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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0d60bd80-95d1-4d88-84ce-c93572a85644 | deepdeform-learning-non-rigid-rgb-d-1 | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Bozic_DeepDeform_Learning_Non-Rigid_RGB-D_Reconstruction_With_Semi-Supervised_Data_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Bozic_DeepDeform_Learning_Non-Rigid_RGB-D_Reconstruction_With_Semi-Supervised_Data_CVPR_2020_paper.pdf | DeepDeform: Learning Non-Rigid RGB-D Reconstruction With Semi-Supervised Data | Applying data-driven approaches to non-rigid 3D reconstruction has been difficult, which we believe can be attributed to the lack of a large-scale training corpus. Unfortunately, this method fails for important cases such as highly non-rigid deformations. We first address this problem of lack of data by introducing a novel semi-supervised strategy to obtain dense inter-frame correspondences from a sparse set of annotations. This way, we obtain a large dataset of 400 scenes, over 390,000 RGB-D frames, and 5,533 densely aligned frame pairs; in addition, we provide a test set along with several metrics for evaluation. Based on this corpus, we introduce a data-driven non-rigid feature matching approach, which we integrate into an optimization-based reconstruction pipeline. Here, we propose a new neural network that operates on RGB-D frames, while maintaining robustness under large non-rigid deformations and producing accurate predictions. Our approach significantly outperforms existing non-rigid reconstruction methods that do not use learned data terms, as well as learning-based approaches that only use self-supervision.
| [' Matthias Niessner', ' Christian Theobalt', ' Michael Zollhofer', 'Aljaz Bozic'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['rgb-d-reconstruction'] | ['computer-vision'] | [ 2.93491453e-01 7.95990750e-02 4.18124013e-02 -6.15716219e-01
-1.22850025e+00 -5.61610579e-01 5.86820543e-01 -3.21664572e-01
-3.39649141e-01 4.45493281e-01 5.91507435e-01 1.92682639e-01
1.36440881e-02 -5.19216299e-01 -1.01208079e+00 -4.81096894e-01
3.38655710e-01 6.55574739e-01 5.32846868e-01 -1.99947417e-01
1.52624503e-01 6.98941588e-01 -1.52587903e+00 1.52974606e-01
3.25334758e-01 8.14538419e-01 7.86636621e-02 4.31652814e-01
3.84294271e-01 6.13958478e-01 -6.39318824e-02 -2.80575365e-01
4.01149362e-01 -3.68265867e-01 -9.16453123e-01 3.14675540e-01
7.27671802e-01 -6.44644380e-01 -4.30561393e-01 6.58674002e-01
6.95575476e-01 3.71747226e-01 3.51912975e-01 -5.71451306e-01
-3.10629338e-01 1.32583827e-01 -4.42716897e-01 -4.74894457e-02
8.16436231e-01 7.99371228e-02 9.60434020e-01 -1.01119149e+00
1.24534047e+00 1.16096103e+00 9.45427120e-01 7.51366735e-01
-1.39408565e+00 -2.46888667e-01 -1.85378522e-01 -5.97868264e-02
-1.29380786e+00 -8.41141284e-01 1.07898402e+00 -4.62002695e-01
1.13733077e+00 1.94564342e-01 7.26054728e-01 1.14891624e+00
-2.29213268e-01 5.15131891e-01 9.65627611e-01 -5.20506740e-01
6.15251735e-02 -3.19718689e-01 -4.04594809e-01 7.00198352e-01
-2.02927932e-01 3.79214436e-01 -4.98467773e-01 -9.61180031e-03
9.34015691e-01 -5.23413010e-02 -3.71143639e-01 -7.22107708e-01
-1.49271524e+00 5.01962245e-01 4.75229263e-01 2.90261686e-01
-2.06086352e-01 4.16323692e-02 1.97277054e-01 1.76316068e-01
5.52914321e-01 2.64105052e-01 -5.54319263e-01 -1.97465971e-01
-1.07949030e+00 3.38182420e-01 7.48016775e-01 7.59894550e-01
9.60735857e-01 -3.42775844e-02 2.11463422e-01 6.29738748e-01
4.07691419e-01 4.36739802e-01 4.16225612e-01 -1.28590608e+00
3.48168910e-01 4.28459823e-01 7.65827820e-02 -9.41639721e-01
-5.31630814e-01 1.38628542e-01 -5.36275804e-01 1.33885220e-01
3.86254221e-01 3.33278924e-01 -9.06980336e-01 1.69864190e+00
5.76164901e-01 1.54175416e-01 3.88246551e-02 1.18156874e+00
8.14757407e-01 1.43450379e-01 -3.29241484e-01 -9.69983488e-02
8.98117542e-01 -1.01548707e+00 -4.78630453e-01 2.02731956e-02
4.24619138e-01 -9.07203197e-01 1.09531951e+00 1.34457737e-01
-1.20921528e+00 -5.10619760e-01 -7.23839462e-01 -3.33187371e-01
7.37530291e-02 -1.29218444e-01 3.17040026e-01 2.89063215e-01
-1.08348191e+00 8.57789993e-01 -1.11086512e+00 -5.46305478e-01
3.98424298e-01 3.76777977e-01 -9.18340266e-01 -2.06414074e-01
-6.48004413e-01 1.05114090e+00 2.68804431e-01 -1.46858832e-02
-9.03153539e-01 -5.04127681e-01 -1.14829540e+00 -4.65180755e-01
3.51030171e-01 -7.36052334e-01 1.07682753e+00 -9.28170979e-01
-1.74759388e+00 1.23620939e+00 -2.04856664e-01 -1.27785102e-01
5.70563138e-01 -3.84492785e-01 -1.70744844e-02 2.88447261e-01
-7.31738331e-03 7.00248897e-01 8.12883973e-01 -1.53841484e+00
-1.40581772e-01 -3.12203348e-01 -2.95129716e-02 2.20703512e-01
-1.05467103e-01 1.57098264e-01 -6.74371839e-01 -7.96928883e-01
4.85197693e-01 -1.24796581e+00 -4.14632916e-01 7.09146708e-02
-5.39664686e-01 1.00581601e-01 6.60753429e-01 -6.11746013e-01
5.93421340e-01 -2.17163467e+00 5.21170855e-01 2.10210308e-01
4.03845683e-02 2.00551841e-03 -1.60358652e-01 1.30601764e-01
-2.28604391e-01 -2.22483054e-01 -3.88884008e-01 -7.02479482e-01
-5.81878498e-02 4.61107165e-01 -2.30684295e-01 6.26986265e-01
3.60971898e-01 8.91765237e-01 -7.73239911e-01 -5.30423760e-01
5.20776629e-01 6.94368243e-01 -8.41570854e-01 5.21555305e-01
-2.60238081e-01 9.32926118e-01 -3.62229168e-01 8.58419418e-01
2.81524301e-01 -2.64937580e-01 -2.07256731e-02 -5.96985519e-01
2.62058545e-02 3.12022924e-01 -1.13631511e+00 2.56077552e+00
-2.00116366e-01 4.46515292e-01 -1.89801633e-01 -9.99846518e-01
8.27964902e-01 3.00031930e-01 9.30097103e-01 -4.98125970e-01
1.57943502e-01 4.11897480e-01 -3.36299390e-01 -5.77814400e-01
3.81584048e-01 -2.92987913e-01 1.61372080e-01 6.39118135e-01
2.43837297e-01 -5.18711329e-01 -1.59634743e-02 -1.51990503e-01
1.35187912e+00 1.02355778e+00 1.59283713e-01 9.24644619e-02
4.72874701e-01 1.01424858e-01 6.32587612e-01 2.35573292e-01
-1.59726024e-01 1.33724129e+00 -5.06537519e-02 -8.45012367e-01
-1.24891651e+00 -8.98252189e-01 -2.23099172e-01 6.93110526e-01
1.48513243e-01 -4.54019070e-01 -7.04876840e-01 -4.54773992e-01
-6.34333342e-02 2.37232968e-01 -5.12345195e-01 -1.23498216e-03
-9.10752833e-01 -5.65273643e-01 2.33830526e-01 4.80122685e-01
2.11540416e-01 -1.09099627e+00 -5.34217477e-01 1.56342596e-01
-2.99125016e-01 -1.51649678e+00 -4.33383882e-01 1.33498877e-01
-1.08867800e+00 -1.16845322e+00 -5.93548298e-01 -5.37322283e-01
8.45202625e-01 2.44558454e-01 1.38908446e+00 6.74106181e-02
-2.53854036e-01 5.96182406e-01 -4.33569610e-01 8.89305547e-02
-5.05518615e-01 7.58936703e-02 2.11954653e-01 -2.24327415e-01
2.33490840e-01 -8.37408006e-01 -4.65129912e-01 5.69424689e-01
-8.21618676e-01 7.76255503e-02 4.93450552e-01 7.16031849e-01
1.11455119e+00 -4.65884596e-01 3.63719203e-02 -6.02286279e-01
-6.58790842e-02 -1.59729421e-01 -5.10470867e-01 1.04072049e-01
-2.76960999e-01 1.28743783e-01 2.62605220e-01 -4.20022577e-01
-9.23339307e-01 7.15421796e-01 -3.78343552e-01 -8.21356773e-01
-2.93234617e-01 1.86921939e-01 -1.37628317e-01 -3.72420490e-01
7.89727211e-01 1.06508315e-01 2.12859899e-01 -7.15882897e-01
3.21896136e-01 3.05375576e-01 7.33446956e-01 -7.51306534e-01
1.18838561e+00 7.06336439e-01 7.01729432e-02 -5.39109528e-01
-1.18234932e+00 -3.38474154e-01 -1.09823596e+00 -1.44275174e-01
8.03678036e-01 -1.16412270e+00 -3.09514850e-01 2.68125534e-01
-1.12938392e+00 -4.88175839e-01 -6.51770473e-01 7.69044220e-01
-1.13798952e+00 4.74830449e-01 -5.71460724e-01 -2.75103986e-01
-2.13217318e-01 -1.39825928e+00 1.53979051e+00 -1.37185916e-01
-2.59175837e-01 -7.80312538e-01 4.70401019e-01 6.31981552e-01
2.73748517e-01 6.40492201e-01 3.65685165e-01 -5.77329993e-01
-7.72634149e-01 -2.54929066e-01 -1.31330602e-02 2.88823813e-01
3.00339758e-01 -6.28103837e-02 -1.00012124e+00 -4.10681814e-01
4.16350476e-02 -7.66183794e-01 5.78723311e-01 1.78220674e-01
1.13317370e+00 1.05080336e-01 -4.83975746e-02 1.06138682e+00
1.35987163e+00 -5.50473869e-01 7.08689511e-01 5.68687677e-01
1.01864064e+00 6.16350710e-01 5.21548033e-01 2.17172444e-01
6.57400548e-01 1.05571830e+00 6.81935012e-01 -1.37049615e-01
-1.32282630e-01 -1.71906829e-01 2.86561608e-01 1.08729684e+00
-7.26045489e-01 1.49957940e-01 -7.53911138e-01 3.62825960e-01
-1.72110415e+00 -8.10279071e-01 1.12732999e-01 2.17538047e+00
1.12149155e+00 2.87909601e-02 8.44124109e-02 1.66537642e-01
3.64144683e-01 8.82724449e-02 -6.44479632e-01 1.02641083e-01
-2.06809506e-01 2.90340662e-01 2.41227731e-01 4.68571901e-01
-1.11210454e+00 1.00457656e+00 6.65888023e+00 3.72814447e-01
-1.04406929e+00 7.05461800e-02 2.48654678e-01 -2.01195836e-01
-4.59518224e-01 1.00405678e-01 -5.16130269e-01 2.26632088e-01
8.24868381e-01 2.71357745e-01 5.08183718e-01 8.31637442e-01
2.49844864e-01 1.00221187e-01 -1.32783961e+00 1.07789397e+00
3.86607677e-01 -1.43799210e+00 -9.35405940e-02 2.35812087e-02
7.80120730e-01 4.27222550e-01 -5.93921542e-01 -2.55192280e-01
3.11962843e-01 -8.76080394e-01 7.86986470e-01 6.92240238e-01
8.40396523e-01 -4.10290420e-01 5.30210853e-01 1.72115311e-01
-9.46002603e-01 3.71042460e-01 -3.23758096e-01 2.25586563e-01
4.20621127e-01 6.08394623e-01 -1.49705485e-01 7.58258402e-01
1.02876890e+00 1.13581812e+00 -3.55723828e-01 7.42954195e-01
-2.42679402e-01 2.04524472e-01 -5.13349891e-01 6.53677881e-01
-1.36162534e-01 -9.48893726e-02 6.36028826e-01 9.40940797e-01
4.06129837e-01 1.98522761e-01 1.87450424e-01 6.06864393e-01
-1.08195737e-01 -3.90261561e-02 -7.33003378e-01 4.78101581e-01
2.38060907e-01 1.29402030e+00 -5.40525734e-01 -1.36916742e-01
-5.89601696e-01 1.10074186e+00 5.37048697e-01 1.32410958e-01
-6.76690340e-01 1.85689643e-01 5.06744027e-01 2.31971428e-01
2.10631028e-01 -5.36892354e-01 1.29942775e-01 -1.53794098e+00
2.45482042e-01 -1.02981651e+00 2.07441077e-01 -9.60268915e-01
-1.20332313e+00 7.12549627e-01 -2.48419736e-02 -1.43256223e+00
-5.64977229e-01 -4.32608187e-01 -2.02665076e-01 5.82855463e-01
-1.75967574e+00 -1.37118542e+00 -6.45038843e-01 8.64963770e-01
4.59425777e-01 -1.16875738e-01 9.60780621e-01 5.29412806e-01
-4.64654684e-01 2.49915421e-01 -2.02160910e-01 8.17505792e-02
9.90848422e-01 -1.18838322e+00 3.52831632e-01 7.34947324e-01
3.02075744e-01 5.57390690e-01 5.60221851e-01 -3.08935374e-01
-1.78058302e+00 -9.55134988e-01 7.74954855e-01 -8.81120384e-01
4.59430218e-01 -2.49402031e-01 -9.54952002e-01 8.98390114e-01
-2.26966873e-01 7.97705829e-01 6.15527153e-01 -1.98630184e-01
-2.81793565e-01 6.73201680e-02 -1.18915284e+00 2.94348389e-01
1.44647086e+00 -6.30756319e-01 -7.53573954e-01 4.49534237e-01
7.17433989e-01 -1.04786003e+00 -1.29150224e+00 5.12796521e-01
6.22400165e-01 -1.10238636e+00 1.25490177e+00 -4.55409080e-01
6.62886024e-01 -3.32705349e-01 -5.25189519e-01 -1.03712583e+00
-1.43676639e-01 -7.03440070e-01 -2.17864916e-01 1.06133819e+00
-5.28027257e-03 -3.09032142e-01 8.54477406e-01 8.03810775e-01
-3.00908744e-01 -6.05694771e-01 -1.04303920e+00 -6.14635587e-01
-1.67140827e-01 -4.98330951e-01 4.71195161e-01 9.85807240e-01
-5.07935941e-01 -6.34415150e-02 -5.29726326e-01 -9.56904441e-02
6.54728711e-01 2.15871319e-01 1.14200389e+00 -1.05756271e+00
-1.52243912e-01 9.37521085e-03 -6.46961808e-01 -1.00576138e+00
4.16628003e-01 -7.83315003e-01 3.23085457e-01 -1.26033413e+00
1.70419604e-01 -5.97627819e-01 -4.66862433e-02 6.99303746e-01
8.84678662e-02 7.69908309e-01 -1.22999914e-01 7.40057826e-01
-6.02460861e-01 6.13678336e-01 1.35781813e+00 2.45735690e-01
-3.23394984e-02 -3.08214724e-01 -3.49573910e-01 9.65224743e-01
3.68828654e-01 -4.64210272e-01 -8.59611258e-02 -6.66392863e-01
1.00990713e-01 -1.41506210e-01 4.41268146e-01 -9.64254677e-01
-5.47044054e-02 -2.09825084e-01 4.17174667e-01 -4.58989888e-01
4.60254341e-01 -1.03906775e+00 4.67010379e-01 1.93623647e-01
-1.65810451e-01 1.25477254e-01 2.22329441e-02 4.42815989e-01
-2.69736618e-01 -3.51735801e-02 7.33275831e-01 -2.63026744e-01
-6.20441020e-01 3.85546356e-01 2.14589119e-01 1.10740975e-01
9.34568286e-01 -4.15865272e-01 6.84622750e-02 -2.08621114e-01
-7.52073944e-01 -1.76081821e-01 1.08564544e+00 4.50613439e-01
6.38948560e-01 -1.57018113e+00 -6.25032663e-01 3.77440333e-01
1.66504458e-01 6.10158682e-01 -1.37066860e-02 8.06497812e-01
-6.32419109e-01 1.88549012e-02 -2.22177729e-01 -9.42924440e-01
-1.12980056e+00 2.17758626e-01 3.93258840e-01 -1.10038549e-01
-1.09825480e+00 5.34231305e-01 -2.13151991e-01 -8.50647330e-01
6.89896718e-02 -2.13193282e-01 5.92027158e-02 -1.74506828e-01
2.19089940e-01 1.09785115e-02 2.16036648e-01 -1.22957361e+00
-4.40694243e-01 1.19423914e+00 3.33433181e-01 -6.88559785e-02
1.81731260e+00 -1.58091187e-01 -1.60465330e-01 3.26195359e-01
1.23726475e+00 -3.27069648e-02 -1.58814549e+00 -5.90664506e-01
-1.02786757e-01 -7.58649826e-01 -8.18917379e-02 -2.90988863e-01
-1.24925673e+00 5.45828044e-01 5.78940928e-01 -2.37717614e-01
1.13010442e+00 1.18204005e-01 8.85263503e-01 4.89634901e-01
4.65734601e-01 -9.58509326e-01 3.36898863e-01 4.59938735e-01
9.63658452e-01 -1.55306101e+00 2.93140978e-01 -3.20196331e-01
-2.52855510e-01 1.17731595e+00 4.23124343e-01 -2.68094331e-01
5.54987669e-01 2.19654053e-01 1.05925292e-01 -2.30357587e-01
-5.41551292e-01 -2.00354353e-01 4.15137500e-01 6.69690430e-01
4.64639127e-01 -5.05348206e-01 1.34819895e-01 6.78632036e-02
-3.04914296e-01 2.77862340e-01 2.46822119e-01 1.11725533e+00
-1.42864525e-01 -1.36597848e+00 -3.85924459e-01 1.11306384e-01
-4.21663225e-01 1.88081279e-01 -1.77819699e-01 7.94313550e-01
-1.49626002e-01 5.49619496e-01 4.03300934e-02 -4.43843722e-01
6.25506699e-01 -1.97807714e-01 8.02057981e-01 -5.83517194e-01
-2.97302634e-01 1.76996142e-01 1.02786228e-01 -1.04596245e+00
-1.28667831e+00 -8.75965536e-01 -1.11738181e+00 -2.40074605e-01
-2.87478328e-01 -3.20797026e-01 6.57110751e-01 1.08752036e+00
3.34438592e-01 2.67405659e-01 7.29347050e-01 -1.57618690e+00
-4.70675021e-01 -7.02976048e-01 -1.05800763e-01 9.63147461e-01
3.72348249e-01 -7.74050891e-01 -4.14999336e-01 5.32546163e-01] | [8.282771110534668, -2.495356321334839] |
d25b5b1b-e1f5-4588-bb9c-dc596ec56819 | bimodal-segnet-instance-segmentation-fusing | 2303.11228 | null | https://arxiv.org/abs/2303.11228v1 | https://arxiv.org/pdf/2303.11228v1.pdf | Bimodal SegNet: Instance Segmentation Fusing Events and RGB Frames for Robotic Grasping | Object segmentation for robotic grasping under dynamic conditions often faces challenges such as occlusion, low light conditions, motion blur and object size variance. To address these challenges, we propose a Deep Learning network that fuses two types of visual signals, event-based data and RGB frame data. The proposed Bimodal SegNet network has two distinct encoders, one for each signal input and a spatial pyramidal pooling with atrous convolutions. Encoders capture rich contextual information by pooling the concatenated features at different resolutions while the decoder obtains sharp object boundaries. The evaluation of the proposed method undertakes five unique image degradation challenges including occlusion, blur, brightness, trajectory and scale variance on the Event-based Segmentation (ESD) Dataset. The evaluation results show a 6-10\% segmentation accuracy improvement over state-of-the-art methods in terms of mean intersection over the union and pixel accuracy. The model code is available at https://github.com/sanket0707/Bimodal-SegNet.git | ['Yahya Zweiri', 'Dimitrios Makris', 'Rajkumar Muthusamy', 'Fariborz Baghaei Naeini', 'Xiaoqian Huang', 'Sanket Kachole'] | 2023-03-20 | null | null | null | null | ['robotic-grasping'] | ['robots'] | [ 4.14771438e-01 -1.74313530e-01 2.28379190e-01 -5.07187366e-01
-8.17377746e-01 -4.92199779e-01 1.28025621e-01 -1.54619798e-01
-5.02911270e-01 5.34775376e-01 -2.12008566e-01 1.92550123e-01
1.30218416e-01 -4.71413106e-01 -1.23211312e+00 -9.86404896e-01
1.58662703e-02 -6.54662997e-02 5.92190206e-01 2.54214078e-01
2.06656024e-01 5.14871359e-01 -1.57663918e+00 6.56590402e-01
8.44935536e-01 1.79396403e+00 6.59796178e-01 7.42780924e-01
1.33189991e-01 6.66717708e-01 -5.40057063e-01 -1.88543305e-01
4.98686761e-01 7.07006231e-02 -5.70851743e-01 -1.24406209e-02
5.90317607e-01 -8.22495699e-01 -7.78187931e-01 1.20865786e+00
6.74358845e-01 -1.68126464e-01 3.50537628e-01 -1.41051352e+00
-7.30013132e-01 4.17923838e-01 -6.83234274e-01 2.37919375e-01
-4.56842817e-02 5.84644675e-01 3.83249760e-01 -6.71154022e-01
6.92479014e-01 1.41879177e+00 6.47390544e-01 3.07125241e-01
-1.26904023e+00 -5.44295669e-01 3.38130966e-02 3.61892402e-01
-8.31015468e-01 -2.23277926e-01 5.50626755e-01 -3.85745585e-01
8.30954611e-01 -1.04806617e-01 4.19579834e-01 1.38043749e+00
5.75561762e-01 1.18736863e+00 1.36712289e+00 6.58758879e-02
8.68677720e-02 -1.94851488e-01 2.70656675e-01 4.50290859e-01
2.89527953e-01 1.09400973e-01 -4.44505811e-01 3.75896782e-01
1.12442768e+00 1.30564272e-01 -3.92162412e-01 -2.43246719e-01
-1.20378327e+00 1.67979375e-01 7.03889430e-01 -1.08644806e-01
-6.26928985e-01 7.91261017e-01 4.73554730e-01 -3.13444324e-02
8.94133747e-02 3.82138491e-02 -7.67627895e-01 -9.03278962e-02
-5.87086737e-01 2.71063924e-01 6.69053316e-01 1.39551616e+00
6.00322127e-01 -3.52489576e-02 -5.05944848e-01 6.88422978e-01
4.35124516e-01 7.69554317e-01 2.82661587e-01 -1.42193377e+00
4.69570786e-01 3.48174572e-01 2.63256162e-01 -6.85520232e-01
-4.01774675e-01 4.76555713e-02 -6.30161643e-01 5.30241489e-01
7.33962238e-01 -2.24913687e-01 -1.58869243e+00 1.64546311e+00
2.30011120e-01 -3.03295795e-02 -1.00425065e-01 1.18378913e+00
1.12451518e+00 5.98464370e-01 1.55790821e-01 8.51380005e-02
1.40921032e+00 -8.77030969e-01 -9.43789899e-01 -2.09499493e-01
-2.85201460e-01 -9.60153759e-01 8.57880831e-01 5.40397882e-01
-1.49327457e+00 -5.88244140e-01 -1.11432409e+00 -4.07267779e-01
-4.27815318e-01 4.80020493e-01 6.04162335e-01 2.25898013e-01
-9.47362423e-01 7.62957692e-01 -1.24064386e+00 -3.04584265e-01
8.10901105e-01 5.20718575e-01 -1.06475838e-01 -2.57673174e-01
-8.29268932e-01 6.77002847e-01 3.93901139e-01 5.10845065e-01
-1.17629015e+00 -6.48356855e-01 -6.48611724e-01 -8.07548910e-02
3.34948361e-01 -3.93153846e-01 1.27709448e+00 -8.82609546e-01
-1.43663919e+00 7.38226950e-01 -1.08830929e-02 -3.39540511e-01
7.08093107e-01 -5.50609827e-01 2.29291692e-01 3.75199586e-01
-1.45092905e-01 9.71357822e-01 7.58983850e-01 -1.46705782e+00
-5.86387396e-01 -5.26001453e-01 9.55632553e-02 7.74659738e-02
1.67464375e-01 9.25672501e-02 -5.85862756e-01 -6.00989938e-01
4.07687247e-01 -6.88853681e-01 9.17022303e-02 4.18193966e-01
-4.52127635e-01 2.63808924e-03 1.18027902e+00 -9.33825314e-01
4.92223650e-01 -2.11127853e+00 1.04205951e-01 -3.74601185e-01
-1.87156796e-01 1.70783609e-01 -1.83778346e-01 1.38546359e-02
5.23105077e-02 -2.66439766e-01 -3.49445611e-01 -3.60862702e-01
7.03352615e-02 1.37645915e-01 -2.09693238e-01 5.70943356e-01
4.64078903e-01 1.04405177e+00 -6.60263658e-01 -3.74311626e-01
4.71637011e-01 5.98472834e-01 -1.76204622e-01 3.81983101e-01
-3.41109157e-01 5.17709553e-01 -1.84167221e-01 1.19008219e+00
1.05757773e+00 3.08360904e-02 -1.64656952e-01 -7.67776310e-01
-2.68415958e-01 -5.32382466e-02 -1.26637399e+00 2.02357697e+00
2.39342935e-02 7.31302083e-01 5.82754552e-01 -8.01558316e-01
6.14644885e-01 1.90236628e-01 6.28744900e-01 -7.05275834e-01
7.18913794e-01 3.39694649e-01 -2.58001357e-01 -1.01620460e+00
3.14771235e-01 4.90774453e-01 1.13555111e-01 3.88263259e-03
3.73227835e-01 -2.31906831e-01 1.44484043e-01 -1.32491395e-01
1.09877741e+00 5.76555133e-01 -5.03708780e-01 -2.66174912e-01
7.15622976e-02 -9.02276337e-02 8.11457336e-01 7.32824445e-01
-7.04521537e-01 9.01512742e-01 5.97622812e-01 -2.33316436e-01
-1.19945800e+00 -1.31325078e+00 -3.18154693e-01 7.18572259e-01
9.05480325e-01 2.88845897e-01 -8.20807040e-01 -1.66543067e-01
2.50910640e-01 3.86775076e-01 -5.89842856e-01 -6.96817786e-02
-7.40641117e-01 -7.86565900e-01 4.75102931e-01 9.00903165e-01
1.07593989e+00 -1.35428905e+00 -1.11266434e+00 8.93396288e-02
-2.40361020e-01 -1.43906128e+00 -2.36307189e-01 6.03988886e-01
-8.72442424e-01 -1.17927384e+00 -6.91411614e-01 -9.48252082e-01
4.15150106e-01 1.08621091e-01 7.38268018e-01 -5.02857566e-01
-8.46105397e-01 3.23512435e-01 -2.32807666e-01 -5.76635420e-01
-7.32336715e-02 -2.23123744e-01 -2.62128592e-01 -3.04718733e-01
2.57954419e-01 -3.64024818e-01 -1.02283573e+00 2.70527989e-01
-1.08264947e+00 9.83503833e-03 6.40434742e-01 7.71768153e-01
7.09264398e-01 -2.86886692e-01 2.63048947e-01 -1.55867666e-01
1.25787348e-01 -2.88921803e-01 -8.51417899e-01 2.36051902e-01
-1.75281823e-01 -2.41012499e-01 2.09494874e-01 -3.69248122e-01
-1.30868983e+00 1.94604144e-01 2.38839090e-01 -5.57995677e-01
-4.58912015e-01 -1.32947087e-01 -2.28324890e-01 -2.28605419e-02
4.27385360e-01 -2.04681177e-02 4.84361462e-02 -4.48112488e-01
2.53176153e-01 8.43687236e-01 9.69738543e-01 -6.10637724e-01
1.40910760e-01 6.33424044e-01 -3.28515679e-01 -4.53936040e-01
-3.97058904e-01 -3.19111317e-01 -6.03795826e-01 -4.24596936e-01
1.22320271e+00 -1.01383269e+00 -8.48924696e-01 1.38453543e+00
-1.53394330e+00 -6.53068364e-01 -2.09525377e-01 4.27524179e-01
-8.50676537e-01 1.92995489e-01 -1.10455036e+00 -6.84809089e-01
-5.32530904e-01 -1.58255529e+00 1.18613434e+00 7.84642935e-01
5.00870287e-01 -2.14054525e-01 -7.12584555e-01 3.99023682e-01
3.52961928e-01 6.68038905e-01 5.14712274e-01 -8.22822750e-02
-1.10582495e+00 5.58200367e-02 -7.82573462e-01 6.37362897e-01
6.97029307e-02 3.19119655e-02 -1.21136689e+00 -3.24628949e-01
5.49618416e-02 -5.42225063e-01 1.01433051e+00 9.74470377e-01
1.44277120e+00 6.44443631e-02 -2.15858027e-01 8.93459797e-01
1.62373161e+00 5.39281309e-01 7.61649132e-01 1.76676229e-01
7.41883218e-01 3.27099234e-01 7.22763717e-01 4.14507926e-01
1.73105985e-01 4.29780632e-01 9.40417290e-01 -2.19138488e-01
-4.99912649e-01 4.18500870e-01 2.73360074e-01 1.64266061e-02
-1.31266676e-02 -3.82142007e-01 -7.51990318e-01 7.60143220e-01
-1.91299307e+00 -6.54898405e-01 -3.71902585e-01 1.84952307e+00
8.28336775e-01 -9.99160856e-02 -1.86185345e-01 -1.79051250e-01
8.73700082e-01 -4.23238613e-02 -1.10707200e+00 -2.74779946e-01
-3.81568372e-01 9.03421268e-02 1.06543326e+00 1.04952820e-01
-1.31125057e+00 9.24728096e-01 5.69315290e+00 4.36350733e-01
-1.22605002e+00 8.59222859e-02 7.37158537e-01 -1.77563906e-01
2.32767910e-01 -4.85262543e-01 -4.68183994e-01 6.18938029e-01
4.91803199e-01 4.80283052e-01 4.66083825e-01 7.25309312e-01
2.30491266e-01 -6.11930311e-01 -9.36108768e-01 1.09393609e+00
-3.53139527e-02 -9.35709178e-01 -6.05413496e-01 -2.92889982e-01
7.86499381e-01 6.34597480e-01 1.79439664e-01 -2.25543097e-01
3.07086349e-01 -8.56426835e-01 1.18638206e+00 5.96833467e-01
7.75894165e-01 -2.40568846e-01 8.14262927e-01 -2.34057710e-01
-9.56300914e-01 -3.93443018e-01 -2.63191760e-01 3.72880757e-01
2.66266167e-01 3.46119076e-01 -2.91102916e-01 3.40195894e-01
1.47755384e+00 7.55194962e-01 -3.49777848e-01 1.23076212e+00
-1.25835076e-01 3.03488523e-01 -4.38419342e-01 3.17707598e-01
1.75554007e-01 -5.30632474e-02 5.39972842e-01 1.24281824e+00
2.00903311e-01 2.25328088e-01 -7.33560547e-02 1.24928439e+00
-2.23857369e-02 -6.58654451e-01 -1.42786667e-01 7.48746470e-02
4.03369814e-01 1.22221375e+00 -8.48714650e-01 -3.64276648e-01
-3.11214983e-01 1.23514342e+00 -7.72226825e-02 7.37432241e-01
-1.11526275e+00 -6.55998409e-01 7.93310881e-01 -3.27625960e-01
7.29253113e-01 -2.93787122e-01 -8.67290139e-01 -9.75828052e-01
3.55700135e-01 -5.60492575e-01 6.68571219e-02 -1.15994227e+00
-1.01960409e+00 3.02092880e-01 7.77217001e-02 -8.86082351e-01
4.14623111e-01 -1.08534610e+00 -2.95260161e-01 7.93473780e-01
-1.56919205e+00 -1.22151875e+00 -9.22149956e-01 3.61244947e-01
7.50480652e-01 2.41130471e-01 2.06545517e-01 4.26595151e-01
-7.51171291e-01 2.22185567e-01 3.02603424e-01 8.56767744e-02
8.32504630e-01 -1.29002714e+00 2.09611759e-01 9.95639145e-01
-8.08822453e-01 1.87714651e-01 6.95583701e-01 -5.72698832e-01
-1.62177336e+00 -1.06625772e+00 9.14302543e-02 -2.40997314e-01
3.89984280e-01 -3.38425726e-01 -7.70008445e-01 6.59985960e-01
3.36162835e-01 2.37301767e-01 -1.88640863e-01 -8.18604708e-01
-1.19807541e-01 -2.81519204e-01 -1.44855094e+00 8.23260322e-02
9.66390252e-01 -1.22130334e-01 -3.55239183e-01 2.03458667e-01
8.31405580e-01 -1.02751935e+00 -1.04326451e+00 6.38076305e-01
8.93390119e-01 -1.04148269e+00 9.54651356e-01 -1.17931485e-01
6.83456957e-01 -4.34890926e-01 -3.53282660e-01 -7.77804554e-01
4.39964086e-02 -2.80052453e-01 -5.47545105e-02 1.12668788e+00
3.80571485e-02 -5.39664626e-01 5.23067832e-01 7.84276009e-01
-5.42693436e-01 -8.15187931e-01 -8.95845175e-01 -7.00776517e-01
-6.62523359e-02 -2.43962824e-01 3.59417707e-01 3.68392050e-01
-4.70332056e-01 -4.75060850e-01 -1.52452039e-02 4.10421997e-01
8.17253411e-01 1.30416662e-01 3.72815222e-01 -7.49424815e-01
8.85873511e-02 -2.41027817e-01 -5.01559198e-01 -1.05781686e+00
-3.67947906e-01 -4.73224342e-01 6.83082044e-01 -1.74156642e+00
2.39402026e-01 -2.51650244e-01 -3.72700632e-01 5.71808517e-01
3.56139033e-03 3.00516665e-01 2.94010699e-01 5.81539012e-02
-5.32986879e-01 4.00718957e-01 1.41031396e+00 -3.19938958e-01
-1.11170635e-01 -5.15865982e-01 -1.35041758e-01 5.04602671e-01
6.48374438e-01 -1.78766027e-01 -2.63866670e-02 -9.68557596e-01
-4.31207508e-01 -3.81420995e-03 8.77351701e-01 -1.25390112e+00
1.77881524e-01 -4.00242321e-02 6.81397021e-01 -7.05504477e-01
4.24453557e-01 -8.33108485e-01 1.67604670e-01 5.66163719e-01
-2.15898469e-01 -1.53379202e-01 5.68139374e-01 5.59699178e-01
-1.35338515e-01 -2.60125585e-02 1.01296508e+00 -1.60826176e-01
-8.56507242e-01 2.68838167e-01 -8.68658945e-02 -1.60602793e-01
1.05770624e+00 -3.69059235e-01 -6.94216132e-01 1.22576244e-01
-6.32335901e-01 3.38878095e-01 5.10895431e-01 5.05526841e-01
6.07879877e-01 -1.11788321e+00 -4.50126022e-01 1.17841914e-01
-1.46173850e-01 4.68805313e-01 4.27210033e-01 1.18222213e+00
-8.24040949e-01 2.15849668e-01 -6.53404057e-01 -9.65493679e-01
-1.02819550e+00 1.99598387e-01 5.03866255e-01 4.43260103e-01
-5.41492403e-01 1.31535041e+00 1.43412128e-01 3.01814266e-02
5.94649553e-01 -9.99884665e-01 3.80705059e-01 -1.33166641e-01
1.76658347e-01 5.92749953e-01 -1.15095310e-01 -3.57061416e-01
-3.53001237e-01 5.98190248e-01 3.31792980e-02 1.07162952e-01
1.43992829e+00 -3.12008560e-01 -1.37167633e-01 3.30642700e-01
1.13714302e+00 -1.09542871e+00 -2.06758142e+00 -4.34631519e-02
-2.92987108e-01 -4.89562839e-01 1.01935238e-01 -1.13787329e+00
-1.29372251e+00 8.96229267e-01 1.19516027e+00 -3.05694323e-02
1.36048794e+00 -3.98680568e-02 1.13723123e+00 8.75718519e-02
1.32635459e-01 -1.36341214e+00 6.44524768e-02 4.18533295e-01
1.06484127e+00 -1.59605801e+00 -2.06914663e-01 -5.19893765e-01
-5.60313523e-01 1.15595222e+00 9.02143896e-01 -3.09761465e-01
4.48345333e-01 5.21414757e-01 1.14877872e-01 -1.78825453e-01
-3.99328351e-01 -1.60669252e-01 -4.43344489e-02 4.26430166e-01
1.80538371e-01 -9.68266279e-02 -2.59338796e-01 6.00957751e-01
2.49495074e-01 2.31850907e-01 3.73459160e-01 1.18853581e+00
-3.35313082e-01 -4.59101140e-01 -3.91422719e-01 4.14455563e-01
-4.48110849e-01 2.10666999e-01 2.45026741e-02 5.54507136e-01
4.94243532e-01 8.60339463e-01 2.64478236e-01 -2.40042001e-01
5.12542725e-01 -1.48859680e-01 7.65437603e-01 -2.79796850e-02
-4.87186164e-01 2.00680956e-01 -1.55047908e-01 -1.16237366e+00
-5.47824442e-01 -8.66620839e-01 -1.42715323e+00 3.91248353e-02
-3.13802838e-01 -7.83414960e-01 1.13727593e+00 6.51782453e-01
3.39767784e-01 9.24349308e-01 2.06532687e-01 -1.27887332e+00
-3.47783655e-01 -1.06600440e+00 -4.99790490e-01 4.19209570e-01
4.82839763e-01 -7.53844380e-01 -3.38946760e-01 5.07748127e-01] | [9.230332374572754, -0.396270751953125] |
1bcb6c36-5a1c-4016-93d3-8e09a6c8c523 | cardiac-segmentation-with-strong-anatomical | 2006.08825 | null | https://arxiv.org/abs/2006.08825v1 | https://arxiv.org/pdf/2006.08825v1.pdf | Cardiac Segmentation with Strong Anatomical Guarantees | Convolutional neural networks (CNN) have had unprecedented success in medical imaging and, in particular, in medical image segmentation. However, despite the fact that segmentation results are closer than ever to the inter-expert variability, CNNs are not immune to producing anatomically inaccurate segmentations, even when built upon a shape prior. In this paper, we present a framework for producing cardiac image segmentation maps that are guaranteed to respect pre-defined anatomical criteria, while remaining within the inter-expert variability. The idea behind our method is to use a well-trained CNN, have it process cardiac images, identify the anatomically implausible results and warp these results toward the closest anatomically valid cardiac shape. This warping procedure is carried out with a constrained variational autoencoder (cVAE) trained to learn a representation of valid cardiac shapes through a smooth, yet constrained, latent space. With this cVAE, we can project any implausible shape into the cardiac latent space and steer it toward the closest correct shape. We tested our framework on short-axis MRI as well as apical two and four-chamber view ultrasound images, two modalities for which cardiac shapes are drastically different. With our method, CNNs can now produce results that are both within the inter-expert variability and always anatomically plausible without having to rely on a shape prior. | ['Pierre-Marc Jodoin', 'Youssef Skandarani', 'Thierry Judge', 'Olivier Bernard', 'Nathan Painchaud', 'Alain Lalande'] | 2020-06-15 | null | null | null | null | ['cardiac-segmentation'] | ['medical'] | [ 4.11181360e-01 5.87045550e-01 2.73262322e-01 -5.01254916e-01
-7.30202734e-01 -9.61561799e-01 3.12583059e-01 1.24249673e-02
-3.84247720e-01 4.96194720e-01 1.00621395e-01 -4.93436068e-01
-1.89080834e-01 -6.68277979e-01 -6.36979401e-01 -7.01698661e-01
3.27036832e-03 8.57154846e-01 5.08108065e-02 1.42171040e-01
-9.97459888e-03 6.48283303e-01 -7.86621332e-01 1.07398547e-01
8.15913260e-01 6.29458606e-01 -3.65525037e-01 8.98949981e-01
2.72082061e-01 2.36940727e-01 -4.35992569e-01 -3.11044544e-01
2.37458855e-01 -4.88510877e-01 -1.06406760e+00 3.09863806e-01
4.37506527e-01 -2.61272281e-01 2.21306115e-01 9.96049821e-01
3.86836708e-01 4.13381401e-03 8.52325439e-01 -6.43835306e-01
-6.79152250e-01 5.25537908e-01 -2.04932749e-01 5.58316037e-02
1.24204122e-01 2.09749222e-01 6.32069826e-01 -6.51327014e-01
9.34096277e-01 9.27749574e-01 1.14053023e+00 6.61697745e-01
-1.64375615e+00 -1.45086423e-01 -3.60725313e-01 -6.89931810e-01
-1.24935603e+00 -2.81404942e-01 6.61912143e-01 -8.87111068e-01
5.28398931e-01 4.63022321e-01 7.81072974e-01 8.06796789e-01
6.83565676e-01 2.63413101e-01 9.06077266e-01 -2.48930246e-01
2.56289423e-01 3.93717661e-02 -1.66215017e-01 7.14939356e-01
-7.25316480e-02 1.52447030e-01 2.65627563e-01 -3.06326538e-01
1.20193958e+00 -4.54703867e-02 -4.83263314e-01 -6.30367517e-01
-1.46839905e+00 1.04689956e+00 5.95020771e-01 6.54749274e-01
-5.30406177e-01 1.94127932e-01 2.42387548e-01 -2.25906938e-01
2.94818223e-01 8.36758316e-01 -3.59392464e-01 1.69799283e-01
-1.37963998e+00 5.25490828e-02 6.82784081e-01 3.62387180e-01
2.47036248e-01 1.16096288e-01 -2.00114578e-01 6.99656248e-01
3.55101198e-01 2.13851258e-01 5.17733216e-01 -1.22618389e+00
-1.39491901e-01 2.94373810e-01 -1.55974060e-01 -1.11160350e+00
-5.63748300e-01 -4.58023816e-01 -1.10608733e+00 3.70023102e-01
5.67667127e-01 -2.61202991e-01 -1.32604575e+00 1.53162324e+00
3.04581046e-01 -6.75158799e-02 7.47079030e-02 1.07161593e+00
6.92372024e-01 4.16875869e-01 1.42954260e-01 -2.14241877e-01
1.29876721e+00 -4.46352899e-01 -4.70706195e-01 -3.21398377e-02
4.41218495e-01 -7.99382746e-01 8.54657292e-01 2.97485620e-01
-1.33290100e+00 -4.45143074e-01 -1.06151628e+00 -2.94831786e-02
1.14738293e-01 -2.43785620e-01 5.16378105e-01 7.13795483e-01
-1.26688778e+00 1.06875956e+00 -1.16032004e+00 -1.05984189e-01
6.20553195e-01 4.05912280e-01 -4.97649193e-01 3.84329885e-01
-9.07584310e-01 1.00885928e+00 5.53329766e-01 3.33758473e-01
-7.22710669e-01 -8.53270352e-01 -9.44674134e-01 -1.47178113e-01
2.02972755e-01 -8.91709447e-01 1.02068627e+00 -1.12027371e+00
-1.50271583e+00 1.12305748e+00 1.86431691e-01 -4.99129623e-01
7.30626404e-01 3.02844733e-01 -2.12059394e-01 2.22048998e-01
6.84519783e-02 8.92681897e-01 1.00076818e+00 -1.33914948e+00
1.49797305e-01 -2.32192993e-01 -2.79754460e-01 -1.61173806e-01
2.78504968e-01 -1.36905998e-01 -2.87271470e-01 -7.49262810e-01
5.72179794e-01 -1.25322974e+00 -6.31711185e-01 2.10982755e-01
-5.31247199e-01 1.53669909e-01 3.72524738e-01 -8.42724919e-01
7.43431330e-01 -2.08158565e+00 4.23516989e-01 5.40761113e-01
5.44181824e-01 1.68284982e-01 2.29329243e-01 -2.98559815e-01
-2.92394757e-01 4.21591312e-01 -9.72773612e-01 -1.17900267e-01
-3.66000623e-01 5.37423134e-01 -6.19255230e-02 6.71613991e-01
2.69003570e-01 1.03043675e+00 -9.07396734e-01 -6.16498113e-01
3.31372648e-01 7.15663791e-01 -7.67724514e-01 1.11786604e-01
2.96167918e-02 1.01980639e+00 -2.46325508e-01 3.80910903e-01
5.62060058e-01 -1.83478430e-01 3.18122149e-01 -3.20575505e-01
2.12143175e-02 -4.32936281e-01 -8.78141284e-01 1.75367534e+00
-3.93370897e-01 4.41795141e-01 1.95348531e-01 -9.96328771e-01
7.85864592e-01 5.78370214e-01 8.01231921e-01 -2.20773965e-01
3.08070660e-01 4.51113492e-01 2.34368026e-01 -3.47684801e-01
6.16458943e-03 -8.07596922e-01 1.60595328e-01 3.62212181e-01
1.83981985e-01 -5.95115304e-01 -2.02984408e-01 -9.19910669e-02
6.21823072e-01 1.24234699e-01 2.02511117e-01 -5.06650031e-01
4.76622850e-01 -7.95742646e-02 5.77206075e-01 4.94183421e-01
-3.77238244e-01 1.26478672e+00 5.74497461e-01 -9.02826905e-01
-1.36155879e+00 -1.46948946e+00 -6.18729949e-01 2.58067966e-01
-2.77645350e-01 1.99081749e-01 -9.32252526e-01 -7.02107310e-01
-1.91362590e-01 6.01713598e-01 -8.55149269e-01 -1.45777062e-01
-7.26913989e-01 -6.35363042e-01 7.12927818e-01 5.70787966e-01
-1.25972917e-02 -1.25034344e+00 -1.04256225e+00 4.83501941e-01
-9.00638998e-02 -9.57752168e-01 -4.42602783e-01 1.06232189e-01
-1.08607984e+00 -9.77590501e-01 -1.14956319e+00 -6.61064744e-01
8.45479608e-01 -8.65218043e-01 1.28321207e+00 3.19990158e-01
-6.19766414e-01 1.73632115e-01 6.51713610e-02 -3.07117134e-01
-9.15081263e-01 -2.08945692e-01 -1.45678401e-01 3.22608580e-03
-3.20167631e-01 -5.38759410e-01 -6.59748316e-01 -3.19654793e-02
-9.69562590e-01 -6.98776124e-03 3.46148998e-01 8.43930662e-01
9.27995503e-01 -2.89245695e-01 2.54515171e-01 -9.53206956e-01
5.19057751e-01 -3.18560064e-01 -3.68109941e-01 7.71162063e-02
-6.15920782e-01 1.62044570e-01 5.06561100e-01 -2.85860360e-01
-7.39762723e-01 2.39820927e-01 -3.92861247e-01 -6.39755607e-01
-3.72769713e-01 4.15758491e-01 5.26391625e-01 3.16771492e-02
9.80856001e-01 1.28605098e-01 2.62748033e-01 -1.57031119e-01
4.40579593e-01 1.60368234e-01 9.04881418e-01 -5.04486263e-01
5.50699353e-01 6.98952734e-01 2.06433743e-01 -5.42976499e-01
-6.39540374e-01 -6.19674362e-02 -1.16433704e+00 -2.76635706e-01
1.53440571e+00 -2.53204077e-01 -4.64648873e-01 1.33159965e-01
-1.14487207e+00 -3.02195489e-01 -5.20956933e-01 4.54507500e-01
-7.10063636e-01 5.11314034e-01 -6.36634827e-01 -4.59537297e-01
-6.07861757e-01 -1.53463912e+00 9.84806299e-01 -2.88303848e-03
-5.47119319e-01 -1.46177256e+00 9.96526033e-02 2.38152176e-01
5.76984227e-01 9.84118700e-01 1.10513294e+00 -6.44677222e-01
-9.84404907e-02 -2.28916109e-01 -5.52928671e-02 4.46502626e-01
1.85151935e-01 2.61568457e-01 -8.43573987e-01 -2.41983846e-01
1.05434164e-01 1.87857833e-03 6.49413884e-01 1.03816545e+00
1.30068660e+00 -1.08865984e-01 -2.08944842e-01 8.91194582e-01
1.30063534e+00 2.37473324e-01 5.67352414e-01 -1.54165938e-01
7.51634777e-01 5.82381785e-01 -1.73103362e-01 -1.57246925e-02
2.28958596e-02 4.38791931e-01 3.45028847e-01 -5.57276368e-01
-3.50743346e-02 9.20101106e-02 -2.79749900e-01 7.32555091e-01
-3.31648648e-01 1.88072369e-01 -1.25552857e+00 6.80061460e-01
-1.45726836e+00 -6.88480079e-01 -2.23198667e-01 2.08260107e+00
1.00733840e+00 2.48702019e-01 -9.12392661e-02 -9.38358367e-04
4.46568578e-01 -4.10568453e-02 -4.36878800e-01 -7.73968756e-01
2.31717840e-01 5.80139935e-01 3.49193275e-01 7.60036469e-01
-1.24593198e+00 6.94667220e-01 6.62719679e+00 2.26405159e-01
-1.46249080e+00 1.42658502e-01 8.70094895e-01 3.14110905e-01
-3.72089535e-01 -2.20610201e-02 1.58731073e-01 1.98751077e-01
8.18876505e-01 1.77660063e-01 3.19529891e-01 6.75898790e-01
-3.78682129e-02 1.00500658e-01 -1.16090620e+00 7.80732751e-01
-8.63800496e-02 -1.70282078e+00 -1.34080559e-01 -7.47195259e-02
7.52581894e-01 -1.36511147e-01 6.13540933e-02 -6.10093176e-02
1.34521571e-03 -1.60691440e+00 6.38574243e-01 6.82959557e-01
1.11203313e+00 -6.52160883e-01 8.26812327e-01 2.76635587e-01
-7.67843544e-01 3.69038075e-01 -2.14086562e-01 4.72565085e-01
2.48281240e-01 5.83180070e-01 -9.56640720e-01 4.38829124e-01
5.85972250e-01 4.00384516e-01 -2.98170716e-01 7.40409732e-01
6.35742471e-02 4.17018473e-01 -2.56608009e-01 5.32743216e-01
3.41229647e-01 -3.06511045e-01 6.10589981e-01 1.06795514e+00
1.76959425e-01 1.73369288e-01 1.69687331e-01 1.48978686e+00
1.43955395e-01 7.10698515e-02 -5.43523669e-01 1.17837921e-01
-6.94653168e-02 1.26220715e+00 -1.23159575e+00 -4.55926061e-01
6.05344027e-02 9.11561608e-01 -1.25837147e-01 2.88865834e-01
-7.25969613e-01 5.85283479e-03 1.44206062e-01 8.07880834e-02
-1.69806462e-02 -1.37775252e-02 -7.38777101e-01 -1.02304590e+00
-2.97666211e-02 -7.27102578e-01 2.74835229e-01 -6.96784973e-01
-1.13279176e+00 8.59873414e-01 -3.50968689e-02 -8.64229441e-01
-3.00960392e-01 -4.99875844e-01 -6.70904160e-01 1.06860077e+00
-9.18627441e-01 -1.06593072e+00 -3.30251381e-02 3.85026753e-01
2.83970058e-01 2.29239002e-01 1.10325170e+00 1.67020902e-01
-5.63458875e-02 3.09218913e-01 -2.66107976e-01 3.57408911e-01
4.48929369e-01 -1.46658444e+00 1.82064533e-01 5.34442842e-01
2.51896560e-01 8.59272420e-01 7.31680691e-01 -6.34333789e-01
-7.81000435e-01 -8.86378527e-01 5.44912517e-01 -6.54977441e-01
2.88191408e-01 8.40988606e-02 -1.10072351e+00 8.83724868e-01
2.15764493e-01 5.07700145e-01 5.92119157e-01 -9.33223143e-02
2.02366654e-02 4.56023306e-01 -1.34301627e+00 4.60137337e-01
5.25493205e-01 -4.36338395e-01 -8.64455462e-01 4.33040142e-01
4.00655031e-01 -8.86354327e-01 -1.34179938e+00 6.24157012e-01
6.48368120e-01 -9.22502160e-01 1.11573958e+00 -6.98599935e-01
5.76497793e-01 -1.22897021e-01 2.02097028e-01 -1.32798600e+00
-1.79466635e-01 -5.18631876e-01 3.00877094e-01 5.60842633e-01
4.49454993e-01 -4.65233445e-01 7.50850379e-01 7.87869573e-01
-3.72842431e-01 -9.95360792e-01 -1.14829111e+00 -4.00956154e-01
7.27377713e-01 -5.32043338e-01 3.01495939e-01 1.24146926e+00
-4.07089382e-01 -1.31938457e-01 -6.57130629e-02 2.08546340e-01
7.15450704e-01 7.39395320e-02 3.05528998e-01 -1.34882927e+00
-2.35248774e-01 -6.11962974e-01 -3.40220392e-01 -2.89034456e-01
-1.14797354e-02 -1.20297563e+00 2.24934131e-01 -1.42603838e+00
1.26501590e-01 -5.52042544e-01 -1.39066756e-01 5.94866335e-01
8.11294839e-03 4.12408382e-01 1.69730470e-01 2.07791626e-01
1.36174589e-01 7.20190722e-03 1.67271256e+00 -4.26058844e-02
-2.83422202e-01 -8.44804794e-02 -3.63964528e-01 1.06315434e+00
5.90183318e-01 -5.83526433e-01 -2.04882652e-01 -3.43132824e-01
3.46126668e-02 4.70973313e-01 6.46451056e-01 -9.41678047e-01
-4.90189679e-02 1.05858296e-01 8.11982870e-01 -3.02559346e-01
1.82079617e-02 -7.68172622e-01 3.73247534e-01 7.26543784e-01
-3.96016747e-01 -1.98645771e-01 2.58151323e-01 1.12925515e-01
-6.49044365e-02 -2.33703166e-01 1.11717510e+00 -3.02965552e-01
-1.02768548e-01 3.91541600e-01 -3.34092438e-01 1.91272438e-01
9.20275569e-01 -3.48489136e-01 5.07019222e-01 -1.41358137e-01
-1.55468595e+00 -8.01766440e-02 4.07448679e-01 2.20252052e-01
6.41884804e-01 -1.13120985e+00 -8.18321764e-01 4.85433251e-01
-1.73674867e-01 3.90326023e-01 3.91890079e-01 9.99747396e-01
-9.78412688e-01 2.98318505e-01 -3.12152088e-01 -1.25645113e+00
-8.51939321e-01 5.70889294e-01 9.97330248e-01 -2.19492793e-01
-9.00351524e-01 8.96576226e-01 1.09265231e-01 -8.04473639e-01
-2.96256930e-01 -7.33816922e-01 -7.27359876e-02 -1.91080868e-01
-8.07984546e-02 -2.22640038e-01 1.81991532e-02 -7.43067741e-01
-4.41815764e-01 8.91277254e-01 3.69216114e-01 -9.49902162e-02
1.26206672e+00 2.74033487e-01 -2.00287998e-01 1.74914986e-01
1.19265270e+00 5.70682511e-02 -1.28389955e+00 2.29511648e-01
-6.59540892e-02 -2.39526421e-01 2.21863434e-01 -8.43714654e-01
-1.33797431e+00 1.00101018e+00 7.73093760e-01 3.10723364e-01
7.72675574e-01 2.20448971e-02 7.91504681e-01 -2.83745915e-01
3.76260132e-02 -8.44385505e-01 -1.40105680e-01 1.40389413e-01
1.05259025e+00 -1.12516284e+00 3.33227846e-03 -3.25243384e-01
-8.21081758e-01 1.33235800e+00 1.06857181e-01 -3.94012392e-01
8.52018058e-01 3.00391376e-01 4.57013965e-01 -4.97788161e-01
-1.09474942e-01 3.03900689e-01 6.84744656e-01 5.81345916e-01
6.03673816e-01 3.83088857e-01 -3.95448916e-02 3.19111228e-01
-2.70907402e-01 9.61167831e-03 4.40929741e-01 6.12837791e-01
-1.05198190e-01 -8.17995906e-01 -4.99917120e-01 2.91357666e-01
-8.58639419e-01 1.23128910e-02 -4.07192670e-02 7.28475332e-01
3.52057040e-01 4.47501391e-01 1.92914203e-01 9.98635069e-02
2.88167953e-01 1.73807383e-01 5.13369501e-01 -7.37735093e-01
-7.07997203e-01 3.36843401e-01 -2.27193087e-01 -6.55693114e-01
-2.15381101e-01 -6.21727407e-01 -1.57641995e+00 -5.61269596e-02
-1.03628084e-01 -1.07824273e-01 6.07046008e-01 1.04963601e+00
-1.51370289e-02 7.61345148e-01 2.32144982e-01 -1.02371144e+00
-4.52048510e-01 -6.68799281e-01 -2.96437234e-01 6.84393942e-01
4.76401478e-01 -4.23332483e-01 -1.87629700e-01 3.06339264e-01] | [14.167277336120605, -2.3564460277557373] |
58402ad2-f019-4ed1-95cb-8049843ebdd9 | impact-of-action-unit-occurrence-patterns-on | 2010.07982 | null | https://arxiv.org/abs/2010.07982v1 | https://arxiv.org/pdf/2010.07982v1.pdf | Impact of Action Unit Occurrence Patterns on Detection | Detecting action units is an important task in face analysis, especially in facial expression recognition. This is due, in part, to the idea that expressions can be decomposed into multiple action units. In this paper we investigate the impact of action unit occurrence patterns on detection of action units. To facilitate this investigation, we review state of the art literature, for AU detection, on 2 state-of-the-art face databases that are commonly used for this task, namely DISFA, and BP4D. Our findings, from this literature review, suggest that action unit occurrence patterns strongly impact evaluation metrics (e.g. F1-binary). Along with the literature review, we also conduct multi and single action unit detection, as well as propose a new approach to explicitly train deep neural networks using the occurrence patterns to boost the accuracy of action unit detection. These experiments validate that action unit patterns directly impact the evaluation metrics. | ['Saandeep Aathreya', 'Shaun Canavan', 'Saurabh Hinduja'] | 2020-10-15 | null | null | null | null | ['action-unit-detection'] | ['computer-vision'] | [ 3.87591958e-01 -3.96385491e-02 -3.82387191e-01 -5.36110163e-01
-1.62916169e-01 -3.75354588e-01 5.47823668e-01 -5.60429990e-01
-1.76687837e-01 4.72187668e-01 1.15563229e-01 2.20197558e-01
2.53862828e-01 -5.97384393e-01 -3.03276211e-01 -7.31385112e-01
-2.06638247e-01 -3.98436278e-01 -3.91580731e-01 -3.13449502e-01
1.88933715e-01 1.00952744e+00 -1.96562660e+00 7.00567126e-01
1.93082631e-01 1.44408846e+00 -6.29262865e-01 3.92464966e-01
1.02451479e-03 1.04825521e+00 -9.95147347e-01 -5.28263807e-01
1.99320763e-01 -7.59479642e-01 -5.69386005e-01 3.06011498e-01
4.71405447e-01 -5.46305180e-01 3.75336153e-03 1.04105842e+00
6.89068675e-01 -7.20901489e-02 8.51494551e-01 -1.46383703e+00
-4.96453166e-01 2.86444992e-01 -7.65577257e-01 3.74551177e-01
5.31247795e-01 1.37497753e-01 1.16171432e+00 -1.06487465e+00
5.77915609e-01 1.55597091e+00 4.72092897e-01 9.21253502e-01
-9.76855695e-01 -1.02507114e+00 -3.11915600e-03 2.52905458e-01
-1.41209543e+00 -8.49344909e-01 9.18112218e-01 -4.76380348e-01
1.31281805e+00 2.09201440e-01 4.23421770e-01 1.33716166e+00
6.55197576e-02 9.93838906e-01 1.24279130e+00 -6.42315388e-01
1.11632027e-01 6.22753473e-03 1.26642868e-01 6.39200211e-01
-7.18071386e-02 -8.88938457e-03 -5.64881861e-01 -1.09151125e-01
7.95543611e-01 -4.41068679e-01 7.23050013e-02 3.20430756e-01
-3.78429949e-01 8.00352991e-01 4.26330827e-02 6.59511089e-01
-5.46436489e-01 2.07574368e-01 5.95256507e-01 4.00292516e-01
6.64266944e-01 4.05398160e-01 -8.96054357e-02 -5.18074334e-01
-8.71764541e-01 8.58720916e-04 7.29405582e-01 6.55756235e-01
8.16515386e-01 3.09828043e-01 -6.86169147e-01 1.06112993e+00
2.83300579e-01 2.30895177e-01 3.84903938e-01 -8.84650588e-01
-9.35143083e-02 9.73208249e-01 -2.38680199e-01 -1.02460206e+00
-3.63354474e-01 1.13573343e-01 -6.09593809e-01 4.57374901e-01
8.00384134e-02 -4.02854353e-01 -8.76202166e-01 1.85392833e+00
1.33769035e-01 3.39271247e-01 -2.19217449e-01 8.09837937e-01
1.03745341e+00 1.89973980e-01 2.35503003e-01 -4.22762066e-01
1.57488322e+00 -8.29064131e-01 -9.46097076e-01 -4.47816141e-02
8.29205334e-01 -7.93242812e-01 7.49713063e-01 4.29389626e-01
-8.78796160e-01 -6.93437397e-01 -1.10341895e+00 3.17180753e-01
-4.02247936e-01 6.44642055e-01 8.98163080e-01 1.07259083e+00
-9.48477447e-01 4.64905947e-01 -7.70741463e-01 -4.54650551e-01
8.27839494e-01 5.25676370e-01 -5.60203910e-01 3.42607886e-01
-1.13565421e+00 1.03893459e+00 6.02849722e-02 7.02736527e-02
-8.28818023e-01 -3.88308972e-01 -7.18885183e-01 9.20461863e-02
5.97280189e-02 5.33547718e-03 1.41000140e+00 -1.68602371e+00
-1.70105863e+00 1.16486096e+00 -2.41100788e-01 -3.25021058e-01
1.25003725e-01 -6.64945394e-02 -5.97969651e-01 3.04962456e-01
-1.69265226e-01 7.45944440e-01 9.77904916e-01 -8.54881763e-01
-4.67011482e-01 -5.12063444e-01 2.44832605e-01 -1.18318722e-01
-5.58912098e-01 7.71480858e-01 -1.98623285e-01 -6.67679548e-01
-5.51256657e-01 -7.71448612e-01 2.30781555e-01 1.23162754e-01
-1.02195762e-01 -8.25633466e-01 8.31585407e-01 -3.71314257e-01
1.59229767e+00 -2.31336474e+00 -1.16015464e-01 2.50870705e-01
1.32853374e-01 4.31892306e-01 -2.86040395e-01 1.04835555e-01
-4.02670711e-01 1.91990107e-01 -5.66528849e-02 -5.68590701e-01
-1.09745273e-02 3.53697538e-01 -6.58879504e-02 4.67425525e-01
7.10937440e-01 8.46444428e-01 -3.88851374e-01 -3.91548067e-01
1.86450571e-01 4.66766149e-01 -5.30996561e-01 2.16102719e-01
1.23088382e-01 4.97262999e-02 -3.24991912e-01 1.09357774e+00
7.45187223e-01 1.99822441e-01 -1.58589222e-02 -2.55303174e-01
-2.63601635e-02 -2.37821578e-03 -8.76850963e-01 1.16600418e+00
-2.33940616e-01 8.59776139e-01 1.72558323e-01 -1.17376018e+00
9.88288820e-01 5.96789360e-01 8.10937047e-01 -7.24334061e-01
6.44127250e-01 1.81632668e-01 3.55784088e-01 -4.83105659e-01
1.95061862e-01 -2.65370131e-01 4.10517484e-01 5.57834148e-01
3.67009431e-01 4.22144532e-01 4.42607582e-01 -3.05314809e-01
1.05214274e+00 -2.50677988e-02 6.62413120e-01 -8.18253085e-02
8.54480028e-01 -4.63566512e-01 4.85663086e-01 4.80920881e-01
-6.65933788e-01 3.21297318e-01 9.43263888e-01 -2.67952472e-01
-5.35521209e-01 -4.14867610e-01 -4.12540138e-01 1.36632264e+00
-4.03830916e-01 -4.62524325e-01 -1.08193409e+00 -7.37416923e-01
1.40835240e-01 4.47256893e-01 -1.22179091e+00 -4.77772951e-01
-4.29540068e-01 -1.02681828e+00 1.09003413e+00 7.00309336e-01
5.49647450e-01 -1.53385234e+00 -7.82241166e-01 5.72570264e-02
1.84874251e-01 -1.22187877e+00 -3.12421303e-02 3.75531502e-02
-4.43523437e-01 -1.04779840e+00 -5.46033442e-01 -4.89409983e-01
5.95246434e-01 1.01815961e-01 1.09506691e+00 2.21745521e-01
-6.48365974e-01 6.67980790e-01 -6.09475553e-01 -7.01348841e-01
-5.39535344e-01 -2.39447087e-01 1.42023563e-01 4.93669003e-01
1.22234499e+00 -4.52991039e-01 -3.55628759e-01 4.48549211e-01
-9.66019332e-01 -3.85415882e-01 8.14120531e-01 6.17284060e-01
2.96499342e-01 -3.22243392e-01 5.13807237e-01 -5.60567439e-01
9.06425059e-01 -3.71175528e-01 -3.11213553e-01 1.87554345e-01
-3.79918665e-01 -3.53473425e-01 5.80579378e-02 -5.19311368e-01
-9.59603488e-01 -3.17611103e-03 -3.60074937e-01 -6.04439795e-01
-3.60038400e-01 3.96689594e-01 -3.40750255e-02 -5.45414507e-01
8.75936508e-01 -2.89769489e-02 3.56326818e-01 -2.42703795e-01
-1.15107752e-01 8.79246354e-01 -9.81031358e-02 -4.20585930e-01
-6.23582937e-02 5.00806034e-01 1.03028737e-01 -1.16300607e+00
-6.56751752e-01 -4.96518672e-01 -7.07669020e-01 -6.10183477e-01
9.38092113e-01 -7.79101968e-01 -7.33150601e-01 6.43686414e-01
-1.27610993e+00 -1.43197358e-01 1.07855357e-01 2.26185590e-01
-4.18785065e-01 1.58497497e-01 -7.05877662e-01 -1.02988267e+00
-4.04052854e-01 -1.21045673e+00 1.27688658e+00 1.96431667e-01
-6.02377713e-01 -7.15967596e-01 -4.99969535e-03 2.83258528e-01
3.29303652e-01 4.70351845e-01 4.85901415e-01 -4.83321637e-01
1.48867816e-01 -2.05841377e-01 -3.34034711e-01 8.75000477e-01
4.33000088e-01 5.14501333e-01 -1.43942523e+00 -9.58049521e-02
3.96850072e-02 -4.23316389e-01 7.47695744e-01 4.15329486e-01
1.17636514e+00 -1.29279912e-01 -1.76753908e-01 2.24757060e-01
9.08549666e-01 3.38844657e-01 9.28593159e-01 1.08137101e-01
3.66658121e-01 7.36852229e-01 6.72675908e-01 8.54610980e-01
-4.41743761e-01 8.72451782e-01 4.34958965e-01 -1.62792906e-01
-2.23717049e-01 4.25463527e-01 8.03525031e-01 2.29021668e-01
-4.18390751e-01 -6.59625530e-02 -6.70679331e-01 1.36682734e-01
-1.49322808e+00 -1.08683050e+00 2.29863733e-01 1.80223918e+00
7.43746996e-01 -1.62710056e-01 3.62220764e-01 3.09184521e-01
4.34354633e-01 1.58961043e-01 -2.23774612e-01 -8.72569740e-01
-2.20626339e-01 5.41112125e-01 8.51352736e-02 1.80746451e-01
-1.21185184e+00 1.06504643e+00 6.85527325e+00 8.33756626e-01
-1.62620115e+00 -3.30320895e-02 8.18340480e-01 -1.35709822e-01
4.08360869e-01 -6.20374918e-01 -9.08848286e-01 2.65595227e-01
7.27592528e-01 -1.01923175e-01 7.85380751e-02 1.13164663e+00
2.12991759e-01 -1.53594255e-01 -1.32290864e+00 1.17712963e+00
4.20590460e-01 -1.00019848e+00 1.22578040e-01 1.63965762e-01
6.90128028e-01 -3.89715672e-01 3.43584307e-02 3.43674839e-01
-2.03900576e-01 -1.26208878e+00 3.24365258e-01 1.39233544e-01
7.70965993e-01 -8.18137407e-01 8.31728637e-01 -2.52599984e-01
-1.04259610e+00 -5.70587739e-02 -3.06286007e-01 -3.69706422e-01
-1.34386480e-01 3.16983610e-01 -8.53121698e-01 1.36232153e-01
6.60319567e-01 9.38798964e-01 -6.31011367e-01 4.19784456e-01
-4.93366927e-01 8.18428934e-01 -2.07928181e-01 -2.74817109e-01
1.73189282e-01 -1.20555572e-01 3.07766318e-01 1.47865558e+00
1.90347657e-01 2.54592925e-01 -2.60770649e-01 8.85561407e-01
-2.01613419e-02 3.35410565e-01 -6.67126179e-01 -4.34992790e-01
1.28196537e-01 1.46375179e+00 -5.46680868e-01 -2.61097282e-01
-8.17770839e-01 1.07562387e+00 1.14093751e-01 2.88763672e-01
-8.87099564e-01 -1.95193917e-01 1.27915776e+00 -6.48584440e-02
2.44826660e-01 8.78921524e-02 -1.62940815e-01 -8.04985285e-01
3.45153920e-02 -1.16497934e+00 3.85422617e-01 -5.62851012e-01
-1.36816490e+00 6.04264081e-01 2.67795986e-03 -1.23495996e+00
-3.88692707e-01 -1.15390778e+00 -7.09926307e-01 5.53393006e-01
-1.29113090e+00 -1.22890913e+00 -3.15075219e-01 6.40815973e-01
3.24897259e-01 -2.73931116e-01 1.05447876e+00 6.17991924e-01
-8.81412625e-01 1.05362749e+00 -6.38490498e-01 6.42589211e-01
6.66391850e-01 -7.30155051e-01 3.92858423e-02 7.50055790e-01
2.31061950e-01 7.09356487e-01 5.67401767e-01 -1.98010981e-01
-1.19829333e+00 -6.50184870e-01 4.66814429e-01 -3.72222394e-01
6.06686413e-01 -4.69410628e-01 -6.88367307e-01 5.82201540e-01
1.02959968e-01 1.09475687e-01 1.15804839e+00 2.06511915e-01
-3.74091476e-01 -1.02074288e-01 -1.25016642e+00 5.95485151e-01
1.07439661e+00 -7.10679591e-01 -2.57295012e-01 -3.84626873e-02
-3.44522297e-02 -8.69076252e-02 -8.61469626e-01 6.60559475e-01
8.06667864e-01 -1.16854346e+00 7.33898342e-01 -8.10106039e-01
7.13491321e-01 1.87786538e-02 -5.70010617e-02 -1.03033864e+00
-2.80127376e-01 -3.89084697e-01 -4.54140127e-01 1.24204636e+00
3.53725970e-01 -5.30258894e-01 9.49360192e-01 6.37156427e-01
1.20764375e-01 -9.27367568e-01 -1.24840081e+00 -6.47463739e-01
-1.53902546e-01 -3.80024970e-01 3.95881891e-01 9.14244950e-01
2.79801991e-02 3.83290350e-01 -3.85613680e-01 -2.42936164e-01
4.12844904e-02 -8.85283798e-02 8.40926528e-01 -1.09485662e+00
1.07967597e-03 -9.78221953e-01 -1.05121291e+00 -6.37350440e-01
4.22287405e-01 -5.50567269e-01 -1.72324419e-01 -8.93081307e-01
5.58745787e-02 1.79847345e-01 -4.43079948e-01 9.20949280e-01
-9.59234536e-02 5.98575890e-01 -4.89853472e-02 -5.51567599e-02
-2.96441883e-01 5.33918619e-01 1.05408311e+00 -2.00512484e-02
-6.41384348e-02 -1.45666122e-01 -5.80019772e-01 6.80450082e-01
7.85704970e-01 -1.01391166e-01 -1.47445828e-01 -4.15806323e-02
1.39700351e-02 -4.61250633e-01 8.28824788e-02 -1.05455804e+00
-3.06868494e-01 -1.17862247e-01 3.98195893e-01 -5.69132417e-02
5.38649380e-01 -6.51058555e-01 -4.79660600e-01 2.70789325e-01
-7.38685280e-02 -3.99100818e-02 6.75135672e-01 -8.04922208e-02
-5.86438596e-01 -1.51142463e-01 9.02593374e-01 2.81352662e-02
-1.05255890e+00 2.11683482e-01 -7.70107150e-01 -3.30349952e-01
1.19976914e+00 -5.54729640e-01 -6.58573136e-02 -4.67656404e-01
-5.50171316e-01 -1.92696616e-01 -2.50494871e-02 7.39729345e-01
4.19451237e-01 -1.47614217e+00 -7.06683815e-01 4.02215332e-01
3.50698382e-01 -7.93053687e-01 8.60489085e-02 1.10830891e+00
-1.93415508e-01 3.64572912e-01 -7.19557047e-01 -6.00917339e-01
-1.93474519e+00 2.21012726e-01 4.74975944e-01 6.64014369e-02
2.48599395e-01 1.18200719e+00 2.40429148e-01 -5.09481430e-02
2.83134997e-01 -2.38074556e-01 -5.36583424e-01 6.18377984e-01
7.96081722e-01 3.89929414e-01 1.40113443e-01 -8.99261415e-01
-5.40888965e-01 4.76843715e-01 -1.69464406e-02 4.39470857e-02
1.03377795e+00 3.70493561e-01 -4.94001806e-01 2.43824989e-01
1.37619281e+00 -1.93148911e-01 -7.68887043e-01 9.88321602e-02
-4.69329171e-02 -3.25646847e-01 5.12155099e-03 -6.84573948e-01
-1.16228914e+00 1.00014400e+00 8.14450145e-01 9.14272666e-02
1.45287204e+00 -8.71804133e-02 1.76405877e-01 3.10759991e-01
2.32313007e-01 -1.30925739e+00 3.78220826e-01 4.06039238e-01
1.03987420e+00 -1.32650125e+00 8.07461143e-02 -5.65416873e-01
-5.37939608e-01 1.20653081e+00 1.08882487e+00 4.98127267e-02
6.95630729e-01 3.83137643e-01 1.81195602e-01 -4.30408061e-01
-6.41250849e-01 -3.15464437e-01 3.25229317e-01 1.40405238e-01
9.60164726e-01 -1.11842845e-02 -4.90543246e-01 6.28956616e-01
8.44068453e-02 2.08583742e-01 1.34087741e-01 1.09870791e+00
-2.45802566e-01 -1.16935825e+00 -2.84994602e-01 6.02086365e-01
-7.48672664e-01 1.09007612e-01 -1.13778818e+00 7.16462433e-01
2.78408736e-01 8.98045182e-01 1.46970376e-01 -6.27651691e-01
3.29068065e-01 1.93556219e-01 6.67976677e-01 -6.97712302e-01
-7.44338572e-01 -2.76778936e-01 3.20560127e-01 -6.92204118e-01
-8.21585536e-01 -5.77044129e-01 -1.08190048e+00 -3.10287446e-01
-2.60585964e-01 -2.81905711e-01 4.03809905e-01 9.12484944e-01
4.92441773e-01 3.44626516e-01 6.10691667e-01 -6.25630081e-01
-2.93369710e-01 -1.59343493e+00 -5.17240226e-01 6.26486659e-01
2.96395496e-02 -9.78638351e-01 -5.08872509e-01 -9.08340365e-02] | [13.57273006439209, 1.7866452932357788] |
a5601d62-762e-4c2c-be67-3fd2b19184c9 | utilizing-graph-measure-to-deduce-omitted | null | null | https://aclanthology.org/C18-2011 | https://aclanthology.org/C18-2011.pdf | Utilizing Graph Measure to Deduce Omitted Entities in Paragraphs | This demo deals with the problem of capturing omitted arguments in relation extraction given a proper knowledge base for entities of interest. This paper introduces the concept of a salient entity and use this information to deduce omitted entities in the paragraph which allows improving the relation extraction quality. The main idea to compute salient entities is to construct a graph on the given information (by identifying the entities but without parsing it), rank it with standard graph measures and embed it in the context of the sentences. | ['Key-Sun Choi', 'Jiho Kim', 'Eun-Kyung Kim', 'Kijong Han'] | 2018-08-01 | utilizing-graph-measure-to-deduce-omitted-1 | https://aclanthology.org/C18-2011 | https://aclanthology.org/C18-2011.pdf | coling-2018-8 | ['relationship-extraction-distant-supervised'] | ['natural-language-processing'] | [ 2.80241102e-01 1.27806306e+00 -4.39478874e-01 -1.54317126e-01
-3.47554624e-01 -6.12635732e-01 5.17005980e-01 1.06664908e+00
-4.44180906e-01 9.10786211e-01 6.18829012e-01 -3.77333581e-01
-6.32948339e-01 -1.35686231e+00 -4.83627290e-01 -1.41362026e-01
-3.43014628e-01 6.61725879e-01 6.96347594e-01 -4.52723265e-01
1.53462648e-01 5.75303614e-01 -1.28207219e+00 3.73673260e-01
7.55845666e-01 5.00597596e-01 9.90293548e-02 4.60609674e-01
-5.73766530e-01 1.01401341e+00 -7.77470052e-01 -9.78532672e-01
-1.23850830e-01 -3.72653842e-01 -1.56092095e+00 -1.58468019e-02
3.00735533e-02 1.83823988e-01 -1.44900689e-02 1.13284898e+00
7.64778182e-02 -7.16341808e-02 6.72919452e-01 -9.48540926e-01
-1.14647955e-01 1.24906051e+00 -1.92559943e-01 4.85977888e-01
6.67280793e-01 -8.96190703e-01 1.61109912e+00 -7.05604494e-01
1.31053305e+00 1.10115325e+00 2.40917042e-01 2.73860782e-01
-7.82521963e-01 1.00389212e-01 -1.63165465e-01 2.64868051e-01
-8.92628133e-01 -1.73035905e-01 8.09104979e-01 -3.25674534e-01
1.05707622e+00 5.30416548e-01 5.21737814e-01 4.07870650e-01
-1.45844743e-01 4.90288764e-01 7.11333692e-01 -9.39336240e-01
1.29186660e-01 3.20841759e-01 9.77028430e-01 8.46322596e-01
1.07535529e+00 -2.94300526e-01 -3.42135251e-01 -1.54123411e-01
1.25978783e-01 -4.95594472e-01 -1.85420796e-01 -2.65290827e-01
-8.34713638e-01 7.82814562e-01 3.76611680e-01 7.66195238e-01
-5.13014495e-01 -2.51089573e-01 4.58275378e-01 1.14764675e-01
4.52739447e-01 6.89921260e-01 -6.29027545e-01 3.88780117e-01
-6.42504454e-01 1.64812848e-01 1.30851829e+00 9.93260205e-01
7.27663636e-01 -4.45379674e-01 1.43686766e-02 2.14271843e-01
6.76947609e-02 1.47388950e-01 2.75884159e-02 -4.10152465e-01
1.04282093e+00 1.16320837e+00 3.91634107e-02 -1.16887581e+00
-4.66071993e-01 -3.51522475e-01 -2.62980640e-01 -1.16862796e-01
2.57526070e-01 -3.83186966e-01 -5.41211069e-01 1.60810852e+00
6.21077120e-01 -2.25840807e-01 4.99845296e-01 3.07922751e-01
1.18029881e+00 3.66972685e-01 3.48567635e-01 -5.38908720e-01
1.51586914e+00 -5.80749750e-01 -9.85446095e-01 -3.88326421e-02
8.66331518e-01 -4.48242962e-01 2.23527536e-01 -5.88427782e-02
-9.12841618e-01 -1.23284526e-01 -1.18135762e+00 -2.13731542e-01
-9.40586984e-01 1.53414860e-01 6.71066582e-01 5.21931469e-01
-5.79911292e-01 7.94371128e-01 -3.89074892e-01 -2.39849716e-01
-2.50738338e-02 5.72142899e-01 -6.09707534e-01 2.90020108e-01
-1.52479804e+00 1.39817107e+00 1.11113703e+00 -6.80115893e-02
6.64062425e-02 -2.77858585e-01 -1.24229515e+00 2.83905268e-01
9.37814772e-01 -8.12152565e-01 6.50108874e-01 -6.81263566e-01
-6.74116015e-01 1.05707932e+00 -1.77083611e-01 -7.80158579e-01
4.19395007e-02 -1.39214456e-01 -7.20011711e-01 4.23317820e-01
4.36830848e-01 1.33556604e-01 3.75876755e-01 -1.13688672e+00
-7.48971343e-01 -4.66895610e-01 5.51115274e-01 4.98014838e-02
-1.15047082e-01 1.55141085e-01 -3.49565417e-01 -4.08809811e-01
2.69878894e-01 -5.05927145e-01 -7.30000809e-02 -1.10350621e+00
-7.85433292e-01 -4.88412380e-01 6.17936432e-01 -8.89146864e-01
1.50940597e+00 -1.86263943e+00 3.21276635e-01 6.50964797e-01
6.50955498e-01 9.57867317e-03 2.19085142e-01 6.92830205e-01
-3.74868095e-01 4.41032499e-01 -1.50879145e-01 1.70208842e-01
-7.45810196e-02 3.62791449e-01 -3.05887163e-01 1.56966336e-02
2.78186917e-01 9.19429660e-01 -8.98344517e-01 -8.52750719e-01
1.31066278e-01 1.08267188e-01 -2.49133676e-01 -1.38177544e-01
-3.69652331e-01 -4.13823634e-01 -7.78760791e-01 2.74630785e-01
3.62872988e-01 -1.34119787e-03 7.03981876e-01 -5.75191557e-01
-3.58761661e-02 7.41232157e-01 -1.46996379e+00 8.03934157e-01
-9.89140645e-02 4.65508699e-01 -3.15797389e-01 -9.01828468e-01
8.29402149e-01 2.94430315e-01 3.91675234e-01 -1.70069546e-01
2.15635404e-01 1.56784549e-01 -2.71478951e-01 -5.81865549e-01
6.92212760e-01 -9.91505384e-02 -2.53078192e-01 -1.05131246e-01
3.85192990e-01 1.37581170e-01 7.51812041e-01 7.24156857e-01
1.10881913e+00 -1.17804833e-01 8.41949999e-01 -3.80409777e-01
8.18699777e-01 3.26597869e-01 3.71611059e-01 3.82463723e-01
5.49417317e-01 -7.57323802e-02 1.15860271e+00 -3.34977627e-01
-8.80093932e-01 -6.23221636e-01 7.83713609e-02 5.76467216e-01
7.65987262e-02 -9.24798548e-01 -7.12126613e-01 -1.30840182e+00
-9.13828146e-03 7.39262044e-01 -7.97745705e-01 8.61933976e-02
-6.28247261e-01 -6.31523132e-01 2.00821429e-01 2.24887609e-01
1.96281761e-01 -1.00688362e+00 -6.02571368e-01 5.20464480e-02
-3.76043767e-01 -1.33696437e+00 3.20855498e-01 5.92620850e-01
-7.54598081e-01 -1.51691198e+00 8.15824047e-02 -8.61638665e-01
8.73193264e-01 -4.28891927e-01 1.42654705e+00 3.30516607e-01
-1.12907616e-02 -7.68411607e-02 -4.24599200e-01 -4.61891204e-01
-4.94265079e-01 2.03180641e-01 -4.37056422e-01 -3.93219978e-01
4.45316434e-01 -1.88094899e-01 2.98930585e-01 -3.98351967e-01
-8.47345233e-01 -1.58087060e-01 4.60347205e-01 4.55013782e-01
4.56117302e-01 6.50730073e-01 4.25252438e-01 -1.71478140e+00
6.90551102e-01 -3.79960477e-01 -5.82046628e-01 6.85826540e-01
-5.81448674e-01 7.45281100e-01 2.42686167e-01 2.11602911e-01
-1.02723753e+00 1.86843239e-03 1.25853373e-02 5.61753333e-01
1.43494681e-01 9.17071640e-01 -6.32267237e-01 2.49085933e-01
5.06828904e-01 -5.24819553e-01 -7.95959473e-01 -2.91287303e-01
6.09597206e-01 1.75042376e-01 3.32257807e-01 -6.48522258e-01
6.98062360e-01 7.34798312e-02 6.89425111e-01 -6.31829560e-01
-1.14217222e+00 -5.86865783e-01 -8.45190644e-01 1.46702170e-01
7.68415332e-01 -5.34042537e-01 -5.42746544e-01 -5.04694819e-01
-1.25451601e+00 4.15292323e-01 -6.69583797e-01 4.38091248e-01
-4.50700037e-02 5.72601914e-01 -5.13558984e-01 -8.19163978e-01
-2.89171249e-01 -4.76590216e-01 6.86837435e-01 1.70026779e-01
-3.63246083e-01 -1.05190015e+00 2.06505090e-01 9.12846029e-02
-4.28969592e-01 4.37249571e-01 1.30180931e+00 -1.18753695e+00
-4.68939126e-01 -3.80584240e-01 -2.87006289e-01 -4.91549559e-02
1.14197113e-01 1.01744249e-01 -4.89227682e-01 4.40234065e-01
-4.80229594e-03 3.84117365e-01 8.45604599e-01 8.84094760e-02
2.03529224e-01 -5.61640501e-01 -7.50186563e-01 -1.09367939e-02
1.83582091e+00 -1.70621634e-01 6.45105600e-01 4.81494278e-01
4.88581985e-01 9.56490219e-01 6.84382498e-01 -4.05914895e-02
2.85729647e-01 5.67881286e-01 3.20293605e-01 9.65282172e-02
-7.83330575e-02 -3.20003182e-01 -1.92662105e-01 6.03741109e-01
-2.44048879e-01 -3.66416186e-01 -7.69472122e-01 7.28014529e-01
-1.71488965e+00 -8.74559939e-01 -5.90971112e-01 1.74334288e+00
8.22965622e-01 5.34150362e-01 1.35868892e-01 4.32299227e-01
8.50408792e-01 6.21208083e-03 3.91110361e-01 -4.24684048e-01
-3.61716717e-01 3.75550061e-01 7.68212557e-01 1.02406120e+00
-9.15409863e-01 9.40589845e-01 6.25771379e+00 4.85087276e-01
-3.97989362e-01 -1.17822312e-01 1.14455372e-01 6.62627876e-01
-5.59480965e-01 5.67209423e-01 -1.18425560e+00 1.29021809e-03
1.01821828e+00 -4.96389180e-01 -2.81600118e-01 7.74269998e-01
-2.15545490e-01 -4.42239493e-01 -9.28654015e-01 1.62169933e-01
1.11284591e-01 -1.31936038e+00 2.84477621e-01 -2.04514712e-02
2.93934047e-01 -4.80741441e-01 -6.56261623e-01 5.13286628e-02
3.39743614e-01 -5.88500142e-01 4.69109058e-01 3.32507759e-01
1.55404717e-01 -1.01868892e+00 1.20543838e+00 4.10710722e-02
-1.22881186e+00 2.33612061e-01 -3.95746142e-01 4.02123109e-02
8.05212408e-02 8.04540396e-01 -1.11631584e+00 1.29493058e+00
1.96589425e-01 4.23716784e-01 -7.61830568e-01 8.23546290e-01
-8.61930788e-01 4.11081254e-01 -3.14354599e-01 -3.05028468e-01
-5.11791445e-02 -1.78975090e-01 6.78258598e-01 1.41013801e+00
9.79148895e-02 3.10957700e-01 -1.94398403e-01 5.94600022e-01
-3.44775945e-01 5.60614347e-01 -7.20939696e-01 6.35630861e-02
2.39930600e-01 1.28165865e+00 -1.06951547e+00 -8.18530500e-01
-3.96581262e-01 7.64123321e-01 4.96703148e-01 7.08082989e-02
-3.50553542e-01 -8.05010438e-01 -6.53134212e-02 6.79673329e-02
4.01184291e-01 -1.56036357e-03 -1.22059233e-01 -9.19927001e-01
3.04208040e-01 -3.58747870e-01 8.38804841e-01 -7.18223035e-01
-7.61283875e-01 7.77630687e-01 3.27298313e-01 -7.47352183e-01
-3.00058037e-01 -6.27630770e-01 -2.53858119e-01 7.40888059e-01
-1.32844937e+00 -9.61269081e-01 7.99868405e-02 4.51866627e-01
2.30869800e-02 3.26466620e-01 8.52558911e-01 6.05842359e-02
-3.30714524e-01 -8.50498974e-02 -5.37488639e-01 5.92564344e-01
1.34667322e-01 -1.62717247e+00 2.80481786e-01 1.15555894e+00
5.22768378e-01 6.86071873e-01 9.70533133e-01 -1.06538475e+00
-8.57184887e-01 -6.06736779e-01 2.02994394e+00 -8.11780930e-01
9.55581844e-01 1.51605541e-02 -6.64363742e-01 8.21085036e-01
3.64400655e-01 -2.28398532e-01 4.40672964e-01 3.20538104e-01
-3.75500470e-01 1.16351441e-01 -1.12964022e+00 3.60754043e-01
8.76840472e-01 -4.13040131e-01 -1.48250854e+00 2.35677987e-01
1.02703393e+00 -3.04690450e-01 -8.35500002e-01 6.26129925e-01
-9.33416709e-02 -4.77279067e-01 9.35788512e-01 -8.80303979e-01
2.65104741e-01 -3.00363243e-01 1.61523804e-01 -1.02654040e+00
-2.50858590e-02 -4.98499453e-01 -2.55690873e-01 1.39000070e+00
1.02629972e+00 -3.15488726e-01 7.93711543e-01 6.75202191e-01
2.18426719e-01 -3.14915210e-01 -7.20727026e-01 -3.66754740e-01
-4.80047703e-01 3.05127073e-02 5.26506484e-01 9.20158803e-01
7.46394455e-01 1.05673993e+00 -6.99602664e-02 3.59243214e-01
6.70388162e-01 4.13565993e-01 1.24616586e-01 -1.64651203e+00
-7.50298873e-02 8.10638294e-02 -5.30784428e-01 -2.91194052e-01
2.27633402e-01 -9.26753044e-01 -4.40304786e-01 -1.89182305e+00
2.67859437e-02 -1.51233539e-01 -1.08647466e-01 3.09210122e-01
-1.87433720e-01 -3.83358359e-01 -8.56683776e-03 -1.88958749e-01
-7.23538816e-01 -8.85628164e-02 8.44202042e-01 -2.72263549e-02
-1.69320136e-01 3.28159817e-02 -8.72011781e-01 8.56263995e-01
6.48707390e-01 -8.75240147e-01 -2.71589667e-01 8.89859125e-02
7.71464944e-01 2.78624147e-01 7.01682419e-02 -6.52141035e-01
2.96944976e-01 -1.60096108e-03 1.83003724e-01 -6.15880609e-01
-1.05423860e-01 -1.22402668e+00 2.18894765e-01 3.99389744e-01
-4.71297443e-01 1.16324827e-01 -7.17335939e-02 3.92805874e-01
-5.70752382e-01 -9.68597651e-01 1.81101024e-01 -1.44934312e-01
-7.29990304e-01 -2.55543947e-01 2.18077198e-01 3.32138598e-01
7.77199745e-01 1.72591001e-01 -5.00798583e-01 7.54751787e-02
-1.30678844e+00 -9.23624709e-02 1.44217074e-01 2.50672661e-02
4.46528018e-01 -9.65059578e-01 -5.44255495e-01 -3.91217530e-01
-1.91722705e-03 -1.54729337e-01 -2.74781018e-01 5.20527840e-01
-5.78724802e-01 3.59705985e-01 6.04825988e-02 2.85361499e-01
-1.81883180e+00 9.67418551e-01 2.51478981e-02 -8.19738090e-01
-5.10433495e-01 5.03470957e-01 -4.27684993e-01 1.33429438e-01
8.49619601e-03 -5.64041555e-01 -1.12231076e+00 4.96105850e-01
1.93212777e-01 1.66168392e-01 2.21803799e-01 -6.94987476e-01
-5.42859197e-01 4.69872385e-01 2.02963844e-01 -8.56295750e-02
1.33907509e+00 -1.05153665e-01 -5.58240592e-01 9.01549011e-02
9.15256560e-01 9.99548376e-01 -1.36689648e-01 -1.80237040e-01
9.77191329e-01 -1.95381027e-02 -2.37278864e-01 -4.51952875e-01
-7.34869719e-01 2.33911559e-01 -2.45322362e-01 7.63036013e-01
8.04308414e-01 5.39594173e-01 2.78611183e-01 6.39346242e-01
1.91711307e-01 -1.12096763e+00 -5.00559390e-01 4.11637455e-01
5.79425514e-01 -7.41330028e-01 5.18605053e-01 -1.47138965e+00
-5.62122047e-01 1.36137819e+00 2.93434132e-02 -3.97466086e-02
7.42853403e-01 4.09360290e-01 -3.79129022e-01 -8.70466232e-01
-4.53843266e-01 -8.46327960e-01 5.22174180e-01 4.39813763e-01
2.53129631e-01 5.96312806e-02 -1.00468028e+00 7.18230188e-01
-1.99636042e-01 -2.26864100e-01 6.59956872e-01 8.48813057e-01
-7.75170267e-01 -1.42737365e+00 1.45919740e-01 2.52722740e-01
-1.03034401e+00 -4.09454882e-01 -9.41253364e-01 9.97677922e-01
4.33835924e-01 9.39605713e-01 -4.72659945e-01 -3.23297024e-01
7.16942251e-01 1.25909373e-01 3.93553734e-01 -9.62045550e-01
-6.65903866e-01 -2.28235081e-01 1.05087423e+00 1.18961949e-02
-8.90180588e-01 -4.00493741e-01 -1.45336425e+00 3.82744111e-02
-6.95751846e-01 8.53445768e-01 5.34600973e-01 1.16442001e+00
4.63566147e-02 6.47291064e-01 3.88129234e-01 2.62594581e-01
2.18189314e-01 -6.13315403e-01 -6.21542096e-01 5.33103764e-01
1.58124179e-01 -5.47035336e-01 -4.61016506e-01 9.24069583e-02] | [9.275533676147461, 8.634690284729004] |
31bf46cc-bd50-4fa1-9d1f-6b07d9fa5a74 | depth-not-needed-an-evaluation-of-rgb-d | 1801.01235 | null | http://arxiv.org/abs/1801.01235v1 | http://arxiv.org/pdf/1801.01235v1.pdf | Depth Not Needed - An Evaluation of RGB-D Feature Encodings for Off-Road Scene Understanding by Convolutional Neural Network | Scene understanding for autonomous vehicles is a challenging computer vision
task, with recent advances in convolutional neural networks (CNNs) achieving
results that notably surpass prior traditional feature driven approaches.
