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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2caf247d-2b30-4cf8-82d3-491e70a811c5 | open-world-story-generation-with-structured | 2212.04634 | null | https://arxiv.org/abs/2212.04634v2 | https://arxiv.org/pdf/2212.04634v2.pdf | Open-world Story Generation with Structured Knowledge Enhancement: A Comprehensive Survey | Storytelling and narrative are fundamental to human experience, intertwined with our social and cultural engagement. As such, researchers have long attempted to create systems that can generate stories automatically. In recent years, powered by deep learning and massive data resources, automatic story generation has shown significant advances. However, considerable challenges, like the need for global coherence in generated stories, still hamper generative models from reaching the same storytelling ability as human narrators. To tackle these challenges, many studies seek to inject structured knowledge into the generation process, which is referred to as structured knowledge-enhanced story generation. Incorporating external knowledge can enhance the logical coherence among story events, achieve better knowledge grounding, and alleviate over-generalization and repetition problems in stories. This survey provides the latest and comprehensive review of this research field: (i) we present a systematical taxonomy regarding how existing methods integrate structured knowledge into story generation; (ii) we summarize involved story corpora, structured knowledge datasets, and evaluation metrics; (iii) we give multidimensional insights into the challenges of knowledge-enhanced story generation and cast light on promising directions for future study. | ['Börje F. Karlsson', 'Wei Hu', 'Zhiwei Yu', 'Jieru Lin', 'Yuxin Wang'] | 2022-12-09 | null | null | null | null | ['story-generation'] | ['natural-language-processing'] | [ 2.18927816e-01 3.83920163e-01 -3.93108428e-01 -5.72699234e-02
-6.86002076e-01 -7.25541234e-01 9.64003801e-01 1.66795552e-02
1.86518118e-01 9.75371003e-01 1.13718700e+00 2.23074600e-01
-2.02793583e-01 -1.15847623e+00 -5.35372257e-01 -2.61259556e-01
1.95519164e-01 3.34136099e-01 -7.71891922e-02 -3.86661053e-01
3.39065522e-01 -1.47080362e-01 -1.70770454e+00 7.39805698e-01
1.04067898e+00 3.71017963e-01 2.13254169e-01 5.96342564e-01
-3.91036719e-01 1.61413705e+00 -8.50059509e-01 -6.02847457e-01
-4.35861945e-01 -1.08359766e+00 -1.13291371e+00 2.51353145e-01
-1.90311193e-01 -3.53921801e-01 -4.62688923e-01 4.86445367e-01
4.91783768e-01 1.80075020e-01 6.23784661e-01 -1.14493752e+00
-1.09402502e+00 1.40786934e+00 -7.57116126e-03 -1.03365697e-01
8.78134787e-01 6.60855621e-02 1.09316838e+00 -7.96581388e-01
1.24883878e+00 8.85561764e-01 6.89739347e-01 4.74716842e-01
-9.31497097e-01 -2.99459130e-01 7.39984512e-02 5.99409759e-01
-1.32246935e+00 -3.57976079e-01 9.83298659e-01 -5.73194563e-01
9.94745493e-01 2.21917301e-01 1.37904274e+00 1.55337608e+00
-2.24495322e-01 1.22631967e+00 6.81649029e-01 -3.95368397e-01
6.75650537e-02 2.02977266e-02 -2.35241726e-01 1.96464285e-01
7.88142234e-02 -1.68741465e-01 -1.02665067e+00 1.68690428e-01
9.68930721e-01 -4.85914826e-01 -1.53263420e-01 8.11281279e-02
-1.53404224e+00 9.56223965e-01 1.03731610e-01 6.35298073e-01
-5.40530205e-01 -9.36586317e-03 5.91066480e-01 -4.50401008e-02
3.09067965e-01 8.87887895e-01 2.16572478e-01 -8.67509127e-01
-9.32865739e-01 9.40147042e-01 8.60471964e-01 1.06663764e+00
2.99163014e-01 3.37197483e-01 -4.83896792e-01 9.18191731e-01
-1.38960585e-01 1.93299130e-01 4.37708765e-01 -9.72175419e-01
4.17253703e-01 7.61941195e-01 6.21930622e-02 -1.21172333e+00
-2.48686805e-01 -4.91855323e-01 -6.51084840e-01 -4.08839434e-01
6.91064727e-03 -3.45728040e-01 -3.54152709e-01 1.68558538e+00
1.33359462e-01 1.36649475e-01 1.98142976e-01 6.43301725e-01
1.46669030e+00 8.10104370e-01 -8.03059861e-02 -2.04907194e-01
1.20793736e+00 -9.80375886e-01 -1.08103538e+00 -2.46381387e-01
5.98087251e-01 -8.27462733e-01 1.16687953e+00 1.69269517e-01
-1.48049200e+00 -3.53638709e-01 -1.08067882e+00 -1.51802853e-01
-2.65162379e-01 -7.38474503e-02 1.03049755e+00 4.61561680e-01
-7.58295357e-01 4.36078936e-01 -5.75533509e-01 -5.29527843e-01
5.66894293e-01 -3.27546746e-01 7.95383658e-03 -8.17769095e-02
-1.47098744e+00 9.79721308e-01 7.96590388e-01 -2.77322054e-01
-8.10681403e-01 -8.49617124e-01 -8.51839483e-01 -2.22395450e-01
5.50613403e-01 -9.09905553e-01 1.38961494e+00 -7.04018295e-01
-1.57459497e+00 6.56487584e-01 -5.92708588e-02 -2.20114082e-01
3.33615899e-01 -4.22227681e-01 -3.78252298e-01 -7.14179873e-02
3.01953495e-01 6.88894391e-01 9.40362141e-02 -1.39668179e+00
-4.55970645e-01 2.83087164e-01 2.46700317e-01 7.17261016e-01
-4.31068093e-01 1.35807753e-01 -2.60720253e-01 -9.69030976e-01
-2.25409225e-01 -4.85790282e-01 -9.86273214e-02 -7.62772739e-01
-5.82219839e-01 -2.56775081e-01 6.54906750e-01 -5.81480086e-01
1.73326647e+00 -1.87199771e+00 3.40784103e-01 -2.76531994e-01
1.27427444e-01 -3.89925647e-03 1.48545150e-02 1.17729163e+00
2.35970199e-01 8.43984336e-02 -3.55651826e-02 -1.86727464e-01
8.99800286e-02 1.97664246e-01 -5.42681336e-01 -2.87443101e-01
1.98640093e-01 1.37577629e+00 -1.27314389e+00 -6.80557966e-01
1.68358788e-01 5.78237712e-01 -5.84877789e-01 1.42519236e-01
-5.19615889e-01 4.37387705e-01 -4.27782565e-01 4.54599589e-01
-1.12803034e-01 -3.91455621e-01 1.10787943e-01 1.88601002e-01
-2.07459271e-01 6.45833850e-01 -1.11470747e+00 1.83096421e+00
-3.00048769e-01 9.89638627e-01 -7.08350778e-01 -5.25945425e-01
9.35301006e-01 6.59955680e-01 3.94512922e-01 -4.23054516e-01
6.18922152e-02 -6.45630732e-02 -2.56750971e-01 -7.95285285e-01
1.06684160e+00 -3.60277086e-01 -3.60647917e-01 7.44880795e-01
-4.63793501e-02 -6.28510237e-01 4.82926846e-01 4.34024930e-01
1.02323771e+00 4.10571665e-01 4.59770471e-01 3.18180025e-01
-5.77272624e-02 4.82270360e-01 3.65718961e-01 6.77778721e-01
3.58508766e-01 7.35376120e-01 5.72746038e-01 -3.59434336e-01
-1.03129876e+00 -1.04505396e+00 2.06879437e-01 8.76285493e-01
1.13139721e-02 -9.22623754e-01 -8.27409029e-01 -1.36334553e-01
-4.68798846e-01 1.14658904e+00 -6.36870444e-01 5.98408189e-03
-6.59873724e-01 -6.40340626e-01 8.67070258e-01 7.65111268e-01
5.53599656e-01 -1.49046826e+00 -5.50300956e-01 5.15922189e-01
-9.47141230e-01 -1.05301797e+00 3.13386694e-02 -4.64367807e-01
-5.08665860e-01 -8.50921690e-01 -5.62742531e-01 -7.11901963e-01
2.26141468e-01 2.42730409e-01 1.40448725e+00 -1.58256859e-01
-1.04234762e-01 3.52622390e-01 -9.45434690e-01 -5.18304408e-01
-5.79360962e-01 2.54031569e-01 -2.94975638e-01 -4.10959274e-01
1.33134291e-01 -7.50221431e-01 -2.41251662e-01 3.86357307e-02
-9.95962739e-01 8.38062704e-01 3.22808027e-01 6.79784536e-01
4.03828859e-01 2.50591248e-01 9.76907849e-01 -5.85998952e-01
1.05571187e+00 -6.90118134e-01 1.83016881e-01 1.34299994e-01
-1.92434087e-01 -3.90540361e-01 4.26197022e-01 -4.53729630e-01
-1.48488903e+00 -4.03957576e-01 -6.29591718e-02 1.92937747e-01
3.74502204e-02 9.83222723e-01 -1.18186094e-01 5.77207088e-01
8.63805592e-01 3.06075543e-01 -4.02550757e-01 -5.46564236e-02
8.22722435e-01 3.57354730e-01 6.80716932e-01 -6.32838726e-01
6.28972888e-01 3.37496996e-01 -4.71955478e-01 -8.73269081e-01
-1.01547945e+00 7.04890266e-02 -4.21589494e-01 -8.96304011e-01
8.43633831e-01 -8.74327540e-01 -2.20938161e-01 4.45506752e-01
-1.32468712e+00 -4.28517461e-01 -7.72045195e-01 4.35892433e-01
-9.75716174e-01 1.00301087e-01 -6.90960526e-01 -5.73838592e-01
-1.74454540e-01 -5.93466103e-01 6.18818998e-01 4.82285976e-01
-1.11921978e+00 -1.19000530e+00 3.06717128e-01 6.76071286e-01
3.41008037e-01 9.47849452e-01 9.64870214e-01 -2.94870555e-01
-6.50493026e-01 -1.97223201e-01 1.44081528e-03 -2.21286938e-01
1.00211479e-01 -1.29580982e-02 -7.29829729e-01 4.56917673e-01
-2.10752085e-01 -6.89705133e-01 3.56456697e-01 2.99025655e-01
5.91904581e-01 -4.69306350e-01 -2.62843668e-01 2.39613160e-01
9.61710691e-01 2.34800175e-01 8.24718714e-01 4.39619392e-01
8.16469371e-01 6.51161313e-01 6.35107458e-01 8.24902534e-01
9.01297092e-01 6.13713443e-01 -4.38185334e-02 2.58194178e-01
-5.26901603e-01 -8.41312766e-01 2.97925442e-01 1.18942034e+00
-5.40562809e-01 -4.44935083e-01 -8.55159819e-01 9.02876318e-01
-2.07537246e+00 -1.50928724e+00 -3.22625190e-01 1.45731986e+00
1.20058095e+00 -1.06798089e-03 4.31210771e-02 2.78998047e-01
4.53793943e-01 4.54929650e-01 -3.61271799e-01 -1.85271770e-01
-6.38830602e-01 -3.40460688e-02 -3.49551082e-01 3.09062421e-01
-7.16629744e-01 1.30349731e+00 7.25317478e+00 9.58918810e-01
-6.63753867e-01 7.83027485e-02 3.66293252e-01 -3.88559967e-01
-7.53550887e-01 3.40290666e-02 -4.96980220e-01 1.60387859e-01
4.14949328e-01 -6.91742897e-01 4.08756584e-01 7.94441760e-01
3.00873160e-01 -1.10576637e-01 -1.07800913e+00 9.01112854e-01
3.74111414e-01 -2.01930499e+00 2.17130333e-01 -9.92654413e-02
1.43450499e+00 -5.52923977e-01 -1.56901479e-01 3.92568618e-01
5.40685713e-01 -1.23139119e+00 1.03680611e+00 5.43532908e-01
6.45390451e-01 -8.82929206e-01 6.25766993e-01 3.25382590e-01
-1.34168577e+00 2.41328135e-01 6.87541291e-02 -6.26154542e-01
7.89613426e-01 5.44348776e-01 -7.40918696e-01 6.60822392e-01
3.90701115e-01 9.25194860e-01 -2.09882438e-01 8.75089169e-01
-7.16546834e-01 6.80081546e-01 1.63322255e-01 -1.90431878e-01
1.91076308e-01 1.44929156e-01 7.16257036e-01 1.22677910e+00
3.55643690e-01 3.59708011e-01 -4.12346656e-03 1.15412068e+00
5.67076504e-02 2.57112324e-01 -7.47762024e-01 -6.90875351e-01
6.29535556e-01 8.71282279e-01 -8.87771070e-01 -3.44489664e-01
-2.42854759e-01 7.61407435e-01 2.17018440e-01 2.66526312e-01
-9.31178272e-01 -3.18173200e-01 4.20405060e-01 1.74010292e-01
-1.24220075e-02 -2.44482413e-01 -8.17653894e-01 -9.81140494e-01
-7.11339489e-02 -8.57516110e-01 1.03860505e-01 -1.04190981e+00
-1.03404760e+00 4.73020583e-01 3.16342205e-01 -1.09332502e+00
-7.75292575e-01 2.61008084e-01 -7.53445387e-01 3.63004386e-01
-7.42277563e-01 -1.36421120e+00 -4.55231488e-01 2.09059685e-01
8.46466184e-01 -7.90916234e-02 9.22822475e-01 2.55268924e-02
-3.03804964e-01 3.11782718e-01 -2.67351657e-01 7.96657354e-02
3.94493312e-01 -8.41212511e-01 4.62905943e-01 7.07094133e-01
1.96886435e-01 3.18929493e-01 7.21131682e-01 -1.06638038e+00
-1.13038552e+00 -1.00892627e+00 1.29000306e+00 -6.65517569e-01
6.99609160e-01 -2.32908845e-01 -7.73048282e-01 6.89518869e-01
5.64868689e-01 -1.01025844e+00 1.18126047e+00 2.07095206e-01
-1.23099402e-01 4.93039101e-01 -7.07627416e-01 1.02901506e+00
1.40181446e+00 -3.07726800e-01 -8.23758066e-01 2.38712311e-01
8.49642038e-01 -5.37168920e-01 -9.92191911e-01 1.21839806e-01
6.32675111e-01 -9.82707620e-01 7.37335980e-01 -4.32375044e-01
1.34588015e+00 -1.87244013e-01 1.23950124e-01 -1.29045939e+00
-4.84457582e-01 -8.56440723e-01 -3.56836230e-01 1.60086727e+00
3.60393494e-01 -1.40527252e-03 7.06901133e-01 4.91032392e-01
-3.99743736e-01 -8.17321897e-01 -4.37669188e-01 -6.32706463e-01
-7.25310594e-02 -7.52521336e-01 8.03949177e-01 1.23012245e+00
6.83992326e-01 5.65110803e-01 -6.06777906e-01 -4.84501004e-01
3.04422140e-01 5.67846410e-02 8.29161823e-01 -9.43869889e-01
-2.82298952e-01 -8.00084293e-01 -1.98273674e-01 -8.54332924e-01
-6.04605637e-02 -8.39088380e-01 -3.04867085e-02 -2.34310627e+00
5.42945385e-01 -3.39408927e-02 4.18248028e-01 4.41838056e-01
-2.66600668e-01 3.11392605e-01 2.82430559e-01 1.05080351e-01
-6.46597683e-01 7.19568908e-01 1.62965405e+00 1.55460894e-01
-5.42877734e-01 -3.66204113e-01 -1.16765881e+00 7.93751955e-01
8.26951206e-01 -1.28591344e-01 -8.48541260e-01 -4.75557268e-01
9.53893065e-01 2.02691168e-01 1.34754613e-01 -1.07883143e+00
2.62674570e-01 -5.49638331e-01 3.22219372e-01 -8.51913273e-01
3.89384031e-01 -1.03624621e-02 6.43757761e-01 1.35505209e-02
-4.48463976e-01 -1.16718695e-01 1.61469847e-01 1.87202603e-01
-4.35736448e-01 -2.02759162e-01 3.29116762e-01 -1.57347694e-01
-6.60995603e-01 -2.32131690e-01 -8.74584436e-01 2.22860232e-01
1.23254049e+00 -6.27102613e-01 -3.49346787e-01 -9.54731584e-01
-4.77852225e-01 2.09928080e-01 3.98123235e-01 7.32272208e-01
7.37665296e-01 -1.83943367e+00 -1.00494790e+00 -3.55834097e-01
2.73611277e-01 2.01502010e-01 5.49109280e-01 3.84260744e-01
-3.87623042e-01 3.26426804e-01 -1.40739605e-01 -7.38260821e-02
-9.61591065e-01 1.95401773e-01 -1.40659228e-01 -3.34750891e-01
-6.82594776e-01 8.90062809e-01 -7.38658905e-02 3.11998911e-02
1.19319670e-02 6.77511096e-02 -5.04307449e-01 2.80594379e-01
7.45932579e-01 6.16128325e-01 -3.91176224e-01 -5.69104135e-01
6.99453875e-02 3.05290967e-01 1.48454994e-01 -4.26371783e-01
1.34617376e+00 -7.86202922e-02 -2.36761682e-02 7.17162490e-01
4.63062346e-01 -9.80714634e-02 -1.02023089e+00 -1.97400212e-01
-2.94933431e-02 -2.56846189e-01 -2.46971369e-01 -9.51447308e-01
-5.79108119e-01 5.19413173e-01 -4.53949571e-01 3.78244966e-01
8.75655890e-01 4.03317779e-01 1.00907743e+00 1.95103213e-01
3.16956908e-01 -1.23905432e+00 6.13750398e-01 8.43299091e-01
1.22022045e+00 -7.93570340e-01 4.61159907e-02 -4.74050403e-01
-1.13502693e+00 8.97411764e-01 6.52688682e-01 1.41051278e-01
1.73275918e-01 2.67584026e-01 -1.80862412e-01 -2.38621742e-01
-8.49439263e-01 -1.50855318e-01 9.78774130e-02 7.16920435e-01
6.09240592e-01 2.72219747e-01 -3.42479467e-01 1.07146406e+00
-1.00688207e+00 2.73833722e-01 6.16181016e-01 8.85393083e-01
-4.33352530e-01 -1.09247077e+00 -3.88924211e-01 3.57871741e-01
-5.48276231e-02 2.63161827e-02 -7.53952265e-01 7.07046747e-01
3.16305965e-01 1.12805307e+00 -1.65545791e-02 -5.45084238e-01
2.12168396e-01 9.23232511e-02 6.58281028e-01 -7.19785988e-01
-6.22518480e-01 -1.50959522e-01 6.47168994e-01 -2.47716874e-01
-4.90153462e-01 -7.73109019e-01 -1.27666080e+00 -7.06197441e-01
-6.08722121e-02 4.99692932e-02 3.22470874e-01 1.13541770e+00
2.91980267e-01 7.94164956e-01 1.48672208e-01 -7.88564205e-01
3.56248289e-01 -1.02482092e+00 -2.46986896e-01 2.58578569e-01
-4.08974528e-01 -5.15124857e-01 1.48856923e-01 5.04848659e-01] | [11.647880554199219, 8.878547668457031] |
51c5c35d-37fb-4cee-96a2-f06a35bf9be8 | learning-action-effect-dynamics-for | 2212.03866 | null | https://arxiv.org/abs/2212.03866v1 | https://arxiv.org/pdf/2212.03866v1.pdf | Learning Action-Effect Dynamics for Hypothetical Vision-Language Reasoning Task | 'Actions' play a vital role in how humans interact with the world. Thus, autonomous agents that would assist us in everyday tasks also require the capability to perform 'Reasoning about Actions & Change' (RAC). This has been an important research direction in Artificial Intelligence (AI) in general, but the study of RAC with visual and linguistic inputs is relatively recent. The CLEVR_HYP (Sampat et. al., 2021) is one such testbed for hypothetical vision-language reasoning with actions as the key focus. In this work, we propose a novel learning strategy that can improve reasoning about the effects of actions. We implement an encoder-decoder architecture to learn the representation of actions as vectors. We combine the aforementioned encoder-decoder architecture with existing modality parsers and a scene graph question answering model to evaluate our proposed system on the CLEVR_HYP dataset. We conduct thorough experiments to demonstrate the effectiveness of our proposed approach and discuss its advantages over previous baselines in terms of performance, data efficiency, and generalization capability. | ['Chitta Baral', 'Yezhou Yang', 'Pratyay Banerjee', 'Shailaja Keyur Sampat'] | 2022-12-07 | null | null | null | null | ['graph-question-answering'] | ['graphs'] | [ 3.52776259e-01 2.24170133e-01 2.80314013e-02 -5.89323819e-01
-2.94445395e-01 -5.99015832e-01 1.17039585e+00 -3.43729034e-02
-6.68970287e-01 4.93021846e-01 6.95816934e-01 -6.08692288e-01
2.54489362e-01 -8.65227938e-01 -8.49077284e-01 -2.86551028e-01
3.69620353e-01 3.68000060e-01 3.97098094e-01 -4.17994380e-01
2.54360616e-01 2.66106457e-01 -1.52230012e+00 7.64201760e-01
6.53300881e-01 6.91733599e-01 3.24165314e-01 7.36240745e-01
-1.77961409e-01 1.89904642e+00 -5.16449094e-01 -6.26388192e-01
-1.44700989e-01 -5.17836750e-01 -1.17107439e+00 6.32920414e-02
3.61852884e-01 -3.91961992e-01 -6.31673694e-01 9.71108258e-01
1.93322316e-01 3.15548003e-01 5.49578607e-01 -1.27373075e+00
-9.96356964e-01 7.81074345e-01 -2.85664320e-01 3.16902548e-01
5.85915625e-01 5.15364945e-01 1.06414473e+00 -4.41704243e-01
7.32212901e-01 1.68424582e+00 1.81429923e-01 7.25009024e-01
-7.47630119e-01 -8.78084749e-02 4.11590278e-01 7.79979527e-01
-8.66938889e-01 -6.10421062e-01 7.72500098e-01 -3.90026808e-01
1.32895029e+00 1.42766073e-01 3.87882978e-01 1.31211245e+00
8.92229825e-02 1.27180076e+00 7.70897269e-01 -5.09019494e-01
1.66028932e-01 -1.65760890e-02 1.69845551e-01 9.78021085e-01
7.92064965e-02 -8.29511210e-02 -5.34068525e-01 4.28136408e-01
7.79449046e-01 -1.64532125e-01 -1.19943015e-01 -4.84712958e-01
-1.59004700e+00 7.09267378e-01 8.77448618e-01 3.23703885e-01
-3.57933581e-01 7.09506810e-01 5.42718530e-01 2.75032848e-01
-2.21558567e-02 7.04023898e-01 -3.10608298e-01 -2.78083861e-01
-2.73695942e-02 1.57626182e-01 6.39415979e-01 8.91930401e-01
1.33615464e-01 5.16814739e-02 -5.09970129e-01 7.43338466e-01
5.13261139e-01 5.38246393e-01 2.62041360e-01 -1.11484337e+00
7.49830782e-01 8.43695045e-01 8.82268623e-02 -7.70704210e-01
-4.51527476e-01 5.66177629e-02 -6.24046564e-01 7.71645596e-03
3.18404049e-01 -9.26557481e-02 -8.46212924e-01 1.90527260e+00
1.66509762e-01 2.79992353e-02 4.95993644e-01 8.81927490e-01
1.21987295e+00 7.27292895e-01 4.80990320e-01 1.19409196e-01
1.47462451e+00 -1.19330275e+00 -9.04402435e-01 -5.48402905e-01
7.45018899e-01 -3.07710558e-01 1.32039773e+00 2.81918291e-02
-1.04365683e+00 -6.24467552e-01 -8.53949904e-01 -6.05446041e-01
-5.44813395e-01 3.20161313e-01 9.79872346e-01 1.84771419e-01
-1.06150222e+00 -7.13127106e-02 -9.06836271e-01 -6.71994448e-01
5.44127584e-01 -5.61387651e-02 -3.07869345e-01 -1.42353505e-01
-1.21578455e+00 1.27984548e+00 5.38241982e-01 -1.88589357e-02
-8.92795384e-01 -3.20510894e-01 -1.13679743e+00 -5.42907566e-02
6.55621588e-01 -9.53168213e-01 1.71696591e+00 -7.84728765e-01
-1.36247599e+00 8.31114948e-01 -1.37853429e-01 -7.68986762e-01
3.49255472e-01 -2.44598284e-01 -3.03057194e-01 2.21344352e-01
1.26780737e-02 9.60535288e-01 4.41679418e-01 -1.06717396e+00
-6.41704500e-01 -4.48631167e-01 9.65587735e-01 5.82059741e-01
1.00006841e-01 -3.85766737e-02 -5.47287643e-01 -3.36802363e-01
-2.11454183e-01 -7.40480423e-01 -1.26959994e-01 8.50710869e-02
-3.21279138e-01 -4.88076687e-01 6.46843553e-01 -6.31891787e-01
9.57505643e-01 -2.19809365e+00 3.98530453e-01 -4.45045143e-01
5.66432700e-02 2.78862596e-01 -2.39097893e-01 5.56247592e-01
2.53444105e-01 -1.16795428e-01 -2.47359440e-01 -1.24601096e-01
1.68265432e-01 2.36552730e-01 -3.77314508e-01 1.09161198e-01
2.70599902e-01 1.35657001e+00 -8.72814715e-01 -3.80628824e-01
6.07598245e-01 4.39947546e-01 -6.99620008e-01 3.30965221e-01
-8.20979416e-01 4.47819322e-01 -6.11490071e-01 5.27853727e-01
7.11797476e-02 -3.79106760e-01 3.08149397e-01 -3.44638884e-01
3.85081843e-02 3.66419226e-01 -6.18665636e-01 1.96683085e+00
-4.65547264e-01 7.78502882e-01 -3.32123131e-01 -1.01538944e+00
3.95278215e-01 1.80206135e-01 1.01832166e-01 -1.18071711e+00
1.29654422e-01 -3.52588683e-01 2.54263461e-01 -1.03314006e+00
2.44990408e-01 1.49915675e-02 -2.92875201e-01 2.74481446e-01
-6.78538810e-03 -2.17563719e-01 3.28541666e-01 5.42556703e-01
1.33217192e+00 5.07468641e-01 5.90201676e-01 2.33620822e-01
8.29886138e-01 3.48187536e-01 5.86264804e-02 6.62044168e-01
-4.14845347e-01 4.52728048e-02 6.15855098e-01 -4.43137437e-01
-6.99930429e-01 -9.24602449e-01 3.49408388e-01 1.34599054e+00
2.05458611e-01 -2.33370617e-01 -5.90212226e-01 -7.15090215e-01
-1.76048025e-01 1.38626254e+00 -7.13314950e-01 -3.68626058e-01
-4.61359650e-01 -3.13087016e-01 6.32968783e-01 8.20252478e-01
1.09335005e+00 -1.72409832e+00 -1.09931290e+00 -8.68317336e-02
-4.38806623e-01 -1.60789418e+00 1.54529080e-01 -1.79826468e-01
-4.11769778e-01 -1.32008064e+00 -7.54281208e-02 -7.31851697e-01
5.25876403e-01 3.94983530e-01 1.20001268e+00 -1.30118161e-01
-1.34077042e-01 7.98377573e-01 -6.02200747e-01 -5.67107558e-01
-4.78918552e-01 -3.10110539e-01 -4.20493096e-01 -2.70388067e-01
4.87521440e-01 -9.44421515e-02 -4.24986154e-01 -1.20068319e-01
-9.95403767e-01 6.30553961e-01 7.51312613e-01 5.13039708e-01
1.93790764e-01 1.16017302e-02 3.42266023e-01 -1.13372862e+00
7.01951444e-01 -3.53017420e-01 -5.46982527e-01 5.75022995e-01
8.78027380e-02 3.69319350e-01 6.14676535e-01 -4.87385057e-02
-1.61642849e+00 1.14035951e-02 -1.65456414e-01 -4.87482622e-02
-4.53206450e-01 6.22737527e-01 -4.61794376e-01 2.25971535e-01
5.79410374e-01 1.71417460e-01 -2.11260691e-01 -1.96502462e-01
9.38710272e-01 5.77571809e-01 6.92977846e-01 -5.81754863e-01
4.85852778e-01 5.30416489e-01 -2.34676898e-02 -7.00643778e-01
-9.55585241e-01 -3.88490230e-01 -4.48405027e-01 -3.01269501e-01
1.24492562e+00 -9.15394783e-01 -1.11089194e+00 2.93485194e-01
-1.34483087e+00 -5.20476341e-01 -1.20302863e-01 4.55642432e-01
-7.57456124e-01 1.82733625e-01 -4.27462339e-01 -6.24767005e-01
-7.01511875e-02 -1.17035162e+00 9.73834991e-01 2.41918206e-01
-8.01824331e-02 -9.97593939e-01 -8.44463333e-02 7.83607543e-01
2.36861199e-01 1.58509463e-01 1.13213193e+00 -4.88778830e-01
-6.79334700e-01 2.24601567e-01 -4.74646598e-01 1.73244983e-01
1.32475644e-01 -3.59932899e-01 -8.44524503e-01 -1.22794658e-01
-1.36182085e-01 -5.95794737e-01 9.83026683e-01 1.84382185e-01
1.27891636e+00 -8.45023096e-02 -1.91441014e-01 3.78707170e-01
1.38145983e+00 6.03485525e-01 9.07487929e-01 3.15481156e-01
8.78416121e-01 5.39917707e-01 6.36585116e-01 2.25981355e-01
8.60824943e-01 4.99921739e-01 7.16692030e-01 1.40888169e-01
-4.03538316e-01 -3.45135450e-01 6.62630677e-01 5.05454123e-01
-2.64813662e-01 -4.53118980e-01 -1.05439556e+00 4.91777658e-01
-1.99127018e+00 -1.21249592e+00 -7.21547082e-02 1.63230026e+00
5.53657413e-01 -4.11004014e-02 -2.16815233e-01 -1.71011016e-01
3.51751715e-01 3.41027230e-01 -6.87312305e-01 -5.52724302e-01
1.13744766e-01 -4.79025394e-02 1.46626279e-01 3.72647107e-01
-1.32653534e+00 1.50197268e+00 5.98255205e+00 6.45652041e-02
-7.01964617e-01 -1.44809932e-01 1.52222991e-01 2.08225414e-01
-1.55209392e-01 -1.49914429e-01 -4.38828677e-01 8.25814530e-02
8.71534884e-01 1.91159293e-01 6.16738498e-01 6.42417073e-01
1.54288888e-01 -2.82252520e-01 -1.28150904e+00 1.12582934e+00
5.78806162e-01 -1.27437139e+00 3.43256235e-01 -4.39300030e-01
3.40775996e-01 -2.30955984e-02 1.02731390e-02 5.85204840e-01
6.06684506e-01 -9.44829643e-01 5.98602891e-01 6.02309465e-01
5.94757497e-01 -4.22505260e-01 5.54339349e-01 3.59553725e-01
-1.05970752e+00 -2.83911854e-01 -3.32760304e-01 -2.46135116e-01
1.67427033e-01 -9.55708027e-02 -9.30847764e-01 4.89450037e-01
5.29730499e-01 9.27095234e-01 -7.98505962e-01 6.41225994e-01
-7.48310745e-01 4.94155198e-01 2.44954929e-01 -2.70056605e-01
2.29432374e-01 -1.14593491e-01 3.21704954e-01 9.24524546e-01
-4.70304452e-02 5.28461039e-01 -5.69455922e-02 7.76067734e-01
-1.89297616e-01 -1.05885841e-01 -8.47985685e-01 -5.68074286e-01
1.63885310e-01 7.23270118e-01 -5.09672105e-01 -4.42293525e-01
-9.83959138e-01 1.01632667e+00 5.18216729e-01 4.30209428e-01
-1.07615674e+00 -1.13033906e-01 6.40477657e-01 -2.53361255e-01
2.74838060e-01 -3.65434557e-01 1.87141746e-02 -1.28083587e+00
-1.24178931e-01 -1.11329162e+00 4.93155599e-01 -1.29839301e+00
-1.10076964e+00 4.05242234e-01 1.26725078e-01 -9.38867748e-01
-2.33209819e-01 -9.89343107e-01 -1.49607986e-01 2.87463248e-01
-1.35065031e+00 -1.60475147e+00 -4.33614820e-01 7.34463215e-01
9.93662357e-01 -1.81813329e-01 7.37876654e-01 -6.66476637e-02
-3.72903734e-01 1.57699227e-01 -5.33812404e-01 3.32951635e-01
3.88854146e-01 -1.23748147e+00 5.51555753e-01 9.90224957e-01
4.44593728e-01 4.71298218e-01 5.49817324e-01 -5.97472608e-01
-1.80582118e+00 -1.11729383e+00 6.83071256e-01 -7.32191682e-01
6.19799733e-01 -1.95224538e-01 -6.35077536e-01 1.31873405e+00
6.12238228e-01 -3.57017845e-01 4.80513573e-01 -4.09971625e-02
-4.63200688e-01 1.74762323e-01 -9.14440274e-01 9.76643384e-01
1.34851742e+00 -6.35820806e-01 -1.17533433e+00 3.93026322e-01
1.01984346e+00 -4.26981360e-01 -4.52053517e-01 3.58533621e-01
2.13351145e-01 -9.06765461e-01 1.06305087e+00 -1.05816269e+00
6.06595218e-01 -5.31101763e-01 -3.50241929e-01 -1.24606860e+00
-3.43757808e-01 -3.57780308e-02 -1.40800640e-01 9.37338531e-01
2.62840867e-01 -4.33007061e-01 2.68801212e-01 5.58055043e-01
-1.60715744e-01 -2.56516814e-01 -4.22606558e-01 -2.72959232e-01
-2.92937577e-01 -6.56966448e-01 5.65168977e-01 7.64400363e-01
1.35854650e-02 8.33128989e-01 -2.01967672e-01 3.02019298e-01
1.90971404e-01 2.19081435e-02 8.54021966e-01 -9.41329718e-01
-2.52331316e-01 -2.24537432e-01 -5.12708843e-01 -1.10966027e+00
3.88486534e-01 -9.55653429e-01 3.59200947e-02 -2.31656551e+00
4.02309209e-01 1.37344033e-01 -2.06152976e-01 7.99637914e-01
-7.75443241e-02 -2.11299792e-01 3.90740603e-01 -8.25643763e-02
-1.01429415e+00 5.43724835e-01 1.51026320e+00 -3.34878504e-01
3.17892767e-02 -3.95148128e-01 -7.73343623e-01 7.98940420e-01
8.55326295e-01 2.41491616e-01 -9.24591064e-01 -1.13014567e+00
4.06655550e-01 5.91658093e-02 5.24055302e-01 -9.89920318e-01
2.67683595e-01 -3.88465613e-01 2.01619759e-01 -3.90002310e-01
5.34368396e-01 -8.26864421e-01 -2.72057712e-01 3.84140372e-01
-6.92639053e-01 2.01004088e-01 2.42954820e-01 5.55595040e-01
-3.05410534e-01 1.59983531e-01 5.39348602e-01 -2.94603735e-01
-1.58041513e+00 -2.56213639e-02 -3.46803695e-01 6.31458759e-02
1.22147143e+00 2.17230439e-01 -7.90044904e-01 -6.44168079e-01
-4.49193478e-01 3.68418485e-01 2.59939134e-01 6.98570609e-01
8.58425736e-01 -1.07608867e+00 -5.17182112e-01 -5.59156425e-02
4.73668426e-01 -2.32909933e-01 1.49856657e-01 6.05033159e-01
-7.71844447e-01 7.30200350e-01 -3.69571924e-01 -2.75549114e-01
-1.10724390e+00 8.46672773e-01 2.87955403e-01 9.07715689e-03
-6.42700553e-01 8.89962018e-01 3.72871578e-01 -4.58812594e-01
2.93948770e-01 -4.61484283e-01 -7.72161603e-01 -2.58596867e-01
5.55601478e-01 5.85918389e-02 -4.24420953e-01 -7.93762147e-01
-4.36014652e-01 1.99911103e-01 7.21058175e-02 -2.21602038e-01
1.13563967e+00 -2.33605299e-02 -1.80068970e-01 4.66736376e-01
8.59969497e-01 -4.62818682e-01 -9.20005977e-01 -2.58481890e-01
-6.14899918e-02 -3.16276073e-01 -7.05431253e-02 -1.12177014e+00
-7.95668304e-01 1.06278479e+00 3.11773092e-01 -3.70441750e-02
1.06297743e+00 3.13215435e-01 4.16475952e-01 8.21629524e-01
4.46329176e-01 -9.56471443e-01 3.69227082e-01 7.10877478e-01
1.12776113e+00 -1.42138207e+00 -1.67703509e-01 -4.71025884e-01
-1.20802104e+00 7.83309877e-01 9.97868299e-01 2.14490265e-01
1.91605046e-01 -2.62789160e-01 1.53462201e-01 -4.28675503e-01
-1.08180666e+00 -7.15060413e-01 1.85443625e-01 5.97094297e-01
5.44953704e-01 1.83893785e-01 -1.66716158e-01 7.71381408e-02
-5.87507598e-02 5.78254871e-02 3.64138454e-01 1.06636798e+00
-3.05453062e-01 -8.39936018e-01 -8.75915363e-02 3.79803896e-01
2.10292321e-02 -5.38395159e-02 -7.52742112e-01 9.29070652e-01
1.37231603e-01 1.14334309e+00 1.46228358e-01 -1.58864558e-01
5.19987464e-01 2.10644081e-01 7.93406069e-01 -6.72076166e-01
-2.47942954e-01 -5.87186396e-01 3.36946994e-01 -9.19671476e-01
-8.33080471e-01 -5.79907358e-01 -1.55441546e+00 -1.85159910e-02
2.97400624e-01 -3.17207009e-01 5.28547943e-01 1.12393880e+00
2.28784814e-01 1.03098011e+00 5.11935167e-02 1.10059953e-03
-4.02832240e-01 -9.12164032e-01 -4.84957434e-02 7.51268864e-01
5.33940569e-02 -7.35446751e-01 -2.54860464e-02 3.36386025e-01] | [10.756982803344727, 1.6076182126998901] |
597a16bf-e9c8-47ff-a374-652221213057 | a-single-camera-3d-scanning-velocimetry | 2102.05787 | null | https://arxiv.org/abs/2102.05787v1 | https://arxiv.org/pdf/2102.05787v1.pdf | A single-camera, 3D scanning velocimetry system for quantifying active particle aggregations | A three-dimensional (3D) scanning velocimetry system is developed to quantify the 3D configurations of particles and their surrounding volumetric, three-component velocity fields. The approach uses a translating laser sheet to rapidly scan through a volume of interest and sequentially illuminate slices of the flow containing both tracers seeded in the fluid and particles comprising the aggregation of interest. These image slices are captured by a single high-speed camera, encoding information about the third spatial dimension within the image time-series. Where previous implementations of scanning systems have been developed for either volumetric flow quantification or 3D object reconstruction, we evaluate the feasibility of accomplishing these tasks concurrently with a single-camera, which can streamline data collection and analysis. The capability of the system was characterized using a study of induced vertical migrations of millimeter-scale brine shrimp (Artemia salina). Identification and reconstruction of individual swimmer bodies and 3D trajectories within the migrating aggregation were achieved up to the maximum number density studied presently, $8 \, \times\,10^5$ animals per $\textrm{m}^3$. This number density is comparable to the densities of previous depth-averaged 2D measurements of similar migrations. Corresponding velocity measurements of the flow indicate that the technique is capable of resolving the 3D velocity field in and around the swimming aggregation. At these animal number densities, instances of coherent flow induced by the migrations were observed. The accuracy of these flow measurements was confirmed in separate studies of a free jet at $Re_D = 50$. | ['John O. Dabiri', 'Isabel A. Houghton', 'Matt K. Fu'] | 2021-02-11 | null | null | null | null | ['3d-object-reconstruction'] | ['computer-vision'] | [-9.59235337e-03 -5.78667819e-01 8.28372836e-01 3.16935956e-01
-9.19916332e-02 -7.61255383e-01 4.07955319e-01 1.80249229e-01
-1.04420233e+00 7.78190017e-01 -2.94149220e-01 -3.53039235e-01
-1.36565492e-01 -7.08525479e-01 -2.87219316e-01 -9.03352499e-01
-7.09918439e-01 6.26564205e-01 4.79688466e-01 -6.80690706e-02
4.58498955e-01 7.43357182e-01 -1.59626031e+00 -2.02118695e-01
3.21413696e-01 7.93855429e-01 3.23004931e-01 1.26797664e+00
-3.23005229e-01 3.11549455e-01 -5.27283370e-01 -2.55391747e-02
3.04285020e-01 -6.73616111e-01 -4.65211123e-01 5.38020730e-02
2.66602904e-01 -6.02011561e-01 2.52579510e-01 6.40214741e-01
4.70071524e-01 3.48883748e-01 7.88716495e-01 -5.54152966e-01
2.94254124e-01 -6.01700060e-02 -6.51818752e-01 9.26952004e-01
2.68111229e-01 3.59092951e-01 4.41629142e-01 -6.79261804e-01
7.81522214e-01 1.01424623e+00 7.13118553e-01 5.71707487e-01
-1.19807804e+00 -5.33455551e-01 -5.10703087e-01 -7.96104133e-01
-1.07428885e+00 -3.88138771e-01 3.59094441e-01 -9.07338798e-01
9.53801095e-01 3.01296145e-01 1.09386718e+00 2.73220986e-01
7.58667171e-01 -2.31531993e-01 1.11769331e+00 -1.70893565e-01
5.77807784e-01 -4.17761914e-02 -1.66555494e-01 6.01775944e-01
8.19253445e-01 2.74914235e-01 -4.17499304e-01 -2.92739600e-01
1.18617046e+00 -3.11127782e-01 -3.44321340e-01 -2.26608351e-01
-8.40393245e-01 5.32926261e-01 5.61601557e-02 2.29504213e-01
-2.62823403e-02 2.08788425e-01 2.85380751e-01 -2.77203321e-01
6.10798120e-01 6.12479091e-01 -5.62232077e-01 -7.67132699e-01
-8.84583473e-01 5.03690302e-01 1.08563888e+00 3.91455054e-01
4.86022860e-01 1.42194778e-01 6.58451498e-01 1.75859317e-01
5.87552786e-01 1.04178047e+00 4.96831954e-01 -1.18074775e+00
3.91507596e-01 7.22592294e-01 4.66115266e-01 -6.35588348e-01
-5.24345577e-01 6.24502264e-02 -2.56208599e-01 7.04705179e-01
7.50552595e-01 -6.93480134e-01 -6.05057120e-01 1.00780511e+00
5.53991139e-01 8.32152441e-02 1.94640353e-01 1.08949852e+00
6.29160643e-01 5.30840933e-01 -1.60452172e-01 -7.08485782e-01
1.38069463e+00 9.58290547e-02 -3.02922726e-01 -6.00518957e-02
7.45025337e-01 -5.33923209e-01 6.39403939e-01 -1.77344978e-01
-1.24266613e+00 -1.78357854e-01 -8.40926766e-01 6.23631179e-01
-1.98933437e-01 -4.42435265e-01 4.32393432e-01 6.83186769e-01
-7.67904460e-01 8.23900044e-01 -1.47940958e+00 -2.72628337e-01
3.52178454e-01 1.32352158e-01 -2.98731059e-01 4.24027026e-01
-4.85632986e-01 5.99820971e-01 -4.36569661e-01 1.16907865e-01
-6.96340084e-01 -8.76328409e-01 -8.10464680e-01 -3.82669121e-01
-6.60218954e-01 -5.65516651e-01 1.24406147e+00 -1.77094012e-01
-1.37985003e+00 8.66549551e-01 -3.50470573e-01 -2.40362197e-01
5.61124444e-01 -1.29160538e-01 -6.17784373e-02 6.03259385e-01
3.95260334e-01 1.35250673e-01 1.41775981e-01 -1.34514642e+00
-7.84688413e-01 -5.08470178e-01 3.40401679e-02 3.60043764e-01
1.82320088e-01 1.99632645e-02 3.68527979e-01 1.28901936e-02
2.52880007e-01 -7.78599143e-01 -4.09264386e-01 -2.84931175e-02
1.61555350e-01 5.33656716e-01 9.31015670e-01 -4.74793091e-02
7.58021832e-01 -1.86719239e+00 -1.61439747e-01 -3.15368809e-02
-1.00930333e-01 3.62038672e-01 3.88888627e-01 7.29326129e-01
5.69485545e-01 1.86700031e-01 -4.95920002e-01 -2.74252594e-01
-5.25081277e-01 2.90275633e-01 -1.26765728e-01 1.03279960e+00
-1.36245385e-01 3.61028254e-01 -1.14508176e+00 -4.63832319e-01
1.70669138e-01 5.41866720e-01 -4.51067567e-01 2.39251554e-01
2.46571645e-01 7.09839761e-01 -4.02011782e-01 5.11619389e-01
8.16528499e-01 -1.29280556e-02 8.93745497e-02 2.15985388e-01
-1.09717488e+00 -1.85356483e-01 -1.19182134e+00 1.15820491e+00
-2.70184815e-01 6.06139243e-01 7.78928339e-01 -4.62437391e-01
8.73233676e-01 2.57868618e-01 8.08513403e-01 -5.05069613e-01
2.86732852e-01 3.77059758e-01 1.35725960e-01 -8.28993380e-01
6.00198209e-01 -9.59293246e-01 1.47222430e-01 6.26353741e-01
-9.51040834e-02 -4.80559617e-01 3.54134530e-01 9.85954255e-02
9.77796137e-01 1.45953581e-01 1.45417741e-02 -1.00534439e+00
1.44290000e-01 2.64837861e-01 2.99167931e-01 4.62563097e-01
-2.93122828e-01 5.45305789e-01 2.33616918e-01 -6.82024598e-01
-9.91033435e-01 -1.08407760e+00 -8.01014602e-01 6.00388825e-01
8.73489738e-01 -2.26681486e-01 -6.29488230e-01 5.01591293e-03
2.26801768e-01 -9.71036777e-02 -6.94579422e-01 3.96160156e-01
-7.97789156e-01 -1.05907464e+00 4.23711121e-01 3.50424111e-01
4.23813254e-01 -9.75930929e-01 -1.78101242e+00 3.97702485e-01
2.92648673e-01 -8.49123299e-01 3.27767551e-01 1.76519617e-01
-1.31464744e+00 -1.44018269e+00 -5.62031686e-01 -3.97547781e-01
7.80923903e-01 1.96687341e-01 8.72803688e-01 1.45109504e-01
-6.90409243e-01 5.26485801e-01 -5.27623534e-01 -2.79465020e-01
-2.81664371e-01 -6.55837834e-01 3.08379799e-01 -5.10254264e-01
2.06599131e-01 -5.36395490e-01 -1.00961936e+00 3.19176704e-01
-6.57141268e-01 -5.00094891e-01 -1.86773032e-01 5.57601333e-01
2.69791514e-01 -1.51372522e-01 8.46950859e-02 -4.05404389e-01
2.18760312e-01 -1.91443413e-01 -8.75461340e-01 -6.38872802e-01
1.65051997e-01 -2.53129780e-01 6.42890155e-01 -1.82152018e-01
-8.76735628e-01 -1.73911929e-01 -1.89566329e-01 8.78943428e-02
-4.09853965e-01 3.54143560e-01 5.83235145e-01 -2.19751969e-01
7.61408448e-01 2.49238431e-01 2.74932444e-01 -4.30430412e-01
-3.25726241e-01 3.90905708e-01 2.66875923e-01 -2.45989010e-01
6.28843963e-01 1.16741514e+00 3.76650989e-01 -1.39432085e+00
-1.47995219e-01 -5.98482013e-01 -7.90030539e-01 -4.75581199e-01
7.69098222e-01 -7.20667362e-01 -1.08031321e+00 4.53240931e-01
-7.00131774e-01 -6.82802200e-01 -4.67074722e-01 1.00982237e+00
-2.68631458e-01 3.40202928e-01 -8.27560723e-01 -1.28192246e+00
-1.72042072e-01 -1.23140097e+00 1.06263912e+00 3.98921281e-01
-5.35685495e-02 -1.36529410e+00 6.08238578e-01 -1.10822052e-01
6.76749527e-01 5.16532063e-01 3.15018535e-01 -1.48791922e-02
-2.29219347e-01 -1.85470045e-01 1.42664924e-01 -3.31349611e-01
1.75964043e-01 3.69150758e-01 -7.15127766e-01 -4.93222147e-01
1.95688173e-01 -5.82805462e-02 6.99982047e-01 8.87789786e-01
-1.35111660e-02 1.12027489e-02 -5.83048403e-01 6.40008628e-01
1.58025467e+00 3.31785709e-01 2.25034803e-01 3.62203836e-01
2.01661319e-01 6.38549149e-01 3.71246785e-01 8.09092224e-01
2.90948749e-02 3.66762429e-01 5.07407665e-01 -5.81571981e-02
8.57331455e-02 3.15584123e-01 2.41192207e-01 4.96035904e-01
-7.89691806e-01 -1.96973965e-01 -9.22619998e-01 5.08407712e-01
-1.04339385e+00 -1.00107360e+00 -5.76378822e-01 2.24335289e+00
3.84856939e-01 -5.97540587e-02 1.41130224e-01 4.17035595e-02
4.43545848e-01 -9.82441232e-02 -2.69153833e-01 -3.71684551e-01
2.92957872e-01 4.06169564e-01 8.71115565e-01 1.09620643e+00
-8.80095303e-01 3.47731203e-01 7.13940907e+00 -2.73014307e-01
-1.31915665e+00 -2.42934570e-01 -1.94767453e-02 -2.70632178e-01
-2.57648796e-01 5.67209646e-02 -1.17335415e+00 6.38868988e-01
9.27010059e-01 -1.72883973e-01 6.77265301e-02 2.18400136e-01
4.94698733e-01 -7.47978270e-01 -4.77060676e-01 3.76220733e-01
-1.75337747e-01 -1.37815201e+00 -1.18015192e-01 4.69383538e-01
4.17865902e-01 9.82323810e-02 -3.67226094e-01 -2.53727287e-01
2.67106056e-01 -5.38235724e-01 6.65187597e-01 3.37839246e-01
9.90543008e-01 -4.50410813e-01 9.49624181e-01 7.26911724e-01
-1.51617312e+00 -2.89001223e-02 -4.73650008e-01 -8.56984496e-01
8.05613875e-01 5.23174345e-01 -7.86697984e-01 1.82522431e-01
8.18325937e-01 3.97079796e-01 -2.30483189e-01 1.08655882e+00
4.81811076e-01 4.84861434e-01 -6.97158456e-01 -3.51526439e-01
3.69244456e-01 -6.83014035e-01 8.00674140e-01 1.30855107e+00
5.75318336e-01 5.77408969e-01 -3.26082051e-01 7.33933270e-01
3.73738170e-01 -1.95979148e-01 -5.94043195e-01 5.86142950e-02
2.65869081e-01 1.14235210e+00 -1.32194495e+00 -2.03802839e-01
2.29144539e-03 1.83602318e-01 -2.62468845e-01 1.40151203e-01
-4.55707222e-01 -4.72685903e-01 1.04758811e+00 4.89347339e-01
3.84822279e-01 -5.89776218e-01 -1.05424233e-01 -9.79106963e-01
-3.22316259e-01 3.59772682e-01 8.09996352e-02 -4.30307597e-01
-9.90030825e-01 4.54841942e-01 2.11887598e-01 -1.30886102e+00
-1.04153909e-01 -9.28389132e-01 -7.10588694e-01 8.68107677e-01
-1.06817067e+00 -3.75514507e-01 -4.01615143e-01 -6.67323067e-04
2.05409855e-01 1.28196359e-01 8.48441601e-01 7.13063031e-03
-1.70367956e-01 -1.82685107e-01 3.61680627e-01 -6.81199133e-03
7.81673267e-02 -1.26781547e+00 1.05922908e-01 8.49685013e-01
-2.81039476e-01 5.21053672e-01 1.03844333e+00 -9.25597191e-01
-1.71656179e+00 -7.79448032e-01 5.36498368e-01 -5.57897568e-01
6.58375084e-01 -3.75679940e-01 -7.76246071e-01 2.38594845e-01
1.73406992e-02 4.42977577e-01 9.12698209e-01 -6.43788278e-01
6.31324470e-01 1.98050752e-01 -1.32641733e+00 1.33881629e-01
9.17805910e-01 3.40514407e-02 -3.94925475e-01 3.06839105e-02
-5.04775681e-02 -6.31134331e-01 -7.94287384e-01 4.76684183e-01
8.82291555e-01 -1.17077363e+00 8.77475500e-01 -3.91262740e-01
3.74245763e-01 -4.30259883e-01 -3.30968276e-02 -1.10062826e+00
6.05584048e-02 -2.08797887e-01 4.78427172e-01 8.82887840e-01
2.21516132e-01 -6.42007887e-01 8.74633014e-01 3.73530567e-01
-2.33786568e-01 -3.89314383e-01 -1.32595384e+00 -6.19011581e-01
3.57873887e-01 -1.02906406e-01 -8.06610286e-02 4.97415811e-01
4.61035371e-01 -6.66279122e-02 2.30433211e-01 4.36338067e-01
7.87600815e-01 3.28065425e-01 4.79410857e-01 -1.30929589e+00
-5.34183532e-02 -9.46879908e-02 -7.70473838e-01 -7.88786113e-01
-3.22458595e-01 -3.66075575e-01 2.91205883e-01 -1.60950398e+00
-2.00854912e-01 -3.52529973e-01 6.45259738e-01 -3.42003316e-01
2.21928507e-01 4.65362072e-01 -1.07119896e-01 3.49618345e-01
-1.53915361e-01 1.36530310e-01 1.36107099e+00 5.54701984e-01
-5.57553887e-01 -1.37828037e-01 8.97464603e-02 6.58258736e-01
5.10496497e-01 -2.74669260e-01 -1.42148510e-01 -4.74175453e-01
2.49814689e-02 2.54845738e-01 2.30954468e-01 -1.31282532e+00
3.37979883e-01 -1.23149149e-01 4.81273264e-01 -5.04861772e-01
5.83251834e-01 -7.11100161e-01 3.03173780e-01 9.56152022e-01
1.08399644e-01 7.49622658e-02 4.43760991e-01 4.71652418e-01
-5.27465492e-02 -3.47587585e-01 1.10912728e+00 -8.40836704e-01
-4.21028286e-01 8.23464841e-02 -9.32897806e-01 2.97701955e-01
1.16666591e+00 -5.62676847e-01 -5.96230686e-01 2.59940714e-01
-4.13213313e-01 -1.99736536e-01 9.49620187e-01 -6.11309588e-01
7.74267912e-01 -6.15726411e-01 -5.40732741e-01 4.53699142e-01
-4.11688656e-01 3.67022395e-01 3.49266052e-01 8.33368838e-01
-1.62350059e+00 6.93516359e-02 -4.12290126e-01 -7.64227986e-01
-1.11620533e+00 7.40271574e-03 6.67190909e-01 1.56960502e-01
-9.24053431e-01 1.22130668e+00 1.32963151e-01 1.03017539e-01
-5.32231271e-01 -3.10137451e-01 -3.95935953e-01 1.13722667e-01
7.74731338e-01 6.90732181e-01 -1.98373020e-01 -9.01261151e-01
-4.57550675e-01 1.15452695e+00 6.28521323e-01 -2.95115024e-01
1.23361850e+00 -3.21424901e-01 -2.50457805e-02 6.48416817e-01
1.05989707e+00 1.82432443e-01 -1.86183691e+00 6.75237000e-01
-6.11284673e-01 -6.62215889e-01 -1.99907646e-01 -2.29940653e-01
-8.29103112e-01 1.04609466e+00 3.32746357e-01 6.32566929e-01
5.69043756e-01 2.12724879e-01 4.96029228e-01 -2.74568379e-01
4.81090218e-01 -5.04770279e-01 -2.05050051e-01 4.81894761e-01
1.64195001e-01 -8.14959824e-01 1.72985733e-01 -3.03861737e-01
-1.06487997e-01 1.26179123e+00 4.60629672e-01 -3.80079806e-01
6.43868923e-01 1.01636291e+00 4.75391626e-01 -4.98191744e-01
-6.31800890e-01 -7.56855831e-02 -5.28614283e-01 5.84689975e-01
3.36823732e-01 -1.17923640e-01 -4.09567535e-01 -2.71532256e-02
-3.09455007e-01 -1.10889435e-01 1.04160893e+00 1.38734448e+00
-7.99953043e-01 -4.69457299e-01 -4.10941005e-01 3.43879253e-01
-5.34358382e-01 3.40265214e-01 3.42389755e-02 8.88851881e-01
1.02686204e-01 6.73956454e-01 6.66524887e-01 -6.17296956e-02
5.47395468e-01 -1.93348557e-01 4.54342335e-01 -3.95193458e-01
-4.39053982e-01 1.02417849e-01 -1.95571687e-02 -3.54748309e-01
-9.28923190e-01 -9.45938408e-01 -1.55671203e+00 -1.31164715e-01
-1.72213286e-01 7.58889198e-01 6.97073221e-01 7.30765045e-01
-1.10652559e-01 1.32282495e-01 4.88231421e-01 -1.53724515e+00
6.80926219e-02 -8.85166526e-01 -1.35131717e+00 2.47081131e-01
4.21703339e-01 -8.31317186e-01 -1.21819103e+00 4.23424065e-01] | [13.315694808959961, -3.0269155502319336] |
c390fdac-1bdc-4ff3-a263-4939c2a84819 | machine-learning-students-overfit-to | 2209.03032 | null | https://arxiv.org/abs/2209.03032v1 | https://arxiv.org/pdf/2209.03032v1.pdf | Machine Learning Students Overfit to Overfitting | Overfitting and generalization is an important concept in Machine Learning as only models that generalize are interesting for general applications. Yet some students have trouble learning this important concept through lectures and exercises. In this paper we describe common examples of students misunderstanding overfitting, and provide recommendations for possible solutions. We cover student misconceptions about overfitting, about solutions to overfitting, and implementation mistakes that are commonly confused with overfitting issues. We expect that our paper can contribute to improving student understanding and lectures about this important topic. | ['Matthia Sabatelli', 'Matias Valdenegro-Toro'] | 2022-09-07 | null | null | null | null | ['misconceptions'] | ['miscellaneous'] | [-2.95269817e-01 2.52623975e-01 -1.04729503e-01 -7.00382352e-01
-4.33890581e-01 -4.72697258e-01 -3.50139588e-01 5.54179251e-01
-2.97192596e-02 7.80041575e-01 -1.23348042e-01 -9.02198493e-01
-2.78319448e-01 -8.35264742e-01 -9.15544748e-01 -3.84393632e-01
2.65794188e-01 1.66766951e-03 1.79009795e-01 -6.38079703e-01
6.51609182e-01 6.45254135e-01 -1.59676790e+00 4.52299625e-01
1.36779392e+00 1.84989870e-01 -1.71249256e-01 7.37466037e-01
-5.36243021e-01 9.45405841e-01 -1.16119146e+00 -6.10111475e-01
-3.71257275e-01 -1.15165301e-01 -1.16503501e+00 1.86142340e-01
9.99148130e-01 7.26326853e-02 -1.17311463e-01 9.08556223e-01
1.17602542e-01 3.99419516e-01 2.20810920e-01 -1.44987309e+00
-4.34699982e-01 4.83977377e-01 -3.47057611e-01 3.70764166e-01
3.59265596e-01 -8.31592008e-02 5.31574249e-01 -3.55492473e-01
-2.38956720e-01 9.15592074e-01 1.34637117e+00 8.24058831e-01
-1.16853452e+00 -6.58083498e-01 3.49437028e-01 -6.86652437e-02
-1.17617488e+00 1.33849040e-01 4.70540494e-01 -6.04982138e-01
8.73087347e-01 7.58284271e-01 9.78685915e-01 6.79835916e-01
4.52008516e-01 8.12844455e-01 9.54218030e-01 -5.18034875e-01
-1.11839190e-01 1.06615996e+00 7.08473086e-01 6.04292810e-01
7.03946352e-01 -1.18467532e-01 -1.09269202e-01 -8.91492441e-02
7.30947554e-01 2.60631561e-01 -2.48119965e-01 -2.39159942e-01
-5.39165199e-01 9.98392999e-01 3.80734056e-01 3.39641124e-01
2.84773797e-01 5.64113148e-02 7.27739707e-02 5.14882505e-01
1.20783756e-02 8.90913486e-01 -7.39958107e-01 -2.98275322e-01
-7.05445409e-01 4.26516712e-01 1.19235289e+00 1.16586351e+00
7.06647396e-01 2.19336048e-01 7.40863681e-01 9.20695901e-01
-2.50368975e-02 7.07911029e-02 6.89983785e-01 -8.21436167e-01
1.51423946e-01 8.30747664e-01 -1.11800067e-01 -8.82816195e-01
-5.40469348e-01 -1.01253271e+00 -3.79456580e-01 2.49646574e-01
4.44469035e-01 -2.81090707e-01 -5.92798710e-01 1.43182206e+00
-2.05087587e-01 1.47860751e-01 -2.63502337e-02 2.26367012e-01
1.29484355e+00 4.53946441e-01 3.74866813e-01 5.92881292e-02
1.17936420e+00 -8.96756351e-01 -8.23559344e-01 -3.60619783e-01
1.20837295e+00 -1.16335428e+00 1.25543416e+00 8.16953659e-01
-1.24724114e+00 -6.29478991e-01 -1.06787765e+00 -3.96684594e-02
-6.09354317e-01 -2.76312798e-01 7.78394222e-01 1.18920302e+00
-5.99720359e-01 1.06296587e+00 -5.57178140e-01 -5.65176308e-01
2.87394702e-01 5.39692283e-01 4.55685556e-02 -1.12831019e-01
-8.28535020e-01 1.17631817e+00 3.11727166e-01 -2.97476888e-01
-5.58885001e-02 -1.14435887e+00 -6.58604920e-01 2.33963594e-01
3.29620242e-01 -6.63955450e-01 1.54397440e+00 -6.00500941e-01
-1.06964779e+00 5.29949963e-01 5.21321520e-02 -1.07920684e-01
2.07800657e-01 -3.83301854e-01 -5.61975837e-01 -6.40827596e-01
-4.55655992e-01 2.90596038e-01 -3.27846110e-02 -9.21985626e-01
-6.02539062e-01 -1.79521874e-01 6.75536469e-02 1.68193251e-01
-6.48928046e-01 -3.60986352e-01 3.08286846e-01 -6.53621614e-01
3.94881994e-01 -6.08589232e-01 -4.16486382e-01 -3.52125376e-01
-7.25435987e-02 -4.31709677e-01 8.34081292e-01 -1.82657659e-01
1.67929447e+00 -1.69936574e+00 -7.51735449e-01 1.71770066e-01
1.80956826e-01 2.20747858e-01 1.26887456e-01 4.99734074e-01
-4.62965697e-01 4.16284770e-01 3.73255342e-01 2.29478449e-01
-1.15430966e-01 4.59618837e-01 -3.86991709e-01 1.48645699e-01
-9.44291577e-02 6.95188105e-01 -9.24264848e-01 -1.87709674e-01
3.00469905e-01 4.49073255e-01 -4.87240195e-01 1.35622859e-01
2.78034866e-01 -1.95009988e-02 -3.06831986e-01 4.78961349e-01
7.06045389e-01 -1.50575653e-01 -1.72373995e-01 1.66814163e-01
-1.22311421e-01 1.04187214e+00 -1.33227623e+00 1.16499424e+00
-6.13193214e-01 7.96990275e-01 -5.38447082e-01 -1.30978167e+00
1.12525499e+00 3.31972629e-01 1.23443067e-01 9.21731293e-02
-7.23714754e-02 3.25512379e-01 -3.85978967e-02 -8.23147833e-01
4.18573946e-01 -5.62142909e-01 2.11734697e-01 4.09598857e-01
8.71072784e-02 -6.07127607e-01 -9.69200656e-02 8.01323950e-02
6.61687970e-01 -1.53244913e-01 4.52725232e-01 -7.59010851e-01
2.76700795e-01 8.63685608e-02 3.32173407e-01 9.07539248e-01
2.52583288e-02 3.91461194e-01 3.24895918e-01 -9.95672524e-01
-5.80527544e-01 -8.54072809e-01 -1.60617962e-01 1.03179634e+00
-3.71765137e-01 -6.12264574e-01 -6.66393936e-01 -7.09003448e-01
-9.19747800e-02 1.03634810e+00 -4.97286946e-01 -3.99680555e-01
-4.57743019e-01 -5.15861571e-01 2.18598619e-01 8.37861300e-01
1.52035549e-01 -6.67512596e-01 -4.21672553e-01 3.04784834e-01
-6.98061362e-02 -7.55331695e-01 -6.76825643e-02 4.38381404e-01
-1.80042207e+00 -1.00255334e+00 -4.01121497e-01 -9.99285400e-01
1.09602141e+00 6.06936395e-01 1.52252603e+00 9.69949007e-01
-5.74547112e-01 5.98758578e-01 1.63995668e-01 -8.58250797e-01
-5.64449191e-01 1.74789608e-01 -8.98549259e-02 -1.25448751e+00
1.01434660e+00 -6.64422631e-01 -2.97302097e-01 2.73656636e-01
-7.66636193e-01 -1.70336947e-01 3.18375260e-01 6.00228071e-01
-3.75715718e-02 3.12277555e-01 3.63399297e-01 -1.43089902e+00
7.85914242e-01 -2.28673339e-01 -4.56937790e-01 3.01567256e-01
-9.19741452e-01 6.93407133e-02 4.72192347e-01 -4.75457549e-01
-7.34406292e-01 -3.47625017e-01 -4.08325970e-01 1.40699700e-01
-6.14176989e-01 2.38077924e-01 -3.27620618e-02 -7.46276081e-01
1.12725461e+00 -3.05900961e-04 -7.23856091e-02 -4.99459028e-01
-4.08201814e-01 4.11472321e-01 -6.36890158e-02 -7.95188189e-01
5.74923754e-01 -2.45170712e-01 2.27776729e-03 -1.14507926e+00
-1.25856805e+00 -4.48819280e-01 -2.73803085e-01 -2.26785652e-02
-1.36127034e-02 -7.03004897e-01 -9.05022502e-01 2.13463996e-02
-8.89966726e-01 -1.25979200e-01 -5.15455723e-01 6.07613146e-01
-4.79396701e-01 1.25419036e-01 -4.54289556e-01 -8.22026372e-01
1.99220866e-01 -1.36251581e+00 5.10169426e-04 8.06412220e-01
-8.15792799e-01 -1.73424780e+00 6.76165745e-02 7.14343011e-01
5.47431231e-01 -9.12603214e-02 8.28096330e-01 -9.36803460e-01
-1.12782732e-01 -5.48080504e-01 5.44199586e-01 4.76053655e-01
1.77008465e-01 1.67170778e-01 -1.22246015e+00 -2.54146039e-01
1.51184216e-01 -4.27502751e-01 6.47398651e-01 2.50567973e-01
1.71636593e+00 -7.42758453e-01 -4.06372637e-01 3.52057010e-01
1.50920236e+00 1.57083020e-01 5.83218694e-01 6.03741050e-01
3.88335109e-01 8.40090096e-01 2.35462949e-01 5.84225096e-02
3.03788722e-01 1.22324660e-01 9.19423103e-02 -1.90353096e-01
3.60726476e-01 -3.29819828e-01 1.70508593e-01 1.00560844e+00
-1.13366462e-01 1.27839640e-01 -9.55938101e-01 5.96811354e-01
-1.32791317e+00 -9.06958520e-01 -6.73597991e-01 2.36455989e+00
7.14508414e-01 4.01731431e-01 3.62347186e-01 4.15153831e-01
4.80116963e-01 -4.76023793e-01 1.64004773e-01 -1.17151630e+00
2.63652205e-01 3.25257331e-01 3.67001206e-01 8.94576013e-01
-7.86274791e-01 6.84975684e-01 7.94814634e+00 6.45299613e-01
-8.47698867e-01 -3.35347295e-01 5.29863954e-01 2.30369359e-01
-7.20595658e-01 2.77633548e-01 -1.25025189e+00 -2.21485049e-01
7.91426599e-01 -1.04418881e-01 9.22718551e-03 1.13093281e+00
-1.71043038e-01 1.72614634e-01 -1.05382216e+00 3.62790942e-01
-1.61283791e-01 -1.21117926e+00 8.18587467e-02 1.83312148e-02
8.26862335e-01 -6.78532779e-01 1.88668817e-01 7.20961392e-01
4.11316425e-01 -1.32844329e+00 -3.10741931e-01 2.28387371e-01
-2.32441887e-01 -9.09201026e-01 7.47448623e-01 5.68971515e-01
-6.33195221e-01 1.33278361e-03 -8.48323584e-01 -7.12025523e-01
-8.49871755e-01 5.65594435e-01 -1.01999688e+00 5.86445667e-02
7.24513292e-01 4.62311327e-01 -7.87093401e-01 1.58166873e+00
-1.90091982e-01 8.30983043e-01 -2.87577569e-01 -4.15161699e-01
7.67524838e-02 1.38114274e-01 1.85618713e-01 1.32238543e+00
6.25984311e-01 5.14852583e-01 -1.43574134e-01 7.62898624e-01
5.74231207e-01 1.29010484e-01 -6.61821008e-01 8.45989510e-02
3.26072484e-01 1.08257449e+00 -6.39044464e-01 -2.19675481e-01
-6.04050875e-01 4.40718830e-01 1.15137011e-01 3.30300212e-01
-5.72424948e-01 -5.96427619e-01 7.93783128e-01 5.91219306e-01
-1.09518826e-01 2.98503369e-01 -7.67598450e-01 -1.08591437e+00
-3.62499803e-01 -9.48023915e-01 5.18681228e-01 -6.59537017e-01
-1.03676760e+00 2.13287622e-01 -5.61289079e-02 -1.24389350e+00
2.73260057e-01 -7.65706658e-01 -1.23799753e+00 7.82668769e-01
-1.19661093e+00 -4.52447444e-01 -5.43035388e-01 2.28283748e-01
4.47830796e-01 1.34262731e-02 1.10608339e+00 -4.72715572e-02
-5.01960456e-01 1.06578577e+00 -2.03585982e-01 -2.95259058e-01
7.19350874e-01 -1.45084321e+00 2.42317453e-01 2.51290202e-01
6.71966653e-03 1.02775860e+00 1.32999277e+00 -1.78124070e-01
-8.47845495e-01 -8.28112841e-01 1.24986410e+00 -5.28247178e-01
3.80277127e-01 2.90433288e-01 -1.49208665e+00 9.34824288e-01
2.46761009e-01 -2.77822822e-01 1.52111101e+00 5.54235935e-01
-1.59671590e-01 -7.85612389e-02 -1.05925143e+00 5.65631866e-01
6.14200652e-01 1.45963877e-01 -9.25293684e-01 5.00051260e-01
2.94784904e-01 -7.82934964e-01 -1.22647452e+00 2.47223288e-01
5.90601385e-01 -1.07745838e+00 9.97497439e-01 -1.14476156e+00
3.83502960e-01 4.45655048e-01 3.04986596e-01 -1.41291940e+00
-1.77024662e-01 -6.86190009e-01 2.32914388e-01 1.05599034e+00
4.90823388e-01 -6.05510712e-01 1.17137432e+00 9.80076492e-01
-1.47396192e-01 -8.92883182e-01 -1.05026707e-01 -9.29858088e-01
8.32935393e-01 -5.13645411e-01 5.56728661e-01 1.41262412e+00
7.34376132e-01 2.33792871e-01 7.91736916e-02 2.99768988e-03
5.75147748e-01 -1.40190171e-02 8.85956824e-01 -1.66351318e+00
-1.22704700e-01 -6.32823646e-01 -5.79558909e-01 -1.16936016e+00
-3.78227860e-01 -4.57495511e-01 -3.16226244e-01 -1.32889307e+00
-5.18061407e-02 -4.81177658e-01 -1.43026546e-01 4.99266028e-01
-3.94464999e-01 1.55518604e-02 -1.01276226e-01 -2.50612199e-01
-1.80159777e-01 -2.36834824e-01 1.48257697e+00 2.87429243e-01
-1.91972136e-01 7.51570284e-01 -1.33443117e+00 1.19883668e+00
1.14855087e+00 -6.22070849e-01 -4.30536985e-01 -2.72291869e-01
4.73055124e-01 -1.71765253e-01 2.52557158e-01 -1.37213278e+00
2.51361907e-01 -6.87732995e-01 6.99388921e-01 -3.58540326e-01
7.41130188e-02 -1.10103464e+00 -2.25830570e-01 7.14505315e-01
-5.63871324e-01 2.30395287e-01 9.71128583e-01 -1.01761147e-01
-6.68337271e-02 -9.78025794e-01 7.42220163e-01 -4.39519942e-01
-2.19295368e-01 -3.00979972e-01 -5.81758142e-01 1.80874482e-01
1.10185683e+00 -3.36212307e-01 -4.13597614e-01 -7.85653055e-01
-1.11958408e+00 3.51832837e-01 2.56708086e-01 1.70024306e-01
8.31268251e-01 -1.14137697e+00 -3.26465338e-01 4.66039777e-01
-9.57638025e-04 -1.18577547e-01 2.63348460e-01 8.15081835e-01
-6.40051544e-01 6.11005723e-01 -1.79416105e-01 -3.80453557e-01
-1.59727335e+00 3.86746734e-01 5.23224354e-01 -8.49046744e-03
-3.98458749e-01 1.18786621e+00 -1.28238067e-01 -8.87033224e-01
7.63428271e-01 -6.19219124e-01 -2.20235571e-01 -4.94204313e-01
7.05909729e-01 7.73712277e-01 3.25709224e-01 2.45098576e-01
-9.20783505e-02 6.22361779e-01 -4.28462595e-01 5.79869926e-01
1.15813720e+00 1.53666273e-01 1.71059743e-01 7.71717906e-01
1.16532004e+00 8.07264671e-02 -6.70909166e-01 -6.89333677e-03
-1.54716209e-01 -3.92114431e-01 -1.44213021e-01 -9.73694265e-01
-6.66326046e-01 1.47615957e+00 2.86044478e-01 4.69558090e-01
1.00630641e+00 -2.67414123e-01 5.27903080e-01 7.91277707e-01
-1.61508456e-01 -8.76588464e-01 1.49663202e-02 4.12894577e-01
7.08594501e-01 -1.22698188e+00 4.42707747e-01 -7.37986445e-01
-4.18716580e-01 1.76250434e+00 1.25968468e+00 -2.22604394e-01
7.66898572e-01 3.65571916e-01 3.17067832e-01 -4.05209586e-02
-9.17990804e-01 3.82039487e-01 4.27191973e-01 4.50937271e-01
1.25734091e+00 -1.03934079e-01 -4.02206987e-01 5.49480200e-01
-6.44246042e-01 2.32135743e-01 1.02336645e+00 1.10579300e+00
-1.07560599e+00 -1.33983445e+00 -8.82500708e-01 4.18261349e-01
-6.25369489e-01 -3.69413779e-03 -4.30269450e-01 1.19148028e+00
6.99755773e-02 7.79899299e-01 -1.42453268e-01 -6.00534201e-01
5.59038579e-01 3.57559711e-01 8.26009452e-01 -8.00687313e-01
-1.05883956e+00 -6.44334406e-02 -2.80708224e-01 -6.94226250e-02
7.34683052e-02 -2.40054920e-01 -9.83706057e-01 -9.14738238e-01
-6.89518094e-01 8.48615587e-01 3.53602171e-01 9.09470439e-01
-1.89510554e-01 7.60210335e-01 1.75209150e-01 1.90700039e-01
-1.03058612e+00 -8.99773359e-01 -3.62744570e-01 9.53431278e-02
4.07865286e-01 -3.00795138e-01 -6.46155477e-01 -2.37615317e-01] | [10.048091888427734, 7.334079742431641] |
995d635f-a6f5-432a-b302-04ba3669b3f0 | a-contrastive-knowledge-transfer-framework | 2303.07599 | null | https://arxiv.org/abs/2303.07599v1 | https://arxiv.org/pdf/2303.07599v1.pdf | A Contrastive Knowledge Transfer Framework for Model Compression and Transfer Learning | Knowledge Transfer (KT) achieves competitive performance and is widely used for image classification tasks in model compression and transfer learning. Existing KT works transfer the information from a large model ("teacher") to train a small model ("student") by minimizing the difference of their conditionally independent output distributions. However, these works overlook the high-dimension structural knowledge from the intermediate representations of the teacher, which leads to limited effectiveness, and they are motivated by various heuristic intuitions, which makes it difficult to generalize. This paper proposes a novel Contrastive Knowledge Transfer Framework (CKTF), which enables the transfer of sufficient structural knowledge from the teacher to the student by optimizing multiple contrastive objectives across the intermediate representations between them. Also, CKTF provides a generalized agreement to existing KT techniques and increases their performance significantly by deriving them as specific cases of CKTF. The extensive evaluation shows that CKTF consistently outperforms the existing KT works by 0.04% to 11.59% in model compression and by 0.4% to 4.75% in transfer learning on various models and datasets. | ['Ming Zhao', 'Yitao Chen', 'Kaiqi Zhao'] | 2023-03-14 | null | null | null | null | ['model-compression'] | ['methodology'] | [ 2.56913334e-01 8.72255638e-02 -6.18823290e-01 -3.44777584e-01
-6.38435245e-01 -2.74842918e-01 4.51886892e-01 -7.08794445e-02
-3.35043997e-01 8.84407163e-01 -1.67705119e-01 -3.28288555e-01
-2.81386763e-01 -8.36052418e-01 -1.03093398e+00 -8.45329404e-01
1.11362472e-01 5.31598508e-01 3.27137142e-01 6.60966486e-02
1.01015612e-01 2.29344726e-01 -1.35744083e+00 4.01785880e-01
1.15291584e+00 1.06950331e+00 5.59113443e-01 3.74015570e-01
-1.63938388e-01 9.53855217e-01 -5.50781369e-01 -5.02904594e-01
2.48981416e-02 -5.46397865e-01 -1.12748194e+00 -3.00399870e-01
4.76859272e-01 -3.76256257e-01 -4.16303664e-01 8.75793159e-01
4.45053577e-01 1.26377821e-01 1.03628099e+00 -1.30576611e+00
-1.28999889e+00 5.37568569e-01 -6.47551775e-01 1.98917359e-01
-2.69661516e-01 -2.12258533e-01 5.99924803e-01 -8.28064322e-01
2.12900087e-01 1.11745441e+00 5.74937701e-01 6.79986775e-01
-1.21676707e+00 -9.26916063e-01 1.39353471e-02 6.04566514e-01
-1.48007786e+00 -2.27751583e-01 4.91221786e-01 -3.95911872e-01
9.33705330e-01 1.23611763e-01 4.25660223e-01 8.06102574e-01
-1.01369783e-01 1.03822398e+00 1.32266545e+00 -5.88898122e-01
2.80608479e-02 5.90402901e-01 7.32132494e-02 7.41869569e-01
6.66117966e-02 7.25549459e-02 -6.54394090e-01 3.96559276e-02
7.65585482e-01 -1.44232020e-01 -3.22288871e-01 -4.69597369e-01
-8.01403403e-01 8.33514631e-01 6.91183627e-01 1.29965633e-01
-7.26806745e-02 1.34763345e-01 2.46801272e-01 5.84370613e-01
5.94146013e-01 1.05904616e-01 -7.19635844e-01 -2.05526985e-02
-7.78520226e-01 -2.95419917e-02 6.86825097e-01 1.23503458e+00
9.69200492e-01 1.34510264e-01 -3.38569343e-01 9.80921924e-01
3.20517927e-01 5.23747385e-01 8.21948409e-01 -7.16040790e-01
6.06964648e-01 4.09190238e-01 -1.24419227e-01 -7.74690092e-01
4.35396552e-01 -6.15410686e-01 -7.18409300e-01 -1.12510785e-01
-1.89714395e-02 3.19081880e-02 -1.00046182e+00 1.78774619e+00
2.81215936e-01 6.54587269e-01 2.54300833e-01 5.07814944e-01
9.33982134e-01 7.45965898e-01 3.15345824e-01 -2.66833454e-01
9.11196947e-01 -1.33364809e+00 -5.20154536e-01 -1.02529854e-01
6.24516845e-01 -7.00191081e-01 1.30108309e+00 3.55779558e-01
-1.20408404e+00 -7.23356545e-01 -8.29817712e-01 -5.37727913e-03
-4.01928306e-01 1.20205581e-01 4.33312476e-01 5.58703542e-01
-1.04927516e+00 6.69892609e-01 -5.72258055e-01 -7.62623325e-02
8.47254097e-01 5.08858442e-01 -1.28744617e-01 -3.06019247e-01
-1.07091582e+00 1.14952457e+00 5.44074595e-01 -3.97836983e-01
-1.12132275e+00 -1.04600275e+00 -4.91658330e-01 2.57083654e-01
2.86939353e-01 -8.96630049e-01 1.50490201e+00 -1.15756536e+00
-1.73489523e+00 6.46133304e-01 -1.12711020e-01 -5.29798925e-01
5.71608126e-01 -4.61815238e-01 1.09964333e-01 2.21459240e-01
-2.32642516e-01 8.55221987e-01 9.66511369e-01 -1.36303258e+00
-5.35094440e-01 -1.38079450e-01 -2.12672260e-02 4.71359283e-01
-1.01097596e+00 -2.56896883e-01 -5.76544583e-01 -7.14465976e-01
-2.86057889e-01 -7.32502282e-01 1.15797445e-01 -1.06461450e-01
-3.53026316e-02 -4.44531739e-01 1.07879293e+00 -6.25754476e-01
1.35892546e+00 -1.88612127e+00 2.80048162e-01 -9.05537158e-02
2.03122288e-01 7.17509627e-01 -1.31851032e-01 2.84647465e-01
5.14277108e-02 1.27234191e-01 -2.21550643e-01 -2.73586333e-01
-9.49779302e-02 5.25552630e-01 -3.13374400e-01 -6.81965100e-03
1.18023284e-01 1.12633693e+00 -8.66174757e-01 -7.16070116e-01
1.26188517e-01 5.11064589e-01 -4.65201497e-01 4.32076901e-01
-1.34182870e-01 3.13749611e-01 -5.97964764e-01 3.25666279e-01
6.38929069e-01 -4.64617014e-01 1.72859652e-03 -2.44577765e-01
2.35950500e-01 2.10161656e-01 -5.93980432e-01 1.40892756e+00
-4.31203455e-01 5.05610108e-01 -3.86827558e-01 -1.59037852e+00
9.99416828e-01 3.56365860e-01 3.24578941e-01 -6.68215930e-01
-3.92975546e-02 -5.83495870e-02 -7.93854967e-02 -4.79006678e-01
2.06573099e-01 -3.70790035e-01 3.38907421e-01 4.18683410e-01
3.50487143e-01 -1.93603203e-01 -1.38728872e-01 1.66802287e-01
7.01274514e-01 8.42314810e-02 1.69944346e-01 -2.88029075e-01
4.40877646e-01 -2.78553516e-01 5.04813313e-01 5.91399789e-01
-1.11626640e-01 3.54165405e-01 1.03820860e-01 -2.89331198e-01
-8.10574770e-01 -1.28590119e+00 -5.02294227e-02 1.13233578e+00
9.87477675e-02 -3.75440001e-01 -8.17350268e-01 -9.88354325e-01
1.58135027e-01 6.93541884e-01 -7.46114612e-01 -6.70479178e-01
-4.04730052e-01 -4.74815965e-01 5.36627114e-01 8.42475176e-01
8.30095232e-01 -8.61287296e-01 -3.50356340e-01 4.66082059e-03
-2.22478271e-01 -9.04366851e-01 -3.91531020e-01 1.78738296e-01
-1.19722629e+00 -7.67924309e-01 -7.68327475e-01 -8.66319001e-01
8.25631976e-01 4.74104375e-01 1.01197064e+00 6.49404293e-03
-8.63095894e-02 4.86565739e-01 -3.22853982e-01 -8.05685103e-01
-3.83179814e-01 1.80228099e-01 4.11161408e-02 -7.60826841e-02
4.47979957e-01 -4.64960188e-01 -3.90546978e-01 3.98480982e-01
-9.14668083e-01 3.14259887e-01 8.20824623e-01 9.17334139e-01
6.50586367e-01 5.82348928e-02 9.31730747e-01 -7.73852885e-01
5.99151313e-01 -5.92175543e-01 -3.46331477e-01 6.53507710e-01
-1.17263889e+00 9.46250558e-02 7.13482261e-01 -7.60278165e-01
-1.30987322e+00 -2.28964895e-01 4.23604131e-01 -9.57124770e-01
9.93401408e-02 5.84195077e-01 -1.62216369e-02 -2.41529360e-01
5.45506418e-01 4.78745520e-01 9.87443328e-02 -6.31949008e-01
2.60348231e-01 5.87196410e-01 2.97451258e-01 -8.71502519e-01
8.50487709e-01 8.41590762e-02 -2.72769541e-01 -4.23681557e-01
-9.63480711e-01 -9.65668783e-02 -6.27653778e-01 3.96184735e-02
5.15543282e-01 -8.64981771e-01 -4.30886835e-01 4.06655282e-01
-7.71021485e-01 -5.29273808e-01 -5.04318714e-01 6.56726062e-01
-7.20031202e-01 3.55254829e-01 -6.42287314e-01 -4.40289795e-01
-2.95170009e-01 -1.07249177e+00 5.14660060e-01 3.50646496e-01
2.61850536e-01 -1.17844450e+00 -1.72659054e-01 5.85406899e-01
6.17587268e-01 -3.47050905e-01 1.24649036e+00 -4.80241567e-01
-4.94658023e-01 1.27386525e-01 -3.34550768e-01 9.34038341e-01
2.27294281e-01 -1.98605418e-01 -1.05351818e+00 -4.35355306e-01
-3.32391374e-02 -8.52519512e-01 1.05782175e+00 3.38653356e-01
1.60081518e+00 -4.66129959e-01 -4.76110905e-01 5.29576659e-01
1.23654246e+00 1.98549971e-01 6.71960294e-01 1.85976967e-01
6.95704579e-01 5.48631072e-01 4.05382156e-01 1.16101503e-02
5.55695951e-01 5.92224538e-01 2.07594305e-01 -6.29658252e-02
-3.60529959e-01 -4.45964426e-01 4.81635630e-01 1.34839642e+00
-2.33596906e-01 -1.73661336e-02 -7.59666860e-01 5.15751600e-01
-1.73174572e+00 -6.72475755e-01 3.96203756e-01 2.28831029e+00
1.29659009e+00 -2.49149278e-01 -1.96202993e-01 -5.21271154e-02
6.81527019e-01 -2.90220052e-01 -6.05027556e-01 -3.70629042e-01
2.33446673e-01 4.06642318e-01 3.76358718e-01 2.95421034e-01
-8.87036681e-01 1.04991210e+00 6.44855165e+00 1.41725683e+00
-1.10722589e+00 3.07137877e-01 7.17552185e-01 6.93734139e-02
-1.30202994e-01 -8.61207172e-02 -8.90611947e-01 3.78258944e-01
1.08643019e+00 -5.86798310e-01 2.88784295e-01 7.87690401e-01
-1.96266979e-01 -2.71808412e-02 -1.18367696e+00 9.35540438e-01
1.27586648e-01 -1.10209656e+00 5.62816024e-01 1.44244015e-01
1.06658936e+00 -6.81100637e-02 5.38091719e-01 6.22550845e-01
3.14167708e-01 -1.34742582e+00 4.06492829e-01 5.93977749e-01
6.83420897e-01 -8.50275218e-01 5.11476576e-01 5.83496630e-01
-9.71397817e-01 -4.38274555e-02 -7.63873160e-01 5.96340708e-02
-2.98274785e-01 3.40463638e-01 -1.17536557e+00 7.89700150e-01
7.73356795e-01 6.66054904e-01 -5.33782899e-01 9.44073319e-01
-2.77210623e-01 1.12324250e+00 -1.43658355e-01 1.96812019e-01
2.10005298e-01 -1.13826439e-01 -5.86154349e-02 1.15246570e+00
3.89759332e-01 -4.81128460e-03 2.46043667e-01 7.88051248e-01
-4.65270430e-01 1.43032566e-01 -5.15822649e-01 3.62030193e-02
5.88982880e-01 8.91114116e-01 -3.74139547e-01 -7.35385537e-01
-4.92001265e-01 8.38525653e-01 6.59944713e-01 4.11473423e-01
-8.76117647e-01 -1.71956405e-01 2.78320044e-01 -1.35271773e-01
6.09298885e-01 1.06586628e-01 -9.60566327e-02 -9.78293598e-01
-7.04160077e-04 -6.71355426e-01 4.69467729e-01 -7.09952295e-01
-1.45251024e+00 3.98710907e-01 4.92743194e-01 -1.23641980e+00
-2.55253404e-01 -4.40528512e-01 -3.71761471e-01 1.07346225e+00
-1.98484111e+00 -1.22475398e+00 -2.36728325e-01 8.72064769e-01
4.05341327e-01 -2.21502230e-01 9.40408885e-01 4.05586869e-01
-4.30795074e-01 1.00445628e+00 4.16846037e-01 6.95784530e-03
7.88138807e-01 -1.13780892e+00 -5.83642870e-02 1.95588365e-01
2.62797028e-02 4.84988689e-01 2.86914736e-01 -4.09277409e-01
-1.20259976e+00 -1.18421674e+00 8.38043630e-01 -3.40651840e-01
3.02063763e-01 7.91709721e-02 -1.29968750e+00 7.13030457e-01
2.03590199e-01 6.51500374e-02 7.59686530e-01 -2.54539233e-02
-6.07807636e-01 -2.93416470e-01 -1.21141887e+00 3.37823272e-01
7.79529214e-01 -4.62626785e-01 -7.53988206e-01 1.65209442e-01
7.18605578e-01 -1.95454508e-01 -1.11016643e+00 6.26833618e-01
5.39234936e-01 -7.85734296e-01 1.02537882e+00 -7.83399820e-01
5.37497759e-01 -6.89595044e-02 -5.04721813e-02 -1.48766756e+00
-2.15103284e-01 -8.22408125e-02 -5.08984625e-01 1.19605923e+00
2.70953298e-01 -5.78552425e-01 7.69209027e-01 3.03237110e-01
-2.12244466e-01 -1.00660825e+00 -8.86972547e-01 -1.10987520e+00
6.44955695e-01 -2.64095247e-01 5.82798421e-01 1.37274420e+00
-1.47175506e-01 4.71963882e-01 -3.62291306e-01 -1.44752949e-01
4.82281506e-01 1.01181254e-01 6.02837861e-01 -1.20575881e+00
-4.10574436e-01 -3.52409303e-01 -2.24858031e-01 -1.34478438e+00
1.62768677e-01 -1.20884514e+00 -2.04808116e-01 -1.33543253e+00
5.88411212e-01 -6.45302534e-01 -6.43942356e-01 6.92714632e-01
-3.32254142e-01 1.19555496e-01 1.94710821e-01 3.51248205e-01
-4.25297469e-01 9.03522134e-01 1.63525200e+00 -2.19801053e-01
-1.10236816e-02 3.81337889e-02 -8.06674242e-01 6.65415347e-01
9.36785102e-01 -5.54922223e-01 -9.97855723e-01 -7.01565087e-01
-3.01400125e-01 -1.92174166e-01 3.00083458e-01 -7.97356367e-01
2.57731199e-01 -2.09858760e-01 5.31333923e-01 -2.33795479e-01
3.40750039e-01 -7.35391259e-01 -2.15606466e-01 5.64806104e-01
-3.99297059e-01 -2.50796616e-01 2.33573675e-01 5.60890138e-01
-3.16362351e-01 -4.20994401e-01 9.19968188e-01 -1.28863692e-01
-6.11253083e-01 4.41182077e-01 4.70568500e-02 2.13272870e-02
1.07006037e+00 -2.59422868e-01 -3.27539682e-01 -4.13297415e-01
-6.05985582e-01 1.67144880e-01 4.71214093e-02 3.02712321e-01
7.64188588e-01 -1.37890732e+00 -7.44181812e-01 2.23759145e-01
5.67479990e-02 1.88556984e-01 1.26553118e-01 7.51414120e-01
-1.00970306e-01 4.64827925e-01 -2.66958147e-01 -6.65313780e-01
-1.34686446e+00 4.97530460e-01 2.16625914e-01 -3.14918339e-01
-3.42232585e-01 9.96056795e-01 5.82623839e-01 -4.30642784e-01
4.07075197e-01 -2.35354200e-01 -5.17794713e-02 -1.48320735e-01
4.48587656e-01 4.74105179e-01 5.31825833e-02 -4.21419173e-01
1.48970589e-01 6.33688331e-01 -5.63546002e-01 2.03286976e-01
1.18335569e+00 1.76803190e-02 1.39533907e-01 3.10734838e-01
1.38741899e+00 -5.30075669e-01 -1.36965334e+00 -6.76310062e-01
-1.07124873e-01 -4.53855455e-01 1.47660219e-04 -9.86569345e-01
-1.12903821e+00 1.14304960e+00 4.96975601e-01 -1.78049922e-01
1.36471915e+00 1.68407828e-01 6.16302133e-01 6.03877664e-01
3.52930129e-01 -1.00721073e+00 4.76480454e-01 5.66209912e-01
7.55939662e-01 -1.04943669e+00 5.74856587e-02 -4.15339798e-01
-5.97261488e-01 8.47379923e-01 9.26236868e-01 4.79003899e-02
5.51349103e-01 1.83489650e-01 -5.78137971e-02 1.81200132e-01
-1.00888586e+00 -7.41157215e-04 4.70531613e-01 7.66942799e-01
3.89121443e-01 1.06244944e-01 -2.01769948e-01 5.64018726e-01
-1.59832805e-01 2.45884359e-01 4.61634472e-02 1.05027652e+00
-6.18785501e-01 -1.27916598e+00 -1.19600073e-01 6.18368924e-01
-3.31104934e-01 -1.55703112e-01 -3.02916169e-01 8.53945851e-01
1.49365544e-01 7.01886415e-01 -1.48696899e-01 -4.19999510e-01
1.32466227e-01 2.66417772e-01 7.51288950e-01 -6.07878208e-01
-5.58101237e-01 -4.25318889e-02 -3.79451960e-01 -7.30152279e-02
-6.61639750e-01 -2.56155550e-01 -9.93360817e-01 -2.46413320e-01
-5.78531504e-01 4.88088369e-01 4.28255469e-01 9.68144298e-01
3.36960614e-01 4.82276142e-01 6.18473649e-01 -4.04675305e-01
-9.43824470e-01 -1.15589261e+00 -4.54876810e-01 2.07240716e-01
1.33555427e-01 -7.41639018e-01 -1.67028010e-01 4.78109211e-01] | [9.492290496826172, 3.2946627140045166] |
9d573ec6-6b21-4b44-9b0d-9056998698f2 | union-subgraph-neural-networks | 2305.15747 | null | https://arxiv.org/abs/2305.15747v1 | https://arxiv.org/pdf/2305.15747v1.pdf | Union Subgraph Neural Networks | Graph Neural Networks (GNNs) are widely used for graph representation learning in many application domains. The expressiveness of vanilla GNNs is upper-bounded by 1-dimensional Weisfeiler-Leman (1-WL) test as they operate on rooted subtrees through iterative message passing. In this paper, we empower GNNs by injecting neighbor-connectivity information extracted from a new type of substructure. We first investigate different kinds of connectivities existing in a local neighborhood and identify a substructure called union subgraph, which is able to capture the complete picture of the 1-hop neighborhood of an edge. We then design a shortest-path-based substructure descriptor that possesses three nice properties and can effectively encode the high-order connectivities in union subgraphs. By infusing the encoded neighbor connectivities, we propose a novel model, namely Union Subgraph Neural Network (UnionSNN), which is proven to be strictly more powerful than 1-WL in distinguishing non-isomorphic graphs. Additionally, the local encoding from union subgraphs can also be injected into arbitrary message-passing neural networks (MPNNs) and Transformer-based models as a plugin. Extensive experiments on 17 benchmarks of both graph-level and node-level tasks demonstrate that UnionSNN outperforms state-of-the-art baseline models, with competitive computational efficiency. The injection of our local encoding to existing models is able to boost the performance by up to 11.09%. | ['Yiping Ke', 'Vijay Prakash Dwivedi', 'Qingtian Bian', 'Aihu Zhang', 'Jiaxing Xu'] | 2023-05-25 | null | null | null | null | ['graph-representation-learning'] | ['methodology'] | [ 1.55373424e-01 3.13434064e-01 -7.10746884e-01 -2.14404076e-01
-3.05913985e-01 -4.92232978e-01 4.86866176e-01 4.71621871e-01
-1.55124009e-01 5.42696655e-01 -1.57901362e-01 -8.26901734e-01
-3.71363819e-01 -1.61246812e+00 -1.08724868e+00 -6.34744763e-01
-9.84496891e-01 3.05838346e-01 4.86866891e-01 -3.93082470e-01
-1.08410187e-01 8.23209643e-01 -1.35433877e+00 2.85130084e-01
7.20569372e-01 1.01469398e+00 -1.06711999e-01 5.89003265e-01
-3.81666660e-01 7.64388323e-01 -3.28592092e-01 -2.99144208e-01
2.56327569e-01 -1.96083456e-01 -1.10889649e+00 -2.87794173e-01
6.28823519e-01 -1.08469076e-01 -9.18121159e-01 1.06382453e+00
1.35697022e-01 -3.42800766e-02 6.05991304e-01 -1.49746311e+00
-5.71882546e-01 1.17472017e+00 -5.85085869e-01 2.67941892e-01
2.28192016e-01 -8.75279158e-02 1.45925295e+00 -3.86645257e-01
7.27651596e-01 1.32546103e+00 7.57594585e-01 1.97596133e-01
-1.26877749e+00 -6.89691007e-01 2.98735440e-01 3.50726813e-01
-1.52154911e+00 1.13362610e-01 8.02862287e-01 8.01305398e-02
1.10350990e+00 3.77483547e-01 5.79983711e-01 9.71219063e-01
4.22105819e-01 7.12125301e-01 7.60075331e-01 -1.52751073e-01
-2.32653953e-02 -4.95210469e-01 3.62675816e-01 1.17239678e+00
4.97450441e-01 2.12100074e-02 -9.31430534e-02 -1.38916254e-01
9.69626844e-01 1.03224866e-01 -3.98462325e-01 -6.40154123e-01
-9.12018001e-01 9.84736085e-01 1.36722350e+00 6.07095420e-01
-1.06315380e-02 4.73482370e-01 4.82379675e-01 6.62084401e-01
3.33167017e-01 1.18118569e-01 -3.10896963e-01 5.36682010e-01
-5.66902161e-01 -1.07817359e-01 9.87102151e-01 8.33722770e-01
1.00138247e+00 2.05139499e-02 -3.80434185e-01 6.17516875e-01
1.58423185e-01 2.04470962e-01 2.35131904e-01 -3.46272171e-01
4.49540168e-01 1.14695072e+00 -8.56438875e-01 -1.30949998e+00
-7.29526699e-01 -8.05567563e-01 -1.46330070e+00 -1.67638347e-01
2.73662120e-01 2.38105521e-01 -1.17590582e+00 1.94095910e+00
3.23140949e-01 5.40994346e-01 -1.44824475e-01 5.06422520e-01
1.27968669e+00 5.87332249e-01 -1.56632170e-01 2.02205643e-01
1.19036293e+00 -8.24008107e-01 -9.80652273e-02 -6.00011609e-02
1.15689754e+00 6.99739233e-02 6.60438478e-01 -1.20128021e-02
-8.38999569e-01 -4.37871695e-01 -1.15891409e+00 7.59662241e-02
-6.73563957e-01 -4.55532104e-01 9.12750781e-01 6.43781543e-01
-1.41058695e+00 9.25626218e-01 -6.96642220e-01 -1.54100344e-01
4.92885560e-01 5.25896549e-01 -8.03631663e-01 -2.92611539e-01
-1.38990164e+00 2.68792838e-01 7.70726740e-01 1.66032948e-02
-8.68662477e-01 -4.92878228e-01 -1.35242701e+00 4.10382926e-01
6.13008618e-01 -4.77026105e-01 7.06854880e-01 -6.91299915e-01
-9.53999341e-01 7.11078644e-01 1.01163104e-01 -6.39953017e-01
9.05883461e-02 5.79966009e-01 -6.16792321e-01 3.07848006e-01
-7.57498369e-02 5.41045249e-01 4.92490828e-01 -9.79328752e-01
-4.00089324e-01 -2.79032975e-01 4.94333208e-01 -1.65983155e-01
-4.52479213e-01 -3.77320617e-01 -6.41086876e-01 -5.92543721e-01
4.01456535e-01 -7.62464285e-01 -2.84729570e-01 1.79724966e-03
-8.97273242e-01 -4.39388186e-01 7.17881918e-01 -2.66125742e-02
1.32993639e+00 -1.95731640e+00 -1.30899325e-02 8.47332954e-01
9.72259343e-01 3.62478226e-01 -6.97739184e-01 6.22574508e-01
-2.09088430e-01 2.53606766e-01 -2.97137052e-01 1.22503757e-01
3.37215625e-02 5.21592915e-01 -6.70639426e-02 5.94037592e-01
2.95484066e-01 1.20838690e+00 -9.65998292e-01 -3.31328392e-01
-7.53465295e-02 4.11809057e-01 -4.31753367e-01 -6.87231943e-02
-1.37469292e-01 -1.98634535e-01 -3.83355469e-01 5.31391799e-01
9.89713132e-01 -6.72558308e-01 5.56660295e-01 -1.35077640e-01
5.00606894e-01 3.95665884e-01 -1.19964027e+00 1.42620921e+00
-2.32137024e-01 3.67676407e-01 -1.18498810e-01 -1.38501048e+00
9.89864707e-01 -1.37344509e-01 1.44287080e-01 -8.54689538e-01
1.24956429e-01 4.09565531e-02 1.36219159e-01 -9.42376554e-02
2.46199951e-01 2.29293779e-01 -2.28747830e-01 2.60629475e-01
2.98136264e-01 5.07927299e-01 5.22980332e-01 4.99055147e-01
1.69052529e+00 -3.60494614e-01 3.73386711e-01 -4.11316782e-01
7.01576591e-01 -4.72958773e-01 4.36665624e-01 1.10908294e+00
2.62577198e-02 2.87324339e-01 1.05044556e+00 -6.42438769e-01
-4.51046318e-01 -1.36355186e+00 6.16642833e-03 1.16610539e+00
1.76899105e-01 -7.32229233e-01 -4.29751813e-01 -9.77885842e-01
2.20814720e-01 1.37161106e-01 -8.17591906e-01 -5.54487407e-01
-6.48340225e-01 -5.90861976e-01 1.01327145e+00 7.77835369e-01
4.50761378e-01 -8.34226131e-01 1.87537875e-02 2.58700192e-01
9.33002606e-02 -1.20532537e+00 -3.12632620e-01 3.91292810e-01
-9.77020681e-01 -1.36014104e+00 -5.28267205e-01 -1.11162782e+00
6.31260276e-01 3.36110264e-01 1.26194048e+00 5.15403926e-01
-8.19915235e-02 1.33087605e-01 -3.18949312e-01 2.68061161e-01
-4.48098361e-01 5.24181783e-01 -2.61744082e-01 3.38130854e-02
1.12684734e-01 -9.90812957e-01 -3.51591080e-01 3.07322770e-01
-1.11444521e+00 -1.67288277e-02 7.34259307e-01 7.13177860e-01
5.84251165e-01 1.21168070e-01 4.01339799e-01 -1.23317802e+00
5.85171819e-01 -6.63869739e-01 -5.76785624e-01 1.83980420e-01
-3.88305753e-01 4.62021023e-01 1.02200401e+00 -3.53633970e-01
-2.81354904e-01 -1.54033795e-01 -1.49464473e-01 -3.21138799e-01
-2.17848554e-01 8.31719100e-01 -8.35524052e-02 -5.12969434e-01
4.35728550e-01 1.89246804e-01 -1.67102382e-01 -3.05029601e-01
3.62150311e-01 2.09313855e-01 5.94682157e-01 -6.89418733e-01
9.27239120e-01 3.78834575e-01 6.08175874e-01 -7.66662300e-01
-4.81148392e-01 -3.94664645e-01 -4.49580133e-01 -9.45539121e-03
2.45009914e-01 -6.52036726e-01 -1.12131500e+00 4.78015661e-01
-1.02584612e+00 -4.06509846e-01 1.06070265e-01 1.35465741e-01
-4.98968996e-02 6.03219151e-01 -1.03568935e+00 -2.68673360e-01
-3.82094055e-01 -1.02379715e+00 8.42913926e-01 -6.96162507e-02
2.45575622e-01 -1.14917243e+00 3.35068745e-03 -3.96931231e-01
3.32718372e-01 6.55034125e-01 1.49602365e+00 -1.03737617e+00
-6.45459414e-01 -1.80152923e-01 -5.92983603e-01 1.14597067e-01
-8.87951851e-02 -1.58696949e-01 -6.01154804e-01 -5.48869848e-01
-6.75214946e-01 -1.77276075e-01 1.30902028e+00 2.66525120e-01
1.35516500e+00 -4.41394836e-01 -6.45021737e-01 8.77305806e-01
1.57100213e+00 -2.52473265e-01 6.45237982e-01 1.41203701e-01
9.82744277e-01 2.71507621e-01 -2.67987400e-01 7.22569302e-02
5.27895391e-01 3.15136850e-01 9.36123848e-01 -1.63470432e-01
-6.65703043e-02 -3.71008635e-01 2.23347396e-01 8.42802525e-01
-1.07363008e-01 -6.02820575e-01 -8.84157717e-01 4.34630930e-01
-1.70037293e+00 -6.91724956e-01 -5.13032973e-01 2.02377343e+00
5.11556983e-01 3.68174851e-01 3.44014615e-02 2.94685632e-01
8.89293075e-01 6.29219413e-01 -3.79305989e-01 -4.90772218e-01
-2.01480314e-01 5.21005630e-01 6.90830946e-01 2.84235179e-01
-1.05578578e+00 8.22483063e-01 5.54951429e+00 1.02973723e+00
-9.14175987e-01 -1.72314972e-01 5.88292062e-01 4.51487869e-01
-3.85182619e-01 -1.49839059e-01 -5.44547796e-01 1.40713692e-01
1.01009178e+00 -1.12402476e-01 4.44912225e-01 7.08994389e-01
-5.63198805e-01 2.93076843e-01 -1.23862898e+00 5.20358086e-01
-3.91141064e-02 -1.44505131e+00 3.84293795e-01 1.33612603e-01
6.01645887e-01 3.64662945e-01 -2.04644680e-01 5.62251806e-01
5.24900734e-01 -1.16116762e+00 2.55028993e-01 3.77156101e-02
9.56378281e-01 -9.78230715e-01 7.65144885e-01 2.50868708e-01
-1.90142846e+00 -9.61142257e-02 -4.70318854e-01 -2.26426125e-02
-2.88140476e-01 6.30381227e-01 -6.75252199e-01 1.11851406e+00
5.72423935e-01 7.00993896e-01 -6.40698075e-01 1.04188716e+00
-2.52342671e-01 6.33347094e-01 -4.50154662e-01 -1.20916098e-01
7.08159268e-01 -1.48481270e-02 5.01664340e-01 1.27605164e+00
1.93870217e-01 -1.29181400e-01 1.74301848e-01 9.21071410e-01
-7.45771945e-01 -2.95529980e-03 -8.84496510e-01 -8.47458318e-02
4.19764936e-01 1.29722726e+00 -9.67347801e-01 -1.94345817e-01
-3.07754606e-01 8.42479944e-01 8.11277986e-01 3.12748492e-01
-8.45832527e-01 -8.39610577e-01 5.60329676e-01 5.45057841e-02
5.94133675e-01 -1.25914410e-01 4.12591845e-01 -8.55114162e-01
8.89975950e-03 -8.26690495e-01 8.02605152e-01 -3.45354259e-01
-1.30861926e+00 9.35184777e-01 -4.28767055e-02 -8.93409550e-01
-5.16871065e-02 -8.82273316e-01 -9.24802721e-01 3.97156954e-01
-1.63817668e+00 -1.19990563e+00 -4.42604154e-01 8.45500350e-01
-1.64614961e-01 1.12175189e-01 8.73582661e-01 2.42636383e-01
-5.10282040e-01 9.67644215e-01 -5.54367304e-02 6.32012904e-01
1.34458676e-01 -1.34041429e+00 7.48415053e-01 7.27360070e-01
5.05667090e-01 5.51282048e-01 1.07560024e-01 -5.13912916e-01
-1.60854948e+00 -1.43226779e+00 6.40550852e-01 2.14062527e-01
8.11430633e-01 -5.84290087e-01 -1.16618145e+00 8.61845076e-01
-1.65517941e-01 4.70674604e-01 3.11637431e-01 1.72232196e-01
-9.37688947e-01 -2.40852207e-01 -9.20015872e-01 6.02528632e-01
1.47363710e+00 -5.53397596e-01 -2.71362841e-01 2.74761289e-01
1.11013174e+00 -4.00759816e-01 -9.55176353e-01 6.87771022e-01
2.96566188e-01 -1.00902879e+00 1.12251151e+00 -8.33164811e-01
2.67838567e-01 -8.06971937e-02 2.06443574e-02 -1.13935924e+00
-6.02679968e-01 -6.04367852e-01 -2.60185301e-01 8.84409368e-01
4.36723650e-01 -1.07715607e+00 8.01880300e-01 -2.57378697e-01
-1.03032611e-01 -9.96218085e-01 -1.03726220e+00 -1.05181336e+00
2.00064287e-01 -6.19274199e-01 9.47493970e-01 9.90713835e-01
-1.28051052e-02 3.63492817e-01 1.61672719e-02 4.44661915e-01
6.02332532e-01 2.47491181e-01 6.33586884e-01 -1.57749689e+00
-2.19605088e-01 -7.25400150e-01 -1.03765976e+00 -1.19898045e+00
4.73792464e-01 -1.67764127e+00 -2.95395702e-01 -1.68539894e+00
7.22696185e-02 -2.63538301e-01 -7.84185112e-01 7.83676445e-01
1.38353467e-01 2.24636540e-01 -8.94369632e-02 -3.12270552e-01
-6.10121012e-01 4.36293006e-01 1.19952738e+00 -5.21300316e-01
-5.02303280e-02 -1.17116816e-01 -5.49866617e-01 5.79530954e-01
7.44380772e-01 -5.86398661e-01 -3.06241095e-01 -2.24064916e-01
3.42860073e-01 5.78134134e-02 6.38231874e-01 -1.05586147e+00
3.66643101e-01 9.31313336e-02 1.15322769e-01 -4.78519708e-01
-3.98742817e-02 -6.48520947e-01 5.91561869e-02 6.27882659e-01
-2.98849076e-01 8.94310325e-02 2.55332798e-01 7.47237206e-01
-1.08755253e-01 2.23167483e-02 4.91936862e-01 1.04772441e-01
-6.91802979e-01 7.54472435e-01 -9.77490246e-02 1.49202377e-01
7.53927767e-01 -2.53048420e-01 -6.24461770e-01 -3.27240616e-01
-4.15254235e-01 2.86063254e-01 6.43996298e-02 3.26959938e-01
5.92828929e-01 -1.47040796e+00 -6.23732626e-01 4.60705131e-01
3.92871261e-01 -6.88854679e-02 2.18518764e-01 7.60201037e-01
-4.89212841e-01 2.97576547e-01 -1.54493377e-01 -5.89049041e-01
-1.12134552e+00 6.26152217e-01 3.71299535e-01 -6.62640333e-01
-8.40671062e-01 8.25095952e-01 3.73650938e-01 -7.60080159e-01
3.45480800e-01 -6.18255854e-01 -1.29197478e-01 -4.30523664e-01
3.27892572e-01 3.70610595e-01 1.30470887e-01 -7.29167104e-01
-4.12768006e-01 4.04667318e-01 -2.15480849e-01 6.58010364e-01
1.24805677e+00 4.12578493e-01 -4.49838102e-01 3.22951227e-02
1.70248103e+00 -3.08662981e-01 -6.08893573e-01 -5.87923050e-01
2.85656173e-02 -1.03214808e-01 5.57115078e-02 -3.13842982e-01
-1.51252592e+00 7.20717013e-01 1.46440879e-01 6.69455409e-01
1.06242692e+00 2.04661533e-01 9.31491315e-01 8.50465000e-01
3.93264532e-01 -4.58886057e-01 -1.67754993e-01 6.74581110e-01
6.25660241e-01 -7.36615360e-01 -2.68259287e-01 -7.22709596e-01
1.51991248e-01 1.29066253e+00 6.23067021e-01 -2.22793728e-01
7.07538068e-01 3.05989742e-01 -5.62373757e-01 -4.08519566e-01
-6.89464509e-01 -4.77422208e-01 5.55382669e-01 5.20333588e-01
1.09818220e-01 2.04226077e-01 -1.34073973e-01 4.76567924e-01
5.54931425e-02 -2.88217068e-01 4.05130386e-01 6.29777014e-01
-4.49847341e-01 -1.00318372e+00 1.78105190e-01 7.51684070e-01
-2.52355069e-01 -2.56901503e-01 -4.11561489e-01 1.05586433e+00
-9.99619290e-02 7.01059401e-01 1.94703281e-01 -7.55513430e-01
1.07321702e-01 -4.20801103e-01 3.69976640e-01 -5.29724956e-01
-4.78553325e-01 -3.77003223e-01 2.15707198e-01 -8.79407525e-01
-1.12918139e-01 1.82326555e-01 -1.42650998e+00 -5.92971265e-01
-3.43157172e-01 1.05890892e-01 1.53027967e-01 7.93251097e-01
4.75153208e-01 6.89264297e-01 6.37107015e-01 -6.05379462e-01
-2.47650191e-01 -7.01753438e-01 -8.06211591e-01 2.39512146e-01
3.52202117e-01 -4.83193636e-01 -5.03111362e-01 -7.13570416e-01] | [6.989717960357666, 6.210718631744385] |
a79d294c-ffc5-4008-9515-4a417174a0dd | addressing-class-imbalance-in-semi-supervised | 2209.00123 | null | https://arxiv.org/abs/2209.00123v1 | https://arxiv.org/pdf/2209.00123v1.pdf | Addressing Class Imbalance in Semi-supervised Image Segmentation: A Study on Cardiac MRI | Due to the imbalanced and limited data, semi-supervised medical image segmentation methods often fail to produce superior performance for some specific tailed classes. Inadequate training for those particular classes could introduce more noise to the generated pseudo labels, affecting overall learning. To alleviate this shortcoming and identify the under-performing classes, we propose maintaining a confidence array that records class-wise performance during training. A fuzzy fusion of these confidence scores is proposed to adaptively prioritize individual confidence metrics in every sample rather than traditional ensemble approaches, where a set of predefined fixed weights are assigned for all the test cases. Further, we introduce a robust class-wise sampling method and dynamic stabilization for a better training strategy. Our proposed method considers all the under-performing classes with dynamic weighting and tries to remove most of the noises during training. Upon evaluation on two cardiac MRI datasets, ACDC and MMWHS, our proposed method shows effectiveness and generalizability and outperforms several state-of-the-art methods found in the literature. | ['Ram Sarkar', 'Sagnik Ghosal', 'Hritam Basak'] | 2022-08-31 | null | null | null | null | ['semi-supervised-medical-image-segmentation'] | ['computer-vision'] | [ 5.73066890e-01 1.30947419e-02 -3.16086531e-01 -6.41071737e-01
-8.17136347e-01 -3.09478521e-01 1.38378039e-01 7.24145710e-01
-4.55379933e-01 8.65117848e-01 -1.44733727e-01 -1.57597765e-01
-6.37656987e-01 -6.16411865e-01 -2.56547064e-01 -1.09179914e+00
9.07624327e-03 6.05540276e-01 3.99655491e-01 3.68229955e-01
4.53993142e-01 2.13229924e-01 -1.78038871e+00 3.84572029e-01
1.47719765e+00 1.02600765e+00 5.60736246e-02 3.27441216e-01
9.97818708e-02 6.59893692e-01 -7.98780680e-01 3.49220484e-02
1.14466302e-01 -4.59581047e-01 -5.32331169e-01 2.45941043e-01
3.67206156e-01 -5.21807522e-02 4.46361244e-01 1.10699642e+00
6.56945169e-01 2.11503536e-01 8.11866224e-01 -1.12505043e+00
5.08112484e-04 6.45646811e-01 -7.88248181e-01 3.75020474e-01
-6.48279935e-02 -8.99031609e-02 6.66321874e-01 -5.34772158e-01
2.40728855e-01 6.23767078e-01 6.54813111e-01 3.94245267e-01
-1.18931687e+00 -8.34048450e-01 1.49645582e-01 1.53144881e-01
-1.38249242e+00 -3.25111538e-01 8.40412557e-01 -5.06449759e-01
3.03192854e-01 3.85081679e-01 3.61202002e-01 5.91662109e-01
2.56493747e-01 5.46920598e-01 1.50383103e+00 -4.35073435e-01
5.05310953e-01 2.38412499e-01 3.28018725e-01 5.81930816e-01
5.50845444e-01 -2.28609324e-01 -3.46023679e-01 -4.18501228e-01
2.89307445e-01 -2.07353141e-02 -2.92492449e-01 -4.68117803e-01
-1.06097627e+00 6.17834747e-01 5.88638820e-02 4.75218266e-01
-6.32865548e-01 -4.28265065e-01 4.50140178e-01 -1.36609241e-01
5.40468633e-01 3.47607821e-01 -4.69904721e-01 9.59060863e-02
-1.42934227e+00 1.17672928e-01 4.70330715e-01 3.28077972e-01
5.54993272e-01 -1.72748417e-01 -5.04125953e-01 1.04915822e+00
7.52125755e-02 4.51743677e-02 6.35704041e-01 -8.68124247e-01
3.84060144e-01 9.07427371e-01 -4.44526486e-02 -1.07191896e+00
-4.20503974e-01 -1.00785983e+00 -8.62873971e-01 2.03160957e-01
4.91986334e-01 -2.24755675e-01 -1.07406950e+00 1.59331751e+00
5.76162338e-01 1.40845433e-01 -8.22784379e-02 8.87244225e-01
5.36593199e-01 2.97598690e-01 2.57500798e-01 -6.95408881e-01
1.13545215e+00 -5.93309164e-01 -7.31861055e-01 -7.13414550e-02
4.26961273e-01 -6.28985643e-01 8.08476388e-01 6.89687908e-01
-8.93621624e-01 -5.56291521e-01 -1.33505213e+00 7.74311543e-01
2.34147366e-02 3.20035040e-01 4.04831380e-01 9.34419215e-01
-6.09753132e-01 7.50857294e-01 -9.14530933e-01 1.41964406e-01
5.63308775e-01 5.29226422e-01 -7.27171674e-02 -1.18875504e-03
-9.18274224e-01 7.30713785e-01 3.72857869e-01 1.75563321e-01
-6.33487761e-01 -7.18889832e-01 -5.62690258e-01 -1.77947655e-01
4.40042526e-01 -3.78473788e-01 9.46265340e-01 -1.00354385e+00
-1.14267957e+00 6.47309840e-01 9.15232152e-02 -4.12556916e-01
6.01040244e-01 7.53872246e-02 -3.64560872e-01 2.97052860e-01
3.08106951e-02 4.16610539e-01 9.10922885e-01 -1.53107071e+00
-7.19215691e-01 -5.52734077e-01 -2.24724188e-01 3.00153911e-01
-4.55497652e-01 -2.00506613e-01 -1.36053011e-01 -7.08418727e-01
5.03754973e-01 -7.42310941e-01 -4.18476373e-01 -4.38913405e-01
-4.29194570e-01 -7.78641105e-02 5.75463951e-01 -4.69724059e-01
1.78598535e+00 -1.84121859e+00 -1.97527409e-01 6.36454642e-01
3.06289136e-01 3.38127822e-01 2.99116045e-01 -9.42044407e-02
-9.98144150e-02 -5.27815484e-02 -6.49874926e-01 -9.46871787e-02
-3.65272760e-01 2.29610920e-01 2.65903056e-01 5.21806598e-01
1.71668798e-01 -1.75308846e-02 -9.25812423e-01 -8.46639931e-01
3.63712341e-01 3.54437053e-01 -5.59954822e-01 1.47159487e-01
9.46267396e-02 4.89774644e-01 -5.99069178e-01 6.59512818e-01
7.62565970e-01 -9.48403031e-02 3.00544024e-01 -3.50576073e-01
1.59844086e-01 -1.88917741e-02 -1.54001939e+00 1.15712011e+00
-1.88668281e-01 -8.82885084e-02 -1.29968509e-01 -1.32741976e+00
1.06760395e+00 3.07689071e-01 8.66313219e-01 -3.80837381e-01
2.77828544e-01 4.18524563e-01 4.24603313e-01 -4.97336507e-01
2.40452677e-01 -1.71448663e-01 6.90279081e-02 3.26381832e-01
-7.06678703e-02 1.52100086e-01 2.15028703e-01 -1.24169812e-02
8.49849105e-01 -2.32252181e-02 3.33685845e-01 -4.39861327e-01
7.56732047e-01 -6.79592416e-02 9.20371771e-01 8.17891061e-01
-6.14183307e-01 7.38275349e-01 3.88515890e-01 -2.32199982e-01
-6.94655836e-01 -7.59031713e-01 -5.25380909e-01 8.00920367e-01
-4.26382311e-02 -6.98544011e-02 -1.05257666e+00 -9.01839912e-01
-1.96953759e-01 6.78974926e-01 -6.63757861e-01 -2.23074883e-01
-3.20527941e-01 -1.38880002e+00 3.64946961e-01 3.91834825e-01
3.72265428e-01 -1.00489521e+00 -9.60740209e-01 3.88510764e-01
-1.39688864e-01 -6.34402275e-01 -2.04824731e-01 4.81979489e-01
-1.08597660e+00 -1.20956719e+00 -6.29769564e-01 -5.98105252e-01
1.06495774e+00 -1.09270833e-01 1.04664516e+00 3.04543912e-01
-8.51217583e-02 -1.85417786e-01 -4.29987848e-01 -4.03208643e-01
-4.33886170e-01 2.36144401e-02 -1.42274434e-02 2.75764525e-01
1.64489791e-01 -4.09508198e-01 -8.91322494e-01 4.79826599e-01
-7.52212226e-01 -1.95425466e-01 4.96474475e-01 1.04306388e+00
8.35387290e-01 5.07204056e-01 9.56940353e-01 -1.16407645e+00
5.82595885e-01 -4.90258098e-01 -1.95592016e-01 3.19899678e-01
-1.06365907e+00 -8.90179127e-02 6.00360155e-01 -5.33879042e-01
-1.07975638e+00 -1.61215160e-02 1.85907498e-01 -1.40568361e-01
-1.85294837e-01 4.92520541e-01 3.39064300e-02 1.12130865e-01
7.15800822e-01 -8.55246037e-02 1.25563964e-01 -2.05923513e-01
-2.56511807e-01 7.71284640e-01 5.03688037e-01 -6.84069633e-01
3.63114208e-01 2.39752620e-01 -1.13197856e-01 -2.75238305e-01
-9.46586072e-01 -4.43666995e-01 -5.88256776e-01 -5.89533687e-01
5.75368881e-01 -6.08036816e-01 -3.18357140e-01 4.22417939e-01
-5.20118773e-01 -1.56029416e-02 -1.86316758e-01 4.99181420e-01
-1.80111274e-01 3.19318324e-01 -2.37599894e-01 -1.21717036e+00
-5.34183562e-01 -1.35185289e+00 8.45426500e-01 4.83820081e-01
-4.26724911e-01 -7.16962755e-01 4.35762014e-03 5.12117922e-01
3.03815037e-01 5.62556624e-01 8.48972559e-01 -9.35588479e-01
1.17760658e-01 -4.13501412e-01 1.60670996e-01 5.83673894e-01
3.63385409e-01 5.55252694e-02 -8.65307331e-01 -3.40620250e-01
9.21281204e-02 -2.12046593e-01 7.77721345e-01 6.03365004e-01
1.44328475e+00 -9.19014290e-02 -3.43540072e-01 2.47360989e-01
1.40442157e+00 4.88265872e-01 3.43877345e-01 3.42758477e-01
3.52006525e-01 6.20993853e-01 8.78538668e-01 5.20303428e-01
1.46726310e-01 3.37243974e-01 3.19164187e-01 -2.07566023e-01
2.06543490e-01 2.41427854e-01 -1.18828811e-01 7.98121512e-01
-9.35587361e-02 -9.33136791e-02 -9.22663093e-01 5.37332654e-01
-1.71013319e+00 -7.25707710e-01 -5.91348298e-02 2.48873019e+00
1.14646506e+00 5.70025086e-01 1.31503329e-01 7.90770710e-01
9.59153175e-01 -1.26564115e-01 -6.04460061e-01 -1.32047132e-01
7.50366319e-03 3.83205682e-01 4.42477912e-01 1.67941004e-01
-1.12269092e+00 2.18993813e-01 6.22230577e+00 9.64658558e-01
-1.02011788e+00 1.14052771e-02 1.13174367e+00 -2.53420603e-02
-1.67000428e-01 -5.22252433e-02 -3.89266282e-01 5.62039733e-01
8.35634410e-01 1.97745096e-02 5.21438429e-03 6.15242600e-01
3.44298929e-01 -5.15447736e-01 -6.64713919e-01 6.16265237e-01
1.01724282e-01 -9.19750929e-01 -1.87940508e-01 -2.50555903e-01
9.26508248e-01 -3.85967851e-01 -8.67944509e-02 3.99540327e-02
1.99740045e-02 -8.38150024e-01 6.06677949e-01 6.25445008e-01
5.27788281e-01 -8.68411303e-01 1.07778096e+00 4.82622176e-01
-6.43443286e-01 -2.33791754e-01 -2.01582178e-01 1.58900991e-01
-1.89401776e-01 1.07087851e+00 -7.78963983e-01 6.82648003e-01
5.93524754e-01 2.85720259e-01 -6.39375746e-01 1.22065198e+00
1.03276320e-01 1.07672787e+00 -2.99522996e-01 6.26797602e-02
-7.18857022e-03 -1.64805263e-01 3.63019854e-01 1.07822347e+00
2.48967230e-01 9.69936997e-02 2.65767425e-01 3.34850460e-01
3.92625064e-01 3.93269747e-01 1.19220488e-01 2.95379043e-01
6.34935558e-01 1.33060634e+00 -1.25789404e+00 -4.87155110e-01
-7.98172969e-03 4.16716814e-01 1.07406368e-02 2.21320782e-02
-6.89177394e-01 -2.43316174e-01 -1.53116428e-03 1.45313770e-01
4.97201756e-02 2.25601345e-01 -7.32859612e-01 -6.94606841e-01
-6.87038153e-02 -9.94337499e-01 9.31029975e-01 -3.16593230e-01
-1.28829598e+00 7.02828050e-01 1.44137785e-01 -1.63095880e+00
-8.29225779e-02 -2.88051575e-01 -5.39184093e-01 5.98717868e-01
-1.17401087e+00 -6.09271348e-01 -4.42055643e-01 3.71782273e-01
4.14205790e-01 -1.15597770e-01 6.84322834e-01 4.26061451e-01
-7.22483993e-01 6.66208863e-01 -1.18685663e-02 -1.38376787e-01
8.81872952e-01 -1.23587644e+00 -5.48576534e-01 8.24287117e-01
-2.97456026e-01 6.31257594e-01 9.48553026e-01 -6.25187039e-01
-6.32958770e-01 -8.47788274e-01 5.96240044e-01 -8.97123143e-02
2.63236523e-01 1.41556114e-01 -1.14631033e+00 3.69348787e-02
-6.26252517e-02 1.66398481e-01 8.76421690e-01 -8.84170607e-02
1.16824254e-01 -4.50866252e-01 -1.55155659e+00 2.48103753e-01
6.27074599e-01 9.00084600e-02 -6.02417171e-01 9.52507555e-02
1.15920603e-01 -5.82710564e-01 -1.02228165e+00 8.80851567e-01
6.73297822e-01 -1.08719647e+00 5.60794413e-01 -2.55720675e-01
3.27665120e-01 -6.15163028e-01 5.50221698e-03 -1.35514903e+00
-1.39092490e-01 -1.99363694e-01 7.67049417e-02 1.32933962e+00
4.90650326e-01 -5.07459283e-01 7.71645963e-01 4.85983372e-01
-1.58061743e-01 -1.09919834e+00 -9.14008439e-01 -3.09003055e-01
-1.23430744e-01 -4.29114819e-01 4.56922919e-01 1.04073215e+00
-1.71687886e-01 -4.93212491e-02 -1.76425099e-01 7.83430859e-02
9.95749056e-01 1.88216358e-01 2.96041548e-01 -1.33996737e+00
-1.43576026e-01 -3.31803888e-01 -3.73513162e-01 -4.41743508e-02
-2.18430251e-01 -7.35144734e-01 1.86601415e-01 -1.34411108e+00
4.00435686e-01 -8.32137883e-01 -8.78311634e-01 5.48688293e-01
-5.74079394e-01 5.39339781e-01 -1.86025500e-01 1.64178520e-01
-6.33551955e-01 1.34211078e-01 1.27162790e+00 2.83123571e-02
-2.07861602e-01 2.27696970e-01 -9.05367911e-01 6.44527316e-01
7.81924188e-01 -6.20517671e-01 -5.97098649e-01 1.10004432e-01
-1.65136352e-01 5.28015383e-02 2.55064592e-02 -1.42754352e+00
8.41491967e-02 -1.80572554e-01 6.50811017e-01 -7.35442281e-01
-3.34681094e-01 -8.60095084e-01 2.23577797e-01 4.49182838e-01
-4.61195797e-01 -1.81886315e-01 2.16932725e-02 3.59179080e-01
-1.90914229e-01 -4.20641452e-01 1.02214992e+00 9.44976509e-02
-2.52918780e-01 8.64036679e-02 -2.58347929e-01 3.57446098e-03
1.04984236e+00 -5.20879567e-01 -4.79378458e-03 -1.87492296e-02
-8.10333431e-01 4.56655949e-01 1.35225892e-01 2.07160041e-01
4.04665172e-01 -1.11953461e+00 -7.36033976e-01 1.29516408e-01
1.36004806e-01 3.20644006e-02 5.98215461e-01 9.84792888e-01
-1.94955543e-01 -1.11923181e-01 -2.51655072e-01 -8.11465859e-01
-1.24059522e+00 2.01956093e-01 3.62639338e-01 -4.58649963e-01
-3.63848209e-01 6.30315781e-01 -3.14154774e-01 -2.71831930e-01
2.98263162e-01 -3.08421552e-01 -6.55912519e-01 3.67656291e-01
3.73523891e-01 5.12389481e-01 5.23508310e-01 -5.51615059e-01
-6.18666291e-01 4.02092397e-01 -3.73592302e-02 1.47691771e-01
1.25272763e+00 -8.10690969e-02 -1.51822837e-02 4.79284376e-01
8.12235177e-01 -7.23689646e-02 -1.20966899e+00 -6.74843118e-02
1.22720622e-01 -3.15151215e-01 2.69111365e-01 -1.03184748e+00
-1.20453882e+00 4.74158496e-01 9.43081498e-01 1.41094357e-01
1.56411922e+00 -4.16500121e-01 4.24763411e-01 -1.75526589e-01
2.80421555e-01 -1.48034513e+00 -1.56321093e-01 -2.58583464e-02
3.98560256e-01 -1.16865945e+00 2.16585785e-01 -3.55635762e-01
-7.51818955e-01 1.11122620e+00 6.96026921e-01 -1.08886443e-01
5.65292597e-01 4.70606953e-01 2.23148197e-01 -4.40548211e-02
-6.47395968e-01 1.12175293e-01 4.00178969e-01 4.75235224e-01
5.44013679e-01 1.54922441e-01 -9.11436617e-01 6.96987331e-01
-5.02456315e-02 1.00449115e-01 3.46999973e-01 1.05537689e+00
-4.77844119e-01 -9.67362404e-01 -6.38182998e-01 1.05109012e+00
-7.71522164e-01 2.37724945e-01 1.54405117e-01 4.29529965e-01
5.46048820e-01 1.07543910e+00 1.22251570e-01 -3.90412211e-01
3.84466618e-01 2.94999659e-01 3.95431817e-01 -4.08589959e-01
-8.07636142e-01 1.99591234e-01 1.12821572e-02 -2.99444377e-01
-8.26227784e-01 -8.87355089e-01 -1.43466020e+00 1.27676740e-01
-7.47878194e-01 3.37529749e-01 4.71826166e-01 9.65899169e-01
4.34909463e-02 8.75744343e-01 7.69469321e-01 -5.95019817e-01
-7.29081988e-01 -1.00802493e+00 -4.87300992e-01 5.90802789e-01
2.59705484e-01 -1.05441606e+00 -5.10073423e-01 -7.45569468e-02] | [8.83029556274414, 4.110901355743408] |
16ebc5e4-552e-4551-8415-6a09cd5f9ebd | accurate-object-association-and-pose-updating | 2012.11368 | null | https://arxiv.org/abs/2012.11368v2 | https://arxiv.org/pdf/2012.11368v2.pdf | Accurate Object Association and Pose Updating for Semantic SLAM | Nowadays in the field of semantic SLAM, how to correctly use semantic information for data association is still a problem worthy of study. The key to solving this problem is to correctly associate multiple object measurements of one object landmark, and refine the pose of object landmark. However, different objects locating closely are prone to be associated as one object landmark, and it is difficult to pick up a best pose from multiple object measurements associated with one object landmark. To tackle these problems, we propose a hierarchical object association strategy by means of multiple object tracking, through which closing objects will be correctly associated to different object landmarks, and an approach to refine the pose of object landmark from multiple object measurements. The proposed method is evaluated on a simulated sequence and several sequences in the Kitti dataset. Experimental results show a very impressive improvement with respect to the traditional SLAM and the state-of-the-art semantic SLAM method. | ['Zhenhua Wang', 'Jianhua Zhang', 'Jialing Liu', 'Kaiqi Chen'] | 2020-12-21 | null | null | null | null | ['semantic-slam'] | ['computer-vision'] | [ 1.69377759e-01 -2.93694139e-01 1.35695368e-01 -3.92663121e-01
-7.20823109e-01 -3.08645785e-01 4.90603685e-01 7.31807828e-01
-6.28259778e-01 6.38509214e-01 -2.30109632e-01 3.90187323e-01
-7.32514441e-01 -5.53608060e-01 -6.91501498e-01 -7.39777088e-01
-2.88954586e-01 1.16441476e+00 9.62474704e-01 3.95799950e-02
5.90892196e-01 5.03469110e-01 -1.54209125e+00 -3.93001050e-01
6.28182471e-01 9.72952664e-01 8.56531203e-01 2.65246838e-01
-3.31265539e-01 2.91822881e-01 -7.59835303e-01 3.05945903e-01
1.71879873e-01 -2.73983389e-01 -6.25901282e-01 1.63232356e-01
4.36191916e-01 9.67902597e-03 3.83479178e-01 1.30385339e+00
2.57977277e-01 3.29252779e-01 2.85367608e-01 -1.36990678e+00
1.79489970e-01 4.13396418e-01 -5.39941072e-01 1.83674887e-01
3.75654221e-01 -4.44562137e-01 8.48488450e-01 -7.31652081e-01
6.01156294e-01 1.19807661e+00 5.77118158e-01 2.01366134e-02
-1.04452038e+00 -6.67353153e-01 2.26659521e-01 6.23874545e-01
-1.88046885e+00 -3.11774135e-01 6.49593711e-01 -3.69306564e-01
2.27816746e-01 1.99137274e-02 6.93857074e-01 2.37282097e-01
1.44400999e-01 5.17000198e-01 1.00822377e+00 -3.18473399e-01
3.41797352e-01 3.40927765e-02 2.62495667e-01 6.63611174e-01
4.87669051e-01 -1.80606216e-01 -7.96045005e-01 -1.79750785e-01
5.99778831e-01 1.82604402e-01 -2.52299547e-01 -1.02729285e+00
-1.44353735e+00 6.02021158e-01 7.10713148e-01 3.96198541e-01
-4.18160349e-01 3.23636323e-01 2.66227480e-02 -1.63788959e-01
1.00457437e-01 3.90788794e-01 -1.45398468e-01 -4.95509729e-02
-8.16040933e-01 4.32118267e-01 5.81475616e-01 1.17937815e+00
1.15814447e+00 -3.85739356e-01 4.14819360e-01 5.66943705e-01
6.62656963e-01 5.90576053e-01 1.81827322e-02 -7.38253236e-01
3.51706713e-01 6.30828083e-01 3.54737580e-01 -1.22943652e+00
-6.67164862e-01 -2.86306947e-01 -2.43611485e-01 3.85081798e-01
4.81149226e-01 4.18703914e-01 -7.31246948e-01 1.44401109e+00
6.78649306e-01 4.81761485e-01 9.14478395e-03 1.26038647e+00
5.88503540e-01 3.73558849e-01 1.06009334e-01 -1.51190624e-01
1.45055652e+00 -6.96604013e-01 -7.74850726e-01 -4.04927820e-01
4.55720901e-01 -8.60238969e-01 1.85736150e-01 3.64270806e-01
-5.27489603e-01 -6.81141436e-01 -1.09171021e+00 2.99923450e-01
-2.21837640e-01 5.63670732e-02 4.63943571e-01 9.89385694e-02
-5.26264429e-01 3.81580055e-01 -8.11512828e-01 -7.63284385e-01
-4.84070219e-02 4.24971819e-01 -5.47298074e-01 -1.57540843e-01
-8.32976282e-01 1.30021143e+00 1.10293698e+00 3.52306873e-01
-9.93162155e-01 -4.18849885e-01 -7.75478423e-01 -2.23989666e-01
8.52589071e-01 -4.43099946e-01 9.10229206e-01 -5.79428673e-01
-1.04274046e+00 5.63158989e-01 -1.49148971e-01 -5.13797283e-01
2.49369755e-01 -5.67541897e-01 -3.33654433e-01 2.34614741e-02
3.70660990e-01 3.99252713e-01 7.30633199e-01 -1.74305904e+00
-1.17976284e+00 -5.82143128e-01 -2.58524925e-01 4.64166641e-01
3.31953794e-01 -6.39533475e-02 -5.33676326e-01 1.12641994e-02
1.05277526e+00 -9.55288470e-01 -2.88212001e-01 -3.35503817e-02
8.40856880e-03 -3.10832590e-01 1.07748568e+00 -5.28259158e-01
7.81089425e-01 -2.05283308e+00 3.78678024e-01 4.17610705e-01
5.53970039e-02 1.17598884e-01 1.02205962e-01 3.38540107e-01
3.91006589e-01 -4.43793178e-01 4.63932054e-03 -4.56468761e-01
-1.53056830e-01 5.16957283e-01 -1.32308472e-02 8.53370070e-01
-3.42076480e-01 3.74127835e-01 -1.21439755e+00 -7.23033071e-01
4.65274543e-01 2.69734114e-01 -6.00057803e-02 2.44960979e-01
-1.68243930e-01 7.34163046e-01 -7.78960109e-01 6.51658297e-01
6.91562235e-01 5.53328618e-02 2.76637077e-01 -2.91674912e-01
-3.29086840e-01 4.95761484e-02 -1.82758069e+00 2.14605737e+00
-5.39670810e-02 2.46522456e-01 2.09790632e-01 -7.83924937e-01
1.37715399e+00 1.65466055e-01 6.16569817e-01 -3.64917308e-01
8.34273547e-02 6.00112200e-01 -1.06277734e-01 -4.05681640e-01
8.41324925e-01 -3.71476300e-02 -1.04792900e-01 5.03346696e-02
4.75399308e-02 -2.65554786e-01 4.35778834e-02 -9.03877914e-02
9.17114437e-01 2.88096249e-01 5.59886992e-01 -3.29833269e-01
8.19301307e-01 1.66569516e-01 6.94990396e-01 7.58244276e-01
-1.66338086e-01 3.66534114e-01 -1.99299380e-01 -5.29649496e-01
-8.39018106e-01 -1.08431399e+00 -1.33199662e-01 7.07983732e-01
1.14338601e+00 -3.97497863e-01 -3.35637987e-01 -6.41230404e-01
7.16551989e-02 3.32292199e-01 -2.49609694e-01 1.39453217e-01
-7.30354846e-01 -4.00870413e-01 3.38063687e-02 4.49528135e-02
4.98478860e-01 -9.12113905e-01 -1.08804369e+00 5.84519029e-01
-4.37804371e-01 -1.14733207e+00 -1.61385849e-01 2.34762952e-01
-7.73906946e-01 -1.35148525e+00 -2.84962952e-01 -6.18243098e-01
7.21386850e-01 5.93386650e-01 8.18621218e-01 4.82055157e-01
-2.02257484e-02 1.18336737e-01 -7.54399598e-01 -5.06895185e-01
-5.02350390e-01 1.12587124e-01 2.14592874e-01 2.62056559e-01
1.83816925e-01 -2.18342096e-01 -3.49982381e-01 7.23766744e-01
-6.84057832e-01 -3.57128493e-02 6.02533638e-01 4.82208759e-01
7.58432627e-01 2.51616925e-01 2.97386974e-01 -3.72164965e-01
-1.32435396e-01 -2.42025077e-01 -9.83332574e-01 3.05417538e-01
-4.86869931e-01 1.62170809e-02 7.64016137e-02 -1.42964065e-01
-7.09981263e-01 5.40281117e-01 1.88682050e-01 -3.66554022e-01
-2.38410741e-01 4.70101267e-01 -2.83417970e-01 -3.94970089e-01
2.22364381e-01 2.25271970e-01 -6.81454614e-02 -7.32868969e-01
2.36908585e-01 3.51396769e-01 6.58003151e-01 -5.63781440e-01
8.62508297e-01 5.99970400e-01 4.27393317e-01 -5.95347285e-01
-8.21253896e-01 -9.56538200e-01 -9.87311542e-01 -4.65042144e-01
9.56631005e-01 -7.47127354e-01 -6.01224542e-01 3.09359103e-01
-1.22484112e+00 3.72026831e-01 1.46897715e-02 7.66326010e-01
-5.53606331e-01 5.99145591e-01 1.33322671e-01 -8.03631723e-01
-6.94931969e-02 -1.29526234e+00 1.24318457e+00 2.49706417e-01
-2.15078026e-01 -8.12292635e-01 1.20294854e-01 3.15154821e-01
4.48632687e-02 1.78159967e-01 4.50059742e-01 -8.33013833e-01
-9.72321570e-01 -2.50567585e-01 -9.91946273e-03 -1.91020221e-01
2.18528524e-01 -3.61417204e-01 -3.83358151e-01 -5.36844730e-01
7.89091811e-02 2.30656698e-01 5.18189371e-01 -2.26242025e-03
3.69610488e-01 2.91657656e-01 -9.22362626e-01 4.70555484e-01
1.65308917e+00 4.01539356e-01 4.31372046e-01 6.44522846e-01
7.82767355e-01 5.52739322e-01 1.42646635e+00 4.33302492e-01
3.15948516e-01 1.31866097e+00 8.64545465e-01 3.56404454e-01
1.19506061e-01 -1.55499429e-01 -1.84057634e-02 6.66812479e-01
-1.67125780e-02 -1.22446030e-01 -9.65048194e-01 5.56940973e-01
-2.25913477e+00 -7.18248725e-01 -4.00048018e-01 2.31034303e+00
1.52871028e-01 4.56991568e-02 -2.08523095e-01 3.04220587e-01
8.95003498e-01 1.67488843e-01 -7.51844645e-02 4.06971365e-01
1.54335067e-01 -2.72940874e-01 7.39736438e-01 6.54776752e-01
-1.16587043e+00 9.86686170e-01 5.24532652e+00 7.05982625e-01
-8.03978503e-01 3.37864012e-01 -5.79053938e-01 3.45575958e-01
2.87429363e-01 5.17401218e-01 -1.21320844e+00 3.89107972e-01
2.97542185e-01 2.73427024e-04 1.67165533e-01 8.31824958e-01
9.93231088e-02 -7.93524563e-01 -1.02580822e+00 1.09509313e+00
3.58817011e-01 -9.90333378e-01 -9.65319127e-02 -1.91373955e-02
4.16400760e-01 -1.85083821e-01 -6.20185256e-01 -2.14043200e-01
-6.26621395e-02 -4.76508290e-01 1.16985381e+00 5.60551941e-01
4.03336175e-02 -6.42138302e-01 9.12760019e-01 5.79488933e-01
-1.64025307e+00 4.71584201e-02 -3.61430883e-01 4.67848638e-03
4.98312771e-01 3.89394045e-01 -1.33778489e+00 9.53860939e-01
8.09654593e-01 6.17426813e-01 -4.59270865e-01 1.71026373e+00
-3.95087600e-01 5.34815341e-03 -4.14741188e-01 -1.08638711e-01
1.41750515e-01 -2.65993744e-01 9.67987001e-01 7.68003225e-01
5.96559227e-01 1.12134509e-03 7.71899223e-01 5.94326556e-01
4.84773517e-01 1.01480849e-01 -3.73819917e-01 5.13568640e-01
6.60791576e-01 1.22138047e+00 -1.14846706e+00 -3.41146529e-01
-5.95735759e-02 8.21525395e-01 2.35395089e-01 -2.69735530e-02
-8.44769418e-01 -1.00923434e-03 6.11180782e-01 2.81647593e-02
2.37396449e-01 -8.17293584e-01 3.18742852e-04 -6.99781895e-01
1.28733786e-02 -4.02667433e-01 4.31919783e-01 -8.43227804e-01
-7.32871473e-01 3.88603181e-01 8.93397629e-02 -1.59491313e+00
1.33894742e-01 -1.70347452e-01 -8.42164904e-02 7.85051107e-01
-1.40081334e+00 -1.11848128e+00 -7.37271190e-01 3.44261497e-01
4.88483697e-01 9.88161489e-02 5.86866379e-01 3.19352537e-01
1.59066096e-01 -2.28534505e-01 -3.16836158e-05 -1.19732052e-01
6.97977483e-01 -9.47744906e-01 -1.17218487e-01 8.82066905e-01
5.35934806e-01 4.35488939e-01 1.07576096e+00 -1.01094270e+00
-1.40901387e+00 -8.51247668e-01 8.65471482e-01 -5.31548679e-01
4.17325824e-01 -1.82115674e-01 -1.00013041e+00 7.76987612e-01
-4.25668269e-01 -3.70616615e-02 5.31839989e-02 -9.21760593e-03
7.47747049e-02 -2.76820838e-01 -9.52503383e-01 9.62702557e-02
1.07561457e+00 -1.70365438e-01 -9.46568370e-01 3.21512163e-01
4.90121722e-01 -7.38872766e-01 -7.31314600e-01 6.50964737e-01
3.11571538e-01 -8.40219438e-01 9.71990705e-01 7.66895562e-02
-5.97363889e-01 -1.30743575e+00 -2.50296623e-01 -1.41820633e+00
-2.85340697e-01 -2.45830223e-01 1.17109872e-01 1.22031093e+00
-1.65583804e-01 -5.18739402e-01 6.72266901e-01 2.51078140e-02
-3.91680479e-01 -8.36892426e-02 -1.39093280e+00 -9.85925555e-01
-8.96884799e-01 -2.98396021e-01 6.67823017e-01 7.13905513e-01
-3.84304613e-01 1.95782542e-01 -4.68334764e-01 7.92519927e-01
1.09214270e+00 3.19632977e-01 1.22913730e+00 -1.57304490e+00
7.96465650e-02 -1.52877271e-01 -1.12394202e+00 -1.05471206e+00
9.40555930e-02 -6.46849275e-01 7.06067502e-01 -1.72630048e+00
6.67367363e-03 -1.07261932e+00 -2.29919016e-01 2.27860734e-01
-1.82643533e-01 2.54253358e-01 5.03576458e-01 3.09023678e-01
-1.08180332e+00 4.45341229e-01 9.33136463e-01 -1.09583493e-02
4.87268949e-03 -5.54757193e-02 1.13416210e-01 7.44788826e-01
3.87647212e-01 -8.95047486e-01 -1.26566917e-01 -5.45065463e-01
-1.53205749e-02 2.79350102e-01 6.07299566e-01 -1.56160378e+00
6.12799823e-01 -2.11152628e-01 1.81882437e-02 -1.02004242e+00
6.77616656e-01 -1.54443431e+00 7.24903464e-01 6.56601608e-01
9.09095332e-02 -5.27624972e-02 -5.97141869e-02 6.04309022e-01
-2.53122061e-01 -6.54570520e-01 8.12456548e-01 -2.38834128e-01
-1.56548274e+00 2.67665178e-01 2.28336477e-03 -3.96483839e-01
1.32061505e+00 -2.05504328e-01 -1.00103542e-01 7.41204396e-02
-7.66694367e-01 3.59501988e-01 4.71826345e-01 6.94993913e-01
7.16816366e-01 -1.28231490e+00 -5.12244999e-01 1.33553892e-01
4.34012145e-01 2.28893384e-01 1.07763551e-01 1.05397654e+00
-6.08249366e-01 2.45818213e-01 -2.05261186e-01 -1.17550421e+00
-1.58264077e+00 6.29971266e-01 2.58708447e-01 1.26252308e-01
-6.56707108e-01 5.29521525e-01 -4.53283973e-02 -1.45022094e-01
2.48693034e-01 -3.01202416e-01 -2.41459265e-01 9.78380665e-02
5.74999273e-01 4.94700938e-01 1.12856507e-01 -1.05305612e+00
-7.69839764e-01 1.18674970e+00 3.52676421e-01 -2.06104871e-02
1.08918607e+00 -5.51873565e-01 -2.50188828e-01 6.10451102e-01
7.62616873e-01 9.95617360e-03 -1.02897990e+00 -5.45634687e-01
4.66117263e-01 -8.49876881e-01 -3.46695870e-01 -4.44946378e-01
-6.62609518e-01 4.47735518e-01 7.07015932e-01 1.36522427e-01
7.19198585e-01 2.19410375e-01 5.04461944e-01 3.39556664e-01
1.29913199e+00 -8.17730188e-01 -9.79525372e-02 5.66455662e-01
7.90109515e-01 -1.23588157e+00 3.24344844e-01 -6.47943735e-01
-2.52989113e-01 1.02795494e+00 7.16429174e-01 -1.23299271e-01
3.90167475e-01 1.47061795e-01 2.20819861e-01 -3.15837353e-01
-6.13572150e-02 -4.10705656e-01 1.65378883e-01 3.73570174e-01
-2.26637289e-01 -1.99789461e-02 -4.97704268e-01 -2.19845376e-03
6.96629733e-02 -1.36140212e-01 1.24833778e-01 1.10578692e+00
-1.03922129e+00 -1.16956306e+00 -9.43274856e-01 -1.87417232e-02
-1.14068456e-01 4.19437647e-01 1.26328722e-01 8.11921060e-01
4.81981009e-01 8.88994217e-01 1.06343225e-01 -3.55780333e-01
3.19446295e-01 4.48400266e-02 5.15099049e-01 -6.19289279e-01
-3.34670842e-01 5.64494990e-02 -3.31898816e-02 -5.76958060e-01
-8.35844517e-01 -7.97487080e-01 -1.81210983e+00 1.19231097e-01
-6.65325820e-01 5.71913481e-01 1.05851543e+00 1.18923509e+00
-1.67978872e-02 2.61111200e-01 2.33359158e-01 -1.12141824e+00
-2.87889987e-01 -7.39980817e-01 -6.70881927e-01 5.45084536e-01
2.48534471e-01 -1.18501127e+00 -2.30407134e-01 -1.20081790e-01] | [7.298513412475586, -2.248411178588867] |
f7b68b9c-6cc8-4b32-943e-f87f32ebf2f5 | transnet-category-level-transparent-object | 2208.10002 | null | https://arxiv.org/abs/2208.10002v1 | https://arxiv.org/pdf/2208.10002v1.pdf | TransNet: Category-Level Transparent Object Pose Estimation | Transparent objects present multiple distinct challenges to visual perception systems. First, their lack of distinguishing visual features makes transparent objects harder to detect and localize than opaque objects. Even humans find certain transparent surfaces with little specular reflection or refraction, e.g. glass doors, difficult to perceive. A second challenge is that common depth sensors typically used for opaque object perception cannot obtain accurate depth measurements on transparent objects due to their unique reflective properties. Stemming from these challenges, we observe that transparent object instances within the same category (e.g. cups) look more similar to each other than to ordinary opaque objects of that same category. Given this observation, the present paper sets out to explore the possibility of category-level transparent object pose estimation rather than instance-level pose estimation. We propose TransNet, a two-stage pipeline that learns to estimate category-level transparent object pose using localized depth completion and surface normal estimation. TransNet is evaluated in terms of pose estimation accuracy on a recent, large-scale transparent object dataset and compared to a state-of-the-art category-level pose estimation approach. Results from this comparison demonstrate that TransNet achieves improved pose estimation accuracy on transparent objects and key findings from the included ablation studies suggest future directions for performance improvements. | ['Odest Chadwicke Jenkins', 'Zeren Yu', 'Jiyue Zhu', 'Xiaotong Chen', 'Anthony Opipari', 'Huijie Zhang'] | 2022-08-22 | null | null | null | null | ['transparent-objects', 'depth-completion'] | ['computer-vision', 'computer-vision'] | [ 3.55474591e-01 1.80357829e-01 3.14760417e-01 -4.61921573e-01
-6.84214056e-01 -7.35115647e-01 3.59843194e-01 8.50816295e-02
-1.60516962e-01 2.33002052e-01 7.00268000e-02 1.32822758e-02
1.96383134e-01 -4.89576072e-01 -9.38181281e-01 -4.89700675e-01
-1.60014421e-01 4.03178871e-01 6.68816149e-01 2.10067838e-01
5.35179198e-01 5.62543392e-01 -1.87131751e+00 5.86515903e-01
5.49755931e-01 1.29654992e+00 5.28540432e-01 6.11107111e-01
6.24631234e-02 3.57166916e-01 -6.11898720e-01 -2.30645284e-01
5.78576267e-01 4.17457283e-01 -4.16868985e-01 1.24982864e-01
1.58231938e+00 -7.65865803e-01 2.35767215e-01 6.08914912e-01
2.74285078e-01 6.11293875e-02 8.98664534e-01 -8.75695646e-01
-7.48048425e-01 -1.03495061e-01 -7.60985374e-01 9.49965641e-02
7.25523591e-01 2.50611693e-01 9.39759016e-01 -1.03615105e+00
5.50729573e-01 1.46198094e+00 5.57213366e-01 4.84039903e-01
-1.04477215e+00 -2.86900252e-01 7.00807989e-01 -8.46515149e-02
-1.05805790e+00 -4.38198119e-01 4.27573442e-01 -5.54049611e-01
1.01289833e+00 4.32341754e-01 6.16659105e-01 9.43359017e-01
3.32154751e-01 4.93838459e-01 1.35218024e+00 -9.22411457e-02
3.01312119e-01 4.25152719e-01 -3.15971971e-02 6.66621625e-01
7.29594171e-01 2.13163979e-02 -7.35958338e-01 -2.76002977e-02
6.23887539e-01 -1.99619472e-01 -5.34301758e-01 -7.32083082e-01
-1.16058624e+00 2.76972055e-01 8.24378192e-01 -2.18539134e-01
-1.47692278e-01 2.13827267e-01 -6.98192045e-02 2.12539621e-02
5.96549690e-01 4.66140985e-01 -4.44132715e-01 -2.78936978e-02
-2.91783601e-01 1.30055010e-01 7.50668287e-01 9.78213906e-01
7.18668997e-01 -1.29314393e-01 -2.96072904e-02 5.93437731e-01
8.72182310e-01 4.50321048e-01 -1.10914558e-01 -7.74413943e-01
6.99104369e-01 6.07073903e-01 3.86440009e-01 -6.38085067e-01
-4.47033674e-01 -2.55166054e-01 1.36812493e-01 8.63538623e-01
6.64291382e-01 3.90143275e-01 -1.17418957e+00 1.40505207e+00
6.54289424e-01 -1.39732584e-01 -2.39951342e-01 1.29651606e+00
1.16938722e+00 1.98262915e-01 -9.16677639e-02 3.36814463e-01
1.65853333e+00 -7.88510203e-01 -1.41282231e-01 -6.74647391e-01
2.06355542e-01 -1.00321841e+00 1.47997010e+00 6.53790534e-01
-1.10750723e+00 -3.63400400e-01 -1.12268043e+00 -3.40818763e-01
-2.24548146e-01 3.51959586e-01 1.11211336e+00 8.84849846e-01
-8.79920185e-01 2.00746208e-01 -8.36830080e-01 -5.52206695e-01
5.20147622e-01 3.34288508e-01 -3.93208236e-01 -7.13009536e-02
-2.83788949e-01 9.22000110e-01 -3.15670148e-02 3.27116191e-01
-1.19883275e+00 -1.00007272e+00 -8.54156315e-01 -1.39976412e-01
2.35564068e-01 -7.94952333e-01 1.33481252e+00 -6.64716244e-01
-1.37377882e+00 1.24791467e+00 -5.29086947e-01 2.55099572e-02
5.88003695e-01 -6.33590341e-01 7.26428255e-03 3.84362429e-01
1.39733348e-02 5.03916144e-01 1.09195757e+00 -1.91447508e+00
-4.16704416e-01 -6.25324547e-01 3.18687677e-01 4.75345492e-01
-1.50622919e-01 1.80234984e-02 -2.82200485e-01 -8.72329101e-02
8.02994967e-01 -8.67154717e-01 2.08079785e-01 8.56503963e-01
-6.39447093e-01 -1.65548831e-01 9.20556903e-01 -3.76699686e-01
2.94835418e-01 -1.91229844e+00 -2.62216270e-01 -1.22477990e-02
7.01986611e-01 -3.96029688e-02 6.05447590e-02 1.58494100e-01
5.13750426e-02 -1.33131921e-01 1.69767253e-02 -8.29612970e-01
-3.13754864e-02 -2.93877512e-01 -3.61971468e-01 8.09713304e-01
1.58488620e-02 6.14363790e-01 -5.83480835e-01 -1.60714999e-01
2.67122477e-01 5.77087879e-01 -3.03596377e-01 6.83862865e-02
-3.95641476e-01 5.14167726e-01 -3.40332359e-01 1.08165109e+00
1.13188612e+00 -1.03839003e-01 -1.80917233e-01 -4.79057431e-01
-2.49917850e-01 6.88510716e-01 -1.04849029e+00 1.46574152e+00
-4.84342963e-01 8.20662498e-01 6.26676306e-02 -9.66252014e-02
6.50507510e-01 1.22958384e-01 -3.74253164e-03 -4.87723649e-01
9.32246726e-03 2.62526512e-01 1.07839495e-01 -4.78959382e-01
4.33782876e-01 1.10756449e-01 4.31908965e-01 3.97923619e-01
-5.67415297e-01 -6.82296157e-01 -2.99844772e-01 2.16572527e-02
9.25538182e-01 4.51277405e-01 2.71420367e-02 -8.22929665e-03
-1.60277814e-01 -3.06000233e-01 2.82737851e-01 6.56111658e-01
-1.40433371e-01 1.15058172e+00 -5.22366092e-02 -5.74884117e-01
-6.42744243e-01 -1.96652055e+00 -5.29539585e-01 9.16688800e-01
5.09213150e-01 -1.08765885e-01 -4.30534005e-01 -3.49041432e-01
3.08502734e-01 1.69910640e-01 -6.88242078e-01 3.89153138e-02
-3.35105121e-01 -3.98516446e-01 -2.44806394e-01 5.21636546e-01
4.81019616e-01 -9.19931114e-01 -1.12980175e+00 -1.90843180e-01
-4.86223912e-03 -1.44239557e+00 -1.66672856e-01 -1.04938775e-01
-9.69914854e-01 -1.07718086e+00 -5.63460231e-01 -5.71475804e-01
9.12902892e-01 8.91910672e-01 1.39355099e+00 -1.33744091e-01
-6.45563841e-01 6.39152527e-01 -1.35691002e-01 -8.24707031e-01
1.08734481e-01 -2.51519591e-01 3.76803279e-01 -4.67506982e-02
1.37392655e-01 -4.76427138e-01 -1.10040677e+00 4.40073550e-01
-4.67803597e-01 -1.72933325e-01 6.49899840e-01 -1.33153796e-01
4.86931384e-01 -4.40513313e-01 -9.78232697e-02 -4.61404234e-01
7.68790990e-02 -2.31763706e-01 -6.95010722e-01 1.31889403e-01
1.49011426e-02 -3.92269462e-01 7.48601779e-02 -6.67474568e-01
-1.17039478e+00 -7.67845362e-02 4.89620596e-01 -2.63298035e-01
-2.67560035e-01 -1.76602542e-01 -1.30723476e-01 -4.87999469e-01
8.58650506e-01 -1.71481103e-01 -4.84715641e-01 -4.33397889e-01
8.81953514e-04 5.40342391e-01 1.61783174e-01 -5.07927418e-01
1.18768597e+00 1.25167000e+00 -1.13616779e-01 -1.11902332e+00
-1.12396717e+00 -4.60701793e-01 -3.85782182e-01 -2.88683385e-01
9.46879804e-01 -1.05342412e+00 -1.39712286e+00 5.25257468e-01
-1.15351737e+00 -4.49437350e-01 5.85647412e-02 7.16143250e-01
-4.41652745e-01 2.58013427e-01 -4.07255471e-01 -9.78606939e-01
5.89102842e-02 -9.99406576e-01 1.61461580e+00 3.02329719e-01
-3.60443234e-01 -6.66773379e-01 -2.12521791e-01 7.48798549e-01
1.65049240e-01 7.39288628e-02 6.54430747e-01 2.84818560e-01
-1.46461391e+00 2.06988771e-03 -5.07020116e-01 1.57275125e-02
3.44141990e-01 1.71302170e-01 -1.64380383e+00 -5.16555548e-01
1.36904761e-01 -5.59061348e-01 7.91546404e-01 5.56948781e-01
1.11059451e+00 3.02098766e-02 -5.05881250e-01 7.97554374e-01
1.34277689e+00 -1.00461476e-01 5.38311720e-01 3.94130737e-01
9.63020802e-01 9.46294367e-01 7.34806597e-01 2.46536210e-01
5.15313983e-01 9.53592479e-01 8.73976707e-01 -9.71482843e-02
-3.91349852e-01 4.21497524e-02 2.32533067e-01 4.46206555e-02
-2.38235429e-01 -1.66454643e-01 -7.97904611e-01 2.98280180e-01
-9.44073915e-01 -5.71602404e-01 -4.19918388e-01 2.57965946e+00
4.71591562e-01 4.58839506e-01 -6.62688166e-02 -1.90650269e-01
1.07207246e-01 -1.88085392e-01 -6.73314512e-01 -4.26670760e-01
-8.40597376e-02 -1.77769035e-01 4.56352025e-01 3.79677773e-01
-8.51610959e-01 7.57567286e-01 6.29827309e+00 1.09820671e-01
-1.09892356e+00 -3.11844423e-02 2.78068841e-01 -3.62935632e-01
-4.85203296e-01 1.01945758e-01 -8.99868667e-01 1.25779197e-01
2.15803728e-01 6.50493443e-01 3.05449694e-01 7.40422606e-01
1.51519209e-01 -8.11861157e-01 -1.59259224e+00 9.78909135e-01
1.83076277e-01 -6.63753450e-01 -5.53407483e-02 1.57827675e-01
6.77426934e-01 3.93474191e-01 5.77694774e-01 -2.19548687e-01
-4.01920313e-03 -1.08284009e+00 1.18349135e+00 2.56076872e-01
5.95865309e-01 -8.64880811e-03 6.85847476e-02 -1.23001384e-02
-1.46412218e+00 -2.42219046e-02 -6.02421641e-01 -2.06191078e-01
-9.93419513e-02 4.91504699e-01 -1.08349633e+00 6.12987876e-02
1.20113564e+00 3.87949973e-01 -8.08272719e-01 1.49671888e+00
-1.87777206e-01 2.63358176e-01 -4.59409475e-01 5.19141853e-02
-1.18170172e-01 2.92665102e-02 6.89942241e-01 7.52181411e-01
4.49896045e-03 -2.10452735e-01 -3.51505056e-02 1.27018154e+00
6.79985359e-02 -2.65387356e-01 -5.97717047e-01 2.88926691e-01
4.76253480e-01 1.29588044e+00 -8.20739150e-01 -2.75419280e-02
-4.94462788e-01 9.49051857e-01 4.94742304e-01 7.57206202e-01
-6.98827922e-01 -8.47243071e-02 1.04401553e+00 3.78715515e-01
1.78796276e-01 -4.54612523e-01 -5.82241952e-01 -1.08669996e+00
8.45904529e-01 -4.21378970e-01 -1.29952803e-01 -1.29801214e+00
-1.13033271e+00 4.61531311e-01 -4.89021558e-03 -1.64015651e+00
4.52006906e-01 -1.07081902e+00 -5.09126425e-01 9.51327503e-01
-1.68535435e+00 -1.21898460e+00 -9.22983825e-01 2.93107301e-01
4.80223209e-01 3.09302181e-01 5.50819695e-01 -2.39576980e-01
-2.95293406e-02 5.42606652e-01 -1.94355398e-01 -3.74811083e-01
6.49622917e-01 -1.21075749e+00 4.73649383e-01 5.20596027e-01
1.21147022e-01 1.04477334e+00 7.82590866e-01 -5.80707967e-01
-1.65562570e+00 -5.63270748e-01 2.19860166e-01 -1.24028707e+00
2.27966666e-01 -1.03766692e+00 -9.16628838e-01 7.00854540e-01
-3.74183089e-01 2.05762193e-01 1.72684982e-01 1.70110658e-01
-8.88892293e-01 -1.21532001e-01 -1.31095111e+00 8.69256556e-01
1.22778165e+00 -6.78959012e-01 -4.31397140e-01 4.80780959e-01
6.80493236e-01 -7.57077456e-01 -6.50481224e-01 4.12517399e-01
8.92108679e-01 -1.43938518e+00 1.34922206e+00 -1.22287031e-02
2.47751132e-01 -3.19932818e-01 -1.65305182e-01 -1.00922763e+00
1.36250481e-01 -3.71023983e-01 -1.83363691e-01 8.14777434e-01
3.44278485e-01 -9.38627183e-01 1.03601265e+00 7.16589570e-01
-4.32323456e-01 -5.97831309e-01 -1.08409989e+00 -8.06902289e-01
-3.79706413e-01 -5.53251445e-01 1.84901655e-01 4.19413596e-01
-5.14487624e-01 -4.64430526e-02 2.96856374e-01 6.86131775e-01
9.71452832e-01 4.36834544e-01 8.51439476e-01 -1.50498760e+00
-3.43179792e-01 -2.39960775e-01 -6.64508164e-01 -1.16952872e+00
-1.13024846e-01 -4.26611155e-01 2.45667621e-01 -2.00089598e+00
3.15840214e-01 -7.28384972e-01 2.69418985e-01 1.88584819e-01
-7.10298717e-02 6.16794705e-01 1.24872454e-01 2.59187251e-01
-3.51252705e-01 4.78496164e-01 1.68266034e+00 -3.09734121e-02
-2.40665913e-01 1.30159616e-01 -6.31170213e-01 9.14161861e-01
5.90161324e-01 -1.97241351e-01 -5.21020830e-01 -8.35301459e-01
3.34478676e-01 -4.54258233e-01 7.18562663e-01 -1.23571563e+00
-3.64685178e-01 -1.21998101e-01 5.01043499e-01 -5.58373868e-01
1.10789192e+00 -9.05971646e-01 -3.82948220e-01 2.72742987e-01
1.73186526e-01 -4.85365808e-01 4.86652076e-01 6.18944049e-01
3.73256147e-01 -4.87383129e-03 6.81941330e-01 -2.17789799e-01
-6.80168927e-01 3.19517761e-01 -2.42998570e-01 -1.71346087e-02
9.80522215e-01 -1.08621764e+00 -8.50384891e-01 -2.30389297e-01
-3.74387354e-01 -2.31999412e-01 8.03404510e-01 7.34100461e-01
1.06429887e+00 -7.72216737e-01 -5.97182751e-01 1.42429294e-02
6.83885813e-01 3.05878997e-01 1.77393377e-01 8.64250481e-01
-5.55957735e-01 3.27209055e-01 -4.30752337e-03 -9.57445502e-01
-1.52250874e+00 1.05693027e-01 5.64972997e-01 5.35648406e-01
-7.31864750e-01 1.30024540e+00 1.12102234e+00 -4.26623434e-01
2.43118882e-01 -9.20194328e-01 2.62622088e-01 -3.88700217e-01
4.56398815e-01 2.70952046e-01 3.29958498e-01 -3.85624975e-01
-5.24482310e-01 1.05809665e+00 -1.38862699e-01 -8.76336768e-02
1.05146146e+00 -2.46625498e-01 -4.38467599e-02 6.72277749e-01
1.06578410e+00 3.61290812e-01 -1.68919790e+00 1.01492479e-02
-3.59543025e-01 -9.26063299e-01 -2.74533778e-01 -8.11455905e-01
-7.96494484e-01 9.81917262e-01 8.10827911e-01 -2.27639675e-01
7.15197682e-01 5.20820677e-01 3.97246957e-01 5.89804292e-01
7.15009570e-01 -7.51146317e-01 5.40224016e-01 2.21339136e-01
1.07161915e+00 -1.51916897e+00 1.64767563e-01 -1.08588064e+00
-3.29056442e-01 8.48556697e-01 1.26778448e+00 9.70640257e-02
2.01385915e-01 1.73360094e-01 3.16695720e-01 -5.39516032e-01
-5.15109956e-01 -1.60811841e-01 7.58097231e-01 1.03406549e+00
5.72321296e-01 -9.29947570e-02 4.10948098e-01 -2.55394459e-01
-3.75283778e-01 -6.22519076e-01 5.49483538e-01 9.69254375e-01
-6.28093302e-01 -4.37027186e-01 -6.83954537e-01 5.46721220e-01
-1.55826747e-01 -5.50762303e-02 -5.08407295e-01 6.57620430e-01
-7.28441477e-02 9.91324961e-01 3.79098654e-01 -2.17520799e-02
2.41850510e-01 -5.83962381e-01 1.08131921e+00 -1.18883705e+00
-3.17485869e-01 -3.00908953e-01 3.81301604e-02 -9.17566001e-01
-3.07246864e-01 -6.22950137e-01 -1.21292794e+00 2.04567283e-01
-2.62463242e-01 -4.72713739e-01 1.10163510e+00 5.49632847e-01
5.19573316e-02 4.32780117e-01 3.03540975e-01 -1.33538747e+00
-3.77131850e-01 -9.67258632e-01 -5.71596026e-01 4.15079385e-01
6.17082715e-01 -1.00452971e+00 -5.72722137e-01 -2.92301148e-01] | [7.178958415985107, -2.1280999183654785] |
03ae5097-fdaf-4dd5-b4d0-86d7f60fe4c9 | deep-audio-visual-speech-recognition | 1809.02108 | null | http://arxiv.org/abs/1809.02108v2 | http://arxiv.org/pdf/1809.02108v2.pdf | Deep Audio-Visual Speech Recognition | The goal of this work is to recognise phrases and sentences being spoken by a
talking face, with or without the audio. Unlike previous works that have
focussed on recognising a limited number of words or phrases, we tackle lip
reading as an open-world problem - unconstrained natural language sentences,
and in the wild videos. Our key contributions are: (1) we compare two models
for lip reading, one using a CTC loss, and the other using a
sequence-to-sequence loss. Both models are built on top of the transformer
self-attention architecture; (2) we investigate to what extent lip reading is
complementary to audio speech recognition, especially when the audio signal is
noisy; (3) we introduce and publicly release a new dataset for audio-visual
speech recognition, LRS2-BBC, consisting of thousands of natural sentences from
British television. The models that we train surpass the performance of all
previous work on a lip reading benchmark dataset by a significant margin. | ['Triantafyllos Afouras', 'Joon Son Chung', 'Andrew Zisserman', 'Oriol Vinyals', 'Andrew Senior'] | 2018-09-06 | null | null | null | null | ['lipreading', 'audio-visual-speech-recognition'] | ['computer-vision', 'speech'] | [ 6.73792422e-01 4.72792566e-01 -1.89521611e-01 -1.69541806e-01
-1.44413662e+00 -4.06947404e-01 8.08990538e-01 -4.38540757e-01
-3.25591743e-01 5.32339215e-01 7.88136363e-01 -3.33549023e-01
4.59253550e-01 4.52927575e-02 -9.67723012e-01 -6.49801493e-01
2.61937290e-01 2.13410854e-01 2.53237784e-01 7.86219090e-02
2.40744546e-01 1.29182845e-01 -2.01304817e+00 4.26198870e-01
3.62203270e-01 1.36297047e+00 1.94005489e-01 1.12986815e+00
-6.66501373e-02 9.95153844e-01 -3.29861343e-01 -3.98258597e-01
3.76528092e-02 -5.99793971e-01 -1.13609350e+00 1.66737571e-01
8.89783859e-01 -2.46474758e-01 -1.65186048e-01 7.99683213e-01
1.25462925e+00 -1.43030956e-01 6.24676108e-01 -1.24308860e+00
-4.42739308e-01 4.54538524e-01 -2.04967096e-01 9.85708367e-03
8.28804612e-01 3.27867359e-01 9.90605056e-01 -1.16569436e+00
4.36443329e-01 1.58168852e+00 6.98257923e-01 8.63202929e-01
-1.17451227e+00 -4.48623687e-01 -6.44173250e-02 4.82253194e-01
-1.32344365e+00 -1.63390625e+00 6.94065571e-01 -2.93603420e-01
1.12646437e+00 2.39632979e-01 3.26932162e-01 1.64415181e+00
-2.01183364e-01 1.32877421e+00 1.06151223e+00 -6.69041514e-01
1.31424397e-01 1.85336992e-01 -1.84162393e-01 2.35773876e-01
-6.81054831e-01 8.53318125e-02 -1.07355130e+00 9.52582732e-02
-3.10929753e-02 -9.46769357e-01 -7.73523450e-01 -3.30259532e-01
-1.26364636e+00 7.79961646e-01 -1.95920333e-01 8.03724490e-03
-1.87867716e-01 1.53528154e-01 7.28997648e-01 3.41103762e-01
4.22786295e-01 -2.40832880e-01 -4.58614379e-01 -3.94463897e-01
-1.17194343e+00 -1.00183986e-01 1.05821764e+00 1.05004525e+00
2.48303771e-01 -4.75241356e-02 -1.75763771e-01 1.10661209e+00
5.32991529e-01 8.17234099e-01 7.70091593e-01 -7.15534508e-01
5.11625290e-01 -4.48217452e-01 -2.67401248e-01 -4.14204836e-01
-6.25685602e-02 1.95757002e-01 -6.21286809e-01 1.01212472e-01
3.66273671e-01 -6.41994849e-02 -7.97043681e-01 1.75819135e+00
-3.98052037e-02 3.71102393e-01 3.49558681e-01 5.59945107e-01
1.30519593e+00 4.96442080e-01 -1.05909847e-01 -2.97454089e-01
1.46915722e+00 -1.19928050e+00 -7.89010882e-01 -1.93708703e-01
-8.61861855e-02 -1.00997436e+00 1.14256775e+00 3.78597766e-01
-1.58024383e+00 -6.16933405e-01 -7.47795284e-01 -4.09797370e-01
-2.60121286e-01 7.87629634e-02 -7.55373761e-02 8.12543929e-01
-1.78011000e+00 1.11598313e-01 -3.49711329e-01 -4.71124440e-01
3.72743398e-01 2.80382812e-01 -3.62186819e-01 1.64131314e-01
-1.22520971e+00 7.23652244e-01 -4.94358167e-02 -6.24047108e-02
-9.78250086e-01 -6.24463558e-01 -1.21066058e+00 7.09588155e-02
4.13988680e-01 -4.89790410e-01 1.59904695e+00 -1.11427641e+00
-2.05670762e+00 1.20617890e+00 -7.38869727e-01 -6.81757927e-01
7.47152925e-01 -1.39824390e-01 -1.76129922e-01 4.21966255e-01
-3.61870080e-02 1.20892894e+00 1.34658372e+00 -1.07535410e+00
-4.11480337e-01 -2.31121830e-03 -3.93178582e-01 3.51614684e-01
2.13693336e-01 3.51225317e-01 -5.48597991e-01 -3.93761218e-01
-4.19983536e-01 -5.55291891e-01 6.42239988e-01 -1.39688794e-02
-5.74345350e-01 -4.20305103e-01 9.38414454e-01 -9.31548119e-01
5.40153801e-01 -2.28262472e+00 2.23199472e-01 -3.94931018e-01
-1.68618292e-01 3.76625836e-01 -3.63776475e-01 3.58210236e-01
-3.45242947e-01 1.90541610e-01 -1.13799170e-01 -7.80517876e-01
2.41405755e-01 -2.21425474e-01 -5.92318237e-01 4.42694008e-01
3.50509167e-01 1.07574999e+00 -5.66080213e-01 -7.17621148e-01
1.36686802e-01 7.22280622e-01 -3.60229522e-01 2.41506919e-01
-1.48320436e-01 1.77366078e-01 2.40329951e-01 8.81556213e-01
6.51701033e-01 1.17918797e-01 -3.62296373e-01 -7.90245682e-02
-6.52223825e-02 6.27064884e-01 -8.62862647e-01 1.71206760e+00
-5.76869786e-01 1.18114901e+00 5.42095780e-01 -9.13180888e-01
5.27162135e-01 9.62762296e-01 2.16260001e-01 -7.33172536e-01
-6.61134347e-03 1.09780051e-01 -4.45476383e-01 -7.75782883e-01
1.17127977e-01 -2.29703054e-01 2.64269203e-01 2.67606109e-01
3.74162465e-01 -4.02978063e-01 -9.89742205e-02 6.98773982e-03
8.03783715e-01 -9.02191699e-02 1.56034455e-01 -2.43460611e-01
1.00003612e+00 -8.02097619e-01 -1.79089848e-02 6.80168867e-01
-4.21758741e-01 9.24889684e-01 7.57631719e-01 2.44028240e-01
-8.32276762e-01 -1.14152598e+00 -4.03133363e-01 1.31413257e+00
-2.07118437e-01 -1.38790041e-01 -9.60200012e-01 -5.26582778e-01
-1.49282008e-01 5.35088718e-01 -3.48412186e-01 1.74041674e-01
-1.42507404e-01 1.58027355e-02 8.93876493e-01 5.50178409e-01
5.44200122e-01 -1.48634350e+00 -3.02731484e-01 -8.66805837e-02
-6.16658926e-01 -1.35666478e+00 -8.90275717e-01 6.66617975e-02
-3.12303733e-02 -8.95056665e-01 -1.31992006e+00 -1.13757479e+00
-2.37826295e-02 -6.07834831e-02 1.36160159e+00 -1.36958495e-01
-1.77779540e-01 1.00360906e+00 -1.72529593e-01 -7.94440925e-01
-5.94697416e-01 1.02411695e-01 9.36450213e-02 2.27209136e-01
3.47256362e-01 -2.51235962e-01 -3.15270752e-01 2.61940569e-01
-5.28108478e-01 -5.68313710e-02 5.79494536e-01 1.03043199e+00
1.70640424e-01 -3.85346860e-01 1.00080466e+00 -1.08987816e-01
6.52922809e-01 -1.75170392e-01 -3.28407466e-01 2.57978201e-01
2.68017743e-02 -1.91090673e-01 2.20810741e-01 -4.94794816e-01
-8.31893086e-01 9.05948877e-02 -8.66710603e-01 -4.00149286e-01
-4.90847081e-01 -8.55543911e-02 -6.04108930e-01 3.95330824e-02
1.27713874e-01 6.16607785e-01 4.54959363e-01 -5.18753231e-01
2.63515323e-01 1.27820492e+00 8.73347819e-01 -3.15615445e-01
2.17986211e-01 2.58378655e-01 -2.85104841e-01 -1.49230492e+00
-3.99321735e-01 -8.65303397e-01 -6.07328892e-01 -2.03447625e-01
1.10984087e+00 -1.05895162e+00 -1.19784701e+00 9.29906666e-01
-1.27937257e+00 -5.38485110e-01 -2.51027256e-01 2.73675621e-01
-1.26806366e+00 5.78183353e-01 -5.63882589e-01 -1.19864511e+00
-3.69333684e-01 -1.42737281e+00 1.78584826e+00 -2.69953534e-02
-7.26043433e-02 -8.28557014e-01 -1.92870870e-02 6.06602550e-01
5.24173379e-01 -3.74181956e-01 5.86958587e-01 -6.99785650e-01
-4.59979832e-01 2.04354107e-01 -2.57998198e-01 7.51020968e-01
-3.67474370e-02 -3.21033657e-01 -1.73547840e+00 -2.94169307e-01
7.62899593e-02 -1.03331351e+00 1.24790132e+00 5.03793478e-01
1.05448353e+00 -3.91993731e-01 -5.49104847e-02 3.95881653e-01
9.80549514e-01 7.78881600e-03 9.75748777e-01 -3.19834977e-01
2.17689201e-01 7.47971892e-01 -7.21954834e-03 -9.28874165e-02
2.47488037e-01 7.77929544e-01 3.72792959e-01 -1.91402778e-01
-5.65999985e-01 -4.62313175e-01 7.45710611e-01 8.06436718e-01
3.71822327e-01 -5.36917388e-01 -8.64358366e-01 9.17519689e-01
-1.39188635e+00 -1.13741100e+00 3.10668379e-01 2.01708698e+00
8.73788059e-01 -4.59735654e-02 3.97972256e-01 4.04486418e-01
7.27693319e-01 4.03613150e-01 -4.54175919e-01 -3.64077568e-01
-2.00565726e-01 4.09702152e-01 1.86130375e-01 7.89431155e-01
-1.25730681e+00 9.08071280e-01 7.17616510e+00 8.87731850e-01
-1.31273782e+00 -1.99710149e-02 5.09530067e-01 -3.05758491e-02
1.00580893e-01 -4.93705004e-01 -8.25543106e-01 5.10440767e-01
1.27870536e+00 1.09508984e-01 3.27036619e-01 6.03485942e-01
1.01074740e-01 -1.26200497e-01 -1.33364570e+00 1.07809043e+00
7.36110210e-01 -1.00287986e+00 -1.73792213e-01 -1.65199675e-02
1.07587583e-01 2.22493336e-01 3.89476538e-01 2.92587310e-01
-1.97617650e-01 -1.44961405e+00 9.57993925e-01 6.52498484e-01
1.28165519e+00 -3.87629956e-01 7.60589063e-01 2.79565483e-01
-1.11695170e+00 1.75976604e-01 -6.20291531e-02 3.80418688e-01
3.57094824e-01 -6.09693006e-02 -9.55569386e-01 2.50012666e-01
6.94347560e-01 6.51034296e-01 -3.76188457e-01 9.99910533e-01
-1.16055235e-01 8.79522920e-01 -2.41644740e-01 -6.12115078e-02
6.71190470e-02 5.06041467e-01 6.94740236e-01 1.47581053e+00
-4.66058543e-03 -3.91163617e-01 -2.06668302e-01 5.54519475e-01
-2.59736329e-01 1.40152067e-01 -6.50004089e-01 1.39793241e-03
8.92009512e-02 7.59277344e-01 -1.19715221e-01 -1.98086783e-01
-5.78494966e-01 9.24441576e-01 -2.22721174e-01 4.22668099e-01
-5.36737084e-01 -4.07759547e-01 6.07974291e-01 1.65958963e-02
5.64464629e-01 2.93947846e-01 2.07441777e-01 -1.00173903e+00
1.19828463e-01 -1.21885324e+00 -1.09725975e-01 -1.12638211e+00
-1.21090531e+00 5.79448402e-01 -3.49483877e-01 -7.64888704e-01
-6.30867362e-01 -6.77493870e-01 -5.22285163e-01 1.03371704e+00
-1.96874142e+00 -1.11208618e+00 -1.11964688e-01 6.78200305e-01
1.21547294e+00 -3.29352647e-01 8.67598653e-01 1.21735379e-01
-2.81536020e-02 7.58395195e-01 -1.05475774e-02 2.61762172e-01
9.11937296e-01 -1.15287292e+00 5.63023329e-01 4.04252142e-01
1.90923601e-01 1.31891305e-02 5.39963424e-01 -1.21123806e-01
-1.14536273e+00 -6.05394363e-01 1.27733040e+00 -4.82383728e-01
5.83497941e-01 -9.75242376e-01 -8.17724228e-01 6.14417493e-01
8.36154759e-01 -3.75325345e-02 5.19729733e-01 -1.90005019e-01
-3.10251445e-01 8.90595242e-02 -1.16171861e+00 2.63734609e-01
8.85480762e-01 -1.15743530e+00 -8.43846500e-01 3.56754571e-01
9.12929833e-01 -2.37781897e-01 -3.96477818e-01 3.11039329e-01
7.70012975e-01 -1.02088058e+00 1.15191102e+00 -5.62849402e-01
2.80091435e-01 1.69877291e-01 -1.88526809e-01 -1.11964583e+00
4.30991352e-01 -1.01445425e+00 2.17212260e-01 1.49040079e+00
3.62641007e-01 -5.71081638e-01 4.99630630e-01 4.74083610e-02
4.09570262e-02 -5.70373297e-01 -1.36010778e+00 -7.41897225e-01
2.98077106e-01 -5.52521110e-01 3.12863559e-01 3.38587523e-01
1.46284893e-01 6.65321887e-01 -5.67772925e-01 -2.36991048e-01
4.42342818e-01 -4.62818980e-01 6.68750465e-01 -1.03442585e+00
-2.46200830e-01 -5.30454159e-01 -5.32065392e-01 -1.63779950e+00
7.63190985e-01 -7.15803504e-01 6.29598320e-01 -1.25888705e+00
1.59866273e-01 3.11641455e-01 1.94606006e-01 4.07558709e-01
1.43537551e-01 2.54090607e-01 3.35491449e-01 -1.22797638e-01
-6.29179120e-01 6.92901790e-01 7.80493200e-01 -5.42104959e-01
2.41173983e-01 3.40647429e-01 -4.29206848e-01 6.82359338e-01
2.85833329e-01 4.10598181e-02 -3.70970517e-01 -2.21604005e-01
-2.56238788e-01 3.57344031e-01 5.50318003e-01 -8.76916647e-01
5.14797628e-01 5.44894576e-01 -1.29010975e-02 -8.75098646e-01
7.32709408e-01 -4.61444169e-01 -5.60495913e-01 -1.10589333e-01
-7.11595535e-01 -3.99549663e-01 3.18871170e-01 5.89765549e-01
-3.88968796e-01 -3.34913850e-01 9.92370009e-01 8.03596899e-02
-4.49283481e-01 -8.78613591e-02 -6.44690454e-01 5.09224951e-01
7.09873080e-01 -9.31017548e-02 -2.22361222e-01 -1.01257706e+00
-8.16322267e-01 1.42041340e-01 2.34816104e-01 4.74116474e-01
6.28987789e-01 -1.01122415e+00 -9.53794062e-01 4.88697857e-01
-6.14393502e-02 -1.62829280e-01 1.02942102e-01 8.93719912e-01
-8.07751343e-02 9.08219457e-01 1.90490872e-01 -9.68084335e-01
-1.73795235e+00 6.33657753e-01 5.09005666e-01 3.03800941e-01
-6.44701660e-01 1.18094826e+00 2.28310481e-01 -2.44577616e-01
9.92110968e-01 -2.36534446e-01 -1.99700534e-01 2.47767359e-01
7.45655715e-01 8.55342224e-02 1.85604408e-01 -1.15632248e+00
-3.72362375e-01 6.59912050e-01 1.61137328e-01 -3.71289045e-01
1.06023991e+00 -3.40753108e-01 1.91635787e-01 6.65592670e-01
1.54167926e+00 3.54024172e-01 -1.14634347e+00 -1.26324728e-01
-2.06044409e-02 -1.98301107e-01 1.18638277e-02 -8.18228066e-01
-6.34291112e-01 1.27782679e+00 5.79244673e-01 2.55759448e-01
1.20517612e+00 3.99849087e-01 6.90526187e-01 1.48354977e-01
2.30917311e-03 -9.74074066e-01 1.98466748e-01 5.67254186e-01
1.23415375e+00 -1.41690862e+00 -6.70667946e-01 -2.98206568e-01
-7.58727431e-01 9.49597120e-01 -1.67744458e-01 4.32493180e-01
4.47280109e-01 5.07171452e-01 1.59473762e-01 2.69925237e-01
-9.00289059e-01 -5.50443769e-01 4.68692809e-01 9.87849295e-01
3.33953023e-01 -4.20038819e-01 3.22279930e-01 3.05835515e-01
-3.48391801e-01 1.23373069e-01 3.00914228e-01 4.22134519e-01
-2.10606277e-01 -8.60580683e-01 -4.00990009e-01 -4.10394855e-02
-6.68685853e-01 -4.94165123e-01 -6.80845499e-01 6.04169726e-01
-2.17521250e-01 1.32682955e+00 1.42878398e-01 -1.92731116e-02
2.48017386e-01 7.56622016e-01 1.53526381e-01 -3.20987552e-01
-1.45783857e-01 1.91913456e-01 2.24348471e-01 -5.23451924e-01
-4.69797611e-01 -9.11105275e-01 -6.42014503e-01 -1.82471592e-02
-4.24988419e-01 3.13271768e-02 8.70100796e-01 7.67159224e-01
2.70954400e-01 2.77858049e-01 4.81458038e-01 -1.13963735e+00
-8.89089584e-01 -1.12365282e+00 -5.46177387e-01 1.38873130e-01
1.07437038e+00 -4.26443249e-01 -9.42689836e-01 2.65208125e-01] | [14.343338966369629, 5.0409369468688965] |
38108657-2b23-4cf4-9ac0-f4262cd847c9 | distribution-level-battery-storage-valuation | 2106.07590 | null | https://arxiv.org/abs/2106.07590v3 | https://arxiv.org/pdf/2106.07590v3.pdf | Decision making under uncertainty for deploying battery storage as a non-wire alternative in distribution networks | The growing demand for electricity in emerging markets and developing economies (EMDE) such as India is causing loading and congestion problems on distribution networks, particularly in urban locations, that adversely impact sustainable development and economic growth. Electric utilities in these economies face unique constraints regarding raising capital required to upgrade their congested networks. Battery storage has emerged as a non-wire alternative (NWA) to feeder upgrades. This article presents a flexible valuation framework for battery storage use in distribution networks and its application in the context of EMDE distribution network planning. We evaluate the value of storage as an NWA using a multi-stage decision making process that combines system optimization with markov-decision processes (MDP) to identify the least-cost network upgrade strategy under demand growth uncertainty. When applied to feeders in Delhi, India, the approach highlights the cost-effectiveness of battery storage to manage load growth while deferring network investments. Across the low, medium and high battery storage capital cost projections for 2030, we estimate that 18 to 29 GWh of battery storage capacity could be deployed to defer 11,752 to 15,914 km of medium voltage distribution feeder lines that are loaded at 60\% or more of their ampere capacity in 2030, resulting in 12 to 16\% capital cost savings. Interestingly, the lowering storage capital costs does not always increase NWA storage deployment, due to network capacity constraints limiting opportunities for off-peak storage charging. | ['Robert Stoner', 'Dharik S. Mallapragada', 'Marc Barbar'] | 2021-06-14 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [-8.00004721e-01 1.26796961e-01 -1.14019208e-01 6.41657487e-02
-2.02330127e-01 -8.60539436e-01 2.61412978e-01 3.35581243e-01
-8.53630248e-03 1.36548102e+00 5.45960665e-02 -1.09630787e+00
-6.42489195e-01 -1.21166492e+00 3.26383933e-02 -7.40830123e-01
-2.91090041e-01 5.19627631e-01 -2.46908247e-01 -2.53899127e-01
3.73736322e-02 1.08211470e+00 -8.68821979e-01 -5.69104731e-01
1.02499235e+00 1.26493311e+00 3.97471130e-01 -5.85748143e-02
1.46304160e-01 2.59271204e-01 -7.55294561e-01 -2.03243375e-01
2.36077845e-01 -1.51524708e-01 -6.19980395e-01 -1.09764069e-01
-1.17989600e+00 -8.19958627e-01 -2.83216864e-01 8.70756745e-01
7.95170128e-01 1.92878589e-01 6.15999103e-01 -1.61563265e+00
-5.99318385e-01 8.25488627e-01 -4.61211324e-01 6.80762589e-01
-5.72993979e-02 4.19033691e-02 8.43481898e-01 -7.50536263e-01
-1.23303764e-01 8.66659164e-01 2.65178114e-01 -7.23653436e-02
-1.06548822e+00 -4.17824566e-01 -1.83325931e-01 1.64229199e-01
-1.42021048e+00 -4.67662245e-01 5.56354702e-01 -1.76889867e-01
1.73432362e+00 3.15874219e-01 1.21725166e+00 -4.48094934e-01
3.18464398e-01 1.87258646e-01 9.80302274e-01 -4.65574443e-01
5.20565808e-01 4.84669171e-02 -3.44925046e-01 -3.33584517e-01
7.22814560e-01 -5.29572189e-01 1.11750148e-01 5.82072884e-02
1.98662609e-01 -6.87428594e-01 -5.62487505e-02 2.04103962e-01
-3.88374001e-01 5.19668818e-01 2.06280991e-01 1.50441527e-01
-4.61501747e-01 -1.86215132e-01 2.65650123e-01 1.16494440e-01
3.32276642e-01 1.71244487e-01 -3.60339880e-01 -2.47881696e-01
-1.05420363e+00 -1.17321685e-02 7.99566448e-01 1.00054717e+00
1.98222533e-01 9.21398222e-01 6.47947490e-02 4.77592260e-01
3.73914838e-01 1.02166009e+00 -1.91334058e-02 -7.74160683e-01
5.08153975e-01 2.90195290e-02 5.44875383e-01 -2.02724084e-01
-6.13531113e-01 -6.84915960e-01 -8.44070375e-01 1.19361721e-01
3.82279344e-02 -5.74364245e-01 -4.42137569e-01 1.36442149e+00
-1.80464208e-01 -4.72361535e-01 7.06404075e-02 2.61806369e-01
-3.62050772e-01 1.06657827e+00 -1.09541684e-01 -7.29684412e-01
9.72744763e-01 -4.46901560e-01 -8.44493270e-01 3.69736463e-01
4.89614040e-01 -3.54765654e-01 2.34275609e-01 3.55637401e-01
-2.08435106e+00 1.86854318e-01 -1.11058748e+00 4.48815912e-01
-5.03241360e-01 -3.06479871e-01 1.92593083e-01 7.36155868e-01
-1.35030115e+00 6.42346025e-01 -7.93682277e-01 -1.01796858e-01
6.21980727e-01 3.33770484e-01 3.24359834e-01 2.78822690e-01
-1.18727934e+00 1.76130199e+00 4.53798801e-01 7.47183740e-01
-6.73832357e-01 -7.84129918e-01 -6.11501575e-01 8.20746958e-01
5.32151833e-02 -1.82809800e-01 9.56841469e-01 1.75453559e-01
-1.19553220e+00 -1.96198344e-01 1.67150006e-01 -7.38122165e-01
3.12376678e-01 3.96442622e-01 -9.42994714e-01 3.76234315e-02
-4.08136174e-02 -2.11472325e-02 -3.51420164e-01 -6.64658844e-01
-7.58273900e-01 -6.82398155e-02 -5.45108840e-02 2.46637791e-01
-4.58917916e-01 1.02811851e-01 6.04652762e-01 -5.45298874e-01
-2.80926794e-01 -8.14655840e-01 -2.96109825e-01 -5.22135735e-01
-4.71060127e-01 -4.18803036e-01 9.91138220e-01 -9.62862849e-01
1.34975421e+00 -1.65916228e+00 -6.07990101e-02 7.67787218e-01
-6.10230982e-01 -1.14704669e-02 4.31842983e-01 8.17237318e-01
-9.97214615e-02 4.62602794e-01 -8.57160389e-02 -1.49318278e-01
4.94514793e-01 5.66032171e-01 -2.59238809e-01 2.60924876e-01
1.26785859e-01 7.17070282e-01 -7.79139161e-01 3.09317350e-01
3.53518456e-01 1.46180307e-02 1.49162635e-01 -2.65993208e-01
8.69139805e-02 8.96663368e-02 -3.03004216e-02 9.87470329e-01
1.14586878e+00 -8.52652192e-02 3.94712776e-01 3.70370597e-03
-5.19537926e-01 2.74645716e-01 -1.04187596e+00 1.19135749e+00
-6.29519343e-01 6.41973853e-01 1.87184408e-01 -1.16407347e+00
7.02976644e-01 5.69165125e-02 5.00653327e-01 -1.14916933e+00
-3.44738960e-01 6.26641393e-01 3.46630156e-01 -2.52248913e-01
4.96573985e-01 -2.06989318e-01 1.06828935e-01 6.07056856e-01
-1.83960691e-01 -3.04198623e-01 5.75960755e-01 1.23858348e-01
6.28140926e-01 -4.61324662e-01 -1.43007964e-01 -1.22943592e+00
3.74850780e-01 -4.49989527e-01 8.50041509e-01 -1.68607131e-01
1.28399998e-01 -4.94574249e-01 7.82210052e-01 3.07284027e-01
-1.31218565e+00 -1.36502206e+00 -6.40778363e-01 1.08975559e-01
-2.46782377e-02 2.74290413e-01 -1.50857121e-01 1.26360327e-01
3.85385484e-01 1.57809722e+00 1.22142844e-01 1.51125655e-01
-4.80554134e-01 -1.17327201e+00 1.36180148e-01 6.91129267e-01
5.83826363e-01 -3.83430332e-01 -6.58341706e-01 4.63615835e-01
2.94520169e-01 -7.19337583e-01 5.74348792e-02 4.76733267e-01
-4.40858513e-01 -5.88787973e-01 -1.10390532e+00 -2.01492548e-01
7.69418657e-01 -1.62190542e-01 9.97576237e-01 5.19542284e-02
4.85892892e-02 1.71828419e-01 2.10064411e-01 -5.73889136e-01
-1.36528509e-02 2.83633620e-02 2.88411766e-01 -8.95755112e-01
-1.06235415e-01 -6.46235108e-01 -7.55983710e-01 3.10740978e-01
-3.39931220e-01 -8.81840661e-02 2.35717744e-03 4.48631018e-01
4.28961724e-01 1.03672075e+00 1.82484460e+00 -1.18512295e-01
7.91512251e-01 -1.10448325e+00 -1.07290351e+00 5.01716673e-01
-1.19642866e+00 -4.89107847e-01 4.30066884e-01 4.00521129e-01
-1.32957053e+00 -6.49879396e-01 -3.15619893e-02 3.19156706e-01
6.74237549e-01 1.17188060e+00 -5.23179412e-01 9.52676088e-02
-5.32674611e-01 -4.78864228e-03 2.19135866e-01 -4.52868521e-01
-5.19619659e-02 5.28182149e-01 3.75535011e-01 -5.39596319e-01
1.01405025e+00 2.04896316e-01 5.95149338e-01 -4.29904610e-01
-1.84731632e-01 2.41500407e-01 -4.78754431e-01 -3.36464643e-01
3.27574998e-01 -1.20131719e+00 -9.29602087e-01 5.32782257e-01
-3.95846426e-01 -5.30882955e-01 -5.41474640e-01 3.02416533e-01
-2.71544665e-01 1.74352765e-01 -3.75666052e-01 -1.22535980e+00
-3.21135581e-01 -9.76248860e-01 -3.48602116e-01 4.90256578e-01
2.03851461e-01 -1.12078738e+00 -5.47118664e-01 -1.17391557e-01
1.09388840e+00 5.30806243e-01 1.35098767e+00 -2.06539854e-01
-7.11801589e-01 2.19503984e-01 -1.38124540e-01 8.48260760e-01
4.53411579e-01 -5.09329187e-03 -1.69722185e-01 -9.49138880e-01
-3.72881234e-01 1.07982427e-01 -4.51160632e-02 4.46372718e-01
1.08528185e+00 -3.43557358e-01 -1.25387415e-01 6.70350576e-03
2.04525661e+00 9.46336746e-01 8.98368716e-01 4.57924217e-01
-2.53403753e-01 3.98638576e-01 6.84805453e-01 1.06638205e+00
6.82855487e-01 3.11367363e-01 6.21512175e-01 -3.52773443e-03
3.64004403e-01 1.01762585e-01 7.45736510e-02 7.27554440e-01
1.47998463e-02 -7.13715851e-01 -1.04512739e+00 9.64999259e-01
-1.26308775e+00 -7.50469446e-01 -8.93305987e-02 2.40998387e+00
4.88081574e-01 5.54577351e-01 3.26469898e-01 4.41124916e-01
4.27092642e-01 -5.12339056e-01 -7.49924242e-01 -9.44239378e-01
-7.66368091e-01 1.83211029e-01 9.92059767e-01 1.47925749e-01
1.57296225e-01 -2.64318347e-01 5.78180647e+00 7.93809414e-01
-6.90930009e-01 6.59337416e-02 1.08322275e+00 -4.97379363e-01
-8.97903323e-01 1.27170697e-01 -6.88647926e-01 1.01145434e+00
1.71528494e+00 -1.27761972e+00 4.12058204e-01 2.18316570e-01
1.10530949e+00 -7.77179062e-01 -5.36656857e-01 2.51224041e-01
-5.44855118e-01 -1.68051791e+00 -2.08470374e-01 5.93108535e-01
9.86751199e-01 -3.97118218e-02 -7.18145892e-02 4.47848067e-02
2.19133407e-01 -8.12388599e-01 1.05763292e+00 6.98710501e-01
8.03104460e-01 -1.77325940e+00 8.92326653e-01 1.82306081e-01
-1.36714065e+00 -7.81185746e-01 -1.72257856e-01 -1.89033940e-01
1.42014384e+00 9.18535829e-01 -4.38962430e-01 6.17855072e-01
1.03817201e+00 1.47324130e-01 -1.64071351e-01 1.06447303e+00
-3.38469818e-02 4.48279917e-01 -6.90338731e-01 1.96150020e-01
2.79157043e-01 -6.00186706e-01 1.03020683e-01 4.96541947e-01
9.06322300e-01 4.38821197e-01 -2.73686469e-01 7.15230763e-01
-1.36021808e-01 -4.64864075e-01 -1.83538720e-01 3.33021767e-02
1.26429951e+00 9.96411264e-01 -8.74133646e-01 -2.19036922e-01
-3.11449140e-01 1.83656886e-01 -3.50776821e-01 3.55288595e-01
-7.28573740e-01 -7.47183204e-01 4.91845548e-01 3.42337519e-01
-2.12853681e-02 -3.54692757e-01 -5.98923743e-01 -6.04096413e-01
8.24966580e-02 1.80625901e-01 1.79849491e-01 -1.08067763e+00
-1.12769675e+00 4.09115255e-02 4.85321999e-01 -8.55925620e-01
-3.53318483e-01 -1.13374315e-01 -1.10817575e+00 1.67405260e+00
-2.08214378e+00 -9.70874310e-01 3.54324132e-01 1.70692995e-01
3.75212580e-01 -2.79762059e-01 4.35641259e-01 7.02479064e-01
-1.01966119e+00 4.67020124e-01 1.02451360e+00 -2.96527565e-01
-6.02583468e-01 -1.34565008e+00 2.45308653e-02 1.10488367e+00
-1.07092333e+00 7.83719495e-02 3.65079790e-01 -6.80282295e-01
-1.29512465e+00 -6.54280663e-01 7.96810567e-01 4.36404854e-01
9.20711875e-01 -1.32937700e-01 -7.75963485e-01 4.65583593e-01
9.31308389e-01 -5.66642165e-01 7.88122654e-01 -4.49897885e-01
8.87682557e-01 -2.11189270e-01 -1.74450529e+00 2.17765316e-01
4.90229040e-01 -2.73049176e-01 2.00360030e-01 2.88544118e-01
-1.45192184e-02 -4.27666157e-02 -1.62327361e+00 2.97536552e-01
3.37198734e-01 -2.71816105e-01 6.82237566e-01 -7.13295117e-02
-3.37929904e-01 -2.06036061e-01 -2.58585364e-01 -1.83615661e+00
-3.23569417e-01 -9.95639741e-01 -3.57035428e-01 1.87687492e+00
6.45855963e-01 -1.12607956e+00 4.36125875e-01 1.23837233e+00
-5.02585113e-01 -1.12875140e+00 -1.72409558e+00 -1.18398058e+00
8.09522629e-01 7.80394375e-02 1.34985995e+00 6.30667567e-01
4.10339117e-01 -2.69203871e-01 1.50287971e-01 3.51658523e-01
9.24179316e-01 -2.51876920e-01 -2.95026809e-01 -9.57622766e-01
2.97270894e-01 -4.88982737e-01 9.44047347e-02 -2.77410686e-01
2.07322344e-01 -7.37691283e-01 -6.44948840e-01 -2.20790124e+00
-1.19992010e-01 -9.54536021e-01 -4.75380480e-01 4.90077615e-01
9.13373530e-01 -1.73876330e-01 5.10239303e-01 1.60012603e-01
5.17343521e-01 8.00809860e-01 6.56950653e-01 -1.06921911e-01
2.20560834e-01 2.49504745e-01 -5.81489146e-01 4.46053743e-01
1.01753449e+00 1.93303585e-01 -8.14794898e-01 -3.92517000e-01
7.14745641e-01 6.01771712e-01 -1.90159362e-02 -7.82198191e-01
2.45927989e-01 -3.54254425e-01 4.34172481e-01 -1.46500540e+00
1.42438993e-01 -1.07599211e+00 1.02279198e+00 7.32650459e-01
2.87255794e-01 5.69297671e-01 4.49511915e-01 1.74128890e-01
1.79644510e-01 -6.25464857e-01 7.39053011e-01 2.56843895e-01
-4.81702924e-01 -9.91350133e-03 -8.81503403e-01 -4.07003999e-01
1.59955168e+00 -4.16794747e-01 -7.34428287e-01 -2.40694016e-01
-1.16267455e+00 1.27813160e+00 3.04349810e-01 1.12897009e-01
3.14124227e-01 -1.22079468e+00 -5.34833074e-01 -2.93315232e-01
-8.69285822e-01 -8.57924297e-02 4.07769561e-01 5.88304996e-01
-5.76877773e-01 6.16400182e-01 -4.65161741e-01 1.70349609e-02
-4.06479985e-01 9.48025882e-02 5.80503583e-01 -1.09017186e-01
-3.00469011e-01 3.90722811e-01 -8.74439120e-01 6.81427956e-01
-2.93108672e-01 -1.86385944e-01 1.66007787e-01 6.85980499e-01
2.17167631e-01 1.27596486e+00 4.87725079e-01 -2.44339362e-01
-2.81226963e-01 -1.74457282e-01 2.74921685e-01 -2.27581441e-01
1.57978106e+00 -9.31127131e-01 -3.80778939e-01 1.80356458e-01
8.25122654e-01 -3.33310783e-01 -1.07075000e+00 3.26228738e-01
2.71729305e-02 -2.80088335e-01 3.89428556e-01 -1.23172939e+00
-1.42160487e+00 4.49037045e-01 3.74948055e-01 7.55105436e-01
1.51007032e+00 -3.67468625e-01 6.55090868e-01 1.52550265e-01
5.71277142e-01 -1.50882423e+00 -6.60103440e-01 6.75757155e-02
9.22917724e-01 -3.94424319e-01 3.98306191e-01 7.37820342e-02
-6.61948398e-02 1.14076018e+00 3.47873151e-01 2.83435911e-01
1.27537310e+00 5.96743762e-01 -1.00031996e+00 1.50141940e-01
-6.24789417e-01 7.06824958e-02 -7.57207692e-01 3.07090223e-01
-3.03560853e-01 3.04948807e-01 -4.99517143e-01 3.76034409e-01
5.52301221e-02 -1.10587150e-01 1.31165206e+00 1.23205864e+00
-2.16836885e-01 -7.73505032e-01 1.92049537e-02 1.10323966e+00
-4.38453823e-01 -9.67606679e-02 9.43540394e-01 6.57527387e-01
9.00192261e-02 9.95365083e-01 7.78923690e-01 6.15470588e-01
6.42284155e-01 -2.29140714e-01 9.78942215e-02 -1.95522144e-01
-1.31578729e-01 1.82385504e-01 4.13495570e-01 4.33516294e-01
-2.22486362e-01 -1.16146433e+00 -1.70725465e+00 -9.44183588e-01
-8.76086473e-01 6.08672202e-01 1.00803363e+00 8.59712899e-01
2.78387517e-01 3.41581911e-01 1.24203467e+00 -8.37375700e-01
-6.95404291e-01 -7.28461027e-01 -1.33540642e+00 -8.02356958e-01
-2.66526073e-01 -7.16961265e-01 -5.53429365e-01 -5.94939649e-01] | [5.689558506011963, 2.4507248401641846] |
b770634c-6ae9-4f10-a67e-fd49b4db6c72 | ellipsis-translation-for-a-medical-speech-to | null | null | https://aclanthology.org/2020.eamt-1.30 | https://aclanthology.org/2020.eamt-1.30.pdf | Ellipsis Translation for a Medical Speech to Speech Translation System | In diagnostic interviews, elliptical utterances allow doctors to question patients in a more efficient and economical way. However, literal translation of such incomplete utterances is rarely possible without affecting communication. Previous studies have focused on automatic ellipsis detection and resolution, but only few specifically address the problem of automatic translation of ellipsis. In this work, we evaluate four different approaches to translate ellipsis in medical dialogues in the context of the speech to speech translation system BabelDr. We also investigate the impact of training data, using an under-sampling method and data with elliptical utterances in context. Results show that the best model is able to translate 88% of elliptical utterances. | ['Hervé Spechbach', 'Pierrette Bouillon', 'Johanna Gerlach', 'Jonathan Mutal'] | null | null | null | null | eamt-2020-11 | ['speech-to-speech-translation'] | ['speech'] | [ 3.39030594e-01 8.63034427e-01 1.19989529e-01 -5.05345106e-01
-1.11570001e+00 -3.37341189e-01 3.03326577e-01 5.17641962e-01
-5.40493846e-01 9.86985266e-01 6.99461460e-01 -6.40278339e-01
1.15008950e-01 -3.55599731e-01 4.04977985e-03 -2.40956411e-01
4.88131195e-01 1.15133703e+00 -5.81806339e-03 -5.17897785e-01
-1.62077129e-01 2.71817923e-01 -5.15873134e-01 9.65207338e-01
8.26469004e-01 1.09716557e-01 -1.04446607e-02 8.30459177e-01
-4.84637976e-01 9.27153587e-01 -1.35120571e+00 -7.23768055e-01
-1.66147798e-01 -7.64899194e-01 -1.28679025e+00 2.81140685e-01
-9.33497101e-02 -4.47170198e-01 1.31623060e-01 7.07481802e-01
8.62547696e-01 -5.79412997e-01 5.06588459e-01 -4.03361976e-01
-1.38856709e-01 1.20057476e+00 1.16704583e-01 4.78565484e-01
1.00411499e+00 -3.01477611e-01 5.78606009e-01 -4.26794738e-01
1.04613197e+00 1.36375451e+00 6.09699368e-01 9.22054291e-01
-1.41048813e+00 -1.99702606e-01 -5.88868320e-01 -2.31362775e-01
-1.26460934e+00 -8.52337122e-01 4.82843965e-01 -2.31394961e-01
1.23668694e+00 6.90066516e-01 5.59234858e-01 1.15163493e+00
2.37739265e-01 6.70673728e-01 1.12005019e+00 -7.34167337e-01
-4.05963161e-04 5.29464006e-01 8.07154551e-02 1.68002963e-01
-2.88281254e-02 -4.11474705e-01 -3.16008598e-01 -5.47832608e-01
3.69362324e-01 -6.22636974e-01 -2.55401641e-01 4.99388248e-01
-1.33399022e+00 9.55165327e-01 -1.88649431e-01 8.00965846e-01
-4.17949408e-01 -5.57062387e-01 7.03610837e-01 5.66930115e-01
4.79934901e-01 9.15150106e-01 -3.46559435e-01 -3.23525399e-01
-8.55056942e-01 3.60023171e-01 1.35785210e+00 9.00845706e-01
-6.69242814e-02 -9.68699381e-02 -2.65875995e-01 1.01201820e+00
-2.79709212e-02 6.13661826e-01 6.69740200e-01 -9.90851641e-01
7.09106028e-01 4.28254098e-01 2.61802133e-02 -8.44560325e-01
-6.68114483e-01 -5.77012040e-02 -3.17878157e-01 -5.63547790e-01
6.23626530e-01 -5.28498232e-01 -2.20635787e-01 1.19260013e+00
4.20297503e-01 -7.98240483e-01 8.36267292e-01 7.94293940e-01
1.15730357e+00 4.67309684e-01 -5.30770011e-02 -7.88623452e-01
1.62832832e+00 -6.72331214e-01 -1.45409679e+00 -1.25212401e-01
1.08172691e+00 -1.53974652e+00 7.38733590e-01 2.13116631e-01
-1.57162774e+00 1.45363539e-01 -4.45651948e-01 -2.10054427e-01
1.55198395e-01 1.30952045e-01 2.05977991e-01 6.24792635e-01
-8.31516027e-01 3.06357890e-01 -7.98684537e-01 -4.37516361e-01
-2.74106562e-01 4.76168841e-01 -3.77468169e-01 2.78435975e-01
-1.25985229e+00 1.22913790e+00 6.71942672e-03 -1.83530152e-02
1.90978765e-01 -2.52725422e-01 -8.40586245e-01 -3.03496897e-01
2.55177826e-01 -6.37254000e-01 1.84279048e+00 -9.41230297e-01
-1.61007226e+00 1.02901137e+00 -3.06173801e-01 -6.41982317e-01
6.99502766e-01 9.35894158e-03 -3.83797526e-01 5.70818782e-01
2.02790394e-01 5.59982002e-01 3.19624513e-01 -7.35551119e-01
-2.43894324e-01 -3.13133389e-01 -1.15170047e-01 4.99720365e-01
1.36471987e-01 7.10344672e-01 6.36171848e-02 -6.66514516e-01
4.40848880e-02 -8.11419427e-01 -7.74368048e-02 -5.86906195e-01
-4.98968065e-01 -2.90444225e-01 3.10648650e-01 -1.10469961e+00
1.42038846e+00 -1.98991096e+00 2.92284757e-01 -2.68725187e-01
8.56067315e-02 3.73231679e-01 2.16988772e-01 8.53870690e-01
8.97311326e-03 1.88899413e-01 -3.90857518e-01 -5.08408427e-01
-2.04952523e-01 6.85408711e-01 -3.44526142e-01 1.23722218e-01
2.29298413e-01 8.45392823e-01 -6.05826855e-01 -1.04832685e+00
6.00586534e-02 3.65843892e-01 -4.70795631e-01 2.63585746e-01
-8.19659233e-02 5.05398273e-01 -4.10506159e-01 6.77295923e-01
1.74946606e-01 5.29252328e-02 6.01901412e-01 3.09821457e-01
-1.06669128e-01 9.40768480e-01 -6.75562322e-01 1.23051822e+00
-3.25996757e-01 5.56074023e-01 3.42466086e-01 -6.84017658e-01
6.49647832e-01 9.50828671e-01 4.04906541e-01 -4.36605364e-01
3.47812295e-01 7.34935641e-01 4.65478390e-01 -8.60526800e-01
5.13042629e-01 -4.80148554e-01 -1.80736631e-01 3.16884190e-01
-3.11914384e-01 -4.03405845e-01 2.13369206e-01 2.52298862e-01
9.84934270e-01 -4.99375194e-01 7.33503520e-01 -6.71098754e-02
4.78520483e-01 4.36516613e-01 2.93867946e-01 4.27111089e-01
-5.60050793e-02 6.79826677e-01 7.18054950e-01 -3.08768362e-01
-9.27394390e-01 -4.77799416e-01 -4.99711186e-01 5.82117140e-01
-4.15727675e-01 -5.16123593e-01 -1.42288876e+00 -6.76870525e-01
-3.85394871e-01 1.00580764e+00 -2.25713074e-01 3.38723958e-01
-1.12900424e+00 -7.11764455e-01 9.80845571e-01 -1.15728742e-02
-1.51328295e-01 -1.34120846e+00 -8.74009788e-01 5.80578506e-01
-8.10264766e-01 -1.45054078e+00 -3.36879015e-01 5.11759333e-03
-8.73179436e-01 -1.00098395e+00 -7.86107481e-01 -8.23997080e-01
6.15249157e-01 -2.90137321e-01 1.07854402e+00 2.52782136e-01
-2.87908435e-01 2.79252946e-01 -5.44934452e-01 -3.20874155e-01
-1.39636672e+00 2.90068775e-01 -1.74758360e-01 -4.66761172e-01
4.88767058e-01 -1.67002842e-01 -3.06899250e-02 1.71423018e-01
-8.76207411e-01 3.72430198e-02 5.49587488e-01 1.02161849e+00
1.28600851e-01 -8.78481865e-01 5.77134967e-01 -1.41490865e+00
1.20325100e+00 -2.47284889e-01 1.75731126e-02 1.73718229e-01
-3.66388232e-01 6.05494762e-03 5.70946515e-01 -4.35867906e-01
-8.69125485e-01 -6.33534491e-02 -8.39880407e-01 4.75262143e-02
-4.31214720e-01 3.89755279e-01 2.20618248e-01 1.61113650e-01
9.55940604e-01 -6.05698004e-02 3.55340838e-01 -4.93001312e-01
-7.90402368e-02 1.11042738e+00 1.53983295e-01 -2.60403156e-01
2.55654920e-02 -2.12913454e-02 -4.46319282e-01 -1.26205254e+00
-4.85074371e-01 -4.71956015e-01 -4.14895445e-01 1.24015607e-01
9.24401522e-01 -4.68966097e-01 -3.49563450e-01 2.02699043e-02
-1.62486660e+00 -1.32466257e-01 -5.01147509e-01 6.56567574e-01
-5.66068351e-01 7.21347094e-01 -9.43532288e-01 -7.32122898e-01
-3.83470565e-01 -1.26825523e+00 1.38565636e+00 -3.73566777e-01
-1.05150020e+00 -8.97058427e-01 8.57755616e-02 7.24298298e-01
3.34756464e-01 5.71586229e-02 9.25100565e-01 -1.20018756e+00
1.32749885e-01 -1.13017291e-01 1.96027622e-01 -1.26047030e-01
3.02699536e-01 -2.78795540e-01 -5.56969702e-01 -5.16791604e-02
4.93151009e-01 -3.77428412e-01 2.08300129e-01 1.93318620e-01
2.00735122e-01 -8.99848223e-01 -2.31954455e-01 2.20614020e-02
6.67711079e-01 2.29667172e-01 4.14970756e-01 1.07097648e-01
1.03957914e-01 1.17713904e+00 6.17166162e-01 1.94375858e-01
3.97758633e-01 7.79046178e-01 -1.60491481e-01 7.55007938e-02
-7.45292604e-02 2.58596867e-01 1.30097270e-01 1.32778561e+00
4.11997348e-01 -2.62115747e-01 -1.20081902e+00 4.80201811e-01
-1.55185735e+00 -5.79157770e-01 -2.58401304e-01 1.60323572e+00
1.33490276e+00 -1.58121914e-01 7.42706209e-02 2.26677224e-01
5.72690010e-01 -2.30249554e-01 1.67871773e-01 -9.65546787e-01
-7.84760993e-03 -2.96678264e-02 2.20505685e-01 1.11322308e+00
-7.62670875e-01 9.38637853e-01 7.26025581e+00 3.46225649e-01
-1.04026937e+00 2.85538197e-01 4.79795605e-01 2.95870658e-02
-3.93651098e-01 -2.78515726e-01 -7.31622636e-01 3.33917856e-01
1.21915877e+00 3.74200125e-03 2.45338693e-01 4.47326362e-01
2.45039225e-01 -1.64313838e-01 -1.05874240e+00 7.89938807e-01
2.31435433e-01 -1.19107056e+00 -1.11741900e-01 2.47058123e-02
1.92750618e-01 -2.99047470e-01 -3.73298943e-01 1.49094716e-01
-2.63410449e-01 -8.58146787e-01 3.97969484e-01 2.15779454e-01
4.95289654e-01 -4.21903938e-01 1.24586296e+00 7.38831043e-01
-3.30452055e-01 1.63620204e-01 -3.01701695e-01 -5.13293594e-02
5.63817203e-01 2.14434445e-01 -1.82660699e+00 4.86888409e-01
1.99640766e-01 -3.16713378e-02 -2.54111230e-01 5.47225356e-01
-5.84163070e-02 6.99363112e-01 -4.48395938e-01 -3.65120023e-01
1.05930030e-01 -9.60313231e-02 1.02098310e+00 1.58587015e+00
1.08451754e-01 5.74492872e-01 5.03017195e-02 2.60931522e-01
1.26411736e-01 7.22038925e-01 -6.83199883e-01 -8.06464404e-02
2.85940051e-01 8.80183518e-01 -6.51036561e-01 -4.39126343e-01
-1.04633965e-01 9.34622049e-01 2.23093480e-01 -7.37577900e-02
-2.45481119e-01 -5.61200492e-02 7.84054548e-02 3.35827738e-01
-2.23428011e-01 1.38870463e-01 -3.46903354e-01 -9.26827192e-01
1.24884836e-01 -1.62960434e+00 4.83375251e-01 -4.50679481e-01
-9.26710606e-01 1.12033570e+00 1.57619387e-01 -7.70512879e-01
-1.12972426e+00 -2.99989998e-01 -8.64578336e-02 7.26342738e-01
-1.01490951e+00 -8.13090563e-01 4.29959357e-01 2.21500278e-01
8.54597688e-01 -1.06449775e-01 1.36828339e+00 3.65081221e-01
-2.21313253e-01 5.46798348e-01 -1.39076471e-01 2.32398286e-01
9.05334592e-01 -1.26246333e+00 2.10978948e-02 3.59038055e-01
-2.69099995e-02 7.94178843e-01 1.07469368e+00 -7.41733313e-01
-9.78463233e-01 -5.86207330e-01 1.85598266e+00 -5.04496336e-01
4.07499403e-01 -1.16483659e-01 -1.09582436e+00 4.48469251e-01
6.44891202e-01 -6.02488756e-01 8.98848355e-01 -1.15139164e-01
3.13879222e-01 2.50950456e-01 -1.32881069e+00 5.47000647e-01
4.21235263e-01 -3.65775704e-01 -1.24300861e+00 8.04728091e-01
9.96536255e-01 -9.79917347e-01 -1.05032957e+00 2.64702946e-01
1.76574066e-01 -5.29218793e-01 7.03382194e-01 -7.11736739e-01
4.41775769e-01 1.91423997e-01 -9.31882411e-02 -1.03352427e+00
3.75243604e-01 -8.77826631e-01 2.42244601e-01 9.66395140e-01
7.73125112e-01 -7.70958304e-01 5.09455442e-01 7.75236368e-01
-2.71464884e-01 -8.09386253e-01 -1.29564977e+00 -1.86607495e-01
2.05681950e-01 8.99891928e-03 3.75850916e-01 9.33497787e-01
5.91963470e-01 6.89751685e-01 -2.84263045e-01 -1.83549255e-01
4.21783701e-02 -4.05319110e-02 4.91451681e-01 -7.41173923e-01
-3.89560521e-01 -1.92471504e-01 -1.66332245e-01 -6.29516006e-01
3.43234912e-02 -6.29547119e-01 7.53597617e-02 -1.56077838e+00
-1.37688607e-01 -2.54654348e-01 8.60765994e-01 3.18470418e-01
-4.28023823e-02 1.29358441e-01 1.81904033e-01 3.27846020e-01
-2.78715938e-01 7.16809183e-02 1.07114398e+00 5.10959923e-02
-3.51453096e-01 2.14450851e-01 -5.24082422e-01 7.61739969e-01
1.08095181e+00 -7.91032434e-01 -6.99811578e-02 -5.40487707e-01
1.42476529e-01 7.98854113e-01 -1.41347423e-01 -3.79063159e-01
2.43058458e-01 6.62817433e-02 -1.28626347e-01 -5.05296171e-01
2.83813238e-01 -6.12306654e-01 1.71898484e-01 7.01522589e-01
-6.54528856e-01 4.90501970e-01 1.24669276e-01 6.94599748e-02
-4.02781010e-01 -9.01664972e-01 5.27678311e-01 -3.09789240e-01
3.40287238e-01 -5.19639611e-01 -1.05441964e+00 2.93600649e-01
7.91474521e-01 -1.61471032e-02 -9.34182554e-02 -4.34124619e-01
-1.05045962e+00 -2.34361161e-02 4.11552578e-01 1.40849844e-01
5.52116275e-01 -8.13178778e-01 -9.61891353e-01 -1.35375457e-02
-2.61949837e-01 5.48952334e-02 -8.10281336e-02 1.10818899e+00
-1.01922286e+00 7.02968240e-01 1.91360503e-01 -6.21062994e-01
-1.78184593e+00 3.33314955e-01 3.86352360e-01 -4.91927683e-01
-7.70217061e-01 4.19228733e-01 -2.36302227e-01 -6.37192249e-01
2.27404803e-01 -4.06519800e-01 -6.01555049e-01 1.60499781e-01
5.06255388e-01 1.93230629e-01 3.81397635e-01 -8.99789691e-01
-3.05430055e-01 -2.23175320e-03 -1.33158371e-01 -4.86988753e-01
8.58355641e-01 -2.01382145e-01 -4.10455406e-01 4.03561831e-01
7.88128138e-01 4.42161143e-01 -4.40714955e-02 -1.31052256e-01
1.05537795e-01 -1.75065875e-01 -5.67967117e-01 -8.56721103e-01
-1.18054517e-01 7.62468696e-01 -1.20723424e-02 6.25908732e-01
8.45588565e-01 1.38313934e-01 8.84326756e-01 6.49062991e-01
8.67264420e-02 -1.09018540e+00 -2.27526650e-01 4.57105339e-01
1.04609382e+00 -1.06483006e+00 -1.87192604e-01 -7.39242196e-01
-9.73262787e-01 1.15959668e+00 1.16159879e-01 3.71846467e-01
1.82751596e-01 4.94989514e-01 7.12446809e-01 -2.65478820e-01
-7.21100807e-01 2.88929254e-01 -2.22706497e-02 4.43375945e-01
8.22004676e-01 1.85193464e-01 -1.01956403e+00 5.60373366e-01
-6.62946403e-01 -3.40186715e-01 8.31735492e-01 9.11268175e-01
-1.63527459e-01 -1.42955565e+00 -7.07836270e-01 2.19364196e-01
-1.11526000e+00 -2.83773154e-01 -1.03937924e+00 8.42033625e-01
-3.01959753e-01 1.11112487e+00 -8.65071937e-02 2.62743801e-01
5.29781222e-01 2.75383472e-01 3.46615821e-01 -9.38187659e-01
-1.22555268e+00 5.48108459e-01 9.70985413e-01 -1.99665532e-01
-5.07347047e-01 -1.06683254e+00 -1.35421586e+00 -2.44690582e-01
-3.76005590e-01 7.41941273e-01 5.25410414e-01 1.01631379e+00
1.45498753e-01 5.83340943e-01 2.44097427e-01 -1.99315175e-01
-9.35883582e-01 -1.22727251e+00 -5.08066975e-02 2.23490268e-01
4.04062003e-01 3.01039610e-02 -5.78414947e-02 -1.15682483e-02] | [11.280023574829102, 9.425070762634277] |
f520c38c-2ed0-4f1b-b9e1-da7b358f6409 | treasure-what-you-have-exploiting-similarity | 2305.06492 | null | https://arxiv.org/abs/2305.06492v1 | https://arxiv.org/pdf/2305.06492v1.pdf | Treasure What You Have: Exploiting Similarity in Deep Neural Networks for Efficient Video Processing | Deep learning has enabled various Internet of Things (IoT) applications. Still, designing models with high accuracy and computational efficiency remains a significant challenge, especially in real-time video processing applications. Such applications exhibit high inter- and intra-frame redundancy, allowing further improvement. This paper proposes a similarity-aware training methodology that exploits data redundancy in video frames for efficient processing. Our approach introduces a per-layer regularization that enhances computation reuse by increasing the similarity of weights during training. We validate our methodology on two critical real-time applications, lane detection and scene parsing. We observe an average compression ratio of approximately 50% and a speedup of \sim 1.5x for different models while maintaining the same accuracy. | ['Smail Niar', 'Ozcan Ozturk', 'Hamza Ouarnoughi', 'Halima Bouzidi', 'Hadjer Benmeziane'] | 2023-05-10 | null | null | null | null | ['scene-parsing', 'lane-detection'] | ['computer-vision', 'computer-vision'] | [ 5.60922027e-02 -3.86997849e-01 -3.67511034e-01 -7.60046601e-01
-3.34329635e-01 -1.04377829e-01 1.15505405e-01 3.21589977e-01
-7.10051239e-01 4.03749377e-01 -7.62827173e-02 -4.23404574e-01
-2.68051736e-02 -8.22209597e-01 -7.04044819e-01 -5.16862214e-01
-2.34062701e-01 -7.11822212e-02 6.20134056e-01 2.35578120e-01
2.49531999e-01 7.45575070e-01 -1.66395068e+00 6.75768256e-01
6.09720826e-01 1.37972260e+00 5.09500861e-01 7.74251640e-01
-2.63752431e-01 9.42739189e-01 -3.15412372e-01 -5.53618610e-01
5.24452269e-01 2.24122107e-01 -5.16685069e-01 1.99912459e-01
4.28134054e-01 -5.67451656e-01 -5.35507321e-01 8.47937286e-01
4.75274205e-01 -1.16222665e-01 -2.23421026e-02 -1.17984188e+00
6.89312965e-02 5.56494534e-01 -6.68803632e-01 4.10091609e-01
-1.66307449e-01 1.09381840e-01 7.17166364e-01 -5.00272751e-01
1.52148113e-01 1.03712344e+00 6.47755682e-01 2.39067614e-01
-7.28533268e-01 -6.91693842e-01 -3.50784361e-02 5.79832196e-01
-1.29468286e+00 -5.92262030e-01 6.82823360e-01 -7.56562501e-02
1.12467623e+00 1.23039177e-02 3.58460873e-01 4.94947910e-01
4.87022877e-01 7.00264454e-01 5.99133253e-01 -2.26623848e-01
2.62972891e-01 -1.46996289e-01 1.53271154e-01 6.68617606e-01
4.21884656e-01 -1.63345292e-01 -4.94210839e-01 4.25562084e-01
8.38800907e-01 2.57146060e-01 1.96484730e-01 -7.70842806e-02
-1.02203572e+00 6.88469112e-01 4.87582028e-01 2.63710648e-01
-5.02500534e-01 5.11853933e-01 8.07860196e-01 1.69552982e-01
3.70342694e-02 -1.47237152e-01 -6.62328005e-01 -2.53950119e-01
-1.04258275e+00 -2.10783318e-01 5.35152972e-01 1.04330134e+00
7.08448052e-01 4.34509635e-01 1.58668175e-01 7.69748807e-01
2.34147936e-01 4.05708045e-01 4.72072423e-01 -1.10122216e+00
7.10024476e-01 5.00800192e-01 -2.78602213e-01 -1.05219471e+00
-6.09777093e-01 -5.51007330e-01 -1.09350133e+00 7.44757950e-02
6.44690618e-02 -9.92355049e-02 -6.31417334e-01 1.35001230e+00
1.89132378e-01 4.44793552e-01 1.33615108e-02 6.65889740e-01
5.71642160e-01 7.85065770e-01 3.46507102e-01 -2.11054951e-01
1.43771207e+00 -1.05629623e+00 -5.85693359e-01 -2.88038254e-01
8.07220519e-01 -9.36272025e-01 8.02157760e-01 4.15886670e-01
-1.15785170e+00 -9.03780639e-01 -1.17367470e+00 -1.21860601e-01
-1.37592465e-01 2.49270692e-01 8.07811201e-01 6.41927361e-01
-8.18077266e-01 6.35462105e-01 -8.02229106e-01 -1.63115248e-01
5.87188601e-01 7.60875583e-01 -1.53148964e-01 -8.86389241e-02
-6.26604915e-01 4.03112620e-01 5.60559988e-01 -2.78504819e-01
-3.12465578e-01 -7.06110716e-01 -7.61869133e-01 2.49509245e-01
3.39504480e-01 -4.72150683e-01 1.09787309e+00 -8.88561785e-01
-1.34863913e+00 4.63564187e-01 -2.23446906e-01 -9.58886087e-01
1.64939806e-01 -4.49390322e-01 -6.65328205e-01 1.10114783e-01
-1.64731920e-01 8.38546097e-01 9.17380393e-01 -9.14990425e-01
-9.38333809e-01 -2.16705322e-01 1.65826768e-01 -1.62673548e-01
-6.43195033e-01 9.68943983e-02 -6.46294415e-01 -6.36321187e-01
1.33097574e-01 -8.45239401e-01 -3.54053676e-01 2.11141273e-01
5.26606478e-02 2.07980862e-03 1.06307185e+00 -4.96958286e-01
1.22783720e+00 -2.09224510e+00 -4.94141221e-01 9.74263810e-03
1.45931095e-01 5.69659770e-01 -9.85027328e-02 -1.02652714e-01
2.16492161e-01 -1.01447336e-01 1.81442983e-02 -2.95330793e-01
-2.70490646e-01 4.02207851e-01 -2.99601853e-02 1.92890704e-01
1.97818264e-01 6.39058292e-01 -5.06550848e-01 -6.10415876e-01
5.60802877e-01 6.17595017e-01 -7.62809992e-01 1.06857270e-01
3.25704180e-02 1.43065408e-01 -3.06056470e-01 4.84863400e-01
8.05725634e-01 -4.26448733e-01 3.03009152e-01 -8.44470918e-01
-2.49208033e-01 2.68333673e-01 -1.23144472e+00 1.46963787e+00
-9.19636250e-01 9.03654277e-01 -5.69135807e-02 -1.01493680e+00
8.50624323e-01 4.18905634e-03 7.85914898e-01 -1.09450865e+00
4.45687413e-01 1.37115996e-02 5.27140126e-02 -4.26378787e-01
7.57626116e-01 2.58778542e-01 1.48300692e-01 3.11463565e-01
