71582770e-01 4.84967411e-01 4.58643615e-01
-3.35660368e-01 4.52789009e-01 7.79070318e-01 -8.27550590e-01
-1.62971687e+00 -8.76631498e-01 3.17081898e-01 -1.61640748e-01
7.02262342e-01 -4.44314718e-01 -1.05978656e+00 2.51526326e-01
9.97392694e-04 2.25512475e-01 1.37389407e-01 -6.55540153e-02
1.84439600e-01 -3.70674074e-01 -1.09349430e+00 6.41489029e-01
9.93373632e-01 -3.90621632e-01 -2.12634370e-01 1.81272000e-01
5.61435580e-01 -4.12353337e-01 -1.29451120e+00 1.06219923e+00
3.99435848e-01 -8.64440441e-01 1.09683824e+00 -5.74396372e-01
6.48565888e-01 -6.79638237e-02 -3.68807130e-02 -1.00779724e+00
-4.07751322e-01 -4.54378426e-01 -1.67341337e-01 9.55495477e-01
5.02696276e-01 -6.86970234e-01 1.13627446e+00 6.15208149e-01
-6.55519664e-01 -1.59360743e+00 -1.05883384e+00 -8.48245859e-01
8.08753073e-01 -4.67306256e-01 4.97328699e-01 8.81792843e-01
-4.40175831e-01 3.22844684e-02 -2.79825926e-01 -2.63498891e-02
5.10546684e-01 9.34174955e-02 3.57512563e-01 -1.31445551e+00
2.47396268e-02 -6.43629134e-01 -4.48333204e-01 -7.40704536e-01
1.86709374e-01 -9.68885779e-01 -1.65579915e-01 -1.72866023e+00
-5.01173913e-01 -8.77395153e-01 -4.62344408e-01 3.65620971e-01
4.24986631e-01 2.70595431e-01 -1.74925521e-01 -2.37649500e-01
6.17973581e-02 8.78050625e-01 1.63947523e+00 -4.16620910e-01
9.71264020e-02 -3.55505168e-01 -2.45884404e-01 5.87756455e-01
7.34859049e-01 -4.70587432e-01 -2.45297864e-01 -3.53870571e-01
5.74450433e-01 3.13778907e-01 8.47926378e-01 -1.36090279e+00
1.47631513e-02 -1.14588395e-01 3.77499759e-01 -6.31342769e-01
2.82155514e-01 -8.18358898e-01 5.54731905e-01 5.12515426e-01
6.27969727e-02 4.84059704e-03 4.52516764e-01 4.66270655e-01
-2.17914078e-02 -3.94641347e-02 9.03170168e-01 7.20867589e-02
-4.43859667e-01 3.00394535e-01 -3.58817935e-01 8.57229307e-02
7.94847250e-01 -5.42142868e-01 -2.83538867e-02 -1.67838037e-01
-9.22524929e-01 6.60854951e-02 6.10006094e-01 3.43196802e-02
7.21835792e-01 -1.50952268e+00 -5.86643159e-01 2.97597826e-01
-5.09245753e-01 1.31932855e-01 4.78511095e-01 8.04807842e-01
-1.07123756e+00 1.22049056e-01 -4.40495223e-01 -5.82411587e-01
-4.60605890e-01 2.69557416e-01 6.88778937e-01 -3.09387952e-01
-5.69719613e-01 1.18465471e+00 2.19649728e-02 -4.43277657e-01
-1.29778594e-01 -4.75539297e-01 1.70858148e-02 -1.63531438e-01
-3.48848909e-01 7.49589920e-01 3.83973747e-01 -5.67324460e-01
-3.94858330e-01 8.41291308e-01 5.62301874e-01 1.73414737e-01
1.63560081e+00 2.11940333e-01 -2.90241659e-01 4.06740397e-01
1.32717144e+00 -7.60711282e-02 -1.18494689e+00 3.49382371e-01
-2.10069820e-01 1.33144677e-01 5.49400568e-01 -7.17193365e-01
-1.67382443e+00 1.01246953e+00 6.52435124e-01 2.58806851e-02
7.90622234e-01 -3.00462991e-01 1.08830738e+00 1.83756143e-01
4.36585456e-01 -1.22863245e+00 1.24686264e-01 4.25592721e-01
1.30738378e+00 -6.81081831e-01 2.14083306e-02 -3.76219720e-01
-2.03530770e-02 1.23128200e+00 8.05502772e-01 -5.81674635e-01
1.16364300e+00 5.45824289e-01 -1.70261756e-01 -8.00699949e-01
-2.97244519e-01 5.37604332e-01 6.07051969e-01 2.79349685e-01
2.98005670e-01 2.02416241e-01 -4.46677923e-01 8.95873189e-01
-2.08366394e-01 -1.88846797e-01 1.36832193e-01 8.04847658e-01
-3.39824498e-01 -8.98087502e-01 -2.05517739e-01 6.36268556e-01
-1.80618078e-01 3.04881483e-03 -1.64548546e-01 9.78377759e-01
7.44090602e-02 4.07671452e-01 -7.57750915e-03 -6.74832940e-01
5.13746560e-01 -1.63653776e-01 3.30536336e-01 -2.79591769e-01
-5.79744399e-01 -1.48700655e-01 -2.36172266e-02 -6.18869424e-01
1.02346487e-01 -1.96821347e-01 -1.50099850e+00 -6.09589458e-01
-3.85031909e-01 7.55400434e-02 4.44063663e-01 4.78815228e-01
3.74271244e-01 8.37518156e-01 5.36148965e-01 -1.08073997e+00
-2.31058732e-01 -6.64631426e-01 -7.50432014e-01 2.51681417e-01
6.39889343e-03 -1.10301769e+00 -3.12209755e-01 -3.43575507e-01] | [6.29596471786499, 3.403398275375366] |
bbcd195f-a2de-47ff-9688-98f4bea55f04 | dually-enhanced-propensity-score-estimation | 2303.08722 | null | https://arxiv.org/abs/2303.08722v1 | https://arxiv.org/pdf/2303.08722v1.pdf | Dually Enhanced Propensity Score Estimation in Sequential Recommendation | Sequential recommender systems train their models based on a large amount of implicit user feedback data and may be subject to biases when users are systematically under/over-exposed to certain items. Unbiased learning based on inverse propensity scores (IPS), which estimate the probability of observing a user-item pair given the historical information, has been proposed to address the issue. In these methods, propensity score estimation is usually limited to the view of item, that is, treating the feedback data as sequences of items that interacted with the users. However, the feedback data can also be treated from the view of user, as the sequences of users that interact with the items. Moreover, the two views can jointly enhance the propensity score estimation. Inspired by the observation, we propose to estimate the propensity scores from the views of user and item, called Dually Enhanced Propensity Score Estimation (DEPS). Specifically, given a target user-item pair and the corresponding item and user interaction sequences, DEPS firstly constructs a time-aware causal graph to represent the user-item observational probability. According to the graph, two complementary propensity scores are estimated from the views of item and user, respectively, based on the same set of user feedback data. Finally, two transformers are designed to make the final preference prediction. Theoretical analysis showed the unbiasedness and variance of DEPS. Experimental results on three publicly available and an industrial datasets demonstrated that DEPS can significantly outperform the state-of-the-art baselines. | ['Ji-Rong Wen', 'Zhenghua Dong', 'Xu Chen', 'Jun Xu', 'Chen Xu'] | 2023-03-15 | null | null | null | null | ['sequential-recommendation'] | ['miscellaneous'] | [-4.46699895e-02 -1.89606175e-01 -5.97523212e-01 -5.22662818e-01
-4.29852307e-01 -5.19805074e-01 4.45503503e-01 7.37988800e-02
-1.36858061e-01 5.12322843e-01 5.32507777e-01 -6.42389134e-02
-3.21588278e-01 -8.52512777e-01 -6.50578558e-01 -6.58899844e-01
-2.05202863e-01 3.46386731e-01 -1.90125648e-02 -2.30737235e-02
1.82489246e-01 -3.25154394e-01 -1.37356532e+00 1.48225620e-01
1.18611574e+00 1.03448915e+00 4.79879469e-01 2.40568653e-01
2.53147364e-01 6.36647940e-01 1.13790624e-01 -5.12349427e-01
2.85871655e-01 -3.87371123e-01 -1.66829020e-01 -5.00888899e-02
2.45856106e-01 -5.38085520e-01 -4.16088730e-01 1.10443485e+00
2.19779268e-01 3.45557123e-01 5.94317019e-01 -1.29609454e+00
-1.05988121e+00 7.00451970e-01 -5.25168002e-01 -2.67334692e-02
6.57434762e-01 -1.78005204e-01 1.51686418e+00 -9.92857456e-01
2.87232310e-01 1.14358211e+00 3.89683723e-01 2.17289314e-01
-1.11745453e+00 -8.92088830e-01 8.35228503e-01 4.08176988e-01
-9.79392111e-01 3.61942686e-02 7.87958205e-01 -5.93709946e-01
1.37981966e-01 2.57695198e-01 7.95390427e-01 1.29337561e+00
1.34833440e-01 1.24091506e+00 1.11401439e+00 7.90267438e-02
4.07857299e-01 6.29950762e-01 5.95184147e-01 4.47532177e-01
1.29577398e-01 5.68906665e-01 -6.54220045e-01 -5.52655876e-01
4.79870021e-01 6.17929041e-01 -4.85803276e-01 -8.34675968e-01
-1.17714500e+00 9.11575496e-01 4.59750384e-01 -2.72984207e-01
-6.57711387e-01 -5.78730583e-01 1.47744909e-01 2.78138906e-01
4.33738232e-01 1.39996931e-01 -7.46950388e-01 1.52724445e-01
-6.27848387e-01 3.64274859e-01 9.07026410e-01 1.16696572e+00
6.85118616e-01 -3.63975346e-01 -3.93608570e-01 5.36259890e-01
7.29986370e-01 6.66033924e-01 5.35588562e-01 -3.11765373e-01
5.51981449e-01 7.95548320e-01 6.03236616e-01 -1.25998890e+00
2.30729301e-03 -6.21650994e-01 -8.14525783e-01 -3.34413588e-01
3.65581691e-01 -1.89981461e-01 -6.64283216e-01 2.04987836e+00
4.11884397e-01 4.84233499e-01 -1.58950895e-01 1.22693062e+00
5.30727088e-01 6.23404503e-01 8.11677948e-02 -7.38446295e-01
1.32023847e+00 -7.94587255e-01 -7.86901712e-01 -4.77253012e-02
4.64812875e-01 -4.06091750e-01 1.17671204e+00 6.37561083e-01
-8.72490346e-01 -5.84996581e-01 -9.59056497e-01 5.47405005e-01
2.90384516e-02 3.15737844e-01 5.78238070e-01 6.22280359e-01
-3.19246948e-01 6.44221485e-01 -5.43358922e-01 -1.70217723e-01
1.94677368e-01 3.61748636e-01 1.61478519e-02 -5.83774373e-02
-1.54582810e+00 4.82935280e-01 9.76678282e-02 2.12339684e-02
-9.23394203e-01 -9.64277565e-01 -5.01258314e-01 2.65368253e-01
5.47591388e-01 -7.46659994e-01 1.30595672e+00 -1.04045999e+00
-1.40720522e+00 1.62165090e-02 -3.05899173e-01 -2.12464735e-01
4.92578179e-01 -5.56380868e-01 -6.99497700e-01 -5.11061311e-01
-5.78431087e-03 -2.08319515e-01 7.66156435e-01 -1.17648351e+00
-1.03504193e+00 -6.79383278e-01 3.94837037e-02 4.82355505e-01
-5.37799239e-01 -3.35739464e-01 -5.54761827e-01 -5.81228435e-01
5.20323925e-02 -1.02756214e+00 -2.16964975e-01 -5.01877367e-01
-2.47960642e-01 -2.46958807e-01 5.61551929e-01 -6.89042032e-01
1.52749097e+00 -2.14319468e+00 2.21809700e-01 3.30639571e-01
1.60578310e-01 -2.15340592e-03 -9.26984921e-02 3.83304328e-01
-8.81112590e-02 -1.95833355e-01 1.68870762e-01 4.30666981e-03
-5.67369396e-03 -2.16217674e-02 -6.96130276e-01 4.45871592e-01
-5.23517311e-01 6.31087482e-01 -1.45511925e+00 -1.84862196e-01
1.16376832e-01 -1.08757786e-01 -8.12139869e-01 7.59256601e-01
-5.76948822e-02 3.95967096e-01 -6.14277005e-01 1.91983402e-01
6.51064575e-01 -3.46831620e-01 5.34865141e-01 -3.10957491e-01
7.65248686e-02 3.31671447e-01 -1.43366778e+00 1.38261044e+00
-5.62335312e-01 -2.54101008e-01 -3.47343296e-01 -7.28303015e-01
7.32061744e-01 5.45066535e-01 5.50788462e-01 -6.37937188e-01
-5.74754663e-02 5.64335138e-02 9.00279731e-02 -4.96160239e-01
3.49269480e-01 3.70846428e-02 -1.28345445e-01 6.86458290e-01
3.06498617e-01 5.96398592e-01 -4.39594612e-02 4.07699525e-01
6.90612972e-01 1.33298546e-01 5.58768094e-01 -2.31791660e-01
5.91596544e-01 -4.82931584e-01 8.43642831e-01 9.15484369e-01
-6.91777244e-02 2.16774270e-01 1.64311707e-01 -3.23038131e-01
-6.02791250e-01 -1.41504526e+00 1.47597209e-01 1.36392045e+00
6.03276908e-01 -4.64124382e-01 -2.75809735e-01 -1.12123501e+00
-8.63413326e-03 8.91149223e-01 -8.93621266e-01 -3.90836507e-01
-4.04246151e-02 -7.82448590e-01 -6.24279022e-01 5.12067199e-01
2.10978031e-01 -7.52899289e-01 4.03691381e-02 3.41693878e-01
-2.20111579e-01 -5.23402035e-01 -1.11335969e+00 -2.10042894e-01
-8.55957568e-01 -1.08002794e+00 -6.58635020e-01 -3.78795892e-01
5.95644057e-01 6.23692572e-01 9.92109418e-01 -3.10869366e-01
6.74046695e-01 3.17209721e-01 -3.99712056e-01 -5.08081377e-01
1.79792106e-01 -1.39683649e-01 6.52400434e-01 7.35761523e-01
4.49638337e-01 -7.84669995e-01 -1.06929386e+00 7.92122245e-01
-4.95030463e-01 1.22045755e-01 7.20432043e-01 9.09727752e-01
4.65521008e-01 -1.47169590e-01 6.74446404e-01 -1.29553592e+00
6.33335233e-01 -1.09685802e+00 -5.58599412e-01 4.06267941e-01
-1.11663616e+00 7.44545832e-02 7.87725508e-01 -9.02588069e-01
-1.43341851e+00 1.50585652e-03 4.16065037e-01 -4.01116371e-01
5.40512018e-02 7.51121223e-01 -4.77905750e-01 7.05535948e-01
3.24936837e-01 6.40201047e-02 -4.04169112e-01 -5.51409781e-01
5.49612820e-01 7.12452114e-01 3.03385884e-01 -3.31508815e-01
6.32905781e-01 3.36847275e-01 -4.36945379e-01 -9.64776501e-02
-1.25471807e+00 -8.34835529e-01 -3.67324352e-01 -2.13510886e-01
4.85882699e-01 -1.06582105e+00 -7.92126298e-01 3.16293478e-01
-7.08194852e-01 7.10320547e-02 -7.48316646e-02 9.51000690e-01
-4.72977161e-01 2.19411775e-01 -1.86426103e-01 -1.09410143e+00
-1.93593889e-01 -8.29958618e-01 8.18638682e-01 4.19191867e-01
-3.23991209e-01 -1.26741517e+00 3.89050364e-01 4.64105941e-02
-8.47874358e-02 -3.92757267e-01 7.43585348e-01 -8.71855021e-01
-5.12662649e-01 -4.67927456e-01 -8.86959955e-03 7.23092556e-02
4.22421694e-01 -3.10541391e-01 -8.24238062e-01 -4.47103083e-01
2.77052283e-01 -5.05196920e-04 3.88644308e-01 5.03419995e-01
1.04008126e+00 -6.38303041e-01 -5.09490252e-01 1.73167109e-01
1.29158449e+00 2.97109693e-01 2.51751840e-01 -2.61031270e-01
7.25726187e-01 6.52284861e-01 1.17143667e+00 7.14610815e-01
4.03458476e-01 7.41351664e-01 3.65527421e-01 3.18897188e-01
5.70071936e-01 -1.04426086e+00 5.24439573e-01 1.15312755e+00
-5.37755303e-02 -3.51754874e-01 -1.63319141e-01 5.09546876e-01
-2.37691927e+00 -9.87515509e-01 -4.74121600e-01 2.72114015e+00
4.80752856e-01 -8.89555663e-02 2.69394010e-01 -2.79590964e-01
7.73241937e-01 -1.10763637e-02 -7.86976695e-01 2.20625296e-01
3.60749692e-01 -4.17601138e-01 4.62510735e-01 4.29136992e-01
-7.65323937e-01 3.39760154e-01 5.41651344e+00 6.66720510e-01
-5.59873521e-01 1.53989881e-01 3.83783221e-01 -3.44021529e-01
-5.91698110e-01 2.10104913e-01 -7.13394105e-01 9.37254906e-01
8.95528257e-01 -5.94907105e-01 5.42050421e-01 8.24271083e-01
3.95636469e-01 1.15097174e-02 -1.33062410e+00 7.53562748e-01
-2.26723462e-01 -6.52251244e-01 2.58258637e-02 2.89041191e-01
9.95322287e-01 -1.72341987e-01 2.92821079e-01 5.59185266e-01
9.14408386e-01 -4.24294204e-01 7.07768381e-01 9.87774670e-01
2.95491219e-01 -5.41647971e-01 8.53499830e-01 7.19631910e-01
-1.27726102e+00 -3.20972055e-01 -3.01159114e-01 -2.57977128e-01
2.81599194e-01 7.12928116e-01 -5.85756838e-01 8.89387846e-01
7.08850265e-01 1.06492472e+00 -2.16175124e-01 9.25979376e-01
-3.34775835e-01 9.30667162e-01 6.68102130e-02 -2.15619043e-01
-2.16664806e-01 -7.06366420e-01 4.98867452e-01 5.09255707e-01
5.86012125e-01 3.43875587e-01 2.60969371e-01 7.69955099e-01
4.44248356e-02 2.91125059e-01 -5.52113533e-01 2.59554684e-01
5.82269430e-01 1.42042720e+00 1.63614064e-01 -5.09060204e-01
-7.73575008e-01 8.18274200e-01 4.76912469e-01 5.99041224e-01
-7.60427773e-01 8.22514668e-02 6.25160992e-01 -7.54615292e-02
2.34805018e-01 2.59174079e-01 -1.54813856e-01 -1.30868602e+00
-2.42341664e-02 -7.12539256e-01 6.45323873e-01 -6.84657454e-01
-1.77592123e+00 1.05484702e-01 -1.43924668e-01 -1.70617545e+00
-8.00136942e-03 -3.07889640e-01 -6.88641250e-01 1.17090178e+00
-1.03011096e+00 -7.22719014e-01 -2.27185339e-01 5.73278904e-01
3.74760985e-01 -1.32693365e-01 6.91585183e-01 1.09838463e-01
-4.98622596e-01 6.25165820e-01 3.01542133e-01 -2.22288504e-01
8.17530751e-01 -1.42182994e+00 1.56126037e-01 5.59876919e-01
2.43637338e-01 1.06080997e+00 6.48702681e-01 -8.80683661e-01
-1.42665398e+00 -1.05042934e+00 7.00555801e-01 -7.80790508e-01
7.38059521e-01 -3.40013057e-01 -9.36925948e-01 6.95522726e-01
-2.20957156e-02 -2.31813014e-01 9.30316985e-01 6.87511325e-01
-2.48004287e-01 -1.84183210e-01 -7.55028427e-01 8.54492843e-01
1.13961089e+00 -5.16317546e-01 -7.63983190e-01 3.05830061e-01
6.32921636e-01 -2.44095623e-01 -1.00738037e+00 2.79104918e-01
8.90474439e-01 -9.02091503e-01 8.09492826e-01 -8.78657281e-01
3.88716161e-01 -2.91583389e-01 -3.41827534e-02 -1.80391192e+00
-9.05364394e-01 -3.54019105e-01 -5.67372561e-01 1.31078601e+00
3.74769896e-01 -5.19266605e-01 7.14366078e-01 8.05777073e-01
1.56617165e-01 -7.01152444e-01 -3.56453896e-01 -6.79021239e-01
-4.20805365e-01 -2.28727162e-01 7.80449986e-01 7.90197432e-01
4.33726966e-01 7.02325761e-01 -1.03538251e+00 5.25675476e-01
6.98647320e-01 5.16373754e-01 7.32223988e-01 -1.48590207e+00
-5.78202963e-01 -3.71378288e-02 1.98265746e-01 -1.30520892e+00
-1.10686675e-01 -6.71729505e-01 2.44408362e-02 -1.23731089e+00
6.03253782e-01 -3.14997315e-01 -9.24286604e-01 -1.13985747e-01
-8.82064462e-01 -4.02318925e-01 -6.15502633e-02 3.92536908e-01
-6.51562810e-01 8.60488415e-01 1.32389903e+00 8.36123750e-02
-3.98030013e-01 6.04381740e-01 -9.32683349e-01 7.17179656e-01
4.62114692e-01 -5.13875723e-01 -8.89782846e-01 7.12816417e-02
3.55573773e-01 3.45335335e-01 5.15643624e-04 -5.32637596e-01
2.04358891e-01 -5.18755794e-01 1.35715783e-01 -6.52405739e-01
-4.87939306e-02 -1.18163276e+00 2.46156096e-01 3.89435291e-01
-5.50604522e-01 -7.84839913e-02 -4.52122539e-01 1.37712443e+00
-3.82392034e-02 -1.15631707e-01 2.57561743e-01 -4.80480716e-02
-4.78226721e-01 8.79257560e-01 -8.56147408e-02 -1.64147526e-01
8.90208960e-01 2.15661585e-01 2.17126179e-02 -6.11196280e-01
-6.75101340e-01 5.50979137e-01 1.93329215e-01 6.74200952e-01
4.98157024e-01 -1.77686453e+00 -4.83230114e-01 1.28080085e-01
4.30407554e-01 -4.55533206e-01 6.29686058e-01 8.98241639e-01
6.98464632e-01 4.12550569e-01 7.33526722e-02 -3.34375083e-01
-1.13575113e+00 1.17656398e+00 -7.49786273e-02 -5.49451947e-01
4.94233929e-02 6.11390710e-01 8.48583162e-01 -6.05097830e-01
9.99406651e-02 -7.96333626e-02 -5.20973802e-01 3.06230009e-01
5.66872716e-01 4.45888132e-01 -1.86230242e-01 -3.13139111e-01
-1.43704340e-01 1.97038308e-01 -4.17861462e-01 -2.27130055e-01
1.17940426e+00 -5.12819231e-01 2.82656640e-01 8.09084952e-01
1.04962468e+00 2.32607648e-01 -1.42721856e+00 -7.17063308e-01
-1.93868384e-01 -8.35223615e-01 3.72370183e-02 -8.69277775e-01
-1.06986630e+00 7.24390090e-01 6.85737371e-01 3.07446569e-01
9.64333713e-01 -3.41496170e-01 5.78131139e-01 -4.21144105e-02
6.46458507e-01 -8.58490348e-01 -1.98271628e-02 1.48475558e-01
6.45790160e-01 -1.32798064e+00 -5.09205423e-02 -3.29262376e-01
-9.04802203e-01 4.83800203e-01 7.67865837e-01 -2.15075061e-01
9.66156960e-01 -4.87894893e-01 -1.68327570e-01 -5.18810144e-03
-9.70529675e-01 7.11845011e-02 8.04537296e-01 3.96393657e-01
4.98778075e-01 4.99679297e-01 -3.54032606e-01 1.28510320e+00
5.83083951e-04 2.15969950e-01 2.29457662e-01 4.49465662e-01
-4.80819829e-02 -1.09695065e+00 -2.71502286e-02 8.80236685e-01
-8.23038071e-03 -1.85743675e-01 -1.00092858e-01 2.78937101e-01
-6.28372431e-02 1.02491605e+00 -1.87998965e-01 -7.88925350e-01
6.36557758e-01 -1.05397001e-01 1.40663922e-01 -6.80165708e-01
-3.66916835e-01 -1.94042876e-01 -1.87069684e-01 -6.09709799e-01
-3.09625089e-01 -9.76134837e-01 -6.51428163e-01 -1.38446614e-01
-7.24187374e-01 4.77621406e-01 8.70969519e-02 9.24300790e-01
4.04245794e-01 3.93108904e-01 1.17649913e+00 -7.00415611e-01
-9.57198918e-01 -1.03082466e+00 -1.00442612e+00 8.32877994e-01
1.13966428e-01 -8.95254314e-01 -3.74702722e-01 -4.72997911e-02] | [9.95627498626709, 5.562069416046143] |
f3aec1f8-8e19-497a-ae52-73d620e62e09 | a2-link-recognizing-disguised-faces-via | null | null | https://ieeexplore.ieee.org/document/9104705 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9104705&tag=1 | A2-LINK: Recognizing Disguised Faces via Active Learning and Adversarial Noise based Inter-Domain Knowledge | Face recognition in the unconstrained environment is an ongoing research challenge. Although several covariates of face recognition such as pose and low resolution have received significant attention“, disguise” is considered an onerous covariate of face recognition. One of the primary reasons for this is the scarcity of large and representative labeled databases, along with the lack of algorithms that work well for multiple covariates in such environments. In order to address the problem of face recognition in the presence of disguise, the paper proposes an active learning framework termed as A2-LINK. Starting with a face recognition machine-learning model, A2-LINK intelligently selects training samples from the target domain to be labeled and, using hybrid noises such as adversarial noise, fine-tunes a model that works well both in the presence and absence of disguise. Experimental results demonstrate the effectiveness and generalization of the proposed framework on the DFW and DFW2019 datasets with state-of-the-art deep learning featurization models such as LCSSE, ArcFace, and DenseNet. | ['Mayank Vatsa', 'Richa Singh', 'Anshuman Suri'] | 2020-06-01 | null | null | null | ieee-transactions-on-biometrics-behavior-and | ['heterogeneous-face-recognition'] | ['computer-vision'] | [ 1.84856236e-01 2.29284674e-01 -1.58462316e-01 -8.16306770e-01
-5.65245926e-01 -2.12227434e-01 7.08223343e-01 -5.90146780e-01
-3.90122265e-01 5.69165826e-01 7.55858496e-02 1.82296664e-01
-4.40126568e-01 -5.78980982e-01 -9.01590168e-01 -9.55667615e-01
-4.57228832e-02 4.03128713e-01 -3.43998283e-01 1.03079244e-01
-2.90508498e-03 9.87959266e-01 -1.70138180e+00 -1.75537542e-01
4.68773454e-01 1.17689312e+00 -3.30538899e-01 1.19346172e-01
-1.44775122e-01 4.38991457e-01 -6.07605696e-01 -6.27053499e-01
6.04885638e-01 -1.49166360e-01 -3.40478361e-01 3.24929893e-01
1.07808864e+00 -3.84304911e-01 -2.53252357e-01 9.67321932e-01
9.58536267e-01 6.89976215e-02 5.53679228e-01 -1.28389299e+00
-6.52696311e-01 3.39875042e-01 -4.05307114e-01 1.98111430e-01
6.40580803e-02 3.60354483e-01 2.89723486e-01 -1.23480678e+00
4.80574936e-01 1.58698583e+00 7.22853124e-01 9.60514784e-01
-1.35644531e+00 -1.06404471e+00 2.98715144e-01 2.35619038e-01
-1.37464368e+00 -1.10038257e+00 1.01192033e+00 -2.95711964e-01
6.19640768e-01 2.11516842e-01 2.06588209e-01 1.70506096e+00
-2.72584438e-01 6.07688963e-01 1.03102505e+00 -5.38073421e-01
3.64974171e-01 1.23140365e-01 -3.98915038e-02 4.09752339e-01
9.09785405e-02 4.35316861e-01 -8.26354563e-01 -2.37634271e-01
5.32789648e-01 -7.75362402e-02 -3.10407668e-01 -4.79616821e-01
-6.44599497e-01 6.79000080e-01 3.63271028e-01 6.86537102e-02
-3.13891470e-01 -2.40875527e-01 1.03754222e-01 2.35419735e-01
7.81347096e-01 3.53924632e-01 -2.23148748e-01 1.86079085e-01
-1.11377311e+00 9.41655133e-03 7.29662120e-01 7.68920779e-01
5.21950543e-01 3.53951067e-01 6.34438694e-02 8.57690811e-01
6.07518792e-01 5.00108600e-01 1.44773141e-01 -4.28512365e-01
2.51068622e-01 5.74916244e-01 5.42951450e-02 -1.12726176e+00
-1.02957800e-01 -4.76022005e-01 -7.59876192e-01 7.06229150e-01
4.73567933e-01 -2.08488673e-01 -1.09211397e+00 1.89949214e+00
3.45738888e-01 5.62932491e-01 -1.77750677e-01 8.50125313e-01
1.15767705e+00 1.68721050e-01 3.59740019e-01 -4.89093751e-01
8.78655732e-01 -7.35609770e-01 -8.07163298e-01 -3.40964735e-01
-2.80449800e-02 -7.41477668e-01 6.13682210e-01 4.07165378e-01
-7.06308365e-01 -7.18460083e-01 -1.07107460e+00 2.72246897e-01
-3.45041960e-01 -3.06961406e-02 4.15597856e-01 1.08783758e+00
-1.11766255e+00 4.19818401e-01 -7.13544130e-01 -4.00086343e-01
8.73809099e-01 6.57290995e-01 -7.99117804e-01 -1.93543151e-01
-9.84097242e-01 1.03423047e+00 -6.32133409e-02 7.17504859e-01
-1.29597938e+00 -7.78166950e-01 -6.91284478e-01 -1.09168850e-01
5.85512400e-01 -3.63204591e-02 6.07419789e-01 -1.29510617e+00
-1.35116589e+00 9.86916602e-01 1.02451913e-01 -2.29505345e-01
7.65962422e-01 -4.83066261e-01 -6.57091379e-01 -1.96948424e-01
-3.61339509e-01 5.26208222e-01 1.33685815e+00 -1.28347170e+00
1.25096664e-01 -8.89229059e-01 -2.20221933e-02 1.43569306e-01
-6.92439973e-01 2.88208038e-01 -2.93226719e-01 -5.96626461e-01
-2.09613606e-01 -7.45927453e-01 5.13653345e-02 4.05025423e-01
-3.19550723e-01 -1.95661575e-01 1.27110684e+00 -5.27064443e-01
7.59873688e-01 -2.36244774e+00 7.83653632e-02 1.53595343e-01
2.02232257e-01 6.07030332e-01 -3.35137218e-01 -8.20675027e-03
-4.26528871e-01 5.22060692e-02 -8.67897421e-02 -5.62552929e-01
-6.12257235e-02 8.04525688e-02 -2.48442709e-01 7.69596517e-01
6.55396700e-01 6.83718681e-01 -5.41292012e-01 -3.00584733e-01
1.11162961e-01 8.42450321e-01 -4.40531880e-01 6.40661418e-01
-8.59003812e-02 7.19353378e-01 -1.27305642e-01 6.59049392e-01
1.06324756e+00 3.41094673e-01 -3.18284743e-02 -4.55842704e-01
-4.19202931e-02 -2.01176524e-01 -1.18685639e+00 1.32136381e+00
-1.81764945e-01 3.91943514e-01 3.89740676e-01 -1.12705648e+00
1.32409918e+00 5.74026763e-01 4.41601515e-01 -6.91494167e-01
3.20144325e-01 1.04372106e-01 2.09998153e-03 -6.88400745e-01
-1.95466608e-01 1.25553042e-01 5.17882466e-01 2.10752174e-01
4.71810371e-01 3.98014724e-01 -1.50816262e-01 -3.08631122e-01
8.92893791e-01 4.31303054e-01 7.44341835e-02 -3.84563208e-01
5.20397127e-01 -6.46737754e-01 6.85012460e-01 3.81055325e-01
-5.98232031e-01 5.10596573e-01 2.78941244e-01 -4.58872736e-01
-7.38455832e-01 -8.63673270e-01 -3.06488782e-01 1.09298813e+00
-2.37226039e-01 1.79642573e-01 -7.17310011e-01 -9.28864121e-01
1.93480268e-01 4.63733822e-01 -1.05660951e+00 -2.60005176e-01
-5.87412238e-01 -9.06858087e-01 6.62114203e-01 3.58654886e-01
7.35088170e-01 -1.15697432e+00 -2.87025154e-01 -1.68435887e-01
3.10853750e-01 -8.61677527e-01 -9.69177410e-02 2.28482485e-01
-5.20638108e-01 -1.19475925e+00 -6.36505365e-01 -8.23719084e-01
7.98942268e-01 -2.87060648e-01 1.15630424e+00 1.05840757e-01
-5.33640504e-01 2.71995693e-01 -2.06511974e-01 -5.38285553e-01
-4.24458869e-02 -2.67015576e-01 1.18429333e-01 6.45666718e-01
5.13020575e-01 -5.21898031e-01 -6.16567194e-01 4.47109491e-01
-7.67486274e-01 -4.94401097e-01 4.23916638e-01 6.68598115e-01
2.19207987e-01 -4.45256159e-02 8.05113375e-01 -1.12308395e+00
2.87712246e-01 -5.24132013e-01 -4.23964471e-01 2.75360405e-01
-4.88896489e-01 -3.28971744e-01 3.71328443e-01 -7.10215807e-01
-1.28055358e+00 2.83577114e-01 -1.86918318e-01 -5.52312672e-01
-5.04252672e-01 8.31929147e-02 -1.02729952e+00 -4.27385271e-01
8.38375926e-01 -8.91673937e-02 1.92052126e-01 -5.27932107e-01
2.41205886e-01 6.16078734e-01 4.02041793e-01 -4.49567258e-01
1.07091415e+00 3.77583712e-01 1.41272591e-02 -7.70419657e-01
-8.25526476e-01 -2.28671655e-02 -7.69275308e-01 -5.37814558e-01
4.91702497e-01 -7.03947723e-01 -4.72220778e-01 8.03290665e-01
-8.68395865e-01 -1.49742559e-01 1.81472376e-02 2.46704593e-01
-2.58834779e-01 7.07755908e-02 5.77385314e-02 -1.08499956e+00
-2.62913704e-01 -1.24567223e+00 6.61918163e-01 3.94038498e-01
-4.97220419e-02 -8.90703201e-01 -2.72750050e-01 4.76810187e-01
7.36385942e-01 5.59625864e-01 7.84988880e-01 -1.01423132e+00
-4.09990519e-01 -1.70105323e-01 5.08217067e-02 6.62486970e-01
3.57458711e-01 1.51270693e-02 -1.50966990e+00 -4.82739121e-01
3.90427262e-01 -7.12679386e-01 4.78161573e-01 -5.83098382e-02
1.13837993e+00 -4.26263422e-01 -6.34697303e-02 7.38408029e-01
1.12130952e+00 2.33480126e-01 8.16106915e-01 -4.89090942e-02
4.78147984e-01 7.90879786e-01 4.40118968e-01 1.81251422e-01
-3.00651968e-01 6.57346785e-01 7.17873573e-01 -2.37379104e-01
-2.66499430e-01 8.23582858e-02 1.72284544e-01 3.20111781e-01
9.24900826e-03 -1.31701648e-01 -9.17302310e-01 3.29122990e-01
-1.51112020e+00 -9.45769668e-01 3.10378700e-01 2.22347355e+00
7.32707441e-01 3.16778123e-02 -5.11080027e-02 3.54445100e-01
8.25438857e-01 5.21652848e-02 -8.02412927e-01 -1.62925627e-02
-1.49185285e-01 4.57778156e-01 2.64461666e-01 1.24109879e-01
-1.35390067e+00 6.51634872e-01 5.37870598e+00 6.79691076e-01
-1.58460462e+00 8.66374597e-02 9.49601233e-01 1.85445007e-02
2.22789288e-01 -4.41365242e-01 -9.01988745e-01 5.36794007e-01
6.26291871e-01 2.79193103e-01 4.64291900e-01 1.03596818e+00
1.19859323e-01 3.38091910e-01 -1.53604436e+00 1.02381372e+00
5.87667644e-01 -7.84918904e-01 -9.74084958e-02 -5.86644337e-02
5.01787364e-01 -3.83857399e-01 4.66651022e-01 3.39791030e-01
-9.78177115e-02 -1.48167205e+00 7.79626608e-01 6.04709804e-01
8.80168974e-01 -6.51180804e-01 5.29415905e-01 3.18153769e-01
-6.63888752e-01 -1.54050797e-01 -2.03563854e-01 2.91790664e-01
-2.67374396e-01 2.61994034e-01 -7.65183866e-01 3.10405284e-01
5.87827504e-01 4.30720419e-01 -8.58588755e-01 1.12041497e+00
-1.30795106e-01 8.46208513e-01 -2.50651419e-01 4.80261683e-01
-2.35103682e-01 7.51277208e-02 4.89964873e-01 9.01609778e-01
8.80478621e-02 -1.77920103e-01 8.86670873e-02 7.47771561e-01
-2.80232280e-01 -3.94124202e-02 -5.19154847e-01 -2.01703429e-01
6.97964549e-01 1.23095953e+00 -4.22713608e-01 1.94579020e-01
-3.67034048e-01 6.57287598e-01 5.13855219e-01 4.36457127e-01
-6.67175651e-01 7.21904039e-02 5.45464754e-01 3.42697442e-01
2.22096801e-01 -5.23947328e-02 -2.27242753e-01 -7.41582394e-01
-2.84697767e-02 -1.11311066e+00 2.12340102e-01 -3.88683826e-01
-1.54630482e+00 7.83641756e-01 -8.33050162e-03 -9.11931813e-01
-9.23917964e-02 -7.27457881e-01 -6.04292214e-01 1.09515560e+00
-1.40202987e+00 -1.38208234e+00 -5.46079934e-01 7.51263440e-01
4.50391859e-01 -7.48389244e-01 1.18096912e+00 6.42376423e-01
-7.22990692e-01 9.56521273e-01 -9.32157561e-02 4.99320000e-01
9.28229213e-01 -7.27677822e-01 5.07917047e-01 8.66974115e-01
2.16240901e-02 8.74879479e-01 6.67342722e-01 -5.58207572e-01
-1.56515372e+00 -1.20525897e+00 5.45214713e-01 -4.31321383e-01
2.88586199e-01 -7.20446765e-01 -8.71820092e-01 4.60546702e-01
-2.71278527e-03 5.26303291e-01 7.77436435e-01 1.10991774e-02
-7.85014749e-01 -4.08318460e-01 -1.64830530e+00 2.27591559e-01
1.25143683e+00 -5.26033878e-01 -4.52302873e-01 1.37895599e-01
8.77950341e-02 -1.65076897e-01 -5.38305104e-01 8.25843990e-01
4.31023031e-01 -9.43410039e-01 1.10635746e+00 -7.38132060e-01
-1.74368601e-02 1.29396319e-01 -1.26285061e-01 -1.17099154e+00
-3.48273337e-01 -5.14423132e-01 -1.50925279e-01 1.64443874e+00
2.10388154e-01 -5.67354143e-01 9.95027363e-01 7.63697565e-01
1.87620059e-01 -6.48068428e-01 -1.23941588e+00 -6.90588653e-01
-1.32070437e-01 -1.53151527e-01 6.03703856e-01 9.92847681e-01
-8.33960414e-01 1.57555848e-01 -3.92058045e-01 5.55120744e-02
9.52379048e-01 -3.22848082e-01 7.82622099e-01 -1.46042228e+00
6.93952590e-02 -1.03785992e-01 -5.92721224e-01 -5.67906499e-01
4.81644839e-01 -6.48079395e-01 6.36311918e-02 -9.14864838e-01
-2.56369524e-02 -5.69854975e-01 -1.96900874e-01 6.23793185e-01
8.65192618e-03 5.76140344e-01 1.36611953e-01 5.49112447e-02
-1.83777779e-01 6.83458507e-01 9.72036123e-01 -2.04737246e-01
2.37661853e-01 -1.14871413e-01 -6.19020700e-01 6.53352797e-01
5.02725959e-01 -2.93460906e-01 -4.36865956e-01 -5.29877186e-01
-1.43487349e-01 -2.69582659e-01 4.03966576e-01 -9.48803484e-01
1.39989868e-01 -1.45255119e-01 8.39888871e-01 -1.83100611e-01
7.07421899e-01 -1.20930862e+00 3.15438122e-01 -1.09865628e-02
-4.27984059e-01 -2.96573430e-01 2.20440164e-01 4.75776851e-01
3.45363468e-02 -1.37161240e-01 1.20319200e+00 1.37359694e-01
-6.24762297e-01 6.39425278e-01 1.61582783e-01 6.44720867e-02
9.64158893e-01 -3.22542012e-01 -3.73469025e-01 -7.68434033e-02
-8.18057537e-01 -1.62350297e-01 6.48524314e-02 8.67755175e-01
5.37119806e-01 -1.28599596e+00 -8.44377100e-01 6.31571412e-01
2.45944317e-02 -1.25603095e-01 1.10958137e-01 5.71682870e-01
-1.58673376e-01 1.70588866e-01 -4.83168006e-01 -4.99331653e-01
-1.44899786e+00 4.52978790e-01 4.33295816e-01 2.87910014e-01
-2.53555179e-01 1.17368281e+00 1.88228592e-01 -4.21965301e-01
7.63517380e-01 6.08140945e-01 -4.17714685e-01 1.95601001e-01
5.90448499e-01 1.12076193e-01 5.46181738e-01 -1.00962114e+00
-5.87455928e-01 1.19337879e-01 -1.18356839e-01 3.48045528e-01
1.51327276e+00 1.03699811e-01 -1.34700164e-02 5.69929443e-02
1.16550255e+00 -1.35314628e-01 -1.35695910e+00 -2.87773788e-01
1.50069267e-01 -8.23876977e-01 1.17338084e-01 -1.02461708e+00
-1.40163243e+00 6.63017809e-01 1.19841623e+00 -1.64971575e-01
9.97004330e-01 -3.57901782e-01 3.72263975e-02 1.39065877e-01
3.22605968e-01 -8.35134029e-01 2.57453412e-01 3.31833512e-01
1.32862151e+00 -1.23741031e+00 -7.34899044e-02 -4.62000430e-01
-2.12245852e-01 9.55993891e-01 1.08064628e+00 -1.20841712e-01
1.00675404e+00 5.03719032e-01 4.31082875e-01 -6.92706108e-02
-4.41168994e-01 1.08348101e-01 3.36060345e-01 7.30384707e-01
4.88069892e-01 -1.90656468e-01 8.10949430e-02 6.27761066e-01
-1.53425366e-01 1.87869295e-02 -2.30247006e-01 9.14979577e-01
-7.18152523e-02 -1.03613043e+00 -6.66801870e-01 2.87679762e-01
-3.72080892e-01 3.82936001e-01 -5.80051899e-01 7.06892312e-01
6.64034724e-01 9.74807084e-01 1.47274077e-01 -3.20730180e-01
3.13898355e-01 3.24824572e-01 5.91670394e-01 -6.99141622e-01
-6.82842731e-01 -2.37200081e-01 4.00399491e-02 -3.79403114e-01
-4.79759008e-01 -4.02902573e-01 -5.26350021e-01 -1.74888268e-01
-3.86356086e-01 -9.06366482e-02 8.49947333e-01 8.87647688e-01
2.43652493e-01 2.24249616e-01 7.75435507e-01 -8.71731877e-01
-8.84349287e-01 -1.17822242e+00 -3.72269601e-01 7.57202506e-01
1.97311044e-01 -1.15171647e+00 -5.79120636e-01 1.83226913e-02] | [13.239324569702148, 0.722039520740509] |
307e2273-3565-47a9-ab36-7dadf36c97f3 | latent-sdes-on-homogeneous-spaces | 2306.16248 | null | https://arxiv.org/abs/2306.16248v1 | https://arxiv.org/pdf/2306.16248v1.pdf | Latent SDEs on Homogeneous Spaces | We consider the problem of variational Bayesian inference in a latent variable model where a (possibly complex) observed stochastic process is governed by the solution of a latent stochastic differential equation (SDE). Motivated by the challenges that arise when trying to learn an (almost arbitrary) latent neural SDE from large-scale data, such as efficient gradient computation, we take a step back and study a specific subclass instead. In our case, the SDE evolves on a homogeneous latent space and is induced by stochastic dynamics of the corresponding (matrix) Lie group. In learning problems, SDEs on the unit $n$-sphere are arguably the most relevant incarnation of this setup. Notably, for variational inference, the sphere not only facilitates using a truly uninformative prior SDE, but we also obtain a particularly simple and intuitive expression for the Kullback-Leibler divergence between the approximate posterior and prior process in the evidence lower bound. Experiments demonstrate that a latent SDE of the proposed type can be learned efficiently by means of an existing one-step geometric Euler-Maruyama scheme. Despite restricting ourselves to a less diverse class of SDEs, we achieve competitive or even state-of-the-art performance on various time series interpolation and classification benchmarks. | ['Roland Kwitt', 'Florian Graf', 'Sebastian Zeng'] | 2023-06-28 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [ 2.45485738e-01 2.18741789e-01 1.19613044e-01 -2.71766156e-01
-9.17630076e-01 -4.93196189e-01 8.26738358e-01 -1.84011415e-01
-5.49432755e-01 8.58319759e-01 6.56863069e-03 -1.25062883e-01
-2.71051168e-01 -5.93753874e-01 -1.08429027e+00 -1.28524983e+00
-1.35584876e-01 6.56968117e-01 -9.00828317e-02 1.13781668e-01
1.94450408e-01 5.19438446e-01 -1.33195543e+00 -4.36787695e-01
8.47841501e-01 1.04611993e+00 8.92025828e-02 4.99174893e-01
9.97966975e-02 3.77518535e-01 -2.96691954e-01 -4.60361481e-01
2.09125355e-01 -3.47406864e-01 -6.53865099e-01 1.45584136e-01
2.79343694e-01 -3.28277200e-02 -3.08469385e-02 1.34429765e+00
2.46428892e-01 2.63273895e-01 1.27382624e+00 -1.08761454e+00
-3.60969752e-01 4.68721360e-01 -4.21851456e-01 1.03872046e-01
-9.93394852e-03 -1.55002372e-02 1.18997467e+00 -9.60343182e-01
8.96131694e-01 1.18449414e+00 5.88124871e-01 4.45575356e-01
-1.78223729e+00 -3.36820334e-01 1.78527236e-01 -6.66022748e-02
-1.51015258e+00 -3.58120471e-01 8.60488296e-01 -6.44728243e-01
3.74721497e-01 4.22499590e-02 4.37164307e-01 1.35966980e+00
2.23821923e-01 9.72140968e-01 9.85803843e-01 -2.27042720e-01
7.10303843e-01 1.18338577e-01 9.21313614e-02 8.55218291e-01
2.92315573e-01 -2.44267642e-01 -4.71309245e-01 -4.63516980e-01
6.45617545e-01 -1.92791801e-02 -3.63175809e-01 -7.44026661e-01
-1.08705389e+00 1.11743867e+00 4.97242995e-02 4.39283401e-02
-4.63841945e-01 2.89406359e-01 1.49740204e-01 1.34764001e-01
6.10549212e-01 7.03282431e-02 -2.12611184e-01 -1.98494226e-01
-1.06412387e+00 4.62349176e-01 1.02958977e+00 5.53485334e-01
6.99994802e-01 1.18259624e-01 -3.89823690e-02 4.67124075e-01
6.17401838e-01 4.78688151e-01 2.28206236e-02 -1.14298010e+00
3.14323813e-01 -1.43603474e-01 5.60778022e-01 -7.23138928e-01
9.39473733e-02 -5.89983225e-01 -1.04924476e+00 2.73602456e-01
8.47327530e-01 -2.96612769e-01 -5.13835490e-01 2.18130279e+00
4.35067594e-01 4.64303613e-01 -1.56464368e-01 7.26389945e-01
-1.76830977e-01 7.27667987e-01 -2.22704008e-01 -6.80896103e-01
9.54429924e-01 -4.71086651e-01 -5.34255207e-01 8.32037553e-02
2.28654653e-01 -3.60653073e-01 9.50980008e-01 5.72871387e-01
-1.32074165e+00 -1.12999499e-01 -1.14108729e+00 -3.60397734e-02
-4.56253737e-02 -1.69155598e-01 3.71303469e-01 4.18996930e-01
-9.53741550e-01 1.02073991e+00 -1.23142707e+00 -2.19365373e-01
1.47828743e-01 5.63603863e-02 -5.25685400e-02 1.10310353e-01
-9.89438534e-01 6.42541349e-01 -5.59304208e-02 3.27998608e-01
-1.09492743e+00 -7.28809595e-01 -5.22438884e-01 1.07427733e-02
5.02581298e-01 -7.12878466e-01 1.04181802e+00 -5.65672159e-01
-1.69111979e+00 5.70756316e-01 -2.86870509e-01 -5.11090279e-01
1.05569792e+00 -1.28003448e-01 6.07990138e-02 1.37942031e-01
4.04624641e-02 1.60455838e-01 1.32788444e+00 -9.88360286e-01
-1.24471918e-01 -4.71382141e-01 -4.50613983e-02 -1.22249611e-01
-7.31583089e-02 -2.32445002e-01 -1.13261744e-01 -5.79261959e-01
2.58313447e-01 -1.05291772e+00 -2.30603635e-01 3.43556732e-01
-1.82347715e-01 -3.92016560e-01 6.19092345e-01 -5.06695986e-01
9.91184890e-01 -2.08712077e+00 7.09739625e-01 1.95554599e-01
1.79661438e-01 -5.93420565e-02 3.06459993e-01 2.80709028e-01
5.14077395e-02 1.39574692e-01 -6.80285096e-01 -7.18885660e-01
3.83828700e-01 1.33651450e-01 -6.61665559e-01 9.88002956e-01
1.65332034e-02 5.67695260e-01 -8.68935168e-01 -3.80853474e-01
-7.49865621e-02 6.21429324e-01 -6.22553766e-01 -5.15212566e-02
-4.74976331e-01 4.79807049e-01 -6.61402643e-01 1.30724028e-01
4.63365018e-01 -5.72457671e-01 -1.28901199e-01 2.42717087e-01
-6.09165952e-02 1.31161600e-01 -1.65647948e+00 1.84797513e+00
-3.93213809e-01 5.84997475e-01 3.66564929e-01 -1.21984017e+00
5.47426999e-01 3.00688326e-01 3.97530496e-01 1.91441819e-01
-7.03814328e-02 3.73281896e-01 -2.72041917e-01 -1.67121142e-01
1.68945760e-01 -5.15968263e-01 8.37523863e-02 4.98181224e-01
7.00095147e-02 -4.83943596e-02 1.41754001e-01 1.09750390e-01
8.53250921e-01 3.10287058e-01 1.91044956e-01 -6.78403258e-01
4.48768556e-01 -5.24639845e-01 5.40047944e-01 9.19214427e-01
-4.11733277e-02 4.82005894e-01 7.19087064e-01 -4.11904603e-02
-1.01231992e+00 -1.56820703e+00 -5.23110747e-01 6.07403696e-01
-3.18373501e-01 -8.82128850e-02 -7.91313946e-01 -3.24771732e-01
6.93334714e-02 7.53666103e-01 -4.97517586e-01 -1.30201252e-02
-4.19558018e-01 -9.71776664e-01 3.26588243e-01 1.62737995e-01
3.78562808e-01 -6.04090273e-01 -6.01642251e-01 3.94601405e-01
-9.89566594e-02 -8.30488086e-01 -4.83239353e-01 1.01395063e-01
-9.42343473e-01 -6.10221207e-01 -9.83049452e-01 -2.27329358e-01
4.63912904e-01 -3.14328045e-01 9.21903431e-01 -6.50693595e-01
-1.74324185e-01 4.79882956e-01 2.87785143e-01 -2.27504104e-01
-2.95706302e-01 -2.11745858e-01 2.76114136e-01 4.67683583e-01
1.59265280e-01 -8.92243147e-01 -6.12967908e-01 3.97844799e-02
-8.79174292e-01 -1.66708872e-01 1.71955466e-01 6.94623828e-01
7.29809165e-01 2.31091515e-03 3.19552600e-01 -4.87047523e-01
4.31756467e-01 -6.52554750e-01 -1.20645356e+00 1.54262334e-01
-6.69746578e-01 6.04700327e-01 5.61167657e-01 -5.40880859e-01
-1.13880968e+00 -4.55034375e-02 3.61019112e-02 -4.89819616e-01
5.00685237e-02 6.99277520e-01 -6.44320101e-02 2.19103262e-01
3.81588489e-01 2.93602705e-01 1.33034564e-03 -6.61004841e-01
3.34991515e-01 4.13386017e-01 4.41932082e-01 -9.31257606e-01
7.24187553e-01 8.99636686e-01 5.31026661e-01 -9.91723716e-01
-9.62075889e-01 -2.35896230e-01 -4.98566270e-01 3.57698686e-02
9.49280202e-01 -6.96568847e-01 -8.59881163e-01 4.42685008e-01
-1.16417766e+00 -3.51440966e-01 -4.94864941e-01 6.53680861e-01
-9.25455987e-01 4.72214282e-01 -7.31267869e-01 -1.16523266e+00
8.12561344e-03 -1.15376484e+00 1.05709052e+00 -1.87070325e-01
-2.33145177e-01 -1.24963808e+00 2.32700109e-01 -6.39256239e-02
1.88309103e-01 4.95627165e-01 1.01381004e+00 -2.20540300e-01
-8.23480308e-01 -1.62611842e-01 1.20811127e-01 4.85873252e-01
-2.47845501e-01 1.58770278e-01 -7.91691720e-01 -1.22378938e-01
6.40697181e-01 -5.40279448e-02 9.74156678e-01 6.43490076e-01
9.79121387e-01 -2.53567874e-01 -1.53317615e-01 6.80451393e-01
1.34744966e+00 -2.77687252e-01 5.61824106e-02 -1.98294625e-01
4.47517544e-01 5.22693813e-01 3.29700798e-01 5.74932396e-01
2.21350312e-01 6.81792200e-01 2.66785622e-01 6.00077450e-01
4.51994240e-01 -2.25632608e-01 6.27115846e-01 8.74901235e-01
-2.16493532e-01 -2.37313453e-02 -9.46918607e-01 3.86520952e-01
-1.96083796e+00 -9.85810757e-01 -1.58777237e-01 2.48500586e+00
1.03875589e+00 2.39171967e-01 -2.58323587e-02 -3.00385188e-02
5.34130335e-01 3.27761620e-01 -8.29074800e-01 7.98190013e-02
-1.34537563e-01 1.11901350e-01 3.41594338e-01 8.10112298e-01
-9.11452472e-01 4.67323065e-01 5.91202593e+00 9.38675880e-01
-1.17073953e+00 4.09285963e-01 3.73216480e-01 -1.73019022e-01
-3.52554142e-01 -2.50128448e-01 -1.02865100e+00 6.24284029e-01
1.14555275e+00 -1.91676974e-01 5.02545238e-01 7.40085125e-01
3.09768200e-01 -6.31983727e-02 -1.34888053e+00 9.56180274e-01
-2.73807913e-01 -1.39073503e+00 -1.57173961e-01 5.24011552e-01
9.00559187e-01 2.14215696e-01 3.12916309e-01 -7.86088686e-03
3.39669704e-01 -7.85072148e-01 9.10212815e-01 8.13461542e-01
6.20904505e-01 -6.10898793e-01 2.68298000e-01 6.46467924e-01
-8.99852216e-01 1.62213787e-01 -4.05614913e-01 -5.81940413e-02
2.57225811e-01 8.55692387e-01 -2.53244251e-01 8.45328495e-02
3.38238478e-01 6.16186440e-01 5.03849685e-02 7.54756391e-01
-2.09871545e-01 7.67915547e-01 -7.22172976e-01 -4.06229384e-02
3.10870796e-01 -8.75603020e-01 9.86641824e-01 9.85216022e-01
5.42274177e-01 4.28535193e-02 -3.24515820e-01 1.21907973e+00
-1.65120304e-01 -1.07234545e-01 -4.87851888e-01 -2.26214498e-01
7.54221827e-02 9.59006727e-01 -7.30548143e-01 -4.01756406e-01
-2.62841374e-01 8.91565979e-01 4.00373608e-01 7.50867963e-01
-8.19955170e-01 5.04358709e-02 8.90154839e-01 -1.73624270e-02
6.57107353e-01 -5.00763535e-01 -2.18898252e-01 -1.56287289e+00
3.85423869e-01 -3.91858608e-01 2.37956867e-01 -6.09924257e-01
-1.37456214e+00 2.27144718e-01 3.04228663e-01 -8.17595005e-01
-7.58342206e-01 -6.64921224e-01 -4.72874016e-01 1.09584522e+00
-1.28342545e+00 -4.48117018e-01 2.43445739e-01 7.37088442e-01
3.40526670e-01 1.50014684e-01 7.17405021e-01 -1.86599549e-02
-3.56853336e-01 9.79586318e-02 8.82819712e-01 -3.70441020e-01
2.60229766e-01 -1.55722857e+00 1.68774426e-01 9.26455617e-01
3.12718183e-01 7.57328689e-01 1.27160215e+00 -5.73617995e-01
-1.52590930e+00 -8.80644917e-01 7.08111167e-01 -5.82753956e-01
1.11440897e+00 -6.98833764e-01 -9.28903222e-01 8.14374208e-01
-2.71037608e-01 1.65218368e-01 3.82746667e-01 1.31343335e-01
-1.90833926e-01 6.23401301e-03 -9.58727121e-01 6.89506948e-01
9.74981606e-01 -6.84664488e-01 -5.11818290e-01 3.45442474e-01
4.01912391e-01 -1.02926806e-01 -6.91393077e-01 1.36858180e-01
5.65081954e-01 -5.98206341e-01 9.52569723e-01 -6.20430946e-01
4.45844203e-01 -2.32313812e-01 -3.63408417e-01 -1.10430264e+00
2.30765998e-01 -1.03510320e+00 -5.00803590e-01 9.55714285e-01
1.56379461e-01 -4.86195773e-01 8.71602178e-01 6.56231344e-01
2.43035287e-01 -8.38769257e-01 -1.15827274e+00 -8.14931750e-01
3.35775435e-01 -5.95660329e-01 1.48320422e-01 7.26519465e-01
-3.73497635e-01 2.45170012e-01 -3.62047464e-01 3.07259381e-01
1.24105406e+00 5.69551550e-02 3.78615677e-01 -1.44463754e+00
-7.60998845e-01 -3.83322388e-01 -6.00741245e-02 -1.51113105e+00
3.98294598e-01 -9.25196648e-01 2.60052711e-01 -1.09721732e+00
1.23279117e-01 -7.15575963e-02 -2.95823842e-01 -3.64430755e-01
1.98615286e-02 -6.52995780e-02 6.19597770e-02 2.30899543e-01
-4.61590022e-01 7.95068741e-01 9.79020238e-01 1.22379325e-01
4.92988415e-02 3.48902404e-01 -2.66685843e-01 1.09519124e+00
3.19750994e-01 -6.02173805e-01 -3.91101927e-01 -2.16768831e-01
4.37990755e-01 4.26392555e-01 5.75176179e-01 -6.01265132e-01
3.28847617e-01 -1.27836868e-01 -1.19559459e-01 -4.39929634e-01
6.75856948e-01 -4.82106447e-01 2.09006965e-01 3.52811843e-01
-3.58883351e-01 -3.96610051e-01 -4.31634486e-01 8.77227902e-01
-6.00701198e-02 -6.06519163e-01 8.76625597e-01 -6.31841645e-02
-6.21156096e-02 4.47306246e-01 -4.45515662e-01 1.47251546e-01
5.47267854e-01 7.46904388e-02 1.49642423e-01 -4.53554302e-01
-1.08218074e+00 -1.23069838e-01 3.19316804e-01 -1.86781183e-01
2.78144926e-01 -1.18286729e+00 -5.73689580e-01 4.07788437e-03
-2.52705246e-01 1.03204049e-01 2.53057294e-03 1.00407374e+00
-8.33840594e-02 3.90882611e-01 1.27928108e-01 -7.02178061e-01
-7.57220507e-01 3.67625982e-01 3.64920080e-01 -4.08123136e-01
-5.47860324e-01 9.28868651e-01 4.15664613e-01 -1.02903709e-01
3.72595221e-01 -4.52936739e-01 3.38746130e-01 3.00046176e-01
3.20790321e-01 5.94976485e-01 -2.58183092e-01 -5.43485522e-01
-1.13066524e-01 3.74008447e-01 1.73662677e-01 -6.37200296e-01
1.37517035e+00 -2.74609536e-01 -9.56656337e-02 1.12674522e+00
1.50554502e+00 1.97174642e-02 -1.87870634e+00 -4.82294112e-01
1.10223614e-01 -5.55388294e-02 3.46954092e-02 -2.08606899e-01
-5.72128892e-01 1.16616690e+00 4.23446089e-01 3.27907890e-01
5.88680446e-01 2.24688247e-01 4.33132976e-01 4.78250146e-01
4.24847603e-01 -9.35703933e-01 -8.11414868e-02 4.31238323e-01
9.19503212e-01 -1.07609236e+00 -1.07021153e-01 -4.21181619e-02
-1.21846057e-01 1.01039958e+00 -2.92248577e-01 -3.93628985e-01
1.03861141e+00 -6.57774284e-02 -4.72828180e-01 -7.20249861e-02
-7.22760856e-01 2.40568683e-01 2.77980834e-01 8.21238980e-02
1.92140117e-01 -3.20559442e-02 -2.75252521e-01 3.57168555e-01
-1.83744594e-01 -6.21838011e-02 5.69158912e-01 6.46593750e-01
-2.58319110e-01 -8.41616988e-01 -1.28607318e-01 2.61171281e-01
-7.27800131e-01 -4.74895947e-02 2.99106628e-01 6.15512013e-01
-1.83048829e-01 4.81594145e-01 -6.98299780e-02 5.39496362e-01
-1.85591906e-01 2.89346635e-01 4.27846998e-01 -4.53318745e-01
2.97485203e-01 3.16823334e-01 -3.47361237e-01 -5.89153051e-01
-4.50555146e-01 -1.40491533e+00 -8.86104167e-01 -1.38616517e-01
-5.19654527e-03 3.83427113e-01 9.87900853e-01 1.22525978e+00
-6.08690316e-03 7.83270001e-02 2.97172338e-01 -8.21175814e-01
-1.29247177e+00 -9.43863273e-01 -1.00855184e+00 2.65001506e-01
6.67400420e-01 -7.56548285e-01 -7.78253913e-01 1.18386365e-01] | [6.933762550354004, 3.840263605117798] |
8696dd92-f5fe-4b94-9666-11b0159853f2 | breadth-first-reasoning-graph-for-multi-hop | null | null | https://aclanthology.org/2021.naacl-main.464 | https://aclanthology.org/2021.naacl-main.464.pdf | Breadth First Reasoning Graph for Multi-hop Question Answering | Recently Graph Neural Network (GNN) has been used as a promising tool in multi-hop question answering task. However, the unnecessary updations and simple edge constructions prevent an accurate answer span extraction in a more direct and interpretable way. In this paper, we propose a novel model of Breadth First Reasoning Graph (BFR-Graph), which presents a new message passing way that better conforms to the reasoning process. In BFR-Graph, the reasoning message is required to start from the question node and pass to the next sentences node hop by hop until all the edges have been passed, which can effectively prevent each node from over-smoothing or being updated multiple times unnecessarily. To introduce more semantics, we also define the reasoning graph as a weighted graph with considering the number of co-occurrence entities and the distance between sentences. Then we present a more direct and interpretable way to aggregate scores from different levels of granularity based on the GNN. On HotpotQA leaderboard, the proposed BFR-Graph achieves state-of-the-art on answer span prediction. | ['Meng Yang', 'Yongjie Huang'] | 2021-06-01 | null | null | null | naacl-2021-4 | ['multi-hop-question-answering'] | ['knowledge-base'] | [-6.88839555e-02 6.36096418e-01 2.84501426e-02 -5.43137789e-01
-4.75808650e-01 -4.05890971e-01 2.80240268e-01 9.56362784e-01
-3.56686205e-01 7.16046691e-01 4.82191592e-01 -4.80487436e-01
-6.46046400e-01 -1.39104939e+00 -4.88607883e-01 -7.19935969e-02
-5.96381538e-02 6.94184601e-01 7.99771786e-01 -6.82708144e-01
2.18744323e-01 2.36056000e-01 -1.39884174e+00 6.51358306e-01
8.27686429e-01 8.17067504e-01 -1.54613838e-01 7.17888415e-01
-6.87158227e-01 1.47130418e+00 -6.88572228e-01 -9.17719901e-01
-6.39219731e-02 -5.62023938e-01 -1.47958815e+00 -4.53179806e-01
2.97251225e-01 -2.56128252e-01 -4.13942248e-01 9.56726789e-01
3.42452168e-01 2.46920913e-01 3.58747065e-01 -1.20993578e+00
-5.29333174e-01 1.15324259e+00 -3.45066160e-01 2.64445454e-01
7.37737358e-01 -2.23748922e-01 1.56884325e+00 -2.95182198e-01
7.81821847e-01 1.39164877e+00 5.31544924e-01 2.18159184e-01
-5.66215396e-01 -1.07138693e-01 4.27961320e-01 8.38939488e-01
-8.45530033e-01 1.52478367e-01 8.02471220e-01 1.87007599e-02
1.17762649e+00 6.72474325e-01 6.14370286e-01 4.58240509e-01
1.57279998e-01 7.28759229e-01 6.42927766e-01 -3.96702081e-01
7.49438033e-02 -2.41970912e-01 9.40021992e-01 1.03467059e+00
1.15067624e-01 -7.31345415e-01 -5.39419532e-01 -4.39128838e-02
2.21210029e-02 -2.38894373e-01 -3.11622262e-01 7.16310740e-02
-6.90313637e-01 9.31379259e-01 7.88592398e-01 4.61093634e-01
-3.72499913e-01 1.94730371e-01 6.17183745e-01 5.93029439e-01
2.83954710e-01 4.27411258e-01 -4.54153091e-01 6.95685446e-02
-7.33450055e-01 2.71343708e-01 1.05804718e+00 6.90018117e-01
6.41968131e-01 -4.47129101e-01 -7.61194825e-01 6.42602265e-01
1.89424068e-01 -6.85685724e-02 2.18548626e-01 -8.73937547e-01
6.53601885e-01 1.37346840e+00 -2.41309896e-01 -1.53891718e+00
-7.68527627e-01 -5.44745684e-01 -8.37648690e-01 -3.29298437e-01
4.17862743e-01 -1.92471705e-02 -4.46686089e-01 1.56113064e+00
5.18748760e-01 -2.19623551e-01 -7.52773955e-02 8.06451857e-01
1.25706422e+00 6.94588184e-01 -6.56594411e-02 -2.05941439e-01
1.67121434e+00 -1.32359457e+00 -7.74354577e-01 -1.26580626e-01
9.95778441e-01 -2.70366490e-01 9.50059116e-01 3.24602216e-01
-1.02376461e+00 -3.99485767e-01 -9.56461608e-01 -5.54825902e-01
-5.31593502e-01 -3.62062901e-01 5.90726733e-01 3.05529088e-01
-1.08098102e+00 6.25346124e-01 -2.53277302e-01 -3.46170247e-01
3.05696633e-02 2.58096606e-02 -2.17530504e-01 -1.52465165e-01
-1.68942702e+00 1.07185757e+00 6.38121128e-01 2.26120129e-01
-1.84551656e-01 -3.35620493e-01 -7.94776380e-01 3.98153543e-01
7.49438286e-01 -9.51993883e-01 1.14421356e+00 -3.31186861e-01
-1.10178316e+00 6.50947094e-01 -1.13741152e-01 -8.08193266e-01
4.02681053e-01 -1.42248765e-01 -5.05462289e-01 3.49772841e-01
-1.58275142e-01 2.94171870e-01 4.66572493e-01 -7.75842190e-01
-6.12309515e-01 -3.76079142e-01 7.25124538e-01 2.83284515e-01
-4.23034519e-01 -8.42478424e-02 -5.19592106e-01 -8.84655416e-02
2.04828262e-01 -4.76600945e-01 3.59877944e-02 -3.03062379e-01
-4.94340032e-01 -9.01058137e-01 5.55717826e-01 -9.67743635e-01
2.05409932e+00 -1.53781831e+00 2.94281274e-01 1.47892490e-01
8.55112612e-01 1.59330860e-01 -1.87104419e-01 7.64475465e-01
1.77444533e-01 1.15825668e-01 -1.30073428e-01 1.83474198e-02
8.67110267e-02 2.41959929e-01 -8.88732821e-02 8.46595913e-02
-1.18011022e-02 1.03799272e+00 -9.79471385e-01 -8.32186222e-01
-1.89327866e-01 -1.30320594e-01 -5.30888438e-01 1.32397085e-01
-5.98900139e-01 -5.77598326e-02 -4.44834620e-01 3.75900328e-01
4.69013005e-01 -4.79262054e-01 1.51043385e-01 -3.93380284e-01
2.26756111e-01 4.38316971e-01 -1.13441014e+00 1.44654655e+00
-4.51067537e-01 4.38533783e-01 -8.34995732e-02 -1.01233101e+00
1.04730034e+00 6.30393298e-03 1.59060702e-01 -8.77122104e-01
9.54541564e-02 1.20893829e-02 1.53696816e-02 -8.87325287e-01
7.16122866e-01 3.05669963e-01 -2.27776200e-01 3.93316716e-01
-4.98904921e-02 2.00499281e-01 8.56046379e-01 6.64594173e-01
1.45427763e+00 -1.57902688e-01 2.37575844e-01 -1.31936418e-02
1.13478315e+00 4.07563262e-02 3.63387913e-01 8.30939353e-01
5.28673008e-02 1.43936515e-01 8.65911305e-01 -6.56642139e-01
-6.92559719e-01 -6.84561253e-01 4.06198740e-01 1.21423733e+00
1.59489632e-01 -7.81009853e-01 -8.95827532e-01 -1.02926183e+00
-2.87035201e-02 8.94031525e-01 -5.32320380e-01 -3.50690544e-01
-8.85486960e-01 -3.40367675e-01 7.04241335e-01 2.42334753e-01
7.29214907e-01 -1.11152804e+00 -4.13323283e-01 3.95351022e-01
-7.45205283e-01 -9.75318789e-01 -2.24851266e-01 -2.12079346e-01
-7.08291411e-01 -1.10851693e+00 -2.35745028e-01 -6.19693160e-01
5.13730705e-01 2.29139291e-02 1.44132423e+00 7.84657121e-01
3.60283884e-03 9.59282443e-02 -7.81109035e-01 -1.33541107e-01
-3.35384607e-01 4.66906011e-01 -3.77408355e-01 1.38856973e-02
3.27964157e-01 -4.82584596e-01 -6.75977647e-01 1.00765504e-01
-8.11270714e-01 1.37431920e-01 3.21032912e-01 5.36822557e-01
3.97431314e-01 1.52757093e-01 7.73294806e-01 -1.08768988e+00
1.22392833e+00 -5.37684262e-01 -3.69702995e-01 7.80969739e-01
-6.21129453e-01 5.12568653e-01 9.05651689e-01 1.32031888e-01
-9.09309566e-01 -6.83060467e-01 -4.47697937e-01 1.40241742e-01
2.04721749e-01 7.98928857e-01 1.00225829e-01 2.13965952e-01
6.51977420e-01 -9.36518088e-02 -3.78003478e-01 -4.04290020e-01
6.99502289e-01 5.38396895e-01 2.45574296e-01 -5.26273489e-01
5.86397767e-01 8.13638344e-02 2.02980459e-01 -4.73295182e-01
-1.14534152e+00 -4.05244797e-01 -3.79974455e-01 -4.41006064e-01
9.62114811e-01 -2.23496482e-01 -1.32155895e+00 2.72182077e-01
-1.40824306e+00 8.30604061e-02 -1.49531409e-01 1.18637281e-02
-9.44817811e-02 6.50577486e-01 -8.68654549e-01 -8.45043063e-01
-7.20075965e-01 -6.78758025e-01 6.32715285e-01 3.57629687e-01
-3.64529073e-01 -9.29182708e-01 -1.74145307e-02 7.56527424e-01
3.84547293e-01 2.78721511e-01 1.40712583e+00 -1.06509960e+00
-6.02387190e-01 -1.99926928e-01 -4.88879114e-01 1.63712755e-01
-2.89769918e-01 -1.46052331e-01 -4.13390905e-01 6.62852824e-02
2.71375757e-02 -1.58608571e-01 8.41735125e-01 -3.69225256e-02
1.16925442e+00 -6.01462543e-01 -5.49865104e-02 8.99414420e-02
1.40749979e+00 -1.54049143e-01 5.25918722e-01 3.86873543e-01
6.05681658e-01 8.71544003e-01 3.63549381e-01 1.26797497e-01
9.05888915e-01 2.95217603e-01 6.50709748e-01 2.37949669e-01
-1.06078148e-01 -4.81117278e-01 -9.00483653e-02 1.34500182e+00
-2.78613288e-02 -7.80357897e-01 -8.13738763e-01 4.28715736e-01
-1.99357975e+00 -8.52394402e-01 -7.51818001e-01 1.87626290e+00
5.88695586e-01 4.66957808e-01 6.39605429e-03 3.91386539e-01
6.58232093e-01 4.75070536e-01 -2.68395066e-01 -8.57829452e-01
-4.76943655e-03 1.59879789e-01 2.30209559e-01 8.09268057e-01
-4.11742479e-01 7.02511370e-01 4.98583412e+00 9.15813267e-01
-3.89987528e-01 3.60016376e-02 3.32259923e-01 2.42067561e-01
-6.88926578e-01 1.53373271e-01 -7.30464876e-01 2.90758312e-01
8.24994445e-01 -2.84748703e-01 3.87823731e-01 4.72364128e-01
-9.99954417e-02 -2.98728764e-01 -1.02218878e+00 6.68198228e-01
5.97366057e-02 -1.45134926e+00 2.92124063e-01 -5.51464081e-01
2.44714633e-01 -4.05796617e-01 -5.61409831e-01 5.93970418e-01
3.09202015e-01 -7.91166842e-01 4.40909445e-01 8.49682331e-01
7.22428039e-02 -6.95569515e-01 9.90684807e-01 7.88443148e-01
-1.38378847e+00 -2.57414013e-01 -2.50022709e-01 -3.03384393e-01
4.86043274e-01 7.68001974e-01 -6.68276310e-01 1.25423312e+00
6.47095680e-01 1.30880982e-01 -8.78052235e-01 7.69853115e-01
-4.65235621e-01 4.28979576e-01 -2.25235224e-01 -7.04690278e-01
3.70040178e-01 -1.13620378e-01 4.78205591e-01 1.08740306e+00
2.97390819e-01 2.64917135e-01 -2.72243116e-02 5.02522945e-01
-4.43891495e-01 3.68046582e-01 -2.81022996e-01 8.33596289e-02
4.46986169e-01 1.30856919e+00 -6.91046774e-01 -3.91661763e-01
-4.18552876e-01 8.34818959e-01 8.63163054e-01 6.90234080e-02
-9.16059971e-01 -7.09902287e-01 -8.75310302e-02 2.00401515e-01
-7.72207528e-02 -5.27264841e-04 -7.22486351e-04 -7.89955616e-01
3.60691071e-01 -6.84979081e-01 8.99387419e-01 -9.36090529e-01
-1.32705629e+00 8.83178771e-01 -6.11402206e-02 -6.70197368e-01
-6.90612569e-02 -3.31136346e-01 -6.33167207e-01 6.97965860e-01
-1.48830783e+00 -9.17048812e-01 -4.38083053e-01 4.82296258e-01
3.23578924e-01 2.13577315e-01 7.62086272e-01 3.01614255e-01
-4.53804582e-01 5.54060519e-01 -4.14107978e-01 2.06345171e-01
2.96920598e-01 -1.21837127e+00 3.09609979e-01 7.07157075e-01
2.58992046e-01 4.76081491e-01 7.55083621e-01 -4.84384716e-01
-1.12066567e+00 -8.04708421e-01 1.39242649e+00 -3.56660604e-01
6.93976104e-01 7.16647953e-02 -1.11749947e+00 5.35289288e-01
3.67370695e-01 -3.83287877e-01 2.60270834e-01 3.74205649e-01
-3.31784368e-01 -4.16928947e-01 -1.02950847e+00 6.08866811e-01
1.09820306e+00 -4.60464835e-01 -9.96912479e-01 4.50648904e-01
1.25361657e+00 -3.35078686e-01 -8.57819378e-01 5.15336096e-01
1.03396431e-01 -1.22991192e+00 6.80333197e-01 -7.27576911e-01
5.09887755e-01 -3.75537664e-01 1.10920392e-01 -9.78905439e-01
-2.99536794e-01 -5.40315092e-01 -6.49392784e-01 1.19751561e+00
7.18799531e-01 -5.86155117e-01 7.42104948e-01 3.54352176e-01
-8.79730582e-02 -1.23211169e+00 -8.99397671e-01 -4.12941337e-01
-2.67308563e-01 -3.79814923e-01 7.56197512e-01 5.94297111e-01
2.68280298e-01 7.64810026e-01 -2.48428673e-01 1.15800500e-01
4.17000532e-01 2.44525358e-01 3.82005483e-01 -1.42696190e+00
-2.91818529e-01 -4.24463063e-01 -2.34194010e-01 -1.20587754e+00
-5.46320081e-02 -1.05889392e+00 -2.72763282e-01 -2.33326483e+00
-3.66765223e-02 -2.31322758e-02 -1.82554424e-01 1.52868167e-01
-1.96069166e-01 -2.79999256e-01 1.27996415e-01 -2.46238008e-01
-1.10374451e+00 1.97577372e-01 1.56001306e+00 -2.29998246e-01
3.94815877e-02 -4.07000370e-02 -5.60247540e-01 7.94759631e-01
8.17289054e-01 -5.02668083e-01 -6.26071751e-01 -6.20334804e-01
1.10290587e+00 4.74913657e-01 4.80907522e-02 -7.87629306e-01
7.73550093e-01 5.26027344e-02 -1.48291886e-01 -8.20038140e-01
1.51745468e-01 -8.12626481e-01 -8.63905996e-03 4.17866796e-01
-5.43631732e-01 4.11932826e-01 -2.82662421e-01 4.67410475e-01
-3.13467234e-01 -6.33639812e-01 2.66452372e-01 -2.52252430e-01
-5.84967554e-01 2.61389881e-01 -9.28328335e-02 3.51671904e-01
7.88620174e-01 -1.86209269e-02 -6.54972553e-01 -6.16039753e-01
-8.10406208e-01 7.64248252e-01 -2.13501737e-01 4.29880351e-01
5.15288830e-01 -1.26550496e+00 -7.04360247e-01 -3.29310119e-01
8.63844380e-02 9.40689296e-02 6.76020622e-01 7.65251040e-01
-7.47178674e-01 4.97653484e-01 1.19173259e-01 -1.30520314e-01
-1.32607114e+00 6.05180681e-01 2.68124819e-01 -1.06608522e+00
-6.33009911e-01 8.86685193e-01 -5.78404069e-01 -4.91000026e-01
3.25809628e-01 -5.18605292e-01 -7.31468618e-01 3.10979664e-01
3.82930458e-01 7.72745371e-01 2.62135625e-01 -2.77420402e-01
-1.40714422e-01 4.20887887e-01 -7.82978609e-02 1.77471280e-01
1.03305304e+00 -2.42250621e-01 -7.61563480e-01 3.08963805e-01
1.01630557e+00 -3.41453068e-02 -4.48901564e-01 -3.32956761e-01
3.89424294e-01 3.83402333e-02 -3.61881673e-01 -6.54521227e-01
-7.38912404e-01 7.67602682e-01 -1.41621545e-01 9.74857926e-01
1.05785418e+00 1.17766015e-01 1.25506139e+00 7.78512776e-01
3.31567198e-01 -1.25098073e+00 8.21984559e-02 7.64870644e-01
9.49842811e-01 -7.79124022e-01 2.37449771e-03 -3.52159649e-01
-3.35615516e-01 1.30814552e+00 6.49198234e-01 1.82179362e-01
4.14861947e-01 -1.43930092e-01 -2.30688125e-01 -5.54920018e-01
-8.10636699e-01 -1.83078051e-01 3.31604213e-01 1.10379234e-01
1.64217860e-01 -5.12769595e-02 -8.98823977e-01 7.31236637e-01
-4.30433929e-01 -1.23205684e-01 4.26955491e-01 5.95537722e-01
-9.28136885e-01 -1.00831211e+00 -6.04243530e-03 6.01417661e-01
-2.51519859e-01 -5.47373593e-02 -4.97801751e-01 6.71254754e-01
4.55611125e-02 1.20731390e+00 -1.10679641e-01 -5.09260356e-01
5.79509676e-01 3.75907511e-01 4.00739104e-01 -3.46076131e-01
-7.30003357e-01 -8.66351962e-01 7.10957110e-01 -4.75101054e-01
-3.21043819e-01 1.18900202e-02 -1.60305512e+00 -6.37781441e-01
-3.62627059e-01 4.61684942e-01 2.73894191e-01 1.06184971e+00
2.40701109e-01 8.68664682e-01 3.47838461e-01 2.55785882e-01
-4.66547102e-01 -1.09025729e+00 -2.15057448e-01 4.64738816e-01
1.65418968e-01 -1.73319474e-01 -3.49625766e-01 -4.59635645e-01] | [10.827146530151367, 7.940967559814453] |
dfa08439-6388-4123-a1b7-cd0986302e41 | simnet-enabling-robust-unknown-object | 2106.16118 | null | https://arxiv.org/abs/2106.16118v1 | https://arxiv.org/pdf/2106.16118v1.pdf | SimNet: Enabling Robust Unknown Object Manipulation from Pure Synthetic Data via Stereo | Robot manipulation of unknown objects in unstructured environments is a challenging problem due to the variety of shapes, materials, arrangements and lighting conditions. Even with large-scale real-world data collection, robust perception and manipulation of transparent and reflective objects across various lighting conditions remain challenging. To address these challenges we propose an approach to performing sim-to-real transfer of robotic perception. The underlying model, SimNet, is trained as a single multi-headed neural network using simulated stereo data as input and simulated object segmentation masks, 3D oriented bounding boxes (OBBs), object keypoints, and disparity as output. A key component of SimNet is the incorporation of a learned stereo sub-network that predicts disparity. SimNet is evaluated on 2D car detection, unknown object detection, and deformable object keypoint detection and significantly outperforms a baseline that uses a structured light RGB-D sensor. By inferring grasp positions using the OBB and keypoint predictions, SimNet can be used to perform end-to-end manipulation of unknown objects in both easy and hard scenarios using our fleet of Toyota HSR robots in four home environments. In unknown object grasping experiments, the predictions from the baseline RGB-D network and SimNet enable successful grasps of most of the easy objects. However, the RGB-D baseline only grasps 35% of the hard (e.g., transparent) objects, while SimNet grasps 95%, suggesting that SimNet can enable robust manipulation of unknown objects, including transparent objects, in unknown environments. | ['Mark Tjersland', 'Brijen Thananjeyan', 'Kevin Stone', 'Michael Laskey', 'Thomas Kollar'] | 2021-06-30 | null | null | null | null | ['transparent-objects'] | ['computer-vision'] | [ 1.41883284e-01 1.13903411e-01 4.04375970e-01 -2.58847624e-01
-3.99893433e-01 -9.23510551e-01 -2.69604009e-02 -2.62184471e-01
-4.99397576e-01 2.95325309e-01 -4.14326489e-01 -1.55122414e-01
8.83507356e-02 -5.39413452e-01 -1.26051199e+00 -5.11798561e-01
-3.51928115e-01 9.90307033e-01 6.64997280e-01 -4.75223988e-01
5.20070158e-02 8.69338989e-01 -1.70854068e+00 4.02188569e-01
4.14787084e-01 1.27469301e+00 1.01922870e+00 8.58087420e-01
3.30055475e-01 2.40158528e-01 -4.26123410e-01 -6.50431844e-04
1.03188264e+00 5.85760057e-01 -4.21705872e-01 -5.09128114e-03
5.61767817e-01 -1.15147042e+00 -3.20063710e-01 7.33613849e-01
4.67426509e-01 2.81459033e-01 7.04273105e-01 -1.43350899e+00
-7.07990348e-01 3.29311669e-01 -5.27587414e-01 -4.15115595e-01
6.30437732e-01 8.67084384e-01 5.92165351e-01 -8.90091896e-01
5.23900151e-01 1.85371971e+00 4.98592436e-01 7.57370949e-01
-1.26054227e+00 -5.93374968e-01 3.39484572e-01 -2.14279100e-01
-8.91426206e-01 -8.81289616e-02 3.35958451e-01 -3.74085128e-01
1.21425545e+00 -7.52307177e-02 6.13862336e-01 1.26832068e+00
1.92487687e-01 8.34905744e-01 9.86945033e-01 -3.87166068e-02
1.98599428e-01 -5.01855277e-02 -1.92498609e-01 7.78077066e-01
4.26249862e-01 3.91241223e-01 -6.04009442e-02 -2.46475693e-02
1.15278327e+00 2.30084524e-01 -2.99173981e-01 -9.18365300e-01
-1.35946083e+00 2.94150919e-01 8.97301674e-01 -4.24031377e-01
-4.31799859e-01 3.79590362e-01 2.81014014e-02 5.09384237e-02
-8.77063423e-02 5.11073172e-01 -6.95093632e-01 1.17430605e-01
1.02258407e-01 7.15119481e-01 9.88150775e-01 1.68273735e+00
6.87480509e-01 -7.31937438e-02 1.41293988e-01 6.49809003e-01
3.48744810e-01 1.16954255e+00 -1.76693752e-01 -1.44013000e+00
7.52237201e-01 5.97727835e-01 6.91793740e-01 -6.90957665e-01
-5.98920405e-01 2.81888664e-01 -1.47190958e-01 1.03606331e+00
6.18576825e-01 -2.30222672e-01 -1.49975443e+00 1.31447911e+00
4.29623038e-01 -6.21253669e-01 4.43436317e-02 1.39599431e+00
8.64072442e-01 4.67773378e-01 -1.34560704e-01 7.59061635e-01
1.05813408e+00 -9.00374770e-01 -1.19577259e-01 -6.75691247e-01
1.47346612e-02 -5.55986047e-01 1.17519808e+00 3.63464594e-01
-1.24228370e+00 -4.53489512e-01 -8.11585724e-01 -2.69493252e-01
-5.30143082e-01 1.32907629e-01 7.70329833e-01 8.24323520e-02
-1.03348935e+00 5.20161331e-01 -1.04529917e+00 -4.13693368e-01
4.68991011e-01 7.18829691e-01 -5.43245494e-01 -6.05959773e-01
-3.73246908e-01 1.18433130e+00 5.06920636e-01 4.50479716e-01
-1.44582844e+00 -4.20899600e-01 -8.47564697e-01 -7.30583593e-02
6.12470984e-01 -6.15465701e-01 1.30486166e+00 -3.43705654e-01
-1.41638052e+00 9.04018819e-01 3.36826384e-01 -1.69386178e-01
7.30053723e-01 -5.85852861e-01 4.71843243e-01 3.26600015e-01
1.43452792e-03 1.27148771e+00 9.23914373e-01 -2.06271410e+00
-6.26574516e-01 -5.69936156e-01 3.32485199e-01 4.38467950e-01
3.23205709e-01 -3.02862555e-01 -3.57722938e-01 -2.40314588e-01
3.72477472e-01 -1.13022852e+00 -1.65423781e-01 7.11149335e-01
-3.27977061e-01 -4.23223972e-02 1.09526229e+00 -5.40062249e-01
-3.93104374e-01 -1.98728585e+00 3.50193530e-01 4.19856533e-02
-5.73475063e-02 -1.39583081e-01 -2.87767977e-01 1.45124808e-01
3.36341530e-01 -3.95935118e-01 -3.01112514e-02 -3.29796910e-01
3.70943338e-01 3.43802065e-01 -4.76065129e-01 3.41223508e-01
1.43569365e-01 9.69652593e-01 -8.37958694e-01 -1.12792820e-01
4.95476991e-01 2.95451134e-01 -4.83784884e-01 6.21852696e-01
-7.03112066e-01 4.05815721e-01 -3.54656518e-01 1.30604923e+00
8.74194980e-01 2.23394811e-01 -1.39794320e-01 -2.23624110e-01
1.11972159e-02 -1.83849901e-01 -1.06252861e+00 1.69632041e+00
-3.88760477e-01 3.35219353e-01 9.09591496e-01 -5.00036240e-01
1.01104546e+00 -7.53487423e-02 3.42380196e-01 -3.97406995e-01
8.52739736e-02 4.56718534e-01 -8.43733177e-02 -7.59961128e-01
5.06314874e-01 1.34022817e-01 2.15764102e-02 1.43070191e-01
-5.77973090e-02 -1.03179681e+00 -1.88442811e-01 -8.43164399e-02
1.22101116e+00 6.42137170e-01 -4.43453163e-01 7.87089244e-02
-5.51444530e-01 4.00091588e-01 1.08002111e-01 9.33296204e-01
-1.54537484e-01 8.69423330e-01 -3.05901300e-02 -4.81696695e-01
-1.19234467e+00 -1.63972723e+00 9.65493321e-02 1.22424352e+00
7.65410066e-01 6.51686966e-01 -4.73011553e-01 -3.28910589e-01
8.46990705e-01 5.59665918e-01 -3.36994857e-01 -1.44396373e-03
-8.70184600e-01 -2.74276167e-01 1.43247964e-02 7.86823452e-01
4.13388073e-01 -1.51148283e+00 -1.31084812e+00 1.97351754e-01
5.81405088e-02 -1.47358620e+00 1.17889680e-02 5.39778650e-01
-7.94048727e-01 -1.17422843e+00 -8.04576457e-01 -8.65440190e-01
8.74989629e-01 7.89838314e-01 8.77521038e-01 -1.38102144e-01
-7.63253212e-01 7.38072395e-01 -5.63566327e-01 -6.98079705e-01
-3.00034702e-01 -2.78054982e-01 1.60033345e-01 -8.79133165e-01
-3.06637329e-03 -2.36648366e-01 -7.22298861e-01 6.12309992e-01
-5.67428589e-01 -3.39869969e-02 7.58869171e-01 3.93528283e-01
2.48284444e-01 -3.79062504e-01 2.01212540e-02 1.47183329e-01
4.02191401e-01 -2.67508596e-01 -8.19729507e-01 1.08800836e-01
1.69324294e-01 -3.28986377e-01 3.34157199e-01 -7.37142444e-01
-9.60976720e-01 2.60338306e-01 2.83252776e-01 -7.15429366e-01
-5.38319170e-01 -2.73478776e-01 4.93061543e-03 -2.23478168e-01
6.30008519e-01 -2.60609895e-01 7.69304782e-02 -4.19881046e-01
2.92622030e-01 6.76722944e-01 7.65753925e-01 -1.00376666e+00
8.20072353e-01 8.42038393e-01 -1.74812123e-01 -6.05261862e-01
-3.50385427e-01 -3.22296262e-01 -7.48824418e-01 -3.75062495e-01
9.56885993e-01 -9.43953693e-01 -1.37680352e+00 6.78367436e-01
-1.40656030e+00 -1.12350070e+00 -2.04860657e-01 4.94553208e-01
-8.43365192e-01 -1.18428670e-01 -5.92402756e-01 -1.00036740e+00
-1.44050375e-01 -1.45504940e+00 1.79378176e+00 1.27006158e-01
7.21591804e-03 -1.73529863e-01 -8.11817229e-01 4.06341970e-01
2.44240150e-01 3.56407374e-01 8.69992673e-01 -2.16900408e-01
-1.15162730e+00 -2.51988381e-01 -5.38644373e-01 1.01608761e-01
2.04839453e-01 -2.93452054e-01 -8.66894424e-01 -7.02201903e-01
-2.95782030e-01 -8.18961442e-01 8.10358584e-01 4.09344167e-01
1.11895990e+00 2.16749720e-02 -6.20099068e-01 4.89240408e-01
1.18131804e+00 3.94040704e-01 1.69714585e-01 2.09287986e-01
7.01631308e-01 5.90267062e-01 9.62431908e-01 2.79372841e-01
4.20102388e-01 6.36841476e-01 1.23027360e+00 -1.41770959e-01
-1.20182987e-02 -2.19723172e-02 3.93608421e-01 -1.81848407e-01
-3.23410094e-01 -3.13014060e-01 -1.03040469e+00 3.98101538e-01
-1.57547295e+00 -5.16746640e-01 1.77986458e-01 2.04164386e+00
4.94043142e-01 4.51977611e-01 9.58705097e-02 -4.41382676e-01
5.16119957e-01 -4.01856989e-01 -1.17667449e+00 -2.54118711e-01
3.62294130e-02 -3.92625891e-02 9.36889827e-01 3.19461524e-01
-8.08029473e-01 1.09841943e+00 6.01286268e+00 5.37510403e-02
-8.85550261e-01 -1.34640187e-01 -1.84054673e-01 -2.12592363e-01
1.44602329e-01 -4.06784415e-01 -8.71632218e-01 9.46111381e-02
-1.08576335e-01 6.69052541e-01 9.16099668e-01 1.20933485e+00
-5.24263270e-02 -6.11483812e-01 -1.51436365e+00 7.76556730e-01
-1.05760414e-02 -6.54629350e-01 -2.11198226e-01 -2.18529820e-01
4.56645638e-01 4.33445603e-01 1.70408368e-01 3.95126730e-01
1.03749144e+00 -9.77759838e-01 1.12943113e+00 3.44088882e-01
7.00009644e-01 -1.34147197e-01 5.94826877e-01 5.51847279e-01
-8.66687894e-01 -6.06946707e-01 -5.62241077e-01 -3.31743136e-02
1.70730010e-01 -7.63416961e-02 -1.20525610e+00 -8.44230279e-02
1.33614516e+00 1.78362504e-01 -3.29856947e-02 9.88562584e-01
-1.60897389e-01 -2.19657853e-01 -7.40051270e-01 -2.28232354e-01
2.24829033e-01 4.34714332e-02 7.30089843e-01 9.19252515e-01
-6.69589862e-02 3.19150597e-01 5.83299994e-01 1.23882055e+00
-9.40026809e-03 -7.09792674e-01 -7.51034617e-01 3.28020304e-01
5.02495468e-01 1.00564110e+00 -7.81347632e-01 -5.31181917e-02
3.25661749e-02 9.54741538e-01 4.76310194e-01 6.35972857e-01
-5.94778717e-01 -3.46110761e-01 7.13029385e-01 -7.05106780e-02
5.35153687e-01 -7.20691085e-01 -2.08820641e-01 -8.15736830e-01
5.03021896e-01 -7.62268543e-01 -2.85004884e-01 -1.39515579e+00
-1.19484437e+00 2.88345814e-01 3.14867705e-01 -9.12601173e-01
1.35592565e-01 -1.34480095e+00 -1.11128174e-01 7.44958341e-01
-1.29822242e+00 -1.33354592e+00 -9.32407677e-01 2.71505266e-01
8.32574964e-01 6.50091469e-02 6.76400304e-01 -3.43287826e-01
1.59658864e-01 4.80849594e-02 -3.93707752e-02 -8.11854973e-02
6.85317814e-01 -1.19662833e+00 4.47342753e-01 1.89418942e-01
-6.96875811e-01 4.22184050e-01 6.78187132e-01 -8.35648954e-01
-1.91772974e+00 -8.58128846e-01 -5.51672935e-01 -7.58594990e-01
3.92307490e-01 -9.73326862e-01 -7.32159138e-01 9.48204696e-01
-2.65819341e-01 1.74426079e-01 -3.25006545e-01 -3.57358992e-01
-2.12341845e-01 -2.34499611e-02 -1.61240697e+00 7.36394644e-01
1.39201653e+00 -1.00917302e-01 -7.45147824e-01 5.77156961e-01
1.04023993e+00 -1.15540612e+00 -7.72210658e-01 7.39468992e-01
1.03240895e+00 -6.99454010e-01 1.37439382e+00 -5.68321586e-01
5.10469019e-01 -2.35967606e-01 -5.91558218e-01 -1.34716129e+00
2.53326930e-02 -2.59495735e-01 -6.52649160e-03 4.63207483e-01
2.26473242e-01 -6.46391332e-01 7.40987897e-01 9.74780381e-01
-3.63688260e-01 -5.29990554e-01 -7.01336622e-01 -8.80607605e-01
-1.86947323e-02 -3.08418393e-01 4.43832308e-01 1.97528332e-01
-2.66573876e-01 -3.90965700e-01 1.34652793e-01 5.51867485e-01
8.65323067e-01 4.11777198e-01 1.04210138e+00 -1.09521639e+00
-1.34894997e-01 -2.12377548e-01 -2.60637015e-01 -1.28670275e+00
2.28655472e-01 -6.65702045e-01 8.56136024e-01 -1.90212345e+00
3.19317207e-02 -1.06634855e+00 4.73925799e-01 7.31428862e-01
1.21979125e-01 -9.88563374e-02 5.64420819e-01 1.73882648e-01
-3.43714297e-01 3.27476978e-01 1.65565801e+00 -3.07873935e-01
-2.90568233e-01 8.37018620e-03 -3.12029608e-02 7.80903399e-01
6.68009996e-01 -1.44344345e-01 4.19900715e-02 -1.01047778e+00
-1.62651971e-01 1.31199107e-01 1.03949976e+00 -1.07120121e+00
3.19294375e-03 -3.38509142e-01 7.29528546e-01 -7.36209571e-01
8.99216115e-01 -1.43586838e+00 -2.77602941e-01 6.19283676e-01
-2.08891258e-01 -9.93281379e-02 5.09211004e-01 5.69275320e-01
6.30992234e-01 -1.19784847e-01 5.93556046e-01 -6.75007164e-01
-7.24574447e-01 1.64363220e-01 -2.11034581e-01 -1.14052333e-01
1.18705165e+00 -6.14050150e-01 -5.17556846e-01 -1.07572503e-01
-8.20011616e-01 6.22212648e-01 8.10726643e-01 8.24977875e-01
9.93245423e-01 -8.47384930e-01 -4.21883613e-01 2.09735021e-01
-5.93579412e-02 9.40194070e-01 -1.45189062e-01 2.08923504e-01
-7.83403277e-01 1.51086763e-01 -3.87931913e-01 -9.68639970e-01
-1.04812574e+00 4.60775226e-01 2.82378912e-01 6.21226847e-01
-6.61319077e-01 9.76555824e-01 2.26770103e-01 -1.00107384e+00
6.89224601e-01 -8.48062575e-01 5.51103771e-01 -7.91088939e-01
-5.65375425e-02 4.71898198e-01 -1.24552481e-01 -1.83147043e-01
-2.95509920e-02 6.65052891e-01 1.30603939e-01 6.26905262e-02
1.58269024e+00 4.18184288e-02 1.33125978e-02 3.08132052e-01
9.06482577e-01 -5.83629787e-01 -1.99793494e+00 3.14769149e-02
-5.15936732e-01 -4.96309817e-01 -4.51374441e-01 -1.08209765e+00
-7.60352612e-01 9.18522775e-01 6.92745984e-01 -1.73448727e-01
5.94772398e-01 3.66497040e-01 7.39650071e-01 1.18353391e+00
1.05882263e+00 -1.06292808e+00 4.54094768e-01 6.04005933e-01
1.41083896e+00 -1.59926820e+00 -2.40947679e-01 -4.80848849e-01
-4.58597511e-01 1.18620765e+00 1.28471828e+00 -2.47094184e-01
1.34625882e-01 6.62841380e-01 2.96774767e-02 -3.34067822e-01
-3.43289405e-01 -2.16198098e-02 -1.05119854e-01 9.46586311e-01
-5.96453369e-01 2.26794079e-01 8.86317253e-01 1.68200973e-02
-2.22107232e-01 -2.73861170e-01 2.21658975e-01 1.38781428e+00
-9.04458642e-01 -7.90921301e-02 -7.27242887e-01 3.62784326e-01
1.35639116e-01 3.25731099e-01 -2.53454119e-01 9.76038814e-01
1.27701059e-01 8.05999696e-01 1.75836861e-01 -2.44327471e-01
7.65501142e-01 -3.98522586e-01 9.48372304e-01 -8.15659821e-01
-4.23164517e-01 -2.54219145e-01 -1.76541165e-01 -7.73092806e-01
-9.64639634e-02 -5.32353282e-01 -1.60300255e+00 1.66654646e-01
-5.18768787e-01 -7.55280793e-01 1.14733005e+00 7.93027341e-01
5.50082512e-02 2.96636373e-01 3.24599981e-01 -1.91727579e+00
-9.39265549e-01 -9.12646949e-01 -4.08302516e-01 4.68900055e-01
5.58937848e-01 -1.01855361e+00 -2.55279303e-01 -8.64710957e-02] | [5.800038814544678, -0.8726003766059875] |
06bcd644-c61f-4b90-8a4d-70e2c845ed5c | yolosa-object-detection-based-on-2d-local | 2206.11825 | null | https://arxiv.org/abs/2206.11825v2 | https://arxiv.org/pdf/2206.11825v2.pdf | YOLOSA: Object detection based on 2D local feature superimposed self-attention | We analyzed the network structure of real-time object detection models and found that the features in the feature concatenation stage are very rich. Applying an attention module here can effectively improve the detection accuracy of the model. However, the commonly used attention module or self-attention module shows poor performance in detection accuracy and inference efficiency. Therefore, we propose a novel self-attention module, called 2D local feature superimposed self-attention, for the feature concatenation stage of the neck network. This self-attention module reflects global features through local features and local receptive fields. We also propose and optimize an efficient decoupled head and AB-OTA, and achieve SOTA results. Average precisions of 49.0% (71FPS, 14ms), 46.1% (85FPS, 11.7ms), and 39.1% (107FPS, 9.3ms) were obtained for large, medium, and small-scale models built using our proposed improvements. Our models exceeded YOLOv5 by 0.8% -- 3.1% in average precision. | ['Lin Huang', 'Weisheng Li'] | 2022-06-23 | null | null | null | null | ['real-time-object-detection'] | ['computer-vision'] | [-2.13693634e-01 -2.24227697e-01 -1.02775199e-02 -3.07975531e-01
-3.87219995e-01 -3.56798857e-01 4.19030458e-01 -1.42552629e-01
-6.31894052e-01 4.02155757e-01 -2.98406690e-01 -1.58611193e-01
2.59648591e-01 -7.92343199e-01 -7.39052713e-01 -4.76271123e-01
-1.08579911e-01 -6.92207143e-02 8.86316836e-01 8.51413608e-02
2.89226204e-01 7.81242907e-01 -1.81853127e+00 2.11662769e-01
6.45637453e-01 1.15330553e+00 5.72289228e-01 1.20623696e+00
1.15975747e-02 6.72297359e-01 -9.12115395e-01 -2.95723349e-01
1.12987012e-01 7.81741217e-02 -3.78882051e-01 -4.22398180e-01
6.15544736e-01 -6.68972552e-01 -5.35339415e-01 9.59685743e-01
8.71654510e-01 -2.31077209e-01 5.07918894e-01 -1.38017285e+00
-5.82774639e-01 4.14252162e-01 -1.09355700e+00 7.28075087e-01
8.23546797e-02 3.25123072e-01 7.15111434e-01 -1.00707495e+00
2.30708018e-01 1.34450614e+00 5.34561932e-01 4.91088808e-01
-7.05104470e-01 -1.06426036e+00 9.53359529e-02 2.30087414e-01
-1.64750576e+00 -6.14228666e-01 -1.59109756e-02 -1.08703129e-01
1.40323186e+00 2.04922557e-01 5.01056910e-01 5.14592409e-01
6.02918506e-01 8.58165443e-01 6.46292031e-01 -3.88854742e-01
-2.05737844e-01 1.56354144e-01 1.88306659e-01 9.93145764e-01
4.61415917e-01 4.25607562e-02 -4.67462420e-01 1.48068637e-01
1.03213096e+00 1.60012349e-01 2.62074620e-01 3.61001164e-01
-8.67137849e-01 6.13038301e-01 7.90684521e-01 2.19767928e-01
-3.41955781e-01 5.82241833e-01 1.72068760e-01 -3.91513109e-02
2.39912733e-01 7.18489662e-02 -2.63007849e-01 -2.89797317e-02
-9.33095813e-01 5.80446161e-02 4.41102803e-01 1.37426412e+00
6.24516368e-01 1.03372142e-01 -6.65802121e-01 5.59527457e-01
4.52648908e-01 1.20132875e+00 2.94777334e-01 -9.59660411e-01
2.67633021e-01 5.31805456e-01 2.09853992e-01 -1.24571717e+00
-6.92886233e-01 -7.52963781e-01 -7.81083405e-01 5.19106016e-02
1.12794168e-01 -2.07748413e-01 -1.08857346e+00 1.59401333e+00
2.64773220e-01 9.06078070e-02 -1.99549183e-01 8.15909982e-01
9.81198251e-01 5.94612837e-01 2.40303591e-01 -8.94142538e-02
1.49775565e+00 -1.22522581e+00 -6.68095767e-01 -2.05718398e-01
3.07070792e-01 -9.38631237e-01 7.37649381e-01 -2.52469294e-02
-1.45871699e+00 -1.02499282e+00 -1.03205967e+00 -7.36139864e-02
-2.70300567e-01 6.57136679e-01 4.89968389e-01 5.16404688e-01
-1.23564553e+00 3.29865068e-01 -7.59948790e-01 -5.19987106e-01
6.03789508e-01 7.14554310e-01 1.03359498e-01 8.38919878e-02
-7.38083184e-01 7.70024061e-01 1.09560512e-01 -1.43556997e-01
-9.91219163e-01 -3.74243021e-01 -4.20004725e-01 4.94156986e-01
3.73403758e-01 -6.37990296e-01 1.42639768e+00 -2.82420546e-01
-1.25547445e+00 6.15127623e-01 -4.27386880e-01 -6.44658387e-01
9.97034162e-02 -4.92427677e-01 -4.69500601e-01 2.70661265e-01
2.28304490e-01 1.01644647e+00 8.55657458e-01 -6.69995308e-01
-1.21579134e+00 -2.72045672e-01 1.23172641e-01 2.41571385e-02
-2.78347164e-01 5.04975080e-01 -8.15741360e-01 -3.39649320e-01
-8.24357849e-04 -6.25515997e-01 -8.78092125e-02 2.62904376e-01
-2.71524936e-01 -2.29190290e-01 1.15343201e+00 -3.99721235e-01
1.29829228e+00 -2.25678968e+00 -7.10563838e-01 -5.89928962e-02
5.49514532e-01 7.83851385e-01 -1.19556382e-01 -1.44042879e-01
3.92660677e-01 1.59180313e-01 2.81390667e-01 -2.55205542e-01
-2.09248692e-01 -5.01651950e-02 3.04256417e-02 2.81184882e-01
4.89221454e-01 1.07830858e+00 -7.05485702e-01 -7.68225253e-01
3.41540903e-01 5.39237797e-01 -6.70740664e-01 3.24315310e-01
2.71099299e-01 -1.28641009e-01 -4.13317859e-01 8.81552517e-01
8.41723144e-01 -4.73253012e-01 -4.12372142e-01 -3.51379603e-01
-4.16838408e-01 4.00151536e-02 -9.99523938e-01 1.44394076e+00
-2.49196306e-01 7.35609889e-01 -2.49522701e-02 -4.49616522e-01
9.17648792e-01 -1.62463933e-02 2.60828227e-01 -9.19683754e-01
4.77242053e-01 -8.85521993e-02 2.68275261e-01 -4.68995750e-01
6.21973336e-01 4.91368204e-01 1.42893672e-01 1.62999004e-01
3.33623707e-01 4.05849636e-01 2.64207929e-01 3.51310074e-01
1.23942471e+00 -1.48938254e-01 3.66333991e-01 -2.41611317e-01
4.75427389e-01 -3.85398626e-01 3.10442507e-01 1.26927257e+00
-5.68660438e-01 6.10118151e-01 2.68938631e-01 -3.07159156e-01
-6.94001615e-01 -9.27742600e-01 -1.36955917e-01 1.42185771e+00
2.12017670e-01 -5.52466452e-01 -8.03974986e-01 -5.84575415e-01
3.40522118e-02 2.66117156e-01 -4.44396824e-01 -3.01877856e-01
-4.30808365e-01 -7.94423580e-01 8.23622525e-01 1.07981014e+00
8.62992525e-01 -1.04062271e+00 -1.18743408e+00 3.49797696e-01
1.12085901e-01 -1.20301664e+00 -4.62152272e-01 1.98877797e-01
-6.26961470e-01 -8.76761734e-01 -7.92611539e-01 -5.34270227e-01
5.59333742e-01 5.56228936e-01 8.53242278e-01 1.18693136e-01
-6.68088436e-01 8.64173844e-02 -1.08042888e-01 -6.39218867e-01
6.44644201e-02 1.41000971e-01 1.01335883e-01 -2.06378639e-01
4.41909224e-01 -1.75603479e-01 -7.31628418e-01 4.22632664e-01
-2.77423322e-01 -1.62597448e-01 1.03755009e+00 5.45014262e-01
4.76433247e-01 -4.74138349e-01 4.80752259e-01 -4.73951846e-01
1.68192446e-01 -2.34239101e-01 -7.26908386e-01 2.23849177e-01
-4.52461839e-01 -8.73748288e-02 3.95009160e-01 -4.07395303e-01
-9.87328887e-01 1.51717022e-01 -1.27913460e-01 -4.83371705e-01
-1.95923880e-01 -2.49219015e-01 4.78904322e-02 -4.63767648e-01
6.72746301e-01 1.42100543e-01 -1.63885146e-01 -2.96061754e-01
1.83358669e-01 8.39988828e-01 6.03592694e-01 -1.10894211e-01
5.43481410e-01 4.10356134e-01 -1.27316728e-01 -8.01563442e-01
-6.45550013e-01 -6.15560055e-01 -4.68612462e-01 -2.60613635e-02
7.20387518e-01 -1.11420965e+00 -1.11476040e+00 5.96065462e-01
-1.23640716e+00 1.22314177e-01 -1.90782979e-01 3.14222187e-01
-2.16108233e-01 1.67165086e-01 -7.29188263e-01 -9.69438374e-01
-8.89946699e-01 -1.01252818e+00 1.16233218e+00 7.60780275e-01
1.50690690e-01 -2.22964376e-01 -5.06706238e-01 -3.00598562e-01
1.04478669e+00 -2.13703349e-01 6.32374436e-02 -4.02115822e-01
-7.13680327e-01 -3.50595921e-01 -1.14565647e+00 1.14457935e-01
-2.39840969e-02 3.35059375e-01 -1.30172789e+00 -2.80969709e-01
-3.44656885e-01 -7.42287561e-02 9.67814982e-01 7.36090422e-01
1.35098338e+00 -1.17166862e-01 -8.45515013e-01 6.10729754e-01
1.29599380e+00 4.55030322e-01 5.90399265e-01 -5.94981201e-02
4.77352411e-01 -1.11785218e-01 5.95086217e-01 6.56511784e-01
3.25041324e-01 6.05128467e-01 5.90713143e-01 -1.89924672e-01
-4.27404284e-01 -3.22709382e-02 2.93726921e-01 5.55568755e-01
-2.19477281e-01 -4.76712376e-01 -8.06330025e-01 4.61241782e-01
-1.77110231e+00 -8.89591992e-01 -1.18084006e-01 1.95739889e+00
3.86148155e-01 5.53001046e-01 2.12582171e-01 -1.58367082e-01
9.73700106e-01 -2.09060442e-02 -6.16672099e-01 -3.30004483e-01
1.49414744e-02 9.88707840e-02 7.33629644e-01 1.58932552e-01
-1.25487900e+00 1.04083550e+00 6.75501728e+00 8.47554386e-01
-9.95038748e-01 1.42293081e-01 4.37227994e-01 -2.63836443e-01
5.73764324e-01 -3.62775892e-01 -1.56912410e+00 5.57876289e-01
1.19166541e+00 -7.55429640e-02 3.32700573e-02 9.90390956e-01
-9.92836431e-02 -3.80271286e-01 -7.99274743e-01 1.09085047e+00
1.32084876e-01 -1.07042980e+00 -8.61678272e-02 -9.75653715e-03
3.39053094e-01 2.72156775e-01 1.10509314e-01 4.14117515e-01
2.43434221e-01 -7.74225473e-01 6.08628452e-01 5.43857634e-01
1.05877626e+00 -7.14739859e-01 8.66419733e-01 2.54269630e-01
-1.57287824e+00 -1.84055224e-01 -6.79341972e-01 -8.33818242e-02
-7.74182901e-02 4.16108340e-01 -7.89470196e-01 6.07537702e-02
1.15461087e+00 5.84319770e-01 -8.80387008e-01 1.35394108e+00
-3.01707909e-02 4.06416953e-01 -7.68455625e-01 -4.50834721e-01
1.55311301e-01 7.83348441e-01 3.03271741e-01 1.67284191e+00
1.45899296e-01 8.91955718e-02 -5.71467392e-02 7.52479613e-01
-1.12860247e-01 -2.28253752e-01 -2.28375167e-01 3.05688947e-01
5.98099470e-01 1.49434769e+00 -7.90340424e-01 -5.39123118e-01
-3.79428267e-01 8.52492213e-01 3.01128924e-01 1.52548283e-01
-1.42024171e+00 -8.99354517e-01 5.38071334e-01 -8.78547281e-02
6.13821447e-01 2.28723530e-02 3.99430692e-02 -9.10666525e-01
-1.12689346e-01 -3.25834155e-01 2.67080545e-01 -7.75143087e-01
-7.91978657e-01 4.57586259e-01 -2.86470413e-01 -9.35750246e-01
1.16179667e-01 -6.18122756e-01 -6.67557657e-01 7.34981716e-01
-1.50140643e+00 -1.00467014e+00 -7.03504980e-01 5.34060001e-01
3.85509670e-01 -1.46302879e-01 6.08778298e-01 4.67616498e-01
-7.74298728e-01 1.20587718e+00 -1.76033154e-02 2.12665498e-01
6.60399675e-01 -8.79736364e-01 7.34714031e-01 8.75769973e-01
-3.84540744e-02 5.64935207e-01 1.78131595e-01 -3.58441740e-01
-1.22094560e+00 -1.24771845e+00 8.12688947e-01 -1.88816786e-01
2.23818198e-01 -2.65552700e-01 -7.32418001e-01 5.22730231e-01
1.08791865e-01 5.96720219e-01 1.79920286e-01 -1.00349307e-01
-2.36223921e-01 -3.18078071e-01 -1.36918247e+00 4.45984244e-01
1.22203088e+00 -2.47740194e-01 -1.71671242e-01 8.94399136e-02
8.68440390e-01 -5.34060538e-01 -5.76198161e-01 3.53279144e-01
8.60035360e-01 -9.10727978e-01 8.99961770e-01 -3.45780820e-01
6.00630604e-02 -3.55119020e-01 -2.41706029e-01 -5.04561722e-01
-9.18734252e-01 -3.69031101e-01 -6.40244484e-01 1.17441809e+00
3.60248506e-01 -6.60704255e-01 7.30258346e-01 1.99690774e-01
-1.49224684e-01 -6.19648278e-01 -8.95999014e-01 -6.00980997e-01
-4.41404969e-01 -3.12131852e-01 5.43769777e-01 2.06491232e-01
-2.93687552e-01 3.98201585e-01 -1.99924409e-01 3.30832899e-01
5.47767401e-01 -4.43872239e-04 6.89198434e-01 -1.04786098e+00
-1.44425765e-01 -4.85655755e-01 -6.31314099e-01 -1.35699332e+00
-5.88509381e-01 -3.35598946e-01 -4.30051014e-02 -1.36429167e+00
5.26963174e-01 -2.89325237e-01 -6.89660728e-01 7.92664766e-01
-2.10620329e-01 3.44618559e-01 3.97412360e-01 9.67653766e-02
-1.11978281e+00 2.44589597e-01 1.07468843e+00 1.38786614e-01
4.98822443e-02 1.11078108e-02 -6.84582174e-01 8.08198273e-01
1.03049350e+00 -6.33028090e-01 1.07318059e-01 -6.19235754e-01
-3.39471132e-01 -2.98447490e-01 4.16186601e-01 -1.52639365e+00
7.17058361e-01 1.23018555e-01 8.54365170e-01 -9.58517015e-01
3.96517634e-01 -6.51072204e-01 -5.20062506e-01 9.04488802e-01
-2.18875762e-02 7.42411390e-02 5.74627817e-01 4.75299895e-01
8.42184797e-02 2.47211941e-02 7.81818807e-01 -3.99907380e-02
-9.68430638e-01 2.91053683e-01 -3.62352669e-01 -1.35030765e-02
1.06933463e+00 -2.31935576e-01 -5.99274516e-01 -1.22360192e-01
-5.70239782e-01 3.47622842e-01 -3.97973359e-02 4.56461638e-01
7.65874803e-01 -1.13626993e+00 -5.56473672e-01 5.84309638e-01
-2.76262462e-02 -3.20653729e-02 2.88475543e-01 7.53947794e-01
-4.09126431e-01 7.24025905e-01 -4.20741767e-01 -8.66205692e-01
-1.27128017e+00 3.72869909e-01 2.66347378e-01 -8.86741728e-02
-2.70675421e-01 1.19936562e+00 2.73105800e-01 1.33355692e-01
4.26429749e-01 -5.32859504e-01 -2.26999089e-01 -1.10593311e-01
9.44657266e-01 7.28406489e-01 -1.87952504e-01 -5.71139157e-01
-6.76073313e-01 8.13092530e-01 -4.39705223e-01 3.58350575e-01
8.67545485e-01 -1.62042677e-01 2.96538055e-01 -7.48631805e-02
1.05633461e+00 -2.42866233e-01 -1.35816181e+00 -3.31978709e-01
-3.40643376e-01 -3.39079440e-01 1.78148210e-01 -7.79229164e-01
-1.21700633e+00 1.10924900e+00 1.02010524e+00 1.30661353e-01
1.24194551e+00 8.02914500e-02 6.67597353e-01 3.86230886e-01
3.91083241e-01 -7.89513409e-01 1.86410118e-02 7.23872364e-01
6.67794168e-01 -1.28029799e+00 5.60280271e-02 -3.27440262e-01
-1.61452547e-01 8.51864278e-01 1.22649789e+00 -2.89876431e-01
5.16573071e-01 7.49465585e-01 -3.23023468e-01 -1.42414197e-01
-8.74722183e-01 -3.52636397e-01 2.74716794e-01 4.75229353e-01
3.67863894e-01 7.00992765e-03 6.88634440e-02 6.33796036e-01
1.10326618e-01 1.68135334e-02 1.13859840e-01 1.05049253e+00
-9.34936464e-01 -2.43283108e-01 -2.29813382e-01 6.03773773e-01
-6.78708076e-01 -9.18950662e-02 -4.03865017e-02 7.57159114e-01
2.81581432e-01 9.93994474e-01 4.69725877e-01 -6.25555634e-01
3.77440780e-01 -2.83304185e-01 2.76872396e-01 -5.72008967e-01
-6.22822881e-01 2.06579998e-01 -3.55012298e-01 -8.15459371e-01
-3.77188951e-01 -2.22997889e-01 -1.34268117e+00 -5.04760504e-01
-9.48493004e-01 -1.96930453e-01 7.25017667e-01 5.53181767e-01
8.54801238e-01 8.40516925e-01 4.75918800e-01 -8.15563202e-01
-3.88661444e-01 -1.06880403e+00 -4.20544833e-01 -3.70032191e-01
4.15445834e-01 -6.56960309e-01 -2.91073829e-01 -1.03420720e-01] | [8.702938079833984, -0.2866772413253784] |
d38a8b36-ebc7-4494-a051-09d53d3e27d9 | discovering-individual-rewards-in-collective | 2305.10548 | null | https://arxiv.org/abs/2305.10548v1 | https://arxiv.org/pdf/2305.10548v1.pdf | Discovering Individual Rewards in Collective Behavior through Inverse Multi-Agent Reinforcement Learning | The discovery of individual objectives in collective behavior of complex dynamical systems such as fish schools and bacteria colonies is a long-standing challenge. Inverse reinforcement learning is a potent approach for addressing this challenge but its applicability to dynamical systems, involving continuous state-action spaces and multiple interacting agents, has been limited. In this study, we tackle this challenge by introducing an off-policy inverse multi-agent reinforcement learning algorithm (IMARL). Our approach combines the ReF-ER techniques with guided cost learning. By leveraging demonstrations, our algorithm automatically uncovers the reward function and learns an effective policy for the agents. Through extensive experimentation, we demonstrate that the proposed policy captures the behavior observed in the provided data, and achieves promising results across problem domains including single agent models in the OpenAI gym and multi-agent models of schooling behavior. The present study shows that the proposed IMARL algorithm is a significant step towards understanding collective dynamics from the perspective of its constituents, and showcases its value as a tool for studying complex physical systems exhibiting collective behaviour. | ['Petros Koumoutsakos', 'Pascal Weber', 'Daniel Waelchli'] | 2023-05-17 | null | null | null | null | ['multi-agent-reinforcement-learning'] | ['methodology'] | [-4.26392704e-02 5.67835895e-03 -5.86980134e-02 4.80396152e-01