However, limited work investigates the application of such methods either
within the highly unstructured off-road environment or to RGBD input data. In
this work, we take an existing CNN architecture designed to perform semantic
segmentation of RGB images of urban road scenes, then adapt and retrain it to
perform the same task with multichannel RGBD images obtained under a range of
challenging off-road conditions. We compare two different stereo matching
algorithms and five different methods of encoding depth information, including
disparity, local normal orientation and HHA (horizontal disparity, height above
ground plane, angle with gravity), to create a total of ten experimental
variations of our dataset, each of which is used to train and test a CNN so
that classification performance can be evaluated against a CNN trained using
standard RGB input. | ['Xiong Wei', 'Toby P. Breckon', 'Christopher J. Holder'] | 2018-01-04 | null | null | null | null | ['road-scene-understanding'] | ['computer-vision'] | [ 7.52024889e-01 2.21439078e-02 2.67728776e-01 -1.01050341e+00
-3.54739815e-01 -4.77104455e-01 7.13053346e-01 -3.87242138e-01
-6.49678230e-01 4.93559361e-01 -1.37330785e-01 -8.81868839e-01
7.59688467e-02 -1.13986695e+00 -7.20648408e-01 -5.43887734e-01
2.93739408e-01 5.72454512e-01 5.09081125e-01 -5.17145574e-01
4.77750480e-01 8.43129039e-01 -2.34380794e+00 1.60054862e-01
4.70958412e-01 1.24358904e+00 -5.75142242e-02 7.91684806e-01
-1.77814424e-01 4.82021183e-01 -2.30039015e-01 -1.27308503e-01
6.80014014e-01 2.44207438e-02 -9.15337205e-01 1.29219308e-01
7.91555285e-01 -5.29772520e-01 -4.53198045e-01 6.67068481e-01
3.85518640e-01 1.82438612e-01 3.72156501e-01 -1.36853945e+00
2.70397253e-02 -2.09746480e-01 -1.12629704e-01 2.92690068e-01
3.63759726e-01 3.39239329e-01 5.88343024e-01 -5.62291920e-01
5.52194953e-01 1.00220191e+00 8.21985126e-01 4.28385437e-01
-1.02620363e+00 -4.41157609e-01 -2.17345163e-01 3.35007846e-01
-9.74257171e-01 -3.42897832e-01 8.04176509e-01 -3.05128127e-01
1.26492560e+00 -5.22456579e-02 8.57894003e-01 8.43890548e-01
-1.63841039e-01 3.10667217e-01 1.48547733e+00 -3.52949202e-02
4.59514469e-01 -1.07792147e-01 2.68586781e-02 5.14985740e-01
2.67058611e-01 4.73409384e-01 -3.99155766e-01 3.48400921e-01
6.39162064e-01 -6.19888268e-02 -5.38483150e-02 -6.01155877e-01
-9.15452778e-01 8.52243304e-01 8.07004213e-01 -5.91550879e-02
-2.23507643e-01 3.11423451e-01 9.29620937e-02 3.84848416e-01
1.87462762e-01 -1.12865698e-02 -6.55939519e-01 -1.98247567e-01
-7.72613525e-01 4.85269219e-01 6.50910914e-01 9.39550400e-01
1.48739338e+00 -2.50006840e-02 3.89276683e-01 4.31938618e-01
4.68793333e-01 8.07141781e-01 2.96422720e-01 -1.24641013e+00
5.67780375e-01 8.46856415e-01 -1.89536259e-01 -8.19460452e-01
-7.48856246e-01 1.74946204e-01 -4.50314462e-01 7.73765147e-01
5.23958683e-01 -5.61734736e-02 -1.47295988e+00 1.10054517e+00
2.31957048e-01 3.67675759e-02 3.61120164e-01 9.48125184e-01
1.10536301e+00 3.36211324e-01 -2.87724167e-01 5.63774168e-01
9.82091546e-01 -6.28274262e-01 3.12773436e-02 -7.07413256e-01
6.25396371e-01 -6.69849753e-01 8.07651341e-01 3.41664374e-01
-5.46871364e-01 -6.48730099e-01 -1.28833389e+00 -4.41872627e-01
-1.04001749e+00 -3.26238960e-01 7.35230744e-01 7.97813296e-01
-1.31949353e+00 4.99289393e-01 -7.20829964e-01 -5.88126779e-01
5.60519457e-01 5.05639434e-01 -5.63660145e-01 -4.15586472e-01
-1.08281243e+00 1.01641810e+00 4.10133481e-01 3.32904994e-01
-7.36954749e-01 -5.12203395e-01 -1.09753406e+00 -5.75136662e-01
2.55246535e-02 -4.85073119e-01 1.20390522e+00 -8.97484481e-01
-1.71660900e+00 1.33493531e+00 -2.79281428e-03 -5.23236871e-01
4.99356598e-01 -1.69834182e-01 1.50161460e-02 1.29340604e-01
1.24055485e-03 1.02139437e+00 5.34827828e-01 -1.33986974e+00
-8.09002519e-01 -4.80625778e-01 3.02285135e-01 2.86399901e-01
2.40575999e-01 -4.95390981e-01 -3.74742299e-01 1.79554641e-01
4.26616639e-01 -1.02911806e+00 -4.25350815e-01 6.83759302e-02
-4.40740168e-01 2.94522256e-01 1.18385911e+00 -2.00937539e-01
1.08286701e-01 -1.83668208e+00 -9.02668834e-02 3.87862206e-01
-1.45411640e-01 3.81696790e-01 1.88668780e-02 2.49741226e-01
3.22040953e-02 -3.99612524e-02 -6.17564619e-01 -2.08919510e-01
-1.82935923e-01 6.78368926e-01 5.38686253e-02 4.56646591e-01
3.60974550e-01 1.06147480e+00 -6.27589703e-01 -3.08964103e-01
7.92258620e-01 6.33393824e-01 -3.72453392e-01 2.26365283e-01
-9.51457769e-02 6.42323434e-01 -4.66367126e-01 6.70442879e-01
9.13558304e-01 2.50896364e-01 -2.49876305e-01 -1.03349060e-01
-3.33677769e-01 3.31443727e-01 -9.88880157e-01 1.60000622e+00
-5.32655835e-01 1.24279630e+00 -1.74682304e-01 -9.93510842e-01
1.08178604e+00 -1.21854834e-01 4.79613036e-01 -1.30628741e+00
4.85288382e-01 2.96178341e-01 -1.62409827e-01 -5.16279995e-01
5.77491403e-01 -1.75198000e-02 9.09191519e-02 1.69195458e-01
-6.97372258e-02 -7.52409220e-01 -1.10373050e-01 -1.62733421e-01
1.05872810e+00 2.86555678e-01 -2.06368476e-01 8.81187171e-02
5.46337605e-01 4.95250136e-01 2.75377072e-02 6.50484443e-01
-3.14799249e-01 9.89907086e-01 3.03047866e-01 -9.56613004e-01
-1.32023895e+00 -1.01240313e+00 -2.97682524e-01 7.60462642e-01
5.28927565e-01 2.66820163e-01 -6.52186275e-01 -3.06723535e-01
1.53970525e-01 5.00424385e-01 -7.99979746e-01 -1.10961460e-01
-6.11299753e-01 -7.28627443e-01 4.47113037e-01 5.39451182e-01
1.12443912e+00 -1.07814837e+00 -1.41834474e+00 5.80143966e-02
7.71564618e-02 -1.42687607e+00 7.36351192e-01 4.94840831e-01
-8.36894751e-01 -1.44727921e+00 -3.77409399e-01 -5.57350993e-01
3.14635098e-01 5.73490143e-01 1.33034754e+00 1.59423575e-01
-4.98750746e-01 4.66732651e-01 -4.36002433e-01 -5.86151421e-01
-1.02217205e-01 1.17657878e-01 -5.08845329e-01 -5.27048968e-02
6.30162597e-01 -5.39410233e-01 -8.48965824e-01 4.77638632e-01
-9.72590327e-01 2.08628803e-01 5.05483925e-01 3.77537221e-01
5.58641553e-01 -2.56826133e-01 -2.61125952e-01 -5.94067454e-01
-1.29996523e-01 -3.62954646e-01 -7.60116994e-01 -4.23091978e-01
-4.21184093e-01 -3.09018612e-01 1.90916702e-01 2.20607266e-01
-8.95932913e-01 4.49450016e-01 -4.45612013e-01 -1.68653816e-01
-7.63380110e-01 1.94139197e-01 -7.90683776e-02 -5.78463197e-01
5.29837847e-01 1.18438385e-01 7.10262209e-02 -1.29480869e-01
4.26783025e-01 7.41816938e-01 5.84492445e-01 -1.73163891e-01
1.02249253e+00 9.56160605e-01 3.18894625e-01 -8.85856688e-01
-7.68525362e-01 -7.17435122e-01 -1.11120224e+00 -3.81432265e-01
9.24457788e-01 -7.32575238e-01 -4.53390449e-01 8.43090534e-01
-8.42571437e-01 -8.17085207e-01 -9.55789462e-02 2.77668774e-01
-9.55886126e-01 -1.87969387e-01 -5.07674692e-03 -4.76408809e-01
6.07049316e-02 -1.23713315e+00 1.37793612e+00 5.22157490e-01
3.39724422e-01 -8.84995997e-01 1.03475668e-01 5.91206193e-01
6.86501324e-01 6.88247383e-01 6.03402734e-01 -2.75369674e-01
-7.51384377e-01 -1.77540228e-01 -4.95994896e-01 1.75457552e-01
-2.97029000e-02 1.62379563e-01 -1.46062124e+00 2.30238914e-01
-2.48263031e-01 -5.78314185e-01 1.07638371e+00 3.23296338e-01
9.36956823e-01 3.70949864e-01 -2.34814227e-01 1.06819093e+00
1.73017371e+00 5.60927726e-02 1.10475123e+00 1.11804199e+00
8.41837883e-01 7.87850976e-01 5.88238597e-01 7.89103061e-02
6.98894858e-01 6.56216562e-01 1.04177713e+00 -4.84517783e-01
-3.15522216e-02 2.15521693e-01 -1.48414835e-01 -9.50123277e-03
-2.98302263e-01 -1.13075674e-01 -1.26244986e+00 5.28604031e-01
-1.52390492e+00 -7.20364273e-01 -4.23317462e-01 2.09832668e+00
3.84035945e-01 3.56131881e-01 5.72182238e-03 5.46633899e-01
1.91132069e-01 1.36884004e-01 -5.94216287e-01 -7.06838489e-01
-2.71379083e-01 3.00748646e-01 1.07485998e+00 3.19414496e-01
-1.35888577e+00 1.11756814e+00 6.46545982e+00 -4.66613621e-02
-1.50410688e+00 -3.63751501e-01 7.07869828e-01 3.48193616e-01
-2.37022996e-01 -2.49618124e-02 -5.89599669e-01 -5.08419238e-02
1.16788685e+00 3.45662564e-01 2.84252018e-01 8.87003064e-01
1.97878823e-01 -5.63895226e-01 -9.02011752e-01 8.05118918e-01
-1.56169236e-01 -1.09058845e+00 -9.87805203e-02 1.96411416e-01
8.21853697e-01 7.69380689e-01 2.92540882e-02 -6.33875141e-03
3.93366128e-01 -1.35668397e+00 5.72354555e-01 3.85867655e-01
6.54963136e-01 -6.40119731e-01 9.89010751e-01 4.09075022e-02
-1.09011245e+00 2.07987186e-02 -2.98916787e-01 -1.84827149e-01
4.47661057e-02 4.50812876e-02 -7.67033398e-01 4.84373569e-01
1.11716187e+00 9.19768155e-01 -8.42211485e-01 1.09719110e+00
-1.22219175e-01 2.14808956e-01 -5.05973220e-01 2.18246594e-01
6.40850186e-01 -1.13877915e-01 1.40651718e-01 9.38302934e-01
2.18049258e-01 2.46027089e-03 -2.44958252e-01 5.66343844e-01
2.63463080e-01 -1.88118115e-01 -1.11341012e+00 4.11453396e-01
1.56776503e-01 1.22276306e+00 -1.01272058e+00 -1.58329383e-01
-5.05687714e-01 7.59769320e-01 5.14876544e-02 5.06047249e-01
-7.79151142e-01 -5.18716991e-01 8.66764069e-01 1.61259118e-02
5.93706667e-01 -4.76571441e-01 -5.47006905e-01 -5.84710240e-01
-8.82449895e-02 -3.71568680e-01 -8.66281763e-02 -9.65445220e-01
-7.09073484e-01 6.88072562e-01 1.52425185e-01 -1.36848760e+00
-2.59873509e-01 -1.17723000e+00 -6.22059822e-01 7.78336465e-01
-2.22891569e+00 -1.02700794e+00 -1.02430880e+00 8.53549957e-01
3.38032782e-01 1.45214479e-02 6.08346760e-01 2.73287177e-01
-1.89977989e-01 1.05980150e-01 2.07745824e-02 1.47008672e-01
2.05339864e-01 -1.20087230e+00 6.67923868e-01 6.52386904e-01
-2.03531221e-01 6.09709546e-02 6.78021073e-01 -1.11878172e-01
-1.56926966e+00 -1.38639653e+00 5.39290190e-01 -4.61280614e-01
3.04030269e-01 -2.50903547e-01 -7.25291669e-01 6.10310256e-01
-8.04018006e-02 3.86995822e-01 2.43987307e-01 -3.84515464e-01
-4.68465090e-01 -1.85898662e-01 -1.38960910e+00 4.20084536e-01
1.11823368e+00 -4.82496262e-01 -4.31526691e-01 7.36328736e-02
4.91741747e-01 -7.28623927e-01 -6.21702373e-01 6.21142864e-01
6.08835280e-01 -1.67631733e+00 1.14974499e+00 -3.00565302e-01
5.32443404e-01 -4.10643160e-01 -4.56290513e-01 -1.12200105e+00
3.62811774e-01 -6.09834604e-02 4.89195049e-01 7.86732197e-01
2.97395706e-01 -8.05958509e-01 1.19968677e+00 6.03049219e-01
-5.38986027e-01 -4.91946250e-01 -1.25119615e+00 -4.11061078e-01
6.56818748e-02 -8.23736489e-01 5.75301468e-01 3.83026034e-01
-7.45115459e-01 -1.14597343e-01 1.50263891e-01 7.10523650e-02
6.05144143e-01 1.77412376e-01 1.19478893e+00 -1.34338534e+00
3.95769298e-01 -5.56259811e-01 -1.15909505e+00 -7.22226024e-01
-3.61795574e-02 -5.21193326e-01 4.53884542e-01 -1.73985672e+00
-3.70443523e-01 -6.17676437e-01 1.05478555e-01 5.75008035e-01
1.83717564e-01 8.81274343e-01 -2.39730269e-01 7.20833847e-03
-2.61387557e-01 5.27368724e-01 1.02535272e+00 -2.46402785e-01
-1.11376256e-01 5.95593415e-02 -2.62701124e-01 7.62529612e-01
9.82285798e-01 -1.54202729e-01 -4.86129254e-01 -4.99290586e-01
3.56717072e-02 -3.73117566e-01 7.71644473e-01 -1.49921966e+00
4.76592965e-02 -5.60036972e-02 4.82087761e-01 -7.07893193e-01
4.46176320e-01 -9.21203613e-01 4.66601690e-03 2.97271371e-01
-2.65503023e-02 -3.91286947e-02 4.33926821e-01 3.18400443e-01
-3.68738443e-01 2.87787709e-02 7.06523657e-01 -1.76754072e-01
-1.38277900e+00 2.92849332e-01 -4.95625198e-01 4.47675735e-02
1.06312883e+00 -1.08263242e+00 -3.66196483e-01 -1.77244991e-01
-2.99775958e-01 1.10357411e-01 7.64213383e-01 4.19992000e-01
8.76946628e-01 -1.15341139e+00 -2.95037031e-01 4.90832031e-01
4.33613896e-01 5.02614498e-01 -6.13729134e-02 5.24533808e-01
-1.20253253e+00 5.52754641e-01 -6.71991765e-01 -1.02430642e+00
-1.08062553e+00 3.35326381e-02 7.08761334e-01 4.23315167e-01
-8.02028537e-01 7.31843948e-01 -2.14786351e-01 -8.74309182e-01
7.34410435e-02 -3.21076572e-01 -2.93774098e-01 -1.99560121e-01
1.71352759e-01 9.81425345e-02 4.83284205e-01 -1.12202370e+00
-4.45407361e-01 1.12414181e+00 5.04263520e-01 -1.58753946e-01
1.26154006e+00 -2.56659776e-01 2.49747440e-01 3.23950142e-01
1.39198267e+00 -7.02264130e-01 -1.62900305e+00 -9.77080241e-02
3.28393616e-02 -5.13569951e-01 2.63061374e-01 -3.88860196e-01
-1.29887140e+00 1.16622138e+00 9.95403528e-01 1.70618773e-01
1.08074403e+00 -1.65160790e-01 7.54921257e-01 6.55387819e-01
5.51656604e-01 -1.02909482e+00 -1.13497302e-01 8.05488229e-01
4.26780343e-01 -1.71470845e+00 -1.37014568e-01 -3.08543503e-01
-3.01968157e-01 1.50406277e+00 6.32938266e-01 -1.13393001e-01
7.41107047e-01 1.62755642e-02 6.11762464e-01 -5.32040298e-01
-2.41781443e-01 -8.37952793e-01 1.13320522e-01 9.99516010e-01
1.92374766e-01 -2.60645509e-01 4.03870255e-01 -4.70729142e-01
-5.67090273e-01 3.37651670e-02 4.58102137e-01 1.23280835e+00
-6.14118993e-01 -8.77750754e-01 -3.07324141e-01 2.84494877e-01
1.07614964e-01 1.08823776e-01 -4.78690773e-01 1.17532027e+00
4.40636516e-01 1.08995426e+00 4.53453869e-01 -6.39107049e-01
6.58036530e-01 -1.33391768e-01 3.82892281e-01 -4.17605966e-01
-2.44729668e-01 -6.21951103e-01 1.90443546e-01 -1.02695823e+00
-9.43858087e-01 -7.71328628e-01 -1.32297373e+00 -1.76535100e-01
2.21865699e-01 -4.57722485e-01 1.29186285e+00 1.18047476e+00
-5.31256646e-02 3.39512080e-01 8.08216155e-01 -1.58027077e+00
3.42827767e-01 -5.64269602e-01 -3.05960387e-01 5.96470296e-01
5.51424384e-01 -9.18724716e-01 -3.89231563e-01 -2.20501218e-02] | [8.468727111816406, -2.2385003566741943] |
2b9616c2-fed8-4b80-95ed-6df109b937a4 | point-gcc-universal-self-supervised-3d-scene | 2305.19623 | null | https://arxiv.org/abs/2305.19623v2 | https://arxiv.org/pdf/2305.19623v2.pdf | Point-GCC: Universal Self-supervised 3D Scene Pre-training via Geometry-Color Contrast | Geometry and color information provided by the point clouds are both crucial for 3D scene understanding. Two pieces of information characterize the different aspects of point clouds, but existing methods lack an elaborate design for the discrimination and relevance. Hence we explore a 3D self-supervised paradigm that can better utilize the relations of point cloud information. Specifically, we propose a universal 3D scene pre-training framework via Geometry-Color Contrast (Point-GCC), which aligns geometry and color information using a Siamese network. To take care of actual application tasks, we design (i) hierarchical supervision with point-level contrast and reconstruct and object-level contrast based on the novel deep clustering module to close the gap between pre-training and downstream tasks; (ii) architecture-agnostic backbone to adapt for various downstream models. Benefiting from the object-level representation associated with downstream tasks, Point-GCC can directly evaluate model performance and the result demonstrates the effectiveness of our methods. Transfer learning results on a wide range of tasks also show consistent improvements across all datasets. e.g., new state-of-the-art object detection results on SUN RGB-D and S3DIS datasets. Codes will be released at https://github.com/Asterisci/Point-GCC. | ['Kaisheng Ma', 'Wenkai Shi', 'Zekun Qi', 'Guofan Fan'] | 2023-05-31 | null | null | null | null | ['unsupervised-3d-semantic-segmentation', '3d-instance-segmentation-1', '3d-semantic-segmentation', 'scene-understanding', 'deep-clustering', 'deep-clustering'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'miscellaneous', 'natural-language-processing'] | [-2.50665635e-01 -4.15201277e-01 -1.00089870e-01 -5.00928998e-01
-7.90619135e-01 -8.19904149e-01 7.29075253e-01 6.06141277e-02
-1.39605477e-01 -4.85973433e-02 -3.22499216e-01 -2.97071129e-01
-1.40081719e-01 -7.48712897e-01 -9.05674279e-01 -6.11749053e-01
-8.62970278e-02 6.28007829e-01 4.45427805e-01 -2.60061711e-01
4.23926592e-01 1.00892043e+00 -1.77330267e+00 1.69058740e-01
9.22613978e-01 1.28596997e+00 4.36984152e-01 4.57688481e-01
-3.14274371e-01 2.22688451e-01 -2.10316017e-01 5.14425300e-02
7.80759633e-01 -5.52807450e-02 -6.30858660e-01 2.34706730e-01
6.24396920e-01 -1.55400664e-01 -1.23741075e-01 1.16846538e+00
3.20543438e-01 -1.20689861e-01 6.66945696e-01 -1.46424174e+00
-6.10091746e-01 -7.52879679e-02 -7.04800427e-01 2.86784917e-01
-3.84474583e-02 5.78461647e-01 1.00606596e+00 -9.76567805e-01
2.80973792e-01 1.21752632e+00 6.91276729e-01 1.81398556e-01
-1.04621661e+00 -5.97079694e-01 2.68540949e-01 1.47686094e-01
-1.29913104e+00 -8.37990374e-04 1.08539963e+00 -5.63049078e-01
8.93785894e-01 2.09309474e-01 7.51090825e-01 9.09234881e-01
-2.79899567e-01 6.69313669e-01 1.25367582e+00 -2.41246119e-01
2.15219170e-01 1.15998246e-01 5.63139655e-02 4.87655461e-01
3.11011642e-01 3.42348158e-01 -4.86623496e-01 1.04921848e-01
8.09410691e-01 1.60014078e-01 -9.17441249e-02 -8.71177137e-01
-1.28705096e+00 7.30302393e-01 1.00608635e+00 6.13654554e-02
-7.99723342e-03 4.24111485e-02 7.79691041e-02 1.35385409e-01
6.54551506e-01 4.47561800e-01 -5.40819287e-01 1.89753413e-01
-7.78599977e-01 1.30491063e-01 3.14837247e-01 1.22979856e+00
1.20477211e+00 -2.09587857e-01 2.74308324e-01 5.11261642e-01
6.92441285e-01 7.41229594e-01 -1.86372604e-02 -1.04125118e+00
3.81437570e-01 9.71839070e-01 -2.73501668e-02 -7.48812795e-01
-4.85919058e-01 -6.84640408e-01 -5.89517057e-01 7.05112219e-01
3.33479255e-01 3.83760363e-01 -1.11863804e+00 1.40807426e+00
7.78104365e-01 2.41546050e-01 -1.82592407e-01 1.24868774e+00
8.96171629e-01 3.49706471e-01 -1.49118707e-01 6.53053522e-01
1.27556682e+00 -8.48865151e-01 1.41625389e-01 -4.50007319e-01
6.30119324e-01 -5.30169606e-01 1.33832598e+00 6.23565279e-02
-8.73795390e-01 -7.42775142e-01 -1.15206707e+00 -3.14590424e-01
-6.86518908e-01 2.65437365e-01 6.63838029e-01 3.57280701e-01
-9.88238990e-01 6.07289970e-01 -1.04334927e+00 -5.62196434e-01
6.70134783e-01 1.79468468e-01 -2.52235353e-01 2.09952500e-02
-6.95662200e-01 8.17483485e-01 2.52269089e-01 5.36568463e-02
-7.31297672e-01 -1.10273612e+00 -7.32075870e-01 -9.21031013e-02
1.74423620e-01 -1.03452182e+00 1.05260766e+00 -7.87103534e-01
-1.26047027e+00 1.36073077e+00 2.33056292e-01 -7.91466460e-02
5.52727163e-01 -1.61591142e-01 7.00478926e-02 3.08773011e-01
3.38397145e-01 7.52125561e-01 7.60854304e-01 -1.73097730e+00
-7.59166896e-01 -6.93414450e-01 -2.40401197e-02 4.39645022e-01
-1.92554127e-02 -2.25515679e-01 -7.67358661e-01 -2.51107633e-01
4.87100959e-01 -1.01665211e+00 -9.08083748e-03 4.57482427e-01
-4.24223214e-01 -1.75163671e-01 8.47858250e-01 -9.23621282e-02
3.63994420e-01 -2.32973456e+00 -6.01769099e-03 1.43742248e-01
2.49220580e-01 9.63659510e-02 -2.52024859e-01 2.36447960e-01
-2.64773309e-01 2.03642510e-02 -3.75322521e-01 -5.07695079e-01
1.75846472e-01 5.99715412e-02 -2.52416432e-01 7.19955266e-01
5.46422303e-01 9.64738786e-01 -8.13731670e-01 -3.17691594e-01
5.60234964e-01 3.74451578e-01 -5.89081883e-01 2.91867614e-01
-4.26093787e-01 6.75167501e-01 -5.75545073e-01 9.90573466e-01
1.04944599e+00 -6.83073342e-01 -4.09355223e-01 -4.09953117e-01
-3.57197940e-01 5.08534431e-01 -9.69434381e-01 1.96815503e+00
-2.53758550e-01 4.64256644e-01 2.25883752e-01 -8.25319052e-01
9.52057958e-01 -3.95210505e-01 4.93269891e-01 -6.31802022e-01
1.66574761e-01 1.74868092e-01 -2.27598727e-01 -2.51264602e-01
2.32226238e-01 -5.56213334e-02 1.29271582e-01 2.50041097e-01
1.17472216e-01 -6.84296489e-01 -2.95271426e-01 3.24320138e-01
8.51100922e-01 4.68909740e-01 -3.67708914e-02 -3.64666015e-01
2.57300735e-01 4.55440313e-01 3.43290508e-01 6.94549024e-01
-1.86585113e-01 9.56160545e-01 2.25965649e-01 -2.10957795e-01
-1.05341756e+00 -1.16977882e+00 -3.78817499e-01 8.70256841e-01
5.96165597e-01 -2.39978850e-01 -2.60189712e-01 -5.81120729e-01
6.62841678e-01 5.06894350e-01 -6.03614211e-01 -1.25757590e-01
-1.83672249e-01 -6.33607686e-01 2.74298638e-01 5.74485481e-01
5.61465323e-01 -5.86340249e-01 -5.15352428e-01 -3.50049973e-01
2.14283198e-01 -1.29466987e+00 -1.16036318e-01 4.67599362e-01
-1.06536984e+00 -1.07073689e+00 -2.50537395e-01 -4.52383667e-01
5.61243534e-01 9.53404665e-01 1.32511640e+00 2.94409394e-01
-9.13634747e-02 6.88179910e-01 -5.11628151e-01 -5.89754045e-01
-7.74621889e-02 2.05662534e-01 -3.05753089e-02 -1.92777961e-01
4.94847894e-01 -7.71682501e-01 -7.27433324e-01 3.96949530e-01
-6.63229406e-01 7.07078055e-02 7.55325675e-01 3.43406081e-01
7.27097213e-01 -3.71503532e-01 -6.76713437e-02 -5.57036996e-01
-1.48360968e-01 -5.10877669e-01 -8.95690441e-01 -1.56832002e-02
-5.91414869e-01 -7.45839104e-02 2.17726454e-01 7.10858032e-02
-8.35322857e-01 2.78044432e-01 -3.59294750e-02 -9.12588000e-01
-4.30247217e-01 2.43066460e-01 -3.44679326e-01 -3.58424634e-01
6.69028819e-01 1.04977816e-01 1.40651911e-01 -6.51738644e-01
5.04154801e-01 5.82868576e-01 6.36054337e-01 -6.44945323e-01
1.35366821e+00 9.84709442e-01 7.87556916e-02 -6.60806298e-01
-1.07154799e+00 -9.00532126e-01 -1.00001919e+00 -1.58119395e-01
8.67238402e-01 -1.33010626e+00 -7.65793025e-01 4.06481802e-01
-1.04816341e+00 -5.49328983e-01 -2.45207161e-01 4.63006079e-01
-4.88520622e-01 4.07785386e-01 -3.91630173e-01 -5.92598379e-01
-1.73336212e-02 -1.01670337e+00 1.65855670e+00 8.78541097e-02
3.72656852e-01 -7.69073725e-01 8.04781318e-02 3.76495004e-01
2.01115489e-01 2.64876336e-01 7.53750443e-01 -4.19845521e-01
-1.20214868e+00 4.19058986e-02 -7.46723890e-01 3.90965939e-01
7.08900690e-02 1.11378238e-01 -1.23634243e+00 -1.13535196e-01
2.58908924e-02 -3.22687924e-01 1.04536867e+00 1.60029203e-01
1.25002646e+00 3.11866045e-01 -3.21858227e-01 1.13337088e+00
1.53034389e+00 -3.29579681e-01 4.32914734e-01 6.87732041e-01
1.13004732e+00 5.54107189e-01 6.82150126e-01 2.81549603e-01
6.73514187e-01 6.13822222e-01 9.38953400e-01 -3.67990226e-01
-2.90343821e-01 -3.00929546e-01 9.02735516e-02 5.52934170e-01
-1.11763828e-01 1.49365529e-01 -1.24612796e+00 4.21574980e-01
-1.76676702e+00 -6.66170895e-01 -4.48767692e-01 2.01042676e+00
6.27773702e-01 2.67142981e-01 1.31814733e-01 -1.14536501e-01
5.02004683e-01 1.45784870e-01 -8.19847763e-01 2.89685756e-01
-2.47527137e-01 1.23758409e-02 7.00091779e-01 1.92613497e-01
-1.30290794e+00 1.02023590e+00 5.19793272e+00 5.76106906e-01
-1.20116866e+00 2.08663996e-02 3.87616992e-01 -6.02014139e-02
-3.86849493e-01 1.21653378e-01 -8.04583251e-01 3.81608963e-01
5.37530303e-01 2.40724847e-01 1.43963814e-01 1.14827836e+00
9.35848504e-02 2.36603916e-02 -1.22573900e+00 1.18527055e+00
-9.25589576e-02 -1.32731247e+00 -2.04162747e-02 9.17055234e-02
5.85530102e-01 8.39825690e-01 9.39375758e-02 9.27527547e-02
2.69168943e-01 -5.72291136e-01 1.08219492e+00 4.15108830e-01
7.32462883e-01 -3.40347767e-01 3.03502858e-01 2.73939371e-01
-1.24517918e+00 8.49342123e-02 -4.97674435e-01 7.72944614e-02
-1.10767886e-01 7.31759012e-01 -5.88138998e-01 8.83610129e-01
1.19426084e+00 1.18002689e+00 -9.56291318e-01 1.18871331e+00
-3.53874356e-01 3.48136216e-01 -6.59441292e-01 2.29251519e-01
3.44147146e-01 -5.07856548e-01 6.17445707e-01 1.22886932e+00
3.20113391e-01 -2.06148140e-02 3.05814967e-02 1.31367648e+00
4.40036990e-02 -3.24595839e-01 -6.63473248e-01 2.62589574e-01
4.87196714e-01 1.37829030e+00 -8.47086012e-01 -1.83571070e-01
-5.64840078e-01 7.75427818e-01 4.39589351e-01 1.95192769e-01
-8.28192234e-01 -5.40227070e-02 1.05603909e+00 2.41609260e-01
5.71519017e-01 -6.74458444e-01 -7.42004514e-01 -1.22314596e+00
1.47010107e-02 -5.76250494e-01 2.78237611e-01 -1.19654858e+00
-1.66004908e+00 3.80123973e-01 3.50119993e-02 -1.81979001e+00
2.98731685e-01 -9.95992064e-01 -7.40673959e-01 8.39420080e-01
-2.04443073e+00 -1.32756996e+00 -8.70314360e-01 5.96556246e-01
1.15180559e-01 9.04976055e-02 3.28988671e-01 1.70404539e-01
-3.63351822e-01 9.84361172e-02 4.39861231e-02 8.97652507e-02
7.55617201e-01 -1.57315445e+00 6.48316383e-01 8.57865572e-01
2.13273004e-01 5.23593009e-01 3.02897781e-01 -4.74665880e-01
-1.68123984e+00 -1.26382542e+00 2.48235166e-01 -1.04775512e+00
6.27524018e-01 -5.36638916e-01 -1.02428091e+00 5.66288590e-01
-1.05309919e-01 2.72486091e-01 4.79154229e-01 1.05177723e-01
-6.70798004e-01 -2.48307988e-01 -8.85898352e-01 2.89203405e-01
1.46378708e+00 -7.50092745e-01 -6.06309056e-01 4.87189800e-01
7.63551772e-01 -5.90064228e-01 -6.80707276e-01 5.83851933e-01
2.36315876e-02 -1.21943879e+00 1.29350996e+00 -4.44789648e-01
4.07645822e-01 -7.89096534e-01 -1.98343605e-01 -1.15375245e+00
-4.73413795e-01 -1.52729079e-01 1.20934539e-01 1.04885948e+00
2.78303474e-01 -5.28244495e-01 9.27745998e-01 4.78976101e-01
-6.80994153e-01 -4.96367723e-01 -7.85088480e-01 -9.84816253e-01
9.95843634e-02 -6.83504641e-01 7.46576905e-01 1.13777995e+00
-5.52730620e-01 3.44869524e-01 3.24743271e-01 7.46863902e-01
9.15642858e-01 7.44577646e-01 1.09957123e+00 -1.46716475e+00
-1.28795192e-01 -6.96250856e-01 -4.73531097e-01 -1.22955787e+00
-2.12872270e-02 -1.26978362e+00 -3.97709422e-02 -1.44101489e+00
7.23143294e-02 -8.41388702e-01 -1.52759209e-01 4.71938401e-01
-2.19732210e-01 3.31720889e-01 3.65428895e-01 6.46785378e-01
-7.38201141e-01 9.03318226e-01 1.05822682e+00 -1.40796512e-01
-1.58239588e-01 -1.02560818e-01 -6.32720947e-01 6.56285107e-01
8.74532819e-01 -3.82585377e-01 -1.29623592e-01 -8.06312263e-01
1.24794267e-01 -4.39007580e-01 8.84692252e-01 -1.10104477e+00
2.74847597e-01 -1.09782025e-01 5.74400663e-01 -1.02787519e+00
3.38216841e-01 -8.83247852e-01 -1.49785906e-01 1.12507477e-01
1.16923526e-01 -8.48290548e-02 3.18531573e-01 6.17158175e-01
-1.09671116e-01 1.41760096e-01 9.07123625e-01 -2.00541511e-01
-8.82599235e-01 5.55746615e-01 3.45422119e-01 -6.95916936e-02
8.43796551e-01 -3.70555431e-01 -4.38444287e-01 1.54187987e-02
-3.25573057e-01 4.08131838e-01 9.77360845e-01 3.66824448e-01
5.34755707e-01 -1.31310689e+00 -5.08035600e-01 3.10126126e-01
6.40640378e-01 6.94243133e-01 4.66348939e-02 7.77692616e-01
-6.27917290e-01 2.63516575e-01 -2.45432675e-01 -1.33061957e+00
-8.44470978e-01 5.81130505e-01 4.60820973e-01 2.33663797e-01
-7.40089834e-01 8.72338057e-01 3.97369504e-01 -7.68632591e-01
3.82544333e-03 -5.33246279e-01 2.43016586e-01 -1.84608340e-01
1.45655945e-01 7.16693550e-02 3.72419745e-01 -4.73680615e-01
-6.04382336e-01 9.00872588e-01 2.29020342e-01 1.42291859e-01
1.53747165e+00 -1.72068849e-01 -1.47160947e-01 5.26640713e-01
1.27964211e+00 -2.33832300e-01 -1.70197046e+00 -2.97988713e-01
4.59844284e-02 -7.87927747e-01 2.43250504e-01 -6.89342141e-01
-1.12546659e+00 1.05705345e+00 6.90737963e-01 1.70767859e-01
1.01346147e+00 4.47597891e-01 2.88196474e-01 2.77571172e-01
3.40446144e-01 -6.82292819e-01 1.03090435e-01 5.03856957e-01
7.86480725e-01 -1.69437647e+00 1.28741577e-01 -4.91229296e-01
-4.41730171e-01 1.01087105e+00 7.44931519e-01 -3.30764264e-01
6.92165434e-01 8.90087038e-02 2.25140005e-01 -5.79402268e-01
-5.04477978e-01 -7.82574892e-01 5.26950657e-01 6.72290921e-01
4.62610386e-02 -1.54511586e-01 4.66628402e-01 4.14061844e-01
-1.91922307e-01 -4.59335685e-01 3.08596157e-02 8.97153378e-01
-4.54695642e-01 -9.32833731e-01 -5.09625554e-01 2.48282775e-01
3.30326468e-01 3.12537886e-02 -6.18216157e-01 9.19138908e-01
1.25011295e-01 6.02653444e-01 2.63679922e-01 -4.21730578e-01
4.69257921e-01 -2.24853322e-01 5.48320174e-01 -7.75175869e-01
-3.77417892e-01 -3.34767848e-02 -3.12613547e-01 -7.97611117e-01
-6.52158916e-01 -6.79304540e-01 -1.21380830e+00 -2.98336565e-01
-2.48047888e-01 -1.15052380e-01 8.71983230e-01 6.20768905e-01
7.30217993e-01 3.00144911e-01 8.00331891e-01 -1.52270007e+00
-5.35358727e-01 -7.56018341e-01 -5.97147703e-01 5.26752889e-01
3.91626596e-01 -8.95526111e-01 -6.56373322e-01 -2.20180571e-01] | [7.985340118408203, -3.1419389247894287] |
085e29ad-31f8-4760-a8e9-b707dca35d4c | detecting-adversarial-attacks-on-audio-visual | 1912.08639 | null | https://arxiv.org/abs/1912.08639v2 | https://arxiv.org/pdf/1912.08639v2.pdf | Detecting Adversarial Attacks On Audiovisual Speech Recognition | Adversarial attacks pose a threat to deep learning models. However, research on adversarial detection methods, especially in the multi-modal domain, is very limited. In this work, we propose an efficient and straightforward detection method based on the temporal correlation between audio and video streams. The main idea is that the correlation between audio and video in adversarial examples will be lower than benign examples due to added adversarial noise. We use the synchronisation confidence score as a proxy for audiovisual correlation and based on it we can detect adversarial attacks. To the best of our knowledge, this is the first work on detection of adversarial attacks on audiovisual speech recognition models. We apply recent adversarial attacks on two audiovisual speech recognition models trained on the GRID and LRW datasets. The experimental results demonstrate that the proposed approach is an effective way for detecting such attacks. | ['Stavros Petridis', 'Pingchuan Ma', 'Maja Pantic'] | 2019-12-18 | null | null | null | null | ['audio-visual-speech-recognition'] | ['speech'] | [ 1.69630185e-01 -1.24304451e-01 2.90933937e-01 1.19781315e-01
-1.13490045e+00 -9.86859381e-01 7.85606563e-01 1.32057471e-02
-2.68895715e-01 4.75311011e-01 1.83200184e-02 -1.80169106e-01
1.48845464e-01 -6.01712525e-01 -9.10269916e-01 -7.73787856e-01
-4.39915061e-01 -3.47208194e-02 5.37235081e-01 -2.04642773e-01
1.18920114e-02 6.22063458e-01 -1.35821223e+00 4.98952657e-01
2.37243965e-01 1.07337165e+00 -3.79116327e-01 1.05704594e+00
2.36992642e-01 9.23963964e-01 -1.11274540e+00 -6.02239251e-01
4.13114786e-01 -2.93582827e-01 -6.50794804e-01 -2.51673877e-01
5.20486355e-01 -5.19885421e-01 -7.79851198e-01 1.30193281e+00
7.64181912e-01 7.32320827e-03 5.90755939e-01 -1.58343434e+00
-3.23887527e-01 8.76235068e-01 -3.73833150e-01 6.07898057e-01
5.96146703e-01 8.06033686e-02 7.01996446e-01 -5.44394553e-01
2.31124967e-01 1.42983043e+00 5.83944380e-01 6.08286619e-01
-8.61693621e-01 -1.16436923e+00 1.48625299e-01 6.00287199e-01
-1.40169871e+00 -5.30640364e-01 9.40605223e-01 -3.49657148e-01
5.68559587e-01 5.00679076e-01 2.81649411e-01 1.73115098e+00
1.18585654e-01 8.02869499e-01 9.95083690e-01 -4.94138747e-01
1.18565120e-01 2.00082973e-01 -4.97063696e-01 3.41527343e-01
-2.87967205e-01 5.50834835e-01 -5.79906940e-01 -5.23883224e-01
4.26705360e-01 -3.05917144e-01 -2.97962755e-01 4.58881073e-02
-6.59988165e-01 1.02545500e+00 1.84947491e-01 4.84299242e-01
-5.86980060e-02 3.33351582e-01 7.48965979e-01 7.00581491e-01
3.66075069e-01 2.00141385e-01 -1.23888008e-01 -7.68044665e-02
-7.12758362e-01 4.34569418e-02 6.88554764e-01 5.77350140e-01
1.64400622e-01 5.08740008e-01 7.54379705e-02 7.30595469e-01
4.24685866e-01 5.70849597e-01 5.07593334e-01 -6.05596185e-01
4.88807708e-01 -2.61888623e-01 -2.92269796e-01 -1.22237647e+00
1.84526457e-03 -1.32644057e-01 -5.68186820e-01 5.59696674e-01
3.97352278e-01 -2.14622810e-01 -5.96849799e-01 1.61627388e+00
1.71833396e-01 7.45965719e-01 2.94106394e-01 6.67262256e-01
7.62276590e-01 5.78604043e-01 3.21699083e-02 -2.38863751e-01
9.36318576e-01 -5.87799132e-01 -7.76435435e-01 4.07950953e-02
3.02563071e-01 -1.26614308e+00 5.64222991e-01 6.05940342e-01
-8.99514794e-01 -3.94815773e-01 -1.26514935e+00 6.64064109e-01
-4.25540656e-01 -5.21412432e-01 3.68631214e-01 1.23336315e+00
-6.25734270e-01 4.10533696e-01 -7.37186313e-01 1.30242249e-03
2.82435745e-01 2.09693760e-01 -3.85481298e-01 9.19347536e-03
-1.69600773e+00 7.15383053e-01 2.53679007e-01 -2.30538353e-01
-1.48578322e+00 -3.48571599e-01 -7.09202230e-01 -2.16175646e-01
2.29938358e-01 1.06319316e-01 1.43217504e+00 -1.25768697e+00
-1.33200192e+00 7.01186597e-01 3.01843494e-01 -8.11494768e-01
8.58716071e-01 -1.19051255e-01 -8.80073905e-01 4.75217879e-01
-3.31726074e-01 4.39488962e-02 1.50852621e+00 -1.28105640e+00
-4.16293055e-01 -1.48242012e-01 1.63179457e-01 -9.33587998e-02
-6.44503236e-01 5.19797504e-01 -1.59560621e-01 -1.06722081e+00
-3.01112831e-01 -8.03764403e-01 1.22934595e-01 -1.59628689e-01
-2.88415521e-01 2.06601068e-01 1.18339801e+00 -5.81327140e-01
1.03992808e+00 -2.39497185e+00 -2.66777396e-01 3.82931679e-01
-1.51289636e-02 6.24661088e-01 -1.69673726e-01 5.26679516e-01
-2.62236804e-01 2.64837563e-01 1.42224833e-01 -2.62315750e-01
3.97718623e-02 1.73087105e-01 -9.46814358e-01 8.24684322e-01
1.21543482e-01 3.48765880e-01 -7.65949190e-01 -4.00402814e-01
2.49956936e-01 6.69558048e-01 -3.50814998e-01 3.11106175e-01
1.53956637e-01 2.79917300e-01 -2.73942530e-01 6.28462493e-01
7.64442861e-01 7.31489360e-01 -1.48282528e-01 -1.83407869e-02
3.56867164e-01 5.68764061e-02 -1.33848894e+00 9.61310923e-01
-3.27146083e-01 9.21235383e-01 1.10497482e-01 -1.08455718e+00
8.39434206e-01 9.50738132e-01 2.93543547e-01 -4.06377673e-01
2.98718244e-01 1.58243492e-01 9.67967361e-02 -3.79394799e-01
-9.03485995e-03 -1.35552585e-01 -9.76596475e-02 1.84969842e-01
2.66310811e-01 -8.17045197e-02 -2.01204151e-01 2.33079925e-01
1.38620067e+00 -3.82267922e-01 3.17961365e-01 3.85744333e-01
7.05601811e-01 -6.89523518e-01 3.17476958e-01 1.01155269e+00
-4.55633759e-01 4.76933360e-01 4.24804956e-01 -6.47797957e-02
-8.85425627e-01 -1.31453991e+00 -7.04038367e-02 9.72588539e-01
-1.97987661e-01 -4.95227516e-01 -6.81972265e-01 -8.93335640e-01
-6.25056103e-02 3.75523537e-01 -4.70344365e-01 -3.46364439e-01
-5.40313363e-01 -3.06205660e-01 1.37616253e+00 5.93395710e-01
2.93481439e-01 -1.00839686e+00 7.11962283e-02 1.29602790e-01
2.80467980e-02 -1.50023437e+00 -4.10377055e-01 6.83152676e-02
-4.08192992e-01 -1.10712767e+00 -4.76348847e-01 -6.86833143e-01
1.52124882e-01 1.27137676e-01 6.54928684e-01 -7.84495473e-03
-4.55911189e-01 5.20253479e-01 -6.15308583e-01 -7.03429639e-01
-1.08266664e+00 -5.91152787e-01 3.90391201e-01 1.62282705e-01
9.27862749e-02 -6.52566791e-01 -9.72167403e-02 5.50695360e-01
-1.08347070e+00 -8.62701237e-01 1.26680389e-01 6.87770903e-01
2.61058033e-01 4.78898585e-01 5.54989517e-01 -5.03967404e-01
6.54551625e-01 -4.70858783e-01 -6.43526316e-01 -4.37384322e-02
-4.65998426e-03 -2.17971891e-01 1.04009330e+00 -9.82698441e-01
-6.65659249e-01 8.39428306e-02 -5.65858781e-01 -9.53941107e-01
-5.04406571e-01 2.40723118e-01 -3.64035726e-01 -5.37743211e-01
7.27584243e-01 2.66589243e-02 -3.72957468e-01 -2.29933023e-01
3.15042943e-01 6.68512464e-01 6.76289797e-01 -2.34647095e-01
1.21366477e+00 3.92648488e-01 -2.21360363e-02 -9.98984158e-01
-3.00908417e-01 -3.86861414e-01 -2.86824465e-01 -5.78433573e-01
6.05115235e-01 -8.05386424e-01 -8.12893331e-01 7.55880415e-01
-1.28338039e+00 4.60432805e-02 1.82558209e-01 6.59570217e-01
-5.43003976e-01 8.24194551e-01 -6.87394619e-01 -1.22085261e+00
-1.24271646e-01 -1.13273084e+00 7.31610775e-01 -3.65936935e-01
-3.91198285e-02 -9.67830002e-01 6.09325394e-02 2.09386975e-01
1.79261327e-01 5.23076475e-01 3.00155014e-01 -1.10757196e+00
-4.24189776e-01 -6.67393565e-01 3.22826505e-01 6.95610464e-01
2.32362263e-02 2.48920709e-01 -1.43419576e+00 -3.66881788e-01
3.97742748e-01 -6.05263829e-01 8.49524856e-01 9.20117460e-03
1.10912502e+00 -6.24092340e-01 -5.12425043e-02 4.33757126e-01
1.17601490e+00 4.73499864e-01 8.56572628e-01 3.25198174e-01
5.47432661e-01 2.47535050e-01 6.60802364e-01 5.28653562e-01
-4.56333369e-01 9.48479295e-01 9.52909946e-01 8.73080641e-02
3.77433449e-02 -4.43727374e-02 1.01119363e+00 6.96114898e-01
1.15239359e-01 -5.39616048e-01 -8.58405471e-01 2.78585255e-01
-1.60340810e+00 -1.54968750e+00 -4.47622407e-03 2.21203327e+00
6.40405595e-01 5.07034302e-01 1.31644011e-01 7.62815654e-01
8.09197247e-01 3.92093867e-01 -6.40967339e-02 -5.97193599e-01
-1.34756476e-01 4.95639443e-01 5.78447104e-01 5.56035161e-01
-1.64905667e+00 8.34236562e-01 6.50998878e+00 1.11849499e+00
-1.10392213e+00 2.52638072e-01 1.85485139e-01 -2.21736908e-01
2.56862074e-01 -4.82731432e-01 -6.20423257e-01 5.07277310e-01
9.66448724e-01 -6.76140338e-02 3.47273111e-01 9.20050979e-01
-2.12291609e-02 4.82565314e-01 -9.44673300e-01 8.72267246e-01
3.93463403e-01 -9.16206777e-01 -7.32405782e-02 -3.04650255e-02
3.27656895e-01 -3.83533277e-02 3.87129128e-01 1.41371876e-01
5.92540264e-01 -1.09604537e+00 7.93763280e-01 5.15359007e-02
4.28720772e-01 -1.17578137e+00 7.97567964e-01 2.46515960e-01
-1.23755026e+00 -3.03622391e-02 -2.21956819e-01 2.65501797e-01
6.32809177e-02 1.24122456e-01 -1.01534283e+00 3.53325188e-01
7.92078316e-01 4.74604338e-01 -3.51408631e-01 9.48437750e-01
-3.06633741e-01 1.00436437e+00 -3.40838313e-01 2.07471564e-01
2.82019287e-01 4.19268817e-01 9.22062218e-01 1.32497859e+00
2.90093958e-01 -2.42501378e-01 6.68421388e-02 2.87651449e-01
-4.70766798e-02 2.51667630e-02 -1.16161501e+00 -9.34712291e-02
5.90643704e-01 9.83275592e-01 -4.37606722e-01 6.24510739e-03
-4.33514416e-01 1.01045465e+00 -7.04211220e-02 1.30397052e-01
-1.18307447e+00 -3.08701158e-01 7.97615826e-01 -2.56083697e-01
5.21515667e-01 -1.29945904e-01 3.12077671e-01 -9.13754523e-01
5.13702482e-02 -1.37752545e+00 5.76868057e-01 -3.79536331e-01
-1.40954852e+00 6.67953730e-01 -2.35514447e-01 -1.49648905e+00
-4.41661477e-01 -5.18973291e-01 -7.90823579e-01 6.08169198e-01
-1.10081017e+00 -1.03494382e+00 6.49811402e-02 1.08372545e+00
3.98082763e-01 -8.02839518e-01 1.07032073e+00 4.85592186e-01
-2.64609426e-01 1.19770479e+00 1.78264469e-01 5.69575012e-01
8.69279385e-01 -1.00469041e+00 5.05177021e-01 1.02372861e+00
4.87591237e-01 1.99319601e-01 1.03248823e+00 -4.17523891e-01
-1.07337701e+00 -1.02992594e+00 3.17386657e-01 -3.32381845e-01
1.10133028e+00 -4.24149603e-01 -9.88420069e-01 6.32171154e-01
3.19610417e-01 -3.19737680e-02 9.54895914e-01 -3.78596336e-01
-8.18323314e-01 8.15723650e-03 -1.23841453e+00 4.46214259e-01
5.59272051e-01 -9.70284462e-01 -6.27301991e-01 5.50968409e-01
7.44193673e-01 -3.19575638e-01 -9.31662917e-01 2.96566218e-01
4.03462857e-01 -9.68294859e-01 1.40604353e+00 -5.90947568e-01
1.49847224e-01 -2.59754986e-01 -4.98361558e-01 -1.30824137e+00
3.19640785e-02 -8.21043193e-01 -2.57632464e-01 1.51650620e+00
7.45440796e-02 -6.21539652e-01 3.82131606e-01 -2.64997840e-01
4.01826687e-02 -6.33467510e-02 -1.33596611e+00 -1.14904106e+00
8.80772900e-03 -9.77804244e-01 3.16084951e-01 1.18251920e+00
-1.15264013e-01 -1.30259350e-01 -7.78490961e-01 7.72916436e-01
6.63468421e-01 -6.45906448e-01 7.06727743e-01 -8.47460449e-01
-7.37803221e-01 -2.85823643e-01 -1.11211240e+00 -2.92252541e-01
3.21671396e-01 -5.73390365e-01 -1.03010178e-01 -5.88049591e-01
-3.47208798e-01 -3.97117995e-02 -3.95389766e-01 2.92478770e-01
-2.39899103e-03 6.35144114e-01 3.84668291e-01 -3.54680270e-02
-3.39291394e-01 2.62111604e-01 5.35223484e-01 -4.15780127e-01
3.15881401e-01 2.90536970e-01 -1.63527414e-01 9.84307885e-01
1.02333546e+00 -8.24330628e-01 -2.39695668e-01 -2.21373909e-03
-1.84375346e-01 -4.76706661e-02 4.77953374e-01 -1.18470991e+00
9.17413011e-02 1.09713838e-01 1.60667554e-01 -3.35235029e-01
5.66256821e-01 -1.04347646e+00 2.20434945e-02 5.57211041e-01
-3.68169338e-01 -5.03414311e-02 4.50666547e-01 7.80099273e-01
-4.64101732e-01 -2.51996189e-01 1.04626787e+00 8.26756179e-04
-4.56866920e-01 2.26834044e-01 -6.62308872e-01 -6.04171865e-02
1.22744191e+00 1.61076486e-01 -3.20462406e-01 -7.72634745e-01
-8.06515157e-01 -2.03065634e-01 1.67134553e-02 6.55824184e-01
8.29375803e-01 -1.52221859e+00 -7.17249930e-01 1.87778860e-01
6.74991906e-02 -8.30564380e-01 4.53542508e-02 4.77933496e-01
-3.79514486e-01 1.48392722e-01 -1.97812527e-01 -4.34405565e-01
-2.11655831e+00 9.87185955e-01 3.69891733e-01 -1.40716255e-01
-2.18289405e-01 8.70998502e-01 8.42025429e-02 6.80941045e-02
7.58880258e-01 2.58257180e-01 -2.24913448e-01 6.83320984e-02
8.35686684e-01 2.78404832e-01 9.27274674e-02 -9.50018525e-01
-5.13023734e-01 3.71776223e-01 -1.65310845e-01 -3.17751169e-01
8.89981925e-01 1.72448188e-01 1.64045706e-01 3.48503590e-01
1.52166331e+00 3.78978759e-01 -8.23214054e-01 -2.05831468e-01
-2.79155165e-01 -6.76620364e-01 -8.80981144e-03 -3.46214890e-01
-1.10191560e+00 9.70096886e-01 9.64686513e-01 7.48888373e-01
1.19500744e+00 4.46881205e-02 6.12414479e-01 1.97092384e-01
1.37981653e-01 -9.07304287e-01 4.63881135e-01 4.27836508e-01
1.02428186e+00 -1.04365838e+00 -3.08796138e-01 -2.97332227e-01
-5.80576777e-01 1.05965531e+00 4.81634736e-01 -1.64750040e-01
6.70230627e-01 5.83651900e-01 3.12872648e-01 2.60761291e-01
-5.41441262e-01 -4.45472039e-02 1.92996845e-01 8.87512803e-01
2.26234570e-01 -1.06826365e-01 -3.11400201e-02 2.56150156e-01
-2.97108650e-01 -5.09940922e-01 4.93811727e-01 7.99641073e-01
-3.33694994e-01 -1.28103423e+00 -1.04235876e+00 -3.52882966e-02
-1.08070064e+00 -1.46784216e-01 -7.43358672e-01 6.38162732e-01
3.78155671e-02 1.26858783e+00 -2.15671226e-01 -7.47427642e-01
3.25986236e-01 -7.54261985e-02 3.97904307e-01 -1.62738025e-01
-7.36015558e-01 1.45792291e-01 6.63322061e-02 -5.96769392e-01
-3.32097232e-01 -6.18993402e-01 -8.84395540e-01 -4.29239482e-01
-4.09704715e-01 -5.24751144e-03 6.42761171e-01 7.58974254e-01
-2.24519864e-01 5.82918525e-01 1.13993990e+00 -8.54787290e-01
-7.16044962e-01 -8.11391532e-01 -6.75722957e-01 5.85215688e-01
7.34237969e-01 -6.16993248e-01 -8.92463505e-01 -3.61637697e-02] | [14.001389503479004, 5.829612731933594] |
fc5f91c3-20ca-403f-a35a-dc600cb1bc1c | learning-representation-for-anomaly-detection | 2303.05000 | null | https://arxiv.org/abs/2303.05000v1 | https://arxiv.org/pdf/2303.05000v1.pdf | Learning Representation for Anomaly Detection of Vehicle Trajectories | Predicting the future trajectories of surrounding vehicles based on their history trajectories is a critical task in autonomous driving. However, when small crafted perturbations are introduced to those history trajectories, the resulting anomalous (or adversarial) trajectories can significantly mislead the future trajectory prediction module of the ego vehicle, which may result in unsafe planning and even fatal accidents. Therefore, it is of great importance to detect such anomalous trajectories of the surrounding vehicles for system safety, but few works have addressed this issue. In this work, we propose two novel methods for learning effective and efficient representations for online anomaly detection of vehicle trajectories. Different from general time-series anomaly detection, anomalous vehicle trajectory detection deals with much richer contexts on the road and fewer observable patterns on the anomalous trajectories themselves. To address these challenges, our methods exploit contrastive learning techniques and trajectory semantics to capture the patterns underlying the driving scenarios for effective anomaly detection under supervised and unsupervised settings, respectively. We conduct extensive experiments to demonstrate that our supervised method based on contrastive learning and unsupervised method based on reconstruction with semantic latent space can significantly improve the performance of anomalous trajectory detection in their corresponding settings over various baseline methods. We also demonstrate our methods' generalization ability to detect unseen patterns of anomalies. | ['Qi Zhu', 'Qi Alfred Chen', 'Xiaowei Yuan', 'Takami Sato', 'Xiangguo Liu', 'Juyang Bai', 'Ruochen Jiao'] | 2023-03-09 | null | null | null | null | ['trajectory-prediction'] | ['computer-vision'] | [ 7.01798424e-02 4.61628735e-02 -2.25616291e-01 -3.80858123e-01
-5.27120948e-01 -5.68943799e-01 7.78918207e-01 3.36613238e-01
-3.53775173e-02 4.38523620e-01 2.69784778e-01 -5.86736083e-01
-3.33182141e-02 -8.09913337e-01 -9.52268422e-01 -7.90646970e-01
-3.03195000e-01 8.48598257e-02 6.16067767e-01 -3.03608268e-01
3.34353328e-01 7.58940518e-01 -1.82392049e+00 2.19902620e-01
9.76484060e-01 7.99360991e-01 -3.12933952e-01 4.47246939e-01
-4.00558919e-01 8.76935899e-01 -3.78148496e-01 -2.08934665e-01
2.45771602e-01 6.72125891e-02 -3.16622078e-01 2.64523923e-02
2.43044257e-01 -4.31198418e-01 -1.00976920e+00 1.18935525e+00
-1.92539141e-01 4.79826391e-01 8.11969459e-01 -1.99327922e+00
-2.69032300e-01 3.72464269e-01 -5.48809320e-02 7.16704071e-01
1.03585020e-01 5.34771025e-01 5.94772875e-01 -8.20506454e-01
1.23204581e-01 1.08089948e+00 5.63757956e-01 6.25605226e-01
-7.86773443e-01 -8.74701560e-01 7.09950149e-01 7.48958766e-01
-1.13393092e+00 -6.65647626e-01 1.01778972e+00 -4.99337196e-01
7.99057364e-01 2.80740082e-01 3.67788106e-01 1.49379373e+00
2.78452486e-01 9.99163151e-01 2.92908728e-01 3.35475028e-01
3.23642373e-01 -2.18983423e-02 1.70520231e-01 3.68047148e-01
1.43548042e-01 3.69182914e-01 -1.19889796e-01 -4.03447658e-01
-1.19555019e-01 6.75678790e-01 2.40056515e-01 -1.16514869e-01
-1.13856053e+00 7.02140749e-01 1.09468609e-01 -4.85066660e-02
-4.39296544e-01 2.17758909e-01 7.66831040e-01 4.54867333e-01
6.72359347e-01 3.22653800e-02 -1.98039472e-01 -3.05149794e-01
-5.67499340e-01 6.04822755e-01 4.44345236e-01 1.14010239e+00
8.50386322e-01 4.86493468e-01 -2.98456967e-01 2.68682212e-01
9.28622633e-02 6.98580325e-01 4.44432259e-01 -7.12458014e-01
5.33889592e-01 7.05671787e-01 3.60972404e-01 -1.46801043e+00
-2.48077020e-01 -7.35064968e-02 -6.95313036e-01 -4.85192426e-02
3.50714564e-01 -3.26352268e-02 -9.70102847e-01 1.45948958e+00
3.48185092e-01 9.44608390e-01 2.76792228e-01 5.32396734e-01
3.42918068e-01 8.96621883e-01 3.16205889e-01 -1.30137458e-01
6.14327490e-01 -8.77008379e-01 -1.01013553e+00 -2.61005133e-01
1.14117408e+00 -4.69758779e-01 9.29763019e-01 -7.72847086e-02
-7.33733118e-01 -3.77304345e-01 -9.08814609e-01 4.00114924e-01
-6.63077891e-01 -3.45041275e-01 4.01207864e-01 2.90500313e-01
-5.06768584e-01 6.12523198e-01 -9.79708731e-01 -1.54292151e-01
5.56723714e-01 1.02280535e-01 -1.60325319e-01 -1.05434239e-01
-1.25379169e+00 6.66772842e-01 4.27006483e-01 6.93920925e-02
-1.14825952e+00 -5.98038912e-01 -9.71974671e-01 -3.64272863e-01
6.30081058e-01 1.30564734e-01 1.16939318e+00 -5.97831786e-01
-9.46647286e-01 3.63573968e-01 -5.57405770e-01 -7.34291017e-01
5.54459929e-01 -2.55947262e-01 -1.21197832e+00 -2.20036954e-01
3.08944643e-01 6.74498230e-02 1.03042006e+00 -8.30451131e-01
-1.02708519e+00 -2.38863945e-01 -3.22302669e-01 -1.68188661e-01
-3.07910204e-01 -2.82486826e-01 -2.30800249e-02 -6.13499999e-01
1.92077413e-01 -1.06209970e+00 -4.48234826e-01 -2.91132361e-01
-3.49885911e-01 -5.90433776e-01 1.54264545e+00 -4.00249004e-01
1.38545811e+00 -2.51404476e+00 -5.19828081e-01 4.80556995e-01
1.27486110e-01 3.27187955e-01 3.37037700e-03 5.11485994e-01
1.37462869e-01 -1.03038102e-01 -3.08704913e-01 1.74356364e-02
-7.85333365e-02 6.78840578e-01 -1.11682487e+00 7.21932292e-01
2.08834574e-01 8.46273541e-01 -1.29874623e+00 -1.11751951e-01
3.55278522e-01 -1.47311047e-01 -2.58459955e-01 2.72576302e-01
-1.91622570e-01 6.96149707e-01 -8.51223171e-01 7.14393437e-01
5.27240813e-01 3.55063379e-01 -5.73848128e-01 2.95881838e-01
-6.65927231e-02 1.16223834e-01 -7.72745073e-01 1.11663353e+00
-1.42581567e-01 7.34307468e-01 -5.99280298e-01 -1.23624086e+00
9.35488462e-01 3.75744522e-01 5.77275455e-01 -6.94259048e-01
-2.82033294e-01 3.01484823e-01 -1.62951797e-01 -8.49930108e-01
5.62730432e-01 2.39857525e-01 -2.54976213e-01 2.91135728e-01
-5.64743102e-01 5.10163784e-01 -1.71926290e-01 3.01398039e-01
1.39591753e+00 -2.13646024e-01 -1.61867350e-01 4.72690240e-02
7.34562278e-01 3.56358528e-01 7.87197232e-01 6.97853684e-01
-6.29863918e-01 6.99981302e-02 3.36257190e-01 -9.91997659e-01
-1.23820651e+00 -1.22537923e+00 -5.95025495e-02 8.64361644e-01
4.32592690e-01 -3.92720520e-01 -5.07950485e-01 -1.20016277e+00
3.05841029e-01 1.32091308e+00 -4.05252159e-01 -8.79735231e-01
-7.43361056e-01 -5.05209446e-01 6.74884558e-01 6.58761680e-01
2.45986655e-01 -1.07424819e+00 -1.10526472e-01 2.32048497e-01
-1.63651124e-01 -1.28192937e+00 -4.36325938e-01 -4.33729976e-01
-6.60624802e-01 -9.45694387e-01 1.05760448e-01 -3.14620137e-01
7.16930687e-01 5.77450573e-01 6.01515591e-01 -3.30423340e-02
1.91389158e-01 3.63685995e-01 -3.90887231e-01 -6.94027603e-01
-8.12340260e-01 -1.79091662e-01 5.99521399e-01 4.82502222e-01
1.00405180e+00 -6.44198179e-01 -6.23635709e-01 6.09638512e-01
-9.01336372e-01 -4.33113694e-01 2.86241978e-01 3.55399013e-01
5.66774786e-01 4.69638675e-01 9.25540626e-01 -8.43821883e-01
4.23954040e-01 -1.26150203e+00 -3.85288030e-01 -2.22351640e-01
-6.74879193e-01 9.86701027e-02 1.00429237e+00 -4.55813527e-01
-8.98723423e-01 -1.41643304e-02 -1.75105691e-01 -8.51966202e-01
-7.97381759e-01 -6.50206655e-02 -1.94403812e-01 2.45952845e-01
5.86776197e-01 6.69862568e-01 9.70026571e-03 -2.24276900e-01
2.69750059e-01 6.02070272e-01 6.84121132e-01 -2.23491117e-01
1.23558974e+00 8.61161590e-01 -6.74963975e-03 -8.74005914e-01
-6.64334536e-01 -8.18289757e-01 -5.17483115e-01 -3.93288583e-01
5.78687549e-01 -9.08152699e-01 -3.10999066e-01 3.17693681e-01
-9.66852665e-01 -6.42156042e-03 -3.60943764e-01 4.02328283e-01
-5.32391906e-01 4.32824939e-01 -1.29282027e-01 -8.88674676e-01
2.61056811e-01 -1.15858543e+00 9.09795165e-01 -2.47189447e-01
-2.36542732e-01 -9.91959810e-01 7.62568489e-02 -3.54873226e-03
3.33400607e-01 3.24394673e-01 9.02452946e-01 -1.23687887e+00
-6.66285872e-01 -5.41972458e-01 6.63321540e-02 2.45436862e-01
2.81361639e-01 -1.58919588e-01 -1.04099655e+00 -2.33447462e-01
-1.03249930e-01 2.55595118e-01 8.36021125e-01 -1.07727565e-01
1.48590386e+00 -7.80177832e-01 -7.75563180e-01 5.64296305e-01
8.01922858e-01 3.01954210e-01 6.84237182e-01 2.30684504e-01
1.04552650e+00 7.92982697e-01 9.89077210e-01 2.67362773e-01
3.37021291e-01 5.32673299e-01 7.19501078e-01 4.56087410e-01
3.90443623e-01 -6.31584466e-01 7.48472035e-01 6.94410443e-01
1.80209517e-01 -9.44295749e-02 -1.09271932e+00 1.04512262e+00
-2.27429724e+00 -1.33759499e+00 -3.88023645e-01 2.22577000e+00
1.00055993e-01 3.13131720e-01 3.21072042e-01 2.01817825e-01
6.70734406e-01 3.38420421e-01 -1.02461052e+00 -4.86454993e-01
1.79599628e-01 -6.96842790e-01 6.61733866e-01 1.37243554e-01
-1.26872766e+00 1.08132899e+00 5.77816057e+00 9.06590819e-01
-9.26885605e-01 1.96437731e-01 4.61670220e-01 -1.40644073e-01
-5.36526084e-01 -7.59269968e-02 -8.49816263e-01 8.79924059e-01
1.31837559e+00 -2.77373463e-01 1.74971059e-01 1.16481304e+00
4.79097486e-01 5.41155636e-01 -1.13213599e+00 8.08183193e-01
-2.11072639e-01 -1.16030586e+00 2.17902690e-01 3.44439894e-02
6.59634888e-01 3.05641830e-01 1.65734097e-01 5.86376905e-01
7.73596317e-02 -1.06927693e+00 5.49503922e-01 7.44259775e-01
2.85161942e-01 -1.19724357e+00 5.57273746e-01 6.67962492e-01
-1.18850744e+00 -4.46114480e-01 -4.80219215e-01 -5.98835535e-02
2.26847753e-01 5.21622717e-01 -8.02115560e-01 2.10119933e-01
6.51064575e-01 1.10121822e+00 -3.75225395e-01 7.30663359e-01
-4.23603840e-02 1.00556195e+00 -2.38233924e-01 2.52670109e-01
5.27613401e-01 -1.40349194e-01 1.42269433e+00 1.03615987e+00
3.98194939e-01 -4.40212861e-02 3.01696092e-01 7.89459348e-01
2.28579432e-01 -1.05222940e-01 -1.61478031e+00 -9.30055082e-02
5.64479113e-01 8.54723454e-01 -3.15728426e-01 -2.40576848e-01
-6.10564590e-01 8.09312940e-01 1.11353986e-01 6.26436412e-01
-1.02911651e+00 -2.54648328e-01 1.24140239e+00 4.37243402e-01
1.01412702e-02 -3.18615407e-01 7.35146478e-02 -1.05670583e+00
3.37087095e-01 -5.00132561e-01 3.76482129e-01 -8.70856717e-02
-1.41917825e+00 4.60360795e-01 7.01065660e-02 -1.74958920e+00
-5.18297434e-01 -2.24503636e-01 -1.21229422e+00 3.58271897e-01
-1.56406462e+00 -9.77564275e-01 -9.09480527e-02 8.91886413e-01
7.28193283e-01 -6.24580622e-01 4.39678311e-01 1.34855613e-01
-6.22879207e-01 6.70802236e-01 2.98395246e-01 -7.61726685e-03
3.51282090e-01 -1.03161454e+00 9.37122405e-01 1.11525297e+00
-3.32205221e-02 1.01790160e-01 7.75656462e-01 -9.07264411e-01
-1.31984413e+00 -2.05173516e+00 7.18506277e-01 -8.00678670e-01
8.94417763e-01 -1.62739679e-01 -1.39535511e+00 1.04842103e+00
-3.78606349e-01 3.51279557e-01 4.58991706e-01 -1.77131474e-01
-7.42994174e-02 -1.44654185e-01 -1.02060056e+00 8.34183216e-01
1.28885221e+00 -5.13137996e-01 -6.24581099e-01 4.89294052e-01
9.04520035e-01 -1.01072565e-01 -2.96336979e-01 6.35627866e-01
1.29686207e-01 -7.75898993e-01 9.11217213e-01 -1.14484453e+00
-5.18145338e-02 -4.42727029e-01 -1.68193877e-01 -1.20933783e+00
-2.42459416e-01 -6.94021285e-01 -6.64608359e-01 9.41643894e-01
2.76636809e-01 -9.11630511e-01 7.74253249e-01 6.22140884e-01
-5.95256507e-01 -6.59135878e-01 -9.72590327e-01 -9.62801337e-01
-4.92896233e-03 -1.09490967e+00 1.02613795e+00 7.14257419e-01
-1.85014814e-01 -2.46829346e-01 -4.31642145e-01 8.73774111e-01
7.08135247e-01 -4.65676263e-02 1.01857805e+00 -9.47573960e-01
4.87184793e-01 -3.62427235e-01 -9.68667150e-01 -8.73991787e-01
6.33201003e-01 -8.93266082e-01 1.70251116e-01 -7.62529492e-01
-3.64883751e-01 -4.76271003e-01 -4.95745808e-01 2.06163853e-01
-1.84719726e-01 -1.37460381e-01 -4.15823340e-01 4.94076937e-01
-7.90194333e-01 8.00356328e-01 6.35276496e-01 -2.30587110e-01
-2.51577318e-01 6.63920343e-01 -2.09491923e-01 1.06562257e+00
9.79028404e-01 -6.43864393e-01 -6.76002800e-01 -7.32933506e-02
1.91717461e-01 -2.24041566e-01 6.65528476e-01 -1.04619384e+00
3.46477687e-01 -4.88809735e-01 -1.45330042e-01 -9.51301932e-01
4.54682484e-02 -9.49041426e-01 -2.51843482e-01 1.94247544e-01
-2.64602661e-01 3.07474285e-01 3.41605067e-01 1.44897294e+00
-2.52477139e-01 1.94889292e-01 3.11817795e-01 1.28040016e-01
-1.26333654e+00 8.23907852e-01 -9.57065284e-01 7.24182576e-02
1.50895107e+00 -3.45828652e-01 -4.44968827e-02 -6.39121175e-01
-7.23530591e-01 6.52045429e-01 3.04958403e-01 8.22078645e-01
9.63153541e-01 -1.67409730e+00 -6.48470879e-01 5.90682209e-01
6.37320995e-01 1.29948286e-02 4.63438869e-01 8.60056460e-01
-8.13763142e-02 4.85161483e-01 8.08162689e-02 -7.69931316e-01
-7.23260939e-01 9.36817467e-01 1.49494246e-01 2.83960611e-01
-1.14784706e+00 2.11870804e-01 3.22898090e-01 -4.42688793e-01
2.96871752e-01 -2.18440711e-01 -2.38479897e-01 -3.74137253e-01
7.34920382e-01 7.79084504e-01 3.35776480e-03 -9.86830115e-01
-2.80348808e-01 6.98926523e-02 -3.61255199e-01 4.84229863e-01
8.09886277e-01 -3.15157264e-01 3.19583207e-01 8.14248681e-01
1.08787715e+00 1.99804716e-02 -1.40302539e+00 -4.22976196e-01
2.54120320e-01 -6.06052756e-01 3.59153794e-03 -1.24882326e-01
-8.81060421e-01 7.08789051e-01 5.78250408e-01 3.92344147e-01
7.76539624e-01 4.80317026e-02 1.52036273e+00 7.15865135e-01
3.14964235e-01 -1.01981807e+00 -2.27144912e-01 4.11128759e-01
5.85755229e-01 -1.52478659e+00 -5.62886477e-01 -2.33728692e-01
-7.54995406e-01 8.88597190e-01 6.94897175e-01 -1.27785265e-01
7.51258254e-01 -2.31066853e-01 2.23234836e-02 -2.94241041e-01
-7.34817207e-01 -1.73787206e-01 4.02799487e-01 5.27387261e-01
-4.35791314e-01 2.53842950e-01 3.05557221e-01 3.61640900e-01
-2.14400534e-02 -5.37297726e-01 4.33373928e-01 8.41633856e-01
-5.60969532e-01 -7.89984941e-01 -1.32310659e-01 6.73642397e-01
-3.99510801e-01 3.03956181e-01 -5.03538288e-02 5.12164950e-01
7.85671994e-02 1.00252485e+00 4.21509534e-01 -6.20296061e-01
4.42106813e-01 4.17792827e-01 -3.11724573e-01 -3.90515894e-01
2.28253543e-01 -6.18625998e-01 -5.42910174e-02 -1.05083466e+00
4.28201854e-02 -9.21886742e-01 -1.45016241e+00 -5.01860797e-01
1.00372896e-01 1.39570296e-01 2.85265684e-01 1.38104630e+00
6.49492979e-01 3.08820665e-01 1.01496792e+00 -4.37016517e-01
-5.92231333e-01 -6.87586904e-01 -4.84256417e-01 9.67662513e-01
8.97247493e-01 -8.90839338e-01 -6.34581149e-01 -2.95542896e-01] | [7.388101100921631, 2.195969820022583] |
2862df61-8933-4c52-96b3-c28582d851e7 | deepchess-end-to-end-deep-neural-network-for | 1711.09667 | null | http://arxiv.org/abs/1711.09667v1 | http://arxiv.org/pdf/1711.09667v1.pdf | DeepChess: End-to-End Deep Neural Network for Automatic Learning in Chess | We present an end-to-end learning method for chess, relying on deep neural
networks. Without any a priori knowledge, in particular without any knowledge
regarding the rules of chess, a deep neural network is trained using a
combination of unsupervised pretraining and supervised training. The
unsupervised training extracts high level features from a given position, and
the supervised training learns to compare two chess positions and select the
more favorable one. The training relies entirely on datasets of several million
chess games, and no further domain specific knowledge is incorporated.
The experiments show that the resulting neural network (referred to as
DeepChess) is on a par with state-of-the-art chess playing programs, which have
been developed through many years of manual feature selection and tuning.