-1.89154059e-01 4.26740378e-01 4.08273101e-01 -6.12621829e-02
1.16492260e+00 -2.01587960e-01 2.35065013e-01 -2.55130172e-01
5.46503305e-01 -3.85102570e-01 6.53387547e-01 5.94018877e-01
-2.62630135e-01 2.08473280e-01 2.01201644e-02 -9.21531439e-01
-1.29488301e+00 -8.13452363e-01 4.31786440e-02 1.15443671e+00
2.24703535e-01 -4.13274884e-01 -4.61451590e-01 -3.34189892e-01
-1.58619300e-01 5.39813280e-01 5.79101406e-02 -1.97059195e-02
-1.06803012e+00 -7.35084951e-01 3.21397156e-01 8.19493651e-01
8.84947658e-01 -8.29361975e-01 -1.26837206e+00 4.51230556e-01
1.73289448e-01 -1.95011270e+00 -1.81325227e-01 1.49930626e-01
-1.29236507e+00 -8.36893439e-01 -9.54282135e-02 -9.23724174e-01
4.55958188e-01 5.69055080e-01 1.24129593e+00 1.13230728e-01
-2.77000576e-01 -5.80927357e-02 -3.28700364e-01 -1.79948747e-01
-2.23071396e-01 3.39866728e-02 -8.60379115e-02 -6.29521757e-02
4.42185462e-01 -6.43985987e-01 -5.75762451e-01 2.91845530e-01
-8.44424605e-01 2.53318667e-01 8.22892725e-01 4.00883824e-01
5.84684193e-01 2.58155137e-01 4.13308203e-01 -5.09260058e-01
1.41827300e-01 -2.24668592e-01 -9.42409694e-01 -4.07092161e-02
-4.10306275e-01 1.71410561e-01 9.61719334e-01 -4.31344151e-01
-9.77445364e-01 3.46918225e-01 -4.41561997e-01 -4.58940089e-01
-1.81894362e-01 1.89049929e-01 -1.07138656e-01 -1.98972732e-01
1.13067560e-01 5.54226851e-03 -4.35427397e-01 -3.60946536e-01
1.56554267e-01 6.06352389e-01 5.29422283e-01 -2.79253274e-01
5.73321044e-01 5.97965360e-01 1.89532921e-01 -1.12454939e+00
-5.85734665e-01 -4.15817112e-01 -5.25314569e-01 -1.33427605e-01
7.53700495e-01 -1.13094997e+00 -9.91966069e-01 3.11781466e-01
-1.27391601e+00 -2.38436714e-01 -1.35061726e-01 7.54674375e-01
-3.10093284e-01 4.03384507e-01 -6.38973117e-01 -4.74899679e-01
-4.59321886e-01 -1.30373025e+00 9.46214259e-01 2.25271106e-01
1.24752335e-01 -7.68434286e-01 -4.85803127e-01 3.23707134e-01
5.27424276e-01 -9.78100970e-02 7.54691184e-01 -2.69458264e-01
-9.78421807e-01 -5.49499393e-02 -7.77781963e-01 3.46525431e-01
4.99184541e-02 -2.15739384e-01 -8.75322223e-01 -1.55195937e-01
-4.24400344e-02 -2.43991241e-02 8.02580059e-01 3.60293716e-01
1.69786167e+00 -2.07463399e-01 -1.43674061e-01 8.61364186e-01
1.68443274e+00 3.90426666e-01 7.63073802e-01 2.86044598e-01
9.10373092e-01 2.71473914e-01 4.76548314e-01 7.24893510e-01
3.06730360e-01 7.35701740e-01 4.93571103e-01 4.16587032e-02
-1.65301576e-01 2.35268071e-01 2.52051800e-01 1.08826768e+00
5.27288578e-02 -3.76679718e-01 -8.43879700e-01 5.40261030e-01
-1.72090769e+00 -7.66822577e-01 -1.56423375e-01 2.16781592e+00
3.32914084e-01 5.38693726e-01 -1.27390325e-01 4.47881371e-01
5.48203290e-01 2.15643749e-01 -6.01600468e-01 -7.09353447e-01
-2.96424655e-03 3.16807359e-01 9.59874213e-01 3.34502101e-01
-1.19414008e+00 8.15276146e-01 5.85697937e+00 7.15001225e-01
-1.36905336e+00 9.00018588e-02 6.88428879e-01 -9.57308933e-02
2.29690015e-01 -3.07437211e-01 -8.03674459e-01 6.16582394e-01
1.18327284e+00 5.27240261e-02 1.83837682e-01 9.90155220e-01
5.23009121e-01 1.12119034e-01 -8.51457059e-01 1.34741127e+00
-2.15126891e-02 -1.58583021e+00 1.95464715e-01 8.51418916e-03
5.46742260e-01 2.79679507e-01 -1.88998103e-01 -6.21495629e-03
-6.46445006e-02 -6.18583202e-01 5.85888088e-01 1.35303169e-01
6.86330438e-01 -9.30980325e-01 9.12704587e-01 6.47990927e-02
-1.57063258e+00 -3.99141133e-01 -4.84213293e-01 -1.19350433e-01
3.87525052e-01 8.83921266e-01 -6.01109326e-01 2.73976654e-01
8.20795834e-01 5.77453613e-01 -4.65610951e-01 1.05633509e+00
2.91809499e-01 5.52482367e-01 -3.44179064e-01 4.07636464e-02
2.01669261e-01 5.49429208e-02 1.76532939e-01 1.50936139e+00
3.52434039e-01 8.69869068e-02 2.42573664e-01 5.38242124e-02
-3.87275159e-01 4.41700332e-02 -2.52761781e-01 3.68934572e-01
4.95220751e-01 1.07944989e+00 -9.66018379e-01 -4.59718227e-01
-7.68696785e-01 9.71235156e-01 5.99436052e-02 -1.00171193e-01
-1.17077100e+00 -3.34626436e-01 8.09009552e-01 8.30890685e-02
6.43214285e-01 -7.36642480e-01 -6.33284271e-01 -8.28344882e-01
1.17884763e-01 -5.30368447e-01 9.30758342e-02 -4.37999606e-01
-8.66215467e-01 5.61458647e-01 -2.17764229e-01 -1.32396662e+00
-7.88410529e-02 -7.86471307e-01 -3.11942488e-01 3.37460339e-02
-1.68878722e+00 -9.54131305e-01 -6.18967354e-01 4.42438394e-01
9.93546069e-01 -2.56290346e-01 3.05016637e-01 8.29434335e-01
-6.87554717e-01 6.27830863e-01 8.02007839e-02 7.05294311e-02
2.62694567e-01 -6.33528829e-01 5.10099590e-01 9.70959663e-01
6.49994761e-02 3.19629371e-01 5.17686009e-01 -3.07094514e-01
-1.60148466e+00 -1.46364677e+00 9.28767502e-01 3.25355947e-01
4.08676475e-01 -8.27039033e-02 -8.53028178e-01 5.01259625e-01
1.60659179e-01 2.49826238e-01 5.61641753e-01 -3.39703530e-01
-3.46804470e-01 -6.72444403e-01 -1.21398365e+00 6.31049216e-01
1.12593305e+00 -2.60067374e-01 3.50025930e-02 2.65582651e-01
7.89151371e-01 -1.38807908e-01 -8.26229155e-01 4.94507790e-01
6.81837618e-01 -1.15219927e+00 1.11234844e+00 -2.56253153e-01
3.29139709e-01 -2.73945153e-01 -4.13748145e-01 -4.40613210e-01
-3.14232320e-01 -3.63306016e-01 -4.42365170e-01 9.80811477e-01
-4.08441983e-02 -2.50913143e-01 9.83714283e-01 4.40229803e-01
-1.58653498e-01 -6.99673712e-01 -8.14478219e-01 -8.86235535e-01
-5.71205497e-01 -1.03170896e+00 4.94978696e-01 5.81464708e-01
-5.78557193e-01 3.75389725e-01 -5.22281766e-01 2.70916551e-01
7.44400680e-01 -5.06606810e-02 6.74198806e-01 -1.05286515e+00
-2.10765600e-01 -3.09224963e-01 -8.20949852e-01 -1.30413604e+00
2.22512763e-02 -4.10968333e-01 -2.08049297e-01 -1.32048070e+00
9.88021307e-03 -4.68327940e-01 -3.72307569e-01 4.84023303e-01
2.99521476e-01 5.08257568e-01 4.25506592e-01 -4.65573072e-02
-9.26838815e-01 1.93472847e-01 6.61061227e-01 3.63204861e-03
-1.13824457e-01 -7.15539679e-02 -2.95204699e-01 1.10398424e+00
1.29354322e+00 -3.60694110e-01 -5.54735959e-01 -1.02340519e+00
-5.60953617e-02 -2.16110781e-01 2.93594182e-01 -1.66630352e+00
2.72731990e-01 -7.63950497e-02 3.98534000e-01 -7.27827251e-01
3.64926845e-01 -1.28317428e+00 5.61817028e-02 6.67925119e-01
-4.15155217e-02 3.88276935e-01 5.34041941e-01 2.52643079e-01
-1.08956397e-01 -1.51857957e-01 9.66845334e-01 7.64672905e-02
-1.24558282e+00 2.98967838e-01 -3.06384116e-01 -3.13078761e-01
1.01807916e+00 -3.54758352e-01 -8.76006708e-02 -3.07738364e-01
-2.11169899e-01 -1.05229206e-01 2.71241248e-01 3.76882017e-01
8.19102645e-01 -1.13344169e+00 -4.95708972e-01 3.15242887e-01
-1.83188856e-01 -1.95780963e-01 4.06888247e-01 4.76445705e-01
-8.21297407e-01 5.27949631e-01 -3.98409963e-01 -7.71896064e-01
-1.48715448e+00 5.33125997e-01 1.89225730e-02 -3.73774499e-01
-5.29827297e-01 5.04397392e-01 -5.22622764e-02 3.77819955e-01
3.38543236e-01 -4.23840612e-01 -2.68295463e-02 -3.15727085e-01
7.17432261e-01 7.41589248e-01 2.57071316e-01 -7.64179945e-01
-4.26739901e-01 7.07258582e-01 -8.42726380e-02 4.64802414e-01
1.32607746e+00 -2.68328160e-01 1.33364990e-01 -2.35042144e-02
1.47775877e+00 -9.25636664e-02 -1.35512817e+00 -6.67897239e-02
2.02697188e-01 -4.91155028e-01 3.00649136e-01 -3.13751996e-01
-1.41862190e+00 8.48826349e-01 1.02774489e+00 1.31655708e-01
1.38876295e+00 -3.65561754e-01 1.31283653e+00 7.22383618e-01
5.65653145e-01 -1.05686748e+00 8.69214758e-02 5.37980080e-01
2.96940595e-01 -1.38364816e+00 2.13177770e-01 -6.28027260e-01
-3.06353956e-01 1.11965191e+00 4.45210755e-01 -1.42042801e-01
6.93618596e-01 7.42169678e-01 -5.39568625e-02 2.66332775e-01
-6.00531638e-01 -2.23001525e-01 2.75118332e-02 4.22427088e-01
4.85865146e-01 -1.06075376e-01 -3.54618520e-01 1.51534095e-01
-4.77586687e-02 1.43989846e-01 2.86474496e-01 1.12209582e+00
-5.91111898e-01 -1.19911051e+00 -5.57586737e-02 5.17939508e-01
-5.80792367e-01 -2.72204615e-02 4.77947742e-01 5.55588007e-01
2.91753948e-01 1.08202469e+00 5.74216425e-01 -7.02947736e-01
1.19577631e-01 -3.27986181e-01 2.63848841e-01 -6.24890812e-02
-4.15836960e-01 3.36080529e-02 6.37132898e-02 -8.88687909e-01
-6.52065337e-01 -2.39390716e-01 -1.50761747e+00 -5.39317012e-01
3.95202450e-03 -2.91006625e-01 1.02729106e+00 9.03570652e-01
6.98224485e-01 7.86180913e-01 7.40822613e-01 -7.97982991e-01
-1.15406543e-01 -4.07872975e-01 -1.67893767e-01 3.77087772e-01
2.69174218e-01 -3.49633873e-01 9.87929478e-02 3.94949704e-01] | [9.062188148498535, -0.266541451215744] |
2ce7c12d-167b-45f9-8536-22e8ea29e283 | implicit-semantic-response-alignment-for | null | null | http://proceedings.neurips.cc/paper/2021/hash/731b03008e834f92a03085ef47061c4a-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/731b03008e834f92a03085ef47061c4a-Paper.pdf | Implicit Semantic Response Alignment for Partial Domain Adaptation | Partial Domain Adaptation (PDA) addresses the unsupervised domain adaptation problem where the target label space is a subset of the source label space. Most state-of-art PDA methods tackle the inconsistent label space by assigning weights to classes or individual samples, in an attempt to discard the source data that belongs to the irrelevant classes. However, we believe samples from those extra categories would still contain valuable information to promote positive transfer. In this paper, we propose the Implicit Semantic Response Alignment to explore the intrinsic relationships among different categories by applying a weighted schema on the feature level. Specifically, we design a class2vec module to extract the implicit semantic topics from the visual features. With an attention layer, we calculate the semantic response according to each implicit semantic topic. Then semantic responses of source and target data are aligned to retain the relevant information contained in multiple categories by weighting the features, instead of samples. Experiments on several cross-domain benchmark datasets demonstrate the effectiveness of our method over the state-of-the-art PDA methods. Moreover, we elaborate in-depth analyses to further explore implicit semantic alignment. | ['Hongfu Liu', 'Zhengming Ding', 'Wenxiao Xiao'] | 2021-12-01 | null | https://openreview.net/forum?id=LNXTIrMqyGz | https://openreview.net/pdf?id=LNXTIrMqyGz | neurips-2021-12 | ['partial-domain-adaptation'] | ['methodology'] | [ 5.02781570e-01 8.34284201e-02 -4.43908483e-01 -6.40201092e-01
-4.43822175e-01 -4.17747259e-01 4.33339566e-01 2.82737732e-01
-4.48975295e-01 5.45779824e-01 4.70538557e-01 3.90217334e-01
-1.72630936e-01 -8.49348843e-01 -3.43267113e-01 -9.16479945e-01
5.08282304e-01 4.94097978e-01 3.92029345e-01 -1.84160873e-01
3.01165640e-01 -4.11930010e-02 -1.57090950e+00 6.23627067e-01
9.96157646e-01 1.07296097e+00 2.44675517e-01 -4.24188584e-01
-7.14710832e-01 4.63026166e-01 -4.78305578e-01 -2.39405766e-01
-3.62004004e-02 -4.61610675e-01 -6.93757355e-01 2.29971066e-01
3.08149815e-01 1.86466753e-01 1.68214321e-01 1.31708062e+00
4.05467123e-01 1.86933860e-01 8.54686975e-01 -1.42293239e+00
-8.68233860e-01 3.51078749e-01 -6.17657304e-01 2.24266089e-02
-2.83491239e-02 -9.69527587e-02 1.18106258e+00 -1.20694625e+00
7.69979477e-01 1.28222191e+00 2.67622083e-01 6.73922181e-01
-1.40958691e+00 -9.44094419e-01 7.31962025e-01 4.29185897e-01
-1.25265741e+00 -1.85747623e-01 1.35968494e+00 -5.46854675e-01
3.70841831e-01 4.85823527e-02 2.66649753e-01 1.25892043e+00
-4.09670293e-01 8.55869949e-01 1.13712800e+00 -5.23590386e-01
3.69254380e-01 8.68203402e-01 6.27996743e-01 1.97652012e-01
7.59164393e-02 -1.82752594e-01 -5.58997214e-01 -3.69122513e-02
2.31847376e-01 8.56061801e-02 -1.70709282e-01 -1.15654182e+00
-1.18470979e+00 1.18281567e+00 5.74404955e-01 4.18629020e-01
-2.60857940e-01 -5.08373022e-01 3.80186111e-01 2.15192914e-01
7.79279292e-01 3.94935280e-01 -7.79439270e-01 5.44461489e-01
-3.46853971e-01 6.18645474e-02 4.12220508e-01 1.06262732e+00
1.05108047e+00 -2.88416147e-01 -3.70847285e-01 1.21359813e+00
4.63057697e-01 3.34951311e-01 6.88580096e-01 -6.17673576e-01
4.90291446e-01 1.02249670e+00 -5.33804931e-02 -1.15332544e+00
-3.63592476e-01 -5.16436279e-01 -5.92422247e-01 3.54127586e-02
3.04310501e-01 1.51098415e-01 -7.47328579e-01 2.07172132e+00
6.23391211e-01 7.02123269e-02 2.00764209e-01 1.04661942e+00
9.04980242e-01 5.07155538e-01 6.08310759e-01 -2.32018366e-01
1.47924578e+00 -1.09719348e+00 -6.70224130e-01 -5.01464128e-01
5.83420753e-01 -6.11825347e-01 1.36265135e+00 9.50765237e-02
-3.81586432e-01 -6.47775710e-01 -9.40592587e-01 1.10981770e-01
-4.49917912e-01 2.21042737e-01 3.03122729e-01 2.96549350e-01
-4.36090916e-01 2.70600379e-01 -2.60997027e-01 -5.48678994e-01
4.94826257e-01 2.80440390e-01 -4.26322341e-01 -2.26416901e-01
-1.41071200e+00 6.63653910e-01 5.04892409e-01 -5.53404629e-01
-4.29655373e-01 -7.64595091e-01 -9.11650658e-01 6.52685687e-02
3.60633403e-01 -4.88579452e-01 8.56819510e-01 -1.60539258e+00
-1.19298959e+00 9.76340294e-01 -2.95989305e-01 7.63335228e-02
-3.39178480e-02 -3.14672031e-02 -5.91835558e-01 -2.55231466e-03
3.97801995e-01 8.21035802e-01 9.06632662e-01 -1.51546061e+00
-9.67141390e-01 -6.09830022e-01 -2.60574371e-01 4.94423151e-01
-9.27141726e-01 -1.11373812e-01 -3.58823955e-01 -8.19312513e-01
2.80420989e-01 -8.76385689e-01 -3.35342437e-02 -4.17159982e-02
-3.11781274e-04 -4.88847494e-01 9.31037664e-01 -3.68181556e-01
1.16944981e+00 -2.44759703e+00 3.89950544e-01 1.16780445e-01
1.60235345e-01 3.64955030e-02 -4.32644010e-01 -5.23237735e-02
-1.95148855e-01 -3.00913900e-01 -3.82871807e-01 -5.92094800e-03
2.46694632e-04 9.14225131e-02 -4.28325295e-01 1.35459423e-01
3.22437853e-01 5.13953686e-01 -1.01766384e+00 -5.58408856e-01
-5.45857660e-02 1.82151288e-01 -5.06327033e-01 2.27066383e-01
-2.83346176e-01 4.18208838e-01 -8.07664752e-01 4.95480090e-01
8.28004420e-01 -2.24821493e-01 2.67324239e-01 -4.40839797e-01
1.49847507e-01 3.21454674e-01 -1.01992714e+00 1.79799998e+00
-1.93509459e-01 2.81259954e-01 -2.02581957e-01 -1.33373332e+00
1.17344213e+00 -1.50115434e-02 6.32396400e-01 -9.53499794e-01
4.56455611e-02 1.57316253e-01 -1.36735678e-01 -4.01721388e-01
3.50394309e-01 -4.24457908e-01 -3.04439068e-01 3.34558010e-01
2.93563038e-01 1.97542056e-01 -9.85663608e-02 9.36325118e-02
4.76110518e-01 9.44377705e-02 3.84846807e-01 -4.07444775e-01
7.08842218e-01 2.41385341e-01 7.36300588e-01 3.33479255e-01
-3.81665081e-01 4.34080005e-01 2.85433620e-01 -2.26946935e-01
-8.67716968e-01 -1.14525056e+00 -1.20207660e-01 1.53506362e+00
5.74409842e-01 -1.53698683e-01 -6.51733100e-01 -1.29309893e+00
2.28420720e-02 9.02897656e-01 -7.93681920e-01 -5.25718451e-01
-2.48621002e-01 -7.58096278e-01 -1.07080102e-01 6.20942831e-01
2.76655495e-01 -1.21444929e+00 -1.14155956e-01 3.01236331e-01
-3.98247540e-01 -7.26193368e-01 -4.30254728e-01 2.16905043e-01
-7.58214474e-01 -9.58401978e-01 -7.01463401e-01 -1.13405454e+00
9.27728951e-01 5.30707419e-01 9.85137701e-01 -4.69826400e-01
8.92838538e-02 2.22797871e-01 -5.83197653e-01 -3.83628488e-01
-5.16845062e-02 6.26891851e-02 7.72163868e-02 3.49267364e-01
1.13856602e+00 -2.79812306e-01 -4.39579517e-01 4.69652086e-01
-7.27332532e-01 2.07603946e-01 3.62987846e-01 9.41873789e-01
7.67781556e-01 3.29596847e-02 7.86732316e-01 -1.20565736e+00
4.96433139e-01 -8.16217959e-01 -1.81079865e-01 2.66855389e-01
-6.56272173e-01 1.99266374e-01 4.16074783e-01 -7.67102540e-01
-1.42922676e+00 1.20863698e-01 4.01691139e-01 -2.66365707e-01
-2.80510306e-01 3.51165205e-01 -7.00279176e-01 3.36430162e-01
6.85019672e-01 1.99735805e-01 -1.08681567e-01 -5.32744110e-01
4.63170171e-01 7.42237806e-01 2.00183466e-01 -7.13520408e-01
5.84477961e-01 5.14588714e-01 -3.73299450e-01 -3.34580630e-01
-1.25517499e+00 -7.25465119e-01 -7.57330358e-01 -1.21047758e-02
8.35919738e-01 -9.16849017e-01 3.53235193e-02 2.95713812e-01
-9.05852079e-01 -6.88335523e-02 -4.62925106e-01 4.70541060e-01
-3.29384238e-01 2.56577730e-01 -1.34474233e-01 -3.16956401e-01
-4.58400063e-02 -9.81542528e-01 9.82215226e-01 2.70617694e-01
-3.78509879e-01 -9.27494705e-01 1.49508134e-01 2.14538440e-01
2.65505046e-01 -1.47177234e-01 1.27712536e+00 -1.10618508e+00
-4.77791578e-02 1.38387784e-01 -4.92650628e-01 2.58528411e-01
5.23430526e-01 -5.47746420e-01 -1.10275936e+00 -2.46444181e-01
1.20223753e-01 -1.41561478e-01 1.00338793e+00 1.95303544e-01
1.09635198e+00 -3.63786250e-01 -5.88446915e-01 2.85379410e-01
1.31524491e+00 2.49523431e-01 3.98902625e-01 4.76679981e-01
8.00818026e-01 1.17761624e+00 1.12355089e+00 3.62900555e-01
4.18072909e-01 7.39736438e-01 2.31806606e-01 -1.35061517e-01
-8.49239454e-02 -2.92839080e-01 3.14991504e-01 7.64260650e-01
5.42625308e-01 -1.12375095e-01 -6.53392136e-01 7.64058828e-01
-1.89071918e+00 -7.08525360e-01 -1.09112918e-01 2.17515540e+00
1.01919448e+00 2.93805022e-02 5.30043505e-02 -3.97858955e-02
1.07436407e+00 -4.14766558e-02 -6.19532168e-01 -5.03676608e-02
-1.40200987e-01 8.33980069e-02 1.78317383e-01 1.45494148e-01
-1.44611537e+00 1.07274806e+00 4.94230318e+00 9.51665044e-01
-9.71758723e-01 1.63173258e-01 4.02621508e-01 3.69047709e-02
-5.81170440e-01 5.29475696e-02 -8.49187732e-01 5.58608651e-01
4.09586191e-01 -2.86748141e-01 7.75909945e-02 1.11043572e+00
-2.37607285e-01 2.07404330e-01 -1.08467197e+00 7.28010356e-01
2.34391823e-01 -7.86398411e-01 2.16095701e-01 -8.91004950e-02
8.28586817e-01 -2.87912458e-01 1.61418021e-01 5.69728017e-01
3.16889435e-01 -4.11276758e-01 5.97855091e-01 3.60787690e-01
6.44123912e-01 -7.17028558e-01 6.25514925e-01 1.90423772e-01
-1.09726465e+00 -1.74692512e-01 -6.81809783e-01 2.08102629e-01
-2.63134569e-01 5.49911618e-01 -6.28324032e-01 3.42329204e-01
7.87123084e-01 1.05118179e+00 -6.62892163e-01 7.19929993e-01
-2.70970613e-01 4.67684418e-01 9.45418999e-02 1.41887441e-01
1.58518851e-01 -1.55749545e-01 4.22052354e-01 1.14937222e+00
3.49648207e-01 3.14444005e-02 2.14374781e-01 7.37214506e-01
-1.54364621e-02 5.32695711e-01 -4.86795217e-01 2.66173661e-01
5.03389597e-01 1.19844174e+00 -4.23815995e-01 -5.21804094e-01
-7.16106117e-01 9.71737564e-01 4.67305809e-01 3.61542165e-01
-6.52532041e-01 -3.58158261e-01 8.22848260e-01 -5.89607097e-02
3.04096878e-01 3.43179286e-01 -4.85708922e-01 -1.21354496e+00
-1.56363502e-01 -7.10121632e-01 8.86527419e-01 -6.34705901e-01
-1.78849852e+00 3.04297566e-01 -1.12490274e-01 -1.70503962e+00
1.82730615e-01 -4.59505439e-01 -2.42709219e-01 8.02235484e-01
-1.56129348e+00 -1.03470814e+00 -4.27560508e-01 7.01306343e-01
7.24655151e-01 -3.67189080e-01 8.88005793e-01 2.34176561e-01
-2.83526629e-01 6.06102347e-01 2.58139402e-01 -4.56468697e-04
1.34421706e+00 -1.06734335e+00 -7.79245123e-02 4.94352937e-01
-4.76384349e-02 5.87173820e-01 4.89888579e-01 -5.93948066e-01
-6.53428614e-01 -1.34671998e+00 9.88863826e-01 -2.95282304e-01
5.72368085e-01 -1.73792452e-01 -1.36336219e+00 5.18028736e-01
1.34407133e-01 2.41155103e-02 1.08649409e+00 2.47958764e-01
-7.93217182e-01 -3.18621993e-01 -1.17532015e+00 5.32736599e-01
9.21766579e-01 -3.47803742e-01 -9.25431073e-01 3.24188359e-02
9.97099578e-01 1.72313347e-01 -5.25188386e-01 3.99389058e-01
3.01172018e-01 -4.60683614e-01 9.98717904e-01 -7.69914865e-01
5.89196742e-01 -4.90322798e-01 -3.29825968e-01 -1.68280351e+00
-8.18139851e-01 3.89440507e-01 1.10993735e-01 1.67206728e+00
2.65996903e-01 -5.28740108e-01 6.56430304e-01 5.28064847e-01
-3.99833545e-02 -1.04242958e-01 -6.83003306e-01 -6.06271625e-01
1.07265800e-01 -2.19349131e-01 6.61764979e-01 1.44512880e+00
2.00726271e-01 6.44959986e-01 -1.98737368e-01 2.29696352e-02
7.46364355e-01 3.83926928e-01 5.81368029e-01 -1.61572909e+00
6.06148317e-03 -4.17525202e-01 -1.88029200e-01 -6.88210249e-01
4.73864287e-01 -1.14685178e+00 8.21282491e-02 -1.29682600e+00
5.73602140e-01 -6.64597332e-01 -1.07247627e+00 6.76059306e-01
-4.37510371e-01 1.61291808e-01 1.04907013e-01 3.73863012e-01
-7.20760703e-01 7.36564457e-01 1.11943150e+00 -5.19266844e-01
-3.07842314e-01 -2.63245285e-01 -1.13213193e+00 7.15685964e-01
7.83542156e-01 -8.20007443e-01 -6.39510334e-01 -3.67291451e-01
-1.81019038e-01 -4.81304973e-01 2.59733915e-01 -8.27717185e-01
4.03107479e-02 -5.41299939e-01 4.55010772e-01 -4.26621884e-01
9.62528214e-02 -1.15390015e+00 -1.17879093e-01 1.98556781e-01
-7.90791154e-01 -7.40879297e-01 1.17238849e-01 6.93626463e-01
-3.58988822e-01 -2.07957610e-01 9.76448536e-01 8.36072937e-02
-1.10913599e+00 1.55516982e-01 -8.65580440e-02 2.26205468e-01
1.14130914e+00 -6.17261901e-02 -3.29649150e-01 1.62599996e-01
-7.92032361e-01 2.20914662e-01 5.04015386e-01 8.18564475e-01
5.42817652e-01 -1.73635685e+00 -6.59982324e-01 2.92523235e-01
9.20141637e-01 -2.29493588e-01 4.35432315e-01 5.03721237e-01
3.47452849e-01 3.19921106e-01 -4.71773952e-01 -4.94429708e-01
-1.34582245e+00 1.06695497e+00 -1.04152195e-01 -1.74880430e-01
-3.88714224e-01 8.02538991e-01 1.17621517e+00 -6.12325847e-01
4.49168943e-02 9.51909944e-02 -7.46136963e-01 5.74515462e-01
5.10220051e-01 1.60248503e-02 -1.86722532e-01 -7.02219725e-01
-5.95070601e-01 6.49632871e-01 -2.29667664e-01 1.32876173e-01
1.30272865e+00 -3.24739009e-01 -2.59734716e-05 3.63918066e-01
1.30128407e+00 -1.17173493e-01 -1.27742147e+00 -9.14944530e-01
2.13347688e-01 -4.83801961e-01 -1.93260178e-01 -9.28652823e-01
-1.14894223e+00 9.85525727e-01 8.33239079e-01 -2.80408531e-01
1.39043665e+00 1.96314350e-01 4.48463887e-01 -1.41669130e-02
2.05088928e-01 -1.36210573e+00 2.02733055e-01 3.61334383e-01
7.72482514e-01 -1.33416605e+00 -1.71564430e-01 -6.28602326e-01
-1.11209691e+00 8.91118467e-01 1.06758785e+00 -6.21721940e-03
5.12003481e-01 -2.69421369e-01 6.14040568e-02 -5.02625778e-02
-6.24449074e-01 -1.97829336e-01 5.20189285e-01 8.74778569e-01
3.64487261e-01 1.42275319e-01 -4.67812747e-01 1.20070612e+00
3.40134501e-01 -2.19578177e-01 4.26450409e-02 6.43196166e-01
-7.13465810e-01 -1.31929815e+00 -2.73477882e-01 2.18968898e-01
-2.41963491e-02 -2.36708242e-02 -5.79951465e-01 5.58847308e-01
3.41614604e-01 8.22325468e-01 2.00612739e-01 -3.56929541e-01
5.18569291e-01 3.28040391e-01 1.77993745e-01 -6.86747909e-01
-3.96699607e-01 1.55363381e-01 -1.46385372e-01 -2.84726083e-01
-5.03451943e-01 -7.31659055e-01 -1.29450405e+00 2.57627845e-01
-2.50789613e-01 1.08040236e-01 4.50374573e-01 7.99841881e-01
4.21239644e-01 5.51837981e-01 7.55381048e-01 -3.36980402e-01
-3.42953116e-01 -9.63411510e-01 -6.08723104e-01 9.78000581e-01
2.06712522e-02 -1.04238451e+00 -2.82698393e-01 2.20391154e-01] | [10.422688484191895, 3.0616345405578613] |
c04684b3-838d-4a19-b2aa-319eac8f457b | convex-aggregation-for-opinion-summarization | 2104.01371 | null | https://arxiv.org/abs/2104.01371v3 | https://arxiv.org/pdf/2104.01371v3.pdf | Convex Aggregation for Opinion Summarization | Recent advances in text autoencoders have significantly improved the quality of the latent space, which enables models to generate grammatical and consistent text from aggregated latent vectors. As a successful application of this property, unsupervised opinion summarization models generate a summary by decoding the aggregated latent vectors of inputs. More specifically, they perform the aggregation via simple average. However, little is known about how the vector aggregation step affects the generation quality. In this study, we revisit the commonly used simple average approach by examining the latent space and generated summaries. We found that text autoencoders tend to generate overly generic summaries from simply averaged latent vectors due to an unexpected $L_2$-norm shrinkage in the aggregated latent vectors, which we refer to as summary vector degeneration. To overcome this issue, we develop a framework Coop, which searches input combinations for the latent vector aggregation using input-output word overlap. Experimental results show that Coop successfully alleviates the summary vector degeneration issue and establishes new state-of-the-art performance on two opinion summarization benchmarks. Code is available at \url{https://github.com/megagonlabs/coop}. | ['Wang-Chiew Tan', 'Stefanos Angelidis', 'Yoshihiko Suhara', 'Xiaolan Wang', 'Hayate Iso'] | 2021-04-03 | null | https://aclanthology.org/2021.findings-emnlp.328 | https://aclanthology.org/2021.findings-emnlp.328.pdf | findings-emnlp-2021-11 | ['unsupervised-opinion-summarization'] | ['natural-language-processing'] | [ 3.37368190e-01 1.88755661e-01 -5.35724498e-02 -2.79587448e-01
-8.67932796e-01 -4.65425253e-01 7.40825772e-01 2.78799951e-01
-8.35762694e-02 7.68012524e-01 8.61297727e-01 -5.32601476e-02
1.97185814e-01 -7.26719975e-01 -6.84135079e-01 -7.79551744e-01
4.12040383e-01 1.70464814e-01 -3.20392877e-01 -1.25642061e-01
4.42919165e-01 -1.24800935e-01 -1.71878588e+00 5.13456404e-01
1.48170125e+00 5.08874953e-01 9.72947404e-02 8.01389039e-01
-4.07551080e-01 6.38808787e-01 -1.17287707e+00 -5.59261143e-01
-7.79714063e-02 -8.70612800e-01 -5.12645006e-01 1.21134147e-01
6.15143061e-01 -4.91699368e-01 -1.79965377e-01 1.07761502e+00
7.03444242e-01 2.73348242e-01 8.73113036e-01 -1.12881875e+00
-1.04537320e+00 1.22747827e+00 -2.63170779e-01 4.63892557e-02
2.56426662e-01 -3.67552065e-03 1.49297047e+00 -9.25318837e-01
5.69360077e-01 1.10739565e+00 4.97626066e-01 5.85237741e-01
-1.33017087e+00 -5.30598402e-01 1.81253612e-01 -4.10543010e-02
-1.11542046e+00 -6.22219443e-01 8.58456850e-01 -2.33811557e-01
1.32868087e+00 4.45954740e-01 6.61484659e-01 1.21498501e+00
5.86466312e-01 1.24024844e+00 6.82806194e-01 -4.37987000e-01
3.60379964e-01 3.23742926e-02 5.19616127e-01 3.91542375e-01
5.64159214e-01 -5.20836413e-01 -7.04259813e-01 -1.58790812e-01
3.23220909e-01 -1.63396344e-01 -3.96931291e-01 -6.74396800e-03
-1.08071566e+00 1.00289357e+00 1.22225359e-01 4.31358069e-01
-5.92226505e-01 1.82985187e-01 3.13999474e-01 1.69922233e-01
8.54424655e-01 6.38521314e-01 -1.96637392e-01 -3.52339298e-01
-1.28495002e+00 3.53117675e-01 8.07790041e-01 7.96218038e-01
5.82288265e-01 6.37440443e-01 -4.82045293e-01 8.63059580e-01
7.88362175e-02 5.10908842e-01 8.99351060e-01 -1.01891541e+00
4.74413097e-01 6.82036102e-01 -1.42441660e-01 -1.01269901e+00
9.55887977e-03 -7.07161903e-01 -1.05045104e+00 -2.83222347e-01
-2.54037771e-02 -2.65043408e-01 -9.56938028e-01 1.69903994e+00
-1.64360508e-01 -2.17425361e-01 5.52541316e-01 4.87868994e-01
1.08992600e+00 1.17893767e+00 -1.33976609e-01 -4.29545760e-01
9.98524606e-01 -1.00777578e+00 -1.15309584e+00 -2.93835491e-01
5.68597376e-01 -7.19596326e-01 1.16599834e+00 2.66726196e-01
-1.36632252e+00 -4.02987480e-01 -1.27738166e+00 -1.13528669e-01
-3.41766417e-01 3.15985203e-01 4.90703672e-01 3.46344471e-01
-1.24481261e+00 8.30994070e-01 -8.52526426e-01 -3.96544307e-01
2.46628508e-01 1.26183257e-01 -9.01462287e-02 4.20642734e-01
-1.18870151e+00 5.64231515e-01 6.18841052e-01 -1.01539850e-01
-3.57265979e-01 -5.97433448e-01 -9.27479625e-01 3.31024051e-01
2.09754109e-01 -1.01787460e+00 1.35318232e+00 -9.62368369e-01
-1.48524535e+00 2.42850170e-01 -5.00879526e-01 -5.53011537e-01
1.25563249e-01 -4.83898014e-01 -2.29180336e-01 -1.68881118e-02
2.38257885e-01 7.64893711e-01 9.76915061e-01 -1.28272772e+00
-3.63043010e-01 -1.16771542e-01 -3.48695904e-01 2.51661956e-01
-7.14245141e-01 -3.59436363e-01 -2.86490917e-01 -9.97043014e-01
-9.31330398e-02 -6.62121177e-01 -3.84725481e-02 -9.11784530e-01
-6.48925900e-01 -5.42925894e-01 4.07821149e-01 -6.73923314e-01
1.72181094e+00 -1.92037821e+00 3.89218539e-01 -1.68747798e-01
2.70753860e-01 1.66129827e-01 -2.35213876e-01 8.17009568e-01
-1.75425839e-02 3.43900770e-01 -3.85348141e-01 -7.59505272e-01
2.44874448e-01 1.61957845e-01 -8.05380583e-01 -7.13316649e-02
1.98215052e-01 1.02460635e+00 -8.61118019e-01 -4.99649137e-01
-1.31642655e-01 4.90446657e-01 -6.25245094e-01 1.30642056e-01
-2.98952550e-01 -1.02451794e-01 -1.79859146e-01 3.13536823e-01
3.18634868e-01 -4.55375373e-01 2.15218719e-02 -1.79138273e-01
-1.50806576e-01 4.12353218e-01 -8.39226484e-01 1.63227439e+00
-1.68173268e-01 9.39901590e-01 -5.58881044e-01 -7.92572081e-01
8.26407850e-01 2.52284288e-01 2.66287774e-01 -3.49356622e-01
2.37560794e-01 2.28794530e-01 -2.22655833e-01 -9.33254212e-02
1.23754370e+00 6.61479086e-02 -1.75096631e-01 6.84224784e-01
3.41362387e-01 -2.95662433e-01 6.24597788e-01 6.39860094e-01
8.38811398e-01 -7.20282868e-02 4.29192036e-01 -1.81933060e-01
1.99050590e-01 -6.61447272e-02 4.23674226e-01 1.02759206e+00
3.22600931e-01 7.48294473e-01 7.05475867e-01 -9.48695168e-02
-1.14532828e+00 -1.17721725e+00 1.43856093e-01 1.00875485e+00
-2.26760909e-01 -9.35387015e-01 -1.04465342e+00 -4.85493183e-01
-7.98091367e-02 1.31735849e+00 -5.52568376e-01 -3.45225573e-01
-4.90291655e-01 -9.01453376e-01 5.62734663e-01 6.29244328e-01
2.82140195e-01 -1.09244788e+00 -4.06647235e-01 2.26753339e-01
-4.35849369e-01 -6.67178988e-01 -6.19936526e-01 -3.82233709e-02
-9.72953916e-01 -2.73839116e-01 -9.51422453e-01 -4.18924183e-01
7.90819108e-01 9.52112526e-02 1.30320370e+00 -1.09157637e-01
1.07722647e-01 3.37368250e-01 -5.23765862e-01 -5.27575374e-01
-7.14110076e-01 4.29370314e-01 3.15289348e-01 -2.45909486e-02
3.61923277e-01 -5.84284902e-01 -4.67209071e-01 -3.27857226e-01
-1.26929855e+00 3.31136793e-01 7.27444887e-01 8.77281308e-01
4.73282218e-01 -5.74348718e-02 8.59667897e-01 -7.10856676e-01
1.47729063e+00 -3.42126608e-01 -2.30917901e-01 2.24632666e-01
-7.69328415e-01 4.65567440e-01 6.92680538e-01 -3.31633627e-01
-8.79996955e-01 -4.20811236e-01 -1.33434117e-01 -1.52373150e-01
9.02042612e-02 7.06623256e-01 7.87981600e-02 9.04738367e-01
5.26075959e-01 7.23525882e-01 -2.34540291e-02 -3.24624628e-01
5.67810953e-01 8.28057468e-01 4.28851604e-01 -2.83342361e-01
5.08586287e-01 1.80425823e-01 -5.97539783e-01 -9.85203683e-01
-8.77566278e-01 -1.34533450e-01 -2.79963285e-01 -5.57179041e-02
8.56250048e-01 -8.10164988e-01 -1.28831431e-01 4.75430846e-01
-1.40138185e+00 -1.78663611e-01 -4.58235174e-01 3.20367098e-01
-4.91291314e-01 5.79030752e-01 -5.34977734e-01 -6.35088921e-01
-1.02613771e+00 -1.04176426e+00 1.12861860e+00 3.44408363e-01
-8.52072299e-01 -8.90753329e-01 1.83722049e-01 3.58677544e-02
5.06363332e-01 1.68610979e-02 9.46901858e-01 -7.98019588e-01
-2.81586319e-01 -2.85342366e-01 5.95891625e-02 6.70985639e-01
1.45172402e-01 1.80323124e-01 -8.12391520e-01 -2.22484559e-01
-8.99315870e-04 3.61391827e-02 1.25080061e+00 5.63966990e-01
9.33108509e-01 -7.65787542e-01 -7.72692487e-02 3.48373622e-01
1.06930959e+00 5.60504105e-03 5.24646938e-01 -6.67803362e-02
6.26291573e-01 4.96798754e-01 1.45047531e-01 5.45190454e-01
2.91003376e-01 4.18858975e-01 -5.75953722e-02 1.70537427e-01
-1.53256059e-01 -3.72318089e-01 7.34289587e-01 1.66067898e+00
8.48066732e-02 -7.37974882e-01 -7.05093563e-01 6.51692688e-01
-1.90920901e+00 -1.04390466e+00 6.77601025e-02 1.89510810e+00
1.05682743e+00 1.50019795e-01 -2.85798192e-01 1.19320631e-01
5.03124714e-01 5.10913491e-01 -5.48054397e-01 -6.22374356e-01
-2.65442282e-01 4.46851999e-02 1.76506177e-01 6.58837497e-01
-7.94891357e-01 9.43636000e-01 5.95828819e+00 1.01957297e+00
-9.38786626e-01 -9.85933915e-02 4.19393629e-01 -3.52935582e-01
-8.60507488e-01 -2.29244173e-01 -9.14636970e-01 5.73630452e-01
9.71987545e-01 -5.83275199e-01 1.12017028e-01 7.03075945e-01
2.04105243e-01 1.31730530e-02 -1.03545630e+00 8.74603212e-01
5.19300699e-01 -1.47280514e+00 7.53427923e-01 -5.69023080e-02
1.06229758e+00 -2.21491218e-01 1.76356986e-01 3.88818920e-01
2.87839293e-01 -8.37811589e-01 6.25148833e-01 6.26418948e-01
5.91900051e-01 -7.15149283e-01 7.61804581e-01 2.48622924e-01
-8.91079068e-01 2.67010368e-02 -3.05753767e-01 1.32212220e-02
2.66422570e-01 7.92847455e-01 -8.28759372e-01 4.43543404e-01
3.47980857e-01 8.58963668e-01 -6.33766711e-01 5.04429638e-01
-3.09169501e-01 7.85091758e-01 -5.81339113e-02 -3.50913405e-01
2.31542841e-01 -1.91531941e-01 9.42239285e-01 1.24463534e+00
5.64438105e-01 -8.68464857e-02 -2.99554318e-01 1.02328587e+00
-1.18377566e-01 2.31826350e-01 -7.18398452e-01 -5.13653696e-01
3.81941438e-01 9.14496183e-01 -5.91837883e-01 -7.27054060e-01
5.65808304e-02 1.21576524e+00 1.40694126e-01 5.45214355e-01
-6.64004266e-01 -6.37243330e-01 5.62924802e-01 -2.93467343e-01
3.88645887e-01 -2.19921172e-01 -4.52491283e-01 -1.56747425e+00
1.88721076e-01 -9.34979200e-01 1.37689620e-01 -8.15645993e-01
-1.12094736e+00 8.25844824e-01 1.55373514e-01 -1.05752563e+00
-7.15084970e-01 -3.99992131e-02 -7.52158165e-01 5.88858902e-01
-9.90701973e-01 -7.82837808e-01 -1.30154938e-01 -1.47167938e-02
1.06483960e+00 -3.93300354e-01 8.72642040e-01 -1.86767265e-01
-7.33187139e-01 7.56510675e-01 4.54325974e-01 -1.29426628e-01
5.86064577e-01 -1.51416874e+00 6.50138676e-01 1.12147415e+00
2.28948459e-01 9.23770547e-01 1.11565161e+00 -7.98426628e-01
-1.17836523e+00 -9.09189701e-01 1.22179759e+00 -5.33217251e-01
5.49287379e-01 -2.97457188e-01 -9.10622954e-01 6.43665791e-01
8.05079043e-01 -9.33665514e-01 9.45017517e-01 -1.17118716e-01
-4.46161628e-02 5.63748367e-03 -5.18034101e-01 1.09395146e+00
7.31264889e-01 -2.95105547e-01 -9.20886993e-01 7.68844485e-02
1.11933601e+00 -7.39204809e-02 -7.15741217e-01 3.76475513e-01
5.83190978e-01 -9.18840230e-01 6.14863813e-01 -3.63409758e-01
9.45896447e-01 -1.12839311e-01 -1.36877403e-01 -1.74925399e+00
-2.25873679e-01 -6.86804175e-01 -4.63622838e-01 1.44413507e+00
5.26267231e-01 -6.77083910e-01 5.06675780e-01 3.44804615e-01
-4.40554887e-01 -9.27307367e-01 -5.11308014e-01 -4.81162965e-01
2.16402575e-01 -2.07682624e-01 5.90457082e-01 5.87601840e-01
9.18739066e-02 7.23733366e-01 -3.64208370e-01 -2.20712230e-01
5.80214620e-01 1.79234937e-01 7.10748374e-01 -9.45792556e-01
-1.64613187e-01 -8.58748972e-01 2.72950158e-02 -1.36364079e+00
2.76363909e-01 -1.01373827e+00 -1.09531730e-02 -1.87130618e+00
7.12762624e-02 3.68346274e-01 -9.74016413e-02 3.96036536e-01
-5.62534273e-01 1.19547009e-01 2.94872135e-01 3.12930375e-01
-7.35018790e-01 9.62850392e-01 9.45423305e-01 -2.41771460e-01
-4.71033275e-01 -1.84129521e-01 -1.06648552e+00 5.54217756e-01
1.16940713e+00 -3.21924746e-01 -4.60240453e-01 -7.43624866e-01
5.03196836e-01 -1.75753251e-01 -1.05023302e-01 -9.47867274e-01
2.78553098e-01 1.84864476e-01 2.85878599e-01 -1.01685929e+00
3.23842525e-01 -1.89530730e-01 -3.19586806e-02 3.02404344e-01
-7.08399653e-01 2.38251224e-01 2.08549738e-01 4.62512553e-01
-3.27972233e-01 -2.55230099e-01 3.29397231e-01 -1.09599940e-01
-3.26886535e-01 -1.35080829e-01 -7.35503376e-01 -7.54819289e-02
5.45895994e-01 -2.43565291e-01 -4.73889112e-01 -6.93861723e-01
-3.33919287e-01 1.69838235e-01 4.75967884e-01 4.61236656e-01
7.59234965e-01 -1.35497010e+00 -1.03800666e+00 3.16200823e-01
2.37785522e-02 -1.42241314e-01 1.89224467e-01 6.74809694e-01
-3.59052390e-01 5.75497866e-01 4.39595282e-02 -5.09469092e-01
-1.20725036e+00 1.41572461e-01 8.18224028e-02 -3.27702761e-01
-6.10717416e-01 7.40592420e-01 1.92303196e-01 -1.27461866e-01
2.61712164e-01 -4.77779925e-01 -2.98169643e-01 4.61827189e-01
6.57749295e-01 3.77967477e-01 -2.58228052e-02 -4.93223667e-01
5.55561036e-02 4.32624400e-01 -3.27911317e-01 -1.63104296e-01
1.25119436e+00 -1.95206106e-01 -3.09072286e-01 7.37043440e-01
1.12671101e+00 1.10070735e-01 -1.00756872e+00 -6.18365705e-02
-2.73480684e-01 -4.61534522e-02 -4.21229191e-02 -4.75915670e-01
-8.04162502e-01 7.73704767e-01 6.99087828e-02 6.25937819e-01
1.04007196e+00 -1.33387581e-01 1.05105531e+00 4.69751388e-01
-2.77603418e-01 -1.15252256e+00 2.27048635e-01 7.18268096e-01
1.06530380e+00 -9.32212055e-01 1.72801793e-01 1.80658832e-01
-8.82784545e-01 9.31006134e-01 3.64285648e-01 -1.11623310e-01
2.14962393e-01 5.29021118e-03 1.02315031e-01 -8.03000703e-02
-1.07285678e+00 8.64218026e-02 3.46704692e-01 7.40451887e-02
6.07197523e-01 1.48006484e-01 -5.78517020e-01 7.70981550e-01
-8.22358191e-01 -2.69672453e-01 7.55891025e-01 7.13536561e-01
-5.80588222e-01 -9.62279320e-01 -2.25805998e-01 6.70471609e-01
-5.95594704e-01 -4.09912676e-01 -6.10418499e-01 2.37520114e-01
-3.78703326e-01 9.82514322e-01 2.86963463e-01 -3.49078655e-01
1.13412112e-01 4.14529264e-01 1.03196740e-01 -5.64839065e-01
-4.80364591e-01 6.95017353e-02 -2.32235505e-03 -2.28909060e-01
-2.17120528e-01 -5.91707468e-01 -1.03356338e+00 -2.45803356e-01
-3.87819499e-01 4.29468513e-01 6.17829621e-01 7.49989152e-01
6.67469501e-01 8.42787802e-01 3.59501362e-01 -9.49518502e-01
-7.83444226e-01 -1.34712744e+00 -2.54276365e-01 4.15799588e-01
4.52658862e-01 -1.03029713e-01 -5.90818763e-01 2.32220352e-01] | [12.357246398925781, 9.34571647644043] |
fd68543e-2fe3-42d7-ab8d-725531d6254c | predict-and-use-latent-patterns-for-short | 2010.13982 | null | https://arxiv.org/abs/2010.13982v2 | https://arxiv.org/pdf/2010.13982v2.pdf | Predict and Use Latent Patterns for Short-Text Conversation | Many neural network models nowadays have achieved promising performances in Chit-chat settings. The majority of them rely on an encoder for understanding the post and a decoder for generating the response. Without given assigned semantics, the models lack the fine-grained control over responses as the semantic mapping between posts and responses is hidden on the fly within the end-to-end manners. Some previous works utilize sampled latent words as a controllable semantic form to drive the generated response around the work, but few works attempt to use more complex semantic patterns to guide the generation. In this paper, we propose to use more detailed semantic forms, including latent responses and part-of-speech sequences sampled from the corresponding distributions, as the controllable semantics to guide the generation. Our results show that the richer semantics are not only able to provide informative and diverse responses, but also increase the overall performance of response quality, including fluency and coherence. | ['Wei-Yun Ma', 'Ta-Hsuan Chao', 'Yu-Chieh Chao', 'Hung-Ting Chen'] | 2020-10-27 | null | null | null | null | ['short-text-conversation'] | ['natural-language-processing'] | [ 1.59970984e-01 2.17262387e-01 -2.41908237e-01 -9.50259149e-01
-8.52433562e-01 -4.76856053e-01 8.45459819e-01 -1.42641798e-01
-2.33659476e-01 7.39176571e-01 9.34994161e-01 2.96848387e-01
2.53688358e-02 -8.57810318e-01 -4.09719467e-01 -3.85303885e-01
6.45068228e-01 9.30726230e-01 3.52723040e-02 -5.35191357e-01
2.40835726e-01 -3.65049064e-01 -1.42717147e+00 8.18012595e-01
9.08851027e-01 6.38672054e-01 6.56213164e-01 7.26020992e-01
-5.17166972e-01 1.02038813e+00 -8.54739368e-01 -4.90250826e-01
-2.16451839e-01 -1.06754553e+00 -8.88107717e-01 -8.58533308e-02
-2.64012188e-01 -3.55884552e-01 -2.87083179e-01 7.63600051e-01
7.36846268e-01 4.55921263e-01 7.44113922e-01 -8.70563924e-01
-1.08356965e+00 1.40853071e+00 3.63180816e-01 -3.88905495e-01
7.45890915e-01 3.63675028e-01 1.04938841e+00 -6.01595521e-01
7.23124325e-01 1.49657071e+00 2.37437710e-01 1.10988915e+00
-1.18541622e+00 -3.81494403e-01 6.53161556e-02 1.87566340e-01
-8.49091411e-01 -4.51035380e-01 8.95036519e-01 -2.54008532e-01
7.39537358e-01 2.97478884e-01 3.87619466e-01 1.82250011e+00
-2.10131332e-01 9.69791591e-01 1.12057292e+00 -4.79121029e-01
3.91887099e-01 3.98830235e-01 1.69231117e-01 1.28345490e-01
-7.08670974e-01 -5.86915016e-02 -8.46122384e-01 7.01718032e-02
4.63548154e-01 1.08204596e-02 -2.79782176e-01 3.10076296e-01
-1.15058470e+00 1.20241666e+00 3.12432110e-01 1.90385118e-01
-6.64844692e-01 6.11112826e-02 3.99341583e-01 2.84394860e-01
7.01644123e-01 7.36932516e-01 -3.01362216e-01 -6.71860039e-01
-9.24379945e-01 5.23748279e-01 1.14951575e+00 1.11402559e+00
6.54796779e-01 -1.35392207e-03 -1.09099734e+00 1.09486234e+00
1.00440241e-01 3.67251843e-01 7.95412719e-01 -1.00799334e+00
6.20263636e-01 4.82059807e-01 3.10520947e-01 -8.24932754e-01
1.78912599e-02 -1.80786878e-01 -6.83803380e-01 -5.21657348e-01
3.22146118e-01 -4.64438200e-01 -8.73002648e-01 1.97151649e+00
7.19677135e-02 -2.68058598e-01 -1.27679497e-01 1.10639036e+00
8.56447816e-01 7.91505575e-01 1.10367648e-01 2.22663563e-02
1.03394127e+00 -1.22639191e+00 -1.08144343e+00 -3.08561623e-01
6.23205662e-01 -7.47033834e-01 1.71425152e+00 2.39184517e-02
-1.33522427e+00 -6.81031346e-01 -5.69754958e-01 -1.95395142e-01
-5.60103506e-02 2.34667078e-01 3.75169784e-01 1.37462407e-01
-1.21211278e+00 5.86684942e-01 -4.09546763e-01 -2.35699713e-01
1.62506998e-02 3.53982560e-02 -1.35130808e-02 -4.20329683e-02
-1.79268312e+00 7.46894896e-01 5.47887981e-02 -9.33480486e-02
-9.82973874e-01 -5.36839783e-01 -7.32428193e-01 2.78637290e-01
2.92422444e-01 -9.82587218e-01 1.73895931e+00 -1.19763744e+00
-2.24911141e+00 5.72819591e-01 -2.69438446e-01 -1.63397476e-01
5.19405425e-01 -1.71093717e-01 7.67222419e-02 4.25228029e-02
1.88275278e-01 9.76294994e-01 6.14568770e-01 -9.06337142e-01
-3.38503271e-01 -4.10108827e-02 1.96381032e-01 3.69312584e-01
-4.53534335e-01 1.69821247e-01 -2.43403792e-01 -6.36760831e-01
-3.52785110e-01 -9.77752507e-01 -2.17623293e-01 -6.80545270e-01
-7.16404378e-01 -7.58072138e-01 2.51421630e-01 -5.48822045e-01
1.30000710e+00 -1.87410891e+00 3.62073153e-01 -2.64917195e-01
-1.84428003e-02 3.14593650e-02 -3.76119047e-01 9.08582091e-01
3.77476066e-01 1.11469910e-01 2.14164063e-01 -7.38476634e-01
5.03093958e-01 9.44288522e-02 -5.81526935e-01 -1.16380498e-01
5.38951643e-02 1.03291011e+00 -8.85606706e-01 -5.28444396e-03
-3.50008011e-02 2.75433570e-01 -7.76322901e-01 8.23854387e-01
-6.73049450e-01 3.46390694e-01 -8.44556987e-01 -2.17677608e-01
6.80697784e-02 -3.58928055e-01 1.34948641e-01 3.81989926e-01
1.80879533e-01 1.11006105e+00 -6.64436698e-01 1.81162357e+00
-7.96862483e-01 2.70497829e-01 -3.54843251e-02 -6.35707378e-01
1.19596064e+00 5.27249396e-01 4.31710808e-03 -7.65208542e-01
2.78474875e-02 -5.43383323e-02 -1.46223649e-01 -6.39417887e-01
7.67844319e-01 -2.45785713e-01 -4.57946032e-01 8.81835878e-01
2.26118285e-02 -2.06195742e-01 2.25322768e-01 4.88451451e-01
8.51085305e-01 7.12557361e-02 -1.69026464e-01 5.32871448e-02
1.50247946e-01 5.25645055e-02 2.39263073e-01 9.90102708e-01
3.95320803e-01 7.88294077e-01 5.19525707e-01 1.08667230e-02
-1.08254910e+00 -8.08971584e-01 5.84197581e-01 1.39044476e+00
-1.11709461e-01 -4.59277391e-01 -1.09671664e+00 -4.22087222e-01
-4.06301051e-01 1.13934410e+00 -5.51527917e-01 -2.77188361e-01
-4.01040614e-01 -1.93463196e-03 5.17171204e-01 6.11464858e-01
2.74814397e-01 -1.63454723e+00 -2.50618964e-01 4.96946782e-01
-9.53243434e-01 -1.07253480e+00 -8.91741693e-01 -1.40430359e-03
-6.19877875e-01 -3.29875737e-01 -8.75814676e-01 -6.44255340e-01
4.63191986e-01 -3.68198641e-02 1.17964113e+00 -9.01861414e-02
2.29574293e-01 1.11985579e-01 -7.21524715e-01 -1.33298337e-01
-7.98419297e-01 3.83374751e-01 -7.01625422e-02 6.69928119e-02
4.29620326e-01 -5.56942344e-01 -4.85147566e-01 2.13785022e-01
-7.71261156e-01 4.50270116e-01 3.83130074e-01 8.64479065e-01
1.38380721e-01 -5.06680131e-01 8.55582178e-01 -1.05958068e+00
1.38106239e+00 -5.25388062e-01 1.40251637e-01 4.30047102e-02
-4.78309721e-01 2.90059060e-01 1.14796376e+00 -6.20904028e-01
-1.33957195e+00 -2.89771110e-01 -2.85079628e-01 -2.50138223e-01
-4.38762575e-01 5.45314312e-01 -5.69350319e-03 9.60822463e-01
7.34724641e-01 3.83101255e-01 4.05435078e-02 -6.11343980e-01
6.33130252e-01 1.03623080e+00 2.29302943e-01 -7.96311438e-01
3.44231516e-01 -1.60718113e-02 -6.72718227e-01 -3.02208662e-01
-1.00412548e+00 -3.88085127e-01 -1.40021101e-01 -1.66013435e-01
9.41228807e-01 -8.02717149e-01 -5.41238785e-01 2.72487402e-01
-1.46383798e+00 -9.21964407e-01 -2.72047758e-01 3.76119733e-01
-9.37805533e-01 -1.18725695e-01 -1.04209936e+00 -7.01957524e-01
-3.95535231e-01 -1.22318065e+00 1.07535267e+00 2.05710873e-01
-8.58263254e-01 -1.00422287e+00 3.88444141e-02 4.89979625e-01
7.75039732e-01 -3.12360585e-01 1.05176175e+00 -7.14690089e-01
-5.25819957e-01 -1.81867063e-01 1.09666556e-01 3.12177807e-01
1.30458832e-01 -1.52764738e-01 -7.92114079e-01 2.55518526e-01
7.61033595e-02 -7.92241633e-01 5.32816768e-01 3.77988994e-01
1.11988437e+00 -8.09318125e-01 1.43975383e-02 2.37484142e-01
8.48239839e-01 2.03782171e-02 6.65145278e-01 -5.71429729e-02
3.68189842e-01 1.19791365e+00 5.65326154e-01 6.38749897e-01
6.68183684e-01 9.18166757e-01 8.48527160e-03 1.75199389e-01
-2.05147654e-01 -9.83322918e-01 7.15319574e-01 1.15771806e+00
2.55099177e-01 -6.04953885e-01 -4.53047782e-01 4.66429263e-01
-2.03229022e+00 -1.20808589e+00 -1.63458154e-01 1.74590909e+00
1.48857880e+00 -2.09661007e-01 7.16600344e-02 -4.05075043e-01
7.37631440e-01 2.82657981e-01 -2.63325125e-01 -4.98904258e-01
1.30547255e-01 2.13129148e-01 -1.38828695e-01 6.61111951e-01
-2.43044376e-01 1.17534041e+00 6.25402784e+00 9.55296576e-01
-1.14169681e+00 2.34391958e-01 5.77283502e-01 -2.48265758e-01
-8.87938559e-01 9.22466349e-03 -8.27161849e-01 7.63611197e-01
1.19307077e+00 -1.75102323e-01 6.09950781e-01 7.43172109e-01
9.79893088e-01 1.24924362e-01 -1.24264097e+00 6.44258916e-01
-7.80917183e-02 -1.41535056e+00 1.80948347e-01 -2.76075900e-01
7.19138741e-01 -2.82784879e-01 3.16170678e-02 5.21055043e-01
7.71035671e-01 -1.08900547e+00 1.03667486e+00 7.73421884e-01
8.57015431e-01 -5.32774389e-01 5.49998343e-01 9.03627813e-01
-5.67609310e-01 1.93549275e-01 -2.48210073e-01 -4.08242106e-01
4.91063416e-01 4.39118952e-01 -1.04352999e+00 3.02110910e-01
2.35501409e-01 5.02399683e-01 -2.94525445e-01 3.68930310e-01
-8.41538787e-01 8.56001318e-01 -5.64463902e-03 -7.63689101e-01
4.06169713e-01 -1.39076978e-01 3.01496685e-01 1.18464661e+00
4.19924527e-01 -3.22596915e-02 4.50038999e-01 1.28108585e+00
-1.29053384e-01 7.35961273e-02 -3.83709788e-01 -3.83942306e-01
7.97970474e-01 9.43309426e-01 -2.83063471e-01 -5.02380192e-01
1.12711847e-01 9.82507229e-01 3.10842365e-01 6.08817279e-01
-8.05600882e-01 -4.07584012e-01 5.69581211e-01 3.75451207e-01
-3.35361175e-02 1.65856890e-02 -4.99336481e-01 -1.01126635e+00
-3.37978043e-02 -1.01648104e+00 -2.95606581e-03 -1.12111199e+00
-1.38066757e+00 5.38873851e-01 -2.70385277e-02 -7.92018294e-01
-9.61632609e-01 -1.20729424e-01 -6.67070150e-01 1.24588990e+00
-1.12826467e+00 -8.80034804e-01 -1.41588658e-01 5.29851735e-01
1.10462141e+00 -6.16142526e-02 1.02760911e+00 7.59119913e-02
-1.76627308e-01 6.32339299e-01 -1.95560977e-01 6.03087731e-02
9.56290364e-01 -1.17614508e+00 1.60868883e-01 4.58655179e-01
-6.04361333e-02 9.51934040e-01 9.17926788e-01 -5.59166908e-01
-1.03106844e+00 -9.18115020e-01 1.42115963e+00 -5.61082542e-01
3.58800590e-01 -5.29570103e-01 -6.43809676e-01 4.50769931e-01
5.76351464e-01 -8.55062902e-01 6.23141587e-01 3.31305087e-01
-8.73374641e-02 1.25613794e-01 -6.65895700e-01 8.53110433e-01
9.36677158e-01 -7.53701925e-01 -5.08809924e-01 3.72209221e-01
1.19581032e+00 -2.65325278e-01 -3.82119805e-01 -1.84288815e-01
1.97444499e-01 -9.19858456e-01 5.01066267e-01 -7.19271958e-01
1.15914500e+00 2.94158533e-02 6.44567609e-02 -1.85662639e+00
-3.89006197e-01 -8.12577605e-01 1.49589136e-01 1.50687540e+00
5.65888226e-01 -2.21846208e-01 6.71638846e-01 7.33509064e-01
-4.76343811e-01 -6.66486204e-01 -3.88844460e-01 -3.56574506e-01
-7.36492798e-02 -3.64269227e-01 7.37372518e-01 7.10414708e-01
1.73765555e-01 9.37391222e-01 -6.07453108e-01 -5.92859626e-01
9.14679393e-02 1.27389938e-01 8.49157751e-01 -8.71797502e-01
-6.31695926e-01 -4.89758283e-01 3.58711004e-01 -1.56480229e+00
3.21019709e-01 -9.60614324e-01 4.76647556e-01 -1.72820199e+00
2.19856068e-01 -3.64607424e-01 2.31154650e-01 3.55799496e-01
-3.83063346e-01 -2.39323571e-01 2.77625740e-01 1.27000809e-01
-6.05971038e-01 9.42233622e-01 1.46420932e+00 1.06628649e-01
-2.19971031e-01 3.03111732e-01 -9.83942270e-01 3.58865082e-01
7.30133891e-01 -6.57198310e-01 -8.10698032e-01 -6.69452071e-01
2.07485572e-01 4.61461037e-01 1.57249302e-01 -4.20776665e-01
3.06776851e-01 -4.11078066e-01 -9.34289917e-02 -2.83710152e-01
4.20899481e-01 -2.48138100e-01 6.25246838e-02 -2.07061172e-02
-1.35071814e+00 5.46881855e-02 -5.34870923e-01 5.48309267e-01
-4.53427017e-01 -4.14883047e-01 4.82790828e-01 -3.52858096e-01
-3.19040090e-01 2.91828603e-01 -5.09380698e-01 1.63037926e-01
5.27480245e-01 -1.74185470e-01 -1.29903048e-01 -1.15352619e+00
-7.99352586e-01 1.86363012e-01 2.81429738e-01 8.09419870e-01
5.75557530e-01 -1.40181649e+00 -7.38341451e-01 4.69981730e-02
4.23630849e-02 3.00087780e-02 4.66568410e-01 5.47669768e-01
-1.62705071e-02 5.48697531e-01 2.17281077e-02 -4.18202609e-01
-9.87453461e-01 9.45734903e-02 2.22296454e-02 -5.05223632e-01
-4.85636383e-01 9.77260113e-01 2.67478615e-01 -6.11223578e-01
3.48149478e-01 -1.15849100e-01 -1.65709421e-01 1.66464701e-01
5.90104282e-01 -1.53325275e-02 -2.24586532e-01 -1.19204290e-01
2.41830677e-01 -1.40121996e-01 -7.95691609e-02 -5.80928028e-01
1.34051824e+00 -2.44100079e-01 6.78843760e-04 6.17183924e-01
1.08701217e+00 -1.95337921e-01 -1.40587306e+00 -3.75102133e-01
-3.15977173e-04 -4.09117669e-01 -3.21989536e-01 -1.05813420e+00
-6.64372325e-01 1.11133158e+00 -3.27498257e-01 3.55021179e-01
7.06319749e-01 1.58733487e-01 1.01032543e+00 1.60734311e-01
4.06766862e-01 -1.46580303e+00 5.77901125e-01 8.41135621e-01
9.42943335e-01 -7.83890903e-01 -8.86765003e-01 -9.30985957e-02
-1.31317556e+00 1.09989321e+00 7.39112914e-01 1.67522132e-02
2.67671961e-02 1.48946539e-01 2.47568458e-01 7.97988996e-02
-1.34465969e+00 -1.39483035e-01 1.41580701e-02 5.79549313e-01
8.73061001e-01 2.92722553e-01 -3.00698489e-01 1.03559840e+00
-6.33445203e-01 1.77999541e-01 5.72065771e-01 3.07023495e-01
-4.38391566e-01 -1.36773551e+00 1.43612191e-01 3.39181572e-01
-3.01265478e-01 -1.69729248e-01 -7.39769995e-01 1.65337846e-01
-2.89228648e-01 1.21647191e+00 -4.05917540e-02 -4.90403265e-01
4.72793818e-01 4.16946411e-01 1.91901594e-01 -1.20394289e+00
-7.91738272e-01 -8.47933814e-02 4.64484215e-01 -5.37094295e-01
-1.36754259e-01 -3.25774908e-01 -1.23268473e+00 -4.98889148e-01
-6.57048076e-02 4.63307142e-01 5.23678720e-01 1.10919344e+00
5.00686526e-01 7.17238367e-01 9.32201743e-01 -5.69110453e-01
-1.30638337e+00 -1.51270938e+00 -4.02864665e-01 8.54119360e-01
-7.38610979e-03 -2.49361992e-01 -2.44481459e-01 1.01353914e-01] | [12.583917617797852, 8.320448875427246] |
00da4a23-a425-44d4-a379-8bbded32d619 | reinforced-multi-task-approach-for-multi-hop | 2004.02143 | null | https://arxiv.org/abs/2004.02143v4 | https://arxiv.org/pdf/2004.02143v4.pdf | Reinforced Multi-task Approach for Multi-hop Question Generation | Question generation (QG) attempts to solve the inverse of question answering (QA) problem by generating a natural language question given a document and an answer. While sequence to sequence neural models surpass rule-based systems for QG, they are limited in their capacity to focus on more than one supporting fact. For QG, we often require multiple supporting facts to generate high-quality questions. Inspired by recent works on multi-hop reasoning in QA, we take up Multi-hop question generation, which aims at generating relevant questions based on supporting facts in the context. We employ multitask learning with the auxiliary task of answer-aware supporting fact prediction to guide the question generator. In addition, we also proposed a question-aware reward function in a Reinforcement Learning (RL) framework to maximize the utilization of the supporting facts. We demonstrate the effectiveness of our approach through experiments on the multi-hop question answering dataset, HotPotQA. Empirical evaluation shows our model to outperform the single-hop neural question generation models on both automatic evaluation metrics such as BLEU, METEOR, and ROUGE, and human evaluation metrics for quality and coverage of the generated questions. | ['Akella Ravi Tej', 'Pushpak Bhattacharyya', 'Hardik Chauhan', 'Deepak Gupta', 'Asif Ekbal'] | 2020-04-05 | null | https://aclanthology.org/2020.coling-main.249 | https://aclanthology.org/2020.coling-main.249.pdf | coling-2020-8 | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 1.87275782e-01 6.98640525e-01 1.97849512e-01 -3.94442886e-01
-1.63721490e+00 -7.14500308e-01 7.93932080e-01 2.07478046e-01
-2.63097793e-01 1.18528533e+00 4.64199275e-01 -5.82016170e-01
-2.24726692e-01 -1.10571957e+00 -7.48674631e-01 1.02299280e-01
3.00413579e-01 8.85182381e-01 4.33345050e-01 -7.41418839e-01
5.23889244e-01 -3.57460767e-01 -1.49182665e+00 7.37871647e-01
1.51510620e+00 9.03586864e-01 3.70874345e-01 9.92552996e-01
-7.50949264e-01 1.48028910e+00 -9.59689558e-01 -6.68581486e-01
-7.90318176e-02 -1.00940454e+00 -1.44016433e+00 -1.38651937e-01
4.58252698e-01 -2.90337801e-01 -7.97599554e-03 6.13599122e-01
4.54563171e-01 2.87849784e-01 4.87794936e-01 -1.21368849e+00
-9.72225189e-01 6.35835171e-01 9.96589884e-02 3.40699881e-01
9.32750404e-01 3.15740287e-01 1.48767853e+00 -7.22506404e-01
5.76979339e-01 1.20474124e+00 2.08061084e-01 8.15345705e-01
-6.96959555e-01 -1.14684906e-02 3.79584059e-02 5.00980258e-01
-6.56804144e-01 -2.84353971e-01 4.70197618e-01 -9.19461250e-02
1.12007248e+00 3.02787870e-01 2.76494354e-01 7.59770870e-01
1.78217113e-01 8.69130075e-01 1.03868151e+00 -7.52930105e-01
3.84676158e-01 -7.15105906e-02 1.12172529e-01 8.54024589e-01
-1.55415773e-01 -1.90300912e-01 -5.85461676e-01 -4.00562704e-01
3.52525532e-01 -4.97785896e-01 -1.84349582e-01 1.98819503e-01
-1.13001037e+00 1.25924194e+00 3.31743568e-01 1.76155850e-01
-6.44661129e-01 1.01569235e-01 1.62366629e-01 4.89477605e-01
1.97838321e-01 1.21612668e+00 -4.67616230e-01 -9.84125659e-02
-7.58116722e-01 7.90020883e-01 1.24229836e+00 9.31719065e-01
6.69039428e-01 -1.65430516e-01 -1.08665049e+00 7.62968719e-01
6.02830313e-02 3.86806250e-01 6.65520608e-01 -1.38287091e+00
8.81630063e-01 7.27181911e-01 5.21984398e-01 -5.40698588e-01
-1.34475499e-01 -1.73917875e-01 -3.72857779e-01 -3.32115978e-01
4.70488071e-01 -3.63226056e-01 -7.44693100e-01 1.65271413e+00
2.78052211e-01 -5.69434687e-02 3.77493382e-01 7.26263106e-01
1.08317602e+00 9.34057117e-01 3.68465707e-02 -3.33758257e-02
1.55167556e+00 -1.32887423e+00 -8.02193820e-01 -2.45878920e-01
5.60871303e-01 -4.52804804e-01 1.40768683e+00 2.09352210e-01
-1.33807790e+00 -5.13876081e-01 -6.43807173e-01 -2.31821880e-01
-1.72842950e-01 1.05099045e-02 3.00961554e-01 4.62972909e-01
-1.05783713e+00 1.88061535e-01 3.36385332e-02 -1.42862678e-01
1.54182330e-01 -8.42160285e-02 2.21834630e-01 -3.49612743e-01
-1.67859638e+00 9.44011807e-01 4.75352436e-01 -1.77983165e-01
-9.54949021e-01 -6.53576791e-01 -7.03775823e-01 2.01858565e-01
8.69008183e-01 -1.27696192e+00 1.77865350e+00 -5.49483120e-01
-1.55495894e+00 3.62927407e-01 -3.04880291e-01 -7.72259951e-01
2.29102910e-01 -1.90288261e-01 -3.39021683e-01 6.25824630e-01
3.93251300e-01 9.15445030e-01 6.00086451e-01 -1.15212536e+00
-9.46110308e-01 1.09830834e-01 6.72411323e-01 3.63038808e-01
-1.44030182e-02 -2.68619388e-01 -9.04545039e-02 -3.19288880e-01
-2.81469584e-01 -4.17888343e-01 -3.18470180e-01 -5.50633907e-01
-3.98411900e-01 -7.53977418e-01 4.35399026e-01 -9.41197813e-01
1.15936565e+00 -1.39253664e+00 -2.43104696e-01 -1.55695096e-01
1.36239871e-01 2.76738644e-01 -6.68790638e-01 7.38652527e-01
5.24364293e-01 4.76833433e-02 -3.50063026e-01 1.92267716e-01
1.85022771e-01 3.14698309e-01 -7.50749648e-01 -6.66846454e-01
5.59286118e-01 1.40507555e+00 -1.25196755e+00 -6.06763303e-01
-3.81013274e-01 -2.06460148e-01 -7.47922957e-01 6.66859090e-01
-1.23090577e+00 1.29764646e-01 -6.11963511e-01 3.40725899e-01
2.58354992e-02 -5.89345634e-01 -7.76300058e-02 3.03001255e-01
4.72448200e-01 7.98424840e-01 -6.20986402e-01 1.61568522e+00
-7.95806646e-01 4.64589559e-02 -3.17606717e-01 -6.68015361e-01
1.14527023e+00 5.69267213e-01 -7.56286010e-02 -1.07460439e+00
-4.19842571e-01 2.26211205e-01 1.42962579e-02 -8.70653152e-01
1.00319743e+00 -1.67641044e-01 -1.70182556e-01 8.26393485e-01
1.98830843e-01 -5.39580345e-01 6.85599148e-01 6.15757823e-01
1.33758092e+00 1.86983183e-01 2.55102515e-01 5.27872257e-02
7.52335191e-01 2.75662154e-01 4.81823757e-02 1.18425703e+00
2.36186329e-02 5.14924169e-01 6.12290025e-01 -1.89765379e-01
-8.37309062e-01 -8.96604955e-01 6.29133642e-01 1.27866006e+00
-1.10808589e-01 -2.76111692e-01 -8.60632300e-01 -1.08027363e+00
-1.87973097e-01 1.33780205e+00 -5.10765851e-01 -1.14091441e-01
-6.73530996e-01 -3.10609758e-01 5.81057847e-01 2.61758983e-01
5.64506233e-01 -1.67226791e+00 -7.09725857e-01 6.55956089e-01
-9.92934227e-01 -1.19866025e+00 -5.81867278e-01 -1.75169632e-01
-6.82759404e-01 -1.20733273e+00 -6.77598417e-01 -6.75116301e-01
2.97260255e-01 1.32645458e-01 1.67631507e+00 2.21507132e-01
1.19812183e-01 6.08613789e-01 -7.24219561e-01 -3.72736007e-01
-6.66768730e-01 3.84866387e-01 -6.60741806e-01 -2.82455325e-01
2.33501434e-01 -1.90742791e-01 -6.84922755e-01 2.16890529e-01
-1.20217550e+00 -3.17456350e-02 5.50798297e-01 1.04048944e+00
4.25921559e-01 -3.69526058e-01 1.58085930e+00 -7.97110975e-01
1.49588990e+00 -7.56241679e-01 -2.95948625e-01 8.44365597e-01
-5.94917774e-01 3.77919853e-01 8.87499511e-01 -6.70400485e-02
-1.27561760e+00 -5.51720202e-01 -3.86405557e-01 2.31748879e-01
-1.30239045e-02 6.01637959e-01 -3.88128273e-02 4.39232349e-01
9.42766249e-01 4.93884832e-01 -2.53205091e-01 -2.15009209e-02
7.92728961e-01 4.10325348e-01 5.15495420e-01 -8.85824680e-01
5.75960636e-01 -5.17400093e-02 -3.61936718e-01 -3.25258791e-01
-1.45784855e+00 -3.36956382e-01 -8.26585963e-02 -2.22087696e-01
8.22717309e-01 -5.62264621e-01 -7.18159318e-01 -2.31815383e-01
-1.35674071e+00 -3.19284409e-01 -6.96141183e-01 -1.60845205e-01
-9.27587926e-01 3.78915548e-01 -4.21911240e-01 -9.21731472e-01
-7.38060236e-01 -7.00440228e-01 9.63190198e-01 3.18816513e-01
-2.59524226e-01 -9.41589534e-01 1.90995902e-01 8.95170748e-01
5.33026934e-01 4.46427427e-02 1.32666171e+00 -8.99928689e-01
-7.36750960e-01 2.25012731e-02 -8.07293802e-02 1.79005101e-01
4.60644066e-02 -5.78928232e-01 -6.03836119e-01 1.85787156e-01
6.18895739e-02 -9.25399661e-01 7.93436229e-01 1.96813103e-02
1.08369553e+00 -7.64290690e-01 2.62899518e-01 -2.46584043e-01
1.27957904e+00 1.61988422e-01 6.00204945e-01 2.88932323e-01
2.71359712e-01 1.01693642e+00 8.78299236e-01 2.78031766e-01
9.83723104e-01 4.37058181e-01 4.48999435e-01 4.39563811e-01
-1.10818721e-01 -7.33108640e-01 8.08529109e-02 5.46900630e-01
3.62234712e-01 -5.13817251e-01 -9.12092865e-01 8.75415564e-01
-1.82637680e+00 -1.08192039e+00 -1.55654669e-01 1.80323672e+00
1.06900203e+00 1.22209247e-02 2.35448256e-01 -9.43865478e-02
5.05367398e-01 -8.55523814e-03 -4.80641454e-01 -4.23676074e-01
-1.48643658e-01 3.52355242e-01 -2.07170665e-01 6.99869990e-01
-4.13438767e-01 9.93277729e-01 6.02172232e+00 7.20918536e-01
-3.62476081e-01 1.10165223e-01 6.48679912e-01 -2.68581491e-02
-9.36299622e-01 5.25114089e-02 -7.93798864e-01 3.25207502e-01
1.14174581e+00 -4.77565378e-01 4.11477894e-01 5.34820199e-01
1.23943910e-02 -2.74164528e-01 -9.13369477e-01 2.95767516e-01
2.00481594e-01 -1.54521763e+00 6.18831635e-01 -2.50365376e-01
8.58488798e-01 -2.96915889e-01 -2.28577718e-01 7.60061920e-01
6.54033899e-01 -9.90848422e-01 6.50003731e-01 7.28551209e-01
3.97935003e-01 -7.63809443e-01 6.95157707e-01 8.55073810e-01
-6.87722325e-01 -3.42613250e-01 -3.93175840e-01 -9.51554102e-04
5.23925722e-01 5.25534093e-01 -1.25916231e+00 8.82479429e-01
2.79916912e-01 -2.21894801e-01 -6.37357533e-01 8.79697502e-01
-6.61269307e-01 7.99818218e-01 8.35570619e-02 -5.43065488e-01
5.93450069e-01 6.34525567e-02 2.05611691e-01 8.92247915e-01
3.73567760e-01 4.12444085e-01 1.56405509e-01 8.94595444e-01
-3.50822687e-01 1.79200649e-01 -3.48451823e-01 4.87776510e-02
5.14550805e-01 1.11888862e+00 -1.85646087e-01 -6.25216246e-01
-2.55280018e-01 8.51029754e-01 4.65946287e-01 3.11958045e-01
-5.41738927e-01 -4.41234559e-01 -1.68642074e-01 1.97883951e-03
3.90288204e-01 4.40986305e-02 -2.16623172e-01 -1.01604378e+00
2.92767882e-01 -1.24118161e+00 6.66742027e-01 -1.04032004e+00
-1.33532953e+00 8.90260875e-01 -1.38104290e-01 -7.61976659e-01
-1.00031757e+00 -1.07893668e-01 -6.72760725e-01 1.11034548e+00
-1.99041593e+00 -9.04147625e-01 -1.83401257e-01 5.38794041e-01
7.18199134e-01 -1.90658405e-01 7.96246111e-01 -2.29520097e-01
4.10737190e-03 4.22202975e-01 -3.33198190e-01 -1.46765009e-01
5.34425914e-01 -1.58288085e+00 7.07274318e-01 6.19137585e-01
3.09849173e-01 3.60174209e-01 6.27778232e-01 -6.34020090e-01
-1.19556952e+00 -1.14317656e+00 1.33562660e+00 -5.78607678e-01
5.53260267e-01 -6.11820370e-02 -1.05223978e+00 2.92758524e-01
5.45267105e-01 -4.14955825e-01 7.27472126e-01 -7.89902657e-02
-2.91706771e-01 1.83242157e-01 -1.24272847e+00 5.27408481e-01
7.39090264e-01 -4.41955566e-01 -1.15328860e+00 5.84898055e-01
1.08207095e+00 -1.59948707e-01 -6.69296086e-01 1.27942145e-01
7.37625286e-02 -8.42628837e-01 6.77995324e-01 -8.65861773e-01
9.74870265e-01 -2.92040467e-01 -1.23961464e-01 -1.37293887e+00
-2.00863749e-01 -6.49306297e-01 -5.52118957e-01 1.11401677e+00
7.39213169e-01 -3.69329572e-01 7.52923965e-01 3.86872560e-01
-1.97726235e-01 -9.88362074e-01 -9.45514739e-01 -6.92048669e-01
1.93383053e-01 -5.05174920e-02 1.02404666e+00 4.01945710e-01
-6.91095293e-02 8.72900724e-01 -3.06256860e-01 -4.50534001e-02
2.96479821e-01 3.96423221e-01 6.46418214e-01 -8.83508801e-01
-6.27489150e-01 -1.95621610e-01 6.06788218e-01 -1.32869160e+00
1.28188044e-01 -7.99820900e-01 4.63651657e-01 -2.21246457e+00
-3.81347686e-02 -1.85214967e-01 7.69351646e-02 2.56691992e-01
-8.88626575e-01 -1.80851266e-01 2.57904440e-01 -1.02578729e-01
-1.03477919e+00 7.81634808e-01 1.65985394e+00 -2.37658322e-02
3.16114388e-02 -2.69542504e-02 -1.16635275e+00 1.59149334e-01
7.83974051e-01 -2.60446966e-01 -6.69884086e-01 -5.43211699e-01
6.09479308e-01 8.52442145e-01 2.81644881e-01 -7.42987514e-01
3.27350557e-01 -2.27792665e-01 -5.85619882e-02 -5.71851730e-01
1.32036939e-01 -2.91110963e-01 -7.07300127e-01 3.88988853e-01
-9.08806741e-01 2.40464970e-01 -1.37960389e-01 4.63993549e-01
-4.85098153e-01 -6.00759506e-01 3.41523945e-01 -5.20274878e-01
-5.15982985e-01 1.63001359e-01 -3.29002678e-01 7.94491112e-01
6.54716790e-01 1.43109009e-01 -7.02360511e-01 -1.01825202e+00
-3.01561952e-01 8.14778984e-01 -2.62414843e-01 4.75549012e-01
8.55604827e-01 -1.12228501e+00 -1.22447288e+00 -4.29729283e-01
4.17231381e-01 -1.62192024e-02 2.45864630e-01 9.98508707e-02
-2.14821354e-01 9.01384890e-01 -2.28207707e-02 -9.36833471e-02
-6.54523015e-01 3.98393422e-01 4.06926632e-01 -1.04297745e+00
-1.47202700e-01 6.90793276e-01 -1.37942940e-01 -7.56179988e-01
-7.32437596e-02 -2.47363165e-01 -4.56948131e-01 -4.16541584e-02
7.13583767e-01 3.44177842e-01 1.94266677e-01 1.33887053e-01
2.21420690e-01 -1.54363275e-01 -1.93763226e-01 -5.34010947e-01
9.38091338e-01 -2.15591602e-02 3.36817428e-02 1.56251252e-01
5.24312854e-01 -1.76406637e-01 -1.01781261e+00 -3.91886711e-01
4.03905123e-01 -1.47401839e-01 -5.13898253e-01 -1.48504210e+00
-2.22808480e-01 6.86441958e-01 -1.01961546e-01 7.07869232e-01
1.04000604e+00 1.24205597e-01 1.25951099e+00 9.00112271e-01
3.47761422e-01 -9.81674492e-01 7.15426207e-01 9.38715100e-01
1.20937371e+00 -1.11065197e+00 -5.37249207e-01 -8.55528563e-02
-7.71452546e-01 8.12707722e-01 9.71465170e-01 -2.74027381e-02
-1.19013637e-01 -2.97607094e-01 2.36676112e-01 -2.84002542e-01
-1.29972947e+00 -3.22350413e-01 2.69034445e-01 6.08722270e-01
3.37274373e-01 -1.73291340e-01 -3.75076890e-01 4.61072803e-01
-6.83544874e-01 1.40518725e-01 7.02528536e-01 9.15091515e-01
-8.88376832e-01 -1.20682061e+00 -3.88063222e-01 8.49983156e-01
-4.19608355e-01 -2.82056808e-01 -4.39952165e-01 3.28154385e-01
-2.68288910e-01 1.49806833e+00 -2.29197457e-01 -2.53030169e-03
5.20303011e-01 4.15806204e-01 5.12500942e-01 -9.23865020e-01
-9.23307419e-01 -6.86736822e-01 5.26322484e-01 -2.66671509e-01
-1.40168443e-01 -3.53150010e-01 -1.10099864e+00 1.06825806e-01
-2.55365252e-01 8.73448610e-01 4.33704019e-01 1.24292362e+00
4.01450127e-01 5.83067238e-01 6.57771230e-01 1.78799972e-01
-9.56083000e-01 -1.21277797e+00 -1.40336260e-01 2.82683730e-01
2.82234281e-01 4.35948698e-03 -2.81620417e-02 -1.36926472e-01] | [11.4198637008667, 8.12186336517334] |
58c740d6-be0b-44cb-9809-2a247a2e7fca | real-world-super-resolution-via-kernel | null | null | https://ieeexplore.ieee.org/document/9150628 | https://ieeexplore.ieee.org/document/9150628 | Real-World Super-Resolution via Kernel Estimation and Noise Injection | Recent state-of-the-art super-resolution methods have achieved impressive performance on ideal datasets regardless of blur and noise. However, these methods always fail in real-world image super-resolution, since most of them adopt simple bicubic downsampling from high-quality images to construct Low-Resolution (LR) and High-Resolution (HR) pairs for training which may lose track of frequency-related details. To address this issue, we focus on designing a novel degradation framework for real- world images by estimating various blur kernels as well as real noise distributions. Based on our novel degradation framework, we can acquire LR images sharing a common domain with real-world images. Then, we propose a real- world super-resolution model aiming at better perception. Extensive experiments on synthetic noise data and real- world images demonstrate that our method outperforms the state-of-the-art methods, resulting in lower noise and better visual quality. In addition, our method is the winner of NTIRE 2020 Challenge on both tracks of Real-World Super-Resolution, which significantly outperforms other competitors by large margins. | ['Feiyue Huang', 'Jilin Li', 'Chengjie Wang', 'Ying Tai', 'Yun Cao', 'Xiaozhong Ji'] | 2020-06-19 | null | null | null | cvprw-2020-6 | ['video-super-resolution'] | ['computer-vision'] | [ 2.50955194e-01 -5.79617560e-01 -5.06802369e-03 -1.50536299e-01
-9.71706808e-01 -1.52913198e-01 4.19128865e-01 -8.84857178e-01
-8.46778303e-02 1.03725731e+00 5.88395059e-01 3.29244137e-01
-7.92757422e-02 -7.01769769e-01 -6.53699994e-01 -6.91288292e-01
1.87513426e-01 -1.02019705e-01 5.69846809e-01 -5.17974913e-01
1.50195837e-01 2.03273416e-01 -1.74237096e+00 5.59311748e-01
1.17989290e+00 7.68313706e-01 6.83367074e-01 4.87004012e-01
3.54384124e-01 7.59290874e-01 -4.24971670e-01 -1.77323818e-01
4.07571942e-01 -3.56938988e-01 -5.73042750e-01 3.54115739e-02
7.25884020e-01 -6.36288643e-01 -6.71617508e-01 1.49232447e+00
6.43522263e-01 7.23683611e-02 1.47711724e-01 -4.81125027e-01
-1.33103955e+00 4.21147883e-01 -9.31484044e-01 6.62178040e-01
4.22569931e-01 3.44684534e-02 5.79529881e-01 -1.09851909e+00
5.28720975e-01 1.45101583e+00 5.94423592e-01 5.03046691e-01
-1.48933196e+00 -8.20396304e-01 -1.02146305e-01 5.32634795e-01
-1.51657200e+00 -5.75630307e-01 5.42627394e-01 1.03706136e-01
2.40064129e-01 3.34692806e-01 1.75748184e-01 1.23879921e+00
-8.35416391e-02 3.95422727e-01 1.69247460e+00 -1.19126715e-01
1.23860454e-02 -1.10732587e-02 -2.53488213e-01 1.98499709e-01
3.86663139e-01 4.18053836e-01 -7.36661494e-01 1.41504973e-01
1.46331406e+00 -1.78963646e-01 -8.82263184e-01 -1.43921271e-01
-1.34809792e+00 2.06778049e-01 5.46552241e-01 4.14700300e-01
-2.92638272e-01 -2.11780816e-01 4.37615924e-02 1.42630890e-01
6.12779915e-01 1.71560183e-01 -6.56788722e-02 1.56802654e-01
-8.75505567e-01 1.70346230e-01 9.95426774e-02 8.91701043e-01
5.88800013e-01 9.28175449e-02 -3.99666578e-01 1.29108560e+00
-2.13061109e-01 4.93462890e-01 5.86545527e-01 -1.19656956e+00
4.08089697e-01 -1.23385854e-01 7.38753319e-01 -7.90083468e-01
-7.51802549e-02 -7.64583647e-01 -1.48466468e+00 1.68085113e-01
2.44110614e-01 2.96084017e-01 -7.73537278e-01 1.59296060e+00
2.01813564e-01 8.03976893e-01 6.79475591e-02 1.44313467e+00
7.12001681e-01 6.15949094e-01 -3.72899562e-01 -5.85607827e-01
1.46345961e+00 -9.74367619e-01 -1.00859952e+00 -1.33410722e-01
-4.98508871e-01 -8.24837565e-01 1.09324610e+00 6.25617146e-01
-1.27273548e+00 -1.18484330e+00 -1.05396044e+00 -2.38986388e-01
2.14058205e-01 1.63505822e-01 3.63474220e-01 4.82144147e-01
-1.12144482e+00 6.86761975e-01 -4.39144671e-01 -1.08766600e-01
3.91651958e-01 -2.51058966e-01 -2.86709577e-01 -5.36863208e-01
-1.47237682e+00 9.81487751e-01 1.19927369e-01 -6.29866077e-03
-8.99650276e-01 -8.83989871e-01 -6.89419746e-01 -8.34786668e-02
3.72190565e-01 -7.63681889e-01 9.72198963e-01 -5.71560442e-01
-1.42114151e+00 6.67410254e-01 -2.65951663e-01 -4.17694360e-01
5.31841397e-01 -5.07616997e-01 -1.03238344e+00 1.52887747e-01
1.44853145e-01 2.12988153e-01 1.25627053e+00 -1.71333516e+00
-7.41168022e-01 -2.94128269e-01 -8.17395523e-02 3.36227417e-01
-2.17148304e-01 -2.82663591e-02 -4.32236224e-01 -8.76318812e-01
2.09064752e-01 -3.49353582e-01 -1.44310489e-01 -1.06521025e-01
-5.14692403e-02 3.71737301e-01 7.99669623e-01 -8.31336439e-01
1.29733741e+00 -2.22699189e+00 5.33232167e-02 -5.45588851e-01
4.13898051e-01 4.51640457e-01 -1.72597095e-01 -1.29411235e-01
-1.20302506e-01 -5.59458807e-02 -6.18474744e-02 -2.05535203e-01
-4.99853939e-01 3.39641534e-02 -5.83416283e-01 5.06197929e-01
4.46961261e-02 5.64752877e-01 -1.13045096e+00 -4.76372242e-01
3.55328768e-01 9.87732053e-01 -1.19822182e-01 2.37171337e-01
1.64265737e-01 8.51428509e-01 -2.53220350e-01 3.51613522e-01
1.32446611e+00 -4.25715923e-01 4.31721434e-02 -6.47724092e-01
-2.43502259e-01 5.55177666e-02 -1.32430649e+00 1.81363308e+00
-5.74459195e-01 3.77665997e-01 2.27099165e-01 -5.24289250e-01
9.05708849e-01 1.47836074e-01 1.73009321e-01 -1.06712997e+00
-4.38155174e-01 3.17627132e-01 -3.91143113e-01 -4.75577801e-01
7.02487528e-01 -2.24839449e-01 4.70136702e-01 7.23971501e-02
-2.40871355e-01 -7.17789754e-02 -3.08507290e-02 -3.49386372e-02
7.66533494e-01 1.90837115e-01 2.45816812e-01 -1.70703068e-01
6.83007240e-01 -5.90165675e-01 6.38732433e-01 7.03108728e-01
-2.36871049e-01 1.31967008e+00 -1.29579306e-01 -3.60973835e-01
-1.43899047e+00 -1.42622209e+00 -3.98618788e-01 7.48578012e-01
7.72509575e-01 -6.40005767e-02 -7.75059581e-01 -6.50613159e-02
-4.11425918e-01 4.71172541e-01 -5.27170241e-01 1.47561029e-01
-6.53585553e-01 -8.76514137e-01 1.98780462e-01 2.45528132e-01
1.18834889e+00 -6.36424959e-01 -3.10859233e-01 7.14655444e-02
-7.62124896e-01 -1.62218153e+00 -6.47317410e-01 -5.48478365e-01
-7.26130903e-01 -8.24926317e-01 -1.24724638e+00 -5.91208577e-01
3.34169060e-01 1.05692005e+00 1.21076286e+00 -2.25020036e-01
-2.92903990e-01 -1.57204047e-01 -2.96217650e-01 2.30586633e-01
-2.96240747e-01 -4.20200139e-01 2.19359159e-01 3.49658281e-01
-9.14556831e-02 -8.28350484e-01 -1.00615835e+00 6.50934279e-01
-9.84491467e-01 3.02852631e-01 7.34396398e-01 9.67148483e-01
8.39223444e-01 6.61237717e-01 5.33024728e-01 -5.63781619e-01
6.32515788e-01 -3.17397356e-01 -5.76594353e-01 1.81586117e-01
-6.75184071e-01 -3.98423597e-02 8.92957807e-01 -5.11955380e-01
-1.71976030e+00 -4.02527899e-01 2.94488698e-01 -7.27412879e-01
-7.79715553e-02 -7.30556399e-02 -1.76674396e-01 -1.05138652e-01
9.38125610e-01 6.16169512e-01 -2.70049483e-01 -8.76394808e-01
3.86430264e-01 7.00251579e-01 1.11159253e+00 -3.25249940e-01
1.15921509e+00 1.00196254e+00 -7.12272301e-02 -6.91849589e-01
-1.26259875e+00 -4.72022563e-01 -4.21523511e-01 -1.82908438e-02
6.51079059e-01 -1.46541655e+00 -2.62865484e-01 5.68831921e-01
-1.00475609e+00 -1.04574009e-03 -2.00725794e-01 4.76127416e-01
-5.81370473e-01 6.01863205e-01 -7.32770801e-01 -7.13164210e-01
-1.59561321e-01 -1.04945087e+00 1.10961306e+00 5.30162275e-01
5.41484356e-01 -4.22020435e-01 -2.45976988e-02 5.15704274e-01
1.02918613e+00 -1.60654500e-01 3.12989503e-01 4.18148279e-01
-8.71032715e-01 4.36275005e-01 -9.79649425e-01 5.53420484e-01
3.82170260e-01 -5.76059937e-01 -1.01237035e+00 -4.46094185e-01
3.57328922e-01 -1.32018879e-01 1.11086881e+00 3.83811891e-01
1.32848227e+00 -1.05922759e-01 -6.55830726e-02 9.79502976e-01
1.77357852e+00 -3.04554701e-01 1.22362924e+00 3.42576772e-01
5.54470718e-01 2.70070642e-01 8.46643031e-01 3.77634555e-01
2.79146075e-01 9.95457351e-01 4.04965311e-01 -2.74936050e-01
-7.79144585e-01 -3.21404189e-01 2.49778301e-01 4.35378522e-01
-4.02390331e-01 1.80493012e-01 -2.89450705e-01 5.47166944e-01
-1.77244675e+00 -1.29359460e+00 -8.08033422e-02 2.50095057e+00
1.10701251e+00 8.44789392e-05 4.28497754e-02 -1.81102499e-01
9.97814178e-01 4.75880235e-01 -7.04542696e-01 4.00689363e-01
-6.58546984e-01 1.93573609e-01 5.38698375e-01 3.65406454e-01
-1.06958032e+00 7.93427050e-01 6.37015533e+00 1.22711313e+00
-9.51016903e-01 3.32873493e-01 8.20953250e-01 -1.06482729e-02
-2.54004151e-01 -3.18094105e-01 -7.30219960e-01 6.32910252e-01
6.90429568e-01 -3.28736931e-01 1.03649735e+00 4.78180200e-01
4.17772382e-01 -9.89922043e-03 -5.94219506e-01 1.41919899e+00
1.45297021e-01 -1.26190734e+00 2.43770890e-02 -1.74619272e-01
1.08002532e+00 -3.50035802e-02 3.55728000e-01 1.90358441e-02
2.75742024e-01 -1.28926790e+00 4.65118349e-01 7.68399775e-01
1.38182509e+00 -5.19722998e-01 6.11546814e-01 2.29684561e-01
-1.30436385e+00 -5.39256670e-02 -7.70316601e-01 1.22924007e-01
2.37341985e-01 8.61447632e-01 8.82024691e-02 1.00649416e+00
1.16200840e+00 8.14054370e-01 -4.72431391e-01 1.03230512e+00
-1.49831504e-01 2.60597199e-01 6.62181601e-02 8.15189660e-01
-3.81803364e-01 -2.61473566e-01 6.58307195e-01 1.04137838e+00
4.76590693e-01 4.62113827e-01 -1.33606553e-01 1.10354674e+00
-7.01588392e-02 -3.66721362e-01 -2.23582789e-01 5.38651109e-01
5.63143432e-01 1.22953963e+00 -1.82163984e-01 -2.73577660e-01
-5.07776380e-01 1.23601472e+00 1.31809130e-01 6.59337580e-01
-1.01681972e+00 -6.54114559e-02 8.12726438e-01 2.81386793e-01
4.41007614e-01 -5.09855151e-02 -2.55397826e-01 -1.58299124e+00
1.42415449e-01 -1.12710690e+00 1.21077895e-01 -1.25905168e+00
-1.50223029e+00 8.21547985e-01 -1.57958910e-01 -1.77401626e+00
2.29635477e-01 -2.45660275e-01 -1.76194340e-01 1.14621794e+00
-1.94216597e+00 -9.97245848e-01 -7.46196628e-01 7.38819301e-01
7.22944260e-01 5.42351380e-02 3.93653333e-01 4.66799200e-01
-2.96741337e-01 3.06478053e-01 4.12056446e-01 -2.15909854e-01
1.11429691e+00 -9.71028626e-01 4.62430805e-01 1.11370885e+00
-8.31813812e-02 4.61638898e-01 9.34345603e-01 -5.44456124e-01
-1.10725701e+00 -1.15060079e+00 2.48654127e-01 -2.02363148e-01
5.31804621e-01 -1.63893271e-02 -1.19147265e+00 2.52996504e-01
9.42449570e-02 4.83624160e-01 -9.17755738e-02 -1.57374993e-01
-6.31998062e-01 -4.30232197e-01 -1.22725332e+00 6.37082458e-01
1.37982631e+00 -6.30414844e-01 -4.11095142e-01 1.25915676e-01
1.01637542e+00 -6.21175349e-01 -9.24226761e-01 6.90299451e-01
4.78593081e-01 -1.60655570e+00 1.56859994e+00 -6.54414371e-02
7.06946671e-01 -7.43423462e-01 -2.48750672e-01 -1.39224648e+00
-7.62322724e-01 -7.23752320e-01 -3.42600644e-01 1.05895960e+00
-2.02680141e-01 -5.51397085e-01 3.01044375e-01 -4.10439307e-03
1.63381889e-01 -5.06146550e-01 -7.58848667e-01 -1.00960886e+00
-3.14978451e-01 -2.32898034e-02 6.65059209e-01 9.22991753e-01
-5.25861442e-01 3.02837461e-01 -9.96913910e-01 5.59563041e-01
1.29767108e+00 2.76968896e-01 6.37587249e-01 -8.52473378e-01
-3.24963033e-01 -1.81779549e-01 -1.03358351e-01 -1.33074129e+00
-3.32633168e-01 -1.17330022e-01 -5.18000200e-02 -1.46357894e+00
6.65848613e-01 -2.12557420e-01 -5.06186128e-01 -2.47791857e-01
-5.62198460e-01 5.98316610e-01 -1.00630731e-03 5.65070987e-01
-7.29006112e-01 5.29444516e-01 1.78195250e+00 9.33166742e-02
4.19376567e-02 -1.75008565e-01 -9.68452871e-01 5.09963691e-01
5.12251496e-01 1.14020638e-01 -3.43944013e-01 -5.61514318e-01
-1.84970886e-01 2.99508035e-01 6.07587397e-01 -1.10003507e+00
5.18593267e-02 -2.99904197e-01 5.60461044e-01 -4.93980646e-01
4.30984080e-01 -5.57649612e-01 4.66737062e-01 2.46275607e-02
-2.13667363e-01 -4.08452272e-01 -4.57674935e-02 7.60271490e-01
-4.06541020e-01 2.68239230e-01 1.39996123e+00 -2.41511375e-01
-7.99555302e-01 3.69975418e-01 4.52197224e-01 5.38745262e-02
6.36924148e-01 -1.90676108e-01 -9.28640068e-01 -4.97557849e-01
-5.10200143e-01 -1.70999337e-02 7.77030468e-01 5.17558873e-01
7.63845742e-01 -1.30833983e+00 -1.19728172e+00 1.72462732e-01
-7.40276724e-02 3.51303490e-03 9.00107920e-01 7.24944651e-01
-2.44137809e-01 2.29822353e-01 -4.28365946e-01 -4.96018887e-01
-1.12391448e+00 9.16567266e-01 3.84422958e-01 -3.67645204e-01
-9.23672974e-01 6.93909287e-01 6.07371211e-01 1.86006054e-01
-2.67258119e-02 3.42653804e-02 -2.84558207e-01 -5.02961636e-01
1.18907523e+00 5.33702970e-01 -1.91502362e-01 -6.44719660e-01
-8.32366794e-02 7.73748398e-01 -1.25624940e-01 -3.65419909e-02
1.35773385e+00 -7.85696566e-01 -3.59256379e-02 1.89071864e-01
8.47859442e-01 1.50231004e-01 -1.53297293e+00 -7.27534652e-01
-4.13133800e-01 -1.23287952e+00 1.89563081e-01 -8.39875340e-01
-1.10386825e+00 5.61834157e-01 8.86766136e-01 1.75821066e-01
1.72956502e+00 -1.58607483e-01 9.50180352e-01 -2.95449108e-01
8.16020012e-01 -9.08542335e-01 2.37034053e-01 1.90514475e-01
1.14375615e+00 -1.37364924e+00 2.76526123e-01 -7.24815488e-01
-5.01827836e-01 8.54441524e-01 6.19939804e-01 -1.62062734e-01
2.88305342e-01 1.90442219e-01 1.19600846e-02 2.92608052e-01
-7.37009227e-01 -4.04345810e-01 2.66203612e-01 8.60978484e-01
2.27380410e-01 -8.67602825e-02 -2.34466910e-01 5.13327181e-01
-4.43895115e-03 3.19119990e-01 7.14102149e-01 7.66289383e-02
-5.74898064e-01 -8.02544713e-01 -8.10928762e-01 1.55729905e-01
-4.92668629e-01 -3.26063573e-01 3.46656770e-01 4.60466802e-01
8.43027458e-02 9.92553592e-01 -1.92162395e-01 -3.43676150e-01
3.52252871e-01 -7.28713691e-01 6.72155976e-01 -3.24255258e-01
1.13483325e-01 2.20689848e-01 -1.53894946e-01 -8.97377968e-01
-5.64380705e-01 -3.64231557e-01 -6.96100235e-01 -3.59326452e-01
-2.23390609e-01 -8.57449174e-02 1.95540324e-01 4.16638762e-01
4.47990268e-01 5.88785827e-01 7.59035408e-01 -1.10891020e+00
-7.17831194e-01 -9.95523036e-01 -1.00535548e+00 7.32778430e-01
6.19945824e-01 -6.48360670e-01 -5.26497662e-01 1.67863742e-01] | [11.101119041442871, -2.138272762298584] |
db900d1c-5eaf-49d9-809b-3de3bbc6faac | supervised-contrastive-learning-for-product | 2202.02098 | null | https://arxiv.org/abs/2202.02098v2 | https://arxiv.org/pdf/2202.02098v2.pdf | Supervised Contrastive Learning for Product Matching | Contrastive learning has moved the state of the art for many tasks in computer vision and information retrieval in recent years. This poster is the first work that applies supervised contrastive learning to the task of product matching in e-commerce using product offers from different e-shops. More specifically, we employ a supervised contrastive learning technique to pre-train a Transformer encoder which is afterward fine-tuned for the matching task using pair-wise training data. We further propose a source-aware sampling strategy that enables contrastive learning to be applied for use cases in which the training data does not contain product identifiers. We show that applying supervised contrastive pre-training in combination with source-aware sampling significantly improves the state-of-the-art performance on several widely used benchmarks: For Abt-Buy, we reach an F1-score of 94.29 (+3.24 compared to the previous state-of-the-art), for Amazon-Google 79.28 (+ 3.7). For WDC Computers datasets, we reach improvements between +0.8 and +8.84 in F1-score depending on the training set size. Further experiments with data augmentation and self-supervised contrastive pre-training show that the former can be helpful for smaller training sets while the latter leads to a significant decline in performance due to inherent label noise. We thus conclude that contrastive pre-training has a high potential for product matching use cases in which explicit supervision is available. | ['Christian Bizer', 'Ralph Peeters'] | 2022-02-04 | null | null | null | null | ['entity-resolution'] | ['natural-language-processing'] | [ 4.28879291e-01 -1.85609668e-01 -5.50351083e-01 -5.85120976e-01
-9.24598873e-01 -6.05025351e-01 7.68181980e-01 4.42490816e-01
-5.63219786e-01 9.70617160e-02 -2.99299836e-01 -3.06962758e-01
-3.73556130e-02 -8.76984656e-01 -1.07027018e+00 -2.95061857e-01
-6.17340095e-02 8.06972980e-01 1.48426965e-01 -6.15274608e-01
-5.21978699e-02 2.87818253e-01 -1.74889696e+00 7.48564959e-01
7.15708196e-01 1.40262985e+00 1.91077411e-01 3.49041551e-01
-3.08550924e-01 5.68436563e-01 -4.60961252e-01 -8.96613061e-01
7.88047493e-01 -1.71368226e-01 -9.56841528e-01 9.20781568e-02
1.05571318e+00 -1.79553196e-01 5.72867654e-02 1.14379215e+00
2.17648283e-01 -5.22169769e-02 3.84361058e-01 -1.42061794e+00
-7.25932419e-01 6.41384244e-01 -7.08298743e-01 1.65962026e-01
3.79442066e-01 1.27208933e-01 1.33092058e+00 -8.23968470e-01
6.45848334e-01 1.14507008e+00 8.40198338e-01 3.47863019e-01
-1.43837690e+00 -8.12378645e-01 -1.79867132e-03 2.84078032e-01
-1.13314474e+00 -2.00303152e-01 7.68114388e-01 -2.43061319e-01
1.28768706e+00 1.45822376e-01 4.04150248e-01 8.16419244e-01
-2.24941611e-01 9.38740790e-01 1.21312213e+00 -7.36916661e-01
1.09639531e-02 5.46441853e-01 1.79640487e-01 5.65549731e-01
-1.93622503e-02 4.73242849e-01 -2.98837334e-01 7.72485808e-02
5.33657491e-01 -5.29325008e-02 3.13496351e-01 -3.88406843e-01
-8.44391763e-01 1.07912481e+00 7.74348974e-01 3.05835694e-01
-4.28321838e-01 -1.99116662e-01 7.12326407e-01 9.45480227e-01
4.70603853e-01 8.15882206e-01 -8.10403347e-01 2.25727201e-01
-1.08991265e+00 5.10344684e-01 6.99942648e-01 1.12013781e+00
8.48339498e-01 -2.00692773e-01 -1.23832151e-01 1.01515925e+00
1.40401095e-01 5.01283050e-01 4.51639384e-01 -8.14359248e-01
5.31715989e-01 6.99689388e-01 8.06773826e-02 -5.84896088e-01
-1.32079944e-01 -5.40199101e-01 -5.10413229e-01 2.51499832e-01
6.88277185e-01 2.53395230e-01 -1.07961619e+00 1.49897969e+00
1.69185445e-01 8.01264402e-03 -7.10786879e-02 6.94036841e-01
9.03044760e-01 4.01156276e-01 1.93730980e-01 1.45860553e-01
1.53280222e+00 -1.31694114e+00 -4.37242210e-01 -2.81838506e-01
8.02578926e-01 -1.08261967e+00 1.34601629e+00 4.10818487e-01
-1.07606196e+00 -1.02350819e+00 -1.34617448e+00 -2.04876065e-02
-6.89567566e-01 -3.30040082e-02 8.29941630e-01 8.08901668e-01
-6.91818357e-01 7.16833591e-01 -4.23618078e-01 -2.76916265e-01
5.94480395e-01 5.29502451e-01 -4.00908381e-01 -4.12565231e-01
-1.27822268e+00 1.00757253e+00 3.54787201e-01 -4.55037326e-01
-5.55391788e-01 -1.17244160e+00 -8.20710123e-01 7.28273466e-02
4.31711763e-01 -5.38993180e-01 1.49895680e+00 -1.45012021e+00
-1.29064679e+00 1.12399197e+00 2.59600401e-01 -8.87524784e-01
6.23976350e-01 -2.31176704e-01 -5.92467189e-01 -3.60202789e-02
2.43185148e-01 1.13809288e+00 7.79633582e-01 -1.10597312e+00
-1.03014278e+00 -3.44572246e-01 2.97199339e-01 1.33260973e-02
-2.45479286e-01 1.78365856e-01 -5.06916165e-01 -5.25281191e-01
-4.23728555e-01 -1.10451531e+00 -3.87353078e-02 5.75864054e-02
1.17517158e-01 -3.76397192e-01 6.23788655e-01 -3.87741596e-01
7.07030356e-01 -2.08862042e+00 -3.86473656e-01 1.41851857e-01
-1.82396457e-01 6.09605312e-01 -5.38111448e-01 5.30941747e-02
-3.88952434e-01 -1.64915472e-01 7.19840033e-03 -4.15492535e-01
8.31088498e-02 3.67741026e-02 -7.67392665e-02 2.19551742e-01
4.85264689e-01 1.10962927e+00 -8.06686819e-01 -1.73411876e-01
3.71463329e-01 3.51996124e-01 -5.51877081e-01 2.95683950e-01
-3.33894551e-01 -2.46918425e-02 1.31472081e-01 5.63146472e-01
9.64302242e-01 -2.85623789e-01 1.19303249e-01 -3.79344076e-01
1.96320698e-01 5.12807310e-01 -1.25672174e+00 1.81798077e+00
-8.71269345e-01 4.95971680e-01 -5.96619733e-02 -1.15858781e+00
8.95385742e-01 2.54868716e-02 4.83192444e-01 -1.36661947e+00
3.63630168e-02 5.50115585e-01 1.79474026e-01 -6.28580302e-02
6.40215337e-01 -1.50440171e-01 5.59911728e-02 1.21402360e-01
3.62204492e-01 6.14185408e-02 4.13814068e-01 -8.34578499e-02
8.56278658e-01 1.01598531e-01 -2.83043068e-02 -3.69815886e-01
6.58150911e-01 2.12498918e-01 2.00577781e-01 6.57741666e-01
-3.73379067e-02 5.02561688e-01 4.95241880e-02 -6.00623310e-01
-1.18895113e+00 -8.15204740e-01 -1.89098656e-01 1.38680780e+00
1.39240861e-01 -4.38052535e-01 -6.68066800e-01 -1.02547348e+00
3.87786478e-01 7.17696548e-01 -5.99995255e-01 -2.52329588e-01
-5.91990709e-01 -4.00140613e-01 2.45671153e-01 6.83011591e-01
7.47444749e-01 -1.13646960e+00 -3.14212292e-01 3.19536060e-01
2.01952010e-01 -1.32984507e+00 -4.76810664e-01 4.03053850e-01
-7.79821634e-01 -1.04966331e+00 -7.49538541e-01 -1.09010160e+00
4.31977123e-01 3.44782978e-01 1.61915743e+00 -5.02726696e-02
-2.59400368e-01 1.53927460e-01 -2.83839732e-01 -4.37315136e-01
-6.56748056e-01 4.62300867e-01 -2.75983572e-01 -1.05007261e-01
8.66834939e-01 -1.19367696e-01 -5.35618603e-01 5.67434430e-01
-7.61660814e-01 -2.40779176e-01 5.31251490e-01 1.09705794e+00
5.98137975e-01 5.55970557e-02 5.14524758e-01 -1.38658571e+00
3.01677614e-01 -3.21141183e-01 -9.22803700e-01 2.35119224e-01
-1.21078110e+00 1.92939967e-01 5.87386191e-01 -5.86505294e-01
-9.77508962e-01 2.48719215e-01 -2.86374807e-01 -4.67956901e-01
-1.15373485e-01 3.24678034e-01 6.40998408e-02 -2.78588776e-02
8.76296222e-01 -1.56421110e-01 1.54139400e-01 -6.27556980e-01
4.59884256e-01 7.98896968e-01 5.22783518e-01 -4.52791363e-01
4.71716970e-01 1.27946571e-01 -7.79227018e-02 -3.00083250e-01
-8.19104373e-01 -7.77378321e-01 -4.88157660e-01 2.94322014e-01
3.36497396e-01 -1.05386674e+00 -7.35812366e-01 5.13731241e-01
-7.55526364e-01 -3.14285427e-01 -5.01553833e-01 2.37619281e-01
-4.25165087e-01 7.83384815e-02 -7.11223483e-01 -3.50896955e-01
-5.72879612e-01 -1.16149390e+00 1.05345690e+00 3.15525793e-02
-1.59057260e-01 -9.36229408e-01 -1.90376788e-01 5.39212584e-01
5.76315939e-01 -1.85111806e-01 7.13333964e-01 -9.97234941e-01
-3.45433801e-01 -2.52039671e-01 -2.49275684e-01 5.99945784e-01
2.14310676e-01 -4.08816904e-01 -9.56497669e-01 -3.25023592e-01
-3.17741424e-01 -3.43679607e-01 9.80376899e-01 8.32952186e-02
7.66942024e-01 -1.66263074e-01 -2.36265436e-01 3.07677686e-01
1.54530621e+00 1.02464549e-01 6.98076189e-01 7.20578492e-01
5.40114760e-01 7.19136596e-01 1.10310650e+00 1.08234130e-01
2.74314433e-01 1.10114729e+00 4.05194044e-01 -4.05943036e-01
-3.37495506e-01 -3.95437330e-01 1.06365532e-01 9.30431113e-02
3.12312782e-01 1.49559885e-01 -5.46136081e-01 7.18666136e-01
-1.75891197e+00 -8.55022430e-01 -1.01176225e-01 2.45069814e+00
9.31684315e-01 4.36676741e-01 5.67716360e-01 2.08061606e-01
6.67932928e-01 -2.16630936e-01 -5.47815144e-01 -5.01116693e-01
4.02749628e-02 7.26315200e-01 8.20963562e-01 3.63105834e-01
-1.44980657e+00 8.79063845e-01 5.85359144e+00 9.31597650e-01
-1.17775226e+00 2.11985961e-01 6.74528897e-01 7.47180730e-02
-2.19176292e-01 -2.23847419e-01 -8.46046031e-01 6.54615104e-01
1.13141334e+00 2.68964559e-01 3.48239005e-01 1.07243407e+00
-5.49348652e-01 9.71476883e-02 -1.46032691e+00 1.18280125e+00
-1.17018878e-01 -1.36629558e+00 -2.14701548e-01 8.61470476e-02
9.03525293e-01 1.59592047e-01 3.99258107e-01 6.87395096e-01
4.14468378e-01 -9.13504541e-01 6.80812061e-01 -2.23810390e-01
8.49598944e-01 -7.89069951e-01 1.09001160e+00 1.58446118e-01
-1.13994288e+00 -4.65287678e-02 -3.54629070e-01 1.31110892e-01
-6.23833500e-02 5.45812964e-01 -8.27323496e-01 5.77503026e-01
8.58939230e-01 7.62440860e-01 -6.87467873e-01 8.93668830e-01
-3.25976871e-02 3.84584069e-01 -4.25386667e-01 1.15158856e-01
2.78631747e-01 8.88442025e-02 -8.07968974e-02 1.38816488e+00
4.01284266e-03 -5.40857315e-01 1.86026603e-01 7.98768759e-01
-3.21907431e-01 2.24704266e-01 -3.38328123e-01 2.79623151e-01
4.27000642e-01 1.31658900e+00 -3.83109152e-01 -4.53381747e-01
-8.40250134e-01 1.17207563e+00 2.95817584e-01 -1.13788545e-01
-8.48141491e-01 -4.29163724e-01 5.32656431e-01 3.72685850e-01
8.94861042e-01 3.81340861e-01 -5.12241619e-03 -7.53656209e-01
1.24271773e-01 -1.11037827e+00 5.99191666e-01 -3.51215243e-01
-1.60239029e+00 6.35867178e-01 1.32325320e-02 -1.24197090e+00
-5.04285693e-01 -8.59958768e-01 -6.11268990e-02 9.08864081e-01
-1.87336326e+00 -1.32452047e+00 -9.99070331e-02 3.59110832e-01
5.59596777e-01 -4.64836508e-01 9.30752397e-01 7.33886540e-01
-1.00738987e-01 1.10580444e+00 -4.34851507e-03 5.39472401e-02
8.33640814e-01 -1.42082882e+00 7.60509968e-01 3.90358925e-01
5.23507774e-01 5.47333241e-01 5.98509610e-01 -2.54647017e-01
-1.28182697e+00 -1.02867150e+00 9.62649822e-01 -3.59031290e-01
5.40865541e-01 -4.53886747e-01 -8.26476336e-01 5.84190547e-01
3.36681485e-01 3.23818982e-01 5.89808106e-01 4.16336149e-01
-1.04997480e+00 -6.30392790e-01 -1.57658422e+00 9.79412571e-02
8.31741929e-01 -6.84655488e-01 -4.26351458e-01 3.57171416e-01
6.48504674e-01 -3.92976195e-01 -1.14208186e+00 5.34499526e-01
6.53542161e-01 -8.20183456e-01 1.04178596e+00 -5.05817711e-01
3.63502562e-01 -1.93337500e-02 -2.15720937e-01 -1.23176944e+00
-3.17035824e-01 -4.34137434e-01 1.13564216e-01 1.43432677e+00
6.62198365e-01 -4.68581140e-01 1.09796560e+00 5.40383995e-01
6.81407377e-02 -6.12660170e-01 -5.12087047e-01 -1.05893564e+00
2.62821108e-01 -3.28400373e-01 7.75069058e-01 1.07182908e+00
-2.53117919e-01 7.05983400e-01 -1.43653318e-01 -2.67267644e-01
3.49870354e-01 2.18979284e-01 6.68697238e-01 -1.30110133e+00
-5.80742776e-01 -4.21118081e-01 -3.79996121e-01 -1.29060197e+00
2.03726798e-01 -1.05827987e+00 -1.42664224e-01 -1.10953522e+00
1.43066689e-01 -6.80982709e-01 -4.97231543e-01 6.11670315e-01
-1.19277880e-01 4.07446712e-01 5.03923118e-01 8.83409008e-02
-4.74180609e-01 3.46629843e-02 9.84118998e-01 -5.95875919e-01
-2.06750169e-01 1.91688448e-01 -7.45450199e-01 6.04273640e-02
7.65133679e-01 -5.05640328e-01 -3.18360835e-01 -2.28607461e-01
1.19484447e-01 -4.98245955e-01 1.73250645e-01 -7.39094198e-01
-5.14137633e-02 3.53804201e-01 2.14698136e-01 -3.79894406e-01
1.66182622e-01 -1.21292090e+00 5.69386669e-02 6.01630449e-01
-5.52959859e-01 1.70796588e-01 3.90049696e-01 2.21681193e-01
-3.72698694e-01 -5.35646200e-01 6.81990445e-01 -2.50766397e-01
-9.16950405e-01 6.93304017e-02 1.37341380e-01 1.54212639e-02
7.61022627e-01 -9.37813520e-02 -2.26883590e-01 -1.40048712e-01
-3.35392952e-01 4.00541723e-02 2.61264145e-01 6.39864266e-01
3.89804356e-02 -1.45372295e+00 -7.63109624e-01 3.43042880e-01
5.30817509e-01 -3.38959396e-01 -5.18785827e-02 3.70347440e-01
-3.47456634e-01 5.15370607e-01 -4.47316021e-01 -6.77912891e-01
-1.40268195e+00 9.64086413e-01 2.53170580e-01 -6.22955620e-01
-2.56435305e-01 8.10691476e-01 -2.32782543e-01 -6.04051292e-01
2.29945540e-01 -3.83349597e-01 -5.65144271e-02 8.67003798e-02
4.82503682e-01 3.36044610e-01 6.58463299e-01 -6.06597006e-01
-3.48408341e-01 4.42923635e-01 -7.47534335e-01 7.15170056e-02
1.30484939e+00 2.25934982e-01 3.84904504e-01 -7.14568123e-02
1.72278070e+00 -2.92361468e-01 -1.12254679e+00 -5.44642210e-01
1.91150054e-01 -5.20581067e-01 1.68413639e-01 -1.15008450e+00
-1.44117534e+00 7.11317301e-01 1.07653964e+00 2.62919426e-01
1.18107069e+00 1.08234547e-01 9.39187884e-01 2.79378235e-01
3.98722559e-01 -1.12372243e+00 2.81101521e-02 2.13980973e-01
5.67722440e-01 -1.70237112e+00 1.66183617e-02 -7.11446822e-01
-6.48051441e-01 7.77315557e-01 5.48849583e-01 -3.89172494e-01
6.72272503e-01 2.20225379e-01 1.27405599e-01 -2.05055401e-01
-5.37490129e-01 -4.04196650e-01 4.53714222e-01 6.56490326e-01
5.09079099e-01 8.83464441e-02 -1.30140111e-01 4.65850919e-01
-2.17764050e-01 2.04610955e-02 -3.45007367e-02 8.22132885e-01
1.17812216e-01 -1.64015484e+00 9.68152136e-02 3.99505466e-01
-5.98148763e-01 -2.16967210e-01 -1.31740525e-01 9.30918157e-01
2.83413321e-01 9.45064127e-01 4.04740036e-01 -2.95059592e-01
7.61257350e-01 -5.59412176e-03 8.94070089e-01 -6.15992069e-01
-1.29439664e+00 1.06542949e-02 3.70735556e-01 -4.96787637e-01
-7.32698619e-01 -6.50066674e-01 -9.24415767e-01 -3.65999579e-01
-5.95382512e-01 -6.44912496e-02 7.01688945e-01 6.20728433e-01
4.60845947e-01 3.45773131e-01 6.34424329e-01 -3.86778802e-01
-8.07797968e-01 -9.57862973e-01 -5.03697336e-01 9.44202602e-01
2.45905474e-01 -7.77757823e-01 -1.12285614e-01 3.68059464e-02] | [9.778712272644043, 8.315797805786133] |
726ef22e-eee7-422a-b68c-9b4f36f61e52 | a-deep-bag-of-features-model-for-music-auto | 1508.04999 | null | http://arxiv.org/abs/1508.04999v3 | http://arxiv.org/pdf/1508.04999v3.pdf | A Deep Bag-of-Features Model for Music Auto-Tagging | Feature learning and deep learning have drawn great attention in recent years
as a way of transforming input data into more effective representations using
learning algorithms. Such interest has grown in the area of music information
retrieval (MIR) as well, particularly in music audio classification tasks such
as auto-tagging. In this paper, we present a two-stage learning model to
effectively predict multiple labels from music audio. The first stage learns to
project local spectral patterns of an audio track onto a high-dimensional
sparse space in an unsupervised manner and summarizes the audio track as a
bag-of-features. The second stage successively performs the unsupervised
learning on the bag-of-features in a layer-by-layer manner to initialize a deep
neural network and finally fine-tunes it with the tag labels. Through the
experiment, we rigorously examine training choices and tuning parameters, and
show that the model achieves high performance on Magnatagatune, a popularly
used dataset in music auto-tagging. | ['Kyogu Lee', 'Juhan Nam', 'Jorge Herrera'] | 2015-08-20 | null | null | null | null | ['music-auto-tagging'] | ['music'] | [ 2.70651996e-01 -3.11069936e-01 -1.75419465e-01 -4.61533070e-01
-1.11068392e+00 -6.70905948e-01 2.54341871e-01 1.61387041e-01
-2.17934653e-01 2.03830391e-01 5.17079651e-01 2.02412263e-01
-3.81510645e-01 -6.92582190e-01 -4.84435260e-01 -6.13061607e-01
-2.15855405e-01 4.87561256e-01 -4.36472781e-02 1.84964359e-01
1.51164278e-01 1.96740761e-01 -1.67540658e+00 5.38587034e-01
2.36901835e-01 1.16814339e+00 -4.96452302e-03 5.58230340e-01
-1.42808817e-02 7.68089712e-01 -4.59049016e-01 -1.45409435e-01
2.54863799e-01 -5.20088375e-01 -9.45035160e-01 8.11521187e-02
4.37838912e-01 1.22608915e-01 -3.63518476e-01 8.87103915e-01
6.66940153e-01 4.75463092e-01 6.32017791e-01 -8.53454590e-01
-3.19568813e-01 1.17641819e+00 -4.32114303e-01 4.24766401e-03
1.18953567e-02 -3.29864353e-01 1.46528482e+00 -8.16921353e-01
1.80852428e-01 1.02967143e+00 7.78794169e-01 1.56695276e-01
-1.26687765e+00 -9.68099952e-01 -1.81378156e-01 2.02795684e-01
-1.58532000e+00 -3.85669410e-01 1.05907989e+00 -7.65259027e-01
5.39826214e-01 3.35464850e-02 6.41961455e-01 6.67215645e-01
-1.65157810e-01 7.85724998e-01 7.06820846e-01 -6.10789120e-01
1.71723366e-01 -2.64247924e-01 -1.36338413e-01 4.04224813e-01
-4.07793343e-01 -6.31993115e-02 -9.99701262e-01 -1.32155642e-01
9.05694425e-01 -1.20711199e-03 1.61168382e-01 -3.18870544e-01
-1.25683987e+00 8.76299083e-01 4.44354713e-01 6.76344633e-01
-3.46527189e-01 5.07775068e-01 5.94147205e-01 2.46643886e-01
3.62083584e-01 7.97096133e-01 -4.09035444e-01 -3.20376158e-01
-1.21285105e+00 1.45093262e-01 4.95611817e-01 4.46940809e-01
7.35697091e-01 3.05268168e-01 -2.07401469e-01 1.39217556e+00
2.05053478e-01 2.69026071e-01 8.24295044e-01 -7.80542076e-01
1.32575825e-01 2.99985886e-01 -2.82229155e-01 -9.35394943e-01
-3.46019447e-01 -8.94723117e-01 -9.76258278e-01 -5.85328713e-02
1.49558544e-01 2.64016129e-02 -5.90397060e-01 1.86410284e+00
1.01948060e-01 5.50800920e-01 -3.04768205e-01 9.80975747e-01
5.62244058e-01 7.31636345e-01 1.48355097e-01 -1.38081126e-02
1.16089737e+00 -9.52646017e-01 -5.08484423e-01 -3.63780186e-02
5.06063998e-01 -1.10597157e+00 1.11078823e+00 6.58428431e-01
-9.37936544e-01 -9.78313267e-01 -8.67319286e-01 3.68756093e-02
-1.61098033e-01 6.00308597e-01 7.70530283e-01 1.67840883e-01
-7.45274723e-01 9.30031776e-01 -7.04760611e-01 -7.15328902e-02
5.21940649e-01 6.19427621e-01 -2.10227653e-01 3.88748556e-01
-1.08368564e+00 1.24558456e-01 6.20163500e-01 -2.85746574e-01
-8.81441236e-01 -6.72685981e-01 -6.55731201e-01 3.76204431e-01
4.28336747e-02 -4.75460023e-01 1.44104671e+00 -1.03378940e+00
-1.78722012e+00 9.04662013e-01 1.37964576e-01 -4.81331974e-01
-2.48423219e-01 -4.48930353e-01 -5.55077672e-01 -1.91889361e-01
1.18601836e-01 7.18456924e-01 1.10958600e+00 -8.89479518e-01
-8.34775507e-01 -1.94200307e-01 -6.68866262e-02 1.43106416e-01
-7.58522928e-01 6.16694167e-02 -2.91954130e-01 -1.17055178e+00
2.73286819e-01 -1.02310860e+00 -7.68510774e-02 -3.64761025e-01
-4.45537597e-01 -3.19348067e-01 3.52516532e-01 -1.32789925e-01
1.55291736e+00 -2.51846433e+00 3.45829278e-01 3.12378258e-01
-7.89477751e-02 1.46373019e-01 -2.59498268e-01 3.01421821e-01
-3.30723852e-01 -3.24991077e-01 -8.86245519e-02 -4.30186510e-01
1.31080955e-01 -7.61699155e-02 -6.98758185e-01 1.76757425e-01
-5.38961031e-02 6.73573196e-01 -8.94388139e-01 -3.26926589e-01
9.71512944e-02 4.16725755e-01 -8.07477832e-01 4.13894773e-01
-2.13208452e-01 5.33779740e-01 -2.47699127e-01 2.29653150e-01
5.69921546e-02 -3.00185263e-01 8.98078308e-02 -3.88711035e-01
-9.70259532e-02 6.27513707e-01 -1.29855800e+00 2.30839920e+00
-6.32143199e-01 6.18982553e-01 -3.46104652e-01 -9.88881767e-01
1.00614858e+00 4.17815834e-01 9.30966318e-01 -4.55902725e-01
7.16534927e-02 2.24618688e-01 -3.62154208e-02 -2.09192052e-01
4.45594072e-01 -4.08399642e-01 -4.60937023e-01 6.80552304e-01
5.12282312e-01 6.53500259e-02 -1.38687953e-01 -2.03529462e-01
8.90282750e-01 9.96817946e-02 2.58099645e-01 3.99977230e-02
6.32481754e-01 -2.22869322e-01 5.14095902e-01 4.78696376e-01
4.46187496e-01 5.46021461e-01 -1.19747885e-01 -5.24254799e-01
-9.21270847e-01 -9.50290143e-01 -2.37641364e-01 1.66977286e+00
-3.49033624e-01 -8.90345573e-01 -4.41488504e-01 -3.28210294e-01
1.23867609e-01 2.63950706e-01 -3.72196347e-01 -3.73181373e-01
-4.50061500e-01 -3.62851709e-01 7.18798518e-01 5.64050853e-01
2.21479148e-01 -1.43489361e+00 -2.67139107e-01 4.71004814e-01
-2.59782597e-02 -6.69972479e-01 -5.38556516e-01 6.56339586e-01
-8.10800076e-01 -6.49745762e-01 -5.39731205e-01 -1.08533120e+00
1.56423137e-01 5.72122149e-02 1.07394576e+00 -3.48555505e-01
-3.25990051e-01 1.79784253e-01 -4.73136365e-01 -2.39409730e-01
-2.75768489e-01 5.40936649e-01 4.24536318e-01 4.11020607e-01
4.47822928e-01 -1.06470287e+00 -4.40876186e-01 2.24273607e-01
-8.92359018e-01 -6.17031157e-02 5.71863711e-01 7.18853831e-01
1.06790555e+00 2.04523236e-01 7.15569615e-01 -8.44150305e-01
4.64272529e-01 -2.32389212e-01 -4.40734863e-01 -1.73266217e-01
-1.99488223e-01 1.23493642e-01 6.07662678e-01 -6.47355199e-01
-4.04879242e-01 5.39501429e-01 -2.81112641e-01 -5.87327659e-01
-1.54481396e-01 6.14013910e-01 -1.08978122e-01 3.98070440e-02
5.85160911e-01 1.31575674e-01 -4.89794463e-01 -9.96588826e-01
5.52103996e-01 9.91244912e-01 8.55326176e-01 -5.55446386e-01
9.50947583e-01 5.83475195e-02 -1.02177374e-01 -5.10426342e-01
-1.56431603e+00 -8.82356703e-01 -8.50791574e-01 -1.92038238e-01
7.95436621e-01 -1.10284817e+00 -6.28088117e-01 2.66032517e-01
-6.40520990e-01 -2.59684414e-01 -7.14582920e-01 7.49342561e-01
-8.13549519e-01 -1.38058811e-01 -6.24172926e-01 -5.99222779e-01
-2.79950917e-01 -7.75577962e-01 1.34590471e+00 1.77765444e-01
-4.94661838e-01 -7.09123909e-01 7.04315841e-01 1.71975330e-01
2.65014023e-01 -2.89441496e-01 1.02218115e+00 -9.05710399e-01
-4.04021055e-01 -8.35149512e-02 5.23067676e-02 3.11399370e-01
3.62569958e-01 -4.77958083e-01 -1.50466681e+00 -3.09443414e-01
-3.41402024e-01 -5.85455060e-01 9.34576750e-01 4.49644983e-01
1.68098581e+00 -6.09372817e-02 -1.22416476e-02 8.89025390e-01
1.28913176e+00 9.28881243e-02 3.03515226e-01 3.16963106e-01
8.32931280e-01 1.08576067e-01 4.84744877e-01 5.77246904e-01
-9.22687948e-02 1.02202034e+00 1.71624929e-01 -9.20026936e-03
-2.30774045e-01 -6.48398399e-01 1.80758595e-01 1.50296533e+00
1.57468915e-01 2.21416965e-01 -6.03219032e-01 4.65515286e-01
-1.87384629e+00 -9.38901305e-01 5.26040733e-01 2.26013827e+00
1.07493567e+00 -1.32963583e-01 4.55640286e-01 6.04265809e-01
5.98163843e-01 1.90309986e-01 -3.43818456e-01 -6.99579939e-02
3.12824845e-02 5.99034488e-01 -1.16427183e-01 1.03016995e-01
-1.57895708e+00 1.19211078e+00 6.10290956e+00 1.04699469e+00
-1.26893723e+00 1.68503389e-01 2.79300541e-01 -1.33401915e-01
4.32395414e-02 -7.44241104e-02 -5.86277187e-01 2.71585345e-01
1.01138520e+00 7.73163810e-02 6.06605411e-01 7.37457335e-01
1.36843473e-02 4.71273541e-01 -1.17078650e+00 1.40940571e+00
-2.48818137e-02 -1.34571505e+00 1.98663846e-01 4.47366945e-02
7.46061146e-01 -1.71801344e-01 2.05840319e-01 4.97618288e-01
1.06108241e-01 -9.47443902e-01 7.62810349e-01 5.98908663e-01
8.99271965e-01 -9.33114231e-01 4.98448461e-01 4.70376611e-02
-1.39296997e+00 -3.25686604e-01 -4.05266076e-01 -1.51809201e-01
1.52596697e-01 6.73827469e-01 -7.00228631e-01 2.99766868e-01
7.13024497e-01 1.05766785e+00 -3.33874464e-01 1.50767541e+00
-1.06506191e-01 9.88458574e-01 -1.52863279e-01 3.82973313e-01
1.31580517e-01 -1.79088078e-02 3.29712629e-01 1.19407344e+00
5.00448346e-01 -1.22718222e-01 4.40433770e-01 5.91659546e-01
-4.08046126e-01 4.19198304e-01 -1.94752261e-01 -4.18190747e-01
4.36036110e-01 1.39487684e+00 -5.54203510e-01 -1.63729221e-01
-1.59504302e-02 7.34422266e-01 3.06770861e-01 -1.31980507e-02
-5.22137284e-01 -3.61710936e-01 6.69996381e-01 1.67819813e-01
4.50718969e-01 -1.79374650e-01 -4.11997773e-02 -9.86816645e-01
-3.99542153e-01 -7.98414171e-01 4.19227719e-01 -8.41589332e-01
-1.44104254e+00 6.48165762e-01 -4.50454742e-01 -1.70008492e+00
-4.65459317e-01 -2.69723028e-01 -3.41444641e-01 6.60101533e-01
-1.18767953e+00 -1.11091101e+00 -9.82807055e-02 6.82380021e-01
4.50799942e-01 -7.02557445e-01 1.35652983e+00 6.39313161e-01
-2.46480793e-01 6.75007105e-01 2.31676072e-01 3.87873441e-01
1.09862399e+00 -1.25406492e+00 1.68041751e-01 2.54097104e-01
1.27817678e+00 4.25342888e-01 2.53384173e-01 -1.75584272e-01
-9.39402878e-01 -1.24556482e+00 7.98175335e-01 -5.71029037e-02
9.36537206e-01 -3.94468725e-01 -7.70611882e-01 5.08857369e-01
-4.88186292e-02 -1.16952330e-01 1.32682526e+00 5.74916661e-01
-5.64404666e-01 -4.55346912e-01 -3.67640376e-01 1.12288527e-01
8.58110189e-01 -1.03081155e+00 -5.67484438e-01 4.56587791e-01
5.88736176e-01 -2.59404570e-01 -1.12678170e+00 2.05366164e-01
6.31637871e-01 -5.57028413e-01 1.03123701e+00 -7.81422615e-01
1.88612908e-01 -5.17439961e-01 -3.36544394e-01 -1.32439780e+00
-7.84316301e-01 -7.87455559e-01 1.63121037e-02 1.38833749e+00
3.82152468e-01 1.47074819e-01 7.31301010e-01 -4.42736506e-01
-1.73246995e-01 -5.50118983e-01 -7.81260550e-01 -4.23289061e-01
-2.67973632e-01 -6.00570083e-01 6.58553064e-01 1.04456341e+00
1.56480640e-01 6.70481503e-01 -5.42713404e-01 -8.50321054e-02
3.48544419e-01 6.36565030e-01 8.43594968e-01 -1.69844031e+00
-7.51084149e-01 -3.74871671e-01 -7.80609608e-01 -1.17072332e+00
1.67406216e-01 -1.23494673e+00 1.89779460e-01 -1.03008974e+00
2.59080470e-01 -7.65668988e-01 -8.72732759e-01 8.25212002e-01
1.62433475e-01 7.41245985e-01 1.17661282e-01 4.91999239e-01
-8.24912906e-01 4.92345333e-01 8.80659699e-01 -2.69172341e-01
-3.29417914e-01 3.56492430e-01 -6.05929196e-01 7.31318295e-01
6.59376144e-01 -7.22497702e-01 -3.73868585e-01 -3.83825451e-01
4.72153574e-01 -5.94534166e-02 1.15441270e-01 -1.40743721e+00
1.35082930e-01 2.04543933e-01 4.24177736e-01 -5.59113860e-01
5.27527273e-01 -7.62517512e-01 1.65972590e-01 -1.11401178e-01
-9.11170781e-01 -2.23963812e-01 1.02296114e-01 4.07342464e-01
-7.06960380e-01 -2.02091277e-01 6.49724185e-01 1.73268914e-01
-5.22057235e-01 4.08100218e-01 -1.12232231e-01 -7.27883354e-02
3.77972096e-01 1.31740674e-01 3.48091602e-01 -3.93811107e-01
-1.21333075e+00 -2.64864326e-01 1.71289481e-02 6.31215811e-01
2.16682613e-01 -1.81331789e+00 -4.85488683e-01 3.35222304e-01
2.21955165e-01 -2.43143186e-01 1.82852834e-01 3.64003152e-01
9.46475286e-03 4.45682704e-01 -2.32654989e-01 -7.77992904e-01
-1.13894784e+00 4.06802446e-01 1.96382672e-01 -2.40553796e-01
-7.56559730e-01 1.03612745e+00 9.02752727e-02 -2.83149302e-01
6.78074121e-01 -1.69875905e-01 -4.76736695e-01 1.43828079e-01
5.08996010e-01 -1.12758331e-01 -5.92656881e-02 -8.49276006e-01
-1.85054094e-01 9.96980309e-01 1.18962802e-01 -1.75868794e-01
1.63345134e+00 1.70203462e-01 7.19741045e-04 8.83008301e-01
1.17742980e+00 1.19311012e-01 -9.41074312e-01 -5.44568062e-01
1.76788703e-01 -3.21088284e-01 2.91116446e-01 -6.42381191e-01
-1.19924736e+00 1.00687206e+00 7.44022548e-01 2.21541807e-01
1.20434022e+00 1.89270526e-01 8.94981682e-01 4.15821135e-01
3.70923311e-01 -1.05143857e+00 2.97097832e-01 5.59149623e-01
5.12925923e-01 -7.39374459e-01 -2.39464864e-01 -1.36899017e-02
-4.06282872e-01 1.08964562e+00 9.08695087e-02 -3.42426836e-01
9.59217966e-01 3.07400167e-01 1.08982340e-01 -6.90876991e-02
-4.88318413e-01 -3.30606848e-01 7.55041838e-01 3.20368379e-01
7.77384758e-01 1.83309823e-01 2.51351804e-01 1.05308795e+00
-7.57523000e-01 1.06590800e-01 -1.25525534e-01 5.56082964e-01
-6.14193320e-01 -1.42993295e+00 -2.06475556e-01 3.90314579e-01
-4.72367167e-01 -9.87377539e-02 -3.69138509e-01 2.23745808e-01
3.70408893e-01 6.30558014e-01 2.49778450e-01 -8.38424265e-01
2.36260250e-01 3.43555540e-01 5.12509704e-01 -9.48137283e-01
-9.92275715e-01 6.76530600e-01 -1.69641480e-01 -3.81269962e-01
-5.64556181e-01 -4.87572134e-01 -1.19802690e+00 3.11770201e-01
-3.58394355e-01 5.95521569e-01 6.26515687e-01 1.00135362e+00
3.10382515e-01 8.69690716e-01 8.53048444e-01 -9.77206647e-01
-5.15391052e-01 -1.04978883e+00 -9.08572376e-01 3.87860239e-01
9.22356471e-02 -6.89554095e-01 -1.20297499e-01 3.93938512e-01] | [15.738375663757324, 5.219289302825928] |
88e62c17-9000-47ec-8f24-84e1154f29ac | glocal-energy-based-learning-for-few-shot | 2304.11855 | null | https://arxiv.org/abs/2304.11855v1 | https://arxiv.org/pdf/2304.11855v1.pdf | Glocal Energy-based Learning for Few-Shot Open-Set Recognition | Few-shot open-set recognition (FSOR) is a challenging task of great practical value. It aims to categorize a sample to one of the pre-defined, closed-set classes illustrated by few examples while being able to reject the sample from unknown classes. In this work, we approach the FSOR task by proposing a novel energy-based hybrid model. The model is composed of two branches, where a classification branch learns a metric to classify a sample to one of closed-set classes and the energy branch explicitly estimates the open-set probability. To achieve holistic detection of open-set samples, our model leverages both class-wise and pixel-wise features to learn a glocal energy-based score, in which a global energy score is learned using the class-wise features, while a local energy score is learned using the pixel-wise features. The model is enforced to assign large energy scores to samples that are deviated from the few-shot examples in either the class-wise features or the pixel-wise features, and to assign small energy scores otherwise. Experiments on three standard FSOR datasets show the superior performance of our model. | ['Yanning Zhang', 'Wei Wei', 'Lei Zhang', 'Peng Wang', 'Guansong Pang', 'Haoyu Wang'] | 2023-04-24 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_Glocal_Energy-Based_Learning_for_Few-Shot_Open-Set_Recognition_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_Glocal_Energy-Based_Learning_for_Few-Shot_Open-Set_Recognition_CVPR_2023_paper.pdf | cvpr-2023-1 | ['open-set-learning'] | ['miscellaneous'] | [ 5.50028443e-01 1.81048941e-02 -4.36160415e-01 -5.80807924e-01
-9.53108430e-01 -3.31038952e-01 5.21521568e-01 3.20935190e-01
-1.22814573e-01 3.18306088e-01 -1.60211712e-01 1.54207364e-01
-2.08087340e-01 -1.00410414e+00 -4.88221586e-01 -1.11590302e+00
6.19009323e-02 3.57783020e-01 3.75040293e-01 1.07182242e-01
3.69670391e-01 3.97820055e-01 -1.92191792e+00 3.07845086e-01
8.40785503e-01 1.58186567e+00 1.16368169e-02 4.90280688e-01
-6.41272590e-02 7.21691787e-01 -3.19956660e-01 2.42010672e-02
3.22521836e-01 -7.23293066e-01 -3.88916254e-01 1.80538055e-02
5.65217972e-01 -3.00236233e-02 -1.70412138e-01 1.37905943e+00
2.91215092e-01 4.64303106e-01 9.15832102e-01 -1.13294029e+00
-4.94608104e-01 7.71200582e-02 -4.31846678e-01 2.10735187e-01
3.55348319e-01 2.62842894e-01 1.30975974e+00 -1.06545758e+00
4.98510659e-01 9.38899755e-01 4.08228815e-01 4.67034459e-01
-1.14925456e+00 -4.89032149e-01 2.13165432e-01 3.73954684e-01
-1.42430377e+00 -3.74765962e-01 9.32860255e-01 -4.57788885e-01
5.72258234e-01 4.42094892e-01 6.75491631e-01 6.52474582e-01
2.33435538e-02 6.52653396e-01 1.03721499e+00 -4.68928337e-01
8.20651531e-01 5.42398691e-02 3.94356430e-01 6.74306095e-01
1.65966749e-01 2.58541524e-01 -5.33222318e-01 -1.71878427e-01
3.29851657e-01 4.86432493e-01 -2.86177784e-01 -6.42933190e-01
-7.41401792e-01 8.30875814e-01 8.49833429e-01 2.25035742e-01
-2.61970133e-01 -6.94767833e-02 6.74997717e-02 1.20322462e-02
4.50984150e-01 3.14673185e-01 -1.16308317e-01 5.08676730e-02
-9.93552208e-01 -8.17345362e-03 5.85373819e-01 3.87516797e-01
1.09521115e+00 -3.63241285e-01 -7.00933158e-01 8.65697622e-01
2.52083480e-01 2.96087682e-01 5.57180941e-01 -6.03792608e-01
8.80990401e-02 1.12488902e+00 -1.09319136e-01 -8.21544409e-01
1.63132459e-01 -3.68045956e-01 -4.48778927e-01 4.19072717e-01
3.63886356e-01 3.98144513e-01 -1.10533226e+00 1.56016123e+00
4.72568989e-01 3.20379138e-01 -3.55918929e-02 1.14553785e+00
8.87947261e-01 6.93310857e-01 -2.15320721e-01 -2.88883150e-01
1.26027882e+00 -8.78824174e-01 -4.26934421e-01 -1.27833188e-01
3.43429238e-01 -2.93241769e-01 1.10845733e+00 2.66537517e-01
-8.29336703e-01 -3.98375720e-01 -1.41869962e+00 1.22252651e-01
-5.42904317e-01 2.06764355e-01 2.01741084e-01 6.53307676e-01
-3.14887047e-01 7.91790426e-01 -7.31831074e-01 -2.44713828e-01
7.73616850e-01 7.52424598e-02 -6.96309935e-03 -2.28802130e-01
-7.69931197e-01 4.72729534e-01 3.30038339e-01 -1.84797272e-01
-9.02012765e-01 -4.01323497e-01 -9.13153708e-01 3.68136764e-01
5.75205803e-01 -1.57603472e-01 8.50512266e-01 -9.39285278e-01
-1.19019544e+00 8.39612722e-01 -1.88623711e-01 -2.43544281e-01
3.74785811e-01 1.49088949e-01 -5.01468778e-01 3.58920634e-01
4.41537052e-02 4.07966852e-01 1.01330900e+00 -1.28822541e+00
-6.83267355e-01 -4.83785570e-01 -7.85587803e-02 1.33918986e-01
-5.76033413e-01 -2.01612070e-01 -3.78775537e-01 -5.74742258e-01
3.60050499e-01 -6.17694080e-01 2.46201735e-02 4.33029592e-01
-4.80719209e-01 -2.43007869e-01 1.01060569e+00 -1.30127609e-01
1.37042391e+00 -2.35881495e+00 -8.31728876e-02 3.74596208e-01
3.18932116e-01 9.75925997e-02 8.94627199e-02 4.68683504e-02
-1.13876332e-02 -6.90919533e-02 -4.20798272e-01 1.16796963e-01
-6.97992146e-02 7.37295076e-02 -4.18984920e-01 5.34911871e-01
5.09690762e-01 6.38025641e-01 -1.20217848e+00 -3.71134102e-01
3.22291851e-01 3.71150017e-01 -3.81369144e-01 2.80921638e-01
-2.14937344e-01 -4.56555374e-02 -4.95627552e-01 8.66280973e-01
5.78146040e-01 -2.24878579e-01 -1.19210884e-01 -2.40479156e-01
4.68020700e-02 1.30284280e-01 -1.34657276e+00 1.07263422e+00
-4.66159023e-02 3.04868519e-01 -1.69308886e-01 -9.23382759e-01
1.00012505e+00 -9.75692719e-02 2.94403762e-01 -5.51547050e-01
2.10648045e-01 2.95049638e-01 9.15806592e-02 -2.82574892e-01
3.94666910e-01 -2.94910908e-01 -9.43601057e-02 3.44077021e-01
1.54260948e-01 -4.18713782e-03 9.18350592e-02 1.22842558e-01
1.22247100e+00 1.55991584e-01 4.60478634e-01 -2.13147834e-01
4.36834812e-01 -3.90714139e-01 8.52008998e-01 7.24207759e-01
-4.79056984e-01 8.55818450e-01 4.26969647e-01 -3.41655105e-01
-4.93520617e-01 -1.30466354e+00 -2.07692012e-01 1.00007725e+00
5.76360703e-01 -4.37327355e-01 -6.12438440e-01 -1.07542872e+00
5.76730296e-02 8.02524328e-01 -8.10965359e-01 -7.48794317e-01
-2.19940066e-01 -5.27197003e-01 -1.51838526e-01 5.34517944e-01
2.22358689e-01 -9.15954351e-01 -9.94646251e-01 -1.12523414e-01
7.15984851e-02 -5.55874527e-01 -7.29584157e-01 6.49939358e-01
-6.16299868e-01 -1.15617013e+00 -4.27333981e-01 -7.02840686e-01
9.93538499e-01 3.37098181e-01 6.88864589e-01 4.86699753e-02
-5.96213818e-01 1.30763769e-01 -5.99840462e-01 -2.78067559e-01
-5.41840261e-03 -4.89079595e-01 -1.10844269e-01 5.96717358e-01
6.63959444e-01 -8.30613822e-02 -6.84678555e-01 5.19955814e-01
-7.11284876e-01 -3.87886435e-01 4.45421517e-01 1.01844966e+00
1.12105691e+00 3.44862826e-02 5.79116523e-01 -6.69893980e-01
4.73588072e-02 -5.13273835e-01 -1.86857551e-01 3.07139933e-01
-7.15162098e-01 -4.64194193e-02 6.78412139e-01 -5.69558144e-01
-6.41389012e-01 1.01019114e-01 2.83754647e-01 -8.06119323e-01
-5.81382103e-02 2.83288985e-01 -5.26525557e-01 -3.22223045e-02
6.14600420e-01 3.32637101e-01 -2.20619425e-01 -2.89685428e-01
2.08907411e-01 6.35545433e-01 7.41239250e-01 -4.05441284e-01
8.02756548e-01 6.62577152e-01 -1.78476378e-01 -6.61209524e-01
-1.15727615e+00 -8.69909585e-01 -5.97004831e-01 -6.23590708e-01
7.39314556e-01 -5.76439142e-01 -4.52661440e-02 2.39806935e-01
-3.21516007e-01 -1.93819925e-01 -9.44042385e-01 2.74081349e-01
-4.48528647e-01 2.00599968e-01 -3.39834057e-02 -1.00363410e+00
-3.08115751e-01 -1.06206882e+00 1.34259868e+00 5.63101590e-01
-2.37356886e-01 -5.54731309e-01 -3.08474321e-02 9.20205861e-02
-1.08340360e-01 3.14155370e-01 7.92867184e-01 -9.97108281e-01
-4.77563322e-01 -6.68206394e-01 -1.43648032e-02 4.94319230e-01
1.52174309e-01 1.59664512e-01 -1.10675001e+00 -4.24737781e-01
1.20354764e-01 -5.76829076e-01 1.33316624e+00 1.91256359e-01
1.58422160e+00 -1.59886196e-01 -4.52720910e-01 6.13185525e-01
1.45555627e+00 2.66450018e-01 5.67521095e-01 -5.85739501e-02
4.49850053e-01 1.92804039e-01 8.25071096e-01 3.81492525e-01
6.55663619e-03 7.09437549e-01 4.42632824e-01 1.45352632e-01
-1.90275759e-01 -3.56261730e-01 5.18750370e-01 3.05815935e-01
2.93048233e-01 -2.63043493e-01 -5.37128985e-01 4.24161106e-01
-1.82233942e+00 -1.21214044e+00 -4.53495979e-02 2.65483236e+00
5.81308126e-01 4.18275952e-01 7.63931572e-02 4.92288083e-01
8.38729322e-01 2.51853019e-01 -8.60549510e-01 -1.02280676e-01
-1.34532064e-01 2.67133445e-01 -1.46966509e-03 7.53332525e-02
-1.28133214e+00 4.31927890e-01 5.05391836e+00 1.08424628e+00
-1.18417823e+00 -6.32906985e-03 7.67100275e-01 -4.09789234e-01
-1.74426183e-01 2.04887077e-01 -7.38342285e-01 5.88204682e-01
5.44981599e-01 1.99003182e-02 1.73037991e-01 9.12780344e-01
-1.38327241e-01 -2.74704069e-01 -1.27013826e+00 8.76852036e-01
3.88187945e-01 -1.13962185e+00 -8.17735866e-02 1.33656710e-01
6.49894655e-01 -1.72328010e-01 3.41463611e-02 4.24766541e-01
5.46943210e-02 -8.36910427e-01 1.00828516e+00 5.89819908e-01
9.06541944e-01 -6.63077056e-01 4.29300159e-01 4.96410131e-01
-1.32291126e+00 -3.42490554e-01 -3.31912547e-01 -4.96725179e-02
-2.43316084e-01 7.00588882e-01 -4.61373657e-01 2.06249297e-01
6.83461547e-01 9.03827310e-01 -4.01540130e-01 1.17395711e+00
-1.77198336e-01 6.98387682e-01 -3.01778734e-01 -6.88823089e-02
1.00867249e-01 -4.63292539e-01 7.27825761e-01 1.02088809e+00
2.35008463e-01 1.73448160e-01 4.73198086e-01 1.05577910e+00
-1.03877626e-01 -2.66679041e-02 -2.98785359e-01 4.15247008e-02
5.14721990e-01 1.57096839e+00 -8.89630854e-01 -4.28466737e-01
-2.64699012e-01 9.05149400e-01 3.67899328e-01 8.69444981e-02
-5.98504245e-01 -8.38054478e-01 4.49045062e-01 1.07196547e-01
4.90864605e-01 3.53699118e-01 -4.67073679e-01 -1.22570777e+00
2.09903196e-01 -2.09366634e-01 7.83925593e-01 -4.91945982e-01
-1.49628186e+00 3.00363511e-01 -3.44263464e-01 -1.53870678e+00
1.50502995e-01 -5.77177405e-01 -1.24471807e+00 5.76048374e-01
-1.52710521e+00 -8.37623000e-01 -6.12524688e-01 2.08117813e-01
4.52534020e-01 -3.56946178e-02 6.33305132e-01 -3.14093456e-02
-9.43820477e-01 7.27572739e-01 3.24199736e-01 1.40265107e-01
4.60188746e-01 -1.45035493e+00 -4.02770340e-02 7.82217741e-01
1.45070300e-01 2.90280998e-01 6.11191988e-01 -5.72813749e-01
-1.06671298e+00 -1.38942409e+00 6.53217912e-01 -3.25302899e-01
4.65674311e-01 -4.95800585e-01 -1.14947748e+00 1.72296554e-01
-4.67903197e-01 7.38369584e-01 7.43256211e-01 -2.69148469e-01
-5.30296266e-01 -1.78032517e-01 -1.34393764e+00 2.09458828e-01
9.30971205e-01 -5.04911542e-01 -6.57216787e-01 2.12980375e-01
1.90119773e-01 -9.56001952e-02 -8.27169716e-01 4.49098527e-01
5.67681968e-01 -9.04702365e-01 8.73679399e-01 -4.77852583e-01
5.76345086e-01 -4.67624754e-01 -4.15533274e-01 -1.23031878e+00
-3.90589476e-01 -1.80481970e-01 -5.94828784e-01 1.18469262e+00
2.38607913e-01 -4.86622840e-01 7.13255525e-01 2.15763584e-01
-1.26917362e-01 -1.20650125e+00 -1.21045387e+00 -1.07292819e+00
-1.48609579e-01 -1.70451015e-01 3.31796139e-01 8.17459524e-01
1.15203775e-01 2.82615900e-01 -2.46803313e-02 3.40905748e-02
5.81535220e-01 5.14701068e-01 3.60735953e-01 -1.34784126e+00
-3.25741142e-01 -6.76023424e-01 -5.85955918e-01 -8.20129395e-01
-5.12622930e-02 -1.10086036e+00 4.40466076e-01 -1.42072403e+00
6.28670573e-01 -4.13169622e-01 -6.71280086e-01 4.97520059e-01
-4.01917249e-01 3.02910179e-01 -1.57987271e-02 3.10895294e-01
-8.44167769e-01 8.10248017e-01 7.55284309e-01 -1.94679365e-01
-1.95769280e-01 -5.45440167e-02 -5.95109224e-01 8.28631282e-01
3.92000020e-01 -5.77107906e-01 -2.27037579e-01 3.45141947e-01
1.57503933e-02 -2.21832782e-01 6.26103580e-01 -1.27728426e+00
1.10203080e-01 -3.81034404e-01 5.90406299e-01 -5.83721042e-01
4.09032583e-01 -7.15425372e-01 -2.90888399e-01 5.65352857e-01
-5.40212512e-01 -8.96136403e-01 -3.55033815e-01 9.43579018e-01
-8.65030512e-02 -3.23407680e-01 1.26254880e+00 1.84623063e-01
-7.78233051e-01 3.71623218e-01 1.07634306e-01 1.38230309e-01
1.46215534e+00 -6.27236485e-01 -2.40388587e-01 -3.56351621e-02
-6.09025240e-01 2.23599926e-01 6.26015604e-01 3.51423740e-01
8.58179450e-01 -1.47909808e+00 -5.49533188e-01 5.78508496e-01
8.48413885e-01 3.96652780e-02 1.40961632e-01 6.50101602e-01
6.48305267e-02 -7.05669746e-02 2.27795303e-01 -8.26587975e-01
-1.06587899e+00 3.93254220e-01 4.75957632e-01 -2.00027470e-02
-7.24344850e-01 8.97664249e-01 3.22425246e-01 -4.53889072e-01
1.92792058e-01 -1.55882448e-01 -2.49001384e-01 8.60382169e-02
7.78324723e-01 5.18835723e-01 1.88405246e-01 -7.45050788e-01
-3.97340506e-01 5.58768094e-01 -7.52466470e-02 3.26149911e-01
1.31581235e+00 2.42396474e-01 2.13413998e-01 7.48128533e-01
1.30065382e+00 -2.85756230e-01 -1.52795255e+00 -4.44924206e-01
-1.60648406e-01 -7.19842315e-01 3.68872851e-01 -1.05625749e+00
-9.41260993e-01 7.72091329e-01 9.31349516e-01 1.53733902e-02
1.24630725e+00 2.77239025e-01 7.49648631e-01 1.93504825e-01
2.27350980e-01 -1.33724117e+00 4.86506701e-01 3.76952291e-01
5.18307805e-01 -1.25951910e+00 -1.79307893e-01 -3.09197128e-01
-5.15366495e-01 1.14379489e+00 7.52192974e-01 -1.99955404e-01
5.08562446e-01 7.58833066e-02 -3.63301575e-01 -4.70824867e-01
-7.72424996e-01 -1.92966864e-01 6.81696177e-01 4.09005135e-01
3.27377878e-02 2.74958581e-01 3.27115208e-02 9.12660837e-01
1.39099479e-01 -1.04201682e-01 2.74715900e-01 1.04542017e+00
-9.73182559e-01 -4.72224027e-01 -3.28148186e-01 1.19680810e+00
-3.06003839e-02 2.57099450e-01 -5.56823075e-01 1.43474549e-01
2.67945081e-01 8.26800883e-01 3.73082817e-01 -5.42432070e-01
2.15435326e-01 1.83828071e-01 2.59725332e-01 -7.70152271e-01
-3.94325852e-01 3.28172594e-02 -3.03230494e-01 -5.72848082e-01
-1.33107409e-01 -8.66665304e-01 -1.43170595e+00 3.30881745e-01
-7.34558642e-01 -1.62322950e-02 4.18173313e-01 9.08712447e-01
1.62759647e-01 4.29498732e-01 1.01689184e+00 -8.85836482e-01
-8.65212619e-01 -5.89648843e-01 -6.92496777e-01 5.52451432e-01
4.72422153e-01 -9.18021202e-01 -6.73579276e-01 -7.05434978e-02] | [9.664225578308105, 2.1920089721679688] |
93fa1c5d-6cec-4186-94e8-a56b3f8ca745 | from-chaos-comes-order-ordering-event | 2304.13455 | null | https://arxiv.org/abs/2304.13455v2 | https://arxiv.org/pdf/2304.13455v2.pdf | From Chaos Comes Order: Ordering Event Representations for Object Detection | Today, state-of-the-art deep neural networks that process events first convert them into dense, grid-like input representations before using an off-the-shelf network. However, selecting the appropriate representation for the task traditionally requires training a neural network for each representation and selecting the best one based on the validation score, which is very time-consuming. In this work, we eliminate this bottleneck by selecting the best representation based on the Gromov-Wasserstein Discrepancy (GWD) between the raw events and their representation. It is approximately 200 times faster to compute than training a neural network and preserves the task performance ranking of event representations across multiple representations, network backbones, and datasets. This means that finding a representation with a high task score is equivalent to finding a representation with a low GWD. We use this insight to, for the first time, perform a hyperparameter search on a large family of event representations, revealing new and powerful representations that exceed the state-of-the-art. On object detection, our optimized representation outperforms existing representations by 1.9% mAP on the 1 Mpx dataset and 8.6% mAP on the Gen1 dataset and even outperforms the state-of-the-art by 1.8% mAP on Gen1 and state-of-the-art feed-forward methods by 6.0% mAP on the 1 Mpx dataset. This work opens a new unexplored field of explicit representation optimization for event-based learning methods. | ['Davide Scaramuzza', 'Mathias Gehrig', 'Daniel Gehrig', 'Nikola Zubić'] | 2023-04-26 | null | null | null | null | ['event-based-vision'] | ['computer-vision'] | [ 3.31162214e-01 7.69610927e-02 -1.85636103e-01 -4.21116620e-01
-1.04786968e+00 -3.49313051e-01 7.51890004e-01 6.60230517e-01
-6.42232180e-01 7.66008675e-01 1.83574721e-01 -3.50971036e-02
-2.62411624e-01 -1.20217860e+00 -7.93009818e-01 -6.76297426e-01
-2.42046893e-01 8.56838703e-01 3.64562333e-01 2.08094995e-02
5.27063832e-02 5.91119885e-01 -1.75274718e+00 5.77497959e-01
5.07873952e-01 1.22913730e+00 1.65838208e-02 3.89765263e-01
-1.51833922e-01 4.41592693e-01 -6.42213702e-01 -2.17157498e-01
9.88424942e-02 -4.13048953e-01 -6.48485363e-01 -4.32162851e-01
3.20683450e-01 5.07840738e-02 -5.75135767e-01 7.32478559e-01
6.52049899e-01 5.52841842e-01 8.09173644e-01 -1.00043011e+00
-6.74937189e-01 7.07912922e-01 -4.32707429e-01 4.15276647e-01
-1.85020342e-02 -1.09613404e-01 1.13051343e+00 -7.88964570e-01
7.25870907e-01 1.03645897e+00 6.97781980e-01 4.23130363e-01
-1.41660583e+00 -7.23882377e-01 1.65724292e-01 3.23563367e-01
-1.35423326e+00 -1.67675957e-01 3.09511542e-01 -3.79281104e-01
1.50678408e+00 9.47808996e-02 6.08098090e-01 1.33227336e+00
1.21617794e-01 7.17172384e-01 5.83385110e-01 -3.61625671e-01
4.85745311e-01 -2.43035734e-01 3.77600312e-01 6.05489552e-01
3.39516699e-01 2.86828995e-01 -7.84851670e-01 -1.68796793e-01
5.08675754e-01 2.64903218e-01 -1.22064322e-01 -9.75661352e-02
-1.15617287e+00 1.04368198e+00 6.81519985e-01 2.85208136e-01
-7.49707043e-01 6.24419212e-01 6.19131207e-01 -3.27915996e-02
4.20408040e-01 6.61127746e-01 -5.39278865e-01 -3.02099109e-01
-9.66771424e-01 3.69093388e-01 5.82542062e-01 3.18405569e-01
7.10089862e-01 1.57127276e-01 -3.76394719e-01 9.85995650e-01
-5.86535409e-02 -5.79380430e-03 7.04424560e-01 -5.31154811e-01
2.97640741e-01 6.81970835e-01 -1.76494971e-01 -6.40140176e-01
-6.76386476e-01 -7.20050275e-01 -9.07905519e-01 1.67270333e-01
5.16089499e-01 -8.87448117e-02 -1.06678700e+00 1.83526444e+00
9.99709219e-02 2.14662686e-01 -1.85866803e-02 7.57021785e-01
8.35012734e-01 9.47023988e-01 3.76745641e-01 -5.22341207e-02
1.45930326e+00 -6.13993168e-01 -3.77542526e-01 -4.28923637e-01
5.95462143e-01 -3.97181034e-01 7.87164032e-01 4.54062164e-01
-9.61930811e-01 -5.12090802e-01 -1.19152892e+00 1.69642456e-02
-5.21645069e-01 2.66850591e-01 1.00849056e+00 3.64507705e-01
-7.26796746e-01 9.32991862e-01 -9.01108980e-01 -5.20929754e-01
4.63584840e-01 3.25395644e-01 -3.58343422e-01 -1.19183913e-01
-1.41439593e+00 1.03617394e+00 9.06364620e-01 -3.50304991e-01
-1.02991831e+00 -1.12280309e+00 -7.15731382e-01 5.30296624e-01
2.81202823e-01 -4.84807551e-01 1.09053254e+00 -6.67910635e-01
-1.20992541e+00 7.91122317e-01 -1.54707057e-03 -9.80021715e-01
1.59725562e-01 -2.91140586e-01 -3.27693760e-01 1.48886412e-01
-3.18854600e-02 7.36429930e-01 4.27233875e-01 -7.33698070e-01
-6.68741286e-01 -9.87957716e-02 -2.59324193e-01 -5.11148088e-02
-2.28751883e-01 -1.58232525e-02 -3.46628159e-01 -6.62182808e-01
7.79583156e-02 -7.78063178e-01 -2.22812369e-01 1.10419393e-01
-1.33699596e-01 -5.26400506e-01 4.20703202e-01 -3.87767375e-01
1.06505597e+00 -2.18236041e+00 1.17727935e-01 1.27633885e-01
9.84093845e-02 2.30025232e-01 -2.46478736e-01 5.47585070e-01
-4.51633066e-01 -1.23556957e-01 -2.62680471e-01 -4.94047761e-01
-2.10664626e-02 2.76475042e-01 -4.82969284e-01 5.25409937e-01
5.17020345e-01 7.62968361e-01 -1.00111032e+00 -2.46594459e-01
1.88429952e-01 5.80704451e-01 -4.78266627e-01 1.10894918e-01
-1.83498994e-01 -2.93248445e-02 -2.38948971e-01 3.85794997e-01
3.54513317e-01 -4.62246120e-01 2.61551350e-01 -1.11551955e-01
-8.39862600e-02 6.10688269e-01 -1.11655295e+00 1.79838252e+00
-4.04076964e-01 7.70452559e-01 -8.04538131e-01 -1.29239583e+00
1.13795805e+00 3.70496899e-01 5.76563358e-01 -6.52400136e-01
7.97370002e-02 1.93587095e-01 -2.35398300e-02 1.37553528e-01
4.98944521e-01 -2.48408765e-01 -1.88553348e-01 5.84085047e-01
4.55510110e-01 4.50486541e-01 3.33798796e-01 1.34597197e-01
1.37087047e+00 -2.52633691e-02 5.26017845e-01 -9.42310691e-02
-8.43489766e-02 -1.91568062e-01 5.98394334e-01 9.70371366e-01
1.76178113e-01 9.73891199e-01 6.21335506e-01 -7.24868059e-01
-7.65478373e-01 -1.22538877e+00 -3.87600362e-01 1.32889485e+00
-3.46376270e-01 -5.62990308e-01 -4.54384804e-01 -5.40465236e-01
1.02939948e-01 1.01093423e+00 -9.57785487e-01 -5.51106155e-01
-5.50299823e-01 -1.22790754e+00 8.09408784e-01 8.73452902e-01
3.45681161e-01 -1.18972301e+00 -7.38126636e-01 5.01435399e-01
-7.28495345e-02 -6.83367193e-01 1.42748117e-01 6.92658961e-01
-7.44938374e-01 -9.46087956e-01 -6.75286412e-01 -3.23920786e-01
3.51951510e-01 -1.34550452e-01 1.35384440e+00 -1.35887861e-01
-4.31771189e-01 -2.41267860e-01 -3.96220744e-01 -4.59023774e-01
-3.95509340e-02 4.19183910e-01 -1.92380562e-01 3.82054895e-02
4.00135666e-01 -6.52907193e-01 -6.05145752e-01 6.09335564e-02
-7.63331532e-01 -1.31457329e-01 6.41162574e-01 9.05702293e-01
8.96898091e-01 -5.96341714e-02 8.46555233e-01 -6.40593171e-01
3.27545226e-01 -9.03470278e-01 -6.00033164e-01 2.16096938e-01
-5.25055945e-01 3.39161336e-01 5.13521433e-01 -6.23009264e-01
-9.62635458e-01 2.56202906e-01 2.94937417e-02 -4.03087288e-01
-1.17664319e-02 6.01901829e-01 2.91287988e-01 4.37276870e-01
1.18173778e+00 1.05849288e-01 -4.15895939e-01 -5.44350982e-01
4.15121078e-01 2.00527027e-01 6.09277427e-01 -6.44819736e-01
2.59137690e-01 3.71409386e-01 1.82569295e-01 -4.62465793e-01
-1.16042471e+00 -4.87105757e-01 -3.56550395e-01 -3.90631752e-03
9.68695223e-01 -7.80583262e-01 -5.72170615e-01 1.75401926e-01
-1.25748599e+00 -2.88084149e-01 -6.51694655e-01 5.88183463e-01
-4.25737053e-01 -2.90282995e-01 -4.41399485e-01 -6.74990296e-01
-4.24588293e-01 -9.49678719e-01 1.12106049e+00 8.84998664e-02
-4.74991292e-01 -6.71293318e-01 2.99348176e-01 -1.32782504e-01
4.12539512e-01 4.50177729e-01 9.62756813e-01 -9.47355151e-01
-4.68987465e-01 -4.54636335e-01 -4.63380307e-01 3.75980325e-02
-1.79526418e-01 -1.71043828e-01 -1.02026749e+00 -5.48284687e-02
-4.58030552e-01 -3.10229838e-01 1.52023399e+00 5.88265240e-01
1.41641724e+00 1.15905181e-01 -4.41732585e-01 7.48162866e-01
1.34935331e+00 2.31086373e-01 8.32987189e-01 4.84511316e-01
3.92867535e-01 3.31377357e-01 2.62126088e-01 6.12192392e-01
7.46629760e-02 7.31438637e-01 4.63494599e-01 1.28156081e-01
-3.13911676e-01 -2.44886264e-01 2.05228135e-01 5.67199774e-02
5.90923242e-02 -4.43276405e-01 -1.08073843e+00 7.69601703e-01
-2.11899018e+00 -1.04641390e+00 4.12466004e-02 2.26921201e+00
7.95379639e-01 2.16253161e-01 -5.39407972e-03 2.81085193e-01
5.27178824e-01 2.20801100e-01 -5.06946862e-01 -2.80146301e-01
-9.14524272e-02 6.67129338e-01 3.61364335e-01 -9.08181667e-02
-1.24230027e+00 8.12842846e-01 6.17470121e+00 9.83704805e-01
-1.14297533e+00 9.16354358e-02 6.95950329e-01 -5.59350193e-01
9.68194604e-02 -1.55651987e-01 -1.01274061e+00 4.63311821e-01
1.46114945e+00 -1.88847050e-01 2.87330419e-01 9.01989043e-01
-7.28743449e-02 -1.77526340e-01 -1.36688781e+00 1.06756532e+00
-4.56562266e-03 -1.82656920e+00 6.81769624e-02 -5.22677265e-02
6.09168470e-01 4.63860303e-01 -1.61379755e-01 6.30305111e-01
3.47789824e-01 -1.25313628e+00 6.91594839e-01 5.18810749e-01
6.75035000e-01 -8.76019597e-01 6.62654877e-01 7.33089149e-02
-1.25528240e+00 -4.42872010e-02 -5.52115023e-01 8.49670023e-02
2.58629471e-01 8.76105905e-01 -7.44724333e-01 3.95008355e-01
7.37927079e-01 7.78010666e-01 -6.10458732e-01 1.20017993e+00
-5.06015867e-02 7.96121120e-01 -5.54034114e-01 6.36875853e-02
3.87318850e-01 1.15709811e-01 2.93986201e-01 1.46713722e+00
5.84731340e-01 -6.41798181e-03 1.65031403e-01 7.81545758e-01
-4.69430983e-01 -1.38305500e-01 -5.38638055e-01 3.76745760e-02
5.00659287e-01 1.06657839e+00 -7.40107954e-01 -5.41582644e-01
-8.09483677e-02 7.20182240e-01 6.83853030e-01 1.82267576e-01
-8.51086318e-01 -5.07900417e-01 6.90077007e-01 -1.99279957e-03
5.17357051e-01 8.43764171e-02 -2.63537675e-01 -8.35706592e-01
-9.66255143e-02 -5.58731496e-01 7.98191965e-01 -6.48373246e-01
-1.39325070e+00 6.61809266e-01 1.86033994e-01 -1.05671763e+00
-4.79346484e-01 -4.63802904e-01 -6.82943106e-01 8.24333012e-01
-1.37210393e+00 -7.36266971e-01 -2.11877227e-01 2.37445503e-01
4.32065994e-01 -2.82539278e-01 1.12503505e+00 4.88455057e-01
-6.51842356e-01 5.55323839e-01 2.40173548e-01 4.71823178e-02
5.27258337e-01 -1.09033322e+00 4.86638248e-01 5.37309647e-01
2.65278727e-01 4.80160892e-01 4.20076638e-01 -3.93011689e-01
-1.15183473e+00 -1.39486170e+00 8.62421274e-01 -1.09965444e-01
5.61777890e-01 -2.62262762e-01 -1.23156178e+00 7.59584844e-01
-9.31567922e-02 3.38070452e-01 8.08734596e-01 6.03257358e-01
-6.56509101e-01 -1.68302655e-01 -1.00533628e+00 4.10007238e-01
1.08378899e+00 -3.72502774e-01 -6.78642750e-01 5.28040528e-01
6.12370670e-01 -1.88867584e-01 -9.32323813e-01 2.18042567e-01
4.61271614e-01 -7.07971215e-01 1.15004480e+00 -8.49376976e-01
3.69883358e-01 -1.36712655e-01 -3.12374711e-01 -1.44087195e+00
-5.08408189e-01 -4.36472446e-01 -3.13454479e-01 1.10207891e+00
6.05196416e-01 -6.22796357e-01 6.48132503e-01 2.40634799e-01
-2.97989130e-01 -9.23857033e-01 -1.16014540e+00 -8.41575801e-01
-1.66743305e-02 -6.27505600e-01 6.52393460e-01 8.16380680e-01
-3.02029997e-01 4.25539285e-01 -1.36675090e-01 1.07064381e-01
4.97317344e-01 4.04445715e-02 2.62908220e-01 -1.64020658e+00
-3.62711370e-01 -5.65591633e-01 -5.33918917e-01 -6.62129164e-01
2.05925107e-01 -1.21762371e+00 6.51094457e-03 -1.58022261e+00
1.43102005e-01 -4.63295400e-01 -7.06281722e-01 8.11851025e-01
-4.96175624e-02 2.19946444e-01 1.61404550e-01 6.19566329e-02
-5.47129571e-01 6.04694545e-01 4.73832250e-01 -4.20079455e-02
-8.80470276e-02 -3.05327773e-01 -7.41687357e-01 6.08737588e-01
1.00149095e+00 -8.82084310e-01 -3.01478654e-01 -4.03806895e-01
4.26425964e-01 6.25731647e-02 4.44122016e-01 -1.22110212e+00
6.29161857e-03 -2.20271070e-02 7.34635115e-01 -7.04170048e-01
7.09182680e-01 -3.60919744e-01 2.31605574e-01 4.30688024e-01
-6.08057797e-01 -6.91737384e-02 4.61169809e-01 6.62795424e-01
-1.50200576e-01 -4.42271262e-01 8.42605770e-01 -1.22077286e-01
-6.91673517e-01 3.06275606e-01 -4.45982665e-01 1.43813789e-01
9.43773568e-01 5.20925075e-02 -4.68324661e-01 -3.63840722e-02
-6.43921435e-01 -1.39468938e-01 -9.04126838e-02 3.00491571e-01
5.50665975e-01 -1.42190349e+00 -9.36426342e-01 1.42285004e-01
2.33316466e-01 -1.17653601e-01 2.22513571e-01 2.89795130e-01
-2.84024447e-01 2.72463053e-01 -2.60437757e-01 -6.02082551e-01
-9.27070439e-01 2.18201548e-01 3.22823435e-01 -6.88011229e-01
-8.38147402e-01 7.86983192e-01 3.16096514e-01 -2.00730413e-01
1.30505651e-01 -3.19140404e-01 -1.05937481e-01 3.87488514e-01
7.35141098e-01 5.24011970e-01 3.75063360e-01 -1.78449109e-01
-4.33525145e-01 2.58362889e-01 -1.92337140e-01 -6.48100153e-02
1.59046125e+00 7.20953822e-01 1.46136776e-01 3.69133353e-01
1.34044504e+00 -7.66314387e-01 -1.22695756e+00 -2.12633297e-01
1.85391828e-02 -3.25970948e-01 3.08771998e-01 -8.84922862e-01
-1.14758015e+00 9.20575500e-01 6.59731805e-01 1.47907004e-01
1.04512763e+00 1.65895075e-01 7.14905620e-01 6.56518877e-01
1.26022562e-01 -8.54024291e-01 1.91665068e-01 6.79948330e-01
1.11184430e+00 -9.98655200e-01 1.00904405e-01 5.40477224e-02
-5.99090874e-01 1.02939737e+00 4.06138122e-01 -5.11404097e-01
6.48762465e-01 -3.33643258e-02 -4.39448833e-01 -3.48720402e-01
-1.12292314e+00 -2.29308873e-01 5.37666500e-01 5.24279952e-01
3.69595766e-01 1.82452708e-01 -1.03318535e-01 6.47096217e-01
-2.17270941e-01 1.40743321e-02 1.84256136e-01 6.95569038e-01
-4.21768814e-01 -9.77910042e-01 -9.39495862e-02 8.02074611e-01
-4.58160698e-01 -2.33482137e-01 -1.86195262e-02 8.26854706e-01
3.50437872e-02 6.61831081e-01 4.64773059e-01 -1.55935481e-01
5.84859669e-01 4.83871818e-01 3.02177519e-01 -6.68677151e-01
-8.63813818e-01 -3.20797741e-01 2.79217988e-01 -6.26017869e-01
2.27799881e-02 -7.56680429e-01 -1.51099432e+00 -2.89260447e-01
-1.11678600e-01 -1.22689180e-01 6.21884704e-01 6.98109925e-01
5.39768636e-01 8.65838885e-01 8.19620714e-02 -1.09503520e+00
-5.28240085e-01 -9.20760095e-01 -3.28594297e-01 3.57476145e-01
1.61397103e-02 -9.84571218e-01 -2.05424055e-01 -2.50452220e-01] | [9.845320701599121, 2.946431875228882] |
7d3f9653-77cd-4c68-9bc3-8755cb10fd21 | evimo2-an-event-camera-dataset-for-motion | 2205.03467 | null | https://arxiv.org/abs/2205.03467v1 | https://arxiv.org/pdf/2205.03467v1.pdf | EVIMO2: An Event Camera Dataset for Motion Segmentation, Optical Flow, Structure from Motion, and Visual Inertial Odometry in Indoor Scenes with Monocular or Stereo Algorithms | A new event camera dataset, EVIMO2, is introduced that improves on the popular EVIMO dataset by providing more data, from better cameras, in more complex scenarios. As with its predecessor, EVIMO2 provides labels in the form of per-pixel ground truth depth and segmentation as well as camera and object poses. All sequences use data from physical cameras and many sequences feature multiple independently moving objects. Typically, such labeled data is unavailable in physical event camera datasets. Thus, EVIMO2 will serve as a challenging benchmark for existing algorithms and rich training set for the development of new algorithms. In particular, EVIMO2 is suited for supporting research in motion and object segmentation, optical flow, structure from motion, and visual (inertial) odometry in both monocular or stereo configurations. EVIMO2 consists of 41 minutes of data from three 640$\times$480 event cameras, one 2080$\times$1552 classical color camera, inertial measurements from two six axis inertial measurement units, and millimeter accurate object poses from a Vicon motion capture system. The dataset's 173 sequences are arranged into three categories. 3.75 minutes of independently moving household objects, 22.55 minutes of static scenes, and 14.85 minutes of basic motions in shallow scenes. Some sequences were recorded in low-light conditions where conventional cameras fail. Depth and segmentation are provided at 60 Hz for the event cameras and 30 Hz for the classical camera. The masks can be regenerated using open-source code up to rates as high as 200 Hz. This technical report briefly describes EVIMO2. The full documentation is available online. Videos of individual sequences can be sampled on the download page. | ['Yiannis Aloimonos', 'Cornelia Fermüller', 'Anton Mitrokhin', 'Levi Burner'] | 2022-05-06 | null | null | null | null | ['motion-segmentation'] | ['computer-vision'] | [ 1.56437978e-01 -3.76275539e-01 -1.59788355e-01 -1.67534307e-01
-7.17025459e-01 -8.52182925e-01 2.72448748e-01 -4.02302712e-01
-5.47799528e-01 6.71003520e-01 -2.05140397e-01 -1.06586918e-01
2.82589883e-01 -4.94459987e-01 -6.76379383e-01 -5.78574955e-01
1.12754166e-01 1.95546061e-01 4.91850823e-01 3.67608577e-01
1.76736012e-01 4.83731061e-01 -1.59180892e+00 -2.87378803e-02
2.23827422e-01 1.10966122e+00 5.49442351e-01 1.30882084e+00
3.98166090e-01 7.39997804e-01 -3.83054197e-01 -2.02772766e-01
5.38044930e-01 -2.20400631e-01 -3.70403945e-01 5.23809493e-01
8.55548322e-01 -7.68378973e-01 -5.58563113e-01 8.47597003e-01
6.00577295e-01 2.24026158e-01 2.85018057e-01 -1.43115795e+00
-1.16278566e-01 -1.78720132e-01 -4.93652195e-01 4.06552464e-01
8.13667655e-01 6.25129879e-01 8.13201427e-01 -8.74024868e-01
8.37767124e-01 8.83379161e-01 6.19167030e-01 4.33393091e-01
-1.03433859e+00 -5.08439481e-01 -1.80031255e-01 3.53459507e-01
-1.42655873e+00 -6.40911102e-01 5.89775622e-01 -6.79589629e-01
1.01438630e+00 1.49346679e-01 8.16574037e-01 1.22779918e+00
2.92002410e-02 7.39864290e-01 6.96084738e-01 -3.16005424e-02
2.58331031e-01 -8.31091329e-02 -3.69063169e-02 6.24743819e-01
3.43101382e-01 2.24684685e-01 -5.50762653e-01 -4.18787263e-03
1.06985736e+00 1.13247491e-01 -5.26475787e-01 -4.63994294e-01
-1.57203197e+00 5.94226480e-01 7.62679949e-02 -2.17841581e-01
-1.96767122e-01 4.90819424e-01 2.37140030e-01 1.78579882e-01
8.45417753e-02 1.63458198e-01 -5.02176702e-01 -6.07969999e-01
-6.00030839e-01 2.46484041e-01 5.60224891e-01 1.51918828e+00
1.06267357e+00 2.25326225e-01 4.15520310e-01 4.20358181e-01
3.06928575e-01 8.96125376e-01 4.20112580e-01 -1.68470395e+00
6.40023470e-01 3.14864159e-01 3.65737677e-01 -1.08316863e+00
-3.61924350e-01 1.81092441e-01 -4.94787157e-01 1.41282827e-01
4.24199015e-01 -3.77867967e-01 -5.38761020e-01 1.14436114e+00
4.27206039e-01 4.18961257e-01 -7.07319006e-02 1.15224171e+00
9.25078511e-01 7.09943712e-01 -5.27135849e-01 -3.15240234e-01
1.20319164e+00 -9.30386066e-01 -5.99560320e-01 -4.40308332e-01
5.07991731e-01 -9.52807426e-01 9.91613328e-01 5.87830126e-01
-1.18695068e+00 -7.56019413e-01 -9.43132341e-01 -1.58837914e-01
-1.67593226e-01 1.95525318e-01 5.10107815e-01 7.12227881e-01
-9.87663805e-01 3.00013989e-01 -1.02937508e+00 -3.36429536e-01
3.90230026e-03 3.41468751e-01 -3.28585625e-01 -3.10116053e-01
-9.01066124e-01 5.55744410e-01 2.46547401e-01 7.05153197e-02
-9.16590393e-01 -5.21802306e-01 -1.26120210e+00 -4.30475384e-01
6.18396103e-01 -6.82014287e-01 1.32406795e+00 -7.91700184e-01
-1.56535745e+00 9.20524418e-01 -1.23767391e-01 -2.10660294e-01
6.30547702e-01 -4.50251758e-01 -2.78700590e-01 6.85575426e-01
2.16974333e-01 7.72685468e-01 7.63736308e-01 -9.55782712e-01
-7.10075617e-01 -2.54742205e-01 7.68265650e-02 3.03580940e-01
1.91661462e-01 -4.73317727e-02 -1.04535615e+00 -3.38338882e-01
-4.72769327e-02 -1.21354830e+00 -7.26601854e-02 1.34895846e-01
-2.43267402e-01 3.99135172e-01 1.11030161e+00 -4.73200828e-01
9.46378708e-01 -2.30521607e+00 -2.82610692e-02 -3.24434936e-01
-1.59527436e-02 1.68052092e-01 1.32444231e-02 1.29513279e-01
9.51088890e-02 -2.90177435e-01 -2.47920081e-02 -5.14914870e-01
-2.31415004e-01 3.62373859e-01 -2.74419435e-03 8.00315142e-01
-2.67797112e-01 7.91160762e-01 -9.22662973e-01 -5.15985072e-01
9.60604072e-01 1.96986958e-01 -5.11278927e-01 2.06032634e-01
7.69960433e-02 6.25345409e-01 -1.48011759e-01 9.95796859e-01
6.46584034e-01 -1.92001775e-01 -2.14239791e-01 -3.29102069e-01
-2.56682187e-01 -2.32828215e-01 -1.66884303e+00 1.81420267e+00
-1.79902613e-01 1.10427558e+00 2.09386125e-01 -5.79089403e-01
5.94437301e-01 3.49666864e-01 8.67017150e-01 -3.25310051e-01
1.69710279e-01 1.04122013e-01 -5.83681703e-01 -6.96513891e-01
8.42389286e-01 2.60165364e-01 4.34048586e-02 1.42447993e-01
-1.00510947e-01 -6.13768876e-01 5.68819761e-01 2.10168481e-01
1.02855456e+00 2.93548942e-01 3.20336819e-01 1.55054092e-01
3.55572492e-01 2.80036867e-01 7.77571142e-01 6.41210496e-01
-4.68855917e-01 1.08474255e+00 -1.24807037e-01 -3.90018433e-01
-1.15633106e+00 -1.12223339e+00 -4.34720665e-02 4.46531862e-01
6.80565119e-01 -5.03220379e-01 -5.13040960e-01 -6.36640713e-02
-2.36428574e-01 2.15295434e-01 -1.43402666e-01 2.34967977e-01
-5.67164421e-01 -3.76783460e-01 2.16024533e-01 6.68169677e-01
6.70392275e-01 -8.84584785e-01 -1.04063857e+00 1.44931763e-01
-3.53653580e-01 -1.72032535e+00 -6.16321802e-01 -1.94845155e-01
-8.97146225e-01 -1.50606692e+00 -6.64651513e-01 -3.42644304e-01
2.85760909e-01 8.75904024e-01 1.02572441e+00 -2.51918405e-01
-5.85403919e-01 9.60196912e-01 -4.44583505e-01 -1.40522465e-01
1.70189798e-01 -9.70570296e-02 2.17464596e-01 -8.72605443e-02
3.37538481e-01 -3.31330568e-01 -7.34205782e-01 6.54869318e-01
-8.35685015e-01 2.03989536e-01 1.50492609e-01 4.06721622e-01
5.49138069e-01 -3.51424575e-01 -2.17255041e-01 -2.73770571e-01
-2.92178482e-01 -3.49236965e-01 -9.55291450e-01 -3.56572658e-01
-1.59269631e-01 -6.09026849e-01 4.76527393e-01 -2.80902863e-01
-9.26537275e-01 4.70353961e-01 2.58406520e-01 -8.50503683e-01
-4.18020517e-01 3.20973396e-02 -1.28482059e-01 -1.42792994e-02
5.91683626e-01 1.24788135e-01 -1.65626913e-01 -2.91383535e-01
3.04165632e-01 6.20624304e-01 9.43223238e-01 -2.69769132e-01
5.81495464e-01 8.77452672e-01 -2.29559079e-01 -1.22916365e+00
-5.23413301e-01 -9.18224216e-01 -6.84099317e-01 -5.26535273e-01
1.10492694e+00 -1.20528841e+00 -8.02012563e-01 8.84905338e-01
-1.06538641e+00 -5.15457034e-01 -2.58545160e-01 9.80532110e-01
-8.67197037e-01 5.25027692e-01 -6.77438021e-01 -6.55932248e-01
2.49326248e-02 -1.28761733e+00 1.36227930e+00 5.12464106e-01
-2.61997227e-02 -9.20406640e-01 2.71499772e-02 6.28533781e-01
-6.33815229e-02 2.94813097e-01 -1.84183970e-01 2.11556986e-01
-1.01149714e+00 -2.61910737e-01 -1.14251278e-01 4.09302026e-01
1.26091391e-01 4.48647648e-01 -9.29423034e-01 -4.36630845e-01
4.29862104e-02 -1.89584106e-01 5.49017012e-01 7.04079151e-01
7.10759878e-01 1.89705506e-01 -2.57109284e-01 1.03676116e+00
1.68404257e+00 5.91248989e-01 6.47111356e-01 3.11794847e-01
8.84780228e-01 2.70264000e-01 7.29102671e-01 6.33324981e-01
4.59095180e-01 9.28155661e-01 6.11562371e-01 3.30249444e-02
7.74564371e-02 6.36599362e-02 8.16980958e-01 7.39193499e-01
-3.31318140e-01 -2.43083537e-01 -8.63663673e-01 5.76428652e-01
-1.69997692e+00 -1.07809842e+00 -6.73471808e-01 2.16971421e+00
3.17061335e-01 -1.99731648e-01 1.87158495e-01 1.38698339e-01
8.05236936e-01 1.84899732e-01 -6.06470227e-01 -4.09349576e-02
-1.36423588e-01 -3.44623834e-01 9.23654735e-01 6.17767632e-01
-1.17078459e+00 8.01087260e-01 6.21959829e+00 2.96693593e-01
-9.56855237e-01 2.05579568e-02 2.71820426e-01 -4.84107286e-01
3.14101934e-01 1.32670730e-01 -1.13559473e+00 7.02840805e-01
8.22056770e-01 5.35328910e-02 3.76483172e-01 1.05311692e+00
4.33930963e-01 -6.08609557e-01 -1.05498075e+00 1.62088203e+00
1.42199382e-01 -1.38777566e+00 -4.96510923e-01 2.85529077e-01
8.82214904e-01 3.75942469e-01 -1.99573725e-01 -1.53559521e-01
1.69287756e-01 -3.97798777e-01 8.82355750e-01 4.28887576e-01
8.92113626e-01 -4.61450100e-01 6.36386991e-01 3.20407689e-01
-1.47649252e+00 -1.14186639e-02 -4.80245769e-01 -3.20620716e-01
6.71250045e-01 3.42944175e-01 -3.92604142e-01 6.85422838e-01
8.33997130e-01 1.19207168e+00 -4.97231811e-01 8.76847088e-01
-2.53051091e-02 3.19753826e-01 -5.87591410e-01 3.84679735e-01
2.04200134e-01 -4.28790480e-01 6.16317272e-01 1.05833375e+00
4.49243337e-01 4.23692673e-01 3.17917138e-01 2.80280143e-01
2.92843044e-01 -3.53989810e-01 -9.06710029e-01 3.54896486e-01
3.46008301e-01 1.23413289e+00 -7.23992825e-01 -6.09881938e-01
-6.96400940e-01 1.18261278e+00 -3.68845969e-01 4.12279874e-01
-1.02480400e+00 -2.78159797e-01 7.06267655e-01 2.77130166e-03
4.01077121e-01 -7.91046798e-01 1.32259680e-02 -1.61046350e+00
7.04157585e-03 -6.15036428e-01 3.24997813e-01 -1.29148865e+00
-7.56234705e-01 2.72742122e-01 2.12935418e-01 -1.45868850e+00
-4.45441693e-01 -8.47688019e-01 -2.75013417e-01 3.07256639e-01
-1.16673005e+00 -7.45706439e-01 -1.01445746e+00 9.71974432e-01
8.29025865e-01 -3.84690166e-02 4.30966616e-01 5.88061213e-01
-7.32000768e-01 8.36981460e-04 2.83696353e-01 1.89268842e-01
7.36395717e-01 -1.09391367e+00 2.73169965e-01 1.04098284e+00
2.07805842e-01 9.78691429e-02 5.30730367e-01 -4.52901214e-01
-1.72689176e+00 -9.22707319e-01 4.09426719e-01 -7.19824135e-01
5.29139519e-01 -1.60755262e-01 -4.38623428e-01 9.91066277e-01
-4.17776741e-02 3.84462565e-01 4.21159595e-01 -7.79527307e-01
3.47247213e-01 -1.55263156e-01 -8.57126176e-01 4.20319498e-01
1.13816714e+00 -4.73020971e-01 -3.20463240e-01 2.30613172e-01
5.51816702e-01 -8.59730601e-01 -7.14661360e-01 -7.44553423e-03
5.03907084e-01 -1.23529255e+00 1.09469974e+00 -7.93572236e-03
2.73637742e-01 -6.15666330e-01 -5.95376730e-01 -7.82196403e-01
5.99676035e-02 -7.79148579e-01 -3.04103017e-01 9.06302571e-01
-6.34472519e-02 -5.01323342e-01 9.08361554e-01 6.30125403e-01
-2.45108888e-01 -1.59866571e-01 -8.88695121e-01 -8.83718848e-01
-5.91194630e-01 -8.75250161e-01 4.39107418e-02 7.77497888e-01
-3.77007037e-01 2.99180686e-01 -5.62107086e-01 1.72545061e-01
8.69232655e-01 2.51189992e-02 1.26345038e+00 -8.27425003e-01
-3.20625365e-01 -5.38668670e-02 -7.99008012e-01 -1.51650059e+00
-1.70318604e-01 -2.80986816e-01 -1.76171020e-01 -1.26571524e+00
-6.81645349e-02 4.23745078e-04 4.82368797e-01 -1.60007793e-02
-3.89522240e-02 6.32997155e-01 4.45207000e-01 4.03427362e-01
-5.46502531e-01 2.84991235e-01 9.79210854e-01 3.06855775e-02
-9.80920792e-02 -1.05353430e-01 -4.83273081e-02 1.03283167e+00
5.61989665e-01 -1.81608975e-01 -4.29283291e-01 -6.12378299e-01
8.56489018e-02 4.21375543e-01 6.04618371e-01 -1.35749090e+00
4.26022232e-01 -2.77752548e-01 2.71340370e-01 -8.78111780e-01
7.41402566e-01 -9.88341749e-01 6.38121724e-01 4.43244785e-01
2.84325480e-01 3.94740105e-01 8.24039951e-02 6.06674075e-01
-2.53281176e-01 -2.28081927e-01 8.57490122e-01 -3.78563493e-01
-1.31451166e+00 3.76385182e-01 -5.04520357e-01 1.59701362e-01
1.38893950e+00 -8.92290056e-01 -1.52325809e-01 -5.23962259e-01
-8.09729636e-01 1.28810689e-01 9.54674840e-01 3.31752688e-01
6.68770313e-01 -1.37257123e+00 -4.17674690e-01 1.84151366e-01
-1.92114385e-04 3.47787321e-01 4.01972413e-01 9.00121987e-01
-1.01471555e+00 5.18643022e-01 -1.18166089e-01 -1.11524248e+00
-1.31547010e+00 4.21874553e-01 2.72886693e-01 3.75403464e-01
-7.65998244e-01 5.28693318e-01 2.46075153e-01 8.98877904e-03
8.69832095e-03 -5.20100296e-01 2.82078534e-01 -2.18122900e-02
6.66003942e-01 8.88187885e-01 -1.55966640e-01 -8.05890858e-01
-4.44687337e-01 9.99307930e-01 3.82903963e-01 -3.73836070e-01
9.29581702e-01 -7.36391008e-01 3.31883371e-01 6.58751786e-01
1.32693076e+00 -1.07707098e-01 -1.81331754e+00 -3.77844572e-02
-4.18190837e-01 -8.71243536e-01 -1.62007675e-01 -1.36809647e-01
-1.08045590e+00 8.59144092e-01 5.85879803e-01 -3.80742997e-02
1.08484340e+00 -4.58579101e-02 6.74884617e-01 2.58522838e-01
5.82415462e-01 -1.12973261e+00 1.59131169e-01 6.00918114e-01
4.34676349e-01 -1.41883707e+00 8.02302510e-02 -3.81255031e-01
-6.81853950e-01 1.09157360e+00 6.62945330e-01 -6.62407875e-02
2.66230404e-01 3.02037954e-01 1.62668660e-01 -8.34560543e-02
-5.66876471e-01 -1.05577089e-01 -2.12966487e-01 6.05493248e-01
1.21009737e-01 -1.30252630e-01 1.53860316e-01 -6.77510053e-02
-8.08224604e-02 1.99725479e-02 9.88243103e-01 1.02791214e+00
-2.90497780e-01 -6.28621995e-01 -6.84692502e-01 4.31341156e-02
-2.78809845e-01 9.74920690e-02 -7.18354285e-02 9.00221050e-01
1.81326851e-01 1.03030860e+00 3.56578588e-01 -2.23660827e-01
3.20244133e-01 -3.21740776e-01 3.90048325e-01 -5.74850261e-01
-2.32969165e-01 -3.88415866e-02 9.69238952e-02 -9.77259338e-01
-7.10269749e-01 -8.77185106e-01 -1.03654170e+00 -5.33012271e-01
-3.08063835e-01 -2.01197267e-01 7.12297022e-01 6.66807055e-01
3.16321820e-01 1.00900315e-01 6.16689980e-01 -1.54263461e+00
9.72076654e-02 -7.33327746e-01 -6.65290475e-01 3.03561002e-01
4.58444834e-01 -7.20076323e-01 -6.15380228e-01 7.94200957e-01] | [8.279294967651367, -1.8203102350234985] |
692b38fa-30ce-49a9-9116-2e773bd11dd6 | augment-features-beyond-color-for-domain | 2307.01703 | null | https://arxiv.org/abs/2307.01703v1 | https://arxiv.org/pdf/2307.01703v1.pdf | Augment Features Beyond Color for Domain Generalized Segmentation | Domain generalized semantic segmentation (DGSS) is an essential but highly challenging task, in which the model is trained only on source data and any target data is not available. Previous DGSS methods can be partitioned into augmentation-based and normalization-based ones. The former either introduces extra biased data or only conducts channel-wise adjustments for data augmentation, and the latter may discard beneficial visual information, both of which lead to limited performance in DGSS. Contrarily, our method performs inter-channel transformation and meanwhile evades domain-specific biases, thus diversifying data and enhancing model generalization performance. Specifically, our method consists of two modules: random image color augmentation (RICA) and random feature distribution augmentation (RFDA). RICA converts images from RGB to the CIELAB color model and randomizes color maps in a perception-based way for image enhancement purposes. We further this augmentation by extending it beyond color to feature space using a CycleGAN-based generative network, which complements RICA and further boosts generalization capability. We conduct extensive experiments, and the generalization results from the synthetic GTAV and SYNTHIA to the real Cityscapes, BDDS, and Mapillary datasets show that our method achieves state-of-the-art performance in DGSS. | ['Yang Tang', 'Michael Felsberg', 'Pavlo Melnyk', 'Qiyu Sun'] | 2023-07-04 | null | null | null | null | ['image-enhancement'] | ['computer-vision'] | [ 3.91510785e-01 -3.85106094e-02 1.35798194e-02 -3.74538124e-01
-5.38080990e-01 -6.98166430e-01 7.00152814e-01 -3.01725745e-01
-4.51086611e-01 5.05981863e-01 -2.04164088e-01 -4.16138351e-01
4.50200588e-01 -1.19374037e+00 -6.89957261e-01 -9.22903955e-01
4.12246585e-01 2.39013106e-01 4.37077850e-01 -4.98332113e-01
-1.98365033e-01 5.37310064e-01 -1.28266919e+00 5.22403866e-02
1.33297336e+00 1.18806469e+00 2.32929543e-01 3.57111841e-01
-2.92111039e-01 4.84468609e-01 -5.41299164e-01 -3.85743201e-01
5.23338318e-01 -5.24414778e-01 -6.97149396e-01 3.20301145e-01
-3.84822786e-02 -1.77627981e-01 -3.24976653e-01 1.25610471e+00
4.19850379e-01 1.24441177e-01 5.18386424e-01 -1.49553323e+00
-9.89699781e-01 1.95352465e-01 -8.47070575e-01 -2.10519135e-01
-1.91645697e-01 2.94267118e-01 5.98782241e-01 -6.84229016e-01
3.59942466e-01 1.23710072e+00 4.41339344e-01 6.13240480e-01
-1.23853707e+00 -7.95397162e-01 4.41290587e-01 -7.68175945e-02
-1.40889728e+00 -5.64696230e-02 9.66743886e-01 -2.16938868e-01
3.13867480e-01 1.73751533e-01 7.89769769e-01 1.19881129e+00
-5.27483284e-01 8.87169778e-01 1.28519642e+00 -3.52777749e-01
3.93200874e-01 1.31139770e-01 -2.40715250e-01 4.83220011e-01
1.89198002e-01 5.20487763e-02 -1.64319634e-01 3.86831880e-01
9.97540534e-01 7.46469423e-02 -4.10601020e-01 -4.60010469e-01
-1.16420710e+00 7.54076302e-01 9.17386770e-01 1.14681832e-01
-2.12450847e-01 -1.73182309e-01 6.23498224e-02 2.01921314e-01
5.19804120e-01 2.92400271e-01 -4.91097808e-01 2.51875341e-01
-8.28291416e-01 9.98041779e-02 3.42042685e-01 1.04484427e+00
1.01765585e+00 3.30812454e-01 -1.14128232e-01 1.10546660e+00
1.66697353e-01 8.61483514e-01 6.89254582e-01 -6.21226609e-01
4.47592616e-01 7.87106335e-01 -5.88801503e-02 -9.51682687e-01
-4.07351315e-01 -6.81737423e-01 -1.01048899e+00 1.51348844e-01
5.21935284e-01 -1.57228351e-01 -1.59191370e+00 1.95358729e+00
2.68818885e-01 1.58852711e-02 1.85935497e-01 1.07745433e+00
9.67763484e-01 7.87957549e-01 1.91781461e-01 2.68975675e-01
1.05312121e+00 -1.16778040e+00 -5.12141228e-01 -5.18041492e-01
4.71025765e-01 -4.84280199e-01 1.48167288e+00 3.43322635e-01
-7.21057594e-01 -6.53792322e-01 -1.06508982e+00 -5.35035357e-02
-6.87151372e-01 2.60984570e-01 7.27973819e-01 7.59424806e-01
-9.36567366e-01 1.18030518e-01 -6.93795741e-01 -3.43065798e-01
6.52888656e-01 2.92636417e-02 -2.39479378e-01 -3.17284048e-01
-1.23019958e+00 4.26957548e-01 4.87872630e-01 2.00384796e-01
-7.58200526e-01 -4.63284850e-01 -1.01692295e+00 -2.66947988e-02
5.46245813e-01 -4.03890342e-01 8.25657368e-01 -1.17922056e+00
-1.58349931e+00 7.53597617e-01 2.40257069e-01 -2.08885685e-01
5.61225712e-01 2.45668098e-01 -6.58293724e-01 2.45009046e-02
1.53872311e-01 1.03902388e+00 9.40971732e-01 -1.49997735e+00
-5.97075820e-01 -3.08337450e-01 3.81206982e-02 2.87039608e-01
-5.28705239e-01 -4.79639828e-01 -1.00640273e+00 -9.04015183e-01
4.08280224e-01 -9.61646199e-01 -3.35996211e-01 -3.17620076e-02
-5.45825899e-01 3.51851642e-01 7.99769223e-01 -7.19746470e-01
8.97626877e-01 -2.23376346e+00 5.30887283e-02 4.19745117e-01
-1.33054242e-01 4.48298097e-01 -4.97179478e-01 -1.68012064e-02
-1.49780780e-01 1.64723352e-01 -6.69072211e-01 -2.13960841e-01
-7.33310133e-02 2.23680615e-01 -2.24081680e-01 1.66329592e-01
2.99225509e-01 1.03526247e+00 -8.15506399e-01 -3.42046708e-01
2.38408908e-01 4.52474445e-01 -5.22428989e-01 1.27418339e-01
-2.79120475e-01 6.54698730e-01 -3.61218959e-01 8.07352960e-01
1.21122766e+00 -2.01906875e-01 -3.80147668e-03 -2.11268544e-01
4.81354212e-03 -4.92768697e-02 -1.16056228e+00 1.86023605e+00
-4.90793049e-01 2.68784791e-01 -1.29431918e-01 -9.91129100e-01
1.06481099e+00 -3.07685882e-01 1.57017961e-01 -1.07626188e+00
9.81559008e-02 1.78532168e-01 -8.71475860e-02 -1.58889547e-01
4.55080032e-01 3.08086202e-02 -1.47550002e-01 1.33953258e-01
-1.21585317e-02 -2.48058408e-01 1.07952587e-01 2.47997090e-01
5.89999795e-01 4.00541067e-01 -1.63718551e-01 -7.83241838e-02
5.12102306e-01 3.49878408e-02 7.66680777e-01 4.15069938e-01
3.18065053e-03 9.96269345e-01 5.56910694e-01 -4.00976427e-02
-9.40326989e-01 -1.20397627e+00 1.03167213e-01 9.51107919e-01
5.57059109e-01 5.90965413e-02 -9.42590594e-01 -1.01522601e+00
-5.31343706e-02 8.33707154e-01 -7.98973441e-01 -4.21081483e-01
-1.92163184e-01 -1.20841777e+00 4.47664827e-01 7.38041520e-01
1.27938497e+00 -9.16478992e-01 -9.54934880e-02 -9.90673527e-02
-3.09956223e-01 -1.22405386e+00 -4.82377976e-01 7.15945736e-02
-6.99549317e-01 -9.58908677e-01 -1.07779837e+00 -7.72563159e-01
9.58922625e-01 4.18050438e-01 7.58154631e-01 -1.86002597e-01
-1.29952747e-02 7.74510251e-03 -6.05598271e-01 -4.85210657e-01
-2.29157835e-01 1.48078173e-01 -4.83929276e-01 3.60583901e-01
2.28220329e-01 -4.17368710e-01 -7.79102147e-01 5.14570415e-01
-1.14398718e+00 3.41532737e-01 8.10694456e-01 9.19657052e-01
6.56759858e-01 -1.06178243e-02 6.09258413e-01 -1.09298074e+00
2.78830826e-01 -3.66570562e-01 -6.62074447e-01 1.04583353e-01
-6.72462702e-01 -1.95645824e-01 6.01684272e-01 -3.81086409e-01
-1.39727843e+00 2.33649582e-01 -2.96881348e-01 -3.37011516e-01
-2.29818031e-01 3.33827347e-01 -6.73464000e-01 -1.83186337e-01
7.19932377e-01 5.35366356e-01 -1.12830333e-01 -5.36040008e-01
5.04908442e-01 6.78203583e-01 6.29858911e-01 -4.26958561e-01
9.84039187e-01 6.11001611e-01 -1.40914887e-01 -7.48990178e-01
-6.26639426e-01 -1.52398229e-01 -5.93580067e-01 -2.66833771e-02
8.36862624e-01 -1.03058481e+00 -1.56166002e-01 9.81914759e-01
-7.77884007e-01 -6.36883080e-01 -4.88302737e-01 1.53311506e-01
-2.39004031e-01 1.88665047e-01 -3.89882505e-01 -5.18079102e-01
2.35820878e-02 -1.14370847e+00 1.00869989e+00 5.34906745e-01
5.10952175e-01 -7.31308520e-01 -3.19707841e-01 2.02687755e-01
4.94733334e-01 4.90392357e-01 7.94206023e-01 -4.77642179e-01
-5.35135150e-01 -1.61227986e-01 -7.30765760e-01 7.54964054e-01
2.02422172e-01 -2.49064997e-01 -1.11332452e+00 -1.38799831e-01
-3.00695240e-01 -2.48619527e-01 1.07697868e+00 2.73926169e-01
1.52407825e+00 5.37250154e-02 -1.27853066e-01 1.00376892e+00
1.47576070e+00 3.45420510e-01 8.88264596e-01 5.56799054e-01
9.68901753e-01 4.29349393e-01 5.87541699e-01 8.10344592e-02
5.63386321e-01 5.45896411e-01 6.97767675e-01 -9.66069579e-01
-3.94008458e-01 -3.07942629e-01 5.93347512e-02 2.22567618e-01
-1.73793629e-01 -3.69902790e-01 -7.34093547e-01 6.42167389e-01
-1.72277570e+00 -3.76235485e-01 -9.94164646e-02 2.09102941e+00
6.70989811e-01 -3.92167009e-02 2.39942238e-01 1.93931282e-01
7.40063488e-01 8.39331597e-02 -8.57528627e-01 -8.23594257e-02
-4.60388839e-01 3.04194540e-01 7.20882654e-01 6.80593103e-02
-1.11330748e+00 1.22715771e+00 4.94330645e+00 9.11440611e-01
-1.38914871e+00 2.01490939e-01 9.45286036e-01 1.71585590e-01
-3.27848494e-01 -2.08034813e-01 -2.74186343e-01 5.98828852e-01
3.01477849e-01 3.07407588e-01 5.89009464e-01 9.58529472e-01
-1.13932088e-01 6.03223927e-02 -5.80888569e-01 1.02878523e+00
-8.33222196e-02 -1.03665471e+00 2.15806231e-01 9.45612937e-02
8.74797940e-01 -1.19747343e-02 2.69020557e-01 4.09057409e-01
4.31619972e-01 -7.54841983e-01 8.32865179e-01 1.79448947e-01
1.03325939e+00 -8.02630544e-01 7.37282455e-01 3.21744483e-05
-9.78467047e-01 -6.95250407e-02 -3.19694042e-01 3.73292416e-01
-5.70759848e-02 5.30612588e-01 -4.84043181e-01 7.36599207e-01
8.36874485e-01 6.46893859e-01 -8.22808146e-01 9.13617432e-01
-6.53095543e-01 6.65453017e-01 -1.68501183e-01 3.48094314e-01
3.42821419e-01 -3.14823627e-01 2.38078922e-01 1.09590328e+00
2.77923882e-01 -1.18739223e-02 1.43343464e-01 1.00693679e+00
-1.67338863e-01 2.88226325e-02 -1.97563350e-01 6.58046752e-02
3.79166394e-01 1.35421765e+00 -1.00582004e+00 -2.84273058e-01
-4.49261427e-01 1.29952371e+00 -5.52513525e-02 8.23554933e-01
-9.79052901e-01 -6.71032071e-01 5.27641475e-01 1.06939666e-01
3.87008458e-01 -2.68507954e-02 -4.04910594e-01 -1.41803718e+00
-9.42755118e-02 -7.95907915e-01 2.56153733e-01 -8.74836028e-01
-1.06440485e+00 7.78279543e-01 -1.91202506e-01 -1.26163006e+00
1.26236677e-01 -5.68500876e-01 -5.06361723e-01 9.95404780e-01
-1.95746183e+00 -1.66064048e+00 -7.78441310e-01 8.52814853e-01
3.37605387e-01 -1.14997908e-01 5.04983902e-01 5.20527363e-01
-7.84205735e-01 7.39043832e-01 -1.53109443e-03 2.22149670e-01
6.22589231e-01 -1.26375139e+00 5.75805187e-01 1.15584707e+00
-2.14013427e-01 1.80166855e-01 3.10114890e-01 -4.95518982e-01
-9.27181900e-01 -1.67620194e+00 2.40766555e-01 1.03146695e-01
2.99429297e-01 -5.54370582e-01 -9.64878976e-01 5.88519216e-01
-1.37973443e-01 2.26512805e-01 3.30934942e-01 -1.97942391e-01
-3.81383091e-01 -3.18643302e-01 -1.26684475e+00 6.69110477e-01
1.11286557e+00 -1.63060442e-01 -1.08796917e-01 1.24275088e-01
8.13204527e-01 -5.01531839e-01 -4.26448554e-01 3.07089180e-01
2.81740695e-01 -9.53377545e-01 1.00928879e+00 -3.55793923e-01
4.23722893e-01 -5.72218359e-01 -2.82086104e-01 -1.62598491e+00
-2.10855842e-01 -3.17433625e-01 4.85626578e-01 1.55615795e+00
5.01699746e-01 -9.19987321e-01 7.13464439e-01 3.87500793e-01
-3.83022428e-01 -4.30722803e-01 -5.12377441e-01 -8.72481108e-01
1.54186562e-01 -4.71390575e-01 1.27419865e+00 1.15941608e+00
-6.17943943e-01 -1.75488647e-02 -2.00664893e-01 2.81730771e-01
4.51790214e-01 1.75681666e-01 8.72078061e-01 -9.57024038e-01
-8.36739317e-02 -5.35370469e-01 -2.27312893e-01 -1.07932615e+00
-2.21606985e-01 -9.44475710e-01 -2.51183882e-02 -1.52702880e+00
-6.35222793e-02 -8.12077224e-01 -2.78095722e-01 7.15163171e-01
-1.65966764e-01 6.64665818e-01 2.28151977e-01 -5.31457700e-02
-4.10865217e-01 8.21465492e-01 1.61672592e+00 -2.83857882e-01
-4.31921333e-01 -6.96304366e-02 -1.00302649e+00 7.27566361e-01
7.54742324e-01 -2.06820294e-01 -5.81796646e-01 -4.72876221e-01
1.76798552e-02 -4.13003355e-01 5.71265459e-01 -9.45901692e-01
-2.13073209e-01 -2.25108474e-01 6.05307996e-01 -3.29983711e-01
2.26946250e-01 -6.93315506e-01 -1.39164895e-01 2.61284113e-01
-3.20389420e-02 -4.42300022e-01 3.39803427e-01 4.88371193e-01
-3.21520507e-01 2.47402355e-01 1.04014456e+00 -7.80468658e-02
-1.05169153e+00 4.72589076e-01 -5.63104004e-02 7.17569366e-02
8.82950664e-01 -2.80106366e-01 -3.29151958e-01 -3.79060805e-01
-6.16663456e-01 2.58097649e-01 6.35805666e-01 3.84305000e-01
3.82499218e-01 -1.45612967e+00 -5.11862338e-01 6.69392109e-01
3.52784991e-01 3.99541527e-01 4.39686269e-01 6.19222105e-01
-4.23122197e-01 5.47747873e-03 -2.60529697e-01 -6.98497176e-01
-7.26578832e-01 5.08048713e-01 2.61451900e-01 1.59051537e-03
-4.05151814e-01 9.26475406e-01 7.21624732e-01 -7.02634692e-01
-3.07375062e-02 -4.05365169e-01 -3.04404438e-01 6.27640933e-02
2.85694212e-01 2.24882305e-01 2.29826778e-01 -6.52910769e-01
-2.01284334e-01 5.60708106e-01 1.72090456e-01 1.08370706e-02
1.35149646e+00 -3.29610616e-01 1.38366580e-01 -5.64115383e-02
1.02105904e+00 -1.56433791e-01 -1.47049272e+00 -2.89591014e-01
-5.66729605e-01 -5.43372750e-01 2.83274710e-01 -1.28847778e+00
-1.66699409e+00 1.05749893e+00 6.72348559e-01 1.94101945e-01
1.76759291e+00 -1.11109942e-01 8.84765446e-01 -3.70205753e-02
1.83475316e-01 -1.15272427e+00 8.18917751e-02 2.84163117e-01
8.43072653e-01 -1.41933620e+00 -3.53522032e-01 -7.25790441e-01
-9.78219330e-01 8.84683669e-01 7.60752499e-01 5.89675754e-02
4.94818568e-01 -9.51561704e-02 1.44469976e-01 1.14460595e-01
-3.12707424e-02 -5.81156194e-01 3.23806435e-01 8.64627481e-01
-6.92981184e-02 1.04622684e-01 -3.17294151e-02 9.18069482e-01
-1.37074128e-01 -4.36781719e-02 3.23479742e-01 7.23107219e-01
-9.20561403e-02 -1.04030716e+00 -3.56175482e-01 2.49938607e-01
-1.66747924e-02 -1.65485874e-01 -2.02603757e-01 1.16464365e+00
3.96336883e-01 8.37655723e-01 1.61140889e-01 -5.61398566e-01
6.00805044e-01 -1.71318188e-01 2.47746944e-01 -3.65348727e-01
-1.69516772e-01 2.85776824e-01 -1.52990550e-01 -5.55164754e-01
-2.66260266e-01 -6.55540109e-01 -1.34290051e+00 -1.07881494e-01
-2.70589381e-01 -1.19935207e-01 9.62244630e-01 7.35684693e-01
2.84710824e-01 8.74974966e-01 7.51000047e-01 -8.51263762e-01
-1.55025143e-02 -7.57166982e-01 -6.18229628e-01 4.61134851e-01
2.16557086e-01 -7.12875426e-01 -2.66420871e-01 1.41346112e-01] | [9.746969223022461, 1.1393702030181885] |
909bd7a8-b421-4f66-a9d7-1cb9b6c5fde8 | relevance-detection-in-cataract-surgery | 2104.14280 | null | https://arxiv.org/abs/2104.14280v1 | https://arxiv.org/pdf/2104.14280v1.pdf | Relevance Detection in Cataract Surgery Videos by Spatio-Temporal Action Localization | In cataract surgery, the operation is performed with the help of a microscope. Since the microscope enables watching real-time surgery by up to two people only, a major part of surgical training is conducted using the recorded videos. To optimize the training procedure with the video content, the surgeons require an automatic relevance detection approach. In addition to relevance-based retrieval, these results can be further used for skill assessment and irregularity detection in cataract surgery videos. In this paper, a three-module framework is proposed to detect and classify the relevant phase segments in cataract videos. Taking advantage of an idle frame recognition network, the video is divided into idle and action segments. To boost the performance in relevance detection, the cornea where the relevant surgical actions are conducted is detected in all frames using Mask R-CNN. The spatiotemporally localized segments containing higher-resolution information about the pupil texture and actions, and complementary temporal information from the same phase are fed into the relevance detection module. This module consists of four parallel recurrent CNNs being responsible to detect four relevant phases that have been defined with medical experts. The results will then be integrated to classify the action phases as irrelevant or one of four relevant phases. Experimental results reveal that the proposed approach outperforms static CNNs and different configurations of feature-based and end-to-end recurrent networks. | ['Klaus Schoeffmann', 'Stephanie Sarny', 'Doris Putzgruber-Adamitsch', 'Mario Taschwer', 'Negin Ghamsarian'] | 2021-04-29 | null | null | null | null | ['spatio-temporal-action-localization'] | ['computer-vision'] | [ 3.62773061e-01 -1.43234581e-02 -2.72769690e-01 3.22650746e-02
-5.00725627e-01 -1.96569487e-01 2.54017800e-01 1.41776547e-01
-7.20559835e-01 5.38670361e-01 2.14553520e-01 -9.36914012e-02
-3.46726298e-01 -4.40886468e-01 -2.70523280e-01 -8.91129851e-01
-1.21750928e-01 -1.55862287e-01 3.07277322e-01 -7.02754185e-02
4.68581796e-01 6.45574391e-01 -2.06345367e+00 3.81389350e-01
5.19640326e-01 1.08976150e+00 5.09420156e-01 8.31344426e-01
1.34802818e-01 7.97532201e-01 -6.00307107e-01 2.75712520e-01
2.82127321e-01 -5.47863543e-01 -4.95300174e-01 2.64438912e-02
4.20397371e-01 -4.04916734e-01 -3.39459062e-01 1.19906557e+00
6.94512606e-01 1.71832532e-01 2.42676705e-01 -3.04915637e-01
6.81855679e-02 2.41689175e-01 -6.13805950e-02 8.21935654e-01
5.03335178e-01 1.79375246e-01 4.40284520e-01 -1.02806461e+00
8.36981058e-01 7.05546379e-01 -7.36475363e-02 5.55170596e-01
-6.32840872e-01 -3.80416185e-01 1.63371250e-01 6.80465698e-01
-1.20083904e+00 -4.51761514e-01 9.77947056e-01 -5.01277328e-01
6.18607163e-01 2.76617825e-01 1.40107286e+00 7.74525523e-01
7.39699841e-01 4.75894302e-01 8.78051460e-01 -4.74721819e-01
6.52513355e-02 9.47435051e-02 3.57706882e-02 7.59709060e-01
1.51935682e-01 4.18775141e-01 -4.99938607e-01 3.32609415e-01
9.07927632e-01 2.67764211e-01 -9.40279961e-01 -9.56936926e-02
-1.29655790e+00 3.94889444e-01 2.74669021e-01 5.82221866e-01
-6.49231791e-01 -2.88423777e-01 2.69513637e-01 3.44835907e-01
2.73042202e-01 5.05672097e-01 -3.30416501e-01 -7.49985576e-02
-1.00881886e+00 -3.32598239e-01 6.02824450e-01 6.02525890e-01
5.80430686e-01 -7.97097385e-02 -6.02863133e-01 3.55593771e-01
1.73925921e-01 -5.88135906e-02 9.59073305e-01 -7.78506935e-01
1.06387757e-01 8.19135547e-01 5.68473637e-02 -8.55527043e-01
-3.98092091e-01 -6.59172058e-01 -7.24732161e-01 4.67951477e-01
1.29359201e-01 7.62275085e-02 -1.19328189e+00 9.03656363e-01
2.18571261e-01 2.32272506e-01 -7.12373555e-02 1.29325557e+00
1.10142112e+00 3.99936646e-01 -1.66550949e-01 -7.56856740e-01
1.53435123e+00 -8.27129662e-01 -8.91977668e-01 7.61666819e-02
3.34474951e-01 -1.09454083e+00 7.00098932e-01 4.13342863e-01
-1.15902400e+00 -8.75492096e-01 -9.05511260e-01 4.45361286e-02
-1.36959583e-01 6.23576522e-01 3.52164149e-01 3.79874781e-02
-1.11986566e+00 4.90267247e-01 -7.41534412e-01 -2.10613608e-01
4.91007268e-02 6.20709956e-01 -3.13010693e-01 3.59426811e-02
-9.13833737e-01 8.66120338e-01 4.09321666e-01 7.83441722e-01
-9.52438474e-01 -2.92044491e-01 -9.53783512e-01 -5.40660284e-02
2.56902963e-01 -6.80542350e-01 1.04362166e+00 -1.21601677e+00
-1.59113789e+00 9.87210810e-01 -3.83046865e-01 -1.84427232e-01
3.83623868e-01 -1.27359807e-01 -2.65268922e-01 7.16592669e-01
-1.64672971e-01 3.65815371e-01 1.05253971e+00 -8.04278553e-01
-9.66880739e-01 -1.75224230e-01 2.45806679e-01 5.05686998e-01
-1.13992698e-01 1.44009262e-01 -1.06477618e+00 -7.44905293e-01
2.83267498e-01 -8.77731502e-01 -1.59631401e-01 9.69578177e-02
-1.58248439e-01 -8.05534199e-02 6.57840610e-01 -8.48700285e-01
1.21903002e+00 -2.11674452e+00 3.54860127e-01 1.55738115e-01
2.54747272e-01 5.04007578e-01 3.07380669e-02 -1.99885473e-01
-3.21965814e-01 -3.45944852e-01 3.58281046e-01 -1.76738173e-01
-9.19894040e-01 7.68863708e-02 3.73035967e-01 6.38950109e-01
1.13837235e-01 5.79440355e-01 -8.67838025e-01 -7.32440352e-01
6.38194859e-01 5.87665975e-01 -2.27237776e-01 5.18640697e-01
1.45609006e-01 8.88215780e-01 -3.26135725e-01 1.00206852e+00
1.17373444e-01 -1.22082710e-01 1.14050977e-01 -3.94206792e-01
-2.45531887e-01 6.91817552e-02 -9.33772504e-01 1.73420250e+00
-3.72565806e-01 9.15185809e-01 1.19851634e-01 -1.00556648e+00
5.99556983e-01 7.20768094e-01 5.83720982e-01 -8.39674413e-01
4.63461220e-01 2.44006321e-01 1.88648596e-01 -1.09019411e+00
3.78086478e-01 3.73062581e-01 5.08544207e-01 -1.63358316e-01
5.11897765e-02 2.38112777e-01 4.14298981e-01 -3.00807565e-01
7.90116251e-01 1.69010296e-01 4.69186991e-01 2.03737795e-01
8.62032771e-01 -1.14335448e-01 6.10386491e-01 5.47733068e-01
-5.29721677e-01 5.70182443e-01 2.07863048e-01 -7.79591322e-01
-5.76993883e-01 -5.57624102e-01 3.38603109e-02 5.74013174e-01
5.33056200e-01 -1.49743289e-01 -4.67899412e-01 -5.24157584e-01
-6.30348384e-01 -9.94599983e-02 -8.04378927e-01 -3.91571760e-01
-8.76036048e-01 -3.23425055e-01 -5.54715931e-01 1.46744743e-01
2.47018725e-01 -1.33352757e+00 -1.07736790e+00 1.76015243e-01
-3.16776305e-01 -8.06160629e-01 -4.79745358e-01 4.77741919e-02
-1.07310903e+00 -1.36699653e+00 -8.72653246e-01 -1.15787148e+00
9.57231879e-01 5.01460016e-01 7.85414457e-01 1.84660301e-01
-6.93655431e-01 2.94266045e-01 -3.38727891e-01 -1.59814686e-01
-1.79639593e-01 -2.67608523e-01 -2.55170316e-01 1.84237614e-01
2.48603508e-01 -3.82502377e-01 -1.17896223e+00 2.81764008e-02
-6.25871122e-01 8.21859986e-02 1.08860481e+00 8.89480412e-01
7.01094270e-01 -1.41826242e-01 -1.86716452e-01 -2.69380957e-01
3.60048234e-01 -4.85381000e-02 -7.99779058e-01 1.81363553e-01
-4.83766437e-01 -1.87168434e-01 4.60702449e-01 -6.25030756e-01
-8.29274237e-01 2.26888850e-01 3.05060863e-01 -9.47089434e-01
-1.31607533e-01 6.66399896e-01 2.18344823e-01 -3.07043731e-01
4.05364633e-01 3.47422600e-01 2.05068797e-01 -3.89672779e-02
-3.02241653e-01 4.03653026e-01 5.58455944e-01 2.04129577e-01
3.29407156e-01 2.55273253e-01 -4.50992025e-02 -6.38485909e-01
-7.36633182e-01 -8.69417369e-01 -4.89142179e-01 -9.64459419e-01
9.76333320e-01 -1.04578328e+00 -8.63658905e-01 3.09448689e-01
-1.24956703e+00 3.07542652e-01 -4.50893492e-01 1.02786958e+00
-2.36333340e-01 3.73122960e-01 -6.35632753e-01 -6.74039483e-01
-4.80992436e-01 -1.52868521e+00 1.05378878e+00 7.69467235e-01
2.15613246e-02 -6.71613216e-01 -2.27168560e-01 3.02476376e-01
1.50618017e-01 3.14300172e-02 5.61940670e-01 -2.23060980e-01
-1.19074523e+00 -5.77658117e-01 1.82679042e-01 3.80972564e-01
3.32037300e-01 1.16437756e-01 -9.05927539e-01 -2.82038540e-01
2.11933419e-01 3.40513974e-01 8.31856966e-01 9.88775730e-01
1.22489321e+00 -1.79616258e-01 -3.23592633e-01 8.65731955e-01
1.43144047e+00 5.30561209e-01 6.59321249e-01 3.82892311e-01
2.19363496e-01 5.75188279e-01 9.42514002e-01 2.31272489e-01
-9.71461684e-02 6.37874603e-01 8.27294469e-01 -3.32512051e-01
-3.31418246e-01 5.18204153e-01 4.54357326e-01 9.10851836e-01
-7.21768320e-01 1.24983042e-01 -5.92344224e-01 6.28360569e-01
-1.71937227e+00 -1.16043818e+00 1.37000918e-01 2.43322229e+00
8.89232397e-01 9.88329500e-02 -3.43894750e-01 2.18572527e-01
7.66281545e-01 -1.25440300e-01 -2.67464101e-01 -1.05751827e-01
8.59400854e-02 5.22173285e-01 2.75693953e-01 3.30248088e-01
-1.29352164e+00 4.87853080e-01 5.72577763e+00 4.85818833e-01
-1.70711827e+00 -1.61679372e-01 4.12709832e-01 -5.70683002e-01
3.08018625e-01 -1.43662184e-01 -5.97165585e-01 4.76591170e-01
5.63458264e-01 1.33889735e-01 4.01686914e-02 6.97990894e-01
7.70194829e-01 -5.47581017e-01 -9.60298121e-01 1.04326105e+00
2.33453766e-01 -1.40235627e+00 -7.67315365e-03 2.30350420e-02
5.31876624e-01 -1.95801765e-01 -7.54579976e-02 -1.07913017e-01
-6.38208985e-01 -8.05859983e-01 4.07523394e-01 1.18356705e+00
8.20712090e-01 -5.87192595e-01 1.19011509e+00 4.43261191e-02
-1.28792059e+00 -3.79611969e-01 -4.00853492e-02 8.90060589e-02
-1.39824569e-01 4.93885994e-01 -8.54530334e-01 6.39761031e-01