-4.10751998e-01 -5.89339256e-01 7.62717426e-01 2.69985616e-01
-5.67004263e-01 1.16054821e+00 -1.88845411e-01 -2.08815008e-01
-7.79027283e-01 -5.09353578e-01 -7.28952646e-01 -1.27588165e+00
-7.31726110e-01 5.83445549e-01 2.42827594e-01 -7.46841133e-01
4.08197731e-01 3.53175163e-01 -1.48054111e+00 -4.78760481e-01
1.11273110e+00 1.82863221e-01 3.88703614e-01 1.21828461e+00
4.04297858e-01 9.60019588e-01 -5.80609560e-01 3.20811063e-01
1.94566101e-01 -5.75842321e-01 -7.76605129e-01 2.86097497e-01
-3.26552600e-01 -8.26018080e-02 -6.65625185e-02 7.02919960e-01
6.72517002e-01 3.55026394e-01 6.80812299e-01 -1.46192694e+00
-4.16213274e-01 4.68319148e-01 -4.90747690e-01 3.10906202e-01
1.43530682e-01 5.68549156e-01 1.03039086e+00 -2.58468390e-02
3.17926496e-01 1.29569638e+00 4.07222092e-01 6.14182830e-01
-1.35467029e+00 -4.69437093e-01 1.64659724e-01 6.00088900e-03
-8.75579596e-01 -7.68515766e-02 5.67909300e-01 -4.27936643e-01
8.95927966e-01 -4.04754579e-02 1.07977676e+00 8.45467329e-01
4.29127604e-01 8.02924335e-01 1.61477268e+00 -5.89829981e-01
7.06971645e-01 -2.80441165e-01 -9.62003767e-02 8.52234483e-01
2.23612309e-01 5.47136724e-01 -3.03887069e-01 -4.09408361e-01
8.50302577e-01 -1.50034979e-01 -3.50784622e-02 -5.72412908e-01
-1.06476676e+00 7.91414857e-01 1.57330409e-01 1.03433624e-01
-4.90571588e-01 4.95770395e-01 -2.02720128e-02 6.56285524e-01
2.62374669e-01 8.46703470e-01 -4.93404597e-01 -3.64745200e-01
-8.83511603e-02 7.63782978e-01 1.02566779e+00 3.86872560e-01
6.61175132e-01 4.68171500e-02 4.01028603e-01 5.39445698e-01
4.20282513e-01 8.48224938e-01 2.27793545e-01 -1.36546540e+00
7.98926223e-03 6.13792479e-01 5.73973477e-01 -6.31644607e-01
-5.81876159e-01 -1.80404991e-01 -5.70847034e-01 7.31070995e-01
6.04483426e-01 -6.13127053e-01 -3.98953468e-01 1.77248728e+00
7.49291599e-01 4.80892539e-01 5.07904947e-01 7.53363788e-01
1.97114170e-01 5.37733316e-01 1.19658457e-02 -3.56644422e-01
1.08086371e+00 -8.30013692e-01 -5.83301842e-01 7.53993839e-02
6.02335155e-01 -3.18132222e-01 7.49843776e-01 1.87728107e-01
-1.10018110e+00 -1.13649324e-01 -7.96019614e-01 6.96533203e-01
-1.60748631e-01 -3.87071609e-01 5.10887384e-01 3.21678847e-01
-1.12772644e+00 8.73999596e-01 -1.19842243e+00 -5.21599710e-01
8.86174962e-02 7.47226298e-01 1.40033066e-01 4.57221925e-01
-9.40823257e-01 7.99373031e-01 3.59554775e-03 -1.71588704e-01
-1.11240768e+00 -5.22408307e-01 -5.79561055e-01 -8.05803761e-02
7.73108125e-01 -6.81026101e-01 1.44086874e+00 -7.16284275e-01
-2.11352897e+00 2.58822203e-01 2.85746068e-01 -4.17887539e-01
3.64535511e-01 1.63551439e-02 -2.01148912e-02 1.32749379e-01
1.63153008e-01 3.90178293e-01 8.22590888e-01 -1.33538330e+00
-7.92215705e-01 -8.40195566e-02 3.62373203e-01 4.99417424e-01
-2.10104570e-01 -3.47054511e-01 6.18208349e-01 -2.98830330e-01
-6.63696706e-01 -1.35372794e+00 -6.68342054e-01 -3.96304607e-01
1.89120248e-01 -6.26159012e-01 8.18732679e-01 1.75104693e-01
7.82575905e-01 -1.80495059e+00 6.57177746e-01 1.23990715e-01
2.84004182e-01 2.95162886e-01 -2.72285163e-01 1.01229465e+00
3.99321198e-01 1.35997459e-02 -1.18969582e-01 3.54273915e-02
1.22044727e-01 5.73805511e-01 -1.06594928e-01 5.19023478e-01
2.39798486e-01 7.85873353e-01 -1.37112331e+00 -2.79178709e-01
1.69810206e-01 1.70839801e-01 -6.25341773e-01 4.43923503e-01
-4.05119032e-01 7.66076267e-01 -9.18686688e-01 2.39832655e-01
2.29313761e-01 -4.91117299e-01 3.67514938e-01 7.74312079e-01
-5.39706767e-01 -3.26449990e-01 -1.34776330e+00 9.91993427e-01
-3.41992348e-01 1.96513191e-01 2.59495050e-01 -1.26802826e+00
6.54844403e-01 3.51869255e-01 8.88413250e-01 -4.83729213e-01
2.48177294e-02 3.09097230e-01 6.97470069e-01 -9.09146726e-01
2.33110711e-01 -1.34692222e-01 -1.10069029e-01 8.77035737e-01
8.51966441e-02 -3.92013311e-01 4.45194960e-01 -1.12329051e-02
1.49249542e+00 3.05872113e-01 4.92985249e-01 -6.03120744e-01
6.36081874e-01 3.13155979e-01 2.91417629e-01 1.03860104e+00
-4.58605140e-01 -3.66164804e-01 6.27765954e-01 -4.44421411e-01
-9.90352809e-01 -7.76827276e-01 2.93131262e-01 1.13485777e+00
4.22583312e-01 -5.46045005e-02 -7.36790538e-01 -2.44794801e-01
1.34816557e-01 4.80796732e-02 -7.92878211e-01 -1.14401303e-01
-9.66123164e-01 -1.05537653e+00 1.70503438e-01 -7.59168481e-03
4.28802550e-01 -1.55497932e+00 -9.55326736e-01 5.95914960e-01
2.01486707e-01 -9.00102496e-01 -2.36614242e-01 1.65629804e-01
-6.20393157e-01 -1.37389410e+00 -6.21840119e-01 -7.94220030e-01
5.13704956e-01 3.38126957e-01 9.33992624e-01 4.39402908e-01
-4.02220935e-01 9.92437840e-01 -4.75518584e-01 -2.98972577e-01
-8.54957759e-01 1.54548332e-01 3.11169028e-01 -1.16448857e-01
-1.90855548e-01 -7.31607676e-01 -7.72972584e-01 3.62148672e-01
-7.63326406e-01 -1.64648131e-01 6.09440506e-01 1.13308251e+00
2.08451524e-01 1.38062298e-01 9.96656358e-01 -3.16021234e-01
1.08526766e+00 -4.96592700e-01 -9.91953075e-01 1.51210904e-01
-4.83412474e-01 3.02701086e-01 8.64849269e-01 -7.11301088e-01
-7.16363430e-01 -3.71511318e-02 1.58675358e-01 1.91023991e-01
-2.25674704e-01 1.69416308e-01 4.97295648e-01 -3.25729579e-01
4.06895936e-01 4.20635343e-01 5.74398160e-01 -1.32857680e-01
1.14468254e-01 3.44112426e-01 -1.33758366e-01 -9.32003677e-01
7.78704047e-01 3.81830990e-01 4.96638149e-01 -1.16120481e+00
-4.63227689e-01 -3.02502006e-01 -3.87346953e-01 -3.93252015e-01
8.88521612e-01 -6.36013210e-01 -1.76900494e+00 6.05220795e-01
-6.77538395e-01 -9.61969078e-01 -3.92313004e-01 3.51825237e-01
-1.10006797e+00 1.48847893e-01 -4.23372686e-01 -1.13120770e+00
2.58935839e-02 -1.26897466e+00 9.58529234e-01 6.45473540e-01
1.32706255e-01 -1.39450371e+00 8.08506191e-01 -6.25064000e-02
2.65032560e-01 3.15160096e-01 7.10829258e-01 -4.54335719e-01
-7.63428330e-01 3.43950182e-01 5.28705299e-01 -1.13176897e-01
2.09510565e-01 1.51896551e-01 -3.63384098e-01 -6.92479670e-01
-3.47552270e-01 -6.83566391e-01 3.65114003e-01 5.35578191e-01
4.57948446e-01 -1.49198964e-01 -1.84343278e-01 -1.31147392e-02
1.30863190e+00 3.16319972e-01 -1.77133810e-02 6.62286818e-01
1.88235328e-01 6.44537449e-01 5.70406139e-01 7.19448566e-01
6.27378702e-01 5.43486953e-01 7.24428296e-01 3.30136828e-02
1.14601947e-01 7.72640333e-02 5.65529585e-01 7.92438805e-01
-4.16466445e-01 -2.13828936e-01 -9.60171938e-01 5.22952497e-01
-2.26327634e+00 -1.09667623e+00 -4.07421887e-02 1.78943348e+00
4.95939940e-01 -2.84657985e-01 6.05824053e-01 -1.03912495e-01
4.65112776e-01 -1.01673558e-01 -7.61255145e-01 -3.75883132e-01
1.02120847e-01 7.61950612e-02 2.23988399e-01 7.05468118e-01
-9.65339482e-01 8.93236220e-01 6.68729734e+00 5.17775834e-01
-7.90468037e-01 -1.76985085e-01 1.19598500e-01 2.84561008e-01
2.61939794e-01 -8.95261019e-03 -7.40068913e-01 2.29142949e-01
8.80435765e-01 -3.82277548e-01 1.03371286e+00 3.99103940e-01
8.37447047e-01 -3.15451533e-01 -6.98975623e-01 2.36161411e-01
-4.42635775e-01 -1.17300761e+00 -5.11500001e-01 5.61714947e-01
1.08074689e+00 -5.41698234e-03 1.39124975e-01 5.01181126e-01
1.16215992e+00 -7.84119427e-01 3.06058764e-01 4.34840858e-01
-4.42183726e-02 -8.00811172e-01 4.72873777e-01 8.38265240e-01
-1.06430912e+00 -5.17037749e-01 -1.07848920e-01 -6.94323957e-01
-1.07684150e-01 -3.00254971e-01 -9.14622962e-01 1.73973754e-01
4.82059002e-01 8.19037318e-01 -2.62437791e-01 9.26201940e-01
8.50790143e-02 6.81959271e-01 -2.39476621e-01 -8.36320341e-01
5.81050217e-01 -5.12398660e-01 7.34429479e-01 5.77360988e-01
9.04892311e-02 1.98983788e-01 6.67678833e-01 5.80242693e-01
4.03190106e-01 -3.05728707e-02 -7.72688746e-01 -1.08191505e-01
7.34395161e-02 1.25767875e+00 -9.15479183e-01 -1.31966889e-01
-8.94217640e-02 3.65519255e-01 7.06459641e-01 4.09132332e-01
-7.62784362e-01 -2.42330730e-02 9.52041984e-01 -3.48918259e-01
4.49991226e-01 -4.12274629e-01 3.93659145e-01 -9.99523282e-01
-5.51172912e-01 -1.05064142e+00 9.91368964e-02 -1.86248362e-01
-1.15920484e+00 1.48088306e-01 1.09914958e-01 -1.33839476e+00
-6.25829816e-01 -5.12081206e-01 -6.04731321e-01 2.60822535e-01
-1.49747324e+00 -8.79621446e-01 6.33215671e-03 4.71256346e-01
4.77978468e-01 -3.89512181e-01 8.58615994e-01 -2.85056710e-01
-5.94747663e-01 -4.80383821e-02 7.17162371e-01 -2.65265077e-01
1.20789081e-01 -1.71301627e+00 -2.94685327e-02 3.45036089e-01
-1.79436579e-01 2.61248380e-01 1.00394547e+00 -4.39706594e-01
-1.81393862e+00 -6.08148634e-01 -1.03221104e-01 -3.54806215e-01
1.23017764e+00 -1.64475501e-01 -7.51087189e-01 2.36686677e-01
6.61069274e-01 5.69743337e-03 5.77973306e-01 -1.48228973e-01
5.61586618e-01 1.55498922e-01 -8.23439181e-01 8.38015020e-01
8.20492625e-01 2.03350380e-01 -2.85673738e-01 3.70940149e-01
5.58044910e-01 -2.00536087e-01 -1.02173853e+00 1.38266340e-01
4.46388423e-01 -6.26656890e-01 7.95415819e-01 -8.99150133e-01
4.95139569e-01 -2.73653537e-01 1.77603334e-01 -1.96418166e+00
-2.63169378e-01 -1.06132758e+00 -3.23233157e-01 7.93861687e-01
9.30170864e-02 -7.94800401e-01 6.22867405e-01 4.14125854e-03
8.03811327e-02 -7.24566996e-01 -1.01592600e+00 -9.00771320e-01
4.66938257e-01 3.07134271e-01 2.38753542e-01 7.40708113e-01
8.99360105e-02 2.69842327e-01 -3.65111798e-01 4.03466254e-01
9.39641893e-01 1.76291227e-01 1.19405770e+00 -9.27131414e-01
-6.92367911e-01 -4.28822100e-01 -1.00808293e-01 -9.57293034e-01
4.77612168e-01 -4.60365146e-01 1.55343056e-01 -1.12074900e+00
8.77503827e-02 -5.96294105e-01 -3.74763422e-02 7.62433410e-02
-2.40425840e-01 -1.71412900e-01 2.72452921e-01 1.46355167e-01
-8.93735051e-01 7.62218177e-01 1.88407779e+00 -2.69244867e-03
-3.55214208e-01 3.04502755e-01 -4.22631145e-01 7.47249067e-01
9.59352374e-01 -6.07097745e-01 -2.99514055e-01 4.10135686e-02
1.34284481e-01 4.06567842e-01 5.54641664e-01 -7.68266439e-01
2.04026282e-01 -7.85200357e-01 -4.18881685e-01 8.55174661e-03
2.80788511e-01 -7.40963757e-01 -2.32463047e-01 1.17892289e+00
-3.96508783e-01 2.51906067e-01 -1.01671629e-01 1.07023597e+00
1.14639647e-01 -1.98910519e-01 7.08030462e-01 -3.67504030e-01
-6.57148600e-01 -5.66627365e-03 -8.87970567e-01 4.19770926e-01
1.39883840e+00 2.06572220e-01 -4.43843305e-01 -3.60829353e-01
-8.15948188e-01 8.26460004e-01 3.30931634e-01 1.28660068e-01
3.32224548e-01 -8.00425768e-01 -8.15620899e-01 3.65618765e-02
-1.86320230e-01 -3.39970529e-01 6.28752932e-02 9.11213577e-01
-5.16605854e-01 -1.37750208e-02 -3.08312595e-01 -7.64958560e-01
-1.26883590e+00 2.77854443e-01 7.19811738e-01 -6.42296553e-01
-6.44255996e-01 2.03397170e-01 -8.15773904e-02 -6.38331771e-01
-1.51689038e-01 -2.18416333e-01 -4.47045773e-01 -2.73203731e-01
2.34187707e-01 5.71199477e-01 -6.47059262e-01 -4.85872060e-01
3.81733589e-02 6.62812591e-01 4.67247367e-02 -2.53835946e-01
1.77730155e+00 -2.55908847e-01 -1.33305704e-02 3.69348139e-01
5.90742767e-01 -3.74367893e-01 -1.63580108e+00 -5.27105369e-02
1.65126603e-02 7.61541771e-03 -2.75910586e-01 -3.89492810e-01
-3.94748658e-01 5.40056288e-01 4.74180251e-01 8.86627436e-01
8.26912403e-01 5.15923388e-02 3.35449278e-01 8.36017191e-01
4.62820619e-01 -1.12060964e+00 7.86259174e-01 7.91966319e-01
6.58800304e-01 -1.45170438e+00 -3.89505252e-02 4.85253036e-02
-5.15652418e-01 1.08482122e+00 7.76176155e-01 -8.14240932e-01
7.31547177e-01 3.26684743e-01 -6.31603375e-02 -2.39614919e-01
-1.24775803e+00 -5.54080665e-01 -2.84012198e-01 6.23695850e-01
-3.10848448e-02 1.08744986e-01 -3.62419248e-01 -2.16840416e-01
-3.79155055e-02 -2.15122670e-01 1.04834175e+00 1.25335574e+00
-6.74959719e-01 -1.31092930e+00 -5.51892042e-01 1.52691305e-01
-2.06783906e-01 5.07906973e-01 -2.54874349e-01 1.14248204e+00
-1.25444695e-01 1.01426673e+00 -2.15200990e-01 8.83824080e-02
1.79572985e-01 -3.51805955e-01 5.76301813e-01 -4.70525682e-01
-6.47024155e-01 1.35658622e-01 -2.79618025e-01 -1.33238420e-01
-1.02597857e+00 -9.17145371e-01 -1.40467691e+00 -3.35966557e-01
-2.72207916e-01 4.46828604e-01 3.30454379e-01 1.15648711e+00
2.55302370e-01 7.46005595e-01 9.73977208e-01 -9.76211607e-01
-1.00396371e+00 -6.64538801e-01 -5.49663663e-01 2.00935692e-01
8.23790431e-01 -9.82347250e-01 -4.09176856e-01 -7.48971179e-02] | [3.8929121494293213, 2.0518603324890137] |
fb033998-16dd-4ab7-82f5-b1c1ab62a491 | driftrec-adapting-diffusion-models-to-blind | 2211.06757 | null | https://arxiv.org/abs/2211.06757v2 | https://arxiv.org/pdf/2211.06757v2.pdf | DriftRec: Adapting diffusion models to blind JPEG restoration | In this work, we utilize the high-fidelity generation abilities of diffusion models to solve blind JPEG restoration at high compression levels. We propose an elegant modification of the forward stochastic differential equation of diffusion models to adapt them to this restoration task and name our method DriftRec. Comparing DriftRec against an $L_2$ regression baseline with the same network architecture and two state-of-the-art techniques for JPEG restoration, we show that our approach can escape the tendency of other methods to generate blurry images, and recovers the distribution of clean images significantly more faithfully. For this, only a dataset of clean/corrupted image pairs and no knowledge about the corruption operation is required, enabling wider applicability to other restoration tasks. In contrast to other conditional and unconditional diffusion models, we utilize the idea that the distributions of clean and corrupted images are much closer to each other than each is to the usual Gaussian prior of the reverse process in diffusion models. Our approach therefore requires only low levels of added noise, and needs comparatively few sampling steps even without further optimizations. We show that DriftRec naturally generalizes to realistic and difficult scenarios such as unaligned double JPEG compression and blind restoration of JPEGs found online, without having encountered such examples during training. | ['Timo Gerkmann', 'Henry N. Chapman', 'Simon Welker'] | 2022-11-12 | null | null | null | null | ['jpeg-artifact-removal'] | ['computer-vision'] | [ 2.79183686e-01 -9.00821760e-02 1.87721983e-01 -1.20047882e-01
-8.11251163e-01 -5.34057319e-01 7.62675345e-01 -4.55363989e-01
-3.75207186e-01 7.35443115e-01 5.73239267e-01 -3.42283875e-01
-1.89253241e-01 -5.64703107e-01 -8.22510123e-01 -9.98339891e-01
-9.99683663e-02 2.62718171e-01 1.14479205e-02 -2.91270435e-01
3.35964173e-01 1.78788126e-01 -1.20969665e+00 2.18772665e-01
8.83574605e-01 6.67696297e-01 4.01080042e-01 1.11860979e+00
2.40930989e-01 1.24172282e+00 -5.63998163e-01 -4.30201262e-01
6.12095952e-01 -6.71026647e-01 -6.94381118e-01 3.60912472e-01
6.49881244e-01 -7.96161413e-01 -1.03764200e+00 1.39214742e+00
5.92151344e-01 1.63203508e-01 7.71309733e-01 -5.32434225e-01
-1.35447645e+00 4.79767382e-01 -7.23964870e-01 3.76042604e-01
3.17677706e-01 5.02164960e-01 5.94725728e-01 -6.76854253e-01
7.77025461e-01 1.30117905e+00 9.02565956e-01 7.94008017e-01
-1.49018919e+00 -2.07642987e-01 8.44139755e-02 1.83924764e-01
-1.11092472e+00 -9.34324861e-01 4.35072809e-01 -3.73055011e-01
8.84138227e-01 8.63651335e-02 3.56285304e-01 1.36743093e+00
1.32735476e-01 7.28070199e-01 1.32035387e+00 -4.43284184e-01
3.09240460e-01 -2.01087922e-01 -1.43277779e-01 4.30501074e-01
2.67603278e-01 2.69008607e-01 -6.11502826e-01 -7.09728301e-02
1.01272368e+00 -7.21951947e-02 -9.24357772e-01 -5.14077365e-01
-1.34224749e+00 7.84647584e-01 4.73090768e-01 2.98388660e-01
-5.02475381e-01 5.32841563e-01 -4.51909006e-03 3.86608034e-01
5.42979181e-01 7.25705549e-02 -1.11874171e-01 -5.23098148e-02
-1.45703578e+00 2.04843938e-01 9.48757946e-01 8.45289707e-01
6.79315567e-01 4.33544070e-01 -4.64746319e-02 7.76464343e-01
4.38866079e-01 6.90954983e-01 7.18093932e-01 -1.47000694e+00
4.70995963e-01 -3.85409534e-01 3.78693044e-01 -8.55061591e-01
2.32141718e-01 -4.61115569e-01 -1.32953525e+00 3.17793608e-01
5.19233346e-01 -7.74427727e-02 -1.28755713e+00 1.65883195e+00
-2.91993976e-01 2.46454850e-01 1.93601474e-01 1.03296411e+00
1.45009220e-01 7.40509510e-01 -3.51068616e-01 -3.65055472e-01
8.83281589e-01 -8.23447645e-01 -7.74431646e-01 -5.25634587e-01
-6.20441400e-02 -8.70688736e-01 8.49717140e-01 5.99062026e-01
-1.50379860e+00 -3.77870977e-01 -1.19126439e+00 -2.31725872e-01
5.52712791e-02 -2.87179112e-01 4.24125046e-01 7.00738192e-01
-1.71294308e+00 1.09228885e+00 -8.90675426e-01 -3.55569363e-01
4.68997538e-01 1.98279902e-01 -1.63251504e-01 -6.51551008e-01
-9.37908232e-01 9.69214678e-01 -2.44762227e-01 2.35458955e-01
-1.28836465e+00 -5.75342417e-01 -8.56963515e-01 -7.93395266e-02
5.04342094e-03 -1.07385433e+00 1.42895412e+00 -1.03442490e+00
-1.49715638e+00 6.20125175e-01 -2.98719555e-01 -7.28252292e-01
8.43833387e-01 -3.32925320e-01 -1.54913083e-01 2.83696890e-01
7.39032105e-02 5.48103929e-01 1.48674142e+00 -1.58411634e+00
4.77377139e-02 -1.69145152e-01 -1.37822837e-01 1.58486903e-01
-1.33786306e-01 -1.89545587e-01 -5.00830770e-01 -8.95493269e-01
8.42541754e-02 -5.48334360e-01 -5.51153183e-01 2.04453737e-01
-3.31461132e-01 6.10498905e-01 4.54823524e-01 -9.68667328e-01
8.23834360e-01 -2.08439589e+00 4.81737643e-01 1.44330665e-01
3.76708150e-01 2.28794202e-01 -2.05725625e-01 4.71827745e-01
-1.45178288e-01 5.34667186e-02 -9.47941124e-01 -7.27033734e-01
3.99311334e-02 3.59316111e-01 -5.85798562e-01 8.98267925e-01
-1.01783365e-01 6.62520111e-01 -9.20402229e-01 -1.21407174e-01
1.10215418e-01 8.36603761e-01 -5.77275634e-01 2.00646326e-01
-1.30990863e-01 3.91610980e-01 1.11707181e-01 3.11818242e-01
1.01780605e+00 -2.96801805e-01 1.56242579e-01 -2.12240577e-01
2.09225386e-01 6.20467216e-02 -1.26389837e+00 1.99838996e+00
-5.35261035e-01 7.61295617e-01 6.38506174e-01 -1.00510597e+00
4.95123953e-01 3.73033851e-01 1.57882422e-01 -6.64472520e-01
-2.47807905e-01 3.14054191e-01 -1.60389572e-01 -4.73704904e-01
5.07302642e-01 -4.90806133e-01 4.97529298e-01 5.88610172e-01
2.21403584e-01 -4.62792635e-01 2.09203571e-01 4.90712553e-01
1.12203050e+00 2.31531247e-01 -2.62797344e-02 -2.17780098e-01
4.98119518e-02 -3.54021102e-01 2.62322724e-01 1.25389385e+00
-1.93108529e-01 1.20002317e+00 4.78894264e-02 -2.79488154e-02
-1.19494784e+00 -1.19871533e+00 5.87618016e-02 4.26938832e-01
2.46931866e-01 -1.99867308e-01 -9.45778608e-01 -3.08562547e-01
-4.32459153e-02 7.85276711e-01 -4.57442254e-01 -4.05382849e-02
-6.65778577e-01 -1.26591480e+00 5.07493258e-01 2.00851604e-01
5.40631771e-01 -7.99024880e-01 -1.25494152e-01 1.70444474e-01
-3.72827411e-01 -9.94466960e-01 -6.37417912e-01 1.30147532e-01
-9.62869465e-01 -9.58366990e-01 -1.35030317e+00 -6.99622750e-01
6.69134140e-01 4.41217154e-01 1.24880922e+00 9.47196186e-02
-1.70284972e-01 6.72542095e-01 2.02967003e-02 2.36589089e-01
-7.98137784e-01 -5.73121786e-01 6.19939603e-02 1.68868840e-01
-5.63294776e-02 -8.45673144e-01 -1.00230122e+00 -5.35740005e-03
-1.21856964e+00 -2.49615759e-01 5.69968402e-01 8.61847460e-01
2.35848725e-01 2.46241137e-01 9.67030451e-02 -5.17569661e-01
9.32142496e-01 -4.96042132e-01 -3.99283350e-01 -8.56137797e-02
-7.50137627e-01 2.15487704e-01 6.17393374e-01 -3.69017839e-01
-1.09944963e+00 -1.48734584e-01 -1.74308613e-01 -4.92379129e-01
-4.27220948e-02 2.49373987e-01 1.25565305e-01 -8.71093571e-02
8.89746130e-01 4.61102962e-01 1.80707157e-01 -8.95594716e-01
7.89685845e-01 5.27393103e-01 9.64780509e-01 -3.81938010e-01
1.13031256e+00 6.80243492e-01 -2.50377476e-01 -8.34289074e-01
-4.24570620e-01 -2.09391892e-01 -3.64165038e-01 1.01945341e-01
6.91411436e-01 -1.10141170e+00 -4.42166626e-01 9.28144515e-01
-1.27327168e+00 -5.71429312e-01 -7.05122709e-01 3.65423620e-01
-8.91336858e-01 1.06119215e+00 -1.21004617e+00 -8.80547404e-01
-7.99141526e-02 -1.18495202e+00 8.57056379e-01 6.01625536e-03
2.05568895e-01 -1.13204968e+00 3.78971696e-02 1.56373769e-01
8.79903674e-01 -3.05422574e-01 6.98242903e-01 2.32265443e-02
-8.68137896e-01 -3.67093757e-02 -3.17919791e-01 8.53885949e-01
2.53137648e-01 -3.36251557e-01 -1.05468607e+00 -4.60874557e-01
7.54256785e-01 -2.18521014e-01 1.57061636e+00 7.29638755e-01
8.82116437e-01 -5.20316124e-01 -1.07908761e-02 8.27376664e-01
1.47685599e+00 -1.48523062e-01 1.07365358e+00 3.03134531e-01
5.67979515e-01 3.13135326e-01 -1.79172739e-01 1.46924898e-01
3.28284889e-01 4.60036248e-01 5.51954448e-01 -2.11409703e-01
-6.75239503e-01 -1.77091643e-01 7.28611290e-01 8.28349173e-01
-1.00202225e-01 -5.17431378e-01 -6.29797697e-01 7.92873681e-01
-1.58357418e+00 -1.20373464e+00 -7.32151121e-02 2.29774475e+00
1.11181808e+00 -1.51541263e-01 -2.42224351e-01 6.11014403e-02
7.47968197e-01 4.90214467e-01 -4.66211170e-01 3.28143612e-02
-5.14934778e-01 2.26408094e-01 7.44897425e-01 1.12485635e+00
-1.05545104e+00 6.62640214e-01 7.64659357e+00 7.61677086e-01
-6.80047870e-01 1.97084382e-01 7.26435363e-01 -1.21318862e-01
-4.02283788e-01 7.98655972e-02 -2.46296033e-01 4.52412397e-01
9.85949159e-01 1.60413757e-01 1.16513002e+00 2.69389123e-01
3.04245502e-01 -1.59139886e-01 -9.87785757e-01 1.00867820e+00
3.20169508e-01 -1.17942321e+00 9.49160755e-02 1.00835957e-01
8.25424314e-01 2.14582652e-01 3.80957305e-01 -3.87254834e-01
9.17669654e-01 -1.16094303e+00 7.26346791e-01 7.97516108e-01
7.59101629e-01 -2.20960565e-02 4.99393672e-01 4.08135086e-01
-4.50617135e-01 -2.34866872e-01 -5.33542871e-01 9.91919264e-02
5.79775333e-01 9.32424605e-01 -2.58679926e-01 4.74717647e-01
7.12996304e-01 1.01323652e+00 -4.34698671e-01 1.19264984e+00
-1.99138865e-01 6.37646079e-01 -2.35686958e-01 6.52368426e-01
3.44266705e-02 -2.52544433e-01 6.59328282e-01 1.32829964e+00
7.39779949e-01 -2.70553946e-01 -3.51202071e-01 8.51399839e-01
-1.08438335e-01 -4.22123849e-01 -5.84646523e-01 1.77342594e-01
-9.48035046e-02 6.08780384e-01 -3.55568260e-01 -4.88698334e-01
-2.69954175e-01 1.55304849e+00 -2.92999409e-02 9.16808903e-01
-5.66420913e-01 -4.69300337e-03 6.09131932e-01 8.81499350e-02
6.85977042e-01 -4.77398098e-01 -1.70434505e-01 -1.65630472e+00
3.23974760e-04 -1.25406051e+00 -5.35723045e-02 -1.11038029e+00
-1.67842507e+00 5.98077834e-01 -1.37943342e-01 -1.13717425e+00
-4.32684243e-01 -5.28888583e-01 -4.44489688e-01 1.22260833e+00
-1.89535618e+00 -7.68323004e-01 -1.71276078e-01 7.87102222e-01
6.56425297e-01 4.15039808e-02 6.54006779e-01 2.40902081e-01
-2.03012630e-01 5.57455830e-02 5.82085073e-01 -1.04482003e-01
8.94086540e-01 -1.42732656e+00 8.09255898e-01 1.45946324e+00
2.63965756e-01 8.20632398e-01 1.09332550e+00 -6.80886626e-01
-1.44863975e+00 -7.14015841e-01 6.77023053e-01 -3.77052426e-01
6.75919533e-01 -1.27304658e-01 -8.94899428e-01 6.05612338e-01
7.30697095e-01 2.14893352e-02 9.11002308e-02 -4.75328505e-01
-5.50619185e-01 6.06029294e-03 -1.05276084e+00 6.15876436e-01
1.11857569e+00 -7.25326002e-01 -4.10540164e-01 5.89669585e-01
4.44255650e-01 -3.84240478e-01 -4.42902714e-01 -4.18409519e-02
1.77963406e-01 -1.33928645e+00 1.41990292e+00 -2.27979332e-01
6.04342401e-01 -3.78235728e-01 -5.23878336e-01 -1.52192402e+00
-3.49691421e-01 -1.02557182e+00 -4.68393177e-01 9.65808392e-01
1.53984800e-01 -4.78101641e-01 4.45958346e-01 5.32193482e-01
-5.35894595e-02 -2.44222414e-02 -9.41219211e-01 -7.73560882e-01
2.89519727e-01 -3.37759793e-01 2.14915946e-01 7.03236341e-01
-4.20976430e-01 -1.60241991e-01 -9.01004791e-01 1.40216023e-01
1.01963401e+00 -2.19766110e-01 4.27470386e-01 -7.59831429e-01
-8.97534668e-01 -4.21978474e-01 -6.75087646e-02 -1.85421062e+00
-9.92858335e-02 -7.48569548e-01 4.62015659e-01 -1.66521931e+00
1.42049164e-01 -1.98639825e-01 -6.25001565e-02 8.20791572e-02
-1.34348705e-01 5.75102806e-01 2.23773882e-01 6.38540745e-01
-1.77676290e-01 6.48990333e-01 1.25141406e+00 -4.46538985e-01
1.30548373e-01 -3.78510058e-02 -1.01677394e+00 6.48943901e-01
3.86730343e-01 -4.76282418e-01 -4.77311462e-01 -9.58121836e-01
7.66890347e-02 7.51036555e-02 6.64473534e-01 -8.87204230e-01
4.70888853e-01 1.62167728e-01 3.95552307e-01 1.79408893e-01
4.56934720e-01 -6.25032604e-01 1.62282974e-01 3.65586311e-01
-2.50768781e-01 -1.19197369e-02 -9.38243642e-02 9.56876576e-01
-1.20345697e-01 -5.13775945e-01 7.89198160e-01 -5.16689897e-01
-3.74478191e-01 1.22270197e-01 -6.25491619e-01 1.15107791e-02
2.62823522e-01 -1.30740523e-01 -5.98380506e-01 -1.05283499e+00
-9.04901505e-01 -4.07442003e-01 8.46160114e-01 1.17944114e-01
6.27322733e-01 -9.45614219e-01 -8.37473333e-01 3.48989576e-01
-5.56907952e-01 -1.00745015e-01 3.43760043e-01 7.67417848e-01
-7.57110059e-01 -7.18246698e-02 2.05319554e-01 -4.86760736e-01
-6.88582838e-01 7.98188508e-01 6.19064987e-01 -1.55648559e-01
-8.73314738e-01 7.90028095e-01 2.43089795e-01 -3.18042114e-02
1.43677860e-01 -9.25445482e-02 2.52893865e-01 -2.95789182e-01
6.32506907e-01 2.09396645e-01 -6.91844570e-03 -6.64993644e-01
-4.13933806e-02 6.23716950e-01 -9.74487066e-02 -4.53416467e-01
1.34284699e+00 -5.87298155e-01 -2.41672844e-01 6.44941553e-02
1.02957261e+00 3.68870705e-01 -1.72148764e+00 -1.12384297e-01
-3.10984015e-01 -7.13194609e-01 2.20582530e-01 -8.68037462e-01
-1.34115350e+00 1.02397752e+00 7.56764829e-01 3.56398433e-01
1.16620338e+00 -2.32367590e-01 6.73725128e-01 1.41143724e-01
1.13747545e-01 -7.47268975e-01 -6.48903102e-02 2.26074338e-01
1.03924549e+00 -1.01444793e+00 8.98667723e-02 -2.09664434e-01
-4.61624324e-01 9.62712765e-01 -2.52346359e-02 -3.11228037e-01
5.72795272e-01 3.65565747e-01 2.45513394e-01 6.84313253e-02
-4.83715296e-01 5.38282096e-02 -1.15983695e-01 1.15034032e+00
2.06071138e-01 -4.38235670e-01 1.87869504e-01 -4.15622592e-02
1.99361797e-02 1.40251657e-02 9.25428689e-01 8.48502159e-01
-2.32296601e-01 -1.11739933e+00 -5.68797708e-01 1.30593613e-01
-4.91042882e-01 -6.88498914e-01 -1.70171540e-02 4.12179291e-01
-2.67814964e-01 1.16433024e+00 -1.22791775e-01 2.06950344e-02
-2.63238866e-02 -5.10426424e-02 5.63573718e-01 -1.90730453e-01
-2.41491273e-01 3.08592290e-01 -1.98081031e-01 -5.41089773e-01
-6.97342277e-01 -6.65173769e-01 -5.46429217e-01 -5.64784467e-01
-1.20917581e-01 -6.00921884e-02 5.77272236e-01 7.67633498e-01
3.11998308e-01 2.84295082e-01 3.58793020e-01 -1.33680284e+00
-9.79650438e-01 -8.81631970e-01 -6.45252049e-01 4.85825568e-01
9.85016227e-01 -1.74233705e-01 -7.79555440e-01 4.73358124e-01] | [11.67971134185791, -2.368962287902832] |
1a9b4ff1-085e-41b6-841e-c2d7a4439904 | tensor-composition-net-for-visual | 2012.05473 | null | https://arxiv.org/abs/2012.05473v2 | https://arxiv.org/pdf/2012.05473v2.pdf | Tensor Composition Net for Visual Relationship Prediction | We present a novel Tensor Composition Net (TCN) to predict visual relationships in images. Visual Relationship Prediction (VRP) provides a more challenging test of image understanding than conventional image tagging and is difficult to learn due to a large label-space and incomplete annotation. The key idea of our TCN is to exploit the low-rank property of the visual relationship tensor, so as to leverage correlations within and across objects and relations and make a structured prediction of all visual relationships in an image. To show the effectiveness of our model, we first empirically compare our model with Multi-Label Image Classification (MLIC) methods, eXtreme Multi-label Classification (XMC) methods, and VRD methods. We then show that thanks to our tensor (de)composition layer, our model can predict visual relationships which have not been seen in the training dataset. We finally show our TCN's image-level visual relationship prediction provides a simple and efficient mechanism for relation-based image-retrieval even compared with VRD methods. | ['Yanwen Guo', 'Xueting Zhang', 'Timothy M. Hospedales', 'Yongxin Yang', 'Yuting Qiang'] | 2020-12-10 | null | null | null | null | ['multi-label-image-classification', 'extreme-multi-label-classification'] | ['computer-vision', 'methodology'] | [ 2.24464417e-01 4.84764725e-02 -5.82631826e-01 -3.33582968e-01
-2.55639970e-01 -7.25006223e-01 5.97253919e-01 3.49664003e-01
5.85328899e-02 -1.08940993e-02 1.58274338e-01 -3.59018117e-01
-3.68564785e-01 -4.02975738e-01 -6.23864651e-01 -3.22080880e-01
-1.30294710e-01 5.47687709e-01 2.00659186e-01 -8.78933668e-02
1.41076446e-01 5.68141401e-01 -1.52825272e+00 7.09881544e-01
1.40224457e-01 1.13132024e+00 5.88681102e-02 6.11690104e-01
-1.36655336e-02 1.41814268e+00 2.19302066e-02 -4.27155465e-01
7.40086362e-02 -5.91453612e-02 -1.32201374e+00 2.17459619e-01
9.82529104e-01 -3.24235827e-01 -4.78321701e-01 8.40405226e-01
8.13974941e-04 4.61477526e-02 8.89804482e-01 -1.90309942e+00
-8.50831330e-01 5.23269236e-01 -7.42919803e-01 6.04428053e-02
3.66394192e-01 -3.97039384e-01 1.55997908e+00 -8.89013946e-01
1.17438519e+00 1.25476027e+00 6.72029316e-01 4.17753249e-01
-1.40706658e+00 -5.06385863e-01 2.46016517e-01 4.57133919e-01
-1.31752086e+00 -4.88236398e-02 7.80740559e-01 -7.10661054e-01
8.52355599e-01 3.52778912e-01 7.02473760e-01 7.95705199e-01
9.11165252e-02 9.10541892e-01 1.18133807e+00 -2.68672824e-01
-3.43272597e-01 8.79947990e-02 3.21582079e-01 1.10153055e+00
-2.51083255e-01 -1.81263581e-01 -7.63667107e-01 -3.81364971e-02
6.04760766e-01 2.65870720e-01 -1.15202516e-01 -8.21048319e-01
-1.48865283e+00 5.16955912e-01 8.33082914e-01 3.45781863e-01
-7.18308166e-02 6.14555597e-01 3.69867295e-01 3.76365364e-01
5.11228740e-01 3.39295477e-01 -3.37052733e-01 2.76910573e-01
-6.91774666e-01 1.50554925e-02 6.45120561e-01 9.99663115e-01
8.56109023e-01 -5.76350570e-01 -2.28382140e-01 1.05187452e+00
3.29506665e-01 2.42621109e-01 1.14700019e-01 -1.26225507e+00
2.24004239e-01 7.44827986e-01 -3.22795689e-01 -1.42267406e+00
-4.84266996e-01 -3.46038878e-01 -8.05437088e-01 -3.14818546e-02
1.54400662e-01 6.24338448e-01 -6.61448061e-01 1.50136554e+00
1.76124975e-01 6.12239055e-02 -2.34808356e-01 8.10331047e-01
1.12766755e+00 2.78247654e-01 1.61038116e-01 -6.47828802e-02
1.47208941e+00 -1.18467689e+00 -5.24350822e-01 -8.39174762e-02
1.05201995e+00 -7.35064447e-01 8.71074200e-01 2.28053443e-02
-6.23690784e-01 -4.38254863e-01 -8.03157151e-01 -4.92660791e-01
-4.76125807e-01 -1.73446015e-02 1.19235718e+00 1.11241810e-01
-1.35389423e+00 4.96969789e-01 -3.43570352e-01 -4.83113289e-01
7.47690856e-01 3.61369818e-01 -7.24044502e-01 -4.69084352e-01
-7.16848373e-01 8.50738823e-01 1.77862957e-01 -1.39040515e-01
-9.71574783e-01 -6.41930461e-01 -8.07466686e-01 -1.29870698e-01
4.93304193e-01 -6.52139306e-01 7.88946211e-01 -6.63071990e-01
-5.79450607e-01 1.26800036e+00 -2.45208442e-01 -1.47042453e-01
1.75265148e-01 1.38813391e-01 -1.36667788e-01 6.08125329e-01
1.92234948e-01 1.23371267e+00 5.55640638e-01 -1.87237561e+00
-4.28737432e-01 -2.32944965e-01 4.07979876e-01 3.05004597e-01
-3.80709827e-01 2.75691524e-02 -6.59354031e-01 -3.15090507e-01
3.80770713e-01 -1.28768599e+00 -5.10412455e-02 4.05443221e-01
-5.88764548e-01 -3.86749417e-01 8.89359355e-01 -3.64690870e-01
8.55306089e-01 -2.08238721e+00 2.95431644e-01 3.55178535e-01
8.87352347e-01 -1.31025791e-01 -5.29439509e-01 3.35906059e-01
-3.06227714e-01 3.08022916e-01 2.38262504e-01 -4.65808243e-01
-1.32144377e-01 6.09392285e-01 -2.75306493e-01 2.72082388e-01
2.57294104e-02 1.24356592e+00 -1.15936303e+00 -1.04346871e+00
1.50288537e-01 4.46058184e-01 -4.22724575e-01 8.15351587e-03
-4.17842597e-01 3.59686226e-01 -2.74186134e-01 9.88168538e-01
4.46590304e-01 -1.03204346e+00 5.34454763e-01 -9.63125408e-01
1.49501532e-01 -1.65199056e-01 -6.62166536e-01 1.55710816e+00
-4.19832110e-01 8.07424009e-01 -5.72271407e-01 -9.07430649e-01
5.19304693e-01 1.02145210e-01 9.57309842e-01 -8.46093953e-01
2.36882400e-02 -9.62548479e-02 -3.20045918e-01 -4.93520260e-01
5.81479132e-01 1.98304415e-01 2.10631579e-01 6.79745734e-01
2.22362444e-01 2.06061658e-02 1.79298371e-01 8.07955325e-01
1.08567536e+00 3.19053590e-01 -7.20173540e-03 2.72201840e-02
3.70642066e-01 8.74013975e-02 2.68488199e-01 7.22044468e-01
-1.23175211e-01 3.97502244e-01 9.04044509e-01 -6.56260669e-01
-9.93907750e-01 -7.54049957e-01 1.32413954e-02 1.32349527e+00
5.37651777e-01 -8.41207206e-01 -5.94021492e-02 -1.05878210e+00
1.21169895e-01 -4.14422201e-03 -9.40321684e-01 5.73585182e-02
-3.71118933e-01 -4.38347310e-01 2.99581468e-01 3.65474075e-01
2.22139671e-01 -7.15229988e-01 6.33030608e-02 -4.08564091e-01
-4.89081144e-01 -1.65914381e+00 -4.22175258e-01 4.80763689e-02
-6.82028592e-01 -1.20587909e+00 -2.82763302e-01 -1.00562024e+00
8.69281232e-01 7.81820714e-01 1.44228041e+00 6.19637549e-01
-2.50953317e-01 8.16702247e-01 -4.25277054e-01 2.50823081e-01
-3.04967642e-01 -1.02237575e-01 -1.65467009e-01 7.58869424e-02
-1.34502966e-02 -5.60859799e-01 -5.29153824e-01 6.32191241e-01
-8.42277586e-01 4.03447777e-01 5.25944829e-01 8.54630113e-01
9.60521460e-01 -7.38763958e-02 -1.24398202e-01 -1.06755340e+00
1.82390854e-01 -3.91780734e-01 -4.31352198e-01 9.09246206e-01
-7.77837574e-01 1.49632707e-01 7.47113153e-02 -6.45394206e-01
-4.65584069e-01 2.69477814e-01 3.83770555e-01 -8.15337777e-01
2.07027689e-01 4.00256962e-01 3.50101799e-01 -6.42102599e-01
3.09686095e-01 -1.01451315e-01 3.27788270e-03 -3.52612942e-01
8.47203374e-01 1.60219774e-01 4.93854791e-01 -4.39169675e-01
8.12065840e-01 6.77955270e-01 7.16388583e-01 -3.37410539e-01
-1.53399658e+00 -8.22418571e-01 -1.06194496e+00 -5.84374011e-01
1.10514820e+00 -1.05084550e+00 -1.25108886e+00 -1.70475747e-02
-1.19666314e+00 -3.71511966e-01 6.29801229e-02 2.48489931e-01
-4.08000767e-01 5.01263797e-01 -7.71921396e-01 -3.92755300e-01
-5.21295965e-02 -9.64284301e-01 1.23444557e+00 -2.47847199e-01
-6.44873530e-02 -1.19649494e+00 -1.90110460e-01 8.47002029e-01
2.89801750e-02 2.22415164e-01 1.32903516e+00 -3.63476604e-01
-9.06760454e-01 1.47711173e-01 -8.67453933e-01 1.82415038e-01
-2.00964496e-01 -9.61640403e-02 -7.90768027e-01 -3.05384081e-02
-4.99909252e-01 -7.27615237e-01 1.03224206e+00 2.70178672e-02
1.29871166e+00 -1.49710074e-01 -6.89331472e-01 5.63870251e-01
1.68026137e+00 -3.10683608e-01 4.24047738e-01 1.69280395e-01
1.42531621e+00 7.94164121e-01 6.52967155e-01 -2.59074848e-02
1.00675893e+00 8.46014202e-01 7.44687140e-01 -2.40955174e-01
-5.62449396e-01 -2.83730119e-01 4.65800650e-02 9.62213457e-01
-1.30006015e-01 -2.54575789e-01 -9.59137917e-01 5.02454638e-01
-2.01558328e+00 -9.00586426e-01 -6.28193378e-01 1.78430533e+00
6.09852850e-01 -1.95302755e-01 8.23451392e-03 -1.85531363e-01
5.39285719e-01 1.46876156e-01 -2.59344131e-01 -1.65585458e-01
-2.73758352e-01 -1.87365204e-01 5.84861636e-01 3.27869833e-01
-1.15514970e+00 1.01823485e+00 7.09497976e+00 6.75297797e-01
-7.11553752e-01 2.50045091e-01 6.41349375e-01 1.69605777e-01
-4.70223695e-01 2.82990396e-01 -5.70930064e-01 -1.62105739e-01
4.24816102e-01 4.14080411e-01 4.43697810e-01 6.35191321e-01
-4.39694673e-01 -2.88771093e-02 -1.40699840e+00 1.24271369e+00
3.48847926e-01 -1.55771732e+00 3.71445060e-01 3.27235907e-01
7.39778221e-01 9.75024179e-02 2.25531161e-01 4.95847547e-03
4.22668278e-01 -9.78905559e-01 6.20775342e-01 5.67682385e-01
8.39125276e-01 -3.43909025e-01 6.10447168e-01 -1.84560686e-01
-1.35131419e+00 -7.40606636e-02 -3.26118469e-01 2.39404693e-01
-5.04002608e-02 5.93735456e-01 -8.17904592e-01 6.45492971e-01
5.38734317e-01 1.32463145e+00 -1.06871545e+00 6.23686552e-01
-1.37397438e-01 2.18162864e-01 -1.61452554e-02 2.09556043e-01
2.02030644e-01 -3.04241683e-02 1.63126305e-01 9.21247602e-01
-1.43409014e-01 -1.37310058e-01 4.32928652e-01 6.18066013e-01
-3.29610705e-01 7.70682842e-02 -7.46669233e-01 -1.03605717e-01
1.90352097e-01 1.68805766e+00 -1.00562298e+00 -2.49303967e-01
-5.65520406e-01 1.10415816e+00 6.08128011e-01 2.70869941e-01
-6.80944264e-01 2.65234470e-01 3.63905489e-01 -2.18830556e-01
1.89923197e-01 -3.11682701e-01 -1.39469013e-01 -1.30117822e+00
-1.42422587e-01 -5.98759413e-01 4.78982747e-01 -1.14972782e+00
-1.36236167e+00 6.67240679e-01 -7.95061812e-02 -1.30491376e+00
-6.23426437e-02 -6.79564416e-01 2.82305330e-01 3.08958381e-01
-1.68582153e+00 -2.01428318e+00 -4.62773234e-01 6.23975456e-01
3.49981338e-03 5.11898287e-02 9.01059508e-01 5.08091331e-01
-3.87522101e-01 4.32705343e-01 -1.84643015e-01 3.00583929e-01
6.52745247e-01 -1.31259179e+00 -1.36809483e-01 4.01105344e-01
6.06522322e-01 4.62365001e-01 3.13849896e-01 -5.28738081e-01
-1.33896804e+00 -1.05259383e+00 1.10046184e+00 -7.46769309e-01
9.54780698e-01 -5.14324427e-01 -5.94880581e-01 8.59881938e-01
-7.00900555e-02 6.10859454e-01 8.35795283e-01 4.39782768e-01
-1.21843040e+00 -1.74892828e-01 -6.60781145e-01 5.75228989e-01
1.45042765e+00 -1.07975173e+00 -6.65194318e-02 7.12647676e-01
9.54638779e-01 -1.43725961e-01 -1.28670788e+00 3.90721738e-01
9.03704882e-01 -8.31206381e-01 1.29373538e+00 -6.20709777e-01
7.86966264e-01 -3.57603878e-01 -4.66956317e-01 -8.89188468e-01
-3.94153297e-01 -1.18505925e-01 -6.42288476e-02 1.29147458e+00
3.45322609e-01 -1.98294222e-01 5.68796575e-01 5.25913656e-01
1.07163198e-01 -8.46309781e-01 -4.34004635e-01 -6.95807755e-01
-3.76165062e-01 -5.78318059e-01 2.80138224e-01 1.26542473e+00
-1.70872137e-01 5.86894751e-01 -6.34579539e-01 2.05334947e-01
8.48049998e-01 5.28188765e-01 5.00775337e-01 -1.29264355e+00
-3.68348360e-01 -2.59488434e-01 -6.62302434e-01 -1.02108335e+00
4.41192806e-01 -1.30793631e+00 -1.52565598e-01 -1.69743919e+00
8.04748952e-01 -8.89284253e-01 -5.25351286e-01 9.92282629e-01
1.36235654e-01 1.01683152e+00 5.25277197e-01 6.96321905e-01
-1.42378640e+00 1.38059974e-01 1.54792929e+00 -5.33559620e-01
3.49323988e-01 -4.74932462e-01 -5.22320151e-01 4.50041026e-01
1.09190591e-01 -5.43851852e-01 -5.12751937e-01 -4.27809626e-01
9.19791222e-01 1.97460473e-01 7.74655402e-01 -4.77596104e-01
2.76610136e-01 -3.77968736e-02 3.56171817e-01 -6.58895969e-01
3.59587282e-01 -8.62674534e-01 2.36192465e-01 2.70731926e-01
-6.74699366e-01 2.57384889e-02 -3.00616890e-01 5.84100723e-01
-2.61565059e-01 4.55633551e-02 3.59405726e-01 -9.27205756e-02
-9.51231956e-01 5.47821999e-01 -4.80218902e-02 -1.94326460e-01
8.06326985e-01 5.21020517e-02 -7.55814195e-01 -3.67778808e-01
-8.51625741e-01 4.20867920e-01 4.80779231e-01 4.81970400e-01
7.88874209e-01 -1.68762994e+00 -2.22372875e-01 -3.36982071e-01
7.49107599e-01 -3.25011462e-01 3.28201085e-01 9.47990477e-01
-3.79554629e-01 2.66336560e-01 -2.03898937e-01 -8.69164646e-01
-1.70767593e+00 8.86336267e-01 2.61174500e-01 -4.78310674e-01
-6.62257195e-01 8.44684064e-01 3.42840880e-01 -3.76191437e-01
1.56957880e-01 -1.23335138e-01 -4.10251439e-01 2.55751431e-01
2.49948829e-01 -5.16108535e-02 -1.61273316e-01 -9.51820135e-01
-5.79419017e-01 9.43874538e-01 -7.30966255e-02 4.71396372e-02
1.34974742e+00 -3.25380653e-01 -7.78713167e-01 5.76888978e-01
1.63335383e+00 -3.83145422e-01 -8.71582270e-01 -4.80152994e-01
1.00444496e-01 -4.91642892e-01 7.74929002e-02 -7.27000654e-01
-1.11308706e+00 7.35784948e-01 4.94381428e-01 2.45172635e-01
9.67422724e-01 5.49161375e-01 5.54310083e-01 3.58772397e-01
3.77000540e-01 -7.44974256e-01 7.10754514e-01 3.38331193e-01
6.59502506e-01 -1.32618976e+00 3.12610298e-01 -9.27600980e-01
-8.47001135e-01 1.04283702e+00 4.34001446e-01 2.61593521e-01
8.60840261e-01 -1.30331010e-01 8.85251164e-02 -8.04913342e-01
-1.00609243e+00 -5.09773791e-01 7.79495895e-01 6.51798368e-01
2.99438119e-01 -1.08445175e-01 1.60939083e-01 -3.58911008e-01
1.96162939e-01 -2.81415850e-01 3.74343306e-01 5.90006173e-01
1.07688218e-01 -1.35942996e+00 8.93539712e-02 5.44772744e-01
-3.62219959e-01 -1.96138471e-01 -4.63314176e-01 5.33630669e-01
3.21486622e-01 8.99729311e-01 8.61116648e-02 -8.19076180e-01
-1.26113728e-01 -1.96677491e-01 7.25142419e-01 -5.21810651e-01
-4.29078490e-01 -1.59608483e-01 2.42268234e-01 -8.95475090e-01
-1.11487651e+00 -4.44786012e-01 -1.00094974e+00 -1.08761467e-01
-2.93799460e-01 -2.71275848e-01 7.65593350e-01 9.92017746e-01
2.86100328e-01 4.72285002e-01 6.70401216e-01 -5.42875290e-01
2.04036847e-01 -5.05777180e-01 -5.50225735e-01 9.47580874e-01
3.42047423e-01 -8.12015116e-01 -1.21148139e-01 1.90215990e-01] | [10.266424179077148, 1.6572456359863281] |
9e0982bf-2cd7-4d54-8654-9c1671422932 | enhancing-safe-exploration-using-safety-state | 2206.02675 | null | https://arxiv.org/abs/2206.02675v2 | https://arxiv.org/pdf/2206.02675v2.pdf | Effects of Safety State Augmentation on Safe Exploration | Safe exploration is a challenging and important problem in model-free reinforcement learning (RL). Often the safety cost is sparse and unknown, which unavoidably leads to constraint violations -- a phenomenon ideally to be avoided in safety-critical applications. We tackle this problem by augmenting the state-space with a safety state, which is nonnegative if and only if the constraint is satisfied. The value of this state also serves as a distance toward constraint violation, while its initial value indicates the available safety budget. This idea allows us to derive policies for scheduling the safety budget during training. We call our approach Simmer (Safe policy IMproveMEnt for RL) to reflect the careful nature of these schedules. We apply this idea to two safe RL problems: RL with constraints imposed on an average cost, and RL with constraints imposed on a cost with probability one. Our experiments suggest that "simmering, a safe algorithm can improve safety during training for both settings. We further show that Simmer can stabilize training and improve the performance of safe RL with average constraints. | ['Haitham Bou Ammar', 'Jun Wang', 'Alexander I. Cowen-Rivers', 'Aivar Sootla'] | 2022-06-06 | null | null | null | null | ['safe-exploration'] | ['robots'] | [ 2.40427092e-01 3.65868151e-01 -8.46633554e-01 3.18657956e-03
-7.12762654e-01 -6.95620418e-01 3.10729086e-01 3.15389901e-01
-7.85887539e-01 1.20672464e+00 3.43786250e-03 -5.65949202e-01
-3.02976012e-01 -5.20916879e-01 -9.18766856e-01 -9.53587413e-01
-4.78203714e-01 3.01384300e-01 1.84828550e-01 -3.29556048e-01
2.93905288e-01 3.04008693e-01 -1.38007748e+00 -1.96220130e-01
9.14660931e-01 6.16796911e-01 6.20058514e-02 3.66646856e-01
4.97143418e-01 6.70282722e-01 -5.77895045e-01 2.56103575e-01
5.20101845e-01 -3.70447487e-01 -8.29923928e-01 -1.46208536e-02
-1.13480121e-01 -2.39380240e-01 -2.51080981e-03 1.19525003e+00
1.84993163e-01 7.07567811e-01 2.27326065e-01 -1.59820843e+00
2.65860051e-01 7.63379574e-01 -6.41353250e-01 6.29479885e-02
3.08505803e-01 4.06206310e-01 9.40500855e-01 2.36821249e-01
4.40221190e-01 8.92292917e-01 3.94086570e-01 1.00415397e+00
-1.31022692e+00 -6.72187328e-01 6.57181025e-01 -2.68861324e-01
-1.09524882e+00 -3.69681627e-01 5.67428470e-01 -2.96222419e-01
7.30268002e-01 3.28270674e-01 6.25029385e-01 9.65356648e-01
2.74509639e-01 6.69823945e-01 1.07690418e+00 -4.94910866e-01
8.55939507e-01 5.53185157e-02 -8.30159485e-02 5.85354507e-01
3.06494087e-01 6.36907518e-01 -5.11589885e-01 -2.40090579e-01
6.38079524e-01 -2.35879973e-01 -2.89131522e-01 -6.53702796e-01
-8.63313556e-01 9.88400340e-01 1.26865447e-01 -8.63871798e-02
-1.09025195e-01 5.27874887e-01 3.58926862e-01 6.09668374e-01
-6.13814183e-02 8.34259033e-01 -4.41208452e-01 -4.71977204e-01
-5.91816247e-01 4.92065072e-01 8.45732868e-01 8.39554965e-01
3.27739060e-01 3.44494641e-01 -1.93065330e-01 4.10298377e-01
6.72817081e-02 3.25266123e-01 1.19371131e-01 -1.21186841e+00
5.07792056e-01 2.08041295e-01 7.40038097e-01 -2.26503894e-01
-3.18060786e-01 -4.34330046e-01 -4.68997896e-01 6.59665346e-01
4.88241643e-01 -5.98017633e-01 -7.96015143e-01 2.38835835e+00
3.90156895e-01 2.82036632e-01 -4.42843465e-03 7.79550135e-01
-3.04463178e-01 6.48617506e-01 1.44969657e-01 -7.89876640e-01
8.84143174e-01 -5.43829560e-01 -6.32308304e-01 -3.36341321e-01
7.51326263e-01 -1.63246065e-01 1.25958133e+00 7.37979233e-01
-1.04196417e+00 3.79577018e-02 -1.08881617e+00 8.21141183e-01
1.30239159e-01 -3.44839871e-01 6.18474007e-01 6.52715445e-01
-8.45112443e-01 8.75737786e-01 -1.00831985e+00 7.20346496e-02
1.39980495e-01 6.52514577e-01 -1.96162760e-01 2.13323399e-01
-1.20352721e+00 1.12349582e+00 4.94374424e-01 -2.80726135e-01
-1.19166553e+00 -6.09566927e-01 -8.75959694e-01 6.85162330e-03
1.06974566e+00 -3.27204496e-01 1.57129085e+00 -5.87976038e-01
-1.39768422e+00 2.74831921e-01 2.58853026e-02 -7.68712103e-01
6.67401493e-01 -3.50609303e-01 -2.51602195e-02 -1.76708296e-01
-1.66041534e-02 4.67980981e-01 1.01892412e+00 -1.29703689e+00
-6.40052557e-01 -4.51355167e-02 4.48103487e-01 4.04625118e-01
-1.36509612e-01 -7.72154555e-02 -2.50227470e-02 -3.72692525e-01
-3.28084648e-01 -1.30595660e+00 -7.94736624e-01 -1.84138164e-01
-2.64547914e-01 -1.06741324e-01 4.99448717e-01 -2.11121216e-02
1.46739995e+00 -2.00351000e+00 1.94737837e-01 4.87349898e-01
-2.52468050e-01 2.15251252e-01 -1.31245390e-01 2.49722719e-01
-1.48867264e-01 1.40242755e-01 -4.63453859e-01 -4.87662386e-03
-4.00668681e-02 6.30533814e-01 -5.56030869e-01 5.79032481e-01
8.88387393e-03 4.02128160e-01 -1.19816291e+00 -2.60273218e-01
1.56343520e-01 -9.39057320e-02 -9.57238495e-01 3.76055390e-01
-6.89141572e-01 6.42938137e-01 -6.35783255e-01 1.63363993e-01
2.27579057e-01 1.13824330e-01 3.58953536e-01 5.38143694e-01
-2.27356464e-01 1.82450950e-01 -1.50485384e+00 1.33116794e+00
-4.63958532e-01 -1.64957326e-02 2.18078896e-01 -8.88694942e-01
5.50859809e-01 2.13499546e-01 6.43364668e-01 -4.67562705e-01
1.92803413e-01 -1.80693969e-01 2.72096712e-02 -2.66286403e-01
3.01750511e-01 -3.73410255e-01 -3.42668355e-01 4.94830519e-01
-5.20510316e-01 -2.85646379e-01 2.42546350e-01 8.10808688e-02
1.11005867e+00 3.55902135e-01 3.65693212e-01 -4.73524630e-01
3.62439126e-01 -2.31421575e-01 9.95932639e-01 9.03795600e-01
-2.62811929e-01 -8.12035501e-02 1.17207253e+00 -1.55712545e-01
-6.50969982e-01 -7.66501248e-01 9.63443369e-02 9.67419386e-01
9.78818387e-02 -3.98532271e-01 -6.32999778e-01 -1.00475550e+00
2.21325144e-01 9.45489407e-01 -7.25991726e-01 -6.44617021e-01
-6.18930280e-01 -3.66855830e-01 1.29989728e-01 4.74516809e-01
1.43041596e-01 -1.01568794e+00 -1.11216700e+00 1.76764280e-01
1.93644434e-01 -5.91725588e-01 -6.81435525e-01 9.40559804e-01
-7.64538229e-01 -9.56354558e-01 -3.11547965e-01 -3.80060345e-01
8.56839597e-01 -1.41345233e-01 8.93043816e-01 1.24389209e-01
1.30271232e-02 1.46452516e-01 -5.18354364e-02 -3.14284563e-01
-3.82588148e-01 -1.62072495e-01 4.37306762e-01 -4.72558469e-01
-3.04989100e-01 -3.11258107e-01 -2.35670641e-01 2.54727840e-01
-9.94444191e-01 -1.90716028e-01 1.38852432e-01 9.97796416e-01
6.15888894e-01 2.34241813e-01 6.89191759e-01 -8.54215503e-01
7.63297856e-01 -4.36187923e-01 -1.13967133e+00 3.06529403e-01
-7.06320405e-01 7.42766917e-01 8.19034696e-01 -6.85573459e-01
-6.75250888e-01 3.08393031e-01 -6.50364235e-02 -6.99644268e-01
1.13922926e-02 3.61249804e-01 -1.38667822e-01 -2.49930501e-01
6.58533096e-01 1.46578010e-02 9.22015235e-02 -9.18489546e-02
2.00609073e-01 1.27672181e-01 2.89612532e-01 -1.23889422e+00
7.40240693e-01 -5.41024879e-02 3.12843949e-01 -4.61352736e-01
-9.94747281e-01 -2.70400554e-01 -6.05211109e-02 -1.53253168e-01
2.97379613e-01 -6.39496982e-01 -1.15657341e+00 -9.80995670e-02
-5.41017056e-01 -9.34932768e-01 -7.10858941e-01 2.87363380e-01
-9.23633754e-01 2.86382347e-01 -2.02676043e-01 -1.31612206e+00
1.32798791e-01 -1.25186431e+00 6.34387255e-01 2.26693541e-01
-3.67589593e-01 -8.58178258e-01 2.23394200e-01 -1.40473604e-01
1.72894478e-01 4.66067225e-01 7.16773272e-01 -4.33660895e-01
-2.82663792e-01 1.42639056e-01 4.75636125e-01 2.53823549e-01
3.01264897e-02 -2.40821779e-01 -4.80732083e-01 -7.22874284e-01
7.47413039e-02 -5.50121903e-01 6.35150909e-01 3.89436632e-01
1.35457289e+00 -9.10124123e-01 -2.35715374e-01 3.14080328e-01
1.34095883e+00 4.79902625e-01 1.43219620e-01 3.42884451e-01
2.70095259e-01 3.78213942e-01 1.10588527e+00 1.04546022e+00
-1.01872839e-01 7.18152821e-01 7.30461538e-01 2.06121817e-01
6.46731973e-01 -2.88895100e-01 4.56111103e-01 3.24344449e-02
2.16592178e-01 -6.32123426e-02 -7.26870596e-01 4.53613430e-01
-2.05416965e+00 -9.12849903e-01 4.91071105e-01 2.84946108e+00
1.40044129e+00 6.72658503e-01 3.24418843e-01 3.07122767e-01
3.57402831e-01 9.44094732e-02 -8.35102797e-01 -8.23509157e-01
4.75939840e-01 2.52994925e-01 7.86339641e-01 9.11031842e-01
-9.43493366e-01 8.22309971e-01 6.43016863e+00 6.69154584e-01
-1.01070440e+00 -3.26203436e-01 4.25709218e-01 -3.73997867e-01
-3.92553002e-01 1.61036462e-01 -8.46714854e-01 5.93734860e-01
1.00234902e+00 -4.76121873e-01 7.22809255e-01 1.06617975e+00
3.13731343e-01 -4.49071288e-01 -1.35510337e+00 3.49775910e-01
-5.08049488e-01 -9.11191344e-01 -3.40810508e-01 -1.33918189e-02
8.17453563e-01 -4.15376663e-01 1.07901677e-01 5.50454795e-01
8.50087523e-01 -9.34308529e-01 6.80277824e-01 6.80440739e-02
8.21875572e-01 -1.43685317e+00 3.31543028e-01 7.07931995e-01
-1.03193235e+00 -2.89434433e-01 2.61275005e-02 -1.45909026e-01
1.14440238e-02 2.32228488e-01 -6.48104250e-01 1.71223789e-01
2.81551778e-01 4.41461414e-01 5.67260310e-02 1.02225029e+00
-4.92204815e-01 4.80965704e-01 -4.74532843e-01 -5.67900203e-02
4.46009129e-01 -2.45512053e-01 5.52688837e-01 7.05608547e-01
-1.00328982e-01 1.33548498e-01 7.02191949e-01 6.37095809e-01
3.64266127e-01 -3.79891574e-01 -8.24048460e-01 -1.43697681e-02
5.87445617e-01 8.68961811e-01 -5.42449653e-01 -4.21020277e-02
3.57069373e-02 3.65953416e-01 3.13060373e-01 2.38209039e-01
-1.04795456e+00 -2.43405849e-01 9.69777644e-01 -6.41506612e-02
1.87230781e-02 -2.16512650e-01 -2.78836250e-01 -8.70263577e-01
-2.17644900e-01 -9.95040655e-01 7.08921075e-01 -2.38005638e-01
-7.21720397e-01 2.75722057e-01 3.01768512e-01 -1.24176621e+00
-5.66140175e-01 -1.05300687e-01 -5.91604292e-01 6.60989285e-01
-1.42359269e+00 -3.93420577e-01 2.99823314e-01 7.74147213e-01
4.63278532e-01 2.13383675e-01 6.50395691e-01 -2.03502506e-01
-8.24440241e-01 5.78066230e-01 -3.85631174e-02 -5.44291139e-01
5.28221548e-01 -1.33690679e+00 -1.67142730e-02 9.60215807e-01
-3.27829808e-01 7.15611100e-01 1.21038389e+00 -7.58011818e-01
-1.33091176e+00 -1.06353104e+00 4.42373633e-01 -2.59329170e-01
6.69073939e-01 -2.05675095e-01 -8.41063559e-01 6.78820252e-01
-1.13432087e-01 -7.75407106e-02 2.80952662e-01 1.70995489e-01
-8.09383839e-02 -1.02018364e-01 -1.21529627e+00 1.00614095e+00
9.40657318e-01 -2.39380933e-02 -6.29792392e-01 4.87468779e-01
9.11522448e-01 -6.12332404e-01 -6.20089471e-01 4.01606113e-01
1.25956982e-01 -6.33660316e-01 6.47240818e-01 -9.88521695e-01
-1.20565258e-02 -3.81106466e-01 2.57101595e-01 -1.57755578e+00
-1.35698631e-01 -1.21777225e+00 -4.49357092e-01 6.05111957e-01
2.95666754e-01 -4.99340326e-01 1.00321186e+00 8.82448196e-01
-1.93115637e-01 -9.62605834e-01 -1.21355247e+00 -1.15846848e+00
2.04737321e-01 -3.48564476e-01 5.61309576e-01 8.18994939e-01
5.18956721e-01 -1.47336572e-01 -6.73032343e-01 4.98778634e-02
6.00720286e-01 -1.33617572e-03 4.21012431e-01 -8.98772597e-01
-5.98419487e-01 -4.06936616e-01 3.44858527e-01 -6.65928543e-01
5.31274259e-01 -4.17569965e-01 4.84485418e-01 -9.88848925e-01
-6.35154769e-02 -7.53032804e-01 -4.18738812e-01 9.01285172e-01
-1.76400483e-01 -5.55045187e-01 1.39612913e-01 -1.99631855e-01
-7.36667275e-01 6.04839206e-01 1.19837618e+00 2.16414019e-01
-7.10677028e-01 2.68692970e-01 -6.25894129e-01 5.52422822e-01
1.06979084e+00 -4.92380291e-01 -7.79810309e-01 1.50861129e-01
3.30922663e-01 5.94157457e-01 -1.40349567e-01 -9.59394693e-01
-1.56437780e-03 -1.05551386e+00 -2.09008306e-01 -2.78696299e-01
1.75270736e-01 -1.05035460e+00 1.79480150e-01 1.01752055e+00
-9.25617814e-01 -2.82589775e-02 4.38176095e-01 5.83507299e-01
2.26529241e-01 -3.37369740e-01 1.08948088e+00 -9.60072316e-03
-3.95810097e-01 2.05755770e-01 -3.88537318e-01 2.82074541e-01
1.40769374e+00 5.18234111e-02 -8.70444700e-02 -4.29072559e-01
-5.75054646e-01 8.72494221e-01 4.35312092e-01 3.39377880e-01
4.85291690e-01 -1.08260000e+00 -1.55414268e-01 4.46643203e-01
3.31047550e-02 1.22485779e-01 -2.01162398e-01 6.94107831e-01
8.82959291e-02 2.35287637e-01 -2.96215355e-01 -2.25432947e-01
-1.16165578e+00 8.14315319e-01 4.15981919e-01 -4.82982248e-01
-4.41512197e-01 8.30357969e-01 -6.22703806e-02 2.57451888e-02
7.65345871e-01 -4.43013698e-01 -9.79932845e-02 -2.07645014e-01
4.07681793e-01 2.05252185e-01 -1.01494454e-01 -7.83772096e-02
-3.21920574e-01 3.25123936e-01 -1.32849097e-01 -3.51256937e-01
1.04749429e+00 1.28124222e-01 3.95206094e-01 2.61168450e-01
5.27301371e-01 -1.06209360e-01 -1.77272296e+00 8.48440528e-02
1.14352942e-01 -4.71495062e-01 2.17202287e-02 -8.02478492e-01
-9.28590536e-01 4.24673617e-01 2.30062500e-01 7.53641501e-02
1.02350390e+00 -4.76267815e-01 4.46818382e-01 5.05172074e-01
7.37888455e-01 -1.28712559e+00 9.83118862e-02 6.09688103e-01
7.22175837e-01 -9.98583972e-01 2.88924426e-02 3.15527208e-02
-9.90252912e-01 6.88057125e-01 1.02857828e+00 -2.27286592e-01
3.37112248e-01 6.76137686e-01 -4.58507955e-01 4.11266088e-01
-1.10850132e+00 -1.05669729e-01 -9.05631036e-02 4.55429912e-01
-5.21570817e-02 5.44699766e-02 -3.42869282e-01 4.15139437e-01
-5.06376848e-02 -9.27443733e-04 6.23987079e-01 1.45547307e+00
-8.03482413e-01 -1.34671903e+00 -3.66248876e-01 3.33042294e-01
-4.19328988e-01 2.47243509e-01 -1.06110595e-01 8.19086373e-01
-1.11160859e-01 8.51751924e-01 -1.50782809e-01 -2.65925616e-01
2.82979488e-01 -3.10754627e-02 5.12525439e-01 -7.17728734e-01
-6.84165657e-01 5.72208203e-02 3.91512185e-01 -9.21496630e-01
1.25271156e-01 -5.75231910e-01 -1.61197579e+00 -2.51670301e-01
-3.76821786e-01 5.36109507e-01 3.13201785e-01 8.63176525e-01
-7.32880086e-02 5.79304278e-01 1.00144887e+00 -3.11699748e-01
-1.32519102e+00 -2.35699996e-01 -7.32499182e-01 -3.38273938e-03
7.53365397e-01 -9.16304946e-01 -5.40531814e-01 -2.60350436e-01] | [4.510612487792969, 2.1386630535125732] |
1a28d488-f4f2-435a-beef-618fe115335c | a-novel-speech-driven-lip-sync-model-with-cnn | 2205.00916 | null | https://arxiv.org/abs/2205.00916v1 | https://arxiv.org/pdf/2205.00916v1.pdf | A Novel Speech-Driven Lip-Sync Model with CNN and LSTM | Generating synchronized and natural lip movement with speech is one of the most important tasks in creating realistic virtual characters. In this paper, we present a combined deep neural network of one-dimensional convolutions and LSTM to generate vertex displacement of a 3D template face model from variable-length speech input. The motion of the lower part of the face, which is represented by the vertex movement of 3D lip shapes, is consistent with the input speech. In order to enhance the robustness of the network to different sound signals, we adapt a trained speech recognition model to extract speech feature, and a velocity loss term is adopted to reduce the jitter of generated facial animation. We recorded a series of videos of a Chinese adult speaking Mandarin and created a new speech-animation dataset to compensate the lack of such public data. Qualitative and quantitative evaluations indicate that our model is able to generate smooth and natural lip movements synchronized with speech. | ['Shiguo Lian', 'Kai Wang', 'Xiang Wang', 'Xiaohong Li'] | 2022-05-02 | null | null | null | null | ['face-model'] | ['computer-vision'] | [ 3.59006077e-02 3.10726404e-01 2.12377116e-01 -2.28959873e-01
-3.35912734e-01 -2.81397969e-01 5.45878768e-01 -9.77290630e-01
-1.88876063e-01 5.05780160e-01 2.27644831e-01 -4.93507320e-03
4.86172885e-01 -4.50888604e-01 -9.62261200e-01 -6.19289875e-01
6.10354766e-02 -1.08394206e-01 3.56430933e-02 -1.28694326e-01
6.44342154e-02 9.18330312e-01 -1.82678103e+00 3.58780086e-01
4.15269673e-01 8.57038319e-01 3.89937401e-01 8.28992903e-01
-2.59453535e-01 2.54412115e-01 -9.38484669e-01 -2.66289920e-01
1.52752757e-01 -5.94130218e-01 -4.29646313e-01 1.72281593e-01
3.79506975e-01 -4.27548796e-01 -4.56943750e-01 9.10640895e-01
9.45343375e-01 7.34444559e-02 5.48007071e-01 -1.37203634e+00
-4.92657632e-01 5.28959215e-01 -2.41512403e-01 -3.10917675e-01
3.99205238e-01 4.23231035e-01 1.87544078e-01 -6.86581671e-01
9.53586936e-01 1.51912153e+00 5.85291922e-01 1.25207114e+00
-9.37101543e-01 -7.89301395e-01 -1.35966942e-01 -4.45341738e-03
-1.43126142e+00 -9.99246538e-01 1.12967861e+00 -2.52467275e-01
6.44342184e-01 2.68517733e-01 8.65303040e-01 1.50665104e+00
6.29412383e-02 5.57429314e-01 6.05933249e-01 -4.77465391e-01
-3.44179012e-02 1.63322568e-01 -6.79550529e-01 5.03774703e-01
-5.31377852e-01 3.10726464e-01 -4.07732010e-01 -3.28066805e-03
1.09605145e+00 -5.83468378e-01 -3.92584831e-01 -4.75186901e-03
-1.10469759e+00 5.13279438e-01 -6.76328968e-03 4.99726325e-01
-2.55343527e-01 2.48531327e-01 2.83829689e-01 1.89380674e-03
3.83435667e-01 -2.33700663e-01 -1.51893944e-01 -4.48525339e-01
-9.57692206e-01 2.18850970e-01 6.44373417e-01 8.28281045e-01
8.72444883e-02 6.78332806e-01 -1.03700243e-01 8.62101674e-01
4.69212651e-01 8.55831623e-01 7.55563915e-01 -1.13894582e+00
1.57390952e-01 1.18892968e-01 -8.19305629e-02 -7.97484517e-01
-2.97557414e-01 2.54099101e-01 -7.65774548e-01 5.82221746e-01
3.46758664e-01 -4.44986105e-01 -9.35991228e-01 2.00264311e+00
3.58774275e-01 4.87023741e-01 1.08120076e-01 9.79511082e-01
1.10142267e+00 7.18501031e-01 -1.46993726e-01 -5.25672972e-01
9.74520802e-01 -4.67181742e-01 -1.14365673e+00 2.12752089e-01
1.11753419e-01 -9.12569940e-01 1.07957458e+00 -7.85782784e-02
-1.41772759e+00 -9.26753998e-01 -7.19198823e-01 9.88881662e-02
-6.40155328e-03 2.67039329e-01 4.26142812e-02 6.75726831e-01
-1.17465222e+00 5.59802473e-01 -6.36202931e-01 7.30614439e-02
1.40162900e-01 2.34992236e-01 -5.14000535e-01 6.42751515e-01
-1.31566417e+00 5.36163330e-01 -3.06267738e-01 1.48629040e-01
-6.02387190e-01 -5.82872212e-01 -1.06894565e+00 -4.14786264e-02
-2.04070583e-01 -4.40654516e-01 1.32857776e+00 -1.24074543e+00
-2.44146252e+00 8.18599284e-01 -4.32358295e-01 -2.34589949e-01
7.33340085e-01 2.43473962e-01 -4.59484518e-01 1.25827432e-01
-3.14619482e-01 9.67270076e-01 1.33632910e+00 -1.23936713e+00
-5.40195964e-02 -3.03223878e-02 -5.43608546e-01 5.88111579e-02
-1.02631822e-02 2.39729881e-01 -4.52818215e-01 -8.28136802e-01
-1.48478642e-01 -9.90602076e-01 2.74782062e-01 3.53320003e-01
-2.63498753e-01 -1.45066440e-01 1.08264923e+00 -8.33101511e-01
9.52635169e-01 -2.44422555e+00 -1.82011202e-02 -3.04807015e-02
-2.82938659e-01 4.97158825e-01 -3.16724032e-01 1.08285509e-01
-1.87943712e-01 1.89992785e-01 -1.74081758e-01 -7.56343782e-01
-1.44859359e-01 4.99074766e-03 -3.88125867e-01 5.16541123e-01
1.27116308e-01 7.39821732e-01 -4.66266066e-01 -5.01516342e-01
1.50769815e-01 1.14052618e+00 -2.38464415e-01 2.77175277e-01
-3.18555266e-01 5.56729198e-01 -4.41795290e-02 4.09636468e-01
9.07249272e-01 4.19362068e-01 -1.36032730e-01 -1.09132417e-01
-2.35146061e-01 1.73528045e-01 -1.19648182e+00 1.61577988e+00
-5.28779566e-01 1.01687205e+00 4.60126042e-01 -4.08497840e-01
1.08901131e+00 7.45245159e-01 4.25624609e-01 -6.71398163e-01
3.54752630e-01 1.33085251e-02 4.94087785e-02 -8.90004694e-01
1.43242553e-01 3.22848484e-02 3.49264979e-01 3.61626744e-01
-2.66851991e-01 -6.14441097e-01 -2.25414068e-01 -2.71604031e-01
3.93819064e-01 2.28153706e-01 -2.70892113e-01 1.78034291e-01
5.94274223e-01 -6.27048969e-01 4.46733981e-01 3.88820693e-02
-3.21648151e-01 8.59466553e-01 4.74847794e-01 -2.17618987e-01
-1.07409430e+00 -8.38388681e-01 7.05243796e-02 5.58422029e-01
-5.75353205e-02 -1.76314060e-02 -1.18851840e+00 -2.94907689e-01
-9.70091298e-02 6.71552420e-01 -4.89232451e-01 2.93215923e-02
-9.17561233e-01 1.05237968e-01 8.96206498e-01 2.67351687e-01
3.85868609e-01 -1.68945003e+00 -5.73465943e-01 1.37678146e-01
-2.28726387e-01 -1.18618548e+00 -9.09276366e-01 -5.31182587e-01
-3.90782148e-01 -7.86910176e-01 -1.11321557e+00 -1.11762035e+00
4.83991444e-01 -1.64603040e-01 4.97719228e-01 -8.67573991e-02
-2.87187338e-01 -1.93038881e-02 1.89785231e-02 -4.43434596e-01
-1.15137851e+00 -3.66428792e-01 4.49533433e-01 1.77950740e-01
-2.16781721e-01 -4.98635322e-01 -2.91316420e-01 4.02158648e-01
-7.63904154e-01 2.49622911e-01 2.94963531e-02 5.56758881e-01
2.55047351e-01 -3.29281181e-01 6.18434727e-01 9.59602445e-02
8.48510981e-01 1.54230595e-01 -6.74653530e-01 -1.25302121e-01
1.88386291e-01 -1.26299366e-01 6.07062399e-01 -1.08377421e+00
-1.10340452e+00 2.91463196e-01 -6.31406844e-01 -7.90724218e-01
-3.55908096e-01 -2.10610271e-01 -3.54823947e-01 -8.16982985e-02
4.12153214e-01 3.14339846e-01 7.34939516e-01 -2.39606231e-01
3.32115501e-01 1.06713736e+00 5.72062135e-01 -1.04260996e-01
5.70104659e-01 4.08022165e-01 -3.06132250e-03 -1.43331015e+00
2.82508135e-01 3.61880958e-01 -4.00371730e-01 -6.92701697e-01
7.81491101e-01 -5.03100216e-01 -1.27193654e+00 1.15013337e+00
-1.66127908e+00 -7.08649993e-01 -2.55237281e-01 5.33183455e-01
-8.94663930e-01 4.24128860e-01 -6.56634510e-01 -9.14828777e-01
-5.56386173e-01 -1.46480477e+00 1.06257999e+00 3.46145302e-01
-1.34522632e-01 -5.65362453e-01 9.31582004e-02 -9.07774828e-03
5.27340055e-01 2.48484284e-01 5.83102465e-01 -3.34380083e-02
-1.85545355e-01 -1.02209061e-01 2.53165603e-01 5.33303142e-01
5.09703577e-01 7.77769029e-01 -1.20879793e+00 5.36437007e-03
-4.08420786e-02 -6.39353022e-02 3.56555432e-01 8.05022597e-01
9.37539637e-01 -4.80015457e-01 -1.58491239e-01 6.76094532e-01
5.01801431e-01 5.46046376e-01 7.78245866e-01 -3.52766693e-01
3.21763307e-01 7.78901637e-01 2.08137229e-01 3.40558797e-01
5.46080433e-02 8.91046941e-01 2.36844927e-01 -4.38570790e-03
-7.74740219e-01 -2.29876310e-01 6.12840235e-01 5.73440492e-01
-1.80245817e-01 -1.15326025e-01 -4.71756518e-01 2.14195624e-01
-1.18185556e+00 -1.28537202e+00 1.81963965e-01 2.10831380e+00
8.63402188e-01 1.11903936e-01 2.47072384e-01 2.40569025e-01
1.18089771e+00 7.07758814e-02 -3.61779928e-01 -6.57667398e-01
-1.67124912e-01 2.15231791e-01 -4.99313660e-02 6.87545180e-01
-7.00776279e-01 1.11123538e+00 6.37221766e+00 7.01776922e-01
-1.98197937e+00 -3.66425395e-01 4.10347313e-01 -2.02394605e-01
-2.73864388e-01 -6.74380541e-01 -6.20029390e-01 7.23355651e-01
8.99292052e-01 -2.24467739e-01 4.18808907e-01 6.38042629e-01
8.84724498e-01 3.18494588e-01 -7.67662525e-01 9.79833543e-01
8.77623856e-02 -1.48047507e+00 -2.38723028e-02 -5.35028689e-02
4.82984930e-01 -3.70365053e-01 3.15359116e-01 -2.15096157e-02
-3.46701384e-01 -1.16597164e+00 8.33658576e-01 6.76596701e-01
1.35867345e+00 -6.53992057e-01 2.50137180e-01 4.75650787e-01
-1.23269081e+00 2.77978241e-01 5.90953827e-02 1.82802320e-01
5.59752941e-01 -1.60354897e-01 -1.08186400e+00 -8.86910260e-02
4.76586729e-01 1.29158407e-01 1.06151268e-01 8.55607092e-01
-1.57126650e-01 3.57586831e-01 -4.42923099e-01 -2.05131650e-01
-1.23644568e-01 3.23625021e-02 8.05243433e-01 9.80129659e-01
5.03934145e-01 -1.05682537e-01 -5.31637132e-01 1.07476926e+00
-1.87001482e-01 1.29008321e-02 -8.54567468e-01 -1.35278001e-01
5.51550508e-01 1.00781512e+00 -2.70424932e-01 -1.05979117e-02
-5.88043146e-02 8.78454268e-01 -3.01835269e-01 4.27589506e-01
-9.51016307e-01 -4.59241748e-01 9.21420991e-01 2.68623799e-01
1.27464488e-01 -3.46633732e-01 -4.64579351e-02 -6.39891624e-01
-1.52616762e-02 -8.34635794e-01 -4.57867652e-01 -1.05512846e+00
-6.03653193e-01 9.60693359e-01 -2.03762338e-01 -1.22064710e+00
-8.47199082e-01 -2.56241888e-01 -9.36782479e-01 1.22535527e+00
-1.15972912e+00 -1.13734162e+00 -3.45843166e-01 6.91670120e-01
7.61266470e-01 -1.54497251e-01 8.37245464e-01 1.47969916e-01
-3.68119240e-01 8.34918618e-01 -4.11260188e-01 3.22578758e-01
7.25027561e-01 -4.64169204e-01 1.00410390e+00 6.06099725e-01
-1.34598926e-01 3.94317448e-01 7.54577816e-01 -6.09893501e-01
-1.17077565e+00 -8.16289663e-01 9.65492666e-01 1.11123659e-01
1.38492614e-01 -4.67120737e-01 -9.43967044e-01 2.34938845e-01
2.10987046e-01 1.09254248e-01 2.13352367e-01 -1.09061599e+00
1.08757123e-01 3.82604524e-02 -1.38397229e+00 8.40512753e-01
9.44572270e-01 -4.64970380e-01 -4.23773140e-01 -8.51878822e-02
8.81397188e-01 -6.20541811e-01 -5.98702550e-01 3.87591243e-01
7.89848268e-01 -8.56790066e-01 8.11092615e-01 -2.01256543e-01
1.43865272e-01 -9.72918943e-02 2.53822953e-01 -1.41533291e+00
2.56905705e-01 -1.21121740e+00 -6.67856336e-02 1.25511885e+00
1.85595840e-01 -4.46921736e-01 9.18418407e-01 4.58999604e-01
-8.24657306e-02 -4.25076276e-01 -1.24555075e+00 -6.28036976e-01
2.05992863e-01 -3.31216037e-01 6.87283993e-01 5.66825688e-01
-8.24136660e-03 -7.02054650e-02 -3.57307673e-01 1.10997483e-01
4.47187841e-01 -2.29620606e-01 8.42060566e-01 -9.06318665e-01
1.98027670e-01 -6.22809529e-01 -2.71267027e-01 -1.00775266e+00
6.52819693e-01 -4.56429869e-01 1.20750710e-01 -1.27278447e+00
-5.80653489e-01 -7.67946988e-02 6.66800141e-01 1.56687826e-01
2.64619917e-01 -7.61405453e-02 2.50444621e-01 -2.62002796e-01
6.51500583e-01 8.32896173e-01 1.60044003e+00 1.09000858e-02
-5.88021457e-01 5.05922794e-01 2.65825659e-01 7.59308040e-01
6.78734660e-01 -1.70876220e-01 -3.34164292e-01 -1.95936933e-01
-3.91833186e-01 6.19445622e-01 2.91094929e-01 -7.97318935e-01
2.07125753e-01 -1.30466953e-01 3.24031800e-01 -4.57475841e-01
8.24776530e-01 -6.84234142e-01 2.86616981e-01 5.45924544e-01
-4.45474714e-01 -1.30878896e-01 4.64485556e-01 -7.45267197e-02
-7.32188746e-02 -6.78630173e-02 1.12174261e+00 2.23674253e-02
-2.17766285e-01 3.70754898e-01 -4.45390999e-01 -1.86584592e-01
9.53226030e-01 -3.38923186e-01 -1.49104178e-01 -7.65232146e-01
-7.09812820e-01 -3.77778023e-01 3.97278577e-01 6.53797925e-01
1.01817143e+00 -1.31891477e+00 -6.85887337e-01 7.97662795e-01
-4.25859958e-01 -6.89051226e-02 3.94230425e-01 3.78658950e-01
-6.06532216e-01 2.21792132e-01 -3.24243963e-01 -4.11697298e-01
-1.67363703e+00 4.20080096e-01 7.47175813e-01 6.03350282e-01
-7.72522569e-01 7.29812026e-01 -1.93749845e-01 -1.96339563e-01
5.87851644e-01 -3.09494317e-01 -3.04459065e-01 -1.26080260e-01
5.26050806e-01 2.52076715e-01 -1.98721781e-01 -9.72643137e-01
-2.35073209e-01 7.62937069e-01 5.72036564e-01 -4.73298460e-01
1.01699710e+00 -8.95478800e-02 3.25700372e-01 2.88353056e-01
1.21248448e+00 4.45743650e-01 -1.52958608e+00 3.28513533e-01
-6.81830108e-01 -2.98437387e-01 -2.28943542e-01 -4.52641785e-01
-1.26203930e+00 1.00979483e+00 5.63139796e-01 1.29681095e-01
8.85926962e-01 -2.53315836e-01 9.41214859e-01 -1.07223079e-01
1.00528382e-01 -8.70885253e-01 1.29412562e-01 4.58167285e-01
1.28793466e+00 -8.30944002e-01 -7.92981625e-01 -3.66332024e-01
-7.48667002e-01 1.37251472e+00 6.17379546e-01 1.79319873e-01
5.65295458e-01 7.60062993e-01 4.37007368e-01 4.32072252e-01
-5.64969003e-01 1.42100928e-02 2.08203122e-01 9.00077760e-01
3.02860558e-01 -1.82308137e-01 -3.49301472e-02 4.20190245e-01
-3.85299772e-01 1.25617594e-01 5.34715831e-01 4.89192784e-01
-1.73457608e-01 -8.26252818e-01 -3.91451180e-01 -3.74390572e-01
-3.61560583e-01 3.93840969e-02 -2.62941867e-01 8.09980392e-01
-8.71585775e-03 7.84727156e-01 3.72712761e-01 -4.07067478e-01
3.99782181e-01 2.84522593e-01 4.72719610e-01 -8.92592520e-02
-4.14276123e-01 2.88765818e-01 -9.68608633e-02 -4.87935275e-01
-1.32950157e-01 -5.38977265e-01 -1.58221698e+00 -4.48109508e-01
3.72623419e-03 -2.33165696e-01 1.21381497e+00 5.95342040e-01
4.32864904e-01 4.16718096e-01 9.15396512e-01 -1.29366148e+00
-5.13633609e-01 -1.12138271e+00 -4.21314985e-01 5.16404986e-01
5.87131679e-01 -4.41061437e-01 -7.41985679e-01 3.18324655e-01] | [13.245064735412598, -0.45447584986686707] |
e22955d9-7840-4a38-a8f7-e7895b163899 | ai-driven-shadow-model-detection-in-agropv | 2304.07853 | null | https://arxiv.org/abs/2304.07853v1 | https://arxiv.org/pdf/2304.07853v1.pdf | AI driven shadow model detection in agropv farms | Agro-photovoltaic (APV) is a growing farming practice that combines agriculture and solar photovoltaic projects within the same area. This emerging market is expected to experience significant growth in the next few years, with a projected investment of $9 billion in 2030. Identifying shadows is crucial to understanding the APV environment, as they impact plant growth, microclimate, and evapotranspiration. In this study, we use state-of-the-art CNN and GAN-based neural networks to detect shadows in agro-PV farms, demonstrating their effectiveness. However, challenges remain, including partial shadowing from moving objects and real-time monitoring. Future research should focus on developing more sophisticated neural network-based shadow detection algorithms and integrating them with control systems for APV farms. Overall, shadow detection is crucial to increase productivity and profitability while supporting the environment, soil, and farmers. | ['Dr. Susan Elias', 'Pascal Brunet', 'Sai Paavan Kumar Dornadula'] | 2023-04-16 | null | null | null | null | ['shadow-detection'] | ['computer-vision'] | [ 3.16636980e-01 5.76675422e-02 9.17335004e-02 8.48848745e-02
4.10740048e-01 -7.65621603e-01 3.60136293e-02 1.11629158e-01
4.08825308e-01 8.31858099e-01 -9.19723064e-02 -8.49797368e-01
2.17195541e-01 -1.42338979e+00 -4.04485554e-01 -9.63816464e-01
-1.58534154e-01 -2.00460672e-01 2.63344914e-01 -3.59150261e-01
-2.78903574e-01 6.62116528e-01 -1.40778899e+00 -8.19408074e-02
1.19675004e+00 8.38977516e-01 7.58095920e-01 7.88960516e-01
1.77042961e-01 1.69205874e-01 -5.25993407e-01 3.96174759e-01
2.77880073e-01 -6.17912412e-01 2.16478467e-01 -3.65949422e-01
2.03734756e-01 -7.82805026e-01 -7.26860855e-03 7.22454071e-01
6.86253428e-01 -4.78659868e-01 4.23743159e-01 -1.13345242e+00
-7.00858653e-01 3.65926743e-01 -6.57461524e-01 -1.28400385e-01
-4.69301164e-01 3.43307197e-01 4.32759434e-01 -4.06745106e-01
9.33791548e-02 7.68125057e-01 7.77440727e-01 -1.26714215e-01
-1.20084536e+00 -1.04850638e+00 -1.13336392e-01 9.23451185e-02
-8.09002876e-01 -1.77538469e-01 4.37791526e-01 -3.43513072e-01
9.87718940e-01 2.71751225e-01 1.33231819e+00 6.79160714e-01
7.82036960e-01 3.65393609e-01 1.11937249e+00 -2.93755770e-01
2.57397801e-01 -5.32986462e-01 -3.90168726e-01 2.98302680e-01
7.58885026e-01 6.49911702e-01 -1.39975861e-01 1.66054174e-01
9.29919362e-01 6.56777397e-02 -5.06846488e-01 -6.52521253e-01
-8.35779667e-01 7.10708559e-01 9.57151711e-01 2.54882067e-01
-6.19785607e-01 6.06058359e-01 2.11717677e-03 -2.04234168e-01
4.75104898e-01 3.10498178e-01 -7.66090870e-01 1.15525469e-01
-9.64676678e-01 1.01044737e-01 6.99367523e-01 7.47214794e-01
5.01427412e-01 4.73633140e-01 -3.11989993e-01 6.04094088e-01
4.97280747e-01 1.75640464e+00 -2.20335141e-01 -1.18218207e+00
-2.73512825e-02 5.37192345e-01 4.00055170e-01 -9.61755335e-01
-3.23378533e-01 -3.02393943e-01 -1.01867199e+00 6.85382187e-01
8.91644582e-02 -5.55040240e-01 -1.23237693e+00 1.41694999e+00
7.82306120e-02 1.11382417e-02 6.09880984e-02 4.59813923e-01
5.34562290e-01 1.08048165e+00 9.49203521e-02 -2.12202013e-01
1.24879766e+00 -5.99310756e-01 -9.18683648e-01 -3.29910547e-01
4.71076250e-01 -4.41043198e-01 4.29518908e-01 -1.36484191e-01
-6.15783036e-01 -1.24353670e-01 -1.23919892e+00 5.59350312e-01
-7.55171895e-01 5.37385285e-01 7.12018907e-01 7.80478179e-01
-9.95889246e-01 5.40384769e-01 -1.14744806e+00 -7.74657726e-01
6.84824586e-01 2.22808897e-01 3.31869423e-01 2.36044660e-01
-9.42919672e-01 1.30212379e+00 1.41420022e-01 8.73976648e-01
-8.65708828e-01 -7.23048270e-01 -7.35788345e-01 5.55749774e-01
1.13581285e-01 -3.70257854e-01 1.11257243e+00 -6.50976121e-01
-1.60275722e+00 3.13382685e-01 -2.27871031e-01 -4.66514111e-01
1.64144173e-01 -3.68283391e-01 -2.49576211e-01 -3.58749568e-01
-4.71112318e-02 5.92856705e-01 3.25824380e-01 -1.60403681e+00
-8.09929252e-01 -6.22412860e-01 -3.57172877e-01 1.97270975e-01
-3.08109552e-01 -2.31292382e-01 5.66054881e-01 -3.41836929e-01
1.07643679e-01 -1.11334419e+00 -1.91775501e-01 2.92443246e-01
1.90573521e-02 4.36062634e-01 1.56920481e+00 -9.57928121e-01
6.33295119e-01 -1.81050038e+00 -4.79644239e-01 -2.07769778e-02
-2.80485094e-01 6.95709944e-01 -1.41236812e-01 5.32579541e-01
3.78897130e-01 -2.42620945e-01 -6.00070953e-01 5.21410704e-01
-2.23351344e-01 4.23326939e-01 -1.82153776e-01 7.83628374e-02
2.05689996e-01 1.26835978e+00 -8.88491809e-01 3.09899300e-01
5.69212198e-01 5.80196798e-01 4.49730754e-01 2.24586353e-02
-2.55876273e-01 1.84681505e-01 -1.10991359e-01 1.12183011e+00
1.30143261e+00 1.08869091e-01 6.48809612e-01 4.82436903e-02
-6.66657090e-01 -3.05816233e-01 -5.62700450e-01 7.21753955e-01
-6.96187913e-01 1.07240033e+00 5.96983194e-01 -8.77284884e-01
9.76307929e-01 2.07196698e-01 2.10422367e-01 -1.03569651e+00
-1.14563018e-01 4.18562591e-01 -6.92326948e-02 -5.04223883e-01
4.14799809e-01 2.31624484e-01 6.02090657e-01 -4.87456508e-02
-4.38451201e-01 -9.55283418e-02 -2.24664420e-01 -4.40997392e-01
8.40755582e-01 5.66123903e-01 2.35989362e-01 -6.08959556e-01
-5.47611574e-03 2.51093566e-01 6.75975621e-01 2.21710041e-01
-2.20169231e-01 1.38033450e-01 2.80321419e-01 -5.92797212e-02
-9.21613932e-01 -1.08511710e+00 -1.86098114e-01 8.87492657e-01
3.91520709e-01 6.76991582e-01 -6.45016789e-01 -1.94909975e-01
5.97947955e-01 1.00616038e+00 -4.51457590e-01 2.33936504e-01
-5.47495961e-01 -1.03770268e+00 4.50641096e-01 9.64216888e-01
9.21279073e-01 -1.50622392e+00 -1.46906400e+00 5.27144134e-01
-1.13409936e-01 -7.04303801e-01 2.50089705e-01 7.55054593e-01
-1.05263090e+00 -9.64929163e-01 -1.27089834e+00 -6.68500841e-01
6.12081528e-01 8.11191976e-01 1.05860734e+00 -1.86403438e-01
-3.45210522e-01 9.17228609e-02 -3.41255546e-01 -1.05565858e+00
-3.37008804e-01 -5.52361039e-03 -6.03086889e-01 -8.34143877e-01
1.95169285e-01 -6.86034679e-01 -1.12817383e+00 2.05193743e-01
-3.74065131e-01 2.14455768e-01 9.24305081e-01 5.45236051e-01
2.21603841e-01 1.07877944e-02 5.73653221e-01 -5.97900867e-01
1.68916732e-01 -3.93222272e-01 -1.19468904e+00 5.78750074e-01
-8.22422504e-01 -6.29854321e-01 1.64240867e-01 -1.26715690e-01
-1.36118782e+00 2.11224303e-01 3.81907970e-01 4.73666370e-01
-4.46506172e-01 5.61349988e-01 -2.94342577e-01 -1.14068814e-01
2.41401210e-01 1.20787853e-02 -2.43681982e-01 1.71097089e-02
2.61417687e-01 4.90194917e-01 6.58485770e-01 2.88160831e-01
9.22693431e-01 2.84663081e-01 2.50741899e-01 -1.26414680e+00
-2.34031335e-01 -1.69406831e-01 -3.36061925e-01 -4.64402169e-01
6.18209660e-01 -1.02886367e+00 -8.63411069e-01 1.03549480e+00
-1.14008009e+00 -9.79878485e-01 6.84543401e-02 2.33210504e-01
1.00273140e-01 -5.51045239e-02 8.24167430e-02 -1.06501341e+00
-8.58451724e-01 -5.28919637e-01 8.68627250e-01 8.47376287e-01
6.23228215e-02 -6.95681453e-01 1.47212818e-01 8.13625529e-02
1.00707924e+00 7.51093090e-01 8.81034553e-01 5.29899538e-01
-7.10907161e-01 1.45985812e-01 -7.17579365e-01 4.87917364e-01
6.50445580e-01 1.39504269e-01 -1.18843877e+00 -3.26377660e-01
-3.29801023e-01 1.31767392e-01 9.73457634e-01 1.25169408e+00
8.19384515e-01 -5.35274930e-02 -6.83451295e-01 5.82871377e-01
1.81431484e+00 8.84801984e-01 7.60277033e-01 4.38463576e-02
5.07349670e-01 4.94233906e-01 8.89285266e-01 2.35041708e-01
2.24192455e-01 7.22471103e-02 1.12469959e+00 -6.12830937e-01
-1.31608263e-01 -5.89383543e-02 3.58703911e-01 3.00094783e-01
-1.67476192e-01 -5.41881859e-01 -9.57488656e-01 9.30606067e-01
-1.89816201e+00 -1.01667631e+00 -5.19602418e-01 2.05476904e+00
3.04799080e-01 -3.21735054e-01 -5.92782021e-01 -1.93910718e-01
5.49708605e-01 1.76158309e-01 -9.20591891e-01 -2.23698065e-01
-3.90563726e-01 5.11417508e-01 1.17606878e+00 1.15729235e-01
-7.44190335e-01 1.01220524e+00 6.00660181e+00 -5.66906072e-02
-1.52652800e+00 -5.43356612e-02 4.86059964e-01 5.46118796e-01
-3.44119549e-01 2.56817311e-01 -4.30637002e-01 3.62135530e-01
6.05246246e-01 1.45278171e-01 5.31232774e-01 7.22615540e-01
7.04698861e-01 -1.00779510e+00 -3.13800186e-01 1.86675683e-01
-2.99283922e-01 -1.18160105e+00 -7.14148104e-01 2.69513279e-01
1.03112590e+00 3.10361385e-01 -3.31387222e-01 1.12052999e-01
7.04376876e-01 -8.27949822e-01 2.20340788e-01 3.86274397e-01
6.43522084e-01 -5.40988624e-01 1.08152270e+00 1.02675125e-01
-1.48071158e+00 -2.26958618e-01 -4.32541579e-01 -1.76503196e-01
-1.04036005e-02 9.66482699e-01 -7.43246019e-01 6.11172616e-01
1.12039399e+00 4.35103565e-01 -3.73355806e-01 8.71121407e-01
-5.58210850e-01 1.02567232e+00 -6.33704424e-01 -1.81578636e-01
-2.94197705e-02 -5.24032474e-01 -9.97111481e-03 1.07037234e+00
8.53653908e-01 2.03967780e-01 -2.93041468e-01 1.03881335e+00
6.87307194e-02 -3.74860913e-01 -8.66379023e-01 -1.91394184e-02
7.90524304e-01 1.37416065e+00 -1.08391476e+00 5.13113774e-02
5.21708019e-02 1.07029212e+00 -6.15035832e-01 5.96134007e-01
-7.79087961e-01 -4.69896019e-01 6.83950901e-01 -2.42500857e-01
6.17714405e-01 -2.97387511e-01 -6.54610336e-01 -6.04061425e-01
8.26238319e-02 -4.02454197e-01 -4.86665428e-01 -1.25854361e+00
-6.71386302e-01 -4.87302482e-01 -2.01685116e-01 -6.07748866e-01
3.79943252e-01 -7.77713656e-01 -1.21117914e+00 1.02049625e+00
-1.77704918e+00 -1.61126792e+00 -1.07053626e+00 -4.38063145e-01
3.12015057e-01 3.12850416e-01 1.37662446e+00 -3.30254018e-01
-2.39084914e-01 -1.10875860e-01 9.31441844e-01 -2.46456444e-01
3.51695389e-01 -1.22073317e+00 5.91881156e-01 9.18256998e-01
-6.23202384e-01 -2.15756774e-01 4.63166773e-01 -9.18654621e-01
-1.33792996e+00 -1.33531153e+00 8.01920176e-01 1.12932742e-01
3.48313481e-01 -1.83025107e-01 -6.14252865e-01 4.76450801e-01
5.91387093e-01 -2.91829288e-01 1.78636327e-01 3.12391073e-02
4.58703905e-01 -6.26370907e-01 -1.22093713e+00 3.97240311e-01
7.44376659e-01 -6.50110096e-02 2.50713110e-01 1.39678299e-01
4.61108476e-01 -4.59589750e-01 -7.37845600e-01 9.04163241e-01
1.27991760e+00 -8.51676166e-01 6.28919005e-01 3.33676368e-01
5.00758529e-01 -3.20718408e-01 -6.01628125e-02 -1.75955188e+00
-5.60084343e-01 -1.29052699e-01 -3.51551294e-01 1.51916409e+00
1.81283101e-01 -8.59215140e-01 8.38038802e-01 3.70590053e-02
1.18373213e-02 -7.27273762e-01 -4.55305398e-01 -8.43715608e-01
8.74442011e-02 1.38238475e-01 6.87850475e-01 6.04420424e-01
-5.58872104e-01 -9.25935134e-02 -7.57509694e-02 8.56079698e-01
4.17401046e-01 3.34259450e-01 6.41354561e-01 -1.25431919e+00
5.12623429e-01 -2.94329554e-01 5.25077693e-02 -4.94193017e-01
-2.51300156e-01 -1.71998814e-01 4.79986906e-01 -2.22546053e+00
6.78646564e-02 -4.03465599e-01 -5.37178516e-02 9.84806776e-01
-1.85512051e-01 8.21064934e-02 1.29107744e-01 -3.52799475e-01
7.92928100e-01 8.87643695e-01 1.21111381e+00 -5.36938131e-01
-5.54291904e-01 2.94197261e-01 -5.63758194e-01 3.95316869e-01
1.29881501e+00 -3.68185133e-01 -3.82740259e-01 -7.86814749e-01
7.97951445e-02 8.80413502e-02 7.05321848e-01 -1.03467047e+00
-2.40088537e-01 -6.28824890e-01 8.50408375e-01 -1.32391429e+00
2.07165256e-02 -1.06427407e+00 3.63565773e-01 9.49138999e-01
3.38010550e-01 -3.31372797e-01 5.60419798e-01 3.19463074e-01
2.88833290e-01 1.84827745e-01 6.82086229e-01 7.85703361e-02
-6.26927137e-01 3.22227478e-02 -7.08929062e-01 -6.40765429e-01
1.20745420e+00 -2.27156207e-01 -6.99050426e-01 -1.63366243e-01
-1.22817002e-01 7.02058733e-01 4.59948391e-01 3.51388544e-01
3.49141598e-01 -1.10654414e+00 -7.46232748e-01 -5.18359505e-02
-4.54839952e-02 1.46898940e-01 1.24495700e-01 2.89713413e-01
-1.10366571e+00 4.46866184e-01 -5.95945597e-01 -6.36298478e-01
-1.17624581e+00 -6.48835525e-02 4.60758358e-01 -1.88011214e-01
-3.98451656e-01 4.59513277e-01 7.08119273e-02 -5.52573204e-01
-1.13840587e-01 -8.06516767e-01 6.54099807e-02 2.69038063e-02
-2.69283801e-02 6.25860333e-01 -5.25191985e-03 -3.29617076e-02
-2.40153059e-01 2.83953160e-01 8.44560802e-01 2.95513332e-01
1.55148935e+00 1.11605421e-01 -2.55365670e-01 3.92446697e-01
3.16617608e-01 -4.23016936e-01 -1.58534789e+00 4.12717372e-01
-1.73522934e-01 -1.04466341e-01 2.70892411e-01 -1.48042655e+00
-1.23775220e+00 7.53204525e-01 1.49455798e+00 2.68600643e-01
1.31956470e+00 -6.02146268e-01 7.45323420e-01 6.08299732e-01
1.77815288e-01 -9.41521823e-01 -3.80405486e-01 4.36893344e-01
9.99013007e-01 -1.21539617e+00 1.55869082e-01 -3.03269356e-01
-1.28009832e-02 1.07534373e+00 7.66799331e-01 1.68161303e-01
6.28893435e-01 5.63670456e-01 4.01924759e-01 -4.12711166e-02
-2.31889665e-01 -1.83142841e-01 -5.67429245e-01 1.14791131e+00
4.75096971e-01 8.27086747e-01 -1.55244350e-01 -2.30561525e-01
1.03723355e-01 2.06870362e-01 2.85421789e-01 1.13261211e+00
-6.74692214e-01 -9.60816920e-01 -5.02413273e-01 7.08089054e-01
2.22279400e-01 -2.86851197e-01 -3.48723590e-01 5.65713704e-01
4.03675258e-01 9.14217710e-01 -6.88839257e-02 1.68485865e-01
3.24447602e-01 -2.55006135e-01 3.97458225e-01 -4.17593688e-01
-3.23857456e-01 -4.68553789e-02 -7.04690516e-02 -1.69225931e-01
-5.34228444e-01 -7.45939672e-01 -8.90000820e-01 -4.86415803e-01
-8.05843353e-01 -4.13421124e-01 1.27179408e+00 6.92507327e-01
3.31756562e-01 7.52609849e-01 7.00491190e-01 -1.10594165e+00
-2.14914009e-01 -1.26091850e+00 -7.35507190e-01 -8.04270804e-01
1.42162055e-01 -5.74833393e-01 1.64017797e-01 2.60577933e-03] | [9.282966613769531, -1.5608569383621216] |
7f1c9054-be1d-4b03-84e7-b10a0c816131 | incremental-graph-based-neural-dependency | null | null | https://aclanthology.org/D17-1173 | https://aclanthology.org/D17-1173.pdf | Incremental Graph-based Neural Dependency Parsing | Very recently, some studies on neural dependency parsers have shown advantage over the traditional ones on a wide variety of languages. However, for graph-based neural dependency parsing systems, they either count on the long-term memory and attention mechanism to implicitly capture the high-order features or give up the global exhaustive inference algorithms in order to harness the features over a rich history of parsing decisions. The former might miss out the important features for specific headword predictions without the help of the explicit structural information, and the latter may suffer from the error propagation as false early structural constraints are used to create features when making future predictions. We explore the feasibility of explicitly taking high-order features into account while remaining the main advantage of global inference and learning for graph-based parsing. The proposed parser first forms an initial parse tree by head-modifier predictions based on the first-order factorization. High-order features (such as grandparent, sibling, and uncle) then can be defined over the initial tree, and used to refine the parse tree in an iterative fashion. Experimental results showed that our model (called INDP) archived competitive performance to existing benchmark parsers on both English and Chinese datasets. | ['Xiaoqing Zheng'] | 2017-09-01 | null | null | null | emnlp-2017-9 | ['transition-based-dependency-parsing'] | ['natural-language-processing'] | [-2.90261954e-03 4.42574501e-01 -1.67287886e-01 -7.92774141e-01
-4.86684233e-01 -4.79356676e-01 2.13464409e-01 3.34593832e-01
-6.49839163e-01 7.19491482e-01 3.50232780e-01 -6.81317747e-01
-5.85145056e-02 -1.03988993e+00 -7.15100527e-01 -5.31391382e-01
-4.71663356e-01 4.92540032e-01 3.34372640e-01 -9.30859819e-02
8.04394037e-02 2.68290430e-01 -1.24664664e+00 1.64938897e-01
9.61571693e-01 7.21400380e-01 5.68986177e-01 5.93328953e-01
-7.29124963e-01 7.64182866e-01 -3.41186464e-01 -7.74810672e-01
-5.29747568e-02 -3.00965875e-01 -8.98505926e-01 -5.00818670e-01
-1.44320846e-01 -3.43541801e-01 -1.94271401e-01 1.04363739e+00
2.73559958e-01 5.59409186e-02 1.13618314e-01 -7.33385444e-01
-7.26046205e-01 1.35509443e+00 -3.99578333e-01 3.94199401e-01
1.51257128e-01 -7.68096223e-02 1.42008233e+00 -7.83511102e-01
5.23966193e-01 1.45934117e+00 4.56819832e-01 5.68220794e-01
-7.92882740e-01 -6.91713750e-01 9.56181049e-01 1.38715506e-01
-7.49769449e-01 -3.42680544e-01 8.31721008e-01 7.15409359e-03
1.24928546e+00 4.49848399e-02 4.28478122e-01 8.77636969e-01
4.29570168e-01 7.75741875e-01 7.61268020e-01 -3.97431582e-01
1.27891660e-01 -3.45085889e-01 9.23343778e-01 1.05604935e+00
2.01559708e-01 1.62464425e-01 -4.45389330e-01 -1.85216352e-01
5.11597097e-01 -3.18673640e-01 -8.59311298e-02 3.66500258e-01
-7.99530804e-01 1.03601408e+00 6.31558120e-01 4.01316255e-01
-4.00970757e-01 -1.04080394e-01 3.27680737e-01 1.87905684e-01
2.80022502e-01 2.72137731e-01 -9.46101546e-01 -6.57608733e-02
-5.21597266e-01 -1.44665614e-01 8.35516036e-01 6.77146018e-01
1.12393320e+00 -2.72888690e-01 -1.61652774e-01 8.01316142e-01
3.06821167e-01 -9.15893540e-02 5.33831120e-01 -6.07931614e-01
8.87507915e-01 7.23567128e-01 -3.00344110e-01 -6.69420600e-01
-6.85260117e-01 -4.60945368e-01 -8.40253592e-01 -2.00211704e-01
3.59983653e-01 -5.29186010e-01 -1.23323631e+00 2.03274608e+00
3.55397820e-01 -1.86370424e-04 3.78910638e-02 7.17042387e-01
7.39581645e-01 7.13047385e-01 5.19699752e-01 -4.40691561e-01
1.30943799e+00 -8.36633861e-01 -5.53445041e-01 -5.67930281e-01
8.16335857e-01 -4.53625381e-01 6.68444157e-01 1.55983284e-01
-8.99904907e-01 -6.22724414e-01 -7.72397816e-01 -1.32061481e-01
-1.47096172e-01 -6.28642663e-02 1.15669215e+00 5.91606081e-01
-9.43670928e-01 8.23708892e-01 -1.12001061e+00 -1.78733796e-01
2.37069279e-01 5.77455521e-01 -5.16288340e-01 -2.26514518e-01
-1.43892932e+00 7.35552847e-01 7.43020654e-01 6.22881293e-01
-4.07887161e-01 -2.51219541e-01 -9.69034493e-01 2.42568970e-01
3.77194971e-01 -5.42211711e-01 9.36500728e-01 -1.02417529e+00
-1.50399160e+00 3.08129549e-01 -4.05670285e-01 -2.97202557e-01
1.18394643e-01 -3.48156214e-01 -2.08294451e-01 -1.33008629e-01
1.18290737e-01 6.01889729e-01 4.86767173e-01 -7.28604257e-01
-8.75214994e-01 -4.60033506e-01 4.66153532e-01 2.52844244e-01
-1.91661894e-01 1.30239144e-01 -5.05811155e-01 -2.11353645e-01
5.91506720e-01 -7.42586017e-01 -4.29299384e-01 -7.27149546e-01
-5.49062371e-01 -5.55638909e-01 4.01795059e-01 -7.88073480e-01
1.42491591e+00 -1.82607245e+00 1.39546290e-01 -1.03049241e-01
1.78689114e-03 3.27517420e-01 -2.62259841e-01 3.19666594e-01
-7.43429363e-02 3.93758893e-01 -2.31315389e-01 -1.59676924e-01
-3.64511132e-01 5.10902762e-01 6.16058633e-02 6.63063973e-02
6.16580546e-01 9.27166939e-01 -8.70900035e-01 -5.95632076e-01
-9.31014493e-02 2.34139964e-01 -7.04186797e-01 2.52667099e-01
-2.56015629e-01 4.05251682e-01 -8.11148167e-01 4.25455153e-01
6.56633377e-01 -2.78923493e-02 4.97163057e-01 -3.86871919e-02
-1.72910511e-01 7.25150049e-01 -9.33592379e-01 1.41372812e+00
-4.67902750e-01 -7.70993605e-02 8.36989284e-02 -1.02759635e+00
8.79642963e-01 2.21454307e-01 5.40268933e-03 -6.25367463e-01
4.20167595e-02 2.84142457e-02 5.02904952e-01 -2.25228116e-01
1.35693669e-01 -5.23937568e-02 -3.00855964e-01 2.46830568e-01
2.88370103e-01 6.63479149e-01 2.85951346e-01 3.54558319e-01
1.17005479e+00 3.73077840e-01 2.97658771e-01 -2.27688253e-01
7.59130299e-01 -1.96362436e-01 1.00259781e+00 7.14221716e-01
3.55289988e-02 2.20412508e-01 7.85977304e-01 -7.86206961e-01
-5.37628949e-01 -7.09486663e-01 5.69841079e-02 1.54291761e+00
-2.35528558e-01 -5.59401453e-01 -7.06257164e-01 -1.14535785e+00
-3.72074276e-01 6.03287637e-01 -7.16732204e-01 -2.41970620e-03
-1.16871631e+00 -1.09524751e+00 3.74022335e-01 8.06545734e-01
2.39214718e-01 -1.41613317e+00 -4.60867703e-01 7.39325464e-01
-1.55146802e-02 -9.77814615e-01 -7.67383501e-02 1.07794118e+00
-9.78652477e-01 -9.57099855e-01 -2.14720592e-01 -9.55356956e-01
7.66488552e-01 -2.86404520e-01 1.10624349e+00 5.36927104e-01
2.03912720e-01 -3.82774442e-01 -5.84532201e-01 -1.61010902e-02
-4.19893414e-01 4.32532638e-01 -1.68465003e-01 -1.42100498e-01
3.48391563e-01 -6.28414571e-01 -2.08529770e-01 -8.71620029e-02
-4.67207670e-01 1.29814193e-01 8.99014890e-01 1.22176921e+00
4.27830637e-01 5.61031140e-02 6.98963940e-01 -1.38413620e+00
4.16021079e-01 -5.39751291e-01 -6.36648118e-01 3.35734963e-01
-3.90519589e-01 7.04067409e-01 9.89024997e-01 -2.47795492e-01
-1.35103214e+00 3.31055373e-01 -6.16334915e-01 3.56632620e-02
-9.95947942e-02 9.17959392e-01 -5.78470767e-01 3.89162451e-01
9.27649885e-02 1.74405411e-01 -6.29716694e-01 -7.61466265e-01
3.06131095e-01 2.45497644e-01 3.66838962e-01 -9.32780266e-01
5.77743769e-01 -1.51501551e-01 -1.17745705e-01 -4.97927576e-01
-1.00450957e+00 -2.25608442e-02 -1.27785099e+00 3.99783790e-01
9.11514282e-01 -6.67626858e-01 -4.03956890e-01 3.70133877e-01
-1.40672815e+00 -4.15645540e-02 2.44476214e-01 3.05707514e-01
1.00712135e-01 5.42374730e-01 -9.05043423e-01 -9.40211356e-01
-4.42140818e-01 -1.04727995e+00 6.10406995e-01 3.71717662e-01
7.95333087e-02 -9.37785685e-01 -7.80122951e-02 -4.75371592e-02