DeepChess is the first end-to-end machine learning-based method that results in
a grandmaster-level chess playing performance. | ['Nathan S. Netanyahu', 'Lior Wolf', 'Eli David'] | 2017-11-27 | null | null | null | null | ['game-of-chess'] | ['playing-games'] | [-1.54844150e-01 -1.83451232e-02 3.50896530e-02 -4.99888450e-01
-5.58204532e-01 -6.39120758e-01 3.94791275e-01 1.40472189e-01
-9.06541467e-01 6.12680197e-01 -1.61165416e-01 -5.08272231e-01
-1.77600175e-01 -1.16681290e+00 -7.82864690e-01 -4.18849468e-01
4.08247411e-02 7.92930126e-01 6.68640673e-01 -1.00200617e+00
5.40978909e-01 1.51739508e-01 -1.60891414e+00 6.04296029e-01
5.38353980e-01 1.11143947e+00 -9.39357132e-02 1.07054508e+00
2.16255173e-01 1.21623266e+00 -9.01719034e-01 -5.79952776e-01
7.59485841e-01 -4.04436916e-01 -9.86491263e-01 -4.01999682e-01
2.62568176e-01 -2.69690096e-01 -4.19971436e-01 7.21123338e-01
4.07786757e-01 3.16916108e-01 5.01562834e-01 -7.50951648e-01
-2.42786855e-01 5.74203968e-01 -3.13770384e-01 2.19987035e-01
1.57794133e-01 3.60985518e-01 1.17912781e+00 -5.29581904e-01
3.61878097e-01 4.50633526e-01 9.86420810e-01 2.84671724e-01
-8.96869242e-01 -3.98997962e-01 -2.65154451e-01 -2.21190527e-02
-1.11135447e+00 -1.43682778e-01 4.78296191e-01 -5.51459253e-01
1.00833118e+00 1.15269840e-01 1.04561675e+00 6.69406772e-01
1.19002067e-01 1.00616634e+00 1.06901443e+00 -7.13314414e-01
4.41225260e-01 -8.08794275e-02 1.09334752e-01 8.02376270e-01
7.59466290e-02 6.70182943e-01 -6.04223132e-01 8.11726600e-02
7.04838336e-01 -4.42338765e-01 2.74133205e-01 -4.26729947e-01
-9.18483019e-01 9.77265298e-01 3.99703324e-01 3.71198833e-01
-2.69545674e-01 1.68870300e-01 7.80503869e-01 6.21895730e-01
9.92725343e-02 1.31616330e+00 -8.03397357e-01 -6.87737703e-01
-1.56628346e+00 6.83598876e-01 9.01669741e-01 4.27852988e-01
7.45589972e-01 -4.71379515e-03 5.02403192e-02 4.92798358e-01
-1.93004757e-01 -1.40379101e-01 6.46087825e-01 -9.78314698e-01
4.75336432e-01 6.95130289e-01 1.34689242e-01 -9.36967850e-01
-4.75800604e-01 -6.42417729e-01 -4.37814415e-01 9.76666927e-01
7.33508468e-01 -4.91166472e-01 -7.45774388e-01 1.38099146e+00
5.04308231e-02 -2.49879584e-01 5.66970296e-02 8.45782697e-01
9.14866269e-01 3.14163268e-01 -4.11032110e-01 4.26631868e-01
9.86606359e-01 -1.17429483e+00 1.33005321e-01 -1.46789163e-01
5.79564393e-01 -4.98359054e-01 1.12566662e+00 8.34382594e-01
-1.37234318e+00 -8.24787021e-01 -1.44669557e+00 -3.30030769e-02
-7.40508199e-01 2.01660439e-01 8.24941278e-01 8.66757512e-01
-9.53531325e-01 9.54566419e-01 -8.08271587e-01 1.71515033e-01
2.75876641e-01 7.49629498e-01 -2.92415440e-01 5.30289292e-01
-1.36037922e+00 7.66015112e-01 9.22080696e-01 -2.62103856e-01
-9.29528177e-01 -5.76137245e-01 -6.77452207e-01 1.42702088e-01
6.75611913e-01 -5.81951976e-01 1.64214063e+00 -1.25392723e+00
-1.95084643e+00 1.05358648e+00 6.44212484e-01 -6.73319817e-01
6.53799415e-01 -2.41591662e-01 9.86694843e-02 -1.24107569e-01
4.02602255e-02 6.08637273e-01 6.83598518e-01 -7.78222799e-01
-9.69407856e-01 -9.54825729e-02 6.95254266e-01 2.16570050e-01
-1.90777004e-01 8.00728500e-02 -5.18747091e-01 -5.60118735e-01
-8.84956345e-02 -6.69147789e-01 -4.56768751e-01 -5.53928196e-01
-4.12172884e-01 -1.29756093e-01 1.90416828e-01 -4.43112433e-01
1.20248997e+00 -1.90536547e+00 1.06579185e-01 3.31727654e-01
3.07524741e-01 4.35898721e-01 -1.45763129e-01 4.00366962e-01
-1.85358703e-01 -2.88851708e-01 5.19422852e-02 1.26306161e-01
6.74112588e-02 -2.46717677e-01 7.65426382e-02 2.82992005e-01
-1.11535832e-01 7.64271438e-01 -1.13465643e+00 -9.77337584e-02
9.42215621e-02 -2.69021571e-01 -9.57863152e-01 3.53190571e-01
-4.85337019e-01 3.73830885e-01 -3.26248288e-01 5.23914993e-01
2.26127729e-01 -2.29466021e-01 3.39252837e-02 1.59385487e-01
-3.64913702e-01 4.74736154e-01 -9.66053903e-01 1.91249204e+00
-3.51985544e-01 7.02117264e-01 -5.27726710e-01 -1.17057431e+00
9.00359631e-01 -9.42047015e-02 4.20184910e-01 -6.42558813e-01
4.79367644e-01 2.18513206e-01 3.82924378e-01 -4.18282002e-01
1.01610506e+00 9.94897038e-02 -5.87274671e-01 5.07754028e-01
3.70499551e-01 -2.75148332e-01 7.59569824e-01 3.79058123e-01
1.25003076e+00 4.33119088e-01 4.58068430e-01 -1.36287436e-01
2.75103569e-01 5.47942579e-01 3.59992474e-01 1.03792024e+00
3.77107970e-02 5.17089367e-01 7.44690776e-01 -1.17151475e+00
-1.23315728e+00 -6.79940104e-01 3.68052959e-01 1.70128012e+00
-1.67173937e-01 -6.46257281e-01 -9.30982411e-01 -7.44309545e-01
-1.58866763e-01 5.07875621e-01 -9.60003138e-01 -2.32608039e-02
-4.41171259e-01 -5.37909985e-01 7.74928749e-01 5.62271118e-01
7.69289911e-01 -1.22549951e+00 -1.15746427e+00 3.40932161e-01
3.83623004e-01 -5.93550742e-01 -7.84786195e-02 5.76198816e-01
-5.95580995e-01 -1.46472931e+00 -6.74408615e-01 -6.66440487e-01
4.44196045e-01 -3.88990909e-01 1.68186891e+00 2.67901599e-01
-3.44105482e-01 -1.68358758e-02 -4.48590547e-01 -4.00241256e-01
-5.12550175e-01 7.59467363e-01 -8.53422284e-02 -5.04700959e-01
4.71729189e-01 -7.25175619e-01 -4.86997545e-01 9.28702112e-03
-6.40498340e-01 1.67979568e-01 8.90384436e-01 1.19798267e+00
2.10157618e-01 3.76836389e-01 1.49670869e-01 -1.24054134e+00
6.70249760e-01 -1.84973076e-01 -1.09177113e+00 2.53194477e-03
-2.15423852e-01 7.26510808e-02 9.05369341e-01 -1.26486063e-01
-6.68495774e-01 3.61991584e-01 -2.92687535e-01 6.95805922e-02
-4.45887834e-01 7.44475305e-01 9.40737575e-02 -2.70976305e-01
1.20993078e+00 1.90536559e-01 -3.12711358e-01 -3.30111206e-01
2.27969483e-01 4.91839141e-01 7.93048441e-01 -7.17135370e-01
7.87131131e-01 -5.52286245e-02 -1.59206614e-01 -4.48904127e-01
-9.54957724e-01 -3.74136508e-01 -1.11184859e+00 -2.91923165e-01
6.46598756e-01 -7.92640626e-01 -1.15248013e+00 7.73673415e-01
-7.50107944e-01 -7.91188598e-01 -5.14611065e-01 2.38485515e-01
-7.93825626e-01 -1.46243349e-02 -6.91209197e-01 -4.99806911e-01
7.86981080e-03 -9.94164705e-01 6.10074520e-01 4.44930732e-01
-2.29828626e-01 -8.36209178e-01 6.15504444e-01 5.51835001e-01
2.31061116e-01 2.69114286e-01 6.98078871e-01 -9.16241288e-01
-3.27253729e-01 -4.86522049e-01 4.66752760e-02 4.77676719e-01
-2.22232908e-01 -8.90451819e-02 -6.29645169e-01 -2.34780878e-01
-4.47346568e-01 -8.94112647e-01 8.39290977e-01 3.09179783e-01
1.11003733e+00 7.44262189e-02 2.66043574e-01 6.89699590e-01
1.38578546e+00 2.39201397e-01 7.08013475e-01 1.11834323e+00
4.35293943e-01 2.96302348e-01 6.10670388e-01 4.74527359e-01
2.34100565e-01 5.11716783e-01 4.55554605e-01 -2.10322320e-01
3.30619603e-01 -3.28125656e-01 1.26996368e-01 4.05738115e-01
-5.53533196e-01 1.58819422e-01 -1.13511145e+00 5.28836966e-01
-1.69438756e+00 -8.75162840e-01 1.49675071e-01 2.16358781e+00
9.93479609e-01 5.85139751e-01 6.65633798e-01 5.63603103e-01
4.16685194e-01 1.14040449e-01 -6.71070814e-01 -4.90459472e-01
2.16162935e-01 6.97245836e-01 7.18000114e-01 1.19567126e-01
-1.45398259e+00 1.35764527e+00 7.43849373e+00 1.00655258e+00
-1.03497612e+00 -2.52665311e-01 6.27606630e-01 -2.88680911e-01
3.22889715e-01 -2.05153767e-02 -5.29015481e-01 3.13730896e-01
9.20130968e-01 7.30988309e-02 6.13906324e-01 1.05491710e+00
-1.18269265e-01 -3.19198489e-01 -9.76992905e-01 6.48186445e-01
-2.17651740e-01 -1.64326298e+00 -3.17816049e-01 -1.57135092e-02
8.63280118e-01 -6.10845350e-02 2.40015119e-01 7.90893555e-01
1.03715825e+00 -1.17395890e+00 9.74697292e-01 3.49883258e-01
6.49152577e-01 -1.05964243e+00 8.99411857e-01 5.20997643e-01
-8.60055685e-01 -1.20631143e-01 -3.76293808e-01 -5.19429326e-01
-3.49616706e-01 2.30119660e-01 -6.92528725e-01 5.72095931e-01
7.22568512e-01 5.23575008e-01 -6.90211594e-01 1.08515084e+00
-3.96061420e-01 6.67945683e-01 -1.69314444e-01 -2.24872321e-01
7.47793555e-01 4.67280634e-02 2.89565146e-01 1.03079021e+00
-6.02038503e-02 2.47799352e-01 2.84409672e-01 6.36057198e-01
-9.61222053e-02 1.20942153e-01 -2.63607293e-01 -1.55513167e-01
7.89065808e-02 1.05323756e+00 -8.47658038e-01 -5.60843050e-01
-2.74565041e-01 8.98681641e-01 4.96216327e-01 -5.37735187e-02
-6.91772282e-01 -7.65350521e-01 1.87237427e-01 1.09347723e-01
6.44516051e-01 -6.90866858e-02 -4.13383454e-01 -9.73230243e-01
-3.60237539e-01 -1.26263607e+00 5.32839894e-01 -6.05515122e-01
-1.10654831e+00 7.02868879e-01 -2.87698448e-01 -1.22695601e+00
-4.52967346e-01 -9.49810266e-01 -8.85123849e-01 5.26031017e-01
-1.08383536e+00 -8.30761194e-01 -3.45142782e-02 4.43202496e-01
3.59306961e-01 -9.97523487e-01 6.83984458e-01 1.39000848e-01
-2.32058555e-01 8.09379518e-01 2.32075661e-01 6.37199938e-01
4.44330126e-01 -1.68179405e+00 5.75467825e-01 5.05738735e-01
2.56005555e-01 4.31821287e-01 8.06609273e-01 -1.84092835e-01
-9.69262183e-01 -8.09204102e-01 5.09810805e-01 -2.52989799e-01
6.14899874e-01 -2.32082948e-01 -4.12097663e-01 4.76112783e-01
2.48907015e-01 -1.33426845e-01 8.44237924e-01 6.01756632e-01
-3.15152258e-01 -2.35395491e-04 -9.60655272e-01 4.22034502e-01
7.41092682e-01 -2.96776235e-01 -9.19962764e-01 2.87435472e-01
2.08626494e-01 -1.01142406e+00 -5.37162006e-01 1.83287919e-01
7.01444387e-01 -1.61035597e+00 8.96938980e-01 -1.21691108e+00
9.48287547e-01 -9.89390016e-02 5.01818210e-02 -1.41784501e+00
-4.02918607e-01 -4.28876996e-01 3.24396610e-01 4.87371087e-01
6.19642735e-01 -7.53244683e-02 1.52283180e+00 4.70506668e-01
-2.57978648e-01 -9.24908459e-01 -5.94991267e-01 -8.15876245e-01
4.48473364e-01 -2.37197131e-01 6.34394705e-01 9.11973298e-01
1.27948239e-01 1.71632782e-01 -4.99762714e-01 -7.22324534e-04
2.41551533e-01 4.07873422e-01 1.22357929e+00 -1.38540173e+00
-9.61822152e-01 -6.34979784e-01 -7.51313090e-01 -1.15447843e+00
-2.00114381e-02 -5.96594036e-01 2.56492853e-01 -9.56940055e-01
-2.03130557e-03 -3.59244585e-01 -4.40006524e-01 7.42142320e-01
9.90820602e-02 5.11923850e-01 7.42091388e-02 -1.58582032e-01
-1.08626544e+00 2.20725358e-01 9.14410710e-01 -1.79570004e-01
-4.15627152e-01 2.07998499e-01 -7.65819967e-01 1.09910297e+00
9.51766610e-01 -4.85247046e-01 -1.59707665e-01 -1.34528890e-01
6.23577356e-01 -5.39247841e-02 4.81699035e-02 -1.35764766e+00
4.29279149e-01 -2.05416545e-01 5.61060011e-01 -4.71675724e-01
2.73155749e-01 -3.66178036e-01 -2.85396010e-01 5.51765263e-01
-4.98228490e-01 -7.97972381e-02 2.54958719e-01 9.19787809e-02
-3.85774821e-01 -5.63628554e-01 7.95942128e-01 -6.39529288e-01
-9.49319899e-01 1.03276268e-01 -5.74676573e-01 2.49956980e-01
9.60024059e-01 -3.45026255e-01 1.85208336e-01 -3.65534425e-01
-9.12167609e-01 -1.73031569e-01 3.75063926e-01 1.08492158e-01
2.12430060e-01 -1.05410945e+00 -7.12445498e-01 1.86083451e-01
-6.67548925e-02 -1.04769208e-01 1.45902276e-01 2.42613822e-01
-9.32091355e-01 2.59920388e-01 -5.93618631e-01 -3.38755727e-01
-1.17703462e+00 2.60351807e-01 4.80024099e-01 -8.09828281e-01
-4.68957782e-01 9.90820229e-01 -2.93244034e-01 -7.43539035e-01
8.16017315e-02 -1.54947832e-01 -2.27197841e-01 -1.77784681e-01
5.63778877e-01 5.28038144e-02 4.31284338e-01 -3.34882736e-01
-5.02262637e-02 1.58753723e-01 -2.60840774e-01 -1.94720998e-01
1.69821775e+00 7.42751718e-01 1.87986881e-01 3.02452296e-01
6.19792402e-01 9.45487246e-02 -1.27810895e+00 4.88801114e-02
1.16690569e-01 -7.33179688e-01 1.75361469e-01 -1.19138920e+00
-1.25765562e+00 7.52395988e-01 1.65941060e-01 2.44216770e-01
9.93233860e-01 -5.92850566e-01 7.83898294e-01 1.01084507e+00
6.55699253e-01 -1.42487395e+00 3.07369977e-01 1.02969754e+00
5.24297655e-01 -1.17023790e+00 9.22155753e-03 2.07097933e-01
-6.28080368e-01 1.39712083e+00 6.86868548e-01 -8.06737959e-01
4.27690953e-01 3.97114486e-01 1.72460407e-01 -4.09602642e-01
-6.89143658e-01 -2.85218269e-01 1.56098083e-01 3.39920789e-01
3.81391197e-01 -1.23659864e-01 -1.45237101e-02 1.01687229e+00
-8.38457346e-01 2.49713302e-01 4.37852532e-01 1.10767901e+00
-5.51220059e-01 -1.26070106e+00 -3.35670829e-01 4.50342298e-01
-4.72132504e-01 -2.92822897e-01 -5.09325206e-01 9.34950769e-01
5.78085184e-01 5.40466964e-01 -5.05734123e-02 -8.24743092e-01
5.91831863e-01 -1.46375343e-01 6.51497900e-01 -6.66566610e-01
-1.29635668e+00 -4.13857549e-01 1.10602468e-01 -5.18746197e-01
-1.84604630e-01 -3.56543541e-01 -1.03490186e+00 -7.80611277e-01
-2.08623767e-01 4.71292764e-01 3.90805066e-01 1.01200640e+00
1.84529051e-02 7.07827747e-01 6.04394913e-01 -1.05218089e+00
-4.77042943e-01 -6.46316946e-01 -7.96301603e-01 1.78839281e-01
-3.79956439e-02 -5.28861403e-01 -2.58444957e-02 -8.57988372e-02] | [3.4441146850585938, 1.430984377861023] |
60c16aff-190f-40c8-ab59-211e0c6ae7b3 | analysis-and-adaptation-of-yolov4-for-object | 2203.10194 | null | https://arxiv.org/abs/2203.10194v1 | https://arxiv.org/pdf/2203.10194v1.pdf | Analysis and Adaptation of YOLOv4 for Object Detection in Aerial Images | The recent and rapid growth in Unmanned Aerial Vehicles (UAVs) deployment for various computer vision tasks has paved the path for numerous opportunities to make them more effective and valuable. Object detection in aerial images is challenging due to variations in appearance, pose, and scale. Autonomous aerial flight systems with their inherited limited memory and computational power demand accurate and computationally efficient detection algorithms for real-time applications. Our work shows the adaptation of the popular YOLOv4 framework for predicting the objects and their locations in aerial images with high accuracy and inference speed. We utilized transfer learning for faster convergence of the model on the VisDrone DET aerial object detection dataset. The trained model resulted in a mean average precision (mAP) of 45.64% with an inference speed reaching 8.7 FPS on the Tesla K80 GPU and was highly accurate in detecting truncated and occluded objects. We experimentally evaluated the impact of varying network resolution sizes and training epochs on the performance. A comparative study with several contemporary aerial object detectors proved that YOLOv4 performed better, implying a more suitable detection algorithm to incorporate on aerial platforms. | ['Satish Shenoy B', 'Karunakar A K', 'Soham Hans', 'Akshatha K R', 'Aryaman Singh Samyal'] | 2022-03-18 | null | null | null | null | ['object-detection-in-aerial-images'] | ['computer-vision'] | [ 2.22630091e-02 -6.44610703e-01 3.43365848e-01 4.01556678e-02
5.38235046e-02 -8.96485746e-01 3.53368014e-01 5.63193671e-02
-5.60358107e-01 4.02557909e-01 -1.00630903e+00 -2.35140651e-01
-3.27574350e-02 -7.88184106e-01 -4.30126637e-01 -5.81041574e-01
-4.83790070e-01 2.77851164e-01 8.58148098e-01 -1.63088739e-01
1.31828291e-03 8.93602133e-01 -1.92737341e+00 -1.90485604e-02
5.26864469e-01 1.20986056e+00 5.01965404e-01 1.03124130e+00
7.25774944e-01 4.69899625e-01 -7.88065791e-01 -6.62920550e-02
9.02811766e-01 1.19654305e-01 -1.95232213e-01 2.06585869e-01
8.64384651e-01 -6.15118861e-01 -8.20045471e-02 9.49819267e-01
4.58726346e-01 2.86780626e-01 4.68258560e-01 -1.20826578e+00
-7.59692639e-02 -2.59116501e-01 -6.67838573e-01 5.69711983e-01
9.87115204e-02 3.18999261e-01 5.49989641e-01 -8.44165087e-01
3.61864299e-01 7.73874223e-01 9.45641994e-01 8.06995854e-03
-8.53994310e-01 -6.29241049e-01 1.64516848e-02 1.12591840e-01
-1.63118327e+00 2.16375459e-02 -5.71018010e-02 -5.87700248e-01
1.13007212e+00 2.45548427e-01 7.68354714e-01 2.88980633e-01
3.16019088e-01 3.13468546e-01 7.80932844e-01 -4.60678339e-01
1.27209634e-01 3.05886596e-01 -3.34371179e-01 1.17021954e+00
6.80375636e-01 2.31317505e-01 -2.03587078e-02 -2.02859938e-02
9.24020052e-01 6.78412020e-02 -1.30507886e-01 -3.96741211e-01
-9.85483825e-01 6.78264320e-01 7.12730050e-01 4.16620746e-02
-5.69125056e-01 -5.73484302e-02 4.93129939e-01 2.22720832e-01
2.49912247e-01 6.33564711e-01 -6.61863089e-01 2.72134244e-01
-1.02572322e+00 1.78285405e-01 3.88735801e-01 1.07078183e+00
4.71327573e-01 4.41944093e-01 1.97753727e-01 3.99820954e-01
5.35265915e-02 1.00934684e+00 1.80903062e-01 -7.24606752e-01
1.35105506e-01 9.44576144e-01 4.32855695e-01 -1.42016602e+00
-5.18106759e-01 -5.95057845e-01 -5.57051659e-01 7.16876626e-01
3.02610338e-01 -4.15629715e-01 -8.54062200e-01 9.15696204e-01
5.57710826e-01 -8.54023397e-02 -1.51526049e-01 1.17165411e+00
4.68304366e-01 9.29255664e-01 -4.92181219e-02 1.50474057e-01
1.24902010e+00 -8.50139558e-01 -4.83753607e-02 -4.22406316e-01
6.53544724e-01 -9.72103894e-01 5.60827374e-01 4.83580559e-01
-7.12609589e-01 -8.81555974e-01 -1.30725944e+00 3.70972633e-01
-4.31400388e-01 1.24538100e+00 5.50250411e-01 5.92736900e-01
-9.66541350e-01 4.98116344e-01 -9.30808127e-01 -7.02314198e-01
2.28399143e-01 8.04183543e-01 -2.24103928e-01 1.90889552e-01
-5.38459659e-01 8.63572121e-01 4.98224884e-01 1.77383497e-01
-8.38228106e-01 -4.15073186e-01 -4.63748336e-01 -1.26364246e-01
4.94466633e-01 -5.93519211e-01 1.08129811e+00 -1.04279625e+00
-1.24042404e+00 8.83681178e-01 6.12479389e-01 -9.36486542e-01
5.36229253e-01 -3.28072250e-01 -2.94928879e-01 2.64964670e-01
2.09786557e-03 9.60416198e-01 1.23072779e+00 -4.74332720e-01
-1.29293442e+00 -5.20472348e-01 1.15565874e-01 2.03129247e-01
-3.40497345e-01 -5.07959444e-03 -1.40055254e-01 -2.58127987e-01
-1.01584354e-02 -1.28966188e+00 1.67302210e-02 4.19687182e-01
2.06771538e-01 3.15973274e-02 1.30479407e+00 -4.60955232e-01
8.06693971e-01 -1.86536264e+00 4.99259643e-02 -1.06395632e-01
-9.67505872e-02 1.08792591e+00 1.69928372e-01 3.89033288e-01
2.10364401e-01 -5.73320329e-01 4.61364649e-02 2.08195299e-01
-5.06573141e-01 -7.72434622e-02 -4.60063547e-01 6.47372663e-01
1.16887026e-01 3.02396804e-01 -7.37576723e-01 -4.47073996e-01
7.30018973e-01 5.09339213e-01 -4.32988316e-01 3.25962007e-01
-1.58062071e-01 1.27469838e-01 -4.31813002e-01 1.10365427e+00
5.73899508e-01 2.24426631e-02 -5.33362217e-02 -1.50935262e-01
-5.71093321e-01 -4.62121308e-01 -1.14881456e+00 1.09759939e+00
-3.71870309e-01 1.03433776e+00 3.04164022e-01 -4.44102496e-01
1.12993789e+00 7.36901313e-02 1.23984866e-01 -8.62745196e-02
5.24893880e-01 4.21285257e-02 1.95568856e-02 -3.25884879e-01
8.67145956e-01 2.87627876e-01 3.61688018e-01 -1.78261802e-01
8.09690282e-02 -3.39767665e-01 4.08568799e-01 -9.86003503e-02
8.69053304e-01 2.06695184e-01 4.11207467e-01 -4.19181973e-01
4.81472701e-01 6.04044616e-01 1.32496417e-01 5.01196384e-01
-3.24129730e-01 1.22155800e-01 -1.71668217e-01 -9.84161079e-01
-7.08414018e-01 -5.27051151e-01 -2.52710849e-01 9.61270869e-01
9.33676884e-02 -2.42976233e-01 -7.16055751e-01 -2.77186692e-01
-1.36123709e-02 2.11244360e-01 -2.69811720e-01 1.25334024e-01
-2.60582685e-01 -6.98985338e-01 4.42933440e-01 5.96100688e-01
7.08458483e-01 -8.64207625e-01 -1.76708817e+00 9.35804844e-02
4.26657051e-01 -1.48968256e+00 2.43260369e-01 7.10505769e-02
-1.09886909e+00 -1.23966992e+00 -5.89894772e-01 -7.28401363e-01
7.58538008e-01 7.60042012e-01 7.29018986e-01 7.79299662e-02
-1.00452638e+00 4.21614498e-01 -3.68551791e-01 -7.14998960e-01
9.03618112e-02 9.14813206e-02 5.51355958e-01 -3.37545604e-01
2.06080139e-01 -5.82432225e-02 -5.92801213e-01 4.94049430e-01
-7.24428535e-01 -1.97078675e-01 5.56629241e-01 5.95421612e-01
4.61235672e-01 3.13668877e-01 -1.36369109e-01 -3.24100435e-01
1.82927728e-01 -2.05701888e-02 -1.66002917e+00 1.26215309e-01
-5.20581782e-01 -4.78083640e-01 8.79587948e-01 -2.46194735e-01
-4.51130092e-01 6.29841685e-01 4.45588857e-01 -6.53225899e-01
-1.72853276e-01 -6.46170452e-02 5.71381450e-01 -6.76157653e-01
8.92565131e-01 9.12685785e-03 -1.03857383e-01 1.25895143e-01
-1.28112093e-01 5.51825941e-01 4.52090055e-01 -1.20041840e-01
7.21241534e-01 4.16004986e-01 3.10262710e-01 -1.49486506e+00
-6.38738275e-01 -5.37405729e-01 -9.12289321e-01 -5.43999135e-01
7.84573853e-01 -1.20408726e+00 -8.27211082e-01 5.51304698e-01
-1.01206291e+00 -2.01254889e-01 6.59784079e-02 5.79963207e-01
-6.47323504e-02 1.50718763e-01 -1.86947584e-01 -9.22203004e-01
-5.50927520e-01 -1.13487542e+00 1.01753616e+00 3.30963284e-01
1.09215118e-01 -7.31008649e-01 -1.59174368e-01 6.00362644e-02
3.38733703e-01 3.19261134e-01 1.93144277e-01 -5.90755753e-02
-7.63208628e-01 -7.90916800e-01 -5.43437004e-01 3.16639751e-01
-3.88243310e-02 4.89437193e-01 -8.21515977e-01 -6.76083386e-01
-3.18627566e-01 -2.50137329e-01 5.97490251e-01 5.16877651e-01
7.01254129e-01 1.23185173e-01 -4.55085903e-01 6.08720303e-01
1.66717255e+00 2.80652225e-01 1.56799592e-02 3.05028528e-01
3.66157264e-01 2.69590437e-01 1.20278203e+00 7.12589562e-01
-3.27693909e-01 7.48339415e-01 8.89777601e-01 -7.66503289e-02
1.38037771e-01 1.57311615e-02 3.18283141e-01 1.58499062e-01
-6.14158988e-01 -3.68245006e-01 -8.89350951e-01 2.08880886e-01
-1.60324919e+00 -7.38091767e-01 -1.98650360e-01 2.27143621e+00
-1.79847963e-02 5.57183623e-02 1.63227484e-01 -7.95729384e-02
6.93672895e-01 -2.78831739e-02 -4.55666065e-01 -1.61363274e-01
1.71924248e-01 -8.98362324e-02 1.21127617e+00 1.47753894e-01
-1.54817796e+00 1.24188149e+00 6.07884407e+00 4.43746507e-01
-1.22791028e+00 -3.68873805e-01 7.68992975e-02 -1.08268604e-01
1.24649441e+00 -3.17457169e-01 -1.24639010e+00 1.13211535e-01
7.32305467e-01 1.63238451e-01 2.59574026e-01 1.37321591e+00
-7.73366988e-02 -3.34622979e-01 -5.04421055e-01 9.94062603e-01
6.40790686e-02 -1.13308442e+00 4.04766239e-02 -7.52960518e-02
6.46023870e-01 2.13258877e-01 9.45755914e-02 1.04475483e-01
1.63441345e-01 -5.61952591e-01 5.32527149e-01 2.03485619e-02
8.55173945e-01 -8.13344061e-01 1.00904703e+00 3.75601739e-01
-1.43116343e+00 -3.51002812e-01 -1.10186875e+00 -5.19965351e-01
-3.09717476e-01 -9.74107161e-02 -1.39577746e+00 1.56675190e-01
8.83611619e-01 6.61129236e-01 -7.26523876e-01 1.06381202e+00
-9.14508030e-02 3.59004170e-01 -5.26509047e-01 -2.91418344e-01
7.11845398e-01 -3.07221770e-01 6.39491677e-01 1.07768190e+00
6.04241610e-01 2.22319499e-01 4.95921224e-01 2.46469468e-01
1.75562724e-01 9.61687509e-03 -8.51075470e-01 -3.37751545e-02
3.75465751e-01 1.64183891e+00 -1.19527471e+00 -1.03363775e-01
-2.42058471e-01 9.45183635e-01 -3.09892483e-02 -2.71301270e-01
-1.08127391e+00 -3.26074541e-01 5.52463770e-01 3.58564109e-01
6.33854210e-01 -6.49012089e-01 2.82940239e-01 -7.68428266e-01
-2.71711379e-01 -6.11621678e-01 2.77626842e-01 -8.85225177e-01
-5.73506355e-01 1.14987087e+00 -4.46499558e-03 -1.52639663e+00
-4.45008948e-02 -1.31202185e+00 -1.80420160e-01 4.25809771e-01
-1.11936021e+00 -1.26873350e+00 -7.84608603e-01 3.14633131e-01
7.82233179e-01 -6.63310826e-01 1.01202905e+00 7.57144392e-02
-7.00592101e-01 1.98762223e-01 -1.38694137e-01 5.78296930e-03
5.46680152e-01 -9.32784200e-01 2.78202951e-01 1.27184308e+00
2.85054028e-01 2.18284041e-01 7.33944178e-01 -5.57705224e-01
-1.56876707e+00 -1.37102032e+00 1.01837926e-01 -4.26245123e-01
5.56357920e-01 5.01753315e-02 -4.59254384e-01 7.09033728e-01
8.84209350e-02 1.55702338e-01 2.51625329e-01 -3.42907101e-01
1.18472941e-01 -2.09516570e-01 -1.04830658e+00 3.05673063e-01
5.22466242e-01 2.26429459e-02 -1.17229693e-01 6.40673280e-01
3.14965606e-01 -6.34114027e-01 -4.56774950e-01 4.21233594e-01
5.42198122e-01 -1.16081429e+00 9.98677433e-01 -2.75223970e-01
-1.01662083e-02 -6.65670872e-01 -1.11229785e-01 -9.48882222e-01
-5.12931943e-01 -3.59347880e-01 2.15709820e-01 5.05618751e-01
9.53034759e-02 -4.92259026e-01 6.59423888e-01 3.29469085e-01
-1.62157983e-01 -4.73679423e-01 -5.37068963e-01 -7.96641290e-01
-7.58787811e-01 4.07239199e-02 5.92891574e-02 4.77580369e-01
-7.08777666e-01 2.63842285e-01 -4.18385834e-01 8.13432574e-01
3.77081871e-01 3.48457485e-01 9.04925942e-01 -1.42577255e+00
-8.68936852e-02 6.78850617e-03 -9.50984895e-01 -7.84952223e-01
-2.95141339e-01 -5.16502917e-01 -1.93867385e-01 -1.15394759e+00
-3.99885565e-01 -3.02947491e-01 1.93553254e-01 4.55989152e-01
2.61457860e-01 6.39782190e-01 2.04184204e-01 2.30231270e-01
-6.24781132e-01 1.08966276e-01 7.84618974e-01 1.28497660e-01
-2.79611647e-01 4.37028289e-01 1.75657347e-01 9.65074956e-01
8.08562875e-01 -3.45541626e-01 -2.74264693e-01 -6.37721956e-01
9.60399210e-02 -4.84509394e-02 5.42267561e-01 -1.80990565e+00
5.06881833e-01 3.85993272e-01 7.98265755e-01 -7.74172246e-01
6.32202744e-01 -1.14194047e+00 -1.09178275e-01 9.62858737e-01
3.54087085e-01 4.88999367e-01 7.93651402e-01 4.66187030e-01
-1.28511280e-01 -1.27140120e-01 9.96111631e-01 -1.11842483e-01
-1.21642160e+00 1.86209053e-01 -5.71302116e-01 -3.87240738e-01
1.58766878e+00 -4.16978121e-01 1.19073458e-01 1.21191032e-01
-3.15991521e-01 -8.82012956e-03 6.37323201e-01 3.29111248e-01
6.17612660e-01 -4.89284605e-01 -5.40824294e-01 5.55951715e-01
-5.95069490e-03 -9.18229818e-02 1.07033998e-01 6.67424917e-01
-1.31708694e+00 8.52975905e-01 -6.26389325e-01 -8.37750256e-01
-1.87486768e+00 4.28127795e-01 3.71776551e-01 2.48859927e-01
-5.31484604e-01 1.08473861e+00 -2.28645816e-01 -8.41265991e-02
1.89224839e-01 -5.85487783e-01 -3.82958353e-02 6.85002208e-02
5.98720491e-01 7.95700610e-01 2.12638751e-01 -5.02815604e-01
-4.70392734e-01 7.00083017e-01 4.07967940e-02 4.36487973e-01
1.12852907e+00 2.90656775e-01 3.34489569e-02 -3.59165780e-02
3.65876496e-01 -3.24403822e-01 -1.44619250e+00 1.25656754e-01
-3.49466980e-01 -6.33828640e-01 5.81395745e-01 -7.31616855e-01
-9.05154109e-01 8.20103526e-01 1.06382704e+00 3.11329126e-01
1.15622747e+00 -5.66819012e-01 5.11246443e-01 8.39153707e-01
8.18264127e-01 -8.82023454e-01 -3.07307601e-01 4.38891023e-01
5.84512413e-01 -1.40840054e+00 5.86512685e-01 -4.00539786e-01
-6.07867539e-01 1.46493614e+00 7.68726647e-01 -5.69563866e-01
2.02134863e-01 3.31732184e-01 9.83942077e-02 -2.24697039e-01
-3.83169800e-01 -2.49442294e-01 2.80677170e-01 5.91121852e-01
8.41409341e-02 2.27003530e-01 1.66182339e-01 -1.99476257e-01
-1.84804592e-02 -7.52145201e-02 3.84797096e-01 1.03131557e+00
-7.97427833e-01 -6.02691650e-01 -5.05230188e-01 2.30643630e-01
-3.47560793e-01 6.94248006e-02 -4.71452743e-01 1.22939801e+00
2.65783280e-01 7.58432746e-01 5.20343035e-02 -4.65250582e-01
3.79769474e-01 -3.85633588e-01 5.69866955e-01 -6.91794395e-01
-7.09111273e-01 -2.93011576e-01 -3.24569434e-01 -3.96431178e-01
-3.11261207e-01 -4.59634930e-01 -9.31031764e-01 -9.71797854e-02
-6.06828690e-01 2.75423820e-03 9.47152138e-01 4.84742939e-01
5.80125511e-01 4.99934107e-01 2.96339691e-01 -1.08642495e+00
-6.88127518e-01 -7.45905399e-01 -5.25421917e-01 -5.47512174e-01
1.77718960e-02 -7.93974400e-01 -1.53954402e-01 1.30576909e-01] | [8.433639526367188, -1.0596481561660767] |
5d4e5d91-de1a-4560-bae4-877befe2c962 | small-object-detection-via-pixel-level | null | null | https://www.frontiersin.org/articles/10.3389/fphys.2022.911297/full#h1 | https://www.frontiersin.org/articles/10.3389/fphys.2022.911297/pdf | Small Object Detection via Pixel Level Balancing With Applications to Blood Cell Detection | Object detection technology has been widely used in medical field, such as detecting the images of blood cell to count the changes and distribution for assisting the diagnosis of diseases. However, detecting small objects is one of the most challenging and important problems especially in medical scenarios. Most of the objects in medical images are very small but influential. Improving the detection performance of small objects is a very meaningful topic for medical detection. Current researches mainly focus on the extraction of small object features and data augmentation for small object samples, all of these researches focus on extracting the feature space of small objects better. However, in the training process of a detection model, objects of different sizes are mixed together, which may interfere with each other and affect the performance of small object detection. In this paper, we propose a method called pixel level balancing (PLB), which takes into account the number of pixels contained in the detection box as an impact factor to characterize the size of the inspected objects, and uses this as an impact factor. The training loss of each object of different size is adjusted by a weight dynamically, so as to improve the accuracy of small object detection. Finally, through experiments, we demonstrate that the size of objects in object detection interfere with each other. So that we can improve the accuracy of small object detection through PLB operation. This method can perform well with blood cell detection in our experiments. | ['Qingjie Kong', 'Minglei Tong', 'Pengzhi Chu', 'Yang Liu', 'Bin Hu'] | 2022-06-17 | null | null | null | frontiers-in-physiology-2022-6 | ['cell-detection', 'small-object-detection', 'blood-cell-detection'] | ['computer-vision', 'computer-vision', 'medical'] | [-1.80060845e-02 -4.07198966e-01 -1.71145350e-02 -1.42445847e-01
1.17584698e-01 1.86954334e-01 -6.73905015e-02 5.14596939e-01
-5.53253949e-01 3.27542424e-01 -7.82035515e-02 1.83490425e-01
1.67642394e-03 -1.03007388e+00 -3.12613785e-01 -1.15950704e+00
1.73306361e-01 2.11248487e-01 8.53465259e-01 1.08403154e-01
2.25334957e-01 6.65236115e-01 -1.15013885e+00 1.71363398e-01
8.71992290e-01 1.08455133e+00 4.80163783e-01 3.65368694e-01
-3.52310956e-01 7.40672708e-01 -9.33061779e-01 1.70176849e-01
2.39909828e-01 -5.08017480e-01 -6.48037493e-02 3.97660077e-01
-2.11191729e-01 -4.53698069e-01 -4.07398976e-02 1.15507174e+00
8.34939659e-01 -3.61229479e-02 7.17495918e-01 -8.78396511e-01
-1.21328823e-01 4.89009351e-01 -1.11133730e+00 8.17644894e-01
-3.36310863e-01 2.97477514e-01 2.15305194e-01 -5.32983661e-01
-2.25536022e-02 1.51236033e+00 3.46581340e-01 4.12177414e-01
-8.06107759e-01 -1.00946724e+00 -1.91485081e-02 4.06205326e-01
-1.48827422e+00 -2.44674578e-01 7.34403908e-01 -4.42284793e-01
2.06688881e-01 3.62328202e-01 6.20713294e-01 1.06499821e-01
2.48596415e-01 1.05484498e+00 7.16097414e-01 -4.67713386e-01
-1.76038831e-01 5.91346562e-01 3.15065175e-01 5.47708094e-01
5.84309220e-01 -1.52153760e-01 1.44819245e-01 1.09498024e-01
9.78075862e-01 4.04016793e-01 -4.44306195e-01 8.60474946e-04
-1.13286364e+00 6.68944776e-01 6.05642319e-01 6.08348489e-01
-1.79766044e-01 -1.93650171e-01 3.33924711e-01 -2.95410976e-02
2.57400900e-01 8.24894607e-02 -2.38234535e-01 2.45271146e-01
-3.33444238e-01 -1.27548967e-02 3.31540018e-01 4.74977016e-01
4.31322187e-01 -1.94050923e-01 -7.30709851e-01 9.12846446e-01
1.77197784e-01 5.03367245e-01 6.79613113e-01 -2.23253414e-01
4.41237569e-01 1.22681606e+00 -1.36654256e-02 -1.22632885e+00
-5.92180789e-01 -6.04069114e-01 -1.08274817e+00 1.79047838e-01
6.09121025e-01 -1.31987154e-01 -7.90276229e-01 1.28430438e+00
8.06258261e-01 2.28745162e-01 -3.01671088e-01 1.03961468e+00
8.53489816e-01 8.31429005e-01 1.43234998e-01 -6.63059473e-01
1.66191947e+00 -7.49731839e-01 -7.04191029e-01 -1.14077389e-01
4.75436479e-01 -7.89018273e-01 8.90240133e-01 2.08305031e-01
-8.02607417e-01 -7.67253160e-01 -9.11541522e-01 3.36990505e-01
-5.75720482e-02 5.13011932e-01 5.16855717e-01 5.66157818e-01
-2.26147875e-01 2.50489086e-01 -7.95166016e-01 -1.80853590e-01
8.15892279e-01 3.45869064e-01 1.52462304e-01 -2.17621699e-02
-9.89009142e-01 7.89944708e-01 5.83257735e-01 1.92614794e-01
-3.60665977e-01 -6.04240060e-01 -3.96486908e-01 1.81369320e-01
4.71841663e-01 -2.48806909e-01 7.39755154e-01 -9.31757629e-01
-7.43206739e-01 6.26850367e-01 1.21688530e-01 -2.46305898e-01
6.60352528e-01 1.21443108e-01 -4.78452891e-01 1.41627520e-01
-1.35972770e-02 4.48463053e-01 6.25860393e-01 -7.64775336e-01
-1.07194328e+00 -6.08056068e-01 -2.02150270e-01 2.03825086e-01
-6.00281775e-01 2.77364492e-01 -6.58475876e-01 -6.44668937e-01
5.41151408e-03 -4.39645350e-01 -3.64693999e-01 2.37917021e-01
-2.95647353e-01 -4.87592071e-01 9.60011423e-01 -4.26387548e-01
1.43503726e+00 -2.37108850e+00 -3.24201941e-01 2.46256053e-01
2.89037675e-01 3.63641918e-01 1.00450970e-01 -4.95985121e-01
3.54865104e-01 -5.27446829e-02 -9.68642235e-02 1.95442364e-01
-5.06007671e-01 -1.23034297e-02 1.12781718e-01 5.02698064e-01
3.93112838e-01 5.72213233e-01 -6.17127657e-01 -1.10888326e+00
3.94869417e-01 4.22794908e-01 -2.94430882e-01 3.76498759e-01
1.21664047e-01 5.05078435e-01 -8.97793233e-01 6.63073361e-01
1.04502928e+00 -1.69301376e-01 -4.05241221e-01 -4.74256277e-01
-5.28845154e-02 -4.05663639e-01 -1.54748559e+00 6.29616797e-01
-1.23047248e-01 2.12774307e-01 9.17054042e-02 -1.09806752e+00
1.01057088e+00 1.14654750e-01 6.24389887e-01 -7.13528395e-01
5.86139858e-01 1.66117385e-01 4.63601649e-01 -8.71525943e-01
-7.49895871e-02 -2.15897501e-01 5.28843403e-01 2.01347604e-01
-6.81892693e-01 2.12825552e-01 4.78608489e-01 -1.02749757e-01
6.84410810e-01 -6.23995245e-01 4.57214713e-01 -1.11084923e-01
8.12413812e-01 -1.50131613e-01 7.93489277e-01 5.03301084e-01
-3.89330328e-01 3.98355454e-01 3.89344096e-01 -3.07676464e-01
-6.71911836e-01 -6.53241575e-01 -6.32199705e-01 6.53802276e-01
6.96652710e-01 3.18224311e-01 -6.38092041e-01 -6.77121043e-01
1.40297323e-01 2.08063200e-01 -5.79143286e-01 -4.11141992e-01
-7.90193141e-01 -1.45298827e+00 1.61912486e-01 5.52784681e-01
6.25899255e-01 -1.48974299e+00 -1.02133524e+00 3.02994877e-01
4.63307947e-02 -8.45043600e-01 -4.17172670e-01 -1.83196545e-01
-9.39530194e-01 -1.25528324e+00 -9.94262934e-01 -9.87470448e-01
1.12382364e+00 4.74307269e-01 6.34412050e-01 6.25566483e-01
-9.03612673e-01 -2.65635371e-01 -3.64584625e-01 -9.05064344e-01
-4.34174538e-01 -1.90339163e-01 -1.73005402e-01 2.36117601e-01
4.82223302e-01 -5.83476610e-02 -9.08573627e-01 7.11168408e-01
-9.40048099e-01 -7.05961660e-02 9.18031394e-01 6.75476074e-01
5.05153120e-01 6.18560672e-01 4.68446046e-01 -9.23440635e-01
5.15385389e-01 -2.22194612e-01 -6.04895532e-01 2.00538546e-01
-3.28468651e-01 -2.45465115e-01 6.66137278e-01 -8.81513953e-01
-8.33535671e-01 -3.93053442e-01 8.21970999e-02 -3.45790893e-01
8.14957321e-02 1.29711583e-01 -3.00044030e-01 -8.23357552e-02
3.65731329e-01 3.09840530e-01 2.24043921e-01 -4.44175839e-01
-4.05039012e-01 8.38177145e-01 1.57621957e-03 6.62376210e-02
4.25986528e-01 4.11670297e-01 3.40762556e-01 -6.65322006e-01
-6.52569830e-01 -7.63650775e-01 -2.83662796e-01 -2.12281406e-01
6.14908040e-01 -5.39057970e-01 -7.62393773e-01 7.54932344e-01
-8.83013368e-01 1.26817793e-01 -2.58376390e-01 6.85674012e-01
2.71890610e-01 3.63314956e-01 -6.35695934e-01 -8.22144568e-01
-5.73318124e-01 -1.25420988e+00 7.20713198e-01 8.62756610e-01
3.09304148e-01 -7.05635607e-01 -3.35253775e-01 1.13578334e-01
3.34838182e-01 9.11154747e-02 9.34064031e-01 -5.81555605e-01
-5.93171656e-01 -4.97257650e-01 -5.46528578e-01 3.50614011e-01
4.37419266e-01 2.30578817e-02 -4.33428049e-01 -3.55856806e-01
3.90469611e-01 1.10157423e-01 1.02317643e+00 8.09873641e-01
1.29410100e+00 -3.62395793e-01 -8.16529274e-01 4.35489416e-01
1.28805232e+00 6.20175064e-01 6.50599957e-01 3.19776356e-01
7.03009605e-01 6.29853308e-01 1.00916028e+00 5.64203382e-01
-2.60250926e-01 3.60288501e-01 2.43121251e-01 -3.75778079e-01
-1.88876197e-01 3.92729789e-01 1.63758663e-03 4.78440821e-01
3.43841985e-02 -1.39219299e-01 -6.43756151e-01 4.21530187e-01
-1.51438510e+00 -9.50065792e-01 -3.38827938e-01 2.33234239e+00
9.63407695e-01 4.08452004e-01 2.93074828e-02 3.01751614e-01
1.11122537e+00 -1.58968642e-01 -7.50556231e-01 3.74967426e-01
-9.84364897e-02 -1.74104184e-01 3.79233658e-01 -5.10674976e-02
-1.03059840e+00 2.51301736e-01 5.00829744e+00 1.33653748e+00
-1.29387057e+00 -1.42960818e-02 1.07227075e+00 -1.54765159e-01
2.46923298e-01 -3.93679619e-01 -1.17887807e+00 9.84949589e-01
-5.54188937e-02 2.00340319e-02 -7.53231859e-03 7.33555079e-01
5.72059631e-01 -4.83003467e-01 -6.90929890e-01 1.04264963e+00
-2.25450024e-02 -8.49239111e-01 2.21541390e-01 -1.39267325e-01
4.26476628e-01 -4.94910687e-01 -1.14446521e-01 1.51441619e-01
-4.93214935e-01 -6.59808457e-01 3.90970856e-01 2.12944284e-01
4.61567611e-01 -6.55265570e-01 1.05751157e+00 5.13895035e-01
-1.15762663e+00 -2.30945066e-01 -8.13683391e-01 1.27098933e-01
3.94357108e-02 1.01706350e+00 -7.69744217e-01 -3.50540224e-03
5.94075322e-01 4.67065394e-01 -7.12414563e-01 1.60801589e+00
5.37684485e-02 5.94148338e-01 -4.05631542e-01 -4.32377130e-01
-2.05885619e-01 -3.29850942e-01 5.27236938e-01 1.06821513e+00
1.82381988e-01 4.20305938e-01 4.57528025e-01 7.72185087e-01
2.41342071e-03 6.20812297e-01 1.00787453e-01 3.51105601e-01
4.13823187e-01 1.36957157e+00 -1.07547629e+00 -6.05242133e-01
-2.00136945e-01 3.33428741e-01 -7.18615428e-02 -9.47002098e-02
-9.01228189e-01 -6.96818531e-01 -4.54722494e-02 4.48453128e-01
2.26266712e-01 4.22679275e-01 -2.02138856e-01 -8.45310926e-01
2.08123237e-01 -5.46630204e-01 6.42607749e-01 -3.95485699e-01
-1.15222025e+00 2.15788454e-01 -2.65767217e-01 -1.48398089e+00
6.06248975e-01 -5.01379907e-01 -9.27077055e-01 7.58870959e-01
-1.47665107e+00 -7.85938084e-01 -7.53656030e-01 4.57298458e-01
6.03082597e-01 4.53778990e-02 1.76631697e-02 5.80030620e-01
-1.03954232e+00 6.74699962e-01 2.10254982e-01 3.03859711e-01
6.04001462e-01 -8.04847658e-01 -3.09652925e-01 7.75515139e-01
-3.52969974e-01 4.31138694e-01 6.04833066e-01 -5.62942445e-01
-8.48312140e-01 -1.05191600e+00 1.78301215e-01 1.27671033e-01
1.41625494e-01 -1.20180398e-02 -1.04242575e+00 7.29802325e-02
-3.87790233e-01 4.08921480e-01 3.62537563e-01 -3.75540644e-01
3.91092658e-01 -5.67752182e-01 -1.44604766e+00 4.81500924e-01
4.96790379e-01 3.98975283e-01 -3.44379872e-01 4.89961684e-01
5.74306190e-01 -4.66166228e-01 -5.54241896e-01 7.08047092e-01
3.91832650e-01 -8.35544944e-01 1.00843930e+00 -2.71942228e-01
1.34069502e-01 -5.51281214e-01 5.26997268e-01 -1.00022829e+00
-3.29231232e-01 9.24192369e-02 -1.68958250e-02 1.30401683e+00
-5.39069436e-02 -6.35652363e-01 7.74200261e-01 2.76809961e-01
1.14253365e-01 -9.26400602e-01 -7.14958966e-01 -5.62778056e-01
-3.13305616e-01 2.11658001e-01 4.99228060e-01 7.19392300e-01
-1.79970950e-01 3.24098051e-01 2.43097730e-02 5.67131527e-02
4.93522465e-01 2.72868931e-01 6.71664178e-01 -1.27654684e+00
-1.43630624e-01 -6.28398955e-01 -5.77028096e-01 -1.05190539e+00
-7.22241402e-01 -3.56346846e-01 1.34916473e-02 -1.56656563e+00
6.95669889e-01 -8.74319255e-01 -6.31605506e-01 2.70040452e-01
-8.73087168e-01 2.78339386e-01 -1.14034735e-01 4.05092478e-01
-4.45694476e-01 3.32565248e-01 1.75777233e+00 -4.88879740e-01
-4.01312798e-01 5.30856133e-01 -5.25263309e-01 7.13864446e-01
6.82519734e-01 -4.62807626e-01 -2.02585563e-01 -1.04142681e-01
-3.55975211e-01 -1.89472064e-01 1.89879254e-01 -1.12629580e+00
2.14381292e-01 -3.15491319e-01 8.39970052e-01 -6.48957074e-01
-3.64368409e-02 -8.87913704e-01 -2.85092801e-01 1.10741353e+00
3.68366241e-02 -5.09040356e-01 3.63192141e-01 3.27875733e-01
-9.91638079e-02 -5.31020880e-01 1.13234735e+00 -3.26491654e-01
-6.62154853e-01 4.34263170e-01 8.32933281e-03 2.71171108e-02
1.47244108e+00 -5.28799534e-01 -2.21113279e-01 2.91745335e-01
-3.14433545e-01 6.20915234e-01 -7.23781157e-03 1.90552637e-01
6.07050657e-01 -1.11259854e+00 -8.99769247e-01 2.33829528e-01
1.60200670e-01 4.19895440e-01 5.43138146e-01 1.21904540e+00
-6.86896682e-01 3.76170427e-02 -2.26602629e-01 -6.75112426e-01
-1.68117428e+00 7.81824768e-01 4.45104539e-01 -3.12204957e-01
-4.79855806e-01 8.71566832e-01 6.29257858e-01 2.81994402e-01
1.89255685e-01 -4.55964237e-01 -8.61535251e-01 1.20351024e-01
1.04113686e+00 5.05514443e-01 -2.04105482e-01 -3.34216148e-01
-2.62589604e-01 7.51638293e-01 -3.74158680e-01 7.15438306e-01
1.10562229e+00 -1.22940294e-01 -2.69969851e-01 3.59933883e-01
8.79752100e-01 2.48227015e-01 -9.89825249e-01 -3.15531075e-01
-4.85356659e-01 -7.45473444e-01 1.15302987e-01 -5.74366629e-01
-1.40892684e+00 9.95217502e-01 9.66548979e-01 2.13773832e-01
1.31758606e+00 1.52968979e-02 8.27782929e-01 5.55462874e-02
1.47146687e-01 -9.05359268e-01 1.37206167e-01 -1.45223245e-01
4.57756728e-01 -1.42590821e+00 3.93409848e-01 -5.86658835e-01
-4.77973223e-01 1.09744966e+00 9.93438303e-01 1.30515620e-01
5.89403570e-01 3.05910826e-01 5.86714521e-02 -1.66630059e-01
-1.15004256e-01 -1.00456871e-01 1.90359414e-01 1.71540588e-01
-4.97331144e-03 2.03462951e-02 -6.14592612e-01 6.05125070e-01
5.04304826e-01 7.11653382e-02 3.48669767e-01 8.87526810e-01
-1.09572423e+00 -6.62059069e-01 -7.90017724e-01 9.26347375e-01
-6.13915741e-01 2.42216140e-02 2.35710308e-01 5.54650962e-01
4.91679460e-01 8.34208310e-01 4.03183162e-01 -1.06314100e-01
4.63261873e-01 -7.83310175e-01 3.34614307e-01 -5.19429922e-01
-3.33447695e-01 2.44848624e-01 -5.85086107e-01 1.18141985e-02
-3.05617362e-01 -4.24520105e-01 -1.59670258e+00 -1.21326663e-01
-9.61578131e-01 1.80591330e-01 1.66156381e-01 7.43446827e-01
-1.87146768e-01 6.11258864e-01 9.50263798e-01 -3.46574455e-01
-5.53292334e-01 -1.14969671e+00 -9.00124550e-01 5.71580231e-01
1.54207468e-01 -6.98420346e-01 -3.26974005e-01 -1.21719800e-02] | [14.780082702636719, -3.052536964416504] |
9942c9a4-4648-401b-9d77-8687602c4292 | characterizing-the-value-of-information-in | 2010.03574 | null | https://arxiv.org/abs/2010.03574v2 | https://arxiv.org/pdf/2010.03574v2.pdf | Characterizing the Value of Information in Medical Notes | Machine learning models depend on the quality of input data. As electronic health records are widely adopted, the amount of data in health care is growing, along with complaints about the quality of medical notes. We use two prediction tasks, readmission prediction and in-hospital mortality prediction, to characterize the value of information in medical notes. We show that as a whole, medical notes only provide additional predictive power over structured information in readmission prediction. We further propose a probing framework to select parts of notes that enable more accurate predictions than using all notes, despite that the selected information leads to a distribution shift from the training data ("all notes"). Finally, we demonstrate that models trained on the selected valuable information achieve even better predictive performance, with only 6.8% of all the tokens for readmission prediction. | ['Chenhao Tan', 'Ziad Obermeyer', 'Sendhil Mullainathan', 'Shantanu Karnwal', 'Chao-Chun Hsu'] | 2020-10-07 | null | https://aclanthology.org/2020.findings-emnlp.187 | https://aclanthology.org/2020.findings-emnlp.187.pdf | findings-of-the-association-for-computational | ['readmission-prediction'] | ['medical'] | [ 7.95482397e-02 5.47475815e-01 -6.98175848e-01 -4.40174997e-01
-1.32788229e+00 -4.61513162e-01 2.54397858e-02 9.20109928e-01
-4.09250319e-01 8.30500066e-01 9.18536961e-01 -5.35014749e-01
-4.80064750e-01 -7.93477952e-01 -3.75813514e-01 -4.90222186e-01
-1.93764679e-02 8.48361135e-01 -2.29909346e-01 2.21507311e-01
1.73133984e-01 2.20416769e-01 -8.48482847e-01 9.09472704e-01
8.71600807e-01 8.78802955e-01 4.75504845e-02 5.96324444e-01
-1.58983558e-01 1.03797424e+00 -5.39011478e-01 -5.18396974e-01
1.30218685e-01 -3.71594518e-01 -1.22420430e+00 -3.13228458e-01
-1.02041945e-01 -4.63298053e-01 -2.48015329e-01 5.12887239e-01
7.57918775e-01 -1.12277016e-01 6.98775828e-01 -6.98084176e-01
-4.47957814e-01 1.14853668e+00 1.51510164e-01 1.85821980e-01
3.47746491e-01 -1.28156230e-01 1.23231041e+00 -7.11439669e-01
6.77441597e-01 5.90219438e-01 9.25296903e-01 8.10500383e-01
-1.08232820e+00 -3.75879139e-01 -2.74279237e-01 -1.53219298e-01
-9.24936891e-01 -5.80968738e-01 3.00772876e-01 -6.21465266e-01
1.13646710e+00 6.57923281e-01 7.27807462e-01 9.07294035e-01
3.97109240e-01 8.34348738e-01 4.75708693e-01 -2.65710473e-01
1.24159463e-01 3.94244403e-01 4.67427731e-01 5.79253614e-01
4.74586725e-01 5.15170302e-03 -4.82247621e-01 -7.40164101e-01
4.76347387e-01 7.65668869e-01 -4.53632921e-01 4.99994680e-02
-1.26963568e+00 6.38716519e-01 6.68647662e-02 3.85365427e-01
-6.08911097e-01 -3.93745780e-01 4.97629583e-01 2.48460636e-01
4.94176775e-01 8.24195027e-01 -1.00583541e+00 -4.77995008e-01
-1.09700418e+00 6.88524917e-02 1.08242440e+00 9.32594121e-01
3.59303534e-01 -5.29896080e-01 -6.72920704e-01 9.76362467e-01
-2.62298405e-01 2.86236018e-01 6.05392992e-01 -8.13253999e-01
6.78813279e-01 7.47243643e-01 2.26095945e-01 -4.60558206e-01
-7.94656873e-01 -6.49126053e-01 -1.12697828e+00 -6.14926994e-01
2.80114144e-01 -2.86799282e-01 -6.11450315e-01 1.31213915e+00
-2.02320680e-01 -3.45968693e-01 2.14065015e-01 1.50117636e-01
8.05335462e-01 4.13027823e-01 8.53129104e-02 -4.39309895e-01
1.09616745e+00 -6.73836887e-01 -6.89516187e-01 1.81174949e-01
1.36441410e+00 -5.07945359e-01 7.48691082e-01 2.73532838e-01
-1.34142840e+00 -1.40332192e-01 -1.44214556e-01 8.44642818e-02
1.29700586e-01 1.24451660e-01 3.41818213e-01 4.00278747e-01
-1.01344931e+00 8.81880999e-01 -8.22825909e-01 -3.18483412e-01
7.04074144e-01 4.97930348e-01 -3.68049681e-01 -6.95142597e-02
-8.06608438e-01 7.46530592e-01 2.76575446e-01 -4.62896585e-01
-2.21296728e-01 -1.31291580e+00 -7.43929267e-01 5.03874063e-01
-5.66809885e-02 -9.56283033e-01 1.27904284e+00 -6.31874919e-01
-8.19989443e-01 9.44101870e-01 -3.52004230e-01 -5.25552630e-01
4.72036302e-01 -3.76610279e-01 -2.34186590e-01 1.20072179e-01
1.84786133e-02 3.61120701e-01 1.75292969e-01 -1.02933729e+00
-8.01840246e-01 -3.32567185e-01 -3.98335338e-01 -2.68971384e-01
-4.54599887e-01 -4.89055589e-02 1.92428008e-02 -5.58774650e-01
-1.29289538e-01 -6.43015206e-01 -4.33394641e-01 -3.96669328e-01
-7.79226720e-01 -2.09618539e-01 -6.21120743e-02 -6.85723484e-01
1.94135046e+00 -2.06990170e+00 -3.45383167e-01 1.76364973e-01
4.98233229e-01 1.09953605e-01 2.32522395e-02 8.13040733e-01
-1.76180864e-03 5.69807470e-01 -3.16219270e-01 -3.78214836e-01
-3.12147945e-01 2.79769748e-01 -4.30629373e-01 -3.01062688e-02
1.15798324e-01 1.05203736e+00 -8.02966058e-01 -5.58479190e-01
4.61648628e-02 1.91264436e-01 -9.95877445e-01 4.74437058e-01
2.86813319e-01 5.52642047e-01 -6.34565055e-01 5.98042071e-01
3.34933221e-01 -8.52186620e-01 1.64610192e-01 8.10244679e-02
7.49992654e-02 9.11005914e-01 -3.53215069e-01 1.33003426e+00
-1.29670143e-01 1.26580903e-02 -4.14614290e-01 -8.30765247e-01
7.00212777e-01 5.57944477e-01 1.01185155e+00 -4.05306667e-01
-1.66057751e-01 1.76978856e-01 6.66270256e-02 -8.94258261e-01
3.06100696e-01 -2.72967666e-01 -4.57378849e-03 4.61446345e-01
-1.22005530e-01 3.42684478e-01 -2.58414567e-01 1.41022310e-01
1.50424063e+00 -5.71429431e-01 7.74517596e-01 -3.70816261e-01
-9.03218091e-02 1.95060045e-01 8.14896107e-01 1.11919510e+00
-1.49751619e-01 8.65556777e-01 4.24746364e-01 -6.90542459e-01
-1.05684376e+00 -8.76291215e-01 -4.85287696e-01 7.02993810e-01
-3.22967708e-01 -6.64628744e-01 -5.28496504e-01 -8.82089734e-01
1.90242231e-01 6.75436080e-01 -6.63032234e-01 -1.69081613e-01
-5.19793749e-01 -8.08670223e-01 7.02529192e-01 7.98478842e-01
1.73212420e-02 -1.18498719e+00 -8.23062062e-01 3.80748659e-01
-3.80802631e-01 -6.20953739e-01 -4.57805455e-01 5.52131832e-01
-1.35428250e+00 -1.16496646e+00 -7.26318896e-01 -6.11693382e-01
6.64267004e-01 -2.12041155e-01 1.54203200e+00 5.51243961e-01
-2.26478323e-01 4.82107103e-01 -4.88321453e-01 -4.42783087e-01
-5.56272209e-01 4.92712826e-01 -2.76755631e-01 -3.84012133e-01
6.47881091e-01 -2.18554825e-01 -7.83010364e-01 -2.84824193e-01
-7.48476326e-01 9.08628404e-02 5.92713356e-01 1.18767643e+00
4.79124576e-01 -2.60742098e-01 9.03107703e-01 -1.46945024e+00
6.29301906e-01 -7.04698682e-01 3.58590752e-01 2.37652108e-01
-1.01034737e+00 1.35984421e-01 6.62822008e-01 -5.85582480e-02
-7.66310155e-01 -9.95255336e-02 -4.75674152e-01 6.25761300e-02
-3.64407748e-01 6.55642271e-01 3.85703474e-01 7.75915027e-01
5.44607759e-01 2.23621860e-01 -1.93654463e-01 -7.90367782e-01
-3.60198915e-01 7.51710176e-01 1.01849367e-03 -2.98301637e-01
1.19480178e-01 1.74395144e-01 -2.25921765e-01 -4.26471442e-01
-1.09751201e+00 -8.35935831e-01 -7.13065445e-01 3.94946486e-01
7.93765008e-01 -5.42762578e-01 -7.91944742e-01 -1.63245783e-03
-1.01330638e+00 -2.63984084e-01 -8.33782673e-01 7.48360395e-01
-4.30483431e-01 1.63313255e-01 -1.10613143e+00 -8.99776340e-01
-6.08521223e-01 -8.56014967e-01 9.79417384e-01 -3.60265881e-01
-7.32544065e-01 -1.15097666e+00 2.11238235e-01 2.80237675e-01
1.74020365e-01 -1.25676155e-01 1.68259823e+00 -1.36464930e+00
-3.36947560e-01 -2.37210646e-01 -2.52002850e-02 1.83037415e-01
5.44648170e-01 -3.97803426e-01 -9.08161521e-01 -1.61713779e-01
4.55814367e-03 -1.73301414e-01 1.03795421e+00 7.71304607e-01
1.49946320e+00 -8.55916440e-01 -6.56718850e-01 5.37881672e-01
1.21685636e+00 4.98398006e-01 5.62097132e-01 2.87018836e-01
3.97242188e-01 7.17656791e-01 3.29193413e-01 1.04936349e+00
4.65793699e-01 4.83418740e-02 9.32776183e-02 -1.37306511e-01
3.91314864e-01 -4.33166325e-01 -1.64771155e-01 1.03865349e+00
-1.32473707e-01 -3.25600356e-01 -1.21627116e+00 6.75665617e-01
-1.77435875e+00 -1.02711189e+00 -2.06379835e-02 2.43775654e+00
1.14596450e+00 -5.46165779e-02 5.10076880e-02 2.79703468e-01
2.59044796e-01 -2.59297520e-01 -3.32885891e-01 -1.28208354e-01
1.03317969e-01 3.05024654e-01 5.74613214e-01 3.22690129e-01
-8.01023543e-01 3.63549054e-01 7.78063536e+00 3.29780996e-01
-7.35893667e-01 -2.04662934e-01 1.08662260e+00 -8.27145427e-02
-6.85776055e-01 -3.76034826e-01 -8.34365487e-01 5.67347527e-01
9.74525750e-01 -4.73129749e-02 -2.14860514e-01 8.00280154e-01
3.56091827e-01 1.45299196e-01 -1.62036264e+00 9.54983950e-01
-2.44160995e-01 -1.69063175e+00 5.00891030e-01 1.27986535e-01
6.29782438e-01 -1.08146131e-01 -4.54727486e-02 2.17961714e-01
2.17562228e-01 -1.25866294e+00 1.72212973e-01 7.39790380e-01
7.51154363e-01 -5.53865671e-01 1.12173200e+00 7.80822575e-01
-7.58382618e-01 -3.13247859e-01 -4.15268123e-01 -6.59566745e-02
1.29584391e-02 9.74011481e-01 -1.14943206e+00 4.35035050e-01
7.48228371e-01 9.16530728e-01 -3.44927460e-01 1.30213141e+00
4.47693944e-01 1.04696155e+00 9.72839966e-02 2.27554142e-01
-1.58732340e-01 1.23421334e-01 1.04027167e-01 1.45454669e+00
6.78185701e-01 4.15144622e-01 1.97686970e-01 4.89139736e-01
-3.54664713e-01 3.14528704e-01 -7.49516726e-01 -2.56040618e-02
4.46730077e-01 5.51703036e-01 -2.95173824e-01 -6.34566963e-01
-4.30561155e-01 6.80367708e-01 2.54361033e-01 1.94881976e-01
-2.15980247e-01 2.83778608e-02 5.30609250e-01 5.01482725e-01
-5.99548630e-02 4.13030148e-01 -6.52419925e-01 -1.16925287e+00
-1.47316113e-01 -6.69601977e-01 7.91620135e-01 -4.20114189e-01
-1.54292166e+00 5.78939319e-01 -3.86094540e-01 -1.29464293e+00
-4.62781817e-01 -3.44165742e-01 -2.29845330e-01 6.72509432e-01
-1.42250776e+00 -6.32427514e-01 -2.81196572e-02 4.23021793e-01
5.19454837e-01 -1.63455084e-01 1.32282698e+00 4.54057097e-01
-2.51873493e-01 7.95739174e-01 3.56381685e-01 5.21734297e-01
8.23981702e-01 -1.24776137e+00 3.41814496e-02 1.85285285e-02
-2.65245795e-01 7.96649635e-01 2.48429134e-01 -6.93045616e-01
-9.49820995e-01 -1.21491206e+00 1.68753946e+00 -8.79477143e-01
1.08772397e-01 2.31736153e-01 -1.07366204e+00 1.03171217e+00
-1.42316267e-01 -1.98836878e-01 1.36847317e+00 3.67950410e-01
-2.84666300e-01 4.05221432e-02 -1.23536360e+00 3.24104130e-01
1.02079844e+00 -4.68095422e-01 -7.91747212e-01 1.32586315e-01
6.63223267e-01 -1.08265273e-01 -1.26343238e+00 6.28951490e-01
6.49736166e-01 -1.07521331e+00 8.30683470e-01 -1.12678301e+00
8.72427583e-01 5.93416691e-01 2.37553809e-02 -1.23248160e+00
-5.95121205e-01 -3.81143332e-01 4.93025361e-03 1.07384729e+00
8.25728357e-01 -6.61364555e-01 1.03196573e+00 1.07402217e+00
-9.67124030e-02 -1.12046874e+00 -7.23745883e-01 -3.45449507e-01
3.04343939e-01 -2.07811683e-01 6.59802020e-01 1.21477914e+00
5.64538121e-01 1.00928821e-01 -4.41522121e-01 -2.85953790e-01
2.68152684e-01 2.52836287e-01 2.57368684e-01 -1.60934043e+00
-4.31852579e-01 -2.57914871e-01 -1.58804819e-01 -1.16218996e+00
-2.00501189e-01 -1.13801515e+00 -1.36423841e-01 -1.99770522e+00
9.21815753e-01 -8.48867714e-01 -7.52953947e-01 7.38285065e-01
-4.37167466e-01 -2.47598603e-01 1.12051675e-02 5.35912514e-01
-4.41710830e-01 1.23700723e-01 1.18556702e+00 -1.47389278e-01
-4.53241438e-01 5.77524960e-01 -8.07136714e-01 6.07795000e-01
8.30296695e-01 -8.21556270e-01 -2.27949888e-01 -3.56709212e-01
2.83723325e-01 7.41630256e-01 -4.30167727e-02 -6.36540592e-01
8.23870972e-02 -2.06004813e-01 5.38428068e-01 -6.74281538e-01
-4.86132652e-02 -9.13214326e-01 -2.71837655e-02 7.52258837e-01
-1.12102866e+00 1.65525615e-01 -7.37685114e-02 2.99973547e-01
-2.94186234e-01 -2.90033758e-01 4.40450251e-01 -3.03311080e-01
1.34855404e-01 4.46090043e-01 -4.73632455e-01 3.25374752e-01
4.94502455e-01 -9.98753458e-02 -1.86200768e-01 -3.04528922e-01
-9.79803205e-01 2.34846082e-02 1.76938757e-01 3.25264856e-02
7.09587276e-01 -9.01340187e-01 -7.95328856e-01 2.82540858e-01
2.97088742e-01 3.80931087e-02 1.23257175e-01 9.06879306e-01
-1.47638395e-01 7.90094435e-01 1.63152725e-01 -3.82085234e-01
-1.37566650e+00 5.89520693e-01 1.47676721e-01 -7.68594980e-01
-1.03635776e+00 6.17844582e-01 2.54255950e-01 -5.05519748e-01
4.32615936e-01 -9.21652615e-01 -1.59999192e-01 -5.09164259e-02
6.49839282e-01 4.37902540e-01 1.04598537e-01 4.26596366e-02
-2.44659305e-01 1.21765532e-01 -1.46207837e-02 2.85427928e-01
1.55440056e+00 3.55591476e-02 -8.17529410e-02 6.26599967e-01
1.06323051e+00 3.15066010e-01 -7.27576673e-01 -1.87861800e-01
3.99662584e-01 -2.49966115e-01 -4.20623958e-01 -1.06450796e+00
-7.38851607e-01 9.80367124e-01 2.18701944e-01 3.30243319e-01
9.31171119e-01 1.27196550e-01 1.01230097e+00 6.74038827e-01
2.87370026e-01 -7.65090108e-01 -2.29540482e-01 4.51785952e-01
4.15556461e-01 -1.35733986e+00 -2.66144603e-01 -2.90106416e-01
-7.30389595e-01 1.01572096e+00 7.38321468e-02 2.50806987e-01
9.78409767e-01 5.95421970e-01 1.47757068e-01 9.56906658e-03
-1.07687497e+00 1.48971558e-01 1.04949147e-01 5.46152711e-01
8.73946607e-01 2.00003251e-01 -1.90921411e-01 1.22581923e+00
-2.77068079e-01 3.91207427e-01 3.70628446e-01 7.33537257e-01
-6.39479280e-01 -1.24485767e+00 -3.97657417e-02 1.40412712e+00
-1.19726074e+00 -6.92837119e-01 -3.33641678e-01 4.05242831e-01
-8.14303532e-02 9.38814282e-01 2.47670095e-02 -3.29980969e-01
2.99750179e-01 5.54149926e-01 1.13296241e-01 -9.62834895e-01
-1.24608111e+00 -1.54227048e-01 2.71217465e-01 -2.86053091e-01
-2.19632655e-01 -3.99694026e-01 -1.52644455e+00 -3.43865544e-01
-5.42424470e-02 6.04659438e-01 -5.82904890e-02 8.58933806e-01
7.08195925e-01 5.81738830e-01 3.74649733e-01 3.61724496e-02
-8.04419160e-01 -8.35853577e-01 -6.64351225e-01 5.20400763e-01
7.13124096e-01 8.07460472e-02 -2.99615771e-01 2.53147244e-01] | [8.025598526000977, 6.58673095703125] |
15f156a7-3888-443a-b0b0-ab55bef7e9dd | joint-spectrum-and-power-allocation-for-v2x | 2302.14704 | null | https://arxiv.org/abs/2302.14704v1 | https://arxiv.org/pdf/2302.14704v1.pdf | Joint Spectrum and Power Allocation for V2X Communications with Imperfect CSI | In Vehicle-to-Everything (V2X) communication, the high mobility of vehicles generates the Doppler shift which leads to channel uncertainties. Moreover, the reasons for channel uncertainties also include the finite channel feedback, channels state information (CSI) loss and latency. With this concern, we formulate a joint spectrum and power allocation problem for V2X communication with imperfect CSI. Specifically, the sum capacity of cellular user equipments (CUEs) is maximized subject to the minimum Signal-to-Interference-and-Noise Ratio (SINR) requirements of CUEs and the outage probability constraints of vehicular user equipments (VUEs). Then, two different robust resource allocation approaches are designed to solve the problem. One is Bernstein Approximation-based Robust Resource Allocation approach. More specifically, Bernstein approximations are employed to convert the chance constraint into a calculable constraint, and Bisection search method is proposed to obtain the optimal allocation solution with low complexity. Then, for further reducing the computational complexity, Self-learning Robust Resource Allocation approach, which includes a learning method and an analytical mapping method, is proposed as the second approach. The learning method is devised to learn the uncertainty set which transforms the chance constraint into calculable constraints, and the analytical mapping method is proposed to obtain closed-form solutions of the resource allocation problem. Finally, the simulation results prove that the proposed approaches can improve the capacity of all CUEs effectively whilst ensuring the reliability of the channel. | ['Li Feng', 'Guanhua Chai', 'Jiayi Liu', 'Weihua Wu', 'Peng Wang'] | 2023-02-21 | null | null | null | null | ['self-learning'] | ['natural-language-processing'] | [-3.08474619e-03 3.85795951e-01 -4.48392123e-01 2.53016442e-01
-8.23736608e-01 -3.02572191e-01 3.54511701e-02 -3.82488370e-01
-1.76357150e-01 1.16732252e+00 -1.98244140e-01 -6.99715793e-01
-6.16168141e-01 -6.42390966e-01 -4.60641086e-01 -1.19203830e+00
-1.91628754e-01 -2.53507495e-01 1.33775130e-01 -2.80221866e-04
1.61307648e-01 4.34525758e-01 -1.00486946e+00 -9.20054078e-01
9.65536654e-01 1.57943964e+00 4.75511342e-01 3.16149533e-01
-2.28411213e-01 4.18691099e-01 -5.39632738e-01 9.48685408e-02
3.11203003e-01 -3.32742542e-01 -2.47077178e-02 1.75612301e-01
-6.42292500e-01 -4.67502803e-01 -7.35265732e-01 1.01422870e+00
5.01491249e-01 9.60448533e-02 7.85229385e-01 -1.67832601e+00
-1.13394514e-01 9.34908018e-02 -6.47254169e-01 1.16682746e-01
-1.87330529e-01 -6.85293302e-02 4.50000584e-01 -7.93911636e-01
2.15859562e-01 1.14809906e+00 2.87337989e-01 4.38768089e-01
-4.61867392e-01 -9.54068601e-01 -7.60438386e-03 3.86960059e-01
-1.78808403e+00 -5.87537348e-01 4.72797751e-01 -2.34401643e-01
2.09950343e-01 4.00056839e-01 5.04800618e-01 1.86848551e-01
1.48272455e-01 6.95189953e-01 5.53815961e-01 -3.66430014e-01
3.05857927e-01 3.01761627e-01 -4.13544059e-01 4.93992269e-01
5.28419912e-01 1.88268080e-01 9.82660428e-03 -1.33104101e-01
7.74507403e-01 -2.65869915e-01 -5.20692050e-01 -3.24034423e-01
-5.59882104e-01 6.29515946e-01 2.47021973e-01 6.66087717e-02
-5.04441857e-01 1.27265990e-01 1.18204109e-01 2.85288453e-01
-3.56148072e-02 -1.77139387e-01 -1.51612893e-01 4.39028777e-02
-6.68910325e-01 2.22077919e-03 6.63200557e-01 1.47084951e+00
5.04124165e-01 4.22939837e-01 -2.92439252e-01 5.39431155e-01
9.20888722e-01 1.05631602e+00 -3.25001270e-01 -8.95233154e-01
7.56934226e-01 6.04191469e-03 5.37932754e-01 -1.01575196e+00
-3.24118257e-01 -8.92420709e-01 -8.72458100e-01 -1.57784596e-01
2.95388818e-01 -7.18260050e-01 -6.06390774e-01 1.63770998e+00
3.09173524e-01 3.89770567e-01 5.08314848e-01 9.53443348e-01
4.63482827e-01 9.43004370e-01 -1.87346339e-01 -1.21295547e+00
9.09482121e-01 -2.05698937e-01 -1.26007295e+00 -6.16560951e-02
3.72700810e-01 -5.30591607e-01 8.83685946e-02 7.37769827e-02
-1.07622814e+00 -1.21883592e-02 -1.53237522e+00 8.41340363e-01
1.29817024e-01 4.04082760e-02 1.92767873e-01 9.44901288e-01
-5.42196214e-01 -3.31723660e-01 -2.34134391e-01 -3.10190744e-03
3.81812721e-01 2.97225267e-01 2.50422031e-01 -2.04695672e-01
-1.48643064e+00 6.53843164e-01 4.28749442e-01 4.26824033e-01
-6.71453357e-01 -6.55554891e-01 -7.07510948e-01 6.01168424e-02
9.03579950e-01 -2.70469904e-01 1.06180406e+00 -5.53846002e-01
-1.62763906e+00 -1.91576153e-01 -7.10198432e-02 -2.30174556e-01
4.19183940e-01 4.89109963e-01 -4.74558324e-01 1.73723847e-01
-1.57613099e-01 -2.87694894e-02 4.69193727e-01 -1.63017809e+00
-9.57433760e-01 -1.90008223e-01 -4.77199145e-02 2.98246384e-01
-1.56006217e-01 -1.12459689e-01 -9.15378034e-01 -2.10172921e-01
2.93553650e-01 -8.36040020e-01 -3.71466398e-01 -1.16957895e-01
-4.22733217e-01 -1.09735811e-02 7.83890367e-01 -6.35687709e-01
1.57288253e+00 -2.24933147e+00 9.65259522e-02 8.92543674e-01
-1.52394786e-01 1.77922368e-01 -4.04657051e-03 4.20761853e-01
6.06874168e-01 2.86263395e-02 4.84076850e-02 1.14451356e-01
-7.69003853e-02 2.87802666e-01 -2.00185087e-02 5.14977634e-01
-3.00391048e-01 4.22622472e-01 -7.77138054e-01 -3.40137482e-01
2.47642696e-01 3.50340426e-01 -2.12192342e-01 9.48285535e-02
2.69351006e-01 3.81319672e-01 -9.04806495e-01 6.92988396e-01
1.00718093e+00 2.99407959e-01 1.11177757e-01 -1.84524566e-01
-3.23052317e-01 -5.42870104e-01 -1.56543195e+00 7.85693944e-01
-6.28606439e-01 3.43318909e-01 7.40550637e-01 -1.05764306e+00
8.31888497e-01 5.26003540e-01 6.16619408e-01 -8.91868949e-01
5.09789944e-01 4.54920739e-01 -3.08412965e-02 -5.47336221e-01
6.22999296e-03 -4.94748466e-02 -9.53788385e-02 -2.16914743e-01
-3.23858082e-01 5.95894121e-02 -1.32871285e-01 2.73236811e-01
7.66362071e-01 -5.35879672e-01 3.92099917e-01 -9.78432894e-02
8.56755972e-01 -4.73652869e-01 9.14032102e-01 4.92404610e-01
-2.14333966e-01 -1.74996614e-01 5.80317616e-01 2.94754028e-01
-8.84578884e-01 -7.48155713e-01 -3.79669033e-02 2.19781399e-01
8.17673087e-01 2.84696370e-01 -4.83903557e-01 -7.41258040e-02
1.77990630e-01 7.93216348e-01 -2.56441623e-01 -3.49809021e-01
-6.03654645e-02 -2.03267515e-01 2.66545326e-01 1.69422477e-01
6.46743655e-01 -1.26142874e-01 -1.38834178e-01 3.39791298e-01
3.55075556e-03 -1.07189953e+00 -3.95802796e-01 -1.88301951e-01
-1.47785693e-01 -1.07305717e+00 -7.96024919e-01 -5.13484359e-01
4.95123953e-01 6.25549436e-01 -1.89726159e-01 1.33186698e-01
7.71232769e-02 5.76699257e-01 -1.51485249e-01 -6.14441752e-01
-1.23262450e-01 -3.75179738e-01 3.02407950e-01 2.36517996e-01
-1.73545972e-01 -2.31922954e-01 -5.42313457e-01 5.57565868e-01
-4.69797462e-01 -3.25154722e-01 8.23387980e-01 5.53426147e-01
6.31661415e-01 5.12093604e-01 1.17980134e+00 -7.48230086e-04
5.90112269e-01 -8.62362564e-01 -9.55633163e-01 2.17653736e-01
-5.04976034e-01 -2.60701418e-01 5.62063813e-01 -1.43300876e-01
-1.14185345e+00 6.15494326e-02 3.95598412e-02 -3.87167931e-01
5.00211835e-01 6.35144114e-01 -8.45561862e-01 -5.21515250e-01
1.48040533e-01 4.51756567e-01 2.00436652e-01 1.66657582e-01
1.71288490e-01 1.15984476e+00 3.72731686e-01 -3.77586663e-01
1.09109807e+00 2.06762895e-01 5.92427552e-01 -1.14598358e+00
-4.51243162e-01 -5.14429033e-01 2.39525829e-02 -6.90715730e-01
5.19484401e-01 -1.01001608e+00 -1.03795052e+00 2.82337904e-01
-9.40666437e-01 1.21690929e-01 4.71694887e-01 8.68673384e-01
-5.82120240e-01 4.24681246e-01 2.02952465e-03 -1.65111685e+00
-7.73063973e-02 -1.02989590e+00 4.13526922e-01 5.47622681e-01
6.60705805e-01 -6.02185190e-01 -6.16284847e-01 5.13735525e-02
5.23240089e-01 2.88738519e-01 6.74450755e-01 7.11552845e-03
-1.00602078e+00 -1.83256686e-01 -5.49757123e-01 1.00993954e-01
-1.66238651e-01 -4.50843245e-01 -2.79042602e-01 -4.40240026e-01
-2.05565870e-01 2.45115757e-01 2.93479949e-01 5.66736877e-01
7.91362286e-01 -5.31658769e-01 -6.62545860e-01 6.25429153e-01
1.68683290e+00 8.69675875e-01 7.90836930e-01 1.96802244e-01
2.58615792e-01 3.59130472e-01 1.00306928e+00 8.20916414e-01
5.38839519e-01 7.31786311e-01 8.39502454e-01 3.51391882e-01
3.03369284e-01 2.69835368e-02 -7.70462006e-02 4.67339545e-01
4.76358682e-02 -6.26251400e-01 -4.20957059e-01 4.33764368e-01
-1.90091586e+00 -8.20509195e-01 -2.43988886e-01 2.48113370e+00
4.11656499e-01 -1.50011126e-02 -1.54473454e-01 3.34007442e-01
8.78667712e-01 -2.63660848e-01 -6.76562130e-01 -3.67766581e-02
-7.41141215e-02 -6.48902893e-01 1.17041695e+00 6.18852139e-01
-6.32837653e-01 3.74448299e-01 4.85762978e+00 1.25628340e+00
-9.13993537e-01 -1.04937376e-02 2.07019985e-01 1.60830885e-01
-3.85507762e-01 -9.12329182e-02 -6.12652123e-01 6.67512178e-01
5.98172724e-01 -6.17789507e-01 5.39812267e-01 4.86687213e-01
7.72048533e-01 -3.14096689e-01 -2.16730505e-01 1.14853966e+00
-2.24740312e-01 -1.19029737e+00 -4.11797255e-01 3.32625538e-01
5.07814467e-01 -5.10907471e-01 -1.50250405e-01 3.26390237e-01
-3.27318221e-01 -5.28369188e-01 5.88684201e-01 9.37580705e-01
9.11640525e-01 -1.27683508e+00 8.36151958e-01 4.98237908e-01
-1.31239700e+00 -7.84978926e-01 -2.65185207e-01 1.56893894e-01
6.96007729e-01 5.75475752e-01 -4.42461520e-01 9.43449795e-01
1.03432052e-01 -1.58023551e-01 2.71798879e-01 1.63871598e+00
-1.34964019e-01 3.76465201e-01 -3.55507344e-01 -3.87017697e-01
3.24954718e-01 -2.20709547e-01 8.91256571e-01 6.75264060e-01
7.39403307e-01 7.50857055e-01 3.27317089e-01 2.03800201e-01
3.04272145e-01 4.32282448e-01 -3.12448621e-01 2.89963335e-01
1.41206026e+00 1.02794838e+00 -3.05529654e-01 -9.26925912e-02
-4.38205063e-01 2.93817699e-01 -3.63304317e-01 8.65432799e-01
-7.90493131e-01 -9.18514669e-01 6.44799948e-01 2.90522072e-02
2.72175491e-01 -4.31616873e-01 -1.81866258e-01 -3.26995701e-01
-1.38539178e-02 -2.97690928e-01 -3.02270409e-02 -4.64703679e-01
-6.42513812e-01 -2.37897430e-02 -7.09663406e-02 -1.43638074e+00
-7.19088912e-02 -2.17289001e-01 -5.72735190e-01 9.24270153e-01
-1.84123087e+00 -7.25424290e-01 -2.09235176e-01 5.47967494e-01
1.47607401e-01 -4.15317774e-01 2.39763275e-01 6.16870403e-01
-6.07324839e-01 7.40468562e-01 5.88158369e-01 -1.95032075e-01
3.06140780e-02 -4.04976755e-01 -9.98296916e-01 7.47676551e-01
-9.25411105e-01 2.37759754e-01 7.50253797e-01 -5.19700706e-01
-1.85942018e+00 -1.04156685e+00 3.82455975e-01 7.13889420e-01
5.16458392e-01 -6.17274176e-03 -5.30379653e-01 9.92044993e-03
-1.13748096e-01 -2.19633803e-02 5.20891726e-01 -6.43653452e-01
2.54170716e-01 -1.30346775e-01 -1.20579433e+00 5.63890278e-01
7.27852345e-01 -3.14635783e-02 2.66212344e-01 1.21109590e-01
6.73695207e-01 -5.66453815e-01 -7.18681395e-01 4.61352974e-01
6.67341709e-01 -2.45686039e-01 7.13862121e-01 -7.11535141e-02
-4.62346226e-01 -5.59385180e-01 -5.47710299e-01 -1.21220386e+00
-2.30152652e-01 -8.72077405e-01 -3.46258134e-01 1.27973914e+00
4.97454047e-01 -7.48552442e-01 5.01440883e-01 3.49103212e-01
-8.59040245e-02 -8.95139933e-01 -1.37889373e+00 -1.09239578e+00
-3.37851673e-01 -4.73598391e-01 4.69661564e-01 3.69596869e-01
1.39762878e-01 2.09134430e-01 -5.77209651e-01 7.71322310e-01
8.24576616e-01 -6.43867850e-01 7.05565333e-01 -9.14716780e-01
2.67372634e-02 -2.43916437e-01 -3.22357386e-01 -1.35180175e+00
1.64455742e-01 -4.82211053e-01 2.88890153e-01 -1.64039886e+00
-2.60192633e-01 -8.67372096e-01 -1.39232233e-01 -9.42658186e-02
-7.21309939e-03 -3.87982160e-01 3.33701372e-01 8.46294314e-02
-4.01087493e-01 8.82748544e-01 1.34185052e+00 -6.87447861e-02
-3.48500311e-01 7.26940393e-01 -7.26377785e-01 3.05225730e-01
7.35423207e-01 -1.98011659e-02 -8.29148531e-01 -1.04692660e-01
1.20483242e-01 9.91208196e-01 2.34738458e-02 -9.44660962e-01
3.08534145e-01 -4.97733772e-01 -1.16601065e-01 -6.86234653e-01
4.64146972e-01 -1.21366072e+00 1.71983182e-01 5.06476402e-01
2.72276253e-01 -9.70796347e-01 8.22164714e-02 1.03456390e+00
7.72111043e-02 -2.85965532e-01 5.75293303e-01 4.75743324e-01
-7.31454432e-01 5.45712173e-01 -6.73221886e-01 -2.71384537e-01
1.47041631e+00 -3.09254199e-01 -1.41561016e-01 -1.02447951e+00
-4.10884976e-01 9.95600820e-01 -3.86536658e-01 3.60987157e-01
5.79615057e-01 -1.44567728e+00 -5.34690559e-01 -5.50528392e-02
-9.10477806e-03 -4.38196748e-01 4.59104657e-01 9.79037702e-01
6.08430095e-02 9.44604993e-01 1.46473318e-01 -4.24341470e-01
-9.47293520e-01 3.56495559e-01 3.12194258e-01 3.27169716e-01
2.80914456e-01 4.69916314e-01 -2.75586516e-01 6.16104752e-02
3.80074829e-01 3.67492646e-01 -3.78553331e-01 5.52244671e-02
5.47795355e-01 6.56035125e-01 -3.13540488e-01 -8.39080095e-01
-3.05619061e-01 6.41591311e-01 4.86167103e-01 -1.84592411e-01
7.59961605e-01 -9.88194048e-01 2.88929015e-01 -3.13079625e-01
9.95488048e-01 2.56192863e-01 -1.30892026e+00 -4.86605912e-01
-3.63823503e-01 -5.65498292e-01 4.14138228e-01 -4.91614312e-01
-1.06641805e+00 4.79905486e-01 6.24690592e-01 3.70409846e-01
1.06285417e+00 -2.09622785e-01 7.56937623e-01 2.30983779e-01
7.98357368e-01 -1.16246760e+00 -2.14847803e-01 4.65502888e-01
7.35253215e-01 -8.39353859e-01 -1.44968659e-01 -7.53594458e-01
-3.48550528e-01 1.00900388e+00 6.37045562e-01 3.60062122e-01
6.96703255e-01 2.85573512e-01 -1.09656572e-01 2.85279810e-01
-5.12938380e-01 -5.77932596e-01 4.21473198e-02 6.83625102e-01
-2.12398931e-01 2.38481015e-01 -8.90226245e-01 9.91479576e-01
3.34538668e-01 -1.00322463e-01 5.99859595e-01 5.34753740e-01
-9.75670636e-01 -6.64407849e-01 -6.31947279e-01 3.98937285e-01
-1.64545983e-01 2.39987865e-01 3.99079293e-01 6.64619744e-01
1.13231689e-01 1.44042909e+00 -2.22506508e-01 -3.10312867e-01
5.32267511e-01 -4.38041121e-01 3.11950892e-01 -1.46142289e-01
8.79170477e-01 1.81230068e-01 2.79952198e-01 -2.57643312e-01
3.04223038e-02 -3.33847463e-01 -1.70812714e+00 -6.13101497e-02
-7.47891307e-01 4.90993828e-01 8.38108599e-01 1.31624830e+00
3.15913320e-01 6.25195444e-01 1.23986185e+00 -5.13523102e-01
-5.24740696e-01 -4.47898358e-01 -7.27489114e-01 -5.53376734e-01
3.35356861e-01 -8.74616921e-01 -4.65323597e-01 -6.97840333e-01] | [6.111091613769531, 1.4518405199050903] |
60281f9e-94d2-41d9-9fa7-54a177b31790 | deeptagger-knowledge-enhanced-named-entity | 2306.17413 | null | https://arxiv.org/abs/2306.17413v1 | https://arxiv.org/pdf/2306.17413v1.pdf | DeepTagger: Knowledge Enhanced Named Entity Recognition for Web-Based Ads Queries | Named entity recognition (NER) is a crucial task for online advertisement. State-of-the-art solutions leverage pre-trained language models for this task. However, three major challenges remain unresolved: web queries differ from natural language, on which pre-trained models are trained; web queries are short and lack contextual information; and labeled data for NER is scarce. We propose DeepTagger, a knowledge-enhanced NER model for web-based ads queries. The proposed knowledge enhancement framework leverages both model-free and model-based approaches. For model-free enhancement, we collect unlabeled web queries to augment domain knowledge; and we collect web search results to enrich the information of ads queries. We further leverage effective prompting methods to automatically generate labels using large language models such as ChatGPT. Additionally, we adopt a model-based knowledge enhancement method based on adversarial data augmentation. We employ a three-stage training framework to train DeepTagger models. Empirical results in various NER tasks demonstrate the effectiveness of the proposed framework. | ['Denis Charles', 'Jian Jiao', 'Qiang Lou', 'Xinyu Hu', 'Pengfei Tang', 'Simiao Zuo'] | 2023-06-30 | null | null | null | null | ['named-entity-recognition-ner', 'cg'] | ['natural-language-processing', 'natural-language-processing'] | [ 7.17171133e-02 2.12603793e-01 -3.71769458e-01 -7.06862152e-01
-1.29602766e+00 -8.76558304e-01 5.91878653e-01 -1.65525585e-01
-7.05903113e-01 6.09838247e-01 3.16363662e-01 -2.75906235e-01
2.42265612e-01 -8.86561930e-01 -7.15922058e-01 -3.77731174e-02
2.29935661e-01 5.36468267e-01 3.19286615e-01 -5.52528739e-01
-8.66198689e-02 2.05275893e-01 -8.36760759e-01 2.39496648e-01
1.06099987e+00 1.18520451e+00 -3.00636411e-01 5.56533098e-01
-6.03056073e-01 8.76553476e-01 -5.69923103e-01 -9.50560212e-01
1.78137645e-01 -9.17861909e-02 -8.82861793e-01 -1.05553471e-01
1.67212233e-01 -4.02999371e-01 -5.09285986e-01 7.86193788e-01
6.55857801e-01 2.37853095e-01 2.29980633e-01 -1.03878736e+00
-1.01360703e+00 7.10276306e-01 -2.62717545e-01 4.70799170e-02
1.45637661e-01 -8.28870758e-02 1.08051336e+00 -6.77216649e-01
5.32587767e-01 1.01401019e+00 6.33083940e-01 7.82600999e-01
-9.29204345e-01 -6.65300488e-01 2.07425371e-01 -1.22381084e-01
-1.21044338e+00 -4.42810267e-01 9.76758003e-01 -1.54031357e-02
4.16586101e-01 1.14961959e-01 1.24716215e-01 1.37946320e+00
-3.55697125e-01 9.37902451e-01 1.09409201e+00 -3.39458555e-01
2.62209684e-01 5.67738533e-01 1.33345380e-01 5.78045011e-01
-2.20465541e-01 1.28773093e-01 -2.71285892e-01 -4.95984882e-01
7.44985461e-01 8.45606476e-02 2.11808965e-01 -1.62826886e-03
-7.39753366e-01 9.53024447e-01 3.93722147e-01 2.82449931e-01
-4.52263415e-01 -1.12967059e-01 2.81938463e-01 9.69837531e-02
4.10629630e-01 3.49791855e-01 -7.72946298e-01 -3.89671065e-02
-7.31619954e-01 1.21285945e-01 1.16733265e+00 1.26758218e+00
8.16161096e-01 3.59169915e-02 -3.22084755e-01 1.08026397e+00
4.17505443e-01 7.18851626e-01 6.20665193e-01 -9.01084483e-01
5.70674479e-01 7.62404680e-01 3.56624216e-01 -7.07993209e-01
-8.21538493e-02 -5.67497194e-01 -4.63527471e-01 -5.02259731e-01
2.91395262e-02 -3.34936649e-01 -9.51341867e-01 1.83323896e+00
5.87751627e-01 2.30496183e-01 3.74135733e-01 7.47589529e-01
9.37142015e-01 5.71508527e-01 5.93672395e-01 1.61079764e-01
1.49191916e+00 -1.33579767e+00 -7.75467694e-01 -4.13205743e-01
7.40631342e-01 -6.71690166e-01 1.24940014e+00 -4.46296707e-02
-7.38555968e-01 -4.52646732e-01 -6.49724841e-01 -8.00005347e-02
-7.01378345e-01 9.91616100e-02 7.77162850e-01 8.25116098e-01
-6.59812987e-01 -1.80576125e-03 -7.55983651e-01 -3.22661668e-01
2.30602756e-01 4.72695887e-01 -1.71087801e-01 -5.93613386e-02