7.94077396e-01 9.58084345e-01 -5.04003942e-01 1.45166767e+00
-3.19858372e-01 3.17840129e-01 1.25637725e-01 8.62457007e-02
-1.50770023e-01 -1.61392540e-01 8.89005482e-01 1.02849078e+00
4.57379252e-01 3.17937620e-02 5.57131879e-02 4.42910492e-01
3.47749710e-01 -2.73222029e-02 -5.52494764e-01 2.93604314e-01
1.29836639e-02 1.41453195e+00 -6.99953973e-01 -3.40315759e-01
-4.81402814e-01 1.01923048e+00 -2.18276098e-01 5.16090810e-01
-4.85762328e-01 -5.92825711e-01 4.27157998e-01 1.11929253e-01
2.42569715e-01 -5.45883588e-02 1.57850757e-01 -1.17621028e+00
3.99847120e-01 -7.20652163e-01 1.48812115e-01 -8.29838812e-01
-4.26572114e-01 5.83823621e-01 -3.69384140e-01 -1.98303819e+00
-4.67229486e-01 -6.79310441e-01 -6.19782329e-01 1.06619358e+00
-1.82971466e+00 -9.04753029e-01 -8.45104337e-01 7.53554106e-01
8.84357214e-01 -2.25561842e-01 6.96097076e-01 1.57550380e-01
-7.79086351e-01 1.14317134e-01 -3.56879771e-01 2.11029813e-01
7.12323248e-01 -9.60722983e-01 -5.93536973e-01 1.22310436e+00
-3.44324768e-01 7.18198478e-01 5.96897364e-01 -5.75978756e-01
-1.18675566e+00 -9.79544461e-01 8.02831709e-01 5.92752509e-02
1.94005623e-01 3.33571285e-01 -5.68805873e-01 2.94692636e-01
1.48288533e-01 3.20897818e-01 4.57536012e-01 -4.64465916e-01
3.87258559e-01 -3.20069551e-01 -6.05132580e-01 5.04247665e-01
6.54294372e-01 -5.96452594e-01 -6.96180224e-01 3.14312577e-01
4.86970276e-01 -6.67320251e-01 -7.69906700e-01 6.45271182e-01
6.31926894e-01 -1.07753277e+00 8.33296597e-01 -1.52482405e-01
6.42983735e-01 -3.07140052e-01 4.46456611e-01 -8.77918720e-01
-1.16295226e-01 -6.41914546e-01 -1.78752601e-01 3.00303102e-01
1.14827365e-01 -2.75051653e-01 7.63966441e-01 4.19063866e-02
-3.80902380e-01 -8.73790443e-01 -9.87826049e-01 -2.90746957e-01
-8.82823169e-01 -2.69538518e-02 -2.64329463e-01 5.69466293e-01
-1.52164578e-01 4.03628089e-02 -8.78869444e-02 2.56800026e-01
3.26391667e-01 2.97565669e-01 3.83820295e-01 -1.29638648e+00
-1.43898457e-01 -5.17673969e-01 -7.74869382e-01 -7.52887845e-01
-2.64116466e-01 -5.92533350e-01 1.25319958e-01 -1.55500221e+00
6.40380457e-02 2.81820685e-01 -6.36297464e-01 2.17613712e-01
-2.63677806e-01 2.60588735e-01 -1.16698138e-01 5.44637918e-01
-2.73564756e-01 1.92941368e-01 1.55424690e+00 -5.98523840e-02
-5.46631753e-01 3.81742120e-01 -3.65165845e-02 6.95784330e-01
5.47080815e-01 -1.31249711e-01 -3.08324605e-01 -2.66775638e-02
1.49973005e-01 5.23143113e-01 5.94909191e-01 -1.28102791e+00
6.14900172e-01 1.31857798e-01 6.19458497e-01 -6.98925138e-01
5.35561323e-01 -8.91840279e-01 -1.55869573e-02 7.90895283e-01
-2.27460071e-01 -1.42899677e-01 9.67660546e-02 4.89791572e-01
-7.72673845e-01 -1.35131896e-01 8.25204432e-01 -4.11858588e-01
-8.28560889e-01 2.83851087e-01 -6.41232550e-01 -5.20762980e-01
1.20498800e+00 -7.63420343e-01 -1.10086098e-01 -2.26076275e-01
-1.23882782e+00 -1.59301117e-01 7.91079327e-02 3.21227401e-01
1.18072629e+00 -7.32483387e-01 -4.79027241e-01 5.28124630e-01
1.89931467e-01 1.60551772e-01 4.67897445e-01 1.37115824e+00
-5.71341097e-01 4.13309902e-01 -1.78088114e-01 -8.22885454e-01
-1.68332481e+00 4.69652057e-01 7.44957566e-01 -1.29486844e-01
-7.80603826e-01 7.92605281e-01 1.16734505e-01 4.87776667e-01
6.38023138e-01 -6.65642798e-01 -1.05528057e+00 3.31836492e-01
5.92375994e-01 -5.36998324e-02 2.73644775e-01 -4.26110834e-01
-6.46404475e-02 9.25846457e-01 -2.05838487e-01 2.37583280e-01
1.04010844e+00 -1.11844100e-01 -3.31540734e-01 2.71080315e-01
9.82654154e-01 2.85942256e-02 -9.65322673e-01 -2.04981655e-01
-3.52108836e-01 -3.75062674e-01 2.15961069e-01 -7.14326799e-01
-1.30237603e+00 7.13667989e-01 1.17290378e+00 -1.64701626e-01
1.79217279e+00 -1.87004119e-01 3.02723080e-01 4.65058023e-03
1.61216974e-01 -1.04580235e+00 6.96128681e-02 2.22586378e-01
8.57161701e-01 -1.25654447e+00 -3.02816462e-02 -2.43130758e-01
-1.73135445e-01 1.51221836e+00 6.15993381e-01 -2.76439432e-02
7.23661125e-01 1.42479932e-03 1.32413641e-01 -3.67881358e-01
-7.23125160e-01 -3.94184917e-01 7.74328530e-01 3.43770623e-01
3.46697837e-01 -2.28058383e-01 -4.13201898e-01 1.15882300e-01
1.18071988e-01 3.76519203e-01 6.86726332e-01 8.97262156e-01
-5.62256098e-01 -6.68216705e-01 -2.51330286e-01 4.91920143e-01
-5.71463406e-01 -2.34518692e-01 4.82430160e-02 7.60128140e-01
4.26848501e-01 9.39583063e-01 8.19031745e-02 -2.52302676e-01
3.35093319e-01 -2.22189471e-01 3.57701242e-01 -7.18526721e-01
-8.33117068e-01 4.62072134e-01 -4.85081375e-02 -1.04900992e+00
-7.64907300e-01 -5.03822923e-01 -9.17747974e-01 3.49738240e-01
-2.91222394e-01 5.45246713e-02 5.87519705e-01 7.38961220e-01
2.42878348e-01 9.45749104e-01 5.45817137e-01 -9.87009764e-01
-1.49091249e-02 -8.55528295e-01 -2.93722361e-01 -1.26296245e-02
9.14784491e-01 -4.77312356e-01 -5.16669631e-01 3.15231264e-01] | [14.107080459594727, -3.333921432495117] |
f973730d-e9f1-4030-ba3e-14a34f5f5545 | learning-by-asking-questions-for-knowledge | 2210.05879 | null | https://arxiv.org/abs/2210.05879v1 | https://arxiv.org/pdf/2210.05879v1.pdf | Learning by Asking Questions for Knowledge-based Novel Object Recognition | In real-world object recognition, there are numerous object classes to be recognized. Conventional image recognition based on supervised learning can only recognize object classes that exist in the training data, and thus has limited applicability in the real world. On the other hand, humans can recognize novel objects by asking questions and acquiring knowledge about them. Inspired by this, we study a framework for acquiring external knowledge through question generation that would help the model instantly recognize novel objects. Our pipeline consists of two components: the Object Classifier, which performs knowledge-based object recognition, and the Question Generator, which generates knowledge-aware questions to acquire novel knowledge. We also propose a question generation strategy based on the confidence of the knowledge-aware prediction of the Object Classifier. To train the Question Generator, we construct a dataset that contains knowledge-aware questions about objects in the images. Our experiments show that the proposed pipeline effectively acquires knowledge about novel objects compared to several baselines. | ['Tatsuya Harada', 'Kohei Uehara'] | 2022-10-12 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [ 6.66801870e-01 3.46033543e-01 9.49383974e-02 -7.22309113e-01
-5.95362663e-01 -6.68633997e-01 5.97700298e-01 1.05519779e-01
-2.84622520e-01 5.65866172e-01 -3.77375335e-01 -2.47275427e-01
6.12344295e-02 -1.20000172e+00 -9.87745106e-01 -3.21998984e-01
4.17705804e-01 5.26238084e-01 6.75778985e-01 2.99502015e-01
3.98334086e-01 4.57174569e-01 -2.01577806e+00 5.05268753e-01
9.97182488e-01 1.30460572e+00 5.07329881e-01 5.25698304e-01
-5.36989570e-01 9.44586635e-01 -7.69588768e-01 -3.60446304e-01
3.74136977e-02 -5.67224205e-01 -1.13893533e+00 5.05260825e-01
4.60937500e-01 -4.56270069e-01 -1.96437389e-02 9.53599155e-01
-1.31312728e-01 2.38289014e-01 5.55062175e-01 -1.22545254e+00
-1.08826864e+00 4.40184504e-01 1.00481905e-01 2.75110424e-01
3.05475831e-01 5.00038981e-01 8.28725755e-01 -1.13899755e+00
5.34677804e-01 1.01914656e+00 5.49953319e-02 7.30993152e-01
-7.56045938e-01 -3.66055906e-01 3.87593061e-01 8.50266397e-01
-1.03934085e+00 -3.01993787e-01 8.15677345e-01 -3.33007216e-01
7.80425072e-01 2.51039952e-01 5.53245127e-01 7.46443331e-01
-2.00071111e-01 1.22421813e+00 9.77939010e-01 -5.69841206e-01
4.72575933e-01 6.66546047e-01 2.95383990e-01 4.28215474e-01
2.10519716e-01 1.00238614e-01 -5.82515121e-01 1.44247741e-01
4.84435230e-01 3.90845202e-02 -1.49386004e-01 -1.51213497e-01
-1.25067961e+00 5.23378134e-01 5.53327858e-01 3.81758451e-01
-6.40823901e-01 -3.05190980e-01 -5.49198575e-02 1.22407600e-01
-1.05152582e-03 6.21648371e-01 -6.55303001e-01 2.49620095e-01
-6.12975895e-01 1.30422547e-01 8.65209043e-01 1.10336506e+00
1.15506136e+00 -2.84843206e-01 -2.83627629e-01 4.06697750e-01
2.57322907e-01 5.83988190e-01 6.51508570e-01 -7.83493936e-01
7.28016645e-02 1.12975073e+00 -6.53920695e-02 -7.57403135e-01
1.78683382e-02 -3.47517073e-01 -3.71988297e-01 -7.94929191e-02
3.09258610e-01 1.78758711e-01 -1.01855969e+00 1.39529574e+00
6.97521091e-01 4.44363266e-01 2.49839500e-01 7.17356741e-01
1.13711405e+00 6.25899136e-01 -4.22572084e-02 -1.40480667e-01
1.33767688e+00 -1.11296821e+00 -5.34501553e-01 -4.93669450e-01
2.28820756e-01 -4.67271090e-01 1.09719932e+00 2.96018273e-01
-6.17575526e-01 -9.10680950e-01 -6.33689046e-01 4.97427732e-02
-6.16987407e-01 2.37953171e-01 5.23765504e-01 4.51435834e-01
-7.05458105e-01 5.59304059e-02 -4.21303391e-01 -4.29991812e-01
6.22901261e-01 1.55918285e-01 -1.80083774e-02 -2.92157650e-01
-8.22680414e-01 8.32244396e-01 9.94170964e-01 1.09151192e-01
-1.17043996e+00 -5.43751180e-01 -8.29096437e-01 2.11356997e-01
8.56415927e-01 -5.73831499e-01 1.43491900e+00 -1.48092759e+00
-1.19846344e+00 9.22828376e-01 -1.82872534e-01 -3.35079044e-01
-1.03906445e-01 -6.44381493e-02 -3.37645054e-01 3.68118614e-01
1.35613933e-01 7.60270178e-01 1.11288106e+00 -1.41848695e+00
-9.85243022e-01 -3.74683350e-01 3.41361314e-01 -8.79652180e-06
-3.60045820e-01 -2.36356571e-01 -3.34478915e-01 -3.70787263e-01
9.77418125e-02 -5.64894497e-01 -6.13374114e-02 -1.06044728e-02
-2.76419729e-01 -5.56048334e-01 1.03353798e+00 -4.39256191e-01
6.80008113e-01 -1.98270500e+00 -4.77793306e-01 2.88093865e-01
-2.51996499e-02 4.75901246e-01 -4.10609215e-01 -6.14647195e-02
-6.12285733e-03 -2.41703019e-01 -9.79974568e-02 4.58683044e-01
-1.52666420e-01 3.96282285e-01 -6.78323567e-01 -4.33154136e-01
8.53950918e-01 1.29145920e+00 -1.17749524e+00 -3.09055150e-01
8.91429931e-02 -9.95048583e-02 -4.06258315e-01 7.59000778e-01
-9.06553149e-01 3.74238670e-01 -6.46085024e-01 6.85514212e-01
5.55760622e-01 -6.77409947e-01 7.42549747e-02 -1.42705232e-01
3.54952037e-01 1.74772546e-01 -1.07277393e+00 1.27473092e+00
-2.23693013e-01 4.00772423e-01 -5.22999763e-01 -1.29169250e+00
1.13429117e+00 -1.50922895e-03 -1.45620689e-01 -5.34702003e-01
4.32125665e-02 4.67606541e-03 8.69732574e-02 -1.06181955e+00
3.07529718e-01 -2.72943318e-01 3.43313873e-01 5.37660897e-01
5.07111609e-01 1.20774109e-03 2.49247521e-01 1.86532110e-01
1.17744696e+00 7.19499961e-02 1.86915591e-01 2.40399942e-01
8.30928445e-01 3.08657080e-01 3.88556927e-01 9.26063418e-01
-1.50254250e-01 2.82393545e-01 1.21230055e-02 -5.87688506e-01
-4.38272834e-01 -1.16681135e+00 2.07912445e-01 1.05764341e+00
2.77247280e-01 2.60276198e-02 -5.58495224e-01 -1.20424092e+00
-2.51742229e-02 9.55313504e-01 -6.83688164e-01 -4.41497922e-01
-2.01125339e-01 -2.35106826e-01 6.95160124e-03 6.16143286e-01
6.17391050e-01 -1.74146855e+00 -8.58354986e-01 5.98578118e-02
-2.22274020e-01 -1.19687688e+00 -1.18137263e-01 -9.70357507e-02
-1.02931392e+00 -1.50271428e+00 -2.60990113e-01 -9.87983286e-01
1.17449760e+00 3.12010884e-01 1.20752025e+00 2.92604893e-01
-3.31105024e-01 8.77415538e-01 -7.24973202e-01 -7.66351759e-01
-5.13637006e-01 -8.75136629e-03 -2.16700748e-01 5.39291918e-01
7.66228318e-01 -2.65565932e-01 -4.11519825e-01 3.29234421e-01
-1.26039529e+00 6.02386147e-02 9.63636816e-01 5.96460938e-01
5.79150200e-01 2.29281962e-01 1.02985609e+00 -8.78012776e-01
1.70462638e-01 -4.24141943e-01 -5.90500116e-01 7.70585954e-01
-2.97151327e-01 8.74363780e-02 5.41385412e-01 -7.59396553e-01
-1.37680852e+00 3.28078091e-01 1.38301803e-02 -1.96601927e-01
-6.40680671e-01 6.72135353e-01 -4.55756068e-01 5.87774552e-02
6.99087441e-01 7.81014502e-01 -2.31889740e-01 -3.74072641e-01
5.89673579e-01 6.10632837e-01 6.56261802e-01 -5.33566535e-01
1.00912583e+00 3.74482602e-01 -5.54639101e-01 -8.01580966e-01
-1.29242468e+00 -3.77967149e-01 -6.99838698e-01 -4.01600331e-01
7.14408040e-01 -6.25606835e-01 -6.88025355e-01 4.41858619e-01
-1.23937881e+00 -1.64334059e-01 -8.25989008e-01 3.21872085e-01
-3.93652320e-01 3.22182700e-02 6.31647855e-02 -8.70991945e-01
-9.56481621e-02 -7.88386643e-01 9.21680391e-01 7.17639625e-01
1.11152768e-01 -6.28980994e-01 -2.08906084e-01 6.81405365e-01
3.54917705e-01 -1.58022553e-01 1.07525718e+00 -1.21150577e+00
-1.23978877e+00 -1.42539158e-01 -2.86076248e-01 6.10901535e-01
3.42913896e-01 -4.07747984e-01 -1.10620499e+00 7.36823678e-02
1.45383239e-01 -6.39062285e-01 6.73180044e-01 -1.10737666e-01
1.32845807e+00 -4.44990695e-01 -3.00233245e-01 -8.72173831e-02
1.00858617e+00 3.92079473e-01 6.70829892e-01 1.44711435e-01
2.55852580e-01 8.25360179e-01 7.67468750e-01 1.27603158e-01
6.60381317e-01 3.18054020e-01 2.82065094e-01 2.04689458e-01
-3.07810586e-02 -1.87785164e-01 2.21956268e-01 5.82833588e-01
3.45865309e-01 -2.35705357e-02 -9.87671077e-01 8.54690373e-01
-1.65310025e+00 -7.52542734e-01 1.72534257e-01 1.84531665e+00
1.12332904e+00 1.09476343e-01 -1.29731596e-01 8.98795798e-02
5.22633255e-01 -4.76706058e-01 -1.04118621e+00 2.08756611e-01
1.00818925e-01 2.46174678e-01 -3.40811968e-01 9.92159843e-02
-8.68106484e-01 1.00825453e+00 5.84773922e+00 4.55747485e-01
-1.02340543e+00 -7.84667805e-02 4.68172759e-01 2.22811192e-01
-2.04571247e-01 2.90448457e-01 -8.93029034e-01 2.53415629e-02
6.54489160e-01 -4.27350283e-01 2.02441067e-02 1.03685498e+00
-2.63048738e-01 -4.17752385e-01 -1.38617134e+00 7.88178205e-01
5.58568120e-01 -1.21851933e+00 7.51519561e-01 -3.71311456e-01
7.05396891e-01 -4.10621613e-01 -1.25899225e-01 4.57830667e-01
7.11828172e-02 -6.93969071e-01 4.95867252e-01 1.08612633e+00
2.47195259e-01 -4.20410395e-01 6.59677446e-01 8.68399978e-01
-8.08408618e-01 -2.90530264e-01 -3.51287335e-01 -3.63272913e-02
-1.02645658e-01 7.65367091e-01 -1.43722534e+00 4.20009047e-01
6.88046932e-01 4.59143758e-01 -1.04352927e+00 1.25805438e+00
-6.86759531e-01 8.82166862e-01 -1.88048378e-01 -9.75576267e-02
-1.92959428e-01 1.55276388e-01 1.05808601e-01 6.52337730e-01
8.05758014e-02 5.69821656e-01 3.22423249e-01 1.06761849e+00
-2.07473502e-01 -1.57182980e-02 -4.08381909e-01 -1.91477910e-01
2.84536600e-01 1.25347817e+00 -8.52419078e-01 -6.84586287e-01
-5.16729176e-01 8.61006320e-01 2.80361980e-01 3.45021605e-01
-3.96965951e-01 -5.01023471e-01 8.39845650e-03 8.68360698e-03
6.67205334e-01 1.13853894e-01 7.62408003e-02 -1.28222728e+00
4.38720137e-01 -1.08098209e+00 4.30019200e-01 -1.01523292e+00
-1.44570589e+00 5.91576278e-01 -1.54678777e-01 -8.63497257e-01
-3.22960198e-01 -8.54367554e-01 -5.75458109e-01 4.97752279e-01
-1.72637618e+00 -1.26200414e+00 -5.73821068e-01 5.16919851e-01
7.01484144e-01 -1.60869032e-01 6.84028089e-01 -9.88323763e-02
-1.46060556e-01 4.15680140e-01 -4.70401794e-01 3.25660139e-01
5.13435781e-01 -9.65997815e-01 3.12220216e-01 8.27867091e-01
5.47308981e-01 5.77045381e-01 2.65620798e-01 -8.36404204e-01
-1.42918241e+00 -1.42375100e+00 8.62290800e-01 -7.15717912e-01
3.91443580e-01 -2.44490877e-01 -1.45273662e+00 7.79755890e-01
-1.31221890e-01 7.43120983e-02 8.93322229e-01 -3.11920065e-02
-5.41903973e-01 -2.07180843e-01 -1.14001656e+00 2.56394953e-01
7.59601533e-01 -5.24533987e-01 -1.20257092e+00 3.30074370e-01
7.00519502e-01 -7.93644935e-02 -4.96957421e-01 5.77030540e-01
4.38679606e-01 -7.22512901e-01 7.09209979e-01 -1.02362168e+00
2.63710588e-01 -5.24890661e-01 -7.43328407e-02 -8.28322947e-01
-1.49581254e-01 5.34406081e-02 -3.29903692e-01 1.28281057e+00
6.53412163e-01 -5.25400400e-01 8.51635933e-01 7.54695475e-01
1.79292008e-01 -5.40917695e-01 -6.76671684e-01 -9.90652978e-01
-2.68829137e-01 -4.36234474e-01 8.04169953e-01 8.63088906e-01
-1.61538944e-01 5.49253643e-01 2.49010816e-01 3.81930172e-01
4.25715983e-01 6.60822928e-01 9.41330194e-01 -1.19019806e+00
-3.98397952e-01 -3.24713066e-02 -5.21029055e-01 -1.27167487e+00
1.87915683e-01 -9.92379248e-01 5.93951404e-01 -1.64471591e+00
5.23157001e-01 -4.62297916e-01 -3.03394586e-01 7.71684408e-01
-7.27508247e-01 -2.41610892e-02 2.34320998e-01 8.85504298e-03
-1.03941751e+00 6.23527110e-01 1.30670011e+00 -2.30683193e-01
-1.13605432e-01 1.72229689e-02 -8.87054443e-01 8.24150980e-01
5.81870496e-01 -4.28669184e-01 -4.60344970e-01 -4.47348058e-01
2.48625547e-01 -2.45021984e-01 7.06740439e-01 -1.10312867e+00
4.28088844e-01 -3.72661799e-01 5.22408426e-01 -7.28131533e-01
-5.19429967e-02 -9.83448625e-01 -2.15135619e-01 3.71723264e-01
-4.81707782e-01 -5.05352557e-01 2.33427301e-01 5.74512422e-01
-2.54511535e-01 -4.61886823e-01 5.18110216e-01 -1.43292814e-01
-1.08931100e+00 2.92064160e-01 -3.28041345e-01 1.23130918e-01
1.12018478e+00 -1.02286741e-01 -3.41508478e-01 -2.97497720e-01
-7.53979445e-01 4.29600269e-01 8.27160329e-02 8.00199986e-01
1.08766949e+00 -1.24667788e+00 -6.60507023e-01 4.06904966e-01
6.36504173e-01 1.10006720e-01 1.40207753e-01 2.04730406e-01
4.41655777e-02 1.73933774e-01 -2.84851231e-02 -6.37143612e-01
-9.41806078e-01 8.66966546e-01 2.24641964e-01 -4.81645428e-02
-9.19219032e-02 8.43912005e-01 3.37723494e-01 -7.73852348e-01
2.58419752e-01 -4.30307031e-01 -3.01454157e-01 -2.15457633e-01
9.56105709e-01 -1.26215279e-01 -2.31881421e-02 -3.38110805e-01
-3.02754253e-01 2.27704868e-01 -3.45068514e-01 5.19178882e-02
1.12851250e+00 5.50517999e-02 -2.36428812e-01 5.62745035e-01
6.59753680e-01 -6.21538341e-01 -1.03103113e+00 -7.67162085e-01
4.37894166e-01 -4.89638537e-01 -4.26420599e-01 -1.27960193e+00
-7.64547944e-01 7.82083094e-01 5.78458428e-01 1.02108449e-01
1.43347883e+00 5.04836857e-01 7.85107017e-01 1.11947691e+00
4.63131100e-01 -8.60690296e-01 6.00091040e-01 7.13480055e-01
1.02958202e+00 -1.31823766e+00 -3.55887085e-01 -6.49798155e-01
-3.80199879e-01 1.09302807e+00 1.08191645e+00 1.99272379e-01
6.19229794e-01 -2.52301961e-01 3.23961467e-01 -2.45760828e-01
-9.00961220e-01 -4.56355631e-01 6.09252334e-01 8.67413819e-01
-1.35982201e-01 -2.68272847e-01 1.99861452e-01 9.83019769e-01
-1.54276446e-01 3.88799459e-01 3.09994489e-01 1.18048000e+00
-8.99260223e-01 -1.05451572e+00 -3.96214604e-01 8.03002238e-01
3.30108479e-02 -3.69184688e-02 -5.58299959e-01 2.23722845e-01
4.64746565e-01 1.19327462e+00 1.21251503e-02 -3.00354451e-01
5.48027635e-01 2.91165411e-01 4.74369794e-01 -1.19502187e+00
-4.31890756e-01 -6.66557252e-01 -2.56894469e-01 -4.31985497e-01
-5.76297283e-01 -4.99909014e-01 -1.25091445e+00 5.89127541e-01
-6.99527264e-01 2.34873280e-01 5.32706141e-01 1.46513331e+00
4.54713196e-01 4.21266645e-01 7.54513383e-01 -2.58575678e-01
-5.56470335e-01 -8.25333416e-01 -1.68727860e-01 4.83872414e-01
2.11522281e-01 -4.05647725e-01 -8.75002071e-02 7.34256148e-01] | [10.050262451171875, 1.903411626815796] |
952aafb7-e99d-445b-ab44-ca9f6673af39 | ovtrack-open-vocabulary-multiple-object | 2304.08408 | null | https://arxiv.org/abs/2304.08408v1 | https://arxiv.org/pdf/2304.08408v1.pdf | OVTrack: Open-Vocabulary Multiple Object Tracking | The ability to recognize, localize and track dynamic objects in a scene is fundamental to many real-world applications, such as self-driving and robotic systems. Yet, traditional multiple object tracking (MOT) benchmarks rely only on a few object categories that hardly represent the multitude of possible objects that are encountered in the real world. This leaves contemporary MOT methods limited to a small set of pre-defined object categories. In this paper, we address this limitation by tackling a novel task, open-vocabulary MOT, that aims to evaluate tracking beyond pre-defined training categories. We further develop OVTrack, an open-vocabulary tracker that is capable of tracking arbitrary object classes. Its design is based on two key ingredients: First, leveraging vision-language models for both classification and association via knowledge distillation; second, a data hallucination strategy for robust appearance feature learning from denoising diffusion probabilistic models. The result is an extremely data-efficient open-vocabulary tracker that sets a new state-of-the-art on the large-scale, large-vocabulary TAO benchmark, while being trained solely on static images. Project page: https://www.vis.xyz/pub/ovtrack/ | ['Fisher Yu', 'Martin Danelljan', 'Henghui Ding', 'Lei Ke', 'Tobias Fischer', 'Siyuan Li'] | 2023-04-17 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_OVTrack_Open-Vocabulary_Multiple_Object_Tracking_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_OVTrack_Open-Vocabulary_Multiple_Object_Tracking_CVPR_2023_paper.pdf | cvpr-2023-1 | ['multiple-object-tracking'] | ['computer-vision'] | [-1.00640237e-01 -4.91331011e-01 -2.65903771e-01 5.01008099e-03
-7.12711036e-01 -5.82435250e-01 8.53587627e-01 -7.77004883e-02
-2.43848845e-01 2.72048622e-01 -9.77589041e-02 1.06542110e-01
-1.39825836e-01 -3.67997319e-01 -8.21866989e-01 -7.36094892e-01
6.30779415e-02 6.96029484e-01 6.42630458e-01 -1.38869435e-01
-1.79265857e-01 4.21835929e-01 -2.00539136e+00 -1.89861171e-02
5.14389753e-01 1.10585606e+00 3.55871499e-01 5.43490171e-01
-1.80877578e-02 6.79256797e-01 -2.52997190e-01 -2.76205719e-01
3.62610996e-01 2.86741331e-02 -2.25853935e-01 -8.67319703e-02
1.14752126e+00 -8.11392367e-02 -5.78014791e-01 1.27148056e+00
3.39169174e-01 6.23128563e-02 5.80647051e-01 -1.45167470e+00
-1.04002464e+00 9.65412855e-02 -4.55556989e-01 4.54509676e-01
-5.62975481e-02 5.32224953e-01 8.64120901e-01 -8.27713013e-01
8.21536541e-01 1.31138444e+00 7.95050263e-01 7.59254217e-01
-1.27588797e+00 -7.81930804e-01 3.22267830e-01 3.17718238e-01
-1.50515270e+00 -5.98388374e-01 5.63024044e-01 -8.72461617e-01
5.45209944e-01 1.31118700e-01 6.78499937e-01 1.22461390e+00
3.79855871e-01 5.85836053e-01 1.07727218e+00 -5.81626855e-02
5.98288327e-02 1.78420261e-01 2.77508527e-01 6.76521838e-01
5.08853078e-01 4.33225274e-01 -5.95432281e-01 7.45575037e-03
5.11721313e-01 1.78576738e-01 -7.30213299e-02 -9.47420299e-01
-1.57727909e+00 7.44541109e-01 4.90516633e-01 2.34859496e-01
-2.98078865e-01 4.93410587e-01 2.09702760e-01 2.65597850e-01
4.50622737e-01 1.01317227e-01 -3.20441872e-01 5.15760332e-02
-8.64569187e-01 2.25961536e-01 6.08249068e-01 1.22775376e+00
7.59154379e-01 9.97490883e-02 -4.15945351e-01 4.01543736e-01
6.95324183e-01 1.15899634e+00 2.91268706e-01 -8.15105438e-01
-6.12092651e-02 3.62543732e-01 9.96630341e-02 -7.28335917e-01
-1.59146413e-01 -5.86205781e-01 -4.48936462e-01 4.60921854e-01
4.40326780e-01 2.74767786e-01 -1.02545905e+00 1.85010910e+00
7.67505884e-01 4.54379469e-01 -8.33405703e-02 9.67402935e-01
1.02542460e+00 2.81105220e-01 6.17142990e-02 6.27499148e-02
1.69014263e+00 -1.11459744e+00 -7.64036357e-01 -2.72153407e-01
4.14325655e-01 -6.58203483e-01 8.13613713e-01 3.34247619e-01
-7.33429134e-01 -6.01545036e-01 -1.00122321e+00 -5.00734448e-02
-5.44832051e-01 3.35570276e-02 6.10281587e-01 6.54637635e-01
-1.10414100e+00 2.01730490e-01 -9.38139677e-01 -6.71714544e-01
6.70002878e-01 3.52566332e-01 -3.41445178e-01 -1.49654374e-01
-7.13071465e-01 1.03758860e+00 7.60888755e-02 -6.86167553e-02
-1.17041874e+00 -8.09660196e-01 -8.04097354e-01 -3.81571025e-01
5.46376824e-01 -9.86712694e-01 1.20075405e+00 -7.96590090e-01
-1.20234311e+00 9.50163126e-01 -2.56815106e-01 -4.93012100e-01
4.56359446e-01 -3.53537023e-01 -6.69061601e-01 -7.68465772e-02
2.41029024e-01 7.41402984e-01 1.26657772e+00 -1.25158083e+00
-5.57420850e-01 -4.25840706e-01 -1.80553317e-01 5.01601547e-02
-3.80219787e-01 1.42383531e-01 -6.67250395e-01 -5.09218991e-01
-2.13337779e-01 -1.15679479e+00 -1.35808317e-02 7.80118167e-01
-1.12868451e-01 -4.47043121e-01 1.30575001e+00 -2.60414064e-01
6.03961766e-01 -2.26569200e+00 1.96764588e-01 -2.58968115e-01
5.83927870e-01 3.07809412e-01 -2.09117353e-01 1.16381114e-02
3.79562646e-01 -3.91651779e-01 2.71239996e-01 -5.27877331e-01
2.38734126e-01 1.17649846e-01 -5.30771077e-01 8.43398094e-01
9.82516110e-02 1.14514387e+00 -1.08877051e+00 -6.65201426e-01
5.14862955e-01 5.81321836e-01 -3.31925243e-01 -1.28224745e-01
-6.36365235e-01 6.24024212e-01 -5.22208154e-01 7.39453971e-01
6.54048979e-01 -4.11013871e-01 -3.70194227e-01 -3.77121717e-01
-3.30572128e-01 -1.82862803e-01 -1.20609546e+00 1.99959087e+00
-1.58524722e-01 8.53762507e-01 8.43770802e-02 -6.20089173e-01
7.97970593e-01 1.47820113e-03 8.02917659e-01 -5.28716028e-01
2.36642599e-01 1.33189887e-01 -5.46521544e-02 -3.86283398e-01
5.63337386e-01 1.09681778e-01 6.17303550e-02 1.87159330e-01
2.56882250e-01 5.74544929e-02 2.75452524e-01 2.96926498e-01
1.23490846e+00 3.41359049e-01 -4.09164978e-03 -2.79003441e-01
1.66903675e-01 2.28128761e-01 5.44926107e-01 9.89282429e-01
-6.56662166e-01 4.63046610e-01 -2.90431142e-01 -3.19164395e-01
-8.43937397e-01 -1.42153561e+00 -3.54109019e-01 1.01309979e+00
5.90339243e-01 -3.35381657e-01 -1.96197599e-01 -5.18125534e-01
2.61769742e-01 3.08207899e-01 -6.60699785e-01 -1.20700084e-01
-1.85640842e-01 -4.28644627e-01 4.65268403e-01 3.25146377e-01
1.27979770e-01 -9.39060390e-01 -6.97819948e-01 -5.69401728e-03
3.98006663e-02 -1.31052732e+00 -5.86277962e-01 -1.65366873e-01
-5.71591854e-01 -1.11864185e+00 -6.57202423e-01 -5.82849681e-01
2.55866647e-01 6.54948592e-01 1.01327753e+00 -1.06548615e-01
-5.14343798e-01 1.02660441e+00 -1.48743466e-01 -6.22803628e-01
-2.69043595e-01 -3.14752966e-01 4.82187599e-01 2.43193626e-01
5.41710317e-01 -2.57962853e-01 -5.79195499e-01 3.72363687e-01
-6.73869610e-01 -5.58585562e-02 7.38368154e-01 5.79374611e-01
7.04439819e-01 -4.13058102e-01 3.17367613e-01 -1.87593207e-01
-1.04528393e-04 -5.54918110e-01 -7.84612954e-01 4.20778245e-01
-3.01811308e-01 7.05797598e-02 1.62939310e-01 -1.12070584e+00
-7.53756106e-01 3.63304615e-01 2.20049933e-01 -1.01678407e+00
-1.35213435e-01 4.66161855e-02 -3.03493179e-02 -4.82158422e-01
6.99853837e-01 3.60902071e-01 1.94072023e-01 -4.30945426e-01
8.32683444e-01 3.82057190e-01 7.11911917e-01 -3.82566601e-01
1.21136975e+00 1.02057719e+00 -1.81574021e-02 -1.02165520e+00
-9.43402946e-01 -9.06713009e-01 -5.85676253e-01 -5.84715426e-01
1.00773025e+00 -1.30461037e+00 -7.89929569e-01 3.64599645e-01
-1.01390767e+00 -3.50134999e-01 -5.23210645e-01 6.61854923e-01
-5.01528561e-01 3.37282896e-01 -9.58008692e-02 -5.88783562e-01
-2.65450329e-01 -1.11122656e+00 1.35570908e+00 2.49308914e-01
1.20505668e-01 -8.88574183e-01 3.18371743e-01 2.36732617e-01
5.76387167e-01 1.16616517e-01 3.26548852e-02 -5.48868299e-01
-1.19709444e+00 -4.73189205e-02 -3.12366545e-01 9.50015262e-02
2.56037563e-02 -1.13909639e-01 -1.04723954e+00 -5.65002203e-01
-9.06843469e-02 -4.35187638e-01 1.12276828e+00 3.39952707e-01
7.89707303e-01 1.76475152e-01 -7.89158940e-01 7.59445131e-01
1.29229414e+00 -2.93006927e-01 2.42839381e-01 1.61845654e-01
8.94303501e-01 2.88130671e-01 6.22458935e-01 1.75263584e-01
5.74250281e-01 1.07032382e+00 5.82437456e-01 1.20559119e-01
-7.21066594e-01 -1.33205364e-02 5.73174357e-01 6.64201736e-01
2.40463659e-01 -1.06761657e-01 -8.67313445e-01 7.42375314e-01
-1.91928339e+00 -1.00324500e+00 -3.57895315e-01 2.17239237e+00
4.88697708e-01 -1.98936928e-02 4.69283164e-02 -5.70760608e-01
7.25975871e-01 2.32548088e-01 -9.05443728e-01 4.60037500e-01
-2.30852768e-01 -1.13577291e-01 6.64127707e-01 1.73341826e-01
-1.39681435e+00 8.91532838e-01 5.44462347e+00 9.45368826e-01
-1.23049641e+00 6.51962161e-01 -3.42364073e-01 -1.92656100e-01
3.66526395e-02 3.65990549e-02 -1.26707935e+00 3.51246685e-01
8.71108115e-01 -1.85067877e-01 3.16123486e-01 7.93870568e-01
-1.30457744e-01 1.56621546e-01 -9.61242318e-01 1.12817371e+00
3.26255947e-01 -1.41884279e+00 -1.13302529e-01 1.88821256e-01
6.38009906e-01 7.55127430e-01 3.61922204e-01 4.30622488e-01
4.42293257e-01 -7.14461446e-01 1.09639287e+00 7.68937051e-01
8.11964571e-01 1.16890945e-01 2.08565325e-01 3.04597229e-01
-1.57766628e+00 -8.69337842e-02 -3.45879257e-01 2.54450351e-01
2.43329927e-01 4.37864333e-01 -3.66717905e-01 5.88369131e-01
9.12563264e-01 1.17840886e+00 -7.77637661e-01 1.47707057e+00
9.35154483e-02 3.17004979e-01 -6.49184465e-01 6.28472716e-02
-2.82589602e-03 1.38455778e-01 8.72552335e-01 9.50650334e-01
2.79150277e-01 -2.26958469e-01 4.48599190e-01 9.10807014e-01
9.99094546e-02 -2.51012474e-01 -8.81751239e-01 3.37210968e-02
4.56795394e-01 1.35278022e+00 -6.60280704e-01 -2.35108167e-01
-7.22791970e-01 6.79287612e-01 3.42687666e-01 1.79801062e-01
-1.05002701e+00 7.71820918e-02 8.57257307e-01 8.34180508e-03
7.52655923e-01 -4.89826143e-01 8.36902484e-02 -1.35640478e+00
6.56335428e-02 -6.72930658e-01 2.77578294e-01 -7.57098556e-01
-1.40333939e+00 4.88214672e-01 -4.74268757e-02 -1.48750913e+00
2.57456541e-01 -7.44731665e-01 -4.23166066e-01 5.31650484e-01
-1.68813241e+00 -1.74149299e+00 -5.92848718e-01 8.29944551e-01
4.65725064e-01 -1.52134940e-01 5.43571591e-01 6.13243699e-01
-4.16201562e-01 4.10405695e-01 5.11475980e-01 -1.00545131e-01
1.00150645e+00 -1.06854427e+00 4.09435719e-01 8.36762369e-01
5.52470982e-01 5.84052563e-01 8.06743920e-01 -7.32686043e-01
-1.90369654e+00 -1.45223105e+00 4.82167035e-01 -1.15976501e+00
1.07885504e+00 -6.81888044e-01 -7.62592733e-01 8.98250103e-01
9.28995293e-03 6.98306382e-01 2.98726767e-01 6.82661757e-02
-6.08891070e-01 -1.88953444e-01 -7.68462837e-01 4.64354604e-01
1.26412070e+00 -4.65243965e-01 -5.61708868e-01 5.47174692e-01
8.14677000e-01 -4.94267493e-01 -9.24734175e-01 3.23643088e-01
7.23754227e-01 -4.49284494e-01 1.36939442e+00 -4.35208350e-01
-2.86301076e-01 -6.97385430e-01 -1.92049369e-01 -1.01797831e+00
-2.70627946e-01 -6.69634759e-01 -6.14488125e-01 1.17236364e+00
-7.14689717e-02 -6.63104415e-01 4.92159367e-01 1.72811136e-01
-2.14493960e-01 -3.12611252e-01 -1.11805606e+00 -1.36956036e+00
-8.13011676e-02 -5.61968923e-01 3.02947551e-01 7.33238101e-01
-6.03354633e-01 1.68976963e-01 -5.44806182e-01 4.25770074e-01
1.31048763e+00 2.31396809e-01 1.00189340e+00 -1.60738039e+00
-2.09745154e-01 -3.98155242e-01 -7.85501778e-01 -1.29382670e+00
1.47851571e-01 -9.57639158e-01 2.81918228e-01 -1.36959493e+00
2.32551038e-01 -8.35976958e-01 -3.04515362e-01 3.44740838e-01
9.03528631e-02 5.01482189e-01 4.06559348e-01 6.07474864e-01
-1.39271498e+00 8.52576435e-01 1.15787411e+00 -4.48763341e-01
1.77277386e-01 1.75558999e-02 -4.26089793e-01 5.03262877e-01
4.29129213e-01 -7.91206121e-01 -2.19804138e-01 -5.89380920e-01
-1.61504343e-01 -4.23442692e-01 1.04921842e+00 -1.24442673e+00
6.59904242e-01 -1.02166355e-01 3.53991203e-02 -7.34152138e-01
7.22977519e-01 -8.06971431e-01 2.84275800e-01 4.88463491e-01
1.60859735e-03 -1.31037951e-01 3.16850126e-01 1.09848046e+00
1.72315106e-01 1.55459538e-01 6.69727981e-01 1.72785506e-01
-1.35788834e+00 6.94342017e-01 -1.31470338e-01 1.43579766e-01
1.33569968e+00 -2.25674763e-01 -6.79964423e-01 -4.36411947e-02
-6.49560511e-01 3.42252016e-01 6.95627987e-01 1.03619838e+00
3.67458820e-01 -1.37696314e+00 -6.71841264e-01 9.52644944e-02
6.63765013e-01 -1.98352143e-01 2.87699431e-01 1.17016399e+00
7.77343893e-03 4.24810410e-01 -8.00392255e-02 -1.21603000e+00
-1.42871189e+00 8.93805981e-01 3.30799252e-01 -1.48945507e-02
-1.03107464e+00 6.41718268e-01 4.09002870e-01 -4.07487065e-01
3.20787907e-01 -3.33246291e-01 -4.17853221e-02 -2.52211571e-01
5.72101533e-01 6.82692528e-02 -1.55048102e-01 -1.01652455e+00
-5.52790821e-01 8.12955916e-01 5.11945970e-02 -3.56437378e-02
1.02433085e+00 -1.89198494e-01 2.38845065e-01 6.04158759e-01
7.97496140e-01 -1.73542321e-01 -1.57351744e+00 -6.77438259e-01
1.32231982e-02 -5.45924246e-01 2.49416277e-01 -5.38205504e-01
-9.06329691e-01 8.03118408e-01 1.15826011e+00 -5.95012344e-02
6.23220563e-01 4.75470692e-01 5.34565747e-01 5.85555196e-01
5.56955576e-01 -6.14992142e-01 3.00171345e-01 4.81618822e-01
6.72440708e-01 -1.49035943e+00 1.35067466e-03 -1.96118101e-01
-4.88004506e-01 7.99156427e-01 5.76621473e-01 4.47606556e-02
8.55514050e-01 1.83950394e-01 1.84036002e-01 -5.66632926e-01
-8.18922400e-01 -6.76729798e-01 6.09189808e-01 8.60734761e-01
-1.07607663e-01 -1.12461306e-01 2.81434685e-01 1.64511934e-01
2.13051409e-01 3.28998677e-02 1.76770151e-01 8.93807352e-01
-5.66718519e-01 -7.67702103e-01 -5.85895479e-01 4.53319401e-01
-1.00899994e-01 9.92057770e-02 -5.12286574e-02 7.80457199e-01
2.61444181e-01 8.86122227e-01 1.20167166e-01 -1.45324156e-01
2.21125036e-01 -2.65378743e-01 6.86006248e-01 -5.44580698e-01
-2.49954939e-01 -2.45021462e-01 -2.88118273e-01 -6.55330420e-01
-6.88307464e-01 -1.04038656e+00 -7.86738634e-01 -1.90932095e-01
-5.17104149e-01 -2.25946411e-01 8.34866107e-01 9.60478842e-01
5.61798990e-01 5.34959733e-01 -8.91503692e-03 -1.05484796e+00
-6.49485409e-01 -7.96197832e-01 -4.25301015e-01 5.29797196e-01
7.45997071e-01 -1.32372594e+00 -1.09165475e-01 1.00630961e-01] | [6.352503776550293, -2.0817127227783203] |
5a36b993-27f2-4ebf-9a35-f0a9c016e752 | learning-on-graphs-under-label-noise | 2306.08194 | null | https://arxiv.org/abs/2306.08194v1 | https://arxiv.org/pdf/2306.08194v1.pdf | Learning on Graphs under Label Noise | Node classification on graphs is a significant task with a wide range of applications, including social analysis and anomaly detection. Even though graph neural networks (GNNs) have produced promising results on this task, current techniques often presume that label information of nodes is accurate, which may not be the case in real-world applications. To tackle this issue, we investigate the problem of learning on graphs with label noise and develop a novel approach dubbed Consistent Graph Neural Network (CGNN) to solve it. Specifically, we employ graph contrastive learning as a regularization term, which promotes two views of augmented nodes to have consistent representations. Since this regularization term cannot utilize label information, it can enhance the robustness of node representations to label noise. Moreover, to detect noisy labels on the graph, we present a sample selection technique based on the homophily assumption, which identifies noisy nodes by measuring the consistency between the labels with their neighbors. Finally, we purify these confident noisy labels to permit efficient semantic graph learning. Extensive experiments on three well-known benchmark datasets demonstrate the superiority of our CGNN over competing approaches. | ['Ming Zhang', 'Wei Ju', 'Yusheng Zhao', 'Yifang Qin', 'Xiao Luo', 'Jingyang Yuan'] | 2023-06-14 | null | null | null | null | ['contrastive-learning', 'contrastive-learning', 'anomaly-detection'] | ['computer-vision', 'methodology', 'methodology'] | [ 2.52480507e-01 1.98472768e-01 -3.01248461e-01 -4.86304998e-01
-1.66725263e-01 -3.77440244e-01 3.88118774e-01 3.46473902e-01
-4.70745750e-02 5.41112840e-01 -1.18471667e-01 -1.26724765e-01
-1.53874129e-01 -1.05733895e+00 -4.91600215e-01 -8.26043367e-01
-1.04767671e-02 2.75107443e-01 1.37810066e-01 1.90376583e-02
5.54506294e-02 2.66241491e-01 -1.17301464e+00 -2.31784001e-01
9.01369810e-01 9.15926576e-01 -7.97248781e-02 -4.38130664e-04
-2.79678822e-01 8.24733377e-01 -5.50993145e-01 -4.57293391e-01
2.15295494e-01 -3.90781760e-01 -6.99365377e-01 3.19536239e-01
2.36575976e-01 3.73062007e-02 -4.10198033e-01 1.70453811e+00
1.76186368e-01 2.12817088e-01 6.27602220e-01 -1.66740847e+00
-6.76549137e-01 6.18736744e-01 -7.26319373e-01 7.90619403e-02
3.16436104e-02 -3.03772479e-01 1.39084244e+00 -5.13756633e-01
6.23453021e-01 1.31902897e+00 7.38352895e-01 4.78367984e-01
-1.28066170e+00 -7.48329282e-01 5.79690993e-01 1.62729368e-01
-1.42546761e+00 -8.77696872e-02 1.26780462e+00 -2.60102451e-01
2.32354999e-02 1.55126169e-01 5.01003802e-01 1.06581891e+00
-2.24273205e-01 6.98977232e-01 9.47494090e-01 -1.74818054e-01
3.98717731e-01 1.33866593e-01 3.74641478e-01 8.91349316e-01
5.42726815e-01 -3.19783837e-01 -2.49723837e-01 -3.23201030e-01
5.85471511e-01 2.48038486e-01 -4.66913939e-01 -6.36371255e-01
-8.88919473e-01 9.29247856e-01 8.47750187e-01 3.65726918e-01
-3.96915197e-01 7.18436390e-02 5.38277626e-01 3.94062698e-01
7.97034323e-01 2.33540624e-01 -6.41578436e-02 5.35615921e-01
-4.33669299e-01 -2.66571313e-01 7.30520844e-01 7.00416744e-01
8.23228002e-01 9.68877748e-02 1.42722368e-01 8.85601640e-01
4.13133323e-01 1.97954580e-01 2.28843227e-01 -8.16334426e-01
2.43047059e-01 1.06144094e+00 -3.47391278e-01 -1.70512784e+00
-3.84180427e-01 -6.86149895e-01 -1.42621994e+00 -9.87213850e-02
2.83995986e-01 1.48313165e-01 -8.09025407e-01 1.94600916e+00
5.04968226e-01 4.79668111e-01 -1.29135430e-01 9.39779997e-01
9.75807548e-01 4.54381168e-01 1.55504584e-01 -3.72709125e-01
8.96942258e-01 -6.68000579e-01 -7.48700261e-01 -1.27777174e-01
7.90857375e-01 -2.12598845e-01 8.16460431e-01 1.41860530e-01
-4.86575991e-01 -2.87484020e-01 -9.56908107e-01 3.40986967e-01
-3.43713820e-01 -3.05268705e-01 7.63980210e-01 5.93534052e-01
-8.97668421e-01 6.75960660e-01 -5.46595991e-01 -3.85011613e-01
5.31831622e-01 1.51512221e-01 -5.40607512e-01 -2.19839528e-01
-1.06055605e+00 3.69584113e-01 5.11053979e-01 2.94370472e-01
-5.79783618e-01 -1.13620520e-01 -1.03550959e+00 1.86407730e-01
8.44132245e-01 -3.04096967e-01 7.54389703e-01 -1.20553339e+00
-8.51665437e-01 7.54554152e-01 -6.80079013e-02 -1.07601181e-01
3.70526850e-01 4.76868540e-01 -5.27368546e-01 1.40222207e-01
3.44430894e-01 2.91250408e-01 8.29811573e-01 -1.54497039e+00
-3.05326372e-01 -5.28810263e-01 4.43073995e-02 1.92657784e-01
-5.51988184e-01 -3.77242893e-01 -3.94060493e-01 -7.16185153e-01
8.27441514e-01 -9.47525740e-01 -1.98088124e-01 7.31539428e-02
-6.93635941e-01 -5.10008454e-01 7.78954864e-01 -4.64928359e-01
1.05021071e+00 -2.24222469e+00 3.90810296e-02 8.65169108e-01
7.56048501e-01 1.84255749e-01 -2.28389829e-01 1.47529006e-01
-1.19013712e-01 2.58300483e-01 -3.48344147e-01 -3.14770252e-01
-1.36364818e-01 3.75104398e-01 -8.75415062e-05 6.82252884e-01
2.95140799e-02 7.47668982e-01 -1.01425672e+00 -4.41801190e-01
-3.45046781e-02 3.35069150e-01 -1.14931688e-01 1.39888689e-01
-1.24303818e-01 4.31221306e-01 -6.90854788e-01 7.20986426e-01
7.84218431e-01 -8.04529667e-01 6.13323748e-01 -1.93088904e-01
6.95027471e-01 -1.03105858e-01 -1.36261356e+00 1.21190703e+00
-6.67649209e-02 3.09225410e-01 2.50854433e-01 -1.49127173e+00
1.18739522e+00 1.45153061e-01 4.62274075e-01 -5.85067749e-01
2.15879738e-01 1.37218207e-01 -6.66308552e-02 -3.86632979e-01
8.04667622e-02 -2.31144018e-03 1.91505417e-01 4.68718171e-01
-1.77401707e-01 2.97492892e-01 1.40251115e-01 4.94495153e-01
1.22202623e+00 -3.68340641e-01 2.01329082e-01 -1.84498936e-01
6.04062557e-01 -3.82996261e-01 8.88887167e-01 6.38431847e-01
-3.68677974e-01 5.04797995e-01 8.45071197e-01 -5.19581854e-01
-5.61381280e-01 -6.54682338e-01 1.81287467e-01 1.06459248e+00
5.55912614e-01 -5.02446949e-01 -6.46638632e-01 -1.20560801e+00
-4.24964800e-02 3.25424612e-01 -5.89126885e-01 -3.76678228e-01
-3.63222748e-01 -8.34843397e-01 1.42200142e-01 3.07940602e-01
5.08166432e-01 -9.30648625e-01 2.35114038e-01 1.02090292e-01
-3.16272110e-01 -1.02084649e+00 -2.83896893e-01 5.65480068e-02
-6.56159520e-01 -1.50667810e+00 -3.57460946e-01 -1.07456005e+00
1.08746183e+00 7.28236556e-01 1.06241775e+00 7.24011183e-01
7.35666603e-02 3.35635453e-01 -4.45243686e-01 5.21492660e-02
-4.44041848e-01 1.57611266e-01 -6.74966536e-03 3.31554592e-01
3.15648198e-01 -7.00568795e-01 -3.26634526e-01 2.92092264e-01
-9.29619014e-01 -2.32651651e-01 4.67018038e-01 8.80460978e-01
5.97002327e-01 4.55622286e-01 8.23679566e-01 -1.36230814e+00
8.28164995e-01 -7.89289057e-01 -4.86412048e-01 4.17649746e-01
-8.91787648e-01 7.15540536e-03 8.18763793e-01 -2.86473632e-01
-7.22218275e-01 -4.66276631e-02 2.10148115e-02 -4.62434649e-01
5.91321178e-02 7.04167724e-01 -2.77380198e-01 -2.66543806e-01
4.27341670e-01 1.01896510e-01 1.88252389e-01 -3.52515280e-01
1.40368909e-01 4.57310617e-01 2.60972351e-01 -3.24342668e-01
8.55944932e-01 5.25007904e-01 3.64780575e-01 -6.55692220e-01
-8.75748754e-01 -5.04244685e-01 -2.84291148e-01 -3.47419530e-01
4.74860311e-01 -7.12281764e-01 -6.93801343e-01 3.06503356e-01
-9.04495239e-01 1.36471823e-01 7.35170692e-02 2.28583589e-01
-2.44345199e-02 6.67885840e-01 -6.26895607e-01 -8.04997087e-01
-1.98687032e-01 -9.78309989e-01 7.09025502e-01 1.99472100e-01
6.55855834e-02 -1.25606287e+00 -2.12566838e-01 1.60037279e-01
1.75475821e-01 3.47107112e-01 1.06309199e+00 -1.04477477e+00
-5.34258127e-01 -2.70381749e-01 -5.63044786e-01 4.28503901e-01
2.69774318e-01 -2.05849588e-01 -7.84495652e-01 -4.76831228e-01
-1.65272012e-01 -3.70017499e-01 1.03131628e+00 2.07044959e-01
1.18282115e+00 -1.29306495e-01 -4.93792206e-01 3.48372042e-01
1.21556973e+00 -1.15294360e-01 1.21182397e-01 8.74622446e-03
1.10827160e+00 8.12835157e-01 3.42770606e-01 2.90697724e-01
3.98470044e-01 2.69449621e-01 8.03543210e-01 -5.97330630e-02
5.80946878e-02 -4.25544053e-01 -1.23114832e-01 9.13325846e-01
2.46364698e-01 -4.55590338e-01 -7.15952098e-01 2.80850738e-01
-1.98954761e+00 -5.38215280e-01 -1.69750407e-01 1.97328937e+00
2.77201533e-01 1.85019895e-01 1.75079517e-02 2.73649842e-01
1.32387650e+00 4.97079700e-01 -6.80889368e-01 2.00342312e-01
-1.18309528e-01 -3.15191478e-01 2.08613127e-01 2.23446190e-01
-1.10749781e+00 7.08748341e-01 5.13045359e+00 7.25429475e-01
-9.44042325e-01 1.84665807e-02 1.01018345e+00 5.39560139e-01
-5.80838323e-01 1.35844462e-02 -3.48663807e-01 5.83390653e-01
4.09180343e-01 -7.67060891e-02 3.75189334e-01 9.83950794e-01
3.98345217e-02 1.95141777e-01 -8.69269013e-01 1.03639317e+00
6.65249452e-02 -1.02426279e+00 4.95568551e-02 7.46519789e-02
7.02335596e-01 -2.45088667e-01 -2.17714459e-01 3.10126215e-01
4.22185361e-01 -8.96078944e-01 2.43650392e-01 2.93490469e-01
4.69162345e-01 -7.51806438e-01 9.23886061e-01 4.18307900e-01
-1.31289065e+00 -5.73879629e-02 -5.18495262e-01 1.61284477e-01
-1.89882323e-01 8.80303621e-01 -6.62375152e-01 5.63174486e-01
5.30701160e-01 1.00160396e+00 -7.01154709e-01 9.03040111e-01
-2.74500042e-01 5.87684035e-01 -2.81318128e-01 -5.68733849e-02
7.98700526e-02 -4.92700458e-01 4.60405767e-01 5.89097261e-01
1.75998792e-01 -3.18977684e-02 5.39165735e-01 8.47539544e-01
-5.58120847e-01 3.06546539e-01 -7.65133440e-01 -1.14336617e-01
7.78258860e-01 1.45632195e+00 -1.13738358e+00 -6.43557012e-02
-4.06568795e-01 8.26944709e-01 7.36163795e-01 4.11140919e-01
-4.69181806e-01 -2.31273860e-01 2.79803276e-01 -1.43504925e-02
-1.52340174e-01 8.57243091e-02 -1.83118489e-02 -1.15533233e+00
3.20791453e-01 -6.90483451e-01 5.81722081e-01 -3.72325718e-01
-1.71411705e+00 4.59620833e-01 -3.47458601e-01 -1.16007292e+00
1.78264275e-01 -2.83593088e-01 -7.82992184e-01 3.86447728e-01
-1.52662897e+00 -1.08059633e+00 -5.75579345e-01 4.63238508e-01
9.58884507e-02 -3.22565101e-02 5.42273223e-01 2.51520365e-01
-7.87519455e-01 4.66005415e-01 1.49997965e-01 3.99879724e-01
5.61146975e-01 -1.20587897e+00 2.90635109e-01 7.70458817e-01
2.64895976e-01 4.87061173e-01 4.20906901e-01 -8.88552964e-01
-1.21157289e+00 -1.30848956e+00 5.63024282e-01 1.88550934e-01
6.01044536e-01 -2.64090836e-01 -1.33973145e+00 7.21515954e-01
-2.76430160e-01 5.91678858e-01 4.52167958e-01 2.80923039e-01
-5.79553425e-01 -2.22949311e-01 -1.13698828e+00 4.23494935e-01
1.25271583e+00 -4.60661232e-01 -5.85854948e-02 5.25879025e-01
6.00428939e-01 1.04748737e-02 -6.57780707e-01 6.40596271e-01
7.70051777e-02 -9.72343147e-01 6.75751507e-01 -3.43607038e-01
-1.19514810e-02 -3.35164487e-01 1.39694484e-02 -1.53883195e+00
-3.26342344e-01 -2.47741461e-01 -2.62566041e-02 1.37467325e+00
9.34238806e-02 -8.73541772e-01 1.10875154e+00 4.01701033e-01
1.13640778e-01 -5.52201271e-01 -8.14928234e-01 -7.22673655e-01
-3.84508431e-01 -1.50424942e-01 5.57585776e-01 1.45097446e+00
-8.98424461e-02 4.21225876e-01 -3.90038967e-01 2.58964717e-01
8.53662312e-01 2.59643272e-02 4.93827671e-01 -1.82070947e+00
-1.19208843e-02 -3.65725815e-01 -6.48697495e-01 -6.60533965e-01
8.08520317e-01 -1.14525187e+00 -4.16188827e-03 -1.60582232e+00
2.41703928e-01 -6.91072345e-01 -5.06856620e-01 4.67153192e-01
-4.47678477e-01 2.36584947e-01 -1.21475823e-01 3.74624074e-01
-9.09834087e-01 6.40476525e-01 1.14026320e+00 -3.69801641e-01
-8.67044255e-02 1.15622580e-01 -7.78350949e-01 7.67749667e-01
8.42659175e-01 -5.58894813e-01 -7.00587094e-01 -5.48762605e-02
3.49497616e-01 -3.72829102e-02 3.69476795e-01 -8.38608742e-01
1.47634074e-01 -5.86615829e-03 9.13896784e-02 -1.45664185e-01
1.04459494e-01 -9.77980733e-01 1.49501666e-01 3.17047745e-01
-4.61634338e-01 -7.37028494e-02 -5.72376370e-01 1.07015288e+00
-3.34862798e-01 -2.88316250e-01 8.12115967e-01 -2.75233299e-01
-7.97694981e-01 4.87372398e-01 4.46528792e-02 1.30558446e-01
9.47369576e-01 1.44752013e-02 -4.57289457e-01 -5.68055570e-01
-7.23663330e-01 4.60862994e-01 4.67532605e-01 2.92843103e-01
6.63598299e-01 -1.40199983e+00 -5.75886488e-01 2.35970601e-01
3.22381020e-01 8.45913962e-02 1.40926316e-01 6.52507007e-01
-2.63389677e-01 -2.07995877e-01 2.31221795e-01 -6.55323625e-01
-1.26972222e+00 5.99449098e-01 2.54696995e-01 -9.63541567e-02
-5.99836171e-01 7.19649434e-01 2.17766568e-01 -5.64021468e-01
4.19160485e-01 3.01256329e-01 -4.88295972e-01 8.52063224e-02
2.37937853e-01 3.61777365e-01 5.36694750e-02 -6.51472509e-01
-2.44234800e-01 3.40804189e-01 -2.68400997e-01 5.54174483e-01
1.19344711e+00 -2.30444491e-01 -3.72350812e-01 4.96874899e-01
1.16332507e+00 -2.92929918e-01 -9.33981121e-01 -6.67364061e-01
2.51283735e-01 -3.70810837e-01 3.70614640e-02 -1.93203777e-01
-1.43681848e+00 5.65628827e-01 4.15052116e-01 7.92893648e-01
9.62945402e-01 1.14912027e-02 6.36937439e-01 5.98428965e-01
1.78268701e-01 -1.07042670e+00 1.92166537e-01 2.02255934e-01
4.23267961e-01 -1.65824819e+00 -6.12493642e-02 -8.95880282e-01
-4.35342669e-01 8.76397669e-01 7.46350527e-01 -1.24170072e-01
7.70186663e-01 -2.88395584e-01 -4.28180816e-03 -3.12593937e-01
-3.83010924e-01 -1.98300675e-01 1.06805906e-01 5.54652393e-01
7.47965276e-02 1.33797854e-01 -2.17962354e-01 3.13448429e-01
3.26964468e-01 -4.92035747e-01 4.26060587e-01 7.41088092e-01
-4.00038779e-01 -9.54275072e-01 -5.22789583e-02 7.09965348e-01
-3.43459219e-01 8.00473839e-02 -5.84632337e-01 6.71295702e-01
-1.42336056e-01 1.01301765e+00 -1.19836718e-01 -5.01395106e-01
1.18135497e-01 -4.06765044e-02 -1.03599038e-02 -5.51270366e-01
-2.96059489e-01 -2.00239476e-02 -1.37745796e-04 -3.74611944e-01
-6.17861390e-01 -2.22266123e-01 -1.12917161e+00 -4.74311769e-01
-5.32803774e-01 4.30263430e-01 4.63221014e-01 9.25471961e-01
2.84801066e-01 3.81248534e-01 9.36165631e-01 -5.05296290e-01
-5.99631667e-01 -7.51405299e-01 -1.02459025e+00 9.27708328e-01
1.52184725e-01 -7.41520107e-01 -7.87948549e-01 -4.46981549e-01] | [7.303913116455078, 6.052811145782471] |
27b71a63-8ad0-439a-8fc4-75e531b13500 | exploring-optimal-granularity-for-extractive | 2209.10041 | null | https://arxiv.org/abs/2209.10041v2 | https://arxiv.org/pdf/2209.10041v2.pdf | Exploring Optimal Granularity for Extractive Summarization of Unstructured Health Records: Analysis of the Largest Multi-Institutional Archive of Health Records in Japan | Automated summarization of clinical texts can reduce the burden of medical professionals. "Discharge summaries" are one promising application of the summarization, because they can be generated from daily inpatient records. Our preliminary experiment suggests that 20-31% of the descriptions in discharge summaries overlap with the content of the inpatient records. However, it remains unclear how the summaries should be generated from the unstructured source. To decompose the physician's summarization process, this study aimed to identify the optimal granularity in summarization. We first defined three types of summarization units with different granularities to compare the performance of the discharge summary generation: whole sentences, clinical segments, and clauses. We defined clinical segments in this study, aiming to express the smallest medically meaningful concepts. To obtain the clinical segments, it was necessary to automatically split the texts in the first stage of the pipeline. Accordingly, we compared rule-based methods and a machine learning method, and the latter outperformed the formers with an F1 score of 0.846 in the splitting task. Next, we experimentally measured the accuracy of extractive summarization using the three types of units, based on the ROUGE-1 metric, on a multi-institutional national archive of health records in Japan. The measured accuracies of extractive summarization using whole sentences, clinical segments, and clauses were 31.91, 36.15, and 25.18, respectively. We found that the clinical segments yielded higher accuracy than sentences and clauses. This result indicates that summarization of inpatient records demands finer granularity than sentence-oriented processing. Although we used only Japanese health records, it can be interpreted as follows: physicians extract "concepts of medical significance" from patient records and recombine them ... | ['Takashi Okumura', 'Yuji Matsumoto', 'Hiromasa Horiguchi', 'Mamoru Komachi', 'Kenichiro Ando'] | 2022-09-20 | null | null | null | null | ['extractive-summarization'] | ['natural-language-processing'] | [ 4.26738769e-01 5.71600199e-01 -1.72133520e-01 -2.92611003e-01
-1.30636132e+00 -5.42899907e-01 5.57002239e-02 1.05487013e+00
-4.69502687e-01 1.23728430e+00 9.93811250e-01 -2.17507899e-01
-2.37904742e-01 -5.59928179e-01 -2.14883298e-01 -5.62100828e-01
2.37204447e-01 7.20962048e-01 -1.20284133e-01 6.87652603e-02
4.97203946e-01 2.49615416e-01 -1.03224337e+00 8.74357879e-01
1.63264358e+00 3.27918619e-01 2.31607631e-01 7.50660181e-01
-2.87155330e-01 6.94400191e-01 -1.21189797e+00 -3.60462248e-01
-2.72451639e-01 -9.62848723e-01 -9.03769910e-01 2.36762822e-01
-4.69046570e-02 -3.40692848e-01 1.15767106e-01 9.90479767e-01
6.73757136e-01 -2.06254587e-01 7.07704604e-01 -5.56747019e-01
-2.75049835e-01 1.01419151e+00 -2.75392771e-01 2.50092983e-01
8.15679967e-01 -4.65774201e-02 8.86307836e-01 -3.68931085e-01
8.38997602e-01 8.61095190e-01 5.08959889e-01 5.94982147e-01
-1.00015342e+00 -2.57454455e-01 -8.63055959e-02 -2.42650047e-01
-1.08612704e+00 -4.50225919e-01 3.67061406e-01 -5.35874665e-01
1.07374287e+00 6.53703511e-01 8.53419483e-01 7.97121704e-01
7.31196463e-01 7.12511241e-01 5.28308928e-01 -4.17870224e-01
3.40347826e-01 1.34214863e-01 5.30904293e-01 4.53841627e-01
1.03490412e+00 -7.17593551e-01 -1.34340137e-01 -6.72358036e-01
2.94087917e-01 -7.56425932e-02 -5.95287979e-01 5.98388612e-01
-1.43809915e+00 9.57639515e-01 -7.95891508e-02 5.88857710e-01
-7.17833877e-01 -5.11795521e-01 7.97064781e-01 -1.72724545e-01
5.18516064e-01 8.36327195e-01 -3.54524285e-01 -2.08439767e-01
-1.17813790e+00 1.99990541e-01 1.07624853e+00 1.08433819e+00
-2.15893872e-02 -2.39424258e-01 -6.71250880e-01 7.57053256e-01
-6.84508607e-02 3.87803912e-01 8.22103322e-01 -7.67100036e-01
8.23932707e-01 7.70051122e-01 9.10846666e-02 -7.93043733e-01
-5.26717961e-01 -3.71260196e-01 -1.10307515e+00 -9.31748509e-01
3.05589698e-02 -6.15910411e-01 -7.47577190e-01 1.19735408e+00
-6.87043518e-02 -4.24022377e-01 5.83166182e-01 5.33985615e-01
1.35044932e+00 7.21605062e-01 1.24026641e-01 -9.98630226e-01
1.64574122e+00 -8.05878997e-01 -1.26613808e+00 1.08693033e-01
1.04636467e+00 -7.99633861e-01 5.13932765e-01 3.48124504e-01
-1.31513083e+00 -2.13547930e-01 -8.51573348e-01 1.83382347e-01
1.92446187e-01 4.20635015e-01 2.12115556e-01 3.99840534e-01
-8.63859177e-01 5.27244985e-01 -8.45286310e-01 -5.43781936e-01
2.08598450e-01 1.41232193e-01 -2.45704785e-01 2.32780814e-01
-1.09247410e+00 7.36498952e-01 7.15978324e-01 -1.32201329e-01
-6.49962500e-02 -5.41602194e-01 -9.01904583e-01 4.08760607e-01
8.26470703e-02 -1.26976085e+00 1.36204004e+00 -5.20166934e-01
-1.08274162e+00 6.57153606e-01 -4.55108821e-01 -4.73433048e-01
1.92241803e-01 -1.66499585e-01 -2.09242374e-01 7.10420966e-01
3.99774224e-01 2.73483843e-01 9.44113880e-02 -9.89083827e-01
-8.29065323e-01 -2.47653961e-01 -3.15310419e-01 2.67184347e-01
-9.40992832e-02 9.09911767e-02 -1.52261600e-01 -6.48101568e-01
6.07450902e-02 -6.49034917e-01 -4.70936656e-01 -1.02501214e+00
-7.69162476e-01 -2.47374430e-01 -2.48519741e-02 -9.91277754e-01
1.88793099e+00 -1.91784632e+00 -1.79796055e-01 -1.20930068e-01
3.95171940e-01 3.43380243e-01 6.06323890e-02 7.45156407e-01
-1.14398956e-01 5.41971743e-01 -4.44635421e-01 -9.17518958e-02
-3.42949361e-01 -1.15956301e-02 -2.65665740e-01 4.05464023e-02
2.29388669e-01 8.26727331e-01 -1.02835178e+00 -1.08519125e+00
-2.45120764e-01 -2.30595674e-02 -6.83091044e-01 2.94449747e-01
2.86750048e-02 2.77011305e-01 -6.98665857e-01 4.72413808e-01
4.09160137e-01 -2.04831094e-01 4.30549353e-01 -1.08183727e-01
-1.12492792e-01 6.12839997e-01 -5.47348440e-01 1.49440897e+00
2.13581875e-01 2.99525529e-01 -2.71640867e-01 -8.64909053e-01
7.82465458e-01 7.50694335e-01 7.69151866e-01 -1.84248656e-01
7.74997324e-02 2.30289698e-01 1.69146195e-01 -1.01379478e+00
8.38424742e-01 -2.70752788e-01 -4.25348729e-01 2.70606726e-01
-3.03546637e-01 -2.38414779e-01 6.82671368e-01 4.80364293e-01
1.26239526e+00 -5.07016599e-01 9.67178762e-01 -1.76088870e-01
4.24808621e-01 5.39806962e-01 7.76482701e-01 7.69364297e-01
-1.96870300e-03 9.12279844e-01 9.15584326e-01 -5.14184013e-02
-6.69132531e-01 -7.59172082e-01 -1.69974014e-01 1.88521177e-01
-2.16568470e-01 -9.01874602e-01 -1.09347343e+00 -7.15210378e-01
-2.86702335e-01 1.04026759e+00 -2.35723317e-01 -2.12748840e-01
-6.33295536e-01 -9.77071643e-01 6.60749495e-01 3.94573778e-01
1.97612241e-01 -1.03870010e+00 -6.63909853e-01 5.76789081e-01
-9.12014902e-01 -9.91408944e-01 -6.98176265e-01 -1.30589142e-01
-1.18121934e+00 -1.13489926e+00 -6.67746186e-01 -6.67695343e-01
8.74152005e-01 2.07380895e-02 8.91845107e-01 2.25627013e-02
-5.71014360e-02 1.71128750e-01 -6.09685659e-01 -5.32865763e-01
-9.10195529e-01 4.25507128e-01 -7.07940087e-02 -5.33492804e-01
3.12729150e-01 -1.86160296e-01 -6.00198865e-01 -3.66897613e-01
-1.09856486e+00 1.96791679e-01 8.46660018e-01 8.38598609e-01
4.67660218e-01 -1.19970135e-01 9.35068071e-01 -1.22144473e+00
1.35784900e+00 -3.65815967e-01 1.69877574e-01 4.21593666e-01
-4.94830549e-01 7.45870247e-02 6.97806716e-01 -1.93807364e-01
-9.89098668e-01 -7.57875368e-02 -3.60061020e-01 2.79186994e-01
-2.79615611e-01 8.19561839e-01 2.00932436e-02 9.29183424e-01
5.64206660e-01 3.29266906e-01 -6.12046570e-02 -3.67165595e-01
-1.31672382e-01 1.05287814e+00 2.87788779e-01 -3.32040489e-01
6.40926659e-02 3.87828760e-02 -5.25525570e-01 -8.63307893e-01
-7.83182442e-01 -6.89657271e-01 -3.78625989e-01 1.58165827e-01
1.18471646e+00 -7.05517590e-01 -5.68539381e-01 -8.05064067e-02
-1.53964972e+00 2.47569725e-01 -4.96464998e-01 7.91630030e-01
-4.28677917e-01 8.58893931e-01 -7.87647128e-01 -5.09896696e-01
-1.08210278e+00 -8.67843091e-01 1.11681199e+00 3.18062961e-01
-9.43052948e-01 -7.32263863e-01 2.38535851e-01 3.26965958e-01
-1.16749384e-01 4.38064039e-01 1.24742377e+00 -1.23127401e+00
7.93753639e-02 -2.26992488e-01 1.72433369e-02 1.71146333e-01
6.74510241e-01 1.39728174e-01 -3.91629964e-01 -1.58208445e-01
3.49085927e-01 1.76049262e-01 7.95052290e-01 7.03338802e-01
8.67180407e-01 -7.54081309e-01 -5.07307231e-01 8.40113238e-02
1.19720113e+00 6.59388304e-01 6.21799350e-01 2.92383060e-02
3.04882288e-01 6.82365775e-01 7.41405368e-01 6.65058613e-01
3.34696263e-01 8.57597217e-02 -3.16858619e-01 3.22708152e-02
2.60389745e-01 -1.57135129e-01 2.69041926e-01 1.51762974e+00
-8.41702893e-02 -4.41737920e-01 -8.69921446e-01 5.35542309e-01
-1.91546524e+00 -1.03830874e+00 -2.61503845e-01 1.93152332e+00
1.12573886e+00 1.47557110e-01 -1.69980770e-03 -4.35125753e-02
6.98490381e-01 -1.31031096e-01 -1.88246787e-01 -7.11043537e-01
1.18190810e-01 -9.14388988e-03 2.01881871e-01 3.48781288e-01
-7.58903563e-01 5.37628114e-01 6.46227121e+00 4.96165603e-01
-7.84766197e-01 -2.27354243e-01 6.33323312e-01 -2.26465121e-01
-3.09066951e-01 -2.38456830e-01 -1.00372481e+00 6.58957124e-01
1.22273040e+00 -6.39860809e-01 -4.85393435e-01 6.01055205e-01
7.51290321e-01 -5.02233624e-01 -1.14341331e+00 5.86741865e-01
2.62532145e-01 -1.53183055e+00 5.05928278e-01 3.10086384e-02
8.06082547e-01 -4.41147506e-01 -6.83121979e-01 1.34659499e-01
-4.26859707e-02 -6.83469117e-01 3.25600356e-01 7.08444417e-01
6.72068298e-01 -5.36560476e-01 1.29305696e+00 6.29703820e-01
-9.22009110e-01 2.04754338e-01 -3.72027159e-01 2.15580076e-01
4.92925972e-01 8.61198425e-01 -1.35355175e+00 1.05903888e+00
2.63853848e-01 4.93780345e-01 -4.02502924e-01 1.12542844e+00
7.51266396e-03 6.61818326e-01 1.61601957e-02 -1.44198984e-01
2.82857865e-02 -2.31774434e-01 5.63904762e-01 1.61725259e+00
4.36978996e-01 6.56672657e-01 1.20610029e-01 5.63308120e-01
-2.55329199e-02 3.63109589e-01 -4.72963631e-01 -3.62801135e-01
3.32062125e-01 1.03450608e+00 -8.17278445e-01 -8.81842434e-01
-8.64694268e-02 6.91471338e-01 -3.60218063e-02 1.99782595e-01
-7.58459449e-01 -7.72549152e-01 2.00367197e-01 1.76644340e-01
-1.43097877e-03 1.71992898e-01 -5.26580393e-01 -1.33627927e+00
-8.45363084e-03 -1.01628494e+00 6.02777600e-01 -4.79505152e-01
-9.69245493e-01 8.70311856e-01 2.54078299e-01 -1.32840586e+00
-5.25510073e-01 4.45905291e-02 -5.68387747e-01 6.97846055e-01
-9.55177128e-01 -3.21851462e-01 -2.00187683e-01 4.81627360e-02
9.17297542e-01 1.75576106e-01 9.60572720e-01 -1.20642222e-02
-7.58581758e-01 4.06526774e-01 1.05294578e-01 1.54878303e-01
7.95713127e-01 -1.08356786e+00 5.50224930e-02 5.80058992e-01
-4.75913137e-01 9.54685807e-01 6.45820498e-01 -1.12749159e+00
-8.18829060e-01 -1.09133983e+00 1.64259148e+00 -1.44638747e-01
2.02549875e-01 2.45749071e-01 -9.99027491e-01 4.93328393e-01
4.89616036e-01 -1.02394104e+00 1.14100599e+00 -3.05185497e-01
4.67307419e-01 3.40979062e-02 -1.14471781e+00 4.94053632e-01
7.06223667e-01 1.77461635e-02 -1.26585209e+00 4.94183034e-01
8.92390609e-01 -3.45007986e-01 -1.17284155e+00 4.43196714e-01
3.02608877e-01 -5.97221911e-01 5.18305004e-01 -6.20927691e-01
8.69857252e-01 -7.55237415e-02 2.73412317e-01 -1.34865165e+00
-3.54661703e-01 -4.70976204e-01 2.32966587e-01 1.25159550e+00
6.62170589e-01 -6.97209299e-01 4.52352315e-01 6.47963643e-01
-4.80178952e-01 -9.60651636e-01 -5.29987395e-01 -2.42522508e-01
-1.77338362e-01 3.15712720e-01 4.41523552e-01 7.19244957e-01
7.58049965e-01 5.93229175e-01 1.39200017e-01 -1.05610631e-01
2.40600526e-01 2.86730677e-01 4.09481436e-01 -1.11892140e+00
2.90771406e-02 -3.98178846e-01 4.52728532e-02 -8.06222320e-01
-4.32225689e-02 -8.16118181e-01 1.20672040e-01 -2.32557487e+00
6.57453954e-01 3.16596031e-02 1.05869412e-01 4.02688384e-01
-5.56654692e-01 -4.15022403e-01 -2.60443296e-02 4.53564823e-01
-5.11114895e-01 3.55061620e-01 1.25707293e+00 -2.25101247e-01
-6.76009476e-01 1.62374288e-01 -1.19097257e+00 6.50122344e-01
1.07586610e+00 -6.18842185e-01 -3.05182248e-01 -1.00251250e-01
-3.56322117e-02 5.02196729e-01 -4.79263484e-01 -5.47682822e-01
2.60273188e-01 -2.54360318e-01 1.68996677e-01 -1.02654803e+00
-2.02149689e-01 -3.36911291e-01 2.83562005e-01 9.57885563e-01
-5.04022777e-01 1.51488289e-01 2.29851082e-01 3.17255467e-01
-3.39901656e-01 -6.44233286e-01 3.81958127e-01 -2.79232234e-01
7.66482949e-02 -3.68299425e-01 -9.34134364e-01 1.11248717e-01
9.19716597e-01 -3.41591090e-01 -4.31576401e-01 -2.38950744e-01
-7.44392633e-01 2.44632781e-01 1.88976765e-01 -6.94373697e-02
8.11834931e-01 -9.37161684e-01 -1.12408447e+00 -1.74831584e-01
-1.43826529e-02 2.28041247e-01 2.55059391e-01 1.19904912e+00
-9.42693055e-01 9.18929160e-01 7.82207493e-03 -4.62165266e-01
-1.48924506e+00 4.71926123e-01 -1.48157567e-01 -7.80412495e-01
-7.27204800e-01 9.86380950e-02 1.87911987e-01 1.18120544e-01
5.59098199e-02 -9.70680058e-01 -5.74247181e-01 4.30634826e-01
6.50480628e-01 3.47879678e-01 2.05891922e-01 -3.24787080e-01
-3.72502327e-01 2.83094645e-01 -3.67979467e-01 2.79965457e-02
1.21623290e+00 -8.49558040e-03 -4.26611811e-01 2.55361140e-01
9.56032217e-01 3.50775540e-01 -2.50815123e-01 1.75602943e-01
2.62590647e-01 7.89677501e-02 -5.60046911e-01 -6.06272757e-01
-4.62853491e-01 2.21079811e-01 -3.15557361e-01 4.87535506e-01
1.30804455e+00 1.09575894e-02 9.76705074e-01 3.13340843e-01
2.03986969e-02 -9.07971442e-01 -2.65452385e-01 3.00141573e-01
9.79273498e-01 -8.40900481e-01 2.81908274e-01 -6.68379664e-01
-8.76545012e-01 1.11746430e+00 2.34265208e-01 2.35149205e-01
1.63253978e-01 7.45170489e-02 -1.66448876e-01 -2.36599773e-01
-8.85526836e-01 1.69127241e-01 2.82194406e-01 1.97628170e-01
7.49282122e-01 2.02270567e-01 -1.35592866e+00 1.16776705e+00
-4.80875254e-01 2.52831392e-02 8.72530341e-01 9.97120976e-01
-6.50478601e-01 -8.72053504e-01 -5.45387745e-01 9.57260668e-01
-7.69328177e-01 -2.85721004e-01 -6.51088536e-01 6.27902567e-01
-4.65847328e-02 1.32135940e+00 -8.86258259e-02 -1.67995393e-01
6.03524208e-01 1.41759977e-01 1.87320337e-01 -1.15059233e+00
-1.00179636e+00 3.31251025e-01 5.83281040e-01 -1.35813072e-01
-5.07308185e-01 -7.60336936e-01 -1.50144136e+00 -1.31474763e-01
-3.57063204e-01 9.40925241e-01 3.18234771e-01 7.89347529e-01
5.19347191e-01 8.51014137e-01 3.70617211e-01 -3.05500805e-01
-5.75032294e-01 -1.24076355e+00 -3.26646030e-01 2.14335963e-01
1.94162056e-01 1.65052146e-01 -3.39876205e-01 3.45808417e-01] | [12.262003898620605, 9.507113456726074] |
5a545b9f-3a77-4b27-acb2-6e3bd5975a18 | mvm3det-a-novel-method-for-multi-view | 2109.10473 | null | https://arxiv.org/abs/2109.10473v1 | https://arxiv.org/pdf/2109.10473v1.pdf | MVM3Det: A Novel Method for Multi-view Monocular 3D Detection | Monocular 3D object detection encounters occlusion problems in many application scenarios, such as traffic monitoring, pedestrian monitoring, etc., which leads to serious false negative. Multi-view object detection effectively solves this problem by combining data from different perspectives. However, due to label confusion and feature confusion, the orientation estimation of multi-view 3D object detection is intractable, which is important for object tracking and intention prediction. In this paper, we propose a novel multi-view 3D object detection method named MVM3Det which simultaneously estimates the 3D position and orientation of the object according to the multi-view monocular information. The method consists of two parts: 1) Position proposal network, which integrates the features from different perspectives into consistent global features through feature orthogonal transformation to estimate the position. 2) Multi-branch orientation estimation network, which introduces feature perspective pooling to overcome the two confusion problems during the orientation estimation. In addition, we present a first dataset for multi-view 3D object detection named MVM3D. Comparing with State-Of-The-Art (SOTA) methods on our dataset and public dataset WildTrack, our method achieves very competitive results. | ['Zhao Dongbin', 'Li Jiaqi', 'Chen Yaran', 'Ma Mingjun', 'Duan Zicheng', 'Li Haoran'] | 2021-09-22 | null | null | null | null | ['multiview-detection'] | ['computer-vision'] | [-2.49704391e-01 -5.89756429e-01 -2.41593406e-01 -1.98260903e-01
-6.06879354e-01 -4.56477553e-01 3.54768127e-01 -2.66436547e-01
-3.18997532e-01 2.61446267e-01 7.14982525e-02 2.52184384e-02
3.75618428e-01 -6.16598368e-01 -4.48104024e-01 -7.61266768e-01
6.35959029e-01 4.31515783e-01 9.55174804e-01 5.18274494e-02
3.16791058e-01 6.65949166e-01 -1.61676490e+00 3.26616377e-01
2.81264216e-01 1.09866929e+00 3.80229473e-01 4.77710694e-01
1.06695004e-01 3.12771916e-01 -5.02705038e-01 -2.40453362e-01
3.63428444e-01 6.35391474e-02 -1.68128893e-01 5.26451349e-01
7.09975481e-01 -8.37919474e-01 -1.58636510e-01 1.17878973e+00
7.70806491e-01 -2.38410220e-01 5.98598063e-01 -1.50915945e+00
-3.11999381e-01 -8.55767950e-02 -1.18466616e+00 3.81224245e-01
3.70821327e-01 2.68137664e-01 6.66639090e-01 -1.24766803e+00
5.30164301e-01 1.51859319e+00 6.26463175e-01 3.87520313e-01
-8.08418751e-01 -7.47062564e-01 3.30494076e-01 4.96018499e-01
-1.48939741e+00 -3.46471995e-01 8.34682167e-01 -4.93283361e-01
6.37693644e-01 2.84583747e-01 9.30739045e-01 7.46331990e-01
2.39902556e-01 1.24092090e+00 1.15203571e+00 -5.43668121e-03
-3.27864498e-01 2.96694577e-01 3.11671257e-01 5.98689377e-01
6.77783966e-01 4.07115966e-01 -3.33180994e-01 -1.41766578e-01
5.32079458e-01 3.63177925e-01 -6.42240345e-02 -7.48589575e-01
-1.33151889e+00 5.40075243e-01 4.36156511e-01 -9.36259627e-02
-2.77687103e-01 -2.05765173e-01 2.68925756e-01 -2.20730245e-01
4.49690431e-01 -2.87042886e-01 -4.19946790e-01 2.43784904e-01
-3.43750566e-01 2.06164688e-01 1.45764723e-01 1.26513875e+00
5.49353838e-01 -1.78217351e-01 -1.96525931e-01 6.56885326e-01
8.02823246e-01 8.27303469e-01 6.43790364e-02 -6.16520882e-01
7.19011188e-01 8.74543071e-01 2.82167852e-01 -1.10063469e+00
-7.79349267e-01 -7.37300813e-01 -7.28349626e-01 2.70166576e-01
4.79614019e-01 5.75556047e-02 -6.76874459e-01 1.31588578e+00
9.99134421e-01 -2.83828288e-01 -2.17247397e-01 1.33678508e+00
1.42048192e+00 4.33239460e-01 -2.55610019e-01 -1.64568931e-01
1.81418502e+00 -8.93786728e-01 -4.51765656e-01 -3.79752696e-01
4.28150862e-01 -1.07924020e+00 4.09469336e-01 4.94824380e-01
-9.80669856e-01 -8.07752311e-01 -1.03343856e+00 1.64542608e-02
-2.15404421e-01 6.25530720e-01 5.42107105e-01 8.24268460e-01
-4.50158864e-01 -4.51697797e-01 -5.51680088e-01 -4.20321703e-01
5.56562960e-01 1.64573461e-01 -4.26913440e-01 -4.41290647e-01
-7.52448738e-01 9.68166947e-01 5.61324358e-01 3.14302981e-01
-6.69032753e-01 -3.76152724e-01 -7.93163836e-01 -2.39132389e-01
7.60729849e-01 -9.68719244e-01 9.23863292e-01 -6.59786761e-02
-9.39387560e-01 1.11500120e+00 -3.77601653e-01 1.07070506e-02
6.56561375e-01 -1.07948065e-01 -3.81667763e-01 -3.27911079e-02
4.65564311e-01 7.28693604e-01 7.51332104e-01 -1.36612415e+00
-1.19611883e+00 -1.00291240e+00 2.46435739e-02 4.57063645e-01
1.11301340e-01 1.89611256e-01 -7.76165485e-01 -1.51320830e-01
7.99340725e-01 -8.41186166e-01 -1.27341241e-01 2.90720791e-01
-5.69004595e-01 -4.85943407e-01 1.08769143e+00 -3.83892953e-01
7.20035613e-01 -2.16127396e+00 -9.81747918e-03 -2.96857476e-01
5.38248479e-01 1.09062776e-01 2.18806580e-01 -7.07497224e-02
2.72553921e-01 -4.13642138e-01 4.70526487e-01 -2.66238093e-01
-3.27432185e-01 -1.84717819e-01 7.58159384e-02 8.97598147e-01
-2.10327487e-02 8.17382753e-01 -9.02081728e-01 -9.22908723e-01
5.56283355e-01 2.96392053e-01 -5.14820397e-01 1.02897465e-01
1.09517865e-01 4.06269968e-01 -7.20290363e-01 1.07230759e+00
1.24517405e+00 -2.58328646e-01 -2.81032890e-01 -7.91375160e-01
-3.24026495e-01 -1.37385711e-01 -1.46328938e+00 1.31053913e+00
-1.37668014e-01 3.88919801e-01 -6.15355000e-02 -6.84913993e-01
9.98522937e-01 1.17426246e-01 4.96850044e-01 -3.29540312e-01
2.92496294e-01 6.99954852e-02 -8.81864205e-02 -5.84843636e-01
6.74526155e-01 -3.40741388e-02 -2.80909017e-02 2.67939150e-01
-1.48442283e-01 -2.71812873e-03 1.97969694e-02 8.30316022e-02
3.41906577e-01 2.25271776e-01 7.44887829e-01 2.10106984e-01
8.32300305e-01 -2.96277087e-02 7.77746797e-01 6.84659660e-01
-6.19598567e-01 5.67360401e-01 1.94437712e-01 -6.17867887e-01
-5.90864599e-01 -1.11916351e+00 -1.66339383e-01 5.33408284e-01
7.53776610e-01 -1.98140785e-01 -1.42644525e-01 -9.39316273e-01
1.88343942e-01 3.85967195e-01 -3.18832666e-01 -2.96154246e-02
-6.00513697e-01 -9.92267311e-01 7.15876445e-02 5.52187383e-01
8.55442762e-01 -4.25613731e-01 -8.24011445e-01 -1.02867307e-02
-4.51253653e-01 -1.44478536e+00 -7.49040067e-01 -2.38629326e-01
-8.44643891e-01 -1.27079701e+00 -6.59919620e-01 -5.20663679e-01
5.62946558e-01 1.33435416e+00 9.00853932e-01 -5.39636649e-02
-4.22229886e-01 3.64760160e-01 -1.04533546e-01 -5.62737942e-01
-9.24363639e-03 -2.47371837e-01 1.83937848e-01 1.85674638e-01
8.25283885e-01 -1.62242100e-01 -8.33259821e-01 8.66577983e-01
-3.33698452e-01 2.13107243e-01 7.30434775e-01 5.92476785e-01
6.47582114e-01 -8.05203989e-02 2.48680100e-01 -3.93143028e-01
-1.82805642e-01 -2.23284408e-01 -9.48460042e-01 1.09991111e-01
-1.63935781e-01 -6.54353380e-01 -2.94865724e-02 -4.38885063e-01
-9.92449641e-01 4.01085645e-01 4.72143777e-02 -4.37512577e-01
-4.11372006e-01 -1.92637429e-01 -6.41905665e-01 -1.92278206e-01
4.32192266e-01 4.64733779e-01 -1.59400403e-02 -5.44172347e-01
2.43079171e-01 9.60951388e-01 2.50558913e-01 1.16599314e-01
7.60515571e-01 7.89906681e-01 2.07230255e-01 -8.46834600e-01
-1.22713757e+00 -9.46625531e-01 -7.09613979e-01 -6.74994469e-01
9.72425878e-01 -1.21852541e+00 -9.71495032e-01 6.32341087e-01
-1.49385738e+00 7.46443212e-01 1.92100048e-01 8.18284571e-01
-3.51259112e-01 5.92574596e-01 -2.89283514e-01 -9.97010767e-01
-1.88624725e-01 -1.30128264e+00 1.45920968e+00 1.79425210e-01
4.25391287e-01 -3.82450730e-01 -5.23305833e-01 7.83890307e-01
5.65285981e-02 9.14900843e-03 3.38678986e-01 -3.19267124e-01
-1.08612263e+00 -4.36537534e-01 -6.53187573e-01 8.05819407e-02
8.77863276e-05 -3.39830369e-01 -9.42005992e-01 -2.02434912e-01
3.32009345e-01 8.24757293e-02 7.78853655e-01 7.15095341e-01
7.13995695e-01 2.57338643e-01 -8.26878667e-01 5.25040209e-01
1.01366901e+00 2.33180955e-01 2.60420710e-01 3.13876808e-01
8.75354886e-01 4.65067327e-01 1.19277835e+00 4.31667954e-01
5.24256647e-01 1.18062055e+00 6.33597255e-01 1.14064939e-01
-2.90185988e-01 -1.34456232e-01 2.53884822e-01 4.92667675e-01
2.95424797e-02 -1.82734653e-01 -6.00519240e-01 2.08327159e-01
-1.71858668e+00 -1.04717124e+00 -6.06384277e-01 2.03655815e+00
-3.92805785e-03 4.26239461e-01 5.12774646e-01 -1.19231246e-01
1.04440129e+00 1.43313065e-01 -5.80295980e-01 5.29424787e-01
-8.61879960e-02 -9.49613094e-01 5.47002017e-01 1.29564047e-01
-1.34901965e+00 5.59338510e-01 5.43906784e+00 8.99357319e-01
-8.60211968e-01 3.74170274e-01 3.45038742e-01 -2.91805025e-02
1.91968217e-01 -1.62095577e-01 -1.75643635e+00 2.84235805e-01
-2.55253822e-01 3.78733903e-01 -1.84018821e-01 9.25816298e-01
2.19731666e-02 -2.93885589e-01 -8.49866807e-01 1.52662134e+00
5.49990118e-01 -1.03246975e+00 -3.20730507e-02 2.60750979e-01
4.03910905e-01 3.77939343e-02 -9.36568826e-02 1.34084985e-01
-2.45587319e-01 -1.57289043e-01 8.86662245e-01 2.32163191e-01
4.32248116e-01 -5.83501101e-01 6.89172804e-01 7.61146784e-01
-1.58184278e+00 -1.13351740e-01 -4.58011806e-01 7.13698789e-02
5.30327439e-01 7.61418998e-01 -8.32133055e-01 6.93790078e-01
5.84463000e-01 1.09281027e+00 -6.57716632e-01 1.37250686e+00
-5.80892898e-02 -4.94186953e-02 -3.59662235e-01 -1.01771720e-01
-8.19446892e-02 -5.89773357e-02 1.03219020e+00 6.65916860e-01
1.87724531e-01 8.33323896e-02 4.93795693e-01 8.93007815e-01
3.69752944e-01 -1.18449740e-01 -6.97803199e-01 7.04477310e-01
3.55410367e-01 1.34465146e+00 -7.74960935e-01 -3.92494708e-01
-6.93110049e-01 6.65367186e-01 -7.09476098e-02 1.15689397e-01
-8.51151466e-01 -9.64291859e-03 3.00525427e-01 8.46836641e-02
5.40584683e-01 -2.70350665e-01 -1.03256851e-01 -1.51019955e+00
2.68886715e-01 -5.06576419e-01 3.98609519e-01 -9.31832373e-01
-1.23234940e+00 4.15397197e-01 2.46785969e-01 -1.95838034e+00
1.12769350e-01 -7.80551016e-01 -2.36575320e-01 6.41474307e-01
-1.61785758e+00 -1.44305003e+00 -4.76685047e-01 2.37361684e-01
6.33299410e-01 -1.82710707e-01 6.86916858e-02 6.14528120e-01
-4.78934407e-01 3.50736231e-01 -4.46531981e-01 -5.48898578e-02
6.75066352e-01 -7.62553990e-01 3.75827253e-01 6.76651955e-01
-1.18121818e-01 1.68318629e-01 3.71161997e-01 -7.29939163e-01
-1.59424675e+00 -1.13198531e+00 8.09789598e-01 -8.10871363e-01
1.71574935e-01 -3.45960349e-01 -4.21943814e-01 5.37813604e-01
-3.84272188e-01 4.18059200e-01 4.34886694e-01 -9.49753672e-02
-1.95075184e-01 -1.74625859e-01 -1.11544168e+00 3.54962021e-01
1.33708155e+00 -8.72682594e-03 -4.88983333e-01 3.83010864e-01
5.42683542e-01 -8.05962443e-01 -6.13488078e-01 6.32413566e-01
8.65659714e-01 -1.21135128e+00 1.45264256e+00 -3.98597181e-01
-1.75086230e-01 -7.73618877e-01 -4.14251179e-01 -7.36086905e-01
-3.40410501e-01 -4.37397026e-02 -3.14906001e-01 1.10196793e+00
-1.19013526e-01 -8.35992634e-01 7.40312636e-01 -1.11639872e-01
-1.02208950e-01 -6.32029116e-01 -1.18342125e+00 -7.69346356e-01
-5.82449436e-01 -4.85817760e-01 5.68194032e-01 3.47601533e-01
-6.00011408e-01 7.13524342e-01 -4.62837636e-01 5.55938900e-01
1.03235292e+00 7.34929323e-01 1.24308395e+00 -1.44923234e+00
-2.53269941e-01 -2.83116907e-01 -8.06376398e-01 -1.82536602e+00
-3.58037949e-01 -6.35251224e-01 -7.06593469e-02 -1.32113254e+00
8.09946954e-01 -3.17075610e-01 7.29369372e-02 -1.67825624e-01
-2.79149562e-01 5.29717326e-01 4.03851658e-01 1.99347615e-01
-8.38066459e-01 5.34410655e-01 1.62188947e+00 -8.22510868e-02
-1.57022048e-02 5.01407802e-01 -5.95267594e-01 9.41103935e-01
4.82125700e-01 -4.77484763e-01 7.38252550e-02 -2.82214910e-01
2.36590989e-02 3.98092926e-01 8.87683213e-01 -8.58266771e-01
2.11449742e-01 -1.63059682e-01 6.86728418e-01 -1.78145921e+00
7.60380924e-01 -9.94108200e-01 -1.36832058e-01 7.37262607e-01
3.22709352e-01 -7.05814138e-02 4.13499773e-02 8.05759430e-01
2.23932266e-02 -8.89386013e-02 7.62904823e-01 -2.92874306e-01
-8.67694080e-01 4.51226056e-01 -2.02920884e-01 -1.49235502e-01
1.16884315e+00 -4.31392610e-01 -5.27788281e-01 -8.84660557e-02
-6.14991903e-01 4.81556565e-01 2.09445462e-01 6.41931951e-01
8.00566316e-01 -1.63719726e+00 -8.94071937e-01 4.89789367e-01
4.41129029e-01 7.01758489e-02 3.90398234e-01 1.16168427e+00
-4.06129770e-02 5.36796570e-01 -4.68162959e-03 -1.33263791e+00
-1.69499075e+00 7.45320380e-01 3.32753181e-01 -6.36701146e-03
-5.71149468e-01 4.93833274e-01 6.29796863e-01 -4.31372076e-01
4.85976003e-02 -2.12097928e-01 -4.05031353e-01 1.43365473e-01
6.44222736e-01 5.46742857e-01 -1.56809494e-01 -1.04998696e+00
-5.34695268e-01 1.21215153e+00 -2.18183592e-01 2.04705805e-01
9.35288966e-01 -5.73936582e-01 7.47723505e-02 3.54246169e-01
9.89009798e-01 -1.25594735e-01 -1.12760103e+00 -4.26699549e-01
-5.52948356e-01 -8.90976012e-01 8.37795734e-02 -3.04273129e-01
-1.09935904e+00 1.09996569e+00 9.12330210e-01 -4.81137931e-02
7.67996252e-01 1.82372630e-01 7.49385834e-01 3.32517058e-01
6.28005385e-01 -6.65874720e-01 2.38189563e-01 3.84271950e-01
5.74185014e-01 -1.65569520e+00 4.22002912e-01 -8.73511732e-01
-4.23006743e-01 1.05297482e+00 1.00529063e+00 2.28943929e-01
7.31457651e-01 -2.08889186e-01 -8.90246034e-02 -2.67777473e-01
-3.33498418e-01 -3.84263009e-01 4.70710993e-01 7.80063927e-01
-1.45019174e-01 2.44425274e-02 -1.34166569e-01 4.51383233e-01
3.88960838e-01 -1.96834490e-01 2.33220264e-01 7.78633356e-01
-6.80628419e-01 -7.39903331e-01 -8.96435559e-01 3.23148727e-01
-1.91788360e-01 2.97370672e-01 -1.19615935e-01 9.44355905e-01
4.30906087e-01 9.94452357e-01 1.26106590e-02 -3.54700029e-01
5.91144323e-01 -2.92535603e-01 6.90877378e-01 -4.44073379e-01
-1.35537863e-01 4.12384033e-01 -1.00571983e-01 -4.24656093e-01
-6.38796508e-01 -9.79377091e-01 -6.66604757e-01 1.28096819e-01
-9.34548438e-01 -4.05667216e-01 5.35700083e-01 8.73139739e-01
2.22146004e-01 2.71971703e-01 7.10999191e-01 -1.16995370e+00
-6.71501040e-01 -7.71741509e-01 -5.32903373e-01 1.09715402e-01
4.35885221e-01 -1.08570516e+00 -1.05605498e-01 -3.09135854e-01] | [7.918281078338623, -2.2829999923706055] |
8f852868-8733-45bc-a378-12ab9a3fbdb6 | conceptbed-evaluating-concept-learning | 2306.04695 | null | https://arxiv.org/abs/2306.04695v1 | https://arxiv.org/pdf/2306.04695v1.pdf | ConceptBed: Evaluating Concept Learning Abilities of Text-to-Image Diffusion Models | The ability to understand visual concepts and replicate and compose these concepts from images is a central goal for computer vision. Recent advances in text-to-image (T2I) models have lead to high definition and realistic image quality generation by learning from large databases of images and their descriptions. However, the evaluation of T2I models has focused on photorealism and limited qualitative measures of visual understanding. To quantify the ability of T2I models in learning and synthesizing novel visual concepts, we introduce ConceptBed, a large-scale dataset that consists of 284 unique visual concepts, 5K unique concept compositions, and 33K composite text prompts. Along with the dataset, we propose an evaluation metric, Concept Confidence Deviation (CCD), that uses the confidence of oracle concept classifiers to measure the alignment between concepts generated by T2I generators and concepts contained in ground truth images. We evaluate visual concepts that are either objects, attributes, or styles, and also evaluate four dimensions of compositionality: counting, attributes, relations, and actions. Our human study shows that CCD is highly correlated with human understanding of concepts. Our results point to a trade-off between learning the concepts and preserving the compositionality which existing approaches struggle to overcome. | ['Yezhou Yang', 'Chitta Baral', 'Tejas Gokhale', 'Maitreya Patel'] | 2023-06-07 | null | null | null | null | ['concept-alignment'] | ['computer-vision'] | [ 4.03196216e-01 1.29088134e-01 2.00610638e-01 -4.12069559e-01
-5.42843640e-01 -8.77802253e-01 1.10449040e+00 2.87188530e-01
-1.02559596e-01 3.28248799e-01 2.11645693e-01 -1.53914243e-01
-6.32572472e-02 -5.74185610e-01 -8.41152847e-01 -3.20612043e-01
2.41375819e-01 6.36230707e-01 1.08040012e-02 -8.00781325e-02
5.34380317e-01 2.24492028e-01 -1.95956874e+00 8.05528939e-01
8.28797102e-01 1.03347230e+00 1.79527432e-01 7.20738530e-01
-5.13978183e-01 9.67267692e-01 -1.00846088e+00 -6.89941585e-01
2.10609928e-01 -5.74781656e-01 -7.37301230e-01 3.64236951e-01
9.41628575e-01 -9.39494371e-02 1.49276122e-01 9.82196748e-01
1.13767006e-01 -9.64879692e-02 1.08667278e+00 -1.84372759e+00
-1.20310020e+00 5.28470218e-01 -2.37285644e-01 -4.17125262e-02
5.47081292e-01 3.50548059e-01 1.25316596e+00 -1.09693146e+00
7.46049404e-01 1.55271482e+00 3.56552720e-01 6.63128734e-01
-1.32744360e+00 -7.28435457e-01 8.18234459e-02 3.01073939e-01
-1.39222026e+00 -2.21225396e-01 5.06884634e-01 -8.18197131e-01
6.94719553e-01 2.33519912e-01 7.50453830e-01 1.27049649e+00
-1.62824154e-01 9.16675270e-01 1.37025023e+00 -6.63102508e-01
2.91661203e-01 5.92333913e-01 -1.94737881e-01 4.99771595e-01
3.77726853e-01 2.82415003e-01 -8.96138608e-01 2.70573795e-01
6.91819251e-01 -2.56560594e-01 -2.49565214e-01 -3.87845576e-01
-1.47236836e+00 7.73338020e-01 3.41952890e-01 1.06120832e-01
-8.16339850e-02 -7.44644403e-02 1.57044366e-01 1.68334261e-01
1.15579061e-01 9.83403623e-01 -2.47417226e-01 7.36955851e-02
-7.42223024e-01 4.15384620e-01 6.15843415e-01 1.54003370e+00
6.15144432e-01 -7.71525279e-02 -4.88615364e-01 6.06358409e-01
-6.63947687e-02 9.13758278e-01 7.02220738e-01 -1.07161140e+00
3.89710635e-01 9.24758792e-01 -1.35626793e-01 -1.08178449e+00
-1.39819384e-01 -1.06909327e-01 -7.81222641e-01 2.37941831e-01
3.11285645e-01 3.87531340e-01 -8.29248071e-01 1.64825130e+00
-1.07662559e-01 -3.04570585e-01 2.61475295e-01 6.24556541e-01
9.57611203e-01 4.08189058e-01 7.42348209e-02 1.59586251e-01
1.20716023e+00 -8.15416217e-01 -4.24525052e-01 -3.15599859e-01
2.93348044e-01 -7.01273203e-01 1.47726297e+00 3.59539390e-01
-9.23184752e-01 -1.03620088e+00 -1.07026017e+00 -5.86034590e-03
-7.19766080e-01 2.24124804e-01 4.59550112e-01 8.03201795e-01
-1.04074585e+00 2.16941774e-01 -1.17413424e-01 -3.88648778e-01
4.71484572e-01 -1.62552848e-01 -4.18722659e-01 -2.54316002e-01
-6.96992815e-01 8.14870596e-01 7.58963346e-01 -5.02876759e-01
-1.04499912e+00 -8.35039794e-01 -7.71754742e-01 8.00354257e-02
5.19711018e-01 -7.53086984e-01 1.10197520e+00 -1.29997241e+00
-9.56759453e-01 1.03127623e+00 1.70269310e-01 -3.71307373e-01
4.41784769e-01 1.33251399e-01 -4.07455266e-01 4.04704452e-01
3.65900397e-01 1.33674145e+00 1.03063774e+00 -1.91523266e+00
-9.64701235e-01 -2.23006248e-01 2.09991068e-01 2.68614739e-01
-4.43248689e-01 -2.77968615e-01 -3.36271793e-01 -6.44567728e-01
1.51995406e-01 -8.64306092e-01 4.12498742e-01 2.48824522e-01
-5.33852875e-01 -8.22131708e-02 5.66844761e-01 -4.67061728e-01
5.96299648e-01 -2.03871703e+00 -3.43947895e-02 1.90614089e-01
3.88994157e-01 -1.42908888e-02 -2.75669783e-01 2.02506334e-01
-1.22181542e-01 3.28765392e-01 -4.88092052e-03 -2.46374398e-01
8.33953638e-03 1.21070758e-01 -4.16971833e-01 -3.77727658e-01
4.00151938e-01 1.06960976e+00 -9.58733499e-01 -6.73823118e-01
2.56542683e-01 1.53050929e-01 -5.16562462e-01 2.13690177e-01
-6.07713223e-01 1.00124188e-01 7.28123486e-02 6.66966736e-01
3.25264573e-01 -4.50866252e-01 2.89970428e-01 -3.42422813e-01
3.47629339e-01 -2.83201396e-01 -9.50324893e-01 1.45983124e+00
-3.73029530e-01 8.78391623e-01 -7.56707132e-01 -7.73467004e-01
1.05630314e+00 1.38802484e-01 9.85532254e-02 -9.54056978e-01
1.60431415e-02 -8.75997022e-02 -1.62482619e-01 -4.93567050e-01
6.94808185e-01 -1.75087810e-01 5.87543212e-02 4.89958972e-01
2.90873021e-01 -6.09647572e-01 4.20157880e-01 5.85261881e-01
6.87726736e-01 1.10681057e-01 3.14535975e-01 -1.58876643e-01
2.59669691e-01 2.42255241e-01 -2.16210224e-02 1.02673566e+00
2.26348490e-02 8.92061651e-01 5.80891013e-01 -2.76182503e-01
-1.22550893e+00 -1.31493986e+00 2.85605133e-01 9.91039217e-01
2.19310194e-01 -4.72418487e-01 -6.70422196e-01 -5.59590280e-01
-3.64761539e-02 9.65546131e-01 -8.69243205e-01 -2.34762534e-01
4.54842821e-02 -3.17406654e-01 5.06951034e-01 7.75060475e-01
4.65215474e-01 -1.08758998e+00 -7.68855214e-01 -2.06267059e-01
-6.54842854e-01 -1.38433862e+00 -3.76204401e-01 -2.50244558e-01
-7.06678867e-01 -1.31785858e+00 -4.52789724e-01 -6.92651629e-01
8.84131491e-01 5.90115249e-01 1.66623986e+00 1.53904453e-01
-5.56105793e-01 7.73432672e-01 -6.65903986e-01 -8.02589118e-01
-8.30763936e-01 -4.44953948e-01 -1.63896218e-01 2.97475047e-02
3.84067655e-01 -1.32824987e-01 -3.43015373e-01 2.75750488e-01
-1.17290092e+00 5.55592656e-01 6.93042397e-01 7.70223856e-01
4.21442598e-01 1.81852803e-01 1.33250713e-01 -6.64273739e-01
6.62359416e-01 -7.03604817e-02 -4.47115511e-01 8.09590757e-01
-7.28212953e-01 2.33338803e-01 2.45092705e-01 -5.27948320e-01
-1.01976907e+00 3.09691560e-02 7.03792512e-01 -5.89786768e-01
-1.98214814e-01 2.14955509e-01 -1.71241522e-01 2.33212814e-01
9.67625082e-01 4.35939997e-01 -2.31962427e-02 -3.09038744e-03
1.05511069e+00 4.14640665e-01 7.81672478e-01 -8.87422979e-01
7.73649037e-01 5.47635019e-01 -2.53278881e-01 -9.04040515e-01
-9.02889371e-01 -4.08221573e-01 -7.06294417e-01 -3.62601221e-01
9.35941160e-01 -9.22352016e-01 -7.94136584e-01 3.31434667e-01
-1.20575869e+00 -5.07939793e-02 -4.58396703e-01 1.08533472e-01
-6.73003972e-01 2.90835768e-01 -1.29527412e-02 -7.29918718e-01
-2.08988726e-01 -8.59459519e-01 1.09245789e+00 1.56576350e-01
-2.39982039e-01 -6.64639413e-01 -4.11133051e-01 4.95855629e-01
2.01723084e-01 2.41193131e-01 1.35548496e+00 -4.07087266e-01
-7.54406810e-01 -3.94753516e-02 -6.56166017e-01 6.23578668e-01
-3.48822284e-03 1.28589019e-01 -1.06289184e+00 -1.47086740e-01
-1.59626916e-01 -7.76847243e-01 6.39992058e-01 8.81353021e-02
1.18408370e+00 -4.70066249e-01 -8.34452137e-02 2.35278755e-01
1.64304578e+00 4.16151077e-01 7.05626905e-01 6.95925578e-02
7.61273682e-01 9.84245121e-01 5.29615939e-01 3.00292492e-01
5.49806178e-01 5.53859830e-01 1.87220931e-01 -4.10938039e-02
-3.22702855e-01 -6.46022499e-01 9.57040861e-02 5.07947028e-01
-2.77066119e-02 -3.61304790e-01 -1.11436033e+00 6.18757784e-01
-1.48614848e+00 -1.09354234e+00 1.47485062e-01 2.03263354e+00
9.51077998e-01 3.45872730e-01 9.08216685e-02 2.76366025e-01
5.75448394e-01 -5.52428842e-01 -3.06997150e-01 -3.30256112e-02
-4.30758297e-01 1.27915099e-01 1.81212217e-01 2.61323065e-01
-7.41513312e-01 9.85037923e-01 6.75709152e+00 7.56445229e-01
-7.50439346e-01 -1.27952218e-01 6.43800914e-01 -3.46189588e-02
-3.95920932e-01 7.29934648e-02 -6.46374583e-01 2.15715364e-01
4.74379301e-01 -3.93154830e-01 3.24498355e-01 8.83569777e-01
-2.70425439e-01 -7.56094605e-02 -1.44866765e+00 1.26043093e+00
7.27599740e-01 -1.38144100e+00 9.48206127e-01 -7.91184306e-02
9.55689311e-01 -5.87682486e-01 2.34262690e-01 2.37677649e-01
4.59151536e-01 -1.26334369e+00 1.20764506e+00 6.80202544e-01
1.10809517e+00 -4.25596058e-01 3.61337751e-01 5.25150336e-02
-9.75625515e-01 -2.45535105e-01 -3.09354961e-01 4.82468158e-02
-3.69161040e-01 1.36869848e-01 -1.20391095e+00 3.86662066e-01
8.43808711e-01 5.50277412e-01 -1.29754364e+00 6.50078893e-01
-2.73056626e-01 1.88591182e-01 1.38711557e-01 -1.66987345e-01
2.36837170e-03 6.59755915e-02 1.02973310e-02 1.01208341e+00
1.24307096e-01 8.09066370e-03 4.70784754e-02 1.31872189e+00
2.01968476e-02 4.49754521e-02 -5.16937137e-01 -3.83360535e-01
5.91001332e-01 1.01992166e+00 -8.37096572e-01 -7.50691891e-01
-3.78211141e-01 9.18169618e-01 3.75829637e-02 2.93232977e-01
-4.67077971e-01 -2.77984858e-01 4.55243170e-01 1.07968494e-01
1.69837385e-01 -1.48882568e-01 -5.58843195e-01 -1.05209148e+00
1.30647078e-01 -1.34187365e+00 2.98045337e-01 -1.43096340e+00
-1.40553606e+00 6.24581814e-01 3.97424102e-01 -1.30975831e+00
-2.45283991e-01 -1.00137579e+00 -2.00852692e-01 5.12220979e-01
-1.08337104e+00 -1.43596661e+00 -9.05731499e-01 6.61129892e-01
6.52392268e-01 -4.23098654e-01 7.63306499e-01 -1.97123572e-01
1.45412564e-01 6.09245181e-01 -2.20043123e-01 2.78904825e-01
5.88783741e-01 -1.48243535e+00 6.96828187e-01 6.69794142e-01
6.90996885e-01 6.29524827e-01 5.72440505e-01 -5.18809259e-01
-8.14318895e-01 -9.05032635e-01 8.09976220e-01 -9.50975955e-01
3.43052059e-01 -5.37985563e-01 -7.34362602e-01 4.57751513e-01
2.18825489e-02 -4.67285603e-01 6.38072193e-01 -1.52852491e-01
-1.05293894e+00 1.10636214e-02 -8.55069220e-01 7.32117534e-01
1.08714795e+00 -7.18110085e-01 -7.05665410e-01 3.66856009e-01
8.91784668e-01 -3.85255665e-02 -6.27814174e-01 1.97898984e-01
6.13003016e-01 -1.13487399e+00 1.16109526e+00 -6.42138958e-01
8.51994157e-01 -2.59114355e-01 -3.76701325e-01 -1.20692801e+00
-1.61650106e-01 3.32823843e-02 3.26203555e-01 1.13140798e+00
5.14689624e-01 -2.20947862e-02 6.17033660e-01 6.28560245e-01
1.67957529e-01 -1.11896351e-01 -4.34604138e-01 -1.02240944e+00
7.18497559e-02 -4.73039329e-01 7.21507967e-01 9.19712722e-01
-1.54506594e-01 6.55813336e-01 -1.88141942e-01 -1.56368554e-01
7.05287695e-01 1.60268471e-01 9.34895813e-01 -1.28763461e+00
-2.78547078e-01 -6.58462465e-01 -6.83898330e-01 -6.55519128e-01
-1.19538240e-01 -8.09491873e-01 3.47364135e-02 -1.51259458e+00
6.06337547e-01 -2.37979576e-01 1.12242430e-01 5.56132376e-01
-2.20742106e-01 1.83748275e-01 6.01236463e-01 3.88802439e-01
-5.15369952e-01 5.19555628e-01 1.29749370e+00 -5.76965332e-01
7.98205510e-02 -4.99241054e-01 -9.66442585e-01 5.41534007e-01
6.43316805e-01 -1.88484222e-01 -7.96850622e-01 -4.22163337e-01
2.76722580e-01 -1.64390102e-01 7.70858347e-01 -1.38473141e+00
3.56422327e-02 -2.56554872e-01 6.90614879e-01 -5.84043860e-01
3.66385132e-01 -6.83741987e-01 -7.67823011e-02 2.02848881e-01
-6.74557745e-01 3.01561713e-01 1.13633625e-01 7.01453447e-01
-2.88975984e-01 -1.09886274e-01 6.52252376e-01 -4.83569771e-01
-1.06415606e+00 -1.01604037e-01 -4.78571616e-02 5.42728543e-01
8.35858643e-01 -4.43793178e-01 -6.18834317e-01 -4.97156113e-01
-4.29256320e-01 -3.07277441e-02 5.61349154e-01 9.52623248e-01
9.91767347e-01 -1.35744393e+00 -8.70842993e-01 1.64780423e-01
1.00865388e+00 -3.56592506e-01 4.54557501e-02 -5.40225096e-02
-4.72480834e-01 4.00016904e-01 -5.20445943e-01 -8.68285179e-01
-1.25650382e+00 8.26693475e-01 2.58236676e-01 3.65888417e-01
-5.08807063e-01 8.18424821e-01 5.51096916e-01 -1.87965542e-01
2.17108294e-01 -4.06648010e-01 -2.04278722e-01 -1.63731519e-02
5.48432171e-01 8.39874595e-02 -1.19325958e-01 -5.37476242e-01
-4.18093987e-02 4.33508545e-01 -5.92074357e-03 -4.61485773e-01
6.47862792e-01 -5.42366086e-03 7.86630586e-02 3.48176837e-01
9.48760748e-01 -4.94692504e-01 -1.11553693e+00 -3.70040894e-01
2.39214331e-01 -5.86288512e-01 -4.74328816e-01 -1.13056219e+00
-7.57919014e-01 9.58189189e-01 6.98335052e-01 -2.05300469e-02
1.05047774e+00 1.20274834e-01 2.19700158e-01 5.56026220e-01
2.95153797e-01 -1.13356853e+00 1.08054650e+00 2.77644694e-01
1.09104538e+00 -1.47120094e+00 -2.04426363e-01 -4.53902394e-01
-1.20963466e+00 9.92076993e-01 9.46399748e-01 4.36344981e-01
2.71970868e-01 -1.84797123e-01 2.86379457e-01 -3.56501371e-01
-7.22030461e-01 -3.64113063e-01 5.59479892e-01 1.06070745e+00
9.66656208e-02 2.84464866e-01 2.62509167e-01 4.12500650e-01
-5.61832368e-01 -1.54766276e-01 7.03154743e-01 6.64981723e-01
-3.48352611e-01 -6.55455589e-01 -3.68897289e-01 4.42021072e-01
3.12803894e-01 -2.62546211e-01 -7.54976213e-01 8.85114133e-01
3.29229206e-01 1.12418962e+00 2.04366267e-01 -5.64651847e-01
3.48627925e-01 1.38457000e-01 6.11771822e-01 -7.01190233e-01
-8.70046392e-02 -6.21064961e-01 -1.92635894e-01 -3.11078191e-01
-5.33792973e-01 -4.14885253e-01 -9.04217184e-01 -4.72057238e-02
-1.56332672e-01 -1.06427021e-01 8.34158063e-01 8.97456527e-01
2.02469021e-01 3.70922416e-01 2.65801132e-01 -5.07594168e-01
-3.81614506e-01 -8.39072347e-01 -4.08209264e-01 1.18202555e+00
-8.72499570e-02 -7.10267365e-01 -3.98077309e-01 5.86163640e-01] | [10.809503555297852, 1.4855324029922485] |
8da25666-7c39-4637-8bda-08c97ad561eb | empirical-effect-of-graph-embeddings-on-fraud | 1903.05976 | null | http://arxiv.org/abs/1903.05976v1 | http://arxiv.org/pdf/1903.05976v1.pdf | Empirical effect of graph embeddings on fraud detection/ risk mitigation | Graph embedding technics are studied with interest on public datasets, such
as BlogCatalog, with the common practice of maximizing scoring on graph
reconstruction, link prediction metrics etc. However, in the financial sector
the important metrics are often more business related, for example fraud
detection rates. With our privileged position of having large amount of
real-world non-public P2P-lending social data, we aim to study empirically
whether recent advances in graph embedding technics provide a useful signal for
metrics more closely related to business interests, such as fraud detection
rate. | ['Sida Zhou'] | 2019-03-05 | null | null | null | null | ['graph-reconstruction'] | ['graphs'] | [-6.95075810e-01 6.55857563e-01 -3.23509902e-01 -9.32684392e-02
-1.45198554e-01 -2.79295027e-01 5.73579848e-01 9.85756099e-01
-4.18877631e-01 7.04831064e-01 4.83842254e-01 -3.92090112e-01
-3.17073464e-01 -1.20066953e+00 -2.93161541e-01 -3.53676051e-01
-6.11876607e-01 3.42544377e-01 4.02806669e-01 -5.06567955e-01
5.10549724e-01 3.31186622e-01 -7.40718007e-01 -9.41623673e-02
4.78318423e-01 7.04233468e-01 -5.66276610e-01 1.91173151e-01
-2.60783676e-02 6.76930308e-01 -3.18770826e-01 -1.31500840e+00
3.31129760e-01 -2.14074835e-01 -6.67622924e-01 -4.95414585e-01
-6.03826940e-02 2.05603391e-02 -6.59466088e-01 1.39902484e+00
3.07746321e-01 -1.24312580e-01 7.02335477e-01 -1.54885590e+00
-3.92561615e-01 8.28239858e-01 -7.64712214e-01 6.22038960e-01
3.60955089e-01 -3.05324774e-02 1.81975484e+00 -4.49779481e-01
6.90050662e-01 1.17658472e+00 8.50301683e-01 -1.98230311e-01
-1.23837888e+00 -5.23296773e-01 -2.57206261e-01 4.09670681e-01
-8.32694709e-01 7.25781918e-02 9.14337218e-01 -4.86386389e-01
5.33272147e-01 2.64544606e-01 7.39360332e-01 9.01060939e-01
3.98847133e-01 3.62162203e-01 9.52449083e-01 -2.61737287e-01
-1.07524887e-01 6.09559178e-01 4.71059620e-01 5.97525835e-01
9.24792588e-01 -5.20033725e-02 -5.28827488e-01 -5.02284646e-01
5.80129921e-01 2.99021900e-01 -4.07902122e-01 -2.73815632e-01
-1.19519448e+00 1.31414509e+00 5.76107621e-01 5.20337760e-01
-4.17316020e-01 2.11360782e-01 5.70337236e-01 8.54933202e-01
7.83833981e-01 5.35753071e-01 -2.52096504e-01 -2.06186518e-01
-6.06486857e-01 1.94574267e-01 1.33009648e+00 4.95606393e-01
7.15950429e-01 -4.15704638e-01 1.61617428e-01 5.02551734e-01
4.76799220e-01 -8.67301300e-02 4.93391931e-01 -5.51448405e-01
8.00291896e-01 7.49230444e-01 -7.54422857e-04 -1.72048008e+00
-2.28254944e-01 -2.76803434e-01 -8.74467015e-01 -4.31864001e-02
6.05710804e-01 -1.00692317e-01 1.87580079e-01 1.31127226e+00
1.58262208e-01 -1.08773880e-01 -1.19858667e-01 6.62712753e-01
3.91505659e-01 2.49973297e-01 -1.32501528e-01 -2.85037816e-01
1.21577251e+00 -5.97807884e-01 -4.43630219e-01 2.14486286e-01
7.76483476e-01 -4.10002589e-01 7.55731225e-01 1.99049503e-01
-6.60314798e-01 1.65952712e-01 -7.27120280e-01 1.30805567e-01
-4.98280376e-01 -7.32513905e-01 1.08908081e+00 8.00728858e-01
-9.48040724e-01 1.17068350e+00 -4.80078220e-01 -7.14495242e-01
5.37032008e-01 -1.36320859e-01 -5.50632060e-01 3.49470116e-02
-1.28905094e+00 8.53100479e-01 3.24688971e-01 -4.57792699e-01
-3.15867364e-01 -4.79991496e-01 -3.89470637e-01 3.57113600e-01
3.58933061e-01 -3.82319659e-01 6.20975912e-01 -6.21691942e-01
-9.89236593e-01 9.11386967e-01 6.14404440e-01 -8.69224250e-01
8.99004102e-01 8.41097534e-02 -4.37912822e-01 6.88120872e-02
-4.47300589e-03 -1.08340949e-01 6.17324173e-01 -5.26442587e-01
-4.78020310e-01 -5.05480766e-01 6.20551221e-02 -2.19399139e-01
-1.05991602e+00 5.17877042e-02 2.69813508e-01 -6.34308040e-01
7.46968687e-02 -5.18695533e-01 -5.08555025e-02 1.88941225e-01
-4.82286096e-01 -4.58359569e-01 7.40056992e-01 -8.72572482e-01
1.18257451e+00 -2.15321922e+00 -6.40248358e-02 3.25401992e-01
6.02849960e-01 -2.39620715e-01 2.25132287e-01 1.00549042e+00
-7.96553716e-02 6.45244241e-01 5.95413968e-02 8.72969180e-02
2.59525329e-02 -2.19676554e-01 -1.77262262e-01 6.60349905e-01
9.33060944e-02 8.05415869e-01 -9.84165072e-01 -7.31891274e-01
-2.89903611e-01 8.74349847e-02 -5.40546715e-01 -4.81337644e-02
1.78977400e-02 8.96839947e-02 -7.85847604e-01 3.41356635e-01
2.78471589e-01 -6.14787698e-01 3.49437684e-01 -9.63714644e-02
1.41458660e-01 3.45833838e-01 -9.14959610e-01 9.85610008e-01
-6.62514591e-04 8.24347854e-01 -1.73103884e-01 -1.38493145e+00
8.75722051e-01 1.68009967e-01 6.60309792e-01 -3.62844616e-01
9.72790271e-02 -3.09198932e-03 1.57641649e-01 -2.18588278e-01
4.91220534e-01 -1.04515672e-01 2.78394073e-01 7.71836817e-01
-2.61542082e-01 2.36235991e-01 1.42590076e-01 5.21862864e-01
1.49298596e+00 -5.36547780e-01 6.63283408e-01 -3.39799941e-01
2.51002330e-03 3.10667250e-02 2.79461652e-01 3.93421143e-01
-3.57611537e-01 4.13447320e-01 1.42004049e+00 -2.48762712e-01
-1.15172446e+00 -6.11223757e-01 -1.46887109e-01 5.54797769e-01
-2.40304053e-01 -4.41554576e-01 -2.18556464e-01 -9.98996496e-01
6.97467327e-01 2.60381728e-01 -5.28583586e-01 -8.88649598e-02
-3.11934590e-01 -1.10740387e+00 3.34763259e-01 6.51291683e-02
3.27087700e-01 -9.50931132e-01 -1.09415546e-01 3.75475228e-01
2.06694305e-01 -8.85269582e-01 -2.71230608e-01 3.98826785e-02
-1.18189728e+00 -1.58861732e+00 -7.65529573e-01 -4.66261446e-01
3.60834092e-01 2.64202148e-01 1.39286542e+00 2.84312695e-01
-1.84873357e-01 1.98525205e-01 -5.47927856e-01 -2.91786253e-01
-4.58523840e-01 1.21599399e-01 -8.06147680e-02 2.79172450e-01
3.71833056e-01 -1.10052419e+00 -6.17896318e-01 2.06256405e-01
-6.50584340e-01 -4.37956780e-01 3.79280388e-01 6.44988120e-01
5.09061068e-02 1.96053848e-01 8.92835081e-01 -1.36635494e+00
1.05703545e+00 -1.13064516e+00 -5.00090539e-01 -5.46732806e-02
-1.29006016e+00 -5.03750183e-02 3.06602746e-01 -1.50382057e-01
-4.24691975e-01 -1.07426679e+00 4.11495447e-01 -1.03910588e-01
4.01226401e-01 1.02621090e+00 1.44110084e-01 -1.87407985e-01
6.77656472e-01 -1.79162234e-01 2.25923210e-01 -6.53564930e-01
5.02501726e-02 5.24174154e-01 2.03352422e-02 -1.84493944e-01
8.09725046e-01 2.73035347e-01 3.45703870e-01 -7.65121043e-01
-5.38346410e-01 -6.13965452e-01 -1.92697749e-01 1.86732695e-01
4.00057286e-01 -4.55784887e-01 -8.25530171e-01 2.69988328e-01
-1.01529944e+00 2.47520030e-01 -4.03608620e-01 4.68576074e-01
-1.93190619e-01 7.86452532e-01 -7.54495561e-01 -7.23639071e-01
-3.64253163e-01 -4.40683693e-01 2.03661114e-01 3.27084102e-02
-6.33189604e-02 -1.34854102e+00 4.95677173e-01 2.31869251e-01
4.50330675e-01 7.37074256e-01 1.23448920e+00 -1.09574366e+00
-7.30599105e-01 -7.66473353e-01 -5.98411739e-01 5.64194024e-01
-2.20125895e-02 -1.46460474e-01 -7.65651047e-01 -3.79512578e-01
-2.29667664e-01 7.54809752e-02 8.02008033e-01 -2.12491192e-02
9.78442490e-01 -6.19817436e-01 -4.41985220e-01 2.00130939e-01
1.49184549e+00 -4.28238839e-01 7.08098412e-01 3.87039423e-01
5.36050439e-01 7.88326144e-01 3.44220191e-01 5.17807841e-01
3.92601848e-01 5.44466257e-01 6.44600511e-01 4.21204627e-01
2.10430548e-01 -5.09818614e-01 2.52572149e-01 1.08523524e+00
-4.22158092e-01 -4.65528309e-01 -9.11976278e-01 5.68708837e-01
-1.98233509e+00 -9.63096201e-01 -1.63066626e-01 2.31704736e+00
7.08477497e-01 3.97045344e-01 5.69907546e-01 3.56627136e-01
8.35274756e-01 4.76959169e-01 -2.08954394e-01 -5.98695781e-03
-3.24370235e-01 -1.29860103e-01 7.00725079e-01 -2.22365689e-02
-5.43038905e-01 3.05615038e-01 5.81211662e+00 4.34331030e-01
-6.95996165e-01 2.89505810e-01 7.52542436e-01 4.08240221e-02
-4.36650425e-01 4.33080822e-01 -3.92840385e-01 7.47858346e-01
9.97428536e-01 -6.14027381e-01 2.89145678e-01 8.76210153e-01
6.88262805e-02 -6.18998259e-02 -1.33587539e+00 7.47549891e-01
-3.13440502e-01 -1.46757007e+00 -2.18422949e-01 8.27324629e-01
4.31255937e-01 3.62403005e-01 -2.57148087e-01 1.69946849e-01
4.89981651e-01 -8.99821043e-01 3.21811765e-01 5.32108068e-01
4.15001512e-01 -7.61795700e-01 9.51979399e-01 2.72306800e-01
-4.81035471e-01 2.92780511e-02 -5.38664639e-01 -5.51856705e-04
1.56101612e-02 1.10069013e+00 -7.99628198e-01 4.68786448e-01
7.33141124e-01 1.08944607e+00 -6.06144071e-01 1.21935463e+00
-5.74876107e-02 8.97437692e-01 -1.96849823e-01 -3.48013818e-01
-2.75903903e-02 -6.74703419e-01 6.31946445e-01 8.42346489e-01
2.72258669e-01 -7.94642717e-02 -3.83283377e-01 7.62777090e-01
-5.85942566e-01 5.16437709e-01 -1.00404322e+00 -6.29606009e-01
3.44942570e-01 1.35885346e+00 -6.40191436e-01 1.49903625e-01
-6.09535575e-01 5.41868806e-01 4.97753888e-01 1.26893684e-01
-4.05149132e-01 -4.32481080e-01 6.25993550e-01 8.13488841e-01
1.66757286e-01 -1.48138881e-01 5.96448071e-02 -1.28069031e+00
9.31956843e-02 -6.25890553e-01 4.61817414e-01 -3.41479897e-01
-1.82327151e+00 7.51704574e-02 -2.98218042e-01 -1.06046438e+00
-7.97775909e-02 -5.52690208e-01 -1.18457472e+00 6.66103661e-01
-1.65102315e+00 -7.91065574e-01 -1.49975970e-01 3.85173231e-01
-3.15564066e-01 -4.97737199e-01 5.36797166e-01 2.67453283e-01
-4.99644667e-01 5.61643660e-01 5.13507783e-01 2.52624929e-01
5.29063940e-01 -1.29255068e+00 4.72395986e-01 4.67707127e-01
5.07459342e-01 2.53367186e-01 9.03878689e-01 -6.94419920e-01
-1.35060811e+00 -7.27599263e-01 9.52639699e-01 -5.10416269e-01
1.43975747e+00 -6.52465820e-02 -9.39693272e-01 6.90812051e-01
-1.09154493e-01 2.59448439e-01 3.41155589e-01 3.78070891e-01
-3.10006410e-01 -1.82432130e-01 -1.36983562e+00 3.14222425e-01
1.02792346e+00 -5.62057912e-01 -3.65870237e-01 9.28347290e-01
5.75782299e-01 3.51742566e-01 -1.20602989e+00 -1.44891560e-01
2.04846963e-01 -1.05695748e+00 9.12355602e-01 -8.64150643e-01
6.26007795e-01 3.45060766e-01 -7.51079470e-02 -1.46619487e+00
-3.98110360e-01 -7.46957541e-01 -2.75142699e-01 1.35767138e+00
5.03247976e-01 -1.23663759e+00 1.10703146e+00 2.95132607e-01
6.96286500e-01 -6.37610078e-01 -1.18968213e+00 -8.02933395e-01
1.68853685e-01 1.43362030e-01 5.23614824e-01 1.29597044e+00
5.10371089e-01 3.46546680e-01 -2.37422913e-01 -2.70831496e-01
6.81376934e-01 3.53817642e-02 8.26735079e-01 -1.83934605e+00
-4.01350975e-01 -5.29940605e-01 -1.01533628e+00 -2.74800330e-01
-4.23024707e-02 -1.11104083e+00 -1.03981304e+00 -1.18791997e+00
3.31967771e-01 -4.12178844e-01 -5.86146951e-01 6.19265176e-02
2.29650170e-01 6.82129264e-02 9.18298215e-02 5.46668708e-01
-3.38450044e-01 5.36224782e-01 1.03820074e+00 -1.79992199e-01
2.60180324e-01 2.26306282e-02 -8.21616948e-01 5.08045673e-01
6.24078155e-01 -7.78804541e-01 -1.74235106e-01 4.36131544e-02
9.52548265e-01 3.78060073e-01 5.69421530e-01 -6.01150990e-01
7.57437870e-02 -1.87743261e-01 -2.87375718e-01 -5.78524880e-02
-1.25414610e-01 -8.00029278e-01 8.19437727e-02 5.21258116e-01
-3.37107740e-02 3.73169154e-01 -4.55606997e-01 1.15380454e+00
-4.86108750e-01 -4.10141200e-01 3.30899060e-01 -1.54669806e-01
-1.96540043e-01 4.82404351e-01 2.99627095e-01 4.91209328e-01
9.57408726e-01 -1.90199628e-01 -6.06694698e-01 -7.25582182e-01
-4.75573659e-01 5.51295616e-02 4.63905334e-01 3.51008892e-01
2.14292675e-01 -1.13395107e+00 -1.08880091e+00 -4.19500053e-01
1.97615296e-01 -3.94408703e-01 -4.39302802e-01 1.09266496e+00
-5.42588115e-01 4.44188029e-01 -1.07446961e-01 -1.17308505e-01
-7.83670485e-01 3.33941400e-01 2.68519938e-01 -6.56546652e-01
-8.50460351e-01 5.77092409e-01 -1.94070280e-01 7.72295594e-02
3.09325438e-02 -1.11566022e-01 -3.49007666e-01 5.36189973e-01
-8.90540052e-03 7.70201504e-01 1.37032643e-01 -7.85720050e-02
-2.83276856e-01 -1.30989715e-01 -1.46462634e-01 1.38430759e-01
1.92182648e+00 5.68279512e-02 -3.46341372e-01 4.79223937e-01
1.27566946e+00 7.86276981e-02 -7.40102053e-01 -2.80941129e-01
6.11557841e-01 -6.84951484e-01 1.20650649e-01 -3.15529019e-01
-1.28371727e+00 8.09424043e-01 3.70992512e-01 1.19900775e+00
3.19354445e-01 2.91025955e-02 7.25736499e-01 4.44137931e-01
7.92659998e-01 -8.87218773e-01 2.03433752e-01 1.48285672e-01
8.59849691e-01 -1.54095030e+00 1.86529309e-01 -3.53055149e-01
-3.52049023e-01 9.49500978e-01 1.91653565e-01 -3.63736451e-01
1.16965055e+00 -2.90564686e-01 -4.80753809e-01 -7.24608660e-01
-4.47728395e-01 3.83740574e-01 -9.73704383e-02 3.35176915e-01
5.13400614e-01 3.28064114e-01 -6.92409039e-01 5.01096725e-01
-2.60819316e-01 -2.48067856e-01 9.02439415e-01 3.95270050e-01
-3.86653900e-01 -1.02987683e+00 2.31563494e-01 1.10741925e+00
-7.82768250e-01 -3.23021412e-01 -3.27248573e-01 9.74208474e-01
-6.85406029e-01 4.31219995e-01 -4.37979363e-02 -2.30027258e-01
8.25891271e-02 5.19342348e-02 -3.87162305e-02 -6.97245896e-01
-6.79551303e-01 -7.61614919e-01 4.20024097e-01 -4.17772681e-01
-2.28382856e-01 -9.80440259e-01 -6.03364944e-01 -8.33196700e-01
-5.00631332e-01 2.79365838e-01 5.35284519e-01 5.51784575e-01
3.32369506e-01 -1.35233337e-02 8.27643692e-01 -1.93077341e-01
-8.21175694e-01 -1.22999692e+00 -1.31628072e+00 7.32859492e-01
1.10254094e-01 -5.13979137e-01 -7.55831838e-01 -5.23270071e-01] | [6.972987651824951, 5.982858180999756] |
0dbb2958-4db7-4b05-8ab1-05ab00bc8c9b | face-parsing-with-roi-tanh-warping-1 | 1906.01342 | null | https://arxiv.org/abs/1906.01342v1 | https://arxiv.org/pdf/1906.01342v1.pdf | Face Parsing with RoI Tanh-Warping | Face parsing computes pixel-wise label maps for different semantic components (e.g., hair, mouth, eyes) from face images. Existing face parsing literature have illustrated significant advantages by focusing on individual regions of interest (RoIs) for faces and facial components. However, the traditional crop-and-resize focusing mechanism ignores all contextual area outside the RoIs, and thus is not suitable when the component area is unpredictable, e.g. hair. Inspired by the physiological vision system of human, we propose a novel RoI Tanh-warping operator that combines the central vision and the peripheral vision together. It addresses the dilemma between a limited sized RoI for focusing and an unpredictable area of surrounding context for peripheral information. To this end, we propose a novel hybrid convolutional neural network for face parsing. It uses hierarchical local based method for inner facial components and global methods for outer facial components. The whole framework is simple and principled, and can be trained end-to-end. To facilitate future research of face parsing, we also manually relabel the training data of the HELEN dataset and will make it public. Experiments on both HELEN and LFW-PL benchmarks demonstrate that our method surpasses state-of-the-art methods. | ['Lu Yuan', 'Fang Wen', 'Dong Chen', 'Hao Yang', 'Jinpeng Lin', 'Ming Zeng'] | 2019-06-04 | face-parsing-with-roi-tanh-warping | http://openaccess.thecvf.com/content_CVPR_2019/html/Lin_Face_Parsing_With_RoI_Tanh-Warping_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Lin_Face_Parsing_With_RoI_Tanh-Warping_CVPR_2019_paper.pdf | cvpr-2019-6 | ['face-parsing'] | ['computer-vision'] | [ 1.59825593e-01 4.37694937e-01 1.99636109e-02 -7.82082915e-01
-3.36782485e-01 -4.27424490e-01 1.77746326e-01 -4.94912803e-01
-2.44520217e-01 3.61500353e-01 8.61671716e-02 1.70803413e-01
2.89809763e-01 -6.54915810e-01 -6.60719097e-01 -8.07559967e-01
2.64177293e-01 -9.43619013e-02 3.13482314e-01 8.03532824e-02
8.95990245e-03 5.92026889e-01 -1.48873794e+00 3.52631956e-01
3.21746528e-01 1.23606467e+00 1.95304036e-01 2.54707068e-01
-3.41269344e-01 5.67206383e-01 -4.19093370e-01 -3.15792114e-01
2.52852410e-01 -5.94721735e-01 -6.11191571e-01 1.68069527e-01
6.70457721e-01 -3.98925751e-01 5.89515418e-02 1.20362651e+00
5.82682490e-01 -1.10052288e-01 2.65318066e-01 -1.13680971e+00
-7.93561518e-01 4.47508007e-01 -1.11014831e+00 1.27207171e-02
8.19774047e-02 -1.52371034e-01 6.20577872e-01 -1.01341474e+00
6.78112209e-01 1.54994392e+00 6.60700977e-01 1.08043170e+00
-9.61488485e-01 -1.00275195e+00 5.71799338e-01 -1.30901501e-01
-1.14322901e+00 -6.56518340e-01 1.02053869e+00 -1.99886426e-01
5.57016790e-01 1.13390826e-01 2.99138963e-01 8.75205278e-01
2.82896996e-01 6.82386577e-01 1.22937834e+00 -2.25443512e-01
1.73812240e-01 -1.23811148e-01 1.01747982e-01 1.01707745e+00
-1.03008315e-01 -2.54621319e-02 -4.82659280e-01 5.80502301e-02
8.10305893e-01 5.87562211e-02 -9.72355083e-02 -2.59514481e-01
-6.70051515e-01 7.09207535e-01 4.88060206e-01 1.96713522e-01
-3.21471095e-02 2.44792864e-01 1.85674340e-01 -2.27970347e-01
6.69490874e-01 -3.04256260e-01 -6.00122511e-01 6.22123241e-01
-1.06292200e+00 -5.93402013e-02 4.02982503e-01 8.41662705e-01
8.67996812e-01 -2.71177024e-01 -2.68307984e-01 1.13262355e+00
6.29952431e-01 2.59192824e-01 3.60701978e-01 -1.20208406e+00
2.34927520e-01 5.82856536e-01 -3.42778981e-01 -7.37979352e-01
-7.40154624e-01 5.95428832e-02 -6.54166698e-01 4.23377246e-01
4.48201060e-01 -3.53649765e-01 -1.22663450e+00 2.12143469e+00
7.42842436e-01 3.79266202e-01 -2.19272807e-01 9.53912497e-01
1.31868410e+00 4.34588194e-01 4.97143179e-01 -3.04449141e-01
1.76846206e+00 -1.20411754e+00 -8.25936079e-01 -6.56594276e-01
1.19873345e-01 -7.80893683e-01 8.62745702e-01 3.69821228e-02
-1.09539986e+00 -7.63947487e-01 -8.83368909e-01 -3.79719466e-01
-4.84955192e-01 3.09788585e-01 6.62853003e-01 7.22359180e-01
-1.29455590e+00 4.05467391e-01 -7.31815577e-01 -4.00568724e-01
8.35041165e-01 3.78953964e-01 -6.39850795e-01 -4.58461419e-02
-8.87394726e-01 4.66023922e-01 6.95726052e-02 5.65684795e-01
-6.68315232e-01 -5.73578119e-01 -9.18185771e-01 -6.80018775e-03
3.59488249e-01 -4.55631107e-01 1.16117716e+00 -1.20649743e+00
-1.63323832e+00 1.09925711e+00 -3.54558945e-01 1.58613816e-01
2.00027257e-01 -1.58291936e-01 -1.29988626e-01 4.22175348e-01
7.57773146e-02 1.29725587e+00 1.18142748e+00 -1.34898555e+00
-3.77994478e-01 -7.93062329e-01 -1.94782734e-01 -3.97242084e-02
-2.38014728e-01 5.63586354e-01 -5.89097679e-01 -4.85418886e-01
4.11184490e-01 -7.49942899e-01 2.98613366e-02 3.17375213e-01
-3.53954732e-01 -3.90880853e-01 1.07573545e+00 -8.76721263e-01
9.30262685e-01 -2.09399939e+00 -1.57130510e-01 -1.13586955e-01
3.35524738e-01 2.40171626e-01 -3.74967366e-01 -5.91007434e-02
-3.97582322e-01 1.99938133e-01 -3.85277241e-01 -5.96217215e-01
-2.76473880e-01 5.52949868e-02 -1.30774915e-01 5.66291392e-01
5.05078018e-01 8.51285696e-01 -5.62693238e-01 -9.14566815e-01
8.28175768e-02 8.02130759e-01 -4.68149126e-01 1.25904024e-01
-1.54855326e-01 4.55169022e-01 -4.86610234e-01 1.15719008e+00
1.21059656e+00 1.38812453e-01 6.33831918e-02 -2.41401717e-01
-6.10054210e-02 -2.21734151e-01 -9.51428056e-01 1.74411094e+00
-2.86724120e-01 3.49045664e-01 6.15785420e-01 -7.26447105e-01
9.57209349e-01 2.09153637e-01 4.15560156e-01 -6.11253083e-01
3.31913173e-01 -2.00373828e-01 -2.64826834e-01 -5.13969243e-01
-2.32839465e-01 -1.53914586e-01 2.02321202e-01 3.31847727e-01
2.94022083e-01 1.26846790e-01 -1.61385924e-01 -6.61278144e-02
8.93165708e-01 6.03946924e-01 1.49286509e-01 -3.37013483e-01
5.53782046e-01 -5.99747837e-01 9.22877133e-01 -1.14655532e-02
-7.32507825e-01 9.16315258e-01 8.04157853e-01 -4.60768670e-01
-5.00448048e-01 -9.85159874e-01 -2.42730781e-01 1.51108801e+00
3.62417907e-01 -3.50420892e-01 -1.54932654e+00 -9.37528789e-01
-1.89313829e-01 1.99738547e-01 -1.02038825e+00 1.31895365e-02
-7.10383236e-01 -6.50028884e-01 3.50244880e-01 7.71388113e-01
8.23696554e-01 -1.46816921e+00 -6.21756196e-01 6.08639382e-02
-1.05771825e-01 -1.38357389e+00 -6.55391634e-01 -5.78691922e-02
-7.04668462e-01 -1.01553583e+00 -5.44069886e-01 -1.13909817e+00
8.22550356e-01 3.65091920e-01 1.04724073e+00 1.04034245e-01
-5.34784496e-01 3.11894089e-01 -6.84101358e-02 -3.56422126e-01
1.40474245e-01 -3.73038709e-01 -4.05647725e-01 1.71104193e-01
4.00336593e-01 -2.73296118e-01 -8.36477995e-01 5.30218422e-01
-6.89706206e-01 5.24924137e-03 5.45472145e-01 7.14895010e-01
6.77614272e-01 -9.26085040e-02 6.54449344e-01 -1.04687500e+00
2.95728862e-01 -3.72677058e-01 -4.50181782e-01 4.97240722e-01
-2.03785807e-01 -3.32007051e-01 3.92889798e-01 -3.84082079e-01
-1.47752964e+00 4.79947418e-01 -1.46185413e-01 -1.89883232e-01
-3.36431742e-01 -1.52065024e-01 -6.08320296e-01 -2.88934857e-01
1.53886393e-01 -3.05258781e-01 5.38766645e-02 -3.82285088e-01
4.86014783e-01 5.83793283e-01 6.13019705e-01 -6.12549245e-01
3.78500253e-01 7.05243766e-01 2.08471902e-02 -7.81808257e-01
-8.32409620e-01 -1.94397733e-01 -7.64622927e-01 -3.72813076e-01
1.18312001e+00 -7.94364214e-01 -8.43567967e-01 3.91248614e-01
-1.37169230e+00 -2.77752459e-01 -2.81480927e-04 1.45636341e-02
-4.54347759e-01 4.06864852e-01 -7.83390701e-01 -7.10835218e-01
-4.82285202e-01 -1.15106559e+00 1.52233481e+00 7.90055096e-01
-4.06590849e-03 -7.47646511e-01 -3.05664897e-01 5.25238037e-01
2.37360269e-01 3.30164611e-01 6.64080203e-01 -2.85058498e-01
-3.74442607e-01 1.04142897e-01 -5.39000154e-01 4.94003415e-01
8.10853089e-04 2.22362652e-01 -1.66008687e+00 -5.75555535e-03
2.54561961e-01 -3.43930572e-01 1.05779850e+00 4.98614609e-01
1.47659135e+00 -6.07746653e-03 -3.90933692e-01 6.22499943e-01
1.26611352e+00 1.51832834e-01 7.08251059e-01 -2.48094678e-01
6.75927579e-01 1.35739005e+00 5.83183348e-01 1.97242677e-01
2.59370893e-01 3.68069023e-01 5.81706285e-01 -3.82275969e-01
-4.72659171e-01 -1.38430312e-01 4.69766259e-01 2.11367935e-01
7.13279322e-02 3.63379084e-02 -4.51456726e-01 3.69628608e-01
-1.64588308e+00 -6.49560392e-01 2.53676593e-01 1.72390056e+00
7.40756631e-01 -1.58509359e-01 -4.73126620e-02 -3.51869285e-01
1.02710044e+00 1.96078524e-01 -6.22586429e-01 -4.08874929e-01
-2.72977129e-02 4.74410027e-01 2.51560718e-01 3.25668097e-01
-1.32835221e+00 1.33851635e+00 6.19298792e+00 8.10952723e-01
-1.01925290e+00 3.14434499e-01 1.07649636e+00 8.06142762e-03
1.92266420e-01 -1.04628250e-01 -1.08954167e+00 4.11687553e-01
6.76167667e-01 6.56944454e-01 4.98330981e-01 9.03759241e-01
1.14645578e-01 -1.63323045e-01 -1.04108143e+00 1.03106368e+00
2.44105503e-01 -8.12655628e-01 -3.96335274e-01 -1.19878970e-01
4.34521526e-01 -2.44713753e-01 5.78209981e-02 1.17055118e-01
-4.48150821e-02 -1.21407473e+00 6.26327336e-01 4.21610475e-01
7.48417199e-01 -5.17104924e-01 6.22805476e-01 2.87516303e-02
-1.36262095e+00 -2.06004515e-01 -6.14119768e-01 2.69021392e-01
-1.92877859e-01 4.52576518e-01 -3.86686176e-01 -3.98624539e-02
9.58740652e-01 3.91060263e-01 -6.90876484e-01 4.53686506e-01
-3.26391429e-01 3.95004660e-01 -3.01862270e-01 3.54733080e-01
4.95809987e-02 -3.38542372e-01 -3.37047279e-02 1.14725542e+00
2.20327690e-01 1.81678265e-01 -2.07518041e-02 1.12837374e+00
-2.90658832e-01 1.86767355e-01 -4.75903243e-01 1.68192118e-01
4.01223153e-01 2.02219200e+00 -1.17521966e+00 -8.03218856e-02
-6.83632195e-01 8.79415929e-01 3.87721986e-01 3.64989877e-01
-8.42813969e-01 -1.54183984e-01 5.47426820e-01 -5.46943396e-03
3.40428531e-01 1.11946233e-01 -4.59807575e-01 -7.55915225e-01
7.69145265e-02 -4.15662080e-01 3.13417137e-01 -8.40876758e-01
-1.19070125e+00 6.11844242e-01 -5.37839383e-02 -6.11348271e-01
1.32811904e-01 -8.59187186e-01 -7.65574753e-01 7.48435378e-01
-1.63133216e+00 -1.54661155e+00 -6.47746742e-01 6.36092842e-01
7.04177141e-01 3.52502093e-02 7.26459861e-01 1.05216034e-01
-8.83763611e-01 6.27299309e-01 -5.14634252e-01 3.72456640e-01
9.15668368e-01 -9.95118439e-01 1.77261978e-01 7.89585888e-01
-1.06090516e-01 6.99192584e-01 2.53746897e-01 -5.49742162e-01
-1.12635171e+00 -1.14322054e+00 8.11742604e-01 -2.98486948e-01
2.69490302e-01 -5.47364771e-01 -7.67022192e-01 6.48288071e-01
3.63402784e-01 5.08834183e-01 5.70123017e-01 -6.10512756e-02
-5.26287556e-01 -4.50486362e-01 -1.57263565e+00 5.26331425e-01
1.15049505e+00 -3.28473777e-01 -3.62722188e-01 3.22097152e-01
6.84961975e-01 -2.46791288e-01 -6.76330924e-01 5.22502303e-01
6.26049280e-01 -1.23353016e+00 9.96352017e-01 -2.47352511e-01
4.15382177e-01 -2.99422264e-01 -9.24914330e-02 -8.69767010e-01
-2.33326733e-01 -5.61104178e-01 1.24947838e-01 1.71661150e+00
4.06627879e-02 -6.25053704e-01 7.62072682e-01 6.58529162e-01
-1.27977639e-01 -8.44450414e-01 -9.72003043e-01 -2.53769785e-01
7.49527961e-02 -1.91256344e-01 6.41214788e-01 5.28193772e-01
-2.57595599e-01 4.12024468e-01 -1.14698589e-01 1.10729866e-01
6.34750903e-01 9.10902619e-02 3.80068928e-01 -1.19353700e+00
1.70584217e-01 -4.86048639e-01 -1.61462694e-01 -7.61330605e-01
4.95072305e-01 -5.51587641e-01 1.87736332e-01 -1.21529412e+00
3.37382346e-01 -1.78530827e-01 -2.06545144e-01 9.61730778e-01
-1.48129582e-01 6.85345113e-01 1.71238959e-01 -1.79024622e-01
-3.51040512e-01 2.23177627e-01 1.59429526e+00 3.66371423e-02
5.43193109e-02 -2.19045654e-01 -9.22648072e-01 9.99667585e-01
8.45225573e-01 -3.07685882e-01 -4.77954119e-01 -2.99932480e-01
-2.78651148e-01 -1.40103281e-01 3.20103168e-01 -7.59323180e-01
2.54610300e-01 -4.53893580e-02 8.12582910e-01 -5.32670021e-01
3.73059422e-01 -8.35646093e-01 -2.78010160e-01 6.89715296e-02
-1.91373408e-01 -2.72364676e-01 1.63509652e-01 4.34791714e-01
-1.54364511e-01 -2.39946604e-01 1.28244460e+00 -3.20932895e-01
-7.29471385e-01 4.10404682e-01 1.51139989e-01 -1.11493938e-01
1.22590172e+00 -2.10750207e-01 -5.31012714e-01 3.31721902e-02
-6.76388085e-01 1.51316628e-01 2.48207241e-01 4.32626456e-01
7.10055470e-01 -1.03857231e+00 -4.96052086e-01 6.49498165e-01
-1.15977958e-01 1.31409407e-01 3.97649616e-01 6.45097077e-01
-5.23777127e-01 1.99602470e-01 -3.59517276e-01 -4.87231165e-01
-1.51870120e+00 6.65308893e-01 4.09859538e-01 2.55996108e-01
-5.07735908e-01 1.13885391e+00 8.62046003e-01 -4.70234632e-01
2.82112956e-01 -4.81862098e-01 -4.42302465e-01 1.55897707e-01
7.41115749e-01 1.14293277e-01 -1.44791320e-01 -8.18066120e-01
-5.50457001e-01 1.13254094e+00 6.62025437e-02 1.96970075e-01
1.19640410e+00 -2.66982138e-01 -5.82204282e-01 4.62508425e-02
1.28351307e+00 4.52906489e-02 -1.57901597e+00 -5.32001220e-02
-2.12815449e-01 -2.66576827e-01 4.38705161e-02 -7.80097306e-01
-1.56147063e+00 1.18513656e+00 7.83302844e-01 -3.84714757e-03
1.54871976e+00 2.09371299e-01 7.87078142e-01 -2.22352728e-01
1.61410436e-01 -1.21398103e+00 3.35609948e-04 2.14876950e-01
9.21172380e-01 -1.30487645e+00 -4.65737171e-02 -9.33710098e-01
-5.53643942e-01 1.37514913e+00 9.23141420e-01 -9.58231986e-02
8.66701007e-01 5.23975730e-01 2.49376833e-01 -1.22306995e-01
-5.97561061e-01 -1.73634648e-01 1.75931677e-01 6.80114090e-01
5.61843336e-01 4.43329066e-02 -3.06327909e-01 8.89640868e-01
8.42501447e-02 -1.63291678e-01 6.31174073e-03 7.56137073e-01
-5.01575410e-01 -9.47223008e-01 -5.07681608e-01 7.66227022e-02
-6.38257802e-01 5.86985163e-02 -5.53121507e-01 8.84539425e-01
6.79661691e-01 1.01497221e+00 1.37220204e-01 -1.54799610e-01
-1.19220227e-01 2.07602218e-01 5.74045122e-01 -6.08664870e-01
-4.67473567e-01 5.79151273e-01 -2.45122954e-01 -8.85248005e-01
-5.16249359e-01 -5.00112772e-01 -1.57627320e+00 -6.94865407e-03
-1.35440633e-01 -3.65301102e-01 9.37399626e-01 8.01110208e-01
2.63680756e-01 4.33123767e-01 5.41885197e-01 -1.04537416e+00
-1.78597823e-01 -1.05980408e+00 -7.31625021e-01 1.71861485e-01
1.58019751e-01 -8.40561569e-01 -2.60132432e-01 2.50925094e-01] | [13.450897216796875, 0.6330985426902771] |
1ff120ac-fa40-41bd-a09d-45e86dd04bd6 | switch-bert-learning-to-model-multimodal | 2306.14182 | null | https://arxiv.org/abs/2306.14182v1 | https://arxiv.org/pdf/2306.14182v1.pdf | Switch-BERT: Learning to Model Multimodal Interactions by Switching Attention and Input | The ability to model intra-modal and inter-modal interactions is fundamental in multimodal machine learning. The current state-of-the-art models usually adopt deep learning models with fixed structures. They can achieve exceptional performances on specific tasks, but face a particularly challenging problem of modality mismatch because of diversity of input modalities and their fixed structures. In this paper, we present \textbf{Switch-BERT} for joint vision and language representation learning to address this problem. Switch-BERT extends BERT architecture by introducing learnable layer-wise and cross-layer interactions. It learns to optimize attention from a set of attention modes representing these interactions. One specific property of the model is that it learns to attend outputs from various depths, therefore mitigates the modality mismatch problem. We present extensive experiments on visual question answering, image-text retrieval and referring expression comprehension experiments. Results confirm that, whereas alternative architectures including ViLBERT and UNITER may excel in particular tasks, Switch-BERT can consistently achieve better or comparable performances than the current state-of-the-art models in these tasks. Ablation studies indicate that the proposed model achieves superior performances due to its ability in learning task-specific multimodal interactions. | ['Wei Chu', 'Kaisheng Yao', 'Qingpei Guo'] | 2023-06-25 | null | null | null | null | ['referring-expression', 'visual-question-answering-1', 'retrieval', 'question-answering'] | ['computer-vision', 'computer-vision', 'methodology', 'natural-language-processing'] | [ 2.52823215e-02 1.11213356e-01 -1.71718583e-01 -4.16818112e-01
-8.78192723e-01 -4.95350152e-01 8.75006318e-01 -5.74560203e-02
-6.19454682e-01 4.38386351e-01 3.01069409e-01 -2.31423140e-01
-1.46019921e-01 -3.31023544e-01 -8.03896785e-01 -4.67663318e-01
6.22428916e-02 6.59585118e-01 2.19001602e-02 -4.09296840e-01
1.92979306e-01 2.78973222e-01 -1.53564048e+00 1.02628517e+00
7.03987360e-01 8.90896440e-01 4.27107781e-01 8.84011745e-01
-4.47175920e-01 1.02744138e+00 -3.14153939e-01 -4.13416654e-01
-2.10769340e-01 -4.30719554e-01 -1.23145640e+00 -2.38037154e-01
7.03993678e-01 -3.44166726e-01 -4.25316304e-01 5.74017525e-01
6.19478941e-01 4.00669314e-02 8.40340018e-01 -1.34243357e+00
-9.70420003e-01 6.15914166e-01 -5.58379173e-01 2.25659549e-01
5.40988028e-01 2.07036898e-01 1.30469823e+00 -9.31858957e-01
2.92625904e-01 1.62231159e+00 2.96903342e-01 9.45656776e-01
-1.40923023e+00 -2.57022679e-01 5.04833162e-01 5.36007285e-01
-1.19930744e+00 -5.37179232e-01 5.84393382e-01 -2.90215850e-01
1.31173623e+00 3.15610617e-01 2.13875994e-01 1.32822669e+00
1.66753054e-01 1.31470025e+00 1.01028955e+00 -4.74521339e-01
-1.69302806e-01 1.06700622e-01 2.00080976e-01 7.37974346e-01
-4.78183240e-01 -1.07376888e-01 -9.50503469e-01 5.00182696e-02
6.58376992e-01 -1.17740452e-01 -4.66467559e-01 -4.69324172e-01
-1.53515339e+00 5.68412840e-01 8.25292587e-01 4.02837276e-01
-2.47905195e-01 3.72342646e-01 3.47315580e-01 4.31456178e-01
-1.20558396e-01 4.68845904e-01 -4.43560362e-01 6.97228014e-02
-5.05216837e-01 9.75675061e-02 6.94373965e-01 8.20147991e-01
5.81090629e-01 -8.36806074e-02 -5.97484648e-01 1.05075562e+00
5.51069677e-01 2.66639888e-01 3.82082433e-01 -9.06056941e-01
6.92797720e-01 6.30198419e-01 -2.24837944e-01 -6.04900360e-01
-7.42443562e-01 -2.30025485e-01 -7.48284400e-01 1.07101440e-01
4.88989592e-01 1.16857089e-01 -1.11675334e+00 2.02277064e+00
-1.80973068e-01 -3.34385961e-01 1.63417876e-01 1.00783360e+00
1.34594929e+00 6.74051762e-01 4.90275294e-01 3.13552350e-01
1.40739357e+00 -1.27825725e+00 -6.67328954e-01 -5.56775212e-01
4.36974049e-01 -5.96802533e-01 1.38576221e+00 1.20126776e-01
-1.41313541e+00 -6.52351797e-01 -7.79185653e-01 -4.03066695e-01
-4.31599349e-01 1.00687400e-01 6.89376712e-01 2.12506324e-01
-1.25003195e+00 8.49363506e-02 -4.68298554e-01 -6.30019546e-01
3.96330774e-01 6.53822541e-01 -5.17178595e-01 -8.76458455e-03
-1.14945865e+00 1.21436417e+00 2.84100205e-01 2.97633648e-01
-1.11810744e+00 -3.87936294e-01 -9.39476490e-01 3.57936382e-01
1.95442960e-01 -1.20670128e+00 1.39363682e+00 -1.24151242e+00
-1.47419643e+00 1.30223846e+00 -2.98771948e-01 -1.68441460e-01
1.87724248e-01 -2.32388705e-01 -1.28479213e-01 4.87901866e-01
-3.18467200e-01 1.30070698e+00 9.19329703e-01 -1.52681303e+00
-2.29980215e-01 -3.00133735e-01 5.38709939e-01 4.51174289e-01
-3.68071139e-01 -9.11866501e-02 -6.22923553e-01 -2.61184722e-01
3.41868773e-02 -7.52614737e-01 1.43300444e-01 3.99671346e-02
-2.33025536e-01 -2.97821164e-01 5.98140180e-01 -3.90973270e-01
8.96925807e-01 -2.03496528e+00 7.91997671e-01 -2.55559146e-01
1.77287951e-01 1.36640787e-01 -7.19267130e-01 8.14104855e-01
-1.35034218e-01 1.21044852e-01 -1.53079525e-01 -6.79441810e-01
1.48853719e-01 3.93680513e-01 -2.15644896e-01 2.22199216e-01
4.26658124e-01 1.41750777e+00 -4.70438302e-01 -4.75256681e-01
2.37173587e-01 3.98537576e-01 -5.52121460e-01 5.21400511e-01
-5.66057801e-01 6.56292260e-01 -1.73469096e-01 8.09881032e-01
4.32490617e-01 -4.43581492e-01 1.08374760e-01 -4.93261963e-01
2.06004322e-01 -4.86500226e-02 -5.80266118e-01 2.09069991e+00
-6.78064823e-01 8.81165802e-01 4.43223417e-01 -1.18194938e+00
3.93061370e-01 3.40770155e-01 -1.21126123e-01 -1.16720521e+00
1.69753671e-01 -1.05740108e-01 2.25791529e-01 -8.82957041e-01
3.33231181e-01 -1.14579715e-01 -6.80854172e-02 2.83383608e-01
4.57062989e-01 -1.37690883e-02 -5.94289675e-02 4.89519000e-01
7.41902888e-01 5.60525060e-02 -6.05002493e-02 -1.68908820e-01
7.62203276e-01 -3.22816283e-01 -7.42706954e-02 9.53677297e-01
-2.06510231e-01 7.05024600e-01 6.65856063e-01 -1.55750573e-01
-6.32948101e-01 -1.24973214e+00 3.49832475e-02 1.85102522e+00
3.11857402e-01 -2.82267988e-01 -4.99518961e-01 -5.29982805e-01
-1.76041156e-01 7.31517792e-01 -7.70732701e-01 -2.24234477e-01
-4.08380061e-01 -4.64102834e-01 6.28472626e-01 5.89964271e-01
5.87687433e-01 -1.26387703e+00 -5.26860714e-01 -2.68899560e-01
-3.52707654e-01 -1.10327816e+00 -1.38756603e-01 1.13899723e-01
-6.90138280e-01 -9.60453749e-01 -9.67908561e-01 -1.07890880e+00
5.98702133e-01 1.56494439e-01 1.50781143e+00 1.44743308e-01
-1.53381720e-01 9.90423203e-01 -2.56953597e-01 -2.63882071e-01
-1.26069501e-01 1.93579897e-01 -5.59781611e-01 4.49487045e-02
1.42558977e-01 -2.70317048e-01 -7.00613379e-01 3.51681948e-01
-1.18851268e+00 2.15654626e-01 6.68882787e-01 1.15817380e+00
1.71504229e-01 -6.50307715e-01 6.01767659e-01 -5.16475379e-01
8.00031066e-01 -3.97377551e-01 -6.65121004e-02 5.32742858e-01
3.40514891e-02 2.36340597e-01 2.58279294e-01 -4.08375800e-01
-1.07511413e+00 -1.75457969e-01 -1.53787285e-01 -2.60736942e-01
-2.94467956e-01 6.39914811e-01 -3.83378923e-01 -3.95466313e-02
5.23464620e-01 9.89920199e-02 5.24905138e-02 -2.82741159e-01
5.30689776e-01 4.34675068e-01 5.14863610e-01 -7.25041866e-01
2.54935980e-01 4.63177443e-01 -1.39335945e-01 -8.16140413e-01
-8.00752282e-01 -3.58977377e-01 -6.04429603e-01 -1.32789895e-01
1.17171025e+00 -9.46301401e-01 -1.04922700e+00 5.53444326e-01
-1.30337918e+00 -5.79888761e-01 1.62989050e-01 1.77228391e-01
-6.39080226e-01 2.46830389e-01 -6.62389636e-01 -8.22193742e-01
-1.41665339e-01 -1.17780745e+00 1.36719501e+00 3.39431554e-01
-2.24994689e-01 -1.23981321e+00 -2.50189275e-01 5.87261319e-01
5.64788401e-01 -1.36671796e-01 1.39894712e+00 -5.02011418e-01
-7.46123970e-01 1.45944521e-01 -4.82684821e-01 1.45181134e-01
-3.53324473e-01 -1.66329041e-01 -1.26067865e+00 -2.93417126e-01
-3.16107064e-01 -7.91888952e-01 1.36756384e+00 3.61032575e-01
1.32770610e+00 2.55012717e-02 -2.56765455e-01 5.31575084e-01
1.12319636e+00 -4.55435365e-02 6.71081185e-01 4.18276936e-01
6.21255636e-01 9.40650463e-01 2.58075178e-01 -1.62499733e-02
8.61935914e-01 6.66407287e-01 9.92200911e-01 -3.43788296e-01
-1.23194262e-01 -7.22225085e-02 3.41402501e-01 4.05330956e-01
2.82443315e-01 -4.06844050e-01 -1.09013438e+00 6.02841675e-01
-2.28672194e+00 -9.25783992e-01 1.42792286e-02 1.88966858e+00
6.79377615e-01 -6.96832761e-02 6.87257946e-02 -2.79011577e-01
3.41662943e-01 2.06937015e-01 -3.85276765e-01 -7.03858018e-01
-3.87884378e-01 -3.81647311e-02 -1.14824884e-01 5.18436253e-01
-9.05510068e-01 1.01133037e+00 6.66691971e+00 5.09526491e-01
-1.05751681e+00 1.12832658e-01 4.16327119e-01 -1.74286589e-01
-5.47539055e-01 -1.86975271e-01 -5.63167393e-01 -1.74934547e-02
6.55731320e-01 4.80828941e-01 5.54028392e-01 1.65794134e-01
-2.64928967e-01 -3.52647781e-01 -1.64323306e+00 1.31421387e+00
4.29789752e-01 -1.21527100e+00 3.31092358e-01 -3.06596696e-01
4.32455778e-01 -1.99465211e-02 3.90209436e-01 6.08148277e-01
-2.28153154e-01 -1.60204899e+00 7.05581963e-01 7.80271292e-01
5.47594249e-01 -5.36407709e-01 6.93541527e-01 5.34879863e-01
-8.37716162e-01 -3.70585740e-01 -1.18714035e-01 -9.47083160e-02
2.00856149e-01 -1.52340323e-01 -3.72911513e-01 3.71688902e-01
6.61052585e-01 3.93586874e-01 -7.66747892e-01 9.50863898e-01
-1.85850263e-01 2.71109492e-01 -5.08727022e-02 -7.08817691e-02
4.39893007e-01 1.97410569e-01 4.01362449e-01 1.39573681e+00
-6.51770383e-02 -9.83752832e-02 -4.74854559e-02 9.16646004e-01
-1.24414779e-01 -6.74161762e-02 -4.80416626e-01 -1.84286535e-01
9.98981595e-02 1.06671274e+00 -1.70699030e-01 -1.50868103e-01
-5.42854011e-01 1.13151395e+00 7.07392454e-01 8.77115667e-01
-7.93153942e-01 -1.54651284e-01 4.41495687e-01 -7.42774457e-02
1.71349019e-01 -3.51992428e-01 -2.95333773e-01 -1.15215707e+00
-1.79588318e-01 -9.96874452e-01 6.75755441e-01 -1.12441313e+00
-1.36511159e+00 6.56691909e-01 1.21730834e-01 -7.28746235e-01
-1.76743805e-01 -1.00731063e+00 -5.44703901e-01 9.41540897e-01
-1.61405218e+00 -1.78425086e+00 -3.40690017e-01 9.27308261e-01
6.40786171e-01 -3.13394278e-01 9.59172606e-01 1.36891603e-01
-4.62858498e-01 7.71257341e-01 -1.97129220e-01 -5.39527163e-02
7.78287768e-01 -1.23113072e+00 -1.17899738e-01 3.09530854e-01
2.58455485e-01 7.91625977e-01 4.40905660e-01 2.70967782e-02
-1.65364623e+00 -3.61743689e-01 8.13648880e-01 -5.95717371e-01
5.67901254e-01 -4.52682376e-01 -9.04120743e-01 7.93035269e-01
9.41861868e-01 -2.10894391e-01 8.17850351e-01 3.94726902e-01
-6.37213290e-01 -4.82002646e-03 -8.16070199e-01 7.76870787e-01
9.24350739e-01 -9.51905310e-01 -8.92681062e-01 2.06634313e-01
3.79920214e-01 -4.10755277e-01 -4.16656584e-01 5.02366781e-01
6.39331639e-01 -1.18026495e+00 1.06361008e+00 -1.00820327e+00
6.58635616e-01 -9.59850624e-02 -1.96798921e-01 -1.20088804e+00
-2.12258697e-01 -3.75365794e-01 -2.07835555e-01 1.22444272e+00
5.36131561e-01 -4.51049298e-01 3.26021403e-01 6.03296340e-01
2.39577349e-02 -6.98632777e-01 -9.77276385e-01 -2.50567615e-01
3.26372296e-01 -2.80075461e-01 1.93977743e-01 7.85076678e-01
1.37236133e-01 8.80361140e-01 -3.63760829e-01 1.17370732e-01
3.54833990e-01 1.97504610e-01 7.83562839e-01 -8.32115948e-01
-4.84478354e-01 -7.45867372e-01 -5.98097891e-02 -1.35200870e+00
5.94220817e-01 -9.89801645e-01 -1.09376214e-01 -1.94933033e+00
3.85611087e-01 -4.33130004e-02 -3.69891316e-01 4.93818223e-01
-1.87262669e-01 1.61464199e-01 5.96518695e-01 6.88568503e-02
-9.66188133e-01 6.38078690e-01 1.42625296e+00 -3.26913804e-01
-1.33240037e-02 -2.29035422e-01 -7.09071994e-01 6.72932923e-01
6.54985130e-01 1.76175892e-01 -4.50863689e-01 -1.10670614e+00
3.86000812e-01 6.00341558e-02 7.80950308e-01 -5.73493838e-01
2.67338395e-01 3.40909548e-02 4.30681437e-01 -5.41897416e-01
7.40307808e-01 -8.75395775e-01 -2.43243635e-01 1.36956349e-01
-7.72885501e-01 1.20271504e-01 6.21368766e-01 4.51320440e-01
-3.83616030e-01 -2.15252325e-01 6.52004421e-01 -2.98926830e-01
-8.88314366e-01 -1.04548065e-02 -6.18301690e-01 1.01723708e-02
6.62729263e-01 -9.57233161e-02 -6.57729506e-01 -6.89200938e-01
-9.60037649e-01 6.32558525e-01 1.55329898e-01 6.97723150e-01
6.76235557e-01 -1.23925197e+00 -6.05486095e-01 -2.85873879e-02
3.95923257e-01 -2.32885957e-01 4.55933452e-01 1.01244104e+00
-3.34616363e-01 6.28210425e-01 -3.45659077e-01 -9.82053876e-01
-1.27389836e+00 4.46913213e-01 5.90591133e-01 -1.72955692e-01
-6.39704987e-02 1.17435849e+00 7.28090048e-01 -5.25965631e-01
5.49997866e-01 -3.91902169e-03 -3.32553804e-01 1.51945189e-01
3.12745035e-01 1.96074434e-02 -1.85321987e-01 -6.28404021e-01
-4.39150482e-01 6.39497757e-01 -1.70627326e-01 -2.10842341e-01
9.60192025e-01 -4.08708990e-01 -1.76499248e-01 5.14309108e-01
1.16336119e+00 -4.57285017e-01 -1.03381133e+00 -1.17487758e-01
-1.51712075e-01 -2.61415452e-01 -7.49633834e-02 -1.19622993e+00
-8.28498363e-01 1.49326587e+00 6.06923997e-01 -3.95427272e-02
1.22450817e+00 4.20176953e-01 3.98753554e-01 5.19909680e-01
2.11690236e-02 -8.59947681e-01 5.09535789e-01 7.63000607e-01
1.28562796e+00 -1.46823776e+00 -3.73222798e-01 8.61748680e-03
-7.87512183e-01 9.90413964e-01 9.81188655e-01 2.13588312e-01
1.87246442e-01 -1.05236307e-01 2.67658949e-01 -2.46146679e-01
-1.08523118e+00 -5.27809441e-01 6.32095516e-01 5.62380970e-01
7.67674506e-01 -2.74256110e-01 -1.01521142e-01 3.62140536e-01
2.79826880e-01 -4.53745782e-01 3.51913832e-02 8.65076780e-01
-2.44905517e-01 -1.07428825e+00 -4.22182411e-01 1.20054156e-01
-2.80229956e-01 -2.49555454e-01 -7.12397099e-01 9.71346557e-01
-1.79260403e-01 9.97993588e-01 1.09364688e-02 -9.00798813e-02
4.23032016e-01 2.25282550e-01 9.57034171e-01 -4.09859538e-01
-8.97273540e-01 4.19362187e-02 1.32184550e-01 -6.16623938e-01
-6.50286376e-01 -2.98040450e-01 -1.25177932e+00 1.39293289e-02
-1.28961444e-01 -7.94573277e-02 4.83448952e-01 9.09983754e-01
4.57994282e-01 6.18431449e-01 5.14375828e-02 -1.04458392e+00
-5.60773909e-01 -8.90542150e-01 -2.68659711e-01 4.80775714e-01
6.30271792e-01 -6.41260862e-01 -1.72947645e-01 -1.50841931e-02] | [10.846662521362305, 1.6249767541885376] |
f9cb513e-401a-4e4c-adc3-15ed1bfbef0e | attention-map-guided-transformer-pruning-for | 2304.01452 | null | https://arxiv.org/abs/2304.01452v1 | https://arxiv.org/pdf/2304.01452v1.pdf | Attention Map Guided Transformer Pruning for Edge Device | Due to its significant capability of modeling long-range dependencies, vision transformer (ViT) has achieved promising success in both holistic and occluded person re-identification (Re-ID) tasks. However, the inherent problems of transformers such as the huge computational cost and memory footprint are still two unsolved issues that will block the deployment of ViT based person Re-ID models on resource-limited edge devices. Our goal is to reduce both the inference complexity and model size without sacrificing the comparable accuracy on person Re-ID, especially for tasks with occlusion. To this end, we propose a novel attention map guided (AMG) transformer pruning method, which removes both redundant tokens and heads with the guidance of the attention map in a hardware-friendly way. We first calculate the entropy in the key dimension and sum it up for the whole map, and the corresponding head parameters of maps with high entropy will be removed for model size reduction. Then we combine the similarity and first-order gradients of key tokens along the query dimension for token importance estimation and remove redundant key and value tokens to further reduce the inference complexity. Comprehensive experiments on Occluded DukeMTMC and Market-1501 demonstrate the effectiveness of our proposals. For example, our proposed pruning strategy on ViT-Base enjoys \textup{\textbf{29.4\%}} \textup{\textbf{FLOPs}} savings with \textup{\textbf{0.2\%}} drop on Rank-1 and \textup{\textbf{0.4\%}} improvement on mAP, respectively. | ['Heng-Tao Shen', 'Fumin Shen', 'Xingguo Huang', 'Zeren Sun', 'Yazhou Yao', 'Junzhu Mao'] | 2023-04-04 | null | null | null | null | ['person-re-identification'] | ['computer-vision'] | [ 8.38134065e-02 -5.14473543e-02 1.19941972e-01 -3.77550274e-01
-6.46335363e-01 -1.24858953e-01 1.65082484e-01 2.36823633e-01
-7.74242163e-01 7.72484422e-01 1.52049989e-01 -1.24868378e-01
-3.15852225e-01 -8.42615664e-01 -5.52723825e-01 -5.10880113e-01
1.41175970e-01 3.91615480e-01 2.81666845e-01 9.01217908e-02
9.41023380e-02 2.17539996e-01 -1.77015209e+00 -4.78761978e-02
1.21666491e+00 1.21972799e+00 3.50708276e-01 2.96357691e-01
3.84533517e-02 6.23758733e-01 -5.70492804e-01 -9.76261258e-01
1.47423819e-01 2.07448914e-03 -5.34337997e-01 -3.25344622e-01
6.36283100e-01 -6.11042559e-01 -7.23693550e-01 1.15367711e+00
1.02264953e+00 1.44098908e-01 5.30358076e-01 -1.04343915e+00
-3.35934311e-01 4.83655304e-01 -7.65212476e-01 5.14972329e-01
-2.67269611e-02 2.74249585e-03 8.11940074e-01 -9.91864324e-01
2.10488334e-01 9.94403720e-01 7.85107851e-01 4.59930331e-01
-8.10989797e-01 -8.07085872e-01 4.58842874e-01 5.84485471e-01
-1.61857724e+00 -4.98424441e-01 3.77380759e-01 -7.85848796e-02
1.20017970e+00 5.35490334e-01 6.64689779e-01 7.70277023e-01
-1.91541538e-01 7.29017317e-01 7.79338300e-01 -2.67987341e-01
-2.47028321e-02 1.30740210e-01 4.40701246e-01 8.12846482e-01
6.16071403e-01 -2.67210096e-01 -7.03932524e-01 9.25347060e-02
7.01957822e-01 -9.33473557e-02 -1.62995771e-01 2.74151176e-01
-8.02781522e-01 5.23331344e-01 2.96604872e-01 -6.43659532e-02
-2.31894836e-01 3.69096994e-01 4.11342531e-01 -1.95616126e-01
4.75542754e-01 -6.09312393e-02 -3.14833850e-01 -4.62765217e-01
-8.85308385e-01 4.13716674e-01 4.01738763e-01 1.18494129e+00
5.01540363e-01 -6.50649443e-02 -5.64248681e-01 1.19259965e+00
8.06446001e-02 7.78569818e-01 1.48370743e-01 -6.26531303e-01
8.11165690e-01 6.66328609e-01 2.83469334e-02 -1.04471242e+00
-5.25816560e-01 -7.67016470e-01 -9.93278980e-01 -3.42510432e-01
4.21917766e-01 -1.84816211e-01 -1.03181505e+00 1.72614670e+00
2.02587843e-01 1.07377879e-01 -4.69890028e-01 7.72794962e-01
8.42672050e-01 3.84017408e-01 2.93350458e-01 -1.24089330e-01
1.77615905e+00 -7.64919043e-01 -4.14124668e-01 -4.61859733e-01
4.85166818e-01 -5.35648942e-01 1.03102934e+00 1.97813347e-01
-1.25821316e+00 -5.52201450e-01 -9.77277637e-01 -4.12419617e-01
-1.30646721e-01 3.94740552e-01 6.06030166e-01 8.99129272e-01
-8.62079680e-01 3.70461345e-01 -6.78041875e-01 -1.52003780e-01
6.47519529e-01 8.06366265e-01 2.74266284e-02 -1.46805495e-01
-1.13496673e+00 6.63660884e-01 1.08005039e-01 3.87727767e-01
-4.71570611e-01 -9.32492137e-01 -7.78642654e-01 4.99850690e-01
4.65825856e-01 -6.73776150e-01 7.76935875e-01 -1.48285821e-01
-1.05172658e+00 5.51664770e-01 -5.72153986e-01 -5.23172319e-01
5.76257110e-01 -3.66266847e-01 -3.09306353e-01 -5.61038218e-02
1.74442858e-01 7.40232766e-01 6.24499261e-01 -7.39225149e-01
-9.46488142e-01 -7.50282764e-01 8.58387202e-02 3.83626789e-01
-9.64004636e-01 -8.39750841e-02 -1.07984984e+00 -7.01089263e-01
9.90873575e-02 -1.01265955e+00 3.54356878e-02 -2.58925438e-01
-5.20353317e-01 -2.79008508e-01 3.52098197e-01 -9.97526646e-01
1.58676910e+00 -1.91348028e+00 -1.27431020e-01 2.45796576e-01
4.44776803e-01 2.54782736e-01 2.18915835e-01 -2.24253729e-01
3.28213453e-01 8.10373500e-02 5.00854887e-02 -5.23167551e-01
-1.54790893e-01 -1.47472676e-02 -4.24498785e-03 2.59326130e-01
-2.00141206e-01 8.74072671e-01 -5.99161267e-01 -6.52986765e-01
2.92803019e-01 5.69985688e-01 -6.98250234e-01 -3.25502217e-01
2.70540595e-01 7.11873770e-02 -5.40077567e-01 7.21480429e-01
8.45937133e-01 -2.77748778e-02 4.85524088e-02 -6.04829073e-01
-1.30126655e-01 3.50084782e-01 -1.35497534e+00 1.31089866e+00
-3.24618310e-01 4.81051862e-01 -8.42184946e-02 -7.52370238e-01
7.52916276e-01 -3.12344823e-02 4.67618972e-01 -1.07282805e+00
1.74625844e-01 4.12211791e-02 -1.79838672e-01 -2.54724294e-01
6.59031332e-01 3.32765251e-01 -1.30992249e-01 1.52835876e-01
-2.17211470e-01 6.60872638e-01 1.80962607e-01 1.66098803e-01
8.54099333e-01 -1.11044899e-01 -1.92688510e-01 -4.64493036e-01
6.29347205e-01 -3.96970868e-01 7.40952492e-01 7.39271224e-01
-1.87693149e-01 3.59729320e-01 3.91823083e-01 -4.28051293e-01
-9.35683310e-01 -8.46783519e-01 -2.46094853e-01 1.07967710e+00
5.46702802e-01 -7.25278437e-01 -9.78691578e-01 -4.89161998e-01
1.09056361e-01 5.59958577e-01 -4.35298502e-01 -2.29301259e-01
-7.76523352e-01 -1.15920937e+00 8.22983384e-01 8.37519348e-01
1.03693271e+00 -8.23829830e-01 -6.74587488e-01 -2.31439844e-02
-5.27907550e-01 -1.19877625e+00 -5.09480238e-01 -3.41777392e-02
-7.63558269e-01 -6.47395968e-01 -9.35381353e-01 -6.01719618e-01
9.05002832e-01 2.73144096e-01 7.80242324e-01 1.32259414e-01
-3.58717978e-01 -1.15259318e-02 4.46833223e-02 -5.81180811e-01
4.97640282e-01 3.34457994e-01 2.22746491e-01 -1.58015177e-01
6.13850713e-01 -4.71058965e-01 -9.39080060e-01 4.65075403e-01
-2.11955234e-01 1.60239026e-01 5.23520231e-01 8.38912189e-01
7.36850679e-01 1.66921720e-01 5.03433347e-01 -5.20041764e-01
4.94899184e-01 9.74717587e-02 -5.66874504e-01 3.45247358e-01
-8.14880729e-01 9.13503692e-02 4.35801148e-01 -3.15609634e-01
-1.10523534e+00 -2.48280644e-01 -1.10792331e-01 -2.02083409e-01
2.61509836e-01 1.70800820e-01 -2.93255776e-01 5.78917526e-02
2.31256485e-01 4.12934154e-01 -5.55051804e-01 -6.65161908e-01
4.98875901e-02 6.86743796e-01 6.72075927e-01 -5.16181350e-01
4.92081493e-01 4.45571721e-01 -8.03329796e-03 -7.55709767e-01
-5.48305869e-01 -2.45615944e-01 -2.44423822e-01 -5.05058765e-02
7.69916594e-01 -1.08069503e+00 -1.26495075e+00 5.61881900e-01
-8.29782367e-01 5.45734949e-02 -1.57781020e-01 3.65502238e-01
1.06280763e-02 5.31803966e-01 -4.76887405e-01 -9.52696264e-01
-8.63186479e-01 -1.15330219e+00 9.12690222e-01 6.34053886e-01
-1.33004814e-01 -4.06902701e-01 -6.35189474e-01 7.45713711e-01
3.61954123e-01 -3.06022197e-01 9.34425116e-01 -2.77682483e-01
-6.15399480e-01 -2.81594813e-01 -7.55349696e-01 1.29523426e-01
-7.76835829e-02 -6.75988972e-01 -9.67065692e-01 -3.62418354e-01
-1.87761396e-01 1.63118809e-01 9.13137972e-01 5.51854551e-01
1.37398791e+00 -1.09960958e-01 -6.01182401e-01 7.58104503e-01
1.23704910e+00 1.51249483e-01 7.28670299e-01 2.75848895e-01
9.61981952e-01 4.05280679e-01 3.14176768e-01 6.54179037e-01
7.07271695e-01 8.50586236e-01 1.46076128e-01 -4.69991267e-02
-3.11458468e-01 -2.91328907e-01 9.26429033e-02 5.27497530e-01
-4.34170306e-01 -3.63439143e-01 -7.49847949e-01 6.95196271e-01
-1.70039403e+00 -7.58874297e-01 -7.92405531e-02 2.53760147e+00
5.88531613e-01 4.02395964e-01 -2.29906403e-02 1.89129919e-01
7.86945164e-01 -8.38393345e-02 -6.79147124e-01 -1.20279364e-01
-5.81335882e-03 3.44487339e-01 8.20987165e-01 3.17891598e-01
-9.72093046e-01 9.48065460e-01 4.39433479e+00 1.00508749e+00
-8.39885235e-01 1.23953462e-01 7.89634705e-01 -5.80626607e-01
4.13511246e-02 -1.51359931e-01 -1.42356384e+00 6.97180390e-01
7.33676910e-01 -1.23837434e-01 4.21303004e-01 7.55729556e-01
5.59647381e-02 -4.60874408e-01 -8.85167897e-01 1.54612982e+00
1.30632102e-01 -1.12492871e+00 1.02179863e-01 1.79162338e-01
3.60233098e-01 -1.28078654e-01 1.34446919e-01 3.61055076e-01
-1.70120150e-01 -8.50270450e-01 7.23948658e-01 3.59310567e-01
1.02920663e+00 -9.05457497e-01 8.19840193e-01 2.78413501e-02
-1.50515878e+00 -2.12676153e-01 -5.20669818e-01 -2.16788705e-02
2.97720820e-01 7.04517186e-01 -4.42300081e-01 5.16672432e-01
1.25432086e+00 2.71369964e-01 -6.55981600e-01 8.68781388e-01
-3.23158056e-02 3.62008572e-01 -6.03103340e-01 -3.98926847e-02
-2.82637984e-01 -2.89926324e-02 4.73028004e-01 1.10629320e+00
2.48160496e-01 2.72580892e-01 -9.72086415e-02 5.93906879e-01
-1.93312094e-01 7.48723093e-03 1.72876015e-01 4.27541137e-01
6.14317894e-01 1.03391075e+00 -7.51438379e-01 -5.25690436e-01
-3.10679078e-01 1.04076076e+00 2.94291675e-01 2.82952189e-01
-1.21603048e+00 -5.92366874e-01 6.47514999e-01 4.17356998e-01
4.39875782e-01 -1.68582182e-02 -7.48475075e-01 -1.08805013e+00
5.04762352e-01 -4.99399424e-01 4.22729731e-01 -4.46506381e-01
-8.03915918e-01 6.52969241e-01 1.57431573e-01 -8.29623640e-01
2.42533684e-02 -2.30561286e-01 -1.20068580e-01 1.08331001e+00
-1.30482626e+00 -1.05108631e+00 -8.16612720e-01 6.86438560e-01
5.10775030e-01 -3.99704538e-02 5.44851065e-01 9.58411396e-01
-9.22441900e-01 1.45309031e+00 -1.41587913e-01 1.83223054e-01
4.01981354e-01 -8.56160223e-01 6.43027246e-01 8.38391483e-01
-6.91525340e-02 7.10933268e-01 3.94040614e-01 -8.56414795e-01
-1.11258888e+00 -9.70223665e-01 1.05714643e+00 -2.00843036e-01
-3.57428640e-02 -6.84800565e-01 -5.82231522e-01 5.14227033e-01
-4.21001911e-01 -7.37160966e-02 2.89928824e-01 2.95732319e-01
-8.03759173e-02 -5.36940098e-01 -1.11212051e+00 9.13724065e-01
1.43667257e+00 -3.90815139e-01 -8.08735713e-02 2.38937467e-01
3.51802409e-01 -3.87963384e-01 -9.21241939e-01 5.58267236e-01
8.07761610e-01 -8.53581548e-01 1.12505114e+00 -5.87998331e-02
-1.53512061e-01 -9.30946469e-02 -1.38987340e-02 -5.55674672e-01
-5.33313096e-01 -5.77565849e-01 -2.09547773e-01 1.38230383e+00
3.42491031e-01 -6.63261056e-01 9.88617301e-01 1.09473503e+00
-1.39442412e-02 -8.77880871e-01 -1.20542419e+00 -4.56185400e-01
-3.90200913e-01 -4.56998587e-01 6.14377737e-01 3.51623118e-01
-1.70042440e-01 3.98988634e-01 -6.00458503e-01 1.45067140e-01
7.30287611e-01 -3.27545136e-01 5.78914881e-01 -1.16915071e+00
-2.55357623e-01 -2.14549333e-01 -4.30796802e-01 -1.28527892e+00
-4.62267220e-01 -6.57415032e-01 -2.55893886e-01 -1.58155632e+00
5.04750252e-01 -8.27310145e-01 -3.41581792e-01 5.91574490e-01
-4.62272197e-01 2.02668667e-01 3.27576399e-01 2.54499167e-01
-3.97311270e-01 5.07895112e-01 7.65691936e-01 -1.66921109e-01
-3.33022684e-01 -5.29043637e-02 -8.09385061e-01 6.65431678e-01
6.23296618e-01 -4.60067749e-01 -6.04766965e-01 -8.73631656e-01
1.74922869e-01 -1.96872279e-01 4.36382264e-01 -1.28633940e+00
4.53256547e-01 3.34957510e-01 7.64417827e-01 -7.65937567e-01
6.33890271e-01 -4.98200417e-01 6.33740658e-03 4.03698444e-01
4.63878550e-02 1.82573915e-01 4.62369114e-01 5.17734826e-01
2.32710168e-01 -8.58719721e-02 5.42556286e-01 4.13930528e-02
-5.97009778e-01 3.76600891e-01 6.13703988e-02 -1.11491099e-01
9.40030336e-01 -5.67319453e-01 -3.78018856e-01 -2.88371384e-01
-6.09823704e-01 3.94149363e-01 1.21647805e-01 3.37141782e-01
5.13851047e-01 -9.88920152e-01 -6.76487505e-01 3.37064415e-01
-1.11858264e-01 5.90617023e-02 8.20655882e-01 9.58934963e-01
-3.14005315e-01 6.19208336e-01 -6.58096522e-02 -3.50389063e-01
-1.38566852e+00 3.25635344e-01 3.72321814e-01 -4.37612593e-01
-8.16786528e-01 1.19333792e+00 3.13571960e-01 1.89155594e-01
4.17306811e-01 -2.86699712e-01 -1.16406485e-01 7.14961365e-02
6.79295301e-01 8.65407348e-01 2.72478968e-01 -5.79232812e-01
-5.48666537e-01 6.42051280e-01 -5.18185139e-01 -4.82207611e-02
1.02591312e+00 -3.08783770e-01 1.16196662e-01 -2.79171586e-01
8.54100704e-01 -4.13598046e-02 -1.12615860e+00 -1.11105695e-01
-2.29628623e-01 -4.31502074e-01 2.36746684e-01 -8.45736206e-01
-1.19789946e+00 7.77321637e-01 9.76292431e-01 -2.43317559e-01
1.26428950e+00 -1.35466561e-01 1.25722241e+00 1.97546884e-01
4.01521921e-01 -1.28317952e+00 -3.34013730e-01 1.97043687e-01
4.39863980e-01 -8.39257360e-01 4.57327694e-01 -6.01324975e-01
-5.15039742e-01 5.72259784e-01 7.81407177e-01 2.43235171e-01
5.38995624e-01 1.32933915e-01 -5.58130562e-01 -1.30483851e-01
-2.90384382e-01 -2.47432724e-01 2.60603040e-01 4.97294277e-01
1.00194559e-01 1.90524440e-02 -4.83745486e-01 9.50512290e-01
-2.68782258e-01 2.96704359e-02 -9.23551898e-03 5.86557984e-01
-1.88167036e-01 -8.67741704e-01 -1.88132882e-01 8.92546296e-01
-5.80883324e-01 -4.20591235e-01 6.88853785e-02 5.72008908e-01
4.81306523e-01 9.20186102e-01 1.16011985e-01 -5.05243659e-01
4.72733259e-01 -9.28409174e-02 5.00390530e-01 -1.75619572e-01
-6.41021609e-01 1.24551482e-01 3.73592019e-01 -4.04132694e-01
4.85178791e-02 -5.52021980e-01 -1.18000841e+00 -4.70525742e-01
-3.61730576e-01 -8.86682644e-02 4.48521078e-01 8.28714967e-01
6.58651650e-01 6.38688862e-01 5.56405857e-02 -5.24731517e-01
-4.14875150e-01 -9.77683663e-01 -3.26609135e-01 2.76792556e-01
-2.29207948e-01 -8.27998817e-01 1.26598533e-02 -1.22287832e-01] | [14.68148422241211, 0.7670210599899292] |
99bf8787-d100-464b-abe0-4100fffb7355 | points-as-queries-weakly-semi-supervised | 2104.07434 | null | https://arxiv.org/abs/2104.07434v1 | https://arxiv.org/pdf/2104.07434v1.pdf | Points as Queries: Weakly Semi-supervised Object Detection by Points | We propose a novel point annotated setting for the weakly semi-supervised object detection task, in which the dataset comprises small fully annotated images and large weakly annotated images by points. It achieves a balance between tremendous annotation burden and detection performance. Based on this setting, we analyze existing detectors and find that these detectors have difficulty in fully exploiting the power of the annotated points. To solve this, we introduce a new detector, Point DETR, which extends DETR by adding a point encoder. Extensive experiments conducted on MS-COCO dataset in various data settings show the effectiveness of our method. In particular, when using 20% fully labeled data from COCO, our detector achieves a promising performance, 33.3 AP, which outperforms a strong baseline (FCOS) by 2.0 AP, and we demonstrate the point annotations bring over 10 points in various AR metrics. | ['Jian Sun', 'Wei zhang', 'Xiangyu Zhang', 'Tong Yang', 'Liangyu Chen'] | 2021-04-15 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Chen_Points_As_Queries_Weakly_Semi-Supervised_Object_Detection_by_Points_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Chen_Points_As_Queries_Weakly_Semi-Supervised_Object_Detection_by_Points_CVPR_2021_paper.pdf | cvpr-2021-1 | ['semi-supervised-object-detection'] | ['computer-vision'] | [-1.24225251e-01 2.76961774e-01 -4.01858389e-01 -1.09136447e-01
-1.27558780e+00 -6.55266941e-01 3.72890294e-01 -1.51308596e-01
-5.15082061e-01 3.39826554e-01 -7.63694048e-02 2.85898864e-01
5.60117781e-01 -1.94838241e-01 -1.14565480e+00 -5.41918337e-01
1.18883088e-01 3.74777198e-01 8.50171447e-01 -2.68150065e-02
4.16704454e-02 3.35442960e-01 -1.26324856e+00 5.98789491e-02
4.97299492e-01 1.15752697e+00 4.20852482e-01 5.24690926e-01
1.77092254e-01 5.70257485e-01 -4.55942124e-01 -5.73120654e-01
6.23550177e-01 2.45158926e-01 -5.79710603e-01 2.32748255e-01
8.85972142e-01 -7.44125962e-01 -3.50269020e-01 1.15723932e+00
5.66803813e-01 -1.88600361e-01 4.11143601e-01 -1.25089276e+00
-6.05767906e-01 7.41706610e-01 -1.15633440e+00 9.93596017e-02
2.72352576e-01 1.76997215e-01 1.20333207e+00 -1.68489289e+00
4.35084671e-01 1.21012986e+00 1.00439167e+00 5.43682218e-01
-1.08038580e+00 -7.32522547e-01 9.49542373e-02 -2.90668935e-01
-1.70688581e+00 -4.73731101e-01 5.59848249e-01 -3.41458529e-01
6.45184994e-01 -6.88266680e-02 2.35868141e-01 1.10270905e+00
-3.94442141e-01 1.26249337e+00 5.76283813e-01 -4.22665805e-01
1.07646100e-01 1.69270143e-01 -1.58618577e-02 7.93874145e-01
2.19325796e-01 1.29182413e-01 -4.18528050e-01 -6.68430999e-02
9.65527117e-01 1.05119392e-01 -8.27620998e-02 -5.45243144e-01
-1.39722919e+00 4.90956068e-01 7.79689550e-01 -2.22665384e-01
-5.40121198e-02 3.36921573e-01 3.80258769e-01 -3.06420535e-01
3.29566658e-01 2.63265610e-01 -4.80498075e-01 1.12383485e-01
-8.64341557e-01 4.79878709e-02 2.99085230e-01 1.74247158e+00
5.19814432e-01 -2.27111280e-01 -3.92631799e-01 9.09566224e-01
3.54521692e-01 8.35049272e-01 7.56271109e-02 -1.01612484e+00
5.82298875e-01 6.98241949e-01 3.86824936e-01 -6.28463209e-01
-1.72539502e-01 -5.44426143e-01 -4.96777654e-01 6.41968623e-02
3.30494165e-01 -1.81310505e-01 -1.06171739e+00 1.51620591e+00
3.83320034e-01 2.51180649e-01 -2.04246208e-01 9.81984138e-01
8.67703378e-01 4.48504955e-01 9.45756808e-02 1.55622587e-01
1.36515856e+00 -1.45871949e+00 -5.04253924e-01 -3.23491037e-01
6.32642686e-01 -6.20914519e-01 1.30010796e+00 1.08456187e-01
-1.19838142e+00 -8.17221880e-01 -1.19059598e+00 -3.07308137e-01
6.78846985e-02 7.67752528e-01 4.88580316e-01 2.05724806e-01
-8.61567140e-01 1.80112600e-01 -1.00042343e+00 -3.29374433e-01
9.79231179e-01 3.44433963e-01 -4.20226932e-01 -8.95455829e-04
-4.43566650e-01 4.49557513e-01 3.94664109e-01 7.07997307e-02
-1.03726673e+00 -7.71669328e-01 -8.22674572e-01 7.38451406e-02
7.22539067e-01 -5.14009535e-01 1.42427993e+00 -5.02923548e-01
-9.50923979e-01 1.07658529e+00 -1.95537299e-01 -6.10270739e-01
8.99632990e-01 -8.26273739e-01 -8.65588710e-02 3.26703370e-01
6.79752350e-01 1.18979895e+00 6.60389721e-01 -1.32519960e+00
-1.13952792e+00 -3.80185068e-01 1.21866629e-01 1.84477791e-01
-3.34549397e-01 1.10507056e-01 -1.25268197e+00 -5.63998282e-01
7.72889704e-02 -1.11362267e+00 -3.17616045e-01 5.56885123e-01
-5.68956077e-01 -5.14549732e-01 8.37766051e-01 1.02372458e-02
7.63172090e-01 -2.45972252e+00 -1.99128136e-01 -2.12712437e-01
4.60221082e-01 1.98540226e-01 -7.85870180e-02 3.57063301e-02
2.23792657e-01 -6.80043772e-02 -7.90325925e-02 -8.88349116e-01
-8.46637487e-02 1.46503821e-01 -3.92733783e-01 4.41838384e-01
4.86801863e-01 1.25861537e+00 -9.71541822e-01 -9.60887492e-01
2.38768592e-01 2.22408563e-01 -5.22414207e-01 1.84497118e-01
-5.21035865e-02 1.27457336e-01 -3.90251607e-01 1.13549829e+00
7.74716854e-01 -6.54409528e-01 -4.31680471e-01 -4.25117403e-01
2.57428866e-02 -3.91247243e-01 -1.06549537e+00 1.95041215e+00
9.30058435e-02 5.39148986e-01 -7.18270540e-02 -4.79834437e-01
8.10194492e-01 7.56272823e-02 5.08771837e-01 -2.76321024e-01
3.18234116e-02 2.41444871e-01 -3.23127478e-01 -2.17420846e-01
4.40435886e-01 1.42712817e-01 4.58286032e-02 -5.19870035e-02
3.04332554e-01 7.53003582e-02 5.60627542e-02 5.13885677e-01
1.18917763e+00 3.42527539e-01 9.86074880e-02 -1.06527731e-01
3.13947886e-01 9.42246988e-02 6.23609126e-01 9.95324433e-01
-6.21588767e-01 1.20774090e+00 2.43387178e-01 -2.83856541e-01
-1.22525072e+00 -1.17609453e+00 -3.95812362e-01 1.12642765e+00
5.13587892e-01 -5.72090745e-01 -6.31845415e-01 -9.40770686e-01
-1.63216721e-02 9.83792692e-02 -6.33503675e-01 1.95352659e-02
-5.57721615e-01 -4.73931193e-01 6.78068042e-01 1.21750212e+00
6.42096996e-01 -7.90659726e-01 -2.94151872e-01 -2.23434314e-01
-1.69554844e-01 -1.76928926e+00 -7.82673657e-01 8.59447569e-02
-6.80440843e-01 -1.00534093e+00 -9.46180403e-01 -1.01961052e+00
7.10562348e-01 6.89753950e-01 1.02156401e+00 1.42160878e-02
-1.81139767e-01 2.59577751e-01 -3.59782100e-01 -5.89511693e-01
1.33649766e-01 1.44453958e-01 2.97097087e-01 -2.07191065e-01
4.94079500e-01 -7.99889565e-02 -7.59463668e-01 6.19587481e-01
-5.76460302e-01 -1.84298560e-01 8.90093088e-01 7.93249547e-01
8.99748683e-01 -4.88190681e-01 3.91670853e-01 -7.87250459e-01
-9.21242535e-02 -2.21593410e-01 -6.82021558e-01 6.42556846e-02
-3.18720937e-01 -1.52128726e-01 1.34994507e-01 -4.18166995e-01
-8.02751958e-01 8.38728905e-01 -1.72424555e-01 -9.44703996e-01
2.71384101e-02 -2.49963492e-01 -1.90422550e-01 -3.61598223e-01
8.38835716e-01 -3.86487739e-03 -2.54122347e-01 -4.65932131e-01
6.30921662e-01 6.60688281e-01 1.06969297e+00 -5.81701040e-01
1.10869968e+00 9.59546983e-01 -2.14932576e-01 -5.85034549e-01
-1.33508980e+00 -9.50166583e-01 -8.68691444e-01 -4.77748290e-02
8.74348283e-01 -1.50943708e+00 -5.90037942e-01 1.08969621e-01
-1.18521678e+00 -8.28052759e-02 -5.40630400e-01 4.65142101e-01
-6.28840864e-01 5.23519933e-01 -7.14444458e-01 -7.48850286e-01
-2.82139838e-01 -1.07725000e+00 1.73689771e+00 3.47346105e-02
4.08843122e-02 -4.75822538e-01 -1.01537779e-01 3.84513974e-01
-1.47106156e-01 1.67378291e-01 -9.79410112e-02 -4.95899558e-01
-6.65310204e-01 -4.94951904e-01 -7.28592157e-01 4.23244119e-01
-2.11464062e-01 -2.01283738e-01 -1.03813171e+00 -3.68928939e-01
-2.68754810e-01 -6.84832215e-01 9.04585481e-01 1.30703837e-01
1.24202609e+00 1.84663951e-01 -6.77267551e-01 7.53703892e-01
1.26340961e+00 -2.68671125e-01 4.07715678e-01 2.08904356e-01
1.07603478e+00 1.69626892e-01 8.47392142e-01 2.41915464e-01
2.42153123e-01 9.04726446e-01 5.49324036e-01 -4.35974211e-01
-3.31972748e-01 -7.56820738e-01 3.68566990e-01 2.61974573e-01
-3.11026603e-01 3.14643867e-02 -8.05318058e-01 6.29418015e-01
-2.07025933e+00 -6.98580027e-01 -2.55775928e-01 1.82990742e+00
6.66264713e-01 5.08153796e-01 3.69083881e-01 -9.79449823e-02
7.71137476e-01 -1.16953038e-01 -6.44341290e-01 6.47440076e-01
-2.28111088e-01 -7.50612468e-02 9.86738682e-01 1.36647671e-01
-1.37199306e+00 1.15811205e+00 7.37140226e+00 8.80451560e-01
-6.22685432e-01 1.73231140e-01 3.60896796e-01 -2.24213317e-01
3.87384117e-01 -1.33955061e-01 -1.14910233e+00 5.01362443e-01
3.52945358e-01 1.22230060e-01 -3.12438995e-01 1.47489536e+00
1.18878275e-01 2.90236086e-01 -1.45659554e+00 1.46228838e+00
1.32872015e-01 -1.22090304e+00 -6.05201721e-02 1.00864232e-01
7.97169447e-01 4.73491371e-01 4.67354320e-02 4.27854151e-01
2.29452983e-01 -8.21102202e-01 9.84678924e-01 4.96958196e-02
9.12571192e-01 -4.06127363e-01 8.83496106e-01 4.02262032e-01
-1.29254079e+00 -1.42709622e-02 -6.94829524e-01 1.42486244e-01
3.48795295e-01 4.26110595e-01 -7.32769608e-01 2.03499645e-01
9.43637908e-01 8.41910064e-01 -9.25797403e-01 1.23066831e+00
-2.98812151e-01 6.13665044e-01 -5.75589478e-01 2.48189345e-01
2.11163059e-01 3.01596463e-01 4.52065498e-01 1.28518319e+00
1.71654165e-01 1.21670410e-01 4.37165588e-01 9.10895586e-01
-4.02549654e-01 -2.55762011e-01 -4.20805991e-01 4.11091596e-01
6.60741329e-01 1.38940895e+00 -6.42505288e-01 -2.92818815e-01
-5.73870182e-01 1.08059943e+00 5.34171641e-01 1.82041034e-01
-1.08160830e+00 -1.74770415e-01 2.09651187e-01 1.29541218e-01
4.76066142e-01 -1.98295012e-01 -1.59555987e-01 -1.35408735e+00
4.85793173e-01 -4.12215918e-01 1.92808419e-01 -9.15228903e-01
-1.34629822e+00 4.93983418e-01 -5.87264337e-02 -1.49739587e+00
1.84172988e-01 -6.90356791e-01 -3.88855159e-01 3.44033211e-01
-1.26359427e+00 -1.52757108e+00 -4.88677263e-01 4.67281967e-01
7.42445648e-01 1.43684754e-02 2.93754488e-01 5.89080215e-01
-6.11432374e-01 9.06953692e-01 -2.08737791e-01 6.01092160e-01
8.23426425e-01 -1.44668877e+00 6.05597258e-01 8.80958736e-01
5.48217595e-01 4.83846098e-01 3.29439700e-01 -3.97916168e-01
-1.30943632e+00 -1.18773031e+00 4.13742512e-01 -1.14969969e+00
5.31631112e-01 -7.01724410e-01 -8.19690287e-01 9.79696929e-01
-1.73548236e-01 8.38743329e-01 3.79160136e-01 -3.06399167e-02
-4.49766159e-01 1.34158656e-01 -1.00064123e+00 4.67537373e-01
1.59653950e+00 -3.84433776e-01 -6.29917085e-01 5.82608521e-01
1.17655444e+00 -7.05237269e-01 -6.09121442e-01 6.81365013e-01
3.86242539e-01 -7.20991790e-01 1.26244342e+00 -5.07343233e-01
3.86497229e-01 -6.15371764e-01 -1.93166018e-01 -7.83039987e-01
-3.84415597e-01 -3.97407174e-01 -4.91737336e-01 1.26341319e+00
3.91970873e-01 -9.06105638e-02 1.22261477e+00 3.92074227e-01
-2.28004858e-01 -7.23741591e-01 -6.65722311e-01 -9.51593637e-01
-1.96248040e-01 -5.13668120e-01 3.34358245e-01 6.45820379e-01
-1.92122310e-01 6.82644010e-01 -4.39371854e-01 3.80414695e-01
9.21554506e-01 -1.33441433e-01 1.09075117e+00 -1.00830328e+00
-3.01082343e-01 -8.64350721e-02 -6.93909943e-01 -1.77331424e+00
-1.09557517e-01 -6.15463793e-01 2.58722425e-01 -1.24153936e+00
5.02639532e-01 -5.05109668e-01 -2.25295842e-01 5.54363012e-01
-4.34264779e-01 8.10764909e-01 2.80448943e-01 7.00107813e-01
-1.23234224e+00 5.43930054e-01 9.85463679e-01 -9.69582945e-02
-3.29400264e-02 8.43336992e-03 -6.70294046e-01 1.10786796e+00
3.36667091e-01 -3.38990957e-01 -1.32832199e-01 -6.39524519e-01
-5.24869096e-03 -3.43467176e-01 5.70372045e-01 -1.23725176e+00
3.97084713e-01 2.71011919e-01 5.74777126e-01 -1.09791327e+00
4.57675219e-01 -8.59916925e-01 -4.58406150e-01 2.94615477e-01
-3.00150961e-01 -1.32919446e-01 8.88465792e-02 1.01961219e+00
-2.06107289e-01 -3.83595750e-02 8.21127772e-01 5.22196144e-02
-9.54684675e-01 6.41140580e-01 6.76458120e-01 1.08324416e-01
1.35365689e+00 -1.32998824e-01 -2.68758506e-01 -2.02913899e-02
-7.45965660e-01 7.01802969e-01 4.80745018e-01 5.02863884e-01
5.96971631e-01 -1.68058968e+00 -7.60718584e-01 -4.81705405e-02
7.47300088e-01 6.68144166e-01 -1.39417976e-01 7.16309130e-01
-5.72423041e-01 2.59495258e-01 1.99686676e-01 -1.14573443e+00
-1.07401800e+00 6.66950881e-01 1.26322076e-01 2.67850667e-01
-9.81628060e-01 9.53019083e-01 5.54632425e-01 -1.31075606e-01
5.66734433e-01 -3.48784775e-01 2.69481577e-02 -2.81973988e-01
6.21818721e-01 2.91027337e-01 -5.67927323e-02 -7.24142671e-01
-5.53908408e-01 8.04244637e-01 -2.63285041e-01 -6.15562387e-02
1.14523780e+00 -1.02660269e-01 4.74838078e-01 3.14130127e-01
1.14393604e+00 2.47328311e-01 -1.74875760e+00 -4.21452135e-01
3.91578041e-02 -6.19115710e-01 -1.35188326e-01 -4.00710613e-01
-8.98909807e-01 6.60902679e-01 6.29514813e-01 -1.08651236e-01
7.06941366e-01 6.70938432e-01 7.83176005e-01 4.55617696e-01
4.89865184e-01 -9.62679148e-01 4.27660197e-01 2.65256643e-01
8.57098341e-01 -1.74993753e+00 -5.09096310e-02 -7.92494953e-01
-7.28654146e-01 7.55007565e-01 1.00493002e+00 -4.23541665e-01
4.18408871e-01 3.17163855e-01 -1.16165699e-02 -3.14100087e-01
-4.70258802e-01 -4.97070312e-01 2.98560172e-01 6.12463772e-01
1.71588272e-01 -7.30176941e-02 -5.87517349e-03 4.81296271e-01
-5.47929369e-02 3.72963287e-02 2.29071945e-01 7.02293277e-01
-5.31158745e-01 -6.54532850e-01 -3.79131407e-01 2.27827713e-01
-5.06353974e-01 6.98264390e-02 -5.32332301e-01 8.79892647e-01
8.19504783e-02 6.80803359e-01 1.20872967e-02 -3.34400028e-01
5.85819602e-01 -3.67024451e-01 2.08741993e-01 -8.49750578e-01
-1.58722118e-01 2.54021823e-01 -3.11046481e-01 -7.69219577e-01
-5.17079175e-01 -5.20143330e-01 -1.31738162e+00 1.95308045e-01
-7.95297861e-01 7.38248415e-03 3.77213776e-01 6.68941259e-01
4.00450498e-01 3.34268659e-01 4.35781687e-01 -9.86210585e-01
-9.07845616e-01 -1.04435515e+00 -4.27152753e-01 5.76416731e-01
4.09232646e-01 -8.21093976e-01 -3.70242536e-01 7.10735321e-02] | [9.295876502990723, 1.142386555671692] |
f6107531-43f5-4abc-90f4-1cd7dea434d1 | visibility-aware-human-object-interaction | 2303.16479 | null | https://arxiv.org/abs/2303.16479v1 | https://arxiv.org/pdf/2303.16479v1.pdf | Visibility Aware Human-Object Interaction Tracking from Single RGB Camera | Capturing the interactions between humans and their environment in 3D is important for many applications in robotics, graphics, and vision. Recent works to reconstruct the 3D human and object from a single RGB image do not have consistent relative translation across frames because they assume a fixed depth. Moreover, their performance drops significantly when the object is occluded. In this work, we propose a novel method to track the 3D human, object, contacts between them, and their relative translation across frames from a single RGB camera, while being robust to heavy occlusions. Our method is built on two key insights. First, we condition our neural field reconstructions for human and object on per-frame SMPL model estimates obtained by pre-fitting SMPL to a video sequence. This improves neural reconstruction accuracy and produces coherent relative translation across frames. Second, human and object motion from visible frames provides valuable information to infer the occluded object. We propose a novel transformer-based neural network that explicitly uses object visibility and human motion to leverage neighbouring frames to make predictions for the occluded frames. Building on these insights, our method is able to track both human and object robustly even under occlusions. Experiments on two datasets show that our method significantly improves over the state-of-the-art methods. Our code and pretrained models are available at: https://virtualhumans.mpi-inf.mpg.de/VisTracker | ['Gerard Pons-Moll', 'Bharat Lal Bhatnagar', 'Xianghui Xie'] | 2023-03-29 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xie_Visibility_Aware_Human-Object_Interaction_Tracking_From_Single_RGB_Camera_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xie_Visibility_Aware_Human-Object_Interaction_Tracking_From_Single_RGB_Camera_CVPR_2023_paper.pdf | cvpr-2023-1 | ['human-object-interaction-detection'] | ['computer-vision'] | [ 3.09694782e-02 -1.70495555e-01 3.82499173e-02 -2.36706212e-01
-2.96165258e-01 -4.02712584e-01 5.40311873e-01 -4.04529989e-01
-3.05911481e-01 3.35459888e-01 9.84947290e-03 2.35848084e-01
4.51138943e-01 -4.86056685e-01 -1.13940525e+00 -4.75073576e-01
2.18200669e-01 5.91069102e-01 6.72691047e-01 1.18908331e-01
-1.37259334e-01 5.78336716e-01 -1.66129363e+00 2.21375763e-01
2.70473421e-01 9.83078003e-01 5.95051944e-01 7.48360336e-01
2.03403130e-01 9.70797122e-01 -1.84814259e-01 -1.37941867e-01
4.69905108e-01 -2.77635068e-01 -7.05868304e-01 4.33921695e-01
6.75638020e-01 -8.23799968e-01 -7.55312443e-01 6.86769605e-01
3.20899010e-01 3.15804601e-01 2.32537493e-01 -1.15732098e+00
-2.40210310e-01 4.94851284e-02 -7.06409514e-01 -1.19261220e-01
7.69460678e-01 3.83249313e-01 5.82630634e-01 -8.79750907e-01
1.08261907e+00 1.60318124e+00 7.79956460e-01 5.65509856e-01
-1.14419174e+00 -3.65382075e-01 3.32464844e-01 3.66326869e-01
-1.26226592e+00 -6.04618013e-01 7.71270633e-01 -6.23746157e-01
8.81516874e-01 7.06549585e-02 1.12750828e+00 1.20669138e+00
-2.04548123e-03 8.94804835e-01 7.87503004e-01 -3.72777104e-01
-1.34767068e-03 -1.15696304e-01 -6.56993389e-02 8.61281753e-01
1.87001273e-01 2.79040813e-01 -7.47032464e-01 1.86590388e-01
1.30142784e+00 2.51610756e-01 -6.22905374e-01 -8.93921554e-01
-1.54485571e+00 2.93256581e-01 5.63896418e-01 -2.68322267e-02
-5.04291892e-01 4.85353380e-01 -1.20760061e-01 -2.23950252e-01
3.95300895e-01 -2.51305819e-01 -4.81297135e-01 -5.95739409e-02
-6.61879361e-01 2.63899803e-01 6.21816039e-01 1.03744304e+00
7.22749293e-01 -8.45574811e-02 1.37328625e-01 3.19457591e-01
5.68762243e-01 6.80466950e-01 6.32295609e-02 -1.43948400e+00
2.10106879e-01 4.24733281e-01 3.35579127e-01 -1.07648909e+00
-5.17138839e-01 -5.22762276e-02 -4.85478729e-01 3.56099188e-01
5.01963735e-01 1.84991628e-01 -1.07990277e+00 1.53990662e+00
8.38975847e-01 6.46377802e-01 -2.17530638e-01 1.21264040e+00
7.74727523e-01 5.76308250e-01 -2.41668731e-01 -6.71047717e-03
1.28058708e+00 -1.12982666e+00 -6.35752082e-01 -6.26239061e-01
2.34566867e-01 -5.78127861e-01 6.85324311e-01 3.16844016e-01
-1.34298360e+00 -7.35788882e-01 -8.43077660e-01 -5.06687760e-01
-1.77007504e-02 5.70709892e-02 6.07665062e-01 1.97632372e-01
-1.07740939e+00 5.63470423e-01 -1.49145198e+00 -5.65607846e-01
2.92468697e-01 2.50626326e-01 -6.12993360e-01 -3.67673248e-01
-6.37098849e-01 1.02667618e+00 2.13573843e-01 4.12053764e-01
-9.44662690e-01 -5.39878011e-01 -1.04885495e+00 -3.12007725e-01
2.93285161e-01 -1.18713164e+00 1.37459445e+00 -7.41420090e-01
-1.60460675e+00 9.55708265e-01 -5.35250366e-01 -3.99488032e-01
8.75580072e-01 -5.59758186e-01 1.54426783e-01 4.10935193e-01
-8.23020563e-02 9.05121148e-01 7.42225289e-01 -1.44905150e+00
-6.59520209e-01 -6.27011359e-01 4.31255028e-02 3.34771872e-01
2.55066276e-01 -1.85122773e-01 -9.85242844e-01 -2.64139861e-01
4.39751714e-01 -1.04403949e+00 -9.31599066e-02 6.54342651e-01
-3.00328434e-01 1.42579868e-01 8.24173212e-01 -6.33396864e-01
4.47548598e-01 -1.78810966e+00 6.19393587e-01 -2.57265598e-01
2.82371521e-01 1.48789555e-01 7.75683001e-02 4.37143892e-02
2.70308524e-01 -5.85394740e-01 -8.99969190e-02 -8.58667314e-01
-1.31491432e-02 4.46055502e-01 -2.26722285e-01 9.29718137e-01
1.28753306e-02 1.02381158e+00 -8.63154888e-01 -4.30824131e-01
6.57006383e-01 1.18555820e+00 -5.88378847e-01 4.16600168e-01
-2.27734581e-01 9.81408894e-01 -1.65358812e-01 7.47598886e-01
6.55960202e-01 -3.15753341e-01 9.94108617e-02 -3.74613374e-01
-7.91921839e-02 1.31018922e-01 -1.21867692e+00 2.12687612e+00
-1.94255471e-01 7.47403085e-01 1.97436824e-01 -7.57112026e-01
3.71059150e-01 3.12389433e-01 6.44043207e-01 -4.14862335e-01
3.18495035e-01 9.11360085e-02 -4.37324524e-01 -5.96384823e-01
3.23630422e-01 1.62299737e-01 3.95141065e-01 2.52455860e-01
4.89783362e-02 -1.13527536e-01 2.79605296e-02 -6.78717904e-03
9.99252200e-01 8.90169144e-01 1.47741705e-01 3.21737200e-01
2.85123229e-01 -1.71741679e-01 6.27395570e-01 4.91799682e-01
-2.85533637e-01 8.70982826e-01 3.14987823e-02 -6.99710786e-01
-9.64634776e-01 -1.12395978e+00 1.63790900e-02 7.93082416e-01
5.83909035e-01 -1.60898641e-01 -6.25830770e-01 -2.50242800e-01
7.95686245e-02 3.21645319e-01 -6.10876679e-01 1.06487773e-01
-8.08002174e-01 -1.14143334e-01 5.10585271e-02 5.67364872e-01
3.71230304e-01 -9.73729610e-01 -1.23813355e+00 8.19420815e-02
-6.01366699e-01 -1.69517815e+00 -2.52414495e-01 -2.79768053e-02
-8.35896373e-01 -1.23302162e+00 -7.52178848e-01 -6.34941518e-01
6.40840828e-01 6.62549317e-01 1.01254702e+00 9.14577045e-04
-3.43004405e-01 7.45595396e-01 -6.04095049e-02 -1.91045608e-02
-2.30863057e-02 -3.82537097e-01 2.78901696e-01 1.18948715e-02
1.07592076e-01 -6.16282821e-01 -7.65063465e-01 5.03382742e-01
-6.49588346e-01 4.07310158e-01 1.66523829e-01 4.40827399e-01
7.68272638e-01 -3.60091686e-01 -4.03401494e-01 -4.60608572e-01
-4.20340329e-01 -2.57914007e-01 -7.49456763e-01 1.18561529e-01
1.17510214e-01 -1.10928416e-01 4.10117544e-02 -5.43291628e-01
-1.05216277e+00 6.85768425e-01 1.50396749e-01 -9.30739939e-01
-2.17858404e-01 -8.21453556e-02 -1.53006420e-01 -8.20812583e-02
4.43655401e-01 -5.87985385e-03 -5.85915521e-02 -5.46530187e-01
4.07797277e-01 1.50804251e-01 1.01613927e+00 -4.30140793e-01
7.98968792e-01 1.03728676e+00 -3.58888060e-02 -7.99337506e-01
-8.24093401e-01 -5.59178650e-01 -1.13118219e+00 -4.18064922e-01
1.01862395e+00 -1.11732948e+00 -9.90906596e-01 6.28739655e-01
-1.56101203e+00 -6.21168673e-01 -1.13345027e-01 7.02078700e-01
-8.05293560e-01 3.69881779e-01 -5.51916718e-01 -8.75353158e-01
-3.78603786e-02 -1.28031611e+00 1.46852267e+00 1.13212556e-01
-1.43045232e-01 -8.79342616e-01 -3.74008119e-02 4.30389106e-01
-4.68491716e-03 5.88402629e-01 -7.58719118e-03 2.35895351e-01
-1.08404922e+00 -9.23194587e-02 -5.44587374e-02 -9.51326564e-02
3.63727994e-02 -5.29227667e-02 -1.03694212e+00 -2.20013604e-01
4.99787144e-02 -2.43045464e-01 8.56748223e-01 6.01810157e-01
6.98950887e-01 -1.68379858e-01 -6.14748657e-01 1.01628113e+00
1.14354098e+00 -1.34197369e-01 4.77295667e-01 4.20563430e-01
1.10124373e+00 7.05298245e-01 5.68513691e-01 3.76317948e-01
6.67052805e-01 1.03643179e+00 6.94701374e-01 1.32022113e-01
-4.59596217e-01 -2.58813739e-01 6.42818034e-01 6.31172419e-01
-2.97936052e-01 -3.36150676e-02 -9.45135295e-01 4.12787884e-01
-2.07812452e+00 -8.94065619e-01 -2.72274822e-01 2.13531160e+00
5.54106295e-01 -8.46903622e-02 4.04336937e-02 -8.71697515e-02
5.60540080e-01 -1.34274922e-02 -7.61833310e-01 4.37637717e-01
-1.03713319e-01 -3.64014506e-02 3.97913069e-01 7.28641033e-01
-9.30103779e-01 1.13324535e+00 5.70156384e+00 -2.49052960e-02
-1.04731941e+00 2.27945924e-01 2.48649414e-03 -4.87403482e-01
1.98235482e-01 6.76243752e-02 -9.52897429e-01 2.12311164e-01
5.28871477e-01 1.98252589e-01 5.43233871e-01 6.01786911e-01
2.10842565e-01 -2.89188981e-01 -1.41655135e+00 1.18273306e+00
1.88787967e-01 -1.10956550e+00 -3.03935677e-01 1.24438055e-01
6.42046034e-01 3.15630138e-01 -9.73268226e-02 -1.55767232e-01
2.41602555e-01 -7.29451239e-01 1.29185569e+00 7.66521096e-01
4.16547090e-01 -3.07329863e-01 4.24886644e-01 4.60143715e-01
-1.21922278e+00 1.66424900e-01 -3.28190923e-01 -2.60163218e-01
5.75433910e-01 4.04040754e-01 -6.75249040e-01 3.39806736e-01
9.59786952e-01 1.01921320e+00 -4.47069883e-01 9.26144481e-01
-3.70909989e-01 -5.32821864e-02 -5.69423139e-01 5.44838548e-01
-1.20712563e-01 -1.04869261e-01 5.61746299e-01 8.34935606e-01
3.01242709e-01 3.14266682e-01 3.66929233e-01 9.18510199e-01
1.61502331e-01 -4.91619706e-01 -5.34456789e-01 3.16328287e-01
2.37387046e-01 1.05365360e+00 -7.96845257e-01 -2.85625517e-01
-3.22582781e-01 1.29586637e+00 4.27797228e-01 5.44563651e-01
-9.87634242e-01 2.65870273e-01 7.46975958e-01 3.90702963e-01
6.34929538e-01 -6.12339318e-01 6.51143566e-02 -1.49995685e+00
3.93347502e-01 -4.16703016e-01 6.72734454e-02 -1.19964445e+00
-7.30933249e-01 6.30015910e-01 -2.15649903e-02 -1.26426136e+00
-3.68886322e-01 -6.86179161e-01 -3.00667286e-02 4.44431186e-01
-1.53839898e+00 -1.48560071e+00 -7.28353024e-01 7.48356879e-01
5.19282520e-01 4.68874395e-01 5.32044113e-01 1.38135836e-01
-4.86085266e-01 6.63401484e-02 -2.51462042e-01 1.31424284e-02
6.23510659e-01 -9.85573351e-01 5.06785870e-01 9.88691092e-01
2.65859812e-01 5.12122571e-01 6.73193574e-01 -6.25656188e-01
-1.75570261e+00 -9.33142304e-01 5.16422570e-01 -1.01279390e+00
2.56430060e-01 -4.68526155e-01 -8.20894420e-01 1.15347850e+00
-3.73946726e-02 5.17619729e-01 2.18704250e-03 -2.73391724e-01
-3.40380400e-01 2.35105325e-02 -9.14430082e-01 5.47465682e-01
1.37391233e+00 -4.62457478e-01 -4.51371193e-01 4.08030301e-01
8.38166535e-01 -1.10318387e+00 -6.71785593e-01 2.88340837e-01
9.37748313e-01 -1.28579223e+00 1.43421745e+00 -2.05040753e-01
3.42248768e-01 -6.04356050e-01 -3.47563356e-01 -9.84025776e-01
-8.48117918e-02 -5.70559621e-01 -7.77396202e-01 6.71125591e-01
-4.21572238e-01 -4.57333744e-01 7.33837605e-01 7.55969405e-01
2.73313969e-02 -4.97099072e-01 -8.82967651e-01 -5.51682830e-01
-4.60573912e-01 -5.97827613e-01 2.64125496e-01 6.56261265e-01
-3.93553704e-01 8.52669328e-02 -6.36253595e-01 4.77390826e-01
9.39590693e-01 7.71300346e-02 1.21983814e+00 -1.27370715e+00
-3.59898835e-01 -1.56094119e-01 -6.88422322e-01 -1.39339948e+00
2.80480027e-01 -4.78432357e-01 3.33817214e-01 -1.60018337e+00
-8.54719430e-02 -1.75168127e-01 2.04743654e-01 5.57072401e-01
-3.24804485e-02 5.39593160e-01 3.63293707e-01 2.98605949e-01
-6.67564690e-01 5.19027472e-01 1.16703796e+00 6.45310879e-02
-1.46218583e-01 -4.49719504e-02 4.76360470e-02 1.11929691e+00
4.21346247e-01 -3.70555788e-01 -9.19583142e-02 -9.17671740e-01
-2.81936545e-02 1.40204668e-01 9.81005788e-01 -1.21222973e+00
5.04879296e-01 -1.58886015e-02 6.98785186e-01 -8.14133584e-01
9.83457029e-01 -1.09069574e+00 6.09666646e-01 5.89575946e-01
-2.51365490e-02 -6.77046413e-03 1.63607389e-01 7.49066114e-01
2.73596704e-01 1.92835718e-01 6.72158122e-01 -3.10441285e-01
-8.72069240e-01 4.55422103e-01 1.78133603e-03 -1.74698070e-01
9.68605816e-01 -3.67854178e-01 -3.75687815e-02 -3.57779115e-01
-7.16989577e-01 1.44523785e-01 9.30990934e-01 6.28319323e-01
8.09306085e-01 -1.27161932e+00 -5.14073789e-01 3.49617213e-01
-3.68342064e-02 5.10348678e-01 3.41496021e-01 1.03194892e+00
-8.18443537e-01 4.40455884e-01 -1.78156301e-01 -1.25411987e+00
-1.31182694e+00 5.03124177e-01 4.30082738e-01 1.94164202e-01
-1.06876254e+00 8.00300062e-01 4.03154343e-01 -4.58507538e-01
6.12781823e-01 -6.50486588e-01 2.91284710e-01 -3.39059085e-01