4.96362001e-02 -1.57659307e-01 1.59835327e+00 -8.93550754e-01
-5.79975963e-01 -5.47850654e-02 -1.41249225e-01 1.55091688e-01
-1.00397943e-02 -9.48997065e-02 -8.21925282e-01 -1.07517190e-01
-1.27675638e-01 -7.56686032e-02 1.18502545e+00 2.85464138e-01
8.88461053e-01 -2.83038259e-01 -5.08534610e-01 6.11107647e-01
1.30643201e+00 1.34264857e-01 2.31868654e-01 8.74621347e-02
8.42977941e-01 6.32325470e-01 5.33568442e-01 1.69866383e-01
6.15604520e-01 1.82849899e-01 6.06103837e-01 2.71734834e-01
8.79621282e-02 -5.61221182e-01 4.32782471e-01 1.15946591e+00
-1.18224725e-01 -3.20921600e-01 -7.81418324e-01 5.35922885e-01
-1.72242928e+00 -4.68558818e-01 -1.97745413e-01 2.07219052e+00
6.94774032e-01 6.09661639e-01 -2.95809209e-01 -1.30541977e-02
9.20813322e-01 1.57761469e-01 -5.02295256e-01 -7.45464087e-01
7.37833381e-02 4.72416788e-01 4.24494594e-01 5.09680510e-01
-1.05013692e+00 1.44644797e+00 5.76726055e+00 2.90258884e-01
-1.01129878e+00 5.69724813e-02 7.57850707e-01 4.66825217e-01
-3.12538594e-01 4.54955012e-01 -1.30790997e+00 2.43437290e-01
1.11263967e+00 1.97032452e-01 2.58692443e-01 7.24967957e-01
-2.82128721e-01 -1.61956072e-01 -1.11820698e+00 1.93103373e-01
-4.32229221e-01 -9.14683461e-01 5.66049367e-02 -7.29746670e-02
1.48685366e-01 2.54893392e-01 -5.29001176e-01 6.47877693e-01
6.55090213e-01 -9.05471325e-01 5.90504885e-01 8.98315832e-02
4.27939981e-01 -8.89579833e-01 9.00125027e-01 8.49593461e-01
-1.25800216e+00 -3.43193561e-01 -8.51564944e-01 -5.84266067e-01
1.73597887e-01 6.36619449e-01 -7.28163362e-01 7.40637183e-01
4.97488171e-01 4.17416036e-01 -5.55509448e-01 5.26803970e-01
-7.70959079e-01 9.50723529e-01 -5.00058115e-01 -2.70447850e-01
5.09325027e-01 -2.38278694e-02 2.04995573e-01 1.10596061e+00
1.40212491e-01 3.88092726e-01 3.23148787e-01 4.55569834e-01
-6.67240694e-02 2.76294500e-01 -1.21517241e-01 1.01029515e-01
4.96024698e-01 1.28356290e+00 -8.94999325e-01 -2.04127207e-01
-5.89356244e-01 7.62803495e-01 1.03147936e+00 9.46141258e-02
-6.75008357e-01 -1.52312189e-01 4.33441997e-01 -2.43355885e-01
6.43299222e-01 -2.65023023e-01 -1.44033074e-01 -1.15542352e+00
6.82640523e-02 -3.54864776e-01 6.75621688e-01 -2.59531945e-01
-1.30474722e+00 1.01895189e+00 -2.21373126e-01 -3.00629050e-01
-4.14007336e-01 -6.33058429e-01 -9.48006511e-01 1.09369051e+00
-1.70452929e+00 -1.18338811e+00 2.75627732e-01 4.32662040e-01
6.69382393e-01 1.38457447e-01 1.13246214e+00 4.13983278e-02
-8.91444743e-01 5.95477641e-01 -3.73616189e-01 6.27344489e-01
2.13598683e-01 -1.17205381e+00 7.85444915e-01 1.16918194e+00
2.96719819e-01 6.54158890e-01 2.56493330e-01 -7.82071352e-01
-1.38011956e+00 -8.51331949e-01 1.37001860e+00 -2.93007016e-01
5.05989790e-01 -4.56152916e-01 -1.10868108e+00 8.59548807e-01
-5.35096265e-02 2.02327818e-01 4.63048667e-01 7.28262067e-01
-3.94577622e-01 1.87728271e-01 -6.87482953e-01 2.26524100e-01
1.26725018e+00 -1.39379799e-02 -9.24261451e-01 -1.11391032e-02
8.38350832e-01 -2.24081963e-01 -7.03523517e-01 3.56932908e-01
5.22892296e-01 -9.58056390e-01 5.55055499e-01 -8.26320767e-01
5.25309801e-01 5.21463230e-02 6.00486882e-02 -1.18752062e+00
-9.32369888e-01 -5.66716552e-01 1.73085891e-02 1.54543984e+00
7.76336789e-01 -5.94663620e-01 6.61448896e-01 7.32931614e-01
-2.71707505e-01 -8.42919588e-01 -1.13636112e+00 -2.78114080e-01
3.64273965e-01 -4.70734417e-01 7.64709413e-01 6.52691901e-01
-2.01087222e-02 7.26793826e-01 -2.94848979e-01 5.55919886e-01
3.29500169e-01 3.82291496e-01 3.13920647e-01 -1.48900414e+00
-4.15156960e-01 -7.39944428e-02 -1.11209966e-01 -1.12868249e+00
5.65257430e-01 -1.01930892e+00 2.46777907e-01 -1.50881457e+00
1.37263879e-01 -7.75854707e-01 -6.87204897e-01 8.72104466e-01
-8.60154688e-01 -3.85255665e-01 1.09807938e-01 -4.04661112e-02
-4.46353912e-01 1.66072056e-01 1.16834784e+00 2.58488089e-01
-2.03699470e-01 1.14335820e-01 -8.76809239e-01 8.99526775e-01
6.99337780e-01 -8.66219044e-01 -2.10620254e-01 -9.04956341e-01
2.45273292e-01 4.22215581e-01 -1.82156622e-01 -6.68714046e-01
3.77552509e-01 -9.22689065e-02 5.12844682e-01 -5.43013394e-01
2.84589641e-02 -6.16970658e-01 -3.01866889e-01 3.65125865e-01
-2.15975672e-01 2.09492832e-01 -8.26705918e-02 5.03355622e-01
-1.78924605e-01 -3.84735674e-01 4.95088369e-01 -5.10647357e-01
-8.83687675e-01 4.30299133e-01 -7.30567500e-02 -9.33371310e-04
4.99706805e-01 1.96878426e-02 -5.37985116e-02 1.74561441e-02
-8.40661526e-01 3.39552701e-01 2.23160181e-02 5.91601312e-01
3.65723163e-01 -8.20163667e-01 -7.44679093e-01 6.03526533e-01
-2.88747877e-01 1.81106672e-01 1.76557809e-01 3.77491921e-01
-2.21300215e-01 3.50571245e-01 -1.67019054e-01 -3.40418488e-01
-9.80528891e-01 5.88212311e-01 -6.81825401e-03 -8.58129323e-01
-7.39704967e-01 1.23834133e+00 4.68837827e-01 -5.17916143e-01
4.86768298e-02 -5.79505444e-01 -4.72745180e-01 -3.65873016e-02
3.57594073e-01 -3.58601928e-01 1.16527952e-01 -6.93107545e-01
-6.19052351e-01 4.13592190e-01 -3.44751656e-01 2.57883966e-01
1.60014915e+00 -7.05030784e-02 -2.30520800e-01 1.26660421e-01
9.12247181e-01 5.08373119e-02 -1.20071995e+00 -1.74944595e-01
4.07647491e-01 -4.62247171e-02 5.61439879e-02 -9.63871360e-01
-1.15587413e+00 1.16183615e+00 -3.72570306e-02 2.47591157e-02
1.11552119e+00 8.41218904e-02 9.00328338e-01 4.46040660e-01
5.59533954e-01 -7.78086007e-01 -5.30909717e-01 1.01799798e+00
2.85914123e-01 -9.55728590e-01 -1.56270817e-01 -6.48569643e-01
-4.87012267e-01 1.35340858e+00 7.44403660e-01 -2.68035859e-01
6.94935143e-01 4.88674492e-01 -7.08846450e-02 1.30055308e-01
-1.06055641e+00 -1.40787557e-01 -2.63864212e-02 5.15546918e-01
7.74834156e-01 1.08846776e-01 -5.46098590e-01 1.25406897e+00
-2.67434269e-01 -2.89211422e-01 1.73638001e-01 8.42101038e-01
-6.46740735e-01 -1.64554870e+00 -3.42278667e-02 4.66652304e-01
-8.06330502e-01 -4.70495999e-01 -1.27949178e-01 6.77698314e-01
3.27391654e-01 7.67637134e-01 -6.15438856e-02 -2.32978806e-01
8.05981979e-02 4.66951728e-01 4.54714596e-01 -8.62393081e-01
-8.95114303e-01 -2.93368883e-02 4.30745184e-01 -4.37700957e-01
-1.28415823e-01 -6.39287889e-01 -1.48579800e+00 1.36509880e-01
-5.81260741e-01 3.36347729e-01 6.52087629e-01 1.22652042e+00
2.47509077e-01 4.92343366e-01 4.72948551e-01 -5.91756165e-01
-4.59480107e-01 -1.03982580e+00 -1.64407074e-01 2.01752201e-01
9.48181897e-02 -5.07281661e-01 -2.04104260e-01 -2.26509154e-01] | [10.415884971618652, 9.549471855163574] |
5b72e45a-1f1a-4e08-9908-3e32c73b562c | deep-big-simple-neural-nets-excel-on | 1003.0358 | null | https://arxiv.org/abs/1003.0358v1 | https://arxiv.org/pdf/1003.0358v1.pdf | Deep Big Simple Neural Nets Excel on Handwritten Digit Recognition | Good old on-line back-propagation for plain multi-layer perceptrons yields a very low 0.35% error rate on the famous MNIST handwritten digits benchmark. All we need to achieve this best result so far are many hidden layers, many neurons per layer, numerous deformed training images, and graphics cards to greatly speed up learning. | ['Juergen Schmidhuber', 'Luca Maria Gambardella', 'Ueli Meier', 'Dan Claudiu Ciresan'] | 2010-03-01 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [-2.94397593e-01 -3.59118829e-04 -4.65230979e-02 -5.55891216e-01
-8.54987949e-02 -3.02322954e-02 3.39554012e-01 -1.90016806e-01
-9.65134740e-01 9.14109111e-01 -3.41720521e-01 -6.02190733e-01
4.53197956e-01 -6.74983859e-01 -8.12831402e-01 -6.38642192e-01
-4.27266508e-02 2.78071493e-01 7.21132040e-01 -3.29093248e-01
2.52760708e-01 8.74604523e-01 -1.05510235e+00 6.96567357e-01
1.63271934e-01 1.03637862e+00 -3.31356004e-02 1.11660039e+00
1.14107296e-01 1.04171681e+00 -6.49982810e-01 -8.20159495e-01
2.85270810e-01 9.05261338e-02 -6.77317441e-01 -9.36346799e-02
5.47652423e-01 -4.91954535e-01 -4.47624892e-01 8.84624302e-01
5.17942488e-01 5.83073534e-02 6.02534294e-01 -6.57918036e-01
-9.26426113e-01 9.47694778e-01 -5.39515257e-01 3.92177105e-01
-1.60696507e-01 1.94812179e-01 7.58028924e-01 -9.52671409e-01
2.35493571e-01 1.23535454e+00 8.77090216e-01 7.31113315e-01
-8.71316493e-01 -5.56582332e-01 6.40768632e-02 -2.75692972e-03
-9.56327617e-01 -2.47002080e-01 7.19129384e-01 -2.48212889e-01
1.57268214e+00 2.60329813e-01 5.69988549e-01 1.06173885e+00
7.43140638e-01 8.22154224e-01 7.86435008e-01 -3.99568141e-01
3.28550227e-02 2.89844215e-01 2.38676280e-01 1.11504161e+00
3.05681199e-01 1.97775751e-01 -1.01696569e-02 1.92100227e-01
1.38667405e+00 2.11796209e-01 1.31938666e-01 1.67591825e-01
-9.66331601e-01 8.47667515e-01 1.02754271e+00 3.32614869e-01
-2.90523946e-01 4.40335453e-01 4.70749050e-01 4.52646166e-01
3.45520712e-02 6.74659312e-01 -5.02611637e-01 1.45938560e-01
-8.18522811e-01 5.59944212e-02 7.26900399e-01 5.07592201e-01
4.49948907e-01 7.10912406e-01 3.21828216e-01 8.14431190e-01
2.15833619e-01 3.61640245e-01 7.53309011e-01 -7.81248391e-01
8.30478191e-01 4.37371790e-01 -7.37003833e-02 -9.88415480e-01
-6.13749325e-01 -5.84003210e-01 -1.41805041e+00 9.97741818e-01
5.22632837e-01 -5.40273547e-01 -1.14570963e+00 7.39350796e-01
-3.32945645e-01 -1.75916746e-01 1.51816383e-01 9.03869748e-01
6.47996664e-01 8.44624102e-01 -2.09902078e-01 4.37787443e-01
8.50535214e-01 -1.40525603e+00 -2.44807661e-01 -8.83701265e-01
2.57825613e-01 -6.89637244e-01 1.00444937e+00 7.85515666e-01
-1.28522980e+00 -7.23052979e-01 -1.50788724e+00 -8.87155607e-02
-4.71528172e-01 3.80873531e-01 1.13673639e+00 5.09268105e-01
-7.82938123e-01 1.05870140e+00 -1.00546753e+00 5.06580353e-01
6.90722287e-01 5.07091463e-01 -2.78074116e-01 4.73169424e-02
-9.03721571e-01 1.15113091e+00 8.89300168e-01 3.65189463e-01
-4.72094148e-01 -3.16801339e-01 -7.53174067e-01 -1.73976589e-02
-1.67070597e-01 -2.28605211e-01 1.09778047e+00 -9.80214298e-01
-1.89217842e+00 6.85951650e-01 1.22742586e-01 -6.51877284e-01
7.68384993e-01 -2.66146660e-01 -5.59620738e-01 -4.79213335e-02
-6.56277597e-01 7.25460529e-01 7.70532429e-01 -8.39656889e-01
-5.22485614e-01 -1.29415646e-01 -4.16741103e-01 -2.24020958e-01
-2.82595843e-01 7.81995431e-02 -2.22241923e-01 -6.97977543e-01
3.22327733e-01 -6.62885368e-01 -5.35847008e-01 2.22933367e-02
-3.57978821e-01 -1.53034851e-01 5.16575873e-01 -6.92420661e-01
6.24870002e-01 -2.05472016e+00 -4.13565561e-02 1.19624369e-01
1.81936637e-01 5.39752662e-01 -8.06389824e-02 -2.94049561e-01
-2.72266865e-01 -1.61574975e-01 1.02731204e-02 -3.58428687e-01
-1.16279066e-01 2.49811411e-01 -4.23211396e-01 6.18046403e-01
2.16304347e-01 1.03635633e+00 -5.85165977e-01 -1.22424379e-01
4.43076253e-01 5.76709151e-01 -6.29005134e-01 -1.05984561e-01
-3.48203480e-02 7.63056334e-03 1.01097256e-01 6.03826702e-01
5.15565991e-01 -3.92369032e-01 -2.64167845e-01 -9.23969038e-03
-3.03811450e-02 -9.58694145e-02 -1.19413948e+00 1.29689300e+00
-8.94171745e-02 1.01171649e+00 -2.28836358e-01 -7.79160738e-01
1.09959149e+00 2.16607645e-01 5.14724255e-02 -4.92445678e-01
3.08739871e-01 3.85253489e-01 1.12279303e-01 -3.15739885e-02
4.36578900e-01 -2.93375373e-01 9.88637060e-02 2.10667834e-01
2.64396071e-01 -2.02413395e-01 7.21002650e-03 -3.87440711e-01
6.77178919e-01 -3.72526705e-01 -1.23198174e-01 -1.35758594e-02
2.18492642e-01 -1.05493441e-01 5.02116024e-01 7.47999132e-01
9.05781239e-03 7.62043178e-01 3.26962709e-01 -1.08529437e+00
-1.48430216e+00 -6.65220857e-01 -1.40625402e-01 1.19373751e+00
-3.13946545e-01 3.26984972e-01 -4.30021346e-01 -1.41021654e-01
1.42041624e-01 5.22915184e-01 -4.35512543e-01 3.87568288e-02
-7.69589305e-01 -9.63068128e-01 1.10329926e+00 1.09954762e+00
7.22778916e-01 -1.34732497e+00 -7.39219189e-01 4.63916630e-01
8.59786630e-01 -9.77652133e-01 1.03179768e-01 6.99134648e-01
-1.17320895e+00 -7.12099075e-01 -1.32918775e+00 -1.25415659e+00
7.52522111e-01 -6.62684262e-01 9.14807737e-01 -2.80687641e-02
-2.72630662e-01 -8.41344833e-01 2.44737700e-01 -2.70943463e-01
-4.35234278e-01 4.80101667e-02 1.96341097e-01 -4.84023511e-01
5.59653416e-02 -2.95962542e-01 -1.78589806e-01 7.81142041e-02
-5.50942540e-01 2.42785335e-01 9.64507222e-01 1.04304373e+00
1.86049446e-01 2.57058233e-01 5.08627705e-02 -1.03968966e+00
4.69651669e-01 9.62337330e-02 -8.37006986e-01 2.19722584e-01
-5.64421415e-01 2.06340030e-01 1.16915143e+00 -6.87834501e-01
-7.69760191e-01 7.50555396e-02 -8.43745351e-01 -2.82445222e-01
-1.73271522e-01 2.79612422e-01 5.54399192e-01 -5.33457577e-01
9.94542956e-01 1.77051276e-01 -4.18687671e-01 -6.52230322e-01
2.17903882e-01 5.06280243e-01 9.36315000e-01 -4.00780998e-02
4.08044547e-01 9.03013423e-02 -2.62572110e-01 -5.75994015e-01
-6.31763101e-01 1.20351225e-01 -7.87095010e-01 3.04040998e-01
5.66943228e-01 -7.22804546e-01 -1.03947794e+00 1.13628936e+00
-1.23954749e+00 -7.73297250e-01 -1.08254083e-01 5.63026428e-01
-1.54082641e-01 6.57045990e-02 -1.52205527e+00 -5.20001292e-01
-4.63461965e-01 -1.07858765e+00 3.04568112e-01 2.68827736e-01
1.65435493e-01 -1.35149872e+00 7.22081540e-03 1.15448520e-01
7.37454832e-01 2.80570865e-01 6.94288433e-01 -2.68482119e-01
-1.29487798e-01 -5.84177971e-01 -3.46083462e-01 8.30367029e-01
-2.92039011e-02 2.29083881e-01 -1.04263926e+00 -4.68931556e-01
-1.09065197e-01 -7.45230258e-01 1.33225882e+00 3.19828629e-01
1.30411875e+00 -5.56427240e-01 -1.19642578e-01 8.68448794e-01
1.53493226e+00 2.91570723e-01 5.95904052e-01 3.78407925e-01
9.18107986e-01 -3.22368026e-01 -3.77323270e-01 3.18747818e-01
-2.71978974e-02 9.87759680e-02 2.01650426e-01 -6.34277225e-01
4.19591889e-02 1.17464714e-01 2.15841562e-01 9.85308766e-01
-3.45453113e-01 5.60259558e-02 -1.25203633e+00 -2.98191532e-02
-1.42674017e+00 -8.68012786e-01 7.73611367e-02 1.50823176e+00
1.11105561e+00 1.18064690e+00 -2.46244669e-01 4.55663651e-01
3.76213938e-01 1.96543083e-01 -7.57576287e-01 -7.10005581e-01
-3.01244587e-01 3.70756924e-01 1.05204642e+00 7.31446862e-01
-1.26987076e+00 8.88569117e-01 8.54690361e+00 3.58274490e-01
-1.69355106e+00 -3.78426999e-01 7.55923092e-01 4.07808721e-02
2.40142822e-01 -7.36571848e-01 -1.05270123e+00 3.15660238e-01
9.88445222e-01 3.75059813e-01 3.50000173e-01 1.15312815e+00
-3.88566822e-01 1.70295350e-02 -9.82210159e-01 1.26326919e+00
-6.25244603e-02 -1.72146261e+00 2.51022220e-01 -2.45490998e-01
8.77325058e-01 4.91194814e-01 3.04213196e-01 3.63714069e-01
6.30085409e-01 -1.39848709e+00 4.76869822e-01 6.47335410e-01
8.09195518e-01 -7.31806815e-01 1.05450737e+00 4.78999168e-01
-7.68600047e-01 -2.15469092e-01 -1.05349493e+00 -4.57304507e-01
-1.87780440e-01 6.80941403e-01 -8.70347142e-01 -7.15000927e-02
3.78211081e-01 5.32647371e-01 -6.25793576e-01 1.20176613e+00
-2.00668350e-01 6.36927843e-01 -4.73157465e-01 -3.32591206e-01
7.67354429e-01 4.34828758e-01 8.08953028e-03 1.54196370e+00
-3.65806818e-01 -1.70689464e-01 -5.13652749e-02 6.41211331e-01
-2.06533045e-01 -5.24605691e-01 -1.04173876e-01 1.57627344e-01
-2.87861619e-02 8.46224725e-01 -7.88297415e-01 -5.76456726e-01
-3.24045539e-01 1.38632011e+00 4.51516598e-01 4.02874082e-01
-6.68230295e-01 -8.91765237e-01 4.09969628e-01 -1.61119208e-01
3.40011805e-01 -6.26520813e-01 -6.27944291e-01 -1.09781051e+00
-1.76212370e-01 -5.36521375e-01 -4.66464013e-02 -7.13843644e-01
-1.18951142e+00 8.14527750e-01 -7.11261809e-01 -4.89376128e-01
-4.07455832e-01 -1.45207310e+00 -7.42450118e-01 9.16853487e-01
-1.53337038e+00 -7.61906028e-01 -4.05498780e-02 5.45410693e-01
4.46393132e-01 -5.53608656e-01 9.97422099e-01 5.06005824e-01
-8.02305460e-01 9.53544855e-01 4.14336443e-01 9.68565285e-01
2.00876743e-01 -1.32776260e+00 1.12043846e+00 7.96445191e-01
2.34536782e-01 2.98690766e-01 4.77549821e-01 -1.52378112e-01
-1.35871696e+00 -8.26222897e-01 6.55685306e-01 -1.77767918e-01
5.70865035e-01 -2.00117320e-01 -1.18314886e+00 1.00520217e+00
-7.21087828e-02 3.59795272e-01 8.89385566e-02 6.97406055e-03
-3.45953614e-01 -4.69341204e-02 -9.81353760e-01 4.66073006e-01
3.08022946e-01 -3.53926241e-01 -9.17791784e-01 2.13339061e-01
4.29485202e-01 -8.51092160e-01 -7.06867635e-01 8.34588483e-02
5.31081140e-01 -8.58679473e-01 1.06675553e+00 -7.79521465e-01
5.52685678e-01 1.79151028e-01 1.56057015e-01 -1.24490619e+00
-7.47854412e-01 -2.68940032e-01 -8.79538506e-02 4.94309336e-01
9.90224183e-01 -7.34925389e-01 1.23720622e+00 7.94625938e-01
-2.16657594e-01 -1.04443228e+00 -6.98575020e-01 -8.33524883e-01
4.84014332e-01 -3.40836227e-01 2.30911091e-01 7.94922709e-01
-1.37678623e-01 3.59761775e-01 -4.89600450e-01 -1.74451303e-02
6.45748079e-01 -2.80249178e-01 2.24341840e-01 -1.07503605e+00
-4.75290149e-01 -9.77837145e-01 -6.34091914e-01 -1.15683746e+00
-1.64685845e-01 -5.77791214e-01 -1.30531386e-01 -1.39820158e+00
-1.63064495e-01 -4.46479648e-01 -4.76356566e-01 9.46062982e-01
5.77937718e-03 6.57637715e-01 7.82748982e-02 1.99814141e-01
-5.49809813e-01 1.13465481e-01 1.06151378e+00 -2.98062086e-01
-2.37419695e-01 1.73403785e-01 -3.36698562e-01 1.19493163e+00
9.32811141e-01 -3.87008190e-01 3.05783838e-01 -9.53685045e-01
1.39040247e-01 -2.19620362e-01 3.11339051e-01 -1.50768697e+00
5.73983073e-01 -7.02299550e-02 1.28095269e+00 -4.25516486e-01
6.89509988e-01 -5.98514020e-01 -1.05462715e-01 1.03342378e+00
-5.20574033e-01 9.18415114e-02 5.48859715e-01 -1.43572330e-01
-3.45061988e-01 -2.81177521e-01 1.21352935e+00 -3.44461620e-01
-9.95063901e-01 7.50736594e-02 -2.49053091e-01 -1.61985084e-01
7.28257179e-01 -2.98897237e-01 -2.70749152e-01 1.83257535e-01
-7.83190846e-01 -1.38622880e-01 7.76359811e-02 1.74188003e-01
7.90052414e-01 -1.10514629e+00 -8.68834078e-01 5.53619087e-01
-5.76501429e-01 -1.05027519e-01 -1.75364703e-01 5.13637289e-02
-1.20871663e+00 6.57034099e-01 -9.08010542e-01 -3.15362632e-01
-1.18425822e+00 3.71391356e-01 7.77205348e-01 -1.92810372e-01
-1.11043787e+00 1.65973508e+00 -7.40355313e-01 -2.95394927e-01
5.57570457e-01 -7.77394772e-01 7.22083300e-02 -2.66564161e-01
8.76877666e-01 3.39183927e-01 2.71686554e-01 -6.20619692e-02
-4.54734921e-01 3.29363793e-01 -2.22937286e-01 -1.48087993e-01
1.54835808e+00 6.24466121e-01 2.66722798e-01 7.34509885e-01
1.17046070e+00 -3.75156701e-01 -1.74866545e+00 -1.80266365e-01
-4.12815250e-02 -1.06057696e-01 2.26777524e-01 -9.29472566e-01
-1.26636863e+00 1.14379752e+00 5.40868223e-01 2.11755723e-01
6.59604490e-01 -4.47746426e-01 8.30779493e-01 1.28263223e+00
2.20357925e-01 -1.17919934e+00 -6.11676164e-02 1.03550160e+00
8.78533006e-01 -1.31373620e+00 2.55971014e-01 4.45950210e-01
-7.69115150e-01 1.61790085e+00 4.17588204e-01 -8.64674509e-01
8.04371893e-01 7.56108582e-01 2.51023889e-01 1.42904624e-01
-7.34785140e-01 4.40239370e-01 4.13375914e-01 1.44591570e-01
5.37627399e-01 1.18252188e-01 5.80649674e-01 3.20635080e-01
-5.42564094e-01 -4.04155580e-03 2.66727954e-01 8.17098498e-01
-8.28560650e-01 -4.80580926e-01 -4.73656446e-01 5.13202488e-01
-7.86231697e-01 -2.18748122e-01 2.38857698e-02 4.66406375e-01
-3.04548085e-01 3.90484989e-01 2.46997938e-01 -4.90941316e-01
2.43481562e-01 6.84519857e-02 6.11075938e-01 -3.55315298e-01
-1.04760575e+00 -4.80797768e-01 -4.47048038e-01 -2.65758395e-01
9.47727934e-02 1.96135283e-01 -1.36939168e+00 -8.92720819e-01
-2.50512689e-01 -1.91099912e-01 9.11980212e-01 8.81782532e-01
-2.43775755e-01 7.31921911e-01 4.31496352e-01 -1.30066156e+00
-8.47227812e-01 -1.08826363e+00 -6.22579217e-01 -2.67662806e-03
6.13629758e-01 -2.02805117e-01 -3.71060610e-01 -5.98872267e-02] | [8.864402770996094, 2.779172897338867] |
bee4c2b7-5502-49d9-831b-2805fff81525 | deep-learning-for-lung-cancer-detection | 1705.09435 | null | http://arxiv.org/abs/1705.09435v1 | http://arxiv.org/pdf/1705.09435v1.pdf | Deep Learning for Lung Cancer Detection: Tackling the Kaggle Data Science Bowl 2017 Challenge | We present a deep learning framework for computer-aided lung cancer
diagnosis. Our multi-stage framework detects nodules in 3D lung CAT scans,
determines if each nodule is malignant, and finally assigns a cancer
probability based on these results. We discuss the challenges and advantages of
our framework. In the Kaggle Data Science Bowl 2017, our framework ranked 41st
out of 1972 teams. | ['Tse Chiang Howe', 'Mathieu Ravaut', 'Jie Lin', 'Cen Chen', 'Zeng Zeng', 'Vijay Chandrasekhar', 'Kingsley Kuan', 'Huiling Chen', 'Gaurav Manek', 'Babar Nazir'] | 2017-05-26 | null | null | null | null | ['lung-cancer-diagnosis'] | ['medical'] | [-1.79178596e-01 1.72206283e-01 -6.13009334e-01 -2.08986878e-01
-1.32121968e+00 -3.68495047e-01 1.68325976e-01 1.20259024e-01
-3.29572797e-01 2.28816301e-01 4.90403086e-01 -7.04559088e-01
-3.71698029e-02 -7.40370393e-01 -8.07592750e-01 -6.06287003e-01
-2.31960826e-02 1.14193690e+00 5.56214213e-01 3.69051844e-01
-3.83114666e-01 6.45600379e-01 -9.39842522e-01 6.40852213e-01
2.25767806e-01 1.00534856e+00 2.27754444e-01 1.05404437e+00
9.16500911e-02 1.10020053e+00 -1.31079048e-01 -2.45425195e-01
1.07501447e-01 -2.09462211e-01 -1.01581478e+00 -6.89734891e-02
5.00699997e-01 -6.85941875e-01 -5.22584438e-01 5.77648520e-01
5.16478658e-01 -8.34945858e-01 9.63973045e-01 -9.48397458e-01
-7.06111267e-02 5.33453465e-01 -5.01367271e-01 6.87008977e-01
-1.53619140e-01 4.67346534e-02 1.10915530e+00 -1.07568073e+00
5.45602977e-01 5.84134221e-01 1.27204573e+00 7.02966213e-01
-7.46428847e-01 -6.20399952e-01 -5.11820674e-01 -1.08514264e-01
-1.10129845e+00 -5.56875244e-02 -1.06735721e-01 -8.04582715e-01
8.90609384e-01 2.27717012e-01 9.80417609e-01 8.76157403e-01
7.66569853e-01 1.03005576e+00 5.75750709e-01 -1.25000894e-01
-3.55765298e-02 -1.46460682e-01 9.27055627e-03 1.45779586e+00
8.81629467e-01 1.48770183e-01 -1.79459170e-01 -5.04436970e-01
8.56001675e-01 3.19017231e-01 1.30412534e-01 -4.07444626e-01
-1.67790091e+00 7.43232608e-01 1.02273583e+00 2.75041699e-01
-4.81471360e-01 6.60432398e-01 4.08433110e-01 -2.92318135e-01
2.78470993e-01 2.44769335e-01 -6.91571012e-02 4.76624578e-01
-9.21948135e-01 2.34343365e-01 5.16522706e-01 6.51537895e-01
2.96052366e-01 -7.95894265e-01 -5.09135067e-01 4.26409245e-01
5.91189981e-01 3.60296935e-01 4.84705150e-01 -6.20504916e-01
-1.12374380e-01 5.66017866e-01 -3.44788581e-01 -1.44014403e-01
-8.68061662e-01 -5.06394982e-01 -1.17547011e+00 1.21011570e-01
2.09393501e-01 -5.28967194e-02 -1.26035082e+00 9.33299422e-01
3.28671694e-01 1.36717647e-01 -1.99995145e-01 7.70807385e-01
1.41415882e+00 -6.21607900e-02 2.14850232e-01 3.38773012e-01
1.70458794e+00 -1.05898917e+00 -2.75387049e-01 -2.41237089e-01
1.06563282e+00 -3.61626774e-01 5.03324687e-01 -1.94964092e-02
-9.27594960e-01 -1.69896692e-01 -7.84905493e-01 -1.45961642e-01
8.49888995e-02 4.83145058e-01 8.59598935e-01 3.48051131e-01
-1.40561080e+00 7.93161392e-02 -1.24764311e+00 -5.68896413e-01
8.42754841e-01 5.12441039e-01 -3.25842679e-01 -2.75554806e-01
-8.74760449e-01 9.64239299e-01 1.52243391e-01 -3.04477721e-01
-1.48465419e+00 -1.07896042e+00 -4.30337429e-01 -1.45165354e-01
2.12860271e-01 -1.28088057e+00 2.03141952e+00 -7.67266527e-02
-6.31236374e-01 1.62302780e+00 -1.90281004e-01 -7.29683280e-01
7.78145790e-01 2.75425732e-01 6.28228784e-02 5.94551675e-02
5.28760731e-01 8.41767490e-01 3.90584886e-01 -6.79092765e-01
-1.14285064e+00 -2.38301560e-01 -3.26050669e-01 2.15398461e-01
6.20879047e-02 -1.47170469e-01 -7.14023948e-01 -3.31119359e-01
2.03011230e-01 -1.17238653e+00 -7.90019214e-01 5.28851688e-01
-5.64564586e-01 -6.27176225e-01 3.74895990e-01 -2.60342121e-01
8.32646728e-01 -1.94418597e+00 -1.51399998e-02 1.71741366e-01
1.09760582e+00 -4.63190854e-01 3.61324638e-01 -3.52877766e-01
-5.20735309e-02 4.02023077e-01 2.30673939e-01 -1.41168132e-01
-3.18147957e-01 1.29421785e-01 5.74849486e-01 3.33163172e-01
1.62891179e-01 1.29577446e+00 -6.88054383e-01 -1.00597537e+00
2.84785293e-02 2.18205839e-01 -5.73743820e-01 1.38782158e-01
6.43343776e-02 -7.02347513e-03 -6.51868224e-01 1.10837388e+00
1.36086509e-01 -1.06725526e+00 -6.50643092e-03 -2.72644043e-01
1.36059687e-01 3.46750677e-01 -2.63776541e-01 1.32747304e+00
-3.15921247e-01 3.57779115e-01 6.59052134e-02 -3.41051906e-01
2.74782807e-01 7.90658951e-01 9.82836723e-01 4.76460792e-02
7.83423930e-02 5.00266552e-01 1.40602425e-01 -3.75923246e-01
-2.67138869e-01 -2.43997261e-01 -7.50527754e-02 5.79204202e-01
-9.75009650e-02 -6.19033217e-01 -1.25244707e-01 3.32870305e-01
1.98462176e+00 -6.47607923e-01 7.53059089e-01 -3.31623733e-01
2.28191525e-01 3.81827652e-01 3.75127316e-01 7.54428864e-01
-5.87048292e-01 7.44516373e-01 7.91372061e-01 -7.65431404e-01
-9.63255048e-01 -1.13728940e+00 -2.01131031e-01 9.24765706e-01
-4.34210390e-01 -1.03922293e-01 -3.54808718e-01 -1.15485001e+00
4.35901225e-01 -6.67536911e-03 -1.16923451e+00 9.77029577e-02
-3.87015998e-01 -7.62714803e-01 5.76065361e-01 7.75385737e-01
2.55889952e-01 -9.16000009e-01 -5.24131954e-01 6.14918061e-02
-1.52882412e-01 -7.03839004e-01 -3.37972701e-01 6.86754346e-01
-1.07048738e+00 -1.46398091e+00 -1.01119876e+00 -1.11626625e+00
8.02027881e-01 4.15742099e-01 1.43401921e+00 3.90436471e-01
-8.19843113e-01 2.00264141e-01 1.44567937e-01 -6.05935276e-01
-9.18313384e-01 4.28121269e-01 -1.00594722e-01 -6.40088320e-01
5.50087214e-01 1.73091665e-01 -5.70352197e-01 1.87497899e-01
-4.76785749e-01 2.67505318e-01 1.24658430e+00 6.51104212e-01
1.04610467e+00 -8.29046294e-02 1.33481398e-01 -1.05258930e+00
3.38508189e-01 -7.41185486e-01 -2.97657818e-01 4.02988613e-01
-3.38891685e-01 -1.22996643e-01 -2.25825578e-01 6.56941459e-02
-4.81027424e-01 7.48697460e-01 -1.40086293e-01 -4.67794031e-01
-8.18104669e-02 3.42834532e-01 5.79072297e-01 -1.30092710e-01
1.10635698e+00 -2.03312427e-01 5.86083997e-03 4.23280932e-02
-2.79681832e-01 3.94103587e-01 5.34377217e-01 -4.49962169e-02
7.40904450e-01 5.26481450e-01 3.02544832e-01 -3.24810386e-01
-1.19782996e+00 -7.41993904e-01 -7.56670535e-01 -3.60896587e-01
1.21035051e+00 -1.35938275e+00 -5.16124547e-01 2.52658993e-01
-9.61229324e-01 -3.70195359e-01 -3.28816086e-01 6.96334779e-01
-1.78788185e-01 -2.57405221e-01 -9.45870221e-01 -7.42980838e-02
-5.49875736e-01 -1.23818374e+00 1.34618032e+00 2.29711775e-02
-4.36601251e-01 -8.29842389e-01 3.01074386e-01 3.77933413e-01
6.43827438e-01 6.93890005e-02 8.68180275e-01 -6.56609237e-01
-6.69879794e-01 -7.51905084e-01 -3.81175816e-01 -1.82096839e-01
1.94249541e-01 7.12220669e-02 -8.66744041e-01 -1.10120036e-01
-3.60687912e-01 -5.64587414e-01 1.08638704e+00 8.49879622e-01
1.43067801e+00 2.82251447e-01 -1.14256787e+00 6.07008636e-01
1.15797782e+00 -1.41091704e-01 1.21369526e-01 1.91588327e-01
7.97367990e-01 -4.51043993e-02 1.13488629e-01 7.95345455e-02
2.27701753e-01 -5.86634688e-02 7.49269843e-01 -3.88713032e-01
-3.33736628e-01 -1.24355301e-01 -1.44354969e-01 5.40122747e-01
-7.68864155e-02 -2.90653527e-01 -1.62588632e+00 6.13834620e-01
-1.34743810e+00 -4.10715163e-01 -3.62183034e-01 1.36029875e+00
5.51161289e-01 4.55791563e-01 -5.00953607e-02 -2.43247271e-01
6.78768039e-01 -1.83392316e-01 -8.70339513e-01 3.21177661e-01
2.47163385e-01 2.09388882e-01 8.94145131e-01 7.21023828e-02
-1.49797797e+00 4.49986756e-01 8.11793518e+00 5.79954743e-01
-7.01566577e-01 3.20510924e-01 9.96452928e-01 -1.90808013e-01
-1.86589554e-01 -5.18687248e-01 -7.91875184e-01 -2.48673052e-01
5.17759502e-01 -3.06479871e-01 -4.19886887e-01 1.11655581e+00
-2.90585041e-01 -1.09100185e-01 -1.47846437e+00 7.58965492e-01
-6.75643310e-02 -1.73456502e+00 -2.99051195e-01 5.70430815e-01
7.17149258e-01 9.79514122e-01 6.91156834e-02 3.64501625e-01
7.44321048e-01 -1.23475611e+00 3.32399011e-02 5.43342233e-01
1.10850620e+00 -2.92629153e-01 1.08973622e+00 3.32024425e-01
-1.13552403e+00 2.36179829e-01 -4.98477407e-02 4.97209251e-01
-4.01236236e-01 6.88269615e-01 -1.83897018e+00 2.76423991e-02
8.23064923e-01 8.81349206e-01 -9.86177742e-01 1.24820650e+00
-2.27359056e-01 7.65793204e-01 -4.26677644e-01 -7.31136650e-02
1.34361774e-01 9.01757240e-01 1.60387635e-01 1.15396595e+00
5.04264772e-01 2.89951026e-01 3.17659229e-01 6.03733182e-01
-4.86706644e-01 -2.52498567e-01 -5.77900410e-01 1.78065728e-02
3.49558681e-01 1.59391558e+00 -1.00430715e+00 -5.38179755e-01
-4.54593182e-01 6.88449264e-01 2.21274653e-03 -4.92082208e-01
-8.00735295e-01 3.31688106e-01 2.45589212e-01 4.84253228e-01
9.48438048e-03 4.62211072e-01 -4.27114069e-01 -8.73701096e-01
-4.04381901e-01 -4.76483494e-01 7.53937006e-01 -8.02861869e-01
-1.52141976e+00 4.75869656e-01 -3.46789598e-01 -1.15598702e+00
-2.40936041e-01 -7.71072090e-01 -7.95480430e-01 4.19623524e-01
-1.10296822e+00 -1.13377357e+00 -6.92021370e-01 3.69882882e-01
4.12730843e-01 -3.26392174e-01 9.34322536e-01 1.28470808e-01
-3.19898069e-01 4.54835474e-01 -2.33507782e-01 3.40905666e-01
6.82397246e-01 -1.53685498e+00 8.79931271e-01 1.77164108e-01
-4.70493942e-01 -2.25385636e-01 8.01446587e-02 -7.45690227e-01
-1.25984800e+00 -1.32821870e+00 1.00532985e+00 -9.29042697e-01
7.41388023e-01 5.73002212e-02 -6.38009191e-01 9.00915623e-01
-1.53813198e-01 6.64831638e-01 7.04084992e-01 -1.37206078e-01
-9.53310132e-02 2.30967164e-01 -1.14231241e+00 1.45392179e-01
9.52833056e-01 -3.80639732e-01 -2.68019438e-01 8.69841397e-01
5.34694254e-01 -7.47909367e-01 -8.25068474e-01 7.07411647e-01
5.32915413e-01 -8.25954437e-01 9.62583065e-01 -3.66955847e-01
5.69670022e-01 2.47625019e-02 1.18734375e-01 -8.82970810e-01
-1.05563521e+00 2.89568186e-01 1.75797299e-01 2.58814283e-02
8.75916481e-01 -1.63602754e-01 1.45756137e+00 3.34994316e-01
-3.62824351e-01 -1.03574300e+00 -9.05538559e-01 -1.01120055e-01
3.10363680e-01 -2.13593721e-01 2.49497503e-01 6.76488578e-01
-5.12371361e-01 -8.75249505e-04 3.10379326e-01 1.56538993e-01
6.26187980e-01 -3.09026986e-01 5.41259944e-01 -1.44457185e+00
-3.38790119e-01 -6.89449310e-01 -3.65801454e-01 -4.86038029e-01
-3.43241751e-01 -1.28286529e+00 1.90907717e-01 -2.00024056e+00
8.98982227e-01 -3.38171631e-01 -5.58367372e-01 7.55898476e-01
-1.80554986e-01 5.20821393e-01 -1.85872748e-01 5.67130923e-01
-7.82581925e-01 -3.98382902e-01 1.13934743e+00 -3.21389794e-01
4.65620041e-01 5.02734065e-01 -8.16908479e-01 8.78508091e-01
6.05832219e-01 -6.87783539e-01 1.91233158e-01 -5.18017173e-01
2.50367641e-01 3.98661554e-01 5.82962990e-01 -1.34440207e+00
4.11682308e-01 2.10333802e-02 9.98883426e-01 -1.24276781e+00
5.77737670e-03 -7.27599263e-01 2.17580840e-01 1.34801066e+00
-4.18302298e-01 -9.27480608e-02 1.57059997e-01 6.06115818e-01
-6.53949231e-02 2.07450017e-02 6.96529090e-01 -4.95070130e-01
-2.83512354e-01 7.40896821e-01 -8.36239815e-01 -2.44357944e-01
1.16042161e+00 1.28932849e-01 -8.50574002e-02 5.27620912e-02
-7.58320510e-01 5.67224205e-01 1.55632704e-01 2.23800942e-01
4.51223463e-01 -1.51196992e+00 -1.09030461e+00 1.20938942e-01
3.59018505e-01 6.19184911e-01 7.05792680e-02 9.47409272e-01
-8.69926810e-01 6.13041461e-01 6.18062094e-02 -1.02592063e+00
-1.39964855e+00 1.41672090e-01 1.08509159e+00 -8.92596245e-01
-4.72751796e-01 1.66761076e+00 5.19748688e-01 -5.05046725e-01
4.14371789e-01 -7.13103890e-01 -3.20911966e-02 -1.53807133e-01
2.53897548e-01 -2.02861317e-02 5.02401888e-01 -2.33296528e-02
-7.97985375e-01 3.90853822e-01 -3.16982955e-01 -8.53675157e-02
1.10167181e+00 7.42820144e-01 -1.73926428e-01 3.69380176e-01
1.14760673e+00 -2.92874396e-01 -5.79864204e-01 -3.91384542e-01
4.66048047e-02 2.18059227e-01 3.81375372e-01 -9.22388434e-01
-1.10653102e+00 6.12251937e-01 7.91557550e-01 -1.30186722e-01
8.60923946e-01 7.02885747e-01 7.47631133e-01 6.45132124e-01
9.21184719e-02 -4.39407915e-01 1.13450214e-01 3.40611160e-01
5.68229198e-01 -1.74286306e+00 3.53939831e-01 -3.10648650e-01
-5.89854717e-01 1.22666454e+00 7.82646954e-01 -3.68747234e-01
1.08828950e+00 6.24066234e-01 1.08625159e-01 -7.63573349e-01
-1.25046468e+00 -2.88594484e-01 4.33691263e-01 2.87636071e-01
7.65823841e-01 4.99334037e-01 2.75108099e-01 3.60900223e-01
-2.59451926e-01 5.00030696e-01 2.16635376e-01 9.18221295e-01
-8.40449929e-01 -4.42914009e-01 -5.11657536e-01 1.09659839e+00
-6.08323932e-01 3.12648639e-02 -8.15805793e-01 9.42272186e-01
1.71406940e-01 1.64961278e-01 2.33327717e-01 -4.58544821e-01
3.73273194e-01 -5.73070832e-02 2.66539872e-01 -9.92348850e-01
-7.70177960e-01 2.10047394e-01 -9.97292548e-02 -3.99013221e-01
-1.42063618e-01 -7.23326027e-01 -1.28072977e+00 -2.47218795e-02
-2.18019247e-01 -5.39685972e-02 6.47636116e-01 6.81764185e-01
2.96383388e-02 9.32793796e-01 5.22337914e-01 -6.18141949e-01
-3.19034576e-01 -9.10645127e-01 -2.10660174e-01 -2.59428173e-01
4.85822767e-01 -3.34750593e-01 -4.97366965e-01 -1.31868601e-01] | [15.392788887023926, -2.161226511001587] |
cce6e77e-1cbc-4150-8ec8-f8ec06f10520 | skew-class-balanced-re-weighting-for-unbiased | 2301.00351 | null | https://arxiv.org/abs/2301.00351v3 | https://arxiv.org/pdf/2301.00351v3.pdf | Skew Class-balanced Re-weighting for Unbiased Scene Graph Generation | An unbiased scene graph generation (SGG) algorithm referred to as Skew Class-balanced Re-weighting (SCR) is proposed for considering the unbiased predicate prediction caused by the long-tailed distribution. The prior works focus mainly on alleviating the deteriorating performances of the minority predicate predictions, showing drastic dropping recall scores, i.e., losing the majority predicate performances. It has not yet correctly analyzed the trade-off between majority and minority predicate performances in the limited SGG datasets. In this paper, to alleviate the issue, the Skew Class-balanced Re-weighting (SCR) loss function is considered for the unbiased SGG models. Leveraged by the skewness of biased predicate predictions, the SCR estimates the target predicate weight coefficient and then re-weights more to the biased predicates for better trading-off between the majority predicates and the minority ones. Extensive experiments conducted on the standard Visual Genome dataset and Open Image V4 \& V6 show the performances and generality of the SCR with the traditional SGG models. | ['Chang D. Yoo', 'Haeyong Kang'] | 2023-01-01 | null | null | null | null | ['scene-graph-generation', 'unbiased-scene-graph-generation'] | ['computer-vision', 'computer-vision'] | [ 4.75591868e-01 3.97886246e-01 -4.29227084e-01 -3.06213200e-01
-5.28179348e-01 -1.24882020e-01 5.74922323e-01 1.32039621e-01
2.67203245e-02 8.16673577e-01 3.28795493e-01 -2.22595245e-01
-1.68858096e-01 -8.49396050e-01 -6.72295570e-01 -9.35474873e-01
2.41058379e-01 3.10442895e-01 7.18109846e-01 1.98941529e-02
2.29576126e-01 -3.99981886e-02 -1.81753111e+00 2.86911070e-01
1.18120313e+00 1.05500245e+00 1.27671123e-01 2.14723691e-01
-1.32267505e-01 8.35183978e-01 -5.82792580e-01 -7.95839429e-01
2.90099919e-01 -3.64773124e-01 -2.06357047e-01 -1.96629971e-01
4.70634311e-01 -5.43532334e-02 -1.79005876e-01 1.30292523e+00
4.90486026e-01 -1.59261480e-01 6.95603430e-01 -1.86516237e+00
-6.86911225e-01 6.56630814e-01 -1.06689334e+00 7.04914182e-02
2.11476848e-01 -1.56549569e-02 1.38825881e+00 -6.57710075e-01
6.69919312e-01 1.24671507e+00 5.51136792e-01 1.62105888e-01
-9.60647643e-01 -8.30170393e-01 3.55271339e-01 5.02416015e-01
-1.47144997e+00 4.90828641e-02 8.36232901e-01 -4.83120829e-01
5.65146267e-01 4.08388078e-01 4.26132977e-01 9.24292922e-01
2.76473433e-01 8.70639861e-01 9.48001623e-01 -1.98532104e-01
-6.35285974e-02 1.97381526e-01 1.04824327e-01 4.64845419e-01
6.94891036e-01 2.39869028e-01 -8.49937975e-01 -8.37228075e-02
1.76675484e-01 -2.70536482e-01 -5.94789982e-01 -8.57732832e-01
-1.06217730e+00 6.10671461e-01 4.96569633e-01 -2.51617134e-01
-2.06677482e-01 -1.91505283e-01 4.38734233e-01 -8.89564455e-02
6.74104571e-01 2.13743821e-01 -4.44404572e-01 5.52400723e-02
-7.94971645e-01 4.44683045e-01 4.48096782e-01 1.38521349e+00
7.53335476e-01 -1.03299931e-01 -9.87900972e-01 6.40043557e-01
3.69216561e-01 4.75409508e-01 3.08965951e-01 -3.27898353e-01
3.96462411e-01 8.50245416e-01 1.07978374e-01 -1.41062295e+00
-1.09599568e-01 -7.16148019e-01 -6.87955081e-01 1.54477268e-01
4.19481009e-01 1.30695879e-01 -7.50313342e-01 1.95949781e+00
5.90218544e-01 -2.73111612e-02 -3.63083519e-02 7.53357291e-01
9.90043104e-01 4.21948582e-01 4.24191892e-01 -1.64571211e-01
1.17976558e+00 -1.05379140e+00 -5.15891254e-01 -1.99477747e-01
4.99573857e-01 -9.29862142e-01 1.25105274e+00 1.02947161e-01
-4.94129479e-01 -5.16864836e-01 -1.06305838e+00 7.24513531e-02
-1.64700001e-01 1.38875142e-01 6.88590646e-01 6.93695962e-01
-6.63282156e-01 5.03966033e-01 -3.93275708e-01 -2.71215379e-01
6.71363771e-01 -1.60094440e-01 -1.05694242e-01 -1.28932893e-01
-9.89098430e-01 5.58754802e-01 6.51845396e-01 -3.82738143e-01
-4.84834254e-01 -1.14440024e+00 -6.88035548e-01 1.54186174e-01
6.08986259e-01 -5.39802015e-01 8.42144489e-01 -9.37811673e-01
-1.09762657e+00 8.50049794e-01 -8.32680464e-02 -4.44901586e-01
1.02038062e+00 -2.34725475e-01 -2.43623942e-01 -2.97968686e-01
3.47436875e-01 4.62279916e-01 8.53263259e-01 -1.38336039e+00
-8.94109726e-01 -4.28483397e-01 -1.24053419e-01 4.67915386e-01
-8.94786492e-02 -4.16903526e-01 -2.11309418e-01 -8.57943714e-01
2.91460156e-01 -7.65914023e-01 2.61382163e-01 -1.82431623e-01
-7.62539387e-01 -2.40296885e-01 9.40670729e-01 -5.04476070e-01
1.41765821e+00 -2.41639543e+00 -4.23091710e-01 1.30704358e-01
1.85743615e-01 1.55405059e-01 -8.40622783e-02 2.79790103e-01
-2.26183191e-01 -1.01901978e-01 3.35521325e-02 1.69293851e-01
4.60792892e-02 4.59200516e-02 -5.38572907e-01 4.43594187e-01
1.70944050e-01 7.60567427e-01 -1.18313825e+00 -5.51726282e-01
3.49900872e-02 2.43056536e-01 -5.05930603e-01 2.15971127e-01
-2.52746999e-01 1.95877314e-01 -3.20297301e-01 8.24562073e-01
1.08241260e+00 -2.06419379e-01 6.49564713e-02 -5.82278073e-01
6.44360110e-02 8.13116357e-02 -9.59724069e-01 8.32901657e-01
1.66743785e-01 2.72801727e-01 -4.42422599e-01 -6.43197894e-01
1.14667773e+00 -3.00160110e-01 2.44644344e-01 -8.13772082e-01
-4.40792069e-02 2.89917558e-01 -1.62912197e-02 -2.16071770e-01
6.72914088e-01 -1.79461986e-01 5.26127368e-02 -3.31593957e-03
-2.41441220e-01 -3.33887599e-02 1.03763245e-01 2.69133329e-01
7.03309894e-01 6.28509402e-01 5.86416543e-01 -4.43510920e-01
4.41806912e-01 1.57393888e-01 9.61971998e-01 6.53512478e-01
-4.70546007e-01 7.75240541e-01 7.76634455e-01 -3.59113738e-02
-8.53228986e-01 -1.10297549e+00 5.73571920e-02 1.13451087e+00
6.03200495e-01 -3.16375226e-01 -4.58580017e-01 -9.39259112e-01
2.51486510e-01 1.01079059e+00 -4.62749958e-01 -5.29571116e-01
-1.25773564e-01 -8.59394908e-01 4.34335738e-01 5.49587190e-01
6.40867591e-01 -7.74170816e-01 -6.60678804e-01 -1.06319048e-01
-1.85118988e-01 -9.48602617e-01 -3.33865792e-01 -9.76357237e-02
-3.03421021e-01 -1.28289938e+00 -7.10443616e-01 -4.64530796e-01
7.21185803e-01 4.57881778e-01 1.21480036e+00 -1.40771881e-01
4.38603349e-02 -1.45338655e-01 -5.75833201e-01 -8.35720122e-01
-6.58899769e-02 -2.16213390e-01 -1.62511960e-01 1.99178860e-01
3.54189247e-01 -2.85684854e-01 -5.89614749e-01 3.64544034e-01
-5.49341381e-01 2.83542931e-01 4.97827262e-01 1.04139316e+00
8.63620937e-01 1.66031927e-01 6.21912420e-01 -1.14530694e+00
3.83909166e-01 -3.69364768e-01 -7.12909341e-01 3.64991635e-01
-9.74775314e-01 -1.01359978e-01 3.21833789e-01 -4.25840646e-01
-1.29489708e+00 -3.36528361e-01 2.23854616e-01 -3.91016483e-01
3.05594504e-01 1.24956965e-01 -5.86690307e-01 1.25025464e-02
2.90563643e-01 1.68045372e-01 -2.23504886e-01 -7.88814053e-02
4.03799415e-01 3.76035303e-01 4.23718512e-01 -3.70953172e-01
8.38843286e-01 5.27266920e-01 6.87635317e-02 -5.48120975e-01
-1.07245386e+00 -4.74243820e-01 9.25099254e-02 -1.77775428e-01
5.44396877e-01 -8.09772015e-01 -4.52396542e-01 7.45046198e-01
-8.14201415e-01 -4.60058032e-03 -7.11687982e-01 3.52986097e-01
-5.24237931e-01 5.39969265e-01 1.12527553e-02 -8.65109026e-01
-2.72768587e-01 -9.29921865e-01 1.09501779e+00 3.82434517e-01
-2.02219412e-01 -4.71921980e-01 -7.95149356e-02 3.19566965e-01
3.34463090e-01 4.40482348e-01 1.28801930e+00 -9.20749307e-01
-4.96680409e-01 -1.52882382e-01 -8.45605850e-01 3.34648281e-01
-3.27349384e-03 1.52141854e-01 -1.06115639e+00 -1.64576590e-01
-2.29884237e-01 -8.58414359e-03 8.43999445e-01 3.91691744e-01
1.12344432e+00 -9.06800926e-02 -4.10903931e-01 5.41772127e-01
1.49731243e+00 1.61481217e-01 6.41862690e-01 4.09846395e-01
7.86988795e-01 5.47403634e-01 1.32446969e+00 5.43869853e-01
6.02046549e-01 4.20413107e-01 5.60894132e-01 7.98120424e-02
-3.92503470e-01 -8.09624851e-01 4.04651053e-02 3.17356884e-01
2.08998442e-01 -5.77179730e-01 -8.17173660e-01 7.10189939e-01
-1.91866446e+00 -5.96072614e-01 -4.15560186e-01 2.37742400e+00
5.47772408e-01 3.99156779e-01 -1.77562639e-01 2.46809244e-01
9.41783488e-01 5.22710502e-01 -5.32745302e-01 -8.42400640e-03
-5.20094812e-01 -1.67655483e-01 7.63124049e-01 1.46688208e-01
-1.07290602e+00 8.89458954e-01 5.44081974e+00 1.31798851e+00
-9.38656092e-01 -1.04294963e-01 7.78099418e-01 2.01024100e-01
-5.19623995e-01 3.50800127e-01 -1.00766456e+00 6.74430549e-01
3.08303773e-01 -6.19136572e-01 -6.80482909e-02 1.10762918e+00
-1.31677911e-01 -2.88273036e-01 -7.95739949e-01 9.75374997e-01
1.65155351e-01 -9.07297671e-01 3.68108243e-01 -1.88127682e-01
7.97455430e-01 -1.77037403e-01 1.71422900e-03 5.07399082e-01
2.10530400e-01 -6.08820617e-01 9.60700095e-01 4.80508775e-01
5.79124093e-01 -6.98037386e-01 1.02481318e+00 2.17972174e-01
-1.09274590e+00 -5.19452728e-02 -4.19115096e-01 6.39734268e-02
1.65188611e-02 1.03108728e+00 -8.50315571e-01 8.58595312e-01
8.07750821e-01 5.22471070e-01 -7.11745560e-01 1.12229049e+00
-4.48129326e-01 5.22733569e-01 -1.76346496e-01 3.39825042e-02
-1.63833618e-01 -1.84675530e-01 7.87023544e-01 8.47317696e-01
2.76853144e-01 -4.43729252e-01 3.01135033e-02 6.54427707e-01
6.69960352e-03 3.20111096e-01 -5.83682954e-01 -1.67089682e-02
4.95137691e-01 1.07584488e+00 -6.69601679e-01 -1.71317607e-01
-2.31149703e-01 5.09197116e-01 1.77327126e-01 2.49569759e-01
-1.03577685e+00 -2.71580666e-01 4.69106883e-01 2.86952198e-01
5.07602930e-01 3.70508820e-01 -7.51684070e-01 -9.61703241e-01
1.23702578e-01 -7.08494484e-01 6.06780648e-01 -7.71216989e-01
-1.53704357e+00 4.02703822e-01 1.83304876e-01 -1.38493049e+00
2.99866050e-01 -2.86547154e-01 -6.49603069e-01 7.99163401e-01
-1.93572199e+00 -1.31940734e+00 -5.37189960e-01 3.03557754e-01
2.82591850e-01 1.16284013e-01 2.18601763e-01 2.77284563e-01
-5.11393845e-01 8.45134616e-01 -2.81826138e-01 -4.79761481e-01
9.56056297e-01 -1.17281318e+00 5.12210429e-02 9.49547470e-01
-2.93627471e-01 1.95025876e-01 9.23956394e-01 -9.38710511e-01
-7.49353647e-01 -1.22516918e+00 1.13827801e+00 -3.17070156e-01
5.44241846e-01 -4.41228552e-03 -8.28578174e-01 3.40721488e-01
-7.82907456e-02 -1.29341483e-01 6.45155549e-01 -1.52939096e-01
-6.23252749e-01 -3.58848512e-01 -1.28440368e+00 7.47702956e-01
1.21823692e+00 -1.41804451e-02 -3.31181377e-01 2.45769575e-01
8.13561320e-01 -3.39955628e-01 -5.17937422e-01 1.12739384e+00
4.47893411e-01 -1.16763115e+00 8.61154497e-01 -3.15631509e-01
4.78441179e-01 -5.07592022e-01 -4.18664187e-01 -1.04281652e+00
-3.47476214e-01 -9.75932330e-02 -1.23793557e-01 1.55409908e+00
3.40234637e-01 -7.00707316e-01 8.70841444e-01 2.65682250e-01
4.50495556e-02 -9.28062677e-01 -7.54553258e-01 -7.73146749e-01
-3.30264866e-01 -1.52746961e-01 8.84476900e-01 7.04701960e-01
-4.10445869e-01 2.21476510e-01 -3.43911856e-01 9.87299234e-02
7.96403408e-01 3.81016642e-01 1.00359643e+00 -1.09973097e+00
-1.87923133e-01 -2.78191090e-01 -5.52868724e-01 -7.04333127e-01
-2.79357433e-01 -7.69869030e-01 1.31998688e-01 -1.28476083e+00
4.69988495e-01 -3.71330202e-01 -2.95174450e-01 2.47383013e-01
-7.46552229e-01 2.08454356e-01 2.48930350e-01 9.79962572e-02
-6.49149060e-01 8.51779997e-01 1.30555642e+00 -5.81188910e-02
1.14558347e-01 -7.11463317e-02 -1.10427284e+00 9.39618826e-01
6.38313711e-01 -3.67018729e-01 -7.96656191e-01 -3.67033780e-02
2.71031350e-01 -4.60537881e-01 3.15183699e-01 -9.79841053e-01
2.99992450e-02 -2.53297240e-01 2.08003566e-01 -1.03285730e+00
-3.39302868e-02 -6.63745701e-01 5.09047918e-02 3.85164350e-01
-3.35517749e-02 -1.56773016e-01 -4.96029854e-03 8.68072569e-01
-2.96655476e-01 1.48305342e-01 8.94581318e-01 1.18661322e-01
-8.79934669e-01 1.86001599e-01 3.16938519e-01 4.18972909e-01
1.37589955e+00 -5.22307515e-01 -5.47385752e-01 -2.67037213e-01
-1.94231108e-01 3.82427990e-01 4.64026511e-01 4.32978153e-01
6.00239694e-01 -1.27042305e+00 -7.25103736e-01 3.36142689e-01
6.42336011e-01 -7.97653347e-02 4.95468348e-01 7.94806004e-01
-3.84325087e-01 1.33807093e-01 -1.04240827e-01 -6.00276709e-01
-1.50435340e+00 5.05313456e-01 -2.40145084e-02 -6.89939320e-01
-4.97053176e-01 1.13562930e+00 6.44968450e-01 -3.83143514e-01
1.77593604e-01 -1.17834382e-01 -3.17757368e-01 2.01661706e-01
2.03855649e-01 6.34257317e-01 -1.89532284e-02 -6.89790726e-01
-5.55195332e-01 2.74403155e-01 -2.33051199e-02 4.56190318e-01
1.08850503e+00 -1.41283020e-01 1.87165719e-02 2.12927297e-01
7.91754246e-01 2.82585293e-01 -1.29933238e+00 -1.08606525e-01
1.01899490e-01 -9.00668561e-01 -2.55839020e-01 -8.25791419e-01
-1.00422323e+00 5.15148699e-01 5.04285812e-01 8.64196047e-02
1.23243690e+00 -1.01325728e-01 6.51470006e-01 -2.96789080e-01
6.32196367e-01 -9.10020471e-01 -7.64197260e-02 4.14996237e-01
6.52959228e-01 -1.09389544e+00 1.75105006e-01 -9.18327451e-01
-1.05582869e+00 4.87754226e-01 7.68747330e-01 -1.36318266e-01
4.81722057e-01 -1.35115907e-02 -5.36025353e-02 -2.04553172e-01
-5.46662331e-01 -1.66549623e-01 3.20378631e-01 8.25874627e-01
2.05453932e-01 4.15140390e-01 -7.71132290e-01 6.38159752e-01
-4.97586787e-01 -1.77254245e-01 3.03228557e-01 6.94368362e-01
-1.13444097e-01 -7.93313622e-01 -1.06536455e-01 6.35326922e-01
-4.30632442e-01 -2.30222061e-01 -3.71727645e-01 7.37258613e-01
3.92629445e-01 8.69605362e-01 -2.21469983e-01 -5.25755227e-01
4.86478478e-01 -7.82101154e-02 1.81586891e-01 -3.50593925e-01
-3.10842156e-01 -1.36464834e-01 2.47892305e-01 -6.25735521e-01
-3.43885362e-01 -3.94264847e-01 -9.48578000e-01 -1.38338849e-01
-6.98868454e-01 -1.05552725e-01 2.59433478e-01 4.87499863e-01
3.79425347e-01 7.89231479e-01 4.45829421e-01 -2.31234506e-01
-7.58475900e-01 -7.01728165e-01 -7.59854674e-01 6.97307944e-01
-4.56561111e-02 -9.57652271e-01 -5.69457471e-01 -4.20874506e-01] | [10.262526512145996, 1.7900077104568481] |
cf848175-aa5e-4607-b2e5-611173997175 | flot-scene-flow-on-point-clouds-guided-by | 2007.11142 | null | https://arxiv.org/abs/2007.11142v1 | https://arxiv.org/pdf/2007.11142v1.pdf | FLOT: Scene Flow on Point Clouds Guided by Optimal Transport | We propose and study a method called FLOT that estimates scene flow on point clouds. We start the design of FLOT by noticing that scene flow estimation on point clouds reduces to estimating a permutation matrix in a perfect world. Inspired by recent works on graph matching, we build a method to find these correspondences by borrowing tools from optimal transport. Then, we relax the transport constraints to take into account real-world imperfections. The transport cost between two points is given by the pairwise similarity between deep features extracted by a neural network trained under full supervision using synthetic datasets. Our main finding is that FLOT can perform as well as the best existing methods on synthetic and real-world datasets while requiring much less parameters and without using multiscale analysis. Our second finding is that, on the training datasets considered, most of the performance can be explained by the learned transport cost. This yields a simpler method, FLOT$_0$, which is obtained using a particular choice of optimal transport parameters and performs nearly as well as FLOT. | ['Renaud Marlet', 'Gilles Puy', 'Alexandre Boulch'] | 2020-07-22 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/6287_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123730528.pdf | eccv-2020-8 | ['scene-flow-estimation'] | ['computer-vision'] | [-1.51206404e-01 -1.11768968e-01 8.12009946e-02 -1.20995320e-01
-2.48963520e-01 -6.06842756e-01 6.93718314e-01 1.80814072e-01
-3.74050200e-01 4.40038264e-01 -1.60225764e-01 -1.47578269e-01
-3.24352682e-01 -9.97661591e-01 -9.37879980e-01 -3.07014793e-01
-3.06384802e-01 6.27830863e-01 4.15211439e-01 -3.18690807e-01
5.34667253e-01 7.12618887e-01 -1.47216558e+00 -2.74243563e-01
8.74691069e-01 9.65208709e-01 2.91701443e-02 7.83455014e-01
-1.40501633e-01 5.73914886e-01 -2.95508325e-01 -4.16242212e-01
8.30208719e-01 -4.13569719e-01 -1.02573121e+00 6.09739386e-02
9.95730221e-01 -2.35464200e-01 -5.41176200e-01 7.90630817e-01
1.25936925e-01 3.38600725e-01 6.45952284e-01 -1.42457604e+00
-3.36231768e-01 1.67670906e-01 -4.08097714e-01 1.80560842e-01
1.90075487e-01 3.21195483e-01 1.18998146e+00 -8.16216648e-01
7.94140637e-01 1.19055331e+00 8.69275510e-01 8.25237408e-02
-1.42980349e+00 -3.25779796e-01 -6.31036311e-02 1.25673115e-01
-1.45582056e+00 -3.54396909e-01 6.78480685e-01 -6.04025364e-01
8.90392423e-01 8.79085734e-02 6.95313692e-01 3.94765228e-01
-5.68256825e-02 3.95746559e-01 7.15353847e-01 -3.54134381e-01
2.34870732e-01 -1.34803459e-01 -1.11064203e-01 9.72610414e-01
3.42006147e-01 -1.23712495e-01 -4.72277761e-01 -5.35129234e-02
8.45250249e-01 -1.36093631e-01 -3.70912105e-01 -9.41871524e-01
-1.57236469e+00 8.56254041e-01 8.47337961e-01 2.06589624e-01
-9.71492380e-04 5.68592608e-01 2.55870223e-01 4.17043746e-01
4.89463627e-01 5.63037455e-01 -2.96668082e-01 4.79493253e-02
-8.67422223e-01 2.12680653e-01 8.75514805e-01 9.77480531e-01
1.23037302e+00 -1.37899518e-01 2.63591051e-01 4.14844602e-01
2.50806570e-01 5.26048839e-01 4.58364040e-02 -1.55511296e+00
5.87538838e-01 5.32212615e-01 3.19307417e-01 -1.38912690e+00
-3.65780562e-01 -3.19668084e-01 -6.10864699e-01 2.90883362e-01
8.02317500e-01 -7.50614405e-02 -5.35199106e-01 1.77615607e+00
2.66328037e-01 6.15191996e-01 -1.94924980e-01 8.33760500e-01
2.96677560e-01 4.72909063e-01 -4.75900590e-01 1.62721500e-01
8.83283556e-01 -9.03416872e-01 -1.35994688e-01 -1.30386218e-01
8.24284434e-01 -6.12262428e-01 9.77794528e-01 1.30193233e-01
-1.01633322e+00 -3.79341573e-01 -1.13979971e+00 -1.41814783e-01
-3.71396273e-01 -2.17816308e-01 6.58824146e-01 4.41567838e-01
-1.38813150e+00 1.22202051e+00 -1.02803266e+00 -5.81437945e-01
4.01995808e-01 3.43622386e-01 -3.80522013e-01 1.62078924e-02
-7.08272278e-01 7.29910135e-01 -1.10655814e-01 7.82383978e-02
-4.44600344e-01 -8.94063473e-01 -8.50204408e-01 7.93259474e-04
2.31380120e-01 -1.14330220e+00 9.92836893e-01 -7.70579100e-01
-1.41648173e+00 7.58429945e-01 -4.16014940e-01 -5.30677080e-01
7.67919719e-01 1.08683780e-02 3.75268787e-01 2.94046372e-01
3.40233773e-01 7.55374491e-01 6.06659949e-01 -1.16984570e+00
-4.01860565e-01 -3.38081032e-01 1.36423588e-01 7.17492476e-02
-7.13584125e-02 -4.43941414e-01 -3.41707557e-01 -1.09061919e-01
2.25728258e-01 -1.19767666e+00 -3.71047080e-01 5.64600229e-01
-4.25659597e-01 -8.41292664e-02 6.65329695e-01 -1.50093049e-01
6.43946826e-01 -1.94300210e+00 2.27711812e-01 4.10296321e-01
5.06226003e-01 5.12493998e-02 -1.62010372e-01 6.09221280e-01
3.44275981e-02 2.84702480e-01 -5.59055805e-01 -5.23706019e-01
4.59875837e-02 2.49206871e-01 -2.36386582e-01 7.38166749e-01
2.40025431e-01 8.52943242e-01 -1.02841568e+00 -3.64062756e-01
5.67491114e-01 3.16859066e-01 -7.31633306e-01 -1.57130715e-02
-1.24689557e-01 4.38170642e-01 -4.30862218e-01 7.30653182e-02
9.89036977e-01 -5.58614314e-01 -5.12511767e-02 -1.29136443e-01
-2.50824898e-01 3.16634446e-01 -1.43029034e+00 1.94408083e+00
-4.98751938e-01 1.01978433e+00 1.26325171e-02 -1.05687201e+00
6.90123498e-01 -1.34405226e-01 8.44278097e-01 -5.23944676e-01
7.32625872e-02 3.73559862e-01 -1.56345353e-01 -4.76853609e-01
5.36928415e-01 -1.28245264e-01 3.08120102e-01 4.36319888e-01
-4.02925648e-02 -4.09456551e-01 2.44422466e-01 4.93136317e-01
1.39383507e+00 1.37778431e-01 -4.00686339e-02 -4.64870572e-01
3.24522406e-01 1.71800971e-01 3.71465772e-01 7.37748146e-01
-9.87185985e-02 6.54291153e-01 5.45799613e-01 -4.07747507e-01
-1.15663719e+00 -9.12568629e-01 -8.70123208e-02 5.21462560e-01
4.55945671e-01 -4.03333962e-01 -7.00582922e-01 -4.77193505e-01
3.52682173e-01 2.39332065e-01 -5.54954052e-01 1.82962313e-01
-7.60145128e-01 -5.03552377e-01 2.79994458e-01 3.66814345e-01
5.08520424e-01 -6.11815631e-01 -7.37575650e-01 5.96745918e-03
-1.85179725e-01 -1.33920681e+00 -4.89349872e-01 -1.46359086e-01
-8.64307106e-01 -1.18901539e+00 -4.78690505e-01 -3.98885667e-01
5.11884630e-01 6.54584944e-01 1.30550158e+00 4.71684366e-01
-1.42167851e-01 3.80236000e-01 -1.64402537e-02 -1.04970587e-02
-1.00642607e-01 2.67342955e-01 -3.23941223e-02 1.20317876e-01
2.99415947e-03 -9.61514235e-01 -7.98796594e-01 4.06007588e-01
-7.77008414e-01 -1.48213610e-01 1.38629392e-01 4.61503476e-01
2.63310075e-01 -2.78313071e-01 8.22492987e-02 -6.16301179e-01
1.88100293e-01 -3.96549970e-01 -8.82998705e-01 -1.65287331e-01
-4.36377078e-01 3.82192761e-01 7.42921472e-01 -3.79613303e-02
-4.52236503e-01 2.19961610e-02 3.54295298e-02 -5.84182739e-01
1.36527289e-02 7.08492845e-02 1.23163015e-01 -6.26623273e-01
6.50446296e-01 -2.81474560e-01 3.26804966e-02 -2.77704775e-01
4.70664620e-01 1.40188694e-01 4.29439366e-01 -5.20390272e-01
1.01733792e+00 1.10555601e+00 5.51445246e-01 -9.06072378e-01
-6.71273649e-01 -5.00988305e-01 -9.29567933e-01 -1.34266883e-01
6.68562591e-01 -5.56687176e-01 -1.22443020e+00 2.80023754e-01
-1.20283973e+00 -5.19630492e-01 -4.71469969e-01 5.46567202e-01
-8.38358998e-01 5.58918893e-01 -3.89702111e-01 -5.66348076e-01
8.01103786e-02 -1.11318839e+00 1.21743023e+00 -2.03310877e-01
-3.08576673e-02 -1.22969139e+00 3.12275648e-01 3.98694836e-02
4.93508518e-01 4.08806533e-01 6.92163289e-01 -2.34678879e-01
-1.30114543e+00 2.11030081e-01 -4.95803148e-01 1.21245988e-01
-2.46270257e-03 2.64491349e-01 -8.47091794e-01 -1.62625760e-01
-4.45998460e-02 8.81378725e-02 9.45335865e-01 4.20227945e-01
8.77343714e-01 -1.87414438e-02 -3.80449533e-01 1.10220134e+00
1.73615193e+00 -2.05489188e-01 4.58870500e-01 3.66255790e-01
9.67456102e-01 6.55403793e-01 9.13434476e-02 1.41602859e-01
5.40056050e-01 8.46963763e-01 7.13447928e-01 2.74817925e-02
-1.74403429e-01 -4.17892724e-01 -3.58358361e-02 7.50182986e-01
-7.64932260e-02 -2.39927828e-01 -1.06543195e+00 7.05864489e-01
-2.02611327e+00 -7.97995985e-01 -3.69739115e-01 2.39510655e+00
2.03730583e-01 1.64611295e-01 6.28748536e-02 2.80182101e-02
3.22230279e-01 2.68123597e-01 -4.68685031e-01 -1.42511383e-01
-8.31815079e-02 2.57967710e-01 9.14660096e-01 8.08546782e-01
-8.90877426e-01 7.66131997e-01 6.58781242e+00 3.86894166e-01
-1.19591379e+00 -2.80324165e-02 3.09299558e-01 -8.04267675e-02
-4.06247944e-01 3.32456350e-01 -4.40776736e-01 4.22301412e-01
8.18381786e-01 -1.33153394e-01 4.51871872e-01 5.86235464e-01
2.43323684e-01 -1.67463571e-01 -1.28860426e+00 9.32113051e-01
-1.18498944e-01 -1.68551445e+00 -2.48656161e-02 1.73501015e-01
6.75074458e-01 5.18457353e-01 -1.78058460e-01 -2.38187373e-01
4.85317051e-01 -7.84401119e-01 6.62309945e-01 4.67019469e-01
4.03913110e-01 -3.85794252e-01 3.87869269e-01 2.39399165e-01
-1.30804265e+00 1.98468819e-01 -4.35661793e-01 -2.33566180e-01
2.42110223e-01 7.26172328e-01 -7.53490150e-01 7.04894781e-01
5.25797129e-01 1.23159993e+00 -5.04630089e-01 1.29767489e+00
6.09467439e-02 3.77420902e-01 -7.53994286e-01 2.01366484e-01
2.06352860e-01 -6.16814077e-01 6.79678798e-01 1.05781972e+00
4.75627571e-01 -3.92314702e-01 1.56841144e-01 9.09600139e-01
-1.91404790e-01 2.14942936e-02 -1.08402979e+00 4.49639350e-01
2.21168920e-01 1.14327419e+00 -8.64612877e-01 -2.56631970e-01
-3.36863875e-01 8.52084517e-01 4.25561726e-01 3.57717276e-01
-5.65712512e-01 -4.04451996e-01 8.21447372e-01 4.07466143e-01
4.70741302e-01 -6.21515691e-01 -2.24725544e-01 -1.31068075e+00
2.86873221e-01 -1.96467623e-01 -2.56861206e-02 -8.70874643e-01
-1.19458175e+00 3.85510296e-01 6.22195266e-02 -1.41629446e+00
-1.66796774e-01 -6.94451809e-01 -7.29620814e-01 6.68998659e-01
-1.77104735e+00 -7.13703632e-01 -4.34797764e-01 4.80756968e-01
1.72081292e-01 4.16812778e-01 2.51121223e-01 3.02960575e-01
-4.81293082e-01 1.15883850e-01 -2.79098493e-03 1.36926442e-01
5.05143166e-01 -1.29316044e+00 8.31323087e-01 9.59281504e-01
2.17533872e-01 6.24960780e-01 6.92943335e-01 -3.70177567e-01
-1.34284031e+00 -9.16200280e-01 9.21700299e-01 -7.27367699e-01
1.00046420e+00 -4.15797144e-01 -8.88415694e-01 7.85749555e-01
1.97690472e-01 4.44775790e-01 -1.09340958e-02 -1.17099322e-01
-3.35050941e-01 -1.71016321e-01 -1.02077317e+00 4.67633575e-01
1.47608423e+00 -4.89164263e-01 -1.19871579e-01 4.73577529e-01
8.80546987e-01 -4.43639785e-01 -7.84097731e-01 2.22122639e-01
2.74669021e-01 -1.24375129e+00 1.00961542e+00 -4.22333717e-01
3.83475125e-01 -4.66203839e-01 -1.07863143e-01 -1.46476591e+00
-2.14280952e-02 -9.06458139e-01 2.74939418e-01 7.72795618e-01
3.15171659e-01 -8.37122381e-01 9.10884321e-01 5.67408621e-01
-9.81186405e-02 -3.38362724e-01 -9.72760201e-01 -9.44303453e-01
4.97039817e-02 -4.51321959e-01 6.32333934e-01 1.08579433e+00
-3.10902089e-01 2.88119197e-01 -1.03241526e-01 1.80425972e-01
8.37089658e-01 1.98046923e-01 1.07181621e+00 -1.44059789e+00
-1.02865689e-01 -3.57977003e-01 -6.78717434e-01 -1.38564682e+00
2.96870679e-01 -8.74862790e-01 -5.43885771e-03 -1.38006866e+00
-2.77769953e-01 -7.36061454e-01 1.22764118e-01 9.64548215e-02
1.94234520e-01 2.81326205e-01 3.24237555e-01 3.45965981e-01
-4.82848227e-01 5.18819928e-01 1.34651005e+00 2.80472077e-02
-1.03774570e-01 1.04741223e-01 -3.71356755e-01 8.02793503e-01
5.94656467e-01 -3.93243879e-01 -3.52790028e-01 -7.96228290e-01
5.55747628e-01 8.34562406e-02 6.26096249e-01 -1.00713754e+00
3.46830666e-01 -2.81147063e-01 -3.76334712e-02 -3.13247323e-01
3.96677345e-01 -8.84016991e-01 1.45230576e-01 3.74197632e-01
-1.48169100e-01 3.34820449e-01 6.94089457e-02 5.57598293e-01
1.42253600e-02 -1.96167231e-01 7.03387737e-01 -2.00019211e-01
-5.09760141e-01 4.10242051e-01 1.28970649e-02 1.12498134e-01
7.13362634e-01 -2.89622098e-01 -5.70897520e-01 -4.18722868e-01
-5.47427416e-01 1.72183052e-01 8.11377823e-01 1.89257011e-01
3.06384146e-01 -1.27518690e+00 -5.30087888e-01 3.12034607e-01
6.76452741e-02 2.65109122e-01 -8.03741962e-02 9.51750994e-01
-9.11105454e-01 4.34775174e-01 -3.52220014e-02 -9.62813795e-01
-7.56968319e-01 3.85107815e-01 7.63518333e-01 -4.50329706e-02
-7.95198917e-01 5.70052207e-01 6.18162863e-02 -4.78712082e-01
-1.92129463e-01 -6.10983670e-01 2.20808193e-01 -1.80314690e-01
1.23297192e-01 4.98856544e-01 2.35318244e-01 -7.22066402e-01
-3.79770249e-01 1.12876284e+00 4.82851446e-01 -1.43068328e-01
1.25080168e+00 -2.39239201e-01 -1.39307827e-01 3.00994754e-01
1.53282392e+00 1.71238959e-01 -1.28645718e+00 -2.73570687e-01
-3.81662920e-02 -8.21125269e-01 -8.44817162e-02 1.14669599e-01
-1.36786711e+00 1.01012886e+00 3.90866786e-01 4.26493108e-01
6.77149534e-01 -8.50450248e-02 7.83247232e-01 5.47727048e-01
4.90866989e-01 -5.44554710e-01 -1.52512453e-03 5.00284374e-01
4.57151979e-01 -1.21777272e+00 -2.39179935e-02 -5.64187944e-01
-7.87328854e-02 8.59319270e-01 3.48431706e-01 -6.69642329e-01
7.62320161e-01 1.32812560e-02 -1.05387293e-01 -2.66693443e-01
-5.90022862e-01 -3.01126540e-01 1.66807145e-01 5.29267669e-01
7.70427426e-03 -2.39161566e-01 1.84565306e-01 -7.28975773e-01
-4.02687341e-01 1.31079480e-01 5.96805990e-01 6.92599952e-01
-4.50409532e-01 -1.06695879e+00 -1.05943650e-01 3.61633003e-01
-4.65707630e-02 4.83097322e-02 -2.77813971e-01 9.10144627e-01
-1.27385437e-01 7.35732496e-01 4.33699995e-01 -2.80394524e-01
5.21283567e-01 -2.76270777e-01 4.90440726e-01 -4.41603005e-01
-3.69209021e-01 -3.79967332e-01 -2.54711568e-01 -9.64282930e-01
-8.00925553e-01 -5.51049650e-01 -1.07295096e+00 -6.79398537e-01
-9.41741839e-02 2.14933127e-01 7.15063393e-01 9.23704207e-01
4.99407262e-01 2.18390569e-01 7.85766661e-01 -1.19074845e+00
-1.40630141e-01 -5.71832538e-01 -4.15037453e-01 5.74904025e-01
7.87234783e-01 -7.17479885e-01 -7.29583859e-01 -1.29026309e-01] | [8.5244722366333, -2.057372808456421] |
24002645-75e5-49af-9bbf-88551dbd10de | double-refinement-network-for-efficient | 1811.08466 | null | http://arxiv.org/abs/1811.08466v2 | http://arxiv.org/pdf/1811.08466v2.pdf | Double Refinement Network for Efficient Indoor Monocular Depth Estimation | Monocular depth estimation is the task of obtaining a measure of distance for
each pixel using a single image. It is an important problem in computer vision
and is usually solved using neural networks. Though recent works in this area
have shown significant improvement in accuracy, the state-of-the-art methods
tend to require massive amounts of memory and time to process an image. The
main purpose of this work is to improve the performance of the latest solutions
with no decrease in accuracy. To this end, we introduce the Double Refinement
Network architecture. The proposed method achieves state-of-the-art results on
the standard benchmark RGB-D dataset NYU Depth v2, while its frames per second
rate is significantly higher (up to 18 times speedup per image at batch size 1)
and the RAM usage per image is lower. | ['Mikhail Romanov', 'Nikita Durasov', 'Valeriya Bubnova', 'Pavel Bogomolov', 'Anton Konushin'] | 2018-11-20 | null | null | null | null | ['indoor-monocular-depth-estimation'] | ['computer-vision'] | [ 4.33631748e-01 -6.96092993e-02 1.19345911e-01 -4.36819822e-01
-4.36215848e-01 5.11357281e-03 3.51512372e-01 1.61816567e-01
-1.07914627e+00 7.06456840e-01 -3.59539688e-01 -1.72529683e-01
3.25696439e-01 -9.46015716e-01 -6.67627573e-01 -7.25289166e-01
2.20294923e-01 3.39949816e-01 6.87903225e-01 3.49505357e-02
5.24567962e-01 7.25563765e-01 -1.83885682e+00 -1.10996058e-02
4.47912693e-01 1.42257524e+00 3.42995554e-01 9.09875095e-01
-6.31096885e-02 8.24964404e-01 -3.35169286e-01 -1.41053498e-01
4.84037817e-01 -3.02678108e-01 -7.71957397e-01 -4.53491230e-03
7.86612928e-01 -7.35199809e-01 -5.86128175e-01 1.23214757e+00
6.92356646e-01 3.10208299e-03 3.16404045e-01 -9.89343345e-01
-2.80003585e-02 1.28899783e-01 -7.23648548e-01 2.27587730e-01
2.15248056e-02 -1.92612931e-01 5.63042223e-01 -7.17550218e-01
5.69506645e-01 9.92350638e-01 4.12020862e-01 7.12708473e-01
-9.54233408e-01 -5.36651313e-01 3.07342652e-02 4.16788012e-01
-1.37691319e+00 -1.74196944e-01 5.20668983e-01 -9.69931260e-02
1.04831886e+00 -4.60773222e-02 8.03817093e-01 4.26111341e-01
1.38727665e-01 7.56490290e-01 1.06845093e+00 -2.74727404e-01
4.37461674e-01 -3.29197466e-01 8.78343955e-02 7.22741246e-01
3.96733880e-01 4.51691151e-02 -5.61578870e-01 4.28052694e-01
1.12501001e+00 2.33985201e-01 -4.78229314e-01 -4.19062018e-01
-9.64188635e-01 8.27542663e-01 5.27010381e-01 2.52280146e-01
-3.08348119e-01 2.94121087e-01 2.66321778e-01 1.29716173e-01
5.14861584e-01 8.04227218e-02 -5.56784034e-01 -4.76712883e-01
-1.14216805e+00 3.48008126e-01 7.22583354e-01 8.01905751e-01
9.41558480e-01 -1.07361957e-01 3.28999549e-01 6.82614207e-01
1.09391481e-01 4.73548949e-01 5.54600716e-01 -1.14826536e+00
3.58928531e-01 6.68108284e-01 -1.02149332e-02 -9.76008534e-01
-6.53084457e-01 -1.70132607e-01 -1.09815109e+00 8.71557713e-01
7.05549002e-01 -1.16969161e-01 -1.10114622e+00 1.23682141e+00
3.14351201e-01 7.91277438e-02 8.19949061e-02 9.72160578e-01
9.31895852e-01 6.44597828e-01 -4.88250226e-01 -2.25532353e-01
1.12452674e+00 -1.20099604e+00 -4.95900273e-01 -4.32331711e-01
4.61616039e-01 -7.28609920e-01 6.74901426e-01 8.14917564e-01
-1.32980740e+00 -6.54603541e-01 -1.30432904e+00 -4.10893768e-01
-1.95262492e-01 -1.07830569e-01 6.33026421e-01 6.30439699e-01
-1.20280004e+00 8.17018449e-01 -8.76715541e-01 -2.61758655e-01
3.96258235e-01 6.26762092e-01 -5.18915892e-01 -2.92113930e-01
-6.70584083e-01 6.15089774e-01 5.90844750e-01 1.27816647e-01
-5.52994251e-01 -6.64898098e-01 -8.84946704e-01 -1.07592262e-01
3.52824003e-01 -6.14619970e-01 1.29916298e+00 -7.87381351e-01
-1.72861218e+00 8.85347128e-01 -3.00489247e-01 -6.89868689e-01
7.94846237e-01 -3.96257758e-01 1.32499248e-01 3.48612756e-01
-2.71889657e-01 1.13478446e+00 7.56699145e-01 -9.33643520e-01
-1.31538665e+00 -6.91434145e-01 5.88363409e-02 3.36220056e-01
-2.24124357e-01 -3.24039012e-01 -7.61556268e-01 -1.99334621e-01
6.62798584e-01 -9.77100909e-01 -3.76581043e-01 2.45194033e-01
-1.14781354e-02 -1.55916139e-01 8.27951968e-01 -3.12846035e-01
9.06071126e-01 -1.89559472e+00 2.37378165e-01 -1.69322938e-01
3.18993300e-01 3.37401450e-01 2.85616517e-01 -1.73015326e-01
1.87707856e-01 -2.58196414e-01 -2.06059471e-01 -6.99346066e-01
-4.43116724e-01 2.65297592e-01 2.21693471e-01 6.59925222e-01
-3.03868324e-01 5.85581005e-01 -6.87584341e-01 -4.28233445e-01
5.31242430e-01 7.12694824e-01 -5.44251263e-01 1.99979752e-01
3.66754830e-03 2.97230035e-01 8.66217613e-02 4.71189648e-01
9.59184766e-01 -1.58836201e-01 -4.20560464e-02 -2.72769898e-01
-3.66603464e-01 6.90355822e-02 -1.44220722e+00 1.99478424e+00
-2.74647295e-01 9.23313498e-01 -1.36245146e-01 -8.07989895e-01
8.06091666e-01 9.17749777e-02 4.52321649e-01 -7.84308851e-01
3.78289789e-01 3.90281528e-01 -1.27411172e-01 -1.95163429e-01
7.38057613e-01 1.01943530e-01 4.11556721e-01 2.70829171e-01
-2.36422956e-01 -3.02200407e-01 4.18978691e-01 -2.74498552e-01
9.95857954e-01 1.75779209e-01 3.49091977e-01 -5.53386584e-02
6.07154131e-01 -8.08766559e-02 5.39456248e-01 5.23897946e-01
-1.71063930e-01 8.37396622e-01 2.95850605e-01 -7.85634995e-01
-1.11489689e+00 -5.60239315e-01 -2.00271115e-01 4.80491132e-01
5.25994480e-01 -1.94528863e-01 -8.59530747e-01 -3.44465137e-01
-2.15166986e-01 8.44173059e-02 -6.18130505e-01 3.59581649e-01
-7.66607046e-01 -6.16960406e-01 4.23979133e-01 8.50993752e-01
1.10506546e+00 -1.01638258e+00 -1.43314695e+00 2.16621369e-01
-2.57913936e-02 -1.51432288e+00 4.54181619e-02 1.82534218e-01
-1.40853393e+00 -1.13146651e+00 -1.18321514e+00 -8.13988864e-01
7.07640052e-01 4.47634846e-01 1.16604197e+00 1.72505043e-02
-3.09102923e-01 -1.64692685e-01 -1.77384511e-01 -3.45902413e-01
1.89143971e-01 3.05376470e-01 -2.61134893e-01 -2.83412158e-01
3.53401214e-01 -5.15822768e-01 -9.57181990e-01 2.51193881e-01
-1.09055674e+00 3.01814884e-01 4.60545480e-01 6.75014496e-01
8.72528791e-01 1.87905326e-01 -2.18376324e-01 -7.98713624e-01
7.12135211e-02 3.13020974e-01 -1.09683967e+00 -3.15629870e-01
-7.89527833e-01 1.23820886e-01 3.75567108e-01 -1.86530039e-01
-8.91171455e-01 3.77081394e-01 -2.11958125e-01 -3.19595009e-01
-3.58288884e-01 1.73059180e-01 2.12823763e-01 -4.59657252e-01
4.33325559e-01 2.70842940e-01 -4.44384962e-02 -3.23400825e-01
1.98221616e-02 4.67569858e-01 6.14293993e-01 -7.00195655e-02
4.68616635e-01 8.69746923e-01 5.37350655e-01 -9.31864023e-01
-8.65102112e-01 -6.63873553e-01 -8.32285941e-01 -1.56050310e-01
1.09156621e+00 -9.92298663e-01 -8.95541489e-01 9.99694288e-01
-1.28583872e+00 -5.26276350e-01 3.59594040e-02 4.60712701e-01
-5.85138977e-01 3.88742298e-01 -7.37752199e-01 -6.57659709e-01
-4.32862520e-01 -1.34384024e+00 1.05977547e+00 5.48913002e-01
2.60397822e-01 -9.12336528e-01 2.69379076e-02 2.66417950e-01
3.65382910e-01 2.12041005e-01 3.87203455e-01 2.31326476e-01
-9.20237601e-01 -8.05082470e-02 -6.25373542e-01 4.01966631e-01
8.58991034e-03 -2.74737239e-01 -1.07934594e+00 -2.69029617e-01
1.67822659e-01 -2.21142247e-01 9.79510963e-01 5.81412375e-01
1.31762636e+00 3.18995804e-01 -4.00241315e-02 8.97746325e-01
1.93845272e+00 4.46933448e-01 1.03935754e+00 5.89311540e-01
8.38844538e-01 4.40547407e-01 7.24562228e-01 3.95537406e-01
2.59743512e-01 6.47382915e-01 7.75333762e-01 -2.13154376e-01
-1.61015809e-01 1.96947455e-01 -4.22208123e-02 5.41702151e-01
-3.68976206e-01 -1.30856618e-01 -9.07840908e-01 4.64809209e-01
-1.95756447e+00 -7.42184639e-01 -3.91185284e-01 2.17821884e+00
5.25530279e-01 3.20892990e-01 -9.74835157e-02 6.61314726e-01
3.29137951e-01 1.87111869e-01 -4.20614630e-01 -2.17819855e-01
-1.83193371e-01 3.33822161e-01 9.47786093e-01 6.29185915e-01
-1.01494980e+00 9.26945329e-01 6.23058319e+00 5.83782375e-01
-1.46054435e+00 -2.42519930e-01 8.06031406e-01 -3.61450344e-01
5.47998071e-01 -3.39417338e-01 -1.08362150e+00 1.43607304e-01
6.25142574e-01 2.51873761e-01 3.13628435e-01 9.52475965e-01
2.63343845e-02 -8.79717469e-01 -1.13920081e+00 1.67340732e+00
3.76081884e-01 -1.41189468e+00 -2.19394132e-01 1.78942174e-01
8.74456584e-01 2.83368647e-01 -1.35247946e-01 -2.55638272e-01
-2.98377216e-01 -1.12847197e+00 5.25249720e-01 2.77991325e-01
7.36616075e-01 -1.08976531e+00 1.16137016e+00 4.01215255e-01
-1.12123299e+00 6.02296889e-02 -7.44315863e-01 -5.21560192e-01
2.29145482e-01 6.66993022e-01 -4.14984971e-01 3.05994987e-01
1.07007718e+00 7.15969086e-01 -5.62898040e-01 1.17126560e+00
-2.80135870e-01 8.72556567e-02 -3.70615810e-01 -1.41279921e-01
3.44338387e-01 -1.91088706e-01 -1.76095217e-02 8.74198675e-01
3.68904561e-01 1.45564988e-01 -1.19862854e-01 2.39421383e-01
-1.26549140e-01 -5.40558659e-02 -4.37991887e-01 4.55653101e-01
8.44234135e-03 1.19195747e+00 -1.00187469e+00 -4.09519404e-01
-4.56551194e-01 1.17278075e+00 3.34838092e-01 -2.15056483e-02
-6.30468547e-01 -4.37186718e-01 5.97150564e-01 1.46890506e-01
5.06866932e-01 -4.67981786e-01 -5.77220380e-01 -7.72649050e-01
5.05257510e-02 -5.72863162e-01 6.11307323e-02 -7.79256761e-01
-6.80161119e-01 6.14402711e-01 -2.65776515e-01 -1.00509846e+00
-2.55446613e-01 -1.06529534e+00 -2.98323259e-02 6.95824981e-01
-1.86767137e+00 -5.14964819e-01 -1.14403641e+00 6.51643038e-01
8.48759711e-01 1.49680212e-01 6.53421879e-01 5.51572442e-01
-2.72223562e-01 3.50898892e-01 -6.29405975e-02 1.29966304e-01
6.25581026e-01 -1.19208491e+00 4.90363926e-01 8.40098977e-01
4.36896048e-02 1.06936067e-01 6.91468418e-01 -3.82923782e-01
-1.14432621e+00 -7.60442793e-01 8.46913934e-01 -1.30791798e-01
1.60771638e-01 -3.96453328e-02 -5.90703428e-01 3.24017525e-01
1.06488913e-01 3.05236250e-01 1.16528921e-01 -2.34507516e-01
-3.11111342e-02 -1.49552301e-01 -1.22238553e+00 5.15683532e-01
1.11700583e+00 -3.32073689e-01 -7.19311535e-02 -5.10424115e-02
4.82224375e-01 -9.28088665e-01 -6.34544432e-01 4.33148921e-01
6.31061554e-01 -1.63387954e+00 8.34605992e-01 2.91099131e-01
6.67778730e-01 -5.07590532e-01 -2.44167671e-01 -8.85182977e-01
5.60459942e-02 -2.17620105e-01 -2.10715503e-01 5.62399745e-01
8.29924121e-02 -5.08491218e-01 1.45116627e+00 4.55529541e-01
1.02467611e-01 -1.04692292e+00 -1.03604019e+00 -5.03269196e-01
-1.89301208e-01 -5.54076135e-01 2.89198518e-01 5.09292245e-01
-4.37638730e-01 3.77107978e-01 -2.59262592e-01 3.37453783e-02
8.24531853e-01 -6.57992288e-02 9.93695796e-01 -1.36591053e+00
-1.58821344e-01 -4.30272222e-01 -8.63772869e-01 -1.47631133e+00
-3.30274075e-01 -1.44069538e-01 3.02862339e-02 -1.72257936e+00
-3.53317684e-03 -2.72764117e-01 1.07020520e-01 1.31901786e-01
-8.38974044e-02 8.50738525e-01 2.30227992e-01 -2.26335544e-02
-4.61211413e-01 2.61612356e-01 1.22761524e+00 3.77825238e-02
-1.68470517e-01 -3.30126919e-02 -1.65133119e-01 1.03649879e+00
8.11865866e-01 -4.23508167e-01 -4.79469538e-01 -7.17372656e-01
1.94147259e-01 4.21089418e-02 2.41322696e-01 -1.53990269e+00
4.82380956e-01 1.67948037e-01 6.06509089e-01 -1.06770408e+00
6.53495014e-01 -8.67274821e-01 -3.95364575e-02 6.70903981e-01
1.67441070e-01 3.30788344e-01 1.63103133e-01 2.87467092e-01
-4.23039049e-01 -4.00650769e-01 1.10648060e+00 -2.86217749e-01
-1.01141179e+00 5.26264608e-01 4.62376280e-03 -1.18683159e-01
1.10458398e+00 -6.02338970e-01 -2.18002692e-01 -2.46642470e-01
-2.73855150e-01 -2.51525313e-01 6.93412721e-01 1.37739778e-01
8.51067483e-01 -9.68768120e-01 -4.35046494e-01 1.86214864e-01
-1.18381716e-01 7.72644162e-01 2.45759070e-01 6.82179928e-01
-1.25300980e+00 4.89468098e-01 -4.22245651e-01 -9.52563643e-01
-1.63450813e+00 1.99710265e-01 4.22542453e-01 -2.28774428e-01
-8.34090650e-01 1.05941498e+00 1.43127605e-01 8.96626636e-02
5.41904986e-01 -4.63441610e-01 -2.64274865e-01 -4.28825431e-02
6.86918855e-01 7.21475959e-01 2.32633650e-01 -4.12916601e-01
-1.24588326e-01 1.03681242e+00 -1.13697452e-02 -2.25141719e-01
1.31955802e+00 -3.55121717e-02 -1.23428561e-01 4.35102344e-01
1.23263896e+00 -3.85077596e-01 -1.59770155e+00 1.68024357e-02
-2.41079986e-01 -7.25551903e-01 4.97228175e-01 -2.57030219e-01
-1.39362335e+00 1.07371700e+00 1.02414358e+00 3.10992030e-03
1.24226236e+00 -3.59366953e-01 9.79250610e-01 5.42515337e-01
6.63694739e-01 -1.09927511e+00 2.11220626e-02 6.06076300e-01
4.20923948e-01 -1.61788678e+00 2.95584440e-01 -4.84321654e-01
-1.52594194e-01 1.18575704e+00 6.25558496e-01 -2.74298698e-01
6.71626568e-01 4.04004931e-01 3.07053894e-01 -5.64895757e-02
-2.02776358e-01 -2.17115432e-01 6.19472750e-02 4.42846864e-01
4.49500024e-01 -2.56871253e-01 -3.69931638e-01 -3.14761639e-01
-2.62065023e-01 2.38855220e-02 5.03728867e-01 1.02490377e+00
-5.29429018e-01 -1.16351116e+00 -2.48933554e-01 1.28179818e-01
-6.75252259e-01 -4.60550152e-02 -1.11541459e-02 8.49210680e-01
7.94477388e-02 8.71710658e-01 2.57643640e-01 -8.50508809e-02
2.76764274e-01 -3.34458411e-01 8.06550205e-01 -2.87345529e-01
-4.38522071e-01 -1.55548647e-01 -1.75030425e-01 -9.27185714e-01
-8.65428925e-01 -4.57807958e-01 -1.39444482e+00 -4.90830868e-01
-9.58243832e-02 -2.95399278e-01 1.06106639e+00 8.65772486e-01
3.96699011e-02 5.85674644e-01 2.50258505e-01 -1.23174107e+00
-6.92853555e-02 -8.27309191e-01 -5.17525256e-01 2.74734676e-01
2.44196162e-01 -5.20326197e-01 -2.39912271e-01 2.49704327e-02] | [8.863543510437012, -2.2871763706207275] |
51ce2bf7-4535-49de-a581-5ad5dd29b811 | active-learning-in-brain-tumor-segmentation | 2302.10185 | null | https://arxiv.org/abs/2302.10185v1 | https://arxiv.org/pdf/2302.10185v1.pdf | Active Learning in Brain Tumor Segmentation with Uncertainty Sampling, Annotation Redundancy Restriction, and Data Initialization | Deep learning models have demonstrated great potential in medical 3D imaging, but their development is limited by the expensive, large volume of annotated data required. Active learning (AL) addresses this by training a model on a subset of the most informative data samples without compromising performance. We compared different AL strategies and propose a framework that minimizes the amount of data needed for state-of-the-art performance. 638 multi-institutional brain tumor MRI images were used to train a 3D U-net model and compare AL strategies. We investigated uncertainty sampling, annotation redundancy restriction, and initial dataset selection techniques. Uncertainty estimation techniques including Bayesian estimation with dropout, bootstrapping, and margins sampling were compared to random query. Strategies to avoid annotation redundancy by removing similar images within the to-be-annotated subset were considered as well. We determined the minimum amount of data necessary to achieve similar performance to the model trained on the full dataset ({\alpha} = 0.1). A variance-based selection strategy using radiomics to identify the initial training dataset is also proposed. Bayesian approximation with dropout at training and testing showed similar results to that of the full data model with less than 20% of the training data (p=0.293) compared to random query achieving similar performance at 56.5% of the training data (p=0.814). Annotation redundancy restriction techniques achieved state-of-the-art performance at approximately 40%-50% of the training data. Radiomics dataset initialization had higher Dice with initial dataset sizes of 20 and 80 images, but improvements were not significant. In conclusion, we investigated various AL strategies with dropout uncertainty estimation achieving state-of-the-art performance with the least annotated data. | ['Zhicheng Jiao', 'Harrison X Bai', 'Ali Nabavizadeh', 'Anahita Fathi Kazerooni', 'Li Yang', 'Beiji Zou', 'Chengzhang Zhu', 'Wei-Hua Liao', 'Haris Sair', 'Craig Jones', 'Chetan Bettegowda', 'Michael Atalay', 'Xue Feng', 'Jing Wu', 'Jian Peng', 'Rajat S Chandra', 'Daniel D Kim'] | 2023-02-05 | null | null | null | null | ['tumor-segmentation', 'brain-tumor-segmentation'] | ['computer-vision', 'medical'] | [-7.80839182e-04 6.37090802e-01 -3.93007487e-01 -6.86262250e-01
-1.55598938e+00 -2.35276580e-01 2.80967206e-01 4.98088986e-01
-1.13361096e+00 1.06456542e+00 2.52991647e-01 -2.66130120e-01
-5.34245431e-01 -4.04601455e-01 -6.97008252e-01 -9.73261714e-01
-3.17020148e-01 8.64368141e-01 4.64662910e-01 6.36990607e-01
8.69045928e-02 5.91269672e-01 -9.80603516e-01 3.52427036e-01
9.08911467e-01 8.66614819e-01 3.82205874e-01 3.54465157e-01
7.15876594e-02 8.53238821e-01 -7.38906324e-01 3.35984305e-02
1.12560824e-01 -2.97676623e-01 -8.46767485e-01 4.99018095e-02
4.52954859e-01 -7.92400420e-01 -1.86040118e-01 7.12817073e-01
1.24903488e+00 2.23681014e-02 8.00063074e-01 -9.79848683e-01
-1.46935046e-01 8.55102897e-01 -3.86713475e-01 4.67786074e-01
-1.45906374e-01 3.50900590e-01 2.27802962e-01 -7.72247553e-01
7.40557134e-01 5.95669568e-01 7.54159689e-01 7.07865000e-01
-1.69839871e+00 -7.06661463e-01 -3.11811626e-01 9.78007987e-02
-1.44908988e+00 -4.75818127e-01 4.12673593e-01 -6.67461097e-01
1.01162314e+00 1.16751045e-01 6.52956784e-01 9.59314585e-01
5.50754964e-01 7.43050694e-01 1.11997342e+00 -4.42372024e-01
6.78414762e-01 1.97243765e-01 4.82102215e-01 6.19495988e-01
4.92377222e-01 1.14343934e-01 -2.97648460e-01 -5.50451577e-01
7.56271601e-01 -4.89741832e-01 -4.49611008e-01 -3.46647859e-01
-8.96322668e-01 9.01104629e-01 3.22191447e-01 2.03547612e-01
-4.81861115e-01 1.06826589e-01 4.90210116e-01 5.31898858e-03
6.08691633e-01 4.17175025e-01 -3.18530649e-01 1.38168875e-02
-1.43669415e+00 -3.09278041e-01 4.68077302e-01 7.58481562e-01
2.46004581e-01 8.23482499e-03 -2.99006373e-01 8.76286507e-01
2.31277421e-01 1.60981212e-02 6.61819935e-01 -1.03896844e+00
8.56008083e-02 3.72955590e-01 -1.99104205e-01 -1.75880298e-01
-8.88518751e-01 -6.54601336e-01 -6.70488238e-01 5.73721230e-01
6.10086858e-01 -5.62492967e-01 -1.32471132e+00 1.60547221e+00
2.63396315e-02 -9.42341983e-02 -2.66261488e-01 5.82143247e-01
7.80338466e-01 5.13617732e-02 8.74774717e-03 -6.45997941e-01
8.97233844e-01 -3.88348311e-01 -6.75609589e-01 8.66720546e-03
8.83668065e-01 -4.13790882e-01 8.28373849e-01 4.74549979e-01
-1.16496348e+00 -2.20196366e-01 -9.89116609e-01 3.72790962e-01
1.80545390e-01 3.45416032e-02 4.27455753e-01 1.17464316e+00
-1.30271971e+00 7.04621494e-01 -1.42736781e+00 1.94351852e-01
1.06396437e+00 8.40331137e-01 -4.15646404e-01 -1.15193054e-01
-9.42931235e-01 1.16004372e+00 5.74431539e-01 5.07149771e-02
-1.19926357e+00 -1.21299791e+00 -8.58623028e-01 -3.21783215e-01
2.60676444e-01 -5.26335418e-01 1.02240038e+00 -7.26067960e-01
-1.10683429e+00 7.28202999e-01 1.66544095e-01 -7.58562326e-01
9.99864340e-01 -4.34250869e-02 1.32468477e-01 2.96486974e-01
3.80116962e-02 9.77170587e-01 4.78676528e-01 -1.18624032e+00
-1.39528796e-01 -4.63925302e-01 -4.16714281e-01 1.80118695e-01
3.72109637e-02 1.00333296e-01 -2.64342517e-01 -4.05718893e-01
3.39411288e-01 -9.48795497e-01 -4.68723923e-01 1.57263428e-01
-8.96447897e-02 1.71060160e-01 3.81156564e-01 -8.66362393e-01
8.66711199e-01 -1.92031240e+00 -2.47613549e-01 2.46351689e-01
4.21904653e-01 -1.33618280e-01 2.16459617e-01 -4.14047718e-01
-3.41121942e-01 2.13718757e-01 -5.26905656e-01 -3.89987797e-01
-4.43226248e-01 1.51359126e-01 4.78409290e-01 7.61030912e-01
1.45783782e-01 5.74299991e-01 -8.57652128e-01 -7.58191466e-01
3.11095595e-01 6.58266306e-01 -6.46562636e-01 6.56666756e-02
2.40091115e-01 7.23803937e-01 1.58862360e-02 4.69903409e-01
6.28854454e-01 -1.07597344e-01 1.26879722e-01 -2.32961148e-01
2.42463857e-01 -3.63950729e-02 -1.12293088e+00 1.67641473e+00
-3.62456381e-01 6.71333253e-01 2.42704675e-02 -8.57268393e-01
8.76959026e-01 5.62994301e-01 9.69493449e-01 -3.80079389e-01
2.51837403e-01 5.24339080e-01 6.87627435e-01 -3.90769809e-01
-1.37466952e-01 -2.38165751e-01 3.06410968e-01 3.52419466e-01
3.18469912e-01 -3.71581167e-01 1.76152736e-02 1.49403960e-01
1.19136298e+00 6.88078329e-02 7.84914047e-02 -6.16801322e-01
-1.47498371e-02 -4.59202752e-02 5.99220753e-01 1.17856205e+00
-6.83129013e-01 8.78001511e-01 7.59665191e-01 -1.23791531e-01
-8.94057930e-01 -1.01573765e+00 -9.75023031e-01 1.79397807e-01
-4.52030182e-01 -8.83083716e-02 -9.05458272e-01 -9.20363724e-01
-3.45585048e-01 1.10241318e+00 -8.54758680e-01 -2.96715975e-01
-3.50485802e-01 -1.53304136e+00 5.86905301e-01 4.71359789e-01
3.77142668e-01 -8.87539506e-01 -9.22444165e-01 2.72029787e-01
-4.65452634e-02 -7.57826567e-01 -7.58640328e-03 7.88933277e-01
-1.48484695e+00 -9.99396265e-01 -1.03595984e+00 -3.11664641e-01
9.92350101e-01 -7.75181532e-01 1.03816807e+00 -3.42088908e-01
-3.79613727e-01 3.69180232e-01 -1.60797879e-01 -3.80325347e-01
-4.94592041e-01 -1.63197666e-01 5.65579198e-02 -6.82648480e-01
1.49224713e-01 -2.19031304e-01 -5.57690322e-01 1.28766462e-01
-1.00323427e+00 -1.52297154e-01 6.24163032e-01 1.26749516e+00
8.39078188e-01 -3.12410127e-02 5.47625184e-01 -8.72567654e-01
3.73337179e-01 -3.15927327e-01 -4.50474262e-01 2.27725357e-01
-7.48151720e-01 1.67250782e-01 1.29529297e-01 -6.60804570e-01
-9.84039545e-01 1.61087975e-01 -7.17296079e-02 -4.51364756e-01
-7.86789209e-02 4.58962470e-01 7.96186253e-02 -1.26176581e-01
9.47123110e-01 -2.59740233e-01 3.71635765e-01 -1.77823186e-01
-1.77363336e-01 4.83545631e-01 7.39939213e-02 -6.39292181e-01
-1.55765131e-01 3.53757888e-01 -2.44996604e-02 -6.29857183e-01
-5.94713628e-01 -2.75890976e-02 -1.03408706e+00 -3.78213227e-01
5.79705834e-01 -5.49251735e-01 -6.68788627e-02 4.93839800e-01
-7.60502219e-01 -6.60618126e-01 -8.12327504e-01 1.33161569e+00
-7.44634271e-01 1.03864163e-01 -4.57783103e-01 -6.71305656e-01
-6.69249415e-01 -2.00193930e+00 6.16899908e-01 -8.73535499e-03
-4.26313579e-01 -9.19815540e-01 1.06744662e-01 3.50048035e-01
3.23563308e-01 3.59725535e-01 9.07370567e-01 -1.07422209e+00
-3.47085863e-01 -2.35331714e-01 4.28971574e-02 4.06308353e-01
1.99241966e-01 -3.42182785e-01 -9.25667703e-01 -4.82281983e-01
3.57136250e-01 -4.43032205e-01 8.42020154e-01 1.25440359e+00
1.29502535e+00 3.17092746e-01 -2.28209540e-01 4.09842670e-01
1.35403049e+00 6.47176683e-01 7.83185303e-01 2.69949287e-01
1.73689932e-01 4.22258109e-01 2.39652216e-01 2.94642866e-01
-3.44444960e-01 4.16800469e-01 2.06583470e-01 -1.19340859e-01
-4.35961962e-01 3.57946724e-01 9.83403102e-02 4.52067286e-01
2.66451716e-01 2.49275997e-01 -1.39128447e+00 5.50618947e-01
-1.41538882e+00 -6.64639473e-01 -1.03169858e-01 2.65538096e+00
9.30088401e-01 7.29635537e-01 -1.39355153e-01 7.86313713e-02
6.81669056e-01 -2.72440284e-01 -7.61100650e-01 -8.07843450e-03
2.10191850e-02 2.80702680e-01 8.01747382e-01 6.63591802e-01
-8.22566211e-01 2.59990245e-01 6.59297228e+00 9.43893731e-01
-1.00331497e+00 4.46981758e-01 1.28923965e+00 -4.55229938e-01
2.91703437e-02 -1.34418374e-02 -7.38000512e-01 5.26791751e-01
1.12114084e+00 1.53882578e-01 -1.98142201e-01 5.97386658e-01
4.82566297e-01 -9.00705278e-01 -1.05925298e+00 7.97039330e-01
3.17913145e-01 -1.32975709e+00 -3.24011981e-01 7.32778385e-02
7.54825592e-01 3.11558276e-01 -2.97771305e-01 2.41333693e-02
1.00941867e-01 -9.79547858e-01 5.04092813e-01 7.42882550e-01
9.92515802e-01 -6.28308237e-01 1.17943096e+00 1.69363633e-01
-3.98031384e-01 2.43693098e-01 -2.50017077e-01 4.07878250e-01
1.76313803e-01 9.07252848e-01 -1.21763170e+00 5.53394139e-01
7.02146232e-01 5.12762032e-02 -7.37751245e-01 1.54846466e+00
5.93805350e-02 8.24748158e-01 -7.10843623e-01 1.98573932e-01
-1.23291938e-02 1.64017960e-01 4.75847036e-01 8.57450724e-01
3.23838480e-02 6.58973530e-02 -1.34079922e-02 7.78755724e-01
1.45418867e-01 -1.54717220e-03 -2.07059503e-01 5.02078891e-01
4.42143887e-01 7.57108927e-01 -1.14123988e+00 -3.02549005e-01
-2.80503631e-01 4.13572788e-01 1.93451926e-01 7.40896612e-02
-6.70965374e-01 -2.20108733e-01 -2.74828672e-01 2.73754925e-01
-2.90169775e-01 1.09327547e-01 -4.96071637e-01 -6.98044419e-01
-2.61853278e-01 -5.26451766e-01 6.90935850e-01 -7.22105742e-01
-1.15107822e+00 6.15035355e-01 5.85336983e-01 -9.36401367e-01
-2.13663891e-01 -5.25121331e-01 -3.52336794e-01 1.11667526e+00
-8.14771175e-01 -5.82537889e-01 -1.03490412e-01 2.74872661e-01
4.08621073e-01 -1.91342607e-01 8.40809464e-01 2.27436468e-01
-4.88132149e-01 5.91788888e-01 5.26355267e-01 3.39612752e-01
7.20647633e-01 -1.18271613e+00 -3.49877119e-01 6.26915634e-01
-3.09053153e-01 5.66367805e-01 4.29392248e-01 -1.01228762e+00
-6.82702065e-01 -6.71413243e-01 3.02457005e-01 -3.79205942e-01
3.98425400e-01 8.21563452e-02 -8.16475272e-01 7.09395945e-01
-8.23408067e-02 1.99468583e-01 8.57232928e-01 -1.46073356e-01
3.71791661e-01 9.16090980e-02 -1.72754180e+00 2.81146348e-01
4.60070908e-01 -1.77705437e-01 -8.19571018e-01 1.43462837e-01
3.76854092e-01 -6.69837773e-01 -1.19848847e+00 8.24568629e-01
4.79319423e-01 -8.15778852e-01 6.77995026e-01 -2.53347516e-01
-2.45453026e-02 1.43165261e-01 4.93460000e-02 -1.30092764e+00
-5.87448664e-02 -1.72064796e-01 2.52440851e-03 8.55245888e-01
7.37300515e-01 -4.83026832e-01 1.25248992e+00 1.11474049e+00
-4.19946134e-01 -1.10860002e+00 -1.37608147e+00 -6.54320955e-01
5.03039420e-01 -6.59780800e-01 -7.17679188e-02 8.05780411e-01
-1.66049048e-01 -8.29227120e-02 1.73587188e-01 -1.63832709e-01
1.03743267e+00 -7.10220039e-01 -6.53120801e-02 -1.10349250e+00
-1.82923272e-01 -3.40259463e-01 -4.73591238e-01 -6.38261437e-02
1.36424959e-01 -9.57398415e-01 -1.58184171e-01 -1.90115297e+00
5.55257142e-01 -6.70189083e-01 -3.80176604e-01 4.96084183e-01
-3.54459626e-03 8.47461224e-02 -7.05894828e-02 3.12526345e-01
1.30962968e-01 2.58988380e-01 1.19184721e+00 -1.91954434e-01
-4.27086711e-01 -4.92962217e-03 -2.51624674e-01 5.62299132e-01
7.98358083e-01 -7.49426365e-01 -7.40679443e-01 -2.77363867e-01
-2.41765410e-01 3.77153605e-01 3.08788866e-01 -1.19973290e+00
3.14772099e-01 3.71728003e-01 1.03648424e+00 -8.70492876e-01
3.52669448e-01 -9.04879212e-01 3.99892360e-01 9.31815207e-01
-4.78684783e-01 -1.92486674e-01 6.26910508e-01 7.35436901e-02
1.22616716e-01 -9.56917524e-01 1.13943756e+00 -2.62071997e-01
-2.63850838e-01 2.45746732e-01 -4.56041634e-01 2.40061596e-01
1.03231955e+00 -3.02617162e-01 -1.59810454e-01 -1.14980713e-01
-1.47400737e+00 8.37270394e-02 1.68008149e-01 -1.37394220e-01
5.65998018e-01 -9.32121575e-01 -8.75587225e-01 -7.76088983e-02
-2.40950361e-01 3.45704049e-01 4.52312499e-01 1.30859709e+00
-7.24738419e-01 3.00666481e-01 -3.22574198e-01 -9.24878240e-01
-1.15255320e+00 5.74946590e-02 9.46467936e-01 -3.01573932e-01
-6.45667553e-01 8.83703709e-01 -2.91651487e-01 -4.74685013e-01
4.46215600e-01 -1.68982252e-01 -1.06802690e-05 3.45342815e-01
4.05609488e-01 4.86785442e-01 6.57099605e-01 -1.49205044e-01
-5.37965357e-01 1.02935202e-01 -3.73881161e-01 -5.61467946e-01
1.37107825e+00 1.98209807e-01 1.57359630e-01 5.45538068e-01
1.27705109e+00 -3.26241493e-01 -1.44714344e+00 -2.87742894e-02
-8.32261168e-04 -3.14812481e-01 8.28683853e-01 -1.13940656e+00
-9.77171004e-01 7.08354056e-01 1.37812400e+00 -4.38491851e-01
8.44133794e-01 -5.53704128e-02 -1.05765730e-01 1.77818447e-01
5.35059154e-01 -1.13087273e+00 -1.57365397e-01 8.16611424e-02
7.28083313e-01 -1.33623707e+00 5.71898222e-01 1.20747328e-01
-7.16134369e-01 1.06949902e+00 5.79409719e-01 4.28860448e-02
8.67150307e-01 6.60516739e-01 4.17503354e-04 -1.97293684e-01
-3.98184747e-01 1.82422176e-01 2.02908158e-01 7.59038925e-01
6.16685748e-01 -1.00401677e-01 -5.22010207e-01 4.86277223e-01
2.31333300e-01 3.69502306e-01 6.37914538e-01 1.03789818e+00
-2.60836095e-01 -8.18776608e-01 -5.06363690e-01 1.05897629e+00
-5.99027514e-01 -1.94680080e-01 1.86779555e-02 1.04108381e+00
-2.67121401e-02 5.99070072e-01 2.72146165e-01 1.37815341e-01
5.69601916e-02 3.59723300e-01 7.72631109e-01 -5.15156388e-01
-6.23035669e-01 3.57231587e-01 2.42138073e-01 -3.65007401e-01
-3.56781840e-01 -9.29070294e-01 -1.27570844e+00 4.13623266e-02
-7.05004454e-01 1.92487195e-01 5.73014736e-01 7.80180693e-01
6.30461127e-02 7.03409672e-01 2.39958137e-01 -6.83965206e-01
-6.37194872e-01 -1.16519821e+00 -4.97093529e-01 -9.22045037e-02
5.78707196e-02 -8.59035850e-01 -5.61111689e-01 -2.32819512e-01] | [14.260316848754883, -2.362783193588257] |
698d4041-0154-48e7-b037-973eb3e431a6 | light-weight-head-pose-invariant-gaze | 1804.08572 | null | http://arxiv.org/abs/1804.08572v1 | http://arxiv.org/pdf/1804.08572v1.pdf | Light-weight Head Pose Invariant Gaze Tracking | Unconstrained remote gaze tracking using off-the-shelf cameras is a
challenging problem. Recently, promising algorithms for appearance-based gaze
estimation using convolutional neural networks (CNN) have been proposed.