-1.52998197e+00 3.90869498e-01 3.31941009e-01 2.35311329e-01
-8.75824690e-01 -4.37572390e-01 -1.06973803e+00 3.12608220e-02
6.98330164e-01 -3.59113455e-01 1.73206103e+00 -9.76449311e-01
-1.50114715e+00 7.65304208e-01 -5.53122722e-02 -4.35966372e-01
3.10231656e-01 -2.62454212e-01 -1.08808947e+00 -3.76004241e-02
9.66371521e-02 5.14479399e-01 4.35879946e-01 -1.33973920e+00
-5.52278221e-01 -2.82554269e-01 3.31882387e-01 -1.13143153e-01
-6.30001307e-01 2.44226336e-01 -7.52010345e-01 -7.58839905e-01
-3.73349756e-01 -9.25222874e-01 -6.08234704e-01 -4.85081941e-01
-4.84059662e-01 -2.42969647e-01 8.04109931e-01 -8.57457161e-01
1.36058581e+00 -1.96910584e+00 -7.83243775e-01 2.12552667e-01
1.67626888e-01 5.73604345e-01 -5.33004880e-01 4.24031526e-01
3.54108810e-01 4.21050519e-01 -2.77257953e-02 -1.30867183e-01
4.83563334e-01 3.94618541e-01 -4.08316970e-01 -4.91216332e-01
4.71926779e-01 1.16623211e+00 -7.75118172e-01 -8.58449101e-01
-1.53625920e-01 5.10082722e-01 -5.98281860e-01 5.36732376e-01
-6.15147650e-01 4.89019305e-01 -1.25829434e+00 7.49426067e-01
7.34384775e-01 -3.41825336e-01 1.26320139e-01 -3.32937300e-01
4.74117026e-02 3.09430510e-01 -9.97563541e-01 1.64105427e+00
-9.77357805e-01 -1.72460079e-01 5.00077605e-02 -6.02243960e-01
9.14359212e-01 3.45241427e-01 1.39795423e-01 -8.88930500e-01
3.27049404e-01 3.34935963e-01 -4.71660435e-01 -6.44497097e-01
5.88880002e-01 3.59942019e-02 -5.33448875e-01 5.42262971e-01
1.48309082e-01 4.45977509e-01 1.22829592e-02 2.59546608e-01
1.38189614e+00 1.31176665e-01 -1.86033640e-02 4.17915881e-01
5.59261739e-01 1.14915334e-02 8.88152480e-01 8.20668399e-01
-3.40933561e-01 1.76243499e-01 -3.15879434e-02 -1.57168776e-01
-6.72049582e-01 -8.13031077e-01 2.14366585e-01 1.63620973e+00
5.04667833e-02 -2.06956342e-01 -8.99712801e-01 -1.37175798e+00
-4.31006849e-01 8.45648050e-01 -5.24866223e-01 -1.51619062e-01
-6.94045246e-01 -4.82008368e-01 9.28630769e-01 6.50272489e-01
9.54737246e-01 -1.22089446e+00 2.55601645e-01 4.78416324e-01
-4.13066328e-01 -1.61656868e+00 -8.00165176e-01 -6.39946237e-02
-7.15657651e-01 -7.88229406e-01 -6.45955503e-01 -1.14181018e+00
5.97556829e-01 -4.47972901e-02 1.15402746e+00 -2.70962805e-01
1.76912487e-01 6.97334111e-01 -5.75177550e-01 -2.41315722e-01
-4.57198441e-01 4.13268030e-01 -2.78379679e-01 2.53416032e-01
8.23831677e-01 -4.72526580e-01 -6.88217103e-01 6.78183258e-01
-1.20970178e+00 -5.09442031e-01 1.14223683e+00 7.79189825e-01
7.02119350e-01 -1.62987158e-01 8.64327610e-01 -1.24473155e+00
6.53000534e-01 -5.48747897e-01 -4.26312238e-01 6.86183095e-01
-4.62798566e-01 2.62977928e-01 6.81966543e-01 -6.88538194e-01
-1.51794374e+00 9.16247517e-02 -5.81464171e-01 -8.31217989e-02
-3.54327261e-01 5.50220668e-01 -6.64260209e-01 -2.85526007e-01
5.83082438e-01 3.03224415e-01 -5.96590459e-01 -8.13686073e-01
6.28382504e-01 9.45469856e-01 5.31876624e-01 -6.79643989e-01
8.59988391e-01 2.78013557e-01 -7.02526748e-01 -5.70652783e-01
-1.28991640e+00 -5.31654358e-01 -2.68285990e-01 2.19944820e-01
8.48354161e-01 -8.87562096e-01 -6.06160522e-01 1.20234273e-01
-1.07097328e+00 -2.07536116e-01 -3.08581799e-01 3.99917543e-01
-2.22313181e-01 2.39899442e-01 -6.88956201e-01 -7.93796301e-01
-6.43314183e-01 -6.96990371e-01 1.03724647e+00 6.26370668e-01
3.33993345e-01 -1.21594977e+00 2.76317179e-01 8.73079419e-01
5.62128723e-01 -3.05717383e-02 7.14173675e-01 -1.68402803e+00
-5.10100126e-01 -6.71926260e-01 -1.95060536e-01 4.43650395e-01
1.29076922e-02 -6.03951395e-01 -8.87527883e-01 1.74796879e-01
-5.27953207e-02 -6.60262942e-01 4.93177325e-01 -3.56023610e-01
1.05645597e+00 -8.09723616e-01 -3.48450840e-01 1.94272488e-01
1.15967357e+00 1.60638690e-01 4.55487072e-01 3.38733435e-01
6.17315471e-01 6.02489829e-01 8.54415178e-01 3.46532196e-01
8.16023469e-01 6.56346798e-01 9.12506804e-02 -2.57345676e-01
1.26108125e-01 -8.74145091e-01 1.72911108e-01 9.90249395e-01
3.29719216e-01 -4.20966566e-01 -7.17263401e-01 5.36227703e-01
-1.73637116e+00 -6.43308818e-01 2.09505364e-01 1.84488714e+00
1.04526639e+00 1.51447877e-01 -3.04439235e-02 -4.86853272e-01
8.79687309e-01 1.57653298e-02 -7.97118366e-01 -4.80170473e-02
7.28770271e-02 3.54222208e-01 4.96533394e-01 3.63140047e-01
-1.25001478e+00 1.37388098e+00 5.58786249e+00 9.32327271e-01
-8.88093233e-01 4.34039563e-01 3.68731499e-01 4.38316524e-01
-5.96060336e-01 1.59087181e-01 -1.08562434e+00 3.81851494e-01
1.12028635e+00 -2.00037118e-02 1.35719508e-01 1.25955260e+00
2.03424260e-01 6.53676450e-01 -7.93924987e-01 7.66385555e-01
8.67232159e-02 -1.01275325e+00 1.76656306e-01 1.44617662e-01
5.65757632e-01 -6.57089502e-02 -4.54289801e-02 1.12285209e+00
8.86195123e-01 -4.03494775e-01 1.45265117e-01 3.92113864e-01
5.04873931e-01 -5.26395023e-01 8.77946556e-01 3.11888486e-01
-1.12453234e+00 3.21606211e-02 -1.81271508e-01 5.72621167e-01
4.62110192e-01 3.71408343e-01 -8.85384619e-01 5.77202201e-01
4.35640544e-01 1.48894757e-01 -6.69533789e-01 7.62467802e-01
-3.27281177e-01 1.09047484e+00 -4.22100753e-01 -2.51655243e-02
3.17933142e-01 5.02206720e-02 1.59773186e-01 1.23832428e+00
3.26078869e-02 1.30463317e-01 5.21521986e-01 7.39343584e-01
-6.45790935e-01 3.93435597e-01 -3.75423878e-01 -6.28147185e-01
7.20779657e-01 1.68975115e+00 -2.58216321e-01 -2.32089281e-01
-7.90391445e-01 1.09697378e+00 1.93298802e-01 5.64484537e-01
-1.02778506e+00 -7.88456321e-01 1.03713676e-01 1.10372536e-01
4.19790387e-01 -6.03655390e-02 2.48301223e-01 -1.25412703e+00
1.07534111e-01 -8.85862112e-01 6.57234609e-01 -6.56545877e-01
-1.50058138e+00 8.42955232e-01 -5.64976394e-01 -1.07604241e+00
-3.44133705e-01 -3.74335855e-01 -6.06608868e-01 3.92540365e-01
-1.93437624e+00 -1.85914719e+00 -2.66385555e-01 6.94419742e-01
4.54029948e-01 -1.71894833e-01 8.05238247e-01 8.13930392e-01
-5.09218156e-01 1.03874123e+00 -1.41679898e-01 7.58246303e-01
8.20243001e-01 -9.83428955e-01 4.72448796e-01 5.76894701e-01
2.31791183e-01 7.40540504e-01 1.98824316e-01 -8.02728951e-01
-1.32651412e+00 -1.55230832e+00 1.16149986e+00 -4.11639810e-01
8.71991396e-01 -5.13558030e-01 -8.08752298e-01 9.86819565e-01
7.26614520e-02 1.74639538e-01 1.31704116e+00 1.72014847e-01
-5.72502613e-01 -2.38636106e-01 -1.21395695e+00 5.13844848e-01
9.97463763e-01 -4.97545660e-01 -6.58245742e-01 3.26175451e-01
1.33704484e+00 -1.96810439e-02 -1.03389490e+00 3.06183696e-01
3.80032003e-01 -3.73869389e-01 7.76956737e-01 -8.95416200e-01
-5.03714755e-02 -1.74961630e-02 -2.77755320e-01 -8.57954383e-01
4.25340123e-02 -8.18241298e-01 -5.40292002e-02 2.00056148e+00
7.00258851e-01 -7.73383558e-01 1.13528466e+00 1.04853928e+00
-1.03611551e-01 -3.10611367e-01 -5.95115006e-01 -8.68437171e-01
-2.70973742e-01 -4.03073132e-01 7.87553549e-01 1.10956252e+00
-3.64009142e-01 7.21195400e-01 -3.35186899e-01 2.01194003e-01
4.56574291e-01 -4.95534018e-02 7.77224839e-01 -1.07075214e+00
-3.17925841e-01 1.72181815e-01 3.22826509e-03 -1.31260478e+00
3.55030239e-01 -6.89207792e-01 2.85002086e-02 -1.42374289e+00
4.51967306e-02 -7.45686531e-01 -3.75304639e-01 8.37701738e-01
-1.17187649e-01 -4.42486703e-02 -1.16275989e-01 1.53264731e-01
-1.14168310e+00 7.92960525e-01 9.66893375e-01 -6.46851957e-02
-2.81078726e-01 1.93718776e-01 -9.58542407e-01 5.81357896e-01
7.97829211e-01 -6.76493645e-01 -3.28341722e-01 -4.94214922e-01
2.31724188e-01 5.92955463e-02 2.94882745e-01 -7.53609121e-01
2.64480799e-01 -1.89717099e-01 -1.28050307e-02 -2.94018775e-01
2.53368765e-01 -9.77426767e-01 -3.75474781e-01 -1.22569338e-03
-7.88587272e-01 -1.99586317e-01 1.06647342e-01 8.98603201e-01
-3.81201863e-01 -1.81809038e-01 3.39050889e-01 -1.60250351e-01
-7.22305417e-01 6.89196289e-01 -1.35855258e-01 6.17064834e-01
6.07670963e-01 2.49991655e-01 -2.72695541e-01 -6.11516356e-01
-6.37067556e-01 4.99364942e-01 -6.38910532e-02 6.02771103e-01
2.87636906e-01 -1.54919088e+00 -5.47807217e-01 -1.15650907e-01
4.20140594e-01 -3.16347629e-01 4.14983422e-01 4.37170088e-01
-7.50308558e-02 3.87179732e-01 3.21487218e-01 -5.83757646e-03
-7.37304986e-01 6.64883554e-01 7.21472278e-02 -7.34226227e-01
9.33880582e-02 6.76924884e-01 2.47897878e-01 -1.14391077e+00
2.81050622e-01 7.60133490e-02 -3.90625596e-01 -3.72147113e-01
5.40951550e-01 1.21428035e-01 -2.13983981e-03 -4.17627513e-01
-1.89455613e-01 2.25649476e-01 -4.98861283e-01 -3.50441664e-01
1.13177514e+00 -2.73100704e-01 3.22368860e-01 -1.03953578e-01
1.22657096e+00 2.98766702e-01 -9.30775762e-01 -7.38502324e-01
2.01862037e-01 -5.59356101e-02 2.67561041e-02 -1.26900065e+00
-9.41900432e-01 8.13578844e-01 6.64181232e-01 9.60419774e-02
1.13335609e+00 1.01186052e-01 1.64465415e+00 7.92946756e-01
2.51969546e-01 -1.31328130e+00 2.19022289e-01 4.15165961e-01
5.10370672e-01 -1.50975800e+00 -7.19641805e-01 -6.00164771e-01
-7.86744714e-01 7.03763843e-01 8.95073116e-01 3.03223550e-01
5.72395802e-01 8.92158747e-02 4.63758618e-01 -3.12092267e-02
-4.81123269e-01 -4.94865537e-01 1.35108694e-01 6.24594748e-01
2.42440701e-01 -2.77370483e-01 -3.45747709e-01 1.53927290e+00
-4.55376431e-02 2.38686800e-01 3.00754346e-02 1.30494773e+00
-2.84573972e-01 -1.51559174e+00 -7.79771106e-03 1.37221783e-01
-7.41057932e-01 -4.00655657e-01 -5.85754693e-01 6.21896744e-01
-2.57288367e-01 1.26377106e+00 -3.61650079e-01 -6.75082266e-01
5.81737518e-01 1.75594196e-01 -8.64123031e-02 -5.09782076e-01
-7.53288627e-01 3.91882025e-02 4.50519681e-01 -3.90279740e-01
-2.81685501e-01 -8.98531899e-02 -1.22379315e+00 2.25649700e-02
-5.53459823e-01 8.29596400e-01 8.12946498e-01 8.61472666e-01
8.40832770e-01 1.57665521e-01 9.02375460e-01 -1.58733144e-01
-7.85959244e-01 -8.28969002e-01 -3.66670102e-01 2.98556834e-01
-9.96548682e-02 -3.19410771e-01 -1.73346445e-01 1.63144633e-01] | [9.898639678955078, 9.589803695678711] |
12017bc3-4b01-436f-ad94-0fa8775e7f4b | hiface-high-fidelity-3d-face-reconstruction | 2303.11225 | null | https://arxiv.org/abs/2303.11225v1 | https://arxiv.org/pdf/2303.11225v1.pdf | HiFace: High-Fidelity 3D Face Reconstruction by Learning Static and Dynamic Details | 3D Morphable Models (3DMMs) demonstrate great potential for reconstructing faithful and animatable 3D facial surfaces from a single image. The facial surface is influenced by the coarse shape, as well as the static detail (e,g., person-specific appearance) and dynamic detail (e.g., expression-driven wrinkles). Previous work struggles to decouple the static and dynamic details through image-level supervision, leading to reconstructions that are not realistic. In this paper, we aim at high-fidelity 3D face reconstruction and propose HiFace to explicitly model the static and dynamic details. Specifically, the static detail is modeled as the linear combination of a displacement basis, while the dynamic detail is modeled as the linear interpolation of two displacement maps with polarized expressions. We exploit several loss functions to jointly learn the coarse shape and fine details with both synthetic and real-world datasets, which enable HiFace to reconstruct high-fidelity 3D shapes with animatable details. Extensive quantitative and qualitative experiments demonstrate that HiFace presents state-of-the-art reconstruction quality and faithfully recovers both the static and dynamic details. Our project page can be found at https://project-hiface.github.io | ['Jiang Bian', 'Chun Yuan', 'Sheng Zhao', 'Runnan Li', 'HsiangTao Wu', 'Tadas Baltrusaitis', 'Xu Tan', 'Tianyu He', 'Tianke Zhang', 'Zenghao Chai'] | 2023-03-20 | null | null | null | null | ['3d-face-reconstruction', 'face-reconstruction'] | ['computer-vision', 'computer-vision'] | [-1.86419263e-02 1.92904800e-01 1.45995291e-02 -5.11871576e-01
-7.92700231e-01 -3.66305113e-01 5.71044326e-01 -4.92856085e-01
2.76315808e-01 5.05495191e-01 2.19986811e-01 3.72218341e-01
3.57717514e-01 -8.20812523e-01 -8.26710820e-01 -7.61690974e-01
1.14888139e-01 5.50148547e-01 -8.18138421e-02 -2.76943773e-01
-3.28402519e-01 7.63944864e-01 -1.50288522e+00 2.65574545e-01
6.72622025e-01 1.06326449e+00 -2.48732850e-01 4.59446311e-01
-1.79820452e-02 3.86898488e-01 -2.87629887e-02 -5.26465297e-01
3.91970128e-01 -5.20172596e-01 -4.05855238e-01 3.74167681e-01
8.02344799e-01 -6.09917700e-01 -4.85678226e-01 9.60056424e-01
4.36221749e-01 -9.75820124e-02 6.40217602e-01 -1.11921728e+00
-8.44603539e-01 -3.88502121e-01 -7.09782779e-01 -4.78428334e-01
4.82603371e-01 2.78375953e-01 6.14002287e-01 -1.05710089e+00
9.56281364e-01 1.55187285e+00 7.65984476e-01 1.01028204e+00
-1.57786012e+00 -7.21165121e-01 6.94052279e-02 -2.14626297e-01
-1.29907465e+00 -7.85357893e-01 1.21688902e+00 -5.84357262e-01
3.42599243e-01 3.42907101e-01 7.89065361e-01 1.26521671e+00
2.70076543e-01 5.91552258e-01 1.31708479e+00 -1.89006031e-01
1.07050044e-02 -3.61579210e-02 -4.80246902e-01 1.09753513e+00
-8.42414871e-02 4.30239826e-01 -4.58132207e-01 -2.11337671e-01
1.36695063e+00 4.05856483e-02 -4.47757870e-01 -5.31838655e-01
-8.32695842e-01 5.20373702e-01 3.12507898e-01 1.08465347e-02
-3.93979132e-01 7.77105391e-02 -1.30367115e-01 1.41209671e-02
7.42838442e-01 2.32301112e-02 -1.31322920e-01 4.96125408e-02
-7.47798681e-01 4.16127443e-01 5.73556960e-01 8.84983182e-01
1.12470436e+00 2.36127123e-01 -2.72251926e-02 9.72202778e-01
5.19398689e-01 9.02839184e-01 -6.14409111e-02 -1.35078466e+00
-1.22520393e-02 4.79442447e-01 1.72681183e-01 -1.23826838e+00
-6.42126799e-02 3.59434751e-03 -9.68469858e-01 5.43400764e-01
3.09009492e-01 2.54673790e-02 -1.14187086e+00 2.04084873e+00
7.55210459e-01 3.22399706e-01 -3.88821781e-01 1.09721529e+00
1.04230237e+00 6.54521048e-01 -7.15125874e-02 -1.60512730e-01
1.15573776e+00 -7.47139513e-01 -8.61115456e-01 -2.16674015e-01
-1.54805496e-01 -6.31610811e-01 9.88803804e-01 -2.08196379e-02
-1.67114389e+00 -4.29893762e-01 -6.05465651e-01 -3.10585886e-01
2.64888376e-01 -1.95897222e-01 3.28733265e-01 9.87456515e-02
-1.16795814e+00 5.08908987e-01 -1.20110512e+00 -4.36201654e-02
6.17584348e-01 8.62775818e-02 -7.96553433e-01 -2.31408954e-01
-9.82751608e-01 4.66619521e-01 -5.64619720e-01 2.17864245e-01
-9.47492242e-01 -1.09127605e+00 -1.13198340e+00 -2.79094011e-01
-1.46248043e-01 -9.32105064e-01 9.90651250e-01 -9.81678486e-01
-1.66052938e+00 1.38666677e+00 -2.92144001e-01 3.67074847e-01
8.55909228e-01 2.43154824e-01 -1.28320754e-01 3.40048462e-01
-6.44776076e-02 6.42259836e-01 9.85845089e-01 -1.84556913e+00
7.00318292e-02 -6.93473458e-01 -2.42705550e-02 8.47728774e-02
-5.09305187e-02 -2.19048083e-01 -6.76492572e-01 -7.65148222e-01
1.73937351e-01 -9.22065079e-01 5.24464101e-02 1.01146460e+00
-2.71705002e-01 4.79434997e-01 7.98802495e-01 -1.05290914e+00
7.89585054e-01 -2.05801797e+00 6.11254156e-01 -2.56778616e-02
3.23944688e-01 -8.12000185e-02 -1.71756610e-01 1.35991260e-01
1.20964304e-01 9.29332525e-02 -5.78536451e-01 -8.75003636e-01
2.31420621e-02 3.30802053e-01 4.08577025e-02 5.63215792e-01
2.69332975e-01 1.18831086e+00 -8.32708120e-01 -4.73786712e-01
1.59812525e-01 1.22889364e+00 -5.52831113e-01 4.49643850e-01
-1.53227210e-01 1.02851796e+00 -5.45444965e-01 1.02674174e+00
1.00649142e+00 -1.88502923e-01 1.00290822e-03 -4.15651739e-01
1.61264330e-01 -1.36541232e-01 -9.09211636e-01 1.74558437e+00
-5.87146044e-01 4.08634275e-01 8.60607982e-01 -4.28788126e-01
8.57118487e-01 4.40221190e-01 6.21709049e-01 -8.41641128e-01
3.20006981e-02 2.79697150e-01 -5.63321173e-01 -4.08297181e-01
7.87351673e-05 -7.70428896e-01 2.23844513e-01 2.14117110e-01
-2.63163131e-02 -4.33656573e-01 -6.32448316e-01 -1.65088251e-01
6.34943962e-01 3.35181653e-01 -1.47804976e-01 -1.41212801e-02
2.87874818e-01 -4.56929088e-01 7.49569178e-01 -1.50459349e-01
-7.01403916e-02 1.06748164e+00 3.44607234e-01 -5.28269768e-01
-1.20151412e+00 -1.30207765e+00 -4.08795893e-01 2.73517877e-01
3.41828555e-01 -1.50475442e-01 -8.71840656e-01 -4.25181180e-01
3.45520347e-01 1.84397519e-01 -9.54690158e-01 -2.76013166e-01
-6.08688056e-01 -3.14598411e-01 2.75557578e-01 2.08703071e-01
5.16788483e-01 -9.93841171e-01 -2.24542528e-01 4.88696508e-02
-2.45707288e-01 -1.23341691e+00 -7.41460085e-01 -5.76691270e-01
-7.55265296e-01 -9.09977794e-01 -7.27960944e-01 -6.48158431e-01
1.00491416e+00 -1.57446433e-02 1.13944960e+00 2.45845810e-01
-4.36489999e-01 4.47482228e-01 9.30501074e-02 4.26918454e-02
-5.17235637e-01 -7.72661388e-01 -2.91198096e-03 4.68354583e-01
-3.82392615e-01 -1.06469035e+00 -7.80061781e-01 3.94186616e-01
-9.03840780e-01 4.28021371e-01 1.81409582e-01 7.80185819e-01
1.06799161e+00 -2.40641460e-01 7.30932653e-02 -7.89628446e-01
4.68175747e-02 -3.34975332e-01 -3.61916006e-01 -8.21254740e-04
-1.92645207e-01 -2.36192316e-01 3.48339617e-01 -5.86450815e-01
-1.16690266e+00 2.71313246e-02 -3.85372072e-01 -9.74155128e-01
-1.27397031e-01 -5.02807461e-02 -5.48971355e-01 -4.00921524e-01
4.52499181e-01 2.49660596e-01 3.83795828e-01 -6.95093751e-01
7.94608146e-02 3.38061452e-01 4.83797640e-01 -9.00898755e-01
1.01604033e+00 9.42477047e-01 1.97003484e-01 -7.28696465e-01
-7.19200373e-01 1.50587440e-01 -5.64407051e-01 -4.85355586e-01
6.53023839e-01 -9.82916713e-01 -5.36100268e-01 8.66274118e-01
-1.03118443e+00 -8.28903139e-01 -4.49839085e-01 -1.41984383e-02
-7.45494485e-01 1.94784343e-01 -8.80696535e-01 -5.21208704e-01
-3.33912700e-01 -1.07850051e+00 1.69749928e+00 6.79984167e-02
-8.11801851e-02 -1.07542312e+00 6.05324470e-02 5.16979158e-01
4.06882972e-01 1.22049093e+00 8.23164642e-01 7.05952406e-01
-7.19812095e-01 -1.01829186e-01 5.17382398e-02 3.81944776e-01
4.55553323e-01 1.79598019e-01 -9.77992058e-01 -3.44408631e-01
8.38092715e-03 -3.20447385e-01 4.55640793e-01 3.48242164e-01
9.34341133e-01 -5.93021035e-01 -6.19949698e-02 1.22034669e+00
1.34300399e+00 -1.42795995e-01 6.97614074e-01 -2.42112309e-01
1.04250538e+00 7.29940355e-01 2.96088248e-01 4.36834604e-01
5.21513939e-01 9.75377917e-01 5.04115701e-01 -2.10693806e-01
-7.09500790e-01 -7.93838739e-01 1.86658010e-01 6.37522638e-01
-3.22884053e-01 1.01610504e-01 -7.93107569e-01 4.25765693e-01
-1.57157362e+00 -8.36829305e-01 4.50703539e-02 2.02046871e+00
1.10134065e+00 -3.21828306e-01 -3.88488248e-02 -2.32052565e-01
4.81431037e-01 2.14402437e-01 -7.65143096e-01 -2.35999804e-02
-1.84011579e-01 2.19364062e-01 -1.68452069e-01 1.02601194e+00
-6.48291349e-01 8.38171482e-01 5.67841816e+00 6.56676769e-01
-1.37973821e+00 2.27954596e-01 7.33342409e-01 -2.02344522e-01
-9.35224950e-01 -1.95019305e-01 -5.37670791e-01 6.21435702e-01
4.98360723e-01 -4.29570302e-02 4.45753634e-01 4.98704284e-01
2.43433848e-01 4.08262700e-01 -8.26919854e-01 9.74061191e-01
-8.60697553e-02 -1.37753332e+00 4.52394515e-01 2.61772364e-01
1.02334797e+00 -4.72630918e-01 2.00643256e-01 -1.91011146e-01
4.70175818e-02 -1.09937823e+00 1.02394640e+00 9.00595069e-01
1.41598845e+00 -4.89297420e-01 1.63549870e-01 2.36477554e-01
-1.01396978e+00 3.63659203e-01 -3.94142084e-02 2.57555783e-01
3.53416890e-01 5.33241451e-01 -1.23535283e-02 3.30400974e-01
7.27264225e-01 1.00819087e+00 -5.05768433e-02 4.71414775e-01
-3.37908924e-01 3.57894689e-01 -2.12601349e-01 7.26904750e-01
-1.86084419e-01 -5.01324177e-01 7.75112391e-01 8.39122713e-01
3.04292619e-01 5.46652615e-01 -6.68042973e-02 1.16960406e+00
-2.08884060e-01 -4.51327413e-02 -5.39133132e-01 1.07777491e-01
2.60794193e-01 1.13292944e+00 -4.85023186e-02 6.79076314e-02
-1.93736866e-01 1.08110845e+00 5.01126349e-01 4.88151789e-01
-6.25221431e-01 2.47374266e-01 1.20254838e+00 8.87567639e-01
4.19034548e-02 -3.33390832e-01 -2.28887439e-01 -1.31928408e+00
2.48922616e-01 -6.96008682e-01 -9.12088305e-02 -8.79682839e-01
-1.29388690e+00 7.11502910e-01 -1.61527127e-01 -1.10531223e+00
-9.38929990e-03 -2.39065200e-01 -5.70650876e-01 9.27396894e-01
-1.73166287e+00 -1.54873741e+00 -6.44727290e-01 6.55223787e-01
1.54312342e-01 2.81858444e-01 8.33067000e-01 3.97153556e-01
-4.26372260e-01 7.42915988e-01 -2.96912909e-01 -2.10475288e-02
6.94352686e-01 -8.67185414e-01 3.49895716e-01 4.06392068e-01
-2.11991280e-01 3.31618309e-01 6.08750939e-01 -5.74001074e-01
-1.70284140e+00 -1.02835965e+00 6.41291440e-01 -5.63833296e-01
2.32064486e-01 -5.08867860e-01 -1.20324516e+00 6.41104996e-01
-3.83595765e-01 7.24156141e-01 3.57323796e-01 -5.13827562e-01
-5.18871248e-01 -1.47799745e-01 -1.66115654e+00 5.29669344e-01
1.35084999e+00 -6.64584041e-01 7.59062218e-03 7.64551908e-02
4.64789271e-01 -8.73322666e-01 -1.12882340e+00 5.04094362e-01
8.70712936e-01 -1.12732553e+00 1.00017977e+00 -4.08410966e-01
6.57448471e-01 -2.60088742e-01 -2.22439364e-01 -1.17150128e+00
-2.61931747e-01 -8.14734042e-01 -4.76704538e-01 1.17775238e+00
5.33294631e-03 -4.91740912e-01 8.25413227e-01 8.92195523e-01
-1.10676251e-01 -1.13610291e+00 -9.99976695e-01 -5.50002694e-01
1.53286383e-01 5.63245872e-03 7.88951814e-01 1.08210075e+00
-4.90330845e-01 -8.35901722e-02 -5.78016818e-01 1.63865134e-01
9.27569091e-01 3.85743320e-01 7.18778074e-01 -1.09523153e+00
-1.88103035e-01 -2.98772722e-01 -4.30165291e-01 -1.07679701e+00
4.58851784e-01 -8.48891914e-01 4.98518012e-02 -1.41834974e+00
2.17984691e-01 -6.22563243e-01 2.67479748e-01 5.89479208e-01
-1.06665544e-01 5.81288636e-01 9.21983330e-04 1.92037284e-01
5.85424639e-02 1.05627871e+00 2.10373616e+00 -5.89481257e-02
-6.89719394e-02 -2.15878636e-01 -6.56042755e-01 6.73224807e-01
4.86623228e-01 -2.52899468e-01 -3.35438341e-01 -6.66850388e-01
-7.89233968e-02 4.85236317e-01 6.88432038e-01 -5.55719733e-01
-9.36671793e-02 -4.32441562e-01 4.63410646e-01 5.49371876e-02
8.56316566e-01 -8.18764150e-01 7.57485926e-01 1.14090331e-01
-2.47192904e-02 -3.75468642e-01 3.28273356e-01 6.26074612e-01
-2.37960249e-01 6.09830737e-01 1.27873409e+00 -9.82750058e-02
-2.26819411e-01 1.12445068e+00 4.27642047e-01 1.74617454e-01
8.25982451e-01 -4.09795821e-01 4.93765064e-02 -5.64422667e-01
-8.11555505e-01 3.65680382e-02 1.26656747e+00 3.99201393e-01
8.30561340e-01 -1.75353360e+00 -9.35712874e-01 6.13620520e-01
2.51221117e-02 3.75995398e-01 6.69791520e-01 6.73424840e-01
-5.32058835e-01 -2.62164861e-01 -4.75521505e-01 -5.90444744e-01
-1.14062715e+00 1.23042025e-01 8.15227985e-01 1.31043687e-01
-8.72185349e-01 8.49326670e-01 5.20368040e-01 -6.70659900e-01
5.07697603e-03 2.61502247e-02 2.96701282e-01 -3.36945832e-01
4.50145692e-01 1.89097822e-02 -2.70940870e-01 -1.16047382e+00
-2.61755586e-01 1.34659719e+00 2.73719549e-01 3.53986472e-02
1.45549560e+00 -1.24374129e-01 -2.23933563e-01 2.70555854e-01
1.55378211e+00 2.03013644e-01 -1.95396757e+00 -2.51377136e-01
-8.27224195e-01 -9.54024673e-01 1.38667583e-01 -6.37711346e-01
-1.60056388e+00 9.06590998e-01 2.28257060e-01 -3.58677953e-01
1.15906048e+00 8.55723321e-02 1.15452015e+00 -4.75112855e-01
5.80023885e-01 -4.44349945e-01 1.71749532e-01 8.32218379e-02
1.33398283e+00 -1.17221165e+00 5.84080443e-03 -6.37264192e-01
-3.95300180e-01 8.89227629e-01 5.48761606e-01 -1.86922729e-01
9.12332177e-01 4.57539022e-01 3.40515524e-02 -3.55736554e-01
-6.59833074e-01 2.88722664e-01 5.19814432e-01 6.27975166e-01
1.71302497e-01 2.10444063e-01 1.54997811e-01 4.99124706e-01
-1.49831668e-01 -1.50987208e-01 1.86499238e-01 6.75708473e-01
1.14805870e-01 -1.02904499e+00 -1.74602374e-01 3.15480046e-02
-2.50960290e-01 3.51043820e-01 -4.43251371e-01 7.51270235e-01
1.98075488e-01 5.16183913e-01 1.19378023e-01 -2.80918628e-01
6.86635733e-01 -2.36635655e-01 8.85097504e-01 -5.37980616e-01
-1.25570402e-01 4.48417813e-02 -1.40075460e-01 -8.78725052e-01
-2.67541707e-01 -7.27842033e-01 -1.31822073e+00 -7.04190135e-01
4.22915161e-01 -2.52510041e-01 6.07828796e-01 3.63030940e-01
5.41143596e-01 1.22626990e-01 8.71438265e-01 -1.29915869e+00
-2.06619367e-01 -4.42382812e-01 -7.43521988e-01 7.70568967e-01
9.44839656e-01 -8.41186583e-01 -7.20456243e-01 1.85738534e-01] | [12.779732704162598, -0.3126075267791748] |
b6aa761d-b079-475a-92ba-914cfaf0510b | automatic-speech-summarisation-a-scoping | 2008.11897 | null | https://arxiv.org/abs/2008.11897v1 | https://arxiv.org/pdf/2008.11897v1.pdf | Automatic Speech Summarisation: A Scoping Review | Speech summarisation techniques take human speech as input and then output an abridged version as text or speech. Speech summarisation has applications in many domains from information technology to health care, for example improving speech archives or reducing clinical documentation burden. This scoping review maps the speech summarisation literature, with no restrictions on time frame, language summarised, research method, or paper type. We reviewed a total of 110 papers out of a set of 153 found through a literature search and extracted speech features used, methods, scope, and training corpora. Most studies employ one of four speech summarisation architectures: (1) Sentence extraction and compaction; (2) Feature extraction and classification or rank-based sentence selection; (3) Sentence compression and compression summarisation; and (4) Language modelling. We also discuss the strengths and weaknesses of these different methods and speech features. Overall, supervised methods (e.g. Hidden Markov support vector machines, Ranking support vector machines, Conditional random fields) performed better than unsupervised methods. As supervised methods require manually annotated training data which can be costly, there was more interest in unsupervised methods. Recent research into unsupervised methods focusses on extending language modelling, for example by combining Uni-gram modelling with deep neural networks. Protocol registration: The protocol for this scoping review is registered at https://osf.io. | ['Ying Wang', 'A. Baki Kocaballi', 'Liliana Laranjo', 'Juan C. Quiroz', 'Shlomo Berkovsky', 'Enrico Coiera', 'Dana Rezazadegan'] | 2020-08-27 | null | null | null | null | ['sentence-compression'] | ['natural-language-processing'] | [ 8.92804146e-01 6.29558682e-01 -6.68733954e-01 -2.10519791e-01
-1.18443775e+00 -3.44742209e-01 6.19649351e-01 8.62706363e-01
-6.00481510e-01 8.19011390e-01 1.14776576e+00 -6.01830125e-01
-2.15916187e-01 -3.44612867e-01 -2.00889364e-01 -3.82514834e-01
1.50832996e-01 3.69727790e-01 2.86170058e-02 7.12535158e-02
6.97687984e-01 1.57070696e-01 -1.32513011e+00 7.15423942e-01
8.42484057e-01 4.13091183e-01 3.63247484e-01 1.13129723e+00
-4.97495532e-01 1.10595059e+00 -1.22857571e+00 -2.24427357e-01
-6.02129340e-01 -9.40523386e-01 -1.16130507e+00 2.68684085e-02
-1.17896669e-01 -1.46206632e-01 -2.83106863e-01 6.68357253e-01
1.09458923e+00 1.58323020e-01 5.27696729e-01 -4.48666483e-01
-4.37961429e-01 7.91745603e-01 -3.65992673e-02 5.63409150e-01
9.38840210e-01 -6.22066259e-02 6.65668190e-01 -5.58135569e-01
5.85902929e-01 1.10950768e+00 5.46469510e-01 6.59569919e-01
-7.88385093e-01 -3.83284658e-01 -2.77165949e-01 -6.72846064e-02
-7.02248037e-01 -1.05355692e+00 3.60782951e-01 -4.21557695e-01
1.85322142e+00 6.82957709e-01 6.40279889e-01 1.09270442e+00
4.71007138e-01 6.78671241e-01 9.44273114e-01 -7.91974664e-01
3.65110576e-01 2.05360696e-01 2.63072044e-01 1.27397016e-01
3.64527732e-01 -1.70196623e-01 -5.08646786e-01 -4.71283972e-01
-1.43128624e-02 -3.02911967e-01 -2.60221541e-01 5.74027419e-01
-1.35145044e+00 9.06174302e-01 -2.95195460e-01 3.81563842e-01
-5.53735971e-01 -2.27480933e-01 1.13615680e+00 3.40950519e-01
6.33143663e-01 5.81040502e-01 -3.17881405e-01 -4.48916405e-01
-1.40029383e+00 -8.72932225e-02 1.14233923e+00 7.06514776e-01
-3.59595194e-02 1.30042359e-01 -3.50350112e-01 1.26209664e+00
3.85885775e-01 3.77894998e-01 1.04041553e+00 -8.21842074e-01
7.98869491e-01 3.33644837e-01 -5.05229473e-01 -6.60404205e-01
-7.14236796e-01 2.38267938e-03 -9.66588259e-01 -4.80136961e-01
-4.63733077e-01 -3.01810235e-01 -1.01907909e+00 1.15734255e+00
-1.84848353e-01 -7.05330849e-01 5.03769994e-01 4.25515294e-01
1.72145033e+00 8.38884532e-01 1.63897365e-01 -9.43412602e-01
1.43856370e+00 -8.63751233e-01 -1.22162795e+00 -2.95908451e-01
1.08910716e+00 -1.09105217e+00 3.05632591e-01 3.79294485e-01
-1.48917949e+00 -4.09987234e-02 -8.54850531e-01 -8.32721889e-02
-3.60056400e-01 2.90454784e-03 3.20574135e-01 1.01216948e+00
-1.27588618e+00 3.97347569e-01 -9.99921322e-01 -6.49301946e-01
4.34592009e-01 4.79250789e-01 -3.79144132e-01 3.73898447e-01
-1.22921252e+00 1.16059685e+00 5.61289251e-01 -1.61689565e-01
1.40205979e-01 -1.97332889e-01 -1.19445419e+00 -2.94405334e-02
9.92942080e-02 -8.90654147e-01 1.54085803e+00 -7.89217114e-01
-1.55717361e+00 8.62114668e-01 -3.97205472e-01 -7.55659521e-01
9.28030685e-02 8.81103575e-02 -2.80919105e-01 4.91451710e-01
2.22564638e-01 4.17070061e-01 3.40208888e-01 -5.22336960e-01
-5.97224653e-01 -2.25985572e-01 -5.78734815e-01 5.16198993e-01
-3.73190306e-02 8.02770793e-01 1.47632301e-01 -7.86923230e-01
-4.53343242e-02 -5.69598794e-01 -2.59080648e-01 -9.96798456e-01
-6.75652146e-01 -3.82557809e-01 3.29668760e-01 -1.15326476e+00
1.82288933e+00 -1.66898286e+00 3.62328961e-02 -2.03071728e-01
1.42533019e-01 5.90436876e-01 3.38375777e-01 1.06161046e+00
-1.78006992e-01 4.39988256e-01 -5.14927089e-01 -4.53583449e-01
-2.16784194e-01 7.04790130e-02 -2.66738962e-02 4.67761159e-01
4.92056832e-02 9.05339718e-01 -7.51699567e-01 -7.63565540e-01
5.47022820e-01 3.51392746e-01 -1.90698385e-01 -1.23553909e-01
4.25776303e-01 1.87349677e-01 -1.64300755e-01 3.45313877e-01
3.59505527e-02 4.54195589e-02 1.30781800e-01 2.80854911e-01
-2.32389495e-01 1.31703758e+00 -7.77216196e-01 1.42909145e+00
-4.93717670e-01 9.25363898e-01 -1.85581341e-01 -1.06391132e+00
8.35892022e-01 7.63368249e-01 3.56115550e-01 -4.83998507e-01
1.34044170e-01 3.98469418e-01 -2.92692669e-02 -7.79768586e-01
5.17907321e-01 -3.81097466e-01 -1.45515949e-01 3.93090278e-01
2.23027095e-01 -3.20509911e-01 2.24786758e-01 2.80717969e-01
1.16812992e+00 -2.46691972e-01 1.04062140e+00 6.13404736e-02
3.45923513e-01 -3.08182724e-02 3.22357297e-01 8.68262887e-01
2.76137628e-02 8.43591154e-01 5.89322805e-01 -2.31748298e-02
-9.15886283e-01 -4.08933371e-01 -2.92056024e-01 7.00240850e-01
-7.34748483e-01 -7.28182614e-01 -9.14543629e-01 -3.53447646e-01
-5.15100479e-01 1.00774205e+00 -3.46983403e-01 -3.96923810e-01
-6.17689192e-01 -7.33095169e-01 8.21120501e-01 4.87976581e-01
-1.70331250e-03 -1.61900985e+00 -9.05449629e-01 3.52338940e-01
-3.43746185e-01 -8.72932076e-01 -3.75289351e-01 4.16993767e-01
-9.95186269e-01 -6.88959658e-01 -9.99000192e-01 -8.52226973e-01
2.42504641e-01 -9.75516811e-02 8.20844293e-01 -9.12041143e-02
-3.42288315e-02 5.01991272e-01 -6.39203548e-01 -7.96863377e-01
-1.02029312e+00 4.27537620e-01 1.05306976e-01 -7.59741545e-01
4.53336418e-01 -1.25780180e-01 -4.24643070e-01 -5.16661167e-01
-9.53140855e-01 -8.06151032e-02 9.92843211e-01 9.22935843e-01
1.68676466e-01 -2.47832328e-01 8.98962438e-01 -8.87806654e-01
1.26117587e+00 -4.34504718e-01 3.62996668e-01 3.75366695e-02
-6.10086381e-01 -3.19476783e-01 9.52283964e-02 -1.12769343e-01
-8.17661166e-01 -2.65004545e-01 -5.18676221e-01 2.56012440e-01
-5.10092854e-01 9.51011002e-01 1.12606928e-01 5.84834397e-01
7.63668180e-01 3.84935051e-01 4.92946804e-01 -2.24238679e-01
-1.15890011e-01 1.46138060e+00 1.80303767e-01 1.32178858e-01
-1.81486905e-02 -3.35596241e-02 -3.62407863e-01 -1.54359436e+00
-3.08917135e-01 -7.82132983e-01 -4.46083665e-01 -4.32674512e-02
1.03033006e+00 -6.11765504e-01 -1.96124017e-01 2.02281386e-01
-1.33272004e+00 -1.74745247e-01 -4.14105028e-01 1.00835192e+00
-5.32797575e-01 7.24834919e-01 -5.79592645e-01 -9.67459977e-01
-9.76869226e-01 -1.20625520e+00 1.15154588e+00 8.93582478e-02
-1.09386039e+00 -9.12196696e-01 2.32893705e-01 5.14771163e-01
1.78314522e-01 1.22146048e-01 6.75363064e-01 -1.14597237e+00
4.82261419e-01 -5.11023343e-01 2.69080579e-01 4.46725577e-01
5.67313194e-01 -1.00890800e-01 -7.05043554e-01 -4.46376801e-02
3.21117908e-01 -1.66176111e-02 1.04622710e+00 1.01987290e+00
7.36196458e-01 -9.66849744e-01 -3.14096361e-01 4.15644161e-02
6.07753277e-01 6.84360147e-01 7.55883992e-01 5.41560531e-01
3.89299750e-01 1.08630502e+00 1.78526431e-01 1.46822289e-01
1.03160329e-01 4.05639589e-01 -3.32709908e-01 9.56971422e-02
-1.70727462e-01 1.98576987e-01 4.51373696e-01 1.30903804e+00
7.80682117e-02 -2.90717393e-01 -1.00272715e+00 6.85615301e-01
-1.51514006e+00 -1.12723446e+00 -1.95629165e-01 2.10017538e+00
7.93309033e-01 3.58533502e-01 4.22698408e-01 5.17992079e-01
7.61984348e-01 3.44100058e-01 -1.23085998e-01 -1.19141293e+00
-6.57314584e-02 3.26281488e-01 5.17864704e-01 3.60980779e-01
-8.19948614e-01 3.78154367e-01 5.93296432e+00 5.64995587e-01
-9.48358655e-01 5.56155108e-02 4.55290437e-01 -3.10089976e-01
-4.38722856e-02 -1.54499561e-01 -3.95806015e-01 5.39026260e-01
1.71854901e+00 -4.17556077e-01 -2.00175434e-01 3.61806124e-01
5.91458023e-01 -3.32926989e-01 -5.76625347e-01 7.48443127e-01
2.57889688e-01 -1.68212163e+00 1.00272521e-01 1.31378740e-01
3.52704793e-01 4.76790555e-02 -1.84074938e-01 2.18528524e-01
-2.10270807e-01 -1.05495739e+00 5.04395902e-01 3.05329502e-01
1.13253438e+00 -6.22074604e-01 1.18490338e+00 3.38524103e-01
-5.46300232e-01 8.53573829e-02 -4.71002012e-02 -1.49576411e-01
5.85429728e-01 4.63517785e-01 -1.01955235e+00 7.84568071e-01
5.77548087e-01 6.17735803e-01 -2.58577079e-01 1.05141461e+00
-3.53862718e-02 1.00023544e+00 -2.00296283e-01 -4.38728362e-01
2.80975461e-01 2.43682891e-01 8.59534860e-01 1.92217946e+00
-2.65626665e-02 3.41640711e-01 -3.37282449e-01 2.55819522e-02
1.63545072e-01 3.54594141e-01 -6.92204058e-01 -4.43452984e-01
3.16439807e-01 7.23360300e-01 -1.05075705e+00 -4.01978493e-01
-5.21517098e-01 6.97422385e-01 -4.43011075e-01 4.63590324e-02
-2.53252164e-02 -7.97379732e-01 -2.70145889e-02 2.63662666e-01
2.83608884e-02 2.45728746e-01 -5.05281508e-01 -6.79941237e-01
2.60334592e-02 -1.02578032e+00 5.30484915e-01 -5.61824143e-01
-7.02300787e-01 7.26695299e-01 4.09165859e-01 -1.03739309e+00
-7.30620444e-01 -3.31394784e-02 -4.80805963e-01 8.50033939e-01
-6.63835645e-01 -5.88459253e-01 3.06448430e-01 -2.28967443e-01
1.11635029e+00 -4.66059655e-01 1.05335855e+00 4.57742922e-02
-4.53988671e-01 3.22739869e-01 2.10751072e-01 6.45337775e-02
5.58953226e-01 -1.20916045e+00 5.02785504e-01 4.00913715e-01
-2.97663212e-01 8.29172075e-01 6.56916559e-01 -1.00367022e+00
-8.83515179e-01 -7.78048098e-01 1.60753381e+00 -3.10564429e-01
4.55070347e-01 4.85686865e-03 -8.54991674e-01 4.53982353e-01
5.76749444e-01 -9.42601264e-01 1.12785745e+00 -2.24314645e-01
4.94881332e-01 4.67987865e-01 -1.00155485e+00 4.10086006e-01
5.24020255e-01 -3.60193908e-01 -1.11219180e+00 5.16111970e-01
8.77547026e-01 -6.78215742e-01 -8.39512408e-01 2.55937725e-01
5.14570475e-01 -5.90814531e-01 6.75954401e-01 -6.83201790e-01
7.00373054e-01 2.94240475e-01 2.10020512e-01 -1.27785802e+00
-6.04390278e-02 -9.29379284e-01 -2.63110369e-01 1.18880892e+00
5.41108072e-01 -8.35076094e-01 4.71497208e-01 3.62366050e-01
-6.45627201e-01 -1.05811238e+00 -1.03542733e+00 -3.31041038e-01
1.38803303e-01 -3.78656864e-01 3.09533149e-01 6.88250124e-01
6.53140426e-01 5.73191106e-01 -1.66127551e-02 -4.77989733e-01
7.27718696e-02 -6.63905442e-01 2.83068597e-01 -8.59892070e-01
1.81234226e-01 -8.39698672e-01 -5.68759978e-01 -5.55123746e-01
1.28904596e-01 -9.96771157e-01 -3.43420133e-02 -2.66103220e+00
1.68682396e-01 3.09466571e-01 3.29080313e-01 3.75091732e-01
-1.33325577e-01 -4.45028126e-01 -1.40748113e-01 3.49083871e-01
-4.18281645e-01 2.86564976e-01 7.16306806e-01 -1.16375953e-01
-7.43741453e-01 4.02658552e-01 -1.06357980e+00 5.27645350e-01
1.12480307e+00 -6.86978042e-01 -3.98888469e-01 -3.02532632e-02
2.48062253e-01 4.50112939e-01 -5.15262596e-02 -5.11579812e-01
4.62412536e-01 4.50778082e-02 2.70256475e-02 -6.90473080e-01
-9.17624161e-02 -2.23775566e-01 -1.80274528e-02 5.31886518e-01
-8.10105205e-01 6.41671047e-02 3.71260315e-01 3.54676515e-01
-3.34638149e-01 -7.75065422e-01 3.52398813e-01 -3.29376280e-01
-3.07120960e-02 -4.44342405e-01 -1.23517644e+00 8.93887579e-02
5.10858953e-01 -6.36533618e-01 -3.43052208e-01 -6.85635209e-01
-8.70771945e-01 1.20043799e-01 2.24420149e-02 4.29760784e-01
7.60865688e-01 -8.74383807e-01 -9.33794320e-01 -2.64748484e-01
4.29089516e-02 3.50614429e-01 3.32529604e-01 1.25995600e+00
-5.90557277e-01 1.02051950e+00 3.29201549e-01 -3.68397802e-01
-1.57379973e+00 2.16184989e-01 1.40537947e-01 -1.74292907e-01
-9.31654155e-01 3.47695261e-01 -3.00805390e-01 -4.74392891e-01
3.64102304e-01 -4.08812702e-01 -6.31203771e-01 4.75984246e-01
7.93642104e-01 7.04839826e-01 5.93005598e-01 -9.48019087e-01
-5.26158333e-01 6.11291118e-02 -1.84578329e-01 -3.99324536e-01
1.30828631e+00 -2.33392179e-01 -2.99576133e-01 6.27099752e-01
1.33117580e+00 -2.03138247e-01 -1.18343689e-01 1.35469660e-01
2.81366557e-01 3.91281515e-01 2.06138164e-01 -6.91764653e-01
-3.24568629e-01 5.98056793e-01 9.46543887e-02 5.24756670e-01
1.14939415e+00 1.67342499e-01 6.91884816e-01 3.74963880e-01
-2.32788846e-01 -1.37073207e+00 -3.85002732e-01 4.28935021e-01
9.48712230e-01 -1.00137770e+00 2.71381974e-01 4.77968827e-02
-9.05544877e-01 1.20885205e+00 -3.11346412e-01 2.83459604e-01
4.45684910e-01 2.46536002e-01 -1.31248072e-01 -4.62006658e-01
-6.56798542e-01 9.27189663e-02 3.45154047e-01 5.88220656e-01
7.93577194e-01 2.48814952e-02 -1.01096225e+00 7.33008981e-01
-7.31672287e-01 4.89850864e-02 7.70348728e-01 1.21297479e+00
-4.16123897e-01 -7.95827091e-01 -5.07204175e-01 1.25071049e+00
-1.05570114e+00 -1.74985647e-01 -7.67983556e-01 4.85324740e-01
-5.10932565e-01 1.39906847e+00 -1.24165878e-01 -5.69102578e-02
5.10410666e-01 8.29592943e-02 2.45130599e-01 -1.16847241e+00
-9.57805276e-01 1.67684764e-01 6.78831935e-01 -1.01665959e-01
-8.07225406e-01 -1.15586281e+00 -1.16806328e+00 6.88338056e-02
-5.07670939e-01 4.45760906e-01 9.20339048e-01 1.13198054e+00
1.71383321e-01 8.01357627e-01 3.11360061e-01 -8.03868473e-01
-3.79476130e-01 -1.49231315e+00 -8.24715048e-02 -2.47695029e-01
6.36149108e-01 -1.26568481e-01 -6.16877377e-01 5.32442741e-02] | [12.356876373291016, 9.560173034667969] |
e6c017a7-94b5-4bf4-bad5-756264eb01c6 | how-to-reduce-change-detection-to-semantic | 2206.07557 | null | https://arxiv.org/abs/2206.07557v2 | https://arxiv.org/pdf/2206.07557v2.pdf | How to Reduce Change Detection to Semantic Segmentation | Change detection (CD) aims to identify changes that occur in an image pair taken different times. Prior methods devise specific networks from scratch to predict change masks in pixel-level, and struggle with general segmentation problems. In this paper, we propose a new paradigm that reduces CD to semantic segmentation which means tailoring an existing and powerful semantic segmentation network to solve CD. This new paradigm conveniently enjoys the mainstream semantic segmentation techniques to deal with general segmentation problems in CD. Hence we can concentrate on studying how to detect changes. We propose a novel and importance insight that different change types exist in CD and they should be learned separately. Based on it, we devise a module named MTF to extract the change information and fuse temporal features. MTF enjoys high interpretability and reveals the essential characteristic of CD. And most segmentation networks can be adapted to solve the CD problems with our MTF module. Finally, we propose C-3PO, a network to detect changes at pixel-level. C-3PO achieves state-of-the-art performance without bells and whistles. It is simple but effective and can be considered as a new baseline in this field. Our code is at https://github.com/DoctorKey/C-3PO. | ['Chengjie Wang', 'Bin-Bin Gao', 'Guo-Hua Wang'] | 2022-06-15 | null | null | null | null | ['scene-change-detection'] | ['computer-vision'] | [ 4.70378548e-01 -1.70518190e-01 -2.36540228e-01 -2.85600096e-01
-5.01908422e-01 -5.65486610e-01 4.76109535e-01 -8.52064788e-02
-3.52374434e-01 2.45681241e-01 -2.65380919e-01 -1.83440223e-01
7.38708004e-02 -7.73599088e-01 -4.68006581e-01 -8.10551763e-01
4.80119102e-02 7.07111433e-02 8.95793200e-01 -3.85036230e-01
1.96818158e-01 3.77868444e-01 -1.41721523e+00 1.82817370e-01
8.86195064e-01 9.64864314e-01 3.51843476e-01 6.29359126e-01
-1.37720346e-01 3.68194431e-01 -4.38331395e-01 -3.28349695e-02
3.77738029e-01 -7.00603604e-01 -1.10822904e+00 1.63805678e-01
3.63311917e-01 -9.33036432e-02 -2.86883414e-01 1.18124568e+00
5.87442458e-01 1.53287705e-02 3.88174295e-01 -1.35055435e+00
-3.13373506e-01 3.51852626e-01 -7.74291635e-01 6.91574156e-01
1.46696553e-01 2.36312136e-01 8.64781022e-01 -5.62423944e-01
7.21195996e-01 9.78086531e-01 7.89339244e-01 6.85100734e-01
-9.56309497e-01 -2.94725597e-01 5.61208546e-01 5.09709060e-01
-1.08932018e+00 -1.26698971e-01 8.74693036e-01 -2.90091723e-01
5.46046257e-01 5.92013359e-01 8.43444347e-01 1.11452997e+00
1.25555724e-01 1.23035324e+00 1.13683820e+00 -2.85176307e-01
2.26961136e-01 -3.06393802e-01 2.79250354e-01 5.90635717e-01
-1.48358315e-01 -1.38326570e-01 -2.89677888e-01 5.85780442e-01
7.86696255e-01 1.21358678e-01 -5.32362163e-01 -1.63015321e-01
-1.23871493e+00 5.15057623e-01 6.33050442e-01 6.48276210e-01
2.83789430e-02 3.06027025e-01 4.50667530e-01 2.66982138e-01
5.77969313e-01 2.61321247e-01 -6.67545497e-01 -3.18832733e-02
-1.01157022e+00 5.75340390e-02 5.01664817e-01 6.94554925e-01
7.74811149e-01 -4.16505575e-01 -3.20965201e-01 8.06122184e-01
-3.98480408e-02 1.96289405e-01 6.23369694e-01 -1.01044774e+00
7.48920068e-02 4.76864040e-01 -1.14940375e-01 -1.12949085e+00
-7.48344421e-01 -5.04360259e-01 -9.09713924e-01 -2.00059672e-04
3.05304199e-01 4.95297164e-02 -1.26216030e+00 1.60407412e+00
6.16383970e-01 5.68928421e-01 -4.06041235e-01 9.20589983e-01
8.16164553e-01 6.00215971e-01 -2.95045942e-01 -9.04883593e-02
1.38369179e+00 -1.36448717e+00 -7.31035948e-01 -3.22761774e-01
4.46158677e-01 -5.18146813e-01 1.19424748e+00 4.33173746e-01
-9.00043249e-01 -5.86653650e-01 -9.42199230e-01 6.40259497e-03
-5.33914030e-01 -1.26473993e-01 6.31103456e-01 3.89244109e-01
-1.40163875e+00 8.88900220e-01 -1.07701659e+00 -5.80748618e-01
4.65017676e-01 2.34409779e-01 -4.96939346e-02 -5.91820031e-02
-1.23151660e+00 6.48510456e-01 3.61046642e-01 4.06693161e-01
-7.61020243e-01 -5.49326301e-01 -6.94244206e-01 -2.90164232e-01
7.49297023e-01 -7.86833107e-01 1.37227142e+00 -1.28735983e+00
-1.36227703e+00 1.14440238e+00 -3.09764534e-01 -2.54451543e-01
7.24921942e-01 2.75452938e-02 -5.11659443e-01 4.45854545e-01
3.61559331e-01 8.11030686e-01 9.89842415e-01 -1.20037115e+00
-9.09699559e-01 -2.46779084e-01 4.90877815e-02 -8.67860671e-03
-7.43833259e-02 -5.71863763e-02 -1.19252145e+00 -8.75927091e-01
4.48856175e-01 -9.13708925e-01 -3.65370363e-01 8.26715976e-02
-5.36898077e-01 -2.89928734e-01 9.30132508e-01 -6.67644143e-01
1.44254279e+00 -2.10398173e+00 6.41663698e-03 2.19883360e-02
4.25516963e-01 4.05389726e-01 -2.29146212e-01 1.67227462e-01
-7.54342973e-02 2.24918023e-01 -7.83584535e-01 -3.18144739e-01
-5.04834317e-02 2.30571121e-01 6.47363737e-02 4.23170477e-01
2.75452942e-01 1.07568169e+00 -1.00916934e+00 -3.80220354e-01
1.26427621e-01 1.40547007e-01 -5.26986897e-01 -1.89290315e-01
-4.00062114e-01 6.58096850e-01 -3.97351205e-01 7.76086807e-01
9.13864791e-01 -3.59932661e-01 -8.30832869e-02 -1.40774585e-02
-1.19621821e-01 -3.67322229e-02 -1.12839973e+00 1.89715838e+00
5.38784191e-02 6.68682814e-01 -9.83041618e-03 -1.40436268e+00
7.04425991e-01 -3.09478901e-02 7.47587681e-01 -7.87214100e-01
1.68950960e-01 3.36496115e-01 -1.75603673e-01 -7.48812258e-01
2.83978105e-01 4.88235205e-02 -1.99358284e-01 7.84100965e-02
-1.93333805e-01 -2.16794595e-01 4.38094854e-01 2.59586759e-02
1.25175440e+00 2.71937519e-01 2.07491145e-01 -2.25481153e-01
6.04598761e-01 2.74308701e-03 9.64456260e-01 9.53541815e-01
-7.05179691e-01 7.98761964e-01 4.98717159e-01 -4.63291347e-01
-4.32279676e-01 -1.02993739e+00 -1.28319114e-01 8.74734223e-01
7.43137777e-01 -2.95438111e-01 -8.38628411e-01 -1.03669131e+00
-1.60442516e-01 2.32767180e-01 -7.21736133e-01 -7.53379464e-02
-7.35786796e-01 -9.00793433e-01 2.50085890e-01 4.29299057e-01
9.54773486e-01 -1.04743207e+00 -5.10426044e-01 2.01449081e-01
-4.90541875e-01 -1.14450192e+00 -7.37870634e-01 2.02896282e-01
-7.72819579e-01 -1.20044208e+00 -7.73530602e-01 -1.12199640e+00
5.34792244e-01 5.15871823e-01 9.50249434e-01 6.01212233e-02
-5.08821428e-01 3.41691554e-01 -5.31600952e-01 -2.68930346e-01
-5.22469133e-02 2.25801468e-01 -4.03975368e-01 1.24306209e-01
2.57715613e-01 -5.82913339e-01 -1.05410409e+00 4.42339122e-01
-1.11835015e+00 1.23949729e-01 3.45220000e-01 4.91661102e-01
7.27665246e-01 3.36499900e-01 4.24508989e-01 -8.80694926e-01
2.52984017e-01 -3.63221645e-01 -3.30445558e-01 1.79589659e-01
-6.78449333e-01 -2.08645284e-01 4.02643830e-01 -2.50827253e-01
-9.19301689e-01 -3.89406132e-03 -3.58710796e-01 -2.19228312e-01
-3.38867247e-01 3.37168217e-01 -1.79506272e-01 6.23465702e-02
3.17946255e-01 2.58260489e-01 -2.84488916e-01 -6.73206091e-01
3.73216301e-01 5.15505910e-01 8.70135427e-01 -2.83161163e-01
5.20726025e-01 9.19720054e-01 -2.31596231e-01 -4.93664980e-01
-1.06531179e+00 -8.81472409e-01 -8.18356335e-01 -2.87644416e-01
9.20095026e-01 -7.72953510e-01 -3.29519749e-01 1.05977213e+00
-9.27544355e-01 -7.51305819e-01 -3.53851616e-01 -8.14691857e-02
-5.38749397e-01 5.90105295e-01 -6.92287266e-01 -1.90261021e-01
-2.55466670e-01 -9.54628050e-01 1.19234002e+00 4.60789800e-01
1.20428046e-02 -1.19107282e+00 1.43682927e-01 2.43981108e-01
3.39206994e-01 5.15327632e-01 4.52789724e-01 -2.49649227e-01
-6.12179399e-01 1.59186527e-01 -2.65247166e-01 3.42791080e-01
3.94809157e-01 3.74248736e-02 -8.64877403e-01 -2.47438893e-01
1.76098555e-01 3.93876046e-01 1.37137961e+00 7.20503986e-01
1.52372050e+00 -1.37613162e-01 -6.25175536e-01 8.24856102e-01
1.55146408e+00 3.52456659e-01 6.83862567e-01 6.14095390e-01
7.61538208e-01 4.60764259e-01 5.27292073e-01 3.08856994e-01
5.86564064e-01 8.31073046e-01 3.27755034e-01 -4.20659631e-01
-4.08072412e-01 -4.48025251e-03 4.21983719e-01 8.83604527e-01
2.11691722e-01 -3.56011808e-01 -1.03619897e+00 7.49091625e-01
-2.01237512e+00 -7.80341625e-01 -4.96574789e-01 1.61932206e+00
1.03553998e+00 3.06657940e-01 1.23141408e-01 2.24348456e-01
9.64542925e-01 3.46967220e-01 -7.62286365e-01 -1.95248351e-01
-2.11078808e-01 3.53615761e-01 3.71562183e-01 2.92466551e-01
-1.36241782e+00 1.14581883e+00 6.03749609e+00 1.01257598e+00
-1.35673535e+00 3.72994781e-01 6.24150932e-01 3.16471867e-02
-2.26258308e-01 4.60937619e-02 -7.30692923e-01 7.43476748e-01
5.46539545e-01 1.65563971e-01 1.71524256e-01 3.11582297e-01
5.65741360e-01 -4.34049636e-01 -9.80869591e-01 1.03214002e+00
7.58127347e-02 -1.20419908e+00 -1.49292320e-01 -3.50883901e-01
9.24662709e-01 1.30625844e-01 -2.28490215e-02 1.80226997e-01
4.80228812e-02 -7.30255365e-01 6.22072041e-01 5.31380475e-01
5.26983619e-01 -3.08135986e-01 6.60272419e-01 1.87443063e-01
-1.27170348e+00 1.17950002e-02 3.77349928e-02 6.34844080e-02
1.93278983e-01 8.50233316e-01 -4.57688183e-01 7.45382488e-01
8.89778912e-01 1.29357183e+00 -7.75938749e-01 1.28684020e+00
-4.05355960e-01 5.96894145e-01 -2.60692894e-01 3.64747882e-01
4.99017298e-01 -1.91592976e-01 4.81840223e-01 1.26166773e+00
3.36429238e-01 -1.75662071e-01 2.01710850e-01 7.23202050e-01
3.81185934e-02 -1.31039977e-01 -1.80872634e-01 2.69161820e-01
1.26598597e-01 1.20836866e+00 -1.42478502e+00 -3.38236272e-01
-3.14009160e-01 1.61945176e+00 -1.11918941e-01 3.34212631e-01
-9.64531302e-01 -4.44755942e-01 5.71116805e-01 -5.29094879e-03
6.37198687e-01 4.62642387e-02 -3.09683323e-01 -1.27620745e+00
1.86467424e-01 -6.99119985e-01 5.53464353e-01 -6.31660938e-01
-1.13227189e+00 3.05890441e-01 -1.81127265e-01 -1.15942526e+00
2.31666982e-01 -4.57285374e-01 -8.85517240e-01 3.76282096e-01
-1.92091739e+00 -8.76580477e-01 -6.03785515e-01 8.25576723e-01
8.30367386e-01 5.03361166e-01 2.07571849e-01 4.42710131e-01
-1.11920989e+00 4.46032345e-01 7.61907473e-02 1.86910272e-01
7.17033803e-01 -1.46063519e+00 6.72727883e-01 1.37364089e+00
5.03230188e-03 5.86094111e-02 5.18625021e-01 -5.17370701e-01
-8.45547020e-01 -1.31565046e+00 7.60090649e-01 -3.60843122e-01
6.42700374e-01 -2.02511191e-01 -8.85173738e-01 4.88509864e-01
6.41934201e-02 2.20999062e-01 2.18449503e-01 -2.43327141e-01
-7.87860379e-02 -2.91265696e-01 -1.00529706e+00 5.25397122e-01
1.42883623e+00 -4.00459975e-01 -4.42088634e-01 2.82256752e-01
9.27344799e-01 -4.95456278e-01 -6.84018254e-01 4.53184694e-01
1.12229168e-01 -1.06412995e+00 8.31185222e-01 -1.42224550e-01
3.70194644e-01 -5.41547954e-01 2.65043646e-01 -1.29940712e+00
-2.28948250e-01 -8.92666757e-01 3.50228488e-03 1.16766751e+00
9.36199501e-02 -7.78944373e-01 6.52325690e-01 1.13587901e-01
-4.05593216e-01 -6.78706765e-01 -8.77491117e-01 -9.92536783e-01
5.73793389e-02 -5.69000602e-01 4.44852114e-01 1.14010072e+00
-2.76728541e-01 1.11248955e-01 -9.13556442e-02 1.74805507e-01
3.50455225e-01 2.13956341e-01 4.54609066e-01 -1.16286004e+00
-1.75697923e-01 -8.25558603e-01 -3.13434631e-01 -1.34463406e+00
-1.00463271e-01 -1.04180038e+00 2.42462575e-01 -1.71422195e+00
1.58748344e-01 -5.40333390e-01 -4.45537239e-01 7.10237145e-01
-4.77586210e-01 3.74439806e-01 2.43520483e-01 3.29830915e-01
-6.96339190e-01 4.18592155e-01 1.70220971e+00 -2.48817801e-01
-4.49766397e-01 1.09213233e-01 -8.16805661e-01 7.09947109e-01
1.06673968e+00 -3.96600306e-01 -3.88729990e-01 -3.35317045e-01
6.51012734e-02 -2.63719976e-01 5.64319730e-01 -1.01277959e+00
2.49516398e-01 -1.01596937e-01 1.28910327e-02 -6.99198663e-01
-1.62986428e-01 -5.48781395e-01 1.03567377e-01 5.80845594e-01
-3.47530693e-02 -1.87707633e-01 1.81074500e-01 5.42727649e-01
-5.39749563e-01 -2.46925473e-01 9.10010219e-01 -3.61294508e-01
-1.45842588e+00 5.86383164e-01 -3.06660920e-01 2.77798653e-01
1.17733312e+00 -2.97336757e-01 -3.36586684e-01 -1.38283193e-01
-1.00443125e+00 5.11299849e-01 4.42606598e-01 5.38752019e-01
4.37605143e-01 -1.17916954e+00 -4.45692450e-01 1.22784130e-01
1.10376656e-01 2.54299223e-01 4.93486255e-01 1.26800418e+00
-5.84620655e-01 1.96116254e-01 1.81327146e-02 -9.33975637e-01
-1.09694541e+00 4.95181084e-01 5.78794777e-01 -2.80346960e-01
-9.23718750e-01 1.03179538e+00 3.00082475e-01 -3.26828480e-01
4.38029096e-02 -7.26556063e-01 -9.81515422e-02 2.94287443e-01
3.48568320e-01 2.22089946e-01 1.17023863e-01 -2.56508023e-01
-4.93423641e-01 8.00510824e-01 7.12840306e-03 2.45882586e-01
1.32876229e+00 -5.30883431e-01 -2.28194669e-01 4.67032582e-01
1.36382365e+00 -3.88044566e-01 -1.58679259e+00 -3.42209667e-01
1.60740912e-01 -2.19059512e-01 8.95663723e-02 -7.60099590e-01
-1.38010681e+00 7.08708525e-01 8.10608387e-01 4.33601946e-01
1.66474926e+00 2.21253172e-01 1.22971129e+00 -1.24405302e-01
1.95435524e-01 -1.38385320e+00 2.16359228e-01 5.72210789e-01
6.15728199e-01 -1.36605310e+00 -3.08039129e-01 -6.45975471e-01
-3.44335943e-01 1.03146923e+00 5.78065932e-01 -7.85622895e-02
7.51552880e-01 1.01765096e-01 6.40959665e-02 -3.60109597e-01
-4.28746998e-01 -7.08058894e-01 3.24164271e-01 6.00255013e-01
-9.34400707e-02 3.68257202e-02 -4.40150529e-01 3.98406237e-01
9.88984779e-02 2.02749167e-02 4.83396918e-01 9.32649314e-01
-5.09992957e-01 -1.15901399e+00 -1.76357463e-01 2.40556255e-01
-3.12243432e-01 6.39322819e-03 -3.38140786e-01 8.15222144e-01
6.61965072e-01 8.18187594e-01 6.38064966e-02 -3.33364964e-01
2.94198334e-01 -1.95785701e-01 4.08991188e-01 -4.37809676e-01
-5.44094861e-01 2.06779540e-01 -2.35307544e-01 -1.02513587e+00
-8.71024311e-01 -8.57689202e-01 -1.37937176e+00 -3.42369080e-02
5.34053929e-02 -2.73357272e-01 3.65471274e-01 8.60856473e-01
2.73683518e-01 6.81774497e-01 8.49139810e-01 -6.45190597e-01
-5.90740368e-02 -6.51526451e-01 -6.54710174e-01 4.69358861e-01
5.82902908e-01 -4.87176538e-01 -4.40883517e-01 2.18394607e-01] | [9.624187469482422, -0.6590315699577332] |
7ec5607c-8755-4b91-aec2-a3e82e345faa | sensing-of-side-lobes-interference-for | 2306.17650 | null | https://arxiv.org/abs/2306.17650v1 | https://arxiv.org/pdf/2306.17650v1.pdf | Sensing of Side Lobes Interference for Blockage Prediction in Dense mmWave Networks | The integration of sensing capability in the design of wireless communication systems is foreseen as a key enabler for efficient radio resource management in next-generation networks. This paper focuses on millimeter-wave communications, which are subject to severe attenuation due to blockages, ultimately detrimental to system performance. In this context, the sensing functionality can allow measuring or even imaging the wireless environment allowing anticipation of possible link failures, thus enabling proactive resource reallocation such as handover. This work proposes a novel mechanism for opportunistic environment sensing, which leverages existing network infrastructure with low complexity. More specifically, our approach exploits the fluctuations of interference, perceived in antenna side lobes, to detect local activity due to a moving blocker around the reference communication link. Numerical evaluations show that the proposed method is promising as it allows effective assessment of the blocker direction, trajectory and possibly, its location, speed, and size. | ['Benoit Denis', 'Hiba Dakdouk', 'Mohamed Sana'] | 2023-06-30 | null | null | null | null | ['management'] | ['miscellaneous'] | [ 4.59424525e-01 3.12999159e-01 -3.03460956e-01 2.13803947e-01
-3.17948371e-01 -7.41215646e-01 1.79898307e-01 8.52131918e-02
-2.59987980e-01 1.09626245e+00 -1.22491628e-01 -8.12394142e-01
-6.36512101e-01 -7.82792449e-01 -6.83039278e-02 -1.18314397e+00
-5.27224958e-01 -9.17346478e-02 -1.68226734e-02 1.35234103e-01
1.67958677e-01 8.11743021e-01 -8.99603903e-01 -6.55706942e-01
6.47670805e-01 1.28922462e+00 4.58134204e-01 5.56480169e-01
1.97212607e-01 9.93258283e-02 -8.86745870e-01 1.32952198e-01
-2.77239457e-02 7.67050460e-02 -2.61472650e-02 -1.11254700e-01
-3.24266344e-01 -2.88039774e-01 -9.13938358e-02 6.63147807e-01
7.62821555e-01 -4.19197559e-01 3.11669469e-01 -9.05776620e-01
4.58733588e-01 5.81819892e-01 -2.78551847e-01 3.69980991e-01
3.40624809e-01 -9.92239416e-02 5.72443843e-01 -3.06625158e-01
3.39658618e-01 4.77280974e-01 4.80602711e-01 -2.32150573e-02
-8.76891553e-01 -9.07562077e-01 -1.79557458e-01 -1.07437782e-01
-1.25367558e+00 -6.14385843e-01 8.05383265e-01 -2.68214881e-01
2.79783487e-01 6.03775978e-01 7.89415300e-01 9.12371695e-01
5.00780165e-01 -3.75982933e-02 6.41349673e-01 -6.24663591e-01
5.17797649e-01 1.10690862e-01 -2.78221488e-01 4.04587924e-01
8.98680091e-01 3.68554145e-01 -2.97917843e-01 -2.92793155e-01
6.66179597e-01 -4.28496152e-01 -6.52915001e-01 -5.34453809e-01
-1.06796551e+00 4.81906272e-02 5.21769464e-01 6.47737503e-01
-5.95790684e-01 3.14670712e-01 -4.51845944e-01 2.41891161e-01
1.25059530e-01 6.07732177e-01 -1.19772762e-01 -3.40941787e-01
-8.61601055e-01 -3.31025422e-01 5.36082089e-01 1.02971828e+00
3.76642525e-01 1.93450432e-02 -8.50442722e-02 2.73003548e-01
6.91325724e-01 9.32841241e-01 9.17640626e-02 -6.59317851e-01
5.74304044e-01 8.53630155e-02 4.54038650e-01 -8.82284880e-01
-1.16031182e+00 -1.73491967e+00 -9.41835403e-01 -1.52888685e-01
1.20588802e-01 -6.25743926e-01 -6.47748411e-01 1.74199498e+00
2.94267893e-01 3.34183097e-01 2.90255278e-01 3.54141265e-01
3.13794136e-01 3.32667559e-01 -1.27963290e-01 -6.17203295e-01
1.01783133e+00 -1.41241774e-01 -6.12236798e-01 -4.61758703e-01
3.00176889e-01 -7.47684300e-01 -1.32886588e-03 5.57592213e-01
-9.50544536e-01 -1.79523155e-01 -1.48228383e+00 1.13350821e+00
5.14807366e-03 9.32728723e-02 5.75980067e-01 1.26985204e+00
-8.82591426e-01 -1.89727977e-01 -6.76547468e-01 -3.80968690e-01
3.25405687e-01 4.33415681e-01 3.50225806e-01 -7.77176768e-02
-8.82463634e-01 1.53495699e-01 8.09879005e-02 3.48283798e-01
-2.87064970e-01 -8.81361723e-01 -3.68617803e-01 2.23365843e-01
1.52435422e-01 -5.98110974e-01 1.05178654e+00 -1.46833763e-01
-1.32974958e+00 -1.19811200e-01 -1.42654613e-01 -5.48822820e-01
3.79348308e-01 1.89158022e-01 -7.47318864e-01 1.29707590e-01
-1.37433052e-01 -2.59161890e-01 5.92430353e-01 -1.29224658e+00
-8.09134901e-01 -2.96239644e-01 3.35526327e-03 -4.21594083e-03
-1.77989155e-01 -5.66525936e-01 -4.66703214e-02 -4.69932467e-01
6.41537905e-01 -1.01994491e+00 -5.28779209e-01 -6.46702349e-01
-7.85547197e-01 5.74403763e-01 6.26658022e-01 3.34172486e-03
1.27251005e+00 -2.14326119e+00 -2.34784648e-01 9.57044780e-01
2.30735224e-02 1.85131371e-01 2.33554065e-01 7.60770261e-01
3.65750700e-01 1.23165935e-01 -3.26295495e-02 1.38246968e-01
-3.68243486e-01 -3.24452788e-01 -1.57980308e-01 5.60563028e-01
-4.79973286e-01 3.10658842e-01 -7.19906449e-01 3.12257320e-01
2.06536338e-01 2.56111026e-01 -1.42252266e-01 -3.41855325e-02
3.52328539e-01 9.05822933e-01 -1.01331365e+00 7.94897437e-01
7.95249879e-01 1.34865403e-01 4.90938455e-01 -1.20627210e-01
-5.87678373e-01 1.45635262e-01 -1.02812111e+00 1.12865758e+00
-9.54379022e-01 5.33538103e-01 5.14075816e-01 -9.04908240e-01
9.73279059e-01 3.78405660e-01 7.59587348e-01 -8.71685266e-01
1.01142950e-01 4.61585045e-01 2.38731951e-01 -3.98533791e-01
5.12345657e-02 8.78712758e-02 -1.21675283e-01 1.00106813e-01
-5.35081029e-01 1.17079444e-01 -1.43006608e-01 5.45689948e-02
1.61914718e+00 -3.13968152e-01 5.51686347e-01 -3.34479570e-01
5.98431408e-01 -3.46212357e-01 4.66989130e-01 8.87228191e-01
-1.00836948e-01 -2.47104824e-01 -8.21899921e-02 1.68001190e-01
-2.81415045e-01 -1.29148948e+00 -2.98659354e-01 2.02264264e-01
7.18216598e-01 1.38492048e-01 -2.98417062e-01 8.12376477e-03
6.33098260e-02 4.53070939e-01 1.33690000e-01 -2.34442845e-01
-4.37589824e-01 -9.06060815e-01 1.89103171e-01 2.29099616e-02
4.91345197e-01 -4.24115837e-01 -9.06835914e-01 7.91108191e-01
1.00595497e-01 -1.33867407e+00 3.52530509e-01 3.10951799e-01
-7.71697819e-01 -9.00270045e-01 -3.17150831e-01 -4.42153156e-01
5.54467976e-01 5.79951346e-01 6.86055183e-01 2.20705166e-01
-2.37209976e-01 7.97816336e-01 -1.38649002e-01 -3.81740302e-01
-8.28757882e-02 2.38979682e-01 -6.00651605e-03 2.35278606e-02
-3.30113113e-01 -1.02451873e+00 -1.07198691e+00 6.79464579e-01
-2.85218030e-01 -2.67226011e-01 8.66219282e-01 7.91200921e-02
1.33399770e-01 6.31641209e-01 9.98848975e-01 -4.30152893e-01
4.29455936e-01 -7.19645917e-01 -9.41002071e-01 6.13638796e-02
-3.54451209e-01 -3.33926231e-01 5.83259583e-01 9.60966870e-02
-1.09182596e+00 -1.34228796e-01 -1.34896830e-01 8.81775081e-01
-1.82357971e-02 7.09349990e-01 -6.27574444e-01 -6.34399474e-01
1.94692284e-01 -8.36657360e-02 -5.69320619e-01 -1.65904701e-01
-1.07773952e-02 6.67894840e-01 3.15597683e-01 -1.70667529e-01
1.40065718e+00 8.19066286e-01 4.42428231e-01 -1.45111704e+00
-2.94641584e-01 -7.56140113e-01 -3.67856473e-02 -4.55122024e-01
2.02805221e-01 -9.58097279e-01 -6.53039396e-01 1.14171527e-01
-8.19105148e-01 -1.70243546e-01 5.55763245e-01 8.72003794e-01
-7.94982687e-02 9.76811629e-03 5.18385135e-02 -1.24889159e+00
-1.64603576e-01 -8.72663498e-01 4.84844863e-01 3.80113006e-01
-2.17283919e-01 -9.92624283e-01 -2.35174835e-01 2.16894910e-01
9.27493989e-01 3.24263453e-01 6.44258201e-01 -5.02471365e-02
-1.06269312e+00 -3.17813694e-01 -3.16924304e-02 -5.30157268e-01
3.14357638e-01 -4.66407210e-01 -7.74244547e-01 -4.80572194e-01
-1.28725305e-01 6.89587653e-01 3.66221994e-01 9.69059408e-01
8.84900808e-01 -5.30241504e-02 -1.04337871e+00 6.96537077e-01
1.62219608e+00 5.08712888e-01 7.73671150e-01 2.39992142e-01
6.86587021e-02 1.91039816e-01 8.76931965e-01 7.77875841e-01
1.02843801e-02 7.66449332e-01 1.03255498e+00 7.29913265e-03
9.76750925e-02 2.62793183e-01 -8.36477987e-03 4.63361830e-01
8.26293156e-02 -9.08341169e-01 -6.18649244e-01 3.13207150e-01
-1.34780550e+00 -6.68513834e-01 -1.28008381e-01 2.45647335e+00
-1.58143878e-01 5.08039296e-01 -2.80036241e-01 6.89938292e-02
3.85426313e-01 2.40213890e-03 -4.37529594e-01 4.82844748e-02
9.57212448e-02 1.42243043e-01 1.24227035e+00 5.96222401e-01
-6.46453857e-01 1.57406822e-01 5.40199614e+00 5.10913551e-01
-1.22512317e+00 -1.09021328e-01 1.46041259e-01 8.49616006e-02
-5.46861172e-01 -2.14817002e-01 -6.62054300e-01 4.30657744e-01
8.78818333e-01 -6.02425188e-02 -2.13729460e-02 3.28324676e-01
8.72945607e-01 -5.52777350e-01 -4.77440923e-01 8.08351696e-01
-5.09723067e-01 -1.37263250e+00 -5.88566601e-01 6.08946085e-01
4.36822802e-01 -2.78459907e-01 2.78568566e-01 -1.92115650e-01
-1.65215537e-01 -3.82448703e-01 4.67029750e-01 6.69525564e-01
2.62810946e-01 -9.08642769e-01 7.92547405e-01 3.12978745e-01
-1.29915893e+00 -5.21202385e-01 2.47936592e-01 -6.48999587e-02
4.92627144e-01 1.26436436e+00 -1.13359559e+00 7.34871864e-01
1.83016255e-01 1.53477103e-01 -7.43398592e-02 1.59778488e+00
-2.57123768e-01 8.54444444e-01 -5.22948623e-01 -1.89999059e-01
5.59800118e-02 -7.73433968e-02 9.17148054e-01 8.23485196e-01
1.06834841e+00 -5.51539324e-02 1.00001812e-01 2.51371771e-01
-7.69145265e-02 -8.00578892e-02 -4.80351895e-01 5.04160821e-01
1.19772673e+00 1.12765360e+00 -7.28785694e-01 2.89099991e-01
-3.04995716e-01 3.65531236e-01 -4.10340756e-01 8.08974326e-01
-6.49353743e-01 -4.85556781e-01 8.16660643e-01 4.00785536e-01
2.94445872e-01 -6.73418343e-01 -5.32487452e-01 -2.47391701e-01
2.21151009e-01 -1.25699922e-01 -1.11859865e-01 -1.74109831e-01
-3.74973804e-01 1.97828382e-01 -4.06121463e-01 -1.34418416e+00
-2.21522003e-01 -2.47109160e-01 -6.51613176e-01 5.57189882e-01
-1.81310451e+00 -7.85411298e-01 -4.28878605e-01 8.69909972e-02
1.78329125e-01 -1.05651967e-01 6.55726254e-01 4.11893398e-01
-4.09596622e-01 4.18131828e-01 5.29327512e-01 -4.16926563e-01
1.79861248e-01 -8.47437084e-01 -1.90269291e-01 1.00475395e+00
-1.63517445e-01 6.21072173e-01 1.03199816e+00 -6.13100946e-01
-1.50999761e+00 -9.36108053e-01 2.77721763e-01 3.53212446e-01
6.18106008e-01 -3.86575341e-01 -1.15085512e-01 1.19206719e-02
7.81457219e-03 -2.95824140e-01 8.93807411e-01 4.53694835e-02
5.07201016e-01 -5.36336362e-01 -1.32470787e+00 7.91210890e-01
1.08648896e+00 1.77301690e-01 5.45769155e-01 1.62298247e-01
3.03742468e-01 -2.41878837e-01 -6.57200336e-01 5.70611417e-01
8.59395802e-01 -9.08976495e-01 1.09212101e+00 3.88939589e-01
-5.69548547e-01 -1.99039638e-01 -2.28410915e-01 -1.29417527e+00
-4.14450169e-01 -1.07242882e+00 -1.65968388e-01 1.28406334e+00
7.57061779e-01 -8.25891495e-01 1.02375185e+00 7.96443969e-02
1.62901506e-02 -5.73010325e-01 -1.37801695e+00 -9.67509091e-01
-6.26207173e-01 -7.39777386e-01 6.08888686e-01 1.90696806e-01
-9.89869516e-03 2.10936721e-02 -2.12704882e-01 1.17013502e+00
8.87259066e-01 -2.42957294e-01 7.51810849e-01 -1.35144353e+00
-2.54081368e-01 -3.27261746e-01 -6.51748657e-01 -1.34980166e+00
-2.79034436e-01 -3.56117994e-01 4.99637276e-02 -1.60945880e+00
-5.66215634e-01 -1.19728386e+00 -4.08811837e-01 -3.42492312e-01
2.81523794e-01 4.70099181e-01 -2.92581171e-01 -1.67709932e-01
-2.38520488e-01 1.84904993e-01 8.08071733e-01 1.52578410e-02
-4.29168433e-01 1.02874565e+00 -5.86953878e-01 6.88963652e-01
1.23032010e+00 -2.41779000e-01 -5.70053875e-01 2.06004605e-02
2.88635045e-01 2.96109587e-01 2.89126515e-01 -1.67063141e+00
2.62712002e-01 -1.60132855e-01 1.01038240e-01 -4.65297669e-01
4.33663666e-01 -1.53466213e+00 3.05350721e-01 8.62714410e-01
2.04289019e-01 -6.06207132e-01 1.39953151e-01 1.06354725e+00
3.10519427e-01 -5.93603514e-02 6.05982065e-01 4.02490973e-01
-3.93798918e-01 5.18207438e-02 -8.78342867e-01 -4.82988536e-01
1.16544783e+00 -2.78541297e-01 -2.04365119e-01 -6.93955183e-01
-4.49675828e-01 2.78899550e-01 -7.13569447e-02 6.05214685e-02
3.48882943e-01 -8.70593607e-01 -1.45479634e-01 -1.31082967e-01
-3.61270085e-02 -6.13441825e-01 2.49441475e-01 8.95626068e-01
-2.16822013e-01 7.74171829e-01 2.44270191e-01 -4.47930336e-01
-1.41515589e+00 -3.33135352e-02 4.37925071e-01 -1.49349064e-01
-2.55566508e-01 6.29483938e-01 -1.42368421e-01 3.31242561e-01
4.84682083e-01 -3.04978967e-01 -1.72420278e-01 -2.71426052e-01
4.85596180e-01 2.99213529e-01 2.57732749e-01 -1.15176030e-01
-5.42405128e-01 6.13496184e-01 3.37847680e-01 -1.51749611e-01
9.78038430e-01 -8.46192300e-01 1.16024666e-01 5.09242043e-02
6.52791440e-01 8.45233500e-01 -7.54380941e-01 -1.89639524e-01
3.04758340e-01 -3.68424803e-01 3.19833010e-01 -9.34925437e-01
-7.52553403e-01 1.37433469e-01 9.25327897e-01 6.53340578e-01
1.16738689e+00 -8.20637047e-02 5.65410256e-01 3.99047375e-01
1.12032604e+00 -7.39757597e-01 -2.05451831e-01 -1.89138159e-01
2.81721771e-01 -8.50686133e-01 2.14706808e-01 -8.47433984e-01
5.15441954e-01 1.10000002e+00 9.55899656e-02 5.45460939e-01
9.30380881e-01 4.02160376e-01 6.48710057e-02 -2.21366107e-01
-2.61874944e-01 -3.04920822e-01 -1.98772624e-01 7.73619056e-01
3.45066428e-01 2.98848540e-01 -3.29552978e-01 1.91174924e-01
-2.43314654e-01 -4.38379794e-01 6.61812723e-01 1.07688570e+00
-9.81810153e-01 -1.23964882e+00 -4.99580383e-01 4.45082307e-01
-4.00758862e-01 6.62050024e-02 1.37771726e-01 7.52086222e-01
1.17403843e-01 1.43198323e+00 -2.78151035e-01 -6.61505200e-03
4.22413051e-01 -5.73596537e-01 2.19084010e-01 -3.43165040e-01
1.04896851e-01 4.01206873e-02 4.91170526e-01 -3.55090261e-01
-4.15185750e-01 -4.57247376e-01 -1.07684386e+00 3.50804180e-01
-3.45123202e-01 4.54512089e-01 1.06395257e+00 1.08871281e+00
3.94468486e-01 9.14825618e-01 1.00718915e+00 -4.76558179e-01
5.43863215e-02 -4.61462319e-01 -6.97719157e-01 -6.52689695e-01
7.01136291e-01 -8.45997751e-01 -3.67118448e-01 -7.44192719e-01] | [6.234353542327881, 1.2528659105300903] |
be50c43e-4ba8-40a0-af77-dbcb1ff48e85 | stock-price-prediction-using-bert-and-gan | 2107.09055 | null | https://arxiv.org/abs/2107.09055v1 | https://arxiv.org/pdf/2107.09055v1.pdf | Stock price prediction using BERT and GAN | The stock market has been a popular topic of interest in the recent past. The growth in the inflation rate has compelled people to invest in the stock and commodity markets and other areas rather than saving. Further, the ability of Deep Learning models to make predictions on the time series data has been proven time and again. Technical analysis on the stock market with the help of technical indicators has been the most common practice among traders and investors. One more aspect is the sentiment analysis - the emotion of the investors that shows the willingness to invest. A variety of techniques have been used by people around the globe involving basic Machine Learning and Neural Networks. Ranging from the basic linear regression to the advanced neural networks people have experimented with all possible techniques to predict the stock market. It's evident from recent events how news and headlines affect the stock markets and cryptocurrencies. This paper proposes an ensemble of state-of-the-art methods for predicting stock prices. Firstly sentiment analysis of the news and the headlines for the company Apple Inc, listed on the NASDAQ is performed using a version of BERT, which is a pre-trained transformer model by Google for Natural Language Processing (NLP). Afterward, a Generative Adversarial Network (GAN) predicts the stock price for Apple Inc using the technical indicators, stock indexes of various countries, some commodities, and historical prices along with the sentiment scores. Comparison is done with baseline models like - Long Short Term Memory (LSTM), Gated Recurrent Units (GRU), vanilla GAN, and Auto-Regressive Integrated Moving Average (ARIMA) model. | ['Anukriti Bansal', 'Vikas Bajpai', 'Priyank Sonkiya'] | 2021-07-18 | null | null | null | null | ['stock-price-prediction'] | ['time-series'] | [-8.22848797e-01 -2.63778001e-01 -1.11284375e-01 -2.41526559e-01
-3.37721705e-01 -7.22964227e-01 6.73197448e-01 -2.26069614e-01
-3.27809572e-01 8.98249030e-01 5.06554842e-01 -6.67618752e-01
4.38207716e-01 -1.25839972e+00 -3.78293484e-01 -5.97459674e-01
-9.47476327e-02 2.26969257e-01 -2.57017344e-01 -7.97617972e-01
5.93792975e-01 3.24461877e-01 -1.07024848e+00 1.21050239e-01
2.64670491e-01 1.54107594e+00 -3.27022016e-01 4.58487034e-01
-5.80670178e-01 1.31902874e+00 -7.68479586e-01 -8.00958157e-01
7.23866701e-01 -1.63485616e-01 -2.80097812e-01 -5.69269001e-01
-4.08214599e-01 -3.52025390e-01 -3.95116694e-02 1.23504698e+00
3.10213357e-01 -2.55245656e-01 2.37965018e-01 -1.11090493e+00
-9.45756018e-01 8.26772451e-01 -6.69977248e-01 6.37195587e-01
-1.24549739e-01 2.53294017e-02 1.15034533e+00 -7.32051671e-01
3.05684716e-01 9.61853981e-01 7.50243187e-01 -6.10076217e-03
-5.65061748e-01 -1.09109664e+00 2.24028584e-02 1.04851071e-02
-7.16576338e-01 3.03262085e-01 8.40472877e-01 -5.44002712e-01
1.22985733e+00 2.03934774e-01 9.94479120e-01 9.98925686e-01
8.71858180e-01 6.41276717e-01 1.53157043e+00 -1.95532084e-01
1.37273222e-01 5.21692097e-01 -4.36981954e-02 9.95720625e-02
8.72903168e-02 3.74013335e-01 -2.35968754e-01 -1.02223810e-02
6.30618095e-01 4.64634746e-01 2.06562623e-01 5.99417150e-01
-1.00268352e+00 1.32269168e+00 5.18059194e-01 7.71554172e-01
-9.85318661e-01 7.07660094e-02 6.56054020e-01 7.90914178e-01
9.94199097e-01 4.70976770e-01 -1.03212774e+00 -4.25699770e-01
-1.06348968e+00 3.35595280e-01 1.03861678e+00 3.21356893e-01
2.50948757e-01 7.60672808e-01 3.65784675e-01 2.18798950e-01
4.06057000e-01 6.53144002e-01 1.12485147e+00 -1.69854000e-01
2.72173524e-01 6.30652964e-01 2.02988684e-01 -1.30313504e+00
-2.02917531e-01 -7.27198899e-01 -9.07946944e-01 6.59684181e-01
9.49977860e-02 -6.05197787e-01 -7.81826675e-01 1.23165715e+00
-3.22878391e-01 7.65555277e-02 2.44245261e-01 5.47879517e-01
3.55532050e-01 1.41823709e+00 -4.47583981e-02 -1.38142377e-01
1.18717456e+00 -7.39168227e-01 -8.65423501e-01 -8.12244415e-02
1.26286611e-01 -8.67651582e-01 5.32232046e-01 5.14231622e-01
-1.08177042e+00 -4.22531903e-01 -9.59217966e-01 4.13089722e-01
-1.08101749e+00 -3.31393808e-01 6.31220937e-01 5.88528991e-01
-1.00461912e+00 6.85060918e-01 -7.81943202e-01 1.80427596e-01
5.89741617e-02 3.69476616e-01 8.12085271e-02 9.06055272e-01
-1.54905653e+00 1.28733814e+00 3.24723303e-01 6.53390288e-02
-3.01869214e-01 -6.69690847e-01 -5.65272689e-01 1.40021145e-01
-2.68564075e-01 -2.37025112e-01 1.19213080e+00 -1.65487027e+00
-1.78509653e+00 5.58349431e-01 4.02251601e-01 -1.11235309e+00
4.19583976e-01 -2.39607632e-01 -8.42806518e-01 -5.26350617e-01
-6.32576123e-02 1.96052372e-01 7.39396989e-01 -3.24982554e-01
-8.66345882e-01 -3.07599664e-01 -7.99088553e-02 -1.91828728e-01
-9.26621407e-02 3.74471039e-01 5.93275011e-01 -1.32933271e+00
-2.44246528e-01 -8.03631604e-01 -2.65030831e-01 -9.41476703e-01
-1.56798452e-01 -2.90771991e-01 8.98001313e-01 -1.25416481e+00
1.24507511e+00 -1.93339252e+00 -2.80481130e-01 4.00919497e-01
-4.15485680e-01 1.66778058e-01 4.61347073e-01 5.16139984e-01
-6.65399611e-01 4.29354221e-01 2.84293026e-01 1.73447862e-01
1.56652004e-01 -1.06916554e-01 -1.07143116e+00 2.89927930e-01
1.11122273e-01 1.20820296e+00 -3.93723279e-01 3.17438304e-01
1.63346633e-01 3.62990499e-01 -2.02370271e-01 2.78085414e-02
-3.08984101e-01 1.27799049e-01 -5.16008198e-01 6.62085295e-01
4.84782279e-01 -2.38627806e-01 -4.17569995e-01 1.01700369e-02
-6.86602116e-01 4.93407130e-01 -8.63982201e-01 1.03202736e+00
-3.73702496e-01 6.86340690e-01 -4.22382802e-01 -9.15747285e-01
1.25490296e+00 6.18588328e-01 3.91243309e-01 -9.18877780e-01
4.01307076e-01 6.71630740e-01 -8.60824808e-02 -1.05863377e-01
4.07376468e-01 -7.26288974e-01 -2.34605730e-01 6.41153574e-01
-2.48341411e-01 1.33598864e-01 -5.57330921e-02 -1.54571384e-01
6.26500249e-01 1.22914910e-01 4.22297031e-01 -1.76602125e-01
3.57595414e-01 2.30984371e-02 4.59882379e-01 2.13995054e-02
1.90930888e-02 5.29275611e-02 4.04962569e-01 -1.06822324e+00
-1.26591980e+00 -6.29731596e-01 1.61894292e-01 8.02576303e-01
-5.45041263e-01 2.41396859e-01 -3.76630753e-01 -2.37761393e-01
-2.54178117e-03 1.05206692e+00 -7.80398488e-01 2.51343697e-01
-3.15622449e-01 -9.37047243e-01 1.15088500e-01 4.69764888e-01
7.09493279e-01 -1.55354595e+00 -6.95207894e-01 5.05391061e-01
3.19856763e-01 -6.19384408e-01 -1.85810612e-03 3.21066707e-01
-7.60385990e-01 -4.07787621e-01 -1.00883293e+00 -3.60967278e-01
1.20840698e-01 -5.78175306e-01 1.17971861e+00 -3.94417793e-01
1.42217547e-01 -7.31803477e-02 -2.97105819e-01 -1.30425227e+00
-4.41554725e-01 -8.02449603e-03 5.19809220e-03 7.82108530e-02
6.76002741e-01 -5.64212859e-01 -5.39677799e-01 -1.63471371e-01
-7.18237877e-01 -1.23777680e-01 5.43378472e-01 5.58873057e-01
1.51777208e-01 4.11473960e-01 9.16098833e-01 -7.66893327e-01
6.33680105e-01 -8.79030824e-01 -1.00220275e+00 -1.17580764e-01
-7.91577876e-01 3.34472246e-02 4.36500788e-01 -1.08085655e-01
-1.10645294e+00 -6.05480075e-01 -2.05727577e-01 -2.27246642e-01