7.19733596e-01 4.82392848e-01 -1.44089252e-01 -1.10667503e+00
-3.66240531e-01 1.07321596e+00 2.79916614e-01 -2.74706393e-01
1.45041895e+00 -4.36269283e-01 -1.01149186e-01 6.07438862e-01
1.20617878e+00 -1.96452096e-01 -1.96381021e+00 -5.83929479e-01
-3.95000786e-01 -8.07352960e-01 3.67635414e-02 -3.52617770e-01
-1.23318565e+00 9.55352128e-01 4.82513011e-01 -2.40543038e-01
9.70071912e-01 1.49115011e-01 8.19518089e-01 3.46531302e-01
7.46635973e-01 -6.03915870e-01 5.42949922e-02 3.54424208e-01
7.68528402e-01 -1.16753459e+00 1.78612694e-01 -4.26272571e-01
-4.25572842e-01 1.06294847e+00 6.79900229e-01 -3.99487466e-02
3.56454581e-01 1.67560771e-01 6.43530628e-03 -1.20617077e-01
-5.91285110e-01 -2.63965786e-01 2.59253651e-01 8.29399526e-01
3.07488516e-02 -1.94175735e-01 5.18442333e-01 -1.40734138e-02
-8.60644057e-02 1.50311545e-01 4.79479194e-01 1.09030879e+00
-3.64849418e-01 -7.03324139e-01 -6.02390945e-01 -2.79252589e-01
-1.07572585e-01 2.36595720e-01 -2.75856405e-01 6.83665395e-01
1.89616457e-01 6.61717176e-01 1.34761915e-01 -1.65199503e-01
3.56067806e-01 -1.37399673e-01 9.92587566e-01 -3.06231171e-01
-1.24339074e-01 1.77802563e-01 -3.05570871e-01 -1.19283521e+00
-9.37357724e-01 -7.87871242e-01 -1.41265380e+00 -3.43056917e-01
-2.12121472e-01 -4.59921926e-01 7.41115749e-01 9.25729871e-01
4.18164760e-01 4.99867171e-01 1.51458487e-01 -1.77150118e+00
-1.41438162e-02 -6.04156971e-01 -3.63699585e-01 3.25707167e-01
7.00050533e-01 -1.00613105e+00 -2.22210810e-01 4.27082717e-01] | [7.39882755279541, -1.3946746587753296] |
d27f09e3-849f-4b9b-8ba6-975cd6d8da8e | generative-models-for-spear-phishing-posts-on | 1802.05196 | null | http://arxiv.org/abs/1802.05196v1 | http://arxiv.org/pdf/1802.05196v1.pdf | Generative Models for Spear Phishing Posts on Social Media | Historically, machine learning in computer security has prioritized defense:
think intrusion detection systems, malware classification, and botnet traffic
identification. Offense can benefit from data just as well. Social networks,
with their access to extensive personal data, bot-friendly APIs, colloquial
syntax, and prevalence of shortened links, are the perfect venues for spreading
machine-generated malicious content. We aim to discover what capabilities an
adversary might utilize in such a domain. We present a long short-term memory
(LSTM) neural network that learns to socially engineer specific users into
clicking on deceptive URLs. The model is trained with word vector
representations of social media posts, and in order to make a click-through
more likely, it is dynamically seeded with topics extracted from the target's
timeline. We augment the model with clustering to triage high value targets
based on their level of social engagement, and measure success of the LSTM's
phishing expedition using click-rates of IP-tracked links. We achieve state of
the art success rates, tripling those of historic email attack campaigns, and
outperform humans manually performing the same task. | ['Philip Tully', 'John Seymour'] | 2018-02-14 | null | null | null | null | ['computer-security'] | ['miscellaneous'] | [-9.08614621e-02 8.88857245e-02 -4.50309068e-01 -1.51936889e-01
-2.87166655e-01 -7.98729599e-01 1.05442679e+00 3.53174955e-02
-5.12917459e-01 4.49941725e-01 1.70189947e-01 -1.01265526e+00
-1.70699090e-01 -9.47399020e-01 -3.47087294e-01 -2.60782957e-01
-5.69857061e-01 6.97512746e-01 3.95087302e-01 -3.31056565e-01
6.71064436e-01 4.48583901e-01 -5.81370413e-01 4.51070815e-01
4.35938597e-01 6.04436457e-01 -1.72094002e-01 8.37655902e-01
-3.24401230e-01 7.84317672e-01 -9.08900619e-01 -6.72463000e-01
1.23685688e-01 9.12515223e-02 -8.86982262e-01 -4.74155128e-01
5.96708693e-02 -5.41493297e-01 -8.69020104e-01 9.72315729e-01
1.62675738e-01 -1.98944643e-01 5.86387575e-01 -1.20360732e+00
-8.68072391e-01 8.26349497e-01 -6.14456415e-01 7.34988511e-01
-2.92472187e-02 5.79307258e-01 1.17761362e+00 -3.17079335e-01
8.47578406e-01 1.36439109e+00 8.39447975e-01 5.99017799e-01
-1.33844495e+00 -6.30419433e-01 -1.23837851e-02 5.14607839e-02
-7.38138199e-01 -2.22050875e-01 6.87840044e-01 -7.57584512e-01
9.59089696e-01 -2.84292549e-03 5.47424078e-01 2.06270504e+00
1.73045725e-01 4.26106393e-01 6.20463550e-01 1.45480260e-01
-7.56448582e-02 4.08690959e-01 3.26276809e-01 7.55194485e-01
2.86547273e-01 2.22404584e-01 -2.32925057e-01 -8.24780822e-01
3.39342892e-01 2.87249744e-01 1.88865196e-02 9.47551727e-02
-1.02691841e+00 1.55209243e+00 8.65348160e-01 6.13994241e-01
-4.92265761e-01 1.43748939e-01 5.34463227e-01 5.82717359e-01
7.24461913e-01 9.69869494e-01 -4.98898268e-01 -1.91227317e-01
-8.42094600e-01 -2.68405434e-02 1.28715360e+00 1.14682682e-01
5.43408990e-01 5.78716956e-02 1.16956331e-01 6.81570709e-01
2.86102057e-01 4.33227211e-01 5.28142750e-01 -6.94941998e-01
2.90044993e-01 4.84371096e-01 -3.82147990e-02 -1.37229097e+00
-3.63559514e-01 -6.60083652e-01 -3.90485197e-01 -1.12229861e-01
5.81025958e-01 -5.63338995e-01 -5.66694736e-01 1.55973244e+00
-6.54930547e-02 1.28793359e-01 -6.80721045e-01 2.28366330e-01
-1.18897498e-01 8.00709903e-01 3.92401725e-01 9.54886526e-02
9.20818865e-01 -4.06658024e-01 -2.85073847e-01 -2.92372495e-01
8.71857882e-01 -3.08869869e-01 8.60403061e-01 2.34582409e-01
-6.76465094e-01 1.02836661e-01 -5.82046747e-01 5.37708223e-01
-1.04099154e+00 -7.62820899e-01 6.22124672e-01 8.90946209e-01
-1.02662528e+00 1.14193022e+00 -3.90167892e-01 -6.44223273e-01
7.93572426e-01 1.23356797e-01 1.26165286e-01 4.79318768e-01
-1.68598723e+00 1.05465460e+00 1.25794157e-01 -5.25652230e-01
-1.04456985e+00 -6.00969374e-01 1.83563158e-02 -1.54078845e-02
3.18258941e-01 -1.43577278e-01 8.94669771e-01 -6.84560001e-01
-9.95601594e-01 9.05428648e-01 3.37436050e-01 -8.45586538e-01
2.95355797e-01 6.01531789e-02 -6.84031427e-01 6.29643276e-02
1.25579610e-01 3.40359479e-01 1.46543872e+00 -9.55094874e-01
-5.29459298e-01 -4.67007935e-01 1.14438377e-01 -5.26137590e-01
-1.29977226e+00 3.75474304e-01 4.92434531e-01 -2.84024626e-01
-5.99722683e-01 -6.88231111e-01 -1.98319554e-01 -4.44548637e-01
-6.69045866e-01 -5.02284765e-01 1.39387822e+00 -9.28528070e-01
1.40033555e+00 -1.81564200e+00 3.05651966e-02 5.74216962e-01
9.71716046e-01 6.23725116e-01 -1.83217615e-01 6.00254297e-01
1.76187530e-01 9.68378663e-01 4.54315186e-01 1.27920151e-01
5.09132743e-02 -4.04815823e-01 -6.85053468e-01 3.27728480e-01
9.83550102e-02 9.48675692e-01 -1.07735753e+00 -1.68889657e-01
-1.80603743e-01 2.96147466e-01 -5.76647162e-01 5.78356646e-02
-4.69667703e-01 3.22880864e-01 -6.37413263e-01 4.33658600e-01
-1.20542794e-02 -6.86224937e-01 -4.86409031e-02 3.68524522e-01
1.13337919e-01 5.30394197e-01 3.26165371e-02 6.90135837e-01
-4.89465445e-01 1.05159521e+00 1.66244015e-01 -6.06860161e-01
7.72010565e-01 4.31292839e-02 4.79716182e-01 -3.87665778e-01
2.85473108e-01 1.31594852e-01 3.46831620e-01 -5.89302957e-01
1.11140616e-01 2.23466158e-01 1.45220146e-01 1.20546603e+00
-1.11672552e-02 4.24578160e-01 -9.35006328e-03 6.69029474e-01
1.69133055e+00 -7.71112442e-01 -4.44663689e-02 -6.68569952e-02
1.02747142e-01 4.58537834e-03 -1.49579629e-01 9.75984871e-01
-5.28918922e-01 -4.06151742e-01 9.55257475e-01 -5.15004873e-01
-1.29174829e+00 -1.24891782e+00 2.00515389e-01 1.73440731e+00
-4.65135127e-01 -2.57340163e-01 -7.87553430e-01 -1.11567008e+00
1.20179847e-01 9.09323871e-01 -5.61833262e-01 -3.42127800e-01
-8.29741359e-01 -6.17062688e-01 9.21502352e-01 -1.25698745e-01
3.56894702e-01 -1.08670425e+00 -1.10170729e-01 4.34520483e-01
-5.55440001e-02 -7.96492815e-01 -3.45981687e-01 -8.35834891e-02
-5.91895103e-01 -1.20053077e+00 -5.15584052e-01 -3.07369202e-01
3.42288494e-01 1.85582668e-01 1.03496814e+00 4.27693933e-01
-4.96202201e-01 2.12938547e-01 -1.93318784e-01 -3.06641161e-01
-6.47653937e-01 6.09412670e-01 2.98476964e-01 1.32699460e-01
7.11682022e-01 -8.62081051e-01 -5.01316190e-01 3.41978788e-01
-7.42584646e-01 -8.34291220e-01 4.49083447e-01 6.48698747e-01
-9.35788691e-01 1.06347464e-01 9.63696599e-01 -1.15454423e+00
1.25352120e+00 -1.34135878e+00 -1.87124431e-01 -1.62957177e-01
-5.13344407e-01 -3.20582330e-01 6.60859525e-01 -9.25934076e-01
-6.07411683e-01 -6.23638988e-01 3.87890674e-02 -2.21837357e-01
3.09986062e-02 3.57005268e-01 4.58073616e-01 -2.80386835e-01
1.40758717e+00 1.68986350e-01 6.42575398e-02 -1.83959574e-01
6.31135643e-01 8.26885700e-01 -2.14088023e-01 -1.89626276e-01
1.36845517e+00 3.68509948e-01 -4.21934158e-01 -1.16119361e+00
-6.44418597e-01 -3.79526049e-01 -3.07665825e-01 -5.87912798e-01
8.49715531e-01 4.63973433e-02 -1.15664852e+00 4.39640343e-01
-1.51662874e+00 -3.32128555e-01 2.46470287e-01 3.63509580e-02
2.17755195e-02 3.38667482e-01 -1.01163518e+00 -8.66347730e-01
-2.75687248e-01 -5.47481120e-01 2.35078499e-01 -6.47386834e-02
-6.88372910e-01 -1.57294512e+00 2.03211695e-01 4.11054879e-01
1.38639224e+00 1.63931370e-01 1.21790850e+00 -1.58369863e+00
-4.78957891e-01 -7.37487316e-01 -6.13006353e-01 2.87213832e-01
-3.17995972e-03 5.29656112e-02 -8.46830308e-01 -2.36572161e-01
1.71050541e-02 -4.70751345e-01 9.18990374e-01 8.68752971e-02
1.18769062e+00 -1.06057036e+00 -7.04383135e-01 2.89341677e-02
8.38735342e-01 6.95081502e-02 3.02194476e-01 2.29794860e-01
6.20260537e-01 1.16000211e+00 -2.65370578e-01 3.08387220e-01
-7.75078312e-02 2.39799157e-01 7.11416781e-01 1.83553353e-01
3.28848302e-01 -5.82380712e-01 5.79046667e-01 4.20675159e-01
1.40143707e-01 -3.32541078e-01 -1.03412819e+00 3.17590982e-01
-1.29222369e+00 -1.39045489e+00 -1.52998447e-01 2.10399532e+00
5.80313802e-01 7.03752279e-01 6.93315685e-01 -1.29848897e-01
8.98787916e-01 6.81652725e-01 -6.11655056e-01 -3.64694774e-01
3.63693506e-01 1.56816244e-01 9.33329046e-01 8.28131676e-01
-9.72146273e-01 1.09605110e+00 6.27292728e+00 8.15706253e-01
-1.03688574e+00 4.21478987e-01 7.01238751e-01 -2.17267647e-01
-1.23268656e-01 -2.08110705e-01 -7.96770215e-01 8.54771435e-01
1.62805200e+00 -2.00740844e-02 9.20329690e-01 8.04851294e-01
3.38043332e-01 5.44072509e-01 -8.17660987e-01 3.37434739e-01
2.41657589e-02 -1.42498183e+00 4.26699631e-02 6.48782134e-01
3.99020404e-01 4.85270441e-01 6.92136705e-01 4.57392901e-01
8.26537907e-01 -1.07387722e+00 -6.90774620e-02 3.47997427e-01
3.72131437e-01 -4.92535651e-01 1.80597633e-01 6.48151875e-01
-4.33191210e-01 -6.40285373e-01 -1.81759521e-01 1.77312404e-01
3.84310246e-01 6.34315610e-01 -1.29639840e+00 -4.49900180e-01
2.41226137e-01 7.88573384e-01 -7.78137207e-01 5.38421094e-01
1.19707379e-02 1.04796064e+00 -1.34400070e-01 -6.32656276e-01
6.33129478e-01 1.31474569e-01 1.08209324e+00 1.12408626e+00
-2.32381284e-01 -3.79972011e-01 -1.83357131e-02 1.14435410e+00
-3.81722897e-01 -2.62291431e-01 -1.20690894e+00 -1.04818475e+00
5.23760915e-01 1.36829495e+00 -4.65350360e-01 -2.92311907e-01
9.22633931e-02 7.47811258e-01 3.57592195e-01 5.31719625e-01
-7.95960903e-01 -6.58520043e-01 6.68291390e-01 8.45953941e-01
-3.45025152e-01 -4.55323040e-01 -2.85265088e-01 -9.37849820e-01
-5.46680987e-01 -7.35464215e-01 2.78369099e-01 -3.97542864e-01
-1.77461410e+00 8.01603556e-01 -4.81135279e-01 -4.73207325e-01
-5.17416835e-01 -6.44715667e-01 -1.15655386e+00 7.49050796e-01
-1.00128782e+00 -1.03093052e+00 1.66039020e-01 4.69731569e-01
4.63231295e-01 -7.43495107e-01 4.60456043e-01 3.56328905e-01
-4.04337704e-01 3.19539756e-01 -2.31033325e-01 4.73518103e-01
4.69413340e-01 -1.07959104e+00 9.24615920e-01 4.50979173e-01
2.41734590e-02 1.01466060e+00 7.05901146e-01 -1.22497714e+00
-9.47978497e-01 -1.04310215e+00 1.02582741e+00 -1.07626104e+00
1.83115566e+00 -4.21074569e-01 -8.09384584e-01 9.60039437e-01
8.65244642e-02 -6.66259885e-01 5.42997777e-01 3.91147912e-01
-7.99849570e-01 3.05136830e-01 -1.27376246e+00 8.17596316e-01
1.05246377e+00 -8.81503880e-01 -3.45893800e-01 1.14492226e+00
1.01430190e+00 7.48815835e-01 -4.49336112e-01 -3.52886975e-01
4.16926980e-01 -7.58620918e-01 1.12406754e+00 -1.56073463e+00
3.80766064e-01 5.61992526e-01 2.24506214e-01 -1.33921635e+00
-5.26926219e-01 -9.05755579e-01 -3.36343884e-01 8.45315337e-01
7.67587245e-01 -1.11516380e+00 1.09997737e+00 3.18052769e-01
3.84562999e-01 -5.65452635e-01 -6.70311570e-01 -6.79109931e-01
2.48283908e-01 -1.31627470e-01 2.02342253e-02 1.20209372e+00
1.10071249e-01 5.94327211e-01 -6.72343731e-01 -1.34933874e-01
1.09402442e+00 -7.74134696e-01 6.30719125e-01 -1.74586999e+00
-4.70364362e-01 -1.04050410e+00 -8.17473754e-02 -7.27016568e-01
3.57402772e-01 -9.92595255e-01 -7.07700551e-01 -8.73378813e-01
-5.40871583e-02 -3.27866524e-01 3.49399261e-03 1.88520148e-01
3.76277775e-01 2.74201453e-01 1.91220064e-02 4.23004806e-01
-5.33640683e-01 -1.12825677e-01 6.52441680e-01 -4.37966496e-01
-3.74987394e-01 4.19861436e-01 -6.82514369e-01 9.32091177e-01
1.10770047e+00 -6.17160738e-01 -1.18326850e-01 -1.21572554e-01
5.22004604e-01 -2.28833616e-01 5.21133661e-01 -5.97031832e-01
2.79122919e-01 -1.13875687e-01 2.10388571e-01 -2.54119128e-01
5.74952543e-01 -4.99674618e-01 -5.31420648e-01 9.59445000e-01
-8.46423030e-01 -1.77768003e-02 -4.58910704e-01 1.02944541e+00
6.02065146e-01 -1.54949486e-01 1.00480604e+00 -2.52777308e-01
-4.67156880e-02 4.80788946e-01 -9.58611131e-01 2.32048824e-01
9.33026195e-01 5.08697005e-03 -8.31905246e-01 -8.68866801e-01
-7.73572087e-01 1.37152541e-02 9.49037671e-02 6.06129944e-01
5.20204425e-01 -8.98492932e-01 -7.07547545e-01 1.31129295e-01
-2.92767465e-01 -1.22776151e+00 -1.57479703e-01 6.80127800e-01
-2.35004008e-01 5.95246673e-01 -3.01035553e-01 -2.52134979e-01
-9.13366675e-01 5.13238490e-01 2.79380739e-01 -4.26580101e-01
-2.21588612e-01 7.54411995e-01 -3.40881884e-01 -4.16954488e-01
4.01126534e-01 4.55678701e-01 -3.15653682e-01 2.35946685e-01
7.19279110e-01 7.28100300e-01 -3.37099731e-01 -2.94290215e-01
-8.20899606e-02 -2.74700254e-01 -4.82219934e-01 -8.88447985e-02
1.17163682e+00 1.84196234e-01 -3.65666807e-01 2.03422189e-01
1.46790445e+00 2.18771026e-03 -6.69155180e-01 -3.63509238e-01
5.29625893e-01 -6.86352730e-01 -1.97536796e-01 -9.01369691e-01
-8.35205972e-01 9.37644482e-01 8.91916975e-02 1.31374598e+00
5.37416339e-02 8.68747607e-02 1.46464527e+00 7.35956430e-01
2.28207722e-01 -8.52344275e-01 9.33215559e-01 7.41081953e-01
4.32191074e-01 -1.00787961e+00 -3.57771516e-01 1.33053690e-01
-2.25907311e-01 1.02028453e+00 5.14165342e-01 -2.98555940e-01
1.09090364e+00 -2.40599960e-01 -3.34598541e-01 -5.78360200e-01
-7.95075297e-01 2.59055167e-01 -2.08551347e-01 7.50641525e-01
-2.03905925e-02 2.46168692e-02 2.93607241e-03 5.55549115e-02
-5.52485399e-02 -7.41096020e-01 5.97883105e-01 3.73859018e-01
-1.15394056e+00 -7.50983953e-01 -1.18633501e-01 1.06036043e+00
-7.75657475e-01 -3.35122257e-01 -1.09270918e+00 4.10943657e-01
-4.19773847e-01 9.72844243e-01 -7.20483512e-02 -9.91212010e-01
-2.95747101e-01 2.13703305e-01 -2.97166020e-01 -6.49362803e-01
-8.72616291e-01 -7.22317934e-01 2.40501061e-01 -4.52010572e-01
4.75580305e-01 -3.83980989e-01 -5.66962719e-01 -1.12761402e+00
-2.46787176e-01 1.36242107e-01 9.86310482e-01 7.95357883e-01
4.40573156e-01 1.65786773e-01 1.12332702e+00 -8.01646054e-01
-1.08381879e+00 -1.18283975e+00 -2.74964958e-01 4.50053096e-01
2.76838809e-01 -5.40496767e-01 -9.80994284e-01 -5.26033819e-01] | [8.07491397857666, 10.0950288772583] |
31e23a75-6bdb-47af-ac59-c18d21b05d82 | expansion-via-prediction-of-importance-with | 2004.14245 | null | https://arxiv.org/abs/2004.14245v2 | https://arxiv.org/pdf/2004.14245v2.pdf | Expansion via Prediction of Importance with Contextualization | The identification of relevance with little textual context is a primary challenge in passage retrieval. We address this problem with a representation-based ranking approach that: (1) explicitly models the importance of each term using a contextualized language model; (2) performs passage expansion by propagating the importance to similar terms; and (3) grounds the representations in the lexicon, making them interpretable. Passage representations can be pre-computed at index time to reduce query-time latency. We call our approach EPIC (Expansion via Prediction of Importance with Contextualization). We show that EPIC significantly outperforms prior importance-modeling and document expansion approaches. We also observe that the performance is additive with the current leading first-stage retrieval methods, further narrowing the gap between inexpensive and cost-prohibitive passage ranking approaches. Specifically, EPIC achieves a MRR@10 of 0.304 on the MS-MARCO passage ranking dataset with 78ms average query latency on commodity hardware. We also find that the latency is further reduced to 68ms by pruning document representations, with virtually no difference in effectiveness. | ['Nicola Tonellotto', 'Raffaele Perego', 'Nazli Goharian', 'Franco Maria Nardini', 'Sean MacAvaney', 'Ophir Frieder'] | 2020-04-29 | null | null | null | null | ['passage-ranking'] | ['natural-language-processing'] | [ 2.50722647e-01 -2.73183495e-01 -5.70565999e-01 1.29796630e-02
-1.72203720e+00 -7.33623266e-01 6.39268935e-01 8.99367511e-01
-6.48813128e-01 5.72756052e-01 9.10524547e-01 -5.45326471e-01
-2.63930976e-01 -5.15363753e-01 -4.56766844e-01 1.04216494e-01
-3.02317590e-01 6.84484839e-01 4.63793755e-01 -5.43971479e-01
9.41615462e-01 2.52775550e-01 -1.56722307e+00 5.87798357e-01
5.65027654e-01 1.00120378e+00 1.15507402e-01 1.09898865e+00
-2.77765036e-01 1.00307083e+00 -7.12173879e-01 2.28740070e-02
1.05960861e-01 -2.96395253e-02 -1.39142787e+00 -7.26362526e-01
3.07380676e-01 -6.06463790e-01 -5.37052751e-01 4.60232049e-01
4.59121734e-01 4.39338505e-01 5.47080696e-01 -4.92580891e-01
-6.76873505e-01 7.24393427e-01 -7.20260382e-01 7.35593438e-01
7.71641791e-01 -6.07973933e-01 1.59436882e+00 -1.06297517e+00
5.41726351e-01 1.02897036e+00 2.60365337e-01 5.30113280e-02
-9.66989934e-01 -1.04742430e-01 2.74464905e-01 1.74275741e-01
-1.68303967e+00 -5.44456959e-01 8.96171033e-02 -4.37942632e-02
1.63540864e+00 8.36252809e-01 5.00926733e-01 2.02035785e-01
-1.73497554e-02 7.96729982e-01 5.41978657e-01 -7.52875268e-01
7.44607076e-02 -2.35371187e-01 5.89968860e-01 4.17659253e-01
2.82544971e-01 -3.65880698e-01 -5.49369395e-01 -7.62599587e-01
2.84250140e-01 5.25985174e-02 -1.27255619e-01 4.13571388e-01
-7.72978902e-01 5.48323035e-01 2.36779675e-01 1.93002135e-01
-4.38451082e-01 4.88008082e-01 5.45330763e-01 4.68304366e-01
3.44658673e-01 8.52554440e-01 -4.37511832e-01 -1.29800722e-01
-1.26049459e+00 3.86033416e-01 8.88614714e-01 1.05042422e+00
6.70286715e-01 -4.57439691e-01 -6.12219691e-01 9.15365875e-01
3.28788191e-01 3.83481324e-01 5.40744960e-01 -9.14630353e-01
4.81474847e-01 4.28206831e-01 3.39111954e-01 -8.51166189e-01
-2.10753366e-01 -4.82055724e-01 -1.24744058e-01 -4.83226687e-01
-2.08677784e-01 2.68735021e-01 -7.46072114e-01 1.45023048e+00
4.53470796e-02 -2.77302474e-01 8.05727243e-02 6.36162639e-01
6.66431367e-01 9.77769077e-01 2.53904253e-01 -2.49675840e-01
1.59399927e+00 -1.09500206e+00 -3.08298737e-01 -3.84771883e-01
7.44920373e-01 -1.24867737e+00 1.14168394e+00 5.81651069e-02
-1.30366433e+00 -1.81999028e-01 -1.11835790e+00 -6.52343750e-01
-1.07370004e-01 -5.80880083e-02 7.87129104e-01 3.42476785e-01
-1.51918828e+00 5.05670309e-01 -6.30005717e-01 -2.87462205e-01
-2.71366090e-01 5.75964332e-01 1.81418836e-01 3.15448157e-02
-1.30338216e+00 7.31873989e-01 2.62591451e-01 -3.97216171e-01
-5.32380581e-01 -8.15543830e-01 -3.96482915e-01 4.78466243e-01
3.00543368e-01 -8.32391500e-01 1.46780813e+00 -3.93854052e-01
-1.09196377e+00 4.97428507e-01 -5.75042129e-01 -4.26958174e-01
-1.94998413e-01 -6.63444459e-01 -3.95679861e-01 4.93054628e-01
6.41931072e-02 5.33552468e-01 5.22046566e-01 -8.87611330e-01
-7.52754390e-01 -5.34545556e-02 5.38613260e-01 6.69655740e-01
-6.55281961e-01 3.43603164e-01 -1.18518686e+00 -6.43307924e-01
1.38760448e-01 -8.69293272e-01 -3.92443389e-01 -3.30202997e-01
-9.69846919e-02 -3.11257273e-01 3.51749867e-01 -8.61576676e-01
2.01291823e+00 -1.71336305e+00 -3.05210263e-01 5.07452965e-01
2.79689193e-01 9.76746902e-02 -3.85073841e-01 8.54681909e-01
2.68294632e-01 4.27482456e-01 3.00251603e-01 -1.36132464e-01
-1.29043937e-01 -1.78963974e-01 -8.80979180e-01 -1.16949588e-01
-2.07152829e-01 1.08131504e+00 -9.64363694e-01 -7.81058788e-01
-2.54410565e-01 4.70138490e-01 -9.15006280e-01 -3.47719155e-02
-2.57750750e-01 -2.42057607e-01 -8.23595047e-01 6.80298150e-01
4.14775126e-02 -5.84813476e-01 2.25551546e-01 -1.09543530e-02
-1.86034329e-02 1.13423657e+00 -7.89294839e-01 1.66407967e+00
-7.21630692e-01 6.53539538e-01 -3.27875346e-01 -2.11913884e-01
5.91644883e-01 2.02972934e-01 6.01336360e-01 -9.05933201e-01
-3.27559143e-01 2.50987768e-01 -5.53853929e-01 -1.50422633e-01
1.58265078e+00 4.47636902e-01 -2.99816459e-01 8.15153658e-01
-5.63296735e-01 -5.95905036e-02 4.77754533e-01 7.93948889e-01
1.42934012e+00 -2.12766171e-01 2.92858809e-01 -2.73848534e-01
3.31422031e-01 1.97126523e-01 -2.03135628e-02 1.07050478e+00
3.86473984e-01 4.56322104e-01 1.41320616e-01 -1.02083169e-01
-9.39232290e-01 -8.16232800e-01 5.71319126e-02 1.72626805e+00
3.05751652e-01 -1.14525628e+00 -3.92276555e-01 -3.15314293e-01
-6.49212077e-02 5.92945039e-01 -1.95877150e-01 -2.21835688e-01
-7.20853806e-01 -4.44557428e-01 4.43687409e-01 7.60730803e-01
-1.34318829e-01 -5.49101889e-01 -4.54714268e-01 2.97605515e-01
-4.37792689e-01 -6.81679070e-01 -7.72589982e-01 5.99023234e-03
-1.06843770e+00 -6.05032682e-01 -5.86143732e-01 -6.84860289e-01
6.35082960e-01 5.27839601e-01 1.52039421e+00 7.14704812e-01
-1.40231788e-01 6.93719983e-01 -5.94994783e-01 1.03823833e-01
-6.53306618e-02 4.03168023e-01 -1.90551683e-01 -9.01039422e-01
4.05036002e-01 -4.30057675e-01 -1.07834387e+00 3.40432860e-02
-9.82232094e-01 -3.61885369e-01 6.18878722e-01 7.03870356e-01
8.63232970e-01 -3.33228469e-01 4.64080989e-01 -6.46196783e-01
1.08170307e+00 -4.61061686e-01 -5.94420373e-01 6.25231624e-01
-1.03718758e+00 3.80613357e-01 2.35085815e-01 -2.23171175e-01
-8.61813784e-01 -3.28915656e-01 -2.10609287e-01 -3.10648046e-02
6.86792433e-01 9.38175082e-01 8.14002454e-01 2.02543531e-02
8.37205768e-01 1.34393468e-01 -5.41856825e-01 -5.24383962e-01
4.72532570e-01 7.22213387e-01 3.65736783e-01 -9.01958585e-01
4.49366570e-01 1.54774934e-01 -1.12757854e-01 -7.79700696e-01
-8.09554160e-01 -9.95660961e-01 -1.49495095e-01 1.49596274e-01
1.26659930e-01 -1.17340553e+00 -4.44110662e-01 -5.18731713e-01
-1.04235411e+00 -1.20333225e-01 -3.52995217e-01 4.19197798e-01
-1.38999641e-01 7.44515061e-01 -1.11600876e+00 -7.40832210e-01
-1.10786974e+00 -8.20897281e-01 1.32854772e+00 1.01272725e-01
-7.16694534e-01 -5.61107039e-01 2.44583800e-01 3.50766420e-01
5.79520285e-01 -5.37020981e-01 1.10252571e+00 -6.20498180e-01
-7.57682741e-01 -6.04550958e-01 -3.60807240e-01 -2.54981846e-01
-2.86701620e-01 -1.81544349e-01 -7.79582322e-01 -4.02682006e-01
-4.52511042e-01 -4.15199786e-01 1.02587426e+00 2.44394690e-01
1.13254237e+00 -2.78673321e-01 -4.23194915e-01 1.35983601e-01
1.40733099e+00 1.65242001e-01 6.55259550e-01 3.41528594e-01
1.97508395e-01 2.82087117e-01 8.18558633e-01 5.33042133e-01
3.56935859e-01 7.27413177e-01 -1.25332355e-01 1.29030362e-01
-1.65761262e-01 -5.16533017e-01 2.28846774e-01 1.21980047e+00
-4.19187024e-02 -4.45101857e-01 -8.01073790e-01 6.77514195e-01
-1.77169108e+00 -8.27350557e-01 1.76101297e-01 2.49450350e+00
9.38950062e-01 1.83753580e-01 -6.31221905e-02 -1.94430351e-01
3.78534317e-01 1.94330454e-01 -2.93918639e-01 -5.66638350e-01
2.58685201e-01 5.62661052e-01 6.03668928e-01 9.56403136e-01
-8.01830709e-01 1.00590277e+00 7.48940945e+00 8.97283435e-01
-7.88621902e-01 6.09829463e-02 6.52791858e-01 -4.42941666e-01
-6.80202484e-01 3.07726771e-01 -1.08044636e+00 -2.30473652e-02
1.13918352e+00 -7.31382430e-01 4.39978302e-01 9.41157579e-01
-2.31548607e-01 -1.73093751e-01 -1.04230320e+00 7.20841587e-01
1.94875747e-01 -1.29197228e+00 3.87291759e-01 -6.49846345e-02
7.22912669e-01 2.12100297e-01 3.48912589e-02 5.99129617e-01
2.17779294e-01 -7.47957468e-01 4.77181077e-01 4.72892016e-01
9.85000670e-01 -6.29322410e-01 3.89652342e-01 3.49648506e-03
-1.54719412e+00 1.28446287e-03 -6.63597643e-01 -3.20881009e-02
1.59180686e-01 3.82747203e-01 -9.00019228e-01 3.19539249e-01
4.48949277e-01 5.60114831e-02 -6.05010748e-01 1.07486618e+00
-9.86547247e-02 5.74920058e-01 -6.59833312e-01 -3.99973631e-01
6.69959038e-02 3.62099141e-01 5.37011921e-01 1.36064541e+00
4.18942481e-01 1.35246545e-01 1.46789268e-01 3.01995784e-01
-3.35672766e-01 4.46500212e-01 -2.95811951e-01 -1.69953600e-01
8.71982813e-01 9.69913423e-01 -6.83711469e-01 -6.98143542e-01
-2.37276793e-01 1.06434178e+00 2.91737109e-01 3.64414662e-01
-5.47770500e-01 -6.52768612e-01 3.56563777e-01 1.73450708e-02
2.37396389e-01 -1.20094724e-01 3.24345492e-02 -1.13357782e+00
2.55481988e-01 -5.71203887e-01 6.34805501e-01 -7.28953719e-01
-7.55601168e-01 7.14504719e-01 2.34108686e-01 -9.33780968e-01
-7.41556227e-01 -1.85111433e-01 -1.62213668e-01 1.02939427e+00
-1.57826948e+00 -5.30650795e-01 1.59717575e-01 2.72453606e-01
6.51146948e-01 3.35177928e-01 1.01492381e+00 4.55107093e-01
-3.13375220e-02 8.31459880e-01 3.00176799e-01 -3.24330181e-01
6.27001226e-01 -1.01909590e+00 7.09180772e-01 6.32563055e-01
3.60416323e-01 1.44475067e+00 4.32804704e-01 -7.01162279e-01
-1.63078356e+00 -7.90642262e-01 1.54929256e+00 -5.03627002e-01
5.39518714e-01 -1.77946948e-02 -8.69557977e-01 4.83591914e-01
9.81015116e-02 -2.60561407e-01 8.41519356e-01 5.26200473e-01
-6.09636664e-01 -1.04244515e-01 -5.98015368e-01 9.62580323e-01
9.26955938e-01 -1.00175571e+00 -8.33427131e-01 5.85011542e-01
1.16502798e+00 -3.47622216e-01 -7.91916490e-01 2.34646320e-01
8.36714804e-01 -3.24513257e-01 1.36354613e+00 -5.36858022e-01
2.70219505e-01 -2.00395554e-01 -3.67305040e-01 -6.73015058e-01
-3.78844112e-01 -6.83197379e-01 -5.95737755e-01 8.77824783e-01
7.71285772e-01 -3.50716829e-01 6.99942768e-01 9.90825355e-01
6.35522082e-02 -9.21311200e-01 -4.37422901e-01 -5.17574251e-01
-1.10129997e-01 -4.70079809e-01 6.97031677e-01 4.00167584e-01
4.91907090e-01 6.47178590e-01 -1.22484215e-01 5.08398041e-02
6.14819601e-02 3.82507384e-01 4.11125869e-01 -8.75422537e-01
-5.85432231e-01 -4.54555482e-01 2.03080866e-02 -1.89391971e+00
-2.89293796e-01 -9.70311582e-01 6.86708000e-03 -1.61593854e+00
5.84952295e-01 -4.97087121e-01 -7.12244213e-01 4.27903652e-01
-3.17339867e-01 2.16764018e-01 -3.48080844e-02 6.47820711e-01
-1.17375720e+00 2.60460675e-01 6.48591220e-01 -1.57593206e-01
-2.83148348e-01 -2.45529264e-01 -1.19143212e+00 4.65699732e-01
5.18652201e-01 -4.84171391e-01 -7.94035017e-01 -6.95686519e-01
7.79391646e-01 3.75772089e-01 -1.14283249e-01 -7.87268639e-01
6.40919268e-01 1.21515520e-01 2.77779430e-01 -8.99611354e-01
4.09571230e-01 -4.24647301e-01 -2.92817205e-01 3.20227832e-01
-9.73096430e-01 5.17394006e-01 3.80237252e-01 5.20808041e-01
-1.88933402e-01 -5.34396172e-01 7.83102885e-02 5.47820851e-02
-8.07182014e-01 1.37892440e-01 -5.23566782e-01 3.44790697e-01
2.03997985e-01 1.79107964e-01 -3.77678066e-01 -6.64409995e-01
-2.77717084e-01 2.81280518e-01 3.61711621e-01 2.02480480e-01
6.71402752e-01 -1.11847103e+00 -4.97554153e-01 -2.99245566e-01
3.11902046e-01 -3.94408613e-01 9.78650749e-02 3.53399187e-01
-6.52849376e-01 9.51107800e-01 5.62775075e-01 -2.78829843e-01
-1.47181797e+00 4.06101376e-01 -2.17574403e-01 -8.36113632e-01
-4.26991671e-01 9.82334733e-01 -2.19572466e-02 1.64770991e-01
2.52961904e-01 -1.32803038e-01 -3.34688693e-01 -1.20837241e-01
1.04311824e+00 3.53436202e-01 3.50198179e-01 -3.08611363e-01
-4.03931439e-01 5.16026556e-01 -6.73728406e-01 -5.80185056e-01
9.91702676e-01 -3.40513080e-01 -2.74245650e-01 7.53085315e-03
1.44284439e+00 4.36482161e-01 -4.88243550e-01 -3.96269441e-01
3.87236863e-01 -4.15800929e-01 4.22433406e-01 -8.91505897e-01
-5.33986390e-01 4.32416439e-01 3.23798776e-01 7.23078102e-02
1.35366178e+00 -1.20308986e-02 1.00064504e+00 1.03948879e+00
5.79642177e-01 -1.08597159e+00 5.71906120e-02 7.15815187e-01
9.70079005e-01 -7.27931917e-01 5.33424437e-01 -1.12457260e-01
-3.28622073e-01 9.32258904e-01 1.29086480e-01 -2.15400830e-02
4.80296582e-01 2.05961585e-01 -3.24333161e-01 -3.47809047e-01
-1.13407815e+00 -1.79528102e-01 7.33661115e-01 -1.54858381e-02
8.53845477e-01 -1.36620149e-01 -6.68375134e-01 4.14270103e-01
-1.66806877e-01 -9.46226157e-03 6.74498454e-02 1.23194468e+00
-7.23748624e-01 -1.36328220e+00 -2.69094914e-01 7.00058043e-01
-8.56607854e-01 -9.53003943e-01 -3.62826020e-01 5.03279865e-01
-7.46010840e-01 1.19232249e+00 5.07061444e-02 -4.33502346e-01
1.90126210e-01 -9.63693671e-03 2.44647026e-01 -8.15341532e-01
-7.40566075e-01 2.39342824e-01 4.44257617e-01 -6.50171757e-01
1.22561470e-01 -3.88328880e-01 -1.31191146e+00 -3.08576852e-01
-4.72147912e-01 7.61681676e-01 5.69121718e-01 6.21996224e-01
9.00476277e-01 3.01533818e-01 4.86596972e-01 -5.08305371e-01
-6.22771502e-01 -7.63190389e-01 -1.06683686e-01 -2.51258407e-02
2.39036128e-01 -1.33229017e-01 -2.41080984e-01 -2.63560265e-01] | [11.543680191040039, 7.657987117767334] |
972886e5-9b5b-4cf9-abad-5a4d9b90ec27 | resfpn-residual-skip-connections-in-multi | 2006.12235 | null | https://arxiv.org/abs/2006.12235v1 | https://arxiv.org/pdf/2006.12235v1.pdf | ResFPN: Residual Skip Connections in Multi-Resolution Feature Pyramid Networks for Accurate Dense Pixel Matching | Dense pixel matching is required for many computer vision algorithms such as disparity, optical flow or scene flow estimation. Feature Pyramid Networks (FPN) have proven to be a suitable feature extractor for CNN-based dense matching tasks. FPN generates well localized and semantically strong features at multiple scales. However, the generic FPN is not utilizing its full potential, due to its reasonable but limited localization accuracy. Thus, we present ResFPN -- a multi-resolution feature pyramid network with multiple residual skip connections, where at any scale, we leverage the information from higher resolution maps for stronger and better localized features. In our ablation study, we demonstrate the effectiveness of our novel architecture with clearly higher accuracy than FPN. In addition, we verify the superior accuracy of ResFPN in many different pixel matching applications on established datasets like KITTI, Sintel, and FlyingThings3D. | ['Oliver Wasenmüller', 'René Schuster', 'Ramy Battrawy', 'Didier Stricker', 'Rishav'] | 2020-06-22 | null | null | null | null | ['scene-flow-estimation'] | ['computer-vision'] | [-3.14862505e-02 -4.22934413e-01 -1.62472963e-01 -2.57271886e-01
-5.87982535e-01 -2.29461566e-01 5.82909048e-01 -2.96285599e-01
-3.78556728e-01 6.59635067e-01 4.71146464e-01 1.74910963e-01
-6.10447153e-02 -8.56545031e-01 -5.90746462e-01 -2.54428387e-01
2.33705938e-02 -8.61612037e-02 5.80948472e-01 -2.95838088e-01
2.82868445e-01 7.64536560e-01 -1.85140479e+00 2.33498722e-01
7.69765198e-01 1.10339236e+00 2.01258108e-01 3.25208843e-01
7.37152621e-02 7.19047010e-01 -2.19862923e-01 -2.52796471e-01
7.12343335e-01 -1.05853379e-01 -6.90164030e-01 -3.03848565e-01
1.26315701e+00 -7.15297937e-01 -7.54031360e-01 8.34005237e-01
4.94852602e-01 1.58736140e-01 1.67766631e-01 -1.19530320e+00
-5.08855522e-01 3.92166317e-01 -6.88243866e-01 4.36825782e-01
4.09240305e-01 4.85006720e-01 1.11973417e+00 -1.10988224e+00
7.25012779e-01 1.41235554e+00 8.86699617e-01 2.49406174e-01
-1.17837596e+00 -9.02013898e-01 -9.97563377e-02 7.23849908e-02
-1.32482648e+00 -5.15771925e-01 5.24151504e-01 -3.67771655e-01
1.18243206e+00 -2.02817768e-01 8.09896529e-01 8.74495924e-01
2.02075109e-01 6.88366473e-01 9.65743363e-01 5.22413366e-02
-1.31268859e-01 -5.03702104e-01 -9.37322825e-02 8.69613647e-01
4.07544583e-01 6.96572959e-01 -1.03509748e+00 1.33444697e-01
1.35911953e+00 2.43277192e-01 -4.54772264e-01 -2.81331688e-01
-1.48070610e+00 6.53235853e-01 1.21800637e+00 3.08777422e-01
-4.14556593e-01 7.04041541e-01 8.76979064e-03 2.64851838e-01
2.82704115e-01 5.44696152e-01 -2.98780322e-01 -1.74924165e-01
-1.03988314e+00 3.77215862e-01 2.29490235e-01 8.43480766e-01
1.26650786e+00 1.87347546e-01 -2.75673926e-01 4.46651995e-01
8.52144137e-02 4.96312886e-01 3.13775420e-01 -1.28108311e+00
3.58675092e-01 6.99263096e-01 1.71666443e-01 -1.21771491e+00
-5.15655637e-01 -4.74259734e-01 -8.05605650e-01 5.10442019e-01
4.43306804e-01 7.86414146e-02 -9.31117773e-01 1.74317694e+00
1.25199854e-01 5.67336202e-01 -9.69889238e-02 1.23217094e+00
1.03513932e+00 3.50921810e-01 -7.73244724e-02 5.40845513e-01
9.16812658e-01 -1.00814331e+00 -2.61464745e-01 -3.52864712e-01
3.35461795e-01 -6.68968201e-01 8.11266422e-01 -1.26163866e-02
-1.00602746e+00 -7.89572001e-01 -1.00881732e+00 -4.70725000e-01
-2.81228334e-01 -4.51654419e-02 1.16979730e+00 2.58827031e-01
-1.39409316e+00 8.26814532e-01 -8.43590617e-01 -4.52835560e-01
7.63917208e-01 4.96568710e-01 -7.66745925e-01 -4.18575555e-01
-1.08948767e+00 7.57978141e-01 -4.67054062e-02 2.57104397e-01
-6.93129838e-01 -1.11990106e+00 -1.18907356e+00 1.74927890e-01
-3.36176246e-01 -1.11178434e+00 8.83519292e-01 -4.65299189e-01
-1.31391716e+00 8.78241837e-01 -2.29072183e-01 -6.69652641e-01
3.72947067e-01 -3.06438237e-01 1.96088329e-02 3.89746189e-01
1.81607857e-01 1.47082078e+00 7.41502881e-01 -7.16335058e-01
-8.29979360e-01 -3.70521873e-01 3.69331568e-01 1.19944073e-01
-1.17963962e-01 -2.07411513e-01 -2.09926903e-01 -3.86898041e-01
3.17213327e-01 -4.97961342e-01 -3.14522386e-01 5.63840568e-01
-6.81574494e-02 6.91042468e-02 7.76655436e-01 -3.21211576e-01
6.75671518e-01 -2.38055944e+00 -8.52942765e-02 -5.29973470e-02
5.58358908e-01 1.78332224e-01 -4.42659855e-01 8.17840695e-02
2.00003237e-01 -1.37531281e-01 -6.58458993e-02 -2.67867267e-01
-1.65160313e-01 -5.60547924e-04 -1.92972794e-01 5.83644032e-01
5.84959090e-01 1.31543684e+00 -9.67952728e-01 -3.85607660e-01
6.85056090e-01 7.92481840e-01 -7.47501135e-01 -7.46337250e-02
1.27351791e-01 4.60648000e-01 -5.26541233e-01 7.61169255e-01
7.70843923e-01 -3.54943901e-01 -5.91421306e-01 -6.17208719e-01
-3.13691348e-01 8.97455886e-02 -1.05832982e+00 2.29458022e+00
-5.28450668e-01 1.01757181e+00 3.51124443e-02 -5.00662982e-01
9.27181780e-01 -1.95584491e-01 6.90895796e-01 -1.03916860e+00
5.10933250e-02 3.59024882e-01 -3.17453332e-02 3.94493863e-02
7.23248899e-01 1.75937608e-01 2.46301353e-01 -1.70385409e-02
2.09808007e-01 -1.97029665e-01 1.81716457e-01 1.80162326e-01
1.26536071e+00 2.36629382e-01 1.48174375e-01 -3.71311903e-01
3.91595125e-01 1.14668207e-03 5.79287946e-01 8.51715624e-01
-4.73270208e-01 1.10306418e+00 -4.08486184e-03 -6.47392809e-01
-7.37216711e-01 -1.05756891e+00 -5.08609474e-01 6.09470308e-01
5.86182714e-01 -3.55295092e-01 -1.16603196e-01 -2.63623089e-01
4.33215559e-01 -2.00000212e-01 -5.48070252e-01 -2.53344718e-02
-5.20027518e-01 -2.63483346e-01 5.52641034e-01 6.40292883e-01
1.12636411e+00 -1.06083155e+00 -9.86641526e-01 2.87922233e-01
-7.99397081e-02 -1.41564071e+00 -4.85550612e-01 -2.10792907e-02
-8.36077392e-01 -1.16074705e+00 -7.01002657e-01 -7.78545439e-01
4.31842893e-01 6.38226926e-01 1.19601536e+00 1.36778623e-01
-5.61805129e-01 2.68050611e-01 -2.09715530e-01 1.14737630e-01
2.02007547e-01 1.09993324e-01 -2.12524921e-01 -1.95158496e-01
2.19413161e-01 -8.59716773e-01 -1.06432235e+00 4.04647499e-01
-7.42386103e-01 1.63772166e-01 5.70660174e-01 7.93188214e-01
4.71906692e-01 -2.70381927e-01 3.00818294e-01 -3.97605062e-01
2.48551905e-01 -1.95257133e-03 -8.08375657e-01 -1.31673202e-01
-2.78823644e-01 4.41153087e-02 3.67804796e-01 -4.82604234e-03
-6.81797683e-01 7.61667341e-02 -1.55906260e-01 -5.73383391e-01
-3.45214196e-02 1.15254633e-01 2.38472953e-01 -7.49651432e-01
7.50214219e-01 -2.71466700e-03 -6.15414372e-03 -2.58300245e-01
2.93440193e-01 1.13827407e-01 7.18854249e-01 -4.21610355e-01
1.01943767e+00 7.93591142e-01 1.91598982e-01 -5.75955629e-01
-7.91432679e-01 -5.39139807e-01 -6.57073021e-01 2.56881844e-02
7.57610381e-01 -1.27742994e+00 -8.77537727e-01 6.24506354e-01
-1.02144349e+00 -3.04184586e-01 -4.43308383e-01 6.19061828e-01
-4.25174803e-01 1.19861811e-01 -7.22165227e-01 -1.80384904e-01
-5.26713133e-01 -1.14181352e+00 1.39107239e+00 6.29063010e-01
8.85219499e-03 -9.10950899e-01 -3.70664801e-03 1.09175682e-01
9.97906506e-01 3.98315698e-01 3.22177619e-01 -5.48550021e-03
-1.04085386e+00 -2.07108185e-01 -8.22181344e-01 6.32201433e-02
2.84680635e-01 4.44883071e-02 -1.16407883e+00 -3.72919381e-01
-6.40057027e-01 -5.10733128e-01 1.27643919e+00 5.23213983e-01
9.78612185e-01 2.90114284e-01 -2.23133683e-01 1.27504468e+00
1.59397519e+00 -3.84663552e-01 6.73576772e-01 4.18632686e-01
6.97669208e-01 2.06986815e-01 3.88785988e-01 3.18328083e-01
3.32512200e-01 6.29699111e-01 5.64059436e-01 -5.95799327e-01
-7.88889229e-01 -3.68651450e-01 3.29167724e-01 2.27459759e-01
1.58439111e-02 2.15033039e-01 -6.41221762e-01 5.16982853e-01
-1.76141727e+00 -9.46467340e-01 1.23501033e-01 2.02303362e+00
4.77878600e-01 5.26650175e-02 -2.54506290e-01 -2.60152280e-01
5.46380460e-01 4.79552895e-01 -5.88503659e-01 1.10023193e-01
-5.84338069e-01 5.47993243e-01 4.31939781e-01 4.64836597e-01
-1.10563576e+00 1.15045416e+00 6.29234982e+00 4.65941787e-01
-1.28428590e+00 -1.74559146e-01 3.81252497e-01 -7.12611750e-02
-3.61264497e-01 9.27633718e-02 -8.57572317e-01 1.99576646e-01
3.03489655e-01 -1.02552310e-01 2.85382420e-01 6.47368193e-01
5.97633794e-02 -2.62306243e-01 -1.11719429e+00 1.31290591e+00
-1.73784837e-01 -1.82516515e+00 9.84586626e-02 1.09672993e-01
9.13276613e-01 6.17368698e-01 2.02907249e-01 1.70558855e-01
3.13158095e-01 -1.08154941e+00 4.13290203e-01 2.38535613e-01
8.47471476e-01 -4.31204319e-01 6.52561784e-01 -2.08102435e-01
-1.43824494e+00 7.02094063e-02 -6.78354144e-01 -1.83850542e-01
1.28903300e-01 7.14950383e-01 -3.57267886e-01 4.58881348e-01
8.43328834e-01 1.45723033e+00 -8.34140837e-01 1.34980178e+00
2.22203205e-03 1.75924432e-02 -6.28335595e-01 4.17793095e-01
4.21218604e-01 -8.01366940e-02 1.74820960e-01 1.01881993e+00
3.94688755e-01 -2.43051574e-01 1.50457725e-01 1.12785053e+00
-2.71471143e-01 -1.03358269e-01 -8.65609765e-01 3.58104527e-01
4.21645612e-01 1.49107373e+00 -4.63707477e-01 -3.01598329e-02
-6.18732274e-01 8.48405778e-01 5.27934968e-01 3.98912340e-01
-5.52182674e-01 -3.33320320e-01 1.22274506e+00 1.83870405e-01
4.18185085e-01 -2.93291867e-01 -8.34522247e-02 -1.42268443e+00
-1.22109517e-01 -3.65132362e-01 1.73279479e-01 -9.58115697e-01
-1.25581253e+00 6.57698572e-01 -3.52855384e-01 -1.29597366e+00
-1.34659320e-01 -6.18309855e-01 -6.11079395e-01 9.50919688e-01
-2.15816975e+00 -1.22563481e+00 -9.30364966e-01 9.94888961e-01
3.36689860e-01 4.79669310e-02 6.02282405e-01 4.72140670e-01
-3.99440229e-01 5.13358414e-01 -2.38945052e-01 3.56942505e-01
7.17030346e-01 -9.89793062e-01 6.21739507e-01 1.09158134e+00
2.47375250e-01 5.41801274e-01 2.31911391e-01 -4.07221049e-01
-1.35827696e+00 -1.14320457e+00 5.87612391e-01 -3.14771742e-01
5.90891659e-01 -1.41884968e-01 -6.67878211e-01 4.74412441e-01
9.52937305e-02 7.30749011e-01 9.54309031e-02 -8.28191563e-02
-4.79336381e-01 -4.56603914e-01 -1.20520794e+00 3.78338993e-01
1.40879393e+00 -6.07111037e-01 -2.03501493e-01 8.74438211e-02
7.87544608e-01 -5.31503797e-01 -9.84789848e-01 6.88542664e-01
6.74100339e-01 -1.49474573e+00 1.21279621e+00 -2.03150839e-01
6.39541090e-01 -4.00660276e-01 -3.47600520e-01 -1.06190860e+00
-6.14432395e-01 -5.25774479e-01 1.58418432e-01 1.03778183e+00
2.00919002e-01 -7.73005068e-01 9.27419364e-01 4.12321210e-01
-1.74313053e-01 -4.72471535e-01 -1.05899882e+00 -6.36954486e-01
-9.84074920e-02 -3.31145406e-01 6.80640340e-01 7.30415702e-01
-3.11306655e-01 8.20413679e-02 -2.59604961e-01 5.29133528e-02
7.87465215e-01 5.97086191e-01 9.20824170e-01 -1.16225994e+00
2.90237367e-02 -5.61481833e-01 -9.81954932e-01 -1.37524974e+00
1.21552110e-01 -8.34326684e-01 -6.32353052e-02 -1.60447967e+00
5.84358908e-02 -5.30078948e-01 -3.84190500e-01 4.53553349e-01
-1.02742940e-01 8.01108122e-01 2.69448549e-01 1.34615391e-01
-6.24111712e-01 6.65774941e-01 1.62878537e+00 -1.11917324e-01
-1.08775627e-02 -2.67109007e-01 -4.95471865e-01 4.60672975e-01
6.56055808e-01 -6.79280162e-02 -2.60460585e-01 -7.30001509e-01
-4.03003544e-02 -4.32203487e-02 7.11468279e-01 -1.38234484e+00
3.61504525e-01 6.23442456e-02 6.65672779e-01 -5.31197011e-01
3.95995319e-01 -7.40737557e-01 -5.67643493e-02 5.40974796e-01
-8.73660017e-03 1.88671961e-01 2.29057014e-01 2.81923652e-01
-6.30530536e-01 4.40606952e-01 7.19030917e-01 -1.02994211e-01
-1.29862583e+00 8.82825553e-01 1.45178661e-01 1.47782207e-01
6.31765783e-01 -4.72752124e-01 -6.23378217e-01 -2.89568037e-01
-1.51148334e-01 2.93289691e-01 7.31609404e-01 6.22596502e-01
9.80018795e-01 -1.39654064e+00 -7.51791358e-01 6.00215554e-01
2.43361980e-01 2.78891236e-01 1.54100686e-01 8.54046822e-01
-8.05908203e-01 5.02728879e-01 -6.26798749e-01 -9.05794263e-01
-6.88911736e-01 2.33497024e-01 6.00577235e-01 -3.08854654e-02
-9.87643361e-01 8.28264356e-01 4.14786309e-01 -1.92719817e-01
8.41471851e-02 -6.20626569e-01 4.49662737e-04 -3.59424323e-01
4.81850684e-01 1.78755343e-01 5.43276258e-02 -6.13711417e-01
-4.33831692e-01 1.04212379e+00 1.37041748e-01 2.12958992e-01
1.33671582e+00 -1.73041433e-01 7.88697321e-03 -1.20023772e-01
1.27458525e+00 -2.05650672e-01 -1.70738780e+00 -4.57239538e-01
-3.36329192e-01 -9.33251262e-01 2.29208842e-01 -3.17886591e-01
-1.55397141e+00 8.83983672e-01 6.10083044e-01 -3.50783318e-01
1.08191717e+00 -7.97040313e-02 9.14484560e-01 2.39174202e-01
8.43052983e-01 -4.61525083e-01 -2.69821770e-02 3.79103184e-01
7.19708920e-01 -1.59939289e+00 -4.52718213e-02 -3.07109743e-01
-5.12396619e-02 1.20197940e+00 9.95874703e-01 -4.39032316e-01
6.70527220e-01 3.60113263e-01 2.78689545e-02 -2.91491628e-01
-5.76152742e-01 -6.07487917e-01 2.32448444e-01 7.40897894e-01
2.47496367e-01 -3.41725707e-01 1.24589838e-01 -1.59717470e-01
-2.78661877e-01 1.09717868e-01 3.07921737e-01 6.79399014e-01
-4.21429783e-01 -7.94597805e-01 2.27470715e-02 3.92183781e-01
-1.34343341e-01 -3.54960173e-01 -1.18215084e-02 8.34813774e-01
6.58482835e-02 7.24109590e-01 2.62564510e-01 -4.97627527e-01
4.30475324e-01 -5.32389700e-01 6.31757975e-01 -3.58439207e-01
-7.27629185e-01 -2.84032941e-01 -2.55266219e-01 -1.36208904e+00
-7.64946580e-01 -3.94979686e-01 -1.20654142e+00 -5.57847142e-01
-6.10541552e-02 -3.20807129e-01 3.77670079e-01 6.88451886e-01
5.75492799e-01 3.10419202e-01 6.10355973e-01 -1.13981497e+00
-4.34131622e-02 -7.04763234e-01 -5.04574955e-01 6.15152657e-01
5.85597873e-01 -7.54157841e-01 -4.08805162e-01 -3.09268355e-01] | [8.8681640625, -2.210829734802246] |
70fa5b58-be0d-44c0-a622-fbbf5bb61099 | multi-channel-nuclear-norm-minus-frobenius | 2209.08094 | null | https://arxiv.org/abs/2209.08094v1 | https://arxiv.org/pdf/2209.08094v1.pdf | Multi-channel Nuclear Norm Minus Frobenius Norm Minimization for Color Image Denoising | Color image denoising is frequently encountered in various image processing and computer vision tasks. One traditional strategy is to convert the RGB image to a less correlated color space and denoise each channel of the new space separately. However, such a strategy can not fully exploit the correlated information between channels and is inadequate to obtain satisfactory results. To address this issue, this paper proposes a new multi-channel optimization model for color image denoising under the nuclear norm minus Frobenius norm minimization framework. Specifically, based on the block-matching, the color image is decomposed into overlapping RGB patches. For each patch, we stack its similar neighbors to form the corresponding patch matrix. The proposed model is performed on the patch matrix to recover its noise-free version. During the recovery process, a) a weight matrix is introduced to fully utilize the noise difference between channels; b) the singular values are shrunk adaptively without additionally assigning weights. With them, the proposed model can achieve promising results while keeping simplicity. To solve the proposed model, an accurate and effective algorithm is built based on the alternating direction method of multipliers framework. The solution of each updating step can be analytically expressed in closed-from. Rigorous theoretical analysis proves the solution sequences generated by the proposed algorithm converge to their respective stationary points. Experimental results on both synthetic and real noise datasets demonstrate the proposed model outperforms state-of-the-art models. | ['Tao Jia', 'Zhi Wang', 'Dong Hu', 'Yiwen Shan'] | 2022-09-16 | null | null | null | null | ['color-image-denoising'] | ['computer-vision'] | [ 4.34608519e-01 -6.50553584e-01 2.21863762e-01 5.72522450e-03
-6.78311348e-01 -1.31769747e-01 3.70363854e-02 -3.49734902e-01
-5.57803512e-01 5.33360124e-01 -1.99903682e-01 8.27815756e-02
-1.71342924e-01 -6.24530971e-01 -3.88307601e-01 -1.32231462e+00
3.66264492e-01 -4.76376295e-01 -8.13687369e-02 -1.19176239e-01
2.58447379e-01 1.18880607e-01 -1.31643212e+00 -1.42450795e-01
1.23588741e+00 1.09749007e+00 4.56811488e-01 4.09218192e-01
-1.47550151e-01 4.45862204e-01 -1.98180333e-01 -1.50567321e-02
6.23860717e-01 -7.11500585e-01 -1.58044174e-01 6.25893295e-01
6.71681538e-02 -2.28988037e-01 -2.84337133e-01 1.69802654e+00
4.38715756e-01 3.14513266e-01 1.82231948e-01 -9.49279368e-01
-4.72475022e-01 3.70659828e-02 -1.18888617e+00 1.65989444e-01
-3.12204808e-02 7.57297277e-02 5.67459702e-01 -9.99107182e-01
4.43179756e-01 1.14774847e+00 6.50260389e-01 1.56694904e-01
-1.44813287e+00 -6.31302536e-01 2.45708376e-01 4.18447644e-01
-1.56302285e+00 -2.36469805e-01 1.24230778e+00 -1.30701251e-02
1.43026769e-01 3.88068140e-01 8.53339374e-01 4.69079703e-01
1.25470221e-01 5.14273882e-01 1.44704556e+00 -4.61016268e-01
8.63041542e-03 -9.47715789e-02 4.84342761e-02 6.40376151e-01
4.47864383e-01 -1.29057288e-01 -3.72654706e-01 -2.79205665e-02
7.41684020e-01 2.65356362e-01 -6.33890629e-01 -2.97344834e-01
-1.20433021e+00 5.22913933e-01 5.70121348e-01 4.56760108e-01
-5.77366769e-01 -5.62507622e-02 8.70764032e-02 1.05976008e-01
3.38523686e-01 -1.73393205e-01 3.82157043e-02 1.71941280e-01
-9.59030747e-01 2.70652696e-02 4.93816376e-01 6.78280473e-01
1.12191963e+00 2.74424732e-01 1.28301427e-01 9.58037138e-01
5.16673386e-01 6.29669368e-01 3.69921803e-01 -1.09292185e+00
5.16385555e-01 3.82964820e-01 6.43118024e-02 -1.61205626e+00
-1.47010028e-01 -6.47679150e-01 -1.55878484e+00 2.54515320e-01
4.29537654e-01 -3.72767001e-02 -6.37926996e-01 1.41955578e+00
5.00253916e-01 5.80875814e-01 5.42802736e-02 1.14629149e+00
3.24742943e-01 1.03319275e+00 -1.79865450e-01 -8.30324888e-01
1.15606773e+00 -5.90355933e-01 -1.10554099e+00 -2.44054034e-01
-1.73480719e-01 -1.08504689e+00 6.80064201e-01 8.47744584e-01
-9.55342114e-01 -7.24478841e-01 -1.21278942e+00 1.69939905e-01
1.78438425e-01 2.85370409e-01 2.55297393e-01 6.63964868e-01
-7.50380635e-01 4.30232912e-01 -7.31022954e-01 9.07130190e-04
2.33291630e-02 -2.44153496e-02 -3.18742841e-01 -3.76765847e-01
-8.43380809e-01 6.05178595e-01 2.40516469e-01 9.80889380e-01
-4.63010281e-01 -2.52708256e-01 -5.27390599e-01 -1.68790683e-01
4.36757922e-01 -4.86770809e-01 5.80767870e-01 -1.20641696e+00
-1.47226822e+00 3.91363055e-01 -4.32299376e-01 6.80259541e-02
3.81578892e-01 -3.15355808e-02 -4.64805424e-01 2.76709110e-01
1.51523143e-01 -8.89462754e-02 1.37167454e+00 -1.60109580e+00
-5.67360580e-01 -4.14281398e-01 -5.19496024e-01 2.93881238e-01
-5.12810230e-01 -3.09547275e-01 -8.47358644e-01 -8.75149786e-01
9.67224240e-01 -8.39553475e-01 -5.48384786e-01 9.54101756e-02
-3.22999239e-01 5.22332907e-01 6.78732932e-01 -1.00705588e+00
1.40567780e+00 -2.39338040e+00 4.21958685e-01 6.35356724e-01
8.95487368e-02 1.09089807e-01 -1.41019061e-01 1.26159966e-01
-1.34044200e-01 -3.90550315e-01 -6.38206005e-01 -8.01829770e-02
-4.24919814e-01 2.09288001e-01 1.50392592e-01 8.86969507e-01
-1.10921122e-01 1.26705363e-01 -7.86011994e-01 -4.59711999e-01
3.94700378e-01 7.32633948e-01 -4.47741240e-01 1.06593579e-01
4.70657736e-01 6.90553427e-01 -4.95207816e-01 5.88817596e-01
1.34308279e+00 1.28794283e-01 2.89333969e-01 -8.54604900e-01
-3.57428432e-01 -6.78806245e-01 -2.01184869e+00 1.57711494e+00
-3.58905226e-01 2.23543569e-01 9.25322175e-01 -1.36810338e+00
9.84440565e-01 1.04305826e-01 6.30141377e-01 -6.84130609e-01
1.56712368e-01 3.76571417e-01 -1.33970156e-01 -5.62783420e-01
2.02078313e-01 -4.05312866e-01 3.49368423e-01 3.55188474e-02
-3.25208127e-01 -5.42504154e-02 2.58307666e-01 -2.62786262e-02
6.34945333e-01 -1.30451075e-03 2.54262120e-01 -2.19939917e-01
1.08004940e+00 -1.83190212e-01 1.06780148e+00 5.40921628e-01
-3.24538141e-01 5.93400002e-01 1.94492377e-02 -3.19910526e-01
-7.23497748e-01 -8.36266518e-01 -7.20977858e-02 4.64411110e-01
6.65325284e-01 -1.07742131e-01 -7.89442718e-01 -1.51481718e-01
-2.36137390e-01 1.98439553e-01 -4.20797735e-01 -9.91354808e-02
-6.34414852e-01 -1.14669979e+00 -4.43015583e-02 -6.84791803e-02
1.09682727e+00 -5.35171688e-01 -3.63172799e-01 4.52406198e-01
-5.09672403e-01 -7.13344097e-01 -4.67811435e-01 -1.47080705e-01
-1.08044064e+00 -1.07272279e+00 -8.74155641e-01 -8.70841265e-01
8.91828895e-01 9.50569689e-01 4.41346675e-01 2.70287037e-01
-2.79515028e-01 3.58845860e-01 -4.25991893e-01 9.38406959e-02
-1.49974301e-01 -5.63211739e-01 -1.28011629e-01 7.93350518e-01
-8.02632247e-04 -6.89707160e-01 -8.23562980e-01 3.33340377e-01
-1.13287163e+00 2.18639113e-02 8.68863881e-01 1.14554536e+00
9.34733391e-01 5.98119259e-01 1.52599424e-01 -3.95172954e-01
5.78641295e-01 -3.18056136e-01 -7.88158655e-01 1.81892484e-01
-6.81082070e-01 -6.09885752e-02 6.73819065e-01 -3.00355911e-01
-1.36764455e+00 3.30914229e-01 1.67127457e-02 -2.57140636e-01
1.38407990e-01 5.57083070e-01 -3.23451132e-01 -4.02762204e-01
1.01095185e-01 7.29703188e-01 3.96234900e-01 -6.79120660e-01
4.52622950e-01 4.24493998e-01 6.99004531e-01 -4.30098981e-01
1.15033054e+00 7.73265660e-01 2.21181646e-01 -1.06930029e+00
-3.83515984e-01 -7.09966660e-01 -4.09584582e-01 -5.10151029e-01
7.71821856e-01 -8.28151107e-01 -7.19361663e-01 7.16196120e-01
-9.21873212e-01 3.31869751e-01 2.29682863e-01 5.22561729e-01
-1.75425157e-01 9.33264911e-01 -5.84387779e-01 -9.25842941e-01
-2.37701312e-01 -1.16539335e+00 5.63521266e-01 4.55057919e-01
5.13644993e-01 -7.05848932e-01 -1.49810528e-02 2.45628223e-01
4.97321934e-01 1.70943365e-01 6.04097664e-01 4.01616216e-01
-5.97132206e-01 -2.85733432e-01 -2.84435332e-01 7.47063637e-01
3.16379905e-01 -6.53097555e-02 -5.51344514e-01 -3.19853842e-01
6.03867412e-01 2.46470258e-01 9.20774221e-01 4.02815670e-01
1.11052036e+00 -2.29963854e-01 -1.04524538e-01 8.18357825e-01
1.97149193e+00 2.64353752e-01 7.07835793e-01 6.66162610e-01
6.19592488e-01 4.46776450e-01 6.18373930e-01 5.89758456e-01
8.88269395e-02 4.77189034e-01 5.46608090e-01 -1.98486000e-01
6.23481274e-02 1.63068146e-01 2.34257191e-01 1.14016235e+00
-2.87365407e-01 2.01975897e-01 -5.05101860e-01 3.40208352e-01
-1.82505608e+00 -9.67074096e-01 -4.18713629e-01 2.24132347e+00
7.97370315e-01 -1.77537560e-01 -3.67385715e-01 4.98987854e-01
8.31680477e-01 2.43833959e-01 -4.07365441e-01 6.56288043e-02
-3.99212182e-01 9.71037149e-02 5.30513406e-01 4.37934846e-01
-1.07351780e+00 3.37315470e-01 5.14917803e+00 9.41534698e-01
-1.05526519e+00 1.18629456e-01 5.23209691e-01 2.38572761e-01
-1.35669127e-01 2.03176424e-01 -2.44843185e-01 5.39982736e-01
2.52272636e-01 -6.77969158e-02 7.70984828e-01 4.41850811e-01
5.79402208e-01 -3.90143275e-01 -2.12542713e-01 1.38057923e+00
4.33882736e-02 -8.56063247e-01 -3.50511633e-02 -1.53692827e-01
7.44397223e-01 -6.91816270e-01 1.12045854e-01 -1.60721257e-01
-3.19955766e-01 -5.11691749e-01 5.98504901e-01 8.89616787e-01
3.79375577e-01 -8.16462755e-01 7.22780585e-01 2.52453566e-01
-1.29540730e+00 -2.33266667e-01 -4.69077975e-01 -9.90280043e-03
1.75507575e-01 8.88972104e-01 9.88572910e-02 9.28444564e-01
8.18130970e-01 9.17152643e-01 -3.33096892e-01 1.21344960e+00
-1.47115752e-01 5.79603136e-01 -2.63253480e-01 3.59076738e-01
2.40570888e-01 -1.26056278e+00 7.48264253e-01 8.72565389e-01
5.79285741e-01 4.56255198e-01 1.59804940e-01 5.67730725e-01
2.78850615e-01 3.22352111e-01 -1.37353778e-01 4.49786782e-01
2.18733907e-01 1.52155221e+00 -7.29863107e-01 -2.25534931e-01
-6.98339880e-01 1.16913915e+00 -1.97523981e-01 7.77405143e-01
-5.77175260e-01 -4.80539978e-01 4.92917717e-01 -2.17989445e-01
2.86716014e-01 -2.29099125e-01 -4.40725923e-01 -1.13264120e+00
2.63261646e-01 -1.11433029e+00 2.35939711e-01 -5.57260036e-01
-1.19813156e+00 4.35301989e-01 -4.08838183e-01 -1.75774062e+00
3.73305976e-01 -4.37183082e-01 -6.01357996e-01 9.89070952e-01
-1.61765420e+00 -8.17830443e-01 -6.22811973e-01 8.67374003e-01
2.91765898e-01 8.90520681e-03 4.12711710e-01 5.47685444e-01
-1.08670270e+00 2.04854280e-01 6.58947885e-01 -1.15341224e-01
6.85519695e-01 -8.92305732e-01 -4.95661765e-01 1.47304463e+00
-2.66816467e-01 7.52899408e-01 8.39339077e-01 -5.32635272e-01
-1.65410161e+00 -8.39570343e-01 1.26044184e-01 5.75650513e-01
4.70523447e-01 2.26225898e-01 -1.07139742e+00 2.10795656e-01
2.98480421e-01 1.96102872e-01 4.00080204e-01 -4.54853415e-01
-1.12495132e-01 -7.06699729e-01 -1.08387959e+00 4.46610510e-01
4.73898262e-01 -1.58364162e-01 -2.98996598e-01 4.04683389e-02
1.00018077e-01 -1.69527099e-01 -7.66762733e-01 2.47752994e-01
4.79750216e-01 -1.11889398e+00 1.14426160e+00 4.56541777e-02
1.32307649e-01 -9.17636096e-01 -2.59100616e-01 -1.28479218e+00
-4.81503427e-01 -6.47400975e-01 1.16139956e-01 1.16589367e+00
2.94526126e-02 -6.58423603e-01 4.96268153e-01 5.26467800e-01
7.09184259e-03 -5.19200087e-01 -1.06739843e+00 -6.26383424e-01
-3.62249255e-01 -2.71315992e-01 1.07765332e-01 7.74886966e-01
-2.87826985e-01 -9.75173041e-02 -5.90200424e-01 4.52185720e-01
1.32592344e+00 1.52991474e-01 5.99133313e-01 -8.89705539e-01
-1.98662281e-01 -3.97404820e-01 -2.03020528e-01 -9.68694925e-01
-1.98659837e-01 -4.82957751e-01 3.02642941e-01 -1.52219188e+00
3.12110603e-01 -3.82359982e-01 -4.57717061e-01 1.96630582e-02
-6.65102959e-01 5.54865539e-01 2.88185984e-01 3.50572228e-01
-3.56509060e-01 6.93913341e-01 1.24480236e+00 -2.71585912e-01
-1.61440566e-01 1.69925310e-03 -6.21247172e-01 7.42203414e-01
6.33198798e-01 -2.46474117e-01 -1.38473049e-01 -4.62148935e-01
1.94036830e-02 1.57475963e-01 3.43414038e-01 -1.05120361e+00
3.46337676e-01 -1.77198187e-01 5.04082382e-01 -3.38195145e-01
3.35138977e-01 -1.20911574e+00 2.79388040e-01 5.21800816e-01
2.11698681e-01 -2.26060882e-01 6.54920712e-02 9.61523831e-01
-6.84359252e-01 -2.02754229e-01 1.01107669e+00 -1.88608021e-01
-8.17578912e-01 1.07892714e-01 -4.73497748e-01 -3.33178461e-01
8.91933262e-01 -4.97259378e-01 1.31370381e-01 -3.29520136e-01
-6.83776736e-01 9.68007669e-02 3.07967961e-01 -1.18166238e-01
8.96921277e-01 -1.39871252e+00 -7.05182135e-01 3.96002710e-01
-2.38902405e-01 -3.00828129e-01 7.99595833e-01 1.09965587e+00
-6.05073810e-01 -1.76657557e-01 -2.18216792e-01 -5.84625900e-01
-1.15326762e+00 5.06468534e-01 3.42447162e-01 -9.17279199e-02
-4.67752934e-01 5.97135842e-01 -5.71493283e-02 -2.66364273e-02
-4.15643619e-04 -5.75255454e-02 -2.15386793e-01 1.39167622e-01
6.77659690e-01 6.98622763e-01 -9.99457315e-02 -9.05645907e-01
-2.10626960e-01 1.06033778e+00 3.25871855e-01 -2.50518233e-01
1.52444005e+00 -5.91989517e-01 -8.13092828e-01 7.66936764e-02
1.39187145e+00 1.87105730e-01 -1.25428915e+00 -2.99592257e-01
-3.37721407e-01 -8.89080882e-01 4.94126201e-01 -2.12984011e-01
-1.50280583e+00 5.92361093e-01 9.01188612e-01 1.80605128e-01
1.86083436e+00 -8.33005130e-01 8.29132199e-01 9.87787768e-02
1.48511603e-01 -1.19999981e+00 2.55608242e-02 -6.40123188e-02
7.31236637e-01 -1.15801978e+00 2.69731253e-01 -6.24232650e-01
-2.93502897e-01 1.29109073e+00 4.23907548e-01 -2.21511483e-01
6.30841076e-01 -1.87343746e-01 1.87621251e-01 1.39477804e-01
9.68019199e-03 -2.21531585e-01 9.70433652e-02 3.28844309e-01
1.11560889e-01 -1.61472589e-01 -6.98448658e-01 6.02824569e-01
5.20529747e-01 -1.22321673e-01 5.78417063e-01 8.41510296e-01
-5.54473102e-01 -1.06724358e+00 -1.15306330e+00 2.26940751e-01
-3.97924811e-01 -1.10294353e-02 3.00111234e-01 4.80248839e-01
1.37324810e-01 1.25450718e+00 -3.64396900e-01 -2.34293774e-01
4.46511090e-01 -2.66826630e-01 3.32745552e-01 -4.75822762e-02
-1.35087237e-01 6.80923462e-01 -4.56737906e-01 -6.36765480e-01
-7.68740237e-01 -7.05686510e-01 -9.93614435e-01 -3.40219215e-02
-3.48034531e-01 2.62520224e-01 6.28021061e-01 5.51005542e-01
-3.29936258e-02 4.13329750e-01 9.25938129e-01 -1.06972778e+00
-5.40266037e-01 -6.59166634e-01 -9.59679365e-01 4.80733931e-01
3.64762574e-01 -4.62605566e-01 -6.62639618e-01 1.77762553e-01] | [11.234712600708008, -2.453395366668701] |
bedffe38-3ab9-41b3-b51d-91ab620ea184 | code2seq-generating-sequences-from-structured | 1808.01400 | null | http://arxiv.org/abs/1808.01400v6 | http://arxiv.org/pdf/1808.01400v6.pdf | code2seq: Generating Sequences from Structured Representations of Code | The ability to generate natural language sequences from source code snippets
has a variety of applications such as code summarization, documentation, and
retrieval. Sequence-to-sequence (seq2seq) models, adopted from neural machine
translation (NMT), have achieved state-of-the-art performance on these tasks by
treating source code as a sequence of tokens. We present ${\rm {\scriptsize
CODE2SEQ}}$: an alternative approach that leverages the syntactic structure of
programming languages to better encode source code. Our model represents a code
snippet as the set of compositional paths in its abstract syntax tree (AST) and
uses attention to select the relevant paths while decoding. We demonstrate the
effectiveness of our approach for two tasks, two programming languages, and
four datasets of up to $16$M examples. Our model significantly outperforms
previous models that were specifically designed for programming languages, as
well as state-of-the-art NMT models. An interactive online demo of our model is
available at http://code2seq.org. Our code, data and trained models are
available at http://github.com/tech-srl/code2seq. | ['Uri Alon', 'Eran Yahav', 'Shaked Brody', 'Omer Levy'] | 2018-08-04 | code2seq-generating-sequences-from-structured-1 | https://openreview.net/forum?id=H1gKYo09tX | https://openreview.net/pdf?id=H1gKYo09tX | iclr-2019-5 | ['code-summarization'] | ['computer-code'] | [ 4.93474811e-01 8.52460414e-02 -4.76140708e-01 -4.40328121e-01
-1.47537315e+00 -8.33427608e-01 3.41260225e-01 1.72475323e-01
1.29491478e-01 3.92538100e-01 4.29963350e-01 -8.82955015e-01
3.90871406e-01 -4.72196966e-01 -9.97958660e-01 1.13423757e-01
-1.09048992e-01 -2.00474216e-03 4.18683626e-02 -3.18723202e-01
8.39396656e-01 -3.55508298e-01 -1.29316723e+00 9.40131724e-01
1.30947232e+00 1.45400599e-01 5.21496415e-01 9.27726805e-01
-8.51840734e-01 1.12924623e+00 -6.56529307e-01 -6.91070914e-01
2.70312969e-02 -6.99908435e-01 -9.87782061e-01 -4.91489619e-01
4.72515315e-01 -6.08041957e-02 -4.66340855e-02 1.36144865e+00
3.15272063e-01 -4.88750905e-01 2.94488430e-01 -1.02995074e+00
-9.83837187e-01 1.32284141e+00 -6.31369233e-01 2.09269315e-01
8.00891101e-01 3.61719221e-01 1.09856677e+00 -9.92283225e-01
7.48179793e-01 1.01410675e+00 7.21086562e-01 8.07963312e-01
-1.34748220e+00 -4.89240348e-01 -2.38355979e-01 -1.99880421e-01
-1.06042206e+00 -4.98719662e-01 4.02082980e-01 -7.30627000e-01
1.61946130e+00 1.39368907e-01 4.42005545e-01 1.04586840e+00
5.81701100e-01 1.07908070e+00 5.48649728e-01 -6.61521673e-01
-5.85510535e-03 -8.84024575e-02 1.86023608e-01 1.07356799e+00
-8.38582143e-02 -3.09762478e-01 -4.93541569e-01 -6.10538363e-01
3.27687889e-01 -2.75143802e-01 -1.09586246e-01 -6.11574501e-02
-1.37277091e+00 8.84759605e-01 -1.34497248e-02 1.28372356e-01
8.76100548e-03 7.89803624e-01 8.25748324e-01 4.08788353e-01
2.06636310e-01 5.59947073e-01 -4.82455522e-01 -8.66472304e-01
-1.04999626e+00 4.34304982e-01 1.05389595e+00 1.67496169e+00
6.24683201e-01 3.47874045e-01 -1.59097359e-01 1.07786512e+00
2.28713378e-01 3.64015996e-01 7.80005515e-01 -1.09732580e+00
1.02496457e+00 7.17335582e-01 -2.17921093e-01 -4.70623821e-01
1.70075223e-01 -1.86234042e-01 -2.21929044e-01 -1.09382600e-01
-3.61159109e-02 -2.03340366e-01 -6.41092181e-01 1.58082521e+00
-3.46207798e-01 -3.52667004e-01 7.54272863e-02 2.96475500e-01
7.51033366e-01 8.69544506e-01 -1.86338067e-01 1.66367859e-01
1.04790819e+00 -1.19559121e+00 -2.87272990e-01 -4.11210448e-01
1.14547527e+00 -1.03467357e+00 1.28693759e+00 1.81903124e-01
-1.47942412e+00 -3.11459750e-01 -9.09393191e-01 -1.98301420e-01
-5.16789071e-02 2.64412731e-01 5.91019452e-01 4.42892849e-01
-1.37832582e+00 5.56903780e-01 -8.10176492e-01 -4.38207924e-01
3.16325009e-01 5.19155450e-02 2.47458834e-02 -1.44494995e-01
-6.63348734e-01 4.71978903e-01 3.47584426e-01 -3.40318948e-01
-9.52446759e-01 -8.50474596e-01 -1.00759923e+00 8.80196691e-02
1.82135567e-01 -6.27953708e-01 1.97580194e+00 -1.14763939e+00
-1.39717948e+00 8.07385325e-01 -3.60931545e-01 -4.00414407e-01
2.27074042e-01 -3.12023461e-01 -2.43137315e-01 -1.95068955e-01
4.58293736e-01 5.24794102e-01 5.03152251e-01 -8.69612217e-01
-5.11761785e-01 2.03773707e-01 2.14615632e-02 -3.93959731e-01
-1.13348162e-03 6.86539590e-01 -4.39826071e-01 -8.17788064e-01
-3.84782821e-01 -9.94703412e-01 -3.01860988e-01 -2.82126665e-01
-4.22196060e-01 -1.95649117e-01 2.14464024e-01 -9.50780809e-01
1.68591404e+00 -2.16710401e+00 3.31035137e-01 -1.40459523e-01
-1.07914999e-01 2.11990565e-01 -4.87913162e-01 9.68044281e-01
-9.36753899e-02 6.12947166e-01 -8.28351438e-01 -1.04207778e-03
1.56169027e-01 -1.71644583e-01 -4.04376566e-01 1.34478547e-02
4.37227249e-01 1.09845352e+00 -1.06764829e+00 -4.29273427e-01
-4.73577410e-01 -1.65609904e-02 -9.94740903e-01 2.30808273e-01
-7.88056254e-01 2.12604776e-02 -4.98997331e-01 5.51344037e-01
1.99919909e-01 -2.88095236e-01 2.84247279e-01 6.75439179e-01
-4.63079870e-01 8.76913786e-01 -5.53819060e-01 2.38589787e+00
-6.66497409e-01 7.90051460e-01 -2.07099706e-01 -6.41995192e-01
9.46041226e-01 1.80969089e-01 6.30025640e-02 -5.51658392e-01
-3.32570195e-01 6.34243906e-01 1.63416207e-01 -9.32267666e-01
6.95885956e-01 4.63014483e-01 -6.02158487e-01 6.68394625e-01
9.92141739e-02 -4.98658270e-01 6.73847556e-01 5.02230287e-01
1.54975247e+00 7.22686410e-01 3.81326318e-01 -4.29381073e-01
3.15017641e-01 4.10863817e-01 5.03678083e-01 7.64664233e-01
3.23037207e-01 4.61156160e-01 9.33318615e-01 -2.33967707e-01
-1.48692667e+00 -8.56041908e-01 2.24445403e-01 1.12200379e+00
-6.10904098e-01 -1.02192318e+00 -1.04387021e+00 -6.93666756e-01
-1.31443188e-01 1.09688878e+00 -3.57232630e-01 -5.98461106e-02
-1.11223388e+00 -4.57493186e-01 1.04523289e+00 5.09540379e-01
1.37520149e-01 -1.14373064e+00 -6.16457522e-01 3.41944247e-01
-3.92101794e-01 -4.75405037e-01 -9.74871814e-01 1.95970070e-02
-8.91403735e-01 -8.18169534e-01 -4.69520181e-01 -8.54110062e-01
5.73607028e-01 -1.15027949e-01 1.49259353e+00 3.37964982e-01
-3.02057624e-01 1.53212488e-01 -4.62998956e-01 -3.42616141e-01
-1.28546047e+00 3.33007723e-01 -5.18796921e-01 -6.92277551e-01
3.34212452e-01 -5.73677778e-01 -1.70096859e-01 -2.30475515e-01
-1.04903018e+00 2.62362361e-01 7.40203798e-01 6.97898388e-01
1.75452679e-01 -6.45469248e-01 5.78942120e-01 -1.22163773e+00
7.51968086e-01 -8.80462468e-01 -6.47615135e-01 3.24982792e-01
-4.59248513e-01 3.56044710e-01 9.18416739e-01 -1.19063787e-01
-9.99433100e-01 -1.01284608e-01 -4.25439119e-01 6.87872767e-02
1.94090996e-02 1.10100651e+00 2.71417081e-01 2.77483106e-01
8.85271013e-01 5.98251283e-01 1.72023219e-03 -4.97841030e-01
3.83929074e-01 8.91507089e-01 4.19337451e-01 -1.04883933e+00
5.73369384e-01 -1.83430463e-01 -5.21019518e-01 -5.18459916e-01
-2.01520413e-01 -2.07993314e-01 -4.70781177e-01 9.12318975e-02
4.46796268e-01 -7.98423946e-01 3.03368792e-02 1.83170766e-01
-1.58439302e+00 -6.12471998e-01 -3.70692611e-02 1.10346779e-01
-8.12246203e-01 4.99207288e-01 -9.00003850e-01 -4.90700990e-01
-5.07398903e-01 -1.38626432e+00 1.04281580e+00 7.84687605e-03
-5.80473483e-01 -6.56723738e-01 1.64145619e-01 1.40418857e-01
5.97147882e-01 1.25535995e-01 1.56912851e+00 -6.11881852e-01
-8.94041479e-01 4.31959983e-03 -5.82569391e-02 1.79636911e-01
-1.23548269e-01 4.57414687e-01 -3.74780208e-01 -1.27659649e-01
-3.97779584e-01 -3.65918547e-01 6.55630469e-01 1.58827435e-02
1.19293332e+00 -5.74307263e-01 -2.73882210e-01 5.30287266e-01
1.57479453e+00 3.34853142e-01 7.26095319e-01 3.15731198e-01
6.32714570e-01 4.74396646e-01 2.66940504e-01 4.58773941e-01
4.80557978e-01 4.75485802e-01 2.91259080e-01 5.06921530e-01
-1.55356988e-01 -5.34589052e-01 9.61085260e-01 1.39986873e+00
3.31851393e-01 -8.03403705e-02 -1.45296419e+00 8.89437139e-01
-1.75339293e+00 -9.43989694e-01 -4.21132535e-01 1.99010444e+00
1.11167705e+00 2.39950623e-02 7.63146579e-02 -5.26392817e-01
5.30401587e-01 -3.92901003e-02 -5.29231668e-01 -1.05720925e+00
2.39881814e-01 2.62560755e-01 3.75649869e-01 3.05684417e-01
-4.44518894e-01 9.32385802e-01 6.21242380e+00 8.19378138e-01
-9.10046458e-01 1.61701962e-01 2.20859334e-01 -6.66806325e-02
-8.98311853e-01 4.02796447e-01 -7.03659356e-01 7.17368126e-01
1.45672584e+00 -6.16762102e-01 6.89932704e-01 1.05482185e+00
1.38335899e-01 8.83360803e-02 -1.32418013e+00 5.75798631e-01
2.06080765e-01 -1.56693494e+00 1.75056234e-01 -1.53530687e-01
8.66454363e-01 4.42262858e-01 -1.20230056e-01 6.67084694e-01
7.71385729e-01 -8.48369896e-01 1.18586648e+00 3.43030453e-01
1.01221514e+00 -5.19652069e-01 4.09765810e-01 3.69045705e-01
-1.15544820e+00 -3.00968587e-01 -1.90000772e-01 7.24496916e-02
7.54918158e-02 4.32366639e-01 -7.97926605e-01 6.57187223e-01
5.94483793e-01 9.41963315e-01 -7.41508782e-01 9.74239469e-01
-2.45677352e-01 7.80800462e-01 2.12591186e-01 -3.76290798e-01
2.41020113e-01 1.27283875e-02 5.72052479e-01 1.84814286e+00
6.50587499e-01 -1.09311879e-01 9.57145244e-02 1.44061565e+00
-2.60004252e-01 2.49871865e-01 -8.43843758e-01 -4.86390829e-01
4.39456344e-01 7.35644341e-01 -3.42575133e-01 -5.24625719e-01
-7.66907573e-01 8.76819015e-01 3.61608356e-01 2.43241057e-01
-8.88537109e-01 -1.03191733e+00 5.75984001e-01 -1.15224756e-01
3.69386196e-01 -4.68692541e-01 -2.70829320e-01 -1.35685515e+00
3.42850834e-01 -1.37579346e+00 1.32081881e-01 -8.87298226e-01
-8.02809536e-01 7.16080308e-01 2.99515910e-02 -1.32928562e+00
-5.29489279e-01 -3.10084313e-01 -7.59020388e-01 8.43100369e-01
-1.19788730e+00 -8.20597589e-01 1.58791900e-01 -1.47986621e-01
1.20501268e+00 -2.94021845e-01 7.05048800e-01 2.82120109e-01
-3.88598949e-01 6.71962917e-01 3.76005858e-01 3.52179408e-01
4.35414314e-01 -1.26277745e+00 1.25778985e+00 1.17165458e+00
-1.25570387e-01 1.16627467e+00 5.30740798e-01 -8.45971704e-01
-1.83158648e+00 -1.28483343e+00 1.03415930e+00 -7.50772476e-01
8.36002827e-01 -4.60665852e-01 -8.61453533e-01 9.22106206e-01
5.95010221e-01 -5.05678117e-01 7.30633497e-01 -3.65564376e-01
-5.23013711e-01 4.07584369e-01 -6.75899565e-01 8.06738913e-01
1.25439441e+00 -5.89129210e-01 -5.98128915e-01 2.18357712e-01
9.95303750e-01 -6.26491189e-01 -8.72964501e-01 -7.51057193e-02
5.25519431e-01 -8.66656065e-01 4.99587357e-01 -5.06911993e-01
1.46076345e+00 -1.50343791e-01 -2.87273228e-01 -1.34293032e+00
-1.37456164e-01 -9.69616055e-01 2.92368662e-02 1.33906579e+00
9.31141675e-01 -3.60709161e-01 3.84938627e-01 2.89547414e-01
-7.20924616e-01 -6.67936563e-01 -4.72548395e-01 -8.31865191e-01
3.65830839e-01 -4.51099545e-01 7.26960182e-01 8.87885690e-01
5.83693266e-01 2.48569995e-01 -5.89462519e-02 -2.69933999e-01
4.78543997e-01 3.66819948e-01 7.95459569e-01 -6.51135921e-01
-6.50826454e-01 -6.53686762e-01 1.39100015e-01 -1.19219327e+00
4.63881195e-01 -1.62888455e+00 2.75655121e-01 -1.63587129e+00
4.46567088e-01 -2.37924367e-01 3.03662091e-01 6.67080283e-01
-5.74398786e-02 -2.75678635e-01 3.03055227e-01 3.11968923e-01
-3.83790642e-01 2.81567127e-01 6.06040120e-01 -2.96188653e-01
-1.08461417e-01 -2.29242712e-01 -7.44414151e-01 4.31895018e-01
1.07107949e+00 -8.49072576e-01 -1.68233514e-01 -1.09939241e+00
5.58461487e-01 6.27709806e-01 -5.65764159e-02 -7.47653186e-01
5.59647866e-02 -2.57792801e-01 -3.13408941e-01 -2.55421281e-01
-3.23564529e-01 -2.11361617e-01 1.98702678e-01 7.11898148e-01
-8.26431096e-01 7.40511775e-01 4.71316963e-01 1.95420906e-01
-6.58569634e-02 -8.61546755e-01 4.33599293e-01 -7.09862590e-01
-6.85031831e-01 -4.20242921e-02 -8.35999906e-01 2.99416274e-01
5.16820550e-01 -3.15673985e-02 -6.80038333e-01 -7.69441724e-02
1.04532309e-01 2.11357549e-01 7.91091383e-01 7.45368600e-01
6.01930797e-01 -1.05219460e+00 -9.50166166e-01 1.11185342e-01
5.10215640e-01 -2.00359702e-01 -4.73657772e-02 6.97466075e-01
-7.24037349e-01 5.86670518e-01 -2.43563980e-01 -4.66934532e-01
-1.15577054e+00 4.79634255e-01 1.14038952e-01 -7.33023807e-02
-4.19771433e-01 7.14502811e-01 -1.17752232e-01 -7.05604374e-01
-2.98602909e-01 -6.13703609e-01 3.06580156e-01 -6.53411806e-01
5.50842106e-01 2.99015582e-01 2.42030546e-02 -2.11908460e-01
-2.02084854e-01 3.72685254e-01 -2.71136105e-01 -1.02781147e-01
1.36621785e+00 2.02245578e-01 -7.57340133e-01 4.96344090e-01
1.34870052e+00 2.47343510e-01 -8.32577884e-01 -1.11007944e-01
5.75110316e-01 -5.57551026e-01 -7.56592810e-01 -8.82556617e-01
-5.13903022e-01 7.70069182e-01 -2.83373054e-02 3.20392065e-02
8.25088561e-01 8.67736712e-03 8.36303890e-01 3.64599615e-01
6.01755798e-01 -7.40502536e-01 1.87111512e-01 8.46640646e-01
7.98783362e-01 -8.99887085e-01 -4.89313334e-01 -2.39394009e-01
-4.90754575e-01 1.35191441e+00 6.85502648e-01 -1.63130350e-02
1.20096914e-02 6.18568122e-01 -9.95899513e-02 -6.17624596e-02
-1.14144886e+00 2.63110131e-01 -1.54046029e-01 3.17695886e-01
1.13013101e+00 -1.33814290e-01 -5.13017237e-01 4.41902220e-01
-2.67656118e-01 2.45169848e-01 1.22041619e+00 1.48477852e+00
-3.56855899e-01 -1.54370701e+00 -1.41362324e-01 7.60209084e-01
-8.07456374e-01 -7.62756288e-01 -2.75225252e-01 3.01593572e-01
-3.55981976e-01 7.26998627e-01 -2.48245060e-01 -2.49889150e-01
2.87856102e-01 4.79083151e-01 3.36723834e-01 -1.30597591e+00
-9.73100185e-01 -2.25486115e-01 2.32388258e-01 -5.69705904e-01
-2.58705802e-02 -8.38973701e-01 -1.33055878e+00 -4.30760443e-01
6.38410822e-02 2.92229146e-01 7.03321457e-01 3.67498696e-01
7.94195354e-01 6.59759581e-01 4.22026575e-01 -6.15485787e-01
-5.88035643e-01 -8.26540232e-01 5.63401310e-03 8.10868293e-02
3.94328356e-01 1.66930676e-01 -1.28850783e-03 5.94688892e-01] | [7.661342620849609, 7.895580768585205] |
fb0839b4-7c1c-4c0a-8c3c-7a7761c5e7cc | learning-nuclei-representations-with-masked | 2306.17116 | null | https://arxiv.org/abs/2306.17116v1 | https://arxiv.org/pdf/2306.17116v1.pdf | Learning Nuclei Representations with Masked Image Modelling | Masked image modelling (MIM) is a powerful self-supervised representation learning paradigm, whose potential has not been widely demonstrated in medical image analysis. In this work, we show the capacity of MIM to capture rich semantic representations of Haemotoxylin & Eosin (H&E)-stained images at the nuclear level. Inspired by Bidirectional Encoder representation from Image Transformers (BEiT), we split the images into smaller patches and generate corresponding discrete visual tokens. In addition to the regular grid-based patches, typically used in visual Transformers, we introduce patches of individual cell nuclei. We propose positional encoding of the irregular distribution of these structures within an image. We pre-train the model in a self-supervised manner on H&E-stained whole-slide images of diffuse large B-cell lymphoma, where cell nuclei have been segmented. The pre-training objective is to recover the original discrete visual tokens of the masked image on the one hand, and to reconstruct the visual tokens of the masked object instances on the other. Coupling these two pre-training tasks allows us to build powerful, context-aware representations of nuclei. Our model generalizes well and can be fine-tuned on downstream classification tasks, achieving improved cell classification accuracy on PanNuke dataset by more than 5% compared to current instance segmentation methods. | ['Katarzyna Bożek', 'Reinhard Büttner', 'Adrian Simon', 'Hussein Naji', 'Piotr Wójcik'] | 2023-06-29 | null | null | null | null | ['whole-slide-images', 'instance-segmentation'] | ['computer-vision', 'computer-vision'] | [ 7.07026839e-01 6.20002329e-01 -1.87171564e-01 -1.48703635e-01
-1.05018330e+00 -5.19259751e-01 3.99005830e-01 4.66589391e-01
-3.40183258e-01 7.49744296e-01 -4.74007428e-02 -1.74691230e-01
2.08007663e-01 -9.60833311e-01 -7.87342489e-01 -1.14562345e+00
1.87506557e-01 7.70799398e-01 3.59236419e-01 1.04840621e-01
1.68061387e-02 6.22621715e-01 -1.39905858e+00 9.00634050e-01
5.33511877e-01 7.87797868e-01 4.15820986e-01 7.99061716e-01
-1.49775237e-01 1.05949211e+00 -4.25301045e-01 -3.06500476e-02
-2.32466504e-01 -5.77837586e-01 -1.16599548e+00 6.55145943e-01
1.82319984e-01 1.03431186e-02 -1.82193518e-01 9.36624825e-01
2.70893484e-01 -3.28305602e-01 1.02181852e+00 -8.64460945e-01
-6.96894884e-01 4.04378355e-01 -6.97439551e-01 4.07298028e-01
-1.92737162e-01 1.49278775e-01 8.80466938e-01 -8.16852450e-01
1.11400223e+00 8.12057912e-01 6.12172842e-01 6.84618533e-01
-1.67343664e+00 -3.77496332e-01 -2.68021166e-01 1.53232962e-01
-1.48756695e+00 -1.91044971e-01 6.27553940e-01 -6.67399943e-01
9.55230474e-01 2.64058977e-01 7.33463526e-01 6.98637724e-01
2.78553247e-01 9.56200659e-01 1.39258420e+00 -5.71170568e-01
3.13887209e-01 3.52084190e-01 -2.21502292e-03 7.67793953e-01
-5.03042005e-02 -1.93971932e-01 -5.62730171e-02 8.36630538e-03
8.84322643e-01 3.04339111e-01 -2.92082638e-01 -2.93239892e-01
-1.31779051e+00 8.12686443e-01 7.38092840e-01 7.62263298e-01
-1.97739020e-01 1.51119068e-01 2.76710004e-01 -1.62544295e-01
5.95911026e-01 9.83276889e-02 -1.48337632e-01 4.12968129e-01
-1.21908772e+00 -8.56298767e-03 2.16697007e-01 4.84985560e-01
9.81119692e-01 -2.50978887e-01 -4.08342391e-01 8.41284633e-01
1.58671588e-01 2.57177390e-02 7.73962915e-01 -7.06558764e-01
-1.61868691e-01 1.00103903e+00 -2.49992490e-01 -7.06863344e-01
-3.76834065e-01 -5.17746389e-01 -1.13002014e+00 2.60676503e-01
6.05665088e-01 5.11099756e-01 -1.32586527e+00 1.49662423e+00
3.77949238e-01 2.05204606e-01 2.66455654e-02 6.39245212e-01
7.79217064e-01 5.63124597e-01 1.62820324e-01 -2.78478086e-01
1.58279765e+00 -9.53243673e-01 -6.20961487e-01 -1.80372104e-01
9.24285054e-01 -4.48774844e-01 7.41008043e-01 1.31582856e-01
-1.03120291e+00 -4.96709853e-01 -9.47303236e-01 -2.10926250e-01
-6.24445677e-01 2.69283563e-01 5.46858788e-01 2.57278860e-01
-1.32966077e+00 4.67238516e-01 -8.27657282e-01 -2.13574722e-01
1.12686801e+00 3.49224091e-01 -5.63126147e-01 -5.04489802e-02
-6.81813896e-01 5.31356037e-01 3.61632109e-01 -2.63074897e-02
-9.33266401e-01 -7.53560483e-01 -8.35984945e-01 7.28385299e-02
6.38487702e-03 -6.82257116e-01 9.84716117e-01 -1.01229656e+00
-1.06189394e+00 1.76121008e+00 -3.87048453e-01 -3.78426075e-01
2.18084157e-01 8.77784252e-01 -6.02172315e-02 7.17887044e-01
2.47575969e-01 1.06925869e+00 6.97824657e-01 -1.57210159e+00
-5.87888718e-01 -4.96279299e-01 -3.21469694e-01 -4.87770401e-02
3.93568631e-03 -3.71921837e-01 -4.58466589e-01 -8.66390467e-01
6.86155036e-02 -7.07764447e-01 -4.48498517e-01 3.51532735e-02
-5.17639756e-01 -5.85457422e-02 6.05643451e-01 -5.87554336e-01
8.70322585e-01 -2.23630500e+00 1.26000434e-01 2.93533504e-01
4.72671390e-01 1.70909464e-01 -1.37138022e-02 1.81499526e-01
-3.16276342e-01 1.39856696e-01 -6.12417579e-01 -4.12941873e-01
-1.68324724e-01 4.46092606e-01 -1.69990942e-01 5.44806063e-01
3.63190323e-01 1.33643675e+00 -8.24635327e-01 -8.58876050e-01
1.37607813e-01 6.08218789e-01 -5.37334383e-01 2.22639769e-01
-2.87935615e-01 7.03384578e-01 -9.17717814e-02 7.61558533e-01
5.38776457e-01 -8.66980076e-01 3.84279907e-01 -1.91235751e-01
3.11623394e-01 -9.28863809e-02 -6.30120933e-01 1.43146622e+00
-2.58030057e-01 5.64694583e-01 3.00707445e-02 -1.38110662e+00
5.68110108e-01 3.04385036e-01 4.36300933e-01 -7.22671807e-01
5.15743531e-02 1.18399635e-01 -2.87696242e-01 -4.39437747e-01
1.10500708e-01 -7.22170889e-01 1.68952286e-01 5.17471492e-01
2.65181690e-01 -1.75233796e-01 2.78522313e-01 1.71804622e-01
9.87560570e-01 -1.59122854e-01 4.96257365e-01 -1.48681358e-01
5.87797403e-01 1.89598456e-01 3.51593226e-01 3.16544801e-01
1.36709521e-02 1.07298028e+00 7.31618166e-01 -3.74678284e-01
-1.00334299e+00 -9.59292054e-01 -4.19198900e-01 8.70138586e-01
-3.30804251e-02 -3.32908034e-02 -9.19763446e-01 -8.06106031e-01
-8.95966031e-03 2.44614840e-01 -1.20132935e+00 -2.08883118e-02
-4.62918818e-01 -9.77610886e-01 5.41876972e-01 6.37589157e-01
2.38051906e-01 -1.16575563e+00 -4.74297106e-01 2.78870940e-01
-2.07088217e-01 -9.11248565e-01 -1.20190889e-01 6.58621669e-01
-9.22888994e-01 -1.17956495e+00 -9.29503918e-01 -1.30985880e+00
1.32099593e+00 7.11168647e-02 1.11254489e+00 4.56384122e-01
-8.53962123e-01 1.82557449e-01 -1.86972201e-01 3.68339904e-02
-5.09280682e-01 -2.00338915e-01 -5.71598291e-01 2.07283020e-01
1.03456981e-01 -4.99796361e-01 -6.88932359e-01 1.39257312e-01
-1.27346396e+00 2.31676117e-01 6.75191998e-01 1.27144063e+00
1.35902202e+00 7.19138756e-02 4.01480794e-01 -1.44506145e+00
-1.19030505e-01 -5.51681042e-01 -7.28462562e-02 3.08025062e-01
-2.88140595e-01 -1.98768526e-01 4.91815299e-01 -3.55934292e-01
-8.49108458e-01 8.88640434e-02 -2.44435474e-01 -2.95199990e-01
-3.98972452e-01 3.33477437e-01 6.60457388e-02 -4.58824961e-03
5.98896444e-01 5.79503536e-01 1.57294646e-01 -1.69031262e-01
2.50295192e-01 7.43985832e-01 7.72714734e-01 -2.83217251e-01
4.93851751e-01 1.09659469e+00 -6.03988394e-02 -7.55408704e-01
-6.59759998e-01 -6.07874572e-01 -6.49340272e-01 1.21826634e-01
1.08073819e+00 -7.36284971e-01 -3.81277919e-01 3.90538335e-01
-7.67917871e-01 -7.92032897e-01 -6.66968107e-01 -1.84593294e-02
-8.74023616e-01 2.92489171e-01 -8.68219793e-01 -3.56873930e-01
-1.35113999e-01 -1.07077491e+00 1.41783690e+00 1.25631377e-01
-1.58897951e-01 -1.22021389e+00 2.67007500e-01 6.09268248e-01
7.68834427e-02 4.51966554e-01 1.30344856e+00 -6.44416451e-01
-5.30472577e-01 -8.91383886e-02 -3.48251402e-01 2.98191309e-01
2.40650415e-01 -9.57482457e-02 -1.19921970e+00 -3.05672944e-01
-2.42505223e-02 -3.62445682e-01 1.18294203e+00 5.63569367e-01
1.44365621e+00 -3.78971279e-01 -9.13013518e-01 7.82924294e-01
1.51913190e+00 4.61539216e-02 8.70698750e-01 1.73679397e-01
5.40753365e-01 7.38602877e-01 2.26388931e-01 5.04980497e-02
2.88362384e-01 4.40053284e-01 2.67811358e-01 -6.38672411e-01
-5.00824690e-01 -1.77877650e-01 -2.55438000e-01 3.09117466e-01
5.95804751e-02 -1.68588728e-01 -8.63723695e-01 1.00523043e+00
-1.49836230e+00 -8.25671136e-01 9.33517888e-02 1.53800893e+00
1.09649467e+00 -6.33333996e-02 -4.39133905e-02 4.31367576e-01
7.04778373e-01 -7.35652214e-03 -3.92836899e-01 -1.22852385e-01
-1.99160755e-01 5.81745684e-01 1.37269706e-01 4.18325543e-01
-9.54188347e-01 7.89281011e-01 5.86371517e+00 1.08093226e+00
-1.19496870e+00 2.74041712e-01 1.41535306e+00 9.29321349e-02
-3.96518677e-01 -1.52944818e-01 -5.65367877e-01 4.85256404e-01
6.54527664e-01 8.09329078e-02 1.26689389e-01 4.80495661e-01
-2.01724589e-01 -1.30506977e-01 -1.19156516e+00 8.38359714e-01
9.47803855e-02 -1.88073659e+00 1.44708872e-01 3.39990944e-01
7.97861755e-01 -2.04877064e-01 1.51458189e-01 1.77534029e-01
1.52541012e-01 -1.51660645e+00 4.11901951e-01 4.23891425e-01
1.17276502e+00 -6.17187679e-01 9.62503314e-01 3.40252072e-01
-9.76563096e-01 9.56254676e-02 -3.92821759e-01 3.38190883e-01
-1.68822095e-01 6.06814146e-01 -1.24058902e+00 2.86150098e-01
5.81517100e-01 7.03903139e-01 -6.57244205e-01 7.31008232e-01
-3.56114917e-02 5.72028279e-01 -1.18036136e-01 3.68089795e-01
3.00033629e-01 4.75331470e-02 -2.06241943e-02 1.48548889e+00
5.35686351e-02 1.75300911e-01 -1.36528686e-01 1.04190373e+00
-1.50250405e-01 -2.09377967e-02 -4.73105609e-01 -1.22775741e-01
-4.00451571e-03 1.53890681e+00 -1.38102806e+00 -5.63728809e-01
-1.89539567e-01 1.04733133e+00 6.24106228e-01 4.71435934e-01
-5.72407126e-01 -1.23147681e-01 1.07942574e-01 5.39947867e-01
7.41251349e-01 4.90353554e-01 -2.99419045e-01 -7.99300849e-01
-2.83149362e-01 -6.78862274e-01 4.85613704e-01 -6.66023493e-01
-1.24143529e+00 6.11339033e-01 -3.30125332e-01 -1.16795647e+00
-2.02909514e-01 -6.18176043e-01 -5.12677073e-01 6.84613824e-01
-1.59220743e+00 -1.64786422e+00 -1.52595446e-01 3.93348962e-01
3.85483772e-01 1.29192278e-01 1.03186870e+00 -2.53745113e-02
-4.18450743e-01 4.68319148e-01 1.69385403e-01 2.07056850e-01
2.49078900e-01 -1.53649771e+00 -1.60021912e-02 4.52978849e-01
1.72953680e-01 4.59828764e-01 3.82410854e-01 -4.68539178e-01
-6.18283451e-01 -1.35288846e+00 9.90162313e-01 -3.77897650e-01
4.69676971e-01 -2.39576310e-01 -1.12089276e+00 7.81750202e-01
1.95355579e-01 6.82229102e-01 1.05927050e+00 -6.46587491e-01
-2.72170067e-01 1.66416243e-01 -1.45145619e+00 3.83253723e-01
5.30722678e-01 -6.66783810e-01 -5.77731609e-01 5.17817974e-01
2.59229213e-01 -3.95386845e-01 -9.18802559e-01 2.40016460e-01
1.84389204e-01 -1.01043570e+00 1.05712795e+00 -4.31524456e-01
5.76216102e-01 -3.77076119e-01 -8.85783881e-02 -1.12355769e+00
-4.18461144e-01 -1.33450955e-01 1.19360894e-01 1.15522444e+00
1.90118968e-01 -5.18288672e-01 1.05820608e+00 2.13244688e-02
-1.74111873e-01 -1.25259066e+00 -1.01811075e+00 -2.31384516e-01
2.91727871e-01 2.25117579e-02 3.34307522e-01 8.20871353e-01
2.46043742e-01 6.11743852e-02 2.32631370e-01 -3.47330123e-02
6.37887061e-01 2.68360108e-01 2.88545728e-01 -9.38249588e-01
-5.20177722e-01 -5.36072254e-01 -5.74418068e-01 -7.85499990e-01
2.09946409e-01 -1.40441990e+00 -1.59815237e-01 -1.67247105e+00
6.90492094e-01 -4.43843216e-01 -4.18192357e-01 7.04219460e-01
-2.09295869e-01 1.08822203e+00 -2.70793945e-01 3.73669654e-01
-5.81729770e-01 2.29589045e-01 1.39326370e+00 -5.64520597e-01
1.94219172e-01 -2.41368368e-01 -8.01717937e-01 6.60157740e-01
5.69717407e-01 -5.04793108e-01 -1.93745226e-01 3.05957114e-03
-2.25485768e-03 9.16024968e-02 6.29084408e-01 -7.91133285e-01
6.63745552e-02 1.60736963e-01 7.94005871e-01 -6.06551111e-01
2.18499467e-01 -7.10428476e-01 1.68729141e-01 4.43825543e-01
-4.56243604e-01 -4.52835321e-01 1.22898728e-01 7.07247019e-01
-4.73027349e-01 -2.05875576e-01 1.06429350e+00 -5.21744967e-01
-2.54969239e-01 1.40486091e-01 -6.17353439e-01 -2.10910589e-01
1.23093367e+00 -6.29449129e-01 -3.89654815e-01 8.96046981e-02
-9.49312866e-01 -2.96744965e-02 7.03901172e-01 -3.08968902e-01
7.42916107e-01 -1.19569790e+00 -4.53396291e-01 3.74914765e-01
2.90616274e-01 4.13955748e-01 6.39875889e-01 9.95222986e-01
-7.92407155e-01 4.37041551e-01 -2.29843035e-01 -8.92051160e-01
-1.19292235e+00 6.99122190e-01 4.60315973e-01 -8.22706878e-01
-6.95051789e-01 1.03524470e+00 8.59981418e-01 -2.74666369e-01
6.38619123e-04 -4.10320640e-01 -4.67752784e-01 4.77598123e-02
6.31950140e-01 6.29689079e-03 1.94115430e-01 -8.03744256e-01
-3.09421688e-01 5.07883191e-01 -3.32075000e-01 1.70429856e-01
1.38441467e+00 -6.60976321e-02 -4.37876582e-01 4.73239094e-01
1.39525712e+00 -1.68263242e-01 -1.25731862e+00 -2.35451564e-01
-8.88876319e-02 -1.35141015e-01 9.13641527e-02 -6.79313898e-01
-1.28472424e+00 9.24656153e-01 5.72277963e-01 2.18839407e-01
1.06761825e+00 4.72040057e-01 7.31483161e-01 -1.93651304e-01
2.32266113e-01 -7.07147539e-01 2.48698473e-01 -1.17827527e-01
5.13632059e-01 -1.08819723e+00 -1.93643257e-01 -5.56396604e-01
-5.74974477e-01 9.96230900e-01 2.30660379e-01 -2.11455256e-01
3.42846602e-01 4.41034466e-01 1.11236066e-01 -5.04944324e-01
-8.02712440e-01 -3.31376612e-01 2.19277844e-01 9.57536280e-01
4.36895132e-01 8.18409249e-02 1.80097818e-01 6.80125177e-01
1.52667642e-01 7.02267960e-02 3.73209387e-01 7.69191146e-01
-3.37208092e-01 -8.81968319e-01 -3.17796946e-01 5.46721160e-01
-6.39447451e-01 3.41672152e-02 -3.00278574e-01 7.34485090e-01
3.92125696e-01 5.24976492e-01 5.13181627e-01 -7.88887963e-02
-7.88982809e-02 3.20456699e-02 6.41806960e-01 -8.91891062e-01
-6.03005648e-01 3.03699166e-01 -2.86040902e-01 -2.16644317e-01
-5.68964303e-01 -4.07536447e-01 -1.60368669e+00 1.33545130e-01
7.58408904e-02 8.62568989e-02 1.06363244e-01 9.34001327e-01
1.20624110e-01 7.23472834e-01 5.26315033e-01 -9.57533121e-01
4.79047000e-02 -8.44645321e-01 -8.97400796e-01 5.64447224e-01
5.32003403e-01 -5.73581815e-01 -3.88998121e-01 5.79686403e-01] | [14.96348762512207, -2.8942995071411133] |
d559ed2a-2832-4849-b2fe-9dfaa67e1dac | neural-fourier-filter-bank | 2212.01735 | null | https://arxiv.org/abs/2212.01735v3 | https://arxiv.org/pdf/2212.01735v3.pdf | Neural Fourier Filter Bank | We present a novel method to provide efficient and highly detailed reconstructions. Inspired by wavelets, we learn a neural field that decompose the signal both spatially and frequency-wise. We follow the recent grid-based paradigm for spatial decomposition, but unlike existing work, encourage specific frequencies to be stored in each grid via Fourier features encodings. We then apply a multi-layer perceptron with sine activations, taking these Fourier encoded features in at appropriate layers so that higher-frequency components are accumulated on top of lower-frequency components sequentially, which we sum up to form the final output. We demonstrate that our method outperforms the state of the art regarding model compactness and convergence speed on multiple tasks: 2D image fitting, 3D shape reconstruction, and neural radiance fields. Our code is available at https://github.com/ubc-vision/NFFB. | ['Kwang Moo Yi', 'Yuhe Jin', 'Zhijie Wu'] | 2022-12-04 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wu_Neural_Fourier_Filter_Bank_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wu_Neural_Fourier_Filter_Bank_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-shape-reconstruction'] | ['computer-vision'] | [ 1.72561482e-01 -4.48597223e-01 1.53031409e-01 -2.57606685e-01
-8.54104459e-01 -3.75062436e-01 5.04110038e-01 6.52370751e-02
-1.35281801e-01 5.43180406e-01 4.45309341e-01 -1.65951729e-01
-2.75169283e-01 -9.78195131e-01 -9.84273553e-01 -8.86625290e-01
-4.23330218e-01 -6.71590790e-02 -1.22784898e-01 2.83627734e-02
2.87908047e-01 7.14282393e-01 -1.61622417e+00 4.89086241e-01
6.63556993e-01 1.28191960e+00 3.49089921e-01 7.32616782e-01
1.07113965e-01 8.33160400e-01 -1.65837511e-01 2.43951738e-01
2.42662758e-01 -1.21721268e-01 -5.54427207e-01 -8.22734926e-03
3.09020966e-01 -3.16616535e-01 -4.37964171e-01 8.41434538e-01
5.19131005e-01 3.93806815e-01 5.81332743e-01 -5.86090624e-01
-9.59795713e-01 2.25048169e-01 -6.13722742e-01 1.34700730e-01
3.56068254e-01 6.58110082e-02 7.94915795e-01 -1.25812674e+00
2.57883012e-01 1.03591263e+00 1.05948734e+00 1.42550126e-01
-1.53775716e+00 -3.37413490e-01 9.99653786e-02 6.27425164e-02
-1.32075381e+00 -6.33403361e-01 1.06991053e+00 -4.28394228e-01
1.43352056e+00 3.70818228e-01 7.76005507e-01 8.12563121e-01
1.91529661e-01 4.99260992e-01 1.08198702e+00 -5.39098024e-01
1.72602400e-01 -3.87518585e-01 4.08996604e-02 7.43450403e-01
-1.90029532e-01 3.42681617e-01 -6.60428047e-01 -2.84489751e-01
1.07111669e+00 -1.50169954e-01 -5.13460279e-01 -5.19798659e-02
-1.11295009e+00 8.66118670e-01 8.60019863e-01 2.79702276e-01
-9.59851325e-01 4.28941548e-01 -2.16509420e-02 7.74989426e-02
8.14826012e-01 4.10216838e-01 -5.13871968e-01 3.13727289e-01
-1.03216624e+00 3.23052734e-01 4.22642618e-01 2.83496588e-01
1.00807011e+00 1.00610830e-01 2.58369800e-02 1.04054487e+00
3.19888681e-01 3.36153626e-01 3.90821427e-01 -1.50030553e+00
-1.33243963e-01 1.74266919e-01 1.17300160e-01 -1.17911947e+00
-5.14303982e-01 -6.30590260e-01 -1.14615560e+00 4.48908448e-01
4.99971174e-02 3.85420173e-02 -8.04476678e-01 1.54453778e+00
2.63770938e-01 6.09304965e-01 -2.17584282e-01 9.92342174e-01
7.55304873e-01 1.00526094e+00 -2.23493576e-01 -1.61574647e-01
1.22476447e+00 -9.55911994e-01 -4.50054526e-01 -1.28547415e-01
6.47072271e-02 -7.80542612e-01 7.65894890e-01 4.37539101e-01
-1.40989614e+00 -7.72411585e-01 -8.78003538e-01 -2.66085446e-01
-1.98369935e-01 1.86676115e-01 8.83853316e-01 1.17179155e-01
-1.46862721e+00 9.96285081e-01 -9.88349557e-01 1.22452050e-01
5.54625690e-01 1.97729934e-02 -1.60135522e-01 1.56824261e-01
-8.89747262e-01 7.97174633e-01 -3.64354290e-02 2.55819540e-02
-6.72334135e-01 -1.07826042e+00 -1.01515627e+00 1.19051605e-01
-2.83132732e-01 -9.71117079e-01 1.28266609e+00 -8.31431150e-01
-1.43271530e+00 7.31220484e-01 -5.59741557e-01 -5.33097088e-01
8.16665683e-03 -2.98971161e-02 -1.29166231e-01 3.13994616e-01
-8.69475678e-03 9.00506616e-01 8.80760789e-01 -1.32317114e+00
-4.36421484e-01 -8.33934322e-02 -9.18271467e-02 1.37862429e-01
-8.61557201e-02 -1.99422583e-01 -2.41064444e-01 -8.15227568e-01
3.22367013e-01 -4.29832458e-01 -1.96080819e-01 3.52143347e-01
-5.09163029e-02 -2.27186933e-01 3.07485878e-01 -8.96090567e-01
1.11815310e+00 -2.14761496e+00 2.55332440e-01 1.06997259e-01
2.04563141e-01 -2.58820355e-01 -2.08836600e-01 3.45025569e-01
-3.21539581e-01 -1.25621796e-01 -5.44566691e-01 -6.68257296e-01
-8.13798159e-02 -4.50521819e-02 -6.18413925e-01 4.99827355e-01
4.63472396e-01 6.75502062e-01 -6.62858427e-01 4.66037765e-02
4.74957943e-01 1.08154845e+00 -6.79976642e-01 -6.62158504e-02
-3.08079809e-01 5.80924571e-01 -1.91517696e-01 6.10784173e-01
8.97763848e-01 -6.06041908e-01 -2.43588924e-01 -4.65950191e-01
-3.84730339e-01 6.09083712e-01 -1.02343059e+00 1.86450958e+00
-7.95037687e-01 6.21425390e-01 2.96377152e-01 -1.00937951e+00
8.09005380e-01 2.59744823e-01 7.75168419e-01 -8.14479411e-01
-2.42089838e-01 2.39006460e-01 -5.22923052e-01 -3.36802930e-01
3.30835432e-01 -6.68089315e-02 3.58903050e-01 3.30789536e-01
1.24009736e-01 -9.65302438e-02 -8.80775005e-02 -3.41629893e-01
9.08236265e-01 5.28538704e-01 1.97485313e-01 -2.13910356e-01
3.04551572e-01 6.45257309e-02 2.57712930e-01 5.42960227e-01
2.04579100e-01 8.64379466e-01 1.25355482e-01 -6.63041234e-01
-1.11031771e+00 -1.11089611e+00 -5.48514664e-01 1.04994857e+00
-2.14909554e-01 -2.87741452e-01 -5.03787696e-01 4.36229073e-02
2.17197046e-01 6.69982374e-01 -5.48169196e-01 9.09271985e-02
-6.20894492e-01 -6.49488866e-01 2.56377578e-01 4.74278808e-01
3.37100387e-01 -1.16813076e+00 -9.58772838e-01 2.95560807e-01
-2.34752476e-01 -6.87053382e-01 -2.75443137e-01 4.91370797e-01
-8.52509737e-01 -6.64250612e-01 -7.38858163e-01 -8.42255533e-01
4.91281152e-01 3.92210364e-01 1.16983604e+00 1.18298121e-01
-3.57706815e-01 1.73752919e-01 -3.03536765e-02 -9.67273489e-02
-7.13354275e-02 -2.67052740e-01 -1.38175577e-01 -8.46350864e-02
8.19474161e-02 -1.13874829e+00 -7.37505555e-01 -2.11514562e-01
-7.44698763e-01 4.19666111e-01 3.62798005e-01 7.29105592e-01
1.04723501e+00 1.45230472e-01 2.92253256e-01 -2.68319815e-01
5.30132651e-01 -5.10603487e-01 -7.52983987e-01 -1.45575047e-01
-8.41612145e-02 7.94775188e-02 7.57801116e-01 -2.26854891e-01
-6.23470128e-01 1.83796927e-01 -3.95126849e-01 -6.31283224e-01
-3.53110909e-01 6.30888462e-01 4.84150380e-01 -3.46891582e-01
8.19916189e-01 4.37496841e-01 -2.16295332e-01 -7.55374551e-01
1.84465885e-01 3.58276635e-01 7.43711114e-01 -4.11811650e-01
4.90746796e-01 5.02655089e-01 6.32206872e-02 -8.80021095e-01
-8.12266290e-01 -2.78642774e-01 -3.93713206e-01 -1.45185843e-01
7.08407164e-01 -9.93517935e-01 -7.07945943e-01 5.14086187e-01
-1.35936975e+00 -5.63025236e-01 -3.01368803e-01 4.85241771e-01
-7.41938174e-01 5.51658534e-02 -6.59465551e-01 -9.04496431e-01
-3.43940765e-01 -8.67468297e-01 1.33059347e+00 1.31225675e-01
-1.35769084e-01 -1.01416171e+00 3.22030991e-01 -4.85492088e-02
5.71231842e-01 3.63960624e-01 8.44545186e-01 2.69735217e-01
-5.18090427e-01 1.88651264e-01 -1.43422008e-01 2.71254420e-01
4.66574468e-02 -2.80541424e-02 -1.29169381e+00 -2.55298346e-01
2.87210703e-01 -3.40942740e-01 1.39952803e+00 1.12794828e+00
1.49526775e+00 -4.02433753e-01 -1.78235725e-01 1.38316584e+00
1.67018425e+00 -2.22810879e-01 4.48926151e-01 2.81201631e-01
3.58356506e-01 4.29681331e-01 -1.39387324e-01 6.25939131e-01
4.42207247e-01 4.16099638e-01 4.68166530e-01 -2.57685184e-01
-2.83328474e-01 -8.94347131e-02 1.64896607e-01 7.40505219e-01
-2.25129545e-01 -1.23554952e-02 -7.75439501e-01 5.53678334e-01
-1.67362392e+00 -1.03678501e+00 8.23019221e-02 1.91846883e+00
8.39984894e-01 -7.03506023e-02 -5.50172552e-02 2.24252746e-01
2.29743734e-01 4.19497579e-01 -4.97508556e-01 -1.45915598e-01
-2.75120497e-01 4.64165807e-01 4.22155499e-01 8.95791471e-01
-1.31081223e+00 5.12642801e-01 6.53113222e+00 6.27054095e-01
-1.32272184e+00 -3.40376012e-02 7.39919484e-01 -2.60883898e-01
-6.51578307e-01 -3.15335989e-01 -3.17103148e-01 3.39081138e-01
8.30923259e-01 1.27653733e-01 1.08490169e+00 4.06741589e-01
2.82625079e-01 -1.13285333e-01 -7.37504005e-01 9.58059311e-01
-2.04551220e-01 -1.82224345e+00 -1.75797045e-01 -1.47293478e-01
6.40435755e-01 4.11744565e-01 1.50297314e-01 -1.15592696e-01
1.46697477e-01 -1.27467334e+00 9.11151886e-01 9.11672771e-01
7.54917145e-01 -5.97353339e-01 1.59394845e-01 4.35450524e-01
-1.26871264e+00 6.41368851e-02 -6.23644948e-01 -2.14789480e-01
2.08051428e-01 8.14405620e-01 -3.24100792e-01 5.13621688e-01
1.06342697e+00 9.19405997e-01 -1.48548275e-01 1.12799907e+00
-5.57384603e-02 5.05149841e-01 -6.13112867e-01 2.26789251e-01
1.65683493e-01 -3.07756573e-01 3.75194848e-01 1.05065715e+00
6.46548867e-01 2.60603726e-01 1.51145920e-01 9.92465615e-01
9.55273658e-02 -3.45825940e-01 -5.14416277e-01 4.28756118e-01
5.19556105e-01 1.26909471e+00 -4.96758252e-01 -1.02280706e-01
-4.73313838e-01 8.76410544e-01 3.71395469e-01 6.44327104e-01
-5.68934381e-01 -1.53409868e-01 7.63963521e-01 2.44836569e-01
6.93906069e-01 -5.02281904e-01 -5.76778829e-01 -9.23876166e-01
8.44637826e-02 -5.37881851e-01 2.18327746e-01 -1.25203180e+00
-1.37699592e+00 7.43337214e-01 -2.03862861e-01 -1.04899192e+00
-1.68782339e-01 -7.60132730e-01 -5.43094337e-01 1.20923662e+00
-1.97069359e+00 -1.09362221e+00 -3.26316029e-01 5.40256083e-01
3.51347268e-01 1.86836779e-01 1.25651538e+00 2.05230236e-01
-1.82165861e-01 7.71626979e-02 1.99767977e-01 -1.60145655e-01
2.67866015e-01 -1.08064008e+00 5.66632807e-01 6.45044148e-01
1.99471250e-01 4.54050928e-01 5.49012601e-01 -4.49233234e-01
-1.37510312e+00 -9.49340224e-01 9.00111973e-01 -5.38753271e-02
5.71116865e-01 -2.90367603e-01 -1.12015998e+00 5.06129682e-01
2.94635922e-01 4.00678158e-01 5.54556429e-01 -1.34766221e-01
-4.17841822e-01 -9.68465880e-02 -1.06828690e+00 3.42648476e-01
9.07970846e-01 -5.42033792e-01 -3.34664226e-01 3.44514430e-01
6.66737497e-01 -4.80950266e-01 -8.51348698e-01 3.68733227e-01
4.78456676e-01 -1.17795217e+00 1.40420580e+00 -1.95010528e-01
5.62693655e-01 -5.19625425e-01 -4.72568601e-01 -1.47246349e+00
-9.09957528e-01 -6.70050442e-01 -3.51109236e-01 5.24257362e-01
3.31013709e-01 -6.33208871e-01 4.75470543e-01 -2.33029798e-01
-3.97157758e-01 -9.46517587e-01 -1.13429642e+00 -2.94897556e-01
8.53172839e-02 -5.65015197e-01 5.93068302e-01 8.58453929e-01
-3.41450632e-01 7.17806742e-02 -2.77714223e-01 5.53488374e-01
8.55274022e-01 4.98033524e-01 2.82917488e-02 -1.02993166e+00
-5.41589320e-01 -7.22005486e-01 1.09805256e-01 -1.13557506e+00
1.07931644e-02 -1.04844511e+00 4.31334004e-02 -1.65735781e+00
-1.34927735e-01 -3.40278357e-01 -3.47239107e-01 6.89088106e-01
5.65243587e-02 6.30849183e-01 1.52371144e-02 2.88246661e-01
-1.45691201e-01 6.11584365e-01 1.15787828e+00 3.32267731e-02
-1.79501504e-01 -1.24086738e-01 -6.38072729e-01 7.41718709e-01
9.00875092e-01 -3.03126067e-01 -7.70130455e-02 -9.20675874e-01
1.33740798e-01 2.53343105e-01 8.54860067e-01 -1.18178141e+00
3.28653455e-01 -1.15727991e-01 1.02275348e+00 -6.22755051e-01
6.22944891e-01 -6.98857784e-01 3.30109864e-01 3.40773970e-01
-2.37412587e-01 -3.08311116e-02 4.73373681e-01 6.84226826e-02
-1.16926976e-01 -3.33816782e-02 8.30312908e-01 -1.47152364e-01
-4.57186520e-01 3.61381412e-01 -3.42983365e-01 -5.26357234e-01
4.63631421e-01 -9.13106203e-02 -2.30714947e-01 -3.98039371e-01
-6.18514299e-01 4.49312292e-02 3.98562789e-01 -1.72699112e-02
7.59988844e-01 -1.57155848e+00 -9.53332186e-01 6.74728811e-01
-3.47026020e-01 1.69975698e-01 3.87321591e-01 8.42673063e-01
-6.27232730e-01 3.00887972e-01 -2.47565791e-01 -6.67541206e-01
-6.97446406e-01 4.75530714e-01 5.89617133e-01 -5.95593899e-02
-8.23061705e-01 9.70915735e-01 1.98471010e-01 -4.07876700e-01
1.97619259e-01 -4.53239143e-01 -8.49137455e-02 -1.06803805e-01
6.05196714e-01 1.86385766e-01 6.20826147e-02 -4.32365596e-01
-4.57051873e-01 1.03985023e+00 5.78859329e-01 -1.61223844e-01
1.93628168e+00 3.98704410e-02 -2.99801618e-01 3.92367363e-01
1.26799154e+00 5.74696437e-02 -1.80124474e+00 -3.35987180e-01
-3.26452434e-01 -4.13359851e-01 4.81059819e-01 -8.72933745e-01
-9.57638144e-01 8.55139554e-01 4.05809760e-01 5.71415782e-01
1.76387179e+00 4.83844317e-02 6.92937911e-01 1.31189495e-01
9.45042819e-02 -5.69036543e-01 -1.60109282e-01 6.33781850e-01
1.24733078e+00 -8.82024229e-01 8.68399069e-02 -1.24550648e-01
-2.16993183e-01 1.05987835e+00 2.62315925e-02 -6.44924939e-01
8.85318756e-01 6.47135377e-01 -4.30518808e-03 -6.53888658e-02
-9.59347367e-01 -3.09493467e-02 5.45688689e-01 6.69175148e-01
5.79872310e-01 -3.88977341e-02 2.85741925e-01 4.64744300e-01
-3.69398803e-01 1.06452098e-02 1.63342953e-02 6.36003077e-01
-6.54847920e-01 -7.31526971e-01 -5.62991619e-01 3.51210952e-01
-2.51246065e-01 -2.61745453e-01 1.91748723e-01 1.57666504e-01
1.20705038e-01 5.81839144e-01 4.83800799e-01 -2.00208232e-01
1.77591980e-01 6.04526550e-02 6.70384288e-01 -3.73613358e-01
-4.99801904e-01 4.52852726e-01 -6.36425540e-02 -6.66806936e-01
-5.41189432e-01 -6.31022871e-01 -1.23895514e+00 -2.96540141e-01
2.06140533e-01 -2.50551969e-01 6.51396930e-01 4.27711040e-01
4.89694417e-01 7.99910128e-01 8.03158402e-01 -1.54662621e+00
-4.17219281e-01 -9.58378434e-01 -2.31173217e-01 4.42477986e-02
1.04603815e+00 -4.18481231e-01 -4.46589202e-01 2.21211255e-01] | [11.015636444091797, -1.9761295318603516] |
a181f5bf-8c8e-4b9e-9abe-474f3ab1fb42 | meta-learning-by-hallucinating-useful | null | null | https://openreview.net/forum?id=rJx8I1rFwr | https://openreview.net/pdf?id=rJx8I1rFwr | Meta-Learning by Hallucinating Useful Examples | Learning to hallucinate additional examples has recently been shown as a promising direction to address few-shot learning tasks, which aim to learn novel concepts from very few examples. The hallucination process, however, is still far from generating effective samples for learning. In this work, we investigate two important requirements for the hallucinator --- (i) precision: the generated examples should lead to good classifier performance, and (ii) collaboration: both the hallucinator and the classification component need to be trained jointly. By integrating these requirements as novel loss functions into a general meta-learning with hallucination framework, our model-agnostic PrecisE Collaborative hAlluciNator (PECAN) facilitates data hallucination to improve the performance of new classification tasks. Extensive experiments demonstrate state-of-the-art performance on competitive miniImageNet and ImageNet based few-shot benchmarks in various scenarios. | ['Karteek Alahari', 'Martial Hebert', 'Yuki Uchiyama', 'Yu-Xiong Wang'] | 2019-09-25 | null | null | null | null | ['novel-concepts'] | ['reasoning'] | [ 2.2302118e-01 1.8980050e-01 -8.9161754e-02 -2.5295949e-01
-8.9963931e-01 2.9025576e-01 8.8495344e-01 3.8060263e-02
-3.0911702e-01 8.6158043e-01 2.2402653e-01 3.6697146e-01
6.2920004e-02 -5.9703583e-01 -6.6177630e-01 -7.0655328e-01
-1.1834326e-01 4.4700670e-01 3.1614760e-01 -3.3703613e-01
1.8137611e-01 1.7567304e-01 -1.8452488e+00 5.3983027e-01
8.5840219e-01 8.8209879e-01 3.2351083e-01 4.7054741e-01
-2.1139977e-02 1.0948628e+00 -6.5853053e-01 -2.9161307e-01
1.7011620e-01 -6.2335116e-01 -5.5988854e-01 4.1361296e-01
3.3850116e-01 -2.6389882e-01 -2.4365954e-01 1.0818626e+00
7.7522367e-01 4.9559852e-01 7.2479743e-01 -1.3843048e+00
-7.1594864e-01 6.3444310e-01 -5.2623516e-01 2.0326497e-01
-8.8064127e-02 6.5858543e-01 7.9379749e-01 -1.3402851e+00
7.0430517e-01 1.0606260e+00 6.0543591e-01 8.6890674e-01
-9.8898566e-01 -6.5453297e-01 -1.0332076e-01 3.7602881e-01
-1.2311326e+00 -5.5955487e-01 6.8385977e-01 -2.4941130e-01
1.0772270e+00 -6.6481881e-02 6.5989375e-01 1.3384601e+00
1.7008795e-01 1.0652350e+00 9.8732734e-01 -2.1228045e-01
7.4313474e-01 5.5935043e-01 1.1944595e-01 5.0199866e-01
1.3654366e-01 6.9615468e-02 -8.0497164e-01 -9.5984936e-02
3.1073865e-01 3.2046872e-01 -2.4066995e-01 -5.5018348e-01
-8.4856051e-01 1.1438702e+00 5.2488512e-01 1.8794502e-01
-4.7054335e-01 2.1534358e-01 6.2963337e-01 3.9721298e-01
5.0400853e-01 9.7405225e-01 -2.8746622e-02 -1.8510722e-02
-8.9417118e-01 2.7077082e-01 6.4580315e-01 1.0205582e+00
7.9556584e-01 3.9380100e-01 -4.6259031e-01 1.1864985e+00
-1.4829502e-01 1.4730941e-01 8.3822554e-01 -8.3990896e-01
5.7261527e-02 4.0775213e-01 -1.5840515e-01 -2.9038200e-01
-3.6473129e-02 -8.7061411e-01 -7.7327079e-01 2.9515991e-01
-1.2190589e-01 -1.0341592e-01 -1.0236789e+00 1.6208798e+00
1.1481733e-02 6.1926800e-01 2.3816080e-01 7.6698554e-01
8.6617178e-01 7.2177297e-01 2.1125259e-01 -4.8885372e-01
1.0191907e+00 -1.3330348e+00 -5.9009242e-01 -4.7981614e-01
4.0909681e-01 -3.7019670e-01 1.2773548e+00 4.2587039e-01
-1.2775108e+00 -6.9318312e-01 -1.2597988e+00 3.0316394e-02
-3.1824026e-01 -4.4250107e-01 6.4211988e-01 5.0383472e-01
-5.4082376e-01 7.6205629e-01 -4.9399644e-01 -3.4933853e-01
9.1909271e-01 3.0693257e-02 -1.4888802e-01 -3.0014762e-01
-1.1080428e+00 1.0090470e+00 5.4244679e-01 -5.5800450e-01
-1.5220112e+00 -1.0102093e+00 -7.5437808e-01 3.4655187e-01
3.8126346e-01 -8.2401866e-01 1.3960886e+00 -7.5195163e-01
-1.1789365e+00 5.7882708e-01 1.5616356e-01 -8.1406683e-01
6.1689645e-01 -2.2378625e-01 -3.8939533e-01 2.1659504e-01
1.4317010e-01 9.5331502e-01 1.0730296e+00 -1.5305455e+00
-4.9336603e-01 -3.6534455e-01 -4.0765113e-01 3.7415555e-01
-6.2127793e-01 -3.5254583e-01 -2.2934228e-01 -5.1317173e-01
-3.7368348e-01 -5.8358729e-01 -3.2865986e-01 1.2796789e-01
-1.5230337e-01 -2.5671592e-01 1.0738451e+00 -1.5639466e-01
8.1014693e-01 -2.0986154e+00 -8.1277311e-02 -1.9852357e-01
4.3263242e-01 7.1950388e-01 -3.2791224e-01 3.4033677e-01
-1.1400971e-01 -2.6272389e-01 -1.4790292e-01 -6.6583484e-01
-1.5653731e-01 1.3560328e-01 -6.4645869e-01 2.7403209e-01
3.5719064e-01 1.1985933e+00 -1.3030246e+00 -1.7550354e-01
3.2690045e-01 4.3659887e-01 -3.5206172e-01 2.4882823e-01
-3.9557305e-01 -9.4866659e-03 -3.1175392e-02 5.5243206e-01
4.6516314e-01 -1.9802856e-01 -2.3270379e-01 1.4147355e-01
3.0576238e-01 -2.0935240e-01 -7.1687901e-01 1.6581594e+00
-5.4281336e-01 5.2430046e-01 -5.1785636e-01 -8.2886159e-01
1.0509881e+00 3.2604656e-01 1.5239277e-01 -5.6844622e-01
-1.7630043e-02 2.1663146e-01 -5.9874836e-02 -5.6672454e-01
5.1864374e-01 -7.8051168e-01 1.3206047e-01 3.5447884e-01
6.5138394e-01 -2.3560043e-01 1.7445406e-01 2.4275255e-01
1.0305880e+00 -2.7997148e-01 6.3568556e-01 -4.1290119e-02
-2.7511453e-02 1.4734100e-01 4.6598130e-01 1.0220923e+00
-2.4576898e-01 7.7518028e-01 3.5381022e-01 -3.1550992e-01
-1.1995541e+00 -1.2261459e+00 8.5875802e-02 9.4930148e-01
1.3450831e-02 -2.4168877e-01 -3.9147338e-01 -5.9719330e-01
-5.3815473e-02 1.3041188e+00 -7.3545820e-01 -7.5277829e-01
3.0622534e-02 -6.8245280e-01 5.0943321e-01 5.9868950e-01
4.6394083e-01 -1.2313272e+00 -8.2499331e-01 3.9375982e-01
1.7548500e-01 -8.6959898e-01 -4.1380587e-01 4.4122514e-01
-8.4259534e-01 -8.6279738e-01 -1.0262042e+00 -7.3303354e-01
4.6966878e-01 6.3319874e-01 8.4893799e-01 -8.4503658e-02
-5.4188281e-01 2.3411790e-01 -4.5995799e-01 -7.0321244e-01
-3.9736596e-01 -2.3739503e-01 -2.1707062e-02 -9.0889670e-03
4.4534734e-01 -7.6618123e-01 -3.5897580e-01 8.0161281e-02
-9.9180323e-01 5.0020352e-02 7.8719485e-01 1.2976186e+00
3.5574353e-01 -4.8043944e-02 9.6952820e-01 -9.7220874e-01
9.9193990e-01 -7.2325015e-01 1.9050639e-02 3.1001231e-01
-6.1741853e-01 -4.5587007e-02 8.8576370e-01 -8.6064619e-01
-1.0703396e+00 -5.7189729e-02 6.4903103e-02 -9.4912076e-01
2.9329978e-02 4.4735777e-01 -2.7438445e-02 9.9996038e-02
1.2341652e+00 6.0627687e-01 2.1771269e-01 -3.2363215e-01
5.5347496e-01 5.6252062e-01 4.4276533e-01 -3.0767372e-01
5.5971605e-01 6.3034922e-01 -2.3924242e-01 -1.1912415e+00
-9.6738571e-01 -8.5431784e-01 -3.7644881e-01 -1.8638684e-01
5.7581812e-01 -1.1477540e+00 -1.7617171e-01 2.7093318e-01
-1.0669800e+00 -2.1095653e-01 -9.0058190e-01 3.8775781e-01
-9.5254171e-01 1.4775461e-01 -3.4993640e-01 -8.9955258e-01
-5.7352614e-01 -1.0550565e+00 7.5268847e-01 2.6469639e-01
-2.3689838e-01 -7.5630432e-01 2.7509484e-01 2.5545096e-01
4.4796610e-01 2.0769614e-01 8.0544126e-01 -8.6301112e-01
-3.4629092e-01 -2.3749267e-01 -2.1522561e-01 5.1875734e-01
-1.7242922e-01 -5.9983838e-01 -1.3207356e+00 -3.4549457e-01
2.2838591e-01 -1.0979840e+00 1.3797032e+00 1.8549524e-02
1.0275363e+00 -3.1930801e-01 -1.4119731e-01 4.9246770e-01
1.4280988e+00 4.2732656e-02 8.3246225e-01 -1.1558160e-01
3.8596961e-01 4.1455123e-01 6.3747239e-01 8.5297686e-01
-3.2205109e-02 4.3349579e-01 3.1194818e-01 1.5426353e-01
-6.2816042e-01 -4.4331235e-01 4.3114999e-01 6.8866593e-01
1.6568752e-01 2.7712164e-02 -8.4413725e-01 7.9558736e-01
-2.0194030e+00 -1.3961231e+00 2.7987716e-01 2.0028923e+00
9.1182232e-01 1.6873096e-01 -1.8991435e-02 -1.6045903e-01
4.5533130e-01 4.1607407e-01 -9.5769536e-01 -1.4931928e-01
8.0599882e-02 4.7785205e-01 5.0346670e-03 1.8731523e-01
-7.5466257e-01 1.0552179e+00 5.7358456e+00 1.0374230e+00
-1.0480446e+00 5.7982099e-01 5.0980192e-01 -3.9946729e-01
-2.2049476e-01 -7.0602290e-02 -7.4594253e-01 1.8665136e-01
7.3904568e-01 -4.7813225e-01 3.5186884e-01 1.2424313e+00
-2.5137864e-02 2.2902090e-02 -1.0824443e+00 1.0682080e+00
4.5504761e-01 -1.4916133e+00 3.9665511e-01 -2.1855736e-01
1.0881377e+00 1.7629279e-01 2.8394222e-01 8.9775187e-01
1.8332291e-01 -1.1741116e+00 5.5149877e-01 6.9135088e-01
6.5133226e-01 -7.4472332e-01 4.1254607e-01 5.3486681e-01
-7.3165911e-01 -2.3956849e-01 -8.1659770e-01 7.0898823e-02
1.6042152e-01 5.9492040e-01 -1.0952587e+00 1.7625453e-01
2.5637582e-01 7.2146648e-01 -5.1413471e-01 1.4557106e+00
-2.7811363e-01 4.0579745e-01 2.0595817e-01 -3.0712217e-01
3.9523914e-01 4.3432009e-01 5.9076196e-01 1.0257246e+00
2.5055858e-01 1.1062538e-02 1.2314548e-01 1.1774482e+00
-2.8191873e-01 -7.7842355e-02 -6.4938444e-01 -1.5265203e-01
4.4897416e-01 1.1290863e+00 -3.5941747e-01 -6.2381548e-01
-2.7773103e-01 9.2808115e-01 5.0925344e-01 2.8028706e-01
-5.9462160e-01 -4.5553303e-01 5.0932771e-01 2.6233867e-01
2.7368015e-01 6.2430635e-02 -3.7178162e-01 -1.3489231e+00
-3.4034455e-01 -9.7793788e-01 2.4814779e-01 -1.0151200e+00
-1.4707751e+00 4.4985816e-01 -1.3343285e-01 -1.2760620e+00
-2.2940670e-01 -3.3225033e-01 -1.0228294e+00 6.2386364e-01
-1.5436209e+00 -1.0870532e+00 -5.7688576e-01 6.0262972e-01
1.1144058e+00 -5.8428037e-01 8.3685821e-01 2.5797145e-02
-1.9017054e-01 4.8949501e-01 8.0665268e-02 -3.0657956e-01
6.7981654e-01 -1.0385158e+00 2.8033555e-01 5.9060150e-01
2.2648931e-01 6.1334550e-01 9.3216223e-01 -7.3601872e-01
-1.0996639e+00 -1.3434241e+00 5.6834918e-01 -2.6414272e-01
5.2513319e-01 -2.4403180e-01 -1.1086440e+00 3.5950682e-01
1.8015052e-01 8.6727023e-02 7.5114512e-01 -2.0714288e-03
-5.5695206e-01 8.0857046e-02 -1.0694181e+00 7.2148091e-01
9.2652190e-01 -5.0311273e-01 -9.9500215e-01 4.8592654e-01
7.2391659e-01 1.7584662e-01 -5.3396231e-01 2.6351929e-01
3.6949912e-01 -1.0122206e+00 9.9580216e-01 -9.1188389e-01
7.4418658e-01 -2.7327310e-02 -1.7170709e-01 -1.8171818e+00
-1.7823932e-01 -5.5316049e-01 -5.2968508e-01 6.5600127e-01
3.2432431e-01 -2.6736337e-01 7.7137667e-01 2.7276221e-01
-3.2372788e-01 -9.6539688e-01 -5.8873838e-01 -1.2952851e+00
1.1601951e-02 -3.3784077e-01 3.8295147e-01 9.4986194e-01
3.1321147e-01 7.3402810e-01 -8.5020411e-01 -4.0738592e-01
9.1067201e-01 -3.0813560e-01 8.8036728e-01 -1.1114522e+00
-6.0097200e-01 -3.6040482e-01 -3.1765509e-01 -6.7995054e-01
9.9561952e-02 -8.8520044e-01 1.2969723e-01 -1.5368327e+00
6.9594854e-01 -1.0770979e-02 -2.4863625e-01 5.5883068e-01
-2.7437413e-01 1.4960974e-01 3.8695595e-01 3.6218205e-01
-9.6343702e-01 1.1457179e+00 1.2360029e+00 -2.2830524e-01
-2.4667211e-01 5.0467547e-02 -8.6581063e-01 5.4354882e-01
7.7320445e-01 -5.0793260e-01 -8.0060250e-01 7.1462154e-02
-7.3131636e-02 1.5578832e-01 4.0718415e-01 -1.2284971e+00
4.4747058e-01 -7.6812603e-02 3.9098671e-01 -4.0576467e-01
9.0068382e-01 -3.6944053e-01 -1.2963103e-01 8.3495879e-01
-5.1403850e-01 -4.7219276e-01 -9.5105648e-02 8.0293083e-01
-2.2688113e-01 -5.2387947e-01 1.2120949e+00 -5.6091923e-01
-1.0574594e+00 5.0501037e-01 -1.4084794e-01 3.2238686e-01
1.3124382e+00 -2.6088396e-01 -5.8774590e-01 -5.4800266e-01
-8.4501666e-01 2.5967881e-01 2.9415172e-01 5.8645207e-01
1.1915944e+00 -1.2801875e+00 -8.3949286e-01 1.8032707e-02
6.9989240e-01 -5.1671213e-01 3.0267543e-01 7.0787752e-01
-3.3043507e-02 6.2817402e-02 -4.6024817e-01 -2.3771352e-01
-1.0543177e+00 7.9374003e-01 1.7621660e-01 -2.4050990e-01
-6.7336649e-01 8.1440085e-01 2.0908959e-02 -2.9210269e-01
4.0054157e-01 2.0736617e-01 -1.4252519e-02 1.9004132e-01
9.5792311e-01 5.7873243e-01 -1.0561438e-01 -9.7065277e-02
1.2716423e-01 -2.7834159e-01 -4.9675915e-01 -1.9021870e-01
1.6326230e+00 1.6477595e-01 3.9235735e-01 6.9874865e-01
1.2565690e+00 -7.3847985e-01 -1.3799660e+00 -4.8125935e-01
-9.5671527e-02 -5.0559020e-01 -7.2712381e-04 -8.5910261e-01
-6.1513174e-01 1.2717072e+00 4.5922509e-01 -1.0915472e-01
7.2313726e-01 6.4254198e-03 9.2340165e-01 5.6517202e-01
4.7593620e-01 -1.2267518e+00 8.9336431e-01 5.2632087e-01
1.0045471e+00 -1.2228410e+00 2.8626675e-02 -2.3161722e-02
-1.1844636e+00 9.8111588e-01 7.9099983e-01 -1.8647736e-01
5.3982586e-01 -1.2939088e-01 -3.4022987e-01 -3.3113024e-01
-1.2227722e+00 -3.0424446e-01 2.6435590e-01 7.2693348e-01
-1.3109245e-02 -6.1505865e-02 -9.3197137e-02 6.6040766e-01
8.8910677e-02 2.5755060e-01 6.7079705e-01 7.8918338e-01
-1.0638883e+00 -5.4762864e-01 -5.1306233e-02 7.8045809e-01
3.8998693e-02 -1.9560610e-01 -2.8121674e-01 5.9373307e-01
3.2387439e-02 7.7759224e-01 -1.3744870e-01 -2.4337535e-01
2.2794279e-01 2.0376898e-01 4.9223918e-01 -1.2270406e+00
-4.9597442e-01 -1.9710560e-01 -1.2463570e-01 -4.6772635e-01
-6.5311678e-02 -4.0097350e-01 -1.0847955e+00 -8.3193809e-02
-3.6885688e-01 1.7030939e-01 4.0651357e-01 9.8653185e-01
3.0575860e-01 5.3299379e-01 5.2832049e-01 -7.1363944e-01
-1.3661017e+00 -1.1087180e+00 -8.1447029e-01 4.3503571e-01
7.9817615e-02 -5.8743161e-01 -2.9591712e-01 -2.3210448e-01] | [9.957225799560547, 2.869290351867676] |
424997a4-76a4-4941-8abc-3dd546b72d68 | dual-task-learning-by-leveraging-both-dense | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Park_Dual_Task_Learning_by_Leveraging_Both_Dense_Correspondence_and_Mis-Correspondence_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Park_Dual_Task_Learning_by_Leveraging_Both_Dense_Correspondence_and_Mis-Correspondence_CVPR_2022_paper.pdf | Dual Task Learning by Leveraging Both Dense Correspondence and Mis-Correspondence for Robust Change Detection With Imperfect Matches | Accurate change detection enables a wide range of tasks in visual surveillance, anomaly detection and mobile robotics. However, contemporary change detection approaches assume an ideal matching between the current and stored scenes, whereas only coarse matching is possible in real-world scenarios. Thus, contemporary approaches fail to show the reported performance in real-world settings. To overcome this limitation, we propose SimSaC. SimSaC concurrently conducts scene flow estimation and change detection and is able to detect changes with imperfect matches. To train SimSaC without additional manual labeling, we propose a training scheme with random geometric transformations and the cut-paste method. Moreover, we design an evaluation protocol which reflects performance in real-world settings. In designing the protocol, we collect a test benchmark dataset, which we claim as another contribution. Our comprehensive experiments verify that SimSaC displays robust performance even given imperfect matches and the performance margin compared to contemporary approaches is huge. | ['Jong-Hwan Kim', 'Seon-Hoon Lee', 'Ue-Hwan Kim', 'Jin-Man Park'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['scene-flow-estimation'] | ['computer-vision'] | [ 3.66001487e-01 -4.52492595e-01 -1.08555883e-01 -2.15326294e-01
-1.19548924e-01 -6.18896544e-01 9.33308482e-01 -1.84807759e-02
-4.96545047e-01 3.20051998e-01 -2.06556857e-01 -4.16431457e-01
2.18122318e-01 -7.65512943e-01 -7.22202241e-01 -4.19768333e-01
-2.49964193e-01 6.78592697e-02 9.96113122e-01 -2.89477080e-01
1.96038276e-01 5.79358935e-01 -1.86104429e+00 -1.31133378e-01
7.38033473e-01 8.92276704e-01 3.45806777e-02 7.84868538e-01
2.02186167e-01 5.84431291e-01 -5.37330985e-01 -2.70855725e-01
8.53468716e-01 -2.36383125e-01 -5.88594675e-01 2.75156051e-01
8.83271992e-01 -5.00706136e-01 -3.19541156e-01 1.09916055e+00
3.27423781e-01 1.65593307e-02 3.70736241e-01 -1.43440592e+00
-1.73310023e-02 -7.02655315e-02 -6.03470981e-01 4.84539151e-01
7.13790417e-01 6.21590078e-01 5.11940420e-01 -5.94747126e-01
9.76971686e-01 1.08194566e+00 9.08432186e-01 3.23871523e-01
-8.63828123e-01 -3.25674266e-01 3.80155504e-01 3.59512091e-01
-1.20383513e+00 -6.35674119e-01 7.24005997e-01 -3.53738099e-01
8.53150070e-01 3.89814615e-01 7.76726246e-01 1.11991882e+00
1.38600796e-01 6.16683543e-01 1.14188313e+00 -1.66633874e-01
4.03064519e-01 -7.25295544e-02 -1.89644024e-01 9.06191945e-01
5.61473668e-01 1.87345684e-01 -2.86991030e-01 -1.18194679e-02
5.42491555e-01 2.31617421e-01 -4.54180658e-01 -6.59678042e-01
-1.50410151e+00 3.10631424e-01 5.66693366e-01 1.33238450e-01
-2.88533866e-01 8.74792188e-02 3.80188525e-01 4.79098946e-01
-3.21226083e-02 3.24948847e-01 -2.06107840e-01 -2.07533926e-01
-9.37500358e-01 1.03595495e-01 6.54920697e-01 1.15307713e+00
6.13978088e-01 -8.40431824e-02 -1.13403156e-01 2.97847420e-01
5.34324870e-02 9.32252467e-01 2.77881712e-01 -1.06761992e+00
4.77200449e-01 6.13450825e-01 3.49421352e-01 -1.48731112e+00
-4.52137202e-01 -3.39619130e-01 -8.89408231e-01 3.04807335e-01
6.51853442e-01 1.97698874e-03 -8.51928234e-01 1.44716632e+00
7.12266147e-01 6.25606418e-01 8.61740559e-02 8.27364266e-01
4.73933101e-01 4.27449435e-01 -2.13598594e-01 -3.64437312e-01
1.18166721e+00 -8.69181395e-01 -6.11059844e-01 -4.48772967e-01
5.22687137e-01 -5.84436834e-01 1.14825046e+00 3.00611407e-01
-7.67977953e-01 -4.38433826e-01 -1.18647075e+00 3.87303680e-01
-4.12006259e-01 -1.23932948e-02 6.56216502e-01 5.72532475e-01
-9.89654183e-01 2.99927950e-01 -1.05542123e+00 -8.28064978e-01
2.96028286e-01 3.57840285e-02 -4.33653116e-01 -6.07565083e-02
-7.72750795e-01 8.83123577e-01 2.41860807e-01 4.94996458e-02
-8.91338825e-01 -4.45033014e-01 -9.26499426e-01 -3.64847153e-01
6.81030810e-01 -6.76110148e-01 1.15481222e+00 -7.24139214e-01
-1.42427886e+00 9.42251980e-01 -2.06014723e-01 -7.16648161e-01
8.53256047e-01 -9.53873023e-02 -3.87744814e-01 1.60413012e-01
7.86303878e-02 5.83093643e-01 9.06692982e-01 -1.21516597e+00
-1.05618572e+00 -1.19432867e-01 3.12895119e-01 9.52894241e-02
-6.77271858e-02 -1.70805722e-01 -8.09646666e-01 -5.19255400e-01
3.34417909e-01 -1.05061007e+00 -3.41318518e-01 2.39586040e-01
-2.45809868e-01 3.27666581e-01 1.12153780e+00 -3.86803895e-01
1.29194605e+00 -2.09671617e+00 -3.47911626e-01 2.62837440e-01
2.26679668e-02 5.55972219e-01 -7.74321258e-02 1.94323465e-01
3.70913506e-01 -2.39916280e-01 -4.92057562e-01 -3.06753635e-01
-9.90950465e-02 2.70704180e-01 -2.57596701e-01 7.87925482e-01
1.08661905e-01 6.32813632e-01 -1.00230682e+00 -4.60409522e-01
5.08876562e-01 2.98105268e-04 -5.95952451e-01 -6.53172061e-02
-1.62417650e-01 2.11077318e-01 -2.48238802e-01 9.27061081e-01
8.84731233e-01 -1.83571801e-01 9.57142487e-02 -1.05611742e-01
-1.33705899e-01 5.22327945e-02 -1.47897696e+00 1.66698885e+00
-1.78668603e-01 9.24522758e-01 -1.65106818e-01 -7.62274444e-01
7.75999486e-01 -1.34943604e-01 4.87178594e-01 -8.70738029e-01
3.05353594e-03 5.52246533e-02 -1.46175548e-01 -5.62961280e-01
7.09034145e-01 5.24387956e-01 -2.04073172e-02 4.98939641e-02
-3.60646963e-01 -1.68678090e-01 2.67129809e-01 1.84169680e-01
1.59903598e+00 1.37270555e-01 7.48839676e-01 -5.18601947e-02
5.91649592e-01 2.93471366e-01 7.69781053e-01 9.85702932e-01
-7.86734700e-01 4.54766095e-01 2.58519232e-01 -6.58448517e-01
-7.37000823e-01 -1.02147841e+00 7.22119659e-02 6.70376182e-01
8.84380043e-01 -4.47113574e-01 -7.54985988e-01 -8.36489797e-01
-2.00685397e-01 3.71681511e-01 -4.49646860e-01 -4.07972448e-02
-6.27149522e-01 -7.30712056e-01 4.62386996e-01 2.65193969e-01
9.69232023e-01 -9.32749748e-01 -1.36049581e+00 -9.56847984e-03
-1.85571492e-01 -1.55044591e+00 -3.89276206e-01 -4.60392952e-01
-6.44253373e-01 -1.51838434e+00 -1.16547309e-01 -6.18973911e-01
6.47247493e-01 8.60044837e-01 1.06987405e+00 2.21691146e-01
-4.66522217e-01 6.20199323e-01 -3.82920176e-01 -2.03580588e-01
-2.98905224e-01 -6.48545623e-02 3.41201983e-02 -7.08166137e-02
2.66825080e-01 -3.67627084e-01 -8.96432877e-01 5.83796740e-01
-9.64229882e-01 5.94820920e-03 4.81236905e-01 4.95189905e-01
3.66604298e-01 -2.94053331e-02 1.31697908e-01 -5.82125664e-01
-1.74037404e-02 -2.46174857e-01 -9.35990274e-01 3.77570689e-01
-6.04406655e-01 -1.10418104e-01 3.68643194e-01 -3.87554377e-01
-1.04818451e+00 2.10777596e-01 1.35737851e-01 -1.53095722e-01
-4.20193672e-01 1.75635695e-01 -2.24192172e-01 -2.28003547e-01
7.46323049e-01 1.73686102e-01 -9.85031128e-02 -1.05747990e-01
1.72816485e-01 4.38249171e-01 1.05918896e+00 -1.27328664e-01
1.09115541e+00 9.22020853e-01 -1.40316123e-02 -8.53616536e-01
-5.78975797e-01 -7.17759550e-01 -6.20431364e-01 -4.78351772e-01
5.00251174e-01 -8.04184437e-01 -5.85538805e-01 7.05944896e-01
-1.13750947e+00 -3.87871802e-01 -2.99787909e-01 3.93459320e-01
-5.58065534e-01 5.16182959e-01 -1.60418540e-01 -6.24821424e-01
-2.19377995e-01 -1.15007591e+00 1.11158657e+00 2.76208639e-01
1.03626885e-02 -8.96842480e-01 1.18111275e-01 -2.74069211e-03
5.43235719e-01 7.95469522e-01 4.17854302e-02 -2.78250396e-01
-9.67184246e-01 -1.68484896e-01 -1.00894354e-01 -1.44440129e-01
4.06589150e-01 2.58736014e-01 -8.61924708e-01 -4.86402929e-01
-2.82435447e-01 3.88756812e-01 7.72828817e-01 2.14599356e-01
7.17669070e-01 -1.22395262e-01 -4.98167306e-01 7.74506927e-01
1.47539723e+00 3.50776762e-01 7.52188265e-01 7.99312472e-01
5.10820031e-01 3.20525944e-01 8.95082474e-01 4.76898760e-01
5.58152854e-01 8.81108701e-01 6.68711305e-01 -1.46221779e-02
-3.45637262e-01 -1.31470203e-01 6.26221538e-01 4.37746942e-01
1.21038496e-01 -3.30218852e-01 -1.01417136e+00 5.57468534e-01
-1.84814715e+00 -1.15651691e+00 -2.44021013e-01 2.35872054e+00
4.82547402e-01 3.71772259e-01 3.64621058e-02 1.53878510e-01
6.40349925e-01 2.35632211e-01 -6.36897326e-01 1.80075839e-01
-3.25109780e-01 -2.70971715e-01 8.40289235e-01 3.69381070e-01
-1.44201791e+00 1.07225847e+00 6.77258682e+00 4.14836645e-01
-1.28400612e+00 1.91329066e-02 3.79134044e-02 8.55470896e-02
-4.92540607e-03 5.01852036e-02 -6.06601834e-01 5.25779605e-01
5.18434882e-01 1.29222587e-01 2.11780548e-01 6.34558380e-01
3.13311368e-01 -3.75681400e-01 -9.16490436e-01 1.18800652e+00
1.59311548e-01 -1.24123740e+00 -1.52898341e-01 -2.39461958e-01
8.21421981e-01 1.30655289e-01 -1.09500147e-01 7.16942623e-02
3.04611564e-01 -4.71026748e-01 6.30959034e-01 6.04929924e-01
7.22717106e-01 -2.14540467e-01 8.16353977e-01 2.31373161e-01
-1.20148456e+00 6.47051707e-02 -1.18849963e-01 -8.32678378e-02
1.57677531e-01 3.30709428e-01 -9.68364239e-01 6.44836605e-01
8.16246927e-01 1.00716233e+00 -1.14616740e+00 1.47709346e+00
-1.23825669e-01 5.88565171e-01 -5.44722676e-01 1.99135110e-01
1.53324336e-01 -1.71143889e-01 8.27444851e-01 1.32954550e+00
4.21307921e-01 -2.39050388e-01 3.62433165e-01 2.12050200e-01
2.22047627e-01 -7.79695511e-02 -8.59466553e-01 6.14256561e-01
6.48014188e-01 9.19367790e-01 -9.51609135e-01 -4.17272210e-01
-3.84273887e-01 1.17821050e+00 -1.50956914e-01 3.35159928e-01
-1.00930536e+00 -3.62537056e-01 7.31027663e-01 1.02530889e-01
2.16645107e-01 -3.36426646e-01 -9.97546613e-02 -1.31325364e+00
3.40000778e-01 -8.42730343e-01 4.05324906e-01 -4.40830976e-01
-7.37673044e-01 5.41571200e-01 2.17241675e-01 -1.58572614e+00
-3.17277938e-01 -5.26531577e-01 -6.26918614e-01 3.23833376e-02
-1.54520833e+00 -9.32677746e-01 -9.88678455e-01 6.23563051e-01
5.78610957e-01 -9.69268605e-02 3.58488917e-01 3.17544460e-01
-7.20838010e-01 4.96393055e-01 -2.00433843e-02 4.38855402e-02
7.65916944e-01 -1.09932327e+00 7.93517232e-01 1.67872083e+00
-1.71464309e-02 2.59819537e-01 1.02941108e+00 -5.89875102e-01
-1.40784383e+00 -1.28777742e+00 3.26982975e-01 -3.93138558e-01
6.02701843e-01 -1.50184944e-01 -8.17693591e-01 6.11177981e-01
5.23348302e-02 3.18798751e-01 6.85231294e-03 -4.74034131e-01
-3.53960425e-01 -3.73899877e-01 -1.29750359e+00 7.34815836e-01
1.54173875e+00 -2.20997229e-01 -3.79161596e-01 4.86648008e-02
6.63041711e-01 -6.65251195e-01 -4.85466421e-01 6.24099553e-01
5.12849689e-01 -1.21514595e+00 9.29885209e-01 -2.53657937e-01
-1.53299049e-01 -9.42832947e-01 -2.16433659e-01 -1.05910373e+00
1.58602312e-01 -7.53740311e-01 -2.51788884e-01 1.07295108e+00
5.14044464e-02 -8.18131149e-01 6.38379097e-01 3.18764508e-01
-8.46525505e-02 -2.68713802e-01 -8.26890349e-01 -9.93809879e-01
-6.57131314e-01 -5.32729626e-01 6.47291780e-01 9.42279756e-01
-4.93207425e-01 -1.84421211e-01 -4.12390500e-01 5.12511075e-01
6.58179879e-01 1.00643218e-01 1.26720810e+00 -1.06944942e+00
-6.89050257e-02 -3.89105022e-01 -8.83593261e-01 -1.12657166e+00
-1.82316497e-01 -2.99987614e-01 2.14455083e-01 -1.46351886e+00
-1.60485704e-03 -6.15486979e-01 1.59434110e-01 2.78465152e-01
-2.36068472e-01 2.51099199e-01 2.25740746e-01 2.11420223e-01
-1.06481028e+00 2.99216211e-01 1.01106334e+00 -3.26657772e-01
-3.12794387e-01 4.64201421e-02 -1.58364162e-01 8.16017210e-01
1.01493168e+00 -2.19160303e-01 -4.79719400e-01 -3.32698494e-01
1.06264409e-02 -3.16838801e-01 7.13781059e-01 -1.49846101e+00
4.05159742e-01 -4.89604831e-01 -3.43217514e-03 -5.87670326e-01
5.21807782e-02 -9.77641940e-01 2.74716884e-01 6.56251132e-01
1.24659814e-01 2.38347501e-01 1.08662799e-01 6.85392439e-01
-1.38903454e-01 -3.37561108e-02 7.16543794e-01 2.09825456e-01
-1.26855111e+00 3.78977984e-01 -2.90861815e-01 2.83147059e-02
1.31657815e+00 -4.70623195e-01 -4.97114986e-01 -2.78155327e-01
-1.86547235e-01 2.56589264e-01 1.01455915e+00 5.08340061e-01
5.36832094e-01 -1.14817536e+00 -3.63694966e-01 2.90012002e-01
3.69481236e-01 1.21931404e-01 3.50397006e-02 8.80017161e-01
-9.63511944e-01 2.68082201e-01 -2.36137241e-01 -9.55017924e-01
-1.40665317e+00 7.20313132e-01 3.88243288e-01 4.90962602e-02
-7.57028699e-01 8.56703967e-02 -8.66767243e-02 -3.04302305e-01
3.81229848e-01 -6.45204306e-01 -1.29692340e-02 -3.03409666e-01
6.22587442e-01 4.78925973e-01 7.99473673e-02 -6.12498462e-01
-5.70443571e-01 6.43328846e-01 2.99999267e-01 8.37627053e-02
9.05087650e-01 -3.66133749e-01 2.24304408e-01 1.13264024e-01
9.10927713e-01 7.38162696e-02 -1.59608769e+00 -4.77046937e-01
1.78149566e-01 -6.52417004e-01 -2.58220285e-01 -3.83820742e-01
-9.58939612e-01 4.01889831e-01 1.01886499e+00 2.33007610e-01
1.50631142e+00 -2.68626064e-01 6.53512299e-01 7.80202448e-01
7.12848127e-01 -1.00482953e+00 9.51444544e-03 5.03968596e-01
7.01791763e-01 -1.47818196e+00 -6.31095245e-02 -5.20665228e-01
-5.62124252e-01 7.96049237e-01 7.50807881e-01 -3.71370614e-02
4.16993409e-01 2.68332273e-01 1.29822373e-01 -1.37552038e-01
-6.42569244e-01 -5.35762131e-01 1.41314477e-01 7.99552202e-01
-2.74878472e-01 -3.73638012e-02 -7.42426142e-02 -3.17483544e-01
-2.01721355e-01 -2.12594405e-01 6.85127914e-01 1.12448871e+00
-4.97346729e-01 -7.31875837e-01 -5.56807578e-01 1.86289921e-01
-7.03275800e-02 3.04336339e-01 -2.15180054e-01 9.70706463e-01
1.19192176e-01 9.99403894e-01 2.01853797e-01 -4.24591154e-01
6.56614721e-01 -3.87834579e-01 2.57926196e-01 -1.94333978e-02
-2.79412061e-01 -3.29863936e-01 -9.35472474e-02 -1.05274081e+00
-8.94044459e-01 -8.73922467e-01 -1.00609446e+00 -3.71879756e-01
-6.86710030e-02 -1.75091401e-01 4.77503389e-01 7.59892046e-01
4.98155832e-01 3.77629220e-01 6.92611217e-01 -6.62132919e-01
-2.32493043e-01 -7.15467036e-01 -6.16342537e-02 5.90640247e-01
6.78888857e-01 -8.68094563e-01 -3.39629233e-01 3.93851250e-01] | [8.581045150756836, -1.2941335439682007] |
a110d0e6-4c00-4707-ad89-ee3c1fa8f276 | from-pixels-to-ui-actions-learning-to-follow | 2306.00245 | null | https://arxiv.org/abs/2306.00245v1 | https://arxiv.org/pdf/2306.00245v1.pdf | From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces | Much of the previous work towards digital agents for graphical user interfaces (GUIs) has relied on text-based representations (derived from HTML or other structured data sources), which are not always readily available. These input representations have been often coupled with custom, task-specific action spaces. This paper focuses on creating agents that interact with the digital world using the same conceptual interface that humans commonly use -- via pixel-based screenshots and a generic action space corresponding to keyboard and mouse actions. Building upon recent progress in pixel-based pretraining, we show, for the first time, that it is possible for such agents to outperform human crowdworkers on the MiniWob++ benchmark of GUI-based instruction following tasks. | ['Kristina Toutanova', 'Kenton Lee', 'Urvashi Khandelwal', 'Hexiang Hu', 'Panupong Pasupat', 'Jonathan Berant', 'James Cohan', 'Mandar Joshi', 'Peter Shaw'] | 2023-05-31 | null | null | null | null | ['instruction-following'] | ['natural-language-processing'] | [ 5.38383961e-01 3.11147392e-01 1.32514626e-01 -4.94000822e-01
-5.26233673e-01 -6.82212651e-01 8.87841940e-01 2.05850735e-01
-7.48988211e-01 6.17780089e-01 3.47083330e-01 -8.77522647e-01
-1.11183479e-01 -7.50667274e-01 -5.22148788e-01 -1.46149814e-01
1.96152061e-01 6.21648192e-01 5.62567651e-01 -5.60634673e-01
5.44748366e-01 5.56919754e-01 -2.23294950e+00 7.94843137e-01
8.73009324e-01 4.61226583e-01 6.20546103e-01 1.28895915e+00
-4.03564900e-01 1.15611339e+00 -1.15102720e+00 -1.32238805e-01
1.11253142e-01 -3.21854502e-01 -6.51787341e-01 -7.68311247e-02
8.57027292e-01 -4.96848583e-01 -1.99590713e-01 6.96221709e-01
5.37324250e-01 5.43552697e-01 4.41819787e-01 -1.27576423e+00
-6.98005378e-01 4.47023660e-01 -1.01375774e-01 7.36853033e-02
1.05841196e+00 6.48725331e-01 7.15887785e-01 -3.60519916e-01
8.89583051e-01 1.46175706e+00 5.35731494e-01 7.41124749e-01
-1.30309939e+00 -4.56141293e-01 2.43759677e-01 -2.07766831e-01
-6.70958579e-01 -3.88074607e-01 3.99521559e-01 -4.65788394e-01
1.72792506e+00 4.74002540e-01 5.40359020e-01 1.40261650e+00
-2.37012625e-01 1.04241860e+00 1.13750052e+00 -1.01072884e+00
-6.84959367e-02 1.99417114e-01 3.79720956e-01 9.79754865e-01
1.59065366e-01 2.89784312e-01 -6.20049477e-01 -4.70403098e-02
1.18188798e+00 -1.34379685e-01 -3.48678648e-01 -5.98237813e-01
-1.25157475e+00 6.20472670e-01 2.15018198e-01 3.02075535e-01
-3.00323546e-01 1.80296570e-01 2.09503636e-01 5.35643578e-01
-3.86159457e-02 8.32744896e-01 -4.38352525e-01 -6.93081915e-01
-2.42210597e-01 4.84720796e-01 1.01121318e+00 1.25704932e+00
4.74981964e-01 -1.08071174e-02 -2.67560869e-01 4.07948732e-01
3.84479105e-01 2.16355965e-01 7.33612001e-01 -1.14217305e+00
8.07380855e-01 7.95234919e-01 7.64562011e-01 -6.87535644e-01
-5.15704155e-01 3.56652945e-01 -1.90195721e-02 1.27580512e+00
1.03045821e+00 -3.43785316e-01 -9.44695532e-01 1.39494443e+00
2.83159256e-01 -3.06051642e-01 4.03821990e-02 3.58050048e-01
9.27870035e-01 3.26748699e-01 4.38351184e-01 3.49017143e-01
1.25130773e+00 -9.82244313e-01 -8.56644332e-01 -1.54194236e-01
1.09035623e+00 -5.51673353e-01 1.85467279e+00 7.06665397e-01
-1.10053134e+00 -8.25948119e-01 -1.30930781e+00 -2.82904029e-01
-1.07667911e+00 -2.57863253e-02 8.26805294e-01 1.04133940e+00
-1.19249737e+00 5.45352340e-01 -9.94727075e-01 -6.54296696e-01
1.27817616e-01 1.25054926e-01 -3.69046807e-01 2.57354170e-01
-6.11627400e-01 1.10992777e+00 3.88729721e-01 -5.78307271e-01
-6.62221670e-01 -6.09470427e-01 -9.60226655e-01 -3.00355434e-01
4.23254073e-01 -6.72036231e-01 1.93750954e+00 -8.90489340e-01
-1.95227253e+00 6.86125994e-01 2.86526024e-01 -2.25464955e-01
7.27167487e-01 -5.66437840e-01 -1.58693463e-01 -8.65754560e-02
-2.93246508e-01 6.29837215e-01 7.17246890e-01 -1.47263026e+00
-1.03217936e+00 -3.87828976e-01 5.89118242e-01 4.13907558e-01
-3.24177265e-01 3.25997114e-01 -3.43430817e-01 -5.75581789e-01
-5.37832022e-01 -8.29238892e-01 -1.65310070e-01 -5.65582924e-02
-2.41387516e-01 -4.46758568e-01 7.62022913e-01 -6.10944927e-01
1.28174710e+00 -1.82841742e+00 -4.55339774e-02 1.08347766e-01
1.12342626e-01 5.21740854e-01 -2.48109773e-01 5.20651877e-01
-4.37662713e-02 -9.43361148e-02 2.22624004e-01 -3.48245889e-01
4.22055364e-01 1.02931648e-01 -7.77925691e-03 -1.80409402e-01
-1.88441023e-01 6.71569169e-01 -1.05803454e+00 -3.64549279e-01
5.76503873e-01 3.73227656e-01 -5.02338350e-01 4.33888882e-01
-6.37022078e-01 1.83184162e-01 -4.73447531e-01 1.89558104e-01
1.92956820e-01 1.07289396e-01 3.13418478e-01 3.65709603e-01
-4.42840576e-01 5.38294315e-01 -1.35556269e+00 2.14736080e+00
-6.38778150e-01 8.14044774e-01 2.15669870e-01 -3.58654946e-01
4.94552314e-01 3.71019274e-01 2.42319349e-02 -7.53254473e-01
-7.12417290e-02 -1.56918809e-01 -1.63365155e-02 -9.17550385e-01
5.70819318e-01 4.41189021e-01 1.70530602e-01 8.59416127e-01
-6.40892610e-02 -4.19065267e-01 5.24211049e-01 1.95087940e-01
1.41496849e+00 8.93127263e-01 3.11294436e-01 1.29951715e-01
1.65242225e-01 4.26703066e-01 -3.65364462e-01 1.31965673e+00
-1.33130290e-02 5.86406589e-01 1.83512077e-01 -4.11703408e-01
-8.96757185e-01 -9.01167572e-01 2.65089095e-01 1.87027550e+00
-3.12081993e-01 -6.93519771e-01 -9.66777921e-01 -8.57322991e-01
-1.30400628e-01 1.34677231e+00 -5.02602637e-01 3.27169329e-01
-7.14276791e-01 -2.57110387e-01 4.64226067e-01 7.71295071e-01
1.28430516e-01 -1.64460588e+00 -1.39188671e+00 3.01919788e-01
4.23530251e-01 -8.52483571e-01 -1.67173594e-01 4.60969776e-01
-7.89447486e-01 -1.23244691e+00 -5.34995377e-01 -7.20702708e-01
7.21844137e-01 3.44017237e-01 1.52226853e+00 4.99456674e-01
-4.41805750e-01 8.44342828e-01 -5.84339976e-01 -8.62792373e-01
-5.68381667e-01 -1.57744363e-01 -3.38176280e-01 -6.64291322e-01
3.87618363e-01 -2.52499372e-01 -2.63240039e-01 1.14077546e-01
-8.60218644e-01 5.13694465e-01 4.85527664e-01 5.78653097e-01
-8.08277875e-02 -1.12124942e-01 -2.08329372e-02 -1.38854468e+00
1.16480911e+00 1.32834539e-01 -6.57508433e-01 2.65898675e-01
-3.08229953e-01 4.85499725e-02 4.46788937e-01 -5.22542834e-01
-1.51573634e+00 1.74360394e-01 -7.22183883e-02 2.90679842e-01
-1.07049930e+00 2.41018340e-01 -2.20181778e-01 -4.22496088e-02
1.20685446e+00 -2.44040772e-01 -7.50185847e-02 -6.07757330e-01
6.52537942e-01 8.73438060e-01 5.12633681e-01 -7.44854510e-01
6.05236053e-01 1.09660558e-01 -5.06744027e-01 -8.46601844e-01
-3.54988188e-01 -3.05696756e-01 -6.51705682e-01 -1.03947751e-01
1.00399435e+00 -3.65940630e-01 -9.84839976e-01 4.25381899e-01
-1.25301671e+00 -1.26705515e+00 -4.20057178e-01 4.20031071e-01
-8.13640893e-01 1.54378498e-02 -4.54412192e-01 -9.68195140e-01
3.01748037e-01 -1.25671661e+00 9.39756334e-01 3.55741858e-01
-8.11534464e-01 -9.40332353e-01 9.56214070e-02 4.45710212e-01
2.47937173e-01 1.99062899e-01 1.15153801e+00 -8.35310817e-01
-3.99989992e-01 -3.62597287e-01 -9.52880234e-02 -6.11739270e-02
3.46864499e-02 1.29541650e-01 -7.99701452e-01 2.27416609e-03
-3.93229485e-01 -6.29000604e-01 3.05319607e-01 -4.33702208e-02
1.04884732e+00 -1.36648744e-01 -5.51325500e-01 2.81526029e-01
9.97841477e-01 7.59343147e-01 5.87825596e-01 8.29065800e-01
6.21648908e-01 5.72467029e-01 4.58411694e-01 3.62402678e-01
2.96262771e-01 7.55095124e-01 1.66266978e-01 -2.12469906e-01
-2.97232240e-01 -3.66299748e-01 3.49289447e-01 -1.79066658e-02
-2.39272013e-01 -5.73513269e-01 -1.02953827e+00 2.01572537e-01
-2.09049201e+00 -1.07418394e+00 -2.23693699e-01 2.30773163e+00
9.40606594e-01 2.90256709e-01 2.22816050e-01 5.52628227e-02
3.38660479e-01 -1.00326024e-01 -3.81945014e-01 -3.58963072e-01
3.18319231e-01 5.09879768e-01 1.66496128e-01 6.05079055e-01
-1.02219796e+00 7.49387860e-01 7.09094095e+00 3.34104091e-01
-4.85776424e-01 -1.32623628e-01 2.01897696e-02 5.76816238e-02
9.19603556e-02 -4.12329674e-01 -7.18120277e-01 1.29164144e-01
8.04879963e-01 2.31906325e-01 6.20336175e-01 8.93088341e-01
2.10101545e-01 -4.39241469e-01 -1.42085707e+00 8.09466839e-01
3.55679020e-02 -1.08586693e+00 2.85214657e-04 -1.90179542e-01
6.21033430e-01 -1.67401448e-01 2.27289200e-01 5.00257969e-01
1.31778073e+00 -1.14727890e+00 6.90048397e-01 5.20150959e-01
6.54511869e-01 -4.13115591e-01 2.25990444e-01 4.88143623e-01
-7.79128432e-01 -6.84105381e-02 9.51256827e-02 -4.07266766e-01
-2.32186362e-01 -6.51183903e-01 -1.07821012e+00 6.72422498e-02
7.62459934e-01 1.01436906e-01 -7.96486437e-01 1.16369402e+00
-4.32291031e-01 2.03164801e-01 -1.46878496e-01 -4.39779639e-01
1.99353293e-01 -1.20130144e-02 2.64619946e-01 1.52845025e+00
2.06632331e-01 1.28117308e-01 2.29288638e-01 3.95762205e-01
2.23036245e-01 -5.10595553e-02 -1.06739974e+00 -9.89633352e-02
2.66284972e-01 1.11813605e+00 -6.52681291e-01 -6.10790670e-01
-8.67093861e-01 8.42899144e-01 4.79454309e-01 6.99882805e-01
-6.46153092e-01 -1.04317987e+00 8.02074254e-01 2.58233905e-01
1.57225549e-01 -6.48160160e-01 -2.13236451e-01 -5.73022187e-01
-3.24618369e-01 -1.49546933e+00 4.27313983e-01 -1.33081698e+00
-9.14945960e-01 5.56307495e-01 1.91322863e-01 -7.42443204e-01
-4.78297055e-01 -1.11347127e+00 -7.67288446e-01 9.19595897e-01
-8.01080525e-01 -9.06012356e-01 -5.93021572e-01 5.80512881e-01
6.94721401e-01 -3.22146207e-01 1.23852730e+00 -1.72543064e-01
-4.23479140e-01 3.33869368e-01 5.62262088e-02 3.72565016e-02
6.59794152e-01 -1.91673362e+00 7.96162009e-01 4.86444980e-01
3.52218151e-01 9.66337025e-01 8.80737185e-01 -6.55852437e-01
-1.41502893e+00 -4.02396321e-01 3.06115061e-01 -1.20181847e+00
5.50692201e-01 -3.42321277e-01 -5.97690701e-01 1.00243604e+00
7.56917000e-01 -5.99000990e-01 7.89874852e-01 2.08257109e-01
-3.04454654e-01 4.64533478e-01 -1.01519632e+00 1.05406356e+00
1.35947967e+00 -5.77095449e-01 -1.07310605e+00 5.59660137e-01
6.03480637e-01 -5.27337968e-01 -1.98450193e-01 -2.10636765e-01
5.35370886e-01 -1.13126028e+00 1.19734836e+00 -1.21440816e+00
-1.14290863e-02 -3.58499646e-01 1.67117581e-01 -1.73864400e+00
-1.98374614e-01 -1.03713012e+00 -4.19691987e-02 1.01393151e+00
4.42128867e-01 -4.58265394e-01 7.40594268e-01 1.06158221e+00
-4.22780007e-01 1.21871494e-01 -3.54574203e-01 -6.00865006e-01
-3.71116489e-01 -5.74581802e-01 7.63047218e-01 6.45254195e-01
4.83423591e-01 3.33750218e-01 3.05835783e-01 -1.61888506e-02
2.12915957e-01 -3.08891028e-01 1.28954887e+00 -1.36632311e+00
-4.12337393e-01 -8.50267589e-01 -7.22529292e-02 -1.19678521e+00
9.42925271e-03 -5.56989551e-01 2.60953039e-01 -1.86947918e+00
-3.37738156e-01 -3.68619204e-01 -1.24360032e-01 4.34308439e-01
-1.21201605e-01 -1.61818400e-01 1.93254083e-01 -3.26251358e-01
-8.64861369e-01 -5.22418357e-02 1.26489449e+00 7.11677894e-02
-4.86305743e-01 -7.47895390e-02 -4.44608033e-01 1.14526343e+00
7.55938351e-01 -1.53726235e-01 -6.56567752e-01 -7.60961711e-01
9.74602029e-02 -1.52179580e-02 1.56420246e-01 -1.37720692e+00
4.27792788e-01 -3.82231236e-01 5.34792006e-01 -2.46446192e-01
3.72130901e-01 -9.05348063e-01 -1.39834553e-01 2.05371052e-01
-6.91679657e-01 3.85614783e-01 2.59540647e-01 4.50649261e-01
2.03412771e-01 -6.88270092e-01 2.21903235e-01 -5.86549997e-01
-9.27356184e-01 -3.27299327e-01 -8.90988946e-01 -1.08979248e-01
1.01359963e+00 -5.12338042e-01 -6.89024627e-01 -4.37792599e-01
-5.49227297e-01 -1.41159773e-01 6.42875731e-01 4.39706326e-01
4.81313348e-01 -1.02557993e+00 -1.75794929e-01 3.48762512e-01
2.31459796e-01 -1.82860613e-01 -3.59084994e-01 1.59043431e-01
-9.35834706e-01 4.65914249e-01 -5.50480902e-01 3.52370143e-02
-1.49059558e+00 7.13953614e-01 -9.47534223e-04 -7.59367868e-02
-5.53234696e-01 1.07063484e+00 1.03131458e-01 -5.67055941e-01
8.29615176e-01 -5.27759433e-01 -3.79445165e-01 -2.06233039e-02
1.02632499e+00 4.32916701e-01 6.50022924e-02 -1.09944910e-01
1.89440012e-01 1.31986409e-01 8.72187987e-02 -6.12935364e-01
1.29403973e+00 2.94825017e-01 6.27195001e-01 4.00859058e-01
4.68756020e-01 1.74865630e-02 -1.63443768e+00 2.90959403e-02
3.41396928e-01 -6.65995121e-01 -2.39039347e-01 -1.24487424e+00
-3.25889289e-01 9.74936187e-01 5.35968184e-01 7.01909482e-01
5.38530827e-01 -3.89097482e-01 3.13336760e-01 1.13658404e+00
4.13370639e-01 -1.42004180e+00 5.26239097e-01 5.06997764e-01
9.77219641e-01 -1.37342620e+00 -1.75982073e-01 -4.38868217e-02
-5.79935670e-01 1.23886073e+00 1.09481514e+00 3.36006612e-01
9.68084261e-02 7.41520643e-01 3.29893589e-01 -3.19756418e-01
-8.10346246e-01 -4.30850446e-01 4.72968502e-04 1.46217549e+00
9.88331139e-01 1.30916415e-02 3.26611727e-01 3.40051860e-01
-1.93061903e-01 5.03687076e-02 3.87321383e-01 1.55473745e+00
-6.14071488e-01 -1.15361619e+00 -6.46014988e-01 4.81988430e-01
-6.27792776e-02 -5.84705472e-02 -5.83954990e-01 1.01926565e+00
-6.92787021e-02 1.25316453e+00 1.72378108e-01 -4.98705134e-02
7.49249876e-01 5.27153552e-01 8.48417640e-01 -9.71112728e-01
-1.05281627e+00 -4.39905256e-01 4.88409698e-01 -6.29621804e-01
-5.62497154e-02 -6.19424284e-01 -1.22640228e+00 -1.00585423e-01
-4.71952930e-02 -2.03373313e-01 6.70951724e-01 4.90295321e-01
7.46868625e-02 7.03997374e-01 -3.59635353e-01 -1.32411897e+00
-4.97572303e-01 -1.16160107e+00 -4.03417870e-02 7.32458234e-01
1.22549996e-01 -6.08439803e-01 1.82437509e-01 2.50677705e-01] | [4.277980804443359, 0.9585561156272888] |
729ebca6-7490-468f-a16e-9e5685cb80f9 | quickest-change-detection-for-unnormalized | 2302.00250 | null | https://arxiv.org/abs/2302.00250v1 | https://arxiv.org/pdf/2302.00250v1.pdf | Quickest Change Detection for Unnormalized Statistical Models | Classical quickest change detection algorithms require modeling pre-change and post-change distributions. Such an approach may not be feasible for various machine learning models because of the complexity of computing the explicit distributions. Additionally, these methods may suffer from a lack of robustness to model mismatch and noise. This paper develops a new variant of the classical Cumulative Sum (CUSUM) algorithm for the quickest change detection. This variant is based on Fisher divergence and the Hyv\"arinen score and is called the Score-based CUSUM (SCUSUM) algorithm. The SCUSUM algorithm allows the applications of change detection for unnormalized statistical models, i.e., models for which the probability density function contains an unknown normalization constant. The asymptotic optimality of the proposed algorithm is investigated by deriving expressions for average detection delay and the mean running time to a false alarm. Numerical results are provided to demonstrate the performance of the proposed algorithm. | ['Vahid Tarokh', 'Jie Ding', 'Taposh Banerjee', 'Enmao Diao', 'Suya Wu'] | 2023-02-01 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 2.94638187e-01 -4.40372318e-01 1.08090080e-01 -1.75344065e-01
-6.90577686e-01 -4.40609008e-01 4.00849015e-01 5.13505638e-01
-3.30810398e-01 8.97569299e-01 -6.99729860e-01 -4.32708591e-01
-3.11355501e-01 -5.59866250e-01 -3.96237582e-01 -8.93115103e-01
-2.45932594e-01 6.97879270e-02 3.71318132e-01 1.93724334e-01
3.06304902e-01 6.27900481e-01 -1.68412769e+00 -5.33468902e-01
1.01306581e+00 1.13685536e+00 2.27613952e-02 1.00264037e+00
4.72342558e-02 2.07900956e-01 -8.00752699e-01 -9.75417495e-02
2.31734991e-01 -8.64675522e-01 -5.43919466e-02 -1.08053178e-01
6.17136294e-03 -1.32828027e-01 1.22906156e-01 1.22824562e+00
5.03867626e-01 3.71187091e-01 8.46784472e-01 -1.33091533e+00
-2.04524994e-02 3.14609617e-01 -8.24075878e-01 8.68805587e-01
2.31522590e-01 -4.38330352e-01 8.27903628e-01 -7.92263925e-01
2.55165637e-01 9.34511900e-01 6.40204787e-01 2.01336414e-01
-1.14350009e+00 -4.75342602e-01 -1.85640022e-01 3.15724611e-01
-1.68732667e+00 -1.62795261e-01 3.65029067e-01 -5.57989001e-01
5.87562382e-01 5.14834404e-01 1.90647721e-01 3.66223335e-01
6.07803702e-01 6.02811754e-01 1.05377185e+00 -7.52447128e-01
5.07045090e-01 3.12599055e-02 4.17137682e-01 5.30914545e-01
8.76295447e-01 5.36854640e-02 1.04942014e-02 -3.43653888e-01
5.55850506e-01 9.02901217e-02 -1.71544567e-01 -3.23586464e-01
-6.57425582e-01 5.95726550e-01 -2.04980940e-01 4.08177942e-01
-2.98605293e-01 2.28738971e-03 2.81852067e-01 2.28201509e-01
4.14306551e-01 4.57403176e-02 -1.11692972e-01 -2.60562629e-01
-1.19634962e+00 2.24949032e-01 8.60766590e-01 9.40114975e-01
4.23223644e-01 2.85659611e-01 -1.71684414e-01 4.25508052e-01
3.28189693e-02 7.49843061e-01 3.57425064e-01 -5.41271448e-01
6.46293312e-02 1.06037281e-01 5.28357744e-01 -7.90184796e-01
-4.13387209e-01 -7.23203182e-01 -7.99973011e-01 1.16698012e-01
5.44475615e-01 -5.41458726e-01 -5.92481494e-01 1.52584469e+00
2.88709998e-01 3.23821217e-01 -1.05354190e-01 5.13586164e-01
-1.18489869e-01 6.72588825e-01 -2.71746278e-01 -8.16265762e-01
7.49165893e-01 -3.39913160e-01 -9.36110079e-01 2.47827381e-01
3.34952444e-01 -7.39100635e-01 6.44200265e-01 4.03318197e-01
-8.97155583e-01 -3.10348362e-01 -1.24316037e+00 7.57279992e-01
-6.88291118e-02 1.39100567e-01 2.75071990e-02 1.00105441e+00
-8.91132534e-01 3.95592391e-01 -1.09182084e+00 -6.06024444e-01
-2.71983415e-01 9.73233283e-02 2.48365805e-01 3.80187221e-02
-8.70272160e-01 7.92331159e-01 2.29350060e-01 1.74479067e-01
-6.06519222e-01 -3.39654416e-01 -6.06104255e-01 4.47944641e-01
3.32558751e-01 -2.11995423e-01 1.40456223e+00 -1.07145500e+00
-1.29323912e+00 1.35607362e-01 -2.19305158e-01 -4.93637174e-01
6.59099817e-01 -3.60627100e-02 -4.71437544e-01 8.96606594e-03
-1.65345930e-02 -5.23699403e-01 9.46356773e-01 -8.49183619e-01
-7.32055068e-01 -1.25007778e-01 -2.84572303e-01 -3.30908522e-02
2.30529848e-02 5.02700694e-02 -1.36580495e-02 -7.32893407e-01
4.04305644e-02 -7.50099897e-01 -2.61997640e-01 -2.79577523e-01
-3.33392859e-01 3.05131283e-02 6.60005212e-01 -4.48747933e-01
1.69454455e+00 -2.25076175e+00 -5.68498313e-01 5.92209995e-01
-3.03746462e-01 1.76950812e-01 1.07897498e-01 6.99991584e-01
-1.57109216e-01 -1.68733418e-01 -2.54229933e-01 1.50151938e-01
-1.05336070e-01 5.38966507e-02 7.05757737e-02 7.47156739e-01
4.89234105e-02 -8.02233815e-02 -9.07412291e-01 -2.06700817e-01
2.01387808e-01 8.42108876e-02 -3.78677785e-01 1.11034423e-01
2.43641362e-01 8.77150595e-02 -2.84731537e-01 4.52514499e-01
6.91767871e-01 3.59214172e-02 1.64214894e-01 2.96907037e-01
-4.92222846e-01 -1.91339225e-01 -1.48649621e+00 6.28809214e-01
-2.80656040e-01 6.21287763e-01 3.36377509e-02 -1.09048653e+00
9.44288313e-01 3.05223465e-01 6.87549531e-01 -2.66367942e-01
4.34376925e-01 2.92163700e-01 1.69078842e-01 -1.49203077e-01
4.07053173e-01 -3.22633177e-01 -2.39208832e-01 2.94956952e-01
-1.85938090e-01 3.31014276e-01 4.74903405e-01 -5.57255074e-02
1.17642868e+00 -2.64129490e-01 9.92884338e-01 -5.45997083e-01
4.50503826e-01 -2.75967181e-01 7.02057958e-01 1.11679339e+00
-4.60700661e-01 1.21030614e-01 5.92150092e-01 8.48133788e-02
-7.39944696e-01 -1.39392281e+00 -3.72295737e-01 5.21491170e-01
2.15570584e-01 -1.09189264e-01 -7.47175336e-01 -1.74308777e-01
-1.37336906e-02 9.30415273e-01 -4.21529353e-01 -2.19844222e-01
-1.46090403e-01 -1.06365335e+00 2.96033740e-01 2.04977110e-01
1.36219248e-01 -3.01598370e-01 -8.92250240e-01 3.05114657e-01
1.07929677e-01 -8.29020262e-01 -1.79708600e-01 9.53541100e-02
-7.69476473e-01 -1.22847795e+00 -7.25095272e-01 -3.46199512e-01
7.07568467e-01 2.25537106e-01 3.95053118e-01 -3.27638805e-01
-3.54066491e-01 6.95466697e-01 -5.58785856e-01 -4.79510397e-01
-4.70722735e-01 -2.20757246e-01 3.65461230e-01 3.24667335e-01
2.27184340e-01 -2.76783615e-01 -3.52810651e-01 3.50328565e-01
-9.61823344e-01 -5.83140850e-01 5.64860821e-01 8.43621254e-01
5.96721411e-01 4.82642233e-01 9.39494550e-01 -6.78754508e-01
7.07574546e-01 -6.86534941e-01 -1.03505886e+00 3.01738828e-01
-1.02579832e+00 -1.08971782e-01 4.72302943e-01 -2.64460117e-01
-1.15214610e+00 -8.85096490e-02 2.34114841e-01 -1.47639990e-01
1.29106389e-02 6.56370282e-01 6.22602515e-02 -1.77555028e-02
4.44969952e-01 9.23892558e-02 -1.54145375e-01 -4.07630920e-01
4.75948267e-02 7.93505013e-01 5.55257201e-01 -2.94367164e-01
6.14770710e-01 1.55517906e-01 1.73266798e-01 -1.07687557e+00
-4.15924966e-01 -7.14278281e-01 -4.10607815e-01 -3.62538993e-01
4.52849925e-01 -4.32815313e-01 -4.56119269e-01 7.10332692e-01
-9.01878178e-01 6.47519110e-03 -1.21623136e-01 7.63201356e-01
-5.38136959e-01 6.40773714e-01 -4.26054567e-01 -1.42975867e+00
-1.34036511e-01 -6.25808001e-01 3.56238693e-01 5.03696442e-01
5.63371740e-03 -8.52210701e-01 1.25523150e-01 -5.44424832e-01
2.11393759e-01 5.93844712e-01 8.20573032e-01 -7.45914221e-01
-2.19325140e-01 -8.99432659e-01 -2.15380296e-01 3.60793203e-01
3.03305686e-01 2.96567619e-01 -4.12490636e-01 -4.11166608e-01
1.14696380e-02 6.71472192e-01 4.18700814e-01 7.84828246e-01
7.29697406e-01 -2.82349408e-01 -3.60588700e-01 7.42423907e-02
1.75224507e+00 6.08123660e-01 3.00231189e-01 3.48583944e-02
1.65587455e-01 2.40109600e-02 9.59469080e-01 1.07233286e+00
1.12442663e-02 5.10011911e-01 2.47979924e-01 2.91126072e-01
4.03718174e-01 2.61773080e-01 5.02348959e-01 7.44419217e-01
9.52947512e-02 -5.01754522e-01 -8.54635477e-01 6.75524116e-01
-1.69541502e+00 -1.09250975e+00 -4.51109976e-01 2.77511525e+00
5.14909625e-01 2.88060606e-01 1.40681505e-01 3.72602463e-01
1.13826060e+00 -3.24462891e-01 -5.52628875e-01 -6.86078250e-01
1.11716978e-01 3.06706309e-01 6.82910860e-01 4.90855455e-01
-1.04774880e+00 2.39388108e-01 6.55831575e+00 7.84993291e-01
-9.35934544e-01 2.23693669e-01 1.87433198e-01 7.24027492e-03
3.10837895e-01 -6.16240613e-02 -7.21273959e-01 6.48141205e-01
1.10755610e+00 -6.54697835e-01 4.03291993e-02 7.36351609e-01
5.12751162e-01 -8.52080166e-01 -7.82982171e-01 9.10288453e-01
1.85074389e-01 -4.20966774e-01 -4.88478869e-01 -1.97008327e-01
7.77882874e-01 -2.97219962e-01 -2.24557728e-01 1.51566118e-01
-8.08658004e-02 -1.37492761e-01 5.61277628e-01 7.71799803e-01
8.52761567e-01 -8.45316648e-01 9.85391855e-01 2.48524636e-01
-1.05623007e+00 -1.93968058e-01 -1.70572653e-01 -2.00354546e-01
3.06738019e-01 9.94632304e-01 -8.79418373e-01 5.95339596e-01
3.78492385e-01 2.66228080e-01 -2.89266139e-01 1.80834770e+00
1.57265902e-01 7.51467884e-01 -6.18862927e-01 -4.42713290e-01
4.66481671e-02 -1.06366590e-01 9.36287403e-01 1.20557308e+00
9.86077428e-01 -7.16032982e-02 8.22546035e-02 3.35760057e-01
4.89531070e-01 1.84962079e-01 -2.10763454e-01 3.73905450e-02
7.50655234e-01 8.53014112e-01 -9.91415858e-01 -3.05648327e-01
-4.77052659e-01 8.74069154e-01 -3.89582008e-01 4.74496603e-01
-7.95873821e-01 -8.08012784e-01 4.88217205e-01 1.83799908e-01
4.75455791e-01 -3.38537574e-01 -2.08103001e-01 -9.65686321e-01
-6.78581360e-04 -2.48290032e-01 6.82645440e-01 -3.20987761e-01
-1.11008298e+00 3.04022819e-01 3.62310469e-01 -1.51316106e+00
-4.17463094e-01 -2.90684491e-01 -5.99151790e-01 7.56618798e-01
-1.17308378e+00 -3.18087906e-01 -8.71593058e-02 4.92844254e-01
3.39154929e-01 2.40980089e-02 5.25134265e-01 4.75915402e-01
-8.94155383e-01 8.05902064e-01 9.00026143e-01 -2.48002365e-01
6.39725745e-01 -1.24938822e+00 -2.57653385e-01 1.43862391e+00
-3.65697414e-01 2.95302808e-01 1.39067817e+00 -6.71341658e-01
-8.43496859e-01 -8.45693886e-01 8.50416362e-01 9.59600881e-02
7.86451459e-01 3.86897512e-02 -6.62134469e-01 4.36860263e-01
-3.27253550e-01 -1.54297575e-01 7.82773614e-01 -6.99243397e-02
1.38020262e-01 -3.98599476e-01 -1.13523388e+00 3.36003929e-01
3.92579973e-01 -6.95768371e-02 -4.17826355e-01 3.25322524e-03
8.03568363e-02 -9.57370996e-02 -6.31062984e-01 2.29936779e-01
6.36168778e-01 -9.73612428e-01 2.63165504e-01 -1.74560368e-01
-3.28839093e-01 -4.45909023e-01 -2.25081742e-01 -1.22803497e+00
-3.28155667e-01 -7.44642794e-01 -1.76788848e-02 1.16989040e+00
2.69213915e-01 -7.47184277e-01 9.11008492e-02 2.94852108e-01
-7.78768808e-02 -4.46761519e-01 -1.19503129e+00 -1.13935709e+00
-4.54232603e-01 -5.09850383e-01 3.33738059e-01 8.28419566e-01
2.67077535e-01 -6.83167651e-02 -3.40498745e-01 4.12649214e-01
5.59977174e-01 7.39667863e-02 3.85396212e-01 -1.34452879e+00
-4.28622335e-01 -4.82039511e-01 -7.03866959e-01 -5.42805493e-01
-2.79000968e-01 -4.50636029e-01 3.77673566e-01 -1.11101460e+00
5.38768852e-03 -3.16313386e-01 -7.21112072e-01 7.46328309e-02
-3.67610514e-01 -2.65739679e-01 -1.23857059e-01 2.10005101e-02
-4.59196359e-01 4.30399537e-01 5.56676507e-01 4.39529985e-01
-3.70242596e-01 5.35497785e-01 -2.39782929e-01 5.59412181e-01
8.93465936e-01 -6.49362028e-01 -3.73463988e-01 1.41524389e-01
7.41357952e-02 3.55518073e-01 1.87196210e-01 -1.26098645e+00
5.89412451e-02 -2.38592505e-01 1.84872076e-02 -6.16377532e-01
-1.82207450e-01 -7.29591250e-01 2.31102586e-01 7.46349335e-01
2.57287566e-02 2.44380772e-01 6.41421080e-02 9.68640983e-01
-1.24351576e-01 -4.98126447e-01 1.02506781e+00 4.48580980e-01
-5.98733783e-01 2.48870999e-03 -9.06368017e-01 -3.29394698e-01
1.33695960e+00 -2.08393589e-01 -9.90222115e-03 -5.59206545e-01
-7.51074493e-01 7.51262382e-02 1.62573263e-01 3.41530368e-02
3.50288808e-01 -1.14421666e+00 -4.54365939e-01 7.82282501e-02
1.05253160e-02 -5.93646228e-01 3.61263186e-01 1.24080276e+00
-6.05121374e-01 2.77690083e-01 -1.23431906e-01 -5.03835857e-01
-1.29113770e+00 5.15135944e-01 3.26280862e-01 -1.78274393e-01
-8.25560614e-02 4.78714108e-01 -8.84380713e-02 2.47757971e-01
-1.29383793e-02 -4.08889800e-01 1.72903568e-01 2.54849475e-02
7.25174367e-01 8.74640048e-01 9.89642218e-02 -3.67215604e-01
-4.53487039e-01 1.70993850e-01 8.36511180e-02 -2.42452487e-01
7.55836368e-01 -4.33130026e-01 -1.46956950e-01 7.73334503e-01
7.89198935e-01 7.37553369e-03 -9.98496175e-01 -2.77144700e-01
2.27429032e-01 -3.85682791e-01 -2.89321784e-02 -5.38954318e-01
-4.51568455e-01 6.53890431e-01 8.79536450e-01 5.27691662e-01
1.39373100e+00 -4.09768373e-01 4.51054871e-01 2.37008423e-01
4.38679338e-01 -1.11257803e+00 -5.17054498e-01 4.42609906e-01
4.46119815e-01 -7.61629701e-01 -8.19834918e-02 -3.71104360e-01
-6.81992471e-02 1.07327306e+00 4.02255386e-01 -1.82530940e-01
1.07864070e+00 3.27357173e-01 -6.05097972e-02 4.60610896e-01
-4.43714648e-01 -5.13430417e-01 2.85932925e-02 5.32898247e-01
3.33193868e-01 4.42378908e-01 -1.18026900e+00 7.72145987e-01
1.31623715e-01 -7.94435516e-02 9.48952079e-01 9.89394546e-01
-6.78919911e-01 -8.39382946e-01 -5.09956837e-01 6.55196726e-01
-6.59947574e-01 1.36934876e-01 2.68629175e-02 6.98696613e-01
-9.70089361e-02 1.13362122e+00 2.11871952e-01 -2.69390494e-01
3.18724692e-01 6.42448366e-02 3.12657714e-01 -5.17208636e-01
7.78378993e-02 3.84467840e-01 -1.01426803e-01 -2.08201483e-01
-2.52249479e-01 -1.15879798e+00 -1.17343771e+00 -2.51040101e-01
-5.81440032e-01 4.24597770e-01 6.61422729e-01 1.01760340e+00
8.94449428e-02 4.84748960e-01 7.80886054e-01 -4.06507820e-01
-8.81242871e-01 -9.61536646e-01 -1.10736573e+00 -1.97457507e-01
3.86149734e-01 -9.35030520e-01 -4.69804645e-01 -1.27909288e-01] | [7.051922798156738, 3.8587937355041504] |
ec067307-174f-4f8d-835d-18e447aee2c5 | spa-gcn-efficient-and-flexible-gcn | 2111.05936 | null | https://arxiv.org/abs/2111.05936v1 | https://arxiv.org/pdf/2111.05936v1.pdf | SPA-GCN: Efficient and Flexible GCN Accelerator with an Application for Graph Similarity Computation | While there have been many studies on hardware acceleration for deep learning on images, there has been a rather limited focus on accelerating deep learning applications involving graphs. The unique characteristics of graphs, such as the irregular memory access and dynamic parallelism, impose several challenges when the algorithm is mapped to a CPU or GPU. To address these challenges while exploiting all the available sparsity, we propose a flexible architecture called SPA-GCN for accelerating Graph Convolutional Networks (GCN), the core computation unit in deep learning algorithms on graphs. The architecture is specialized for dealing with many small graphs since the graph size has a significant impact on design considerations. In this context, we use SimGNN, a neural-network-based graph matching algorithm, as a case study to demonstrate the effectiveness of our architecture. The experimental results demonstrate that SPA-GCN can deliver a high speedup compared to a multi-core CPU implementation and a GPU implementation, showing the efficiency of our design. | ['Jason Cong', 'Yuze Chi', 'Atefeh Sohrabizadeh'] | 2021-11-10 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [-1.88659981e-01 -9.53705013e-02 -5.07144518e-02 -2.16081023e-01
1.26183107e-01 -1.09574415e-01 3.57386976e-01 3.32256883e-01
-4.79488611e-01 -4.41424660e-02 -1.21390246e-01 -8.62379789e-01
-3.83199267e-02 -1.19382966e+00 -7.17488945e-01 -6.79522514e-01
-3.59036833e-01 1.41494378e-01 2.35272110e-01 -1.76534951e-01
-7.50923306e-02 8.01419318e-01 -1.86865282e+00 2.91309834e-01
3.85702699e-01 8.11852276e-01 1.11685842e-01 7.32408345e-01
-2.67842501e-01 7.66369343e-01 -4.05997187e-01 -5.75752914e-01
4.63856190e-01 -7.12819099e-02 -4.77382272e-01 -2.10284755e-01
8.42342377e-01 -3.46956342e-01 -6.42163932e-01 1.16324627e+00
5.84899247e-01 9.65175852e-02 1.28146350e-01 -1.40466440e+00
-1.97912097e-01 7.10082352e-01 -7.59523392e-01 4.12959814e-01
9.21182111e-02 -1.03663363e-01 1.02602220e+00 -4.33170468e-01
3.46162885e-01 1.23732769e+00 1.10277498e+00 4.11731362e-01
-7.77298450e-01 -7.33061612e-01 8.25488791e-02 2.60161370e-01
-1.14084530e+00 -7.99135193e-02 6.86297834e-01 -2.06313893e-01
1.33599079e+00 3.79211493e-02 9.98163700e-01 7.36779213e-01
6.64263427e-01 4.98000056e-01 5.50780356e-01 -4.60369468e-01
3.54147673e-01 -6.09206498e-01 6.05438232e-01 1.01341975e+00
8.04481804e-01 1.04162455e-01 -3.99171352e-01 -1.64742723e-01
7.30563819e-01 5.87689914e-02 7.22982511e-02 -3.89298052e-01
-6.42913878e-01 9.16107357e-01 9.27915573e-01 1.42505243e-01
-3.31668049e-01 6.94370985e-01 1.04397452e+00 5.36017656e-01
2.65066117e-01 -7.75428489e-02 -1.37388855e-01 1.77292600e-02
-6.62034690e-01 2.89070010e-01 1.00964570e+00 1.09411335e+00
7.59971142e-01 1.97205469e-01 1.75645635e-01 5.34033597e-01
2.26329699e-01 1.00880839e-01 4.23572570e-01 -2.63428956e-01
4.50393707e-01 8.45162332e-01 -9.98743832e-01 -1.38263667e+00
-7.55097687e-01 -4.57956940e-01 -1.27058876e+00 3.70205045e-01
2.96420485e-01 -1.20836794e-01 -8.95430267e-01 1.43910038e+00
4.44379658e-01 2.55822390e-01 8.36151615e-02 9.37478602e-01
1.34513557e+00 5.26906073e-01 3.55997920e-01 3.15820545e-01
1.58244669e+00 -1.24966538e+00 -4.19428378e-01 -3.84418398e-01
1.23395312e+00 -5.87044716e-01 1.15936446e+00 1.25682518e-01
-7.89407015e-01 -5.94625771e-01 -1.30256581e+00 -3.81715655e-01
-4.73462731e-01 -5.90981953e-02 1.37393200e+00 8.01130295e-01
-1.46426785e+00 7.45437920e-01 -1.18357527e+00 -4.66391951e-01
4.01494026e-01 6.62027538e-01 -3.31158429e-01 -1.54680803e-01
-8.57207000e-01 3.22949976e-01 6.44171417e-01 -1.22439312e-02
-3.92382979e-01 -7.23955154e-01 -7.80697823e-01 5.38876951e-01
-9.81151089e-02 -8.15465152e-01 9.30819690e-01 -1.01134539e+00
-1.05466855e+00 9.18656290e-01 3.84721160e-01 -6.88197076e-01
2.31086016e-01 1.18594624e-01 -3.01919132e-01 -3.67688090e-02
-3.96670580e-01 4.80582237e-01 6.30044878e-01 -1.72067568e-01
-5.45319021e-01 -5.60326755e-01 2.78905571e-01 1.01563282e-01
-5.50367415e-01 1.50549844e-01 -5.90822697e-01 -6.50038540e-01
3.53494170e-03 -1.16497815e+00 -4.55880493e-01 -1.79951146e-01
-8.62303656e-03 -2.83489168e-01 1.08978367e+00 -1.19624741e-01
1.08842862e+00 -2.23331809e+00 -1.77044436e-01 3.01198393e-01
6.19757593e-01 5.35841942e-01 -1.57881215e-01 3.44854265e-01
-1.67073771e-01 -4.39477444e-01 2.57026374e-01 -7.51511380e-02
-2.96011299e-01 4.53978390e-01 -1.30418167e-01 4.30110127e-01
-1.37490243e-01 9.63967800e-01 -7.47803688e-01 -2.46545836e-01
2.11036932e-02 6.74130142e-01 -6.95208788e-01 1.21364750e-01
5.87910265e-02 -3.39678615e-01 -5.31279802e-01 3.60595196e-01
8.89319241e-01 -6.19253576e-01 4.54842955e-01 -3.01024556e-01
5.74207157e-02 -2.85999607e-02 -1.08658266e+00 1.63292825e+00
-3.15718442e-01 6.62178814e-01 1.58560678e-01 -8.88108492e-01
8.36817324e-01 3.25650088e-02 2.02304080e-01 -8.44301939e-01
3.21954161e-01 1.52153179e-01 2.61541039e-01 -1.38491794e-01
6.67915583e-01 3.18952799e-01 2.90621638e-01 4.69905257e-01
4.39872257e-02 3.12507719e-01 3.81790310e-01 2.99480885e-01
1.45319355e+00 -2.17943743e-01 2.06314579e-01 -7.04720020e-01
3.72482389e-01 1.66027650e-01 2.88488865e-01 6.80462420e-01
5.15844338e-02 1.20003857e-01 6.19244993e-01 -1.16356027e+00
-1.06240666e+00 -4.83898014e-01 1.11331522e-01 1.45499003e+00
-1.02561556e-01 -1.06124604e+00 -9.75003242e-01 -4.09410954e-01
1.18379788e-02 1.95692554e-02 -3.68007034e-01 -2.37656981e-02
-6.84724033e-01 -1.00516164e+00 4.67398167e-01 7.14243412e-01
6.43587589e-01 -9.87348795e-01 -1.00399542e+00 2.50152230e-01
6.79157436e-01 -1.29803228e+00 -3.31396371e-01 1.52099118e-01
-1.18124127e+00 -1.18505132e+00 -1.77580819e-01 -1.02942717e+00
6.36830151e-01 6.58876359e-01 1.32742214e+00 6.01834238e-01
-4.53849167e-01 2.51405895e-01 -1.54413193e-01 -4.47702885e-01
-4.75172162e-01 4.12072271e-01 -1.56361654e-01 -2.13969544e-01
5.04899681e-01 -8.44506741e-01 -6.07287109e-01 -1.53781831e-01
-8.99214745e-01 4.93140697e-01 4.89052415e-01 7.39005804e-01
4.43719506e-01 1.79472283e-01 7.47011229e-02 -1.28808606e+00
7.83194244e-01 -1.45306647e-01 -1.10632718e+00 -5.82643747e-02
-7.76429951e-01 1.28612146e-01 8.93499315e-01 -3.63573223e-01
-4.24026579e-01 2.49327257e-01 -3.08316618e-01 -4.97840434e-01
1.85345292e-01 7.71214008e-01 5.02797067e-02 -6.87076747e-01
7.32117951e-01 -2.30971232e-01 2.63664395e-01 -1.63079679e-01
1.17057204e-01 2.37522721e-01 4.40353990e-01 -5.83686233e-01
3.14641654e-01 2.61642486e-01 7.32522547e-01 -8.09599638e-01
-6.37826264e-01 -3.98268908e-01 -3.71699244e-01 -2.02596471e-01
6.85349643e-01 -1.09778178e+00 -9.66820955e-01 6.09968662e-01
-1.06668007e+00 -4.85467106e-01 1.56787351e-01 4.77115124e-01
-2.90213674e-01 3.57499599e-01 -9.20108497e-01 -1.91689298e-01
-9.27371204e-01 -1.40461290e+00 8.84695113e-01 2.71083027e-01
1.65421620e-01 -1.28179085e+00 -1.42180160e-01 -1.82253588e-02
6.42811835e-01 4.55026418e-01 1.04727280e+00 -6.02049291e-01
-7.50182152e-01 -1.81561232e-01 -5.95894516e-01 -3.98370586e-02
-3.14333349e-01 -2.63586640e-02 -9.38355267e-01 -8.53793085e-01
-2.74245739e-01 -2.63774037e-01 7.24842727e-01 2.59895086e-01
1.44249403e+00 -1.45071000e-01 -2.47587278e-01 1.19743764e+00
1.77593815e+00 -1.58142462e-01 5.32176733e-01 5.33847451e-01
1.26465285e+00 9.58073512e-02 1.45854309e-01 2.48070896e-01
1.43308833e-01 5.46941698e-01 7.27311790e-01 -3.39578241e-01
-3.52745295e-01 -4.23531048e-02 1.16374500e-01 1.26589692e+00
-2.33359233e-01 -2.51657397e-01 -1.23589456e+00 7.99305364e-02
-1.96107042e+00 -4.38962847e-01 -6.25612676e-01 1.84975755e+00
1.91613674e-01 5.72642684e-02 -3.86198498e-02 1.70072541e-01
6.17610395e-01 4.04726475e-01 -2.32496336e-01 -9.11636829e-01
1.93199024e-01 5.73668897e-01 8.42525482e-01 3.37972380e-02
-1.07167280e+00 1.00254941e+00 6.57446432e+00 8.76672029e-01
-1.39655626e+00 8.11371114e-03 4.47330862e-01 1.75243482e-01
1.19492956e-01 -1.27585754e-01 -8.79301071e-01 1.55509233e-01
1.20180845e+00 -2.32748911e-01 3.67295593e-01 1.12386024e+00
-2.07203031e-01 3.12998325e-01 -1.15085161e+00 1.28681433e+00
-1.21168971e-01 -1.51289499e+00 3.34724993e-01 2.92830825e-01
4.40860659e-01 4.49934959e-01 -1.55394509e-01 2.58983999e-01
2.71096945e-01 -1.04569626e+00 3.83095652e-01 -3.44458818e-01
7.20587790e-01 -1.16523314e+00 7.92322338e-01 1.02159739e-01
-1.43454075e+00 1.06225431e-01 -7.72259951e-01 -4.63047773e-01
-3.28083783e-01 6.72801316e-01 -8.28336895e-01 5.61378002e-01
8.19273651e-01 7.01762021e-01 -7.48598874e-01 8.76239538e-01
6.13715984e-02 6.25622213e-01 -1.87955573e-01 -2.04756901e-01
5.40670574e-01 -3.79916996e-01 6.64161667e-02 1.48755276e+00
3.75905275e-01 9.98822525e-02 1.58052281e-01 4.30722564e-01
-2.46905684e-01 4.45093632e-01 -8.44205678e-01 3.46126705e-02
1.40900416e-02 1.55496109e+00 -1.09890163e+00 -2.33636454e-01
-9.25477922e-01 6.72306776e-01 7.20274627e-01 -3.02544478e-02
-8.94783318e-01 -5.51712632e-01 7.27071524e-01 5.19586988e-02
2.01575056e-01 -4.31686133e-01 -2.86350012e-01 -9.70619678e-01
-1.32093191e-01 -9.75865901e-01 8.48668098e-01 -4.94854808e-01
-9.36558723e-01 7.77379096e-01 -3.24990541e-01 -7.33889878e-01
1.03853740e-01 -1.03763640e+00 -7.57626057e-01 6.58226728e-01
-1.28843439e+00 -1.13016510e+00 -8.49857330e-01 7.11361229e-01
6.59882650e-02 -5.09197176e-01 7.55654335e-01 5.71407497e-01
-5.35723269e-01 8.79372895e-01 -1.53560415e-01 2.58437157e-01
1.61835805e-01 -9.71051991e-01 1.11426163e+00 1.06338322e+00
2.44666263e-01 4.92852449e-01 4.44501072e-01 -4.67020601e-01
-2.18044925e+00 -1.38088739e+00 5.54595709e-01 2.24913254e-01
7.34728158e-01 -5.71074784e-01 -1.18237388e+00 7.43344963e-01
8.99487138e-02 4.59824085e-01 6.54161870e-01 3.58297259e-01
-4.57490087e-01 -2.78069019e-01 -8.45979869e-01 7.48629987e-01
1.32492435e+00 -4.48641866e-01 -4.19792496e-02 5.70921183e-01
7.82373250e-01 -1.00233316e+00 -8.09967279e-01 2.69079894e-01
4.01298434e-01 -1.05684960e+00 8.90968800e-01 -6.10446155e-01
2.80362636e-01 -9.29373726e-02 3.77476402e-02 -1.02454317e+00
-6.08563662e-01 -4.28931683e-01 -1.80549219e-01 6.95288658e-01
-1.45837605e-01 -7.85411060e-01 1.31620181e+00 4.47542280e-01
-2.91256338e-01 -8.66019249e-01 -6.49512172e-01 -7.81507730e-01
-2.59042829e-01 -3.39479297e-01 1.04139650e+00 9.39255893e-01
-4.12985772e-01 4.09072906e-01 -1.38985654e-02 2.42247447e-01
5.81581235e-01 2.53030032e-01 1.11458266e+00 -1.28011394e+00
-3.95689398e-01 -4.85683560e-01 -1.07743168e+00 -7.49721348e-01
3.11370969e-01 -1.28380823e+00 -4.66391087e-01 -1.32178414e+00
2.43310973e-01 -4.35357809e-01 -1.71051756e-01 6.32315099e-01
-2.01296974e-02 2.10032701e-01 1.94815263e-01 -3.48124243e-02
-3.94681454e-01 1.38186336e-01 1.11350191e+00 -7.88269565e-02
-1.65278446e-02 -1.71253875e-01 -4.79238957e-01 5.44141352e-01
6.46496773e-01 -4.14147139e-01 -6.93514824e-01 -8.34637344e-01
4.56461817e-01 -1.54185608e-01 2.44434834e-01 -1.41157401e+00
6.79356873e-01 2.03427240e-01 3.12827416e-02 -4.03443187e-01
-1.47535235e-01 -8.14729273e-01 4.75410193e-01 8.41593564e-01
1.42476335e-02 9.01741028e-01 5.00955403e-01 4.76116568e-01
-2.34503090e-01 -2.22657993e-02 8.01628351e-01 8.44162628e-02
-9.14518356e-01 8.85531723e-01 -3.65207754e-02 -6.62740245e-02
8.51580918e-01 -3.03413540e-01 -4.40976709e-01 7.26975426e-02
-3.12542140e-01 2.57533435e-02 4.73519564e-01 3.54341358e-01
5.64985573e-01 -1.39458907e+00 -5.62996745e-01 5.16640961e-01
1.45167753e-01 1.82987988e-01 1.62788883e-01 5.40174782e-01
-1.35710430e+00 4.16332960e-01 -6.93352640e-01 -7.23132670e-01
-1.76523101e+00 8.69528532e-01 2.71179587e-01 -4.31965113e-01
-1.34636676e+00 7.35707819e-01 3.95092398e-01 -8.13422725e-02
3.43111366e-01 -4.19927508e-01 -1.48675129e-01 -2.05091193e-01
4.90825057e-01 2.82265574e-01 7.08801746e-01 -2.77038097e-01
-1.82572082e-01 4.06772286e-01 -2.38086894e-01 7.23354161e-01
1.29738665e+00 4.49984431e-01 -5.15076578e-01 -2.43365273e-01
1.26510155e+00 -3.23172420e-01 -8.55042934e-01 -1.64475322e-01
4.35394756e-02 -3.47275734e-01 3.67380738e-01 2.03408673e-01
-1.76834428e+00 8.68103266e-01 6.82206810e-01 1.85250178e-01
1.25971031e+00 -3.68251234e-01 9.86706972e-01 6.62877679e-01
5.58666229e-01 -7.88930237e-01 -7.53419921e-02 8.66761744e-01
4.61869687e-01 -8.37719381e-01 2.22951174e-01 -5.90174735e-01
6.51486740e-02 1.45015717e+00 7.91917264e-01 -4.13052887e-01
7.45851755e-01 5.64624250e-01 -9.60604921e-02 -5.83396375e-01
-7.36809433e-01 9.28336158e-02 1.54823661e-01 4.78191286e-01
5.75388610e-01 1.54754803e-01 -3.20566863e-01 9.55377985e-03
-4.87142354e-01 -1.41428992e-01 5.08945465e-01 9.21816111e-01
-1.16260685e-01 -1.05186653e+00 3.30225639e-02 6.26349986e-01
-6.12162828e-01 -1.87838838e-01 3.02156955e-02 9.09915149e-01
-2.01247409e-01 4.59822029e-01 2.76482582e-01 -4.94983673e-01
2.30871633e-01 -4.26963836e-01 6.23577833e-01 -5.71220398e-01
-1.09549725e+00 -4.31694955e-01 9.81471911e-02 -9.07540977e-01
-1.78140834e-01 -1.41334102e-01 -1.35040474e+00 -9.88595486e-01
-2.91697178e-02 -3.21973972e-02 8.74259710e-01 4.17789102e-01
7.03849018e-01 8.46622944e-01 1.22670420e-01 -7.36284792e-01
-2.73543388e-01 -3.64538431e-01 -7.33901680e-01 4.56272840e-01
6.16939664e-02 -3.79462808e-01 -1.46119937e-01 -4.90137219e-01] | [7.013814926147461, 5.6533427238464355] |
ce40110f-2ca4-4755-94ab-1d6e52ce98b2 | reward-gaming-in-conditional-text-generation | 2211.08714 | null | https://arxiv.org/abs/2211.08714v3 | https://arxiv.org/pdf/2211.08714v3.pdf | Reward Gaming in Conditional Text Generation | To align conditional text generation model outputs with desired behaviors, there has been an increasing focus on training the model using reinforcement learning (RL) with reward functions learned from human annotations. Under this framework, we identify three common cases where high rewards are incorrectly assigned to undesirable patterns: noise-induced spurious correlation, naturally occurring spurious correlation, and covariate shift. We show that even though learned metrics achieve high performance on the distribution of the data used to train the reward function, the undesirable patterns may be amplified during RL training of the text generation model. While there has been discussion about reward gaming in the RL or safety community, in this discussion piece, we would like to highlight reward gaming in the natural language generation (NLG) community using concrete conditional text generation examples and discuss potential fixes and areas for future work. | ['He He', 'Ankur P. Parikh', 'Thibault Sellam', 'Vishakh Padmakumar', 'Richard Yuanzhe Pang'] | 2022-11-16 | null | null | null | null | ['conditional-text-generation'] | ['natural-language-processing'] | [ 4.86785710e-01 9.14366722e-01 -1.49815485e-01 -4.24465388e-01
-1.08005285e+00 -5.76953530e-01 7.75518298e-01 1.98887922e-02
-1.76639602e-01 1.06623530e+00 6.82774246e-01 -3.25995475e-01
7.50157088e-02 -6.34437203e-01 -7.39157021e-01 -3.17803830e-01
-8.03020690e-03 3.96485776e-01 -3.97819668e-01 -1.53536901e-01
2.97069609e-01 -9.92300436e-02 -1.37881732e+00 2.34069631e-01
1.04193223e+00 3.72116208e-01 -1.71776831e-01 8.84665251e-01
1.02351792e-01 1.32917929e+00 -1.07749867e+00 -5.94969094e-01
2.25730151e-01 -8.83633316e-01 -7.85532176e-01 -1.68355316e-01
1.97262913e-01 -3.46188039e-01 9.72872376e-02 8.75091732e-01
5.67816913e-01 1.86368227e-01 7.90071428e-01 -1.62288022e+00
-8.24632168e-01 1.00591648e+00 -3.26555192e-01 -5.61280288e-02
4.49198335e-01 7.20152855e-01 1.39014912e+00 -4.57914680e-01
6.88372135e-01 1.32663059e+00 6.37514472e-01 8.96227956e-01
-1.49399090e+00 -9.42874014e-01 1.54853255e-01 -4.62487817e-01
-8.37334931e-01 -3.12512696e-01 3.69091302e-01 -6.50104105e-01
1.16814160e+00 1.54789343e-01 5.40926754e-01 1.61643219e+00
8.51447731e-02 8.77947688e-01 9.69338655e-01 -3.56907785e-01
2.34841332e-01 6.38367832e-02 -6.67728037e-02 4.30990666e-01
2.88763285e-01 7.80225635e-01 -6.49058700e-01 -4.46837068e-01
6.66247964e-01 -5.51356673e-01 1.19865723e-01 -1.79389164e-01
-1.11486077e+00 1.07827067e+00 1.57344326e-01 4.57192361e-02
-2.60355622e-01 8.17772686e-01 2.73865879e-01 1.85294151e-01
5.99486768e-01 1.01382923e+00 -4.22839582e-01 -6.47618294e-01
-7.50967383e-01 9.05825198e-01 6.39036238e-01 1.10246694e+00
6.94822609e-01 2.93178707e-01 -6.88188910e-01 7.41626084e-01
2.10943148e-01 5.60288668e-01 4.16116536e-01 -1.13623440e+00
6.38655663e-01 3.96462053e-01 4.42140162e-01 -6.61595047e-01
-3.66615385e-01 -2.29403391e-01 -4.61634338e-01 2.18810678e-01
6.82993948e-01 -8.18316102e-01 -6.01620376e-01 2.07993340e+00
-2.27091521e-01 -7.72123635e-02 2.88386941e-01 6.47126257e-01
3.53907943e-01 3.21639687e-01 4.76536483e-01 -1.26948446e-01
6.74488842e-01 -5.96695244e-01 -7.03281343e-01 -3.17644238e-01
8.82024467e-01 -6.47773027e-01 1.22695780e+00 2.32001141e-01
-1.15508449e+00 -4.17503208e-01 -8.20852637e-01 1.03385173e-01
2.30200253e-02 3.07336636e-02 7.99215496e-01 7.69953072e-01
-1.00401711e+00 8.48567963e-01 -6.78234994e-01 -2.87427545e-01
3.72315258e-01 2.66223639e-01 3.10472131e-01 2.79786348e-01
-1.39110911e+00 9.25326288e-01 -7.93629326e-03 -2.09964633e-01
-9.82321918e-01 -6.92969084e-01 -6.86514854e-01 -3.63888848e-03
2.41703719e-01 -5.93108416e-01 1.75724483e+00 -1.36153984e+00
-1.43536079e+00 5.98035634e-01 1.31584138e-01 -5.22346139e-01
7.62613297e-01 -3.38776290e-01 -2.10335299e-01 -5.69599807e-01
2.94541508e-01 8.63630712e-01 6.19636059e-01 -1.10794222e+00
-6.72791541e-01 4.52789366e-02 1.77258309e-02 2.79700965e-01
2.13869944e-01 4.47896682e-02 3.94395322e-01 -7.86940515e-01
-6.29226804e-01 -9.86761928e-01 -4.34301466e-01 -5.60645163e-01
-4.31475788e-01 -5.07098496e-01 2.73638368e-01 -4.02996600e-01
1.20996284e+00 -1.88383806e+00 -2.39816859e-01 1.16673954e-01
-1.29587397e-01 -1.83045402e-01 -2.26798922e-01 5.25393426e-01
-2.29232162e-01 5.09788394e-01 -1.80455595e-01 -2.69582629e-01
1.63844213e-01 2.79562958e-02 -5.94424605e-01 6.77451268e-02
4.91688073e-01 1.08524966e+00 -1.33260167e+00 -1.26486301e-01
-1.81979612e-01 4.27863039e-02 -7.68085718e-01 4.61231828e-01
-5.91312826e-01 3.33869725e-01 -2.87242979e-01 4.14596409e-01
1.81323588e-01 -2.08598390e-01 1.77267626e-01 4.41139549e-01
-2.00066771e-02 6.31741405e-01 -8.53872120e-01 1.27747989e+00
-3.63974273e-01 5.42997003e-01 -4.11127627e-01 -4.53979760e-01
8.84632885e-01 2.89037675e-01 4.18841213e-01 -4.53696311e-01
-8.82963687e-02 1.58380434e-01 2.36268312e-01 -4.61387306e-01
8.58362973e-01 -4.45720375e-01 -2.76355773e-01 9.90198791e-01
-2.26351991e-03 -3.44565779e-01 5.07935844e-02 1.50620267e-01