Improving their robustness to various confounding factors including variable
head pose, subject identity, illumination and image quality remain open
problems. In this work, we study the effect of variable head pose on machine
learning regressors trained to estimate gaze direction. We propose a novel
branched CNN architecture that improves the robustness of gaze classifiers to
variable head pose, without increasing computational cost. We also present
various procedures to effectively train our gaze network including transfer
learning from the more closely related task of object viewpoint estimation and
from a large high-fidelity synthetic gaze dataset, which enable our ten times
faster gaze network to achieve competitive accuracy to its current
state-of-the-art direct competitor. | ['Rajeev Ranjan', 'Jan Kautz', 'Shalini De Mello'] | 2018-04-23 | null | null | null | null | ['viewpoint-estimation'] | ['computer-vision'] | [ 5.25332056e-02 1.28603941e-02 2.99795736e-02 -6.04731262e-01
-4.06989902e-01 -2.28640497e-01 1.27762482e-01 -7.05062509e-01
-5.94076693e-01 5.55470943e-01 -9.27909389e-02 -1.94842011e-01
3.95874470e-01 1.00979351e-01 -9.85712707e-01 -6.54798150e-01
2.42111072e-01 -5.75376078e-02 4.20326591e-02 5.14521115e-02
3.46743107e-01 1.76822171e-01 -2.07103777e+00 -4.00065392e-01
7.36510038e-01 1.17328918e+00 -3.61236669e-02 7.08630562e-01
5.49398720e-01 7.56477773e-01 -5.74215472e-01 -3.14592212e-01
-1.80509947e-02 -3.68442237e-01 -6.09914184e-01 4.53786179e-02
1.53843534e+00 -5.25218427e-01 6.77355900e-02 9.30189073e-01
7.10317373e-01 1.41237220e-02 6.85926616e-01 -1.62559485e+00
-8.93559575e-01 -1.59496278e-01 -9.03945267e-01 2.91118920e-01
4.16446388e-01 3.23190540e-01 6.96064711e-01 -9.47727561e-01
4.22447890e-01 1.03742433e+00 7.53083169e-01 9.92050052e-01
-1.12402999e+00 -1.26813376e+00 2.45721400e-01 5.11894345e-01
-1.34617972e+00 -9.89338160e-01 6.64287448e-01 -3.32635552e-01
6.90253139e-01 1.27011761e-01 4.43747610e-01 1.52792990e+00
6.66092560e-02 7.80290306e-01 1.25275886e+00 -5.30974388e-01
-1.28285185e-01 8.82624090e-02 1.89661589e-02 9.16923106e-01
2.22205937e-01 1.62071567e-02 -9.69914198e-01 1.74101740e-01
4.99466360e-01 -2.91725576e-01 -7.64094949e-01 -5.99127591e-01
-8.36771250e-01 7.18165278e-01 8.98255825e-01 -2.44826555e-01
-4.99262288e-02 2.68049747e-01 -1.11569524e-01 3.69586535e-02
8.12436640e-01 5.08946359e-01 -6.65370047e-01 2.76468433e-02
-9.29605424e-01 2.34029919e-01 5.80980718e-01 1.13246644e+00
5.01128554e-01 4.74555120e-02 6.79878742e-02 5.28023779e-01
8.41395915e-01 8.39917362e-01 1.93862915e-01 -8.95458221e-01
3.01801503e-01 2.40347952e-01 1.54459670e-01 -7.27846742e-01
-8.32347155e-01 -2.11474493e-01 -3.28946501e-01 6.00288332e-01
9.41410244e-01 -3.23875338e-01 -9.01462018e-01 1.96468842e+00
4.71693218e-01 2.73399293e-01 -4.42078084e-01 1.15776014e+00
1.08788466e+00 -4.93172742e-02 1.86066478e-01 -3.03340554e-01
1.53285980e+00 -9.92502689e-01 -7.09070981e-01 -4.30654615e-01
4.53906357e-01 -8.07431817e-01 1.09922159e+00 4.20594513e-01
-1.17615759e+00 -3.06093484e-01 -1.00845313e+00 -4.22749400e-01
-5.73076382e-02 2.41187826e-01 5.82506299e-01 9.28597510e-01
-1.52290154e+00 8.17653984e-02 -7.71501362e-01 -5.71672559e-01
7.25012243e-01 8.93992305e-01 -3.91129225e-01 2.20951572e-01
-6.96480632e-01 1.14109147e+00 -4.38281089e-01 2.95609415e-01
-6.11764669e-01 -8.06745529e-01 -1.04216504e+00 6.59300154e-03
4.42012995e-01 -8.86616111e-01 1.62242782e+00 -1.40544629e+00
-1.73588431e+00 1.15805793e+00 -8.21386933e-01 3.31241861e-02
2.71187663e-01 -4.44381446e-01 -2.01571271e-01 -9.28256437e-02
-2.56932408e-01 1.04322970e+00 1.46341133e+00 -9.64461207e-01
-4.03142929e-01 -7.69888997e-01 -7.37406537e-02 1.75193250e-01
-3.22514623e-01 4.89439189e-01 -2.98117846e-01 -1.52681202e-01
-1.10009208e-01 -1.39406931e+00 4.36919808e-01 5.03137469e-01
-2.67709255e-01 -3.84779483e-01 7.11014807e-01 -4.61632639e-01
7.44415998e-01 -1.84541202e+00 2.46884316e-01 -2.80390620e-01
6.87597990e-01 7.26208985e-02 4.44583222e-02 -5.53101420e-01
-3.00500542e-01 -1.85933620e-01 1.94044977e-01 -7.28703678e-01
-2.49649614e-01 -4.71063256e-01 1.98318228e-01 8.08924377e-01
3.38787049e-01 1.13879287e+00 -5.67699790e-01 -5.40573120e-01
-1.54710282e-02 7.29866147e-01 -5.60124695e-01 3.50344539e-01
-2.51942140e-04 6.57311559e-01 1.14866281e-02 7.93057859e-01
6.44793749e-01 -5.74876726e-01 -9.52171385e-02 -2.66227901e-01
3.40546817e-02 1.62338749e-01 -5.23242056e-01 1.54224539e+00
-4.45328206e-01 1.52109504e+00 1.48288891e-01 -1.03527240e-01
6.26440823e-01 1.67093739e-01 -4.87730652e-03 -7.14545608e-01
8.49668860e-01 -6.39031455e-02 1.56150475e-01 -7.43315279e-01
6.37014508e-01 -5.09076985e-03 4.35363859e-01 5.61280131e-01
3.09452474e-01 3.18175375e-01 -4.83615041e-01 -3.08654547e-01
5.49249649e-01 4.44210887e-01 6.22348115e-02 -3.07486326e-01
4.00537014e-01 -4.99504805e-01 3.25436741e-02 2.10178077e-01
-7.05567777e-01 9.19704795e-01 2.80413985e-01 -3.59799594e-01
-7.57409453e-01 -5.17411411e-01 -3.05195361e-01 1.67357492e+00
1.28600031e-01 1.56533197e-01 -1.15440071e+00 -7.34596014e-01
-1.62684828e-01 4.04386520e-01 -1.01924860e+00 -1.97589654e-03
-6.02803946e-01 -6.70402467e-01 2.93328166e-01 7.37303317e-01
1.66242793e-01 -1.04244590e+00 -6.95754886e-01 -4.52809989e-01
-1.22575499e-01 -1.02830493e+00 -7.14452326e-01 -1.77656770e-01
-4.54230487e-01 -1.29587770e+00 -1.12730002e+00 -6.89107239e-01
7.99518108e-01 4.29577798e-01 1.20494938e+00 1.61627859e-01
1.45272696e-02 2.64829814e-01 -2.00847033e-02 -9.45824265e-01
2.19731435e-01 4.98892695e-01 2.70030051e-01 1.50182039e-01
6.56586587e-01 -1.36217669e-01 -7.74041235e-01 4.30066884e-01
-2.55975813e-01 -1.97253913e-01 1.48503050e-01 6.21903062e-01
-4.00821194e-02 -1.05629301e+00 1.89147905e-01 -7.09344566e-01
5.08016050e-01 -2.54969686e-01 -8.72006953e-01 2.79615283e-01
-6.82292521e-01 1.46144748e-01 -1.42857596e-01 -5.50237954e-01
-1.10012436e+00 -3.35377529e-02 7.10040778e-02 -7.37530589e-01
-2.37807646e-01 -8.70029405e-02 1.94150768e-02 -6.98868752e-01
8.57141256e-01 -3.25749606e-01 2.60659397e-01 -3.16039711e-01
2.63478070e-01 7.21222579e-01 3.98384035e-01 2.25337758e-03
6.27855003e-01 4.39805210e-01 1.35215223e-01 -6.02954566e-01
-1.28198910e+00 -1.89390033e-01 -9.40841973e-01 -3.31892788e-01
1.08929288e+00 -1.14383340e+00 -1.47021341e+00 9.16276217e-01
-1.17628062e+00 -3.00860703e-01 6.31642222e-01 4.99014825e-01
-4.18368995e-01 -1.50068015e-01 -2.00032845e-01 -7.80278146e-01
-5.48449457e-01 -1.25814855e+00 1.36376739e+00 6.78766191e-01
-3.99860352e-01 -8.55529785e-01 -6.23749159e-02 4.87803996e-01
5.52031994e-01 2.70366464e-02 1.92590386e-01 -1.87167913e-01
-7.80104578e-01 1.55353755e-01 -4.23686922e-01 -7.22878799e-03
-1.13025680e-01 2.14685112e-01 -1.59564102e+00 -5.31929493e-01
-5.12132570e-02 -6.06996000e-01 7.25733817e-01 7.51865208e-01
1.01551580e+00 -1.37696475e-01 -4.84422415e-01 1.21756864e+00
9.18962538e-01 -3.00470382e-01 4.52324718e-01 4.76998240e-01
1.12574542e+00 7.52400756e-01 3.81025672e-01 2.32321899e-02
7.38574684e-01 7.64153183e-01 6.29582882e-01 -2.84698904e-01
-1.31493568e-01 1.37486473e-01 1.81851491e-01 3.87983829e-01
-2.44735047e-01 -3.84640992e-01 -9.32281673e-01 2.25870118e-01
-1.46227527e+00 -7.05889463e-01 -2.45260075e-01 2.11268139e+00
6.26820862e-01 -1.53053641e-01 4.86994714e-01 -2.60861427e-01
5.28122485e-01 6.66504279e-02 -7.44952440e-01 -2.18156621e-01
1.98742256e-01 1.77621037e-01 5.73859751e-01 1.85229361e-01
-1.27129018e+00 9.36617851e-01 6.86447525e+00 -1.28614813e-01
-1.64159727e+00 2.10366160e-01 5.30139685e-01 -7.24232435e-01
2.61742264e-01 -6.40897512e-01 -1.12721789e+00 4.55405235e-01
1.08367169e+00 3.12409669e-01 3.97024751e-01 8.79620433e-01
-3.04430898e-04 -1.40257135e-01 -1.18878794e+00 1.31715965e+00
9.54715371e-01 -8.46259832e-01 -8.31770658e-01 2.25794524e-01
5.90454459e-01 5.02258599e-01 6.08278155e-01 8.92801806e-02
-1.46884784e-01 -1.21803987e+00 8.05136144e-01 5.46494603e-01
1.29091418e+00 -4.35362130e-01 4.90573823e-01 -1.05493218e-01
-7.13647008e-01 -1.85020521e-01 2.77790036e-02 -8.78764540e-02
-1.82566792e-02 -3.84599686e-01 -6.90605760e-01 -2.05664262e-01
1.04697418e+00 7.57512569e-01 -1.12097061e+00 1.16673183e+00
-3.33719105e-01 5.44457316e-01 -2.37356201e-01 -1.98045000e-01
-1.75442651e-01 3.03314030e-01 3.40691864e-01 5.68304598e-01
8.41230899e-02 -3.55415419e-02 -7.57860541e-01 9.19833422e-01
-3.58055532e-01 -1.82437018e-01 -4.09702361e-01 6.40322983e-01
2.33565196e-01 1.19387877e+00 -2.72824019e-01 1.50595129e-01
-5.75508416e-01 6.25954807e-01 6.65647805e-01 5.83515346e-01
-8.75930429e-01 -1.31568074e-01 1.09578764e+00 2.67481953e-01
5.10289788e-01 -4.41992842e-02 -3.12015176e-01 -1.41269863e+00
7.87272230e-02 -8.86982381e-01 -5.35697490e-02 -1.39046645e+00
-1.02008712e+00 7.77976573e-01 -2.04007775e-01 -1.18506873e+00
-3.59651834e-01 -9.20819342e-01 -5.64624488e-01 1.09341848e+00
-1.87401724e+00 -1.37310553e+00 -7.45413661e-01 8.15015733e-01
4.54956144e-01 -1.50465220e-01 7.49762475e-01 1.00051291e-01
-9.21259940e-01 1.30169404e+00 -1.61804855e-01 -2.80096335e-03
1.26976776e+00 -1.05252707e+00 4.75980043e-01 7.22905159e-01
-4.91359532e-02 7.29340076e-01 7.40326822e-01 -3.50280926e-02
-1.27840269e+00 -7.58351266e-01 8.71673226e-01 -1.19023824e+00
3.67527783e-01 -7.12102950e-01 -9.07761157e-01 9.86506343e-01
4.34290051e-01 1.11465208e-01 5.41366041e-01 7.41955519e-01
-6.13350689e-01 -1.49525031e-01 -9.44649935e-01 6.35412455e-01
9.88760114e-01 -6.65764987e-01 -4.58045065e-01 4.30116877e-02
5.01972735e-01 -7.76805818e-01 -4.19849724e-01 2.39850700e-01
1.03766668e+00 -8.79276276e-01 7.17096627e-01 -6.15300298e-01
2.31557921e-01 -8.31495747e-02 5.27129531e-01 -1.08292389e+00
-2.26656601e-01 -7.52122581e-01 -2.93994308e-01 9.08780694e-01
3.93080026e-01 -6.64287686e-01 8.59120011e-01 1.07703447e+00
2.62799263e-01 -5.72476447e-01 -7.44077861e-01 -3.07189763e-01
-2.34945845e-02 -1.84245750e-01 5.79199910e-01 7.47196436e-01
-2.80294903e-02 8.03792834e-01 -6.19167686e-01 8.09327662e-02
7.27511823e-01 -2.14830562e-01 1.04220414e+00 -1.45892274e+00
2.04551116e-01 -3.91673237e-01 -5.16306341e-01 -1.06527209e+00
5.81691086e-01 -2.69443154e-01 1.15754440e-01 -7.67559171e-01
1.71983421e-01 -8.80660862e-02 -2.30442569e-01 4.10140812e-01
-5.70645034e-01 7.30756760e-01 2.80687332e-01 2.68235028e-01
-8.11846733e-01 4.56776738e-01 1.31422091e+00 9.85935479e-02
-1.62604332e-01 2.53100604e-01 -8.15113485e-01 7.12522924e-01
5.69395781e-01 -5.49857616e-01 -2.21695736e-01 -8.29445362e-01
3.22711110e-01 -2.43196040e-01 3.91496390e-01 -8.53225827e-01
4.44917202e-01 2.40833014e-01 6.58742666e-01 -3.80916893e-01
6.40251577e-01 -7.33518720e-01 -3.78470689e-01 -1.54950365e-01
-3.01923007e-01 3.19208056e-01 4.14511949e-01 3.88745159e-01
1.77308545e-01 -8.14208854e-03 8.45772028e-01 2.55684763e-01
-3.24613094e-01 3.81523371e-01 8.05733725e-02 1.14540577e-01
9.11807537e-01 -2.88211226e-01 -5.54831445e-01 -4.63936448e-01
-5.42349100e-01 7.78925195e-02 8.40844870e-01 7.10846186e-01
3.26080084e-01 -8.82331550e-01 -5.44874310e-01 4.96439874e-01
2.76441127e-01 -2.12266937e-01 -1.46714449e-01 1.22734690e+00
-3.27026129e-01 5.13367593e-01 -3.16529632e-01 -9.36837435e-01
-1.80299366e+00 6.51029587e-01 6.47469938e-01 6.40256464e-01
-3.20184901e-02 1.46895504e+00 4.57860023e-01 -2.17075691e-01
4.80978101e-01 -4.23543453e-01 -4.86607403e-01 6.80002989e-03
7.55933940e-01 2.58876860e-01 1.25111058e-01 -1.10726261e+00
-4.68838066e-01 8.29064071e-01 -1.01541787e-01 2.32063249e-01
1.17607832e+00 -4.75766510e-01 1.77117795e-01 3.07350934e-01
1.31255794e+00 -3.23848158e-01 -1.45838785e+00 -1.37958005e-01
-2.23829761e-01 -5.29252529e-01 2.78287351e-01 -6.98982239e-01
-1.31130195e+00 1.05123973e+00 1.01101863e+00 -2.19908357e-01
1.09805930e+00 1.38031673e-02 5.05566955e-01 2.59524435e-01
1.53452665e-01 -6.22852683e-01 -3.23335244e-03 4.15816367e-01
7.41051376e-01 -2.04598475e+00 6.73977658e-02 -2.05739245e-01
-7.56068647e-01 9.30636406e-01 9.80707109e-01 1.22059815e-01
8.10613692e-01 -1.13230780e-01 5.46495557e-01 -4.66755152e-01
-7.03599572e-01 -4.26056236e-01 8.49636257e-01 7.49451876e-01
6.77738070e-01 -2.46242642e-01 2.68393040e-01 3.84776108e-02
-4.04050559e-01 1.70361921e-01 4.02816951e-01 5.15632510e-01
-5.83767407e-02 -6.41635180e-01 -5.01094759e-01 4.02683496e-01
-7.32248187e-01 -2.70715356e-01 -4.14666623e-01 8.23898971e-01
-5.02771176e-02 9.57228065e-01 3.96514207e-01 -1.61036164e-01
2.65890211e-01 1.25596628e-01 8.42541039e-01 -4.11972374e-01
-6.67450905e-01 -7.16313533e-03 -9.82372612e-02 -6.61859035e-01
-8.15537870e-01 -9.99099076e-01 -4.00704652e-01 -3.91937435e-01
-1.00588536e+00 -6.47051752e-01 7.71597981e-01 9.79701400e-01
4.08340901e-01 3.36764187e-01 3.74606282e-01 -1.38552690e+00
-2.70648569e-01 -1.31891596e+00 -3.28854829e-01 1.57675922e-01
8.94810975e-01 -1.13693619e+00 -4.03620780e-01 2.72913158e-01] | [14.14551067352295, 0.07579293102025986] |
331d179d-083d-4aaa-acbd-40c73701d598 | biogan-an-unpaired-gan-based-image-to-image | 2306.06217 | null | https://arxiv.org/abs/2306.06217v1 | https://arxiv.org/pdf/2306.06217v1.pdf | BioGAN: An unpaired GAN-based image to image translation model for microbiological images | A diversified dataset is crucial for training a well-generalized supervised computer vision algorithm. However, in the field of microbiology, generation and annotation of a diverse dataset including field-taken images are time consuming, costly, and in some cases impossible. Image to image translation frameworks allow us to diversify the dataset by transferring images from one domain to another. However, most existing image translation techniques require a paired dataset (original image and its corresponding image in the target domain), which poses a significant challenge in collecting such datasets. In addition, the application of these image translation frameworks in microbiology is rarely discussed. In this study, we aim to develop an unpaired GAN-based (Generative Adversarial Network) image to image translation model for microbiological images, and study how it can improve generalization ability of object detection models. In this paper, we present an unpaired and unsupervised image translation model to translate laboratory-taken microbiological images to field images, building upon the recent advances in GAN networks and Perceptual loss function. We propose a novel design for a GAN model, BioGAN, by utilizing Adversarial and Perceptual loss in order to transform high level features of laboratory-taken images into field images, while keeping their spatial features. The contribution of Adversarial and Perceptual loss in the generation of realistic field images were studied. We used the synthetic field images, generated by BioGAN, to train an object-detection framework, and compared the results with those of an object-detection framework trained with laboratory images; this resulted in up to 68.1% and 75.3% improvement on F1-score and mAP, respectively. Codes is publicly available at https://github.com/Kahroba2000/BioGAN. | ['Anastasios D. Tsaousis', 'Sadiya Maxamhud', 'Md. Moinul Hossain', 'Gianluca Marcelli', 'Chee Siang Ang', 'Saber Mirzaee Bafti'] | 2023-06-09 | null | null | null | null | ['image-to-image-translation', 'image-to-image-translation'] | ['computer-vision', 'miscellaneous'] | [ 9.98656690e-01 -1.86018571e-01 5.85985541e-01 1.11660575e-02
-5.56144714e-01 -7.91165113e-01 5.00881672e-01 -1.69237554e-01
-3.32716405e-01 9.29251909e-01 -6.48313522e-01 -2.49067798e-01
1.40981629e-01 -1.13278353e+00 -1.17951000e+00 -1.14652944e+00
2.96055436e-01 2.90636003e-01 -1.36251107e-01 -1.81969497e-02
1.81121886e-01 5.83900213e-01 -1.43775368e+00 3.56415749e-01
9.10761833e-01 7.93489575e-01 6.59424007e-01 9.72459495e-01
1.52424080e-02 4.83075470e-01 -1.00519586e+00 -2.98057079e-01
4.26867634e-01 -9.53523159e-01 -4.43419158e-01 2.54821301e-01
2.09211051e-01 -2.00322822e-01 2.63566840e-02 8.57627809e-01
7.05200255e-01 -1.23858862e-01 7.96698153e-01 -1.28933370e+00
-1.08041990e+00 8.29268470e-02 -3.72623175e-01 -1.10965893e-01
1.82280600e-01 5.48863113e-01 1.26895204e-01 -6.69504881e-01
7.71601856e-01 9.94071245e-01 6.13987505e-01 6.37923717e-01
-1.31022847e+00 -6.75268710e-01 -4.80748713e-01 -1.19878806e-01
-1.35352337e+00 -1.38140887e-01 6.93148017e-01 -6.10850036e-01
5.44392705e-01 3.07825297e-01 6.64533913e-01 1.32898736e+00
4.82977837e-01 4.55152661e-01 1.54906309e+00 -6.12694323e-01
1.42151162e-01 3.31289291e-01 -8.17698002e-01 5.51824927e-01
2.92045593e-01 2.83437848e-01 -2.29317725e-01 2.18490914e-01
1.12587941e+00 2.43782312e-01 -4.02699351e-01 -1.59810841e-01
-1.17215824e+00 8.25671136e-01 5.40788352e-01 2.25564599e-01
-3.96411270e-01 -1.49914101e-01 8.57384428e-02 2.97809601e-01
3.99412572e-01 8.13255250e-01 -2.00492263e-01 3.10319811e-01
-5.67844748e-01 1.64906867e-02 5.53382933e-01 9.59164500e-01
9.39991295e-01 2.47045815e-01 -2.14398280e-01 7.83055961e-01
-9.52445194e-02 1.05806005e+00 4.60616797e-01 -7.34265208e-01
2.43277866e-02 5.80302894e-01 -4.06579599e-02 -1.01807964e+00
-4.07143459e-02 -4.40598637e-01 -1.02820587e+00 2.48564929e-01
3.12384397e-01 -1.86461210e-01 -1.13033319e+00 1.59726989e+00
3.24952394e-01 1.63700029e-01 3.31578076e-01 6.88459575e-01
6.51524663e-01 8.53416204e-01 -1.34405270e-01 -1.47697851e-01
1.04491651e+00 -8.12456846e-01 -4.81224656e-01 -7.39027485e-02
2.78215945e-01 -1.06086242e+00 1.10466385e+00 2.34619677e-01
-8.29586744e-01 -8.16053748e-01 -8.55767727e-01 3.42571080e-01
-6.55519605e-01 8.89972150e-02 2.59220868e-01 7.78442383e-01
-1.01219356e+00 2.67883122e-01 -6.33698761e-01 -6.68037653e-01
3.69667023e-01 2.93972552e-01 -4.19406831e-01 -3.41320902e-01
-1.00331235e+00 6.63364172e-01 4.68690991e-01 2.36704536e-02
-1.31427288e+00 -6.60254776e-01 -8.09415340e-01 -3.24066818e-01
2.79658705e-01 -7.44182348e-01 7.65649140e-01 -1.09123397e+00
-1.58287370e+00 9.79617417e-01 1.67003617e-01 -2.86927372e-01
6.07192934e-01 8.35704282e-02 -1.09248042e-01 1.81871459e-01
1.23726003e-01 9.71989810e-01 9.05013442e-01 -1.52200055e+00
-4.52656358e-01 -3.28324258e-01 -1.14620067e-01 -1.94256544e-01
-1.09889485e-01 -2.28608236e-01 -2.03007087e-01 -8.45008910e-01
-4.51370448e-01 -9.03388083e-01 -1.45611186e-02 -6.18603975e-02
-2.72228837e-01 3.77426177e-01 1.14178503e+00 -8.39499652e-01
3.08445454e-01 -2.01870251e+00 4.61914092e-02 -1.70627952e-01
-1.65712819e-01 5.48373342e-01 -5.40052593e-01 5.29846966e-01
5.07225581e-02 2.07704455e-01 -6.14538193e-01 -1.03946716e-01
-4.44904894e-01 2.59922773e-01 -2.23095924e-01 3.39114338e-01
6.45220220e-01 1.09112632e+00 -9.88607943e-01 -3.71163756e-01
3.80865842e-01 9.14852798e-01 -4.64121938e-01 6.75424337e-01
-1.14207610e-01 1.06512630e+00 -1.95168495e-01 8.09851587e-01
8.11657250e-01 -3.74756753e-02 -9.09539014e-02 -1.39630079e-01
1.25223339e-01 -6.26164734e-01 -6.11961365e-01 1.43518865e+00
-6.71773612e-01 5.60150743e-01 -7.59381056e-02 -1.00030422e+00
1.22117233e+00 1.28610358e-01 4.13619876e-01 -6.20027959e-01
1.70199737e-01 2.37378985e-01 -1.09075576e-01 -4.74321634e-01
2.10722774e-01 -2.39888638e-01 1.21514007e-01 1.57767519e-01
2.53633037e-02 -7.25384653e-01 1.88693985e-01 -2.27430612e-01
8.12518060e-01 4.77987051e-01 4.15460058e-02 -2.02589538e-02
7.03383386e-01 2.47954398e-01 4.45411712e-01 7.29335070e-01
8.93186331e-02 1.01731646e+00 2.36379296e-01 -4.28190321e-01
-1.40045452e+00 -9.96471286e-01 -3.98817956e-02 6.31401062e-01
3.78326140e-02 4.14376080e-01 -1.12746739e+00 -4.86143857e-01
-1.45960286e-01 5.71846545e-01 -5.45021176e-01 -3.74653786e-01
-5.69593966e-01 -9.18386459e-01 7.98102677e-01 3.43389124e-01
8.23634386e-01 -1.28395522e+00 -5.47585070e-01 2.19894409e-01
-1.62485763e-01 -1.03904855e+00 -2.74143904e-01 1.82394385e-01
-5.50570786e-01 -1.24891663e+00 -1.05456519e+00 -9.60616946e-01
9.48718965e-01 1.42879382e-01 9.21186805e-01 -6.28873110e-02
-6.69730484e-01 2.61136293e-01 -5.02120733e-01 -7.48329520e-01
-9.60734248e-01 -6.49597868e-02 -1.09663017e-01 6.91959783e-02
1.96606353e-01 -1.70540243e-01 -7.46731877e-01 6.08832598e-01
-1.56204975e+00 -2.17437046e-03 7.55952895e-01 1.32757866e+00
8.38085413e-01 1.06183745e-01 6.31017089e-01 -8.89161766e-01
2.95321167e-01 -2.22332016e-01 -7.75860786e-01 3.58283430e-01
-3.78036708e-01 -2.73033172e-01 1.03703892e+00 -5.06319225e-01
-1.11779094e+00 1.58396542e-01 -3.99027504e-02 -4.57321823e-01
-3.43111664e-01 2.27449924e-01 -2.25731313e-01 -4.22227323e-01
8.16112041e-01 6.26504302e-01 3.85525852e-01 -7.63281882e-02
1.68445587e-01 8.55002582e-01 6.78780019e-01 -5.60254097e-01
8.44188392e-01 4.82571095e-01 1.45687357e-01 -8.85226309e-01
-3.80839616e-01 -2.45488226e-01 -4.73528028e-01 -1.23583861e-01
9.90822256e-01 -7.86811352e-01 -2.88647294e-01 1.05378234e+00
-1.15411794e+00 -6.12203002e-01 -3.48947585e-01 3.49448174e-01
-6.71526730e-01 2.04044238e-01 -4.69145149e-01 -5.00856340e-01
-3.92072469e-01 -1.42482984e+00 1.27025807e+00 2.29057416e-01
2.93345571e-01 -1.01338351e+00 -8.98659751e-02 5.59856713e-01
6.62598968e-01 8.01892281e-01 6.34843230e-01 -3.60622764e-01
-6.41543031e-01 -3.07708949e-01 -1.54334009e-01 8.36516559e-01
6.15620732e-01 4.45559621e-02 -9.37169194e-01 -6.08428776e-01
3.03308994e-01 -4.68196839e-01 6.63411081e-01 3.15810710e-01
1.10821986e+00 -1.30152643e-01 -2.55023748e-01 8.76821697e-01
1.60932899e+00 6.88165307e-01 9.40120459e-01 1.88382849e-01
8.06384742e-01 5.38676381e-01 5.17565846e-01 1.46880552e-01
-1.67695045e-01 4.75584865e-01 4.11803931e-01 -4.63478953e-01
-3.25288117e-01 -2.19847888e-01 1.83983043e-01 5.30403078e-01
-2.84136474e-01 -7.39315510e-01 -7.81337380e-01 3.56961191e-01
-1.30858743e+00 -8.38950932e-01 2.01114729e-01 2.15874720e+00
8.11930597e-01 -3.43542188e-01 -2.41909698e-01 -4.68909217e-04
1.06436658e+00 -3.36167663e-01 -5.95413744e-01 -2.57091761e-01
-4.02434200e-01 4.03738528e-01 6.51855528e-01 2.77819037e-01
-1.09361112e+00 8.42947006e-01 5.89533472e+00 8.07841599e-01
-1.62350261e+00 7.67909884e-02 9.75399315e-01 3.37859064e-01
-2.95534208e-02 -3.64492923e-01 -4.57723379e-01 4.69679773e-01
7.98995495e-01 1.83723662e-02 4.70349163e-01 6.77585065e-01
8.30610991e-02 1.11153655e-01 -1.01638234e+00 8.35454643e-01
3.12951267e-01 -1.10012984e+00 3.37207437e-01 1.51979163e-01
1.08581042e+00 -4.14923012e-01 1.92736015e-01 5.64592071e-02
1.58758029e-01 -1.26446617e+00 2.31538728e-01 3.91705930e-01
1.11739206e+00 -6.11650646e-01 1.14675748e+00 2.30731070e-01
-8.67628992e-01 2.45085999e-01 -4.53081220e-01 2.02748016e-01
-6.41091317e-02 4.14100766e-01 -1.29852188e+00 6.61702454e-01
5.14512420e-01 3.75306159e-01 -5.07584989e-01 8.27161372e-01
-2.20130295e-01 4.58377570e-01 -7.43876770e-02 -8.29780381e-03
8.50374252e-02 -3.72010022e-01 3.35951954e-01 1.16705108e+00
6.58781528e-01 -3.06084067e-01 9.77935567e-02 1.09509563e+00
-4.54784408e-02 -5.84728122e-02 -1.06425774e+00 -1.92473620e-01
2.84138650e-01 1.07850683e+00 -8.44347775e-01 -2.08848402e-01
-5.82447611e-02 1.18082917e+00 -2.27132469e-01 4.80575770e-01
-9.64756370e-01 -6.29232287e-01 2.65875280e-01 1.08177297e-01
3.29552144e-01 5.65805174e-02 2.11190549e-03 -8.70707452e-01
-9.88588557e-02 -1.04195201e+00 9.26034153e-02 -8.48866343e-01
-1.24946356e+00 6.90983295e-01 -8.66256729e-02 -1.20089805e+00
-3.02344471e-01 -8.61257732e-01 -3.34652394e-01 1.04871595e+00
-1.36793339e+00 -1.52601731e+00 -8.52313876e-01 3.75833392e-01
5.69409907e-01 -3.68900627e-01 8.91331971e-01 2.11653709e-01
-3.21659118e-01 5.94466865e-01 4.43912327e-01 1.24732554e-01
7.26042688e-01 -9.54902649e-01 3.39311093e-01 8.73286784e-01
-2.02340871e-01 4.57370043e-01 4.49417949e-01 -5.66823900e-01
-1.43615162e+00 -1.71443582e+00 1.73789844e-01 -4.34507698e-01
1.18472658e-01 -3.01006019e-01 -8.28323305e-01 5.37547708e-01
3.85902822e-01 -2.73252632e-02 6.52410865e-01 -8.86477709e-01
-7.87473023e-02 -1.31983057e-01 -1.67422366e+00 4.09239471e-01
8.44440937e-01 -3.71621519e-01 -1.66855410e-01 4.62978721e-01
8.23110759e-01 -3.98874819e-01 -9.13714886e-01 4.51028168e-01
3.73647422e-01 -7.73194253e-01 1.10973930e+00 -3.41937751e-01
6.87287509e-01 -5.94760180e-01 -2.10412174e-01 -1.59248877e+00
-2.13390008e-01 -4.51349974e-01 5.47380388e-01 1.20015323e+00
2.01351315e-01 -9.03850496e-01 5.87009132e-01 -1.12720266e-01
-1.66560501e-01 -4.38832730e-01 -5.44787347e-01 -9.99213457e-01
2.23097563e-01 2.57562548e-01 7.58967221e-01 8.15289021e-01
-7.44701684e-01 7.03754798e-02 -3.73565584e-01 1.02982029e-01
5.12662411e-01 -1.14617199e-02 1.09927285e+00 -6.91748142e-01
-3.26872766e-01 -9.64386687e-02 -6.27177298e-01 -4.88481909e-01
-4.55963723e-02 -7.85343766e-01 2.06982329e-01 -1.47404206e+00
8.44215155e-02 -4.51431066e-01 -1.78644046e-01 3.29718769e-01
-2.18129337e-01 7.33171344e-01 1.25053346e-01 1.39199317e-01
7.73911104e-02 3.62177581e-01 1.72182703e+00 -4.93214071e-01
-6.50263578e-02 -2.39456847e-01 -4.46702600e-01 2.22865820e-01
1.07427239e+00 -3.74355555e-01 -4.59073633e-01 -3.96585822e-01
-2.92678505e-01 -1.28309280e-01 5.75780034e-01 -1.15549231e+00
-1.96083173e-01 -3.54109347e-01 6.47984624e-01 -1.84930801e-01
2.48193488e-01 -7.55694866e-01 7.25293875e-01 8.75846982e-01
5.50140887e-02 -1.65875360e-01 3.16226721e-01 3.66201997e-01
-3.62955451e-01 -2.56261975e-01 9.29084301e-01 -4.19322133e-01
-5.38047969e-01 1.47228047e-01 -1.79797858e-01 -1.87011957e-01
1.42686749e+00 -3.19559425e-01 -4.61659759e-01 -1.57430321e-01
-2.23868996e-01 -2.88219720e-01 9.07390535e-01 2.01614693e-01
7.10679412e-01 -1.02713990e+00 -9.32674527e-01 4.29505378e-01
1.90709248e-01 3.88653040e-01 2.11898074e-01 5.59282005e-01
-1.21010613e+00 2.91092157e-01 -7.30910122e-01 -8.65469754e-01
-1.02924049e+00 8.39899540e-01 3.56577456e-01 -1.62146986e-01
-5.44578321e-02 7.14735031e-01 5.90008318e-01 -7.75307059e-01
-2.24926934e-01 -2.02959299e-01 9.50522572e-02 -3.79789531e-01
3.41066033e-01 2.44396567e-01 1.83114693e-01 -4.65240389e-01
-1.25646666e-01 7.54524589e-01 2.50131458e-01 3.37357610e-01
1.21751809e+00 2.84779102e-01 -2.16353968e-01 -6.83931680e-03
1.25362957e+00 -1.62719667e-01 -1.22513926e+00 6.15562014e-02
-6.22486234e-01 -6.02411866e-01 -4.08118069e-01 -7.64696479e-01
-1.16461372e+00 8.26603115e-01 8.79984021e-01 1.42105833e-01
1.43794668e+00 -2.97129393e-01 6.38387322e-01 3.13271224e-01
4.67476875e-01 -6.21263683e-01 3.06183130e-01 2.51286328e-01
1.01988351e+00 -1.29771841e+00 -3.37380528e-01 -5.54751635e-01
-4.07839566e-01 9.32402611e-01 6.31003499e-01 2.26280801e-02
-2.36090533e-02 3.01764369e-01 3.74938875e-01 6.73299059e-02
-1.47174478e-01 1.50173694e-01 -1.28328279e-01 1.20041347e+00
3.65320742e-01 1.29479825e-01 -1.10278815e-01 -7.48026893e-02
-7.71178454e-02 2.82893002e-01 4.89771396e-01 1.13551700e+00
-1.05323628e-01 -1.22173405e+00 -8.25913072e-01 2.56144375e-01
-4.99579042e-01 5.72942942e-02 -3.38733584e-01 7.36311674e-01
4.31932986e-01 9.77089703e-01 1.31045058e-01 -3.62453282e-01
2.31088594e-01 -1.29711814e-02 4.68020499e-01 -5.49319386e-01
-4.94110852e-01 -7.71855265e-02 -3.51023942e-01 -1.07728112e-02
-5.89641511e-01 -4.60147232e-01 -8.12779546e-01 -1.97735637e-01
-3.10419381e-01 -9.95171163e-03 8.32084417e-01 4.75486249e-01
4.44542885e-01 7.29175210e-01 8.58401895e-01 -8.33678842e-01
-3.88996154e-01 -1.06545079e+00 -3.47713381e-01 6.28760278e-01
6.54179677e-02 -5.23525178e-01 -3.79766494e-01 7.59045422e-01] | [11.78054141998291, -0.5019364953041077] |
d6d46b19-b53c-421b-8a95-c1352486f73c | multi-agent-continual-coordination-via | 2305.13937 | null | https://arxiv.org/abs/2305.13937v1 | https://arxiv.org/pdf/2305.13937v1.pdf | Multi-agent Continual Coordination via Progressive Task Contextualization | Cooperative Multi-agent Reinforcement Learning (MARL) has attracted significant attention and played the potential for many real-world applications. Previous arts mainly focus on facilitating the coordination ability from different aspects (e.g., non-stationarity, credit assignment) in single-task or multi-task scenarios, ignoring the stream of tasks that appear in a continual manner. This ignorance makes the continual coordination an unexplored territory, neither in problem formulation nor efficient algorithms designed. Towards tackling the mentioned issue, this paper proposes an approach Multi-Agent Continual Coordination via Progressive Task Contextualization, dubbed MACPro. The key point lies in obtaining a factorized policy, using shared feature extraction layers but separated independent task heads, each specializing in a specific class of tasks. The task heads can be progressively expanded based on the learned task contextualization. Moreover, to cater to the popular CTDE paradigm in MARL, each agent learns to predict and adopt the most relevant policy head based on local information in a decentralized manner. We show in multiple multi-agent benchmarks that existing continual learning methods fail, while MACPro is able to achieve close-to-optimal performance. More results also disclose the effectiveness of MACPro from multiple aspects like high generalization ability. | ['Yang Yu', 'Cong Guan', 'Fuxiang Zhang', 'Ziqian Zhang', 'Lihe Li', 'Lei Yuan'] | 2023-05-07 | null | null | null | null | ['multi-agent-reinforcement-learning'] | ['methodology'] | [ 2.82818154e-02 1.15466006e-02 -3.06647927e-01 -6.77985977e-03
-7.12991774e-01 -3.21878612e-01 7.38681078e-01 3.79614532e-01
-9.17467713e-01 1.19884205e+00 -9.29226428e-02 1.19926736e-01
-7.62807846e-01 -3.77119720e-01 -6.07379913e-01 -1.14940131e+00
-3.85591835e-01 8.85621011e-01 2.01748371e-01 -2.91960537e-01
2.24391326e-01 1.88848883e-01 -1.51018989e+00 1.33986071e-01
1.06736159e+00 8.41206193e-01 7.06260443e-01 3.92388284e-01
6.28130510e-02 1.06906605e+00 -6.99683666e-01 -1.97548747e-01
2.87303388e-01 -1.02068856e-01 -8.08587313e-01 1.89935192e-01
-4.37297225e-02 -3.08912527e-02 3.14612031e-01 7.59337783e-01
4.99697626e-01 3.17472100e-01 5.80626547e-01 -1.78429186e+00
-1.96707264e-01 6.85463190e-01 -7.32553184e-01 1.99825644e-01
2.92524155e-02 3.41231227e-02 1.10089779e+00 -3.67530107e-01
5.46292901e-01 1.11566448e+00 3.03559273e-01 4.72906858e-01
-1.27578068e+00 -3.50679398e-01 7.58755445e-01 4.87067223e-01
-8.58748317e-01 -5.35148159e-02 7.16885805e-01 -3.60193312e-01
7.89729238e-01 -4.68556844e-02 4.50591385e-01 1.18408251e+00
4.48411763e-01 1.07625854e+00 1.57032132e+00 -3.79538000e-01
5.59232652e-01 1.98619246e-01 4.45869006e-02 3.12393606e-01
2.32497483e-01 -1.77081257e-01 -7.16294169e-01 -2.38660023e-01
4.35202241e-01 7.40410909e-02 4.09375280e-02 -5.03459334e-01
-1.36670101e+00 8.27983618e-01 9.82217267e-02 4.70379263e-01
-8.90487850e-01 1.11022875e-01 7.32206464e-01 7.22959876e-01
5.18046856e-01 5.69915652e-01 -7.33335555e-01 -1.73879355e-01
-5.73110104e-01 5.33084989e-01 6.66418672e-01 5.73987961e-01
9.69784021e-01 -1.30793396e-02 -2.45651647e-01 5.74893653e-01
-5.06032519e-02 2.95689017e-01 6.00078523e-01 -1.04056716e+00
5.69227695e-01 5.21511912e-01 2.78182119e-01 -7.43226647e-01
-8.11394870e-01 -6.80627167e-01 -8.54927301e-01 5.82936645e-01
5.55852056e-01 -5.33305347e-01 -1.57307908e-01 1.79139996e+00
5.08351207e-01 8.80331546e-02 2.24230647e-01 8.55529010e-01
-1.44923478e-01 3.62932026e-01 3.75574917e-01 -5.78609109e-01
1.47285628e+00 -1.12913764e+00 -6.54897153e-01 -3.05659771e-01
5.37472129e-01 -6.62349939e-01 6.92215323e-01 9.44322765e-01
-7.22122908e-01 -3.66418421e-01 -6.93950772e-01 6.49478853e-01
-2.56115139e-01 4.45212536e-02 7.48644590e-01 4.23838913e-01
-9.53459024e-01 4.41856444e-01 -7.18665183e-01 -1.50049582e-01
3.88807476e-01 5.21119535e-01 -3.93509120e-01 -1.29748946e-02
-1.11887813e+00 8.20013940e-01 5.36710024e-01 -1.33598983e-01
-1.05449760e+00 -4.62175667e-01 -3.30238998e-01 1.00576401e-01
1.21975875e+00 -6.01116300e-01 1.29437947e+00 -1.26599717e+00
-1.43426073e+00 1.55824438e-01 2.40108088e-01 -7.97887206e-01
7.92776346e-01 -2.75946796e-01 -7.26010464e-03 6.82218447e-02
2.71661520e-01 4.47564632e-01 1.16766143e+00 -1.36283541e+00
-1.25116634e+00 -4.18006897e-01 4.03029084e-01 7.32375562e-01
-6.89923763e-01 -7.29189515e-02 -3.09845433e-04 -6.66806579e-01
-5.33367872e-01 -1.08795917e+00 -3.91882688e-01 -6.37162983e-01
-2.35451519e-01 -7.61077583e-01 7.20271826e-01 -2.26146594e-01
1.08172429e+00 -1.74934530e+00 4.67858315e-01 -7.10722357e-02
4.04023051e-01 2.17345670e-01 -1.31745607e-01 6.36352062e-01
1.66755512e-01 -2.79341698e-01 -1.16832908e-02 -8.26569557e-01
1.78871095e-01 3.26975346e-01 4.54917364e-02 5.09117603e-01
-4.21186127e-02 5.31285405e-01 -9.41849828e-01 -5.45889497e-01
-1.05299437e-02 1.97098255e-02 -4.28932846e-01 9.75771174e-02
-4.75503117e-01 5.89782000e-01 -9.82725441e-01 3.99490029e-01
5.90700746e-01 -1.56349450e-01 4.24887180e-01 2.78935164e-01
-2.07254887e-01 -2.72329271e-01 -1.45290852e+00 1.69712114e+00
-6.50936961e-01 2.61510313e-01 4.12374467e-01 -1.37899292e+00
7.20482111e-01 5.50759435e-01 1.10432315e+00 -8.66328597e-01
-1.66397452e-01 3.43865097e-01 -2.28914786e-02 -3.24974895e-01
4.36211258e-01 4.11063917e-02 -5.12127616e-02 7.81816006e-01
-6.11299016e-02 4.26859796e-01 4.05979812e-01 1.27707273e-01
1.06935143e+00 1.84325248e-01 3.77773196e-01 -3.94595623e-01
6.63208723e-01 2.11674184e-01 7.04687238e-01 9.11052108e-01
-3.70221555e-01 -8.88965875e-02 6.74135685e-01 -4.93797988e-01
-7.02317715e-01 -4.21737134e-01 2.75097579e-01 1.58225799e+00
1.21278971e-01 -1.17687851e-01 -6.07206225e-01 -8.94593656e-01
2.25019991e-01 2.98333049e-01 -8.33995938e-01 -2.61933450e-02
-7.01484084e-01 -1.00012803e+00 7.27308691e-02 3.87564786e-02
5.92781425e-01 -1.45680833e+00 -1.13484991e+00 6.42313540e-01
-1.31568342e-01 -1.09398556e+00 -2.72628129e-01 5.78931689e-01
-5.97391605e-01 -1.23643208e+00 -9.21414137e-01 -6.08418941e-01
2.73801267e-01 3.71507823e-01 1.01030779e+00 -1.50232732e-01
-8.07222426e-02 6.23344302e-01 -5.84969759e-01 -4.41123545e-01
-3.07997316e-01 5.76934338e-01 2.11463124e-01 4.39654499e-01
6.34654164e-02 -5.41719556e-01 -5.06893396e-01 4.80686247e-01
-8.54377151e-01 -1.47818342e-01 8.29081655e-01 9.22384083e-01
4.27178144e-01 2.46801555e-01 1.35541379e+00 -1.03645611e+00
8.85323942e-01 -6.32092237e-01 -5.83711445e-01 5.11469662e-01
-9.63963628e-01 1.59816474e-01 9.51036632e-01 -5.52874327e-01
-1.26960373e+00 1.05107032e-01 2.92884707e-01 -2.37809718e-01
-1.70632899e-01 4.86699283e-01 2.30697505e-02 6.53994605e-02
5.95919609e-01 3.86218905e-01 2.29265362e-01 -3.10291171e-01
1.69365793e-01 4.02647763e-01 -7.94720836e-03 -9.70695257e-01
5.81967413e-01 4.07312512e-01 9.16626900e-02 -4.72244918e-01
-5.67093372e-01 -4.78160530e-01 -3.88117492e-01 -3.21426749e-01
6.94615245e-01 -8.74648035e-01 -1.20719290e+00 6.54600024e-01
-1.06645322e+00 -6.68404520e-01 -3.08527529e-01 6.41829133e-01
-8.33176136e-01 2.91913301e-01 -3.19223642e-01 -9.42380309e-01
-2.11465180e-01 -1.27263880e+00 8.39432001e-01 1.44855499e-01
3.04425538e-01 -1.11731601e+00 3.27157140e-01 3.28505933e-01
5.51715195e-01 -1.48137696e-02 8.09055328e-01 -1.00782120e+00
-5.35953224e-01 1.52769208e-01 1.43356293e-01 2.51388788e-01
1.35642439e-01 -5.40434182e-01 -9.07383740e-01 -6.07014656e-01
-1.81762055e-02 -6.34492517e-01 7.76943326e-01 3.95881951e-01
5.98476291e-01 -1.55339137e-01 -2.62323529e-01 -7.64898807e-02
1.41765738e+00 3.36502165e-01 1.75681889e-01 8.29353034e-01
3.00051391e-01 8.38433981e-01 1.06201458e+00 9.66224551e-01
4.82372761e-01 9.66974258e-01 8.67887259e-01 6.27815202e-02
1.20240241e-01 3.44042897e-01 4.99035716e-01 3.46426666e-01
-2.91547388e-01 -2.09458530e-01 -6.36708021e-01 3.40953231e-01
-2.41959500e+00 -9.30711091e-01 1.62404865e-01 2.18953419e+00
5.64239383e-01 9.40443575e-02 5.40576696e-01 -3.89748416e-03
6.63550675e-01 1.05250403e-01 -5.79477489e-01 -2.34696895e-01
-3.39191020e-01 -2.91318178e-01 3.67389500e-01 2.65543908e-01
-1.18930674e+00 9.39937890e-01 4.50166702e+00 1.21995568e+00
-9.79247749e-01 4.71912950e-01 5.81372678e-01 3.29947360e-02
7.13397861e-02 -1.16850674e-01 -8.52176070e-01 3.62984300e-01
5.40945411e-01 -2.67052174e-01 6.45040512e-01 7.79565454e-01
3.73001456e-01 -5.24108052e-01 -8.50066900e-01 7.03247070e-01
-3.28718945e-02 -1.02920437e+00 -2.99445212e-01 2.87611306e-01
8.00587296e-01 4.13589999e-02 8.11176524e-02 6.39733076e-01
4.01855886e-01 -5.83252311e-01 7.77023315e-01 4.78550404e-01
2.84354389e-01 -7.81890810e-01 7.65050232e-01 8.21634293e-01
-1.20658207e+00 -5.75130820e-01 -3.21680784e-01 -1.33851737e-01
-4.60204780e-02 4.33333129e-01 -8.40721965e-01 1.00958121e+00
5.28613746e-01 4.54021305e-01 -4.99641240e-01 9.72680151e-01
5.53504601e-02 2.99673110e-01 -7.42276013e-02 -1.32927418e-01
5.40005565e-01 -1.63265273e-01 6.32592261e-01 7.74957538e-01
-5.88581711e-03 -4.49997932e-01 7.39103496e-01 2.35389709e-01
2.20936865e-01 3.93829614e-01 -4.08369094e-01 3.61896932e-01
2.44778529e-01 1.42728364e+00 -8.52566838e-01 -1.85236037e-01
-4.23853874e-01 9.09622967e-01 5.63346148e-01 3.59005630e-01
-7.49654055e-01 -4.09643501e-02 4.76739109e-01 -1.68969184e-01
2.28998631e-01 -2.44836226e-01 1.04481071e-01 -8.93072963e-01
2.02091739e-01 -1.05454016e+00 5.20722806e-01 -1.00612178e-01
-1.37765300e+00 8.26461077e-01 6.75299838e-02 -1.23228347e+00
-5.22845805e-01 -3.80208582e-01 -5.15323758e-01 5.14555097e-01
-1.96689296e+00 -1.14422154e+00 -1.42961755e-01 1.01809406e+00
9.03405666e-01 -6.18286490e-01 7.60565817e-01 3.34705770e-01
-7.59859383e-01 4.01767582e-01 3.40027750e-01 -4.74831402e-01
1.02538967e+00 -1.54995871e+00 -3.95746231e-01 4.01974261e-01
-7.46630803e-02 5.82873374e-02 6.71933830e-01 -5.05799115e-01
-1.38567019e+00 -1.00353813e+00 5.95988154e-01 2.33975407e-02
7.38037705e-01 -2.20701814e-01 -6.70771182e-01 4.63668585e-01
4.63052064e-01 -4.27147225e-02 2.46661276e-01 2.26432338e-01
8.60154554e-02 -4.21105713e-01 -8.83825302e-01 2.57201880e-01
6.52502596e-01 3.94265056e-02 -2.16690779e-01 6.10136926e-01
6.17411256e-01 7.96598103e-03 -6.99275613e-01 -7.46198595e-02
8.48609880e-02 -1.04474807e+00 6.12155676e-01 -6.17044032e-01
1.57609046e-01 -1.59566663e-03 1.42874688e-01 -1.63767540e+00
-1.59603208e-01 -8.74249339e-01 -5.74134588e-02 1.13213086e+00
2.87298620e-01 -9.61305499e-01 8.80497336e-01 3.12111735e-01
-2.25713268e-01 -7.66518354e-01 -1.29560840e+00 -1.10423684e+00
1.43996045e-01 -1.83441237e-01 2.72109509e-01 9.23713446e-01
-5.94388656e-02 3.15763980e-01 -7.85691977e-01 -2.34707445e-03
8.07411909e-01 2.26108089e-01 7.74092376e-01 -1.46952891e+00
-5.57746410e-01 -4.45358574e-01 5.82253411e-02 -7.75408387e-01
4.72181976e-01 -7.05095351e-01 -1.20033488e-01 -1.33702850e+00
9.75554138e-02 -9.10105109e-01 -4.93171453e-01 7.28282690e-01
-1.58005595e-01 -2.34108299e-01 5.30144453e-01 4.35722172e-01
-1.35221601e+00 6.31612480e-01 1.31408250e+00 -8.94130692e-02
-4.37331378e-01 4.51935709e-01 -5.80582023e-01 4.32708889e-01
9.73243237e-01 -6.85496271e-01 -4.95298564e-01 -3.33130360e-01
2.95498431e-01 3.15187812e-01 2.49724925e-01 -1.09263074e+00
6.02663398e-01 -4.53751981e-01 -3.20551783e-01 6.72084168e-02
2.74571121e-01 -1.12195051e+00 -4.60931845e-02 6.28276587e-01
-3.51769030e-01 2.67286837e-01 -1.28955901e-01 9.15399194e-01
-2.79038399e-01 -4.51602221e-01 4.39983845e-01 -1.81972325e-01
-7.22568333e-01 2.22353026e-01 -5.25470138e-01 -4.07147221e-02
1.49476242e+00 1.28120825e-01 -4.56651121e-01 -2.10866526e-01
-8.77094269e-01 7.11607635e-01 -5.16849868e-02 3.40385020e-01
9.03553814e-02 -8.61145973e-01 -9.22874987e-01 -3.45498860e-01
5.03605455e-02 -5.74585944e-02 5.57506204e-01 1.10790968e+00
1.36362165e-01 4.39687401e-01 -5.17269850e-01 -4.91947770e-01
-1.07523441e+00 6.08996630e-01 2.82386035e-01 -1.09285879e+00
-4.14388299e-01 3.15777063e-01 2.01963365e-01 -1.16001964e-01
3.77792925e-01 1.29266024e-01 -6.10765576e-01 6.56332612e-01
2.14910552e-01 6.25911474e-01 2.62190402e-01 -2.41592571e-01
-2.61300325e-01 3.26243281e-01 -3.01589370e-01 -4.42310013e-02
1.54703712e+00 -2.32662290e-01 3.64381224e-02 2.70521343e-01
7.16095030e-01 -1.35836408e-01 -1.71519470e+00 -5.65190077e-01
3.11479747e-01 -9.72615778e-02 -3.91417593e-02 -8.45997334e-01
-1.17960179e+00 4.76513088e-01 5.35518408e-01 6.08733654e-01
1.10194707e+00 -1.92077830e-01 3.30578744e-01 6.18739367e-01
7.71119952e-01 -1.45107245e+00 4.43583012e-01 4.94811743e-01
7.55694091e-01 -1.41009212e+00 -4.75599393e-02 -1.08145848e-01
-1.13780499e+00 1.16557920e+00 7.08400726e-01 -1.07192472e-01
4.61725771e-01 1.72199756e-01 -1.04728296e-01 -1.08992681e-01
-1.10768008e+00 -2.68032640e-01 -2.96235621e-01 6.01781964e-01
-5.62953576e-02 1.88730314e-01 -4.94713664e-01 5.99081516e-01
5.07111013e-01 -5.12086004e-02 5.43116212e-01 1.01544583e+00
-4.71122563e-01 -1.35656583e+00 -4.35647666e-01 3.85617197e-01
-3.72179061e-01 1.54417872e-01 1.92159668e-01 8.96614671e-01
2.90323973e-01 8.45201313e-01 -1.45761862e-01 -1.79340113e-02
2.69941688e-01 -4.99858242e-03 1.89283684e-01 -4.37129498e-01
-1.01817191e+00 3.45075250e-01 -8.73227492e-02 -4.11427736e-01
-7.17902303e-01 -8.44206095e-01 -1.08608246e+00 -2.41538361e-02
-1.30922228e-01 4.34424758e-01 6.63329959e-01 1.03740406e+00
3.99522901e-01 7.22265542e-01 9.71609116e-01 -8.73824835e-01
-9.97753143e-01 -8.57690573e-01 -7.95965016e-01 2.36984909e-01
3.05720925e-01 -9.90594506e-01 -2.46169716e-01 -2.14812458e-01] | [3.7770791053771973, 1.9724757671356201] |
44a7d03d-b7c5-4e9c-830d-2a05af40a7ef | simultaneous-traffic-sign-detection-and | 1802.10019 | null | http://arxiv.org/abs/1802.10019v1 | http://arxiv.org/pdf/1802.10019v1.pdf | Simultaneous Traffic Sign Detection and Boundary Estimation using Convolutional Neural Network | We propose a novel traffic sign detection system that simultaneously
estimates the location and precise boundary of traffic signs using
convolutional neural network (CNN). Estimating the precise boundary of traffic
signs is important in navigation systems for intelligent vehicles where traffic
signs can be used as 3D landmarks for road environment. Previous traffic sign
detection systems, including recent methods based on CNN, only provide bounding
boxes of traffic signs as output, and thus requires additional processes such
as contour estimation or image segmentation to obtain the precise sign
boundary. In this work, the boundary estimation of traffic signs is formulated
as a 2D pose and shape class prediction problem, and this is effectively solved
by a single CNN. With the predicted 2D pose and the shape class of a target
traffic sign in an input image, we estimate the actual boundary of the target
sign by projecting the boundary of a corresponding template sign image into the
input image plane. By formulating the boundary estimation problem as a
CNN-based pose and shape prediction task, our method is end-to-end trainable,
and more robust to occlusion and small targets than other boundary estimation
methods that rely on contour estimation or image segmentation. The proposed
method with architectural optimization provides an accurate traffic sign
boundary estimation which is also efficient in compute, showing a detection
frame rate higher than 7 frames per second on low-power mobile platforms. | ['Hee Seok Lee', 'Kang Kim'] | 2018-02-27 | null | null | null | null | ['traffic-sign-detection'] | ['computer-vision'] | [ 2.97891825e-01 -4.13204804e-02 -4.09310907e-01 -5.07312536e-01
-5.48014164e-01 -3.92433017e-01 3.34393263e-01 -6.93685174e-01
-5.19144118e-01 2.59392887e-01 -4.07792777e-01 -7.62112081e-01
2.22313479e-01 -5.36596775e-01 -7.07541466e-01 -4.18426514e-01
3.39962631e-01 6.87694907e-01 7.82012284e-01 3.27011682e-02
5.60085535e-01 6.66988671e-01 -1.63491106e+00 -1.13770060e-01
8.36386263e-01 1.56009984e+00 -1.34991750e-01 7.90103436e-01
-3.95106971e-01 2.61870980e-01 -5.45714080e-01 -3.12383354e-01
4.83177841e-01 -2.79149443e-01 -1.77410871e-01 4.24698405e-02
1.34754264e+00 -5.37770808e-01 -2.50580817e-01 8.37419271e-01
4.03642923e-01 -2.98941165e-01 6.97810709e-01 -1.41897058e+00
6.23427108e-02 -3.14135164e-01 -5.99252105e-01 -2.36112371e-01
-1.39660910e-01 1.99606284e-01 6.93634212e-01 -8.56697500e-01
7.30985105e-01 7.62437999e-01 9.00291979e-01 6.68104649e-01
-6.00095868e-01 -9.94970560e-01 1.13753304e-01 4.05217230e-01
-1.41901803e+00 -3.54603559e-01 9.02221918e-01 -3.98790449e-01
3.71227682e-01 1.71087742e-01 8.13341916e-01 2.28798285e-01
-5.16076759e-02 1.04998136e+00 8.93311262e-01 -1.53420985e-01
9.86701772e-02 -3.90601277e-01 1.67057708e-01 9.90600586e-01
4.20566916e-01 7.67739862e-02 -1.73944652e-01 2.26471484e-01
9.15604472e-01 -2.95867562e-01 -1.78299546e-01 -5.56566536e-01
-1.05920792e+00 3.61299336e-01 6.30691350e-01 -2.59676576e-01
-2.89676517e-01 7.77238607e-01 2.41569370e-01 -6.55174404e-02
7.81347752e-02 -1.94732487e-01 -4.96147066e-01 -3.44229877e-01
-9.65944231e-01 1.09091336e-02 6.19309306e-01 9.81667638e-01
6.91814542e-01 1.16825335e-01 -5.30495457e-02 4.76609766e-01
6.05991602e-01 1.26001096e+00 -1.35592028e-01 -8.77438247e-01
6.68926418e-01 7.09156990e-01 2.25916490e-01 -8.14656556e-01
-6.14840984e-01 -6.35879785e-02 -4.36523676e-01 5.71906030e-01
1.07800567e+00 -3.25207591e-01 -1.51345563e+00 1.20339334e+00
6.27706409e-01 4.40885633e-01 -4.19450819e-01 1.17277706e+00
9.02341843e-01 3.63282084e-01 -1.08442837e-02 4.45475489e-01
1.48649573e+00 -8.87474954e-01 -5.16908109e-01 -4.98698950e-01
7.28043020e-01 -7.59011269e-01 5.74377716e-01 3.06086808e-01
-8.14856470e-01 -4.13277864e-01 -1.03584826e+00 -2.66154081e-01
-4.27888691e-01 7.88488746e-01 5.36513567e-01 9.39279139e-01
-8.29962790e-01 -5.54431267e-02 -4.99448746e-01 -8.93850848e-02
6.26042426e-01 6.10128045e-01 -8.67955163e-02 -4.85296138e-02
-7.89440334e-01 9.35233474e-01 2.18935460e-02 8.60536933e-01
-1.46818981e-01 -5.25740325e-01 -8.40589404e-01 -1.49837941e-01
3.84146988e-01 -4.38902795e-01 1.14694893e+00 -7.66575634e-01
-1.54736817e+00 9.14992213e-01 -4.58817571e-01 -1.42440915e-01
8.77152562e-01 -1.08703477e-02 -2.55883425e-01 6.84067979e-02
-1.96764663e-01 8.75831425e-01 1.05535865e+00 -1.13499951e+00
-9.16603267e-01 -8.00670162e-02 -3.34149599e-01 -1.09872691e-01
3.58055472e-01 -1.09554594e-02 -8.13473225e-01 4.45975885e-02
5.27969837e-01 -9.63029385e-01 -1.35912432e-03 9.46270764e-01
-2.00267240e-01 -2.74926633e-01 1.54026318e+00 -8.24638665e-01
8.97978604e-01 -1.79591644e+00 -7.06670582e-01 7.23235130e-01
9.79914442e-02 8.16187143e-01 -1.13453604e-01 -4.98773873e-01
3.51163596e-01 9.25068744e-03 -7.00193122e-02 -1.16982698e-01
7.42336065e-02 2.58976314e-02 -1.36544898e-01 5.54728091e-01
1.36355892e-01 1.41051936e+00 -6.91243947e-01 -7.88792670e-01
4.33637410e-01 6.48775280e-01 -2.39879936e-01 -2.32520625e-01
-1.47219941e-01 2.05199331e-01 -4.75571990e-01 9.72186148e-01
1.13725495e+00 6.05013706e-02 -4.81505990e-02 -4.92732882e-01
-2.31612995e-01 1.59881070e-01 -1.22730494e+00 7.83469319e-01
-4.41280425e-01 1.30518723e+00 2.46591598e-01 -6.97855115e-01
1.31300020e+00 2.35130545e-02 3.37007284e-01 -9.65517759e-01
3.69290799e-01 7.72425413e-01 -3.44329067e-02 -4.08278614e-01
3.96591485e-01 2.34856352e-01 -9.68794674e-02 4.70656425e-01
-6.11811399e-01 -2.75145620e-01 3.09063643e-01 -4.33434844e-01
7.27704585e-01 3.58588845e-01 -3.30967933e-01 4.27709043e-01
6.16068780e-01 1.30897462e-01 6.97806299e-01 5.65810621e-01
-4.52048838e-01 6.33189499e-01 4.66905057e-01 -6.63296223e-01
-1.14662111e+00 -8.37107956e-01 -2.83856899e-01 4.07253414e-01
5.25758088e-01 3.03049386e-01 -6.70913994e-01 -9.31679547e-01
1.84979483e-01 1.73234180e-01 -3.16703200e-01 3.38983297e-01
-1.41483176e+00 -1.66036516e-01 6.09039426e-01 8.47735524e-01
1.00566006e+00 -9.40912008e-01 -6.34160876e-01 -6.24437369e-02
-1.58395662e-04 -1.27584124e+00 -8.20954978e-01 -1.94919795e-01
-6.86999619e-01 -1.50742936e+00 -9.73281026e-01 -1.18436623e+00
1.08867788e+00 2.30482668e-01 3.99043381e-01 4.14825052e-01
-1.87255383e-01 -1.79062528e-03 -7.29806162e-03 -2.84442186e-01
-3.26938555e-02 -2.08600730e-01 -4.89887595e-01 2.98048317e-01
2.14197084e-01 5.04890364e-03 -9.73021984e-01 1.04146004e+00
-3.64146799e-01 3.08140814e-01 7.03111112e-01 5.98148465e-01
5.74090421e-01 -5.21134198e-01 3.48104805e-01 -3.34169775e-01
7.31181875e-02 2.41757736e-01 -1.09212065e+00 3.23751211e-01
-1.99224412e-01 -3.17394920e-02 2.74802119e-01 -4.01546359e-01
-9.55572963e-01 5.10501683e-01 -5.71046621e-02 -2.79805720e-01
-2.26269439e-01 9.46961641e-02 -7.02277869e-02 -6.16068304e-01
1.99308306e-01 1.66335881e-01 2.29223400e-01 -2.06448123e-01
4.01235908e-01 8.67582500e-01 5.85814595e-01 -6.63161278e-02
8.77682924e-01 6.45898700e-01 4.63250846e-01 -9.29079652e-01
-4.14563060e-01 -5.72406113e-01 -8.73403013e-01 -7.92480588e-01
7.09508419e-01 -3.99059564e-01 -1.33445203e+00 8.43659699e-01
-1.47143364e+00 -4.59464759e-01 1.93538681e-01 4.99644637e-01
-4.30241346e-01 4.34280455e-01 -5.60796186e-02 -6.99406326e-01
-5.29709637e-01 -9.82479930e-01 1.48312688e+00 5.14144540e-01
-1.69554073e-02 -7.39686668e-01 -4.71652865e-01 6.76058173e-01
5.57533085e-01 1.76759303e-01 6.95302308e-01 -1.14960730e-01
-1.35586321e+00 -9.55464423e-01 -7.88856089e-01 6.80676177e-02
-2.59831876e-01 1.20969981e-01 -7.23896444e-01 3.22721511e-01
-7.50155091e-01 3.70268635e-02 8.72030318e-01 6.23939455e-01
7.54850924e-01 1.05184115e-01 -5.10892630e-01 7.68379152e-01
1.09861887e+00 3.71333838e-01 5.89182377e-01 2.07267612e-01
8.64976764e-01 4.00903195e-01 8.33144486e-01 -7.20616281e-02
4.33604509e-01 8.13760638e-01 2.56520212e-01 -3.01942527e-01
-7.36836731e-01 -2.07450271e-01 2.40129307e-01 3.78043205e-01
4.89500165e-02 -6.96463510e-02 -1.05433571e+00 4.74726617e-01
-1.81108427e+00 -8.09894562e-01 -8.33033442e-01 2.20111108e+00
2.98087418e-01 1.86210170e-01 1.00228608e-01 1.35241821e-01
9.72422302e-01 -1.76608220e-01 -7.19208658e-01 -4.99428958e-01
1.86864510e-01 -4.59347665e-03 9.33511198e-01 5.83680689e-01
-1.05713511e+00 1.12023628e+00 6.01671600e+00 7.45698929e-01
-1.57853770e+00 -3.90238792e-01 4.57826018e-01 3.69178802e-01
-1.47117663e-03 -5.86622916e-02 -1.13114035e+00 4.38988566e-01
3.23364079e-01 3.69760752e-01 -3.27828862e-02 9.01919007e-01
3.90136153e-01 -4.79835719e-01 -8.42563331e-01 8.80988538e-01
8.25261623e-02 -1.27213216e+00 -3.06586623e-01 8.60820413e-02
6.62647605e-01 1.54765666e-01 2.45207045e-02 8.96313414e-02
-1.63062379e-01 -8.04961801e-01 8.03374887e-01 6.15144014e-01
1.13742483e+00 -4.42601800e-01 9.18210268e-01 3.02965552e-01
-1.64588273e+00 1.45662487e-01 1.55790774e-02 1.55097380e-01
6.90357983e-01 1.62403300e-01 -1.18814266e+00 -1.45955637e-01
1.00267000e-01 4.22223508e-01 -5.69020152e-01 1.89937615e+00
-7.01804161e-01 5.65736055e-01 -8.36212575e-01 -4.94423956e-01
2.38506183e-01 -2.32471645e-01 2.97481090e-01 1.06343794e+00
2.97964633e-01 -3.36915612e-01 2.58352011e-02 8.03723514e-01
1.68108325e-02 -4.84878235e-02 -2.62928605e-01 4.26126570e-01
3.68166864e-01 1.05961394e+00 -1.15246391e+00 -3.65272760e-01
-2.06769064e-01 6.94846094e-01 -2.39392877e-01 4.10335183e-01
-1.13320041e+00 -8.44422519e-01 5.66934347e-01 1.19063891e-01
8.46629858e-01 -3.89882535e-01 -8.03240120e-01 -7.48719692e-01
5.41215539e-01 -1.70524240e-01 -7.52782822e-02 -9.49646831e-01
-4.39547509e-01 1.23119771e-01 -4.26811695e-01 -1.74116278e+00
2.00362541e-02 -1.05034626e+00 -9.61650491e-01 6.36842430e-01
-1.95106351e+00 -1.36988628e+00 -6.64025784e-01 7.11269528e-02
1.92474574e-01 3.03317308e-01 2.74124086e-01 4.76380318e-01
-4.25738126e-01 7.53710270e-01 -6.30014464e-02 6.07806742e-01
3.45524877e-01 -8.63112867e-01 7.04706490e-01 6.88635528e-01
-1.68903217e-01 -6.71423413e-03 3.90336737e-02 -9.11878407e-01
-1.20667720e+00 -9.24798429e-01 1.26662242e+00 -4.06844914e-01
3.40384275e-01 -2.77944326e-01 -4.46993619e-01 2.91605294e-01
-5.49220741e-01 4.27097738e-01 1.97669826e-02 -3.20134193e-01
-2.13726401e-01 -3.25846165e-01 -9.85240459e-01 6.69206738e-01
1.07710326e+00 -2.36683369e-01 -1.84390724e-01 1.83795951e-02
-4.72926870e-02 -7.32096493e-01 -1.94847867e-01 4.48601156e-01
1.12510657e+00 -3.98555309e-01 9.74024534e-01 -1.74684867e-01
-1.44732147e-01 -8.65569949e-01 5.53136587e-01 -5.44667482e-01
3.75122011e-01 -4.95783776e-01 -4.31737304e-02 8.00161660e-01
5.63390732e-01 -4.71790791e-01 1.54056835e+00 9.72117126e-01
-2.20539510e-01 -8.74589980e-01 -1.38519096e+00 -6.22895300e-01
-5.02468407e-01 -9.59243834e-01 4.64564443e-01 1.38990283e-01
-2.92172551e-01 -1.59145683e-01 -2.10264146e-01 1.47950336e-01
6.96483016e-01 3.43409568e-01 1.20922232e+00 -1.18013299e+00
5.90614438e-01 -8.72114420e-01 -9.32541132e-01 -1.95710075e+00
3.41470838e-02 -7.28974044e-01 4.79669064e-01 -1.73936832e+00
-5.43431342e-01 -7.06088185e-01 2.56538659e-01 5.53105593e-01
1.27006680e-01 6.71664476e-01 1.93030253e-01 -6.62822947e-02
-6.39982879e-01 2.86034852e-01 1.67847252e+00 -2.54675537e-01
-1.95992976e-01 4.44652945e-01 1.58176586e-01 9.72753048e-01
5.82507312e-01 -2.18971595e-01 1.26245484e-01 -3.13520789e-01
1.61922351e-01 -8.84592608e-02 6.09429955e-01 -8.07127893e-01
6.88415170e-01 -2.22732462e-02 3.12059790e-01 -1.30019820e+00
4.48532552e-01 -9.84364688e-01 -4.98638153e-01 6.40501320e-01
1.05565831e-01 -5.54049015e-01 2.13997170e-01 3.53376687e-01
-7.69318268e-02 -2.99570948e-01 5.73598981e-01 4.24301177e-01
-1.13722694e+00 3.82016957e-01 -4.37764138e-01 -1.07376896e-01
1.07249510e+00 -1.12813246e+00 -3.39518487e-01 -4.88297135e-01
-3.61268282e-01 5.60300708e-01 1.34747133e-01 4.38018352e-01
1.06309319e+00 -1.47332942e+00 -4.91032869e-01 5.78776062e-01
-7.09785968e-02 2.92345554e-01 6.71177590e-03 1.03811610e+00
-1.16245711e+00 6.71381533e-01 4.97088991e-02 -9.18850303e-01
-1.61420798e+00 -2.72190928e-01 6.76300108e-01 2.29827613e-01
-4.98424530e-01 6.69792593e-01 -1.53369918e-01 -2.70162284e-01
3.99881691e-01 -7.81245768e-01 -3.85907553e-02 -9.24900826e-03
4.78519708e-01 4.59653586e-01 -5.08772023e-02 -9.08432543e-01
-4.32252973e-01 1.50626814e+00 2.50361472e-01 6.99108839e-02
8.19408178e-01 1.31502613e-01 3.58075708e-01 -3.71359497e-01
1.25738180e+00 -2.18776062e-01 -1.67966032e+00 -1.06702544e-01
1.04484558e-01 -5.95175326e-01 1.57705948e-01 -6.96596205e-01
-1.31686461e+00 9.30299461e-01 4.94842768e-01 -4.36457187e-01
6.28696144e-01 -1.61981016e-01 1.24750090e+00 5.28355777e-01
2.64075100e-01 -1.33081412e+00 -2.25846469e-01 8.06907713e-01
6.66464150e-01 -1.29038191e+00 -1.19224392e-01 -6.16623342e-01
-3.30365986e-01 1.45835602e+00 8.42352271e-01 -3.88037390e-03
7.86185145e-01 1.01062603e-01 4.82815862e-01 -2.08444253e-01
4.39780243e-02 -6.22301221e-01 9.15590763e-01 7.07435429e-01
-9.94003490e-02 -1.50828343e-02 -3.50555539e-01 -9.96126607e-02
-5.42110705e-04 9.33120921e-02 1.36755586e-01 7.93941438e-01
-6.76813722e-01 -9.70930338e-01 -5.33196032e-01 2.82588929e-01
2.55962074e-01 1.07796259e-01 -4.73605752e-01 9.82569456e-01
4.80636835e-01 5.72867274e-01 4.55443352e-01 -1.36388078e-01
5.35057724e-01 1.94274738e-01 3.73958796e-01 -1.92300111e-01
-1.09581068e-01 -1.19649895e-01 3.53902310e-01 -4.61195827e-01
-5.49303032e-02 -6.49759829e-01 -1.58695602e+00 -6.40848428e-02
-7.01988220e-01 -4.18906569e-01 1.22687042e+00 1.04418182e+00
2.62327909e-01 -4.72812504e-02 4.85547930e-01 -1.10361505e+00
-7.14792311e-02 -5.91973901e-01 -1.06607243e-01 2.62396812e-01
3.18669766e-01 -6.17997468e-01 -1.59832478e-01 -1.71497524e-01] | [7.984344959259033, -0.8149138689041138] |
dfd524cd-e718-4d7f-8574-3c778f1136e5 | robust-person-re-identification-through | 2009.07491 | null | https://arxiv.org/abs/2009.07491v1 | https://arxiv.org/pdf/2009.07491v1.pdf | Robust Person Re-Identification through Contextual Mutual Boosting | Person Re-Identification (Re-ID) has witnessed great advance, driven by the development of deep learning. However, modern person Re-ID is still challenged by background clutter, occlusion and large posture variation which are common in practice. Previous methods tackle these challenges by localizing pedestrians through external cues (e.g., pose estimation, human parsing) or attention mechanism, suffering from high computation cost and increased model complexity. In this paper, we propose the Contextual Mutual Boosting Network (CMBN). It localizes pedestrians and recalibrates features by effectively exploiting contextual information and statistical inference. Firstly, we construct two branches with a shared convolutional frontend to learn the foreground and background features respectively. By enabling interaction between these two branches, they boost the accuracy of the spatial localization mutually. Secondly, starting from a statistical perspective, we propose the Mask Generator that exploits the activation distribution of the transformation matrix for generating the static channel mask to the representations. The mask recalibrates the features to amplify the valuable characteristics and diminish the noise. Finally, we propose the Contextual-Detachment Strategy to optimize the two branches jointly and independently, which further enhances the localization precision. Experiments on the benchmarks demonstrate the superiority of the architecture compared the state-of-the-art. | ['Jane Shen', 'Zhikang Wang', 'Xinbo Gao', 'Lihuo He'] | 2020-09-16 | null | null | null | null | ['human-parsing'] | ['computer-vision'] | [-5.36579303e-02 -2.63958335e-01 2.64870644e-01 -3.25536907e-01
-5.35890460e-01 -3.48163992e-01 6.95044279e-01 4.63777669e-02
-6.08338296e-01 6.22980177e-01 3.40854317e-01 1.57559395e-01
1.34969845e-01 -6.79071665e-01 -7.47404933e-01 -8.94584715e-01
2.57618815e-01 1.34916613e-02 2.82924831e-01 -8.31571035e-03
9.50961560e-03 2.22368285e-01 -1.65862489e+00 1.81058362e-01
1.05406284e+00 1.00963497e+00 2.95074195e-01 3.58122349e-01
3.62232476e-02 4.88023967e-01 -5.61557770e-01 -5.31132698e-01
2.40888938e-01 -4.22785908e-01 -3.49296957e-01 6.43546209e-02
2.55807817e-01 -2.01287583e-01 -3.53269905e-01 1.09837198e+00
7.98258126e-01 1.88883737e-01 3.11712503e-01 -1.10484529e+00
-4.95121986e-01 4.42893237e-01 -8.89819026e-01 2.25898594e-01
3.69768530e-01 3.40527207e-01 6.95908010e-01 -1.02927256e+00
1.39169946e-01 1.41506112e+00 8.47640157e-01 5.18991172e-01
-1.20862436e+00 -7.00242937e-01 5.94861567e-01 4.77465719e-01
-1.69496334e+00 -3.94604415e-01 9.03338909e-01 -2.92726666e-01
3.73580217e-01 2.51894385e-01 5.72232485e-01 1.31560779e+00
-2.45338961e-01 8.01750302e-01 1.12663388e+00 -3.86138737e-01
-2.68518087e-02 2.20404044e-01 1.02194272e-01 5.88450849e-01
2.87637353e-01 1.14878438e-01 -4.27734077e-01 5.57150170e-02
6.56720281e-01 1.16713233e-01 -2.29876846e-01 -4.78871554e-01
-1.07605410e+00 4.20279324e-01 7.52337337e-01 2.33954966e-01
-3.24631274e-01 -1.18530486e-02 3.27911288e-01 -2.57162958e-01
1.59748495e-01 2.77580060e-02 -2.42809147e-01 1.94221944e-01
-6.65148616e-01 3.11115295e-01 3.05285275e-01 8.26020837e-01
7.19203174e-01 -2.31785834e-01 -5.22268951e-01 8.89580667e-01
4.28648621e-01 6.13029063e-01 4.69044596e-01 -1.96140975e-01
7.10565150e-01 6.56775177e-01 2.51395822e-01 -1.30229795e+00
-5.47030330e-01 -9.02399480e-01 -9.46928263e-01 -1.18343741e-01
5.60475945e-01 -2.37889022e-01 -8.54030192e-01 2.01271939e+00
5.23587346e-01 5.21682978e-01 -2.28617027e-01 1.24460554e+00
6.78130388e-01 2.98306793e-01 4.76396620e-01 1.71069205e-01
1.52150893e+00 -1.06066346e+00 -4.58819121e-01 -4.48551565e-01
2.72620380e-01 -5.91580808e-01 7.36628890e-01 9.89692956e-02
-7.76259422e-01 -1.04311728e+00 -1.05336392e+00 1.48540065e-01
-2.42569745e-01 6.64759517e-01 3.23994577e-01 7.50336170e-01
-7.72685826e-01 2.74750620e-01 -7.86364853e-01 -3.79661083e-01
4.47723061e-01 3.98536861e-01 -3.64008099e-01 8.33154172e-02
-1.19176877e+00 6.72589839e-01 2.65231371e-01 6.40825510e-01
-7.40346849e-01 -3.46998334e-01 -7.19179273e-01 6.04159832e-02
3.62659693e-01 -9.14417565e-01 7.09513068e-01 -1.06710386e+00
-1.54439592e+00 5.78173041e-01 -1.65377200e-01 -3.51722538e-01
8.22444737e-01 -4.35765356e-01 -3.26792091e-01 -1.96863458e-01
2.22973764e-01 5.90705574e-01 9.07321215e-01 -1.38174582e+00
-9.18187976e-01 -6.48513377e-01 -1.10347077e-01 3.63691986e-01
-3.47277045e-01 -1.05387352e-01 -8.12079966e-01 -7.80529559e-01
2.13047892e-01 -8.48746121e-01 -3.82988393e-01 -4.55210179e-01
-6.01986051e-01 -4.98474911e-02 5.81571579e-01 -9.70873892e-01
1.09085882e+00 -2.18521237e+00 3.54303837e-01 3.32985699e-01
1.52396217e-01 2.65313894e-01 7.60565549e-02 -5.72945550e-02
-6.05404451e-02 -1.34506613e-01 8.59439839e-04 -6.94167972e-01
-5.34500219e-02 -8.51516575e-02 -1.60588939e-02 6.03117943e-01
2.21326336e-01 8.54134023e-01 -8.46730828e-01 -4.69137371e-01
4.84288365e-01 6.58698499e-01 -5.32665431e-01 3.53262842e-01
2.08317593e-01 9.45820868e-01 -3.72782648e-01 4.96067196e-01
9.65746343e-01 -1.94933470e-02 1.05443463e-01 -5.50994754e-01
-2.20972493e-01 -7.59947002e-02 -1.45653844e+00 1.58260524e+00
-3.49261969e-01 8.75383019e-02 2.06446454e-01 -1.11757421e+00
9.61974621e-01 -5.63635863e-02 1.74182102e-01 -7.98812687e-01
3.68078768e-01 7.00609609e-02 -1.23699300e-01 -4.02537614e-01
1.94618359e-01 3.04364294e-01 -5.45132197e-02 1.34720802e-02
-4.48575765e-02 9.45534766e-01 -1.15519568e-01 -1.27185106e-01
6.45540118e-01 3.55322182e-01 1.98718578e-01 -1.55838281e-01
1.16207111e+00 -5.53700387e-01 6.85042381e-01 7.60145307e-01
-3.72992337e-01 4.25893843e-01 1.48026496e-01 -4.88086700e-01
-7.86095917e-01 -9.74993587e-01 3.79105881e-02 1.17461193e+00
6.44906461e-01 -3.43972921e-01 -1.01298344e+00 -7.97554553e-01
-1.78763624e-02 2.53990650e-01 -7.50061274e-01 -2.82519490e-01
-8.17153931e-01 -1.01687551e+00 4.46137130e-01 7.51340032e-01
1.00466621e+00 -7.22812295e-01 -5.00487089e-01 1.07578605e-01
-3.74333441e-01 -1.20217192e+00 -4.52472597e-01 1.06325205e-02
-3.81196231e-01 -7.67179370e-01 -9.68531668e-01 -7.47165978e-01
7.21549809e-01 4.82229263e-01 7.13398635e-01 1.04180492e-01
-3.14223468e-01 1.82551071e-01 -3.26149106e-01 -1.11640416e-01
2.04033986e-01 1.57852381e-01 7.81860352e-02 6.79131329e-01
4.69208270e-01 -6.26268983e-01 -9.28568184e-01 5.20882308e-01
-3.22793692e-01 1.62531957e-01 8.24035466e-01 9.19452190e-01
4.87453431e-01 9.00221840e-02 4.69840407e-01 -4.35611278e-01
2.59831309e-01 -4.64495748e-01 -4.90478635e-01 1.41547531e-01
-2.62268156e-01 7.02024903e-03 5.90277255e-01 -3.94513935e-01
-1.38225007e+00 3.75176877e-01 -1.49535805e-01 -1.64139420e-01
-4.89132911e-01 5.61962575e-02 -8.38064492e-01 2.88272761e-02
4.27021861e-01 3.46312732e-01 -3.45759153e-01 -7.16100574e-01
5.21189153e-01 6.31655157e-01 8.64224434e-01 -6.45458162e-01
9.15579736e-01 5.82097948e-01 -1.88779756e-01 -6.12373531e-01
-6.95629239e-01 -5.92905521e-01 -7.95081496e-01 -2.70546943e-01
1.07323229e+00 -1.01856983e+00 -7.84805059e-01 6.88455462e-01
-1.18927097e+00 9.63551849e-02 5.15915416e-02 3.46925735e-01
-1.14452755e-02 4.93938416e-01 -2.20664501e-01 -9.00016010e-01
-2.54030943e-01 -1.23423934e+00 1.23509872e+00 5.69536567e-01
1.55878484e-01 -5.57787716e-01 -1.66532472e-01 2.60922641e-01
3.37432742e-01 2.29505673e-01 4.74495888e-01 -5.89697063e-01
-6.45658433e-01 -1.85897693e-01 -5.37498295e-01 1.84499040e-01
1.26933351e-01 -5.22229195e-01 -1.18831480e+00 -2.50015557e-01
-3.33801806e-01 1.27893195e-01 8.90595734e-01 2.98674136e-01
1.20053136e+00 -2.70996511e-01 -6.74478650e-01 7.56133676e-01
1.15400350e+00 4.17550541e-02 5.38265407e-01 5.31851590e-01
9.77730274e-01 7.60535896e-01 3.04639786e-01 6.17718160e-01
6.30573153e-01 9.54891682e-01 3.69711727e-01 -3.17030966e-01
-1.95340484e-01 -4.58770335e-01 4.74532843e-02 4.33918267e-01
-3.18204552e-01 3.63457315e-02 -7.14400232e-01 3.63882840e-01
-2.08099413e+00 -8.32075298e-01 6.43994194e-03 2.21126390e+00
6.19876683e-01 1.25360504e-01 3.46798807e-01 1.13208495e-01
1.09924245e+00 3.11221033e-02 -3.96032453e-01 3.55974406e-01
-1.51803359e-01 -3.44750322e-02 5.22814929e-01 4.00696516e-01
-1.46205533e+00 9.25233603e-01 4.54290724e+00 7.97393858e-01
-9.85141814e-01 1.35805935e-01 6.77012503e-01 1.33108407e-01
1.70218974e-01 -1.76992223e-01 -1.11160016e+00 7.77633846e-01
3.79990369e-01 3.73134047e-01 3.67434293e-01 9.07186925e-01
3.59822027e-02 7.76133835e-02 -9.15307045e-01 1.27459264e+00
1.18007749e-01 -8.47662270e-01 -5.55319376e-02 -5.48457019e-02
5.25721371e-01 -4.81753886e-01 1.39423311e-01 3.73711050e-01
5.03280722e-02 -8.86975348e-01 9.89993453e-01 6.60651624e-01
3.72100443e-01 -9.97311592e-01 1.00985360e+00 3.25836480e-01
-1.59422696e+00 -3.29684407e-01 -2.28174344e-01 8.66746530e-02
9.01328549e-02 4.00596023e-01 -4.75122422e-01 6.79028034e-01
9.70167160e-01 5.16436994e-01 -9.20104027e-01 1.06714213e+00
-3.01023751e-01 3.09612513e-01 -1.88408226e-01 1.00929193e-01
-2.62570772e-02 -1.84756801e-01 6.31757557e-01 1.44885302e+00
4.78992127e-02 -1.36457786e-01 4.16006833e-01 9.44746912e-01
1.41306981e-01 1.25065893e-01 -1.23551197e-01 6.28768682e-01
4.90186304e-01 1.39628661e+00 -6.08775556e-01 -1.76784590e-01
-3.55009377e-01 1.27415812e+00 3.89920026e-01 4.33429152e-01
-1.17489243e+00 -3.25400412e-01 4.75201935e-01 5.74037135e-02
4.53860164e-01 -1.33146346e-01 -2.55389422e-01 -1.08836305e+00
3.09224933e-01 -8.15953135e-01 2.41689757e-01 -2.38301858e-01
-1.39075017e+00 5.80339015e-01 -1.23891257e-01 -1.05546296e+00
4.70044091e-02 -4.38482493e-01 -4.87663954e-01 1.11340094e+00
-1.37723374e+00 -1.58416128e+00 -8.67535710e-01 7.11067140e-01
3.22703809e-01 -1.27216846e-01 4.34064835e-01 7.20867455e-01
-1.02928054e+00 1.00298476e+00 -2.85553277e-01 3.63521546e-01
7.38564432e-01 -1.16506350e+00 3.84738803e-01 1.02795613e+00
-1.16337068e-01 8.96636784e-01 5.00600517e-01 -6.27442181e-01
-1.12615597e+00 -1.20966160e+00 5.16583800e-01 -3.07863176e-01
3.20570379e-01 -6.56920910e-01 -6.54201746e-01 5.09618878e-01
-1.00287817e-01 7.85271227e-02 4.45850909e-01 6.78546652e-02
-3.44328165e-01 -4.12849486e-01 -8.96951616e-01 6.83867574e-01
1.20717561e+00 -2.72819817e-01 -2.82101780e-01 1.15750730e-01
4.85707492e-01 -3.43215287e-01 -2.62467980e-01 3.38798732e-01
6.70361817e-01 -1.04244924e+00 1.30946314e+00 -3.75955045e-01
-3.98245007e-02 -6.34830475e-01 -1.03017747e-01 -1.08491766e+00
-6.42958283e-01 -4.45152581e-01 1.36750517e-02 1.69605255e+00
8.16518813e-02 -6.47478223e-01 6.95959628e-01 4.56462294e-01
1.21433601e-01 -6.28762424e-01 -9.73805606e-01 -6.03582323e-01
-3.40113401e-01 -1.76324889e-01 8.73708665e-01 6.59627259e-01
-2.75481641e-01 3.85481387e-01 -6.94918752e-01 4.18224126e-01
8.20128202e-01 -9.83330607e-02 1.06721354e+00 -1.10026157e+00
-4.13991988e-01 -4.66765285e-01 -5.38384080e-01 -1.38124013e+00
-9.46213305e-03 -6.34623826e-01 1.01289362e-01 -1.19957197e+00
3.45474184e-01 -6.00673139e-01 -3.63433838e-01 2.28568420e-01
-7.05595136e-01 6.95266649e-02 2.20806420e-01 1.67708129e-01
-7.30323434e-01 6.47081017e-01 9.00407910e-01 -1.13116920e-01
-3.05042118e-01 1.07764609e-01 -8.35288107e-01 8.35934401e-01
6.01305127e-01 -7.89469928e-02 -1.57436684e-01 -4.41086411e-01
-1.03300288e-01 -4.41104501e-01 8.98225069e-01 -1.29670048e+00
4.40362930e-01 2.33199388e-01 9.04017031e-01 -5.86639285e-01
2.70016521e-01 -8.20687413e-01 -7.10671768e-02 2.70800799e-01
-1.39797568e-01 -1.85147841e-02 3.76095027e-02 8.37721169e-01
-1.08394079e-01 1.65491514e-02 8.72139335e-01 -1.59518018e-01
-7.35807121e-01 2.17371076e-01 1.55454343e-02 -2.24723980e-01
9.69268620e-01 -1.54853180e-01 5.55628017e-02 -8.83138776e-02
-5.95540166e-01 2.09730789e-01 1.49335518e-01 3.76056612e-01
4.41234291e-01 -1.49721467e+00 -6.65352225e-01 4.83960897e-01
-5.66197112e-02 -9.42433551e-02 4.71033841e-01 1.01882064e+00
-1.12600729e-01 3.24835360e-01 -2.13309765e-01 -8.20043802e-01
-1.11124229e+00 5.83928406e-01 4.60376292e-01 -3.00251871e-01
-6.34697318e-01 9.43415701e-01 5.70085764e-01 -2.24088624e-01
3.99352878e-01 -6.98701516e-02 -5.74863672e-01 5.26021980e-03
6.96597099e-01 5.75951815e-01 -6.97226152e-02 -9.69360948e-01
-6.18932486e-01 7.11528122e-01 -1.08460911e-01 1.32767914e-03
1.04815841e+00 -4.99287099e-01 7.12963566e-02 -1.03973694e-01
9.57894027e-01 2.17950627e-01 -1.37477493e+00 -3.25000077e-01
2.35393606e-02 -5.61608553e-01 -9.30337980e-02 -6.73861802e-01
-1.01994109e+00 8.53430331e-01 1.00741267e+00 -1.04669325e-01
1.09558105e+00 -2.26736382e-01 8.11791301e-01 7.67595833e-03
3.76597762e-01 -1.02152169e+00 8.50040242e-02 1.92499623e-01
6.08696759e-01 -1.15711164e+00 -2.28851825e-01 -4.50760633e-01
-5.84186852e-01 8.62539530e-01 8.43678892e-01 -5.93374521e-02
4.67184335e-01 1.17751673e-01 -1.32526174e-01 2.41874009e-01
-1.19385384e-02 -4.47210729e-01 4.31702077e-01 8.38653684e-01
1.30425066e-01 1.11999504e-01 -2.12188661e-01 1.40339613e+00
-6.29200637e-02 -3.25077742e-01 -3.00408959e-01 6.08194530e-01
-3.46185833e-01 -1.03597224e+00 -7.27293730e-01 -9.44945961e-02
-2.50037313e-01 -7.16668786e-03 -2.71074235e-01 5.72398365e-01
8.74730527e-01 9.45049226e-01 -1.35275111e-01 -5.34274936e-01
5.07266879e-01 -1.17053442e-01 3.51825029e-01 -1.84956744e-01
-5.52283347e-01 2.59376615e-01 -1.14149645e-01 -5.09118855e-01
-3.16721410e-01 -6.95684016e-01 -8.28398943e-01 -3.95763591e-02
-3.71637106e-01 -8.69514607e-03 5.31219959e-01 9.53826070e-01
4.39894676e-01 6.49505377e-01 5.63383341e-01 -1.09832847e+00
-6.11172378e-01 -1.02148509e+00 -1.66292608e-01 5.24605870e-01
2.36174762e-01 -1.02129257e+00 -1.89681977e-01 1.98458806e-02] | [14.654603958129883, 0.8837879300117493] |
72bb1e72-ed6e-4231-add5-930880ff9312 | directed-acyclic-graph-structure-learning | 2211.17029 | null | https://arxiv.org/abs/2211.17029v1 | https://arxiv.org/pdf/2211.17029v1.pdf | Directed Acyclic Graph Structure Learning from Dynamic Graphs | Estimating the structure of directed acyclic graphs (DAGs) of features (variables) plays a vital role in revealing the latent data generation process and providing causal insights in various applications. Although there have been many studies on structure learning with various types of data, the structure learning on the dynamic graph has not been explored yet, and thus we study the learning problem of node feature generation mechanism on such ubiquitous dynamic graph data. In a dynamic graph, we propose to simultaneously estimate contemporaneous relationships and time-lagged interaction relationships between the node features. These two kinds of relationships form a DAG, which could effectively characterize the feature generation process in a concise way. To learn such a DAG, we cast the learning problem as a continuous score-based optimization problem, which consists of a differentiable score function to measure the validity of the learned DAGs and a smooth acyclicity constraint to ensure the acyclicity of the learned DAGs. These two components are translated into an unconstraint augmented Lagrangian objective which could be minimized by mature continuous optimization techniques. The resulting algorithm, named GraphNOTEARS, outperforms baselines on simulated data across a wide range of settings that may encounter in real-world applications. We also apply the proposed approach on two dynamic graphs constructed from the real-world Yelp dataset, demonstrating our method could learn the connections between node features, which conforms with the domain knowledge. | ['Chuan Shi', 'Xiao Wang', 'Shuyang Zhang', 'Shaohua Fan'] | 2022-11-30 | null | null | null | null | ['graph-structure-learning'] | ['graphs'] | [ 1.17864758e-01 1.46006659e-01 -4.76381570e-01 -4.13641453e-01
-2.52956420e-01 -5.37867367e-01 6.57450855e-01 2.03802615e-01
1.58106789e-01 8.05388510e-01 2.18910307e-01 -3.95491153e-01
-6.71777844e-01 -8.91325474e-01 -7.87455916e-01 -8.10524940e-01
-8.72738123e-01 3.29101145e-01 4.20128703e-02 2.69516017e-02
7.84802735e-02 2.57887512e-01 -1.11748362e+00 -2.45464757e-01
8.62625062e-01 7.89230585e-01 2.30248600e-01 5.16726613e-01
-1.10690832e-01 6.94955409e-01 -2.47432202e-01 -2.57244378e-01
2.07795441e-01 -2.94375122e-01 -7.56631553e-01 3.79962325e-01
8.02764744e-02 7.76461512e-03 -4.36495543e-01 1.04750419e+00
1.03775881e-01 2.32505426e-02 6.92509532e-01 -1.71542251e+00
-5.11791468e-01 7.84756064e-01 -8.44714046e-01 3.27109605e-01
4.23324794e-01 -7.56955966e-02 1.68208611e+00 -7.10049808e-01
6.94038570e-01 1.51461816e+00 3.44459772e-01 -3.69386040e-02
-1.28716266e+00 -7.72724509e-01 6.65894270e-01 1.77416399e-01
-1.12773728e+00 -5.21345995e-02 1.14218199e+00 -5.52011847e-01
5.48757851e-01 1.19646035e-01 8.11924398e-01 1.21050668e+00
6.25760138e-01 7.64071107e-01 9.51284826e-01 -4.78223301e-02
2.15310957e-02 -6.16377518e-02 3.76753300e-01 1.05228651e+00
4.63604629e-01 2.20634326e-01 -5.27597547e-01 -3.91815126e-01
6.25205994e-01 1.37407213e-01 -8.64272863e-02 -5.40335000e-01
-1.05227184e+00 9.21373427e-01 4.85867590e-01 1.78816449e-02
-2.85839468e-01 1.58445373e-01 2.80535609e-01 5.87619185e-01
4.66506690e-01 2.48767555e-01 -6.27895594e-01 2.41188139e-01
-2.11877227e-01 2.08457679e-01 8.81835401e-01 1.01273870e+00
9.41204071e-01 4.80257012e-02 -1.54514626e-01 3.73448431e-01
6.24148309e-01 4.19897318e-01 1.51778966e-01 -3.31508130e-01
8.00581634e-01 1.03485012e+00 -1.37272328e-01 -1.57495880e+00
-5.31531990e-01 -5.60570061e-01 -1.04629529e+00 -3.91807854e-01
4.29716110e-01 -4.06830192e-01 -5.91491938e-01 2.03362203e+00
6.16355121e-01 6.85351312e-01 -1.77843511e-01 5.76250553e-01
4.82234120e-01 7.66535282e-01 -1.18194580e-01 -5.94550967e-01
1.04067683e+00 -5.72566152e-01 -7.12279499e-01 -1.63919851e-01
5.04842401e-01 -2.67947674e-01 9.31334317e-01 1.58084363e-01
-5.74745953e-01 -2.84507602e-01 -9.54433799e-01 2.78715402e-01
-2.10563205e-02 -1.86103344e-01 1.15730119e+00 2.82553762e-01
-8.85743618e-01 5.89236379e-01 -8.83855224e-01 -1.70377672e-01
1.82924777e-01 5.37868381e-01 -4.06924248e-01 8.34661871e-02
-1.43179715e+00 1.77650318e-01 5.56878746e-01 5.01373708e-01
-1.05228221e+00 -6.18811786e-01 -8.54448497e-01 1.71718106e-01
1.03017330e+00 -6.68537438e-01 7.80777812e-01 -8.03903222e-01
-1.07784891e+00 4.03810710e-01 -7.24324286e-02 -1.97608441e-01
3.48519355e-01 6.97880313e-02 -5.90819895e-01 -3.58803153e-01
1.95882604e-01 -1.09718226e-01 1.00694275e+00 -1.17321825e+00
-7.56626725e-01 -3.99251610e-01 2.37101942e-01 4.90181483e-02
-4.84862447e-01 -3.45855981e-01 -4.27166879e-01 -5.66339552e-01
7.53867701e-02 -9.23280180e-01 -3.31885040e-01 -3.23084950e-01
-7.29018748e-01 -4.76536542e-01 9.39819753e-01 -4.93176877e-01
1.57692313e+00 -1.92840016e+00 2.55858839e-01 5.54358959e-01
4.78078932e-01 -3.05865467e-01 -2.85030335e-01 6.44714892e-01
-1.44320488e-01 2.00294197e-01 -2.62864411e-01 -7.79750645e-02
-1.64001435e-01 2.32284784e-01 -3.43916446e-01 5.46345651e-01
2.91598171e-01 8.98697376e-01 -1.24955130e+00 -4.55704689e-01
-2.20103785e-01 1.95959322e-02 -3.59115332e-01 4.33596313e-01
-3.18953067e-01 3.49926651e-01 -1.01143801e+00 6.14932060e-01
4.41042632e-01 -5.34397840e-01 6.61777794e-01 1.36264041e-01
1.86014235e-01 9.84232780e-03 -1.38661468e+00 1.42605460e+00
-1.50791675e-01 3.78557295e-01 -5.83750010e-02 -1.37060571e+00
1.06808066e+00 9.14893672e-02 6.86468482e-01 -4.45207149e-01
-1.62799001e-01 -7.77005106e-02 1.66203007e-01 -6.00674689e-01
-3.69903911e-03 2.05153786e-02 -3.13943923e-01 3.74969691e-01
-7.95929730e-02 2.41018295e-01 3.19447637e-01 4.91231143e-01
1.35996163e+00 -6.76764846e-02 4.01255280e-01 -4.72748458e-01
6.67503417e-01 -2.10782856e-01 8.88607264e-01 6.22010469e-01
2.24896401e-01 -1.59302935e-01 1.23063779e+00 -3.38343054e-01
-6.92335069e-01 -8.59531224e-01 1.42330766e-01 8.49188566e-01
1.59178227e-01 -5.65678179e-01 -1.54744029e-01 -1.03251159e+00
2.72952884e-01 2.20519096e-01 -7.76999295e-01 -2.90658295e-01
-5.13338149e-01 -9.23156977e-01 4.47707660e-02 3.31355095e-01
1.78859130e-01 -9.09923017e-01 5.98979089e-03 2.43711933e-01
1.87005922e-01 -9.10246789e-01 -6.26349628e-01 2.09855765e-01
-8.82630169e-01 -1.55547297e+00 9.99004543e-02 -8.49114478e-01
7.65656412e-01 1.77209139e-01 1.09801269e+00 1.43269897e-01
-2.09710643e-01 2.21797839e-01 -1.64599791e-01 -1.53634623e-01
-1.59592181e-01 2.24459976e-01 3.58210355e-02 3.58030379e-01
-1.33908421e-01 -9.09168482e-01 -3.76162589e-01 1.87889203e-01
-8.31388474e-01 6.62023202e-02 7.07020581e-01 1.13894606e+00
5.63429415e-01 4.90499258e-01 7.99060047e-01 -1.36515009e+00
7.96701908e-01 -1.04196548e+00 -9.78020012e-01 5.15405893e-01
-9.59233284e-01 5.69441617e-01 7.40822732e-01 -4.39191997e-01
-9.37996030e-01 8.37692544e-02 4.25809115e-01 -2.66960472e-01
4.14231986e-01 1.16384983e+00 -4.53996718e-01 3.24777335e-01
1.51088595e-01 1.43736765e-01 -1.89187601e-01 -2.49594167e-01
3.72220099e-01 9.25679877e-02 2.69639611e-01 -8.31047595e-01
1.17759955e+00 2.42436185e-01 6.13723695e-01 -5.55275083e-01
-7.82680213e-01 -3.81657928e-01 -5.53842664e-01 -4.69351001e-02
3.96218330e-01 -6.84471548e-01 -9.39712346e-01 1.98776767e-01
-8.80399406e-01 -1.24298871e-01 1.54390216e-01 3.20875406e-01
-2.94492453e-01 3.32191557e-01 -3.17210287e-01 -7.54425049e-01
-1.17541812e-01 -8.40034604e-01 8.30585837e-01 2.99984943e-02
1.66323349e-01 -1.51637113e+00 3.97042900e-01 -2.39849851e-01
-2.02857912e-01 7.22817183e-01 1.33791554e+00 -5.53355038e-01
-7.35088766e-01 -1.17274195e-01 -1.61082879e-01 -1.19686879e-01
5.37457407e-01 2.46819183e-01 -3.84977818e-01 -5.87148190e-01
-9.00914147e-02 -5.59028983e-02 7.79792011e-01 2.54057407e-01
1.12218833e+00 -6.72671318e-01 -6.96888685e-01 6.26266778e-01
1.28106260e+00 2.61736512e-01 2.06150915e-02 -2.42744442e-02
8.86997521e-01 6.50918961e-01 6.62387252e-01 5.61907411e-01
5.06226778e-01 5.16213715e-01 6.15678430e-01 -9.39077362e-02
2.11004302e-01 -7.08658814e-01 4.35331702e-01 9.05005753e-01
9.88629386e-02 -4.10482883e-01 -8.97887766e-01 3.35118264e-01
-2.25900841e+00 -8.11207473e-01 -4.27588403e-01 2.00316906e+00
7.26545274e-01 1.85133472e-01 1.20342351e-01 -1.33799121e-01
7.52043128e-01 4.14441347e-01 -7.72466719e-01 -1.04877921e-02
1.01318873e-01 -1.76713735e-01 2.18754798e-01 4.12661105e-01
-1.07759416e+00 7.13398874e-01 6.08426857e+00 5.62012136e-01
-1.15292442e+00 -2.43758291e-01 5.93125522e-01 3.49774450e-01
-6.87880635e-01 5.03033698e-01 -8.14906716e-01 4.39965308e-01
7.26833284e-01 -6.87811136e-01 3.62535685e-01 7.50040293e-01
4.42360133e-01 2.02776194e-01 -1.33606577e+00 7.00934172e-01
-3.57375413e-01 -1.06297338e+00 6.82467222e-02 3.13433498e-01
8.39844167e-01 -3.53196740e-01 5.13733327e-02 3.34354252e-01
6.54813826e-01 -9.55051243e-01 3.10475856e-01 5.83987415e-01
6.10573292e-01 -6.70896113e-01 3.90829742e-01 3.89187336e-01
-1.62768340e+00 -3.91242683e-01 -3.22572529e-01 -1.98326528e-01
8.41112435e-02 8.79428506e-01 -1.04686296e+00 1.00653958e+00
2.15410084e-01 1.23363411e+00 -6.66975439e-01 7.94596195e-01
-3.93379480e-01 7.95179069e-01 -8.95987973e-02 -4.29091416e-02
2.73353308e-01 -4.96690005e-01 6.45102561e-01 8.72544527e-01
1.25465274e-01 -1.18831426e-01 6.03203654e-01 7.75159895e-01
-1.90968797e-01 2.27272525e-01 -6.93283617e-01 -2.89158463e-01
4.87012655e-01 1.37514496e+00 -7.31169522e-01 6.01754896e-02
-4.62394238e-01 5.40515602e-01 6.39603913e-01 4.43967462e-01
-7.62267113e-01 -9.94826555e-02 5.65276444e-01 1.56020209e-01
-5.01906313e-02 -2.36728176e-01 1.14069797e-01 -1.29176092e+00
1.61810145e-01 -7.57210791e-01 7.91587114e-01 -6.48391619e-02
-1.52986681e+00 4.04396743e-01 1.69622302e-01 -9.65869367e-01
-4.63211656e-01 -3.28738958e-01 -9.76699889e-01 6.49811745e-01
-1.40003920e+00 -9.63647485e-01 -3.01399529e-01 8.68045568e-01
3.56021225e-01 -1.28750056e-01 4.08800304e-01 1.00589477e-01
-9.82010841e-01 3.99355143e-01 1.34044349e-01 8.55625793e-02
4.32744026e-01 -1.35398221e+00 3.46884459e-01 1.01203120e+00
3.19669455e-01 6.62403703e-01 5.75402975e-01 -1.00135052e+00
-1.87092149e+00 -1.17098141e+00 6.97359204e-01 -1.86954767e-01
1.19146955e+00 -6.38051867e-01 -8.59381199e-01 8.70788813e-01
-1.52280614e-01 2.67773062e-01 3.22983772e-01 4.85098273e-01
-2.48392701e-01 -3.31863284e-01 -6.51034713e-01 3.32180738e-01
1.44853342e+00 -1.76086619e-01 -3.36179018e-01 4.13741767e-01
8.88307750e-01 -1.66773975e-01 -8.16017509e-01 4.69291329e-01
4.69014436e-01 -3.78649861e-01 7.71449566e-01 -1.13834345e+00
4.84784365e-01 -2.17827946e-01 2.17514381e-01 -1.38909423e+00
-6.03057861e-01 -7.90451467e-01 -4.91700768e-01 1.44645977e+00
4.92000908e-01 -8.10702026e-01 6.10847652e-01 3.92529726e-01
6.71960711e-02 -1.04472482e+00 -7.53958881e-01 -7.53014266e-01
-3.52221817e-01 -3.36519629e-02 6.94522142e-01 1.18390667e+00
-7.24509135e-02 6.62473619e-01 -5.77612400e-01 3.08908373e-01
6.76987290e-01 4.28646803e-01 8.44166398e-01 -1.52952862e+00
-4.73543435e-01 -2.76201844e-01 -3.59225512e-01 -7.86799788e-01
6.11769021e-01 -1.26314497e+00 -2.06808835e-01 -1.42205477e+00
4.91804868e-01 -7.13751733e-01 -3.17629337e-01 4.19104218e-01
-6.30926430e-01 -8.15110505e-01 -1.65846571e-01 2.08056197e-01
-4.35708284e-01 7.50701249e-01 1.28128040e+00 -1.79189980e-01
-3.76557827e-01 2.71288216e-01 -7.54255593e-01 4.98912245e-01
4.29186225e-01 -6.23117268e-01 -1.05342746e+00 -7.97964707e-02
4.62884218e-01 4.72416759e-01 2.56941468e-01 -2.46279433e-01
2.70934284e-01 -6.68665946e-01 6.86648265e-02 -4.74728823e-01
-2.17262402e-01 -8.74182045e-01 4.65329766e-01 4.61913556e-01
-3.58932763e-01 3.56834680e-01 -2.90894508e-01 1.09077466e+00
-2.49796033e-01 1.15639724e-01 1.02843858e-01 1.39889941e-01
-7.81095386e-01 8.77717972e-01 2.20025748e-01 1.81018740e-01
1.09858906e+00 1.20405637e-01 -1.86063692e-01 -2.89851397e-01
-5.59855521e-01 7.90412724e-01 -4.26885709e-02 6.53833866e-01
5.10449767e-01 -1.42346013e+00 -8.11900198e-01 2.26992413e-01
1.92061509e-03 2.06510685e-02 -3.38807069e-02 7.33940184e-01
6.31958246e-02 2.90615857e-01 1.23263158e-01 -5.78094006e-01
-1.06057739e+00 7.63727665e-01 -5.82484948e-03 -9.43433821e-01
-6.19938314e-01 5.67183316e-01 3.00216496e-01 -2.58922279e-01
1.65604293e-01 -2.28596702e-01 -3.25635046e-01 2.81809568e-01
-7.44686462e-03 2.89055467e-01 -2.60071397e-01 -2.02198654e-01
-4.08911645e-01 3.30812603e-01 -1.08596884e-01 2.28697166e-01
1.49523425e+00 -2.73827821e-01 -1.93852812e-01 5.37748039e-01
1.24735808e+00 -4.03727740e-02 -1.48533571e+00 -4.28032517e-01
4.47136164e-01 -5.52123308e-01 -1.80714294e-01 -2.46319249e-01
-1.25655365e+00 4.05786186e-01 5.37366718e-02 4.69831198e-01
1.02890348e+00 1.24580532e-01 4.68356699e-01 4.02697355e-01
3.83334130e-01 -6.11646593e-01 4.25638646e-01 4.62255985e-01
7.96145797e-01 -1.20480716e+00 4.64641191e-02 -7.07040131e-01
-4.71144795e-01 1.12400234e+00 6.45779788e-01 -1.95634902e-01
9.94970262e-01 1.49005964e-01 -5.37942708e-01 -4.06235754e-01
-1.13622558e+00 -1.01378617e-04 5.25343776e-01 3.26519161e-01
2.31852382e-01 2.06690535e-01 -5.39495707e-01 7.02852368e-01
-2.40040526e-01 -4.11918610e-01 3.32531124e-01 5.88612616e-01
-1.58046395e-01 -1.17315805e+00 -1.38960853e-02 6.54620230e-01
-2.37057686e-01 9.19665396e-02 -3.21402252e-01 8.45813334e-01
-1.01445101e-01 8.43631566e-01 -1.42198280e-01 -5.36067367e-01
3.20074409e-01 -1.94985300e-01 9.75061059e-02 -6.79562390e-01
-2.17013463e-01 -2.26707514e-02 1.33588225e-01 -5.60103536e-01
-3.31513852e-01 -8.73029768e-01 -1.02410626e+00 -2.14852959e-01
-5.34612179e-01 1.95461094e-01 2.59165794e-01 9.35148060e-01
1.50365874e-01 6.13940179e-01 1.31442690e+00 -2.55815864e-01
-6.45264685e-01 -6.71174943e-01 -8.61657739e-01 5.12950301e-01
2.54263520e-01 -8.72321308e-01 -4.56613481e-01 -9.45099145e-02] | [7.382645606994629, 5.827086448669434] |
9861dbb2-a720-4b28-9b14-496b6baba783 | fast-robust-tensor-principal-component | 2108.10448 | null | https://arxiv.org/abs/2108.10448v1 | https://arxiv.org/pdf/2108.10448v1.pdf | Fast Robust Tensor Principal Component Analysis via Fiber CUR Decomposition | We study the problem of tensor robust principal component analysis (TRPCA), which aims to separate an underlying low-multilinear-rank tensor and a sparse outlier tensor from their sum. In this work, we propose a fast non-convex algorithm, coined Robust Tensor CUR (RTCUR), for large-scale TRPCA problems. RTCUR considers a framework of alternating projections and utilizes the recently developed tensor Fiber CUR decomposition to dramatically lower the computational complexity. The performance advantage of RTCUR is empirically verified against the state-of-the-arts on the synthetic datasets and is further demonstrated on the real-world application such as color video background subtraction. | ['Deanna Needell', 'Longxiu Huang', 'Zehan Chao', 'HanQin Cai'] | 2021-08-23 | null | null | null | null | ['video-background-subtraction'] | ['computer-vision'] | [ 7.25931898e-02 -7.92340934e-01 1.36217102e-01 1.45307913e-01
-1.00641739e+00 -4.07102793e-01 2.51257300e-01 -6.62620962e-01
-7.32980222e-02 2.14167744e-01 2.64507502e-01 -2.52718121e-01
-4.82089609e-01 1.16837852e-01 -5.76660335e-01 -1.04109740e+00
-4.13166523e-01 1.57963887e-01 -1.87810391e-01 -2.64080055e-02
1.20914385e-01 3.14720660e-01 -1.11033809e+00 2.59007871e-01
8.53802919e-01 1.05587077e+00 -2.80108571e-01 6.43295884e-01
3.30424875e-01 9.35679734e-01 7.21947402e-02 -4.44689214e-01
6.25340402e-01 5.42218387e-02 -3.99525285e-01 5.75197995e-01
6.65722489e-01 -1.66213572e-01 -3.21564853e-01 1.18838048e+00
1.27193972e-01 2.02006936e-01 3.27576876e-01 -1.39341831e+00
-4.65822786e-01 3.74744087e-01 -1.17531633e+00 1.86877355e-01
2.22810298e-01 -2.02360749e-01 1.09368670e+00 -1.39837897e+00
4.65758950e-01 1.28030121e+00 6.56001627e-01 -1.38171734e-02
-1.43074238e+00 -4.48272049e-01 2.84002483e-01 2.59158820e-01
-1.40830934e+00 -3.37828100e-01 1.19328535e+00 -8.04722667e-01
3.58728647e-01 6.35968685e-01 5.45965195e-01 8.94913614e-01
-4.14301595e-03 1.01692808e+00 1.71534097e+00 -6.52092919e-02
1.54359177e-01 -5.67059278e-01 5.44604897e-01 8.36612582e-01
4.75563109e-01 -6.66183159e-02 -7.12443173e-01 -8.23807657e-01
4.75091070e-01 1.54093727e-01 -4.23574239e-01 -6.71827614e-01
-1.94955695e+00 6.04819834e-01 6.59084767e-02 -4.07466553e-02
-7.26982653e-01 2.25151256e-01 3.81383449e-01 -3.94038297e-02
5.69988966e-01 1.33603051e-01 -4.82408464e-01 -1.48992643e-01
-8.07189226e-01 1.63978219e-01 5.14571011e-01 9.28575695e-01
6.88138127e-01 4.97075886e-01 1.00363558e-03 8.24676096e-01
4.39561367e-01 1.00997150e+00 3.30288321e-01 -1.15764713e+00
5.98985970e-01 5.09027600e-01 9.85958055e-02 -1.40731835e+00
-2.64976442e-01 -2.24284261e-01 -9.91140246e-01 -4.28863466e-02
4.41650808e-01 1.97420549e-02 -6.90700650e-01 1.32898903e+00
5.74238539e-01 6.69596314e-01 1.13641694e-01 1.35218096e+00
2.42644981e-01 4.92334872e-01 -4.32389975e-01 -5.69273829e-01
1.38365638e+00 -9.32872117e-01 -7.42921174e-01 1.83724351e-02
3.52944046e-01 -1.09357190e+00 8.63737583e-01 8.40627253e-01
-5.44405341e-01 -1.91164896e-01 -9.60890174e-01 -4.32716385e-02
1.27148300e-01 5.88585675e-01 1.08030260e+00 6.51324570e-01
-6.33310139e-01 4.62167323e-01 -1.25655222e+00 -8.55714604e-02
1.85119018e-01 1.25230402e-01 -6.21317565e-01 -4.16428894e-01
-4.25743490e-01 2.60443151e-01 -2.35885590e-01 8.32238793e-01
-9.66486871e-01 -6.16235077e-01 -5.29982567e-01 -5.50567627e-01
6.79495096e-01 -5.93680978e-01 6.33059382e-01 -7.57085502e-01
-1.71206295e+00 4.87327129e-01 -3.41052592e-01 2.08426686e-03
3.68190616e-01 -6.77017093e-01 -3.29945952e-01 3.46712500e-01
1.72714412e-01 -2.31448531e-01 1.51733625e+00 -1.25663829e+00
-1.45312995e-01 -5.83922029e-01 -1.88241556e-01 8.86916928e-03
-3.27551305e-01 3.07416350e-01 -6.93426609e-01 -1.08017612e+00
8.29910338e-01 -1.68150878e+00 -5.77498257e-01 -2.98635960e-01
-6.62236631e-01 1.38389200e-01 9.54473138e-01 -1.10216868e+00
8.82572532e-01 -2.01085949e+00 9.35509980e-01 4.18509811e-01
4.39931244e-01 -2.02859402e-01 -1.82469815e-01 1.69367999e-01
-4.89254594e-01 -2.31646106e-01 -3.07065338e-01 -3.59485716e-01
-1.60602823e-01 2.20875636e-01 -6.11185968e-01 9.73078191e-01
-9.56291892e-03 4.43885416e-01 -9.20540988e-01 -3.05914402e-01
2.47115269e-02 4.88284647e-01 -3.55967402e-01 9.03242007e-02
1.47579595e-01 3.07666928e-01 -6.92435622e-01 1.00958705e+00
1.01650918e+00 -1.59071103e-01 1.14434391e-01 -9.03607786e-01
-1.61207959e-01 -3.22662294e-01 -1.46293819e+00 1.80285954e+00
2.06698567e-01 4.25694913e-01 4.69346195e-01 -7.61653483e-01
4.13089961e-01 8.28370824e-02 1.02690363e+00 -2.18273044e-01
1.44884408e-01 2.91614383e-01 -1.49535820e-01 -3.99338901e-01
7.11447716e-01 2.26877347e-01 1.48326978e-01 4.43327755e-01
-2.83748060e-01 2.66353995e-01 4.46471870e-01 5.69704831e-01
1.05359137e+00 5.14341295e-01 -7.66876489e-02 -5.78482926e-01
6.28858507e-01 2.65608821e-02 8.28402102e-01 1.89396903e-01
-1.35540321e-01 5.92316151e-01 5.10883689e-01 -3.17410052e-01
-7.54918516e-01 -9.80157018e-01 -6.89737592e-03 1.05972540e+00
-1.65781334e-01 -7.03566611e-01 -4.61337537e-01 -7.19629109e-01
-1.27275474e-02 -4.49200347e-02 -3.67263407e-01 4.22294468e-01
-7.26352990e-01 -1.38379180e+00 7.75512680e-02 2.12120607e-01
1.11094795e-01 -5.15160076e-02 1.42430579e-02 -1.04287669e-01
-5.12581646e-01 -1.63941646e+00 -6.22137666e-01 -3.19806725e-01
-1.11021197e+00 -1.25616360e+00 -7.97278523e-01 -1.04245551e-01
8.94010842e-01 1.03689480e+00 6.68327034e-01 -3.51905584e-01
-6.04617409e-02 7.94889092e-01 -4.92088258e-01 7.86089674e-02
2.42514238e-01 -5.35152435e-01 5.64634621e-01 8.55974734e-01
-2.45343447e-02 -6.10784888e-01 -4.14320260e-01 5.86536646e-01
-8.36316884e-01 1.48889825e-01 6.65213883e-01 8.58910978e-01
9.63573039e-01 5.28893434e-02 -3.40193093e-01 -6.00752950e-01
6.10370040e-01 -2.93200314e-01 -8.72710049e-01 2.16674879e-01
-4.16045636e-01 1.50016740e-01 4.56570089e-01 -6.70042396e-01