3.27075809e-01 1.06563377e+00 4.10986334e-01 3.03413540e-01
2.00105011e-01 1.97803363e-01 7.53465220e-02 -2.01769143e-01
-1.20208552e-02 7.01425374e-01 4.04440820e-01 1.35889262e-01
9.44772959e-01 2.58462191e-01 -3.61325681e-01 -2.97634184e-01
-8.05096507e-01 -1.07470751e-01 -1.77368656e-01 -2.15640947e-01
9.10478175e-01 -1.06145108e+00 -6.10163987e-01 1.01046264e+00
-1.02688658e+00 3.53079811e-02 -2.84979820e-01 5.12194574e-01
-1.22461565e-01 -3.75707209e-01 -1.02225804e+00 -1.09157896e+00
-8.45090926e-01 -9.20754373e-01 2.81315655e-01 4.66039896e-01
-3.19218308e-01 -1.41881692e+00 3.17461729e-01 -6.45214692e-02
1.09373283e+00 7.53990054e-01 7.58888781e-01 -1.19622576e+00
-4.47779715e-01 -6.93110406e-01 5.60397282e-02 8.62139702e-01
2.60535419e-01 8.04759488e-02 -7.96311975e-01 -1.23533651e-01
5.95993161e-01 -7.46798068e-02 4.18050766e-01 6.99217081e-01
3.56436431e-01 -6.24581933e-01 3.19823474e-01 3.69041950e-01
1.59671760e+00 8.46082747e-01 7.48654246e-01 9.96273458e-01
1.04701415e-01 3.05161566e-01 4.33248520e-01 5.59958875e-01
1.59179479e-01 -5.15879430e-02 3.26799184e-01 -1.13774724e-01
7.86533237e-01 -2.12099925e-01 9.06425893e-01 1.04358768e+00
-2.81997532e-01 3.27168480e-02 -7.10129619e-01 2.36391127e-01
-1.49867308e+00 -1.28252280e+00 2.49057338e-01 1.84926975e+00
6.07534230e-01 6.04377329e-01 1.78218067e-01 2.95186281e-01
4.29122090e-01 3.02822679e-01 -5.85905671e-01 -8.46179962e-01
-2.73154706e-01 4.62600648e-01 1.00866854e+00 9.40218866e-02
-9.83723223e-01 7.96606421e-01 5.86993980e+00 4.38975811e-01
-1.82891750e+00 -1.12821773e-01 1.23349202e+00 1.57654569e-01
-3.97085845e-01 -2.45948419e-01 -6.84455574e-01 8.39820087e-01
1.40519536e+00 -5.96175432e-01 2.87917614e-01 1.18155789e+00
3.26875240e-01 1.23021096e-01 -4.57587212e-01 6.28036201e-01
-2.18015611e-01 -1.54243183e+00 -9.84800830e-02 2.21331760e-01
8.07304442e-01 3.96499157e-01 6.33073747e-01 5.61945975e-01
4.47754145e-01 -9.62560475e-01 8.79339039e-01 6.75533056e-01
2.44939402e-01 -1.00877261e+00 1.31039870e+00 2.71811932e-01
-8.50121617e-01 -3.27515662e-01 -2.00830981e-01 -4.08797979e-01
2.47116834e-01 5.05047679e-01 -5.83698988e-01 2.63719261e-01
9.71431673e-01 7.20495820e-01 -7.02883080e-02 3.73670518e-01
-6.95688510e-03 7.14644074e-01 -2.99857646e-01 -2.88499326e-01
8.35840285e-01 -5.35680175e-01 2.80561596e-01 1.00302696e+00
7.55731881e-01 2.19826669e-01 -4.98006046e-01 8.18519354e-01
7.87670910e-02 2.47141361e-01 -5.55455327e-01 -5.01722336e-01
-9.51471180e-02 1.08652258e+00 -6.56447411e-01 -4.12511110e-01
-8.25187087e-01 6.47292614e-01 -4.29499775e-01 2.39020884e-01
-7.94664800e-01 -1.68864042e-01 5.98117352e-01 1.55500218e-01
3.33346575e-01 8.76510069e-02 -3.33994687e-01 -1.09076810e+00
-1.37712002e-01 -9.92580354e-01 3.76899958e-01 -8.03938806e-01
-1.50369835e+00 7.47545719e-01 -3.78053278e-01 -1.12865174e+00
-6.25919104e-01 -8.11605573e-01 -9.91052866e-01 1.32868648e+00
-1.71141565e+00 -7.39834011e-01 4.82274145e-01 5.36813736e-01
7.11567700e-01 -1.00748408e+00 6.97267771e-01 4.07217368e-02
-4.26109791e-01 -9.83046219e-02 1.01626508e-01 5.27475595e-01
1.26070663e-01 -1.37414038e+00 9.27123666e-01 8.23519349e-01
1.26507401e-01 4.01296705e-01 9.01483536e-01 -6.78177595e-01
-9.03024018e-01 -8.14487696e-01 1.00791037e+00 -1.07499279e-01
1.39333797e+00 1.30484954e-01 -7.21614122e-01 1.08124936e+00
8.01157534e-01 -2.96540916e-01 6.58525288e-01 -4.14604366e-01
-1.69655204e-01 -2.15063989e-01 -1.05441427e+00 2.74655282e-01
-3.69833320e-01 -2.48295546e-01 -8.18846345e-01 1.05236113e-01
6.37191236e-01 -1.44079700e-01 -7.40640700e-01 1.64347291e-01
5.12646198e-01 -1.23667884e+00 6.37383580e-01 -6.24882936e-01
4.75781530e-01 1.00832075e-01 -7.51700625e-02 -1.57851982e+00
-2.30500370e-01 -6.97208583e-01 1.46434873e-01 1.15714920e+00
7.48841524e-01 -1.14053369e+00 7.57539392e-01 9.51106429e-01
2.70159930e-01 -7.09650218e-01 -6.48834467e-01 -3.65079671e-01
2.67645597e-01 -4.11858618e-01 7.36363053e-01 1.04865062e+00
-8.56825039e-02 1.53855547e-01 -4.88699287e-01 -1.16062440e-01
1.27594456e-01 3.05167019e-01 4.12675709e-01 -1.22398913e+00
-3.35663766e-01 -7.31742799e-01 -4.62129205e-01 -4.64970618e-01
1.02826310e-02 -5.50697267e-01 -8.70187461e-01 -1.19769347e+00
-3.76413345e-01 -9.44614261e-02 -8.30873609e-01 2.85821706e-01
3.83064508e-01 -1.22303560e-01 2.79130280e-01 1.84389278e-01
3.44852775e-01 1.31180391e-01 9.17482495e-01 -6.99286442e-03
-7.88122267e-02 3.70313883e-01 -6.13204837e-01 8.21785867e-01
1.15413725e+00 -2.39758298e-01 4.96899849e-03 1.48575887e-01
1.07654512e+00 1.77772835e-01 2.98644844e-02 -7.48460412e-01
3.91381159e-02 -9.29224677e-03 5.62253237e-01 -9.20055687e-01
4.33156863e-02 -9.32486415e-01 5.66529274e-01 8.30775201e-01
-1.40500918e-01 9.98346329e-01 1.81031078e-01 1.07220009e-01
-7.95299649e-01 -2.02258341e-02 5.87805092e-01 -5.86954296e-01
-6.74358845e-01 3.24235708e-01 -4.23487723e-01 -2.64012516e-01
1.13070643e+00 -2.01418605e-02 -2.06946403e-01 -8.38630736e-01
-4.53700870e-01 -3.47956605e-02 -7.18392953e-02 6.58931732e-01
3.57679367e-01 -1.21079934e+00 -8.87721598e-01 3.43271226e-01
-6.44326389e-01 -4.71319109e-01 -1.28173932e-01 6.51534855e-01
-8.95910203e-01 7.15607345e-01 -5.18356800e-01 2.15481177e-01
-6.40309751e-01 6.49832070e-01 4.78738636e-01 -7.30615854e-01
-4.06268090e-01 8.51248264e-01 -2.57692277e-01 3.79851833e-02
-4.06391993e-02 -7.45636106e-01 -4.69404459e-01 5.64367890e-01
6.30833685e-01 2.48795718e-01 1.33138252e-02 -5.89835048e-01
1.37135328e-03 3.89060050e-01 -6.62534982e-02 -2.77084082e-01
2.03276777e+00 1.55715272e-01 -2.69833416e-01 1.06776118e+00
1.27905774e+00 -7.23698735e-02 -7.51883626e-01 -1.75512973e-02
3.23110133e-01 4.35622893e-02 3.06004643e-01 -1.03805935e+00
-1.65062881e+00 7.18234479e-01 4.96709108e-01 8.26380134e-01
1.03079987e+00 -5.80988050e-01 1.25267792e+00 4.13588434e-02
3.97588938e-01 -1.18483591e+00 -3.32737237e-01 3.88805956e-01
8.52305532e-01 -1.21764374e+00 -2.52606660e-01 4.78827417e-01
-8.94529939e-01 1.28656483e+00 -1.40854299e-01 -5.57485104e-01
1.26799321e+00 7.29487538e-01 5.92928410e-01 -9.36444327e-02
-8.12992513e-01 4.70100164e-01 -1.08746774e-01 -1.24774262e-01
6.83853030e-01 1.69926867e-01 -1.76942050e-02 7.79129207e-01
-6.83083355e-01 1.62009314e-01 5.52473783e-01 6.84788585e-01
-2.70976454e-01 -9.13138270e-01 -4.60419774e-01 5.26202142e-01
-1.40398836e+00 -2.76607335e-01 -7.43459687e-02 8.71874452e-01
-2.00412795e-01 5.50496161e-01 4.67070341e-01 -2.26793423e-01
1.14409499e-01 2.56293923e-01 -5.11063457e-01 -1.75299793e-01
-1.11308241e+00 1.77184209e-01 -2.56783009e-01 -1.83158919e-01
-5.01578867e-01 -7.70517707e-01 -1.10440338e+00 -5.68062961e-01
-4.09179665e-02 3.37493807e-01 9.75399196e-01 7.55754411e-01
-3.72525230e-02 3.26565981e-01 1.06240904e+00 -6.64176404e-01
-7.90649414e-01 -1.15912068e+00 -9.24950778e-01 1.59218580e-01
3.68543386e-01 -1.53079689e-01 -6.89413667e-01 -8.15327838e-02] | [4.433347702026367, 4.255926609039307] |
8ab3fbb0-d420-4301-882c-d082d2e91f4c | streaming-robust-submodular-maximization-a | 1711.02598 | null | http://arxiv.org/abs/1711.02598v1 | http://arxiv.org/pdf/1711.02598v1.pdf | Streaming Robust Submodular Maximization: A Partitioned Thresholding Approach | We study the classical problem of maximizing a monotone submodular function
subject to a cardinality constraint k, with two additional twists: (i) elements
arrive in a streaming fashion, and (ii) m items from the algorithm's memory are
removed after the stream is finished. We develop a robust submodular algorithm
STAR-T. It is based on a novel partitioning structure and an exponentially
decreasing thresholding rule. STAR-T makes one pass over the data and retains a
short but robust summary. We show that after the removal of any m elements from
the obtained summary, a simple greedy algorithm STAR-T-GREEDY that runs on the
remaining elements achieves a constant-factor approximation guarantee. In two
different data summarization tasks, we demonstrate that it matches or
outperforms existing greedy and streaming methods, even if they are allowed the
benefit of knowing the removed subset in advance. | ['Ashkan Norouzi-Fard', 'Slobodan Mitrović', 'Volkan Cevher', 'Jakub Tarnawski', 'Ilija Bogunovic'] | 2017-11-07 | streaming-robust-submodular-maximization-a-1 | http://papers.nips.cc/paper/7042-streaming-robust-submodular-maximization-a-partitioned-thresholding-approach | http://papers.nips.cc/paper/7042-streaming-robust-submodular-maximization-a-partitioned-thresholding-approach.pdf | neurips-2017-12 | ['data-summarization'] | ['miscellaneous'] | [ 5.28346360e-01 3.48039597e-01 -6.82571352e-01 -2.05946863e-01
-9.09628630e-01 -9.79804397e-01 -1.26119377e-02 7.82942951e-01
-3.66255194e-01 7.69201636e-01 3.71914774e-01 8.11492279e-02
-5.09892941e-01 -7.10135043e-01 -9.71623957e-01 -7.46481180e-01
-4.58877832e-01 9.53439116e-01 2.99105823e-01 1.59820188e-02
4.02824879e-01 3.38454604e-01 -1.19934773e+00 2.74218827e-01
7.78108060e-01 1.00467253e+00 1.64222851e-01 7.49515355e-01
-1.41027709e-02 5.92770755e-01 -4.41895306e-01 -1.00008182e-01
7.69466162e-01 -5.73861122e-01 -8.50047052e-01 6.74236536e-01
4.39389020e-01 -5.59041202e-01 -5.53702533e-01 7.69794703e-01
2.68037438e-01 6.90691769e-02 3.06170613e-01 -1.30755389e+00
-9.07848328e-02 1.08039176e+00 -8.43434691e-01 1.35435253e-01
4.04079884e-01 -2.49726281e-01 1.43262053e+00 -6.25134647e-01
1.00848472e+00 8.28034997e-01 4.02547479e-01 1.46634519e-01
-1.29354906e+00 -1.15432568e-01 4.45903093e-01 -1.97718754e-01
-1.28697097e+00 -5.34675002e-01 3.89824182e-01 6.44970611e-02
8.76007318e-01 8.25658619e-01 7.10443556e-01 -4.30448204e-02
-1.48471375e-03 1.03315067e+00 5.93056083e-01 -9.83800963e-02
5.27477860e-01 2.33556051e-02 4.66621697e-01 4.59975749e-01
8.92055929e-01 -5.45833230e-01 -9.33623910e-01 -5.83038867e-01
1.43962055e-02 -8.52439646e-03 -4.23057199e-01 -5.97080946e-01
-1.16823125e+00 7.30079472e-01 -3.76632698e-02 7.44627789e-02
-5.86803555e-01 3.72856408e-01 4.10551071e-01 6.04656279e-01
4.47311699e-01 2.45646656e-01 -4.51269120e-01 3.19840536e-02
-1.55018497e+00 7.11151421e-01 1.12736046e+00 1.54589260e+00
7.11574018e-01 -2.44016394e-01 -3.00401956e-01 3.79251301e-01
-2.01826274e-01 4.87506777e-01 -1.15285397e-01 -1.20199358e+00
7.48778582e-01 5.34217477e-01 2.52657354e-01 -8.92126322e-01
-4.01133716e-01 -3.50025952e-01 -5.76255083e-01 -3.28632653e-01
1.75829515e-01 -1.69418439e-01 -7.40029871e-01 1.65117085e+00
4.53865528e-01 -2.87675738e-01 -1.36220276e-01 7.15428948e-01
5.50228477e-01 8.35702300e-01 -6.66544974e-01 -1.17142987e+00
8.34476471e-01 -8.34169984e-01 -6.29171371e-01 -3.67901698e-02
6.05306983e-01 -4.05954868e-01 2.42908821e-01 7.75188982e-01
-1.93783903e+00 3.14020902e-01 -1.07101822e+00 -2.02677876e-01
2.53439456e-01 -2.98210233e-01 5.29265165e-01 5.57844162e-01
-1.12736750e+00 6.02082849e-01 -7.48075306e-01 -2.37023458e-01
3.23527902e-01 4.73696351e-01 -2.49849886e-01 -3.95883173e-01
-3.39459658e-01 1.55688703e-01 3.70071441e-01 -2.25821629e-01
-9.42786813e-01 -8.16282272e-01 -6.66030943e-01 3.16824704e-01
8.68516386e-01 -1.13333714e+00 1.33383834e+00 -7.07913995e-01
-9.24797952e-01 7.43379056e-01 -5.71672976e-01 -6.51835620e-01
7.31922925e-01 -1.26719669e-01 3.39470565e-01 5.10235369e-01
1.50307983e-01 2.67277509e-01 7.07341075e-01 -1.22822332e+00
-9.80900943e-01 -7.67571449e-01 4.32705842e-02 4.25570518e-01
-4.87225085e-01 -1.23925380e-01 -7.60685146e-01 -5.22055209e-01
5.17434180e-01 -9.04946744e-01 -5.67170382e-01 -2.09202200e-01
-6.29540265e-01 -3.49825434e-02 4.81788903e-01 -4.18938339e-01
1.72222328e+00 -1.84220505e+00 4.96203870e-01 4.38240439e-01
6.66596055e-01 -3.14179868e-01 2.45304424e-02 9.26426530e-01
3.69367242e-01 1.66709438e-01 -6.37565553e-01 -4.50304508e-01
-3.47707383e-02 2.21236184e-01 -2.41950184e-01 7.46251166e-01
-5.79039991e-01 6.09867096e-01 -8.26993048e-01 -3.23489130e-01
-5.33154070e-01 -6.69913292e-01 -7.98286438e-01 -2.41165772e-01
-6.31566107e-01 -2.28331909e-01 -2.19631970e-01 5.67071676e-01
1.01792800e+00 -2.31127635e-01 5.27283728e-01 2.23095357e-01
-2.28146508e-01 2.95934767e-01 -1.67184138e+00 1.78622961e+00
1.05148941e-01 3.39905024e-01 6.78188205e-01 -9.54217911e-01
5.23100793e-01 2.04202086e-01 9.88013268e-01 -2.49713376e-01
-2.13595703e-02 5.20069301e-01 -4.17662531e-01 -3.07753235e-01
9.13126707e-01 -7.66385272e-02 -3.06017280e-01 9.80936766e-01
-3.38490099e-01 -2.48224195e-02 6.36858165e-01 9.81739819e-01
1.50440943e+00 -7.11531222e-01 3.35247427e-01 -3.45073104e-01
1.71616808e-01 5.29672682e-01 9.25768912e-01 1.07309759e+00
2.91779608e-01 7.56455362e-01 8.54941547e-01 -1.82091415e-01
-1.01747906e+00 -8.78231585e-01 1.09108254e-01 1.03797150e+00
3.47359151e-01 -7.80221522e-01 -5.60878813e-01 -5.59165061e-01
4.54301566e-01 5.73060930e-01 -5.89234412e-01 1.85350507e-01
-5.27380526e-01 -5.67707479e-01 -6.21341616e-02 3.64471525e-01
-5.10606728e-02 -4.62371707e-01 -4.97482896e-01 3.92883122e-01
-1.35555610e-01 -9.22460198e-01 -1.12363291e+00 3.30880851e-01
-1.28686953e+00 -8.99390280e-01 -4.76075441e-01 -7.78463125e-01
9.83631194e-01 1.09268963e+00 9.51766551e-01 1.17729023e-01
7.20286220e-02 4.67956990e-01 -2.80198097e-01 -3.39086175e-01
-3.40816081e-02 2.97092944e-01 -5.80430999e-02 -2.05725580e-01
-5.17106913e-02 -4.98811722e-01 -4.52011764e-01 5.62885739e-02
-1.12717235e+00 -1.94734752e-01 2.88428366e-01 3.76753360e-01
9.48441565e-01 3.69858563e-01 6.23166800e-01 -1.30868685e+00
5.90916276e-01 -9.18779016e-01 -5.54423690e-01 1.16327159e-01
-5.50555289e-01 -5.68401441e-02 5.64787924e-01 -1.08498640e-01
-5.97002566e-01 1.08847171e-01 4.81890172e-01 -2.44419813e-01
4.94496316e-01 6.94342673e-01 -2.35785738e-01 2.60731101e-01
3.86488885e-01 4.41876382e-01 -7.43164793e-02 -3.06566000e-01
2.97340214e-01 5.82673728e-01 6.42454326e-01 -4.06998456e-01
8.59128058e-01 8.58443022e-01 1.52916223e-01 -9.55074310e-01
-7.00716972e-01 -8.74813080e-01 -3.84443372e-01 8.80702659e-02
1.86742604e-01 -9.22095478e-01 -5.30271947e-01 3.25968146e-01
-1.03903425e+00 -1.08410791e-01 -6.66452706e-01 2.26244643e-01
-7.14804530e-01 6.25031352e-01 -3.55539531e-01 -8.34037602e-01
-4.82829303e-01 -4.57194418e-01 7.85038054e-01 -1.17033474e-01
-1.05687745e-01 -5.04185200e-01 8.97136033e-02 3.48166823e-01
-1.96851958e-02 2.00298935e-01 8.44077289e-01 -9.16258037e-01
-1.05788815e+00 -2.30966911e-01 6.02552071e-02 1.24631099e-01
2.49604564e-02 -2.68010616e-01 -2.34164000e-01 -6.74933970e-01
4.14878093e-02 1.16697522e-02 1.13475084e+00 6.29498601e-01
1.10153544e+00 -9.77621973e-01 -4.87101048e-01 7.85678089e-01
1.49237025e+00 4.04524943e-03 2.26938486e-01 1.02096364e-01
4.01043296e-01 4.52327222e-01 7.90699065e-01 1.06374669e+00
6.48925900e-01 2.30682507e-01 5.47059953e-01 3.86800885e-01
3.83967400e-01 -1.29517138e-01 4.10845578e-01 9.18424428e-01
2.37650931e-01 -7.86329806e-01 -2.32278302e-01 1.22063446e+00
-2.20961142e+00 -1.02350914e+00 -5.08694112e-01 2.67277265e+00
9.23029661e-01 1.78031310e-01 6.23896182e-01 4.03017431e-01
5.52558303e-01 3.61872464e-01 -8.08585882e-01 -7.47196555e-01
-3.28102708e-01 -6.25434518e-02 1.16028452e+00 6.21402383e-01
-6.48088992e-01 5.87008655e-01 6.58355427e+00 5.95597982e-01
-5.67117929e-01 -1.04237255e-02 3.70097131e-01 -9.08117294e-01
-1.08651769e+00 1.70185581e-01 -7.88260818e-01 2.93481886e-01
6.35177314e-01 -8.44501495e-01 5.29788315e-01 6.56948149e-01
3.42734307e-01 -4.21701401e-01 -1.32357943e+00 7.71472931e-01
2.97626138e-01 -1.67517960e+00 7.46323727e-03 2.30199516e-01
1.21865571e+00 -1.66611090e-01 -2.25206196e-01 -3.54723096e-01
3.83524120e-01 -7.08769500e-01 8.76303673e-01 1.18699402e-01
5.32705128e-01 -9.56678748e-01 3.13627988e-01 7.54834890e-01
-1.08058977e+00 -3.55900019e-01 -3.81417394e-01 3.40156741e-02
4.48094279e-01 8.96434605e-01 -7.48608589e-01 7.98422277e-01
5.47699988e-01 4.51321393e-01 1.82125390e-01 1.32153213e+00
1.74855858e-01 6.90251291e-01 -8.78795981e-01 9.08728093e-02
1.97958291e-01 -3.25477153e-01 1.20013916e+00 1.17163742e+00
3.79732490e-01 4.46563035e-01 4.05035615e-01 3.55375886e-01
-4.19634998e-01 2.03260079e-01 -4.89902765e-01 4.80808020e-02
7.67164409e-01 1.10245323e+00 -6.84392333e-01 -4.88069713e-01
-4.28359032e-01 8.58411252e-01 4.80222367e-02 1.83560416e-01
-4.80063081e-01 -3.60848099e-01 3.52426559e-01 5.99942982e-01
7.76553571e-01 -4.37696964e-01 -8.25787365e-01 -7.87554681e-01
4.44355607e-01 -7.23035216e-01 9.00028944e-01 -2.20324382e-01
-8.13796103e-01 4.15853187e-02 9.82982852e-03 -8.01002145e-01
3.24411318e-02 1.59173146e-01 -4.43629622e-01 3.42922479e-01
-1.31235790e+00 -7.26211131e-01 -2.10509405e-01 5.15803874e-01
4.93253201e-01 2.23000273e-01 1.17167950e-01 2.21206620e-03
-3.75899196e-01 5.11221647e-01 3.85855317e-01 -6.50882244e-01
3.35472465e-01 -1.38662040e+00 6.19306788e-02 1.05701852e+00
-1.09068677e-01 3.75231564e-01 9.53741133e-01 -7.35629022e-01
-1.87635303e+00 -1.03081918e+00 1.16304564e+00 2.51007359e-03
3.90502185e-01 -2.89133072e-01 -7.95203209e-01 9.57758307e-01
2.36804008e-01 -2.31505007e-01 5.90378404e-01 -1.61212876e-01
-6.89973980e-02 -3.93292248e-01 -1.28399289e+00 2.24845111e-01
1.33097875e+00 1.93350971e-01 -5.69071054e-01 5.86228251e-01
7.75438845e-01 -4.12429988e-01 -6.84933424e-01 2.32435331e-01
4.98205096e-01 -6.40818715e-01 5.21556318e-01 -7.76752830e-01
3.01312894e-01 -3.06953996e-01 -3.77912968e-01 -1.08661163e+00
-2.86194891e-01 -1.45452428e+00 -5.61883867e-01 8.68333578e-01
4.62549388e-01 -5.00655830e-01 9.84833300e-01 4.73385721e-01
-1.95048198e-01 -7.96291351e-01 -8.92453134e-01 -9.23554480e-01
-1.39660209e-01 -1.57271162e-01 5.29615641e-01 5.45403242e-01
2.90304393e-01 3.21687251e-01 -5.58011949e-01 2.62861639e-01
8.46510112e-01 7.12368190e-01 1.02507889e+00 -1.18552637e+00
-3.06150436e-01 -2.10474923e-01 1.74809694e-01 -1.66268468e+00
-3.81617934e-01 -1.04877043e+00 -1.19205900e-01 -1.81010818e+00
7.58898139e-01 -4.64267284e-01 -7.85338953e-02 2.65916526e-01
-4.33874130e-02 -1.98374256e-01 2.16394380e-01 2.99690634e-01
-1.14368713e+00 3.76355678e-01 1.10788643e+00 -2.28449404e-02
-6.58516645e-01 3.12795728e-01 -1.25274372e+00 3.74639511e-01
5.85148215e-01 -8.85181069e-01 -4.45789039e-01 -5.22830665e-01
5.77402711e-01 6.18214548e-01 -3.52526963e-01 -3.45032394e-01
7.52398252e-01 -4.01651978e-01 -2.42338464e-01 -1.16495466e+00
1.09307542e-02 -8.10390294e-01 2.29027793e-01 5.01850307e-01
-3.75978649e-01 2.00105041e-01 -1.02791421e-01 8.10262620e-01
-7.32975602e-02 -4.59445357e-01 7.01593459e-01 -8.65271091e-02
-1.76236212e-01 6.61698878e-01 -4.39028144e-01 2.86766976e-01
1.28541124e+00 -6.10728979e-01 -3.79158944e-01 -6.19111836e-01
-5.88675499e-01 8.74648035e-01 7.31895089e-01 -7.24350289e-02
4.03355598e-01 -1.06395769e+00 -9.49104667e-01 -2.23189935e-01
6.20908886e-02 5.38503289e-01 3.08696538e-01 1.12556159e+00
-4.01634037e-01 5.65309525e-01 3.11514974e-01 -4.17643905e-01
-1.46606338e+00 8.30428898e-01 -2.15471491e-01 -3.05281609e-01
-6.94212317e-01 7.48521745e-01 1.12757489e-01 7.62782618e-02
2.65028328e-01 -2.76903749e-01 3.85206670e-01 3.60412866e-01
4.54416782e-01 9.19733465e-01 2.02987090e-01 -1.80745140e-01
-3.29470605e-01 1.76417872e-01 -3.58999997e-01 -2.47536555e-01
1.70910680e+00 -4.61352468e-01 -4.43898767e-01 2.83346027e-01
9.25836205e-01 7.32337594e-01 -8.11914980e-01 -4.57255274e-01
-9.46727023e-03 -7.00503767e-01 -1.56532750e-01 -3.28425288e-01
-1.12533998e+00 -6.56792372e-02 -5.09909749e-01 6.62043869e-01
1.30672562e+00 1.28702536e-01 1.35720074e+00 3.65623653e-01
5.22617579e-01 -1.27931440e+00 -1.43554509e-01 2.24554434e-01
7.86143541e-01 -6.28107131e-01 5.63563883e-01 -6.95778430e-01
-3.24439168e-01 8.84785652e-01 8.77104625e-02 -1.13535695e-01
4.07739848e-01 4.73884553e-01 -6.06983602e-01 -2.17461720e-01
-1.35284185e+00 -1.15087271e-01 -2.63774961e-01 1.93440184e-01
-2.01204047e-01 6.22412339e-02 -8.88669789e-01 7.93384910e-01
-2.49491990e-01 -6.64994493e-02 9.98689473e-01 1.18196225e+00
-1.10968924e+00 -8.56268823e-01 -3.70738417e-01 8.73443067e-01
-5.06101668e-01 1.24084622e-01 -6.76455498e-01 4.91504163e-01
-2.14046195e-01 1.07983935e+00 1.31126389e-01 -4.86596189e-02
3.15227449e-01 -3.92849922e-01 5.45568168e-01 -8.34163845e-01
-4.17260617e-01 1.47979677e-01 3.32163721e-01 -4.60624337e-01
-8.62734020e-02 -1.14062488e+00 -1.27925992e+00 -7.68293977e-01
-2.90233284e-01 3.54776740e-01 4.70273465e-01 5.90199947e-01
4.10829246e-01 1.48128420e-01 1.07971621e+00 -3.68260115e-01
-8.69957149e-01 -5.10715902e-01 -9.90701079e-01 2.86104865e-02
5.56573451e-01 -3.08580175e-02 -4.81161028e-01 -7.86577463e-02] | [6.535688400268555, 4.949305057525635] |
57929ee6-bbbc-4cf7-bdcb-51117c31e2fe | gan-based-joint-activity-detection-and | 2204.01731 | null | https://arxiv.org/abs/2204.01731v1 | https://arxiv.org/pdf/2204.01731v1.pdf | Gan-Based Joint Activity Detection and Channel Estimation For Grant-free Random Access | Joint activity detection and channel estimation (JADCE) for grant-free random access is a critical issue that needs to be addressed to support massive connectivity in IoT networks. However, the existing model-free learning method can only achieve either activity detection or channel estimation, but not both. In this paper, we propose a novel model-free learning method based on generative adversarial network (GAN) to tackle the JADCE problem. We adopt the U-net architecture to build the generator rather than the standard GAN architecture, where a pre-estimated value that contains the activity information is adopted as input to the generator. By leveraging the properties of the pseudoinverse, the generator is refined by using an affine projection and a skip connection to ensure the output of the generator is consistent with the measurement. Moreover, we build a two-layer fully-connected neural network to design pilot matrix for reducing the impact of receiver noise. Simulation results show that the proposed method outperforms the existing methods in high SNR regimes, as both data consistency projection and pilot matrix optimization improve the learning ability. | ['Yong Zhou', 'Yinan Zou', 'Shuang Liang'] | 2022-04-04 | null | null | null | null | ['activity-detection'] | ['computer-vision'] | [ 4.56730813e-01 1.89139515e-01 -1.77286580e-01 7.66208693e-02
-7.34465003e-01 -2.12221041e-01 3.23429972e-01 -5.79065561e-01
-1.52519047e-01 9.52897370e-01 3.09464544e-01 -5.04165590e-01
8.21197778e-02 -8.58863413e-01 -6.99158251e-01 -1.17365980e+00
1.71806179e-02 -2.83801466e-01 -2.02053800e-01 1.20929137e-01
-2.77351767e-01 1.08830288e-01 -5.13542533e-01 -3.21252972e-01
7.96739399e-01 1.27878821e+00 9.64444354e-02 5.61852336e-01
3.66691351e-01 8.29498529e-01 -7.76377141e-01 -2.24312529e-01
4.40137714e-01 -9.50556338e-01 -9.11317244e-02 -8.93251374e-02
-3.80161524e-01 -5.32706797e-01 -9.20451581e-01 9.87122536e-01
9.43365514e-01 -3.56851310e-01 4.66784894e-01 -1.11771691e+00
-2.97440320e-01 8.00523818e-01 -4.28831756e-01 -1.19059198e-02
3.23300958e-01 2.71987647e-01 6.54102147e-01 -4.68344241e-01
6.25530481e-02 8.10748518e-01 6.60625517e-01 4.25512433e-01
-1.08242381e+00 -1.29016745e+00 -1.39906621e-02 9.12946016e-02
-1.70601964e+00 -4.32152987e-01 8.91814649e-01 -3.24745402e-02
2.25797862e-01 7.95839131e-02 8.58441234e-01 1.43945003e+00
3.48272994e-02 5.09668529e-01 1.07381809e+00 -3.12434405e-01
3.76651645e-01 -3.77488658e-02 -5.98443985e-01 7.27819622e-01
3.47770840e-01 1.60884887e-01 -1.55624568e-01 -1.84814557e-01
1.14964473e+00 3.70022207e-02 -6.04788661e-01 -2.98525095e-01
-1.34283710e+00 6.10443890e-01 5.71683407e-01 8.64053369e-02
-7.16310978e-01 5.31005740e-01 -1.06612444e-01 3.86242181e-01
-2.48801306e-01 4.39476483e-02 -1.18531426e-02 -8.77988264e-02
-8.49465072e-01 -3.35629106e-01 8.91084850e-01 1.02615690e+00
6.66836917e-01 6.93056583e-01 -5.56186974e-01 2.88127720e-01
5.47232270e-01 8.06525111e-01 4.72562164e-01 -6.07660711e-01
6.98050618e-01 1.45037130e-01 4.81654704e-02 -7.94771969e-01
-2.84227639e-01 -1.12603247e+00 -1.61331356e+00 -1.27343699e-01
1.15092769e-01 -9.05353248e-01 -6.29324019e-01 1.97553074e+00
2.43945375e-01 8.90091181e-01 4.13325042e-01 5.92923760e-01
3.80022794e-01 7.95349181e-01 -1.67873457e-01 -4.48330730e-01
8.16375613e-01 -8.68891418e-01 -8.20730746e-01 -1.45189390e-01
3.19060653e-01 -3.01518917e-01 6.63821816e-01 3.39799017e-01
-7.50933707e-01 -4.86789942e-01 -1.72194314e+00 4.99401122e-01
1.82109758e-01 4.19100851e-01 4.16986138e-01 1.17449307e+00
-7.28628993e-01 7.35853761e-02 -5.81491292e-01 1.51723385e-01
5.67235768e-01 4.66850728e-01 -9.69003588e-02 2.38915114e-03
-1.41586959e+00 3.60846132e-01 4.85550284e-01 3.08561027e-01
-1.08830583e+00 -1.98256567e-01 -7.05342174e-01 2.84220308e-01
3.20479125e-01 -7.39215791e-01 9.24169123e-01 -7.58490801e-01
-1.94138265e+00 -1.79798931e-01 2.34069526e-01 -7.01412559e-01
6.42417789e-01 2.09706470e-01 -6.85377061e-01 -1.89917423e-02
-2.35998347e-01 2.09714696e-01 1.06466997e+00 -9.87299383e-01
-4.69722420e-01 3.18614766e-02 1.29319131e-01 1.87584817e-01
-4.04025108e-01 -7.96047509e-01 -6.42251968e-01 -8.65974188e-01
2.64651120e-01 -9.02605414e-01 -4.71644193e-01 -2.32559383e-01
-5.73533475e-01 5.05568206e-01 8.88536394e-01 -8.26518774e-01
1.51976252e+00 -2.09121752e+00 -1.18203849e-01 8.14288914e-01
7.42419362e-02 2.08817601e-01 3.75323072e-02 4.52273697e-01
2.25840226e-01 -8.94294530e-02 -2.08085015e-01 -6.83161467e-02
-6.67043179e-02 5.07820547e-02 -2.72567809e-01 5.65546155e-01
-1.09528890e-02 9.69722629e-01 -8.89742494e-01 -1.22763723e-01
1.55348241e-01 5.84108114e-01 -5.37674069e-01 2.99046934e-01
1.88265759e-02 9.83099163e-01 -6.00297332e-01 2.83820778e-01
6.90939546e-01 -4.16840583e-01 4.96480316e-01 -3.55461329e-01
2.29451835e-01 2.12549239e-01 -1.37872100e+00 1.57272124e+00
-8.43951166e-01 1.50014862e-01 1.46770209e-01 -9.67465460e-01
8.34495425e-01 6.36040092e-01 3.77406001e-01 -9.18045759e-01
2.60027885e-01 1.52616903e-01 2.18232602e-01 -4.19708900e-02
-2.61574358e-01 1.47691950e-01 -2.87600845e-01 6.02999330e-01
6.21803254e-02 1.71205118e-01 -4.05432791e-01 2.23728910e-01
1.15225410e+00 -1.23816393e-01 5.60018778e-01 1.17537640e-01
8.70029569e-01 -9.75527227e-01 8.24562013e-01 9.29207325e-01
2.50433862e-01 3.27899903e-01 5.12413859e-01 1.60775840e-01
-8.24335217e-01 -8.60305727e-01 1.65255964e-01 2.96809971e-01
2.46894062e-01 -3.11553359e-01 -7.61670172e-01 -7.59222627e-01
-4.04655814e-01 5.23584723e-01 -3.30354214e-01 -4.82738525e-01
-3.58986676e-01 -8.55144680e-01 7.75669992e-01 4.82441187e-01
1.36706281e+00 -6.40962660e-01 -3.97928385e-03 4.98779833e-01
-7.90038556e-02 -1.11991155e+00 -5.57654619e-01 2.36190856e-01
-4.17912066e-01 -7.57327914e-01 -7.63607502e-01 -6.00123823e-01
6.92318201e-01 1.44730344e-01 2.61637866e-01 4.98261210e-03
4.92844790e-01 2.27059573e-01 -2.76103854e-01 -6.32593334e-02
-5.28065860e-01 3.37101519e-01 -2.30861753e-02 4.89743710e-01
-1.25704303e-01 -9.51982498e-01 -7.76731431e-01 3.93773288e-01
-7.27698863e-01 1.09938085e-01 1.12276697e+00 9.62544084e-01
4.16465551e-01 2.15636820e-01 9.39835429e-01 -6.90341830e-01
6.03299677e-01 -6.75001860e-01 -6.31945968e-01 1.67924017e-01
-6.26068592e-01 2.62623459e-01 9.36177254e-01 -5.06670356e-01
-8.87571871e-01 2.02516571e-01 -6.35546386e-01 -1.47523820e-01
4.70760942e-01 3.72599483e-01 -8.97801697e-01 -4.59865540e-01
4.19881403e-01 5.86948156e-01 -2.91752040e-01 -9.71780643e-02
3.25051308e-01 1.00567758e+00 4.69087869e-01 -1.52881935e-01
1.19331729e+00 3.13684732e-01 1.72288105e-01 -6.92475915e-01
-7.01215386e-01 1.40442904e-02 -1.95173562e-01 -2.41252631e-02
5.38491368e-01 -1.41725659e+00 -6.19688809e-01 5.13219476e-01
-8.20454478e-01 -3.19267422e-01 -1.38931200e-01 7.04305947e-01
-3.97298694e-01 3.84769768e-01 -3.52384597e-01 -6.62702084e-01
-7.27593780e-01 -1.02787817e+00 6.59286737e-01 2.56061792e-01
3.69133800e-01 -7.44940817e-01 -1.49774775e-01 2.44577732e-02
5.89645684e-01 3.27422619e-01 6.73299253e-01 -2.69937754e-01
-8.20126176e-01 -4.04627711e-01 -4.25185002e-02 5.12318671e-01
1.32332638e-01 -7.98494458e-01 -8.84436429e-01 -6.28833055e-01
4.29685175e-01 -1.27005717e-03 4.94531810e-01 2.38940910e-01
1.31306958e+00 -8.46374094e-01 -3.16701651e-01 1.00000668e+00
1.29844677e+00 2.86370873e-01 1.11282933e+00 -1.17919520e-01
8.02653432e-01 -3.97580475e-01 7.52722770e-02 6.58891499e-01
3.71485859e-01 5.19449234e-01 4.19540167e-01 -6.28316253e-02
-1.00777112e-01 -6.73094034e-01 3.20416659e-01 9.53077972e-01
1.40177086e-01 -3.76291037e-01 -2.78842092e-01 -1.49186194e-01
-1.71549332e+00 -8.87302101e-01 1.52529731e-01 2.51407862e+00
6.72173738e-01 3.88858140e-01 -1.27616093e-01 4.87306684e-01
5.84262192e-01 1.82239428e-01 -7.53394723e-01 3.66891831e-01
3.66137654e-04 2.53254592e-01 8.72180104e-01 3.20867479e-01
-9.54713166e-01 3.93744230e-01 5.56075048e+00 9.00763929e-01
-1.31301856e+00 4.12004948e-01 4.54692870e-01 3.22860450e-01
-2.63543785e-01 -2.97789183e-02 -5.04378796e-01 8.61098826e-01
8.80281806e-01 1.81595266e-01 5.92825949e-01 6.37663305e-01
2.95109570e-01 1.66566506e-01 -7.12178826e-01 1.22745085e+00
-8.02782550e-02 -1.00928915e+00 -8.98927972e-02 1.74062505e-01
7.64734507e-01 -2.60415494e-01 -5.18850721e-02 5.71180522e-01
1.83292672e-01 -8.30676377e-01 5.66916049e-01 5.93171656e-01
1.10119534e+00 -7.39187777e-01 7.72621214e-01 5.55839479e-01
-1.32649016e+00 -3.65936637e-01 -1.80188000e-01 1.10824622e-01
1.56836063e-01 7.34963357e-01 -8.17332327e-01 6.33773685e-01
-8.06480087e-03 4.44592953e-01 -4.34044898e-01 1.13300288e+00
-8.64148557e-01 1.03156483e+00 -3.53160053e-01 -7.28029013e-02
1.02862701e-01 -1.80749655e-01 4.07347679e-01 9.19203401e-01
6.94211781e-01 -3.74125950e-02 2.33621255e-01 5.31249046e-01
-4.19398606e-01 -4.67043445e-02 -4.03798312e-01 2.07679257e-01
8.12857509e-01 1.17852557e+00 -1.79038376e-01 -1.91315711e-01
-5.26557982e-01 1.06920815e+00 3.46808252e-03 5.42625070e-01
-1.01259840e+00 -4.70705986e-01 1.07277267e-01 2.32132971e-02
4.31051195e-01 -3.70505363e-01 -5.01657315e-02 -1.23363066e+00
7.61474967e-02 -9.00909364e-01 8.33391994e-02 -5.78747272e-01
-8.02351296e-01 2.95728922e-01 -4.94174272e-01 -1.61615407e+00
-4.43542212e-01 5.10368086e-02 -7.04550087e-01 9.34004188e-01
-1.45092320e+00 -1.28871024e+00 -5.08034825e-01 7.88212061e-01
-1.29740074e-01 -2.25603163e-01 8.63665938e-01 4.98321116e-01
-6.69098318e-01 9.15350378e-01 2.00060531e-01 2.60253221e-01
1.79054752e-01 -9.75533247e-01 2.33033940e-01 1.10519457e+00
9.76894572e-02 5.34030795e-01 2.49376729e-01 -5.87309837e-01
-1.49614465e+00 -1.31238365e+00 1.22871771e-01 3.21464330e-01
4.03819233e-01 -6.05756700e-01 -5.04311085e-01 7.20504284e-01
2.38377497e-01 1.36575580e-01 6.26421392e-01 -4.73307610e-01
-9.08616632e-02 -5.40161490e-01 -1.15925860e+00 6.27373517e-01
8.63471031e-01 -5.11830211e-01 1.57500058e-01 5.44934124e-02
6.07578099e-01 -3.86462361e-01 -7.11025715e-01 3.64922345e-01
6.26990080e-01 -5.75904965e-01 7.25526869e-01 8.69051144e-02
-3.46282005e-01 -5.67283273e-01 -1.14035957e-01 -1.26307142e+00
-2.73374915e-01 -1.29123890e+00 -7.90212810e-01 1.14118731e+00
3.32999527e-01 -7.51871049e-01 8.85542214e-01 -2.54273325e-01
6.01256303e-02 -6.48158967e-01 -1.04208291e+00 -6.66187525e-01
-4.34774548e-01 -2.86745220e-01 8.42126667e-01 5.89297235e-01
-2.32690647e-01 7.45885909e-01 -9.50211883e-01 5.23526251e-01
5.36189497e-01 -4.28996027e-01 9.99253690e-01 -7.57030249e-01
-7.55030513e-01 -3.74640338e-02 -3.91062826e-01 -1.69911909e+00
-2.14216396e-01 -7.28001595e-01 -1.95927992e-01 -1.41362214e+00
-2.38429293e-01 -6.11866593e-01 -4.86995220e-01 2.68815815e-01
-1.16678037e-01 1.88605949e-01 -7.31279477e-02 1.12200052e-01
-4.69175994e-01 7.83778012e-01 1.22315907e+00 -1.71165347e-01
-2.66347349e-01 6.48409426e-01 -9.63106096e-01 2.63457179e-01
1.06595385e+00 -2.05662340e-01 -6.10819280e-01 -4.07207198e-02
1.58744544e-01 3.83741409e-01 2.84146488e-01 -1.80309522e+00
2.54298180e-01 3.30866307e-01 5.72783172e-01 -2.16889933e-01
3.32294643e-01 -1.15089440e+00 4.32229340e-01 7.59433448e-01
-2.16639899e-02 -2.91180909e-01 -3.00233573e-01 1.00527585e+00
1.04297154e-01 -4.93798479e-02 7.10576475e-01 2.77465254e-01
-2.37244561e-01 5.88269591e-01 -3.32246512e-01 -1.47885174e-01
9.44604933e-01 5.79632297e-02 2.29230579e-02 -1.02598095e+00
-3.99855167e-01 2.94796646e-01 1.69002488e-02 6.54025143e-03
4.97675270e-01 -1.60519111e+00 -4.42953557e-01 5.47579288e-01
-1.38094783e-01 -1.12629719e-01 2.36472517e-01 8.82724106e-01
-2.08453700e-01 3.82011384e-01 1.27672806e-01 -2.13215962e-01
-5.73805869e-01 2.96301335e-01 4.33177590e-01 -4.77035850e-01
-4.10408080e-01 2.73924917e-01 -2.02103496e-01 -2.75650173e-01
4.43973571e-01 -7.31412172e-02 -1.24006413e-01 -3.75058860e-01
4.60918009e-01 1.53341100e-01 -3.29315066e-02 -2.22295150e-01
6.88384548e-02 1.81952134e-01 2.00194642e-01 -1.82927817e-01
8.59178245e-01 -2.70588368e-01 2.74823904e-01 -2.02709120e-02
1.11476195e+00 3.89969647e-01 -1.42699718e+00 -3.73861670e-01
-4.95525569e-01 -3.62069830e-02 2.32288256e-01 -7.27182627e-01
-1.30258238e+00 7.05908775e-01 7.90892839e-01 1.57810241e-01
1.28594637e+00 -5.24341702e-01 9.50312078e-01 4.74773675e-01
6.13138318e-01 -7.26958156e-01 1.49137020e-01 1.61576882e-01
4.26838517e-01 -7.92751789e-01 -3.31454873e-02 -2.44776443e-01
-2.95310527e-01 8.50304544e-01 4.50582355e-01 1.13795333e-01
7.72286236e-01 2.39796147e-01 -1.67049974e-01 2.15580776e-01
-2.14227736e-01 -1.43446863e-01 -1.32321328e-01 7.79031813e-01
-1.27328038e-01 -1.23216817e-02 -3.97408098e-01 8.90841007e-01
-1.88707441e-01 -1.15847280e-02 6.46611810e-01 8.57547879e-01
-2.84040302e-01 -1.10847640e+00 -1.91620648e-01 6.24950171e-01
-4.38731611e-01 -6.51293248e-02 1.10600010e-01 4.24136192e-01
1.01876296e-01 9.02050912e-01 -3.16183746e-01 -7.63460100e-01
1.18966304e-01 -2.67782032e-01 2.96617150e-01 -3.47263455e-01
-1.17223121e-01 1.09187849e-01 -5.97404428e-02 -3.07578981e-01
-2.58040398e-01 -3.08396518e-01 -1.08422303e+00 -1.09037738e-02
-6.47781491e-01 2.26250201e-01 5.70018828e-01 8.16656530e-01
3.60876948e-01 8.30057740e-01 1.34498847e+00 -4.42034990e-01
-7.10622370e-01 -9.81493592e-01 -5.53200543e-01 -1.58014879e-01
2.84958363e-01 -3.37067842e-01 -2.67083526e-01 -3.61567885e-01] | [6.3975067138671875, 1.5150320529937744] |
aa333897-3f1a-4ec3-ae7a-fc35f97d5fd7 | coordinate-based-texture-inpainting-for-pose | 1811.11459 | null | https://arxiv.org/abs/1811.11459v2 | https://arxiv.org/pdf/1811.11459v2.pdf | Coordinate-based Texture Inpainting for Pose-Guided Image Generation | We present a new deep learning approach to pose-guided resynthesis of human photographs. At the heart of the new approach is the estimation of the complete body surface texture based on a single photograph. Since the input photograph always observes only a part of the surface, we suggest a new inpainting method that completes the texture of the human body. Rather than working directly with colors of texture elements, the inpainting network estimates an appropriate source location in the input image for each element of the body surface. This correspondence field between the input image and the texture is then further warped into the target image coordinate frame based on the desired pose, effectively establishing the correspondence between the source and the target view even when the pose change is drastic. The final convolutional network then uses the established correspondence and all other available information to synthesize the output image. A fully-convolutional architecture with deformable skip connections guided by the estimated correspondence field is used. We show state-of-the-art result for pose-guided image synthesis. Additionally, we demonstrate the performance of our system for garment transfer and pose-guided face resynthesis. | ['Victor Lempitsky', 'Alexander Vakhitov', 'Artur Grigorev', 'Artem Sevastopolsky'] | 2018-11-28 | null | null | null | null | ['pose-guided-image-generation'] | ['computer-vision'] | [ 6.39591455e-01 5.44061065e-01 1.69461608e-01 -3.94547313e-01
-6.38838887e-01 -3.11855406e-01 2.48170257e-01 -6.09327495e-01
-9.40704793e-02 5.69852769e-01 1.12583555e-01 5.39921641e-01
4.99224156e-01 -9.24188673e-01 -1.26898074e+00 -5.40263116e-01
4.93657827e-01 6.26673222e-01 2.95520313e-02 -2.52602041e-01
-1.45644560e-01 6.76562965e-01 -1.49242032e+00 4.25136328e-01
2.20393494e-01 1.11614871e+00 -7.25205317e-02 7.75015175e-01
1.25361145e-01 2.88901061e-01 -6.56907499e-01 -5.92531860e-01
5.76932907e-01 -8.27123344e-01 -8.16757679e-01 7.33706594e-01
1.09084713e+00 -8.00819039e-01 -3.78007740e-01 7.61362195e-01
2.44605258e-01 3.20937857e-02 4.06717360e-01 -8.26062739e-01
-5.20043552e-01 1.39173731e-01 -7.13717282e-01 -4.50144142e-01
6.24610424e-01 -1.51749909e-01 5.26361704e-01 -1.03042901e+00
1.10728955e+00 1.33062756e+00 6.36555672e-01 9.29785132e-01
-1.14812601e+00 -4.92771000e-01 6.04512878e-02 -1.76994398e-01
-1.22220600e+00 -4.73398596e-01 1.15317094e+00 -2.52597481e-01
4.00657892e-01 2.50091910e-01 1.01882327e+00 8.63845944e-01
2.90225953e-01 3.35224330e-01 1.06687987e+00 -6.97850943e-01
-1.08883679e-02 -2.94335276e-01 -6.49319649e-01 1.10726249e+00
-1.88508213e-01 8.95497724e-02 -7.64468849e-01 2.48621672e-01
1.39211547e+00 6.12064227e-02 -2.75328547e-01 -4.26106036e-01
-1.03567338e+00 2.81035364e-01 4.77342397e-01 -1.12828180e-01
-2.82710731e-01 4.35198992e-01 -6.00712001e-02 2.05880940e-01
7.08200634e-01 1.36956379e-01 -1.33285031e-01 2.95144349e-01
-1.07220602e+00 1.75543755e-01 6.41465127e-01 7.11378455e-01
9.32926297e-01 1.56215250e-01 7.65040740e-02 6.99906051e-01
1.92926005e-01 5.32665730e-01 -1.33166626e-01 -1.20757461e+00
3.78896326e-01 6.05193257e-01 2.05927938e-01 -9.18033183e-01
-1.95377082e-01 -6.45360947e-02 -6.04461670e-01 6.34945095e-01
7.60138154e-01 -2.41571203e-01 -1.21681738e+00 1.59471703e+00
7.49982357e-01 9.24646631e-02 -2.35318914e-01 1.13180661e+00
5.91621339e-01 5.12765884e-01 -2.88710624e-01 1.32314786e-01
1.34641802e+00 -9.01131034e-01 -6.39334917e-01 -5.16752243e-01
-9.87548828e-02 -8.21222782e-01 8.89908552e-01 2.33846292e-01
-1.34465170e+00 -6.63464427e-01 -1.04906952e+00 -4.46970433e-01
8.57781768e-02 4.49747473e-01 1.56674996e-01 2.77958125e-01
-1.25356102e+00 8.45614254e-01 -8.56419444e-01 -3.27916294e-01
1.88864559e-01 4.76599783e-01 -7.12118924e-01 -1.11747809e-01
-9.70501781e-01 7.43236184e-01 -8.48686770e-02 4.55157787e-01
-7.42414653e-01 -4.84331995e-01 -9.36871409e-01 -1.06877618e-01
2.88266748e-01 -9.52023268e-01 1.15778637e+00 -1.62589347e+00
-1.96019804e+00 1.22574294e+00 -7.77087137e-02 -4.38879468e-02
9.10381973e-01 -3.32897872e-01 -4.54567224e-02 4.89787817e-01
-8.63157064e-02 9.58877921e-01 1.47258985e+00 -1.39155281e+00
-3.61959487e-01 -3.86682421e-01 -3.06357536e-03 4.44999188e-01
8.53400975e-02 -1.60645723e-01 -9.04012442e-01 -8.79617214e-01
1.84109196e-01 -1.00920141e+00 2.23627180e-01 8.36307108e-01
-4.39210206e-01 2.59994715e-01 7.85139859e-01 -1.17560613e+00
6.72262132e-01 -2.04599452e+00 5.55885971e-01 2.57102907e-01
1.24652073e-01 -1.12444282e-01 -1.33095562e-01 3.27769309e-01
-2.91128576e-01 -4.36193407e-01 -3.95901144e-01 -7.37745166e-01
-3.91730845e-01 3.22349630e-02 -3.01205933e-01 6.95578694e-01
1.16176389e-01 8.59511137e-01 -6.48458362e-01 -3.67041737e-01
4.04171169e-01 6.52486086e-01 -5.02000451e-01 4.74031568e-01
-3.37780714e-01 7.78779447e-01 -9.79408175e-02 5.74634492e-01
7.30946183e-01 4.58614081e-02 3.29636306e-01 -5.54217935e-01
1.50885254e-01 -1.00984998e-01 -1.11435664e+00 2.01746774e+00
-5.15887260e-01 4.93861854e-01 1.16411529e-01 -4.75383848e-01
9.03959692e-01 3.18882704e-01 5.30879557e-01 -3.90630573e-01
1.04295820e-01 9.59965214e-03 -2.94445246e-01 -2.39492804e-01
2.07743034e-01 -3.08410823e-01 1.25429511e-01 5.68580151e-01
5.40118925e-02 -3.70880157e-01 -3.45207065e-01 -2.79255480e-01
3.90203506e-01 7.94975281e-01 -2.80119851e-02 4.19277176e-02
3.83632809e-01 -2.07184210e-01 1.56011358e-01 -3.08792628e-02
3.55075389e-01 1.25435114e+00 4.45688039e-01 -9.00793791e-01
-1.33002567e+00 -1.10448778e+00 2.55287230e-01 8.20272684e-01
8.16119239e-02 -2.32608110e-01 -1.39273262e+00 -5.81124008e-01
-1.93595782e-01 1.70393273e-01 -1.20288432e+00 -1.41829804e-01
-8.82540941e-01 1.63278077e-02 1.99836835e-01 4.89729226e-01
8.52396846e-01 -1.10051441e+00 -6.27869606e-01 8.65707397e-02
-3.66226256e-01 -9.30745959e-01 -9.37125981e-01 -4.73570228e-01
-7.85054564e-01 -1.12282670e+00 -9.05616581e-01 -8.94789159e-01
1.24876988e+00 -1.12180494e-01 8.28407347e-01 1.73905790e-01
-4.43776369e-01 2.83743024e-01 9.49727744e-02 6.90194890e-02
-4.77371275e-01 -3.76219481e-01 -1.49044573e-01 5.89618087e-01
-4.32386160e-01 -4.18728620e-01 -9.39385355e-01 2.49829724e-01
-8.28871310e-01 4.62354094e-01 1.95847720e-01 8.25805366e-01
6.73014402e-01 -2.72304893e-01 -4.25266176e-02 -6.27723873e-01
-3.02943727e-03 2.14083135e-01 -5.21423876e-01 2.59607494e-01
1.80204213e-01 5.69859073e-02 3.86899978e-01 -6.17280543e-01
-1.19958925e+00 5.26097953e-01 -2.35879302e-01 -7.43526518e-01
9.13093686e-02 -9.57272723e-02 -3.14755589e-01 -6.47568703e-02
5.53965807e-01 8.26591924e-02 2.38177627e-01 -5.74744701e-01
4.61009979e-01 2.78124213e-01 8.75883400e-01 -4.35549676e-01
8.28085303e-01 9.33953643e-01 1.33863181e-01 -6.77701712e-01
-7.57728040e-01 2.76257187e-01 -1.02474070e+00 -5.61383188e-01
1.10477877e+00 -7.85774291e-01 -6.80892408e-01 7.81526506e-01
-1.31752622e+00 -6.14372671e-01 -5.09400308e-01 1.51440591e-01
-7.25677490e-01 2.07000971e-01 -6.22683525e-01 -3.67733359e-01
-4.51047182e-01 -1.00720084e+00 1.65853989e+00 -1.30047230e-02
-3.58862817e-01 -7.79435694e-01 4.45635850e-03 4.52054232e-01
-1.19666807e-01 5.93051791e-01 8.36963773e-01 2.50461906e-01
-6.45062625e-01 -2.94926345e-01 7.49029219e-02 4.42937195e-01
3.44643325e-01 3.35861072e-02 -1.05646408e+00 -3.63326490e-01
-4.05098200e-02 -2.70051032e-01 5.04118562e-01 3.37091476e-01
1.00675392e+00 -3.63434583e-01 -1.59001887e-01 8.70829105e-01
1.28029811e+00 -2.04880554e-02 8.05481791e-01 9.91193801e-02
1.01235056e+00 8.89366686e-01 4.39712822e-01 2.22766161e-01
1.87842384e-01 9.17141914e-01 3.29574227e-01 -5.29121578e-01
-7.43185639e-01 -6.80117249e-01 3.68422627e-01 3.25089991e-01
-3.43446553e-01 1.09323718e-01 -5.49838901e-01 2.11161956e-01
-1.33329272e+00 -6.04783595e-01 3.13750118e-01 2.43199396e+00
8.90654027e-01 -2.10596487e-01 2.03314796e-02 -8.01692903e-02
7.44481325e-01 -3.55678126e-02 -5.55768728e-01 -3.91660333e-01
2.54962802e-01 4.18966621e-01 3.46645385e-01 8.89787614e-01
-8.57112229e-01 1.06868243e+00 6.57903242e+00 5.69605350e-01
-1.37474287e+00 -1.41779169e-01 7.19937325e-01 -1.07501410e-01
-1.18446529e-01 -2.57685721e-01 -6.07148707e-01 2.12764114e-01
4.20449585e-01 2.93825030e-01 6.03216052e-01 6.26411617e-01
-1.18160076e-01 -2.17023641e-01 -1.33963072e+00 1.02202833e+00
4.76387680e-01 -1.29764414e+00 1.43527135e-01 1.55557152e-02
8.76090288e-01 -6.49613917e-01 1.14091046e-01 -3.74858350e-01
-6.42846301e-02 -9.78485823e-01 1.06812930e+00 7.56329656e-01
1.63278830e+00 -7.65683949e-01 2.29055822e-01 -7.42698908e-02
-1.05438232e+00 2.52994865e-01 -1.95560277e-01 -1.92369781e-02
4.80399951e-02 2.38051891e-01 -7.62478471e-01 2.87517488e-01
6.49838328e-01 5.76465964e-01 -4.29969043e-01 4.39232051e-01
-4.67475384e-01 6.35547787e-02 -2.49740109e-01 7.11475134e-01
-3.70740771e-01 -2.21186042e-01 4.02377129e-01 6.50070190e-01
3.33347350e-01 2.73536835e-02 -1.49843143e-02 8.63432944e-01
-4.53412473e-01 -9.11306739e-02 -6.07718110e-01 2.81766444e-01
7.50739351e-02 1.21011055e+00 -8.56013656e-01 -4.46456254e-01
-2.59998620e-01 1.61798823e+00 3.58428478e-01 3.87623370e-01
-6.83017254e-01 -6.95114955e-02 5.08980870e-01 5.44819474e-01
2.55125850e-01 -1.11424267e-01 -4.65795137e-02 -9.80469406e-01
1.55270919e-01 -8.64935219e-01 7.76865855e-02 -1.14955676e+00
-7.97705889e-01 7.91273654e-01 4.74964618e-04 -1.32742739e+00
-3.58888149e-01 -4.17787939e-01 -4.62099701e-01 1.06026781e+00
-9.42861497e-01 -1.50860691e+00 -4.70769286e-01 6.54454887e-01
6.72588229e-01 2.14181751e-01 8.97196114e-01 -3.60313319e-02
-1.58641890e-01 5.51160634e-01 -2.65642703e-01 2.51561671e-01
8.72521162e-01 -9.84617651e-01 8.65179539e-01 7.28996456e-01
-1.50735274e-01 3.62869322e-01 4.23377454e-01 -9.73488867e-01
-1.40540528e+00 -1.10136914e+00 6.87688529e-01 -4.79175806e-01
8.90929997e-02 -6.94219172e-01 -6.24937117e-01 8.64236355e-01
7.78671876e-02 1.24714859e-01 7.46423304e-02 -5.11348426e-01
-2.43437171e-01 -3.85197937e-01 -1.20100331e+00 7.87880957e-01
9.79393899e-01 -4.90199953e-01 -5.12593746e-01 2.67445952e-01
4.28565502e-01 -1.12229621e+00 -7.28670239e-01 8.38183835e-02
9.89208817e-01 -7.13600099e-01 1.05367815e+00 -3.02964807e-01
7.90212572e-01 -3.79823059e-01 8.51788297e-02 -1.24486625e+00
-5.64088896e-02 -7.51112998e-01 1.63048372e-01 8.81521463e-01
7.22470731e-02 -2.45406851e-01 1.06821406e+00 7.09878564e-01
3.77768278e-02 -6.87176347e-01 -9.56835389e-01 -1.54089764e-01
-8.44656155e-02 1.19365625e-01 5.77273726e-01 6.38240635e-01
-3.11817527e-01 -8.20535235e-03 -8.39266241e-01 7.87188783e-02
6.04775369e-01 4.01191622e-01 9.44130301e-01 -7.85958052e-01
-4.24672842e-01 2.99544424e-01 -3.03253233e-01 -9.15641725e-01
7.27937445e-02 -5.90108395e-01 1.35538727e-01 -1.44984627e+00
-5.07615954e-02 2.27333084e-02 4.86327827e-01 6.59373879e-01
-5.36492746e-03 9.04160738e-01 2.35536203e-01 9.72204730e-02
1.66729733e-01 2.25263089e-01 1.85890996e+00 8.31445307e-02
-2.08454564e-01 -1.10754013e-01 -3.01886499e-01 8.27823758e-01
5.60679257e-01 -2.04390123e-01 -2.52415240e-01 -5.30129433e-01
4.79008071e-02 4.26182717e-01 5.50773203e-01 -9.15074050e-01
-1.07211709e-01 -1.40807450e-01 1.10176957e+00 -2.38525867e-01
8.80030274e-01 -1.01535761e+00 6.79619193e-01 5.00583827e-01
-3.63781333e-01 3.12910341e-02 1.17467463e-01 3.86256188e-01
8.44337195e-02 2.08517194e-01 1.00910270e+00 -2.14152277e-01
-2.67596662e-01 4.22262937e-01 9.37200412e-02 -2.74906814e-01
8.81159961e-01 -5.30614793e-01 7.31043443e-02 -5.12112498e-01
-1.15198243e+00 -4.04355615e-01 9.36674953e-01 3.73916805e-01
9.33492362e-01 -1.39942670e+00 -6.62988722e-01 8.01297486e-01
-2.53024191e-01 8.45371485e-02 4.24374849e-01 4.86041784e-01
-9.53272521e-01 -1.56578466e-01 -6.19567454e-01 -5.25847077e-01
-1.41405022e+00 3.35311979e-01 8.31253350e-01 1.66943207e-01
-9.66089308e-01 6.92686439e-01 6.28424525e-01 -1.43437117e-01
1.21084362e-01 -3.09042394e-01 2.70545006e-01 -2.20175654e-01
5.50801694e-01 2.32604176e-01 1.25717744e-01 -8.42222989e-01
-1.48713693e-01 1.15294456e+00 1.72978491e-01 -5.85411251e-01
1.15504766e+00 -3.09233535e-02 -2.42734760e-01 2.30205670e-01
1.25467670e+00 1.80436015e-01 -1.87203515e+00 -8.15552324e-02
-8.11776519e-01 -7.17926860e-01 -3.59335244e-01 -8.27382922e-01
-1.38441694e+00 8.58506441e-01 5.46805859e-01 -6.40452266e-01
1.18136859e+00 -1.51897445e-01 7.23003149e-01 -8.41016024e-02
4.50286418e-01 -1.01460028e+00 3.20482612e-01 1.47014912e-02
1.37877178e+00 -7.31803298e-01 1.45907357e-01 -6.37426257e-01
-5.15376866e-01 1.40067828e+00 6.78956568e-01 -4.84011143e-01
4.81756091e-01 3.22950304e-01 2.72158742e-01 -1.91799715e-01
-2.12373406e-01 2.33352691e-01 5.95825613e-01 6.65057719e-01
2.66185224e-01 -1.00396149e-01 -1.40852053e-02 -1.18326426e-01
-4.36364770e-01 3.62166837e-02 3.59352797e-01 7.98164308e-01
-1.42604813e-01 -1.14355302e+00 -7.07960129e-01 7.70141110e-02
-3.97148192e-01 8.61958712e-02 -6.92987621e-01 8.24791849e-01
2.37435743e-01 5.20648718e-01 2.83568233e-01 -2.23556355e-01
3.99843693e-01 5.85681014e-02 1.06760240e+00 -7.86912203e-01
-3.44209462e-01 4.66736376e-01 1.49810269e-01 -7.76635826e-01
-2.48166606e-01 -3.80742162e-01 -1.17320061e+00 -2.30858400e-01
1.50169909e-01 -3.50197792e-01 6.17610335e-01 7.87301421e-01
1.22989550e-01 4.03013080e-01 5.40524304e-01 -1.36927307e+00
-8.21299851e-02 -7.48217583e-01 -5.57568014e-01 5.69212258e-01
5.70837080e-01 -5.66654980e-01 -4.04584184e-02 5.96517861e-01] | [11.753835678100586, -0.8038618564605713] |
796b9df1-ab9c-452d-9ee0-a35f68ca2503 | clam-selective-clarification-for-ambiguous | 2212.07769 | null | https://arxiv.org/abs/2212.07769v2 | https://arxiv.org/pdf/2212.07769v2.pdf | CLAM: Selective Clarification for Ambiguous Questions with Generative Language Models | Users often ask dialogue systems ambiguous questions that require clarification. We show that current language models rarely ask users to clarify ambiguous questions and instead provide incorrect answers. To address this, we introduce CLAM: a framework for getting language models to selectively ask for clarification about ambiguous user questions. In particular, we show that we can prompt language models to detect whether a given question is ambiguous, generate an appropriate clarifying question to ask the user, and give a final answer after receiving clarification. We also show that we can simulate users by providing language models with privileged information. This lets us automatically evaluate multi-turn clarification dialogues. Finally, CLAM significantly improves language models' accuracy on mixed ambiguous and unambiguous questions relative to SotA. | ['Sebastian Farquhar', 'Yarin Gal', 'Lorenz Kuhn'] | 2022-12-15 | null | null | null | null | ['triviaqa'] | ['miscellaneous'] | [ 3.67109388e-01 7.26885259e-01 2.43502948e-02 -7.30243862e-01
-1.13333094e+00 -1.25530183e+00 6.62764549e-01 2.08251402e-01
-3.20768505e-01 1.09355354e+00 6.20415270e-01 -1.16517377e+00
-4.32082117e-02 -4.11356419e-01 -1.77606419e-02 3.14775527e-01
7.63098240e-01 7.50295997e-01 2.62019247e-01 -6.87544882e-01
4.67753768e-01 2.06618294e-01 -1.00046122e+00 8.18375528e-01
1.36882579e+00 -2.27711070e-02 2.24958345e-01 1.33484888e+00
-7.97809303e-01 1.08345616e+00 -8.62246990e-01 -3.41775507e-01
-2.82451540e-01 -8.09350431e-01 -1.74258661e+00 1.19536623e-01
6.80909991e-01 -6.59228146e-01 2.91402459e-01 5.97565472e-01
9.03129727e-02 8.67122188e-02 5.62624991e-01 -1.37321615e+00
-4.72650319e-01 7.73402274e-01 3.38601202e-01 1.73058689e-01
1.10907960e+00 4.21327025e-01 1.19162977e+00 -7.36437976e-01
3.14331979e-01 1.55364478e+00 5.38839340e-01 1.14388180e+00
-1.25216520e+00 -1.54962689e-01 3.80274117e-01 -8.95829126e-02
-5.40207267e-01 -6.05216920e-01 3.58967632e-01 -3.52675438e-01
8.66589487e-01 1.07369328e+00 3.24052066e-01 8.12027812e-01
-1.22481421e-01 7.78134346e-01 1.21070421e+00 -5.80531776e-01
-1.08880505e-01 4.66130197e-01 1.00769520e+00 8.32043886e-01
-1.84923843e-01 -3.17378789e-01 -1.47975445e-01 -5.28645694e-01
4.61445749e-01 -4.82647240e-01 -3.36415052e-01 2.55679846e-01
-8.46903265e-01 8.24580967e-01 -3.06986958e-01 4.84553277e-01
-1.28972903e-01 -2.25239500e-01 9.99541059e-02 6.78441584e-01
-7.31247589e-02 1.35187888e+00 -8.34868670e-01 -4.98224705e-01
-3.39183301e-01 5.65924525e-01 1.59511375e+00 7.45214641e-01
6.57741725e-01 -3.58265191e-01 -5.06781578e-01 9.86949921e-01
4.47302252e-01 3.55293423e-01 1.80597633e-01 -1.71097386e+00
3.12573731e-01 7.23477840e-01 7.65717208e-01 -4.91349727e-01
-4.28728998e-01 -1.09954411e-02 -1.06782384e-01 -7.48173818e-02
8.59800339e-01 -5.77153742e-01 -1.16498999e-01 1.73536122e+00
1.26788542e-01 -4.85984206e-01 3.33168507e-01 6.55943811e-01
8.75680327e-01 6.19562626e-01 2.53418535e-01 -2.14103028e-01
1.59488738e+00 -7.79610574e-01 -9.84204769e-01 -3.11245561e-01
1.10409713e+00 -9.95327175e-01 1.65182817e+00 2.15440929e-01
-1.22727454e+00 -4.17058855e-01 -4.57375884e-01 -3.85380983e-01
7.74962362e-04 2.07362458e-01 3.11990380e-01 7.39682734e-01
-1.15754712e+00 -7.75482878e-02 -2.19001368e-01 -2.73115098e-01
-5.06405175e-01 6.08707964e-02 1.80535272e-01 1.78919762e-01
-1.43454754e+00 1.11933291e+00 -6.84032440e-02 -4.64591533e-02
-2.79567122e-01 -6.37326777e-01 -1.06279409e+00 7.98328519e-02
4.13517505e-01 -8.35113704e-01 2.47451973e+00 -7.53036976e-01
-1.63227236e+00 6.35200322e-01 -6.28546536e-01 -3.88485402e-01
5.07342398e-01 -6.09962940e-01 -1.04193553e-01 1.21083960e-01
3.81040424e-01 7.62116432e-01 2.39193663e-01 -1.38775718e+00
-7.69919336e-01 -1.80001274e-01 8.95118892e-01 4.93255854e-01
2.92382270e-01 2.78494865e-01 1.11468427e-01 3.01624602e-03
1.86355934e-02 -8.11475277e-01 -1.54477239e-01 -2.33185217e-01
-6.31802857e-01 -5.57365298e-01 7.87991345e-01 -7.98902333e-01
1.53524923e+00 -1.46760762e+00 -3.19112211e-01 -1.87768303e-02
3.98940146e-01 5.18091798e-01 -3.40467989e-01 6.05654240e-01
-2.19180770e-02 5.30780971e-01 -1.54945418e-01 -1.04641058e-01
4.15091008e-01 3.01027596e-01 -4.93571877e-01 -4.54414666e-01
8.25090036e-02 1.03010631e+00 -8.10202539e-01 -4.29544806e-01
2.47705355e-01 -1.67034581e-01 -7.04106867e-01 8.22496295e-01
-6.99548125e-01 2.32178092e-01 -3.91237229e-01 2.10940942e-01
4.66309845e-01 -2.36593276e-01 3.47948045e-01 2.67148465e-01
-7.13389143e-02 1.14024627e+00 -1.04975128e+00 8.75872374e-01
-9.29443955e-01 6.93497241e-01 6.41211331e-01 -3.84718627e-01
6.19243145e-01 4.42795843e-01 -2.86564469e-01 -2.50239938e-01
-1.72400698e-01 2.34426126e-01 4.03776653e-02 -6.65024281e-01
6.42907619e-01 -1.95702240e-01 -2.05390528e-01 1.09015298e+00
-3.79431009e-01 -5.89483082e-01 9.09757912e-02 6.23069704e-01
7.08428383e-01 -1.81310907e-01 3.27991575e-01 -2.84156740e-01
1.10101151e+00 -2.00744364e-02 -2.13026002e-01 1.25886118e+00
-1.45378768e-01 1.09350979e-01 6.43489838e-01 -2.08163604e-01
-5.81653774e-01 -9.36978161e-01 4.48367774e-01 1.52699399e+00
-1.42612100e-01 -6.41939163e-01 -1.19955218e+00 -1.01301098e+00
-2.77502865e-01 1.46850550e+00 6.45648018e-02 1.28646880e-01
-7.66907513e-01 2.11948380e-02 6.00906372e-01 1.51104197e-01
5.52010000e-01 -1.07067192e+00 -5.77611744e-01 3.56131464e-01
-8.74904811e-01 -8.38755548e-01 -9.24878359e-01 -2.10187629e-01
-6.74441516e-01 -1.46033978e+00 -2.58351088e-01 -9.09923971e-01
5.10507286e-01 3.70491624e-01 1.28172779e+00 9.19105113e-01
4.03252482e-01 1.13963473e+00 -1.91202924e-01 -1.15737699e-01
-1.16088808e+00 4.11866128e-01 -4.03347164e-01 -4.09949064e-01
3.04369837e-01 -1.85536578e-01 -3.14847857e-01 6.33495748e-01
-9.06775117e-01 3.65593970e-01 -4.56346273e-02 6.60544753e-01
-3.54746163e-01 -8.95073533e-01 9.62034702e-01 -1.24785995e+00
1.82686210e+00 -1.70334116e-01 -4.16100681e-01 7.07225919e-01
-6.13539875e-01 5.76580226e-01 7.28770256e-01 -9.12673622e-02
-1.35186267e+00 -2.61031866e-01 -6.79411113e-01 8.50853145e-01
-6.16638958e-01 3.17440003e-01 -7.34760985e-02 1.55167282e-01
9.90928352e-01 1.59129336e-01 2.17226624e-01 -4.74922687e-01
6.12587452e-01 9.12371576e-01 4.24920589e-01 -1.18144810e+00
6.60422504e-01 -4.91308808e-01 -7.64576435e-01 -9.45658505e-01
-9.50561643e-01 -6.33186936e-01 -3.74863863e-01 -2.16646895e-01
5.85574687e-01 -3.10722828e-01 -1.11460674e+00 7.40825608e-02
-1.60048878e+00 -5.54699898e-01 7.75104389e-03 -6.45339116e-02
-5.32887280e-01 8.84006321e-01 -6.22076809e-01 -1.30263901e+00
-3.50703150e-01 -1.03952372e+00 8.77865970e-01 3.27253610e-01
-1.32895958e+00 -1.13681817e+00 -1.48845226e-01 7.75280416e-01
5.15029430e-01 -4.96411175e-01 1.44846570e+00 -1.28684092e+00
-3.15707117e-01 3.12267691e-01 -5.02879452e-03 1.13890134e-01
4.12991166e-01 -4.57751490e-02 -5.66041768e-01 2.05540016e-01
4.44665030e-02 -4.33831006e-01 3.25155444e-02 -8.50920901e-02
7.28831112e-01 -1.01601005e+00 1.78422895e-03 -2.77925044e-01
4.75895643e-01 1.53579831e-01 5.33467174e-01 1.25389500e-02
8.92817080e-02 9.74361002e-01 5.72229028e-01 -2.78239977e-02
8.45599115e-01 4.17364985e-01 -2.10103616e-01 1.06278472e-01
1.67234138e-01 -4.83573705e-01 1.99553534e-01 4.38418329e-01
7.13066518e-01 -2.12220520e-01 -7.94076741e-01 5.14033020e-01
-1.65611935e+00 -8.24441254e-01 -6.14659548e-01 1.77582633e+00
1.49202180e+00 2.58938652e-02 -2.80810823e-03 -9.47505236e-03
5.95448375e-01 9.93650332e-02 -2.45973393e-01 -7.76284158e-01
2.52573818e-01 3.45923156e-01 -3.81871313e-01 2.07943797e+00
-4.05732363e-01 1.13720858e+00 6.74167204e+00 7.64825419e-02
-5.63784122e-01 -1.36323407e-01 5.55737734e-01 5.93530536e-01
-1.07355702e+00 4.11990583e-01 -7.86398172e-01 9.47115943e-02
9.92329240e-01 -2.35741585e-01 4.88229752e-01 4.93589699e-01
4.54866797e-01 -3.84518057e-01 -1.32000268e+00 4.72874790e-01
-1.53053582e-01 -1.17078424e+00 2.56228834e-01 -5.76662540e-01
1.68845922e-01 -9.11037624e-01 -3.07514608e-01 6.79422200e-01
6.28967702e-01 -1.02599585e+00 8.20856914e-02 4.56710905e-01
3.14767569e-01 -4.24454898e-01 4.61003512e-01 1.06230474e+00
-6.48962677e-01 1.04062922e-01 1.67268947e-01 -5.59547722e-01
2.19240516e-01 -3.45829159e-01 -1.76451194e+00 -1.14947855e-01
-2.37593323e-01 -3.50292772e-01 -6.14073575e-01 6.58624530e-01
-6.43651307e-01 6.69201791e-01 -2.64516205e-01 -6.57958329e-01
3.11760634e-01 -1.32475749e-01 4.96043593e-01 1.27636242e+00
6.22319616e-02 6.91854060e-01 2.28968620e-01 7.90747285e-01
1.75738558e-01 1.33868948e-01 -2.74890900e-01 1.48869172e-01
6.67604685e-01 1.05157316e+00 3.71525809e-02 -6.59083128e-01
6.04842827e-02 1.04878271e+00 5.36631532e-02 5.61900496e-01
-2.58598506e-01 -3.10956717e-01 5.46278715e-01 3.11578423e-01
-6.93396389e-01 -2.48326257e-01 -2.72149593e-01 -1.09593725e+00
-1.88640416e-01 -1.55911791e+00 3.22111845e-01 -1.34168363e+00
-9.17168438e-01 3.98479223e-01 -1.21393404e-03 -3.42936009e-01
-8.91581774e-01 -5.72287083e-01 -7.75859416e-01 1.39796352e+00
-1.21532845e+00 -5.25969744e-01 -2.27963567e-01 2.86878437e-01
1.05030763e+00 3.04934710e-01 1.09236562e+00 -2.02140912e-01
6.41987659e-03 4.48834777e-01 -8.20018589e-01 1.39456898e-01
8.61088991e-01 -1.67290199e+00 4.34509486e-01 5.10397911e-01
-3.50393206e-02 1.19814587e+00 1.15991211e+00 -4.49914932e-01
-9.02270496e-01 -4.93490219e-01 1.67766488e+00 -9.44726944e-01
3.09709698e-01 7.16191856e-03 -1.37039936e+00 8.46262872e-01
7.47632265e-01 -8.94876182e-01 1.03975606e+00 1.29920080e-01
-2.16508985e-01 5.04881084e-01 -1.02951336e+00 9.68417108e-01
6.42582417e-01 -9.65783596e-01 -1.44774556e+00 4.50643986e-01
8.99278522e-01 -5.30741751e-01 -3.17112774e-01 -8.41306597e-02
3.96211147e-01 -9.17781949e-01 5.81304193e-01 -1.30261981e+00
6.49022684e-02 -3.15855891e-01 -2.18020827e-02 -1.32404041e+00
1.31171897e-01 -1.27784562e+00 1.41111434e-01 1.03282082e+00
9.50197339e-01 -8.21716368e-01 4.91354734e-01 1.60632598e+00
-4.09975238e-02 -3.17638904e-01 -5.01509130e-01 8.14042427e-03
2.07491204e-01 -3.06683928e-01 7.39731610e-01 8.41980398e-01
6.89743996e-01 1.03522027e+00 -1.57158017e-01 9.69227180e-02
-2.38890145e-02 1.30428821e-01 9.23872292e-01 -1.10497320e+00
-1.17104478e-01 -4.04339552e-01 8.97095740e-01 -1.98282766e+00
2.87450701e-01 -5.02202213e-01 2.52818942e-01 -1.88255465e+00
-3.54174972e-01 -3.28863740e-01 6.65216327e-01 4.80824828e-01
-5.78496039e-01 -5.75279057e-01 1.33169293e-01 -1.96688548e-01
-6.62304163e-01 1.52103931e-01 1.11926389e+00 2.19163373e-02
-4.06436026e-01 5.54940581e-01 -1.44258654e+00 7.32285440e-01
1.00123489e+00 -6.63998649e-02 -5.95907629e-01 -2.10504889e-01
2.30533093e-01 8.22769463e-01 3.42718035e-01 -4.53567117e-01
2.60974050e-01 -6.43337071e-01 -1.30967513e-01 -3.84911925e-01
5.95404357e-02 -5.64354539e-01 -4.42722887e-01 5.89642048e-01
-1.28562653e+00 4.25215125e-01 2.54312873e-01 -1.54913023e-01
-5.44520319e-02 -8.34259510e-01 4.26969677e-01 -3.25200826e-01
-1.89073503e-01 -2.88914233e-01 -1.26974940e+00 3.47872734e-01
3.32580924e-01 7.40687475e-02 -4.05297995e-01 -1.35034323e+00
-8.22547078e-01 7.71134913e-01 2.34721035e-01 4.30113435e-01
5.87286234e-01 -9.30002630e-01 -7.01297402e-01 -1.20804772e-01
-6.38003349e-02 -2.93353230e-01 -6.91906363e-02 2.36939624e-01
-4.45486456e-01 8.67701888e-01 2.95507669e-01 -2.41193816e-01
-1.73730946e+00 -1.53185487e-01 7.39381611e-01 -3.07877690e-01
1.75835356e-01 6.05130494e-01 8.48112330e-02 -1.21399319e+00
1.09459184e-01 -7.48524427e-01 -4.16938394e-01 -1.23582840e-01
6.77109241e-01 9.44147110e-02 -1.58933997e-01 -1.74575582e-01
-1.22489870e-01 1.03061639e-01 -1.00598887e-01 -6.68260098e-01
4.68627334e-01 -6.06874347e-01 -1.39424995e-01 2.76776880e-01
7.90873587e-01 6.15203977e-01 -8.53155613e-01 -3.65345955e-01
2.10272744e-01 -2.23783940e-01 -9.18935180e-01 -1.51422846e+00
2.73663014e-01 8.50766003e-01 -5.20887189e-02 8.18645120e-01
5.31544507e-01 -6.59294650e-02 7.30242074e-01 1.16058397e+00
-3.16488906e-03 -9.56573665e-01 2.08011046e-01 1.06690812e+00
1.22317624e+00 -1.02279496e+00 -4.01832372e-01 -4.73659724e-01
-7.49603450e-01 1.26775324e+00 1.16008115e+00 5.58962226e-01
-5.48277888e-03 6.32288808e-04 6.31660640e-01 -8.18635300e-02
-1.06107771e+00 1.70351893e-01 1.10380501e-01 5.06679058e-01
8.40759218e-01 -4.72586751e-02 -5.63498318e-01 7.23358572e-01
-6.90049827e-01 -2.11378381e-01 1.00044489e+00 7.90210307e-01
-9.72045004e-01 -1.44176829e+00 -5.41781723e-01 2.44996578e-01
-2.20935822e-01 -3.97482634e-01 -1.26099288e+00 5.89646876e-01
-7.16455042e-01 1.52540243e+00 -4.29734141e-01 -2.00770088e-02
4.20964211e-01 8.20554912e-01 2.79179424e-01 -1.03087652e+00
-1.20335865e+00 -5.12996495e-01 8.70872498e-01 -1.17582247e-01
1.62510034e-02 -3.82119805e-01 -1.31652141e+00 -1.43509850e-01
-3.27542156e-01 1.06752920e+00 1.30198628e-01 1.29171288e+00
4.57895190e-01 8.75041857e-02 5.23788512e-01 1.00653574e-01
-1.04248929e+00 -1.14951801e+00 3.87241155e-01 1.10586002e-01
7.85209060e-01 1.06966242e-01 -5.06314516e-01 -3.33991386e-02] | [12.058337211608887, 7.971707820892334] |
7bf96b58-3942-42f8-808f-f44a4d24177e | three-stream-joint-network-for-zero-shot | 2204.05666 | null | https://arxiv.org/abs/2204.05666v1 | https://arxiv.org/pdf/2204.05666v1.pdf | Three-Stream Joint Network for Zero-Shot Sketch-Based Image Retrieval | The Zero-Shot Sketch-based Image Retrieval (ZS-SBIR) is a challenging task because of the large domain gap between sketches and natural images as well as the semantic inconsistency between seen and unseen categories. Previous literature bridges seen and unseen categories by semantic embedding, which requires prior knowledge of the exact class names and additional extraction efforts. And most works reduce domain gap by mapping sketches and natural images into a common high-level space using constructed sketch-image pairs, which ignore the unpaired information between images and sketches. To address these issues, in this paper, we propose a novel Three-Stream Joint Training Network (3JOIN) for the ZS-SBIR task. To narrow the domain differences between sketches and images, we extract edge maps for natural images and treat them as a bridge between images and sketches, which have similar content to images and similar style to sketches. For exploiting a sufficient combination of sketches, natural images, and edge maps, a novel three-stream joint training network is proposed. In addition, we use a teacher network to extract the implicit semantics of the samples without the aid of other semantics and transfer the learned knowledge to unseen classes. Extensive experiments conducted on two real-world datasets demonstrate the superiority of our proposed method. | ['Xin-Shun Xu', 'Zhen-Duo Chen', 'Yongxin Wang', 'Xin Luo', 'Yu-Wei Zhan'] | 2022-04-12 | null | null | null | null | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 2.29297012e-01 -3.07327151e-01 -2.83310890e-01 -3.94909441e-01
-4.45591271e-01 -5.92090011e-01 8.37366164e-01 -2.74835557e-01
-2.43653178e-01 4.92709339e-01 1.53584689e-01 2.63899833e-01
-8.63291100e-02 -9.57859695e-01 -6.26663446e-01 -4.45587516e-01
5.24137437e-01 2.22549915e-01 5.47005057e-01 -2.73170680e-01
2.68528521e-01 2.62359679e-01 -1.60598433e+00 3.95211399e-01
4.36501831e-01 1.12049484e+00 3.54419947e-01 2.60030061e-01
-7.19327331e-01 4.27027225e-01 -4.70324010e-01 -3.28162193e-01
2.60976881e-01 -3.11693877e-01 -4.90876228e-01 2.62155145e-01
9.00618374e-01 -6.05313003e-01 -7.09370315e-01 1.06911111e+00
4.12352175e-01 2.12208226e-01 6.52916074e-01 -1.68635082e+00
-1.30292976e+00 1.35178834e-01 -4.33051348e-01 -1.81677833e-01
1.39106765e-01 -1.08881593e-01 8.74910474e-01 -1.32746172e+00
9.10870969e-01 1.32627916e+00 3.64926785e-01 7.19828367e-01
-1.02108669e+00 -1.02029157e+00 3.72997701e-01 3.44085962e-01
-1.67077076e+00 -3.34933698e-01 1.28366351e+00 -2.88302302e-01
3.71275753e-01 4.35042940e-02 4.75629419e-01 1.26181960e+00
-5.37372768e-01 1.01155853e+00 7.68244445e-01 -2.87119061e-01
2.43532360e-01 5.55073261e-01 -1.13876581e-01 6.57584012e-01
3.56239043e-02 -5.47304899e-02 -6.98879242e-01 -1.55062944e-01
1.17351508e+00 6.91645801e-01 -4.11908627e-01 -6.33709311e-01
-1.18748713e+00 7.18818188e-01 5.15635014e-01 2.96010762e-01
-1.27660513e-01 2.12358594e-01 3.78124595e-01 2.95425147e-01
3.32318068e-01 1.15189612e-01 -1.63652837e-01 4.16415274e-01
-8.49880397e-01 4.39303778e-02 4.85975057e-01 1.41570699e+00
9.38108861e-01 -1.31593272e-01 -1.35752499e-01 1.08595741e+00
9.76968035e-02 5.75643241e-01 4.22481120e-01 -6.29864633e-01
3.76195014e-01 7.18467295e-01 1.32511556e-01 -1.43814647e+00
6.31438851e-01 -1.68949172e-01 -8.69244158e-01 9.92845073e-02
1.51021346e-01 4.66427892e-01 -9.31555748e-01 1.77760744e+00
1.43743798e-01 5.09216487e-01 1.01285279e-02 1.18633306e+00
1.12002826e+00 5.45086443e-01 1.29122138e-01 3.13146949e-01
1.40520418e+00 -1.10362434e+00 -6.74711525e-01 -3.25353563e-01
-1.23650275e-01 -6.50332808e-01 1.29913747e+00 -3.83723825e-02
-8.39288712e-01 -8.32639217e-01 -1.20158648e+00 -4.25771683e-01
-9.43816960e-01 3.15904528e-01 3.55187356e-01 2.59253055e-01
-7.52658963e-01 4.63726819e-01 -2.26654530e-01 -3.62345606e-01
5.63114703e-01 -5.93371838e-02 -5.34657180e-01 -4.87435281e-01
-1.31707251e+00 3.69165391e-01 3.44849020e-01 -1.87178738e-02
-7.36138701e-01 -6.77809358e-01 -9.95145380e-01 1.73650727e-01
6.82613730e-01 -5.86715102e-01 6.34023309e-01 -1.16591108e+00
-1.12024581e+00 8.60426366e-01 -8.71662572e-02 3.75025809e-01
4.48793948e-01 -1.11951433e-01 -5.08495867e-01 3.99853498e-01
2.44279981e-01 9.02525067e-01 1.29949486e+00 -1.54583085e+00
-5.96789837e-01 -4.33071256e-01 -3.52998707e-03 2.44725212e-01
-7.56202638e-01 -3.44783038e-01 -7.78111160e-01 -8.66013169e-01
2.44686395e-01 -6.27727568e-01 2.82912791e-01 7.82308817e-01
-3.66232656e-02 -3.83061051e-01 1.26871133e+00 -4.87715214e-01
9.16987956e-01 -2.37583113e+00 3.93896624e-02 8.04349333e-02
2.36452326e-01 2.21257046e-01 -6.43642962e-01 4.65871692e-01
-5.16314805e-03 5.45821711e-02 -2.78117597e-01 -3.17246109e-01
4.81744222e-02 4.96014237e-01 -7.92586386e-01 1.62016153e-01
4.30241376e-01 1.11783445e+00 -1.35898399e+00 -6.40530050e-01
2.93931752e-01 5.65338194e-01 -1.49799049e-01 4.15934324e-01
-1.10336453e-01 1.42378166e-01 -5.24408937e-01 6.92333341e-01
9.07872498e-01 -1.35272339e-01 -6.60480931e-02 -5.26328802e-01
1.81630880e-01 -1.99222073e-01 -1.19782412e+00 2.12667251e+00
-4.85566616e-01 5.63669443e-01 -9.61948782e-02 -9.92728591e-01
1.13665342e+00 2.62488037e-01 1.19629890e-01 -8.09131503e-01
-2.32646972e-01 1.52054146e-01 -7.59359896e-01 -3.93768072e-01
3.59686643e-01 -3.84310693e-01 -1.30010739e-01 4.58324850e-01
3.18907708e-01 -2.95755476e-01 -1.15774244e-01 4.12999600e-01
5.85165918e-01 1.72762692e-01 -6.93555251e-02 -4.73314412e-02
6.02159679e-01 -3.15470695e-01 4.46485996e-01 6.79061472e-01
-8.94264653e-02 8.36614192e-01 6.34074509e-02 -6.07212484e-01
-1.06438243e+00 -1.57297432e+00 1.04869427e-02 9.83043373e-01
8.46929669e-01 -3.38659793e-01 -2.86177963e-01 -7.24967003e-01
1.96528599e-01 3.48795921e-01 -6.50312722e-01 -3.71080846e-01
-3.25674266e-01 1.23920068e-02 3.10438275e-01 6.31105304e-01
6.71682835e-01 -1.12411880e+00 -1.33570388e-01 -6.81617558e-02
-2.80210972e-01 -1.23545456e+00 -9.25419986e-01 -4.42438811e-01
-5.84802806e-01 -1.09403396e+00 -9.93821979e-01 -1.21064699e+00
8.14811528e-01 8.03452432e-01 1.08912170e+00 2.80877054e-01
-5.69533765e-01 6.51178718e-01 -4.10706788e-01 -4.30492461e-02
1.25936836e-01 -3.07875425e-01 -1.77551121e-01 3.24354082e-01
4.55508113e-01 -6.32321179e-01 -7.73348510e-01 3.90701622e-01
-1.25066125e+00 6.96061999e-02 6.09019101e-01 1.21252513e+00
5.54126322e-01 -7.22927004e-02 7.05087066e-01 -6.35392904e-01
5.89538515e-01 -5.45484781e-01 -2.71012902e-01 6.81867599e-01
-3.05422783e-01 1.35413811e-01 6.29655898e-01 -7.12947190e-01
-1.04169512e+00 -5.57212811e-03 4.60446924e-01 -1.06651187e+00
-2.35767365e-01 1.26871407e-01 -3.30753893e-01 5.97812682e-02
1.49638504e-01 5.82510948e-01 1.21260174e-01 -5.04989922e-01
7.36016273e-01 8.93509090e-01 7.94054091e-01 -7.09422648e-01
8.26955676e-01 8.24261367e-01 -2.83210754e-01 -1.04208231e+00
-7.50829875e-01 -6.76575661e-01 -5.32102048e-01 6.97280979e-03
7.26724505e-01 -9.84890997e-01 -2.94856131e-01 2.05677822e-01
-1.26153493e+00 1.01386033e-01 -4.29389417e-01 2.41070256e-01
-3.87969196e-01 6.67049348e-01 -4.62794036e-01 -6.61833107e-01
-3.40366811e-01 -7.65671372e-01 1.46449232e+00 1.67049900e-01
1.28251031e-01 -8.66503060e-01 -2.54185259e-01 6.74500391e-02
3.93511027e-01 3.27716358e-02 9.43685532e-01 -3.76782537e-01
-8.13909054e-01 -2.01564029e-01 -9.13685739e-01 3.66991073e-01
4.07210320e-01 -2.63376206e-01 -8.57746899e-01 -2.11363539e-01
-6.80001900e-02 -5.23878694e-01 9.70312715e-01 -1.35076553e-01
1.23927021e+00 -1.28424034e-01 -3.76580417e-01 4.58111912e-01
1.67431045e+00 8.89687538e-02 5.34881115e-01 -5.13361059e-02
7.26891816e-01 6.97073102e-01 4.63246971e-01 3.20352346e-01
2.58927941e-01 5.40041447e-01 1.92067832e-01 -2.03597307e-01
-5.28745651e-01 -7.84601569e-01 5.77521101e-02 8.80946815e-01
3.56614381e-01 -1.11925956e-02 -4.21818495e-01 8.61549616e-01
-1.86537170e+00 -9.35171008e-01 3.54278415e-01 2.22698641e+00
6.92891479e-01 -1.93316191e-01 -2.59481937e-01 -1.53837651e-01
9.87559378e-01 2.09536418e-01 -6.17765963e-01 1.59965754e-01
-1.10651068e-01 2.11847946e-01 1.12263739e-01 2.34029084e-01
-9.79656279e-01 1.06689966e+00 5.21546268e+00 1.23729384e+00
-1.11810446e+00 8.87547433e-02 2.76455045e-01 6.69561327e-02
-4.53899801e-01 9.34665352e-02 -3.67929012e-01 5.52746832e-01
2.51237601e-02 -7.16432184e-02 5.45988798e-01 8.34935427e-01
-4.39793319e-01 2.40285411e-01 -1.24601293e+00 1.40299129e+00
3.94429982e-01 -1.23673391e+00 4.67099726e-01 -3.79970253e-01
5.79440773e-01 -3.90133470e-01 5.59572019e-02 3.21323305e-01
3.29096504e-02 -7.77694881e-01 6.39609039e-01 6.56698585e-01
1.30624509e+00 -4.79058862e-01 5.22328973e-01 1.51872233e-01
-1.52023053e+00 3.92271904e-05 -6.82224691e-01 9.83232558e-02
-1.26646534e-01 2.40424410e-01 -3.54064167e-01 5.93927801e-01
5.73220253e-01 1.05739403e+00 -4.62132215e-01 6.11032724e-01
-2.15609297e-01 -5.94908334e-02 -2.52579629e-01 1.25418022e-01
1.54447198e-01 -3.85245293e-01 2.71081269e-01 1.03643334e+00
3.22022736e-01 1.52372718e-01 2.18470261e-01 1.40148962e+00
-2.48953775e-01 -1.29206702e-01 -9.15891290e-01 -2.93461502e-01
6.83857143e-01 1.19859946e+00 -6.25937581e-01 -7.01522112e-01
-5.69366574e-01 1.50234938e+00 2.18972415e-01 6.98982358e-01
-5.11562169e-01 -8.01745117e-01 6.79965198e-01 -1.48600101e-01
3.11760932e-01 -1.51857585e-01 -6.65713623e-02 -1.37797809e+00
3.13689232e-01 -2.56372243e-01 2.99814105e-01 -9.46372211e-01
-1.83553123e+00 4.70650911e-01 -1.00997776e-01 -1.37958372e+00
2.30996653e-01 -4.88343418e-01 -6.41494751e-01 8.76185179e-01
-1.80470908e+00 -1.40031588e+00 -6.99852228e-01 8.01167309e-01
7.93684721e-01 -1.56732738e-01 7.67926872e-01 4.69105363e-01
-7.75892735e-02 5.69050729e-01 7.11418018e-02 5.36355495e-01
8.87353420e-01 -1.03774500e+00 3.68480086e-01 3.26146215e-01
3.20945352e-01 6.08823836e-01 1.91219941e-01 -6.53746843e-01
-1.46313548e+00 -1.13322330e+00 6.09216392e-01 -3.34769547e-01
5.25980949e-01 -6.86377168e-01 -1.00454831e+00 3.34866226e-01
5.18237390e-02 4.07449335e-01 4.12885755e-01 -1.90744609e-01
-8.50027800e-01 -1.59251153e-01 -9.10623789e-01 6.70412838e-01
1.32372320e+00 -1.10290480e+00 -8.15088034e-01 6.01447187e-02
7.61426628e-01 1.45364419e-01 -6.87901556e-01 2.08770156e-01
8.87135088e-01 -7.58728623e-01 1.43629289e+00 -5.64650774e-01
5.58054745e-01 -2.86176205e-01 -3.54827315e-01 -1.19419837e+00