1.18042064e+00 5.42234659e-01 1.46600336e-01 4.00849916e-02
-2.21834660e-01 3.09501827e-01 4.76477593e-01 1.03599358e+00
-1.15932748e-01 5.33921778e-01 8.95462573e-01 -1.03304602e-01
-1.21963263e+00 -1.11949873e+00 1.43346712e-01 1.14922440e+00
-4.18020517e-01 -5.76191008e-01 -6.35831833e-01 -1.01008046e+00
-4.54897098e-02 1.22748959e+00 -7.40204811e-01 -3.84448677e-01
-4.82836634e-01 -9.21377003e-01 8.84743929e-01 6.53615594e-01
-1.00295851e-02 -1.48781323e+00 -6.93498492e-01 2.58093625e-01
-3.50982845e-01 -6.71836078e-01 -5.09705782e-01 3.90919000e-01
-7.52692878e-01 -1.02646422e+00 -6.52695894e-01 -3.01779151e-01
6.29133523e-01 -1.73327997e-01 1.35739720e+00 4.30495478e-02
-1.82662718e-02 5.62294126e-01 -5.35772681e-01 -7.07855880e-01
-1.00531292e+00 1.29503444e-01 -9.17188525e-02 -5.43875217e-01
3.56665760e-01 -2.09647894e-01 -4.10198927e-01 7.06431866e-02
-5.97792804e-01 1.57750189e-01 3.18520814e-01 9.60678458e-01
2.62910817e-02 -5.18702209e-01 9.93823469e-01 -1.11468017e+00
1.26441371e+00 -6.08640969e-01 -5.81469595e-01 -3.90356630e-02
-9.13205862e-01 3.14996362e-01 5.28574646e-01 -4.18122470e-01
-1.04639137e+00 -6.44784793e-02 6.07279837e-02 -6.63549751e-02
-1.24714039e-01 3.64008397e-01 2.10529119e-01 6.70372188e-01
1.19530487e+00 -2.30112642e-01 6.20857142e-02 9.68156159e-02
4.33968186e-01 5.18296063e-01 1.09798945e-01 -7.67050743e-01
6.55592918e-01 -1.24562711e-01 -3.20012331e-01 -2.16794401e-01
-8.49111378e-01 1.65050715e-01 -6.44889697e-02 -1.87424332e-01
7.54175305e-01 -8.08952451e-01 -5.06910801e-01 5.29802293e-02
-1.06176734e+00 -1.10162294e+00 -8.00864577e-01 2.96556801e-01
-1.05873919e+00 -3.61147933e-02 -2.88557172e-01 -1.26992929e+00
-2.32168600e-01 -1.09916937e+00 1.05410707e+00 1.39274999e-01
-9.43081319e-01 -8.96454394e-01 4.04619008e-01 1.98054910e-01
3.49602550e-01 2.51156896e-01 1.01351893e+00 -6.77605033e-01
-4.93861675e-01 2.93734651e-02 9.18064788e-02 1.70225218e-01
3.14424574e-01 2.54711092e-01 -1.05585861e+00 4.43694852e-02
-4.85354125e-01 -7.26759732e-01 5.16052604e-01 4.60323811e-01
8.74758005e-01 -5.14736712e-01 -1.14643939e-01 8.52398053e-02
1.01637483e+00 1.85821727e-01 4.22883868e-01 1.11739144e-01
4.14734542e-01 8.15140963e-01 7.72906184e-01 5.37195742e-01
3.39048326e-01 4.84182149e-01 2.67886817e-01 -1.14664979e-01
8.38320032e-02 -7.87212789e-01 7.19652891e-01 -6.52379915e-02
-7.11336173e-03 -3.86758894e-01 -7.38584936e-01 4.09280986e-01
-2.08910346e+00 -1.20956767e+00 -8.27348307e-02 2.26579690e+00
9.47284043e-01 2.63537109e-01 5.50584555e-01 -2.42303580e-01
5.67906141e-01 7.95114115e-02 -5.70896745e-01 -6.81468606e-01
-4.47802357e-02 1.33275941e-01 4.47135776e-01 6.55037344e-01
-6.15086555e-01 1.12160861e+00 7.95846987e+00 3.56889516e-01
-8.17207932e-01 -1.90513700e-01 1.01033795e+00 -1.45328984e-01
-6.18020654e-01 2.39172176e-01 -5.40467978e-01 2.27237895e-01
9.93964195e-01 -4.17857081e-01 6.82242870e-01 8.06553960e-01
5.39855182e-01 -1.56984359e-01 -1.42831266e+00 6.06715381e-01
-2.08052397e-01 -1.01290095e+00 -2.08711818e-01 1.32804275e-01
9.70826685e-01 1.50661003e-02 5.34293652e-02 7.50065327e-01
1.38003421e+00 -1.36372685e+00 6.84725642e-01 4.14680541e-01
7.46913970e-01 -8.80805671e-01 4.91960645e-01 3.33924919e-01
-5.04589260e-01 -1.26164958e-01 -2.61658639e-01 -3.67192447e-01
-8.48972350e-02 4.50162172e-01 -1.35063660e+00 1.66863054e-01
2.69518048e-01 6.60969436e-01 -4.94501770e-01 5.19157946e-01
-4.43597436e-01 8.04814041e-01 1.08173393e-01 -3.71638209e-01
1.73032492e-01 1.77497175e-02 2.93782949e-01 1.09737754e+00
3.64180654e-01 -2.22134262e-01 -9.39273462e-02 1.27561700e+00
2.14807075e-02 2.22780630e-02 -9.50322986e-01 -3.11826468e-01
8.15301761e-02 1.07209194e+00 -2.91881442e-01 -2.03036472e-01
-2.93455511e-01 8.09124708e-01 4.21330631e-01 4.68963057e-01
-9.49934900e-01 -2.28797272e-02 7.18239725e-01 1.11246750e-01
-1.95888236e-01 -1.09200887e-02 -6.56712532e-01 -9.14209664e-01
-4.20996159e-01 -1.08170009e+00 3.21243435e-01 -9.58440006e-01
-1.34784293e+00 2.18691543e-01 1.58108443e-01 -1.08732986e+00
-1.01180637e+00 -2.50389040e-01 -6.95942640e-01 9.29379344e-01
-9.55858290e-01 -6.47287548e-01 8.47019628e-03 1.44958377e-01
4.85432446e-01 -1.98794261e-01 7.62395918e-01 -3.10301095e-01
-3.43997896e-01 7.36529112e-01 -1.88645631e-01 5.37802875e-02
9.00438726e-01 -1.52846301e+00 6.89994931e-01 5.36308408e-01
1.23987541e-01 4.84587878e-01 1.10175371e+00 -8.94338608e-01
-7.71085382e-01 -1.09272981e+00 9.96274590e-01 -7.67802775e-01
5.46955407e-01 -3.79483074e-01 -4.28725630e-01 9.46476936e-01
4.85925585e-01 -5.41456282e-01 8.81220162e-01 2.71772504e-01
-6.78658709e-02 2.11665154e-01 -1.07698464e+00 1.01087666e+00
1.03575766e+00 -1.80014580e-01 -2.11380422e-01 4.58767742e-01
6.32625043e-01 -4.50717390e-01 -5.00144005e-01 5.88846616e-02
4.17943299e-01 -8.86916816e-01 5.17745435e-01 -8.10708940e-01
8.31186891e-01 -2.07613725e-02 9.73771140e-02 -1.71805239e+00
-1.16744094e-01 -1.02264798e+00 1.81298196e-01 1.21961999e+00
9.27194238e-01 -4.08306330e-01 9.00951147e-01 1.20483375e+00
4.54393914e-03 -6.35231137e-01 -5.51365137e-01 -5.86570323e-01
4.39277500e-01 -7.17811525e-01 6.32258117e-01 8.33409607e-01
5.30096829e-01 5.50185859e-01 -6.32110536e-01 -3.32820565e-01
3.42687160e-01 -1.57747567e-01 1.14711308e+00 -8.98009121e-01
-5.64728260e-01 -5.29750109e-01 2.85562813e-01 -9.43577468e-01
2.04000786e-01 -1.02258623e+00 3.92139912e-01 -1.35216570e+00
2.17883602e-01 -4.84808892e-01 1.12585798e-02 5.24198830e-01
-4.23683882e-01 -1.84840038e-01 3.06947410e-01 -8.05462822e-02
-4.37138975e-01 6.02964818e-01 1.34121084e+00 6.03007674e-02
-3.69917512e-01 2.77862042e-01 -1.19380283e+00 5.04787683e-01
1.04528975e+00 -7.31975913e-01 -5.67210972e-01 -2.07570642e-01
6.48264229e-01 1.98947057e-01 2.90742546e-01 -7.55344987e-01
-2.71118969e-01 -6.63783908e-01 4.24356192e-01 -1.82645857e-01
-5.95768401e-03 -5.13258457e-01 6.18563034e-03 4.12047178e-01
-1.08735359e+00 3.30068260e-01 2.04212770e-01 4.78399068e-01
2.25982472e-01 -3.79510701e-01 5.38404405e-01 -2.54705757e-01
1.37207165e-01 -1.88141651e-02 -7.96454310e-01 4.03341830e-01
8.75646889e-01 -1.60184633e-02 -3.64044517e-01 -9.38435733e-01
-5.89647174e-01 4.17033106e-01 3.96425813e-01 6.07486963e-01
3.75748724e-01 -1.30482984e+00 -7.53167212e-01 -2.96663269e-02
1.15675196e-01 6.91777887e-03 -1.94434866e-01 4.13393348e-01
8.71602893e-02 2.47864872e-01 1.42788157e-01 -1.84171572e-01
-8.77448916e-01 2.30057165e-01 4.84216899e-01 -6.94566250e-01
-1.59128755e-01 5.82199872e-01 1.35023296e-01 -4.87415105e-01
1.49029836e-01 -4.88530725e-01 7.18612522e-02 -1.74476534e-01
2.24360511e-01 4.01207298e-01 -2.47824833e-01 -2.02092320e-01
-1.45690646e-02 -7.83164725e-02 1.32209927e-01 -6.03034914e-01
1.09204173e+00 1.25438303e-01 5.11008203e-01 5.62301993e-01
6.25003755e-01 5.29742204e-02 -1.64178228e+00 2.40401849e-01
2.10504636e-01 -1.89013213e-01 -4.68721956e-01 -1.14835882e+00
-6.13135755e-01 7.10629344e-01 4.63030398e-01 3.00733417e-01
6.79896653e-01 -9.28797498e-02 3.09101373e-01 2.23371446e-01
1.88783765e-01 -1.45199919e+00 5.47222674e-01 6.11870110e-01
1.09916878e+00 -1.19749904e+00 -4.51231189e-02 5.43393195e-02
-1.25282335e+00 9.30248260e-01 1.01782060e+00 -3.01654875e-01
2.88929194e-01 2.75894552e-01 1.44089997e-01 5.80996946e-02
-1.15936768e+00 -8.41707438e-02 -1.74293056e-01 1.04221785e+00
9.58871126e-01 1.37963891e-01 -3.32734972e-01 4.79870558e-01
-7.23123908e-01 1.03993654e-01 8.29132855e-01 6.06853485e-01
-1.47953272e-01 -1.42378998e+00 -2.93001831e-01 7.66309023e-01
-5.86938858e-01 -2.82166570e-01 -7.51228094e-01 7.79799283e-01
-3.28061031e-03 1.20446134e+00 9.63130072e-02 -3.76729727e-01
3.07581812e-01 2.13191405e-01 3.82223248e-01 -8.67564380e-01
-1.25334251e+00 8.45633820e-02 4.74300742e-01 -5.06314158e-01
-1.45757973e-01 -9.08403933e-01 -1.31910658e+00 -3.23786139e-02
-2.96523362e-01 4.01055701e-02 2.11486906e-01 7.51129925e-01
1.49050698e-01 5.60351610e-01 5.23559630e-01 -4.57029879e-01
-7.42702007e-01 -1.25474727e+00 -4.74636048e-01 4.88523334e-01
3.24777365e-01 -4.03734922e-01 -3.67154837e-01 -3.70536447e-02] | [11.710933685302734, 8.932479858398438] |
e6179fd1-0447-4730-bb41-60b25b82dcc1 | combining-state-of-the-art-models-with | 2211.10808 | null | https://arxiv.org/abs/2211.10808v1 | https://arxiv.org/pdf/2211.10808v1.pdf | Combining State-of-the-Art Models with Maximal Marginal Relevance for Few-Shot and Zero-Shot Multi-Document Summarization | In Natural Language Processing, multi-document summarization (MDS) poses many challenges to researchers above those posed by single-document summarization (SDS). These challenges include the increased search space and greater potential for the inclusion of redundant information. While advancements in deep learning approaches have led to the development of several advanced language models capable of summarization, the variety of training data specific to the problem of MDS remains relatively limited. Therefore, MDS approaches which require little to no pretraining, known as few-shot or zero-shot applications, respectively, could be beneficial additions to the current set of tools available in summarization. To explore one possible approach, we devise a strategy for combining state-of-the-art models' outputs using maximal marginal relevance (MMR) with a focus on query relevance rather than document diversity. Our MMR-based approach shows improvement over some aspects of the current state-of-the-art results in both few-shot and zero-shot MDS applications while maintaining a state-of-the-art standard of output by all available metrics. | ['Yllias Chali', 'Gandharv Suri', 'David Adams'] | 2022-11-19 | null | null | null | null | ['document-summarization'] | ['natural-language-processing'] | [ 4.58800465e-01 9.74802375e-02 -2.88985908e-01 -1.77178264e-01
-1.31040549e+00 -2.93610901e-01 9.24314737e-01 7.37474203e-01
-4.19137239e-01 7.74950981e-01 9.20083523e-01 -5.21338545e-02
-2.59087205e-01 -6.01009250e-01 -3.38590115e-01 -3.63624990e-01
7.02096671e-02 5.65650225e-01 2.87633598e-01 -6.12105906e-01
8.01568270e-01 2.08914191e-01 -1.75361574e+00 5.85172951e-01
1.19657636e+00 4.00444031e-01 2.82940596e-01 8.90105426e-01
-5.38845778e-01 7.50931680e-01 -1.03639019e+00 -2.22075820e-01
-1.01897664e-01 -6.50223970e-01 -9.81310129e-01 -1.60950229e-01
7.73716748e-01 -3.48512203e-01 -2.98380405e-01 7.54758358e-01
1.09165716e+00 5.23463190e-01 6.04199052e-01 -6.81417227e-01
-9.23727155e-01 7.56822467e-01 -5.68468630e-01 6.80609763e-01
7.45093822e-01 -6.96936920e-02 1.16187024e+00 -7.06655741e-01
7.43990183e-01 1.24435675e+00 5.18715024e-01 5.38767219e-01
-1.22977328e+00 -1.20168202e-01 -5.70485964e-02 1.00405753e-01
-9.66975689e-01 -7.33269453e-01 5.92775643e-01 -7.01344982e-02
1.73174357e+00 4.59380955e-01 3.47739398e-01 1.00127828e+00
3.40058655e-01 1.06728470e+00 4.05415922e-01 -5.86728871e-01
3.90939921e-01 -9.22160782e-03 2.24463716e-01 3.29886556e-01
5.76901257e-01 -5.23391426e-01 -8.32845509e-01 -2.04469755e-01
2.10146233e-01 -2.72301912e-01 -3.84677313e-02 -1.61987945e-01
-1.14385068e+00 9.24001396e-01 1.36418581e-01 5.71730614e-01
-4.17916387e-01 -1.13933250e-01 8.53699148e-01 3.59675318e-01
7.81033993e-01 1.02728367e+00 -9.69230384e-02 -3.56993228e-01
-1.77372766e+00 6.77741051e-01 9.00519609e-01 6.84337974e-01
3.93996626e-01 2.34166473e-01 -6.36986911e-01 8.95598114e-01
-1.28267750e-01 -9.52742472e-02 8.86019766e-01 -8.55706930e-01
8.33818316e-01 7.50527561e-01 -4.64558974e-02 -8.47430050e-01
-4.35901910e-01 -5.32152593e-01 -8.56329560e-01 -2.99863130e-01
-2.02225700e-01 -3.24171185e-02 -8.16874743e-01 1.37878025e+00
-1.25136644e-01 -2.28506118e-01 4.59812433e-01 3.85643274e-01
1.17033303e+00 8.71196151e-01 -1.60788924e-01 -4.09900516e-01
9.31496680e-01 -1.01778305e+00 -6.81508422e-01 -4.39655393e-01
7.45445251e-01 -6.32135987e-01 1.05754626e+00 3.54591340e-01
-1.44770706e+00 -5.17602980e-01 -1.40259218e+00 -3.68812650e-01
-3.43296081e-01 -1.00888841e-01 5.82113862e-01 5.81605732e-01
-1.40615427e+00 9.24075782e-01 -7.66236782e-01 -8.29164147e-01
3.68879408e-01 2.97660679e-01 -2.59322822e-01 -4.52679954e-02
-1.09503949e+00 1.28728056e+00 5.85527182e-01 -5.24494112e-01
-4.90962297e-01 -6.67172849e-01 -8.13835442e-01 4.51081663e-01
2.97528863e-01 -1.03342152e+00 1.57787633e+00 -4.95856971e-01
-1.40034044e+00 7.04804361e-01 -1.88520864e-01 -7.46774733e-01
4.28988069e-01 -5.18648386e-01 -2.22928137e-01 3.87773544e-01
3.49585235e-01 7.59174526e-01 3.64908457e-01 -1.00765026e+00
-5.83421052e-01 -1.63422570e-01 7.66664669e-02 6.15245461e-01
-5.58958173e-01 2.11240113e-01 -8.34923312e-02 -6.04892373e-01
-9.50503126e-02 -3.64649475e-01 -2.52680272e-01 -5.79325557e-01
-5.04467189e-01 -3.30546141e-01 7.32523978e-01 -7.11731255e-01
1.73167777e+00 -1.68918157e+00 3.58024955e-01 -6.31018162e-01
-1.62034556e-01 6.50145173e-01 -3.26337516e-01 1.13837075e+00
9.76945236e-02 3.16326499e-01 -2.69487411e-01 -8.17540228e-01
-7.07700625e-02 -4.35980372e-02 -3.38675767e-01 1.02446072e-01
3.71633083e-01 8.18710029e-01 -1.02431130e+00 -6.21183932e-01
7.04096481e-02 8.47676694e-02 -4.90388066e-01 8.66377205e-02
-3.31259668e-01 4.90615480e-02 -2.52746642e-01 5.36176980e-01
2.91369826e-01 -4.53752354e-02 -9.11142752e-02 2.14736342e-01
-2.89103031e-01 6.81063175e-01 -1.05671275e+00 2.15861297e+00
-2.48854756e-01 7.32669890e-01 -4.36658263e-01 -1.05811965e+00
7.77773678e-01 3.18846315e-01 4.00279313e-01 -6.58037901e-01
-5.09808101e-02 1.53133482e-01 2.94412253e-04 -4.93072987e-01
1.53201985e+00 -5.41591411e-03 -1.24326080e-01 7.02282548e-01
4.04859364e-01 -1.18393041e-01 7.97994018e-01 6.78248167e-01
1.20307577e+00 -3.31893712e-02 6.98869050e-01 -3.27979803e-01
2.58739740e-01 1.42088354e-01 5.24626300e-02 1.07010484e+00
7.17493221e-02 8.52504313e-01 4.45374519e-01 1.10641145e-03
-1.09571683e+00 -8.14433157e-01 2.66972929e-01 1.33027673e+00
-8.86096582e-02 -6.95650160e-01 -7.03680396e-01 -2.16072395e-01
-1.77173018e-01 1.28134692e+00 -3.07715505e-01 -4.00559694e-01
-6.25436902e-01 -8.17400277e-01 9.16662872e-01 5.69873869e-01
3.42447311e-01 -1.14356756e+00 -1.04880023e+00 4.98843461e-01
-1.96333438e-01 -5.44672251e-01 -2.36544937e-01 2.80338347e-01
-1.28557944e+00 -4.90097702e-01 -8.83641839e-01 -5.02149940e-01
1.14960067e-01 5.42074323e-01 1.27178597e+00 -2.21963659e-01
-2.72092164e-01 3.61467570e-01 -4.87101197e-01 -6.30317807e-01
-5.79522908e-01 5.84324658e-01 8.07935576e-05 -5.72570801e-01
2.93438584e-01 -5.91186225e-01 -6.26650155e-01 -5.39228201e-01
-1.18153358e+00 -4.78240438e-02 8.26166034e-01 6.22555554e-01
1.85388431e-01 -2.55880296e-01 1.19640195e+00 -7.00085461e-01
1.51104426e+00 -4.54031438e-01 2.48993725e-01 3.92117023e-01
-7.21850097e-01 3.69352102e-01 5.56066573e-01 -3.14728469e-01
-1.16021156e+00 -4.20460105e-01 -4.44576815e-02 -9.58261490e-02
-7.80099863e-03 9.33782399e-01 8.38001445e-02 4.50262040e-01
1.04257822e+00 4.21338797e-01 -9.18045416e-02 -4.96942610e-01
5.88000059e-01 7.86595702e-01 5.98374665e-01 -2.64029384e-01
1.34584427e-01 1.83249384e-01 -2.02072054e-01 -1.12649095e+00
-6.99529469e-01 -7.74438620e-01 -6.32461846e-01 1.53515562e-01
5.68894744e-01 -7.67677009e-01 9.80801657e-02 2.09835291e-01
-1.25996459e+00 2.35110164e-01 -6.30774677e-01 5.78011498e-02
-5.06119549e-01 8.84940803e-01 -4.58777606e-01 -7.69521177e-01
-1.08822596e+00 -8.73490214e-01 1.17616701e+00 3.88253450e-01
-8.17594707e-01 -8.53829980e-01 4.06507403e-01 2.24856272e-01
6.61833704e-01 2.92132217e-02 1.18114507e+00 -1.18497360e+00
-2.57067531e-01 -4.58933383e-01 1.19205385e-01 2.84188747e-01
1.41660690e-01 -2.26312429e-02 -7.67456055e-01 -3.59603941e-01
7.37408083e-03 -4.17391449e-01 1.26335120e+00 6.25581920e-01
5.94695091e-01 -3.78180563e-01 -1.83119312e-01 -5.45209227e-03
1.32362282e+00 2.86981091e-03 6.40791297e-01 5.20978749e-01
2.76168436e-01 4.98799235e-01 5.46599507e-01 6.09456718e-01
4.48171794e-01 4.43940729e-01 1.81299403e-01 1.19592607e-01
-1.53753459e-01 -2.21348524e-01 3.82812947e-01 8.70603323e-01
5.94432056e-02 -7.37604797e-01 -7.70232320e-01 7.16017008e-01
-2.04857111e+00 -1.33217287e+00 1.98179528e-01 2.12903070e+00
8.26029062e-01 2.38679960e-01 1.88887358e-01 1.82926089e-01
5.99814236e-01 7.72426426e-01 -6.73947155e-01 -9.89132702e-01
-2.65535861e-01 6.98014274e-02 1.30640462e-01 2.08553627e-01
-1.01396465e+00 7.92899370e-01 6.90230179e+00 7.75679290e-01
-9.26770031e-01 -2.18905121e-01 3.60305160e-01 -5.22784233e-01
-3.35459530e-01 -1.52735472e-01 -7.76310265e-01 2.15276584e-01
1.25416148e+00 -7.54887044e-01 1.99251905e-01 8.24240804e-01
2.61580169e-01 -4.28875327e-01 -1.31637383e+00 7.84426332e-01
6.97344780e-01 -1.61783397e+00 4.63638484e-01 -2.10657254e-01
9.51487899e-01 1.18377514e-01 -9.59680006e-02 5.83085895e-01
1.03217781e-01 -8.43419433e-01 4.73214298e-01 4.57759649e-01
6.78103924e-01 -7.60118783e-01 7.31908679e-01 6.66419685e-01
-7.55900860e-01 1.38925621e-02 -5.22606492e-01 -1.35625958e-01
2.30104417e-01 2.52198994e-01 -8.07019711e-01 9.44585204e-01
3.00569981e-01 5.81455946e-01 -7.84019589e-01 1.07218993e+00
2.04101175e-01 3.50820541e-01 -8.05820152e-02 -1.96443409e-01
3.32567036e-01 2.36951858e-01 8.60080838e-01 1.55071485e+00
4.33997691e-01 -1.33965924e-01 1.37962192e-01 4.44571018e-01
-1.46773651e-01 2.74567723e-01 -6.21340096e-01 -1.98822275e-01
4.91897523e-01 9.57485557e-01 -6.01742804e-01 -6.12776458e-01
-3.21336150e-01 1.03215349e+00 2.53951073e-01 2.76344307e-02
-3.01200718e-01 -6.56588197e-01 1.92165017e-01 3.44461277e-02
1.52288482e-01 -8.10156837e-02 -5.34134448e-01 -1.07023168e+00
-2.53888313e-02 -1.04907703e+00 5.36148489e-01 -5.83464384e-01
-1.09042120e+00 4.64587033e-01 3.53292286e-01 -1.11522877e+00
-8.15122724e-01 3.33496660e-04 -1.09518433e+00 8.33061635e-01
-1.17836082e+00 -8.86266947e-01 6.80057555e-02 -4.84487377e-02
1.26741362e+00 -4.45065588e-01 1.02220941e+00 -5.51328659e-02
-2.62127817e-01 4.06930447e-01 5.68379581e-01 -5.43482184e-01
7.73167193e-01 -1.24127090e+00 7.52286017e-01 1.02432680e+00
2.06491333e-02 6.56331420e-01 1.05306149e+00 -6.84698880e-01
-1.27882016e+00 -8.03173006e-01 1.16928518e+00 -4.32672501e-01
3.91164392e-01 1.40237778e-01 -9.23245847e-01 3.64346951e-01
6.90991640e-01 -8.43566477e-01 8.18697870e-01 1.59187660e-01
3.50569934e-02 2.74249196e-01 -1.03596926e+00 7.71764576e-01
7.36221373e-01 -3.57366294e-01 -1.24538159e+00 2.45161414e-01
8.58970523e-01 -2.56832629e-01 -4.45551127e-01 2.42274433e-01
4.27508324e-01 -9.89955068e-01 9.02307272e-01 -7.92657733e-01
9.33009744e-01 6.90665469e-02 -1.71826348e-01 -1.66951251e+00
-3.33661109e-01 -4.94264305e-01 -4.90685374e-01 1.37191081e+00
2.71098703e-01 -1.16096549e-01 7.24992394e-01 4.15739983e-01
-6.27026141e-01 -8.61742914e-01 -9.24624920e-01 -8.12998295e-01
2.60356426e-01 2.26174034e-02 3.44256580e-01 6.62489593e-01
4.34294403e-01 7.61566699e-01 -3.98373276e-01 -4.42915201e-01
1.71785802e-01 -5.36076911e-02 7.04826415e-01 -1.26569021e+00
-1.52735204e-01 -7.02691376e-01 -2.19547719e-01 -9.14212584e-01
-7.65039176e-02 -1.00615823e+00 -1.52810551e-02 -2.51063609e+00
4.12582487e-01 4.97307301e-01 -1.78408280e-01 8.60051513e-02
-3.12542021e-01 -2.40768045e-01 3.04300100e-01 1.26858220e-01
-9.93269622e-01 7.39997447e-01 8.21709692e-01 -2.12622285e-01
-6.60568297e-01 -1.02543518e-01 -1.21353817e+00 4.70398426e-01
8.54330301e-01 -3.47310781e-01 -5.77014983e-01 -6.13170445e-01
6.46162555e-02 1.05914667e-01 -1.73527285e-01 -1.25613070e+00
5.87588727e-01 1.07538931e-01 1.60301492e-01 -1.01658583e+00
3.57340723e-01 1.20134823e-01 -2.28025287e-01 4.22657311e-01
-7.96724856e-01 2.64336765e-01 3.35516423e-01 5.20425379e-01
-3.29445869e-01 -5.99799156e-01 5.28529763e-01 -4.76610392e-01
-6.63419425e-01 -2.70298988e-01 -5.82415998e-01 1.94006920e-01
6.12266541e-01 -4.83051240e-01 -6.97843432e-01 -5.49856961e-01
-2.73449659e-01 1.91569731e-01 3.15262944e-01 5.35309672e-01
5.75073481e-01 -8.70030463e-01 -1.06005132e+00 -2.83466369e-01
1.77813172e-01 -3.10906060e-02 3.91182780e-01 4.03364569e-01
-4.12376136e-01 8.34248483e-01 -2.83292592e-01 -3.14319342e-01
-1.33743846e+00 4.49292839e-01 -1.40894771e-01 -7.47683704e-01
-7.07594693e-01 6.83506608e-01 -3.97275209e-01 -1.38120264e-01
2.90559649e-01 -1.64662957e-01 -3.95932764e-01 5.66799641e-01
6.89000785e-01 8.45833361e-01 3.70469391e-01 -4.07433420e-01
-1.54393598e-01 3.13701704e-02 -5.63765287e-01 -1.93365514e-01
1.43601680e+00 -1.29726961e-01 3.88377830e-02 5.67824781e-01
9.27772522e-01 -2.79591739e-01 -6.74915493e-01 -2.20374331e-01
3.69239420e-01 -2.09519431e-01 9.05680563e-03 -9.46678340e-01
-2.00621322e-01 9.76019144e-01 2.99674302e-01 3.47544432e-01
1.01264465e+00 -3.02459374e-02 7.11421788e-01 9.30936933e-01
-2.08809916e-02 -1.30881262e+00 3.94200474e-01 6.30757213e-01
1.04333687e+00 -1.12272906e+00 5.72320044e-01 2.80396402e-01
-9.06516731e-01 1.03705466e+00 3.90696168e-01 -1.87880620e-01
-6.31877109e-02 -1.27796561e-01 -1.86834842e-01 -1.17634296e-01
-1.10475814e+00 -1.22581869e-01 3.15791935e-01 4.21487868e-01
7.62369633e-01 -2.79297054e-01 -5.92617333e-01 4.88845319e-01
-2.90787458e-01 1.95795111e-02 7.22314060e-01 1.19473016e+00
-9.37457681e-01 -9.49622929e-01 -8.52978826e-02 9.54102814e-01
-6.55541301e-01 -3.23718935e-01 -6.38099790e-01 6.51023865e-01
-4.48642373e-01 1.14075494e+00 3.67777310e-02 -2.06907168e-01
4.49052095e-01 2.43673950e-01 4.17727500e-01 -1.06003225e+00
-7.47849703e-01 -7.83368722e-02 3.55918527e-01 -1.91717640e-01
-4.34430212e-01 -7.91103959e-01 -9.91713345e-01 -2.54051834e-01
-3.61884594e-01 3.77512686e-02 6.59267068e-01 1.00142777e+00
6.14749134e-01 6.73372328e-01 1.58089682e-01 -1.27941787e+00
-1.02344084e+00 -1.36017609e+00 -3.68701339e-01 1.84712738e-01
2.43026122e-01 -1.80301905e-01 5.22714527e-03 -1.28105521e-01] | [12.40308666229248, 9.439666748046875] |
6fcb840a-de3e-4dd6-b067-af45874e8cc2 | video-representation-learning-by-dense | 1909.04656 | null | https://arxiv.org/abs/1909.04656v3 | https://arxiv.org/pdf/1909.04656v3.pdf | Video Representation Learning by Dense Predictive Coding | The objective of this paper is self-supervised learning of spatio-temporal embeddings from video, suitable for human action recognition. We make three contributions: First, we introduce the Dense Predictive Coding (DPC) framework for self-supervised representation learning on videos. This learns a dense encoding of spatio-temporal blocks by recurrently predicting future representations; Second, we propose a curriculum training scheme to predict further into the future with progressively less temporal context. This encourages the model to only encode slowly varying spatial-temporal signals, therefore leading to semantic representations; Third, we evaluate the approach by first training the DPC model on the Kinetics-400 dataset with self-supervised learning, and then finetuning the representation on a downstream task, i.e. action recognition. With single stream (RGB only), DPC pretrained representations achieve state-of-the-art self-supervised performance on both UCF101(75.7% top1 acc) and HMDB51(35.7% top1 acc), outperforming all previous learning methods by a significant margin, and approaching the performance of a baseline pre-trained on ImageNet. | ['Weidi Xie', 'Tengda Han', 'Andrew Zisserman'] | 2019-09-10 | null | null | null | null | ['self-supervised-action-recognition'] | ['computer-vision'] | [ 5.35054088e-01 2.50994712e-01 -4.89516258e-01 -3.43257576e-01
-5.42024374e-01 -1.92699686e-01 9.10931766e-01 -6.08620942e-02
-5.33499241e-01 5.16658664e-01 7.76189864e-01 1.58432975e-01
7.12488368e-02 -4.89949793e-01 -1.03105366e+00 -5.80426812e-01
-6.02016211e-01 2.30774611e-01 5.41625082e-01 -4.52261120e-02
3.06256264e-02 2.52796561e-01 -1.87947297e+00 9.00092065e-01
3.13516438e-01 1.23258984e+00 9.83287171e-02 9.08488333e-01
3.12993884e-01 1.51304519e+00 -2.13359147e-01 -9.09221247e-02
4.73797023e-02 -4.80771899e-01 -1.07669234e+00 2.77803212e-01
4.54916984e-01 -4.38762635e-01 -8.17943990e-01 4.44111735e-01
1.99364801e-03 3.52004021e-01 7.26385176e-01 -9.72903550e-01
-9.53760684e-01 3.58039647e-01 -2.15245411e-01 5.28501570e-01
3.59748065e-01 4.71649349e-01 1.07976699e+00 -7.64505029e-01
8.98246348e-01 1.17009997e+00 5.27080178e-01 8.26482296e-01
-1.23709595e+00 -3.69159698e-01 4.71861809e-01 6.76246405e-01
-1.26383507e+00 -5.35104930e-01 4.78602380e-01 -4.21176136e-01
1.55561697e+00 -1.83620408e-01 8.35809886e-01 1.60885477e+00
1.99460030e-01 1.18884516e+00 9.35169697e-01 -2.02563480e-01
3.62612933e-01 -2.60045916e-01 -2.12688968e-01 6.02607369e-01
-4.45230663e-01 4.25291181e-01 -7.72445798e-01 3.42136919e-01
6.98252082e-01 1.28924087e-01 -9.66822356e-02 -4.25590724e-01
-1.46521676e+00 6.77407146e-01 6.35245800e-01 3.85876149e-01
-4.28846776e-01 7.30814576e-01 7.54807293e-01 2.44698793e-01
5.84913373e-01 2.43713841e-01 -6.59486055e-01 -5.72023153e-01
-8.53651762e-01 1.02838218e-01 2.51388282e-01 8.33303273e-01
6.93532526e-01 1.32230878e-01 -3.50041747e-01 5.01621127e-01
7.50647709e-02 2.94536799e-01 9.96520102e-01 -1.11799347e+00
4.72384959e-01 3.62872481e-01 -9.28429887e-02 -6.45311296e-01
-3.08171928e-01 -3.67031753e-01 -7.03519881e-01 9.50824767e-02
1.29922092e-01 9.78670791e-02 -1.08433652e+00 1.60743058e+00
-1.21982902e-01 7.68118799e-01 3.73558968e-01 7.32072055e-01
4.85387534e-01 1.01549745e+00 5.45230985e-01 -1.72573730e-01
9.48616862e-01 -1.37733316e+00 -5.31763256e-01 -2.57418841e-01
9.33574021e-01 -5.00176623e-02 7.95169413e-01 3.99507284e-01
-7.90449739e-01 -1.17927337e+00 -9.07579839e-01 -1.30959246e-02
-3.28652561e-01 2.36800492e-01 6.86244071e-01 -4.82792929e-02
-1.26840961e+00 1.08769119e+00 -1.26800931e+00 -4.86677438e-01
7.22992063e-01 9.62491184e-02 -6.85143948e-01 -1.79490030e-01
-1.12946951e+00 7.80903339e-01 6.78562701e-01 -3.27546418e-01
-1.21853137e+00 -8.80336583e-01 -1.08511996e+00 -2.31072698e-02
8.57311413e-02 -3.56991738e-01 1.04182291e+00 -1.38343501e+00
-1.51201713e+00 9.23894703e-01 -1.98406845e-01 -1.14004803e+00
3.57252896e-01 -5.28606296e-01 -7.18212545e-01 6.04698300e-01
1.34524837e-01 1.31622052e+00 1.03662384e+00 -8.79989266e-01
-7.73835897e-01 -1.22408085e-01 -9.53530241e-03 1.51347056e-01
-4.13603276e-01 -2.89871782e-01 -3.57908577e-01 -7.38311529e-01
-2.81701595e-01 -9.40782070e-01 -2.40787804e-01 1.70423493e-01
1.49953514e-01 -5.93455732e-01 9.96468782e-01 -5.01789510e-01
9.87471104e-01 -2.34257674e+00 3.96416873e-01 -3.64717603e-01
-1.36852026e-01 4.97544795e-01 -4.99161333e-01 3.85578334e-01
-5.55847824e-01 -2.60181069e-01 -1.39707714e-01 -4.56305414e-01
-1.68647438e-01 5.59754550e-01 -4.65797991e-01 4.94747549e-01
7.64759004e-01 1.17884135e+00 -1.19015503e+00 -3.63973707e-01
3.90962332e-01 5.35906315e-01 -7.55547225e-01 3.37004721e-01
-4.29950356e-01 4.66513962e-01 -1.82045743e-01 4.63232368e-01
6.09672517e-02 -5.11792004e-01 1.43679991e-01 -2.24612400e-01
-7.05628246e-02 2.97454178e-01 -6.75934374e-01 2.20641398e+00
-4.13963765e-01 8.38499010e-01 -7.23819852e-01 -1.37141287e+00
7.18840897e-01 3.16010386e-01 7.51271844e-01 -8.92842174e-01
-2.47794166e-01 -8.90928134e-02 -3.64908993e-01 -6.07503295e-01
2.39076093e-01 -5.35605922e-02 1.04848653e-01 3.13904315e-01
7.23051906e-01 4.07378078e-01 2.04093829e-01 2.31433406e-01
1.48477304e+00 7.64783204e-01 9.96391028e-02 -2.73920246e-03
5.35136759e-01 -3.42226662e-02 4.05476928e-01 2.79065698e-01
-4.69733417e-01 6.67658985e-01 5.22749066e-01 -7.59202242e-01
-7.85487473e-01 -9.61734891e-01 1.07372887e-01 1.34743917e+00
-1.63580582e-01 -6.93928182e-01 -3.91334653e-01 -9.85416651e-01
1.06384039e-01 6.60650313e-01 -1.08487463e+00 -3.93217206e-01
-7.40078747e-01 -1.50026500e-01 1.98402524e-01 1.08944988e+00
4.77065980e-01 -1.28963828e+00 -8.64540637e-01 2.57224947e-01
1.82110555e-02 -1.33746576e+00 -8.29600692e-02 5.06738245e-01
-1.13200998e+00 -9.88635540e-01 -8.45278919e-01 -6.62630022e-01
3.07277113e-01 5.56340292e-02 1.11871541e+00 -1.06825016e-01
-3.16349238e-01 8.25327516e-01 -8.34092081e-01 1.62610516e-01
-1.60884231e-01 -2.89863665e-02 -1.18417311e-02 1.96148381e-01
4.19828862e-01 -6.68498337e-01 -7.16105521e-01 3.90008204e-02
-8.12077820e-01 4.90484498e-02 6.05741441e-01 8.31293106e-01
7.14058697e-01 -1.18791923e-01 4.50577497e-01 -6.02484405e-01
-2.05247223e-01 -6.75598919e-01 -9.43049937e-02 1.79585796e-02
-4.42434698e-01 1.80094391e-01 7.36104250e-01 -5.51026821e-01
-9.81153309e-01 4.86544043e-01 -1.67778835e-01 -8.64108264e-01
-4.79240656e-01 1.29868031e-01 2.52311498e-01 3.31632346e-01
6.72533453e-01 5.69268823e-01 4.93882857e-02 -4.20999348e-01
5.83846152e-01 3.45239848e-01 7.85589933e-01 -2.71263242e-01
4.76290107e-01 6.51849687e-01 -2.04804242e-01 -6.72924042e-01
-1.09689331e+00 -5.88162363e-01 -1.08654499e+00 -2.86350757e-01
1.22529995e+00 -1.21040130e+00 -4.19704437e-01 3.44894290e-01
-8.83862436e-01 -8.89942586e-01 -7.07992375e-01 6.01519644e-01
-1.13299263e+00 2.31096625e-01 -6.00702703e-01 -5.25983870e-01
1.46416828e-01 -7.34331250e-01 1.17814374e+00 -6.64779842e-02
-4.04692739e-01 -1.02762055e+00 3.13397944e-01 3.00989419e-01
2.13577285e-01 3.25972378e-01 4.57047015e-01 -5.01984835e-01
-4.68320400e-01 -1.22919045e-02 -1.81526959e-01 6.41996562e-01
1.45212635e-02 -2.45912448e-01 -1.13972580e+00 -3.63232523e-01
-4.30557519e-01 -9.38427746e-01 1.48490834e+00 2.09964126e-01
1.50905836e+00 -2.13469610e-01 -3.80306453e-01 8.63049328e-01
1.25726652e+00 5.17235883e-02 1.03444684e+00 4.33114827e-01
5.88066757e-01 4.98156250e-01 6.67082608e-01 5.74233830e-01
3.35706264e-01 6.85528219e-01 6.44102693e-01 1.06379479e-01
-5.02720892e-01 -5.85086048e-01 6.22577727e-01 4.30774748e-01
-2.97103316e-01 -2.54949871e-02 -6.43883169e-01 8.29561055e-01
-2.07094622e+00 -1.11292922e+00 2.80729949e-01 1.62897408e+00
7.23017097e-01 1.32427961e-01 2.84445643e-01 2.62961000e-01
2.96433091e-01 5.99270761e-01 -6.82271361e-01 -2.23058075e-01
1.16577763e-02 4.60085690e-01 5.91059089e-01 1.17116898e-01
-1.67916131e+00 1.21883237e+00 6.43146181e+00 5.52627325e-01
-1.11258662e+00 1.48488462e-01 6.33154929e-01 -2.05313042e-01
1.24552295e-01 -1.22242026e-01 -6.23623013e-01 5.47323167e-01
1.44729638e+00 3.26526940e-01 1.45778134e-01 9.41741765e-01
2.39622202e-02 8.52409750e-02 -1.33901584e+00 9.60761368e-01
2.96013594e-01 -1.57062531e+00 7.70400017e-02 -2.00724110e-01
8.41792047e-01 2.79788792e-01 8.41826573e-02 7.30140805e-01
2.79736638e-01 -1.19685817e+00 6.98657632e-01 6.56675220e-01
1.00249934e+00 -5.50208211e-01 4.57895011e-01 1.07604787e-01
-1.13792300e+00 -3.51540476e-01 -3.27779949e-01 -2.16263384e-01
1.80827409e-01 7.36107901e-02 -3.40845615e-01 2.52472550e-01
8.72318327e-01 1.79682684e+00 -7.19829679e-01 6.49367630e-01
-3.78165662e-01 7.98952043e-01 1.67416990e-01 2.16662601e-01
6.71632826e-01 2.99004048e-01 1.04767747e-01 1.52015221e+00
6.83826730e-02 2.45964319e-01 6.41265363e-02 4.81540293e-01
8.59905630e-02 -2.04273120e-01 -5.71864963e-01 -2.35390082e-01
-5.50377667e-02 7.54879892e-01 -4.24316734e-01 -4.69941765e-01
-4.62924510e-01 1.57008350e+00 6.10462427e-01 4.93854463e-01
-8.42333972e-01 1.76300630e-02 7.15623021e-01 1.47966772e-01
8.93587708e-01 -2.05001399e-01 3.10496897e-01 -1.14433563e+00
-1.65167242e-01 -6.22279584e-01 5.94236732e-01 -8.48763943e-01
-1.00331855e+00 5.03441870e-01 -8.26032087e-02 -1.50448000e+00
-5.88970244e-01 -9.36266959e-01 -4.38137025e-01 3.11558366e-01
-1.79215074e+00 -1.17817688e+00 -7.60893151e-02 6.99318051e-01
7.72209764e-01 -4.22761798e-01 1.05668557e+00 1.07327200e-01
-2.56944478e-01 3.98960501e-01 4.35442217e-02 1.30202577e-01
5.87864041e-01 -1.17896223e+00 4.52130675e-01 6.13977432e-01
4.20515597e-01 1.39823779e-01 3.67697477e-01 -3.86134118e-01
-1.34678948e+00 -1.51006055e+00 9.01586056e-01 -4.63449538e-01
9.41471756e-01 -2.26320654e-01 -9.13860977e-01 9.16572809e-01
1.29104659e-01 7.49774992e-01 6.44659638e-01 -5.59073174e-03
-6.41121030e-01 -1.79779381e-01 -7.49319613e-01 2.21711934e-01
1.39239788e+00 -6.99566007e-01 -6.92704558e-01 4.93149996e-01
8.69088948e-01 -1.56754941e-01 -8.51956844e-01 2.64919847e-01
5.14774323e-01 -1.01738012e+00 1.06368792e+00 -1.09456885e+00
9.29516733e-01 -6.23497032e-02 -2.43885800e-01 -1.21947467e+00
-6.55828416e-01 -4.78583634e-01 -6.29315078e-01 7.38880396e-01
1.66257471e-01 -2.02165440e-01 9.87409651e-01 -4.45350409e-02
-2.98244834e-01 -8.55566740e-01 -9.79933798e-01 -1.06201506e+00
1.37666360e-01 -6.51954234e-01 6.09894618e-02 6.96175694e-01
6.64722472e-02 1.28205895e-01 -6.15353405e-01 -1.56900406e-01
3.18795294e-01 -8.01689327e-02 5.89030623e-01 -7.63315141e-01
-5.29081702e-01 -1.68787703e-01 -9.72341359e-01 -1.53963959e+00
5.13357341e-01 -9.20073032e-01 6.04156964e-02 -1.41051877e+00
9.21152532e-02 3.27055007e-02 -5.45281589e-01 5.94062567e-01
1.24874361e-01 4.02610123e-01 1.49541482e-01 3.39929789e-01
-1.23507869e+00 9.63697374e-01 1.07696724e+00 -2.91673422e-01
1.43917173e-01 -3.18923503e-01 -2.35647485e-01 5.22051334e-01
5.03483891e-01 -2.57779717e-01 -5.81222415e-01 -5.58278024e-01
-2.80461937e-01 -8.74853209e-02 6.16284311e-01 -1.43844771e+00
1.46153541e-02 -1.11972123e-01 6.60722971e-01 -5.08246779e-01
4.52111363e-01 -6.39689207e-01 -3.02160054e-01 6.33073330e-01
-7.61171937e-01 -1.98092073e-01 2.02816218e-01 1.09925342e+00
-3.73882562e-01 6.81325793e-02 7.83307314e-01 -1.39029488e-01
-1.42788720e+00 3.04084778e-01 -5.51514208e-01 2.80766729e-02
1.18569922e+00 -2.13174000e-01 -1.33114249e-01 -3.82468939e-01
-1.09597266e+00 1.67972267e-01 2.76091874e-01 6.61719143e-01
7.78507650e-01 -1.54729414e+00 -5.45653701e-01 3.17178071e-01
5.27308822e-01 -4.19668496e-01 4.12676901e-01 6.45007312e-01
-3.74209374e-01 6.83492661e-01 -3.83659691e-01 -8.17327559e-01
-8.96050930e-01 8.14644873e-01 1.60123885e-01 -2.45608822e-01
-1.05087447e+00 1.09102929e+00 5.57611249e-02 -3.42499698e-03
4.74663347e-01 -3.98532361e-01 -3.40713143e-01 -4.96443622e-02
7.79135346e-01 1.30339131e-01 -2.13390350e-01 -8.67754877e-01
-4.63928223e-01 5.29707372e-01 -7.12801442e-02 3.24323736e-02
1.63023067e+00 9.57951099e-02 4.28480923e-01 5.35122097e-01
1.77313590e+00 -8.78509700e-01 -1.92726469e+00 -3.18273664e-01
1.14711054e-01 -4.42004889e-01 2.98544001e-02 -6.27364039e-01
-1.04948032e+00 9.98395979e-01 7.56115198e-01 -1.87789902e-01
1.02006161e+00 2.43690684e-01 8.04264188e-01 3.65533233e-01
1.66624799e-01 -1.25289202e+00 8.03458512e-01 7.48147607e-01
9.73595977e-01 -1.27622080e+00 -6.58310726e-02 7.17160404e-02
-1.02145219e+00 1.26158988e+00 5.39985240e-01 -6.01920426e-01
7.07311749e-01 -1.29494652e-01 -2.54337132e-01 -2.32607618e-01
-1.24701798e+00 -4.90077853e-01 4.70072687e-01 8.00688267e-01
4.74360496e-01 -1.18923970e-01 1.85729682e-01 3.12248111e-01
2.45612353e-01 2.84637868e-01 1.42248377e-01 1.01244855e+00
-2.77235091e-01 -8.10833514e-01 2.81769514e-01 3.12549621e-01
-2.04334304e-01 1.19622596e-01 -1.02415219e-01 6.41726911e-01
2.82314032e-01 5.90762317e-01 5.89233935e-01 -5.97561002e-01
3.17539603e-01 1.42984658e-01 4.82843131e-01 -7.71291077e-01
-1.22299477e-01 -1.47546113e-01 4.08123359e-02 -1.17951369e+00
-7.63583481e-01 -8.41432571e-01 -1.20806539e+00 1.57037050e-01
3.50389451e-01 -1.58823326e-01 2.13914573e-01 9.69114542e-01
5.32442987e-01 6.73077881e-01 5.66545010e-01 -1.06862807e+00
-4.14055884e-01 -9.97591615e-01 -4.34668899e-01 7.34269738e-01
3.84825408e-01 -8.14963162e-01 -3.25732619e-01 5.55381119e-01] | [8.647808074951172, 0.8116270899772644] |
a7c9aea5-d353-4f46-988b-cb6a9e0bbd86 | a-label-attention-model-for-icd-coding-from | 2007.06351 | null | https://arxiv.org/abs/2007.06351v1 | https://arxiv.org/pdf/2007.06351v1.pdf | A Label Attention Model for ICD Coding from Clinical Text | ICD coding is a process of assigning the International Classification of Disease diagnosis codes to clinical/medical notes documented by health professionals (e.g. clinicians). This process requires significant human resources, and thus is costly and prone to error. To handle the problem, machine learning has been utilized for automatic ICD coding. Previous state-of-the-art models were based on convolutional neural networks, using a single/several fixed window sizes. However, the lengths and interdependence between text fragments related to ICD codes in clinical text vary significantly, leading to the difficulty of deciding what the best window sizes are. In this paper, we propose a new label attention model for automatic ICD coding, which can handle both the various lengths and the interdependence of the ICD code related text fragments. Furthermore, as the majority of ICD codes are not frequently used, leading to the extremely imbalanced data issue, we additionally propose a hierarchical joint learning mechanism extending our label attention model to handle the issue, using the hierarchical relationships among the codes. Our label attention model achieves new state-of-the-art results on three benchmark MIMIC datasets, and the joint learning mechanism helps improve the performances for infrequent codes. | ['Anthony Nguyen', 'Thanh Vu', 'Dat Quoc Nguyen'] | 2020-07-13 | null | null | null | null | ['medical-code-prediction'] | ['medical'] | [-9.35724820e-04 3.24549526e-02 -5.41777074e-01 -4.03149247e-01
-6.23074055e-01 -2.20185354e-01 -1.14854895e-01 8.36918652e-01
-2.10223466e-01 4.31796104e-01 3.29273641e-01 -4.82778281e-01
-1.64295331e-01 -7.34795272e-01 -2.67529845e-01 -5.04806280e-01
-7.32602999e-02 8.46111417e-01 -5.22072949e-02 1.12104207e-01
1.36813655e-01 1.23994686e-01 -1.21946013e+00 6.60999537e-01
1.06458640e+00 1.12963665e+00 -2.84554306e-02 2.71241784e-01
-6.45706832e-01 1.18981934e+00 -5.68021357e-01 -3.68245214e-01
-5.72822057e-02 -3.69143784e-01 -8.68189752e-01 -5.46613894e-02
-1.21700186e-02 -4.93463546e-01 -3.02132666e-01 9.55170333e-01
4.45083618e-01 -3.13177913e-01 8.37863624e-01 -1.20244253e+00
-9.25130188e-01 6.00118160e-01 -6.82005286e-01 2.57256091e-01
2.15005636e-01 -2.92117417e-01 1.13921058e+00 -6.55938208e-01
2.73837209e-01 9.51272130e-01 9.97875035e-01 3.14372778e-01
-8.59244704e-01 -1.06984234e+00 1.26414165e-01 3.45430195e-01
-1.55309618e+00 -1.53464794e-01 6.89946711e-01 -8.09842050e-01
9.96453226e-01 -6.91284612e-02 5.64918399e-01 8.06962848e-01
4.19995666e-01 5.14441788e-01 3.82846981e-01 -5.09029925e-01
1.27396673e-01 -1.02771036e-01 3.92267764e-01 6.24629319e-01
3.48561317e-01 -3.52975547e-01 6.06422648e-02 -5.16667783e-01
6.19972408e-01 6.44288301e-01 -1.43730253e-01 -1.20321967e-01
-1.08514142e+00 1.08239031e+00 4.94406521e-01 4.67340350e-01
-3.09556097e-01 -4.17949166e-03 9.53710020e-01 2.10069623e-02
4.85359788e-01 2.64832497e-01 -5.90168417e-01 -5.38444966e-02
-1.01094294e+00 -3.72234248e-02 4.92479920e-01 9.65073228e-01
3.82006735e-01 -5.33933759e-01 -2.62910783e-01 1.11501205e+00
3.97932351e-01 -3.03243455e-02 6.56136155e-01 -5.53850830e-01
8.95975709e-01 1.06403613e+00 -6.44535720e-02 -1.10168767e+00
-7.64783859e-01 -4.18686360e-01 -1.41072226e+00 -2.70491838e-01
1.24061801e-01 -1.10090010e-01 -1.04657531e+00 1.46718538e+00
4.73873541e-02 -3.67812961e-02 -2.73473352e-01 5.24278700e-01
9.02140498e-01 6.06265604e-01 3.58396828e-01 -4.07995671e-01
1.57187533e+00 -7.71921277e-01 -9.86932039e-01 -8.16300288e-02
1.06809855e+00 -5.91252685e-01 6.31134868e-01 1.41076192e-01
-7.61812329e-01 -6.02430582e-01 -5.91959476e-01 -2.92212427e-01
-1.86203703e-01 3.77855122e-01 6.05765402e-01 3.36877972e-01
-7.51858354e-01 2.11958796e-01 -6.93392634e-01 -2.37200886e-01
6.29009902e-01 4.06031877e-01 -2.56003916e-01 -1.31517708e-01
-1.40245664e+00 5.13830543e-01 5.81396043e-01 -8.47306624e-02
-2.30235353e-01 -6.91105306e-01 -8.70239198e-01 4.33756202e-01
9.50280353e-02 -5.47773361e-01 1.02256811e+00 -1.00857270e+00
-7.68302560e-01 1.00176036e+00 8.11316893e-02 -3.57524335e-01
3.53851378e-01 5.09409979e-02 -3.86493951e-01 7.25028478e-03
3.31344336e-01 6.51323080e-01 4.22506869e-01 -7.80913413e-01
-6.96532547e-01 -3.02038044e-01 2.37435494e-02 3.89132872e-02
-5.13448238e-01 1.40332967e-01 -4.28772241e-01 -9.99911785e-01
1.04651600e-01 -9.13333416e-01 -1.66976005e-01 1.31553054e-01
-2.77765542e-01 -6.60945415e-01 2.91100830e-01 -6.79542840e-01
1.82637858e+00 -2.49820256e+00 -5.96759990e-02 -8.37419406e-02
4.60936636e-01 1.91807181e-01 2.62603194e-01 4.32662934e-01
-4.06208694e-01 4.14903492e-01 -2.87322164e-01 -2.10529029e-01
-2.36593977e-01 2.81837076e-01 -3.30335230e-01 3.30460489e-01
-8.11929330e-02 5.82988262e-01 -8.14117968e-01 -9.23085213e-01
-1.04100347e-01 2.66615957e-01 -8.46090853e-01 4.59202617e-01
1.60901308e-01 1.91377938e-01 -4.65258539e-01 7.63795853e-01
7.34261870e-01 -7.56804049e-01 2.50587642e-01 -2.27675542e-01
1.38483897e-01 2.90054262e-01 -8.15306723e-01 1.50745368e+00
-3.36816341e-01 3.29914451e-01 -4.36828852e-01 -1.16102064e+00
7.34625280e-01 5.94611943e-01 9.57841873e-01 -3.42879355e-01
2.72867650e-01 3.08344990e-01 2.43973181e-01 -7.97942281e-01
1.21063506e-02 -1.95227012e-01 -2.37424806e-01 4.02858198e-01
-2.02393919e-01 4.29554015e-01 1.87702313e-01 7.14154914e-02
1.18198526e+00 -5.71957648e-01 7.28196502e-01 -1.15604788e-01
2.73775548e-01 -1.46269217e-01 9.65162158e-01 6.42354071e-01
-2.98476845e-01 9.25211430e-01 7.86633134e-01 -1.04089987e+00
-9.97352123e-01 -5.97784281e-01 -5.74024141e-01 9.74545896e-01
-8.61069486e-02 -4.09633815e-01 -5.23607731e-01 -8.64189088e-01
1.44164711e-01 2.69167602e-01 -9.87953365e-01 -2.42672741e-01
-3.97449642e-01 -9.08027351e-01 6.38606071e-01 9.23596919e-01
2.69300580e-01 -1.20771694e+00 -5.74399412e-01 5.68365395e-01
-5.81380486e-01 -8.23291481e-01 -7.37561405e-01 3.89158577e-01
-7.36991584e-01 -1.31170213e+00 -7.32703209e-01 -9.87275839e-01
7.25129187e-01 2.50444189e-02 1.24662459e+00 5.43246746e-01
-2.73170859e-01 -3.39194685e-01 -7.47664392e-01 -5.02028227e-01
-2.28490174e-01 3.45407307e-01 -1.19176313e-01 9.02889743e-02
4.47665691e-01 -3.36686522e-01 -6.63436353e-01 -1.60469636e-02
-1.02881420e+00 1.17028631e-01 4.66165692e-01 1.08850873e+00
4.98489439e-01 1.04220532e-01 6.22985721e-01 -1.08132374e+00
4.96365786e-01 -7.77435839e-01 -2.13081554e-01 2.08558694e-01
-6.58498466e-01 4.83151749e-02 7.94745803e-01 -3.05197775e-01
-5.58979511e-01 -7.52928853e-02 -4.99285340e-01 -3.51541609e-01
-2.34146059e-01 6.64474428e-01 9.36504162e-04 3.02208483e-01
3.02498907e-01 -5.84143363e-02 -3.21280658e-01 -4.83483374e-01
-2.63562888e-01 1.20369565e+00 -2.09273677e-02 -1.41078189e-01
-3.43780890e-02 2.23033398e-01 -3.07688326e-01 -1.27085879e-01
-1.03302324e+00 -6.29618466e-01 -4.63261843e-01 2.93508977e-01
1.07654369e+00 -9.17459309e-01 -5.13754129e-01 4.79781657e-01
-1.40494752e+00 -1.71612948e-02 7.42778331e-02 4.14997846e-01
-3.10600340e-01 4.31201011e-01 -1.04169059e+00 -4.12326157e-01
-5.76968372e-01 -1.34533536e+00 1.05531681e+00 -2.67358273e-01
-6.04434431e-01 -8.51625979e-01 1.72414079e-01 3.19743216e-01
1.29785806e-01 6.53375909e-02 1.63824177e+00 -7.55275369e-01
8.10051784e-02 -2.27419287e-01 -7.11144447e-01 -5.64489188e-03
4.61335838e-01 -6.60341680e-02 -5.48319280e-01 -2.56540418e-01
-2.32011870e-01 -2.23973364e-01 8.71549845e-01 5.95715821e-01
1.77252817e+00 -2.88250297e-01 -6.65029526e-01 7.63862431e-01
1.19234741e+00 6.84857905e-01 5.83297372e-01 2.77948201e-01
9.73152101e-01 3.88634294e-01 3.70637417e-01 7.16503620e-01
6.38753951e-01 7.05898106e-01 4.03054178e-01 -2.99925238e-01
2.73531735e-01 1.64348722e-01 -3.55503738e-01 8.99466336e-01
1.50337085e-01 -4.07958657e-01 -1.35859716e+00 6.73749983e-01
-1.86734474e+00 -6.63382769e-01 -1.66306734e-01 1.86395001e+00
1.12107658e+00 1.20010711e-01 -1.57237917e-01 3.18642616e-01
7.42413759e-01 -1.39606863e-01 -5.89035392e-01 -4.21935856e-01
5.40482223e-01 -1.19429335e-01 3.59043807e-01 3.18216346e-02
-1.25284326e+00 3.26139599e-01 6.24261236e+00 6.71134770e-01
-9.94895458e-01 1.21974811e-01 9.69242752e-01 5.44739189e-03
-1.80624370e-02 -4.00031865e-01 -6.47233486e-01 1.10804331e+00
8.13043952e-01 1.47951975e-01 -6.08414188e-02 8.68997753e-01
-6.83966130e-02 2.75067419e-01 -1.07469642e+00 1.24834359e+00
9.63249523e-03 -1.20423412e+00 1.34550214e-01 2.27842897e-01
6.62262142e-01 -1.36872577e-02 -1.37350738e-01 4.14034665e-01
2.06604362e-01 -8.73135567e-01 5.20663559e-01 1.52044684e-01
1.17763257e+00 -8.53879631e-01 1.38905776e+00 2.65030533e-01
-1.27889919e+00 -4.87501949e-01 -3.62755150e-01 1.31914079e-01
-1.30727291e-01 9.09493506e-01 -5.70404232e-01 2.93887049e-01
9.24767137e-01 8.41850460e-01 -4.58518863e-01 1.14315474e+00
1.79720894e-01 4.96416777e-01 2.14305781e-02 4.38254744e-01
1.78960502e-01 2.21374422e-01 -2.40180165e-01 1.33328581e+00
5.97684801e-01 1.01648167e-01 5.10374606e-01 4.23011780e-01
-4.68385905e-01 3.53518695e-01 -4.03162211e-01 4.18102778e-02
4.05300051e-01 8.66759241e-01 -8.94334972e-01 -5.47819078e-01
-5.43189049e-01 7.10191369e-01 6.01298749e-01 6.91367835e-02
-1.13798892e+00 -4.73368168e-01 5.37223756e-01 2.07089290e-01
2.00436220e-01 3.42393935e-01 -4.61410940e-01 -1.15799749e+00
-1.83935277e-02 -8.64532113e-01 8.91489923e-01 -4.69263345e-01
-1.37822080e+00 6.75815105e-01 -2.87607968e-01 -1.65932178e+00
-6.15620799e-02 -2.86111563e-01 -3.00635278e-01 4.46113765e-01
-1.38620734e+00 -8.63408685e-01 -2.71783113e-01 3.32686126e-01
6.10084891e-01 -1.96149969e-03 9.23208952e-01 8.84401977e-01
-5.83331704e-01 8.64863575e-01 1.80199698e-01 6.86539888e-01
1.01214588e+00 -1.17437410e+00 1.83238760e-01 9.34431627e-02
-2.88859159e-01 6.04759753e-01 1.89262584e-01 -5.93107700e-01
-6.88442886e-01 -1.31238306e+00 1.30607474e+00 -1.04083396e-01
3.69090170e-01 -1.01825729e-01 -1.22603059e+00 7.03282714e-01
5.93251130e-03 3.13672632e-01 1.21504962e+00 2.10859179e-01
-4.59739298e-01 -1.88606769e-01 -1.06603563e+00 1.74163848e-01
9.15686548e-01 -3.33398461e-01 -6.13157570e-01 4.40118045e-01
7.46918917e-01 -3.21964502e-01 -8.64481330e-01 4.34889197e-01
5.37553132e-01 -8.16003203e-01 8.24373186e-01 -5.27574182e-01
7.76922941e-01 -1.46107435e-01 2.42779210e-01 -1.13556147e+00
-8.59842241e-01 2.25953683e-01 -2.43518297e-02 1.17373633e+00
3.24795753e-01 -4.99464363e-01 4.12389904e-01 4.41246450e-01
-1.23524703e-01 -9.91636634e-01 -1.07855368e+00 -3.49694848e-01
7.59205520e-02 -1.68668911e-01 7.03886390e-01 1.32787001e+00
3.16225082e-01 7.11110383e-02 -4.79306281e-01 -3.59363034e-02
2.33227864e-01 2.92429298e-01 4.35420088e-02 -1.52708411e+00
-1.40440106e-01 -5.33889472e-01 -4.23737466e-01 -6.43424690e-01
5.58258407e-03 -1.03114331e+00 4.09285985e-02 -1.77838290e+00
5.37298918e-01 -7.58979619e-01 -6.38911009e-01 7.82575607e-01
-4.26985919e-01 9.36339051e-02 -3.65558341e-02 6.34663939e-01
-7.49441624e-01 2.21104056e-01 8.54859710e-01 -4.01131392e-01
-1.88814431e-01 -2.50906378e-01 -7.22719312e-01 7.07924485e-01
8.00041795e-01 -9.43305910e-01 -2.29414970e-01 -6.19511187e-01
5.13070703e-01 2.80781358e-01 -1.62568018e-01 -1.05732405e+00
2.74794191e-01 -1.13088019e-01 4.62609380e-01 -5.99340260e-01
-2.28106171e-01 -1.05687487e+00 1.24497339e-01 6.70084834e-01
-4.42496032e-01 2.41685182e-01 1.55189320e-01 4.87888575e-01
-4.34521824e-01 -3.26656908e-01 8.04425538e-01 -1.70792252e-01
-2.20714271e-01 6.17080450e-01 -5.73531747e-01 1.17273629e-01
9.25959945e-01 1.29501358e-01 2.92730443e-02 -2.85244826e-02
-5.77230096e-01 3.38538319e-01 2.56607801e-01 3.47936898e-01
4.31302547e-01 -1.49017608e+00 -6.48215950e-01 3.33683908e-01
6.40726388e-01 1.68608889e-01 4.51902062e-01 7.65158534e-01
-8.01397800e-01 4.81873453e-01 -1.58701569e-01 -4.30429578e-01
-1.09080052e+00 9.85158384e-01 2.24827528e-01 -8.36068392e-01
-6.18013442e-01 7.03837395e-01 3.25078368e-01 -4.11061317e-01
3.56767565e-01 -7.52684236e-01 -4.49029356e-01 4.52314556e-01
6.93525612e-01 7.41692036e-02 2.32381999e-01 -3.48590970e-01
-4.60755497e-01 7.44220316e-01 -4.71969277e-01 6.90617621e-01
1.13309610e+00 -4.06915694e-02 -2.63056904e-01 2.65053988e-01
1.24906123e+00 -4.06995326e-01 -7.48904586e-01 -1.63150609e-01
-1.66519172e-02 -2.47226551e-01 1.79375000e-02 -5.03969848e-01
-1.26010287e+00 1.04424775e+00 6.02487206e-01 4.46157485e-01
1.18016064e+00 -8.89851972e-02 1.09475148e+00 5.62876984e-02
2.22859725e-01 -1.06160712e+00 -1.91404581e-01 5.05026221e-01
6.23974204e-01 -1.46795011e+00 -1.10490866e-01 -3.09339464e-01
-4.98684675e-01 1.20127416e+00 4.68170881e-01 2.17960417e-01
8.54194939e-01 3.61371219e-01 2.27245271e-01 -7.65949413e-02
-6.93935573e-01 1.13387182e-01 9.84031931e-02 2.78161228e-01
9.04967248e-01 1.52992740e-01 -5.32416403e-01 7.41314948e-01
2.44118646e-01 2.81430304e-01 2.77290821e-01 8.11284423e-01
-3.07647884e-01 -1.10854900e+00 -3.78661841e-01 8.92293572e-01
-9.69192624e-01 -3.56821567e-01 4.96030645e-03 3.47151369e-01
7.20182896e-01 8.51036370e-01 3.71631533e-01 -2.67806500e-01
1.40408680e-01 1.24983869e-01 -1.75219730e-01 -7.69314051e-01
-7.18205035e-01 -2.60806875e-03 -3.20228070e-01 -1.11642674e-01
-9.68326628e-02 -4.59854573e-01 -1.52923942e+00 -1.36335447e-01
-2.37338334e-01 1.36363566e-01 4.36617322e-02 8.99966240e-01
6.83688402e-01 8.19803238e-01 4.64339614e-01 -3.90553445e-01
-3.00460488e-01 -1.06832743e+00 -5.70070624e-01 7.36445010e-01
5.91225147e-01 -7.90125668e-01 -2.20475018e-01 1.36684611e-01] | [8.008199691772461, 6.852161884307861] |
8761c3bd-c267-457c-a38f-daa501d89b5b | unpaired-image-to-image-translation-with-2 | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xie_Unpaired_Image-to-Image_Translation_With_Shortest_Path_Regularization_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xie_Unpaired_Image-to-Image_Translation_With_Shortest_Path_Regularization_CVPR_2023_paper.pdf | Unpaired Image-to-Image Translation With Shortest Path Regularization | Unpaired image-to-image translation aims to learn proper mappings that can map images from one domain to another domain while preserving the content of the input image. However, with large enough capacities, the network can learn to map the inputs to any random permutation of images in another domain. Existing methods treat two domains as discrete and propose different assumptions to address this problem. In this paper, we start from a different perspective and consider the paths connecting the two domains. We assume that the optimal path length between the input and output image should be the shortest among all possible paths. Based on this assumption, we propose a new method to allow generating images along the path and present a simple way to encourage the network to find the shortest path without pair information. Extensive experiments on various tasks demonstrate the superiority of our approach. | ['Kun Zhang', 'Mingming Gong', 'Yanwu Xu', 'Shaoan Xie'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['image-to-image-translation', 'image-to-image-translation'] | ['computer-vision', 'miscellaneous'] | [ 7.15270817e-01 3.51222217e-01 -2.98106968e-01 -3.51463795e-01
-2.12710410e-01 -8.46722245e-01 4.87158120e-01 -4.29116368e-01
-3.11545193e-01 7.46234357e-01 -3.49656008e-02 -2.62385577e-01
-1.14040107e-01 -1.05135822e+00 -1.09203744e+00 -6.98021710e-01
2.65677840e-01 3.00274014e-01 1.57801002e-01 -8.29466954e-02
1.74903378e-01 4.21253860e-01 -1.04748189e+00 2.15574384e-01
8.09626460e-01 5.15873313e-01 4.68902320e-01 3.48644197e-01
-8.13912153e-02 3.88058275e-01 -4.35797036e-01 -3.48492503e-01
8.26870024e-01 -9.65849698e-01 -1.04416454e+00 3.23230505e-01
3.67041528e-01 -4.71301049e-01 -5.47543705e-01 1.42223907e+00
2.27260455e-01 5.82837835e-02 5.59332490e-01 -1.61296761e+00
-1.07352614e+00 6.51415348e-01 -5.23922026e-01 -2.38639005e-02
2.30358645e-01 1.15983658e-01 8.06241333e-01 -5.11617541e-01
8.65629971e-01 9.76297081e-01 7.01407716e-02 7.80873775e-01
-1.36267519e+00 -7.44920969e-01 2.44024962e-01 1.69837717e-02
-1.34144199e+00 -3.43828768e-01 8.78263474e-01 -9.06081051e-02
3.78528386e-01 1.02786876e-01 5.14080286e-01 8.86485279e-01
-1.05461486e-01 3.85823607e-01 1.41094911e+00 -6.25327885e-01
2.80603059e-02 3.65270108e-01 -4.30919975e-01 6.22847378e-01
2.83538252e-01 -1.56819392e-02 -3.45436603e-01 1.19262345e-01
1.12731874e+00 -9.76569578e-02 -4.65040982e-01 -6.13418698e-01
-1.49337554e+00 6.06531084e-01 6.01004779e-01 3.37987602e-01
-1.25987083e-01 8.10139850e-02 -9.60522369e-02 5.71255267e-01
-8.29848573e-02 5.89294314e-01 -1.78099290e-01 4.05620337e-01
-5.98001957e-01 1.30983397e-01 6.21305764e-01 1.04267645e+00
8.74601364e-01 -2.70937324e-01 -1.15667909e-01 5.59632003e-01
3.94417457e-02 3.41008365e-01 1.42583296e-01 -1.17342496e+00
7.26196289e-01 2.87318349e-01 1.71713740e-01 -1.23728979e+00
1.56168908e-01 -1.03953689e-01 -8.26345444e-01 1.61113143e-01
7.41401136e-01 -1.34829178e-01 -7.65572309e-01 2.19082856e+00
1.84987113e-01 2.99012512e-01 -3.69375362e-03 1.02131939e+00
2.64883190e-01 8.14996600e-01 -2.48314112e-01 -7.85713866e-02
1.02422404e+00 -9.44020987e-01 -4.76810694e-01 -2.12420896e-01
4.32076693e-01 -5.51176310e-01 1.15076399e+00 1.97019074e-02
-1.18887162e+00 -4.81913567e-01 -1.14497411e+00 6.32878765e-02
-1.82556853e-01 1.69372678e-01 1.48231372e-01 4.11580205e-01
-1.28416288e+00 5.14811695e-01 -3.35303903e-01 -2.81014025e-01
3.62443954e-01 4.88914490e-01 -6.21987522e-01 -4.01341200e-01
-1.23335457e+00 6.46116555e-01 6.62744045e-01 -6.84793591e-02
-7.77228534e-01 -3.10872078e-01 -6.76541209e-01 7.35669285e-02
3.87397677e-01 -8.77729416e-01 1.00924253e+00 -1.61165118e+00
-1.50164866e+00 9.65329468e-01 -2.99179494e-01 -3.76267880e-01
6.77604795e-01 3.35476518e-01 -1.28618777e-01 3.62102598e-01
1.30638421e-01 1.26025498e+00 9.82471824e-01 -1.40794730e+00
-5.70573568e-01 -1.99411418e-02 4.34770077e-01 4.20746237e-01
-5.56323290e-01 -1.94893867e-01 -6.18189394e-01 -5.36901116e-01
2.30627552e-01 -1.22999465e+00 -2.11408019e-01 2.90496379e-01
-7.15835512e-01 2.02967636e-02 6.39785290e-01 -3.37102264e-01
7.57186949e-01 -2.30969191e+00 1.32160097e-01 4.36297745e-01
1.90634936e-01 1.31954953e-01 -5.46698749e-01 3.36760998e-01
-2.88384348e-01 2.53077507e-01 -3.20884377e-01 -1.78614825e-01
-1.95260301e-01 3.85933638e-01 -4.26614642e-01 5.43138742e-01
1.80358395e-01 9.07887936e-01 -9.51774299e-01 -5.01622021e-01
-1.34807840e-01 2.38389567e-01 -6.30012810e-01 2.68662006e-01
-8.25226381e-02 8.14714968e-01 -3.61215085e-01 2.55472809e-02
9.63073373e-01 -3.49870861e-01 3.73787165e-01 -7.01913387e-02
9.58321989e-02 2.14396223e-01 -1.07285774e+00 1.46913922e+00
-2.36477464e-01 6.06009960e-01 -1.61173627e-01 -1.12337852e+00
9.05715168e-01 2.69838840e-01 4.01038140e-01 -8.04757595e-01
4.35862839e-02 9.47495475e-02 2.12986931e-01 -2.63804972e-01
2.23874032e-01 -3.25724721e-01 4.58201580e-02 7.42868185e-01
-2.11876720e-01 1.78008780e-01 6.91653267e-02 1.07544132e-01
7.11198807e-01 3.02921259e-03 5.51879480e-02 -1.16932295e-01
5.21666944e-01 -1.41875371e-01 6.88790858e-01 8.67816985e-01
-1.87787339e-01 8.28478396e-01 8.24379444e-01 -3.52079093e-01
-1.28913915e+00 -1.24985719e+00 4.13974911e-01 6.83413565e-01
5.40191472e-01 7.42436061e-03 -9.16973829e-01 -9.17779028e-01
-2.99878895e-01 4.70254213e-01 -6.42184436e-01 -2.16040298e-01
-7.33927250e-01 -1.41593367e-01 4.65915352e-01 2.43615612e-01
8.59115779e-01 -1.06399202e+00 -2.52404660e-01 -1.33003414e-01
-4.81218606e-01 -1.31541979e+00 -8.86439085e-01 -1.86207190e-01
-8.20239425e-01 -9.03964400e-01 -9.00125563e-01 -1.32674897e+00
1.47611511e+00 5.57516396e-01 8.24665964e-01 1.24112539e-01
2.15667352e-01 4.04384062e-02 -1.16785154e-01 9.16423649e-02
-5.99145293e-01 6.12940714e-02 -5.60477339e-02 1.40385404e-01
-9.14120525e-02 -5.23037732e-01 -7.12838590e-01 5.98948061e-01
-1.21115971e+00 4.50809717e-01 5.73067248e-01 5.74170709e-01
7.18824208e-01 3.73237759e-01 5.31606495e-01 -9.14320588e-01
5.55136859e-01 -4.82254297e-01 -5.58979750e-01 4.07523364e-01
-4.64166522e-01 2.32538596e-01 1.10903275e+00 -5.10192156e-01
-9.21438515e-01 3.62406552e-01 3.16952646e-01 -4.61676717e-01
-2.25741357e-01 1.17238224e-01 -6.31620049e-01 -7.89233372e-02
2.97862768e-01 5.31553924e-01 1.38356984e-01 -1.57428741e-01
6.40301883e-01 2.66697377e-01 6.37736559e-01 -5.37169456e-01
1.09988678e+00 7.12979138e-01 5.88350222e-02 -4.03026640e-01
-4.74060535e-01 1.28058106e-01 -7.43143976e-01 -9.45989266e-02
7.13698208e-01 -5.96844614e-01 -5.30658185e-01 2.72232145e-01
-1.23447084e+00 -3.60247105e-01 -2.75208980e-01 3.06800276e-01
-7.38047957e-01 4.66039956e-01 -2.21909091e-01 -2.37226173e-01
1.43437698e-01 -1.28782630e+00 5.59127033e-01 2.44358674e-01
-1.81877594e-02 -8.86922002e-01 -1.85353532e-01 -1.49901146e-02
2.47885600e-01 7.94674680e-02 1.05653214e+00 -3.74437094e-01
-1.01880181e+00 8.43339711e-02 -4.29047912e-01 3.73723805e-01
3.95889848e-01 -2.08600104e-01 -4.46662664e-01 -3.40914100e-01
-6.71694502e-02 -4.22275402e-02 7.23032475e-01 2.95995682e-01
1.36857319e+00 -7.32250631e-01 -3.77816141e-01 7.46716261e-01
1.53635442e+00 2.34760374e-01 9.60697055e-01 3.98554027e-01
5.19481122e-01 6.84548736e-01 3.54448438e-01 6.37558624e-02
3.32389474e-01 6.32891834e-01 4.74063367e-01 -2.68765002e-01
-1.26818791e-01 -6.47077978e-01 4.39617097e-01 4.58405346e-01
3.05028766e-01 -4.61722940e-01 -5.00334799e-01 6.82232380e-01
-1.60107160e+00 -8.46491337e-01 2.95564353e-01 2.38818860e+00
7.71045268e-01 9.12330970e-02 -5.09047098e-02 -8.31461921e-02
1.02595854e+00 8.94354135e-02 -4.71219331e-01 -4.05633479e-01
-1.99927345e-01 -4.16492335e-02 6.88822150e-01 4.48303491e-01
-8.15361023e-01 8.14124584e-01 6.95663786e+00 5.92851043e-01
-1.28721511e+00 -1.96680203e-01 7.43608594e-01 -1.02561936e-01
-6.24681115e-01 2.90976822e-01 -5.42645276e-01 6.18213296e-01
5.50225079e-01 -3.40303391e-01 7.07687318e-01 3.52300465e-01
2.49784306e-01 1.22344263e-01 -1.21310461e+00 7.07970440e-01
-8.04232284e-02 -1.37427413e+00 3.21554214e-01 1.90928638e-01
1.05834663e+00 -3.90657395e-01 3.59573424e-01 -2.91333854e-01
3.25591475e-01 -1.06068850e+00 6.24290228e-01 3.95538181e-01
1.01453185e+00 -8.27789128e-01 1.60592780e-01 5.50706208e-01
-8.87505412e-01 1.92374974e-01 -3.28967541e-01 -5.51076941e-02
1.43892970e-02 2.95526475e-01 -9.46392655e-01 4.32071835e-01
3.34824920e-01 5.64862430e-01 -3.70202810e-01 8.58730376e-01
-4.79015857e-01 2.77419776e-01 -8.10086429e-02 3.58048409e-01
2.11227790e-01 -5.04923582e-01 5.06377757e-01 7.93769360e-01
5.19946635e-01 -5.66834882e-02 1.16342500e-01 1.12126160e+00
-5.23653507e-01 -6.34849966e-02 -1.17539310e+00 6.42638654e-02
6.34028614e-01 9.56098795e-01 -1.03094375e+00 -3.10705811e-01
-3.91824216e-01 1.35739589e+00 2.85014063e-01 5.25277913e-01
-9.74839926e-01 -4.40806776e-01 5.70043564e-01 1.92198485e-01
2.88319707e-01 -1.77712038e-01 -1.59885719e-01 -8.80838275e-01
1.69374391e-01 -8.29201996e-01 2.23993018e-01 -6.99213445e-01
-1.12997723e+00 4.48850363e-01 1.35906443e-01 -1.44014204e+00
-3.20842909e-03 -2.58567274e-01 -6.89706564e-01 9.71144497e-01
-1.71034718e+00 -9.61906731e-01 -1.36374727e-01 7.67980456e-01
2.60148734e-01 5.42647615e-02 5.86916149e-01 1.47013202e-01
-3.17863762e-01 6.82599187e-01 5.57423867e-02 3.16522419e-01
7.09137738e-01 -9.43387091e-01 4.77964520e-01 9.54439521e-01
1.73312500e-01 7.84808993e-01 5.12258351e-01 -4.36641842e-01
-1.12154281e+00 -1.10211444e+00 9.42105889e-01 -2.42489688e-02
2.46553108e-01 -2.15234116e-01 -8.72494757e-01 8.27804863e-01
5.32180488e-01 4.15734313e-02 4.63055938e-01 -5.59627593e-01
-3.71684223e-01 -5.10117225e-02 -1.15995133e+00 9.01642680e-01
1.10446739e+00 -4.54160988e-01 -2.95757234e-01 2.52115518e-01
7.59276509e-01 -3.03804338e-01 -4.61263299e-01 1.19844504e-01
4.66155171e-01 -8.76739383e-01 9.50772762e-01 -4.95973855e-01
7.86054373e-01 -5.44095278e-01 5.93111617e-03 -1.54033148e+00
-3.17254156e-01 -7.66031742e-01 3.53232831e-01 1.01319957e+00
4.86082464e-01 -8.60280335e-01 8.98288131e-01 3.21929812e-01
2.27824017e-01 -4.50912386e-01 -8.31696033e-01 -1.02402771e+00
2.13395536e-01 -3.60780279e-04 8.72175217e-01 8.68863523e-01
-1.22780956e-01 1.71864435e-01 -5.31927228e-01 3.15214604e-01
5.17392874e-01 3.24711233e-01 6.82504237e-01 -8.23285222e-01
-2.99298465e-01 -3.94952416e-01 -1.35750145e-01 -1.43095779e+00
4.43018019e-01 -9.81384814e-01 1.95943460e-01 -1.49435914e+00
3.41633171e-01 -6.33230209e-01 -2.98467159e-01 4.57916349e-01
-1.70915723e-02 4.22433257e-01 4.75134850e-01 3.70322496e-01
-2.94752091e-01 3.24266136e-01 1.83982193e+00 -6.24367967e-02
-5.92353605e-02 -1.02610476e-02 -1.06448877e+00 4.22885984e-01
1.05133259e+00 -6.56854987e-01 -8.47416282e-01 -7.42963076e-01
2.40001351e-01 4.54403050e-02 2.94086277e-01 -6.49972320e-01
1.65609106e-01 -5.02565145e-01 2.80165911e-01 -1.88962713e-01
1.02524109e-01 -1.07802272e+00 2.77018338e-01 5.32291651e-01
-6.21000051e-01 3.68857086e-02 -1.09930977e-01 5.85539162e-01
-1.46781221e-01 -3.06673557e-01 9.83704448e-01 -1.52799323e-01
-6.34565115e-01 2.95524538e-01 -1.25067443e-01 -1.03603519e-01
1.36146653e+00 -4.03499871e-01 -3.53971511e-01 -6.01276636e-01
-5.15517414e-01 1.17914930e-01 8.61671150e-01 5.06808162e-01
7.49817848e-01 -1.51701915e+00 -6.09770238e-01 4.49794441e-01
1.95341762e-02 -1.24691330e-01 1.95949376e-02 6.66012049e-01
-4.89906728e-01 4.06821460e-01 -6.18884325e-01 -3.04483861e-01
-1.17165244e+00 6.82139456e-01 3.78710568e-01 -1.61062166e-01
-6.13888025e-01 5.59898913e-01 5.80191433e-01 -3.73524696e-01
-9.46419835e-02 -1.41448408e-01 -4.59485762e-02 -3.18636298e-01
3.72295469e-01 7.12685138e-02 -3.23985398e-01 -7.68488944e-01
-1.21199347e-01 5.50819039e-01 -1.36110857e-01 -2.66228706e-01
1.03724706e+00 -3.60754639e-01 -3.13654304e-01 -2.37244926e-02
1.50690746e+00 -1.31665885e-01 -1.48634148e+00 -4.12225544e-01
-3.79588008e-01 -8.85503113e-01 -3.57574195e-01 -3.87582600e-01
-1.35752809e+00 6.55411065e-01 4.06388074e-01 2.00644433e-01
1.35985947e+00 3.11851054e-02 8.36926699e-01 1.80930674e-01
2.50740886e-01 -9.29306567e-01 2.92953670e-01 3.05716842e-01
6.30046308e-01 -9.46170568e-01 -2.47838214e-01 -5.12324750e-01
-7.19775259e-01 1.01566744e+00 7.33621120e-01 -2.67680943e-01
3.47493529e-01 -3.17527987e-02 -8.27736259e-02 1.30994424e-01
-5.78919947e-01 -7.11431727e-02 1.60210475e-01 8.99135172e-01
1.50943652e-01 1.70205776e-02 -2.28062376e-01 -1.03751607e-01
-5.92364222e-02 2.64958084e-01 7.93353915e-01 8.19984257e-01
-2.80993670e-01 -1.47439098e+00 -3.79605651e-01 1.67840093e-01
-3.24635118e-01 6.78464696e-02 -4.13423359e-01 6.05184495e-01
2.21947938e-01 7.83672452e-01 2.62110353e-01 -2.17727289e-01
2.23679721e-01 -2.30702385e-01 6.96122110e-01 -6.06806457e-01
-4.88664471e-02 -1.22429803e-01 -4.94357914e-01 -4.08092886e-01
-4.52113211e-01 -4.48614329e-01 -1.22044480e+00 -4.20753747e-01
-9.24198478e-02 -1.01435602e-01 3.66905719e-01 7.80415237e-01
4.09904808e-01 3.17674130e-01 1.06333709e+00 -4.82025981e-01
-3.67207915e-01 -3.48146379e-01 -6.11736178e-01 5.16741395e-01
3.80993783e-01 -3.16615522e-01 -9.93636325e-02 3.72680217e-01] | [11.648972511291504, -0.39622703194618225] |
277a2318-4d58-4c05-b049-d5816c97fe24 | distributional-reinforcement-learning-with | 1902.03149 | null | http://arxiv.org/abs/1902.03149v1 | http://arxiv.org/pdf/1902.03149v1.pdf | Distributional reinforcement learning with linear function approximation | Despite many algorithmic advances, our theoretical understanding of practical
distributional reinforcement learning methods remains limited. One exception is
Rowland et al. (2018)'s analysis of the C51 algorithm in terms of the Cram\'er
distance, but their results only apply to the tabular setting and ignore C51's
use of a softmax to produce normalized distributions. In this paper we adapt
the Cram\'er distance to deal with arbitrary vectors. From it we derive a new
distributional algorithm which is fully Cram\'er-based and can be combined to
linear function approximation, with formal guarantees in the context of policy
evaluation. In allowing the model's prediction to be any real vector, we lose
the probabilistic interpretation behind the method, but otherwise maintain the
appealing properties of distributional approaches. To the best of our
knowledge, ours is the first proof of convergence of a distributional algorithm
combined with function approximation. Perhaps surprisingly, our results provide
evidence that Cram\'er-based distributional methods may perform worse than
directly approximating the value function. | ['Subhodeep Moitra', 'Marc G. Bellemare', 'Pablo Samuel Castro', 'Nicolas Le Roux'] | 2019-02-08 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-2.20462531e-02 3.36466521e-01 -4.59073722e-01 -3.12975556e-01
-9.06881511e-01 -8.72095048e-01 6.07536435e-01 2.91970015e-01
-7.45448291e-01 1.05243146e+00 1.02596499e-01 -9.27808583e-01
-4.35225576e-01 -7.33124197e-01 -7.71707475e-01 -8.91819358e-01
-4.22350541e-02 6.84293628e-01 6.26695305e-02 -2.31936976e-01
2.00172514e-01 2.23412097e-01 -1.39428723e+00 -2.42543686e-02
6.12647355e-01 9.59721088e-01 -1.86866522e-03 7.19284892e-01
5.40100746e-02 7.90081084e-01 -6.66555464e-01 -3.71338904e-01
2.59441286e-01 -3.56458724e-01 -6.88745320e-01 -4.82677311e-01
1.74907476e-01 -5.24545372e-01 -2.28508413e-01 1.18388748e+00
5.64222097e-01 2.64852941e-01 9.20685828e-01 -1.36303246e+00
-7.50267863e-01 1.01510835e+00 -4.96951520e-01 3.04988977e-02
2.28536963e-01 -1.19997635e-02 1.26892948e+00 -2.43313879e-01
4.67592955e-01 1.28536892e+00 6.97446704e-01 4.65468287e-01
-1.36771560e+00 -4.35989708e-01 2.24375233e-01 -5.68457991e-02
-1.12125242e+00 -5.58614358e-02 4.64117348e-01 -4.50737298e-01
7.40466595e-01 1.60208866e-01 3.06290776e-01 9.95497942e-01
-1.78340733e-01 1.13836992e+00 1.30172169e+00 -5.63871503e-01
5.06548703e-01 1.31844327e-01 9.95839834e-02 3.20204616e-01
2.68762469e-01 3.95026296e-01 -1.00862838e-01 -5.39994121e-01
4.89697307e-01 -1.73594385e-01 -1.43123507e-01 -4.54578608e-01
-7.34391689e-01 1.25491333e+00 4.18392941e-02 -1.50575433e-02
-2.13023961e-01 6.80954099e-01 5.77607989e-01 3.35462064e-01
5.26495814e-01 1.58898681e-01 -5.79002082e-01 -6.43664956e-01
-8.44500840e-01 7.18321085e-01 8.68849456e-01 9.66620922e-01
2.99674749e-01 1.51833221e-01 -1.54835552e-01 6.08395278e-01
2.68848360e-01 5.97714424e-01 4.28986996e-01 -1.27506685e+00
1.80663198e-01 -2.66100734e-01 4.95134681e-01 -5.09798169e-01
-3.42273831e-01 -4.14685667e-01 -3.71022075e-01 2.57269949e-01
9.61490154e-01 -4.30671036e-01 -4.05880034e-01 2.09717751e+00
1.43356502e-01 -7.11315200e-02 -5.38587719e-02 7.59816289e-01
-2.37462320e-03 4.83769834e-01 9.78280157e-02 -3.54381263e-01
9.16394055e-01 -3.82596910e-01 -4.78338212e-01 2.79230952e-01
8.51722121e-01 -6.08980477e-01 1.29348385e+00 6.57359302e-01
-1.00386238e+00 -1.40980676e-01 -8.26521397e-01 1.95488185e-01
-1.73246264e-01 -2.57794112e-01 9.43555653e-01 9.49326873e-01
-1.06080770e+00 7.12844551e-01 -7.12104678e-01 -1.71624586e-01
4.50141609e-01 2.19586253e-01 1.09099984e-01 9.68115255e-02
-1.19491458e+00 1.03142250e+00 2.37679988e-01 -4.45446074e-01
-8.04611206e-01 -7.25132227e-01 -6.99350774e-01 -2.89711505e-02
5.62938333e-01 -5.61546922e-01 1.87226462e+00 -9.89384353e-01
-1.66286981e+00 2.82224178e-01 -9.84921828e-02 -9.00865018e-01
5.85010052e-01 -1.32264733e-01 1.37404382e-01 -3.78392972e-02
-3.97858359e-02 1.51633427e-01 4.31463718e-01 -1.08212328e+00
-5.81549525e-01 -2.75336385e-01 3.31705570e-01 8.75332505e-02
-6.64336607e-02 -8.33541974e-02 3.40382695e-01 -6.98162675e-01
-5.72195649e-01 -7.60244608e-01 -2.18040362e-01 -2.20619485e-01
-1.56647310e-01 -6.67219222e-01 3.03755432e-01 -1.30097553e-01
1.20991635e+00 -2.09287715e+00 -1.48892000e-01 3.24925542e-01
2.18318537e-01 7.20689595e-02 -1.35993704e-01 6.37931168e-01
-5.17946742e-02 1.99287102e-01 -3.79738271e-01 -2.92758673e-01
8.25365722e-01 4.17640179e-01 -6.82371020e-01 8.56292844e-01
-3.08401972e-01 8.30660939e-01 -9.92707312e-01 -2.40330085e-01
1.67956993e-01 2.25161240e-01 -8.30116272e-01 -1.45631552e-01
-5.54147601e-01 -2.95269251e-01 -2.78967053e-01 8.22159201e-02
5.36357582e-01 1.84079513e-01 2.79206514e-01 2.74332196e-01
-2.26681367e-01 4.00250852e-01 -1.08098745e+00 1.38509142e+00
-4.00769800e-01 4.87768114e-01 1.80586904e-01 -1.32000566e+00
5.01519620e-01 -1.07133444e-02 5.87243259e-01 -3.78348231e-01
2.11399481e-01 2.26383418e-01 7.49801174e-02 -2.02427283e-01
5.62702060e-01 -6.61027610e-01 -2.22368822e-01 7.76230514e-01
-1.00058675e-01 -9.48249102e-02 1.36681378e-01 2.06905618e-01
8.62181127e-01 5.21096289e-01 2.18909636e-01 -6.29139900e-01
-1.93926096e-01 -8.61260667e-02 2.62064695e-01 1.08497918e+00
-2.79921144e-01 2.47243941e-01 9.48368549e-01 -5.82830645e-02
-1.10763800e+00 -1.19033909e+00 -3.75298381e-01 1.52137244e+00
-3.28164935e-01 -4.81928349e-01 -7.07818925e-01 -7.09412217e-01
5.15304923e-01 1.20564044e+00 -8.97938132e-01 -1.42577663e-01
-9.35181826e-02 -7.46449888e-01 1.01798856e+00 7.03945816e-01
-1.99075595e-01 -6.18656456e-01 -7.40518272e-01 1.21543422e-01
2.49916628e-01 -4.39364582e-01 -6.02992237e-01 5.46024263e-01
-5.99681258e-01 -9.69673812e-01 -5.15419304e-01 -2.58424580e-01
7.61677027e-02 -2.92479098e-01 9.89646494e-01 -3.70190501e-01
9.67334062e-02 5.89325368e-01 -2.34233826e-01 -7.58026719e-01
-4.34425026e-01 -1.45415261e-01 1.49853125e-01 -4.52174485e-01
5.18083274e-01 -7.03131139e-01 -5.62084317e-01 -6.06946871e-02
-9.82404470e-01 -6.43946350e-01 3.34617466e-01 9.87800717e-01
5.12184620e-01 1.89163357e-01 7.46964097e-01 -1.09777534e+00
9.90156770e-01 -7.53574133e-01 -8.50520134e-01 -3.05270795e-02
-8.48482609e-01 4.93990630e-01 9.04322624e-01 -5.70203900e-01
-7.86059678e-01 -2.80920058e-01 -2.87451625e-01 -3.80211800e-01
2.32925382e-03 5.27334034e-01 -1.85388476e-02 4.46214527e-01
7.14852870e-01 3.00128311e-01 2.35636085e-01 -3.66560876e-01
7.37533271e-01 5.96866488e-01 4.13613588e-01 -1.27279961e+00
5.51774681e-01 2.06509247e-01 4.86096628e-02 -6.28356934e-01
-8.41527998e-01 -1.93389133e-01 -1.11765591e-02 1.31899327e-01
4.58533913e-01 -5.79100013e-01 -1.28202271e+00 9.29588675e-02
-7.27461159e-01 -7.77839601e-01 -7.68307805e-01 5.56179345e-01
-1.12325907e+00 4.93565798e-01 -4.22687143e-01 -1.33201885e+00
1.55863568e-01 -9.93151367e-01 8.30876768e-01 -1.21889800e-01
-2.46456638e-01 -1.10045207e+00 2.86620110e-01 -1.70959502e-01
4.72727209e-01 5.14686331e-02 9.81235743e-01 -8.51774991e-01
2.69463420e-01 1.50014758e-01 -5.36409067e-03 4.98420775e-01
-2.52902120e-01 -1.24224566e-03 -8.90537679e-01 -1.89036638e-01
8.85840207e-02 -3.32292587e-01 8.41809809e-01 6.95984900e-01
1.33026743e+00 -6.35358930e-01 1.45257369e-01 3.82367581e-01
1.37837839e+00 1.85760289e-01 5.01602590e-01 3.97338390e-01
3.02643120e-01 4.58000213e-01 5.62880278e-01 8.17466021e-01
4.64765728e-01 4.30152148e-01 4.44332093e-01 3.79678965e-01
1.82412237e-01 -6.52334690e-01 5.74567556e-01 2.54517078e-01
9.33820233e-02 -2.13445425e-01 -6.57554507e-01 5.24763405e-01
-1.88354027e+00 -1.25629473e+00 8.67737010e-02 2.47918940e+00
1.15766120e+00 2.65424669e-01 7.27896571e-01 -1.33169284e-02
5.31512856e-01 1.98568720e-02 -6.90830588e-01 -8.64408195e-01
3.85412984e-02 5.95992804e-01 9.78446305e-01 7.97207832e-01
-8.34146917e-01 6.55499756e-01 6.85080719e+00 1.23783576e+00
-7.58123517e-01 1.23365909e-01 3.88937354e-01 -4.01134998e-01
-7.00718522e-01 8.00210238e-02 -4.84824926e-01 6.80605352e-01
1.15555108e+00 -2.38014504e-01 6.86515212e-01 1.09889281e+00
8.77718776e-02 -1.95107773e-01 -1.13819218e+00 8.56020272e-01
-2.23557189e-01 -8.49897265e-01 -3.17285508e-01 2.64492035e-01
5.15848100e-01 3.91321704e-02 4.16728586e-01 5.61359346e-01
1.02975214e+00 -1.42098594e+00 9.33579504e-01 4.20759201e-01
6.85256839e-01 -1.32808399e+00 6.80109799e-01 4.69986767e-01
-5.84247530e-01 -1.25075340e-01 -4.44315851e-01 -3.42089176e-01
-6.03287853e-02 6.62876606e-01 -5.34282148e-01 2.66487926e-01
1.76307097e-01 4.42769170e-01 4.19607423e-02 6.96563482e-01
-4.55603972e-02 1.11665714e+00 -6.88133955e-01 -2.86793828e-01
5.95546663e-01 -2.31301114e-01 4.71523613e-01 1.15783131e+00
3.13534707e-01 -5.39177246e-02 3.20301056e-01 7.53077805e-01
8.98495317e-02 6.45431131e-02 -7.87203670e-01 -2.12750226e-01
5.93050063e-01 6.39533997e-01 -1.50712639e-01 -1.67665020e-01
-2.98629016e-01 4.39122915e-01 3.63880187e-01 3.61011565e-01
-9.62557673e-01 -4.42819923e-01 7.60988355e-01 6.86683580e-02
4.58180845e-01 -2.41485387e-01 -2.35133350e-01 -7.65214503e-01
9.04253647e-02 -8.36286068e-01 4.86403555e-01 -4.60882902e-01
-1.43697572e+00 -2.90184915e-02 3.73397350e-01 -7.86083162e-01
-5.02823710e-01 -8.09553921e-01 -3.39803010e-01 7.55947590e-01
-1.30339754e+00 -7.55291224e-01 8.01535845e-01 5.78065991e-01
1.92214534e-01 -2.82118414e-02 7.39824116e-01 -9.20371115e-02
-2.11624146e-01 8.34520280e-01 7.65492857e-01 -1.02731109e-01
6.90818787e-01 -1.85457933e+00 -2.58336842e-01 4.70334500e-01
-8.41607898e-03 6.79005027e-01 1.13393116e+00 -2.16832414e-01
-1.54620790e+00 -8.53740335e-01 4.03357804e-01 -5.23705781e-01
1.06266379e+00 -9.93950740e-02 -4.01782781e-01 8.48927796e-01
1.52987748e-01 -2.24930182e-01 7.92985260e-01 4.22617495e-01
-6.04166627e-01 4.35299575e-02 -1.20637977e+00 5.70188999e-01
7.23699331e-01 -4.89964187e-01 -5.70298851e-01 1.91681713e-01
5.94865143e-01 -2.41209954e-01 -1.09593618e+00 1.36029556e-01
6.81783020e-01 -8.93337905e-01 7.95466602e-01 -8.27508807e-01
3.05904955e-01 -2.98245817e-01 -6.62005424e-01 -1.47165740e+00
7.14799538e-02 -8.04857790e-01 -2.57178336e-01 1.04130733e+00
1.41283184e-01 -9.36714590e-01 3.82152677e-01 4.58940536e-01
-1.42223909e-01 -9.28248882e-01 -1.07387078e+00 -1.05638874e+00
1.06533718e+00 -9.91123438e-01 5.91307282e-01 7.58458197e-01
3.88263494e-01 6.24723285e-02 -2.95304447e-01 -1.94137484e-01
7.47507751e-01 1.36180356e-01 5.53184986e-01 -8.69100869e-01
-8.84641409e-01 -1.06037307e+00 -1.78848699e-01 -1.27153015e+00
4.72036898e-01 -9.66748059e-01 9.34858322e-02 -1.09026110e+00
1.84145868e-01 -6.35337234e-01 -3.62695456e-01 4.58342254e-01
6.96040690e-02 -9.63673219e-02 1.73608959e-01 -9.74780172e-02
-5.48258245e-01 6.49665117e-01 9.70829368e-01 4.30225953e-02
9.85395610e-02 9.51057076e-02 -1.13235414e+00 6.58594012e-01
1.12824297e+00 -5.33653438e-01 -8.14865351e-01 -1.26234174e-01
3.63395691e-01 -4.76994514e-02 4.33419079e-01 -3.35891515e-01
-3.03519946e-02 -4.39765543e-01 1.10589445e-01 -1.36829838e-01
-1.28041327e-01 -4.88741308e-01 -1.88291952e-01 4.00470376e-01
-6.86764240e-01 8.47732723e-02 2.73926020e-01 6.15185857e-01
1.37986600e-01 -3.43401492e-01 8.06873262e-01 8.41905847e-02
-5.13531156e-02 1.55814365e-01 -5.66785872e-01 4.74166483e-01
8.86830211e-01 1.10796414e-01 -2.88925052e-01 -6.42878771e-01
-6.16453350e-01 1.36724263e-01 6.24823749e-01 -1.88977361e-01
2.81023681e-01 -1.25186324e+00 -7.56238639e-01 -2.67804563e-01
-1.75469741e-01 -1.29424036e-01 -5.76659739e-02 9.44239318e-01
-1.04679249e-01 3.97018909e-01 2.94716954e-01 -2.71171093e-01
-6.91257119e-01 8.67994905e-01 2.70313144e-01 -2.55995899e-01
-4.36771721e-01 4.39235300e-01 2.41088197e-01 -3.55489105e-01
2.84999400e-01 -4.39780682e-01 2.71470368e-01 5.06596733e-03
4.80058044e-01 3.96463692e-01 -3.85947764e-01 -2.60922551e-01
-2.69865930e-01 1.54815227e-01 9.76407677e-02 -6.43741250e-01
1.37140882e+00 8.19770619e-02 5.21603636e-02 7.37118125e-01
1.08957326e+00 2.53398806e-01 -1.33561373e+00 -1.10947028e-01
-2.04958469e-01 -3.10201943e-01 -1.32555300e-02 -9.51973379e-01
-8.38987648e-01 8.05877984e-01 3.35846573e-01 3.52794141e-01
1.00818241e+00 1.74704622e-02 5.38097680e-01 4.04154867e-01
4.31656778e-01 -1.13257563e+00 -5.14127672e-01 5.21916807e-01
5.66412687e-01 -7.84927487e-01 -3.62315625e-02 3.43259156e-01
-7.33998835e-01 1.01923585e+00 1.08448818e-01 -2.37211943e-01
6.54869795e-01 5.07439435e-01 -2.87268430e-01 1.76916018e-01
-8.80647182e-01 -4.03280467e-01 -1.69421837e-01 8.65895689e-01
4.50349867e-01 3.66821587e-01 -5.44161618e-01 8.87056172e-01
-6.19822383e-01 -8.62002075e-02 5.78866124e-01 8.02969515e-01
-5.41751623e-01 -1.32026339e+00 -2.08162770e-01 4.99532223e-01
-6.68108821e-01 -2.76118040e-01 -3.40697356e-02 1.04601002e+00
-2.09602937e-01 9.19880331e-01 -1.16197921e-01 -1.27315313e-01
9.68366414e-02 3.27216387e-01 7.77499914e-01 -3.14262241e-01
-2.56578445e-01 1.06632961e-02 5.53612523e-02 -4.15797621e-01
-2.95777529e-01 -9.05269206e-01 -1.26540697e+00 -7.92575479e-01
-1.86568685e-02 4.65670615e-01 4.88608599e-01 8.84790659e-01
7.58702978e-02 1.72310501e-01 4.84582156e-01 -3.46324980e-01
-1.50014746e+00 -8.96721363e-01 -9.20312762e-01 3.37528914e-01
3.13751429e-01 -6.90648079e-01 -4.46301788e-01 -3.21233422e-01] | [4.099818706512451, 2.6142802238464355] |
1d10f85f-b36f-45f1-a7fb-c280ff6e5518 | branchconnect-large-scale-visual-recognition | 1704.06010 | null | http://arxiv.org/abs/1704.06010v3 | http://arxiv.org/pdf/1704.06010v3.pdf | BranchConnect: Large-Scale Visual Recognition with Learned Branch Connections | We introduce an architecture for large-scale image categorization that
enables the end-to-end learning of separate visual features for the different
classes to distinguish. The proposed model consists of a deep CNN shaped like a
tree. The stem of the tree includes a sequence of convolutional layers common
to all classes. The stem then splits into multiple branches implementing
parallel feature extractors, which are ultimately connected to the final
classification layer via learned gated connections. These learned gates
determine for each individual class the subset of features to use. Such a
scheme naturally encourages the learning of a heterogeneous set of specialized
features through the separate branches and it allows each class to use the
subset of features that are optimal for its recognition. We show the generality
of our proposed method by reshaping several popular CNNs from the literature
into our proposed architecture. Our experiments on the CIFAR100, CIFAR10, and
Synth datasets show that in each case our resulting model yields a substantial
improvement in accuracy over the original CNN. Our empirical analysis also
suggests that our scheme acts as a form of beneficial regularization improving
generalization performance. | ['Karim Ahmed', 'Lorenzo Torresani'] | 2017-04-20 | null | null | null | null | ['image-categorization'] | ['computer-vision'] | [ 5.87268770e-02 2.18699947e-01 -2.65118271e-01 -6.49849772e-01
-4.40863788e-01 -6.53090000e-01 4.31109309e-01 2.26601288e-01
-5.07866144e-01 4.48032320e-01 -1.46203250e-01 -2.53735960e-01
-2.25131400e-02 -8.64281774e-01 -6.96493208e-01 -8.50730062e-01
-3.37315381e-01 1.81473032e-01 3.80006224e-01 -6.85303509e-02
2.77794510e-01 7.67515600e-01 -1.60525405e+00 6.24328613e-01
5.07081687e-01 1.63401020e+00 1.13442309e-01 6.05264604e-01
6.13590644e-04 9.59279716e-01 -5.71924686e-01 -1.41601339e-01
5.54780900e-01 -2.48216540e-01 -9.74485159e-01 2.88659066e-01
6.99236989e-01 -1.03727624e-01 -2.70413935e-01 7.69789934e-01
2.39425063e-01 1.30688980e-01 6.63368881e-01 -1.14383590e+00
-8.16534817e-01 6.70145273e-01 -3.72622669e-01 3.10772777e-01
-2.49652758e-01 -9.81681719e-02 1.33930600e+00 -7.39581287e-01
4.83115017e-01 1.09528553e+00 5.75541556e-01 5.23290336e-01
-1.36442113e+00 -6.66188955e-01 3.72817248e-01 -4.28563841e-02
-1.26793838e+00 -2.85273075e-01 6.86720788e-01 -4.91176754e-01
9.21442091e-01 1.92164779e-01 5.22568226e-01 8.45199347e-01
5.05065136e-02 7.77645946e-01 8.49617720e-01 -3.79982829e-01
2.90849477e-01 3.55078846e-01 6.99285567e-01 8.37005138e-01
2.45851725e-01 1.41508514e-02 -1.10839024e-01 -1.31208137e-01
6.64284229e-01 2.42901444e-01 -1.80719569e-01 -6.24562562e-01
-8.60490561e-01 9.29336846e-01 1.17168379e+00 3.18020970e-01
-2.33085245e-01 2.12038010e-01 4.29696411e-01 5.20408630e-01
2.36779496e-01 4.21145618e-01 -4.58890080e-01 7.33258665e-01
-7.65434265e-01 9.51727033e-02 5.75827062e-01 7.48382986e-01
1.08015418e+00 -2.80236248e-02 -4.30346042e-01 9.27662969e-01
1.07237712e-01 -5.68166338e-02 6.58367574e-01 -7.33253121e-01
3.08399528e-01 9.94830251e-01 -1.45060614e-01 -8.28390777e-01
-4.60737497e-01 -9.07826841e-01 -8.50421429e-01 4.53016132e-01
4.03342783e-01 1.19547583e-02 -1.17585218e+00 1.79844916e+00
8.82155541e-03 -3.56590077e-02 5.81990443e-02 8.25233042e-01
8.19479644e-01 4.18728769e-01 3.19946975e-01 3.76212358e-01
1.33880413e+00 -1.11382592e+00 -6.44795820e-02 -2.49271303e-01
6.28423035e-01 -4.28533971e-01 1.01361191e+00 3.79225045e-01
-7.42617071e-01 -8.55054617e-01 -1.21083736e+00 -4.05874178e-02
-5.85391104e-01 4.63284910e-01 7.10862696e-01 3.86810362e-01
-1.23352706e+00 6.78195894e-01 -5.87337792e-01 -3.32963139e-01
8.76526296e-01 6.53424025e-01 -4.63623762e-01 1.28205419e-01
-7.40597010e-01 5.25992930e-01 4.91578937e-01 -6.98688533e-03
-1.10175371e+00 -3.25861543e-01 -7.92423546e-01 4.64761913e-01
-6.57486171e-02 -6.04561627e-01 1.09006071e+00 -1.71039903e+00
-1.19090474e+00 1.11655080e+00 4.18476313e-02 -5.49162149e-01
2.80003101e-01 6.24611638e-02 1.03200721e-02 3.04444760e-01
1.18568717e-02 9.65637982e-01 1.14070261e+00 -1.39053178e+00
-9.82082486e-01 -3.05254608e-01 3.65328163e-01 -7.10937604e-02
-6.61013663e-01 -4.36760932e-02 -2.50508189e-01 -6.34174705e-01
6.56543970e-02 -8.86534214e-01 -2.87725747e-01 1.15253543e-02
-6.36376262e-01 -3.24866623e-01 7.46432722e-01 -1.36064580e-02
1.04412878e+00 -2.20883012e+00 1.72494397e-01 4.15315181e-01
4.86667275e-01 5.70406467e-02 -3.25649768e-01 2.58803934e-01
-3.46851230e-01 1.59952238e-01 -1.67277426e-01 -2.40693346e-01
-3.16505253e-01 1.27879322e-01 -3.00694078e-01 4.63485688e-01
4.05397475e-01 8.47474337e-01 -5.56066990e-01 -6.53596893e-02
4.56736833e-02 3.13709527e-01 -5.79591632e-01 4.03822869e-01
8.20892453e-02 1.64662540e-01 -6.08445644e-01 4.25502777e-01
5.65647542e-01 -5.46002328e-01 1.23503856e-01 -3.21519077e-02
-1.42673433e-01 1.76394433e-01 -1.00247598e+00 1.27146971e+00
-4.16946620e-01 5.52918196e-01 8.19320828e-02 -1.41045213e+00
1.02874804e+00 4.37895581e-02 1.05771469e-02 -3.67401004e-01
3.98020983e-01 8.44102502e-02 2.23882005e-01 -2.08694384e-01
2.58869559e-01 -1.88534930e-02 -3.38535719e-02 2.66750544e-01
4.24460143e-01 2.79914469e-01 1.51377544e-01 1.01185329e-01
8.84460449e-01 -2.84424633e-01 3.41461390e-01 -5.96498132e-01
6.57423019e-01 -1.75795536e-02 3.88043851e-01 9.44291353e-01
-2.47347370e-01 5.51945806e-01 4.61818099e-01 -9.20638263e-01
-9.41006124e-01 -8.23628664e-01 -2.72747785e-01 1.47846627e+00
-6.34955019e-02 -1.75009161e-01 -5.42200923e-01 -8.49124372e-01
1.71258062e-01 2.51461208e-01 -1.14417696e+00 -2.58049071e-01
-4.73391712e-01 -7.02329159e-01 3.45121205e-01 7.44338989e-01
4.33884650e-01 -1.24378371e+00 -8.28076899e-01 -2.34212596e-02
2.81239271e-01 -9.02968585e-01 -1.10957257e-01 9.16504562e-01
-1.00676978e+00 -1.23870933e+00 -5.13026953e-01 -1.30324018e+00
8.45841825e-01 3.19339037e-01 1.03115511e+00 3.29764605e-01
-1.25367761e-01 2.35090833e-02 -4.10394400e-01 -2.09839657e-01
-2.89234877e-01 3.92119437e-01 -2.82894224e-01 4.55218673e-01
1.99273035e-01 -5.05443871e-01 -7.18534827e-01 2.10465252e-01
-7.45580018e-01 -2.58865029e-01 5.01346290e-01 9.66054440e-01
5.13441563e-01 -2.32875124e-01 5.81442595e-01 -9.83870447e-01
3.86255294e-01 -5.54123402e-01 -5.54908454e-01 1.53970465e-01
-2.52173126e-01 2.18429700e-01 1.08695018e+00 -4.42542046e-01
-5.98607481e-01 3.12091678e-01 7.92372674e-02 -2.99993247e-01
-3.40287536e-01 2.60049909e-01 1.14609875e-01 -3.69785935e-01
7.39886105e-01 1.95150346e-01 -3.15203995e-01 -4.93456692e-01
4.10720795e-01 8.06410670e-01 3.87460083e-01 -4.17076796e-01
5.84482431e-01 5.20021141e-01 -1.88816994e-01 -6.70452833e-01
-9.06198859e-01 -3.36173058e-01 -7.94790506e-01 1.26915276e-01
9.06662822e-01 -9.56431329e-01 -6.78538978e-01 4.35070723e-01
-9.19511437e-01 -4.28998262e-01 -3.83722872e-01 1.00977272e-01
-3.19459558e-01 8.47239569e-02 -7.27149844e-01 -5.05802929e-01
-3.64426553e-01 -1.28605473e+00 9.22165453e-01 5.25874853e-01
-1.57534927e-02 -9.19014335e-01 -1.46843329e-01 9.96200368e-03
3.44535381e-01 4.30583991e-02 1.18055892e+00 -1.05930603e+00
-4.94473219e-01 -3.47991168e-01 -2.92536467e-01 6.06125534e-01
7.67937005e-02 2.67616436e-02 -1.41382802e+00 -5.59417248e-01
-2.33521521e-01 -7.93743134e-01 1.46759522e+00 3.61030072e-01
1.62802052e+00 -2.01221108e-02 -4.63473380e-01 9.43138421e-01
1.56833100e+00 2.46881008e-01 4.51505601e-01 5.62178552e-01
6.70102835e-01 3.83421659e-01 1.01610415e-01 1.48475751e-01
9.79921371e-02 3.73857617e-01 6.25408173e-01 -4.17328656e-01
4.51937132e-02 1.19002707e-01 1.17566600e-01 3.94884735e-01
-4.55509461e-02 -2.21181467e-01 -7.06933737e-01 6.10823870e-01
-1.61052489e+00 -7.52653539e-01 2.67759264e-01 2.21728873e+00
5.70390046e-01 1.66914478e-01 1.18034258e-01 1.66043341e-01
7.08478391e-01 1.78709000e-01 -4.20234889e-01 -6.28068209e-01
-8.21277499e-02 5.70290267e-01 4.11629647e-01 3.01040977e-01
-1.45637453e+00 9.67338264e-01 6.65450525e+00 7.72873461e-01
-1.48804224e+00 -2.62738734e-01 1.03295588e+00 2.75895238e-01
7.21325874e-02 -1.41123876e-01 -7.93215871e-01 -7.42036924e-02
7.43344367e-01 1.67672619e-01 1.46723181e-01 1.05686235e+00
-1.57203004e-01 1.50416836e-01 -1.24139249e+00 7.07826316e-01
-2.16711536e-01 -1.38158619e+00 4.08032209e-01 -1.49347410e-01
6.02230430e-01 1.46419838e-01 9.95352864e-02 2.48932257e-01
4.61705983e-01 -1.15872467e+00 8.77204239e-01 -6.89548180e-02
7.12015212e-01 -8.32525253e-01 5.10326684e-01 1.96789280e-01
-1.23900068e+00 -7.39337623e-01 -6.80628657e-01 -1.67442262e-01
-5.82470357e-01 2.18088716e-01 -6.56377137e-01 3.31925780e-01
6.62804663e-01 7.74185598e-01 -7.67269254e-01 1.04682827e+00
-2.04922467e-01 5.64701557e-01 -4.85849596e-04 1.62224278e-01
6.46983802e-01 -1.03183025e-02 1.82865709e-01 1.37817323e+00
6.06835075e-03 -1.98580995e-01 4.45758879e-01 6.45723283e-01
-3.97211343e-01 1.47167325e-01 -6.18138194e-01 1.40814200e-01
2.96893269e-01 1.60503244e+00 -9.60098982e-01 -6.01847172e-01
-5.75999558e-01 9.46952701e-01 1.00398970e+00 4.59785104e-01
-4.69831049e-01 -7.25632250e-01 6.91677153e-01 -8.48056972e-02
7.78305888e-01 1.33986607e-01 -3.67171824e-01 -1.10078704e+00
-1.10208757e-01 -8.38267565e-01 7.56898820e-01 -3.91271532e-01
-1.32957149e+00 1.22237682e+00 -3.92979175e-01 -1.22077489e+00
8.91783834e-02 -9.57566440e-01 -7.32686937e-01 9.26637590e-01
-1.66560280e+00 -1.16167557e+00 -3.50484639e-01 7.14483738e-01
4.95397061e-01 -3.22565407e-01 8.56283426e-01 1.28840609e-02
-6.48604274e-01 7.30122983e-01 1.20591111e-01 4.89868671e-01
4.36192513e-01 -1.23089325e+00 2.29428679e-01 6.83466017e-01
2.32979327e-01 6.42191589e-01 2.68542826e-01 -1.92856237e-01
-7.10399210e-01 -1.26811540e+00 7.41070867e-01 -1.10859431e-01
5.39163589e-01 -6.17760122e-01 -8.78778219e-01 8.48952889e-01
2.25694865e-01 3.12817514e-01 6.16423309e-01 1.14911899e-01
-8.31425726e-01 -3.45029563e-01 -1.09467161e+00 2.39081323e-01
8.14534605e-01 -4.75717604e-01 -6.02879822e-01 9.02100727e-02
4.75048423e-01 3.52522880e-02 -4.74083304e-01 8.61808881e-02
5.95281661e-01 -1.14294600e+00 9.78983819e-01 -1.05632591e+00
5.77490091e-01 -8.15675706e-02 -2.00767264e-01 -1.40477645e+00
-7.11888790e-01 -1.71218276e-01 3.19364339e-01 8.94157946e-01
5.53880453e-01 -6.48269951e-01 6.29357934e-01 2.63526171e-01
-2.04368964e-01 -8.77132118e-01 -6.02978766e-01 -6.48925900e-01
3.13951284e-01 -2.64644057e-01 3.36612880e-01 8.98180127e-01
-1.95912749e-01 4.61396188e-01 -7.87882507e-02 1.05332673e-01
5.27688682e-01 5.97171366e-01 4.20399547e-01 -1.39920139e+00
-3.43909115e-01 -6.07806921e-01 -5.89682639e-01 -1.09877467e+00
1.41515911e-01 -1.27662051e+00 -8.98667797e-02 -1.21108902e+00
5.23598552e-01 -6.38462484e-01 -7.17617571e-01 8.56841326e-01
-1.69663414e-01 4.74326193e-01 1.92050651e-01 2.41910115e-01
-4.74221617e-01 2.51519889e-01 1.16286683e+00 -2.83308744e-01
-1.92124978e-01 1.13724537e-01 -9.69220340e-01 6.98454678e-01
7.45915174e-01 -5.72169960e-01 -3.14026594e-01 -5.07727325e-01
-2.93856204e-01 -4.21406865e-01 5.13544142e-01 -1.01917565e+00
-1.06073126e-01 7.56340101e-02 7.46207058e-01 -2.20679771e-02
-8.85810237e-03 -8.19110215e-01 -4.20766473e-01 6.52789950e-01
-7.93045342e-01 -6.12659305e-02 1.62831903e-01 4.49528813e-01
-3.78260672e-01 -5.98615408e-02 1.23557448e+00 -1.84154987e-01
-7.28284597e-01 3.72683883e-01 -2.38017082e-01 -1.08676046e-01
8.84690762e-01 -2.04122186e-01 -1.60148546e-01 -2.40666837e-01
-9.52849329e-01 1.48968786e-01 2.77378231e-01 3.70466173e-01
3.97706926e-01 -1.21356940e+00 -5.34104407e-01 3.68688464e-01
2.73173124e-01 -1.44027203e-01 -1.56565771e-01 3.87659103e-01
-4.23161060e-01 4.28595126e-01 -5.84792972e-01 -6.86380923e-01
-1.17414010e+00 7.41766632e-01 6.62761927e-01 -3.27452689e-01
-5.17430425e-01 1.06988204e+00 8.21252465e-01 -2.77255684e-01
3.63895059e-01 -6.09139681e-01 -5.33466756e-01 8.42039064e-02
5.68480015e-01 -3.86004299e-02 1.12720214e-01 -6.68323100e-01
-3.47551703e-01 4.71817136e-01 -2.22373858e-01 4.22056049e-01
1.59908545e+00 -3.24440859e-02 -6.77864105e-02 2.91919082e-01
1.51551795e+00 -2.20769569e-01 -1.30743349e+00 -1.35550559e-01
5.35570793e-02 -1.52544096e-01 -5.95592223e-02 -7.05283999e-01
-1.46760046e+00 9.95775580e-01 6.35277152e-01 3.77302676e-01
1.33590865e+00 3.34083401e-02 3.95019412e-01 6.45489752e-01
2.56564975e-01 -7.03253865e-01 -1.15928734e-02 6.41377866e-01
7.20660806e-01 -1.20101082e+00 -4.99996692e-01 -2.31661484e-01
-5.87977469e-01 1.36405289e+00 6.63493216e-01 -7.89997637e-01
5.96551239e-01 1.32253170e-01 2.99705297e-01 -3.19134176e-01
-7.34293342e-01 -2.99148142e-01 3.44262511e-01 4.84821647e-01
5.00332415e-01 -3.49993482e-02 -1.56776011e-01 7.23279715e-01
1.29664540e-01 -1.56022608e-01 3.03811431e-01 8.25524271e-01
-7.10059702e-01 -1.05234730e+00 -1.30798608e-01 4.86395091e-01
-4.53376412e-01 -6.06570877e-02 -4.96070147e-01 9.09313858e-01
2.89305568e-01 6.75873935e-01 3.38245958e-01 -4.06695366e-01
3.23395669e-01 2.33072773e-01 3.32328945e-01 -7.81010687e-01
-1.06615007e+00 -2.95192022e-02 -1.83470517e-01 -5.22355616e-01
-2.85221606e-01 -6.95294365e-02 -1.32009220e+00 1.44235164e-01
-1.32875085e-01 1.39062405e-01 2.15813324e-01 6.92660511e-01
2.67612725e-01 5.47893405e-01 9.46884930e-01 -8.92456412e-01
-6.14236474e-01 -8.97258401e-01 -6.64363563e-01 7.36786604e-01
6.17412150e-01 -6.84152663e-01 -3.30053747e-01 7.47924522e-02] | [9.410053253173828, 2.2724761962890625] |
479f2592-d004-41c2-a898-022dacc2740d | the-genea-challenge-2022-a-large-evaluation | 2208.10441 | null | https://arxiv.org/abs/2208.10441v1 | https://arxiv.org/pdf/2208.10441v1.pdf | The GENEA Challenge 2022: A large evaluation of data-driven co-speech gesture generation | This paper reports on the second GENEA Challenge to benchmark data-driven automatic co-speech gesture generation. Participating teams used the same speech and motion dataset to build gesture-generation systems. Motion generated by all these systems was rendered to video using a standardised visualisation pipeline and evaluated in several large, crowdsourced user studies. Unlike when comparing different research papers, differences in results are here only due to differences between methods, enabling direct comparison between systems. This year's dataset was based on 18 hours of full-body motion capture, including fingers, of different persons engaging in dyadic conversation. Ten teams participated in the challenge across two tiers: full-body and upper-body gesticulation. For each tier we evaluated both the human-likeness of the gesture motion and its appropriateness for the specific speech signal. Our evaluations decouple human-likeness from gesture appropriateness, which previously was a major challenge in the field. The evaluation results are a revolution, and a revelation. Some synthetic conditions are rated as significantly more human-like than human motion capture. To the best of our knowledge, this has never been shown before on a high-fidelity avatar. On the other hand, all synthetic motion is found to be vastly less appropriate for the speech than the original motion-capture recordings. Additional material is available via the project website at https://youngwoo-yoon.github.io/GENEAchallenge2022/ | ['Gustav Eje Henter', 'Mihail Tsakov', 'Teodor Nikolov', 'Carla Viegas', 'Taras Kucherenko', 'Pieter Wolfert', 'Youngwoo Yoon'] | 2022-08-22 | null | null | null | null | ['gesture-generation'] | ['robots'] | [ 2.93783154e-02 1.67163521e-01 2.05131933e-01 -9.84784588e-02
-9.18195069e-01 -7.85296619e-01 1.10727620e+00 -7.27968633e-01
-3.65129471e-01 5.13528943e-01 1.00525737e+00 3.35385948e-02
2.94329256e-01 -5.29037751e-02 -3.54716867e-01 -4.50415283e-01
1.11867920e-01 4.45531845e-01 2.18239039e-01 -3.62983406e-01
-3.85426916e-02 2.69437790e-01 -1.58141541e+00 6.44699931e-01
-4.05980125e-02 3.17828566e-01 8.41100141e-02 1.43792236e+00
2.45782763e-01 3.93757135e-01 -8.97047400e-01 -1.28150597e-01
3.36944669e-01 -8.38774741e-01 -8.67545724e-01 -2.50112284e-02
4.79708612e-01 -5.65588653e-01 -3.98893416e-01 3.81139368e-01
1.27054083e+00 3.01301569e-01 2.99342662e-01 -1.43539453e+00
-2.23669171e-01 2.18559369e-01 7.94911981e-02 4.40810174e-02
1.32100677e+00 8.73662531e-01 8.53446782e-01 -7.85114229e-01
1.22612059e+00 1.41730535e+00 6.88107252e-01 1.19232357e+00
-9.64986086e-01 -4.62030202e-01 -2.74371505e-01 -2.62029856e-01
-1.14541233e+00 -8.33597600e-01 4.14900392e-01 -6.59783244e-01
1.09173584e+00 7.78293550e-01 1.13080597e+00 1.99772537e+00
-4.00444746e-01 8.60307932e-01 9.98202443e-01 -4.28278983e-01
4.34295908e-02 -2.11802930e-01 -4.27269340e-01 3.55992913e-01
-2.04115629e-01 3.45375776e-01 -9.64697003e-01 -1.60746962e-01
1.02427232e+00 -5.49516022e-01 -6.32645190e-01 -2.42645722e-02
-1.86342335e+00 3.61538976e-01 -2.75600940e-01 5.74900270e-01
-1.66903853e-01 4.62575465e-01 3.28741252e-01 2.84057051e-01
1.33285776e-01 2.79329509e-01 -1.42785192e-01 -1.31622505e+00
-8.88999164e-01 8.38818431e-01 1.11901379e+00 1.01474881e+00
-2.45130494e-01 -2.84378808e-02 -4.04642105e-01 4.77552474e-01
3.57223690e-01 5.09508073e-01 7.03375161e-01 -1.16742170e+00
5.15016615e-01 1.99741974e-01 3.63965690e-01 -9.36662614e-01
-5.96069634e-01 3.85654807e-01 -3.42006803e-01 6.12460554e-01
1.10039520e+00 -4.48503941e-01 -7.92122245e-01 1.63440406e+00
3.72112900e-01 -4.50281203e-02 -7.46474341e-02 1.50854647e+00
1.22774637e+00 5.82383394e-01 1.77984871e-03 1.64672248e-02
1.36006999e+00 -9.02434289e-01 -8.53886664e-01 1.13990568e-01
7.24834502e-01 -1.01206970e+00 1.50512397e+00 3.62183630e-01
-1.30027223e+00 -4.40570444e-01 -9.07443643e-01 -2.56294552e-02
4.12964113e-02 -7.72744790e-02 4.84410435e-01 9.11422849e-01
-1.24246609e+00 5.59479058e-01 -8.13751698e-01 -8.26828301e-01
4.13809158e-02 -4.04598266e-02 -6.30526245e-01 3.57721120e-01
-1.10639310e+00 7.81507671e-01 -1.08369939e-01 -6.04744852e-02
-6.69136167e-01 -4.58877981e-01 -6.89910769e-01 -7.08760500e-01
7.43738264e-02 -6.79623902e-01 1.94208395e+00 -8.46058428e-01
-1.90901339e+00 9.33618724e-01 -1.86689183e-01 3.82332429e-02
1.31629455e+00 -3.57769430e-01 -4.55789566e-01 1.19570494e-01
-5.54421321e-02 7.49535322e-01 5.52050114e-01 -1.12532282e+00
-4.64371026e-01 5.51141687e-02 -1.33516729e-01 5.53587437e-01
1.20536342e-01 3.62808317e-01 -5.57057798e-01 -9.60411727e-01
-8.20742473e-02 -1.20774364e+00 1.50830910e-01 -3.78840975e-02
-3.26374173e-01 -4.16988647e-03 8.55890691e-01 -9.29137170e-01
1.17162406e+00 -1.92216825e+00 2.38069057e-01 -1.16329603e-01
1.37833044e-01 2.43318051e-01 1.67280361e-02 7.31647134e-01
9.38750282e-02 4.35597509e-01 -7.73746371e-02 -5.33440590e-01
2.02012256e-01 -8.62491429e-02 -1.02435246e-01 5.19525647e-01
-1.66605160e-01 1.12475765e+00 -1.01676404e+00 -3.39069188e-01
2.65828371e-01 6.44096196e-01 -2.78860092e-01 1.89119369e-01
-1.09982863e-02 1.01401234e+00 -7.03409687e-02 6.08823121e-01
2.04360951e-02 1.79027125e-01 4.85833064e-02 2.63493191e-02
-1.82529911e-01 2.67083496e-01 -1.22834086e+00 2.05610275e+00
-2.04874530e-01 1.16787601e+00 4.82926751e-03 -2.86527481e-02
5.57871222e-01 9.77784634e-01 4.94470388e-01 -5.17459035e-01
-1.66866090e-02 2.70128667e-01 1.21748060e-01 -9.73340929e-01
6.13215446e-01 -2.92969383e-02 -1.95737422e-01 7.12224185e-01
-2.11271062e-01 -5.24680018e-01 4.55091409e-02 3.87437530e-02
1.16688776e+00 6.98058665e-01 2.13884115e-01 1.60014644e-01
-1.52970105e-01 3.45014483e-01 9.29320231e-02 6.79076672e-01
-5.05112290e-01 1.36982155e+00 2.33023465e-01 -3.54505360e-01
-1.22687709e+00 -9.38186109e-01 3.24754864e-01 1.07128131e+00
-1.65503342e-02 -6.59066021e-01 -1.02933335e+00 -2.73470998e-01
-4.23356682e-01 4.57755178e-01 -4.87608790e-01 3.06817830e-01
-8.01717281e-01 -3.78701180e-01 1.27408123e+00 4.78125185e-01
2.19587207e-01 -1.54196024e+00 -1.25313532e+00 -3.42034339e-03
-4.53244209e-01 -1.15626061e+00 -7.59439468e-01 -7.25928605e-01
-4.49355990e-01 -8.74797046e-01 -1.27600574e+00 -4.36742783e-01
1.15617305e-01 9.24210995e-02 1.05124807e+00 2.09235504e-01
-4.18528706e-01 7.29904056e-01 -7.29493022e-01 -3.63037407e-01
-6.24148369e-01 -1.34361774e-01 2.27581799e-01 -2.53240705e-01
1.75415680e-01 -3.99025977e-01 -6.35820985e-01 5.78850985e-01
-5.62750697e-01 3.21060449e-01 1.84743598e-01 6.01022720e-01
-1.97091606e-02 -8.07482243e-01 4.99402508e-02 -1.86975926e-01
1.04959667e+00 -2.33446434e-01 9.36534032e-02 -2.09150761e-01
5.66919392e-04 -3.59148473e-01 1.01361261e-03 -6.63572431e-01
-8.90043736e-01 1.52995855e-01 -1.04258731e-01 -2.58961558e-01
-5.39472103e-01 -9.54323858e-02 -2.73429826e-02 3.39505494e-01
8.67671072e-01 -1.08875759e-01 1.74005374e-01 -3.65688413e-01
4.44766164e-01 8.60392213e-01 8.71699214e-01 -5.02956629e-01
6.43775344e-01 6.17811978e-01 -3.95248473e-01 -1.08847368e+00
2.58466423e-01 -3.38318706e-01 -7.07789600e-01 -7.63760328e-01
1.09691882e+00 -7.18795955e-01 -8.94538820e-01 7.55765378e-01
-1.32189345e+00 -9.85496759e-01 -4.15412873e-01 6.69867694e-01
-8.02823901e-01 2.28804767e-01 -3.89588416e-01 -9.82334197e-01
-1.20086312e-01 -1.16723311e+00 1.25153875e+00 9.79863331e-02
-1.25272453e+00 -7.77574480e-01 3.65105599e-01 4.25133914e-01
1.81691080e-01 7.79586315e-01 -1.35327622e-01 -5.55030525e-01
-7.37777576e-02 -3.66005540e-01 2.09446162e-01 -3.55380386e-01
1.75526753e-01 3.06819886e-01 -1.11952972e+00 -1.70735329e-01
-4.28906679e-01 -3.90438318e-01 2.58678198e-01 1.58049941e-01
6.05140403e-02 -2.81170219e-01 -1.17102847e-01 2.45997071e-01
4.94606674e-01 2.25338340e-01 7.35880256e-01 3.64285707e-01
9.17973161e-01 9.10170615e-01 6.73743248e-01 5.39733768e-01
3.24908525e-01 1.37136757e+00 4.12573293e-02 1.12209640e-01
-6.60243094e-01 -3.05464894e-01 6.99273169e-01 6.53419554e-01
-7.65331268e-01 -4.48011816e-01 -1.26465821e+00 6.23984754e-01
-1.84922826e+00 -1.29717529e+00 -6.37537837e-01 2.05008459e+00
6.35621190e-01 -3.37492555e-01 8.92891705e-01 1.50573701e-01
6.34945869e-01 1.80093855e-01 -7.06324205e-02 -4.62334454e-01
-1.77995235e-01 -9.20497105e-02 -1.82543620e-02 6.66193902e-01
-8.06946695e-01 8.52665126e-01 6.46672726e+00 4.56105798e-01
-1.06309414e+00 2.56255329e-01 2.20111713e-01 -8.79445910e-01
-2.49067754e-01 -1.15487620e-01 -4.01825815e-01 5.18094003e-01
9.35949326e-01 6.51883259e-02 3.70673358e-01 4.48013246e-01
8.41315091e-01 -9.18802917e-02 -1.11839914e+00 1.11843383e+00
7.03983232e-02 -1.15677094e+00 -2.91886836e-01 8.67043957e-02
5.19621372e-01 -1.47059187e-01 -2.08338737e-01 1.17655974e-02
2.42175981e-01 -1.35732520e+00 1.25387073e+00 6.72066808e-01
1.08144653e+00 -3.15409601e-01 3.84993523e-01 3.28850359e-01
-1.09132349e+00 4.10203755e-01 3.59543473e-01 -3.44452441e-01
8.26935828e-01 -1.84779018e-01 -9.83407855e-01 1.48351401e-01
9.52657521e-01 2.59981871e-01 -2.67685533e-01 6.04544938e-01
-2.50190526e-01 4.74843681e-01 -3.56148422e-01 -3.52090001e-01
1.20505571e-01 1.65357426e-01 1.18718112e+00 1.67876947e+00
3.18460971e-01 2.92879134e-01 -1.57020137e-01 4.45282757e-01
3.16111714e-01 -2.95142997e-02 -8.49638343e-01 -2.86604911e-01
4.98748213e-01 1.03902984e+00 -6.33500218e-01 -1.96952030e-01
-1.82193324e-01 1.28064728e+00 -2.61850685e-01 3.49817097e-01
-7.05407560e-01 -6.80134818e-02 7.57394254e-01 2.12262571e-01
-2.00936243e-01 -4.42436874e-01 -2.14512244e-01 -8.59584510e-01
2.82272756e-01 -1.09025431e+00 6.17006943e-02 -1.01268733e+00
-7.33775735e-01 5.86445153e-01 2.85036057e-01 -1.53232026e+00
-9.81258929e-01 -3.81896615e-01 -6.02555156e-01 9.39591587e-01
-2.04086795e-01 -1.09352076e+00 -7.52329707e-01 3.84679645e-01
7.37182260e-01 -1.26918048e-01 9.35104787e-01 1.22682549e-01
-3.32044423e-01 5.79657137e-01 -4.64501798e-01 1.52396828e-01
8.80045831e-01 -1.07429969e+00 1.11252081e+00 6.71847343e-01
3.58559579e-01 4.96819824e-01 9.88904834e-01 -6.76751792e-01
-1.22823775e+00 -4.05382633e-01 7.48690009e-01 -1.18041027e+00
4.44549292e-01 -4.68249828e-01 -5.01361489e-01 5.48881710e-01
3.55638593e-01 -2.65696216e-02 3.85930866e-01 -4.96535838e-01
4.63932473e-03 7.98566520e-01 -1.04596531e+00 9.91970301e-01
1.64515674e+00 -5.11412442e-01 -5.10947168e-01 1.55246690e-01
5.35157144e-01 -9.20915246e-01 -7.62009382e-01 7.65504390e-02
1.36984038e+00 -1.13667262e+00 7.90496826e-01 -6.77019596e-01
4.22312081e-01 -3.27191353e-01 -1.43192679e-01 -1.30699778e+00
7.36228302e-02 -1.39470661e+00 -1.18249647e-01 1.24616385e+00
2.98822790e-01 -2.11919367e-01 7.60084808e-01 8.54399741e-01
-7.14685991e-02 -5.02658486e-01 -9.97452319e-01 -7.72133946e-01
-1.58436328e-01 -8.42625439e-01 4.44615006e-01 9.81511891e-01
3.64000201e-01 8.53516087e-02 -5.96432924e-01 -2.56687433e-01
8.25258791e-02 -4.57421005e-01 1.54928505e+00 -9.10774291e-01
-4.47244465e-01 -7.33816028e-01 -5.94788313e-01 -8.54233325e-01
-3.41178596e-01 -6.45684600e-01 2.29654118e-01 -1.63022959e+00
-2.29398027e-01 6.92006052e-02 8.35959911e-01 4.30479378e-01
8.52607116e-02 4.82890546e-01 6.75430894e-01 6.04993224e-01
-1.22789316e-01 1.36205480e-01 1.41541004e+00 2.78085887e-01
-6.56431198e-01 1.89444590e-02 -2.67864853e-01 7.11263299e-01
7.88664937e-01 -1.92312285e-01 -3.90492409e-01 -3.69882941e-01
-3.73858176e-02 1.88267112e-01 7.94167638e-01 -1.06198156e+00
1.70873210e-01 -1.01634033e-01 1.06592529e-01 -1.71630904e-01
6.19933903e-01 -2.93781996e-01 6.77334070e-01 3.99815410e-01
-2.72714823e-01 2.39908531e-01 2.61612803e-01 2.39186481e-01
1.67267591e-01 1.78811431e-01 2.67242223e-01 -3.36466849e-01
-8.44220579e-01 -1.78732231e-01 -6.42352521e-01 3.34792435e-01
8.41514111e-01 -7.82478690e-01 -2.22698182e-01 -1.09866905e+00
-9.44940984e-01 -2.06341192e-01 5.47605038e-01 8.71132851e-01
4.86234009e-01 -1.51523519e+00 -9.31761146e-01 -1.49660483e-01
1.14572577e-01 3.11208516e-02 5.61991967e-02 9.33972478e-01
-7.66146362e-01 2.20997185e-01 -2.65829980e-01 -7.52625406e-01
-1.52317595e+00 -2.57912040e-01 2.28458583e-01 1.79195225e-01
-1.03174496e+00 7.99244344e-01 -2.84416199e-01 -3.07587653e-01
1.30625963e-01 -2.78104275e-01 1.04427621e-01 5.96995465e-02
7.36778378e-01 7.39212573e-01 -1.43650025e-01 -1.19762254e+00
-4.27502155e-01 4.77498919e-01 5.97825825e-01 -1.27108216e+00
8.92950296e-01 -8.23264718e-02 5.87820411e-01 7.90305078e-01
8.58003080e-01 2.14122519e-01 -1.47230291e+00 5.14974177e-01
-1.26657933e-01 -5.21829307e-01 -4.66082990e-01 -8.15648973e-01
-4.91160750e-01 6.79227352e-01 5.80952704e-01 2.54849702e-01
6.08426690e-01 3.66038717e-02 7.19339848e-01 1.38700008e-01
4.28435743e-01 -1.03410041e+00 1.93431377e-01 3.51740032e-01
1.52137280e+00 -9.88867640e-01 -2.52561688e-01 -1.28265336e-01
-1.12844622e+00 9.69883680e-01 5.31090438e-01 2.13122785e-01
8.49331245e-02 3.91279846e-01 5.05902767e-01 -6.59555271e-02
-4.93986487e-01 -1.59099266e-01 4.07394558e-01 9.12000179e-01
7.16929376e-01 2.32353181e-01 -4.32512105e-01 5.34489095e-01
-8.63785207e-01 1.61191806e-01 5.92033267e-01 1.00803971e+00
1.88449219e-01 -8.79675567e-01 -7.05913365e-01 -1.16092630e-01
-1.54441670e-01 1.79900020e-01 -9.77854490e-01 1.15007091e+00
-1.81293145e-01 1.18121648e+00 -2.30544806e-02 -7.38872468e-01
6.84774935e-01 1.69910610e-01 6.47403479e-01 -5.04855752e-01
-7.93520093e-01 -1.07578747e-02 6.09370828e-01 -9.40215886e-01
-5.19820869e-01 -1.05644417e+00 -1.43934631e+00 -5.09526789e-01
3.34722191e-01 -3.03711891e-01 7.47600675e-01 7.77312100e-01
3.14262033e-01 3.72937948e-01 1.99496392e-02 -1.77352476e+00
-1.85544416e-01 -1.26035655e+00 -1.03966489e-01 9.24638152e-01
3.05906355e-01 -5.61244488e-01 -4.78340507e-01 2.85939157e-01] | [5.60507869720459, -0.08735744655132294] |
fde3f535-8fad-475a-9f4d-cced5de9759a | seedformer-patch-seeds-based-point-cloud | 2207.10315 | null | https://arxiv.org/abs/2207.10315v1 | https://arxiv.org/pdf/2207.10315v1.pdf | SeedFormer: Patch Seeds based Point Cloud Completion with Upsample Transformer | Point cloud completion has become increasingly popular among generation tasks of 3D point clouds, as it is a challenging yet indispensable problem to recover the complete shape of a 3D object from its partial observation. In this paper, we propose a novel SeedFormer to improve the ability of detail preservation and recovery in point cloud completion. Unlike previous methods based on a global feature vector, we introduce a new shape representation, namely Patch Seeds, which not only captures general structures from partial inputs but also preserves regional information of local patterns. Then, by integrating seed features into the generation process, we can recover faithful details for complete point clouds in a coarse-to-fine manner. Moreover, we devise an Upsample Transformer by extending the transformer structure into basic operations of point generators, which effectively incorporates spatial and semantic relationships between neighboring points. Qualitative and quantitative evaluations demonstrate that our method outperforms state-of-the-art completion networks on several benchmark datasets. Our code is available at https://github.com/hrzhou2/seedformer. | ['Chengjie Wang', 'Ying Tai', 'Tong Lu', 'Junwei Zhu', 'Wenqing Chu', 'Yun Cao', 'Haoran Zhou'] | 2022-07-21 | null | null | null | null | ['point-cloud-completion'] | ['computer-vision'] | [-1.06078550e-01 -2.80277818e-01 7.24623203e-02 -2.78015465e-01
-7.83566952e-01 -6.93750381e-01 6.18071139e-01 -3.10246330e-02
5.10213263e-02 3.64612162e-01 1.56991169e-01 4.93262662e-03
-1.66031718e-02 -9.91327703e-01 -8.73809099e-01 -4.26292390e-01
2.67933369e-01 4.41806078e-01 1.53563678e-01 -1.86624214e-01
1.92815155e-01 1.04819047e+00 -1.39904487e+00 5.11088073e-02
1.04000354e+00 9.47812676e-01 5.45380354e-01 2.38028802e-02
-2.23554134e-01 9.56763327e-02 -2.06786767e-01 -4.50718515e-02
4.84685063e-01 7.11470023e-02 -4.15577322e-01 2.91416168e-01
4.88167524e-01 -4.19564486e-01 -4.89434987e-01 1.04104567e+00
2.78887808e-01 1.16098970e-01 5.71269453e-01 -1.13625479e+00
-9.30191457e-01 2.39913613e-02 -7.71145403e-01 -1.70616210e-01
3.18026394e-01 2.79805005e-01 1.00887847e+00 -1.43389750e+00
6.85166001e-01 1.20950532e+00 7.31679022e-01 2.13090420e-01
-1.30100346e+00 -9.65992451e-01 3.00480247e-01 -1.66520283e-01
-1.63782275e+00 -2.04255491e-01 1.22591746e+00 -3.60349566e-01
6.77832723e-01 2.12114125e-01 7.14097440e-01 6.53107703e-01
-1.04162075e-01 7.11695433e-01 7.34181643e-01 -8.47957730e-02
1.69592891e-02 -3.65102261e-01 -3.22566539e-01 5.15867293e-01
2.89461136e-01 1.04334161e-01 -3.49103957e-01 -2.56494641e-01
1.47270167e+00 6.66963577e-01 -4.66400295e-01 -7.09330380e-01
-1.46223617e+00 6.50003076e-01 9.72450137e-01 1.90919951e-01
-6.42014503e-01 2.04161838e-01 -2.30730236e-01 4.16266872e-03
7.64383674e-01 2.56524533e-01 -2.78815359e-01 3.14202845e-01
-9.73909616e-01 4.97240156e-01 3.05327803e-01 1.43439877e+00
1.21549928e+00 -1.23896919e-01 -1.96605623e-01 6.96815789e-01
3.35265309e-01 7.38107502e-01 -1.68481514e-01 -1.00929809e+00
6.05198443e-01 1.01349354e+00 2.56017447e-01 -1.06346118e+00
-2.14677960e-01 -5.98946810e-01 -1.16901886e+00 2.42995024e-01
-5.47010340e-02 2.42511734e-01 -1.05307961e+00 1.47798395e+00
6.47186756e-01 6.73789799e-01 -4.09805864e-01 9.46442902e-01
9.80895042e-01 7.52844036e-01 -3.79698634e-01 8.05797502e-02
1.14363599e+00 -7.50784755e-01 -3.61800700e-01 -5.31408787e-02
-3.38175371e-02 -8.08687449e-01 9.01744664e-01 9.09434482e-02
-1.22558415e+00 -6.47168994e-01 -8.54830325e-01 -5.29444635e-01
-7.86726829e-03 1.57509074e-01 6.80821419e-01 -5.08110039e-02
-9.52842772e-01 6.98177159e-01 -9.56584334e-01 5.76955527e-02
8.61552298e-01 2.25809738e-01 -5.50698459e-01 -4.65629041e-01
-5.87849557e-01 3.80248547e-01 -1.88472003e-01 1.32494718e-01
-7.70452082e-01 -1.20113635e+00 -9.44338858e-01 2.12144062e-01
2.27236599e-01 -1.08752823e+00 1.04104805e+00 -1.57702580e-01
-1.08168590e+00 6.99013352e-01 -4.38627332e-01 1.26893714e-01
3.99435371e-01 -1.61159158e-01 7.24799186e-02 1.55726552e-01
4.51107413e-01 7.43232548e-01 8.43644142e-01 -1.65135038e+00
-5.49664259e-01 -5.11559486e-01 -1.22923469e-02 2.06094310e-01
5.01765385e-02 -4.16048199e-01 -8.61688852e-01 -8.64125490e-01
7.14779258e-01 -7.69119561e-01 -4.32240337e-01 5.13840437e-01
-3.93425643e-01 -2.65250921e-01 6.54995024e-01 -4.58460987e-01
8.28117490e-01 -2.26870131e+00 2.61197865e-01 4.05973583e-01
6.75553322e-01 -7.71066025e-02 -3.23114485e-01 5.13440907e-01
-1.07704706e-01 1.22650601e-01 -5.43973148e-01 -8.19998562e-01
8.84657055e-02 3.85731198e-02 -6.47668898e-01 4.70753163e-01
3.93617809e-01 1.09235001e+00 -9.10367966e-01 -8.75698254e-02
4.19905990e-01 8.20010424e-01 -6.63172066e-01 1.45570114e-01
-1.83659002e-01 6.81504965e-01 -8.53742719e-01 8.80246878e-01
1.28831995e+00 -4.64600980e-01 -4.90603477e-01 -3.73919338e-01
-2.42839754e-01 1.04735665e-01 -1.04716539e+00 2.38020372e+00
-4.89746481e-01 1.20878190e-01 1.12014048e-01 -5.52034199e-01
1.19213653e+00 2.05252439e-01 6.41147077e-01 -2.63763428e-01
-1.01136439e-01 2.15715393e-01 -5.21635592e-01 7.05153719e-02
5.13087094e-01 -1.30434677e-01 1.09194010e-01 1.97977617e-01
-9.83153954e-02 -7.94057429e-01 -2.59844214e-01 2.27953732e-01
1.06957185e+00 2.77414620e-01 1.17689911e-02 -1.06687164e-02
4.23018456e-01 -1.92406103e-02 6.82196319e-01 3.51267606e-01
2.52606422e-01 1.33559728e+00 1.73855394e-01 -4.09585297e-01
-1.09286654e+00 -1.32511067e+00 -8.21044147e-02 1.62535682e-01
4.63357747e-01 -3.73675525e-01 -3.21428269e-01 -3.91270638e-01
3.89486730e-01 4.22083318e-01 -5.24272978e-01 5.88329732e-02
-5.43258667e-01 -2.05990762e-01 6.56597018e-02 5.15999377e-01
3.88022244e-01 -8.82824779e-01 2.53247451e-02 1.79892987e-01
-2.60327250e-01 -1.14070487e+00 -8.34766567e-01 -2.57630944e-01
-1.24168098e+00 -9.02869046e-01 -8.06579292e-01 -7.45682716e-01
1.00503731e+00 9.57243860e-01 1.19073510e+00 5.03871322e-01
-9.54270735e-02 1.55772328e-01 -4.32861835e-01 -1.73777178e-01
1.02390163e-01 7.99844638e-02 -1.01248577e-01 -7.84397200e-02
1.35533154e-01 -1.22765303e+00 -8.20130467e-01 3.10163587e-01
-1.01694739e+00 3.28001112e-01 7.42329478e-01 6.09193265e-01
1.13784993e+00 -6.17867447e-02 2.79069394e-01 -6.01441860e-01
3.76675069e-01 -4.54131067e-01 -6.20547771e-01 -2.86549460e-02
-1.59354582e-01 -7.69612268e-02 4.21121538e-01 -1.30784482e-01
-7.63949990e-01 3.01484615e-01 -2.56355941e-01 -1.22711921e+00
-1.33847013e-01 2.31929079e-01 -3.99276584e-01 -2.49596059e-01
3.18298697e-01 4.43179965e-01 6.95203692e-02 -9.13205266e-01
4.94334072e-01 1.33982375e-01 4.75762039e-01 -6.64421141e-01
1.48470676e+00 8.66762161e-01 1.54596224e-01 -6.18357420e-01
-6.63092732e-01 -6.74784720e-01 -9.15388346e-01 5.95519133e-02
5.26412547e-01 -1.16890538e+00 -5.30888259e-01 3.95202249e-01
-1.50188863e+00 -6.24221414e-02 -3.93686026e-01 1.71553522e-01
-4.37533706e-01 3.19635570e-01 -4.03027892e-01 -2.10907161e-01
-3.46895486e-01 -1.05923200e+00 1.51330340e+00 3.74837741e-02
2.84637034e-01 -6.36721730e-01 1.33014619e-01 1.51788339e-01
1.91749796e-01 4.57966000e-01 5.76579213e-01 1.79243740e-02
-1.26199090e+00 -2.03690499e-01 -4.32179391e-01 2.95157015e-01
4.46465790e-01 -1.32354110e-01 -7.76781797e-01 -3.42076272e-01
3.74913253e-02 1.80128232e-01 8.37069392e-01 3.26707602e-01
1.26977146e+00 -1.25935778e-01 -5.00258505e-01 1.11824012e+00
1.48675179e+00 -3.17572355e-01 6.33062005e-01 -6.49625063e-02
1.03526163e+00 3.39379013e-01 4.75198179e-01 4.80323404e-01
5.55687904e-01 5.85150361e-01 7.01308727e-01 -2.18077242e-01
-2.82310247e-01 -7.46068597e-01 -1.35611326e-01 8.19079518e-01
-3.45224053e-01 4.50719260e-02 -8.05376232e-01 5.28730869e-01
-1.74445736e+00 -7.25950658e-01 -1.23633474e-01 2.13500929e+00
6.19391680e-01 -1.37673408e-01 -2.79860437e-01 -1.60158813e-01
7.02116311e-01 2.88315892e-01 -5.96885741e-01 5.75071633e-01
7.41119459e-02 4.50767845e-01 2.48799294e-01 5.81967235e-01
-9.02519703e-01 8.77364635e-01 4.92601156e+00 9.56538498e-01
-8.28644454e-01 8.83471966e-02 4.02269334e-01 -8.83635283e-02
-7.89283216e-01 1.12329058e-01 -6.98856831e-01 4.33419108e-01
-1.09953061e-01 -1.81830093e-01 3.11070532e-01 7.42596388e-01
2.01150998e-01 3.70210201e-01 -1.04082644e+00 1.07210326e+00
-1.05643861e-01 -1.69792604e+00 3.76720816e-01 2.49479011e-01
9.07075346e-01 2.34411955e-01 -4.99506146e-02 3.91585007e-02
2.73301989e-01 -8.92425120e-01 7.70392537e-01 6.67258382e-01
8.83096576e-01 -8.18752289e-01 3.42771620e-01 4.83981878e-01
-1.55363810e+00 2.88768411e-01 -6.39692426e-01 1.04477100e-01
2.82652617e-01 9.03417945e-01 -4.71525431e-01 9.05101776e-01
6.35155439e-01 1.08176839e+00 -4.42514658e-01 1.32162738e+00
-5.07858992e-01 7.38355145e-02 -4.65719998e-01 3.99402767e-01
3.77372876e-02 -4.39505368e-01 7.33176768e-01 6.99623585e-01
7.24538088e-01 4.99933243e-01 2.38726914e-01 1.38204682e+00
-3.01188797e-01 -9.66141298e-02 -7.88634658e-01 3.95792514e-01
8.57831955e-01 1.47235322e+00 -6.47924900e-01 -6.57337457e-02
-3.68989408e-01 8.84946465e-01 4.23517466e-01 3.96535218e-01
-5.26560009e-01 -2.13209331e-01 9.21482444e-01 3.92000169e-01
4.76690054e-01 -5.62480152e-01 -4.35326606e-01 -1.35181475e+00
3.93356532e-01 -5.02894878e-01 -2.27798030e-01 -1.03130507e+00
-1.40357685e+00 6.16248667e-01 -4.49057408e-02 -1.88735461e+00
2.82985508e-01 -2.97078580e-01 -8.21560323e-01 1.14120138e+00
-1.71127546e+00 -1.53109062e+00 -8.23317111e-01 7.06299424e-01
3.58113587e-01 1.91284433e-01 5.11123776e-01 2.21575364e-01
-9.38056931e-02 1.50247395e-01 -1.29910231e-01 -3.22440602e-02
4.70938444e-01 -9.20474231e-01 8.56457889e-01 9.02016997e-01
1.08662963e-01 8.04918408e-01 1.71871215e-01 -8.18474591e-01
-1.46785998e+00 -1.23377335e+00 6.41343236e-01 -4.99624670e-01
2.60253727e-01 -5.10385036e-01 -1.08602762e+00 6.21676683e-01
-1.68356612e-01 3.16463172e-01 7.21974224e-02 -1.72880039e-01
-3.32404912e-01 -1.83054194e-01 -1.07035160e+00 4.81575757e-01
1.39311862e+00 -3.80637050e-01 -5.08308232e-01 3.37028384e-01
1.13970530e+00 -6.63481593e-01 -8.54897797e-01 5.71293116e-01
1.28551260e-01 -8.62268269e-01 1.31487691e+00 -9.91991758e-02
6.70801520e-01 -7.63151646e-01 -1.65678740e-01 -1.39164746e+00
-7.16382086e-01 -5.93350112e-01 6.95084855e-02 1.20535481e+00
1.66618258e-01 -5.94980955e-01 9.33930874e-01 3.63045573e-01
-5.95112801e-01 -8.61870766e-01 -9.14749920e-01 -6.18632138e-01
9.33058485e-02 -3.41961771e-01 1.26594079e+00 9.31170344e-01
-5.95501065e-01 1.53126240e-01 -8.46108422e-02 4.95434672e-01
8.43070567e-01 5.64252257e-01 9.96276617e-01 -1.32233942e+00
3.74961197e-02 -3.94067228e-01 -4.05027330e-01 -1.59791338e+00
2.79914085e-02 -9.95266020e-01 -1.40442736e-02 -1.84712505e+00
7.63662830e-02 -8.53001654e-01 -9.60681662e-02 5.08333564e-01
-2.16085568e-01 3.64927739e-01 2.33082250e-01 5.43895602e-01
-1.67150855e-01 1.14539397e+00 1.78330529e+00 -2.27763597e-02
-2.22947374e-01 1.71852082e-01 -7.96875775e-01 6.16894901e-01
6.02950573e-01 -3.90506536e-01 -3.51743937e-01 -7.81090260e-01
1.04506455e-01 1.14595599e-01 6.72324657e-01 -9.71501291e-01
2.76916176e-01 -2.56986767e-01 6.08101785e-01 -1.14365256e+00
7.25320160e-01 -1.05297148e+00 4.38311785e-01 -9.27764457e-03
1.88699961e-01 1.80753410e-01 1.47104323e-01 6.59792125e-01
-2.68122882e-01 1.11360595e-01 5.35629630e-01 -2.19491988e-01
-4.21296209e-01 1.15976989e+00 6.13401175e-01 -2.60405302e-01
9.37425375e-01 -6.70239404e-02 -2.44134918e-01 -1.46143943e-01
-4.78369534e-01 4.18417871e-01 1.04266441e+00 5.11193871e-01
1.09081376e+00 -1.74522102e+00 -8.50798249e-01 5.18858075e-01
1.51964083e-01 1.01545894e+00 4.14707154e-01 6.25393033e-01
-7.00253844e-01 2.25235462e-01 -1.22199893e-01 -7.90257633e-01
-8.96279156e-01 4.62911457e-01 5.65020293e-02 8.62938073e-03
-1.01980901e+00 8.32989037e-01 5.95937073e-01 -7.25703955e-01
-9.53136310e-02 -6.46854877e-01 1.26197338e-01 -3.64353120e-01
3.91097724e-01 5.19633889e-02 1.41416445e-01 -5.82684457e-01
-3.23956817e-01 1.04928601e+00 3.64553817e-02 9.72934365e-02
1.67196500e+00 1.20658569e-01 -2.75127292e-01 1.24657452e-01
1.04312265e+00 3.06156904e-01 -1.51265454e+00 -4.99192744e-01
-5.63250363e-01 -9.44074214e-01 1.19095877e-01 -2.99120396e-01
-1.29046631e+00 8.24344575e-01 7.00650290e-02 -2.50207577e-02
1.03882456e+00 1.61976680e-01 8.31187248e-01 6.81535080e-02
5.88504791e-01 -2.71196634e-01 -6.54506311e-02 2.77029544e-01
1.39629126e+00 -1.06082070e+00 3.14477652e-01 -8.88921797e-01
-2.75017858e-01 8.51892710e-01 5.14756322e-01 -6.10754967e-01
9.04717743e-01 -9.34998691e-03 -3.51933420e-01 -4.22409207e-01
-6.40549123e-01 -5.84725104e-02 5.04285634e-01 6.30740762e-01
3.06779109e-02 1.53261587e-01 1.44211471e-01 5.35623550e-01
-2.46503174e-01 4.52984199e-02 1.93696618e-01 7.74553657e-01
-3.65070194e-01 -1.22563589e+00 -4.74498391e-01 4.15991217e-01
1.02730080e-01 -2.25782916e-01 -1.34001240e-01 7.04553425e-01
7.63071999e-02 5.85840762e-01 2.94627756e-01 -4.60014045e-01
5.93889117e-01 -4.86894518e-01 3.47932637e-01 -9.40335393e-01
-1.24133848e-01 1.58576146e-01 -4.90429908e-01 -6.52336895e-01
-3.16438764e-01 -8.07085812e-01 -1.22702122e+00 -4.00409490e-01
-2.08829209e-01 -2.46085729e-02 5.19560516e-01 5.55466652e-01
7.72653520e-01 4.07478094e-01 9.66318130e-01 -1.59149122e+00
-3.00715506e-01 -8.69893789e-01 -5.88921666e-01 3.64735305e-01
4.92887080e-01 -8.14246297e-01 -4.15046543e-01 -7.24489465e-02] | [8.33443546295166, -3.5887677669525146] |
d8ee1f24-087b-45d8-b439-c933c0fd864d | sparsity-based-convolutional-kernel-network | 1807.05648 | null | https://arxiv.org/abs/1807.05648v4 | https://arxiv.org/pdf/1807.05648v4.pdf | Convolutional Sparse Kernel Network for Unsupervised Medical Image Analysis | The availability of large-scale annotated image datasets and recent advances in supervised deep learning methods enable the end-to-end derivation of representative image features that can impact a variety of image analysis problems. Such supervised approaches, however, are difficult to implement in the medical domain where large volumes of labelled data are difficult to obtain due to the complexity of manual annotation and inter- and intra-observer variability in label assignment. We propose a new convolutional sparse kernel network (CSKN), which is a hierarchical unsupervised feature learning framework that addresses the challenge of learning representative visual features in medical image analysis domains where there is a lack of annotated training data. Our framework has three contributions: (i) We extend kernel learning to identify and represent invariant features across image sub-patches in an unsupervised manner. (ii) We initialise our kernel learning with a layer-wise pre-training scheme that leverages the sparsity inherent in medical images to extract initial discriminative features. (iii) We adapt a multi-scale spatial pyramid pooling (SPP) framework to capture subtle geometric differences between learned visual features. We evaluated our framework in medical image retrieval and classification on three public datasets. Our results show that our CSKN had better accuracy when compared to other conventional unsupervised methods and comparable accuracy to methods that used state-of-the-art supervised convolutional neural networks (CNNs). Our findings indicate that our unsupervised CSKN provides an opportunity to leverage unannotated big data in medical imaging repositories. | ['Euijoon Ahn', 'Jinman Kim', 'Michael Fulham', 'Dagan Feng', 'Ashnil Kumar'] | 2018-07-16 | null | null | null | null | ['medical-image-retrieval', 'medical-image-retrieval'] | ['computer-vision', 'medical'] | [ 5.41741729e-01 -4.11153547e-02 -2.07168445e-01 -4.75305259e-01
-1.16862679e+00 -5.23813605e-01 4.17988628e-01 5.75681329e-01
-6.19478524e-01 2.30064258e-01 3.80020142e-01 -2.29510479e-02
-3.56099904e-01 -3.87677491e-01 -6.15637302e-01 -7.26854205e-01
-4.46784496e-01 1.03359692e-01 3.14578742e-01 1.93012834e-01
1.05560586e-01 7.13375151e-01 -1.32892489e+00 7.96198964e-01
4.90384132e-01 1.18188095e+00 3.01252633e-01 6.72586024e-01
1.84302807e-01 1.02826440e+00 -1.25291944e-01 1.58323243e-01
3.04709524e-01 -1.34416282e-01 -1.06334019e+00 3.58410865e-01
6.31881535e-01 -2.40405366e-01 -3.11916411e-01 8.57599914e-01
6.93206966e-01 -4.42943946e-02 7.77227938e-01 -8.49750876e-01
-5.85349023e-01 3.55567038e-02 -4.99965101e-01 4.21736330e-01
-8.96979645e-02 2.13808656e-01 8.67282331e-01 -8.36444557e-01
8.64107490e-01 7.04483747e-01 1.09974337e+00 1.51522160e-01
-1.41235042e+00 -3.74352694e-01 -4.61702555e-01 -7.52051845e-02
-1.36312640e+00 -3.91596496e-01 5.72114706e-01 -6.94280386e-01
9.08491492e-01 1.13679759e-01 3.25620860e-01 7.57982254e-01
3.25359404e-01 7.18467534e-01 1.24072313e+00 -5.24245620e-01
1.64855778e-01 6.44676387e-02 -1.79806560e-01 1.00988662e+00
-1.61219053e-02 5.40268868e-02 -4.24698949e-01 -4.43310201e-01
8.54272664e-01 3.49138141e-01 -9.74959731e-02 -7.53650367e-01
-1.41295433e+00 9.44678545e-01 7.80629814e-01 4.23430949e-01
-5.46645939e-01 3.30820903e-02 6.05250776e-01 1.59351587e-01
3.90224755e-01 5.58152974e-01 -6.31331801e-01 1.79406658e-01
-1.23492801e+00 -1.44777775e-01 5.42159200e-01 5.50474763e-01
8.48535001e-01 -3.62498134e-01 -2.97261328e-01 1.03206217e+00
3.98975052e-02 1.16677389e-01 7.04193115e-01 -7.57757783e-01
-2.23026238e-02 7.16369808e-01 -3.06039602e-01 -1.06033313e+00
-5.94835103e-01 -4.20328110e-01 -9.68406498e-01 2.02746958e-01
3.12706292e-01 2.19600677e-01 -1.16103017e+00 1.23698926e+00
2.39204928e-01 5.99790588e-02 -6.81091771e-02 6.95182085e-01
9.98178005e-01 7.44972453e-02 2.24286407e-01 2.24300593e-01
1.41526294e+00 -9.37339067e-01 -2.15065807e-01 4.79453951e-02
9.44044769e-01 -8.03512335e-01 1.01347148e+00 2.40291670e-01
-6.15026832e-01 -3.55597466e-01 -8.12267721e-01 -1.91832650e-02
-5.75053215e-01 4.11225379e-01 8.43559861e-01 4.89282459e-01
-1.19190633e+00 5.14981925e-01 -1.01593447e+00 -5.25815070e-01
1.14118648e+00 5.80794334e-01 -9.25321817e-01 -3.07456881e-01
-6.03284061e-01 6.44459486e-01 3.55694801e-01 8.88741836e-02
-8.88364434e-01 -1.08361566e+00 -1.09375274e+00 -1.47386029e-01
2.35866442e-01 -4.86485094e-01 1.06945968e+00 -1.25792289e+00
-1.01337206e+00 1.21649516e+00 2.93233782e-01 -3.08938831e-01
1.76212698e-01 7.18921870e-02 -5.57399029e-03 7.45320678e-01
3.11554223e-01 8.11821282e-01 7.65142202e-01 -1.05619133e+00
-4.24507499e-01 -1.54887602e-01 -3.14957142e-01 -1.31203681e-01
-4.21454757e-01 8.46406743e-02 -4.09208834e-01 -8.27149272e-01
-1.11933351e-01 -8.98631871e-01 -6.68804765e-01 3.39246601e-01
-3.32388550e-01 8.58442038e-02 7.39359796e-01 -5.63394725e-01
8.69686544e-01 -2.26648474e+00 -6.33156225e-02 3.60795647e-01
4.81152624e-01 3.29915285e-01 -1.62556879e-02 5.00140607e-01
-1.21750392e-01 -1.74948722e-01 -3.92834365e-01 -2.26493031e-01
-2.99212635e-01 2.39829779e-01 6.41507432e-02 7.11541653e-01
4.48424369e-01 1.16630185e+00 -9.99098778e-01 -8.83535206e-01
2.91698366e-01 5.15029907e-01 -6.10657692e-01 2.34505340e-01
2.04111576e-01 5.60309768e-01 -4.21782374e-01 1.01926053e+00
3.14630628e-01 -7.49616504e-01 4.53331089e-03 -5.14538586e-01
5.08818589e-02 -1.10905878e-01 -8.43290091e-01 2.00930858e+00
-3.28733683e-01 4.97315377e-01 -1.30985588e-01 -1.25285280e+00
4.12433982e-01 2.28893384e-01 8.43599081e-01 -4.11849886e-01
1.98212698e-01 2.26537079e-01 -1.97439238e-01 -7.42848039e-01
4.65732850e-02 -2.00819254e-01 -6.70911074e-02 1.91418931e-01
5.41454494e-01 1.22818507e-01 -1.73669960e-02 2.29420125e-01
1.59222734e+00 -1.62200183e-01 4.74882603e-01 -5.36343575e-01
3.76357287e-01 9.15808901e-02 3.86230886e-01 8.43198776e-01
-3.02205265e-01 1.01346529e+00 4.23555315e-01 -8.69357109e-01
-1.03772938e+00 -1.07388186e+00 -3.78487438e-01 1.12893474e+00
-4.01639849e-01 -3.55999023e-01 -4.20396447e-01 -8.93273830e-01
1.31752789e-01 -2.26408988e-01 -1.06589913e+00 4.40675616e-02
-3.45796317e-01 -7.80346513e-01 6.77490830e-01 7.00060546e-01
2.42571592e-01 -1.20217907e+00 -8.94725800e-01 1.20225497e-01
2.77515501e-01 -1.26643991e+00 -4.37805802e-01 4.79716837e-01
-8.28713000e-01 -1.26616871e+00 -7.36484110e-01 -1.01258445e+00
1.13066733e+00 3.61446850e-02 1.15368903e+00 1.16758838e-01
-1.24271858e+00 7.40813911e-01 -4.55671161e-01 -4.57947046e-01
-1.76069379e-01 2.07363367e-01 -2.31715634e-01 3.56080495e-02
3.36666256e-01 -4.58064944e-01 -8.92162621e-01 4.08031307e-02
-1.21585417e+00 -9.51829329e-02 1.20283842e+00 1.11748767e+00
8.68495345e-01 -1.12456731e-01 4.35673594e-01 -1.25665116e+00
3.08628172e-01 -4.67091262e-01 -3.54821175e-01 1.96812227e-01
-4.57196683e-01 1.02592975e-01 4.85602975e-01 -2.31568903e-01
-5.38439155e-01 4.23221171e-01 1.21117294e-01 -3.64920229e-01
-3.69114220e-01 7.34568834e-01 3.43833506e-01 -5.44657826e-01
7.31644154e-01 1.45002767e-01 4.05094981e-01 -4.01340187e-01
2.80576468e-01 6.75951481e-01 6.10717058e-01 -4.38570529e-01
5.84555745e-01 8.08016002e-01 1.40524656e-01 -9.22517121e-01
-9.82519627e-01 -1.03477705e+00 -1.00685453e+00 8.45143124e-02
1.05570185e+00 -9.45012867e-01 -2.96812505e-01 3.67004097e-01
-5.83984017e-01 -4.13457960e-01 -3.79854888e-01 5.18472850e-01
-4.92911041e-01 5.19006371e-01 -7.14187801e-01 -1.43650576e-01
-6.16988420e-01 -1.21397603e+00 1.45719492e+00 7.52026439e-02
-2.22773373e-01 -1.06330276e+00 2.30619177e-01 3.15254480e-01
6.48318172e-01 5.45806229e-01 8.06138873e-01 -7.11272597e-01
-4.35139745e-01 -4.15073752e-01 -4.76355582e-01 5.10264814e-01
4.71011609e-01 -3.19181979e-01 -9.05324161e-01 -4.82255787e-01
-5.23953915e-01 -8.81547749e-01 9.86007094e-01 4.06066149e-01
1.45111465e+00 -1.16697259e-01 -3.39914024e-01 8.23371351e-01
1.43925130e+00 -5.15950382e-01 5.17747581e-01 4.88463372e-01
8.11416149e-01 4.98691410e-01 2.26204589e-01 2.49219581e-01
2.09426388e-01 5.07136822e-01 9.52194408e-02 -7.36220360e-01
-1.50206819e-01 1.00559853e-01 -1.11051299e-01 6.61549687e-01
-1.76894646e-02 6.17241323e-01 -1.00502920e+00 8.74425828e-01
-1.75659513e+00 -6.04686916e-01 2.76182085e-01 2.04427648e+00
1.06294274e+00 -1.02074906e-01 3.80541757e-02 -2.09747910e-01
2.78327167e-01 -6.65197521e-02 -2.51942039e-01 -6.53377250e-02
9.33763757e-02 6.86436236e-01 8.03594410e-01 3.88210197e-03
-1.61612964e+00 7.52338290e-01 6.25258684e+00 8.15534532e-01
-1.23973334e+00 2.31130734e-01 7.41387486e-01 3.46915536e-02
3.44589233e-01 -3.09108526e-01 -3.37861121e-01 2.08872333e-01
8.42147231e-01 4.02674824e-01 -6.65611923e-02 9.10909832e-01
-4.76861745e-02 -7.59671703e-02 -1.04275346e+00 1.00948906e+00
2.02850133e-01 -1.60931361e+00 -3.74026597e-02 7.24402666e-02
8.07440698e-01 4.02910531e-01 1.72598407e-01 -4.99924719e-02
3.09742630e-01 -1.43558419e+00 6.21818081e-02 4.00309265e-01
9.03351426e-01 -6.35698557e-01 8.87320459e-01 -8.54760595e-03
-1.08540750e+00 3.36887799e-02 -4.36981261e-01 3.38657528e-01
-3.11598748e-01 5.46302676e-01 -1.07626987e+00 6.22650743e-01
8.97604048e-01 8.00949991e-01 -8.68796885e-01 1.31887281e+00
1.03986852e-01 6.36233985e-01 -1.60612628e-01 4.83272642e-01
6.68657959e-01 3.67554575e-01 7.67417718e-03 1.66956747e+00
-9.02378336e-02 -1.78870916e-01 4.78970081e-01 5.15460968e-01
-1.47672981e-01 3.79955202e-01 -5.33137619e-01 -2.30875462e-02
-1.66282672e-02 1.74347913e+00 -1.19923484e+00 -8.02133232e-02
-6.11949325e-01 1.01435089e+00 5.45807362e-01 1.43073380e-01
-2.73384959e-01 -3.54378939e-01 3.22581947e-01 7.44499937e-02
7.51196444e-01 -4.79553789e-02 -1.02679217e-02 -1.00664616e+00
-1.49921075e-01 -8.18522394e-01 7.27364302e-01 -4.19299036e-01
-1.67123806e+00 5.78656018e-01 -6.74553514e-02 -1.21551454e+00
-5.39227538e-02 -8.72666001e-01 -3.60292226e-01 6.20125592e-01
-1.72933710e+00 -1.68856978e+00 -3.56321603e-01 6.42832160e-01
2.07430437e-01 -3.60233665e-01 1.25938046e+00 2.62873024e-01
-1.43320620e-01 6.09104097e-01 2.74241835e-01 5.49061835e-01
8.85957479e-01 -1.27593696e+00 -2.65509218e-01 5.25535941e-01
2.86016971e-01 6.51100397e-01 1.81194350e-01 -3.21724594e-01
-1.28345263e+00 -1.23093474e+00 5.69812179e-01 -6.34162724e-01
5.91785014e-01 -3.58810782e-01 -8.16938281e-01 6.22560441e-01
3.10838800e-02 9.99844551e-01 1.41740298e+00 -3.87741365e-02
-5.43893516e-01 -7.49927983e-02 -1.16053343e+00 1.30030572e-01
4.93300140e-01 -7.17929542e-01 -4.37277138e-01 4.10237014e-01
2.71920562e-01 -2.57669806e-01 -1.16263950e+00 5.05663633e-01
6.87976360e-01 -7.93323934e-01 1.18079448e+00 -7.33357549e-01
5.58489919e-01 -2.15639457e-01 -1.76418573e-01 -1.00421488e+00
-4.28442508e-01 -3.43619436e-01 3.04220170e-01 7.05918014e-01
3.65998864e-01 -4.22317356e-01 7.32717693e-01 4.12797064e-01
-8.99298489e-02 -1.17350507e+00 -8.98475051e-01 -5.03516555e-01
1.21226773e-01 -2.88366586e-01 -4.00263304e-03 1.04751170e+00
-1.88426543e-02 -9.43175778e-02 -2.25627944e-01 1.53649122e-01
6.00535214e-01 1.39405102e-01 5.56480229e-01 -1.07886064e+00
-4.48919237e-01 -2.28656977e-01 -1.05970955e+00 -3.04697037e-01
7.29509816e-02 -1.18137121e+00 -5.61177870e-03 -1.40482116e+00
5.84529400e-01 -5.16595781e-01 -8.07239175e-01 7.46081293e-01
-1.12646721e-01 9.10999835e-01 -1.94254920e-01 4.48268056e-01
-1.00605273e+00 1.21144421e-01 8.89970958e-01 -1.11631185e-01
-3.84042934e-02 -3.75709474e-01 -6.84364140e-01 7.16217875e-01
5.37893534e-01 -5.78833520e-01 -2.39142805e-01 -2.23470792e-01
-8.10540766e-02 -5.66897213e-01 6.80169404e-01 -9.59290981e-01
2.60700881e-01 2.07694024e-01 7.77149856e-01 -2.91297674e-01
-1.78605884e-01 -8.30907166e-01 -2.69158393e-01 5.12906909e-01
-4.36400950e-01 -2.47913674e-01 2.38578439e-01 6.36443794e-01
-2.64072001e-01 1.21124890e-02 8.17334652e-01 -4.30560201e-01
-8.45939636e-01 4.09906179e-01 -3.29023302e-01 -2.14113183e-02
9.47716713e-01 -1.78282201e-01 6.48049563e-02 2.95873042e-02
-5.94799340e-01 4.67371754e-02 4.20586318e-01 1.58302352e-01
6.12882853e-01 -1.30200529e+00 -6.69331908e-01 1.31670430e-01
6.54057860e-01 1.42889500e-01 2.89827764e-01 1.20731997e+00
-8.83413136e-01 4.25631672e-01 -2.37979576e-01 -9.52072740e-01
-1.20978653e+00 5.47539592e-01 2.10490555e-01 -6.37927830e-01
-1.07343638e+00 8.13846290e-01 3.56763899e-01 -6.09589756e-01
1.07664838e-01 -3.25157374e-01 -1.02216333e-01 -2.41201725e-02
6.42970681e-01 -2.02720508e-01 3.65089983e-01 -7.50164151e-01
-5.23343801e-01 4.24269080e-01 -5.46195388e-01 4.32340324e-01
1.79378164e+00 2.28168339e-01 -1.61632702e-01 2.55258024e-01
1.63326299e+00 -1.24687091e-01 -1.13141119e+00 -4.67761695e-01
1.81066051e-01 -4.58932102e-01 2.47642875e-01 -7.57562280e-01
-9.24990058e-01 6.26351416e-01 7.85099804e-01 -2.66277671e-01
1.14296865e+00 4.10352349e-01 7.07355917e-01 4.59415734e-01
1.98249578e-01 -1.10026217e+00 3.17296147e-01 1.06909610e-01
5.13979197e-01 -1.57045507e+00 1.80846840e-01 -3.51099938e-01
-6.01296425e-01 1.21018732e+00 1.32376790e-01 -2.59969682e-01
8.19637835e-01 3.13464284e-01 3.00220639e-01 -7.24321187e-01
-3.84894550e-01 -3.09082896e-01 6.85431004e-01 5.35006821e-01
5.89386642e-01 -1.58628106e-01 8.46589357e-02 5.61123252e-01
1.20391332e-01 2.81099170e-01 1.31596595e-01 1.42745411e+00
-1.94929048e-01 -9.91465807e-01 -9.96963978e-02 7.42308080e-01
-9.08446670e-01 -2.68725216e-01 -1.26808286e-01 6.26677632e-01
1.40600443e-01 4.71445173e-01 -1.12005644e-01 -2.99425311e-02
1.20669134e-01 -3.70929763e-03 5.20372808e-01 -9.64060724e-01
-6.50770426e-01 5.25224581e-02 -2.34969109e-01 -7.15262294e-01
-6.40203953e-01 -5.65887988e-01 -1.05720377e+00 4.39928144e-01
-3.09512485e-02 -2.11085409e-01 3.60565275e-01 9.97428596e-01
5.33049166e-01 5.38413048e-01 4.27139342e-01 -9.61090088e-01
-3.59446824e-01 -7.87996948e-01 -6.62511766e-01 7.87425339e-01
4.44806576e-01 -6.56224310e-01 -7.06679374e-02 4.30390120e-01] | [14.907489776611328, -2.4762940406799316] |
0e188fde-032e-4332-8e90-149fc882e98c | reconstructing-the-mind-s-eye-fmri-to-image | 2305.18274 | null | https://arxiv.org/abs/2305.18274v1 | https://arxiv.org/pdf/2305.18274v1.pdf | Reconstructing the Mind's Eye: fMRI-to-Image with Contrastive Learning and Diffusion Priors | We present MindEye, a novel fMRI-to-image approach to retrieve and reconstruct viewed images from brain activity. Our model comprises two parallel submodules that are specialized for retrieval (using contrastive learning) and reconstruction (using a diffusion prior). MindEye can map fMRI brain activity to any high dimensional multimodal latent space, like CLIP image space, enabling image reconstruction using generative models that accept embeddings from this latent space. We comprehensively compare our approach with other existing methods, using both qualitative side-by-side comparisons and quantitative evaluations, and show that MindEye achieves state-of-the-art performance in both reconstruction and retrieval tasks. In particular, MindEye can retrieve the exact original image even among highly similar candidates indicating that its brain embeddings retain fine-grained image-specific information. This allows us to accurately retrieve images even from large-scale databases like LAION-5B. We demonstrate through ablations that MindEye's performance improvements over previous methods result from specialized submodules for retrieval and reconstruction, improved training techniques, and training models with orders of magnitude more parameters. Furthermore, we show that MindEye can better preserve low-level image features in the reconstructions by using img2img, with outputs from a separate autoencoder. All code is available on GitHub. | ['Tanishq Mathew Abraham', 'Kenneth A. Norman', 'David Weisberg', 'Elad Yundler', 'Nathalie Verlinde', 'Aidan J. Dempster', 'Ethan Cohen', 'Alex Nguyen', 'Stepan Shabalin', 'Jimmie Goode', 'Atmadeep Banerjee', 'Paul S. Scotti'] | 2023-05-29 | null | null | null | null | ['image-reconstruction'] | ['computer-vision'] | [ 2.59124395e-02 -1.40405998e-01 -7.83831701e-02 -2.67452359e-01
-9.62324023e-01 -5.35488188e-01 7.40436792e-01 -2.96641558e-01
-6.66819513e-01 4.87850636e-01 5.86538017e-01 2.50304729e-01
-1.14700496e-01 -5.18187761e-01 -7.88706124e-01 -8.03600192e-01
-6.73026443e-02 4.45274174e-01 -1.25712365e-01 2.86142170e-01
6.68111593e-02 6.98158205e-01 -1.31969297e+00 5.05931318e-01
2.31842622e-01 8.53428781e-01 5.19116342e-01 5.45897186e-01
2.64967918e-01 6.76433206e-01 -2.95803457e-01 -2.03104183e-01
2.80214369e-01 -3.41936558e-01 -7.36221731e-01 -2.22703516e-01
5.41411221e-01 -6.81845903e-01 -8.05949926e-01 7.87654340e-01
7.40884960e-01 1.13863565e-01 9.50399220e-01 -8.56656194e-01
-1.13131106e+00 2.92658806e-03 -2.80948490e-01 6.32340014e-01
1.94146410e-01 3.38139862e-01 7.06564367e-01 -1.32810807e+00
9.59360898e-01 1.21613693e+00 4.81005609e-01 6.72603011e-01
-1.67106318e+00 -4.92714524e-01 -2.07078934e-01 2.15237841e-01
-1.53270626e+00 -7.90328681e-01 4.40924257e-01 -4.51471299e-01
1.26878834e+00 3.10189605e-01 7.28410184e-01 1.42591870e+00
5.69650114e-01 7.79763937e-01 1.32223928e+00 -1.51002720e-01
3.37150209e-02 1.25339746e-01 -4.82198671e-02 8.35695505e-01
-6.03123344e-02 2.24030033e-01 -7.97691345e-01 -1.39617682e-01
1.23055947e+00 3.20624799e-01 -5.58979928e-01 -3.30657363e-01
-1.58911312e+00 7.95509994e-01 6.37700140e-01 3.85573238e-01
-6.53392136e-01 3.38880122e-01 1.43060967e-01 3.16985756e-01
3.88544530e-01 4.23534602e-01 4.62152213e-02 1.53271168e-01
-1.32272756e+00 -7.10805207e-02 5.64050615e-01 7.22130537e-01
5.09314060e-01 1.41284764e-01 -4.02277052e-01 8.25906336e-01
3.45170170e-01 6.52338386e-01 9.63838160e-01 -1.25051534e+00
4.33730632e-02 -5.38383648e-02 -8.82998258e-02 -1.01394701e+00
-2.26195157e-01 -3.54012579e-01 -8.47302735e-01 -2.32218485e-02
-1.56725273e-01 2.79888839e-01 -9.23155427e-01 1.81188989e+00
-2.50465453e-01 1.58318534e-01 -1.04746334e-01 1.11327648e+00
1.02150679e+00 6.55494213e-01 1.21910684e-01 -2.04535440e-01
1.51178265e+00 -9.28174138e-01 -7.08557367e-01 -4.01184499e-01
2.99203992e-02 -3.36504370e-01 1.11852300e+00 2.20739305e-01
-1.44398165e+00 -5.18919408e-01 -1.03794694e+00 -4.36997801e-01
-2.79003471e-01 2.02790990e-01 7.12638855e-01 2.35973194e-01
-1.61286342e+00 4.52367932e-01 -1.04842091e+00 -3.66655856e-01
6.11417830e-01 3.12315851e-01 -9.40593839e-01 -2.06401482e-01
-1.00428689e+00 9.90282059e-01 2.05230027e-01 9.39680114e-02
-1.38997304e+00 -7.34701991e-01 -8.85187089e-01 2.61377126e-01
-1.86300188e-01 -1.05532372e+00 8.64151657e-01 -7.64943480e-01
-1.32814050e+00 1.18141723e+00 -3.16338837e-01 -4.39186752e-01
1.03311330e-01 3.97606194e-02 -2.69669712e-01 9.65900779e-01
1.10824324e-01 1.25895762e+00 1.12968075e+00 -1.01550317e+00
4.45984870e-01 -4.12778318e-01 -3.10024261e-01 2.95254141e-01
-4.80041116e-01 -9.65780169e-02 -7.20701635e-01 -6.92824781e-01
3.45701396e-01 -8.44299018e-01 1.35475099e-01 4.07516837e-01
8.16533118e-02 1.32547706e-01 3.53895873e-01 -8.37497532e-01
8.18595052e-01 -2.17538309e+00 4.40105200e-01 -6.99767917e-02
7.39820063e-01 1.47107635e-02 -4.38232809e-01 2.58640677e-01
-3.31986755e-01 1.09328426e-01 -1.54826850e-01 -5.24518430e-01
-9.62820724e-02 2.83348858e-01 -3.02409112e-01 8.55112314e-01
1.38039693e-01 1.57216203e+00 -7.62641966e-01 -4.24963325e-01
2.23446190e-01 7.49703765e-01 -5.39456427e-01 2.01184481e-01
2.99131751e-01 2.86079317e-01 -9.43683535e-02 6.26674831e-01
5.00610948e-01 -5.51250100e-01 1.76977813e-01 -5.30171394e-01
1.13291837e-01 -9.54620764e-02 -4.07581002e-01 2.05874705e+00
-5.25113106e-01 9.69673991e-01 9.62090790e-02 -1.03714585e+00
3.24069262e-01 3.46163660e-01 4.80627626e-01 -1.04631150e+00
1.44004775e-02 -7.58803450e-03 -3.00868511e-01 -4.92790490e-01
1.67438120e-01 -3.25011283e-01 2.48120770e-01 7.28053510e-01
6.44175112e-01 4.46554758e-02 -1.03152841e-01 5.05222261e-01
1.11608386e+00 -5.29808663e-02 3.11581679e-02 -2.80699015e-01
-1.40129775e-02 -3.38297635e-01 3.06000840e-02 9.62501884e-01
-1.37104988e-01 9.31660652e-01 2.46184677e-01 -2.41788030e-01
-9.67850924e-01 -1.46752405e+00 -4.10488069e-01 8.10828984e-01
-1.61998004e-01 -2.75962740e-01 -6.64587557e-01 -3.91491264e-01
-1.12499274e-01 4.22040224e-01 -9.30250764e-01 -3.75714988e-01
-2.25866929e-01 -8.80792379e-01 5.64581394e-01 5.17984867e-01
4.94358659e-01 -1.33121037e+00 -7.16584861e-01 1.45431850e-02
-3.74594629e-01 -1.03643870e+00 -4.92001981e-01 1.34797290e-01
-1.13760149e+00 -7.77627826e-01 -1.23643053e+00 -6.62531257e-01
8.77135217e-01 4.63445604e-01 1.26949978e+00 4.97916564e-02
-6.80832624e-01 9.16848481e-01 5.08757457e-02 1.44214615e-01
-8.39523301e-02 -2.60138690e-01 2.30191201e-01 -1.95865825e-01
1.73883438e-01 -7.73066461e-01 -9.57473695e-01 1.22918218e-01
-1.20256841e+00 1.00986227e-01 9.05456543e-01 1.01795924e+00
9.12816525e-01 -3.63786846e-01 3.06993932e-01 -6.64350748e-01
7.31621325e-01 -5.53609788e-01 -3.76670212e-01 1.93612784e-01
-7.55844533e-01 2.77460545e-01 1.55083105e-01 -4.47231978e-01
-7.86573887e-01 -5.38456887e-02 -7.18998834e-02 -9.75309134e-01
1.07648574e-01 5.13038874e-01 2.50396550e-01 -1.35334834e-01
7.53245175e-01 6.09226644e-01 2.69707739e-01 -5.42600393e-01
5.16092002e-01 3.46280932e-01 7.58719802e-01 -2.88306296e-01
4.61091459e-01 7.32305765e-01 -2.52317041e-01 -6.39507949e-01
-6.01858318e-01 -4.36491996e-01 -6.97282553e-01 -1.40022486e-01
1.03618848e+00 -1.22872341e+00 -7.80711412e-01 2.24756196e-01
-9.57735181e-01 -3.18432242e-01 -2.99389601e-01 7.90198267e-01
-5.93329251e-01 1.81331038e-01 -9.50491250e-01 -2.60960937e-01
-5.77451944e-01 -1.13694167e+00 1.20983577e+00 -1.06753260e-01
-1.72998399e-01 -1.00070775e+00 1.84891030e-01 1.89189389e-01
5.34702063e-01 -8.39408115e-02 9.15732861e-01 -3.52923751e-01
-7.99481690e-01 -1.48767024e-01 -4.16042715e-01 3.64792168e-01
-1.80352643e-01 -6.10915065e-01 -1.17189670e+00 -4.05710220e-01
3.63483250e-01 -4.92535204e-01 1.28656375e+00 5.19229412e-01
1.23874092e+00 -3.68385166e-01 -3.68621826e-01 9.78623748e-01
1.38029361e+00 -3.05017829e-01 1.01650763e+00 -5.99415749e-02
4.15160835e-01 2.24146664e-01 -1.23979189e-01 -2.47018095e-02
1.55422464e-01 5.87355673e-01 1.33932233e-01 -2.58270621e-01
-5.95575571e-01 -2.07793325e-01 6.11709356e-01 1.17846787e+00
1.56958371e-01 -5.57497852e-02 -6.37816310e-01 5.91200829e-01
-1.54265451e+00 -1.20540309e+00 2.99393356e-01 2.07043266e+00
8.97391856e-01 -3.54165852e-01 -2.74288833e-01 -5.02829134e-01
1.97296441e-01 2.58122414e-01 -6.98946357e-01 -4.14226614e-02
-2.80950040e-01 3.94032627e-01 3.72929633e-01 3.76733929e-01
-7.61176646e-01 8.05385649e-01 8.16102409e+00 7.65476227e-01
-1.22516346e+00 6.46250427e-01 5.31055689e-01 -6.04624689e-01
-5.19838274e-01 -2.68013030e-01 -1.90196857e-01 3.36300462e-01
1.12969899e+00 1.25505060e-01 9.42371190e-01 3.85856986e-01
-9.84292664e-03 -2.85261810e-01 -1.15597439e+00 1.52464819e+00
6.12098515e-01 -1.56803155e+00 1.13289706e-01 4.75865677e-02
5.85199416e-01 4.48098749e-01 3.52100849e-01 1.86560407e-01
-1.64944515e-01 -1.34845638e+00 5.94517171e-01 1.04331231e+00
1.15549636e+00 -1.82964161e-01 3.69858503e-01 2.95233637e-01
-6.69806004e-01 1.76627278e-01 -5.64501405e-01 4.69397187e-01
-1.78117126e-01 5.66575527e-01 -5.94844401e-01 3.01127046e-01
7.54078686e-01 8.33644331e-01 -7.90969193e-01 7.74638891e-01
-1.69629350e-01 5.19823790e-01 -1.52267084e-01 4.17304337e-01
-1.22858115e-01 -6.28027916e-02 3.83537769e-01 1.19590008e+00
2.80733198e-01 2.63041973e-01 -1.97130919e-01 1.37006593e+00
-3.82989824e-01 -1.57293260e-01 -6.80948019e-01 -2.72569001e-01
1.55658096e-01 1.38387668e+00 -6.15829170e-01 -4.60731953e-01
-2.53373176e-01 1.48418319e+00 4.38399136e-01 6.75005019e-01
-6.63998246e-01 -4.41497080e-02 1.83396310e-01 1.33794323e-01
1.24068558e-01 -3.36634606e-01 4.43322472e-02 -1.40909648e+00
-5.61000183e-02 -6.37761831e-01 1.46588326e-01 -1.36230135e+00
-1.34100831e+00 7.55710185e-01 2.82281816e-01 -8.05473804e-01
-2.89068729e-01 -6.01599693e-01 -1.51605427e-01 1.09617019e+00
-1.42369032e+00 -9.55045879e-01 -3.41291189e-01 8.03435564e-01
3.31966788e-01 -1.77745745e-01 1.14827538e+00 4.01087284e-01
-3.17088872e-01 3.52061987e-01 1.95944503e-01 1.08167805e-01
8.99532676e-01 -9.77798879e-01 7.98532292e-02 5.57081699e-01
6.16304874e-01 1.17095268e+00 1.87360153e-01 -5.41875541e-01
-1.68174791e+00 -7.02387989e-01 4.99909371e-01 -6.31274402e-01
3.14133465e-01 -5.70302367e-01 -8.22749555e-01 9.37335551e-01
4.49880600e-01 3.64219338e-01 6.53472185e-01 -5.56123368e-02
-4.11625832e-01 6.70516416e-02 -1.02632737e+00 4.95708793e-01
1.07608008e+00 -1.17928326e+00 -6.15496695e-01 5.51110446e-01
2.46307552e-01 -2.47972637e-01 -9.10620093e-01 1.50820735e-04
7.84119964e-01 -8.10209930e-01 1.19001317e+00 -4.54709798e-01
3.89443547e-01 -5.69394007e-02 -1.72297925e-01 -1.23292089e+00
-5.73240697e-01 -2.67152637e-01 -2.66732365e-01 6.11134589e-01
2.39043590e-02 -5.53736746e-01 3.86501253e-01 5.56199610e-01
6.19248934e-02 -6.42045259e-01 -9.64346349e-01 -6.61316156e-01
-5.13396552e-03 -1.50263026e-01 6.96944296e-02 7.60727465e-01
-2.51721114e-01 3.56186986e-01 -4.08946246e-01 -1.03113219e-01
5.28072059e-01 1.60763443e-01 2.01708674e-01 -9.39446628e-01
-3.49255860e-01 -2.67076880e-01 -3.33771259e-01 -1.11163497e+00
4.31595683e-01 -1.52988088e+00 -6.20413795e-02 -1.57149434e+00
7.42353737e-01 9.09858570e-02 -4.54179913e-01 7.69822001e-01
1.79004312e-01 8.02356601e-01 1.92636132e-01 7.21955061e-01
-6.76803410e-01 7.27036357e-01 1.26654780e+00 -2.39549100e-01
1.78498045e-01 -6.77822709e-01 -7.42189348e-01 3.08123201e-01
5.22872925e-01 -6.60322964e-01 -5.21619976e-01 -6.84980452e-01
6.32575080e-02 1.69142142e-01 9.32830930e-01 -9.74879324e-01
1.83964670e-01 2.22749084e-01 1.04225481e+00 -2.55332798e-01
6.74042463e-01 -5.93414724e-01 4.45632935e-01 2.33881503e-01
-3.47904205e-01 1.49503216e-01 4.11729127e-01 5.22871971e-01
-2.04037100e-01 -1.59386650e-01 6.85367107e-01 -4.65871096e-01
-6.40275240e-01 4.30383682e-01 -6.27116382e-01 -1.16184667e-01
5.11744618e-01 -7.10388944e-02 -2.86920130e-01 -4.58944201e-01
-1.04599690e+00 -2.12710872e-01 4.05456781e-01 2.59407073e-01
9.74795103e-01 -1.50520790e+00 -5.98208189e-01 3.75248551e-01
4.29864861e-02 -6.49975717e-01 4.57838148e-01 1.18687904e+00
-3.76852930e-01 6.07619345e-01 -5.98975360e-01 -6.26266241e-01
-9.83954132e-01 5.73994994e-01 4.20518160e-01 -2.20713466e-01
-8.18334222e-01 8.12011063e-01 5.36631882e-01 -1.95931971e-01
-7.09539056e-02 -1.89232342e-02 1.24149881e-01 1.48207933e-01
6.33961201e-01 -1.65948570e-01 1.42450973e-01 -5.25273800e-01
-4.94497687e-01 2.91291445e-01 -4.13271599e-02 -5.19867718e-01
1.51818478e+00 -9.88857523e-02 -2.47640952e-01 4.19400275e-01
1.58285105e+00 -3.34618650e-02 -1.17483020e+00 -1.12043902e-01
-5.89720964e-01 -4.27965045e-01 6.51706398e-01 -9.88067985e-01
-1.24418497e+00 1.06230640e+00 9.24041152e-01 -3.75552058e-01
1.18030369e+00 1.74286768e-01 6.50668919e-01 3.72824878e-01
2.45226458e-01 -8.07392716e-01 4.20790195e-01 7.60701448e-02
1.33759570e+00 -1.06422687e+00 2.63070852e-01 2.00701565e-01
-5.85290611e-01 8.46588194e-01 1.45499751e-01 -2.25736484e-01
6.29257023e-01 1.60085305e-03 -1.14989646e-01 -7.31917739e-01
-8.29497933e-01 8.48872364e-02 5.99295735e-01 4.49908584e-01
2.48589337e-01 -5.71353249e-02 1.97480600e-02 4.97857094e-01
1.72108456e-01 4.63660248e-02 1.18258201e-01 6.43378854e-01
-6.12175763e-02 -9.09895062e-01 -2.47089744e-01 6.65678442e-01
-3.39062631e-01 -5.42387784e-01 -1.08081676e-01 6.63387954e-01
-2.66639501e-01 3.87053818e-01 2.94125289e-01 -4.53010388e-02
4.38619889e-02 2.26734892e-01 9.72803354e-01 -5.93721628e-01
-4.17029351e-01 2.39373460e-01 -3.95534098e-01 -9.34032321e-01
-4.81769800e-01 -6.07942522e-01 -1.01540947e+00 -1.69784725e-01
-5.63153590e-04 -1.76635548e-01 6.36788785e-01 6.07569933e-01
8.03652942e-01 4.70799178e-01 2.30949745e-01 -1.04361725e+00
-2.68034160e-01 -9.93714869e-01 -6.22894168e-01 4.16107118e-01
2.86261797e-01 -7.84276068e-01 -1.60932392e-01 2.23137766e-01] | [10.743611335754395, 2.4993081092834473] |
2ff35440-d2d1-4faa-b7ed-22838ec2edfe | mcua-multi-level-context-and-uncertainty | 2108.10709 | null | https://arxiv.org/abs/2108.10709v1 | https://arxiv.org/pdf/2108.10709v1.pdf | MCUa: Multi-level Context and Uncertainty aware Dynamic Deep Ensemble for Breast Cancer Histology Image Classification | Breast histology image classification is a crucial step in the early diagnosis of breast cancer. In breast pathological diagnosis, Convolutional Neural Networks (CNNs) have demonstrated great success using digitized histology slides. However, tissue classification is still challenging due to the high visual variability of the large-sized digitized samples and the lack of contextual information. In this paper, we propose a novel CNN, called Multi-level Context and Uncertainty aware (MCUa) dynamic deep learning ensemble model.MCUamodel consists of several multi-level context-aware models to learn the spatial dependency between image patches in a layer-wise fashion. It exploits the high sensitivity to the multi-level contextual information using an uncertainty quantification component to accomplish a novel dynamic ensemble model.MCUamodelhas achieved a high accuracy of 98.11% on a breast cancer histology image dataset. Experimental results show the superior effectiveness of the proposed solution compared to the state-of-the-art histology classification models. | ['Saeid Nahavandi', 'Abbas Khosravi', 'U Rajendra Acharya', 'Moloud Abdar', 'Mohamed Medhat Gaber', 'Mohammed M. Abdelsamea', 'Zakaria Senousy'] | 2021-08-24 | null | null | null | null | ['breast-cancer-histology-image-classification'] | ['medical'] | [ 4.91594672e-02 -1.48460850e-01 5.52560203e-02 -4.40852553e-01
-1.00056279e+00 -1.71683535e-01 4.30691510e-01 5.21197796e-01
-5.03927767e-01 7.31119335e-01 -1.95979461e-01 -2.29114935e-01
-4.56618071e-01 -7.23642409e-01 -5.92551947e-01 -1.35005319e+00
-1.59179792e-01 2.49182522e-01 2.06911445e-01 1.02656998e-01
-9.34517663e-03 7.77283013e-01 -1.44321120e+00 6.43583715e-01
7.71912754e-01 1.31160414e+00 1.66176811e-01 7.32278645e-01
-2.41169930e-01 5.78879595e-01 -2.07594037e-01 -8.20220336e-02
-1.86544895e-01 -1.36473119e-01 -4.46838498e-01 -2.20142841e-01
3.46944183e-01 -4.12665121e-02 3.08058765e-02 1.15038192e+00
5.15373290e-01 -4.49287325e-01 7.56100178e-01 -8.68315756e-01
-2.57209927e-01 4.18449521e-01 -6.81845665e-01 4.93998289e-01
-5.93307316e-01 -1.11697674e-01 5.12062073e-01 -7.66341507e-01
7.29369819e-01 6.71247244e-01 9.22732651e-01 4.15501028e-01
-1.34336150e+00 -7.17034638e-01 -1.06838264e-01 3.95915419e-01
-1.59061086e+00 -2.49908864e-01 7.82637417e-01 -5.06288409e-01
6.58666313e-01 2.78978318e-01 6.19209826e-01 9.33609664e-01
9.82487679e-01 4.80149746e-01 1.40050793e+00 -3.89201671e-01
3.69391114e-01 2.19475269e-01 2.32217506e-01 8.19844782e-01
2.25484803e-01 4.49823476e-02 -2.68471837e-01 7.38582388e-02
7.77631581e-01 8.31019282e-02 -1.02453828e-01 -3.38972926e-01
-8.88651133e-01 4.83753920e-01 8.88061941e-01 7.71285713e-01
-1.42234683e-01 2.86596417e-01 4.62515831e-01 -4.89347018e-02
4.16918129e-01 9.12126675e-02 -2.01913044e-01 2.65653282e-01
-9.83074009e-01 -8.84702578e-02 4.70637590e-01 3.89989257e-01
5.11270821e-01 -2.82133460e-01 -2.28563622e-01 6.19261801e-01
2.87818432e-01 1.95738137e-01 5.30296504e-01 -3.39933395e-01
-1.32380694e-01 7.42147923e-01 -3.17513734e-01 -1.08084023e+00
-8.20010662e-01 -1.00177109e+00 -1.55230749e+00 5.02315819e-01
3.90538454e-01 1.23351000e-01 -1.20595193e+00 1.39734495e+00
3.98319870e-01 3.41349840e-01 2.65814308e-02 7.21679211e-01
9.60332632e-01 3.56614083e-01 3.17754865e-01 -2.09696963e-01
1.36546993e+00 -4.27054524e-01 -8.21960628e-01 2.89570242e-01
6.18242025e-01 -3.55971187e-01 3.85099292e-01 2.37134814e-01
-5.35338640e-01 -3.06501359e-01 -1.31246507e+00 8.50159526e-02
-4.53313947e-01 4.98374701e-01 6.56522334e-01 4.73831207e-01
-1.09298837e+00 3.78231138e-01 -1.04390228e+00 -3.36693883e-01
8.36871207e-01 4.47019339e-01 -7.36874759e-01 -1.08652443e-01
-7.80815661e-01 7.03736961e-01 4.23707396e-01 3.68074030e-01
-8.57808769e-01 -8.82082522e-01 -6.55621171e-01 1.77783087e-01
8.57645869e-02 -4.52686578e-01 9.36929226e-01 -9.30869639e-01
-1.16599894e+00 7.84598649e-01 8.07323828e-02 -2.95764565e-01
7.08907604e-01 4.24678415e-01 -1.57857895e-01 2.70270586e-01
-2.94779807e-01 4.93332237e-01 4.95852739e-01 -1.46396315e+00
-6.14077628e-01 -5.68095446e-01 -4.04299170e-01 -2.06756994e-01
-3.97268772e-01 -4.41593200e-01 -2.91446120e-01 -5.78309298e-01
2.24312857e-01 -6.13032103e-01 -2.90423661e-01 3.81731182e-01
-2.51119524e-01 -5.94020747e-02 8.25888157e-01 -5.06101727e-01
1.15564620e+00 -2.19978166e+00 1.49187699e-01 3.96763563e-01
2.79138565e-01 6.22960515e-02 6.93892017e-02 1.25035606e-02
-3.51562388e-02 1.88126922e-01 -3.19262654e-01 -2.74270743e-01
-2.39960983e-01 1.32428870e-01 5.16996443e-01 4.61223155e-01
2.91295737e-01 9.39487040e-01 -8.51296902e-01 -9.59639251e-01
2.11269215e-01 7.12111533e-01 -1.63689405e-01 -3.82841416e-02
-8.01961720e-02 4.05807257e-01 -3.30591321e-01 1.14997089e+00
9.29584920e-01 -3.45537335e-01 4.76458639e-01 -5.74184000e-01
-2.51875985e-02 -5.77217877e-01 -8.55800688e-01 1.62584162e+00
-4.72128868e-01 6.85262203e-01 1.82632133e-01 -8.56380284e-01
6.70924723e-01 3.81739944e-01 5.96485078e-01 -2.50117898e-01
4.25202608e-01 3.82004410e-01 2.58885622e-01 -5.74863911e-01
3.98540497e-02 -3.67458820e-01 1.11235201e-01 -2.67504007e-02
2.38232300e-01 1.75441787e-01 -4.28864127e-03 -7.55507052e-02
1.21082962e+00 -1.20247997e-01 6.36618972e-01 -6.59441471e-01
5.85841358e-01 -7.42077604e-02 6.77099407e-01 6.22220278e-01
-6.92114353e-01 6.09222054e-01 5.80865145e-01 -5.62494755e-01
-7.59207189e-01 -9.07424986e-01 -6.24945223e-01 3.42851877e-01
-1.25900090e-01 8.99559259e-02 -7.15829670e-01 -8.30266953e-01
1.71212569e-01 9.90152135e-02 -1.22834122e+00 -2.18229238e-02
-2.85260081e-01 -1.21111310e+00 5.44847190e-01 7.35474646e-01
8.42544019e-01 -5.66125393e-01 -4.63776499e-01 1.95976570e-01
1.00962006e-01 -8.18459272e-01 1.13502376e-01 5.10916770e-01
-9.49360430e-01 -1.16294730e+00 -4.50710773e-01 -7.63436615e-01
9.10283148e-01 -2.54059166e-01 8.70248497e-01 1.52987614e-01
-9.26524341e-01 -2.84955297e-02 -1.93454787e-01 -5.16483009e-01
-4.21951085e-01 8.78502652e-02 -4.54991549e-01 2.16546178e-01
3.85162950e-01 -5.23578942e-01 -6.81311905e-01 -7.49459043e-02
-1.12920177e+00 -5.79751022e-02 1.21840870e+00 1.37837696e+00
9.98884976e-01 3.82935315e-01 4.97423321e-01 -9.67840433e-01
1.77795812e-01 -5.67594945e-01 -4.49453026e-01 4.69430447e-01
-3.37462574e-01 -1.50814399e-01 3.17416698e-01 -2.77437449e-01
-1.15974617e+00 2.93879628e-01 -2.29830638e-01 -1.46432742e-01
-3.50848645e-01 7.75741398e-01 -4.62483168e-02 -4.58509445e-01
5.52522957e-01 -8.91591329e-03 1.02850400e-01 -2.13034973e-01
-3.48583221e-01 6.05915725e-01 4.61385280e-01 -2.73903579e-01
3.71573597e-01 5.89823306e-01 4.46724385e-01 -5.68349421e-01
-5.78212678e-01 -2.70892411e-01 -6.31057739e-01 -4.82246071e-01
1.02236223e+00 -6.31811082e-01 -6.61440969e-01 7.94102907e-01
-8.23794067e-01 -2.41325751e-01 1.22075878e-01 3.82513136e-01
-1.06836699e-01 -1.47350160e-02 -6.83724701e-01 -5.75082362e-01
-3.60878080e-01 -1.16741967e+00 1.05136907e+00 4.33273554e-01
3.00598592e-01 -1.06700838e+00 6.19489560e-03 -3.03287059e-03
5.18873155e-01 6.87437594e-01 1.27671385e+00 -4.51895982e-01
-6.79219246e-01 -4.95300263e-01 -4.74313974e-01 2.20955864e-01
1.35246769e-01 3.91668409e-01 -1.10192430e+00 -1.93607584e-01
-2.01703623e-01 -1.81002781e-01 1.15820897e+00 6.68291807e-01
1.62303579e+00 2.19246075e-01 -7.40392864e-01 7.70911455e-01
2.01814175e+00 1.84028447e-01 6.96715295e-01 4.02543217e-01
4.01299864e-01 2.94997483e-01 3.84816736e-01 2.78436303e-01
1.85069010e-01 3.94765586e-01 7.77130306e-01 -1.84590414e-01
-1.40438572e-01 2.69223481e-01 -4.12545919e-01 5.72950363e-01
-4.52518351e-02 -2.77590185e-01 -1.23692381e+00 6.35650694e-01
-1.89305747e+00 -7.28073120e-01 1.18021639e-02 1.82797468e+00
6.05303943e-01 -5.61584495e-02 -6.83848560e-01 2.47793749e-01
6.12789810e-01 -1.54350281e-01 -4.23244745e-01 -1.68743268e-01
-2.89565530e-02 1.14786923e-01 3.65794212e-01 2.03513265e-01
-1.40003145e+00 2.99872547e-01 5.61451101e+00 1.16633546e+00
-1.40727365e+00 1.14222817e-01 1.11276186e+00 1.12327047e-01
7.11237872e-03 -5.80851495e-01 -5.75071990e-01 3.35955054e-01
5.73819578e-01 1.66299507e-01 -3.27710241e-01 5.66401601e-01
-1.46149367e-01 -3.60359192e-01 -1.00718534e+00 8.35354626e-01
1.10160910e-01 -1.53901851e+00 -6.49210513e-02 9.27039310e-02
7.73809135e-01 -1.82295248e-01 1.56458199e-01 5.23568243e-02
-2.94641759e-02 -1.22334588e+00 8.05546120e-02 9.49360132e-01
8.90044034e-01 -9.25238192e-01 1.49654436e+00 2.41159439e-01
-9.38885987e-01 -1.96620211e-01 -2.63213485e-01 5.53532243e-01
-4.32966441e-01 8.42634201e-01 -9.34232295e-01 7.49616921e-01
9.64786291e-01 5.82098782e-01 -7.63178170e-01 1.15066063e+00
3.06556016e-01 4.57801819e-01 -2.44127706e-01 -1.62247524e-01
1.60671234e-01 2.57162780e-01 3.63155365e-01 1.45731199e+00
5.55342972e-01 1.24669157e-01 -1.47805586e-01 5.53127527e-01
-1.04674868e-01 7.01460391e-02 -1.97141483e-01 4.17042226e-02
1.28866181e-01 1.74502718e+00 -9.81739521e-01 -1.82071757e-02
-8.34335666e-03 5.49404800e-01 4.04466838e-01 1.05144180e-01
-6.61688089e-01 -2.06062391e-01 4.04577047e-01 -9.80151892e-02
1.72860742e-01 1.00271128e-01 -5.04953086e-01 -7.62809515e-01
-1.60612479e-01 -5.97763717e-01 4.18427616e-01 -5.58505714e-01
-1.35573709e+00 8.05728853e-01 -1.50716245e-01 -1.08243835e+00
1.20710634e-01 -8.69498253e-01 -4.75934535e-01 5.31769574e-01
-1.69500887e+00 -1.53639424e+00 -6.99761689e-01 2.99071193e-01
2.83899546e-01 -1.23487853e-01 1.15887952e+00 1.07803978e-01
-6.18483365e-01 6.89271033e-01 4.38842535e-01 1.73379481e-01
7.23140895e-01 -1.49959934e+00 -6.98565304e-01 6.62939370e-01
-4.30870146e-01 3.05527210e-01 4.23223168e-01 -4.36565310e-01
-1.07741916e+00 -1.26214659e+00 6.47855341e-01 -8.65969360e-02
4.50299114e-01 -2.16708377e-01 -8.40416133e-01 3.48584861e-01
2.48442933e-01 5.83366930e-01 1.09900045e+00 -1.39882281e-01
-1.90890476e-01 -6.31401360e-01 -1.67987442e+00 3.03634554e-01
4.93838996e-01 -3.17214876e-01 -1.01538124e-02 -1.94698479e-02
2.07740381e-01 -4.98421878e-01 -1.11628842e+00 9.46354687e-01
8.95845234e-01 -1.13588083e+00 6.15086615e-01 -1.57535866e-01
5.73372424e-01 -3.63815516e-01 -1.08423688e-01 -1.40148985e+00
-4.89219904e-01 4.31087583e-01 8.63715112e-02 1.00215471e+00
2.67985582e-01 -5.04938722e-01 7.92433202e-01 2.90367395e-01
-2.59292960e-01 -1.16888535e+00 -1.33478510e+00 -4.34911817e-01
1.92652628e-01 -1.73890129e-01 5.27476132e-01 8.10151875e-01
2.76136640e-02 -3.44322771e-01 9.00294259e-02 4.09762293e-01
7.92509735e-01 -8.26831758e-02 1.45740494e-01 -1.28647721e+00
-1.63722932e-01 -5.77477157e-01 -1.17468071e+00 3.43040265e-02
-4.71297791e-03 -8.45252812e-01 9.36679468e-02 -1.35433578e+00
6.14507198e-01 -6.46487892e-01 -1.06131625e+00 4.09464598e-01
-2.23060504e-01 5.83178818e-01 -2.18976215e-01 -6.49923459e-02
-5.07022798e-01 9.48315039e-02 1.13300467e+00 -6.16930366e-01
2.01193944e-01 -2.78179079e-01 -5.26694298e-01 4.62002277e-01
8.24591219e-01 -3.14593077e-01 -1.71687663e-01 -2.29012832e-01
-1.66137982e-02 -1.82353240e-02 6.31778777e-01 -1.42983913e+00
3.74531746e-01 -1.29024491e-01 9.98336017e-01 -8.46196890e-01
1.29326433e-01 -1.17319500e+00 5.65163851e-01 5.73681831e-01
-3.27412605e-01 -4.81845856e-01 4.91166085e-01 8.37592900e-01
-4.44049269e-01 -2.97227819e-02 9.57929313e-01 -1.54761091e-01
-5.71566164e-01 3.10657710e-01 -3.38078141e-01 -7.86770523e-01
1.40449750e+00 -5.70613742e-02 -3.68547559e-01 2.06734896e-01
-8.09620678e-01 6.64229318e-02 2.94232070e-01 6.06012829e-02
5.59645712e-01 -1.37095785e+00 -6.62491977e-01 1.29313409e-01
4.81944561e-01 1.62799060e-01 9.43157017e-01 9.58901227e-01
-6.71503723e-01 2.73489654e-01 -3.45095277e-01 -1.07563889e+00
-1.43920290e+00 1.93505809e-01 9.19401646e-01 -7.38867939e-01
-2.15799823e-01 1.06559229e+00 9.71697196e-02 -1.04799710e-01
1.58517942e-01 -5.31496048e-01 -5.00824749e-01 -7.82572106e-02
5.78662336e-01 9.11904201e-02 5.60421169e-01 -3.23868752e-01
-4.48789269e-01 4.64496106e-01 -3.73051137e-01 4.02836531e-01
1.38248789e+00 1.88471243e-01 -2.97969729e-01 4.12263423e-01
1.25974929e+00 -3.79031986e-01 -1.21240580e+00 -2.66542342e-02
-1.32971397e-02 -3.36459041e-01 5.07984698e-01 -1.28732669e+00
-1.15584695e+00 9.15345728e-01 1.19260538e+00 -2.38111541e-01
1.33350241e+00 -2.58444548e-01 2.97223955e-01 2.66523331e-01
2.77050525e-01 -9.67676461e-01 -4.40541655e-02 3.84329781e-02
7.84092426e-01 -1.57246315e+00 5.15913665e-02 -4.55516815e-01
-1.51603892e-01 1.47641468e+00 7.34080493e-01 -4.73460965e-02
1.13760984e+00 6.51991069e-01 3.72695297e-01 -1.34562358e-01
-8.99359882e-01 -1.05518781e-01 2.13262260e-01 5.83407938e-01
4.38072473e-01 2.12289274e-01 -2.41263673e-01 8.04205716e-01
4.20555949e-01 3.33072186e-01 1.86285689e-01 9.82276976e-01
-3.56863946e-01 -8.47253263e-01 -2.28921846e-01 4.88027185e-01
-6.13403380e-01 1.54872701e-01 -1.63629711e-01 9.81958866e-01
4.55870062e-01 5.78048885e-01 1.53207287e-01 -4.16436285e-01
-5.40096462e-02 1.03567354e-01 5.44517517e-01 -2.97722578e-01
-5.99876881e-01 -4.16295975e-03 -3.11836064e-01 -2.65391022e-01
-5.40304840e-01 -5.55021107e-01 -1.23317087e+00 1.59910411e-01
-2.71999300e-01 -6.14395291e-02 1.11878312e+00 8.61489773e-01
3.17660809e-01 1.00182414e+00 4.42771792e-01 -6.96213901e-01
-4.17280346e-01 -1.05136836e+00 -6.56939924e-01 -8.99534598e-02
5.60740948e-01 -6.92077935e-01 -2.22297311e-01 4.87091541e-02] | [15.07246208190918, -2.918498992919922] |
198e0339-8248-41d7-81ef-affc4109ad2b | pyramid-diffusion-models-for-low-light-image | 2305.10028 | null | https://arxiv.org/abs/2305.10028v1 | https://arxiv.org/pdf/2305.10028v1.pdf | Pyramid Diffusion Models For Low-light Image Enhancement | Recovering noise-covered details from low-light images is challenging, and the results given by previous methods leave room for improvement. Recent diffusion models show realistic and detailed image generation through a sequence of denoising refinements and motivate us to introduce them to low-light image enhancement for recovering realistic details. However, we found two problems when doing this, i.e., 1) diffusion models keep constant resolution in one reverse process, which limits the speed; 2) diffusion models sometimes result in global degradation (e.g., RGB shift). To address the above problems, this paper proposes a Pyramid Diffusion model (PyDiff) for low-light image enhancement. PyDiff uses a novel pyramid diffusion method to perform sampling in a pyramid resolution style (i.e., progressively increasing resolution in one reverse process). Pyramid diffusion makes PyDiff much faster than vanilla diffusion models and introduces no performance degradation. Furthermore, PyDiff uses a global corrector to alleviate the global degradation that may occur in the reverse process, significantly improving the performance and making the training of diffusion models easier with little additional computational consumption. Extensive experiments on popular benchmarks show that PyDiff achieves superior performance and efficiency. Moreover, PyDiff can generalize well to unseen noise and illumination distributions. | ['Yi Yang', 'Zongxin Yang', 'Dewei Zhou'] | 2023-05-17 | null | null | null | null | ['image-enhancement', 'low-light-image-enhancement'] | ['computer-vision', 'computer-vision'] | [ 2.39282787e-01 -3.71569723e-01 2.26384684e-01 -4.74475659e-02
-5.44626176e-01 -1.56983435e-01 3.50109249e-01 -3.01913172e-01
-3.37194413e-01 6.84380710e-01 3.11065525e-01 9.39828753e-02
2.32843935e-01 -1.09533405e+00 -5.71896434e-01 -1.05782819e+00
4.07906294e-01 -2.28978559e-01 7.92659342e-01 -2.77522892e-01
2.90783495e-01 5.49501896e-01 -1.34367168e+00 4.34108168e-01
9.30122197e-01 7.19879150e-01 5.66170990e-01 7.48884320e-01
-3.21561128e-01 1.04496288e+00 -4.93337423e-01 -3.38068336e-01
5.76867044e-01 -6.78342700e-01 -6.50627911e-01 3.17309856e-01
5.24486005e-01 -6.73843861e-01 -4.17986780e-01 1.27156079e+00
6.72145605e-01 1.88233569e-01 3.60215068e-01 -8.09799254e-01
-1.10460091e+00 1.23397946e-01 -1.19423187e+00 3.37997764e-01
1.73397675e-01 4.78502870e-01 6.75668240e-01 -1.01700652e+00
6.13729000e-01 1.25973058e+00 9.03866708e-01 5.79936445e-01
-1.48313856e+00 -6.53953493e-01 3.86944525e-02 1.38616666e-01
-1.37028921e+00 -4.49526101e-01 7.13357210e-01 -4.37049791e-02
7.59741843e-01 2.67646790e-01 6.97753310e-01 7.95550644e-01
2.77917057e-01 6.90071881e-01 1.52219296e+00 -2.44252458e-01
2.18652070e-01 -1.89328238e-01 -1.34066179e-01 6.25571430e-01
3.66702527e-01 2.40151361e-01 -4.98426050e-01 8.26212242e-02
1.25408673e+00 1.14221118e-01 -6.54474437e-01 2.02444778e-03
-9.95225549e-01 5.79263568e-01 6.92935109e-01 3.87503952e-01
-5.46532631e-01 1.44971848e-01 -3.81562859e-02 3.58043790e-01
6.42027855e-01 1.78161547e-01 -1.37701467e-01 1.42069519e-01
-1.13691306e+00 1.56617388e-01 4.06226635e-01 8.43080163e-01
1.14210653e+00 5.77980131e-02 -2.77506381e-01 1.07993257e+00
5.66955246e-02 5.41716695e-01 2.46773452e-01 -1.40861666e+00
2.64222443e-01 1.68909892e-01 1.56313837e-01 -1.05887318e+00
-2.29855910e-01 -2.15444863e-01 -1.33820617e+00 7.32537925e-01
4.33980644e-01 8.47474113e-02 -1.02095449e+00 1.38213921e+00
2.81245917e-01 3.21543366e-01 1.18977893e-02 9.60304916e-01
7.30619788e-01 1.00161564e+00 -1.54002666e-01 -4.41669554e-01
1.12260592e+00 -1.31692719e+00 -9.17819381e-01 -1.23438381e-01
2.41817057e-01 -1.09491372e+00 1.16335440e+00 7.07870185e-01
-1.52727401e+00 -6.06039584e-01 -7.93069839e-01 -4.32325602e-01
1.08066034e-02 -2.70061761e-01 6.31242216e-01 5.69132745e-01
-1.41911221e+00 8.36609304e-01 -7.50655651e-01 -1.87767774e-01
3.87222081e-01 4.36105430e-02 -1.33017287e-01 -6.47065759e-01
-8.57722223e-01 6.33900702e-01 9.36179608e-02 2.62548681e-02
-5.95533431e-01 -6.87920928e-01 -5.45505404e-01 -6.42452016e-02
2.94893026e-01 -9.21654642e-01 9.10616159e-01 -8.76268744e-01
-1.56609964e+00 4.94482934e-01 -4.29715961e-01 -1.36833429e-01
6.76314712e-01 -1.23661488e-01 -1.85641035e-01 2.56163090e-01
2.92795803e-02 7.51700103e-01 1.08372629e+00 -1.55789685e+00
-6.92737877e-01 -1.85824037e-01 7.27920532e-02 2.21513256e-01
-3.21970940e-01 -1.60140485e-01 -9.03590560e-01 -8.50408196e-01
2.52629817e-01 -6.76607132e-01 -3.90071929e-01 4.70191002e-01
-8.06083456e-02 1.00534454e-01 8.31050754e-01 -7.21367478e-01
1.26172829e+00 -2.28086996e+00 -4.17899899e-02 -1.32719889e-01
4.66510683e-01 3.30243737e-01 -3.25988203e-01 1.19793728e-01
2.03371689e-01 1.78389270e-02 -5.36234200e-01 -5.61370969e-01
-4.48693216e-01 5.34518838e-01 -2.02044576e-01 3.26318562e-01
8.50180686e-02 6.86827958e-01 -8.95210564e-01 -4.85453397e-01
1.59761176e-01 8.41578424e-01 -8.06609213e-01 7.61886835e-02
8.68707821e-02 4.67723161e-01 -9.38250050e-02 5.78317165e-01
1.28851891e+00 -1.82970479e-01 -3.01920801e-01 -5.36982715e-01
-3.59512568e-01 -1.49388984e-01 -1.46381724e+00 1.50023246e+00
-6.74351156e-01 6.92905188e-01 3.49904090e-01 -4.45922613e-01
7.47655094e-01 -6.87043890e-02 2.80765831e-01 -8.87041688e-01
-1.79781064e-01 2.56088078e-01 -3.60090852e-01 -4.18231010e-01
7.15198040e-01 -3.46103877e-01 7.58885264e-01 4.93395239e-01
-3.90172094e-01 -5.02418220e-01 2.40097255e-01 1.91834733e-01
1.10357940e+00 1.15156081e-02 3.50489467e-02 -1.42869800e-02
3.89975131e-01 -1.88787460e-01 8.25761378e-01 8.30323398e-01
-2.15571180e-01 1.09483159e+00 9.27241985e-03 -3.52460474e-01
-1.00721145e+00 -1.01840174e+00 2.23963782e-02 7.13246286e-01
6.04670405e-01 -2.76428461e-01 -6.84946716e-01 -4.34174001e-01
-3.52435678e-01 6.00026190e-01 -4.34300780e-01 2.45766044e-02
-6.66910231e-01 -1.06969047e+00 2.54788429e-01 5.15449226e-01
1.11896765e+00 -8.44176114e-01 -3.69598985e-01 2.04947278e-01
-4.70041692e-01 -1.19425344e+00 -6.85635030e-01 -6.63695782e-02
-1.09462214e+00 -7.53718555e-01 -1.19123781e+00 -7.31879294e-01
6.93153799e-01 1.02045751e+00 1.00416434e+00 3.41761976e-01
-3.15923244e-01 2.89779127e-01 -2.59792536e-01 4.12998609e-02
-2.80236870e-01 -5.64408779e-01 -3.20819139e-01 -3.95682175e-03
1.65473651e-02 -5.87577581e-01 -1.11532497e+00 4.21254545e-01
-1.32721472e+00 3.13418776e-01 6.78014815e-01 9.87964272e-01
8.56792927e-01 6.51789248e-01 -9.47662592e-02 -7.01650858e-01
7.77238965e-01 -1.19975291e-01 -4.87745404e-01 3.60195190e-02
-8.55329275e-01 2.80923350e-03 6.09310508e-01 -6.34211779e-01
-1.58438230e+00 -2.30439737e-01 -3.24062675e-01 -3.68408203e-01
1.34569956e-02 -8.63255188e-02 1.63246125e-01 -3.89455229e-01
6.81553185e-01 5.28832376e-01 -1.30189285e-01 -7.02848077e-01
3.85344505e-01 3.70792717e-01 4.77615416e-01 -3.22530657e-01
8.86865318e-01 1.03270149e+00 5.62262535e-02 -9.44812417e-01
-7.83999085e-01 -3.15617949e-01 -3.77616972e-01 -1.42421812e-01
8.56122017e-01 -8.95517647e-01 -4.33179170e-01 7.76443303e-01
-1.14317906e+00 -7.22293973e-01 -5.09850919e-01 3.27634186e-01
-3.91349137e-01 7.64045298e-01 -1.14207995e+00 -4.86483067e-01
-2.69615412e-01 -1.25128996e+00 1.03609169e+00 4.42664325e-01
1.96050227e-01 -1.00750470e+00 -1.17087267e-01 3.04274350e-01
7.94650912e-01 -1.96179733e-01 6.33564532e-01 6.87905967e-01
-9.40673947e-01 3.48383069e-01 -7.62487233e-01 6.74558401e-01
2.12176964e-01 -1.62161469e-01 -9.37364399e-01 -2.99692094e-01
3.66724700e-01 -1.82326473e-02 1.15690362e+00 4.75291610e-01
1.16569507e+00 -1.19857199e-01 -9.24831331e-02 9.83407795e-01
1.58827364e+00 -1.83156529e-03 1.08822536e+00 4.63954419e-01
6.32221818e-01 3.30430359e-01 5.65860927e-01 1.35410607e-01
3.07571769e-01 6.16988778e-01 3.00567925e-01 -7.75453925e-01
-1.04773915e+00 -1.66724846e-02 4.55735862e-01 7.36119688e-01
-3.76500338e-01 -9.95512083e-02 -4.87599909e-01 3.90196770e-01
-1.46845710e+00 -1.01727808e+00 -5.47908008e-01 2.00800824e+00
1.06263936e+00 -3.08776591e-02 -2.04280719e-01 2.19474584e-01
5.88938534e-01 3.18876714e-01 -5.25045335e-01 -2.21411139e-01
-5.29703796e-01 1.40617192e-01 6.84695542e-01 7.62226522e-01
-5.36738217e-01 9.28007603e-01 6.26243114e+00 1.01144826e+00
-1.12367499e+00 4.03063923e-01 7.17297554e-01 -5.47973365e-02
-4.27838624e-01 2.26618443e-02 -7.21043169e-01 3.16366494e-01
1.73638105e-01 1.48930460e-01 6.89689934e-01 3.51503253e-01
3.83775115e-01 -4.64410722e-01 -4.35372561e-01 1.19989467e+00
1.29623622e-01 -1.17187715e+00 1.51489466e-01 1.09503962e-01
1.09180832e+00 2.36060843e-02 2.18427498e-02 -2.02818289e-02
3.82298142e-01 -5.51136434e-01 5.48871458e-01 5.30594349e-01
7.47095168e-01 -5.58010638e-01 5.41155815e-01 4.37634856e-01
-1.08583522e+00 -5.51123917e-02 -7.71143198e-01 1.24535978e-01
3.96471113e-01 9.85447824e-01 -2.07885310e-01 4.39721644e-01
9.93346334e-01 7.42762148e-01 -5.02639174e-01 1.12844431e+00
-3.92105401e-01 4.79934335e-01 -2.80240089e-01 4.24870163e-01
1.41983420e-01 -5.12337923e-01 6.24836385e-01 1.27883124e+00
4.88751113e-01 2.89090544e-01 -6.48205802e-02 8.44228983e-01
-5.18440716e-02 -1.31419748e-01 -4.20081377e-01 4.81529236e-01
-3.05108004e-03 1.13627517e+00 -8.96078229e-01 -2.80148715e-01
-6.41318977e-01 1.59744525e+00 1.79818086e-02 6.98907256e-01
-6.10829115e-01 -3.18599224e-01 5.21758616e-01 4.35022384e-01
3.42507869e-01 -3.38238150e-01 -4.24368382e-01 -1.26751757e+00
-7.71068707e-02 -6.73619211e-01 2.13753030e-01 -1.13007236e+00