-7.91913509e-01 3.54151219e-01 7.11162686e-01 -1.19175875e+00
6.71616912e-01 8.82549286e-01 6.94393069e-02 -5.75554907e-01
3.35316122e-01 2.98370451e-01 -1.57717571e-01 -7.52794743e-01
5.78702331e-01 1.01051509e-01 4.76084471e-01 -1.10144567e+00
1.43539453e+00 1.02212954e+00 4.58571076e-01 -9.92118895e-01
-7.02108741e-01 -7.51195908e-01 -6.60601258e-01 -2.33426124e-01
5.71732283e-01 -1.11448359e+00 -7.51613557e-01 5.45782328e-01
-8.99941981e-01 2.14108545e-02 2.14081615e-01 8.72280121e-01
-4.01716053e-01 8.70468915e-01 -6.81451142e-01 -8.23573232e-01
-3.30927849e-01 -1.16618598e+00 9.30933058e-01 -3.59365016e-01
5.77628314e-01 -5.29105127e-01 2.20153496e-01 8.44945908e-01
-4.75023035e-03 3.61278147e-01 3.82782340e-01 6.73522502e-02
-1.00990307e+00 -2.21160293e-01 -4.31731492e-01 6.86755478e-01
-1.98574632e-01 2.93475896e-01 -8.99175882e-01 -5.53030670e-01
8.96746591e-02 3.09351254e-02 7.32978046e-01 1.19033515e-01
9.67426479e-01 -3.15953672e-01 -3.56001891e-02 8.59442592e-01
1.35464489e+00 -3.90847445e-01 4.07707453e-01 1.35857344e-01
1.62692630e+00 5.39845884e-01 8.18466663e-01 3.37205410e-01
4.44965333e-01 7.35871911e-01 1.99646220e-01 -1.90759413e-02
1.42435327e-01 1.40208676e-01 8.11339557e-01 1.90380287e+00
-7.74554610e-01 4.85107869e-01 -8.76811802e-01 4.24767673e-01
-2.28417492e+00 -6.60942852e-01 -6.99042857e-01 2.23025823e+00
3.17465127e-01 -2.37431616e-01 8.26611519e-02 9.93961394e-02
4.17995781e-01 3.02922249e-01 -3.56524527e-01 -3.47524285e-02
-3.42517257e-01 -1.01318516e-01 8.57085049e-01 2.41385445e-01
-1.19723165e+00 6.82881474e-01 6.42106962e+00 6.32498562e-01
-9.79764581e-01 3.36180925e-01 1.09738037e-01 8.74499157e-02
9.00444686e-02 3.08263868e-01 -2.05654711e-01 1.02149226e-01
7.98083782e-01 -1.23062894e-01 9.88089681e-01 9.33096051e-01
3.62965971e-01 1.39273793e-01 -8.04994643e-01 1.26135874e+00
3.68095040e-01 -9.83814776e-01 1.56545326e-01 1.42194390e-01
8.27604651e-01 4.08918083e-01 1.58989742e-01 -2.64383610e-02
1.23619840e-01 -3.23559463e-01 6.50636494e-01 5.47615886e-01
3.65237594e-01 -3.92983675e-01 3.95835191e-01 -2.29777634e-01
-1.39603043e+00 -7.76852071e-02 -4.60704118e-01 7.26823732e-02
1.99608225e-02 1.00040591e+00 -4.58117783e-01 1.00646472e+00
9.19083118e-01 1.12220299e+00 -6.53972268e-01 9.94693041e-01
-5.64178340e-02 8.92120361e-01 -3.78439873e-01 5.74776888e-01
1.33496881e-01 -1.00609839e+00 1.14018583e+00 1.08324802e+00
2.44296998e-01 1.72342494e-01 4.81259614e-01 3.96048337e-01
1.44412741e-01 3.50516915e-01 -1.50372982e-01 -3.61628354e-01
-2.40998268e-01 1.72779620e+00 -5.88799953e-01 -1.16357632e-01
-4.11602318e-01 9.63251770e-01 3.71331163e-02 7.58675456e-01
-6.95621133e-01 1.07190646e-01 9.20180023e-01 -3.92222255e-01
5.12955725e-01 -6.68973386e-01 -3.66509221e-02 -1.89874363e+00
4.99354303e-01 -1.25510263e+00 4.40644652e-01 -8.49956930e-01
-1.58423221e+00 4.10301983e-01 -7.81995133e-02 -1.68529129e+00
3.29292595e-01 -9.14405525e-01 -2.19602242e-01 5.78450501e-01
-1.43558133e+00 -1.63377941e+00 -2.94316560e-01 9.51080143e-01
1.39033929e-01 -1.20140612e-01 6.71625257e-01 5.49554706e-01
-1.25675023e+00 1.98377803e-01 5.75808585e-01 1.82829380e-01
5.05692482e-01 -1.15776742e+00 1.41823133e-02 1.49901354e+00
1.07922807e-01 9.11978126e-01 6.84976459e-01 -6.08910382e-01
-2.63669944e+00 -1.09181368e+00 -7.29184924e-03 -4.66865182e-01
1.35199344e+00 -3.86146009e-01 -6.93592250e-01 9.04742539e-01
-1.58610240e-01 4.77104098e-01 6.15362406e-01 3.37327361e-01
-1.02428758e+00 -4.64203596e-01 -6.37652695e-01 5.75818419e-01
8.20161521e-01 -6.53179407e-01 -2.17430606e-01 5.94525397e-01
6.01547718e-01 -3.73446882e-01 -1.23813939e+00 1.82682097e-01
6.27411008e-01 -6.12566590e-01 1.21782851e+00 -5.96904576e-01
-4.43727523e-02 -1.02747428e+00 -6.74070001e-01 -1.12916052e+00
-4.73447651e-01 -1.32920337e+00 -4.57618833e-01 9.35968041e-01
-3.48062627e-03 -5.23769498e-01 6.56452239e-01 4.63943869e-01
-9.14432183e-02 -3.90528351e-01 -1.12532949e+00 -8.87648642e-01
-4.90509599e-01 -8.10334980e-01 1.21993132e-01 1.34040070e+00
-5.28626889e-02 1.71412244e-01 -9.78554308e-01 6.27302587e-01
1.39284015e+00 2.00924426e-01 1.06159365e+00 -9.38867927e-01
-2.91144013e-01 -5.38020656e-02 -5.88714182e-01 -8.21031094e-01
-1.51460484e-01 -6.63745642e-01 -6.12286814e-02 -8.42744529e-01
5.07843375e-01 -4.07655209e-01 -5.70918262e-01 2.33424351e-01
-3.28344792e-01 3.02980751e-01 4.68491793e-01 6.42823398e-01
-8.62831295e-01 6.76581740e-01 1.17442596e+00 -3.32201034e-01
-2.10429169e-02 -2.97877222e-01 -6.51005447e-01 7.52198339e-01
1.55011058e-01 -5.84214926e-01 -3.09432298e-01 -4.20559317e-01
3.78678739e-01 -6.56951815e-02 2.03543559e-01 -8.91897321e-01
3.13524157e-02 -4.23180848e-01 2.73654219e-02 -7.50512838e-01
6.01693988e-01 -8.27729046e-01 2.31827959e-01 9.48917866e-02
1.88455373e-01 3.13028008e-01 1.62835028e-02 8.90506923e-01
-3.51969004e-02 4.73845959e-01 2.85077602e-01 2.80677676e-01
-4.89215583e-01 4.44763988e-01 -7.54713342e-02 -2.12058499e-01
7.24659383e-01 1.42131060e-01 -6.41751885e-01 -1.11733591e-02
-4.80470479e-01 5.14702871e-02 2.98139423e-01 2.90371418e-01
8.57595205e-01 -1.56458950e+00 -9.39352930e-01 8.48420337e-02
2.79202908e-01 -3.48896444e-01 3.45598251e-01 1.74729109e+00
-4.49014157e-01 3.53301764e-01 -4.04158682e-02 -8.32793832e-01
-1.38524449e+00 9.07586098e-01 -1.39025208e-02 -3.21240932e-01
-7.12281644e-01 5.06180823e-01 1.93533391e-01 -3.70093167e-01
-1.28007472e-01 -5.06781220e-01 -1.17135562e-01 -1.82710677e-01
6.47006929e-01 9.15323198e-01 -7.64383599e-02 -1.21860337e+00
-4.04290497e-01 1.00924158e+00 4.41430472e-02 -7.16248080e-02
1.39195406e+00 -8.57117027e-02 -8.60222816e-01 5.78026474e-01
1.30640113e+00 3.82182986e-01 -1.14964378e+00 -3.22811663e-01
6.79367259e-02 -8.76118839e-01 4.07148093e-01 -1.30596533e-01
-1.35034120e+00 4.36937362e-01 6.13414407e-01 3.34956944e-02
1.17873180e+00 -5.26995957e-01 7.31336117e-01 7.59557962e-01
4.33188349e-01 -8.69133174e-01 1.66763559e-01 2.88178980e-01
1.16599166e+00 -9.67639446e-01 7.65022457e-01 -7.46602356e-01
-9.05726850e-01 1.03095901e+00 9.89026800e-02 -3.05295765e-01
6.69215441e-01 -1.56807378e-01 4.57726270e-02 -2.71856606e-01
-6.26820624e-01 -1.50498614e-01 6.52600586e-01 4.03383195e-01
3.51879634e-02 4.32297051e-01 -1.53129429e-01 7.78649971e-02
7.71192461e-02 -1.06673181e-01 6.63802624e-01 9.46625113e-01
3.53915155e-01 -9.32475567e-01 -1.13384271e+00 4.09869283e-01
-5.77511549e-01 -3.04295076e-03 -3.26480150e-01 2.39307120e-01
-2.12565243e-01 9.97356892e-01 -5.33154011e-01 -6.91075265e-01
1.31664649e-01 -3.34731698e-01 3.33232611e-01 3.31037119e-02
-1.27506420e-01 6.85492814e-01 1.55170515e-01 -1.15032220e+00
-9.80972409e-01 -1.07211578e+00 -7.55046308e-01 -2.39372402e-01
-2.92149633e-01 9.10623297e-02 9.77880597e-01 6.77858412e-01
3.79325747e-01 1.78305238e-01 8.54458332e-01 -1.04992235e+00
-3.60430330e-01 -8.44281495e-01 -6.66807234e-01 6.62515521e-01
4.42754269e-01 -7.77994394e-01 -4.93603975e-01 1.81891367e-01] | [7.484620094299316, 4.447385311126709] |
14a3564d-14d1-4974-ad0e-632b457d7115 | image-retrieval-on-real-life-images-with-pre | 2108.04024 | null | https://arxiv.org/abs/2108.04024v1 | https://arxiv.org/pdf/2108.04024v1.pdf | Image Retrieval on Real-life Images with Pre-trained Vision-and-Language Models | We extend the task of composed image retrieval, where an input query consists of an image and short textual description of how to modify the image. Existing methods have only been applied to non-complex images within narrow domains, such as fashion products, thereby limiting the scope of study on in-depth visual reasoning in rich image and language contexts. To address this issue, we collect the Compose Image Retrieval on Real-life images (CIRR) dataset, which consists of over 36,000 pairs of crowd-sourced, open-domain images with human-generated modifying text. To extend current methods to the open-domain, we propose CIRPLANT, a transformer based model that leverages rich pre-trained vision-and-language (V&L) knowledge for modifying visual features conditioned on natural language. Retrieval is then done by nearest neighbor lookup on the modified features. We demonstrate that with a relatively simple architecture, CIRPLANT outperforms existing methods on open-domain images, while matching state-of-the-art accuracy on the existing narrow datasets, such as fashion. Together with the release of CIRR, we believe this work will inspire further research on composed image retrieval. | ['Stephen Gould', 'Damien Teney', 'Cristian Rodriguez-Opazo', 'Zheyuan Liu'] | 2021-08-09 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Liu_Image_Retrieval_on_Real-Life_Images_With_Pre-Trained_Vision-and-Language_Models_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Liu_Image_Retrieval_on_Real-Life_Images_With_Pre-Trained_Vision-and-Language_Models_ICCV_2021_paper.pdf | iccv-2021-1 | ['composed-image-retrieval'] | ['computer-vision'] | [ 3.57414603e-01 -2.29257956e-01 -2.52664804e-01 -5.02487361e-01
-1.02170730e+00 -1.21693289e+00 9.01238680e-01 8.92026126e-02
-4.72765058e-01 4.96862791e-02 3.27937394e-01 -2.98231781e-01
1.75090268e-01 -5.49814224e-01 -9.36651170e-01 -4.29481789e-02
4.73577768e-01 4.46274519e-01 2.09008485e-01 -6.57108784e-01
2.75024980e-01 2.96612203e-01 -1.66253161e+00 7.96327353e-01
2.95875967e-01 1.04504681e+00 2.42894247e-01 5.90769053e-01
1.08030759e-01 9.57445085e-01 -4.05456364e-01 -1.09372783e+00
5.85572302e-01 2.17137896e-02 -8.39857519e-01 2.97534704e-01
1.12469161e+00 -8.14509511e-01 -7.42231011e-01 1.07031131e+00
5.17186165e-01 2.02199429e-01 7.70244598e-01 -1.26123464e+00
-1.91145599e+00 5.89039505e-01 -2.69190133e-01 3.02561730e-01
7.01900184e-01 5.31309247e-01 1.17378998e+00 -1.03913152e+00
1.34038687e+00 1.61060655e+00 4.64911550e-01 5.17627060e-01
-1.15773845e+00 -4.44692492e-01 1.95792347e-01 2.63782024e-01
-1.49040306e+00 -5.80162585e-01 5.63671529e-01 -4.25666660e-01
1.07478261e+00 8.92547891e-02 3.98649782e-01 1.18026614e+00
7.55862743e-02 1.19247591e+00 8.96952748e-01 -2.57313788e-01
-1.26841351e-01 2.39920855e-01 -1.34756148e-01 7.06851602e-01
-8.62718448e-02 1.23679183e-01 -3.12724322e-01 1.39141291e-01
6.05635226e-01 4.48950268e-02 -6.96291775e-02 -6.36859357e-01
-1.37067866e+00 9.70778942e-01 6.79045200e-01 1.17726110e-01
-1.46723896e-01 1.30486280e-01 5.14856339e-01 6.12221479e-01
4.25262779e-01 6.43659234e-01 -1.74603805e-01 9.30750594e-02
-8.49214315e-01 5.35176218e-01 7.84402728e-01 1.37116742e+00
7.53432393e-01 -4.03258830e-01 -3.75009269e-01 9.06979680e-01
1.52531043e-01 8.68512869e-01 3.24119151e-01 -1.22826278e+00
6.39163613e-01 4.67309922e-01 1.82023451e-01 -1.20734465e+00
2.85360605e-01 -4.72741295e-03 -3.86991799e-01 -8.31114799e-02
2.93927908e-01 5.66102087e-01 -8.87502313e-01 1.35456622e+00
2.51091361e-01 -6.00627005e-01 2.17123747e-01 1.21520996e+00
1.26826370e+00 6.46320701e-01 3.27057280e-02 4.62720394e-01
1.57032955e+00 -1.42001295e+00 -5.64998746e-01 -3.99254709e-01
4.31853294e-01 -1.17929780e+00 1.37164259e+00 1.54974446e-01
-1.24130130e+00 -5.95024526e-01 -8.94183695e-01 -9.81606424e-01
-9.57589865e-01 5.67599498e-02 4.98758435e-01 2.43256167e-01
-1.21070111e+00 2.89784789e-01 -1.36704117e-01 -6.85247898e-01
5.71660161e-01 -1.85921803e-01 -7.71168888e-01 -9.55968499e-01
-1.23802555e+00 9.26281929e-01 2.36998890e-02 -7.91456643e-03
-1.15862203e+00 -9.27280188e-01 -1.16283512e+00 -2.71507174e-01
4.59933579e-01 -7.79693365e-01 1.52486312e+00 -1.22059417e+00
-8.97217035e-01 1.62063718e+00 9.39138606e-02 -4.10409003e-01
6.48405433e-01 -1.80389449e-01 -5.08690894e-01 5.51674962e-01
4.98556376e-01 1.21138656e+00 1.14896894e+00 -1.45364642e+00
-3.40907693e-01 -3.85236949e-01 7.54550397e-01 2.32751176e-01
-2.21722364e-01 1.75618067e-01 -1.01461577e+00 -9.51061130e-01
-4.45364952e-01 -1.08880568e+00 1.12574873e-02 5.73557675e-01
-3.22912961e-01 -1.73322365e-01 6.43016934e-01 -8.71154249e-01
8.05873454e-01 -2.43476653e+00 2.15951577e-02 -8.92947987e-02
1.57176822e-01 2.57255107e-01 -8.33599091e-01 6.35918975e-01
9.85307023e-02 7.66235888e-02 2.25278921e-02 -2.43224353e-01
3.42550039e-01 5.53977899e-02 -7.14354992e-01 3.08935434e-01
1.34382218e-01 1.55142891e+00 -1.04337811e+00 -6.48419619e-01
4.05553937e-01 3.32080781e-01 -4.69336987e-01 2.05637634e-01
-4.21491921e-01 -1.49030343e-01 -4.51976597e-01 8.97522807e-01
6.21247172e-01 -2.18503833e-01 -2.80916005e-01 -7.45286763e-01
2.33433291e-01 -3.44441205e-01 -6.25962555e-01 2.13157940e+00
-4.30457383e-01 8.55486572e-01 -1.08061753e-01 -5.56489646e-01
6.00697815e-01 -1.88497454e-01 1.14562891e-01 -1.33393824e+00
-8.07829127e-02 -1.33555084e-01 -4.98902917e-01 -7.95356870e-01
8.63406420e-01 2.94698268e-01 -5.26350379e-01 5.68776429e-01
-1.03131309e-01 -5.90090632e-01 4.97712433e-01 7.63533175e-01
1.18702996e+00 1.05025232e-01 1.24006569e-01 3.85537595e-02
3.03201795e-01 4.66490924e-01 -2.22219706e-01 1.07907450e+00
-4.01267111e-01 7.04701424e-01 8.12280551e-02 -5.38457990e-01
-1.26459098e+00 -1.33586061e+00 -1.31260650e-02 1.34476566e+00
5.31547308e-01 -4.30986881e-01 -5.42686164e-01 -5.37361264e-01
4.18489993e-01 6.15314424e-01 -8.14358354e-01 -1.59475669e-01
-3.72594714e-01 1.08121939e-01 6.94098771e-01 3.96973640e-01
7.05639482e-01 -1.12166369e+00 -2.88447767e-01 -1.49397850e-01
-1.73277795e-01 -1.44934928e+00 -9.83707190e-01 -4.74656880e-01
-3.58766645e-01 -1.13654435e+00 -8.79206657e-01 -1.25858927e+00
5.50808668e-01 6.25915170e-01 1.61362827e+00 -3.67770642e-02
-7.47959852e-01 1.12701464e+00 -4.59224343e-01 1.61295347e-02
-6.19570017e-01 -1.22452028e-01 -4.85386550e-01 -1.67619437e-01
4.33787316e-01 -3.77911367e-02 -7.68673837e-01 3.93851310e-01
-1.37288797e+00 8.69189948e-03 7.06691027e-01 7.02831566e-01
6.19205117e-01 -2.26806328e-01 1.74105838e-01 -8.32437158e-01
6.93230033e-01 -4.67086792e-01 -3.92538697e-01 6.57538414e-01
-3.41560513e-01 -3.07453480e-02 2.86666989e-01 -7.37527966e-01
-1.11296833e+00 5.03190747e-03 3.04398954e-01 -7.96967566e-01
-3.14047560e-02 2.18555912e-01 6.44178465e-02 -7.59024993e-02
6.39027596e-01 2.99962372e-01 -1.49110779e-01 -2.82634526e-01
1.14683902e+00 6.61390960e-01 9.02636886e-01 -4.09747094e-01
9.65339780e-01 6.51061893e-01 -4.13097858e-01 -3.93695503e-01
-9.57593441e-01 -4.92061496e-01 -4.45595235e-01 -1.43296927e-01
9.04367626e-01 -1.34037888e+00 -5.18437445e-01 2.87330776e-01
-1.10800004e+00 -4.34752315e-01 -2.49103025e-01 -8.93918797e-02
-5.57503164e-01 3.77688140e-01 -8.84673417e-01 -1.65082708e-01
-3.19440037e-01 -1.18396592e+00 1.71438146e+00 -2.58328229e-01
-1.19968317e-01 -8.08465004e-01 -2.64086694e-01 1.00119519e+00
4.02338803e-01 -4.08761874e-02 9.71068382e-01 -4.34734434e-01
-9.58652258e-01 -2.97426462e-01 -5.76215386e-01 4.05591995e-01
-7.45136812e-02 -1.20222956e-01 -7.37155557e-01 -2.68686146e-01
-5.58560014e-01 -9.90177631e-01 1.11548233e+00 -1.03906944e-01
1.19806397e+00 -1.31072089e-01 -2.56416857e-01 4.00346041e-01
1.41934943e+00 -3.77277173e-02 8.10273230e-01 3.97564203e-01
7.30130374e-01 5.72244167e-01 9.97635126e-01 2.58013397e-01
7.89475203e-01 6.48190796e-01 4.61657107e-01 -2.46767357e-01
-6.39877617e-01 -6.75667346e-01 3.28482598e-01 3.65253180e-01
4.71016288e-01 -4.70221668e-01 -6.79765940e-01 8.44791174e-01
-1.84497654e+00 -1.04154289e+00 2.54678458e-01 1.60692549e+00
9.01854515e-01 -2.91311145e-01 -3.05201203e-01 -4.24745649e-01
7.20415831e-01 5.41664362e-01 -8.32409859e-01 -2.85701066e-01
-3.52424920e-01 -1.22964896e-01 4.31611240e-01 3.27910334e-01
-1.23936915e+00 1.26933575e+00 6.55878925e+00 1.01397991e+00
-7.15997756e-01 2.25753695e-01 4.64542627e-01 -2.99564034e-01
-6.57022715e-01 -8.90188292e-02 -5.56896746e-01 1.49801642e-01
3.75686198e-01 7.55034536e-02 9.48909402e-01 7.96276748e-01
-2.54804641e-01 -2.16009486e-02 -1.43656957e+00 1.38601255e+00
8.74937832e-01 -1.32898736e+00 6.01788461e-01 -2.41928935e-01
8.98642302e-01 -2.88199876e-02 4.73944306e-01 5.82822382e-01
4.27588284e-01 -9.61654723e-01 1.05451572e+00 4.02357250e-01
9.37144399e-01 -3.18382114e-01 3.64040524e-01 -1.34461418e-01
-9.94621992e-01 -6.60643429e-02 -3.77094686e-01 2.40689471e-01
1.42041013e-01 3.78386170e-01 -5.81049323e-01 2.69988090e-01
9.41028714e-01 1.17428648e+00 -1.20841932e+00 3.75970066e-01
-9.27557051e-02 -1.72436133e-01 4.41950113e-02 5.30801862e-02
2.53519863e-01 1.35725170e-01 2.82853007e-01 1.02616787e+00
-7.30955973e-02 -1.30511656e-01 1.46944508e-01 1.03056085e+00
-4.10410941e-01 -7.20229000e-03 -9.72573638e-01 -4.23391163e-01
4.72740740e-01 1.33705342e+00 -4.84140426e-01 -4.82571363e-01
-5.63356638e-01 1.50284100e+00 4.04080808e-01 5.49022317e-01
-7.97064185e-01 -3.18797708e-01 5.06314576e-01 7.59396404e-02
7.17999935e-01 3.53385545e-02 3.17563266e-01 -1.34802091e+00
3.07630897e-01 -1.20898020e+00 4.36898023e-01 -1.56238282e+00
-1.77474594e+00 5.49354017e-01 8.73565003e-02 -1.00794888e+00
-2.69423932e-01 -7.24621356e-01 6.43365309e-02 3.99377614e-01
-1.73759389e+00 -1.67370248e+00 -3.96596640e-01 1.00869906e+00
8.69398952e-01 -2.39616334e-01 6.20399415e-01 4.26093489e-01
-2.90676691e-02 6.79529369e-01 7.33020976e-02 3.30202550e-01
1.17758536e+00 -8.61787021e-01 6.32126629e-01 5.40443897e-01
3.73312652e-01 6.18167162e-01 4.79753703e-01 -6.51358128e-01
-2.00119638e+00 -1.28688169e+00 7.25454807e-01 -9.22162175e-01
9.09012735e-01 -6.19336247e-01 -3.86240512e-01 8.89046788e-01
4.78458196e-01 3.71636420e-01 4.31315213e-01 -3.31059366e-01
-1.04933381e+00 -1.55009285e-01 -1.17532361e+00 9.34916258e-01
1.46827972e+00 -1.09924710e+00 -8.67799163e-01 5.46562850e-01
1.02153230e+00 -4.41119462e-01 -1.12572598e+00 1.54943481e-01
7.59288013e-01 -6.27590239e-01 1.31021988e+00 -6.42783165e-01
8.61724079e-01 -2.71276236e-01 -6.88281715e-01 -1.02385008e+00
-3.25556964e-01 -3.01357388e-01 2.68624574e-01 1.51089358e+00
2.61388004e-01 -3.23679149e-01 1.87248066e-01 8.33565772e-01
1.32704660e-01 -3.79528135e-01 -4.90990490e-01 -5.64485610e-01
-5.57132661e-02 -3.90754163e-01 5.50708115e-01 7.62769163e-01
-5.09149253e-01 3.37067544e-01 -1.58274740e-01 -1.07887208e-01
4.78760988e-01 4.90639716e-01 9.08687830e-01 -6.23608410e-01
-2.65450150e-01 -4.74564470e-02 -5.90681911e-01 -1.09912932e+00
6.59282327e-01 -1.04018950e+00 2.17504781e-02 -1.48056924e+00
6.07774496e-01 -2.19865263e-01 2.16556657e-02 4.73391354e-01
1.47158712e-01 8.80328000e-01 5.12306750e-01 3.28227162e-01
-1.20372999e+00 3.99918735e-01 1.53214681e+00 -8.09868515e-01
3.97152871e-01 -7.86540568e-01 -9.44973707e-01 3.50134581e-01
2.42742255e-01 -1.87541306e-01 -6.10270560e-01 -9.52966213e-01
3.83232594e-01 -1.32645309e-01 8.42665017e-01 -4.86088485e-01
1.26348078e-01 3.94292809e-02 5.17451108e-01 -5.28959632e-01
4.60571170e-01 -8.82473350e-01 5.56287393e-02 1.02783337e-01
-9.14944649e-01 4.88206327e-01 7.70620629e-02 7.62328088e-01
-3.49138230e-01 -4.12527248e-02 3.45945299e-01 -3.97402555e-01
-1.33881474e+00 3.61212194e-01 -1.64113179e-01 2.82628447e-01
1.04901493e+00 -6.17995113e-02 -7.70398259e-01 -6.04550481e-01
-4.36619252e-01 2.67074049e-01 8.75332832e-01 1.12944424e+00
7.22478926e-01 -1.53383458e+00 -5.22898257e-01 3.54319327e-02
8.32366586e-01 -3.16600621e-01 4.13673162e-01 1.43258885e-01
-4.89466727e-01 5.09252250e-01 -7.32633471e-02 -5.65671265e-01
-1.23176670e+00 1.28941250e+00 6.50020093e-02 -7.31047764e-02
-5.20327747e-01 5.37876129e-01 5.15954912e-01 -4.95771259e-01
7.83518702e-02 -2.95373917e-01 2.14766070e-01 -1.25611136e-02
6.51115119e-01 -8.57615098e-02 -1.18010350e-01 -8.97183836e-01
-2.91652322e-01 7.19500303e-01 -3.89291435e-01 -1.91049233e-01
1.00507784e+00 -5.48656106e-01 -1.07558526e-01 2.64848977e-01
1.69043636e+00 -2.86328852e-01 -1.13702393e+00 -4.91165817e-01
-4.21993524e-01 -7.57331967e-01 5.75084314e-02 -1.07581413e+00
-1.10775101e+00 7.07224429e-01 6.73224032e-01 -2.94965148e-01
1.13567603e+00 4.07398134e-01 9.55329716e-01 7.45217502e-01
4.28535432e-01 -1.00065136e+00 5.20165920e-01 2.50813305e-01
1.29051197e+00 -1.68611419e+00 -9.42392461e-03 -3.42552453e-01
-9.22940075e-01 8.66515517e-01 5.83371937e-01 -2.40753412e-01
4.43380624e-01 -1.14622220e-01 2.72456348e-01 -2.85209686e-01
-7.00448930e-01 -3.28244179e-01 3.57151598e-01 8.08278918e-01
-1.16390120e-02 -3.06827910e-02 3.02200645e-01 3.76829416e-01
-8.23267475e-02 1.41822129e-01 1.43703252e-01 9.32629108e-01
1.39774725e-01 -8.20981681e-01 -2.76991099e-01 3.15510243e-01
-2.32328519e-01 -4.67767745e-01 -5.86658359e-01 8.72276843e-01
-1.61062542e-03 9.73732531e-01 2.85400510e-01 1.26173347e-02
4.99717236e-01 -3.86447817e-01 8.19958389e-01 -4.19853210e-01
-6.31042540e-01 -4.61521864e-01 2.02454254e-01 -1.00589144e+00
-5.12458324e-01 -4.34639335e-01 -6.90622449e-01 -3.19947392e-01
1.53141513e-01 -4.21958685e-01 5.55479050e-01 6.79688871e-01
5.94556570e-01 1.33073092e-01 3.01343560e-01 -9.10068870e-01
-5.73449075e-01 -5.91780245e-01 -6.10626876e-01 1.22546899e+00
3.30286354e-01 -4.64603543e-01 -1.77660555e-01 3.74976575e-01] | [10.851000785827637, 1.4210984706878662] |
d243613c-6321-49fe-808e-582f8d5ec839 | near-optimal-experimental-design-under-the | 2302.05005 | null | https://arxiv.org/abs/2302.05005v1 | https://arxiv.org/pdf/2302.05005v1.pdf | Near-Optimal Experimental Design Under the Budget Constraint in Online Platforms | A/B testing, or controlled experiments, is the gold standard approach to causally compare the performance of algorithms on online platforms. However, conventional Bernoulli randomization in A/B testing faces many challenges such as spillover and carryover effects. Our study focuses on another challenge, especially for A/B testing on two-sided platforms -- budget constraints. Buyers on two-sided platforms often have limited budgets, where the conventional A/B testing may be infeasible to be applied, partly because two variants of allocation algorithms may conflict and lead some buyers to exceed their budgets if they are implemented simultaneously. We develop a model to describe two-sided platforms where buyers have limited budgets. We then provide an optimal experimental design that guarantees small bias and minimum variance. Bias is lower when there is more budget and a higher supply-demand rate. We test our experimental design on both synthetic data and real-world data, which verifies the theoretical results and shows our advantage compared to Bernoulli randomization. | ['Zheng Cai', 'Zhihua Zhu', 'Yuqing Kong', 'Jinshan Zhang', 'Yuan Yuan', 'Yongkang Guo'] | 2023-02-10 | null | null | null | null | ['experimental-design'] | ['methodology'] | [-1.61358997e-01 -2.59015411e-01 -8.63423049e-01 -1.95380598e-01
-3.83564740e-01 -9.19504941e-01 1.91493571e-01 6.65709078e-02
-4.34285998e-01 9.91171300e-01 -3.22621942e-01 -8.72143328e-01
-3.41083884e-01 -7.41869748e-01 -1.19816911e+00 -3.78792316e-01
-5.89345813e-01 5.14009416e-01 1.28535405e-01 1.20752059e-01
1.64936051e-01 5.10035455e-02 -1.37366605e+00 -5.24280667e-02
5.50879121e-01 9.49104786e-01 2.39960551e-02 5.53129792e-01
7.87706897e-02 2.14783624e-01 -6.29273593e-01 -6.44388139e-01
9.42384124e-01 -4.75533873e-01 -2.37582386e-01 -2.85396054e-02
-6.46221489e-02 -6.09102786e-01 2.81393170e-01 1.20808601e+00
4.03103769e-01 -5.45791388e-01 3.64075840e-01 -2.01431704e+00
-4.24014509e-01 1.08408296e+00 -1.41601670e+00 3.08221877e-01
3.80945951e-01 1.40881196e-01 1.17103314e+00 -2.26679370e-01
5.02554834e-01 1.21131730e+00 4.96917039e-01 1.92618102e-01
-1.40756285e+00 -1.07592130e+00 3.18152428e-01 -2.73307621e-01
-1.10329509e+00 -5.46461456e-02 3.43022943e-01 -4.25887227e-01
3.71540189e-01 3.54957312e-01 6.93335235e-01 9.77991045e-01
4.49653059e-01 7.35466957e-01 1.39071620e+00 -4.16709661e-01
4.12520260e-01 4.19347733e-01 -8.58198926e-02 8.37585703e-02
1.32009900e+00 4.24743533e-01 -6.14179850e-01 -3.98583025e-01
6.76962018e-01 2.40814220e-02 -5.39169423e-02 -4.00634438e-01
-1.02411675e+00 8.99823368e-01 5.23176976e-04 -5.89912571e-02
-4.99169201e-01 2.26827055e-01 2.76513904e-01 7.31616259e-01
1.57961458e-01 5.02604306e-01 -6.17365301e-01 -1.90022022e-01
-8.38590562e-01 5.41045070e-01 9.36282873e-01 1.35818744e+00
3.21249574e-01 -4.31044549e-01 -3.04078192e-01 1.53543323e-01
2.77753264e-01 7.21218348e-01 1.45571649e-01 -6.65737092e-01
8.44454765e-01 2.94532865e-01 9.13710356e-01 -7.03455091e-01
-5.49095832e-02 -4.04699862e-01 -3.35846812e-01 -4.31074314e-02
5.51048636e-01 -6.33683145e-01 -2.66190380e-01 1.74881530e+00
1.85584828e-01 -1.09167434e-02 -5.22920847e-01 1.03905809e+00
-7.07724616e-02 3.08743477e-01 -1.17229275e-01 -7.12639928e-01
1.26376677e+00 -6.64081097e-01 -7.36474812e-01 -2.11817682e-01
6.50420785e-01 -7.28947341e-01 1.23142636e+00 4.05905575e-01
-1.29523647e+00 1.81569546e-01 -1.04889870e+00 9.46089387e-01
2.12219149e-01 -4.77835745e-01 6.50662541e-01 1.18549418e+00
-7.69875288e-01 2.85564542e-01 -7.29950249e-01 7.58990422e-02
3.93969536e-01 3.17223877e-01 1.32104158e-01 -1.33867323e-01
-1.14220130e+00 4.18424010e-01 -2.73944378e-01 -1.06590882e-01
-1.02506948e+00 -8.53217900e-01 -4.45192099e-01 3.03741485e-01
6.67189181e-01 -5.03958881e-01 1.42713404e+00 -9.20481503e-01
-1.10901451e+00 3.84007007e-01 2.88382798e-01 -4.27038193e-01
8.92979383e-01 -3.97093967e-02 2.55732741e-02 -5.17435730e-01
5.46161711e-01 2.21524134e-01 2.37771481e-01 -1.07535219e+00
-7.73829997e-01 -4.86807108e-01 1.62011370e-01 6.84640259e-02
-4.80308503e-01 2.85182714e-01 -6.19416796e-02 -3.53569746e-01
-1.87176004e-01 -9.44664061e-01 -4.09630835e-01 -5.99685431e-01
-6.33434951e-01 2.21170723e-01 1.20425425e-01 9.32647064e-02
1.13396955e+00 -1.98369813e+00 -5.96111834e-01 3.26243073e-01
-1.14652246e-01 -4.64192390e-01 -2.19799519e-01 4.66994822e-01
1.03947483e-01 8.19364786e-01 5.43897390e-01 1.27399772e-01
2.58583248e-01 -1.99419826e-01 -1.67912155e-01 6.80980623e-01
-5.20724095e-02 5.22459090e-01 -6.47756696e-01 -2.84240901e-01
-5.47615230e-01 -7.07833827e-01 -7.55784154e-01 9.00309011e-02
-2.21376941e-01 7.52012506e-02 -7.01601267e-01 7.46853828e-01
9.96010005e-01 -4.82336402e-01 6.35418415e-01 7.25765109e-01
-3.08115751e-01 1.80586994e-01 -1.29282725e+00 6.14085495e-01
-3.68026435e-01 3.01468492e-01 2.88075119e-01 -6.61058009e-01
4.26268309e-01 -2.23961212e-02 2.17577651e-01 -9.62407231e-01
2.34371111e-01 3.83351713e-01 4.72421169e-01 -4.53060627e-01
1.05583817e-01 -2.44386569e-01 -2.48540893e-01 7.43439913e-01
-7.62861788e-01 1.89729780e-01 2.42161721e-01 1.27399191e-02
1.35689592e+00 -2.30385378e-01 2.41679057e-01 -5.69771707e-01
-2.90832609e-01 6.17715269e-02 1.18828940e+00 1.26739442e+00
-3.32874924e-01 -1.16980769e-01 1.36961222e+00 -5.56528643e-02
-1.10016358e+00 -5.80068231e-01 -1.20343983e-01 1.07576942e+00
6.29447341e-01 1.18157953e-01 -2.97297806e-01 -5.74382365e-01
8.96289825e-01 5.77500820e-01 -8.91553581e-01 2.95327306e-01
5.65628847e-03 -7.93152094e-01 1.64344251e-01 4.73291039e-01
3.16370368e-01 -3.18215072e-01 -7.21913874e-01 -3.13254558e-02
2.63080090e-01 -7.97698498e-01 -6.30894363e-01 -3.16221733e-03
-6.24139667e-01 -1.04137945e+00 -6.24565244e-01 -2.26888478e-01
7.83828199e-01 7.70190060e-01 1.01023901e+00 -1.03779562e-01
-6.50731474e-02 -1.12334816e-02 -2.13482365e-01 -8.37393522e-01
-6.24760836e-02 -3.02356809e-01 2.63678640e-01 -2.67006159e-01
2.84628451e-01 -8.82342681e-02 -1.03908157e+00 1.17251205e+00
-6.51566446e-01 -4.16510820e-01 7.29898989e-01 7.88101792e-01
3.16586196e-01 2.11907342e-01 7.33127952e-01 -1.04977906e+00
6.27271295e-01 -8.66146922e-01 -1.45626724e+00 3.18289071e-01
-9.47908163e-01 -3.82085651e-01 3.11525404e-01 -7.96921313e-01
-7.84608126e-01 -4.09870952e-01 9.57080007e-01 -3.86942923e-01
4.29943472e-01 8.03567290e-01 -2.36916617e-01 3.32747847e-02
6.42613351e-01 -5.22941470e-01 1.39836401e-01 -4.82727885e-02
-4.40141968e-02 5.19788384e-01 -4.22858685e-01 -7.02803552e-01
6.58541024e-01 2.62651473e-01 9.32866186e-02 -6.47284389e-02
-3.44991118e-01 -3.20073701e-02 3.57173532e-01 -8.87919962e-02
2.32654214e-01 -9.73937929e-01 -1.34873497e+00 -3.54694054e-02
-6.68565869e-01 -5.95910013e-01 2.75805853e-02 7.33064711e-01
-4.35543537e-01 -7.06550181e-02 -3.75865281e-01 -1.05207515e+00
3.94955009e-01 -1.36254847e+00 5.34777999e-01 3.24318945e-01
-5.56879938e-02 -6.00958943e-01 -4.22250591e-02 2.46871203e-01
4.78313148e-01 -4.81520733e-03 6.67268276e-01 -8.86728466e-01
-9.34550345e-01 -4.06319231e-01 -3.29011530e-02 -2.25461930e-01
-1.70022234e-01 1.26841098e-01 -7.00214654e-02 -6.06343806e-01
7.31586516e-02 -2.53776520e-01 3.52273998e-03 7.31576622e-01
1.23463202e+00 -6.74727499e-01 -6.02489650e-01 9.03177783e-02
1.38631082e+00 2.96479851e-01 1.91076443e-01 7.29954720e-01
-2.12018773e-01 6.45885170e-01 1.31575775e+00 9.59853053e-01
1.47797406e-01 6.50210083e-01 5.50275683e-01 1.36026263e-01
6.85617864e-01 -4.68602568e-01 3.75630289e-01 8.42781737e-02
3.63138527e-01 -5.80225587e-01 -1.08189225e+00 6.41896665e-01
-1.79644656e+00 -6.19846404e-01 -2.67442018e-01 2.81918645e+00
7.62676001e-01 4.37642574e-01 6.62373245e-01 -1.43396363e-01
1.16473424e+00 -4.99518842e-01 -5.72062254e-01 -3.48122209e-01
1.16695650e-01 -3.87617081e-01 1.19159055e+00 -1.11620678e-02
-4.68391091e-01 1.76913857e-01 7.56475639e+00 5.44707835e-01
-9.67870295e-01 3.40416491e-01 1.20452452e+00 -8.14994991e-01
-6.14369273e-01 2.55876064e-01 -1.11021733e+00 8.05893481e-01
1.09151340e+00 -7.54966378e-01 3.21485996e-02 9.97750044e-01
5.67449450e-01 -3.51169407e-01 -1.53397536e+00 5.87773621e-01
-3.36524367e-01 -1.27619672e+00 -4.68165010e-01 5.75135589e-01
1.08213007e+00 -2.43124738e-01 2.05580533e-01 3.16821933e-01
7.86235571e-01 -7.19257414e-01 9.06398714e-01 -2.69750178e-01
7.16337442e-01 -6.94469571e-01 1.12738371e+00 4.86632645e-01
-3.28444749e-01 -3.43721420e-01 -2.05813199e-01 -4.98731464e-01
-2.24168319e-02 8.33138287e-01 -8.69250417e-01 2.36736506e-01
7.42917180e-01 -1.42618760e-01 6.83934242e-02 1.24580753e+00
-2.16336310e-01 7.12794065e-01 -2.20179468e-01 -6.59516215e-01
2.74568461e-02 -2.30976894e-01 3.00706446e-01 5.85761428e-01
4.64232177e-01 -3.13840457e-03 3.41186285e-01 9.46569622e-01
-3.55122089e-01 1.86559618e-01 -6.98791981e-01 -1.69919565e-01
9.47457373e-01 1.05853403e+00 -9.61022317e-01 -2.28667893e-02
-6.47939980e-01 6.05919026e-02 -3.52143675e-01 2.86395729e-01
-1.23010898e+00 -2.18499750e-01 3.83805245e-01 5.79213023e-01
2.29754433e-01 8.61863643e-02 -6.20119512e-01 -7.48493552e-01
1.69356123e-01 -9.85161781e-01 3.84412616e-01 -3.64552885e-01
-1.39631081e+00 -1.66525319e-01 6.78037032e-02 -1.27662635e+00
1.18485503e-01 -4.00158584e-01 -5.22016764e-01 7.66319036e-01
-1.33395684e+00 -4.38898087e-01 1.38994465e-02 9.58089679e-02
1.56174973e-01 5.96656604e-03 6.41342625e-02 2.35423952e-01
-8.23432267e-01 9.69023764e-01 1.66206136e-01 -2.47722343e-01
8.32079172e-01 -1.00272727e+00 -9.99223664e-02 7.45044470e-01
-5.55454016e-01 8.40307295e-01 8.70117664e-01 -1.01552188e+00
-1.85963035e+00 -7.56277978e-01 4.24036145e-01 -1.40157148e-01
1.05995047e+00 -5.72064996e-01 -2.34788999e-01 7.45310783e-01
4.71556038e-02 3.10816877e-02 1.01132405e+00 6.05559766e-01
-1.66582003e-01 -4.10967618e-01 -1.18810642e+00 8.44446957e-01
8.56875598e-01 1.71000868e-01 -2.42373701e-02 5.52891195e-01
6.81241214e-01 -9.70588699e-02 -6.85043573e-01 4.14291471e-01
5.57439446e-01 -9.26106036e-01 1.84537902e-01 -8.43767583e-01
6.17649972e-01 -2.17094086e-02 -1.25633910e-01 -1.15082693e+00
-1.19240813e-01 -1.08433342e+00 7.42444992e-01 1.17285824e+00
9.73786592e-01 -1.03360188e+00 7.79134393e-01 9.43593264e-01
4.55123156e-01 -4.79365170e-01 -9.36634898e-01 -1.43182337e+00
1.20160125e-01 -1.05911486e-01 7.64062047e-01 8.79707992e-01
3.87347937e-01 -1.73328325e-01 -1.38247475e-01 1.86692759e-01
5.88839233e-01 4.36302006e-01 9.72685099e-01 -7.74682999e-01
-5.95037937e-01 -5.36293566e-01 -3.60711396e-01 -7.51648605e-01
-5.80664054e-02 -3.71103138e-01 1.69746980e-01 -7.23975241e-01
8.57183337e-01 -1.01891530e+00 -1.52869701e-01 3.74249279e-01
-4.19709682e-02 -7.62860179e-02 -1.02368124e-01 2.44116023e-01
-5.95884502e-01 2.68631101e-01 1.22664499e+00 1.63080096e-01
-3.01560134e-01 3.47665846e-01 -1.10177267e+00 3.48467618e-01
4.94757891e-01 -6.93085313e-01 -3.94841671e-01 -1.33881494e-01
9.04068232e-01 7.37117946e-01 -1.70827769e-02 -1.63803026e-01
1.23788662e-01 -7.49182880e-01 1.36352912e-01 -2.15943098e-01
-6.01926565e-01 -7.62089849e-01 5.04109085e-01 7.56810308e-01
-4.31116700e-01 5.18782496e-01 1.18054688e-01 7.92949796e-01
1.01457380e-01 -2.43784964e-01 3.45714897e-01 2.62521207e-03
2.67406762e-01 2.81007886e-01 -1.91774070e-01 -9.68227983e-02
1.58589423e+00 7.92615265e-02 -8.33637774e-01 -5.51642716e-01
-2.19503835e-01 8.91597450e-01 7.44891167e-01 1.96846157e-01
9.78717506e-02 -1.31708944e+00 -7.28239119e-01 -7.24510849e-02
3.48608345e-02 -4.11857575e-01 1.98483393e-01 1.17716002e+00
-4.64600056e-01 3.80660474e-01 -2.55913347e-01 -6.05450273e-01
-1.12670541e+00 9.29081082e-01 7.45202824e-02 -1.97640106e-01
2.16142967e-01 7.70985603e-01 3.80403817e-01 1.21646769e-01
-2.43739318e-02 5.82649112e-02 4.17092532e-01 9.78529975e-02
3.63867790e-01 4.88396257e-01 3.35687003e-03 2.91897178e-01
-5.90171576e-01 -1.08532220e-01 -2.57222086e-01 -5.42836785e-01
1.16034353e+00 -1.94612980e-01 -2.15755820e-01 3.52892399e-01
4.29912657e-01 4.24450785e-01 -1.09200037e+00 1.23904176e-01
-2.79547484e-03 -1.25511622e+00 -1.46159276e-01 -6.58825934e-01
-8.76143634e-01 4.04798150e-01 1.54454663e-01 8.76563966e-01
7.87527680e-01 -1.89198226e-01 -2.56006271e-02 -1.23236194e-01
9.01567578e-01 -9.80151594e-01 9.88835990e-02 -2.42732167e-01
7.47764945e-01 -1.17386603e+00 1.17111623e-01 -6.81186378e-01
-5.44334471e-01 3.99263829e-01 7.86161184e-01 4.74931709e-02
6.20831192e-01 5.65842748e-01 -2.08379105e-01 1.64616749e-01
-1.17978287e+00 1.45673588e-01 -3.93203199e-01 4.35273238e-02
4.95072693e-01 5.23750901e-01 -1.06473720e+00 9.30280566e-01
-3.63141522e-02 1.15614934e-02 1.17152393e+00 1.11014330e+00
-4.58588116e-02 -1.27901816e+00 -7.06140637e-01 6.68069839e-01
-7.94652522e-01 1.29688904e-01 -1.90893948e-01 1.11812830e+00
-1.04446068e-01 1.13857543e+00 3.47770870e-01 -2.27811798e-01
3.28731745e-01 -2.68481493e-01 4.10338968e-01 -5.67612588e-01
-4.15345609e-01 5.48024714e-01 7.14392494e-03 -5.57004988e-01
-2.08328933e-01 -1.03995979e+00 -6.06430948e-01 -6.61507308e-01
-8.76881003e-01 2.23033234e-01 9.48496938e-01 4.14511681e-01
5.66230237e-01 3.14492494e-01 1.31216323e+00 -2.92495072e-01
-1.27560759e+00 -8.01821053e-01 -1.13392258e+00 2.16899514e-01
-9.71373543e-02 -1.01103699e+00 -6.17617846e-01 -6.24812186e-01] | [4.582613945007324, 3.3271994590759277] |
a99da37b-26a1-4797-a6ce-7b634968d959 | a-multi-level-annotated-corpus-of-scientific | null | null | https://aclanthology.org/2020.lrec-1.824 | https://aclanthology.org/2020.lrec-1.824.pdf | A Multi-level Annotated Corpus of Scientific Papers for Scientific Document Summarization and Cross-document Relation Discovery | Related work sections or literature reviews are an essential part of every scientific article being crucial for paper reviewing and assessment. The automatic generation of related work sections can be considered an instance of the multi-document summarization problem. In order to allow the study of this specific problem, we have developed a manually annotated, machine readable data-set of related work sections, cited papers (e.g. references) and sentences, together with an additional layer of papers citing the references. We additionally present experiments on the identification of cited sentences, using as input citation contexts. The corpus alongside the gold standard are made available for use by the scientific community. | ['Horacio Saggion', "Ahmed Abura{'}ed", 'Luis Chiruzzo'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['scientific-article-summarization'] | ['natural-language-processing'] | [ 4.90410149e-01 4.80997026e-01 -3.83366853e-01 1.69379056e-01
-1.24153244e+00 -8.94315898e-01 9.23159659e-01 8.36716890e-01
-3.38998228e-01 1.27940035e+00 6.99173272e-01 -4.80199277e-01
-4.71993893e-01 -4.36648607e-01 -4.07261670e-01 -3.61690968e-01
4.04668838e-01 3.64713132e-01 2.09462792e-01 8.47566798e-02
1.23965764e+00 4.95984852e-01 -1.48623943e+00 1.26946509e-01
1.16617727e+00 2.78438598e-01 4.22322094e-01 9.00059283e-01
-7.70868421e-01 3.71504992e-01 -1.40907252e+00 -5.91367483e-01
-3.55817497e-01 -6.07698143e-01 -1.13808227e+00 5.13736457e-02
6.36328399e-01 3.58788669e-01 2.23318249e-01 1.00010288e+00
3.46183389e-01 -6.52893856e-02 7.67099619e-01 -1.03480029e+00
-4.68532026e-01 9.52951074e-01 -6.46246612e-01 7.65524685e-01
6.27040029e-01 -3.76120120e-01 9.20984089e-01 -6.11754656e-01
1.09877884e+00 1.16330719e+00 2.39315748e-01 3.32158327e-01
-9.23757911e-01 -1.84531257e-01 -2.85289474e-02 -6.28439188e-02
-1.02189469e+00 -6.41359985e-01 8.73310387e-01 -6.14701927e-01
9.70521450e-01 4.99625474e-01 3.18818092e-01 1.14784634e+00
6.67755127e-01 3.03583264e-01 9.35835183e-01 -9.62863386e-01
2.72792876e-01 2.42304876e-01 8.68360877e-01 1.89088032e-01
1.10316110e+00 -7.28221476e-01 -3.04925472e-01 -4.17907268e-01
1.59529373e-01 -3.22517246e-01 -2.94004530e-01 4.59540725e-01
-1.37443471e+00 3.39534253e-01 -2.65484989e-01 8.63665640e-01
-6.68337226e-01 -5.26160598e-02 6.19933903e-01 1.12911098e-01
4.90355849e-01 7.83780038e-01 -1.18661180e-01 -2.55446315e-01
-1.19774437e+00 5.28888822e-01 1.21127105e+00 1.10823226e+00
3.07156235e-01 -3.63321483e-01 -6.94120824e-01 6.08747661e-01
-1.34763539e-01 3.04491222e-01 3.80608201e-01 -1.11914575e+00
9.04953361e-01 8.45365405e-01 2.65599340e-01 -1.12524664e+00
-1.22739092e-01 -5.49875975e-01 -6.95671439e-01 -4.80206847e-01
-7.13062510e-02 -2.58758098e-01 -3.18297267e-01 1.03962564e+00
-1.38259098e-01 -4.05962169e-01 3.87310922e-01 1.26393870e-01
1.24840319e+00 7.51295567e-01 2.86291894e-02 -1.07566500e+00
1.50428891e+00 -7.74413526e-01 -1.34215033e+00 2.07905307e-01
4.63057250e-01 -1.18761957e+00 3.30118686e-01 3.91279429e-01
-1.47289240e+00 -4.44477916e-01 -9.71105695e-01 -1.65946662e-01
-5.26323140e-01 2.91403145e-01 3.93701494e-02 1.14568993e-01
-1.06925547e+00 5.71519554e-01 -1.40939415e-01 -5.69051623e-01
3.03465098e-01 -2.19536051e-02 -2.52683789e-01 7.76722431e-02
-8.20671737e-01 1.24236500e+00 3.81566972e-01 -2.13932469e-01
1.59121547e-02 -6.09332860e-01 -4.78784174e-01 1.20468289e-02
5.40345550e-01 -7.76787341e-01 1.34607577e+00 -4.37531710e-01
-8.47512543e-01 1.09988642e+00 -4.80833143e-01 -2.52931386e-01
2.91123301e-01 -1.20891005e-01 -5.67233026e-01 4.60979611e-01
4.95881766e-01 3.19563970e-02 4.15415108e-01 -1.39075875e+00
-7.47510970e-01 -2.79672176e-01 -5.58310598e-02 -5.23873083e-02
-2.13721111e-01 5.34423828e-01 -5.11259735e-01 -8.06002617e-01
-4.44585562e-01 -6.33572936e-01 -2.30221823e-01 -9.03911829e-01
-9.17270362e-01 -8.12297463e-01 5.15232384e-01 -8.46889913e-01
1.88400471e+00 -1.69411206e+00 4.30654883e-01 4.71848249e-02
2.47880012e-01 1.42750829e-01 -2.58614607e-02 9.49182510e-01
1.37118548e-01 7.25591063e-01 -4.25501876e-02 -5.31209558e-02
-1.48696616e-01 -1.30692065e-01 -3.44969124e-01 -1.41040385e-02
1.42362759e-01 7.31726944e-01 -9.95766699e-01 -9.34378862e-01
-1.77992031e-01 4.38363887e-02 3.36565316e-01 -1.25923418e-02
-6.89176172e-02 1.12456143e-01 -9.04440284e-01 4.71537381e-01
2.96421140e-01 2.11171228e-02 -2.75603503e-01 -8.20216537e-02
-6.78042531e-01 4.05482799e-01 -9.68434751e-01 1.37229514e+00
-4.49131966e-01 7.79524088e-01 -3.17014009e-02 -8.46163750e-01
1.00380313e+00 4.54730123e-01 4.09144491e-01 -3.22315484e-01
1.12703569e-01 5.88786066e-01 -7.53731355e-02 -6.09282494e-01
9.22104001e-01 1.57448336e-01 -2.03708827e-01 5.32319427e-01
4.18975055e-02 -4.18180287e-01 1.20540249e+00 9.03568506e-01
1.04641175e+00 -1.88295141e-01 5.78764915e-01 -4.75705892e-01
8.43863130e-01 4.04893428e-01 1.59015566e-01 7.77839243e-01
3.27341050e-01 5.34594953e-01 7.88879573e-01 -3.30456831e-02
-1.22405839e+00 -4.96728450e-01 -2.27628663e-01 2.81483561e-01
-2.29212701e-01 -8.90468001e-01 -1.04619753e+00 -4.32483405e-01
-1.05109878e-01 9.13047552e-01 -4.76203114e-01 1.99531376e-01
-6.02923989e-01 -3.37623715e-01 2.25416407e-01 2.16114804e-01
2.11518064e-01 -1.21660209e+00 -7.25849390e-01 3.23495477e-01
-3.33076149e-01 -8.82379711e-01 -3.85619015e-01 2.28272438e-01
-6.44158661e-01 -1.27982569e+00 -1.17474687e+00 -6.52514875e-01
6.34646535e-01 2.94912845e-01 1.37052274e+00 2.20031023e-01
-2.84656435e-01 6.69975817e-01 -4.86089498e-01 -8.86673748e-01
-9.33804095e-01 4.58340287e-01 -2.17276812e-01 -7.14996219e-01
2.26358607e-01 -2.56273896e-01 8.82775262e-02 -3.26899469e-01
-9.34259236e-01 -1.25371128e-01 5.89519382e-01 4.98434156e-01
6.14473104e-01 -1.91213503e-01 8.88332546e-01 -9.93954480e-01
1.45680737e+00 -4.61534619e-01 -5.03571928e-01 6.42627299e-01
-5.72681427e-01 1.26940399e-01 3.83332044e-01 -9.31081995e-02
-1.07173467e+00 -4.57885951e-01 1.43868193e-01 1.76975027e-01
-2.57819444e-01 8.73126090e-01 -8.24359581e-02 1.72308013e-01
5.96822262e-01 -4.58372533e-02 -2.54813343e-01 -6.39567137e-01
2.50865012e-01 1.09753942e+00 5.84349096e-01 -5.87994516e-01
6.18749738e-01 -2.84364939e-01 3.81847262e-01 -9.36806679e-01
-1.04917145e+00 -7.29596972e-01 -9.55504417e-01 -3.41082364e-01
6.20829225e-01 -4.36468154e-01 -3.26634906e-02 -4.42974746e-01
-1.76112700e+00 5.03583729e-01 -4.67690736e-01 1.44367024e-01
-1.12898350e-01 4.82009500e-01 -1.18186280e-01 -8.33611548e-01
-6.82073832e-01 -8.85835469e-01 1.07650435e+00 5.82800388e-01
-8.53934050e-01 -9.97010469e-01 2.80546665e-01 3.85416150e-01
1.06483951e-01 4.96515810e-01 8.96477997e-01 -8.18672001e-01
1.99558996e-02 -3.86754096e-01 -1.68185774e-02 1.03845522e-01
2.45484337e-01 6.04915798e-01 -6.20750129e-01 4.70520109e-02
-2.54709095e-01 1.68780889e-02 1.12843931e+00 4.51359689e-01
9.95435894e-01 -6.46291375e-01 -7.32547462e-01 -2.50078022e-01
1.25497031e+00 3.68570864e-01 4.73509759e-01 5.23279309e-01
5.03009498e-01 9.54644024e-01 4.68100011e-01 2.68859327e-01
2.98668653e-01 4.08371627e-01 -1.59275457e-01 3.56279552e-01
-1.17250919e-01 2.66117573e-01 -1.67240933e-01 1.38323951e+00
-2.73943841e-01 -7.34175205e-01 -1.02072418e+00 8.27189445e-01
-1.59792721e+00 -1.25503564e+00 -7.98915505e-01 1.82655632e+00
8.71698260e-01 2.15054318e-01 1.28642619e-01 3.69285703e-01
8.85777593e-01 1.23231202e-01 -6.27560820e-03 -9.15698469e-01
-1.67989135e-01 1.30255222e-01 2.98666179e-01 4.71597344e-01
-7.48932242e-01 4.64087635e-01 6.61103773e+00 6.68995917e-01
-5.21568954e-01 -1.99183539e-01 2.65737236e-01 2.20219344e-01
-4.92419243e-01 3.55672240e-02 -1.08660567e+00 4.93045360e-01
1.30921996e+00 -1.04484355e+00 -2.96905398e-01 5.55256009e-01
4.62785512e-01 -7.24333882e-01 -8.88869822e-01 6.86428130e-01
2.43444309e-01 -1.73604906e+00 3.74024361e-01 8.64001587e-02
9.05289710e-01 -5.33083737e-01 -4.38179314e-01 -2.57850766e-01
8.08964595e-02 -5.22492170e-01 6.87970042e-01 6.85160995e-01
6.62299633e-01 -7.07213044e-01 1.06381440e+00 3.63451332e-01
-6.53754294e-01 2.06033364e-01 -2.40768105e-01 -4.75275479e-02
1.71595454e-01 8.09684515e-01 -2.80802995e-01 1.27185619e+00
4.50737864e-01 8.57716560e-01 -8.10293376e-01 1.20212400e+00
-2.49452397e-01 6.67911649e-01 2.36761585e-01 -4.63386655e-01
2.20085546e-01 -1.34655520e-01 7.91143656e-01 1.61012566e+00
3.85070443e-01 2.78036445e-01 6.05374807e-03 7.41872787e-01
-3.79339665e-01 3.30805957e-01 -7.10692286e-01 -4.55192149e-01
6.04784429e-01 1.36971772e+00 -1.07668281e+00 -6.26516283e-01
-1.21785887e-02 8.07088196e-01 -9.22960490e-02 1.78990960e-01
-1.17484052e-02 -1.19332337e+00 -1.00213543e-01 -9.71017629e-02
-4.03106138e-02 1.75388679e-01 -4.30399269e-01 -9.43494499e-01
4.18519050e-01 -5.64186215e-01 1.80595219e-01 -6.85857177e-01
-1.22919464e+00 7.61451423e-01 2.20592856e-01 -1.07324731e+00
-3.23942006e-01 -2.66218185e-01 -9.79493141e-01 1.20739579e+00
-1.19835222e+00 -6.44903839e-01 -3.07737529e-01 -2.50097096e-01
7.75600910e-01 -3.09590518e-01 7.97740400e-01 -5.81763908e-02
-7.94933796e-01 -8.86498690e-02 2.18444377e-01 -2.91764557e-01
5.86814821e-01 -1.41825461e+00 3.33067328e-01 8.69199932e-01
-1.10546593e-02 9.14601028e-01 9.43999767e-01 -9.76826131e-01
-1.18281889e+00 -8.41876328e-01 1.68688333e+00 -6.33547068e-01
8.59786451e-01 1.70497358e-01 -1.01873410e+00 9.42742079e-02
9.39221621e-01 -7.97788620e-01 7.19962418e-01 -1.29718315e-02
3.24241638e-01 5.18370457e-02 -7.99101174e-01 5.20010591e-01
8.71624768e-01 -1.51809812e-01 -1.32642412e+00 6.04858577e-01
6.68336332e-01 -3.02014291e-01 -8.89621317e-01 -1.77188247e-01
6.56230897e-02 -3.96632642e-01 5.98904312e-01 -3.31237972e-01
9.99025881e-01 -8.13488960e-02 5.41753292e-01 -1.33005631e+00
-2.76462346e-01 -8.69681835e-01 -1.26142830e-01 2.00043201e+00
5.15176952e-01 -1.46269083e-01 1.69181198e-01 4.25665766e-01
-2.96647608e-01 -5.71911991e-01 -7.67843723e-01 -6.94885135e-01
2.91234910e-01 2.17095222e-02 1.58232644e-01 7.77776599e-01
3.72395992e-01 7.09091783e-01 1.44274071e-01 -4.30663884e-01
6.29124641e-01 1.79561764e-01 5.31570137e-01 -1.78913999e+00
5.55551112e-01 -9.64477062e-01 -6.13926016e-02 -3.08740467e-01
3.36433291e-01 -7.86236703e-01 -1.87874362e-01 -2.52250791e+00
5.19556820e-01 9.91484746e-02 4.46964763e-02 3.38739865e-02
-3.41003627e-01 -1.92169487e-01 1.45790800e-01 4.54229742e-01
-7.60572195e-01 -1.25397891e-01 1.18759346e+00 -1.97682187e-01
-1.67546004e-01 3.97872962e-02 -1.17623532e+00 4.99250412e-01
7.77823389e-01 -5.56228399e-01 -6.14029095e-02 -7.82007258e-03
2.45597288e-01 9.73018035e-02 6.23775087e-03 -8.16346765e-01
3.80739212e-01 -2.93218315e-01 1.87017143e-01 -1.11183178e+00
-3.15606683e-01 -3.48681152e-01 8.32596719e-02 2.70227849e-01
-8.21823895e-01 5.42025030e-01 3.10368329e-01 3.66890430e-01
-3.34251761e-01 -9.74611580e-01 1.41454309e-01 -4.39542860e-01
-1.57468081e-01 -3.51689667e-01 -5.81707418e-01 1.92619666e-01
9.51431334e-01 -2.75852650e-01 -4.92043525e-01 3.25503126e-02
-1.70764953e-01 3.84927809e-01 4.89309013e-01 4.49488968e-01
3.83193076e-01 -8.26083839e-01 -1.02829075e+00 -7.35141456e-01
1.93371490e-01 -1.71851397e-01 -1.63878322e-01 6.49572909e-01
-4.92227167e-01 9.29810286e-01 -2.02647403e-01 1.16974756e-01
-1.36771584e+00 6.42621756e-01 -4.22032833e-01 -4.60947752e-01
-5.14763951e-01 3.87387514e-01 -4.19369191e-01 3.80386859e-01
3.29302877e-01 -4.22326416e-01 -8.85707617e-01 5.08292079e-01
9.78101552e-01 6.97274029e-01 2.15829968e-01 -7.16912329e-01
-3.12491238e-01 5.20872355e-01 -6.96650967e-02 -2.93258309e-01
1.37999034e+00 -3.00688475e-01 -6.90912902e-01 1.02667987e+00
9.40003574e-01 3.40169996e-01 -2.16716528e-01 3.62152234e-02
8.24985743e-01 9.73946974e-02 2.06113279e-01 -9.99020875e-01
-3.47630918e-01 6.52494669e-01 -2.77088135e-01 7.39349604e-01
8.29445183e-01 2.07962885e-01 2.62683094e-01 4.25800115e-01
1.58690333e-01 -1.14262247e+00 -3.16744119e-01 3.61543208e-01
1.22154784e+00 -9.91470814e-01 6.19730651e-01 -4.10628915e-01
-4.11435068e-01 1.61432540e+00 1.49905086e-01 2.19492495e-01
1.59881264e-01 2.52806276e-01 -1.15424052e-01 -4.01829123e-01
-8.36271226e-01 7.16292188e-02 6.25465095e-01 3.39691788e-01
8.13678324e-01 -3.87985617e-01 -1.32334638e+00 5.57305038e-01
1.69937480e-02 5.78962155e-02 1.09519339e+00 1.15035129e+00
-4.74837154e-01 -1.17989528e+00 -6.13602281e-01 8.81971776e-01
-1.02534163e+00 3.04142218e-02 -1.23808920e+00 5.17848074e-01
-1.23141892e-01 1.14020467e+00 -3.84164415e-02 2.31762663e-01
6.52335346e-01 6.03433140e-02 4.27935719e-01 -7.96405852e-01
-1.10517812e+00 6.38071820e-02 4.93340880e-01 1.94146946e-01
-8.78611684e-01 -8.38810265e-01 -1.07373464e+00 -2.74633676e-01
-3.32162112e-01 9.47963715e-01 8.20176423e-01 1.10948515e+00
4.52956498e-01 1.05640769e+00 2.56102890e-01 -8.28452170e-01
-2.19307125e-01 -1.24209547e+00 -4.03136790e-01 2.62434632e-01
3.59957546e-01 -3.42591733e-01 -3.80439132e-01 5.34303963e-01] | [12.348226547241211, 9.535038948059082] |
e1b3a521-a9d0-4f75-8806-4cdfa85fbb34 | depth-estimation-by-learning-triangulation | 2003.08933 | null | https://arxiv.org/abs/2003.08933v2 | https://arxiv.org/pdf/2003.08933v2.pdf | DELTAS: Depth Estimation by Learning Triangulation And densification of Sparse points | Multi-view stereo (MVS) is the golden mean between the accuracy of active depth sensing and the practicality of monocular depth estimation. Cost volume based approaches employing 3D convolutional neural networks (CNNs) have considerably improved the accuracy of MVS systems. However, this accuracy comes at a high computational cost which impedes practical adoption. Distinct from cost volume approaches, we propose an efficient depth estimation approach by first (a) detecting and evaluating descriptors for interest points, then (b) learning to match and triangulate a small set of interest points, and finally (c) densifying this sparse set of 3D points using CNNs. An end-to-end network efficiently performs all three steps within a deep learning framework and trained with intermediate 2D image and 3D geometric supervision, along with depth supervision. Crucially, our first step complements pose estimation using interest point detection and descriptor learning. We demonstrate state-of-the-art results on depth estimation with lower compute for different scene lengths. Furthermore, our method generalizes to newer environments and the descriptors output by our network compare favorably to strong baselines. Code is available at https://github.com/magicleap/DELTAS | ['Andrew Rabinovich', 'Vijay Badrinarayanan', 'Zak Murez', 'James Bartolozzi', 'Ayan Sinha'] | 2020-03-19 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3649_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123660103.pdf | eccv-2020-8 | ['interest-point-detection'] | ['computer-vision'] | [ 5.29916510e-02 -4.49434817e-02 -7.33449161e-02 -4.00412381e-01
-1.07475555e+00 -5.74045360e-01 7.08703935e-01 7.20872208e-02
-5.87718308e-01 4.10844803e-01 1.69065982e-01 -5.08128013e-03
2.80907929e-01 -8.20239723e-01 -7.91544080e-01 -3.91989499e-01
6.52331337e-02 5.80934227e-01 4.12716895e-01 1.26436546e-01
5.71194887e-01 8.10771942e-01 -1.81951559e+00 -3.71856568e-03
4.83627707e-01 1.40086102e+00 1.81533724e-01 7.89768577e-01
1.12870678e-01 7.53487468e-01 -4.30275463e-02 -1.21982887e-01
5.43610096e-01 2.45966569e-01 -7.22565830e-01 -2.80851368e-02
1.20484412e+00 -1.15673852e+00 -5.17828405e-01 6.51416957e-01
6.91909850e-01 1.12129999e-02 3.97326291e-01 -9.32035387e-01
-1.96792752e-01 -2.45709762e-01 -6.72302008e-01 6.67624101e-02
6.46630645e-01 1.76654369e-01 9.78238940e-01 -1.37751961e+00
7.51198649e-01 1.11563432e+00 7.81447053e-01 4.95830715e-01
-9.97656167e-01 -4.61113304e-01 7.76721984e-02 -1.02253929e-01
-1.44451833e+00 -8.22476566e-01 8.00690711e-01 -4.36267674e-01
1.32061708e+00 -2.27733850e-01 9.22580481e-01 8.40861380e-01
7.23168813e-03 9.18736339e-01 6.31507218e-01 -1.74982727e-01
2.62252778e-01 -2.95528561e-01 -1.38510451e-01 8.35922897e-01
1.13471352e-01 2.61670411e-01 -6.52174473e-01 -4.95033786e-02
1.04664409e+00 2.56472588e-01 -1.63605839e-01 -9.70957994e-01
-1.12724411e+00 8.31367552e-01 5.76339066e-01 -2.08409727e-01
-3.14591616e-01 5.13826251e-01 2.60297954e-01 -8.23242683e-03
6.57577813e-01 1.85131669e-01 -4.94521111e-01 -3.24368685e-01
-9.88162100e-01 3.59894276e-01 4.25124317e-01 9.73871648e-01
1.10938919e+00 -2.70363748e-01 4.51209962e-01 6.82227194e-01
2.51907080e-01 4.73803908e-01 -2.21835058e-02 -1.48497438e+00
5.57833850e-01 7.04533935e-01 1.36719242e-01 -7.39449382e-01
-4.28458095e-01 -2.35560000e-01 -4.91542220e-01 6.01213574e-01
5.55342615e-01 2.40164604e-02 -8.61024559e-01 1.32575452e+00
5.33659756e-01 -1.38842553e-01 -2.26566523e-01 9.28411007e-01
7.86066115e-01 2.50610441e-01 -4.46534187e-01 4.17661369e-01
8.41059446e-01 -8.33647132e-01 7.78739676e-02 -5.61065316e-01
7.60616481e-01 -7.01681495e-01 7.87074685e-01 5.32849848e-01
-1.38091207e+00 -3.54414970e-01 -1.04431415e+00 -8.93092811e-01
-2.80103445e-01 3.66852954e-02 7.54081309e-01 3.15313756e-01
-1.46242118e+00 6.70772672e-01 -1.03379512e+00 -3.31291467e-01
7.40035236e-01 6.76332593e-01 -5.84759712e-01 -4.65960413e-01
-6.35748744e-01 6.41409874e-01 1.11804031e-01 6.62640482e-02
-9.53320265e-01 -8.33005846e-01 -1.24589741e+00 -2.28605106e-01
1.35714114e-01 -1.08116424e+00 1.24980712e+00 -4.88194466e-01
-1.24231958e+00 1.23012924e+00 -3.51978481e-01 -3.37648302e-01
7.64748812e-01 -5.36944091e-01 4.01348710e-01 4.68088835e-01
3.24304849e-01 1.22182047e+00 5.01713753e-01 -1.15768397e+00
-8.76605034e-01 -7.46923089e-01 3.78454179e-01 4.83943403e-01
5.29718511e-02 -4.06468451e-01 -5.96830308e-01 1.28215700e-01
8.15710783e-01 -7.07500935e-01 -1.07288122e-01 7.57800996e-01
-2.64667034e-01 -3.81807387e-02 7.77404666e-01 -3.34522367e-01
5.05921364e-01 -2.10255480e+00 5.94190042e-03 5.65057360e-02
6.10476315e-01 -1.61793381e-01 1.11550249e-01 2.96563685e-01
1.76028460e-01 -2.50573575e-01 1.08946323e-01 -8.65979612e-01
-1.41691074e-01 -5.00443168e-02 -4.29658592e-02 7.97476172e-01
7.14241564e-02 8.28006446e-01 -9.75795507e-01 -3.60349178e-01
8.43048990e-01 7.62744606e-01 -8.69973063e-01 5.27053624e-02
-1.42296404e-02 3.72136384e-01 -1.67825446e-01 1.09150803e+00
7.67533898e-01 -4.35931414e-01 -2.37625763e-01 -3.21104437e-01
-3.65248293e-01 5.68036437e-01 -1.21132183e+00 2.17872357e+00
-6.01690650e-01 7.17055857e-01 4.33658622e-02 -6.70347393e-01
8.10351610e-01 -6.96319491e-02 7.62038171e-01 -6.37395978e-01
8.40022340e-02 4.62464482e-01 -4.79787111e-01 -1.75644293e-01
6.53887391e-01 3.34858373e-02 2.25512683e-01 1.99435622e-01
7.45382980e-02 -5.99805236e-01 -2.16785267e-01 1.91800505e-01
1.07829750e+00 5.25229812e-01 3.95928502e-01 1.00971743e-01
2.38151655e-01 -6.24468997e-02 2.77602792e-01 3.79578561e-01
-2.73145020e-01 9.78376448e-01 3.05155009e-01 -7.14557469e-01
-1.29669261e+00 -1.20008063e+00 -2.69334197e-01 6.16427481e-01
3.59233946e-01 -1.18743941e-01 -3.02834690e-01 -3.41273189e-01
3.51266325e-01 1.17386200e-01 -6.76622629e-01 2.84768105e-01
-5.10509789e-01 -4.33664676e-03 2.63736308e-01 8.49738896e-01
6.47163868e-01 -5.09866774e-01 -1.00775683e+00 2.05066297e-02
-6.07182123e-02 -1.23866856e+00 -4.13307697e-01 4.23715800e-01
-1.13933527e+00 -1.11084473e+00 -7.66797960e-01 -5.95568120e-01
7.00742662e-01 6.35394216e-01 1.17455363e+00 -5.02789253e-03
-1.60102680e-01 5.65127194e-01 -9.58931893e-02 -1.68700576e-01
3.41210425e-01 2.07471564e-01 -7.19972551e-02 -4.81144249e-01
4.37839776e-01 -8.65207791e-01 -1.23305893e+00 7.75081813e-02
-5.84647477e-01 8.42616707e-02 5.17272353e-01 5.52469432e-01
8.76262128e-01 -6.47709608e-01 -1.93098962e-01 -4.85855669e-01
-2.23140821e-01 -1.28408387e-01 -9.65054810e-01 -3.41755152e-01
-4.54834223e-01 -1.09886393e-01 1.62860841e-01 6.20958731e-02
-5.88752151e-01 6.19705319e-01 -2.77425706e-01 -7.09294200e-01
-1.41187325e-01 4.48470451e-02 -8.52262825e-02 -4.65447128e-01
7.13240683e-01 2.27102548e-01 1.49242088e-01 -2.46449918e-01
2.09070474e-01 4.68276799e-01 5.10276914e-01 -3.56924832e-01
7.05616772e-01 1.13820565e+00 2.53594071e-01 -8.28427196e-01
-9.48680520e-01 -7.58946478e-01 -1.20194399e+00 -2.68861264e-01
6.96161985e-01 -1.54174995e+00 -7.72007227e-01 5.80012619e-01
-1.28957510e+00 -4.20169801e-01 -1.74123615e-01 5.59437096e-01
-7.90257156e-01 4.46895897e-01 -6.13939762e-01 -7.90054262e-01
-3.24454367e-01 -1.05981469e+00 1.78400338e+00 -1.16077319e-01
-2.23478228e-01 -1.00399077e+00 2.45081540e-02 5.76679885e-01
1.61103427e-01 5.44421196e-01 2.22195834e-01 -7.93637484e-02
-1.11607599e+00 -3.32680196e-01 -3.82598132e-01 2.05804542e-01
-1.24222197e-01 -3.07758208e-02 -1.41970575e+00 -3.56033146e-01
-1.80204034e-01 -6.35663390e-01 8.71317148e-01 7.26124227e-01
9.39674318e-01 2.07609549e-01 -3.16701055e-01 1.20574570e+00
1.64212632e+00 -3.85618545e-02 7.07948506e-01 6.08547330e-01
9.70585704e-01 6.02741003e-01 5.58043599e-01 5.67173004e-01
8.18713307e-01 6.16702020e-01 8.41684878e-01 -1.82510585e-01
-9.00498182e-02 -3.42754811e-01 8.30856413e-02 4.74951088e-01
-1.38193928e-02 1.85960695e-01 -1.09253931e+00 7.45347381e-01
-1.50833333e+00 -8.27210009e-01 -2.16660625e-03 2.25303411e+00
4.90915745e-01 2.09243849e-01 1.23285115e-01 2.05579102e-01
2.41007835e-01 2.57543176e-01 -9.14962769e-01 -6.43228740e-02
-2.23517329e-01 1.05834059e-01 7.37638295e-01 7.24168658e-01
-1.18800271e+00 9.01230156e-01 5.39071608e+00 3.60012352e-01
-1.25252759e+00 -1.22561082e-01 7.35507190e-01 -5.78518271e-01
-2.84593999e-01 -3.28240804e-02 -8.71135592e-01 3.67659591e-02
3.97409439e-01 2.30188087e-01 2.78664559e-01 1.11611295e+00
1.22051962e-01 -5.24224818e-01 -1.51438498e+00 1.48030603e+00
9.04190019e-02 -1.45011103e+00 -1.87237963e-01 4.30474758e-01
8.64756405e-01 7.17091024e-01 -7.04324171e-02 -1.46226168e-01
1.41374111e-01 -8.67094815e-01 8.58025908e-01 1.94409505e-01
1.04886413e+00 -7.57361710e-01 6.12073660e-01 2.11134791e-01
-1.27013457e+00 -1.43727036e-02 -4.97908562e-01 -3.19425642e-01
-3.77100147e-02 7.44390070e-01 -6.51286125e-01 2.57629067e-01
7.13878572e-01 1.04691231e+00 -3.99824739e-01 9.83984768e-01
-7.12300986e-02 -2.65236408e-01 -5.76661468e-01 1.40631825e-01
3.16623002e-01 1.02297060e-01 2.01211050e-01 7.64008760e-01
3.76020253e-01 2.75646616e-02 -1.43009603e-01 7.79891431e-01
-1.79060906e-01 -1.59071162e-01 -1.04215372e+00 3.87565315e-01
5.92432201e-01 1.14401615e+00 -6.41710162e-01 -1.99003115e-01
-5.56281924e-01 9.84687209e-01 6.14435673e-01 1.45842731e-02
-4.13754761e-01 -3.19839060e-01 7.72422552e-01 4.84467179e-01
3.15061152e-01 -4.92586046e-01 -6.71228230e-01 -1.21014833e+00
3.35541695e-01 -3.00022751e-01 1.03213340e-01 -9.16476727e-01
-9.42227304e-01 2.32643276e-01 -3.46843362e-01 -1.51762962e+00
-3.42618704e-01 -6.51459932e-01 -2.98309833e-01 8.73235881e-01
-1.85981953e+00 -1.01498222e+00 -7.80449092e-01 5.99475741e-01
5.48870385e-01 3.00904661e-01 5.75335741e-01 3.52763355e-01
-5.22767417e-02 3.46634090e-01 -4.27609384e-02 2.13663906e-01
5.63467443e-01 -1.22105587e+00 7.29024768e-01 7.28520274e-01
-1.25276491e-01 4.86023575e-01 1.94784030e-01 -4.09477353e-01
-1.55390811e+00 -9.60815907e-01 8.19642663e-01 -7.66884804e-01
3.77110541e-01 -5.32706738e-01 -5.65105379e-01 6.86365187e-01
-3.17315072e-01 3.25109422e-01 2.08801717e-01 6.88182116e-02
-4.22716737e-01 -1.25783443e-01 -1.18337333e+00 4.22073990e-01
1.38131976e+00 -8.30764234e-01 -1.88253000e-01 3.65147918e-01
4.53309298e-01 -9.70944464e-01 -8.34190428e-01 3.98056895e-01
9.36978757e-01 -1.49796820e+00 1.33103776e+00 1.81895614e-01
9.06963527e-01 -1.73507407e-01 -4.29895848e-01 -7.90468931e-01
-2.13521663e-02 -2.64889956e-01 -1.48957565e-01 6.18955791e-01
1.71510726e-01 -5.56322873e-01 1.34307146e+00 6.68811560e-01
-2.64862239e-01 -1.03006184e+00 -1.01128781e+00 -5.69198251e-01
9.56368819e-02 -5.61454833e-01 2.74207681e-01 6.99305117e-01
-2.48091817e-01 1.35935009e-01 -4.89061885e-02 2.35037640e-01
9.41601872e-01 1.16329834e-01 1.05325210e+00 -1.30407155e+00
-4.78515849e-02 -3.42272967e-01 -7.34644830e-01 -1.67282045e+00
-1.59102082e-01 -6.07846856e-01 -6.19824976e-02 -1.72095323e+00
1.67848989e-02 -3.92871320e-01 2.51290083e-01 2.55335689e-01
1.89430416e-01 4.71217662e-01 -1.59686655e-01 4.18195426e-01
-6.34945035e-01 4.56432790e-01 1.14898515e+00 1.80337653e-01
-7.73321986e-02 -2.44949147e-01 -4.96420622e-01 8.41414809e-01
6.61247492e-01 -1.18436240e-01 -3.91390741e-01 -8.18173468e-01
2.69764990e-01 1.75286293e-01 7.02730179e-01 -1.22768581e+00
4.12092179e-01 1.22526370e-01 7.78547227e-01 -1.09863889e+00
8.80387366e-01 -6.52169704e-01 -2.29838476e-01 4.78799164e-01
1.87038053e-02 3.17098238e-02 1.06193483e-01 3.29856187e-01
-1.07102066e-01 1.22303560e-01 9.42011952e-01 -3.92750502e-01
-9.21467066e-01 7.12979138e-01 2.45087355e-01 1.02910377e-01
8.68880570e-01 -8.85676265e-01 -2.77828157e-01 -4.39910918e-01
-4.19955075e-01 1.14687510e-01 9.79770899e-01 1.05726764e-01
9.41336393e-01 -1.14688194e+00 -4.50617522e-01 3.16948324e-01
3.13219070e-01 7.87786543e-01 2.35501796e-01 7.91506827e-01
-1.04876661e+00 5.42496622e-01 -7.13181123e-02 -1.10242558e+00
-1.07510805e+00 1.55630365e-01 5.53147376e-01 9.67839658e-02
-7.35320330e-01 1.14979649e+00 3.16502035e-01 -5.08083105e-01
4.59381849e-01 -3.28761131e-01 1.54757813e-01 -5.06051779e-02
3.54633808e-01 4.51789021e-01 2.42082670e-01 -4.98487711e-01
-5.02786398e-01 1.07729483e+00 -7.35929534e-02 -1.29266366e-01
1.55766761e+00 -3.13504249e-01 7.85839111e-02 2.68804640e-01
1.65058732e+00 -1.11986183e-01 -1.83228445e+00 -2.57020801e-01
-2.19165295e-01 -7.62327492e-01 3.69803369e-01 -3.35973054e-01
-1.09342384e+00 1.07597160e+00 5.65648317e-01 -3.91143113e-01
9.36155498e-01 3.38881910e-02 8.10627162e-01 3.69743615e-01
6.38028800e-01 -1.00176978e+00 1.54569417e-01 5.04130900e-01
6.37291253e-01 -1.66597390e+00 2.12629840e-01 -3.77228469e-01
-2.04348087e-01 1.20492458e+00 6.79046929e-01 -2.93113858e-01
6.86237752e-01 3.99039268e-01 1.70956984e-01 -5.16146898e-01
-6.02431059e-01 -1.15714200e-01 1.61569454e-02 6.36366844e-01
3.76616627e-01 -3.82364839e-01 3.47790092e-01 -3.13937575e-01
-2.97120482e-01 8.69456381e-02 3.86949599e-01 9.87533391e-01
-4.58600610e-01 -7.25350738e-01 -1.74439967e-01 3.72549921e-01
-2.46824980e-01 -2.79642493e-01 -2.92176038e-01 9.21521187e-01
1.56083077e-01 7.03356266e-01 3.62545341e-01 -4.11185205e-01
4.07634556e-01 -3.22592765e-01 6.31071508e-01 -7.24913776e-01
-1.94141462e-01 -4.92820479e-02 -2.27208510e-02 -1.08863556e+00
-3.82202297e-01 -9.29725289e-01 -1.06531131e+00 -5.96486151e-01
-3.50207761e-02 -5.17843008e-01 7.60140181e-01 7.11864173e-01
4.83487427e-01 -8.37851316e-03 6.96887255e-01 -1.37830019e+00
-2.59814471e-01 -5.94519556e-01 -4.06487525e-01 1.89007744e-01
5.94310880e-01 -6.72419190e-01 -5.05604148e-01 -3.00714344e-01] | [8.571086883544922, -2.66267991065979] |
8f03db2f-09a7-4b8b-adf5-c0a59dce4b35 | order-sensitive-neural-constituency-parsing | 2211.00421 | null | https://arxiv.org/abs/2211.00421v1 | https://arxiv.org/pdf/2211.00421v1.pdf | Order-sensitive Neural Constituency Parsing | We propose a novel algorithm that improves on the previous neural span-based CKY decoder for constituency parsing. In contrast to the traditional span-based decoding, where spans are combined only based on the sum of their scores, we introduce an order-sensitive strategy, where the span combination scores are more carefully derived from an order-sensitive basis. Our decoder can be regarded as a generalization over existing span-based decoder in determining a finer-grain scoring scheme for the combination of lower-level spans into higher-level spans, where we emphasize on the order of the lower-level spans and use order-sensitive span scores as well as order-sensitive combination grammar rule scores to enhance prediction accuracy. We implement the proposed decoding strategy harnessing GPU parallelism and achieve a decoding speed on par with state-of-the-art span-based parsers. Using the previous state-of-the-art model without additional data as our baseline, we outperform it and improve the F1 score on the Penn Treebank Dataset by 0.26% and on the Chinese Treebank Dataset by 0.35%. | ['Cong Liu', 'Liyin Xiao', 'Tianyu Shi', 'Zhicheng Wang'] | 2022-11-01 | null | null | null | null | ['constituency-parsing'] | ['natural-language-processing'] | [ 2.80388296e-01 4.12843227e-01 -3.69790033e-03 -4.78966355e-01
-1.38975549e+00 -8.09936821e-01 6.67116642e-02 6.67639077e-01
-6.29649460e-01 7.27415204e-01 4.49889809e-01 -6.71065688e-01
1.94712371e-01 -8.09225440e-01 -7.45733261e-01 -4.49546099e-01
6.62501082e-02 4.80723768e-01 7.63943791e-01 -5.71120739e-01
3.09561044e-01 1.56802565e-01 -1.25248492e+00 5.24646938e-01
8.53253841e-01 7.78986216e-01 2.19467729e-01 8.39158952e-01
-2.06645340e-01 4.35454100e-01 -5.10344684e-01 -6.44114017e-01
7.93666765e-03 -5.31669855e-01 -7.48051465e-01 -6.11430228e-01
6.19276047e-01 -2.04497367e-01 -7.55445883e-02 9.13453460e-01
3.64594877e-01 -2.02629134e-01 8.90459344e-02 -2.04481632e-01
-2.64766306e-01 1.02588296e+00 -4.23100889e-01 2.99441934e-01
3.53827327e-01 -2.94686079e-01 1.44957864e+00 -4.55959946e-01
5.13291895e-01 9.94181812e-01 5.89942038e-01 5.81325650e-01
-1.13433123e+00 -3.32218498e-01 3.98067087e-01 4.19842154e-02
-7.45157480e-01 -3.87429804e-01 6.13864660e-01 -1.11609675e-01
1.39200389e+00 1.82291210e-01 5.63766658e-01 5.67669213e-01
4.37378705e-01 7.74218023e-01 1.05151701e+00 -9.41387951e-01
1.83234856e-01 -6.51275814e-01 6.94426835e-01 9.75939810e-01
2.30976328e-01 2.13592604e-01 -7.26801038e-01 -7.64810741e-02
3.72498244e-01 -6.89622700e-01 -2.79855896e-02 1.92601904e-01
-1.10940444e+00 1.02968657e+00 -1.68404784e-02 4.57897633e-01
-1.03268676e-01 2.44930834e-01 6.90830410e-01 1.82747915e-01
4.82839823e-01 2.78931767e-01 -7.46689439e-01 -4.29269582e-01
-1.32254696e+00 4.15405005e-01 1.02783024e+00 8.01862240e-01
6.46567047e-01 -1.38884634e-01 -4.74746674e-01 6.50536835e-01
1.27594426e-01 -1.28783375e-01 2.17745662e-01 -7.97696352e-01
9.23789620e-01 5.24647415e-01 -1.88063577e-01 -2.18576014e-01
-5.16392469e-01 -3.97436112e-01 -1.59334630e-01 1.98697224e-02
8.76552045e-01 -3.10909629e-01 -8.04534793e-01 2.11127520e+00
2.45758221e-01 3.54709141e-02 6.21248558e-02 7.95200109e-01
5.38670361e-01 7.66405225e-01 3.28591973e-01 -4.15280491e-01
1.89365804e+00 -7.79376447e-01 -5.57108521e-01 -4.38629985e-02
1.02173984e+00 -7.54690409e-01 1.01340163e+00 5.41099608e-01
-1.41990030e+00 -3.77988935e-01 -1.40437603e+00 -3.15710783e-01
9.95853096e-02 1.69163451e-01 7.86959946e-01 7.87626684e-01
-1.14763367e+00 9.49301660e-01 -9.84050512e-01 -3.45408805e-02
-8.73903334e-02 1.70744002e-01 -2.08151028e-01 3.85561019e-01
-1.26445115e+00 9.15656924e-01 6.66175961e-01 -9.65217203e-02
-3.65511298e-01 -8.05877149e-01 -9.09111559e-01 1.85486928e-01
2.47074485e-01 -3.00084829e-01 1.39735663e+00 -2.46162757e-01
-1.72610760e+00 9.64884520e-01 -2.12711424e-01 -6.02501750e-01
1.87916160e-01 -3.99947017e-01 -2.02942684e-01 1.77254990e-01
-1.00680433e-01 5.09670079e-01 1.29258618e-01 -5.94503522e-01
-8.15592527e-01 -2.72368491e-01 2.58518577e-01 1.11911088e-01
-7.63592348e-02 4.71799910e-01 -3.62394720e-01 -5.86502671e-01
2.81792521e-01 -9.01893497e-01 -2.13590577e-01 -3.65376830e-01
-2.72748351e-01 -4.47729409e-01 2.72185296e-01 -1.04234660e+00
1.81765401e+00 -1.81367624e+00 1.39534608e-01 -2.99874336e-01
4.26005200e-02 4.15971458e-01 -9.81388241e-02 5.80998898e-01
1.10378183e-01 -2.29953498e-01 -6.31153524e-01 -3.61302525e-01
3.20955105e-02 1.73598453e-01 -7.57936314e-02 2.34474182e-01
4.54927206e-01 5.31647742e-01 -1.00332642e+00 -5.38030744e-01
-1.25048131e-01 1.68531477e-01 -7.07345605e-01 7.69115314e-02
-3.97283524e-01 2.66514122e-01 -2.62212127e-01 4.16216165e-01
6.29240453e-01 3.76614422e-01 5.69933772e-01 3.25719416e-02
-5.13281167e-01 1.12071347e+00 -8.01427484e-01 2.12984061e+00
-5.45347393e-01 2.20085114e-01 4.81821178e-03 -8.62183154e-01
1.18483126e+00 3.59284610e-01 6.22828212e-03 -5.65627754e-01
-7.68585876e-02 4.49551493e-01 3.73112977e-01 1.37374708e-02
9.37647104e-01 -2.63074487e-01 -8.42236876e-01 3.98170292e-01
1.75696358e-01 1.39019921e-01 5.83443701e-01 1.29602015e-01
1.30522943e+00 6.10826492e-01 4.71370131e-01 -5.78149080e-01
6.07581019e-01 8.32709223e-02 8.26211929e-01 3.91886950e-01
-2.48796985e-01 5.97357929e-01 9.82358038e-01 -5.13284266e-01
-1.04921019e+00 -9.61127043e-01 -2.47305796e-01 1.56212747e+00
-2.53249288e-01 -5.92604101e-01 -1.22050738e+00 -9.35214579e-01
-4.35021996e-01 8.96208644e-01 -4.14272457e-01 1.99426174e-01
-1.45062613e+00 -9.68204737e-01 1.07321024e+00 7.22366273e-01
3.58484596e-01 -1.02380550e+00 -8.29492331e-01 7.88332462e-01
-9.07145962e-02 -1.20695066e+00 -5.57136834e-01 7.16709971e-01
-1.01770210e+00 -8.41648161e-01 -2.21490562e-01 -9.42618728e-01
1.78678364e-01 -5.65578759e-01 1.47031498e+00 5.78890257e-02
3.25472534e-01 -5.22572756e-01 -6.21114552e-01 -1.20367236e-01
-7.70503342e-01 5.17633021e-01 -3.06275755e-01 -3.88733417e-01
3.33615899e-01 -5.61061203e-01 -4.93968219e-01 -1.56533554e-01
-5.85467815e-01 1.24278374e-01 4.83134329e-01 9.06201303e-01
5.09712934e-01 -6.89518213e-01 7.17225552e-01 -1.10938787e+00
4.09796447e-01 -1.34788483e-01 -7.67484248e-01 3.80984128e-01
-4.18814421e-01 5.54736376e-01 8.63613248e-01 -1.69956475e-01
-1.00489569e+00 2.48585641e-01 -7.83452749e-01 4.24292564e-01
-4.46258485e-02 3.18513960e-01 -3.03610146e-01 1.89024374e-01
3.94217998e-01 9.14891362e-02 -5.01681149e-01 -6.03014588e-01
4.16137666e-01 3.30688000e-01 4.78877842e-01 -9.38572168e-01
9.56589580e-02 -8.66556838e-02 1.48227468e-01 -1.66173220e-01
-8.11129868e-01 -1.75713509e-01 -8.10038090e-01 6.22509681e-02
9.45778310e-01 -7.56143749e-01 -5.66228509e-01 1.31872103e-01
-1.62004268e+00 -1.85891330e-01 -7.67794997e-02 3.74032944e-01
-6.09022200e-01 5.52184045e-01 -1.17926192e+00 -6.11858606e-01
-7.05648065e-01 -1.06167114e+00 1.29395115e+00 3.85334082e-02
-3.54537934e-01 -8.57975304e-01 3.09431076e-01 5.60008325e-02
-4.49902266e-02 3.29704493e-01 1.16915834e+00 -6.90457463e-01
-2.63204008e-01 -7.79159600e-03 -1.57127038e-01 1.95652500e-01
-5.11647165e-01 -4.00285065e-01 -7.31599391e-01 -2.56027579e-02
-4.74089794e-02 -1.04838997e-01 1.21465576e+00 2.23219395e-01
5.45451641e-01 -9.05147791e-02 -6.63842410e-02 4.35721576e-01
1.69523335e+00 1.26235470e-01 5.91664195e-01 4.37370509e-01
4.99222547e-01 5.31213284e-01 6.12808526e-01 3.47120136e-01
4.99495983e-01 8.13359201e-01 3.27692300e-01 3.55718136e-01
-2.29635894e-01 -4.32472885e-01 5.45358717e-01 1.26518631e+00
-1.23204969e-01 -1.17330626e-01 -6.62162542e-01 7.01905608e-01
-1.84586108e+00 -5.73354006e-01 -3.48154604e-01 2.11471415e+00
1.17612123e+00 3.81583989e-01 2.83061326e-01 4.61017519e-01
7.31408954e-01 4.68862563e-01 -9.13895741e-02 -1.40869033e+00
1.26983762e-01 7.66330421e-01 6.59938455e-01 7.13588178e-01
-1.05475831e+00 1.39106834e+00 6.30004835e+00 8.52068007e-01
-8.94316137e-01 3.23877662e-01 4.84130263e-01 -1.27790838e-01
-3.74434590e-01 2.80419409e-01 -1.39024568e+00 4.47517574e-01
1.29412878e+00 2.82406986e-01 1.24152593e-01 6.85291171e-01
-9.23904777e-03 -1.46514580e-01 -1.20599747e+00 2.88787603e-01
-1.66304663e-01 -1.14601398e+00 -3.04462373e-01 -1.13915585e-01
3.18958700e-01 -8.74317065e-02 -3.63073170e-01 4.40200090e-01
4.73887026e-01 -5.71243346e-01 1.08047378e+00 1.82955220e-01
7.79487073e-01 -7.92096317e-01 6.12735331e-01 4.29955631e-01
-1.43878436e+00 1.68267459e-01 -4.68091846e-01 -4.78571773e-01
5.87304950e-01 7.00274706e-01 -5.36725163e-01 5.88224530e-01
3.06056619e-01 1.51320755e-01 -1.71250075e-01 5.81322491e-01
-6.99846864e-01 1.16001308e+00 -3.86037916e-01 -3.06496143e-01
4.57770109e-01 -7.99977779e-02 5.27280927e-01 1.70482874e+00
4.14721906e-01 1.68632880e-01 -4.19842452e-02 3.75248164e-01
-2.55244202e-03 2.06622958e-01 7.32306316e-02 4.80924904e-01
6.19300485e-01 1.02687836e+00 -6.08887076e-01 -3.13758403e-01
-4.99799788e-01 7.68687725e-01 8.97612631e-01 -3.06319237e-01
-8.29242229e-01 -3.52176070e-01 4.39971656e-01 -1.10215895e-01
8.48587096e-01 -5.11554599e-01 -7.16474950e-01 -9.39224303e-01
1.96173862e-01 -6.60397887e-01 5.65123081e-01 -9.25996602e-02
-7.82190859e-01 9.85440373e-01 -1.59644298e-02 -1.05250919e+00
-3.45340520e-01 -6.43313706e-01 -6.22098446e-01 9.66725707e-01
-1.76769078e+00 -1.10855448e+00 5.60754597e-01 5.21269813e-03
5.53191781e-01 1.19816974e-01 1.21061206e+00 -4.20455933e-02
-3.84428084e-01 9.57353532e-01 -1.57818630e-01 1.16746664e-01
4.16583180e-01 -1.37422323e+00 8.26956093e-01 1.31264913e+00
2.29568481e-01 5.30478358e-01 5.97283602e-01 -4.70149696e-01
-1.08873963e+00 -8.83166254e-01 1.67339873e+00 -4.33312148e-01
6.84643507e-01 -6.09311640e-01 -8.97196651e-01 4.95509535e-01
2.13962063e-01 -1.39251411e-01 6.24702871e-01 3.51070285e-01
-5.53798795e-01 -1.25441924e-01 -1.00074708e+00 4.17987436e-01
1.26500106e+00 -1.38581052e-01 -1.00381279e+00 -1.72350109e-01
9.31542456e-01 -7.12232530e-01 -1.00251782e+00 3.75160635e-01
6.86346471e-01 -1.18599570e+00 4.47522014e-01 -3.43715936e-01
5.42873859e-01 -2.09987551e-01 -3.26331645e-01 -1.15088332e+00
-4.46821272e-01 -7.07263827e-01 -1.81658179e-01 1.17434442e+00
6.91625118e-01 -3.07523966e-01 8.70237410e-01 1.17911376e-01
-7.35012054e-01 -9.99810755e-01 -1.39423585e+00 -6.20885670e-01
4.42817956e-01 -6.05695248e-01 7.04324245e-01 2.93712884e-01
4.14758116e-01 5.22227466e-01 -3.82966459e-01 2.00736448e-01
4.96009141e-01 4.01499450e-01 8.64399076e-02 -9.33848858e-01
-6.89479053e-01 -4.39629674e-01 -2.20929891e-01 -1.10775459e+00
1.07419834e-01 -9.80315626e-01 3.17908585e-01 -1.34282100e+00
-7.07696453e-02 -4.63232815e-01 -4.35542762e-01 5.57176113e-01
-4.44577277e-01 2.73217857e-01 5.32514691e-01 -1.17686227e-01
-5.35399199e-01 2.93391254e-02 1.08616781e+00 3.63815546e-01
-1.82574630e-01 -2.61685759e-01 -8.49431455e-01 5.30480325e-01
8.01348925e-01 -6.70275509e-01 1.67372122e-01 -6.79233968e-01
2.84996927e-01 5.00140011e-01 -3.50239217e-01 -9.46308911e-01
8.56795684e-02 2.28025466e-01 -2.17337564e-01 -7.41166949e-01
1.50080383e-01 -1.99666142e-01 -3.52454245e-01 6.76806092e-01
-4.29651439e-01 2.25253284e-01 3.36801916e-01 2.29451284e-01
-1.67552650e-01 -5.20343304e-01 5.39998531e-01 -1.30847648e-01
-6.70529366e-01 -8.27956423e-02 -2.91227996e-01 6.37062415e-02
6.01347506e-01 -1.31435141e-01 -2.65805244e-01 2.01905370e-01
-7.39704370e-01 -1.66201890e-01 1.40361100e-01 8.29365849e-02
3.21880370e-01 -9.94019508e-01 -1.09927666e+00 9.79466736e-02
3.15195173e-02 -1.10488467e-01 2.97610704e-02 5.24109066e-01
-7.95161247e-01 5.69216371e-01 -8.52146000e-02 -3.30293417e-01
-1.21777523e+00 2.10389659e-01 -1.50479063e-01 -1.23838210e+00
-5.20648956e-01 9.12157059e-01 4.20267470e-02 -2.92530954e-01
-1.25715300e-01 -7.15301692e-01 -2.74434447e-01 -1.00553267e-01
5.25109589e-01 2.71967143e-01 3.81073684e-01 -5.42820215e-01
-5.72819948e-01 7.31921196e-01 -3.64751071e-02 -2.27119714e-01
1.34120774e+00 1.76659539e-01 -7.56167099e-02 3.45209241e-01
9.27183092e-01 1.79508641e-01 -1.36427689e+00 -8.30304027e-02
4.35515046e-01 2.02709630e-01 -3.12101431e-02 -9.28801239e-01
-6.57433093e-01 8.51740181e-01 6.58384338e-02 1.42607596e-02
1.22786474e+00 1.07161567e-01 1.30546963e+00 -8.52279738e-02
5.11007667e-01 -9.89639819e-01 -3.88036340e-01 1.04715300e+00
5.54422915e-01 -7.71036983e-01 -1.42672107e-01 -6.80703998e-01
-4.85013366e-01 1.35893846e+00 2.40190238e-01 -6.06915474e-01
2.62217581e-01 7.62538731e-01 -2.81936489e-02 2.18158424e-01
-9.47986960e-01 -3.00269634e-01 9.03858393e-02 3.45799118e-01
9.56717670e-01 3.47097158e-01 -1.39512706e+00 9.44293380e-01
-6.38594389e-01 -2.37677068e-01 2.58975416e-01 7.77113199e-01
-6.68074012e-01 -1.73805654e+00 -3.23471814e-01 9.72070992e-02
-9.68402445e-01 -6.49157047e-01 2.48933703e-01 6.20975018e-01
2.32983634e-01 9.39916909e-01 1.69192255e-01 -2.95568556e-01
3.46580803e-01 3.07120681e-01 8.96564186e-01 -8.20029616e-01
-1.04339707e+00 2.34372728e-03 8.87804866e-01 -4.42096323e-01
-1.98504433e-01 -8.86746407e-01 -1.41384256e+00 -1.93778232e-01
-4.05681700e-01 1.90244079e-01 8.05910945e-01 1.15810847e+00
8.40273127e-02 4.94354188e-01 3.70131344e-01 -5.98976970e-01
-7.66656101e-01 -1.03339136e+00 -4.29420948e-01 -9.47965160e-02
8.05845037e-02 -2.21257165e-01 8.61119330e-02 -1.35556147e-01] | [10.362837791442871, 9.696105003356934] |
da964864-77c9-4eac-9cfe-78833d856ab0 | 3d-surface-reconstruction-from-multi-date | 2102.02502 | null | https://arxiv.org/abs/2102.02502v2 | https://arxiv.org/pdf/2102.02502v2.pdf | 3D Surface Reconstruction From Multi-Date Satellite Images | The reconstruction of accurate three-dimensional environment models is one of the most fundamental goals in the field of photogrammetry. Since satellite images provide suitable properties for obtaining large-scale environment reconstructions, there exist a variety of Stereo Matching based methods to reconstruct point clouds for satellite image pairs. Recently, the first Structure from Motion (SfM) based approach has been proposed, which allows to reconstruct point clouds from multiple satellite images. In this work, we propose an extension of this SfM based pipeline that allows us to reconstruct not only point clouds but watertight meshes including texture information. We provide a detailed description of several steps that are mandatory to exploit state-of-the-art mesh reconstruction algorithms in the context of satellite imagery. This includes a decomposition of finite projective camera calibration matrices, a skew correction of corresponding depth maps and input images as well as the recovery of real-world depth maps from reparameterized depth values. The paper presents an extensive quantitative evaluation on multi-date satellite images demonstrating that the proposed pipeline combined with current meshing algorithms outperforms state-of-the-art point cloud reconstruction algorithms in terms of completeness and median error. We make the source code of our pipeline publicly available. | ['Michael Arens', 'Christoph Bodensteiner', 'Sebastian Bullinger'] | 2021-02-04 | null | null | null | null | ['point-cloud-reconstruction'] | ['computer-vision'] | [ 2.88480878e-01 -3.21734726e-01 2.97359288e-01 -3.90671909e-01
-9.42836821e-01 -5.41954160e-01 6.82252884e-01 1.88148573e-01
-4.82027858e-01 4.31609869e-01 -3.08825940e-01 -8.46285373e-02
-2.87954956e-01 -1.20122015e+00 -8.54110897e-01 -4.77089405e-01
2.58134268e-02 1.31222963e+00 5.67733228e-01 -3.70612383e-01
4.51693535e-01 1.16899800e+00 -1.88296163e+00 -4.28892151e-02
8.80879581e-01 7.91973829e-01 5.41377068e-01 3.99424791e-01
-3.38821024e-01 -2.04759955e-01 2.01422736e-01 -1.85187444e-01
5.02428770e-01 1.48326054e-01 -7.67681062e-01 3.80826294e-01
8.86992037e-01 -5.42294979e-01 -8.58470798e-02 1.04380262e+00
3.00264239e-01 5.72025850e-02 4.85217243e-01 -8.40315878e-01
2.51002133e-01 -2.74268370e-02 -5.56579590e-01 -4.04846191e-01
3.75258386e-01 -2.01162416e-02 6.07717752e-01 -1.00652361e+00
8.22800577e-01 1.12163782e+00 8.05840850e-01 -1.29631549e-01
-1.54433608e+00 -4.26854402e-01 -4.74881262e-01 9.95923728e-02
-1.65760434e+00 -2.95204371e-01 6.31195605e-01 -5.91767550e-01
8.93019080e-01 5.27727067e-01 1.01636016e+00 3.09501559e-01
9.79362614e-03 9.64286551e-03 1.18282580e+00 -4.70833093e-01
3.45981628e-01 -1.41015634e-01 -2.04054087e-01 2.29289636e-01
2.86546379e-01 2.11443871e-01 -3.72548670e-01 -4.36975300e-01
1.21612453e+00 9.27588865e-02 -2.75853902e-01 -6.75942481e-01
-1.24353278e+00 6.73434377e-01 4.01734591e-01 6.63894862e-02
-6.48477137e-01 9.53783765e-02 3.27824503e-02 8.79599825e-02
7.45310605e-01 -6.36520982e-02 -1.31310552e-01 7.49711096e-02
-1.36609232e+00 4.51441824e-01 5.41028559e-01 9.78865504e-01
1.37852526e+00 -1.29021794e-01 6.93717182e-01 5.32317102e-01
5.74206173e-01 1.10072780e+00 -2.43429374e-02 -1.14648414e+00
4.64900821e-01 6.07340753e-01 1.65010035e-01 -1.14572513e+00
-2.83684552e-01 -6.13989495e-02 -8.19403350e-01 2.62702435e-01
-2.46572942e-02 5.14850199e-01 -5.66984653e-01 7.71629274e-01
6.76465690e-01 3.17816526e-01 -1.14035318e-02 7.72115171e-01
5.55955589e-01 5.71651518e-01 -4.05584931e-01 -7.42419735e-02
1.07766104e+00 -1.51317477e-01 -3.41148227e-01 -1.96375743e-01
2.63570070e-01 -9.49358702e-01 6.54245734e-01 3.50742489e-01
-8.67013395e-01 -2.62433439e-01 -7.93498456e-01 -7.64479116e-02
2.23337524e-02 -7.38691241e-02 4.13236439e-01 2.98190236e-01
-1.12458324e+00 9.98526990e-01 -1.04190350e+00 -6.83756173e-01
7.88480192e-02 2.88299322e-01 -7.18959570e-01 -1.24592960e-01
-6.28593147e-01 9.14842308e-01 3.16794544e-01 2.67876089e-01
-5.60709000e-01 -5.42816520e-01 -7.38651395e-01 -2.67280221e-01
-3.90165113e-02 -7.71772325e-01 9.44498241e-01 -8.42828989e-01
-1.20057738e+00 1.27198720e+00 -1.74597055e-01 -2.51389623e-01
6.28778577e-01 2.08099589e-01 1.13678668e-02 3.30331445e-01
8.51314887e-02 5.02572775e-01 5.85020244e-01 -1.46838415e+00
-5.02197862e-01 -7.41444767e-01 -3.93095404e-01 1.67732149e-01
1.08200938e-01 -1.26002342e-01 -4.16206509e-01 -1.82841301e-01
9.06398952e-01 -9.27727997e-01 -1.96634859e-01 4.29598957e-01
9.69143733e-02 3.99029136e-01 5.71217120e-01 -8.10495734e-01
6.26804590e-01 -1.85447121e+00 4.96938586e-01 5.05644441e-01
-3.43713790e-01 -1.20529786e-01 9.69718620e-02 6.91591740e-01
5.80670405e-03 -1.74019054e-01 -6.89988911e-01 -6.69383943e-01
-2.09045932e-01 7.02822626e-01 -2.97655493e-01 9.73225772e-01
-2.37039760e-01 3.39102030e-01 -5.13815522e-01 -3.85463148e-01
7.40454733e-01 5.84507048e-01 -2.23213837e-01 6.68453947e-02
-2.45856270e-01 7.89316118e-01 -1.78493217e-01 6.84659123e-01
1.46903229e+00 2.82740086e-01 1.01276442e-01 -1.86522126e-01
-6.90145850e-01 9.32196062e-03 -1.66344202e+00 2.03352928e+00
-3.72295916e-01 3.32397342e-01 4.69406635e-01 -5.79377949e-01
1.08299851e+00 3.40457022e-01 5.23620129e-01 -4.14502978e-01
-1.56009076e-02 6.52044058e-01 -5.97290456e-01 -2.40225613e-01
8.15394044e-01 -4.50673372e-01 4.11866307e-01 1.36888977e-02
-2.31861249e-01 -1.07920432e+00 -2.26062700e-01 -5.41810095e-02
4.37802285e-01 5.34765542e-01 2.18018577e-01 -4.91630703e-01
6.09081507e-01 5.46397388e-01 3.19741517e-01 -2.86648665e-02
5.05727887e-01 9.47258651e-01 -9.29161981e-02 -5.01992226e-01
-1.54584432e+00 -7.74307549e-01 -6.57288074e-01 1.21683646e-02
1.38014525e-01 -1.77077010e-01 -6.20615363e-01 3.53273988e-01
2.67649889e-01 3.38222116e-01 -4.11927283e-01 7.38494635e-01
-4.67157990e-01 -5.36061764e-01 1.54846281e-01 5.94686642e-02
4.84617859e-01 -4.99180317e-01 -5.58782578e-01 1.02799922e-01
-2.31190771e-01 -1.06115890e+00 2.09909722e-01 -3.58409047e-01
-1.77669346e+00 -1.10467482e+00 -4.80028063e-01 -4.74848926e-01
7.16835201e-01 5.00398457e-01 1.07996953e+00 2.59008735e-01
1.08463438e-02 3.65900397e-01 -2.89926976e-01 -1.28439695e-01
-5.59576690e-01 -1.65239811e-01 6.94539305e-03 8.12118798e-02
-4.65789139e-02 -1.01475215e+00 -3.85844380e-01 5.11775255e-01
-1.30954087e+00 2.58282661e-01 4.12094176e-01 2.38316268e-01
1.26972008e+00 -8.83185565e-02 -5.82023442e-01 -6.00339592e-01
-2.58882135e-01 -2.39223465e-01 -1.23365963e+00 6.28316626e-02
-3.17981154e-01 -1.37075841e-01 8.84659365e-02 1.39473617e-01
-6.95695698e-01 6.41268194e-01 -4.81531024e-01 -4.94189262e-01
-2.50920922e-01 6.04985237e-01 -7.61182308e-02 -5.72375953e-01
6.12102091e-01 4.82964188e-01 2.95933604e-01 -9.45455551e-01
6.73848242e-02 7.30549514e-01 6.40505731e-01 -5.11842549e-01
1.09204280e+00 1.15997708e+00 5.60260534e-01 -1.33497226e+00
-8.65744054e-02 -9.26213920e-01 -1.25885332e+00 -3.23988616e-01
6.17162228e-01 -1.06292081e+00 -2.80929416e-01 6.65971398e-01
-1.25006771e+00 -1.97524741e-01 -1.54188693e-01 4.42211330e-01
-7.02138305e-01 8.11148047e-01 -1.01718575e-01 -6.50023401e-01
-2.16632560e-01 -1.26654422e+00 1.45966887e+00 -1.70572817e-01
2.39175245e-01 -8.52336645e-01 4.21236008e-01 3.20802152e-01
2.10795611e-01 5.58390975e-01 3.42936754e-01 2.73414254e-01
-1.05877209e+00 -1.94763765e-01 1.47519037e-01 1.64169818e-01
-1.60894006e-01 3.40014458e-01 -8.19987595e-01 -3.78687084e-01
1.08821981e-01 2.65262634e-01 5.35721660e-01 2.61550277e-01
4.20242667e-01 1.73742250e-01 -2.72947073e-01 1.16050828e+00
2.07063055e+00 -5.46501815e-01 1.19795489e+00 9.12953138e-01
6.16141975e-01 6.68820083e-01 8.18228126e-01 5.51321507e-01
5.23141742e-01 1.05166125e+00 1.02315319e+00 6.91811740e-02
1.64990768e-01 -1.28470376e-01 -1.09475583e-01 6.96087480e-01
-5.64439952e-01 5.25978804e-01 -1.35188818e+00 4.65639591e-01
-1.71327412e+00 -8.53671968e-01 -1.05199301e+00 2.72890520e+00
4.29761529e-01 -3.63000244e-01 -2.60805219e-01 2.07902074e-01
6.21736526e-01 -8.50674957e-02 -6.90416545e-02 1.26515822e-02
-2.53336221e-01 3.07349473e-01 9.87329781e-01 9.76048589e-01
-9.26695406e-01 9.50979590e-01 5.56920147e+00 4.62489307e-01
-1.13255703e+00 1.67902082e-01 -3.00575763e-01 4.71915036e-01
-5.76862156e-01 5.51415920e-01 -6.70780838e-01 1.35704711e-01
9.91816998e-01 5.88644892e-02 3.06415826e-01 7.92188823e-01
5.39363801e-01 -5.25302410e-01 -6.10376418e-01 1.17638063e+00
2.49380302e-02 -1.46458805e+00 1.10067040e-01 4.09708500e-01
6.76532626e-01 4.94047642e-01 -4.86559480e-01 -5.34281850e-01
-1.85420871e-01 -5.81534505e-01 9.40157890e-01 8.07494700e-01
9.65336442e-01 -6.48634851e-01 5.60845852e-01 6.32556856e-01
-1.31523955e+00 4.55916971e-01 -7.61079907e-01 -1.55692711e-01
4.82338041e-01 8.30065966e-01 -5.65215051e-01 1.17777669e+00
7.46502280e-01 8.79308939e-01 -5.51802456e-01 1.35534441e+00
-6.53622076e-02 4.87269796e-02 -6.90935910e-01 7.38340735e-01
-9.83113125e-02 -9.16540325e-01 5.34132183e-01 6.74039781e-01
7.26734161e-01 2.58930445e-01 -4.34646718e-02 7.41535187e-01
3.78991246e-01 3.02128911e-01 -5.43200970e-01 5.25305390e-01
4.65119183e-01 1.20598137e+00 -6.95208311e-01 -2.68693835e-01
-1.49759889e-01 9.63152468e-01 -1.00922480e-01 -2.35868469e-01
-3.89602721e-01 2.79435337e-01 6.24118268e-01 6.57403529e-01
1.03699334e-01 -8.15101802e-01 -4.13165152e-01 -1.16980958e+00
1.66161153e-02 -6.17133856e-01 -3.93289588e-02 -1.03787756e+00
-7.91073561e-01 5.46982348e-01 2.88943619e-01 -1.80098474e+00
-1.04813091e-01 -4.66562271e-01 -3.10156673e-01 1.21550488e+00
-1.70156968e+00 -1.34449661e+00 -7.91525722e-01 4.94648010e-01
3.67540449e-01 2.59096473e-01 9.91350472e-01 1.99707031e-01
4.00306433e-02 -4.95674908e-01 6.92767978e-01 -3.24172467e-01
5.27230740e-01 -8.63561571e-01 5.91155112e-01 8.86719704e-01
-1.53088523e-03 3.54092956e-01 9.73728538e-01 -8.96356165e-01
-1.75503457e+00 -8.28894556e-01 7.76004970e-01 -2.97059059e-01
3.70341688e-01 7.79771730e-02 -1.07853627e+00 8.56211007e-01
-3.98681581e-01 8.02883357e-02 2.15103552e-01 -5.26211619e-01
8.13386068e-02 -2.25769445e-01 -1.25910199e+00 -2.19398811e-02
6.56243443e-01 -4.73407924e-01 -5.97465992e-01 4.09952849e-01
-7.61259161e-03 -8.47857952e-01 -1.11632323e+00 4.16696846e-01
4.84668046e-01 -1.40637124e+00 1.06362653e+00 4.33407456e-01
2.44516551e-01 -7.22490489e-01 -5.41012168e-01 -1.04577899e+00
-6.34438125e-03 -1.66387618e-01 5.08384585e-01 1.01617861e+00
-1.23517387e-01 -6.56240463e-01 7.25271642e-01 5.61853468e-01
-3.14185888e-01 2.85616945e-02 -1.19582593e+00 -6.85488701e-01
7.36334994e-02 -5.08152485e-01 6.34600937e-01 9.17594731e-01
-5.60004354e-01 -2.45009080e-01 -7.04286993e-02 7.26552844e-01
1.07521164e+00 3.50645274e-01 1.11871099e+00 -1.86142313e+00
-8.43648091e-02 2.81521864e-02 -6.94780171e-01 -8.14676940e-01
6.75952286e-02 -8.05787802e-01 -8.00573379e-02 -1.76103067e+00
5.14689125e-02 -6.49067163e-01 8.54536295e-01 4.71510738e-02
5.31907678e-01 4.13169771e-01 -3.41562033e-02 9.11620438e-01
2.07493380e-01 4.70203429e-01 8.94206643e-01 1.50346801e-01
4.42727283e-02 -1.48689955e-01 2.36459956e-01 6.77997708e-01
4.15596783e-01 -5.36406338e-01 5.00652418e-02 -9.60729122e-01
4.11763579e-01 1.71308041e-01 5.99315763e-01 -1.19194818e+00
1.99738935e-01 -3.51775765e-01 2.29301061e-02 -1.13116896e+00
6.75732374e-01 -1.44703484e+00 1.26511765e+00 4.35817897e-01
4.62229401e-01 6.06604367e-02 2.55076498e-01 4.41409200e-01
-2.60442257e-01 -4.37009960e-01 9.43026662e-01 -3.94321918e-01
-8.29683244e-01 4.17426556e-01 1.10925287e-01 -7.08809793e-01
8.83293390e-01 -5.67428172e-01 6.46136850e-02 -1.19475581e-01
-4.17499959e-01 -5.64137921e-02 1.56316066e+00 -1.18128344e-01
7.34295905e-01 -1.14773285e+00 -8.94731700e-01 3.45922470e-01
1.65716231e-01 6.26214683e-01 4.75284755e-01 8.57480109e-01
-1.56386685e+00 3.07319671e-01 -4.31468546e-01 -1.17427957e+00
-1.44766796e+00 2.05162894e-02 5.94846666e-01 2.51282007e-01
-7.85831928e-01 4.32917714e-01 -4.59558904e-01 -7.46360898e-01
-3.63493353e-01 -3.77254456e-01 1.11749992e-02 -2.68855859e-02
3.70728493e-01 4.63325888e-01 6.16828322e-01 -1.41831136e+00
-3.91750783e-01 1.28084540e+00 7.99718380e-01 -4.09819692e-01
1.65219450e+00 -2.56473631e-01 -7.47187197e-01 3.87119383e-01
8.83044720e-01 -1.27899230e-01 -1.07259238e+00 -1.89902827e-01
-1.61530510e-01 -1.12141931e+00 4.93966758e-01 -7.54527003e-02
-8.55261922e-01 1.05747569e+00 6.82147980e-01 -2.15470240e-01
9.22942877e-01 -1.52639538e-01 4.32850331e-01 1.58672966e-02
1.13324046e+00 -7.98079550e-01 -7.89195001e-01 4.31325465e-01
1.17612386e+00 -1.13530815e+00 5.29309630e-01 -7.23942995e-01
-1.99181378e-01 1.40025961e+00 -1.22810774e-01 -2.05525488e-01
6.10784709e-01 3.33087817e-02 -1.10168561e-01 -2.55379498e-01
-8.69611278e-02 -2.73676753e-01 2.74812356e-02 5.81672311e-01
1.26433611e-01 4.85977270e-02 -3.84163409e-01 -3.03964734e-01
-1.39103726e-01 1.03431210e-01 5.90428412e-01 1.00916421e+00
-7.07777619e-01 -1.51151872e+00 -1.25830472e+00 9.42383781e-02
2.61516184e-01 -3.34386006e-02 -1.56869963e-01 6.32494509e-01
-1.70556992e-01 4.04617369e-01 3.98729175e-01 -3.18686575e-01
5.70894003e-01 -1.14548661e-01 6.47853196e-01 -5.84193707e-01
-2.96811551e-01 1.12716049e-01 -1.03093617e-01 -5.55718064e-01
-8.01709473e-01 -1.10901558e+00 -1.15559220e+00 -6.65082514e-01
-2.19240010e-01 4.70399521e-02 1.28447270e+00 7.33358562e-01
1.81237996e-01 -3.25053155e-01 5.35253763e-01 -1.70142984e+00
-3.37259352e-01 -9.01211858e-01 -1.02640176e+00 4.00828034e-01
-3.11434362e-02 -6.58757687e-01 -4.23942208e-01 1.52584031e-01] | [8.440211296081543, -2.6218910217285156] |