-6.11487180e-02 7.85891488e-02 2.28179440e-01 1.25608516e+00
-2.24274725e-01 -3.90576988e-01 8.17583263e-01 4.74822938e-01
2.61264592e-01 -4.28631872e-01 -5.72819948e-01 -9.76870120e-01
-5.78082614e-02 -5.77798970e-02 7.31211185e-01 1.18134487e+00
-1.65035889e-01 4.73668724e-01 -4.47035640e-01 -3.72374244e-02
7.83359170e-01 5.97063601e-01 7.39007533e-01 -1.13506210e+00
-4.50082868e-02 -3.68030965e-01 -5.78483701e-01 -1.03194749e+00
2.92080015e-01 -8.77127409e-01 9.11763590e-03 -1.32864094e+00
4.15472597e-01 -4.70200479e-01 -4.39926565e-01 4.38639015e-01
-2.39207700e-01 4.82446700e-01 3.19494784e-01 2.96411663e-01
-6.32406890e-01 1.00514090e+00 1.39494133e+00 -5.52647173e-01
1.63241610e-01 -6.12034023e-01 -4.28485364e-01 4.01877165e-01
3.61132950e-01 -3.75421971e-01 -8.01303625e-01 -5.68090856e-01
-2.00329185e-01 1.16111204e-01 6.63377166e-01 -7.40389049e-01
3.76691490e-01 -2.41447315e-01 4.22980160e-01 -4.33583468e-01
6.09389782e-01 -1.06070924e+00 -1.19002186e-01 1.34692386e-01
-3.97991925e-01 -3.34784657e-01 -9.36213136e-02 1.01624846e+00
-5.98849237e-01 -1.53195605e-01 4.93937552e-01 -3.06648701e-01
-1.07396233e+00 6.16159558e-01 3.74418944e-01 -5.84245883e-02
8.99710774e-01 -4.56599206e-01 -3.07107359e-01 -3.78013670e-01
-6.32152975e-01 2.30685234e-01 5.33897817e-01 8.56325328e-01
1.16115570e+00 -1.71636581e+00 -4.20740813e-01 6.26698315e-01
6.32346451e-01 1.86210424e-02 5.01162887e-01 1.06372423e-01
-1.12674102e-01 2.73735255e-01 -4.70192134e-01 -4.58637178e-01
-1.01713216e+00 9.01465535e-01 -4.46147248e-02 3.88266474e-01
-6.82665646e-01 6.83436453e-01 6.68174684e-01 -5.04269063e-01
2.33774155e-01 -5.13753258e-02 1.05829746e-01 7.66096413e-02
6.06236637e-01 4.49542925e-02 -2.92477846e-01 -3.95163834e-01
-2.29187995e-01 7.61102557e-01 -7.00172484e-02 -7.45600611e-02
1.12297881e+00 -2.53197134e-01 -1.34241402e-01 5.24981916e-01
1.63916254e+00 -3.84704173e-01 -1.24596369e+00 -7.16866016e-01
-6.48933277e-02 -1.00373232e+00 3.10273822e-02 -5.36606610e-01
-1.11770248e+00 1.24729085e+00 6.06652141e-01 -1.97330102e-01
1.03236866e+00 6.43476620e-02 1.08812821e+00 2.47233868e-01
3.05791378e-01 -1.07629395e+00 6.42045319e-01 7.05594718e-02
9.82221007e-01 -1.44869983e+00 -2.62342006e-01 -4.81680900e-01
-4.13373798e-01 1.13053584e+00 6.87354088e-01 -4.16961312e-01
6.04636908e-01 -3.79142106e-01 -8.22233111e-02 -1.90610096e-01
-5.92606187e-01 -2.13543549e-01 3.12074691e-01 6.58278048e-01
-1.13109827e-01 -3.22918333e-02 -7.79609829e-02 7.65960038e-01
5.24371684e-01 2.01294228e-01 1.46305233e-01 9.86644089e-01
-2.89902240e-01 -1.14661634e+00 -8.66919011e-02 2.22662508e-01
1.90250233e-01 -1.62460148e-01 -5.81760585e-01 6.92794085e-01
1.55207038e-01 6.37796938e-01 3.45600545e-01 -3.36430877e-01
2.11555108e-01 1.61438376e-01 4.48636860e-01 -5.36060154e-01
9.67767239e-02 -1.76057339e-01 -5.17662525e-01 -3.25350374e-01
-3.30503225e-01 1.61974020e-02 -9.77246940e-01 1.38159571e-02
-3.04326773e-01 1.46886781e-01 7.69275725e-01 7.46682286e-01
4.83094543e-01 1.68204173e-01 6.23847246e-01 -8.48544419e-01
-5.83701968e-01 -6.35494888e-01 -8.71890247e-01 9.75687206e-01
3.23569775e-01 -8.91722977e-01 -3.64986897e-01 6.94184899e-02] | [11.613899230957031, 0.6836234331130981] |
bf502ed2-944b-48e4-844d-7af60b370ac8 | efficient-video-instance-segmentation-via | 2203.01853 | null | https://arxiv.org/abs/2203.01853v1 | https://arxiv.org/pdf/2203.01853v1.pdf | Efficient Video Instance Segmentation via Tracklet Query and Proposal | Video Instance Segmentation (VIS) aims to simultaneously classify, segment, and track multiple object instances in videos. Recent clip-level VIS takes a short video clip as input each time showing stronger performance than frame-level VIS (tracking-by-segmentation), as more temporal context from multiple frames is utilized. Yet, most clip-level methods are neither end-to-end learnable nor real-time. These limitations are addressed by the recent VIS transformer (VisTR) which performs VIS end-to-end within a clip. However, VisTR suffers from long training time due to its frame-wise dense attention. In addition, VisTR is not fully end-to-end learnable in multiple video clips as it requires a hand-crafted data association to link instance tracklets between successive clips. This paper proposes EfficientVIS, a fully end-to-end framework with efficient training and inference. At the core are tracklet query and tracklet proposal that associate and segment regions-of-interest (RoIs) across space and time by an iterative query-video interaction. We further propose a correspondence learning that makes tracklets linking between clips end-to-end learnable. Compared to VisTR, EfficientVIS requires 15x fewer training epochs while achieving state-of-the-art accuracy on the YouTube-VIS benchmark. Meanwhile, our method enables whole video instance segmentation in a single end-to-end pass without data association at all. | ['Gerard Medioni', 'Jayan Eledath', 'Junsong Yuan', 'Tian Lan', 'Hui Liang', 'Sudhir Yarram', 'Jialian Wu'] | 2022-03-03 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Wu_Efficient_Video_Instance_Segmentation_via_Tracklet_Query_and_Proposal_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wu_Efficient_Video_Instance_Segmentation_via_Tracklet_Query_and_Proposal_CVPR_2022_paper.pdf | cvpr-2022-1 | ['video-instance-segmentation'] | ['computer-vision'] | [-7.00748935e-02 -1.71084091e-01 -4.69225645e-01 -3.01744521e-01
-1.36086607e+00 -6.30070925e-01 3.16630572e-01 1.63303148e-02
-4.86407131e-01 4.04266268e-01 -1.50004759e-01 7.67285675e-02
-2.87470296e-02 -4.58587438e-01 -1.30475521e+00 -2.33041242e-01
-1.47179067e-01 4.40662950e-01 7.88451195e-01 2.23068669e-01
-1.49441957e-01 2.56987542e-01 -1.49679458e+00 6.10761642e-01
5.17363131e-01 1.26761293e+00 3.37686926e-01 6.50797188e-01
-1.01893097e-01 9.67858434e-01 -9.28629041e-02 -4.05979633e-01
5.61010480e-01 -3.44877839e-01 -5.89315295e-01 2.17617378e-01
1.23689651e+00 -4.35565680e-01 -6.32564545e-01 8.66404414e-01
3.69624943e-01 2.13024899e-01 1.67721942e-01 -1.30271244e+00
-3.00646544e-01 7.28703499e-01 -8.30045581e-01 2.99070030e-01
4.44875866e-01 2.83630610e-01 1.00165904e+00 -1.17554355e+00
9.75115001e-01 9.31347787e-01 7.81247318e-01 3.42362106e-01
-1.03724456e+00 -6.75921381e-01 5.50986946e-01 4.38534200e-01
-1.36179233e+00 -3.38368922e-01 7.69417763e-01 -5.71414292e-01
7.18859971e-01 2.13915154e-01 1.02886844e+00 8.59955072e-01
-1.41020760e-01 1.44319379e+00 6.52293921e-01 1.00511633e-01
2.85041686e-02 -3.10535997e-01 -3.78017165e-02 8.74585509e-01
-2.74716884e-01 6.80444390e-02 -6.81811690e-01 3.65809381e-01
9.78278577e-01 2.89685905e-01 -3.49926591e-01 -4.74476308e-01
-1.49123943e+00 3.24888051e-01 4.90509659e-01 1.05204239e-01
-6.03025675e-01 4.53625202e-01 7.60461807e-01 2.39243016e-01
4.36881840e-01 3.30518708e-02 -5.86511910e-01 -3.11299264e-01
-1.75759852e+00 2.60889024e-01 4.09650713e-01 1.29683685e+00
6.08877599e-01 6.78029284e-02 -5.94483137e-01 5.27171433e-01
-2.06074817e-03 4.43566322e-01 5.36225401e-02 -1.08945286e+00
6.21777177e-01 2.79401422e-01 -6.42884597e-02 -6.64713264e-01
-2.04840153e-01 -6.58021986e-01 -5.17295122e-01 5.53887635e-02
4.38435972e-01 -9.89556685e-02 -9.68509138e-01 1.56487894e+00
5.68636656e-01 1.15191996e+00 -4.32402790e-01 1.22819710e+00
9.15581167e-01 7.49891400e-01 2.30299518e-01 -3.11430275e-01
1.43777704e+00 -1.60511005e+00 -6.83814645e-01 -7.68249333e-02
4.68807757e-01 -5.85335076e-01 9.09824550e-01 4.63719904e-01
-1.62924862e+00 -9.99725342e-01 -6.70091748e-01 -3.15409690e-01
-1.35639802e-01 2.69227419e-02 4.10499245e-01 -1.23615218e-02
-7.51903415e-01 2.46730536e-01 -1.06654418e+00 -3.82810794e-02
7.98923373e-01 3.14654380e-01 -4.23974991e-01 -8.04391205e-02
-7.85046458e-01 2.10165039e-01 4.07220870e-01 3.14784981e-02
-1.10580230e+00 -1.44502413e+00 -8.84326458e-01 4.22611348e-02
9.43764627e-01 -7.37559199e-01 1.23732078e+00 -1.12090576e+00
-1.23470235e+00 8.33896458e-01 -2.74093419e-01 -7.10375011e-01
9.88405168e-01 -6.60515189e-01 -2.71876365e-01 5.20076036e-01
1.49799258e-01 9.66879308e-01 8.55648577e-01 -1.07071853e+00
-1.00374639e+00 4.50174175e-02 2.35082969e-01 1.20959133e-02
-1.10364959e-01 1.01688407e-01 -1.53645551e+00 -8.27225447e-01
-9.79258865e-02 -8.21947217e-01 4.56450656e-02 4.52934533e-01
-2.69249648e-01 -2.80883908e-01 1.13251185e+00 -6.08055234e-01
1.31737936e+00 -2.17144990e+00 2.27410808e-01 -1.58148363e-01
3.03509086e-01 5.15990198e-01 -3.86443168e-01 2.15770498e-01
3.99793237e-02 -3.84688973e-01 -2.75681950e-02 -5.59558809e-01
7.51072615e-02 -6.68094009e-02 -1.87493309e-01 5.15595436e-01
6.27115816e-02 1.26119411e+00 -1.01030195e+00 -8.57805431e-01
5.27080953e-01 5.29428005e-01 -8.25544477e-01 3.21516246e-01
-5.74594617e-01 3.78960341e-01 -3.20395321e-01 8.24241996e-01
5.92718184e-01 -2.65475661e-01 -2.23776177e-01 -7.65098691e-01
-1.82807073e-01 -2.50864387e-01 -1.07403862e+00 2.35495663e+00
-2.14500755e-01 8.52391183e-01 4.69097160e-02 -1.08713543e+00
4.47393507e-01 3.67974669e-01 1.05976820e+00 -8.03066313e-01
1.09834215e-02 -1.84889901e-02 -5.16235173e-01 -7.99695432e-01
3.46990824e-01 4.29712713e-01 -1.41213173e-02 1.02336057e-01
2.18104169e-01 4.55243647e-01 5.57117701e-01 2.83171892e-01
9.57560241e-01 6.79022372e-01 -1.20042361e-01 1.12221636e-01
4.74531084e-01 1.06077105e-01 7.95807958e-01 6.47094190e-01
-1.66599616e-01 9.13560450e-01 2.49633789e-01 -5.27626216e-01
-9.57146943e-01 -1.18805122e+00 -6.35886937e-02 1.23215938e+00
5.03320277e-01 -4.54787433e-01 -8.51058781e-01 -9.52064991e-01
-7.64674917e-02 3.98381382e-01 -5.76255679e-01 2.16992259e-01
-8.81403685e-01 1.17644839e-01 5.09965122e-01 6.72974169e-01
5.43005347e-01 -1.09335089e+00 -7.44862497e-01 5.67276180e-01
-3.26837301e-01 -1.55382538e+00 -1.08167160e+00 -2.06749097e-01
-7.06098735e-01 -1.26908755e+00 -8.85249197e-01 -7.88754940e-01
5.18137515e-01 4.09287244e-01 1.24518001e+00 3.40751535e-03
-4.50190574e-01 5.80581784e-01 -4.15116906e-01 -8.87728203e-03
1.79989487e-01 1.09685920e-01 -3.23847562e-01 1.64094999e-01
1.81355074e-01 -3.14474046e-01 -8.94872189e-01 5.48943937e-01
-9.47729886e-01 2.34066993e-01 4.73193079e-01 5.81973016e-01
1.22749305e+00 -3.82866383e-01 4.29676175e-01 -8.69133234e-01
-3.19005698e-01 -4.78846759e-01 -7.22235739e-01 5.07281601e-01
-8.45270008e-02 -4.82441664e-01 4.22746450e-01 -5.51331401e-01
-6.89252555e-01 3.44172299e-01 -2.29154658e-02 -1.22754622e+00
-8.89307410e-02 4.79765296e-01 6.42027780e-02 8.04842412e-02
3.05051804e-01 2.46233448e-01 -2.63539865e-03 -4.08596516e-01
4.66953784e-01 1.61354944e-01 1.04121375e+00 -5.51240683e-01
8.14783275e-01 3.90476733e-01 -2.54268795e-01 -5.34439862e-01
-1.22271276e+00 -8.05028737e-01 -8.64906609e-01 -6.37482822e-01
1.17453754e+00 -1.38710296e+00 -7.64478683e-01 2.16734022e-01
-9.13647175e-01 -6.48924232e-01 -4.33836192e-01 5.75316429e-01
-8.19944620e-01 1.78125635e-01 -6.90203488e-01 -3.12608868e-01
-4.01790053e-01 -1.08955312e+00 1.29590976e+00 7.69072101e-02
3.92197222e-02 -9.01825666e-01 -2.74632335e-01 5.45100451e-01
5.85916489e-02 5.13685822e-01 8.02529603e-02 -2.37280235e-01
-1.15311682e+00 -4.08306748e-01 -3.88547450e-01 6.92764148e-02
-1.74533293e-01 9.69921574e-02 -6.94771290e-01 -5.26277304e-01
-4.72188622e-01 -3.21263194e-01 9.19708133e-01 6.73321068e-01
1.47664297e+00 -2.91498423e-01 -3.64128500e-01 9.61838305e-01
1.58917642e+00 9.44468677e-02 5.20901382e-01 1.92748517e-01
1.05640864e+00 2.73696929e-01 1.17678368e+00 2.70505548e-01
4.14419711e-01 1.12059402e+00 3.86823624e-01 -4.74270761e-01
-5.22600532e-01 -2.27008492e-01 5.94554305e-01 5.86342216e-01
-4.32002321e-02 -2.31222749e-01 -4.79614764e-01 6.15562677e-01
-2.23375058e+00 -1.38993537e+00 -1.71155646e-01 2.08140707e+00
6.34744287e-01 1.81832775e-01 5.96718490e-01 -1.81845143e-01
6.39859319e-01 1.67799771e-01 -7.90346682e-01 4.18923467e-01
8.62289965e-03 3.14117856e-02 4.25870061e-01 3.35005730e-01
-1.34188879e+00 1.16597939e+00 5.07057190e+00 1.00100505e+00
-1.23434055e+00 3.26870948e-01 4.89355236e-01 -5.89900970e-01
1.46932244e-01 -1.67986140e-01 -6.94730937e-01 5.36449730e-01
5.13713777e-01 8.25026247e-04 2.61814713e-01 8.13981831e-01
4.31307822e-01 2.80944239e-02 -1.32439113e+00 1.18934023e+00
7.13415146e-02 -1.81367552e+00 -5.40300608e-02 -2.89751917e-01
6.86031520e-01 2.67082870e-01 -3.97626683e-02 4.63306069e-01
-3.47589403e-01 -5.48603714e-01 1.30876386e+00 6.85670197e-01
9.78338301e-01 -7.53917575e-01 4.62035924e-01 1.29674271e-01
-1.74661040e+00 -1.49832442e-02 -2.79443562e-01 3.40220749e-01
6.88106537e-01 4.88399059e-01 -4.17674780e-01 6.70803487e-01
7.43889511e-01 1.13523173e+00 -5.13557017e-01 1.48295891e+00
1.05258510e-01 7.20572352e-01 -2.77528822e-01 5.88057697e-01
5.43774426e-01 -1.85118560e-02 4.75804389e-01 1.64231384e+00
3.92520934e-01 1.47217408e-01 8.80047023e-01 6.18895590e-01
-9.39932317e-02 -2.23812789e-01 -1.19693175e-01 2.21640900e-01
4.60813105e-01 1.30235946e+00 -8.26958835e-01 -7.08191335e-01
-6.42873824e-01 1.03490055e+00 1.29321396e-01 4.78133529e-01
-1.35293174e+00 3.91828502e-03 5.59479058e-01 5.08434057e-01
8.70886087e-01 -1.45995438e-01 1.01846062e-01 -1.17283022e+00
1.75127536e-01 -6.55610800e-01 5.01335680e-01 -6.44160867e-01
-1.18593311e+00 4.76834178e-01 1.53890960e-02 -1.64587867e+00
-2.71547902e-02 -1.94354922e-01 -4.02867109e-01 1.79283902e-01
-1.53423214e+00 -1.43323612e+00 -4.77838993e-01 9.40553427e-01
1.11247659e+00 9.81067345e-02 2.43154868e-01 9.44627523e-01
-7.49944806e-01 9.86580670e-01 -2.28633121e-01 5.45211315e-01
7.65357852e-01 -1.12512219e+00 2.70466357e-01 8.90247464e-01
4.47018981e-01 1.34001940e-01 4.99489784e-01 -5.88932574e-01
-1.56293488e+00 -1.61225760e+00 3.90587866e-01 -3.82997751e-01
5.14938891e-01 -3.54253143e-01 -9.08047557e-01 7.36203492e-01
2.26509556e-01 7.98134863e-01 3.31479311e-01 -2.41950706e-01
-3.84121865e-01 -3.68173093e-01 -8.36491048e-01 3.12445194e-01
1.37377751e+00 -5.76981246e-01 -2.59803534e-01 4.60302383e-01
8.95850420e-01 -8.96355748e-01 -1.08064508e+00 3.58419359e-01
5.30785382e-01 -9.35154557e-01 1.18826258e+00 -5.28234899e-01
4.50561315e-01 -7.04523802e-01 6.59944266e-02 -6.58880651e-01
-3.53698581e-01 -7.76855767e-01 -5.27073383e-01 1.29271972e+00
2.55044550e-01 1.17791593e-01 9.22614515e-01 3.49328786e-01
-5.27473867e-01 -1.03509700e+00 -8.72890890e-01 -9.84638453e-01
-4.53627974e-01 -7.12368667e-01 3.33057553e-01 9.85916317e-01
-2.38684118e-01 1.61766484e-02 -5.35613179e-01 2.14909121e-01
9.38182771e-01 3.50275755e-01 9.79178786e-01 -8.55467916e-01
-4.10820723e-01 -3.98272634e-01 -4.89646435e-01 -1.35538793e+00
5.79924919e-02 -8.56640160e-01 5.92090711e-02 -1.49360073e+00
1.09692201e-01 -4.17417109e-01 -5.32661617e-01 5.52325368e-01
-2.66659945e-01 6.11059904e-01 6.91803396e-01 2.34810308e-01
-1.39734066e+00 2.58808643e-01 1.40434027e+00 -2.76224375e-01
-2.20768422e-01 -1.18850671e-01 -7.36125186e-02 5.10251641e-01
3.00421298e-01 -5.77205062e-01 -5.53460300e-01 -6.30003870e-01
-7.42091164e-02 6.88543737e-01 5.65899968e-01 -1.38801563e+00
5.60355842e-01 -2.76626237e-02 3.50655913e-01 -1.10632861e+00
5.22228718e-01 -9.56810713e-01 3.76453072e-01 2.42407501e-01
-4.37700361e-01 1.33924231e-01 2.45155290e-01 6.83942199e-01
-4.59330559e-01 2.22950175e-01 7.27909684e-01 7.27625713e-02
-1.07459295e+00 1.01344824e+00 1.46056235e-01 3.59760076e-01
1.49939299e+00 -4.50022280e-01 8.15687701e-02 1.67221259e-02
-8.17993939e-01 5.13543308e-01 2.88689137e-01 5.56576908e-01
4.96188194e-01 -1.29207873e+00 -8.49271238e-01 -1.79065645e-01
7.21951053e-02 3.17588896e-01 7.38749325e-01 1.17271757e+00
-5.17430246e-01 1.61723465e-01 -1.03951849e-01 -1.07718968e+00
-1.35963154e+00 9.06948924e-01 1.47603348e-01 -8.28280076e-02
-1.07137525e+00 1.05030346e+00 3.67769927e-01 1.74910232e-01
5.47278583e-01 -2.55557090e-01 1.09608091e-01 2.72747785e-01
6.62874997e-01 2.16343805e-01 -1.32060245e-01 -5.42777538e-01
-1.02030687e-01 9.51544166e-01 -2.25416139e-01 9.45826322e-02
1.18566823e+00 -9.82337072e-02 3.02585512e-01 3.09750527e-01
1.35565472e+00 -1.72476232e-01 -2.07417679e+00 -4.14829552e-01
-3.18684071e-01 -6.08824074e-01 -1.72049683e-02 -6.25260770e-01
-1.67474210e+00 6.99827611e-01 5.51049709e-01 -1.60809442e-01
1.12669933e+00 -2.73237973e-02 1.36977339e+00 2.05529809e-01
3.12588304e-01 -1.10317504e+00 1.42703682e-01 1.53947875e-01
7.29786813e-01 -1.11379576e+00 9.49805509e-03 -4.54456568e-01
-5.98398924e-01 9.52202916e-01 7.53668606e-01 -2.16903731e-01
6.16974294e-01 2.85960823e-01 5.95481955e-02 -2.18551129e-01
-7.68031240e-01 -2.33398229e-01 6.01149201e-01 3.81286532e-01
1.87404722e-01 -1.98991612e-01 -4.06090133e-02 2.41414025e-01
2.32390076e-01 3.97027969e-01 2.85992678e-02 6.71749949e-01
-1.91900760e-01 -7.53391385e-01 -1.33440703e-01 3.54403555e-01
-5.19748390e-01 1.91494495e-01 3.07165176e-01 8.01352203e-01
3.00902605e-01 6.53038323e-01 2.31692642e-01 -1.96034670e-01
4.38902378e-01 -2.61969626e-01 3.97941649e-01 -3.33164394e-01
-7.91644156e-01 3.69148523e-01 -2.10455969e-01 -1.12387061e+00
-6.89231694e-01 -7.63161182e-01 -1.45746899e+00 -1.73031822e-01
-2.13410303e-01 4.77299243e-02 2.11412624e-01 8.44553351e-01
5.74676931e-01 7.60336816e-01 4.43886071e-01 -1.07317400e+00
1.48342222e-01 -4.95108157e-01 -1.24993823e-01 5.94559669e-01
4.11334246e-01 -5.36244750e-01 3.04192811e-01 5.27727783e-01] | [9.132604598999023, -0.02941025421023369] |
b534b910-0dad-4f5e-9ca9-b8c84b28cafc | contrastive-language-action-pre-training-for | 2204.12293 | null | https://arxiv.org/abs/2204.12293v1 | https://arxiv.org/pdf/2204.12293v1.pdf | Contrastive Language-Action Pre-training for Temporal Localization | Long-form video understanding requires designing approaches that are able to temporally localize activities or language. End-to-end training for such tasks is limited by the compute device memory constraints and lack of temporal annotations at large-scale. These limitations can be addressed by pre-training on large datasets of temporally trimmed videos supervised by class annotations. Once the video encoder is pre-trained, it is common practice to freeze it during fine-tuning. Therefore, the video encoder does not learn temporal boundaries and unseen classes, causing a domain gap with respect to the downstream tasks. Moreover, using temporally trimmed videos prevents to capture the relations between different action categories and the background context in a video clip which results in limited generalization capacity. To address these limitations, we propose a novel post-pre-training approach without freezing the video encoder which leverages language. We introduce a masked contrastive learning loss to capture visio-linguistic relations between activities, background video clips and language in the form of captions. Our experiments show that the proposed approach improves the state-of-the-art on temporal action localization, few-shot temporal action localization, and video language grounding tasks. | ['Loris Bazzani', 'Michael Donoser', 'Bernard Ghanem', 'Maksim Lapin', 'Erhan Gundogdu', 'Mengmeng Xu'] | 2022-04-26 | null | null | null | null | ['few-shot-temporal-action-localization', 'action-localization'] | ['computer-vision', 'computer-vision'] | [ 3.61100912e-01 -1.17176391e-01 -5.68789482e-01 -3.00415993e-01
-7.80858457e-01 -6.12454176e-01 5.37597299e-01 -1.13497026e-01
-4.66691792e-01 5.79765975e-01 3.54483426e-01 -4.69160601e-02
2.08206818e-01 -2.57643998e-01 -1.09372687e+00 -3.55461866e-01
-1.71196401e-01 -8.41981769e-02 4.89403695e-01 2.86828160e-01
-5.18142283e-02 -1.81530640e-02 -1.38487971e+00 8.09615612e-01
6.16819203e-01 1.18217194e+00 3.93197060e-01 5.15310109e-01
5.53651201e-03 1.40281713e+00 -2.46312261e-01 -1.67485282e-01
3.02944899e-01 -6.23467863e-01 -6.99747741e-01 3.73332143e-01
7.50898063e-01 -6.87124312e-01 -7.02665150e-01 7.83435941e-01
7.61534423e-02 4.02623564e-01 1.57642633e-01 -1.31519306e+00
-5.16949475e-01 4.99634773e-01 -5.94687283e-01 3.84323508e-01
4.42033291e-01 2.49188811e-01 8.19671631e-01 -6.84506238e-01
8.26744616e-01 1.07615733e+00 5.38256168e-01 7.32051134e-01
-8.88575971e-01 -6.82364941e-01 8.36599827e-01 3.88523340e-01
-1.26667678e+00 -6.60736144e-01 6.52591884e-01 -6.80365980e-01
9.74025548e-01 -9.88023803e-02 4.82872427e-01 1.52544415e+00
-1.46520555e-01 9.00010765e-01 6.71057820e-01 -2.87069559e-01
3.19188833e-01 -1.56726643e-01 -1.65555537e-01 8.53208244e-01
-3.13212872e-01 -3.38840783e-02 -1.06946290e+00 2.43092149e-01
8.15802872e-01 1.76840663e-01 -2.41680890e-01 -5.25557399e-01
-1.29712105e+00 3.11797768e-01 2.41768077e-01 4.42010939e-01
-3.05782855e-01 4.42826182e-01 7.38543153e-01 2.38611341e-01
4.02008444e-01 6.82661384e-02 -5.67500353e-01 -4.16411549e-01
-1.28243804e+00 -3.86387467e-01 4.62351203e-01 1.12032735e+00
7.08361447e-01 -4.87824865e-02 -3.69732618e-01 4.19311017e-01
5.47819509e-05 2.28832796e-01 4.74235982e-01 -1.12915480e+00
9.34364855e-01 4.06896740e-01 7.69310668e-02 -7.55640447e-01
7.57918227e-03 -7.72059932e-02 -3.91787559e-01 -2.16814280e-01
5.55730402e-01 -5.16127162e-02 -9.82768476e-01 1.93939984e+00
1.29071474e-01 9.10496175e-01 -3.23240273e-02 1.00411785e+00
3.25982392e-01 5.60551405e-01 3.51287156e-01 -3.67191792e-01
1.20984983e+00 -1.23324907e+00 -7.79671371e-01 -6.38256967e-01
7.66435206e-01 -3.65089566e-01 1.27369809e+00 2.46925324e-01
-9.96165276e-01 -7.69190669e-01 -9.13490951e-01 -2.39367098e-01
-1.41797498e-01 2.57147849e-01 4.51868117e-01 1.34843141e-01
-7.61421323e-01 4.85246330e-01 -1.46659160e+00 -4.64039028e-01
4.70731705e-01 2.00940058e-01 -4.34303194e-01 -3.10357273e-01
-1.07100224e+00 4.98346001e-01 5.31971633e-01 1.00984760e-01
-1.22220027e+00 -6.48452401e-01 -9.99014914e-01 -3.93449739e-02
8.31285954e-01 -4.96979058e-01 1.18014777e+00 -1.62757969e+00
-1.41879618e+00 7.06776857e-01 -2.68278837e-01 -5.86904526e-01
7.14381158e-01 -5.79822540e-01 -3.58944029e-01 5.83428085e-01
1.34624988e-01 7.33927488e-01 9.54114556e-01 -7.98758745e-01
-7.76306450e-01 -1.97077855e-01 4.07507926e-01 2.29735553e-01
-5.69344759e-01 1.45054877e-01 -1.02120852e+00 -6.65914118e-01
-2.25752100e-01 -9.78184581e-01 -3.89974238e-03 2.81700283e-01
2.88138911e-03 1.17575452e-01 1.07740676e+00 -8.26255322e-01
1.25155902e+00 -2.45621753e+00 1.28780216e-01 -2.28292972e-01
-1.44838661e-01 1.65408310e-02 -3.29427958e-01 1.68652982e-01
3.25442175e-03 -1.11238062e-01 1.17455050e-03 -4.47962701e-01
-2.16712564e-01 3.34009409e-01 -4.45608974e-01 4.16291237e-01
1.58325195e-01 8.35276186e-01 -1.14809263e+00 -6.30932391e-01
2.37994835e-01 5.60269594e-01 -7.46728361e-01 2.01868683e-01
-5.88656664e-01 7.22594202e-01 -6.07817709e-01 5.89431167e-01
2.16379121e-01 -4.06868666e-01 3.30078930e-01 -3.01873356e-01
6.40733317e-02 3.81899297e-01 -9.32837605e-01 2.52194118e+00
-6.34024680e-01 8.39152157e-01 2.58637182e-02 -1.11650741e+00
2.31939182e-01 5.40346563e-01 7.38316715e-01 -7.77261913e-01
-5.40311225e-02 4.91133332e-02 -3.35402280e-01 -8.31477106e-01
1.32486284e-01 6.38843700e-02 -1.18407577e-01 2.62733489e-01
2.65619606e-01 5.27877986e-01 2.62131393e-01 1.78441688e-01
1.27064776e+00 7.35697329e-01 3.11568771e-02 1.55609295e-01
3.59671623e-01 -1.26488209e-01 7.61779666e-01 5.59891343e-01
-2.23223045e-01 5.94160497e-01 4.51933026e-01 -2.69662857e-01
-7.17764795e-01 -8.56082201e-01 3.85922134e-01 1.64442301e+00
2.00329959e-01 -6.41273439e-01 -7.39504278e-01 -9.51923609e-01
-3.05346191e-01 3.68577331e-01 -6.37622178e-01 -2.54612207e-01
-8.73115122e-01 -6.98058084e-02 4.09376770e-01 8.87396932e-01
4.29040074e-01 -8.92267406e-01 -8.02413940e-01 2.23903701e-01
-4.64452893e-01 -1.87167680e+00 -8.69315386e-01 3.34394425e-02
-9.17975128e-01 -1.10813630e+00 -4.67850834e-01 -7.61551678e-01
6.24134898e-01 2.95748621e-01 1.06102192e+00 -1.46052748e-01
-3.70702264e-03 5.68010390e-01 -4.64117587e-01 2.54303485e-01
-2.30236426e-01 -7.75931478e-02 -4.54897434e-02 3.35367769e-01
3.68852079e-01 -6.49653316e-01 -7.08819568e-01 3.21241617e-01
-8.55719626e-01 2.00408161e-01 4.80075866e-01 6.06305063e-01
6.01770461e-01 -1.24028780e-01 2.17366293e-01 -5.83589971e-01
-1.94540977e-01 -3.95011038e-01 -6.10691369e-01 5.03787756e-01
2.68595647e-02 7.53139779e-02 5.70447445e-01 -8.66335392e-01
-1.01863790e+00 4.68571484e-01 4.93785620e-01 -1.08246100e+00
-6.41803369e-02 4.31066513e-01 -2.26482525e-01 1.66257903e-01
3.54337633e-01 1.58387974e-01 -3.73325586e-01 -4.25596833e-01
4.03648943e-01 4.71484423e-01 7.56524861e-01 -6.35230482e-01
4.52372164e-01 6.29705131e-01 -3.03142488e-01 -6.66050851e-01
-1.14458144e+00 -6.01139367e-01 -8.81590664e-01 -2.46152028e-01
1.21425939e+00 -1.45678484e+00 -3.50263983e-01 7.89701492e-02
-1.13966060e+00 -6.30329072e-01 -1.25426427e-01 6.91301584e-01
-7.74177372e-01 3.73752236e-01 -6.25210047e-01 -5.58411896e-01
5.21015339e-02 -1.06783772e+00 1.20888579e+00 -1.35052040e-01
-1.69865310e-01 -8.19367886e-01 -3.42178822e-01 5.89995325e-01
-1.51049867e-02 2.00301871e-01 5.83969295e-01 -2.86848903e-01
-9.46029067e-01 -6.79405453e-03 -1.80472717e-01 3.55808407e-01
1.01800025e-01 -3.45063746e-01 -9.45203424e-01 -3.51546884e-01
2.26775985e-02 -4.90130007e-01 9.84748960e-01 2.84140378e-01
1.31572449e+00 -3.98160964e-01 -4.49116588e-01 6.28639817e-01
1.18920624e+00 2.18870848e-01 4.52248245e-01 1.71310648e-01
9.56417978e-01 5.65878808e-01 8.41820598e-01 3.82971138e-01
2.41575584e-01 8.58620524e-01 1.28747165e-01 1.89017445e-01
-2.64886051e-01 -5.74152231e-01 8.76980305e-01 5.15249074e-01
1.89141482e-02 -2.09533244e-01 -9.12024975e-01 6.34479523e-01
-2.18171406e+00 -1.20281267e+00 5.23394883e-01 2.11652994e+00
8.14696670e-01 1.89271241e-01 1.14103317e-01 -2.23894730e-01
6.11027300e-01 2.02712342e-01 -7.09247172e-01 2.73375958e-01
1.41800955e-01 -1.35169089e-01 4.07133400e-01 4.82106894e-01
-1.34782982e+00 1.09357357e+00 5.09636211e+00 6.40294313e-01
-1.41880620e+00 4.86265600e-01 4.20833051e-01 -5.87973654e-01
1.61046356e-01 7.32305869e-02 -6.16052210e-01 5.53456604e-01
9.96556044e-01 1.78185955e-01 4.73776579e-01 7.76320755e-01
5.69245934e-01 -1.10837817e-01 -1.70923424e+00 1.12838376e+00
2.77310163e-01 -1.10136425e+00 -2.03802124e-01 -2.00276569e-01
7.33442783e-01 1.88849956e-01 -7.28256479e-02 4.37824130e-01
-1.87960029e-01 -7.32571006e-01 1.02312648e+00 3.35455328e-01
1.03337753e+00 -3.10213894e-01 1.82823539e-01 2.93269128e-01
-1.34129965e+00 -3.33917946e-01 -4.35230024e-02 -1.70298055e-01
2.86255985e-01 2.89913058e-01 -5.25797844e-01 2.53275543e-01
7.33716249e-01 1.10405171e+00 -3.88599694e-01 6.19751513e-01
-1.88339695e-01 6.55727744e-01 -2.97454566e-01 6.55698538e-01
3.26826662e-01 3.30835320e-02 2.14376554e-01 1.25372684e+00
2.86254197e-01 1.16714686e-01 6.85861766e-01 5.52471519e-01
-1.11546457e-01 -2.77791858e-01 -5.28892159e-01 -3.50502253e-01
3.06387156e-01 6.98336184e-01 -7.96672642e-01 -3.48405927e-01
-9.52461302e-01 1.35008740e+00 1.64802700e-01 6.87658489e-01
-1.21371782e+00 2.10503638e-01 6.28035665e-01 3.30022842e-01
5.44698298e-01 -4.04034197e-01 4.89888750e-02 -1.60436809e+00
3.81010830e-01 -8.80958498e-01 5.28009415e-01 -6.71451330e-01
-9.49784815e-01 4.52784181e-01 -1.12722698e-03 -1.41894174e+00
-4.17944402e-01 -4.21736628e-01 -1.95337072e-01 2.35498220e-01
-1.41659057e+00 -1.32928014e+00 -2.99782544e-01 7.68207252e-01
9.02870893e-01 5.19985631e-02 3.92255574e-01 6.47744119e-01
-5.59897900e-01 5.51844597e-01 -2.06209436e-01 3.05497766e-01
8.66887033e-01 -7.93306530e-01 -1.15432873e-01 1.15334094e+00
3.56241673e-01 4.18749332e-01 5.56907713e-01 -6.24469995e-01
-1.30873799e+00 -1.27950191e+00 7.37503052e-01 -4.96429116e-01
8.14418435e-01 -7.83006489e-01 -9.35291052e-01 9.43603635e-01
7.63364509e-02 5.37058473e-01 4.97506380e-01 5.25085256e-02
-5.95968306e-01 -1.69317916e-01 -5.28750002e-01 4.92229939e-01
1.42254591e+00 -1.11233795e+00 -4.27985638e-01 5.09263098e-01
8.26678634e-01 -3.65575433e-01 -5.46898723e-01 2.24074483e-01
5.65055370e-01 -7.24335551e-01 8.50642741e-01 -7.12246358e-01
6.12887681e-01 -3.65536392e-01 -1.95132613e-01 -6.75687253e-01
1.06035240e-01 -7.95404792e-01 -4.75422144e-01 1.27111292e+00
2.37723470e-01 2.28065819e-01 9.16356683e-01 7.78246284e-01
-1.22754663e-01 -4.44306612e-01 -9.42369878e-01 -9.58154082e-01
-2.99485087e-01 -5.62146902e-01 1.47953955e-02 1.00248063e+00
1.87448829e-01 3.24652761e-01 -6.42349422e-01 2.69040883e-01
2.30202496e-01 1.71102300e-01 5.11262834e-01 -7.30747283e-01
-5.92573106e-01 -2.87626777e-03 -3.41212302e-01 -1.35595691e+00
5.01532972e-01 -5.92451751e-01 2.83429503e-01 -1.13982368e+00
2.44886860e-01 -7.66494423e-02 -6.06668413e-01 7.80573606e-01
2.78041475e-02 2.48715013e-01 3.08206499e-01 2.36901522e-01
-1.28248572e+00 4.66864914e-01 1.00042903e+00 -9.06358585e-02
-3.21126550e-01 -3.44795316e-01 -9.49331000e-02 7.18240380e-01
4.05505270e-01 -4.67771560e-01 -8.35016847e-01 -8.31389844e-01
3.03753987e-02 3.65570933e-01 4.89731967e-01 -1.21579337e+00
3.01794976e-01 -3.16860050e-01 1.38148665e-01 -2.81037807e-01
4.80503678e-01 -9.73559558e-01 -6.37795180e-02 1.68670222e-01
-5.59637427e-01 -6.46117628e-02 2.89076149e-01 9.05478179e-01
-4.40226167e-01 2.37489656e-01 6.00398540e-01 -9.25204605e-02
-1.12097204e+00 3.24916154e-01 -2.33372912e-01 2.69946545e-01
1.18709648e+00 -3.82949144e-01 -9.51938629e-02 -3.46407026e-01
-8.88818860e-01 2.87276834e-01 5.56821883e-01 7.43200064e-01
3.95966530e-01 -1.11121094e+00 -2.02723667e-01 5.09840362e-02
2.38929302e-01 -1.49363846e-01 4.70700145e-01 1.00911260e+00
-2.20669582e-01 4.78645593e-01 -1.08093083e-01 -6.47712350e-01
-1.28660524e+00 9.40859079e-01 3.40276718e-01 -1.61443546e-01
-7.43313074e-01 8.25512052e-01 7.00948715e-01 2.77800739e-01
6.15720510e-01 -6.12108111e-01 9.30248201e-02 3.75738740e-02
5.86957872e-01 -2.47489233e-02 -1.48750275e-01 -6.06113851e-01
-4.52871591e-01 4.91965026e-01 -4.55887523e-03 -1.88243434e-01
1.10965180e+00 -4.38938439e-01 2.84446746e-01 5.55734992e-01
1.39324880e+00 -2.05161601e-01 -2.03682137e+00 -3.73417169e-01
1.46257713e-01 -4.19124424e-01 1.25713736e-01 -7.85862982e-01
-1.10678220e+00 9.64099824e-01 5.79895556e-01 -5.46198487e-01
1.39056718e+00 7.60522336e-02 1.00424993e+00 4.32678938e-01
3.42149615e-01 -1.24269223e+00 5.23787439e-01 4.70631987e-01
6.66724861e-01 -1.31365192e+00 -2.63754010e-01 -4.37198579e-01
-7.98726976e-01 9.09163475e-01 7.57375956e-01 1.34816512e-01
5.36371529e-01 2.66439289e-01 -5.57747856e-02 9.34974551e-02
-8.97345841e-01 -2.26850823e-01 3.87148649e-01 3.46074045e-01
3.64869505e-01 -3.16748708e-01 1.01214185e-01 6.05409384e-01
3.68186444e-01 5.16422510e-01 1.93163425e-01 1.10848105e+00
-7.38334209e-02 -9.19009030e-01 8.36987197e-02 -7.73539627e-03
-5.70268095e-01 -2.81368457e-02 -1.86973602e-01 5.83369970e-01
4.04329330e-01 8.39798689e-01 1.93092436e-01 -2.44378328e-01
1.11192219e-01 2.03620270e-01 6.28556788e-01 -7.17012465e-01
-1.87010467e-01 2.64314234e-01 8.95171314e-02 -1.00656736e+00
-7.21069574e-01 -6.68069601e-01 -1.26504123e+00 2.17203125e-01
-6.51497394e-02 2.40652189e-02 2.23226890e-01 1.11691427e+00
5.35922766e-01 5.20163953e-01 1.72732741e-01 -7.43477702e-01
-2.15646639e-01 -6.74468815e-01 -2.14518383e-01 6.28512561e-01
5.36140084e-01 -6.77266181e-01 -9.34844166e-02 7.13064134e-01] | [8.566803932189941, 0.6729558706283569] |
ba758621-171f-4185-b597-222b32ca343a | enrich-multi-purpose-dataset-for-benchmarking | null | null | https://www.sciencedirect.com/science/article/pii/S0924271623000539 | https://doi.org/10.1016/j.isprsjprs.2023.03.002 | ENRICH: Multi-purposE dataset for beNchmaRking In Computer vision and pHotogrammetry | The availability of high-resolution data and accurate ground truth is essential to evaluate and compare methods and algorithms properly. Moreover, it is often difficult to acquire real data for a given application domain that is sufficiently representative and heterogeneous in terms of scene representation, acquisition conditions, point of view, etc. To overcome the limitations of available datasets, this paper presents a new synthetic, multi-purpose dataset called ENRICH for testing photogrammetric and computer vision algorithms. Compared to existing datasets, ENRICH offers higher resolution images rendered with different lighting conditions, camera orientations, scales, and fields of view. Specifically, ENRICH is composed of three sub-datasets: ENRICH-Aerial, ENRICH-Square, and ENRICH-Statue, each exhibiting different characteristics. We show the usefulness of the proposed dataset on several examples of photogrammetry and computer vision-related tasks such as: evaluation of hand-crafted and deep learning-based local features, effects of ground control points (GCPs) configuration on the 3D accuracy, and monocular depth estimation. | ['Fabio Remondino', 'Gianluigi Ciocca', 'Simone Bianco', 'Elisa Mariarosaria Farella', 'Luca Morelli', 'Davide Marelli'] | 2023-04-01 | null | null | null | isprs-journal-of-photogrammetry-and-remote-10 | ['3d-scene-reconstruction', '3d-reconstruction', 'monocular-depth-estimation', 'key-point-matching'] | ['computer-vision', 'computer-vision', 'computer-vision', 'natural-language-processing'] | [ 1.75669134e-01 -5.21607757e-01 1.94757834e-01 -4.25609291e-01
-6.99022055e-01 -6.31155014e-01 7.35054910e-01 1.83262572e-01
-3.21369171e-01 6.05836451e-01 -1.69053778e-01 -4.85606119e-02
-3.51291001e-01 -1.04830539e+00 -5.30962408e-01 -5.83398879e-01
1.01866171e-01 5.99161744e-01 1.28254652e-01 -3.20947051e-01
4.15424287e-01 1.13205218e+00 -2.07540035e+00 -1.98721990e-01
8.49445939e-01 1.06756806e+00 6.13278151e-01 2.94387460e-01
3.16773862e-01 1.45566314e-01 -1.31358966e-01 -1.03793196e-01
6.65947676e-01 5.06366044e-02 -4.24834549e-01 4.88520831e-01
1.10091317e+00 -5.15541911e-01 -3.36084098e-01 1.01850498e+00
7.75625467e-01 2.16818437e-01 6.38058424e-01 -1.03167617e+00
-4.53664482e-01 -3.33161876e-02 -6.34290636e-01 5.97474761e-02
3.65723908e-01 1.95976332e-01 7.59559512e-01 -8.51159155e-01
6.82942927e-01 9.70841467e-01 6.23644471e-01 5.13528176e-02
-1.11801219e+00 -3.28950405e-01 -1.61032975e-01 1.38810903e-01
-1.59959149e+00 -2.24149927e-01 8.48763525e-01 -7.28465974e-01
6.49386048e-01 3.99538353e-02 4.73990262e-01 8.06874871e-01
-1.95556916e-02 1.19801007e-01 1.26696074e+00 -5.02650976e-01
4.05261964e-01 4.87153158e-02 -1.74937218e-01 5.39488852e-01
4.18026268e-01 4.16766524e-01 -2.67342627e-01 7.52980784e-02
1.04752088e+00 -3.01682670e-02 -5.98484576e-01 -7.39542067e-01
-1.33618283e+00 6.99120641e-01 5.97907603e-01 2.70735472e-01
-5.48291266e-01 -2.53508151e-01 8.20675641e-02 -2.25988711e-04
2.47565895e-01 5.56276679e-01 -4.97851431e-01 2.26851001e-01
-9.39639330e-01 1.16298750e-01 3.45243454e-01 1.10577703e+00
1.29743409e+00 1.71801075e-01 3.19596440e-01 7.73989141e-01
-1.13400808e-02 7.99874663e-01 3.60381395e-01 -8.82572770e-01
5.27583838e-01 6.59172654e-01 3.04797202e-01 -1.45826256e+00
-7.19012201e-01 -2.30558768e-01 -8.90486479e-01 3.49425584e-01
3.34783077e-01 9.99413058e-02 -5.59695661e-01 1.06865871e+00
2.65377104e-01 -5.49325533e-02 9.54761636e-03 1.21960747e+00
9.72680449e-01 5.16891658e-01 -3.00037801e-01 1.59335025e-02
9.83614206e-01 -6.04130447e-01 -1.96255982e-01 -4.67045575e-01
5.10253608e-01 -7.27890015e-01 1.28232908e+00 4.80388790e-01
-7.48707414e-01 -7.02679276e-01 -1.06207335e+00 -2.37602778e-02
-2.93045849e-01 7.20679820e-01 5.76501727e-01 3.53198647e-01
-8.18967700e-01 5.63232422e-01 -4.28795695e-01 -6.22034788e-01
2.04129785e-01 2.44769268e-02 -6.94780231e-01 -4.02480781e-01
-8.11913192e-01 9.70946074e-01 4.25241917e-01 1.62668735e-01
-9.26784575e-01 -4.65485275e-01 -9.75615144e-01 -5.93297482e-02
1.20110400e-01 -5.55002153e-01 8.31174970e-01 -8.02432060e-01
-1.15737391e+00 1.04524302e+00 4.47577506e-01 -1.86080292e-01
4.94801670e-01 -4.76812571e-02 -2.40301907e-01 1.85085326e-01
5.41850142e-02 4.38889652e-01 7.65379369e-01 -1.34378266e+00
-6.06047153e-01 -9.17594552e-01 1.49650201e-01 4.05529380e-01
-4.16502878e-02 -3.36620390e-01 -2.88601041e-01 -1.89890727e-01
4.42969680e-01 -8.78218114e-01 -2.23759755e-01 -7.55754411e-02
-1.68947771e-01 4.88285065e-01 6.32746994e-01 -5.06768465e-01
6.07540131e-01 -2.14000535e+00 8.52920637e-02 6.68051168e-02
-2.08939314e-01 1.78446993e-01 -1.18922710e-01 5.06978035e-01
-7.10679218e-02 -1.51370510e-01 -2.05702081e-01 7.32657537e-02
-2.68950731e-01 2.31991664e-01 5.24481349e-02 7.70659685e-01
-2.82925576e-01 3.86581093e-01 -7.21997797e-01 -4.57444251e-01
1.06466103e+00 3.94086987e-01 -1.50245979e-01 3.40160616e-02
-1.06799372e-01 6.98437631e-01 -5.08510709e-01 8.99762809e-01
9.56487358e-01 1.26665622e-01 -9.53838453e-02 -4.34407741e-01
-3.28922242e-01 -2.00748175e-01 -1.41025078e+00 1.72858179e+00
-8.47270548e-01 7.87859678e-01 -1.42143965e-01 -6.69503748e-01
1.31558406e+00 -1.53592788e-02 3.04349869e-01 -8.86367917e-01
2.64456689e-01 3.01771581e-01 -4.96753573e-01 -4.52595383e-01
8.84224057e-01 4.62269410e-02 2.91821748e-01 -2.26436138e-01
-2.22062506e-02 -6.56783521e-01 5.79797812e-02 -4.20688808e-01
6.77222013e-01 2.89532822e-02 5.17196417e-01 -1.52235657e-01
6.21897936e-01 2.05152184e-01 2.95576602e-01 3.00886124e-01
1.12346876e-02 9.63563025e-01 4.51961830e-02 -6.68891490e-01
-1.34479344e+00 -6.92296326e-01 -4.78337467e-01 4.13035661e-01
5.12774706e-01 3.92612368e-02 -2.58667886e-01 -1.20415337e-01
1.48855820e-01 5.91934860e-01 -3.34521979e-01 2.69952506e-01
-2.75369793e-01 -6.18828893e-01 2.19302356e-01 3.56430411e-01
8.12075973e-01 -9.19976473e-01 -9.94482696e-01 -6.09244518e-02
-9.54061225e-02 -1.47521293e+00 1.90178648e-01 -2.86339819e-01
-1.09654868e+00 -1.37950313e+00 -5.09949386e-01 -5.91193259e-01
6.67966425e-01 6.63862526e-01 1.21999884e+00 -2.40542337e-01
-3.10988635e-01 5.15410006e-01 -5.06391764e-01 -6.01351410e-02
-9.76950377e-02 9.29746255e-02 -1.98562130e-01 3.78650092e-02
1.40383895e-02 -6.56692088e-01 -5.42890191e-01 5.74004710e-01
-9.70141172e-01 -7.48614175e-03 7.14374244e-01 7.48180330e-01
6.87564433e-01 2.40204632e-01 -6.23363443e-02 -3.94206852e-01
1.76741749e-01 -1.45701185e-01 -1.05965638e+00 1.00013331e-01
-4.93105888e-01 -3.14177096e-01 4.06556278e-01 2.11662967e-02
-8.98965955e-01 2.00550959e-01 6.34483546e-02 -4.14613783e-01
-6.72202826e-01 5.28949142e-01 -3.10360938e-01 -4.64804947e-01
1.08308840e+00 3.11456501e-01 -3.18326116e-01 -4.90191460e-01
1.27722234e-01 6.07468545e-01 6.78953588e-01 -5.19032836e-01
9.00357187e-01 5.75032294e-01 2.90102661e-01 -1.27120709e+00
-4.94950742e-01 -6.17008448e-01 -1.00528991e+00 -1.91124305e-01
6.32131159e-01 -1.13904428e+00 -1.36240959e-01 5.95313609e-01
-1.11046886e+00 -1.10799961e-01 -2.13289186e-01 7.38017201e-01
-6.10813856e-01 4.53710496e-01 4.50832658e-02 -6.74197435e-01
-2.22734958e-01 -1.30018318e+00 1.42995405e+00 1.38453126e-01
2.74723530e-01 -8.45680535e-01 -4.99386117e-02 4.59583610e-01
2.13296190e-01 7.72431493e-01 8.51601064e-01 3.70134339e-02
-7.65610933e-01 -3.16189170e-01 -3.00914407e-01 4.88216609e-01
1.06523737e-01 2.87259042e-01 -9.66781735e-01 -3.31046104e-01
-3.03701401e-01 -3.39832902e-01 4.56964612e-01 5.61251640e-01
9.05548215e-01 5.20760939e-02 -2.88748275e-02 8.92826080e-01
2.08342695e+00 6.80118278e-02 7.89673209e-01 8.29344332e-01
5.93651354e-01 5.65829575e-01 9.83203053e-01 6.84924185e-01
2.99144715e-01 8.51329565e-01 9.74246025e-01 -1.20177634e-01
1.74931258e-01 -2.26660490e-01 -5.28450385e-02 2.62896836e-01
-4.01858777e-01 -1.01498134e-01 -1.13487899e+00 5.77673614e-01
-1.50087440e+00 -7.87605286e-01 -2.54432499e-01 2.52858567e+00
1.51977643e-01 -3.99259150e-01 -2.33704165e-01 3.04348677e-01
7.03093171e-01 2.62317330e-01 -4.36868876e-01 1.07731983e-01
-4.55462486e-01 -2.25090101e-01 8.89068186e-01 4.07746822e-01
-1.22523153e+00 9.32780206e-01 5.62837982e+00 6.09976828e-01
-1.49675477e+00 -2.58685261e-01 1.91566601e-01 3.60246718e-01
-2.11245134e-01 -2.72585172e-02 -7.56760597e-01 2.11582080e-01
4.30074662e-01 9.41915214e-02 3.76469672e-01 1.12484169e+00
1.08649544e-01 -4.57279414e-01 -1.00385213e+00 1.36385036e+00
2.92614579e-01 -1.36328065e+00 -2.29862351e-02 5.40155172e-02
1.00632906e+00 4.03775901e-01 -9.07509308e-03 -1.27952144e-01
-1.04706533e-01 -7.30758131e-01 6.37512326e-01 3.61740649e-01
9.19781625e-01 -7.03721702e-01 9.89152789e-01 4.61476922e-01
-8.98934901e-01 -1.39788643e-01 -6.96476221e-01 -8.25974159e-03
-1.68356314e-01 5.20923138e-01 -6.76516652e-01 1.03688598e+00
8.17304850e-01 7.49833226e-01 -8.14055145e-01 1.14810896e+00
-3.14516217e-01 1.55351369e-03 -3.31982434e-01 3.00684988e-01
3.17236841e-01 -4.90772098e-01 3.82327437e-01 7.08837807e-01
6.73766196e-01 7.07633197e-02 1.65412754e-01 6.88277006e-01
2.30138808e-01 2.18645871e-01 -1.04162002e+00 2.34718412e-01
6.02136850e-01 1.46898890e+00 -5.33048987e-01 3.22691351e-02
-4.22473103e-01 5.79589307e-01 -7.24584162e-02 4.03713286e-01
-5.28372705e-01 -4.76059504e-02 3.88768256e-01 4.45708781e-01
1.17903762e-01 -6.13184988e-01 -2.20022038e-01 -1.26295745e+00
5.74947856e-02 -9.67108130e-01 1.23459943e-01 -1.20014238e+00
-9.98044848e-01 8.46967638e-01 1.54353946e-01 -1.69198084e+00
-2.11209849e-01 -7.14048803e-01 -3.41892779e-01 8.87938619e-01
-1.67174709e+00 -1.27002776e+00 -1.23583424e+00 6.37512803e-01
4.83601868e-01 -4.21410531e-01 7.60005355e-01 2.87487090e-01
-3.45201641e-01 5.66526428e-02 3.05182189e-01 -1.33250430e-01
5.36750078e-01 -8.81711066e-01 2.65406370e-01 8.32047522e-01
1.43984050e-01 -4.78736982e-02 6.37196898e-01 -1.25778601e-01
-1.42469168e+00 -1.15150201e+00 3.00090641e-01 -2.60713339e-01
2.42428884e-01 5.20912968e-02 -5.95646322e-01 6.14034593e-01
-3.08038533e-01 2.18160480e-01 2.18592852e-01 -1.26380071e-01
2.45167431e-03 -3.22116733e-01 -1.33930981e+00 3.69804859e-01
7.70170927e-01 -3.56512219e-01 -3.78644824e-01 2.87056506e-01
1.02701500e-01 -5.85980237e-01 -9.22281623e-01 7.02756584e-01
4.07155335e-01 -1.49428940e+00 1.10239136e+00 5.95897473e-02
5.00149190e-01 -3.74458164e-01 -7.60985017e-01 -1.46911061e+00
-2.39685029e-01 -3.00112963e-02 5.30757070e-01 1.02833796e+00
8.66974667e-02 -5.87506950e-01 6.74072683e-01 3.37668955e-01
-2.62414277e-01 -5.67727983e-01 -7.30606794e-01 -8.07848692e-01
-1.85892403e-01 -3.55663031e-01 9.01106298e-01 1.10224009e+00
-6.23531938e-01 2.38140121e-01 -2.23626778e-01 6.34596884e-01
4.86667126e-01 5.12708604e-01 1.24745250e+00 -1.48258352e+00
1.37798160e-01 -1.71045914e-01 -9.12465930e-01 -6.34694636e-01
-6.07605204e-02 -5.53024530e-01 -2.51869589e-01 -1.68559325e+00
-3.21531802e-01 -7.03715026e-01 4.05488491e-01 1.52466074e-01
3.70640159e-01 2.16093466e-01 1.04641952e-01 4.27998364e-01
6.00799965e-03 7.05063522e-01 1.32344556e+00 -9.74832326e-02
-1.38024032e-01 -3.12350895e-02 -2.37630337e-01 9.92072523e-01
7.53055871e-01 -4.19532321e-02 -3.78345907e-01 -8.29443395e-01
1.58846408e-01 2.34088033e-01 5.73539853e-01 -1.24794304e+00
1.11706264e-01 -3.00424635e-01 4.51788783e-01 -9.67001557e-01
4.45935398e-01 -1.12448359e+00 2.79267788e-01 4.35183719e-02
6.39057308e-02 1.31542996e-01 2.24753857e-01 3.04314315e-01
-5.09325266e-01 -3.64770532e-01 9.40123558e-01 -4.01948869e-01
-1.36432791e+00 5.64466715e-01 2.26119369e-01 -1.19056977e-01
1.18852055e+00 -6.41117394e-01 -2.05145031e-01 -3.17653328e-01
-2.80436605e-01 -1.48959517e-01 9.65142488e-01 2.14254126e-01
7.73744345e-01 -1.30440903e+00 -8.20393622e-01 5.70753813e-01
4.99720871e-01 4.70018089e-01 5.01197159e-01 5.62742472e-01
-1.16978824e+00 4.14509177e-01 -6.99583948e-01 -8.19321275e-01
-1.18665934e+00 4.28119510e-01 4.43207473e-01 1.03274278e-01
-6.38487875e-01 4.08324093e-01 1.39696866e-01 -7.72589624e-01
2.22893469e-02 -4.50560421e-01 -1.38328597e-01 3.17781828e-02
2.01508746e-01 4.93114591e-01 3.67392361e-01 -9.06763673e-01
-2.75191277e-01 1.19750202e+00 5.69327176e-01 1.06280252e-01
1.25061154e+00 -2.26485819e-01 -3.85636166e-02 2.32935175e-01
8.62326086e-01 -1.86018944e-01 -1.28305054e+00 -3.78689766e-01
-2.19081879e-01 -1.05752480e+00 3.46567214e-01 -5.41567564e-01
-1.07174039e+00 1.10664916e+00 7.86412239e-01 -9.29423645e-02
1.32712555e+00 -2.35706642e-01 2.71105945e-01 5.23318827e-01
9.62728500e-01 -9.62638974e-01 -2.02028856e-01 3.58813614e-01
1.24280524e+00 -1.53320754e+00 2.29361996e-01 -2.99318641e-01
-7.00343251e-01 1.21202147e+00 5.79125464e-01 -9.01212692e-02
3.90772879e-01 -1.47155538e-01 1.61026925e-01 -2.34027535e-01
-8.95688087e-02 -2.78299630e-01 2.34650612e-01 8.74163270e-01
8.67841244e-02 3.03679705e-03 6.93210810e-02 -1.29521713e-01
-3.67328852e-01 -1.16947688e-01 6.86449647e-01 6.51462555e-01
-4.78466094e-01 -6.09086633e-01 -6.68660581e-01 9.70823765e-02
1.79349277e-02 1.24471746e-01 -1.98021814e-01 1.21051991e+00
2.25815862e-01 7.00155795e-01 5.39673604e-02 -4.32740718e-01
5.37757754e-01 -4.55981731e-01 5.98599374e-01 -4.61511612e-01
2.75455527e-02 -2.35334843e-01 5.66782095e-02 -3.02033603e-01
-5.22198796e-01 -7.23394096e-01 -7.85071075e-01 -3.57534468e-01
-4.35099095e-01 -1.74715668e-01 9.85539496e-01 7.47056067e-01
2.61150301e-01 2.61607826e-01 8.10707629e-01 -1.37169063e+00
-4.91411805e-01 -1.00492275e+00 -9.85628426e-01 3.34283084e-01
1.48501411e-01 -8.95767927e-01 -3.71569663e-01 -3.58609594e-02] | [8.550796508789062, -2.4482765197753906] |
3e060078-1140-4957-8618-a0db909ebe02 | data-driven-predictive-latency-for-5g-a | 2307.02329 | null | https://arxiv.org/abs/2307.02329v2 | https://arxiv.org/pdf/2307.02329v2.pdf | Data-driven Predictive Latency for 5G: A Theoretical and Experimental Analysis Using Network Measurements | The advent of novel 5G services and applications with binding latency requirements and guaranteed Quality of Service (QoS) hastened the need to incorporate autonomous and proactive decision-making in network management procedures. The objective of our study is to provide a thorough analysis of predictive latency within 5G networks by utilizing real-world network data that is accessible to mobile network operators (MNOs). In particular, (i) we present an analytical formulation of the user-plane latency as a Hypoexponential distribution, which is validated by means of a comparative analysis with empirical measurements, and (ii) we conduct experimental results of probabilistic regression, anomaly detection, and predictive forecasting leveraging on emerging domains in Machine Learning (ML), such as Bayesian Learning (BL) and Machine Learning on Graphs (GML). We test our predictive framework using data gathered from scenarios of vehicular mobility, dense-urban traffic, and social gathering events. Our results provide valuable insights into the efficacy of predictive algorithms in practical applications. | ['Roberto Verdone', 'Simone Bizzarri', 'Giorgio Ghinamo', 'Davide Micheli', 'Andrea Orsi', 'Nicol Sarcone Grande', 'Francesca Conserva', 'Marco Skocaj'] | 2023-07-05 | null | null | null | null | ['anomaly-detection', 'management', 'decision-making'] | ['methodology', 'miscellaneous', 'reasoning'] | [ 1.89632382e-02 2.46719569e-01 -3.52095723e-01 -4.64230895e-01
-5.57053864e-01 -3.64203244e-01 3.77748758e-01 1.33869663e-01
1.34649530e-01 1.01024926e+00 -2.67154455e-01 -1.19341838e+00
-7.80795515e-01 -7.01320171e-01 -4.12302971e-01 -4.76221442e-01
-1.10796118e+00 9.02267277e-01 3.90686005e-01 6.08352311e-02
6.58229664e-02 9.73815620e-01 -1.03123903e+00 -5.30444756e-02
7.72305727e-01 1.49753821e+00 -2.46864811e-01 8.06021452e-01
2.80049175e-01 8.33411694e-01 -5.45649230e-01 -5.21694839e-01
8.01078677e-02 2.97035038e-01 -7.74073303e-01 9.01306123e-02
-3.05114120e-01 -2.31231883e-01 -6.19983792e-01 4.33792353e-01
2.59216666e-01 -2.84889508e-02 7.27737606e-01 -2.12312007e+00
-5.55677265e-02 5.11435747e-01 -3.19922328e-01 7.45138526e-01
-6.43013685e-04 -2.98216492e-02 9.86420453e-01 -2.94434756e-01
3.23625118e-01 1.18521345e+00 6.07645214e-01 1.78058878e-01
-1.19116557e+00 -4.88024175e-01 3.47535342e-01 3.86940688e-01
-1.33371997e+00 -7.62534618e-01 5.01516581e-01 -4.70964938e-01
7.41996586e-01 3.06843311e-01 4.14474666e-01 8.57555509e-01
4.60962534e-01 5.20057023e-01 3.69317889e-01 -3.19074839e-01
4.48270649e-01 3.45622122e-01 -8.92618075e-02 4.75978613e-01
2.79100955e-01 3.43875974e-01 -8.78282636e-02 -5.42929709e-01
4.44311798e-01 -1.30529061e-01 1.12058803e-01 -3.49397063e-01
-4.03789669e-01 7.43493795e-01 5.14303520e-02 -2.09825024e-01
-7.20157802e-01 6.71439230e-01 1.19435102e-01 5.32048583e-01
8.50645602e-01 -2.95055240e-01 -5.36873698e-01 -4.96290684e-01
-1.03962076e+00 8.19872171e-02 1.20761085e+00 1.39822888e+00
5.86959004e-01 1.06739849e-01 8.74765515e-02 3.50955099e-01
8.03604484e-01 5.01961768e-01 -3.81393731e-01 -1.12951350e+00
3.61398429e-01 9.60161984e-02 2.22466275e-01 -9.51585650e-01
-6.01144135e-01 -8.36024761e-01 -5.34312606e-01 -2.93409169e-01
3.83130819e-01 -4.54913378e-01 -4.42473859e-01 1.57174397e+00
-1.12200916e-01 6.12733185e-01 -2.27056026e-01 2.66904593e-01
2.11642936e-01 5.28398275e-01 1.35091275e-01 -6.67474508e-01
6.91189110e-01 -5.19693255e-01 -4.43556309e-01 -1.11573495e-01
9.43205714e-01 -5.93694985e-01 3.01483989e-01 3.60135674e-01
-1.06134081e+00 -6.96562044e-03 -5.91561615e-01 5.67409396e-01
4.84501161e-02 -4.24570471e-01 8.50608408e-01 1.24547875e+00
-1.13826227e+00 3.38143706e-01 -9.67922032e-01 -7.17224121e-01
5.95166981e-01 4.80139226e-01 1.35180891e-01 -1.38321385e-01
-9.27869022e-01 4.60579574e-01 -4.26867492e-02 -4.53083776e-02
-9.25820351e-01 -7.19287813e-01 -3.67742479e-01 1.23279333e-01
5.56832850e-01 -5.18591166e-01 1.26649272e+00 -4.70689029e-01
-1.06527364e+00 4.00738448e-01 -3.83631438e-01 -7.56519437e-01
3.89809608e-01 -9.69810709e-02 -8.99398565e-01 2.50025213e-01
-1.44682363e-01 -2.49725804e-02 5.35858631e-01 -1.13938367e+00
-9.18285906e-01 -3.42470556e-01 1.72910571e-01 -7.01691151e-01
2.74168421e-02 9.80534926e-02 -2.35719845e-01 -2.36714929e-01
2.13496342e-01 -1.10179102e+00 -4.66667265e-01 -4.71098930e-01
-4.15492117e-01 -1.65288836e-01 1.03058541e+00 -5.49628496e-01
1.64304411e+00 -1.67960989e+00 -2.42713079e-01 1.06056070e+00
4.82370734e-01 -2.64439940e-01 1.31190762e-01 8.23158383e-01
3.17249715e-01 3.56576324e-01 3.27153563e-01 -3.22973520e-01
-1.34600237e-01 3.03983808e-01 -3.72663349e-01 3.86911988e-01
-3.36829513e-01 7.13697016e-01 -8.42342973e-01 -4.01763886e-01
8.02568495e-02 -5.59583604e-02 -7.50719130e-01 4.27124426e-02
-2.90646285e-01 2.91963190e-01 -6.27648532e-01 1.03477657e+00
5.22837996e-01 -5.38840532e-01 5.48319399e-01 -1.69165894e-01
1.20969497e-01 9.07358751e-02 -8.52932036e-01 7.65871108e-01
-5.92179239e-01 6.25274301e-01 -5.19840792e-02 -1.07382345e+00
6.46008193e-01 3.19022596e-01 7.07973182e-01 -3.58603328e-01
1.66925147e-01 1.92954212e-01 1.93124428e-01 -2.85034478e-01
-1.97854377e-02 1.97845146e-01 7.19282478e-02 5.03878474e-01
9.72937979e-03 3.70437473e-01 8.25598612e-02 4.51876998e-01
1.47251701e+00 -5.77200890e-01 1.50256351e-01 -1.87911943e-01
2.74896622e-01 -5.47215521e-01 3.60919714e-01 1.03936505e+00
-6.27929211e-01 -1.70744926e-01 1.03726256e+00 -4.46968108e-01
-8.39074016e-01 -1.52049339e+00 -4.83028479e-02 1.07557833e+00
6.00295626e-02 -4.41380322e-01 -4.67544913e-01 -5.80451488e-01
7.64716342e-02 9.56271589e-01 -2.37149000e-01 -4.71215472e-02
-2.50290573e-01 -1.04832160e+00 2.87865341e-01 3.20657879e-01
9.90120098e-02 -3.30280900e-01 1.73281431e-01 5.94415843e-01
-1.21486476e-02 -1.90257013e+00 6.95879981e-02 -2.57260650e-02
-9.37120914e-01 -8.85485947e-01 3.24431211e-01 -8.83578584e-02
1.96708024e-01 2.89363950e-01 1.23823476e+00 6.32417500e-02
-7.03084515e-03 8.52860093e-01 -1.27047837e-01 -2.59556800e-01
-5.50390720e-01 4.02904809e-01 4.33918774e-01 1.62454695e-01
3.52281183e-01 -1.33803582e+00 -6.54200673e-01 8.67445230e-01
-1.83481723e-01 -4.29737866e-01 6.30559564e-01 1.30108282e-01
2.20492035e-01 2.04978928e-01 1.03829634e+00 -8.61042023e-01
5.44880629e-01 -1.21374607e+00 -7.79762924e-01 2.42781177e-01
-1.19250917e+00 -3.17954510e-01 1.57817304e-01 -4.21275236e-02
-7.82822728e-01 -7.35736489e-01 8.10537860e-02 -1.99521899e-01
6.62534684e-03 6.28702581e-01 -2.37328783e-01 -2.53715008e-01
4.66181993e-01 -2.00501621e-01 -1.46972775e-01 -1.71204880e-01
3.43908191e-01 9.73070741e-01 4.31308858e-02 -8.58483195e-01
9.11730349e-01 7.05178142e-01 4.69278544e-01 -1.18273318e+00
-6.67798519e-01 -4.19508278e-01 -4.55990285e-01 -6.04293644e-01
9.94016528e-02 -4.52862084e-01 -9.49212909e-01 4.68419343e-02
-9.21129405e-01 -2.34400138e-01 2.31487691e-01 5.34470618e-01
-7.17865467e-01 2.95559287e-01 -8.01311791e-01 -1.39967382e+00
2.30125040e-02 -7.49207616e-01 4.77355927e-01 -2.85573602e-02
-1.56999901e-01 -1.37632680e+00 -2.74246335e-01 5.82147300e-01
6.13669157e-01 1.34033198e-02 1.13833833e+00 -9.00293231e-01
-1.18060625e+00 -3.19289654e-01 -6.63581669e-01 9.69255567e-02
-1.28494620e-01 4.94869918e-01 -8.91953349e-01 -3.45665216e-01
-5.37070632e-01 2.57803530e-01 2.22829074e-01 8.29141140e-01
1.23895323e+00 -4.14519191e-01 -8.20995986e-01 5.88165522e-01
1.15174258e+00 4.19242054e-01 7.01323688e-01 5.83717898e-02
3.76285255e-01 8.27374697e-01 5.76650202e-01 7.37990320e-01
6.85123861e-01 5.76693535e-01 7.91882157e-01 3.70456189e-01
3.56255531e-01 -1.70425758e-01 3.64852026e-02 4.78033394e-01
-4.22895581e-01 -8.47576439e-01 -9.18562233e-01 2.90101558e-01
-2.02744436e+00 -1.05681467e+00 -2.21232742e-01 2.18782592e+00
-1.50817066e-01 7.42136598e-01 5.01070559e-01 1.63785085e-01
5.57711959e-01 -6.27446175e-02 -2.60896295e-01 -3.63931805e-01
2.92385578e-01 -9.64440554e-02 9.77458119e-01 5.16915023e-01
-9.14058447e-01 7.73083687e-01 6.89588451e+00 1.03166616e+00
-8.11931431e-01 1.44366801e-01 9.72688675e-01 1.25784986e-02
-2.04954103e-01 2.06124812e-01 -5.73517442e-01 5.88615477e-01
1.75412488e+00 -2.98011631e-01 4.45110291e-01 7.52091765e-01
7.76495874e-01 -6.75588176e-02 -1.06339335e+00 8.05018127e-01
-5.68981528e-01 -1.41690564e+00 -5.90973459e-02 5.64485788e-01
4.53494430e-01 3.68624330e-01 6.49787709e-02 2.72570729e-01
2.52915859e-01 -1.04947805e+00 3.02874804e-01 9.03970242e-01
4.76311445e-01 -1.02752864e+00 9.09233212e-01 2.20557854e-01
-9.79191780e-01 -4.74250764e-01 1.09962307e-01 -5.83638251e-02
6.42826319e-01 9.52507496e-01 -1.03278935e+00 4.95273530e-01
4.51510549e-01 5.44220805e-01 -2.65789270e-01 1.36831498e+00
2.99576342e-01 1.18725729e+00 -4.31808084e-01 1.53614610e-01
1.28360674e-01 -8.78650919e-02 8.24058354e-01 1.11116660e+00
3.76404583e-01 -9.14798900e-02 1.89700902e-01 5.15307784e-01
9.44328308e-02 -9.10811871e-02 -3.47395390e-01 -1.95280418e-01
9.49677229e-01 1.12150395e+00 -7.60340571e-01 1.29708126e-01
-6.10295713e-01 2.27706820e-01 -1.29256770e-01 8.64646554e-01
-8.10742557e-01 -2.48459667e-01 9.15859818e-01 8.40521395e-01
8.73358473e-02 -6.18809283e-01 -3.92734766e-01 -6.67357266e-01
-3.04686815e-01 -1.78049505e-01 4.28971797e-01 -4.03222889e-01
-1.31337535e+00 4.60604101e-01 2.29593173e-01 -9.90153611e-01
-6.00741029e-01 -5.37290156e-01 -6.13691688e-01 4.35669333e-01
-1.26532781e+00 -1.08162892e+00 -8.45022574e-02 2.61632115e-01
-8.61123055e-02 -5.61668694e-01 5.22230268e-01 7.82763898e-01
-6.19433999e-01 6.15050375e-01 1.62684232e-01 -3.35647553e-01
1.90542266e-01 -9.03337002e-01 3.66131604e-01 6.94208741e-01
-1.56369016e-01 1.38386920e-01 9.24151719e-01 -6.10431373e-01
-1.07072997e+00 -1.10265315e+00 5.73121667e-01 -5.28348982e-01
1.01407933e+00 -1.47974834e-01 -4.67887849e-01 1.03713119e+00
-5.99362493e-01 1.85118020e-01 8.07975888e-01 4.91410613e-01
1.16752356e-01 -4.82787937e-01 -1.26946068e+00 3.42833966e-01
1.24149513e+00 -4.28824604e-01 4.68727410e-01 3.91098976e-01
5.75641751e-01 2.00661551e-02 -9.05005395e-01 4.99166816e-01
5.84815562e-01 -1.00818634e+00 8.78241360e-01 -9.36898470e-01
-4.66753334e-01 1.29308496e-02 -6.94373965e-01 -8.47427845e-01
-1.63720176e-01 -1.16199982e+00 -8.43121648e-01 9.96891797e-01
5.67431033e-01 -7.28445530e-01 1.08834291e+00 6.43487513e-01
2.14436240e-02 -1.05957925e+00 -1.22114050e+00 -8.83881927e-01
-3.32237810e-01 -1.21310246e+00 4.41081107e-01 3.47958624e-01
-2.77600110e-01 6.35755062e-02 -4.23376262e-01 5.65380335e-01
8.48064840e-01 -2.33167544e-01 8.30601513e-01 -1.67213130e+00
-1.23235770e-01 -1.69175938e-01 -9.88748908e-01 -1.21194482e+00
4.29105051e-02 -6.38267577e-01 -5.79131603e-01 -1.17850757e+00
-1.65388718e-01 -9.83034670e-01 -2.66282469e-01 -2.12983206e-01
6.97657287e-01 -6.59185573e-02 -3.59519601e-01 2.16183111e-01
-8.11344266e-01 3.66108298e-01 4.47498649e-01 4.69177693e-01
-1.42114773e-01 1.14412165e+00 -5.50594568e-01 8.26172233e-01
7.71021962e-01 -4.03042048e-01 -8.05332065e-01 2.01501623e-01
4.28404242e-01 5.66568017e-01 5.80252051e-01 -9.28719044e-01
1.58849984e-01 -2.66918242e-01 -3.77778448e-02 -7.47526407e-01
3.49016309e-01 -8.80019128e-01 -5.54485358e-02 2.06237510e-01
2.63795376e-01 -8.68474469e-02 -8.91409814e-02 1.14770007e+00
3.46341074e-01 1.82620019e-01 5.55203199e-01 5.42073309e-01
-6.38589859e-01 8.05420756e-01 -8.26993525e-01 3.21186334e-01
1.16910970e+00 -2.29649857e-01 -3.18331629e-01 -1.21937180e+00
-1.02663374e+00 4.23983365e-01 1.20535623e-02 -1.66881233e-02
5.97513735e-01 -9.45958555e-01 -4.24895823e-01 -1.55521169e-01
4.18780334e-02 -8.27769041e-01 2.63919473e-01 1.15944719e+00
-3.82470161e-01 6.88138664e-01 1.56663924e-01 -7.63224781e-01
-9.53797400e-01 3.07106167e-01 5.12636483e-01 -1.71009511e-01
1.50248809e-02 4.96758491e-01 -3.81476641e-01 -5.72735444e-02
4.17219758e-01 -5.77701628e-02 1.53503150e-01 -1.24921218e-01
1.16805471e-01 1.06359851e+00 1.29217669e-01 -3.20663512e-01
-5.26752353e-01 -1.19486347e-01 -1.05659731e-01 -1.32987991e-01
1.03312218e+00 -7.95482099e-01 2.14808863e-02 4.64434534e-01
8.26296389e-01 1.71208143e-01 -8.33223522e-01 -4.03446436e-01
5.29533684e-01 -4.16534662e-01 1.94188327e-01 -6.37560725e-01
-8.94483328e-01 5.07338762e-01 5.71448565e-01 8.88538301e-01
7.81760156e-01 3.58647346e-01 8.15229416e-01 4.25787717e-01
8.00118685e-01 -7.69715488e-01 -1.45827696e-01 3.53296638e-01
2.55692899e-01 -1.08563077e+00 -1.66770548e-01 -8.23013902e-01
3.61214653e-02 1.12822831e+00 3.51339221e-01 2.38922194e-01
1.56075609e+00 3.39883156e-02 -1.98791713e-01 -2.27355570e-01
-1.21665347e+00 -2.02211425e-01 2.33338013e-01 5.92514873e-01
1.49277896e-01 3.77796412e-01 1.77927032e-01 8.31564665e-01
-7.90676326e-02 -1.20140333e-02 7.31687367e-01 6.22468114e-01
-5.66248298e-01 -1.03006530e+00 1.95587069e-01 9.24828589e-01
-4.53458250e-01 4.25922731e-03 2.60279238e-01 8.03103268e-01
-3.73315096e-01 1.49874437e+00 1.71033278e-01 -5.66423953e-01
-3.88869084e-02 -2.50452667e-01 3.46867859e-01 -3.33292365e-01
4.46260959e-01 -2.18747780e-01 6.95286989e-01 -6.49319649e-01
-2.76151026e-04 -6.40681326e-01 -8.60240996e-01 -9.94672775e-01
-1.95867881e-01 3.98863167e-01 7.01687098e-01 1.27195787e+00
4.52129453e-01 4.95570153e-01 8.77894402e-01 -4.37028885e-01
-1.56426102e-01 -7.49557734e-01 -9.13595438e-01 -3.37231815e-01
2.12100655e-01 -7.78320134e-01 -7.27567732e-01 -4.29485470e-01] | [6.092363357543945, 1.638914704322815] |
7e98ba38-03c7-4bb3-b564-32c2dbd54ccf | toward-more-accurate-and-generalizable | 2306.03984 | null | https://arxiv.org/abs/2306.03984v2 | https://arxiv.org/pdf/2306.03984v2.pdf | Toward More Accurate and Generalizable Evaluation Metrics for Task-Oriented Dialogs | Measurement of interaction quality is a critical task for the improvement of spoken dialog systems. Existing approaches to dialog quality estimation either focus on evaluating the quality of individual turns, or collect dialog-level quality measurements from end users immediately following an interaction. In contrast to these approaches, we introduce a new dialog-level annotation workflow called Dialog Quality Annotation (DQA). DQA expert annotators evaluate the quality of dialogs as a whole, and also label dialogs for attributes such as goal completion and user sentiment. In this contribution, we show that: (i) while dialog quality cannot be completely decomposed into dialog-level attributes, there is a strong relationship between some objective dialog attributes and judgments of dialog quality; (ii) for the task of dialog-level quality estimation, a supervised model trained on dialog-level annotations outperforms methods based purely on aggregating turn-level features; and (iii) the proposed evaluation model shows better domain generalization ability compared to the baselines. On the basis of these results, we argue that having high-quality human-annotated data is an important component of evaluating interaction quality for large industrial-scale voice assistant platforms. | ['Anuj Goyal', 'Timothy Leffel', 'Aram Galstyan', 'Spyros Matsoukas', 'Angeliki Metallinou', 'Nagesh Panyam Chandrasekarasastry', 'Abishek Komma'] | 2023-06-06 | null | null | null | null | ['domain-generalization'] | ['methodology'] | [-3.90205115e-01 4.04419988e-01 -8.62584263e-02 -1.02003741e+00
-9.80431497e-01 -9.18546617e-01 7.05141425e-01 4.15774792e-01
-2.85507083e-01 7.87615895e-01 7.98794448e-01 -1.10671453e-01
-5.75935505e-02 -5.41663587e-01 1.16753317e-01 -3.66469137e-02
3.85119498e-01 8.88926744e-01 1.28419593e-01 -6.76921606e-01
2.77049124e-01 4.03561532e-01 -1.08770049e+00 3.49214226e-01
1.09411025e+00 1.27458012e+00 -8.76568854e-02 8.91145289e-01
-6.33825660e-01 1.11473966e+00 -1.08624196e+00 -4.76377904e-01
-5.24805486e-02 -5.54730058e-01 -1.51839447e+00 3.51639777e-01
1.06604360e-01 -6.79110706e-01 -6.33073971e-02 8.13978493e-01
2.77754545e-01 1.86850712e-01 5.47693014e-01 -1.57371378e+00
-1.47341445e-01 8.59908998e-01 6.05926037e-01 -1.79479122e-01
8.84693146e-01 4.59278136e-01 1.46021903e+00 -6.22495770e-01
3.75318646e-01 1.58352864e+00 3.11008185e-01 6.30211055e-01
-1.09879792e+00 -2.26160124e-01 1.10624686e-01 -2.56759852e-01
-5.70372045e-01 -6.67186797e-01 4.87911135e-01 -7.27784157e-01
8.15156460e-01 2.02258199e-01 2.73821145e-01 1.04189491e+00
-9.66463760e-02 7.97734141e-01 1.34554982e+00 -3.09739500e-01
4.17979062e-01 6.63319170e-01 9.18403506e-01 5.58945477e-01
-6.34689808e-01 -1.22434877e-01 -7.29952753e-01 -3.61896157e-01
1.72994941e-01 -4.64526415e-01 -2.08925217e-01 2.20032390e-02
-1.14539266e+00 9.30348992e-01 8.64496604e-02 5.19444823e-01
-2.60488987e-01 -5.80519259e-01 5.12477219e-01 6.94259465e-01
3.87473643e-01 1.00787568e+00 -5.68175793e-01 -9.23545063e-01
-5.28476596e-01 3.65405887e-01 1.59819710e+00 9.33881879e-01
7.18711495e-01 -1.95208594e-01 -6.60141349e-01 1.07550490e+00
2.84162045e-01 4.05970991e-01 4.08977032e-01 -1.30084229e+00
4.08727944e-01 1.10174632e+00 6.75346315e-01 -3.87365043e-01
-5.95615149e-01 4.89861071e-01 -3.06129426e-01 5.32015443e-01
1.02535439e+00 -3.87789607e-01 -2.18070135e-01 1.70357943e+00
2.13369355e-01 -9.32916343e-01 3.75323802e-01 7.41717696e-01
1.14700854e+00 4.52292383e-01 4.89438847e-02 -3.85387629e-01
1.50428462e+00 -1.11937690e+00 -1.35571909e+00 -1.22032948e-02
4.17535961e-01 -9.22052145e-01 1.67687249e+00 5.87046146e-01
-1.20301461e+00 -7.69598365e-01 -9.47603047e-01 2.95668915e-02
-4.12966847e-01 -1.30202606e-01 2.60065883e-01 8.72618139e-01
-1.17199945e+00 5.43810666e-01 -3.73097867e-01 -2.83158541e-01
-3.71076018e-01 3.37286472e-01 -2.11529687e-01 4.58986819e-01
-1.21239018e+00 1.25440824e+00 -9.42920595e-02 -2.85483360e-01
-6.70696020e-01 -3.31856340e-01 -7.60486424e-01 4.33128253e-02
3.83121550e-01 -1.92131102e-01 2.15511990e+00 -7.78269887e-01
-2.41187167e+00 6.52087629e-01 -2.14675263e-01 -1.05318084e-01
3.59338284e-01 -5.10527611e-01 -2.79963881e-01 -2.28345096e-01
-2.47449636e-01 4.77413028e-01 3.80962610e-01 -1.31860888e+00
-9.44343269e-01 -2.87942141e-01 6.60982072e-01 4.52770919e-01
-5.10481834e-01 7.91137144e-02 -2.60523885e-01 7.78901204e-02
-2.41483867e-01 -7.63067722e-01 -3.72979753e-02 -2.94401497e-01
-3.94789547e-01 -8.22050333e-01 6.61598861e-01 -5.74474275e-01
1.44526815e+00 -1.88408947e+00 1.70459375e-01 -1.18107051e-01
4.75929290e-01 2.48745129e-01 3.51829454e-02 6.04601681e-01
5.48337877e-01 1.33035863e-02 2.18520425e-02 -4.97760296e-01
5.85073352e-01 4.45073564e-03 -1.62634686e-01 -2.49082625e-01
1.20017134e-01 7.11159587e-01 -9.50833797e-01 -3.51655453e-01
5.41714370e-01 -1.59614563e-01 -3.21110785e-01 1.06328559e+00
-4.98501062e-01 6.86783493e-01 -5.19004703e-01 3.36744756e-01
2.41920575e-01 -2.09722206e-01 2.12070584e-01 -2.17802092e-01
-2.26100504e-01 7.95287013e-01 -9.59445298e-01 1.59380412e+00
-8.05537105e-01 4.23488587e-01 4.57901031e-01 -2.03634784e-01
1.17846847e+00 6.52687311e-01 3.60860914e-01 -5.64235926e-01
2.99616724e-01 -5.76144122e-02 1.30023628e-01 -4.72139299e-01
8.74617696e-01 6.57677129e-02 -5.82520962e-01 7.46069431e-01
4.80337113e-01 -5.59070528e-01 1.33782223e-01 3.72936785e-01
1.10898781e+00 -4.30142760e-01 4.87776011e-01 -1.96150050e-01
7.78306842e-01 -1.24935322e-02 2.30869845e-01 4.38755125e-01
-7.91794181e-01 2.71377265e-02 7.66221046e-01 -5.30009717e-02
-7.86562443e-01 -9.02521372e-01 -3.79202627e-02 1.41308081e+00
1.27767697e-01 -6.28344119e-01 -1.05017102e+00 -9.75037932e-01
-2.77222574e-01 7.54227281e-01 -1.12239406e-01 8.93293321e-02
-2.56400943e-01 -9.81337428e-02 4.95742679e-01 2.83663005e-01
6.16774857e-01 -1.26116681e+00 -1.30940646e-01 2.96300650e-01
-6.14515960e-01 -1.36962128e+00 -5.06123126e-01 7.99234807e-02
-5.16279042e-01 -9.39825356e-01 -2.21192822e-01 -4.56837595e-01
-8.33407640e-02 -7.56490752e-02 1.55086339e+00 -9.22364965e-02
4.47863430e-01 7.18351185e-01 -6.84288740e-01 -1.34011507e-01
-1.23500216e+00 3.41169201e-02 1.41362548e-01 -3.80762339e-01
4.81461138e-01 -1.79071650e-01 -3.77673984e-01 7.85405815e-01
-4.30723548e-01 -2.84845263e-01 3.39554965e-01 8.37749124e-01
-1.60288528e-01 -1.07048221e-01 8.68809104e-01 -8.52425873e-01
1.56753290e+00 9.15259123e-02 -4.60783392e-01 3.43266368e-01
-7.78552055e-01 1.48412928e-01 5.01349986e-01 -7.23049045e-02
-1.36714733e+00 -1.12128533e-01 -5.27410686e-01 2.55433559e-01
-6.21823132e-01 1.55241862e-01 -6.13334894e-01 2.23694310e-01
7.27823138e-01 -3.89913976e-01 2.87572742e-01 -2.53763705e-01
6.63853884e-01 1.23748648e+00 2.61161119e-01 -6.84791625e-01
3.25311869e-01 -1.91683397e-01 -6.05874598e-01 -8.60499799e-01
-7.73874104e-01 -7.68132329e-01 -8.95024776e-01 -5.74209690e-01
1.14361238e+00 -6.71453059e-01 -1.42163861e+00 3.67941976e-01
-1.12421882e+00 -7.70494998e-01 -2.99069166e-01 3.86543930e-01
-6.15884364e-01 3.43713492e-01 -9.22335804e-01 -1.33765280e+00
-3.37263823e-01 -1.48493934e+00 1.06628489e+00 3.83259863e-01
-8.53915513e-01 -1.09533048e+00 9.55655947e-02 8.33427668e-01
3.22730184e-01 -2.39769608e-01 8.28897595e-01 -1.04932332e+00
-1.31823137e-01 -1.87741786e-01 -1.36631578e-02 7.43934393e-01
4.73769218e-01 -1.09054886e-01 -1.23800170e+00 7.19217062e-02
2.51615256e-01 -1.03111279e+00 -6.14450388e-02 1.21770769e-01
3.05045247e-01 -3.09485823e-01 3.02894771e-01 -4.63004887e-01
7.06435740e-01 4.32944834e-01 2.20678851e-01 -1.72403932e-03
1.86905041e-01 1.06586015e+00 1.13613534e+00 6.53771341e-01
5.31249285e-01 1.05296195e+00 2.03783676e-01 1.17114941e-02
-2.98122521e-02 2.28103865e-02 6.66741312e-01 9.81885552e-01
2.61862487e-01 -4.64349926e-01 -8.00873041e-01 1.77569106e-01
-1.78849924e+00 -6.18630946e-01 -1.48523733e-01 2.00660253e+00
1.05159223e+00 3.05185437e-01 8.13721299e-01 8.84913802e-02
4.55656320e-01 8.50299820e-02 -3.07851315e-01 -6.08908951e-01
2.16218024e-01 -7.51869008e-02 -6.56852797e-02 1.17558217e+00
-8.87538910e-01 1.06584692e+00 6.65349531e+00 4.74905878e-01
-5.30808389e-01 1.01196356e-01 5.13525248e-01 3.79648328e-01
-1.28130596e-02 -1.93476528e-01 -9.73074377e-01 2.17660517e-01
1.10516107e+00 -1.65842369e-01 5.80400586e-01 9.17242408e-01
4.88026470e-01 -1.98238403e-01 -1.57666647e+00 5.15145481e-01
-1.83750257e-01 -6.96612477e-01 -3.32925498e-01 1.04271628e-01
3.26968342e-01 -4.86212224e-01 -9.15099010e-02 6.98836684e-01
9.67533827e-01 -1.01317883e+00 4.10458177e-01 4.29986149e-01
3.00786823e-01 -5.39951146e-01 9.87194896e-01 4.42569762e-01
-8.98484707e-01 3.11476160e-02 -8.26895088e-02 -9.29750726e-02
4.46616441e-01 4.65312511e-01 -1.29256344e+00 3.71551454e-01
4.46884215e-01 1.00689553e-01 -4.40800786e-01 2.48496547e-01
-1.74222216e-01 4.91925776e-01 1.08160168e-01 -2.30181217e-01
1.09112218e-01 -3.37004095e-01 3.77083242e-01 1.15968752e+00
-1.36601225e-01 2.27512568e-01 5.04336476e-01 8.46875370e-01
-8.72122198e-02 2.27410614e-01 -4.95734990e-01 -3.13967824e-01
5.83226085e-01 1.52082598e+00 -2.86502391e-01 -3.80553722e-01
-3.47947627e-01 8.53564620e-01 2.30500326e-01 -4.09379452e-02
-4.24343079e-01 -1.80946112e-01 8.41603041e-01 -4.26195413e-01
-3.60007077e-01 -1.21700116e-01 -4.70715046e-01 -7.16191292e-01
-1.97141767e-02 -1.35670686e+00 2.01209247e-01 -3.92736316e-01
-1.53419113e+00 1.08868718e+00 -2.23842323e-01 -1.18184555e+00
-6.27243400e-01 -7.70043194e-01 -6.56725645e-01 8.39788198e-01
-1.03959787e+00 -8.40972841e-01 -7.00023055e-01 5.19487560e-01
1.06824505e+00 -2.74505109e-01 1.22973478e+00 -1.01463959e-01
-2.89146125e-01 5.18842816e-01 -3.51287365e-01 1.75204977e-01
1.15251231e+00 -1.81617093e+00 1.66172057e-01 1.07159257e-01
-8.82727951e-02 3.86772782e-01 9.21033680e-01 -5.00591576e-01
-1.05307138e+00 -6.20611429e-01 9.51770246e-01 -8.94147456e-01
6.91742718e-01 -3.17741513e-01 -9.70305979e-01 1.85584754e-01
6.84102595e-01 -8.76965702e-01 8.33866537e-01 8.20196271e-01
-1.36043310e-01 5.79365529e-02 -1.26283514e+00 2.48383358e-01
6.66804671e-01 -7.29230464e-01 -8.74910831e-01 5.26142240e-01
8.24753761e-01 -3.45262021e-01 -1.45867646e+00 2.29613274e-01
4.31492478e-01 -1.23407197e+00 5.97815990e-01 -3.97614926e-01
2.89503217e-01 1.46841779e-02 -1.70829147e-01 -1.41903949e+00
8.45818892e-02 -6.56981230e-01 9.30819213e-02 1.57786238e+00
5.97796142e-01 -2.20726520e-01 6.84381843e-01 1.25068128e+00
-2.91427106e-01 -3.81152332e-01 -5.13086736e-01 -6.65348411e-01
2.19125256e-01 -1.48723423e-01 6.32762671e-01 6.54654324e-01
7.93429792e-01 1.21347237e+00 -4.98522282e-01 -6.68933392e-02
-3.25362459e-02 -7.59878680e-02 1.19319856e+00 -1.44509900e+00
-1.75074458e-01 -6.97995782e-01 -1.03203868e-02 -1.24036348e+00
3.07249516e-01 -2.67818928e-01 5.76465726e-01 -1.47569287e+00
-1.64332643e-01 -4.32087183e-01 1.43067464e-01 -3.30602378e-02
-3.84879440e-01 -4.09820080e-01 1.22239247e-01 1.26763538e-01
-8.83054793e-01 6.61823869e-01 1.26639521e+00 -6.02621622e-02
-6.29779756e-01 4.64580178e-01 -4.33240294e-01 7.14069784e-01
7.47471571e-01 1.72714189e-01 -4.66886193e-01 1.69857636e-01
-1.94474488e-01 7.58257210e-01 -1.77366942e-01 -9.36254025e-01
2.86166102e-01 -3.42606723e-01 -2.81346917e-01 -4.84765947e-01
7.06664920e-01 -6.98625863e-01 -4.51085329e-01 1.57517299e-01
-8.50182235e-01 1.79015994e-02 -1.53297469e-01 3.74822438e-01
-6.12560689e-01 -3.85797143e-01 7.09430039e-01 -1.41338438e-01
-4.27366525e-01 -1.56478852e-01 -7.51820266e-01 9.10922661e-02
6.82053864e-01 5.69519848e-02 -4.11713541e-01 -1.05669034e+00
-8.76346648e-01 4.19069558e-01 4.12963897e-01 5.01392484e-01
1.60396203e-01 -1.05731225e+00 -3.44563335e-01 -3.04373085e-01
3.46750706e-01 -3.31678838e-01 6.46874309e-02 4.58559066e-01
-9.81352553e-02 6.75146997e-01 -1.45438626e-01 -6.13872170e-01
-1.51239729e+00 2.27113217e-01 1.76838174e-01 -7.21640110e-01
-9.57311466e-02 6.97328746e-01 1.32142246e-01 -8.52886319e-01
8.76195610e-01 -3.93857837e-01 -5.65812409e-01 1.85126990e-01
5.87900400e-01 3.09151113e-01 2.53352940e-01 -6.11578107e-01
-2.40076575e-02 -9.37578604e-02 1.21305756e-01 -6.50704384e-01
7.84068465e-01 -4.16304082e-01 1.34821728e-01 1.00320458e+00
6.79486811e-01 3.71537097e-02 -1.05715656e+00 -4.29185212e-01
4.09370720e-01 -2.51389891e-01 -3.09471220e-01 -1.21988893e+00
-3.83907497e-01 9.25647795e-01 6.53559804e-01 7.13272214e-01
7.86014795e-01 6.74964041e-02 7.02321649e-01 8.07202995e-01
3.54963332e-01 -1.51573646e+00 8.35170150e-01 8.81627798e-01
1.02351665e+00 -1.68567705e+00 -3.72558534e-01 -5.40977061e-01
-1.34737277e+00 1.03029406e+00 1.01339388e+00 4.80770320e-01
4.23200279e-01 1.98011845e-01 5.56905925e-01 -2.11379766e-01