-1.36526859e+00 6.32312059e-01 -2.23893598e-01 -1.09194815e+00
3.88884097e-01 -2.70477593e-01 -4.21864510e-01 9.38128710e-01
-2.19930530e+00 -9.06691909e-01 -7.08100200e-01 8.33202183e-01
9.15694535e-01 5.65441966e-01 3.03507447e-01 5.31214416e-01
-3.53207141e-01 2.31152788e-01 1.29476428e-01 -2.45067358e-01
9.10876155e-01 -1.14127982e+00 3.15120995e-01 1.10479116e+00
-1.32234290e-01 4.48912621e-01 6.20174229e-01 -6.48149848e-01
-1.16705012e+00 -9.96942103e-01 7.47575104e-01 3.08918003e-02
4.02969986e-01 4.64314297e-02 -1.18990171e+00 3.00481141e-01
3.04567099e-01 1.36249751e-01 2.39618018e-01 -3.85684282e-01
-2.57409841e-01 -2.45458931e-01 -1.31925213e+00 8.19573224e-01
1.34277725e+00 -3.57945681e-01 -1.83916941e-01 2.04251364e-01
4.97071117e-01 -4.68284518e-01 -6.99288666e-01 2.25605533e-01
2.83293843e-01 -1.53927565e+00 1.20052254e+00 3.13160568e-01
5.02082527e-01 -6.65195107e-01 -1.02171153e-01 -1.33717906e+00
-6.30858779e-01 -6.81349695e-01 -2.59394497e-01 1.26870871e+00
9.46888421e-03 -7.10600436e-01 3.36221844e-01 5.39955080e-01
1.25054330e-01 -6.26643598e-01 -5.81738234e-01 -8.99124980e-01
-7.47766271e-02 -2.69892484e-01 4.94782478e-01 9.32908833e-01
-8.47454309e-01 1.34147406e-01 -5.36228836e-01 2.21014470e-01
9.81479049e-01 9.99490693e-02 6.86129272e-01 -8.90159309e-01
-4.14753884e-01 -4.75616515e-01 -9.22210421e-03 -1.70967424e+00
-4.58336204e-01 -4.75015998e-01 1.15143910e-01 -1.84290934e+00
3.33549798e-01 -5.05956233e-01 2.94162799e-02 2.95176864e-01
-5.17911017e-01 7.50195861e-01 2.71500498e-01 5.69559574e-01
-2.78058320e-01 5.08615375e-01 1.84278619e+00 -2.24128485e-01
-2.99204797e-01 -1.39119551e-01 -8.52133572e-01 8.94188106e-01
6.85700536e-01 -5.23775041e-01 -4.05610085e-01 -9.18607712e-01
1.65926456e-01 -1.01726130e-01 3.11698496e-01 -9.93264735e-01
3.63139302e-01 -1.90112203e-01 5.32695472e-01 -3.18638325e-01
5.20037532e-01 -8.20315003e-01 2.80074533e-02 4.86953259e-01
8.18467438e-02 -7.29858950e-02 1.50463372e-01 7.05520570e-01
-3.18781525e-01 -2.47062162e-01 1.31598842e+00 -2.56573975e-01
-7.41550863e-01 2.46154860e-01 -2.99427271e-01 -1.27563313e-01
8.10930550e-01 -5.51370680e-01 -3.11271369e-01 -5.71791887e-01
-6.75399184e-01 -2.84009036e-02 8.66468549e-01 5.53292260e-02
7.00891256e-01 -1.08800137e+00 -7.36790597e-01 2.26386592e-01
-3.25961351e-01 9.31362957e-02 5.14529347e-01 9.62889850e-01
-7.04503596e-01 -3.87696117e-01 4.35524546e-02 -5.25961936e-01
-1.33714867e+00 4.34469134e-01 3.11976880e-01 -3.00181359e-01
-1.11411870e+00 9.96641576e-01 5.78763008e-01 5.80878370e-02
3.18983644e-02 -3.25702250e-01 5.90285547e-02 -1.74270079e-01
9.47309017e-01 5.63947797e-01 -1.25134187e-02 -3.60655397e-01
6.33253306e-02 9.88763154e-01 -1.51100203e-01 -7.34135807e-02
1.49330997e+00 -6.45659804e-01 -2.44632006e-01 7.30909556e-02
8.97203565e-01 3.94060463e-01 -1.59081793e+00 -4.10873890e-01
-4.93816197e-01 -1.01114082e+00 6.40095294e-01 -7.31083095e-01
-1.44824934e+00 1.00444961e+00 6.53579533e-01 2.48783275e-01
1.76886106e+00 -3.20724368e-01 1.27027619e+00 2.74049956e-02
2.69543886e-01 -9.51654732e-01 2.50788599e-01 4.22826976e-01
8.29262674e-01 -1.14056420e+00 2.27831751e-01 -6.31345034e-01
-6.94562256e-01 1.15887904e+00 5.13049126e-01 1.73509583e-01
5.04148364e-01 5.77071190e-01 1.48768842e-01 -7.54874647e-02
-5.91243207e-01 -3.81588966e-01 -1.31442606e-01 7.11783946e-01
2.45979249e-01 -4.00011122e-01 -3.34144741e-01 4.71618230e-04
3.68051171e-01 2.54868656e-01 6.46569431e-01 7.70830095e-01
-5.42784452e-01 -1.03675091e+00 -7.22113907e-01 7.65995681e-02
-3.78020942e-01 -3.79969329e-01 -5.30469604e-02 5.16182423e-01
2.43124560e-01 1.18038273e+00 -8.14523324e-02 -2.39722505e-02
4.56025779e-01 -5.41910827e-01 6.06885910e-01 -2.85328478e-01
-3.81931335e-01 3.41381043e-01 -3.43529344e-01 -6.90055430e-01
-5.76773345e-01 -2.46691182e-01 -1.13065994e+00 -5.27957797e-01
-3.58148307e-01 -1.33223951e-01 4.60755348e-01 4.44565535e-01
3.33328396e-01 5.98340929e-01 5.41621983e-01 -1.09985566e+00
-2.61362761e-01 -7.51819551e-01 -7.14540243e-01 5.36627948e-01
4.55700457e-01 -5.29645979e-01 -5.32488942e-01 2.47728139e-01] | [11.022858619689941, -2.3973701000213623] |
58f93127-af64-487d-926c-b7fa273acbf0 | domain-adversarial-spatial-temporal-network-a | 2202.03630 | null | https://arxiv.org/abs/2202.03630v2 | https://arxiv.org/pdf/2202.03630v2.pdf | Domain Adversarial Spatial-Temporal Network: A Transferable Framework for Short-term Traffic Forecasting across Cities | Accurate real-time traffic forecast is critical for intelligent transportation systems (ITS) and it serves as the cornerstone of various smart mobility applications. Though this research area is dominated by deep learning, recent studies indicate that the accuracy improvement by developing new model structures is becoming marginal. Instead, we envision that the improvement can be achieved by transferring the "forecasting-related knowledge" across cities with different data distributions and network topologies. To this end, this paper aims to propose a novel transferable traffic forecasting framework: Domain Adversarial Spatial-Temporal Network (DASTNet). DASTNet is pre-trained on multiple source networks and fine-tuned with the target network's traffic data. Specifically, we leverage the graph representation learning and adversarial domain adaptation techniques to learn the domain-invariant node embeddings, which are further incorporated to model the temporal traffic data. To the best of our knowledge, we are the first to employ adversarial multi-domain adaptation for network-wide traffic forecasting problems. DASTNet consistently outperforms all state-of-the-art baseline methods on three benchmark datasets. The trained DASTNet is applied to Hong Kong's new traffic detectors, and accurate traffic predictions can be delivered immediately (within one day) when the detector is available. Overall, this study suggests an alternative to enhance the traffic forecasting methods and provides practical implications for cities lacking historical traffic data. | ['Wei Ma', 'S. C. Wong', 'William H. K. Lam', 'Andy H. F. Chow', 'Ao Qu', 'Yihong Tang'] | 2022-02-08 | null | null | null | null | ['time-series-regression', 'spatio-temporal-forecasting'] | ['time-series', 'time-series'] | [-1.84045862e-02 -2.67212149e-02 -4.92205977e-01 -3.70705247e-01
-4.93521452e-01 -2.66980231e-01 8.15650642e-01 -4.44723904e-01
-9.95066836e-02 8.93652737e-01 2.70039201e-01 -8.52363527e-01
-1.47599205e-01 -1.25523055e+00 -5.62421083e-01 -6.50579810e-01
7.88149983e-02 7.46426523e-01 6.22258663e-01 -6.98189020e-01
-2.44631842e-01 7.98302710e-01 -1.24159873e+00 -8.65792297e-03
1.10187900e+00 9.35847461e-01 -1.62350297e-01 5.19497931e-01
-3.96632612e-01 8.51403236e-01 -4.66001630e-01 -6.22764289e-01
3.97559583e-01 1.82154194e-01 -6.09407067e-01 -1.88780382e-01
4.13306415e-01 -4.82637137e-01 -1.20161235e+00 6.15560770e-01
2.97883153e-01 3.83863240e-01 8.05508792e-01 -1.88140500e+00
-7.66258121e-01 2.01890558e-01 -2.34135032e-01 3.52528155e-01
-2.93193817e-01 4.00504291e-01 6.57373607e-01 -4.22343940e-01
4.16383535e-01 1.23054266e+00 8.11126709e-01 6.32961571e-01
-9.71402168e-01 -1.04952216e+00 4.55797911e-01 3.88585120e-01
-1.25001717e+00 -5.40918827e-01 9.87923086e-01 -5.50876498e-01
7.51582861e-01 -4.72771190e-02 2.58100718e-01 1.42714250e+00
2.05892161e-01 6.42861187e-01 5.79862416e-01 1.17389694e-01
8.37860480e-02 1.11897640e-01 -1.01836234e-01 4.88845438e-01
9.29007158e-02 4.54429179e-01 1.39126778e-01 1.74824983e-01
4.24996763e-01 1.89573348e-01 3.22689831e-01 -2.79930755e-02
-8.52673709e-01 8.42636228e-01 7.25468516e-01 2.48294234e-01
-5.76935470e-01 3.65487605e-01 6.16373241e-01 3.71573716e-01
9.91266549e-01 -2.68425852e-01 -3.59489292e-01 -1.02735899e-01
-8.58606935e-01 5.14630266e-02 5.84995210e-01 9.66972649e-01
6.88424349e-01 6.74242914e-01 -6.54248223e-02 7.23671198e-01
1.78844646e-01 9.26378071e-01 7.88237825e-02 -7.55399287e-01
8.03966284e-01 5.02137601e-01 -4.62090932e-02 -1.23173428e+00
-4.62842494e-01 -3.67619991e-01 -9.06125307e-01 7.86048770e-02
4.40667331e-01 -4.96929049e-01 -9.75753784e-01 1.60977483e+00
2.95650542e-01 8.23453903e-01 5.71999401e-02 5.60268044e-01
6.88737094e-01 8.34760487e-01 4.00410295e-01 4.24804717e-01
7.41364300e-01 -1.11769295e+00 -4.65794861e-01 -2.67019928e-01
7.54848063e-01 -5.09074867e-01 5.97523332e-01 -1.88837931e-01
-5.92728078e-01 -6.82832122e-01 -6.44868433e-01 8.08685347e-02
-1.00603926e+00 -1.53527379e-01 5.54290473e-01 8.45508099e-01
-1.03634191e+00 7.54811242e-02 -7.49228477e-01 -6.27792001e-01
7.91000843e-01 3.85758907e-01 -1.23162597e-01 -5.08480728e-01
-1.60381734e+00 1.03470373e+00 8.65212977e-02 7.10828751e-02
-8.76171052e-01 -8.62144768e-01 -6.59053504e-01 -1.59542039e-01
1.89891919e-01 -7.77101696e-01 9.67741966e-01 -8.54985774e-01
-1.41753411e+00 3.44547242e-01 -2.15348452e-01 -6.90904319e-01
3.64880025e-01 3.58178824e-01 -1.44674599e+00 -1.56067163e-01
3.60526443e-01 6.18162334e-01 8.59363198e-01 -1.08464026e+00
-8.11927378e-01 -6.69138432e-02 9.13993344e-02 -3.35782796e-01
-4.74881649e-01 -1.64869934e-01 -2.55222589e-01 -5.97304106e-01
-5.72966635e-01 -1.01939404e+00 -3.53944063e-01 1.29517674e-01
-1.87211335e-01 -3.42652768e-01 1.39603484e+00 -6.66671515e-01
1.20006168e+00 -1.88203955e+00 -4.40478444e-01 3.58220160e-01
2.25321084e-01 8.30180228e-01 -5.15859306e-01 5.29259026e-01
5.19068651e-02 -3.28655494e-03 -6.90214857e-02 -1.80052444e-01
1.69156834e-01 5.31467080e-01 -5.14879525e-01 3.42249751e-01
2.14849204e-01 1.19875932e+00 -1.01213431e+00 -2.76165485e-01
6.15739167e-01 6.38669252e-01 -3.73566002e-01 -6.87958160e-03
-1.85153633e-01 7.12330461e-01 -7.49054253e-01 6.05007708e-01
8.87748837e-01 -4.63042073e-02 -1.54825017e-01 -8.71957317e-02
-5.89272156e-02 2.06938803e-01 -7.15498686e-01 1.11082983e+00
-7.44438767e-01 1.01670420e+00 -1.08565055e-01 -1.43623054e+00
1.17427897e+00 2.11759105e-01 7.39041984e-01 -1.15346289e+00
-2.06238013e-02 1.70572266e-01 -1.61338583e-01 -4.03560936e-01
4.56902146e-01 -6.40346557e-02 -9.88577306e-02 1.89369664e-01
-3.27754557e-01 2.02825278e-01 -3.91373374e-02 1.44470766e-01
1.11748159e+00 -2.81217337e-01 -3.31158608e-01 1.39470458e-01
7.07843840e-01 2.79918820e-01 4.51449841e-01 4.04992610e-01
-6.35591865e-01 -3.90919000e-02 2.41232023e-01 -7.70573616e-01
-1.00595522e+00 -1.22198665e+00 3.05881053e-02 1.28238165e+00
-1.19804025e-01 1.37213245e-01 -3.89526188e-01 -1.05537653e+00
3.02884787e-01 8.81241977e-01 -5.49199462e-01 -2.27088273e-01
-8.03234041e-01 -2.44236156e-01 9.11655247e-01 7.58858502e-01
6.56741619e-01 -5.56746840e-01 4.98512715e-01 4.43684936e-01
-4.48412970e-02 -1.45334399e+00 -5.53228021e-01 -4.65549946e-01
-6.49562657e-01 -9.48710501e-01 -6.93888783e-01 -6.73415482e-01
5.49051285e-01 6.30751848e-01 9.96326983e-01 -3.49427909e-02
2.44357139e-01 4.23629284e-01 -1.89311713e-01 -3.83966506e-01
-4.64819342e-01 4.57172841e-01 1.41266063e-01 2.28347972e-01
5.82314610e-01 -7.41961122e-01 -5.96988618e-01 7.13059664e-01
-6.32576525e-01 -2.05324143e-01 3.99767905e-01 6.18506968e-01
1.41327083e-01 4.31300014e-01 1.04137790e+00 -8.62568736e-01
5.42295635e-01 -1.07441473e+00 -4.71337497e-01 3.67161036e-02
-5.62550664e-01 -1.55358657e-01 1.03004181e+00 -3.35831404e-01
-1.29142714e+00 -2.96679348e-01 -3.74555528e-01 -4.90784019e-01
-3.62488896e-01 3.79513025e-01 -2.34277453e-02 -3.08491886e-01
5.16395688e-01 1.55253664e-01 -4.78019416e-02 -9.33594033e-02
5.94803810e-01 8.20670009e-01 3.86609048e-01 -5.29243112e-01
1.29374230e+00 7.42449999e-01 2.87002206e-01 -8.96372199e-01
-6.32054031e-01 -3.46473187e-01 -6.32963002e-01 -4.94054735e-01
7.50541687e-01 -9.17287648e-01 -4.79038835e-01 3.62861514e-01
-9.49150324e-01 -5.30514181e-01 -1.11306705e-01 3.84743869e-01
-1.97513342e-01 1.42292738e-01 -1.96228758e-01 -6.13633394e-01
4.76509659e-03 -9.66912210e-01 7.92217672e-01 -5.65666966e-02
2.51061589e-01 -1.75651157e+00 8.50321501e-02 5.60023725e-01
1.01676261e+00 2.98256636e-01 6.20989442e-01 -7.66358256e-01
-7.19892859e-01 -3.48384231e-01 -5.58425188e-01 4.26586568e-01
2.80852556e-01 9.86342207e-02 -9.21829462e-01 -7.52603561e-02
-8.47959816e-01 2.09123999e-01 9.17845726e-01 4.61612254e-01
1.14974368e+00 -3.95190656e-01 -6.98125124e-01 5.31035483e-01
9.75705624e-01 2.60608703e-01 7.52997279e-01 3.62914056e-01
9.33385134e-01 7.03712583e-01 5.00948668e-01 4.35522683e-02
9.92673039e-01 7.41336167e-01 4.90286261e-01 -3.51411730e-01
-5.10411084e-01 -5.55180788e-01 4.24631715e-01 7.42632210e-01
1.28034679e-02 -6.44286752e-01 -1.21459317e+00 7.69326568e-01
-1.88228977e+00 -1.23486328e+00 -3.08974743e-01 1.76336861e+00
-1.98651981e-02 3.09531838e-01 4.59817410e-01 -1.49112597e-01
7.76053846e-01 4.20153320e-01 -7.10528135e-01 -5.42334855e-01
7.86321759e-02 2.28400439e-01 9.86466765e-01 4.33981806e-01
-1.05854845e+00 1.26736999e+00 5.96857309e+00 1.08361244e+00
-1.28450775e+00 2.72787243e-01 4.57811207e-01 2.67502874e-01
-3.17929447e-01 -2.15496689e-01 -7.37676680e-01 7.93334544e-01
1.65945196e+00 -7.96540231e-02 4.68900621e-01 8.34632456e-01
6.29900157e-01 5.54781318e-01 -6.31817698e-01 5.49769938e-01
-2.83369869e-01 -1.54089630e+00 1.03118278e-01 3.29913437e-01
9.22215164e-01 5.12504578e-01 3.61446112e-01 8.32735658e-01
6.22501075e-01 -1.01638091e+00 -3.92199401e-03 7.48952270e-01
7.87400246e-01 -8.96823466e-01 6.83802128e-01 1.90820917e-01
-1.71482110e+00 -2.01161757e-01 -2.35815808e-01 1.42825115e-03
5.29550493e-01 3.43828976e-01 -9.15098786e-01 6.28503978e-01
1.85019374e-01 1.19480407e+00 -3.35462689e-01 8.02473307e-01
7.61916265e-02 9.73939300e-01 -1.28987104e-01 1.55750006e-01
5.98719060e-01 -1.52003691e-01 4.67862934e-01 1.15971351e+00
3.79170805e-01 -4.96951938e-02 2.84688562e-01 5.93770146e-01
-2.49252513e-01 -3.01674336e-01 -1.06287515e+00 7.40796700e-02
7.47826815e-01 9.16618645e-01 -1.97441667e-01 -3.34887266e-01
-7.96718776e-01 6.41437709e-01 2.00934187e-01 7.07123578e-01
-1.16171873e+00 -3.30876648e-01 1.09602582e+00 3.84776920e-01
3.93396378e-01 -4.38734502e-01 -2.99591154e-01 -8.25519562e-01
-4.34820384e-01 -4.29621786e-01 2.09212959e-01 -5.82698822e-01
-1.66217017e+00 5.10865808e-01 -6.57765120e-02 -1.45981443e+00
-7.59529769e-02 -7.45034456e-01 -1.00588882e+00 7.86079109e-01
-2.06859970e+00 -1.70669460e+00 -1.90810800e-01 9.34221327e-01
5.49244702e-01 -6.64976180e-01 4.56991524e-01 9.53705907e-01
-8.07223499e-01 8.62367570e-01 4.30381685e-01 2.85114437e-01
5.53558111e-01 -7.20371127e-01 8.97320032e-01 8.00893068e-01
-2.37796560e-01 9.42770615e-02 4.52933162e-01 -5.20688832e-01
-1.01583731e+00 -2.01155114e+00 1.05262709e+00 -5.87083578e-01
1.08649123e+00 -8.22847933e-02 -7.63632774e-01 8.26525688e-01
1.53596336e-02 3.38576943e-01 5.40205002e-01 -7.03869164e-02
-3.19188118e-01 -7.35532105e-01 -1.28779054e+00 5.59931695e-01
1.14826131e+00 -6.25463068e-01 -1.51438519e-01 4.13816869e-01
8.09009075e-01 -9.72935781e-02 -9.94367540e-01 1.53298229e-01
4.05632675e-01 -5.04215777e-01 1.17573225e+00 -7.74266005e-01
-6.94052652e-02 -2.22867116e-01 -2.48500854e-01 -1.52842104e+00
-3.74689251e-01 -3.61401588e-01 -2.00069457e-01 1.28882658e+00
3.47141415e-01 -1.22621858e+00 9.07770455e-01 7.68117547e-01
-4.28058147e-01 -3.46635371e-01 -1.20297933e+00 -1.04764473e+00
4.82108295e-01 -8.95517826e-01 1.19428146e+00 9.70559239e-01
-6.24571323e-01 1.12158246e-01 -5.95697582e-01 3.95103335e-01
5.60644090e-01 -3.52666706e-01 1.08877409e+00 -1.40064859e+00
5.28688431e-01 -5.54433882e-01 -6.38342857e-01 -1.30926251e+00
8.04505169e-01 -8.65626693e-01 -4.61058319e-01 -1.53429961e+00
-5.35160601e-01 -8.22777569e-01 -4.07144070e-01 4.06309068e-01
3.51998955e-01 1.48780495e-02 -1.18722014e-01 2.56154360e-03
-4.18596447e-01 8.50592852e-01 1.37935495e+00 -5.71242392e-01
5.71400374e-02 5.17731130e-01 -6.27040505e-01 4.03118223e-01
1.17971861e+00 -4.10324752e-01 -6.03264749e-01 -5.17842650e-01
-2.80211329e-01 -1.30199924e-01 5.70355415e-01 -1.05468142e+00
6.04476511e-01 -4.81719971e-01 -5.03227822e-02 -8.01602244e-01
2.80502468e-01 -1.21293271e+00 2.40526367e-02 2.30028704e-01
2.06682179e-02 -1.13625638e-02 3.88673067e-01 8.97553921e-01
-3.02009955e-02 6.07445717e-01 4.24930215e-01 3.51592332e-01
-1.10445058e+00 9.22671616e-01 -5.77176094e-01 9.23789740e-02
1.10193503e+00 -4.30888206e-01 -6.84863746e-01 -4.84280288e-01
-3.79902810e-01 4.62927341e-01 6.36997074e-02 6.77401125e-01
5.74895799e-01 -1.67874444e+00 -6.91695213e-01 2.43690193e-01
1.89344600e-01 -3.38547200e-01 4.21680450e-01 7.37882793e-01
-2.35329166e-01 7.74997473e-01 -1.75908506e-01 -3.50359768e-01
-5.93550920e-01 5.70965469e-01 3.78371388e-01 -1.79762080e-01
-4.46112812e-01 3.58180970e-01 -4.13135327e-02 -8.47127557e-01
5.52156419e-02 -3.19676310e-01 -9.76091921e-02 -1.63241506e-01
1.52353466e-01 9.09720957e-01 1.68956048e-03 -1.06548476e+00
-4.56784964e-01 5.10257483e-01 3.89223956e-02 2.78073490e-01
1.26266170e+00 -3.36362302e-01 4.63975251e-01 -4.13113534e-02
1.26113260e+00 -2.85061032e-01 -1.32122815e+00 -5.25020301e-01
-1.73363850e-01 -4.77836996e-01 2.76064366e-01 -7.36892283e-01
-1.63065636e+00 1.00367939e+00 5.68350494e-01 1.80524558e-01
1.03923881e+00 -2.69846767e-01 1.51562655e+00 4.07182634e-01
3.22062016e-01 -9.80774164e-01 -1.07541233e-01 7.06945777e-01
5.40602088e-01 -1.52557850e+00 -4.76289093e-01 -2.00545147e-01
-7.10295022e-01 9.95692909e-01 6.92136049e-01 -9.64786634e-02
1.09707451e+00 -1.08473249e-01 -3.85759547e-02 -1.39408797e-01
-7.47729063e-01 -5.91381073e-01 4.25046355e-01 1.19009268e+00
-4.10282286e-03 4.31288689e-01 3.42042148e-01 2.50880271e-01
7.50018284e-02 1.20488837e-01 9.65942368e-02 4.78924513e-01
-4.25325692e-01 -1.36873186e+00 -1.08984537e-01 5.26315868e-01
9.16256905e-02 8.77497531e-03 2.22479608e-02 9.85705137e-01
1.63418084e-01 1.26937354e+00 1.33445665e-01 -7.31939018e-01
5.48574269e-01 -2.48831779e-01 -1.75790891e-01 -1.39659315e-01
-2.85076529e-01 -7.40083516e-01 4.15765703e-01 -5.03774762e-01
-4.36249018e-01 -4.38729316e-01 -9.58778679e-01 -1.17806637e+00
9.21820551e-02 -6.15783408e-02 4.94831622e-01 9.56823945e-01
6.77594543e-01 7.13365376e-01 9.68371809e-01 -7.06189692e-01
-6.25437424e-02 -5.72567940e-01 -2.60739475e-01 2.90797591e-01
6.14540875e-01 -1.02701533e+00 -2.48079240e-01 -2.87552685e-01] | [6.47177267074585, 2.0692965984344482] |
6b8ea2c9-0e42-4861-bd0a-1516651b7535 | benchmarking-zero-shot-and-few-shot | 2208.01814 | null | https://arxiv.org/abs/2208.01814v2 | https://arxiv.org/pdf/2208.01814v2.pdf | Benchmarking zero-shot and few-shot approaches for tokenization, tagging, and dependency parsing of Tagalog text | The grammatical analysis of texts in any written language typically involves a number of basic processing tasks, such as tokenization, morphological tagging, and dependency parsing. State-of-the-art systems can achieve high accuracy on these tasks for languages with large datasets, but yield poor results for languages which have little to no annotated data. To address this issue for the Tagalog language, we investigate the use of alternative language resources for creating task-specific models in the absence of dependency-annotated Tagalog data. We also explore the use of word embeddings and data augmentation to improve performance when only a small amount of annotated Tagalog data is available. We show that these zero-shot and few-shot approaches yield substantial improvements on grammatical analysis of both in-domain and out-of-domain Tagalog text compared to state-of-the-art supervised baselines. | ['Franz de Leon', 'Angelina Aquino'] | 2022-08-03 | null | null | null | null | ['morphological-tagging'] | ['natural-language-processing'] | [-1.58833891e-01 1.80163443e-01 -2.48553529e-01 -6.38427138e-01
-1.25970352e+00 -7.94708252e-01 6.04059458e-01 7.10548341e-01
-7.61915982e-01 5.93648255e-01 3.35225642e-01 -4.53459382e-01
4.58150655e-01 -6.63767040e-01 -4.99494940e-01 -2.18185946e-01
8.96249991e-03 6.86868489e-01 2.75110096e-01 -2.64051139e-01
1.62404925e-01 -7.85127282e-02 -1.08063626e+00 2.21965745e-01
1.11392522e+00 2.26760775e-01 2.73124397e-01 5.31325579e-01
-7.65895426e-01 5.68397582e-01 -4.12657976e-01 -5.99600017e-01
-1.31570771e-01 -1.59884579e-02 -1.04741800e+00 -1.06268667e-01
5.89212358e-01 -3.21142189e-02 1.74684282e-02 1.04033840e+00
4.26372051e-01 -5.99493720e-02 5.26220322e-01 -6.88238561e-01
-6.65039837e-01 1.04553866e+00 -4.00190592e-01 4.35528666e-01
3.33171725e-01 -9.17912796e-02 1.33027232e+00 -8.50495815e-01
7.10431695e-01 1.35941100e+00 7.28747785e-01 5.65719903e-01
-1.35824990e+00 -3.93176258e-01 3.36503893e-01 -3.92768741e-01
-1.02725589e+00 -6.25600100e-01 5.15845537e-01 -5.95596552e-01
1.57840490e+00 -4.04923528e-01 5.29148504e-02 8.35790813e-01
5.34418374e-02 6.89470410e-01 7.93358743e-01 -1.02093899e+00
1.77716687e-02 -9.30838063e-02 6.63633943e-01 9.17311072e-01
3.21870416e-01 -2.57169694e-01 -4.98054862e-01 -2.95852870e-01
2.02118576e-01 -5.05507231e-01 1.67778388e-01 9.01297294e-03
-9.43232238e-01 9.28339481e-01 -3.65990788e-01 6.17105603e-01
2.84505170e-03 5.47181778e-02 8.26306760e-01 3.10997218e-01
1.14334750e+00 6.51704729e-01 -1.00211704e+00 -3.64975184e-01
-1.03807926e+00 1.08995281e-01 1.00061119e+00 9.93955791e-01
7.83056676e-01 9.85946879e-02 1.07499301e-01 1.06746650e+00
2.44176999e-01 2.61528850e-01 5.89030027e-01 -5.54414392e-01
1.06888127e+00 4.72641259e-01 9.12297517e-02 -1.08738668e-01
-4.13888693e-01 -9.93703306e-02 2.26794109e-02 -2.61037290e-01
8.21650684e-01 -3.78051728e-01 -1.09902537e+00 1.84785414e+00
4.41318423e-01 -3.16431910e-01 2.86329716e-01 2.51227617e-01
9.31898117e-01 5.62666178e-01 5.94030857e-01 2.37948615e-02
1.55193508e+00 -8.06346297e-01 -6.67471349e-01 -7.19066799e-01
1.56197739e+00 -9.19523716e-01 1.30880499e+00 8.26142542e-03
-1.13488722e+00 -2.92157888e-01 -6.76508427e-01 -5.19934714e-01
-6.18871212e-01 5.88470772e-02 6.60738289e-01 6.48853242e-01
-8.88328493e-01 5.95679164e-01 -1.06828868e+00 -7.29023635e-01
8.04558098e-02 1.00736216e-01 -5.95356703e-01 -2.94932544e-01
-1.25016093e+00 1.02427268e+00 5.88480830e-01 -3.82972062e-01
-5.43902576e-01 -7.62981534e-01 -1.37529397e+00 -2.74285786e-02
3.31753820e-01 -6.75036311e-02 1.52991426e+00 -5.59711397e-01
-9.78534758e-01 1.43434799e+00 -2.52268344e-01 -1.99071407e-01
-3.79542932e-02 -4.57650423e-01 -2.34910324e-01 -2.37786442e-01
3.74563813e-01 4.49426979e-01 4.48722482e-01 -6.19739354e-01
-9.44664955e-01 -4.27743316e-01 -3.00018713e-02 1.03017673e-01
-5.23615062e-01 7.37375796e-01 -5.01127660e-01 -6.31529450e-01
-1.16608299e-01 -7.84328759e-01 -1.57514811e-01 -7.03228652e-01
-1.39434129e-01 -8.52382243e-01 5.70183694e-01 -1.01846921e+00
1.41469193e+00 -2.15778637e+00 -1.27361447e-01 -2.93360978e-01
-1.65728912e-01 3.89274508e-01 -3.25051665e-01 5.95497668e-01
4.16739509e-02 5.58054447e-01 -1.42832741e-01 -6.69169009e-01
1.63448021e-01 4.58256543e-01 -1.08086824e-01 3.07637930e-01
4.01503146e-01 9.27102447e-01 -1.11456382e+00 -7.06579566e-01
-4.68318090e-02 3.51579748e-02 -4.56215024e-01 1.90571800e-01
-4.02369529e-01 1.08837433e-01 -4.11113262e-01 8.35137844e-01
3.15047026e-01 9.76527110e-02 6.47681296e-01 2.65473276e-01
-3.79494488e-01 1.29085743e+00 -7.24434257e-01 1.76863050e+00
-7.40843832e-01 3.39852989e-01 9.25678462e-02 -9.31199670e-01
6.19155228e-01 3.70586127e-01 9.15629044e-02 -4.65990841e-01
9.84743163e-02 4.96749431e-01 1.59461305e-01 -4.43335354e-01
6.79783940e-01 -2.86095858e-01 -7.24296153e-01 6.96475208e-01
4.52631235e-01 -1.44286320e-01 6.95578814e-01 2.76741475e-01
1.44648015e+00 2.80778348e-01 4.51954544e-01 -5.34953058e-01
9.85658616e-02 3.94789696e-01 7.28272021e-01 6.53125584e-01
-2.64853686e-02 2.49693155e-01 4.97831970e-01 -2.64358044e-01
-1.28442454e+00 -7.03821719e-01 -1.86347783e-01 1.79210579e+00
-4.55082059e-01 -7.05531597e-01 -8.77040744e-01 -9.78622258e-01
-9.11335170e-04 9.39601481e-01 -3.57242674e-01 2.06183508e-01
-1.00199950e+00 -9.70592558e-01 8.98236513e-01 8.06024134e-01
5.48115186e-03 -1.07402563e+00 -9.86067206e-02 7.13703930e-01
-1.09975792e-01 -1.41366255e+00 -3.87533128e-01 5.88224053e-01
-9.08691049e-01 -9.60697412e-01 -1.31390274e-01 -1.18036556e+00
6.06996536e-01 -3.41397196e-01 1.53167391e+00 1.74592748e-01
-1.13392577e-01 1.76444307e-01 -6.33146286e-01 -4.48894978e-01
-7.23556936e-01 4.84675229e-01 -1.05152063e-01 -7.02263653e-01
8.72130632e-01 -2.61445165e-01 2.90528208e-01 -2.00411960e-01
-6.31752610e-01 -3.36577535e-01 2.06641316e-01 8.34920704e-01
3.08330357e-01 -3.89197946e-01 3.79655242e-01 -1.47341740e+00
6.28017128e-01 -4.00004596e-01 -6.51774526e-01 3.60875636e-01
-2.83883303e-01 4.14279014e-01 7.01440632e-01 -3.97658855e-01
-1.00852072e+00 3.87406833e-02 -2.80124754e-01 1.30347610e-01
-1.18618198e-01 7.39713490e-01 -2.81956315e-01 2.08711997e-01
4.76189613e-01 -1.99396640e-01 -2.04952672e-01 -9.39747095e-01
4.08252686e-01 7.78982759e-01 5.26216805e-01 -9.86756802e-01
6.52767241e-01 -4.56363298e-02 -3.69951963e-01 -8.37393284e-01
-1.26450133e+00 -6.82928383e-01 -1.02070606e+00 4.84708935e-01
9.19752479e-01 -1.06910992e+00 8.77177268e-02 4.09884483e-01
-1.26257133e+00 -5.97840786e-01 -1.88142911e-01 3.41482550e-01
-1.77035421e-01 5.37718832e-01 -9.46041226e-01 -6.24504328e-01
-3.73276025e-01 -7.20405579e-01 1.18377805e+00 -1.80327117e-01
-5.41518390e-01 -1.22956681e+00 1.95097432e-01 1.55699432e-01
7.71837682e-03 -2.29671299e-01 1.38683712e+00 -1.17820227e+00
-2.19130680e-01 -2.18471527e-01 3.70008573e-02 3.05431902e-01
1.53868198e-01 5.74354157e-02 -8.49497676e-01 -2.17399776e-01
-4.44820702e-01 -6.71884358e-01 9.51780021e-01 2.13864118e-01
6.13472402e-01 -9.23845768e-02 -2.62766778e-01 4.65565592e-01
1.20595980e+00 -1.09170221e-01 3.86589915e-01 3.25357735e-01
8.32845151e-01 6.85236812e-01 8.59984398e-01 1.82250649e-01
6.60449505e-01 4.32879388e-01 3.17008346e-02 8.23928267e-02
-1.27733991e-01 -3.16298187e-01 4.12107408e-01 1.08701968e+00
3.24723303e-01 -5.64107060e-01 -1.27886915e+00 1.18643224e+00
-1.75448895e+00 -5.35291314e-01 -1.31205827e-01 2.07085514e+00
1.19289279e+00 2.13458270e-01 -1.13761626e-01 -6.73416331e-02
7.61474073e-01 2.75472373e-01 1.10597638e-02 -8.84242237e-01
5.94408214e-02 4.44488615e-01 6.00534558e-01 6.52303219e-01
-1.57065499e+00 1.75629044e+00 6.92973804e+00 7.58892894e-01
-7.43320346e-01 4.20009136e-01 2.17312381e-01 1.59945980e-01
-3.26485217e-01 3.91565353e-01 -1.47279131e+00 5.17090082e-01
1.28268874e+00 2.18730167e-01 1.50501147e-01 8.63428116e-01
2.21512169e-02 -1.95894837e-01 -1.18366718e+00 4.88819957e-01
-1.92986690e-02 -9.59610581e-01 -3.44020933e-01 6.46482036e-02
4.97240871e-01 4.15716618e-01 -4.99295592e-01 5.32937706e-01
9.67617691e-01 -8.46183717e-01 5.63734233e-01 -1.77612439e-01
1.02302969e+00 -5.80296338e-01 8.08744431e-01 4.87948418e-01
-1.17849505e+00 2.10588127e-01 -4.69942778e-01 -2.22675920e-01
3.69418830e-01 6.08385921e-01 -7.60028183e-01 2.37887651e-01
4.18856472e-01 6.58624172e-01 -5.97258449e-01 5.40939808e-01
-6.96598828e-01 1.14299595e+00 -4.65515524e-01 -1.98558390e-01
4.54489440e-01 9.71639007e-02 4.30914342e-01 1.54232049e+00
1.71352923e-01 7.16038719e-02 5.81907094e-01 3.94048750e-01
-3.02013010e-01 4.07221556e-01 -6.48091197e-01 -5.21971226e-01
6.99697375e-01 1.18087649e+00 -6.03828490e-01 -6.12599313e-01
-8.36818933e-01 6.50751233e-01 1.01832545e+00 -8.57360587e-02
-2.80230135e-01 -5.58298826e-01 6.54666841e-01 2.20401585e-01
5.26198447e-01 -6.79175556e-01 -3.15148115e-01 -1.19420052e+00
-8.86640046e-03 -7.32642591e-01 7.72533357e-01 -3.85954589e-01
-1.47877347e+00 2.37665772e-01 2.27870476e-02 -6.76817656e-01
-5.81228077e-01 -6.64626360e-01 -7.04798400e-01 9.43956316e-01
-1.59678566e+00 -1.44102061e+00 2.42912278e-01 8.14541429e-02
6.46623373e-01 -2.41449401e-01 1.01722646e+00 2.86602050e-01
-5.32179058e-01 7.16093600e-01 3.85405630e-01 5.76452911e-01
9.60896969e-01 -1.44321990e+00 9.88921881e-01 1.08569825e+00
2.71345705e-01 6.39677584e-01 5.46513975e-01 -9.67569649e-01
-1.11551070e+00 -1.24110973e+00 1.81014168e+00 -6.72209442e-01
1.06603742e+00 -8.91664863e-01 -1.21411836e+00 9.89763439e-01
9.00289118e-02 1.43016756e-01 7.40900695e-01 9.81703281e-01
-4.39542800e-01 3.10035735e-01 -9.31595385e-01 3.22853535e-01
1.07120502e+00 -7.20379174e-01 -1.08721471e+00 3.41529757e-01
6.47890091e-01 -2.97037363e-01 -8.20569277e-01 -2.32762774e-03
3.18922490e-01 -1.64083064e-01 4.90592688e-01 -9.99728560e-01
3.85958761e-01 7.41325021e-02 -9.40568075e-02 -1.43469548e+00
-2.93977767e-01 -4.84457433e-01 3.18945974e-01 1.75004232e+00
6.93539619e-01 -5.96311450e-01 5.21463156e-01 7.88362443e-01
-6.35835886e-01 -3.51900190e-01 -8.32962632e-01 -9.82046485e-01
5.48375547e-01 -4.12017733e-01 4.44516838e-01 1.04767144e+00
4.39565539e-01 4.05640870e-01 1.94079001e-02 3.89963798e-02
4.92442161e-01 -1.29261650e-02 6.71006680e-01 -1.34616244e+00
-6.70782924e-02 1.16858386e-01 -3.44218940e-01 -7.03057468e-01
9.62260425e-01 -1.11871088e+00 3.91185313e-01 -1.64189923e+00
1.44308612e-01 -6.69136047e-01 -1.42351508e-01 9.63147044e-01
-3.40380937e-01 1.35868996e-01 -1.28941074e-01 -2.08289139e-02
-5.93551219e-01 2.00070500e-01 5.63297510e-01 1.03878044e-01
-9.79450345e-02 -6.84349597e-01 -7.29568362e-01 9.36636150e-01
6.70518637e-01 -9.59703445e-01 9.99893993e-02 -8.27047944e-01
-1.78275467e-03 -2.94888347e-01 -2.04188183e-01 -5.74857175e-01
-9.27943066e-02 -1.81790367e-01 1.19893812e-02 -3.94833565e-01
-7.71692581e-03 -1.71897262e-01 -6.26362681e-01 8.38404223e-02
-1.67482615e-01 1.64707854e-01 3.92065257e-01 2.24133596e-01
-3.04220855e-01 -7.03521490e-01 8.50137353e-01 -5.09156346e-01
-8.89869630e-01 7.36648068e-02 -6.76189184e-01 9.00275409e-01
6.10051453e-01 -8.32839906e-02 -4.86445278e-01 1.34849548e-01
-4.56643909e-01 3.08168322e-01 6.63814127e-01 3.38146180e-01
-1.00537769e-01 -9.38614964e-01 -7.41465449e-01 2.22363621e-01
1.81931198e-01 -4.10171859e-02 -1.93363711e-01 4.57423121e-01
-4.45094138e-01 4.79545146e-01 1.21060364e-01 -1.46731392e-01
-1.45982158e+00 3.50775421e-01 1.86219029e-02 -7.44913936e-01
-5.90363324e-01 8.71426105e-01 1.09973259e-01 -7.94934750e-01
-5.85708842e-02 -4.12693232e-01 -1.24995284e-01 2.27164388e-01
2.69062519e-01 -1.49368858e-02 3.34813565e-01 -6.95770860e-01
-6.20129526e-01 3.73639911e-01 -4.33360845e-01 -2.22416803e-01
1.46109653e+00 -1.00960784e-01 -1.27218664e-01 4.81053889e-01
9.43397522e-01 2.71941453e-01 -8.45053315e-01 -3.46804917e-01
5.41699290e-01 -2.66769618e-01 6.62245229e-02 -6.54850841e-01
-4.18645829e-01 1.15152490e+00 -6.79626390e-02 4.15383428e-02
5.17624795e-01 3.63606215e-01 1.07251120e+00 3.67623419e-01
3.56974721e-01 -1.58308256e+00 -2.92876810e-01 1.33079147e+00
2.48624519e-01 -1.45554829e+00 -9.73298103e-02 -5.51620424e-01
-4.96164858e-01 9.04462636e-01 5.16515911e-01 -4.52731289e-02
6.25448406e-01 5.95310450e-01 2.12446764e-01 -1.74875811e-01
-8.64441276e-01 -5.04062235e-01 1.23659879e-01 4.10447925e-01
1.11238348e+00 2.85535213e-02 -6.63228273e-01 8.09515476e-01
-2.36059129e-01 -3.90148818e-01 4.57840741e-01 1.40199399e+00
-5.40043592e-01 -1.66032004e+00 -2.05260143e-02 5.37933409e-01
-1.06347835e+00 -7.76130438e-01 -3.72838378e-01 1.05255008e+00
-1.32661507e-01 9.90120113e-01 3.14908445e-01 1.64178640e-01
7.80945346e-02 8.13613713e-01 6.19707108e-01 -1.43961883e+00
-7.41887510e-01 3.03644717e-01 9.43805397e-01 -2.11590871e-01
-1.14126734e-01 -9.26829696e-01 -1.45965159e+00 -8.96990523e-02
-4.27109718e-01 2.07196295e-01 6.42334580e-01 1.27112269e+00
7.46611208e-02 1.60239846e-01 8.74906313e-03 -6.63717270e-01
-5.68314135e-01 -1.26906121e+00 -8.14292669e-01 5.10085464e-01
-1.00742746e-02 -4.62164253e-01 -3.28364611e-01 1.20845310e-01] | [10.43354606628418, 9.85705852508545] |
e8da2db8-a1de-413b-8042-b78194e01a44 | when-do-you-need-chain-of-thought-prompting | 2304.03262 | null | https://arxiv.org/abs/2304.03262v2 | https://arxiv.org/pdf/2304.03262v2.pdf | When do you need Chain-of-Thought Prompting for ChatGPT? | Chain-of-Thought (CoT) prompting can effectively elicit complex multi-step reasoning from Large Language Models~(LLMs). For example, by simply adding CoT instruction ``Let's think step-by-step'' to each input query of MultiArith dataset, GPT-3's accuracy can be improved from 17.7\% to 78.7\%. However, it is not clear whether CoT is still effective on more recent instruction finetuned (IFT) LLMs such as ChatGPT. Surprisingly, on ChatGPT, CoT is no longer effective for certain tasks such as arithmetic reasoning while still keeping effective on other reasoning tasks. Moreover, on the former tasks, ChatGPT usually achieves the best performance and can generate CoT even without being instructed to do so. Hence, it is plausible that ChatGPT has already been trained on these tasks with CoT and thus memorized the instruction so it implicitly follows such an instruction when applied to the same queries, even without CoT. Our analysis reflects a potential risk of overfitting/bias toward instructions introduced in IFT, which becomes more common in training LLMs. In addition, it indicates possible leakage of the pretraining recipe, e.g., one can verify whether a dataset and instruction were used in training ChatGPT. Our experiments report new baseline results of ChatGPT on a variety of reasoning tasks and shed novel insights into LLM's profiling, instruction memorization, and pretraining dataset leakage. | ['Tianyi Zhou', 'Heng Huang', 'Lichang Chen', 'Jiuhai Chen'] | 2023-04-06 | null | null | null | null | ['memorization', 'arithmetic-reasoning'] | ['natural-language-processing', 'reasoning'] | [ 2.15306699e-01 3.05730581e-01 -2.32141986e-02 -4.10271198e-01
-9.95612264e-01 -7.74941325e-01 3.43098640e-01 3.80284518e-01
-5.38314402e-01 5.60820699e-01 1.96954787e-01 -1.18527102e+00
8.62448066e-02 -9.68845904e-01 -1.06317079e+00 -2.75319606e-01
2.79059142e-01 6.23413682e-01 2.77380377e-01 -5.19619226e-01
3.08124691e-01 1.45649061e-01 -1.11462045e+00 7.68683553e-01
1.04582560e+00 3.74457628e-01 2.70470679e-01 7.75654197e-01
-2.63846248e-01 1.52627075e+00 -9.06884432e-01 -7.39203513e-01
-1.01873390e-01 -6.63074404e-02 -1.07589853e+00 -5.63011765e-01
7.88962603e-01 -3.54375243e-01 -1.54478252e-01 7.98504114e-01
2.37565815e-01 2.42476478e-01 5.31876445e-01 -9.83655214e-01
-6.20992541e-01 1.26741576e+00 -4.54596907e-01 2.47880757e-01
4.87717420e-01 5.30423462e-01 1.03275228e+00 -7.83120334e-01
3.68028969e-01 1.36338246e+00 6.90063059e-01 3.91146749e-01
-1.47108209e+00 -7.07885623e-01 7.00751990e-02 -2.47440502e-01
-1.17861497e+00 -3.11330914e-01 3.30134571e-01 -2.33463317e-01
1.52069604e+00 3.95169109e-01 2.40350351e-01 1.11068177e+00
6.35244489e-01 1.11126006e+00 1.29426432e+00 -4.82506514e-01
-1.34184603e-02 5.60336299e-02 5.09082973e-01 8.76814067e-01
1.22146964e-01 -8.14661458e-02 -6.08671129e-01 -1.11078106e-01
4.33122456e-01 -2.33532161e-01 -1.02032140e-01 2.34926939e-01
-1.25513387e+00 1.01807094e+00 3.95355582e-01 5.27655959e-01
4.72392561e-03 2.52960563e-01 4.34952229e-01 5.72212994e-01
7.28427917e-02 9.88736928e-01 -9.15016353e-01 -3.66778821e-01
-6.71494901e-01 3.08570296e-01 1.00420475e+00 1.11910200e+00
7.02237368e-01 7.16450140e-02 -2.67729133e-01 4.91547227e-01
-7.04990029e-02 6.66121721e-01 6.07727051e-01 -8.52037847e-01
9.54002202e-01 7.33068347e-01 -2.37340465e-01 -7.51190960e-01
-6.17099464e-01 -4.49806571e-01 -4.73180175e-01 -1.38437271e-01
8.04294705e-01 -2.02505559e-01 -4.62273419e-01 1.99636376e+00
-1.76494926e-01 -2.00597748e-01 2.70831674e-01 5.67580223e-01
8.37155521e-01 5.75563669e-01 4.01461810e-01 1.09381028e-01
1.51397729e+00 -8.62265170e-01 -4.65894967e-01 -9.09059286e-01
1.50508940e+00 -7.70934165e-01 1.76580906e+00 4.57303345e-01
-1.23213363e+00 -7.34644532e-01 -9.40073371e-01 -4.06312972e-01
-4.17724311e-01 -1.09948553e-01 1.07819891e+00 6.62780344e-01
-1.10067713e+00 4.11052555e-01 -5.81641555e-01 -3.84905599e-02
2.12593719e-01 3.72269034e-01 -9.30355862e-02 -5.25857151e-01
-1.38405073e+00 1.01026297e+00 3.75618219e-01 -2.73125798e-01
-7.92193890e-01 -9.28705394e-01 -8.87595773e-01 1.80045053e-01
6.92152679e-01 -6.32271469e-01 1.70443344e+00 -6.02597594e-01
-1.42624474e+00 8.25343132e-01 -2.24599689e-01 -4.85887736e-01
3.23762357e-01 -5.35865009e-01 -2.31285840e-01 -1.68637201e-01
1.78103790e-01 6.39230192e-01 6.40989244e-01 -7.63460279e-01
-2.04777956e-01 -1.55652806e-01 4.96206880e-01 1.18238144e-01
6.91588223e-02 8.26199874e-02 8.62841215e-03 -4.57267970e-01
1.26477014e-02 -8.25132072e-01 1.20112993e-01 -6.91681027e-01
-3.51135194e-01 -4.47429031e-01 3.52786303e-01 -4.37815636e-01
1.31174755e+00 -2.15265989e+00 -8.73607919e-02 2.61972826e-02
5.11501253e-01 3.19739074e-01 -2.60238051e-01 3.01209122e-01
-9.00460184e-02 4.09939855e-01 4.59318347e-02 -1.52854025e-01
1.81457475e-01 2.64754713e-01 -5.71052730e-01 -7.83017278e-02
1.93292618e-01 1.40442514e+00 -9.27198887e-01 -3.74569476e-01
7.64727145e-02 -7.53147900e-02 -9.52684939e-01 2.46153101e-01
-8.09453547e-01 2.49321446e-01 -2.70946831e-01 3.80300701e-01
3.66459221e-01 -4.40477848e-01 1.30744338e-01 4.74888496e-02
1.12259321e-01 9.00167644e-01 -6.87578619e-01 1.59435582e+00
-7.65738308e-01 6.27091050e-01 -3.28736156e-01 -8.14240754e-01
5.57293177e-01 1.22798786e-01 -3.04554760e-01 -9.32984412e-01
1.10846683e-01 3.07205439e-01 5.14182210e-01 -4.61536556e-01
6.10912025e-01 -2.90817767e-01 -4.41406012e-01 6.93556130e-01
-9.71674398e-02 -3.72836858e-01 6.23200797e-02 4.84078765e-01
1.31624663e+00 -1.00932911e-01 1.12814188e-01 -3.71480018e-01
5.68044186e-01 2.67386824e-01 9.58111063e-02 1.15634537e+00
1.10114530e-01 8.70400816e-02 6.16106689e-01 -2.34843671e-01
-5.72457254e-01 -9.04557407e-01 9.04407576e-02 1.74044335e+00
-3.86561275e-01 -7.14689553e-01 -7.62319088e-01 -7.39183664e-01
-2.08111852e-02 1.41772664e+00 -2.99527913e-01 -5.39085031e-01
-8.91433597e-01 -5.88820994e-01 1.05040264e+00 6.07417226e-01
7.52870262e-01 -1.03327775e+00 -6.80943608e-01 3.20267022e-01
-4.53551263e-01 -1.11234915e+00 -4.45616305e-01 5.21714866e-01
-1.01002729e+00 -9.24589992e-01 9.04159248e-02 -6.24426007e-01
6.13111138e-01 2.11642891e-01 1.73820519e+00 5.73918223e-01
1.89801097e-01 1.97322518e-01 -1.22774638e-01 -3.07673842e-01
-7.27903008e-01 4.77859348e-01 -2.04517350e-01 -7.40879834e-01
6.58860683e-01 -2.55011946e-01 -2.53521740e-01 2.21274287e-01
-7.74299204e-01 3.31281126e-01 7.27973163e-01 7.87866712e-01
-6.38970435e-02 1.33788869e-01 4.16080654e-01 -1.35878575e+00
7.03311384e-01 -4.18072194e-01 -3.81215185e-01 2.99412072e-01
-3.98108929e-01 4.38267976e-01 8.97033870e-01 -7.15859890e-01
-1.07399702e+00 -6.57643318e-01 -2.22545043e-01 -6.54363260e-02
-1.46117702e-01 6.68309510e-01 -1.74027160e-02 9.79960989e-03
1.03773654e+00 2.02430099e-01 -3.28637123e-01 -1.28325939e-01
3.09506714e-01 3.09607357e-01 3.56991708e-01 -1.37836409e+00
6.92761242e-01 -2.86390215e-01 -3.64589125e-01 -5.63868225e-01
-1.00226784e+00 1.25220278e-02 -8.18223432e-02 2.75196940e-01
7.89762914e-01 -8.84711981e-01 -1.30927837e+00 5.25202990e-01
-1.07120037e+00 -1.16458046e+00 -9.76981036e-03 2.16002882e-01
-2.67375588e-01 -3.06548793e-02 -1.04165995e+00 -5.52857518e-01
-2.71229446e-01 -1.54290450e+00 8.98776770e-01 1.40045539e-01
-8.77486825e-01 -1.13004541e+00 -4.59700346e-01 6.79535627e-01
4.80667502e-01 -4.44182545e-01 1.73436868e+00 -9.45934892e-01
-7.13450193e-01 7.31986538e-02 -1.24121606e-01 2.64501750e-01
-1.48767024e-01 -3.05604964e-01 -1.04681575e+00 -2.41906330e-01
4.14504021e-01 -7.02781379e-01 4.41882730e-01 5.28917387e-02
9.40079272e-01 -4.69476104e-01 -1.53770924e-01 4.30077612e-01
1.26233971e+00 2.31645286e-01 5.23163855e-01 2.14219719e-01
6.42009556e-01 3.48204374e-01 5.22305906e-01 -3.13039154e-01
6.24021113e-01 4.33030993e-01 -9.07417983e-02 3.61097127e-01
1.17244534e-01 -4.79280353e-01 7.52496362e-01 8.96505535e-01
3.83071780e-01 -1.50895908e-01 -1.30370736e+00 8.71971548e-02
-1.39780545e+00 -4.38274086e-01 -3.25561851e-01 2.09310341e+00
1.38543999e+00 6.33584797e-01 -4.58121449e-01 1.00317836e-01
-1.01817809e-01 1.07534900e-02 -3.32789958e-01 -8.86500776e-01
-4.23313454e-02 6.62964106e-01 3.76264811e-01 8.12223196e-01
-5.13669431e-01 1.32667208e+00 5.96752834e+00 8.00336421e-01
-1.03498638e+00 1.47745028e-01 6.34725213e-01 1.89964250e-01
-5.29112279e-01 5.27247824e-02 -9.89160061e-01 2.07208708e-01
1.34707236e+00 -6.86294660e-02 5.48136532e-01 7.18039930e-01
-3.08835089e-01 -6.50646448e-01 -1.49425077e+00 5.19822359e-01
-3.55461566e-03 -1.04950380e+00 1.55946299e-01 -2.76944399e-01
4.51669097e-01 -2.26263195e-01 1.66843012e-01 1.25439358e+00
6.61846936e-01 -1.29414189e+00 6.61891818e-01 2.01625079e-01
5.39189696e-01 -7.01807439e-01 7.49305725e-01 9.68310297e-01
-8.51279438e-01 4.27273521e-03 -2.77727783e-01 -5.74345469e-01
-4.44564432e-01 3.64027649e-01 -1.10149825e+00 2.51877666e-01
3.37189406e-01 2.70742923e-02 -1.18248510e+00 2.43674114e-01
-5.77077210e-01 8.84483159e-01 -2.50619680e-01 -3.00295174e-01
2.88252026e-01 7.17041269e-02 1.52124003e-01 1.07937133e+00
-1.51362987e-02 4.29101199e-01 2.50226170e-01 8.99137914e-01
-1.00018121e-01 -2.08029076e-02 -6.21430635e-01 -3.16459864e-01
4.23056543e-01 8.73775363e-01 -4.83989835e-01 -6.46935523e-01
-5.32547355e-01 7.50169754e-01 6.85313106e-01 3.95816505e-01
-8.66671383e-01 -1.37218505e-01 4.83886480e-01 5.62296249e-02
-1.90064579e-01 -2.40428895e-01 -6.27216697e-01 -1.15398622e+00
-1.97547480e-01 -1.47063923e+00 4.49801296e-01 -1.18972576e+00
-9.28635657e-01 4.03915852e-01 1.91188484e-01 -2.76688129e-01
-4.17263001e-01 -8.32322717e-01 -4.92732286e-01 1.05744779e+00
-9.24897909e-01 -7.43288815e-01 -5.37047014e-02 5.42031765e-01
6.92890346e-01 -6.86676055e-03 9.79918659e-01 1.19748831e-01
-4.63163257e-01 9.84773993e-01 -5.87232172e-01 2.06506327e-01
7.66588449e-01 -1.44405556e+00 4.62638050e-01 8.08506668e-01
8.20557401e-02 1.36864746e+00 8.13180864e-01 -6.92198992e-01
-1.78265440e+00 -8.93016219e-01 1.12401879e+00 -1.01363373e+00
7.85491228e-01 -4.74043161e-01 -1.29702532e+00 1.33590853e+00
1.59915790e-01 -5.21392405e-01 6.26697898e-01 3.17373723e-01
-5.61190903e-01 7.74664506e-02 -9.22412455e-01 1.11486721e+00
9.62512553e-01 -8.59668255e-01 -1.09158468e+00 4.75148618e-01
1.04923928e+00 -8.80129337e-01 -8.41324985e-01 2.12696508e-01
8.94575119e-02 -7.97642529e-01 1.08998561e+00 -7.00710177e-01
7.35300779e-01 5.82250468e-02 -2.86953270e-01 -1.31935179e+00
-1.74446002e-01 -5.32086194e-01 -7.35130534e-02 1.12627685e+00
4.87111062e-01 -9.74911213e-01 4.02045786e-01 9.47287381e-01
-2.00927109e-01 -7.45920897e-01 -4.14299071e-01 -4.30222899e-01
6.99283063e-01 -9.56962407e-01 6.32279456e-01 8.19236100e-01
1.36311188e-01 8.32267106e-01 9.23931003e-02 1.42833158e-01
2.98546642e-01 1.34970713e-02 8.89159143e-01 -8.39175344e-01
-5.02644598e-01 -3.44086170e-01 2.79878289e-01 -1.23686099e+00
4.25150901e-01 -1.26055670e+00 7.30605721e-02 -1.03833127e+00
3.16850431e-02 -6.46291018e-01 1.49354741e-01 7.09454954e-01
-5.32035768e-01 -2.70155102e-01 2.33070701e-01 -1.18937925e-01
-4.39875871e-01 -5.59362397e-02 1.43372536e+00 -8.50921497e-02
6.58628941e-02 -7.28765363e-03 -9.59196627e-01 6.70590281e-01
8.90062630e-01 -4.40803409e-01 -7.05244124e-01 -7.38524318e-01
5.94695032e-01 3.13567013e-01 3.95118654e-01 -9.79568541e-01
2.68388331e-01 -1.15798183e-01 2.84111023e-01 -3.79940361e-01
-2.91973464e-02 -5.72643220e-01 -8.17557648e-02 5.97893476e-01
-4.94303882e-01 3.84395599e-01 8.27810109e-01 -3.99719290e-02
4.08301763e-02 -4.12405223e-01 4.40333039e-01 -4.01359528e-01
-6.59447432e-01 -3.39359581e-01 -6.50359392e-01 6.05737686e-01
4.10286307e-01 8.01231489e-02 -8.27015400e-01 -1.68653324e-01
-3.86597186e-01 3.31475407e-01 2.88885117e-01 2.60437846e-01
4.75485444e-01 -9.01370168e-01 -5.35932004e-01 4.07526046e-01
4.56088185e-02 2.68096477e-01 8.44518393e-02 9.51804340e-01
-4.63945806e-01 7.47321725e-01 2.30930001e-01 -5.60570598e-01
-1.00780320e+00 6.59477830e-01 2.49843612e-01 -6.51029408e-01
-5.52415013e-01 1.04345727e+00 4.62165028e-01 -7.61835754e-01
2.27251761e-02 -1.03313243e+00 2.08527356e-01 -2.00461894e-01
4.67334151e-01 -7.80310556e-02 2.15786740e-01 -9.40067023e-02
-2.31204927e-01 2.62517363e-01 -3.28772277e-01 -6.13357201e-02
8.44365656e-01 2.94644654e-01 -1.62065953e-01 6.83157444e-01
1.14569938e+00 1.59602538e-01 -7.10722506e-01 -2.97902167e-01
1.57695308e-01 -1.90972611e-01 -4.19679672e-01 -1.14339089e+00
-7.76510179e-01 1.03991246e+00 -7.10473433e-02 -1.08369872e-01
7.55022645e-01 -1.30454719e-01 7.27553964e-01 9.17971909e-01
8.58536482e-01 -5.85223019e-01 3.52769077e-01 9.98987079e-01
8.43071103e-01 -1.25744200e+00 -2.62854904e-01 -3.11002970e-01
-4.36956912e-01 1.01188684e+00 9.48140085e-01 1.85962737e-01
1.02688991e-01 4.94260222e-01 -6.08930662e-02 -2.19138429e-01
-1.09005618e+00 3.03714573e-01 1.46501094e-01 1.87288925e-01
7.71422863e-01 2.03473657e-01 2.79219355e-02 7.12157488e-01
-8.39330673e-01 -2.86743995e-02 4.05406237e-01 1.06832528e+00
-3.40180218e-01 -8.73804748e-01 -4.78629500e-01 6.19657755e-01
-3.71033520e-01 -7.31744766e-01 -1.19081102e-02 1.01254511e+00
-5.73685169e-02 8.62215102e-01 -4.69500348e-02 -4.07016039e-01
4.00551081e-01 6.80403948e-01 5.85845232e-01 -1.02747667e+00
-1.15638506e+00 -3.32882971e-01 2.58812219e-01 -5.50689399e-01
3.59820515e-01 -4.34623778e-01 -1.53649867e+00 -8.17295730e-01
-8.60162228e-02 1.91726953e-01 -1.05290888e-02 1.21917510e+00
-5.93788363e-02 7.62086868e-01 -1.55703306e-01 -2.09768221e-01
-7.28974879e-01 -1.13052297e+00 9.33533709e-04 3.26541901e-01
9.01128650e-02 -4.66909945e-01 -4.04657423e-01 -2.40785390e-01] | [9.739147186279297, 7.435482025146484] |
0d9c24cd-2651-48b0-9ead-1cde330c2d00 | socially-aware-robot-crowd-navigation-with | 2203.01821 | null | https://arxiv.org/abs/2203.01821v4 | https://arxiv.org/pdf/2203.01821v4.pdf | Intention Aware Robot Crowd Navigation with Attention-Based Interaction Graph | We study the problem of safe and intention-aware robot navigation in dense and interactive crowds. Most previous reinforcement learning (RL) based methods fail to consider different types of interactions among all agents or ignore the intentions of people, which results in performance degradation. To learn a safe and efficient robot policy, we propose a novel recurrent graph neural network with attention mechanisms to capture heterogeneous interactions among agents through space and time. To encourage longsighted robot behaviors, we infer the intentions of dynamic agents by predicting their future trajectories for several timesteps. The predictions are incorporated into a model-free RL framework to prevent the robot from intruding into the intended paths of other agents. We demonstrate that our method enables the robot to achieve good navigation performance and non-invasiveness in challenging crowd navigation scenarios. We successfully transfer the policy learned in simulation to a real-world TurtleBot 2i. Our code and videos are available at https://sites.google.com/view/intention-aware-crowdnav/home. | ['Katherine Driggs-Campbell', 'D. Livingston McPherson', 'Weihang Liang', 'Kaiwen Hong', 'Junyi Geng', 'Neeloy Chakraborty', 'Zhe Huang', 'Peixin Chang', 'Shuijing Liu'] | 2022-03-03 | null | null | null | null | ['social-navigation'] | ['robots'] | [-3.83726090e-01 5.34855425e-01 2.29608119e-01 -8.38196427e-02
-2.50589073e-01 -3.56292129e-01 5.04855216e-01 -2.65229821e-01
-6.58722341e-01 9.94351864e-01 2.82429218e-01 -2.45401889e-01
1.94612786e-01 -7.21217215e-01 -7.76708663e-01 -6.38875365e-01
-5.92181146e-01 7.20532954e-01 4.71175641e-01 -6.22329950e-01
7.37357736e-02 3.59003574e-01 -1.26652968e+00 -5.13146698e-01
9.86183584e-01 2.11655557e-01 5.52742004e-01 8.79483700e-01
4.82981533e-01 1.54730487e+00 -2.21739650e-01 4.27118391e-02
2.63318211e-01 -3.67997557e-01 -6.90156221e-01 -1.25622496e-01
-5.00327229e-01 -7.33955503e-01 -8.17283690e-01 8.72441709e-01
4.35368657e-01 5.96212327e-01 6.12083554e-01 -1.48069251e+00
-4.74416524e-01 6.01728022e-01 -4.27316278e-01 -1.23173997e-01
4.68260378e-01 6.66402817e-01 5.03883302e-01 -2.72701353e-01
5.66228330e-01 1.34954476e+00 5.56816459e-01 8.96703959e-01
-6.50465786e-01 -5.21446943e-01 4.45046633e-01 2.72551924e-01
-1.07583404e+00 -4.28633571e-01 5.09880185e-01 -4.11075085e-01
1.03320336e+00 -3.20412606e-01 8.83848488e-01 1.28333020e+00
5.64231873e-01 8.89002562e-01 6.45877838e-01 1.05569735e-01
4.58706617e-01 -4.31857973e-01 -2.18548626e-01 9.91689265e-01
2.17477784e-01 2.19954744e-01 -3.27296048e-01 -3.58884901e-01
8.35271537e-01 1.51344895e-01 -2.01456800e-01 -5.12816548e-01
-1.17842793e+00 9.72011149e-01 7.51925588e-01 -2.49799073e-01
-9.15707469e-01 5.68108857e-01 2.12517187e-01 1.00502566e-01
-3.33226472e-02 3.55389208e-01 -1.76437780e-01 -3.23972672e-01
1.70878381e-01 5.16143143e-01 9.84768212e-01 1.09802449e+00
7.19093084e-01 1.49040326e-01 1.61502004e-01 5.25625706e-01
4.43674296e-01 8.76163065e-01 2.63775438e-01 -1.58645141e+00
2.76905507e-01 4.15539414e-01 8.47121835e-01 -1.19985771e+00
-7.90910602e-01 1.56444490e-01 -4.83899146e-01 4.94390130e-01
2.78943032e-01 -8.40225756e-01 -7.24138081e-01 1.66577148e+00
4.18627858e-01 1.24800690e-01 3.24686259e-01 1.11221790e+00
4.66776848e-01 6.58107877e-01 3.01470868e-02 -7.39828795e-02
9.57176864e-01 -1.27190995e+00 -5.64481854e-01 -7.44607151e-01
6.85418248e-01 6.16787672e-02 7.56835699e-01 2.96415329e-01
-8.73048484e-01 -1.19585365e-01 -7.13361561e-01 2.19101101e-01
-7.53663331e-02 -3.05056512e-01 5.52139342e-01 -1.28944293e-01
-1.16778541e+00 3.72581989e-01 -1.28007340e+00 -6.76868558e-01
2.40779370e-01 5.17280161e-01 -3.05770874e-01 -2.83386232e-03
-1.09999466e+00 1.07698750e+00 6.22361936e-02 1.13374196e-01
-1.46504438e+00 3.98254722e-01 -1.01571703e+00 -3.86396229e-01
8.34364116e-01 -5.85712016e-01 1.58739460e+00 -6.15958750e-01
-1.89363682e+00 2.22811282e-01 -7.52414167e-02 -7.14642644e-01
5.58775544e-01 -5.31197846e-01 2.08033785e-01 9.08602104e-02
5.15568674e-01 8.23044538e-01 5.36182463e-01 -1.42459154e+00
-7.59571970e-01 -1.47116512e-01 2.23708272e-01 7.33251929e-01
1.92727163e-01 -4.49807435e-01 -4.24192637e-01 -2.38680262e-02
-2.93391734e-01 -1.47459149e+00 -7.94983566e-01 -2.05801800e-01
-3.92363608e-01 -1.70018122e-01 5.24028122e-01 -2.99762815e-01
2.69263268e-01 -1.83715892e+00 2.96731293e-01 -1.09920800e-01
2.55441040e-01 -5.00352914e-03 -3.58937621e-01 6.71488822e-01
8.15114260e-01 -3.73660117e-01 -5.82334101e-02 -3.45206201e-01
-4.95478809e-02 5.72222590e-01 -1.96888119e-01 6.91563427e-01
-2.86343753e-01 8.99231315e-01 -1.38628972e+00 -2.15893641e-01
1.56683087e-01 4.91303861e-01 -6.60667300e-01 5.46671748e-01
-3.22786629e-01 1.04067886e+00 -9.43566442e-01 1.85906410e-01
2.74759144e-01 -2.33316481e-01 4.09397066e-01 7.43277073e-01
-2.20272288e-01 2.63952941e-01 -6.79033220e-01 1.27833366e+00
-3.78431290e-01 4.30592686e-01 4.63372916e-01 -5.61136961e-01
7.27564275e-01 1.79311320e-01 3.89452964e-01 -6.77145898e-01
3.06539148e-01 2.45812107e-02 -8.69885683e-02 -6.50333881e-01
7.13170886e-01 2.06295818e-01 -2.76531845e-01 4.87333864e-01
-3.90113890e-01 -1.51575860e-02 -1.68093458e-01 2.30229557e-01
1.31082094e+00 2.41572872e-01 3.40857387e-01 -8.50956608e-03
8.88187140e-02 4.53909606e-01 6.49779379e-01 1.28524864e+00
-4.90977824e-01 1.83479153e-02 2.72178143e-01 -4.85257506e-01
-7.13144362e-01 -8.86007011e-01 8.28501761e-01 1.52490437e+00
9.56232607e-01 -2.50644591e-02 -4.87633020e-01 -5.32390594e-01
-4.15818281e-02 8.96920085e-01 -7.39490926e-01 -3.31295207e-02
-9.57555413e-01 -3.77276450e-01 3.16704065e-01 3.01314205e-01
4.78491873e-01 -1.50409377e+00 -1.17146504e+00 3.77811462e-01
-3.77856523e-01 -9.49876428e-01 -3.60186249e-01 1.31830081e-01
-3.64825666e-01 -1.05721080e+00 -4.59854931e-01 -7.50533223e-01
7.02732146e-01 4.90529329e-01 6.35576069e-01 2.22211137e-01
2.33685017e-01 7.31005132e-01 -4.71006662e-01 -2.62262553e-01
-4.49779302e-01 -6.37570843e-02 4.23350424e-01 -4.48077619e-01
2.51875948e-02 -5.08692443e-01 -7.42278159e-01 4.11704689e-01
-1.48210078e-01 1.20201930e-01 1.84105292e-01 7.25821376e-01
7.33051673e-02 -1.17042206e-01 4.45699543e-01 -3.19580823e-01
8.66843164e-01 -7.86032856e-01 -8.05289984e-01 -1.99600920e-01
-1.41074210e-01 -9.51226428e-02 7.40615606e-01 -5.42023540e-01
-9.28895354e-01 3.18887264e-01 8.22871849e-02 7.29936063e-02
-3.77186596e-01 2.18091026e-01 2.22058326e-01 1.48303504e-03
6.07717395e-01 3.63340795e-01 2.83203989e-01 6.22557215e-02
4.82806057e-01 4.56177115e-01 4.62737292e-01 -3.87954712e-01
4.75554407e-01 6.71315432e-01 -3.12232912e-01 -9.12963986e-01
-1.63177714e-01 -1.27638295e-01 -2.77259827e-01 -6.67600036e-01
8.74973357e-01 -1.17740214e+00 -1.42275453e+00 5.84066093e-01
-1.22566545e+00 -1.15657115e+00 3.68588790e-02 3.67739707e-01
-1.00216937e+00 3.31580698e-01 -7.85488248e-01 -1.25409734e+00
-2.34174170e-02 -1.29352450e+00 8.58179748e-01 5.65309763e-01
-6.23392202e-02 -7.96353579e-01 2.52306193e-01 3.77036184e-02
4.63536710e-01 -2.78591160e-02 1.01261124e-01 -6.57754719e-01
-7.05622375e-01 7.78424814e-02 1.29938751e-01 -5.21538436e-01
-2.02503968e-02 -4.16997313e-01 -3.82717758e-01 -3.79218131e-01
-2.18000442e-01 -5.81685424e-01 5.91039479e-01 5.54268718e-01
6.83891177e-02 -7.95591116e-01 -8.33148539e-01 1.21624671e-01
9.17697728e-01 6.17255151e-01 3.92143667e-01 4.98510897e-01
7.36313105e-01 7.82266915e-01 7.52721846e-01 7.33002007e-01
1.20309699e+00 5.03811836e-01 1.00245059e+00 3.15066010e-01
5.13395667e-01 -4.54355747e-01 7.12893426e-01 4.55463320e-01
-2.64016449e-01 -7.05011070e-01 -1.07006514e+00 7.08564043e-01
-2.47455239e+00 -8.45869243e-01 4.50777188e-02 1.83358264e+00
2.40444854e-01 -8.33280943e-03 4.30921525e-01 -7.58462667e-01
6.42762423e-01 -6.06854334e-02 -7.96191037e-01 -1.48507521e-01
3.70313615e-01 -1.02750063e+00 7.81369150e-01 9.76484299e-01
-9.45081174e-01 1.39899337e+00 5.77970552e+00 1.15121514e-01
-8.67238879e-01 9.60958749e-02 2.11966693e-01 -3.10463402e-02
2.16949694e-02 -6.39434010e-02 -6.42505467e-01 1.76976502e-01
7.79688835e-01 3.62837128e-02 8.03502381e-01 9.39148903e-01
5.44030666e-01 -4.73650008e-01 -6.13019347e-01 4.72537935e-01
-2.95880139e-01 -9.86686766e-01 -5.16354620e-01 1.99067235e-01
4.81615990e-01 6.39581680e-01 -2.62842447e-01 5.72053611e-01
1.51477456e+00 -9.66546357e-01 8.59582543e-01 7.15873897e-01
-5.59344217e-02 -8.12252641e-01 6.75480366e-01 9.21174705e-01
-1.03147161e+00 -4.56878334e-01 -3.97701234e-01 -6.81142211e-01
7.06732750e-01 -1.97999716e-01 -1.13549650e+00 2.04866529e-01
6.68821335e-01 7.58291543e-01 -2.46458910e-02 8.59276593e-01
-3.38299483e-01 2.90142447e-01 -4.36261445e-01 -7.57920861e-01
5.14745176e-01 -3.16781759e-01 9.28710520e-01 6.73961282e-01
1.99888796e-01 3.71188015e-01 5.96283793e-01 5.56884527e-01
3.35448176e-01 -2.21294120e-01 -1.28841555e+00 2.20733643e-01
6.00783348e-01 7.45772183e-01 -6.73694313e-01 -6.84579760e-02
-1.64555863e-01 1.15386307e+00 7.60172129e-01 5.74916244e-01
-8.88186991e-01 -1.48003370e-01 7.12411106e-01 -2.26021066e-01
3.02130610e-01 -5.93051016e-01 2.46312395e-01 -9.68288660e-01
-2.97629744e-01 -5.16617119e-01 -3.29392403e-02 -8.40030491e-01
-1.11273170e+00 7.76720583e-01 -2.21068591e-01 -1.04985046e+00
-5.84796727e-01 -2.58578628e-01 -5.89252353e-01 3.90884519e-01
-1.22086251e+00 -9.27986085e-01 -2.37884298e-01 4.34552252e-01
5.00280440e-01 -2.30335742e-01 5.54651022e-01 -5.11935949e-01
-3.76766264e-01 -1.19418338e-01 -1.00368112e-01 2.03251168e-01
4.09639388e-01 -9.72071528e-01 5.75922012e-01 7.11000860e-01
-5.89881241e-01 4.72242713e-01 1.08399642e+00 -1.14290118e+00
-1.56520665e+00 -1.02081680e+00 4.43351239e-01 -3.30500305e-01
6.35507882e-01 -2.58400768e-01 -6.87728643e-01 1.09317720e+00
3.69563937e-01 -1.17909104e-01 1.03030026e-01 -1.99117303e-01
1.93016827e-01 4.66696382e-01 -9.08957720e-01 1.28993368e+00
1.06969893e+00 2.26060376e-02 -4.07872707e-01 3.27605575e-01
8.03192794e-01 -4.45706636e-01 -1.62480950e-01 1.63436607e-01
4.68638062e-01 -9.19471622e-01 7.17612207e-01 -3.52342546e-01
-6.08905219e-02 -3.82310957e-01 -8.46898183e-03 -1.34352899e+00
-3.23349059e-01 -8.33150804e-01 7.92088285e-02 4.47022259e-01
3.61251771e-01 -9.34497058e-01 7.50581622e-01 6.08501017e-01
2.43781470e-02 -3.51141810e-01 -1.01881969e+00 -6.53184593e-01
-5.54737821e-02 -2.23819971e-01 2.14992315e-01 3.99507254e-01
5.17660201e-01 4.31741714e-01 -1.01573563e+00 6.20402575e-01
6.13645077e-01 -2.33190581e-01 1.12669361e+00 -8.31882536e-01
-9.84959975e-02 -1.38154373e-01 -5.32293320e-02 -1.51178622e+00
4.67662007e-01 -3.19891244e-01 1.03100502e+00 -1.87876701e+00
-1.10515735e-04 -4.26597148e-01 4.71678048e-01 6.12397075e-01
-4.07011248e-02 -1.86221123e-01 2.57389992e-01 3.13684374e-01
-1.35235000e+00 1.10533178e+00 1.32826686e+00 -1.77184783e-03
-5.04444003e-01 6.82736635e-02 -4.13663983e-01 9.90070462e-01
1.13753283e+00 -5.49436986e-01 -4.69186485e-01 -4.59684342e-01
1.98962495e-01 7.22911835e-01 4.54019040e-01 -1.07674205e+00
7.75418162e-01 -4.55480218e-01 -1.96214393e-01 -3.01623076e-01
6.14369094e-01 -7.28150606e-01 2.30726339e-02 1.02277279e+00
-2.66466528e-01 1.54343426e-01 3.38882059e-02 1.16547036e+00
3.76666188e-01 -2.47977972e-02 4.87355351e-01 -3.61517459e-01
-8.33791733e-01 2.32095733e-01 -1.40223718e+00 -1.46902222e-02
1.18159676e+00 1.61620438e-01 -5.12927234e-01 -1.23018324e+00
-6.46811903e-01 1.00409973e+00 7.60311425e-01 3.35967273e-01
7.87331641e-01 -8.31326842e-01 -6.08493090e-01 -7.73227960e-02
-8.51953253e-02 -2.28142850e-02 3.11381072e-01 6.97029650e-01
-7.58840859e-01 2.23196730e-01 -3.59308600e-01 -3.73365968e-01
-8.97596061e-01 4.77173448e-01 6.33461058e-01 -2.50282348e-03
-8.66729379e-01 7.16737986e-01 3.07717472e-01 -9.11109447e-01
3.70511621e-01 4.21955995e-02 -5.28524935e-01 -6.17080927e-01
4.04071867e-01 4.68702853e-01 -9.69642878e-01 -7.99041450e-01
-4.75239515e-01 2.21752793e-01 1.91009734e-02 -3.88615906e-01
1.26704931e+00 -6.15913272e-01 1.27403811e-01 3.31016600e-01
4.31079000e-01 -2.99333986e-02 -1.94408512e+00 -3.40314321e-02
-1.41251504e-01 -4.64130705e-03 -4.78732526e-01 -4.57142591e-01
-6.29379392e-01 3.22078198e-01 1.56136826e-01 2.60497421e-01
5.13244271e-01 1.85919069e-02 6.97678924e-01 9.19426978e-01
1.14503396e+00 -8.45057249e-01 2.02168062e-01 1.15735471e+00
8.77620339e-01 -1.46961927e+00 -3.65957499e-01 2.77414359e-02
-1.29848206e+00 7.67187476e-01 9.58118558e-01 -4.65554148e-01
4.40890729e-01 3.23148638e-01 3.67131293e-01 -2.09823102e-01
-9.21331763e-01 -3.60281289e-01 -6.00881338e-01 1.08208942e+00
-3.44997942e-01 1.99538335e-01 2.43985921e-01 3.36662620e-01
-1.69745207e-01 -8.61412063e-02 9.62395191e-01 1.12824404e+00
-9.92562354e-01 -5.65048337e-01 -2.85270274e-01 1.11577608e-01
-2.19749715e-02 3.41560185e-01 -2.60350466e-01 7.40688562e-01
-2.71774858e-01 1.38190711e+00 1.47465430e-02 -5.95713198e-01
1.98817223e-01 -3.72784138e-01 9.96325761e-02 -4.30360883e-01
-2.05872342e-01 -1.14750445e-01 2.76593208e-01 -7.38677263e-01
-2.83620179e-01 -7.36053050e-01 -1.77477193e+00 -4.28646773e-01
3.09763495e-02 2.20278293e-01 1.47268474e-01 9.11669314e-01
5.02536058e-01 2.49342427e-01 4.54492718e-01 -1.38612807e+00
-1.57882690e-01 -8.68407786e-01 -4.05297279e-01 -9.54779461e-02
7.69541502e-01 -8.35364044e-01 -3.29653114e-01 -2.84219444e-01] | [4.765652179718018, 1.0544382333755493] |
343e5e3a-9ce3-4cbc-8d50-de2731835265 | repairing-adversarial-texts-through | 2201.02504 | null | https://arxiv.org/abs/2201.02504v1 | https://arxiv.org/pdf/2201.02504v1.pdf | Repairing Adversarial Texts through Perturbation | It is known that neural networks are subject to attacks through adversarial perturbations, i.e., inputs which are maliciously crafted through perturbations to induce wrong predictions. Furthermore, such attacks are impossible to eliminate, i.e., the adversarial perturbation is still possible after applying mitigation methods such as adversarial training. Multiple approaches have been developed to detect and reject such adversarial inputs, mostly in the image domain. Rejecting suspicious inputs however may not be always feasible or ideal. First, normal inputs may be rejected due to false alarms generated by the detection algorithm. Second, denial-of-service attacks may be conducted by feeding such systems with adversarial inputs. To address the gap, in this work, we propose an approach to automatically repair adversarial texts at runtime. Given a text which is suspected to be adversarial, we novelly apply multiple adversarial perturbation methods in a positive way to identify a repair, i.e., a slightly mutated but semantically equivalent text that the neural network correctly classifies. Our approach has been experimented with multiple models trained for natural language processing tasks and the results show that our approach is effective, i.e., it successfully repairs about 80\% of the adversarial texts. Furthermore, depending on the applied perturbation method, an adversarial text could be repaired in as short as one second on average. | ['Jin Song Dong', 'Jie Shi', 'Ting Dai', 'Xinyu Wang', 'Sudipta Chattopadhyay', 'Jun Sun', 'Jingyi Wang', 'Guoliang Dong'] | 2021-12-29 | null | null | null | null | ['adversarial-text'] | ['adversarial'] | [ 7.54421890e-01 4.33370382e-01 3.67899150e-01 -1.59731865e-01
-6.14332318e-01 -1.18732762e+00 5.86299241e-01 2.65961528e-01
-2.22140566e-01 6.58864796e-01 -2.39370808e-01 -4.93085414e-01
4.51565146e-01 -1.10960495e+00 -1.24758697e+00 -6.15884006e-01
3.79688084e-01 3.77895594e-01 3.53802651e-01 -3.20132524e-01
1.38102919e-01 6.71344221e-01 -1.46239424e+00 5.85355997e-01
9.10835743e-01 5.59996545e-01 -2.62057841e-01 8.63586068e-01
-1.52662814e-01 8.99417222e-01 -1.29396307e+00 -6.92554116e-01
3.88631642e-01 -3.73990625e-01 -7.89421499e-01 4.31181956e-03
4.24359232e-01 -3.18764955e-01 -4.43495095e-01 1.60501921e+00
1.36768952e-01 2.64279842e-02 5.02190292e-01 -1.36453390e+00
-6.10244811e-01 7.03568459e-01 4.21042554e-02 -4.78920750e-02
4.40158755e-01 4.83684897e-01 5.99938750e-01 -6.86060667e-01
4.26072210e-01 1.31193519e+00 2.36170620e-01 9.88962471e-01
-1.15507019e+00 -7.61931479e-01 2.70737231e-01 -1.39462948e-01
-1.22209799e+00 -4.45626438e-01 7.97997832e-01 -3.02552134e-01
8.01943481e-01 6.91297829e-01 3.86277959e-02 1.58720291e+00
6.22740865e-01 5.82490683e-01 7.59512484e-01 -4.53979462e-01
3.59557241e-01 3.55759144e-01 -3.94314796e-01 4.81104046e-01
2.20397890e-01 2.54453868e-01 -1.96537375e-02 -3.26314718e-01
1.48802400e-01 -7.76639432e-02 -3.05241615e-01 2.80882984e-01
-9.65343416e-01 7.75025129e-01 3.32149059e-01 3.00091565e-01
-2.96027005e-01 -9.56711546e-02 7.68355131e-01 5.89695990e-01
4.73231375e-01 8.07650566e-01 -3.87144357e-01 2.29340836e-01
-5.05488694e-01 2.57761478e-01 8.20061922e-01 7.64619231e-01
3.53385568e-01 4.31912184e-01 -3.81390005e-02 6.86255455e-01
3.14269736e-02 6.53212249e-01 5.69545627e-01 -3.18086922e-01
4.95658070e-01 6.16753280e-01 6.52107596e-02 -1.16705811e+00
-6.87488988e-02 -5.86279370e-02 -9.47864354e-01 5.06422162e-01
3.60062689e-01 -2.73624748e-01 -1.10609400e+00 1.66039169e+00
2.23556295e-01 9.96391550e-02 4.23926741e-01 7.65516818e-01
5.96523821e-01 6.33554101e-01 1.28461093e-01 -4.82680909e-02
9.57637906e-01 -6.13632500e-01 -5.48155189e-01 -6.55888736e-01
4.12600219e-01 -8.73315275e-01 1.14956892e+00 5.77436686e-01
-8.62155437e-01 -2.98906177e-01 -1.15092897e+00 8.15308869e-01
-5.70483625e-01 -3.76606077e-01 9.65878069e-02 8.64383221e-01
-6.20093524e-01 6.27286971e-01 -6.39228284e-01 -6.10160567e-02
1.22137442e-01 2.68478602e-01 -3.95167828e-01 -1.80123702e-01
-1.47727573e+00 7.04456866e-01 7.50003099e-01 -1.01445794e-01
-1.23879576e+00 -2.76176602e-01 -7.15907097e-01 8.40834305e-02
4.04095709e-01 -2.84104168e-01 1.18833745e+00 -1.57928157e+00
-1.12659013e+00 5.78993738e-01 8.73577893e-02 -6.10268831e-01
8.25206101e-01 3.10665052e-02 -8.08836341e-01 2.01556355e-01
-1.37442097e-01 6.28723204e-02 1.25834632e+00 -1.36559677e+00
-3.50543171e-01 -1.60228059e-01 4.27260369e-01 -2.07026452e-01
-7.44607031e-01 2.66223758e-01 -3.17464173e-02 -1.02300942e+00
-1.51143465e-02 -9.95867968e-01 -3.06720883e-01 -2.43335143e-01
-8.40571880e-01 3.09931725e-01 7.87601054e-01 -4.94304359e-01
1.03576183e+00 -2.15394521e+00 -1.50599197e-01 4.39399213e-01
1.33670896e-01 7.90689826e-01 -1.25065520e-01 3.33110958e-01
-4.32621688e-01 6.36609495e-01 -3.53560030e-01 5.92716187e-02
-3.28798927e-02 3.81964684e-01 -9.60133731e-01 4.24151838e-01
3.84790272e-01 6.01770997e-01 -8.93224239e-01 -4.56807725e-02
7.86708593e-02 2.09616899e-01 -3.49973738e-01 4.67123002e-01
-6.03744209e-01 3.26976746e-01 -3.79753143e-01 6.57577872e-01
5.66017807e-01 1.29817948e-01 4.62738983e-02 6.29985109e-02
4.28672105e-01 3.33666392e-02 -1.10582864e+00 5.99610806e-01
-3.17272723e-01 5.29578447e-01 -1.70192301e-01 -1.08083224e+00
8.64509404e-01 6.33590162e-01 -1.62687615e-01 -2.60772943e-01
2.68967003e-01 2.33138874e-01 2.62964312e-02 -3.40560466e-01
2.82213807e-01 -1.86689347e-01 -2.59002715e-01 4.47771996e-01
-2.29174063e-01 -3.64160761e-02 -1.11303098e-01 2.06147283e-01
1.63699138e+00 -4.00876403e-01 2.30090186e-01 2.66823888e-01
8.78076613e-01 6.96324036e-02 7.26926506e-01 1.00659657e+00
-8.03077593e-02 4.64841664e-01 4.32795525e-01 -3.81387442e-01
-1.00209773e+00 -1.05114830e+00 3.27480346e-01 8.49059463e-01
-2.53483001e-02 -2.36405432e-01 -9.99181271e-01 -1.24193966e+00
-6.32347837e-02 8.53982329e-01 -5.06095290e-01 -7.94751704e-01
-4.83063608e-01 -3.30148190e-01 1.19069064e+00 4.03014868e-01
3.71373653e-01 -1.44597626e+00 -2.03401953e-01 1.83755606e-01
1.64481811e-02 -1.08504498e+00 -3.34968120e-01 2.21197471e-01
-6.71400070e-01 -1.24106896e+00 -2.33066842e-01 -5.00826955e-01
1.13551581e+00 5.35098538e-02 9.03627574e-01 4.99153167e-01
-3.95571999e-02 9.04180929e-02 -5.01605809e-01 -5.01435876e-01
-1.39177072e+00 -2.21615568e-01 3.58819544e-01 2.12103929e-02
1.58189893e-01 -3.51452142e-01 -3.69138122e-02 3.36257070e-01
-1.47447026e+00 -4.74167585e-01 5.29494941e-01 8.75086546e-01
4.97896582e-01 6.34532094e-01 7.52504170e-01 -1.31973243e+00
8.72379243e-01 -5.16206563e-01 -5.72764516e-01 3.39731067e-01
-4.43011552e-01 -8.83819237e-02 1.72406244e+00 -9.94723439e-01
-7.88818359e-01 7.18880147e-02 -5.12773097e-01 -8.56689036e-01
-6.03492677e-01 5.00974536e-01 -6.46802366e-01 -4.17280853e-01
1.16334116e+00 2.86915570e-01 -2.93005913e-01 1.48014268e-02
1.24633841e-01 6.58318102e-01 4.81234759e-01 -4.05299157e-01
1.39951301e+00 1.03692636e-01 -2.89253742e-01 -5.47628224e-01
-6.91802680e-01 1.96867928e-01 -1.49078116e-01 -2.75421739e-01
4.29999560e-01 -4.20830935e-01 -5.72518706e-01 7.72746503e-01
-1.24918795e+00 -7.33319819e-02 -7.63681009e-02 1.51473224e-01
-2.02868462e-01 4.09813404e-01 -5.42831600e-01 -8.15666676e-01
-4.76280510e-01 -1.22901070e+00 3.99990112e-01 1.04491606e-01
-2.56483793e-01 -7.44614780e-01 -3.53083074e-01 6.31095022e-02
3.59313577e-01 5.41151881e-01 9.76366282e-01 -1.39623165e+00
-3.71056557e-01 -7.63775706e-01 3.23712081e-01 7.26734757e-01
1.72575772e-01 2.90241748e-01 -9.61253405e-01 -3.66512001e-01
8.90528113e-02 -3.75401050e-01 4.04712379e-01 -3.47994566e-01
1.14818680e+00 -8.93629968e-01 -2.24713966e-01 3.08695585e-01
1.11008656e+00 5.64036071e-01 7.12009311e-01 3.68406475e-01
6.61660492e-01 3.83790374e-01 7.21174717e-01 1.82577804e-01
-3.12392801e-01 2.71415561e-01 8.99796665e-01 4.98012863e-02
4.26964760e-01 -2.57450402e-01 6.18299067e-01 3.01846743e-01
5.34584641e-01 -7.14310765e-01 -9.43187654e-01 3.22562784e-01
-1.49236214e+00 -1.01845789e+00 8.52745250e-02 2.20629573e+00
9.20505345e-01 7.69967794e-01 -2.66334772e-01 4.44386572e-01
8.91733229e-01 8.42637196e-02 -9.16744649e-01 -8.40230405e-01
-8.12194198e-02 2.07542062e-01 5.28409779e-01 3.50340903e-01
-1.13262975e+00 1.11306798e+00 5.50832891e+00 4.97956514e-01
-1.34076762e+00 -5.47677539e-02 4.84396756e-01 1.55164981e-02
-4.66799617e-01 -1.95032313e-01 -5.00760078e-01 5.17524719e-01
9.77252066e-01 -3.27142626e-01 5.00965953e-01 1.02543175e+00
-1.01882845e-01 3.46064121e-01 -9.04424131e-01 3.67457628e-01
3.13093007e-01 -1.02820635e+00 2.73426712e-01 -2.78088689e-01
5.11648178e-01 -2.16616124e-01 -9.14734602e-02 4.22573000e-01
5.95472634e-01 -9.85897183e-01 7.75347590e-01 3.65271002e-01
5.97020507e-01 -1.13971424e+00 8.67759287e-01 8.22095692e-01
-7.70508051e-01 -4.92170118e-02 -3.64098221e-01 1.68258205e-01
-2.75870502e-01 5.01631916e-01 -1.16225338e+00 3.39308500e-01
5.07207930e-01 4.89487350e-02 -5.82174242e-01 6.09442949e-01
-6.78150594e-01 7.77848780e-01 -8.98378119e-02 7.87554309e-03
-3.71222682e-02 2.19634131e-01 8.08425426e-01 8.79676938e-01
2.19121560e-01 2.44465396e-02 2.25011244e-01 7.72371352e-01
-2.58648336e-01 -1.46979894e-02 -1.01737952e+00 -3.97990465e-01
5.85430741e-01 8.72878194e-01 -6.49879158e-01 -3.10176522e-01
-8.14582109e-02 1.28767586e+00 1.52895138e-01 3.65327090e-01
-1.02478588e+00 -6.55015469e-01 7.74081647e-01 -1.15077712e-01
1.06251799e-01 2.06353009e-01 -1.34522006e-01 -1.18780601e+00
3.07274699e-01 -1.34428751e+00 3.12676728e-01 -6.26959741e-01
-1.42570364e+00 1.07825899e+00 -4.46471661e-01 -1.32874465e+00
-4.35344368e-01 -4.00823236e-01 -8.41573060e-01 8.60093117e-01
-9.01325047e-01 -7.06735969e-01 -6.61925822e-02 6.56108379e-01
4.66969043e-01 -4.23265159e-01 1.09767723e+00 1.10389419e-01
-6.19405210e-01 1.10705352e+00 -1.63024679e-01 3.70448798e-01
8.18638384e-01 -1.05555582e+00 6.99430823e-01 1.56952298e+00
3.09630837e-02 7.03865588e-01 1.03331900e+00 -8.66630733e-01
-1.34481847e+00 -1.63451540e+00 6.44532144e-01 -2.14052245e-01
6.74916267e-01 -5.14206886e-01 -1.34629369e+00 8.53126347e-01
-1.08392119e-01 4.16462421e-02 6.14410579e-01 -4.22505796e-01
-5.76984823e-01 8.74437541e-02 -1.66176856e+00 9.05462980e-01
6.62130654e-01 -5.81722915e-01 -6.80021882e-01 6.64798975e-01
1.02732432e+00 -6.24195576e-01 -6.21806681e-01 4.03172880e-01
2.75844485e-02 -7.15295732e-01 9.03550923e-01 -8.42534959e-01
5.20963073e-01 -4.68593478e-01 -1.62611336e-01 -1.46283507e+00
4.03498933e-02 -4.06118780e-01 -1.32347852e-01 1.36220253e+00
4.37032968e-01 -9.64379430e-01 5.69187999e-01 6.53364122e-01
-1.55153304e-01 -5.47282696e-01 -8.20316255e-01 -8.34636569e-01
4.77273166e-02 -5.91685414e-01 7.69927025e-01 1.22706914e+00
-1.91675514e-01 4.61133244e-03 -3.94329280e-01 7.55318463e-01
2.47028589e-01 -2.69360930e-01 7.93466687e-01 -9.22370195e-01
-2.86847800e-01 -2.25389704e-01 -6.08066559e-01 -4.48919415e-01
6.68459654e-01 -8.79416585e-01 4.59295124e-01 -9.69214678e-01
-4.87950921e-01 -2.69596338e-01 -1.27304837e-01 8.17769706e-01
-4.01145607e-01 2.21675038e-01 1.03536092e-01 1.19908459e-01
-2.48598494e-02 2.09359571e-01 7.78166711e-01 -5.35402775e-01
1.04188293e-01 3.08859527e-01 -7.01704741e-01 8.97441149e-01
1.11645532e+00 -7.51670241e-01 -3.37855577e-01 -1.88324854e-01
2.59990990e-01 8.95225536e-03 3.59745026e-01 -9.45452809e-01
1.37868404e-01 -3.68509889e-01 3.24763775e-01 -1.34698957e-01
-5.58850318e-02 -9.02744830e-01 9.30667669e-02 5.90273917e-01
-3.39822680e-01 3.14469375e-02 2.80264258e-01 6.08756900e-01
-2.24270895e-01 -5.31345665e-01 9.15909648e-01 -1.61052138e-01
-4.23920512e-01 7.12619051e-02 -7.65759408e-01 -3.56730893e-02
1.26111805e+00 1.35419993e-02 -5.33846796e-01 -2.98976988e-01
-7.08640397e-01 -4.15420383e-02 6.68197811e-01 5.30016661e-01
9.10401881e-01 -1.07583630e+00 -5.21688819e-01 4.46796268e-01
2.84010857e-01 -1.48529321e-01 9.47863981e-02 5.91184981e-02
-3.69250715e-01 -7.40786046e-02 5.14867716e-02 -2.15195253e-01
-1.64815974e+00 1.03801942e+00 4.57961380e-01 -6.39095381e-02
-5.05725324e-01 7.41636992e-01 -1.36873536e-02 -6.23030245e-01
2.68159807e-01 -4.85969484e-02 -1.04631342e-01 -3.23496968e-01
7.51208425e-01 -1.87762320e-01 3.69158745e-01 -6.55817151e-01
-3.91343743e-01 -4.31639627e-02 -3.27759653e-01 1.30230531e-01
8.97232771e-01 3.88639957e-01 -7.15692565e-02 2.00291872e-01
9.07619834e-01 2.42004827e-01 -7.26783097e-01 -2.19353765e-01
-1.96922913e-01 -6.59951925e-01 -3.77620488e-01 -8.88479769e-01
-1.11031735e+00 6.07131660e-01 4.24339622e-01 6.14497006e-01
1.37058640e+00 -3.15889090e-01 7.56285727e-01 6.24422073e-01
2.95053720e-01 -6.46345794e-01 -7.60069564e-02 6.23765945e-01
8.25586498e-01 -1.11133969e+00 -4.35639739e-01 -4.11636621e-01
-6.55066252e-01 1.20562637e+00 8.50620270e-01 -1.06812961e-01
2.35681653e-01 3.59515995e-01 2.25224361e-01 1.02553926e-01
-7.99257994e-01 3.44449610e-01 1.03239849e-01 4.05664325e-01
6.07682168e-02 1.79344594e-01 8.98516737e-04 5.14530301e-01
-3.64166170e-01 -4.32134718e-01 8.08824539e-01 1.08010328e+00
-4.06643510e-01 -1.06404269e+00 -8.06381643e-01 4.87629890e-01
-7.46218324e-01 -1.79580197e-01 -7.67045200e-01 3.25215816e-01
-2.15603244e-02 1.27162910e+00 -3.67987901e-01 -9.03021038e-01
7.18317926e-01 1.62887514e-01 -2.42414996e-02 -7.88561583e-01
-1.00851202e+00 -4.46093619e-01 -2.74313781e-02 -4.25950080e-01
2.49116719e-01 -2.56369948e-01 -1.36849904e+00 -3.61361444e-01
-2.40162894e-01 7.32536018e-02 4.54956174e-01 1.14487910e+00
1.77732512e-01 7.49377072e-01 1.00051785e+00 -4.28721756e-01
-8.75807047e-01 -6.95298851e-01 -1.35078847e-01 5.54287255e-01
2.57491559e-01 -3.16033572e-01 -9.46811676e-01 7.54729658e-02] | [5.903046131134033, 8.01942253112793] |
f455c3e7-3873-4be7-aec2-84acde5cfd63 | hallucinated-adversarial-control-for | 2303.01076 | null | https://arxiv.org/abs/2303.01076v2 | https://arxiv.org/pdf/2303.01076v2.pdf | Hallucinated Adversarial Control for Conservative Offline Policy Evaluation | We study the problem of conservative off-policy evaluation (COPE) where given an offline dataset of environment interactions, collected by other agents, we seek to obtain a (tight) lower bound on a policy's performance. This is crucial when deciding whether a given policy satisfies certain minimal performance/safety criteria before it can be deployed in the real world. To this end, we introduce HAMBO, which builds on an uncertainty-aware learned model of the transition dynamics. To form a conservative estimate of the policy's performance, HAMBO hallucinates worst-case trajectories that the policy may take, within the margin of the models' epistemic confidence regions. We prove that the resulting COPE estimates are valid lower bounds, and, under regularity conditions, show their convergence to the true expected return. Finally, we discuss scalable variants of our approach based on Bayesian Neural Networks and empirically demonstrate that they yield reliable and tight lower bounds in various continuous control environments. | ['Andreas Krause', 'Parnian Kassraie', 'Tobias Birchler', 'Bhavya Sukhija', 'Jonas Rothfuss'] | 2023-03-02 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-1.84922934e-01 3.87925148e-01 -3.17281634e-01 9.02520120e-02
-1.00156379e+00 -9.37340498e-01 5.63102722e-01 4.00141418e-01
-5.30044436e-01 1.05857539e+00 -1.39072955e-01 -6.59287810e-01
-4.56718355e-01 -6.37472868e-01 -1.01522243e+00 -7.67460704e-01
-6.05954230e-01 6.80940390e-01 2.86078066e-01 1.31118387e-01
9.58080664e-02 4.24770027e-01 -1.14456165e+00 -3.60942364e-01
5.94691157e-01 1.20692360e+00 -1.00871153e-01 1.01361001e+00
7.61178255e-01 8.47181916e-01 -6.78692877e-01 -7.86279514e-02
2.94526517e-01 -7.31157372e-03 -4.26347584e-01 -5.80332018e-02
-3.12540114e-01 -7.86028266e-01 -8.89831036e-02 1.49955606e+00
3.12022328e-01 3.21763903e-01 7.07402170e-01 -1.33822834e+00
-5.45190498e-02 8.37265015e-01 -2.34154195e-01 1.53866693e-01
2.18454868e-01 4.64503169e-01 6.27083361e-01 -1.50170259e-03
4.65505809e-01 1.21794343e+00 4.89313573e-01 5.05318403e-01
-1.35731053e+00 -4.29603249e-01 5.32233477e-01 -2.25901008e-01
-1.08140326e+00 -4.99962568e-01 2.00381711e-01 -4.67893392e-01
6.11914456e-01 1.86309278e-01 5.92692912e-01 1.26115382e+00
2.33494952e-01 6.22957587e-01 1.20750642e+00 -2.10247800e-01
7.78899610e-01 2.41998225e-01 -2.11720765e-01 4.96403098e-01
5.59104741e-01 7.83825219e-01 -1.40816718e-01 -5.80524743e-01
4.53003764e-01 -3.33437711e-01 -4.01698142e-01 -4.35009360e-01
-8.59104216e-01 5.87304950e-01 -3.11738960e-02 -3.09663981e-01
-4.90361899e-01 4.48294133e-01 2.62496114e-01 2.19283581e-01
2.46144325e-01 4.16304350e-01 -3.41739863e-01 -3.71986389e-01
-4.42315191e-01 6.05337381e-01 1.03848433e+00 1.03587842e+00
2.59601802e-01 -7.67891258e-02 -3.11227649e-01 3.20158191e-02
2.76373506e-01 7.33738601e-01 -2.83630401e-01 -1.24983406e+00
3.97965163e-01 -2.46087641e-01 1.18066823e+00 -6.42201781e-01
-1.61395520e-01 -4.89272922e-01 -6.16440624e-02 3.39006215e-01
7.03470349e-01 -5.45861542e-01 -5.07912993e-01 2.05080581e+00
4.49193299e-01 3.38323146e-01 2.02834561e-01 7.12607026e-01
-5.39485335e-01 7.13526368e-01 1.36254877e-01 -5.97869992e-01
8.69385719e-01 -3.39868575e-01 -4.47330594e-01 -1.02456547e-01
2.87980229e-01 -1.78761065e-01 7.18873382e-01 6.64238632e-01
-1.16825819e+00 1.68840811e-01 -1.01463771e+00 7.84805119e-01
9.02589224e-03 -2.31279075e-01 2.28565782e-01 5.71942806e-01
-6.97825372e-01 6.80679262e-01 -1.18759239e+00 -5.67915216e-02
1.76626623e-01 2.75163770e-01 1.33112803e-01 4.66810852e-01
-8.83393466e-01 1.07182920e+00 5.33803403e-01 1.99544221e-01
-1.97351265e+00 -4.43843156e-01 -5.24150312e-01 5.41199334e-02
1.15310609e+00 -2.95489252e-01 1.98389673e+00 -5.32008767e-01
-1.64663637e+00 1.60412133e-01 3.29884827e-01 -9.00047123e-01
9.97797430e-01 -3.24323833e-01 -1.47988304e-01 1.52493253e-01
-6.09314702e-02 -4.21291143e-02 8.68458807e-01 -1.38765180e+00
-9.67000544e-01 -2.06534073e-01 5.47518611e-01 1.09019391e-01
2.32581422e-02 -1.24440044e-01 -9.60654616e-02 -7.77156502e-02
-5.72403848e-01 -1.16876364e+00 -6.58130467e-01 -2.34170407e-01
-5.93206763e-01 -1.29372194e-01 2.72076607e-01 -4.22638834e-01
1.07933199e+00 -1.78464675e+00 -1.73717260e-01 5.72717130e-01
-1.15271606e-01 2.94314381e-02 1.93974823e-01 5.71849167e-01
5.25265574e-01 1.29827201e-01 -1.71172172e-01 -9.87337828e-02
4.42282557e-01 1.93176195e-01 -7.92050064e-01 1.00173163e+00
-9.70596150e-02 3.44418377e-01 -1.04403651e+00 -1.05246291e-01
1.41679540e-01 -1.19668841e-01 -4.89632934e-01 4.10703480e-01
-7.75763273e-01 6.53326988e-01 -9.32542801e-01 1.47196636e-01
3.07220608e-01 1.24339417e-01 4.95477706e-01 3.77196461e-01
-1.79019004e-01 -5.37934434e-03 -1.04359865e+00 8.94232512e-01
-5.54427087e-01 2.47993916e-01 4.11437899e-01 -9.75489676e-01
3.78093570e-01 1.58618510e-01 3.98168623e-01 -1.63120717e-01
5.53468525e-01 1.20140240e-01 -1.80414140e-01 -4.72457975e-01
2.73673594e-01 -1.87895149e-01 -3.13277274e-01 5.66542804e-01
-2.83498108e-01 7.27798836e-03 -4.82206093e-03 -9.61543545e-02
1.12146902e+00 3.25577050e-01 3.62381786e-01 -4.98991907e-01
1.51778713e-01 -1.07208574e-02 6.40174329e-01 1.27127731e+00
-4.51313108e-01 -3.28452677e-01 1.02233875e+00 -7.90616199e-02
-9.45826173e-01 -1.07126594e+00 -6.85118604e-03 8.34119141e-01
2.94095337e-01 -8.84536579e-02 -9.07078445e-01 -6.39289677e-01
1.62654743e-01 9.45177734e-01 -7.35153496e-01 -1.73824534e-01
-2.45154172e-01 -5.07668734e-01 2.89484352e-01 3.88911873e-01
1.10420711e-01 -4.56922740e-01 -1.31603062e+00 2.99208999e-01
2.10595027e-01 -1.02607989e+00 -4.51873541e-01 1.74720570e-01
-4.57375139e-01 -1.37212813e+00 -3.76996398e-01 5.04479045e-03
5.40452838e-01 -1.97590202e-01 7.83376992e-01 -4.65216279e-01
4.45093781e-01 7.37195849e-01 -1.60562864e-03 -5.99560618e-01
-6.48209512e-01 -1.78364396e-01 4.66619670e-01 -1.50013298e-01
-1.21335469e-01 -4.55685437e-01 -4.65322793e-01 1.68140367e-01
-6.59593701e-01 -3.31082165e-01 2.44712949e-01 5.84371924e-01
5.48321366e-01 4.33471411e-01 4.51131672e-01 -6.45494878e-01
1.08841491e+00 -5.69034696e-01 -1.77213717e+00 4.58296090e-01
-5.81618190e-01 3.05420518e-01 8.04959774e-01 -6.82091951e-01
-1.11424565e+00 -1.46860912e-01 4.07761276e-01 -4.09970105e-01
-1.45973504e-01 4.63712186e-01 -7.29217306e-02 2.26574406e-01
6.06169522e-01 -6.21066652e-02 -9.85116586e-02 -3.78349215e-01
3.28542739e-01 5.88913262e-01 8.37668896e-01 -1.46146846e+00
7.20149338e-01 5.55438161e-01 1.63157269e-01 -4.68740880e-01
-9.68730569e-01 -8.55551437e-02 2.18021348e-02 -4.27255034e-01
4.39686507e-01 -7.29235351e-01 -1.59868193e+00 -2.01019756e-02
-8.94651055e-01 -8.20273817e-01 -2.03528851e-01 5.66099346e-01
-1.09344232e+00 9.99389961e-02 -3.43867898e-01 -1.82921290e+00
2.46416613e-01 -1.06642425e+00 6.67761803e-01 2.74263531e-01
1.98428199e-01 -9.18639243e-01 3.29208106e-01 -3.03604662e-01
1.52274162e-01 5.71810007e-01 6.23180985e-01 -8.19384634e-01
-5.21929860e-01 -1.27394900e-01 1.35271445e-01 2.17684612e-01
-4.15775836e-01 9.76488814e-02 -8.34234715e-01 -5.49933493e-01
1.23299204e-01 -1.62578627e-01 6.31478846e-01 5.64431608e-01
1.09213638e+00 -1.10614121e+00 -4.68030751e-01 2.54405528e-01
1.26932836e+00 5.70635319e-01 -2.09585167e-02 4.97406036e-01
-1.45937085e-01 6.08756363e-01 8.51932347e-01 1.09204233e+00
2.36614883e-01 5.41993558e-01 7.27742255e-01 7.85705268e-01
6.63649797e-01 -5.02410710e-01 7.93537676e-01 2.15342715e-01
-1.30202330e-03 -2.61537254e-01 -6.16619468e-01 5.34074187e-01
-2.11366343e+00 -9.62947786e-01 5.23270845e-01 2.91627336e+00
8.75509560e-01 5.65992355e-01 5.64827263e-01 -3.90035033e-01
7.74495244e-01 -1.75477505e-01 -1.22998500e+00 -4.42186773e-01
3.23815882e-01 -1.84734181e-01 1.08399701e+00 6.84793174e-01
-1.06248236e+00 5.30297279e-01 7.02379608e+00 6.09659255e-01
-7.88277566e-01 1.79365799e-01 7.28187323e-01 -1.71715961e-04
-9.30522010e-02 9.77038667e-02 -7.44540274e-01 5.10336936e-01
1.46059179e+00 -6.56907976e-01 7.26017118e-01 1.16831052e+00
4.63086993e-01 -4.35218483e-01 -1.29497182e+00 4.05260682e-01
-6.17739439e-01 -1.07085204e+00 -6.47049963e-01 3.47517222e-01
7.31055796e-01 -3.41847874e-02 8.89815390e-02 4.09086525e-01
1.15177441e+00 -8.32144141e-01 1.28766119e+00 5.81150830e-01
5.45843780e-01 -1.17540753e+00 5.77719688e-01 7.55517006e-01
-9.00168836e-01 -5.78217566e-01 -3.42877746e-01 -1.21418491e-01
1.67210370e-01 4.70998675e-01 -9.22357321e-01 3.10517073e-01
2.38070473e-01 2.55078256e-01 1.31178066e-01 1.09864724e+00
-4.57606673e-01 6.26373649e-01 -7.34069049e-01 -2.34839663e-01
4.41919267e-01 -1.65363014e-01 8.10423136e-01 8.53835166e-01
3.51786256e-01 4.24121320e-03 6.28521681e-01 9.17657137e-01
1.53470710e-01 -5.26209712e-01 -7.01170743e-01 -2.09473714e-01
7.07937598e-01 7.77025342e-01 -5.56075573e-01 -3.39211166e-01
1.55504689e-01 2.44873866e-01 2.70299077e-01 5.38708806e-01
-1.06923711e+00 -1.30962044e-01 8.09940398e-01 -4.96977448e-01
1.81707487e-01 -1.50619164e-01 1.78192183e-01 -8.99615228e-01
1.43279687e-01 -8.01024973e-01 4.12725419e-01 -2.34451860e-01
-1.09709406e+00 2.71396726e-01 4.11306053e-01 -1.08493257e+00
-5.66798568e-01 -6.38519466e-01 -5.05110979e-01 4.82423991e-01
-1.14575112e+00 -5.26224852e-01 5.16069353e-01 1.69730261e-01
1.79630160e-01 2.22354814e-01 5.52868545e-01 -3.85329932e-01
-7.46591687e-01 1.17153347e-01 4.09181625e-01 -2.27999002e-01
1.48616835e-01 -1.37652302e+00 -4.81853969e-02 1.05110204e+00
-3.58640879e-01 2.86703318e-01 1.28593600e+00 -7.91236639e-01
-1.52022064e+00 -1.14649439e+00 -2.02260807e-01 -4.44761992e-01
1.08228052e+00 -4.52797525e-02 -4.81598288e-01 9.12225187e-01
-1.37965158e-01 1.13776937e-01 4.42696437e-02 5.55950925e-02
-1.94669023e-01 -3.49356197e-02 -1.19136953e+00 7.43039429e-01
7.96384573e-01 -3.49736273e-01 -4.25237328e-01 3.09450209e-01
6.27512634e-01 -4.61073071e-01 -8.03209662e-01 3.76578331e-01
7.53952324e-01 -8.16217840e-01 5.63773155e-01 -1.13828206e+00
-2.54393458e-01 -3.75933766e-01 -2.93879002e-01 -1.46167767e+00
1.61707103e-01 -1.32559597e+00 -5.09424865e-01 7.75834024e-01
2.41055503e-01 -8.27121198e-01 3.99674833e-01 7.42083728e-01
1.32208973e-01 -6.92123413e-01 -1.25079381e+00 -1.29284811e+00
2.43155509e-01 -7.56577849e-01 6.73481166e-01 2.88750470e-01
1.07805170e-01 -4.09034342e-01 -5.64549565e-01 8.04198921e-01
1.03791845e+00 2.15921551e-02 5.51390707e-01 -8.11211050e-01
-5.37766576e-01 -5.21785438e-01 7.69476667e-02 -9.70081687e-01
4.66422647e-01 -1.65727973e-01 6.34522557e-01 -9.86965835e-01
1.41767144e-01 -6.20778680e-01 -4.10766453e-01 9.02899504e-02
1.84546307e-01 -5.66133440e-01 2.17281178e-01 6.74332073e-03
-9.28679049e-01 6.36220515e-01 9.09016192e-01 -7.14155212e-02
-1.99077010e-01 4.64411765e-01 -5.56699634e-01 8.40384066e-01
7.54189730e-01 -5.12682974e-01 -5.07278740e-01 -8.27911571e-02
3.58763158e-01 5.34769177e-01 2.99230248e-01 -7.86318958e-01
1.41754076e-01 -8.60007286e-01 -1.14918537e-01 -3.48260343e-01
1.39817089e-01 -6.83082640e-01 2.37501323e-01 7.63778985e-01
-8.99978936e-01 -1.11113280e-01 -2.55603958e-02 1.15021074e+00
5.14151335e-01 -2.88298130e-01 7.67986953e-01 1.58956682e-03
-2.52612531e-01 3.61187905e-01 -5.91945052e-01 9.36618224e-02
1.20570362e+00 4.81612235e-01 -3.24727774e-01 -6.96177423e-01
-8.81746650e-01 6.68807745e-01 3.77975345e-01 -1.05318449e-01
2.87585288e-01 -9.71947134e-01 -3.48142236e-01 -1.85706049e-01
-1.17765227e-02 -1.31302968e-01 -6.07497618e-02 8.94217551e-01
-2.80614551e-02 4.44403887e-01 1.97126210e-01 -3.61888796e-01
-5.19775808e-01 8.80549729e-01 6.07139170e-01 -3.07142705e-01
-4.35703814e-01 2.99424708e-01 1.96443483e-01 -1.10185042e-01
6.02374494e-01 -5.80515862e-01 1.89010456e-01 -4.00856555e-01
7.38444448e-01 5.11439681e-01 -2.87499189e-01 -2.30673417e-01
-1.99845120e-01 -5.82334632e-03 2.87008941e-01 -7.51399159e-01
1.17221427e+00 -2.59228289e-01 2.81625807e-01 4.93908763e-01
5.88165700e-01 9.80544090e-03 -2.06511402e+00 -1.03369541e-01
4.31377977e-01 -5.07902861e-01 1.00046419e-01 -8.07425320e-01
-5.91446340e-01 4.63253558e-01 4.56924915e-01 6.22556031e-01
8.43913198e-01 -1.47641018e-01 3.34895194e-01 8.05255353e-01
7.55326688e-01 -1.29174590e+00 -5.63456178e-01 2.35433072e-01
6.67229474e-01 -7.81703115e-01 -1.23381250e-01 2.16693878e-01
-4.27022785e-01 1.00519073e+00 2.33033866e-01 -1.64980590e-01
5.45332313e-01 3.79755557e-01 -3.76722962e-01 2.54420400e-01
-1.29274416e+00 -2.97770083e-01 -1.57315210e-01 7.60455966e-01
-5.07617176e-01 6.11399055e-01 1.47812828e-01 5.21686971e-01
4.65286011e-03 -5.12442179e-02 7.66349852e-01 8.31244886e-01
-7.48594105e-01 -5.97571552e-01 -4.55230951e-01 2.39485323e-01
-5.64911366e-01 4.40336168e-01 6.77356347e-02 6.88033104e-01
-2.89814889e-01 1.16106081e+00 -2.56726652e-01 5.97907305e-02
1.83150887e-01 -3.26358199e-01 5.19905925e-01 -3.03230017e-01
7.53042921e-02 1.99664280e-01 4.64161932e-01 -7.38578081e-01
-2.35453740e-01 -6.13276720e-01 -8.56705010e-01 -3.22361201e-01
-2.36644179e-01 5.08014619e-01 5.97715914e-01 1.15943301e+00
1.06762312e-01 2.24746376e-01 8.43710780e-01 -7.63895452e-01
-1.62363303e+00 -7.95111418e-01 -4.86235648e-01 -2.07227722e-01
6.97086871e-01 -9.00141239e-01 -5.24137497e-01 -2.00721905e-01] | [4.3677802085876465, 2.4467296600341797] |
340335e6-b664-4f7d-b8d5-1e08f65957c4 | secoda-sense-complexity-dataset | null | null | https://aclanthology.org/2020.lrec-1.730 | https://aclanthology.org/2020.lrec-1.730.pdf | SeCoDa: Sense Complexity Dataset | The Sense Complexity Dataset (SeCoDa) provides a corpus that is annotated jointly for complexity and word senses. It thus provides a valuable resource for both word sense disambiguation and the task of complex word identification. The intention is that this dataset will be used to identify complexity at the level of word senses rather than word tokens. For word sense annotation SeCoDa uses a hierarchical scheme that is based on information available in the Cambridge Advanced Learner{'}s Dictionary. This way we can offer more coarse-grained senses than directly available in WordNet. | ['Shiva Taslimipoor', 'David Strohmaier', 'Sian Gooding', 'Ekaterina Kochmar'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['complex-word-identification'] | ['natural-language-processing'] | [-3.14476527e-02 8.77648592e-02 -1.71545163e-01 -9.89747420e-02
-4.38726127e-01 -1.03402567e+00 6.91744685e-01 1.13476086e+00
-1.17995787e+00 6.20123565e-01 7.09559321e-01 -4.64153767e-01
-2.54585803e-01 -7.34676898e-01 4.28000301e-01 -2.19428912e-01
8.68590996e-02 4.87530828e-01 1.77197248e-01 -8.24136257e-01
4.16492850e-01 2.45274946e-01 -1.72407436e+00 4.14273664e-02
8.86820972e-01 6.34428620e-01 7.92026579e-01 4.16888565e-01
-6.20111406e-01 2.65676379e-01 -5.40039480e-01 -2.66240418e-01
-3.03439908e-02 1.09799497e-01 -1.21643782e+00 -3.49273741e-01
1.09063111e-01 5.20958602e-01 7.45977834e-02 1.10820687e+00
5.45798898e-01 3.55991006e-01 5.61455488e-01 -6.10919714e-01
1.06932819e-01 9.40861583e-01 2.97647357e-01 6.52389526e-01
1.01749861e+00 -1.37140781e-01 1.67903483e+00 -7.09167778e-01
8.22176337e-01 1.42133176e+00 3.86026442e-01 2.49237105e-01
-8.58906150e-01 -5.87091804e-01 3.13489825e-01 2.83103526e-01
-1.44759440e+00 -1.90589335e-02 3.68270338e-01 -4.90258902e-01
1.64298272e+00 2.82662004e-01 1.11278105e+00 8.77034485e-01
-6.49748445e-02 5.99691212e-01 1.15647495e+00 -7.54558802e-01
2.88939446e-01 -8.11119154e-02 7.59576619e-01 6.14495337e-01
6.14607394e-01 -1.21501468e-01 -3.92991424e-01 -1.74726337e-01
5.49930274e-01 -3.06166619e-01 -1.56986713e-02 1.54916853e-01
-1.14804447e+00 8.28573287e-01 1.69776648e-01 1.13524699e+00
-2.71914989e-01 -1.88011751e-01 8.28478038e-01 4.15334612e-01
4.05629843e-01 1.24890792e+00 -9.63849187e-01 -4.07501161e-01
-6.58035994e-01 5.21500707e-01 9.75631416e-01 7.56951809e-01
6.81368232e-01 -5.04489578e-02 1.65386438e-01 1.04545438e+00
5.37680984e-01 2.69419640e-01 1.00659096e+00 -5.44517457e-01
2.23053247e-01 1.02707088e+00 -1.19816758e-01 -7.41753995e-01
-7.33039737e-01 -2.76774496e-01 -3.31834972e-01 -1.56773061e-01
1.30310386e-01 2.59106785e-01 -8.85477304e-01 1.75193393e+00
8.88936147e-02 -7.06334561e-02 1.81562632e-01 3.22084159e-01
1.14774156e+00 6.53568059e-02 8.27451468e-01 -3.79144460e-01
1.96783876e+00 -1.20614491e-01 -7.96267331e-01 -3.53180856e-01
9.06347334e-01 -7.78059006e-01 1.40584040e+00 5.49041927e-01
-7.80699313e-01 -3.66359770e-01 -1.20329559e+00 -2.11542457e-01
-1.32499301e+00 -5.83077788e-01 9.55517888e-01 7.53054559e-01
-1.15968692e+00 2.30085939e-01 -4.76960063e-01 -5.19615591e-01
8.10232908e-02 -2.19830610e-02 -4.37105060e-01 2.38482635e-02
-1.90266156e+00 1.45962441e+00 1.20474708e+00 -6.79676771e-01
-2.54636645e-01 -6.73175991e-01 -1.61217105e+00 -1.00345507e-01
3.86274874e-01 -7.25834191e-01 1.00211656e+00 -3.93104553e-01
-6.90889657e-01 1.25960135e+00 -8.35157558e-03 -3.86475921e-01
-2.36605838e-01 -7.08080456e-02 -6.89111531e-01 5.66461869e-02
5.87854207e-01 6.43875420e-01 2.10152254e-01 -8.20441186e-01
-9.42251503e-01 -3.65486771e-01 3.13420236e-01 5.32563806e-01
-3.97444099e-01 1.23415947e-01 -1.27516776e-01 -1.09377933e+00
-1.09852865e-01 -6.72113836e-01 -3.67231786e-01 -6.24875069e-01
-4.63183969e-02 -1.15125299e+00 1.86503679e-01 -2.72881418e-01
1.89054525e+00 -1.84287310e+00 3.12759634e-03 2.64135301e-01
5.94391346e-01 3.74126405e-01 -2.30697185e-01 6.31674826e-01
-4.12548453e-01 4.21305805e-01 -1.92265481e-01 -1.95501879e-01
3.66607904e-01 7.22772539e-01 -1.07752923e-02 -4.83676326e-03
-1.21698104e-01 7.43215621e-01 -1.40141904e+00 -6.70023799e-01
4.22023326e-01 9.72659048e-03 -5.60765326e-01 -2.31577039e-01
-1.97836965e-01 -2.77552605e-01 -5.09272456e-01 4.89844590e-01
1.27445698e-01 2.50322878e-01 3.57293278e-01 -4.32402007e-02
-2.88046360e-01 7.62039602e-01 -1.23393524e+00 1.82213092e+00
-9.89155054e-01 4.95941997e-01 -5.33471227e-01 -6.35902584e-01
6.00371599e-01 4.84724015e-01 4.30700600e-01 -5.12444913e-01
1.95182934e-01 3.27725857e-01 1.71465352e-01 -3.59898150e-01
9.62952614e-01 -4.09107596e-01 -6.68359041e-01 1.53058752e-01
3.83065760e-01 -6.98309004e-01 8.91613126e-01 4.10941601e-01
1.08298600e+00 -4.62122232e-01 1.13675702e+00 -1.00713766e+00
6.43295169e-01 9.23671797e-02 4.43937689e-01 4.10720646e-01
1.08904235e-01 -5.53875007e-02 -1.64064311e-03 -3.29744726e-01
-8.02885056e-01 -1.14819741e+00 -5.76107085e-01 1.35739684e+00
-2.57178582e-02 -1.23866618e+00 -4.09981459e-01 -2.62514740e-01
-1.54771984e-01 7.72092819e-01 -3.93866569e-01 1.05109341e-01
-2.13477030e-01 -2.19515070e-01 5.77779710e-01 3.98605615e-01
1.24118134e-01 -1.35812473e+00 -5.50075114e-01 3.71609658e-01
-1.48227364e-01 -1.15357971e+00 -3.23218644e-01 2.81635821e-01
-4.38759059e-01 -1.32230890e+00 -1.10813402e-01 -9.56476271e-01
1.60202369e-01 -1.13473460e-01 1.72305453e+00 4.44503129e-01
-5.14286280e-01 6.46899343e-01 -8.77337277e-01 -7.23602653e-01
-1.93983406e-01 3.93211782e-01 4.45408911e-01 -9.40531433e-01
6.37346208e-01 -5.33484519e-01 -2.22875699e-01 -2.96266854e-01
-1.34866476e+00 -5.49156129e-01 1.78924009e-01 6.31895483e-01
3.35134417e-01 2.86828488e-01 4.96008456e-01 -9.57821429e-01
1.08577621e+00 -4.30602044e-01 -3.47791106e-01 1.07215587e-02
-1.09056604e+00 2.14920685e-01 8.44472945e-02 -2.14878067e-01
-8.30167115e-01 -5.44130206e-02 -7.01801717e-01 3.93088609e-01
-4.35975164e-01 7.73628712e-01 -4.10305917e-01 5.52728064e-02
6.78723693e-01 1.51421577e-01 -6.73162222e-01 -6.20936632e-01
5.56294799e-01 5.49020648e-01 3.71539772e-01 -6.80766761e-01
4.65185493e-01 -1.94950894e-01 -1.21101223e-01 -1.31720448e+00
-9.85964954e-01 -1.22270930e+00 -8.44469249e-01 4.23566215e-02
1.00935960e+00 -9.46147740e-01 -6.05849445e-01 7.59171173e-02
-9.49820578e-01 1.83024913e-01 -3.76713604e-01 3.07730883e-01
-3.24092835e-01 6.65443718e-01 -3.06466579e-01 -5.96321881e-01
-2.84452647e-01 -6.99672520e-01 1.06214833e+00 5.08944802e-02
-1.12869644e+00 -1.76487744e+00 2.53432840e-01 7.29247704e-02
1.33059397e-01 2.37794995e-01 1.12558222e+00 -1.06757748e+00
2.49560907e-01 -8.43128413e-02 2.95021266e-01 2.38585606e-01
2.03253925e-01 -4.32416528e-01 -4.96863008e-01 -3.42716575e-02
-1.57705933e-01 -3.78105432e-01 9.49317813e-01 -1.59746632e-01
6.37129784e-01 -3.97326559e-01 -2.26816982e-01 3.09187099e-02
1.76258838e+00 -1.18039690e-01 4.87148046e-01 5.65279007e-01
6.59280121e-01 4.22981352e-01 6.72033906e-01 4.55455869e-01
6.85334921e-01 3.35430801e-01 2.95095611e-04 2.02525333e-01
1.55502588e-01 -1.08419493e-01 -1.03562243e-01 1.00820911e+00
2.37050354e-01 -1.69566244e-01 -1.47615731e+00 1.05894113e+00
-1.38313007e+00 -7.49045014e-01 -2.65885741e-01 1.78901935e+00
1.24996209e+00 4.08184946e-01 2.13796720e-01 6.11461341e-01
4.49673206e-01 2.85706401e-01 2.04781577e-01 -6.08303666e-01
-7.33062848e-02 8.05836856e-01 3.05504441e-01 9.70265985e-01
-7.20841408e-01 1.36690378e+00 7.27877617e+00 1.19333696e+00
-3.35719585e-01 1.13494908e-02 -1.89699650e-01 3.61178905e-01
-6.95970118e-01 -2.62643304e-03 -6.28084362e-01 3.26510638e-01
5.64468324e-01 -4.19258475e-01 8.02355483e-02 5.65594196e-01
-8.90156701e-02 -4.04955536e-01 -9.71911430e-01 1.11140656e+00
-1.37747511e-01 -1.01456499e+00 4.35423851e-01 -2.74198741e-01
3.82205784e-01 4.66679633e-02 -3.29176128e-01 4.17729497e-01
6.84012175e-01 -1.04027653e+00 5.33507228e-01 4.04730171e-01
9.42970872e-01 -8.23524773e-01 7.43280590e-01 2.98356950e-01
-1.65529406e+00 1.49758160e-01 -2.19000086e-01 -5.18998146e-01
1.03705637e-02 6.84199452e-01 -6.76014543e-01 4.30534840e-01
3.31685185e-01 7.54513860e-01 -6.00792110e-01 7.65913725e-01
-3.72661889e-01 4.28798914e-01 -3.31362635e-01 -2.86221743e-01
2.65645564e-01 1.60395816e-01 8.33974183e-01 1.68302894e+00
-1.31429136e-01 4.60429490e-01 7.33365059e-01 2.72721618e-01
3.08228850e-01 3.79466712e-01 -6.16466641e-01 -3.91565979e-01
1.00240076e+00 1.20582044e+00 -8.53876054e-01 -6.81810558e-01
1.97055861e-02 5.73200285e-01 1.32762209e-01 -1.47788778e-01
9.68855321e-02 -6.58797204e-01 1.11626518e+00 4.31783684e-02
-6.24799803e-02 -5.74711442e-01 -3.43392521e-01 -1.08158374e+00
-4.24371123e-01 -5.98173797e-01 9.47449803e-01 -6.15751743e-01
-1.47844124e+00 6.11882925e-01 4.23090369e-01 -7.87591100e-01
-6.08531356e-01 -9.95705247e-01 -2.56151915e-01 8.89590442e-01
-1.18330264e+00 -8.48709106e-01 5.26470393e-02 6.67169392e-01
7.28774905e-01 -7.21499175e-02 1.25923097e+00 1.15189776e-02
1.53710321e-01 4.72817421e-01 -6.31382704e-01 3.88603806e-02
1.95305556e-01 -1.94467628e+00 4.89483118e-01 3.11382443e-01
2.45079726e-01 7.82959402e-01 1.01108623e+00 -8.26621890e-01
-7.21830606e-01 -7.27089405e-01 1.68412638e+00 -8.96300316e-01
1.13570607e+00 -1.91152707e-01 -6.67623758e-01 2.27541298e-01
1.96810290e-01 -3.80419761e-01 1.14377391e+00 2.41712213e-01
-3.71343106e-01 2.12012932e-01 -1.15683424e+00 4.69181448e-01
1.26772106e+00 -7.09740698e-01 -1.44164515e+00 2.59488583e-01
9.61697876e-01 -2.05284148e-01 -1.06663692e+00 8.10698122e-02
2.23625585e-01 -3.91114712e-01 9.49872196e-01 -4.72757846e-01
4.42689434e-02 -1.89619914e-01 -2.80102968e-01 -1.61285400e+00
-2.56720215e-01 -4.67495829e-01 3.57276022e-01 1.12639964e+00
5.10265291e-01 -6.11761391e-01 2.95863487e-02 1.67054698e-01
-2.51915246e-01 -4.20260012e-01 -1.03685844e+00 -7.69678116e-01
2.15095639e-01 -9.88531709e-01 5.76412916e-01 1.33557105e+00
6.58169746e-01 7.67322302e-01 5.30808628e-01 -3.31714123e-01
3.37303132e-01 -4.47875708e-01 -3.36746186e-01 -1.67787266e+00
2.27946088e-01 -8.40622902e-01 -8.81965995e-01 -6.47684276e-01
5.31638563e-01 -1.11123824e+00 -1.66270193e-02 -1.62000120e+00
-1.90494001e-01 -4.80196506e-01 -5.34454405e-01 6.57501698e-01
-5.08091629e-01 2.24569842e-01 1.69661999e-01 -2.81690896e-01
-6.51380122e-01 2.18060553e-01 9.12352324e-01 2.18294114e-01
-1.17473453e-02 -4.87462908e-01 -9.05230701e-01 9.38178599e-01
7.55346894e-01 -4.09004837e-01 -4.43153232e-01 -2.11625233e-01
8.39236975e-01 -2.71796018e-01 -7.54793435e-02 -7.87571907e-01
1.51955821e-02 -2.36605972e-01 3.92924249e-03 -3.80014360e-01
9.08569470e-02 -7.58475363e-01 -3.67023379e-01 5.08769631e-01
-3.61185938e-01 6.29366219e-01 3.18272710e-01 2.90101528e-01
-2.74116784e-01 -4.37692314e-01 6.06221914e-01 -7.28595495e-01
-1.11462569e+00 9.68544781e-02 -9.10819948e-01 5.16207576e-01
4.91540343e-01 -1.02537476e-01 2.24871606e-01 -9.19076502e-02
-9.06831026e-01 3.14114869e-01 4.33394134e-01 6.26525342e-01
4.15610969e-01 -1.49422646e+00 -5.23396254e-01 -2.56403953e-01
6.42651379e-01 -1.45291343e-01 -3.38296086e-01 -2.48759575e-02
-3.22236896e-01 5.61076343e-01 -9.14604068e-02 -6.93965554e-02
-1.23726010e+00 5.44603407e-01 8.04904029e-02 -4.96338487e-01
-4.86835450e-01 8.17749739e-01 -2.64389783e-01 -4.33254391e-01
1.02714241e-01 -4.88138914e-01 -9.16870654e-01 5.77160299e-01
7.39332497e-01 3.26589018e-01 1.30594134e-01 -7.90238261e-01
-4.81052905e-01 2.60245800e-01 9.58296135e-02 -3.17361087e-01
1.22989500e+00 -3.60418379e-01 -4.50993001e-01 8.06858599e-01
9.61270988e-01 1.27092600e-01 -1.01287134e-01 -2.99052566e-01
7.16760039e-01 -2.49847770e-01 6.88738897e-02 -9.60532367e-01
-1.86059758e-01 3.20403636e-01 6.01832926e-01 5.26469529e-01
1.03773534e+00 2.93664664e-01 7.24508286e-01 6.78432405e-01
3.26095998e-01 -1.26941299e+00 -1.42478704e-01 1.31731033e+00
7.55566597e-01 -8.08269441e-01 -8.11510012e-02 -5.56622267e-01
-3.00170660e-01 9.11644578e-01 3.62414598e-01 -1.31621733e-01
8.41290772e-01 1.97008491e-01 2.55399439e-02 -4.94712204e-01
-6.23268247e-01 -1.10966265e+00 5.59944391e-01 9.01748538e-01
7.13148773e-01 2.80856282e-01 -1.18274891e+00 9.66711164e-01
-8.77773464e-01 -5.58062017e-01 4.52485979e-01 7.34474301e-01
-8.65079463e-01 -1.54231882e+00 -8.94844234e-02 5.07673264e-01
-5.02937436e-01 -8.18936944e-01 -6.75881863e-01 8.85035753e-01
4.44069654e-01 9.92676795e-01 4.30775322e-02 -1.91971451e-01
3.43755126e-01 8.14621523e-02 5.17202497e-01 -1.33716142e+00
-7.64959276e-01 -3.47854108e-01 5.57214618e-01 -4.79361802e-01
-3.69450480e-01 -4.87108439e-01 -1.43111575e+00 -7.96363652e-02
-8.56073052e-02 5.41939199e-01 4.31626171e-01 1.50799036e+00
-3.46034706e-01 4.82473642e-01 1.89540416e-01 -5.80404460e-01
-3.29117887e-02 -1.18868363e+00 -8.56364608e-01 5.65097153e-01
1.43855408e-01 -7.61053801e-01 -7.97909349e-02 -2.37959608e-01] | [10.177331924438477, 9.194952964782715] |
87bbac80-d006-436f-8bcb-847d0d6bba04 | a-framework-for-information-extraction-from | 1902.10031 | null | http://arxiv.org/abs/1902.10031v1 | http://arxiv.org/pdf/1902.10031v1.pdf | A framework for information extraction from tables in biomedical literature | The scientific literature is growing exponentially, and professionals are no
more able to cope with the current amount of publications. Text mining provided
in the past methods to retrieve and extract information from text; however,
most of these approaches ignored tables and figures. The research done in
mining table data still does not have an integrated approach for mining that
would consider all complexities and challenges of a table. Our research is
examining the methods for extracting numerical (number of patients, age, gender
distribution) and textual (adverse reactions) information from tables in the
clinical literature. We present a requirement analysis template and an integral
methodology for information extraction from tables in clinical domain that
contains 7 steps: (1) table detection, (2) functional processing, (3)
structural processing, (4) semantic tagging, (5) pragmatic processing, (6) cell
selection and (7) syntactic processing and extraction. Our approach performed
with the F-measure ranged between 82 and 92%, depending on the variable, task
and its complexity. | ['Robert Hernandez', 'Goran Nenadic', 'Nikola Milosevic', 'Cassie Gregson'] | 2019-02-26 | null | null | null | null | ['table-detection'] | ['miscellaneous'] | [ 4.27278817e-01 3.13637972e-01 -2.19599888e-01 -3.24427128e-01
-6.25473499e-01 -5.69176555e-01 2.67397970e-01 1.42494702e+00
-4.30560410e-01 9.78422344e-01 6.14829242e-01 -6.44686401e-01
-2.96433151e-01 -6.30503416e-01 -1.36200666e-01 -1.20696455e-01
9.20160487e-02 6.92676365e-01 2.22478360e-01 1.19457029e-01
6.74745977e-01 3.78815591e-01 -1.53393066e+00 6.58018529e-01
8.33853722e-01 9.70369101e-01 6.08895421e-02 5.14108658e-01
-7.97522545e-01 9.70416427e-01 -6.23616040e-01 -3.40191275e-01
1.09770838e-02 -4.88652021e-01 -8.99861872e-01 6.89096525e-02
-5.57371914e-01 9.30874348e-02 3.13190162e-01 9.36252713e-01
5.29144526e-01 -1.97679088e-01 7.54349053e-01 -1.00939953e+00
-2.55746424e-01 9.91515279e-01 -3.95049930e-01 2.16224387e-01
7.56251395e-01 -3.12783033e-01 4.83927220e-01 -7.50196099e-01
1.03843105e+00 1.02281976e+00 5.20010591e-01 3.27269793e-01
-5.73897481e-01 -6.82192802e-01 -5.62506840e-02 -2.15749577e-01
-1.36395109e+00 -2.70621955e-01 2.67580122e-01 -6.38734579e-01
1.28290987e+00 4.20206010e-01 6.54869735e-01 5.52536249e-01
4.91800845e-01 4.13410291e-02 1.19163442e+00 -7.49675393e-01
4.06299144e-01 7.66959250e-01 3.14160079e-01 3.28923017e-01
7.35308886e-01 -5.92216194e-01 -4.75863844e-01 -4.06079799e-01
2.74724901e-01 -1.36865780e-01 2.74558991e-01 2.84847081e-01
-1.08081532e+00 7.58050799e-01 -5.68140566e-01 5.72404563e-01
-4.97406960e-01 -7.64929593e-01 8.69828641e-01 2.68193722e-01
2.48456389e-01 5.37702501e-01 -8.44529271e-01 -2.29328349e-01
-8.66189957e-01 2.01072469e-01 1.15793574e+00 1.33188713e+00
1.84735805e-02 -4.43572342e-01 -1.70400769e-01 4.35863584e-01
3.17169189e-01 2.41988689e-01 4.95351911e-01 -2.27473199e-01
6.57308578e-01 1.20990312e+00 7.71697313e-02 -9.86851931e-01
-6.77607000e-01 -4.97006699e-02 -6.96842372e-01 -3.74230564e-01
3.64178419e-01 -4.28071886e-01 -1.03252327e+00 1.19954264e+00
2.57866770e-01 -1.06874752e+00 1.82586327e-01 3.09080660e-01
1.28285539e+00 3.73263061e-01 5.07370651e-01 -7.94475436e-01
2.18465972e+00 -2.61677742e-01 -1.34475291e+00 -5.53752705e-02
7.52724469e-01 -1.39665771e+00 4.75037992e-01 5.30224085e-01
-1.32847130e+00 -6.45976216e-02 -9.75712419e-01 -1.26112223e-01
-1.05559433e+00 6.46699294e-02 7.02716112e-01 8.97191644e-01
-7.39957511e-01 1.81265146e-01 -3.44346046e-01 -8.01961362e-01
3.97916615e-01 5.35173714e-01 -4.50074226e-01 2.92983592e-01
-1.07510543e+00 1.03953314e+00 5.87537825e-01 -1.36220500e-01
1.64744645e-01 -6.39258027e-01 -7.42130995e-01 -4.14538160e-02
5.68158805e-01 -8.83479178e-01 6.53106689e-01 -5.01038134e-01
-9.22524452e-01 1.09559786e+00 -1.39608338e-01 -2.38540530e-01
1.91688985e-01 1.59861684e-01 -5.09757936e-01 6.14129417e-02
2.88275450e-01 1.99132830e-01 1.65754244e-01 -9.09475446e-01
-9.01616335e-01 -6.64795518e-01 -4.01771307e-01 1.67465284e-01
-3.01305074e-02 7.50356555e-01 -2.36209005e-01 -6.68544829e-01
2.13328838e-01 -3.89523119e-01 -3.37297797e-01 -3.07700366e-01
-5.28434098e-01 -2.86410004e-01 3.24028164e-01 -1.07759750e+00
2.11111259e+00 -1.82552528e+00 -3.10463488e-01 4.55829203e-01
2.45660841e-01 1.19333982e-01 8.37038219e-01 1.03544152e+00
-1.87928781e-01 7.15394557e-01 -1.49725392e-01 1.89824939e-01
-1.05302252e-01 -2.16860786e-01 -2.48121154e-02 9.94379520e-02
-5.01228087e-02 3.74881715e-01 -6.61407769e-01 -1.10738325e+00
-1.15996331e-01 3.83368552e-01 -4.06653315e-01 -5.26509136e-02
4.63686548e-02 -4.15106816e-03 -7.51952291e-01 9.69124198e-01
4.78590637e-01 -1.27274066e-01 4.52325433e-01 -1.00486718e-01
-3.70625257e-01 6.16311491e-01 -1.56013989e+00 1.26992249e+00
6.48769736e-02 -1.07334554e-01 1.37470558e-01 -7.12969661e-01
8.27425838e-01 6.18834794e-01 8.06967080e-01 -5.74471951e-01
4.17878568e-01 3.44600827e-01 1.16731808e-01 -1.04544652e+00
2.26521716e-01 -1.53202385e-01 -2.29622036e-01 4.51861113e-01
-1.00208022e-01 1.97672695e-01 5.65189540e-01 2.65653342e-01
1.03844118e+00 -1.64522827e-01 1.12293434e+00 -3.91935527e-01
7.86561370e-01 3.25826287e-01 5.37189722e-01 2.99695998e-01
-5.86867929e-02 2.25560978e-01 7.10318565e-01 -4.46201146e-01
-9.42312717e-01 -5.64792812e-01 -2.59771854e-01 2.48683363e-01
-4.08676296e-01 -6.47511780e-01 -9.22257066e-01 -4.67926443e-01
6.34402856e-02 5.07952869e-01 -6.86048210e-01 1.36301205e-01
-3.09761882e-01 -9.10024822e-01 2.49998108e-01 2.72744477e-01
1.21111117e-01 -1.21266055e+00 -9.73548949e-01 4.83799309e-01
-1.35410294e-01 -1.06935310e+00 -4.14335161e-01 6.16128802e-01
-1.07473111e+00 -1.08016479e+00 7.46759549e-02 -8.21313679e-01
8.73787820e-01 -3.70887578e-01 9.86119330e-01 1.57300889e-01
-5.34607828e-01 -1.07330576e-01 -5.29553831e-01 -1.27492094e+00
-4.29638624e-01 8.20915103e-02 -2.76987553e-01 -4.33239579e-01
8.89324784e-01 -2.60137200e-01 -4.28685695e-01 -1.26315430e-01
-1.04910326e+00 -6.99389949e-02 9.00633693e-01 4.34155107e-01
6.62462592e-01 3.40802133e-01 7.13021517e-01 -1.34820557e+00
9.47687685e-01 -5.27353942e-01 -2.00978100e-01 3.36688042e-01
-1.11107755e+00 6.37404397e-02 4.62397724e-01 6.15822449e-02
-9.29807127e-01 1.82363495e-01 -1.87461480e-01 6.09754741e-01
-4.07853603e-01 8.85805786e-01 -2.24237919e-01 4.84619528e-01
5.84391654e-01 5.59298918e-02 2.44036227e-01 -4.08030480e-01
-2.28883654e-01 1.08172262e+00 -2.91096196e-02 -3.00333083e-01
4.06126946e-01 2.80924529e-01 -1.46353571e-02 -5.06033480e-01
-4.63978797e-01 -6.83191001e-01 -6.94520712e-01 -3.95028479e-02
1.11965775e+00 -6.14891887e-01 -8.56411457e-01 -4.82033081e-02
-8.00509751e-01 3.40937018e-01 -2.58526832e-01 6.20673358e-01
-6.94357678e-02 2.35372335e-01 -5.20052969e-01 -1.22329760e+00
-7.06463635e-01 -7.82029927e-01 6.67067945e-01 2.33464047e-01
-6.93201780e-01 -7.82969534e-01 -3.05022657e-01 4.17596757e-01
3.05300355e-01 6.45830035e-01 1.51480091e+00 -1.08665299e+00
-7.79549256e-02 -4.28218246e-01 -1.95775583e-01 -4.35016394e-01
4.92923856e-01 1.29386410e-01 -3.61041754e-01 2.68566012e-01
3.43368918e-01 -1.37921134e-02 1.27778783e-01 3.25220227e-01
8.24832678e-01 -5.90994716e-01 -6.27810121e-01 6.79558516e-02
1.51855111e+00 1.12601376e+00 8.24240327e-01 4.99806702e-01
2.15521187e-01 9.36203897e-01 8.03025961e-01 5.67777574e-01
3.39172959e-01 2.12522507e-01 -6.33659363e-02 6.94593191e-02
1.71458408e-01 -9.36122239e-02 -2.40620524e-02 8.24567914e-01
-1.28387973e-01 -5.92838228e-02 -1.01355946e+00 4.86624926e-01
-1.46858835e+00 -7.67139852e-01 -2.13433743e-01 2.17731977e+00
1.19717574e+00 5.83836555e-01 4.24874365e-01 5.61086953e-01
5.72046041e-01 -5.71508408e-01 -1.43490702e-01 -9.42724526e-01
1.30004555e-01 3.35020930e-01 5.19597530e-01 9.53253955e-02
-7.22671688e-01 5.00088453e-01 6.31493235e+00 5.73443711e-01
-7.12617993e-01 -1.71284780e-01 6.90294802e-01 2.58311272e-01
-2.58092731e-01 -9.14532095e-02 -9.42998588e-01 5.13411880e-01
9.67756748e-01 -3.45403224e-01 -1.04193270e-01 6.17341161e-01
3.67956251e-01 -5.58655560e-01 -8.80522728e-01 7.92884886e-01
7.28228837e-02 -1.20590770e+00 2.82671899e-01 1.83851853e-01
2.34337538e-01 -7.92314231e-01 -3.76556635e-01 1.25927161e-02
-2.30706751e-01 -1.16697478e+00 5.71700692e-01 5.01244247e-01
6.37579143e-01 -6.54887080e-01 1.26548243e+00 2.37406954e-01
-9.37930882e-01 4.38249931e-02 -3.59744281e-02 -6.70742542e-02
-1.71005931e-02 6.63188159e-01 -9.64019954e-01 8.69079828e-01
8.95905912e-01 3.11941236e-01 -5.09070635e-01 8.96200418e-01
3.95541191e-01 3.47127616e-01 -3.13662529e-01 -4.26216692e-01
-1.39708996e-01 -3.34695041e-01 1.40187174e-01 1.34399176e+00
2.99968719e-01 2.79503703e-01 -1.47744834e-01 4.07327294e-01
1.76489726e-01 9.43578720e-01 -5.82649708e-01 -3.94381881e-01
5.67189157e-01 1.04267097e+00 -1.68914950e+00 -3.05493832e-01
-5.48544466e-01 1.49445772e-01 -4.15157050e-01 -1.69729397e-01
-4.44533855e-01 -8.38516533e-01 3.45408209e-02 4.67427284e-01
1.81337342e-01 1.57225072e-01 -1.07656634e+00 -5.57068944e-01
2.17283696e-01 -1.12534153e+00 1.05679810e+00 -3.59751225e-01
-9.38225448e-01 6.61094427e-01 2.16986433e-01 -1.13084602e+00
-1.18459940e-01 -5.09283483e-01 -2.20665075e-02 8.84546459e-01
-7.22571671e-01 -6.59075201e-01 2.16288865e-02 3.64846945e-01
4.68838632e-01 -1.68318510e-01 9.98183250e-01 6.17785990e-01
-4.00373101e-01 3.88939440e-01 -3.99306208e-01 1.50829822e-01
7.66804636e-01 -1.25167298e+00 -3.46280843e-01 5.88610530e-01
-4.58800524e-01 9.39496934e-01 7.35042214e-01 -1.02829647e+00
-1.41996205e+00 -5.19477069e-01 1.91634464e+00 -5.87735593e-01
4.32784021e-01 -4.46777493e-01 -6.94286525e-01 3.70787054e-01
2.79949844e-01 -7.27628350e-01 1.14216518e+00 -6.25210628e-02
9.61846188e-02 -1.79717600e-01 -1.58793378e+00 4.70991373e-01
7.64311552e-01 -1.09229550e-01 -7.36611545e-01 4.29361641e-01
3.33949864e-01 -5.83281100e-01 -1.11995840e+00 1.93038464e-01
6.40417099e-01 -5.19825876e-01 7.08772957e-01 -7.15342462e-01
4.35782582e-01 -2.40003496e-01 1.92980707e-01 -5.49327135e-01
5.77936172e-02 -6.28925264e-01 1.82324588e-01 1.34970820e+00
1.13425791e+00 -5.29812813e-01 5.77380240e-01 1.00609791e+00
-8.29562321e-02 -1.01065052e+00 -5.61985910e-01 -1.06282055e-01
-2.55929381e-01 -1.99696496e-01 4.86923873e-01 9.64060545e-01
6.23546660e-01 5.39375067e-01 5.45214154e-02 -3.05576622e-01
3.15495282e-01 -4.57436182e-02 1.41171545e-01 -1.42256880e+00
2.19523937e-01 -5.23448825e-01 -2.58435905e-01 1.73160285e-01
-6.91460609e-01 -7.55044937e-01 -4.30461913e-01 -2.27608323e+00
5.18352091e-01 -2.26754591e-01 3.20405364e-02 4.09921557e-01
-3.59834433e-02 -4.18336481e-01 -2.06952617e-01 1.48083121e-01
-3.20891231e-01 -4.91929978e-01 9.89281595e-01 2.70587921e-01
-4.17521477e-01 -2.15687171e-01 -1.33116090e+00 5.25731444e-01
8.07475150e-01 -1.02123022e+00 -6.16538823e-01 1.99475601e-01
8.36447358e-01 1.75197721e-01 -5.68496466e-01 -6.80839896e-01
5.02170444e-01 -2.14256704e-01 6.54536188e-01 -8.20629299e-01
-4.10089105e-01 -1.00784576e+00 3.83319050e-01 7.50973821e-01
-2.09816605e-01 6.20872140e-01 4.26420450e-01 1.95843756e-01
-3.11465532e-01 -5.27734041e-01 3.34917516e-01 -3.30516636e-01
-1.13228098e-01 -7.02778772e-02 -8.94378066e-01 1.69989437e-01
1.05777788e+00 -5.42529702e-01 -2.07614273e-01 2.23881323e-02
-1.01984823e+00 3.00150216e-01 1.91819176e-01 1.65216967e-01
4.99485821e-01 -7.39617229e-01 -4.74191785e-01 3.30455229e-02
8.16596672e-02 -5.39321154e-02 -1.33658834e-02 1.09505177e+00
-9.72897053e-01 7.47767210e-01 -3.66813660e-01 -1.42274657e-02
-1.50019455e+00 9.60531652e-01 -3.22060436e-01 -7.88566947e-01
-4.65908438e-01 4.30658638e-01 -3.58637959e-01 7.62758031e-02
4.76424783e-01 -6.57285929e-01 -9.44390416e-01 6.17962241e-01
5.09997427e-01 2.49675840e-01 4.88909662e-01 -3.24453592e-01
-7.96886504e-01 4.39923584e-01 -2.64127821e-01 -9.84933674e-02
1.28304422e+00 -1.05348945e-01 -3.85324776e-01 4.73713070e-01
7.43541241e-01 2.41164744e-01 7.29773268e-02 3.12439919e-01
6.29005015e-01 1.16735613e-02 -3.74739170e-01 -1.12517631e+00
-5.39657116e-01 3.44630390e-01 1.13139398e-01 5.08479297e-01
1.30984712e+00 -2.33552501e-01 5.19254088e-01 1.49677182e-02
1.52467847e-01 -1.60832202e+00 -5.17620027e-01 2.78633386e-01
6.81795299e-01 -1.00680244e+00 3.96463841e-01 -9.12997723e-01
-7.43891120e-01 1.25588155e+00 1.57619953e-01 5.22728264e-01
1.27785122e+00 9.99951661e-01 1.24487855e-01 -5.93666673e-01
-9.83616948e-01 -9.76854190e-02 1.01848245e-01 4.69690919e-01
9.42642212e-01 -2.23143131e-01 -1.42233241e+00 1.15811133e+00
-2.61681169e-01 3.20260555e-01 4.75694031e-01 1.39110720e+00
-3.30964148e-01 -1.33311474e+00 -5.37908554e-01 9.26491737e-01
-1.32170570e+00 -1.71453923e-01 -7.44659066e-01 1.04252481e+00
4.54604089e-01 1.18608427e+00 -2.25723952e-01 -2.29321674e-01
7.00799465e-01 3.27673346e-01 3.29798102e-01 -9.20596302e-01
-1.21536207e+00 2.27927312e-01 4.98159796e-01 -2.63873726e-01
-5.01433253e-01 -9.72831368e-01 -1.56276166e+00 -5.74611910e-02
1.03595071e-01 6.56585157e-01 8.79640996e-01 9.33084905e-01
3.52658421e-01 5.66648066e-01 3.37890722e-02 2.70901740e-01
-1.30554736e-01 -1.01044154e+00 -5.13352633e-01 3.49254429e-01
-1.84139535e-01 -4.49637622e-01 -3.58084947e-01 4.55242902e-01] | [8.537771224975586, 8.692288398742676] |
2594ace5-7ab0-4738-b2d2-de3466083f63 | recognizing-disguised-faces-in-the-wild | 1811.08837 | null | http://arxiv.org/abs/1811.08837v1 | http://arxiv.org/pdf/1811.08837v1.pdf | Recognizing Disguised Faces in the Wild | Research in face recognition has seen tremendous growth over the past couple
of decades. Beginning from algorithms capable of performing recognition in
constrained environments, the current face recognition systems achieve very
high accuracies on large-scale unconstrained face datasets. While upcoming
algorithms continue to achieve improved performance, a majority of the face
recognition systems are susceptible to failure under disguise variations, one
of the most challenging covariate of face recognition. Most of the existing
disguise datasets contain images with limited variations, often captured in
controlled settings. This does not simulate a real world scenario, where both
intentional and unintentional unconstrained disguises are encountered by a face
recognition system. In this paper, a novel Disguised Faces in the Wild (DFW)
dataset is proposed which contains over 11000 images of 1000 identities with
different types of disguise accessories. The dataset is collected from the
Internet, resulting in unconstrained face images similar to real world
settings. This is the first-of-a-kind dataset with the availability of
impersonator and genuine obfuscated face images for each subject. The proposed
dataset has been analyzed in terms of three levels of difficulty: (i) easy,
(ii) medium, and (iii) hard in order to showcase the challenging nature of the
problem. It is our view that the research community can greatly benefit from
the DFW dataset in terms of developing algorithms robust to such adversaries.
The proposed dataset was released as part of the First International Workshop
and Competition on Disguised Faces in the Wild at CVPR, 2018. This paper
presents the DFW dataset in detail, including the evaluation protocols,
baseline results, performance analysis of the submissions received as part of
the competition, and three levels of difficulties of the DFW challenge dataset. | ['Nalini Ratha', 'Mayank Vatsa', 'Maneet Singh', 'Richa Singh', 'Rama Chellappa'] | 2018-11-21 | null | null | null | null | ['disguised-face-verification'] | ['computer-vision'] | [ 3.96327376e-01 -7.68292025e-02 2.66606003e-01 -5.78429341e-01
-1.75964221e-01 -9.34789300e-01 9.83245850e-01 -7.13747084e-01
-3.24456662e-01 5.32012105e-01 -1.24322481e-01 -1.66542530e-01
-9.39587206e-02 -5.23617983e-01 -6.71704352e-01 -8.38381827e-01
-2.69468963e-01 2.86109000e-01 -4.04184818e-01 -2.02025875e-01
1.95282623e-01 1.02312219e+00 -1.74025929e+00 1.18007787e-01
2.30980754e-01 1.04640555e+00 -3.80668670e-01 5.82751870e-01
4.30496633e-01 1.37369499e-01 -9.58721936e-01 -1.03529561e+00
8.30249608e-01 -3.21600020e-01 -6.69951081e-01 2.70889580e-01
1.23444617e+00 -5.70743918e-01 -3.82853627e-01 1.22669947e+00
8.36498976e-01 -1.90416157e-01 3.67155284e-01 -1.63417172e+00
-7.65833855e-01 1.02894500e-01 -4.59696978e-01 3.02024275e-01
5.06165862e-01 4.72858638e-01 3.30938250e-01 -8.07324827e-01
5.62727153e-01 1.60127223e+00 5.73185146e-01 8.93381655e-01
-1.14599562e+00 -1.11315203e+00 -2.14407638e-01 2.89828926e-01
-1.55706036e+00 -9.19353902e-01 6.13612115e-01 -4.02939945e-01
7.95192957e-01 4.35456008e-01 3.40273976e-01 1.70619619e+00
-2.48796880e-01 3.57843041e-01 1.40230346e+00 -4.39264774e-01
4.48141992e-02 3.15718740e-01 -8.10837150e-02 2.81941295e-01
3.15436661e-01 5.59435606e-01 -3.71370465e-01 -3.78948063e-01
4.94433522e-01 -1.43068805e-01 -4.98539567e-01 -5.50802536e-02
-8.09711814e-01 7.39863873e-01 2.63626873e-01 2.62844235e-01
-1.84552848e-01 -3.52708697e-01 3.29759955e-01 5.37787557e-01
1.59514830e-01 2.40116403e-01 -1.98856443e-01 -1.27715260e-01
-1.04203522e+00 2.81972289e-01 1.04764926e+00 7.56894052e-01
4.01721001e-01 1.90237537e-01 1.69206917e-01 8.88273835e-01
2.41119057e-01 6.16282642e-01 4.68869239e-01 -4.88822877e-01
2.70683289e-01 7.61570558e-02 1.79625452e-02 -1.43955839e+00
3.43706086e-02 4.94447388e-02 -7.61106491e-01 5.39989710e-01
6.12318337e-01 -1.90437526e-01 -1.09372663e+00 1.66443789e+00
2.95542628e-01 4.68680471e-01 9.65246409e-02 8.57853413e-01
7.52387464e-01 4.39595580e-01 -8.87629092e-02 -2.22862050e-01
1.47832572e+00 -5.20999134e-01 -8.18327725e-01 -5.78445643e-02
7.97735713e-03 -9.73283291e-01 6.62436247e-01 6.37146950e-01
-6.83513463e-01 -4.18424368e-01 -1.11672008e+00 4.87491727e-01
-6.75145924e-01 -2.46370122e-01 3.19831371e-01 1.61063862e+00
-1.29695165e+00 3.50694031e-01 -3.30408186e-01 -5.63191116e-01
9.16318834e-01 5.83179832e-01 -1.01220500e+00 -3.49972814e-01
-1.16527057e+00 1.00300717e+00 8.19639340e-02 4.94010478e-01
-1.30297017e+00 -6.15006328e-01 -7.36922979e-01 -3.50043714e-01
3.93300891e-01 1.81533858e-01 7.51810610e-01 -1.41754985e+00
-1.11864233e+00 1.39662910e+00 7.34201968e-02 -2.70316273e-01
7.67166615e-01 -1.97891369e-01 -1.02761006e+00 1.03698455e-01
-3.85642678e-01 2.55650848e-01 1.35524917e+00 -1.31224167e+00
5.22966795e-02 -7.46378422e-01 -1.20536275e-02 -3.04339826e-01
-4.07733440e-01 5.15358388e-01 4.36040713e-03 -8.01415384e-01
-5.15089273e-01 -9.04718816e-01 3.93730283e-01 9.74234119e-02
-3.93230915e-01 6.22386970e-02 1.48531914e+00 -6.30084515e-01
7.38264680e-01 -2.29147959e+00 -1.39906719e-01 1.37334332e-01
-5.06485589e-02 1.05848503e+00 -2.47706860e-01 4.68807071e-01
-4.62169528e-01 4.70190495e-01 -1.32782951e-01 -2.97483355e-01
-3.33004221e-02 3.63616258e-01 -5.21789372e-01 7.56748557e-01
3.45212668e-01 6.26637101e-01 -5.99934697e-01 -1.43090989e-02
9.21913683e-02 6.76328182e-01 -2.51348317e-01 4.98294443e-01
2.80353367e-01 3.15906376e-01 -1.78295299e-02 8.97135139e-01
1.35592902e+00 6.21036887e-01 1.02109678e-01 -3.26156497e-01
2.01496676e-01 -3.91969830e-01 -1.14066291e+00 9.83304858e-01
-3.69169377e-02 6.19228959e-01 5.17000854e-01 -1.00997257e+00
7.27466583e-01 6.26412570e-01 -1.74327858e-03 -2.64353782e-01
1.79121494e-01 2.86404818e-01 1.45499483e-01 -6.73848450e-01
1.53964102e-01 -2.45918125e-01 3.07526886e-01 4.07362133e-01
4.13003951e-01 -2.65761893e-02 -8.08360875e-02 7.65503966e-04
9.73324120e-01 -8.13201964e-02 2.63094127e-01 -3.34330946e-01
6.52794182e-01 -6.13941848e-01 2.76587963e-01 5.97483397e-01
-8.24321985e-01 6.70671880e-01 3.72707069e-01 -5.72066128e-01
-6.63226724e-01 -1.05406034e+00 -4.59959686e-01 5.51942289e-01
-1.38474852e-01 -7.08483830e-02 -1.00497878e+00 -9.32148933e-01
1.55301809e-01 3.09747100e-01 -8.22350085e-01 -1.41212270e-01
-4.62976068e-01 -6.78902149e-01 1.26378775e+00 7.78993890e-02
8.55655074e-01 -1.17518067e+00 -2.83595145e-01 -3.97911698e-01
1.87780708e-01 -1.20306051e+00 -4.32332516e-01 -3.92568707e-01
-2.16181099e-01 -1.33412898e+00 -7.00701833e-01 -6.18387103e-01
7.33640194e-01 2.52582639e-01 8.93319786e-01 2.34425306e-01
-7.74214745e-01 5.12221158e-01 -3.34653139e-01 -5.41003585e-01
-3.60539287e-01 -6.52041674e-01 4.37802911e-01 5.75137079e-01
7.40472555e-01 -4.67457503e-01 -4.78683442e-01 7.49209940e-01
-1.17091727e+00 -8.18067968e-01 2.45611325e-01 8.43268752e-01
-9.75552201e-02 1.93612322e-01 4.83145446e-01 -8.22335899e-01
4.03057247e-01 -6.79223478e-01 -3.66050243e-01 2.17373207e-01
-2.01828510e-01 -5.58001578e-01 5.57897270e-01 -7.25632787e-01
-9.20938432e-01 -1.62803397e-01 -2.95799941e-01 -4.52648163e-01
-8.51863682e-01 -1.83460966e-01 -7.21441269e-01 -6.03050053e-01
6.01112962e-01 1.81956932e-01 2.11740106e-01 -3.99705082e-01
1.19753882e-01 1.02592862e+00 5.02117872e-01 -5.33533514e-01
1.18927765e+00 4.45331395e-01 -1.61016420e-01 -1.28347707e+00
-3.85439277e-01 -5.04829139e-02 -6.40442669e-01 -5.62082767e-01
5.11333346e-01 -7.20216155e-01 -6.81959987e-01 1.32757950e+00
-8.35235834e-01 -8.22076574e-03 1.39719054e-01 1.79309577e-01
-1.03926331e-01 5.46838343e-01 -2.94780433e-01 -9.26240444e-01
-7.95115381e-02 -1.25004244e+00 7.64523447e-01 3.03168148e-01
-3.97856683e-02 -8.99135053e-01 -2.32511818e-01 6.10124111e-01
7.73277640e-01 7.11697936e-01 4.37957078e-01 -8.82787466e-01
-8.13203901e-02 -5.03189802e-01 -5.77576272e-02 8.00297558e-01
4.37037110e-01 1.39431357e-01 -1.51483512e+00 -6.80690169e-01
3.29540789e-01 -8.49425077e-01 4.97459829e-01 -2.25456223e-01
9.62021112e-01 -6.49722338e-01 3.64899673e-02 6.05453372e-01
1.41454387e+00 1.68137431e-01 1.01665139e+00 3.09255309e-02
4.57712889e-01 8.17301333e-01 3.29122812e-01 2.88316756e-01
-1.80705130e-01 7.83231020e-01 8.58443677e-01 1.76098272e-01
-9.74889174e-02 1.44859813e-02 5.57934225e-01 4.44145966e-03
-2.22507149e-01 -3.45424145e-01 -8.93852830e-01 4.65390772e-01
-1.13169503e+00 -1.24371922e+00 1.51960745e-01 2.37746406e+00
6.21828139e-01 -2.84485519e-01 2.92283863e-01 2.15777740e-01
9.67972457e-01 2.96894908e-01 -2.99377203e-01 -7.19754457e-01
-1.82859063e-01 4.28161293e-01 1.34127438e-01 2.78080285e-01
-1.22603357e+00 8.06991637e-01 6.13147354e+00 7.03983665e-01
-1.29521024e+00 4.84136418e-02 6.31576180e-01 -1.87886402e-01
3.29016775e-01 -3.70750964e-01 -8.97317410e-01 7.43053854e-01
1.03693807e+00 8.82923678e-02 6.72879517e-01 7.65106857e-01
-1.25720084e-01 1.84729010e-01 -9.79636550e-01 1.12844431e+00
7.71904767e-01 -6.64184332e-01 1.09469809e-01 3.94284755e-01
4.32980835e-01 -2.25441366e-01 4.54122216e-01 9.89130810e-02
1.16256118e-01 -1.66072762e+00 6.45065248e-01 1.29265487e-01
9.82024193e-01 -7.99616277e-01 7.85976350e-01 1.93864554e-01
-7.44826138e-01 -1.25183612e-01 -4.39540178e-01 -2.06377194e-03
-9.91583988e-02 2.97314465e-01 -5.41517317e-01 4.02648032e-01
7.70893157e-01 4.90210235e-01 -7.16531038e-01 7.32638240e-01
1.49638178e-02 8.16855371e-01 -5.29979348e-01 2.87572533e-01
4.19135951e-02 -3.55923586e-02 6.57298088e-01 1.19533896e+00
2.84659535e-01 1.27269939e-01 -2.68340558e-01 6.22736156e-01
-3.22299391e-01 -1.61726952e-01 -1.08036470e+00 -3.01574677e-01
5.60727656e-01 1.37435079e+00 -3.37257743e-01 9.96791795e-02
-2.80720025e-01 1.09561396e+00 1.44276828e-01 3.38516265e-01
-7.70119190e-01 -2.28372097e-01 1.09904504e+00 6.66659698e-02
3.10690373e-01 -1.23292938e-01 2.62358457e-01 -1.12240851e+00
2.21418902e-01 -1.46439993e+00 4.50573564e-01 -3.71860594e-01
-1.50011933e+00 8.43458652e-01 -6.07813671e-02 -8.61814201e-01
-6.88103214e-02 -9.04484510e-01 -6.11779034e-01 9.97967660e-01
-1.31661308e+00 -1.39455950e+00 -4.47306991e-01 7.75814772e-01
4.03064042e-01 -6.59661293e-01 1.12351537e+00 4.20622170e-01
-6.50080144e-01 1.13270366e+00 -1.37640640e-01 3.36892813e-01
9.36448395e-01 -7.46883452e-01 3.39462280e-01 9.44274306e-01
-1.23625789e-02 8.65970314e-01 5.73882222e-01 -3.64801407e-01
-1.61014926e+00 -1.00034761e+00 4.64764595e-01 -5.47003508e-01
4.93490964e-01 -6.14543140e-01 -8.83244336e-01 7.20681310e-01
3.18756610e-01 3.60456109e-01 8.81698847e-01 -3.22181880e-01
-8.67400169e-01 -4.02561761e-02 -1.95568025e+00 3.19973141e-01
9.43826973e-01 -5.76013625e-01 -5.81926942e-01 2.61484534e-01
-7.41589293e-02 -9.24786925e-02 -7.90112317e-01 3.30737591e-01
8.45646858e-01 -1.12023854e+00 1.07514203e+00 -6.98478222e-01
2.83696115e-01 -1.14311248e-01 -3.39302510e-01 -1.17149389e+00
2.72781327e-02 -8.96708429e-01 -1.65368065e-01 1.71626294e+00
-1.13447897e-01 -9.32359219e-01 5.52098930e-01 7.01452255e-01
5.83590865e-01 -3.94428641e-01 -1.15963280e+00 -9.77854729e-01
1.45290419e-01 -2.77508229e-01 6.76832616e-01 1.20917654e+00
-3.87737334e-01 -3.43321443e-01 -6.87022090e-01 3.38557839e-01
9.25721407e-01 -4.30115402e-01 8.92673552e-01 -9.06919181e-01
-1.17877565e-01 -2.34737918e-01 -1.10931778e+00 -3.66197973e-01
2.61784494e-01 -7.32822120e-01 -2.54078060e-01 -4.80473846e-01
8.59582350e-02 -1.28640473e-01 1.32616311e-01 6.21247947e-01
-1.04905948e-01 9.25076187e-01 3.84833187e-01 1.17461219e-01
1.02464348e-01 8.25804174e-02 7.91278064e-01 -1.90084800e-01
5.31953931e-01 -8.93047079e-02 -6.54620588e-01 4.70387220e-01
8.59072447e-01 -2.14844599e-01 -3.22909832e-01 -1.98613137e-01
-4.15573508e-01 -5.89720786e-01 6.84492409e-01 -8.99302244e-01
-1.60215437e-01 -5.04478179e-02 4.91206229e-01 3.73566039e-02
4.62208778e-01 -1.09505904e+00 3.56806666e-01 3.33505780e-01
8.65984932e-02 -2.47836888e-01 4.86485660e-01 3.52980703e-01
-1.60783321e-01 -1.93171978e-01 1.24760044e+00 -1.43225640e-01
-6.41545653e-01 3.66678119e-01 -1.72915846e-01 1.63532533e-02
1.38295531e+00 -4.81941700e-01 -5.28285503e-01 -3.13597977e-01
-6.00023210e-01 -1.25942603e-01 5.50022840e-01 6.91369772e-01
6.43844366e-01 -1.19020116e+00 -9.25271273e-01 7.78804004e-01
-2.86450908e-02 -6.03312552e-01 2.55981922e-01 4.39190686e-01
-5.73208570e-01 1.34664908e-01 -6.37038231e-01 -2.80410051e-01
-1.78126943e+00 6.67293429e-01 4.38543916e-01 1.83506474e-01
-2.49190778e-01 1.00676000e+00 5.21753281e-02 -3.48884761e-01
1.61073983e-01 7.03228951e-01 -1.65965021e-01 9.90521163e-02
1.11961055e+00 3.02354604e-01 1.86937347e-01 -1.33359325e+00
-4.84177113e-01 3.83255273e-01 -3.63645226e-01 1.91616371e-01
1.19113064e+00 6.79656938e-02 -3.58581632e-01 -2.08429098e-01
1.47489965e+00 -1.00378841e-01 -9.81166959e-01 8.83332163e-04
-2.52645463e-01 -1.18626142e+00 -2.00761810e-01 -1.04677427e+00
-1.26565921e+00 8.67549956e-01 9.36670363e-01 3.16466779e-01
1.14390802e+00 -3.38061512e-01 6.71083152e-01 6.34359121e-02
5.61907530e-01 -4.98190492e-01 -3.76458652e-02 2.75788546e-01
1.24568033e+00 -1.19385707e+00 -1.02019235e-01 -3.61317873e-01
-4.92933571e-01 1.11058640e+00 6.51927650e-01 -9.52456985e-03
8.83168817e-01 4.76934105e-01 3.25459242e-01 -1.29161060e-01
-3.00528616e-01 2.56611437e-01 1.09464288e-01 1.13801754e+00
1.16672352e-01 -7.57104382e-02 4.99073155e-02 3.38170409e-01
-4.23434049e-01 -1.62511334e-01 4.79459286e-01 1.03200889e+00
4.01667729e-02 -1.10993552e+00 -7.84665227e-01 3.65200877e-01
-8.76081645e-01 1.95567876e-01 -8.59482110e-01 7.69828498e-01
3.21818560e-01 1.15822339e+00 -1.92828402e-01 -4.13685143e-01
3.30133229e-01 2.04945624e-01 5.99544942e-01 -3.03305060e-01
-7.71971464e-01 -6.21173263e-01 -2.89157871e-02 -6.25489354e-01
-3.25822026e-01 -5.63408554e-01 -3.85069847e-01 -7.88974464e-01
-2.70314932e-01 -1.36924356e-01 8.49593222e-01 8.03749382e-01
2.58832246e-01 -1.63792342e-01 7.87029684e-01 -1.34861851e+00
-7.47582555e-01 -1.04049957e+00 -8.34392488e-01 8.04912984e-01
5.08095622e-01 -7.24092305e-01 -7.86060750e-01 2.10581332e-01] | [12.984569549560547, 1.0391535758972168] |
e1119827-0c8c-4494-a4fc-8fbdd12e1e6f | towards-accurate-ground-plane-normal | 2212.04224 | null | https://arxiv.org/abs/2212.04224v1 | https://arxiv.org/pdf/2212.04224v1.pdf | Towards Accurate Ground Plane Normal Estimation from Ego-Motion | In this paper, we introduce a novel approach for ground plane normal estimation of wheeled vehicles. In practice, the ground plane is dynamically changed due to braking and unstable road surface. As a result, the vehicle pose, especially the pitch angle, is oscillating from subtle to obvious. Thus, estimating ground plane normal is meaningful since it can be encoded to improve the robustness of various autonomous driving tasks (e.g., 3D object detection, road surface reconstruction, and trajectory planning). Our proposed method only uses odometry as input and estimates accurate ground plane normal vectors in real time. Particularly, it fully utilizes the underlying connection between the ego pose odometry (ego-motion) and its nearby ground plane. Built on that, an Invariant Extended Kalman Filter (IEKF) is designed to estimate the normal vector in the sensor's coordinate. Thus, our proposed method is simple yet efficient and supports both camera- and inertial-based odometry algorithms. Its usability and the marked improvement of robustness are validated through multiple experiments on public datasets. For instance, we achieve state-of-the-art accuracy on KITTI dataset with the estimated vector error of 0.39{\deg}. Our code is available at github.com/manymuch/ground_normal_filter. | ['Cong Yang', 'Tao Chen', 'Qian Zhang', 'Wei Sui', 'Jiaxin Zhang'] | 2022-12-08 | null | null | null | null | ['trajectory-planning'] | ['robots'] | [-1.64054886e-01 -2.61582229e-02 -3.05881083e-01 -3.26931983e-01
-4.61550891e-01 -4.16573465e-01 4.88363624e-01 -1.57770678e-01
-3.15163344e-01 4.48974103e-01 -3.44915748e-01 -3.92469645e-01
8.92523453e-02 -1.00487220e+00 -9.00854766e-01 -7.77364671e-01
2.00652376e-01 3.00966620e-01 5.17250180e-01 -5.68839312e-01
5.22302866e-01 5.93010068e-01 -1.60714447e+00 -8.23814332e-01
1.06627655e+00 8.65488827e-01 3.16542238e-01 4.85143602e-01
2.31872872e-01 9.86955166e-02 -8.89778212e-02 -2.53058404e-01
1.54807016e-01 3.31445783e-01 -8.11002403e-02 6.43600225e-02
3.43618035e-01 -5.11304915e-01 -5.99803805e-01 1.15459800e+00
3.52116883e-01 1.65572777e-01 4.61340636e-01 -1.29654467e+00
2.26359993e-01 -3.67269963e-01 -5.04147947e-01 4.01235223e-02
4.03431743e-01 4.09631848e-01 4.36262637e-01 -1.05307961e+00
5.78458965e-01 9.02143300e-01 6.91202343e-01 7.86920488e-02
-6.75626993e-01 -8.65145504e-01 2.78631896e-02 5.57506502e-01
-1.81901956e+00 -6.37458384e-01 7.58787811e-01 -5.29623926e-01
5.99728584e-01 7.00956881e-02 6.10398471e-01 7.77934074e-01
6.51548624e-01 4.62168932e-01 5.97904682e-01 -9.80441552e-03
3.94898914e-02 1.53231129e-01 1.75536707e-01 8.13310146e-01
9.19834912e-01 1.01941437e-01 -2.45175049e-01 2.00712904e-01
5.54211676e-01 -1.20077366e-02 -2.12825745e-01 -7.00785756e-01
-1.15923214e+00 5.08277774e-01 2.06448227e-01 -3.99724573e-01
-3.96221727e-01 2.46480643e-03 2.00316742e-01 -1.10448681e-01
2.65373796e-01 -2.07767949e-01 -9.82463583e-02 -4.31842864e-01
-3.32085580e-01 6.28031552e-01 5.48782468e-01 1.28063416e+00
1.21890640e+00 8.57711434e-02 4.72225010e-01 4.69450235e-01
5.82681000e-01 1.21261716e+00 1.85152635e-01 -7.41032660e-01
8.42415929e-01 5.19694209e-01 4.01833266e-01 -1.51539040e+00
-7.25160122e-01 -2.88123995e-01 -6.98696613e-01 6.01298027e-02
3.17231715e-01 -3.63250256e-01 -6.82529330e-01 1.38695669e+00
8.66748810e-01 5.47861040e-01 1.64817318e-01 8.66024315e-01
6.56531274e-01 3.75896633e-01 -3.58607948e-01 2.79875714e-02
1.35639346e+00 -5.97522020e-01 -1.07632995e+00 -7.42466450e-01
7.79518545e-01 -8.00517917e-01 5.47622323e-01 1.99204534e-01
-5.44530451e-01 -5.52326202e-01 -1.48740435e+00 7.11663291e-02
-2.79513687e-01 3.95362765e-01 3.92463863e-01 5.37049949e-01
-6.51236176e-01 3.16353530e-01 -1.05241072e+00 -5.57500541e-01
-2.07812786e-01 1.95073739e-01 -3.59526753e-01 -1.23948038e-01
-1.37574196e+00 1.20426500e+00 1.78562671e-01 4.28006202e-01
-4.22103703e-01 -1.75166130e-01 -1.08273947e+00 -5.25655150e-01
5.76179087e-01 -5.06464779e-01 1.13257790e+00 8.83031562e-02
-1.73101735e+00 5.24650633e-01 -5.52071273e-01 -2.81119674e-01
5.59630334e-01 -2.81689256e-01 -7.74439633e-01 -1.30756527e-01
2.07526356e-01 -6.09035790e-02 7.40068734e-01 -1.01717269e+00
-8.16934288e-01 -6.78251326e-01 -3.32630903e-01 4.12208200e-01
2.02218533e-01 -7.07774162e-01 -7.17708945e-01 1.30189601e-02
8.62187922e-01 -1.38222110e+00 -2.20872641e-01 -1.39866963e-01
-3.28589976e-01 7.25164684e-03 7.78351486e-01 -4.73246783e-01
1.19096673e+00 -2.23891068e+00 -4.28084344e-01 2.94213504e-01
5.78920767e-02 1.41095534e-01 4.29792374e-01 3.63431305e-01
3.97276491e-01 -3.81804943e-01 1.24142386e-01 -2.94555482e-02
-6.82249367e-02 1.69979006e-01 -1.30570650e-01 1.04495060e+00
-2.89398488e-02 6.22147739e-01 -9.09813046e-01 -1.89507976e-01
4.78001595e-01 4.78289038e-01 -2.41439417e-01 -9.72204506e-02
4.85976249e-01 2.35849112e-01 -7.73632050e-01 4.63995665e-01
1.28399563e+00 3.09409171e-01 -5.13410419e-02 -2.78314710e-01
-5.76816678e-01 5.15331864e-01 -1.68724918e+00 1.36635542e+00
-3.66900533e-01 7.09032476e-01 -4.86812741e-02 -4.70791608e-01
1.20422268e+00 7.84878284e-02 1.88816100e-01 -6.91060901e-01
2.80603468e-01 4.01462018e-01 -2.42461696e-01 -6.48237050e-01
8.93837035e-01 4.04602557e-01 -6.69560432e-02 -1.09545574e-01
-6.06785774e-01 -4.25841808e-01 -8.75542834e-02 -9.66739506e-02
8.44239235e-01 2.59712845e-01 4.35473561e-01 -2.01658458e-01
7.37659693e-01 1.99955389e-01 7.88282335e-01 3.14721286e-01
-3.72028798e-01 1.79389432e-01 1.24840379e-01 -1.97720319e-01
-9.55128670e-01 -9.48391318e-01 -2.77356803e-01 3.05101156e-01
1.01004303e+00 -2.78210670e-01 -5.67576706e-01 -4.93690148e-02
3.97214711e-01 6.45785570e-01 -3.06522876e-01 -3.71228993e-01
-6.57499969e-01 -4.93561029e-01 5.00322342e-01 3.17001641e-01
6.14319205e-01 -7.65463486e-02 -4.21935290e-01 2.44255587e-01
-2.74693429e-01 -1.33816838e+00 -2.94836968e-01 -4.54190344e-01
-9.24582183e-01 -1.14878225e+00 -1.04059264e-01 -3.67638648e-01
6.65916085e-01 1.05878508e+00 4.01087970e-01 -9.89327505e-02
3.40495199e-01 -3.39184962e-02 -4.92960624e-02 -5.57990074e-01
-3.05672269e-02 -5.28430566e-02 3.85437310e-01 1.30622655e-01
6.19694352e-01 -1.90700814e-01 -6.55967295e-01 8.84014130e-01
-2.49477178e-01 2.66967088e-01 3.27976853e-01 3.99945229e-01
4.81107354e-01 3.64605151e-02 1.97191969e-01 -4.41085339e-01
1.24944545e-01 -6.14292502e-01 -1.05014360e+00 -4.35185552e-01
-7.28718519e-01 -1.63660660e-01 1.87939763e-01 2.99199000e-02
-9.95841861e-01 1.42291218e-01 -3.13326627e-01 -8.11425820e-02
-1.47416785e-01 4.22815949e-01 -2.11442962e-01 -2.42467478e-01
5.04072368e-01 1.71187714e-01 3.91934782e-01 -2.87561595e-01
1.15389794e-01 8.65462303e-01 5.81738234e-01 -1.98506452e-02
1.19986355e+00 7.69231617e-01 3.90224963e-01 -1.14488006e+00
-3.42301577e-01 -6.42785311e-01 -5.48776388e-01 -4.34380203e-01
5.32019138e-01 -1.29716933e+00 -1.01765752e+00 8.99071634e-01
-1.05256069e+00 -4.58628200e-02 5.28641462e-01 8.57337475e-01
-4.55647677e-01 5.07918477e-01 -2.13978887e-01 -7.72099018e-01
-3.18540893e-02 -1.30714667e+00 1.00668228e+00 2.14979142e-01
7.71166012e-02 -7.56699383e-01 8.84869769e-02 2.14709342e-01
1.46719724e-01 2.66971231e-01 1.32984877e-01 1.18625648e-01
-8.07747483e-01 -7.18115389e-01 -1.62209690e-01 -1.82810843e-01
1.32775471e-01 2.28952035e-01 -8.62508774e-01 -3.61534178e-01
-8.16929117e-02 4.48649019e-01 4.58796024e-01 1.97267994e-01
4.66221660e-01 -7.73150846e-02 -5.73210299e-01 7.83054054e-01
1.25665581e+00 9.21475366e-02 7.25277424e-01 7.71770954e-01
8.14067125e-01 4.56848532e-01 1.35591197e+00 4.38937604e-01
9.77627039e-01 1.00250757e+00 6.82045996e-01 2.78847963e-01
2.62545913e-01 -2.47716352e-01 5.02646267e-01 8.23491514e-01
-3.40304226e-01 -1.74591746e-02 -9.21115041e-01 3.37844998e-01
-1.88171768e+00 -7.82542229e-01 -7.70130277e-01 2.56112313e+00
1.84075221e-01 4.03164148e-01 -2.78822720e-01 1.78237200e-01
8.73531818e-01 1.78412065e-01 -6.13348961e-01 -2.47317433e-01
2.35685110e-01 -6.78402424e-01 1.36059380e+00 7.31438696e-01
-9.52688992e-01 8.76759291e-01 4.70953560e+00 5.27153611e-01
-1.29644036e+00 -2.01165318e-01 -2.32343495e-01 3.72890234e-01
-1.20005712e-01 -1.03725895e-01 -1.44354582e+00 5.26998878e-01
8.57749403e-01 -4.43186939e-01 1.54178530e-01 9.81290638e-01
7.34983444e-01 -4.03159171e-01 -4.30877656e-01 9.01207924e-01
-1.13165468e-01 -1.10566986e+00 -4.23572540e-01 2.95574814e-01
5.11774838e-01 3.56505841e-01 -1.11577109e-01 2.27201685e-01
-1.59579515e-01 -2.03866497e-01 7.73023665e-01 3.72578382e-01
8.45522344e-01 -8.66451859e-01 9.04691637e-01 6.25689924e-01
-1.55119312e+00 2.01002181e-01 -2.87157208e-01 -3.41904610e-01
3.23616028e-01 5.91778815e-01 -1.15356469e+00 8.46733034e-01
4.15893257e-01 8.80080462e-01 -3.59675378e-01 1.00009978e+00
-4.44015086e-01 3.05546075e-01 -6.04258358e-01 2.12880895e-02
2.48732820e-01 -4.32701886e-01 9.16698992e-01 7.96151221e-01
6.23179197e-01 1.78292334e-01 4.19594944e-02 2.20199332e-01
4.99866545e-01 5.15854508e-02 -1.03655362e+00 5.00696838e-01
6.40982985e-01 1.18081284e+00 -3.27401876e-01 -2.59156883e-01
-3.62072617e-01 4.00413871e-01 -1.32856339e-01 4.22085166e-01
-9.83323634e-01 -6.42214179e-01 1.21623993e+00 4.07125652e-01
2.65592694e-01 -7.19975650e-01 -3.05362493e-01 -1.11909366e+00
2.00433478e-01 -3.13231260e-01 -2.37873405e-01 -6.06289864e-01
-4.02165622e-01 3.01506758e-01 -7.88080841e-02 -1.77328515e+00
-4.26536083e-01 -5.50905526e-01 -4.72770184e-01 8.68929386e-01
-1.62306261e+00 -6.50605142e-01 -8.56131673e-01 3.76637787e-01
2.79157758e-01 1.91710189e-01 3.81902367e-01 4.59485859e-01
-7.65915155e-01 4.07875419e-01 3.93992186e-01 -7.13625364e-03
6.86221480e-01 -7.18724191e-01 9.35463905e-01 1.03887689e+00
-5.95966816e-01 5.15297115e-01 1.08423567e+00 -7.95275390e-01
-2.02846384e+00 -1.16020608e+00 8.98667216e-01 -3.58169317e-01
7.22317755e-01 -2.94893444e-01 -6.79551005e-01 7.41267562e-01
-5.20187974e-01 1.29521936e-02 -5.65711968e-02 -1.98368013e-01
1.86047643e-01 -3.86631191e-01 -9.31497037e-01 7.46747553e-01
8.80052865e-01 -2.85423577e-01 -2.21843734e-01 1.46223847e-02
4.57884073e-01 -1.19096375e+00 -5.87622464e-01 6.94524705e-01
7.71059453e-01 -7.55120575e-01 8.29685330e-01 2.07864016e-01
-4.03072715e-01 -7.61339605e-01 -1.29728422e-01 -1.05621195e+00
-2.22335801e-01 -6.16678953e-01 -1.73899189e-01 7.52651930e-01
1.21580936e-01 -1.19013655e+00 8.99932683e-01 3.12667042e-01
-4.25454110e-01 -7.06500232e-01 -8.99748623e-01 -6.45305395e-01
-7.41050482e-01 -8.39084506e-01 7.40844369e-01 5.57637632e-01
-9.43128094e-02 1.90552533e-01 -3.12851280e-01 9.79198098e-01
8.08569968e-01 -1.75206274e-01 1.48777103e+00 -1.11683214e+00
4.12421852e-01 -5.27701974e-02 -8.98142993e-01 -1.55885386e+00
-8.95323828e-02 -4.95811820e-01 3.17418814e-01 -1.31429780e+00
-5.42921841e-01 -4.60450649e-01 7.99981579e-02 -7.15385601e-02
-1.17593355e-01 1.45299450e-01 -1.72082216e-01 1.01586625e-01
-2.79329389e-01 4.33001161e-01 1.30673289e+00 2.85487100e-02
-3.20356190e-01 5.37805140e-01 -2.14103609e-01 9.68389452e-01
9.47853208e-01 -2.83242077e-01 -3.23805660e-01 -4.02351797e-01
3.08385730e-01 9.77539122e-02 4.37554628e-01 -1.10136664e+00
3.64429563e-01 -1.89978719e-01 -5.32584898e-02 -1.10683644e+00
4.74286377e-01 -6.43997312e-01 1.43207684e-01 4.62061942e-01
5.95464647e-01 9.66286138e-02 3.29725415e-01 6.51622474e-01
-1.50003936e-02 -3.68173122e-02 5.10573149e-01 3.64664495e-01
-1.11003458e+00 5.11875093e-01 -4.99667019e-01 -2.95381606e-01
1.01166737e+00 -6.27516508e-01 -3.92264545e-01 -3.40405196e-01
-1.18828744e-01 4.32450682e-01 7.24765539e-01 4.56440508e-01
6.25808358e-01 -1.33959937e+00 -4.66951281e-01 5.55663228e-01
5.60859323e-01 2.56304026e-01 4.19981748e-01 1.17650843e+00
-7.57937253e-01 6.76507056e-01 7.57833850e-03 -1.03941643e+00
-1.02864349e+00 9.78665203e-02 1.73405930e-01 3.62811863e-01
-6.63626492e-01 2.62108415e-01 -1.99467726e-02 -4.05899197e-01
-6.70105219e-02 -4.76718843e-01 -2.70065248e-01 -5.80131374e-02
4.68436509e-01 7.32868850e-01 3.54352832e-01 -1.08680630e+00
-5.79598069e-01 1.02843380e+00 1.25446469e-01 3.82410698e-02
8.48097801e-01 -8.54714692e-01 2.83471197e-01 4.96207952e-01
1.09400058e+00 8.40363652e-02 -1.43786776e+00 -5.69687597e-02
-1.36332586e-01 -6.87951624e-01 1.71336323e-01 2.11400717e-01
-6.60055459e-01 6.26824796e-01 6.20790899e-01 -1.08491898e-01
5.56519032e-01 -5.12631297e-01 8.57590318e-01 7.22707391e-01
7.44681418e-01 -1.13873827e+00 -5.55244327e-01 7.95009434e-01
4.18003082e-01 -1.33181608e+00 2.55842388e-01 -6.69016540e-01
-4.67269421e-01 1.02461410e+00 6.24113739e-01 -1.09491222e-01
5.91411829e-01 2.51064133e-02 3.13116133e-01 -9.72227156e-02
-5.10160089e-01 -2.96788990e-01 9.37248021e-02 4.68446791e-01
-4.50929068e-02 1.83124870e-01 -2.35516801e-01 -6.28987234e-03
-2.65244633e-01 -1.68620497e-01 7.44799733e-01 9.04246807e-01
-8.59475255e-01 -5.87252736e-01 -6.32947087e-01 1.65869176e-01
8.69061351e-02 4.08267498e-01 5.03039420e-01 1.06322205e+00
7.52205104e-02 9.31164801e-01 2.05390185e-01 -6.32132769e-01
7.40931988e-01 -3.28132808e-01 1.36124596e-01 -2.96464533e-01
4.79080170e-01 -1.43073797e-01 2.59493619e-01 -8.20112526e-01
-7.91727472e-03 -7.98884511e-01 -1.46412086e+00 -5.64695418e-01
-5.15077651e-01 -2.88706110e-03 1.15953183e+00 9.34079647e-01
4.55020815e-01 2.02784017e-01 9.43643749e-01 -1.04615438e+00
-4.30131614e-01 -7.80840218e-01 -6.88445270e-01 6.74333423e-02
4.87132907e-01 -1.20353794e+00 -4.55680609e-01 -4.27167416e-01] | [7.523440837860107, -2.0063624382019043] |
01736970-5076-4626-aa35-37a796bca6f1 | a-comparative-study-on-application-of-class | 2206.09752 | null | https://arxiv.org/abs/2206.09752v1 | https://arxiv.org/pdf/2206.09752v1.pdf | A Comparative Study on Application of Class-Imbalance Learning for Severity Prediction of Adverse Events Following Immunization | In collaboration with the Liaoning CDC, China, we propose a prediction system to predict the subsequent hospitalization of children with adverse reactions based on data on adverse events following immunization. We extracted multiple features from the data, and selected "hospitalization or not" as the target for classification. Since the data are imbalanced, we used various class-imbalance learning methods for training and improved the RUSBoost algorithm. Experimental results show that the improved RUSBoost has the highest Area Under the ROC Curve on the target among these algorithms. Additionally, we compared these class-imbalance learning methods with some common machine learning algorithms. We combined the improved RUSBoost with dynamic web resource development techniques to build an evaluation system with information entry and vaccination response prediction capabilities for relevant medical practitioners. | ['Tong Jia', 'Zhengke Sun', 'Ning Chen'] | 2022-06-20 | null | null | null | null | ['severity-prediction'] | ['computer-vision'] | [-3.02640408e-01 -4.14021879e-01 -8.03344190e-01 -7.51034915e-01
-6.25402212e-01 -2.65549988e-01 -3.28597724e-01 9.42373991e-01
-1.99468791e-01 6.78523362e-01 2.60473073e-01 -6.22415543e-01
-3.75806451e-01 -1.13375890e+00 -5.69645643e-01 -4.52562124e-01
3.74261435e-04 7.40548790e-01 2.25109890e-01 -8.91285613e-02
1.89863354e-01 1.75348431e-01 -1.44429684e+00 1.09412813e+00
1.42788255e+00 8.96392524e-01 -4.03881341e-01 5.58841407e-01
-3.65669578e-01 9.07288790e-01 -7.22442806e-01 -2.72410750e-01
6.71479702e-02 -2.86048323e-01 -6.16128564e-01 -7.35949814e-01
-2.48698547e-01 -4.34007823e-01 2.89358705e-01 5.75727344e-01
8.70809853e-01 -3.83128703e-01 6.87815189e-01 -1.53218436e+00
-3.16254973e-01 4.64533210e-01 -4.54237461e-01 5.40140331e-01
7.44131446e-01 -2.52518088e-01 5.22196770e-01 -4.36190665e-01
4.63235497e-01 9.05465364e-01 9.90971565e-01 5.52824914e-01
-8.36311221e-01 -1.00352609e+00 -1.49833569e-02 4.30233061e-01
-9.14626062e-01 3.61900628e-01 3.81314903e-01 -7.05873907e-01
1.08990109e+00 4.85911369e-01 8.18059444e-01 5.25391638e-01
3.58555824e-01 8.82626891e-01 1.07297921e+00 -3.94266903e-01
5.89224324e-02 1.13810107e-01 5.91572046e-01 6.71172619e-01
3.73450845e-01 7.64908046e-02 -3.78771156e-01 -8.55210125e-01
-8.68969113e-02 6.06795728e-01 7.34941959e-02 1.36577189e-02
-6.50734067e-01 1.06696427e+00 2.65603572e-01 -3.08744721e-02
-5.79976857e-01 -8.46478522e-01 4.09011692e-01 6.80039048e-01
1.06599665e+00 1.54268295e-01 -1.26509476e+00 2.11902201e-01
-1.46493956e-01 2.14616030e-01 7.63888776e-01 4.44889516e-01
4.20373797e-01 -6.19090080e-01 -2.41646513e-01 1.02562284e+00
8.04357305e-02 7.35992432e-01 3.83609086e-01 -5.23629002e-02
5.79780400e-01 1.23309207e+00 1.39398808e-02 -6.75348163e-01
-8.00896823e-01 -1.59503520e-01 -7.20723152e-01 -3.34092706e-01
1.22120745e-01 -5.81321239e-01 -9.22684312e-01 1.28951883e+00
8.88170540e-01 -2.21833915e-01 1.20029241e-01 5.71561754e-01
1.11981702e+00 6.09286070e-01 4.54811394e-01 -6.02171540e-01
1.03402913e+00 -7.24202752e-01 -8.28389943e-01 6.52154028e-01
1.26396370e+00 -7.37474918e-01 5.74636638e-01 7.35742033e-01
-7.05179274e-01 -9.21858847e-02 -6.27268791e-01 7.49667883e-01
-6.40313625e-01 -2.07376808e-01 7.58143544e-01 7.44486451e-01
-6.24299705e-01 3.67825270e-01 -6.17850244e-01 -3.69063258e-01
2.32594952e-01 4.54366326e-01 -2.29061007e-01 -2.48904243e-01
-1.48641944e+00 9.81851578e-01 3.28417599e-01 -6.70216918e-01
-2.04326928e-01 -1.11441779e+00 -5.61659098e-01 -3.91169935e-01
-2.54539311e-01 -6.27410948e-01 9.10410523e-01 -8.09201896e-01
-9.43809032e-01 7.45498776e-01 1.55332014e-01 -1.19057000e-01
1.31892905e-01 -3.10143650e-01 -7.68951476e-01 4.77737673e-02
-6.06457852e-02 1.54177612e-02 4.98516634e-02 -4.08834100e-01
-8.80522907e-01 -8.19389045e-01 -1.52023867e-01 1.50750615e-02
-2.98460037e-01 5.67653656e-01 3.49523216e-01 -4.57814395e-01
2.84813970e-01 -4.89086002e-01 -3.12699050e-01 -6.66219831e-01
1.32635444e-01 -7.48633206e-01 5.32366455e-01 -9.13396537e-01
1.62063909e+00 -1.59027421e+00 -6.69913828e-01 4.24344540e-01
4.60784435e-02 5.34359813e-01 -2.62433112e-01 5.94342470e-01
-4.99009162e-01 -2.87447660e-03 5.79497993e-01 9.34048712e-01
-7.50203967e-01 -1.55016467e-01 3.34123224e-02 2.74510175e-01
2.72306085e-01 5.85618377e-01 -8.14813197e-01 -4.02272761e-01
1.13542542e-01 2.54400969e-01 -8.41546416e-01 9.41878498e-01
-2.32792646e-01 4.08579081e-01 -7.24180520e-01 7.29112506e-01
9.08701003e-01 -1.63744316e-01 3.10356736e-01 8.55459773e-04
-1.45490728e-02 4.92049724e-01 -6.80168569e-01 8.04358780e-01
7.83026218e-02 -2.87208557e-01 -5.64309657e-01 -1.05004311e+00
1.20199597e+00 3.76178592e-01 9.49330628e-01 -8.48790526e-01
-5.02127893e-02 4.44782935e-02 -5.49853481e-02 -1.21870565e+00
-6.77688241e-01 1.95809767e-01 2.54716814e-01 5.68776309e-01
-2.87673891e-01 4.45532709e-01 -4.79880199e-02 -3.72719765e-02
1.19801474e+00 -2.37746939e-01 4.48792607e-01 -2.49113947e-01
2.97610939e-01 4.24427986e-01 8.95681918e-01 5.18048704e-01
-1.37259707e-01 2.37664849e-01 4.73619729e-01 -1.28968632e+00
-5.48966825e-01 -8.62948716e-01 -3.95823628e-01 1.45869243e+00
-4.14218634e-01 -2.85284698e-01 -8.84771585e-01 -1.08750892e+00
-2.31745746e-02 2.41597384e-01 -5.36168754e-01 6.27667233e-02
-3.50474060e-01 -1.48452449e+00 2.99891621e-01 2.96112776e-01
1.08650096e-01 -7.65595436e-01 -3.56943071e-01 3.32789660e-01
-3.68199497e-01 -1.19368643e-01 -1.42710194e-01 6.26973957e-02
-7.79079556e-01 -1.82478905e+00 -3.67196232e-01 -8.94975424e-01
7.82855392e-01 1.71787098e-01 1.15510106e+00 4.18149620e-01
-4.22476918e-01 9.75135937e-02 -7.56224036e-01 -1.08375919e+00
-3.55445564e-01 2.83854622e-02 6.31361976e-02 -4.43137676e-01
1.04097891e+00 -4.42605168e-02 -1.00990629e+00 2.74388701e-01
-8.76745820e-01 -7.39120990e-02 1.38731375e-01 8.82443726e-01
2.81651020e-01 1.94655024e-02 1.12352800e+00 -1.10222805e+00
2.70018786e-01 -1.00051200e+00 -5.99333107e-01 6.13746881e-01
-1.01969039e+00 -2.52453864e-01 3.37052077e-01 -5.23420632e-01
-9.07467723e-01 -1.12632185e-01 -5.12402534e-01 3.58534217e-01
-1.72687382e-01 5.44286191e-01 2.01452617e-02 1.86203972e-01
5.81252694e-01 -3.46959263e-01 6.58043101e-02 -7.12833643e-01
-3.81621510e-01 1.19979298e+00 -5.40857792e-01 -2.30346978e-01
-1.44435942e-01 -7.48915896e-02 -3.95148128e-01 -1.85686439e-01
-1.05371356e+00 -8.32343936e-01 -1.78057238e-01 -7.66866207e-02
7.76252270e-01 -9.79784369e-01 -1.03290653e+00 8.56711745e-01
-1.27848887e+00 1.03755884e-01 4.21643347e-01 9.09626067e-01
-7.27591813e-02 -1.45890281e-01 -8.52538586e-01 -7.50165880e-01
-1.01017129e+00 -6.82926714e-01 3.15307826e-01 1.99051455e-01
-1.68805823e-01 -8.86639178e-01 8.84100378e-01 7.27155387e-01
3.83748084e-01 3.84796113e-01 1.45611703e+00 -1.41637862e+00
4.89492789e-02 -3.81604791e-01 -7.45454729e-02 1.11994877e-01
7.08990097e-02 3.81663740e-01 -6.06782675e-01 -4.27633703e-01
-3.31766829e-02 -3.49178761e-01 4.49472487e-01 5.28704941e-01
1.34032357e+00 -6.14591062e-01 -4.28563505e-01 3.72585088e-01
1.51622117e+00 8.71668637e-01 4.56970811e-01 2.35333234e-01
3.66882861e-01 1.05777466e+00 9.67415452e-01 4.45856065e-01
8.21924090e-01 4.02578294e-01 4.07837257e-02 -3.45086992e-01
5.51691711e-01 -2.77397096e-01 3.50451544e-02 1.17052531e+00
-1.29168034e-01 -8.39275960e-03 -1.44580925e+00 3.43282074e-01
-1.82162774e+00 -6.08039141e-01 -4.22373533e-01 2.22231054e+00
1.19942844e+00 -1.80908084e-01 3.17674488e-01 4.06067558e-02
7.69160390e-01 -3.92005652e-01 -3.86239350e-01 -8.44159782e-01
5.65710329e-02 5.48802853e-01 2.53711760e-01 2.04972345e-02
-1.06984699e+00 2.96709001e-01 7.44345427e+00 6.86144352e-01
-9.70026910e-01 2.79894829e-01 9.80539620e-01 1.96062922e-02
-8.63818601e-02 -4.07397002e-01 -7.08023846e-01 5.35887122e-01
1.04664707e+00 -1.10813174e-02 -4.46280912e-02 7.34252810e-01
8.01941603e-02 -3.35778557e-02 -8.11520875e-01 4.30655926e-01
6.92315251e-02 -1.11370909e+00 -4.01075855e-02 -5.33958375e-01
1.01466465e+00 2.14704663e-01 -2.89285451e-01 2.19732955e-01
5.65369010e-01 -8.91986132e-01 -3.96061927e-01 5.15048027e-01
5.73158145e-01 -8.09212327e-01 1.15282607e+00 6.19374275e-01
-7.28734970e-01 -1.09341100e-01 -2.92469710e-01 -3.10749412e-01
-6.10970855e-01 9.37074721e-01 -1.07311714e+00 5.20189941e-01
1.01570928e+00 3.87540787e-01 -3.67466271e-01 1.14975142e+00
1.66501686e-01 7.35916317e-01 -5.25268205e-02 -3.02333474e-01
-3.79720002e-01 9.65161771e-02 -2.06578851e-01 1.21142602e+00
1.24521419e-01 6.87608719e-01 2.22966030e-01 -1.82385325e-01
2.80015677e-01 9.17174578e-01 -6.24197423e-01 3.47565949e-01
2.61671841e-01 7.98500717e-01 -2.26867869e-01 -4.70099509e-01
-5.98581791e-01 -1.99904621e-01 2.74295360e-01 1.11718640e-01
-7.64586151e-01 -3.84366274e-01 4.51175302e-01 2.12972417e-01
-1.53359249e-01 7.73458302e-01 -4.63900715e-01 -9.72356617e-01
-3.80883187e-01 -1.04356325e+00 1.23231161e+00 -4.11153585e-01
-1.54222047e+00 3.93045783e-01 -5.98211959e-02 -1.09558725e+00
-3.53758559e-02 -3.74124736e-01 -6.90506577e-01 6.38794065e-01
-1.25696015e+00 -7.48001277e-01 -3.05270460e-02 5.34207344e-01
2.52191603e-01 -1.24258153e-01 1.04228187e+00 4.35560256e-01
-7.48149335e-01 7.01835334e-01 9.26742405e-02 2.09842429e-01
6.75852656e-01 -8.00828934e-01 -1.22767545e-01 1.73916981e-01
-7.69610643e-01 3.82053316e-01 2.41123497e-01 -1.01089597e+00
-1.07170558e+00 -1.15001869e+00 1.34735990e+00 -4.88268614e-01
2.65554100e-01 1.26350731e-01 -9.61389899e-01 2.38326833e-01
1.30932769e-02 -3.91345143e-01 1.45190334e+00 3.31514865e-01
-5.06871164e-01 -4.19402778e-01 -1.54613471e+00 -2.05774859e-01
5.69026113e-01 -1.61564071e-02 -7.09527969e-01 8.01875114e-01
9.50119615e-01 -2.56330103e-01 -1.27054071e+00 1.14794052e+00
8.18590760e-01 -9.12005544e-01 9.16981757e-01 -1.58498096e+00
5.11843204e-01 4.34766039e-02 2.50427634e-01 -1.24925983e+00
-2.77907550e-01 5.74942380e-02 1.53607056e-01 9.97256100e-01
5.36711216e-01 -8.96397769e-01 8.22853744e-01 6.97497487e-01
5.07672429e-01 -1.30261374e+00 -5.93383133e-01 -2.61390388e-01
-5.20418677e-03 -1.05281621e-01 1.21379173e+00 1.25364089e+00
2.44416073e-01 1.31049883e-02 -2.14028195e-01 1.02737240e-01
2.95090854e-01 3.81580353e-01 4.72254068e-01 -1.19680703e+00
-3.94781157e-02 2.37726018e-01 -2.77924985e-01 -1.54841319e-01
-2.97857076e-01 -6.61853909e-01 -3.17034334e-01 -1.49212790e+00
6.99086130e-01 -7.60596037e-01 -9.89812851e-01 6.19323254e-01
-5.91967583e-01 -1.42256558e-01 -4.31624025e-01 -9.18208212e-02
-6.28830135e-01 -7.87510872e-02 9.86786485e-01 -1.77513301e-01
-4.43492740e-01 7.29371250e-01 -4.38059866e-01 7.28397787e-01
9.42230642e-01 -7.84402847e-01 -2.26596788e-01 -3.46446425e-01
4.86304730e-01 4.61447090e-01 -4.17648822e-01 -4.55544502e-01
5.73445670e-02 -6.68869793e-01 6.84608161e-01 -1.13181305e+00
-6.15909219e-01 -6.36426866e-01 5.04222929e-01 1.23508143e+00
-5.13348222e-01 4.49063748e-01 -1.47989556e-01 9.75026749e-03
-1.32273734e-01 1.57025367e-01 4.79569018e-01 2.55028278e-01
-1.21866353e-02 4.68321562e-01 -2.06136510e-01 5.68601117e-02
1.15604687e+00 2.79621631e-01 -8.82731974e-01 2.38730296e-01
-4.58770812e-01 4.30250734e-01 1.71616763e-01 4.38977927e-01
6.53887749e-01 -1.06335306e+00 -9.34809625e-01 7.19563961e-01
3.10518146e-01 -3.31033230e-01 3.32401484e-01 8.18223715e-01
-8.82697463e-01 5.14247537e-01 -3.86907309e-01 -5.31194389e-01
-1.57521856e+00 7.50032127e-01 2.56279064e-03 -5.18832922e-01
1.42068833e-01 5.52503228e-01 5.05107408e-03 -8.16578329e-01
4.02421594e-01 -8.62200558e-02 -8.33522081e-01 2.23767608e-01
8.34280849e-01 7.74064362e-01 5.05847096e-01 -2.57980227e-02
-6.91096902e-01 3.63564551e-01 -2.25044101e-01 7.20255315e-01
1.58342552e+00 3.62087041e-01 -7.22486734e-01 1.15163691e-01
1.20038927e+00 -2.65306145e-01 -3.50661486e-01 -8.43608677e-02
1.79215491e-01 -4.04973477e-01 -5.54590404e-01 -1.18698502e+00
-1.05525064e+00 5.89947164e-01 1.04236674e+00 3.45906734e-01
1.60711133e+00 -2.81286716e-01 8.68648410e-01 2.22686201e-01
4.70090389e-01 -1.10722876e+00 -2.35368505e-01 3.25155407e-01
3.93622458e-01 -1.49238205e+00 2.33542025e-02 -5.04684687e-01
-4.40943837e-01 8.92269373e-01 8.21023941e-01 7.35843331e-02
1.12119663e+00 2.74044693e-01 5.96133590e-01 -5.41141182e-02
-1.53139150e+00 1.89407706e-01 2.75710851e-01 9.33381736e-01
5.74244022e-01 5.18159151e-01 -1.20279467e+00 7.94800818e-01
2.45132551e-01 3.90366644e-01 -6.42975420e-03 7.50250936e-01
-5.31825304e-01 -1.35101402e+00 -4.28201050e-01 1.14020145e+00
-9.75884855e-01 -1.22424833e-01 -9.60074961e-02 2.68780112e-01
4.94485825e-01 1.21183741e+00 -3.50111350e-02 -7.09131300e-01
5.01143873e-01 2.62902170e-01 1.05117477e-01 -5.18379331e-01
-1.21135652e+00 -1.30825147e-01 1.66804478e-01 -5.48146605e-01
-4.96034890e-01 -2.79842019e-01 -1.04865873e+00 -3.44869673e-01
-4.29261744e-01 5.50583005e-01 6.99967861e-01 7.87544250e-01
3.45409989e-01 1.11391366e-01 1.43092012e+00 4.77124602e-01
-3.19767892e-01 -7.17253149e-01 -2.16484517e-01 5.57152689e-01
1.61772951e-01 -2.70873070e-01 -1.50628716e-01 -3.98233145e-01] | [8.293390274047852, 5.376340389251709] |
e23bb3c5-f9b0-47c4-8913-9dd1aa586cb0 | physics-informed-computer-vision-a-review-and | 2305.18035 | null | https://arxiv.org/abs/2305.18035v2 | https://arxiv.org/pdf/2305.18035v2.pdf | Physics-Informed Computer Vision: A Review and Perspectives | Incorporation of physical information in machine learning frameworks are opening and transforming many application domains. Here the learning process is augmented through the induction of fundamental knowledge and governing physical laws. In this work we explore their utility for computer vision tasks in interpreting and understanding visual data. We present a systematic literature review of formulation and approaches to computer vision tasks guided by physical laws. We begin by decomposing the popular computer vision pipeline into a taxonomy of stages and investigate approaches to incorporate governing physical equations in each stage. Existing approaches in each task are analyzed with regard to what governing physical processes are modeled, formulated and how they are incorporated, i.e. modify data (observation bias), modify networks (inductive bias), and modify losses (learning bias). The taxonomy offers a unified view of the application of the physics-informed capability, highlighting where physics-informed learning has been conducted and where the gaps and opportunities are. Finally, we highlight open problems and challenges to inform future research. While still in its early days, the study of physics-informed computer vision has the promise to develop better computer vision models that can improve physical plausibility, accuracy, data efficiency and generalization in increasingly realistic applications. | ['George Karniadakis', 'Clinton Fookes', 'Kien Nguyen', 'Chayan Banerjee'] | 2023-05-29 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [ 4.58632767e-01 8.90999734e-02 -5.34683406e-01 -3.41964275e-01
-2.42665276e-01 -4.57762718e-01 1.08144379e+00 3.16321328e-02
-2.66867220e-01 4.88966465e-01 1.57221053e-02 -6.24485075e-01
-6.86383963e-01 -5.38749516e-01 -8.23802710e-01 -9.80227053e-01
1.11262709e-01 2.42323592e-01 3.15008342e-01 -2.69011455e-03
6.52520061e-01 8.95500362e-01 -1.52263761e+00 1.95238560e-01
8.73604059e-01 1.07070374e+00 2.53906757e-01 9.51320887e-01
-1.26502052e-01 9.79305685e-01 3.21760662e-02 -3.86171162e-01
1.96738601e-01 -2.52098441e-01 -7.51102209e-01 -2.69248247e-01
9.00734544e-01 -2.35744089e-01 -7.91869044e-01 9.27326322e-01
1.66039795e-01 3.71153742e-01 1.05082786e+00 -1.34925163e+00
-1.45535755e+00 1.80002272e-01 -5.86258650e-01 7.59716094e-01
-1.91054016e-01 6.39524519e-01 1.08047068e+00 -4.72117871e-01
3.04312855e-01 1.53058529e+00 9.48897183e-01 8.31846476e-01
-1.34970343e+00 -1.35005519e-01 4.62117255e-01 6.26344085e-01
-5.62838256e-01 -1.94295630e-01 8.46155047e-01 -7.94071436e-01
1.28975868e+00 -1.55703323e-02 5.39500237e-01 1.09972656e+00
5.40186942e-01 5.97396672e-01 1.17497849e+00 -6.79022908e-01
3.18883717e-01 2.34498098e-01 6.50453329e-01 9.01314259e-01
2.90524781e-01 1.04628050e+00 -8.96537423e-01 -1.28288642e-01
7.58074284e-01 -4.87720340e-01 1.84641201e-02 -5.20363510e-01
-7.47300088e-01 8.96160424e-01 7.82988667e-01 -3.14803809e-01
-3.83756101e-01 6.40987992e-01 2.53871650e-01 -2.53604263e-01
1.58509880e-01 6.30112410e-01 -7.06174076e-01 3.76150399e-01
-6.24512792e-01 4.72594559e-01 5.45960724e-01 4.64279473e-01
6.43524766e-01 2.85011590e-01 9.76987258e-02 7.96324670e-01
7.41113067e-01 7.49354780e-01 2.87875105e-02 -1.24546027e+00
-6.94986507e-02 -9.98229161e-02 -2.45860964e-03 -6.35835648e-01
-3.98919016e-01 -2.29886889e-01 -5.27404249e-01 6.91333473e-01
2.85924166e-01 8.49868695e-04 -1.18770182e+00 1.59820235e+00
3.44207644e-01 4.91014361e-01 -5.13978228e-02 7.52894640e-01
1.09289348e+00 8.38739395e-01 6.88105881e-01 -8.50796252e-02
1.47051704e+00 -9.44272161e-01 -3.14319372e-01 -3.38229775e-01
7.51735494e-02 -5.49802780e-01 9.95714366e-01 4.82192457e-01
-9.47682619e-01 -9.08440471e-01 -9.16389883e-01 -5.02741516e-01
-4.58739966e-01 -3.31590801e-01 1.07984817e+00 6.21736169e-01
-9.48231816e-01 1.01857817e+00 -1.19296479e+00 -5.36829412e-01
6.18828237e-01 1.39787853e-01 2.39493579e-01 4.12426353e-01
-1.09885907e+00 1.40320718e+00 3.56888771e-01 -2.19937697e-01
-7.81367838e-01 -1.41324759e+00 -8.01292002e-01 -3.81930351e-01
1.00826956e-01 -1.20866179e+00 1.51623511e+00 -5.59709132e-01
-1.46435535e+00 8.38860631e-01 -1.60599008e-01 -8.26054931e-01
1.89271063e-01 -2.39127114e-01 -2.11517453e-01 2.98885942e-01
-3.62663627e-01 1.08087754e+00 6.62432730e-01 -1.67254221e+00
-8.50498855e-01 -1.08366147e-01 3.49597842e-01 1.45237714e-01
2.66949832e-02 -6.59053624e-02 -2.40373164e-01 -3.08633387e-01
-1.66275159e-01 -9.27717865e-01 -2.39430353e-01 7.07329631e-01
-5.97129464e-02 -1.93214223e-01 7.14945793e-01 -5.73309541e-01
5.65083325e-01 -1.64438808e+00 1.01094410e-01 -2.44186185e-02
3.53156418e-01 2.81517804e-01 5.34019619e-02 3.26463282e-01
4.04856056e-02 9.92956534e-02 -2.59215266e-01 -6.51296182e-03
3.97226959e-02 4.12482828e-01 -7.03429580e-01 3.88447642e-01
3.28199655e-01 1.11151862e+00 -9.27406371e-01 -3.51342618e-01
9.02201772e-01 8.01609755e-01 -4.11904752e-01 6.53203717e-03
-3.79405946e-01 4.91031110e-01 -6.08932316e-01 4.13586766e-01
6.65871561e-01 -2.43640259e-01 -4.28456634e-01 -5.89025855e-01
-2.07712874e-01 5.28156281e-01 -8.02391708e-01 1.19123435e+00
-5.27636588e-01 8.60568047e-01 -1.36290953e-01 -1.17245519e+00
6.65879905e-01 -2.25020558e-01 3.97355795e-01 -6.23437047e-01
-1.39097825e-01 -1.12901218e-01 1.71995029e-01 -7.79353440e-01
1.87478244e-01 -6.11693621e-01 5.89300752e-01 2.42990613e-01
2.26291284e-01 -5.88695228e-01 -3.62301506e-02 5.73775694e-02
7.42254615e-01 7.50083387e-01 3.84811491e-01 -3.30098629e-01
5.80742240e-01 3.99281949e-01 1.81584984e-01 1.15980220e+00
-5.53005219e-01 1.95728928e-01 7.54684582e-02 -6.26837611e-01
-1.16780496e+00 -1.54836404e+00 -5.35270810e-01 1.06825387e+00
2.75720984e-01 1.01224601e-01 -4.07558709e-01 -3.08034331e-01
4.44511503e-01 9.73228693e-01 -8.89970362e-01 -6.93177104e-01
-2.59207189e-01 -1.12962282e+00 5.10684133e-01 8.57898891e-01
4.42752212e-01 -1.32062876e+00 -7.53218770e-01 -5.84586672e-02
2.82147378e-01 -1.12476480e+00 2.82680005e-01 1.66338727e-01
-1.07753539e+00 -1.08984053e+00 -1.51987046e-01 -2.38367945e-01
2.25116774e-01 2.94003010e-01 1.21295428e+00 -2.19063219e-02
-6.51908755e-01 9.27352548e-01 -6.30825385e-02 -9.45115745e-01
-6.11222208e-01 -3.73193443e-01 2.57337451e-01 -5.05878031e-01
5.55330873e-01 -5.08460164e-01 -7.19818473e-01 -1.56169161e-01
-5.88566244e-01 9.97272134e-02 5.47683954e-01 5.83973944e-01
4.21312213e-01 -1.96351305e-01 1.95602670e-01 -6.62314832e-01
3.75751615e-01 -3.44466627e-01 -5.88533580e-01 3.52579683e-01
-8.32679868e-01 2.03790784e-01 3.69673461e-01 -4.57596898e-01
-1.73421240e+00 -1.07408561e-01 -4.66560163e-02 -2.40383297e-01
-4.77108598e-01 2.68889457e-01 2.09868670e-01 -4.09956932e-01
1.04610229e+00 6.05502576e-02 -1.69613764e-01 -4.87805009e-01
8.42796504e-01 1.42476276e-01 8.13723922e-01 -8.97715747e-01
8.40721965e-01 7.44973123e-01 4.36495036e-01 -1.04998684e+00
-1.23605287e+00 -3.16177636e-01 -9.04556155e-01 -3.17289680e-01
9.35852945e-01 -4.57401097e-01 -6.80772841e-01 6.06996953e-01
-1.09083188e+00 -4.60279286e-01 -4.92268443e-01 6.65735841e-01
-7.59473860e-01 4.45316941e-01 -6.42279685e-01 -9.90910113e-01
-9.17271972e-02 -9.37017024e-01 7.30009079e-01 6.67660594e-01
-6.35005757e-02 -1.76035440e+00 2.78566539e-01 5.32477617e-01
9.47839916e-02 4.12627608e-02 1.23643160e+00 1.91968028e-02
-4.53555554e-01 4.23251897e-01 -4.87142235e-01 2.82857388e-01
9.82007384e-02 6.00383639e-01 -1.48298979e+00 1.46088868e-01
8.63380209e-02 -4.21638817e-01 1.29394841e+00 1.01549792e+00
1.18472588e+00 2.48077959e-01 -3.61361980e-01 7.61646509e-01
1.50839198e+00 2.72401065e-01 3.79378617e-01 2.57907331e-01
8.01782250e-01 8.98739696e-01 1.98759586e-01 8.95329341e-02
1.44985184e-01 3.18707049e-01 4.15936291e-01 -1.55798540e-01
-5.74208856e-01 -2.11516395e-01 3.47150832e-01 4.26616192e-01
-3.25913399e-01 -6.12198450e-02 -1.06903362e+00 2.40359947e-01
-1.78521836e+00 -9.59228873e-01 -5.67530930e-01 1.82886493e+00
7.80700326e-01 3.42338204e-01 -2.30472777e-02 -2.43837193e-01
5.96338093e-01 6.21148050e-02 -1.20417595e+00 -6.26013756e-01
-2.34927405e-02 1.59667164e-01 5.47769964e-01 8.66770566e-01
-1.09241998e+00 9.99375105e-01 7.78359890e+00 4.57013398e-01
-1.14864290e+00 6.57323971e-02 4.97123361e-01 9.23344493e-02
-7.66655579e-02 2.57469982e-01 -9.86301184e-01 1.01763114e-01
9.55018938e-01 1.42131329e-01 4.35028017e-01 6.45650208e-01
6.10110283e-01 -3.66071761e-01 -1.33288109e+00 8.17472398e-01
-1.58148855e-01 -1.68948615e+00 3.66864949e-01 5.35244122e-02
7.14648426e-01 5.34048200e-01 2.67637908e-01 -4.49542366e-02
6.25156462e-01 -8.98710251e-01 7.44167686e-01 9.75718796e-01
4.70379472e-01 -1.32322013e-01 1.54296756e-01 1.37765616e-01
-8.66156220e-01 -7.97094330e-02 -8.48493874e-01 -3.81218046e-01
1.28942356e-01 6.65138841e-01 -7.27726370e-02 4.55524325e-01
8.81349742e-01 8.52662563e-01 -5.33599138e-01 1.14895415e+00
-3.44770581e-01 8.26118290e-01 -4.19980198e-01 -2.89480388e-02
3.05366993e-01 -7.69172087e-02 5.10003984e-01 1.35312200e+00
-3.54638904e-01 1.80078045e-01 8.35413486e-02 1.22389627e+00
3.23624909e-01 -3.91857058e-01 -4.63915288e-01 2.37356961e-01
2.71009982e-01 1.32837379e+00 -6.99689925e-01 -2.89977908e-01
-5.30414999e-01 3.78766984e-01 3.86133134e-01 4.41464990e-01
-9.82747674e-01 2.28588417e-01 8.56098652e-01 -2.24403650e-01
1.19927093e-01 -2.44095489e-01 -5.85760951e-01 -9.16398823e-01
-2.34797388e-01 -3.11795443e-01 2.69967347e-01 -1.15073657e+00
-1.86007655e+00 -1.12327173e-01 5.04234433e-01 -3.95264447e-01
3.00131947e-01 -1.53263009e+00 -8.25456977e-01 8.14322114e-01
-1.72414339e+00 -1.29018962e+00 -2.88557410e-01 1.73000351e-01
4.81341451e-01 -1.45681482e-02 5.02592325e-01 -2.88263142e-01
-3.52709025e-01 1.57170027e-01 2.76510537e-01 -4.21464890e-01
4.02509868e-01 -1.33505642e+00 5.43912172e-01 4.64416146e-01
3.36735219e-01 6.68335199e-01 1.15707004e+00 -7.95901120e-01
-1.33515275e+00 -7.79706895e-01 2.19194084e-01 -7.98104644e-01
9.88432288e-01 -1.60189822e-01 -1.18239880e+00 3.69990289e-01
2.10020915e-01 8.95755589e-02 6.56653404e-01 2.07275227e-01
-5.92006385e-01 -4.60103480e-03 -1.19318914e+00 4.64772016e-01
1.21217644e+00 -5.55372417e-01 -9.73148108e-01 4.01704431e-01
2.63085395e-01 -1.41819045e-01 -6.01368725e-01 3.02409828e-01
6.15614235e-01 -1.03472292e+00 1.23096979e+00 -1.07834148e+00
5.86791039e-01 -1.15200832e-01 1.68613002e-01 -1.14881432e+00
-6.88287914e-01 -3.53381127e-01 -5.88683188e-01 1.00242460e+00
2.92628378e-01 -6.06910765e-01 6.56063914e-01 8.88933897e-01
-3.01989745e-02 -8.08994472e-01 -6.92680120e-01 -8.16128492e-01
7.43667603e-01 -8.65736008e-01 -6.11621961e-02 5.61083078e-01
-3.62265617e-01 2.61626005e-01 -4.83116135e-02 2.70848364e-01
9.86890435e-01 -2.00146481e-01 2.49667466e-01 -1.48548341e+00
-4.70358938e-01 -7.64796674e-01 -3.44206631e-01 -1.04186034e+00
1.98785931e-01 -7.75199294e-01 -8.02656859e-02 -1.84389710e+00
5.43185651e-01 -2.83737689e-01 -5.90023637e-01 2.00967014e-01
-2.58102477e-01 1.46792039e-01 2.71469742e-01 3.46140742e-01
-1.56503338e-02 4.68120605e-01 1.12005329e+00 -1.27041310e-01
2.17093201e-03 -1.71616316e-01 -8.74743640e-01 1.15816438e+00
8.35101962e-01 -1.92825109e-01 -7.13493466e-01 -4.22534674e-01
1.65528089e-01 -7.13502944e-01 1.16509438e+00 -9.08593416e-01
1.87295645e-01 -6.87734246e-01 4.98296112e-01 -5.16955674e-01
4.02635157e-01 -4.83310163e-01 -4.88143027e-01 5.56297541e-01
-5.17438710e-01 -3.37004662e-01 5.98186910e-01 6.57458723e-01
4.60572124e-01 -5.40699005e-01 1.28235197e+00 -1.35327980e-01
-1.06911600e+00 1.97934434e-01 -5.53679645e-01 -4.93819118e-02
9.34561551e-01 -2.33209923e-01 -2.76577264e-01 -2.34033033e-01
-7.58557260e-01 1.59588903e-01 1.09304860e-01 4.11210150e-01
3.82735312e-01 -9.33555961e-01 -4.13635522e-01 -4.36420888e-01
1.54857924e-02 -4.65765208e-01 4.27351624e-01 5.36750078e-01
-4.16985124e-01 5.45890391e-01 -2.20316395e-01 -7.62493372e-01
-1.01396918e+00 7.71775186e-01 9.01560724e-01 4.59710844e-02
-6.41755760e-01 1.06097269e+00 7.01350391e-01 -4.11019742e-01
1.11198701e-01 -6.32951617e-01 -1.07025228e-01 -3.66229087e-01
3.96398574e-01 5.24428308e-01 -3.35976392e-01 -4.23505574e-01
-2.79502839e-01 9.98213410e-01 -1.29916713e-01 -1.02211580e-01
1.35059845e+00 -2.19426140e-01 8.02402943e-02 7.12702990e-01
6.68376267e-01 -4.28397208e-01 -1.99752688e+00 -3.22644264e-01
5.89893013e-03 -3.96550447e-02 5.25783896e-01 -1.30466402e+00
-9.22949016e-01 1.33405125e+00 7.88932920e-01 2.88043227e-02
7.75569499e-01 3.35608333e-01 4.83563185e-01 6.44436479e-01
-7.48048797e-02 -1.10078096e+00 -1.07239828e-01 5.23107767e-01
8.52128148e-01 -1.37523627e+00 5.71345329e-01 -6.90536022e-01
-3.70220333e-01 1.31896377e+00 8.34783018e-01 -3.91013443e-01
1.13841748e+00 1.90711260e-01 7.44325146e-02 -4.08273429e-01
-7.27818310e-01 -1.28850281e-01 6.24441683e-01 1.21041858e+00
2.73101300e-01 -3.37957382e-01 1.34058923e-01 1.11687236e-01
-1.90458208e-01 -1.34440452e-01 1.73940644e-01 7.21096694e-01
-5.87819695e-01 -9.28539991e-01 -3.86629492e-01 4.89757568e-01
-9.12300274e-02 -2.50518650e-01 -6.08879149e-01 7.37772524e-01
5.00140369e-01 8.62653494e-01 -1.45440642e-02 4.32920828e-02
1.79853231e-01 -1.16302520e-02 9.93027687e-01 -4.90224063e-01
-2.64628857e-01 -2.39758268e-01 -1.76247567e-01 -1.94354221e-01
-8.15954387e-01 -8.26069772e-01 -1.43786180e+00 -3.08254063e-01
-3.37921500e-01 -3.35112840e-01 7.95734644e-01 1.28547394e+00
1.41554907e-01 7.44230866e-01 6.61100447e-03 -1.11820006e+00
-5.95262706e-01 -5.73014677e-01 -1.63235202e-01 7.56711066e-02
5.25583029e-01 -1.21728075e+00 -3.92655313e-01 6.33003771e-01] | [6.526979923248291, 3.6115894317626953] |