8c089800-7978-4675-ab05-557a8354a4b1 | visible-infrared-person-re-identification-via | 2302.08212 | null | https://arxiv.org/abs/2302.08212v1 | https://arxiv.org/pdf/2302.08212v1.pdf | Visible-Infrared Person Re-Identification via Patch-Mixed Cross-Modality Learning | Visible-infrared person re-identification (VI-ReID) aims to retrieve images of the same pedestrian from different modalities, where the challenges lie in the significant modality discrepancy. To alleviate the modality gap, recent methods generate intermediate images by GANs, grayscaling, or mixup strategies. However, these methods could ntroduce extra noise, and the semantic correspondence between the two modalities is not well learned. In this paper, we propose a Patch-Mixed Cross-Modality framework (PMCM), where two images of the same person from two modalities are split into patches and stitched into a new one for model learning. In this way, the modellearns to recognize a person through patches of different styles, and the modality semantic correspondence is directly embodied. With the flexible image generation strategy, the patch-mixed images freely adjust the ratio of different modality patches, which could further alleviate the modality imbalance problem. In addition, the relationship between identity centers among modalities is explored to further reduce the modality variance, and the global-to-part constraint is introduced to regularize representation learning of part features. On two VI-ReID datasets, we report new state-of-the-art performance with the proposed method. | ['Bo Du', 'Yutian Lin', 'Zhihao Qian'] | 2023-02-16 | null | null | null | null | ['person-re-identification'] | ['computer-vision'] | [ 4.24986064e-01 -1.70510143e-01 -1.29201829e-01 -2.57387400e-01
-6.16934955e-01 -5.10410309e-01 7.30503321e-01 -4.30962741e-01
-2.92912692e-01 6.48578942e-01 2.23218173e-01 3.04595947e-01
8.18163306e-02 -7.23244727e-01 -7.20164180e-01 -1.01427972e+00
8.50685656e-01 9.88932401e-02 -1.12071343e-01 -2.32933894e-01
5.25188334e-02 1.18176199e-01 -1.65983629e+00 3.47862303e-01
1.10024321e+00 9.94875848e-01 1.11690693e-01 1.07270584e-01
-3.35389435e-01 4.55027431e-01 -5.02162755e-01 -6.87658131e-01
4.43476319e-01 -7.42716074e-01 -3.90933484e-01 3.60855848e-01
7.29748785e-01 -1.60998166e-01 -5.57163239e-01 1.28779554e+00
6.09798968e-01 1.09696113e-01 5.99761963e-01 -1.52566123e+00
-8.85115266e-01 2.86765039e-01 -1.01298594e+00 -2.13211745e-01
4.17207301e-01 2.29255632e-01 3.99231941e-01 -6.98936820e-01
4.01274294e-01 1.60798120e+00 4.86513704e-01 8.46193612e-01
-1.27175450e+00 -9.23725069e-01 4.12093937e-01 5.41408241e-01
-1.40912294e+00 -4.97694105e-01 1.05950534e+00 -3.78300458e-01
1.05709404e-01 3.06638628e-01 8.40562522e-01 1.45045507e+00
-2.32492328e-01 7.14000881e-01 1.46217179e+00 -4.43112999e-01
-1.20472305e-01 3.30947369e-01 -2.26163357e-01 4.72113431e-01
2.59585917e-01 2.06573933e-01 -6.17176116e-01 -3.56375892e-03
7.22619295e-01 4.07779396e-01 -3.91716510e-01 -3.03147048e-01
-1.34544861e+00 4.38184679e-01 5.99098444e-01 4.24597412e-02
-1.31008044e-01 -3.54315937e-01 2.46668935e-01 6.98016733e-02
1.42598420e-01 -6.66036084e-03 -5.74124381e-02 2.33549699e-01
-5.26423872e-01 1.64967030e-01 2.55757183e-01 9.04203236e-01
9.08420444e-01 -1.02773085e-01 -4.24430788e-01 1.04326594e+00
4.04550701e-01 7.99810052e-01 5.29717207e-01 -8.17962885e-01
8.17615747e-01 7.49222100e-01 -1.25761074e-03 -9.32168067e-01
-2.09221635e-02 -3.23572993e-01 -1.29839444e+00 1.42736554e-01
4.67228144e-01 -1.48581669e-01 -1.21566546e+00 2.03475118e+00
4.58230257e-01 2.80881137e-01 1.31082356e-01 1.12099802e+00
1.06794572e+00 5.64510226e-01 2.30601013e-01 -1.53643593e-01
1.68701589e+00 -1.03014207e+00 -5.81281662e-01 -4.12654251e-01
-1.21150605e-01 -6.82457089e-01 6.98730826e-01 1.91738568e-02
-9.65465724e-01 -1.02863240e+00 -9.24020767e-01 5.32441474e-02
-1.67978913e-01 1.84271187e-01 3.29209596e-01 6.43405080e-01
-6.86387897e-01 2.31664628e-02 -2.38163590e-01 -3.61104637e-01
4.42660689e-01 -1.25200953e-02 -6.47437871e-01 -5.04659593e-01
-1.14039421e+00 6.08285367e-01 5.22566974e-01 3.77744585e-01
-5.30746281e-01 -5.92010856e-01 -8.78329575e-01 -1.43784553e-01
2.90223360e-01 -1.04819584e+00 6.35885596e-01 -1.35378718e+00
-1.45654964e+00 1.06795168e+00 -1.92821413e-01 1.54299617e-01
7.28138089e-01 3.03009480e-01 -7.24379957e-01 2.38443129e-02
1.05906792e-01 8.61389518e-01 1.14124274e+00 -1.66743231e+00
-6.27193451e-01 -6.59167886e-01 4.76341397e-02 5.34079134e-01
-2.51468807e-01 -1.11250468e-01 -7.08690166e-01 -8.02086592e-01
3.25817168e-01 -9.37027752e-01 1.06917344e-01 2.65842471e-02
-4.62210715e-01 5.37765154e-04 3.57765943e-01 -8.80557060e-01
5.93256891e-01 -2.26807427e+00 3.77512813e-01 2.10618824e-01
1.21835314e-01 1.04851015e-01 -4.79001671e-01 8.13321993e-02
-1.80061013e-01 -1.73665464e-01 -1.67767718e-01 -4.17559057e-01
-2.01331511e-01 1.71038345e-01 -6.68172687e-02 2.99165189e-01
2.22417768e-02 8.81226301e-01 -7.34149039e-01 -6.22939765e-01
1.46289393e-01 6.61062002e-01 -1.66920125e-01 4.10099924e-01
9.56866294e-02 9.49829280e-01 -3.53235513e-01 7.35588431e-01
1.18877804e+00 2.15650089e-02 9.97411646e-03 -6.22887909e-01
1.51491702e-01 -3.84195566e-01 -1.22963059e+00 2.02810740e+00
-2.26214260e-01 6.74834996e-02 1.43903065e-02 -8.31804037e-01
1.03618944e+00 8.47394317e-02 3.30184817e-01 -1.04534984e+00
3.70737053e-02 -3.23081426e-02 -1.31502792e-01 -5.78100562e-01
3.04029465e-01 -1.44482598e-01 1.27557099e-01 1.17163889e-01
-1.50843069e-01 3.97241890e-01 -1.67631246e-02 -1.77760109e-01
3.57746989e-01 2.84419984e-01 -2.24665672e-01 2.06072956e-01
8.06584239e-01 -2.34336078e-01 8.36677432e-01 7.17506468e-01
-1.90704197e-01 1.02598047e+00 1.50819331e-01 -1.52213648e-01
-9.17998314e-01 -1.22816932e+00 -1.22558102e-01 9.33819413e-01
8.22772980e-01 1.07374787e-01 -7.49957383e-01 -4.93254751e-01
-1.24629833e-01 3.22218537e-01 -6.47895992e-01 -3.51827174e-01
-3.96069318e-01 -7.93477535e-01 4.85430449e-01 4.39627469e-01
1.12002647e+00 -7.76383638e-01 -1.07919918e-02 -1.70113191e-01
-5.97910941e-01 -9.54099298e-01 -7.36016154e-01 -6.00453258e-01
-5.64986885e-01 -1.05204320e+00 -1.25018334e+00 -9.79509652e-01
9.47551548e-01 6.10664189e-01 7.64810979e-01 -7.07355812e-02
-7.04471320e-02 3.82815689e-01 -4.00199354e-01 6.68853745e-02
-2.61876583e-01 -1.88991129e-01 -2.03834139e-02 6.24049962e-01
2.89941877e-01 -5.27493000e-01 -9.43898678e-01 4.16184038e-01
-8.85892153e-01 4.80268955e-01 6.29575968e-01 1.13385558e+00
6.48732543e-01 -1.18369740e-02 3.09470385e-01 -5.83575964e-01
1.62689641e-01 -3.94117385e-01 -3.08302850e-01 7.06052303e-01
-5.12319922e-01 -1.33236587e-01 5.11557579e-01 -7.78635263e-01
-1.53002918e+00 2.13106163e-02 9.94916782e-02 -5.99037111e-01
-4.44476724e-01 1.25315040e-01 -9.28480625e-01 -1.77631840e-01
3.34218740e-01 3.84043634e-01 2.54233450e-01 -5.27044237e-01
5.09162247e-01 6.30067766e-01 7.13502705e-01 -6.38752759e-01
9.30960000e-01 4.07608598e-01 -2.04135418e-01 -3.17435384e-01
-5.64461231e-01 -1.69003382e-01 -4.59849626e-01 -1.88323811e-01
8.89343619e-01 -1.27454412e+00 -5.38368583e-01 9.83639061e-01
-1.10565507e+00 4.17754352e-02 -1.77965969e-01 4.09774721e-01
-1.19192362e-01 5.01484513e-01 -4.77908790e-01 -5.37978411e-01
-1.79545373e-01 -1.01444471e+00 1.04761434e+00 9.46360230e-01
5.48105776e-01 -5.07152915e-01 -1.10216565e-01 7.54620314e-01
2.92158663e-01 7.43531436e-02 7.27769852e-01 -1.42288983e-01
-4.63480115e-01 8.16532003e-04 -6.78968668e-01 2.25917771e-01
2.49256074e-01 -4.52510267e-01 -1.05677414e+00 -4.10544097e-01
-2.12074995e-01 -7.33980983e-02 8.71460259e-01 1.31308332e-01
1.13677478e+00 -2.88301945e-01 -3.82214069e-01 8.06780100e-01
1.18226326e+00 1.84915394e-01 8.41982841e-01 3.98556620e-01
1.10847902e+00 8.21823835e-01 3.49764377e-01 1.62101269e-01
6.88097239e-01 7.88692951e-01 9.58334431e-02 -1.98058829e-01
-4.47370738e-01 -6.69879436e-01 1.89807296e-01 4.03703868e-01
-2.18986601e-01 -2.08836704e-01 -4.11361933e-01 2.20360577e-01
-1.85599899e+00 -1.19581020e+00 2.32481763e-01 2.45457125e+00
7.06408918e-01 -3.22105497e-01 2.48347461e-01 -2.13373572e-01
1.24013448e+00 4.94719930e-02 -7.58708358e-01 3.93897504e-01
-4.77816314e-01 -4.12767142e-01 3.06418031e-01 2.28005931e-01
-9.48962986e-01 5.46937168e-01 5.13785028e+00 9.57659125e-01
-1.02632320e+00 1.90564647e-01 7.24094152e-01 1.05354324e-01
-4.04297441e-01 3.11852228e-02 -6.90840900e-01 8.39900136e-01
1.97149843e-01 -2.94512436e-02 6.90472782e-01 4.94637787e-01
-8.68396088e-02 6.00193487e-03 -7.95741320e-01 1.67487574e+00
4.69259232e-01 -8.51443052e-01 1.15795568e-01 1.40307471e-01
7.78080106e-01 -5.48328578e-01 2.42733583e-01 2.37688318e-01
-2.24052832e-01 -8.62450480e-01 8.17705393e-01 9.24509466e-01
9.01381791e-01 -7.08249032e-01 6.77604318e-01 1.42741501e-01
-1.42928076e+00 -1.57065973e-01 -4.25735712e-01 3.55573773e-01
1.21501736e-01 3.38560551e-01 -1.56365752e-01 9.29445267e-01
7.82982051e-01 6.65105283e-01 -9.85838294e-01 1.02487636e+00
-4.80224602e-02 3.03716701e-03 -1.80203244e-01 5.17210364e-01
-5.27794659e-01 -5.51732421e-01 4.93738890e-01 7.81424820e-01
4.18908656e-01 -1.16665696e-03 4.27881539e-01 1.08168530e+00
-1.48497242e-02 -1.82811931e-01 -2.84595162e-01 3.26059133e-01
6.49296343e-01 1.18266022e+00 -2.26746440e-01 -2.49405488e-01
-4.35671002e-01 1.35452747e+00 2.66678026e-03 7.17660487e-01
-8.38077545e-01 -2.78988510e-01 5.24533212e-01 5.88262342e-02
5.37287593e-02 1.67858511e-01 -1.28553867e-01 -1.57853055e+00
3.93196315e-01 -9.72042084e-01 5.09917617e-01 -8.58285129e-01
-1.89786172e+00 5.72765410e-01 1.71049148e-01 -1.57128668e+00
-1.53156007e-02 -2.74651259e-01 -4.87110734e-01 1.17923748e+00
-1.53289604e+00 -1.74161804e+00 -7.38532484e-01 8.94411922e-01
2.08431572e-01 -4.34283733e-01 5.05704105e-01 4.99927133e-01
-7.07970619e-01 1.09872746e+00 4.53467993e-03 2.77095407e-01
8.67713571e-01 -8.83440733e-01 -1.09756645e-02 8.20890725e-01
-1.97319254e-01 5.97119272e-01 3.89745623e-01 -6.50461316e-01
-1.49026740e+00 -1.03309464e+00 2.97697604e-01 -1.04667850e-01
9.80594829e-02 -2.00340465e-01 -9.84574556e-01 3.02765071e-01
2.20527127e-01 1.50747448e-01 5.84498465e-01 -8.64561740e-03
-7.93237984e-01 -4.71212596e-01 -1.22849607e+00 5.87254703e-01
1.21219206e+00 -6.46363378e-01 -4.94277269e-01 5.10676429e-02
4.54228967e-01 -4.16665822e-01 -7.73049891e-01 3.82611215e-01
6.87674165e-01 -9.96692777e-01 1.11322486e+00 -3.55086714e-01
2.47609332e-01 -6.37383521e-01 -1.84452515e-02 -1.35960853e+00
-2.77441740e-01 -1.45216286e-01 2.60558948e-02 1.87010241e+00
-5.07982783e-02 -9.39726651e-01 5.90279639e-01 7.44687319e-01
3.12572867e-01 -1.89170465e-01 -9.20704365e-01 -6.37884557e-01
4.09968384e-02 2.52474546e-01 9.77476954e-01 8.58007491e-01
-3.38066787e-01 2.16866344e-01 -6.14942372e-01 2.26190761e-01
8.91750276e-01 3.13921303e-01 7.74659216e-01 -1.10398519e+00
-5.19517064e-01 -4.89519387e-01 -4.24879491e-01 -1.11555982e+00
1.93239838e-01 -7.38212466e-01 -1.39799323e-02 -1.26587164e+00
7.78498769e-01 -5.59720457e-01 -4.03178543e-01 4.64499861e-01
-5.41039348e-01 4.00748938e-01 4.00211424e-01 3.63986403e-01
-3.39727700e-01 8.67580652e-01 1.46163356e+00 -6.18234038e-01
-1.45378247e-01 -3.34554203e-02 -9.65724707e-01 2.97572047e-01
7.08163142e-01 -1.32060140e-01 -4.14575368e-01 -5.47113895e-01
-1.19874887e-01 1.86587617e-01 7.47401178e-01 -9.01043117e-01
2.79104322e-01 -1.99361935e-01 8.73954892e-01 -4.17536706e-01
5.41742325e-01 -8.56958270e-01 5.56290865e-01 2.76562329e-02
-2.48275891e-01 -1.45633286e-02 -2.28882227e-02 8.46335113e-01
-2.20459715e-01 2.35240739e-02 7.93211758e-01 -1.04989551e-01
-6.27459943e-01 5.48861384e-01 1.47120908e-01 -8.72374251e-02
8.09264243e-01 -4.01317298e-01 -6.02203846e-01 -1.68801278e-01
-4.20747519e-01 1.61619455e-01 8.43051314e-01 8.10134530e-01
5.78931987e-01 -1.83491004e+00 -8.60204577e-01 4.68480706e-01
3.89260918e-01 -6.11879900e-02 1.02729714e+00 7.08253920e-01
-3.97486053e-03 -2.08436862e-01 -4.58168060e-01 -6.87776268e-01
-1.14132357e+00 7.38058746e-01 5.41778207e-01 2.15461016e-01
-5.42038143e-01 5.85775435e-01 5.19208729e-01 -6.12495840e-01
6.35955483e-02 5.28757811e-01 -3.52938980e-01 5.20694517e-02
8.12610924e-01 2.44460076e-01 -2.26502880e-01 -1.03939342e+00
-3.28672588e-01 8.98744583e-01 -2.34688982e-01 -1.46581456e-01
8.10884178e-01 -4.33985412e-01 -1.73735678e-01 1.64462358e-01
1.08880258e+00 -5.14103360e-02 -1.40293384e+00 -4.84754950e-01
-7.41681933e-01 -8.95120621e-01 -3.46457541e-01 -8.35045695e-01
-1.40272057e+00 6.14716709e-01 1.24157989e+00 -1.38950974e-01
1.34622073e+00 -5.52216060e-02 7.87951946e-01 -4.09098081e-02
3.84775132e-01 -9.02517259e-01 1.06708273e-01 1.05542518e-01
6.20683014e-01 -1.38090611e+00 -1.64270565e-01 -4.32594210e-01
-8.28251243e-01 9.32449162e-01 9.03429389e-01 3.20237875e-01
3.17704797e-01 -3.61444771e-01 1.51055709e-01 2.65669584e-01
-6.01181835e-02 -2.80333281e-01 4.10090029e-01 1.04709554e+00
6.80692419e-02 1.48015946e-01 -6.46899939e-02 6.80997074e-01
1.16964586e-01 -1.95264786e-01 5.61986044e-02 5.99772096e-01
3.48363332e-02 -1.27566981e+00 -8.28048050e-01 1.02143258e-01
9.21416804e-02 7.72223547e-02 -1.74192369e-01 3.04235458e-01
6.15536809e-01 1.16828799e+00 -7.71058053e-02 -6.24387145e-01
3.82435232e-01 -2.07384363e-01 7.64482617e-01 -3.47488262e-02
-2.46342391e-01 9.01795924e-02 -3.51951927e-01 -1.81453362e-01
-6.80853248e-01 -6.16587818e-01 -7.06502259e-01 -3.87392551e-01
-1.85403332e-01 -9.53697786e-03 3.75058681e-01 8.79159451e-01
5.41083694e-01 3.19789559e-01 8.46742690e-01 -1.05654347e+00
-2.83670843e-01 -8.45835447e-01 -5.55746078e-01 7.37422943e-01
1.78797647e-01 -7.26166725e-01 -1.39915541e-01 6.64590150e-02] | [14.70380973815918, 0.9636226296424866] |
71ff7c27-21c1-4633-b6cb-22db3ae4916e | learning-to-transpile-amr-into-sparql | 2112.07877 | null | https://arxiv.org/abs/2112.07877v2 | https://arxiv.org/pdf/2112.07877v2.pdf | Learning to Transpile AMR into SPARQL | We propose a transition-based system to transpile Abstract Meaning Representation (AMR) into SPARQL for Knowledge Base Question Answering (KBQA). This allows us to delegate part of the semantic representation to a strongly pre-trained semantic parser, while learning transpiling with small amount of paired data. We depart from recent work relating AMR and SPARQL constructs, but rather than applying a set of rules, we teach a BART model to selectively use these relations. Further, we avoid explicitly encoding AMR but rather encode the parser state in the attention mechanism of BART, following recent semantic parsing works. The resulting model is simple, provides supporting text for its decisions, and outperforms recent approaches in KBQA across two knowledge bases: DBPedia (LC-QuAD 1.0, QALD-9) and Wikidata (WebQSP, SWQ-WD). | ['Salim Roukos', 'Radu Florian', 'Pavan Kapanipathi', 'Ibrahim Abdelaziz', 'Nandana Mihindukulasooriya', 'Tahira Naseem', 'Ramon Fernandez Astudillo', 'Mihaela Bornea'] | 2021-12-15 | null | null | null | null | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-2.39160061e-01 1.04014194e+00 -1.70981124e-01 -7.21025765e-01
-8.95361423e-01 -7.65265226e-01 5.43130696e-01 5.23684263e-01
-2.91566432e-01 8.68221521e-01 4.43917006e-01 -5.95249712e-01
-3.14551950e-01 -1.50741577e+00 -1.12365162e+00 6.84555545e-02
1.74507815e-02 1.00078630e+00 7.69304931e-01 -8.13164353e-01
-1.11439586e-01 1.54076174e-01 -1.44846439e+00 6.50998414e-01
8.74448180e-01 9.56587911e-01 6.26441762e-02 6.58902228e-01
-8.04714084e-01 1.55883980e+00 -4.07203764e-01 -1.12960386e+00
-1.72564074e-01 -1.54067367e-01 -1.93195963e+00 -5.94598591e-01
4.75050330e-01 6.98249340e-02 -1.53342620e-01 9.10491526e-01
1.42626405e-01 2.03371286e-01 2.04045430e-01 -1.15369487e+00
-1.03494036e+00 9.01615679e-01 2.59592235e-01 1.60505846e-02
7.70214796e-01 -3.31111193e-01 1.41570783e+00 -2.96518862e-01
9.12822664e-01 1.56120992e+00 6.87627077e-01 8.16478968e-01
-7.40960896e-01 -1.62845626e-01 7.49173835e-02 4.42945510e-01
-1.02213085e+00 -4.05518383e-01 2.21440300e-01 -4.71077859e-02
1.60006440e+00 3.50916326e-01 4.74330217e-01 5.82403243e-01
-1.92636088e-01 6.01488650e-01 8.61095488e-01 -7.55490065e-01
4.28311259e-01 1.92573056e-01 6.78247869e-01 1.03814840e+00
2.11831719e-01 -5.35953283e-01 -4.07755136e-01 -3.11170012e-01
5.31228244e-01 -4.60858464e-01 -5.56088798e-02 -6.24667585e-01
-6.27568305e-01 8.03523123e-01 3.24680746e-01 -5.86554073e-02
-1.80766925e-01 4.39784348e-01 5.36945760e-01 5.88551462e-01
3.88179719e-02 6.53312743e-01 -1.13172483e+00 -6.07339069e-02
-1.01954758e-01 5.72818696e-01 1.35150421e+00 1.35375583e+00
9.87432420e-01 -5.09985685e-01 -2.22269997e-01 8.24964345e-01
4.93392289e-01 2.23824069e-01 2.07265273e-01 -1.41139567e+00
6.85827136e-01 8.45308602e-01 3.10715824e-01 -4.33920145e-01
-1.85407490e-01 1.16168998e-01 1.74717054e-01 -2.18495980e-01
3.99368882e-01 6.35103285e-02 -1.00090253e+00 1.71930397e+00
3.89990568e-01 -1.34223655e-01 7.87864923e-01 6.94285870e-01
1.26693344e+00 5.34157336e-01 7.14584231e-01 3.67586404e-01
2.01044250e+00 -8.64608169e-01 -5.60735047e-01 -4.43377554e-01
1.10388529e+00 9.38693509e-02 1.23118472e+00 8.89747068e-02
-1.24925852e+00 -6.71217814e-02 -8.68686438e-01 -8.44788432e-01
-9.54473972e-01 -3.94343734e-01 1.06992865e+00 4.90706801e-01
-1.08032155e+00 4.04393643e-01 -5.63835144e-01 -8.50052059e-01
5.89193225e-01 2.37580284e-01 -5.81184149e-01 -6.34017348e-01
-1.73925281e+00 1.17601669e+00 9.43261385e-01 -2.81213284e-01
-7.07455695e-01 -7.83711731e-01 -1.21227860e+00 1.57308280e-01
6.09860539e-01 -1.12891018e+00 1.40634120e+00 -5.12864649e-01
-1.54858780e+00 1.12402952e+00 -2.35530287e-02 -7.64106989e-01
-2.62899429e-01 -4.64483619e-01 -6.30773365e-01 2.96325266e-01
2.42188856e-01 8.54220629e-01 1.44937068e-01 -1.05207014e+00
-6.61099553e-01 -5.89303970e-01 1.14192677e+00 3.27901363e-01
2.37900808e-01 2.66041875e-01 -4.98632014e-01 1.45486109e-02
-3.46098170e-02 -3.04974377e-01 -2.22786084e-01 -2.81290144e-01
-6.10780083e-02 -5.55999279e-01 3.38471264e-01 -8.47457945e-01
9.12094951e-01 -1.70617092e+00 -6.23437688e-02 1.29090939e-02
-7.68217444e-02 1.58313781e-01 -1.56894878e-01 6.59251332e-01
-1.49786258e-02 1.39030114e-01 -4.06835675e-01 2.06592157e-01
2.48469278e-01 1.14834666e+00 -5.94149709e-01 -2.63990492e-01
4.34045434e-01 1.20742249e+00 -1.13013959e+00 -6.54043376e-01
-1.48110762e-01 -5.87475598e-02 -1.01972723e+00 2.85808444e-01
-1.00395989e+00 -1.97206482e-01 -8.40132117e-01 7.05241024e-01
5.32927752e-01 -1.94748908e-01 5.33486605e-01 -1.83915764e-01
3.26629430e-01 1.01140726e+00 -9.47132230e-01 2.25371695e+00
-6.93487287e-01 -8.21538121e-02 9.06138718e-02 -1.21174705e+00
8.11470389e-01 4.14559811e-01 -6.15549311e-02 -8.94151509e-01
-4.38431203e-01 2.02830538e-01 -4.73099828e-01 -8.36722553e-01
5.91247320e-01 -3.74322265e-01 -5.10757208e-01 4.15919907e-02
7.39426017e-01 -3.34510714e-01 2.37389132e-01 3.89212489e-01
1.42911541e+00 6.18356705e-01 4.18799311e-01 -3.44052613e-01
5.56556582e-01 5.36143661e-01 5.05903542e-01 7.09573805e-01
2.16329336e-01 1.48927838e-01 7.68872201e-01 -4.87658739e-01
-7.11092591e-01 -1.23725975e+00 -1.62432104e-01 1.41370463e+00
7.54553005e-02 -7.81933188e-01 -6.46613538e-01 -1.24921286e+00
1.12626784e-01 1.17766547e+00 -3.66534084e-01 -1.42829031e-01
-8.14135075e-01 -4.86397475e-01 8.99227858e-01 6.46192133e-01
4.49349642e-01 -1.28178406e+00 -6.35891020e-01 3.64644259e-01
-1.90960824e-01 -1.26797104e+00 5.42916775e-01 4.28838581e-01
-7.47027278e-01 -1.11227155e+00 1.69066489e-01 -6.76613867e-01
2.54226267e-01 -3.92069697e-01 1.89959061e+00 1.53757976e-02
1.20174550e-02 7.00436175e-01 -5.79734206e-01 -4.05152321e-01
-4.37606990e-01 1.01183236e-01 -6.29606843e-01 -7.00941980e-01
7.77190447e-01 -5.25629878e-01 -1.82977095e-01 -1.30524009e-01
-8.51877451e-01 -8.51509571e-02 1.96353182e-01 5.63160241e-01
4.17039216e-01 -1.29169166e-01 7.86013126e-01 -1.36669493e+00
5.05629063e-01 -7.66758978e-01 -4.09463316e-01 7.14275420e-01
-3.88735831e-01 4.77894425e-01 7.53961444e-01 5.29289305e-01
-1.36078215e+00 -2.97695279e-01 -6.10653222e-01 1.97031423e-01
-4.86105978e-01 6.18074417e-01 -7.14972138e-01 2.88355231e-01
6.91801250e-01 -1.52933940e-01 -2.39201128e-01 -6.38422608e-01
1.01957679e+00 4.79829073e-01 7.26455629e-01 -1.46524608e+00
4.80110824e-01 2.56625623e-01 -1.06019415e-01 -2.36260533e-01
-1.30502164e+00 -3.54656219e-01 -4.15154010e-01 6.47098064e-01
1.09253490e+00 -9.91339326e-01 -7.73319125e-01 -3.27323824e-01
-1.11243606e+00 -3.51458997e-01 -8.64540339e-01 -5.38461143e-03
-8.27478409e-01 1.94902211e-01 -7.13587880e-01 -5.15582144e-01
-3.47107977e-01 -4.99215931e-01 1.02266598e+00 4.37422469e-02
-1.79123938e-01 -1.03861344e+00 1.53958976e-01 7.22766638e-01
1.84299856e-01 -4.45695892e-02 1.59763813e+00 -1.04946387e+00
-4.94351417e-01 -2.81351595e-03 -3.95457655e-01 2.75451452e-01
-1.47593305e-01 -6.38309956e-01 -8.92695367e-01 2.56572306e-01
-5.02625942e-01 -7.51977921e-01 5.03491879e-01 -2.43428543e-01
1.10903668e+00 -5.58206439e-01 -3.14597264e-02 4.23988938e-01
1.70761919e+00 6.59984425e-02 9.93752122e-01 7.76552200e-01
5.14887631e-01 8.01858604e-01 6.53507769e-01 -1.18829660e-01
1.17654467e+00 5.09447396e-01 4.75355387e-01 3.60482633e-01
-1.98870286e-01 -5.24133682e-01 1.75136924e-01 4.30330038e-01
-1.96523905e-01 9.90872979e-02 -1.09125173e+00 6.96958423e-01
-1.94583821e+00 -7.67781138e-01 -1.07220471e-01 1.81344056e+00
1.11088669e+00 -2.25410119e-01 -1.75178140e-01 -4.18924272e-01
1.50026217e-01 2.20562629e-02 -1.93188578e-01 -8.55706692e-01
1.42313763e-02 8.76452267e-01 4.83279020e-01 5.92663169e-01
-6.66317821e-01 1.56311941e+00 6.37391233e+00 5.02219915e-01
-1.82510197e-01 3.31432462e-01 2.27688830e-02 4.21770662e-01
-6.28557324e-01 4.74993587e-01 -9.05322909e-01 8.77222344e-02
1.40773547e+00 9.26968753e-02 5.26005924e-01 8.22377264e-01
-6.24349773e-01 5.35493717e-02 -1.35396385e+00 3.03871095e-01
-2.90106326e-01 -1.51700759e+00 5.51330984e-01 -5.01709521e-01
-9.78233069e-02 -1.64323729e-02 -6.09959483e-01 9.11950469e-01
1.05514681e+00 -1.15967631e+00 6.03867412e-01 5.50648808e-01
5.39169490e-01 -6.69265449e-01 7.35450149e-01 2.20185127e-02
-1.05299294e+00 -3.82382840e-01 -7.61027992e-01 -2.94522615e-04
-4.74173622e-03 1.79462895e-01 -6.91746891e-01 1.32411337e+00
8.91387761e-01 4.08184826e-01 -5.65872431e-01 3.90014142e-01
-6.63864672e-01 4.87105906e-01 -1.50888056e-01 3.20664316e-01
1.54688641e-01 -1.18674718e-01 1.40990719e-01 1.27432847e+00
1.42010346e-01 4.39567357e-01 -8.16182494e-02 8.11646104e-01
-1.66714832e-01 9.41156447e-02 -3.35115761e-01 1.20983915e-02
5.66028833e-01 1.03018963e+00 -5.12902886e-02 -6.89389825e-01
-6.62032545e-01 8.08890343e-01 7.96651661e-01 3.42540026e-01
-5.57725728e-01 -4.46245015e-01 7.48630464e-01 1.54186532e-01
4.29921180e-01 2.67866462e-01 1.47756562e-01 -1.25931621e+00
9.97488573e-02 -8.78510714e-01 1.12244523e+00 -1.21186626e+00
-1.23443592e+00 2.14179054e-01 2.86613554e-01 -2.75870562e-01
-3.36804718e-01 -7.44825482e-01 -4.01103169e-01 8.97113264e-01
-1.92852437e+00 -1.51898694e+00 2.75075305e-02 7.19354808e-01
1.19855337e-01 5.11190556e-02 1.39402735e+00 2.01791272e-01
-1.87639773e-01 1.80430844e-01 -6.96860731e-01 2.02356875e-01
4.63390827e-01 -1.62786937e+00 6.33154154e-01 4.94913936e-01
-7.93916062e-02 8.22473049e-01 6.02772653e-01 -4.29956704e-01
-1.66661775e+00 -1.10860193e+00 1.05485523e+00 -9.65658367e-01
8.38915050e-01 -3.24694723e-01 -1.22978294e+00 1.16392148e+00
2.10416302e-01 4.20793384e-01 6.56545341e-01 4.64890391e-01
-8.06392789e-01 -1.87416852e-01 -1.32850754e+00 1.68408871e-01
1.42149544e+00 -6.93584144e-01 -1.55482531e+00 7.41641223e-02
1.27556241e+00 -4.38522220e-01 -1.36151648e+00 3.19185764e-01
2.94240773e-01 -6.15692496e-01 1.21244979e+00 -1.51818967e+00
4.03383225e-01 -3.72582287e-01 -6.19834542e-01 -1.12309587e+00
-9.33346301e-02 -1.03486426e-01 -6.88267171e-01 1.29816151e+00
4.84037578e-01 -6.53941035e-01 8.99497807e-01 8.65601420e-01
-4.94131267e-01 -5.50149798e-01 -9.49610651e-01 -6.08320236e-01
3.24928790e-01 -5.64233124e-01 1.07348740e+00 1.09638834e+00
2.06205800e-01 5.97000480e-01 3.86330932e-01 5.57879567e-01
5.02327681e-01 1.51968136e-01 5.27186096e-01 -1.34939039e+00
-4.60900307e-01 1.69253469e-01 -2.67397553e-01 -9.03080642e-01
4.21689332e-01 -1.20373821e+00 -2.04212874e-01 -2.27672696e+00
-1.99097157e-01 -8.60135853e-01 -3.07730556e-01 1.09413946e+00
1.67415775e-02 -2.66739398e-01 -1.79283112e-01 -3.16967547e-01
-8.62138391e-01 3.22260469e-01 9.59572554e-01 6.76178560e-02
3.21642280e-01 -3.80717665e-01 -1.14725590e+00 5.90024054e-01
6.10995889e-01 -4.71122831e-01 -6.08607709e-01 -6.04770482e-01
8.54461670e-01 2.41768375e-01 5.87992311e-01 -6.39224410e-01
1.31248876e-01 -2.22482875e-01 -8.11362714e-02 1.93938036e-02
2.40677699e-01 -7.96874642e-01 -1.38059407e-01 -2.98358896e-03
-2.99851567e-01 -1.79301545e-01 2.46496245e-01 1.69013664e-01
-4.15515482e-01 -6.84029520e-01 4.67240155e-01 -6.76842809e-01
-1.21806753e+00 -7.69743783e-05 -3.04442970e-03 6.34692907e-01
4.93832886e-01 2.63131391e-02 -7.90361702e-01 2.43866388e-02
-7.94656396e-01 5.14629483e-01 2.97547340e-01 2.66165644e-01
2.66533405e-01 -9.70783710e-01 -2.73804814e-01 -1.48806170e-01
5.50324440e-01 5.22591710e-01 1.45595685e-01 3.15407246e-01
-6.96747780e-01 5.93703389e-01 -3.01436454e-01 3.91915888e-02
-6.39375329e-01 6.04026556e-01 4.91258264e-01 -4.73940820e-01
-7.29619563e-01 8.31245482e-01 -3.78335118e-02 -1.14914393e+00
-2.16250848e-02 -4.55985606e-01 -3.85012507e-01 -1.68286949e-01
3.34380895e-01 8.54001492e-02 4.21670467e-01 -1.74464628e-01
-5.06517887e-01 3.13814342e-01 2.42067669e-02 1.32294670e-01
1.55407286e+00 -6.83578178e-02 -5.42980850e-01 6.68975785e-02
8.82915199e-01 -2.73263101e-02 -6.92536235e-01 -2.17681453e-01
5.76927781e-01 -1.66939886e-03 -2.70548224e-01 -1.22774494e+00
-4.19218242e-01 6.27561569e-01 7.96148553e-02 2.45907262e-01
9.15571928e-01 6.91406786e-01 6.98679566e-01 1.01889205e+00
6.07539594e-01 -1.03020239e+00 -3.66781175e-01 8.34610045e-01
7.29141414e-01 -8.31862986e-01 -2.39844248e-01 -5.82889438e-01
-5.15720427e-01 9.86610651e-01 8.22090149e-01 5.12679666e-02
1.98121786e-01 2.61512756e-01 -6.93977475e-02 -7.72893131e-01
-9.88732576e-01 -4.90228385e-01 -3.39375734e-02 9.48213398e-01
3.71050209e-01 -6.65769875e-02 -1.14362843e-01 1.01211834e+00
-3.84204894e-01 1.48408428e-01 3.33842576e-01 1.23542833e+00
-6.55723572e-01 -1.40057397e+00 1.16389789e-01 2.89501846e-01
-8.00593495e-01 -4.48622972e-01 -1.47240773e-01 1.08311152e+00
1.61394894e-01 5.08913517e-01 1.71989292e-01 1.90036923e-01
8.55256498e-01 6.82744741e-01 7.95008481e-01 -1.10281789e+00
-6.35247588e-01 -1.01736343e+00 1.01159549e+00 -7.58551240e-01
-4.49958771e-01 -1.72468901e-01 -1.77028060e+00 -7.37885805e-03
9.93480459e-02 7.41464674e-01 3.45030338e-01 1.08272016e+00
5.29375315e-01 3.61180812e-01 -1.50713503e-01 3.72315943e-01
-3.81990969e-01 -6.49509132e-01 -4.04428214e-01 4.52766150e-01
-1.18707038e-01 -3.98053974e-01 1.34041309e-01 2.25929376e-02] | [10.286165237426758, 8.016737937927246] |
7735c60c-2385-49b6-b476-aa26c7447ff2 | sequential-neural-models-with-stochastic | 1605.07571 | null | http://arxiv.org/abs/1605.07571v2 | http://arxiv.org/pdf/1605.07571v2.pdf | Sequential Neural Models with Stochastic Layers | How can we efficiently propagate uncertainty in a latent state representation
with recurrent neural networks? This paper introduces stochastic recurrent
neural networks which glue a deterministic recurrent neural network and a state
space model together to form a stochastic and sequential neural generative
model. The clear separation of deterministic and stochastic layers allows a
structured variational inference network to track the factorization of the
model's posterior distribution. By retaining both the nonlinear recursive
structure of a recurrent neural network and averaging over the uncertainty in a
latent path, like a state space model, we improve the state of the art results
on the Blizzard and TIMIT speech modeling data sets by a large margin, while
achieving comparable performances to competing methods on polyphonic music
modeling. | ['Søren Kaae Sønderby', 'Ulrich Paquet', 'Ole Winther', 'Marco Fraccaro'] | 2016-05-24 | sequential-neural-models-with-stochastic-1 | http://papers.nips.cc/paper/6039-sequential-neural-models-with-stochastic-layers | http://papers.nips.cc/paper/6039-sequential-neural-models-with-stochastic-layers.pdf | neurips-2016-12 | ['music-modeling'] | ['music'] | [ 5.64632490e-02 2.04709470e-01 -2.40088016e-01 -2.47283310e-01
-6.66942596e-01 -5.45360565e-01 8.33207369e-01 -7.34087348e-01
1.48651870e-02 5.65547109e-01 7.60340571e-01 -4.18450117e-01
-9.70597789e-02 -4.17806864e-01 -6.84805930e-01 -7.97969639e-01
5.07957451e-02 4.92274344e-01 3.57985683e-02 4.11981065e-03
-1.77966014e-01 1.50046408e-01 -1.30371404e+00 3.82077575e-01
3.84913564e-01 7.75667548e-01 1.38064146e-01 1.01530206e+00
-2.71400303e-01 1.36514151e+00 -4.10721362e-01 -1.89368889e-01
-1.31693140e-01 -7.80361950e-01 -4.71098125e-01 -1.73321590e-01
1.30185142e-01 -2.79748529e-01 -6.35786474e-01 9.58113730e-01
3.23994577e-01 3.51910740e-01 9.40867782e-01 -7.30162203e-01
-6.31724477e-01 1.42965651e+00 5.18236905e-02 1.53972402e-01
-1.11525953e-01 -2.66485721e-01 1.28415394e+00 -6.50249124e-01
5.54321706e-01 1.48489761e+00 9.38126385e-01 6.08837485e-01
-1.53984642e+00 -5.05776227e-01 4.24608618e-01 -1.37032047e-01
-1.12566257e+00 -8.76822412e-01 9.59996641e-01 -5.10955155e-01
1.11102200e+00 5.62777072e-02 5.49543560e-01 1.62428594e+00
1.51316002e-01 1.22420049e+00 5.61969697e-01 -3.39635253e-01
1.95664942e-01 -1.94230124e-01 1.24214254e-01 6.66817605e-01
-3.20337236e-01 5.83560348e-01 -6.64483428e-01 -3.05706561e-01
9.39066768e-01 1.05327323e-01 5.45619056e-02 -4.19085890e-01
-1.01261735e+00 9.26880658e-01 1.41571954e-01 1.66862324e-01
-4.13324594e-01 9.50740755e-01 4.00000840e-01 1.92202166e-01
5.64304233e-01 4.91568334e-02 -3.24673653e-01 -3.27537060e-01
-1.59767139e+00 1.41155615e-01 9.15283203e-01 6.30420268e-01
1.43364415e-01 9.24760938e-01 -3.50424886e-01 8.42750967e-01
7.17921376e-01 6.86588883e-01 6.31549239e-01 -1.29787338e+00
1.27624884e-01 -2.02021286e-01 1.60894781e-01 -5.37730575e-01
-1.36061370e-01 -8.63468111e-01 -9.84032035e-01 1.39153153e-01
1.99940696e-01 -5.11512816e-01 -1.13913834e+00 2.07375431e+00
-3.43906224e-01 7.16357708e-01 2.38745391e-01 3.23306412e-01
4.37617600e-01 1.04793310e+00 -3.60906988e-01 -5.18508375e-01
8.19480181e-01 -9.97628927e-01 -9.92328048e-01 -1.55859500e-01
-1.16678059e-01 -6.64662242e-01 4.32002932e-01 3.98548543e-01
-1.45314837e+00 -7.14874923e-01 -1.06042159e+00 1.35074213e-01
-3.26265092e-03 3.50707531e-01 5.36151946e-01 6.72011912e-01
-1.08830321e+00 9.52658713e-01 -1.58651042e+00 1.54586285e-01
-9.45632607e-02 7.18643516e-02 3.70879441e-01 4.62884575e-01
-1.31197536e+00 7.28036880e-01 4.83122170e-01 3.88503313e-01
-1.43709540e+00 -5.89986682e-01 -7.30497658e-01 1.17288940e-01
3.97954024e-02 -8.71449113e-01 1.67017102e+00 -3.65352452e-01
-2.20543575e+00 1.27088562e-01 -5.97359538e-01 -9.56704855e-01
4.09160286e-01 -3.60176027e-01 -5.63820362e-01 -2.74120480e-01
-3.19810629e-01 4.09847558e-01 1.29005015e+00 -8.73511672e-01
-4.52707767e-01 1.22800350e-01 -4.72146869e-01 -8.72711837e-03
1.12012587e-01 -2.79185381e-02 -3.83653313e-01 -8.20384741e-01
3.69915456e-01 -1.14217687e+00 -4.78967011e-01 -7.43427992e-01
-4.88311708e-01 -7.87004679e-02 7.00321138e-01 -4.57614481e-01
1.56634319e+00 -2.00216699e+00 7.13449121e-01 -2.51260158e-02
-1.04971163e-01 6.35153726e-02 6.49990067e-02 4.74033505e-01
-2.41796747e-01 -1.64154187e-01 -6.29631057e-02 -9.36458528e-01
3.72953117e-01 4.63192612e-01 -1.20754480e+00 1.39355361e-01
-4.54057716e-02 1.24046624e+00 -7.23598480e-01 5.35777546e-02
1.76614478e-01 6.44331694e-01 -5.46011865e-01 1.56766206e-01
-6.17811680e-01 2.01234445e-01 -4.54305261e-02 1.10344432e-01
-7.34365284e-02 -2.87407368e-01 2.06691727e-01 1.27761900e-01
-9.54859629e-02 8.01053286e-01 -1.40268397e+00 1.95203853e+00
-4.23771083e-01 7.61644065e-01 -5.45895770e-02 -5.97567320e-01
8.44271481e-01 8.52846384e-01 1.87858641e-01 1.95382401e-01
-2.01756377e-02 1.84445783e-01 -1.89254090e-01 2.29183361e-01
5.22920012e-01 -4.98069376e-01 -2.34487876e-01 8.83636951e-01
4.95193809e-01 -1.73147574e-01 3.86805013e-02 2.61603594e-01
6.86789989e-01 4.97082204e-01 -1.50546953e-01 3.21884304e-02
8.04713145e-02 -4.54937339e-01 5.94005764e-01 1.10444069e+00
9.53200534e-02 7.23592997e-01 4.80617672e-01 -4.35000062e-01
-9.53612864e-01 -1.59410191e+00 5.63268736e-02 1.14531112e+00
-6.95841491e-01 -5.57201624e-01 -4.14744258e-01 -1.36582360e-01
-1.88709110e-01 1.08506334e+00 -7.09608674e-01 -2.54727244e-01
-4.56775427e-01 -7.83247530e-01 8.35652113e-01 6.84195101e-01
-4.97584864e-02 -1.10235679e+00 -5.31526133e-02 6.61543727e-01
-1.57434046e-01 -7.79206276e-01 -4.51219499e-01 7.77351379e-01
-1.03797722e+00 -2.58869320e-01 -8.43158543e-01 -3.70412797e-01
-5.23540936e-02 -3.95325631e-01 1.11147916e+00 -8.96594346e-01
1.62178233e-01 -4.29385109e-03 2.66712815e-01 -5.11632740e-01
-8.38148952e-01 6.88696206e-02 3.26071620e-01 -1.47110060e-01
-3.01466826e-02 -9.81159210e-01 -1.34655699e-01 4.02782746e-02
-6.94556236e-01 -2.29717232e-03 1.94111928e-01 9.89870787e-01
4.50008392e-01 1.99717522e-01 2.99712121e-01 -6.92284763e-01
6.76969707e-01 -3.00090015e-01 -7.48177528e-01 1.39200300e-01
-5.81274092e-01 5.96666694e-01 1.67721793e-01 -5.32836139e-01
-1.34729254e+00 1.16867729e-01 -1.80561081e-01 -8.90289009e-01
1.34633437e-01 6.41295612e-01 4.15861309e-01 7.64505684e-01
5.55662692e-01 4.50806051e-01 7.24085718e-02 -5.91651976e-01
9.36724722e-01 2.22440615e-01 8.80290747e-01 -4.08985704e-01
6.57342255e-01 4.78304505e-01 -2.94976741e-01 -4.31277394e-01
-1.17595494e+00 -2.77526915e-01 -5.58241248e-01 1.25470191e-01
8.23750496e-01 -1.10377467e+00 -5.78732371e-01 5.95962226e-01
-1.27781773e+00 -2.50518233e-01 -8.56328785e-01 7.51675248e-01
-9.80983973e-01 1.85363106e-02 -1.18048239e+00 -1.26830924e+00
-2.06439972e-01 -9.70789135e-01 8.54405880e-01 1.24067888e-02
-5.15101790e-01 -1.16790473e+00 6.52004480e-01 -3.08692664e-01
5.41845560e-01 -1.55777097e-01 7.87982464e-01 -3.18522245e-01
-5.88003278e-01 -1.78207025e-01 2.85337150e-01 6.65143847e-01
7.51481280e-02 3.36370915e-01 -1.09259176e+00 -5.95142096e-02
2.56807655e-01 -3.76363583e-02 1.44212651e+00 1.08215606e+00
3.77908260e-01 -1.85856819e-01 -1.21214457e-01 7.57655680e-01
1.02770042e+00 2.12943554e-01 4.46867853e-01 -3.50096673e-01
5.92723966e-01 1.37318879e-01 -3.93677741e-01 3.26802313e-01
-3.26160155e-02 2.59988070e-01 1.92837119e-01 3.94580096e-01
-2.21491739e-01 -7.12527037e-01 8.52499664e-01 1.48759961e+00
-1.28784835e-01 -2.07053423e-01 -7.38477051e-01 5.76045454e-01
-2.18847179e+00 -1.54076374e+00 1.93069369e-01 1.80443740e+00
8.03735793e-01 4.67574686e-01 5.13807312e-02 1.15359817e-02
4.45103526e-01 6.40431643e-01 -6.06801689e-01 -5.32724679e-01
-2.14026511e-01 6.64063841e-02 2.25186735e-01 7.79451072e-01
-1.01608169e+00 1.06558919e+00 8.70631027e+00 9.12837923e-01
-8.87065232e-01 1.74572691e-01 2.22176358e-01 -5.15813708e-01
-5.69255352e-01 -1.53165013e-01 -1.22253346e+00 2.71142870e-01
1.63006151e+00 -2.53181085e-02 5.88112652e-01 6.50614440e-01
3.54766518e-01 4.69862968e-01 -1.04510224e+00 7.45662272e-01
-1.21388622e-01 -1.72375667e+00 1.54606342e-01 2.79266033e-02
1.04427278e+00 7.33793080e-01 5.85799873e-01 5.10066748e-01
1.44621515e+00 -1.11631656e+00 1.01157284e+00 9.85678494e-01
4.94862646e-01 -7.90832937e-01 3.85527372e-01 5.53417802e-01
-9.06649172e-01 -1.38292551e-01 -2.90500373e-01 -1.58352301e-01
5.77955604e-01 5.65973341e-01 -5.52623153e-01 1.72779858e-01
4.36567068e-01 8.96902204e-01 1.18587546e-01 6.85930312e-01
-5.14524579e-01 1.13503301e+00 -4.80282396e-01 1.02711983e-01
6.14883423e-01 -2.25321934e-01 9.43925023e-01 1.11741257e+00
2.62768745e-01 -3.52074444e-01 -1.98835567e-01 1.32881820e+00
6.00070804e-02 -6.20488286e-01 -4.72922027e-01 -5.22682130e-01
-1.75935104e-02 6.71957910e-01 -4.59672481e-01 -3.41692686e-01
6.28649518e-02 9.26808655e-01 1.19670816e-01 8.40750098e-01
-7.53157973e-01 1.63073130e-02 7.88829267e-01 -4.54374611e-01
8.41068029e-01 -6.32870734e-01 -1.48758128e-01 -1.46577096e+00
-4.66853589e-01 -4.50294256e-01 2.61558384e-01 -8.89015019e-01
-1.04404485e+00 7.93408751e-01 -3.90067920e-02 -9.64680314e-01
-1.40577555e+00 -3.77361149e-01 -7.09511399e-01 1.34004700e+00
-1.17094791e+00 -1.00624120e+00 7.55190670e-01 3.33899945e-01
5.99709034e-01 -4.88912374e-01 1.16139460e+00 -3.08666378e-01
-6.11408174e-01 1.73559502e-01 7.61628747e-01 2.65688188e-02
1.92558184e-01 -1.33995986e+00 9.11907136e-01 1.05308568e+00
9.49972808e-01 1.02903152e+00 8.17430735e-01 -6.56416893e-01
-9.69026625e-01 -8.68758202e-01 8.58686805e-01 -8.54766548e-01
1.09840095e+00 -3.64381254e-01 -8.17611694e-01 1.08844495e+00
2.87935227e-01 -1.66560560e-01 5.59824765e-01 6.66108966e-01
-5.16272366e-01 4.10542369e-01 -3.31357956e-01 6.71864271e-01
7.54753411e-01 -1.15906572e+00 -1.00145197e+00 -8.88141096e-02
1.18481851e+00 -3.36003095e-01 -5.05777359e-01 1.57953933e-01
8.79119575e-01 -7.85381675e-01 1.03946936e+00 -9.21159267e-01
3.92607838e-01 -2.88546145e-01 -3.47412318e-01 -1.42827868e+00
-6.90256655e-01 -1.28044307e+00 -9.08492804e-01 9.62452114e-01
6.80372655e-01 -1.92320615e-01 9.58389819e-01 1.80626884e-01
-8.93570408e-02 -3.74513566e-01 -9.20392215e-01 -7.76120305e-01
1.63676381e-01 -1.04641306e+00 3.31428975e-01 4.07526702e-01
-3.26800853e-01 4.68870461e-01 -7.80909598e-01 1.52767107e-01
6.44636810e-01 1.60307735e-01 1.41860202e-01 -1.19968092e+00
-6.77597940e-01 -6.78573370e-01 -1.62452087e-02 -1.36407113e+00
3.12594682e-01 -9.44274902e-01 1.89541966e-01 -1.41697645e+00
-1.61578536e-01 3.96081284e-02 -8.80008519e-01 2.02411085e-01
9.35698971e-02 1.17168806e-01 5.63374832e-02 4.20012802e-01
-5.50409019e-01 7.69534469e-01 6.99702501e-01 -3.20182480e-02
-5.01719773e-01 5.29742837e-01 -5.17722309e-01 1.13989782e+00
4.04262900e-01 -5.83179355e-01 -5.92790484e-01 -4.56760824e-01
5.00660658e-01 4.85403836e-01 1.55643567e-01 -9.30267572e-01
5.20153821e-01 1.24780573e-01 2.77527869e-01 -1.06883574e+00
7.24311113e-01 -3.94983292e-01 4.79582816e-01 4.22008276e-01
-6.86342180e-01 -2.33003229e-01 1.47894278e-01 9.48553741e-01
-3.64695758e-01 -1.48969859e-01 5.51605463e-01 -8.23070481e-02
-3.32136214e-01 2.14245483e-01 -6.53046310e-01 -1.88035488e-01
1.17208496e-01 8.17356557e-02 7.69110471e-02 -6.62099957e-01
-1.50694823e+00 -1.80717111e-01 -1.55889884e-01 6.54642820e-01
3.68981630e-01 -1.45099163e+00 -7.30402946e-01 3.88244957e-01
-4.47238177e-01 -4.37322944e-01 1.86037377e-01 4.51873749e-01
1.48673251e-01 8.02014530e-01 1.53118744e-01 -6.90678656e-01
-7.17126906e-01 2.57663369e-01 5.84579945e-01 -4.65417683e-01
-5.89752436e-01 1.17416811e+00 -5.12186363e-02 -5.23998797e-01
5.17466307e-01 -6.50414526e-01 -9.58755612e-02 3.40927124e-01
4.86037672e-01 2.73234397e-01 -2.41463676e-01 -4.06973600e-01
9.20979492e-03 2.21961901e-01 9.11648944e-02 -8.02158237e-01
1.30838835e+00 -3.19582343e-01 4.47600409e-02 1.53190911e+00
1.06663859e+00 -5.44303702e-03 -1.54307592e+00 -5.89369655e-01
-2.19702106e-02 4.40514624e-01 4.12044734e-01 -7.15932190e-01
-7.81517744e-01 1.01853764e+00 3.51684839e-01 2.13040963e-01
5.55680811e-01 -1.21111078e-02 6.56717956e-01 4.19826210e-01
3.74508314e-02 -9.14405107e-01 -1.72563970e-01 1.09063530e+00
8.00113440e-01 -8.19351196e-01 -2.98930764e-01 3.56870443e-01
-5.16565502e-01 1.00342560e+00 -5.35308421e-01 -4.20000881e-01
1.12599683e+00 6.00014627e-01 1.45474449e-01 -7.85696581e-02
-1.31228447e+00 -3.36483978e-02 6.24308050e-01 1.19955197e-01
2.83091068e-01 1.99437484e-01 6.09211624e-01 7.90877402e-01
-4.87755418e-01 9.99106169e-02 3.58216137e-01 4.63760972e-01
-4.62889820e-01 -9.74089384e-01 -1.79104820e-01 1.98449641e-01
-6.23365700e-01 -3.79492581e-01 9.50863212e-02 1.99899450e-01
-3.19570243e-01 7.83259571e-01 1.72589913e-01 -2.15312034e-01
-1.32310223e-02 6.71741426e-01 2.55885720e-01 -7.27830827e-01
-4.31582093e-01 7.68440366e-01 6.45719767e-02 -4.67662245e-01
-2.89728522e-01 -9.01988089e-01 -9.19086337e-01 -3.40027288e-02
-3.14898938e-01 3.21879894e-01 8.02151322e-01 9.21247005e-01
6.27173111e-02 9.55218613e-01 2.84081250e-01 -7.30477989e-01
-8.76454175e-01 -1.03395545e+00 -7.51322269e-01 -3.12611729e-01
5.13333559e-01 -2.84458488e-01 -5.00744998e-01 2.47997209e-01] | [15.387662887573242, 5.8761515617370605] |
4f648dbf-b77e-4732-b8ca-f835c340a251 | a-benchmark-for-compositional-visual | 2206.05379 | null | https://arxiv.org/abs/2206.05379v1 | https://arxiv.org/pdf/2206.05379v1.pdf | A Benchmark for Compositional Visual Reasoning | A fundamental component of human vision is our ability to parse complex visual scenes and judge the relations between their constituent objects. AI benchmarks for visual reasoning have driven rapid progress in recent years with state-of-the-art systems now reaching human accuracy on some of these benchmarks. Yet, a major gap remains in terms of the sample efficiency with which humans and AI systems learn new visual reasoning tasks. Humans' remarkable efficiency at learning has been at least partially attributed to their ability to harness compositionality -- such that they can efficiently take advantage of previously gained knowledge when learning new tasks. Here, we introduce a novel visual reasoning benchmark, Compositional Visual Relations (CVR), to drive progress towards the development of more data-efficient learning algorithms. We take inspiration from fluidic intelligence and non-verbal reasoning tests and describe a novel method for creating compositions of abstract rules and associated image datasets at scale. Our proposed benchmark includes measures of sample efficiency, generalization and transfer across task rules, as well as the ability to leverage compositionality. We systematically evaluate modern neural architectures and find that, surprisingly, convolutional architectures surpass transformer-based architectures across all performance measures in most data regimes. However, all computational models are a lot less data efficient compared to humans even after learning informative visual representations using self-supervision. Overall, we hope that our challenge will spur interest in the development of neural architectures that can learn to harness compositionality toward more efficient learning. | ['Thomas Serre', 'Sebastian Musslick', 'Julien Colin', 'Mohit Vaishnav', 'Aimen Zerroug'] | 2022-06-11 | null | null | null | null | ['visual-reasoning', 'visual-reasoning'] | ['computer-vision', 'reasoning'] | [ 4.15217519e-01 3.20839822e-01 -5.17814048e-03 -5.63290834e-01
-2.55307406e-01 -6.53957784e-01 1.06325448e+00 1.31463349e-01
-3.24040532e-01 4.64326739e-01 3.75201792e-01 -4.22531366e-01
-2.76423246e-01 -8.61715019e-01 -8.20223331e-01 -2.85216480e-01
1.08126486e-02 6.11841202e-01 2.92365968e-01 -3.06078374e-01
1.95985496e-01 5.44421315e-01 -1.68890977e+00 7.41382360e-01
7.55450130e-01 8.75475109e-01 6.57030195e-02 7.73289561e-01
-1.28546417e-01 1.65828526e+00 -3.09136122e-01 -8.11893702e-01
3.07065934e-01 -4.61593360e-01 -1.16384637e+00 -2.62881279e-01
8.06757510e-01 -1.40183508e-01 -4.48731422e-01 7.50427902e-01
7.48498067e-02 1.14798188e-01 8.69467437e-01 -1.12393439e+00
-1.20117676e+00 7.85851240e-01 -2.08129480e-01 4.23074752e-01
1.42784476e-01 5.50193548e-01 1.38588667e+00 -6.72712862e-01
8.09791684e-01 1.25864363e+00 4.71392900e-01 7.03607976e-01
-1.40668952e+00 -6.13423228e-01 1.87826201e-01 5.80759406e-01
-7.63755500e-01 -6.00326002e-01 5.64119220e-01 -5.47804415e-01
1.45494461e+00 1.14960603e-01 8.44649315e-01 8.89171898e-01
-4.24574427e-02 9.44639623e-01 1.24048591e+00 -4.55129296e-01
2.97483597e-02 -1.01210468e-01 1.38834625e-01 1.17003059e+00
2.56179541e-01 9.39206406e-02 -7.17048109e-01 4.44446027e-01
6.15328372e-01 -8.88649002e-02 -1.33081257e-01 -6.10020399e-01
-1.34426880e+00 6.97933853e-01 9.73729849e-01 2.82843918e-01
-1.31389976e-01 2.94288397e-01 4.85132545e-01 5.14410138e-01
-3.91539447e-02 9.69326854e-01 -2.76800156e-01 3.93739603e-02
-6.05097651e-01 2.98800379e-01 6.60410881e-01 8.19852114e-01
6.80174112e-01 1.84229583e-01 -2.56236225e-01 7.40600646e-01
9.39017907e-03 3.85159135e-01 3.48973691e-01 -1.25596631e+00
4.13450837e-01 1.06340826e+00 -5.33421278e-01 -7.24819303e-01
-3.73325825e-01 -4.68211025e-01 -8.91349673e-01 4.86343652e-01
5.64522862e-01 3.70312899e-01 -9.47412789e-01 1.76260543e+00
-1.55319393e-01 -3.09168488e-01 3.34424376e-01 6.71939611e-01
8.44519675e-01 4.90888476e-01 1.70143068e-01 8.09123144e-02
1.39500976e+00 -8.38616073e-01 -1.71323448e-01 -5.23246586e-01
4.46706623e-01 -3.57776254e-01 1.35939002e+00 5.50082624e-01
-1.36813521e+00 -7.33600855e-01 -1.18964887e+00 -6.02836609e-01
-4.97130811e-01 -2.94001848e-01 9.37861323e-01 2.65556991e-01
-1.07875013e+00 4.85170245e-01 -4.67674136e-01 -3.81367534e-01
1.11709714e+00 4.39365208e-01 -3.97096097e-01 -3.55724066e-01
-7.66172111e-01 1.26850677e+00 4.50926930e-01 -3.28070223e-01
-9.78618741e-01 -9.17160690e-01 -6.84991002e-01 2.72267282e-01
3.62733692e-01 -1.06379461e+00 1.33329725e+00 -1.23225129e+00
-1.05558169e+00 1.17582023e+00 1.17397634e-02 -8.92676115e-01
3.68375540e-01 -7.87125677e-02 -1.14912868e-01 3.52743506e-01
-7.08029941e-02 1.05006349e+00 7.48023450e-01 -1.30978262e+00
-7.32096553e-01 -2.16868967e-01 3.98889333e-01 1.12411290e-01
-1.79616898e-01 -1.94878846e-01 -1.26420096e-01 -4.38042819e-01
-2.58007258e-01 -8.67887020e-01 2.27888584e-01 3.20563853e-01
-4.50814106e-02 -5.85193574e-01 5.30254483e-01 -3.81360412e-01
4.86088634e-01 -2.02428722e+00 3.32386017e-01 -3.87874432e-02
6.45968199e-01 5.43247640e-01 -1.63620487e-01 1.60220608e-01
-1.19283600e-02 2.72354055e-02 -1.91757053e-01 -5.89089915e-02
1.15498886e-01 4.31020826e-01 -7.62974739e-01 -9.57584679e-02
5.75945616e-01 1.45785737e+00 -9.34287608e-01 -3.92606795e-01
2.05752641e-01 3.31362575e-01 -6.85304761e-01 1.99790314e-01
-5.38819432e-01 1.08428039e-01 2.92825103e-02 3.74485791e-01
6.91430420e-02 -6.17515206e-01 3.09152842e-01 -2.79766679e-01
1.14665113e-01 3.18553567e-01 -4.46495175e-01 1.51012659e+00
-5.18311203e-01 1.04547036e+00 -5.52736700e-01 -1.29548407e+00
8.00578952e-01 -3.55646759e-02 -2.74506006e-02 -1.20524156e+00
3.41391861e-02 -5.21529280e-02 5.58134198e-01 -4.27594572e-01
1.50858372e-01 -2.20261395e-01 3.57695043e-01 3.65714341e-01
2.41685659e-01 -3.64332467e-01 4.45019782e-01 3.45143050e-01
1.16345298e+00 2.01845139e-01 5.22930801e-01 -1.12288222e-01
3.57725322e-01 2.77435213e-01 1.36818156e-01 7.52264440e-01
-1.05461039e-01 1.99843809e-01 5.55740595e-01 -1.06680155e+00
-1.20634913e+00 -1.22765195e+00 2.24924400e-01 1.45893407e+00
-2.30464935e-01 -4.46889549e-01 -3.78722340e-01 -5.02802789e-01
5.13666198e-02 7.48390615e-01 -8.61665010e-01 -3.18345368e-01
-6.75926208e-01 -3.08439970e-01 6.89200163e-01 8.75308514e-01
8.24359596e-01 -1.42087221e+00 -1.10860872e+00 -8.73974636e-02
1.73715457e-01 -1.30759084e+00 1.46627367e-01 2.22772256e-01
-8.79581869e-01 -1.20874822e+00 -3.78126293e-01 -8.44539821e-01
6.11348391e-01 2.78497636e-01 1.58305848e+00 3.14149857e-01
-5.34647703e-01 4.72846776e-01 -1.37059867e-01 -5.24271727e-01
-5.24601281e-01 -1.05300676e-02 -2.25170404e-01 -2.97180861e-01
3.43622923e-01 -6.33246660e-01 -5.11608899e-01 9.17689279e-02
-7.88365364e-01 4.98858333e-01 8.54487181e-01 7.60154605e-01
3.36131990e-01 -1.49065226e-01 4.14124846e-01 -9.21215713e-01
5.90037465e-01 -1.95341662e-01 -5.53159416e-01 5.28686941e-01
-6.46436274e-01 6.93840146e-01 9.10241425e-01 -2.75646240e-01
-1.04745054e+00 -1.46712974e-01 3.79949987e-01 -2.71994859e-01
-5.09470440e-02 2.65717179e-01 1.57595575e-01 -4.51569594e-02
1.00333834e+00 2.09850132e-01 -5.63614108e-02 7.24825114e-02
7.57918477e-01 1.56553373e-01 8.04767191e-01 -9.66177225e-01
7.57917285e-01 5.18361628e-01 4.27809149e-01 -7.35917091e-01
-1.19622827e+00 -1.90364514e-02 -5.73161662e-01 3.03300284e-02
1.05658472e+00 -9.32143748e-01 -1.25909889e+00 6.06198981e-02
-1.10614789e+00 -6.70521617e-01 -4.69736516e-01 1.16202936e-01
-7.68892348e-01 -3.34098339e-02 -4.37058330e-01 -4.88942087e-01
-3.28455776e-01 -1.09170699e+00 5.95948219e-01 2.67267913e-01
-4.58294392e-01 -8.72288346e-01 -5.24520427e-02 7.01520503e-01
5.29392481e-01 4.45248149e-02 1.56416869e+00 -5.05791903e-01
-9.93058503e-01 3.52566451e-01 -8.62064779e-01 3.90005648e-01
-2.77091652e-01 -1.18695751e-01 -1.05340493e+00 -4.94873524e-02
-4.71163601e-01 -9.90213633e-01 1.28105843e+00 -3.13826948e-02
1.15028846e+00 -2.31487453e-01 -8.98997262e-02 7.17258573e-01
1.27422822e+00 1.23214632e-01 4.91512179e-01 2.38700330e-01
7.33711064e-01 6.08568907e-01 -9.81487855e-02 -8.99914429e-02
6.87555790e-01 4.58251357e-01 3.80627781e-01 1.42226398e-01
-7.10127532e-01 -2.91447908e-01 8.52242336e-02 4.68903869e-01
-6.59287512e-01 1.40458763e-01 -1.34486747e+00 4.67962861e-01
-1.73886824e+00 -1.24684143e+00 3.41400236e-01 1.87954414e+00
1.02480376e+00 4.25559670e-01 4.85354401e-02 1.72315463e-01
1.20293237e-02 -9.96799488e-03 -8.08102906e-01 -5.69131315e-01
-2.03351080e-01 5.48972547e-01 1.29395306e-01 3.95046353e-01
-7.71705747e-01 1.07068419e+00 6.41056299e+00 2.98062801e-01
-8.58448684e-01 -2.37405136e-01 5.33264697e-01 -2.36529838e-02
-2.53030449e-01 4.23644632e-02 -5.87064326e-01 -3.05843234e-01
9.17710960e-01 -1.10321706e-02 8.29844177e-01 7.37549484e-01
-6.26576245e-01 -7.99010228e-03 -1.61140013e+00 1.12925065e+00
4.07021046e-01 -1.84137213e+00 5.28668225e-01 -8.07066262e-03
7.99580693e-01 1.28089294e-01 2.10119337e-01 5.75705409e-01
8.15656900e-01 -1.42238069e+00 6.34623289e-01 6.46590710e-01
6.67391658e-01 -5.77927828e-01 2.50537544e-01 5.50072640e-02
-1.07211304e+00 -4.15880084e-01 -4.42829192e-01 -4.66747910e-01
-4.77866381e-01 1.93825349e-01 -9.89772797e-01 1.27919465e-01
7.05548942e-01 8.63640785e-01 -1.08058572e+00 6.49201214e-01
-4.98824656e-01 3.39695632e-01 6.82170242e-02 -8.84874687e-02
1.78393915e-01 2.71373153e-01 5.21330945e-02 1.09785736e+00
-1.48890704e-01 1.30176231e-01 -1.00046389e-01 1.01788628e+00
-3.74083370e-01 -2.87020892e-01 -8.18794012e-01 -1.99877381e-01
2.90651649e-01 1.00990188e+00 -6.69026256e-01 -5.13173223e-01
-6.13234222e-01 7.01882601e-01 9.98530924e-01 2.91871279e-01
-5.28659105e-01 -6.25691563e-02 6.06535435e-01 -5.28354235e-02
3.51796865e-01 -4.02459860e-01 -4.85498756e-01 -1.15391159e+00
-3.32598947e-02 -1.22948813e+00 4.94400859e-01 -1.08050525e+00
-1.25271392e+00 5.55420637e-01 2.84353569e-02 -5.55546701e-01
-3.40891868e-01 -1.13818216e+00 -3.30580384e-01 4.53513414e-01
-1.50949395e+00 -1.40336120e+00 -4.70334888e-01 8.32895815e-01
5.75405777e-01 -6.09099925e-01 8.66859853e-01 -2.68489033e-01
-8.29110593e-02 4.98027176e-01 -5.12910366e-01 4.68711644e-01
4.59752709e-01 -1.40881109e+00 6.28546476e-01 6.78284705e-01
7.87673891e-01 6.64753199e-01 5.20101428e-01 -2.03953102e-01
-1.43271661e+00 -7.28105128e-01 4.88251090e-01 -8.16835701e-01
6.36414111e-01 -5.32271683e-01 -8.28791797e-01 9.39792395e-01
4.46050107e-01 2.16906875e-01 5.78388870e-01 4.16826904e-01
-1.28401756e+00 -3.72165799e-01 -6.90230966e-01 9.65953588e-01
1.47760141e+00 -8.30321729e-01 -1.19598401e+00 2.17120633e-01
7.47859359e-01 3.07920054e-02 -7.00992584e-01 3.87454867e-01
7.66844988e-01 -1.32050192e+00 1.16940546e+00 -9.68156278e-01
7.66113341e-01 -1.46942809e-01 -2.12686688e-01 -1.16742539e+00
-6.03805661e-01 -4.30843771e-01 -2.56775796e-01 8.78312528e-01
3.14068675e-01 -4.63175654e-01 5.84172785e-01 5.70710897e-01
3.63832414e-02 -6.38831794e-01 -5.34661293e-01 -6.39883876e-01
2.52391011e-01 -4.31268215e-01 4.22641218e-01 7.89102137e-01
-4.16737013e-02 8.86276960e-01 1.03382662e-01 -2.67533422e-01
7.10029483e-01 3.33668917e-01 9.12262261e-01 -1.41847360e+00
-3.34775925e-01 -9.48658645e-01 -5.90271652e-01 -7.07562625e-01
3.21323872e-01 -1.19880915e+00 -3.13352078e-01 -1.71846354e+00
3.56038213e-01 -1.14559650e-01 -2.17931688e-01 8.25469732e-01
-1.52464407e-02 3.65203917e-01 5.98667800e-01 2.85683364e-01
-6.63962781e-01 2.05643207e-01 1.46095836e+00 -3.92652363e-01
2.05865458e-01 -4.82339174e-01 -9.09453511e-01 8.10974658e-01
6.77448750e-01 1.42673338e-02 -8.09173584e-01 -6.73412979e-01
8.17183793e-01 -3.39622766e-01 7.70985961e-01 -1.39345467e+00
2.62129813e-01 -2.01185465e-01 8.21164489e-01 -4.71985191e-02
2.01207355e-01 -7.64486313e-01 -2.29035854e-01 5.15731514e-01
-7.69829333e-01 1.59321681e-01 3.16317827e-01 3.17125171e-01
-2.17232406e-02 2.02610314e-01 8.03216219e-01 -3.60582948e-01
-1.05720699e+00 1.04449615e-01 4.80773076e-02 5.42033195e-01
8.65735948e-01 -6.44835606e-02 -1.01954842e+00 -2.04068556e-01
-5.12206018e-01 1.64114218e-02 3.19012940e-01 5.01958311e-01
6.93037927e-01 -1.05579233e+00 -7.62578845e-01 1.49924219e-01
4.30288345e-01 -9.30033848e-02 -6.53067743e-03 5.11077642e-01
-6.55391157e-01 4.82733727e-01 -7.39039183e-01 -5.01605749e-01
-1.09536183e+00 8.21681261e-01 3.31158936e-01 -2.36611307e-01
-8.02838206e-01 9.00829852e-01 4.28375363e-01 -1.49190158e-01
1.85139969e-01 -6.65465474e-01 -4.67649736e-02 -7.07678422e-02
6.99155629e-01 2.15729214e-02 -1.90553218e-01 -2.65260279e-01
-2.30687454e-01 6.42959833e-01 -2.40864858e-01 1.41027510e-01
1.38558531e+00 4.23618376e-01 -1.22155413e-01 4.04302001e-01
9.03168797e-01 -3.77292424e-01 -1.36245060e+00 -4.02295113e-01
1.16191827e-01 -1.22356944e-01 -3.35521907e-01 -1.11960542e+00
-8.49587679e-01 1.17310166e+00 3.13248634e-01 1.35865241e-01
1.10390687e+00 2.39601538e-01 4.05257493e-01 1.00909889e+00
1.37826145e-01 -7.75605977e-01 5.23522973e-01 7.62836695e-01
9.65817988e-01 -1.37813079e+00 1.68646872e-01 -3.24927759e-03
-7.52323389e-01 1.20072293e+00 6.30293131e-01 -3.21117938e-01
2.40388080e-01 1.70275077e-01 -1.02285244e-01 -4.39290553e-01
-1.19137323e+00 -4.79474694e-01 6.53037548e-01 9.14023936e-01
5.17236948e-01 1.07147433e-02 4.91000384e-01 8.39407444e-02
-4.89408910e-01 8.94287005e-02 1.67989030e-01 6.06846392e-01
-5.19861639e-01 -7.63792992e-01 9.50227156e-02 5.39913833e-01
-3.56506854e-02 -2.23995760e-01 -5.51415384e-01 9.06033814e-01
2.71951318e-01 5.89054406e-01 2.00871736e-01 -1.00303613e-01
2.99080491e-01 3.46176982e-01 1.05714297e+00 -5.39508581e-01
-5.74045002e-01 -9.00599360e-01 1.08676277e-01 -6.94720268e-01
-6.82620287e-01 -3.62943172e-01 -1.40619516e+00 -2.87405044e-01
4.24443156e-01 -3.59086037e-01 2.47694045e-01 1.16775799e+00
1.86132893e-01 6.83548510e-01 -1.77285939e-01 -5.67998588e-01
-3.72979015e-01 -5.88049293e-01 5.48127741e-02 7.33415365e-01
3.07398319e-01 -5.51533282e-01 1.41947800e-02 3.12537879e-01] | [10.561445236206055, 2.2422144412994385] |
a00b7256-f71a-4f39-ab33-8b46a3cf6a69 | iris-presentation-attack-detection-based-on | 1811.07252 | null | http://arxiv.org/abs/1811.07252v1 | http://arxiv.org/pdf/1811.07252v1.pdf | Iris Presentation Attack Detection Based on Photometric Stereo Features | We propose a new iris presentation attack detection method using
three-dimensional features of an observed iris region estimated by photometric
stereo. Our implementation uses a pair of iris images acquired by a common
commercial iris sensor (LG 4000). No hardware modifications of any kind are
required. Our approach should be applicable to any iris sensor that can
illuminate the eye from two different directions. Each iris image in the pair
is captured under near-infrared illumination at a different angle relative to
the eye. Photometric stereo is used to estimate surface normal vectors in the
non-occluded portions of the iris region. The variability of the normal vectors
is used as the presentation attack detection score. This score is larger for a
texture that is irregularly opaque and printed on a convex contact lens, and is
smaller for an authentic iris texture. Thus the problem is formulated as binary
classification into (a) an eye wearing textured contact lens and (b) the
texture of an actual iris surface (possibly seen through a clear contact lens).
Experiments were carried out on a database of approx. 2,900 iris image pairs
acquired from approx. 100 subjects. Our method was able to correctly classify
over 95% of samples when tested on contact lens brands unseen in training, and
over 98% of samples when the contact lens brand was seen during training. The
source codes of the method are made available to other researchers. | ['Kevin W. Bowyer', 'Zhaoyuan Fang', 'Adam Czajka'] | 2018-11-18 | null | null | null | null | ['cross-domain-iris-presentation-attack'] | ['computer-vision'] | [ 5.82928002e-01 9.37360227e-02 1.20429543e-03 -1.51210815e-01
-1.30750716e-01 -5.95228910e-01 3.88179392e-01 -2.21782446e-01
-2.89495051e-01 3.03489238e-01 -1.71855576e-02 -3.67332488e-01
-5.45574389e-02 -4.42376912e-01 -5.44281662e-01 -8.39867651e-01
2.69914567e-01 5.76121688e-01 3.24337743e-02 5.05602844e-02
4.72509742e-01 8.53573859e-01 -1.90930843e+00 8.52159709e-02
6.39993668e-01 8.79126787e-01 -1.27755731e-01 9.54517543e-01
3.22123855e-01 2.85062231e-02 -7.34631062e-01 -3.42944562e-01
6.60306752e-01 -6.03422761e-01 -5.10411739e-01 7.53047943e-01
1.15305853e+00 -4.07742649e-01 2.33944163e-01 1.22433710e+00
3.16869169e-01 -3.53752047e-01 5.71029305e-01 -5.02197146e-01
-1.27102196e-01 -3.02648842e-01 -7.53802538e-01 2.07141060e-02
7.34273911e-01 3.98081303e-01 1.99016273e-01 -4.11537826e-01
6.31566942e-01 8.44971240e-01 3.44158053e-01 5.90971351e-01
-1.28160036e+00 -3.75812590e-01 -5.60283959e-01 -1.75856411e-01
-1.33621860e+00 -4.57832724e-01 4.87325370e-01 -7.37431169e-01
4.10571486e-01 7.81515360e-01 9.22285676e-01 6.88375711e-01
3.25674295e-01 2.56190419e-01 1.92837989e+00 -6.25261664e-01
-8.99900198e-02 6.67667449e-01 1.06450962e-02 5.81316650e-01
4.97467995e-01 7.20339954e-01 -5.24685569e-02 -3.69888186e-01
9.52752709e-01 -1.36423230e-01 -6.07231677e-01 -2.37689033e-01
-9.63249087e-01 2.55018592e-01 -1.20422803e-01 4.32263762e-01
-1.93534955e-01 -7.21778989e-01 -1.43112570e-01 4.26088214e-01
2.49214455e-01 6.80929542e-01 -1.39190897e-01 -2.90903360e-01
-5.79903662e-01 -1.28802449e-01 8.83682370e-01 6.20041668e-01
3.66662115e-01 -4.14334416e-01 2.25540191e-01 6.74934328e-01
3.48581523e-01 7.39054143e-01 4.33040500e-01 -4.30948973e-01
9.63125601e-02 8.57314348e-01 3.29326123e-01 -6.50161624e-01
-9.61207300e-02 -1.45794407e-01 -4.30649519e-01 8.00619543e-01
9.59059477e-01 -1.71180561e-01 -9.86505687e-01 7.55338609e-01
5.49209177e-01 3.40554267e-01 -5.58365099e-02 1.20141208e+00
6.45070195e-01 6.42858818e-02 -7.62282014e-01 -2.33392581e-01
1.48727918e+00 -4.07236665e-01 -5.56241274e-01 1.48936450e-01
3.40937853e-01 -1.60395038e+00 8.30773294e-01 7.97139823e-01
-1.10086453e+00 -4.52889293e-01 -1.00902212e+00 2.66919672e-01
-1.85186625e-03 3.93344551e-01 1.49559110e-01 1.27676702e+00
-9.73546982e-01 3.44345063e-01 -7.65445411e-01 -3.85783583e-01
-1.48108423e-01 7.55954385e-01 -4.68469381e-01 1.59367353e-01
-4.37695593e-01 7.69395232e-01 -1.38449237e-01 1.51041612e-01
8.36242214e-02 -1.63611501e-01 -8.77677560e-01 -5.12427270e-01
-5.27760200e-02 -3.78514647e-01 8.48221719e-01 -1.33593369e+00
-1.97108948e+00 1.45814621e+00 -5.56510389e-01 -2.24472508e-02
5.26793122e-01 6.69652447e-02 -7.54445314e-01 1.58473283e-01
-4.85789984e-01 -3.47999096e-01 1.11879981e+00 -1.09902418e+00
-5.29959440e-01 -7.93083847e-01 -1.78679034e-01 1.45714656e-01
7.06321150e-02 3.93752873e-01 -5.50075531e-01 -3.93328071e-01
3.36607695e-01 -1.28253877e+00 2.17715740e-01 -6.32380247e-02
-6.00865841e-01 -5.64689897e-02 5.26359916e-01 -5.68136156e-01
9.43120003e-01 -2.15358543e+00 -2.72582561e-01 6.09782457e-01
1.12023130e-01 6.47109866e-01 1.33200973e-01 9.67632085e-02
-2.76965618e-01 -1.95409209e-01 1.58400118e-01 -2.91225631e-02
-3.74906987e-01 -2.76034717e-02 -1.25016525e-01 1.03467917e+00
-2.51876235e-01 2.01338202e-01 -5.29305577e-01 -3.15688640e-01
3.86257380e-01 5.52203596e-01 -2.85263956e-02 3.06833774e-01
2.65162081e-01 4.91987705e-01 -2.55225331e-01 8.41323733e-01
8.64915252e-01 -1.12788724e-02 3.92602384e-02 -3.89913693e-02
-2.27306888e-01 -5.01757637e-02 -1.51872027e+00 1.09311950e+00
-9.80884805e-02 7.49032617e-01 -1.13118492e-01 -3.02113801e-01
1.09290159e+00 5.91906011e-01 -1.14881424e-02 -5.35326004e-01
1.68729156e-01 4.48168427e-01 1.79293513e-01 -7.63109088e-01
2.64312506e-01 -1.70576930e-01 8.12415183e-01 3.48543614e-01
-4.93391275e-01 1.87925156e-03 -3.11771650e-02 -6.26731396e-01
4.23160672e-01 8.55241995e-03 2.12363034e-01 -2.96124279e-01
7.73177862e-01 -2.95840621e-01 2.60755539e-01 4.85795945e-01
-8.56608078e-02 6.65484726e-01 4.62134361e-01 -8.67542565e-01
-8.57739389e-01 -9.22253013e-01 -1.10100722e+00 3.56597044e-02
3.89305681e-01 4.98186350e-02 -8.45067084e-01 -2.77291954e-01
9.55261663e-02 2.82798018e-02 -5.60535669e-01 2.77374148e-01
-2.94302553e-01 -4.60063010e-01 2.70806663e-02 -2.30474025e-01
2.00270712e-01 -4.84851062e-01 -3.90999079e-01 -2.76196599e-01
1.75900891e-01 -6.41079247e-01 -4.67054069e-01 -8.57989788e-01
-8.94400477e-01 -1.68931568e+00 -7.68128216e-01 -6.88964486e-01
1.16277087e+00 1.83575191e-02 9.38092232e-01 2.99610645e-01
-6.83923125e-01 5.09091377e-01 9.97250825e-02 -5.23182988e-01
-4.99163091e-01 -7.22228646e-01 3.40326160e-01 7.65906930e-01
9.86944199e-01 8.88257623e-02 -7.00871170e-01 8.08963597e-01
-5.67748845e-01 -2.01922596e-01 2.10762799e-01 7.10192680e-01
5.62776268e-01 1.20170042e-01 -6.60890341e-01 -7.94508457e-01
2.30201855e-01 7.92450011e-02 -1.14497304e+00 1.17359534e-01
-5.39361596e-01 -2.40064234e-01 1.99160084e-01 -6.61770582e-01
-7.89844990e-01 -2.02316623e-02 2.23467752e-01 -4.02544230e-01
-6.48469567e-01 -2.07006603e-01 1.55417264e-01 -5.57438195e-01
8.49524498e-01 1.25690326e-01 4.41050321e-01 -6.10562444e-01
-4.62588161e-01 1.16432691e+00 4.62058842e-01 -4.00169045e-01
7.91997671e-01 6.88836873e-01 2.30601400e-01 -1.17655563e+00
-3.04426342e-01 -7.44251966e-01 -4.06340718e-01 -2.89601862e-01
7.07822919e-01 -6.29638314e-01 -1.19838488e+00 1.00048745e+00
-6.78817332e-01 -1.66330785e-02 -8.59165639e-02 1.06781840e+00
-2.89173752e-01 4.46723551e-01 -4.33519542e-01 -8.49229038e-01
-1.07785931e-03 -1.50094366e+00 1.11355472e+00 5.08808970e-01
-2.94570085e-02 -1.11764336e+00 2.28711948e-01 6.70747340e-01
2.61752643e-02 2.89791435e-01 4.68979836e-01 -2.37425044e-01
-4.95246232e-01 -6.47376657e-01 1.20361574e-01 4.41715300e-01
4.26865280e-01 5.02172589e-01 -1.29196739e+00 -6.31696343e-01
4.72934216e-01 3.63154672e-02 3.76825124e-01 6.05227053e-01
9.54912245e-01 -2.05304578e-01 -3.71812552e-01 8.69000673e-01
1.63042760e+00 4.37898487e-01 1.00582969e+00 2.72692561e-01
3.23542953e-01 8.99401546e-01 6.20652318e-01 4.17444333e-02
-3.11180562e-01 8.52464974e-01 1.76415533e-01 -6.64637148e-01
2.73973122e-02 3.25362265e-01 2.00160630e-02 1.31056383e-01
-6.22595191e-01 -1.17193616e-03 -9.72307503e-01 3.34073067e-01
-9.09620345e-01 -7.69345820e-01 -5.57492256e-01 3.05430984e+00
9.51733768e-01 -3.29587758e-02 6.79945499e-02 1.17876492e-01
8.77538502e-01 -5.41823328e-01 -3.07488501e-01 -5.92127621e-01
4.49735904e-03 4.62733209e-01 5.12018383e-01 8.68997037e-01
-9.70886230e-01 3.41750801e-01 6.22353220e+00 3.11830759e-01
-1.52076924e+00 -7.10544646e-01 6.43748939e-01 -2.48611286e-01
8.53994563e-02 3.30729783e-02 -9.09901321e-01 5.28193474e-01
7.02916265e-01 1.54078320e-01 3.08232963e-01 2.45408908e-01
4.90543395e-02 -5.49245059e-01 -1.00184083e+00 1.37627530e+00
1.62820905e-01 -9.25335288e-01 -3.44405115e-01 6.43159509e-01
6.55287564e-01 -1.74719110e-01 4.00235444e-01 -7.49371052e-01
-4.93012458e-01 -1.03692865e+00 -2.25996628e-01 8.58435154e-01
1.32741892e+00 -3.74959946e-01 8.47907484e-01 1.10412599e-03
-6.70058846e-01 4.13165480e-01 -2.15748206e-01 -4.58022058e-02
-5.49012721e-01 3.18446398e-01 -1.04273510e+00 1.62274882e-01
5.07798433e-01 5.01516104e-01 -5.32388926e-01 1.34031141e+00
1.03519261e-01 5.39198816e-01 -4.52058285e-01 1.75238863e-01
-2.24820733e-01 -7.53543019e-01 9.89402413e-01 6.30169511e-01
1.46260500e-01 2.35062968e-02 -3.43138814e-01 6.57655120e-01
4.55742717e-01 2.47266680e-01 -7.51665652e-01 1.72706917e-01
-5.44169098e-02 9.82468963e-01 -3.52877080e-01 -3.05737525e-01
-6.29472911e-01 8.45333159e-01 -6.25682652e-01 4.77153599e-01
-2.81575918e-02 -4.57534999e-01 8.68123949e-01 4.75023031e-01
-2.12340802e-01 4.37984347e-01 -1.58755049e-01 -1.07759976e+00
3.19760084e-01 -1.05622554e+00 -3.49469520e-02 -7.12964177e-01
-9.29201901e-01 7.76121378e-01 -3.35557520e-01 -1.66542566e+00
-2.75330752e-01 -1.02613783e+00 -4.25593644e-01 1.76473582e+00
-1.11531293e+00 -6.37220621e-01 -4.00831908e-01 6.97589695e-01
-4.79003154e-02 -4.97480750e-01 1.03919697e+00 -1.27153486e-01
-4.62642342e-01 5.81646621e-01 2.33411491e-01 9.13892016e-02
8.72361898e-01 -1.52308679e+00 6.24261796e-02 6.88926578e-01
-4.12834436e-03 7.87295818e-01 7.05959499e-01 -4.76715863e-01
-1.38172197e+00 -4.54069167e-01 1.06389666e+00 -6.69855654e-01
2.44090363e-01 -8.69211406e-02 -7.89685071e-01 4.10680681e-01
2.55339622e-01 -1.17430136e-01 9.59764481e-01 1.38869181e-01
-1.80793628e-01 -1.03670336e-01 -1.22507691e+00 5.37957668e-01
2.41716772e-01 -6.28061712e-01 -6.24890089e-01 6.37621522e-01
-2.58796662e-01 -1.01392436e+00 -1.19275343e+00 2.68371969e-01
8.95247400e-01 -1.26444149e+00 6.15794718e-01 -3.93266410e-01
-1.68405753e-03 -4.44511890e-01 4.08304960e-01 -8.16099167e-01
2.30750099e-01 -1.05272496e+00 5.09107649e-01 6.15440607e-01
1.75097749e-01 -1.25744498e+00 7.81416714e-01 6.71799719e-01
2.91918904e-01 -6.91565156e-01 -6.33896232e-01 -6.47701561e-01
-4.57883656e-01 2.77094811e-01 4.36342150e-01 9.54424322e-01
6.86203986e-02 -2.35417962e-01 3.31604667e-02 5.53978860e-01
7.97118664e-01 4.35018569e-01 9.82687116e-01 -1.52779317e+00
-4.37757909e-01 -2.21194223e-01 -1.00907362e+00 -7.80770242e-01
-3.58687758e-01 -3.92301381e-01 -5.37111402e-01 -7.72375643e-01
-5.31902313e-02 -3.20339382e-01 1.28260434e-01 1.73602864e-01
5.56327924e-02 4.91714150e-01 -2.76039124e-01 2.89042056e-01
7.02341139e-01 -4.43321705e-01 1.55809462e+00 1.52217969e-01
-4.66479301e-01 7.47760415e-01 -2.33461246e-01 8.55691373e-01
6.25740409e-01 -5.36618233e-02 -2.90997505e-01 -5.37960511e-03
1.40714303e-01 3.40181559e-01 3.21075171e-01 -9.25815523e-01
1.09858081e-01 -3.34713864e-03 4.51745987e-01 -2.44124919e-01
2.31019467e-01 -1.10426033e+00 4.99236673e-01 5.20585299e-01
1.31937474e-01 -3.87844145e-01 2.98225492e-01 3.19961339e-01
-2.57373631e-01 -2.50407606e-01 9.91963923e-01 1.56242605e-02
-3.49870659e-02 1.00731552e-01 -7.91134611e-02 -3.30031008e-01
9.49303210e-01 -9.25170422e-01 -4.28992599e-01 9.04947221e-02
-7.53793478e-01 -4.65058327e-01 1.28528559e+00 2.75042683e-01
6.12537563e-01 -8.33403409e-01 -6.72177196e-01 1.22771299e+00
3.55531126e-01 -3.85305732e-01 -1.17024601e-01 1.05927396e+00
-9.18148637e-01 3.63017231e-01 -4.01764065e-02 -1.02565479e+00
-1.91121304e+00 2.68276602e-01 9.26240563e-01 4.07251567e-01
-6.89476252e-01 7.45150149e-01 1.05547503e-01 1.13134548e-01
2.46039480e-01 -3.16549122e-01 -2.91951478e-01 -3.80184412e-01
1.02725363e+00 1.39496937e-01 2.11779639e-01 -7.11862445e-01
1.17473163e-01 1.22287560e+00 -1.14969619e-01 1.35080338e-01
6.65553749e-01 -1.72854178e-02 -4.52787876e-01 4.17518705e-01
1.03896034e+00 5.90421617e-01 -8.40562820e-01 -2.80142277e-01
-5.36375523e-01 -1.19754446e+00 1.22940935e-01 -8.69386852e-01
-9.08320904e-01 5.76433539e-01 1.21022379e+00 2.86954731e-01
1.36455345e+00 -3.20466191e-01 2.36671612e-01 -6.72967583e-02
3.39261830e-01 -7.54821420e-01 -6.58148885e-01 -2.23327860e-01
6.55948818e-01 -1.25434804e+00 -9.15454924e-02 -4.56087351e-01
-4.63515669e-01 1.37212276e+00 1.68890670e-01 -4.41212021e-02
6.73038781e-01 9.41620171e-02 5.61945796e-01 -3.43984663e-01
-2.64868736e-01 -1.35674939e-01 9.76349294e-01 5.02295375e-01
5.42298555e-01 1.33474067e-01 -4.02277917e-01 -6.64852500e-01
-1.64293408e-01 1.18885180e-02 7.87364662e-01 6.06605291e-01
-1.19329698e-01 -1.24344206e+00 -8.45409453e-01 6.20197237e-01
-5.93589425e-01 5.41922860e-02 -3.44422728e-01 6.89419389e-01
1.65424302e-01 1.02431715e+00 4.48730290e-01 -1.18383653e-01
4.00716931e-01 -1.43348992e-01 6.68993771e-01 -3.90127629e-01
-5.30468166e-01 5.88977754e-01 -6.37897849e-02 -6.94911361e-01
-5.33663571e-01 -8.11532736e-01 -5.67402184e-01 -3.20316106e-01
-3.08400512e-01 3.22160423e-02 9.23685491e-01 5.72684824e-01
-9.80872288e-02 -3.29955608e-01 8.19421828e-01 -3.35565716e-01
-3.17003459e-01 -9.45192397e-01 -1.31668293e+00 5.20065904e-01
9.52194333e-01 -4.68683392e-01 -8.35060835e-01 2.34644622e-01] | [3.7451512813568115, -3.630770206451416] |
b23941d7-6390-4440-b792-536e24349769 | estimating-treatment-effects-from-irregular | 2302.09446 | null | https://arxiv.org/abs/2302.09446v2 | https://arxiv.org/pdf/2302.09446v2.pdf | Estimating Treatment Effects in Continuous Time with Hidden Confounders | Estimating treatment effects plays a crucial role in causal inference, having many real-world applications like policy analysis and decision making. Nevertheless, estimating treatment effects in the longitudinal setting in the presence of hidden confounders remains an extremely challenging problem. Recently, there is a growing body of work attempting to obtain unbiased ITE estimates from time-dynamic observational data by ignoring the possible existence of hidden confounders. Additionally, many existing works handling hidden confounders are not applicable for continuous-time settings. In this paper, we extend the line of work focusing on deconfounding in the dynamic time setting in the presence of hidden confounders. We leverage recent advancements in neural differential equations to build a latent factor model using a stochastic controlled differential equation and Lipschitz constrained convolutional operation in order to continuously incorporate information about ongoing interventions and irregularly sampled observations. Experiments on both synthetic and real-world datasets highlight the promise of continuous time methods for estimating treatment effects in the presence of hidden confounders. | ['Yan Liu', 'James Enouen', 'Defu Cao'] | 2023-02-19 | null | null | null | null | ['irregular-time-series'] | ['time-series'] | [ 3.36408466e-01 -3.11187152e-02 -8.59597981e-01 -2.28660703e-01
-6.57724023e-01 -1.97065398e-01 3.24624777e-01 8.31556693e-02
-2.43236035e-01 1.22634804e+00 7.01769412e-01 -6.66019678e-01
-3.64102811e-01 -7.36279190e-01 -8.61269414e-01 -6.29448771e-01
-5.44963002e-01 3.35748911e-01 -4.58751023e-01 1.72800660e-01
-1.65792197e-01 2.46177718e-01 -9.40712094e-01 -2.17484698e-01
9.65503931e-01 1.40093058e-01 -4.38583225e-01 3.58656675e-01
3.97076875e-01 6.49944603e-01 -2.11752877e-01 -1.21868104e-01
1.55401140e-01 -3.23263258e-01 -2.99987704e-01 -1.50506243e-01
1.95775345e-01 -4.43352580e-01 -2.02711031e-01 5.50776541e-01
6.95277691e-01 6.88726604e-02 5.74222744e-01 -1.11089277e+00
-7.51596451e-01 7.28764117e-01 -6.22612417e-01 2.45019138e-01
-1.06024727e-01 3.92429322e-01 6.76125348e-01 -3.43342096e-01
3.31161350e-01 1.50480270e+00 8.63230228e-01 4.42711234e-01
-1.42381990e+00 -8.45218420e-01 4.88329947e-01 -2.08851606e-01
-8.96031439e-01 -2.93306023e-01 4.94469970e-01 -7.62115896e-01
5.47382951e-01 5.44952303e-02 3.03507775e-01 1.60641742e+00
5.20623505e-01 5.76007485e-01 1.19098365e+00 -1.34555414e-01
3.69357228e-01 -3.88455868e-01 2.47532651e-01 2.85974145e-01
2.74963588e-01 6.66199565e-01 1.09543577e-02 -5.09880841e-01
1.05297899e+00 5.43282270e-01 -1.82333931e-01 -1.04190163e-01
-1.25680864e+00 1.14740372e+00 8.90459120e-02 -6.47957921e-02
-7.87892163e-01 4.53508824e-01 2.37350881e-01 3.14098507e-01
1.08184671e+00 7.86852464e-02 -5.05260587e-01 1.19392514e-01
-7.75053382e-01 3.82494718e-01 6.53990865e-01 5.49089670e-01
1.75381124e-01 1.41262040e-01 -6.63804710e-01 8.10563415e-02
-1.75724357e-01 6.51301146e-01 -9.17528011e-03 -9.88208413e-01
4.62329119e-01 4.00473773e-01 6.03124857e-01 -6.94929898e-01
-6.12400949e-01 -5.31172276e-01 -1.47034085e+00 1.62052456e-02
7.42878556e-01 -6.78477108e-01 -8.55564833e-01 2.14228463e+00
6.78365827e-01 7.50275552e-01 -2.58662134e-01 7.09405005e-01
1.70563385e-01 2.71005660e-01 3.18736315e-01 -5.53831458e-01
1.03953898e+00 -3.72831881e-01 -1.07515049e+00 2.43038118e-01
6.69957876e-01 -2.86189646e-01 7.24550426e-01 1.02318697e-01
-1.22304833e+00 -2.41162151e-01 -3.42029482e-01 -1.40456930e-01
-3.01492363e-01 -5.34792989e-02 8.16629946e-01 4.70888317e-01
-7.51554787e-01 5.57590246e-01 -9.11120296e-01 -1.37562980e-03
5.84257245e-01 6.16623878e-01 -8.56944695e-02 -2.56125592e-02
-1.48988402e+00 4.47273225e-01 -3.48937750e-01 3.37452412e-01
-1.24521601e+00 -1.50120533e+00 -7.22690403e-01 3.37744832e-01
7.74304926e-01 -1.10591543e+00 1.12673891e+00 -7.77287722e-01
-1.22077024e+00 2.87156999e-01 -2.83722460e-01 -4.96645153e-01
9.62443769e-01 -7.30710924e-02 -2.12266833e-01 -4.48469967e-01
2.44959772e-01 -4.17992398e-02 6.28384888e-01 -4.74619389e-01
-5.07315099e-01 -6.77240670e-01 1.21911004e-01 -1.14440486e-01
-4.26643267e-02 -2.58766860e-02 1.74083322e-01 -8.05892110e-01
-4.97952253e-01 -9.60082054e-01 -4.79923457e-01 1.25714242e-02
-4.71530259e-01 -2.90153176e-01 4.54373509e-01 -7.48416126e-01
1.28318810e+00 -1.93102026e+00 3.22607487e-01 -2.56800562e-01
4.29745615e-01 1.24471188e-02 1.83734596e-01 4.56482768e-01
-3.43904108e-01 2.50860304e-01 -4.34337139e-01 -5.95360637e-01
-8.49874392e-02 5.61082438e-02 -3.70566338e-01 8.11536193e-01
1.76233724e-01 1.06089389e+00 -9.75499451e-01 -1.34549141e-01
2.44572498e-02 6.89989686e-01 -6.94063365e-01 -3.90181653e-02
-2.36511841e-01 1.09953976e+00 -6.17406905e-01 1.37213349e-01
5.17699063e-01 -3.60501766e-01 1.79167330e-01 5.20030737e-01
-3.85487467e-01 6.94711357e-02 -1.19166458e+00 1.25704443e+00
-6.71689808e-01 3.93811733e-01 2.55615592e-01 -1.26433635e+00
-1.55244609e-02 9.45165634e-01 6.62857354e-01 -5.29295802e-01
1.78548738e-01 -8.41291994e-02 -5.82141131e-02 -7.03479290e-01
7.52787990e-03 -6.61291242e-01 -1.98770195e-01 4.24897939e-01
-4.91642207e-01 6.81139886e-01 -1.75612018e-01 -7.18126670e-02
1.06370926e+00 -1.41570270e-01 1.48014560e-01 -2.54464000e-01
8.79188031e-02 -2.81021953e-01 7.06515372e-01 7.93557286e-01
-3.22203524e-03 4.80576545e-01 8.08924615e-01 -4.17707562e-01
-9.01257038e-01 -9.29916143e-01 -3.48871261e-01 7.23224998e-01
-5.19090295e-01 4.56921399e-01 -4.67636555e-01 -4.32046086e-01
4.38306630e-01 5.42210340e-01 -1.33491564e+00 -1.93011492e-01
-5.50826490e-01 -1.18084085e+00 4.00035232e-01 7.07588613e-01
1.60155855e-02 -7.72720397e-01 -3.54591846e-01 4.69277352e-01
-2.16947973e-01 -6.09398186e-01 -7.02674150e-01 -1.40282258e-01
-1.02304542e+00 -1.09151387e+00 -1.09273458e+00 -2.62285918e-01
4.89288181e-01 6.87766820e-03 9.45327282e-01 -2.04627469e-01
-2.48789713e-01 4.10645992e-01 3.88104796e-01 -7.45749593e-01
-3.15296590e-01 -1.88971404e-02 5.70581714e-03 2.76937217e-01
2.46919870e-01 -5.42321503e-01 -9.24047410e-01 -3.63488011e-02
-9.33902264e-01 -7.37662837e-02 2.13313207e-01 9.43035901e-01
3.55276167e-01 -1.80781901e-01 1.07822227e+00 -1.34507263e+00
5.19701660e-01 -1.10152030e+00 -9.22719181e-01 -4.33313586e-02
-6.39648855e-01 1.42065987e-01 4.51654375e-01 -8.92067373e-01
-1.22406828e+00 -3.59985381e-01 3.23425084e-01 -2.75362015e-01
-9.65422988e-02 6.57926500e-01 -1.10754386e-01 7.23161161e-01
2.66416669e-01 -4.91055012e-01 3.43798064e-02 -6.23503327e-01
2.73377150e-01 4.86000150e-01 1.20742686e-01 -5.54520011e-01
2.60561615e-01 1.01953208e+00 3.86559635e-01 -3.93494457e-01
-8.85100484e-01 -5.03325045e-01 -3.26630324e-01 3.50371391e-01
1.01513982e+00 -1.04146326e+00 -1.06584013e+00 3.71657997e-01
-9.50209379e-01 -7.15351343e-01 -5.07109880e-01 8.67861092e-01
-5.45561790e-01 -7.82715008e-02 -5.65543771e-01 -1.04105902e+00
-1.55573934e-01 -1.11758626e+00 1.16137755e+00 -4.03994434e-02
-9.42833945e-02 -1.59042573e+00 4.06356663e-01 8.84721577e-02
3.57124180e-01 6.03398263e-01 1.13651848e+00 8.50953981e-02
-5.26287913e-01 -3.21012139e-01 -4.09521386e-02 -1.73482537e-01
3.42174232e-01 -2.13176504e-01 -7.72158802e-01 -1.75282910e-01
9.12115276e-02 2.40568116e-01 8.32901537e-01 1.32920659e+00
1.13244855e+00 -4.75005776e-01 -5.10158956e-01 6.25633061e-01
1.19201827e+00 9.57666412e-02 4.42530811e-01 -3.27029496e-01
6.03355289e-01 6.96814299e-01 -1.69567987e-02 5.52920222e-01
6.33033216e-01 4.40253168e-01 2.93171912e-01 -6.04066849e-01
2.08963379e-01 -3.36251289e-01 1.25205964e-01 1.16621673e-01
-1.03601590e-01 -3.29073995e-01 -7.07310319e-01 9.74392712e-01
-1.99451816e+00 -1.21788383e+00 -5.73325396e-01 2.61484098e+00
9.54291224e-01 -1.18429236e-01 4.30422783e-01 -1.91890046e-01
8.18655014e-01 -3.88111472e-01 -7.30089366e-01 -2.02098742e-01
-1.18538640e-01 3.16406667e-01 7.87429273e-01 5.48306882e-01
-9.81743634e-01 2.55788058e-01 6.17358923e+00 1.15505166e-01
-1.19076717e+00 3.82251143e-01 8.98559391e-01 -1.73808545e-01
-1.63785979e-01 1.50566518e-01 -6.28229916e-01 5.19043446e-01
1.21097219e+00 -1.63827598e-01 1.33295089e-01 -8.73955041e-02
1.04791427e+00 -5.64170219e-02 -1.16683137e+00 2.11301267e-01
-5.97492397e-01 -1.08451378e+00 -5.74656665e-01 4.89934355e-01
1.19099343e+00 -1.02288000e-01 5.13939321e-01 2.99464643e-01
7.64305174e-01 -1.18509924e+00 1.35684475e-01 7.42858887e-01
9.90796864e-01 -4.82703716e-01 5.77089071e-01 4.16610628e-01
-7.35259056e-01 -2.31561273e-01 -1.19668723e-03 -5.85864425e-01
3.74238729e-01 7.50941455e-01 -6.39514506e-01 4.06113148e-01
2.84915894e-01 8.19615006e-01 5.29333688e-02 9.24248934e-01
-1.49829000e-01 1.23674369e+00 -3.20253640e-01 4.48325127e-01
1.93155482e-01 -9.99662429e-02 3.74694347e-01 7.91206181e-01
2.99882084e-01 1.96646467e-01 9.98732913e-03 9.32143986e-01
-1.21869326e-01 -1.09977342e-01 -7.45438278e-01 1.52135879e-01
-1.69217795e-01 6.34584069e-01 -3.70626777e-01 -3.59302044e-01
-6.63298249e-01 5.03094614e-01 7.57588148e-02 7.32652962e-01
-9.73852575e-01 3.90139937e-01 6.48745000e-01 4.84805077e-01
1.10257484e-01 -1.85538605e-01 -3.55027020e-01 -1.29150474e+00
-1.08608507e-01 -6.83311760e-01 7.70965278e-01 -1.66942760e-01
-1.35164070e+00 -4.51745480e-01 2.58240938e-01 -6.74612403e-01
-9.01931897e-02 -4.65347134e-02 -6.85984850e-01 1.31363344e+00
-1.58558357e+00 -9.18047369e-01 2.37659186e-01 5.19309938e-01
4.80839819e-01 7.09522307e-01 5.66796541e-01 4.30210024e-01
-9.28804100e-01 1.79369614e-01 4.86887813e-01 -1.16318464e-01
7.44456649e-01 -1.21309257e+00 3.43415678e-01 5.73483527e-01
-4.92587388e-01 6.39203906e-01 5.64851880e-01 -1.09868777e+00
-1.36053288e+00 -1.21306980e+00 9.34002101e-01 -6.68506980e-01
6.73751473e-01 -4.30813909e-01 -1.04493284e+00 1.07304370e+00
4.84085232e-02 -2.14743894e-02 5.02756000e-01 4.81388181e-01
-8.28166604e-02 1.25039041e-01 -9.92401063e-01 6.54767692e-01
9.88020241e-01 -4.11216825e-01 -1.70630231e-01 4.47632760e-01
7.67358541e-01 -1.46226078e-01 -8.01070452e-01 4.18481082e-01
5.51786065e-01 -3.96217257e-01 9.62822258e-01 -1.14974439e+00
3.43148202e-01 1.33839101e-01 5.12414753e-01 -1.30316758e+00
-2.44564503e-01 -9.30345953e-01 -1.30888999e-01 1.00197089e+00
2.61037022e-01 -8.78590763e-01 5.91577530e-01 9.34598386e-01
3.52131754e-01 -4.26900178e-01 -1.03043211e+00 -5.43377936e-01
6.53324008e-01 -2.80079603e-01 7.13855147e-01 1.22433913e+00
-2.23397598e-01 2.28255153e-01 -7.78983295e-01 4.13297981e-01
8.55417311e-01 4.91575561e-02 5.01528263e-01 -1.28576756e+00
-2.86324471e-01 -2.02562526e-01 2.03743726e-01 -5.74769676e-01
4.66856778e-01 -5.17351985e-01 -1.33503363e-01 -1.31481266e+00
4.57818747e-01 -5.39512455e-01 -3.59878749e-01 1.67404830e-01
-7.92056441e-01 -2.90172160e-01 -2.61227608e-01 -2.13445872e-01
-3.56284417e-02 6.54025018e-01 1.21739733e+00 -1.36523500e-01
-3.45666885e-01 4.62931931e-01 -5.52089214e-01 4.67065811e-01
7.25399792e-01 -8.12029481e-01 -5.85067451e-01 -5.15875220e-01
1.89830884e-01 6.43208146e-01 6.82043195e-01 -4.02283520e-01
3.01848669e-02 -7.24633574e-01 2.09275391e-02 -2.00508878e-01
5.87448142e-02 -7.53515244e-01 5.04918993e-01 5.47231019e-01
-5.13597190e-01 3.28979611e-01 1.60140976e-01 9.60191905e-01
1.49084523e-01 3.60552460e-01 3.06930006e-01 -1.07425541e-01
1.00654684e-01 5.79253078e-01 -3.42130780e-01 4.08821642e-01
7.69854128e-01 3.08267057e-01 -8.85622054e-02 -6.35115147e-01
-8.39914858e-01 5.48488855e-01 1.40541464e-01 2.06106856e-01
1.67256549e-01 -7.96686172e-01 -1.03055549e+00 1.06114909e-01
-1.91074103e-01 -4.26306129e-02 7.00988948e-01 1.28153694e+00
3.37229788e-01 4.69725281e-01 4.65944648e-01 -2.48315677e-01
-7.93965161e-01 1.12627041e+00 3.31425846e-01 -5.74419916e-01
-6.45712972e-01 2.38505408e-01 9.30170953e-01 -2.15491757e-01
2.82261610e-01 -6.63201690e-01 2.87074912e-02 3.85938631e-03
5.37958622e-01 6.98575675e-01 -2.52825260e-01 -3.01905990e-01
1.78622007e-01 1.44177586e-01 3.53657067e-01 -1.04565978e-01
1.47075737e+00 -2.92599857e-01 1.94866732e-01 7.67224967e-01
1.11668026e+00 -7.02716708e-02 -1.54945803e+00 -3.23857248e-01
-3.66360396e-02 5.43512516e-02 2.24174052e-01 -7.32551038e-01
-9.44982648e-01 1.16440332e+00 6.47596478e-01 1.73776045e-01
9.23456073e-01 -2.81052530e-01 5.03664553e-01 -4.34057236e-01
9.26528275e-02 -6.27483308e-01 -4.74160224e-01 8.67657363e-02
5.80184221e-01 -1.31276882e+00 -1.42381862e-01 -1.41445264e-01
-1.32076725e-01 5.81416965e-01 1.20074593e-01 -1.73408225e-01
9.89520371e-01 3.84972543e-01 -1.70274138e-01 -1.71583712e-01
-9.16610479e-01 -5.82802929e-02 1.39442056e-01 4.08248365e-01
5.82430184e-01 4.99179095e-01 -3.88659269e-01 5.38376033e-01
4.32238549e-01 6.06438220e-01 6.98680580e-01 7.92104304e-01
4.97190088e-01 -8.19543600e-01 -4.24079001e-01 4.55271602e-01
-1.16501534e+00 -2.28833705e-01 2.05474302e-01 8.41514528e-01
8.27545524e-02 8.42929423e-01 1.31145924e-01 7.22351253e-01
3.59485775e-01 1.36948884e-01 8.47304240e-02 -6.14675343e-01
-5.70156872e-01 1.56586275e-01 -2.73353100e-01 -2.59784967e-01
-5.51581144e-01 -8.87834251e-01 -8.45836401e-01 -4.84179020e-01
-4.13193226e-01 -3.51987593e-02 4.81774718e-01 1.04220688e+00
4.19074804e-01 8.44040871e-01 5.71592093e-01 -4.98943835e-01
-8.53445053e-01 -8.41808677e-01 -3.49705040e-01 3.76826137e-01
1.09624457e+00 -7.66513944e-01 -4.84347552e-01 2.87750140e-02] | [8.009182929992676, 5.328891277313232] |
36dd0c9e-7f8e-4e84-afa7-a23628ce61a4 | simple-yet-effective-neural-ranking-and | 2304.01019 | null | https://arxiv.org/abs/2304.01019v1 | https://arxiv.org/pdf/2304.01019v1.pdf | Simple Yet Effective Neural Ranking and Reranking Baselines for Cross-Lingual Information Retrieval | The advent of multilingual language models has generated a resurgence of interest in cross-lingual information retrieval (CLIR), which is the task of searching documents in one language with queries from another. However, the rapid pace of progress has led to a confusing panoply of methods and reproducibility has lagged behind the state of the art. In this context, our work makes two important contributions: First, we provide a conceptual framework for organizing different approaches to cross-lingual retrieval using multi-stage architectures for mono-lingual retrieval as a scaffold. Second, we implement simple yet effective reproducible baselines in the Anserini and Pyserini IR toolkits for test collections from the TREC 2022 NeuCLIR Track, in Persian, Russian, and Chinese. Our efforts are built on a collaboration of the two teams that submitted the most effective runs to the TREC evaluation. These contributions provide a firm foundation for future advances. | ['Xinyu Zhang', 'Jheng-Hong Yang', 'Nandan Thakur', 'Mehdi Rezagholizadeh', 'Odunayo Ogundepo', 'Rodrigo Nogueira', 'Carlos Lassance', 'Ehsan Kamalloo', 'Vitor Jeronymo', 'David Alfonso-Hermelo', 'Jimmy Lin'] | 2023-04-03 | null | null | null | null | ['cross-lingual-information-retrieval'] | ['natural-language-processing'] | [-2.85454601e-01 -7.00995684e-01 -3.85735154e-01 -2.39364937e-01
-1.83596170e+00 -1.09650135e+00 1.21605706e+00 3.32204491e-01
-9.47130680e-01 5.73867738e-01 5.01933515e-01 -3.20091814e-01
-2.44552881e-01 1.78475771e-02 -4.23388571e-01 -2.06450433e-01
-1.02788024e-01 8.21581662e-01 3.02545220e-01 -7.37456083e-01
3.61624241e-01 3.18567932e-01 -1.30165410e+00 7.26669014e-01
4.60669369e-01 5.57170689e-01 4.05246913e-02 5.73295176e-01
-1.55785218e-01 5.89493990e-01 -2.82138199e-01 -5.14713943e-01
1.10194363e-01 -6.25804588e-02 -1.28886104e+00 -7.01036215e-01
6.71385467e-01 1.29713848e-01 -1.52669296e-01 6.75231755e-01
7.37739623e-01 7.00717419e-02 7.25253284e-01 -6.35855377e-01
-8.43643844e-01 5.43246090e-01 -5.39506555e-01 2.35817388e-01
4.40778583e-01 -4.79494482e-01 1.34017003e+00 -9.54274118e-01
8.71805727e-01 1.45826542e+00 3.62039298e-01 4.36076999e-01
-8.89042020e-01 -4.23491597e-01 2.65410505e-02 2.56004691e-01
-1.73813856e+00 -5.51990688e-01 1.48453191e-01 -2.40895137e-01
1.31235480e+00 3.29618543e-01 -2.36021802e-02 9.59068060e-01
8.55325907e-02 1.14906955e+00 1.07360184e+00 -9.54065084e-01
-2.87531346e-01 3.55855614e-01 2.52036542e-01 9.35986936e-02
2.53946073e-02 1.24957934e-02 -3.59306961e-01 -3.38137746e-01
4.09564860e-02 -3.45718861e-01 -8.30543339e-02 -1.51839852e-01
-1.10496938e+00 7.20907152e-01 1.69862270e-01 8.20475161e-01
-1.07724980e-01 -5.81266172e-02 8.62299681e-01 6.95718825e-01
7.05756426e-01 6.69272721e-01 -5.24408162e-01 -1.36563197e-01
-9.94220197e-01 5.36095917e-01 8.23042750e-01 9.30079341e-01
3.66742700e-01 -7.15441346e-01 -4.13141921e-02 1.21971881e+00
6.25407100e-01 7.78157175e-01 6.17432952e-01 -6.77465141e-01
5.06050289e-01 2.47466072e-01 2.79832721e-01 -6.49611533e-01
-7.42813870e-02 -1.21620603e-01 -2.00125545e-01 -2.86066711e-01
1.69281751e-01 1.54820740e-01 -6.72993481e-01 1.47028720e+00
-1.14923120e-01 -7.27205098e-01 2.81815827e-01 6.66370988e-01
5.83327949e-01 7.09816158e-01 2.85605699e-01 -7.05548152e-02
1.42653871e+00 -9.29552317e-01 -6.15928710e-01 -8.52590576e-02
1.19373071e+00 -1.58080161e+00 9.13119614e-01 1.66145280e-01
-1.24239981e+00 -2.72961915e-01 -8.72995734e-01 -4.26683158e-01
-9.82771754e-01 4.25405763e-02 3.84841055e-01 4.54178482e-01
-1.53261995e+00 9.99244899e-02 -6.46422863e-01 -8.26226830e-01
-4.37499106e-01 3.92480381e-03 -3.06708604e-01 -4.23850864e-01
-1.53554893e+00 1.29331517e+00 4.32039559e-01 1.08365780e-02
-4.14445311e-01 -3.39631647e-01 -5.56052506e-01 -3.31916481e-01
1.43206462e-01 -1.86112262e-02 1.42131090e+00 -6.39700174e-01
-9.90170658e-01 1.24108577e+00 -2.43146881e-01 -2.66327292e-01
2.46690676e-01 -4.36900854e-01 -6.72958553e-01 -2.43685186e-01
3.60563338e-01 6.86906099e-01 1.30427644e-01 -9.75861549e-01
-7.19288647e-01 -4.61814404e-01 -7.91399851e-02 6.38816833e-01
-1.18226826e-01 1.01057804e+00 -1.12200677e+00 -4.21478271e-01
-2.74806231e-01 -1.12839627e+00 1.49178356e-01 -5.73733032e-01
2.16531847e-02 -8.23179722e-01 5.95808506e-01 -5.61887681e-01
1.48945355e+00 -2.05618906e+00 7.92955793e-03 3.38032395e-02
-3.04824114e-01 4.07161206e-01 -3.70739430e-01 1.09319258e+00
-3.96782206e-03 4.59214032e-01 3.62884849e-01 -5.36913514e-01
1.68639630e-01 -5.54302074e-02 -7.00759470e-01 1.66430488e-01
-2.18285426e-01 9.65920389e-01 -9.11469460e-01 -6.60271347e-01
-1.55475633e-02 5.56774914e-01 -3.14271487e-02 -1.55972153e-01
-1.62242368e-01 -1.75566748e-02 -6.72106147e-01 5.91995478e-01
2.53463835e-01 -1.08154319e-01 2.37640500e-01 9.09634233e-02
-6.41322374e-01 8.53582621e-01 -5.11289239e-01 1.99704981e+00
-4.53256190e-01 4.73609209e-01 4.34640124e-02 -4.32526290e-01
4.93043959e-01 7.40852654e-01 5.21473408e-01 -1.10137355e+00
7.46716571e-04 6.90352082e-01 -3.86791945e-01 -1.27359346e-01
9.74083185e-01 1.61742106e-01 -4.19748366e-01 8.63832772e-01
-3.82342525e-02 -2.08856076e-01 6.50305927e-01 3.55970532e-01
8.26472819e-01 2.40974650e-01 1.87970251e-01 -5.39832652e-01
7.13430166e-01 2.60804653e-01 -1.80953383e-01 8.39889109e-01
-6.73433691e-02 3.59335244e-01 -2.41204455e-01 -5.35604119e-01
-9.56759095e-01 -8.07898700e-01 -4.61544156e-01 1.54150641e+00
-3.63714784e-01 -7.56964564e-01 -3.67853373e-01 -6.18498445e-01
-1.92993239e-01 3.34259748e-01 -3.51118416e-01 2.40134701e-01
-6.94537461e-01 -6.21533751e-01 9.35869336e-01 2.52438821e-02
1.80867359e-01 -1.13320053e+00 2.62636207e-02 4.86877188e-02
-4.70780045e-01 -1.07303298e+00 -7.60675967e-01 -9.50096175e-04
-5.78330934e-01 -8.38816822e-01 -1.06405294e+00 -9.01476860e-01
-8.56928900e-02 5.73435247e-01 1.50493121e+00 8.76052976e-02
-3.04891616e-01 8.19035769e-01 -5.74393511e-01 -5.41593075e-01
-3.70480031e-01 5.43247998e-01 -3.20108160e-02 -6.49802506e-01
9.23465848e-01 1.62299797e-01 -3.89970601e-01 3.22711229e-01
-1.00297761e+00 -5.01236379e-01 4.06780362e-01 3.64910096e-01
6.07630491e-01 -3.94800365e-01 4.63407457e-01 -7.76992321e-01
1.12974739e+00 -4.52313125e-01 -7.46941030e-01 9.11436558e-01
-8.86824012e-01 3.02689910e-01 1.93494931e-01 1.36521235e-01
-9.97010291e-01 -2.77614415e-01 -7.46893361e-02 7.72814900e-02
9.38481688e-02 9.60968971e-01 4.40978229e-01 -2.36742627e-02
8.20220888e-01 6.98806569e-02 -4.97272998e-01 -8.53597641e-01
6.86785102e-01 9.28966820e-01 1.92010343e-01 -9.39294219e-01
4.08007324e-01 6.62038848e-02 -5.18299639e-01 -8.19808781e-01
-8.13807786e-01 -1.10811305e+00 -6.61794126e-01 3.42762587e-03
8.12219381e-01 -1.19880295e+00 -4.19169188e-01 2.79026598e-01
-1.17012966e+00 -2.85541229e-02 1.25088543e-01 4.23767924e-01
-1.02785908e-01 4.54790145e-01 -7.97725856e-01 -6.20588958e-01
-6.38306975e-01 -1.14323783e+00 1.38917840e+00 -1.22812629e-01
-3.30504477e-01 -1.34868407e+00 8.80691051e-01 4.52726692e-01