-8.81585538e-01 -3.82932156e-01 4.03076231e-01 5.18062115e-01
8.97120237e-01 1.64251745e-01 -2.83286393e-01 8.06782126e-01
-5.77825367e-01 -3.19194764e-01 3.72026712e-01 7.12393343e-01
-8.54990900e-01 -1.54993939e+00 -2.27623686e-01 3.29150319e-01
-1.27505675e-01 2.28016108e-01 -1.05787098e+00 6.02774799e-01
-4.27664548e-01 1.93124676e+00 -3.17491651e-01 -6.96259975e-01
8.06823611e-01 6.15417123e-01 5.99822924e-02 -8.33939075e-01
-1.32766831e+00 -9.99015868e-02 7.79325962e-01 -6.05904996e-01
-3.91964167e-01 -3.30527455e-01 -9.59678233e-01 -3.97575289e-01
-5.32792211e-01 5.72848141e-01 6.19871736e-01 1.08872199e+00
-1.09256946e-01 6.35225236e-01 8.54622960e-01 -4.18335021e-01
-8.96864355e-01 -1.54552877e+00 -4.93366957e-01 7.21245885e-01
2.07445785e-01 -6.39394045e-01 -3.64904374e-01 -3.51157854e-03] | [12.870697021484375, 8.016271591186523] |
6c73eb1d-7d2a-4d3d-8cae-94d3c4482a17 | nmtpy-a-flexible-toolkit-for-advanced-neural | 1706.00457 | null | http://arxiv.org/abs/1706.00457v1 | http://arxiv.org/pdf/1706.00457v1.pdf | NMTPY: A Flexible Toolkit for Advanced Neural Machine Translation Systems | In this paper, we present nmtpy, a flexible Python toolkit based on Theano
for training Neural Machine Translation and other neural sequence-to-sequence
architectures. nmtpy decouples the specification of a network from the training
and inference utilities to simplify the addition of a new architecture and
reduce the amount of boilerplate code to be written. nmtpy has been used for
LIUM's top-ranked submissions to WMT Multimodal Machine Translation and News
Translation tasks in 2016 and 2017. | ['Loïc Barrault', 'Mercedes García-Martínez', 'Walid Aransa', 'Ozan Caglayan', 'Adrien Bardet', 'Fethi Bougares'] | 2017-06-01 | null | null | null | null | ['multimodal-machine-translation'] | ['natural-language-processing'] | [ 1.62313864e-01 2.84715593e-01 -3.24778587e-01 -6.31147981e-01
-9.90441322e-01 -5.71374178e-01 6.86662257e-01 -3.38423103e-01
-4.17374760e-01 9.11672473e-01 2.66582936e-01 -1.07211995e+00
5.69189787e-01 -3.60561252e-01 -1.11070657e+00 -1.95348382e-01
4.78086382e-01 9.03564751e-01 -3.62822741e-01 -3.05209011e-01
-2.67659813e-01 -1.16780266e-01 -8.37422371e-01 7.57180989e-01
7.17940629e-01 5.40493548e-01 2.59077400e-01 7.63739109e-01
-1.85613945e-01 3.80057991e-01 -4.55597997e-01 -8.13863516e-01
1.25587761e-01 -7.09585845e-01 -1.10837758e+00 -6.13459468e-01
3.90604556e-01 -1.92430183e-01 -1.66196927e-01 6.24064088e-01
8.23088646e-01 -1.03613101e-01 5.22629440e-01 -1.03803194e+00
-9.57632363e-01 9.55131114e-01 -7.90988803e-02 1.38065174e-01
1.16937585e-01 3.40628564e-01 7.23281503e-01 -1.24368727e+00
8.95186365e-01 1.22945201e+00 6.44683778e-01 9.51937199e-01
-1.21074164e+00 -5.57088256e-01 -3.94008011e-01 -5.46032339e-02
-8.49969208e-01 -8.58043313e-01 1.73918754e-01 -3.10466260e-01
1.73346090e+00 3.47494513e-01 2.95266569e-01 1.75620401e+00
1.85971498e-01 8.51632655e-01 8.02905977e-01 -5.05181909e-01
2.20794659e-02 -8.17024261e-02 -8.92770812e-02 6.24264240e-01
-3.29344988e-01 5.89088239e-02 -6.02011859e-01 -1.81665659e-01
5.03687143e-01 -7.79963374e-01 5.26589900e-02 2.22221583e-01
-1.46289182e+00 7.32399225e-01 1.47495702e-01 2.41742536e-01
-2.28742763e-01 2.89178818e-01 8.28447342e-01 5.43306887e-01
5.94842136e-01 6.44931853e-01 -7.19104230e-01 -6.93970144e-01
-8.12074065e-01 1.58399120e-01 8.95833790e-01 1.09002912e+00
3.98250550e-01 2.79825538e-01 -2.39139020e-01 1.10880625e+00
7.89089650e-02 4.20991093e-01 6.36833429e-01 -9.80900049e-01
9.62793529e-01 2.97675550e-01 -8.65176395e-02 -2.71119267e-01
-2.65449971e-01 -3.50039691e-01 -6.66194797e-01 -3.19687039e-01
3.13259155e-01 -7.15536892e-01 -9.71156657e-01 1.59580004e+00
-3.08132451e-02 -2.24667802e-01 3.13928545e-01 9.54828501e-01
1.16429389e+00 1.09699225e+00 -1.02918200e-01 1.30951628e-01
8.75718117e-01 -1.26684248e+00 -5.61575532e-01 -4.46180910e-01
8.95138562e-01 -1.11736810e+00 1.16435790e+00 -3.73793356e-02
-1.53436399e+00 -3.09854567e-01 -1.00581837e+00 -4.48887795e-01
-5.44020712e-01 3.09561402e-01 6.36120558e-01 6.98187351e-02
-1.40572870e+00 6.62622988e-01 -9.41752613e-01 -6.99555039e-01
6.07789420e-02 4.64753598e-01 -3.66863966e-01 1.16649546e-01
-1.52083635e+00 1.67633379e+00 7.42709279e-01 2.42024064e-01
-6.90573156e-01 -5.80793679e-01 -7.68543303e-01 -9.05806497e-02
-2.08341360e-01 -1.07434821e+00 1.87160599e+00 -1.08403838e+00
-1.99028802e+00 1.01050508e+00 -3.48457336e-01 -6.07383490e-01
5.43048382e-01 -1.80986807e-01 -3.11156929e-01 -2.38399327e-01
-8.07975158e-02 1.03245914e+00 3.74822557e-01 -4.77495372e-01
-1.75985083e-01 1.86753333e-01 -2.40183637e-01 2.85125673e-01
-3.25761549e-02 5.71838915e-01 -5.25218546e-01 -3.89858842e-01
-4.84338939e-01 -9.57281888e-01 1.05068982e-01 -4.57581162e-01
-4.62361783e-01 -3.22368979e-01 6.26830757e-01 -1.02706277e+00
7.13792503e-01 -1.89681387e+00 6.67214692e-01 -3.07745159e-01
-2.32546717e-01 4.05308485e-01 -5.66851676e-01 7.02676594e-01
-4.71411571e-02 8.81621912e-02 -3.82484078e-01 -5.89751780e-01
2.07970753e-01 4.92307097e-01 -3.88653219e-01 1.26132935e-01
4.22711849e-01 1.48295355e+00 -5.81549644e-01 -2.35840395e-01
-1.02120116e-02 5.32596469e-01 -1.72053814e-01 4.19797152e-01
-6.59542620e-01 4.95381892e-01 4.08365279e-02 4.60314989e-01
3.96782637e-01 -2.02116936e-01 1.22947417e-01 2.87565380e-01
-2.85328329e-01 1.05223656e+00 -8.68441686e-02 2.09639478e+00
-6.43146098e-01 8.31477761e-01 2.05452114e-01 -5.11238337e-01
6.92173183e-01 7.64703870e-01 -3.30152772e-02 -6.36816502e-01
2.10369632e-01 5.04829645e-01 1.36130303e-02 -4.88190740e-01
6.57953858e-01 1.74854696e-02 -9.41659287e-02 7.35096931e-01
5.56492090e-01 3.00055016e-02 1.82513893e-01 -6.14271201e-02
5.99281311e-01 6.86807454e-01 1.25883138e-02 -2.32904971e-01
5.70237190e-02 2.79595524e-01 2.01399431e-01 4.40308809e-01
3.42949718e-01 3.93342406e-01 3.57742310e-01 -5.01878083e-01
-1.60732710e+00 -9.04286683e-01 4.55402676e-03 1.67005908e+00
-6.73352778e-01 -3.71091664e-01 -1.05409193e+00 -3.97204429e-01
-3.58195633e-01 8.93085837e-01 -5.00084162e-01 8.56116638e-02
-8.17970932e-01 -8.68212342e-01 1.31298518e+00 3.80981475e-01
2.60482997e-01 -1.21972477e+00 -2.82970428e-01 1.06907018e-01
-6.40907824e-01 -1.09922504e+00 -6.31458461e-01 4.63074028e-01
-8.61396432e-01 -3.18557769e-01 -7.05395699e-01 -9.95026946e-01
3.74009877e-01 -3.05777013e-01 1.35068893e+00 -2.53989309e-01
1.96035147e-01 -2.79901356e-01 -1.70969084e-01 -5.61915278e-01
-8.22571814e-01 7.80274630e-01 -1.22019621e-02 -6.25952780e-01
4.63254184e-01 -4.92820084e-01 -7.63688162e-02 -9.78680477e-02
-6.62714124e-01 7.54537344e-01 5.82802832e-01 9.52176452e-01
7.31773898e-02 -1.05696619e+00 4.48593080e-01 -5.58127820e-01
1.00031841e+00 -3.76695961e-01 -5.78350723e-01 2.24568009e-01
-5.28792322e-01 -3.38190049e-02 6.50176287e-01 -3.88854235e-01
-7.84744978e-01 -1.99423015e-01 -2.65209198e-01 -1.13877982e-01
7.16136619e-02 8.10054600e-01 7.25366324e-02 2.00315639e-01
5.93340456e-01 4.14903373e-01 1.75403804e-01 -5.12054861e-01
6.53511047e-01 8.94325018e-01 7.87251294e-01 -7.10114956e-01
3.79960060e-01 -4.34384465e-01 -3.92493397e-01 -3.96949291e-01
-3.14716548e-01 1.62030175e-01 -6.04584754e-01 9.57593843e-02
9.33984876e-01 -8.96666765e-01 -5.91668427e-01 2.15978771e-01
-1.68649030e+00 -6.73364580e-01 1.60443366e-01 4.56493884e-01
-5.83656192e-01 -9.07619148e-02 -1.21716499e+00 -1.63721398e-01
-9.93349254e-01 -1.28820503e+00 8.58620226e-01 -7.53405690e-02
-5.58085024e-01 -9.78646159e-01 2.31489047e-01 1.94245502e-01
8.27741206e-01 -1.24087751e-01 1.03103137e+00 -6.36808872e-01
-3.09615374e-01 -2.35575959e-02 -1.96746930e-01 3.73310059e-01
-4.59980518e-01 3.81452739e-02 -9.00024831e-01 -2.30736986e-01
-2.87588060e-01 -6.24950647e-01 6.89871252e-01 2.61036128e-01
6.38154268e-01 -6.18828773e-01 -2.16961559e-02 7.98699558e-01
8.87381852e-01 -5.96407913e-02 5.09678125e-01 5.24939656e-01
5.74338555e-01 5.61529815e-01 -8.47182199e-02 -1.66326135e-01
3.94704133e-01 8.42446625e-01 3.01219851e-01 -3.50755274e-01
1.74050499e-02 -3.94905746e-01 8.97416472e-01 1.32482791e+00
2.61822366e-03 -2.22626850e-01 -9.84097242e-01 4.20665443e-01
-2.01825833e+00 -6.47242963e-01 -1.46768898e-01 1.81678164e+00
1.24472857e+00 -2.65497446e-01 1.88364331e-02 -7.69650340e-01
5.64623117e-01 -1.82415545e-01 -3.62059295e-01 -1.27494502e+00
-2.28743315e-01 4.27645355e-01 3.80936682e-01 8.31532180e-01
-7.80669332e-01 1.37846947e+00 7.70643711e+00 6.31410301e-01
-1.28355062e+00 4.99062210e-01 4.88355547e-01 -4.01507467e-01
-2.35415101e-01 7.47124553e-02 -6.44772589e-01 4.31242645e-01
1.59973466e+00 -1.81466118e-01 8.24190617e-01 6.08546376e-01
2.06529036e-01 3.89752597e-01 -1.34810126e+00 6.21060431e-01
-1.92818027e-02 -1.50032330e+00 2.04731692e-02 -1.17806360e-01
4.85489607e-01 9.81675506e-01 8.30025896e-02 4.83334154e-01
6.20291233e-01 -1.53121448e+00 5.86750209e-01 2.78932959e-01
9.37547266e-01 -6.51417255e-01 8.10717702e-01 2.05814183e-01
-3.10263455e-01 3.51772636e-01 -4.16442811e-01 -2.58507103e-01
2.91169792e-01 3.19823951e-01 -1.22864699e+00 5.46496451e-01
4.14789468e-01 5.72318912e-01 -2.91507989e-01 5.58541894e-01
-3.77474576e-01 9.37966585e-01 -2.11112142e-01 -1.26781717e-01
4.85371351e-01 -4.84397560e-02 5.02465069e-01 1.74078548e+00
3.74038845e-01 -4.88370955e-01 5.13030998e-02 9.78104293e-01
-4.05390948e-01 2.16759548e-01 -5.54566920e-01 -2.45390788e-01
4.14490491e-01 1.12025046e+00 1.01330569e-02 -5.58833539e-01
-3.17064166e-01 1.28365481e+00 4.16540563e-01 5.67593992e-01
-7.98656523e-01 -2.21645728e-01 6.74697876e-01 -1.33729637e-01
5.12692966e-02 -5.06206870e-01 -4.44641054e-01 -1.33245301e+00
5.00487909e-02 -1.16452980e+00 2.06765141e-02 -1.14987278e+00
-1.06990743e+00 9.90448296e-01 -1.28869489e-01 -8.44805181e-01
-7.93929338e-01 -6.46340966e-01 -5.73626041e-01 1.59115040e+00
-1.01369810e+00 -1.38972723e+00 3.35481554e-01 1.40171677e-01
5.39102554e-01 -3.49107265e-01 1.37422311e+00 3.36572587e-01
-6.18145883e-01 8.38848948e-01 1.80772752e-01 3.96452457e-01
8.51298034e-01 -1.03411615e+00 1.32634902e+00 7.76728094e-01
-2.69565254e-01 9.24910188e-01 6.83131158e-01 -4.85898286e-01
-1.40781355e+00 -1.06297350e+00 1.47213244e+00 -7.58176684e-01
8.16354871e-01 -6.75828755e-01 -6.65282071e-01 1.04862952e+00
1.18295586e+00 -6.43748820e-01 5.42926967e-01 1.50505394e-01
-4.88328665e-01 2.71872997e-01 -6.19643927e-01 7.16617346e-01
6.69409871e-01 -9.02638912e-01 -5.54294467e-01 5.23867726e-01
1.04863667e+00 -8.72212708e-01 -1.04979658e+00 3.71092051e-01
6.59990728e-01 -2.37434253e-01 6.22165799e-01 -9.10524964e-01
8.78952265e-01 -2.03265250e-01 -4.67701554e-02 -1.55144334e+00
-1.89739555e-01 -9.64951754e-01 -2.42592111e-01 9.78255212e-01
1.29230046e+00 -6.94711685e-01 3.38739663e-01 3.26725483e-01
-5.21601021e-01 -7.80943036e-01 -1.17144263e+00 -5.74406683e-01
6.59124553e-01 -2.55259782e-01 6.15590394e-01 1.20776355e+00
3.75295997e-01 8.30288172e-01 -6.54147625e-01 -4.48521256e-01
2.34278999e-02 -2.82714158e-01 8.49327087e-01 -5.52430987e-01
-7.91687727e-01 -6.06989503e-01 7.68150985e-02 -1.01343226e+00
1.64281473e-01 -1.56119311e+00 2.09045812e-01 -1.62562180e+00
3.17192078e-01 2.95121878e-01 4.15162221e-02 1.19182372e+00
1.61508918e-01 3.32530797e-01 2.56870449e-01 3.36007506e-01
-2.08398879e-01 4.03606445e-01 1.01735032e+00 -2.11618066e-01
1.85452234e-02 -2.17965379e-01 -6.23833656e-01 1.23599432e-01
9.66764450e-01 -4.47328806e-01 -1.47428617e-01 -1.39829755e+00
3.82631660e-01 -2.75172107e-02 1.71016395e-01 -4.63415056e-01
-5.34230173e-02 7.31303915e-02 4.11042154e-01 -4.84103501e-01
3.64205331e-01 -1.68425187e-01 1.41017705e-01 2.29181305e-01
-7.15719402e-01 7.88778186e-01 6.06311917e-01 -4.44958359e-01
-1.61307707e-01 3.82796489e-02 5.24872839e-01 -2.91616768e-01
-5.18654510e-02 -2.16094572e-02 -5.88253856e-01 -1.98314130e-01
4.33026850e-01 1.41365409e-01 -7.59042680e-01 -4.42800760e-01
-6.83646739e-01 3.56611580e-01 5.46359360e-01 8.03395689e-01
3.24419856e-01 -1.22956073e+00 -1.00229800e+00 -1.12047844e-01
1.06550932e-01 -2.88670212e-01 -2.09880576e-01 1.00777328e+00
-5.81010699e-01 8.26262832e-01 -3.35709214e-01 -4.48346198e-01
-1.08148813e+00 1.32512793e-01 4.60079134e-01 -2.25225583e-01
-6.36522710e-01 6.95670426e-01 -3.22144032e-01 -1.29101443e+00
3.81238274e-02 -1.26106709e-01 2.84256309e-01 -4.96457696e-01
5.10473847e-01 1.01383910e-01 3.65099996e-01 -5.61071754e-01
-2.17520893e-01 -8.91230628e-03 -2.81314194e-01 -7.12262034e-01
1.22819138e+00 1.23142615e-01 -6.45906091e-01 5.98819971e-01
1.13693750e+00 -5.39690971e-01 -7.22569168e-01 -1.23180419e-01
1.56526923e-01 3.86946082e-01 -1.70552373e-01 -1.59078670e+00
-4.83401090e-01 1.13029361e+00 3.64704758e-01 -5.13804197e-01
7.82149673e-01 -1.27744883e-01 1.23744392e+00 6.23251736e-01
-9.57273785e-03 -1.02785206e+00 -7.07404137e-01 1.24644554e+00
8.99966478e-01 -8.21485281e-01 -3.50342244e-01 1.02121547e-01
-5.28431833e-01 1.28033686e+00 6.54829443e-01 4.04183239e-01
-5.60868010e-02 4.37586784e-01 5.04282892e-01 2.10226640e-01
-1.24687886e+00 2.90123552e-01 3.82526457e-01 5.42668104e-01
1.06489968e+00 3.30954105e-01 -4.28508639e-01 4.21465248e-01
-6.32653236e-01 2.15171516e-01 2.49130577e-01 5.96467793e-01
-7.87459463e-02 -1.41333604e+00 -1.15731418e-01 3.78495343e-02
-6.02463365e-01 -7.82318115e-01 -9.25288975e-01 5.51221728e-01
-9.60131586e-02 7.02027917e-01 7.32124671e-02 -5.95406413e-01
9.56236199e-03 5.11096001e-01 5.69823444e-01 -5.09785354e-01
-1.00204170e+00 -3.25864144e-02 6.93994045e-01 -4.50393736e-01
-9.85893160e-02 -3.49072933e-01 -1.11958861e+00 -6.45806491e-01
1.44223005e-01 2.89729357e-01 1.15324867e+00 1.03969991e+00
8.02733362e-01 3.03830892e-01 -7.43769333e-02 -8.81903350e-01
-4.75478083e-01 -1.62870681e+00 4.42964196e-01 -1.90595314e-01
7.64380693e-02 8.33119452e-02 2.85478681e-01 -5.90581633e-02] | [11.605818748474121, 10.3318510055542] |
c7889377-51ff-40c3-a0d6-e4d4f8a79db3 | average-outward-flux-skeletons-for | 2111.13826 | null | https://arxiv.org/abs/2111.13826v1 | https://arxiv.org/pdf/2111.13826v1.pdf | Average Outward Flux Skeletons for Environment Mapping and Topology Matching | We consider how to directly extract a road map (also known as a topological representation) of an initially-unknown 2-dimensional environment via an online procedure that robustly computes a retraction of its boundaries. In this article, we first present the online construction of a topological map and the implementation of a control law for guiding the robot to the nearest unexplored area, first presented in [1]. The proposed method operates by allowing the robot to localize itself on a partially constructed map, calculate a path to unexplored parts of the environment (frontiers), compute a robust terminating condition when the robot has fully explored the environment, and achieve loop closure detection. The proposed algorithm results in smooth safe paths for the robot's navigation needs. The presented approach is any time algorithm that has the advantage that it allows for the active creation of topological maps from laser scan data, as it is being acquired. We also propose a navigation strategy based on a heuristic where the robot is directed towards nodes in the topological map that open to empty space. We then extend the work in [1] by presenting a topology matching algorithm that leverages the strengths of a particular spectral correspondence method [2], to match the mapped environments generated from our topology-making algorithm. Here, we concentrated on implementing a system that could be used to match the topologies of the mapped environment by using AOF Skeletons. In topology matching between two given maps and their AOF skeletons, we first find correspondences between points on the AOF skeletons of two different environments. We then align the (2D) points of the environments themselves. We also compute a distance measure between two given environments, based on their extracted AOF skeletons and their topology, as the sum of the matching errors between corresponding points. | ['Kaleem Siddiqi', 'Gregory Dudek', 'Ioannis Rekleitis', 'Elham Karimi', 'Babak Samari', 'Morteza Rezanejad'] | 2021-11-27 | null | null | null | null | ['loop-closure-detection'] | ['computer-vision'] | [ 4.57709730e-01 5.66456318e-01 3.70027184e-01 -2.60516733e-01
-2.40502372e-01 -6.52898252e-01 5.89637756e-01 3.97069991e-01
-5.00645578e-01 5.50706565e-01 -3.67231816e-01 -3.53390038e-01
-7.43043959e-01 -1.15913391e+00 -8.85628819e-01 -3.43879968e-01
-3.80506337e-01 1.07967639e+00 5.57689011e-01 -5.18229187e-01
6.09757304e-01 9.91682053e-01 -1.73539460e+00 -5.44939876e-01
9.13073778e-01 7.20020115e-01 5.58695018e-01 6.00212872e-01
-2.56640822e-01 -2.20036119e-01 9.94336531e-02 1.28335923e-01
6.10654414e-01 -1.42014086e-01 -9.58552837e-01 3.17417122e-02
3.03633749e-01 4.62183580e-02 -1.36355340e-01 1.00472581e+00
-1.16525002e-01 3.05823654e-01 7.98938930e-01 -1.24341083e+00
2.15559885e-01 2.36847848e-01 -2.57276833e-01 -3.62623721e-01
5.68005800e-01 -8.71001855e-02 3.71544838e-01 -7.82265007e-01
1.04291105e+00 9.24694479e-01 5.85983217e-01 3.52237299e-02
-1.13748729e+00 -2.14451000e-01 -1.40779018e-01 1.05781637e-01
-1.49339926e+00 -4.26863819e-01 8.64046752e-01 -4.75937605e-01
6.67482793e-01 8.28309283e-02 5.98984122e-01 6.04876041e-01
3.31051081e-01 -9.74805579e-02 9.91716921e-01 -8.20566356e-01
5.73650956e-01 -1.49164414e-02 -1.66756973e-01 8.08183849e-01
2.91433990e-01 3.15467119e-01 -1.97371021e-01 -5.93835376e-02
8.49268794e-01 -2.96329200e-01 -5.32402769e-02 -1.26415300e+00
-1.11956620e+00 5.01425564e-01 5.25166810e-01 4.14007068e-01
-3.67983222e-01 -8.43280181e-02 3.50694582e-02 3.23771626e-01
-1.22086756e-01 5.25606275e-01 -5.64847551e-02 -4.15715799e-02
-5.17032385e-01 3.13798785e-01 9.34996903e-01 1.09477842e+00
1.41294706e+00 -4.17410344e-01 7.38398433e-01 2.43633002e-01
1.65088981e-01 4.98576194e-01 5.63598238e-02 -1.11626887e+00
5.35876870e-01 7.69126832e-01 3.36269736e-01 -1.49083447e+00
-6.12021744e-01 7.48630837e-02 -4.06020135e-01 5.83228827e-01
5.24037123e-01 1.82740688e-01 -7.29944766e-01 1.61624551e+00
6.00564003e-01 -4.42229882e-02 2.68540859e-01 4.93786544e-01
-1.40270293e-01 2.00702831e-01 -5.55509269e-01 3.07166614e-02
9.70207334e-01 -5.85566521e-01 -3.46863121e-01 -3.35917115e-01
6.05385184e-01 -4.49032128e-01 6.83666468e-01 2.84470350e-01
-8.93387496e-01 -4.19903934e-01 -1.27515996e+00 8.88444558e-02
-6.59829080e-01 -2.29444318e-02 4.72354777e-02 2.84246415e-01
-1.33810747e+00 8.82642210e-01 -8.23281527e-01 -9.17689383e-01
-1.68049186e-01 2.97474653e-01 -7.22867072e-01 1.30266637e-01
-8.56662631e-01 1.29399431e+00 8.32413614e-01 3.00016612e-01
-5.68402231e-01 -8.13084468e-02 -9.11299169e-01 -2.44883776e-01
6.33623302e-01 -6.22566402e-01 7.54032671e-01 -5.79545140e-01
-1.45930266e+00 7.41311550e-01 -7.67540038e-02 -4.24849480e-01
7.61832952e-01 -2.85983295e-03 -8.63299221e-02 2.45186552e-01
3.02213907e-01 5.29347301e-01 5.33653319e-01 -1.36358011e+00
-6.76757395e-01 -5.58251619e-01 -6.40647635e-02 4.02709872e-01
1.16995357e-01 -6.02396548e-01 -3.72424841e-01 -1.93978231e-02
9.66319680e-01 -1.05852127e+00 -6.27023995e-01 1.27544820e-01
-4.03589189e-01 1.51638672e-01 7.77688682e-01 -4.78753954e-01
7.95905769e-01 -2.08872533e+00 3.89632910e-01 9.52251554e-01
-6.35271370e-02 -3.35018009e-01 -8.19835514e-02 8.09162736e-01
2.81077921e-01 -4.40472364e-02 -7.53535748e-01 -1.32314637e-01
-1.29398420e-01 3.89180750e-01 -2.96229303e-01 7.22478092e-01
-2.11721256e-01 3.54672670e-01 -1.01486540e+00 -2.80195534e-01
3.22897077e-01 1.44098952e-01 -1.66607350e-01 -1.53231313e-02
6.52018860e-02 4.32881296e-01 -5.80010653e-01 2.04132721e-01
7.60310650e-01 7.10908771e-01 1.51274979e-01 7.24570304e-02
-8.19809616e-01 1.37405947e-01 -1.60892284e+00 1.97435224e+00
-3.93242508e-01 4.58430618e-01 2.88129121e-01 -9.41785455e-01
1.56869650e+00 -6.80325106e-02 4.89323586e-01 -5.45662820e-01
1.69477642e-01 8.04499805e-01 -4.67277974e-01 -3.73715043e-01
7.11726248e-01 2.14880601e-01 -1.45399898e-01 3.85990173e-01
-3.70479882e-01 -4.38373774e-01 1.31974608e-01 -1.22071393e-01
1.27558923e+00 4.97668624e-01 4.11411971e-01 -4.04995859e-01
7.05945790e-01 4.16702658e-01 7.05676898e-02 7.30226934e-01
1.15447477e-01 3.11673135e-01 1.80464119e-01 -4.52160150e-01
-1.29109776e+00 -1.27514696e+00 -1.31105155e-01 2.60599941e-01
7.73096263e-01 1.68098994e-02 -7.61573434e-01 -3.41317624e-01
7.72897527e-02 5.61380267e-01 -5.96579313e-01 -4.98693623e-02
-8.99751723e-01 1.12274617e-01 3.37821543e-01 9.46176127e-02
5.04292190e-01 -7.78097034e-01 -1.51417422e+00 3.23851824e-01
-1.81652278e-01 -9.04829979e-01 1.53013229e-01 4.27789301e-01
-1.21623647e+00 -1.21834433e+00 -1.87118620e-01 -8.33381355e-01
9.88230288e-01 2.79350311e-01 4.23095018e-01 -1.94993988e-01
-1.28865659e-01 3.86388898e-01 -3.30660433e-01 -5.71993887e-02
-5.38775861e-01 9.91014838e-02 2.40778089e-01 -5.39438538e-02
-1.45283699e-01 -8.46194625e-01 -1.95462272e-01 6.73133194e-01
-5.82424700e-01 9.55728665e-02 5.64814508e-01 2.85302013e-01
8.79023731e-01 3.74017328e-01 1.00349344e-01 -4.20568138e-01
2.08998159e-01 -5.46622157e-01 -8.64195406e-01 4.75079454e-02
-5.98353565e-01 4.15622532e-01 4.73236501e-01 -9.00844783e-02
-7.25040376e-01 6.26412511e-01 2.18250990e-01 -2.75520951e-01
-3.28423381e-01 4.89902109e-01 -8.27053115e-02 -3.03180605e-01
7.31423438e-01 9.95960161e-02 2.24829018e-01 -6.34980619e-01
5.28345287e-01 4.71394777e-01 9.70937610e-01 -5.12446523e-01
1.15586519e+00 8.65679920e-01 5.37524462e-01 -6.12571657e-01
1.16163455e-01 -5.88568509e-01 -1.30885017e+00 -3.00403804e-01
5.89844823e-01 -1.96411520e-01 -4.89877343e-01 1.22952372e-01
-1.18545699e+00 -9.46589261e-02 -2.87435442e-01 4.13950562e-01
-1.13019967e+00 4.06521112e-01 1.89350456e-01 -9.38355863e-01
4.06177603e-02 -9.32891488e-01 7.88369417e-01 1.18492432e-02
-1.16742708e-01 -8.65180612e-01 5.95883071e-01 -2.60082334e-01
1.66142017e-01 8.20078969e-01 7.53050327e-01 -4.86072361e-01
-6.75681174e-01 -1.74557805e-01 1.00074060e-01 -4.07580465e-01
6.97831139e-02 2.96381377e-02 -7.47719228e-01 -2.59171426e-01
5.47653362e-02 3.06902170e-01 4.77213770e-01 1.26051337e-01
1.76285371e-01 -1.09417751e-01 -8.81005466e-01 4.77122933e-01
1.68504429e+00 3.90889525e-01 5.88467777e-01 1.07732928e+00
2.59662896e-01 1.28087449e+00 9.64862883e-01 6.43891692e-02
3.42625588e-01 9.26999927e-01 8.75006497e-01 9.59014595e-02
3.13912392e-01 -5.13735116e-01 2.78010488e-01 5.11313915e-01
-2.52617318e-02 2.96834648e-01 -1.15470994e+00 7.02780545e-01
-2.06205797e+00 -6.00736737e-01 -3.89956951e-01 2.59308219e+00
1.68728903e-01 2.52280235e-01 -4.77660112e-02 1.90433562e-01
8.27416658e-01 -4.23842818e-01 -3.71088952e-01 -7.24998951e-01
1.53104082e-01 7.82330707e-02 9.18913662e-01 9.17591453e-01
-9.67952073e-01 6.61544681e-01 5.52712774e+00 1.49410173e-01
-8.05068493e-01 -3.19507003e-01 -2.77184516e-01 4.06738549e-01
-1.86108142e-01 5.84914625e-01 -4.51258481e-01 2.84210891e-02
9.97115731e-01 -2.06260666e-01 3.97904843e-01 9.84359443e-01
3.17857772e-01 -6.08998120e-01 -1.10516369e+00 4.25601989e-01
-6.44816458e-02 -1.05918074e+00 -2.47019023e-01 3.51232350e-01
4.12581086e-01 -5.89675568e-02 -2.49700502e-01 -2.06342101e-01
-7.44065046e-02 -6.95274293e-01 1.10185683e+00 7.50698984e-01
7.21190214e-01 -7.24938154e-01 5.45274079e-01 7.34188139e-01
-1.51219785e+00 -1.74226075e-01 -3.36826652e-01 -1.64887756e-01
4.02007043e-01 3.39572728e-01 -1.40560138e+00 1.08276284e+00
5.53371072e-01 2.24891186e-01 -3.66515398e-01 1.24122989e+00
4.55306992e-02 -3.76821697e-01 -7.44980335e-01 1.05687417e-01
2.57860959e-01 -7.54466534e-01 9.43093538e-01 8.31912398e-01
6.98305726e-01 -2.33050674e-01 2.55082637e-01 9.97112513e-01
5.33097982e-01 1.98276654e-01 -1.15708983e+00 5.21966517e-01
7.42467523e-01 1.02392125e+00 -1.22296607e+00 -6.78005442e-02
1.60702392e-01 7.65962005e-01 2.07904339e-01 1.48466066e-01
-4.63221282e-01 -8.82775426e-01 2.18009993e-01 2.33975470e-01
4.92427126e-02 -6.34326041e-01 -3.22628707e-01 -4.39407527e-01
2.13627532e-01 -1.15214124e-01 -7.32832178e-02 -7.08251715e-01
-3.77443284e-01 6.08370841e-01 2.90138334e-01 -1.32502735e+00
-4.27002966e-01 -4.65484709e-01 -5.67852437e-01 8.54688823e-01
-1.40579629e+00 -6.76205933e-01 -4.65605974e-01 3.43422264e-01
2.68074960e-01 2.79046237e-01 7.44390428e-01 -1.75880745e-01
4.27586213e-02 -1.21618800e-01 3.09285879e-01 -2.34311804e-01
4.18537349e-01 -1.12044215e+00 4.70596582e-01 9.75047231e-01
-1.02836140e-01 7.00249076e-01 9.03896630e-01 -1.01823950e+00
-1.28694153e+00 -7.92420387e-01 8.12623680e-01 -3.38319272e-01
6.05655551e-01 -3.39519709e-01 -8.34311068e-01 6.60134256e-01
-4.91023034e-01 -5.06731331e-01 -2.96626508e-01 -3.98773909e-01
1.14797875e-01 -5.79019776e-03 -1.37014627e+00 5.98731697e-01
1.25260913e+00 -3.05174321e-01 -7.83408463e-01 -5.54816313e-02
3.96142751e-01 -5.99116981e-01 -5.65933526e-01 5.13196588e-01
6.96873605e-01 -1.07118857e+00 7.47724950e-01 -4.25183289e-02
-1.72954544e-01 -7.49351680e-01 -9.63981897e-02 -1.27866495e+00
-3.11420746e-02 -7.17259347e-01 4.24490541e-01 8.70512486e-01
2.44190127e-01 -8.44501376e-01 8.11166942e-01 9.03928876e-02
-4.06194597e-01 -5.30027866e-01 -1.33498430e+00 -9.04742599e-01
-2.24479035e-01 -3.80899966e-01 4.89673257e-01 5.64537466e-01
1.02739297e-01 -2.40881905e-01 2.84057140e-01 6.05604351e-01
6.97881162e-01 3.77773121e-02 1.00788939e+00 -1.54718292e+00
2.31545761e-01 -2.96729803e-01 -6.97951019e-01 -6.78247333e-01
5.22026308e-02 -8.76095414e-01 4.63082880e-01 -1.58880901e+00
-5.23599982e-01 -9.00160611e-01 1.90723106e-01 3.09213251e-01
7.21309900e-01 -1.88901439e-01 -4.57898295e-03 4.32285905e-01
-1.13620749e-03 3.06290567e-01 7.97137260e-01 2.56460011e-01
-6.22053266e-01 -1.18576668e-01 -2.53979545e-02 7.69445419e-01
7.33828187e-01 -4.43283051e-01 -3.26528251e-01 -2.57097870e-01
2.92396516e-01 1.53598383e-01 2.19555452e-01 -1.33160102e+00
5.01295209e-01 -1.80541009e-01 -1.18230015e-01 -6.70908034e-01
3.50528985e-01 -1.20805228e+00 7.72635341e-01 7.55172431e-01
-1.21875755e-01 2.40506932e-01 1.19634911e-01 7.51609981e-01
1.51663244e-01 -7.14352548e-01 5.14753580e-01 -1.36382043e-01
-8.86609674e-01 -1.29874662e-01 -3.90908003e-01 -5.41444242e-01
1.37071955e+00 -9.20367181e-01 2.71756649e-02 -3.04291770e-02
-7.63502419e-01 3.88781905e-01 1.26400697e+00 1.13619387e-01
7.82464385e-01 -1.02854013e+00 -3.14904332e-01 3.13444376e-01
2.21543163e-01 2.48803467e-01 -1.34687692e-01 8.23285460e-01
-9.73201454e-01 3.43441099e-01 -5.86900473e-01 -6.32096171e-01
-9.53577876e-01 6.57935083e-01 3.62044185e-01 7.86381438e-02
-7.64174879e-01 1.74621642e-01 -2.13351846e-01 -7.59195447e-01
-8.41648281e-02 -4.02929306e-01 -2.69901663e-01 -1.48217767e-01
1.63595304e-01 8.96403074e-01 1.80399880e-01 -9.00281727e-01
-4.20306921e-01 1.16348124e+00 4.91859019e-01 -5.26415408e-01
1.15953314e+00 -5.86683154e-01 -2.11108297e-01 4.08934772e-01
1.03849268e+00 2.45286271e-01 -1.11385918e+00 4.09419313e-02
5.16547561e-01 -4.53792334e-01 -2.95330346e-01 -3.58555853e-01
-3.71978581e-01 4.50743735e-01 7.06698358e-01 1.84691265e-01
8.65630686e-01 -6.47999644e-02 4.54326123e-01 5.97650349e-01
1.09037948e+00 -1.15409088e+00 -2.53498495e-01 6.53528452e-01
7.90335596e-01 -5.68954229e-01 -1.88397214e-01 -5.75943112e-01
-1.89981386e-02 1.65179765e+00 1.80813879e-01 -1.75889388e-01
3.16429079e-01 1.42344996e-01 -5.15976548e-02 -2.78185755e-01
-9.28041264e-02 -3.10749978e-01 -1.33510023e-01 8.40933859e-01
-5.88914573e-01 -5.13655841e-02 -3.36462349e-01 -1.84240341e-01
-6.01613283e-01 -1.26763672e-01 9.41138029e-01 1.27539098e+00
-1.15634716e+00 -9.90634978e-01 -8.63118172e-01 -6.19980954e-02
6.73161626e-01 4.90330964e-01 -5.03176153e-01 9.32854831e-01
2.13462427e-01 7.79288650e-01 3.27918559e-01 -4.97092366e-01
6.89302862e-01 5.10658547e-02 3.91408592e-01 -4.68914121e-01
8.95542949e-02 -2.99735278e-01 1.26635134e-01 -8.09375167e-01
-3.16769540e-01 -7.93419361e-01 -1.67770410e+00 2.23418362e-02
-2.08270401e-01 3.43022615e-01 1.29410434e+00 9.07042801e-01
3.59118134e-01 -8.98840129e-02 8.28578234e-01 -1.14047527e+00
-3.48504156e-01 -4.31454957e-01 -4.45413888e-01 6.41550357e-03
2.20017329e-01 -1.14571035e+00 -3.70446503e-01 -1.58272654e-01] | [7.245418548583984, -2.0703253746032715] |
0b4607ff-44c0-4c8b-9fb6-49ee2d38d035 | summarize-before-aggregate-a-global-to-local | null | null | https://aclanthology.org/2020.coling-main.367 | https://aclanthology.org/2020.coling-main.367.pdf | Summarize before Aggregate: A Global-to-local Heterogeneous Graph Inference Network for Conversational Emotion Recognition | Conversational Emotion Recognition (CER) is a crucial task in Natural Language Processing (NLP) with wide applications. Prior works in CER generally focus on modeling emotion influences solely with utterance-level features, with little attention paid on phrase-level semantic connection between utterances. Phrases carry sentiments when they are referred to emotional events under certain topics, providing a global semantic connection between utterances throughout the entire conversation. In this work, we propose a two-stage Summarization and Aggregation Graph Inference Network (SumAggGIN), which seamlessly integrates inference for topic-related emotional phrases and local dependency reasoning over neighbouring utterances in a global-to-local fashion. Topic-related emotional phrases, which constitutes the global topic-related emotional connections, are recognized by our proposed heterogeneous Summarization Graph. Local dependencies, which captures short-term emotional effects between neighbouring utterances, are further injected via an Aggregation Graph to distinguish the subtle differences between utterances containing emotional phrases. The two steps of graph inference are tightly-coupled for a comprehensively understanding of emotional fluctuation. Experimental results on three CER benchmark datasets verify the effectiveness of our proposed model, which outperforms the state-of-the-art approaches. | ['Haozhuang Liu', 'Haitao Zheng', 'Ying Shen', 'Dong Wang', 'Dongming Sheng'] | 2020-12-01 | null | null | null | coling-2020-8 | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [ 2.41895090e-03 2.38067999e-01 -1.29344672e-01 -8.22478831e-01
-6.49627030e-01 -3.44774723e-01 4.28433776e-01 5.48795938e-01
1.74354225e-01 4.97111768e-01 8.40654671e-01 3.88152599e-01
1.50924981e-01 -5.15563428e-01 -4.41389322e-01 -7.59593785e-01
-1.54582158e-01 6.68133097e-03 -2.05505162e-01 -4.92245942e-01
1.80964395e-01 -9.92538184e-02 -1.17221153e+00 5.40100813e-01
9.62278605e-01 1.10281324e+00 -2.87021786e-01 6.24118865e-01
-6.88866198e-01 1.17760789e+00 -8.18287849e-01 -5.96661210e-01
-7.82527447e-01 -6.44336939e-01 -8.04105103e-01 2.81211108e-01
-2.72177428e-01 3.25130016e-01 1.96195185e-01 1.13020658e+00
4.82454002e-01 3.91839206e-01 5.60077727e-01 -1.32202792e+00
-5.19357562e-01 9.50683057e-01 -6.35763288e-01 8.50337967e-02
6.83737993e-01 -3.83803308e-01 1.49278867e+00 -8.36183727e-01
4.88681942e-01 1.65663946e+00 3.82507741e-01 3.51401448e-01
-7.36567855e-01 -4.86643106e-01 8.60212147e-01 2.66870230e-01
-9.94205058e-01 -1.76460654e-01 1.44522333e+00 -2.18676195e-01
1.30153358e+00 3.96473229e-01 5.88289440e-01 1.09564543e+00
3.82722348e-01 1.10213435e+00 6.94886148e-01 -8.29659253e-02
2.43607342e-01 2.21811950e-01 6.67918921e-01 3.35168421e-01
-4.28993940e-01 -8.21888447e-01 -7.36372352e-01 -2.13649258e-01
-6.59558699e-02 -1.62198186e-01 -4.45928514e-01 2.23807767e-01
-8.04117620e-01 1.00957513e+00 3.90506148e-01 5.15487254e-01
-7.45172381e-01 -8.40159431e-02 1.05773771e+00 3.16801965e-01
1.20387125e+00 2.09969759e-01 -5.58147192e-01 -1.96211040e-01
-5.17848074e-01 -6.52907863e-02 1.12175155e+00 9.07867432e-01
8.68006825e-01 -1.93967387e-01 -2.81470358e-01 1.10509968e+00
3.78172994e-01 1.27664804e-01 3.83365393e-01 -5.34319222e-01
3.99809897e-01 8.88530552e-01 -3.24754238e-01 -1.71158803e+00
-4.25858051e-01 -3.49887431e-01 -9.24155474e-01 -7.61589944e-01
-6.03461325e-01 -5.78537881e-01 -9.23631936e-02 1.84884095e+00
5.93194723e-01 2.05547974e-01 6.46932304e-01 8.04674983e-01
1.39674997e+00 1.11724603e+00 3.49936962e-01 -6.38934135e-01
1.74623656e+00 -1.24449015e+00 -1.29005122e+00 -2.20194548e-01
6.34208739e-01 -6.28453255e-01 7.77875841e-01 1.33961037e-01
-8.96952569e-01 -2.12195128e-01 -7.79938400e-01 -1.23802334e-01
-4.84632015e-01 -2.51036406e-01 6.46833539e-01 1.42036393e-01
-8.93379867e-01 6.42635971e-02 -3.61804187e-01 -4.44129169e-01
1.18615001e-01 1.00352727e-01 -9.82331559e-02 2.82680899e-01
-1.78204477e+00 6.98896766e-01 4.13406700e-01 4.87938851e-01
-2.79085636e-01 -4.51238871e-01 -1.21147811e+00 2.93667734e-01
5.76016843e-01 -5.57413697e-01 1.17750645e+00 -1.13220942e+00
-1.70104361e+00 5.86526930e-01 -7.69075334e-01 -5.94593696e-02
-1.71218485e-01 -2.32216403e-01 -7.52266824e-01 3.39679629e-01
-4.23139185e-02 5.04430652e-01 7.33467162e-01 -1.33993983e+00
-4.57925737e-01 -3.83419186e-01 -2.14323658e-03 8.76574278e-01
-4.37378854e-01 5.17362893e-01 -6.11347795e-01 -5.34704804e-01
-1.35722801e-01 -4.58790094e-01 -8.58055353e-02 -9.67335463e-01
-6.08331203e-01 -9.27833438e-01 1.08936191e+00 -5.96599102e-01
1.47437239e+00 -2.02000499e+00 4.55860257e-01 1.12934165e-01
7.44819120e-02 -2.69417882e-01 -7.22764134e-02 6.30074799e-01
-1.10056579e-01 1.03222735e-01 -9.64231566e-02 -5.01610696e-01
3.27025086e-01 1.85347557e-01 -5.46522737e-01 -4.33272272e-02
5.99020243e-01 9.77959216e-01 -1.09843612e+00 -7.54854620e-01
7.45042637e-02 6.14267707e-01 -3.42841417e-01 3.87644112e-01
-2.92196125e-01 4.84445989e-01 -9.10266578e-01 6.63603842e-01
4.52266037e-01 -1.27826165e-02 1.46278992e-01 -2.87508637e-01
1.69575140e-01 3.38649958e-01 -5.29323757e-01 1.61860943e+00
-6.10373676e-01 4.17869568e-01 2.75924206e-01 -1.16319764e+00
1.28269112e+00 5.96148849e-01 2.83393025e-01 -4.17117983e-01
5.08152843e-01 -2.80144185e-01 -5.11247993e-01 -7.55483091e-01
7.10589886e-01 -3.67168546e-01 -8.33226323e-01 4.18844134e-01
1.67037889e-01 -3.69202554e-01 7.31550157e-02 5.72921515e-01
7.04273164e-01 -2.75116384e-01 5.46444952e-01 -2.14227870e-01
8.79680812e-01 -3.11296999e-01 6.82210684e-01 2.08788142e-01
-3.65189373e-01 1.11744598e-01 1.08211648e+00 1.29887208e-01
-1.50432706e-01 -5.03963530e-01 1.74215421e-01 1.46058083e+00
4.51058656e-01 -6.16858125e-01 -8.51956844e-01 -8.27280343e-01
-4.79114980e-01 8.70400667e-01 -8.26321721e-01 -2.34583229e-01
-2.38998860e-01 -7.93366671e-01 3.44405353e-01 3.45075339e-01
4.92340893e-01 -1.51055157e+00 -8.64953026e-02 2.81532258e-01
-7.81564832e-01 -1.21920037e+00 -4.97324437e-01 1.30804598e-01
-4.24015462e-01 -6.34968281e-01 -3.99170130e-01 -7.54685640e-01
6.04629397e-01 3.89359109e-02 1.16091430e+00 -8.27804282e-02
1.87561333e-01 5.98279774e-01 -7.16573238e-01 -4.64026541e-01
-3.09666991e-01 -6.62163123e-02 -1.26528740e-01 3.52741152e-01
7.24986732e-01 -5.50391257e-01 -4.19005483e-01 1.63123757e-01
-7.58200049e-01 -8.55309740e-02 4.02091533e-01 7.54723251e-01
3.47751886e-01 3.14908147e-01 1.04799235e+00 -1.02401280e+00
1.24739587e+00 -9.48653460e-01 3.37762535e-01 4.39728618e-01
5.66940233e-02 -5.89882955e-02 6.11586332e-01 -2.44190052e-01
-1.75932634e+00 -6.01468921e-01 -1.55610323e-01 -1.13013357e-01
-2.86708206e-01 1.01766658e+00 -5.16471922e-01 5.16125262e-01
-8.94870535e-02 1.36436403e-01 -4.58816916e-01 9.10387412e-02
6.48118854e-01 8.87326121e-01 4.56673950e-01 -7.82537401e-01
-7.93453231e-02 2.61554807e-01 -5.52347064e-01 -1.01980352e+00
-1.29499400e+00 -8.19487512e-01 -2.55253971e-01 -6.97024047e-01
1.13300574e+00 -9.95186388e-01 -7.29404509e-01 3.63486886e-01
-1.45557034e+00 2.14676395e-01 -3.19104679e-02 2.95128763e-01
-3.06420684e-01 3.96729231e-01 -7.76690722e-01 -9.75275695e-01
-6.88958168e-01 -8.43742132e-01 1.34054565e+00 5.79533815e-01
-6.60370290e-01 -1.45317495e+00 1.11250252e-01 4.12387937e-01
9.19520780e-02 3.92943799e-01 8.13233316e-01 -7.94042706e-01
2.65282899e-01 -9.81734842e-02 -2.19230726e-01 3.51794988e-01
1.60985917e-01 1.86440036e-01 -1.02888441e+00 1.24508515e-01
3.66218328e-01 -6.10258698e-01 8.21793258e-01 3.05351108e-01
9.84299123e-01 -4.04855222e-01 -2.24831894e-01 -1.06512681e-01
9.41132367e-01 1.33445188e-01 3.14170390e-01 -2.70710528e-01
6.60251737e-01 1.12566662e+00 8.34402025e-01 5.97943604e-01
7.97472179e-01 3.12187314e-01 2.32737377e-01 -1.80275872e-01
4.77423549e-01 -1.22947112e-01 7.38738596e-01 1.74898553e+00
1.34861648e-01 -4.42752510e-01 -3.43539089e-01 4.77240205e-01
-2.25479031e+00 -9.41917479e-01 -2.47745827e-01 1.19240153e+00
9.63231385e-01 -1.14836469e-01 -4.23655778e-01 -2.46835262e-01
8.90631855e-01 7.73601592e-01 -6.11295879e-01 -1.11468112e+00
-2.33870387e-01 -1.40479624e-01 -5.04895747e-01 5.59105337e-01
-1.08002126e+00 1.09182739e+00 5.00793648e+00 8.73913884e-01
-9.07163739e-01 3.83168384e-02 8.07209849e-01 1.71962187e-01
-5.00093579e-01 -1.84520900e-01 -5.56572378e-01 2.58402914e-01
7.67540872e-01 -4.03837800e-01 -1.81213692e-02 9.06848848e-01
4.21418279e-01 -1.15791664e-01 -8.16007555e-01 7.21253097e-01
6.02635145e-01 -8.36959422e-01 -4.58914638e-02 -6.04309499e-01
8.26625764e-01 -2.44126782e-01 -2.48947605e-01 6.89901471e-01
2.08523571e-01 -5.05935490e-01 4.66526359e-01 5.35937607e-01
1.32267997e-01 -1.09102333e+00 1.11513710e+00 2.66856939e-01
-1.52424777e+00 3.49729300e-01 -3.04794401e-01 -1.59689963e-01
4.10371363e-01 8.97171259e-01 -6.10193908e-01 9.90250111e-01
8.02898824e-01 1.25154996e+00 -6.26296103e-02 -7.58423954e-02
-6.92524016e-01 7.37234056e-01 1.23420283e-02 -5.43437779e-01
3.88474017e-01 -3.82212669e-01 7.12116420e-01 1.85224891e+00
-6.95366710e-02 5.92484415e-01 6.14379272e-02 8.03255379e-01
-3.86460900e-01 5.53890765e-01 -3.99194241e-01 -2.62951910e-01
2.77440220e-01 1.73744094e+00 -6.82238758e-01 -5.39043009e-01
-3.80784601e-01 1.14571702e+00 4.78088170e-01 4.10020947e-01
-9.10539269e-01 -6.05326772e-01 6.88361466e-01 -1.04083312e+00
1.91883400e-01 3.19402009e-01 4.04301248e-02 -1.25748551e+00
1.54431850e-01 -6.08066380e-01 5.01534879e-01 -8.88619483e-01
-1.72868335e+00 6.94356382e-01 -2.35867187e-01 -7.79129446e-01
-1.56755760e-01 1.87767837e-02 -1.32101715e+00 6.37522817e-01
-1.54212177e+00 -9.20792341e-01 -2.17461050e-01 5.50351024e-01
8.21122766e-01 3.70767325e-01 9.92384315e-01 -2.28673220e-01
-9.46282744e-01 3.31359744e-01 -5.41399539e-01 -5.17849587e-02
6.35386050e-01 -1.21605396e+00 -1.75066054e-01 7.20028102e-01
-1.93873256e-01 6.30591691e-01 8.65797758e-01 -5.75432122e-01
-1.11766958e+00 -1.15634048e+00 1.26694286e+00 -7.27582350e-02
8.49260628e-01 -4.31301534e-01 -1.18278635e+00 5.15267015e-01
8.84068251e-01 -5.87319911e-01 1.04933202e+00 6.04459405e-01
-2.36712635e-01 1.34094641e-01 -8.53948176e-01 4.95835513e-01
6.30471826e-01 -6.36326909e-01 -1.04734707e+00 3.36358517e-01
1.28097200e+00 -1.23753630e-01 -9.80403841e-01 3.57483178e-01
1.95659027e-01 -9.01584327e-01 4.83039200e-01 -5.65906227e-01
7.71073282e-01 -5.08417450e-02 -3.97462584e-02 -1.73652637e+00
6.88320547e-02 -7.29164243e-01 3.39876972e-02 1.87613332e+00
3.86934668e-01 -5.97738385e-01 1.86680451e-01 5.35992265e-01
-2.36278415e-01 -9.10155416e-01 -3.86955053e-01 -3.28825898e-02
-3.58274817e-01 -7.55353868e-01 5.30725598e-01 1.37177336e+00
9.92932975e-01 1.00796986e+00 -3.14572752e-01 3.04394066e-01
1.00876503e-01 3.60701621e-01 5.23509324e-01 -9.19475257e-01
-2.39540767e-02 -7.24703312e-01 -2.43524849e-01 -9.01668787e-01
8.68407190e-01 -7.77221143e-01 3.84528607e-01 -1.65015113e+00
4.78314281e-01 6.57585934e-02 -3.61758083e-01 3.41613531e-01
-6.47956848e-01 -2.10511625e-01 -1.64731026e-01 -1.82520747e-01
-1.18166196e+00 1.11576164e+00 1.07413077e+00 -2.01892614e-01
-2.65953958e-01 -3.48180801e-01 -1.03362846e+00 9.38485861e-01
7.78338552e-01 -3.30903322e-01 -5.05217969e-01 -3.72739881e-02
4.88650799e-01 2.11412534e-01 -1.22226432e-01 -3.53362709e-01
2.81576931e-01 -8.91834348e-02 -1.80557787e-01 -7.96355546e-01
2.86359936e-01 -6.25789106e-01 -2.83142954e-01 -1.51615128e-01
-6.51300609e-01 -2.61051029e-01 1.26946181e-01 7.75333345e-01
-9.60883319e-01 5.75906411e-02 2.66838610e-01 2.06588700e-01
-7.05497086e-01 1.83847278e-01 -4.17249918e-01 2.33380944e-01
1.07113647e+00 2.31431738e-01 -2.87967622e-01 -9.05746222e-01
-8.34130943e-01 8.48235428e-01 -1.73934609e-01 7.86996424e-01
7.15132236e-01 -1.20734084e+00 -8.46442938e-01 -1.12881161e-01
3.10176313e-01 2.25159913e-01 8.71579647e-01 9.81447399e-01
2.33455852e-01 3.87745231e-01 3.73226911e-01 -5.37226617e-01
-1.51763952e+00 3.14281851e-01 1.00855768e-01 -5.42809129e-01
-3.37157965e-01 1.10528922e+00 5.06995440e-01 -5.06111026e-01
1.97365358e-01 -4.05140966e-01 -7.25162327e-01 5.52647173e-01
3.98742348e-01 -3.60904001e-02 -2.10482582e-01 -1.02477813e+00
-4.39981967e-01 6.45237565e-01 -1.35538131e-01 7.08988383e-02
1.22098196e+00 -5.65116227e-01 -7.99332261e-01 8.79240513e-01
1.45628119e+00 -8.59261677e-02 -7.95154393e-01 -3.26845795e-01
-5.23580983e-02 2.70886749e-01 1.20072998e-01 -6.51553035e-01
-1.03834879e+00 9.74679410e-01 -4.39447939e-01 5.31693280e-01
1.31668448e+00 4.24443364e-01 8.99082661e-01 3.28534752e-01
-6.57748878e-02 -1.34444022e+00 1.38925970e-01 6.28846645e-01
1.12664771e+00 -1.12926590e+00 -1.58445105e-01 -6.99175954e-01
-1.33741474e+00 1.02402294e+00 5.98796546e-01 1.21774286e-01
5.41719139e-01 1.06713004e-01 1.80257797e-01 -5.62636375e-01
-1.24078131e+00 2.42934860e-02 3.22214931e-01 -6.27806736e-03
5.67998469e-01 3.49157184e-01 -4.70916152e-01 1.19766331e+00
-1.12364523e-01 -4.94894058e-01 2.00660780e-01 7.30042815e-01
-4.26718593e-01 -6.48328304e-01 -7.54185766e-02 8.20822120e-02
-3.57611775e-01 -1.43425629e-01 -9.55524683e-01 3.64146709e-01
-1.52225822e-01 1.45019722e+00 2.48662308e-01 -3.49507987e-01
2.53120124e-01 3.80333841e-01 -2.06885263e-01 -4.46715593e-01
-8.80191386e-01 3.46206218e-01 4.78918314e-01 -6.15989208e-01
-9.17714834e-01 -6.81316912e-01 -1.70800972e+00 -2.84681618e-02
-4.65184301e-01 6.12928510e-01 6.02519095e-01 1.11335433e+00
5.59918761e-01 1.03366065e+00 1.01934612e+00 -6.80525839e-01
5.46037294e-02 -1.17105186e+00 -6.15905881e-01 5.68924129e-01
1.09843142e-01 -2.86162436e-01 -6.20537400e-01 -4.90807109e-02] | [12.925463676452637, 6.22299861907959] |
8cdf42d4-11db-44cb-bd09-805024fd8ac3 | variational-bayesian-sequence-to-sequence | 2102.06143 | null | https://arxiv.org/abs/2102.06143v1 | https://arxiv.org/pdf/2102.06143v1.pdf | Variational Bayesian Sequence-to-Sequence Networks for Memory-Efficient Sign Language Translation | Memory-efficient continuous Sign Language Translation is a significant challenge for the development of assisted technologies with real-time applicability for the deaf. In this work, we introduce a paradigm of designing recurrent deep networks whereby the output of the recurrent layer is derived from appropriate arguments from nonparametric statistics. A novel variational Bayesian sequence-to-sequence network architecture is proposed that consists of a) a full Gaussian posterior distribution for data-driven memory compression and b) a nonparametric Indian Buffet Process prior for regularization applied on the Gated Recurrent Unit non-gate weights. We dub our approach Stick-Breaking Recurrent network and show that it can achieve a substantial weight compression without diminishing modeling performance. | ['Dimitris N. Metaxas', 'Sotirios Chatzis', 'Dimitrios Kosmopoulos', 'Andreas Voskou', 'Harris Partaourides'] | 2021-02-11 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 7.17807293e-01 4.08800066e-01 -9.90730599e-02 -3.43845874e-01
-9.04718041e-01 1.50279507e-01 4.51705426e-01 -5.75000107e-01
-6.64671481e-01 7.27121115e-01 6.89611554e-01 -4.33134943e-01
-7.70699531e-02 -4.16149229e-01 -7.55048394e-01 -1.01858950e+00
2.56839246e-01 6.63845420e-01 1.97795406e-01 6.70405179e-02
8.26089606e-02 3.06706935e-01 -1.55458534e+00 3.18784744e-01
7.26626754e-01 6.94881439e-01 5.63165426e-01 1.02071297e+00
1.04951248e-01 8.61104846e-01 -1.68258309e-01 -2.77045459e-01
5.96073642e-02 -6.09749079e-01 -2.26557255e-01 -4.84350801e-01
-1.34021034e-02 -6.71036780e-01 -7.09090948e-01 8.74865472e-01
1.00994742e+00 2.80519873e-01 1.10855985e+00 -4.98200387e-01
-4.20729846e-01 7.63772964e-01 -2.38599271e-01 -1.67588905e-01
-3.58255148e-01 1.46170244e-01 8.27955663e-01 -6.96323454e-01
6.58761084e-01 1.25218904e+00 6.64191484e-01 9.74802792e-01
-1.31912434e+00 -5.00336170e-01 -5.15078641e-02 1.38985619e-01
-1.21322083e+00 -9.12674546e-01 8.70275140e-01 -2.72017270e-01
1.23920989e+00 -4.22055945e-02 7.98120379e-01 1.50215757e+00
1.92274541e-01 1.00591266e+00 7.03784108e-01 -6.28656864e-01
5.72474897e-01 -3.95746738e-01 5.90561815e-02 5.73751271e-01
2.76435111e-02 4.70452368e-01 -1.06306422e+00 -3.05212587e-01
7.68846035e-01 -2.77913839e-01 1.22098707e-01 -1.67442501e-01
-2.92860627e-01 9.09409046e-01 -1.75375804e-01 -6.64275736e-02
-5.23277760e-01 8.14534843e-01 3.67450535e-01 -4.92543690e-02
4.64489698e-01 -6.00516260e-01 -1.85664326e-01 -5.77347636e-01
-1.38321257e+00 2.62304544e-01 7.97164142e-01 9.71096516e-01
3.85652333e-02 6.38029099e-01 -4.18459713e-01 1.01257062e+00
9.33568180e-01 7.53520787e-01 4.59953696e-01 -1.07974517e+00
1.64869413e-01 -2.34896123e-01 -1.78770632e-01 -1.98745787e-01
-2.68009365e-01 -4.11202967e-01 -8.86189640e-01 2.83544391e-01
2.28925481e-01 -3.19714732e-02 -1.54842198e+00 1.97564816e+00
-2.72570670e-01 2.87939757e-01 -1.20647661e-01 6.07662201e-01
4.61129725e-01 5.14241099e-01 2.06512406e-01 -3.01877201e-01
8.58644068e-01 -4.85689431e-01 -8.88778269e-01 -1.11852728e-01
2.70624936e-01 -5.21573305e-01 7.67613590e-01 4.40673888e-01
-1.66462076e+00 -4.79775481e-02 -9.13272381e-01 -3.46738935e-01
2.94443637e-01 1.82048485e-01 3.22407246e-01 6.46913409e-01
-1.24737918e+00 7.40572691e-01 -1.29794157e+00 -2.03702122e-01
4.63182449e-01 4.56369311e-01 1.69736072e-01 1.72343746e-01
-7.49345660e-01 9.48799551e-01 3.13827455e-01 5.42349458e-01
-8.04899395e-01 -3.92655283e-01 -6.71958268e-01 -7.19572827e-02
-3.81818831e-01 -8.70073915e-01 1.33767951e+00 -4.20006424e-01
-2.18130422e+00 9.36736822e-01 -4.77628112e-01 -9.38925147e-01
7.61734128e-01 -3.99151266e-01 1.50909200e-01 2.16182340e-02
-5.56209564e-01 5.69834113e-01 1.33555496e+00 -9.99971747e-01
-3.71344507e-01 -1.56702250e-01 -1.07193696e+00 -8.25205073e-02
-1.18073829e-01 -2.23512538e-02 -2.99576163e-01 -8.69511425e-01
4.36596930e-01 -1.04395545e+00 -1.84919253e-01 1.41925082e-01
-1.71596497e-01 -7.98445717e-02 3.27660412e-01 -1.38384032e+00
1.27995622e+00 -1.98898661e+00 4.24880505e-01 4.40901488e-01
2.42828205e-02 3.08123708e-01 -1.19247653e-01 2.34015658e-01
4.15113181e-01 -3.15907180e-01 -5.10019362e-01 -7.30525017e-01
1.87231347e-01 5.46162367e-01 -6.50808394e-01 5.34968495e-01
-5.18878065e-02 9.78273749e-01 -5.18358767e-01 -3.40086281e-01
2.76558064e-02 1.02103055e+00 -8.37027729e-01 2.68905759e-01
-4.04132932e-01 1.51543558e-01 -7.06227943e-02 4.71659154e-01
6.21485174e-01 2.30591223e-01 5.09506285e-01 1.37338946e-02
-4.76083681e-02 3.59939665e-01 -7.11638033e-01 1.87120521e+00
-2.39934489e-01 7.31135547e-01 1.72491461e-01 -8.58135462e-01
1.01342773e+00 1.78298309e-01 1.46870121e-01 -6.90769494e-01
4.10850763e-01 4.90952224e-01 -2.47763712e-02 -3.94461781e-01
6.61549449e-01 -5.46815753e-01 2.44852051e-01 5.40459454e-01
2.99189717e-01 -9.96106565e-02 -6.51594996e-02 -5.24666756e-02
1.10069966e+00 5.41652799e-01 -3.98207575e-01 -1.46299899e-01
-8.90122056e-02 -7.21467435e-01 7.08804846e-01 9.57867801e-01
-1.35639220e-01 8.55390251e-01 1.58485666e-01 -6.00895844e-02
-1.56235385e+00 -1.39624906e+00 8.78801420e-02 1.09092867e+00
-5.22108614e-01 6.82305405e-03 -7.53608823e-01 2.48541862e-01
-1.63509533e-01 9.97914612e-01 -3.81841570e-01 -3.55579436e-01
-8.16220403e-01 -6.37118101e-01 8.96942139e-01 7.19595253e-01
2.72824317e-01 -1.26199067e+00 -6.70218468e-01 4.63651657e-01
-3.55678946e-01 -7.16432512e-01 -4.22187448e-01 5.23993611e-01
-1.17346704e+00 -1.46617278e-01 -9.41950381e-01 -9.33199167e-01
3.00452411e-01 -4.68366742e-01 8.09935927e-01 -2.07130805e-01
-1.11710533e-01 1.78133547e-01 -3.97719555e-02 -5.42307377e-01
-6.64082348e-01 9.68020782e-02 1.74652100e-01 -3.19461167e-01
4.49064821e-01 -1.10959351e+00 -5.28680861e-01 -4.86639142e-01
-8.73167157e-01 1.99425757e-01 5.85513711e-01 1.10362983e+00
5.63193500e-01 -6.72523022e-01 3.24120998e-01 -1.67698234e-01
6.97522640e-01 -1.47764562e-02 -4.76968408e-01 1.46784768e-01
-3.93176883e-01 7.12216675e-01 -1.20402485e-01 -6.93037570e-01
-1.08419776e+00 8.78927037e-02 -4.53777790e-01 -3.68059993e-01
3.99284661e-01 3.12458932e-01 9.68353599e-02 2.14331746e-01
4.78044838e-01 5.41919708e-01 4.07955676e-01 -5.85850656e-01
5.92208803e-01 7.64784992e-01 7.62172759e-01 -7.59693682e-01
4.29816931e-01 4.72609937e-01 -5.96740693e-02 -1.00147283e+00
-3.21410060e-01 -5.02494425e-02 -4.00525987e-01 -3.33293527e-01
8.37693274e-01 -7.28540778e-01 -8.77694011e-01 7.10264325e-01
-1.31783855e+00 -7.47032285e-01 -6.05481625e-01 6.28337562e-01
-1.32785690e+00 1.21833660e-01 -1.07956755e+00 -1.53104198e+00
-6.31028056e-01 -6.50961876e-01 7.92449296e-01 4.18863446e-02
-3.32423449e-01 -6.89661026e-01 3.47409338e-01 1.78435847e-01
6.12925708e-01 -3.13160092e-01 1.10084593e+00 -2.31959999e-01
-6.57260180e-01 -1.30306685e-03 -9.62002203e-02 6.55853629e-01
-4.91054446e-01 -2.08842345e-02 -1.18912923e+00 -2.85026670e-01
-2.74065826e-02 -2.56708920e-01 1.37350237e+00 1.08568215e+00
5.57888806e-01 -1.35991484e-01 -8.75868872e-02 7.46106207e-01
1.06857455e+00 8.43261629e-02 9.73210156e-01 -8.17196667e-02
3.45191032e-01 4.34817135e-01 -1.28827035e-01 6.53953850e-01
4.45903689e-02 3.45145255e-01 -3.00403824e-03 4.69172537e-01
-4.46225911e-01 -7.57619739e-01 6.12314105e-01 1.32242811e+00
-3.83325577e-01 -1.18903793e-01 -6.92571163e-01 7.38697469e-01
-2.11628890e+00 -1.12229371e+00 3.11093628e-01 2.13951135e+00
9.83484924e-01 2.28839040e-01 -2.25606300e-02 1.57101601e-02
5.75743854e-01 -2.27740519e-02 -7.41319954e-01 -5.71084261e-01
-3.89718205e-01 6.42650247e-01 5.31409562e-01 6.70348465e-01
-6.35808408e-01 9.12244320e-01 7.27741098e+00 1.13050508e+00
-7.66044855e-01 2.77422249e-01 2.16015577e-01 -4.38764185e-01
-4.52773511e-01 -1.66316643e-01 -9.73037362e-01 4.45171326e-01
1.35684311e+00 3.54977906e-01 4.40600097e-01 3.94625783e-01
5.13038099e-01 2.77065448e-02 -6.61457479e-01 9.40745950e-01
3.69988345e-02 -1.31276345e+00 1.88841283e-01 3.28504920e-01
4.75393891e-01 3.11824054e-01 2.96556473e-01 1.02066278e-01
4.88325536e-01 -1.04150224e+00 1.09538186e+00 1.09113646e+00
9.21922684e-01 -7.18967497e-01 2.35115200e-01 3.36459279e-01
-8.00621927e-01 -1.33545339e-01 -3.25830847e-01 3.52575853e-02
6.25275254e-01 4.49862093e-01 -4.80630547e-01 -2.42365211e-01
3.54592383e-01 3.02742183e-01 4.05623555e-01 1.02729023e+00
-2.49125659e-01 1.12950480e+00 -6.96393490e-01 -1.51678622e-01
-4.55973558e-02 -2.77777642e-01 7.64011621e-01 1.35517967e+00
6.32362664e-01 5.39803281e-02 -4.81469184e-01 1.01213825e+00
3.58742923e-02 -3.01133364e-01 -4.63355452e-01 -1.10975862e-01
1.76363006e-01 2.84417897e-01 -5.55999756e-01 -1.15911514e-01
9.54434499e-02 1.01160598e+00 3.90083998e-01 6.64261639e-01
-3.69894654e-01 -1.45089746e-01 5.96122265e-01 1.30260497e-01
6.97928429e-01 -6.05479896e-01 -5.93200803e-01 -9.61533129e-01
1.52600119e-02 -2.84948200e-01 6.11439906e-02 -7.13014007e-01
-1.06814682e+00 2.53003716e-01 -1.56981662e-01 -9.73795295e-01
-6.64309680e-01 -6.36553943e-01 -3.34091932e-01 1.22377217e+00
-1.35818076e+00 -1.37000418e+00 4.28780407e-01 3.18156004e-01
2.25541130e-01 -3.27278823e-01 9.46471393e-01 2.84942359e-01
-1.55711815e-01 8.92333329e-01 4.16155487e-01 -7.77082518e-02
2.64076710e-01 -6.22369289e-01 4.77376193e-01 7.51234353e-01
4.14841175e-02 7.24948525e-01 1.11285102e+00 -9.32110250e-01
-1.05164957e+00 -6.38493657e-01 1.06475246e+00 -9.71584618e-02
5.48075736e-01 -4.42824304e-01 -9.55221534e-01 5.14769733e-01
4.11084108e-02 -4.75544125e-01 5.14948726e-01 -3.68405543e-02
-2.91653395e-01 2.05100358e-01 -1.11506009e+00 6.67677104e-01
1.18005228e+00 -9.74474013e-01 -6.28763258e-01 -9.89623889e-02
5.98861456e-01 -3.24888974e-01 -2.44009122e-01 3.87513429e-01
1.24939287e+00 -6.17221653e-01 6.74301982e-01 -6.14763081e-01
2.21878007e-01 4.93592843e-02 -4.45704281e-01 -8.12928855e-01
-6.79243430e-02 -1.18739974e+00 -9.01614189e-01 8.64096105e-01
3.33150715e-01 -3.18491071e-01 1.20131481e+00 8.54744256e-01
-2.32533678e-01 -6.30213022e-01 -1.42251706e+00 -7.60329187e-01
3.75757098e-01 -8.88299584e-01 -2.26757571e-01 -2.08083197e-01
-1.09064773e-01 3.76537442e-02 -8.41629446e-01 -1.97470650e-01
1.00784934e+00 -4.75038975e-01 1.80412605e-01 -1.06197894e+00
-4.62535143e-01 -5.56697786e-01 -3.48870337e-01 -1.49068391e+00
2.62033522e-01 -8.49691153e-01 6.83546603e-01 -1.48453879e+00
3.86781514e-01 -4.95968238e-02 -2.17824146e-01 3.15605327e-02
3.36830407e-01 1.16742842e-01 6.72042966e-02 9.91648622e-03
4.41255383e-02 9.73462582e-01 6.25872135e-01 -1.60043929e-02
-4.25233036e-01 2.93850422e-01 -1.01319298e-01 7.41461039e-01
5.84122300e-01 -6.18870735e-01 -3.56284797e-01 -4.38240349e-01
3.56880873e-01 8.67994353e-02 3.91386896e-01 -8.10931206e-01
4.65422481e-01 2.34086692e-01 -1.49826571e-01 -9.39894080e-01
7.28296518e-01 -4.44990695e-01 1.18674055e-01 6.29875183e-01
-5.00230670e-01 -5.25458634e-01 -1.96689889e-02 6.89587235e-01
1.10194586e-01 -3.10886353e-01 8.54433358e-01 2.31119707e-01
-1.51321083e-01 7.49064535e-02 -8.12837064e-01 -5.36052696e-02
9.72517282e-02 -3.50912750e-01 1.11298956e-01 -5.82706332e-01
-1.06836271e+00 -2.88455367e-01 -1.86683778e-02 1.11219920e-01
9.85798419e-01 -1.32722580e+00 -8.97857904e-01 4.58452553e-01
-3.46785069e-01 -2.67925441e-01 2.69204170e-01 6.36321127e-01
-4.40732807e-01 2.05702737e-01 -6.84236661e-02 -5.52916050e-01
-1.16757798e+00 -3.47100616e-01 3.40858012e-01 -2.18902305e-02
-9.96229827e-01 1.12712562e+00 -3.13491791e-01 -2.74663776e-01
6.09165549e-01 -3.89135599e-01 2.93272167e-01 8.79352987e-02
4.71825063e-01 4.67478573e-01 -2.57997066e-01 -5.13244867e-01
5.51394559e-02 5.10631502e-01 9.53465030e-02 -9.74922478e-01
1.60796952e+00 -1.00258999e-01 -4.13950086e-02 7.27211237e-01
8.43106687e-01 -5.15089929e-01 -1.68252146e+00 -4.05124187e-01
2.90635973e-02 8.40234682e-02 1.89259231e-01 -6.61136866e-01
-7.28815198e-01 1.04495573e+00 9.03178394e-01 -5.20669222e-01
9.30580616e-01 -2.66274303e-01 1.07239556e+00 3.42804819e-01
3.59215915e-01 -1.59784949e+00 -2.22915277e-01 8.89995277e-01
9.15144742e-01 -7.05892026e-01 -1.86423630e-01 2.11747885e-01
-2.77734935e-01 1.02933490e+00 -2.34447733e-01 -2.95102686e-01
1.07666612e+00 6.09133065e-01 -4.81640473e-02 1.58707380e-01
-7.08709002e-01 -2.12788373e-01 2.51279563e-01 7.78994560e-01
3.64265621e-01 1.19863294e-01 -5.12308061e-01 6.43697202e-01
-2.02759445e-01 3.56608748e-01 1.23265028e-01 1.02032316e+00
-5.85762203e-01 -9.19553399e-01 1.48818225e-01 5.11167765e-01
-2.13151366e-01 -5.05894244e-01 1.91775665e-01 8.95303488e-02
-2.76815265e-01 6.48594201e-01 2.23751478e-02 -3.43677163e-01
1.74854711e-01 5.53406537e-01 8.32550347e-01 -1.14278384e-01
-1.55645847e-01 4.53568131e-01 9.87398252e-02 -2.09865689e-01
-2.87476540e-01 -9.11292970e-01 -1.35582709e+00 -3.64187151e-01
-1.87391013e-01 -2.88109243e-01 9.78266656e-01 1.05571890e+00
2.35254765e-01 4.19333726e-01 -1.57315612e-01 -1.13199186e+00
-7.99948454e-01 -1.27101755e+00 -8.06514740e-01 -1.70800552e-01
3.79283428e-01 -4.28797603e-01 -3.58410448e-01 1.87473208e-01] | [7.072738170623779, 3.8194572925567627] |
44e3abbc-deed-4db7-9c92-12417e4d0241 | backdoor-attacks-against-transfer-learning | 2001.03274 | null | https://arxiv.org/abs/2001.03274v2 | https://arxiv.org/pdf/2001.03274v2.pdf | Backdoor Attacks against Transfer Learning with Pre-trained Deep Learning Models | Transfer learning provides an effective solution for feasibly and fast customize accurate \textit{Student} models, by transferring the learned knowledge of pre-trained \textit{Teacher} models over large datasets via fine-tuning. Many pre-trained Teacher models used in transfer learning are publicly available and maintained by public platforms, increasing their vulnerability to backdoor attacks. In this paper, we demonstrate a backdoor threat to transfer learning tasks on both image and time-series data leveraging the knowledge of publicly accessible Teacher models, aimed at defeating three commonly-adopted defenses: \textit{pruning-based}, \textit{retraining-based} and \textit{input pre-processing-based defenses}. Specifically, (A) ranking-based selection mechanism to speed up the backdoor trigger generation and perturbation process while defeating \textit{pruning-based} and/or \textit{retraining-based defenses}. (B) autoencoder-powered trigger generation is proposed to produce a robust trigger that can defeat the \textit{input pre-processing-based defense}, while guaranteeing that selected neuron(s) can be significantly activated. (C) defense-aware retraining to generate the manipulated model using reverse-engineered model inputs. We launch effective misclassification attacks on Student models over real-world images, brain Magnetic Resonance Imaging (MRI) data and Electrocardiography (ECG) learning systems. The experiments reveal that our enhanced attack can maintain the $98.4\%$ and $97.2\%$ classification accuracy as the genuine model on clean image and time series inputs respectively while improving $27.9\%-100\%$ and $27.1\%-56.1\%$ attack success rate on trojaned image and time series inputs respectively in the presence of pruning-based and/or retraining-based defenses. | ['Tianle Chen', 'Shangyu Chen', 'Carsten Rudolph', 'Surya Nepal', 'Marthie Grobler', 'Shuo Wang'] | 2020-01-10 | null | null | null | null | ['electrocardiography-ecg'] | ['methodology'] | [ 5.13177395e-01 -1.40658572e-01 -4.81793657e-02 -2.53313273e-01
-8.61687601e-01 -1.06742072e+00 1.94747746e-01 -5.26414812e-02
-6.27668619e-01 8.73761594e-01 -7.73675382e-01 -7.41218030e-01
-4.26159114e-01 -8.50945950e-01 -1.14626682e+00 -9.29745734e-01
-3.93476158e-01 3.80813587e-03 1.99804649e-01 -2.51702994e-01
3.64316463e-01 6.07645750e-01 -1.28798747e+00 4.63964462e-01
9.04670596e-01 1.15821397e+00 -4.22463715e-01 7.86873758e-01
5.59308648e-01 4.53210682e-01 -1.19752872e+00 -2.74496347e-01
5.46223640e-01 -8.92840996e-02 -4.65685070e-01 -7.84080803e-01
6.92683995e-01 -4.61663187e-01 -3.62350285e-01 1.29845059e+00
5.92161238e-01 -2.87561994e-02 5.86276889e-01 -1.51562405e+00
-6.82423711e-01 6.04596078e-01 -4.36269730e-01 7.76636541e-01
-1.91610157e-01 5.47816157e-01 2.18715191e-01 -1.67636499e-01
1.31152108e-01 7.89000213e-01 5.30818939e-01 9.80917335e-01
-1.34986389e+00 -1.62610006e+00 -1.97279394e-01 9.99627821e-03
-1.23706484e+00 9.07600392e-03 7.30745971e-01 -3.32167387e-01
9.22977328e-01 5.63388467e-01 1.42445952e-01 1.88374746e+00
5.92349350e-01 2.20869467e-01 1.51048458e+00 -7.33691528e-02
1.81068048e-01 2.34933585e-01 3.40103030e-01 5.82781136e-01
3.37844044e-01 7.76475012e-01 -3.40466768e-01 -5.54815352e-01
8.51666033e-01 -9.49532017e-02 -1.98507220e-01 4.37254697e-01
-5.93499124e-01 8.32285762e-01 2.95027554e-01 7.53870979e-02
-1.33836284e-01 3.13545704e-01 5.64572155e-01 6.91831112e-01
4.68965061e-02 7.00357735e-01 -8.83962274e-01 6.66328818e-02
-6.95371687e-01 3.05509008e-02 3.31655294e-01 6.80645347e-01
5.50856888e-01 9.15564835e-01 -1.20263264e-01 3.66098225e-01
-1.50338024e-01 8.22869539e-01 7.48665094e-01 -4.67936367e-01
4.13179189e-01 3.52447420e-01 -4.14248228e-01 -8.63909245e-01
-2.51786411e-01 -9.29753304e-01 -8.15937281e-01 5.34220040e-01
5.13865910e-02 -4.89474267e-01 -1.13707709e+00 2.09559774e+00
3.75472367e-01 7.99158037e-01 1.59952924e-01 3.70502949e-01
4.97867584e-01 5.42683899e-01 6.56154975e-02 -1.66998252e-01
1.37732756e+00 -2.11038888e-01 -1.34535581e-01 2.72462457e-01
3.49003285e-01 -4.62251097e-01 9.07964408e-01 8.39870930e-01
-7.94504821e-01 -6.64900780e-01 -1.53169549e+00 6.65563643e-01
-5.39665341e-01 -1.45418823e-01 3.99301231e-01 1.22189498e+00
-5.59404671e-01 5.73112965e-01 -7.66854048e-01 2.25924432e-01
6.36624277e-01 8.95995498e-01 -2.45978281e-01 5.37934363e-01
-1.39341688e+00 6.54179752e-01 3.75311911e-01 -3.55722785e-01
-1.83175385e+00 -9.98324215e-01 -4.24878418e-01 6.77923560e-02
8.44911113e-02 -3.81239653e-01 7.79199839e-01 -7.71423161e-01
-1.54979515e+00 5.91335177e-01 8.60674977e-01 -1.00391495e+00
2.37448215e-01 -1.40133113e-01 -6.67840064e-01 4.66890007e-01
-2.78185129e-01 5.46634197e-01 1.59591758e+00 -1.05953741e+00
-4.04917032e-01 -4.21448767e-01 -1.62311479e-01 -4.08461332e-01
-8.88113976e-01 5.65864965e-02 6.51032329e-01 -1.05556118e+00
-1.78629279e-01 -1.01564491e+00 2.59405375e-01 -5.27983785e-01
-6.43936217e-01 7.01439604e-02 1.25133026e+00 -3.82513493e-01
1.23212934e+00 -1.96317530e+00 -3.17308277e-01 5.91856897e-01
6.28456995e-02 9.97025490e-01 -2.17063814e-01 2.62665868e-01
-4.77701217e-01 3.07618231e-01 6.49319962e-02 4.83294606e-01
-3.40182841e-01 1.03204355e-01 -1.14586353e+00 4.84883726e-01
1.66044489e-01 6.23437285e-01 -5.06544292e-01 -1.95315052e-02
-1.67034212e-02 4.31138754e-01 -5.54835796e-01 1.55777544e-01
4.78767157e-02 6.26475215e-01 -7.48376369e-01 5.70392907e-01
4.76415694e-01 2.79540122e-01 -1.86776653e-01 -1.86586753e-01
3.06295246e-01 5.81301041e-02 -8.95032048e-01 8.63461912e-01
-3.49045157e-01 3.75739455e-01 -8.76980554e-03 -1.36127949e+00
9.79937673e-01 4.58557278e-01 3.20043802e-01 -5.95461786e-01
5.30316412e-01 1.00326136e-01 3.13397706e-01 -4.23899382e-01
-1.72678351e-01 1.20010920e-01 -2.70741880e-01 7.11099684e-01
4.13152337e-01 -5.52196391e-02 -6.95332289e-01 4.14235480e-02
1.20807815e+00 -3.19386199e-02 -2.60513335e-01 -5.89066923e-01
5.03956258e-01 -1.46684840e-01 5.81814528e-01 9.13567126e-01
-2.39680961e-01 1.49489148e-02 2.41769001e-01 -4.72715169e-01
-7.16262400e-01 -1.30252826e+00 -3.30840051e-01 1.29607987e+00
-2.94725955e-01 1.82886384e-02 -1.07552576e+00 -8.14447463e-01
-3.43424976e-02 7.39872038e-01 -9.46252286e-01 -8.65176797e-01
-6.53717279e-01 -7.13229775e-01 1.85509539e+00 5.11544704e-01
6.90614641e-01 -7.95669258e-01 -1.05379117e+00 -3.82429175e-02
3.49919230e-01 -6.50741994e-01 -2.49451071e-01 7.34430313e-01
-9.72712100e-01 -1.10067809e+00 2.09332048e-03 -4.26247895e-01
6.89522326e-01 -2.35887304e-01 4.91097599e-01 2.72300541e-01
-5.72028279e-01 3.58802378e-01 3.55961244e-03 -9.77293193e-01
-3.90548199e-01 -1.41113997e-01 6.77619040e-01 -7.44618177e-02
7.35184774e-02 -1.02369797e+00 -5.15509367e-01 3.96679610e-01
-1.16061962e+00 -6.11214638e-01 4.30597246e-01 1.29939532e+00
3.73611718e-01 2.37843335e-01 1.06330454e+00 -7.45813906e-01
7.02405035e-01 -2.61029989e-01 -9.08305943e-01 2.20208526e-01
-9.09013093e-01 3.97965349e-02 1.02602673e+00 -1.29109812e+00
-6.27543092e-01 -4.18855608e-01 8.18793029e-02 -1.09136736e+00
-1.38154849e-01 -2.27941610e-02 4.16468270e-02 -4.35944647e-01
1.37188053e+00 6.15698695e-01 -2.65159875e-01 -2.36646518e-01
-1.21551909e-01 4.14548606e-01 8.03674340e-01 -1.06149459e+00
1.34976268e+00 3.80713463e-01 1.18354775e-01 -5.43160796e-01
-4.38298196e-01 3.36452812e-01 -6.06203824e-02 3.15237008e-02
7.67972946e-01 -6.44942045e-01 -1.05697370e+00 4.89414096e-01
-6.96136951e-01 -3.15326512e-01 1.00402005e-01 4.65562046e-01
-2.79439330e-01 4.51264381e-02 -6.53345346e-01 -7.44600832e-01
-9.80331719e-01 -1.27682626e+00 5.50172925e-01 2.65567929e-01
7.08384961e-02 -5.68342090e-01 -1.64851919e-01 4.90116507e-01
5.16429365e-01 6.12645566e-01 1.41882336e+00 -1.36634982e+00
-3.15461606e-01 -2.96388924e-01 2.68605828e-01 8.77447069e-01
-1.87926322e-01 3.72952148e-02 -1.38566613e+00 -5.46607673e-01
4.53398198e-01 -5.50749719e-01 4.55901980e-01 2.66757235e-02
1.37016082e+00 -6.78556561e-01 -2.26728410e-01 7.93624043e-01
1.13011146e+00 6.39224589e-01 5.78291595e-01 2.79480070e-01
4.13583249e-01 1.07418492e-01 6.52090609e-02 1.18875742e-01
-3.66711259e-01 2.91086465e-01 5.66933990e-01 3.19909006e-01
3.92030537e-01 -3.00047100e-01 5.76360643e-01 4.23017293e-01
2.65580356e-01 1.25882447e-01 -8.28426957e-01 1.70352712e-01
-1.21198225e+00 -1.13752425e+00 1.78428710e-01 2.29674482e+00
1.14043522e+00 5.51542401e-01 -1.62248462e-01 3.22710037e-01
5.94101310e-01 -5.35063446e-01 -8.49781215e-01 -7.37471700e-01
2.90998705e-02 1.13834655e+00 6.03634417e-01 8.83071423e-02
-1.16216469e+00 8.35221171e-01 4.72990942e+00 1.12185550e+00
-1.81230950e+00 1.03332497e-01 7.91345239e-01 -2.56214201e-01
-1.70425158e-02 -2.59317040e-01 -8.54706526e-01 4.73609716e-01
1.14812398e+00 -9.80404019e-02 3.57430607e-01 9.50522602e-01
-9.46195647e-02 2.26154253e-01 -1.14050829e+00 7.58787692e-01
-6.41363189e-02 -1.45722008e+00 1.44503459e-01 1.04476906e-01
6.87809706e-01 -1.58284768e-01 8.09921086e-01 8.81319582e-01
1.53834000e-01 -1.32899034e+00 5.24118304e-01 -1.74179557e-03
9.20237124e-01 -9.89625812e-01 1.67914048e-01 5.53922117e-01
-6.36237562e-01 -3.74884963e-01 -1.92829445e-01 3.20138961e-01
-8.20677578e-01 1.35409892e-01 -8.76527011e-01 3.98894101e-01
1.05913675e+00 -4.50175643e-01 -4.80491340e-01 3.07466209e-01
-3.43671978e-01 1.30458403e+00 -4.13134426e-01 1.64560631e-01
3.08111995e-01 2.50077695e-01 6.25369549e-01 9.17207837e-01
5.63123450e-02 3.29761952e-01 1.24296509e-01 7.77718008e-01
-1.23004057e-01 -3.60903442e-01 -5.79585195e-01 1.71464175e-01
7.55606294e-01 1.18219030e+00 -3.92631352e-01 -3.53872746e-01
3.87323916e-01 4.64502215e-01 1.15984663e-01 3.26504350e-01
-1.32068706e+00 -7.49698281e-01 6.59132898e-01 -1.74128473e-01
2.34206110e-01 -3.34918015e-02 -2.82929659e-01 -9.57582712e-01
-1.92910030e-01 -1.41197956e+00 7.02412486e-01 -4.84787464e-01
-1.17171180e+00 8.47378790e-01 2.72261411e-01 -1.12073123e+00
-2.33189911e-01 -5.01870990e-01 -9.11738217e-01 7.84403682e-01
-1.04666650e+00 -1.14215159e+00 1.40140116e-01 1.14356470e+00
9.45360214e-02 -5.37213206e-01 1.00469255e+00 2.30560035e-01
-7.90156245e-01 1.44755459e+00 -1.30217671e-01 3.89420599e-01
6.00035012e-01 -7.80445158e-01 -1.27946004e-01 9.29275990e-01
1.43622383e-01 1.17952418e+00 5.38334072e-01 -5.46681464e-01
-1.52988708e+00 -1.29729581e+00 -1.63847834e-01 -5.16734123e-01
6.02897465e-01 -4.01203364e-01 -1.04805779e+00 6.65302277e-01
1.18312456e-01 -6.10452741e-02 9.88093674e-01 -4.00402933e-01
-8.44361901e-01 -3.93308073e-01 -1.81902230e+00 6.31014526e-01
4.17931050e-01 -5.72551191e-01 -6.84830546e-01 4.00216617e-02
5.82812607e-01 -4.00484771e-01 -9.81554806e-01 6.13348246e-01
4.73757982e-01 -5.92798293e-01 1.10520589e+00 -1.03300786e+00
2.72955233e-03 -1.18577614e-01 -1.47474766e-01 -1.00702477e+00
1.06385000e-01 -1.15768671e+00 -2.70939469e-01 1.14744818e+00
4.01431173e-01 -8.36825371e-01 7.11708546e-01 3.41083705e-01
-1.90798372e-01 -8.46424103e-01 -1.34260476e+00 -1.01313734e+00
5.81180751e-01 -4.36785430e-01 4.76396918e-01 1.24915862e+00
-2.03393459e-01 7.67163634e-02 -3.10271204e-01 6.86874807e-01
7.55097270e-01 -3.46325994e-01 7.54525542e-01 -8.24957371e-01
-3.92728001e-01 -2.11287722e-01 -3.50098312e-01 -2.54552454e-01
2.48191729e-02 -8.80648077e-01 -4.41313684e-01 -1.33189678e-01
-3.08726221e-01 -5.38319349e-01 -8.50316226e-01 9.27942216e-01
1.47045888e-02 4.95881140e-01 -1.06198162e-01 -4.13821824e-03
1.50776222e-01 1.58193097e-01 7.92191684e-01 -1.63798720e-01
-7.72635415e-02 1.62825547e-02 -6.53509676e-01 5.59823811e-01
9.42773044e-01 -9.33372080e-01 -7.93202221e-01 -2.55879045e-01
-2.10393593e-02 -3.46938446e-02 6.85186386e-01 -1.09401190e+00
2.52961278e-01 -1.82373092e-01 7.65690506e-01 -2.57461488e-01
2.01359853e-01 -7.51458406e-01 -6.76147565e-02 8.85022283e-01
-3.51856351e-01 2.41590723e-01 7.13727713e-01 4.80151623e-01
2.62318462e-01 -7.83127770e-02 1.10433090e+00 1.67596549e-01
-1.53625444e-01 2.73521692e-01 -4.45560634e-01 1.82846487e-02
1.36836600e+00 -2.45004058e-01 -5.94241858e-01 -3.96642685e-02
-5.33063293e-01 5.80970757e-02 -2.59008676e-01 3.81713212e-01
7.22158253e-01 -8.80783379e-01 -5.82421243e-01 7.13937044e-01
-1.00272745e-01 -4.59709585e-01 2.32775033e-01 7.06026018e-01
-1.08124264e-01 2.86235809e-01 -5.60182512e-01 -6.42129540e-01
-1.26872635e+00 6.44263983e-01 5.01026690e-01 -7.48453215e-02
-2.73803085e-01 9.64138806e-01 8.88041407e-02 -2.84278780e-01
4.19658393e-01 -3.12558621e-01 1.00073569e-01 -4.11852151e-01
4.18908298e-01 4.14471477e-01 1.72837391e-01 -2.71969959e-02
-3.05575997e-01 1.61508769e-01 -3.65224719e-01 -2.86334492e-02
1.27623880e+00 6.93202078e-01 1.24444775e-01 -2.38486648e-01
1.00433052e+00 -9.34502482e-02 -1.06921661e+00 1.71665296e-01
-2.06954643e-01 3.86041701e-02 -6.24863766e-02 -1.25289929e+00
-9.97545898e-01 1.00639296e+00 9.03508544e-01 4.21514325e-02
1.35710263e+00 -8.55218649e-01 8.82914662e-01 5.96759081e-01
5.08542001e-01 -9.74546969e-01 3.77611607e-01 3.01944405e-01
5.72760463e-01 -7.19722688e-01 -9.26377401e-02 1.91292986e-01
-3.17045957e-01 1.14484251e+00 1.08943903e+00 -5.89950085e-01
7.84126818e-01 5.26147783e-01 1.02888756e-01 -6.69743568e-02
-1.00976038e+00 6.84887707e-01 3.38674694e-01 8.96478891e-01
-3.32181484e-01 -6.75224662e-02 4.29819003e-02 9.80931044e-01
-2.64510602e-01 -2.57731229e-01 3.73545676e-01 9.86354172e-01
-3.36577237e-01 -1.02903771e+00 -6.77442789e-01 5.99686325e-01
-8.99488151e-01 -1.63240626e-01 -2.35849306e-01 7.40251780e-01
4.21200991e-01 9.94873345e-01 -4.03442353e-01 -8.20677459e-01
5.56573831e-02 2.77953416e-01 5.26581049e-01 -5.00548661e-01
-1.53053677e+00 2.04777215e-02 -3.11142713e-01 -4.04445797e-01
5.30443668e-01 -9.16784182e-02 -1.26682222e+00 -9.88855287e-02
-4.93293852e-01 3.40355068e-01 6.67907357e-01 6.11206412e-01
5.52087188e-01 5.17198384e-01 8.51576746e-01 -3.55771333e-01
-1.35734487e+00 -8.45335960e-01 -1.31819338e-01 2.40277484e-01
2.56874979e-01 -6.04385853e-01 -4.37562644e-01 3.35237980e-02] | [5.5751953125, 7.8257670402526855] |
8099dc34-dc9e-4d80-9c8b-a47f054eed97 | lessons-from-computational-modelling-of | 2011.07398 | null | https://arxiv.org/abs/2011.07398v2 | https://arxiv.org/pdf/2011.07398v2.pdf | Lessons from Computational Modelling of Reference Production in Mandarin and English | Referring expression generation (REG) algorithms offer computational models of the production of referring expressions. In earlier work, a corpus of referring expressions (REs) in Mandarin was introduced. In the present paper, we annotate this corpus, evaluate classic REG algorithms on it, and compare the results with earlier results on the evaluation of REG for English referring expressions. Next, we offer an in-depth analysis of the corpus, focusing on issues that arise from the grammar of Mandarin. We discuss shortcomings of previous REG evaluations that came to light during our investigation and we highlight some surprising results. Perhaps most strikingly, we found a much higher proportion of under-specified expressions than previous studies had suggested, not just in Mandarin but in English as well. | ['Kees Van Deemter', 'Guanyi Chen'] | 2020-11-14 | null | https://aclanthology.org/2020.inlg-1.33 | https://aclanthology.org/2020.inlg-1.33.pdf | inlg-acl-2020-12 | ['referring-expression-generation'] | ['computer-vision'] | [ 1.30899519e-01 5.87499201e-01 -5.00896014e-02 -5.70205986e-01
-1.06204808e+00 -8.81390274e-01 6.73327506e-01 -7.19029009e-02
-2.54435390e-01 1.08617210e+00 8.26865613e-01 -5.90285361e-01
-9.51702744e-02 -4.42736119e-01 -3.45633000e-01 -2.53696978e-01
1.79706618e-01 2.71864712e-01 -6.87223673e-02 -7.19310701e-01
4.16855216e-01 4.94073778e-01 -1.07319748e+00 1.90240681e-01
5.56500018e-01 2.03329474e-01 5.34452684e-02 3.33370715e-01
-2.76688576e-01 1.29346466e+00 -1.16817176e+00 -7.54071236e-01
-2.98612952e-01 -7.61059284e-01 -1.33199584e+00 8.17075893e-02
1.18426286e-01 1.72467008e-01 -1.19284309e-01 9.08573687e-01
3.93158346e-01 3.29994112e-01 3.41963917e-01 -7.36504257e-01