7cf4bd05-883d-470b-8cf0-c85b3849d1bb | sentence-level-propaganda-detection-in-news | null | null | https://aclanthology.org/D19-5022 | https://aclanthology.org/D19-5022.pdf | Sentence-Level Propaganda Detection in News Articles with Transfer Learning and BERT-BiLSTM-Capsule Model | In recent years, the need for communication increased in online social media. Propaganda is a mechanism which was used throughout history to influence public opinion and it is gaining a new dimension with the rising interest of online social media. This paper presents our submission to NLP4IF-2019 Shared Task SLC: Sentence-level Propaganda Detection in news articles. The challenge of this task is to build a robust binary classifier able to provide corresponding propaganda labels, propaganda or non-propaganda. Our model relies on a unified neural network, which consists of several deep leaning modules, namely BERT, BiLSTM and Capsule, to solve the sentencelevel propaganda classification problem. In addition, we take a pre-training approach on a somewhat similar task (i.e., emotion classification) improving results against the cold-start model. Among the 26 participant teams in the NLP4IF-2019 Task SLC, our solution ranked 12th with an F1-score 0.5868 on the official test data. Our proposed solution indicates promising results since our system significantly exceeds the baseline approach of the organizers by 0.1521 and is slightly lower than the winning system by 0.0454. | ['Dumitru-Clementin Cercel', 'Cristian Onose', 'Mircea-Adrian Tanase', 'ru', 'George-Alex Vlad'] | 2019-11-01 | null | null | null | ws-2019-11 | ['propaganda-detection'] | ['natural-language-processing'] | [-9.37438309e-02 1.66100696e-01 -7.00183287e-02 -2.85563260e-01
-7.57019639e-01 -3.25613588e-01 8.01080346e-01 5.00464916e-01
-7.41342962e-01 7.37941682e-01 4.93618339e-01 -3.55975807e-01
4.75593895e-01 -5.70480525e-01 -5.66347182e-01 -6.81108892e-01
2.74512112e-01 9.87447873e-02 -1.91593871e-01 -5.34059227e-01
6.03628039e-01 -8.46511405e-03 -9.85278547e-01 8.25032771e-01
1.10331309e+00 9.67109442e-01 -2.69317627e-01 6.22989476e-01
-1.50382429e-01 1.36603177e+00 -1.05049205e+00 -6.95747554e-01
-3.08705300e-01 -5.07037103e-01 -9.46853161e-01 -4.30720568e-01
4.45088029e-01 3.46057899e-02 2.78841674e-01 1.01629353e+00
5.79833508e-01 2.46249475e-02 6.36928380e-01 -7.66203701e-01
-7.76611984e-01 9.45922732e-01 -7.41197586e-01 2.02492267e-01
3.38976473e-01 -2.90251017e-01 1.00255406e+00 -8.28303993e-01
9.20020163e-01 1.25249839e+00 6.73395157e-01 6.11187875e-01
-1.05478489e+00 -5.89090228e-01 1.46767154e-01 2.95780480e-01
-8.24473202e-01 -1.58145234e-01 7.81921208e-01 -6.77664638e-01
1.12825656e+00 5.82225509e-02 6.37000084e-01 1.80944097e+00
6.51242614e-01 8.00337791e-01 1.36032820e+00 -4.44532543e-01
1.93880290e-01 4.26640898e-01 3.36234510e-01 6.38527989e-01
-5.43549657e-02 -4.21838939e-01 -7.55042195e-01 -2.45568290e-01
-5.23676127e-02 -5.66163361e-01 -9.92522985e-02 5.51202118e-01
-1.01922071e+00 1.29638302e+00 6.47917688e-01 7.78659999e-01
-3.10186207e-01 1.64384186e-01 6.98978961e-01 4.86427039e-01
1.31395292e+00 7.37218976e-01 -4.94783074e-01 -1.77984253e-01
-8.72615159e-01 3.41292173e-01 9.95148957e-01 -1.86313853e-01
1.11298271e-01 -4.69046235e-02 -2.26754919e-01 8.24988067e-01
8.07974190e-02 6.01852715e-01 4.28945959e-01 -1.93544909e-01
4.98523504e-01 3.61491203e-01 7.82945007e-02 -1.48114514e+00
-8.18839312e-01 -6.78587496e-01 -9.92802620e-01 5.07470034e-02
3.61011982e-01 -6.36220634e-01 -8.68270457e-01 1.74531984e+00
2.57198960e-01 -2.99208909e-01 -1.92731455e-01 7.65061855e-01
9.78683770e-01 1.25686133e+00 2.52444357e-01 -2.93070227e-01
1.18329358e+00 -1.13464534e+00 -8.95537734e-01 -1.84457377e-01
9.91947711e-01 -9.48417723e-01 7.56692708e-01 7.28750467e-01
-6.72880709e-01 -2.53830373e-01 -8.79906654e-01 7.62492567e-02
-6.23921275e-01 2.01296419e-01 4.54622418e-01 5.47298908e-01
-7.79747427e-01 4.79091942e-01 -3.53001297e-01 -3.58471066e-01
3.08921844e-01 -1.46522611e-01 -2.62068450e-01 4.75799799e-01
-1.53215635e+00 1.39722562e+00 2.29537711e-01 1.97654784e-01
-6.16333902e-01 -6.22940898e-01 -6.24819994e-01 -3.28732461e-01
3.32268417e-01 -2.48922855e-01 1.08087528e+00 -1.28786218e+00
-1.41549027e+00 1.08355165e+00 1.81673557e-01 -7.05091119e-01
4.49687153e-01 -6.97027326e-01 -6.33316576e-01 -1.22352317e-01
2.06933185e-01 4.92403060e-01 1.02373934e+00 -9.21882987e-01
-4.30277735e-01 1.45954713e-02 -2.33433902e-01 -1.48931935e-01
-3.79021108e-01 6.57177627e-01 5.12141526e-01 -5.63451767e-01
-4.02658105e-01 -7.39829898e-01 -5.85294999e-02 -5.70628643e-01
-7.50480413e-01 -6.24882400e-01 6.48153424e-01 -9.80514526e-01
1.21295691e+00 -1.99451983e+00 1.34895384e-01 -2.94759125e-03
2.68118680e-01 3.27135473e-01 -1.11607693e-01 6.64919138e-01
-2.21924946e-01 5.18601120e-01 5.31854518e-02 -1.81647226e-01
-4.88973260e-02 -3.59562457e-01 -6.36136115e-01 6.30740643e-01
4.15076792e-01 5.62971473e-01 -9.16199505e-01 -3.30993742e-01
-2.94835389e-01 4.81278658e-01 -3.41208190e-01 -7.14432001e-02
-3.36721659e-01 2.63567746e-01 -2.50793487e-01 3.75377178e-01
5.85767508e-01 -3.82926285e-01 4.26765792e-02 6.63984045e-02
-5.37576318e-01 6.45472288e-01 -4.48304355e-01 1.43608046e+00
-3.54474664e-01 9.54003632e-01 2.99034685e-01 -9.58295166e-01
9.20187831e-01 3.44141573e-01 2.49270126e-01 -6.58651173e-01
6.71126425e-01 2.28254393e-01 9.39099342e-02 -4.83528435e-01
3.88303220e-01 -2.88503230e-01 -6.42669559e-01 6.09368980e-01
7.44228885e-02 -7.22884610e-02 2.09986821e-01 5.47226012e-01
1.16295385e+00 -2.35590294e-01 5.19642949e-01 -4.17725176e-01
4.45291579e-01 9.54262391e-02 3.75303864e-01 8.36726427e-01
-3.54404837e-01 2.05325380e-01 8.04854810e-01 -7.98222244e-01
-5.96930981e-01 -7.38935947e-01 -1.07440121e-01 1.22497714e+00
-2.26947367e-01 -4.91488516e-01 -6.61325157e-01 -1.00620842e+00
-3.77177715e-01 8.92056942e-01 -1.12579095e+00 9.39888880e-02
-5.15247107e-01 -1.10881412e+00 4.93718326e-01 -1.44487455e-01
6.59471631e-01 -1.47579992e+00 -6.27512395e-01 4.10459995e-01
-5.09737909e-01 -8.16475153e-01 1.02765761e-01 5.28149426e-01
-2.47245654e-01 -8.59109521e-01 -5.63947320e-01 -7.47482181e-01
2.74884582e-01 -1.20192789e-01 1.06011891e+00 3.34676653e-02
-2.26560369e-01 -1.67350203e-01 -7.60453939e-01 -7.61712134e-01
-6.04868352e-01 3.17747533e-01 -1.76143393e-01 -1.15587311e-02
2.09077924e-01 -1.46728650e-01 -1.11996807e-01 -3.67008924e-01
-6.13878787e-01 3.81817132e-01 3.13035488e-01 1.00371695e+00
-2.26792961e-01 -3.25353682e-01 9.76208627e-01 -8.24090481e-01
1.06928110e+00 -5.55276275e-01 -1.24479473e-01 -8.79248828e-02
-3.28732997e-01 -3.15111905e-01 7.16803670e-01 -2.25355878e-01
-1.07024932e+00 -2.96277344e-01 -4.19819534e-01 5.78700960e-01
-2.05223203e-01 8.50964248e-01 4.59184498e-01 2.22330466e-01
1.00363183e+00 -2.78670877e-01 -1.53711826e-01 -5.86478055e-01
4.65001494e-01 6.60499752e-01 1.06948830e-01 -3.66992690e-02
5.48110902e-01 2.03360915e-01 -4.46581542e-01 -8.10097277e-01
-1.69262373e+00 -3.19070488e-01 -2.76888132e-01 -3.70728254e-01
1.10686457e+00 -8.44934702e-01 -6.61195576e-01 8.84436846e-01
-1.79069138e+00 -2.52321541e-01 2.43090436e-01 2.18731314e-01
1.07968859e-01 1.32208258e-01 -1.05498362e+00 -9.27577436e-01
-7.30175912e-01 -4.80736226e-01 7.31664658e-01 -3.22144553e-02
-4.60008532e-01 -9.73515451e-01 4.82181668e-01 5.17029583e-01
5.79890132e-01 6.31372750e-01 8.92813683e-01 -9.85067785e-01
5.07496774e-01 -3.13857377e-01 -1.11807704e-01 5.68886817e-01
-1.42322809e-01 1.07585862e-01 -9.81713057e-01 7.68361613e-02
2.87627935e-01 -8.57188523e-01 1.35332084e+00 2.83388227e-01
7.45383620e-01 -5.81168175e-01 1.37708317e-02 -3.17470431e-01
1.07454050e+00 -1.78494323e-02 3.96635711e-01 5.99343061e-01
5.05002737e-01 7.85924077e-01 5.09398639e-01 5.16986132e-01
2.18679726e-01 4.70329940e-01 5.47325909e-01 -2.19950899e-01
1.70242935e-01 -7.12745450e-03 8.78609300e-01 9.37296569e-01
3.15282010e-02 -5.61564267e-01 -1.06825829e+00 2.94946551e-01
-1.86836481e+00 -1.30655265e+00 -5.76696277e-01 1.54496229e+00
9.70928371e-01 1.34572491e-01 -1.92045781e-03 7.09260553e-02
7.67596424e-01 6.36980534e-01 1.23903252e-01 -1.25322950e+00
-3.00602496e-01 2.43246198e-01 -1.79326758e-01 6.83472931e-01
-1.67737412e+00 1.08390820e+00 5.76452637e+00 1.05197597e+00
-1.44348931e+00 4.32577074e-01 8.57943296e-01 -3.04985493e-01
2.79599801e-02 -4.98827249e-01 -6.43296659e-01 5.96520722e-01
1.17753792e+00 2.87580285e-02 7.23908097e-02 8.36494565e-01
3.45306218e-01 -3.21142972e-01 -5.13688445e-01 6.22905791e-01
6.46697819e-01 -1.44746995e+00 -1.47986561e-01 -9.80150625e-02
9.43453908e-01 3.89700294e-01 1.18923880e-01 6.81981921e-01
1.70374066e-01 -1.16148460e+00 8.75995100e-01 1.05676226e-01
-4.61887307e-02 -6.04453564e-01 1.10115170e+00 6.49580598e-01
-3.14702302e-01 6.65061846e-02 -7.01331645e-02 -5.54072082e-01
5.68587959e-01 1.17184627e+00 -8.44072223e-01 3.68888646e-01
6.15586340e-01 7.75274098e-01 -5.14223516e-01 5.86209178e-01
-7.07931101e-01 1.06706417e+00 -1.14283085e-01 -6.71545684e-01
7.33214021e-01 1.64217278e-01 5.70004582e-01 1.66850650e+00
1.19038172e-01 -3.74530613e-01 7.72550106e-02 5.71218669e-01
-2.06258386e-01 5.35241067e-01 -4.74012524e-01 -2.28735194e-01
-1.38933092e-01 1.56879246e+00 -6.42277062e-01 -3.94528270e-01
1.35288551e-01 8.51562738e-01 6.26001716e-01 1.06158983e-02
-1.23159575e+00 -3.01760733e-01 1.43714279e-01 -1.43473148e-01
-4.88563068e-02 -2.37387400e-02 -2.00287193e-01 -1.10002434e+00
-1.70497268e-01 -7.66167402e-01 1.84419185e-01 -5.58832467e-01
-1.49898303e+00 9.20873225e-01 -4.70691532e-01 -6.29790068e-01
-2.19479278e-01 -7.01223671e-01 -7.65953422e-01 6.15869284e-01
-1.04842973e+00 -1.06014299e+00 3.31895193e-04 2.85002682e-02
4.70565319e-01 8.44909027e-02 9.31193352e-01 1.58443153e-01
-5.48769236e-01 8.14258158e-02 -2.17046402e-02 1.92273259e-01
1.06071484e+00 -1.08324683e+00 1.03328943e-01 8.34132671e-01
2.20509484e-01 3.94930899e-01 1.01109993e+00 -6.54124498e-01
-7.99466729e-01 -9.08375144e-01 1.66619432e+00 -4.49240118e-01
1.19155347e+00 -6.11545265e-01 -6.91124558e-01 1.50302038e-01
7.22055614e-01 -5.50820231e-01 8.23228776e-01 4.98633444e-01
-6.78707600e-01 2.65099645e-01 -1.05592883e+00 3.67606640e-01
4.79195356e-01 -2.91462451e-01 -7.56420255e-01 8.00690472e-01
6.24247909e-01 -1.58903956e-01 -4.00076121e-01 1.44199058e-01
4.14011419e-01 -8.40275526e-01 3.44669759e-01 -7.28684545e-01
1.24295688e+00 1.39342219e-01 2.10219085e-01 -1.63199091e+00
-3.58953595e-01 -3.93462390e-01 1.59584880e-01 1.17564547e+00
7.68657446e-01 -5.63959420e-01 4.42956030e-01 -1.69698172e-03
-8.03652331e-02 -8.92010272e-01 -1.19146550e+00 -5.30445933e-01
3.06349099e-01 -4.44523454e-01 -5.08879840e-01 1.18588567e+00
3.80774617e-01 1.00613320e+00 -9.22326326e-01 -4.51662391e-01
2.43200615e-01 1.45545945e-01 4.23470169e-01 -1.18144917e+00
-7.86370412e-02 -6.00648046e-01 -3.17464396e-02 -6.06709957e-01
3.11335683e-01 -1.00261402e+00 2.42796674e-01 -1.75217879e+00
3.26707482e-01 1.45973116e-01 -3.07183474e-01 6.22236371e-01
-3.81277576e-02 2.89204866e-01 2.63626873e-01 -1.59403682e-01
-8.48461092e-01 5.14027178e-01 1.06982172e+00 -3.41384381e-01
-1.45941854e-01 -4.90909547e-01 -9.09306645e-01 8.23798001e-01
1.19901824e+00 -6.88780248e-01 3.28862160e-01 -3.49722475e-01
8.34547102e-01 -4.69759524e-01 4.33274031e-01 -7.13862956e-01
-2.11313870e-02 7.54207671e-02 2.99779028e-01 -6.62958205e-01
2.85630643e-01 -1.52307570e-01 -3.11828196e-01 8.59512389e-01
-4.78675991e-01 -1.68423310e-01 1.32475108e-01 3.31384033e-01
-2.37436533e-01 -2.27146357e-01 6.15123153e-01 2.87887026e-02
-3.00703108e-01 -3.69232744e-01 -1.14280987e+00 9.89854634e-02
7.52626896e-01 4.81533438e-01 -1.02953792e+00 -4.15143579e-01
-6.53368533e-01 -7.68919438e-02 -2.09114894e-01 5.67647159e-01
2.15329856e-01 -1.00981390e+00 -1.18829083e+00 -4.43102598e-01
-2.04983473e-01 -6.23924136e-01 2.23113865e-01 1.20710766e+00
-4.76876527e-01 3.50726098e-01 -1.31319300e-03 -1.40968740e-01
-1.29165173e+00 1.41695738e-01 -3.32920551e-02 -7.72832930e-01
-4.52109456e-01 1.14591682e+00 -1.40286103e-01 -2.40000248e-01
-5.34499176e-02 -1.25180468e-01 -4.04883474e-01 6.20016277e-01
7.66285181e-01 3.09121579e-01 1.80467978e-01 -6.62322819e-01
-5.79965889e-01 7.54606649e-02 -3.63749504e-01 -2.31325656e-01
1.47572351e+00 4.77914691e-01 -5.32198966e-01 6.47645533e-01
1.12374067e+00 3.07139605e-01 -2.53536850e-01 1.68754548e-01
8.75224695e-02 1.22567691e-01 3.16616833e-01 -1.50034130e+00
-6.03462338e-01 8.71600211e-01 2.04660773e-01 6.73645794e-01
5.37536502e-01 3.57840993e-02 7.14641094e-01 5.92942595e-01
1.46470159e-01 -1.49956918e+00 2.47974187e-01 1.17840385e+00
1.46217525e+00 -1.26345062e+00 -6.48652017e-02 -3.50664347e-01
-6.19745493e-01 9.86686587e-01 4.83395785e-01 -3.89924198e-01
5.37762463e-01 2.03312352e-01 3.32626641e-01 -3.82626474e-01
-9.88206923e-01 2.61182934e-01 3.04056019e-01 1.37468800e-01
9.22485352e-01 1.47792101e-01 -1.05898118e+00 8.02695930e-01
-1.46740854e-01 -3.65477413e-01 4.17074770e-01 6.20320499e-01
-7.79527664e-01 -7.13330150e-01 -3.27818543e-01 4.71301496e-01
-9.53975737e-01 -3.43781710e-01 -1.06430423e+00 6.29089475e-01
1.85020015e-01 1.30278528e+00 -1.53962687e-01 -5.77918351e-01
-7.71216229e-02 7.62039274e-02 1.82890162e-01 -5.73259652e-01
-1.44843197e+00 1.37719750e-01 7.60592937e-01 -5.64017892e-01
-6.32321477e-01 -3.56653333e-01 -9.43376184e-01 -4.95836437e-01
-4.40866113e-01 4.39307213e-01 9.42067623e-01 9.45256829e-01
5.71519546e-02 5.53832114e-01 4.88259554e-01 -9.00246799e-01
-5.93396366e-01 -1.42299461e+00 -2.51846880e-01 2.64081001e-01
1.06673479e-01 -3.99966687e-01 -7.22392499e-01 -1.58888549e-01] | [8.494525909423828, 10.654776573181152] |
af894ebc-39d5-4574-979f-ad0fea584211 | multiple-discrimination-and-pairwise-cnn-for | 2002.11977 | null | https://arxiv.org/abs/2002.11977v1 | https://arxiv.org/pdf/2002.11977v1.pdf | Multiple Discrimination and Pairwise CNN for View-based 3D Object Retrieval | With the rapid development and wide application of computer, camera device, network and hardware technology, 3D object (or model) retrieval has attracted widespread attention and it has become a hot research topic in the computer vision domain. Deep learning features already available in 3D object retrieval have been proven to be better than the retrieval performance of hand-crafted features. However, most existing networks do not take into account the impact of multi-view image selection on network training, and the use of contrastive loss alone only forcing the same-class samples to be as close as possible. In this work, a novel solution named Multi-view Discrimination and Pairwise CNN (MDPCNN) for 3D object retrieval is proposed to tackle these issues. It can simultaneously input of multiple batches and multiple views by adding the Slice layer and the Concat layer. Furthermore, a highly discriminative network is obtained by training samples that are not easy to be classified by clustering. Lastly, we deploy the contrastive-center loss and contrastive loss as the optimization objective that has better intra-class compactness and inter-class separability. Large-scale experiments show that the proposed MDPCNN can achieve a significant performance over the state-of-the-art algorithms in 3D object retrieval. | ['Z. Gao', 'K. X Xue', 'S. H Wan'] | 2020-02-27 | null | null | null | null | ['3d-object-retrieval'] | ['computer-vision'] | [-3.39796305e-01 -8.04370522e-01 -9.38938260e-02 -4.31211114e-01
-8.43219995e-01 -4.16307420e-01 4.93777752e-01 -2.40207035e-02
-4.73356068e-01 1.54388383e-01 -3.24792802e-01 2.34857291e-01
-4.68219161e-01 -6.50164723e-01 -3.26124787e-01 -8.67335439e-01
2.18465164e-01 5.77612877e-01 2.69025683e-01 1.71760216e-01
1.75493151e-01 9.53229070e-01 -1.71871996e+00 7.83337653e-02
7.65706658e-01 1.56522202e+00 3.23995113e-01 6.43977374e-02
-4.95598689e-02 1.97458059e-01 -4.52508926e-01 -1.97863579e-01
5.96724153e-01 -1.35868475e-01 -4.27627802e-01 1.66789770e-01
5.99947453e-01 -4.86018300e-01 -3.41985375e-01 1.09614837e+00
9.19358671e-01 1.78126812e-01 7.87663341e-01 -1.15110278e+00
-8.10603499e-01 -8.59409571e-03 -9.02333975e-01 2.55147010e-01
-1.26714975e-01 -1.27876371e-01 9.94133651e-01 -1.19751716e+00
4.69786316e-01 1.35154068e+00 2.04736918e-01 3.12202722e-01
-8.78503919e-01 -7.47895241e-01 9.30583477e-02 4.28355932e-01
-1.71738017e+00 -1.31926268e-01 1.03206801e+00 -1.89952463e-01
7.80140102e-01 1.73291385e-01 7.71908939e-01 5.64751029e-01
6.62389249e-02 1.17823064e+00 8.70444894e-01 -2.34335810e-01
-8.24505910e-02 1.45010650e-01 4.19042073e-02 5.21141469e-01
2.86033809e-01 -5.12080826e-02 -8.97618905e-02 -7.45368078e-02
7.07295537e-01 6.57965422e-01 -3.21800888e-01 -7.45267749e-01
-9.79299426e-01 7.04401553e-01 7.38696635e-01 2.80313373e-01
-3.56605053e-01 -7.50458390e-02 4.79351193e-01 7.67187029e-02
6.71197474e-01 1.33555353e-01 -2.10854277e-01 3.20472538e-01
-1.13275528e+00 2.74502605e-01 2.32021928e-01 9.92601275e-01
6.24681771e-01 -2.08795011e-01 -4.66775522e-02 1.21542585e+00
3.52625608e-01 7.52519608e-01 3.49305809e-01 -7.17477441e-01
3.46522748e-01 9.86284673e-01 -1.33121461e-01 -1.32548201e+00
-2.49547914e-01 -7.18361437e-01 -1.12872291e+00 -2.24575829e-02
8.56710672e-02 5.01642108e-01 -9.43516731e-01 1.29408050e+00
5.03772378e-01 -1.76232368e-01 -1.35062546e-01 1.41643310e+00
9.59817231e-01 6.35102272e-01 -3.17368299e-01 -1.24527551e-01
1.13967133e+00 -9.95823264e-01 -3.90958309e-01 9.32582021e-02
3.41845334e-01 -1.05604482e+00 7.35752463e-01 3.33107591e-01
-1.26661563e+00 -7.38541663e-01 -1.06081295e+00 -1.78799808e-01
-3.58227074e-01 3.46345037e-01 5.37409306e-01 2.43487686e-01
-6.56853557e-01 4.10281837e-01 -6.13489985e-01 -1.50610238e-01
6.54788017e-01 4.22601521e-01 -5.16725779e-01 -5.84017992e-01
-9.45907652e-01 7.00955391e-01 4.48278487e-01 4.82078284e-01
-7.97742009e-01 -4.68358278e-01 -5.72731614e-01 2.34566778e-01
3.17861676e-01 -6.95980251e-01 6.25311494e-01 -7.35459924e-01
-1.17183065e+00 1.14760351e+00 1.59711763e-01 6.98893890e-02
5.05506992e-01 -2.13019982e-01 -2.61551052e-01 4.81147826e-01
-2.49605738e-02 6.23315454e-01 8.15718591e-01 -1.37555087e+00
-5.82683802e-01 -7.85619974e-01 -7.53898546e-03 4.67844307e-01
-4.60383117e-01 1.03158336e-02 -9.51232910e-01 -4.47017342e-01
5.07349074e-01 -9.59455431e-01 7.21254498e-02 4.19915199e-01
-3.16701353e-01 -4.45608497e-01 1.05553484e+00 -3.95106673e-01
9.02411520e-01 -2.18169212e+00 3.61660600e-01 3.17021757e-01
2.28516817e-01 6.30064845e-01 -2.52383381e-01 1.32780492e-01
9.68759358e-02 1.03870807e-02 1.14755891e-03 -3.54503989e-01
-9.56749395e-02 -4.36306000e-02 1.43399119e-01 7.08363056e-01
4.45970371e-02 7.15467095e-01 -6.56166553e-01 -7.73962021e-01
4.22514319e-01 5.45698524e-01 -5.54346442e-01 2.86050171e-01
1.03289217e-01 2.50583410e-01 -7.31638193e-01 8.31996024e-01
1.18890655e+00 -5.09570003e-01 -2.74557590e-01 -3.26117784e-01
1.08967476e-01 -2.53062844e-01 -1.19996905e+00 1.77499831e+00
-3.36300611e-01 3.63887846e-01 3.71851586e-02 -1.05936611e+00
1.01260900e+00 6.19917363e-03 4.94424134e-01 -6.79404140e-01
2.12353349e-01 4.66550589e-01 -1.86605871e-01 -6.16140842e-01
3.88615668e-01 3.86691019e-02 1.80652022e-01 3.97951782e-01
5.97243197e-02 -2.61989772e-01 3.95625532e-02 1.44691030e-02
5.22272110e-01 -2.21896410e-01 -4.31720503e-02 -6.22630976e-02
7.91525543e-01 -3.03785264e-01 5.65020680e-01 4.28778887e-01
-1.44322559e-01 9.42484140e-01 1.00998737e-01 -4.14413214e-01
-9.79553282e-01 -8.94842446e-01 -1.37219921e-01 5.68652630e-01
5.91338992e-01 1.08686283e-01 -3.65252823e-01 -7.68724382e-01
1.64343044e-01 5.07348850e-02 -3.96995336e-01 -3.69749188e-01
-5.31551301e-01 -5.44758618e-01 3.23056132e-01 3.50244552e-01
7.91659117e-01 -7.81046271e-01 -4.17507648e-01 -1.82533413e-01
2.02692784e-02 -1.07852721e+00 -6.43928170e-01 -2.07274742e-02
-9.95394170e-01 -1.12824321e+00 -1.21849024e+00 -1.06351316e+00
8.63037229e-01 8.25853169e-01 8.74091685e-01 3.27701449e-01
-3.37211192e-01 3.53505105e-01 -4.94659781e-01 -2.97344655e-01
2.03271791e-01 2.36596584e-01 -3.98542918e-03 1.27374247e-01
4.46631491e-01 -5.34853339e-01 -1.06936955e+00 4.69404042e-01
-1.11787713e+00 -1.88977599e-01 7.68474281e-01 1.05568528e+00
8.67382050e-01 9.31169689e-02 3.45377386e-01 -4.02045608e-01
4.01003361e-01 -2.40103602e-01 -6.80208564e-01 5.53133130e-01
-6.39114141e-01 -3.53749692e-01 4.99420404e-01 -3.92967314e-01
-6.36308074e-01 -1.95463061e-01 5.90654165e-02 -1.17931116e+00
-1.72347203e-03 4.80243534e-01 -4.16749954e-01 -2.19656557e-01
8.18898156e-02 4.63496268e-01 2.09236935e-01 -6.05220079e-01
1.82489991e-01 7.30646431e-01 -1.25460669e-01 -2.31291607e-01
8.06293964e-01 5.51537335e-01 7.83839598e-02 -6.53308213e-01
-8.40574861e-01 -6.50796294e-01 -5.40352702e-01 -2.02286944e-01
7.03154206e-01 -9.61773396e-01 -7.04602599e-01 6.73872650e-01
-1.18688583e+00 3.14576030e-01 2.96983831e-02 7.47820020e-01
-6.46569505e-02 4.74377573e-01 -4.10046399e-01 -7.65351772e-01
-6.65213525e-01 -1.29894209e+00 1.24331856e+00 4.09802794e-01
4.81611669e-01 -6.41380072e-01 -4.09331381e-01 3.79272878e-01
4.03860033e-01 -7.74444640e-02 1.14357412e+00 -8.51136446e-01
-8.95668805e-01 -5.32223105e-01 -7.03123033e-01 6.12847209e-01
7.72659034e-02 9.59778950e-03 -8.35392118e-01 -6.07364655e-01
2.89565837e-03 -3.90835881e-01 1.04633665e+00 5.35873711e-01
1.42108512e+00 5.35013899e-02 -3.48561764e-01 5.87769449e-01
1.53923345e+00 2.65043437e-01 4.07555878e-01 1.10786758e-01
7.00037003e-01 5.91371119e-01 6.28937602e-01 3.70145172e-01
1.96062133e-01 6.36264145e-01 6.14260733e-01 -5.36108501e-02
3.10439356e-02 6.55129552e-03 -2.62816042e-01 1.09847379e+00
2.26098597e-02 -4.83788967e-01 -7.13541746e-01 6.12532139e-01
-1.70368373e+00 -9.37064469e-01 2.35903755e-01 2.17141581e+00
2.38067389e-01 7.29729682e-02 -1.92707181e-01 1.50274009e-01
8.47209096e-01 2.74665415e-01 -6.41517937e-01 1.56592503e-01
-2.01978922e-01 -3.44252251e-02 2.34822541e-01 3.12136449e-02
-1.15308118e+00 5.76453865e-01 4.78541851e+00 1.37209880e+00
-1.26480925e+00 -9.28785503e-02 6.85250640e-01 -3.32778305e-01
-3.07509243e-01 -3.15802395e-01 -8.59634995e-01 3.84707600e-01
-3.04620489e-02 2.69974291e-01 2.40171909e-01 7.48988330e-01
5.94303161e-02 -1.25090197e-01 -1.08956921e+00 1.40271997e+00
3.96167099e-01 -1.13711619e+00 4.93421376e-01 2.34646350e-02
7.51035094e-01 -5.60099296e-02 2.72384763e-01 8.03371146e-02
-4.44071412e-01 -8.14979672e-01 6.19962990e-01 5.83855748e-01
4.90804672e-01 -9.71684098e-01 9.39747930e-01 3.20227027e-01
-1.09890151e+00 -1.16375595e-01 -6.51353836e-01 4.85066384e-01
-4.24420014e-02 8.85602117e-01 -2.58492887e-01 8.18786860e-01
9.57631230e-01 8.76224399e-01 -5.60440302e-01 1.54601526e+00
2.37872496e-01 -1.19663095e-02 -3.84042531e-01 -2.83012360e-01
4.00777340e-01 -3.75204921e-01 6.95012689e-01 7.53106356e-01
5.36592245e-01 5.23332059e-02 2.78965950e-01 6.94818735e-01
-2.02366322e-01 2.78944701e-01 -5.80131710e-01 7.52825513e-02
3.63825083e-01 1.22838771e+00 -7.64240026e-01 -2.21605837e-01
-3.49635780e-01 8.27150464e-01 2.57597804e-01 3.08129728e-01
-4.93107498e-01 -4.77054924e-01 4.28639740e-01 1.99398454e-02
5.53280413e-01 -1.67891607e-01 1.15681946e-01 -1.12970877e+00
3.51170838e-01 -6.87620461e-01 3.52819115e-01 -7.58827806e-01
-1.72159386e+00 5.94765782e-01 -7.19554350e-03 -1.65696537e+00
2.88151622e-01 -5.33680320e-01 -3.27538967e-01 8.61368537e-01
-1.84473848e+00 -1.17627311e+00 -3.98789555e-01 5.21535695e-01
4.41157937e-01 -3.74642253e-01 5.31825900e-01 8.04467797e-01
-5.40143728e-01 5.44953465e-01 4.16938692e-01 9.32328179e-02
8.42020750e-01 -7.53978133e-01 -3.77566665e-01 4.29266751e-01
-5.27385101e-02 5.83688498e-01 3.85011770e-02 -1.77422360e-01
-1.35855985e+00 -1.08832169e+00 6.73038960e-01 7.33328685e-02
1.34855539e-01 -1.66150004e-01 -8.56960297e-01 1.11486495e-01
5.51437289e-02 3.99902284e-01 5.28304815e-01 -1.95598960e-01
-2.26308167e-01 -5.89847565e-01 -1.10579002e+00 2.79194415e-01
7.72830009e-01 -4.87758219e-01 -3.14867377e-01 3.48300755e-01
5.32088637e-01 -4.69025135e-01 -8.13053250e-01 7.21858263e-01
6.58439577e-01 -1.05355382e+00 1.22493577e+00 -5.05943120e-01
4.95886296e-01 -3.70520025e-01 -2.66617894e-01 -9.97234344e-01
-1.72932655e-01 8.89321864e-02 1.90274119e-02 1.25737846e+00
1.90701067e-01 -5.36134064e-01 7.29140341e-01 3.06138217e-01
-1.85747534e-01 -1.18126881e+00 -9.49996591e-01 -7.60076106e-01
7.37830773e-02 -8.28707591e-03 4.76519465e-01 8.28557491e-01
-6.92105412e-01 2.61616945e-01 -2.47350663e-01 1.42799139e-01
6.70955658e-01 7.19769955e-01 6.63656831e-01 -1.56678176e+00
-6.32004216e-02 -5.73479056e-01 -5.28424859e-01 -1.32630992e+00
1.17124096e-01 -1.12228978e+00 -2.09121376e-01 -1.60859311e+00
5.72718382e-01 -8.08783531e-01 -5.86952388e-01 7.62085170e-02
-7.45224655e-02 4.03703272e-01 3.69765282e-01 4.22296524e-01
-8.38495135e-01 9.01984811e-01 1.76361644e+00 -3.99919182e-01
-1.42295465e-01 1.14043996e-01 -3.05714011e-01 5.44662952e-01
4.75197315e-01 -4.74908531e-01 -3.21418524e-01 -7.15338349e-01
1.13925852e-01 7.08584785e-02 5.70404947e-01 -8.03476930e-01
4.34533209e-01 1.56541809e-01 7.12700367e-01 -1.12190390e+00
6.23934746e-01 -1.17417812e+00 -8.70342255e-02 2.37997964e-01
-1.70202643e-01 5.82706630e-02 6.30235896e-02 6.77325249e-01
-5.89782536e-01 -2.89941877e-01 7.93869436e-01 -2.70024896e-01
-4.50342089e-01 7.86179662e-01 9.14112255e-02 -1.05041852e-02
9.53645527e-01 -3.39766860e-01 -1.52931094e-01 -1.26990080e-01
-3.73088568e-01 5.25326133e-01 3.34800512e-01 4.59927142e-01
9.33572292e-01 -1.55773711e+00 -5.52980661e-01 9.13601145e-02
2.73995012e-01 3.85768354e-01 5.57094336e-01 8.03174973e-01
-5.78922451e-01 3.85291576e-01 -1.03456482e-01 -9.30782914e-01
-1.33642340e+00 5.20336986e-01 3.11518729e-01 -3.04808915e-01
-4.03759778e-01 8.88076186e-01 1.75078452e-01 -4.30991322e-01
3.89453650e-01 1.46751907e-02 -1.18752971e-01 2.22040236e-01
2.57096231e-01 2.96818495e-01 5.63938394e-02 -5.30878186e-01
-4.29205298e-01 1.04108059e+00 -3.08125556e-01 3.31910014e-01
1.51229680e+00 -1.05930813e-01 -3.59639019e-01 2.51249075e-01
1.80915439e+00 -3.49171638e-01 -8.94763947e-01 -4.04757112e-01
-4.48794693e-01 -6.47166610e-01 2.08399624e-01 -5.50756693e-01
-1.43461573e+00 1.21973395e+00 9.03951168e-01 1.68997258e-01
1.21034575e+00 -3.72923762e-02 1.10376811e+00 4.95755583e-01
2.56619453e-01 -1.16808796e+00 3.05081844e-01 4.20249999e-01
8.95157695e-01 -1.49924994e+00 3.39638054e-01 -2.48730361e-01
-2.10971788e-01 1.01496816e+00 6.56585038e-01 -4.35945094e-01
9.60168719e-01 -4.52553242e-01 -1.13347899e-02 -3.09950948e-01
-3.37562054e-01 -1.30562574e-01 6.01659060e-01 2.07905412e-01
5.04536182e-02 -2.30781391e-01 -2.19713420e-01 3.37239206e-01
4.41216826e-01 -2.47381076e-01 -2.10845858e-01 6.92326188e-01
-2.33010486e-01 -8.97061229e-01 -2.44900271e-01 6.74265265e-01
-5.44501066e-01 6.55631870e-02 -2.79546082e-01 8.72889996e-01
3.45592722e-02 6.69195175e-01 1.34195223e-01 -3.62641692e-01
3.39712203e-01 -2.15408430e-01 5.18092036e-01 -2.94932604e-01
-3.96383941e-01 3.24819148e-01 -5.10962069e-01 -2.89092690e-01
-8.67936194e-01 -3.48673582e-01 -9.09997225e-01 -1.32498890e-01
-9.27361071e-01 2.22163409e-01 5.22404015e-01 8.96229446e-01
5.42734623e-01 2.34907180e-01 1.08694565e+00 -1.11126113e+00
-4.45979893e-01 -7.34733105e-01 -7.60900319e-01 2.46465325e-01
2.86534667e-01 -8.35700452e-01 -4.38127518e-01 -4.54795718e-01] | [8.18588638305664, -3.880969762802124] |
44359e65-802e-4b07-ade7-bb40124a7d5f | maven-multi-agent-variational-exploration | 1910.07483 | null | https://arxiv.org/abs/1910.07483v2 | https://arxiv.org/pdf/1910.07483v2.pdf | MAVEN: Multi-Agent Variational Exploration | Centralised training with decentralised execution is an important setting for cooperative deep multi-agent reinforcement learning due to communication constraints during execution and computational tractability in training. In this paper, we analyse value-based methods that are known to have superior performance in complex environments [43]. We specifically focus on QMIX [40], the current state-of-the-art in this domain. We show that the representational constraints on the joint action-values introduced by QMIX and similar methods lead to provably poor exploration and suboptimality. Furthermore, we propose a novel approach called MAVEN that hybridises value and policy-based methods by introducing a latent space for hierarchical control. The value-based agents condition their behaviour on the shared latent variable controlled by a hierarchical policy. This allows MAVEN to achieve committed, temporally extended exploration, which is key to solving complex multi-agent tasks. Our experimental results show that MAVEN achieves significant performance improvements on the challenging SMAC domain [43]. | ['Shimon Whiteson', 'Tabish Rashid', 'Mikayel Samvelyan', 'Anuj Mahajan'] | 2019-10-16 | maven-multi-agent-variational-exploration-1 | http://papers.nips.cc/paper/8978-maven-multi-agent-variational-exploration | http://papers.nips.cc/paper/8978-maven-multi-agent-variational-exploration.pdf | neurips-2019-12 | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [-9.02584195e-02 2.32189104e-01 -4.30467308e-01 1.68401629e-01
-8.53753507e-01 -3.94583225e-01 8.37488472e-01 3.74320149e-02
-8.19080293e-01 1.31746018e+00 6.39724254e-04 -1.61967263e-01
-5.75093746e-01 -4.90693003e-01 -6.96685493e-01 -1.15185308e+00
-8.29415798e-01 8.40670586e-01 7.20083341e-02 -3.79651278e-01
1.47866711e-01 2.28544310e-01 -1.36333966e+00 -4.50021997e-02
7.83532560e-01 7.84824967e-01 3.91569614e-01 8.93935323e-01
4.36838090e-01 1.12200296e+00 -8.68766844e-01 2.50234008e-01
2.91658849e-01 -5.24059176e-01 -1.05302680e+00 1.75346538e-01
-2.00384632e-01 -3.98779392e-01 6.91362545e-02 6.88587666e-01
6.94112718e-01 5.88784575e-01 3.66440535e-01 -1.59937882e+00
1.10232711e-01 9.29222047e-01 -4.22732890e-01 1.63515255e-01
-9.58226621e-02 5.31715155e-01 1.03969443e+00 1.33954003e-01
7.47666776e-01 1.33422041e+00 2.51961172e-01 8.76841486e-01
-1.40096259e+00 -2.45734259e-01 5.81568778e-01 3.13920677e-01
-8.42604995e-01 -3.26704562e-01 4.76372242e-01 -8.65314752e-02
1.29023349e+00 2.08636865e-01 9.83161747e-01 1.14267802e+00
4.09701645e-01 1.19221318e+00 1.43211985e+00 -5.16525209e-01
9.79300916e-01 -2.81345755e-01 -6.52518153e-01 6.91575289e-01
-2.98841119e-01 5.70566773e-01 -6.28521681e-01 -3.35475773e-01
8.82622421e-01 -3.34925443e-01 3.78557108e-02 -8.67164671e-01
-1.29350746e+00 1.22631574e+00 4.09076810e-01 8.32606852e-02
-7.85814583e-01 8.29214096e-01 7.23789811e-01 6.55579984e-01
1.60438269e-01 9.03181314e-01 -6.98522568e-01 -5.12533307e-01
-7.29996920e-01 6.66338027e-01 8.59364152e-01 6.96678817e-01
3.69555026e-01 2.47176707e-01 -3.25109184e-01 4.97030497e-01
4.16809618e-01 9.53618586e-02 4.93329257e-01 -1.71321750e+00
3.86498332e-01 4.56586964e-02 3.32098484e-01 -2.78059959e-01
-7.00545490e-01 -6.53494298e-01 -5.95707536e-01 7.57251978e-01
1.67337373e-01 -6.25990510e-01 -6.20051265e-01 2.06057358e+00
5.92081368e-01 1.66092440e-01 5.86866558e-01 9.24239218e-01
-5.96761238e-03 5.18769205e-01 1.29898638e-01 -5.31213224e-01
9.71500516e-01 -1.43088710e+00 -6.56707942e-01 -2.17717767e-01
7.40949273e-01 -3.38951461e-02 8.03350687e-01 5.50578177e-01
-1.36928022e+00 7.62331625e-03 -9.81208324e-01 6.58398151e-01
-4.96691950e-02 -4.83950168e-01 9.14614320e-01 2.21890792e-01
-1.32063425e+00 8.75462770e-01 -1.31195068e+00 -2.66921967e-02
5.32670140e-01 6.60564244e-01 9.00386050e-02 3.00799727e-01
-1.18825948e+00 1.13413930e+00 5.63358486e-01 -1.31375894e-01
-1.67051864e+00 -3.62643659e-01 -5.67188203e-01 5.81106022e-02
9.82132137e-01 -7.08026648e-01 1.90073895e+00 -9.14720953e-01
-2.18980360e+00 1.89206660e-01 4.08110887e-01 -8.33125472e-01
7.69446969e-01 -1.26805782e-01 3.25918227e-01 2.11723775e-01
1.57225758e-01 7.29167283e-01 8.32366347e-01 -1.32935929e+00
-7.73360729e-01 -2.87161637e-02 3.40907902e-01 7.66817927e-01
-1.50648385e-01 -2.24536672e-01 5.82340499e-03 -3.35848927e-01
-5.09206653e-01 -1.06290793e+00 -9.70183849e-01 -3.53115410e-01
4.36501168e-02 -5.55467427e-01 5.75459301e-01 -1.13059953e-01
8.10148478e-01 -1.70642412e+00 8.57988417e-01 1.73018426e-01
2.15138480e-01 6.84015602e-02 -2.63736576e-01 8.66672993e-01
4.78217691e-01 -9.57723781e-02 -1.88317820e-01 -5.54072738e-01
6.46307766e-01 7.55583048e-01 -1.59814656e-02 5.74366868e-01
-6.95806593e-02 1.06224608e+00 -1.14434409e+00 -4.73345011e-01
1.85030341e-01 1.19159795e-01 -6.90786362e-01 3.08617592e-01
-9.74533856e-01 6.88731551e-01 -6.90247893e-01 2.45212793e-01
8.46451595e-02 -2.44321883e-01 6.69351697e-01 7.21779227e-01
-1.97571307e-01 8.25430527e-02 -1.02283037e+00 1.98834908e+00
-5.75955510e-01 3.28739047e-01 6.34562135e-01 -1.09610987e+00
3.36807549e-01 5.32392383e-01 7.48369813e-01 -8.66319656e-01
1.75894514e-01 3.36432271e-02 1.14250459e-01 -2.93319345e-01
3.27386886e-01 -3.76248732e-02 -1.99770972e-01 6.72344029e-01
-1.17964044e-01 -3.78733844e-01 3.06237131e-01 2.28662059e-01
1.33047664e+00 4.19330746e-01 3.74357551e-01 -4.44494754e-01
2.71148384e-01 1.89271525e-01 5.73400438e-01 1.12880194e+00
-3.86589199e-01 -3.45353752e-01 7.69457817e-01 -2.30596036e-01
-1.05810452e+00 -6.94648266e-01 2.73239434e-01 1.58408082e+00
-1.05272671e-02 -3.43154609e-01 -7.29731739e-01 -5.98095477e-01
-7.12536201e-02 5.84945321e-01 -8.23710620e-01 1.84944458e-02
-8.63203585e-01 -6.01073921e-01 2.04868764e-01 4.23140436e-01
4.70492601e-01 -1.62773693e+00 -1.69428849e+00 6.23007774e-01
2.11683631e-01 -6.71104610e-01 -2.06481144e-01 7.72327662e-01
-7.26510763e-01 -8.36443007e-01 -7.40961373e-01 -4.53982294e-01
2.27702409e-01 -3.67644012e-01 1.27511191e+00 7.69938827e-02
-7.31988251e-02 6.56009614e-01 -3.82550746e-01 -3.13043922e-01
-5.13385534e-01 3.96892667e-01 -6.09744228e-02 -5.44918776e-01
-4.44566488e-01 -5.72087169e-01 -6.18705451e-01 1.00212879e-01
-8.95892024e-01 1.04134358e-01 5.94784558e-01 1.23968232e+00
4.24524069e-01 9.68725532e-02 6.75822198e-01 -5.37079215e-01
9.28228676e-01 -4.36242193e-01 -9.64747965e-01 5.31098694e-02
-7.13253796e-01 4.03072923e-01 6.23972237e-01 -6.24692082e-01
-9.78599310e-01 -6.87429905e-02 1.92822307e-01 -3.25382233e-01
-3.78177948e-02 5.80930948e-01 3.01810235e-01 3.42824161e-02
4.56704557e-01 1.74197420e-01 4.25486147e-01 -3.78686413e-02
4.20839816e-01 2.29466125e-01 2.46772990e-01 -1.08926666e+00
2.18064085e-01 3.54666561e-01 3.35936069e-01 -4.80469048e-01
-5.22061169e-01 -1.91953182e-01 -2.68073440e-01 -2.80042171e-01
7.24119782e-01 -8.33760619e-01 -1.40113699e+00 2.04000905e-01
-9.53241050e-01 -1.26800835e+00 -5.00751495e-01 3.01539421e-01
-1.44684219e+00 4.97166663e-02 -7.10485280e-01 -1.03939521e+00
-2.70077348e-01 -1.28377664e+00 9.93084848e-01 1.80532023e-01
1.07991288e-03 -1.20215487e+00 4.88139421e-01 -9.03428439e-03
6.67573631e-01 4.74894315e-01 6.55003488e-01 -4.00523543e-01
-7.06505656e-01 5.98878801e-01 4.49192166e-01 -1.77176148e-01
-4.26939547e-01 -4.61513996e-01 -5.57581007e-01 -9.47479188e-01
-1.56904981e-01 -8.70392680e-01 7.29559779e-01 5.46328306e-01
7.72531092e-01 -5.83349049e-01 -2.61771351e-01 3.69244993e-01
1.42599654e+00 2.79832363e-01 1.62308022e-01 1.04236591e+00
1.39801368e-01 5.55410504e-01 9.66105461e-01 1.04515159e+00
3.47117811e-01 8.99501324e-01 1.05454278e+00 2.23207846e-02
4.52400476e-01 1.27788588e-01 4.57661688e-01 3.93299073e-01
-4.19292264e-02 -1.47000864e-01 -8.12999308e-01 5.91157794e-01
-2.68544292e+00 -8.48529339e-01 3.67029816e-01 1.71917367e+00
1.16493690e+00 -9.80953574e-02 4.38818336e-01 -2.48258740e-01
2.47046813e-01 2.04234436e-01 -9.11407471e-01 -5.46346664e-01
1.75693572e-01 3.17654461e-01 5.61627388e-01 6.03388131e-01
-1.00111961e+00 1.11361635e+00 6.46784496e+00 9.00240779e-01
-7.53477335e-01 4.44789439e-01 1.67297512e-01 -4.98129278e-01
7.54669532e-02 -1.80699751e-01 -4.95713025e-01 3.23214322e-01
1.14919066e+00 -8.54690000e-02 8.98899198e-01 8.32247317e-01
3.65629435e-01 -3.47485602e-01 -1.08656800e+00 5.12534022e-01
-3.24493557e-01 -1.26412618e+00 -8.20601523e-01 3.63380909e-01
1.06840670e+00 3.47668678e-01 1.46056905e-01 5.66666842e-01
1.15932596e+00 -1.14696240e+00 8.33485365e-01 2.09180683e-01
2.85700023e-01 -1.15793633e+00 6.03450537e-01 6.47989273e-01
-8.20980549e-01 -4.25732523e-01 -1.66886598e-01 -2.74809182e-01
1.54873252e-01 -1.81368530e-01 -7.77932167e-01 3.82260829e-01
5.69304287e-01 3.92326117e-01 -9.00970623e-02 7.68931448e-01
-2.92615563e-01 3.58635992e-01 -3.15098822e-01 -3.84223491e-01
9.66240227e-01 -7.88204297e-02 4.82773662e-01 7.89048016e-01
-2.24086717e-01 -1.55770913e-01 6.65504694e-01 6.93845570e-01
3.10481638e-01 -2.20548436e-01 -5.17922103e-01 7.94920139e-03
3.06976646e-01 1.06765079e+00 -6.68827415e-01 -3.45351130e-01
9.07449797e-02 7.91895926e-01 7.57299721e-01 3.71132493e-01
-7.82575846e-01 1.15485333e-01 6.97212398e-01 -7.00333774e-01
5.74135482e-01 -4.75736260e-01 1.07581168e-01 -9.70400870e-01
-2.32372731e-01 -1.22010136e+00 3.78268778e-01 -1.19142175e-01
-9.78532791e-01 5.35975516e-01 1.62514597e-01 -8.02280664e-01
-9.86660242e-01 -2.72067487e-01 -4.25405383e-01 5.06370902e-01
-1.66850865e+00 -8.32814455e-01 2.00826377e-01 5.67200065e-01
6.66994750e-01 -3.92691731e-01 9.85282958e-01 -3.45642120e-01
-5.48847556e-01 2.22981751e-01 5.70361316e-01 -4.93060768e-01
2.33907297e-01 -1.64233160e+00 1.57454863e-01 3.51689786e-01
-3.14925969e-01 1.37258291e-01 9.41298544e-01 -4.40836638e-01
-1.57651603e+00 -8.23438525e-01 -2.54712813e-02 -1.60968423e-01
7.70525813e-01 -2.58496612e-01 -4.61906731e-01 5.92597842e-01
7.95347333e-01 -1.95526809e-01 3.24594736e-01 1.84673712e-01
3.01022887e-01 2.41421580e-01 -8.98805380e-01 6.37379587e-01
7.29318082e-01 -8.08461830e-02 -2.89017111e-01 4.39353913e-01
7.10555494e-01 -5.12882948e-01 -9.25136149e-01 1.11278675e-01
2.70926058e-01 -9.66113448e-01 7.35284269e-01 -7.74745703e-01
2.62856275e-01 4.72078286e-02 -6.17748573e-02 -1.89132631e+00
-1.32139295e-01 -1.11982596e+00 -5.55966198e-01 5.67040503e-01
1.15263440e-01 -7.59658813e-01 1.10292864e+00 3.26332778e-01
-7.46335238e-02 -1.01642990e+00 -1.17996848e+00 -1.02562129e+00
4.75179464e-01 2.71613756e-03 5.06316364e-01 7.98005342e-01
2.63335675e-01 1.89905077e-01 -5.81853747e-01 -2.86671426e-02
8.95828843e-01 3.12026516e-02 6.75473630e-01 -7.45959044e-01
-8.79092395e-01 -5.70553243e-01 1.57942370e-01 -9.31313097e-01
5.56584060e-01 -5.83046854e-01 3.67604315e-01 -1.42095160e+00
2.94718370e-02 -5.67013681e-01 -2.06791148e-01 5.51542401e-01
2.39117190e-01 -3.59736115e-01 5.16925693e-01 1.57331124e-01
-1.31034684e+00 1.05642641e+00 1.50172162e+00 2.45742989e-03
-4.55485135e-01 -1.45201042e-01 -2.59545892e-01 3.40381205e-01
1.14750254e+00 -3.42915446e-01 -5.66293776e-01 -3.59368503e-01
2.89860904e-01 6.46372259e-01 3.00426513e-01 -7.41171062e-01
4.01868999e-01 -7.14536309e-01 -8.43485147e-02 -1.28739342e-01
5.64223588e-01 -6.98845267e-01 -1.14029773e-01 9.91184473e-01
-8.13257575e-01 2.94475347e-01 1.00112945e-01 6.41004562e-01
-3.29235569e-02 -2.77629524e-01 8.18257749e-01 -4.89587009e-01
-6.74187481e-01 1.39584765e-01 -6.45033002e-01 2.00045258e-01
1.36336839e+00 3.63689095e-01 -2.49235556e-01 -5.04619658e-01
-8.61974716e-01 9.86424804e-01 2.20785469e-01 1.38685182e-02
4.50702488e-01 -1.15061986e+00 -7.12907374e-01 -1.25917926e-01
-3.52219313e-01 1.73686981e-01 3.88331115e-02 8.61582637e-01
-3.10921133e-01 4.16857660e-01 -4.67416137e-01 -5.83399832e-01
-1.22010088e+00 4.44511116e-01 5.22630215e-01 -7.81501293e-01
-6.34888828e-01 6.35887325e-01 -1.94224596e-01 -5.04404604e-01
4.63221639e-01 2.41334587e-02 -5.86601011e-02 -1.29952893e-01
1.61398783e-01 4.12638217e-01 -3.45418632e-01 4.67258878e-02
-1.75783649e-01 4.11901325e-02 -1.99148536e-01 -8.29778314e-01
1.57178319e+00 -7.21860006e-02 5.86883016e-02 1.85443655e-01
7.80743003e-01 -7.77459264e-01 -1.89099908e+00 -2.99600691e-01
5.80338202e-02 -1.67702019e-01 3.88130844e-01 -9.00706053e-01
-7.95859933e-01 5.24625897e-01 4.50410515e-01 3.71141970e-01
8.47421706e-01 2.55166218e-02 2.95391828e-01 6.53606534e-01
6.47472322e-01 -1.56529224e+00 2.58095652e-01 7.17899203e-01
9.08576012e-01 -1.16675282e+00 1.79871563e-02 4.02018815e-01
-9.39849675e-01 7.95235515e-01 8.13514173e-01 -1.01411194e-01
1.34146571e-01 3.31541002e-01 -9.35342088e-02 -2.99770951e-01
-1.68147123e+00 -4.42382097e-01 -4.46746677e-01 6.70334458e-01
-1.44097775e-01 2.17271566e-01 -2.30371386e-01 -8.07732940e-02
-1.30468503e-01 -2.16951251e-01 4.67649788e-01 1.46329880e+00
-5.13291717e-01 -1.24994862e+00 -3.33704770e-01 1.65844977e-01
-3.25435370e-01 2.10507378e-01 -1.22656254e-02 9.04645324e-01
-2.11005822e-01 9.06408012e-01 -1.62967909e-02 2.58221656e-01
-1.03140697e-01 -2.34278426e-01 7.90554404e-01 -4.21658814e-01
-8.78222764e-01 3.45231324e-01 1.41073555e-01 -8.55548263e-01
-5.21164834e-01 -8.68816555e-01 -1.47352147e+00 -1.33981824e-01
-1.58132780e-02 5.60314655e-01 7.53469348e-01 1.02938831e+00
2.64702827e-01 9.46791708e-01 6.59507453e-01 -1.08532953e+00
-1.31386817e+00 -8.31619084e-01 -5.45736015e-01 1.08750975e-02
6.85842633e-01 -7.70402431e-01 -2.36597762e-01 -4.83558536e-01] | [3.8386924266815186, 1.874169945716858] |
904f342c-fca5-42f0-bf12-4e8abe941996 | boundary-smoothing-for-named-entity-1 | 2204.12031 | null | https://arxiv.org/abs/2204.12031v1 | https://arxiv.org/pdf/2204.12031v1.pdf | Boundary Smoothing for Named Entity Recognition | Neural named entity recognition (NER) models may easily encounter the over-confidence issue, which degrades the performance and calibration. Inspired by label smoothing and driven by the ambiguity of boundary annotation in NER engineering, we propose boundary smoothing as a regularization technique for span-based neural NER models. It re-assigns entity probabilities from annotated spans to the surrounding ones. Built on a simple but strong baseline, our model achieves results better than or competitive with previous state-of-the-art systems on eight well-known NER benchmarks. Further empirical analysis suggests that boundary smoothing effectively mitigates over-confidence, improves model calibration, and brings flatter neural minima and more smoothed loss landscapes. | ['Jinpeng Li', 'Enwei Zhu'] | 2022-04-26 | null | https://aclanthology.org/2022.acl-long.490 | https://aclanthology.org/2022.acl-long.490.pdf | acl-2022-5 | ['nested-named-entity-recognition', 'chinese-named-entity-recognition'] | ['natural-language-processing', 'natural-language-processing'] | [-1.38906762e-01 3.50373030e-01 -2.54869640e-01 -6.43190920e-01
-8.73596311e-01 -5.63033521e-01 2.01450154e-01 2.84089327e-01
-9.92730319e-01 8.30747724e-01 5.28831542e-01 -1.09836847e-01
1.36574477e-01 -6.25280797e-01 -6.88005805e-01 -2.07762092e-01
-6.92218542e-02 1.69635743e-01 2.66233295e-01 1.61346525e-01
-6.33419007e-02 4.25542235e-01 -9.72763300e-01 5.54865301e-02
1.11518919e+00 8.79417956e-01 -3.33904624e-01 2.83609658e-01
-5.82126260e-01 8.27906668e-01 -7.15197623e-01 -9.40935850e-01
2.37766765e-02 1.26111857e-03 -8.37257504e-01 -5.60620248e-01
7.86665678e-01 9.65138376e-02 -5.15494645e-01 1.19308794e+00
5.25490284e-01 4.73580867e-01 6.64781094e-01 -7.19382942e-01
-1.18350708e+00 1.08131838e+00 -4.44479108e-01 5.15233427e-02
-3.07633638e-01 -3.73314917e-01 1.03340614e+00 -9.44886029e-01
5.10563016e-01 1.07883275e+00 1.54069853e+00 8.46309185e-01
-1.38990808e+00 -6.15655422e-01 3.20529997e-01 -2.14427918e-01
-1.71912432e+00 -4.35859770e-01 3.65029037e-01 -6.78706616e-02
1.01293051e+00 1.14243068e-01 2.70246225e-03 1.20934939e+00
9.95193198e-02 7.45139301e-01 5.70899725e-01 -4.04868186e-01
3.41483831e-01 -1.02297969e-01 6.35391831e-01 4.60434616e-01
4.16738778e-01 1.79544434e-01 -2.88680524e-01 -2.46182784e-01
5.86143970e-01 -2.23569229e-01 -1.73991144e-01 -6.58530593e-02
-8.41616750e-01 6.31657064e-01 5.86388886e-01 2.92699069e-01
-3.73248160e-01 3.66090685e-01 4.46242988e-01 -2.24875316e-01
5.48260212e-01 5.75989127e-01 -7.79677868e-01 -5.54622710e-02
-1.20421886e+00 -2.07034945e-01 9.37667668e-01 1.04594743e+00
6.16833627e-01 2.54653901e-01 -6.64289415e-01 1.08468580e+00
2.67610788e-01 1.97440878e-01 1.38526738e-01 -8.79272938e-01
5.00087857e-01 5.59152067e-01 1.42795503e-01 -5.65156937e-01
-5.88371396e-01 -5.76037228e-01 -9.97438431e-01 -1.86262041e-01
5.02241790e-01 -5.84327340e-01 -1.30242443e+00 2.09476972e+00
1.40703484e-01 4.27529693e-01 -4.16135713e-02 5.51607966e-01
7.23584652e-01 4.54261690e-01 9.31890368e-01 1.00525044e-01
1.40989876e+00 -9.96116877e-01 -7.13986635e-01 -5.45059741e-01
5.21929860e-01 -2.42720887e-01 1.03947365e+00 -3.23400684e-02
-7.70280182e-01 -4.69980061e-01 -9.24066782e-01 -2.73638844e-01
-5.88186741e-01 3.72302264e-01 6.22718573e-01 1.01285899e+00
-8.54389131e-01 8.41668487e-01 -7.46228635e-01 -2.83218086e-01
4.73363876e-01 1.47238165e-01 -4.21671093e-01 1.71917245e-01
-1.49487436e+00 1.04001844e+00 7.66956031e-01 2.09753677e-01
-3.53024215e-01 -9.58811700e-01 -9.96800363e-01 3.15514237e-01
2.86411256e-01 -5.19085407e-01 1.22259331e+00 -3.79362047e-01
-1.26185644e+00 6.42835081e-01 -1.29920855e-01 -6.15598917e-01
3.39686275e-01 -6.36233389e-01 -7.85505176e-01 -4.52331662e-01
-1.85721651e-01 9.51554239e-01 2.79506117e-01 -1.21630061e+00
-3.50088835e-01 -7.38469735e-02 -2.18080178e-01 -9.11714584e-02
-6.83867753e-01 3.06794792e-02 -5.96216798e-01 -9.08984780e-01
-2.36023739e-01 -6.10368371e-01 -6.50662899e-01 -4.55569267e-01
-6.56367004e-01 -4.34204131e-01 1.89147413e-01 -6.95148110e-01
1.90459037e+00 -2.08846140e+00 -3.77732456e-01 1.45741552e-01
1.48259252e-01 2.66067624e-01 -2.19244838e-01 1.55284271e-01
-8.44091922e-02 6.08095884e-01 -3.66445601e-01 -5.34763992e-01
3.18542868e-01 1.28528371e-01 -2.45652288e-01 1.48099378e-01
4.25244451e-01 8.56534660e-01 -7.62206018e-01 -7.48583972e-01
-2.43766561e-01 7.30489373e-01 -2.49445692e-01 1.12253301e-01
-1.00202300e-01 1.50400504e-01 -7.65979886e-02 4.87048328e-01
6.27700925e-01 -2.63501614e-01 2.35919785e-02 -2.65156478e-01
-8.78189728e-02 3.18996966e-01 -1.24893856e+00 1.71494985e+00
-2.89235294e-01 5.58941901e-01 -2.73206472e-01 -4.32010949e-01
1.14251959e+00 2.05538630e-01 9.25618485e-02 -3.57590348e-01
2.11745754e-01 -1.20715953e-01 -5.34905851e-01 -1.62500918e-01
9.01795566e-01 4.17232588e-02 -3.85566771e-01 1.71239991e-02
3.66242796e-01 4.89303708e-01 2.14698508e-01 -1.01914510e-01
1.15244186e+00 1.00563802e-01 1.81460500e-01 -2.38864064e-01
8.21600407e-02 -3.28336328e-01 1.13679862e+00 9.54621732e-01
-3.77601057e-01 5.74366629e-01 2.80885577e-01 -4.28203076e-01
-1.08813179e+00 -1.28800869e+00 -3.37551296e-01 1.50779343e+00
-1.05695955e-01 -4.47497487e-01 -1.18032432e+00 -9.80412006e-01
-9.45642889e-02 1.13970399e+00 -7.73923218e-01 -1.74331203e-01
-6.27075732e-01 -7.26290524e-01 1.44587421e+00 1.10178041e+00
3.97857219e-01 -1.04050672e+00 -1.54827274e-02 4.12798226e-01
-9.36784893e-02 -1.04572833e+00 -9.11497176e-01 4.91907567e-01
-9.91659939e-01 -4.93226886e-01 -8.73666942e-01 -8.70713651e-01
5.14835238e-01 -4.24921662e-01 1.53146279e+00 -2.07393050e-01
8.71269256e-02 -4.85199690e-02 -4.63153757e-02 -5.67467630e-01
-2.53010809e-01 4.19741869e-01 1.02035172e-01 -4.20957804e-01
7.25561619e-01 -2.73151308e-01 -3.51642102e-01 3.06800783e-01
-8.75019968e-01 -5.64756930e-01 6.32690012e-01 8.84315133e-01
6.62496209e-01 -6.61171004e-02 8.43407154e-01 -1.00807822e+00
6.79618597e-01 -3.76734018e-01 -4.01986182e-01 7.70500362e-01
-9.58569705e-01 4.75374460e-01 5.95520198e-01 -5.87515533e-01
-1.39459503e+00 1.65307671e-01 -1.85624436e-01 -4.12299156e-01
-4.02439356e-01 4.28282022e-01 -2.93808550e-01 2.52090365e-01
1.01714456e+00 -3.10883582e-01 -7.79953778e-01 -8.08732748e-01
7.09437668e-01 5.92916489e-01 9.95570660e-01 -6.69475079e-01
4.03478652e-01 5.38005978e-02 -4.14642870e-01 -5.61878741e-01
-1.19246209e+00 -3.75345618e-01 -6.59042001e-01 5.33085167e-02
8.53098094e-01 -9.01234329e-01 -3.98835957e-01 3.36416513e-01
-1.35067320e+00 -2.18492001e-01 -3.82359236e-01 3.61588925e-01
-6.56462014e-02 4.89682823e-01 -9.42505777e-01 -9.60583866e-01
-6.42910480e-01 -3.78642589e-01 5.44476807e-01 7.93160677e-01
-3.89123112e-01 -1.12891591e+00 2.86229372e-01 -8.86888951e-02
5.45741796e-01 8.68443325e-02 7.20157087e-01 -1.04698026e+00
9.78687704e-02 -2.94276893e-01 -3.76051575e-01 4.01014090e-01
-3.50483298e-01 -3.94027717e-02 -1.21449471e+00 9.72708985e-02
-4.07003552e-01 -1.19090760e-02 1.11335468e+00 3.96590501e-01
1.08982420e+00 -1.79855600e-01 -3.58891189e-01 5.25333941e-01
1.35914528e+00 -1.57350793e-01 8.29713345e-01 4.05802995e-01
6.75275803e-01 5.37382603e-01 2.73182631e-01 3.33886415e-01
4.97501493e-01 8.42756629e-02 1.39110908e-01 -3.39131683e-01
2.80552711e-02 -6.04930341e-01 1.59829706e-01 5.99761665e-01
1.01735488e-01 -3.95459950e-01 -1.06006360e+00 7.41771519e-01
-1.79842019e+00 -8.18224728e-01 -4.73950803e-02 1.99679875e+00
1.26341712e+00 3.74794573e-01 -4.36870754e-02 -4.13770735e-01
1.24540889e+00 1.49235696e-01 -7.31115222e-01 -3.20975542e-01
-2.56733298e-01 2.09148243e-01 9.66341376e-01 1.26065731e-01
-1.51200092e+00 1.17698920e+00 7.53023911e+00 9.77717876e-01
-5.00848413e-01 1.92500323e-01 7.10682690e-01 3.85642767e-01
-2.12795839e-01 -9.21290070e-02 -1.30483603e+00 2.46833608e-01
1.26058853e+00 -6.24413192e-02 1.41855076e-01 1.18193591e+00
-2.35017627e-01 3.62062722e-01 -1.16797078e+00 6.24105096e-01
-1.31119922e-01 -1.32554853e+00 -8.42271447e-02 -3.01104933e-01
8.75598431e-01 4.51821625e-01 -3.23520273e-01 5.75493336e-01
8.94794226e-01 -1.18146157e+00 8.12929153e-01 9.61603522e-01
9.12142456e-01 -8.56997311e-01 1.04777241e+00 2.13594943e-01
-1.46774602e+00 -1.46050155e-01 -5.97726941e-01 3.59663069e-01
3.55190396e-01 7.70676792e-01 -6.02496088e-01 3.26002538e-01
7.61197627e-01 2.19232351e-01 -4.92071867e-01 1.38710546e+00
-3.72813016e-01 8.76910686e-01 -4.21438307e-01 -7.83287063e-02
1.50220320e-02 1.66660115e-01 2.44486362e-01 1.90533376e+00
8.86516646e-02 -7.93127269e-02 -2.25943364e-02 1.06677198e+00
-6.14649415e-01 8.31575096e-02 -3.01287174e-01 -1.31339356e-01
1.14414644e+00 1.15214837e+00 -6.19333863e-01 -2.15904415e-01
-2.88753331e-01 8.59181881e-01 8.87066960e-01 5.04941583e-01
-1.10340679e+00 -1.01094949e+00 4.82999533e-01 -3.60909343e-01
4.83902872e-01 -2.05813289e-01 -8.80961299e-01 -1.06748688e+00
-2.49467462e-01 -2.09812984e-01 7.45429337e-01 -3.90534163e-01
-1.86010134e+00 9.25743103e-01 -2.17363209e-01 -8.25880647e-01
2.31482863e-01 -5.40735066e-01 -8.13271701e-01 6.67802811e-01
-1.36444271e+00 -9.80576038e-01 -3.30356844e-02 9.34701040e-02
2.21741155e-01 1.40932724e-01 1.02662206e+00 4.23016876e-01
-7.80400753e-01 1.21186233e+00 1.98993921e-01 8.08025956e-01
9.12467062e-01 -1.40884471e+00 9.57260013e-01 9.77781832e-01
2.94333279e-01 8.93750250e-01 3.75353992e-01 -7.51493454e-01
-5.47817409e-01 -1.33476722e+00 1.09466755e+00 -7.54623115e-01
6.38624191e-01 -1.65661082e-01 -1.15133262e+00 6.73653662e-01
9.19774249e-02 -6.68085888e-02 1.02441096e+00 4.56728280e-01
-8.03462505e-01 7.02155009e-02 -1.07289219e+00 6.06775761e-01
1.16088438e+00 -4.82223183e-01 -9.23140049e-01 -2.88170099e-01
9.77754712e-01 -2.62899637e-01 -1.28864312e+00 5.55850387e-01
4.58094180e-01 -6.25575900e-01 1.04401278e+00 -8.35407794e-01
-1.81501046e-01 -1.22322462e-01 -5.56787401e-02 -1.17101383e+00
-6.38970554e-01 -4.85019356e-01 -2.85366237e-01 1.89304399e+00
8.21824789e-01 -2.21846014e-01 8.19991052e-01 1.16955423e+00
-2.52726823e-01 -4.62284476e-01 -9.56299484e-01 -1.09899545e+00
3.57579976e-01 -5.89941025e-01 6.60185635e-01 1.08480561e+00
2.48716306e-02 3.64766240e-01 -3.61474097e-01 2.92496979e-01
7.88930953e-01 -6.41144574e-01 9.34609100e-02 -1.49394453e+00
1.32775083e-01 -4.39688772e-01 -4.50629853e-02 -1.10316873e+00
4.67362344e-01 -6.40580118e-01 5.08227289e-01 -1.62234426e+00
-6.46299347e-02 -5.72765231e-01 -8.06097090e-01 9.45026159e-01
-4.44307119e-01 2.08523467e-01 1.15169227e-01 5.75897982e-03
-9.40283597e-01 4.41221774e-01 5.43795407e-01 3.72357666e-02
-3.04428220e-01 -2.07300529e-01 -7.55053222e-01 9.21194553e-01
7.30365217e-01 -8.60949099e-01 1.19943082e-01 -7.15911806e-01
3.74304652e-01 -3.22364360e-01 -2.48485187e-04 -1.07160163e+00
5.21941602e-01 -1.07968248e-01 5.18802047e-01 -5.99303722e-01
-1.39013067e-01 -6.83764517e-01 1.35011807e-01 2.54973024e-02
-6.34478211e-01 -9.13649201e-02 2.46066093e-01 8.68843675e-01
1.74345225e-02 -6.14370406e-01 7.46611178e-01 4.82615037e-03
-8.87145460e-01 1.87682793e-01 -1.86068192e-01 5.21956861e-01
5.17729700e-01 -1.25911623e-01 -4.86362696e-01 1.00044459e-01
-8.29561114e-01 3.00741911e-01 3.49936604e-01 3.83759737e-01
2.36465231e-01 -1.40563905e+00 -5.74752510e-01 -1.75291449e-01
1.23835623e-01 7.24436492e-02 2.75544345e-01 3.90895844e-01
-1.63943052e-01 1.52162597e-01 1.68210343e-01 -1.28375396e-01
-7.44296551e-01 3.88372689e-01 4.18171674e-01 -5.46842337e-01
-5.08776009e-01 1.08778954e+00 -1.70690700e-01 -6.95893466e-01
8.22463751e-01 -4.88939226e-01 -3.81085008e-01 -7.19187856e-02
5.64906120e-01 6.61018372e-01 8.15301575e-03 -3.36746752e-01
-3.94967258e-01 2.89653838e-01 -8.58358219e-02 8.61993581e-02
1.17738700e+00 -1.22313388e-01 2.04970494e-01 4.22527641e-01
8.66580844e-01 1.58842444e-01 -1.42030156e+00 -3.85206908e-01
8.77010822e-01 1.93500832e-01 7.11869746e-02 -1.00497937e+00
-7.05894768e-01 6.76585674e-01 4.44049597e-01 2.09264949e-01
8.63313675e-01 -6.37107193e-02 1.08918512e+00 7.38228858e-01
1.01416640e-01 -1.50298500e+00 -4.39423651e-01 1.15822363e+00
3.73893321e-01 -1.00168324e+00 -3.30631375e-01 -1.57079086e-01
-8.51975918e-01 9.97288406e-01 7.51651287e-01 -1.61812901e-01
6.50450826e-01 6.23716116e-01 2.03531414e-01 2.05506101e-01
-5.72012961e-01 -8.86452794e-02 4.79324579e-01 7.06555188e-01
6.45704985e-01 -5.39659970e-02 -2.51736075e-01 1.41159713e+00
4.04519476e-02 7.62308366e-04 1.31308168e-01 5.78949451e-01
-5.31275749e-01 -9.27519143e-01 -2.28950724e-01 3.19844395e-01
-8.22113812e-01 -4.66998369e-01 -3.13553095e-01 6.09479606e-01
6.76325113e-02 7.98881233e-01 1.42708406e-01 -3.48713160e-01
6.03575468e-01 5.12181759e-01 -1.66651048e-02 -4.47617531e-01
-8.32934797e-01 -2.02037215e-01 5.11636078e-01 -2.82918841e-01
-1.44155145e-01 -1.48307070e-01 -1.70113575e+00 -1.42836362e-01
-8.08485270e-01 4.61000919e-01 6.57753825e-01 6.98539257e-01
7.18267024e-01 4.67097551e-01 1.73165008e-01 -5.53723752e-01
-1.03439116e+00 -1.17002547e+00 -5.27496040e-01 3.33940178e-01
-3.53511423e-02 -4.46378738e-01 -3.71816069e-01 7.05423355e-02] | [9.55909252166748, 9.378582000732422] |
851be997-198d-4201-a77c-6f1b5fc63f3d | prompter-zero-shot-adaptive-prefixes-for | 2306.04724 | null | https://arxiv.org/abs/2306.04724v1 | https://arxiv.org/pdf/2306.04724v1.pdf | Prompter: Zero-shot Adaptive Prefixes for Dialogue State Tracking Domain Adaptation | A challenge in the Dialogue State Tracking (DST) field is adapting models to new domains without using any supervised data, zero-shot domain adaptation. Parameter-Efficient Transfer Learning (PETL) has the potential to address this problem due to its robustness. However, it has yet to be applied to the zero-shot scenarios, as it is not clear how to apply it unsupervisedly. Our method, Prompter, uses descriptions of target domain slots to generate dynamic prefixes that are concatenated to the key and values at each layer's self-attention mechanism. This allows for the use of prefix-tuning in zero-shot. Prompter outperforms previous methods on both the MultiWOZ and SGD benchmarks. In generating prefixes, our analyses find that Prompter not only utilizes the semantics of slot descriptions but also how often the slots appear together in conversation. Moreover, Prompter's gains are due to its improved ability to distinguish "none"-valued dialogue slots, compared against baselines. | ['Nancy F. Chen', 'Min-Yen Kan', 'Taha Aksu'] | 2023-06-07 | null | null | null | null | ['dialogue-state-tracking'] | ['natural-language-processing'] | [ 2.91271120e-01 4.73298371e-01 -4.08853978e-01 -4.03401405e-01
-8.13300073e-01 -6.74011528e-01 1.03454828e+00 1.12530164e-01
-5.33141434e-01 9.52491045e-01 8.27520370e-01 -2.75165230e-01
1.24466509e-01 -5.81920922e-01 -1.79861009e-01 -2.40753442e-01
1.00009805e-02 9.84012544e-01 4.50347066e-01 -8.78767014e-01
1.68111235e-01 -9.47054252e-02 -1.24867702e+00 4.07562166e-01
6.05262399e-01 5.33953428e-01 1.37934223e-01 5.72163522e-01
-6.60030246e-01 6.59824073e-01 -6.85105324e-01 -3.56790453e-01
4.71127182e-02 -7.10873008e-01 -1.29547250e+00 8.78459141e-02
1.29331291e-01 -6.02437437e-01 -3.49230021e-01 5.01080990e-01
5.95808923e-01 5.63557386e-01 6.15859509e-01 -1.12564182e+00
-1.96960106e-01 7.43007243e-01 -2.03180715e-01 2.33849093e-01
5.89590430e-01 3.34309161e-01 1.24951911e+00 -5.73533356e-01
1.12910664e+00 1.55181503e+00 5.06108463e-01 1.02029455e+00
-1.39516020e+00 -4.13894385e-01 2.05332115e-01 5.55069372e-02
-5.62028229e-01 -8.19139779e-01 6.93161190e-01 -3.31959546e-01
1.37807345e+00 2.85831466e-03 2.43853301e-01 1.47939503e+00
-3.28404695e-01 1.00960958e+00 1.09198225e+00 -6.31898761e-01
2.91825801e-01 2.27661923e-01 3.28666121e-01 2.56812066e-01
-4.77028996e-01 4.32731770e-02 -7.75752008e-01 -3.16857100e-01
5.07812440e-01 -5.07024229e-01 -1.11078862e-02 -3.82853925e-01
-1.18550014e+00 9.79411006e-01 -3.49648148e-02 3.49997371e-01
-2.88191199e-01 -3.61725718e-01 7.61372805e-01 5.50623417e-01
6.80288255e-01 8.44799101e-01 -6.90639496e-01 -9.05367851e-01
-4.55603272e-01 5.41466236e-01 1.20952928e+00 9.72466588e-01
7.47637749e-01 -1.25532970e-01 -5.62654674e-01 1.21215117e+00
-1.31664857e-01 8.66726320e-03 7.09081173e-01 -1.22938728e+00
5.73620141e-01 6.04794204e-01 3.97178113e-01 -3.04542214e-01
-5.13598502e-01 1.49994373e-01 -1.99102283e-01 -1.78013086e-01
6.97267115e-01 -6.07505262e-01 -6.86779618e-01 2.04828763e+00
3.78722280e-01 7.95529981e-04 4.51443076e-01 6.15487039e-01
7.77411520e-01 6.15327477e-01 4.00433689e-01 -8.95229056e-02
1.38120210e+00 -9.22583520e-01 -7.11641490e-01 -6.43390059e-01
9.06958818e-01 -5.86302817e-01 1.23680258e+00 -1.14468880e-01
-9.67911601e-01 -3.65457237e-01 -7.05210268e-01 -1.33189425e-01
-4.37523633e-01 -4.92956042e-01 6.79185987e-01 4.77599561e-01
-8.31225336e-01 6.28579736e-01 -6.01291120e-01 -8.55330169e-01
6.24870062e-02 1.13335259e-01 -2.52048701e-01 3.97125036e-02
-1.78307879e+00 1.20115650e+00 4.75182295e-01 -5.87739587e-01
-4.97930586e-01 -5.76872706e-01 -1.22493505e+00 2.07927138e-01
6.15394413e-01 -4.21084166e-01 2.00009847e+00 -8.59567523e-01
-1.97707617e+00 7.22193241e-01 -3.41930002e-01 -5.85380197e-01
3.06386501e-01 -7.33290687e-02 -1.14872769e-01 -2.15967856e-02
1.53532237e-01 7.71245897e-01 3.97911102e-01 -8.98903251e-01
-7.53845513e-01 -5.85210063e-02 4.13358629e-01 5.70478380e-01
-3.13909113e-01 1.38605423e-02 -3.82908940e-01 -3.24856758e-01
-2.07647994e-01 -9.49795902e-01 -2.05301687e-01 -3.70125383e-01
-1.21524617e-01 -6.53574526e-01 8.06686163e-01 -5.69272637e-01
1.27869856e+00 -2.13929796e+00 3.82793397e-02 -1.17159940e-01
5.90770431e-02 4.88828152e-01 -2.70608187e-01 8.69799078e-01
2.48306617e-01 -1.12821393e-01 -1.04714438e-01 -3.73029888e-01
3.88116777e-01 3.80156934e-01 -1.65046468e-01 -4.96644266e-02
3.21145058e-01 9.16684628e-01 -1.13554311e+00 -4.19665188e-01
3.05156142e-01 5.03678247e-02 -7.53613949e-01 3.39605272e-01
-5.89868128e-01 2.62734711e-01 -3.40459824e-01 1.14970788e-01
1.46424443e-01 -2.78017968e-01 5.40428698e-01 1.41858026e-01
-2.85477471e-02 1.00969148e+00 -8.82790864e-01 1.79344797e+00
-4.94526535e-01 5.66228688e-01 2.01635435e-02 -7.42231905e-01
9.05328929e-01 5.53716481e-01 2.99070328e-01 -8.91032696e-01
-4.97585461e-02 1.89567497e-03 1.27782166e-01 -4.48902607e-01
7.27315068e-01 -3.70341152e-01 -4.97347027e-01 8.40241969e-01
4.13940281e-01 1.09496817e-01 3.80109727e-01 5.10418534e-01
1.20781994e+00 1.91914409e-01 5.25156617e-01 -2.82422900e-02
1.47381991e-01 3.08259100e-01 6.26025021e-01 8.26179147e-01
-3.81807536e-01 1.68076769e-01 7.18845069e-01 -1.05921395e-01
-1.14199185e+00 -8.52602363e-01 7.92524144e-02 1.80388868e+00
9.21407156e-03 -4.02274460e-01 -7.42529213e-01 -7.29375958e-01
1.65051356e-01 1.03962255e+00 -4.79952872e-01 -3.66574973e-01
-4.85415906e-01 -2.93133169e-01 4.32205617e-01 3.83705616e-01
4.94464070e-01 -1.28211069e+00 -4.87504512e-01 6.83411121e-01
-5.13241768e-01 -1.09494758e+00 -4.93678719e-01 3.74254048e-01
-6.62497580e-01 -8.46349537e-01 -6.55289590e-01 -5.73733389e-01
7.34766573e-02 1.65906250e-01 1.26829433e+00 -2.94476658e-01
8.22059214e-02 5.50841451e-01 -5.94974935e-01 -2.20620885e-01
-8.70872080e-01 5.38550794e-01 -4.91925068e-02 -1.92185745e-01
8.91078532e-01 -3.76758456e-01 -1.97329849e-01 4.09509897e-01
-5.23249149e-01 6.98716417e-02 3.11083436e-01 1.09573150e+00
-1.50409326e-01 -6.04011118e-01 1.00448418e+00 -1.50283539e+00
1.15471470e+00 -5.76696754e-01 -1.76870719e-01 1.59563631e-01
-5.88273644e-01 2.81831175e-01 2.45634973e-01 -4.59058791e-01
-1.40744877e+00 -2.35968903e-01 -7.40618780e-02 4.17768061e-02
-3.59525949e-01 3.61327082e-01 -6.40391111e-02 4.63515490e-01
8.52473140e-01 6.95384443e-02 3.53383631e-01 -6.28499985e-01
5.21858513e-01 8.84466767e-01 5.38654506e-01 -8.63506138e-01
4.69405860e-01 -1.03516646e-01 -7.37539649e-01 -9.30414438e-01
-7.21201658e-01 -7.80266464e-01 -5.71516156e-01 5.87635338e-02
7.64718413e-01 -8.96316469e-01 -4.82567281e-01 2.29144484e-01
-1.10024714e+00 -8.53599191e-01 -3.75086188e-01 1.41718492e-01
-7.74287879e-01 4.62008357e-01 -9.07602310e-01 -9.23801601e-01
-2.49310195e-01 -7.86727667e-01 7.63484418e-01 2.63473660e-01
-8.97041976e-01 -1.16866827e+00 1.54328316e-01 2.61970550e-01
6.83724225e-01 -1.45366862e-01 1.08210361e+00 -1.28900611e+00
-9.67989191e-02 5.95874265e-02 1.07448570e-01 8.71950313e-02
1.92101300e-01 -2.96270221e-01 -1.10631680e+00 -2.24499628e-01
-3.22080344e-01 -6.77695453e-01 5.94675899e-01 1.61011107e-02
3.95489395e-01 -4.56931204e-01 -3.24687928e-01 -1.69748515e-02
7.48637557e-01 2.89219797e-01 4.10540432e-01 5.58630109e-01
1.12164728e-01 8.57494712e-01 7.88105011e-01 5.46777189e-01
7.84704626e-01 9.79938388e-01 -1.28635123e-01 -6.65798411e-02
-1.02432661e-01 -3.30349326e-01 3.36226285e-01 4.66223717e-01
3.42839777e-01 -1.68389603e-01 -1.00376439e+00 6.36873424e-01
-1.94701672e+00 -1.17432892e+00 3.48455936e-01 2.02734351e+00
1.16883075e+00 4.43154007e-01 3.30421984e-01 -2.24324986e-01
6.39408708e-01 3.25642705e-01 -6.97301984e-01 -6.83580399e-01
5.75323105e-02 2.30480075e-01 1.32925376e-01 7.12893128e-01
-1.04137933e+00 1.41115594e+00 6.22309113e+00 5.27429163e-01
-8.40292573e-01 1.23389393e-01 3.67702246e-01 8.97733867e-02
-2.44142205e-01 1.17214911e-01 -1.01349449e+00 4.30800706e-01
1.13359630e+00 -3.21688801e-01 4.25312817e-01 8.06205869e-01
4.00494784e-02 -2.23714367e-01 -1.15166259e+00 4.44298118e-01
-1.69663146e-01 -1.12125027e+00 -1.52260214e-01 -1.29932150e-01
5.48057377e-01 -8.95927474e-03 -2.49581635e-01 1.04642951e+00
8.43631387e-01 -5.77068090e-01 3.12895536e-01 5.15277386e-02
6.28520131e-01 -5.98888278e-01 5.45136213e-01 5.76946080e-01
-5.90227008e-01 -4.29862924e-02 -3.12180459e-01 -2.24672988e-01
2.30908319e-01 -4.04063389e-02 -1.71024716e+00 1.33562475e-01
2.32537672e-01 4.87926304e-01 -3.17651540e-01 6.39303625e-01
-1.06383093e-01 7.97825396e-01 -2.37846076e-01 -1.24999709e-01
4.79723513e-01 1.28551692e-01 5.09622753e-01 1.42366743e+00
-1.00336500e-01 2.12918773e-01 4.46663916e-01 6.01920426e-01
5.98965120e-03 1.45408930e-02 -6.55366957e-01 -2.90081054e-01
9.32534277e-01 1.01558757e+00 -2.14383677e-01 -6.08183861e-01
-5.35976529e-01 9.32815552e-01 2.43493870e-01 3.88177127e-01
-1.55841738e-01 -4.20228928e-01 1.00124645e+00 1.34351090e-01
1.45024523e-01 -1.79381341e-01 -9.75872278e-02 -9.85400438e-01
-4.52695936e-01 -1.12972164e+00 6.09186888e-01 -5.61669767e-01
-1.41034877e+00 5.26904464e-01 2.82443970e-01 -8.71429205e-01
-1.08723438e+00 -3.37480456e-01 -6.46673501e-01 9.71213818e-01
-1.42717373e+00 -9.86910820e-01 7.28962291e-03 6.49699748e-01
9.75668609e-01 -2.32847989e-01 1.25463581e+00 -2.11145710e-02
-4.09349948e-01 6.58230007e-01 7.30853453e-02 3.30481529e-01
1.18651867e+00 -1.39628124e+00 9.83358622e-01 3.81909758e-01
-1.05270751e-01 5.20000041e-01 9.32645798e-01 -5.64034581e-01
-9.11053717e-01 -6.29206061e-01 1.17588949e+00 -6.17686689e-01
6.73965633e-01 -4.26078379e-01 -1.26923525e+00 8.64409864e-01
3.35320681e-01 -4.39327210e-01 8.32391322e-01 7.16363609e-01
-4.66582119e-01 3.07353765e-01 -9.89323854e-01 5.69795132e-01
8.77267778e-01 -6.96108222e-01 -1.12440312e+00 1.26699239e-01
8.46802115e-01 -5.22151232e-01 -8.59501600e-01 1.56194612e-01
3.02179128e-01 -7.94551134e-01 8.31935942e-01 -9.13993657e-01
2.81402886e-01 2.85251498e-01 5.06768003e-02 -1.54586232e+00
-3.01563233e-01 -9.02674615e-01 -9.76202935e-02 1.47653961e+00
6.01432920e-01 -6.37507677e-01 7.53446698e-01 9.64121103e-01
-1.68383062e-01 -3.75473976e-01 -8.93574834e-01 -6.44125104e-01
1.86540917e-01 1.43949939e-02 6.24934852e-01 1.20273912e+00
5.68805516e-01 9.36090946e-01 -4.73760962e-01 -4.39258337e-01
3.32831472e-01 4.22217026e-02 1.08960319e+00 -1.43840742e+00
-4.78701651e-01 -3.73759001e-01 1.01243168e-01 -1.21139836e+00
2.87775755e-01 -7.27959454e-01 1.63909122e-01 -1.49181998e+00
-8.09865892e-02 -5.49580514e-01 -1.10967062e-01 7.45445013e-01
-3.98209959e-01 -2.97006458e-01 3.77803177e-01 6.62575737e-02
-8.86726856e-01 5.47276974e-01 1.09074867e+00 -6.54764008e-03
-5.06157517e-01 -1.89780723e-02 -6.37238920e-01 4.35396880e-01
9.13591802e-01 -2.23029807e-01 -4.31235254e-01 -9.80027318e-02
-3.33618999e-01 3.52979869e-01 -5.59826940e-02 -7.60401905e-01
2.09310398e-01 -2.70974815e-01 -4.30152416e-02 -1.57423258e-01
5.99370897e-01 -2.94618756e-01 -2.97045916e-01 6.48957267e-02
-7.82059550e-01 -1.21282220e-01 4.04749304e-01 4.48063672e-01
-4.44438383e-02 -3.10848355e-01 7.66932547e-01 -3.50673944e-01
-1.19443536e+00 -1.18633136e-01 -6.55229390e-01 5.34893930e-01
7.46249318e-01 -2.35021204e-01 -4.24822539e-01 -6.43116891e-01
-7.63227165e-01 4.70688313e-01 4.44330066e-01 7.08442152e-01
1.76620446e-02 -1.18632758e+00 -5.94105005e-01 1.43763870e-01
4.87176478e-01 -7.07866997e-02 2.98113734e-01 4.30352986e-01
2.31308579e-01 3.99579614e-01 -3.21150690e-01 -4.41498190e-01
-1.11834455e+00 3.56104970e-01 2.22980142e-01 -4.58837271e-01
-7.04862118e-01 6.86446309e-01 -2.08530910e-02 -7.02924013e-01
4.43238556e-01 6.74190447e-02 -2.89746851e-01 2.19519690e-01
5.31208217e-01 5.85251898e-02 -1.15693115e-01 -3.56350899e-01
-4.67371345e-02 -1.42177284e-01 -4.95129257e-01 -4.13964599e-01
1.22943151e+00 -3.28888923e-01 3.62101525e-01 7.61396646e-01
9.19059277e-01 -3.12821209e-01 -1.54775989e+00 -8.45139503e-01
5.28149366e-01 -3.08312505e-01 -3.91907364e-01 -1.07419741e+00
-1.46980405e-01 7.75713086e-01 2.28610903e-01 2.67894328e-01
5.53422093e-01 4.53544296e-02 9.51357841e-01 6.28649950e-01
3.89239132e-01 -1.42160678e+00 2.01913059e-01 1.05966830e+00
4.10730273e-01 -1.29086351e+00 -5.18776178e-01 1.80090796e-02
-1.16826320e+00 8.81742597e-01 8.64168525e-01 2.78562486e-01
1.28215566e-01 1.03342369e-01 3.84101301e-01 7.04676984e-03
-1.23943877e+00 -4.85932440e-01 -1.03765644e-01 8.22579384e-01
6.08306944e-01 -5.32094538e-02 -2.55861193e-01 4.23833072e-01
-2.18910515e-01 -2.08560824e-01 5.09822369e-01 9.93234932e-01
-7.41021216e-01 -1.48863029e+00 -6.78818971e-02 4.71567094e-01
-1.43617555e-01 4.18390264e-04 -5.81952691e-01 8.12182963e-01
-3.86560142e-01 1.08481002e+00 1.16003864e-01 -3.15277845e-01
5.43148637e-01 8.01532090e-01 9.69129279e-02 -1.04767692e+00
-7.16287613e-01 4.17823493e-02 6.49980843e-01 -4.45256829e-01
-1.58103958e-01 -6.48736954e-01 -1.15032721e+00 -3.64705890e-01
-1.29734740e-01 5.79798937e-01 3.93735677e-01 1.02796245e+00
4.93933856e-01 4.06447530e-01 4.79187578e-01 -7.85983026e-01
-1.00788295e+00 -1.24054754e+00 -3.48507881e-01 7.49925733e-01
2.44702086e-01 -9.02078331e-01 -1.39191493e-01 -2.46522233e-01] | [12.773920059204102, 7.893133640289307] |
1c990ab0-f5eb-4079-bed0-26ccb5d0ba8f | prompt-federated-learning-for-weather | 2301.09152 | null | https://arxiv.org/abs/2301.09152v2 | https://arxiv.org/pdf/2301.09152v2.pdf | Prompt Federated Learning for Weather Forecasting: Toward Foundation Models on Meteorological Data | To tackle the global climate challenge, it urgently needs to develop a collaborative platform for comprehensive weather forecasting on large-scale meteorological data. Despite urgency, heterogeneous meteorological sensors across countries and regions, inevitably causing multivariate heterogeneity and data exposure, become the main barrier. This paper develops a foundation model across regions capable of understanding complex meteorological data and providing weather forecasting. To relieve the data exposure concern across regions, a novel federated learning approach has been proposed to collaboratively learn a brand-new spatio-temporal Transformer-based foundation model across participants with heterogeneous meteorological data. Moreover, a novel prompt learning mechanism has been adopted to satisfy low-resourced sensors' communication and computational constraints. The effectiveness of the proposed method has been demonstrated on classical weather forecasting tasks using three meteorological datasets with multivariate time series. | ['Jing Jiang', 'Tao Shen', 'Guodong Long', 'Shengchao Chen'] | 2023-01-22 | null | null | null | null | ['weather-forecasting'] | ['miscellaneous'] | [-2.55351990e-01 -3.97498280e-01 6.88442215e-02 -6.56618595e-01
-5.41674852e-01 -6.88915968e-01 6.65418744e-01 3.11906040e-01
-1.78467795e-01 8.82155120e-01 3.44784945e-01 -4.19499844e-01
-4.82182771e-01 -1.16044271e+00 -3.80569875e-01 -7.86137998e-01
-5.63399017e-01 -1.56216174e-01 -1.29053026e-01 -3.46535206e-01
-1.51018864e-02 3.18889081e-01 -1.60299134e+00 1.62232250e-01
1.19069254e+00 9.78593886e-01 4.08038795e-01 5.92615724e-01
-2.99941957e-01 7.06772923e-01 -7.10195005e-01 -5.42134233e-03
3.66393119e-01 7.35384300e-02 -4.55101430e-01 -2.72682697e-01
2.69349992e-01 -1.56404123e-01 1.33567497e-01 7.40565062e-01
7.18633175e-01 3.67867082e-01 9.63775814e-02 -1.32408607e+00
-3.73297811e-01 5.95166385e-01 -2.19946340e-01 3.93843889e-01
2.04849616e-01 -2.49733776e-02 8.68880272e-01 -7.01249957e-01
1.96522266e-01 7.94552386e-01 8.33019733e-01 2.36270010e-01
-7.20879436e-01 -7.09509909e-01 8.60808551e-01 1.23824015e-01
-1.19625342e+00 -9.64636579e-02 6.29728258e-01 -3.89099896e-01
9.21209097e-01 7.19146371e-01 5.62806785e-01 9.88975286e-01
3.74147117e-01 1.34534210e-01 1.35585356e+00 -6.74287453e-02
3.40194941e-01 2.83380777e-01 -1.71863958e-01 9.36549678e-02
2.38949418e-01 3.86025816e-01 -6.32494926e-01 -2.52339721e-01
1.39466032e-01 3.76026034e-01 -9.54237133e-02 3.26845735e-01
-1.23979211e+00 4.36858654e-01 5.62071264e-01 3.76325488e-01
-8.05266380e-01 -2.87258595e-01 4.53483760e-01 2.99081057e-01
1.09869957e+00 1.52917489e-01 -9.69889283e-01 9.51835215e-02
-8.90542328e-01 2.45461926e-01 7.43075430e-01 6.27224207e-01
7.35320926e-01 3.75502169e-01 2.72652626e-01 2.89292753e-01
1.70205116e-01 1.10906446e+00 5.89375570e-02 -5.98867595e-01
8.12633991e-01 5.70831120e-01 6.17541313e-01 -1.31721473e+00
-7.30900288e-01 -5.36184072e-01 -1.46363366e+00 -4.33698185e-02
5.11786491e-02 -8.70897591e-01 -1.86952576e-01 1.59700477e+00
8.90786409e-01 6.85665131e-01 1.74430907e-01 1.20918059e+00
6.90087199e-01 1.14255142e+00 2.20012695e-01 -3.19241643e-01
1.06028116e+00 -5.71683884e-01 -8.53440762e-01 2.86609828e-01
2.73213267e-01 -6.28445745e-01 8.54744852e-01 4.42067057e-01
-5.59112251e-01 -7.45386422e-01 -6.33142591e-01 5.58979571e-01
-8.54714334e-01 2.31580772e-02 8.09070110e-01 5.11261344e-01
-9.85848069e-01 1.75253168e-01 -4.77562755e-01 -3.37811261e-01
-1.15742497e-01 -6.89943656e-02 -1.23177037e-01 -1.49291789e-03
-1.67902327e+00 8.13149214e-01 2.24004567e-01 5.46590209e-01
-9.56155241e-01 -1.24111402e+00 -4.04082835e-01 1.34334946e-02
-7.33769592e-03 -7.61787295e-01 7.08893180e-01 -5.94687760e-01
-1.31684518e+00 1.07096303e-02 2.98657324e-02 -7.02993050e-02
2.83195555e-01 -1.86522424e-01 -1.38150859e+00 -3.72465640e-01
-3.08443569e-02 -1.69516236e-01 7.88158655e-01 -1.09077382e+00
-1.13748324e+00 -3.44217062e-01 -1.41656369e-01 3.07561427e-01
-8.80319893e-01 2.79686563e-02 4.35242563e-01 -7.69956350e-01
-3.82956654e-01 -4.07423496e-01 -7.77597070e-01 -5.20608008e-01
1.73162669e-01 -2.73405552e-01 8.99082899e-01 -7.88796782e-01
1.39834988e+00 -2.01337004e+00 -3.47680263e-02 5.19966066e-01
1.80393741e-01 1.95680767e-01 -2.64650285e-01 8.18086028e-01
1.05758667e-01 -6.38032183e-02 -1.21817570e-02 -3.08409363e-01
2.78853183e-03 3.87169868e-01 -9.59134400e-01 4.49474782e-01
-7.84991533e-02 3.58134478e-01 -8.72808218e-01 -1.07623696e-01
2.78001308e-01 5.88613868e-01 -3.31142917e-02 7.24026024e-01
-4.20216352e-01 9.52819169e-01 -7.45866954e-01 6.96552992e-01
1.07527375e+00 -2.03688487e-01 2.48053744e-01 -1.71255454e-01
-8.61689508e-01 -2.56829448e-02 -1.49593353e+00 1.60406983e+00
-9.07396972e-01 -4.98228595e-02 6.26505196e-01 -9.68350172e-01
1.16853559e+00 6.58974528e-01 7.26267755e-01 -9.63027656e-01
-2.47181013e-01 1.41954839e-01 -5.93364716e-01 -7.88028538e-01
8.13467443e-01 1.25790596e-01 -3.50051045e-01 7.50576258e-01
-4.43480551e-01 1.13375448e-01 -3.54938745e-01 2.85891127e-02
7.21436858e-01 -8.83747078e-03 -9.80555266e-02 -4.11352962e-01
6.32865489e-01 1.58143565e-01 8.81547928e-01 5.42632818e-01
-1.99989572e-01 -2.39734009e-01 -5.00504792e-01 -1.32471228e+00
-6.55390620e-01 -7.40078509e-01 9.97821987e-02 1.37426090e+00
-2.06323508e-02 -4.13534641e-01 -6.56517074e-02 -3.34039211e-01
2.30254605e-01 3.02964419e-01 -3.75893205e-01 4.92365003e-01
-2.85652965e-01 -1.00348222e+00 5.98727047e-01 1.05116896e-01
5.17694294e-01 -6.14026129e-01 -7.69503295e-01 4.78918642e-01
-3.77063096e-01 -8.53404164e-01 -2.77213356e-03 -1.59444418e-02
-8.02980006e-01 -7.94878662e-01 -4.16903615e-01 -1.46918565e-01
4.39806163e-01 6.27624869e-01 1.09941196e+00 -1.98193520e-01
-2.84423262e-01 5.19035518e-01 -4.06649113e-01 -6.76198363e-01
2.21825078e-01 2.13462159e-01 3.09675366e-01 2.81584829e-01
-9.58214179e-02 -6.96929872e-01 -8.26986194e-01 2.70959288e-01
-9.50983465e-01 -3.28937918e-01 5.76218963e-02 4.99940395e-01
3.01547438e-01 5.02015054e-01 1.04013705e+00 -5.27531624e-01
6.14066422e-01 -1.09117401e+00 -1.02187121e+00 5.60986042e-01
-8.66943359e-01 -4.89584595e-01 1.03332818e+00 -1.77025527e-01
-1.59201300e+00 -3.99911329e-02 3.75127316e-01 -2.92181373e-02
-3.71776640e-01 1.03134739e+00 1.20005876e-01 -4.04523090e-02
5.68693697e-01 2.57694989e-01 -3.65416825e-01 -6.66559756e-01
6.09575689e-01 7.75549293e-01 5.30483782e-01 -7.24350214e-01
9.28223670e-01 5.11272550e-01 -5.99685162e-02 -5.10993183e-01
-8.40869009e-01 -5.45931816e-01 -2.33551234e-01 -5.51571786e-01
6.31689072e-01 -1.68929887e+00 -7.84939051e-01 6.10763371e-01
-1.16455340e+00 -2.03230679e-01 1.03954859e-01 6.81521535e-01
3.07651490e-01 1.51100188e-01 -1.91447437e-01 -1.03516543e+00
-7.04041839e-01 -3.40018511e-01 5.52655935e-01 2.02889562e-01
5.14410026e-02 -1.45420051e+00 7.49191880e-01 1.89543024e-01
1.32519579e+00 5.80640972e-01 4.15460348e-01 -2.47019649e-01
-6.15747750e-01 8.99175182e-02 -8.02305862e-02 6.73490912e-02
6.78921461e-01 -4.97921072e-02 -1.19275057e+00 -5.97773015e-01
-1.33390054e-01 -2.43814334e-01 4.30957437e-01 -6.41379952e-02
1.11726427e+00 -6.25990868e-01 -2.47403190e-01 6.06317759e-01
1.47032285e+00 -1.48862332e-01 -9.86725911e-02 1.64365336e-01
6.20808423e-01 8.79346788e-01 8.09013784e-01 1.16174924e+00
1.01761210e+00 1.44581616e-01 6.55313969e-01 -3.83325756e-01
4.54681635e-01 3.51946391e-02 3.43843699e-01 1.09202814e+00
-1.77199855e-01 -3.42339069e-01 -8.63259614e-01 7.16027141e-01
-1.87545550e+00 -1.11996496e+00 -2.52935916e-01 1.95717525e+00
6.74081683e-01 -6.48292899e-01 -4.93471809e-02 -1.43347412e-01
4.29957718e-01 6.95290506e-01 -4.98280913e-01 -3.46674562e-01
-3.63676697e-01 1.17475674e-01 5.22701979e-01 4.03185368e-01
-1.12923861e+00 4.91866887e-01 6.38318682e+00 3.84803772e-01
-1.69851780e+00 4.89150912e-01 3.24515551e-01 2.52020312e-03
-5.97285986e-01 -1.60289600e-01 -6.72570109e-01 5.45324504e-01
1.36946571e+00 -2.07529560e-01 4.98403698e-01 6.38040185e-01
7.35719085e-01 1.34146810e-01 -3.98259223e-01 6.50582612e-01
-4.95257884e-01 -1.58518314e+00 -2.40978360e-01 -2.54629284e-01
1.14058518e+00 5.57001233e-01 4.87086503e-03 1.61885004e-02
6.42592430e-01 -8.13702524e-01 2.87864745e-01 9.07523274e-01
6.06526077e-01 -6.12328649e-01 4.69399482e-01 5.53925812e-01
-1.85842955e+00 -3.73009145e-01 -3.93034786e-01 -5.77255130e-01
1.68149248e-01 1.17152071e+00 -4.56473082e-01 1.40039647e+00
1.10290349e+00 9.45809484e-01 -1.40213475e-01 7.93732762e-01
1.98655710e-01 7.33098984e-01 -4.05853599e-01 1.82435066e-01
9.04965475e-02 -3.33397061e-01 3.33377212e-01 1.05922604e+00
7.73368955e-01 3.23792070e-01 5.14960766e-01 3.03193212e-01
3.12745422e-01 2.15898171e-01 -8.79199743e-01 2.98945755e-01
7.93799102e-01 1.69867134e+00 6.64092675e-02 -3.66222352e-01
-5.04707217e-01 3.61174911e-01 9.17583480e-02 4.51496094e-01
-5.98044991e-01 -9.73844342e-03 9.35751975e-01 -3.81399721e-01
2.09905192e-01 -3.32137316e-01 -2.66806334e-01 -1.24964952e+00
5.31600676e-02 -7.90670753e-01 6.87082231e-01 -5.50902724e-01
-1.68763685e+00 6.83189511e-01 -2.33610928e-01 -1.47718644e+00
2.38544047e-02 -5.63226789e-02 -1.16526866e+00 1.31329465e+00
-2.20637608e+00 -1.48410130e+00 -6.04818821e-01 1.25469172e+00
6.61598817e-02 -3.78334433e-01 1.27400231e+00 6.84932232e-01
-6.25442982e-01 5.87495230e-02 2.34832242e-01 -5.20677924e-01
7.99388885e-01 -9.36010897e-01 1.87306970e-01 1.04607320e+00
-3.32462251e-01 6.26601160e-01 5.71682572e-01 -6.65066838e-01
-1.66869771e+00 -1.81164610e+00 1.30903828e+00 -1.71567708e-01
6.83155894e-01 -3.39804977e-01 -9.05946136e-01 2.98691779e-01
6.23869061e-01 3.35409552e-01 9.19739127e-01 3.05224121e-01
-6.23002648e-01 -9.75726485e-01 -1.14289165e+00 -7.55421296e-02
6.40789747e-01 -4.83847290e-01 -3.01115960e-01 6.42245114e-01
7.31344938e-01 -1.51947096e-01 -1.43679786e+00 3.90652955e-01
2.09932834e-01 -6.88744247e-01 6.18731022e-01 -6.46050751e-01
-2.03735307e-01 -6.51068628e-01 -3.48581195e-01 -1.83997035e+00
-3.49156827e-01 -7.83427954e-01 1.99301727e-02 1.36946142e+00
3.43181908e-01 -9.17796075e-01 4.03885841e-01 5.93177915e-01
-1.28989577e-01 -8.67170617e-02 -1.14557004e+00 -5.33237338e-01
-9.64200720e-02 -4.72709268e-01 1.40567875e+00 1.51001835e+00
2.79772151e-02 -2.63163805e-01 -8.09939921e-01 8.88648570e-01
5.82953036e-01 6.00227892e-01 8.07756007e-01 -1.42615068e+00
-5.82616730e-03 1.38100550e-01 4.14775312e-01 -4.20963973e-01
7.24038407e-02 -4.25016403e-01 -2.44145691e-01 -1.21583247e+00
-5.03789723e-01 -1.00819993e+00 -8.59371483e-01 5.63251138e-01
1.72091443e-02 -2.50028610e-01 -1.61190763e-01 1.55603945e-01
-4.82263178e-01 6.41524673e-01 9.27681863e-01 -2.29138345e-01
-7.55801201e-02 2.04655707e-01 -4.52705622e-01 9.46481898e-02
1.10023224e+00 -2.64304161e-01 -6.15211546e-01 -8.43368828e-01
8.40096712e-01 3.25144470e-01 3.81198615e-01 -8.27546418e-01
6.92183971e-01 -7.72166491e-01 1.43229738e-01 -7.36157179e-01
-9.67288017e-02 -1.32038867e+00 7.45821953e-01 3.43050748e-01
-1.12337977e-01 7.55142331e-01 1.95562020e-01 5.64602971e-01
-4.64072227e-01 9.41907346e-01 -3.67836189e-03 -6.71655387e-02
-9.09135818e-01 5.41124821e-01 -4.07107949e-01 -2.89627880e-01
1.12903965e+00 4.80338991e-01 -6.11807704e-01 -1.62799228e-02
-5.61723948e-01 7.93309212e-01 -9.52506810e-02 5.33269882e-01
5.18994093e-01 -1.20989370e+00 -9.27776635e-01 4.24384087e-01
1.78924933e-01 -9.86433029e-02 9.38827813e-01 7.94115067e-01
5.69967292e-02 4.50217545e-01 -2.31372580e-01 -3.34663987e-01
-8.52034867e-01 4.28370029e-01 3.09424341e-01 -3.19564104e-01
-3.98173064e-01 5.47417939e-01 -4.72506404e-01 -1.09314823e+00
2.24582925e-01 -5.52599609e-01 -1.43419921e-01 4.39555585e-01
6.42198384e-01 4.92062211e-01 3.15000743e-01 -2.40253806e-01
-3.97700340e-01 3.46202075e-01 5.62502980e-01 1.67674005e-01
1.68056250e+00 -7.51630783e-01 -3.77976805e-01 3.91057163e-01
5.55992484e-01 9.34293866e-02 -1.26275611e+00 -5.13783038e-01
-1.19365290e-01 -4.67535943e-01 2.15708315e-01 -1.16492689e+00
-1.32560682e+00 6.33576453e-01 7.30084181e-01 5.17902076e-01
1.31159580e+00 -6.52803361e-01 9.38593626e-01 3.60583395e-01
5.15616536e-01 -1.16862035e+00 -5.39979637e-01 4.38790917e-01
5.78506112e-01 -1.38086331e+00 -2.23342255e-02 5.69204837e-02
-5.23321867e-01 8.99932265e-01 4.69451606e-01 1.94683567e-01
1.12821710e+00 4.21213835e-01 5.88036478e-01 -2.08860770e-01
-1.21930540e+00 1.14928715e-01 1.78973749e-02 7.57660806e-01
3.04110020e-01 6.09225452e-01 5.56702632e-03 5.90268672e-01
1.64822385e-01 1.04315169e-01 1.83311164e-01 7.40492523e-01
-2.46121809e-01 -1.15056384e+00 -6.59271061e-01 6.29981235e-02
-3.33810151e-01 2.30493009e-04 2.43574351e-01 2.81008601e-01
4.74134922e-01 1.29042923e+00 3.45477574e-02 -5.22645295e-01
2.41657853e-01 -2.29081810e-01 -4.37600553e-01 -9.79169831e-02
-1.10441387e+00 -2.42721483e-01 -8.15278888e-02 -7.37098336e-01
-1.09167778e+00 -4.32261020e-01 -8.06882679e-01 -6.65353596e-01
1.75348967e-01 5.46325445e-01 9.53450441e-01 8.09692740e-01
9.70013618e-01 3.50506842e-01 1.55230856e+00 -7.36743808e-01
-6.63082123e-01 -7.25200295e-01 -4.98734742e-01 1.17901694e-02
5.67359626e-01 -2.80049860e-01 -2.66433716e-01 -9.61465687e-02] | [6.725500583648682, 2.8042893409729004] |
6602ea92-1df5-4b2a-b943-d6806a50556c | making-person-search-enjoy-the-merits-of | 2108.10536 | null | https://arxiv.org/abs/2108.10536v2 | https://arxiv.org/pdf/2108.10536v2.pdf | Making Person Search Enjoy the Merits of Person Re-identification | Person search is an extended task of person re-identification (Re-ID). However, most existing one-step person search works have not studied how to employ existing advanced Re-ID models to boost the one-step person search performance due to the integration of person detection and Re-ID. To address this issue, we propose a faster and stronger one-step person search framework, the Teacher-guided Disentangling Networks (TDN), to make the one-step person search enjoy the merits of the existing Re-ID researches. The proposed TDN can significantly boost the person search performance by transferring the advanced person Re-ID knowledge to the person search model. In the proposed TDN, for better knowledge transfer from the Re-ID teacher model to the one-step person search model, we design a strong one-step person search base framework by partially disentangling the two subtasks. Besides, we propose a Knowledge Transfer Bridge module to bridge the scale gap caused by different input formats between the Re-ID model and one-step person search model. During testing, we further propose the Ranking with Context Persons strategy to exploit the context information in panoramic images for better retrieval. Experiments on two public person search datasets demonstrate the favorable performance of the proposed method. | ['Shibao Zheng', 'Qin Zhou', 'Hua Yang', 'Chuang Liu'] | 2021-08-24 | null | null | null | null | ['person-search'] | ['computer-vision'] | [-2.21423358e-01 -2.71406054e-01 -1.93263546e-01 -2.64330387e-01
-3.97084475e-01 -3.73598844e-01 9.65369403e-01 -3.07536006e-01
-8.65721226e-01 4.07466501e-01 4.06482488e-01 1.71042047e-02
-5.22399545e-01 -6.55977190e-01 -2.59398520e-01 -3.87390912e-01
6.45935535e-01 9.74619627e-01 4.31249261e-01 -2.57578701e-01
1.00728281e-01 1.83202893e-01 -1.61365509e+00 8.38137195e-02
1.30618358e+00 5.29224992e-01 4.02006090e-01 4.29116100e-01
2.00699419e-02 1.57032251e-01 -7.06424415e-01 -5.72329283e-01
1.92925498e-01 -4.16271091e-01 -9.05409992e-01 -3.48906100e-01
4.88570064e-01 -7.21210778e-01 -5.94342828e-01 1.12302864e+00
1.03930426e+00 3.58394653e-01 5.35521269e-01 -1.27872539e+00
-1.18903053e+00 4.86626923e-01 -3.94718766e-01 4.37582493e-01
7.12486625e-01 1.65594235e-01 5.73657811e-01 -8.43695819e-01
1.66275874e-01 1.75364220e+00 6.39453351e-01 8.80184829e-01
-8.62535536e-01 -9.56969142e-01 2.68451810e-01 5.76414526e-01
-1.46577680e+00 -1.47711903e-01 6.05553806e-01 -4.12953645e-01
7.16780782e-01 1.66776031e-01 6.16725922e-01 1.10796201e+00
-6.19254053e-01 8.37655663e-01 1.04726231e+00 -4.23964828e-01
-5.66551745e-01 6.38538480e-01 6.11649156e-01 5.94149113e-01
4.48978931e-01 4.30714488e-01 -5.29430866e-01 -3.14974673e-02
8.56257319e-01 2.81593144e-01 -2.97355115e-01 -1.18858017e-01
-1.32085335e+00 4.28003311e-01 5.92271864e-01 3.80947679e-01
-1.97963506e-01 -2.45273679e-01 3.82081658e-01 3.38114560e-01
1.34539291e-01 3.58197570e-01 -6.59116879e-02 2.15054348e-01
-8.27197969e-01 3.84163588e-01 6.01857066e-01 8.98688376e-01
6.51747286e-01 -4.21789289e-01 -9.09955680e-01 1.01581836e+00
3.34300429e-01 8.82128656e-01 9.38341498e-01 -2.93439925e-01
5.44755757e-01 9.16342258e-01 3.36730570e-01 -1.06657636e+00
-2.46527001e-01 -8.71537745e-01 -9.18488145e-01 -1.56514436e-01
3.14766973e-01 1.71408340e-01 -9.71644282e-01 1.84060109e+00
4.10692364e-01 7.14168698e-02 4.40068282e-02 1.29320586e+00
1.34212494e+00 3.25969070e-01 3.22530150e-01 1.47296607e-01
1.89115798e+00 -1.50409257e+00 -5.06989300e-01 -3.58395934e-01
2.85973370e-01 -5.53826094e-01 1.01962531e+00 3.71900212e-04
-9.65678096e-01 -1.32680571e+00 -9.69649851e-01 -1.96112245e-01
-6.37523830e-01 5.22300303e-01 2.55774647e-01 8.03913593e-01
-8.27183604e-01 1.57478645e-01 -1.13936111e-01 -6.01803124e-01
2.18636692e-01 6.54267311e-01 -4.15470332e-01 -2.27038488e-01
-1.65107906e+00 1.03398895e+00 4.62046951e-01 2.95273364e-01
-7.67890930e-01 -6.35827959e-01 -3.37008357e-01 3.09379935e-01
4.08297479e-01 -1.30418026e+00 1.00798094e+00 -6.93763196e-01
-1.07760060e+00 1.02620649e+00 -2.09104255e-01 -1.59753561e-02
6.56809390e-01 -3.39343399e-01 -6.04591846e-01 -2.24280432e-02
2.74615437e-01 5.58303356e-01 6.06886566e-01 -1.23397338e+00
-9.23047125e-01 -6.46389008e-01 8.80371556e-02 8.31604838e-01
-6.08681738e-01 1.39465466e-01 -7.06504524e-01 -6.09909534e-01
-3.26871097e-01 -7.60289550e-01 2.59737194e-01 -2.55599618e-01
-1.34646982e-01 -9.21227932e-01 6.07715070e-01 -8.90043795e-01
1.29013157e+00 -1.84826672e+00 2.48241857e-01 1.35529995e-01
2.50399828e-01 7.62819171e-01 -4.32596803e-01 2.57514864e-01
4.65573706e-02 -9.75100882e-03 5.52102208e-01 -4.48076367e-01
7.50329122e-02 -1.36696383e-01 3.18394378e-02 -2.22372077e-02
-4.29311424e-01 1.20476425e+00 -9.29153979e-01 -5.44774055e-01
-1.63748600e-02 3.12782824e-01 -1.95732027e-01 4.09800917e-01
3.41445535e-01 4.72171366e-01 -4.10983533e-01 4.02572066e-01
7.11544871e-01 -1.72947943e-01 -4.83101755e-01 -4.26277697e-01
5.03161699e-02 -2.90008187e-02 -1.05816042e+00 1.59677398e+00
-2.86278546e-01 6.52942732e-02 -1.14905059e-01 -7.35593259e-01
8.85822773e-01 2.18520209e-01 8.07790533e-02 -1.07207692e+00
-1.43031344e-01 1.79447606e-01 6.62946776e-02 -6.13638878e-01
4.81391668e-01 1.28660291e-01 1.07117057e-01 5.30756891e-01
-2.71924008e-02 5.65465808e-01 3.58374342e-02 1.54835299e-01
5.21688461e-01 -5.92647493e-02 1.70549154e-02 -2.98288077e-01
1.25707984e+00 -1.04756959e-01 4.35132802e-01 1.07886469e+00
-4.26831961e-01 1.71089873e-01 -3.80882084e-01 -4.80570406e-01
-7.01833785e-01 -7.94409215e-01 3.03847402e-01 1.48580039e+00
7.33123362e-01 -2.22311735e-01 -7.78477788e-01 -8.84751916e-01
3.68159451e-02 2.25849465e-01 -4.92503732e-01 -4.11591530e-01
-5.70865631e-01 -6.40575647e-01 8.10948789e-01 4.82323557e-01
1.31943536e+00 -1.11189961e+00 -8.57353210e-02 -5.75103983e-02
-5.39648771e-01 -7.59880960e-01 -1.06603277e+00 -4.62761432e-01
-3.58610243e-01 -1.06981719e+00 -1.50439394e+00 -1.34848833e+00
7.31569529e-01 5.74088037e-01 8.40154767e-01 5.10423958e-01
-6.03418332e-03 6.69373214e-01 -3.10957968e-01 -2.62399852e-01
-3.41244563e-02 4.03779060e-01 3.25847715e-01 -1.18505895e-01
8.78532350e-01 -2.82279372e-01 -1.01808059e+00 9.15113628e-01
-4.99773562e-01 3.31546336e-01 6.40524805e-01 1.02502072e+00
1.77040964e-01 2.45502934e-01 5.87980092e-01 -3.01694214e-01
1.09198213e+00 -2.11523250e-01 -2.06485569e-01 8.74541044e-01
-1.17727029e+00 -2.92691998e-02 3.96707833e-01 -8.99475932e-01
-1.45847929e+00 -4.86703277e-01 1.43956142e-02 -1.30689293e-01
-1.69682264e-01 2.90346414e-01 -3.63805890e-01 -2.12157086e-01
4.63229418e-01 7.08733082e-01 -1.67241976e-01 -7.92266071e-01
1.61215305e-01 8.90424848e-01 8.21116686e-01 -6.47938132e-01
1.07951164e+00 2.25101173e-01 -4.41650391e-01 -1.71345577e-01
-7.16827333e-01 -6.71132803e-01 -5.81598282e-01 -1.15914002e-01
9.49618280e-01 -1.14736569e+00 -1.15903127e+00 8.55444312e-01
-1.32252479e+00 -4.47110608e-02 -7.21087381e-02 4.49370027e-01
-5.27691934e-03 5.82707405e-01 -3.15631807e-01 -6.80672467e-01
-8.35637748e-01 -1.07423878e+00 1.06869829e+00 9.42311764e-01
1.63135856e-01 -8.06094825e-01 1.41200632e-01 6.82383657e-01
5.56040823e-01 -6.99995995e-01 8.46266448e-01 -1.04862130e+00
-4.99063402e-01 -3.17812204e-01 -6.69982195e-01 2.92630922e-02
9.95855555e-02 -1.03027725e+00 -7.08000302e-01 -5.20930290e-01
-2.25384951e-01 -1.84410274e-01 9.91827369e-01 -2.27151394e-01
7.68989265e-01 -3.28004152e-01 -7.50236750e-01 5.23053646e-01
1.02317953e+00 2.30827004e-01 5.63224554e-01 6.69877291e-01
9.52335358e-01 6.63157940e-01 6.20033920e-01 -5.27121015e-02
9.77140486e-01 9.96553123e-01 -2.29071051e-01 -3.23236406e-01
-6.43087268e-01 -8.15132320e-01 3.88254002e-02 4.77822483e-01
-4.87428844e-01 -2.24223599e-01 -7.30308533e-01 4.96805727e-01
-1.97083306e+00 -1.13307905e+00 1.60309628e-01 2.14392352e+00
6.41548097e-01 -3.15539300e-01 4.58944768e-01 -1.10936239e-01
1.04595447e+00 -1.54763415e-01 -5.05140364e-01 2.65915811e-01
-6.54300675e-02 -3.76327276e-01 1.55483037e-01 4.50945884e-01
-9.04807448e-01 9.94922876e-01 5.27607298e+00 1.17813003e+00
-5.80090702e-01 1.72051758e-01 1.42558053e-01 3.10933053e-01
-3.43936026e-01 -5.60129322e-02 -1.45249498e+00 6.77145183e-01
2.74078131e-01 -5.63271284e-01 5.08127272e-01 7.34243870e-01
6.90613389e-02 9.17437002e-02 -1.25248337e+00 1.46281910e+00
3.68819386e-01 -6.94035649e-01 6.05251491e-01 9.83084142e-02
4.93255377e-01 -5.96074700e-01 4.73751947e-02 7.51912594e-01
4.82535996e-02 -9.50521052e-01 4.08402473e-01 8.10202599e-01
7.28793681e-01 -5.76567411e-01 8.99196386e-01 7.51669586e-01
-1.37711227e+00 -3.69376987e-01 -3.25484693e-01 1.01888828e-01
2.40832448e-01 7.34037012e-02 -5.73829710e-01 8.29603255e-01
8.65592182e-01 3.85994941e-01 -1.08353901e+00 1.23744094e+00
-2.02861488e-01 -1.26092598e-01 -2.24125743e-01 -6.82483539e-02
-1.98350132e-01 -6.76048622e-02 5.81934512e-01 1.00208354e+00
3.81868839e-01 1.74727261e-01 3.02148342e-01 8.35616827e-01
1.14419304e-01 -2.85545457e-02 -2.32933953e-01 2.35911876e-01
6.43475354e-01 9.46594179e-01 -3.22168946e-01 -5.82705915e-01
-2.79532731e-01 1.46647048e+00 2.20459893e-01 6.34364486e-01
-5.18396020e-01 -5.17763138e-01 2.53876358e-01 1.47726029e-01
-5.31991124e-02 6.42905757e-02 1.29594997e-01 -1.09632421e+00
8.78991000e-03 -1.03684199e+00 7.12737799e-01 -8.66037011e-01
-1.55837715e+00 5.81056356e-01 2.89538831e-01 -9.25938368e-01
-1.33393243e-01 -3.57852012e-01 -6.14911914e-01 1.42504609e+00
-1.57079113e+00 -1.70296276e+00 -7.89747953e-01 9.94567394e-01
5.46444118e-01 -5.32790959e-01 6.99725568e-01 6.59505904e-01
-5.61988533e-01 1.18487227e+00 -4.26667899e-01 1.63607329e-01
9.34014857e-01 -8.37032199e-01 3.35917354e-01 7.80463219e-01
-1.23087861e-01 1.15621781e+00 3.61992389e-01 -8.89935195e-01
-9.95866120e-01 -5.36377668e-01 1.04895782e+00 -5.73600709e-01
-5.09830229e-02 -1.33746505e-01 -8.13107610e-01 4.66380119e-01
6.16052561e-02 -6.10525429e-01 4.18359667e-01 3.28777492e-01
-3.51799548e-01 -4.39600348e-01 -1.09857452e+00 6.42020047e-01
1.45374918e+00 -6.84592068e-01 -8.92800272e-01 6.62222505e-03
7.14892507e-01 -8.25363621e-02 -5.38640857e-01 5.34163237e-01
8.82061303e-01 -6.92193687e-01 1.56964922e+00 -3.23718399e-01
-1.65163487e-01 -4.52457368e-01 3.96983773e-01 -1.26483154e+00
-7.85782814e-01 -1.78698257e-01 2.58623268e-02 1.55476427e+00
-5.59739694e-02 -8.09273541e-01 6.76621974e-01 7.55468488e-01
2.95147508e-01 -2.04726100e-01 -8.69188845e-01 -9.27901447e-01
-3.68417874e-02 4.14425671e-01 1.10519993e+00 6.87853813e-01
-1.08375058e-01 6.19757652e-01 -5.70737898e-01 2.62494147e-01
5.93034625e-01 -8.06923658e-02 1.02429783e+00 -1.54829788e+00
-3.12243521e-01 -5.78328252e-01 -3.13270651e-02 -1.61524951e+00
1.67605970e-02 -9.22282934e-01 -1.40245765e-01 -1.81647122e+00
8.91588449e-01 -6.79655731e-01 -5.56461334e-01 3.71594250e-01
-7.67009735e-01 -9.65740085e-02 2.28694484e-01 6.41989231e-01
-6.98624909e-01 6.31794989e-01 1.46176374e+00 -2.61627555e-01
-2.53870249e-01 1.28610253e-01 -9.19567883e-01 3.48557442e-01
3.91988218e-01 -2.41082311e-01 -6.94579720e-01 -8.23937714e-01
1.00839570e-01 -6.15813993e-02 7.93601990e-01 -9.49683070e-01
9.72625732e-01 1.24864735e-01 6.03034735e-01 -7.08298266e-01
2.02356234e-01 -6.82907701e-01 -1.87670868e-02 4.77217495e-01
-2.72327393e-01 1.18710391e-01 -3.92511599e-02 5.62928736e-01
-1.30189881e-01 -2.51272172e-01 3.52487147e-01 -3.59789908e-01
-8.71387482e-01 4.53312337e-01 9.96503532e-02 -1.45521760e-01
7.32288063e-01 -3.52079093e-01 -6.68004632e-01 -3.41065586e-01
-5.31566024e-01 7.43455470e-01 2.18085632e-01 7.01890886e-01
5.54846048e-01 -1.44007158e+00 -6.49575531e-01 2.98828542e-01
1.77660972e-01 -2.99406558e-01 5.86282253e-01 5.72254717e-01
-1.99425202e-02 6.74461663e-01 -2.31773615e-01 -2.92447329e-01
-1.54809546e+00 7.51965821e-01 5.34872949e-01 -5.94579101e-01
-3.36381674e-01 9.88556862e-01 4.24884200e-01 -6.56922340e-01
5.22105634e-01 5.13479829e-01 -6.56355560e-01 -3.07076909e-02
7.51743257e-01 5.74334979e-01 -4.30129141e-01 -7.16458738e-01
-4.49606448e-01 9.64412689e-01 -3.84608597e-01 -2.31709570e-01
6.82431400e-01 -4.39003587e-01 -7.62432590e-02 -2.10062623e-01
7.69158065e-01 -1.87320635e-01 -7.64728487e-01 -4.56587583e-01
-2.47399315e-01 -4.52608973e-01 -3.09969723e-01 -1.26013815e+00
-6.09252810e-01 7.44122207e-01 1.02905548e+00 -1.41835973e-01
1.08394325e+00 1.19419940e-01 9.20535982e-01 4.93931592e-01
4.02471572e-01 -1.07447386e+00 3.05842191e-01 3.44331563e-01
9.01645899e-01 -1.37264645e+00 2.19061747e-02 -3.44415903e-01
-3.27800274e-01 7.25457668e-01 1.07632422e+00 2.06060618e-01
4.13784057e-01 -5.23070812e-01 -2.47581638e-02 -1.74153656e-01
-1.02462478e-01 -5.12467384e-01 8.55980158e-01 8.24536860e-01
-1.18062951e-01 -1.51813418e-01 -5.75721562e-01 1.23206067e+00
-2.22085595e-01 2.60538518e-01 -3.40107858e-01 3.01070213e-01
-3.44161183e-01 -1.27684164e+00 -5.01790643e-01 7.71792158e-02
1.04804717e-01 -1.94689274e-01 -3.98419589e-01 6.93937659e-01
5.66267669e-01 1.01734543e+00 -2.37375349e-01 -6.40443623e-01
5.77616513e-01 7.10769147e-02 5.46073020e-01 -3.27446729e-01
-8.12019765e-01 -4.24459606e-01 -2.51264460e-02 -1.65864170e-01
-4.78425115e-01 -9.85206589e-02 -6.73781991e-01 -2.59134650e-01
-3.95136714e-01 4.25166190e-01 3.01282763e-01 1.05730522e+00
3.15975666e-01 3.41568977e-01 2.17353478e-01 -5.78729331e-01
-5.75101316e-01 -1.11012423e+00 -2.25042373e-01 6.23195052e-01
9.95362177e-02 -7.69359827e-01 -2.91693598e-01 -3.36764336e-01] | [14.820977210998535, 0.798926055431366] |
a75da425-4d89-4474-90ae-ee81ac9680c2 | ptt-point-track-transformer-module-for-3d | 2108.06455 | null | https://arxiv.org/abs/2108.06455v3 | https://arxiv.org/pdf/2108.06455v3.pdf | PTT: Point-Track-Transformer Module for 3D Single Object Tracking in Point Clouds | 3D single object tracking is a key issue for robotics. In this paper, we propose a transformer module called Point-Track-Transformer (PTT) for point cloud-based 3D single object tracking. PTT module contains three blocks for feature embedding, position encoding, and self-attention feature computation. Feature embedding aims to place features closer in the embedding space if they have similar semantic information. Position encoding is used to encode coordinates of point clouds into high dimension distinguishable features. Self-attention generates refined attention features by computing attention weights. Besides, we embed the PTT module into the open-source state-of-the-art method P2B to construct PTT-Net. Experiments on the KITTI dataset reveal that our PTT-Net surpasses the state-of-the-art by a noticeable margin (~10%). Additionally, PTT-Net could achieve real-time performance (~40FPS) on NVIDIA 1080Ti GPU. Our code is open-sourced for the robotics community at https://github.com/shanjiayao/PTT. | ['Yubo Cui', 'Zheng Fang', 'Sifan Zhou', 'Jiayao Shan'] | 2021-08-14 | null | null | null | null | ['3d-single-object-tracking'] | ['computer-vision'] | [-3.65637571e-01 -1.87637240e-01 8.01149532e-02 -6.72867522e-02
-6.20123267e-01 -3.94941807e-01 4.68366504e-01 7.07577094e-02
-3.32903653e-01 1.82705373e-02 -1.19758314e-02 -1.54193074e-01
9.51250829e-03 -7.49453723e-01 -1.16415203e+00 -4.88472074e-01
-2.10735753e-01 2.94959277e-01 4.94917691e-01 -1.61711693e-01
2.59287089e-01 7.81459570e-01 -1.72697198e+00 -1.64382439e-02
5.27006149e-01 1.09861720e+00 5.87191999e-01 6.41090035e-01
-3.46946232e-02 4.12492901e-01 -2.72821248e-01 -2.27061242e-01
5.78502834e-01 3.59826565e-01 -4.91272151e-01 -2.47880712e-01
5.20349979e-01 -2.65148610e-01 -7.79762506e-01 1.18924046e+00
4.78794843e-01 -5.58137819e-02 3.08709025e-01 -1.57691061e+00
-1.07648253e+00 1.92663878e-01 -6.41431689e-01 2.65471846e-01
1.43334612e-01 5.92289090e-01 8.63921881e-01 -1.21640706e+00
4.15010452e-01 1.46762478e+00 8.30051422e-01 4.65637416e-01
-8.95791471e-01 -7.60280013e-01 1.64403319e-01 2.08671287e-01
-1.26736128e+00 -4.49635759e-02 7.55407691e-01 -5.06688476e-01
1.34000754e+00 1.63613886e-01 1.05468261e+00 7.54986763e-01
5.40575087e-01 8.77580702e-01 6.45621896e-01 8.63498896e-02
-1.20478362e-01 -2.93130130e-01 2.86289036e-01 7.79260814e-01
3.94944638e-01 3.13408196e-01 -3.56382787e-01 -1.09554932e-01
1.15844953e+00 3.82619292e-01 -1.36761442e-01 -5.56325495e-01
-1.66670549e+00 6.91048086e-01 1.18779123e+00 1.76475104e-02
-4.40818876e-01 6.71647370e-01 3.84086221e-01 1.16216607e-01
4.46107268e-01 2.04458892e-01 -3.00629199e-01 -1.68074340e-01
4.55198176e-02 3.84682387e-01 2.65625000e-01 1.64422607e+00
6.65407240e-01 -8.02186355e-02 -2.41646275e-01 3.89637262e-01
5.83182991e-01 7.71238387e-01 4.93624300e-01 -9.81624722e-01
3.54649663e-01 7.65534282e-01 1.76956967e-01 -1.12678826e+00
-2.78520167e-01 -2.87188351e-01 -4.34156209e-01 4.98499870e-01
-3.05743236e-02 6.05408065e-02 -1.00987232e+00 1.32595360e+00
6.28125608e-01 3.53254527e-01 -1.54557899e-01 1.36224473e+00
1.02555037e+00 7.77222514e-01 -1.19926505e-01 5.87794721e-01
1.39234102e+00 -1.27104819e+00 -3.07647109e-01 -2.48475745e-01
4.04423177e-01 -5.45716405e-01 9.66637015e-01 -1.54929385e-01
-9.56172943e-01 -8.96663249e-01 -9.84470785e-01 -5.37749290e-01
-2.81118274e-01 1.84959769e-01 7.63645768e-01 1.94253892e-01
-9.83234704e-01 7.17317462e-01 -1.37147582e+00 -3.80468130e-01
6.75649166e-01 5.22400618e-01 -4.98254806e-01 5.54414652e-03
-5.34184396e-01 8.45037043e-01 1.14401408e-01 1.14639312e-01
-8.40801954e-01 -8.74487758e-01 -1.08577096e+00 8.71655867e-02
2.28668489e-02 -9.27900314e-01 1.26917005e+00 -1.70212552e-01
-1.57171619e+00 9.44647849e-01 -1.00174300e-01 -3.15558851e-01
3.18873316e-01 -5.74812174e-01 1.82419196e-02 -2.71556322e-02
2.53842473e-01 9.42646503e-01 8.28814507e-01 -1.06730783e+00
-6.80272818e-01 -5.75223446e-01 -9.01803151e-02 2.17398211e-01
-6.34109452e-02 -3.90795246e-02 -7.12538600e-01 -4.63201612e-01
3.34721327e-01 -1.16889513e+00 -2.44192153e-01 7.17373013e-01
-2.70247132e-01 -5.53971767e-01 1.21531379e+00 -1.34076387e-01
3.21343154e-01 -2.41190982e+00 8.62256438e-02 -4.02808785e-01
5.63062966e-01 2.47143373e-01 -1.24067441e-01 7.00592771e-02
6.94735125e-02 -3.12315583e-01 5.47688529e-02 -4.02136177e-01
3.02844286e-01 1.47889197e-01 -2.93183684e-01 7.71294415e-01
6.36375785e-01 1.30237114e+00 -1.01544023e+00 -3.20058644e-01
6.75029516e-01 8.55900466e-01 -6.35161042e-01 1.08401164e-01
-1.02426134e-01 2.64305830e-01 -7.68484592e-01 7.01673746e-01
7.83158660e-01 -3.50348771e-01 -5.18076122e-01 -3.11464459e-01
-5.74943125e-01 2.90135294e-01 -6.75268173e-01 1.98466170e+00
-9.78636593e-02 7.18148530e-01 -2.73479491e-01 -5.05514324e-01
1.16565728e+00 -2.03619823e-01 5.49584746e-01 -3.07747990e-01
5.13306081e-01 6.72139227e-02 -1.06664419e-01 -3.21212023e-01
8.53742719e-01 4.14706945e-01 -1.70211554e-01 9.75433588e-02
1.12652667e-01 -1.58333644e-01 -3.01467896e-01 -3.71163152e-02
1.17664421e+00 4.66153502e-01 6.37524202e-02 -3.46945167e-01
1.91846341e-01 2.43696183e-01 5.68550885e-01 4.29413438e-01
-5.51422715e-01 5.75154185e-01 5.70000820e-02 -5.26907623e-01
-1.17018557e+00 -1.13711917e+00 -1.50268093e-01 8.63686323e-01
6.14470899e-01 -3.27055156e-01 -3.04605514e-01 -4.87861335e-01
6.54192269e-01 2.55160958e-01 -6.31207526e-01 -2.42452592e-01
-4.64608580e-01 -1.78026766e-01 2.03763932e-01 9.15465891e-01
4.51023012e-01 -9.41586912e-01 -1.18737197e+00 2.66523868e-01
2.40235761e-01 -1.02801132e+00 -5.35831451e-01 2.29832664e-01
-1.06354988e+00 -8.19026291e-01 -7.56521881e-01 -9.57095563e-01
7.26958811e-01 8.17896962e-01 7.78912306e-01 -8.39690864e-02
-2.13900357e-01 3.57245803e-01 -4.62171733e-01 -7.78898835e-01
2.49857590e-01 -4.21788022e-02 3.55079211e-02 -4.56874162e-01
7.40170777e-01 -4.24762070e-01 -7.52038956e-01 2.11257964e-01
-3.64397049e-01 1.12302393e-01 5.66872954e-01 7.05768406e-01
9.58658814e-01 -4.75385994e-01 -8.11002329e-02 -1.11702129e-01
1.26953647e-01 -3.83655787e-01 -8.63722682e-01 -2.54388660e-01
1.15363477e-02 1.51866436e-04 1.66289419e-01 -6.58872426e-01
-3.74475688e-01 3.50952804e-01 -2.06470996e-01 -1.22229385e+00
4.38680593e-03 4.86394353e-02 4.17149588e-02 -5.73565602e-01
2.13793904e-01 1.26960203e-01 2.48317838e-01 -5.58468997e-01
2.79555947e-01 6.20329142e-01 7.24089801e-01 -4.26348209e-01
8.34083617e-01 5.86303651e-01 -2.07375899e-01 -4.88908499e-01
-5.92153847e-01 -4.42436159e-01 -6.10082269e-01 -2.21233219e-01
8.57190907e-01 -1.11876333e+00 -1.16805637e+00 5.67712307e-01
-1.31382954e+00 -3.30893636e-01 -4.09725308e-01 5.69196522e-01
-7.16514826e-01 1.25491675e-02 -6.58963621e-01 -5.47511935e-01
-6.45882905e-01 -1.34848583e+00 1.62861240e+00 4.06121314e-01
1.81744933e-01 -4.52150941e-01 -4.47429297e-03 -5.63966259e-02
2.03351542e-01 3.13078314e-01 3.92543048e-01 -2.69710839e-01
-9.24829900e-01 -3.18642080e-01 -6.29085600e-01 -1.43563539e-01
3.89615707e-02 -7.07389861e-02 -9.60484087e-01 -4.99262035e-01
3.05259880e-02 -1.37946662e-02 7.09196270e-01 4.46750075e-01
1.20759773e+00 7.75589794e-02 -5.89808106e-01 8.76466036e-01
1.48590374e+00 1.64076105e-01 3.74651164e-01 5.13070226e-01
9.75284517e-01 -2.33950671e-02 7.00768948e-01 3.52261662e-01
8.17019880e-01 7.60365486e-01 8.19358885e-01 6.41626120e-02
-1.42544225e-01 -3.38045865e-01 3.05957347e-01 9.87743616e-01
-1.26294326e-02 1.29589349e-01 -9.58017051e-01 7.53917217e-01
-2.04718566e+00 -6.96550369e-01 -1.73925653e-01 1.91224349e+00
2.43350476e-01 2.39546135e-01 -7.00287707e-03 -1.31818771e-01
7.93420017e-01 1.32976929e-02 -8.29392731e-01 -6.34271353e-02
2.85788268e-01 -6.39327019e-02 6.03305161e-01 2.13464811e-01
-1.28323066e+00 9.89253700e-01 4.79932928e+00 4.07499701e-01
-1.34265697e+00 2.68472910e-01 5.66473380e-02 -1.40281275e-01
1.23466654e-02 -2.27262884e-01 -9.17095244e-01 5.64489305e-01
5.94148576e-01 -4.25802380e-01 -1.25355786e-04 1.31501245e+00
-3.09586436e-01 3.75794590e-01 -9.66960549e-01 1.04547298e+00
-9.53741893e-02 -1.27148545e+00 -1.50137991e-01 1.61235958e-01
3.21592182e-01 8.65310192e-01 7.44877905e-02 4.92001086e-01
3.55133355e-01 -6.19873822e-01 9.78103042e-01 2.40360096e-01
6.53802514e-01 -7.34136999e-01 6.46452129e-01 2.76083369e-02
-1.60012782e+00 -1.17375955e-01 -9.43352401e-01 -9.91307870e-02
6.86171502e-02 4.09098953e-01 -5.58880866e-01 5.68996429e-01
1.21275878e+00 1.22959518e+00 -4.38389063e-01 1.41477072e+00
-3.77503410e-03 -6.11103922e-02 -6.15167022e-01 -3.03589493e-01
4.84486759e-01 7.89391845e-02 8.55640233e-01 8.32869232e-01
6.05269969e-01 1.64461508e-01 2.12756291e-01 1.03040969e+00
1.31382728e-02 -3.59972328e-01 -8.32090020e-01 8.03968087e-02
5.00431180e-01 1.17159593e+00 -7.11688578e-01 -1.98833302e-01
-4.07788813e-01 1.07257783e+00 4.66944635e-01 -9.44385976e-02
-9.94864702e-01 -7.25726426e-01 1.27751255e+00 -1.81141227e-01
8.56298149e-01 -5.21638572e-01 -3.02487910e-01 -9.91768360e-01
9.00452733e-02 -1.86038226e-01 -6.05574884e-02 -1.17616320e+00
-1.32214940e+00 8.92659485e-01 -3.04206610e-01 -1.66232622e+00
4.72732782e-01 -9.58060801e-01 -4.62582350e-01 8.41933608e-01
-1.52779245e+00 -1.13863087e+00 -7.80261815e-01 5.05221725e-01
5.86826265e-01 1.28251880e-01 6.34673655e-01 1.99762717e-01
-3.80393565e-01 3.28220725e-01 -1.65426899e-02 2.24113896e-01
2.49532059e-01 -1.06168067e+00 1.24973488e+00 6.06570005e-01
7.45195313e-04 7.04711735e-01 3.63601208e-01 -7.47237980e-01
-2.01289511e+00 -1.41150689e+00 6.44317567e-01 -7.38361001e-01
7.82819808e-01 -4.23238456e-01 -8.51403713e-01 8.96439850e-01
-1.35017289e-02 6.51646495e-01 1.43537849e-01 -3.35239589e-01
-2.69931942e-01 6.44341260e-02 -1.10569406e+00 5.89183092e-01
1.38423359e+00 -4.25234348e-01 -7.62794316e-01 2.80928701e-01
1.39996576e+00 -1.01702642e+00 -1.01009595e+00 4.53683138e-01
6.18848681e-01 -5.65615654e-01 1.14680326e+00 -2.97160238e-01
3.63459080e-01 -6.12232327e-01 -2.09534019e-01 -1.17661035e+00
-8.79332066e-01 -4.87806648e-01 -2.41198391e-01 7.47694552e-01
1.93030704e-02 -7.38255501e-01 7.35338509e-01 3.49600911e-01
-6.92647815e-01 -8.44027460e-01 -1.09639621e+00 -9.40063059e-01
-2.34623905e-02 -4.91788626e-01 9.32503581e-01 6.86699569e-01
-1.07557224e-02 6.66913518e-04 1.09738506e-01 5.10680735e-01
7.74867892e-01 4.13112611e-01 9.52188492e-01 -1.20124006e+00
7.06832707e-02 -4.07587498e-01 -1.08740938e+00 -1.44256508e+00
3.67144905e-02 -9.95227635e-01 2.97364831e-01 -1.53189921e+00
3.73588726e-02 -5.60557485e-01 -3.18196356e-01 7.25647748e-01
-1.20156787e-01 3.88062328e-01 6.60355031e-01 2.36294985e-01
-5.31457365e-01 1.09169817e+00 1.36776721e+00 -2.63058960e-01
-4.24988829e-02 -2.73743063e-01 -6.20604634e-01 5.42781651e-01
6.40975595e-01 -6.69238269e-01 6.21138550e-02 -1.03122818e+00
-3.65963489e-01 -1.99967444e-01 8.00450206e-01 -1.44271684e+00
4.17260468e-01 1.47496939e-01 3.97123992e-01 -1.12483323e+00
6.35198355e-01 -1.09812391e+00 8.12351704e-02 7.07482696e-01
8.06892589e-02 4.69013691e-01 6.42283440e-01 5.78563631e-01
1.75804839e-01 1.31633177e-01 4.66129810e-01 1.64139811e-02
-9.87322867e-01 6.86547756e-01 -1.89105533e-02 -3.63784403e-01
1.15110493e+00 -2.67062157e-01 -5.79646111e-01 2.20393002e-01
-2.86401004e-01 4.56571579e-01 5.73888481e-01 8.47939312e-01
8.71070981e-01 -1.71266007e+00 -5.19273102e-01 3.50246310e-01
3.59533966e-01 5.16479254e-01 1.76777050e-01 6.80777490e-01
-6.53799832e-01 5.64038098e-01 -6.31958544e-01 -1.21546745e+00
-1.05976164e+00 5.84215522e-01 4.07590568e-02 3.67694795e-01
-1.39732957e+00 1.04062355e+00 1.18256696e-01 -6.60807133e-01
1.46958366e-01 -8.01537573e-01 2.25078344e-01 -6.19308949e-01
4.84854430e-01 1.39176384e-01 -1.24531448e-01 -6.64926767e-01
-5.57730973e-01 9.89998162e-01 -5.88737847e-03 3.50212127e-01
1.58811259e+00 1.02852724e-01 1.39949486e-01 3.84427309e-01
1.26902938e+00 -4.74818707e-01 -1.75921285e+00 -1.83782682e-01
-2.27816373e-01 -8.18850935e-01 2.39456117e-01 -1.71726108e-01
-1.17133880e+00 8.72977257e-01 8.06004763e-01 -5.51818013e-02
7.13330626e-01 1.22082911e-01 9.96949434e-01 4.68692929e-01
8.54693532e-01 -3.37542236e-01 4.53531556e-02 7.07830727e-01
1.07959390e+00 -1.31795251e+00 -5.12284860e-02 -4.70579147e-01
-5.49486339e-01 7.85196006e-01 8.95988882e-01 -7.69589603e-01
6.84059620e-01 3.91959608e-01 -5.17830439e-02 -4.12039906e-01
-6.81267262e-01 -3.58992487e-01 3.51037025e-01 7.40937352e-01
7.32850209e-02 -4.55139875e-02 2.11565197e-01 5.68854332e-01
-4.03671294e-01 -1.39778173e-02 1.12758912e-01 1.14787579e+00
-4.17164534e-01 -4.89239156e-01 -2.45107606e-01 2.61340588e-01
1.12833958e-02 1.69191688e-01 -1.27526611e-01 7.82509029e-01
-8.39255378e-02 4.15502757e-01 5.32237649e-01 -7.80732632e-01
5.80017090e-01 -3.28543395e-01 5.20691574e-01 -4.94491756e-01
-6.99667156e-01 -1.11935791e-02 -4.52964395e-01 -8.89966965e-01
-2.83574760e-01 -6.19525969e-01 -1.33069956e+00 -3.54582429e-01
-2.46616647e-01 -1.07778788e-01 8.74314845e-01 4.75896597e-01
9.38726544e-01 7.14453042e-01 4.51528192e-01 -1.64879429e+00
-2.59691179e-01 -8.42369854e-01 -6.83682263e-02 2.11098477e-01
4.86313105e-01 -1.06712890e+00 7.83589035e-02 -2.22567096e-01] | [6.714693069458008, -2.4302709102630615] |
3969aac5-0ad8-4483-a656-22f0d215628e | time-out-of-mind-generating-emotionally | 2301.12331 | null | https://arxiv.org/abs/2301.12331v2 | https://arxiv.org/pdf/2301.12331v2.pdf | Time out of Mind: Generating Rate of Speech conditioned on emotion and speaker | Voice synthesis has seen significant improvements in the past decade resulting in highly intelligible voices. Further investigations have resulted in models that can produce variable speech, including conditional emotional expression. The problem lies, however, in a focus on phrase-level modifications and prosodic vocal features. Using the CREMA-D dataset we have trained a GAN conditioned on emotion to generate worth lengths for a given input text. These word lengths are relative to neutral speech and can be provided, through speech synthesis markup language (SSML) to a text-to-speech (TTS) system to generate more expressive speech. Additionally, a generative model is also trained using implicit maximum likelihood estimation (IMLE) and a comparative analysis with GANs is included. We were able to achieve better performances on objective measures for neutral speech, and better time alignment for happy speech when compared to an out-of-box model. However, further investigation of subjective evaluation is required. | ['Paige Tuttosi', 'Navjot Kaur'] | 2023-01-29 | null | null | null | null | ['speech-synthesis'] | ['speech'] | [ 2.55826294e-01 7.22881377e-01 1.02866642e-01 -6.01489604e-01
-1.00781631e+00 -4.82526034e-01 7.88461328e-01 -4.61682260e-01
-6.22709766e-02 9.27948236e-01 6.36412144e-01 -4.09076223e-03
3.51235092e-01 -4.39184725e-01 -2.17003390e-01 -7.07237959e-01
1.02859087e-01 5.00399590e-01 -2.74924994e-01 -3.71879905e-01
-2.67828643e-01 4.77347523e-01 -1.47197890e+00 3.32865745e-01
5.46364546e-01 6.14651322e-01 2.34075606e-01 1.12516522e+00
-1.47517711e-01 5.60201168e-01 -1.19299042e+00 -3.09134185e-01
-3.71023677e-02 -9.04578269e-01 -4.58855629e-01 2.52234221e-01
-2.79637128e-02 8.97836387e-02 2.92798668e-01 7.68912077e-01
1.01456237e+00 2.95103103e-01 6.68797433e-01 -1.08845830e+00
-4.79157656e-01 7.41240978e-01 2.88371090e-02 -3.97956640e-01
6.11253142e-01 2.31491044e-01 9.15332496e-01 -7.21049905e-01
8.37085903e-01 1.36319613e+00 2.24776313e-01 1.16425145e+00
-1.45926321e+00 -5.47997415e-01 -3.19532722e-01 -3.76333952e-01
-1.09379840e+00 -9.34305131e-01 8.84025037e-01 -2.91683167e-01
1.32807040e+00 7.25529790e-01 6.77254796e-01 1.43688416e+00
5.17929569e-02 5.14097452e-01 1.19371724e+00 -7.10316241e-01
3.50531250e-01 5.11490583e-01 -7.07606435e-01 2.02267095e-01
-7.55740523e-01 4.08195406e-01 -3.91884327e-01 2.35907704e-01
4.31245893e-01 -8.28104436e-01 -3.57506692e-01 3.35948706e-01
-9.97119129e-01 9.56051230e-01 6.53061643e-02 8.08486044e-01
-4.58437711e-01 9.55682341e-03 5.49298227e-01 4.67391968e-01
7.44768023e-01 8.03874612e-01 -2.68781513e-01 -5.32280147e-01
-1.40288377e+00 2.90767461e-01 1.03130078e+00 7.65280902e-01
3.47320795e-01 7.88612366e-01 -3.82234305e-01 1.22632980e+00
2.47499809e-01 5.86719632e-01 6.38415396e-01 -1.10569859e+00
1.34451434e-01 -1.38336912e-01 -2.19476312e-01 -5.14511287e-01
-2.66753465e-01 -4.24904555e-01 -7.37997293e-01 6.59535944e-01
-1.43689647e-01 -6.94620609e-01 -1.00912201e+00 1.96226907e+00
-3.05042360e-02 -1.35243550e-01 5.75348556e-01 8.36809874e-01
6.95478559e-01 1.22330284e+00 -2.00926103e-02 -6.82790339e-01
8.84757102e-01 -8.79080117e-01 -1.35414410e+00 -2.15502262e-01
4.36001658e-01 -1.06462574e+00 1.20421267e+00 4.40196067e-01
-1.46200871e+00 -5.98125756e-01 -9.79151487e-01 1.50615871e-01
-2.30466977e-01 1.32338658e-01 1.78835049e-01 1.19308317e+00
-1.39723194e+00 3.77512485e-01 -6.75039530e-01 -3.85694057e-01
-1.64398313e-01 3.56696844e-01 -2.78544426e-01 7.62470305e-01
-1.29640758e+00 1.02748656e+00 2.40326777e-01 -4.41305619e-03
-7.39501476e-01 -4.08026755e-01 -1.07351434e+00 -8.75952095e-02
-1.38098255e-01 -6.19905770e-01 1.45258963e+00 -1.35851967e+00
-2.49807692e+00 7.67276943e-01 -1.43146068e-01 -5.43745935e-01
4.99209166e-01 1.09053843e-01 -7.32932150e-01 -4.29596230e-02
-3.41147572e-01 1.21632183e+00 9.70948696e-01 -1.38318574e+00
-3.19265783e-01 3.15058231e-01 -4.06132460e-01 2.66825140e-01
2.67970897e-02 4.52685058e-01 -1.12752154e-01 -9.32548225e-01
-5.71849167e-01 -1.09285951e+00 -1.54105023e-01 -4.64578092e-01
-5.03575385e-01 -1.52794532e-02 8.00432563e-01 -9.51820552e-01
1.17889690e+00 -2.00666738e+00 5.42221248e-01 2.57711541e-02
-4.97147977e-01 4.03105110e-01 -1.37769535e-01 3.48002583e-01
-3.94600958e-01 2.93561667e-01 -3.00027102e-01 -8.55393529e-01
2.35746562e-01 4.20257360e-01 -5.18454909e-01 -3.45022790e-02
4.31819648e-01 7.38059044e-01 -4.10759211e-01 -4.21214134e-01
4.67798650e-01 6.82207108e-01 -7.16163099e-01 4.97110903e-01
-4.93944377e-01 6.60237372e-01 1.04660697e-01 3.25280190e-01
2.67168134e-01 7.89704502e-01 -1.70057252e-01 8.92620459e-02
-3.28224629e-01 2.85620362e-01 -7.57570267e-01 1.61199856e+00
-9.80760455e-01 7.59648561e-01 3.04013938e-01 -4.09204185e-01
1.22856426e+00 1.00889111e+00 9.19373557e-02 -5.09254515e-01
2.75538981e-01 3.35894793e-01 2.15894833e-01 -3.78646940e-01
5.81090391e-01 -7.16843307e-01 -1.10179991e-01 7.98778459e-02
2.76679963e-01 -9.39897954e-01 1.57381400e-01 -2.18819991e-01
4.88153249e-01 1.47478521e-01 2.34048814e-02 8.48635659e-03
5.58028400e-01 -3.70646507e-01 2.73277313e-01 1.30892217e-01
3.01054232e-02 8.72105241e-01 5.11747420e-01 4.13131863e-01
-1.11364639e+00 -9.14860964e-01 5.33679649e-02 6.62461638e-01
-7.34611928e-01 -4.30081457e-01 -1.12207055e+00 -2.62373716e-01
-6.62584007e-01 1.55990374e+00 -2.89209217e-01 -3.23186032e-02
-5.66663742e-01 -4.54930604e-01 6.96242332e-01 2.38336056e-01
-7.07328692e-02 -1.70735800e+00 -3.16275626e-01 4.47601795e-01
-7.40957186e-02 -1.00175858e+00 -4.29757088e-01 3.99031222e-01
-6.17833912e-01 4.47102487e-02 -1.00547910e+00 -7.38048851e-01
2.15199053e-01 -9.01123047e-01 9.03743923e-01 -4.68256325e-01
-8.59658197e-02 2.09005341e-01 -4.51898962e-01 -6.40352309e-01
-1.29349971e+00 1.03736877e-01 1.22524582e-01 -1.93797186e-01
-2.85747856e-01 -6.86861515e-01 -9.12682414e-02 1.68234780e-01
-9.68547165e-01 2.32850388e-01 5.21656215e-01 7.82346725e-01
3.65941972e-01 -1.04793064e-01 1.00380743e+00 -6.78096592e-01
1.02051854e+00 -3.41334641e-02 -2.94049263e-01 -2.60259658e-01
-5.49057066e-01 9.47180986e-02 6.25938594e-01 -4.66283202e-01
-1.63608909e+00 -5.61578199e-02 -1.07497740e+00 -3.59845787e-01
-4.59525734e-01 3.68605644e-01 -4.36537921e-01 3.51010084e-01
6.40052319e-01 3.73297445e-02 1.50521696e-01 -2.15480760e-01
6.15764022e-01 1.14655530e+00 5.98387361e-01 -4.74973023e-01
5.84082484e-01 -1.99490041e-01 -2.91107833e-01 -1.25384808e+00
-3.39481205e-01 9.26764458e-02 -3.31775755e-01 -3.53804618e-01
1.03521442e+00 -7.24348128e-01 -1.91810489e-01 4.41886008e-01
-1.10216022e+00 -6.71398938e-01 -6.91368282e-01 4.41397339e-01
-9.75926936e-01 -6.75977543e-02 -7.17662752e-01 -1.16834879e+00
-5.19908130e-01 -1.19917607e+00 1.02544296e+00 2.31795758e-01
-8.18902194e-01 -9.89877462e-01 3.27221006e-01 3.61246049e-01
6.13880038e-01 2.31978253e-01 8.04682136e-01 -5.51511049e-01
2.17409078e-02 4.34167404e-03 4.97066826e-01 9.19407547e-01
3.15587014e-01 3.10816139e-01 -1.22207487e+00 -1.46440834e-01
1.09600984e-01 -3.79718989e-01 3.78557861e-01 4.25410062e-01
5.89906096e-01 -4.95088428e-01 1.51922375e-01 4.09865558e-01
8.19024026e-01 6.03233635e-01 8.81320894e-01 -1.87797129e-01
2.00526655e-01 7.25508571e-01 6.10635638e-01 2.59743184e-01
-2.71255702e-01 7.66954660e-01 1.20938413e-01 -4.06162798e-01
-5.44289291e-01 -1.31782055e-01 7.51122594e-01 1.15162206e+00
1.16644517e-01 -6.97656810e-01 -3.80820423e-01 5.06071627e-01
-1.29831004e+00 -1.03724205e+00 3.28497924e-02 1.85493100e+00
1.14641535e+00 2.55347729e-01 1.41461357e-01 2.80030251e-01
6.48844659e-01 3.90210927e-01 -1.68848395e-01 -1.23359370e+00
-2.20621571e-01 5.86505711e-01 1.91807970e-02 9.29340899e-01
-6.56649530e-01 1.06513202e+00 6.55058098e+00 9.02092040e-01
-1.57840252e+00 3.78495902e-02 6.71878397e-01 -2.51453757e-01
-4.45241034e-01 -2.01696694e-01 -5.45138001e-01 4.78898287e-01
1.53345823e+00 -2.19805509e-01 5.84175587e-01 6.91847324e-01
7.19489813e-01 7.71064684e-02 -8.58591318e-01 7.43761718e-01
1.48962080e-01 -9.84675586e-01 -3.10974475e-02 -1.67234585e-01
7.60586858e-01 -2.89450258e-01 2.25812227e-01 5.94315350e-01
-1.78356573e-01 -1.20504117e+00 8.40750515e-01 5.59775531e-01
1.08425641e+00 -1.08465409e+00 6.86671555e-01 3.47352237e-01
-8.06297302e-01 5.16622305e-01 2.46165786e-02 1.18792854e-01
6.39849424e-01 2.96214759e-01 -1.52730060e+00 4.30462897e-01
2.32790992e-01 1.77415788e-01 -6.48939461e-02 5.56371450e-01
-4.42033440e-01 9.59780872e-01 -1.63528174e-01 -2.64691085e-01
1.50135264e-01 -1.11271530e-01 8.83139014e-01 1.47031188e+00
6.38599813e-01 -1.81501687e-01 -2.55864918e-01 1.05783212e+00
5.43298014e-02 3.76054347e-01 -3.65206659e-01 -5.55306375e-01
5.16631734e-03 1.38215125e+00 -3.97549212e-01 -1.50432467e-01
1.18586386e-03 1.31329918e+00 -1.87721148e-01 3.40373397e-01
-6.91066682e-01 -5.08651674e-01 6.94366872e-01 1.04441522e-02
1.21211261e-01 -1.49540171e-01 -1.66153125e-02 -6.68831587e-01
-3.61456990e-01 -1.19242358e+00 -1.86013475e-01 -1.26450682e+00
-1.02466369e+00 1.06000030e+00 -7.68260360e-02 -8.46707463e-01
-1.07001162e+00 -3.52561414e-01 -7.49532819e-01 1.34775293e+00
-1.01080680e+00 -9.89727378e-01 2.21017301e-01 1.48377970e-01
9.27685082e-01 -2.32510611e-01 1.25439680e+00 2.90275484e-01
-3.00086588e-01 5.52550733e-01 -2.80682325e-01 -2.02823520e-01
8.57691467e-01 -1.38674772e+00 4.01987255e-01 5.95888972e-01
2.00954959e-01 2.38622159e-01 1.37276566e+00 -5.02881825e-01
-7.93732226e-01 -1.00560164e+00 1.14862859e+00 -1.13713056e-01
3.39601904e-01 -4.87791508e-01 -7.06465960e-01 5.39575934e-01
8.77721190e-01 -6.22034311e-01 8.26953113e-01 -1.96430013e-01
3.93379062e-01 1.59623474e-01 -1.28237760e+00 6.50987327e-01
5.09491265e-01 -5.14100790e-01 -6.64869785e-01 -7.35572800e-02
1.03122926e+00 -5.10703087e-01 -8.47089887e-01 3.55923384e-01
1.63049683e-01 -9.17314053e-01 4.57564265e-01 -2.29270279e-01
3.09559673e-01 -3.18168312e-01 -1.80748329e-02 -1.94225562e+00
1.34300128e-01 -1.10976970e+00 2.97929049e-01 1.87812757e+00
8.36811244e-01 -3.45864117e-01 5.79357326e-01 4.62369949e-01
-4.91509527e-01 -4.51818764e-01 -9.17261124e-01 -6.02621913e-01
6.10027499e-02 -6.13582790e-01 2.95747310e-01 5.70627809e-01
9.37733129e-02 7.01576412e-01 -7.50537932e-01 -8.53911787e-02
-8.07841718e-02 -9.44517627e-02 5.72681129e-01 -7.92642713e-01
-3.87369752e-01 -5.66221893e-01 -1.20153129e-01 -5.91761649e-01
5.47151148e-01 -9.75589454e-01 3.70120794e-01 -1.36996567e+00
-4.64320928e-01 -4.73985784e-02 1.55434296e-01 4.55091208e-01
1.75908402e-01 1.61243603e-01 3.00236017e-01 -4.93148088e-01
1.94851562e-01 9.96293247e-01 1.28007460e+00 9.70635042e-02
-4.43672776e-01 1.78376123e-01 -3.31685811e-01 4.59929138e-01
9.89010215e-01 -4.54212636e-01 -4.70487088e-01 2.70325422e-01
1.53241665e-04 5.69148719e-01 -1.37874916e-01 -1.06886446e+00
-1.45478442e-01 7.67290667e-02 1.73495978e-01 -4.28594738e-01
8.33333254e-01 -6.27720892e-01 5.00371814e-01 2.50710934e-01
-6.35224164e-01 -2.56114125e-01 4.16327894e-01 -3.49043868e-03
-5.31373799e-01 -4.69841599e-01 9.61594164e-01 1.04054250e-01
-1.80722281e-01 -1.75843284e-01 -8.86727571e-01 -1.55387819e-01
8.10789943e-01 -7.70454034e-02 1.93335757e-01 -9.06583548e-01
-1.13727307e+00 -2.46973321e-01 4.05424327e-01 5.53292751e-01
3.96926403e-01 -1.27324176e+00 -7.99643099e-01 3.40735376e-01
-1.84497103e-01 -3.12456846e-01 2.48038515e-01 5.00396907e-01
-2.69943625e-01 5.00067592e-01 -1.57863140e-01 -3.95484596e-01
-1.51963615e+00 3.42600048e-01 5.35919487e-01 -1.79748088e-02
-1.50694221e-01 7.18832731e-01 -2.19077781e-01 -6.45529628e-01
1.52608275e-01 -3.51015925e-01 -1.69734627e-01 1.76748499e-01
7.20608532e-02 1.27491191e-01 2.70336401e-02 -8.05728376e-01
-9.77633893e-02 2.13815302e-01 5.45881629e-01 -9.69097435e-01
1.18610024e+00 -6.54887035e-02 3.38657834e-02 6.11500382e-01
1.16627598e+00 6.44873917e-01 -1.05358016e+00 5.23999691e-01
-1.14303410e-01 -1.16802357e-01 1.37421653e-01 -1.15703249e+00
-8.89925003e-01 8.00050199e-01 6.30017936e-01 2.22967476e-01
1.19478476e+00 -3.36563177e-02 8.16398203e-01 2.48053074e-02
-5.75102344e-02 -1.15070379e+00 -4.37658764e-02 5.20769954e-01
1.29690313e+00 -8.23981047e-01 -6.59976840e-01 -2.52784878e-01
-1.01015592e+00 1.03684568e+00 3.18823248e-01 1.84733704e-01
3.32345515e-01 8.29744220e-01 5.15064538e-01 2.74054527e-01
-1.01743746e+00 -1.39542043e-01 3.00318480e-01 6.72868729e-01
8.77084017e-01 2.28139341e-01 -3.98257315e-01 5.13775289e-01
-8.73516500e-01 -1.42950311e-01 4.62836057e-01 4.78534907e-01
-2.96277344e-01 -1.53080606e+00 -2.95444220e-01 2.66686846e-02
-6.99345231e-01 -1.01349175e-01 -5.82481444e-01 6.38486445e-01
6.09022528e-02 1.25078058e+00 1.06281541e-01 -3.35447848e-01
3.14089864e-01 5.69390357e-01 3.46496969e-01 -8.13242197e-01
-8.68209839e-01 5.78840613e-01 6.96855068e-01 -1.42641976e-01
-2.91296273e-01 -6.69352472e-01 -1.49665940e+00 1.01285353e-01
-2.72287250e-01 4.43111032e-01 1.05744088e+00 7.77294815e-01
5.31370528e-02 1.06333315e+00 7.03858733e-01 -1.02339792e+00
-2.39228666e-01 -1.21152866e+00 -6.02012336e-01 1.09509386e-01
2.55574465e-01 1.39636239e-02 -7.42723227e-01 1.43147603e-01] | [14.991663932800293, 6.506266117095947] |
72672991-2bc8-48c0-a444-50cac090dd3f | design-and-implementation-of-real-time-1 | 2112.04839 | null | https://arxiv.org/abs/2112.04839v1 | https://arxiv.org/pdf/2112.04839v1.pdf | Design and Implementation of Real-Time Localization System (RTLS) based on UWB and TDoA Algorithm | Nowadays, accurate localization plays an essential role in many fields, like target tracking and path planning. The challenges of indoor localization include inadequate localization accuracy, unreasonable anchor deployment in complex scenarios, lack of stability, and high cost. So the universal positioning technologies cannot meet the real application requirements scarcely. To overcome these shortcomings, a comprehensive Ultra Wide-Band (UWB) based RTLS is presented in this paper. We first introduce the architecture of the real-time localization system, then propose a new wireless clock synchronization (WCS) scheme, finally discuss the time difference of arrival (TDoA) algorithm. We define the time-base selection strategy for the TDoA algorithm, and analyze the relationship between anchor deployment and positioning accuracy. The Extended Kalman Filter (EKF) method is presented for non-linear dynamic localization estimation, and it performs well in terms of stability and accuracy in moving targets. | ['Hao Li', 'Shuang-Hua Yang', 'Yulong Ding', 'Yuhuan Liu', 'Li Yang', 'Fengyun Zhang'] | 2021-12-09 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [-1.87413380e-01 -5.57274044e-01 -2.27397159e-01 -1.44841865e-01
-7.08238900e-01 -6.03839815e-01 1.63962469e-01 1.57993451e-01
-4.49900120e-01 1.14826608e+00 -3.53197128e-01 -5.58623910e-01
-5.74747622e-01 -7.65864253e-01 -1.09279357e-01 -1.10169971e+00
-5.21518469e-01 -2.21976787e-01 4.04456854e-01 -1.13064587e-01
8.10912102e-02 6.15896344e-01 -7.48827577e-01 -9.41305459e-01
1.13799393e+00 1.38717866e+00 1.78385004e-01 5.23651481e-01
-8.01995490e-03 1.04897581e-01 -1.24716604e+00 2.28144810e-01
-2.03095883e-01 -1.29486889e-01 1.11878134e-01 -6.83945060e-01
-2.74860710e-01 -5.05872704e-02 -2.44931668e-01 6.26897931e-01
1.10441804e+00 1.10149041e-01 2.31561005e-01 -1.44848764e+00
-1.49760127e-01 5.49227595e-01 -4.51413631e-01 3.30535859e-01
6.49047971e-01 -7.37887681e-01 5.49757928e-02 -4.43671793e-01
2.82143503e-02 7.55246580e-01 1.42591906e+00 3.31105590e-02
-7.61696637e-01 -8.80974889e-01 -2.73414887e-03 4.69843261e-02
-1.80462813e+00 -6.07202888e-01 4.15987819e-01 1.55189127e-01
1.21988147e-01 3.38568985e-01 6.18151486e-01 1.04245567e+00
8.71442080e-01 -4.08927351e-02 7.56151974e-01 -5.14849246e-01
4.99665678e-01 -2.17092559e-01 -1.10075712e-01 1.77541584e-01
9.49277461e-01 -6.36528879e-02 -3.75813067e-01 -3.65966290e-01
7.65380383e-01 -1.47349074e-01 -4.46099013e-01 -2.00696498e-01
-1.40433013e+00 1.36862829e-01 5.65496385e-01 9.20446634e-01
-2.89140761e-01 6.85575664e-01 -1.61475435e-01 1.69833302e-01
2.56468713e-01 2.43873239e-01 -4.74383891e-01 -2.35300258e-01
-9.32568550e-01 -1.22787587e-01 7.18443632e-01 1.29059935e+00
3.10847431e-01 3.18341464e-01 -1.25660775e-02 3.27081412e-01
6.21420145e-01 1.30651402e+00 3.21706802e-01 -8.60928714e-01
8.09528306e-02 -4.01686400e-01 7.61235416e-01 -1.32339299e+00
-7.74193704e-01 -1.21324265e+00 -8.78702164e-01 -5.73065341e-01
3.90505821e-01 -6.62707090e-01 -4.19103354e-01 1.56821883e+00
4.74893808e-01 4.71963942e-01 1.46976575e-01 2.08506674e-01
4.08084154e-01 7.08619535e-01 6.01845533e-02 -8.04593921e-01
1.11572850e+00 -2.77039468e-01 -1.40622270e+00 -2.65471041e-01
4.64797288e-01 -1.06698775e+00 6.65953532e-02 1.79033935e-01
-6.19550824e-01 -6.25950873e-01 -1.36710334e+00 7.04403639e-01
-3.08304965e-01 2.24127755e-01 6.18800759e-01 1.29439616e+00
-9.70279396e-01 -8.25398788e-02 -8.67061734e-01 -5.42377949e-01
-3.22901011e-01 5.71961224e-01 6.65119365e-02 -2.62835085e-01
-1.41741133e+00 8.83021235e-01 -6.29813597e-02 5.88036060e-01
1.28002360e-01 -4.06359404e-01 -7.70579219e-01 -2.49363780e-01
2.59149214e-03 -4.31034058e-01 1.21699440e+00 -2.68928915e-01
-1.37381494e+00 -2.09137112e-01 -6.17714763e-01 -5.07867575e-01
1.98655248e-01 3.68059650e-02 -1.18899095e+00 -2.50448465e-01
5.68903506e-01 -2.81852007e-01 3.83698106e-01 -1.19982278e+00
-8.04066420e-01 -1.27757445e-01 -3.25832963e-01 -2.90394574e-01
3.12376209e-02 -4.85111922e-01 -2.23921742e-02 -5.18017769e-01
1.08798611e+00 -7.37440228e-01 -6.68590963e-01 -2.96087861e-01
-2.86414772e-02 1.88628942e-01 7.00684786e-01 -1.96507677e-01
1.59921110e+00 -2.15061235e+00 -7.83340275e-01 5.12855947e-01
-2.15772897e-01 -1.81975752e-01 3.17719221e-01 5.73929250e-01
4.62529808e-01 -2.52584338e-01 3.20578933e-01 -1.22019246e-01
-1.51651250e-02 1.83764443e-01 -3.05110186e-01 9.90221977e-01
-7.85996735e-01 2.89670169e-01 -1.19221365e+00 -3.71659577e-01
2.11757362e-01 4.66689885e-01 1.47957370e-01 -1.85481608e-01
5.95329046e-01 8.66167486e-01 -9.06846762e-01 9.72374022e-01
1.24743032e+00 1.65032327e-01 6.03116900e-02 -4.24214959e-01
-1.02866948e+00 -5.41654415e-02 -1.46477640e+00 1.64809513e+00
-6.85480475e-01 4.53615963e-01 2.31384963e-01 -5.09815872e-01
1.06226122e+00 4.31011885e-01 7.47301221e-01 -9.96067703e-01
2.55646020e-01 9.16656137e-01 -1.30605400e-01 -3.21687698e-01
4.42024559e-01 3.07420135e-01 -4.24431175e-01 2.31293738e-01
-2.59084344e-01 2.40600884e-01 -1.78591445e-01 -6.64820299e-02
1.21959651e+00 -2.51081511e-02 6.25145972e-01 -3.71714234e-01
7.05241621e-01 -1.68158546e-01 8.81970227e-01 9.18722212e-01
-3.47657144e-01 -1.76829785e-01 -2.97244549e-01 -2.01343089e-01
-2.59227127e-01 -9.15733278e-01 -2.39923373e-01 5.94622314e-01
9.53139842e-01 -3.30984712e-01 -1.07449479e-01 8.09809789e-02
1.60253078e-01 4.60420996e-01 1.17757556e-03 -1.21637985e-01
-5.92848480e-01 -6.08214438e-01 8.91682327e-01 5.03110401e-02
8.86206090e-01 1.43231034e-01 -5.49321771e-01 7.23184586e-01
-3.96729827e-01 -1.28714931e+00 -8.28629509e-02 1.18845090e-01
-4.33042347e-01 -6.14208996e-01 -6.38285279e-01 -5.92373073e-01
6.48218513e-01 1.04143798e+00 3.94573659e-01 5.61601147e-02
6.96434220e-03 6.53334796e-01 -3.93653959e-01 -5.41939676e-01
3.96966070e-01 1.93568408e-01 7.15655327e-01 -7.89068341e-02
-3.44624817e-02 -6.98169947e-01 -8.97702038e-01 7.92307675e-01
-6.06952645e-02 -7.24046469e-01 6.31044507e-01 4.43033963e-01
3.75378489e-01 4.34115916e-01 9.16674435e-01 -8.62689875e-03
3.42828155e-01 -5.89228511e-01 -9.47754264e-01 2.90101022e-01
-5.76077938e-01 -4.79688197e-01 2.37252250e-01 -2.17464715e-01
-9.80615616e-01 1.33169218e-04 -3.18430334e-01 5.79611123e-01
2.29000762e-01 3.83517325e-01 -2.71436095e-01 -1.05209720e+00
6.02866411e-01 2.87321806e-01 -5.01183212e-01 -2.02876493e-01
3.67657579e-02 7.87885368e-01 5.01501501e-01 -5.43362856e-01
1.17037416e+00 6.37112200e-01 3.51631165e-01 -9.36572373e-01
-6.55446470e-01 -5.51512897e-01 -3.69021982e-01 -3.17894012e-01
1.07361853e-01 -1.02466583e+00 -9.25877810e-01 1.99184537e-01
-1.29024017e+00 2.92167980e-02 3.01881969e-01 9.08498764e-01
-1.77903417e-02 4.17938530e-01 4.02363427e-02 -1.26088083e+00
-5.70317470e-02 -8.74057829e-01 5.26518106e-01 6.08020425e-01
-2.83554234e-02 -1.10072851e+00 9.41138417e-02 -3.44623655e-01
1.07330048e+00 6.72306836e-01 -1.56883933e-02 -1.59387812e-01
-7.73244977e-01 -6.83073163e-01 -7.20756426e-02 -4.88939941e-01
2.50744402e-01 -2.85029918e-01 -5.86557209e-01 -4.36791569e-01
7.85176754e-02 5.49799860e-01 3.31061482e-02 9.51115370e-01
6.48387015e-01 -2.10720440e-03 -1.07251251e+00 8.12564313e-01
1.47584713e+00 6.39071226e-01 6.55356407e-01 6.60962880e-01
7.45839477e-02 -1.30740777e-01 1.32930505e+00 4.88732964e-01
4.94612068e-01 6.02607250e-01 4.28925812e-01 4.79902700e-02
3.81347686e-01 7.71784112e-02 1.67756811e-01 5.40575624e-01
2.72008870e-02 -6.81756794e-01 -7.45125055e-01 1.99136212e-01
-1.82316673e+00 -7.42953122e-01 -6.05239809e-01 2.32047391e+00
2.40844056e-01 1.37129501e-01 -4.66285914e-01 1.02647968e-01
8.60987246e-01 1.99867666e-01 -6.20880760e-02 1.56566560e-01
5.45711517e-02 -8.16978700e-03 1.43997514e+00 6.45191729e-01
-8.99546683e-01 5.01286626e-01 6.48089123e+00 8.86521578e-01
-1.13992667e+00 4.43652391e-01 -2.04077750e-01 7.76791930e-01
-1.15317754e-01 5.04702851e-02 -8.46701562e-01 7.61418462e-01
9.61241245e-01 -1.80964634e-01 -2.58480370e-01 7.44379401e-01
3.62973899e-01 -6.95725858e-01 -4.02053833e-01 1.15449286e+00
-1.78575575e-01 -1.13899279e+00 -8.88143122e-01 1.26118317e-01
4.56995040e-01 -2.69818306e-01 -4.55632769e-02 6.31007329e-02
-2.68587679e-01 -4.21882302e-01 6.02798045e-01 4.70144928e-01
7.32202947e-01 -7.09890783e-01 9.98644292e-01 2.33559161e-01
-1.77239013e+00 -1.77489311e-01 -5.04791081e-01 -1.74756199e-01
5.83973706e-01 1.08642912e+00 -4.60386246e-01 1.10119212e+00
4.57041562e-01 2.35796705e-01 -1.68383405e-01 1.84445894e+00
-2.14621618e-01 2.95233309e-01 -7.84111083e-01 -2.33828455e-01
2.15106711e-01 2.32512355e-02 5.27380705e-01 9.41526115e-01
1.25687397e+00 2.41417870e-01 1.17760837e-01 9.48856957e-03
4.61154252e-01 -2.47571424e-01 -4.35159028e-01 6.91097200e-01
1.52749872e+00 1.16563344e+00 -7.96661735e-01 1.43052563e-01
-6.67140111e-02 3.72283608e-01 -8.88775349e-01 5.88519573e-01
-1.00128710e+00 -9.21509027e-01 2.26029366e-01 -4.51775119e-02
-1.20193966e-01 -1.06645155e+00 -1.82503670e-01 -7.23626554e-01
-3.04800183e-01 -3.11684981e-02 -3.65829207e-02 -4.34594214e-01
-5.66753030e-01 3.41182441e-01 -2.24809542e-01 -1.72416031e+00
2.00897634e-01 -3.94491144e-02 -6.23954356e-01 6.90962374e-01
-1.62668371e+00 -9.80247915e-01 -7.81805694e-01 4.36814994e-01
1.54093355e-01 8.80449191e-02 7.93232620e-01 8.61291230e-01
-3.59105587e-01 5.65759301e-01 7.29673207e-01 -3.60903442e-01
1.00139928e+00 -6.96879148e-01 1.50371835e-01 1.05870295e+00
-5.86187661e-01 9.32941377e-01 1.09292352e+00 -6.40329301e-01
-1.60176826e+00 -7.37651527e-01 1.03472328e+00 -5.28357327e-02
6.19063735e-01 -2.11857393e-01 -3.82275544e-02 4.35586780e-01
-4.74561006e-02 1.98174089e-01 8.21259379e-01 -3.32802385e-02
3.29105675e-01 -8.83590281e-01 -1.19084001e+00 4.01944011e-01
7.36287653e-01 2.53305674e-01 -1.04936004e-01 2.93008000e-01
6.45349920e-01 -5.84940135e-01 -6.66635215e-01 4.56909090e-01
1.10634458e+00 -6.43640637e-01 1.25020790e+00 7.31927633e-01
-1.15183651e+00 -8.68599832e-01 -4.45554584e-01 -1.05227053e+00
-4.18483704e-01 -9.03640687e-01 -6.73524756e-03 1.42911100e+00
1.30706981e-01 -1.23065913e+00 2.86486119e-01 2.53673811e-02
-1.10274695e-01 -2.59701937e-01 -1.57669139e+00 -1.06082535e+00
-1.03683317e+00 -5.13129592e-01 8.24474812e-01 8.31464767e-01
-1.96150497e-01 5.30200675e-02 -5.54987073e-01 1.04958892e+00
1.03388071e+00 -3.06743920e-01 8.53009820e-01 -1.35181439e+00
1.89890876e-01 3.41927260e-01 -4.08330172e-01 -1.39977205e+00
-3.96423548e-01 1.40004665e-01 2.38687411e-01 -1.82270861e+00
-1.02546608e+00 -1.30381787e+00 -3.66626024e-01 -1.29632995e-01
4.65262234e-01 1.56685010e-01 -5.09803593e-01 1.73785478e-01
-9.09281492e-01 3.23460251e-01 6.34677708e-01 1.47833422e-01
-1.84096873e-01 8.94523561e-01 -2.65486956e-01 5.37566662e-01
7.84198463e-01 -5.41328788e-01 -4.76622283e-01 -6.20538473e-01
3.41892570e-01 2.89754450e-01 9.62571353e-02 -1.68583953e+00
9.03648853e-01 -1.13610223e-01 7.16448963e-01 -7.89228678e-01
3.32628816e-01 -1.43279421e+00 4.64317828e-01 1.13172448e+00
4.98088628e-01 6.55464306e-02 -5.91446608e-02 9.04451072e-01
1.18327998e-01 7.18420278e-03 4.34673488e-01 3.26870114e-01
-6.26796305e-01 3.05751354e-01 -6.29577816e-01 -7.72935390e-01
1.19619000e+00 -5.36911666e-01 -5.89326620e-01 -5.98846436e-01
-3.86657298e-01 4.38143522e-01 1.19053915e-01 -1.33142127e-02
3.28638881e-01 -1.66399229e+00 2.38627680e-02 -5.62855527e-02
-3.55023965e-02 -2.63807833e-01 4.49022114e-01 1.27506554e+00
-7.47922957e-01 7.51296997e-01 8.73771124e-03 -6.62149429e-01
-8.67935836e-01 -1.27658486e-01 4.15073097e-01 1.39213145e-01
1.27967820e-01 7.72172213e-01 -5.72020054e-01 -3.70108671e-02
5.04749060e-01 -2.39259511e-01 -6.73808753e-02 -7.76114389e-02
7.03671038e-01 5.42974293e-01 -3.33877429e-02 -5.31960547e-01
-9.04161751e-01 1.08411849e+00 6.69895947e-01 -1.95329398e-01
7.41083443e-01 -9.65347886e-01 -3.56985271e-01 3.58048916e-01
5.76759696e-01 7.04879701e-01 -6.30812347e-01 -1.15088895e-01
2.58406132e-01 -6.91325784e-01 -1.48919746e-01 -4.55275595e-01
-2.35918149e-01 2.00547278e-01 8.92210484e-01 5.41841030e-01
8.04167032e-01 -5.05245864e-01 8.85202110e-01 6.35976493e-01
1.35089886e+00 -8.85661423e-01 -2.61204004e-01 4.43064094e-01
1.64836273e-01 -7.82466650e-01 2.41896570e-01 -4.84976500e-01
3.62212807e-01 1.26763022e+00 3.94261092e-01 1.26696704e-02
7.12013006e-01 3.93349707e-01 1.76711202e-01 2.68482447e-01
5.66971600e-02 -1.53181463e-01 -3.58658999e-01 1.01782501e+00
3.86383653e-01 -1.44824550e-01 -6.73500597e-01 6.74107492e-01
-1.18747793e-01 -3.06266874e-01 2.70455778e-01 1.30306995e+00
-1.00988019e+00 -1.27021253e+00 -1.01761639e+00 -7.83006325e-02
-3.89903635e-01 4.24103945e-01 5.16443074e-01 8.16455543e-01
4.84628290e-01 1.31341970e+00 -2.54738092e-01 -3.35530639e-01
3.79726708e-01 -5.75987101e-01 3.36112261e-01 -1.50292888e-02
2.50599682e-01 3.01948637e-01 1.13452852e-01 -5.24429679e-01
-6.33503675e-01 -2.86052346e-01 -1.28545225e+00 -4.16300178e-01
-6.12154424e-01 9.29634869e-01 1.17776501e+00 9.09692109e-01
2.76144892e-01 7.49981701e-01 9.28980231e-01 -7.99195409e-01
7.97890872e-03 -4.72012311e-01 -5.91869891e-01 -1.03271699e+00
6.96654916e-01 -9.14767742e-01 -3.76906931e-01 -7.17029810e-01] | [6.31514835357666, 1.0585548877716064] |
1e9c30fe-bede-4aad-9883-b3a84a907884 | faster-r-cnn-features-for-instance-search | 1604.08893 | null | http://arxiv.org/abs/1604.08893v1 | http://arxiv.org/pdf/1604.08893v1.pdf | Faster R-CNN Features for Instance Search | Image representations derived from pre-trained Convolutional Neural Networks
(CNNs) have become the new state of the art in computer vision tasks such as
instance retrieval. This work explores the suitability for instance retrieval
of image- and region-wise representations pooled from an object detection CNN
such as Faster R-CNN. We take advantage of the object proposals learned by a
Region Proposal Network (RPN) and their associated CNN features to build an
instance search pipeline composed of a first filtering stage followed by a
spatial reranking. We further investigate the suitability of Faster R-CNN
features when the network is fine-tuned for the same objects one wants to
retrieve. We assess the performance of our proposed system with the Oxford
Buildings 5k, Paris Buildings 6k and a subset of TRECVid Instance Search 2013,
achieving competitive results. | ["Shin'ichi Satoh", 'Xavier Giro-i-Nieto', 'Ferran Marques', 'Amaia Salvador'] | 2016-04-29 | null | null | null | null | ['instance-search'] | ['computer-vision'] | [-3.21460254e-02 5.42515032e-02 -1.62414983e-01 -5.46796322e-01
-1.09601569e+00 -5.38165748e-01 1.23929310e+00 5.28414667e-01
-9.22172785e-01 4.90054458e-01 2.64112711e-01 7.51421824e-02
-6.39615953e-01 -8.51597548e-01 -8.60199630e-01 -3.14115316e-01
-2.61652261e-01 7.43041217e-01 5.58504105e-01 -2.83820003e-01
5.83221912e-01 1.01079309e+00 -2.02132940e+00 7.64640450e-01
2.41105124e-01 1.57812190e+00 2.89355367e-01 6.67718232e-01
-8.30468349e-03 4.64140177e-01 -7.19558358e-01 1.26313180e-01
4.03462827e-01 4.55792516e-01 -1.12936628e+00 -3.84548634e-01
9.49171960e-01 -3.48513246e-01 -4.88780826e-01 5.56691408e-01
5.18337905e-01 4.43731785e-01 9.09676492e-01 -8.24349582e-01
-8.04892600e-01 4.94136661e-01 -2.38365427e-01 6.35767877e-01
1.66932255e-01 6.42983317e-02 1.18467903e+00 -1.08140481e+00
8.73846054e-01 1.27014089e+00 4.30510491e-01 -4.33350494e-03
-8.90885413e-01 -5.04272819e-01 -1.48135442e-02 4.19537634e-01
-1.87850916e+00 -2.76001632e-01 3.83213311e-01 -2.43373021e-01
1.54978085e+00 2.06396252e-01 5.43732285e-01 7.83566535e-01
-9.36987624e-02 1.05692863e+00 6.86678231e-01 -2.77549922e-01
8.00368786e-02 2.52408713e-01 1.85294926e-01 5.61616123e-01
1.08597472e-01 1.73738360e-01 -2.57330567e-01 -2.87611097e-01
7.36850500e-01 1.26744241e-01 1.31080905e-02 -4.13500369e-01
-1.32959461e+00 8.20818007e-01 1.54695821e+00 6.12288237e-01
-6.42389596e-01 5.59539557e-01 6.10226631e-01 1.77858844e-01
5.34084141e-01 9.33568239e-01 -6.38485312e-01 4.71950233e-01
-1.43541777e+00 5.55225372e-01 3.46608132e-01 1.03253376e+00
6.89045131e-01 -5.22396088e-01 -8.68389070e-01 9.59819257e-01
2.97042251e-01 1.49992272e-01 5.21090031e-01 -5.16045094e-01
4.06613797e-01 7.22060442e-01 1.36474311e-01 -8.25057149e-01
-4.83624548e-01 -8.16244245e-01 -5.74059248e-01 -2.08793133e-01
2.27027386e-01 7.52997279e-01 -1.40066969e+00 1.18806505e+00
1.39647007e-01 -6.50751367e-02 -4.38336357e-02 9.74300861e-01
1.22664034e+00 7.12282419e-01 1.99127018e-01 5.72410047e-01
1.38270259e+00 -1.16956496e+00 8.67389515e-02 1.72168270e-01
6.04366541e-01 -6.74949229e-01 6.35071158e-01 1.49330005e-01
-9.95007038e-01 -9.81456041e-01 -9.21790481e-01 -3.65943342e-01
-1.03535044e+00 5.02104044e-01 3.59947175e-01 1.43253356e-01
-1.32386541e+00 7.11101830e-01 -1.59198344e-01 -7.25004613e-01
6.42739475e-01 6.01821661e-01 -4.95906472e-01 -3.16313773e-01
-1.12579036e+00 1.00680244e+00 8.70557964e-01 2.82689601e-01
-1.29882634e+00 -6.12980962e-01 -6.00130200e-01 3.41350257e-01
7.41604641e-02 -4.36247766e-01 9.31928039e-01 -6.94475591e-01
-7.90907323e-01 1.14510334e+00 3.98990005e-01 -6.55882239e-01
4.05045062e-01 -4.14404005e-01 -2.12306052e-01 2.94700712e-01
1.30234152e-01 1.15090215e+00 8.16191792e-01 -1.02399552e+00
-8.40669274e-01 -3.40523511e-01 2.42831141e-01 3.68105210e-02
-1.56346887e-01 9.73933563e-02 -6.82752371e-01 -4.07460749e-01
-3.62092629e-02 -8.04542243e-01 -4.03345823e-01 -3.03353393e-03
-4.02427107e-01 -8.46385717e-01 5.96490920e-01 -3.81656796e-01
9.38495159e-01 -1.81553268e+00 6.73559383e-02 6.57224715e-01
-1.01891288e-03 5.11809289e-01 -6.12928689e-01 3.05406928e-01
-2.37835363e-01 1.31132975e-01 3.99638504e-01 -7.45311826e-02
-6.15353025e-02 -1.13355339e-01 -2.27472946e-01 4.87963647e-01
5.43812871e-01 1.30523086e+00 -8.84029567e-01 -5.35879910e-01
2.22834125e-01 6.17412388e-01 -2.93518722e-01 -1.22713204e-02
-3.61842096e-01 7.26344138e-02 -8.09163034e-01 7.83125103e-01
4.43988025e-01 -4.77712274e-01 -2.67148376e-01 -4.47577089e-01
-2.39359632e-01 2.15394169e-01 -8.51492226e-01 1.81468999e+00
-3.20452422e-01 5.44292748e-01 -5.24038374e-01 -9.48044658e-01
1.02452326e+00 1.64498016e-01 1.37099296e-01 -1.19451714e+00
-4.05390300e-02 4.56678599e-01 -4.30708736e-01 -1.91927269e-01
1.06406772e+00 6.04347765e-01 -8.45635608e-02 1.70093969e-01
4.47735310e-01 5.15015796e-02 1.85419887e-01 2.88598120e-01
1.23933566e+00 1.73910737e-01 3.04134756e-01 -5.60471475e-01
6.12446249e-01 2.18664154e-01 -2.80838162e-01 1.29099178e+00
1.86120830e-02 1.13345563e+00 -7.80997053e-02 -1.06078875e+00
-1.39356065e+00 -8.22712183e-01 -3.99964124e-01 1.35954547e+00
-1.33878008e-01 -2.71634638e-01 -1.60064399e-01 -8.80496323e-01
3.12133223e-01 1.77203521e-01 -9.79938328e-01 -1.26066953e-01
-6.45759881e-01 -3.42501312e-01 6.84835076e-01 6.35310590e-01
4.43636149e-01 -1.49463737e+00 -6.29287004e-01 1.38949841e-01
2.31407255e-01 -7.58984506e-01 -9.94629636e-02 4.62213039e-01
-7.43801057e-01 -8.62081885e-01 -1.27698183e+00 -6.67394936e-01
6.04629397e-01 1.79136932e-01 1.45925367e+00 4.17502850e-01
-6.80279911e-01 5.84312797e-01 -4.40848321e-01 -2.89059579e-01
2.47945890e-01 8.78293335e-01 -2.99305528e-01 -2.58454621e-01
1.71108395e-01 6.67878613e-02 -1.03902531e+00 2.37808734e-01
-1.05865204e+00 -5.69486797e-01 7.79624045e-01 6.06904268e-01
7.44999111e-01 -1.29452005e-01 1.78286299e-01 -5.72232127e-01
5.64101517e-01 -5.28395593e-01 -6.98689342e-01 7.08362401e-01
-4.44977731e-01 3.80867571e-01 1.96316645e-01 -3.33919883e-01
-6.49208426e-01 2.28115410e-01 1.76004469e-01 -6.32189870e-01
-3.28037351e-01 4.86146867e-01 5.05559027e-01 -1.82814866e-01
9.91671264e-01 2.58805335e-01 -6.20365500e-01 -6.01280332e-01
6.92241609e-01 5.53110719e-01 3.09685618e-01 -6.37012005e-01
7.53853917e-01 2.86935389e-01 4.96606668e-03 -6.32178962e-01
-8.63670886e-01 -1.19370127e+00 -8.23491156e-01 -2.03481942e-01
9.94944513e-01 -1.06596768e+00 -5.46841264e-01 -1.58451661e-01
-1.26722264e+00 8.17313790e-02 -4.21129405e-01 3.13076377e-01
-3.51459026e-01 -3.14506888e-01 -3.93973738e-01 -5.70523560e-01
-5.50539792e-01 -1.01928830e+00 1.73276126e+00 3.99263918e-01
1.38775870e-01 -3.88421625e-01 3.06555420e-01 2.13238358e-01
8.59140456e-01 -1.86047599e-01 7.81112790e-01 -1.25145662e+00
-1.08529532e+00 -6.22883916e-01 -8.96624088e-01 -6.78879842e-02
-4.40060437e-01 -1.30600482e-01 -1.22836995e+00 -3.97079319e-01
-9.60666716e-01 -4.73017693e-01 1.57708406e+00 3.58459562e-01
1.29594648e+00 -1.61893815e-01 -5.89264750e-01 3.65262896e-01
1.68903756e+00 -2.93342769e-01 7.51127303e-01 7.40297019e-01
4.46453810e-01 5.68660736e-01 6.40150666e-01 3.10030520e-01
6.38653561e-02 8.23391140e-01 7.26179004e-01 -5.83796874e-02
-1.47649616e-01 -8.35395604e-03 -3.86386007e-01 1.80375099e-01
-3.93844277e-01 -2.12786749e-01 -1.00172579e+00 1.11521661e+00
-1.90669930e+00 -7.57007122e-01 2.78757304e-01 2.07638741e+00
3.93128932e-01 4.82154340e-02 2.12915093e-02 -2.63266474e-01
3.95140439e-01 3.08633298e-01 -1.35259032e-01 -2.78150082e-01
1.44539569e-02 6.84000134e-01 7.83437073e-01 -1.14461273e-01
-1.37461305e+00 1.02576244e+00 5.96144819e+00 1.10257423e+00
-1.04037440e+00 -5.25959134e-02 6.68257952e-01 -1.51599750e-01
2.57758163e-02 -2.47193858e-01 -1.01351655e+00 -9.84557867e-02
8.34461749e-01 4.17899817e-01 -7.24270642e-02 1.22975314e+00
-3.30431432e-01 3.20033059e-02 -1.22314429e+00 7.47662425e-01
1.18429661e-01 -1.67978299e+00 3.44787419e-01 1.13184609e-01
6.29707396e-01 6.67859495e-01 3.14519614e-01 6.55181348e-01
1.24456115e-01 -1.36111677e+00 6.44058108e-01 8.61567438e-01
4.32625502e-01 -6.10941947e-01 9.19915318e-01 1.00232195e-02
-1.46210456e+00 -2.97391325e-01 -7.70833075e-01 5.43240190e-01
-2.70150870e-01 1.89214110e-01 -1.24280119e+00 6.10958755e-01
1.27148855e+00 5.85825145e-01 -9.91401613e-01 1.66218376e+00
7.47488737e-02 1.65016547e-01 -5.55423737e-01 -1.43953562e-01
8.09924603e-01 4.01047438e-01 2.81093121e-01 1.34069729e+00
2.65289783e-01 -1.88025057e-01 2.18242347e-01 8.96932244e-01
-2.09631026e-01 3.59755814e-01 -6.07088745e-01 1.15781955e-01
1.65292606e-01 1.59838533e+00 -1.01106894e+00 -5.04884660e-01
-5.35935722e-02 7.17753172e-01 7.10342586e-01 2.70528048e-01
-4.92080510e-01 -2.91404873e-01 3.90505612e-01 1.95312575e-01
8.51676047e-01 1.05921336e-01 4.29534167e-01 -6.68307006e-01
-1.49752274e-01 -4.70135808e-01 5.22365093e-01 -1.11433470e+00
-1.23944640e+00 9.20026362e-01 2.33017266e-01 -1.00669432e+00
-1.25894025e-01 -6.69171333e-01 -3.21947157e-01 8.32146943e-01
-1.90217471e+00 -1.51354218e+00 -2.87913203e-01 4.09993768e-01
6.55833304e-01 -3.07601839e-01 7.46322095e-01 3.25654596e-01
-1.69422358e-01 4.23014313e-01 5.46542555e-02 1.96023837e-01
5.40240645e-01 -1.08741081e+00 4.24143016e-01 1.30860820e-01
4.80460048e-01 8.00954282e-01 2.11761519e-01 -3.07060599e-01
-8.45235348e-01 -1.38041949e+00 9.24583495e-01 -5.62358141e-01
2.99443364e-01 -1.53907761e-01 -6.81229770e-01 4.25429314e-01
1.03826143e-01 5.45466363e-01 6.00829311e-02 1.92147687e-01
-7.13010550e-01 -1.59483224e-01 -1.16555440e+00 1.29660338e-01
7.97370315e-01 -6.77613854e-01 -5.32238424e-01 5.40960073e-01
6.56825662e-01 -1.53662175e-01 -9.16121483e-01 6.32058322e-01
7.64325857e-01 -5.51707804e-01 1.63138759e+00 -9.48794663e-01
3.36582482e-01 -2.29639277e-01 -4.28915769e-01 -8.03205848e-01
-5.95268369e-01 1.52687564e-01 2.45472252e-01 8.02769899e-01
6.61703646e-01 -9.02451277e-02 8.24390769e-01 3.31364453e-01
4.92427982e-02 -6.04538858e-01 -8.97235692e-01 -7.28284180e-01
-1.13180451e-01 -2.29382813e-01 6.50531709e-01 6.38464332e-01
-6.33166254e-01 7.92158619e-02 -3.34086344e-02 2.27361605e-01
3.68993849e-01 9.11003500e-02 6.14791393e-01 -1.58421671e+00
5.04910611e-02 -4.78584528e-01 -6.81819320e-01 -9.27775443e-01
-1.39459789e-01 -1.12050748e+00 2.14388385e-01 -1.55120420e+00
4.02722716e-01 -7.63842583e-01 -1.12270141e+00 5.35233676e-01
2.06093073e-01 6.55849278e-01 3.46207410e-01 4.64811534e-01
-1.41108191e+00 3.62658650e-01 8.66346776e-01 -4.40264046e-01
-8.43339711e-02 -4.22298387e-02 -1.19391426e-01 2.12546036e-01
4.17744964e-01 -5.86352408e-01 -1.01489000e-01 -3.96259755e-01
5.36274970e-01 -2.48500705e-01 5.77297628e-01 -1.12428176e+00
3.53081942e-01 4.25675005e-01 8.98211837e-01 -1.06131232e+00
4.23445165e-01 -9.23154354e-01 -8.36200044e-02 3.13159555e-01
-8.11137974e-01 1.93665043e-01 2.71312565e-01 5.96162736e-01
-3.76385957e-01 -2.81480551e-01 4.12209064e-01 -5.71046233e-01
-9.70324218e-01 4.67210352e-01 3.55134979e-02 -3.18154871e-01
6.90473497e-01 -1.91899523e-01 -3.31800699e-01 -1.80514790e-02
-7.36713648e-01 7.85464868e-02 5.98010421e-02 7.14557707e-01
7.38919258e-01 -1.27963328e+00 -6.15049124e-01 -3.90128940e-02
8.19393277e-01 -2.14770772e-02 2.62978990e-02 4.79904830e-01
-7.61709988e-01 1.09527278e+00 -2.23172233e-01 -8.68833423e-01
-1.30113375e+00 6.85641289e-01 1.33617699e-01 -7.19358921e-01
-4.64070261e-01 1.20802128e+00 1.54955655e-01 -4.99410719e-01
3.68468285e-01 -5.30585647e-01 -5.46687901e-01 2.11550578e-01
5.00978529e-01 6.49414435e-02 4.14700419e-01 -7.78363049e-01
-5.20693421e-01 5.78801274e-01 -5.49957395e-01 1.22784115e-01
1.57529962e+00 2.82149494e-01 -6.41951784e-02 -8.77827257e-02
1.37661672e+00 -7.88728476e-01 -6.63498223e-01 -5.32403171e-01
4.10525501e-01 -2.67168611e-01 2.25803047e-01 -7.49583662e-01
-1.18827999e+00 5.55878818e-01 9.69770968e-01 1.13375463e-01
7.70980299e-01 4.67397571e-01 3.18426520e-01 8.27875674e-01
4.79615003e-01 -1.08320475e+00 2.94733584e-01 5.75289190e-01
1.14515543e+00 -1.28441453e+00 2.52632052e-01 3.23056668e-01
-1.32830203e-01 1.13159120e+00 3.23197484e-01 -6.58243418e-01
8.98023665e-01 -5.29365897e-01 -3.01010072e-01 -7.04433382e-01
-6.61831379e-01 -6.70671523e-01 1.12280738e+00 3.72089267e-01
2.91191280e-01 -1.52555197e-01 -4.47707111e-03 5.83873391e-02
1.81118146e-01 -7.73770213e-02 -3.22814196e-01 9.11361694e-01
-6.40348017e-01 -7.98904240e-01 -3.12683851e-01 6.40298665e-01
-4.63232875e-01 -3.71876061e-01 -4.50578213e-01 9.66008961e-01
3.14249806e-02 3.99033844e-01 1.61588207e-01 -1.43440068e-01
4.48330045e-01 -2.58978754e-02 3.98785293e-01 -7.67153800e-01
-1.22433221e+00 -2.91801035e-01 -6.82836249e-02 -9.52287316e-01
-4.38693702e-01 -2.23155648e-01 -6.60064518e-01 3.41857851e-01
-6.90784454e-01 8.37216061e-03 9.16860878e-01 9.07392442e-01
4.51461732e-01 5.50136328e-01 1.69773534e-01 -1.21052837e+00
-3.15958172e-01 -1.13113499e+00 -3.33430052e-01 4.36274856e-01
3.11327577e-01 -6.15543365e-01 -1.56267046e-03 -5.47717392e-01] | [10.645170211791992, 0.6165434122085571] |
a53a92ab-6a1c-420e-908e-86246a14f172 | a-comparative-study-between-full-parameter | 2304.08109 | null | https://arxiv.org/abs/2304.08109v2 | https://arxiv.org/pdf/2304.08109v2.pdf | A Comparative Study between Full-Parameter and LoRA-based Fine-Tuning on Chinese Instruction Data for Instruction Following Large Language Model | Recently, the instruction-tuning of large language models is a crucial area of research in the field of natural language processing. Due to resource and cost limitations, several researchers have employed parameter-efficient tuning techniques, such as LoRA, for instruction tuning, and have obtained encouraging results In comparison to full-parameter fine-tuning, LoRA-based tuning demonstrates salient benefits in terms of training costs. In this study, we undertook experimental comparisons between full-parameter fine-tuning and LoRA-based tuning methods, utilizing LLaMA as the base model. The experimental results show that the selection of the foundational model, training dataset scale, learnable parameter quantity, and model training cost are all important factors. We hope that the experimental conclusions of this paper can provide inspiration for training large language models, especially in the field of Chinese, and help researchers find a better trade-off strategy between training cost and model performance. To facilitate the reproduction of the paper's results, the dataset, model and code will be released. | ['Xiangang Li', 'Baochang Ma', 'Yunjie Ji', 'Xianghui Sun'] | 2023-04-17 | null | null | null | null | ['instruction-following'] | ['natural-language-processing'] | [-3.45922917e-01 -5.44515371e-01 -8.29564333e-01 -6.15476370e-01
-7.53649771e-01 -4.49272126e-01 1.21264949e-01 2.73105502e-01
-6.71899796e-01 6.25751078e-01 -1.43877591e-03 -9.22828019e-01
6.12011105e-02 -7.41550744e-01 -5.85170329e-01 -3.78315151e-01
-3.62099186e-02 2.44292051e-01 3.78837377e-01 -2.11329058e-01
7.56118059e-01 5.69928586e-01 -1.26847637e+00 5.07569574e-02
9.18605506e-01 5.58307827e-01 3.59546870e-01 4.12928939e-01
-4.99853075e-01 6.11205399e-01 -6.01389289e-01 -1.57011449e-01
-3.45615298e-02 -1.67038769e-01 -6.44314408e-01 -3.20419133e-01
3.70147452e-02 -1.91213697e-01 -5.96393161e-02 8.20073664e-01
5.42322755e-01 1.21558078e-01 4.04144645e-01 -8.68511856e-01
-2.88812906e-01 8.39335203e-01 -7.11466789e-01 5.82623839e-01
-1.12976857e-01 1.25392854e-01 7.32350588e-01 -4.54339355e-01
-3.60377803e-02 1.40326059e+00 4.90727752e-01 4.03978825e-01
-9.90515172e-01 -8.59623075e-01 1.76844835e-01 -1.11068964e-01
-1.59253383e+00 -6.40455604e-01 4.33256686e-01 -2.89299339e-01
1.24941957e+00 -6.03810418e-03 1.50885731e-01 5.85100949e-01
7.77685165e-01 7.88060188e-01 1.02488422e+00 -9.61261451e-01
2.87854105e-01 3.91478598e-01 3.38923782e-01 7.95213282e-01
4.40200388e-01 4.37640771e-02 -5.49693227e-01 -3.00845802e-01
8.41576159e-01 -4.32366729e-01 2.25991070e-01 -2.77366191e-01
-1.07965612e+00 1.07393098e+00 1.68166310e-01 5.20677507e-01
1.01934761e-01 3.22735041e-01 4.85252589e-01 3.81671607e-01
3.14799756e-01 6.92762494e-01 -6.94654346e-01 -5.90896904e-01
-8.93273473e-01 9.51922610e-02 8.26653719e-01 1.03569674e+00
7.76763618e-01 2.54413813e-01 -7.71559775e-02 8.93909216e-01
4.22454476e-01 3.61521989e-01 9.54572082e-01 -9.42177653e-01
6.53062105e-01 5.23815870e-01 3.35337222e-02 -6.45614982e-01
-4.77794856e-01 -2.44926825e-01 -4.04480815e-01 -3.52870017e-01
3.22374046e-01 -3.15971315e-01 -4.53040242e-01 1.60810685e+00
9.88245476e-03 -2.27183253e-01 -7.44942129e-02 4.41052139e-01
4.42056298e-01 8.62666667e-01 3.88651490e-01 -2.77551681e-01
1.50683725e+00 -1.03810000e+00 -7.35124707e-01 -4.32476044e-01
1.27091384e+00 -1.13120461e+00 1.67442119e+00 6.18048832e-02
-1.11250103e+00 -7.13506937e-01 -1.21375203e+00 3.57353240e-02
-3.22119027e-01 2.57101893e-01 1.02364814e+00 1.04011071e+00
-9.22080994e-01 3.03492635e-01 -1.21663332e+00 -4.69684094e-01
2.39609182e-02 3.96331728e-01 3.56131613e-01 9.66181979e-03
-1.13454235e+00 8.25652778e-01 6.16088867e-01 -2.33372539e-01
-5.42136252e-01 -6.03143156e-01 -7.37304032e-01 2.42909908e-01
3.48013252e-01 -2.54424751e-01 1.44902802e+00 -3.55856210e-01
-1.56402445e+00 5.18274367e-01 -3.28786492e-01 -3.09874415e-01
-4.69703265e-02 -2.00845927e-01 -4.02559429e-01 -3.80249530e-01
-2.04010487e-01 5.10926545e-01 4.80791599e-01 -8.93041313e-01
-4.48593110e-01 -9.54340622e-02 7.22012892e-02 9.84048024e-02
-6.86469138e-01 4.50642586e-01 -6.82331502e-01 -5.61503887e-01
-2.50632405e-01 -8.93380702e-01 -2.48976454e-01 -6.96812928e-01
3.69485468e-02 -3.75179619e-01 3.82314950e-01 -1.12857252e-01
2.00492620e+00 -2.14386153e+00 -5.55831492e-01 2.76046604e-01
-1.30019322e-01 4.20122951e-01 -3.17619056e-01 3.39765757e-01
4.52938139e-01 3.83664429e-01 1.40801847e-01 -4.26178472e-03
-2.02761460e-02 1.26415044e-01 -1.90650761e-01 5.83421774e-02
3.30614038e-02 8.49308968e-01 -7.53286481e-01 -7.26741791e-01
6.49081841e-02 1.79696977e-01 -6.97556794e-01 2.51952797e-01
-2.03888163e-01 -1.04485817e-01 -9.09180582e-01 4.57783818e-01
2.61861414e-01 -3.45216602e-01 1.40218496e-01 -2.31042970e-02
-3.82046729e-01 4.96717393e-01 -9.64365005e-01 1.49011922e+00
-8.19881856e-01 6.10473990e-01 5.49268955e-03 -5.11514962e-01
1.14179468e+00 1.17556207e-01 5.56692332e-02 -8.26446652e-01
2.91978478e-01 3.06453258e-01 2.85752475e-01 -4.33401585e-01
9.51432347e-01 1.54022038e-01 -3.13857198e-01 7.75514185e-01
-3.41420352e-01 -4.15523797e-01 4.81313050e-01 1.52753457e-01
6.42163277e-01 1.31114706e-01 3.87247264e-01 -8.42393577e-01
4.11311924e-01 1.35530367e-01 2.10578427e-01 8.21807325e-01
-4.23821300e-01 -1.10361047e-01 4.11546856e-01 -2.60705769e-01
-9.65173721e-01 -5.14451563e-01 -4.09060478e-01 1.60159504e+00
-8.78434703e-02 -6.36945784e-01 -9.89651322e-01 -1.68229401e-01
-1.59539595e-01 8.62901449e-01 -2.35907853e-01 -1.77908674e-01
-1.05521572e+00 -1.00600040e+00 6.21466279e-01 6.96136892e-01
5.15667021e-01 -8.71285200e-01 -7.05500841e-01 1.97340831e-01
-7.91967437e-02 -8.43185067e-01 -8.21213186e-01 4.80282634e-01
-1.18349898e+00 -5.58346212e-01 -3.05757433e-01 -9.44379389e-01
5.08061767e-01 4.46875721e-01 1.27071679e+00 4.42635179e-01
2.99723428e-02 1.05399713e-01 -1.34260550e-01 -6.32062316e-01
-7.86860526e-01 6.28910661e-01 1.28918245e-01 -6.79877639e-01
9.17026639e-01 -6.26261085e-02 -2.36487523e-01 5.23968935e-01
-9.44294035e-01 -1.90592274e-01 6.71388090e-01 7.51522899e-01
4.96003270e-01 1.91816986e-01 6.61530495e-01 -1.07271266e+00
1.19598246e+00 -2.02281207e-01 -9.46629763e-01 5.68035960e-01
-1.00506032e+00 6.12714589e-01 6.30172133e-01 -4.17357862e-01
-1.10731614e+00 -3.28865349e-01 -1.02816746e-01 9.61147845e-02
-6.18867874e-02 7.30418503e-01 -1.10429712e-01 -2.24429354e-01
8.14540207e-01 2.04541609e-02 -2.45803937e-01 -3.66080940e-01
3.65964733e-02 7.21465468e-01 7.87489787e-02 -1.01271737e+00
4.05430615e-01 -2.12880298e-01 -3.72226268e-01 -8.28338742e-01
-6.82445168e-01 -4.69850540e-01 -5.64438343e-01 2.17091486e-01
5.28343320e-01 -1.08055544e+00 -3.17297012e-01 3.38741899e-01
-7.25142896e-01 -7.07809508e-01 1.51098341e-01 6.79943204e-01
-4.34516221e-01 2.65286386e-01 -7.86788523e-01 -4.77986604e-01
-2.12612659e-01 -1.54429650e+00 8.00175607e-01 4.28451598e-01
-3.39189678e-01 -1.29287457e+00 -7.25252032e-02 4.36465412e-01
8.67299139e-01 -5.65709233e-01 1.33836949e+00 -4.99790162e-01
-5.08038580e-01 -1.31999090e-01 -1.02310866e-01 6.07892945e-02
3.36910225e-02 1.92758650e-01 -7.88713396e-01 -5.62409759e-01
6.01914972e-02 -5.22613883e-01 4.19170439e-01 5.72648168e-01
1.33168292e+00 2.23943815e-02 -3.39884907e-01 5.18196702e-01
1.40574384e+00 2.96698093e-01 2.69128978e-01 5.05462110e-01
4.68843788e-01 2.65159458e-01 8.14218938e-01 3.70422900e-01
4.63723928e-01 5.86117506e-01 -2.87300050e-01 1.07598536e-01
1.35234132e-01 -2.13179737e-01 4.47000474e-01 1.58291864e+00
2.50293314e-01 -1.83352470e-01 -1.17496943e+00 3.19090396e-01
-1.45351362e+00 -4.30278897e-01 3.30586374e-01 2.36387420e+00
1.13924515e+00 4.40462768e-01 -2.53707096e-02 -7.67812654e-02
4.10003930e-01 5.83890937e-02 -4.64498222e-01 -9.59492445e-01
1.89333677e-01 6.27434021e-03 8.03718925e-01 7.83769011e-01
-7.47527063e-01 1.28954494e+00 7.23459387e+00 1.17204559e+00
-1.33006144e+00 -8.26522782e-02 8.12326550e-01 -7.25734383e-02
-3.11631501e-01 -2.42351182e-02 -1.49796426e+00 2.68221498e-01
1.78995132e+00 -5.16946554e-01 5.32158613e-01 7.03267038e-01
5.80264390e-01 -2.66238868e-01 -1.12588227e+00 6.20079875e-01
-1.73316002e-01 -1.10522330e+00 -1.84007604e-02 -4.75612804e-02
7.61237264e-01 9.52643454e-02 -1.26698181e-01 7.94109762e-01
3.48525941e-01 -8.55755746e-01 4.43117589e-01 -2.78068464e-02
5.95484972e-01 -8.42728674e-01 6.01328254e-01 6.23465598e-01
-1.27244389e+00 -8.34431127e-02 -6.94035888e-01 -2.42287934e-01
-9.68447998e-02 2.64471799e-01 -6.83519304e-01 3.84796411e-02
5.42744100e-01 1.46362245e-01 -7.12884665e-01 9.90526795e-01
8.05401728e-02 1.13020539e+00 -2.44144693e-01 -5.03356040e-01
2.21731678e-01 -7.07334429e-02 -2.15196207e-01 1.40798354e+00
7.13435858e-02 1.37439981e-01 3.12465668e-01 7.08049476e-01
-4.69374992e-02 4.66711819e-01 -3.11006993e-01 -4.12210405e-01
8.99112344e-01 9.46594834e-01 -8.88213873e-01 -3.09425563e-01
-5.88404775e-01 2.68915407e-02 3.83072317e-01 1.97377056e-01
-7.85027623e-01 -4.28546309e-01 3.90679955e-01 6.20883144e-02
7.62729496e-02 -5.71527839e-01 -7.22138405e-01 -1.13272178e+00
-1.84245333e-01 -1.07269895e+00 1.73033506e-01 -4.63810474e-01
-7.46843874e-01 5.30354679e-01 5.10719538e-01 -7.43533671e-01
-3.34238499e-01 -7.10847259e-01 -5.59794188e-01 9.34949219e-01
-1.55433917e+00 -6.13143027e-01 8.19860995e-02 1.71349823e-01
8.03854346e-01 -3.44964087e-01 8.90095055e-01 2.18786106e-01
-7.94146478e-01 1.15486896e+00 5.09403825e-01 1.20530669e-02
7.94260621e-01 -7.25179136e-01 3.38107705e-01 7.26407051e-01
-6.51122183e-02 1.12764025e+00 5.31703591e-01 -4.27891076e-01
-1.49319053e+00 -7.84915626e-01 1.05774677e+00 -4.13790673e-01
8.74118090e-01 -3.22540462e-01 -8.51813376e-01 6.09704971e-01
1.40753970e-01 -4.75298077e-01 8.28337371e-01 3.75449777e-01
1.38363969e-02 -2.56143987e-01 -7.86115229e-01 6.16272509e-01
4.20422673e-01 -4.46936429e-01 -3.74015868e-01 6.30190074e-02
8.99779677e-01 -5.15202105e-01 -1.20920289e+00 1.62923276e-01
4.41495776e-01 -5.95369577e-01 7.20138311e-01 -4.73067701e-01
4.40224335e-02 5.53308660e-03 -2.20478788e-01 -1.11798787e+00
-3.71813416e-01 -4.20943052e-01 4.97074351e-02 1.14971185e+00
6.21030509e-01 -5.20234823e-01 6.92273438e-01 7.66996980e-01
-8.52907822e-02 -9.60912764e-01 -4.04484957e-01 -7.96305001e-01
6.98725939e-01 -5.42074800e-01 8.47182214e-01 6.58329606e-01
-1.08819112e-01 5.08400977e-01 -7.14097694e-02 -9.33570340e-02
2.58156359e-01 1.09513253e-01 7.48236120e-01 -8.14590335e-01
-3.18937361e-01 -5.25604367e-01 2.39346087e-01 -1.31228042e+00
1.49271578e-01 -3.95299852e-01 1.29022926e-01 -8.55960310e-01
2.20819578e-01 -9.65602994e-01 -3.69090378e-01 5.28115809e-01
-3.51129115e-01 -1.84783012e-01 2.90128976e-01 3.75207841e-01
-4.81642723e-01 2.84276485e-01 1.30425787e+00 6.05074726e-02
-5.85132241e-01 1.31824479e-01 -8.72584462e-01 6.09723091e-01
1.05361009e+00 -3.78287017e-01 -6.59446120e-01 -7.79851317e-01
8.58507156e-02 -1.85266342e-02 -7.00868845e-01 -9.42696869e-01
2.79508621e-01 -5.92725098e-01 1.00982666e-01 -2.65766501e-01
4.69002873e-02 -5.24512947e-01 -4.63096410e-01 5.48367918e-01
-7.27206171e-01 5.87859511e-01 7.56972373e-01 1.33310303e-01
-2.29875416e-01 -5.61432123e-01 8.60684633e-01 -8.30204338e-02
-8.59247863e-01 5.41463345e-02 -4.98231411e-01 3.29216957e-01
7.79811263e-01 -1.76111862e-01 -2.87438720e-01 -4.21056375e-02
-1.24418281e-03 3.42221290e-01 5.33399940e-01 5.37854791e-01
2.28374138e-01 -1.03253424e+00 -6.37056768e-01 4.93562579e-01
6.94990903e-02 -1.93289205e-01 -1.22979984e-01 6.14506364e-01
-7.23548532e-01 1.09872997e+00 -5.57097122e-02 -5.25438428e-01
-1.18194723e+00 4.81782854e-01 3.62828113e-02 -4.90825981e-01
-2.78548896e-03 6.11952186e-01 1.53381422e-01 -4.24648613e-01
4.40491319e-01 -5.80747306e-01 -4.96196225e-02 -4.07026768e-01
6.54608190e-01 2.28719443e-01 1.16706736e-01 -3.61772746e-01
-1.93709299e-01 3.79900426e-01 -2.84944534e-01 1.73708826e-01
9.06680346e-01 -2.14473829e-01 -9.40870494e-02 6.79405212e-01
1.01846039e+00 2.65268534e-01 -9.39106286e-01 -4.33183134e-01
8.52371752e-02 -3.22359741e-01 2.98676670e-01 -6.67176306e-01
-8.04277003e-01 9.68194008e-01 6.33131027e-01 3.56438075e-04
1.02206683e+00 -3.17585289e-01 5.83855391e-01 6.32171631e-01
6.91737413e-01 -1.21182847e+00 -6.22729212e-02 9.39128458e-01
2.22956896e-01 -1.28810179e+00 1.86630175e-01 -1.10925980e-01
-3.73181492e-01 1.15822542e+00 9.91215825e-01 2.52144605e-01
7.26795316e-01 7.96227813e-01 3.25079322e-01 1.70542642e-01
-1.02038503e+00 4.70293641e-01 7.36143589e-02 2.28902161e-01
1.14643943e+00 2.31253952e-01 -5.13419092e-01 5.62905729e-01
-4.28742200e-01 1.26935229e-01 4.08999652e-01 1.06527758e+00
-8.41023147e-01 -1.50899816e+00 -5.48545241e-01 6.09211087e-01
-4.10227984e-01 -4.08298612e-01 2.21414194e-01 8.76656473e-01
-3.21688443e-01 9.70505178e-01 1.14112370e-01 -2.10913956e-01
1.66493759e-01 1.28725111e-01 3.97480130e-01 -5.73683023e-01
-8.59628558e-01 6.09998927e-02 2.27343477e-02 -3.85598838e-01
-1.07339911e-01 -2.65346169e-01 -1.36392796e+00 -9.00358796e-01
-6.81929529e-01 5.95710456e-01 7.38967478e-01 9.36697662e-01
3.28170627e-01 4.43632662e-01 4.63396758e-01 -5.12889624e-01
-1.07169139e+00 -1.02239668e+00 -4.41993177e-01 -2.29652196e-01
-1.35234401e-01 -3.85602802e-01 -5.52362837e-02 -5.46968505e-02] | [10.703971862792969, 8.361773490905762] |
10faa59d-d2fe-425b-8057-03a27f54904d | do-we-actually-need-dense-over | 2102.02887 | null | https://arxiv.org/abs/2102.02887v3 | https://arxiv.org/pdf/2102.02887v3.pdf | Do We Actually Need Dense Over-Parameterization? In-Time Over-Parameterization in Sparse Training | In this paper, we introduce a new perspective on training deep neural networks capable of state-of-the-art performance without the need for the expensive over-parameterization by proposing the concept of In-Time Over-Parameterization (ITOP) in sparse training. By starting from a random sparse network and continuously exploring sparse connectivities during training, we can perform an Over-Parameterization in the space-time manifold, closing the gap in the expressibility between sparse training and dense training. We further use ITOP to understand the underlying mechanism of Dynamic Sparse Training (DST) and indicate that the benefits of DST come from its ability to consider across time all possible parameters when searching for the optimal sparse connectivity. As long as there are sufficient parameters that have been reliably explored during training, DST can outperform the dense neural network by a large margin. We present a series of experiments to support our conjecture and achieve the state-of-the-art sparse training performance with ResNet-50 on ImageNet. More impressively, our method achieves dominant performance over the overparameterization-based sparse methods at extreme sparsity levels. When trained on CIFAR-100, our method can match the performance of the dense model even at an extreme sparsity (98%). Code can be found https://github.com/Shiweiliuiiiiiii/In-Time-Over-Parameterization. | ['Mykola Pechenizkiy', 'Decebal Constantin Mocanu', 'Lu Yin', 'Shiwei Liu'] | 2021-02-04 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [-1.20572396e-01 2.80324996e-01 -2.02752769e-01 -2.40929395e-01
-5.60396433e-01 -3.06552291e-01 4.18515086e-01 -3.44071716e-01
-4.26274866e-01 5.99138618e-01 1.99614108e-01 -2.28690043e-01
-2.84094602e-01 -6.29585385e-01 -1.06974268e+00 -6.97622061e-01
-5.37373424e-01 5.00511348e-01 8.78343061e-02 -1.21558771e-01
-8.78015459e-02 5.20558357e-01 -1.24408007e+00 2.79289991e-01
7.01219320e-01 7.49192953e-01 2.73627669e-01 3.95908833e-01
-4.32547219e-02 6.30882680e-01 -3.10330510e-01 -7.38107115e-02
5.49238503e-01 -3.21661353e-01 -7.18919337e-01 1.11085162e-01
7.24683583e-01 -1.73584148e-01 -6.49558604e-01 1.13338697e+00
3.76308650e-01 1.97120905e-01 2.40499914e-01 -1.00752938e+00
-4.49484020e-01 1.10539377e+00 -6.35498285e-01 5.80380201e-01
-1.65494338e-01 2.05023631e-01 1.02457809e+00 -1.08203554e+00
5.85807085e-01 1.06860054e+00 8.66075516e-01 3.27005446e-01
-1.40258694e+00 -8.16119671e-01 6.09136283e-01 2.80598309e-02
-1.70106232e+00 -5.97012281e-01 7.22890019e-01 -2.99392194e-01
1.09201074e+00 -6.09880686e-02 9.15016055e-01 8.31169248e-01
-9.31624845e-02 7.50493288e-01 7.81965554e-01 -2.17125803e-01
3.42608035e-01 -5.19410521e-02 2.03953445e-01 8.94098938e-01
4.74384934e-01 3.21831852e-02 -5.91197491e-01 -1.25134975e-01
1.08292449e+00 1.27033414e-02 -3.28373581e-01 -3.72217536e-01
-1.13403106e+00 9.40240026e-01 8.00351799e-01 7.21800029e-01
-2.81066447e-01 3.02403808e-01 2.95278192e-01 5.05099535e-01
3.82621557e-01 7.10646868e-01 -4.29862201e-01 -6.59034029e-02
-1.46881330e+00 3.17617208e-02 8.29696119e-01 8.32633078e-01
9.22425151e-01 5.85949481e-01 1.63511574e-01 8.56737375e-01
-1.03424275e-02 1.12146340e-01 6.41029775e-01 -1.08234167e+00
5.84303141e-01 3.56821895e-01 -4.18324202e-01 -8.93304408e-01
-2.47742102e-01 -1.05127192e+00 -1.23115373e+00 -7.33495178e-03
3.70949894e-01 -3.55887532e-01 -9.29798782e-01 1.87480795e+00
7.05404356e-02 6.15050316e-01 -3.17968652e-02 8.19190562e-01
6.02158427e-01 6.18506670e-01 -1.25581786e-01 -8.02206341e-03
1.00702524e+00 -1.05182612e+00 -2.12692246e-01 -4.40079361e-01
9.56708789e-01 -2.52130657e-01 1.08771110e+00 3.97272378e-01
-1.26846063e+00 -4.58711833e-01 -1.03044724e+00 2.22357348e-01
4.98430850e-03 -6.56314418e-02 8.81411314e-01 2.92245775e-01
-1.37749493e+00 8.54910731e-01 -1.13472199e+00 -4.07671571e-01
8.09142888e-01 5.48059881e-01 -5.01921117e-01 -3.55098367e-01
-9.53540504e-01 5.81588209e-01 4.85856533e-01 3.63907702e-02
-1.15195358e+00 -1.08914137e+00 -6.97225809e-01 4.08865333e-01
3.02821487e-01 -8.21078956e-01 1.02242982e+00 -1.27128172e+00
-1.21461630e+00 6.43211365e-01 -1.51248187e-01 -8.59276831e-01
3.32974494e-01 -2.12962687e-01 -5.36013804e-02 4.14424896e-01
-4.02135700e-02 8.80618691e-01 7.23477781e-01 -1.08654165e+00
-1.22684509e-01 -1.24404535e-01 -3.70060094e-02 2.80796409e-01
-4.44152296e-01 -4.11650628e-01 -5.45374870e-01 -7.00166702e-01
2.61085153e-01 -1.17681932e+00 -5.09367108e-01 -1.23269662e-01
-5.45431554e-01 1.49994381e-02 6.31661475e-01 -1.96574926e-01
1.08299077e+00 -2.22729278e+00 2.24523962e-01 3.96088004e-01
5.40266573e-01 3.24695170e-01 -4.46121424e-01 3.12647492e-01
-3.97187084e-01 2.33210936e-01 -2.74133474e-01 -6.05465174e-01
-3.68986130e-01 3.85634333e-01 -2.19741032e-01 6.24029458e-01
1.94615453e-01 9.37215805e-01 -6.98161721e-01 -3.79223973e-01
-3.99270803e-02 6.33887291e-01 -8.79245520e-01 -5.49163856e-02
1.35973888e-02 3.89591217e-01 -5.35659909e-01 5.32189548e-01
5.61948955e-01 -7.41117716e-01 2.42200315e-01 -1.85447797e-01
3.12097929e-02 2.01532677e-01 -1.08496034e+00 1.88016152e+00
-4.19181228e-01 6.86642706e-01 2.49291897e-01 -1.38759971e+00
5.54463565e-01 2.21686244e-01 6.82151139e-01 -5.09207845e-01
1.00947589e-01 2.73486733e-01 2.05099881e-01 -1.09602146e-01
1.83782175e-01 -8.00339803e-02 4.10919011e-01 3.40924114e-01
3.75139773e-01 1.12808518e-01 2.98739225e-01 5.59463382e-01
1.19190943e+00 -4.61980879e-01 -4.03065085e-02 -7.16969252e-01
1.05205245e-01 -1.21284299e-01 3.92779976e-01 9.05189276e-01
1.55455828e-01 6.29151464e-01 6.38207972e-01 -3.72775882e-01
-1.17446280e+00 -7.90650070e-01 -1.92058444e-01 1.00342941e+00
-2.69941926e-01 -3.22908759e-01 -6.53962553e-01 -5.53807020e-01
5.92352934e-02 3.44875276e-01 -7.94132888e-01 5.04212230e-02
-6.88509941e-01 -7.18096137e-01 4.91099894e-01 4.49506432e-01
6.27588511e-01 -7.95573533e-01 -1.57124788e-01 1.37575373e-01
2.43049070e-01 -1.16047525e+00 -4.77770269e-01 5.22133529e-01
-1.30211830e+00 -7.78298080e-01 -9.97618973e-01 -8.56928825e-01
1.02042210e+00 5.08696795e-01 1.27739632e+00 4.25822258e-01
2.46102945e-03 1.69590965e-01 -1.59631476e-01 1.80284157e-01
-7.12864697e-02 6.63242579e-01 -1.53307021e-02 -2.62371033e-01
-5.00808284e-02 -1.21360528e+00 -6.74391031e-01 8.01660866e-02
-9.25070524e-01 2.38283530e-01 6.72084451e-01 8.36205781e-01
6.82028234e-01 1.85474427e-03 6.41750216e-01 -1.17280746e+00
3.20167542e-01 -9.02329862e-01 -4.75159168e-01 -8.47921968e-02
-5.33913076e-01 2.99313515e-01 7.39001572e-01 -6.60685718e-01
-6.60770714e-01 5.73137328e-02 -3.47045422e-01 -9.23451841e-01
1.31883070e-01 8.05811405e-01 1.96282744e-01 -3.95301074e-01
7.62289584e-01 2.83282250e-01 8.40642825e-02 -5.82519889e-01
3.66472840e-01 -1.42461881e-01 2.98922658e-01 -7.83416629e-01
8.65096688e-01 5.38798034e-01 -6.73784316e-02 -8.44884455e-01
-1.15564120e+00 -3.55883747e-01 -4.42914456e-01 1.60450906e-01
2.65971512e-01 -1.31559038e+00 -2.79444605e-01 3.14556062e-01
-6.62684321e-01 -6.85557246e-01 -7.07058966e-01 4.42579538e-01
-3.98375869e-01 2.83387780e-01 -6.75101638e-01 -2.82165259e-01
-3.22298169e-01 -9.97606933e-01 6.68520272e-01 2.59248484e-02
-5.11455052e-02 -1.21520805e+00 1.02023892e-02 -4.48454022e-02
6.70918882e-01 8.05684477e-02 6.74924195e-01 -8.46430123e-01