6.11896217e-01 -5.49259961e-01 8.48683178e-01 -7.96352983e-01
-3.52653354e-01 -6.75437510e-01 -2.91947454e-01 1.70641795e-01
-2.55837534e-02 1.47016617e-02 -8.73008013e-01 -5.79776406e-01
-3.50755900e-01 -1.05989087e+00 7.99150467e-01 -2.29948778e-02
5.89718878e-01 1.61788493e-01 -3.63433391e-01 2.81192968e-03
1.45821202e+00 2.44828314e-01 4.93787080e-01 5.31110764e-01
2.12214395e-01 7.65860975e-01 4.63078558e-01 -2.06867576e-01
6.81971312e-01 1.07046843e+00 -4.43869770e-01 3.23092192e-02
-1.73604667e-01 -1.34297177e-01 3.10015440e-01 1.27389419e+00
-1.04242839e-01 -4.35210407e-01 -1.12570822e+00 6.26228273e-01
-1.78166342e+00 -8.38241577e-01 2.24525332e-01 2.54178309e+00
1.04670060e+00 -3.46149951e-01 -7.60262385e-02 -5.86077869e-01
3.10288548e-01 2.85469919e-01 -2.69237254e-02 -4.90328282e-01
-3.45616400e-01 3.05896193e-01 4.38800454e-01 8.17986071e-01
-1.11838126e+00 1.31800294e+00 7.34915400e+00 8.76090527e-01
-1.18487537e+00 7.43285492e-02 2.72571683e-01 -2.44205818e-02
-2.59487271e-01 -4.85666916e-02 -1.22827363e+00 5.54293543e-02
1.34366691e+00 -5.91042876e-01 4.52570260e-01 5.05633771e-01
-1.23760894e-01 -4.41361684e-03 -9.90337908e-01 7.64990270e-01
4.67814207e-01 -1.04165661e+00 -3.04221567e-02 -8.38358235e-03
5.69737434e-01 1.13864636e+00 1.69831961e-01 6.83281362e-01
8.39986742e-01 -8.80988836e-01 2.46281981e-01 1.77381009e-01
7.13324010e-01 -4.14050698e-01 6.06724679e-01 7.14979768e-02
-1.07143342e+00 5.14993489e-01 -3.12657058e-01 4.60948348e-01
6.25526682e-02 -2.23548263e-01 -6.87448859e-01 8.01065683e-01
7.65651584e-01 7.47470915e-01 -6.78420126e-01 9.45969582e-01
6.85225651e-02 3.16796869e-01 -1.71704873e-01 -3.21361758e-02
5.57770789e-01 5.52491695e-02 2.86390305e-01 1.57911038e+00
-2.33832039e-02 -2.91878402e-01 3.58009726e-01 1.07798167e-01
-2.06243128e-01 7.58686006e-01 -7.27815866e-01 -2.75527239e-01
4.16911960e-01 1.19725788e+00 -3.12404752e-01 -3.11173856e-01
-6.35033071e-01 7.15181649e-01 5.24520814e-01 4.37287152e-01
-1.56784028e-01 -2.21581608e-01 3.19320351e-01 -2.02044338e-01
-2.64525771e-01 -3.00433099e-01 2.03947142e-01 -1.25476336e+00
1.01012051e-01 -1.18619919e+00 9.71470654e-01 -7.13629246e-01
-1.42952943e+00 9.49523628e-01 3.13469887e-01 -1.02674580e+00
-7.34451950e-01 -4.94113147e-01 5.49270287e-02 1.41170526e+00
-1.88018763e+00 -1.02735567e+00 3.73224705e-01 5.65643787e-01
6.32721007e-01 -1.93195701e-01 1.28630579e+00 7.41638660e-01
-1.27321377e-01 7.20756948e-01 7.03808129e-01 1.20207429e-01
1.56639457e+00 -9.04819369e-01 4.34675545e-01 5.87076306e-01
4.38313186e-01 1.10752571e+00 2.65939653e-01 -5.27910590e-01
-1.41724420e+00 -7.76254714e-01 1.71122706e+00 -8.68642449e-01
1.14337242e+00 -2.33350769e-01 -8.25805843e-01 8.98739636e-01
6.49200141e-01 -5.17341137e-01 5.75249493e-01 6.32303357e-01
-7.34990060e-01 -2.52528638e-02 -5.23411155e-01 6.08743787e-01
2.19061390e-01 -1.04736960e+00 -6.43504202e-01 7.18538880e-01
5.95158815e-01 -2.01869830e-01 -9.84312713e-01 2.76109189e-01
8.11373711e-01 -2.19797745e-01 1.10010755e+00 -6.46047533e-01
5.15720621e-02 -8.58861357e-02 -3.88325721e-01 -1.10226262e+00
-1.09935522e-01 -5.37450254e-01 7.36766875e-01 1.08411694e+00
7.62305439e-01 -5.01425505e-01 1.79299027e-01 6.82865977e-01
-1.64294153e-01 -3.63928974e-01 -7.96557009e-01 -7.18391836e-01
7.81897604e-01 -4.95225966e-01 7.84620792e-02 8.90796244e-01
2.49714255e-01 6.48076773e-01 -2.49830902e-01 -3.89652461e-01
4.63125706e-01 -1.80291995e-01 5.73017240e-01 -1.21119976e+00
-1.54660447e-02 -5.56036532e-01 -3.18976864e-02 -1.11032295e+00
1.35827020e-01 -1.18148696e+00 1.32043302e-01 -1.32160747e+00
3.53081942e-01 -6.89458787e-01 -7.68295944e-01 5.79382300e-01
3.79346684e-02 4.06530708e-01 8.91053900e-02 8.45689774e-01
-1.09578443e+00 1.76775694e-01 7.61574507e-01 -1.92302972e-01
-1.82209704e-02 -1.64639309e-01 -7.15469837e-01 3.33514392e-01
4.64330763e-01 -3.92192394e-01 -4.45589870e-01 -1.03534079e+00
5.49834430e-01 -1.86340548e-02 -2.15243846e-01 -4.77174729e-01
4.05375779e-01 2.46483207e-01 -5.47698848e-02 -7.78517902e-01
2.93642998e-01 -5.25465906e-01 -2.06271201e-01 1.20351892e-02
-7.46545315e-01 6.71506405e-01 4.74822611e-01 1.43543854e-01
-5.93768239e-01 -1.67925045e-01 4.78222430e-01 -2.81902969e-01
-7.14380264e-01 1.86788544e-01 -4.11166072e-01 4.46387410e-01
3.25090617e-01 5.44702113e-01 -5.52102089e-01 -3.20495397e-01
-2.26077721e-01 4.98946190e-01 4.54634517e-01 1.02677202e+00
4.02982086e-02 -1.24352741e+00 -9.57974434e-01 -1.47028893e-01
6.59904718e-01 -7.02974439e-01 -6.63492922e-03 7.50078976e-01
-3.16295594e-01 1.53816700e+00 1.47969753e-01 -4.48102891e-01
-1.46452880e+00 3.84342909e-01 1.40291870e-01 -7.78262019e-01
-1.65266797e-01 5.77193499e-01 2.16914173e-02 -6.61979020e-01
2.65947640e-01 3.25446129e-01 -3.60724688e-01 1.12383880e-01
9.27296340e-01 -4.65999171e-03 5.78149796e-01 -8.01949799e-01
-4.49729443e-01 6.23322070e-01 -8.69184852e-01 -7.12316513e-01
1.00292706e+00 -3.58581513e-01 -4.24823254e-01 7.28043079e-01
1.46658134e+00 3.11098937e-02 -1.10732101e-01 -6.69186175e-01
5.91254115e-01 -4.56702374e-02 2.54301518e-01 -1.25762498e+00
-4.47190315e-01 6.90797627e-01 5.59701085e-01 1.42061874e-01
9.01018620e-01 2.25846916e-01 5.50720453e-01 8.58164072e-01
6.23048663e-01 -1.23052335e+00 -4.56476986e-01 1.06878591e+00
9.64076161e-01 -1.27548504e+00 1.81798972e-02 1.93583623e-01
-4.70247477e-01 9.27215397e-01 2.50625927e-02 1.97651461e-01
7.03360617e-01 -6.89195767e-02 5.67731500e-01 -4.80782986e-01
-1.05310309e+00 -2.38972262e-01 7.84346819e-01 1.09397374e-01
1.44529152e+00 -1.99205220e-01 -8.36335957e-01 -1.74488783e-01
3.31405774e-02 -1.54743820e-01 -9.57566872e-02 1.06802273e+00
-2.97931194e-01 -1.75197411e+00 -1.91618383e-01 -1.59392446e-01
-9.18446362e-01 -6.77748442e-01 -7.95299351e-01 1.09566832e+00
-7.06779897e-01 9.27453756e-01 -2.73026466e-01 -2.51443386e-02
2.25032791e-01 3.83043021e-01 3.58975679e-01 -5.68427324e-01
-6.99431717e-01 4.39619124e-01 2.40110934e-01 -5.38283944e-01
-4.76011425e-01 -8.41155350e-01 -5.52940905e-01 -8.96747038e-02
-9.13860872e-02 7.84858465e-01 1.07107413e+00 9.22128439e-01
4.70360577e-01 -2.16131568e-01 3.99219871e-01 -4.05322999e-01
-4.78223890e-01 -1.17705417e+00 -3.63578767e-01 2.17793494e-01
-3.70756760e-02 -1.37652725e-01 -1.98943660e-01 -1.38943627e-01] | [11.383012771606445, 9.824538230895996] |
38ec8a10-f3e5-4e44-868d-6b336bd5bd19 | feature-representation-matters-end-to-end | null | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1761_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123490222.pdf | Feature Representation Matters: End-to-End Learning for Reference-based Image Super-resolution | In this paper, we are aiming for a general reference-based super-resolution setting: it does not require the low-resolution image and the high-resolution reference image to be well aligned or with a similar texture. Instead, we only intend to transfer the relevant textures from reference images to the output super-resolution image. To this end, we engaged neural texture transfer to swap texture features between the low-resolution image and the high-resolution reference image. We identify the importance of designing a super-resolution task-specific features rather than classification oriented features for neural texture transfer, making the feature extractor more compatible with the image synthesis task. We develop an end-to-end training framework for the reference-based super-resolution task, where the feature encoding network prior to matching and swapping is jointly trained with the image synthesis network. We also discover that learning the high-frequency residual is an effective way for the reference-based super-resolution task. Without bells and whistles, the proposed method E2ENT2 achieved better performance than state-of-the method (i.e., SRNTT with five loss functions) with only two basic loss functions. Extensive experimental results on several datasets demonstrate that the proposed method E2ENT2 can achieve superior performance to existing best models both quantitatively and qualitatively. | ['Ming-Jie Sun', 'Kai-Zhu Huang', 'Chao Yao', 'Yanchun Xie', 'Jimin Xiao'] | null | null | null | null | eccv-2020-8 | ['reference-based-super-resolution'] | ['computer-vision'] | [ 6.57406747e-01 7.35388473e-02 -4.58897604e-03 -3.50607902e-01
-1.26616323e+00 -1.88796893e-02 6.05849564e-01 -8.21505606e-01
-4.12642024e-02 6.64148152e-01 1.51190892e-01 2.66617477e-01
-2.23583087e-01 -9.65809405e-01 -9.44884658e-01 -9.82242882e-01
2.07157120e-01 9.04304758e-02 4.17744666e-01 -4.92419511e-01
2.24664986e-01 2.49013901e-01 -1.59486175e+00 6.28003418e-01
8.08781505e-01 1.27555454e+00 6.79784715e-01 4.15380180e-01
2.69313306e-01 6.84566915e-01 2.27955286e-03 -1.63048133e-01
4.52197492e-01 -5.25533557e-01 -8.49396884e-01 2.28883013e-01
8.48607600e-01 -3.54779333e-01 -4.09252942e-01 1.12442279e+00
4.75082606e-01 2.05491930e-01 4.48649585e-01 -4.16391015e-01
-1.14576018e+00 4.75627035e-01 -9.66256082e-01 4.81783837e-01
6.69932887e-02 -1.81398705e-01 8.59048426e-01 -1.28541553e+00
8.48049700e-01 1.33861828e+00 4.90011662e-01 4.93750125e-01
-1.37569702e+00 -5.49861252e-01 5.37281409e-02 1.58304363e-01
-1.32237887e+00 -6.52436793e-01 9.01962757e-01 -1.74110845e-01
4.47875679e-01 1.24452896e-01 7.95895793e-03 1.19782078e+00
3.47239166e-01 3.19913238e-01 1.48026645e+00 -3.36103469e-01
-1.11613616e-01 8.47838446e-03 -1.60050437e-01 6.42955840e-01
-7.93599337e-02 4.05730516e-01 -7.21484244e-01 3.41163099e-01
1.57994437e+00 -4.07583490e-02 -4.35072184e-01 -2.87404805e-01
-1.26274276e+00 5.34860134e-01 5.67831755e-01 4.83398438e-01
-4.71271306e-01 -5.29234298e-02 4.03623134e-02 4.13555652e-01
8.59605372e-01 2.90360361e-01 -2.49166802e-01 3.53448272e-01
-7.21525311e-01 4.71904455e-03 -2.78856717e-02 8.00439000e-01
8.35196555e-01 1.10426240e-01 -3.57838631e-01 1.14979351e+00
-8.27071518e-02 3.07985485e-01 4.10034597e-01 -1.06340170e+00
5.09705842e-01 -2.45694946e-02 3.85401994e-01 -9.64875638e-01
-9.18252766e-02 -6.39178813e-01 -1.30811000e+00 2.13029861e-01
2.65794426e-01 2.23696932e-01 -8.49507272e-01 1.78155196e+00
2.19228983e-01 3.78332347e-01 1.18340053e-01 1.28545570e+00
6.98155105e-01 5.36715269e-01 -4.31209803e-01 -2.32062876e-01
1.35298645e+00 -1.05569577e+00 -7.65979767e-01 -1.08037874e-01
-1.31790504e-01 -9.01672006e-01 1.24945998e+00 2.78266013e-01
-1.40445018e+00 -1.01093745e+00 -1.13968623e+00 -3.85226279e-01
1.07705079e-01 1.00711040e-01 5.13443530e-01 5.89858089e-03
-1.09462261e+00 8.97699714e-01 -6.01309836e-01 -1.05499633e-01
4.83301312e-01 1.24882780e-01 -4.90231991e-01 -1.34843558e-01
-1.23438132e+00 7.82355368e-01 1.97545245e-01 2.08533943e-01
-5.70243537e-01 -6.83065414e-01 -8.05616558e-01 3.09593580e-03
2.87143260e-01 -8.07708979e-01 9.32044387e-01 -1.06697655e+00
-1.74614453e+00 8.70232642e-01 -2.49450177e-01 -1.12602077e-01
3.37481916e-01 -2.13456124e-01 -4.52012688e-01 2.43788764e-01
2.14839846e-01 3.58240455e-01 1.20740998e+00 -1.50702107e+00
-8.17858756e-01 -2.68304795e-01 -1.82968915e-01 3.06360632e-01
-8.88540149e-02 -5.62017448e-02 -4.41135257e-01 -8.49529982e-01
3.72262448e-01 -5.97931266e-01 5.05449735e-02 -5.04527390e-02
-1.09864846e-01 2.24158809e-01 7.96386838e-01 -7.57494807e-01
8.26929271e-01 -2.18059015e+00 3.79628897e-01 -7.87046626e-02
2.36965284e-01 2.49916967e-02 -3.53720009e-01 -1.07780971e-01
-2.60431886e-01 -1.33833617e-01 -1.25516504e-01 -2.95241624e-01
-3.12447041e-01 -1.32471099e-01 -5.34025788e-01 4.33439076e-01
4.25532341e-01 8.63095045e-01 -8.50499630e-01 -3.70259225e-01
2.72922426e-01 9.89431620e-01 -4.42187518e-01 3.02772313e-01
1.31911829e-01 9.41230059e-01 -5.38486123e-01 3.72584969e-01
8.56109381e-01 -5.56147814e-01 -3.23232785e-02 -7.67450154e-01
-2.39498720e-01 2.28262320e-01 -1.02830946e+00 1.88709104e+00
-8.20841312e-01 4.00608450e-01 1.18286185e-01 -8.01262498e-01
1.12661338e+00 1.66223228e-01 4.18798655e-01 -1.40593565e+00
-1.45296961e-01 2.18319550e-01 -3.04060072e-01 -3.19211543e-01
6.57360077e-01 -3.72655869e-01 1.15043536e-01 2.41411403e-01
3.24579030e-02 1.61294222e-01 -2.28385165e-01 -3.11193407e-01
5.79092085e-01 4.70190197e-01 2.25063588e-04 -2.20916256e-01
5.28619528e-01 -5.55680394e-01 5.15186191e-01 6.42606139e-01
2.55381048e-01 1.09479690e+00 7.79629499e-02 -3.96561593e-01
-1.40308416e+00 -1.18694067e+00 -3.84071141e-01 1.17777824e+00
4.76999968e-01 7.93718696e-02 -6.23106420e-01 -2.22004637e-01
-5.40512800e-01 2.21373543e-01 -7.62135804e-01 -5.55892736e-02
-8.78005028e-01 -6.22378707e-01 -2.10641697e-02 3.95970225e-01
9.74175572e-01 -8.48241091e-01 -3.71204764e-01 1.52407676e-01
-6.01593256e-01 -1.30841744e+00 -7.87817717e-01 3.00901867e-02
-7.90949643e-01 -7.11616039e-01 -9.11433876e-01 -9.11027431e-01
4.80527401e-01 5.69273829e-01 1.05111265e+00 -1.53229371e-01
-4.56449687e-02 -5.82405441e-02 -3.43503863e-01 2.98398167e-01
-1.23683460e-01 9.78960246e-02 -2.53397822e-01 4.80295748e-01
-1.96647018e-01 -7.94243693e-01 -7.79996216e-01 5.24821341e-01
-8.84826005e-01 6.26214445e-01 8.21701527e-01 1.23935223e+00
1.02867174e+00 2.40063637e-01 5.91453910e-01 -8.28544378e-01
2.30384424e-01 -2.09331796e-01 -4.12927896e-01 2.20379829e-01
-4.19124097e-01 2.12605640e-01 7.12380111e-01 -4.67986107e-01
-1.46606493e+00 -1.64756089e-01 -6.24293722e-02 -6.37908518e-01
1.23416662e-01 1.50318637e-01 -2.35990405e-01 -2.75756776e-01
5.94135404e-01 4.66822624e-01 -2.26878509e-01 -6.87069058e-01
2.49399468e-01 4.93336290e-01 8.45500529e-01 -6.16997182e-01
1.03559709e+00 7.80920208e-01 6.94546178e-02 -6.99878335e-01
-1.31417990e+00 -9.22768041e-02 -8.31245959e-01 9.61863399e-02
9.08931196e-01 -1.10076296e+00 -4.29331154e-01 4.53425437e-01
-9.03541207e-01 -4.38064158e-01 -2.21820056e-01 2.89032876e-01
-8.07805598e-01 1.98262990e-01 -7.77837276e-01 -5.33145249e-01
-3.62022996e-01 -1.11523032e+00 1.48895311e+00 2.01013073e-01
3.19232434e-01 -7.91182756e-01 -1.75517201e-01 4.39606488e-01
8.44386339e-01 2.95968622e-01 8.27926219e-01 1.28360599e-01
-8.77494812e-01 4.35183823e-01 -8.19942772e-01 2.77293891e-01
4.71126080e-01 -5.16642749e-01 -1.10381722e+00 -3.60677510e-01
2.60365307e-01 -4.36467171e-01 1.05905092e+00 3.25586706e-01
1.41642010e+00 -3.70704681e-01 8.83599892e-02 1.11225998e+00
1.65956092e+00 -1.45867333e-01 9.77152526e-01 4.12887007e-01
8.68393123e-01 6.79466248e-01 8.15410614e-01 1.27879426e-01
3.75259727e-01 1.29017520e+00 1.57856911e-01 -4.78057057e-01
-5.46776056e-01 -2.00383574e-01 3.13885003e-01 6.63595140e-01
-3.53809565e-01 2.24351272e-01 -2.94266671e-01 3.16590488e-01
-1.81758678e+00 -1.17218721e+00 1.26320109e-01 2.28887415e+00
1.15343583e+00 -5.16308509e-02 -1.84845999e-01 -1.75735667e-01
8.16761971e-01 3.59908640e-01 -5.90113878e-01 -3.16173397e-02
-3.76082182e-01 5.19971550e-01 2.33487234e-01 6.83948517e-01
-1.23379457e+00 1.10876012e+00 5.69918633e+00 1.11593235e+00
-1.38553965e+00 2.83118427e-01 8.88215363e-01 5.73231280e-02
-2.48401627e-01 -2.85881132e-01 -6.76476598e-01 4.09877270e-01
6.84850097e-01 5.46039790e-02 7.56986737e-01 3.75875622e-01
2.32160807e-01 1.27163485e-01 -1.02935600e+00 1.03342342e+00
-3.16828415e-02 -1.34975195e+00 8.18439722e-02 -1.08817719e-01
9.00658369e-01 -1.54865578e-01 5.15523493e-01 2.96327546e-02
8.16749260e-02 -1.26843286e+00 7.01262414e-01 7.90154397e-01
1.49920106e+00 -7.00872004e-01 5.12064874e-01 -2.07234293e-01
-1.48218954e+00 9.96718407e-02 -6.40401065e-01 1.96120143e-01
8.75849575e-02 5.17509401e-01 -5.47441021e-02 9.44013655e-01
8.51619124e-01 8.98673475e-01 -3.94506454e-01 4.26587135e-01
-1.25541247e-03 1.52461857e-01 1.68463722e-01 8.83156598e-01
-1.22138321e-01 -3.02532554e-01 3.53680521e-01 9.98217046e-01
3.90108526e-01 2.14541867e-01 6.54776245e-02 1.14046657e+00
-4.77565750e-02 -2.32744768e-01 -3.71283233e-01 4.70822990e-01
3.55814487e-01 1.41005635e+00 -4.55198318e-01 -1.15781143e-01
-3.76869470e-01 1.27111459e+00 3.65359575e-01 6.52728975e-01
-7.73233652e-01 -2.76214302e-01 5.32083452e-01 2.81013817e-01
6.27477169e-01 9.21045765e-02 -4.01193947e-01 -1.34075606e+00
1.63452446e-01 -8.77453744e-01 -4.91111819e-03 -1.06692898e+00
-1.54440320e+00 9.19649839e-01 -2.32510954e-01 -1.29763663e+00
-1.76877245e-01 -3.36389303e-01 -3.42807084e-01 1.32467067e+00
-1.87324011e+00 -1.55063331e+00 -4.40724194e-01 7.90615916e-01
5.40937126e-01 5.69731742e-02 5.54096222e-01 4.37534004e-01
-4.25781876e-01 6.57416701e-01 1.06976300e-01 5.62084392e-02
8.82624984e-01 -9.99628365e-01 2.98956782e-01 8.47468793e-01
-2.51627386e-01 4.93640095e-01 6.48796141e-01 -4.60597217e-01
-1.41061997e+00 -1.29317749e+00 5.61445832e-01 -2.85812557e-01
5.61392128e-01 -5.15677035e-01 -1.24550247e+00 4.19136077e-01
5.41298985e-02 2.02593774e-01 2.51874924e-02 -8.40053111e-02
-4.76821095e-01 -3.52009386e-01 -1.05613840e+00 6.21456683e-01
1.21995735e+00 -7.92369008e-01 -4.26337153e-01 6.91945851e-02
8.50557506e-01 -6.26832545e-01 -1.18308508e+00 6.18937075e-01
5.37435710e-01 -1.06482732e+00 1.33928990e+00 -2.08867610e-01
8.85481596e-01 -3.32317531e-01 -3.65281761e-01 -1.06362593e+00
-7.74149656e-01 -5.35696745e-01 6.29404709e-02 1.17183185e+00
1.96310848e-01 -6.03082895e-01 3.39441746e-01 2.00445354e-01
-1.21152036e-01 -6.87241018e-01 -8.91278207e-01 -6.68302178e-01
2.14302629e-01 2.06376448e-01 4.33709413e-01 1.05752170e+00
-5.60156941e-01 4.80387628e-01 -7.59907067e-01 4.37526762e-01
9.73416328e-01 5.59333682e-01 4.97473061e-01 -1.01066279e+00
-4.38255012e-01 -4.33426529e-01 8.62862542e-02 -1.03689897e+00
7.28586242e-02 -7.48063982e-01 1.12769939e-01 -1.38059962e+00
5.33833683e-01 -4.75297540e-01 -5.17971575e-01 2.76437253e-01
-2.91242748e-01 5.67056894e-01 -7.03588203e-02 4.01014179e-01
-5.68766296e-01 7.33290732e-01 1.93048477e+00 1.73684359e-01
-8.65827054e-02 -2.70944029e-01 -1.02519786e+00 3.81129026e-01
5.40810823e-01 -7.31680691e-02 -4.15352285e-01 -4.36681807e-01
4.84347567e-02 3.29990983e-01 6.19192064e-01 -6.14989460e-01
1.37611181e-02 -3.02963763e-01 5.48836768e-01 -2.21876338e-01
3.85174125e-01 -6.21853232e-01 1.57594070e-01 -2.22173363e-01
-5.72171986e-01 -3.44216645e-01 -8.22438672e-02 3.50543350e-01
-3.13269854e-01 2.58808792e-01 1.24609625e+00 7.48531744e-02
-6.74412787e-01 5.62574506e-01 2.08439305e-01 5.64088672e-02
4.41266447e-01 -3.10653001e-01 -6.51935697e-01 -2.44534805e-01
-5.83853245e-01 -1.78415373e-01 7.51596868e-01 5.38516819e-01
7.70985305e-01 -1.46478069e+00 -9.69083190e-01 3.01343054e-01
2.77924296e-02 1.55000374e-01 6.23765111e-01 8.85213435e-01
2.29335651e-02 2.10730627e-01 -5.07885933e-01 -5.81932724e-01
-1.00974798e+00 6.55523777e-01 5.27307570e-01 -4.10513163e-01
-1.10405052e+00 6.12151980e-01 9.29511130e-01 -2.15107694e-01
-8.44292715e-03 -6.94398582e-02 -2.30119392e-01 -3.74697596e-01
8.01323175e-01 1.07889906e-01 -2.26862486e-02 -7.43354321e-01
-3.54519184e-03 1.12698483e+00 -3.66431981e-01 -1.81977496e-01
1.60078323e+00 -5.30835211e-01 -1.22376166e-01 4.19008613e-01
1.28513753e+00 -1.41008452e-01 -1.68129206e+00 -7.41731882e-01
-2.28917405e-01 -7.58777022e-01 3.91793698e-01 -6.74461901e-01
-1.42526281e+00 6.68312490e-01 7.57474303e-01 -1.61222935e-01
1.57501507e+00 -1.54062770e-02 8.19191575e-01 3.34940329e-02
7.20484078e-01 -9.05623436e-01 3.47135305e-01 4.75005537e-01
1.12314355e+00 -1.33845675e+00 -8.41032937e-02 -5.72518051e-01
-6.09164655e-01 8.93495739e-01 6.64892018e-01 -3.26070428e-01
4.57482964e-01 1.29467994e-01 -7.56256357e-02 -1.48175225e-01
-9.59905207e-01 -2.62018174e-01 5.47063410e-01 6.51117086e-01
5.56578219e-01 -1.96423993e-01 8.25713295e-03 5.87834835e-01
-1.26757383e-01 1.36044398e-01 4.03809100e-01 3.69725257e-01
-3.73241395e-01 -9.82727528e-01 -3.27133030e-01 2.68160015e-01
-4.68270004e-01 -2.85174727e-01 2.96886880e-02 5.08192122e-01
-2.49465480e-02 6.83828115e-01 1.19292416e-01 -4.42700177e-01
3.09046179e-01 -4.27447319e-01 7.25756943e-01 -4.14655834e-01
-2.37974629e-01 4.37666953e-01 -7.24799335e-02 -8.98951650e-01
-6.45916879e-01 -3.11552107e-01 -9.78468001e-01 -3.23257506e-01
-1.11212991e-01 -1.98045194e-01 2.17977524e-01 7.08763003e-01
4.58455354e-01 7.53857672e-01 1.00789022e+00 -1.27580297e+00
-3.62254500e-01 -8.72502744e-01 -7.67484188e-01 4.28010195e-01
5.77449918e-01 -7.00395942e-01 -1.39261901e-01 1.44878849e-01] | [10.95870590209961, -2.0814220905303955] |
cba39a41-4147-4724-9170-c700adfe2a28 | label-structure-preserving-contrastive | 2209.01314 | null | https://arxiv.org/abs/2209.01314v1 | https://arxiv.org/pdf/2209.01314v1.pdf | Label Structure Preserving Contrastive Embedding for Multi-Label Learning with Missing Labels | Contrastive learning (CL) has shown impressive advances in image representation learning in whichever supervised multi-class classification or unsupervised learning. However, these CL methods fail to be directly adapted to multi-label image classification due to the difficulty in defining the positive and negative instances to contrast a given anchor image in multi-label scenario, let the label missing one alone, implying that borrowing a commonly-used way from contrastive multi-class learning to define them will incur a lot of false negative instances unfavorable for learning. In this paper, with the introduction of a label correction mechanism to identify missing labels, we first elegantly generate positives and negatives for individual semantic labels of an anchor image, then define a unique contrastive loss for multi-label image classification with missing labels (CLML), the loss is able to accurately bring images close to their true positive images and false negative images, far away from their true negative images. Different from existing multi-label CL losses, CLML also preserves low-rank global and local label dependencies in the latent representation space where such dependencies have been shown to be helpful in dealing with missing labels. To the best of our knowledge, this is the first general multi-label CL loss in the missing-label scenario and thus can seamlessly be paired with those losses of any existing multi-label learning methods just via a single hyperparameter. The proposed strategy has been shown to improve the classification performance of the Resnet101 model by margins of 1.2%, 1.6%, and 1.3% respectively on three standard datasets, MSCOCO, VOC, and NUS-WIDE. Code is available at https://github.com/chuangua/ContrastiveLossMLML. | ['Songcan Chen', 'Qirong Mao', 'Lisha Li', 'Zhongchen Ma'] | 2022-09-03 | null | null | null | null | ['multi-label-image-classification', 'multi-label-learning'] | ['computer-vision', 'methodology'] | [ 8.07290792e-01 1.78563073e-01 -3.17575514e-01 -4.69949305e-01
-1.24097478e+00 -5.59690595e-01 5.14372230e-01 2.92838424e-01
-3.39578360e-01 8.47739041e-01 -4.27063346e-01 4.44229729e-02
-3.03818822e-01 -4.10472244e-01 -6.95895612e-01 -1.10372329e+00
2.22735018e-01 5.19773185e-01 6.45126170e-03 1.66793168e-01
-1.76310670e-02 3.11253458e-01 -1.82954443e+00 5.45759499e-01
5.50994754e-01 1.14080203e+00 7.71580115e-02 4.19107407e-01
3.48686166e-02 9.99628246e-01 -4.02746469e-01 -2.95395017e-01
3.15582871e-01 -4.22056288e-01 -8.50301802e-01 2.68932343e-01
1.01302207e+00 8.80273655e-02 2.34372586e-01 1.13405132e+00
4.82126743e-01 -5.39832376e-02 1.10136151e+00 -1.63360405e+00
-6.02161348e-01 2.68547207e-01 -1.07343316e+00 -1.79398879e-01
9.27214473e-02 -3.39587718e-01 1.24988317e+00 -1.00756252e+00
5.25956571e-01 1.27240336e+00 8.30560327e-01 6.92868650e-01
-1.28389084e+00 -8.89813781e-01 7.41584376e-02 9.87405237e-03
-1.32499051e+00 -2.48831987e-01 7.65333414e-01 -5.13526261e-01
4.77579087e-01 4.61564451e-01 1.43003277e-02 1.13197124e+00
1.62485123e-01 7.38882661e-01 1.73904598e+00 -7.30185986e-01
-2.08439231e-01 4.91396397e-01 2.77108371e-01 7.90390491e-01
2.86581041e-03 4.52357531e-02 -2.28577942e-01 -1.14999980e-01
2.44230554e-01 7.38437176e-02 -1.14194646e-01 -6.29796624e-01
-9.51912582e-01 9.58748996e-01 3.53736162e-01 2.65548021e-01
8.76433123e-03 2.57141203e-01 4.45781380e-01 3.56503338e-01
6.39483690e-01 1.05599105e-01 -5.44755876e-01 6.21965170e-01
-8.12834501e-01 -1.38891876e-01 3.53245825e-01 7.39533544e-01
9.60179627e-01 -2.26128995e-01 -1.30737454e-01 1.23803949e+00
3.80144030e-01 4.92912978e-01 5.22651374e-01 -8.77224743e-01
2.69378066e-01 5.62923372e-01 -8.63295048e-02 -1.03993595e+00
-6.56809211e-01 -7.59018779e-01 -1.03843880e+00 4.46221948e-01
3.81381482e-01 1.47128224e-01 -9.37545657e-01 1.89086413e+00
2.57798135e-01 2.72144794e-01 4.10513543e-02 5.05263031e-01
8.22166681e-01 4.32193249e-01 3.09578925e-01 -3.31308812e-01
1.26280951e+00 -1.14651740e+00 -6.23514235e-01 -3.23528528e-01
8.68743718e-01 -8.62302423e-01 1.21833599e+00 4.03281838e-01
-7.37596452e-01 -5.22747815e-01 -1.09393311e+00 6.96467236e-02
-4.36150521e-01 3.41393977e-01 5.86400628e-01 8.03266406e-01
-1.01532686e+00 2.92382538e-01 -1.94917813e-01 -1.40969217e-01
6.87734246e-01 3.73032093e-01 -5.92189491e-01 -3.63217860e-01
-1.03879595e+00 9.26116168e-01 3.08111846e-01 -1.23122536e-01
-8.61260653e-01 -7.91280985e-01 -7.38887250e-01 -1.99550316e-01
3.21489543e-01 -4.63003486e-01 8.96718323e-01 -1.37319314e+00
-8.93548965e-01 1.43857825e+00 2.91627087e-02 -2.36372426e-01
6.87880576e-01 9.46494415e-02 -4.02203679e-01 2.17976958e-01
5.24471879e-01 1.06568670e+00 8.88242185e-01 -1.82435453e+00
-7.36486733e-01 -3.08517426e-01 2.51540802e-02 2.64694095e-01
-2.88306028e-01 -9.05992240e-02 3.60867754e-02 -7.49235868e-01
3.21101874e-01 -1.15204322e+00 -5.51979207e-02 8.49307105e-02
-3.41972709e-01 -1.93385825e-01 8.51027608e-01 -3.12067538e-01
7.04051197e-01 -2.15770292e+00 -1.01354755e-01 1.35983720e-01
1.76006883e-01 4.34941351e-02 -3.24708849e-01 9.81982872e-02
-4.50601250e-01 1.90985218e-01 -3.47196370e-01 -5.54754913e-01
-1.35478973e-01 2.25686029e-01 -1.07203647e-01 7.17775583e-01
1.93834946e-01 7.61113644e-01 -8.65962803e-01 -5.91216624e-01
2.99966156e-01 5.58792830e-01 -2.67958641e-01 -2.46752948e-01
4.38716598e-02 5.25362611e-01 1.20285312e-02 7.37757504e-01
7.45834053e-01 -5.62452674e-01 8.81123021e-02 -2.91993499e-01
3.70042861e-01 -4.37540382e-01 -1.24667764e+00 1.42085814e+00
-6.75462842e-01 2.76345372e-01 -1.65266752e-01 -1.24892592e+00
6.85916841e-01 3.90311092e-01 6.76434278e-01 -7.14058816e-01
-4.15675342e-02 3.54901165e-01 -4.82958674e-01 -1.50788829e-01
1.77762106e-01 -5.31074047e-01 -1.29718661e-01 3.01689148e-01
2.77027398e-01 9.20255706e-02 1.49137661e-01 4.39547338e-02
6.92776561e-01 -1.47082135e-01 3.03789407e-01 -1.56957150e-01
7.08718896e-01 -3.39883476e-01 5.83027840e-01 8.39054823e-01
-3.34021926e-01 7.90497243e-01 3.47412616e-01 -1.75777256e-01
-8.71924937e-01 -1.10669279e+00 -4.88429934e-01 1.26883304e+00
1.65202588e-01 7.25689018e-03 -3.34286958e-01 -1.01577973e+00
-3.57487723e-02 4.49122101e-01 -5.87339401e-01 -2.01477796e-01
-3.08054775e-01 -1.10653365e+00 6.35695994e-01 8.63776281e-02
4.60952193e-01 -9.25940454e-01 -1.42401189e-01 -1.11432917e-01
-2.60880023e-01 -1.06298423e+00 -2.20650911e-01 5.08328974e-01
-6.46159232e-01 -1.26052225e+00 -7.22635806e-01 -1.10194445e+00
9.20873284e-01 2.67341644e-01 1.05750871e+00 1.08176790e-01
-4.02346879e-01 3.50851476e-01 -3.63762558e-01 -1.46067128e-01
-5.44028223e-01 -5.90149537e-02 -6.56916806e-03 3.17828894e-01
-8.51050764e-03 -4.44947064e-01 -4.23525989e-01 5.27291954e-01
-1.11344993e+00 4.74572591e-02 5.72890997e-01 1.18785787e+00
8.08169305e-01 1.91886082e-01 1.05677950e+00 -1.32932115e+00
1.12806171e-01 -5.88100791e-01 -2.85297781e-01 4.90080953e-01
-1.03511906e+00 -1.44219235e-01 4.60260302e-01 -4.92290795e-01
-8.51010382e-01 3.25293630e-01 -1.40829995e-01 -3.60889524e-01
-3.77446741e-01 8.19837302e-02 -4.75096107e-02 -3.99914443e-01
5.33985913e-01 -8.85463059e-02 -1.39773250e-01 -2.64292419e-01
5.18324375e-01 6.86129034e-01 2.23337546e-01 -3.92913818e-01
5.37908673e-01 6.74804330e-01 3.27634871e-01 -3.80522549e-01
-1.36966884e+00 -8.34415972e-01 -7.13969588e-01 -3.67874175e-01
7.33226955e-01 -1.05604804e+00 -4.28293854e-01 6.68059230e-01
-7.40295231e-01 -1.22314461e-01 -3.07806730e-01 1.54553890e-01
-6.91864192e-01 5.30462205e-01 -5.62559903e-01 -5.65995932e-01
-1.43571898e-01 -1.16859007e+00 1.08717084e+00 7.07949698e-02
5.00335358e-02 -1.15720487e+00 -2.33369783e-01 7.58997381e-01
1.09954968e-01 4.22222614e-01 1.05097961e+00 -6.04925990e-01
-1.99036866e-01 -1.77640036e-01 -3.68922919e-01 7.69274056e-01
1.41817287e-01 -4.69871253e-01 -1.31481600e+00 -6.16007745e-01
3.25620966e-03 -8.21058214e-01 1.01416957e+00 3.50953698e-01
1.06156230e+00 -1.82223693e-01 -3.30687881e-01 4.12851363e-01
1.86317956e+00 -7.34534347e-03 4.51873660e-01 1.90287471e-01
7.72788525e-01 6.13120258e-01 6.83598161e-01 1.91841274e-01
2.64758408e-01 6.55480742e-01 6.26887977e-01 -3.23329002e-01
-5.09766102e-01 -2.12076586e-03 1.52978256e-01 6.96405828e-01
4.69316572e-01 -2.93483824e-01 -7.53583252e-01 4.67405170e-01
-1.68943405e+00 -7.83403099e-01 -3.19029599e-01 2.08242607e+00
8.90123129e-01 -7.65350135e-03 -1.24141812e-01 3.82257313e-01
8.99318576e-01 2.37457052e-01 -4.63832885e-01 -8.73661600e-03
-4.84304488e-01 2.20785111e-01 8.29246223e-01 5.76006472e-01
-1.59828889e+00 7.47277319e-01 5.70523310e+00 1.22662878e+00
-9.49233890e-01 6.49625003e-01 9.51806664e-01 -2.41860207e-02
-1.24103174e-01 -1.19101882e-01 -9.56606150e-01 3.17262322e-01
8.39647770e-01 5.20833313e-01 6.18105568e-02 7.10031211e-01
-1.49569720e-01 -1.33063734e-01 -9.58090961e-01 1.21714604e+00
3.85883808e-01 -9.46105063e-01 1.65176749e-01 -1.49101555e-01
1.04256999e+00 -1.81936666e-01 4.37666357e-01 2.45576859e-01
1.50050506e-01 -1.08510506e+00 7.01631188e-01 3.91742408e-01
1.07675064e+00 -7.26373911e-01 6.79840982e-01 2.79962033e-01
-9.45349216e-01 -2.79017180e-01 -2.73223132e-01 1.35757968e-01
1.00548878e-01 7.75121987e-01 -5.47670066e-01 5.51745534e-01
4.88364160e-01 7.76718855e-01 -9.28519070e-01 8.48470688e-01
-1.08489722e-01 4.10027325e-01 -3.56870629e-02 5.24397910e-01
3.21402907e-01 4.07417677e-02 1.91636220e-01 1.05186009e+00
1.83269963e-01 -4.59846944e-01 4.02428597e-01 4.41856891e-01
-1.69261709e-01 2.59865493e-01 -5.91751456e-01 4.81793046e-01
1.80049494e-01 1.44235539e+00 -8.31627369e-01 -1.96761504e-01
-4.85658735e-01 1.14078486e+00 3.55117530e-01 2.15798095e-01
-9.18675065e-01 -6.89997151e-02 1.65921614e-01 1.43498583e-02
-6.34829886e-03 3.70363295e-01 -4.55176234e-01 -9.02088284e-01
6.26540706e-02 -7.46285021e-01 6.86061740e-01 -6.89864337e-01
-1.56044757e+00 3.95124823e-01 -7.57532120e-02 -1.36658168e+00
2.26627174e-03 -6.39752448e-01 8.09629709e-02 6.66122854e-01
-1.85506201e+00 -1.60995579e+00 -2.23681152e-01 4.50806111e-01
3.55971515e-01 -1.88034758e-01 9.95494366e-01 8.28131437e-01
-2.88669437e-01 9.36605334e-01 3.87937576e-01 -1.16968140e-01
1.08355486e+00 -1.26663959e+00 -4.51348811e-01 4.79017824e-01
3.13751340e-01 5.51529974e-02 4.08516437e-01 -2.97289521e-01
-4.88974214e-01 -1.35064399e+00 8.17663133e-01 -4.74391490e-01
4.80806708e-01 -2.41053432e-01 -7.38033652e-01 7.23257124e-01
5.35093546e-02 3.85378242e-01 8.22657228e-01 -1.23822734e-01
-6.80984080e-01 -2.70612568e-01 -1.40771031e+00 2.58753926e-01
8.50810945e-01 -5.55131257e-01 -1.52552128e-01 8.16795766e-01
4.35630947e-01 -1.41341344e-01 -8.88177276e-01 7.65581429e-01
4.64851707e-01 -9.21397090e-01 1.28006351e+00 -2.76521206e-01
3.84712756e-01 -3.57534438e-01 -3.77595901e-01 -1.17634237e+00
-2.50002921e-01 1.90698430e-01 3.65893364e-01 1.41092455e+00
5.11907995e-01 -6.87500417e-01 6.24601066e-01 -2.24435274e-02
-1.38818309e-01 -8.12073290e-01 -1.03875208e+00 -7.27185369e-01
3.25990885e-01 -2.54708886e-01 -2.60711350e-02 1.31107354e+00
-6.32000983e-01 3.36313665e-01 -6.23461962e-01 2.97579944e-01
9.27688658e-01 -8.37037340e-02 2.27556184e-01 -1.45003426e+00
-2.54649460e-01 -3.63085568e-01 -4.16607291e-01 -5.87987721e-01
5.46105862e-01 -1.36434686e+00 -8.56552795e-02 -1.34739280e+00
5.73092103e-01 -1.07744801e+00 -6.96260571e-01 6.40100181e-01
-6.81577846e-02 9.05373096e-01 2.21112937e-01 3.39105159e-01
-7.17003345e-01 2.11570725e-01 1.09553182e+00 -4.65960562e-01
2.96328425e-01 3.54621857e-02 -6.77180350e-01 7.53293872e-01
8.82109642e-01 -8.71345818e-01 -5.48162460e-01 -4.30283509e-02
2.71867961e-01 -1.24767944e-01 5.09638071e-01 -1.02906907e+00
8.78612569e-04 6.42338768e-02 2.63826668e-01 -4.16081876e-01
4.01730359e-01 -8.92625272e-01 3.31804097e-01 3.32780838e-01
-6.39373064e-01 -2.20255464e-01 4.23427206e-03 6.84136569e-01
-2.81495363e-01 -5.49264789e-01 1.25183737e+00 -3.13849956e-01
-8.20655048e-01 5.45043573e-02 -2.01373585e-02 1.44859686e-01
1.27349699e+00 -2.24196121e-01 -3.94659877e-01 -4.94410209e-02
-9.25619721e-01 1.13518044e-01 2.88585901e-01 4.70717311e-01
4.23684001e-01 -1.51641345e+00 -6.45327866e-01 1.32213756e-01
3.75258923e-01 -2.11470410e-01 5.54451942e-01 9.84551609e-01
-2.19536811e-01 3.36368263e-01 -2.90507823e-01 -7.98436642e-01
-1.53162837e+00 4.88197833e-01 4.71600384e-01 -5.54800510e-01
-3.58438343e-01 9.83019710e-01 5.42880416e-01 -7.86456168e-01
2.24642113e-01 3.24865580e-02 -2.68431008e-01 4.40492034e-01
3.25674862e-01 2.48120606e-01 2.32084870e-01 -9.49974537e-01
-3.30790281e-01 9.76092815e-01 -2.35267684e-01 2.23992810e-01
1.05748427e+00 -3.30298066e-01 -2.80209243e-01 7.92741835e-01
1.58490062e+00 -1.50026262e-01 -1.10063183e+00 -1.49856105e-01
1.53614447e-01 -3.26195925e-01 -3.85988988e-02 -1.02934957e+00
-1.17397225e+00 7.07714617e-01 1.11653960e+00 2.10026950e-02
1.15720022e+00 1.85182646e-01 4.19486582e-01 -3.70473005e-02
3.81757826e-01 -1.01764297e+00 3.93605083e-01 1.76898718e-01
7.33000994e-01 -1.68874955e+00 -9.54914615e-02 -6.14235044e-01
-5.55044293e-01 6.41695619e-01 4.28355813e-01 4.41138335e-02
6.20647967e-01 5.88160791e-02 2.86756694e-01 -1.37276456e-01
-4.84580487e-01 -4.09237780e-02 2.30115801e-01 6.45810544e-01
3.69061977e-01 2.00595230e-01 -3.44314277e-01 1.51299343e-01
3.24558437e-01 -2.98929632e-01 5.18904626e-01 6.36820376e-01
-3.57607156e-01 -1.28994989e+00 -3.91407162e-01 5.91322482e-01
-7.34447360e-01 -4.88029420e-02 3.16935219e-02 7.50322819e-01
4.72632647e-01 1.01277280e+00 -1.51882797e-01 -1.68794587e-01
1.85441956e-01 2.92477220e-01 4.69696641e-01 -5.85695088e-01
-4.64508832e-01 -1.46286100e-01 -7.10365996e-02 -2.19398350e-01
-8.58583570e-01 -5.88819385e-01 -1.17994785e+00 2.23061636e-01
-4.87695634e-01 -1.59204155e-01 5.35500705e-01 6.87002182e-01
-1.49479648e-02 5.01702011e-01 6.55756712e-01 -6.05715692e-01
-5.66292286e-01 -8.40614855e-01 -8.05220664e-01 8.23457003e-01
4.43176866e-01 -1.03185344e+00 -7.73838997e-01 -1.08565949e-02] | [9.642746925354004, 4.092779159545898] |
e1f5a943-e723-465c-b416-ab14ccd80d95 | heterogeneity-aware-deep-embedding-for-mobile | 1811.00846 | null | http://arxiv.org/abs/1811.00846v1 | http://arxiv.org/pdf/1811.00846v1.pdf | Heterogeneity Aware Deep Embedding for Mobile Periocular Recognition | Mobile biometric approaches provide the convenience of secure authentication
with an omnipresent technology. However, this brings an additional challenge of
recognizing biometric patterns in unconstrained environment including
variations in mobile camera sensors, illumination conditions, and capture
distance. To address the heterogeneous challenge, this research presents a
novel heterogeneity aware loss function within a deep learning framework. The
effectiveness of the proposed loss function is evaluated for periocular
biometrics using the CSIP, IMP and VISOB mobile periocular databases. The
results show that the proposed algorithm yields state-of-the-art results in a
heterogeneous environment and improves generalizability for cross-database
experiments. | ['Rishabh Garg', 'Nalini Ratha', 'Mayank Vatsa', 'Richa Singh', 'Yashasvi Baweja', 'Soumyadeep Ghosh'] | 2018-11-02 | null | null | null | null | ['mobile-periocular-recognition'] | ['computer-vision'] | [-1.51234269e-01 -7.45599151e-01 -2.90611267e-01 -4.00457919e-01
-6.92528009e-01 -5.30061662e-01 3.36436123e-01 -4.64528948e-01
-4.50574428e-01 4.67710465e-01 -1.85596868e-01 -1.37426332e-01
-2.00404912e-01 -1.79472104e-01 -5.22248924e-01 -8.93775284e-01
1.48504404e-02 -2.39823043e-01 -5.68991780e-01 2.30675980e-01
2.83468574e-01 7.76376665e-01 -1.58704627e+00 -1.05478778e-01
4.82213140e-01 1.38420010e+00 -7.60293305e-01 5.88330507e-01
6.63327813e-01 -1.40920222e-01 -5.81362963e-01 -9.18070972e-01
7.86364138e-01 2.30299786e-01 -1.91157237e-01 -3.17091681e-02
1.05860567e+00 -4.33792412e-01 -6.36936247e-01 9.63748455e-01
1.26037157e+00 -3.78987342e-01 4.30393398e-01 -1.23556304e+00
-7.01112270e-01 2.13544182e-02 -7.64632583e-01 1.45449415e-01
3.48919034e-01 2.27120996e-01 5.75694680e-01 -9.94373500e-01
1.89326271e-01 7.97807217e-01 1.02560985e+00 5.69144845e-01
-7.71323919e-01 -7.95186341e-01 -4.55133587e-01 2.38419607e-01
-1.84586060e+00 -8.14652205e-01 5.50192893e-01 -2.60203760e-02
9.17189896e-01 -1.66241135e-02 5.29563427e-01 1.10982347e+00
4.76604372e-01 6.29385710e-01 1.48045337e+00 -3.68490964e-01
-2.75364161e-01 2.09456116e-01 6.49039075e-02 7.98332214e-01
5.40417314e-01 4.22700793e-01 -8.72162223e-01 -3.36057059e-02
3.91152084e-01 4.00279798e-02 -1.22230507e-01 -2.92511672e-01
-7.40196943e-01 2.15965718e-01 1.66743353e-01 -7.86695406e-02
-3.51934880e-01 7.67320693e-02 2.46348500e-01 2.95402974e-01
1.06260687e-01 4.02510576e-02 -4.86770093e-01 -4.39011276e-01
-7.18395650e-01 -1.95368543e-01 7.59327829e-01 8.39689732e-01
3.60232323e-01 1.50522470e-01 1.47720119e-02 7.12644815e-01
6.24046683e-01 1.06149387e+00 4.54354435e-01 -3.25608343e-01
4.57303613e-01 1.88875407e-01 3.46716270e-02 -1.17596292e+00
-2.24788576e-01 -4.63365167e-01 -8.52194548e-01 8.86786282e-02
4.41287696e-01 -3.93076777e-01 -8.73592198e-01 1.43656027e+00
2.04608917e-01 6.10738993e-01 6.91681206e-02 6.50328159e-01
8.22422504e-01 -9.28548053e-02 -2.82924801e-01 -5.06995097e-02
1.26113451e+00 -6.75723970e-01 -6.78962171e-01 2.85150260e-01
4.48604636e-02 -9.26112354e-01 9.88952398e-01 4.33915496e-01
-8.98988008e-01 -4.73403364e-01 -1.19588828e+00 6.68601245e-02
-3.66240561e-01 5.07111669e-01 8.22770536e-01 1.67154551e+00
-1.09477603e+00 -9.50909406e-03 -6.70612633e-01 -2.77295262e-01
6.97482586e-01 1.12016332e+00 -5.52887797e-01 7.57132545e-02
-8.76573324e-01 6.25007033e-01 -2.31622934e-01 3.50323439e-01
-5.32387972e-01 -4.35119390e-01 -7.40015507e-01 1.21544577e-01
-3.45414616e-02 -4.47946638e-01 6.93109810e-01 -6.51858270e-01
-1.99799669e+00 1.12923586e+00 -2.87054628e-01 -4.11971897e-01
3.83697927e-01 -7.93541521e-02 -6.84320092e-01 -2.59723097e-01
-5.56924522e-01 1.88586906e-01 9.10488546e-01 -7.47757018e-01
-3.41534048e-01 -8.36513698e-01 -1.22041501e-01 2.02425882e-01
-3.94422024e-01 1.10861570e-01 -5.99510968e-01 -3.61782074e-01
-6.10614792e-02 -1.14690411e+00 3.57662439e-01 -2.52895147e-01
-2.43638188e-01 1.56062722e-01 8.89666557e-01 -6.39281869e-01
1.05048478e+00 -2.29107690e+00 -3.58979791e-01 4.00878578e-01
4.25991118e-02 5.44390142e-01 -9.19714570e-02 -4.56490032e-02
2.69085646e-01 -7.48489648e-02 3.88365775e-01 -5.71462035e-01
4.45542037e-02 -2.62538075e-01 2.07081735e-01 8.36829662e-01
-2.69819707e-01 8.79107416e-01 -2.97306836e-01 -2.49179140e-01
2.30305314e-01 8.71366262e-01 -3.68776292e-01 -4.14561108e-02
7.80361354e-01 2.09946498e-01 -7.09259063e-02 1.54337370e+00
1.08178067e+00 -9.73896459e-02 -1.45955041e-01 -3.41972411e-01
3.13350677e-01 -1.51382357e-01 -1.22081709e+00 1.53254187e+00
-4.62890148e-01 8.67588162e-01 6.25557601e-02 -4.79549199e-01
7.32540667e-01 2.56502956e-01 3.57736945e-01 -5.81815064e-01
5.50084412e-01 3.20074707e-01 2.24791601e-01 -5.19549370e-01
4.75819230e-01 9.14116502e-02 1.67998672e-01 7.13729262e-02
2.27239355e-01 4.68690366e-01 -5.69521844e-01 -5.30710161e-01
6.61986828e-01 -1.47740236e-02 3.49222600e-01 -6.02401840e-03
7.29060113e-01 -8.88155639e-01 4.13426846e-01 7.53339887e-01
-7.95363545e-01 3.69727850e-01 -8.60528350e-02 -4.96613324e-01
-5.67113042e-01 -9.90874290e-01 -5.73566258e-01 4.69500154e-01
4.48632330e-01 -1.48039281e-01 -6.69012427e-01 -8.13880086e-01
3.14127296e-01 -4.27452356e-01 -4.76344228e-01 8.09211135e-02
-1.49634257e-01 -9.53309476e-01 1.12282085e+00 3.38633835e-01
9.74081278e-01 -3.25682133e-01 -2.24062875e-01 -4.06162292e-01
1.46720067e-01 -1.34816289e+00 -6.11527622e-01 -5.22749960e-01
-4.13428932e-01 -1.13741970e+00 -5.53046286e-01 -6.53279722e-01
5.25421560e-01 2.26210982e-01 6.42927945e-01 -1.50707066e-01
-3.79514128e-01 7.15136170e-01 2.38711223e-01 -6.88007057e-01
3.23085666e-01 1.97355986e-01 5.48183143e-01 7.36315906e-01
1.03873885e+00 -1.34790838e-01 -9.03060436e-01 3.35099041e-01
-3.41105133e-01 -7.58099616e-01 3.74579966e-01 1.16146898e+00
5.90447247e-01 -7.51291662e-02 3.10963035e-01 -4.47932094e-01
7.16932416e-01 -2.31108010e-01 -8.63995373e-01 3.54860216e-01
-1.08902669e+00 -3.61785620e-01 1.51062399e-01 -6.13844872e-01
-7.72029102e-01 -4.03012848e-03 8.95592477e-03 -2.14656487e-01
2.96115149e-02 4.69466686e-01 -5.67716002e-01 -1.09396195e+00
4.21526700e-01 3.01092714e-01 7.08618313e-02 -2.79159099e-01
-3.94179225e-02 1.28534925e+00 5.64068556e-01 -4.13555145e-01
5.89725614e-01 4.50195432e-01 3.19694698e-01 -7.87592113e-01
4.27226312e-02 -4.12676007e-01 -3.38722467e-01 -2.21544385e-01
1.65363580e-01 -1.09140038e+00 -1.48981583e+00 1.43722439e+00
-6.85804963e-01 3.24308902e-01 2.47074202e-01 6.18112445e-01
-1.76318511e-01 5.48570573e-01 -5.98457694e-01 -9.07523692e-01
-5.07089078e-01 -1.50136435e+00 1.15467978e+00 8.23358834e-01
2.46612936e-01 -8.36986840e-01 -2.26090699e-01 6.00656986e-01
6.26538157e-01 -4.45861444e-02 5.94429113e-02 -4.06857699e-01
-8.32506716e-01 -9.20633197e-01 -3.28417510e-01 4.14934665e-01
6.17395103e-01 -1.62982047e-02 -1.43279743e+00 -6.49427295e-01
-2.60319829e-01 -2.49294072e-01 4.09278750e-01 4.79606867e-01
1.31732702e+00 -2.95552999e-01 -1.43162698e-01 1.36018264e+00
1.58572745e+00 3.03590357e-01 8.89299989e-01 3.14155579e-01
7.11162627e-01 2.24043205e-01 3.02910149e-01 4.75940734e-01
5.73190212e-01 8.69349539e-01 1.47887111e-01 -6.20188899e-02
1.73624992e-01 4.53191660e-02 1.22642741e-01 3.39557022e-01
-2.82140493e-01 -2.50481904e-01 -9.36348319e-01 4.28897023e-01
-1.47067547e+00 -9.11389351e-01 4.80894715e-01 2.45883083e+00
5.86133897e-01 -4.73710626e-01 1.13024376e-01 3.76630910e-02
5.37991107e-01 -3.97451920e-03 -6.91132665e-01 -3.39757830e-01
-5.13975084e-01 4.47409004e-01 1.08909070e+00 4.71342951e-01
-1.10195780e+00 1.02612031e+00 7.08288145e+00 6.83979571e-01
-1.85517132e+00 9.35702994e-02 8.06715846e-01 -2.46962681e-01
3.83347809e-01 -5.03714979e-01 -1.23214996e+00 7.36171007e-01
1.09313631e+00 -1.55136306e-02 4.36566859e-01 6.39894366e-01
-4.85547706e-02 1.14651538e-01 -7.76951432e-01 1.83586919e+00
6.08982027e-01 -1.29280400e+00 -2.19094053e-01 4.25135314e-01
5.58628976e-01 1.17019668e-01 9.21356976e-01 -1.16442330e-01
-5.04225314e-01 -1.22365606e+00 -1.16424732e-01 6.64407611e-01
1.23743916e+00 -6.15978360e-01 1.19289172e+00 -4.14248347e-01
-9.68945622e-01 -2.70589981e-02 -1.56762749e-01 1.62616283e-01
-3.09253663e-01 -2.57782992e-02 -7.91999936e-01 4.81912851e-01
6.30289257e-01 5.99788904e-01 -7.91410029e-01 1.35667157e+00
3.70918214e-01 4.79978651e-01 -3.66531968e-01 -1.08893856e-01
-3.31402987e-01 -2.08124854e-02 3.16779852e-01 1.09641325e+00
5.22086143e-01 -2.17541382e-01 -3.54359239e-01 7.51257464e-02
-4.34913099e-01 2.56686628e-01 -7.91745961e-01 -2.56734565e-02
6.04889870e-01 1.01216066e+00 -4.03980985e-02 2.29320139e-01
-5.79487026e-01 8.82241964e-01 -2.97564566e-01 4.41261858e-01
-6.08645976e-01 -4.41751897e-01 9.19863820e-01 1.24253090e-02
6.91780150e-02 -2.33544394e-01 -6.08855903e-01 -1.55787063e+00
2.02780500e-01 -1.32426715e+00 1.40066639e-01 -3.38225991e-01
-1.24213970e+00 4.08568978e-01 -4.04420763e-01 -1.38657629e+00
1.46033271e-04 -9.89951432e-01 -1.50311276e-01 1.19870305e+00
-1.56456327e+00 -1.66161287e+00 -5.52243054e-01 9.58000898e-01
-1.13945834e-01 -1.24532783e+00 8.42290401e-01 6.25325024e-01
-5.53572059e-01 1.76945686e+00 1.53978869e-01 2.63911217e-01
9.27054405e-01 -9.29651678e-01 1.95262998e-01 1.13536024e+00
8.86841044e-02 1.06701803e+00 1.35172442e-01 -1.91331506e-01
-1.94227505e+00 -4.12286252e-01 7.14732945e-01 -7.13266313e-01
2.06552923e-01 -1.48436412e-01 -2.43396789e-01 4.12088841e-01
2.65085250e-01 3.94028991e-01 1.36569846e+00 2.73266375e-01
-6.08504176e-01 -6.44162238e-01 -1.63720989e+00 5.23830891e-01
7.88242996e-01 -9.29928720e-01 8.52868631e-02 -6.72598556e-02
-1.46203086e-01 -8.06294978e-01 -9.28684413e-01 6.44646764e-01
1.46006119e+00 -7.30804503e-01 9.78147328e-01 -4.68141258e-01
-2.79801548e-01 -2.86886334e-01 -4.59778935e-01 -7.11096525e-01
2.50801474e-01 -1.10944963e+00 -4.45110321e-01 1.14722204e+00
4.06778663e-01 -8.49313140e-01 1.16233039e+00 8.10688317e-01
4.11577851e-01 -8.90417814e-01 -1.08141029e+00 -6.35726810e-01
-2.73527801e-01 -1.86375633e-01 8.62757862e-01 8.71949077e-01
-2.89179027e-01 -9.42588300e-02 -9.46646214e-01 4.83070314e-01
8.57810736e-01 -3.77504945e-01 1.19729710e+00 -1.05164111e+00
-2.57629752e-01 -1.42419875e-01 -1.19234002e+00 -8.62832785e-01
9.13851559e-02 -2.85574377e-01 -4.64300185e-01 -6.05633438e-01
1.77847579e-01 -2.98836499e-01 -8.34999681e-01 2.38302201e-01
-1.22150615e-01 7.96456456e-01 4.86647263e-02 3.77221033e-02
-2.97801703e-01 2.32970148e-01 8.21754694e-01 -4.29790437e-01
-2.03706101e-01 3.53401303e-01 -7.51150966e-01 3.71042579e-01
6.09057009e-01 1.02430984e-01 -4.00525004e-01 -5.22831142e-01
5.43238111e-02 -2.09761605e-01 1.54287249e-01 -1.11949229e+00
6.01062357e-01 1.97370872e-01 4.92096037e-01 -3.60672176e-01
4.45839912e-01 -8.91814232e-01 -8.47659707e-02 1.79854840e-01
2.64737248e-01 3.75240743e-01 5.65870643e-01 4.35412765e-01
-3.21073800e-01 5.38968563e-01 8.63126338e-01 6.35254085e-01
-4.66216981e-01 7.55259633e-01 1.27613366e-01 -3.64159375e-01
6.58090055e-01 -8.90406907e-01 -3.40973020e-01 -3.47034961e-01
-3.83952200e-01 -1.34721875e-01 5.41925728e-01 4.99531209e-01
7.94592083e-01 -1.24112189e+00 -4.37735230e-01 8.34739208e-01
2.29629025e-01 -7.10768878e-01 2.26149365e-01 9.71090138e-01
-5.63894093e-01 4.82274860e-01 -4.97501910e-01 -7.26841986e-01
-2.01823258e+00 -1.32776974e-02 9.64217186e-01 3.05405855e-01
-6.81600198e-02 9.85480368e-01 -4.98042554e-01 -2.64907122e-01
6.62762582e-01 -3.01501136e-02 7.63275428e-03 -1.70876205e-01
5.92884541e-01 1.44827679e-01 3.37930888e-01 -8.12902689e-01
-4.51384455e-01 8.37148011e-01 -1.31326854e-01 4.03606482e-02
8.26884091e-01 -2.86288470e-01 2.68315729e-02 -5.83449192e-02
1.16307724e+00 4.84564751e-01 -1.04818213e+00 -2.19294563e-01
-2.44510144e-01 -1.14076960e+00 -8.10825527e-02 -1.00768554e+00
-1.13582587e+00 7.16903210e-01 1.64216614e+00 -4.06769902e-01
1.29737544e+00 -6.46833837e-01 6.90981805e-01 5.14086306e-01
6.32602632e-01 -8.24777842e-01 -2.47940868e-01 4.18635964e-01
2.76757061e-01 -1.75186086e+00 -1.43521205e-01 -1.39195234e-01
-4.20651764e-01 7.93448687e-01 5.47838807e-01 1.89840227e-01
1.17073500e+00 2.62972027e-01 5.52391768e-01 -2.73235478e-02
-1.09313950e-01 1.48372740e-01 6.98024094e-01 9.86213207e-01
6.37259781e-01 8.04885104e-02 -5.74295163e-01 4.09833431e-01
-1.70724615e-01 8.23150650e-02 4.42393541e-01 6.21501386e-01
5.73233485e-01 -1.39209282e+00 -2.04923987e-01 2.81161547e-01
-1.18518555e+00 -1.89024612e-01 -5.10212362e-01 7.41369188e-01
2.08306670e-01 8.96724641e-01 -3.20678741e-01 -8.05365086e-01
2.03890987e-02 -5.84454797e-02 7.17230439e-01 -1.51040256e-01
-5.79934716e-01 -6.13065474e-02 -1.48232684e-01 -5.31450629e-01
-7.22739160e-01 -5.36892831e-01 -3.63309860e-01 -6.12375498e-01
-4.66595531e-01 -9.40850899e-02 9.88196790e-01 8.51091802e-01
8.23659480e-01 -1.31169483e-01 8.57364655e-01 -5.11629701e-01
-6.18598342e-01 -8.19219410e-01 -7.17471480e-01 2.16849074e-01
6.07353985e-01 -5.55756450e-01 -1.08645864e-01 -1.71861589e-01] | [13.460844039916992, 1.1096090078353882] |
161329ed-08b2-4f85-85d0-9e7247c41565 | self-supervised-class-cognizant-few-shot | 2202.08149 | null | https://arxiv.org/abs/2202.08149v1 | https://arxiv.org/pdf/2202.08149v1.pdf | Self-Supervised Class-Cognizant Few-Shot Classification | Unsupervised learning is argued to be the dark matter of human intelligence. To build in this direction, this paper focuses on unsupervised learning from an abundance of unlabeled data followed by few-shot fine-tuning on a downstream classification task. To this aim, we extend a recent study on adopting contrastive learning for self-supervised pre-training by incorporating class-level cognizance through iterative clustering and re-ranking and by expanding the contrastive optimization loss to account for it. To our knowledge, our experimentation both in standard and cross-domain scenarios demonstrate that we set a new state-of-the-art (SoTA) in (5-way, 1 and 5-shot) settings of standard mini-ImageNet benchmark as well as the (5-way, 5 and 20-shot) settings of cross-domain CDFSL benchmark. Our code and experimentation can be found in our GitHub repository: https://github.com/ojss/c3lr. | ['Hadi Jamali-Rad', 'Ojas Kishore Shirekar'] | 2022-02-15 | null | null | null | null | ['unsupervised-few-shot-image-classification'] | ['computer-vision'] | [ 3.37132871e-01 -4.43652421e-02 -3.47040385e-01 -6.73636615e-01
-6.46178424e-01 -5.12889922e-01 9.05536473e-01 1.48641288e-01
-8.45040560e-01 6.85256422e-01 1.45837948e-01 -1.89044103e-01
-3.69983166e-01 -3.96340042e-01 -4.61973667e-01 -5.67612886e-01
-8.19218606e-02 6.00446641e-01 4.22869980e-01 -2.16745421e-01
3.36296707e-01 2.55558133e-01 -1.67569959e+00 2.19148830e-01
7.12090075e-01 6.51212513e-01 7.29776323e-02 7.03481436e-01
1.35102674e-01 9.36295211e-01 -2.57671803e-01 -4.40672636e-01
3.63335907e-01 -3.85679960e-01 -1.07093942e+00 2.74088591e-01
4.39957052e-01 -8.39578509e-02 -4.99162339e-02 9.85254645e-01
6.68301284e-01 5.25592625e-01 8.61251354e-01 -1.22425795e+00
-7.32630372e-01 7.09204018e-01 -6.15475237e-01 6.75229669e-01
-2.30086908e-01 5.24542332e-01 1.06136143e+00 -9.42424595e-01
6.18288875e-01 9.17882144e-01 4.47141588e-01 6.52891636e-01
-1.24002767e+00 -6.18087769e-01 1.94734961e-01 3.26702386e-01
-1.21140897e+00 -6.75900459e-01 5.65338194e-01 -5.56489050e-01
1.16722310e+00 -5.36445603e-02 1.09427147e-01 1.25876606e+00
-2.26287022e-01 8.85894299e-01 1.34300387e+00 -7.12092280e-01
4.00128216e-01 1.22592926e-01 4.84240144e-01 6.44286931e-01
-5.17471544e-02 3.94478858e-01 -3.88375282e-01 2.19622537e-01
4.28315610e-01 -9.59445536e-02 8.05746168e-02 -4.73797888e-01
-9.50124502e-01 9.76537704e-01 3.24241757e-01 4.35205579e-01
-1.74808115e-01 -5.22616580e-02 5.47983766e-01 5.14012992e-01
6.80681944e-01 5.86525202e-01 -8.05105567e-01 -2.26176396e-01
-1.01984954e+00 -2.28174254e-01 7.77775228e-01 7.93416500e-01
8.17481041e-01 -4.33698818e-02 -3.43924910e-01 1.05236459e+00
1.66285485e-01 -1.36152685e-01 7.50373304e-01 -8.84234786e-01
2.33615488e-01 3.85817051e-01 -2.42986664e-01 -2.03098446e-01
-3.54233116e-01 -6.42025352e-01 -6.85241759e-01 4.46361601e-01
5.35155356e-01 -3.30010921e-01 -1.27131057e+00 1.76991630e+00
1.62671477e-01 5.74576974e-01 1.54840648e-01 8.06069732e-01
8.16153467e-01 2.75262326e-01 3.87006253e-01 -1.73954591e-01
1.17224526e+00 -1.48090804e+00 -3.29866618e-01 -3.97659451e-01
7.91919827e-01 -6.57898247e-01 1.21383345e+00 4.47479099e-01
-6.15318477e-01 -7.92429030e-01 -1.06584561e+00 -1.64273009e-01
-7.42554963e-01 -2.82283071e-02 6.22526526e-01 5.99982083e-01
-9.11660135e-01 6.76595211e-01 -6.42084122e-01 -7.38007545e-01
6.97299361e-01 1.84093431e-01 -2.36524597e-01 -2.56007224e-01
-1.18033206e+00 8.64441454e-01 6.06434345e-01 -4.46825236e-01
-9.86778200e-01 -7.86689997e-01 -5.07413328e-01 -1.54496089e-01
7.37285316e-01 -5.08845448e-01 1.44134879e+00 -1.05394959e+00
-1.49119496e+00 1.24678397e+00 1.71935305e-01 -6.19281530e-01
5.66565812e-01 -4.16236788e-01 -3.99131209e-01 7.45566338e-02
2.32539579e-01 9.29018378e-01 6.74592197e-01 -1.06275368e+00
-6.53055906e-01 -4.23554957e-01 1.70910433e-01 2.75323778e-01
-2.62428761e-01 3.59783974e-03 -2.84190148e-01 -5.78329504e-01
-3.75837803e-01 -7.75235713e-01 -2.64226347e-01 -2.36818254e-01
-1.93307549e-01 -3.21003735e-01 6.38770640e-01 -1.42897546e-01
1.03086090e+00 -2.12238598e+00 -1.13441803e-01 -1.86097875e-01
6.44258931e-02 5.54333627e-01 -2.45709568e-01 2.83363342e-01
-3.88367563e-01 8.59744623e-02 -4.03401047e-01 -4.84830439e-01
9.23768207e-02 6.08572699e-02 4.02104296e-02 4.77619261e-01
4.04519290e-01 9.72570062e-01 -1.13523471e+00 -3.13272864e-01
2.28544757e-01 2.63402730e-01 -4.71159607e-01 -1.40849710e-03
-1.91096872e-01 4.39311177e-01 -1.45053044e-01 4.97657657e-01
4.26013440e-01 -4.82887983e-01 -1.47437751e-01 -1.24912955e-01
-1.44455254e-01 4.88664098e-02 -1.02059889e+00 1.99042034e+00
-2.25945532e-01 5.03872275e-01 -3.65248531e-01 -1.25517666e+00
6.06525362e-01 1.20382421e-01 3.87059897e-01 -7.70013988e-01
3.14882994e-01 1.46560222e-02 2.39713103e-01 -3.60475689e-01
1.63742691e-01 -2.56135106e-01 1.42861456e-02 6.09250128e-01
7.50118315e-01 3.15103233e-02 5.57471275e-01 3.29198450e-01
1.13242507e+00 2.03874782e-01 5.70425272e-01 -4.98373181e-01
4.34768081e-01 1.25537321e-01 2.29287297e-01 1.07774150e+00
-5.29512227e-01 6.38786256e-01 2.15465307e-01 -1.80192694e-01
-9.53147113e-01 -1.11156225e+00 -2.01323971e-01 1.68587220e+00
-1.77047089e-01 -2.16577068e-01 -6.89120412e-01 -9.28743243e-01
-1.41098097e-01 1.07000434e+00 -7.16538131e-01 -3.27872753e-01
-2.47072093e-02 -8.94996583e-01 6.07554317e-01 5.40510178e-01
5.66905856e-01 -1.14626443e+00 -7.13273585e-01 1.12287715e-01
3.58888298e-01 -1.03437662e+00 -4.14749622e-01 8.50477874e-01
-7.67325819e-01 -1.03606784e+00 -5.84282935e-01 -7.91834414e-01
5.27357340e-01 6.61247447e-02 1.38109672e+00 -9.35833529e-02
-5.18875480e-01 5.43468177e-01 -6.82962298e-01 -4.45726663e-01
-1.23908989e-01 2.18904510e-01 1.04373172e-02 -8.74045417e-02
7.51528084e-01 -7.13587999e-01 -5.71447074e-01 3.58679831e-01
-9.13235307e-01 -5.19970953e-02 5.55440366e-01 8.86391580e-01
5.45777321e-01 3.11064180e-02 6.75526261e-01 -1.34092844e+00
5.31788588e-01 -8.35399508e-01 -5.77653170e-01 1.37310609e-01
-8.88891220e-01 8.57663453e-02 5.10110199e-01 -5.05670130e-01
-1.29977536e+00 1.65929243e-01 1.15701966e-02 -4.41145867e-01
-5.96806765e-01 3.77164006e-01 1.93844974e-01 1.61581501e-01
1.05728161e+00 5.21220826e-02 -2.79542863e-01 -5.38117409e-01
8.72507155e-01 6.70599580e-01 4.90703672e-01 -6.36307895e-01
6.97180390e-01 5.52791893e-01 -3.63604009e-01 -5.28048158e-01
-1.15872192e+00 -9.32718277e-01 -1.07660079e+00 -2.15181604e-01
7.97281206e-01 -9.64430332e-01 -1.93659365e-01 4.81234580e-01
-5.52400947e-01 -7.60614634e-01 -6.35759950e-01 3.87721777e-01
-6.40977085e-01 1.54006302e-01 -5.59580028e-01 -6.17097497e-01
-3.77868742e-01 -1.03265131e+00 7.02091932e-01 3.94954503e-01
-1.54406771e-01 -1.04385471e+00 3.22205544e-01 4.06280816e-01
2.93198764e-01 -2.09727243e-01 4.45874453e-01 -1.11309290e+00
-1.17148265e-01 9.21539888e-02 -3.22126746e-01 4.86378074e-01
2.55417190e-02 -2.21209407e-01 -1.32069719e+00 -2.57431149e-01
-4.95639779e-02 -8.78824472e-01 1.26477981e+00 2.56193578e-01
9.35386479e-01 1.89090017e-02 2.44684592e-02 5.70382416e-01
1.62952292e+00 2.92844865e-02 5.59787273e-01 3.75667542e-01
5.35043299e-01 6.31666839e-01 8.07208300e-01 4.51888055e-01
2.05787808e-01 3.12307388e-01 2.59137005e-01 -2.12934092e-02
-3.26351553e-01 1.47617087e-02 1.65991243e-02 5.51265299e-01
-7.11759254e-02 -3.28237087e-01 -1.06793690e+00 7.24800467e-01
-1.84169221e+00 -1.08422899e+00 2.41120886e-02 2.01101303e+00
9.75218773e-01 4.22065169e-01 2.72165954e-01 2.63926247e-03
7.68123686e-01 1.72472656e-01 -7.77736723e-01 -2.58122236e-01
-3.33965593e-03 3.87608707e-01 5.76767445e-01 3.75588804e-01
-1.41201222e+00 1.22923148e+00 5.45997047e+00 1.09620559e+00
-1.02759647e+00 5.49744189e-01 9.46116686e-01 -2.50321120e-01
2.10056245e-01 1.37670850e-02 -7.65399039e-01 4.31259423e-01
1.15671015e+00 -1.42125130e-01 4.51299399e-01 8.60443354e-01
2.96031311e-03 -2.42120117e-01 -1.25878692e+00 8.09906006e-01
8.39963183e-02 -1.18731880e+00 -4.02942717e-01 -1.95580661e-01
1.03767467e+00 8.23833764e-01 1.18355662e-01 6.97380304e-01
7.32443213e-01 -8.79945040e-01 5.31069279e-01 3.07281792e-01
7.39280820e-01 -5.45610309e-01 5.36975145e-01 4.65491354e-01
-8.96061182e-01 -1.06578730e-01 -1.63977981e-01 -1.05218716e-01
-8.26782435e-02 5.15550673e-01 -7.10458457e-01 3.34178835e-01
7.47796953e-01 8.71990204e-01 -8.30364585e-01 1.11282861e+00
-3.01853806e-01 8.62503529e-01 -2.40543008e-01 1.50313616e-01
4.59083736e-01 5.86735755e-02 3.66257310e-01 1.53377652e+00
-7.42815807e-02 2.06573978e-01 2.92627543e-01 6.33067906e-01
-1.94905132e-01 2.49447841e-02 -3.74246180e-01 -3.75296660e-02
3.13661516e-01 1.35949230e+00 -1.08897221e+00 -6.49923265e-01
-4.25857216e-01 9.67920542e-01 6.22703791e-01 3.78467828e-01
-7.23281324e-01 -3.27550501e-01 3.84287089e-01 -2.28415802e-02
5.72916508e-01 7.98009112e-02 -2.57731944e-01 -1.33406913e+00
-4.20673430e-01 -7.60551691e-01 7.79660523e-01 -6.20152593e-01
-1.80935323e+00 5.07857621e-01 1.47455826e-01 -1.19360101e+00
-2.15379328e-01 -5.89585304e-01 -6.44776285e-01 4.31270510e-01
-1.69780231e+00 -8.69691670e-01 -1.42089337e-01 6.96144819e-01
9.89327550e-01 -3.80949974e-01 7.32426107e-01 3.83288532e-01
-5.50076902e-01 5.97314119e-01 2.74227977e-01 1.96371764e-01
9.58883703e-01 -1.31141162e+00 3.56930047e-01 9.22510743e-01
3.39989036e-01 4.03742671e-01 7.18084395e-01 -3.85624141e-01
-8.20875943e-01 -1.19196939e+00 4.84733611e-01 -5.19860566e-01
9.98336256e-01 -2.54709691e-01 -8.49224627e-01 7.13755250e-01
4.21568811e-01 2.95958072e-01 7.82738388e-01 1.76101893e-01
-5.57195365e-01 3.67957391e-02 -1.27357042e+00 4.52204257e-01
1.29609299e+00 -4.57684368e-01 -6.47950888e-01 4.61401641e-01
5.36338866e-01 -5.56631424e-02 -9.21012819e-01 4.68337476e-01
3.16678613e-01 -1.04935741e+00 8.47914517e-01 -7.01553345e-01
6.91202998e-01 -1.23876579e-01 -1.42545179e-01 -1.34337628e+00
-4.52012151e-01 -4.02031362e-01 -5.15936874e-02 1.30820680e+00
4.64541554e-01 -5.17700493e-01 8.09544504e-01 1.81680217e-01
-1.59588084e-01 -6.86053276e-01 -6.74652696e-01 -9.61896777e-01
1.38146251e-01 -5.28282821e-01 4.58129309e-02 1.14800847e+00
-5.75201586e-02 7.92846918e-01 -2.08969906e-01 -9.12932754e-02
9.38788712e-01 -1.34912908e-01 4.90566313e-01 -1.25547194e+00
-5.49503207e-01 -5.16019762e-01 -3.12883735e-01 -7.69093156e-01
7.85940513e-02 -1.03556740e+00 1.51595756e-01 -1.18986773e+00
3.73377234e-01 -2.04773858e-01 -7.79429257e-01 5.44617951e-01
-1.53749570e-01 5.13647139e-01 1.90023482e-01 1.77286237e-01
-1.19017851e+00 3.44407737e-01 8.29896986e-01 -9.55018476e-02
-1.04261443e-01 -1.94370329e-01 -8.07033122e-01 6.64964259e-01
1.06456244e+00 -5.65543175e-01 -7.31834531e-01 -3.48750740e-01
-2.85550803e-01 -4.76361126e-01 2.09829733e-01 -1.03798389e+00
3.73836637e-01 -1.23209253e-01 3.29521775e-01 -4.19466585e-01
3.58360261e-01 -5.16319692e-01 -3.14887404e-01 3.30207855e-01
-6.87966764e-01 -3.36633027e-01 3.09537888e-01 5.44783771e-01
-7.00669885e-02 -5.39519429e-01 1.10629869e+00 -4.96183991e-01
-1.41925704e+00 3.08933437e-01 -1.19669855e-01 5.96037328e-01
1.03355265e+00 -1.28412113e-01 -5.50036252e-01 -9.07953009e-02
-9.92594123e-01 2.69869149e-01 3.40407491e-01 3.80754590e-01
3.10517401e-01 -9.91493165e-01 -8.27115357e-01 8.63483772e-02
4.59944665e-01 -2.31931418e-01 2.00810581e-01 8.73610556e-01
-1.28510669e-01 2.29346663e-01 -3.39677721e-01 -5.38200319e-01
-1.07894075e+00 5.33866405e-01 1.29654795e-01 -4.07284886e-01
-3.86999935e-01 1.24518883e+00 1.63808420e-01 -5.57124555e-01
2.86712527e-01 1.90900981e-01 -3.41541678e-01 2.77117163e-01
5.58792353e-01 2.55927593e-01 7.13432357e-02 -3.96979660e-01
-3.70299399e-01 2.29210243e-01 -4.00582403e-01 -1.74029633e-01
1.54211092e+00 -1.67343885e-01 2.87982762e-01 6.38101578e-01
1.34024012e+00 -3.28829646e-01 -1.37787616e+00 -4.13845986e-01
3.34282964e-01 -2.17689201e-01 1.70878381e-01 -1.18025970e+00
-9.22983348e-01 7.14832008e-01 8.24931026e-01 -9.29982215e-02
1.15264583e+00 3.03455800e-01 1.70807123e-01 5.75462759e-01
1.62766963e-01 -1.39061737e+00 2.49340191e-01 6.81951761e-01
4.84023064e-01 -1.65665150e+00 9.94528532e-02 2.24557240e-03
-8.84551823e-01 7.77602613e-01 8.44741166e-01 -3.81740123e-01
9.22506690e-01 2.04060301e-01 9.81363952e-02 -2.12984741e-01
-1.07469833e+00 -7.18089581e-01 2.40569010e-01 6.35182977e-01
5.84873974e-01 -1.14905804e-01 -2.25933731e-01 3.92033458e-01
4.67469878e-02 3.14837813e-01 3.36119741e-01 1.10209298e+00
-4.78216171e-01 -9.94766057e-01 7.06881657e-02 5.95634699e-01
-4.78389084e-01 -4.55586314e-01 -2.42619470e-01 7.00524092e-01
2.83945769e-01 9.90351796e-01 5.28823771e-02 -3.29573095e-01
1.58202693e-01 2.27809548e-01 4.38861936e-01 -1.05702686e+00
-5.55857062e-01 2.15674266e-02 1.56913072e-01 -2.88053453e-01
-6.69143736e-01 -5.38864732e-01 -1.10131764e+00 1.34654060e-01
-1.72083929e-01 -4.69306484e-02 4.04728681e-01 1.07480645e+00
2.60737300e-01 6.23385370e-01 3.50040764e-01 -7.59523749e-01
-6.56669438e-01 -1.13534248e+00 -6.08140528e-01 6.64491653e-01
-1.25422943e-02 -7.83750355e-01 -5.93508065e-01 2.44391948e-01] | [9.895495414733887, 2.7290849685668945] |
928a0a31-70e6-417a-b1a6-b6ab2b126cfe | assembly101-a-large-scale-multi-view-video | 2203.14712 | null | https://arxiv.org/abs/2203.14712v2 | https://arxiv.org/pdf/2203.14712v2.pdf | Assembly101: A Large-Scale Multi-View Video Dataset for Understanding Procedural Activities | Assembly101 is a new procedural activity dataset featuring 4321 videos of people assembling and disassembling 101 "take-apart" toy vehicles. Participants work without fixed instructions, and the sequences feature rich and natural variations in action ordering, mistakes, and corrections. Assembly101 is the first multi-view action dataset, with simultaneous static (8) and egocentric (4) recordings. Sequences are annotated with more than 100K coarse and 1M fine-grained action segments, and 18M 3D hand poses. We benchmark on three action understanding tasks: recognition, anticipation and temporal segmentation. Additionally, we propose a novel task of detecting mistakes. The unique recording format and rich set of annotations allow us to investigate generalization to new toys, cross-view transfer, long-tailed distributions, and pose vs. appearance. We envision that Assembly101 will serve as a new challenge to investigate various activity understanding problems. | ['Angela Yao', 'Robert Wang', 'Dipika Singhania', 'Kun He', 'Daniel Shelepov', 'Dibyadip Chatterjee', 'Fadime Sener'] | 2022-03-28 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Sener_Assembly101_A_Large-Scale_Multi-View_Video_Dataset_for_Understanding_Procedural_Activities_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Sener_Assembly101_A_Large-Scale_Multi-View_Video_Dataset_for_Understanding_Procedural_Activities_CVPR_2022_paper.pdf | cvpr-2022-1 | ['action-anticipation', 'mistake-detection', '3d-human-action-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.53275001e-01 -1.23514168e-01 -4.03092802e-02 -3.71531844e-01
-6.96397603e-01 -9.53531027e-01 7.03362048e-01 -4.07638997e-01
-1.91162884e-01 6.06895506e-01 6.63551748e-01 2.89329559e-01
4.14425619e-02 6.60678819e-02 -9.93982494e-01 -4.72663730e-01
-2.55702853e-01 7.91715086e-01 2.78632194e-01 -5.28080389e-02
3.74607831e-01 4.08207357e-01 -1.59072030e+00 7.33058512e-01
4.89985377e-01 7.68344402e-01 1.56861365e-01 1.02861345e+00
5.40714383e-01 9.38252687e-01 -6.37706757e-01 -5.92773736e-01
4.99271661e-01 -2.43235663e-01 -1.10255659e+00 4.93301958e-01
7.72138476e-01 -5.89726746e-01 -4.68779385e-01 7.12445080e-01
3.01237911e-01 6.35839999e-01 6.41821623e-01 -1.47066367e+00
-3.42530519e-01 2.39029586e-01 -5.82650900e-01 2.58741558e-01
1.05030036e+00 7.80686140e-01 7.90370047e-01 -7.00509965e-01
8.46130848e-01 1.38208890e+00 7.62170315e-01 6.51910365e-01
-8.91941011e-01 -3.50253105e-01 4.67775732e-01 5.15947282e-01
-9.53272462e-01 -4.38532501e-01 5.11895955e-01 -8.60834956e-01
1.29719865e+00 3.87503326e-01 1.05675375e+00 2.14729643e+00
1.82348624e-01 1.35119355e+00 8.97172511e-01 1.22192219e-01
8.50765854e-02 -4.92370605e-01 -1.77871048e-01 7.18611598e-01
1.90788731e-01 -2.05332667e-01 -7.62469292e-01 5.05096838e-02
9.71412718e-01 1.78222895e-01 -2.60446310e-01 -7.01272070e-01
-1.68741500e+00 3.03795159e-01 -2.00480953e-01 6.18964508e-02
-1.91848919e-01 4.17093903e-01 6.76619411e-01 1.23876609e-01
2.32204363e-01 6.76356614e-01 -7.17736781e-01 -1.13223004e+00
-3.93219322e-01 4.99713033e-01 7.02813864e-01 1.40026557e+00
3.36319178e-01 -1.28064543e-01 -3.15413684e-01 5.95289052e-01
-2.36795470e-01 4.62504417e-01 3.01062465e-01 -1.67034125e+00
1.04422605e+00 4.16367680e-01 5.60773075e-01 -8.31323564e-01
-6.14505649e-01 2.19863094e-02 -4.88233149e-01 -4.51817028e-02
9.01956081e-01 -1.36903552e-02 -8.82266939e-01 1.50728047e+00
2.38804981e-01 1.50669202e-01 -2.94557035e-01 9.71448600e-01
4.02018696e-01 6.66131824e-02 -1.15484066e-01 -3.78366816e-03
1.35784125e+00 -1.54752326e+00 -6.06859803e-01 -2.63205558e-01
8.68705034e-01 -5.00447154e-01 1.42961848e+00 6.79489672e-01
-1.19134951e+00 -6.48222744e-01 -5.60204446e-01 -2.05108419e-01
-1.42090604e-01 4.07409877e-01 7.14325845e-01 2.78973103e-01
-5.06961942e-01 8.11752498e-01 -1.09822631e+00 -5.26495934e-01
5.64071298e-01 5.84882796e-02 -7.92320073e-01 -1.96819097e-01
-5.54242551e-01 7.71812797e-01 8.44736472e-02 1.50991097e-01
-1.15529859e+00 -7.02287972e-01 -9.70008790e-01 -3.80011082e-01
7.88993120e-01 -7.18056679e-01 1.63418698e+00 -1.04816949e+00
-1.58971012e+00 1.03646564e+00 4.45304997e-02 -1.96200445e-01
1.10352123e+00 -7.53549278e-01 -2.41682619e-01 1.75590843e-01
3.43747228e-01 4.49017286e-01 6.02566957e-01 -9.01589036e-01
-5.26212215e-01 -7.08812594e-01 2.84260482e-01 5.06566823e-01
3.51638824e-01 -9.47192386e-02 -4.05090660e-01 -8.62888932e-01
-9.55055580e-02 -1.26669276e+00 5.75495092e-03 -2.54610538e-01
-3.84288877e-01 -1.26085743e-01 5.34079313e-01 -8.28334033e-01
7.04475105e-01 -2.19609022e+00 6.14120305e-01 -2.52803981e-01
5.95029481e-02 -5.92841804e-02 -2.36198768e-01 4.53329563e-01
1.42673752e-03 -1.20226979e-01 9.65666026e-02 -6.03140473e-01
1.14115618e-01 3.54727268e-01 9.79453605e-03 6.03648186e-01
-1.39721766e-01 1.09653759e+00 -1.16367042e+00 -3.13330382e-01
1.34219527e-01 -8.56594741e-02 -7.27384031e-01 3.36615831e-01
-3.63593698e-01 1.09105146e+00 -3.87110621e-01 8.11899781e-01
2.96392918e-01 -2.35432759e-01 8.54134094e-03 -1.37861460e-01
6.48423359e-02 1.44614398e-01 -1.07406437e+00 2.30066824e+00
-1.29243478e-01 7.63006926e-01 -2.12149069e-01 -6.58299863e-01
2.81216204e-01 2.29608580e-01 7.76455343e-01 -3.41461778e-01
1.70982659e-01 -1.62026212e-01 -5.70420362e-02 -1.05194438e+00
6.48540735e-01 5.25496483e-01 -3.89501601e-01 3.75818133e-01
1.53779447e-01 -2.89963037e-01 4.06743735e-01 9.06916335e-02
1.38404417e+00 5.79221010e-01 3.12517285e-01 1.35780454e-01
-8.32810402e-02 1.05137691e-01 4.53293264e-01 8.06866705e-01
-5.87307036e-01 8.62596154e-01 7.42760479e-01 -7.58041978e-01
-1.04669714e+00 -1.20967007e+00 5.24102867e-01 1.15244186e+00
2.59783328e-01 -4.02150154e-01 -8.23963106e-01 -8.41673970e-01
-1.06097519e-01 4.60921258e-01 -9.39505935e-01 5.35725951e-02
-7.92398691e-01 -9.83440205e-02 5.97680211e-01 1.06909871e+00
6.17688656e-01 -1.28623927e+00 -7.82258689e-01 -1.07718363e-01
-7.32134998e-01 -1.38055193e+00 -9.34495807e-01 -2.21741006e-01
-7.89387584e-01 -1.57806456e+00 -6.30542457e-01 -4.42598075e-01
5.51498592e-01 2.87721395e-01 1.28268909e+00 -4.69423026e-01
-3.40287060e-01 1.02138638e+00 -5.13075650e-01 -1.67691752e-01
-6.38094395e-02 -1.13709264e-01 4.66560006e-01 6.68046325e-02
1.31689087e-01 -6.04084611e-01 -6.62619770e-01 5.30070603e-01
-1.32448345e-01 2.59237617e-01 5.94100893e-01 5.27801037e-01
5.56740105e-01 -5.35834849e-01 -1.09492928e-01 -6.29269361e-01
1.84891239e-01 -2.33866632e-01 -2.08584204e-01 3.26254487e-01
3.14759552e-01 -3.68058652e-01 3.61929625e-01 -5.18848538e-01
-1.35693264e+00 3.11000258e-01 2.74195284e-01 -7.84052789e-01
-5.99304318e-01 -4.32498544e-01 -3.62367272e-01 2.69044250e-01
5.81368566e-01 2.78127454e-02 -2.32600242e-01 -6.40125990e-01
3.91817063e-01 2.23947182e-01 8.28310609e-01 -7.03429997e-01
3.27187032e-01 5.72786689e-01 -4.63883996e-01 -7.70695508e-01
-7.82380760e-01 -6.20445192e-01 -1.19861507e+00 -4.49109733e-01
1.27621901e+00 -1.00483680e+00 -1.25373900e+00 9.66076136e-01
-1.26283538e+00 -7.45807528e-01 -4.42067116e-01 6.50607169e-01
-1.40920842e+00 4.11128402e-01 -7.20576346e-01 -7.13421345e-01
2.54896432e-01 -1.20377660e+00 1.44552112e+00 -4.29507941e-02
-7.84578025e-01 -4.16801095e-01 1.14102654e-01 1.03621721e+00
-4.60363507e-01 3.54477286e-01 2.65007675e-01 -5.64592421e-01
-6.89809322e-01 4.54388261e-02 1.36478350e-01 1.04284592e-01
2.59065896e-01 -7.37419501e-02 -6.10118866e-01 -2.28632793e-01
-2.00638145e-01 -6.16371512e-01 5.93507707e-01 3.58437985e-01
1.43276620e+00 -3.47312927e-01 -2.36578554e-01 5.73413312e-01
4.84520763e-01 2.86867946e-01 6.14551246e-01 1.69416964e-02
1.10568953e+00 6.25906050e-01 1.08055484e+00 6.65963590e-01
3.46136153e-01 9.23211098e-01 4.86973494e-01 6.01671994e-01
2.10552156e-01 -4.18365806e-01 5.81224859e-01 6.37389839e-01
-7.44456410e-01 -2.63576269e-01 -9.49640870e-01 6.15146399e-01
-1.89269876e+00 -1.46546555e+00 -1.27964839e-01 1.87384403e+00
4.47586030e-01 4.77338433e-02 5.78188539e-01 -1.74737126e-01
2.81255186e-01 2.85940856e-01 -7.23143637e-01 -8.17148089e-02
2.21173540e-01 -2.75623053e-01 3.73337209e-01 2.05813140e-01
-1.29702222e+00 8.41609061e-01 6.49572754e+00 5.11225224e-01
-4.77931261e-01 3.36481370e-02 3.49938661e-01 -4.78438556e-01
3.35675571e-03 -4.13810641e-01 -4.83512074e-01 5.63230276e-01
3.23803514e-01 5.68716347e-01 4.22975838e-01 9.48372900e-01
2.39195228e-01 -4.97436851e-01 -1.55721390e+00 1.23760998e+00
3.49466920e-01 -1.11019552e+00 -2.33276293e-01 -2.50093728e-01
9.13006604e-01 1.43641919e-01 -6.50386512e-02 2.43687302e-01
4.31090415e-01 -1.01267135e+00 9.37302113e-01 7.18189120e-01
6.79852545e-01 -3.69092733e-01 3.35625112e-01 4.83884424e-01
-1.23619735e+00 -2.59986669e-01 1.65314808e-01 -3.48756731e-01
3.79549801e-01 -2.74495989e-01 -5.49429119e-01 4.30876881e-01
1.19968355e+00 1.32264996e+00 -6.23860121e-01 6.71565413e-01
-1.49759455e-02 1.00595482e-01 -1.59569934e-01 1.13813326e-01
1.04931228e-01 -2.90155828e-01 6.94346189e-01 1.06057525e+00
2.07394093e-01 3.24493438e-01 4.35701668e-01 2.74514019e-01
1.22402325e-01 -3.24599922e-01 -8.03093910e-01 -1.95611358e-01
1.43163562e-01 7.41655648e-01 -8.52773547e-01 -4.41511035e-01
-3.44914168e-01 1.51331592e+00 1.65525123e-01 4.39989060e-01
-1.16544056e+00 1.22298747e-01 1.26306939e+00 3.22211385e-02
3.63152742e-01 -6.97272778e-01 -2.26076007e-01 -1.49267411e+00
3.69992584e-01 -1.21257389e+00 1.80937514e-01 -1.06095147e+00
-1.13586318e+00 2.46045038e-01 3.37951362e-01 -1.51724124e+00
-2.62008518e-01 -5.96564174e-01 -2.36269787e-01 7.36103952e-02
-4.18801457e-01 -1.13680112e+00 -7.45073676e-01 6.06463969e-01
1.24871194e+00 -2.39732787e-01 4.49257433e-01 1.57837301e-01
-4.82216179e-01 4.39168990e-01 -1.12779640e-01 1.59481555e-01
7.45887280e-01 -1.17958045e+00 6.27765834e-01 3.92671973e-01
4.45624113e-01 1.43098921e-01 6.40636742e-01 -6.88878417e-01
-1.22070169e+00 -1.03744078e+00 4.00069445e-01 -1.48716140e+00
4.56433266e-01 -6.38745964e-01 -4.95424718e-01 1.41189301e+00
7.13193975e-03 2.73803156e-02 3.71575415e-01 -5.00930250e-02
-2.90072531e-01 3.01823765e-01 -9.12957668e-01 9.14528847e-01
2.20351672e+00 -4.75491434e-01 -6.20482802e-01 5.95368266e-01
6.26794636e-01 -9.21173096e-01 -7.76233673e-01 3.36718768e-01
9.46004748e-01 -1.23837221e+00 9.90992308e-01 -1.13203359e+00
6.51947916e-01 -7.96674341e-02 -1.85320929e-01 -1.26761734e+00
-2.99340367e-01 -6.84691966e-01 -1.70452312e-01 8.35790038e-01
3.87932584e-02 -9.05345604e-02 8.94907713e-01 3.90854567e-01
-3.93768966e-01 -7.33589053e-01 -7.20746100e-01 -9.56727386e-01
-4.49930906e-01 -6.10663593e-01 4.58053559e-01 7.32214153e-01
-8.83310474e-03 -6.34652227e-02 -7.99474001e-01 -1.52322397e-01
3.28867882e-01 1.14587871e-02 1.30440462e+00 -9.10423219e-01
-5.15332878e-01 -3.02028477e-01 -5.09183586e-01 -1.61056077e+00
1.19317509e-01 -3.31432164e-01 2.07566410e-01 -1.17006612e+00
3.09654117e-01 -2.39429735e-02 3.19263548e-01 2.89841920e-01
9.51782688e-02 1.88995585e-01 2.58146733e-01 2.18084365e-01
-1.13100386e+00 5.87620199e-01 1.64072812e+00 -1.73672050e-01
-7.08665177e-02 2.89716870e-01 1.87228201e-03 1.08814013e+00
6.09075189e-01 -1.20434478e-01 -6.14022791e-01 -6.54921412e-01
1.92758739e-02 1.79319590e-01 6.10972524e-01 -1.09108341e+00
-1.18197873e-01 -5.75670779e-01 3.35171193e-01 -7.23371267e-01
7.64549434e-01 -7.14042366e-01 4.13079828e-01 2.77760476e-01
-3.02328676e-01 3.26173931e-01 1.00978380e-02 9.60031450e-01
6.89791068e-02 2.78587699e-01 3.26734841e-01 -5.29343665e-01
-1.08765674e+00 3.64035368e-01 -4.61991370e-01 6.19880795e-01
1.44771767e+00 -7.10737765e-01 -2.22826809e-01 -4.08239365e-01
-1.16921449e+00 2.90335864e-01 5.55247843e-01 9.37647343e-01
2.92030632e-01 -1.30170476e+00 -4.36835557e-01 1.63701609e-01
2.82346010e-01 7.69750997e-02 8.10023010e-01 1.19479024e+00
-5.31428874e-01 3.03605378e-01 -5.22531509e-01 -8.07896912e-01
-1.41275144e+00 3.79579604e-01 2.96287626e-01 -1.65816337e-01
-8.47889364e-01 1.17506075e+00 4.69803333e-01 -5.02607882e-01
3.54721725e-01 -6.31944120e-01 9.56917405e-02 8.94992948e-02
3.16152722e-01 8.40170979e-01 -2.56188393e-01 -5.86326301e-01
-4.31133449e-01 6.58582449e-01 1.41527519e-01 9.80015844e-02
1.09927726e+00 -1.00158855e-01 3.95856202e-01 9.30083632e-01
9.02174294e-01 -1.26230031e-01 -1.89161301e+00 1.13091335e-01
-2.08702032e-02 -8.50574672e-01 -1.20606542e+00 -6.42675877e-01
-7.89051592e-01 8.98361862e-01 5.32359779e-02 -2.91687250e-01
6.49746597e-01 3.57971638e-01 7.60219455e-01 5.74528337e-01
7.36445606e-01 -1.43911660e+00 9.99379516e-01 8.18933725e-01
1.37875891e+00 -1.40641987e+00 -2.09637076e-01 -3.84313285e-01
-1.16366553e+00 6.42965078e-01 1.17228246e+00 2.34393571e-02
7.70161077e-02 5.98207675e-02 -2.34782547e-01 -2.73696721e-01
-6.75814688e-01 1.67771205e-02 1.64430156e-01 6.90558791e-01
-2.13486962e-02 2.45861202e-01 2.14294925e-01 7.87383854e-01
-3.76963943e-01 -1.61376372e-01 5.38698852e-01 7.86149323e-01
-4.92108092e-02 -5.45067549e-01 -1.39267638e-01 4.68155056e-01
-6.79025576e-02 5.71217358e-01 -5.97262919e-01 9.13311422e-01
2.73217827e-01 5.90472043e-01 3.90505552e-01 -5.57364643e-01
7.41876781e-01 5.77193722e-02 1.00991154e+00 -7.27582335e-01
-4.53468353e-01 -5.14672697e-01 3.65902752e-01 -1.29937100e+00
-4.66709822e-01 -1.29315376e+00 -1.00035524e+00 -4.24019307e-01
1.20719470e-01 -3.04318428e-01 1.28306508e-01 1.14531207e+00
4.51909274e-01 5.76368093e-01 2.05477193e-01 -1.28647625e+00
-5.55592000e-01 -1.14588070e+00 -5.42781413e-01 8.30696583e-01
3.39080662e-01 -1.07382250e+00 -1.26290992e-01 3.47949713e-01] | [7.965557098388672, 0.28837329149246216] |
c4eadbda-b26f-40e8-a0b8-ff866c0d8a5f | event-aided-direct-sparse-odometry | 2204.07640 | null | https://arxiv.org/abs/2204.07640v2 | https://arxiv.org/pdf/2204.07640v2.pdf | Event-aided Direct Sparse Odometry | We introduce EDS, a direct monocular visual odometry using events and frames. Our algorithm leverages the event generation model to track the camera motion in the blind time between frames. The method formulates a direct probabilistic approach of observed brightness increments. Per-pixel brightness increments are predicted using a sparse number of selected 3D points and are compared to the events via the brightness increment error to estimate camera motion. The method recovers a semi-dense 3D map using photometric bundle adjustment. EDS is the first method to perform 6-DOF VO using events and frames with a direct approach. By design, it overcomes the problem of changing appearance in indirect methods. We also show that, for a target error performance, EDS can work at lower frame rates than state-of-the-art frame-based VO solutions. This opens the door to low-power motion-tracking applications where frames are sparingly triggered "on demand" and our method tracks the motion in between. We release code and datasets to the public. | ['Davide Scaramuzza', 'Guillermo Gallego', 'Javier Hidalgo-Carrió'] | 2022-04-15 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Hidalgo-Carrio_Event-Aided_Direct_Sparse_Odometry_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Hidalgo-Carrio_Event-Aided_Direct_Sparse_Odometry_CVPR_2022_paper.pdf | cvpr-2022-1 | ['monocular-visual-odometry'] | ['robots'] | [-9.79309529e-03 -1.94165096e-01 -5.46516776e-02 -1.20781869e-01
-8.89981985e-01 -6.87939763e-01 6.86462998e-01 -4.57371831e-01
-4.99038547e-01 5.92317164e-01 4.17120934e-01 -2.99684778e-02
2.95926064e-01 -3.97358298e-01 -9.04926360e-01 -6.55919611e-01
1.80244416e-01 2.90460825e-01 4.46745813e-01 1.24717966e-01
1.08269833e-01 5.18670559e-01 -1.76033950e+00 -1.13324247e-01
9.14806128e-02 7.70304680e-01 3.74227643e-01 1.04774654e+00
8.99652168e-02 1.02868962e+00 -3.31421345e-01 -1.47600770e-01
5.30784309e-01 -3.25067312e-01 -4.07647520e-01 4.99346763e-01
8.59379172e-01 -8.03502738e-01 -8.52556407e-01 9.25237954e-01
6.38288319e-01 -8.52512643e-02 1.40345931e-01 -1.10479188e+00
-1.39605654e-02 -3.25553328e-01 -4.50155169e-01 1.63983136e-01
1.07947552e+00 4.77591753e-01 7.03135729e-01 -1.08074272e+00
1.17508042e+00 1.08324075e+00 8.78218353e-01 5.35692513e-01
-1.30424118e+00 -3.00753653e-01 6.63887197e-03 1.55465916e-01
-1.31832290e+00 -9.70983088e-01 4.12462711e-01 -5.62009633e-01
9.27797914e-01 2.02193603e-01 9.41546857e-01 9.45947468e-01
8.31009224e-02 7.45270491e-01 8.79462600e-01 -5.52078366e-01
1.05931707e-01 -2.72042096e-01 -3.72021586e-01 7.99171388e-01
3.07783540e-02 5.18948793e-01 -1.15360188e+00 -1.60268456e-01
1.02708006e+00 -1.06340796e-02 -7.60239601e-01 -6.96453512e-01
-1.56054485e+00 4.60877687e-01 4.07015607e-02 -3.72071266e-01
-2.97476649e-01 6.37473702e-01 -1.58187225e-01 6.62828758e-02
3.91802251e-01 -1.09697632e-01 -3.29006016e-01 -6.02176905e-01
-1.15606582e+00 2.75516808e-01 5.82731903e-01 1.13842201e+00
8.63984048e-01 -7.02222586e-02 3.13163064e-02 1.04113355e-01
7.59896040e-01 7.76904345e-01 2.33892947e-02 -1.51694441e+00
2.37298951e-01 7.86895305e-02 5.15456915e-01 -5.75238049e-01
-2.01403603e-01 7.05222487e-02 -1.00268126e-01 6.75462425e-01
6.18008733e-01 -2.53885180e-01 -1.03830183e+00 1.48102224e+00
6.84554935e-01 4.85565513e-01 -2.50735223e-01 1.05647933e+00
6.59229815e-01 4.50973868e-01 -3.84045929e-01 -4.23876822e-01
1.06283581e+00 -5.96074522e-01 -8.46094072e-01 -2.51725405e-01
2.59607136e-01 -1.10591543e+00 4.55025584e-01 4.07862842e-01
-1.36151552e+00 -1.32022232e-01 -8.19840789e-01 -1.74096271e-01
3.17361623e-01 9.03348848e-02 4.87331271e-01 7.50182092e-01
-1.48755443e+00 4.37070429e-01 -1.22995901e+00 -6.34470165e-01
1.16643272e-01 3.21509868e-01 -2.61031628e-01 -8.79593417e-02
-5.55322349e-01 8.14566731e-01 -2.01956660e-01 -1.08397454e-01
-1.02982521e+00 -6.56397104e-01 -9.99674976e-01 -5.11366069e-01
-2.67446879e-02 -9.51638997e-01 1.58829570e+00 -7.50860512e-01
-1.63863993e+00 9.82460916e-01 -1.06079030e+00 -4.69077408e-01
7.31315911e-01 -2.98482180e-01 -1.09139964e-01 3.20922345e-01
1.02112316e-01 7.05546558e-01 1.01052892e+00 -1.19650185e+00
-8.14703047e-01 -1.72652125e-01 6.60801083e-02 6.25303268e-01
3.00301880e-01 6.11010864e-02 -1.01204646e+00 -3.13683063e-01
5.61768889e-01 -9.59996939e-01 -8.94251391e-02 6.35660052e-01
-1.16903462e-01 3.47863019e-01 7.41623580e-01 -6.47706330e-01
6.87053561e-01 -2.16529965e+00 2.12901384e-02 -1.93596512e-01
3.12320143e-01 -1.96603760e-01 4.05760497e-01 2.59003311e-01
2.81585604e-01 -5.64517319e-01 3.36475410e-02 -9.14636552e-01
-9.85748172e-02 1.83402568e-01 -2.96179533e-01 1.01595187e+00
-1.62677705e-01 5.82039058e-01 -1.04814446e+00 -4.74971145e-01
8.44602704e-01 6.07095659e-01 -5.49182236e-01 2.92461693e-01
-1.40006348e-01 6.01363122e-01 1.35819957e-01 7.92011380e-01
6.57178938e-01 -1.65349498e-01 1.02892086e-01 -2.17647657e-01
-4.07853216e-01 4.76235539e-01 -1.48524547e+00 2.22297788e+00
-1.79541528e-01 1.21794331e+00 1.50515646e-01 -4.47780490e-02
5.71955323e-01 3.96537662e-01 5.67831218e-01 -3.79185617e-01
-9.28661749e-02 1.24325640e-02 -7.91037560e-01 -2.64206052e-01
9.12841082e-01 2.56567635e-02 2.25848854e-01 2.16612220e-01
7.12526441e-02 -2.77935654e-01 -1.88574120e-01 3.69202107e-01
1.36424541e+00 7.36174226e-01 2.60011643e-01 1.98501185e-01
3.87859456e-02 3.48553121e-01 6.82936132e-01 6.69722199e-01
-4.36611235e-01 1.01028109e+00 -1.38511717e-01 -3.72527629e-01
-1.26420188e+00 -1.49661934e+00 -2.52415717e-01 3.09501320e-01
7.32694626e-01 -5.40596366e-01 -4.08836782e-01 -2.02635117e-02
-4.37999405e-02 2.34862730e-01 -1.21550366e-01 4.43044454e-01
-3.23361605e-01 -5.30805707e-01 2.31003195e-01 2.87804991e-01
4.03301090e-01 -4.75595653e-01 -8.81385446e-01 3.12316149e-01
-4.06755835e-01 -1.33320010e+00 -6.27302289e-01 -1.70828864e-01
-8.04582775e-01 -1.02646172e+00 -6.12560928e-01 -6.74122632e-01
6.45311177e-01 8.21888447e-01 1.00764871e+00 -2.52856821e-01
-3.17253768e-01 1.05660498e+00 -1.20175824e-01 -1.95224464e-01
-1.37890819e-02 -8.44756484e-01 2.87667215e-01 -7.18728220e-03
5.48968911e-01 -4.35773671e-01 -8.39263916e-01 3.26397210e-01
-3.52314413e-01 2.81472534e-01 6.22683354e-02 7.09316909e-01
8.37695301e-01 -4.12831932e-01 -4.94387150e-01 -2.79902853e-02
-2.80023456e-01 -1.75323859e-01 -1.14634228e+00 -2.16970876e-01
-3.26468676e-01 -5.63047417e-02 -2.90462881e-01 -3.23779166e-01
-1.19363129e+00 7.75198281e-01 3.48995090e-01 -7.96883643e-01
-2.01590899e-02 -1.66631833e-01 8.36616457e-02 -4.32389259e-01
6.81134403e-01 1.37399346e-01 1.00422792e-01 -3.47230226e-01
6.30313754e-01 4.59382594e-01 1.09900391e+00 -2.12208763e-01
8.92690003e-01 1.26958251e+00 -6.68750629e-02 -6.52711451e-01
-4.10861313e-01 -8.96117151e-01 -4.51555222e-01 -5.82319140e-01
9.33063090e-01 -1.62158763e+00 -8.72649014e-01 6.02766097e-01
-1.52005434e+00 -4.69898999e-01 -3.39978039e-01 8.02264035e-01
-7.55593002e-01 6.80531800e-01 -6.34348750e-01 -1.09168804e+00
9.26240310e-02 -1.04692495e+00 1.50919497e+00 1.92203775e-01
-1.46780178e-01 -6.25501752e-01 2.97045916e-01 5.94440252e-02
-1.44912094e-01 2.76023418e-01 -3.36882502e-01 5.28232753e-01
-1.40239906e+00 -1.25629650e-02 -1.21989965e-01 -2.70008057e-01
8.15935899e-03 -3.98007743e-02 -1.34305227e+00 -2.91028112e-01
8.81923363e-02 1.89162806e-01 6.07067883e-01 9.09757078e-01
4.02975947e-01 1.03150113e-02 -4.44316089e-01 9.86962974e-01
1.62484205e+00 1.68728903e-02 7.90205598e-01 6.52044356e-01
7.34589458e-01 2.01859742e-01 6.51015699e-01 6.78663254e-01
6.66854322e-01 1.13457263e+00 4.68543977e-01 -3.31439637e-02
-5.60799181e-01 -3.32964629e-01 7.29450703e-01 4.73436296e-01
-1.36478484e-01 -3.64055224e-02 -7.87940621e-01 9.87293124e-01
-1.98018694e+00 -1.32450640e+00 -5.62916338e-01 2.56226563e+00
6.37145281e-01 -1.34720787e-01 1.04232412e-02 -1.06025852e-01
8.74162257e-01 2.22313643e-01 -3.40642363e-01 1.36979297e-01
-4.19397682e-01 -3.81429717e-02 1.23128533e+00 8.28975916e-01
-1.01177871e+00 9.54642594e-01 6.85912514e+00 1.04453251e-01
-6.52480960e-01 3.04310620e-01 -1.47198215e-01 -6.79615498e-01
-1.37965113e-01 5.20620167e-01 -1.11293399e+00 4.80464906e-01
8.39943886e-01 1.65034421e-02 6.68572962e-01 4.96868610e-01
5.90811431e-01 -6.10909820e-01 -8.60669792e-01 1.63035500e+00
2.07230940e-01 -1.56535113e+00 -5.61608791e-01 4.06251103e-01
8.53610337e-01 5.45205116e-01 -3.23538393e-01 -5.92276990e-01
7.32788682e-01 -4.88651752e-01 1.17789996e+00 7.56639719e-01
8.30335438e-01 -3.45364779e-01 2.19343081e-01 8.71427264e-03
-1.42672110e+00 1.47977933e-01 -2.98486441e-01 -1.72535732e-01
1.04899037e+00 7.40585923e-01 -9.87831712e-01 3.60935003e-01
9.82476234e-01 9.51194882e-01 -2.10919201e-01 1.56054068e+00
-3.91264081e-01 3.27033818e-01 -6.31091416e-01 4.61230487e-01
-1.62218451e-01 1.29487375e-02 1.01805663e+00 8.16341519e-01
5.87289691e-01 8.51493105e-02 -7.83202797e-03 5.74045002e-01
1.80220470e-01 -6.00621343e-01 -8.34933400e-01 5.65857053e-01
7.69846618e-01 8.49860489e-01 -1.42136395e-01 -3.15674126e-01
-5.46837628e-01 1.29505575e+00 -2.00895935e-01 3.18335146e-01
-6.40994906e-01 -4.72160764e-02 9.98994648e-01 7.18419477e-02
4.47913766e-01 -7.59166300e-01 -1.50381118e-01 -1.41862237e+00
1.98013678e-01 -4.60731179e-01 7.18421787e-02 -1.34892464e+00
-8.94504905e-01 1.87751070e-01 -2.95716494e-01 -1.74380481e+00
-4.06897724e-01 -5.65963328e-01 -2.43236974e-01 7.86103547e-01
-1.58527470e+00 -8.05005193e-01 -7.77331173e-01 9.85453129e-01
6.07317388e-01 1.97514057e-01 5.61033607e-01 2.81245083e-01
1.08207837e-02 1.38906494e-01 3.14228356e-01 -7.63665661e-02
9.87294734e-01 -1.18854499e+00 7.35017955e-01 1.50402653e+00
3.88340503e-01 8.25520679e-02 9.31871712e-01 -7.29745805e-01
-1.84719133e+00 -7.91584074e-01 9.70708728e-01 -1.08606422e+00
5.06224692e-01 -3.14245105e-01 -3.33980888e-01 9.41899598e-01
6.37191087e-02 2.60737151e-01 6.16192594e-02 -2.77344972e-01
6.10274747e-02 8.47633481e-02 -1.02083516e+00 5.60251057e-01
1.49712622e+00 -9.84337807e-01 -4.92651045e-01 3.61270040e-01
5.61167061e-01 -1.05400681e+00 -6.54627502e-01 -2.05605887e-02
6.05443835e-01 -1.09297049e+00 1.14384949e+00 1.48693159e-01
-3.17630470e-01 -9.16815221e-01 -3.98104668e-01 -8.17149162e-01
2.66515221e-02 -1.22723997e+00 -4.60703939e-01 1.01017904e+00
-8.92386362e-02 -4.22741652e-01 9.45911884e-01 7.93353915e-01
1.14739500e-01 2.61513025e-01 -1.24611914e+00 -7.68598080e-01
-9.28587675e-01 -7.22852230e-01 3.72108012e-01 5.60613394e-01
-2.00891063e-01 -2.31730521e-01 -7.54250884e-01 6.92063510e-01
1.17830431e+00 -1.75443336e-01 1.18585348e+00 -1.09897578e+00
-4.80578333e-01 1.57051116e-01 -9.89345670e-01 -1.50270510e+00
-2.06469595e-01 -3.18184406e-01 4.61827129e-01 -1.42956102e+00
-1.45954713e-01 -1.77218616e-01 4.08156931e-01 9.77227092e-02
1.37395054e-01 4.25630182e-01 1.52709261e-01 5.60691297e-01
-5.75897932e-01 3.65289032e-01 6.63048923e-01 1.75155550e-01
-3.71431768e-01 -3.04017216e-01 5.34350760e-02 7.03270614e-01
2.50764102e-01 -4.67445999e-01 -6.11425750e-02 -7.63636470e-01
7.59440064e-02 1.96028844e-01 8.89454842e-01 -1.14787090e+00
7.69061506e-01 9.62150693e-02 4.77064222e-01 -8.54689062e-01
8.75725150e-01 -6.52466416e-01 6.33655429e-01 1.47874430e-01
3.38241428e-01 1.30961865e-01 1.47260919e-01 8.79222155e-01
1.44511551e-01 1.43520311e-01 5.67897856e-01 -9.99226049e-02
-1.19560111e+00 3.59593362e-01 -4.26935762e-01 -2.32286707e-01
8.03943813e-01 -5.15621543e-01 -3.52149367e-01 -6.95576489e-01
-6.69482946e-01 -2.19866768e-01 1.11919725e+00 2.11750805e-01
6.38249159e-01 -1.41898143e+00 -3.98639262e-01 1.53560862e-02
7.05342516e-02 1.19668148e-01 1.22646078e-01 9.85319853e-01
-7.40391850e-01 1.40457809e-01 1.67649329e-01 -1.25394177e+00
-1.45969927e+00 2.52033949e-01 4.46811616e-01 4.79292035e-01
-1.07572103e+00 7.68305421e-01 4.59172428e-02 3.89761403e-02
2.88695961e-01 5.04189357e-02 4.66435343e-01 -3.23661625e-01
7.44274855e-01 3.65556538e-01 -4.04043458e-02 -7.90754199e-01
-3.55345249e-01 7.14471698e-01 4.64612871e-01 -8.76270950e-01
1.09819305e+00 -8.83730471e-01 2.07983091e-01 3.54872465e-01
8.79422069e-01 2.59635061e-01 -2.01677704e+00 -3.02844584e-01
-2.83635527e-01 -1.04237807e+00 5.20233214e-01 -3.76891345e-01
-7.07010925e-01 5.64783514e-01 1.04042828e+00 -2.94320613e-01
1.06329274e+00 9.49620530e-02 6.35976553e-01 3.33448239e-02
7.61649787e-01 -7.70870864e-01 -4.36140090e-01 2.73610026e-01
2.95080483e-01 -1.16598153e+00 1.15053892e-01 -3.35733443e-01
-3.51389676e-01 1.00101221e+00 1.99301928e-01 1.17778018e-01
3.08544159e-01 3.12972724e-01 2.19419315e-01 -1.72440767e-01
-4.99877900e-01 -5.15812635e-01 -3.29113603e-02 9.80408728e-01
8.64746198e-02 -1.55764893e-01 1.84923872e-01 -5.18580139e-01
-1.60809383e-02 1.05150081e-01 7.97066867e-01 9.56963420e-01
-3.06001991e-01 -1.14288521e+00 -8.49229813e-01 -8.55168030e-02
-2.07752120e-02 -2.08732516e-01 -8.37238953e-02 6.52676642e-01
-5.61621115e-02 1.17937028e+00 2.32014671e-01 -1.89625844e-01
8.51375088e-02 -1.03693321e-01 7.29447603e-01 -3.00170124e-01
-3.73789668e-02 2.17035264e-01 3.96768004e-01 -1.10695398e+00
-9.68635440e-01 -1.21018481e+00 -9.93299544e-01 -6.44939184e-01
-2.60751307e-01 -4.10079956e-01 8.88164341e-01 7.94532418e-01
5.87319076e-01 -3.48697305e-02 6.06869161e-01 -1.45057559e+00
1.23991996e-01 -3.51822197e-01 -4.94835049e-01 2.37857103e-01
8.01640630e-01 -7.37963676e-01 -5.66216707e-01 4.81864929e-01] | [8.219805717468262, -1.889723539352417] |
1291d6bd-973a-430b-8c4c-b65ce347d399 | extracting-food-substitutes-from-food-diary | 1607.08807 | null | http://arxiv.org/abs/1607.08807v1 | http://arxiv.org/pdf/1607.08807v1.pdf | Extracting Food Substitutes From Food Diary via Distributional Similarity | In this paper, we explore the problem of identifying substitute relationship
between food pairs from real-world food consumption data as the first step
towards the healthier food recommendation. Our method is inspired by the
distributional hypothesis in linguistics. Specifically, we assume that foods
that are consumed in similar contexts are more likely to be similar dietarily.
For example, a turkey sandwich can be considered a suitable substitute for a
chicken sandwich if both tend to be consumed with french fries and salad. To
evaluate our method, we constructed a real-world food consumption dataset from
MyFitnessPal's public food diary entries and obtained ground-truth human
judgements of food substitutes from a crowdsourcing service. The experiment
results suggest the effectiveness of the method in identifying suitable
substitutes. | ['Weber Ingmar', 'Achananuparp Palakorn'] | 2016-07-29 | null | null | null | null | ['food-recommendation'] | ['miscellaneous'] | [-6.07136674e-02 8.97872746e-02 -6.68866158e-01 -6.91352367e-01
-3.64535779e-01 -7.60967076e-01 2.77176857e-01 1.04224110e+00
-5.77510715e-01 4.71307874e-01 9.07975912e-01 -1.47767276e-01
2.03155532e-01 -9.73975420e-01 -7.79150546e-01 -4.16189700e-01
3.11680794e-01 3.57807398e-01 7.56043568e-02 -4.07260835e-01
-3.71539071e-02 -2.69458055e-01 -1.35579336e+00 4.93524432e-01
9.39956665e-01 5.51881075e-01 6.02147460e-01 1.08901961e-02
-2.14016289e-01 5.45634866e-01 -6.63846731e-03 -9.17099178e-01
5.50573528e-01 -6.10140026e-01 -7.87597179e-01 1.59072027e-01
6.00883365e-02 -2.44767323e-01 1.42323211e-01 1.44919062e+00
3.55428040e-01 2.63710350e-01 7.85819530e-01 -1.02393901e+00
-1.33696103e+00 1.53288543e+00 -2.67075121e-01 -1.39580872e-02
8.85770440e-01 -4.06661779e-02 1.02911675e+00 -6.09670043e-01
2.63309777e-01 1.45978749e+00 7.39589274e-01 3.80124927e-01
-1.12417948e+00 -5.62914371e-01 1.10610515e-01 1.38010889e-01
-1.33301473e+00 -2.70695806e-01 5.69808066e-01 -4.65890080e-01
7.08413780e-01 3.46095502e-01 8.92690778e-01 7.27478504e-01
-6.09531486e-03 7.57887781e-01 1.17835891e+00 -6.66925192e-01
2.61913002e-01 3.21506292e-01 3.30966055e-01 3.51849228e-01
6.46463811e-01 2.72120200e-02 -2.42327482e-01 -3.16842705e-01
2.61432797e-01 2.19472408e-01 -1.57971844e-01 -5.94955906e-02
-1.24896026e+00 1.25876057e+00 5.40649354e-01 4.33730185e-01
-6.42595172e-01 -2.48985693e-01 3.77850533e-01 3.21180373e-01
7.17988014e-01 3.98599982e-01 -3.68032247e-01 5.42178392e-01
-4.67253536e-01 5.22088230e-01 1.10872602e+00 1.10950589e+00
4.87433434e-01 -6.42514586e-01 1.29258454e-01 8.35264325e-01
8.23164165e-01 7.84078240e-01 6.54770017e-01 -6.70678377e-01
1.65923133e-01 5.11455953e-01 7.39464700e-01 -1.38411915e+00
-5.04426301e-01 3.62812668e-01 -4.31387752e-01 -5.88876724e-01
8.21355045e-01 1.16626367e-01 -1.93368956e-01 1.56843364e+00
6.89362168e-01 -3.96368146e-01 1.86482608e-01 1.24538648e+00
9.37185109e-01 3.30443054e-01 8.19920540e-01 -1.43907145e-01
1.83219075e+00 -8.49357605e-01 -9.27166104e-01 6.31935708e-03
8.32399249e-01 -9.96470094e-01 1.09807813e+00 1.75923258e-01
-8.01087737e-01 -6.61543906e-01 -8.22251379e-01 -2.67677009e-01
-5.52340806e-01 -3.50483917e-02 6.79780841e-01 7.05468059e-01
-4.60086763e-01 6.47710502e-01 -5.22697568e-01 -9.34538484e-01
-3.82484645e-01 -1.41060472e-01 -2.50505894e-01 -3.97976190e-01
-1.43389571e+00 1.02694011e+00 6.79916978e-01 -5.56885675e-02
-2.71487385e-01 -4.42814261e-01 -1.30895615e+00 -3.69600654e-01
2.69110054e-01 -5.17089844e-01 1.42384422e+00 -1.16953564e+00
-1.02033091e+00 1.17751479e+00 -3.68582725e-04 -6.80335224e-01
1.18271716e-01 -2.46502668e-01 -9.55078721e-01 1.20961049e-03
7.33707190e-01 5.62700748e-01 1.63022101e-01 -9.37646449e-01
-7.19230652e-01 -3.32572579e-01 3.54711443e-01 7.35373199e-02
9.08149034e-02 4.58127916e-01 5.19457817e-01 -8.35795820e-01
1.19612232e-01 -8.79581034e-01 -3.15226704e-01 8.25319514e-02
-2.46122926e-01 -6.60312772e-01 -3.20995986e-01 -7.00885713e-01
1.10694206e+00 -2.24183941e+00 -4.14990067e-01 1.37455478e-01
-1.66081637e-01 5.06374575e-02 2.33149761e-03 5.41245282e-01
9.24762413e-02 1.46890387e-01 8.80321562e-02 2.65920877e-01
6.54219568e-01 3.43464017e-01 5.74909942e-03 6.40470624e-01
-2.40181461e-01 6.42898083e-01 -1.61581433e+00 -5.71925402e-01
1.02649428e-01 1.39675230e-01 -3.46092224e-01 5.07644787e-02
-4.39008087e-01 1.51888179e-02 -3.41775119e-01 6.30569220e-01
6.51607871e-01 3.66747454e-02 5.93390048e-01 -7.43228018e-01
-3.44178796e-01 4.54874814e-01 -1.24255955e+00 1.45692885e+00
-1.93059351e-02 -3.54712009e-01 -1.28668204e-01 -6.32292926e-01
8.06990862e-01 2.19089359e-01 2.09726408e-01 -7.67399549e-01
2.26826832e-01 4.73883480e-01 1.11119509e-01 -9.29112017e-01
5.96979141e-01 -3.62557858e-01 -1.64917737e-01 5.30470014e-01
-4.92308527e-01 -1.54739156e-01 5.78685462e-01 -2.80131489e-01
3.34020138e-01 3.05422664e-01 1.02695036e+00 -1.05773115e+00
5.41710019e-01 5.18362284e-01 6.38151467e-01 1.73763633e-01
-5.01521647e-01 7.73186535e-02 -4.08750921e-01 -5.93878388e-01
-9.91267025e-01 -9.95852590e-01 -1.07521884e-01 1.35387850e+00
3.20240855e-01 -2.56272495e-01 -4.82002765e-01 -6.04200840e-01
1.63769841e-01 1.11154735e+00 -6.12335980e-01 2.78757513e-02
-4.98492360e-01 -5.54878175e-01 2.61092901e-01 4.98247415e-01
6.05920732e-01 -7.12958872e-01 -6.38518453e-01 4.27399933e-01
-6.11104906e-01 -6.20028913e-01 -1.29764903e+00 1.32760584e-01
-2.87671983e-01 -1.32907486e+00 -7.46514201e-01 -1.17211294e+00
3.99338275e-01 9.04717207e-01 1.55686080e+00 1.44246250e-01
3.16540003e-01 -6.50189910e-03 -7.65982866e-01 -7.36481249e-01
-8.71779323e-01 -4.36971039e-01 4.37302887e-01 -2.25605384e-01
1.50858760e+00 -1.70144737e-01 -9.41800594e-01 5.12256503e-01
-1.02304709e+00 -1.35689974e-01 -2.11766213e-01 4.80395317e-01
6.40585363e-01 1.76308043e-02 5.36562502e-01 -8.32105160e-01
5.15644431e-01 -1.12787712e+00 -1.65007457e-01 2.33494744e-01
-4.53408867e-01 -7.12418109e-02 5.49493968e-01 -7.96587706e-01
-6.98240697e-01 4.87378865e-01 -1.57423645e-01 5.96010447e-01
-2.43146718e-01 4.42817599e-01 -1.63639411e-01 4.79379714e-01
8.28767955e-01 -2.08331689e-01 -3.26557070e-01 -4.75289941e-01
6.08594954e-01 7.67578304e-01 3.19299400e-01 -6.76056623e-01
1.44276008e-01 1.81320310e-01 -6.86857820e-01 -5.81721961e-01
-9.32352841e-01 -7.07772672e-01 -4.40653801e-01 5.77156506e-02
1.24214458e+00 -1.08773446e+00 -9.64791834e-01 -6.48510829e-02
-7.67950296e-01 -2.12481603e-01 -9.74044725e-02 9.38480496e-01
-2.31941268e-01 4.09621239e-01 -5.99964023e-01 -6.97011590e-01
-2.69810975e-01 -7.17157483e-01 8.07493031e-01 5.27762510e-02
-8.41125667e-01 -1.08493125e+00 3.14185113e-01 1.12406418e-01
-3.23970593e-03 2.31693864e-01 9.76325870e-01 -9.93410468e-01
3.82543981e-01 7.29256570e-02 -1.07972875e-01 8.55247453e-02
5.47133088e-01 -2.87658244e-01 -3.74510527e-01 5.77368923e-02
2.37229332e-01 -1.91217005e-01 4.37750369e-01 3.80913556e-01
2.06825256e-01 -5.92705488e-01 -2.22692668e-01 -2.18660057e-01
1.43153536e+00 2.56671250e-01 1.83017477e-01 1.42593130e-01
3.43327761e-01 9.97647345e-01 9.95429277e-01 3.46226633e-01
1.06165171e+00 5.64126074e-01 -8.95878896e-02 8.12799037e-02
-8.05923939e-02 -6.97243214e-01 4.53612149e-01 1.03549576e+00
3.25670600e-01 -4.44541126e-01 -6.09242439e-01 8.87740195e-01
-1.67202246e+00 -8.91363561e-01 -5.89185417e-01 2.03900647e+00
1.26826859e+00 -4.93182629e-01 7.46259093e-01 -5.17725646e-02
7.62943387e-01 -4.19935435e-01 -2.41156831e-01 -1.86780676e-01
-2.00877283e-02 -1.52546898e-01 5.26912093e-01 3.71271193e-01
-1.08381438e+00 6.29109025e-01 6.84394836e+00 3.93039644e-01
-4.54530686e-01 3.25626969e-01 4.24499899e-01 4.05639023e-01
-6.71716332e-01 -5.52471578e-01 -5.35324812e-01 7.71200478e-01
8.71438563e-01 -1.80089310e-01 3.73059988e-01 5.13461351e-01
2.62138307e-01 -3.21694285e-01 -1.41397357e+00 7.87341535e-01
-4.11927141e-02 -6.96821332e-01 -2.05266356e-01 -4.76420701e-01
5.01675367e-01 -2.36635014e-01 -4.51337457e-01 6.85162023e-02
1.01498413e+00 -5.87576449e-01 1.24529576e+00 2.10442632e-01
1.24201067e-01 -6.29306555e-01 6.91132307e-01 4.95773405e-01
-1.47589421e+00 1.19446963e-01 -9.20858204e-01 -3.56378585e-01
2.01970309e-01 7.62256205e-01 -6.15480423e-01 4.72053230e-01
7.12990582e-01 6.69265747e-01 -2.14390919e-01 8.07859123e-01
-2.44916201e-01 3.24003696e-01 -4.45349455e-01 -4.05902773e-01
8.04148167e-02 -5.31746686e-01 -5.80929704e-02 1.25568843e+00
4.93963450e-01 4.96831924e-01 5.93060374e-01 9.96831894e-01
1.31392807e-01 7.21530259e-01 -7.68174052e-01 -1.39223129e-01
3.92928153e-01 9.07484591e-01 -8.71817172e-01 -3.01066667e-01
-6.27367914e-01 9.34184313e-01 -9.75449337e-04 -5.21760248e-02
-8.12869191e-01 2.58126706e-01 5.86130142e-01 4.80446666e-02
-1.07549421e-01 1.10239089e-01 1.14261203e-01 -9.28269804e-01
-1.95394248e-01 -7.95373619e-01 4.18115795e-01 -6.35062993e-01
-1.95529687e+00 3.59418452e-01 1.14617594e-01 -1.11167109e+00
-8.72610509e-02 -3.73741001e-01 9.81042860e-04 8.51183116e-01
-1.20297778e+00 -1.10131764e+00 5.78040965e-02 5.74803233e-01
5.07753730e-01 3.94544929e-01 1.09560072e+00 4.51503307e-01
-3.04590445e-02 3.24963868e-01 -6.21460825e-02 9.96094048e-02
8.25330079e-01 -1.10762882e+00 5.69578767e-01 5.87493300e-01
1.65716723e-01 8.00911188e-01 1.19976878e+00 -1.18845379e+00
-1.27008355e+00 -1.06980145e+00 1.48726106e+00 -3.47735971e-01
8.08166623e-01 -2.11018458e-01 -6.78409040e-01 4.84702557e-01
2.89953232e-01 -6.03621960e-01 1.19458878e+00 -1.97177052e-01
-6.23554647e-01 2.88972497e-01 -1.73931563e+00 4.70629305e-01
1.16693926e+00 -4.94570404e-01 -9.74261582e-01 7.98562884e-01
8.85283649e-01 -1.50620965e-02 -1.12162185e+00 -1.49138018e-01
6.65374219e-01 -5.19807041e-01 1.01299894e+00 -4.01483566e-01
3.28405619e-01 -4.75679427e-01 -7.19537437e-01 -1.61629379e+00
-6.12955391e-01 -6.35694563e-02 5.48304319e-01 1.15333951e+00
3.29989880e-01 -4.81500357e-01 8.27906877e-02 1.07619619e+00
-4.27240692e-02 4.21222262e-02 -3.80112588e-01 -7.99458861e-01
9.24699828e-02 -7.48481974e-02 1.25106239e+00 1.30472958e+00
3.60437810e-01 1.15110211e-01 -3.13909531e-01 7.04454482e-02
3.50025535e-01 2.06163615e-01 2.18980461e-01 -9.15997207e-01
-3.48647416e-01 -5.76018803e-02 1.14344910e-01 -9.89627421e-01
1.61983311e-01 -1.17657614e+00 4.24634963e-01 -1.40784824e+00
3.74936879e-01 -3.10252279e-01 -4.16910015e-02 2.75931090e-01
-2.33269721e-01 3.65089625e-01 3.23612720e-01 1.07971309e-02
-2.71992266e-01 1.05675887e-02 1.07631576e+00 -4.01998132e-01
-2.49265701e-01 -1.60551686e-02 -1.12114978e+00 9.06500697e-01
8.19864750e-01 -7.59748995e-01 -5.02863526e-01 -5.16473591e-01
6.71516299e-01 -2.00265899e-01 1.91688433e-01 -3.43634725e-01
-9.40121189e-02 -4.48951215e-01 1.91669315e-02 -3.15967262e-01
-2.87023216e-01 -1.35091424e+00 8.26580048e-01 7.44292915e-01
-7.48884797e-01 4.02579427e-01 -3.02144527e-01 6.48876607e-01
1.91535518e-01 -5.97500980e-01 4.14469719e-01 -5.62388957e-01
-5.13080478e-01 -1.48338318e-01 -3.05388123e-01 -1.18862994e-01
9.05653536e-01 -3.02433432e-03 -2.19630703e-01 1.24030747e-01
-6.05590105e-01 1.65421158e-01 7.06928670e-01 5.00925362e-01
1.76043555e-01 -1.80540824e+00 -1.02162337e+00 4.21759002e-02
4.97312069e-01 -8.18423748e-01 -2.14848042e-01 6.73946023e-01
-5.99388897e-01 1.19118899e-01 -9.23145935e-02 -1.12289667e-01
-8.87301385e-01 1.35731268e+00 1.01871207e-01 1.81898683e-01
-5.09558141e-01 4.15925920e-01 2.64040560e-01 -5.52885950e-01
-9.58472565e-02 -1.03159857e+00 -2.86922663e-01 2.31160372e-02
9.60344613e-01 4.39810306e-01 -2.15107456e-01 -8.96070182e-01
-3.23998421e-01 2.34644562e-01 2.90684283e-01 3.49816710e-01
1.08285785e+00 -5.17217457e-01 -2.62530178e-01 6.97597146e-01
9.47097063e-01 1.37273207e-01 -5.88676035e-01 -3.62591475e-01
2.78789639e-01 -5.93263328e-01 -4.53994244e-01 -8.55808020e-01
-8.04777503e-01 -3.53140160e-02 5.51024079e-01 6.86159492e-01
9.63719964e-01 1.33972198e-01 1.12389326e+00 1.35341018e-01
7.54113495e-01 -1.00933766e+00 -6.48115993e-01 -1.12531900e-01
7.24557877e-01 -1.49928343e+00 -1.18115686e-01 -7.21531868e-01
-6.77885056e-01 9.39769208e-01 3.15444693e-02 -4.01032150e-01
1.04865038e+00 -2.30441600e-01 3.18238810e-02 -1.90467551e-01
-1.29227236e-01 -3.07388514e-01 4.49767113e-01 8.50082576e-01
9.35694695e-01 8.77470732e-01 -1.14000356e+00 1.10375595e+00
-2.40803063e-01 6.00591749e-02 1.68652788e-01 5.96171200e-01
-4.32851225e-01 -1.04696631e+00 -3.16537023e-01 3.26531649e-01
-3.91503274e-01 -2.88300395e-01 -3.06729406e-01 6.60441637e-01
5.31666517e-01 1.54021621e+00 -1.88628137e-01 -1.41157493e-01
4.74866897e-01 -2.57760491e-02 6.30645037e-01 -6.01793587e-01
-1.10143709e+00 2.86585063e-01 3.86842251e-01 -4.52857673e-01
-1.43538713e+00 -6.87660873e-01 -1.29762959e+00 -6.42904699e-01
-2.51947373e-01 2.99614012e-01 3.46737891e-01 9.75036919e-01
-1.26683757e-01 -1.26038447e-01 3.84865761e-01 -5.05624712e-01
-6.79285228e-01 -1.01076102e+00 -9.49015617e-01 1.11532354e+00
-8.32058564e-02 -6.35582387e-01 -2.55086310e-02 3.86747509e-01] | [11.530132293701172, 4.524235725402832] |
f90e9554-08ac-4d70-acb9-1af5bf4d2fa5 | integration-of-deep-learning-and-traditional | null | null | https://aclanthology.org/D19-5724 | https://aclanthology.org/D19-5724.pdf | Integration of Deep Learning and Traditional Machine Learning for Knowledge Extraction from Biomedical Literature | In this paper, we present our participation in the Bacteria Biotope (BB) task at BioNLP-OST 2019. Our system utilizes fine-tuned language representation models and machine learning approaches based on word embedding and lexical features for entities recognition, normalization and relation extraction. It achieves the state-of-the-art performance and is among the top two systems in five of all six subtasks. | ['Wanli Liu', 'Jihang Mao'] | 2019-11-01 | null | null | null | ws-2019-11 | ['medical-concept-normalization'] | ['medical'] | [ 2.68974662e-01 1.72646761e-01 -2.42713302e-01 7.44746476e-02
-3.28479111e-01 -2.97598243e-01 9.01012540e-01 1.01082706e+00
-9.52548087e-01 1.14624119e+00 4.03486311e-01 -5.50424576e-01
-1.24292776e-01 -6.86908841e-01 -6.66283965e-01 -5.15839636e-01
-4.08652067e-01 5.36144555e-01 5.09431064e-02 -6.15623474e-01
3.04968238e-01 7.85175383e-01 -1.00669086e+00 5.76351345e-01
1.68388650e-01 5.76677084e-01 -1.08770601e-01 9.24707234e-01
-3.74142528e-01 9.30924594e-01 -5.04717886e-01 -6.43979549e-01
-2.99888581e-01 1.87722445e-01 -1.07817340e+00 -9.59976077e-01
-3.36636752e-01 4.97374028e-01 -3.45390379e-01 7.96479166e-01
8.68532836e-01 -1.92880422e-01 8.59657407e-01 -8.79274845e-01
-8.28162253e-01 7.85400510e-01 -1.45962864e-01 6.60572290e-01
8.08293700e-01 -1.27401561e-01 9.19443429e-01 -1.18770981e+00
9.99673247e-01 1.28093398e+00 8.92424226e-01 5.55298686e-01
-1.24279714e+00 -4.72310215e-01 -2.40210414e-01 2.76728928e-01
-1.43917727e+00 -7.24879503e-01 -4.86898199e-02 -6.35903478e-01
2.42745423e+00 6.90685064e-02 4.37384903e-01 1.32934988e+00
8.58113110e-01 2.31434003e-01 8.62439096e-01 -4.83702838e-01
4.33596708e-02 -9.50342342e-02 3.05761904e-01 1.00108027e+00
1.04373741e+00 1.76505819e-02 -7.56632268e-01 -5.81264257e-01
4.25927758e-01 -2.65994161e-01 -2.75401652e-01 -1.64363027e-01
-1.62862349e+00 6.87826335e-01 1.52123734e-01 6.62629545e-01
-7.03499198e-01 3.23617272e-02 7.07252562e-01 1.20669536e-01
6.58764124e-01 1.11291814e+00 -1.22273219e+00 -7.05927759e-02
-1.54087096e-01 1.95288301e-01 1.50051808e+00 9.49038386e-01
3.31213564e-01 -2.01836511e-01 -1.06965788e-01 7.72797167e-01
3.45694155e-01 7.51444921e-02 8.52252483e-01 5.05897962e-02
3.13135922e-01 5.11186838e-01 8.05912688e-02 -8.33227575e-01
-9.07556355e-01 -7.59358704e-02 -6.19453311e-01 -4.52871442e-01
-8.65388215e-02 -1.33440867e-01 -1.02398181e+00 1.34648943e+00
2.55488008e-01 1.80086628e-01 7.67039955e-01 7.79194534e-02
1.43720710e+00 6.00786507e-01 6.67913735e-01 -2.38944679e-01
1.88171387e+00 -7.90175259e-01 -1.06785047e+00 -9.71267000e-02
8.45195651e-01 -9.29163933e-01 1.74923480e-01 2.71296084e-01
-7.55167663e-01 -9.09963027e-02 -1.19175994e+00 -2.53850281e-01
-1.56658161e+00 -5.23559093e-01 6.87844515e-01 4.06627923e-01
-7.89534390e-01 5.26429892e-01 -9.65079784e-01 -1.00751436e+00
1.08726941e-01 4.68669325e-01 -8.74937415e-01 4.24947217e-02
-1.40082633e+00 1.29942632e+00 6.42963171e-01 -4.03791666e-01
-7.85055697e-01 -7.10925698e-01 -1.11632764e+00 -2.47261479e-01
1.54005125e-01 -8.44322503e-01 8.16728771e-01 6.57802463e-01
-1.43653941e+00 1.07469034e+00 -2.14725956e-01 -6.95988476e-01
-1.36346683e-01 -4.68406498e-01 -8.28245401e-01 -1.66806594e-01
8.83933008e-02 3.08212638e-01 -4.39866707e-02 -6.64232254e-01
-4.65780497e-01 -2.48544946e-01 -2.35532299e-01 -1.81889504e-01
-3.44833374e-01 4.57950503e-01 -1.57179579e-01 -5.35225749e-01
-3.67675930e-01 -4.99630392e-01 -4.25230682e-01 -6.59509540e-01
-4.33889031e-01 -6.05092049e-01 1.89738199e-01 -8.03930521e-01
1.04657972e+00 -1.76447928e+00 3.04910451e-01 -2.33141243e-01
1.53058633e-01 5.08448422e-01 -4.73415196e-01 9.32091236e-01
-3.29416901e-01 4.16495949e-01 1.29773065e-01 -8.73839017e-03
2.17604283e-02 2.72654712e-01 -4.34577912e-02 4.61563855e-01
7.90258408e-01 9.41495955e-01 -1.22804642e+00 -2.90181756e-01
5.63674495e-02 8.11246097e-01 -1.43009111e-01 3.58296633e-01
-1.03071719e-01 -1.31170750e-02 -4.72368985e-01 7.03851283e-01
2.21364602e-01 -2.22986057e-01 3.66193503e-01 -4.64437842e-01
-2.72329479e-01 6.49166346e-01 -6.74016416e-01 1.62143803e+00
-2.34906211e-01 4.36074495e-01 -1.12549677e-01 -8.50947618e-01
8.05973828e-01 5.54450691e-01 5.74990511e-01 -1.11592479e-01
-3.52026187e-02 2.27444291e-01 2.81115677e-02 -1.00129724e+00
3.86679769e-01 2.20327899e-01 1.55525785e-02 -1.85445901e-02
3.86697739e-01 -2.44049914e-02 4.22869027e-01 -1.83047831e-01
1.59792209e+00 3.83499414e-01 1.58404434e+00 -5.77452362e-01
7.18822062e-01 -1.64714269e-02 4.81771618e-01 5.10659993e-01
-1.48396283e-01 -4.58717048e-02 2.61846155e-01 -8.35262060e-01
-9.00789738e-01 -6.85342252e-01 -4.11861837e-01 1.45672190e+00
-2.11253792e-01 -8.46311986e-01 -4.87842649e-01 -5.56615055e-01
2.05778927e-01 5.75217664e-01 -7.95145214e-01 7.51550719e-02
-4.97761816e-01 -1.17873907e+00 1.26218510e+00 1.89923793e-01
-1.71407327e-01 -1.30036008e+00 -5.18938243e-01 7.72885978e-01
-1.57739565e-01 -1.18459225e+00 2.05561534e-01 1.11303699e+00
-2.86424398e-01 -1.50640333e+00 -4.45159078e-01 -8.11737716e-01
2.01680213e-01 -4.21965897e-01 1.21750712e+00 -4.06099170e-01
-8.77636373e-01 -1.34154916e-01 -3.72765183e-01 -1.05165362e+00
-4.24499780e-01 3.62369210e-01 1.33466542e-01 -7.15265393e-01
9.98075366e-01 -3.90324295e-02 -7.53703713e-02 -2.51907766e-01
-8.65072072e-01 -2.44918361e-01 5.24135292e-01 9.29099381e-01
7.27253020e-01 -3.07812750e-01 7.37394750e-01 -8.63180161e-01
9.07132924e-01 -6.04856908e-01 -2.79586732e-01 5.51913917e-01
-7.72411644e-01 3.83407056e-01 3.46526653e-01 -2.63358235e-01
-5.67015290e-01 -6.88686669e-02 -6.00417435e-01 6.32512569e-01
-2.99520433e-01 8.51822615e-01 -1.22666784e-01 -1.66802436e-01
8.70427489e-01 1.13309138e-01 -4.63661283e-01 -4.84584212e-01
4.10167634e-01 9.00752962e-01 1.54845744e-01 -3.09043974e-01
7.12030157e-02 7.08019286e-02 1.10050410e-01 -8.90918016e-01
-7.47264743e-01 -8.10996532e-01 -6.00589275e-01 4.92241561e-01
1.13259745e+00 -1.07821083e+00 -7.00028121e-01 2.77631283e-01
-1.58685374e+00 5.00447489e-02 -3.71901304e-01 6.48329198e-01
-3.49442273e-01 6.73254719e-03 -9.39869225e-01 -5.77406168e-01
-7.31619358e-01 -9.03791487e-01 1.11022842e+00 -2.31983989e-01
-4.35560942e-01 -9.32180226e-01 7.25336373e-01 -1.39085382e-01
3.18118095e-01 4.85951394e-01 1.02767479e+00 -1.43389273e+00
2.32187942e-01 -2.72799075e-01 -2.65639812e-01 -1.67697594e-01
5.76187015e-01 -9.81715322e-02 -8.56516302e-01 -2.11426675e-01
-3.60289216e-01 -3.10434908e-01 1.19190192e+00 -9.73841771e-02
7.98873007e-01 -3.07147622e-01 -7.61287570e-01 7.40446568e-01
1.42725956e+00 2.44854435e-01 5.41370094e-01 7.35117793e-01
6.53923452e-01 6.04124665e-01 2.83906460e-01 4.56751317e-01
3.23533267e-01 3.08603734e-01 3.18959594e-01 1.69419631e-01
4.28928547e-02 7.87172168e-02 1.81690589e-01 1.04960465e+00
-1.24740832e-01 -6.33359671e-01 -1.46571016e+00 7.44380891e-01
-1.81828070e+00 -6.54065311e-01 -8.40328336e-02 1.52006567e+00
1.35318935e+00 -3.13435733e-01 -4.16671485e-01 -1.89608797e-01
4.22685087e-01 1.13397114e-01 -2.76603937e-01 -7.85932541e-01
-3.60364407e-01 6.85816228e-01 6.85405850e-01 3.68181825e-01
-1.40587354e+00 1.21168661e+00 7.46297836e+00 5.35421789e-01
-5.95799267e-01 1.59315854e-01 2.37722740e-01 4.15762395e-01
2.07933217e-01 -5.89121878e-01 -1.12432241e+00 -1.27119407e-01
1.45736432e+00 -3.01243067e-01 4.74103361e-01 4.35214251e-01
-3.10523480e-01 4.18911397e-01 -1.38331640e+00 6.73911154e-01
3.11394811e-01 -1.54249036e+00 2.65022933e-01 1.39210895e-01
4.18294787e-01 8.44353855e-01 -5.11997759e-01 2.86811471e-01
7.48933077e-01 -1.56136739e+00 8.96942392e-02 5.35452187e-01
6.12437248e-01 -6.52045727e-01 1.19730687e+00 -4.94321473e-02
-1.25069892e+00 1.41416356e-01 -6.43442988e-01 1.51395515e-01
-4.62971553e-02 5.90362251e-01 -1.10967636e+00 9.41987455e-01
6.87043786e-01 8.39395523e-01 -3.03840786e-01 8.14877689e-01
-1.67203471e-01 3.94126743e-01 -1.35668471e-01 -4.21199024e-01
-9.20420687e-04 4.54255700e-01 5.31413496e-01 2.27938533e+00
9.83193889e-02 1.84535652e-01 3.69986296e-02 2.05175787e-01
-3.72739345e-01 5.78923285e-01 -9.15769100e-01 -7.38983691e-01
3.67467135e-01 1.05593252e+00 -4.48985368e-01 -6.01237357e-01
-2.38799825e-01 4.86007839e-01 5.86505771e-01 2.10302144e-01
-3.81875813e-01 -7.44520545e-01 9.71348763e-01 -4.00438070e-01
4.19729203e-01 -1.74752548e-02 -1.59688979e-01 -9.83010232e-01
-6.27202511e-01 -8.58070672e-01 4.78629977e-01 -4.11613911e-01
-1.70189166e+00 9.64144588e-01 -2.41312191e-01 -3.66166174e-01
-2.16770306e-01 -1.22799873e+00 2.34290615e-01 8.28154206e-01
-1.86477184e+00 -1.00623035e+00 3.44179459e-02 3.52549046e-01
1.24059334e-01 -5.00967264e-01 1.79085445e+00 3.26317489e-01
-6.21085644e-01 2.99533635e-01 3.17274928e-01 6.78920820e-02
8.85146379e-01 -1.05591083e+00 9.34477448e-01 2.48711273e-01
-3.09924874e-02 1.06962001e+00 6.63522124e-01 -8.92595589e-01
-1.46922386e+00 -1.45189548e+00 1.78135848e+00 -5.48673630e-01
1.19693339e+00 -4.05692488e-01 -6.63797557e-01 6.28114283e-01
4.57302243e-01 4.44420846e-04 1.23655200e+00 -3.92661951e-02
-6.68545008e-01 3.86025488e-01 -1.17184544e+00 3.58977318e-01
1.04787004e+00 -5.99487603e-01 -9.62403059e-01 9.34364617e-01
9.42238152e-01 -2.95126528e-01 -1.46336055e+00 7.74107695e-01
6.83587253e-01 1.91366635e-02 1.32030237e+00 -1.33168685e+00
3.69658232e-01 -4.34323549e-02 -2.68575311e-01 -1.05576444e+00
-5.96431255e-01 -5.43711901e-01 -2.38025025e-01 1.02581286e+00
6.43994510e-01 -8.41767371e-01 1.91772759e-01 -1.63572490e-01
1.17692485e-01 -7.99699366e-01 -9.37342286e-01 -8.65233004e-01
8.02101120e-02 -1.33368313e-01 9.06081676e-01 1.21500814e+00
3.38201374e-01 6.52916849e-01 -7.84939826e-02 4.57273871e-02
2.68065840e-01 -2.52853036e-01 5.03101468e-01 -1.20056283e+00
-1.10938519e-01 -2.59679049e-01 -6.64746821e-01 -4.46532011e-01
1.09231837e-01 -9.14761722e-01 2.07103088e-01 -1.90308642e+00
3.00063848e-01 2.10136876e-01 -9.31352735e-01 6.86496556e-01
-2.33372018e-01 1.62182927e-01 -4.33251381e-01 -8.62436146e-02
-6.25776529e-01 -6.68302923e-03 4.90522027e-01 -2.86368042e-01
9.19702798e-02 -7.60429382e-01 -7.21108079e-01 6.02881491e-01
8.46568108e-01 -8.25133622e-01 3.40991646e-01 -3.43250722e-01
4.27723348e-01 -4.61843938e-01 -1.54324900e-02 -4.56001699e-01
1.29613757e-01 -1.78332955e-01 3.17836732e-01 -3.64279062e-01
2.82903194e-01 -5.18686533e-01 1.81369752e-01 8.84949863e-01
-4.64136243e-01 3.33319843e-01 5.26136458e-01 6.18575692e-01
-5.49959578e-02 -4.47526686e-02 3.93528581e-01 -2.27806345e-01
-4.26790267e-01 1.06768921e-01 -7.62257695e-01 -7.22261071e-02
8.29015136e-01 2.44254082e-01 -8.55798721e-01 5.60619652e-01
-5.52793503e-01 -9.31497216e-02 1.44327998e-01 3.75825495e-01
5.32799482e-01 -8.99155319e-01 -9.62972045e-01 1.49669945e-01
6.59385681e-01 -3.66251767e-01 -7.95324862e-01 6.09347045e-01
-7.96724260e-01 9.85543787e-01 -1.61772609e-01 -8.43421891e-02
-1.36157799e+00 6.33499622e-01 2.73283795e-02 -1.01053166e+00
-2.13829368e-01 1.08223093e+00 5.50763905e-02 -5.77081501e-01
2.34202698e-01 -3.06527615e-01 -8.87422323e-01 8.16929042e-02
7.02858686e-01 3.38282734e-01 3.39876920e-01 -6.61384881e-01
-1.00387251e+00 2.44968459e-01 -1.41209111e-01 2.97953516e-01
1.86344004e+00 3.77635628e-01 -6.83503032e-01 6.56826317e-01
1.20478344e+00 -2.11766995e-02 -7.46644735e-02 -2.23247483e-01
6.83880687e-01 1.51093647e-01 -3.08043629e-01 -1.14141989e+00
-8.46344680e-02 6.60092294e-01 4.27743286e-01 1.84217617e-01
4.14963424e-01 5.67090474e-02 7.07810879e-01 1.08853102e+00
4.36267704e-01 -8.22208107e-01 -4.35840458e-01 9.37357545e-01
8.71980488e-01 -1.03251100e+00 4.40388143e-01 -5.68255723e-01
9.60193500e-02 1.18745589e+00 1.79424867e-01 9.00470689e-02
8.58359635e-01 7.08548784e-01 1.04120269e-03 -5.64103127e-01
-1.12684774e+00 -2.93902397e-01 3.84008795e-01 8.57405722e-01
1.20174026e+00 2.44711861e-01 -6.72327936e-01 6.34696722e-01
1.29981525e-02 -5.52669093e-02 6.96958378e-02 1.25253356e+00
-4.16018099e-01 -1.33733964e+00 2.25407295e-02 3.48545343e-01
-1.09276760e+00 -8.04130733e-01 -9.96874034e-01 7.29267597e-01
1.39022097e-01 1.04133797e+00 -4.19903815e-01 -3.43167663e-01
5.57495117e-01 3.80554646e-01 4.76406276e-01 -8.33467782e-01
-1.02540851e+00 -2.44341969e-01 6.54289782e-01 -8.14578712e-01
-7.43583500e-01 -6.02090776e-01 -1.32234800e+00 -1.80522949e-01
-4.27455813e-01 2.65626878e-01 1.00007045e+00 8.24724853e-01
6.90635026e-01 9.17880058e-01 -1.41734481e-01 -4.94385123e-01
-2.29290187e-01 -1.23677838e+00 -1.94589958e-01 1.85086295e-01
2.18871444e-01 -5.27246296e-01 5.20657040e-02 3.18047523e-01] | [8.490418434143066, 8.764745712280273] |
1e5238fd-fb55-481f-a045-61bf7b9b2b96 | a-fully-convolutional-neural-network-for-1 | 1604.00494 | null | http://arxiv.org/abs/1604.00494v3 | http://arxiv.org/pdf/1604.00494v3.pdf | A Fully Convolutional Neural Network for Cardiac Segmentation in Short-Axis MRI | Automated cardiac segmentation from magnetic resonance imaging datasets is an
essential step in the timely diagnosis and management of cardiac pathologies.
We propose to tackle the problem of automated left and right ventricle
segmentation through the application of a deep fully convolutional neural
network architecture. Our model is efficiently trained end-to-end in a single
learning stage from whole-image inputs and ground truths to make inference at
every pixel. To our knowledge, this is the first application of a fully
convolutional neural network architecture for pixel-wise labeling in cardiac
magnetic resonance imaging. Numerical experiments demonstrate that our model is
robust to outperform previous fully automated methods across multiple
evaluation measures on a range of cardiac datasets. Moreover, our model is fast
and can leverage commodity compute resources such as the graphics processing
unit to enable state-of-the-art cardiac segmentation at massive scales. The
models and code are available at
https://github.com/vuptran/cardiac-segmentation | ['Phi Vu Tran'] | 2016-04-02 | null | null | null | null | ['cardiac-segmentation'] | ['medical'] | [ 1.92562565e-01 1.26496628e-01 4.90221158e-02 -5.76486409e-01
-9.22894359e-01 -4.47259098e-01 -7.13545904e-02 2.14000568e-01
-6.21306717e-01 4.89525378e-01 -2.56606877e-01 -7.27025330e-01
3.22117805e-01 -4.10436720e-01 -4.41708118e-01 -3.96947801e-01
-3.04672211e-01 9.13761497e-01 2.05190063e-01 2.16077045e-01
-1.94969680e-02 7.29450345e-01 -6.26272142e-01 1.51309639e-01
4.86488700e-01 1.01721084e+00 -4.50504087e-02 1.17060935e+00
2.66659468e-01 8.43562961e-01 2.55897772e-02 6.30366355e-02
3.59404713e-01 -4.09295946e-01 -1.06620431e+00 3.20680618e-01
3.28540504e-01 -7.75562108e-01 -1.53445795e-01 7.79070199e-01
9.29612577e-01 -3.57440293e-01 2.95479894e-01 -7.35550165e-01
-1.48514047e-01 5.83986700e-01 -4.76256281e-01 8.79053116e-01
-5.27189612e-01 3.53383392e-01 6.21733725e-01 -9.09551859e-01
5.87550104e-01 5.51799297e-01 9.48435128e-01 3.82236302e-01
-1.22316337e+00 -3.48144770e-01 -3.31188589e-01 -3.26393455e-01
-1.17362738e+00 -3.41722041e-01 5.18981516e-01 -8.13310921e-01
7.11454988e-01 -1.95719838e-01 7.44599819e-01 2.58232683e-01
1.35652617e-01 7.34048367e-01 1.00269210e+00 -2.68340796e-01
3.55726667e-02 -5.18130660e-01 3.59191746e-01 1.23925173e+00
2.80022085e-01 -1.23639405e-01 1.10072382e-02 -2.05441475e-01
1.28650093e+00 -3.37048178e-03 -1.03729993e-01 -4.94150072e-01
-1.56873918e+00 7.57811606e-01 4.05589342e-01 1.83884591e-01
-6.03329301e-01 6.55833244e-01 6.41943514e-01 -5.23690991e-02
6.48230910e-01 1.35207459e-01 -6.76754475e-01 -1.09951511e-01
-1.13730526e+00 9.96743962e-02 5.86121619e-01 5.07554293e-01
3.65210354e-01 3.73762362e-02 -1.14620559e-01 6.50580704e-01
1.48677707e-01 3.99581552e-01 2.48384804e-01 -1.42638743e+00
5.35063855e-02 2.99282789e-01 -2.66186923e-01 -4.59478766e-01
-1.02851939e+00 -6.24599695e-01 -8.89285564e-01 3.83672446e-01
7.13641346e-01 -8.14980805e-01 -1.12678099e+00 1.17345726e+00
6.35697722e-01 3.30854863e-01 -4.30943489e-01 1.33658969e+00
9.21953142e-01 5.75309023e-02 1.85261846e-01 8.03254619e-02
1.52586508e+00 -8.94426644e-01 -2.46682003e-01 -8.50814506e-02
1.06044972e+00 -5.71993589e-01 6.31945491e-01 1.92678839e-01
-1.29932392e+00 -2.18052268e-01 -7.92051136e-01 -2.03669712e-01
2.36005396e-01 3.39032531e-01 7.20353305e-01 6.22549176e-01
-1.24859941e+00 7.93464720e-01 -1.53425539e+00 -3.51924226e-02
1.10593462e+00 3.76375258e-01 -8.97795483e-02 4.70038727e-02
-7.13132799e-01 5.31593502e-01 2.76915103e-01 2.70859748e-01
-9.33077455e-01 -1.14649200e+00 -5.16638398e-01 -2.19979420e-01
3.45550805e-01 -9.98031795e-01 1.61193264e+00 -9.29361403e-01
-1.15842092e+00 1.30675447e+00 1.05173424e-01 -6.78185999e-01
8.30810368e-01 -7.26923496e-02 1.86671987e-01 7.53679216e-01
1.42565250e-01 8.10938537e-01 5.65231025e-01 -8.95944476e-01
-4.18094128e-01 -3.24879080e-01 -3.47561598e-01 -4.43296395e-02
3.28004181e-01 1.95268601e-01 -3.89852375e-01 -7.24060297e-01
2.96014786e-01 -1.04330158e+00 -7.04180539e-01 3.88153493e-01
-2.51304418e-01 2.53420889e-01 5.18882215e-01 -9.60657597e-01
6.92585289e-01 -1.71323073e+00 -2.18185991e-01 1.50922686e-01
7.85406351e-01 3.80739331e-01 2.91816652e-01 -1.72137380e-01
-2.17203781e-01 1.42446384e-01 -7.01875210e-01 -2.24942490e-01
-3.51007402e-01 7.42671490e-02 1.64691299e-01 7.40828395e-01
9.81463045e-02 1.30587351e+00 -7.51065791e-01 -8.43336642e-01
3.54133338e-01 6.58331633e-01 -6.21825039e-01 2.40111724e-02
5.14534079e-02 1.04106295e+00 -4.29109693e-01 6.58612072e-01
3.76893938e-01 -7.58818746e-01 2.75674939e-01 2.33077053e-02
8.72846395e-02 -1.08942285e-03 -7.05783069e-01 1.99615109e+00
-1.84894174e-01 3.85765046e-01 4.16759670e-01 -1.30762327e+00
4.04755741e-01 6.48594260e-01 1.08747077e+00 -4.83977795e-01
4.64442730e-01 4.89503890e-01 2.39142925e-01 -3.71624202e-01
-7.62046203e-02 -2.99322128e-01 1.52592391e-01 7.81879723e-01
1.00054264e-01 -1.31141350e-01 3.22569340e-01 1.85150996e-01
1.24880993e+00 1.11500047e-01 4.57237149e-03 -3.90971988e-01
3.55768114e-01 3.26148540e-01 6.91723466e-01 7.84825385e-01
-5.09223402e-01 1.05359113e+00 5.57043076e-01 -1.12623477e+00
-1.39932311e+00 -9.48959768e-01 -4.30391341e-01 8.88673246e-01
-3.79867315e-01 7.95562714e-02 -1.04345059e+00 -6.85780704e-01
-2.23259702e-01 1.76356416e-02 -3.60146940e-01 6.40906096e-01
-1.01840460e+00 -8.87262881e-01 7.02042341e-01 9.28676724e-01
5.05187511e-01 -1.15011990e+00 -1.35874200e+00 5.05721092e-01
-1.52525261e-01 -1.34404254e+00 -4.04489398e-01 5.62405102e-02
-1.33069730e+00 -1.38572276e+00 -1.02010894e+00 -8.06730509e-01
8.46665978e-01 -2.96195209e-01 1.50627518e+00 6.43058181e-01
-1.03514493e+00 2.73371130e-01 -3.98664661e-02 -3.21683168e-01
-3.93698484e-01 2.15616599e-01 -5.14289021e-01 -4.08012271e-01
1.82985421e-02 -4.19587284e-01 -1.14818192e+00 -1.97991524e-02
-5.80721915e-01 3.34186852e-01 4.39891994e-01 7.50557721e-01
9.43846345e-01 -4.34564888e-01 3.67530942e-01 -1.32232618e+00
3.25829834e-01 -3.64183664e-01 -6.39902294e-01 -1.40872151e-01
-4.54779059e-01 -1.95939407e-01 2.87276953e-01 1.98807061e-01
-5.63057721e-01 4.89994526e-01 -3.72204453e-01 -4.69110489e-01
-5.62554598e-01 5.20613194e-01 8.04839373e-01 -8.80809054e-02
4.84599411e-01 -4.49293591e-02 2.23916873e-01 -3.18033606e-01
3.32227468e-01 2.11348161e-01 7.41493285e-01 -5.10241866e-01
2.60303289e-01 8.06103885e-01 4.08603400e-01 -4.29702848e-01
-7.36635864e-01 -4.55465555e-01 -1.15332806e+00 -4.04488295e-01
1.22586894e+00 -8.25686634e-01 -3.49269658e-01 5.42344689e-01
-1.09678447e+00 -7.34991491e-01 -2.72145331e-01 4.01804775e-01
-7.60289609e-01 3.94788444e-01 -1.07260466e+00 -2.52995938e-01
-9.65130210e-01 -1.23482096e+00 1.14990735e+00 -7.43172839e-02
-1.65901199e-01 -1.26107252e+00 -3.05102766e-02 4.56864297e-01
4.85186070e-01 6.87433958e-01 7.72290289e-01 -6.56490386e-01
-6.29760623e-01 -2.38568336e-01 -3.99317384e-01 4.44703430e-01
-5.07251211e-02 -1.94156915e-01 -7.06223726e-01 -3.15779716e-01
-2.63410002e-01 -4.00914639e-01 8.54551435e-01 1.10774040e+00
1.43030107e+00 5.68410568e-02 -2.03703001e-01 1.00938582e+00
1.28576159e+00 -1.43894836e-01 2.26347551e-01 3.37155089e-02
9.10616517e-01 1.86183199e-01 2.20041826e-01 4.97582823e-01
4.28759694e-01 5.04606776e-02 2.51523942e-01 -9.33502138e-01
-2.13634431e-01 4.33414757e-01 -5.69555342e-01 8.59162509e-01
-1.08831719e-01 4.31045204e-01 -1.61844385e+00 7.17155159e-01
-1.69483507e+00 -5.98683655e-01 -3.28653485e-01 1.85640264e+00
7.60677755e-01 1.43942207e-01 2.12678820e-01 -2.13118330e-01
6.77048922e-01 -1.13410607e-01 -6.24914348e-01 -2.38062561e-01
4.75129217e-01 6.49792790e-01 7.16033757e-01 4.73855525e-01
-1.59793425e+00 6.88702941e-01 6.45566082e+00 2.66778946e-01
-1.24234235e+00 4.41635251e-01 1.40310943e+00 -1.63043052e-01
3.56801391e-01 -1.13985665e-01 -2.91187942e-01 2.17690557e-01
7.79778242e-01 1.67879075e-01 1.96612239e-01 7.11240172e-01
3.21083367e-01 -2.36444995e-02 -8.43594849e-01 7.68775582e-01
-1.75486892e-01 -1.73055494e+00 -5.50223291e-01 -5.27243838e-02
7.39580750e-01 7.20008671e-01 -3.12434524e-01 -1.99291587e-01
2.18299791e-01 -1.09200692e+00 3.33989352e-01 4.72695738e-01
9.69423234e-01 -7.23865151e-01 8.23805571e-01 1.78390712e-01
-9.94293153e-01 3.38807285e-01 1.18259549e-01 1.75625876e-01
4.17903185e-01 7.72455990e-01 -1.03882110e+00 1.67385831e-01
7.02464283e-01 5.73490918e-01 -4.49666917e-01 1.00784743e+00
4.09216955e-02 9.29078758e-01 -2.85937876e-01 6.72004104e-01
2.40530625e-01 -1.16511201e-02 3.26235205e-01 1.26550210e+00
1.06748663e-01 4.36091840e-01 5.60225904e-01 1.00766969e+00
-1.98437437e-01 9.55754966e-02 -2.64334798e-01 6.98171556e-02
1.50986277e-02 1.62813735e+00 -1.45560682e+00 -7.41833150e-01
-3.73525143e-01 6.13895893e-01 1.77836925e-01 1.71721831e-01
-8.22691500e-01 -1.71684191e-01 9.35023874e-02 4.45586354e-01
3.63821685e-01 -3.49358529e-01 -9.12767768e-01 -9.88855124e-01
-1.06438674e-01 -6.60407603e-01 5.31723440e-01 -6.40452743e-01
-1.07356799e+00 5.54288685e-01 -2.27730766e-01 -7.98112035e-01
-2.81985611e-01 -5.94581425e-01 -6.07863784e-01 9.50461090e-01
-1.65470421e+00 -1.07017481e+00 -4.13070023e-01 4.96592432e-01
3.22368830e-01 8.76095966e-02 7.47020721e-01 3.52515250e-01
-3.89207631e-01 7.73286074e-02 -2.51907200e-01 8.41426313e-01
3.33970368e-01 -1.43797457e+00 6.99045956e-01 1.01674247e+00
-1.73872545e-01 3.44759613e-01 1.74691007e-01 -5.88693500e-01
-9.14203703e-01 -1.16915917e+00 4.99352664e-01 -3.22049528e-01
5.53310692e-01 1.30299285e-01 -8.17002773e-01 7.48445272e-01
1.36299551e-01 8.78023803e-01 6.67369545e-01 -3.10928345e-01
2.44441763e-01 3.57070267e-01 -1.05428016e+00 2.47852713e-01
7.06420839e-01 -3.29881161e-01 -2.47444838e-01 5.99672675e-01
4.80063707e-01 -1.05003202e+00 -1.15462399e+00 5.94981611e-01
3.48455191e-01 -8.69045734e-01 1.05586278e+00 -4.05226380e-01
6.82936728e-01 -2.47269884e-01 4.26906735e-01 -7.70076513e-01
-1.16647750e-01 -4.68175054e-01 -5.65789640e-02 3.20526898e-01
4.39089179e-01 -5.26012778e-01 9.09981728e-01 6.89083219e-01
-3.45991582e-01 -1.16327012e+00 -9.77159560e-01 -9.50657055e-02
4.90189016e-01 -4.89672631e-01 7.44585693e-02 9.35707867e-01
-5.30281246e-01 -1.26866519e-01 1.33906528e-01 1.99985087e-01
1.00650895e+00 2.08072841e-01 3.96418095e-01 -1.20591342e+00
-3.08043897e-01 -4.39035714e-01 -3.03422332e-01 -5.55790424e-01
-3.49652730e-02 -1.16068113e+00 -1.14774674e-01 -1.69235182e+00
2.57696390e-01 -5.59284627e-01 -4.04894024e-01 7.32258201e-01
-2.00945944e-01 8.56991231e-01 1.46355197e-01 2.59538472e-01
-4.88341063e-01 -2.11649939e-01 1.43141782e+00 1.70838997e-01
8.03127214e-02 -1.59093469e-01 -4.55309927e-01 9.57016289e-01
1.07093525e+00 -5.42311966e-01 -1.16487868e-01 -5.56229711e-01
-1.73260689e-01 4.48435456e-01 8.34387898e-01 -1.10311580e+00
1.66023433e-01 3.90561432e-01 5.85155904e-01 -5.83107471e-01
-1.25710323e-01 -5.47818124e-01 -3.19029152e-01 8.38980138e-01
-4.24505651e-01 1.81225687e-01 2.77132392e-01 -2.55381055e-02
1.22059233e-01 1.85328051e-02 1.03515446e+00 -5.14741957e-01
-5.40262997e-01 6.81024432e-01 -3.12793255e-01 5.74878156e-01
9.37315106e-01 4.58062142e-02 -3.67294252e-02 1.95683613e-02
-1.06765914e+00 2.61556327e-01 1.41646162e-01 -1.28496587e-01
4.40990627e-01 -9.32703435e-01 -1.07516587e+00 4.42983136e-02
-3.14718217e-01 4.22988415e-01 3.24215710e-01 1.36993742e+00
-1.52334106e+00 3.74590755e-01 -1.80311248e-01 -1.16772377e+00
-1.15165830e+00 1.61672950e-01 9.71171796e-01 -3.83238137e-01
-1.18855548e+00 6.88308775e-01 4.01948318e-02 -6.21285737e-01
-9.59852114e-02 -5.70888221e-01 3.02179247e-01 -5.21030068e-01
4.22879100e-01 1.28890827e-01 2.22779617e-01 -5.76099932e-01
-4.42362547e-01 4.59466428e-01 1.94744207e-03 -1.12276167e-01
1.37308276e+00 -2.89261807e-02 -1.72752544e-01 1.48911044e-01
1.02860105e+00 -6.49988949e-01 -1.46083856e+00 -2.32812345e-01
-1.16714962e-01 -8.04363862e-02 5.59998214e-01 -9.85367000e-01
-1.66755021e+00 1.14577913e+00 8.17559838e-01 -2.51610965e-01
9.83376801e-01 -1.09964646e-01 1.05039978e+00 1.87432200e-01
2.17278320e-02 -1.09736872e+00 -2.26059720e-01 2.66397595e-01
4.00094360e-01 -1.35978997e+00 1.40411109e-01 -3.23461771e-01
-8.24592471e-01 1.19048357e+00 3.95743072e-01 -3.68110627e-01
8.36377978e-01 7.06544757e-01 6.27361774e-01 -4.47573215e-01
-5.42958319e-01 -1.47930488e-01 1.04867168e-01 3.01418602e-01
9.43113208e-01 1.88532144e-01 -1.86353669e-01 5.93048185e-02
1.85433760e-01 5.14325321e-01 4.07213956e-01 1.16267586e+00
-3.00531089e-01 -8.79802585e-01 -1.27517208e-01 5.38428545e-01
-1.09565651e+00 -3.49498451e-01 1.35301292e-01 5.87888956e-01
-2.71783881e-02 4.50532347e-01 1.14551708e-01 4.80267107e-01
5.71361035e-02 2.14535996e-01 3.89348030e-01 -6.22987628e-01
-7.83413827e-01 2.87492573e-01 -1.12655610e-01 -6.65036500e-01
-3.89369190e-01 -6.86700702e-01 -1.87539780e+00 -1.72245085e-01
2.04613253e-01 -4.12318289e-01 7.06381977e-01 9.57860410e-01
4.87746865e-01 9.73425865e-01 3.33522260e-01 -1.02419996e+00
-3.32036048e-01 -6.41417086e-01 -4.16891903e-01 4.68235999e-01
3.05638462e-01 -1.11909933e-01 1.46368623e-01 3.95463318e-01] | [14.277265548706055, -2.451770544052124] |
bf6f28c7-d844-440c-a28f-d19b87bbfea6 | 3dcapsule-extending-the-capsule-architecture | 1811.02191 | null | http://arxiv.org/abs/1811.02191v1 | http://arxiv.org/pdf/1811.02191v1.pdf | 3DCapsule: Extending the Capsule Architecture to Classify 3D Point Clouds | This paper introduces the 3DCapsule, which is a 3D extension of the recently
introduced Capsule concept that makes it applicable to unordered point sets.
The original Capsule relies on the existence of a spatial relationship between
the elements in the feature map it is presented with, whereas in point
permutation invariant formulations of 3D point set classification methods, such
relationships are typically lost. Here, a new layer called ComposeCaps is
introduced that, in lieu of a spatially relevant feature mapping, learns a new
mapping that can be exploited by the 3DCapsule. Previous works in the 3D point
set classification domain have focused on other parts of the architecture,
whereas instead, the 3DCapsule is a drop-in replacement of the commonly used
fully connected classifier. It is demonstrated via an ablation study, that when
the 3DCapsule is applied to recent 3D point set classification architectures,
it consistently shows an improvement, in particular when subjected to noisy
data. Similarly, the ComposeCaps layer is evaluated and demonstrates an
improvement over the baseline. In an apples-to-apples comparison against
state-of-the-art methods, again, better performance is demonstrated by the
3DCapsule. | ['Lars Petersson', 'Ali Cheraghian'] | 2018-11-06 | null | null | null | null | ['classify-3d-point-clouds'] | ['computer-vision'] | [-9.55423340e-02 -1.69763267e-02 -1.98964924e-02 -2.02722281e-01
-6.23608589e-01 -9.35769975e-01 9.75098670e-01 2.87219644e-01
-2.56001502e-01 3.12290728e-01 1.49410218e-01 2.23495904e-02
-6.00539863e-01 -4.71155107e-01 -9.91096199e-01 -7.17633307e-01
-5.13054252e-01 5.83609223e-01 3.27191502e-01 -2.92801589e-01
3.08461070e-01 9.17883158e-01 -1.60423040e+00 2.40972653e-01
2.29007453e-01 1.18330646e+00 1.66046783e-01 1.22529216e-01
-1.36694480e-02 1.62072390e-01 -5.67589521e-01 1.60216376e-01
6.54058516e-01 9.35247913e-02 -7.40671039e-01 8.83122981e-02
8.68499577e-01 1.43288538e-01 -2.60504093e-02 5.22767425e-01
3.48229557e-01 -4.90238257e-02 8.21203470e-01 -1.30750310e+00
-4.84919280e-01 1.46997228e-01 -2.36760005e-01 2.15742830e-02
5.76947033e-01 -2.06522435e-01 8.16636682e-01 -1.20555246e+00
9.90455210e-01 1.25459504e+00 1.11548626e+00 3.13355446e-01
-1.38576162e+00 -3.53176504e-01 2.35991463e-01 2.43450589e-02
-1.39783931e+00 -2.79497374e-02 9.42491651e-01 -4.13757473e-01
1.23911703e+00 4.70198661e-01 9.26694930e-01 1.01451063e+00
3.24965149e-01 8.39716256e-01 1.22651470e+00 -4.42809284e-01
4.50070560e-01 8.49941000e-02 1.99961454e-01 2.18520388e-01
2.32610866e-01 3.84027183e-01 -6.33100212e-01 -1.09343939e-01
6.97391689e-01 3.82865667e-02 -3.35137904e-01 -1.36844456e+00
-1.29864192e+00 6.36109054e-01 8.80925894e-01 5.05745649e-01
-7.64860958e-02 6.96484372e-02 2.32708767e-01 3.61971915e-01
4.66883361e-01 8.40734124e-01 -3.84943038e-01 -8.99327267e-03
-1.08841181e+00 5.34677804e-01 9.35355902e-01 1.24798572e+00
5.61707795e-01 -5.10534227e-01 -5.06263748e-02 4.25780416e-01
3.06147039e-01 2.51576990e-01 2.26503551e-01 -6.82463109e-01
2.76601613e-01 9.56369460e-01 -1.34658724e-01 -9.27822292e-01
-6.33491874e-01 -7.39955306e-01 -4.35017645e-01 7.93890655e-01
7.26202577e-02 4.63044345e-01 -9.47942257e-01 1.39453959e+00
2.76215434e-01 1.71507895e-01 -8.69391039e-02 8.99722517e-01
8.39229167e-01 1.97823748e-01 -3.37848246e-01 3.89580667e-01
1.10758793e+00 -7.76747465e-01 -9.99583155e-02 -1.77443754e-02
4.94207948e-01 -5.88502467e-01 8.68021965e-01 3.54143232e-01
-9.52827036e-01 -5.74238181e-01 -1.50442171e+00 -1.42886624e-01
-9.86972749e-01 -2.95806766e-01 5.15665412e-01 4.15381283e-01
-1.30164480e+00 8.18540096e-01 -7.26474583e-01 -6.36462152e-01
6.98891521e-01 5.94917238e-01 -8.62682998e-01 2.07304042e-02
-6.68995142e-01 1.19519567e+00 3.31161737e-01 1.21365581e-02
-6.19632125e-01 -1.07742333e+00 -8.73081028e-01 -2.42440011e-02
2.16459408e-01 -5.79210341e-01 9.06367362e-01 -3.61052603e-01
-1.28676307e+00 1.11344516e+00 1.84533596e-01 -3.89370412e-01
5.34811616e-01 -3.08281332e-01 -5.62508665e-02 3.01027536e-01
2.09555745e-01 8.66837084e-01 7.72354305e-01 -1.48861802e+00
-5.05284131e-01 -4.08132434e-01 1.08277380e-01 2.67332107e-01
-1.19454395e-02 -3.75797689e-01 -5.37016332e-01 -6.76283598e-01
4.43865210e-01 -1.15857613e+00 -2.53707618e-02 3.48304838e-01
-3.69563133e-01 -4.65671599e-01 1.16012549e+00 -2.80076284e-02
6.17528737e-01 -2.39400506e+00 4.93967026e-01 5.79948604e-01
2.33920649e-01 5.68999723e-02 -4.32487093e-02 6.16919637e-01
-4.91196364e-01 2.08368450e-01 -5.76960325e-01 -5.03379285e-01
2.23361924e-01 8.21042284e-02 -1.15744822e-01 7.55837679e-01
4.52015042e-01 8.85703683e-01 -7.52661884e-01 -1.66963771e-01
5.55198193e-01 5.43251634e-01 -5.46516716e-01 -2.67443985e-01
-1.56047903e-02 2.81011164e-01 -3.59370381e-01 6.64009929e-01
9.38399792e-01 -1.95792690e-01 -3.44965965e-01 -2.18446791e-01
-1.63307309e-01 1.90573841e-01 -9.75853622e-01 2.50776911e+00
-1.88731611e-01 5.50768197e-01 -1.33222312e-01 -8.96963000e-01
1.01963460e+00 2.67857939e-01 7.26801634e-01 -3.79296303e-01
-5.29359467e-02 3.36520016e-01 -6.33589774e-02 -1.14478379e-01
2.57967442e-01 1.10919505e-01 -3.61325949e-01 1.88941479e-01
2.11956248e-01 -5.69295645e-01 -2.87097469e-02 1.15540743e-01
1.21336424e+00 7.03793824e-01 1.99776888e-01 -8.10200334e-01
6.08574688e-01 3.14629346e-01 -2.07666438e-02 6.76977813e-01
-8.70243832e-02 1.13855433e+00 3.64090562e-01 -3.83476526e-01
-9.35293496e-01 -1.32257330e+00 -6.61880016e-01 2.33768225e-01
4.15075690e-01 -5.20294249e-01 -4.96340901e-01 -9.12721455e-01
5.87849021e-01 4.77577120e-01 -9.37419116e-01 -2.78588921e-01
-6.03783607e-01 -3.30220431e-01 2.33458936e-01 5.63876629e-01
4.65990782e-01 -7.20117092e-01 -6.89161062e-01 1.26386704e-02
6.07659280e-01 -9.15273011e-01 -1.58821225e-01 4.89268988e-01
-1.06006908e+00 -1.19996727e+00 -6.30787373e-01 -9.73976731e-01
5.88532984e-01 3.03774983e-01 1.30341601e+00 1.92303322e-02
-4.90426198e-02 6.99280918e-01 -5.70489943e-01 -6.34377718e-01
1.18274376e-01 4.12007481e-01 3.48372757e-02 -4.19070184e-01
8.22231412e-01 -6.88097298e-01 -6.02749586e-01 2.08128244e-01
-7.13995457e-01 -3.56704473e-01 5.66993594e-01 7.48724401e-01
6.63416862e-01 -2.96054095e-01 4.23796117e-01 -6.01572096e-01
2.83235550e-01 -3.70962292e-01 -3.23646128e-01 -7.05567375e-02
-3.10209960e-01 5.78672178e-02 3.53111982e-01 -1.52001396e-01
-3.10813636e-01 3.68299335e-01 -1.57560278e-02 -7.30599046e-01
-2.91319221e-01 3.49103004e-01 -1.30967557e-01 -5.10729551e-01
6.11484945e-01 -3.91890667e-02 4.24243659e-01 -8.96785676e-01
4.16835755e-01 5.97868025e-01 5.41412711e-01 -4.03823853e-01
1.14448106e+00 8.69120240e-01 2.71581292e-01 -5.61616004e-01
-4.41698045e-01 -6.26214981e-01 -1.26196814e+00 -2.43986733e-02
7.54837275e-01 -8.62392306e-01 -5.22801757e-01 2.14919820e-01
-8.94216835e-01 3.72587964e-02 -8.31151783e-01 5.01569986e-01
-8.56603622e-01 1.28167748e-01 3.70923132e-02 -3.27546716e-01
8.15048665e-02 -1.17654109e+00 1.45542097e+00 -1.70783162e-01
-3.57155263e-01 -1.01150274e+00 -3.49799320e-02 -2.49359936e-01
1.92205876e-01 6.42429173e-01 9.25569296e-01 -9.46811855e-01
-5.44799805e-01 -7.47545660e-01 1.48526311e-01 3.05371046e-01
9.54198465e-02 -5.27202368e-01 -9.65147138e-01 -5.92842340e-01
1.93741068e-01 -2.08089277e-02 7.84013748e-01 1.34267852e-01
1.08068907e+00 1.79605439e-01 -6.36412680e-01 6.44050419e-01
1.40571511e+00 -4.78656664e-02 5.80180645e-01 6.92225993e-01
3.81585270e-01 4.51962084e-01 3.49816471e-01 -6.99688047e-02
2.58472770e-01 9.88170028e-01 8.59050632e-01 -2.89743900e-01
-3.60050440e-01 -1.58018082e-01 9.93690342e-02 5.57623208e-01
8.98800604e-03 4.31142673e-02 -7.36773193e-01 3.64514172e-01
-1.62926793e+00 -6.96710467e-01 -2.34278035e-03 2.34895873e+00
4.24578488e-01 2.10318327e-01 1.32952213e-01 4.27175313e-01
3.50433826e-01 2.12273121e-01 -3.97505462e-01 -1.35142267e-01
-2.37060919e-01 4.59254891e-01 4.94667768e-01 1.48541912e-01
-1.42553985e+00 5.22611797e-01 6.73075533e+00 6.78469896e-01
-1.12894738e+00 -5.59290051e-02 3.34423804e-03 -4.42407906e-01
-6.87261596e-02 -3.77549455e-02 -7.33890772e-01 4.42518741e-01
3.15514684e-01 1.40074506e-01 1.60344556e-01 9.15411472e-01
-1.42976165e-01 -1.15332482e-02 -1.83727503e+00 1.06037867e+00
5.12444854e-01 -1.46948147e+00 1.59532875e-02 3.15552622e-01
4.06609744e-01 1.02530040e-01 1.60500169e-01 1.66011766e-01
-1.95886135e-01 -1.14642060e+00 1.03437233e+00 5.23050368e-01
4.86682951e-01 -4.98629004e-01 7.82077312e-01 7.67801553e-02
-1.07281673e+00 2.45029689e-03 -2.66179711e-01 5.17294966e-02
-3.59494356e-03 2.78597057e-01 -8.00458908e-01 8.19776416e-01
1.04342794e+00 1.20183802e+00 -9.65871096e-01 1.48408103e+00
-1.15349934e-01 3.75972986e-02 -6.85539782e-01 3.01850699e-02
4.94642943e-01 6.26685396e-02 1.01339066e+00 1.17508566e+00
4.29506153e-01 -3.41895849e-01 1.35243386e-01 9.98291850e-01
3.77598479e-02 -1.85414389e-01 -9.37405944e-01 4.44461793e-01
4.17164147e-01 1.21172273e+00 -7.55882442e-01 -8.95757694e-03
-5.11976004e-01 6.43884957e-01 2.63114363e-01 1.65159047e-01
-4.51345146e-01 -3.93058330e-01 6.78310573e-01 3.80815387e-01
8.54068220e-01 -3.82809758e-01 -4.83138800e-01 -6.33222342e-01
3.03862184e-01 -6.36843562e-01 1.73503101e-01 -8.43977749e-01
-1.48318303e+00 4.74218667e-01 3.18315715e-01 -1.95025384e+00
-1.74191564e-01 -1.03112912e+00 -5.03329039e-01 8.09385359e-01
-1.55225551e+00 -1.14568663e+00 -3.52712572e-01 6.01845026e-01
2.72774965e-01 -2.58614331e-01 9.07435298e-01 7.50325397e-02
1.44235068e-03 4.44091976e-01 1.33843914e-01 -1.68195456e-01
5.29205799e-01 -1.50859237e+00 3.96728396e-01 5.70516765e-01
4.29155141e-01 6.57215953e-01 4.47020054e-01 -4.32935029e-01
-1.52206552e+00 -8.73746514e-01 5.88773370e-01 -9.66662109e-01
2.17988893e-01 -9.33306277e-01 -8.81837130e-01 5.40251315e-01
2.44576141e-01 1.94411457e-01 6.08729780e-01 8.63730982e-02
-4.73079532e-01 -3.20976824e-02 -1.20655596e+00 2.73043871e-01
1.36139262e+00 -4.67366725e-01 -1.03778648e+00 3.79287779e-01
6.28090620e-01 -5.15723705e-01 -1.23864567e+00 5.17359257e-01
3.60210806e-01 -8.67564440e-01 1.28010225e+00 -3.85651529e-01
2.73024321e-01 -6.31751478e-01 -3.05467516e-01 -1.41522741e+00
-4.20196354e-01 -4.30624306e-01 -1.20114371e-01 7.00925231e-01
4.36527550e-01 -3.95111829e-01 9.61259246e-01 1.30651280e-01
-6.84627593e-01 -8.56043339e-01 -1.42585778e+00 -1.17162859e+00
4.85062897e-01 -3.24860752e-01 7.89452493e-01 7.43207872e-01
1.55300006e-01 1.00843593e-01 2.05926299e-01 6.82198703e-02
4.75745916e-01 2.03153864e-01 7.80178785e-01 -1.59489405e+00
-4.94392738e-02 -5.16359985e-01 -1.06601119e+00 -1.09539235e+00
-1.80007860e-01 -1.40942836e+00 -2.63805278e-02 -1.60533917e+00
-1.87899590e-01 -5.48095107e-01 -3.28625560e-01 4.27166283e-01
2.88920909e-01 4.07853931e-01 6.00224078e-01 4.75710392e-01
-5.34938514e-01 4.75389808e-01 1.11718726e+00 -1.96939021e-01
-2.88417041e-01 -3.44830118e-02 -5.89080691e-01 5.08296728e-01
4.63052124e-01 -3.89411986e-01 -1.58822283e-01 -3.80315393e-01
-5.64806908e-02 -7.40289629e-01 8.36174190e-01 -1.39743257e+00
3.03732574e-01 6.82293892e-01 6.80218697e-01 -9.47725534e-01
7.81529009e-01 -1.28686392e+00 1.85643658e-01 3.03926438e-01
4.02218141e-02 1.69332698e-01 4.77675617e-01 6.09722793e-01
-2.62102723e-01 -1.39635980e-01 4.62509364e-01 -2.29678258e-01
-8.65469933e-01 1.97198540e-01 7.96362571e-03 -2.94451118e-01
1.23779321e+00 -8.67336571e-01 -1.72981858e-01 5.59819192e-02
-7.99561739e-01 1.60688370e-01 7.28241563e-01 8.47814679e-01
5.60521066e-01 -1.64562643e+00 -6.21531069e-01 4.83641088e-01
4.41712499e-01 3.76602352e-01 -1.55219451e-01 8.08615148e-01
-4.46941197e-01 5.50455928e-01 -2.14160219e-01 -1.21555853e+00
-9.12211657e-01 6.35994971e-01 4.65889156e-01 9.78979617e-02
-1.09405589e+00 6.92132056e-01 3.89823206e-02 -6.05946660e-01
3.97666276e-01 -6.23213887e-01 -2.73054928e-01 1.19206905e-01
1.86517672e-03 7.95888603e-02 6.33837640e-01 -5.86503744e-01
-7.94742644e-01 9.74423170e-01 -2.17368994e-02 1.33051248e-02
1.61052525e+00 2.27230832e-01 1.12891505e-02 5.24572372e-01
1.39164245e+00 -1.09697193e-01 -1.34704387e+00 -6.90186173e-02
2.98109889e-01 -4.88515556e-01 -1.89680457e-02 -9.13381994e-01
-7.50169039e-01 5.67571223e-01 8.92810106e-01 2.64754981e-01
8.47316742e-01 3.84611279e-01 1.42233521e-01 1.83043897e-01
6.43509805e-01 -7.12630570e-01 -1.82924360e-01 3.21907699e-01
1.41743255e+00 -1.10887849e+00 2.42729560e-01 -5.65093219e-01
-1.86139882e-01 1.08656526e+00 2.18718544e-01 -7.20023990e-01
1.13301027e+00 2.84063250e-01 -1.79886237e-01 -5.84467947e-01
-3.43554497e-01 3.48775573e-02 5.10078132e-01 9.44462240e-01
4.85345013e-02 -2.67614126e-01 -5.64036220e-02 3.64587128e-01
-5.07191360e-01 -1.92075983e-01 6.93408921e-02 1.18505979e+00
-1.04100205e-01 -1.04716194e+00 -4.01336849e-01 4.04562593e-01
1.41830057e-01 9.96179134e-02 -6.55235171e-01 1.54293060e+00
2.52178907e-01 6.57287717e-01 5.11685491e-01 -4.45306569e-01
7.81605363e-01 2.70522870e-02 5.00929356e-01 -6.97161019e-01
-9.89743948e-01 -9.34739411e-02 -2.61377275e-01 -8.38571131e-01
-7.78557241e-01 -8.79479885e-01 -1.03542471e+00 1.14513993e-01
-2.36887291e-01 1.09453164e-01 9.48670685e-01 7.68073440e-01
6.19075358e-01 5.23730338e-01 4.78631467e-01 -1.35275209e+00
-3.59396130e-01 -8.88422191e-01 -5.73504984e-01 4.72778767e-01
4.05164063e-01 -1.09569907e+00 -4.11358267e-01 -3.48190099e-01] | [7.972347259521484, -3.314114570617676] |
f5b3af08-0436-46ed-af22-fea5238e1efc | representation-driven-reinforcement-learning | 2305.19922 | null | https://arxiv.org/abs/2305.19922v2 | https://arxiv.org/pdf/2305.19922v2.pdf | Representation-Driven Reinforcement Learning | We present a representation-driven framework for reinforcement learning. By representing policies as estimates of their expected values, we leverage techniques from contextual bandits to guide exploration and exploitation. Particularly, embedding a policy network into a linear feature space allows us to reframe the exploration-exploitation problem as a representation-exploitation problem, where good policy representations enable optimal exploration. We demonstrate the effectiveness of this framework through its application to evolutionary and policy gradient-based approaches, leading to significantly improved performance compared to traditional methods. Our framework provides a new perspective on reinforcement learning, highlighting the importance of policy representation in determining optimal exploration-exploitation strategies. | ['Shie Mannor', 'Guy Tennenholtz', 'Ofir Nabati'] | 2023-05-31 | null | null | null | null | ['multi-armed-bandits'] | ['miscellaneous'] | [ 1.91654470e-02 1.40222564e-01 -1.22379947e+00 1.24825602e-02
-6.95112765e-01 -6.44943058e-01 7.28376627e-01 -1.25525072e-01
-6.24939322e-01 1.23161030e+00 4.91764635e-01 -7.27027476e-01
-4.82541114e-01 -7.64985740e-01 -5.93605280e-01 -6.71776354e-01
-3.26341867e-01 2.86227971e-01 -2.73740798e-01 -3.19346011e-01
4.90202755e-01 5.03287077e-01 -1.40163231e+00 -1.30632013e-01
1.03481174e+00 1.11105692e+00 -3.23638916e-02 4.58353758e-01
-1.65852442e-01 7.13966787e-01 -4.42105234e-01 -7.41761997e-02
4.14657772e-01 -4.18339223e-01 -6.43093884e-01 -5.80964750e-03
-2.47642756e-01 -4.91850615e-01 -2.75235176e-01 1.18687856e+00
1.95216596e-01 5.79384685e-01 6.68026984e-01 -1.18347514e+00
-8.10537159e-01 6.59298718e-01 -6.32445395e-01 2.72234172e-01
3.89776647e-01 2.30782434e-01 1.23965383e+00 -3.78038973e-01
5.58596909e-01 1.54167116e+00 2.18958139e-01 5.23917198e-01
-1.51203954e+00 -4.44241703e-01 7.98164606e-01 2.72941291e-01
-6.33738339e-01 -2.24567816e-01 7.50613451e-01 -1.85613737e-01
9.97761130e-01 2.50933111e-01 1.26443934e+00 1.21502864e+00
-5.33488318e-02 1.45139456e+00 1.30118370e+00 -6.47021711e-01
9.15026605e-01 1.08624369e-01 -3.98547858e-01 4.84700024e-01
3.34628969e-01 1.13270068e+00 -4.94658440e-01 -4.67938602e-01
9.45689738e-01 -2.16836762e-02 -1.18699066e-01 -9.96028721e-01
-7.10514009e-01 1.42564738e+00 4.73910779e-01 -1.12357207e-01
-6.52474999e-01 5.26144385e-01 2.85669863e-01 5.19644022e-01
4.22055066e-01 9.32599664e-01 -2.63992041e-01 -5.22578895e-01
-6.88601792e-01 7.29951262e-01 8.31421673e-01 5.82470357e-01
8.13855052e-01 4.49613184e-01 -4.16027725e-01 6.60733998e-01
3.99324208e-01 2.79021710e-01 5.30079603e-01 -1.25692534e+00
3.26418281e-01 3.02938908e-01 6.28994048e-01 -4.74075019e-01
-8.72482061e-02 -5.86911917e-01 -5.25017977e-02 6.85581803e-01
1.08286239e-01 -4.68576789e-01 -8.07156563e-01 1.77181494e+00
4.85257596e-01 -8.64020288e-02 1.17931619e-01 6.20589912e-01
-3.23315471e-01 5.50034225e-01 1.68369904e-01 -6.02983892e-01
7.03053951e-01 -1.01299191e+00 -7.43218541e-01 -3.84385616e-01
4.18857723e-01 -8.04789141e-02 9.78232145e-01 5.72557271e-01
-1.12402344e+00 -4.78836335e-02 -9.87198591e-01 7.19931126e-01
-2.87372231e-01 -2.99084604e-01 8.97939563e-01 8.25130999e-01
-1.15010619e+00 8.71429086e-01 -8.15200746e-01 -1.24577552e-01
8.08704972e-01 1.85981184e-01 2.23922372e-01 1.92272365e-01
-1.02936983e+00 9.81804073e-01 8.44393790e-01 -1.35450825e-01
-1.07728636e+00 -5.21261454e-01 -7.74807215e-01 1.87983558e-01
8.89477432e-01 -4.47193921e-01 1.58398402e+00 -1.10264361e+00
-1.98289001e+00 4.00916487e-02 2.06074659e-02 -8.81836414e-01
7.47224212e-01 -3.18175346e-01 -3.48575532e-01 5.42986318e-02
-9.14640576e-02 4.28945422e-01 1.05475712e+00 -1.09525168e+00
-9.86397862e-01 -8.56528431e-02 2.59741992e-01 5.07198393e-01
-4.35697138e-01 -2.18548387e-01 -5.12057776e-03 -8.93800676e-01
-3.35136354e-01 -7.90840209e-01 -8.11658382e-01 -1.15826353e-01
-2.65546471e-01 -2.01827452e-01 6.22924089e-01 -1.60042971e-01
1.41001487e+00 -1.81623149e+00 1.65167451e-01 5.91062605e-01
-2.63626105e-04 1.41241938e-01 -1.91716865e-01 5.40029049e-01
1.35479048e-01 2.32015073e-01 -1.16640732e-01 1.01911865e-01
4.31946933e-01 4.78858382e-01 -8.21202874e-01 4.37347770e-01
1.45590633e-01 1.18782842e+00 -1.23337471e+00 -4.61039692e-02
2.67087579e-01 -9.88921225e-02 -8.11330736e-01 3.39993894e-01
-6.41566396e-01 2.70041943e-01 -9.71392810e-01 8.36541057e-01
1.89985961e-01 -6.46038651e-02 5.50939202e-01 6.85794413e-01
-2.18903512e-01 2.79421687e-01 -9.99509692e-01 1.40994918e+00
-3.71650845e-01 2.57681757e-01 -6.45691296e-03 -1.33663630e+00
9.38292086e-01 6.06022077e-03 6.15722835e-01 -8.11643898e-01
-1.52001917e-01 2.92037427e-01 -2.82070130e-01 -1.57044247e-01
4.92168576e-01 -9.32429209e-02 -6.12021843e-03 9.28705812e-01
-1.11438230e-01 -9.77332741e-02 3.48016292e-01 -9.93513018e-02
7.12274551e-01 6.85584724e-01 8.13407242e-01 -4.41558063e-01
1.78583357e-02 2.94667538e-02 5.84248424e-01 1.36923265e+00
-3.37534845e-01 -2.62398988e-01 7.42823958e-01 -5.86548686e-01
-8.13693345e-01 -1.06224430e+00 -9.23342351e-03 1.41649270e+00
-1.10173196e-01 -2.97356158e-01 -2.74508327e-01 -1.06732142e+00
6.92891002e-01 8.40044200e-01 -1.08796847e+00 -2.76475370e-01
-5.79811037e-01 -7.51062512e-01 7.60234669e-02 6.47996724e-01
2.21759528e-01 -1.05173218e+00 -1.07857835e+00 4.66966003e-01
4.77355778e-01 -2.96042442e-01 -2.53672302e-01 4.04547483e-01
-1.04325604e+00 -1.02252042e+00 -9.01470959e-01 -9.35914889e-02
3.43292445e-01 -1.02998674e-01 1.03972542e+00 -2.62702137e-01
-6.72366023e-02 6.51471198e-01 -1.68877587e-01 -4.96028602e-01
-3.26043576e-01 -3.39548737e-02 2.35087276e-01 -2.79636353e-01
1.67275921e-01 -6.57275856e-01 -8.92607987e-01 4.32506800e-02
-5.51261365e-01 -4.49262738e-01 4.44376856e-01 1.12138188e+00
4.57429081e-01 -3.82936627e-01 7.22128630e-01 -7.08828747e-01
1.34102798e+00 -7.82514393e-01 -1.06541073e+00 4.35638309e-01
-1.18502486e+00 5.47761977e-01 4.52415913e-01 -6.07113063e-01
-1.06156230e+00 -3.04914802e-01 2.65238792e-01 -4.81760591e-01
2.38156125e-01 4.08559889e-01 4.08273339e-01 -1.73621684e-01
7.57370830e-01 2.81200558e-01 2.04099000e-01 -4.75635260e-01
8.65873098e-01 1.88595697e-01 1.98201329e-01 -1.04470348e+00
4.08813834e-01 2.67715991e-01 -1.76611453e-01 -3.32562327e-01
-5.94635248e-01 1.05727529e-02 1.07878760e-01 -1.34916127e-01
1.64714232e-01 -5.78720093e-01 -9.32087719e-01 -4.25828308e-01
-4.04749572e-01 -5.32279193e-01 -1.05312526e+00 5.99754989e-01
-1.17394865e+00 1.74806848e-01 -1.41394779e-01 -1.46918607e+00
-6.34368286e-02 -1.27057660e+00 5.38848400e-01 4.37875539e-01
-1.48194671e-01 -1.15720284e+00 4.89613026e-01 -4.17319953e-01
5.22725224e-01 2.89434195e-01 8.55123401e-01 -5.41304350e-01
-4.32836145e-01 2.34477654e-01 2.05325466e-02 -9.57437903e-02
9.95099470e-02 -1.45260677e-01 -6.49821281e-01 -6.35007679e-01
-2.73314089e-01 -7.87201583e-01 1.13930595e+00 7.27862716e-01
1.19105494e+00 -6.61266208e-01 -3.97912174e-01 7.39783287e-01
1.39865458e+00 4.85524744e-01 1.51943028e-01 9.27683711e-01
-5.05786203e-02 3.89073610e-01 8.09039772e-01 1.03075159e+00
1.22780828e-02 5.59078991e-01 6.19389057e-01 2.28638828e-01
4.75421429e-01 -7.57980704e-01 2.77377933e-01 -2.39799678e-01
3.82983349e-02 8.14374238e-02 -6.41424179e-01 5.59112847e-01
-2.34470439e+00 -9.35491443e-01 1.07879102e+00 2.09453177e+00
8.55127931e-01 6.96160505e-03 7.65749335e-01 -3.68790984e-01
4.02085513e-01 3.65917981e-01 -1.29416811e+00 -8.60669911e-01
2.02218629e-02 2.64495134e-01 4.56511229e-01 5.40204048e-01
-1.04771197e+00 1.00769448e+00 8.77810192e+00 9.04451191e-01
-8.67052734e-01 -3.50983322e-01 6.61960661e-01 -4.19394165e-01
-7.74303198e-01 1.64871924e-02 -6.53298140e-01 3.59688789e-01
7.16583490e-01 -6.51372492e-01 1.00318992e+00 1.18970954e+00
1.52217299e-01 -1.84330881e-01 -9.11591470e-01 6.66035056e-01
-6.47666991e-01 -1.61786354e+00 1.80113297e-02 3.01403940e-01
9.95660305e-01 -7.90661350e-02 5.11759996e-01 6.10114038e-01
1.37065899e+00 -1.14516401e+00 7.44850516e-01 1.90253302e-01
7.44200349e-01 -1.04217088e+00 -1.14838652e-01 1.87409937e-01
-8.85718286e-01 -8.78738284e-01 -4.04951513e-01 -2.25041747e-01
1.62897799e-02 6.32628277e-02 -7.31325448e-01 3.73096794e-01
4.39712852e-01 7.63454854e-01 6.56588823e-02 1.05374372e+00
-5.28835237e-01 5.17356575e-01 -3.16892952e-01 -3.40083271e-01
9.09470081e-01 -5.18819809e-01 6.01422429e-01 1.09905350e+00
3.82692248e-01 -3.54783088e-01 5.10392725e-01 1.05822337e+00
1.56845301e-01 1.58980340e-01 -8.15529644e-01 -2.39328191e-01
7.28238344e-01 6.82439089e-01 -4.08035576e-01 -2.65006483e-01
-9.51149836e-02 3.19739312e-01 7.98267186e-01 9.31532443e-01
-5.16096175e-01 -1.28016233e-01 1.03070951e+00 -5.16076684e-01
7.19603598e-01 1.49166966e-02 -4.40753661e-02 -1.13610184e+00
-3.45740885e-01 -1.07231891e+00 6.89652920e-01 -3.97456028e-02
-1.00388527e+00 2.02038005e-01 3.29729795e-01 -9.52288747e-01
-1.05714738e+00 -4.98028815e-01 -5.62500119e-01 7.14547575e-01
-1.69690871e+00 -4.84605938e-01 6.63414001e-01 3.51940662e-01
5.13526380e-01 -5.74561477e-01 6.99723661e-01 -6.56192303e-01
-5.31569839e-01 5.10775328e-01 8.47133934e-01 -4.85434562e-01
7.40313902e-02 -1.48537779e+00 3.90682131e-01 6.01453602e-01
3.27742994e-01 4.99257267e-01 7.21497357e-01 -4.73903865e-01
-1.47443068e+00 -6.87242568e-01 -2.50597447e-01 1.03032380e-01
9.94363189e-01 -4.68383394e-02 -5.01923025e-01 6.47223890e-01
1.60102755e-01 -5.27876616e-02 5.35226941e-01 4.65167046e-01
-2.34413147e-01 1.23074353e-01 -1.06493795e+00 8.88853252e-01
9.39643204e-01 -3.23066205e-01 -6.20511115e-01 2.75368728e-02
5.50446391e-01 -2.36125201e-01 -5.86813211e-01 6.67740628e-02
8.18398833e-01 -9.04603004e-01 1.05140996e+00 -1.17518282e+00
1.60778731e-01 2.59062499e-01 -9.92057025e-02 -1.79655576e+00
-5.02172649e-01 -1.31339431e+00 -1.02180147e+00 3.87650371e-01
2.62694716e-01 -9.81389523e-01 9.51692760e-01 4.49191064e-01
3.62634778e-01 -1.17857623e+00 -1.00215852e+00 -9.52079356e-01
4.08986926e-01 -3.14789355e-01 9.71107185e-01 7.28908360e-01
3.97487938e-01 -5.95338941e-01 -4.60683018e-01 -4.61539328e-01
6.93663001e-01 6.04094744e-01 3.50227356e-01 -7.55614161e-01
-5.87028980e-01 -9.69824970e-01 2.02463865e-01 -1.28235102e+00
4.49494213e-01 -6.55502558e-01 -2.19296291e-01 -1.13380051e+00
-1.10471204e-01 -3.61074746e-01 -8.58238041e-01 4.08109397e-01
-1.21613093e-01 -1.76354960e-01 3.38469237e-01 1.10157616e-01
-5.20994961e-01 1.04624450e+00 1.11397243e+00 -7.99799040e-02
-5.00566542e-01 1.76460728e-01 -1.16738605e+00 5.54487050e-01
9.44059134e-01 -3.56163532e-01 -6.84260845e-01 -5.73094077e-02
3.28823179e-01 2.13842839e-01 1.48597313e-02 -4.32234138e-01
-2.04065099e-01 -8.89498770e-01 3.49415392e-01 -1.92105860e-01
2.17868626e-01 -4.23957109e-01 -3.41246337e-01 7.18648374e-01
-7.91828811e-01 1.31579861e-01 3.15570742e-01 1.07715249e+00
-1.10663675e-01 -1.51296824e-01 5.02281785e-01 -3.92887264e-01
-8.68702769e-01 2.57820725e-01 -5.59565544e-01 1.67434499e-01
1.06348896e+00 -2.69053936e-01 -1.27558604e-01 -6.09064639e-01
-7.61833608e-01 6.13331199e-01 3.18279654e-01 2.13436648e-01
6.95976198e-01 -1.51641059e+00 -5.49383640e-01 2.20381036e-01
-2.05192319e-03 -6.60719335e-01 -2.11549789e-01 3.89633775e-01
-3.44587602e-02 5.45064330e-01 -3.39006662e-01 -2.58741647e-01
-5.15838861e-01 7.21764743e-01 4.42018420e-01 -6.41144037e-01
-5.64893126e-01 7.32190549e-01 1.75914258e-01 -2.66108721e-01
4.21946973e-01 -1.66519269e-01 -2.73162156e-01 -4.81113829e-02
6.01086318e-01 5.03965735e-01 -5.80322146e-01 3.36963445e-01
-2.94912420e-02 1.15049981e-01 -2.68970191e-01 -6.22075081e-01
1.36382568e+00 3.36558521e-02 3.01853687e-01 2.25679994e-01
7.58514762e-01 -4.27851319e-01 -1.94956005e+00 -3.56522471e-01
3.85118961e-01 -8.03828061e-01 2.25794002e-01 -9.01667833e-01
-7.13840365e-01 4.80052620e-01 5.05759120e-01 4.72193897e-01
8.99456859e-01 -3.22519839e-01 4.44874577e-02 7.52709329e-01
3.35804433e-01 -1.65251756e+00 7.70492246e-03 4.15408164e-01
8.84316862e-01 -1.08646846e+00 2.54437625e-01 5.26519060e-01
-8.01259041e-01 1.17330575e+00 3.87939811e-01 -4.31239426e-01
5.42758405e-01 1.35690674e-01 -2.08041936e-01 1.00592650e-01
-1.13403928e+00 -3.89269412e-01 2.19158038e-01 8.43328118e-01
7.45107383e-02 2.02363417e-01 -2.17056707e-01 -3.46101001e-02
-4.14702818e-02 1.26021773e-01 1.85814574e-01 1.21801388e+00
-6.90130055e-01 -1.48664236e+00 -3.36077213e-01 5.54063678e-01
-3.51435632e-01 1.79133639e-01 -5.89687973e-02 8.58521938e-01
-7.53676653e-01 6.40699267e-01 1.29462220e-02 -1.91637389e-02
4.55018394e-02 -5.86593412e-02 6.74273908e-01 -3.00292969e-01
-3.97680759e-01 2.72679746e-01 1.15893543e-01 -1.07539344e+00
-2.46200830e-01 -6.94917202e-01 -6.31634057e-01 2.05043028e-03
-7.08212890e-03 5.94914675e-01 4.74309713e-01 7.65362918e-01
3.27290595e-01 4.12849307e-01 9.56611931e-01 -6.49894238e-01
-1.59701407e+00 -5.21513522e-01 -8.01024675e-01 1.63277145e-02
6.95779681e-01 -1.08960128e+00 -6.62704930e-02 -8.19690287e-01] | [4.129539489746094, 2.1736249923706055] |
a8f20e9a-d0c7-4de3-96db-d09bbeb2e611 | debiasing-scores-and-prompts-of-2d-diffusion | 2303.15413 | null | https://arxiv.org/abs/2303.15413v2 | https://arxiv.org/pdf/2303.15413v2.pdf | Debiasing Scores and Prompts of 2D Diffusion for Robust Text-to-3D Generation | The view inconsistency problem in score-distilling text-to-3D generation, also known as the Janus problem, arises from the intrinsic bias of 2D diffusion models, which leads to the unrealistic generation of 3D objects. In this work, we explore score-distilling text-to-3D generation and identify the main causes of the Janus problem. Based on these findings, we propose two approaches to debias the score-distillation frameworks for robust text-to-3D generation. Our first approach, called score debiasing, involves gradually increasing the truncation value for the score estimated by 2D diffusion models throughout the optimization process. Our second approach, called prompt debiasing, identifies conflicting words between user prompts and view prompts utilizing a language model and adjusts the discrepancy between view prompts and object-space camera poses. Our experimental results show that our methods improve realism by significantly reducing artifacts and achieve a good trade-off between faithfulness to the 2D diffusion models and 3D consistency with little overhead. | ['Seungryong Kim', 'Donghoon Ahn', 'Susung Hong'] | 2023-03-27 | null | null | null | null | ['text-to-3d'] | ['computer-vision'] | [ 1.82792004e-02 1.78289972e-02 1.50804684e-01 -2.47198164e-01
-7.89110422e-01 -8.73087645e-01 6.14777327e-01 -2.58259147e-01
2.91870445e-01 3.55297267e-01 5.04640996e-01 -1.75374925e-01
-3.82539965e-02 -3.55790168e-01 -4.32748824e-01 -3.62893492e-01
5.19737303e-01 4.08164620e-01 2.57180095e-01 -4.47530933e-02
7.18080163e-01 4.53178257e-01 -1.31237364e+00 1.89193234e-01
9.66777027e-01 7.73290336e-01 4.15486783e-01 1.01832509e+00
-1.92813158e-01 8.80383492e-01 -9.69283640e-01 -4.50160801e-01
4.35762793e-01 -6.79518163e-01 -4.17062134e-01 2.44435608e-01
6.68362617e-01 -8.91181171e-01 -2.48267185e-02 1.05138934e+00
7.82971799e-01 -1.12818806e-02 5.89320540e-01 -1.26982033e+00
-8.50094080e-01 1.34620413e-01 -8.87320399e-01 -6.46511093e-02
9.07303989e-01 1.89041883e-01 7.46169806e-01 -1.10799611e+00
8.02427828e-01 1.55836022e+00 4.66994584e-01 6.13346636e-01
-1.15553939e+00 -5.51654339e-01 9.34923887e-02 -1.67518154e-01
-1.49639070e+00 -4.77044910e-01 8.95825088e-01 -7.64718175e-01
6.44239366e-01 5.24991274e-01 8.92308533e-01 1.07270920e+00
2.63115585e-01 6.46486104e-01 1.11056709e+00 -2.98213035e-01
3.51429582e-01 3.03793103e-01 -3.20330113e-01 6.87290370e-01
2.44674236e-01 2.10873574e-01 -8.64229620e-01 -3.74807626e-01
1.48006344e+00 -2.59093285e-01 -3.26852322e-01 -6.76400304e-01
-1.33383024e+00 5.74346006e-01 -1.74206212e-01 -2.15742126e-01
-2.01777562e-01 1.29011363e-01 2.94607300e-02 1.75769806e-01
7.31914222e-01 5.92839360e-01 -2.48001590e-01 -4.21846986e-01
-9.73847866e-01 7.14172959e-01 6.50389612e-01 1.31221163e+00
4.36078429e-01 7.21275359e-02 -3.60694230e-01 7.22937882e-01
4.31849539e-01 7.62242854e-01 3.00703436e-01 -1.39290345e+00
5.43941200e-01 5.93356669e-01 5.12224257e-01 -1.16836536e+00
-1.05300099e-01 -1.01902829e-02 -5.15600383e-01 4.61794674e-01
3.70814592e-01 -2.44705841e-01 -6.10556722e-01 1.62939310e+00
6.23468578e-01 -3.51241142e-01 -2.07504377e-01 1.21686327e+00
6.79450214e-01 5.84639132e-01 -4.24019635e-01 -2.94446737e-01
9.92704451e-01 -9.12364066e-01 -8.99053812e-01 1.23456847e-02
5.95960200e-01 -1.27966464e+00 1.29624891e+00 4.41358328e-01
-1.56248891e+00 -4.99697745e-01 -9.64007795e-01 -1.32227927e-01
2.10286334e-01 7.72526190e-02 2.60751456e-01 5.03205538e-01
-1.11888719e+00 4.31063563e-01 -4.79204923e-01 -4.03121024e-01
-9.22950506e-02 5.69142327e-02 -6.16194643e-02 1.13506041e-01
-5.90358675e-01 9.92584944e-01 -3.47810872e-02 -3.09969991e-01
-6.47920072e-01 -7.82408118e-01 -5.47486305e-01 -1.96305051e-01
6.53653741e-01 -8.87183845e-01 1.54008567e+00 -8.41838658e-01
-1.83520079e+00 8.06819856e-01 -2.42779747e-01 2.13879511e-01
8.18934023e-01 -5.03698409e-01 6.50454685e-02 -9.15079936e-02
2.24641055e-01 7.91818619e-01 9.23263669e-01 -1.74632394e+00
-2.69111872e-01 -3.57777238e-01 1.95221439e-01 7.80239940e-01
-9.05154869e-02 -1.96931690e-01 -6.42824709e-01 -9.68353510e-01
4.46327627e-01 -1.08351874e+00 -1.70958146e-01 4.37420845e-01
-4.62586164e-01 1.56293269e-02 8.51360977e-01 -4.90439773e-01
1.42168045e+00 -1.98997164e+00 4.67910379e-01 8.41548517e-02
4.66657758e-01 -1.72793791e-01 -1.19454250e-01 3.60618889e-01
6.97579607e-02 6.15332983e-02 2.99999416e-01 -4.20952439e-01
-9.27319154e-02 1.23609766e-01 -3.35218400e-01 1.48847565e-01
-1.65174857e-01 7.01710939e-01 -1.03908384e+00 -5.37763953e-01
3.39165419e-01 2.46195555e-01 -8.62723112e-01 5.32083094e-01
-2.11640656e-01 4.46394444e-01 -3.87851745e-01 4.46246147e-01
9.81165230e-01 -3.53368849e-01 1.07783951e-01 -2.95898139e-01
-3.66256684e-01 1.78375781e-01 -1.29792857e+00 1.90103805e+00
-4.05306429e-01 5.04948437e-01 -2.36366969e-02 1.07333921e-01
8.46140087e-01 1.48891196e-01 5.30303895e-01 -4.98824745e-01
-1.10281907e-01 1.80324882e-01 -3.40239972e-01 -5.46279252e-01
9.92346525e-01 -1.28042847e-01 3.15408036e-02 5.74962556e-01
-2.83958554e-01 -1.04140639e+00 -6.89449757e-02 6.06258690e-01
6.32154465e-01 4.07219768e-01 4.23493207e-01 -3.29046577e-01
1.62872151e-01 -1.26201063e-01 2.39791974e-01 6.52350545e-01
-8.44785720e-02 1.05906796e+00 6.46537304e-01 -3.45449924e-01
-1.16220081e+00 -1.17888558e+00 4.58379954e-01 6.17443144e-01
6.43071413e-01 -6.42567277e-01 -9.73635674e-01 -7.24833786e-01
-2.29691297e-01 8.90375555e-01 -5.15169144e-01 -1.61958300e-02
-4.09511596e-01 -1.71217635e-01 1.87415823e-01 4.84378576e-01
2.35659197e-01 -2.92433858e-01 -8.76335740e-01 -3.59453037e-02
-4.15576249e-01 -7.66171753e-01 -1.17111850e+00 -2.99160361e-01
-9.75316823e-01 -9.29478705e-01 -1.02849925e+00 -1.39325380e-01
1.02791941e+00 7.29547501e-01 1.11843419e+00 4.07852195e-02
2.13558033e-01 4.39934433e-01 -2.90134579e-01 -3.75937998e-01
-6.44834161e-01 -2.73244560e-01 5.99364601e-02 -3.44077408e-01
-1.33960083e-01 -2.50236839e-01 -8.30802143e-01 7.03175426e-01
-8.98032904e-01 6.04143202e-01 2.89488345e-01 4.34667438e-01
4.38782513e-01 -5.00196159e-01 -1.10336589e-02 -4.95868653e-01
8.25623155e-01 -3.00443731e-03 -7.91989505e-01 3.13199498e-02
-6.77492559e-01 2.71597117e-01 3.57438833e-01 -7.52807319e-01
-1.11178434e+00 -1.38991386e-01 4.31740321e-02 -5.90944290e-01
2.41153330e-01 -7.80899078e-02 -1.06637232e-01 5.85623756e-02
7.30135322e-01 -9.33845714e-02 -1.82427332e-01 -2.71496773e-01
5.61614037e-01 3.74639541e-01 1.59571454e-01 -5.89251041e-01
8.66684079e-01 6.16613626e-01 -3.22529614e-01 -5.29134035e-01
-9.70750213e-01 -2.06950203e-01 -3.88755322e-01 -5.70562720e-01
7.86524534e-01 -6.97175086e-01 -4.94750559e-01 4.85197484e-01
-1.58353102e+00 -3.16784859e-01 -4.04724687e-01 3.02517354e-01
-7.21250713e-01 4.19196010e-01 -3.49003911e-01 -7.95296371e-01
1.35434186e-02 -1.22350693e+00 1.38621795e+00 1.34576231e-01
-6.99813366e-01 -7.42713690e-01 2.57144272e-01 2.95360267e-01
4.64961320e-01 7.65030757e-02 8.98301125e-01 4.77036578e-04
-7.30890453e-01 3.12004182e-02 -2.79499143e-01 8.97581503e-02
1.86905190e-01 4.77249175e-03 -6.00576639e-01 -2.04698771e-01
3.15037161e-01 -1.05495214e-01 3.65837850e-02 4.60477948e-01
9.22478020e-01 -4.54897374e-01 -9.29530114e-02 4.85288382e-01
1.14582765e+00 3.54750901e-01 4.94208783e-01 6.37177331e-03
7.52456069e-01 4.31196123e-01 7.25560784e-01 9.76547480e-01
4.79144007e-01 1.05208528e+00 4.77156520e-01 2.32631639e-02
-4.25641149e-01 -6.75129950e-01 1.85239702e-01 1.06476355e+00
-1.47231564e-01 -5.41436732e-01 -5.88415623e-01 4.71086949e-01
-1.69737530e+00 -7.95390904e-01 -2.37498134e-01 2.42903876e+00
6.66972339e-01 8.03290606e-02 -1.45260978e-03 -2.76685208e-01
6.18036985e-01 3.61759722e-01 -5.64280689e-01 -3.32116872e-01
1.98021084e-01 -5.65166175e-01 3.48704457e-01 7.69236684e-01
-3.08527946e-01 9.13377702e-01 7.11473751e+00 6.32610023e-01
-9.51057434e-01 8.77598375e-02 4.90782470e-01 -5.36418617e-01
-6.69548690e-01 1.23226933e-01 -6.84843600e-01 4.99683052e-01
3.53071153e-01 -2.06639081e-01 5.71406662e-01 6.78809464e-01
7.49963820e-01 -2.75866508e-01 -1.23022830e+00 1.19569325e+00
3.53613466e-01 -1.21289206e+00 3.51631731e-01 9.75091159e-02
1.02684736e+00 -5.89898109e-01 2.07363695e-01 -1.46026254e-01
5.16506076e-01 -4.93888438e-01 1.40880525e+00 4.48611051e-01
9.39612567e-01 -4.68929887e-01 2.29943350e-01 3.98431867e-01
-9.84983921e-01 2.75676936e-01 -1.64140850e-01 4.76916097e-02
6.06686532e-01 6.22404873e-01 -7.25479722e-01 2.48599544e-01
4.70150352e-01 3.19454700e-01 -3.36030394e-01 6.89652145e-01
-2.49588713e-01 1.49249062e-01 -1.21353164e-01 4.26311232e-02
-4.83362973e-02 -1.31555468e-01 9.03798342e-01 6.82555437e-01
8.49301517e-01 2.38120690e-01 -1.50956027e-02 1.10998309e+00
1.64279059e-01 -1.45646438e-01 -7.89253294e-01 2.86594540e-01
6.02773666e-01 8.15796494e-01 -5.05996585e-01 -3.63923281e-01
-4.88827080e-02 1.37684643e+00 -7.30610359e-03 3.31919670e-01
-1.14910686e+00 3.84903233e-03 4.49034810e-01 4.74515259e-01
1.03367291e-01 -5.49683630e-01 -5.29488444e-01 -1.16173029e+00
9.67716873e-02 -8.62147093e-01 -1.88503802e-01 -1.47414339e+00
-9.73741531e-01 3.36038202e-01 3.32296491e-01 -1.50262225e+00
-3.54786456e-01 -1.38736963e-01 -3.33371192e-01 7.65609801e-01
-1.10751069e+00 -9.14673865e-01 -5.47888458e-01 4.67705131e-01
1.03371584e+00 2.46739253e-01 4.95925546e-01 5.22045158e-02
1.12368725e-02 3.91377389e-01 -7.96562061e-02 -6.13473177e-01
1.07507169e+00 -1.21330178e+00 6.38738334e-01 5.99422157e-01
-1.28339976e-01 6.72638714e-01 8.55493605e-01 -7.92628646e-01
-1.38129842e+00 -6.04558587e-01 8.43161583e-01 -9.18385446e-01
3.00441146e-01 -3.27874094e-01 -5.18759847e-01 3.50874990e-01
2.65435934e-01 -3.72484624e-01 4.74771410e-01 -3.05093318e-01
-2.23822936e-01 1.11901112e-01 -1.06620586e+00 1.02398777e+00
1.21243119e+00 -3.77647907e-01 -2.96387017e-01 1.46404594e-01
8.48799229e-01 -1.01711726e+00 -5.29571652e-01 7.30558438e-03
6.91351652e-01 -1.28321171e+00 8.61726284e-01 -4.41723280e-02
6.78694904e-01 -3.85963917e-01 -1.86146155e-01 -1.41951978e+00
-4.00975317e-01 -9.93350029e-01 -3.77292782e-01 9.80887175e-01
2.14385435e-01 -1.07247114e-01 7.19371378e-01 7.18880296e-01
-1.73727702e-02 -4.77366507e-01 -7.64019907e-01 -7.77192235e-01
-1.97672084e-01 -2.14274034e-01 5.80401599e-01 8.48903954e-01
-1.46258071e-01 3.72112691e-01 -6.31337464e-01 6.83659464e-02
5.81145406e-01 -4.47986014e-02 1.08362186e+00 -9.82298613e-01
-4.42978859e-01 -2.93785602e-01 1.62728563e-01 -1.90414572e+00
-4.62888330e-01 -2.81966567e-01 5.97562715e-02 -1.46571159e+00
2.76076764e-01 -2.66562849e-01 4.23090726e-01 -6.79342598e-02
-1.83739379e-01 -1.33163640e-02 5.25896668e-01 4.12672669e-01
-4.95165616e-01 5.23953855e-01 1.76852536e+00 1.24258950e-01
-6.42311335e-01 -2.27838099e-01 -8.27669382e-01 9.11625504e-01
3.91913235e-01 -4.79723483e-01 -8.12245905e-01 -1.02104843e+00
6.24710083e-01 4.96668845e-01 3.98963928e-01 -7.99147725e-01
3.31206508e-02 -4.90098029e-01 7.78361559e-02 -9.82755303e-01
4.38296288e-01 -8.60977888e-01 2.93587983e-01 2.88656622e-01
-4.17777121e-01 4.62891906e-01 9.71868113e-02 5.61379850e-01
2.09159642e-01 -2.23883331e-01 7.06907213e-01 -2.83632010e-01
-2.35820591e-01 5.07443957e-03 -5.26553273e-01 1.11164577e-01
1.00780857e+00 -6.12051010e-01 -2.60577470e-01 -8.29611301e-01
-4.08196837e-01 -1.45933321e-02 8.65333200e-01 6.55686796e-01
8.56588781e-01 -1.65464127e+00 -5.59288740e-01 3.51108670e-01
1.29924202e-02 8.63330290e-02 3.16468984e-01 5.46654880e-01
-6.39629066e-01 2.30218813e-01 2.54992872e-01 -7.44484425e-01
-1.46318984e+00 2.97155857e-01 1.26023248e-01 -2.28835061e-01
-3.87478560e-01 8.36438358e-01 5.35120428e-01 -2.13761806e-01
7.26317614e-02 -4.77512240e-01 3.22480828e-01 -2.61893123e-01
3.95053834e-01 5.80583572e-01 -1.94839120e-01 -4.00693059e-01
-2.24052012e-01 7.81643152e-01 -1.83762595e-01 -6.57194853e-01
7.84970582e-01 -5.43892145e-01 2.13977158e-01 2.82238305e-01
8.15039217e-01 4.55377609e-01 -1.67977190e+00 2.08879173e-01
-4.39017534e-01 -1.02573073e+00 -3.19945849e-02 -9.87387955e-01
-7.72787333e-01 6.26747429e-01 4.78811532e-01 2.39440635e-01
8.12147081e-01 -5.48573770e-02 9.19441164e-01 1.77122392e-02
3.54770064e-01 -1.26347876e+00 6.74532890e-01 3.90168667e-01
1.19455290e+00 -9.93114710e-01 9.57862660e-02 -4.54632878e-01
-9.90007937e-01 8.08322489e-01 9.41703022e-01 2.53528237e-01
1.79368421e-01 3.76467586e-01 2.42642373e-01 -2.63127625e-01
-8.18802476e-01 3.98960590e-01 2.40941823e-01 5.55323601e-01
4.35424298e-01 -1.01856820e-01 -2.80110240e-01 3.28772783e-01
-1.77495942e-01 -3.17991711e-02 7.79773295e-01 8.16460848e-01
-3.47762585e-01 -9.20729756e-01 -7.13535786e-01 -5.72136976e-02
-1.70861818e-02 -9.16723609e-02 -7.20306396e-01 6.12190485e-01
9.06055644e-02 1.01954114e+00 -1.15148365e-01 -4.18774486e-01
6.06308997e-01 -3.02953273e-01 7.62132883e-01 -5.15839040e-01
-4.02287364e-01 3.90110344e-01 -1.83294594e-01 -7.10233152e-01
-3.25307339e-01 -4.23788846e-01 -9.49615061e-01 -5.64828277e-01
-6.66871071e-01 -2.28727639e-01 5.78953564e-01 6.31994188e-01
5.87515235e-01 3.00058872e-01 6.47720277e-01 -9.65574741e-01
-7.99118459e-01 -6.81399465e-01 -3.49311501e-01 5.20968616e-01
2.35928297e-01 -6.36441410e-01 -3.25227976e-01 1.53606907e-01] | [11.282845497131348, -0.41237348318099976] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.