-7.12987244e-01 -7.89225921e-02 -2.49411985e-01 3.90417457e-01
-2.47291625e-01 -2.48022288e-01 -7.82796264e-01 -6.80420458e-01
6.17084838e-02 -4.27942425e-01 1.23883104e+00 6.68711007e-01
1.49234509e+00 -6.92668974e-01 -2.38297403e-01 1.25270998e+00
1.67012930e+00 -3.01225811e-01 6.26292229e-01 2.40411967e-01
8.50992203e-01 1.69321492e-01 -1.01906694e-01 3.76501828e-01
1.78409010e-01 3.47939849e-01 4.11186069e-01 -2.34847814e-01
-3.45196754e-01 -3.42326522e-01 1.34499580e-01 1.01678336e+00
7.13178068e-02 -5.18628322e-02 -8.75061870e-01 7.80309498e-01
-1.62547624e+00 -8.34832430e-01 2.86814541e-01 1.99786270e+00
1.07398927e+00 2.56943345e-01 8.15941170e-02 3.20161767e-02
3.96871626e-01 3.48740578e-01 -7.38719940e-01 -1.03264369e-01
-1.89121246e-01 3.10977995e-01 7.65787840e-01 6.77651763e-01
-7.21291959e-01 1.13733625e+00 6.83433199e+00 1.07103968e+00
-1.23782599e+00 1.85813785e-01 9.32637095e-01 -5.70088625e-01
-4.91291255e-01 -1.09308593e-01 -9.66060221e-01 3.73358577e-01
9.70912635e-01 -5.52571602e-02 6.83380485e-01 1.01519692e+00
3.49059477e-02 -3.50399725e-02 -1.07067311e+00 9.72452998e-01
-1.02235928e-01 -1.90623307e+00 1.10259190e-01 1.01703472e-01
1.12013984e+00 7.80300677e-01 1.51132926e-01 2.60343254e-01
4.00539279e-01 -1.25158644e+00 5.35941899e-01 1.69350430e-01
8.59949529e-01 -4.25946236e-01 3.83367002e-01 4.30964708e-01
-1.14482927e+00 -5.67006283e-02 -5.07124245e-01 1.10994659e-01
-5.85455634e-02 1.19950056e+00 -6.63475335e-01 6.51782379e-02
6.32039607e-01 1.07891870e+00 -5.67425013e-01 1.20824015e+00
-1.64641365e-02 9.58524048e-01 -7.37781763e-01 2.97075123e-01
4.96328503e-01 -1.22163840e-01 5.47025740e-01 1.27679431e+00
5.29918492e-01 5.45099117e-02 1.97150707e-01 8.64924729e-01
-3.76316339e-01 -1.34022251e-01 -6.57807767e-01 -1.64302066e-01
6.37549460e-01 9.60502207e-01 -8.17529619e-01 -5.11716008e-01
-2.50920892e-01 5.30168056e-01 5.67419767e-01 6.41483307e-01
-5.86702049e-01 -1.53984092e-02 5.91965199e-01 3.29002857e-01
6.45035982e-01 -4.99485016e-01 -4.53257680e-01 -1.38756347e+00
-1.24099150e-01 -9.89148021e-01 2.34811783e-01 -4.77238387e-01
-1.19588470e+00 7.75069892e-01 6.09506220e-02 -9.03941453e-01
-5.10453433e-02 -2.29447201e-01 -6.61408186e-01 7.57325292e-01
-1.55291820e+00 -8.16942036e-01 -1.83526188e-01 7.76289284e-01
5.61418831e-01 -2.43328094e-01 4.91209239e-01 4.07556415e-01
-6.41856432e-01 7.00292110e-01 1.79147109e-01 -4.36462574e-02
2.11529568e-01 -9.06991243e-01 4.61962819e-01 8.44657123e-01
3.57149750e-01 7.95789838e-01 5.99780977e-01 -4.86841530e-01
-1.28408062e+00 -9.55287933e-01 5.91431320e-01 1.34585099e-02
9.34337854e-01 -2.65050083e-01 -1.19437945e+00 8.66554260e-01
1.25532225e-01 2.82863706e-01 3.82668406e-01 4.12183583e-01
-5.26992142e-01 -1.06068872e-01 -9.40604746e-01 3.47687989e-01
1.29160750e+00 -3.62307489e-01 -3.53295296e-01 5.35726130e-01
8.58725190e-01 -4.06980544e-01 -8.88894737e-01 2.69449919e-01
2.85288364e-01 -9.35784459e-01 9.93331075e-01 -4.20466900e-01
3.80971581e-01 4.62220609e-03 -6.22618459e-02 -1.33397269e+00
-5.05946636e-01 -7.73438990e-01 -2.88330019e-01 7.10959792e-01
4.12978858e-01 -7.94288218e-01 1.20317376e+00 2.62345284e-01
-3.92601907e-01 -1.14816988e+00 -1.00701010e+00 -9.13906515e-01
3.60541075e-01 -3.49657387e-01 3.72530878e-01 1.09891844e+00
-2.14281380e-01 6.84576705e-02 -2.24404678e-01 1.09718949e-01
6.34698153e-01 -2.08721533e-01 3.75545293e-01 -1.08656394e+00
-5.53478420e-01 -5.07840574e-01 -2.34394863e-01 -1.38239980e+00
2.34906301e-01 -9.88972783e-01 -2.67552286e-01 -1.25171804e+00
1.77331269e-01 -7.92937398e-01 -2.46943548e-01 7.74905562e-01
1.22170240e-01 2.83810556e-01 3.28161508e-01 5.93908191e-01
-5.26722670e-01 5.60002208e-01 1.39294958e+00 5.04630841e-02
-2.11787745e-01 -2.18788713e-01 -9.99126852e-01 6.03002846e-01
9.35107529e-01 -5.93589485e-01 -5.91995001e-01 -7.20306814e-01
3.39130402e-01 -9.21597611e-03 2.51788586e-01 -1.26778305e+00
2.77536273e-01 2.42523476e-02 2.29469463e-01 -1.81623966e-01
3.17973018e-01 -6.82878375e-01 1.98990047e-01 5.09114027e-01
-5.37970304e-01 1.10381193e-01 3.03846955e-01 4.25141275e-01
-2.57485539e-01 -1.89040557e-01 9.74125981e-01 -3.28012794e-01
-4.58284497e-01 6.98895872e-01 -1.82391524e-01 4.20185030e-01
6.00642085e-01 -3.24700862e-01 -4.38588619e-01 -2.43657410e-01
-9.28584397e-01 1.79110736e-01 4.86515254e-01 1.60693284e-02
4.84068871e-01 -1.22293723e+00 -5.97020805e-01 4.57905024e-01
-3.71207118e-01 2.50157565e-01 4.03621674e-01 9.84227300e-01
-5.06860495e-01 3.22046548e-01 -9.56874564e-02 -6.34270966e-01
-8.33439410e-01 3.02285433e-01 5.50202727e-01 -3.84311825e-01
-9.47057307e-01 1.17810249e+00 3.35195392e-01 6.22927211e-02
4.03494090e-01 -2.64237553e-01 9.18482840e-02 -2.24829882e-01
3.03235114e-01 -1.75527409e-02 1.64897472e-01 -3.04190606e-01
-1.68524444e-01 4.66809273e-01 -1.85309857e-01 6.27160221e-02
1.62035084e+00 1.45640343e-01 3.19714695e-02 2.88219750e-01
1.52622831e+00 -1.80079639e-01 -1.60250175e+00 -4.07004058e-01
-3.80874425e-01 -4.30706531e-01 3.29102337e-01 -4.80223328e-01
-1.80341804e+00 8.27154279e-01 3.54195356e-01 3.31078291e-01
8.82441878e-01 1.34026229e-01 9.03741002e-01 6.33158445e-01
2.19881594e-01 -7.23669231e-01 2.34957144e-01 7.48233855e-01
8.63956690e-01 -1.07933939e+00 9.60679948e-02 -2.21539080e-01
-6.03464723e-01 7.35532939e-01 4.38514143e-01 -7.87249267e-01
1.06528711e+00 4.41411287e-01 -3.80237103e-01 -3.39781433e-01
-9.70599294e-01 5.84011339e-02 2.41700813e-01 4.18983579e-01
3.84913564e-01 -1.21584974e-01 1.09747477e-01 2.21429661e-01
-3.78862828e-01 -6.55461997e-02 3.28771770e-01 6.21585190e-01
-6.49280548e-01 -7.97843814e-01 -3.78796309e-02 6.85520589e-01
-4.36745852e-01 -5.26495516e-01 6.99624419e-02 8.40491772e-01
-1.90053791e-01 3.72378737e-01 1.38374031e-01 -2.66134650e-01
3.05491127e-02 -1.27804622e-01 4.53055114e-01 -7.81111479e-01
-5.61735868e-01 7.34792848e-05 6.25675917e-02 -7.38163054e-01
-3.34122837e-01 -5.04079819e-01 -1.16575980e+00 -5.55597186e-01
-1.65944427e-01 2.97981530e-01 4.65102941e-01 8.28641832e-01
6.35993242e-01 4.70088124e-01 5.75033247e-01 -1.08886182e+00
-5.83807230e-01 -8.33844781e-01 -7.01577723e-01 2.32020065e-01
3.71803135e-01 -5.10389864e-01 -7.84708500e-01 -2.24156484e-01] | [8.526068687438965, 3.406461000442505] |
869cde5c-fa77-4b43-9051-2f389a103843 | ranking-news-feed-updates-on-social-media-a | null | null | https://www.researchgate.net/publication/339043426_Ranking_news_feed_updates_on_social_media_A_comparative_study_of_supervised_models | https://www.researchgate.net/publication/339043426_Ranking_news_feed_updates_on_social_media_A_comparative_study_of_supervised_models | Ranking news feed updates on social media: A comparative study of supervised models | Social media users are overwhelmed by a large number of updates displayed chronologically in their news feed. Moreover, most updates are irrelevant. Ranking news feed updates by relevance has been proposed to help users catch up with the content they may find interesting. For this matter, supervised learning models have been commonly used to predict relevance. However, no comparative study was made to determine the most suitable models. In this work, we select, analyze, and compare six supervised learning algorithms applied to this case study. Experimental results on Twitter highlight that ensemble learning models are the most appropriate to predict the relevance of updates. | ['Omar Boussaid', 'Kamel Boukhalfa', 'Sami Belkacem'] | 2020-01-01 | null | null | null | conference-on-knowledge-extraction-and | ['social-media-popularity-prediction', 'social-media-popularity-prediction'] | ['miscellaneous', 'time-series'] | [-2.41014376e-01 -6.55793399e-02 -4.83573586e-01 -2.27871016e-01
-2.67610729e-01 -2.38528207e-01 8.87823641e-01 1.18853748e+00
-5.79565465e-01 1.06550419e+00 2.96382308e-01 1.16554208e-01
-3.27235222e-01 -8.20052445e-01 -2.34553635e-01 -1.25738621e-01
-1.72015548e-01 4.27811623e-01 2.71381199e-01 -7.32701540e-01
7.27234006e-01 -9.68688801e-02 -2.11176467e+00 5.37825286e-01
1.03487396e+00 8.57587159e-01 1.58920601e-01 2.38242671e-01
-5.88721037e-01 1.10797822e+00 -9.02470171e-01 -2.98528045e-01
-3.26342493e-01 -4.64981407e-01 -7.21733212e-01 -5.78134298e-01
2.68579900e-01 -1.40046179e-01 2.37322062e-01 7.12416470e-01
3.59001487e-01 1.52760431e-01 5.84691346e-01 -1.07788610e+00
-1.42843321e-01 9.73568261e-01 -2.68828660e-01 7.68272460e-01
8.82460177e-01 -8.78046274e-01 1.02132869e+00 -8.54138613e-01
6.72182441e-01 7.36327887e-01 4.85272437e-01 4.61205319e-02
-7.47107387e-01 -5.22264004e-01 5.39676011e-01 4.93289262e-01
-1.11966348e+00 -9.30024758e-02 7.84324348e-01 -4.95961875e-01
7.39153922e-01 7.01760411e-01 9.93245304e-01 1.02201557e+00
4.84859407e-01 4.58443135e-01 1.13174617e+00 -6.16958618e-01
3.70625526e-01 6.47659779e-01 4.37691361e-01 2.39842162e-02
5.28898895e-01 -3.19644064e-01 -8.61838996e-01 -4.54564184e-01
-1.24549471e-01 1.52202800e-01 -1.44552186e-01 3.22641551e-01
-4.72473651e-01 9.36048031e-01 3.70332062e-01 6.89024985e-01
-5.11287928e-01 -6.78768396e-01 3.52842242e-01 6.97857797e-01
1.27687430e+00 8.39405417e-01 -2.99629301e-01 -2.20871642e-01
-8.11615825e-01 1.96665540e-01 9.72650290e-01 3.94997120e-01
6.46592677e-01 -2.94617563e-01 -7.24530369e-02 7.28220522e-01
1.45029277e-01 1.30772501e-01 6.50468946e-01 -2.28948951e-01
1.55666731e-02 8.12942863e-01 1.91196665e-01 -1.39100385e+00
-5.97367465e-01 -8.54831994e-01 -4.59417522e-01 -9.42639336e-02
2.28689052e-02 -2.57905453e-01 -1.11466132e-01 9.00331497e-01
1.71916112e-01 1.07041001e-01 -1.20545000e-01 6.04185641e-01
1.14506936e+00 8.17347109e-01 2.70576805e-01 -7.77367830e-01
1.07103455e+00 -4.42308038e-01 -9.65192139e-01 -2.07497910e-01
3.68523538e-01 -1.10320568e+00 6.87297583e-01 4.96626645e-01
-7.88748145e-01 -4.00255650e-01 -1.00108671e+00 4.39794779e-01
-6.42538130e-01 -2.64110088e-01 4.73629355e-01 4.18215215e-01
-6.74776018e-01 8.13626230e-01 -2.88843781e-01 -6.23959363e-01
-1.30919576e-01 7.50772804e-02 2.12944314e-01 4.24995780e-01
-1.73276567e+00 1.31624079e+00 6.54530942e-01 -2.91554719e-01
1.27332941e-01 -3.28652054e-01 -2.04310358e-01 -2.43939549e-01
1.44810513e-01 -4.78258371e-01 1.37122238e+00 -1.09041142e+00
-1.24872255e+00 4.53685910e-01 -2.05828343e-02 -5.57389796e-01
4.90957290e-01 -4.63185042e-01 -9.28684056e-01 -9.36287493e-02
1.12251706e-01 -8.39130133e-02 8.09643209e-01 -9.75231707e-01
-1.16507101e+00 -1.44689009e-01 1.13645993e-01 5.41621625e-01
-6.83858454e-01 2.83171028e-01 -2.57485926e-01 -6.02348089e-01
1.50552407e-01 -5.78396857e-01 -1.99109111e-02 -8.39858294e-01
1.71545774e-01 -3.68356377e-01 6.87742293e-01 -4.32823509e-01
2.00101256e+00 -1.67214632e+00 -4.36787426e-01 3.43351513e-01
2.25947812e-01 1.04235888e-01 5.01608610e-01 7.82120407e-01
2.91857809e-01 2.63344049e-01 5.31385779e-01 6.39512166e-02
-4.72554684e-01 2.13919915e-02 -3.58876199e-01 8.93624201e-02
-2.76142478e-01 2.69106388e-01 -1.00268304e+00 -4.64677721e-01
-7.08612241e-03 4.95947868e-01 -8.42366442e-02 9.04609859e-02
-2.04128072e-01 4.62519258e-01 -7.48779297e-01 3.44506443e-01
2.38196522e-01 -4.97536451e-01 2.18450770e-01 -2.77752001e-02
-3.37277859e-01 4.78870481e-01 -8.09859931e-01 7.00023890e-01
-3.60498488e-01 7.19684422e-01 -5.65552473e-01 -6.58998847e-01
9.97338474e-01 9.08997133e-02 6.22384667e-01 -9.56680954e-01
3.11392993e-01 1.75457373e-01 -1.84347734e-01 -6.15829349e-01
9.22787905e-01 2.08403081e-01 3.21361534e-02 6.69569671e-01
-5.80270588e-01 4.31371391e-01 4.63296562e-01 3.51767331e-01
6.19463503e-01 -3.25332969e-01 7.68922329e-01 -2.00191498e-01
5.29461920e-01 1.59991875e-01 1.74400121e-01 8.01574349e-01
1.52164742e-01 1.88269243e-01 2.50076592e-01 -6.45575285e-01
-5.35936356e-01 -4.65204686e-01 -1.92091510e-01 1.57001722e+00
5.09102643e-01 -8.37764561e-01 -5.38541436e-01 -5.43515563e-01
-1.95845157e-01 9.05720055e-01 -5.40559411e-01 1.36236459e-01
-3.15728873e-01 -8.56108725e-01 -3.09442312e-01 -2.67275628e-02
9.72921923e-02 -9.18131471e-01 -6.79848790e-01 5.21101892e-01
-5.82500577e-01 -6.33637607e-01 2.26327907e-02 7.05950558e-02
-9.27042425e-01 -1.09626067e+00 -4.42272842e-01 -3.91826779e-01
6.21470690e-01 3.44252259e-01 1.15819120e+00 2.97698736e-01
2.95518994e-01 3.22706878e-01 -1.04899538e+00 -7.32374907e-01
-3.55460078e-01 4.99880582e-01 2.09792435e-01 7.08966851e-02
4.81151998e-01 -2.06886560e-01 -4.03919071e-01 3.80528659e-01
-7.28662312e-01 -1.68049619e-01 3.31425577e-01 4.93720055e-01
3.21492016e-01 2.24775329e-01 8.98860574e-01 -1.30905521e+00
1.31047678e+00 -1.08336079e+00 -4.02186066e-01 2.57694691e-01
-1.30266428e+00 -8.49371031e-02 5.25875807e-01 -4.10356253e-01
-1.07570875e+00 -5.21453083e-01 -1.86177462e-01 6.01747334e-01
1.73466206e-01 1.14588785e+00 7.58339524e-01 1.75822005e-02
1.11223233e+00 -4.95732650e-02 -1.86261073e-01 -5.37745953e-01
-2.03094468e-01 8.24528217e-01 -1.49799272e-01 9.04484245e-04
2.47645512e-01 1.58953831e-01 -3.96257013e-01 -8.78726423e-01
-9.75246072e-01 -7.70487189e-01 -2.83268929e-01 -8.51031899e-01
1.60281211e-01 -8.51894736e-01 -2.95442373e-01 1.57230616e-01
-9.16042507e-01 4.30170000e-01 -4.97513302e-02 6.90372527e-01
-4.21565063e-02 5.24070449e-02 -3.12160701e-01 -8.41080964e-01
-6.11114144e-01 -5.11827648e-01 1.28338322e-01 5.44887424e-01
-8.15953076e-01 -9.18741643e-01 2.55541742e-01 -2.27847807e-02
1.07513070e+00 2.54991204e-01 4.01603818e-01 -1.25223434e+00
-6.07206188e-02 -7.40189850e-01 2.23793149e-01 -2.99930036e-01
2.89212704e-01 5.25101125e-02 -8.75274003e-01 -8.67845416e-02
-2.49671657e-02 2.67773449e-01 6.87955797e-01 2.98003405e-01
7.35971332e-01 -7.31141508e-01 -6.56748533e-01 -1.83923349e-01
1.08039916e+00 2.32710123e-01 2.57683963e-01 9.00902629e-01
-9.38482489e-03 7.17239857e-01 1.08462906e+00 8.02875996e-01
2.84066141e-01 6.20551825e-01 3.87368381e-01 4.27062213e-01
3.95649254e-01 -2.52583295e-01 3.36894095e-01 8.81830156e-01
-1.96664959e-01 -3.61855298e-01 -7.35217988e-01 -6.91085383e-02
-1.93923807e+00 -1.07796359e+00 -4.66133684e-01 2.22609019e+00
6.32077813e-01 6.31466210e-01 4.18543220e-01 2.95748264e-01
8.45397830e-01 1.18015707e-01 -6.28776625e-02 -2.17701867e-01
-2.52752323e-02 -1.40203595e-01 1.93835586e-01 4.95918423e-01
-1.08603811e+00 3.00706953e-01 6.83213902e+00 5.65081239e-01
-1.42559135e+00 1.68779522e-01 6.60556734e-01 -2.20092148e-01
-3.47840041e-01 -1.55866891e-01 -9.96138871e-01 8.35776925e-01
1.04840863e+00 -6.55435205e-01 -2.87761331e-01 1.03260541e+00
3.09850007e-01 -4.26364034e-01 -5.87332845e-01 8.93988609e-01
2.55762011e-01 -1.34595788e+00 -1.04897775e-01 -2.93344408e-01
7.82861471e-01 -2.36752424e-02 -9.61121321e-02 2.65621156e-01
-1.80594191e-01 -3.91236305e-01 4.27654743e-01 9.18636382e-01
1.24458991e-01 -7.45039225e-01 1.06804585e+00 5.39092124e-01
-7.31897593e-01 -1.08968861e-01 -1.21198744e-01 -6.16003871e-01
1.05496936e-01 8.53411078e-01 -1.04894555e+00 3.83455336e-01
9.51355577e-01 6.60970330e-01 -8.21364880e-01 1.45568359e+00
-2.87737548e-01 6.50801301e-01 -3.02757889e-01 -7.86662757e-01
-1.07430227e-01 9.49587375e-02 5.64347863e-01 1.13252664e+00
5.06248236e-01 -2.37401817e-02 1.75225791e-02 5.41931167e-02
2.08722964e-01 9.24906671e-01 -4.47511494e-01 1.53301507e-01
3.85970831e-01 1.22744584e+00 -9.72623825e-01 -4.90267724e-01
-3.35191280e-01 5.03723204e-01 4.89748642e-02 3.44417244e-02
-5.00492156e-01 -1.02488600e-01 1.16033494e-01 8.38709891e-01
-3.66763741e-01 2.92155236e-01 -8.35172012e-02 -8.51006866e-01
-1.86594784e-01 -4.78576332e-01 6.89908504e-01 -4.82347757e-01
-1.48023570e+00 1.06034315e+00 3.15320492e-01 -1.59034157e+00
-5.07391632e-01 -1.74234733e-01 -8.73519182e-01 5.59560776e-01
-1.49437141e+00 -3.88460726e-01 -5.43897152e-01 1.66597664e-01
4.67656076e-01 -2.56632775e-01 6.37967050e-01 1.48280784e-01
-3.39525253e-01 2.21058950e-01 3.91937137e-01 -5.41643262e-01
8.24111581e-01 -1.07045722e+00 -8.02325178e-03 3.03916693e-01
-4.59361523e-02 5.32407284e-01 1.11448812e+00 -9.27867651e-01
-6.32113636e-01 -7.04846382e-01 1.41710246e+00 -2.56186992e-01
6.09342217e-01 4.31828976e-01 -1.00228047e+00 2.48297915e-01
4.37316775e-01 -4.49885905e-01 1.03742063e+00 5.03524601e-01
1.32837677e-02 -3.04146379e-01 -8.63516927e-01 4.38575745e-01
3.23100507e-01 -5.23857363e-02 -7.13945389e-01 2.99193829e-01
2.18859583e-01 -3.71815473e-01 -8.16311777e-01 3.85672957e-01
7.21455336e-01 -1.00590718e+00 6.31493568e-01 -2.77550459e-01
2.40004465e-01 -1.42846450e-01 3.28238249e-01 -1.55005503e+00
-2.37217069e-01 -4.77528036e-01 -3.46020043e-01 1.11330819e+00
6.74581707e-01 -7.06872225e-01 6.70398653e-01 4.79721069e-01
2.68810183e-01 -6.05510235e-01 -4.57210183e-01 -4.25220191e-01
-4.28893417e-01 -3.10484827e-01 5.18607855e-01 1.10454535e+00
6.14992142e-01 6.25998497e-01 -5.68502069e-01 -2.24008456e-01
2.77202576e-01 -1.14513654e-02 5.31679630e-01 -1.97373295e+00
1.51675671e-01 -7.75235653e-01 -1.61497772e-01 -4.69511241e-01
-3.65244657e-01 -6.24756634e-01 -2.94930458e-01 -1.69408882e+00
1.49400234e-01 -2.68928826e-01 -6.36353135e-01 7.23237079e-03
-4.80426580e-01 -1.52351204e-02 -6.39964491e-02 5.94587982e-01
-8.91642392e-01 6.07668236e-02 6.90338254e-01 -1.11885689e-01
-5.69301009e-01 8.00609350e-01 -9.46396470e-01 8.23677480e-01
1.23166728e+00 -7.05728054e-01 -5.14791429e-01 4.14396487e-02
1.09504211e+00 -2.54805654e-01 -3.54151756e-01 -8.62200141e-01
6.83458865e-01 -2.11751357e-01 4.06420827e-01 -7.97524810e-01
4.56032492e-02 -8.69085908e-01 3.98708433e-01 3.89980674e-01
-5.31263173e-01 1.82998732e-01 3.54736112e-02 5.07455945e-01
-5.40143788e-01 -7.08999634e-01 3.03560346e-01 5.40043600e-02
-9.84837413e-01 -2.83421427e-02 -7.24986672e-01 -6.87588453e-02
1.01185822e+00 -1.89419582e-01 -2.33595058e-01 -8.56233180e-01
-7.60325134e-01 1.86573222e-01 7.79675543e-02 7.00422108e-01
6.26973689e-01 -1.18862617e+00 -7.96340466e-01 -3.05975139e-01
5.47405362e-01 -5.95621884e-01 1.62854850e-01 7.36541271e-01
-5.30292451e-01 2.87269264e-01 -2.05447674e-01 -3.66668344e-01
-1.41191149e+00 3.23122174e-01 -2.09886104e-01 -2.96735138e-01
-2.88145393e-01 6.12335861e-01 -6.93107963e-01 1.51077330e-01
2.98631221e-01 1.46485955e-01 -1.21000195e+00 1.09536898e+00
9.73460615e-01 6.75248981e-01 5.16555667e-01 -6.88836992e-01
-2.83683479e-01 3.01878214e-01 -4.17371869e-01 2.46062949e-01
1.16567159e+00 -4.53142971e-01 -4.24576737e-02 7.03597128e-01
8.55153799e-01 8.28321204e-02 -4.72230077e-01 -3.41571748e-01
4.78600174e-01 -3.68442297e-01 8.56248662e-02 -7.18783200e-01
-5.91288328e-01 1.68455765e-01 5.06687760e-01 1.09798467e+00
1.27260435e+00 -5.52244596e-02 3.02271336e-01 4.70929444e-01
4.07774985e-01 -1.49723172e+00 3.09080929e-02 6.75186038e-01
8.78209591e-01 -1.42761528e+00 4.05485541e-01 -2.67033756e-01
-7.24346220e-01 1.31072044e+00 5.58354080e-01 2.50156611e-01
1.07735252e+00 6.96062967e-02 -3.27000045e-03 -1.10556230e-01
-8.01820874e-01 -2.32119337e-01 4.41853076e-01 2.28878051e-01
8.34772289e-01 -1.15959860e-01 -1.13076758e+00 5.49085021e-01
-2.99101651e-01 -2.68207025e-02 6.21999502e-01 1.09358156e+00
-9.49549139e-01 -1.08908153e+00 -5.22753298e-01 9.04389262e-01
-8.17574441e-01 8.81518051e-02 -5.16341388e-01 5.22298813e-01
-8.73248428e-02 1.30994558e+00 9.23164487e-02 -6.46989584e-01
3.11280549e-01 -7.60783814e-03 -1.24034233e-01 -4.66653675e-01
-7.58349597e-01 8.61829668e-02 2.30107963e-01 -2.19158344e-02
-5.87780058e-01 -5.07424772e-01 -9.24701333e-01 -4.11518335e-01
-5.90655625e-01 7.31434107e-01 7.29615271e-01 9.54520524e-01
2.83295631e-01 3.78948599e-01 8.20420682e-01 -3.81733835e-01
-1.61769688e-01 -1.09408116e+00 -3.26465100e-01 3.23175669e-01
7.61366040e-02 -5.82852304e-01 -4.20797408e-01 -1.62480712e-01] | [10.197062492370605, 6.156817436218262] |
cf957b2b-9d86-4517-86cd-290bd0e9a78e | ca-centripetalnet-a-novel-anchor-free-deep | 2307.04103 | null | https://arxiv.org/abs/2307.04103v1 | https://arxiv.org/pdf/2307.04103v1.pdf | CA-CentripetalNet: A novel anchor-free deep learning framework for hardhat wearing detection | Automatic hardhat wearing detection can strengthen the safety management in construction sites, which is still challenging due to complicated video surveillance scenes. To deal with the poor generalization of previous deep learning based methods, a novel anchor-free deep learning framework called CA-CentripetalNet is proposed for hardhat wearing detection. Two novel schemes are proposed to improve the feature extraction and utilization ability of CA-CentripetalNet, which are vertical-horizontal corner pooling and bounding constrained center attention. The former is designed to realize the comprehensive utilization of marginal features and internal features. The latter is designed to enforce the backbone to pay attention to internal features, which is only used during the training rather than during the detection. Experimental results indicate that the CA-CentripetalNet achieves better performance with the 86.63% mAP (mean Average Precision) with less memory consumption at a reasonable speed than the existing deep learning based methods, especially in case of small-scale hardhats and non-worn-hardhats. | ['Han Wang', 'Nili Tian', 'Chengbin Zhang', 'Wensheng Ouyang', 'Nian Cai', 'Zhijian Liu'] | 2023-07-09 | null | null | null | null | ['management'] | ['miscellaneous'] | [-1.00449391e-01 1.09740056e-01 8.66235420e-03 2.33811815e-03
-5.73782384e-01 2.38706544e-01 4.59590964e-02 -8.96904692e-02
-3.49017709e-01 4.94334400e-01 1.83525309e-01 5.39720641e-04
-2.53336370e-01 -8.79194617e-01 -6.73384070e-01 -1.16964197e+00
-1.13514662e-01 -1.40555441e-01 6.62022650e-01 -4.54857051e-01
-1.57928810e-01 4.94924635e-01 -1.40509784e+00 2.17087880e-01
4.09324378e-01 1.44886911e+00 3.06353837e-01 1.87357694e-01
5.86241245e-01 4.06919897e-01 -4.56196606e-01 -1.41044810e-01
4.06065494e-01 4.60548937e-01 -1.35658145e-01 1.86275169e-02
1.70797899e-01 -9.54732895e-01 -4.76241171e-01 8.49135101e-01
9.10877228e-01 -3.82438064e-01 4.09209043e-01 -1.21777010e+00
-5.98264217e-01 3.28201443e-01 -9.66786623e-01 1.65122479e-01
-7.51679316e-02 2.26721197e-01 6.69630766e-01 -9.28921878e-01
3.15551370e-01 9.57391441e-01 1.17465377e+00 1.90485060e-01
-4.81111795e-01 -7.92664051e-01 1.08311996e-01 3.26867014e-01
-1.88850296e+00 -7.17898086e-02 8.88705373e-01 -3.57355535e-01
3.57244939e-01 4.96025123e-02 8.99349749e-01 9.76039886e-01
4.40922678e-01 1.01803744e+00 6.92340314e-01 -2.73403674e-01
-7.00609535e-02 5.65294002e-04 1.65426716e-01 1.09018397e+00
5.16843200e-01 -1.23946436e-01 -3.87342907e-02 2.00978834e-02
1.04568553e+00 2.68289298e-01 -4.18139458e-01 -3.68600339e-01
-1.12066627e+00 9.36970234e-01 7.08825707e-01 1.74149975e-01
-5.75977147e-01 2.90698618e-01 5.30811727e-01 -2.00723872e-01
4.78623629e-01 -1.80102233e-02 -6.57763600e-01 3.30486327e-01
-5.80613434e-01 2.98955232e-01 3.37836653e-01 1.10242665e+00
8.00104439e-01 1.59772262e-02 -4.95841682e-01 8.25754404e-01
3.12744021e-01 6.15068257e-01 4.67080921e-02 -5.14906466e-01
4.78053063e-01 6.85067773e-01 1.42838182e-02 -1.34816396e+00
-5.64022899e-01 -4.98854578e-01 -7.93095171e-01 1.71721548e-01
7.20897615e-02 -5.00059307e-01 -1.19489038e+00 1.34893680e+00
4.61406648e-01 -6.39467910e-02 -4.80542570e-01 8.56507003e-01
7.26914585e-01 6.90831959e-01 3.17128003e-02 -2.68126905e-01
1.42524529e+00 -8.13701749e-01 -9.74856496e-01 -1.08553253e-01
6.56393290e-01 -5.92426717e-01 1.01745331e+00 4.42806065e-01
-6.24939620e-01 -5.33427477e-01 -1.53593636e+00 -5.98474964e-02
-5.48151851e-01 4.30260003e-01 5.17887115e-01 6.32513702e-01
-6.22985363e-01 9.59011316e-02 -6.47787511e-01 -2.49340877e-01
1.02392161e+00 3.95332307e-01 -3.49030077e-01 -2.27832317e-01
-1.25184381e+00 6.52604878e-01 3.54516685e-01 6.42871678e-01
-1.15631664e+00 -4.38398123e-01 -7.36511290e-01 2.08375275e-01
6.94547892e-01 -1.82343870e-01 1.05123544e+00 -4.07565624e-01
-1.17107642e+00 4.12222236e-01 5.96916020e-01 -8.18116963e-02
6.41864717e-01 -5.74431121e-01 -2.21396089e-01 -7.76585238e-03
1.43556431e-01 3.77318144e-01 8.17644835e-01 -1.16468787e+00
-7.40347266e-01 -4.09235060e-01 -1.23430928e-02 8.68638419e-03
-9.50191379e-01 -1.78825080e-01 -7.39569843e-01 -7.50303984e-01
1.84200779e-01 -9.25505817e-01 -1.41819447e-01 5.17143786e-01
-5.06980777e-01 -4.99039084e-01 1.33906758e+00 -8.79462659e-01
1.46860313e+00 -1.97859526e+00 -2.76465684e-01 1.75233439e-01
1.14886902e-01 5.54842532e-01 2.03636885e-01 3.38886708e-01
5.62216751e-02 -2.99895108e-01 -8.13983604e-02 -2.67152973e-02
-2.34242216e-01 2.40210623e-01 4.44072127e-01 8.17562699e-01
9.16225910e-02 4.71499801e-01 -6.05102122e-01 -8.03790212e-01
3.83095771e-01 5.63276827e-01 -3.98721397e-01 4.66410490e-03
1.96091920e-01 -2.18950763e-01 -6.42486393e-01 9.75584686e-01
1.03698206e+00 5.24035916e-02 -1.97369397e-01 -4.55095291e-01
-7.18986839e-02 -5.02436280e-01 -1.25204825e+00 1.60269392e+00
-1.94340140e-01 5.04423559e-01 2.06895277e-01 -9.02498662e-01
8.78554165e-01 2.82971829e-01 5.20514965e-01 -5.57532012e-01
5.58979154e-01 1.51366293e-02 -4.21215832e-01 -1.10796010e+00
1.01299636e-01 -2.29263213e-02 -1.76192760e-01 -3.55094075e-01
-2.22227827e-01 3.94612461e-01 -2.36769989e-01 -9.99584198e-02
1.09417856e+00 3.82941626e-02 1.82937354e-01 -5.65912008e-01
4.62700754e-01 -2.86951095e-01 8.58147919e-01 5.11556625e-01
-2.43882522e-01 6.23578787e-01 3.01675439e-01 -8.67129445e-01
-8.71712387e-01 -8.54259074e-01 -1.60531819e-01 9.02872026e-01
2.84049451e-01 -1.53323516e-01 -9.27902877e-01 -9.96080041e-01
1.20728798e-01 -6.13406524e-02 -8.71518314e-01 -3.69832158e-01
-7.54230201e-01 -7.01569736e-01 3.87553632e-01 9.38593626e-01
9.22414541e-01 -7.87890375e-01 -7.11720645e-01 2.91633576e-01
-1.13942130e-02 -1.02579474e+00 -1.95178241e-01 1.56807378e-01
-4.63797480e-01 -1.12376821e+00 -1.03300905e+00 -1.01049697e+00
5.93864620e-01 3.68874043e-01 3.76503110e-01 3.55178803e-01
-5.62884867e-01 -2.40883619e-01 -4.14397895e-01 -9.57473278e-01
4.35791254e-01 1.83252394e-01 -6.19557463e-02 1.29825115e-01
3.73813152e-01 -3.05777937e-01 -8.79467726e-01 3.90148133e-01
-6.81486368e-01 4.48243367e-03 1.04042101e+00 8.58794987e-01
2.99795747e-01 1.40032604e-01 4.94455576e-01 -6.55101299e-01
3.28674704e-01 -6.65098846e-01 -4.90339220e-01 2.23623440e-01
-4.90162939e-01 -4.51963603e-01 2.67425418e-01 -3.12701523e-01
-8.76726747e-01 1.44132271e-01 -2.75519937e-01 -4.21932429e-01
-4.51672785e-02 4.54013020e-01 -4.47167933e-01 -2.85413325e-01
5.07218778e-01 -2.46056244e-02 -2.13727474e-01 -6.64331138e-01
-2.50518978e-01 9.97247994e-01 4.56082582e-01 -4.19506997e-01
8.51446569e-01 5.03307045e-01 -1.76570088e-01 -9.43757474e-01
-9.54197764e-01 -3.61887574e-01 -5.28457224e-01 -5.69185376e-01
1.30510533e+00 -1.08366036e+00 -7.13140607e-01 7.77091324e-01
-1.23690951e+00 2.35151853e-02 1.47591323e-01 3.91002446e-01
-1.08331293e-01 3.65035772e-01 -7.28063583e-01 -8.69024396e-01
-6.36723995e-01 -8.82952929e-01 1.42813706e+00 3.31108987e-01
5.69434643e-01 -5.11191368e-01 -2.23787919e-01 2.98658341e-01
1.24056377e-01 7.14561880e-01 7.32307971e-01 -2.17933387e-01
-5.15499115e-01 -7.35073090e-01 -4.44145769e-01 7.24412560e-01
2.26109046e-02 -2.11594835e-01 -1.03830600e+00 -4.04892415e-01
-1.48433834e-01 -9.93702859e-02 7.07272291e-01 4.65004236e-01
1.52069747e+00 -2.80829102e-01 -5.01928151e-01 5.90673983e-01
1.46122348e+00 2.16176569e-01 9.30138350e-01 6.01374984e-01
8.95030737e-01 3.96495432e-01 9.02247906e-01 6.45239413e-01
2.63200819e-01 5.60639203e-01 8.20970058e-01 -7.72739589e-01
5.76149374e-02 -3.42391990e-02 8.19165632e-02 6.00366592e-01
-3.01512778e-01 -1.03314824e-01 -8.83677423e-01 8.64143729e-01
-1.92563105e+00 -6.07150912e-01 -2.36229137e-01 1.70060766e+00
4.13895607e-01 3.68910581e-01 -3.03959828e-02 4.29808378e-01
1.04636145e+00 1.95328847e-01 -1.74314335e-01 8.90684649e-02
8.47319737e-02 -3.88290100e-02 7.24363685e-01 -2.02174276e-01
-1.57027912e+00 5.48423290e-01 5.67270184e+00 1.06274235e+00
-8.11751366e-01 4.03517485e-01 3.19543958e-01 1.73689812e-01
5.04156113e-01 -4.62087095e-01 -8.13335836e-01 6.02794409e-01
1.02414966e-01 5.80297709e-01 -3.92596543e-01 1.35034811e+00
9.36856270e-02 5.14592528e-02 -7.27179527e-01 9.49510872e-01
7.73248821e-02 -1.28232622e+00 -2.29180172e-01 1.65993094e-01
5.50900519e-01 -7.66935199e-02 -1.53072432e-01 2.92016685e-01
-2.24293485e-01 -3.33747566e-01 7.19128788e-01 3.25677246e-01
7.00315714e-01 -9.47543442e-01 1.44168818e+00 2.96159953e-01
-1.34206343e+00 -4.78794843e-01 -5.19394517e-01 -1.70758635e-01
1.69680044e-01 5.79066873e-01 -5.86210907e-01 6.58144176e-01
1.12452221e+00 5.18861353e-01 -3.72153699e-01 1.24928212e+00
-2.34492034e-01 3.63643616e-01 -3.18942249e-01 -6.72229053e-03
4.72129554e-01 3.29772949e-01 2.43225053e-01 1.33316946e+00
3.75280887e-01 -1.57858524e-02 5.86188316e-01 7.05491081e-02
9.84562635e-02 4.69641834e-02 -5.37438929e-01 5.68371534e-01
3.82278591e-01 1.37233555e+00 -5.44165909e-01 -2.47681916e-01
-5.21393776e-01 6.23842359e-01 1.04052775e-01 9.00922939e-02
-1.25349164e+00 -8.74525547e-01 4.19881016e-01 6.41966403e-01
6.14784598e-01 -2.32315212e-01 -1.18792586e-01 -5.82989752e-01
2.02415273e-01 -4.32530135e-01 5.21117687e-01 -7.62642682e-01
-1.12367487e+00 3.09624732e-01 2.38733828e-01 -1.25303507e+00
8.52962613e-01 -9.15278077e-01 -6.60263956e-01 2.64237314e-01
-1.47041333e+00 -1.78655410e+00 -9.64333475e-01 7.40123153e-01
7.21616030e-01 7.67979473e-02 4.26517665e-01 7.83156812e-01
-1.04407799e+00 7.59462714e-01 -9.65988860e-02 4.14580107e-01
4.83614117e-01 -6.23952746e-01 -2.47130752e-01 8.83145511e-01
-7.49510825e-01 3.34350675e-01 4.63800430e-01 -6.86356187e-01
-1.23039639e+00 -1.17331147e+00 3.17067266e-01 -1.00279655e-02
3.94326717e-01 -5.48543990e-01 -6.71436906e-01 5.99883437e-01
-2.32075900e-02 3.49709034e-01 2.58497536e-01 -1.78157464e-01
4.48697014e-03 -6.08140945e-01 -1.07038248e+00 2.58652121e-01
8.77259195e-01 -1.24676861e-01 -2.95505732e-01 6.81652069e-01
7.06023574e-01 -3.23794276e-01 -9.94358480e-01 7.10369766e-01
7.64742076e-01 -5.78504860e-01 9.91676211e-01 -1.06634244e-01
2.22871795e-01 -5.35291255e-01 -2.21252516e-01 -5.63140392e-01
-7.74443805e-01 -4.29357737e-01 -4.99075167e-02 1.09299636e+00
-2.36104131e-02 -2.73737043e-01 8.84718418e-01 1.42564893e-01
-7.20184326e-01 -1.03862178e+00 -9.20175433e-01 -8.14930677e-01
-3.86907727e-01 -1.44314364e-01 5.71752369e-01 8.48245382e-01
-4.09257174e-01 1.39192089e-01 -8.42919111e-01 6.07097149e-01
5.47840953e-01 -5.26610076e-01 6.94726825e-01 -1.40786040e+00
1.83876753e-01 2.48067737e-01 -7.75750697e-01 -9.55070019e-01
-4.56827313e-01 -2.14942291e-01 4.47697759e-01 -1.56341648e+00
1.98448017e-01 -4.03572828e-01 -5.87476075e-01 8.77983809e-01
-7.56558925e-02 4.44946289e-01 -1.93558946e-01 -9.41687152e-02
-5.89065611e-01 6.48977637e-01 1.32973683e+00 -2.13965461e-01
2.14186639e-01 -1.79925859e-01 -4.12058860e-01 9.63172495e-01
6.53213561e-01 -5.08776009e-01 -2.27953076e-01 -7.58593798e-01
-8.18958953e-02 -1.70411885e-01 5.28826416e-01 -1.49524391e+00
1.45442739e-01 -1.46833912e-01 7.08722651e-01 -7.85448670e-01
3.49197388e-01 -1.05421829e+00 -1.62374869e-01 8.48937154e-01
2.06774503e-01 -6.74335808e-02 5.49370497e-02 7.89277077e-01
1.96343198e-01 1.95713975e-02 9.29384410e-01 -5.00207916e-02
-8.88869882e-01 4.73459542e-01 -4.10402507e-01 -5.18736541e-01
1.52060175e+00 -3.02125841e-01 -4.88200068e-01 -8.30952972e-02
-5.35573363e-01 2.55750924e-01 -6.36410192e-02 3.97008926e-01
7.87429452e-01 -1.56791270e+00 -5.79148591e-01 1.36568397e-01
3.89833093e-01 1.36224046e-01 6.06957793e-01 9.30249751e-01
-8.42402577e-01 1.15303151e-01 -3.98261756e-01 -6.46658301e-01
-1.18442953e+00 6.94125473e-01 1.37287915e-01 6.59342250e-03
-9.02032852e-01 9.49787319e-01 3.67353231e-01 6.69115931e-02
5.06968260e-01 -2.12808847e-01 -2.33043075e-01 -1.62341679e-03
3.12768430e-01 5.70372999e-01 2.44796827e-01 -3.94527078e-01
-4.88401592e-01 8.46035063e-01 -2.21783713e-01 5.87787449e-01
1.54758549e+00 3.19892466e-02 7.88858160e-02 -9.73348096e-02
1.31216502e+00 -3.31782371e-01 -1.45022058e+00 -9.03685912e-02
-2.27839723e-01 -6.01035535e-01 5.24386525e-01 -4.49144036e-01
-1.41008759e+00 8.11677516e-01 1.20704985e+00 1.86178803e-01
1.11003208e+00 -2.26743460e-01 1.36636078e+00 4.28256184e-01
4.60426062e-01 -1.51813436e+00 3.31418902e-01 1.16686799e-01
9.84897554e-01 -1.15874541e+00 3.34429204e-01 -4.99480933e-01
-2.96935976e-01 1.04511487e+00 1.02502716e+00 -3.17908287e-01
9.25006151e-01 2.99493372e-01 -1.25372633e-01 -5.89511037e-01
-1.39929771e-01 -1.22589514e-01 5.44233853e-03 6.20794833e-01
-2.43351609e-01 -2.46782750e-01 -5.01321435e-01 7.10248947e-01
4.37927097e-01 -1.07273810e-01 1.90775827e-01 1.39963770e+00
-8.20733070e-01 -3.71212214e-01 -2.99850702e-01 4.01806206e-01
-4.38604653e-01 2.84575641e-01 1.04441166e-01 1.17328262e+00
7.97408700e-01 7.59161711e-01 -1.91113770e-01 -7.47605205e-01
6.77454352e-01 -4.86646742e-01 1.92427486e-02 -4.40191865e-01
-3.83116037e-01 3.27364117e-01 1.54020265e-01 -5.79318881e-01
-2.30487242e-01 -4.06690627e-01 -8.80491614e-01 -1.67682126e-01
-1.02426457e+00 -1.69344202e-01 8.06739450e-01 8.54777396e-01
3.16353887e-02 7.00412333e-01 6.62511230e-01 -9.37145233e-01
-4.53443110e-01 -1.05405593e+00 -5.86862028e-01 9.09132212e-02
4.61166203e-01 -1.10883892e+00 -2.35788245e-02 -2.61045881e-02] | [8.796307563781738, -0.5515021681785583] |
ef0018d8-5882-42b2-aa03-99af38c34050 | cross-individual-recognition-of-emotions-by-a | 2009.12525 | null | https://arxiv.org/abs/2009.12525v2 | https://arxiv.org/pdf/2009.12525v2.pdf | Cross-individual Recognition of Emotions by a Dynamic Entropy based on Pattern Learning with EEG features | Use of the electroencephalogram (EEG) and machine learning approaches to recognize emotions can facilitate affective human computer interactions. However, the type of EEG data constitutes an obstacle for cross-individual EEG feature modelling and classification. To address this issue, we propose a deep-learning framework denoted as a dynamic entropy-based pattern learning (DEPL) to abstract informative indicators pertaining to the neurophysiological features among multiple individuals. DEPL enhanced the capability of representations generated by a deep convolutional neural network by modelling the interdependencies between the cortical locations of dynamical entropy based features. The effectiveness of the DEPL has been validated with two public databases, commonly referred to as the DEAP and MAHNOB-HCI multimodal tagging databases. Specifically, the leave one subject out training and testing paradigm has been applied. Numerous experiments on EEG emotion recognition demonstrate that the proposed DEPL is superior to those traditional machine learning (ML) methods, and could learn between electrode dependencies w.r.t. different emotions, which is meaningful for developing the effective human-computer interaction systems by adapting to human emotions in the real world applications. | ['Zhong Yin', 'Xiaolong Zhong'] | 2020-09-26 | null | null | null | null | ['eeg-emotion-recognition'] | ['miscellaneous'] | [-1.86475351e-01 -2.60620415e-01 4.00978029e-01 -5.94874382e-01
-1.05261400e-01 -1.45985886e-01 4.53883380e-01 1.74078062e-01
-3.98271680e-01 9.30373609e-01 -3.71010927e-03 3.93083662e-01
-5.13650775e-01 -4.85214233e-01 -4.38929319e-01 -8.31477940e-01
-6.27070367e-01 3.27096768e-02 -3.25657547e-01 -2.41422772e-01
2.36537516e-01 4.84426200e-01 -1.61782694e+00 5.91092646e-01
7.74189413e-01 1.53920698e+00 -1.65686324e-01 2.26040408e-01
8.68796259e-02 5.51704586e-01 -6.28256440e-01 -2.63532281e-01
-1.74189061e-01 -5.17662823e-01 -4.29845065e-01 -4.48303699e-01
-6.47356331e-01 4.00792927e-01 -1.51154190e-01 9.32452083e-01
8.96965504e-01 1.14019841e-01 9.68664587e-01 -1.63307977e+00
-5.95748305e-01 3.50846648e-01 -3.65760118e-01 4.47665572e-01
5.86024940e-01 -7.70423040e-02 5.36398411e-01 -8.13655257e-01
1.10439956e-01 8.33018541e-01 8.15808594e-01 4.33196634e-01
-8.26159656e-01 -9.74707365e-01 -1.65036485e-01 6.12277150e-01
-1.55485916e+00 -1.54991388e-01 1.10210705e+00 -6.11621320e-01
1.10315180e+00 2.96913981e-01 1.18488169e+00 1.49292207e+00
9.32560980e-01 5.71159303e-01 1.53432870e+00 -1.37896538e-01
4.18864965e-01 4.61113930e-01 3.89441013e-01 3.85782659e-01
-1.10398538e-01 6.91581890e-02 -1.00521874e+00 -3.71427625e-01
3.80989224e-01 -1.94930762e-01 -2.26504311e-01 1.14236936e-01
-8.99570763e-01 4.49610621e-01 3.49867940e-01 6.88786149e-01
-7.24244237e-01 -1.45021424e-01 8.28072011e-01 2.53904283e-01
5.79501569e-01 6.17536783e-01 -7.48107374e-01 -4.01213855e-01
-5.64685285e-01 -1.49775982e-01 7.39946187e-01 3.85753244e-01
5.69493294e-01 1.44028468e-02 6.52722940e-02 7.79822946e-01
1.87784597e-01 3.13923508e-01 9.13277268e-01 -2.33885348e-01
2.01198719e-02 7.94207692e-01 -3.41157824e-01 -1.29217887e+00
-9.47840035e-01 -1.47423655e-01 -1.25586820e+00 -1.07454114e-01
-3.53438705e-01 -4.59107339e-01 -3.14743400e-01 1.64052784e+00
-7.02032670e-02 2.97755539e-01 1.38107345e-01 6.04490936e-01
8.36973250e-01 7.08802938e-01 3.11229765e-01 -4.52214986e-01
1.25062573e+00 -1.52017325e-01 -1.05817306e+00 -7.28765801e-02
3.51338714e-01 -1.20962687e-01 7.72119880e-01 6.40042007e-01
-6.51554942e-01 -6.12575769e-01 -8.84224474e-01 6.16403162e-01
-8.45971584e-01 6.36081547e-02 8.19197357e-01 6.07261837e-01
-9.21030283e-01 5.14003098e-01 -7.81581461e-01 -2.18129680e-01
5.91849685e-01 8.46071541e-01 -5.84310651e-01 7.24826396e-01
-1.74282897e+00 1.01888084e+00 6.33443534e-01 4.32008266e-01
-2.41455346e-01 -6.27759695e-01 -6.66046739e-01 2.86051840e-01
-5.81652582e-01 -3.33947301e-01 4.76959169e-01 -1.48468816e+00
-1.68831027e+00 5.01277208e-01 1.86706096e-01 -2.13367715e-01
-5.70315458e-02 2.56247334e-02 -8.09300005e-01 -1.21557321e-02
-3.71712416e-01 5.69779694e-01 5.92130363e-01 -9.02807832e-01
-1.28026143e-01 -5.60871065e-01 -6.13395870e-01 1.78927347e-01
-7.90527523e-01 1.85278162e-01 1.02065966e-01 -4.13007796e-01
-3.53820324e-01 -8.09637964e-01 2.36475080e-01 -5.60137868e-01
-1.94262132e-01 -4.95177954e-01 6.04784727e-01 -7.13796675e-01
1.27717602e+00 -2.16451097e+00 2.84318894e-01 6.26850486e-01
-2.61748172e-02 6.42694682e-02 6.44683614e-02 3.49541903e-01
-4.62803274e-01 -6.01303950e-02 -1.66407347e-01 1.63846821e-01
9.82551053e-02 -4.15563583e-02 8.34683254e-02 4.00489867e-01
3.01345617e-01 9.45580781e-01 -5.42022109e-01 -3.52505147e-01
2.16352105e-01 7.36918092e-01 -3.35397184e-01 3.27305108e-01
4.08948720e-01 6.80418551e-01 -4.62686270e-01 5.08250654e-01
4.09075141e-01 8.79953727e-02 4.71569039e-02 -3.84821773e-01
4.49409429e-03 -1.25324622e-01 -8.46840858e-01 1.45690763e+00
-3.83313775e-01 8.61413121e-01 -4.45207000e-01 -1.16088092e+00
1.17277133e+00 6.79319501e-01 9.78436828e-01 -8.98449659e-01
5.07689238e-01 6.00163862e-02 2.22034827e-01 -8.93937051e-01
-1.44812196e-01 -5.21217957e-02 -3.36294770e-01 1.51118666e-01
3.78769904e-01 2.33848006e-01 -4.10956949e-01 -3.96852225e-01
9.60662723e-01 -1.11702435e-01 5.95609486e-01 -6.09817863e-01
5.97659111e-01 -6.69111371e-01 6.78949416e-01 3.07816595e-01
-2.63304055e-01 2.79258583e-02 5.66931307e-01 -6.10856056e-01
-4.96754497e-01 -6.67430997e-01 -6.91051245e-01 6.77321196e-01
1.82727635e-01 -2.21496120e-01 -7.82458186e-01 -3.93535763e-01
-2.30184466e-01 5.38305879e-01 -9.74657416e-01 -7.57238388e-01
1.27232283e-01 -1.43612278e+00 6.14625812e-01 5.95373094e-01
7.00390935e-01 -1.57907462e+00 -7.88277030e-01 2.45712653e-01
-8.48849341e-02 -8.81071627e-01 2.39702404e-01 5.57918489e-01
-4.54220623e-01 -8.59639704e-01 -2.52410799e-01 -5.25337398e-01
2.86740839e-01 -8.50398719e-01 8.41570616e-01 -2.62467384e-01
-3.59922200e-01 5.82254291e-01 -4.30389613e-01 -6.84619665e-01
4.56105843e-02 1.33615136e-01 4.89471823e-01 4.39895123e-01
8.18702340e-01 -9.90640700e-01 -5.94900370e-01 1.31763324e-01
-8.28767776e-01 -1.97344810e-01 6.36324346e-01 9.46423113e-01
3.49515080e-01 3.86536330e-01 1.12875068e+00 -3.58886868e-01
1.19871783e+00 -8.23906898e-01 -6.92411810e-02 4.26288307e-01
-5.49203575e-01 -8.07359219e-02 6.21028304e-01 -5.98104119e-01
-1.02002072e+00 6.77130967e-02 -2.87040174e-01 -1.54925540e-01
-4.87063259e-01 6.27688050e-01 -4.33818847e-01 -2.26587608e-01
4.20215011e-01 3.31261486e-01 -4.56008077e-01 -1.00106718e-02
-3.31412137e-01 1.10678852e+00 2.31583118e-01 -3.94920975e-01
-5.45409918e-02 -1.79875478e-01 -6.61284924e-02 -7.24983156e-01
-1.86700910e-01 -1.46665826e-01 -7.96066880e-01 -5.20209193e-01
1.09396470e+00 -7.18036950e-01 -1.07381237e+00 5.98628163e-01
-1.15804064e+00 -1.13356248e-01 6.21242598e-02 7.16705024e-01
-6.24931395e-01 -5.12945578e-02 -5.64384282e-01 -9.99252737e-01
-6.63797081e-01 -1.04627192e+00 7.97692239e-01 4.53280121e-01
-5.23637176e-01 -1.10264528e+00 4.16427761e-01 -2.56039977e-01
3.24490428e-01 4.41353112e-01 9.99926984e-01 -9.75215971e-01
3.50162268e-01 -3.80610466e-01 2.31974006e-01 2.89436817e-01
7.25298524e-02 -1.00264333e-01 -1.30288458e+00 6.29075989e-02
2.16516599e-01 -4.53350991e-01 2.87932783e-01 2.25183740e-01
1.55418146e+00 -9.41792056e-02 -1.59816742e-01 5.14639080e-01
1.14893174e+00 8.07221830e-01 9.10064697e-01 2.09688991e-01
3.30676764e-01 5.04783034e-01 8.37148950e-02 8.27146471e-01
3.21211398e-01 5.45029640e-01 9.85461622e-02 -6.17635101e-02
7.27096856e-01 2.47285619e-01 4.18240547e-01 1.22470343e+00
-2.76047409e-01 -5.86579740e-02 -1.01998436e+00 1.38502002e-01
-1.87596607e+00 -1.01074827e+00 1.25073537e-01 1.76139724e+00
6.79795682e-01 -2.24271774e-01 -1.24981351e-01 2.76057780e-01
4.33614433e-01 -2.28147104e-01 -6.53642535e-01 -7.16189265e-01
-1.75951675e-01 4.59784359e-01 -2.15332583e-01 -1.28175825e-01
-9.57392931e-01 5.05844176e-01 5.57197142e+00 7.09317446e-01
-1.28454947e+00 1.87952176e-01 8.28930140e-01 1.01301245e-01
5.63124865e-02 -7.16145635e-01 -3.34705859e-01 7.95857906e-01
1.31917155e+00 -2.14140609e-01 5.55815518e-01 5.39158285e-01
2.09456861e-01 -9.78226215e-02 -1.03433979e+00 1.54460466e+00
2.18579367e-01 -7.16104269e-01 -3.15795362e-01 -1.31599069e-01
4.26802635e-01 -1.78024560e-01 6.92250207e-02 4.93314445e-01
-4.41574842e-01 -1.21241391e+00 2.07014307e-01 1.26263750e+00
5.64036548e-01 -1.01068449e+00 1.17261088e+00 3.20781499e-01
-1.12899125e+00 -3.22825193e-01 -2.39481285e-01 -7.17492178e-02
-2.05839619e-01 4.93881047e-01 -3.61249626e-01 6.95767641e-01
1.11681104e+00 9.02191639e-01 -6.94586813e-01 8.87472630e-01
1.68481126e-01 4.57493126e-01 -1.77220419e-01 -3.97390664e-01
-2.56942540e-01 -2.53152907e-01 1.31173521e-01 1.30863690e+00
5.57164729e-01 3.22977990e-01 -3.33438784e-01 8.56808722e-01
1.08910173e-01 3.97462964e-01 -7.54629076e-01 -1.85940951e-01
2.88133740e-01 1.44771838e+00 -6.62594497e-01 -1.69909224e-02
-2.12256253e-01 1.16033876e+00 2.54656255e-01 2.79546320e-01
-1.01867354e+00 -6.02947295e-01 4.51766074e-01 -5.10778964e-01
-4.60644752e-01 2.64590800e-01 -3.38665038e-01 -1.21877229e+00
-7.77985156e-02 -7.71340311e-01 1.92157120e-01 -1.07697248e+00
-1.68615174e+00 1.19619775e+00 7.85665289e-02 -1.05280507e+00
-1.28828019e-01 -7.00532079e-01 -8.98103118e-01 8.03724408e-01
-1.04972935e+00 -7.25293517e-01 -2.92725116e-01 9.45844829e-01
2.04509705e-01 -3.40536386e-01 1.25045431e+00 4.60287750e-01
-8.44851851e-01 5.33932924e-01 -1.22805722e-01 -7.24369148e-03
5.94368339e-01 -9.84890759e-01 -4.84909892e-01 2.84770042e-01
3.89807075e-02 3.78737450e-01 3.45478863e-01 -3.00142854e-01
-1.18294048e+00 -7.57295191e-01 6.75300419e-01 -3.22514027e-01
5.78968108e-01 -5.27035832e-01 -9.79104698e-01 4.27475721e-01
6.57076299e-01 -1.07286245e-01 1.27528858e+00 1.19025670e-01
2.83543557e-01 -4.17566121e-01 -1.15145910e+00 2.89946645e-01
6.31646931e-01 -6.70085788e-01 -7.57391572e-01 1.91820815e-01
-5.57276309e-02 3.82740013e-02 -1.28622866e+00 6.50244951e-01
7.57539213e-01 -9.92432594e-01 6.66170418e-01 -6.31214321e-01
2.64619052e-01 2.25148827e-01 -9.45688561e-02 -1.78047466e+00
-3.98072332e-01 -3.07656229e-01 -1.03880607e-01 1.27883112e+00
2.68916160e-01 -7.56893098e-01 1.24211334e-01 1.07130134e+00
7.07422793e-02 -1.13751614e+00 -1.09293795e+00 -3.17112803e-01
-5.54896481e-02 -5.20128608e-01 7.82397270e-01 1.00878870e+00
8.23823690e-01 2.70207524e-01 -3.03090751e-01 1.23288624e-01
-5.50955394e-03 -2.88760841e-01 1.26460657e-01 -1.51663685e+00
-1.07264435e-02 -4.89297658e-01 -1.00646830e+00 4.13969234e-02
7.36977935e-01 -9.92504358e-01 -2.96894759e-02 -1.04823649e+00
3.28156173e-01 -2.47820705e-01 -9.88233268e-01 4.93410259e-01
-1.23764165e-01 -3.79612520e-02 -2.30066642e-01 -1.42932460e-01
-5.57711363e-01 1.12156296e+00 7.22581327e-01 6.04435988e-03
-3.89490604e-01 -5.62754609e-02 -4.80279416e-01 7.41883278e-01
1.19865632e+00 -3.94373715e-01 -4.64125037e-01 2.04869181e-01
4.27083820e-01 1.62682265e-01 3.04243624e-01 -1.36888564e+00
2.48688638e-01 1.56214193e-01 9.66274500e-01 -2.67457753e-01
4.09286976e-01 -1.03065562e+00 2.97444910e-01 2.27321491e-01
-3.81743968e-01 3.86844993e-01 5.04122138e-01 4.60011154e-01
-2.84291893e-01 1.64564222e-01 5.66866875e-01 1.51704744e-01
-5.33095717e-01 4.79116499e-01 -7.02920437e-01 -1.53261542e-01
1.11669314e+00 -7.52513632e-02 1.41242156e-02 -2.39778504e-01
-8.39322329e-01 -4.77227978e-02 -3.40269446e-01 4.05668914e-01
7.88380027e-01 -1.52702796e+00 -1.94747493e-01 5.63178718e-01
1.11479573e-01 -8.51970494e-01 4.23170924e-01 1.13001359e+00
3.44560631e-02 3.61040682e-01 -8.64647806e-01 -4.10197288e-01
-9.95796800e-01 3.56132776e-01 7.55207360e-01 -1.75482601e-01
-3.32642555e-01 6.03304565e-01 7.79925883e-02 -4.89171624e-01
1.84799969e-01 -2.68684104e-02 -7.36934602e-01 2.56888747e-01
3.33299696e-01 1.69732451e-01 1.97301120e-01 -6.30587459e-01
-6.14759147e-01 3.13329905e-01 3.81218582e-01 -7.95312971e-02
1.58758569e+00 5.51071055e-02 -4.00784791e-01 9.37765539e-01
1.33580923e+00 -4.13168460e-01 -9.21998024e-01 2.77892113e-01
1.16079852e-01 1.33083835e-01 -5.52200526e-02 -1.04766679e+00
-1.09578335e+00 1.12919450e+00 1.13159871e+00 2.04604477e-01
1.51037002e+00 -3.39612573e-01 3.95049512e-01 6.08214021e-01
4.74796742e-01 -1.42214942e+00 5.48414961e-02 2.84280896e-01
9.97268140e-01 -1.13516569e+00 -2.36675620e-01 2.16858253e-01
-1.00249910e+00 1.45471525e+00 7.77189553e-01 1.75629500e-02
1.20662427e+00 4.39843118e-01 -2.16930598e-01 -4.03313607e-01
-7.83842981e-01 3.40476155e-01 6.14079356e-01 4.30867761e-01
4.41146165e-01 1.59459963e-01 -3.51494312e-01 1.51168406e+00
-1.62039846e-01 1.88046560e-01 -5.75058497e-02 5.33670247e-01
3.55479009e-02 -8.42739820e-01 -1.32138997e-01 6.40786469e-01
-3.81639987e-01 -8.87720361e-02 -6.31291270e-01 7.61026263e-01
4.59741443e-01 7.12870300e-01 8.84705558e-02 -9.27695751e-01
2.40548417e-01 6.86650276e-01 3.97724569e-01 -1.49325311e-01
-9.92948115e-01 -2.95823634e-01 -2.36355379e-01 -4.47611064e-01
-6.30595148e-01 -5.26328266e-01 -1.27934706e+00 1.83305308e-01
-2.24762931e-01 3.03355098e-01 5.03490925e-01 1.19819653e+00
7.24594235e-01 7.29754746e-01 8.78729045e-01 -9.42582011e-01
8.11907649e-02 -1.17424357e+00 -8.91105294e-01 4.22229737e-01
-6.58000857e-02 -8.73478949e-01 -2.88132936e-01 1.09607894e-02] | [13.167445182800293, 3.4370369911193848] |
401652a4-5614-4469-84e3-848ed5474afd | affinity-attention-graph-neural-network-for | 2106.04054 | null | https://arxiv.org/abs/2106.04054v1 | https://arxiv.org/pdf/2106.04054v1.pdf | Affinity Attention Graph Neural Network for Weakly Supervised Semantic Segmentation | Weakly supervised semantic segmentation is receiving great attention due to its low human annotation cost. In this paper, we aim to tackle bounding box supervised semantic segmentation, i.e., training accurate semantic segmentation models using bounding box annotations as supervision. To this end, we propose Affinity Attention Graph Neural Network ($A^2$GNN). Following previous practices, we first generate pseudo semantic-aware seeds, which are then formed into semantic graphs based on our newly proposed affinity Convolutional Neural Network (CNN). Then the built graphs are input to our $A^2$GNN, in which an affinity attention layer is designed to acquire the short- and long- distance information from soft graph edges to accurately propagate semantic labels from the confident seeds to the unlabeled pixels. However, to guarantee the precision of the seeds, we only adopt a limited number of confident pixel seed labels for $A^2$GNN, which may lead to insufficient supervision for training. To alleviate this issue, we further introduce a new loss function and a consistency-checking mechanism to leverage the bounding box constraint, so that more reliable guidance can be included for the model optimization. Experiments show that our approach achieves new state-of-the-art performances on Pascal VOC 2012 datasets (val: 76.5\%, test: 75.2\%). More importantly, our approach can be readily applied to bounding box supervised instance segmentation task or other weakly supervised semantic segmentation tasks, with state-of-the-art or comparable performance among almot all weakly supervised tasks on PASCAL VOC or COCO dataset. Our source code will be available at https://github.com/zbf1991/A2GNN. | ['Yao Zhao', 'Yunchao Wei', 'Jianbo Jiao', 'Jimin Xiao', 'Bingfeng Zhang'] | 2021-06-08 | null | null | null | null | ['box-supervised-instance-segmentation'] | ['computer-vision'] | [ 3.40102464e-01 5.07777870e-01 -3.60639274e-01 -6.12594485e-01
-8.22820723e-01 -5.00898600e-01 1.71353802e-01 5.35203293e-02
-4.71654266e-01 6.46229506e-01 -4.22910511e-01 -2.00894237e-01
8.50530863e-02 -8.27767491e-01 -1.02746856e+00 -7.18717575e-01
2.34965384e-01 4.64035034e-01 5.54965377e-01 -5.08716479e-02
4.67458740e-02 3.85703407e-02 -1.28602493e+00 -1.64783634e-02
1.18062556e+00 1.27206767e+00 4.10063446e-01 2.68196702e-01
-2.55144596e-01 5.01316786e-01 -4.62261885e-01 -4.52192247e-01
1.65597245e-01 -3.25401604e-01 -9.87501919e-01 8.35578889e-03
2.93666482e-01 -1.68772846e-01 4.66186590e-02 1.44946170e+00
1.91277951e-01 2.09264368e-01 3.76111060e-01 -1.12036669e+00
-4.87883747e-01 7.04627573e-01 -7.41419613e-01 -8.89204219e-02
-8.35787319e-03 1.52598873e-01 1.14361751e+00 -7.06647456e-01
4.44195151e-01 9.49308991e-01 5.54149628e-01 6.99794590e-01
-9.49584603e-01 -7.77873516e-01 5.27792454e-01 1.21528599e-02
-1.28783023e+00 -1.25739872e-01 8.36688697e-01 -2.88866013e-01
5.64667225e-01 5.00781722e-02 5.36679924e-01 8.40899467e-01
-3.86612773e-01 1.01134419e+00 9.03286934e-01 -1.79301232e-01
2.12389261e-01 -2.86406241e-02 3.55499834e-01 9.07707155e-01
1.62680492e-01 -2.15655655e-01 -2.36717656e-01 2.81072617e-01
6.90089047e-01 -8.46640095e-02 -2.97515422e-01 -2.15162501e-01
-8.12013924e-01 8.01533401e-01 9.22122896e-01 2.26121321e-01
-9.82075855e-02 2.72749931e-01 3.84562343e-01 -2.17222065e-01
6.10662520e-01 3.25621903e-01 -6.08350933e-01 -1.99971423e-02
-8.90637398e-01 -4.79516981e-04 4.37433988e-01 1.06451595e+00
9.98511553e-01 -1.31075203e-01 -4.83380556e-02 9.19193268e-01
4.67908323e-01 2.53783166e-01 1.89512968e-01 -8.15223396e-01
5.80912113e-01 8.22410345e-01 -1.15239009e-01 -7.53361404e-01
-3.72676522e-01 -5.60106695e-01 -5.97061098e-01 -4.38457429e-02
5.35453916e-01 2.17087250e-02 -1.36623800e+00 1.67095709e+00
4.10333961e-01 2.66653210e-01 -6.53623715e-02 1.24294770e+00
8.40004206e-01 5.45251131e-01 2.11581066e-01 1.62530303e-01
1.36117303e+00 -1.37004077e+00 -4.79334682e-01 -5.67699671e-01
8.14175487e-01 -5.26469290e-01 1.42627776e+00 2.01406658e-01
-9.14197147e-01 -6.07198238e-01 -9.85071838e-01 -2.56213635e-01
-3.26323807e-01 3.64295036e-01 7.09020913e-01 4.00760621e-01
-7.35692561e-01 5.32217503e-01 -1.10739148e+00 -1.08022667e-01
9.73171294e-01 3.95247757e-01 -1.05589591e-01 -2.43048057e-01
-1.01905501e+00 3.40925634e-01 7.59299576e-01 3.93997729e-01
-8.58140886e-01 -5.97194374e-01 -1.11764908e+00 -4.25268039e-02
6.99296117e-01 -3.66690546e-01 1.03508210e+00 -1.13262796e+00
-1.29919457e+00 1.09780574e+00 3.36862616e-02 -4.62729752e-01
5.35262525e-01 -2.78844506e-01 -1.76771488e-02 1.72024041e-01
4.39794630e-01 1.00846589e+00 5.44778645e-01 -1.38443077e+00
-6.35785222e-01 -4.02613401e-01 1.60654619e-01 1.07198991e-01
-2.11112633e-01 -2.55705565e-01 -9.91242945e-01 -6.22377038e-01
3.76960397e-01 -8.99989426e-01 -2.78724402e-01 -6.45591989e-02
-7.48746336e-01 -2.53062189e-01 8.42512727e-01 -6.82077706e-01
1.07743073e+00 -2.09931350e+00 3.41701843e-02 3.58513921e-01
1.26810491e-01 4.56011474e-01 -9.73842070e-02 -2.90860236e-01
1.99388996e-01 2.32270852e-01 -8.21707428e-01 -6.10898554e-01
-4.13560085e-02 3.00575703e-01 2.60873530e-02 3.07906508e-01
3.72423053e-01 1.04674029e+00 -9.93628204e-01 -5.25935948e-01
1.96900636e-01 3.27408373e-01 -4.61743653e-01 2.92531490e-01
-6.40198350e-01 4.75508064e-01 -7.01660872e-01 8.23687911e-01
7.87024617e-01 -5.25319099e-01 -7.96260759e-02 -1.59153700e-01
1.84050202e-01 1.70737833e-01 -8.71799290e-01 1.95630860e+00
-3.99284095e-01 2.34110594e-01 2.58310854e-01 -1.19852018e+00
9.97752666e-01 -1.34211838e-01 2.97108084e-01 -5.52221239e-01
3.32556576e-01 3.27087373e-01 -1.27415404e-01 -1.78381920e-01
2.28676930e-01 8.30630884e-02 -1.47147089e-01 8.69929567e-02
1.78592547e-03 -2.69620657e-01 2.58370787e-01 1.86949745e-01
8.36356223e-01 3.92405927e-01 -1.37459412e-01 -2.50398904e-01
6.55442357e-01 1.79587290e-01 8.54482412e-01 4.61014241e-01
-2.98036665e-01 7.74833202e-01 6.68014586e-01 -1.93222657e-01
-8.04511011e-01 -8.10693085e-01 -1.65696219e-01 9.23261285e-01
5.17029166e-01 -3.74456793e-01 -1.16509497e+00 -1.15964508e+00
-1.80348352e-01 5.18902183e-01 -6.49945438e-01 -1.90434396e-01
-5.79208970e-01 -7.26861119e-01 4.23646271e-01 9.17269170e-01
8.44344497e-01 -1.16081965e+00 -2.76509881e-01 1.36583179e-01
-3.13104726e-02 -1.33134294e+00 -5.47389805e-01 1.70768395e-01
-7.65362799e-01 -1.16042650e+00 -7.01152682e-01 -9.10302401e-01
1.02097738e+00 -5.73371239e-02 8.97019923e-01 2.90091902e-01
-8.67634490e-02 -1.14157736e-01 -4.55144316e-01 -2.40973487e-01
-5.89535572e-02 3.12168688e-01 -4.70456034e-01 -7.14350268e-02
3.43540162e-01 -2.54726917e-01 -7.15590298e-01 4.57055807e-01
-8.75881791e-01 2.31175840e-01 4.54240471e-01 1.01605356e+00
1.12134278e+00 -1.01672672e-01 5.05640924e-01 -1.22509718e+00
4.83548231e-02 -3.36747378e-01 -9.50792491e-01 3.98350433e-02
-5.89830935e-01 -1.32621214e-01 8.36418808e-01 -1.17756978e-01
-9.26566362e-01 2.95282274e-01 -4.30455506e-01 -5.08871913e-01
-2.57270247e-01 5.03826499e-01 -5.68829775e-01 8.81123021e-02
3.04147333e-01 4.07109410e-02 -2.04376176e-01 -4.71078306e-01
4.29083735e-01 5.23905396e-01 5.96048176e-01 -6.90479457e-01
6.96954608e-01 5.40052056e-01 -2.50619024e-01 -3.79937142e-01
-1.25156009e+00 -4.55566317e-01 -5.08263469e-01 9.82568264e-02
1.16224861e+00 -7.84238040e-01 -3.96677554e-01 5.70357561e-01
-9.08692777e-01 -7.92135358e-01 -1.78279489e-01 2.88864821e-01
-4.50726151e-01 2.65512705e-01 -6.85998201e-01 -5.39513290e-01
-4.00216192e-01 -1.36467421e+00 1.30797768e+00 4.83810544e-01
8.68729129e-02 -9.25445020e-01 -3.73937428e-01 9.66795981e-01
4.28772578e-03 2.37856552e-01 7.10362256e-01 -6.74217463e-01
-6.67119503e-01 -1.22795194e-01 -7.22380519e-01 6.10317707e-01
5.13300821e-02 -1.58181921e-01 -1.03403890e+00 -2.15949878e-01
-4.53494787e-01 -5.55558503e-01 1.16467130e+00 4.25454199e-01
1.61773109e+00 -9.87863839e-02 -4.51948732e-01 9.73743379e-01
1.24115002e+00 1.08771577e-01 4.84699488e-01 3.78005028e-01
1.18426895e+00 5.75540185e-01 9.67878044e-01 1.68825507e-01
4.30915803e-01 4.87754166e-01 6.46908045e-01 -3.46367598e-01
-1.57115072e-01 -3.57939720e-01 -5.04497625e-02 4.84280080e-01
1.57856733e-01 -2.54360586e-01 -9.83306885e-01 5.72159886e-01
-1.84361613e+00 -2.71509379e-01 -1.48713127e-01 1.96781600e+00
9.03310001e-01 4.95287389e-01 -5.38053475e-02 1.51878176e-02
8.77419949e-01 2.27070794e-01 -7.35669136e-01 -6.44930303e-02
5.17913997e-02 3.48447144e-01 5.11653066e-01 5.38840234e-01
-1.26229942e+00 1.43652570e+00 3.92045760e+00 1.13212264e+00
-1.10952413e+00 1.42289534e-01 1.09434271e+00 1.54855460e-01
-3.82823229e-01 5.27833439e-02 -9.19345677e-01 6.80932164e-01
4.33412820e-01 3.28827858e-01 1.63074732e-01 9.57875192e-01
1.60112366e-01 -4.67878394e-02 -8.24310124e-01 7.93217242e-01
-1.36398494e-01 -1.07755184e+00 -1.90177485e-01 -1.79206312e-01
9.28860545e-01 -4.42350376e-03 -9.62254107e-02 2.91576266e-01
2.23055720e-01 -1.06732702e+00 7.07037747e-01 -3.82112898e-02
8.58982444e-01 -7.62155056e-01 8.56191337e-01 2.38483623e-01
-1.31972885e+00 2.46596158e-01 -3.51455837e-01 3.18457514e-01
1.84670448e-01 8.91730428e-01 -5.44332266e-01 5.96927881e-01
8.76268327e-01 8.27928722e-01 -3.84177774e-01 7.94369400e-01
-6.86134040e-01 7.81632543e-01 -3.99764717e-01 1.08984418e-01
5.30840218e-01 -2.52622157e-01 1.62618637e-01 1.02809823e+00
1.26357570e-01 3.45622785e-02 4.02718544e-01 1.15910137e+00
-4.62935269e-01 1.56899303e-01 -1.32960245e-01 5.52565232e-02
3.96547407e-01 1.36951721e+00 -1.21637368e+00 -2.63557345e-01
-2.89981812e-01 1.14273894e+00 5.22732735e-01 3.30800533e-01
-1.08526039e+00 -4.36517596e-01 4.37834561e-01 -8.03722255e-03
3.52765203e-01 4.08606082e-02 -4.39339161e-01 -1.08430529e+00
2.15840667e-01 -5.76938152e-01 4.83681887e-01 -6.30748451e-01
-1.18172753e+00 5.79563618e-01 -1.30501091e-01 -8.07983577e-01
2.57103890e-01 -6.61473930e-01 -6.67785466e-01 7.61163533e-01
-1.71294904e+00 -1.33503783e+00 -5.65735638e-01 3.11644822e-01
4.57648456e-01 2.27514192e-01 3.78541082e-01 3.45138907e-01
-8.96418273e-01 6.60739601e-01 -3.30359846e-01 4.12199765e-01
4.66677845e-01 -1.42290878e+00 4.54521358e-01 9.27984715e-01
6.67243125e-03 4.07120436e-01 2.26817086e-01 -7.02287376e-01
-7.75109887e-01 -1.54107904e+00 4.33097184e-01 -1.22115001e-01
5.09591222e-01 -4.56987351e-01 -1.11387277e+00 6.40127778e-01
-2.78831124e-01 5.56481183e-01 3.62813473e-01 -1.31093204e-01
-2.91431785e-01 -1.18431568e-01 -1.15647280e+00 4.86472130e-01
1.25549603e+00 -3.50651026e-01 -2.57321656e-01 4.16442275e-01
1.06609404e+00 -7.04510510e-01 -7.62182295e-01 6.91255987e-01
2.12414354e-01 -7.20333397e-01 7.89042890e-01 -2.25064725e-01
4.66852546e-01 -4.36418295e-01 8.10834318e-02 -1.02089059e+00
1.34897813e-01 -3.26304048e-01 7.40363002e-02 1.41329134e+00
7.03346848e-01 -6.53094530e-01 1.31345546e+00 4.28356797e-01
-6.03142858e-01 -1.11159897e+00 -8.21916938e-01 -6.52859032e-01
7.44335428e-02 -5.59257388e-01 5.18676400e-01 9.17008221e-01
-4.17309254e-01 4.67642546e-02 4.13026586e-02 3.03344280e-01
6.34850264e-01 1.66128337e-01 5.63874483e-01 -8.83261383e-01
-3.30143094e-01 -4.40509826e-01 -2.25753844e-01 -1.31190348e+00
4.49465513e-01 -9.65807736e-01 3.49340171e-01 -1.56079769e+00
6.44514859e-02 -9.06553507e-01 -3.79561603e-01 7.80178547e-01
-5.07086992e-01 5.14482081e-01 3.28850187e-02 2.82489508e-02
-7.40232527e-01 6.53543293e-01 1.48832726e+00 -1.78787410e-01
-2.14302868e-01 4.16312553e-02 -6.68144524e-01 8.87980759e-01
1.00821424e+00 -3.99895102e-01 -5.62748075e-01 -5.39545000e-01
-5.62509336e-03 -2.75955498e-01 5.21921575e-01 -8.42850685e-01
-1.64162833e-02 2.71796305e-02 8.69483426e-02 -5.17640471e-01
1.92547038e-01 -6.84448481e-01 -4.00315762e-01 2.65933454e-01
-2.01457590e-01 -3.95571500e-01 1.62410066e-01 4.86349493e-01
-3.46452892e-01 -3.23485404e-01 8.83890629e-01 -1.24296732e-01
-8.42428505e-01 5.46803057e-01 3.07966381e-01 3.51204038e-01
1.14185333e+00 -1.81289107e-01 -3.47578377e-01 -1.22380666e-02
-6.73683584e-01 6.96944416e-01 5.45658350e-01 2.39336282e-01
4.84767467e-01 -1.01141477e+00 -3.75467360e-01 2.04176635e-01
3.23548853e-01 8.94789159e-01 2.61036605e-01 7.87065506e-01
-6.53921545e-01 2.12999076e-01 1.51312783e-01 -8.06976557e-01
-9.80445802e-01 3.30670983e-01 3.47865760e-01 -2.19187453e-01
-5.34283280e-01 1.30549788e+00 3.58861595e-01 -5.73635578e-01
3.05275321e-01 -5.45945108e-01 -9.02836397e-03 -2.59006709e-01
1.07573941e-01 7.11174533e-02 -3.10971010e-02 -5.94459713e-01
-4.92574275e-01 7.14890420e-01 -2.30647415e-01 2.85591155e-01
1.12661827e+00 -6.86898008e-02 -5.95792048e-02 7.22988918e-02
1.19656861e+00 -2.05673918e-01 -1.65449083e+00 -2.71623284e-02
-2.40988545e-02 -3.01308990e-01 1.18547231e-01 -8.05355728e-01
-1.49160802e+00 9.07956541e-01 4.70533609e-01 -9.10482928e-02
1.13818920e+00 3.59553397e-01 1.10056365e+00 1.45964948e-02
2.62787730e-01 -1.04668558e+00 -2.87100598e-02 2.98300892e-01
4.37416434e-01 -1.49087739e+00 -2.35750124e-01 -8.44225526e-01
-4.65731651e-01 7.54223049e-01 1.04043770e+00 -5.43500632e-02
3.79624546e-01 1.79210186e-01 1.61445066e-01 -2.38260061e-01
-1.57954976e-01 -4.19929117e-01 3.12784851e-01 4.66567397e-01
3.71600389e-01 9.11342427e-02 -3.28359395e-01 9.49872255e-01
-1.25471458e-01 -1.80200309e-01 1.50008991e-01 7.24739730e-01
-4.32834566e-01 -1.15910912e+00 -5.13539389e-02 4.18088675e-01
-5.02261937e-01 -2.21507967e-01 -2.34631807e-01 7.38258302e-01
3.66167784e-01 8.34098935e-01 -5.79009689e-02 -1.54390693e-01
3.50417852e-01 -7.42765889e-02 2.29665905e-01 -8.02923322e-01
-4.42045301e-01 1.48520485e-01 4.59499955e-02 -7.02834845e-01
-3.76231700e-01 -3.68132383e-01 -1.87467372e+00 3.70448716e-02
-5.52476108e-01 2.57605731e-01 5.41412711e-01 8.74712408e-01
1.98386952e-01 6.64158762e-01 2.96378642e-01 -7.02653766e-01
-2.27692917e-01 -8.40666234e-01 -4.59694952e-01 5.69891155e-01
6.61474019e-02 -5.53255081e-01 -3.56659174e-01 2.15831562e-03] | [9.534878730773926, 0.5471881628036499] |
11064626-bdd5-470e-84c6-78f84572734b | language-conditioned-goal-generation-a-new-1 | null | null | https://openreview.net/forum?id=OeLMp3kWT8y | https://openreview.net/pdf?id=OeLMp3kWT8y | Language-Conditioned Goal Generation: a New Approach to Language Grounding in RL | In the real world, linguistic agents are also embodied agents: they perceive and act in the physical world. The notion of Language Grounding questions the interactions between language and embodiment: how do learning agents connect or ground linguistic representations to the physical world ? This question has recently been approached by the Reinforcement Learning community under the framework of instruction-following agents. In these agents, behavioral policies or reward functions are conditioned on the embedding of an instruction expressed in natural language. This paper proposes another approach: using language to condition goal generators. Given any goal-conditioned policy, one could train a language-conditioned goal generator to generate language-agnostic goals for the agent. This method allows to decouple sensorimotor learning from language acquisition and enable agents to demonstrate a diversity of behaviors for any given instruction. We propose a particular instantiation of this approach and demonstrate its benefits. | ['Olivier Sigaud', 'Mohamed Chetouani', 'Pierre-Yves Oudeyer', 'Ahmed Akakzia', 'Cédric Colas'] | 2020-06-12 | null | null | null | icml-workshop-larel-2020-7 | ['language-acquisition'] | ['natural-language-processing'] | [ 1.78721473e-01 4.76726413e-01 -7.86761567e-02 -2.58968603e-02
-1.04900785e-01 -7.99672663e-01 1.03828216e+00 1.78386003e-01
-6.94114149e-01 7.74338126e-01 2.41698116e-01 -3.16457063e-01
-8.23628306e-02 -1.17309153e+00 -6.21436357e-01 -6.52174532e-01
-2.29931369e-01 2.24213183e-01 3.54402438e-02 -5.01610398e-01
2.74314046e-01 6.42231166e-01 -1.49478471e+00 -9.08854306e-02
7.57398486e-01 1.75529033e-01 5.22261858e-01 8.40185642e-01
-1.02591194e-01 1.28904927e+00 -3.09093177e-01 3.35772097e-01
4.03392762e-02 -9.59986687e-01 -7.97029555e-01 1.51298806e-01
-2.01656967e-01 -2.23984256e-01 -1.68083042e-01 1.11777639e+00
4.50381599e-02 4.24062192e-01 5.16966522e-01 -1.29804039e+00
-8.64628971e-01 8.54388297e-01 4.37079147e-02 -1.42796353e-01
7.92002857e-01 4.97618824e-01 9.13130105e-01 -1.01328962e-01
5.80800533e-01 1.51105320e+00 -3.40115763e-02 1.08524513e+00
-1.54241168e+00 -1.48701563e-01 3.22015375e-01 -9.85405520e-02
-1.01933944e+00 -3.30805153e-01 6.71512544e-01 -5.82252920e-01
1.11367357e+00 -1.87719613e-01 7.97792673e-01 1.08100665e+00
3.72228056e-01 5.58359802e-01 1.55694866e+00 -9.53460515e-01
6.62913680e-01 9.94514674e-02 -2.71232631e-02 1.00596261e+00
1.36729881e-01 1.11937642e+00 -5.74436784e-01 5.81126362e-02
1.02525079e+00 -4.76552486e-01 -1.24505542e-01 -6.50799036e-01
-1.34846950e+00 8.69222581e-01 3.82497966e-01 6.12434864e-01
-6.15162969e-01 8.25845957e-01 4.22898941e-02 6.39109313e-01
-3.29685956e-01 9.67567503e-01 -1.99486583e-01 5.46893664e-03
-1.20453581e-01 2.99924195e-01 8.75641406e-01 4.71765637e-01
8.01843345e-01 4.12600666e-01 5.17601194e-03 9.39506516e-02
6.56780839e-01 6.81271732e-01 5.37718773e-01 -1.30795479e+00
-1.49476677e-01 4.58707601e-01 3.15710813e-01 -6.13806903e-01
-4.40418690e-01 -7.59728700e-02 3.63552719e-02 1.09716225e+00
3.51555437e-01 -3.96590739e-01 -5.93491197e-01 2.38724542e+00
4.02558446e-01 8.24922100e-02 5.76782286e-01 8.61478388e-01
1.75140709e-01 6.42654419e-01 4.62914437e-01 -7.60048926e-02
1.08000112e+00 -6.18789852e-01 -5.55895269e-01 -4.34441507e-01
8.38871777e-01 -9.85923558e-02 1.23717821e+00 1.62654012e-01
-1.04835582e+00 -4.75020021e-01 -1.20153117e+00 2.11793095e-01
-3.33287209e-01 -3.50090057e-01 6.22646689e-01 4.26404208e-01
-1.36201429e+00 6.44699037e-01 -9.30567563e-01 -6.60978198e-01
-1.32331446e-01 4.61565673e-01 -3.30474734e-01 6.45807207e-01
-1.10897398e+00 1.38815987e+00 7.37579048e-01 -3.06783050e-01
-1.45524740e+00 -8.20568576e-03 -1.14298999e+00 -3.18353288e-02
5.52201807e-01 -8.32160056e-01 1.63344860e+00 -1.28669846e+00
-2.17406154e+00 9.74351466e-01 2.43434221e-01 -5.36279976e-01
-2.36928985e-02 -2.16805432e-02 -2.36368954e-01 1.13820627e-01
2.42586434e-02 6.41927063e-01 6.72079027e-01 -1.43559563e+00
-6.11268878e-01 -1.93442062e-01 7.53200471e-01 3.81241471e-01
1.04839094e-01 1.71686616e-02 3.44391257e-01 -3.76901656e-01
-2.23576695e-01 -9.88406539e-01 -4.19639498e-01 -6.63313046e-02
8.56269747e-02 -2.81225175e-01 2.06824526e-01 3.22773829e-02
5.49881876e-01 -2.08691025e+00 6.30622327e-01 7.11890981e-02
7.40559623e-02 -1.30972266e-01 -6.82203948e-01 7.98972368e-01
-1.04145892e-01 -8.46637711e-02 2.44907662e-02 5.00941724e-02
5.68169117e-01 5.64192951e-01 -3.95775199e-01 5.51662207e-01
1.33463383e-01 1.08657408e+00 -1.28200316e+00 -3.67215097e-01
3.48129004e-01 2.71342576e-01 -7.65171111e-01 5.56759715e-01
-7.19688654e-01 7.93962538e-01 -9.64362502e-01 -9.72441211e-02
-4.02359635e-01 1.53121769e-01 6.09650075e-01 5.24143040e-01
-2.76547968e-01 5.58591604e-01 -1.07000995e+00 1.83348775e+00
-7.32473433e-01 2.94057637e-01 2.22043440e-01 -1.01662362e+00
8.45721960e-01 5.67372322e-01 1.00158833e-01 -7.77573705e-01
3.74760330e-01 1.01161368e-01 5.11508465e-01 -7.30420291e-01
4.30672765e-02 -7.77669489e-01 -3.12819809e-01 7.83534944e-01
1.11682467e-01 -5.57568312e-01 8.24284926e-02 7.50566125e-02
8.96426201e-01 9.02888119e-01 6.47846460e-01 -3.69733900e-01
6.21073365e-01 1.77328829e-02 1.59321830e-01 8.63696456e-01
-1.17784277e-01 -4.20051128e-01 2.69441813e-01 -1.47576198e-01
-7.33623266e-01 -1.35357618e+00 4.20818061e-01 1.48857474e+00
1.99994683e-01 -1.12233218e-02 -8.41710985e-01 -2.75968581e-01
-2.41418034e-01 1.18131566e+00 -7.78078318e-01 -4.22649413e-01
-8.12366903e-01 5.19245602e-02 2.47482121e-01 2.62692273e-01
8.59679654e-02 -1.89576948e+00 -1.47367847e+00 4.00580287e-01
3.93151343e-01 -9.25776482e-01 -1.33944377e-01 4.45413947e-01
-6.76168561e-01 -7.95779943e-01 -6.61758631e-02 -8.28788340e-01
6.78393185e-01 -2.50809550e-01 1.07778120e+00 3.73365313e-01
1.55641362e-01 9.89310265e-01 -2.49080032e-01 -3.65138769e-01
-1.05648696e+00 -3.00923824e-01 2.59013921e-01 -2.08344549e-01
2.09347039e-01 -7.47117043e-01 -1.89840332e-01 -3.10785592e-01
-9.94589090e-01 1.91075131e-01 3.22549522e-01 6.97312474e-01
7.87937641e-02 -1.32309824e-01 5.87734997e-01 -2.49058664e-01
1.00718677e+00 -1.96291327e-01 -9.54904258e-01 2.41550803e-01
-2.47308105e-01 7.22035229e-01 7.89234519e-01 -5.28283477e-01
-1.02394283e+00 3.90442908e-02 1.42697290e-01 2.35787243e-01
-5.49274027e-01 5.79989195e-01 -2.86844730e-01 -1.76738519e-02
6.83073103e-01 3.55872929e-01 2.15299606e-01 -9.16401371e-02
8.29627991e-01 1.43924698e-01 5.61175883e-01 -1.35486305e+00
8.72351825e-01 1.31150201e-01 1.63428113e-01 -5.84164083e-01
-4.32392389e-01 2.77337462e-01 -4.18178111e-01 -1.78056955e-01
1.14382088e+00 -5.67842245e-01 -1.18728054e+00 7.78782442e-02
-9.67786372e-01 -9.65976179e-01 -8.34224701e-01 7.04351962e-01
-1.36766624e+00 4.20142673e-02 -3.07156086e-01 -9.38508153e-01
2.64141619e-01 -1.39121509e+00 7.08264291e-01 2.95327395e-01
-4.95139182e-01 -1.17683625e+00 4.04548466e-01 -4.98849213e-01
3.84381264e-01 2.20853627e-01 1.05594742e+00 -6.06969774e-01
-4.90485758e-01 2.53360182e-01 4.47646171e-01 -6.10056929e-02
4.07517552e-01 -1.66149735e-01 -7.27067828e-01 -3.11013699e-01
2.20734015e-01 -6.10217631e-01 6.70000762e-02 1.69587553e-01
4.21092689e-01 -2.33658597e-01 -2.18127891e-01 6.82962760e-02
1.52819574e+00 5.93653560e-01 3.76098424e-01 5.05460262e-01
6.54711500e-02 8.60490561e-01 1.81135371e-01 8.52644369e-02
2.55472064e-01 7.54045904e-01 3.38305056e-01 2.79514283e-01
-3.43870334e-02 -5.40191233e-01 8.32750618e-01 4.97819364e-01
-4.58231345e-02 2.02710591e-02 -8.09736371e-01 3.74263883e-01
-1.76772594e+00 -1.17926562e+00 5.26067495e-01 1.85070300e+00
1.15102959e+00 3.77558805e-02 1.87083438e-01 -2.02572361e-01
3.70958984e-01 -9.35256183e-02 -4.54814434e-01 -6.97137713e-01
1.89555481e-01 3.58319849e-01 -1.16908818e-01 1.17144001e+00
-5.78994751e-01 1.24176455e+00 7.01412773e+00 1.58929706e-01
-1.14570189e+00 -5.82633391e-02 -4.78475057e-02 2.37491921e-01
-4.36598122e-01 1.39636099e-01 -3.65234286e-01 8.15770179e-02
1.06099987e+00 -6.47260249e-01 1.00792801e+00 4.98664498e-01
5.09208500e-01 -2.28082791e-01 -1.69583392e+00 2.96704978e-01
-3.96864951e-01 -1.13782072e+00 8.75574723e-03 1.07324190e-01
4.06009376e-01 -1.74144730e-01 -6.17800746e-03 4.88285989e-01
1.10060871e+00 -1.25363624e+00 9.43790853e-01 6.33596539e-01
3.53592992e-01 -4.11948711e-01 -2.03770697e-02 7.09358275e-01
-9.41531062e-01 -1.58188626e-01 1.23162203e-01 -5.65429628e-01
2.48225585e-01 -4.72059280e-01 -4.59024996e-01 2.01144233e-01
-7.56647438e-02 1.73737213e-01 -9.04216617e-02 5.32180309e-01
-6.83311105e-01 3.52602750e-01 -3.34735572e-01 -4.98470277e-01
5.08978844e-01 -5.66186130e-01 5.09722292e-01 8.30991149e-01
2.37317815e-01 2.39430413e-01 4.69712555e-01 1.16876757e+00
4.61751729e-01 -4.21915092e-02 -9.62994754e-01 -2.53370583e-01
2.03562751e-01 8.16765726e-01 -7.03728318e-01 -3.99047673e-01
-3.07057053e-01 6.01196706e-01 5.16599417e-01 5.60295701e-01
-7.28253722e-01 1.15402587e-01 7.73267329e-01 -2.29062930e-01
-2.54718307e-02 -6.18409634e-01 1.39149681e-01 -9.80303586e-01
-5.55083096e-01 -1.21932471e+00 -1.48223117e-01 -8.71417046e-01
-9.58581507e-01 5.01937866e-01 1.71153024e-01 -7.10508168e-01
-7.66936123e-01 -7.81223357e-01 -5.34940124e-01 8.46134961e-01
-1.25240731e+00 -7.90053546e-01 1.59611121e-01 6.58355653e-01
1.53681815e-01 -3.34571362e-01 1.21875334e+00 -4.16076779e-01
-1.81466460e-01 -3.77095230e-02 -4.83751267e-01 -9.61934589e-03
-8.70556608e-02 -1.40549219e+00 -5.35601452e-02 7.33417749e-01
4.54482496e-01 8.51341426e-01 1.16161716e+00 -2.59309173e-01
-1.56489599e+00 -5.50038636e-01 5.48433244e-01 -4.81718272e-01
1.11363661e+00 -1.41520053e-01 -7.37189054e-01 9.36613739e-01
8.39625001e-01 -2.33255759e-01 5.03314912e-01 -2.75604159e-01
-2.40821302e-01 1.65487394e-01 -9.71342385e-01 1.24900782e+00
1.03756559e+00 -8.61812413e-01 -1.26246822e+00 -9.48714931e-03
8.19049239e-01 -9.65527073e-02 -4.54672962e-01 -3.36054601e-02
3.03844780e-01 -7.46687949e-01 7.86121011e-01 -1.04915619e+00
2.04069152e-01 -7.28163302e-01 -3.29406917e-01 -1.47337973e+00
-3.61071736e-01 -7.37576246e-01 8.69128108e-02 8.19270134e-01
2.66071200e-01 -9.14561450e-01 4.16774571e-01 4.68113482e-01
-2.85627097e-02 -1.61984846e-01 -8.18408906e-01 -7.91654170e-01
6.49135172e-01 -3.33965957e-01 6.63779974e-01 6.92352116e-01
6.41182542e-01 4.89335626e-01 2.16652434e-02 1.76637888e-01
4.13258046e-01 2.44615182e-01 6.20642900e-01 -8.84678423e-01
-7.70044804e-01 -6.07373893e-01 -1.33032337e-01 -1.02264702e+00
7.19393134e-01 -1.25469077e+00 4.24590439e-01 -1.44506550e+00
-1.48071483e-01 -3.54350328e-01 -2.58116841e-01 7.42122293e-01
1.19291298e-01 -4.50415134e-01 5.20711780e-01 -1.28390029e-01
-3.47132355e-01 5.54035485e-01 1.42910480e+00 1.65240332e-01
-2.41548985e-01 -5.54946184e-01 -6.58978939e-01 8.31750453e-01
1.09631932e+00 -3.90487134e-01 -7.79866397e-01 -3.88948739e-01
4.71433252e-01 2.71484464e-01 6.05780244e-01 -1.01972961e+00
1.03641503e-01 -8.49329233e-01 -2.27743611e-01 2.94670999e-01
1.08428448e-01 -1.07582796e+00 1.02722809e-01 9.48841870e-01
-8.31786573e-01 2.85633445e-01 2.01183632e-01 3.75310928e-01
3.78605314e-02 -4.77212518e-01 7.15503633e-01 -3.96401644e-01
-9.59280610e-01 -2.58686572e-01 -1.12583995e+00 7.82586113e-02
1.33746779e+00 1.68358523e-03 -2.05218211e-01 -3.10526431e-01
-9.86777902e-01 1.41060382e-01 6.11912608e-01 3.34553778e-01
4.27760512e-01 -1.16237235e+00 -4.40957367e-01 6.74906969e-02
-1.48368469e-02 -5.49581110e-01 -4.86180782e-01 3.63107979e-01
-3.84180039e-01 3.34124267e-01 -5.46923935e-01 -2.88903058e-01
-6.48963988e-01 8.87467861e-01 8.59256685e-01 -7.49598593e-02
-5.03219187e-01 5.33457041e-01 5.67076325e-01 -5.52466631e-01
-3.17816399e-02 -4.97828513e-01 -2.74570853e-01 -4.93170619e-01
3.93097728e-01 -3.98825079e-01 -7.52931476e-01 -6.01731658e-01
-1.81294724e-01 4.37546074e-01 5.13867378e-01 -7.81807482e-01
1.07505214e+00 1.33987498e-02 -2.12273270e-01 6.32673979e-01
7.10358262e-01 1.30842477e-01 -1.30820620e+00 -9.81224775e-02
2.60534644e-01 9.40678418e-02 -1.89457744e-01 -7.59639919e-01
-5.51871896e-01 6.76216841e-01 4.41289932e-01 5.64866602e-01
8.56874764e-01 1.49177745e-01 -1.98923238e-02 4.80419427e-01
7.83037603e-01 -9.29118872e-01 5.45792222e-01 5.55062175e-01
1.04362452e+00 -7.89276659e-01 -3.41217071e-01 1.92797687e-02
-4.42008674e-01 1.04047763e+00 6.17863774e-01 -5.09467602e-01
3.40588868e-01 5.94186962e-01 1.80091754e-01 -2.50037789e-01
-8.48241925e-01 -5.03168821e-01 -3.17291892e-03 1.00789416e+00
3.95569026e-01 2.11920470e-01 -3.08127999e-01 -2.40999963e-02
-3.10371757e-01 2.97006607e-01 5.65870762e-01 1.18364859e+00
-7.16317117e-01 -1.46208727e+00 -4.38940406e-01 -3.18957210e-01
-5.32038063e-02 1.04770705e-01 -2.66128570e-01 8.22549343e-01
1.52035773e-01 1.00503397e+00 -4.28918190e-02 7.08765816e-03
2.17530429e-01 1.97508171e-01 1.01486850e+00 -1.06636667e+00
-5.73131204e-01 -2.18945220e-01 2.87405010e-02 -6.59009337e-01
-7.71907032e-01 -6.26395047e-01 -2.01916742e+00 -8.22193697e-02
6.96836635e-02 3.98532391e-01 4.87916023e-01 1.10505843e+00
2.52472721e-02 6.59137547e-01 3.27895910e-01 -7.14854062e-01
-7.55052805e-01 -4.56770211e-01 -4.73306537e-01 3.64398986e-01
6.28681481e-01 -6.33717716e-01 -3.06376845e-01 2.28917018e-01] | [4.2674407958984375, 1.2729872465133667] |
42c79ea1-3b17-4477-81f7-0b9cdcce141e | proposalclip-unsupervised-open-category | 2201.06696 | null | https://arxiv.org/abs/2201.06696v1 | https://arxiv.org/pdf/2201.06696v1.pdf | ProposalCLIP: Unsupervised Open-Category Object Proposal Generation via Exploiting CLIP Cues | Object proposal generation is an important and fundamental task in computer vision. In this paper, we propose ProposalCLIP, a method towards unsupervised open-category object proposal generation. Unlike previous works which require a large number of bounding box annotations and/or can only generate proposals for limited object categories, our ProposalCLIP is able to predict proposals for a large variety of object categories without annotations, by exploiting CLIP (contrastive language-image pre-training) cues. Firstly, we analyze CLIP for unsupervised open-category proposal generation and design an objectness score based on our empirical analysis on proposal selection. Secondly, a graph-based merging module is proposed to solve the limitations of CLIP cues and merge fragmented proposals. Finally, we present a proposal regression module that extracts pseudo labels based on CLIP cues and trains a lightweight network to further refine proposals. Extensive experiments on PASCAL VOC, COCO and Visual Genome datasets show that our ProposalCLIP can better generate proposals than previous state-of-the-art methods. Our ProposalCLIP also shows benefits for downstream tasks, such as unsupervised object detection. | ['Jianfei Cai', 'Yicheng Wu', 'Munawar Hayat', 'Hengcan Shi'] | 2022-01-18 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Shi_ProposalCLIP_Unsupervised_Open-Category_Object_Proposal_Generation_via_Exploiting_CLIP_Cues_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Shi_ProposalCLIP_Unsupervised_Open-Category_Object_Proposal_Generation_via_Exploiting_CLIP_Cues_CVPR_2022_paper.pdf | cvpr-2022-1 | ['object-proposal-generation'] | ['computer-vision'] | [ 2.16123790e-01 4.12090659e-01 -1.57310665e-01 -5.22709727e-01
-8.18009555e-01 -5.90496659e-01 5.85733175e-01 4.46463943e-01
-4.23813522e-01 2.17890888e-01 -8.80623609e-02 -1.45330891e-01
1.78924456e-01 -6.85835958e-01 -7.83490539e-01 -2.33319089e-01
1.52704865e-01 6.76120579e-01 1.04154730e+00 -3.40438187e-02
6.12907708e-01 3.41746032e-01 -1.85084260e+00 4.89935726e-01
8.79974663e-01 1.00223899e+00 4.69677746e-01 4.16152179e-01
-5.15816867e-01 3.13752890e-01 -4.31675583e-01 -4.57161039e-01
4.94309723e-01 -8.26119930e-02 -7.81685889e-01 5.03426790e-01
9.73434389e-01 -1.56272128e-01 2.59360015e-01 9.94067788e-01
4.87037182e-01 7.74537120e-03 9.26137745e-01 -1.18546236e+00
-2.15639532e-01 9.89215255e-01 -5.89740753e-01 -6.83167204e-02
8.18595737e-02 4.14872736e-01 1.20082092e+00 -1.42966151e+00
9.58176613e-01 1.42162776e+00 6.25361800e-01 5.31127214e-01
-1.43175519e+00 -7.21438169e-01 5.58887124e-01 2.06388444e-01
-1.52627623e+00 -2.68892765e-01 7.69842267e-01 -6.21770144e-01
7.03222930e-01 1.60897300e-01 5.54632068e-01 6.69308901e-01
-2.52418697e-01 1.15507400e+00 5.98753929e-01 -4.29833561e-01
2.92294413e-01 3.03118408e-01 1.76326677e-01 5.82923174e-01
2.80343622e-01 -1.54373586e-01 -8.15082565e-02 -5.28918058e-02
5.96666396e-01 -6.30843639e-02 2.30924282e-02 -9.43779111e-01
-1.45586538e+00 9.40573275e-01 9.11114752e-01 3.67760062e-02
-1.36910975e-01 8.86278301e-02 4.45398569e-01 -4.22554277e-02
2.13661432e-01 7.27403820e-01 -3.53800058e-01 5.92680335e-01
-1.26795173e+00 4.36840832e-01 7.05493927e-01 1.25233448e+00
9.44782734e-01 -1.45143762e-01 -7.09747255e-01 1.01591051e+00
5.28334796e-01 2.56372809e-01 3.01722437e-01 -8.28339338e-01
5.53539753e-01 8.70053113e-01 -1.12722833e-02 -8.06281328e-01
-4.50161397e-01 -5.43765962e-01 -4.63076144e-01 4.56903316e-02
3.32313240e-01 2.20895723e-01 -1.23898077e+00 1.20356023e+00
4.17394459e-01 -1.34523779e-01 -2.74808437e-01 1.15168822e+00
1.33960390e+00 5.49444139e-01 1.27587706e-01 1.39896031e-02
1.31060445e+00 -1.39194834e+00 -2.49998346e-01 -2.51771688e-01
3.78697902e-01 -8.14262211e-01 7.93837607e-01 4.01124030e-01
-9.21590924e-01 -8.50135446e-01 -6.77586913e-01 -1.00892775e-01
-2.29887605e-01 5.59484601e-01 5.78737795e-01 3.38464975e-01
-8.25817585e-01 1.21358193e-01 -4.33181137e-01 -4.13330734e-01
8.24866831e-01 4.00429994e-01 -1.22187249e-01 -6.39017522e-02
-4.05308753e-01 4.91558492e-01 9.84370708e-01 -1.09172270e-01
-9.29587007e-01 -6.13725543e-01 -8.93798292e-01 1.42464131e-01
6.71570361e-01 -6.21631920e-01 1.12428761e+00 -9.37364042e-01
-1.14665127e+00 6.92804515e-01 2.56760538e-01 -5.86149931e-01
5.77398121e-01 1.64934754e-01 1.85767606e-01 2.43105739e-01
3.38658571e-01 1.93463302e+00 1.13150859e+00 -1.46039689e+00
-9.77112353e-01 -3.42550501e-02 -1.31574959e-01 -4.71870713e-02
-4.04335409e-02 -4.87229368e-03 -9.89280164e-01 -7.03444183e-01
4.99083549e-01 -9.34247255e-01 -7.74275005e-01 2.49498248e-01
-7.57832110e-01 -7.30817437e-01 5.64073682e-01 -2.41093934e-01
9.54378724e-01 -1.95049107e+00 2.20743958e-02 3.07990581e-01
2.22715512e-01 2.85627574e-01 -4.14954782e-01 -6.66598277e-03
3.02647114e-01 6.13105483e-02 -3.54850501e-01 -3.81630123e-01
1.52025610e-01 -1.92123458e-01 -3.36932302e-01 8.20974857e-02
6.05320752e-01 9.03422177e-01 -7.69824922e-01 -9.63170111e-01
2.48199642e-01 9.73057654e-03 -1.12462223e+00 2.28048608e-01
-7.63655245e-01 1.26727521e-01 -1.82918951e-01 7.64397800e-01
8.49505186e-01 -1.16156742e-01 8.67091268e-02 -2.62289792e-01
-2.06148878e-01 7.49823451e-02 -1.41189969e+00 1.77680123e+00
-1.03522472e-01 5.76692104e-01 -2.60427147e-01 -9.14700449e-01
1.17611873e+00 -1.91571042e-01 4.57704887e-02 -1.02481894e-01
2.73396224e-01 3.22908372e-01 1.17049649e-01 -2.58251131e-01
8.70758891e-01 2.96749890e-01 -2.27892622e-01 2.99222499e-01
4.69173282e-01 -5.40618002e-01 6.20439351e-01 4.04992253e-01
9.09020603e-01 3.58451337e-01 2.82558292e-01 -1.87946290e-01
5.28909087e-01 1.99563324e-01 6.12120211e-01 1.04522777e+00
-3.51448417e-01 9.36489582e-01 4.93229806e-01 -4.74137247e-01
-1.03265917e+00 -8.97947848e-01 -2.21912801e-01 1.21330369e+00
3.16855371e-01 -2.77616471e-01 -6.44208968e-01 -1.16943610e+00
1.84012018e-02 5.14020741e-01 -2.73294210e-01 3.03741872e-01
-4.53909218e-01 -3.63506794e-01 2.17230767e-01 5.89316487e-01
1.16730593e-01 -1.35014689e+00 -5.25153339e-01 2.02991724e-01
-2.46788204e-01 -1.08851719e+00 -5.35227537e-01 1.67425990e-01
-8.89270663e-01 -1.04044950e+00 -6.82230353e-01 -1.01699317e+00
1.16908157e+00 3.26219469e-01 1.09855974e+00 7.63233975e-02
-6.94357395e-01 2.16130897e-01 -4.97662425e-01 -4.98365641e-01
-3.42022419e-01 2.87070841e-01 -2.81960815e-01 8.87145326e-02
3.06116194e-01 -5.34475483e-02 -5.51711380e-01 5.25555372e-01
-7.51582623e-01 3.68811011e-01 8.70646834e-01 6.29369676e-01
8.23767900e-01 -4.85543400e-01 6.42105818e-01 -8.25366378e-01
1.47418380e-01 -2.29937136e-01 -9.40605879e-01 1.27320647e-01
-2.55993247e-01 2.20909014e-01 1.76730618e-01 -5.37379622e-01
-9.91672695e-01 6.36127949e-01 -6.30788654e-02 -6.01855218e-01
-5.07298052e-01 -4.39721793e-02 -1.70693442e-01 3.29531468e-02
7.47583866e-01 3.55472341e-02 -2.50357270e-01 -4.65976715e-01
8.35911393e-01 5.96898019e-01 4.95857298e-01 -3.67674798e-01
8.47596407e-01 4.70471412e-01 -3.67170155e-01 -6.83969975e-01
-7.92979896e-01 -9.83434498e-01 -1.01598287e+00 -2.61318922e-01
7.93908715e-01 -9.59015250e-01 -3.54260594e-01 -4.85929959e-02
-1.39282715e+00 -1.72143728e-01 -4.28306967e-01 3.35787624e-01
-5.37052453e-01 2.87064105e-01 -2.63493717e-01 -6.04246378e-01
-4.84679073e-01 -1.13813794e+00 1.37727833e+00 2.79907435e-01
-8.29134732e-02 -4.55480218e-01 -3.40285420e-01 4.26808119e-01
9.42186564e-02 -1.26858428e-01 4.52983588e-01 -9.60169554e-01
-1.17588282e+00 -1.60781860e-01 -7.49905646e-01 2.00372487e-01
-4.38404739e-01 2.00663134e-02 -7.72691607e-01 -2.15597779e-01
-7.06771314e-01 -5.16899467e-01 1.39513838e+00 2.98576206e-01
1.37284005e+00 -2.92851120e-01 -6.17383361e-01 5.25616169e-01
1.46402812e+00 -1.68406814e-01 5.37373543e-01 7.46707320e-02
7.97054470e-01 8.31595421e-01 9.01709735e-01 5.01309693e-01
4.33462739e-01 6.89619839e-01 6.22835219e-01 -3.27891521e-02
-3.29861313e-01 -2.61841297e-01 8.70970488e-02 5.02585292e-01
2.48134539e-01 -1.73646301e-01 -8.15201104e-01 9.79075968e-01
-1.88403141e+00 -7.23540664e-01 -4.74631697e-01 1.88776684e+00
6.86463416e-01 4.23417836e-01 3.20373982e-01 -3.52964625e-02
8.84162545e-01 -2.60145962e-01 -4.13160712e-01 -1.47519499e-01
1.21540353e-01 1.04018010e-01 4.08386588e-01 8.37730616e-02
-1.34448969e+00 1.26323903e+00 5.62704849e+00 1.07745266e+00
-9.55349684e-01 1.90530494e-01 3.76948535e-01 1.63568109e-01
-1.99984223e-01 2.38236040e-01 -1.28015172e+00 2.12796733e-01
-2.02656887e-03 3.71951014e-01 -3.99505645e-01 1.29727602e+00
-1.21501036e-01 -1.32151425e-01 -1.25995922e+00 7.91314483e-01
4.19102550e-01 -1.41514862e+00 3.44026417e-01 -1.14412948e-01
8.73203635e-01 1.08232856e-01 -2.81598538e-01 5.37405550e-01
2.93884218e-01 -4.67252702e-01 1.06823385e+00 3.90201956e-01
4.79005754e-01 -4.74226236e-01 7.54827440e-01 4.10624743e-01
-1.23215306e+00 -2.87056327e-01 -7.20409989e-01 2.93460131e-01
5.01897968e-02 5.06925344e-01 -1.47305751e+00 2.71827340e-01
4.28427458e-01 5.32606006e-01 -1.10205615e+00 1.76874125e+00
-4.27082121e-01 6.22221649e-01 -2.72315830e-01 -2.13773102e-01
2.62155592e-01 2.24005356e-01 5.58061421e-01 1.58727169e+00
8.95447657e-02 -1.06312975e-01 8.32865894e-01 1.26969326e+00
-5.83339259e-02 3.92111003e-01 -1.61624283e-01 3.06291819e-01
4.05707717e-01 1.56681883e+00 -1.24714577e+00 -5.39052248e-01
-1.50340170e-01 6.56567633e-01 4.08733934e-01 -1.23476297e-01
-8.11272383e-01 -5.51536560e-01 -9.07889567e-03 1.94748551e-01
9.36090112e-01 1.56583842e-02 -1.06990241e-01 -1.12404537e+00
-1.14539348e-01 -4.28484380e-01 4.11342025e-01 -7.64884949e-01
-1.20734000e+00 4.92548257e-01 8.71400312e-02 -1.48367119e+00
-3.55411433e-02 -4.90720600e-01 -7.31365561e-01 1.85120046e-01
-1.46059608e+00 -1.53706670e+00 -3.71418566e-01 1.41867310e-01
1.19445360e+00 -1.68900639e-01 2.27498680e-01 1.93359360e-01
-4.22861069e-01 4.67797637e-01 -4.07398582e-01 2.83475459e-01
7.35464871e-01 -1.20908368e+00 4.18827951e-01 8.92457306e-01
4.41268206e-01 3.35541844e-01 5.20030558e-01 -6.75666332e-01
-8.98785949e-01 -1.61040246e+00 6.46871686e-01 -4.01327163e-01
3.79234225e-01 -4.82310414e-01 -8.35907042e-01 3.27502578e-01
-2.00389735e-02 4.68245834e-01 3.22540700e-01 -1.17757887e-01
-5.26357532e-01 3.25026340e-03 -8.98259938e-01 4.74853724e-01
1.06352997e+00 2.16016889e-01 -7.71672130e-01 4.26391542e-01
9.75031793e-01 -3.05363387e-01 -3.58052075e-01 5.05675077e-01
3.87284994e-01 -9.27273810e-01 8.98335099e-01 -3.51643503e-01
3.51731181e-01 -5.92698812e-01 5.38627803e-02 -8.80606413e-01
-3.74810249e-01 -2.83731610e-01 1.00499317e-01 1.37447119e+00
6.78801119e-01 -5.06725498e-02 7.38760769e-01 8.10774267e-02
-3.43802989e-01 -4.70454812e-01 -6.30132437e-01 -7.06217706e-01
-3.05957109e-01 -7.25048840e-01 2.97888428e-01 4.47570235e-01
-2.37855554e-01 3.12474370e-01 1.44572947e-02 -3.68196629e-02
8.37827981e-01 3.07350606e-01 1.24735832e+00 -1.36557806e+00
-1.22751944e-01 -6.24760866e-01 -6.13593757e-01 -1.09101176e+00
1.25279784e-01 -1.26471281e+00 6.41648889e-01 -1.70981801e+00
4.39025074e-01 -9.50399876e-01 -1.39157996e-01 6.75783277e-01
-2.29502723e-01 6.96405947e-01 7.13336766e-01 3.18662494e-01
-1.39122069e+00 4.70435202e-01 1.04812193e+00 -4.49705213e-01
-4.51765716e-01 2.16407869e-02 -5.07177591e-01 9.09163892e-01
4.89941835e-01 -5.82093537e-01 4.67534550e-02 -7.17781559e-02
-1.11469887e-02 -3.02449465e-01 5.04852176e-01 -1.25443840e+00
1.66057721e-01 -1.59887761e-01 2.38896519e-01 -1.25247252e+00
9.29298922e-02 -5.27580082e-01 -3.11074644e-01 4.85983580e-01
-4.62301016e-01 -7.07476258e-01 1.33935004e-01 7.21775353e-01
-2.97787897e-02 -4.58658367e-01 7.38725722e-01 -2.42421821e-01
-9.82701242e-01 2.59118021e-01 -2.60565639e-01 2.67767012e-02
1.08868599e+00 -2.48394251e-01 -1.56146377e-01 2.15905696e-01
-7.44253457e-01 4.91754234e-01 3.01679611e-01 6.64963722e-01
8.17821026e-01 -1.22396243e+00 -8.00143182e-01 1.76147282e-01
6.26366436e-01 4.30459589e-01 -1.11530041e-02 7.59183288e-01
-5.03287494e-01 4.54000711e-01 3.17923576e-02 -1.19153500e+00
-1.30792558e+00 7.44024754e-01 -8.51052031e-02 -1.28852382e-01
-5.68152666e-01 1.08270240e+00 5.25622547e-01 -7.97372162e-01
3.32281321e-01 -5.56772172e-01 -4.08114284e-01 2.44934991e-01
3.26528788e-01 1.18420459e-01 -3.98228914e-02 -6.97479129e-01
-2.56875426e-01 5.99054456e-01 -2.33781114e-01 1.20363183e-01
1.11126804e+00 -1.45700555e-02 -8.55298489e-02 2.31656134e-02
7.21361339e-01 -2.73772806e-01 -1.30626798e+00 -3.07749808e-01
2.17440292e-01 -3.86842310e-01 -2.04749659e-01 -5.06493747e-01
-8.50787640e-01 8.50420535e-01 6.84131563e-01 -6.22399412e-02
7.29634583e-01 3.39527100e-01 4.22161460e-01 5.65625250e-01
4.57872331e-01 -1.17382252e+00 4.81996328e-01 4.18344587e-01
1.06770396e+00 -1.58891618e+00 1.53522953e-01 -7.88772583e-01
-6.07921124e-01 1.19972789e+00 1.08342946e+00 -1.03116632e-01
4.91697401e-01 -1.01084642e-01 -1.57536954e-01 -2.63036583e-02
-7.42010891e-01 -6.35233819e-01 6.25584185e-01 5.83447278e-01
1.21161610e-01 1.06797807e-01 -4.26802397e-01 6.26347601e-01
-1.05674066e-01 -2.13417262e-01 6.70754910e-01 6.51533067e-01
-1.00357711e+00 -9.28271830e-01 -6.83390856e-01 6.08966827e-01
-9.29913074e-02 1.18530445e-01 -4.22026128e-01 5.07571518e-01
7.19764352e-01 7.52096117e-01 2.90258497e-01 -6.28191978e-02
1.25971928e-01 -5.96978329e-02 2.61348277e-01 -1.10318661e+00
-6.67292118e-01 2.48448685e-01 -7.77824372e-02 -4.65999544e-01
-5.08561194e-01 -5.34263968e-01 -1.31178725e+00 5.79053700e-01
-1.09956741e+00 -2.22717691e-02 8.72776210e-01 8.13585758e-01
4.48970109e-01 5.60592473e-01 3.48192245e-01 -1.32483780e+00
-2.85765320e-01 -1.08072484e+00 -8.16720501e-02 3.81805182e-01
-9.86170396e-02 -6.16294563e-01 -1.34812266e-01 3.07867199e-01] | [9.311633110046387, 0.701696515083313] |
6fd6113f-762e-4873-a7c7-e50e0eb7381a | evaluating-mixed-initiative-conversational | 2204.08046 | null | https://arxiv.org/abs/2204.08046v2 | https://arxiv.org/pdf/2204.08046v2.pdf | Evaluating Mixed-initiative Conversational Search Systems via User Simulation | Clarifying the underlying user information need by asking clarifying questions is an important feature of modern conversational search system. However, evaluation of such systems through answering prompted clarifying questions requires significant human effort, which can be time-consuming and expensive. In this paper, we propose a conversational User Simulator, called USi, for automatic evaluation of such conversational search systems. Given a description of an information need, USi is capable of automatically answering clarifying questions about the topic throughout the search session. Through a set of experiments, including automated natural language generation metrics and crowdsourcing studies, we show that responses generated by USi are both inline with the underlying information need and comparable to human-generated answers. Moreover, we make the first steps towards multi-turn interactions, where conversational search systems asks multiple questions to the (simulated) user with a goal of clarifying the user need. To this end, we expand on currently available datasets for studying clarifying questions, i.e., Qulac and ClariQ, by performing a crowdsourcing-based multi-turn data acquisition. We show that our generative, GPT2-based model, is capable of providing accurate and natural answers to unseen clarifying questions in the single-turn setting and discuss capabilities of our model in the multi-turn setting. We provide the code, data, and the pre-trained model to be used for further research on the topic. | ['Fabio Crestani', 'Mohammad Aliannejadi', 'Ivan Sekulić'] | 2022-04-17 | null | null | null | null | ['user-simulation', 'conversational-search'] | ['natural-language-processing', 'natural-language-processing'] | [ 5.40499203e-02 3.47240180e-01 3.18370104e-01 -4.88479257e-01
-1.12810373e+00 -1.14370394e+00 8.41788888e-01 1.16208918e-01
-2.86723584e-01 8.52442622e-01 3.08358163e-01 -5.11182189e-01
-2.03669779e-02 -3.14315289e-01 -3.12804341e-01 -9.05070975e-02
4.79539990e-01 1.04422510e+00 1.99190393e-01 -7.28981197e-01
3.31640452e-01 -3.06925755e-02 -1.45554554e+00 4.67933357e-01
1.16498590e+00 4.55966473e-01 4.81035292e-01 1.16902888e+00
-2.74593741e-01 7.49561071e-01 -1.14773214e+00 -5.15895724e-01
-1.71699122e-01 -6.60447717e-01 -1.55267060e+00 9.21165347e-02
2.33009309e-01 -4.47591007e-01 1.57610610e-01 4.10076082e-01
6.43452525e-01 4.89280194e-01 2.54237562e-01 -1.41521657e+00
-5.36916256e-01 4.76953447e-01 3.10061157e-01 2.32484370e-01
1.19844627e+00 3.72552097e-01 1.09103239e+00 -8.48183095e-01
5.03588736e-01 1.36888063e+00 3.97749245e-01 7.37714708e-01
-8.77698541e-01 -2.64083773e-01 6.52162880e-02 1.30028367e-01
-1.01325333e+00 -4.63688254e-01 5.20714581e-01 -3.59709382e-01
8.72841477e-01 7.05249727e-01 4.33891743e-01 1.15550947e+00
-2.50620782e-01 8.45670342e-01 1.11139810e+00 -4.75247025e-01
2.08953395e-01 4.95982140e-01 3.93344700e-01 4.56804246e-01
-2.22889140e-01 -2.13313490e-01 -6.78519309e-01 -5.80018818e-01
2.91334510e-01 -5.72238564e-01 -4.82226282e-01 1.17885120e-01
-1.09950292e+00 8.92715633e-01 -1.95592418e-02 4.19943601e-01
-3.07892203e-01 -1.44542575e-01 7.45703429e-02 3.47574025e-01
4.82694983e-01 8.32870960e-01 -2.71298975e-01 -7.51743436e-01
-5.20814419e-01 7.93745637e-01 1.57893801e+00 1.21456027e+00
6.21168375e-01 -6.17384434e-01 -7.16924846e-01 9.94257927e-01
8.11142549e-02 5.17803729e-01 5.04080594e-01 -1.26954818e+00
4.40678716e-01 7.10206509e-01 8.61693799e-01 -8.55165482e-01
-3.82192284e-01 -2.89851986e-02 -2.09333345e-01 -4.87849057e-01
5.27804196e-01 -4.77002203e-01 -2.04900414e-01 1.65477717e+00
4.74329412e-01 -3.85695279e-01 2.24222645e-01 1.07169509e+00
1.14945936e+00 4.46254581e-01 -5.65244928e-02 -1.52555138e-01
1.83123779e+00 -1.06903219e+00 -8.89879644e-01 -1.47961125e-01
6.93955243e-01 -9.80513632e-01 1.54925394e+00 6.83417320e-02
-1.21783352e+00 -4.85842824e-01 -4.09375191e-01 -3.61027807e-01
-6.51923716e-02 1.04900077e-01 2.53245682e-01 6.43826246e-01
-1.32024169e+00 -1.03171449e-02 -1.78087518e-01 -6.56925023e-01
-3.66928726e-01 2.11775050e-01 6.42400160e-02 4.05710377e-02
-1.60993993e+00 9.15998101e-01 -2.94803619e-01 5.78888506e-02
-5.84850729e-01 -4.96682584e-01 -7.77585685e-01 1.05312593e-01
5.05127966e-01 -6.91139817e-01 2.30718780e+00 -4.71260786e-01
-1.70989704e+00 7.72984505e-01 -7.29018867e-01 -2.68289298e-01
4.80408967e-01 -1.72749698e-01 2.43867636e-02 1.85644940e-01
4.60766524e-01 7.26954043e-01 3.89649153e-01 -1.52281272e+00
-6.04893386e-01 -2.56861430e-02 8.20984900e-01 5.24320364e-01
1.59187466e-01 1.16207540e-01 -4.07189012e-01 -1.72271281e-01
-3.97623122e-01 -1.22549450e+00 -3.47686522e-02 -5.83979368e-01
-6.12766385e-01 -7.79235125e-01 6.37039959e-01 -6.52043283e-01
1.15481472e+00 -1.57174933e+00 -2.34677345e-01 -1.17239237e-01
3.08892250e-01 5.16607940e-01 -3.48660618e-01 9.80650425e-01
3.11680824e-01 2.78889477e-01 -1.01294108e-01 -5.00860095e-01
1.59613490e-01 1.24474965e-01 -3.13764334e-01 -2.42974237e-01
3.87533419e-02 1.27631557e+00 -1.23577797e+00 -4.74884450e-01
-5.39792813e-02 1.38235679e-02 -3.99684459e-01 7.36223876e-01
-4.86263126e-01 7.46712625e-01 -4.02474403e-01 2.82884628e-01
2.99985468e-01 -3.90874743e-01 -5.35204187e-02 1.99091628e-01
4.55377027e-02 4.92899418e-01 -8.59341621e-01 1.44281948e+00
-9.23121095e-01 7.39161611e-01 2.43760034e-01 -3.94812435e-01
8.24579418e-01 5.11110783e-01 -2.43111118e-03 -7.34823048e-01
7.73173720e-02 7.62121081e-02 -7.26252124e-02 -8.94140661e-01
9.48797762e-01 9.72226635e-02 -2.17710704e-01 1.23865068e+00
-2.60626942e-01 -6.03450239e-01 4.34479415e-01 5.95743775e-01
9.81557369e-01 -4.10855502e-01 1.56759366e-01 -3.58386815e-01
7.64023125e-01 1.70299202e-01 -2.39714190e-01 1.21247780e+00
-3.20390970e-01 4.62590396e-01 4.30765599e-01 -2.72100657e-01
-5.41885614e-01 -6.11237884e-01 3.67644340e-01 1.31147671e+00
3.73028755e-01 -4.79103267e-01 -1.18655348e+00 -5.62108397e-01
-2.62566239e-01 1.06119132e+00 -3.36047292e-01 2.15803877e-01
-4.26317245e-01 -8.63295645e-02 5.55477142e-01 1.15717061e-01
6.05568469e-01 -1.14146090e+00 -7.62758493e-01 2.37984359e-01
-1.21157324e+00 -1.34911382e+00 -8.91528845e-01 -4.71619576e-01
-2.62221605e-01 -1.29277539e+00 -8.07529390e-01 -7.81802237e-01
3.91070187e-01 8.31171036e-01 1.38198662e+00 5.63993156e-01
6.22520447e-02 1.14680719e+00 -5.59004962e-01 -4.75899905e-01
-7.52452314e-01 3.35912913e-01 -2.57226974e-01 -1.35697603e-01
3.63633662e-01 -2.65966892e-01 -6.90442264e-01 9.62534010e-01
-9.25293505e-01 1.95418894e-01 6.36245757e-02 5.28791070e-01
-1.79484010e-01 -5.24947822e-01 8.14452946e-01 -7.71811545e-01
2.01919699e+00 -4.61770654e-01 -2.91763216e-01 5.65111578e-01
-3.51680100e-01 1.31304711e-01 3.25122207e-01 -3.59386861e-01
-1.17067516e+00 -3.37013155e-01 -1.05285585e-01 3.20522994e-01
-1.49210289e-01 3.47538918e-01 9.95162055e-02 9.61998329e-02
1.04256463e+00 2.35361621e-01 -5.77635318e-03 -2.48741284e-01
5.01210988e-01 1.10081244e+00 3.33539933e-01 -8.52524400e-01
6.10498607e-01 5.44609427e-02 -7.22930729e-01 -9.18746769e-01
-9.19143498e-01 -7.31662989e-01 -2.57548004e-01 -4.42112207e-01
6.84110284e-01 -6.68006897e-01 -1.32245898e+00 2.62254864e-01
-1.72810543e+00 -7.16787934e-01 3.23064439e-02 -8.13125893e-02
-6.32808745e-01 5.46919048e-01 -4.34832513e-01 -1.31386900e+00
-5.15047967e-01 -1.31537366e+00 1.42788053e+00 3.67591143e-01
-1.00015771e+00 -9.67762172e-01 3.73887032e-01 1.09092963e+00
7.33547091e-01 -2.86935955e-01 7.55430698e-01 -1.06361055e+00
-4.07092184e-01 -2.26072311e-01 -5.82450964e-02 -4.91501614e-02
2.38254011e-01 -2.59224325e-01 -9.53516006e-01 -2.45381277e-02
1.66629419e-01 -5.98351419e-01 4.71451506e-03 -1.31655067e-01
7.15309858e-01 -6.30943835e-01 -1.33575439e-01 -4.99109894e-01
5.28766334e-01 1.58822507e-01 3.64069968e-01 -1.43413851e-02
5.86342402e-02 1.08549750e+00 7.69852579e-01 3.69497001e-01
1.08157444e+00 8.54865909e-01 7.85505213e-03 1.66179478e-01
2.11287856e-01 -1.78108409e-01 -7.02149868e-02 5.79775155e-01
1.12056144e-01 -7.37535834e-01 -9.02867138e-01 7.30958879e-01
-1.87060130e+00 -8.93803000e-01 -3.16268414e-01 1.88151300e+00
1.17027080e+00 -2.97546118e-01 2.71744013e-01 -1.07751302e-01
8.02515447e-01 -1.01049259e-01 -3.12502176e-01 -6.56899452e-01
1.76783070e-01 9.09701809e-02 -4.69674379e-01 1.30042970e+00
-4.18041229e-01 1.01055336e+00 5.94001293e+00 4.12973404e-01
-6.48754954e-01 2.98432223e-02 5.23521066e-01 2.37383634e-01
-5.94369888e-01 1.11478865e-01 -6.30197823e-01 4.76290137e-01
9.79749858e-01 -5.76637149e-01 7.10255563e-01 5.58050811e-01
6.48246050e-01 -6.75742090e-01 -1.31906974e+00 8.93817902e-01
1.87948748e-01 -1.02010870e+00 -1.49934236e-02 -2.85480261e-01
4.51971054e-01 -5.15563428e-01 -3.75828415e-01 5.95559001e-01
5.42852521e-01 -8.29706192e-01 3.41710508e-01 2.75218934e-01
4.16855603e-01 -2.36643538e-01 7.81700194e-01 9.84195411e-01
-8.24326873e-01 1.66009605e-01 1.79168493e-01 -2.87467927e-01
4.73124474e-01 2.63460815e-01 -1.46189308e+00 3.06497574e-01
4.11491036e-01 -2.09299892e-01 -5.80039382e-01 6.77897990e-01
-4.64592218e-01 4.99211401e-01 -2.30729654e-01 -7.14304864e-01
1.93847895e-01 -4.21314612e-02 6.09939873e-01 1.10919726e+00
4.44675013e-02 5.97127140e-01 2.06025660e-01 8.53268266e-01
-1.76967159e-01 1.08250439e-01 -5.35921574e-01 5.95956892e-02
7.11399555e-01 1.34777021e+00 -2.42103189e-01 -5.00030339e-01
1.42146900e-01 1.09696400e+00 2.22466916e-01 5.37681282e-01
-6.30833864e-01 -4.01273936e-01 3.78092676e-01 1.27328426e-01
-3.61567408e-01 -1.91082343e-01 -1.06815323e-01 -8.96159828e-01
3.64819467e-01 -1.44575059e+00 9.71377045e-02 -1.24338067e+00
-9.17405188e-01 8.38453054e-01 2.52447039e-01 -8.52883935e-01
-9.53518629e-01 -9.71588567e-02 -8.25917304e-01 1.25690591e+00
-1.39734089e+00 -7.71701396e-01 -9.10569131e-01 4.23477381e-01
1.02576160e+00 3.20451260e-01 7.78518081e-01 -3.54732871e-02
-1.88922331e-01 4.69014049e-01 -6.58649087e-01 -1.25460982e-01
9.42177832e-01 -1.11451530e+00 6.47005737e-01 3.01579356e-01
1.46782070e-01 8.98801446e-01 9.99827802e-01 -6.30791843e-01
-1.18015087e+00 -6.17032230e-01 1.44064295e+00 -9.35741305e-01
5.31995237e-01 -6.37613237e-01 -1.03478444e+00 3.87243181e-01
6.05970025e-01 -7.35442698e-01 8.82898510e-01 7.60531574e-02
2.58233279e-01 4.65310901e-01 -1.04963231e+00 8.55887771e-01
9.81506169e-01 -7.88407445e-01 -8.43677700e-01 8.79950702e-01
9.38293278e-01 -7.06631839e-01 -3.56681794e-01 -6.38667271e-02
4.00255024e-01 -8.40043128e-01 5.53161919e-01 -7.79880583e-01
2.75788486e-01 -2.91687436e-02 2.09224328e-01 -1.49337065e+00
1.35297284e-01 -1.27072883e+00 2.40409613e-01 1.21217728e+00
6.97649896e-01 -6.83226883e-01 5.82811415e-01 1.44103146e+00
8.55764002e-02 -6.26031578e-01 -8.12252343e-01 -4.31980938e-01
-2.11444110e-01 -3.43896300e-01 5.82231164e-01 4.82560664e-01
4.12344396e-01 8.73852611e-01 -2.51172870e-01 1.00583896e-01
-3.66110355e-02 1.41920641e-01 1.30753577e+00 -1.06608641e+00
-7.21602589e-02 -1.18752539e-01 4.47853118e-01 -1.56853998e+00
2.04539299e-01 -4.47694182e-01 4.26324785e-01 -1.78901517e+00
5.41824289e-02 -2.21817851e-01 8.66776943e-01 1.53918594e-01
-5.84554970e-01 -2.73963571e-01 1.68680176e-01 3.90767187e-01
-8.73105943e-01 5.53402305e-01 1.44673240e+00 3.20746936e-02
-5.66966236e-01 3.84315640e-01 -9.66243386e-01 3.98612410e-01
8.29577684e-01 -1.20651349e-01 -7.07894564e-01 -5.54361582e-01
2.37075448e-01 4.90691930e-01 3.95650119e-01 -6.97401941e-01
6.46820903e-01 -1.55313641e-01 -5.20955980e-01 -4.78795081e-01
5.12148082e-01 -4.38655615e-01 -3.40813845e-01 1.62406057e-01
-9.14193094e-01 3.66775095e-01 1.56285897e-01 4.55684215e-01
-2.16183171e-01 -5.12918472e-01 9.34344903e-02 -2.47650355e-01
-1.53994098e-01 -7.60414600e-02 -8.91951501e-01 6.26310110e-01
8.13668549e-01 -9.19953063e-02 -4.42463756e-01 -1.45991349e+00
-4.78892565e-01 7.46046126e-01 1.97365880e-01 5.16069233e-01
4.14492130e-01 -9.28917587e-01 -8.76758158e-01 -2.89893478e-01
2.25297481e-01 1.06334849e-03 1.55708969e-01 4.51847315e-01
-2.51257896e-01 8.61039579e-01 4.31592673e-01 -6.09890461e-01
-1.35037792e+00 1.74106985e-01 2.88785189e-01 -3.42407525e-01
8.34304467e-02 6.48687184e-01 1.68553561e-01 -7.57173181e-01
1.05898343e-01 -4.79420960e-01 -2.68408448e-01 3.51018049e-02
6.78212523e-01 3.23183239e-01 1.99273482e-01 -4.22303140e-01
-1.32797554e-01 1.63381904e-01 8.68645459e-02 -6.48496211e-01
5.44341385e-01 -5.55265069e-01 5.71984611e-02 2.41052836e-01
8.89176607e-01 2.16097474e-01 -6.45993054e-01 -3.67445797e-01
1.02987602e-01 -2.17148602e-01 -7.17696249e-01 -1.09303737e+00
-7.56147355e-02 6.90729141e-01 5.20942919e-02 7.40749240e-01
6.17728174e-01 2.32396305e-01 1.02901554e+00 1.06952488e+00
3.52833092e-01 -9.64891434e-01 3.45725983e-01 6.96915209e-01
1.41388929e+00 -1.58046699e+00 -5.63407898e-01 -5.25292099e-01
-9.06186998e-01 8.77985954e-01 7.59900510e-01 4.40011024e-01
1.15007423e-01 -1.95042431e-01 5.08359253e-01 -3.98095518e-01
-9.76403117e-01 -2.37249687e-01 2.77384371e-01 6.04803741e-01
4.45370317e-01 -6.92568049e-02 -3.39827418e-01 7.16061711e-01
-6.86868846e-01 -6.25041081e-03 7.04058468e-01 8.99906635e-01
-4.50187713e-01 -1.07671583e+00 -5.11899173e-01 1.32925466e-01
-9.76221729e-03 -1.70618206e-01 -1.25340176e+00 7.25745320e-01
-6.18743360e-01 1.94655657e+00 -2.98950613e-01 -1.47864833e-01
7.25943506e-01 2.59141117e-01 2.10572127e-02 -8.04790378e-01
-1.03506839e+00 -7.07411110e-01 6.71197176e-01 -3.78414690e-01
-3.73606592e-01 -2.68579334e-01 -9.14698660e-01 -1.61004171e-01
-4.55864191e-01 8.61934364e-01 6.20311856e-01 1.19231820e+00
7.14438140e-01 -1.24008477e-01 6.23705328e-01 -6.62561178e-01
-7.01385200e-01 -1.31504583e+00 2.66650885e-01 6.84869528e-01
2.50505775e-01 -3.92803580e-01 -4.40023482e-01 -1.63210128e-02] | [12.09553337097168, 7.888439655303955] |
6a07780e-163f-4c95-adf0-2f94d9f54490 | long-term-person-re-identification-a | 2105.14685 | null | https://arxiv.org/abs/2105.14685v4 | https://arxiv.org/pdf/2105.14685v4.pdf | DeepChange: A Large Long-Term Person Re-Identification Benchmark with Clothes Change | Existing person re-identification (re-id) works mostly consider short-term application scenarios without clothes change. In real-world, however, we often dress differently across space and time. To solve this contrast, a few recent attempts have been made on long-term re-id with clothes change. Currently, one of the most significant limitations in this field is the lack of a large realistic benchmark. In this work, we contribute a large, realistic long-term person re-identification benchmark, named as DeepChange. It has several unique characteristics: (1) Realistic and rich personal appearance (e.g., clothes and hair style) and variations: Highly diverse clothes change and styles, with varying reappearing gaps in time from minutes to seasons, different weather conditions (e.g., sunny, cloudy, windy, rainy, snowy, extremely cold) and events (e.g., working, leisure, daily activities). (2) Rich camera setups: Raw videos were recorded by 17 outdoor varying resolution cameras operating in a real-world surveillance system. (3) The currently largest number of (17) cameras, (1, 121) identities, and (178, 407) bounding boxes, over the longest time span (12 months). Further, we investigate multimodal fusion strategies for tackling the clothes change challenge. Extensive experiments show that our fusion models outperform a wide variety of state-of-the-art models on DeepChange. Our dataset and documents are available at https://github.com/PengBoXiangShang/deepchange. | ['Xiatian Zhu', 'Peng Xu'] | 2021-05-31 | null | null | null | null | ['person-identification'] | ['computer-vision'] | [-9.42335576e-02 -8.85557234e-01 9.39994454e-02 -4.28555608e-01
-6.20548986e-02 -6.32407427e-01 5.93044937e-01 -3.44454885e-01
-3.41442406e-01 8.35421324e-01 2.79614002e-01 4.62297022e-01
1.46121472e-01 -5.50564170e-01 -6.42079890e-01 -5.40716112e-01
4.38728631e-02 3.58521760e-01 -1.01314239e-01 -5.06679952e-01
-2.30567276e-01 3.64840299e-01 -1.66700280e+00 -4.31343801e-02
6.68456733e-01 7.70392120e-01 -6.58929497e-02 7.36872733e-01
4.42899615e-01 9.88051668e-02 -6.18592024e-01 -1.02937865e+00
4.67397481e-01 -2.04806060e-01 -3.95332187e-01 2.76988029e-01
9.41301942e-01 -6.59567237e-01 -5.74289143e-01 1.10986078e+00
7.53358364e-01 2.73288339e-01 2.70811677e-01 -1.44233215e+00
-1.04023933e+00 1.07087716e-01 -8.04550648e-01 7.98434615e-02
7.12673068e-01 3.23949456e-01 4.18186158e-01 -8.40694785e-01
4.76854295e-01 1.30721259e+00 1.07131362e+00 8.99280667e-01
-1.02037740e+00 -8.59770477e-01 5.58413088e-01 2.70846456e-01
-1.58329129e+00 -5.55317938e-01 6.02938235e-01 -2.23031208e-01
5.68006277e-01 5.90752244e-01 1.04099631e+00 1.85557151e+00
-1.67910337e-01 4.58442897e-01 1.42602360e+00 -2.27350816e-01
-1.19816087e-01 2.73890376e-01 2.32338339e-01 4.57715750e-01
3.96115124e-01 -1.29179582e-01 -3.68751854e-01 -2.87414849e-01
5.18686056e-01 6.19499922e-01 -3.54809970e-01 9.26287994e-02
-1.21756268e+00 2.69836903e-01 1.97206050e-01 3.77348103e-02
-4.76244614e-02 -9.32388157e-02 2.89264917e-01 2.89152414e-01
4.56613719e-01 -1.29489139e-01 -2.91348577e-01 -1.87412396e-01
-6.29069507e-01 5.78942239e-01 7.31909573e-01 1.12868834e+00
4.53853458e-01 -1.05688587e-01 -1.86483145e-01 1.19388783e+00
1.01813965e-01 1.13941693e+00 4.10759300e-01 -6.19247913e-01
3.72133911e-01 1.86839595e-01 4.40695643e-01 -1.16319013e+00
-4.68431890e-01 -1.38530374e-01 -1.17615068e+00 -1.70487046e-01
4.90928024e-01 -3.54025036e-01 -7.49995410e-01 1.87225866e+00
2.81243205e-01 5.08675337e-01 -1.26557037e-01 1.19817448e+00
9.95719612e-01 2.77474552e-01 3.17980051e-02 -7.00764656e-02
1.76027906e+00 -1.07610166e+00 -6.68038189e-01 -3.44414741e-01
-1.58695176e-01 -7.88054049e-01 9.66758072e-01 4.13862675e-01
-9.28982794e-01 -8.37724268e-01 -8.35460246e-01 2.02199996e-01
-6.70231879e-01 2.24093378e-01 4.49298501e-01 8.84632647e-01
-1.06441474e+00 3.77835095e-01 -3.47693592e-01 -1.10194528e+00
-5.45625500e-02 1.39382407e-01 -5.46735466e-01 -3.53340298e-01
-1.34177446e+00 7.22377300e-01 -7.83768818e-02 3.27522337e-01
-6.12542212e-01 -4.33059007e-01 -6.78501785e-01 -4.29027945e-01
3.56269181e-01 -8.60452235e-01 9.14398730e-01 -1.09234107e+00
-1.35453677e+00 1.12285411e+00 -2.68432856e-01 -1.32078871e-01
7.80473650e-01 -4.35324132e-01 -1.03264892e+00 -2.82852024e-01
-1.33925453e-01 4.37258512e-01 7.66522825e-01 -1.48025453e+00
-4.39751148e-01 -7.54913211e-01 2.08875313e-01 3.96928191e-01
-6.28747642e-01 3.53777677e-01 -8.26944768e-01 -9.03861046e-01
-5.32697558e-01 -1.44191432e+00 5.18861562e-02 -1.21776663e-01
-2.64109194e-01 7.09067136e-02 6.06846631e-01 -1.00312400e+00
1.17831635e+00 -2.14498448e+00 2.38252833e-01 -1.30488098e-01
8.23334306e-02 2.31820583e-01 1.23206384e-01 3.50369126e-01
-3.67925577e-02 1.12907015e-01 5.24610393e-02 -8.37671161e-01
2.32123002e-01 2.51431614e-01 -6.86815009e-02 5.73900819e-01
-5.11958778e-01 7.63604224e-01 -7.79409409e-01 -3.58263105e-01
2.95170218e-01 6.53812885e-01 7.55160898e-02 1.95678487e-01
4.28760678e-01 4.96694177e-01 1.01947658e-01 1.08853722e+00
9.86282229e-01 1.11777015e-01 5.47101609e-02 -3.59189838e-01
-4.82300669e-02 -5.49981534e-01 -1.37947083e+00 1.66415620e+00
-2.19955176e-01 5.08598208e-01 3.06978405e-01 -3.10204387e-01
7.01930523e-01 1.71012580e-01 4.08890098e-01 -4.82367218e-01
1.04200833e-01 -7.52485916e-02 -6.00932837e-01 -3.45076084e-01
7.05787838e-01 9.24836248e-02 -2.49813572e-01 1.07554488e-01
-2.00943768e-01 4.84544635e-01 2.71061897e-01 -5.73224649e-02
6.32293582e-01 5.55534139e-02 1.12100080e-01 -2.67303865e-02
3.28604221e-01 -3.88476878e-01 7.74527669e-01 7.01311350e-01
-7.61091173e-01 9.49190855e-01 -2.69598693e-01 -6.79561615e-01
-9.40211415e-01 -1.03179991e+00 1.56068001e-02 1.17311943e+00
6.19947016e-01 -3.35709244e-01 -7.82848001e-01 -2.90706068e-01
1.59184784e-01 1.24503680e-01 -7.16890514e-01 -8.46717693e-03
-5.54357290e-01 -1.17956913e+00 7.82135665e-01 5.31719506e-01
1.05823624e+00 -8.06806862e-01 -1.50544629e-01 -1.55543312e-02
-5.46612740e-01 -1.39632511e+00 -8.89711678e-01 -7.59813786e-01
-4.16018099e-01 -8.74039292e-01 -1.26549578e+00 -4.36850190e-01
5.43979526e-01 7.70822465e-01 1.32780075e+00 4.04542014e-02
-3.94100189e-01 8.34626675e-01 -3.94090116e-01 -3.35054368e-01
2.97062069e-01 -3.34228843e-01 6.62930191e-01 4.66847539e-01
4.60482329e-01 -2.60127783e-01 -8.99987519e-01 7.62465298e-01
-5.10540485e-01 -2.42387634e-02 1.42233938e-01 6.53760791e-01
3.23845744e-01 7.75572211e-02 2.98294663e-01 -5.33422649e-01
5.82464159e-01 -4.49884206e-01 -2.54009783e-01 6.82574809e-01
-4.67665941e-01 -8.96178305e-01 3.87167305e-01 -7.70568490e-01
-1.18509912e+00 -3.00290138e-01 3.87668312e-02 -3.92034352e-01
-6.01009011e-01 -1.70091882e-01 -2.95869023e-01 -6.71167597e-02
4.21743453e-01 1.88786343e-01 -1.94772974e-01 -6.43490016e-01
1.84789181e-01 8.07711899e-01 8.36282730e-01 -6.95611000e-01
9.13726926e-01 6.72151089e-01 -4.63016123e-01 -7.79078960e-01
-5.62212586e-01 -4.40801919e-01 -6.33104205e-01 -5.70150912e-01
8.94872606e-01 -1.33139467e+00 -7.87463546e-01 1.35010958e+00
-8.31115544e-01 -3.03213596e-01 2.00243771e-01 2.97785252e-01
7.38826320e-02 6.43972933e-01 -6.80765569e-01 -8.94577146e-01
-6.02763534e-01 -8.09832215e-01 9.35625196e-01 7.78208733e-01
-2.11588591e-01 -8.31332743e-01 -4.60486636e-02 7.13479221e-01
3.62203717e-01 6.44535601e-01 -8.11222941e-02 5.41612096e-02
-7.75792450e-02 -1.66569158e-01 -3.24407697e-01 1.72773704e-01
4.82000738e-01 8.10454786e-02 -9.86770868e-01 -6.98552072e-01
-4.28839654e-01 -1.78830504e-01 7.54714072e-01 1.42031774e-01
1.02199483e+00 -3.03057164e-01 -3.32987487e-01 8.86703432e-01
1.24369335e+00 4.87449318e-02 4.77173477e-01 3.44472229e-01
1.02757859e+00 5.88662446e-01 5.60188949e-01 7.94751167e-01
9.50033545e-01 9.63195443e-01 2.00440794e-01 -3.34244013e-01
-3.25172603e-01 1.26663717e-02 6.88328683e-01 5.01287341e-01
-7.51388490e-01 -3.50908756e-01 -6.43207610e-01 5.87640822e-01
-1.99229085e+00 -1.16107249e+00 -1.87928930e-01 2.42086744e+00
6.18736267e-01 -2.05472350e-01 6.24798834e-01 -3.16273183e-01
1.00776732e+00 1.15660913e-01 -6.88655317e-01 3.08768987e-03
-5.05128205e-01 -4.00571078e-01 6.46763325e-01 2.43008554e-01
-1.28405070e+00 7.69798696e-01 5.61770201e+00 4.00739402e-01
-8.86294246e-01 1.70607090e-01 4.28794742e-01 -3.71113628e-01
1.26331061e-01 -4.88726854e-01 -1.03334451e+00 1.01648152e+00
6.16924286e-01 1.58978432e-01 9.07050371e-01 5.49817741e-01
-7.50482306e-02 1.06798746e-01 -8.24159920e-01 1.64311635e+00
6.82231307e-01 -7.61388958e-01 -3.44880760e-01 -2.10831687e-02
7.62413085e-01 -1.42136052e-01 1.15852378e-01 3.26682240e-01
3.33819091e-01 -8.07764769e-01 7.99677134e-01 7.89734900e-01
1.03779101e+00 -6.77818894e-01 7.14665949e-01 -4.75321896e-03
-1.46297848e+00 -1.46122590e-01 -2.02286541e-01 -3.75772733e-03
2.33015925e-01 3.84276897e-01 6.44033030e-02 7.41818130e-01
1.52405560e+00 9.38066065e-01 -8.60342622e-01 7.78637528e-01
1.19734034e-01 2.36486182e-01 -3.52922082e-01 2.26877511e-01
-3.30450624e-01 -2.23170221e-01 4.77350265e-01 1.48333406e+00
3.52740854e-01 1.97820127e-01 4.22541648e-01 3.50148201e-01
-1.02885745e-01 -2.98918664e-01 -4.52930331e-01 3.96877736e-01
2.76978850e-01 1.34907281e+00 -4.23455715e-01 -4.38786030e-01
-5.96121848e-01 1.53325665e+00 -3.63752358e-02 7.02462256e-01
-1.23206818e+00 9.61059108e-02 1.18087053e+00 9.43861008e-02
-6.03749789e-02 -2.87934035e-01 -6.03437563e-03 -1.65078759e+00
3.00496489e-01 -1.05423820e+00 6.14172399e-01 -7.97722280e-01
-1.90718138e+00 5.90966046e-01 1.35612175e-01 -1.08911419e+00
1.91963121e-01 -5.38218081e-01 -4.87735033e-01 8.55533242e-01
-1.43143332e+00 -1.71564436e+00 -1.11585820e+00 1.03516304e+00
5.58708012e-01 -2.77175725e-01 8.76125336e-01 7.02120483e-01
-1.04752195e+00 9.27890956e-01 9.21582431e-02 2.61548638e-01
1.35195935e+00 -1.20331454e+00 5.32849967e-01 8.76397073e-01
-1.86938539e-01 7.38148332e-01 6.04383349e-01 -6.79377139e-01
-1.62235415e+00 -1.13056457e+00 5.97502410e-01 -6.31526232e-01
3.68232548e-01 -8.11746895e-01 -6.49739623e-01 7.18502700e-01
2.16757655e-01 1.55784085e-01 7.48433590e-01 3.56067479e-01
-4.88095880e-01 -5.00460148e-01 -1.24902582e+00 7.53288090e-01
1.36704266e+00 -3.08043122e-01 -1.77826628e-01 3.66447181e-01
2.17334718e-01 -5.68214059e-01 -1.03165197e+00 3.74407262e-01
1.18648136e+00 -9.30757701e-01 1.33033502e+00 -3.46705586e-01
-1.36140630e-01 -3.29456151e-01 -1.59028739e-01 -1.22940254e+00
-5.66122055e-01 -6.99917912e-01 -2.96462834e-01 1.59458375e+00
-1.84574232e-01 -8.00346196e-01 4.77871209e-01 1.05882514e+00
2.59189844e-01 -4.71185565e-01 -4.93190676e-01 -8.39486122e-01
-2.91527450e-01 4.93653156e-02 9.49244738e-01 1.01923239e+00
-5.05142808e-01 -1.37342885e-01 -1.17035317e+00 2.56106257e-01
7.72206485e-01 1.71089992e-01 1.19007897e+00 -1.22996712e+00
-2.58766711e-01 -2.20563382e-01 -2.37925559e-01 -8.19635093e-01
-7.08762649e-03 -2.84940064e-01 -3.35817307e-01 -1.34381318e+00
7.03971386e-01 -4.16288108e-01 -3.53321552e-01 5.81387997e-01
-4.32874888e-01 5.09829283e-01 5.29547215e-01 3.32323581e-01
-6.91391885e-01 3.23194861e-01 9.14633691e-01 -3.36153507e-01
2.38438081e-02 1.20829856e-02 -7.41620839e-01 7.54038632e-01
9.08711195e-01 5.21050990e-02 -1.27874821e-01 -6.50113881e-01
2.80524362e-02 -2.70691693e-01 8.19535434e-01 -1.13053083e+00
1.35176823e-01 -2.83983320e-01 7.13180959e-01 -2.81428695e-01
7.35071182e-01 -8.94240737e-01 6.61695421e-01 1.71160758e-01
2.07788840e-01 4.33732033e-01 1.30173430e-01 5.93083322e-01
1.00420594e-01 2.09728733e-01 6.29225016e-01 -3.11428785e-01
-1.10012841e+00 5.51737905e-01 7.92369917e-02 -9.26643014e-02
1.08954501e+00 -5.76566935e-01 -5.32530308e-01 -3.63214701e-01
-7.50642896e-01 2.63542712e-01 9.18983579e-01 9.47957158e-01
5.44746518e-01 -1.52678514e+00 -1.00794005e+00 7.02453032e-02
3.40131015e-01 -4.63095605e-01 6.91275597e-01 6.98411286e-01
-1.77072048e-01 -1.44835874e-01 -5.14973342e-01 -3.29009324e-01
-1.77192891e+00 4.72804368e-01 2.94963360e-01 -1.07087329e-01
-6.53432608e-01 6.86735868e-01 2.84593441e-02 -4.67884362e-01
3.01184714e-01 7.92039745e-03 -1.90155134e-01 8.21860954e-02
8.08528244e-01 7.13224888e-01 -2.13088214e-01 -1.08790815e+00
-5.75775802e-01 9.52936769e-01 5.50599061e-02 1.71526805e-01
1.06437910e+00 -5.04830956e-01 -6.22813888e-02 3.85678291e-01
7.60270596e-01 2.67868806e-02 -1.32650363e+00 -2.36710757e-01
-5.56684554e-01 -6.81405783e-01 -5.27731299e-01 -8.74423325e-01
-1.13599145e+00 3.53756338e-01 1.09287691e+00 9.84701812e-02
1.26019800e+00 -3.18630308e-01 1.26917887e+00 1.18861936e-01
6.74888372e-01 -1.30873299e+00 -1.43803686e-01 3.53355795e-01
9.75523055e-01 -1.57052743e+00 2.08659023e-01 -2.29194328e-01
-8.32509995e-01 7.21696615e-01 7.85141110e-01 2.46874616e-01
5.28688550e-01 6.31516799e-02 1.52545616e-01 8.28294009e-02
-1.88465104e-01 -2.45160043e-01 2.12971479e-01 7.94910550e-01
1.15445755e-01 4.57163841e-01 8.88475254e-02 7.32946873e-01
-2.58707374e-01 -2.02603862e-01 2.45096892e-01 6.30671322e-01
1.73003286e-01 -9.11263883e-01 -7.93437541e-01 3.57153267e-02
-4.38576490e-01 2.03405172e-01 -6.20904922e-01 6.34714425e-01
6.46243632e-01 1.23879874e+00 -1.07580729e-01 -4.95255709e-01
5.52176952e-01 -2.38124788e-01 5.68400204e-01 -1.10807866e-01
-7.36038268e-01 -2.72496879e-01 1.41100034e-01 -4.94804829e-01
-6.01713061e-01 -8.79914701e-01 -5.81776440e-01 -9.26587105e-01
5.31101506e-03 -4.43203866e-01 5.92407227e-01 5.79962552e-01
2.89787889e-01 3.05430859e-01 4.85255808e-01 -9.78845239e-01
-1.75036207e-01 -1.07744682e+00 -5.75027764e-01 1.04040432e+00
2.75662869e-01 -8.63917828e-01 -1.40740469e-01 3.38657618e-01] | [14.621301651000977, 0.9557335376739502] |
ce096816-cce9-4e84-bc81-6b22624df7b6 | clovacall-korean-goal-oriented-dialog-speech | 2004.09367 | null | https://arxiv.org/abs/2004.09367v2 | https://arxiv.org/pdf/2004.09367v2.pdf | ClovaCall: Korean Goal-Oriented Dialog Speech Corpus for Automatic Speech Recognition of Contact Centers | Automatic speech recognition (ASR) via call is essential for various applications, including AI for contact center (AICC) services. Despite the advancement of ASR, however, most publicly available call-based speech corpora such as Switchboard are old-fashioned. Also, most existing call corpora are in English and mainly focus on open domain dialog or general scenarios such as audiobooks. Here we introduce a new large-scale Korean call-based speech corpus under a goal-oriented dialog scenario from more than 11,000 people, i.e., ClovaCall corpus. ClovaCall includes approximately 60,000 pairs of a short sentence and its corresponding spoken utterance in a restaurant reservation domain. We validate the effectiveness of our dataset with intensive experiments using two standard ASR models. Furthermore, we release our ClovaCall dataset and baseline source codes to be available via https://github.com/ClovaAI/ClovaCall. | ['Sunghun Kim', 'Kyoungtae Doh', 'Sang-Woo Lee', 'Jung-Woo Ha', 'Soojin Kim', 'Hyunhoon Jung', 'Eunmi Kim', 'Kihyun Nam', 'Hyeji Kim', 'Sohee Yang', 'Nako Sung', 'Jingu Kang', 'Hyun Ah Kim', 'Chan Kyu Lee'] | 2020-04-20 | null | null | null | null | ['goal-oriented-dialog', 'open-domain-dialog'] | ['natural-language-processing', 'natural-language-processing'] | [-0.33612233 -0.07518198 0.04552832 -0.70384085 -1.3514099 -0.7051899
0.4340225 -0.22042365 -0.34681004 0.72759885 0.7571383 -0.5508018
0.48802057 -0.3182082 -0.10283002 -0.22762619 0.2867657 0.80350906
0.17289315 -0.86525464 -0.1639951 0.14195043 -0.73932403 0.39753973
0.70731044 0.6960069 0.4222375 0.91656977 -0.18305214 0.55793977
-1.0607868 -0.545275 -0.01538549 -0.32405695 -1.0626336 0.10978039
-0.06366516 -0.5217745 -0.7391209 0.6454791 0.97761154 0.5926541
0.2090083 -0.9664731 -0.79759043 1.0982065 0.20131633 0.15721448
0.71435124 0.19235523 1.2079176 -0.9332795 0.42052302 1.3594412
0.28657386 0.98126143 -0.8170834 -0.60282123 0.2745435 -0.05381203
-1.5187688 -1.2485533 0.7975469 0.06148891 1.1452578 0.6157751
0.248381 1.5248814 -0.6920584 1.2289109 0.8738897 -0.29602695
0.20505278 0.36173746 0.4310139 0.17650129 -0.4889584 -0.37959114
-0.5839192 -0.2215734 0.54364926 -0.12621933 -0.45660603 0.5701471
-1.2676166 0.8207806 -0.03790979 0.4136617 -0.13678572 -0.43392536
0.45571736 0.55213696 0.42959943 0.06479002 -0.6241386 -0.5895216
-0.39414883 0.11379939 1.2048541 1.5873939 0.3837572 0.29511893
-0.19081548 1.6979809 0.31832108 0.81629795 0.65119237 -0.7821841
0.8036374 0.25681686 0.03468755 -0.48092175 -0.22216867 0.07513656
-0.5987203 -0.7553393 0.26281136 -0.5772209 -0.35863343 1.4634786
0.14705501 -0.29469848 0.61494666 0.95249784 1.404609 0.8992784
-0.36994636 -0.19117905 1.417504 -1.3433307 -0.9376539 -0.4495707
0.66695017 -0.91177994 1.6391906 0.14534517 -0.96192 -0.24840003
-0.4743685 -0.19476712 -0.25329077 0.45387483 0.50739175 0.7529276
-1.0754337 -0.28843343 -0.6400542 -0.68844455 -0.25179702 0.09354714
-0.18934189 0.01442353 -1.3512704 0.40273884 0.0130395 0.1482385
-0.790848 -0.27815518 -0.8345899 -0.10120279 0.61811304 0.1382549
1.9437661 -0.5118509 -2.1413963 0.78277737 -0.42415485 -0.32726124
0.26980692 -0.4597794 -0.9060126 -0.1052753 -0.250988 0.42774916
0.2951844 -0.9687381 -0.46256402 -0.22382522 0.15980405 0.23044367
-0.25356904 0.6219401 -0.6359223 -0.7992674 -0.148853 -0.9829942
-0.12139263 -0.88626164 -0.6969567 -0.39952588 0.64987665 -0.9487374
1.4680742 -2.3578508 -0.29006466 -0.25624588 -0.11494324 0.2935552
-0.09093066 0.950944 0.2398236 0.02364605 -0.32076797 -0.55871177
0.09523852 0.14293122 -0.59339046 0.07276125 -0.13324028 0.7568909
-0.72696674 -0.22364688 0.21101785 0.20631582 -0.5120449 0.68069535
-0.11407141 0.6612799 -0.48163 0.86848366 0.3727598 0.02677855
0.20941351 0.23751362 -0.11874633 0.9402386 -0.7956848 1.8834809
-0.6685575 0.64202034 0.6249104 -0.7278425 1.2616686 0.8231073
0.05354933 -0.22393122 0.11376953 0.43529952 0.11819685 -0.19086093
0.8546078 0.05829256 -0.36035603 0.3609963 0.0529228 -0.34386674
-0.13621853 0.47048625 1.0734679 -0.62102354 0.29593292 0.01278703
0.68815404 -0.04558766 0.50388527 0.5777517 -0.6047633 0.7818984
0.02102502 0.2530712 -0.52782094 -0.9015045 -0.01278837 1.334825
-0.32825047 -0.4677916 -0.81350404 -0.651606 -0.4000803 0.93668807
0.33301383 0.2520005 -0.77230793 -0.30833802 0.8837903 0.40119088
0.6218123 -1.3850448 0.4109817 0.27411237 -0.581987 -1.6139045
-1.2146004 -0.16721924 -0.41482222 -0.6182213 -0.94752973 -0.95175624
0.15323025 0.52315944 1.0178766 -0.15362032 0.04902419 0.4848985
-0.7335191 -0.16958247 -0.7576204 0.36549324 0.2818298 0.02064891
0.5658035 -0.28370902 -0.23543873 0.72287184 -0.25039288 -0.27984056
0.28213042 0.70947 0.0137701 -0.51972306 1.242357 -0.7668562
1.0406728 -0.33920527 -0.34368098 0.336326 0.03405557 -0.5629225
0.6041799 -0.45073688 -1.3315823 -0.15626438 -0.90126294 -0.15185589
-0.5275456 0.5045573 -0.5147442 0.5547397 0.34897435 0.48690084
-0.16252029 -0.6735378 0.47921386 1.8349693 0.4796545 -0.5904526
0.29974657 -0.04217381 -1.2292432 -1.6008866 -0.5207667 -0.8757019
-0.31683224 -0.13357233 0.7681434 -1.0019633 -1.0646003 0.6422621
-1.2922778 -0.6634077 -0.05068933 0.72192836 -0.41690123 0.21107373
-0.9625714 -1.3361951 -0.38003564 -1.2149823 1.1087766 0.04195037
-0.38172352 -0.929329 0.03712004 0.88444513 0.3883501 -0.8365678
0.49306238 -1.3332225 -0.26844153 -0.26491612 0.15202619 0.48267686
0.3824103 -0.31262705 -1.1658449 -0.33518323 -0.01024112 -0.7573171
0.32891417 0.12489804 0.7476009 -0.4833544 -0.04496458 -0.01182705
0.5601659 0.69612217 0.33165205 -0.3379408 0.43490505 0.8137317
0.6132978 0.4218467 0.56581825 0.86745715 -0.29750082 0.30660188
0.02422447 -0.40605086 0.7635043 2.0055091 0.28825703 -0.5936965
-1.0032403 0.6444504 -1.7237108 -0.7455337 0.02942907 2.256188
1.1330315 -0.11522178 0.3811122 -0.31608766 0.8895391 0.33249176
-0.38127866 -0.20587566 -0.27240217 -0.2333162 0.09838434 0.9372493
-0.78057444 1.437437 6.0716047 0.9565793 -0.80464005 0.41176957
0.61175656 -0.02381745 -0.21359305 -0.12220866 -1.2005775 0.31269544
1.2017646 -0.15671806 0.6682174 0.9706569 0.35238522 0.21215683
-1.0352226 1.2207258 0.11266527 -1.0479548 -0.18927577 -0.15491344
0.19586776 0.26979294 -0.14050679 0.8335941 0.698393 -0.8353437
0.43288067 -0.07374507 0.8766471 -0.43279716 0.6306657 0.54752845
-1.168151 0.16028771 -0.2887891 0.1368105 0.6194401 0.11447947
-0.9959617 0.4210917 0.5339439 0.55932134 -0.1342864 0.3935791
-0.02374691 1.2552599 -0.3200297 -0.35974583 0.09164716 -0.33873102
0.78930664 1.3222778 0.11228399 0.57380825 0.39020824 0.4472956
-0.2530618 0.3731446 -0.6289318 -0.40116805 1.0077149 1.0738746
-0.33128247 -0.33098444 -0.6827272 1.1967319 0.22934897 0.5743418
-0.5700084 -0.56687856 0.81463724 -0.11101022 -0.17416136 -0.4096901
0.14827551 -1.3266337 -0.00966079 -1.296691 0.19134153 -0.63134617
-1.4472756 0.98849785 -0.22770648 -0.9100509 -0.36056378 -0.4462292
-0.6909258 0.85649514 -1.0957648 -0.92987704 -0.0713767 0.7440768
1.489707 -0.6853293 1.0216469 0.5554007 -0.85879076 0.848381
0.17797871 0.64206517 1.0137556 -1.0407909 0.7591498 0.35312203
0.05626683 0.8491278 0.38863257 -0.5894932 -1.5233468 -0.9271303
0.99635965 -0.6081823 0.8563947 -0.91373134 -0.9674975 1.0426925
0.41293287 -0.30559823 0.75052303 0.29474613 -0.06407401 -0.19775389
-0.92629474 0.8315435 1.1080561 -0.86977726 -0.58072215 0.43902552
1.3887786 -0.40943024 -0.75255215 -0.04478619 0.24946684 -0.4833979
0.6484641 -0.6116535 -0.12106224 0.18336086 -0.7254098 -1.4135871
0.18693624 -1.2586609 0.06895036 1.723922 0.6083039 -0.92401403
0.62089634 0.63231015 -0.647502 -0.3077234 -0.974803 -0.84111893
-0.16558708 -0.6424417 0.6739614 0.95814586 0.29581204 0.8639359
-0.5200761 0.12446777 0.03073361 -0.15852086 0.9468386 -0.6432017
-0.46499285 -0.21341416 0.13311736 -1.7727696 0.35018292 -0.86547256
0.3870262 -1.2664319 -0.26719737 -0.695661 0.23795006 0.3828657
-0.03507523 -0.3933391 0.13368315 0.2234826 -0.44990298 0.9780682
0.8322201 -0.11934092 -0.62333846 0.40551385 -0.6338778 0.5602236
1.0067158 -0.15662722 -0.37308928 -0.45473716 -0.48368937 0.6007613
-0.13782556 -0.5899265 0.2772529 -0.24787873 -0.58660734 -0.5909827
0.7850355 -0.50509465 -0.28382596 -0.02694571 -0.727352 -0.04925561
-0.03377336 0.39576057 -0.43845558 -0.14348632 0.4935996 0.01031857
-0.51260906 0.23814741 -0.6507682 0.4741135 0.5931769 0.20520508
-0.55108577 -1.0836585 -0.6169479 0.4621523 -0.05872792 0.7817007
0.7544256 -1.1428455 -0.84893644 0.13247323 0.4030886 -0.08327395
0.36656415 0.7545009 -0.16977671 0.82238126 0.53703576 -0.1750207
-1.4558419 0.1549806 0.13182431 0.21659057 -0.58583903 0.90239555
0.24495846 -1.1349746 0.65149504 -0.3809883 -0.05650264 -0.28916094
0.65220934 0.3966606 -0.08649738 -0.81167704 -0.47751492 -0.22145906
-0.16640843 -0.7362288 0.9046395 -0.5158733 0.15697305 0.9639247
1.059624 0.52710813 -0.80912787 -0.43323138 -0.16867891 -0.2582293
-0.27698794 -0.77626216 -0.90241224 0.8469719 0.14844032 0.19118194
0.65161735 0.35315403 1.2678474 1.0139619 0.47586685 -1.209764
0.02403979 0.96681035 1.2125224 -1.5012617 -0.7334461 -0.5642775
-1.3275335 0.5668203 0.6427798 0.42931324 0.76882905 0.11316057
0.54829866 0.16958585 -0.80519253 -0.23729448 -0.22360793 0.67251134
0.88926315 0.39898366 -0.17338108 1.1791565 -0.5840759 -0.41792133
0.5520437 0.8378582 -0.40061614 -1.0702734 -0.25481725 0.147725
-0.4372457 -0.4138101 -0.7787453 0.4915349 -0.9348657 1.7318262
-0.1119132 -0.4361685 0.45137265 0.1972567 -0.29252073 -1.0208364
-0.6978359 0.34085646 0.78134423 -0.15412003 -0.07605809 -0.53149986
-1.4241959 -0.40703487 -0.37216312 0.5732152 0.55131435 0.6334024
0.33781832 0.2098463 0.8393474 -0.34558007 -0.6831986 -1.4279135
-0.73984814 0.01880963 0.18137595 -0.1884963 -0.20891044 -0.08281154] | [14.18829345703125, 6.907529354095459] |
eb0b058d-989e-4f1b-a798-8f2ba5da3f21 | rethinking-the-evaluation-of-unbiased-scene | 2208.01909 | null | https://arxiv.org/abs/2208.01909v2 | https://arxiv.org/pdf/2208.01909v2.pdf | Rethinking the Evaluation of Unbiased Scene Graph Generation | Current Scene Graph Generation (SGG) methods tend to predict frequent predicate categories and fail to recognize rare ones due to the severe imbalanced distribution of predicates. To improve the robustness of SGG models on different predicate categories, recent research has focused on unbiased SGG and adopted mean Recall@K (mR@K) as the main evaluation metric. However, we discovered two overlooked issues about this de facto standard metric, which makes current unbiased SGG evaluation vulnerable and unfair: 1) mR@K neglects the correlations among predicates and unintentionally breaks category independence when ranking all the triplet predictions together regardless of the predicate categories. 2) mR@K neglects the compositional diversity of different predicates and assigns excessively high weights to some oversimple category samples with limited composable relation triplet types. In addition, we investigate the under-explored correlation between objects and predicates, which can serve as a simple but strong baseline for unbiased SGG. In this paper, we refine mR@K and propose two complementary evaluation metrics for unbiased SGG: Independent Mean Recall (MR) and weighted IMR (wIMR). These two metrics are designed by considering the category independence and diversity of composable relation triplets, respectively. We compare the proposed metrics with the de facto standard metrics through extensive experiments and discuss the solutions to evaluate unbiased SGG in a more trustworthy way. | ['Jun Xiao', 'Songyang Zhang', 'Shaoning Xiao', 'Jian Shao', 'Long Chen', 'Xingchen Li'] | 2022-08-03 | null | null | null | null | ['scene-graph-generation', 'unbiased-scene-graph-generation'] | ['computer-vision', 'computer-vision'] | [ 3.37124407e-01 4.41378474e-01 -4.19271648e-01 -4.42780584e-01
-6.21099234e-01 -5.64654827e-01 7.79357791e-01 2.31232300e-01
6.13022521e-02 8.70341003e-01 4.05293763e-01 -2.14046732e-01
-4.32397097e-01 -1.04620290e+00 -5.65637827e-01 -6.19294226e-01
8.44928175e-02 6.67779326e-01 5.02600610e-01 -1.89837635e-01
2.66894192e-01 3.03492963e-01 -2.00533414e+00 5.30063212e-01
1.14984548e+00 1.03325808e+00 -2.53799796e-01 1.78325772e-01
-2.77475655e-01 9.47574198e-01 -7.68364966e-01 -9.18811738e-01
1.72009259e-01 -4.13930148e-01 -1.00563240e+00 -8.96614492e-02
6.16250217e-01 1.83895305e-01 -1.39153540e-01 1.26260769e+00
3.39941204e-01 2.84151696e-02 8.54273140e-01 -1.58002257e+00
-8.33071291e-01 9.15180624e-01 -5.74577093e-01 1.28577426e-01
4.65534687e-01 -1.70328710e-02 1.51758206e+00 -8.60235214e-01
6.79688036e-01 1.37449217e+00 5.80598116e-01 3.77970636e-01
-9.63463187e-01 -5.18489838e-01 4.51776206e-01 4.93194371e-01
-1.57002795e+00 -1.36138260e-01 6.30998671e-01 -3.85581672e-01
6.37314618e-01 8.16468894e-01 3.88637394e-01 1.03609967e+00
-1.55088499e-01 9.13460612e-01 1.11991107e+00 -3.79788727e-01
7.79359639e-02 2.09819496e-01 5.54573417e-01 3.84637445e-01
9.56946969e-01 9.08751488e-02 -6.00248933e-01 -2.06165612e-01
2.80089289e-01 -3.26569140e-01 -3.46667469e-01 -4.15973425e-01
-1.37546980e+00 6.10983849e-01 4.82379496e-01 2.70744801e-01
-9.02210623e-02 -3.32770914e-01 2.97002345e-01 8.36397633e-02
4.05255586e-01 5.69535315e-01 -5.32825351e-01 1.44174248e-01
-5.15946090e-01 4.32415694e-01 8.09290648e-01 1.35083270e+00
8.47271383e-01 -2.78474450e-01 -8.66124272e-01 7.94129968e-01
1.66022569e-01 4.05613810e-01 2.53121525e-01 -5.28330863e-01
5.65480709e-01 1.00728273e+00 7.90754985e-03 -1.43053985e+00
-2.87489265e-01 -6.11904442e-01 -7.90472686e-01 -3.77209574e-01
4.96992379e-01 3.85825664e-01 -8.20042014e-01 1.77186418e+00
4.89397228e-01 -3.00539672e-01 -5.59967905e-02 9.19245839e-01
1.26628208e+00 9.57801938e-02 2.74631321e-01 -2.77063698e-01
1.28075755e+00 -1.03741908e+00 -5.61876833e-01 -1.20079905e-01
6.44096851e-01 -7.42390454e-01 1.32821321e+00 6.32489398e-02
-5.99264503e-01 -4.02018309e-01 -9.40542996e-01 8.37557390e-02
-5.74377716e-01 -1.39335304e-01 1.00925100e+00 7.96546459e-01
-6.36322558e-01 5.53298414e-01 -3.22852910e-01 -3.57889354e-01
3.80872428e-01 1.88361168e-01 -1.45791471e-01 -1.90535367e-01
-1.29172850e+00 8.09183657e-01 6.99323058e-01 -1.17814623e-01
-5.37895381e-01 -6.83177710e-01 -7.16071069e-01 4.83893715e-02
8.37946177e-01 -5.65547228e-01 8.82965088e-01 -6.47093296e-01
-9.51550603e-01 9.22971666e-01 -1.49757847e-01 -1.24015599e-01
5.62645257e-01 -9.82135162e-02 -7.82171071e-01 -1.78364381e-01
2.89983690e-01 3.70979488e-01 3.42814267e-01 -1.42397785e+00
-6.85161352e-01 -3.45950454e-01 2.20372126e-01 3.58541042e-01
-6.90897852e-02 -3.45828801e-01 -1.07354105e-01 -7.09441900e-01
6.09651983e-01 -5.88869035e-01 1.31346077e-01 -6.89495921e-01
-1.00401843e+00 -5.54732740e-01 4.83301401e-01 -1.56345621e-01
1.48622429e+00 -1.98484659e+00 -2.85521418e-01 2.64016002e-01
4.14463282e-01 1.62365586e-01 -9.99727994e-02 3.65219176e-01
-8.76777023e-02 2.48401642e-01 -1.30742952e-01 1.95232481e-01
1.52213916e-01 1.59016222e-01 -3.90542805e-01 1.50425389e-01
1.25803992e-01 9.84938145e-01 -1.35711026e+00 -6.63709044e-01
2.79099178e-02 6.04907945e-02 -3.09942782e-01 8.56666416e-02
-2.54110336e-01 1.34291425e-01 -4.11358565e-01 1.04922235e+00
8.29004407e-01 -3.50865304e-01 3.28681946e-01 -6.45956218e-01
2.92038679e-01 5.91253042e-01 -1.29287100e+00 8.44006121e-01
2.49461606e-01 8.28040093e-02 -8.98038805e-01 -9.02552009e-01
1.07351661e+00 -9.60895792e-02 3.31997037e-01 -5.21664381e-01
4.14457656e-02 3.89213562e-01 8.53030905e-02 -2.88706332e-01
7.49962449e-01 -1.35689169e-01 -2.28770226e-01 1.84730202e-01
9.21663828e-03 8.02690610e-02 3.79017472e-01 3.46233159e-01
1.12888694e+00 1.31935671e-01 6.31695092e-01 -4.99408215e-01
5.21808565e-01 1.01048134e-01 7.98264980e-01 9.22084451e-01
-1.93057761e-01 7.36947179e-01 6.71071887e-01 -4.04914439e-01
-6.13602638e-01 -1.19606566e+00 -8.32282379e-02 9.15223181e-01
8.03602695e-01 -5.81471741e-01 -3.54628354e-01 -1.25217724e+00
1.29602551e-02 8.80693078e-01 -5.45355916e-01 -4.35754240e-01
-2.33465210e-01 -1.16297591e+00 5.92090130e-01 3.74998182e-01
5.16262293e-01 -9.86120522e-01 -3.14414613e-02 9.24356375e-03
-5.40638566e-01 -1.16917610e+00 -2.10423037e-01 -1.37604550e-01
-4.97557372e-01 -1.55755222e+00 -2.39611313e-01 -5.34005702e-01
6.23526692e-01 5.11003792e-01 1.64295757e+00 5.45727350e-02
1.91587076e-01 -5.92122674e-02 -7.29970157e-01 -4.40952778e-01
-2.92226076e-01 1.09703526e-01 1.16837904e-01 1.31614774e-01
7.71795452e-01 -4.91152763e-01 -5.17254055e-01 5.51956713e-01
-7.19633937e-01 1.40113369e-01 5.71795523e-01 6.48389161e-01
8.95316720e-01 1.80404574e-01 5.06556511e-01 -1.52091205e+00
4.43457067e-01 -4.16365653e-01 -2.65065968e-01 6.64649248e-01
-8.88544559e-01 5.74150980e-02 4.30347353e-01 -5.00411749e-01
-9.80482161e-01 -5.53243876e-01 1.73015773e-01 -1.13252506e-01
1.07808234e-02 3.09164882e-01 -6.94151759e-01 9.87103805e-02
6.50632322e-01 5.06988540e-02 -5.95030725e-01 -3.04436386e-01
3.38787884e-01 5.10299921e-01 4.46544707e-01 -7.35068023e-01
7.00418949e-01 2.97440886e-01 9.56099406e-02 -4.36910003e-01
-1.32980776e+00 -4.25026923e-01 -3.04648131e-01 -9.86221582e-02
2.67135799e-01 -7.94708014e-01 -5.69499671e-01 3.48005295e-01
-7.52511859e-01 1.03170425e-01 -5.55074036e-01 3.26833516e-01
-3.34083110e-01 5.85383296e-01 -3.28245193e-01 -6.84731007e-01
-3.06251019e-01 -8.59831214e-01 1.07268584e+00 1.06450051e-01
-3.61565620e-01 -6.38458908e-01 -2.12948680e-01 2.40993917e-01
2.06490025e-01 5.78429103e-01 1.07852387e+00 -9.76745188e-01
-6.80672407e-01 -1.44644901e-01 -6.04838133e-01 1.72254637e-01
2.71967411e-01 7.41014024e-03 -9.03414488e-01 7.08460286e-02
-2.28071943e-01 -8.83194879e-02 8.58419955e-01 8.65763798e-02
1.10233402e+00 -4.59800720e-01 -5.13145745e-01 3.59440088e-01
1.44825399e+00 4.40842360e-02 7.25270927e-01 3.56407821e-01
1.02373505e+00 6.80610836e-01 1.13338614e+00 1.62255332e-01
7.09575355e-01 6.90185130e-01 4.08552527e-01 1.73035607e-01
-4.30975318e-01 -5.60494483e-01 4.93075363e-02 8.41293573e-01
-2.80048013e-01 -3.79114330e-01 -7.53628016e-01 5.98502874e-01
-1.87160110e+00 -9.16065395e-01 -5.68326414e-01 2.42350245e+00
8.49172056e-01 2.64739633e-01 3.82862333e-03 4.00507808e-01
9.19306993e-01 9.65864062e-02 -1.69609174e-01 -1.00398265e-01
-5.65518081e-01 1.32437060e-02 4.27199751e-01 2.52103508e-01
-1.11260366e+00 9.96106505e-01 6.09958792e+00 1.20515406e+00
-6.14613950e-01 2.22058501e-02 6.16997600e-01 2.26066694e-01
-6.45154417e-01 2.58879036e-01 -9.96785641e-01 6.40762866e-01
4.46782768e-01 -3.44600320e-01 2.22593173e-02 8.38406920e-01
-5.38812518e-01 -1.17193617e-01 -1.14604092e+00 8.59854281e-01
7.55434344e-03 -9.64031219e-01 4.88003522e-01 3.06269759e-03
8.90539348e-01 -3.33048701e-01 -1.42352819e-01 4.80059177e-01
4.91319388e-01 -7.47093141e-01 7.92881846e-01 4.21970874e-01
6.72869265e-01 -6.13844872e-01 9.44574177e-01 1.65593475e-01
-1.22979498e+00 1.77664116e-01 -4.46640462e-01 -1.28473014e-01
-1.31743386e-01 1.07386422e+00 -8.24188232e-01 9.97392178e-01
6.35446310e-01 5.14050126e-01 -9.56754088e-01 9.63095486e-01
-4.56834435e-01 5.17230392e-01 -2.60208368e-01 -1.78616494e-01
-7.87427090e-03 5.19477390e-02 6.54822648e-01 1.11709690e+00
1.94510981e-01 -1.59680080e-02 1.44670352e-01 6.09950423e-01
2.02058023e-03 2.25107610e-01 -5.45283496e-01 1.52688473e-01
7.71072805e-01 1.06737840e+00 -9.46977794e-01 -4.67086256e-01
-2.41230696e-01 4.22229797e-01 3.70294422e-01 1.78432196e-01
-8.56808364e-01 -8.10769424e-02 5.65621197e-01 1.00596428e-01
3.41563523e-02 3.57895881e-01 -6.03139997e-01 -1.40212440e+00
2.14957118e-01 -7.63177335e-01 7.73851931e-01 -4.47860152e-01
-1.56992459e+00 6.49546742e-01 3.56062531e-01 -1.49404907e+00
7.82287419e-02 -3.60614926e-01 -1.40000626e-01 5.08102953e-01
-1.45844293e+00 -1.18429542e+00 -4.94057745e-01 4.88596261e-01
1.90632582e-01 1.78594552e-02 6.07274055e-01 3.02761406e-01
-5.84772587e-01 9.73000944e-01 -4.11104560e-01 -1.54177755e-01
6.46990240e-01 -1.45851624e+00 3.42077076e-01 9.89548266e-01
2.24033102e-01 5.81429958e-01 9.16461229e-01 -7.80241847e-01
-8.71516705e-01 -1.26492989e+00 1.28485727e+00 -7.42111683e-01
4.26529378e-01 -9.52097401e-02 -9.20138061e-01 6.32599056e-01
-4.85891491e-01 -2.96842866e-02 6.68197572e-01 4.99843687e-01
-6.83467865e-01 -2.64543921e-01 -1.18865299e+00 6.85536563e-01
1.64009345e+00 -1.22171797e-01 -6.17798567e-01 5.61973393e-01
9.35582459e-01 -2.35814705e-01 -8.93968642e-01 1.18603516e+00
4.00921434e-01 -1.24293935e+00 8.23704541e-01 -4.14847821e-01
3.19778651e-01 -5.98325610e-01 -4.74593908e-01 -9.50759113e-01
-4.31103230e-01 -1.98178157e-01 -5.26618838e-01 1.61933410e+00
5.55466652e-01 -7.36584663e-01 7.51369238e-01 5.51831245e-01
-7.29055330e-02 -9.37298477e-01 -5.35039246e-01 -1.00747108e+00
-3.41042310e-01 -3.89544964e-01 1.16841578e+00 1.02562261e+00
-2.22370178e-02 3.45489621e-01 -2.74021506e-01 2.29528144e-01
7.12291121e-01 4.60134029e-01 7.25923002e-01 -1.20859909e+00
-1.79399803e-01 -4.52499568e-01 -6.37799680e-01 -6.29734695e-01
-3.33566517e-01 -8.75500143e-01 -6.72289431e-02 -1.37118399e+00
5.56056976e-01 -6.69747651e-01 -4.00140077e-01 4.61696804e-01
-6.45164669e-01 4.36886877e-01 8.22461247e-02 3.76929641e-01
-8.19095433e-01 4.41200048e-01 1.38778234e+00 -9.24881324e-02
-2.03297995e-02 -2.05480140e-02 -1.30741012e+00 7.62143254e-01
5.76492965e-01 -3.41622919e-01 -5.63948989e-01 -1.01909317e-01
3.24790776e-01 -5.97334266e-01 2.25498945e-01 -1.00493371e+00
-5.16829416e-02 -3.42357934e-01 1.51320145e-01 -7.24916995e-01
-6.46281242e-02 -6.26022696e-01 3.53433371e-01 1.13364846e-01
-8.71648863e-02 -2.78977752e-01 -3.78178269e-01 5.26969731e-01
-3.90863895e-01 1.02464207e-01 4.11874413e-01 -2.40338385e-01
-9.27394748e-01 2.74048597e-01 3.84532899e-01 2.46958181e-01
9.00203705e-01 -4.73859280e-01 -7.49886394e-01 -3.19736451e-02
-4.80133832e-01 8.40685740e-02 5.61998188e-01 4.93824244e-01
4.64352667e-01 -1.52518749e+00 -6.74353421e-01 5.39319851e-02
6.35380745e-01 9.02623907e-02 1.45175904e-01 8.38093877e-01
-3.62525821e-01 2.94998229e-01 1.66297093e-01 -5.14099479e-01
-1.14047742e+00 7.66846478e-01 1.29217401e-01 -6.54933214e-01
-3.97674263e-01 9.74748135e-01 4.42199916e-01 -4.63629127e-01
1.63017005e-01 -2.51150727e-01 -4.90634054e-01 9.27171633e-02
3.53801608e-01 5.11768818e-01 2.82801777e-01 -7.50150502e-01
-6.02271140e-01 3.70747685e-01 -1.47086471e-01 5.13310969e-01
9.04732525e-01 -8.45777169e-02 -1.64009482e-01 3.94817591e-01
8.13955545e-01 7.30722696e-02 -7.05576837e-01 -2.59362668e-01
3.38551790e-01 -6.74320161e-01 -5.77573299e-01 -9.36933696e-01
-8.62527013e-01 3.19029510e-01 1.05005808e-01 6.07584238e-01
1.22978592e+00 1.86366081e-01 4.92348582e-01 3.37290093e-02
8.52071941e-01 -8.48770499e-01 -1.69673204e-01 3.98324400e-01
6.32697225e-01 -1.13753259e+00 3.50450099e-01 -1.35775101e+00
-8.33698869e-01 6.29123926e-01 8.77520919e-01 1.95752919e-01
3.83430988e-01 -1.83701128e-01 -1.22437939e-01 -4.10992175e-01
-5.39037645e-01 -5.58328629e-01 6.21566415e-01 7.51622021e-01
4.37458992e-01 4.43278641e-01 -8.52294564e-01 6.56404555e-01
-5.75207293e-01 -2.76223421e-01 3.34297240e-01 5.66698730e-01
-2.35126063e-01 -1.08523095e+00 -1.14917137e-01 7.90253937e-01
-3.95316690e-01 -1.37588173e-01 -5.65724552e-01 8.68226111e-01
4.67563480e-01 9.88255918e-01 -3.51847649e-01 -7.39357352e-01
5.62246442e-01 -2.28764247e-02 4.84680116e-01 -6.24149323e-01
-3.26813072e-01 -4.76565450e-01 4.37908202e-01 -5.90994835e-01
-7.36066997e-01 -5.39979756e-01 -7.69129217e-01 -3.14341962e-01
-6.58846200e-01 2.15813041e-01 1.92548800e-02 9.83528733e-01
3.59126329e-01 4.47337627e-01 4.96140391e-01 -2.57441878e-01
-4.22847629e-01 -1.05761218e+00 -7.05235243e-01 8.90455961e-01
-1.67994022e-01 -1.10653722e+00 -4.60842639e-01 -3.55129987e-01] | [10.299129486083984, 1.7773667573928833] |
33b702e9-958c-4dde-9f22-3f75907a84d5 | an-encoder-decoder-based-audio-captioning | 2108.02752 | null | https://arxiv.org/abs/2108.02752v1 | https://arxiv.org/pdf/2108.02752v1.pdf | An Encoder-Decoder Based Audio Captioning System With Transfer and Reinforcement Learning | Automated audio captioning aims to use natural language to describe the content of audio data. This paper presents an audio captioning system with an encoder-decoder architecture, where the decoder predicts words based on audio features extracted by the encoder. To improve the proposed system, transfer learning from either an upstream audio-related task or a large in-domain dataset is introduced to mitigate the problem induced by data scarcity. Besides, evaluation metrics are incorporated into the optimization of the model with reinforcement learning, which helps address the problem of ``exposure bias'' induced by ``teacher forcing'' training strategy and the mismatch between the evaluation metrics and the loss function. The resulting system was ranked 3rd in DCASE 2021 Task 6. Ablation studies are carried out to investigate how much each element in the proposed system can contribute to final performance. The results show that the proposed techniques significantly improve the scores of the evaluation metrics, however, reinforcement learning may impact adversely on the quality of the generated captions. | ['Wenwu Wang', 'Mark D. Plumbley', 'Xi Shao', 'H Lilian Tang', 'Tom Ko', 'Shengchen Li', 'Jinzheng Zhao', 'Yusong Wu', 'Jingqian Wu', 'Gengyun Chen', 'Xubo Liu', 'Qiushi Huang', 'Xinhao Mei'] | 2021-08-05 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 4.75914419e-01 4.98915076e-01 1.20389633e-01 -3.86415303e-01
-1.42528725e+00 -2.16396376e-01 3.08227897e-01 4.32652831e-02
-3.70388985e-01 7.71426082e-01 7.04862833e-01 1.91970468e-01
1.30755156e-01 -3.88530910e-01 -8.84649754e-01 -4.97139215e-01
-4.07550372e-02 1.65346652e-01 6.98871762e-02 -1.02894366e-01
9.69918147e-02 -3.72809827e-01 -1.54401803e+00 6.37175858e-01
6.93940282e-01 1.25252569e+00 4.13212001e-01 7.14519680e-01
9.89950374e-02 9.57651138e-01 -8.34766984e-01 -4.05723989e-01
-2.40326277e-03 -6.73302174e-01 -7.59094715e-01 -1.95657149e-01
3.46983254e-01 -2.98916548e-01 -2.61179000e-01 8.37805092e-01
9.19859767e-01 -1.43981948e-01 5.41598976e-01 -1.33770025e+00
-4.07836795e-01 9.41450000e-01 -2.32681394e-01 8.94695967e-02
4.21291530e-01 5.91962934e-02 1.23103583e+00 -7.64443874e-01
8.07057843e-02 1.00023496e+00 5.07554710e-01 5.68801284e-01
-9.20513570e-01 -7.26149142e-01 -2.27981564e-02 4.87116039e-01
-1.45931673e+00 -7.52249599e-01 9.75979924e-01 -4.71769422e-01
6.78150773e-01 1.06714956e-01 2.51486897e-01 1.17337584e+00
-1.69651225e-01 7.51140714e-01 6.55187011e-01 -4.95940030e-01
1.86637327e-01 4.62282777e-01 -3.41808856e-01 1.93348944e-01
-3.41761410e-01 3.28923017e-01 -8.92630398e-01 -2.56005842e-02
1.87158674e-01 -9.28495944e-01 -2.83326209e-01 -9.56881121e-02
-8.90885711e-01 6.96636379e-01 4.98054892e-01 7.20358491e-02
-3.95933509e-01 3.15665483e-01 6.94335639e-01 3.18882585e-01
4.33485091e-01 8.45553398e-01 -3.13078821e-01 -4.94885325e-01
-9.10237908e-01 1.70690268e-01 3.34441602e-01 8.25334966e-01
3.79017442e-01 2.27169901e-01 -5.91549754e-01 1.02412462e+00
2.91184455e-01 3.27328324e-01 5.55035651e-01 -8.00354540e-01
8.82479966e-01 1.25065088e-01 1.44673124e-01 -7.28295803e-01
-2.27799714e-01 -5.92986286e-01 -4.22875881e-01 -2.13866547e-01
1.85643345e-01 -5.53723812e-01 -7.19177008e-01 2.11737561e+00
-2.19506696e-02 3.03063899e-01 2.23540843e-01 1.08236003e+00
7.52419651e-01 9.02453661e-01 2.62478203e-01 -2.24357724e-01
1.20745242e+00 -9.45228040e-01 -9.26044941e-01 -3.55531305e-01
5.42420328e-01 -9.25240040e-01 1.11600637e+00 2.07156196e-01
-1.31625068e+00 -6.98640108e-01 -1.11508846e+00 2.58037359e-01
2.16836363e-01 2.60359406e-01 4.34469953e-02 5.87740481e-01
-8.14992189e-01 2.67351389e-01 -3.00100029e-01 1.22815976e-02
1.66090593e-01 3.40789407e-01 6.65841997e-02 3.24030370e-01
-1.78940511e+00 7.15091348e-01 4.04789835e-01 5.95504120e-02
-1.10323834e+00 -7.92888403e-01 -5.84128201e-01 4.23791677e-01
9.48565528e-02 -3.84856045e-01 1.63140202e+00 -1.21711636e+00
-1.53655398e+00 3.95521700e-01 1.54314280e-01 -7.19226062e-01
4.47836608e-01 -4.69139785e-01 -4.21744853e-01 2.59960562e-01
-5.37026078e-02 1.04910100e+00 8.51106346e-01 -9.73202705e-01
-7.41952479e-01 1.37039542e-01 -6.93521053e-02 5.60717523e-01
-5.22101462e-01 -1.16731159e-01 -2.25134760e-01 -8.07186782e-01
-5.06549716e-01 -8.90079439e-01 1.17823839e-01 -3.48531812e-01
-2.77572662e-01 -8.51956159e-02 5.68544507e-01 -8.80505204e-01
1.37291181e+00 -2.52626014e+00 -6.66687861e-02 5.94616234e-02
-4.88441080e-01 2.02629432e-01 -3.04618359e-01 4.24921006e-01
-1.07622489e-01 3.10287941e-02 -2.39852995e-01 -9.61897746e-02
-1.70732755e-02 -2.32357457e-01 -4.67116773e-01 8.35820213e-02
4.93268341e-01 4.07492995e-01 -8.47399652e-01 -3.67589206e-01
-2.13060737e-01 5.18468738e-01 -7.79040217e-01 6.94587946e-01
-3.96660507e-01 3.21656168e-01 -2.57898569e-01 1.34408191e-01
1.92003533e-01 1.16933517e-01 -2.03311130e-01 -1.34518325e-01
2.04308912e-01 7.05510139e-01 -8.45157206e-01 1.71492445e+00
-6.80131733e-01 6.67525887e-01 -1.37975952e-02 -1.01993752e+00
9.59510803e-01 9.71816063e-01 4.16196316e-01 -8.41977954e-01
1.13509223e-01 1.00425012e-01 1.89901620e-01 -6.17362678e-01
3.11522692e-01 -1.85541019e-01 -1.86116531e-01 3.01723301e-01
1.06990002e-01 -6.95152730e-02 -1.76233724e-01 8.03033909e-05
9.84544456e-01 -2.21598782e-02 -9.19144154e-02 1.32999942e-01
6.19555295e-01 -1.14353277e-01 5.12713552e-01 5.70911884e-01
-2.91367650e-01 9.42941189e-01 4.58965480e-01 1.28258675e-01
-1.02145112e+00 -9.19205666e-01 3.02143376e-02 1.26792610e+00
-1.49471670e-01 -2.39369571e-01 -8.69786799e-01 -4.50330973e-01
-4.23027337e-01 9.20992494e-01 -4.81491685e-01 -6.52069986e-01
-2.62350559e-01 -5.06370187e-01 8.44689786e-01 5.70486903e-01
3.46067727e-01 -1.19671977e+00 -4.72018778e-01 4.45894927e-01
-7.23123133e-01 -1.25653350e+00 -6.53878033e-01 1.03365943e-01
-4.93779242e-01 -5.06806791e-01 -8.44875455e-01 -9.71337974e-01
3.43340039e-01 -1.06411792e-01 8.42627823e-01 -4.05890048e-01
1.65310130e-01 1.48697793e-01 -7.18675971e-01 -6.15628242e-01
-6.01379216e-01 3.90008450e-01 3.14393453e-02 2.47254252e-01
2.38244280e-01 -6.16773903e-01 -5.26289165e-01 2.74058253e-01
-6.70197606e-01 1.53109863e-01 7.60429919e-01 8.71738374e-01
2.19157264e-01 -1.05046987e-04 1.30475771e+00 -5.62035382e-01
9.79825795e-01 -4.71867561e-01 -2.65611559e-01 5.57324092e-05
-6.25976801e-01 2.95877486e-01 6.65576696e-01 -5.65598547e-01
-1.09230196e+00 1.13301173e-01 -2.93286979e-01 -3.68959397e-01
-1.81293022e-02 5.15552282e-01 -4.04912978e-01 4.93371516e-01
7.27543354e-01 2.08388731e-01 -1.86287407e-02 -3.37821245e-01
2.17002034e-01 1.16834474e+00 6.00075662e-01 -3.87935460e-01
6.31185770e-01 -1.47959828e-01 -5.10004103e-01 -5.15906155e-01
-8.84971559e-01 -3.60218912e-01 4.30381782e-02 -3.28775167e-01
8.42002690e-01 -1.30218160e+00 -3.46570194e-01 1.02635622e-01
-1.26569724e+00 -3.18169780e-02 -3.40345770e-01 6.18314028e-01
-9.88083363e-01 -1.94861427e-01 -4.23581988e-01 -8.98659885e-01
-5.23438871e-01 -1.10097075e+00 9.31251287e-01 1.49272338e-01
-5.56615710e-01 -4.43806261e-01 2.33186930e-01 5.39498568e-01
4.31447953e-01 -1.82691664e-01 9.52715933e-01 -8.44286263e-01
-1.92887068e-01 -2.01319531e-01 -1.45299137e-01 6.76982403e-01
3.30908038e-02 -4.41455513e-01 -1.42060697e+00 -2.19919413e-01
-1.30574569e-01 -5.86327195e-01 3.64302874e-01 1.86548173e-01
1.13361084e+00 -5.65754592e-01 1.55022323e-01 1.32400528e-01
1.11733460e+00 3.97315115e-01 6.82731688e-01 2.47390732e-01
3.59896600e-01 7.69537687e-01 7.49641657e-01 4.35169250e-01
3.81863490e-02 9.28498089e-01 4.17887330e-01 7.36810565e-02
-2.91796416e-01 -7.23375857e-01 8.50767374e-01 1.04231572e+00
4.42009509e-01 -2.67938256e-01 -8.27394187e-01 7.98807323e-01
-1.62862706e+00 -8.54225814e-01 2.82724679e-01 2.16806674e+00
1.03713012e+00 4.29716796e-01 1.63935766e-01 4.31830913e-01
8.99814725e-01 -4.05949727e-02 -2.82145232e-01 -6.63792729e-01
8.94717574e-02 2.75898213e-03 1.58432439e-01 5.00405073e-01
-7.93766439e-01 6.95733190e-01 6.00291586e+00 7.50316381e-01
-1.21129990e+00 7.95597583e-02 4.20311838e-01 -2.52691031e-01
-1.27578750e-01 -1.38400346e-01 -4.25720215e-01 7.11898446e-01
1.46798348e+00 -4.28291708e-01 2.91126013e-01 6.02079928e-01
6.18107498e-01 2.65744060e-01 -1.21258020e+00 8.96743834e-01
8.42600837e-02 -9.46112514e-01 1.24418870e-01 -1.73634186e-01
4.17196602e-01 -1.49292797e-01 3.62111986e-01 5.22485197e-01
-2.43409634e-01 -8.28337550e-01 8.85634482e-01 3.26031893e-01
8.11265826e-01 -9.07721817e-01 9.03752685e-01 1.88532710e-01
-7.75172949e-01 -2.05725402e-01 -1.31023392e-01 -5.88895828e-02
2.74038285e-01 4.03021216e-01 -1.55797136e+00 2.56791711e-01
4.88564253e-01 1.62973121e-01 -4.34490472e-01 1.28448844e+00
-2.64666855e-01 1.16992974e+00 -1.15976758e-01 -1.41495481e-01
1.83902934e-01 3.56248558e-01 6.33027911e-01 1.22723174e+00
3.98756802e-01 -2.43012801e-01 -2.18388960e-01 5.81413150e-01
-2.67390162e-01 3.28904420e-01 -4.66556579e-01 -1.47988945e-01
4.77159292e-01 7.33510673e-01 5.64483590e-02 5.66709414e-03
-3.18609267e-01 6.33704305e-01 2.52048001e-02 2.94670343e-01
-9.80329633e-01 -6.09798193e-01 3.70893180e-01 4.46873367e-01
2.83619434e-01 4.27283019e-01 -1.69364884e-01 -6.29309297e-01
9.81634557e-02 -9.28335369e-01 2.03340322e-01 -1.09866154e+00
-1.07384992e+00 7.31699586e-01 -4.18569483e-02 -1.49780381e+00
-6.73253655e-01 -2.36205403e-02 -4.82497334e-01 7.63783991e-01
-1.47627604e+00 -5.58244228e-01 8.33707377e-02 2.96464682e-01
7.27539957e-01 -3.91426533e-01 7.52274036e-01 6.82002842e-01
-3.72471303e-01 1.03328800e+00 -4.58138026e-02 6.18117675e-02
9.71558809e-01 -9.76676106e-01 7.82710388e-02 6.12520874e-01
1.43253371e-01 -1.19686626e-01 1.09693551e+00 -1.65567800e-01
-6.23133600e-01 -1.27467465e+00 9.99790788e-01 -1.16398707e-01
4.86218214e-01 -4.98354852e-01 -7.53045321e-01 1.90676033e-01
3.95891100e-01 -3.46179843e-01 9.92866516e-01 1.16064094e-01
-1.64907455e-01 -4.66339409e-01 -9.45837855e-01 3.46228093e-01
5.81840754e-01 -7.41852164e-01 -6.43172920e-01 1.41418308e-01
1.03069878e+00 -1.45280838e-01 -7.11639583e-01 4.69650656e-01
4.47537929e-01 -4.11774516e-01 6.18391216e-01 -6.94114625e-01
7.85516143e-01 -7.46408775e-02 -2.11286381e-01 -1.56501043e+00
-1.96235031e-01 -6.45225227e-01 -1.52796805e-01 1.62653089e+00
9.40469325e-01 9.55732539e-02 7.55106270e-01 2.06480607e-01
-2.09236696e-01 -4.94568408e-01 -1.16197646e+00 -6.75448596e-01
4.46980260e-02 -4.34275866e-01 4.37027425e-01 4.42506015e-01
2.48846710e-01 8.94951701e-01 -7.45367587e-01 2.83696920e-01
2.81530797e-01 -3.57774347e-01 4.52914625e-01 -8.77266169e-01
-3.44365627e-01 -6.45598173e-02 -3.09648931e-01 -9.14787531e-01
6.54503405e-02 -7.38178968e-01 6.43781602e-01 -1.11326575e+00
-1.63500402e-02 -3.46619338e-01 -6.13528550e-01 2.92096794e-01
-1.07996225e-01 1.00678198e-01 3.40457022e-01 -1.24010347e-01
-6.37201369e-01 9.97799695e-01 1.04821825e+00 -2.88752913e-01
-2.97579914e-01 2.47589871e-01 -7.82711804e-01 3.80825460e-01
9.16934907e-01 -7.37185478e-01 -7.07381248e-01 -5.90299547e-01
2.44117513e-01 3.89953762e-01 7.80947953e-02 -1.24903488e+00
2.14935064e-01 1.83034837e-01 1.47382431e-02 -1.92289919e-01
5.11800647e-01 -9.07221138e-01 -1.31105378e-01 1.63022622e-01
-9.67048943e-01 -2.10664934e-03 3.50394666e-01 4.84480023e-01
-6.75580323e-01 -3.63319159e-01 8.20187926e-01 3.07913840e-01
-1.49160191e-01 -1.59132510e-01 -5.22784472e-01 2.43440643e-01
7.12804496e-01 1.78270057e-01 1.37048354e-02 -9.40616727e-01
-7.85269380e-01 2.57357448e-01 -1.93977565e-01 7.52208829e-01
4.10496771e-01 -1.59957707e+00 -1.04850352e+00 -1.07923128e-01
3.20328802e-01 -2.13926941e-01 1.11666143e-01 3.88739407e-01
-1.65933281e-01 6.28705263e-01 -2.04790026e-01 -4.74823445e-01
-9.38241720e-01 2.68496871e-01 2.84629434e-01 -2.52827942e-01
-1.95515096e-01 8.75749648e-01 1.00029483e-01 -3.17435712e-02
8.38192940e-01 -4.31448333e-02 -2.36581475e-01 1.83545485e-01
4.81652439e-01 1.84799343e-01 2.83312738e-01 -5.63147187e-01
-1.72679856e-01 -5.48637398e-02 -2.05134004e-01 -5.91470718e-01
1.20855522e+00 -2.82078832e-01 6.63936496e-01 4.17278558e-01
1.26509023e+00 -1.96309149e-01 -1.26211655e+00 -2.43265063e-01
-2.66209897e-02 -1.40895650e-01 3.13146889e-01 -1.06947207e+00
-9.52675223e-01 1.15126467e+00 9.08400655e-01 2.05875398e-03
1.39971566e+00 -2.37147152e-01 9.33828056e-01 4.99884635e-02
-1.96080897e-02 -1.27268124e+00 2.78284431e-01 3.82360697e-01
1.03700161e+00 -1.13400602e+00 -4.69009250e-01 -1.41142562e-01
-9.02092576e-01 7.31014371e-01 7.62708306e-01 1.53334662e-01
3.05479646e-01 1.20089412e-01 2.23835409e-01 3.22956532e-01
-1.00249887e+00 -3.56873013e-02 2.53768742e-01 5.21173179e-01
4.92429376e-01 -1.79741845e-01 -3.19176853e-01 9.22294080e-01
-5.64653456e-01 2.15328243e-02 6.19160116e-01 5.78611195e-01
-3.72006774e-01 -1.06514549e+00 -3.63647074e-01 2.56270200e-01
-6.52264833e-01 -9.61953327e-02 -4.25002903e-01 2.28736788e-01
1.42481789e-01 1.03772140e+00 1.46262020e-01 -6.06737494e-01
5.45517266e-01 3.06385279e-01 5.89602720e-03 -6.31455004e-01
-5.74890435e-01 1.65794000e-01 2.49245033e-01 -1.33172080e-01
-2.36275852e-01 -4.74424005e-01 -1.03978467e+00 3.80183846e-01
-4.58316773e-01 7.57807374e-01 6.05692267e-01 7.55440414e-01
4.80260402e-01 7.73782432e-01 1.01455903e+00 -3.08087587e-01
-7.69820035e-01 -1.22746408e+00 -2.75343299e-01 4.32623684e-01
5.18198967e-01 -4.14581895e-01 -4.42531317e-01 2.44406462e-01] | [15.268620491027832, 4.900907039642334] |
fbddfd8e-3e4d-4579-b10b-53bdc9438993 | retrieve-and-refine-exemplar-based-neural | 2010.04459 | null | https://arxiv.org/abs/2010.04459v1 | https://arxiv.org/pdf/2010.04459v1.pdf | Retrieve and Refine: Exemplar-based Neural Comment Generation | Code comment generation which aims to automatically generate natural language descriptions for source code, is a crucial task in the field of automatic software development. Traditional comment generation methods use manually-crafted templates or information retrieval (IR) techniques to generate summaries for source code. In recent years, neural network-based methods which leveraged acclaimed encoder-decoder deep learning framework to learn comment generation patterns from a large-scale parallel code corpus, have achieved impressive results. However, these emerging methods only take code-related information as input. Software reuse is common in the process of software development, meaning that comments of similar code snippets are helpful for comment generation. Inspired by the IR-based and template-based approaches, in this paper, we propose a neural comment generation approach where we use the existing comments of similar code snippets as exemplars to guide comment generation. Specifically, given a piece of code, we first use an IR technique to retrieve a similar code snippet and treat its comment as an exemplar. Then we design a novel seq2seq neural network that takes the given code, its AST, its similar code, and its exemplar as input, and leverages the information from the exemplar to assist in the target comment generation based on the semantic similarity between the source code and the similar code. We evaluate our approach on a large-scale Java corpus, which contains about 2M samples, and experimental results demonstrate that our model outperforms the state-of-the-art methods by a substantial margin. | ['Zhi Jin', 'Xin Xia', 'Ge Li', 'Yongmin Li', 'Bolin Wei'] | 2020-10-09 | null | null | null | null | ['code-comment-generation', 'comment-generation'] | ['computer-code', 'natural-language-processing'] | [ 4.22595114e-01 2.50997603e-01 -2.43101522e-01 -2.94357359e-01
-1.03361368e+00 -4.50360805e-01 5.32736123e-01 3.32250088e-01
1.88047767e-01 3.84951204e-01 6.66739225e-01 -4.23637599e-01
3.39047402e-01 -7.18845308e-01 -6.84907794e-01 -1.48550808e-01
1.20771199e-01 -3.23667843e-03 -2.81314994e-03 -3.84059966e-01
8.85500908e-01 -5.15035033e-01 -1.47516716e+00 7.24287391e-01
1.41971719e+00 5.24138510e-01 4.72980618e-01 5.77882648e-01
-7.09442019e-01 1.12768018e+00 -8.74554813e-01 -6.44673407e-01
4.21315664e-03 -8.53208363e-01 -8.43419731e-01 -2.96637833e-01
2.17955902e-01 -2.04396456e-01 6.51213378e-02 1.06871128e+00
5.46202898e-01 -7.70235881e-02 7.73208201e-01 -1.12187552e+00
-1.31950176e+00 1.30861437e+00 -4.79578674e-01 -1.77546367e-01
6.77094221e-01 8.81927013e-02 1.23597562e+00 -1.09462130e+00
7.35684991e-01 8.80325019e-01 6.60591483e-01 9.92110312e-01
-9.29507375e-01 -5.78963697e-01 -2.27909070e-02 -7.66685531e-02
-9.40621853e-01 -2.19377458e-01 1.06374371e+00 -6.58462405e-01
9.96967912e-01 -1.59465492e-01 5.13486922e-01 1.10302973e+00
5.58825910e-01 1.05024421e+00 2.68272251e-01 -5.38935423e-01
3.05319995e-01 1.82136610e-01 -7.45322034e-02 5.52262366e-01
5.28270490e-02 -1.96292996e-01 -3.56406331e-01 -4.24156129e-01
1.65446416e-01 2.80396700e-01 -3.12954605e-01 -5.52767292e-02
-1.24712288e+00 1.13088441e+00 5.63733280e-01 2.00577796e-01
-3.30526590e-01 2.24845439e-01 7.79196978e-01 3.80934596e-01
4.91474926e-01 7.47762740e-01 -3.24283808e-01 -2.78087258e-01
-9.44367528e-01 4.72801000e-01 9.80755806e-01 1.38745952e+00
8.85765195e-01 1.01740696e-02 -6.03689134e-01 9.56730425e-01
4.44964707e-01 4.63111728e-01 8.78348053e-01 -5.66145062e-01
9.34051514e-01 1.05425060e+00 -1.75363366e-02 -9.56913292e-01
1.73501432e-01 -1.56483710e-01 -6.92470253e-01 9.54299346e-02
-1.73010111e-01 -3.22931051e-01 -5.44460595e-01 1.36833560e+00
-1.59316972e-01 -3.06290954e-01 4.58446920e-01 5.75768471e-01
9.43671167e-01 8.09305370e-01 -2.65377909e-01 7.32435100e-03
8.63424182e-01 -1.25087547e+00 -2.44984582e-01 -2.66319335e-01
8.80570471e-01 -7.97486544e-01 1.26161838e+00 1.23052858e-01
-8.25691462e-01 -5.74153781e-01 -9.61619139e-01 1.15357943e-01
-3.09233442e-02 3.87412637e-01 3.83903354e-01 2.07836747e-01
-9.41832900e-01 5.40040731e-01 -2.58434087e-01 -2.12159723e-01
3.36037725e-01 -2.48919427e-01 4.83006239e-02 -1.17408998e-01
-9.61668551e-01 3.63552928e-01 3.56768489e-01 -1.48526549e-01
-1.03785241e+00 -9.62956190e-01 -1.11459255e+00 2.01107621e-01
1.67653725e-01 -5.99948525e-01 1.61192346e+00 -1.42798579e+00
-1.34896016e+00 3.96923125e-01 -7.48318061e-02 -4.20978040e-01
2.05005072e-02 -1.83732137e-01 -3.34030449e-01 -2.69489646e-01
3.80653381e-01 6.22623742e-01 8.00306082e-01 -1.20979810e+00
-5.74920237e-01 3.21852505e-01 2.61872679e-01 -3.68081070e-02
-3.64264190e-01 2.15370297e-01 1.55580658e-02 -8.13956976e-01
-5.53400278e-01 -8.15258563e-01 -3.61606032e-01 -3.08633000e-01
-5.98671377e-01 -5.07326782e-01 3.66472244e-01 -6.02177262e-01
1.51618254e+00 -2.18280315e+00 7.62914831e-04 -2.66167000e-02
1.41900420e-01 2.59325027e-01 -5.10030329e-01 9.20341730e-01
-5.74688949e-02 2.77863443e-01 -5.72892129e-01 5.18412422e-03
1.60655439e-01 -3.75427008e-01 -6.27411366e-01 -2.47558638e-01
4.91175115e-01 9.29163456e-01 -1.30775189e+00 -3.38357359e-01
-5.94736099e-01 2.03678682e-01 -9.55704510e-01 6.77243233e-01
-7.69152939e-01 -5.57558285e-03 -6.28768563e-01 2.20977396e-01
2.47198284e-01 -3.38356405e-01 -1.35471776e-01 2.40165636e-01
-1.12908445e-01 6.02598250e-01 -4.72796589e-01 2.17310071e+00
-8.67461264e-01 7.32010901e-01 -6.91401303e-01 -6.06487870e-01
1.54442978e+00 3.98516536e-01 6.02252819e-02 -6.56465173e-01
-2.44105548e-01 4.81186688e-01 1.49869442e-01 -9.29367363e-01
6.26525879e-01 9.71852392e-02 -3.38607788e-01 1.17447770e+00
-1.95702195e-01 -1.96959719e-01 3.86115521e-01 5.28969169e-01
1.51965570e+00 4.29077744e-01 4.88122165e-01 -2.52585597e-02
6.49999738e-01 1.56474128e-01 3.47364128e-01 6.16764069e-01
4.44917053e-01 8.36440742e-01 6.30397320e-01 -4.47832376e-01
-1.13839579e+00 -8.43609929e-01 3.70132655e-01 1.02114737e+00
-1.75719947e-01 -8.58921170e-01 -8.26525688e-01 -1.15697432e+00
-1.79714188e-01 7.94025958e-01 -7.33544469e-01 -4.36920911e-01
-6.87170744e-01 -2.69147128e-01 4.77728754e-01 5.56190789e-01
2.09752977e-01 -1.63821805e+00 -6.58170462e-01 4.55269217e-01
-2.03888118e-01 -2.40301237e-01 -9.55990672e-01 -1.58322468e-01
-5.59295774e-01 -9.19501901e-01 -1.03172827e+00 -9.31198001e-01
1.00773907e+00 5.19350730e-02 1.48709762e+00 4.32635486e-01
-6.11609854e-02 -7.64767230e-02 -9.55410957e-01 -3.54483098e-01
-1.19851935e+00 3.07663411e-01 -5.57704091e-01 -3.22247654e-01
4.09853548e-01 -4.57412422e-01 -5.82499266e-01 -1.88261226e-01
-1.00070226e+00 2.57294625e-01 1.03556132e+00 9.00847971e-01
2.33063325e-02 -4.39513862e-01 1.09053111e+00 -1.22597206e+00
1.20578766e+00 -1.02513552e+00 -2.87293345e-01 3.28478456e-01
-7.26462007e-01 4.53848720e-01 1.22303653e+00 -3.43020052e-01
-1.20782554e+00 -1.07098982e-01 -1.16760314e-01 -5.18842898e-02
3.60925682e-02 1.14390218e+00 2.25488767e-01 4.20313090e-01
1.19059730e+00 5.58254957e-01 -7.26864859e-02 -2.95844674e-01
3.65067869e-01 1.08874226e+00 2.49050707e-01 -7.24238813e-01
7.59592533e-01 -1.17094591e-01 -7.31721103e-01 -3.31408866e-02
-6.63884401e-01 -3.41708362e-01 -4.19006735e-01 -1.08310111e-01
5.82356930e-01 -7.40656495e-01 -3.02304737e-02 1.85444176e-01
-1.74846756e+00 -2.64313787e-01 -2.41750434e-01 1.10230379e-01
-5.74041247e-01 2.58318543e-01 -4.34638828e-01 -6.46149397e-01
-7.97204554e-01 -1.19762695e+00 1.18801236e+00 3.34093958e-01
-3.72716099e-01 -8.45228553e-01 4.97204781e-01 -1.57624662e-01
6.56362891e-01 8.44148695e-02 1.25342405e+00 -9.38031614e-01
-6.34419799e-01 -1.07373849e-01 -2.69816220e-01 3.47621948e-01
4.17270988e-01 1.11231118e-01 -5.40089965e-01 -1.97466835e-01
-2.88321674e-01 -3.02031606e-01 5.58329701e-01 -2.49082655e-01
9.36241269e-01 -7.00455606e-01 -1.51222497e-01 2.52189696e-01
1.45747304e+00 3.18984449e-01 5.77860177e-01 9.63434428e-02
6.42224729e-01 6.40559316e-01 6.29283607e-01 7.20773458e-01
5.48027456e-01 3.38382304e-01 3.95891935e-01 2.62294888e-01
-3.95140462e-02 -5.22997856e-01 7.82287896e-01 1.40658891e+00
3.43410879e-01 -2.12012872e-01 -1.01758289e+00 8.32883775e-01
-1.85092735e+00 -9.15019989e-01 -2.97053922e-02 2.05435085e+00
1.29002666e+00 -5.28098308e-02 -2.02322137e-02 -2.54997849e-01
6.10104680e-01 -7.62944892e-02 -6.37739003e-01 -5.58896840e-01
5.64320683e-01 3.50240506e-02 -1.65466011e-01 -1.07994966e-01
-4.86509174e-01 5.51222146e-01 5.18671751e+00 6.15886271e-01
-9.85619247e-01 -1.09388635e-01 2.32379198e-01 2.50732034e-01
-9.21002507e-01 1.14206970e-01 -5.98017812e-01 7.68088996e-01
1.07607985e+00 -8.90909314e-01 4.64373678e-01 1.12940431e+00
1.36283785e-01 1.44947276e-01 -1.53207088e+00 6.07194245e-01
6.43747747e-01 -1.50721347e+00 3.71172935e-01 -3.18181843e-01
1.16000521e+00 -6.55494491e-03 -4.47425358e-02 6.01529121e-01
6.35145903e-01 -8.56280744e-01 6.58691823e-01 5.69126308e-01
7.66009748e-01 -6.44660830e-01 9.44036484e-01 4.45099086e-01
-1.21680260e+00 -1.87872931e-01 -5.43008745e-01 -9.18026716e-02
-4.99412529e-02 5.71605682e-01 -1.09829998e+00 6.00932837e-01
4.19943869e-01 1.21473670e+00 -7.27128744e-01 1.25385153e+00
-3.84563088e-01 5.43021321e-01 4.40014154e-01 -5.25117159e-01
2.29694337e-01 8.54171664e-02 2.34040752e-01 1.31971395e+00
6.59633279e-01 -4.47761089e-01 4.93135601e-02 1.57241309e+00
-3.42654228e-01 4.33045805e-01 -8.41462851e-01 -1.23310447e-01
4.87841308e-01 1.21367347e+00 -2.53304094e-01 -3.77024055e-01
-6.78828776e-01 8.45653296e-01 3.81067425e-01 3.67101222e-01
-6.48900270e-01 -1.10656142e+00 3.13641071e-01 -2.23475307e-01
3.32283556e-01 2.74508357e-01 9.12685413e-03 -1.12178433e+00
3.85878772e-01 -1.05235219e+00 2.62763910e-02 -9.34210420e-01
-1.36274242e+00 8.12801659e-01 -2.55227029e-01 -1.78666461e+00
-8.29433203e-01 -5.85605428e-02 -1.32605803e+00 9.67021823e-01
-1.28249788e+00 -9.65007663e-01 -2.89920151e-01 7.10794404e-02
1.02919757e+00 -5.90216458e-01 6.99023306e-01 1.41261101e-01
-1.59362152e-01 6.73532963e-01 1.32541060e-01 4.86674637e-01
6.32003069e-01 -1.37642050e+00 9.56522107e-01 9.85342681e-01
9.21067595e-02 1.13779712e+00 4.85985845e-01 -8.60874712e-01
-1.20651448e+00 -1.52266943e+00 9.37223196e-01 -3.26990575e-01
6.26879394e-01 -3.06018054e-01 -9.52392757e-01 4.78235602e-01
5.67937195e-01 -4.00557160e-01 8.34510148e-01 -1.78556323e-01
-5.70293427e-01 9.66430306e-02 -6.95181489e-01 7.30557323e-01
9.07187581e-01 -6.41476929e-01 -7.81319916e-01 3.11995387e-01
1.14250505e+00 -2.12821215e-01 -7.47523129e-01 2.83127110e-02
4.40136224e-01 -8.81734550e-01 3.81608874e-01 -5.59906721e-01
1.39768422e+00 -3.78027678e-01 1.91068888e-01 -1.81332493e+00
-4.23417352e-02 -6.96375847e-01 9.69653055e-02 1.64566278e+00
8.57341945e-01 -1.90469623e-01 5.46159625e-01 2.10782349e-01
-6.37850881e-01 -8.13837349e-01 -2.67829597e-01 -4.93448406e-01
1.93534762e-01 -1.96478710e-01 8.48709881e-01 6.74392700e-01
5.95576704e-01 4.96675342e-01 -2.79112339e-01 -4.49656516e-01
1.67384461e-01 5.11070490e-01 1.00359690e+00 -1.08474207e+00
-4.50426161e-01 -4.77079332e-01 -2.85889804e-02 -1.10793245e+00
4.73652244e-01 -1.36930346e+00 6.53903663e-01 -1.69075847e+00
5.54730356e-01 -4.42359298e-01 5.22346571e-02 4.22365338e-01
-4.64065760e-01 -1.52302399e-01 1.35750562e-01 3.73806208e-01
-6.23774350e-01 6.07711315e-01 9.95903969e-01 -5.04544377e-01
-1.47274688e-01 1.13559701e-01 -1.05428433e+00 3.06012124e-01
5.93345582e-01 -7.65499473e-01 -5.97398818e-01 -5.64960957e-01
8.73163462e-01 2.10062191e-01 -3.72093767e-02 -9.31059897e-01
3.00859064e-01 -6.31688610e-02 -2.26621121e-01 -3.13632250e-01
-5.10093093e-01 -5.39640546e-01 -5.53739071e-02 5.01167953e-01
-1.05589247e+00 2.87095636e-01 -2.32699811e-01 4.89274889e-01
-3.56768221e-01 -8.52224410e-01 3.45987231e-01 -2.55916387e-01
-5.90481758e-01 2.12242231e-01 -4.10958588e-01 3.50172281e-01
6.73126101e-01 1.08717300e-01 -6.99782073e-01 -2.78253436e-01
1.31982788e-01 1.35156572e-01 6.99499309e-01 9.19427693e-01
8.44215810e-01 -1.46962762e+00 -1.14830852e+00 2.71596253e-01
8.84116173e-01 1.21144697e-01 -1.36428297e-01 4.21364367e-01
-3.36081773e-01 1.70982555e-01 -1.15938663e-01 -3.80484998e-01
-8.10083389e-01 6.33960068e-01 4.72629396e-03 -2.11605176e-01
-5.62260151e-01 5.93189538e-01 1.89991876e-01 -5.59138417e-01
-1.59620672e-01 -5.00489950e-01 -3.19334179e-01 -3.07238907e-01
8.11272681e-01 -1.03258908e-01 -7.86044355e-03 -2.14077950e-01
-3.31989899e-02 4.10010129e-01 -3.99890572e-01 2.29920864e-01
1.22523391e+00 2.96747714e-01 -2.46237993e-01 5.22447050e-01
1.34125185e+00 1.72091499e-01 -1.01384127e+00 -3.40052307e-01
3.04319769e-01 -3.77824157e-01 -6.96734488e-01 -8.15888822e-01
-1.04264235e+00 8.16287518e-01 -1.31260976e-01 2.78816164e-01
8.52945447e-01 1.81234330e-01 9.72123265e-01 5.26249945e-01
4.49979216e-01 -7.31176257e-01 5.84390819e-01 6.92198634e-01
1.05247366e+00 -1.13673747e+00 -5.05135238e-01 -2.36453172e-02
-7.13230610e-01 1.33842897e+00 8.58254492e-01 -2.12709591e-01
1.04779251e-01 2.22210497e-01 1.54298455e-01 -9.13556218e-02
-1.17153728e+00 2.19901159e-01 2.51833111e-01 7.66124606e-01
1.06424975e+00 -3.19422185e-01 -3.00173998e-01 7.01330364e-01
-2.85293251e-01 8.13890100e-02 1.02096677e+00 8.50957453e-01
-4.47069913e-01 -1.26417637e+00 4.51037660e-02 8.91954601e-01
-2.31328800e-01 -6.92060292e-01 -3.83563042e-01 1.83453321e-01
-1.15149587e-01 8.57915282e-01 -8.36546347e-02 -5.23005307e-01
2.22097233e-01 -4.00421508e-02 -1.43019155e-01 -1.36073792e+00
-1.08539808e+00 -6.02355480e-01 -1.27817407e-01 -2.97412485e-01
-1.84817657e-01 -6.50028646e-01 -1.44619918e+00 5.14398776e-02
-2.67480791e-01 5.83695948e-01 5.83253920e-01 7.84526825e-01
7.41577387e-01 6.43979371e-01 8.74113977e-01 -6.57993376e-01
-6.43870175e-01 -1.15463448e+00 -2.41975542e-02 5.75802982e-01
3.79372030e-01 -1.24072239e-01 -2.84568846e-01 5.17373800e-01] | [7.667696952819824, 7.9518280029296875] |
0d54634f-3eff-450a-b9fc-32f8434ed046 | understanding-engagement-with-insurgents | null | null | https://aclanthology.org/U15-1015 | https://aclanthology.org/U15-1015.pdf | Understanding engagement with insurgents through retweet rhetoric | null | ['Timothy Baldwin', 'Joel Nothman', 'Christoph Breidbach', 'Atif Ahmad', 'David Malet'] | 2015-12-01 | understanding-engagement-with-insurgents-1 | https://aclanthology.org/U15-1015 | https://aclanthology.org/U15-1015.pdf | alta-2015-12 | ['dialogue-act-classification'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.536954879760742, 3.546748638153076] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.