-1.13121665e+00 1.20165265e+00 -1.32468462e-01 4.85326856e-01
8.76048386e-01 -1.38544470e-01 1.00884354e+00 -6.47235096e-01
1.16130769e+00 1.28485370e+00 4.55164999e-01 1.01923883e+00
-9.41617429e-01 -1.99474916e-01 1.13546968e-01 -3.66536885e-01
-1.58830225e+00 -7.24930346e-01 7.14458525e-01 -1.84012875e-01
1.10307205e+00 2.80137271e-01 4.51151669e-01 7.50197649e-01
-7.73578063e-02 8.15193057e-01 1.07416832e+00 -9.49328899e-01
-1.24595210e-01 2.53390744e-02 2.52230614e-01 9.87380818e-02
-4.95599173e-02 -2.71434277e-01 -3.28279108e-01 -1.58261135e-02
7.45638907e-01 -8.54218125e-01 -3.12867343e-01 5.72472751e-01
-1.21557975e+00 7.90029705e-01 2.92381961e-02 1.15196335e+00
-3.80973428e-01 5.38524054e-03 4.91985172e-01 1.63900256e-01
3.75768363e-01 4.80177104e-01 -3.98915708e-01 -5.29688120e-01
-4.90876675e-01 3.58833462e-01 8.13788533e-01 1.45388019e+00
3.90083671e-01 1.20630585e-01 -2.61464883e-02 8.52520823e-01
8.46494064e-02 1.23158835e-01 2.34857708e-01 -1.25461304e+00
4.40260917e-01 3.31602126e-01 3.22437584e-01 -7.23020613e-01
-1.89165875e-01 2.13623151e-01 -1.84155256e-02 -4.92718011e-01
5.24629712e-01 -4.85381633e-01 -2.20666021e-01 1.92724979e+00
9.84335467e-02 -9.61179018e-01 3.99319649e-01 6.99531555e-01
9.68434334e-01 4.72588599e-01 5.96296310e-01 -5.21645486e-01
1.56672502e+00 -3.79800022e-01 -1.27122641e+00 1.83999501e-02
1.11403096e+00 -1.17760015e+00 1.17863321e+00 7.18581453e-02
-1.43222940e+00 -2.53166974e-01 -6.42192721e-01 -3.25497270e-01
-3.68802488e-01 -2.87086386e-02 1.09793520e+00 8.40649843e-01
-1.15919554e+00 3.94726634e-01 -5.39559245e-01 -7.63076663e-01
-1.87366512e-02 3.12884212e-01 -2.01092437e-01 2.82641917e-01
-1.15706682e+00 1.12837410e+00 3.13172638e-01 2.32461572e-01
2.21225172e-01 -3.54483053e-02 -1.05414224e+00 -6.75142825e-01
3.28385681e-01 -2.44490534e-01 1.68639886e+00 -1.18978167e+00
-1.34855509e+00 1.48894548e+00 -7.36086309e-01 -1.47918567e-01
-8.44944939e-02 -2.53862560e-01 -8.85127246e-01 -1.35457201e-03
4.13254052e-01 6.47821128e-01 -6.16434477e-02 -1.01007438e+00
-5.95830679e-01 -2.35786185e-01 5.16183853e-01 -5.14928717e-03
6.38392508e-01 1.29776800e+00 -2.50374436e-01 -8.99901807e-01
1.44505188e-01 -6.50633812e-01 3.84173505e-02 -9.34234440e-01
-1.03874758e-01 -8.83019090e-01 4.41661388e-01 -5.18737495e-01
1.65724409e+00 -2.33505797e+00 -2.26202339e-01 -2.61609405e-01
-1.77741662e-01 -8.84622559e-02 3.19912404e-01 6.56778216e-01
-4.59170878e-01 5.83621800e-01 -2.05333278e-01 2.50979681e-02
1.41558439e-01 6.26788318e-01 -3.50174397e-01 3.16025406e-01
6.70652747e-01 1.28469419e+00 -9.13539231e-01 -7.95663774e-01
-5.45472093e-02 4.57857043e-01 -1.24617323e-01 -5.83384447e-02
-1.28431320e-01 4.05984044e-01 -4.82750148e-01 1.06966496e+00
1.64461091e-01 2.91976422e-01 2.38113448e-01 2.99256835e-02
-5.79760969e-01 9.35643315e-01 -6.81899905e-01 1.43214548e+00
-3.32791775e-01 6.77499950e-01 -3.74949016e-02 -6.08900726e-01
8.09857607e-01 7.23304331e-01 8.54538903e-02 -6.00271761e-01
3.61260921e-01 4.94685978e-01 1.94595203e-01 -3.89217079e-01
9.91937041e-01 -6.05254412e-01 -6.30121291e-01 7.67759740e-01
-1.32625401e-01 -4.44732666e-01 4.42208856e-01 7.89338648e-02
8.00723612e-01 7.06795573e-01 7.90724695e-01 -5.66180408e-01
6.60702527e-01 5.44550478e-01 4.43643600e-01 3.13775122e-01
-1.08491614e-01 5.17937839e-01 6.69508815e-01 -5.09399295e-01
-6.08325779e-01 -6.96804941e-01 -5.81916273e-01 1.02583122e+00
-2.68058747e-01 -7.12027967e-01 -1.19904685e+00 -5.58258235e-01
-8.33904028e-01 1.52806902e+00 -6.35830760e-01 3.72035056e-01
-1.33465445e+00 -7.35489428e-01 7.87033617e-01 6.15140855e-01
-6.16514981e-02 -1.59943461e+00 -7.76225507e-01 4.34587926e-01
-4.11179155e-01 -1.39429343e+00 -2.31255516e-01 1.76130250e-01
-7.41021514e-01 -9.44498897e-01 -2.97910988e-01 -8.37774277e-01
5.34651458e-01 -1.44730911e-01 1.30732358e+00 2.19462588e-01
3.36066753e-01 3.93504322e-01 -6.84883595e-01 -6.57114387e-01
-8.07554364e-01 -1.03676476e-01 -4.75656539e-01 -7.34022915e-01
1.03908575e+00 -2.26127759e-01 2.84564823e-01 3.03565264e-01
-1.01761138e+00 -4.61908787e-01 5.07454611e-02 4.54488516e-01
4.85316157e-01 -6.10403419e-01 5.87069094e-01 -1.21834219e+00
8.00006449e-01 -3.84213090e-01 -4.05213684e-01 2.15455946e-02
-9.54069123e-02 3.71673424e-03 1.68509930e-01 -1.11792631e-01
-1.21763492e+00 -2.70179838e-01 -5.06479084e-01 4.94453281e-01
-2.56734610e-01 3.83888394e-01 -2.40553871e-01 1.75594434e-01
6.74677134e-01 -2.40419313e-01 -2.60241657e-01 -2.71083385e-01
3.55234832e-01 6.84909284e-01 6.54943109e-01 -1.11851156e+00
1.74519420e-01 4.88750041e-02 -1.43701375e-01 -9.48995173e-01
-9.56996381e-01 -8.67140964e-02 -7.16708064e-01 -1.02598295e-01
8.15554619e-01 -8.36540401e-01 -3.71433824e-01 -7.58382380e-02
-1.55641544e+00 -4.74072874e-01 -7.04974830e-01 3.30145478e-01
-8.70108843e-01 2.81970978e-01 -6.99100256e-01 -1.00898576e+00
-2.35531494e-01 -8.86884272e-01 1.15581429e+00 -5.68298018e-03
-1.27168155e+00 -1.24804997e+00 -1.78136602e-01 -1.28784463e-01
4.33061123e-01 4.67135161e-01 1.00215876e+00 -6.06434762e-01
-1.70849591e-01 1.66044250e-01 -6.37521669e-02 -9.18751135e-02
4.07035679e-01 1.42357051e-01 -6.87080801e-01 4.69115555e-01
3.59737664e-01 -1.90099359e-01 -9.01771039e-02 2.37612531e-01
2.85605013e-01 -1.50358260e-01 -1.37778064e-02 1.83486119e-01
1.35477841e+00 5.92019320e-01 7.26983845e-01 7.15514958e-01
1.58808842e-01 1.17425716e+00 1.03937888e+00 7.85144642e-02
5.39338291e-01 6.82062984e-01 -1.43988896e-02 2.65378118e-01
-1.89525649e-01 -8.46368745e-02 4.01480198e-01 6.04820967e-01
-4.99489427e-01 -2.11146876e-01 -8.00303221e-01 8.70111108e-01
-1.56855154e+00 -6.30502701e-01 -5.44861794e-01 1.87329590e+00
9.19821262e-01 -9.26211327e-02 3.04369658e-01 -9.77796316e-02
9.50072050e-01 1.59200951e-01 4.35026556e-01 -8.39164495e-01
-6.76607192e-01 7.77480662e-01 -6.57882169e-02 6.94047034e-01
-8.11460197e-01 1.59131408e+00 7.93099546e+00 2.81505287e-01
-7.65082002e-01 1.53018832e-01 3.48026991e-01 1.60290346e-01
-4.77543443e-01 3.19881767e-01 -8.35185707e-01 -8.98979604e-02
1.20276666e+00 -2.72161365e-01 2.12943941e-01 7.22082794e-01
2.37324908e-01 -2.18406528e-01 -1.14770019e+00 7.53691435e-01
7.59081542e-02 -1.00674582e+00 -1.11933261e-01 -2.89311439e-01
4.51729238e-01 -9.59818959e-02 -5.58241248e-01 1.72065005e-01
1.80904225e-01 -8.01572442e-01 9.76185799e-01 2.75845468e-01
5.74438572e-01 -8.38074028e-01 1.04038143e+00 -1.49398595e-01
-7.69003510e-01 4.09364492e-01 -2.33874232e-01 -4.57859904e-01
7.32511938e-01 -3.68702412e-03 -1.38076127e-01 4.17823344e-01
4.96533245e-01 3.20149124e-01 -4.63111877e-01 3.80875975e-01
-6.83831990e-01 7.03846931e-01 -3.65029633e-01 -4.54655707e-01
4.83861744e-01 -3.74099612e-01 6.60056949e-01 1.41034853e+00
1.06525488e-01 6.71839774e-01 -4.76109326e-01 9.80498075e-01
2.10656762e-01 6.16193414e-01 -8.01778615e-01 -1.76583841e-01
3.73914123e-01 1.09223223e+00 -7.53298759e-01 -3.75883639e-01
-5.13942063e-01 7.81353593e-01 2.13412836e-01 2.32266203e-01
-7.36582220e-01 -4.66688395e-01 2.54736453e-01 2.35259399e-01
1.29844353e-01 -2.92119116e-01 -8.33434463e-02 -7.18429387e-01
1.86283395e-01 -9.03009534e-01 4.23025757e-01 -1.16111863e+00
-1.08339739e+00 9.68322575e-01 4.93368745e-01 -7.09721446e-01
-8.29532266e-01 -5.18846929e-01 -6.22695923e-01 1.16783214e+00
-1.05684030e+00 -1.07864642e+00 2.87034094e-01 3.86805058e-01
4.82287437e-01 2.39060104e-01 1.16217363e+00 2.80822683e-02
-5.10862947e-01 3.66110116e-01 -6.86919987e-01 4.08604592e-01
5.66176057e-01 -1.13341534e+00 6.28819942e-01 8.56803358e-01
2.15370893e-01 1.31904638e+00 8.34441960e-01 -5.22562683e-01
-1.07088995e+00 -4.94004518e-01 1.77628624e+00 -6.67538762e-01
9.52348411e-01 7.29445294e-02 -9.11924362e-01 1.22458053e+00
7.72569656e-01 -2.48652980e-01 8.03060174e-01 1.56122461e-01
1.22527972e-01 5.92369258e-01 -1.19692791e+00 8.64757895e-01
1.28524899e+00 -5.79603076e-01 -1.38600910e+00 1.47558138e-01
7.42842615e-01 -6.40090764e-01 -1.02863431e+00 1.01581655e-01
3.95295620e-01 -8.90124738e-01 1.68686718e-01 -8.47969353e-01
3.22141290e-01 -3.33710015e-01 -4.49763447e-01 -7.57293820e-01
-1.65399369e-02 -1.29069769e+00 4.93477464e-01 1.75461709e+00
5.25951207e-01 -5.10767400e-01 4.27975506e-01 1.05689430e+00
-4.72909242e-01 -4.95248258e-01 -7.97646105e-01 -4.46405500e-01
3.73638570e-01 -8.64199698e-01 6.80577934e-01 9.20352757e-01
6.93178177e-01 7.35712945e-01 3.29756230e-01 -4.88796383e-01
1.13570862e-01 -6.50891140e-02 5.37982762e-01 -8.15593183e-01
-9.33968350e-02 -3.27326000e-01 -4.05502677e-01 -9.92283046e-01
4.81455088e-01 -5.71066976e-01 -4.75795567e-03 -1.26872909e+00
-2.58139133e-01 -3.46973479e-01 4.45221603e-01 6.10723734e-01
-6.40707612e-02 1.71296641e-01 1.98725387e-01 7.22488910e-02
-2.99959451e-01 8.48082826e-02 1.09031010e+00 5.29826641e-01
-3.46912861e-01 -4.40741107e-02 -1.34103918e+00 9.27919924e-01
9.64599729e-01 -5.09024799e-01 7.91973397e-02 -5.69768369e-01
2.84444213e-01 4.91491407e-02 -1.01575777e-01 -3.76374692e-01
-1.88883916e-01 -1.23846769e-01 -1.26660764e-01 -6.96999013e-01
-9.92126986e-02 -6.41934812e-01 2.62113184e-01 -1.61398008e-01
1.95604041e-02 7.20006585e-01 5.67654192e-01 -4.66141671e-01
-4.67934370e-01 -8.27003002e-01 4.45037901e-01 -3.20255429e-01
-7.73475647e-01 -6.03007376e-01 -8.25715005e-01 4.15762275e-01
8.41357231e-01 -5.01249373e-01 -3.52226317e-01 -4.06834006e-01
-8.16277325e-01 -1.68025121e-01 6.07648134e-01 2.63746321e-01
2.50432670e-01 -1.25880134e+00 -4.20812726e-01 -3.25651497e-01
1.79003626e-01 8.00237805e-03 -2.47281060e-01 1.02911186e+00
-6.38527453e-01 5.96111417e-01 -3.12746922e-03 4.86920923e-02
-9.08782065e-01 6.33373857e-01 2.20390171e-01 1.11254983e-01
-5.17992973e-01 3.25313836e-01 -1.96838558e-01 -2.37701952e-01
-2.49010444e-01 -3.35194498e-01 -3.42760414e-01 1.87038686e-02
4.81027216e-01 1.28263772e-01 -2.29952037e-01 -1.63021469e+00
-3.89880687e-01 6.34667277e-01 1.06905736e-01 -4.74955380e-01
1.09927130e+00 -3.42895240e-01 -6.81223333e-01 8.78892839e-01
7.73313522e-01 7.06024706e-01 -1.66961059e-01 3.60047147e-02
4.42659110e-01 -5.13897017e-02 -4.37549651e-01 -3.95437360e-01
-4.60088640e-01 4.06927973e-01 -2.20353499e-01 4.97885287e-01
1.08917880e+00 6.25260890e-01 5.74715674e-01 3.73061985e-01
4.43177491e-01 -1.29075241e+00 -7.97128797e-01 7.33186543e-01
9.49752092e-01 -6.68250918e-01 -7.50130862e-02 -7.27650583e-01
-8.28268051e-01 1.26135421e+00 2.94741631e-01 6.88946992e-02
3.00529689e-01 2.62859434e-01 4.89348441e-01 -6.82321966e-01
-4.80752677e-01 -3.72294635e-01 -2.25995734e-01 6.83125079e-01
1.43312538e+00 -1.01430088e-01 -1.24026120e+00 7.72302628e-01
-7.88145781e-01 -8.09894204e-02 7.91123271e-01 1.32368648e+00
-4.30295616e-02 -1.51199341e+00 -5.78174055e-01 -3.41332108e-02
-1.24985921e+00 -4.85933334e-01 -8.22330713e-01 1.55359602e+00
-3.56137864e-02 1.19760537e+00 3.63760442e-01 8.01052600e-02
8.22717309e-01 3.04217041e-01 7.15278625e-01 -7.71330297e-01
-8.48235786e-01 2.67242342e-01 9.50307310e-01 -4.40569192e-01
-1.21336496e+00 -1.15407634e+00 -1.59148514e+00 -1.94135323e-01
-4.61117804e-01 5.56844771e-01 3.76421720e-01 1.17474163e+00
-1.43564984e-01 2.51454055e-01 4.70306538e-02 -4.13049549e-01
-4.01195548e-02 -1.10196102e+00 -7.39817977e-01 2.52945304e-01
-2.87823379e-01 -3.96041930e-01 -4.58780408e-01 2.87317246e-01] | [10.413121223449707, 9.192953109741211] |
88f29f25-0a09-45b3-b77f-4ecf113b418f | uniflg-unified-facial-landmark-generator-from | 2302.14337 | null | https://arxiv.org/abs/2302.14337v2 | https://arxiv.org/pdf/2302.14337v2.pdf | UniFLG: Unified Facial Landmark Generator from Text or Speech | Talking face generation has been extensively investigated owing to its wide applicability. The two primary frameworks used for talking face generation comprise a text-driven framework, which generates synchronized speech and talking faces from text, and a speech-driven framework, which generates talking faces from speech. To integrate these frameworks, this paper proposes a unified facial landmark generator (UniFLG). The proposed system exploits end-to-end text-to-speech not only for synthesizing speech but also for extracting a series of latent representations that are common to text and speech, and feeds it to a landmark decoder to generate facial landmarks. We demonstrate that our system achieves higher naturalness in both speech synthesis and facial landmark generation compared to the state-of-the-art text-driven method. We further demonstrate that our system can generate facial landmarks from speech of speakers without facial video data or even speech data. | ['Kei Sawada', 'Yukiya Hono', 'Kentaro Mitsui'] | 2023-02-28 | null | null | null | null | ['talking-face-generation', 'face-generation', 'speech-synthesis'] | ['computer-vision', 'computer-vision', 'speech'] | [ 4.34297800e-01 6.37721479e-01 2.41914347e-01 -6.29129052e-01
-1.06651533e+00 -2.27927864e-01 9.61195171e-01 -8.74510169e-01
3.56013447e-01 3.49732816e-01 6.02524042e-01 1.82563350e-01
6.07110977e-01 -6.65325999e-01 -5.00660956e-01 -6.32034421e-01
2.57268190e-01 3.13487351e-01 -7.97111094e-02 -9.56970304e-02
-5.68219200e-02 5.30583858e-01 -1.97451782e+00 2.93256581e-01
4.49738145e-01 1.01234448e+00 1.56164896e-02 6.00606740e-01
-5.53233385e-01 4.66792405e-01 -5.85030079e-01 -4.64387268e-01
1.40310302e-01 -6.62502408e-01 -5.20318985e-01 5.27592719e-01
6.13829553e-01 -4.00812030e-01 -3.52427185e-01 7.75253296e-01
7.25096047e-01 -4.80478704e-02 7.40503490e-01 -1.34859848e+00
-5.16728461e-01 5.23211241e-01 -4.57619816e-01 -4.49858755e-01
7.73387372e-01 -1.48637488e-01 5.01495004e-01 -1.28932059e+00
7.31668770e-01 1.72756517e+00 3.15245539e-01 1.09864426e+00
-8.47732961e-01 -1.03726697e+00 4.92850952e-02 -3.88203025e-01
-1.74816346e+00 -1.62238252e+00 8.88031006e-01 -3.54909003e-01
6.03494823e-01 5.34494929e-02 4.06695157e-01 1.15801656e+00
1.35177091e-01 5.67760408e-01 6.74406826e-01 -5.59654832e-01
4.22246754e-02 -4.90031093e-02 -7.30877280e-01 1.03951633e+00
-4.91346866e-01 1.64752781e-01 -1.02157581e+00 -9.91749316e-02
8.22848439e-01 -3.39879423e-01 -8.38968828e-02 1.73537612e-01
-1.11138988e+00 5.89940786e-01 -1.18330225e-01 4.60804924e-02
-3.56375933e-01 1.90970600e-01 2.52379060e-01 3.70326489e-02
6.43291354e-01 -4.16521519e-01 3.67301941e-01 -7.20938668e-02
-1.42261112e+00 2.88369171e-02 7.76321411e-01 1.39100766e+00
6.01642311e-01 4.98513728e-01 -2.19807535e-01 9.88543391e-01
7.82172441e-01 8.38711679e-01 5.90222836e-01 -8.48403752e-01
3.10277045e-01 2.15822116e-01 -2.93774068e-01 -1.00907171e+00
-9.86084267e-02 3.64010632e-01 -6.28504574e-01 8.93847197e-02
-2.16900513e-01 -3.37993950e-01 -1.11536479e+00 1.81710958e+00
4.93641168e-01 4.30110455e-01 2.49436453e-01 7.07038164e-01
1.32144129e+00 7.54691064e-01 -8.99623185e-02 -2.52445191e-01
1.31687915e+00 -9.61452842e-01 -9.49245334e-01 -1.33259743e-01
1.00829445e-01 -1.09621012e+00 8.35753024e-01 1.03081286e-01
-1.25262308e+00 -6.82117343e-01 -7.73049712e-01 -2.23573670e-01
-1.26106858e-01 5.80708683e-01 1.57156855e-01 8.57545137e-01
-1.42014956e+00 -3.35601680e-02 -6.43310964e-01 -5.28278589e-01
3.30330670e-01 4.17797685e-01 -6.11230731e-01 2.90592760e-01
-8.27921808e-01 5.61915219e-01 -1.24606177e-01 1.50453508e-01
-1.10428011e+00 -2.11428136e-01 -1.17729902e+00 4.69576903e-02
2.34697685e-01 -7.22736835e-01 1.44220352e+00 -9.69973981e-01
-2.40849638e+00 9.66223359e-01 -9.27692354e-01 -3.21664698e-02
3.43975544e-01 1.32714793e-01 -5.46371102e-01 4.30232286e-01
3.24224859e-01 1.12465954e+00 1.39998579e+00 -1.19682193e+00
-4.32995230e-01 -3.43224615e-01 -4.53899503e-01 2.48571843e-01
-1.82617828e-01 1.97844580e-01 -7.81741202e-01 -6.86798692e-01
2.79283017e-01 -9.45140183e-01 2.07230046e-01 2.46165141e-01
-5.73780596e-01 -7.70285949e-02 1.18239546e+00 -4.42134827e-01
8.20661068e-01 -2.12503910e+00 -9.32411849e-02 1.25244424e-01
-6.24645837e-02 8.67643356e-02 -3.77418697e-01 5.47594309e-01
-9.87951756e-02 4.35481034e-02 5.53551912e-02 -1.03208661e+00
-6.53191656e-02 5.08628972e-02 -6.40629113e-01 4.27188873e-01
1.97382942e-01 6.70550287e-01 -5.20081937e-01 -7.59644568e-01
4.27883536e-01 8.83454382e-01 -3.01570415e-01 1.39736086e-01
-2.13164836e-01 5.18460929e-01 -4.18503106e-01 9.03445303e-01
7.02484250e-01 2.34256759e-01 1.67465791e-01 -4.84460816e-02
-1.87055796e-01 2.04374358e-01 -9.82899725e-01 1.75457013e+00
-7.00168788e-01 5.48264563e-01 4.09017384e-01 -3.07644486e-01
1.41874766e+00 8.59571457e-01 2.62424350e-01 -2.75435507e-01
3.02851468e-01 8.13530982e-02 -5.76643586e-01 -4.60681915e-01
3.90820622e-01 -3.40992779e-01 -1.16869517e-01 3.75778586e-01
3.51402372e-01 -3.99186552e-01 5.86996190e-02 4.40811180e-02
5.93330145e-01 3.46758127e-01 9.69880745e-02 -7.66155049e-02
9.57888126e-01 -5.09661376e-01 3.49880606e-01 -1.66530192e-01
8.38759467e-02 9.56141591e-01 4.17822540e-01 5.72775714e-02
-6.45072162e-01 -1.21234536e+00 1.58569634e-01 1.03734457e+00
-6.61706626e-02 -6.82478428e-01 -1.14661884e+00 -4.47209448e-01
-2.55118638e-01 4.63040560e-01 -4.96191829e-01 4.32488509e-02
-1.72627598e-01 7.65589997e-02 8.74437511e-01 3.54006112e-01
6.34166360e-01 -9.82039690e-01 1.83355194e-02 7.45430496e-03
-3.03004980e-01 -1.37517917e+00 -9.39987957e-01 -7.76113629e-01
-3.67577314e-01 -7.54339337e-01 -8.84720564e-01 -9.92687523e-01
1.00833118e+00 2.35332191e-01 6.61252797e-01 -1.60943776e-01
-2.29483351e-01 5.41232228e-01 -1.92136705e-01 -5.07923245e-01
-7.23141372e-01 -5.06965965e-02 3.35652590e-01 6.22830451e-01
1.31303176e-01 -4.84445393e-01 -3.29686463e-01 5.74814200e-01
-7.67719388e-01 1.42282382e-01 3.95664245e-01 4.38390911e-01
4.87876475e-01 -3.20718318e-01 8.30200732e-01 -5.07948279e-01
6.60585582e-01 -1.33744121e-01 -7.18491077e-01 2.37159237e-01
-6.60788938e-02 -1.64253473e-01 6.87730789e-01 -8.77450258e-02
-1.36046553e+00 5.11041820e-01 -5.47298133e-01 -6.54884934e-01
-2.82257527e-01 2.14505941e-02 -4.56713289e-01 -1.29331186e-01
3.44201684e-01 5.44048488e-01 4.42743123e-01 -2.06242591e-01
6.32689655e-01 1.23796499e+00 7.81510592e-01 -6.27095163e-01
8.57767701e-01 7.02821136e-01 -2.51549985e-02 -1.38208008e+00
-1.88418999e-01 -2.38478884e-01 -7.58915424e-01 -4.94247347e-01
9.65866446e-01 -1.05964303e+00 -7.79279470e-01 5.01427591e-01
-1.22172284e+00 4.73845890e-03 3.82450735e-03 2.84418255e-01
-9.11511362e-01 1.73022404e-01 -3.14143807e-01 -9.98404622e-01
-5.19278884e-01 -1.42410445e+00 1.89868832e+00 2.75168121e-01
-1.83226600e-01 -6.44094408e-01 -9.25278068e-02 2.77604878e-01
3.25650364e-01 2.39279419e-01 2.67031163e-01 -1.49341062e-01
-5.21142066e-01 -6.48828074e-02 -3.70764881e-02 2.71816496e-02
6.10378027e-01 3.81286711e-01 -1.32945812e+00 -4.07368727e-02
-2.95760542e-01 -2.14336485e-01 3.79196435e-01 3.13761085e-01
5.99503279e-01 -3.68177474e-01 -4.70791638e-01 7.82780051e-01
8.20080280e-01 3.91268879e-01 6.36457741e-01 -4.09880221e-01
5.37003994e-01 9.28272486e-01 2.98099339e-01 3.61431718e-01
4.30428594e-01 8.41152251e-01 -3.73509079e-02 -3.75777967e-02
-3.78317624e-01 -5.43978214e-01 6.60333812e-01 8.33716631e-01
2.44911402e-01 -7.86765963e-02 -8.09146762e-01 4.67272639e-01
-1.48209989e+00 -1.04694867e+00 3.02624255e-01 1.74835372e+00
6.65340900e-01 -4.76718873e-01 2.06580222e-01 -1.55526355e-01
9.17059898e-01 3.60579878e-01 -3.28769028e-01 -4.15069610e-01
3.08176696e-01 2.47095779e-01 -6.32118881e-02 6.22594059e-01
-8.74764860e-01 1.46652770e+00 7.07311201e+00 7.71614015e-01
-1.64008009e+00 -7.68875517e-03 5.96046031e-01 -8.07552561e-02
-1.50157586e-01 -3.65797460e-01 -9.70172226e-01 2.54572064e-01
1.05786788e+00 -5.76185703e-01 2.33349189e-01 1.04422593e+00
6.19255245e-01 2.19342053e-01 -1.03054965e+00 1.20759690e+00
6.12068892e-01 -1.45434093e+00 5.20748973e-01 -3.19032222e-02
4.65552807e-01 -4.96881127e-01 1.24513261e-01 -1.19420784e-02
-4.32125330e-02 -1.18921244e+00 9.20343816e-01 4.28742409e-01
1.49156511e+00 -7.39804924e-01 2.16406211e-01 1.21530853e-01
-1.65664744e+00 4.34319317e-01 -1.87203050e-01 4.06018972e-01
4.61013615e-01 2.44087741e-01 -1.38580334e+00 5.19173682e-01
2.58896619e-01 4.55695897e-01 -1.81303471e-01 3.96133721e-01
-2.98734426e-01 2.90026248e-01 -3.32132071e-01 1.45769238e-01
-2.48037558e-02 -2.22220302e-01 3.83291185e-01 1.11241806e+00
7.49501705e-01 1.78206097e-02 2.67540187e-01 9.49707031e-01
-2.62185991e-01 2.57966936e-01 -9.74254489e-01 -1.62391737e-01
7.29263246e-01 1.37714660e+00 -7.17079997e-01 -3.44527543e-01
-3.99724543e-01 1.11329639e+00 -3.32286835e-01 1.82407558e-01
-6.89285576e-01 -5.68391323e-01 7.38108575e-01 2.88665920e-01
3.16392370e-02 -3.29398692e-01 -7.65631795e-02 -9.27710176e-01
4.69426671e-03 -7.63602674e-01 -1.05673127e-01 -9.50612724e-01
-9.05694187e-01 1.20357585e+00 5.14667667e-03 -1.27118886e+00
-8.46705854e-01 -2.53953189e-01 -8.17045510e-01 1.10721684e+00
-1.33380866e+00 -1.68825281e+00 -4.39168364e-01 9.35810030e-01
8.32003295e-01 -5.37435174e-01 9.11656260e-01 1.19327329e-01
-4.82536137e-01 8.68422270e-01 -4.58747745e-01 2.86433816e-01
7.40613937e-01 -6.61414146e-01 9.39363480e-01 8.24534297e-01
1.47050455e-01 7.03386605e-01 2.66617715e-01 -7.16622114e-01
-1.39152241e+00 -1.33337665e+00 9.57265437e-01 -1.06567189e-01
9.20531228e-02 -8.00659716e-01 -2.31292740e-01 7.15676546e-01
1.35626704e-01 -1.17122583e-01 6.73457623e-01 -4.02179420e-01
-2.29293823e-01 -2.31185421e-01 -1.32012987e+00 9.46547031e-01
1.16413248e+00 -8.44733953e-01 -4.38764334e-01 2.67428994e-01
5.39373934e-01 -5.13723314e-01 -6.07221603e-01 1.62639186e-01
7.05276489e-01 -9.17789698e-01 7.37254798e-01 2.33187363e-01
2.90030748e-01 -4.38263893e-01 -1.44526079e-01 -1.24091983e+00
1.78432494e-01 -1.29132533e+00 3.97024006e-01 1.74509215e+00
4.04305011e-01 -6.16803348e-01 9.77790594e-01 3.04489791e-01
-2.71387547e-01 -4.45291489e-01 -1.10307312e+00 -5.08475482e-01
-2.62694746e-01 -2.11014420e-01 9.13413465e-01 6.68131948e-01
6.20745355e-03 4.47007000e-01 -4.26957548e-01 3.28618199e-01
3.40137005e-01 -4.88523245e-02 1.18061078e+00 -8.79846990e-01
4.28749055e-01 -1.78551883e-01 -4.74997759e-01 -1.10686600e+00
5.76593399e-01 -9.38851953e-01 4.14890021e-01 -1.51155865e+00
-4.25378442e-01 -6.57018125e-02 8.31830680e-01 4.43482399e-01
2.94192582e-01 3.41131717e-01 1.46081403e-01 -2.77622906e-03
-5.88835590e-02 9.15219843e-01 1.12431180e+00 1.21971965e-01
-2.54108071e-01 -1.68402746e-01 -6.11795843e-01 8.62272859e-01
3.97202492e-01 -1.48249492e-01 -7.43413627e-01 -2.30496049e-01
-5.49979925e-01 4.53873277e-01 4.07658797e-03 -1.16886067e+00
2.93710530e-01 -2.90150046e-02 3.13219845e-01 -5.86452425e-01
7.86977172e-01 -6.43892765e-01 9.30367559e-02 -9.67643186e-02
-2.54494131e-01 6.51796209e-03 1.66119516e-01 2.64737755e-01
-4.36064482e-01 1.74864009e-01 8.97700608e-01 -7.96854049e-02
-3.35780621e-01 4.70319569e-01 -4.72738445e-01 -3.41252148e-01
1.22154033e+00 -3.77195537e-01 -2.56742299e-01 -9.64343727e-01
-8.91166449e-01 -7.83425272e-02 3.07808965e-01 6.97519898e-01
1.03094995e+00 -1.49138427e+00 -8.10551822e-01 8.80071342e-01
-7.65680522e-03 -1.24085456e-01 1.08262502e-01 4.99545097e-01
-4.89434391e-01 4.89872634e-01 -9.71419960e-02 -7.12362468e-01
-1.56855857e+00 1.94471881e-01 3.49171191e-01 6.09827876e-01
-6.01310730e-01 8.28603148e-01 4.15836632e-01 -3.71023804e-01
2.94525802e-01 -1.63765088e-01 7.45910183e-02 1.21157721e-01
7.54341364e-01 2.84211878e-02 1.35957360e-01 -1.32625842e+00
-3.88704360e-01 9.68220413e-01 1.72468498e-01 -6.24244452e-01
8.23785305e-01 -2.59252578e-01 6.84154630e-02 1.75815061e-01
1.09876812e+00 3.39744687e-01 -1.01514661e+00 9.50358957e-02
-3.75742018e-01 -4.81987029e-01 -1.20108113e-01 -3.26274306e-01
-1.11031175e+00 8.77966046e-01 3.37388873e-01 -1.92680031e-01
1.17256558e+00 -8.18099901e-02 9.40689087e-01 1.07493646e-01
6.43638194e-01 -8.48981559e-01 3.86486240e-02 3.93005401e-01
1.18787241e+00 -8.02829862e-01 -4.10655797e-01 -9.16806340e-01
-6.94499731e-01 1.31328094e+00 4.83448327e-01 6.83597624e-02
7.11779892e-01 5.49359679e-01 4.53891456e-01 -1.29020944e-01
-8.60820651e-01 -2.49905080e-01 3.47497821e-01 8.16252649e-01
5.79752684e-01 -3.20001036e-01 7.90226310e-02 3.22928041e-01
-7.47157753e-01 2.36014992e-01 3.20209205e-01 6.32176340e-01
-3.54958594e-01 -1.09432864e+00 -6.32664025e-01 4.47582453e-02
-3.99233103e-01 -8.28702897e-02 -7.75040388e-01 5.27409852e-01
5.50176576e-03 1.33863020e+00 1.64350420e-01 -6.10985160e-01
2.89339513e-01 4.83540267e-01 2.57217288e-01 -8.75883102e-01
-2.75605500e-01 4.15408701e-01 2.89392740e-01 -6.98799729e-01
-3.43542337e-01 -5.43975294e-01 -1.40450561e+00 -2.55798489e-01
-2.04872176e-01 1.10519238e-01 8.71056795e-01 8.43509912e-01
6.72672212e-01 3.11492503e-01 8.93793881e-01 -1.35436571e+00
-8.14078189e-03 -9.93145227e-01 -5.02124250e-01 2.69232579e-02
2.75864452e-01 -6.83221817e-01 -2.87101924e-01 5.56172669e-01] | [13.235435485839844, -0.40774106979370117] |
67fda977-f4d1-4cd6-8fb7-50836f6dba60 | xbd-a-dataset-for-assessing-building-damage | 1911.09296 | null | https://arxiv.org/abs/1911.09296v1 | https://arxiv.org/pdf/1911.09296v1.pdf | xBD: A Dataset for Assessing Building Damage from Satellite Imagery | We present xBD, a new, large-scale dataset for the advancement of change detection and building damage assessment for humanitarian assistance and disaster recovery research. Natural disaster response requires an accurate understanding of damaged buildings in an affected region. Current response strategies require in-person damage assessments within 24-48 hours of a disaster. Massive potential exists for using aerial imagery combined with computer vision algorithms to assess damage and reduce the potential danger to human life. In collaboration with multiple disaster response agencies, xBD provides pre- and post-event satellite imagery across a variety of disaster events with building polygons, ordinal labels of damage level, and corresponding satellite metadata. Furthermore, the dataset contains bounding boxes and labels for environmental factors such as fire, water, and smoke. xBD is the largest building damage assessment dataset to date, containing 850,736 building annotations across 45,362 km\textsuperscript{2} of imagery. | ['Howie Choset', 'Eric Heim', 'Nirav Patel', 'Ritwik Gupta', 'Matthew Gaston', 'Jigar Doshi', 'Sandra Sajeev', 'Richard Hosfelt', 'Bryce Goodman'] | 2019-11-21 | null | null | null | null | ['2d-semantic-segmentation'] | ['computer-vision'] | [ 3.13930959e-01 -2.07220271e-01 -4.64605540e-02 -5.07197976e-02
-7.72626400e-01 -5.54623485e-01 3.62936348e-01 9.78345990e-01
-6.28081858e-01 5.26142538e-01 1.13615906e+00 -3.90587658e-01
-3.09621811e-01 -1.60480845e+00 -1.97805807e-01 -4.97385651e-01
-3.83549124e-01 1.94577381e-01 -4.62571159e-02 -4.65622038e-01
-8.38322658e-03 8.25050890e-01 -1.22005904e+00 8.37728232e-02
3.15544665e-01 7.37062037e-01 2.37711981e-01 5.86490273e-01
5.10418057e-01 4.11934137e-01 -2.59178638e-01 3.20125729e-01
3.94240111e-01 2.06153661e-01 -9.47572112e-01 5.01716845e-02
3.87740433e-01 -1.21055996e+00 -5.71320713e-01 6.30258560e-01
7.12664247e-01 1.03688268e-02 7.01626241e-01 -8.36784005e-01
-4.33370322e-01 2.72156954e-01 -7.35603392e-01 7.32208550e-01
6.11323118e-01 3.32950979e-01 7.36483634e-01 -8.01693678e-01
3.95396709e-01 9.78742957e-01 1.17765594e+00 -2.06904054e-01
-1.22647834e+00 -3.11853021e-01 -7.14234263e-02 -1.16947696e-01
-1.79962564e+00 -2.90159166e-01 2.82606691e-01 -8.58026862e-01
1.18586814e+00 4.67853725e-01 6.46454036e-01 3.79884779e-01
-2.77350754e-01 1.59775600e-01 8.24795067e-01 -2.67038345e-01
1.20158233e-01 -7.97160208e-01 7.73165300e-02 5.77187240e-01
4.37014341e-01 1.87684759e-01 -2.37629890e-01 -5.86486936e-01
6.47668004e-01 5.40903509e-01 -2.87537187e-01 3.75273019e-01
-1.35686171e+00 6.63868487e-01 1.08888674e+00 1.01985812e-01
-8.93813670e-01 7.68645704e-02 4.98920470e-01 -2.89119184e-01
5.94481289e-01 7.15491474e-02 -9.07984376e-02 3.00422549e-01
-1.11360466e+00 2.39112124e-01 -1.19016230e-01 2.42712334e-01
9.00786400e-01 -1.42297506e-01 1.04445271e-01 6.88878834e-01
2.55000979e-01 8.74696612e-01 -2.66998410e-01 -8.50537777e-01
7.30754614e-01 7.96676874e-01 3.62881124e-01 -1.76519454e+00
-1.00956023e+00 1.23752102e-01 -1.27782404e+00 1.13709070e-01
-1.93558130e-02 -4.17397842e-02 -8.35208476e-01 1.30323792e+00
6.67800426e-01 -2.64731765e-01 -3.35208535e-01 8.48698676e-01
5.47743261e-01 7.44900286e-01 5.80788493e-01 -1.03332393e-03
1.07110894e+00 2.39243656e-02 -1.65502787e-01 -4.74812061e-01
4.40676391e-01 -1.28142029e-01 9.55346406e-01 2.51252595e-02
-4.78429049e-01 -2.57219076e-01 -7.32706189e-01 1.28070250e-01
-4.37215865e-01 -1.70670256e-01 2.25362509e-01 2.37701952e-01
-1.06112111e+00 2.79835373e-01 -8.03777337e-01 -7.99967706e-01
8.16729546e-01 -6.64852038e-02 -5.32007515e-01 -4.11521882e-01
-1.11900949e+00 1.09098101e+00 3.78399074e-01 3.16946149e-01
-1.02155817e+00 -7.37472355e-01 -9.75985825e-01 -2.01172248e-01
-6.42078966e-02 -5.01737237e-01 6.28544867e-01 5.15735000e-02
7.78426304e-02 1.06278849e+00 5.13215959e-01 -3.90698552e-01
2.22069070e-01 -2.48373002e-01 -3.94362837e-01 2.91258752e-01
7.85643399e-01 6.74353540e-01 3.22567344e-01 -1.18440354e+00
-8.65029454e-01 -9.58864927e-01 1.05966404e-02 4.66420621e-01
-3.78169000e-01 3.89764220e-01 4.13095117e-01 -8.66244078e-01
2.32455254e-01 -8.00517619e-01 -3.58521044e-01 2.87671783e-03
-2.15656221e-01 1.88451141e-01 7.16838419e-01 -1.32184947e+00
1.61024892e+00 -2.05520511e+00 -1.87905282e-01 7.53954425e-02
-1.12279862e-01 7.29981810e-02 1.10719450e-01 9.61723447e-01
6.06843308e-02 6.41429961e-01 -8.39960933e-01 1.86840907e-01
-2.14823067e-01 3.48548770e-01 -6.16932273e-01 5.46768904e-01
-2.22844556e-01 4.76513714e-01 -9.98468995e-01 -3.70563984e-01
4.33206409e-01 2.69260913e-01 1.51649686e-02 -1.48768723e-01
2.91690975e-01 2.50161678e-01 -4.53567833e-01 9.76999223e-01
6.94238424e-01 4.61677909e-01 -1.83391973e-01 9.07642320e-02
-6.86108828e-01 1.08822972e-01 -1.07963908e+00 1.28935301e+00
-7.45225996e-02 3.90413254e-01 2.52204597e-01 -5.02048910e-01
6.53379858e-01 2.57964551e-01 8.09284627e-01 -3.94600689e-01
-1.84885606e-01 -9.59260538e-02 -9.92175579e-01 -5.11623919e-01
6.30372345e-01 -3.05669576e-01 -5.85032344e-01 7.79020190e-01
-7.79757082e-01 -3.92742425e-01 -1.54706597e-01 9.59162489e-02
1.52687275e+00 -4.91009831e-01 5.91071963e-01 -2.51292706e-01
-2.95250058e-01 7.06936181e-01 2.86798447e-01 5.01466095e-01
-4.64700252e-01 6.25405073e-01 -2.83898979e-01 -1.10221171e+00
-1.10506153e+00 -1.18622458e+00 -4.25376594e-01 1.15508187e+00
-1.39093801e-01 -9.23092961e-02 -4.87849116e-01 -1.31746143e-01
2.57774264e-01 7.83241332e-01 -4.30022508e-01 5.08523323e-02
-3.81783724e-01 -1.36357856e+00 1.15628028e+00 6.14130497e-01
8.76674235e-01 -1.09746122e+00 -9.52592075e-01 3.35568070e-01
-6.79761708e-01 -5.46764433e-01 -5.21214977e-02 -7.51630813e-02
-8.23751509e-01 -1.25620162e+00 -3.28677058e-01 -5.72393000e-01
6.52145565e-01 7.21849263e-01 8.87849212e-01 4.39620644e-01
-7.82095373e-01 8.66244078e-01 -3.05820495e-01 -9.74164531e-03
-5.47039434e-02 -9.64892209e-02 2.67826498e-01 -5.52574515e-01
1.76655874e-01 -6.86150074e-01 -7.68932164e-01 2.66873330e-01
-1.03860569e+00 -1.10908955e-01 1.09291293e-01 2.21858442e-01
5.71650863e-01 6.14210904e-01 3.47989291e-01 2.49752194e-01
3.92478734e-01 -9.04463172e-01 3.00270710e-02 1.03111535e-01
-2.61366397e-01 -7.96142280e-01 -1.69800967e-01 4.02896464e-01
-9.27270055e-01 2.94865310e-01 6.21313937e-02 6.56927168e-01
-9.67693865e-01 1.02119303e+00 -1.90010853e-02 4.44078624e-01
1.05724001e+00 -2.17543498e-01 -6.36934757e-01 -4.68675017e-01
1.34876564e-01 9.50725198e-01 1.08409154e+00 -3.57001722e-01
8.95492375e-01 8.13319743e-01 -1.35466933e-01 -1.23373663e+00
-9.12988245e-01 -9.40225065e-01 -1.20838833e+00 -4.20253932e-01
8.79407167e-01 -1.24276698e+00 -1.62415832e-01 9.87115681e-01
-9.87126827e-01 -5.62770307e-01 -1.68618768e-01 2.24398464e-01
-3.75908166e-02 2.92058498e-01 -2.87037820e-01 -6.60306573e-01
-4.87743109e-01 -2.11556986e-01 9.63411689e-01 -1.96312934e-01
-2.71992981e-01 -5.42346537e-01 5.64086914e-01 5.30416071e-01
2.69142449e-01 1.19684005e+00 7.81137407e-01 4.31275159e-01
-2.35306248e-01 -2.33389154e-01 -4.77198094e-01 1.20143220e-01
6.97664738e-01 -2.01911870e-02 -3.60913515e-01 -3.20490271e-01
-2.97960758e-01 -2.60347664e-01 1.19395590e+00 4.87354457e-01
5.34086704e-01 -6.21108770e-01 -4.42262590e-01 1.24494463e-01
1.60460484e+00 -1.37722060e-01 9.01607096e-01 7.98208117e-01
6.53476775e-01 5.74728966e-01 6.25791550e-01 9.95846570e-01
7.95650959e-01 4.34931964e-01 9.11600769e-01 -3.04030478e-01
-2.08669007e-02 -7.61149079e-02 2.00327381e-01 -9.25480202e-02
-3.37117434e-01 -1.45730108e-01 -2.11001825e+00 1.04356050e+00
-1.39036191e+00 -1.53327465e+00 -5.81453025e-01 2.21914792e+00
6.90576911e-01 -2.95705378e-01 1.09557033e-01 3.44797969e-01
7.93455422e-01 4.19715464e-01 -3.50739509e-01 3.12513411e-01
-2.02100992e-01 -1.04460016e-01 8.64009917e-01 5.28695643e-01
-1.71024680e+00 5.97867846e-01 7.05845022e+00 1.46215856e-01
-5.23696065e-01 1.79837812e-02 4.52758670e-01 -5.92409261e-02
-1.25782266e-01 -3.01588494e-02 -5.19090593e-01 -5.05878683e-03
5.40310979e-01 2.48127356e-01 4.51292545e-01 6.37823761e-01
7.00713813e-01 -7.18787074e-01 -1.41016051e-01 5.38596630e-01
-6.52633011e-02 -1.05948436e+00 -1.57991111e-01 7.47206807e-02
7.80455589e-01 3.87855738e-01 -2.23270237e-01 -3.49825859e-01
6.09256983e-01 -7.23584533e-01 6.67864084e-01 6.34516835e-01
1.08922291e+00 -8.23534787e-01 4.11517382e-01 5.53192437e-01
-1.41714084e+00 -4.14550871e-01 -8.13729018e-02 -6.86064363e-01
4.82716411e-01 5.42380571e-01 -8.31486821e-01 2.61551142e-01
1.19327497e+00 5.82610846e-01 -8.02279234e-01 8.89609277e-01
-4.67615217e-01 7.28591502e-01 -6.83310509e-01 8.06856036e-01
3.03103328e-01 7.20679164e-02 5.19398928e-01 1.00196958e+00
5.02212703e-01 8.97173882e-01 6.14185452e-01 2.08445668e-01
2.84038663e-01 -3.14290613e-01 -9.56851304e-01 3.59032899e-01
6.28632247e-01 9.33355212e-01 -7.34834731e-01 -8.33731070e-02
4.10851724e-02 8.17452013e-01 8.54130611e-02 1.04563139e-01
-5.67947209e-01 7.36716688e-02 7.07966268e-01 7.36718416e-01
-2.14143395e-01 -8.29729140e-01 -3.06053072e-01 -5.01272619e-01
-3.03638786e-01 -4.50804085e-01 6.65285408e-01 -1.00547731e+00
-1.06542230e+00 1.57400146e-01 4.78739083e-01 -9.55427110e-01
2.85694957e-01 -1.88212041e-02 -5.67969441e-01 6.22700274e-01
-1.24349451e+00 -1.59722090e+00 -8.14517736e-01 7.04426467e-01
1.45277172e-01 2.67991096e-01 1.15974271e+00 2.89674371e-01
-6.56397879e-01 -5.36122441e-01 1.64157912e-01 2.85159796e-01
3.88337821e-01 -8.89084280e-01 6.06775105e-01 7.99436331e-01
-5.04887342e-01 6.49452955e-02 3.13873053e-01 -1.37743843e+00
-6.43353403e-01 -1.66821635e+00 8.44733119e-01 -3.58495951e-01
6.06212199e-01 2.48380125e-01 -6.97140455e-01 7.54683137e-01
-2.68898904e-01 -2.52779514e-01 7.15665877e-01 4.82524261e-02
-1.60621822e-01 -1.46905318e-01 -1.49609017e+00 2.73850948e-01
1.26749837e+00 -7.20873356e-01 -6.22624397e-01 5.16030133e-01
3.81358057e-01 1.75328106e-02 -9.23863530e-01 4.24212575e-01
2.42536455e-01 -7.40709424e-01 1.29472935e+00 -1.15339033e-01
3.19722891e-01 -5.64314544e-01 -7.16674626e-01 -1.19337785e+00
-8.65779221e-01 1.85418963e-01 6.08160436e-01 1.22220898e+00
-1.09451190e-01 -1.08363368e-01 1.30813420e-01 1.02018440e+00
-4.41996217e-01 9.78939682e-02 -1.08672929e+00 -5.31836212e-01
-5.65735921e-02 -6.83995545e-01 6.94325149e-01 8.84657264e-01
-2.35697210e-01 -1.72110900e-01 -2.65396088e-01 1.03042793e+00
6.24478161e-01 -1.53894171e-01 5.16132116e-01 -1.44115162e+00
7.01744854e-01 -1.87803090e-01 -3.79283220e-01 -9.84486639e-02
-3.50024045e-01 -5.15611410e-01 -1.49133271e-02 -2.06406617e+00
3.60807925e-01 -6.70642495e-01 -4.36910316e-02 1.37732506e+00
-9.60011110e-02 3.79871100e-01 -1.14832148e-01 5.48239708e-01
-4.14922424e-02 5.24465203e-01 3.04137290e-01 -6.00040019e-01
-2.06519097e-01 -2.23177537e-01 -4.29231584e-01 8.86140525e-01
9.96852040e-01 -7.49313951e-01 2.07095534e-01 -8.53537977e-01
3.98673803e-01 -1.65260568e-01 1.26984370e+00 -1.33399999e+00
-2.00639479e-02 -6.74485862e-01 1.92682847e-01 -1.20228219e+00
-3.92159484e-02 -9.29055929e-01 8.67553949e-01 6.80156171e-01
1.00495689e-01 6.88188151e-02 3.91168028e-01 5.71514368e-01
9.15838331e-02 -9.18096676e-02 6.93689287e-01 -2.14944720e-01
-8.27010930e-01 4.47645098e-01 -8.51433396e-01 -1.72957838e-01
1.01566052e+00 -5.21153696e-02 -4.97747660e-01 1.00666359e-02
-7.16534376e-01 2.27584243e-01 7.98107266e-01 1.15441836e-01
5.81368923e-01 -1.53866565e+00 -1.10773146e+00 -2.06283018e-01
2.04373777e-01 1.70408219e-01 6.56191587e-01 3.55344713e-01
-9.15347278e-01 1.43636484e-02 -5.46451449e-01 -1.40635222e-01
-1.16954601e+00 4.44816470e-01 7.43914023e-02 -1.74744874e-01
-6.41674459e-01 5.22558868e-01 -3.54967296e-01 -4.61487353e-01
-1.89111561e-01 -2.91466296e-01 -1.07492164e-01 5.81516027e-01
9.38295424e-01 9.41342294e-01 -3.01794335e-02 -1.11967564e+00
-6.78348899e-01 4.63404596e-01 6.85190797e-01 -2.32248843e-01
1.86407721e+00 -6.23517752e-01 -3.36585462e-01 1.55576974e-01
6.14708245e-01 -4.39914972e-01 -1.31287718e+00 -1.63936555e-01
6.13224097e-02 -4.46533173e-01 6.19556963e-01 -9.34517920e-01
-6.83822691e-01 6.79286778e-01 8.42795849e-01 2.47423202e-01
1.33865798e+00 3.23457015e-03 8.54116857e-01 3.74762386e-01
7.20548928e-01 -1.27134860e+00 -1.17771760e-01 2.75063604e-01
1.71899867e+00 -1.20026660e+00 5.87069631e-01 1.64405584e-01
-3.00466985e-01 7.87959158e-01 1.86156556e-01 2.95288056e-01
6.74281895e-01 2.49486919e-02 -2.34177142e-01 -3.96983474e-01
1.06302850e-01 -5.13025880e-01 -2.50810295e-01 9.71962512e-01
-2.27012187e-01 6.08631074e-01 1.97384253e-01 2.78256774e-01
-3.58358510e-02 -2.86839783e-01 2.79140949e-01 1.08462310e+00
-1.13036311e+00 -3.83263379e-01 -1.21531653e+00 4.53384608e-01
9.69539359e-02 -4.64494705e-01 -3.76706213e-01 5.97178698e-01
4.10146773e-01 1.19439793e+00 2.95751896e-02 -1.92696288e-01
4.96164411e-01 -3.15489501e-01 -9.73929539e-02 -6.55811310e-01
-5.37373602e-01 -4.30929571e-01 8.86645988e-02 -2.14415461e-01
-5.49723983e-01 -1.22512734e+00 -1.34000897e+00 -7.35953689e-01
9.98503342e-02 -6.24940038e-01 6.33151114e-01 6.74037576e-01
2.87152618e-01 -1.62825644e-01 7.10077643e-01 -1.14389968e+00
-2.86538340e-02 -1.07291853e+00 -7.36793995e-01 2.30877593e-01
3.99108887e-01 -5.58340609e-01 -1.97098225e-01 2.29706123e-01] | [9.52580451965332, -1.2854713201522827] |
495cacc0-9004-4905-b95b-29e99de5d427 | maskreid-a-mask-based-deep-ranking-neural | 1804.03864 | null | http://arxiv.org/abs/1804.03864v2 | http://arxiv.org/pdf/1804.03864v2.pdf | MaskReID: A Mask Based Deep Ranking Neural Network for Person Re-identification | Person retrieval faces many challenges including cluttered background,
appearance variations (e.g., illumination, pose, occlusion) among different
camera views and the similarity among different person's images. To address
these issues, we put forward a novel mask based deep ranking neural network
with a skipped fusing layer. Firstly, to alleviate the problem of cluttered
background, masked images with only the foreground regions are incorporated as
input in the proposed neural network. Secondly, to reduce the impact of the
appearance variations, the multi-layer fusion scheme is developed to obtain
more discriminative fine-grained information. Lastly, considering person
retrieval is a special image retrieval task, we propose a novel ranking loss to
optimize the whole network. The proposed ranking loss can further mitigate the
interference problem of similar negative samples when producing ranking
results. The extensive experiments validate the superiority of the proposed
method compared with the state-of-the-art methods on many benchmark datasets. | ['Yang Gao', 'Lei Qi', 'Jing Huo', 'Yinghuan Shi', 'Lei Wang'] | 2018-04-11 | null | null | null | null | ['person-retrieval'] | ['computer-vision'] | [ 1.24278881e-01 -8.86232316e-01 3.77286196e-01 -4.87748563e-01
-4.07442689e-01 -2.20988303e-01 4.95360047e-01 -9.78747904e-02
-5.99078834e-01 6.19390368e-01 1.58752158e-01 4.42912459e-01
-2.09305793e-01 -6.11294150e-01 -4.77343112e-01 -9.16705549e-01
3.50821435e-01 -2.79881675e-02 3.02447319e-01 2.81439647e-02
1.13350295e-01 3.52545649e-01 -1.69645357e+00 2.25390628e-01
9.12159622e-01 9.68120515e-01 1.57441795e-01 -8.68890584e-02
1.01065569e-01 5.13456345e-01 -9.02111650e-01 -4.50357378e-01
3.47183555e-01 -8.58269259e-03 -5.73264211e-02 3.53782415e-01
6.27165556e-01 -7.28363216e-01 -5.72241366e-01 1.27441061e+00
7.92639911e-01 5.34682274e-01 4.18119848e-01 -9.49691772e-01
-6.90140247e-01 1.29283026e-01 -1.02060449e+00 3.06473315e-01
1.79379076e-01 9.05249864e-02 6.32022023e-01 -9.02918935e-01
2.96083927e-01 1.63280416e+00 1.46135092e-01 3.39471400e-01
-8.28274965e-01 -7.93024778e-01 6.35021806e-01 4.73647237e-01
-1.70690536e+00 -3.47016871e-01 9.22082305e-01 -2.95219094e-01
1.58728257e-01 2.30918720e-01 3.95832837e-01 9.11819935e-01
-9.62181315e-02 9.56928134e-01 1.02840102e+00 -1.33908153e-01
-2.50080496e-01 3.63071889e-01 4.08420205e-01 7.57252932e-01
5.11205971e-01 -5.97863942e-02 -1.13071017e-01 -1.50724113e-01
5.54391265e-01 5.27095079e-01 -4.77298290e-01 -4.21044439e-01
-1.04832721e+00 4.09771293e-01 5.80255210e-01 1.98599488e-01
-3.36844504e-01 3.48642655e-02 5.09580612e-01 -5.04273251e-02
2.48970389e-01 -2.19434202e-01 -2.06472054e-01 6.21639609e-01
-9.24335539e-01 3.65211934e-01 1.45932794e-01 6.42496884e-01
5.95703661e-01 5.16367704e-03 -6.41047955e-01 1.19575953e+00
5.01768827e-01 4.74975258e-01 4.69510376e-01 -4.76655334e-01
7.02563882e-01 6.34636045e-01 3.60506117e-01 -1.44503689e+00
-2.34955877e-01 -9.63776946e-01 -9.27385211e-01 -2.41562575e-01
2.25581273e-01 -4.16134819e-02 -9.08260524e-01 1.70645690e+00
3.55140835e-01 3.15269977e-02 -1.04038119e-01 1.50963843e+00
1.13105834e+00 7.28138983e-01 6.74318671e-02 -2.75445759e-01
1.43185782e+00 -1.07169032e+00 -7.07583487e-01 -1.99386373e-01
-1.11924924e-01 -8.18227410e-01 6.92514658e-01 2.76416153e-01
-9.84459162e-01 -9.19803202e-01 -1.14642155e+00 1.63957193e-01
-2.63700485e-01 5.43007195e-01 5.14022887e-01 5.00293076e-01
-6.76509798e-01 2.08470762e-01 -3.59201521e-01 -1.35262683e-01
3.92362416e-01 5.08763909e-01 -3.11457992e-01 -4.72123593e-01
-1.34020853e+00 5.66797018e-01 4.57140148e-01 6.33575261e-01
-5.62949002e-01 -2.14572981e-01 -4.90686268e-01 1.70080975e-01
4.27788228e-01 -8.16280961e-01 7.58688986e-01 -1.15178847e+00
-1.11380482e+00 5.81976712e-01 -2.11023986e-01 9.10985768e-02
6.37205482e-01 -4.68491495e-01 -5.93151748e-01 -7.44706206e-03
3.14100198e-02 4.35747415e-01 8.48042130e-01 -1.38671684e+00
-5.57957590e-01 -6.17797375e-01 8.84822160e-02 5.60226142e-01
-5.08492827e-01 2.23945454e-01 -1.10785758e+00 -6.53656781e-01
-1.10149160e-02 -7.94699550e-01 -1.23017952e-01 -2.70394776e-02
-2.67650753e-01 -1.59972951e-01 7.43459284e-01 -7.66834855e-01
1.22275043e+00 -2.11978912e+00 1.81122392e-01 3.92982155e-01
-6.63635135e-02 4.40369457e-01 -2.84502655e-01 -7.23015293e-02
8.53863582e-02 -2.19300166e-01 1.35529354e-01 -2.38327101e-01
-3.93786244e-02 -1.61237776e-01 1.82599276e-01 4.83889520e-01
7.40138888e-02 5.29961348e-01 -6.82539940e-01 -6.81273282e-01
4.15684789e-01 6.85545206e-01 -9.43885222e-02 1.70700833e-01
1.90339416e-01 4.25769717e-01 -6.17386878e-01 6.53821528e-01
1.21506417e+00 -1.35696605e-01 -3.50956798e-01 -4.84005868e-01
1.15703724e-01 -3.71080011e-01 -1.52564764e+00 1.50598419e+00
-1.99011594e-01 1.24354802e-01 2.44897380e-01 -7.92576373e-01
9.57470059e-01 1.09978423e-01 2.47343391e-01 -7.95866311e-01
3.11008215e-01 7.47965351e-02 1.71918646e-01 -5.07303298e-01
4.00440574e-01 1.73756242e-01 8.71951059e-02 1.49828583e-01
-2.55816519e-01 8.58237147e-01 1.22751184e-01 3.24149840e-02
6.97117031e-01 9.91435535e-03 -6.40476942e-02 -1.17802180e-01
1.16383672e+00 -5.61529219e-01 8.59947145e-01 6.92079008e-01
-4.76200908e-01 6.24003708e-01 -5.87156788e-02 -6.41706109e-01
-6.31142080e-01 -9.41799462e-01 3.93096022e-02 1.05218410e+00
7.73972869e-01 2.54122578e-02 -5.99996805e-01 -5.17178297e-01
-1.20617719e-02 5.97417727e-02 -4.16033536e-01 -1.83189437e-01
-5.60406625e-01 -1.24673605e+00 3.13793540e-01 3.41787308e-01
1.03273237e+00 -9.91575360e-01 -2.65976131e-01 4.45162468e-02
-3.30411613e-01 -1.13780844e+00 -7.29138315e-01 -3.82231504e-01
-5.76062441e-01 -8.59229803e-01 -1.18664920e+00 -9.83015656e-01
8.01995635e-01 7.15566516e-01 8.26632559e-01 3.48039478e-01
-3.20885360e-01 1.00940958e-01 -1.80355772e-01 -1.88603759e-01
3.24618191e-01 -1.40225962e-01 2.02793572e-02 4.27473217e-01
4.46275800e-01 -1.70970276e-01 -1.07389987e+00 3.13777953e-01
-1.05965257e+00 -7.83867612e-02 7.16633677e-01 1.02315557e+00
5.34974992e-01 7.15156317e-01 4.08321917e-01 -5.60469031e-01
6.45190477e-01 -2.51314133e-01 -6.10131741e-01 5.56298673e-01
-2.06555232e-01 -2.17412964e-01 5.87710440e-01 -6.44114435e-01
-1.41676831e+00 -7.81694241e-03 2.77488679e-01 -4.25051689e-01
-1.83921874e-01 1.93974584e-01 -6.24594331e-01 5.41313458e-03
8.29910412e-02 2.49388382e-01 -2.62720108e-01 -3.80160034e-01
5.28458841e-02 5.51285207e-01 4.58163023e-01 -5.96654058e-01
9.76662993e-01 3.86334002e-01 -1.29972011e-01 -4.20582175e-01
-7.83283770e-01 -5.01803696e-01 -3.36403012e-01 -2.47191206e-01
7.50310659e-01 -1.24160874e+00 -5.21438062e-01 7.85766602e-01
-1.37428844e+00 4.66531456e-01 5.34037054e-01 3.31308961e-01
2.73445785e-01 5.92377901e-01 -6.43335700e-01 -9.98674512e-01
-7.05971897e-01 -1.22943306e+00 9.84473944e-01 8.54901135e-01
3.06025416e-01 -4.63852972e-01 -4.50036615e-01 6.10698521e-01
2.59347081e-01 1.65888637e-01 7.14275301e-01 -5.61596572e-01
-7.48299062e-01 -3.48943889e-01 -7.40189731e-01 3.89732510e-01
3.23991150e-01 -1.13699540e-01 -9.13899660e-01 -5.13763309e-01
-5.21521010e-02 -8.93192962e-02 9.93001461e-01 2.20930517e-01
1.25009680e+00 -2.77734160e-01 -3.69995028e-01 4.47177857e-01
1.40229344e+00 2.85166234e-01 4.26128477e-01 4.39808905e-01
9.37304556e-01 8.28531206e-01 8.10461998e-01 3.82805407e-01
1.23704068e-01 7.39401162e-01 2.07997367e-01 -3.66739303e-01
7.51798749e-02 1.34453923e-01 7.62324110e-02 4.59168315e-01
6.75324118e-03 -3.86602372e-01 -3.17788303e-01 3.89885902e-01
-1.99290204e+00 -9.48610008e-01 3.64335328e-02 2.33198762e+00
2.87189484e-01 9.57098082e-02 -2.84934044e-02 -1.01144731e-01
1.29065788e+00 3.09189111e-01 -5.68170547e-01 2.20030740e-01
-3.86419266e-01 -5.46859391e-02 2.46006578e-01 2.50915382e-02
-1.43415093e+00 7.03406215e-01 4.65953827e+00 1.06657743e+00
-1.03200006e+00 -2.49444786e-02 7.05498874e-01 -2.90930182e-01
-1.85662836e-01 -3.54230285e-01 -8.44932616e-01 7.89844871e-01
3.14467698e-02 2.15082750e-01 4.26307678e-01 3.82775128e-01
2.71028191e-01 3.11982515e-03 -6.53677523e-01 1.27807117e+00
3.98952037e-01 -5.86649477e-01 2.99705207e-01 -1.47125825e-01
6.39282286e-01 -3.89469832e-01 2.53185272e-01 2.47248441e-01
-1.45367011e-01 -7.34625399e-01 5.31009138e-01 6.98217034e-01
2.57896751e-01 -1.10924447e+00 1.12113452e+00 4.67376970e-02
-1.26352870e+00 -2.08568051e-01 -5.70453107e-01 1.86183885e-01
6.94165751e-02 7.35949039e-01 -7.03955218e-02 7.06646860e-01
8.21310878e-01 5.26261568e-01 -8.06415617e-01 1.21340346e+00
7.78763592e-02 -7.08582774e-02 -2.74284244e-01 1.30449623e-01
1.44728124e-01 -2.44661435e-01 5.58277905e-01 1.14739394e+00
1.24610171e-01 7.26525858e-02 4.07129914e-01 6.92587435e-01
-6.15265481e-02 3.06304485e-01 -2.18863115e-01 3.35933149e-01
3.95591825e-01 1.44657302e+00 -5.03591001e-01 -3.41784507e-01
-6.24188960e-01 1.19859731e+00 4.01277281e-02 6.84365213e-01
-7.85406947e-01 -2.72982538e-01 4.03536111e-01 9.61041730e-03
3.13055307e-01 3.06089759e-01 2.98064739e-01 -1.34628928e+00
4.39912736e-01 -1.02082002e+00 4.19213295e-01 -6.62274063e-01
-1.44666457e+00 4.24504340e-01 -7.49488324e-02 -1.27240789e+00
2.89710075e-01 -3.38944763e-01 -6.06254995e-01 1.00235486e+00
-1.64423037e+00 -1.19360113e+00 -5.75826645e-01 6.82112753e-01
4.53665644e-01 -3.25532794e-01 3.17143798e-01 9.23061013e-01
-1.02553582e+00 7.69192040e-01 9.78939533e-02 2.73518980e-01
1.08501053e+00 -8.22720349e-01 -1.95326731e-01 9.56888914e-01
-4.46956962e-01 1.08401179e+00 3.54027599e-01 -4.75760460e-01
-1.13179028e+00 -1.18350875e+00 3.64887208e-01 1.42509401e-01
1.66020930e-01 -1.10197276e-01 -9.24412608e-01 1.12465002e-01
-3.82091105e-02 3.74815352e-02 5.41807294e-01 -2.74435654e-02
-3.31369758e-01 -6.49824619e-01 -1.10064149e+00 5.77074170e-01
6.79737508e-01 -1.93522751e-01 -4.35220420e-01 8.89676139e-02
5.37213326e-01 -3.55383545e-01 -5.95206141e-01 7.14739442e-01
9.36783135e-01 -9.94444132e-01 1.25744963e+00 -2.66207289e-02
1.06516168e-01 -7.55680442e-01 -3.13087367e-02 -9.09882724e-01
-5.79964638e-01 -1.11477293e-01 1.08332843e-01 1.62443876e+00
-1.86426744e-01 -5.19864142e-01 6.15344882e-01 5.88726342e-01
2.33297214e-01 -5.65603614e-01 -5.63384295e-01 -5.16359806e-01
-3.28442097e-01 5.08386254e-01 6.68737650e-01 5.86673319e-01
-7.04634488e-01 4.44996357e-01 -6.44589365e-01 2.86163360e-01
8.08975637e-01 1.72743708e-01 7.25830019e-01 -1.37125516e+00
-3.36481541e-01 -1.99529842e-01 -3.84314686e-01 -9.05170500e-01
4.38997671e-02 -4.44424987e-01 1.32008642e-01 -1.45037234e+00
7.83556402e-01 -2.79177338e-01 -9.54635322e-01 1.11721642e-02
-7.90822268e-01 2.68029898e-01 3.90519589e-01 2.91623414e-01
-9.23767030e-01 6.12141192e-01 1.30948997e+00 -4.74127918e-01
-1.78168744e-01 -4.16845977e-02 -6.88918829e-01 6.51707113e-01
5.95132530e-01 -2.50164479e-01 -4.69540715e-01 -7.45655239e-01
-2.07934707e-01 -4.43355255e-02 5.61502576e-01 -1.13346672e+00
2.62189835e-01 7.03636333e-02 1.08989847e+00 -8.81771028e-01
4.76763666e-01 -1.02689278e+00 1.18226148e-01 4.05059099e-01
-1.58384547e-01 7.43408725e-02 2.00239286e-01 7.14419067e-01
-3.99527431e-01 -1.18099235e-01 7.69186080e-01 -2.90428430e-01
-6.52829766e-01 5.09144843e-01 6.78508878e-02 -2.44872198e-01
7.41885364e-01 -1.56624287e-01 -5.02343655e-01 -2.63374418e-01
-2.31172949e-01 5.63018620e-01 3.92958552e-01 5.60891509e-01
6.40526891e-01 -1.36760271e+00 -8.22844803e-01 4.12493525e-03
1.98179632e-01 -8.61011073e-02 8.15123081e-01 6.14032030e-01
-2.33359084e-01 3.05576622e-01 -1.22248255e-01 -4.49047118e-01
-1.55425394e+00 6.42315507e-01 3.69178981e-01 -4.84429568e-01
-3.04629326e-01 7.20913708e-01 6.55250251e-01 -2.44109631e-01
3.76632124e-01 2.85156697e-01 -4.30837214e-01 2.08572485e-02
6.80827796e-01 3.07462126e-01 -6.90079406e-02 -7.74915099e-01
-5.30175686e-01 7.87203312e-01 -5.47209382e-01 1.59689277e-01
1.01026881e+00 -2.86858559e-01 -1.88059762e-01 7.97778443e-02
1.17406213e+00 -2.06927117e-02 -1.00736177e+00 -3.96839947e-01
-4.59419101e-01 -5.21756351e-01 8.03222880e-02 -7.87878394e-01
-1.27177429e+00 8.32196355e-01 1.10987866e+00 -3.70755494e-01
1.23950505e+00 -5.69419205e-01 9.69327271e-01 3.62515897e-01
2.22495392e-01 -1.11708617e+00 1.00854732e-01 2.13660598e-01
6.30696654e-01 -1.38118672e+00 4.17011797e-01 -3.36520225e-01
-3.83044034e-01 7.86985517e-01 9.20278132e-01 -2.53477037e-01
3.29798430e-01 -4.21687037e-01 1.33179471e-01 6.17202893e-02
-3.47751737e-01 -2.99132109e-01 5.96322000e-01 3.33171517e-01
3.04594994e-01 -2.17698529e-01 -5.57608902e-01 6.84954941e-01
3.32361400e-01 -4.40334007e-02 1.68874264e-02 6.15939081e-01
-2.80033648e-01 -9.19736505e-01 -7.52728105e-01 4.27436173e-01
-6.11000955e-01 -1.03150457e-01 -2.68807143e-01 6.55907094e-01
4.99828219e-01 1.12405860e+00 -2.96435356e-02 -3.35332960e-01
3.04251343e-01 -1.15507737e-01 4.45477605e-01 -3.49592119e-01
-6.54992342e-01 4.31924641e-01 -2.00643361e-01 -2.30378643e-01
-6.40181184e-01 -5.55467069e-01 -8.40010285e-01 -5.42753004e-02
-5.87894022e-01 4.94494289e-02 2.08787099e-01 9.03512836e-01
1.26775473e-01 7.11688101e-01 6.66471124e-01 -6.64815366e-01
-4.49274898e-01 -9.14139986e-01 -4.85449553e-01 7.17171609e-01
2.35429958e-01 -8.10865879e-01 -3.01069051e-01 -1.77454099e-01] | [14.7142915725708, 0.8891558647155762] |
74af0af5-41ba-4352-ab0b-38afbdaa7531 | metal-artifact-correction-in-cone-beam | 2208.08288 | null | https://arxiv.org/abs/2208.08288v1 | https://arxiv.org/pdf/2208.08288v1.pdf | Metal artifact correction in cone beam computed tomography using synthetic X-ray data | Metal artifact correction is a challenging problem in cone beam computed tomography (CBCT) scanning. Metal implants inserted into the anatomy cause severe artifacts in reconstructed images. Widely used inpainting-based metal artifact reduction (MAR) methods require segmentation of metal traces in the projections as a first step which is a challenging task. One approach is to use a deep learning method to segment metals in the projections. However, the success of deep learning methods is limited by the availability of realistic training data. It is challenging and time consuming to get reliable ground truth annotations due to unclear implant boundary and large number of projections. We propose to use X-ray simulations to generate synthetic metal segmentation training dataset from clinical CBCT scans. We compare the effect of simulations with different number of photons and also compare several training strategies to augment the available data. We compare our model's performance on real clinical scans with conventional threshold-based MAR and a recent deep learning method. We show that simulations with relatively small number of photons are suitable for the metal segmentation task and that training the deep learning model with full size and cropped projections together improves the robustness of the model. We show substantial improvement in the image quality affected by severe motion, voxel size under-sampling, and out-of-FOV metals. Our method can be easily implemented into the existing projection-based MAR pipeline to get improved image quality. This method can provide a novel paradigm to accurately segment metals in CBCT projections. | ['Simo Särkkä', 'Ari Hietanen', 'Harshit Agrawal'] | 2022-08-17 | null | null | null | null | ['metal-artifact-reduction'] | ['medical'] | [ 2.74344236e-01 1.35298893e-01 5.52542269e-01 -2.28580683e-01
-1.13513255e+00 -7.64394179e-02 1.53678000e-01 4.77733202e-02
-3.90113354e-01 5.44532835e-01 3.83209512e-02 -2.11331427e-01
1.04027679e-02 -7.07308412e-01 -8.94843340e-01 -5.40354013e-01
1.48692176e-01 9.75785255e-01 7.89288759e-01 -6.29817788e-03
3.30094807e-02 8.04730237e-01 -7.67351866e-01 6.81801260e-01
4.66792107e-01 8.09955359e-01 4.35974747e-01 3.39121401e-01
1.79178938e-02 5.34021497e-01 -2.86858737e-01 -1.35677487e-01
5.84547877e-01 -5.98820627e-01 -7.10214615e-01 1.05192691e-01
2.62515664e-01 -3.95093769e-01 -1.39821634e-01 6.71910405e-01
7.96251893e-01 -3.07144731e-01 6.01310194e-01 -6.49151683e-01
3.92985232e-02 5.57675421e-01 -6.41943634e-01 1.48131609e-01
6.12737313e-02 2.29851872e-01 -8.73527303e-02 -9.65877175e-01
6.99774265e-01 5.76662600e-01 1.24044180e+00 7.09398270e-01
-1.23966467e+00 -6.68723822e-01 -5.18101275e-01 5.02808616e-02
-9.60484982e-01 -6.87939152e-02 7.71159887e-01 -7.36424685e-01
6.43319130e-01 4.78001058e-01 8.79590273e-01 7.73180604e-01
5.70993304e-01 2.83634722e-01 1.30729616e+00 -5.53207636e-01
3.47408891e-01 -9.74934269e-03 -7.67971948e-02 6.80004656e-01
3.18890572e-01 2.23245278e-01 2.59398110e-02 -3.05175185e-01
1.19597292e+00 3.63208711e-01 -6.39231205e-01 -4.87401783e-01
-1.07408214e+00 6.51567400e-01 6.70359850e-01 3.56044769e-01
-6.74222410e-01 4.45295423e-01 3.62576991e-01 -7.39628375e-02
3.31009477e-01 5.91254234e-01 -2.38021255e-01 2.11157694e-01
-1.24159658e+00 2.92352915e-01 3.27319562e-01 5.50328374e-01
2.79809922e-01 -1.37630487e-02 -2.77914643e-01 9.10973549e-01
3.56427640e-01 2.94569731e-01 7.28841186e-01 -7.96627641e-01
2.48966157e-03 3.37227166e-01 -5.45432456e-02 -3.32315415e-01
-5.53693533e-01 -2.91241765e-01 -5.92551768e-01 7.56206751e-01
5.49148738e-01 -1.32937878e-01 -1.71476352e+00 9.90135312e-01
3.86478901e-01 1.61562711e-01 -5.36413610e-01 1.12151563e+00
6.48329794e-01 3.47043037e-01 -6.86190724e-02 -3.68967980e-01
1.32846761e+00 -7.24357188e-01 -6.30961359e-01 -2.32056871e-01
6.12045169e-01 -1.09068060e+00 1.12994385e+00 6.10512078e-01
-1.49040008e+00 -2.46351492e-02 -1.14582324e+00 1.94690973e-01
3.04806173e-01 -8.43760967e-02 5.57229042e-01 7.01310754e-01
-7.08884001e-01 1.00043166e+00 -1.35487878e+00 9.19136778e-02
9.33775783e-01 8.08427811e-01 -1.81659013e-01 -4.00396258e-01
-6.42773628e-01 9.98282671e-01 -2.85618275e-01 4.73472960e-02
-9.05208945e-01 -1.11762834e+00 -4.80670244e-01 -4.15949762e-01
2.10758984e-01 -7.59608209e-01 1.59819746e+00 -7.89514184e-01
-1.53592360e+00 7.77885675e-01 2.00295642e-01 -3.87819618e-01
1.05133390e+00 -1.82608455e-01 7.90965855e-02 2.36734226e-01
-1.00785540e-03 2.83244818e-01 7.30621219e-01 -1.46911955e+00
1.92768648e-01 -4.36982572e-01 -6.27913833e-01 -1.25059173e-01
3.41443509e-01 1.50113136e-01 -4.36284155e-01 -7.22351491e-01
7.14969456e-01 -1.16491365e+00 -8.14801693e-01 4.06549811e-01
-1.92408621e-01 5.80916584e-01 5.65169394e-01 -9.55860496e-01
6.49453819e-01 -1.66697621e+00 -4.56080079e-01 2.97581375e-01
8.69814828e-02 4.60131168e-02 2.84490734e-01 5.28794974e-02
-3.58708322e-01 -1.93154797e-01 -7.28711784e-01 -2.82885045e-01
-7.27751851e-01 3.99299264e-02 2.17812061e-01 6.52717173e-01
-2.94956207e-01 8.18771660e-01 -5.62508225e-01 -4.63032752e-01
3.28470349e-01 6.00747526e-01 -5.66904068e-01 7.93855712e-02
-1.51797131e-01 9.04242814e-01 -1.75901592e-01 3.50887150e-01
1.06636798e+00 -1.27922863e-01 3.48245539e-02 -4.05880809e-01
1.35019407e-01 2.91034251e-01 -1.03771663e+00 1.89679646e+00
-7.73725033e-01 3.65633935e-01 -7.33495166e-04 -5.73404610e-01
4.05668795e-01 5.87747991e-01 1.00080156e+00 -7.32817173e-01
2.63130426e-01 5.99160314e-01 3.17255527e-01 -5.30079961e-01
-9.00566876e-02 -1.01418674e+00 4.75024343e-01 6.46005809e-01
-3.79982889e-01 -8.56154382e-01 -6.19200408e-01 -1.07259423e-01
1.19787979e+00 2.50392482e-02 -1.77517816e-01 -2.22913086e-01
-4.07061279e-02 3.03060442e-01 3.74384999e-01 3.11556816e-01
7.63369948e-02 1.41606832e+00 -2.33863555e-02 -7.00654566e-01
-1.38227439e+00 -1.08716214e+00 -5.12041032e-01 -9.29918699e-03
-7.05320984e-02 1.27499253e-01 -7.85070777e-01 -5.59782147e-01
-4.31462079e-01 5.80376744e-01 -5.79164147e-01 -7.14874268e-02
-1.08811235e+00 -1.20711589e+00 1.61347196e-01 6.81285203e-01
1.07747681e-01 -1.19544125e+00 -9.55489397e-01 4.49740589e-01
-1.38340563e-01 -1.01578450e+00 -1.39593363e-01 4.80992824e-01
-1.39963710e+00 -1.01715636e+00 -1.08435392e+00 -7.31320381e-01
9.06785846e-01 -1.31698340e-01 1.14609957e+00 2.86082447e-01
-7.21032321e-01 1.01588547e-01 -1.10259213e-01 -5.18335342e-01
-8.62899780e-01 -5.68246007e-01 -2.43724242e-01 -6.32876873e-01
-2.34184787e-01 -5.71004868e-01 -1.00077152e+00 3.19553256e-01
-1.05259347e+00 3.04097801e-01 5.52489519e-01 7.86273062e-01
8.52812231e-01 -8.66399184e-02 1.41718805e-01 -1.12297010e+00
3.82005692e-01 -6.86849505e-02 -5.34905434e-01 -6.71563074e-02
-4.60401148e-01 1.63301155e-01 2.95512587e-01 -4.83558238e-01
-1.09782481e+00 3.77011925e-01 -6.03947401e-01 -3.42392653e-01
-7.81000883e-04 1.22439556e-01 2.23258913e-01 -5.42276382e-01
9.36055958e-01 -1.13501407e-01 1.33814365e-01 -3.84220779e-01
-1.61462754e-01 3.47101003e-01 2.59519070e-01 -3.50458831e-01
4.79737014e-01 8.93683314e-01 2.37156644e-01 -5.87714612e-01
-3.50116372e-01 -2.59408206e-01 -6.90407634e-01 -2.20582888e-01
7.09206402e-01 -6.34519935e-01 -2.47555926e-01 3.41788024e-01
-1.22702944e+00 -6.88406646e-01 -5.86297989e-01 7.79836297e-01
-5.30188382e-01 5.96437752e-01 -1.02568281e+00 -3.72903734e-01
-6.94588959e-01 -1.71749437e+00 1.02914655e+00 -2.21940652e-01
-1.72340676e-01 -6.11918271e-01 1.98161900e-01 4.02004868e-01
4.05044138e-01 4.23927993e-01 9.09181356e-01 -1.86101869e-01
-6.03622139e-01 -2.29054824e-01 2.53741201e-02 3.82369220e-01
1.07822731e-01 -4.39103633e-01 -8.56254399e-01 1.71159171e-02
6.03601754e-01 -5.72726615e-02 7.50066459e-01 9.45396721e-01
1.38659716e+00 1.68985054e-01 -4.76445645e-01 6.14338100e-01
1.63636875e+00 3.27750415e-01 9.42060292e-01 3.33509207e-01
7.15115845e-01 2.71785617e-01 4.08904701e-01 2.74472237e-01
-2.79977292e-01 8.93420160e-01 3.98059964e-01 -2.66239911e-01
-5.43729186e-01 2.47328505e-01 -1.80208698e-01 6.34195328e-01
-2.16835573e-01 3.82791936e-01 -1.19761956e+00 4.73173380e-01
-1.38090456e+00 -5.60592234e-01 -7.96793938e-01 2.46972942e+00
9.26024556e-01 5.83851039e-02 -5.86748719e-02 2.83379734e-01
2.97711015e-01 -8.09458613e-01 -2.60390222e-01 -1.72938764e-01
3.94710213e-01 9.26771283e-01 7.19855011e-01 5.93429863e-01
-6.74123466e-01 3.63185525e-01 6.47534323e+00 6.11986041e-01
-1.37472761e+00 7.75292873e-01 5.23987532e-01 -2.00645313e-01
-2.57612675e-01 -1.71578288e-01 -2.43471846e-01 5.47598422e-01
7.22471058e-01 5.32438576e-01 -5.67517877e-02 5.32236934e-01
3.90269965e-01 -5.97465813e-01 -1.02270293e+00 1.02871156e+00
-6.59873113e-02 -1.47935021e+00 -2.38778129e-01 1.43315747e-01
7.65521228e-01 2.06021354e-01 -2.17999130e-01 -1.29858643e-01
6.02583289e-02 -1.09369397e+00 6.12382054e-01 2.61962950e-01
7.40030468e-01 -4.12363917e-01 7.12380707e-01 1.43304273e-01
-4.86415088e-01 2.78578758e-01 -4.89459425e-01 3.52377146e-01
4.94637281e-01 1.01126146e+00 -1.23583555e+00 4.13215846e-01
7.76013494e-01 -6.23980165e-02 -5.04474580e-01 1.67475939e+00
1.06757537e-01 4.61755991e-01 -6.65846825e-01 4.69555438e-01
-1.09174624e-01 -9.14743021e-02 2.34409958e-01 9.88138378e-01
4.44355488e-01 2.55179375e-01 -9.72938910e-02 7.56444752e-01
-3.84228118e-02 9.55325067e-02 -3.64405483e-01 7.82961309e-01
-1.37333438e-01 1.03422523e+00 -1.13968313e+00 -3.34588200e-01
-2.38509342e-01 1.13126254e+00 -1.49541825e-01 -3.74220721e-02
-9.17223454e-01 3.95903170e-01 4.14542481e-02 1.01792514e+00
-1.77757256e-02 -1.75436884e-01 -8.68714213e-01 -5.81465304e-01
1.00921884e-01 -5.84950149e-01 1.81997165e-01 -9.05976176e-01
-1.10371196e+00 8.82060945e-01 3.22878803e-03 -1.28594911e+00
1.46302003e-02 -4.76322651e-01 -7.22754955e-01 7.25994349e-01
-1.20279717e+00 -1.04348862e+00 -5.09257615e-01 4.64741379e-01
6.08244717e-01 4.34749722e-01 6.19301915e-01 5.64488888e-01
4.00200300e-02 3.02513868e-01 -8.73624813e-03 -5.85888103e-02
6.19057655e-01 -1.04127455e+00 1.31166965e-01 5.70621967e-01
-3.32416408e-02 9.09369886e-02 7.68305957e-01 -9.95553017e-01
-1.06109810e+00 -9.87121999e-01 4.70841601e-02 -5.22156954e-01
1.91089839e-01 -1.69951141e-01 -9.44263637e-01 7.68648744e-01
7.66566843e-02 4.84229088e-01 6.49663746e-01 -5.49833953e-01
3.27775925e-01 2.07548514e-01 -1.50881660e+00 3.80333632e-01
5.29028893e-01 7.18094558e-02 -5.35008788e-01 6.49570286e-01
4.05037463e-01 -9.11949873e-01 -7.69407928e-01 7.17768967e-01
5.19350946e-01 -9.58918631e-01 1.05825555e+00 -2.06509888e-01
4.95865494e-01 -2.08238419e-02 2.82061189e-01 -1.35268736e+00
-9.93234962e-02 -1.23826601e-01 5.00433624e-01 3.80848080e-01
4.65428919e-01 -2.81134456e-01 1.33449435e+00 8.58715951e-01
-6.89674258e-01 -8.16945910e-01 -1.08083999e+00 -5.90953887e-01
4.30946678e-01 -7.68046558e-01 5.62998243e-02 8.84865046e-01
-3.04210603e-01 -2.70473897e-01 1.38631845e-02 2.57220894e-01
6.49493098e-01 -3.75631750e-02 3.07363003e-01 -1.20764935e+00
-4.83773947e-01 2.89201830e-02 -2.12887019e-01 -4.06891108e-01
-6.06441557e-01 -8.42758596e-01 3.13174397e-01 -1.92388332e+00
2.26629958e-01 -8.56878519e-01 3.27965915e-02 3.43205661e-01
7.32613653e-02 7.51727283e-01 -1.96487382e-02 3.28600526e-01
3.17705482e-01 2.28785992e-01 1.71422195e+00 2.83857565e-02
-3.94958973e-01 2.47186556e-01 6.66168481e-02 9.78416979e-01
6.90462470e-01 -8.98971438e-01 -7.38606155e-02 -4.23331350e-01
1.94358855e-01 1.73611209e-01 5.19756079e-01 -1.38413727e+00
5.19746765e-02 2.39464641e-01 7.62850404e-01 -5.66677213e-01
4.54573542e-01 -1.20899808e+00 6.38132095e-01 1.00888908e+00
1.19261131e-01 -3.43099311e-02 4.96119350e-01 1.91110879e-01
1.46021515e-01 -4.74903226e-01 1.33570099e+00 -7.55745590e-01
8.35727341e-03 2.01355666e-01 -4.16885316e-01 -3.15105945e-01
9.58695412e-01 -3.86017203e-01 4.52517837e-01 -9.36013833e-02
-1.15155292e+00 -5.17228901e-01 6.13278925e-01 -1.63881436e-01
8.81871879e-01 -1.19926929e+00 -5.74346304e-01 9.59777534e-02
-2.56795973e-01 3.19165319e-01 3.99488747e-01 1.30377519e+00
-1.35211051e+00 -2.76795942e-02 -2.24538967e-01 -7.86778152e-01
-1.26042581e+00 4.58615035e-01 9.10265803e-01 -5.03386378e-01
-1.06412101e+00 8.65480661e-01 2.92387873e-01 -2.33898371e-01
-2.22382009e-01 -7.21652865e-01 4.70088840e-01 -5.73542476e-01
2.07320541e-01 1.04788274e-01 9.25108612e-01 -2.26905242e-01
-3.30845773e-01 7.42557704e-01 -2.05996007e-01 -2.22910583e-01
1.52892089e+00 2.71624297e-01 1.61050051e-01 8.22781473e-02
8.07551622e-01 1.20231867e-01 -1.12611091e+00 2.80795749e-02
-2.08216175e-01 -3.92805904e-01 3.40062529e-01 -1.00969648e+00
-1.38839722e+00 9.81727898e-01 1.27136695e+00 -5.12779891e-01
9.32346523e-01 4.31584381e-02 1.09943795e+00 -3.07966709e-01
5.42183161e-01 -8.17235947e-01 3.22416276e-01 -2.62341410e-01
1.00304675e+00 -1.18617094e+00 4.17828351e-01 -9.45453703e-01
-5.00280619e-01 1.14501655e+00 5.18736720e-01 -3.46952111e-01
8.56909633e-01 7.73548782e-01 2.69620031e-01 -6.07944846e-01
-4.73619322e-04 2.94546038e-01 1.57427594e-01 5.73985696e-01
5.89299858e-01 -8.59626085e-02 -5.46220601e-01 4.49344277e-01
-4.35005538e-02 1.91118553e-01 9.06716526e-01 1.17138791e+00
-2.36351430e-01 -1.26942146e+00 -8.73648107e-01 5.73389471e-01
-8.53657663e-01 -1.73450291e-01 -5.58529003e-03 6.43188715e-01
5.68177029e-02 4.13901925e-01 -1.93545163e-01 1.92854822e-01
4.56800073e-01 -6.21962994e-02 9.97134686e-01 -9.54587758e-01
-9.48073089e-01 4.07533884e-01 -3.08583498e-01 -5.62169731e-01
-1.73197910e-01 -4.46433872e-01 -1.56459844e+00 2.56696939e-01
-4.44841772e-01 -2.32366487e-01 1.04613781e+00 8.30525100e-01
-3.17101814e-02 7.73045599e-01 2.29901716e-01 -1.05592036e+00
-2.26916581e-01 -9.73603368e-01 -4.81375188e-01 5.54242253e-01
1.50308823e-02 -6.73680127e-01 -1.59155801e-01 1.26164138e-01] | [13.508267402648926, -2.605604410171509] |
8d7065c2-808b-4314-a8c5-83c1df188034 | question-aware-memory-network-for-multi-hop | 2104.13173 | null | https://arxiv.org/abs/2104.13173v1 | https://arxiv.org/pdf/2104.13173v1.pdf | Question-Aware Memory Network for Multi-hop Question Answering in Human-Robot Interaction | Knowledge graph question answering is an important technology in intelligent human-robot interaction, which aims at automatically giving answer to human natural language question with the given knowledge graph. For the multi-relation question with higher variety and complexity, the tokens of the question have different priority for the triples selection in the reasoning steps. Most existing models take the question as a whole and ignore the priority information in it. To solve this problem, we propose question-aware memory network for multi-hop question answering, named QA2MN, to update the attention on question timely in the reasoning process. In addition, we incorporate graph context information into knowledge graph embedding model to increase the ability to represent entities and relations. We use it to initialize the QA2MN model and fine-tune it in the training process. We evaluate QA2MN on PathQuestion and WorldCup2014, two representative datasets for complex multi-hop question answering. The result demonstrates that QA2MN achieves state-of-the-art Hits@1 accuracy on the two datasets, which validates the effectiveness of our model. | ['Quanjun Yin', 'Keping Yu', 'Qian Li', 'Mamoun Alazab', 'Xinmeng Li'] | 2021-04-27 | null | null | null | null | ['graph-question-answering', 'multi-hop-question-answering'] | ['graphs', 'knowledge-base'] | [-1.49970412e-01 5.80096006e-01 -2.28420556e-01 -2.86962092e-01
-4.29866940e-01 -4.24205661e-01 3.33974093e-01 4.18278456e-01
-5.33863485e-01 5.99081397e-01 5.11953235e-01 -5.61141670e-01
-3.37096781e-01 -1.16128159e+00 -7.85236239e-01 4.62984741e-02
3.35141689e-01 9.56045210e-01 1.08138371e+00 -6.01044595e-01
4.35507018e-03 -1.74638018e-01 -1.12765777e+00 4.27309275e-01
1.06865728e+00 8.54582369e-01 2.03555986e-01 5.99160373e-01
-5.40067554e-01 1.60134697e+00 -5.19271135e-01 -8.33335161e-01
-3.98787856e-01 -4.05394465e-01 -1.83459496e+00 -5.77994585e-01
3.47562969e-01 -3.45248014e-01 -8.05716515e-01 9.34758484e-01
2.36532196e-01 4.95201975e-01 4.42666352e-01 -1.27066672e+00
-1.23411381e+00 9.27472949e-01 -1.11339785e-01 3.88225228e-01
7.81985462e-01 -3.26594641e-03 1.39065957e+00 -6.80194795e-01
5.80895782e-01 1.57111883e+00 3.01180184e-01 5.35467267e-01
-4.61383909e-01 -1.56657130e-01 2.63349175e-01 1.28602469e+00
-1.20637810e+00 1.03460494e-02 7.17828870e-01 1.85655691e-02
1.33887470e+00 1.95414051e-01 3.55712265e-01 6.56900764e-01
-1.58057168e-01 8.98652256e-01 2.83981025e-01 -3.20254534e-01
-2.05852643e-01 -2.21783504e-01 8.33133340e-01 1.06732845e+00
2.70883828e-01 -4.57211524e-01 -4.54428256e-01 -1.66636184e-01
4.23078984e-01 -3.74793202e-01 -4.42508310e-01 -1.42622590e-01
-1.23007774e+00 8.03315759e-01 8.73953879e-01 1.75055951e-01
-4.38007325e-01 3.38389993e-01 4.78938133e-01 3.89755785e-01
-2.37684727e-01 6.66830540e-01 -7.29089618e-01 -2.54710559e-02
3.00956517e-01 1.86894253e-01 1.15520477e+00 1.05502498e+00
7.46896148e-01 -6.86135888e-01 -5.91194749e-01 6.70609415e-01
4.69515175e-01 2.05443218e-01 3.85658234e-01 -9.20918524e-01
8.11065972e-01 1.38046896e+00 2.35661641e-02 -1.12971854e+00
-5.41183829e-01 -2.16537691e-03 -3.98775905e-01 -6.96985364e-01
4.55296487e-01 -3.68848294e-02 -7.75254846e-01 1.53941584e+00
8.12488139e-01 1.91217825e-01 4.48492080e-01 9.50740039e-01
1.61888850e+00 6.95470393e-01 2.97953695e-01 2.26244465e-01
1.95992053e+00 -1.58171010e+00 -8.99271488e-01 -3.78253907e-01
9.41170931e-01 -2.62325525e-01 1.35481274e+00 -1.26348853e-01
-5.63057005e-01 -4.06449199e-01 -8.44838321e-01 -7.99728751e-01
-5.52703500e-01 -1.81977615e-01 6.43715382e-01 7.74473622e-02
-7.37145245e-01 1.77410111e-01 -4.45520520e-01 -4.98404890e-01
2.02360928e-01 7.48179331e-02 -1.83621034e-01 -5.82874656e-01
-1.96484125e+00 1.11650300e+00 7.56295264e-01 2.46778339e-01
-6.44932151e-01 -5.48792005e-01 -9.09861326e-01 2.85389334e-01
1.01492357e+00 -1.03793025e+00 1.37489188e+00 -1.31261989e-01
-1.31222296e+00 5.10533273e-01 -2.17216149e-01 -3.32380533e-01
-1.31600231e-01 -3.28164786e-01 -5.91068745e-01 4.42833304e-01
1.22345120e-01 5.11026859e-01 4.03568715e-01 -9.55754399e-01
-5.58008552e-01 -2.50355005e-01 8.85756791e-01 5.13691008e-01
-4.30862792e-02 -2.79336512e-01 -8.31345141e-01 4.45071906e-02
1.13099486e-01 -4.61411506e-01 1.65286660e-02 -4.28622037e-01
-4.00877863e-01 -1.06917036e+00 7.81698585e-01 -8.66491795e-01
1.40279794e+00 -1.74428308e+00 2.04435378e-01 -5.80528527e-02
3.84901106e-01 3.04304838e-01 -3.67651582e-01 5.04963517e-01
2.79326916e-01 -6.03671093e-03 4.14277054e-02 3.19226056e-01
1.47410557e-01 5.23224056e-01 -1.90327436e-01 -2.56700099e-01
3.09814006e-01 1.59065020e+00 -1.25167215e+00 -8.20900142e-01
-2.46517360e-01 8.31422880e-02 -4.62694287e-01 3.91267806e-01
-7.63733864e-01 1.56949162e-01 -8.90544713e-01 4.40400898e-01
2.62801468e-01 -8.86653543e-01 1.76449582e-01 -6.04837656e-01
7.71582782e-01 7.19015718e-01 -8.07055235e-01 1.74781072e+00
-2.78430223e-01 1.30535111e-01 -3.69517207e-01 -6.23676658e-01
5.87057710e-01 2.67851502e-01 -6.64631352e-02 -1.09337986e+00
4.16160077e-02 6.33119885e-03 2.15940759e-01 -1.09942198e+00
7.38738179e-01 2.79021561e-01 1.09195612e-01 2.68591940e-01
1.87878683e-01 -1.30579164e-02 2.80583829e-01 6.44114733e-01
1.37975311e+00 -1.03059076e-02 1.94664806e-01 -4.14451025e-02
8.58619452e-01 9.40814167e-02 4.85050529e-02 4.73886669e-01
-2.40143478e-01 5.30548580e-02 5.92430890e-01 -2.83290893e-01
-2.84272492e-01 -8.55346262e-01 5.27233005e-01 1.27144456e+00
6.29781246e-01 -4.18366164e-01 -6.16599858e-01 -1.17331088e+00
-3.92624550e-02 9.40486670e-01 -6.35580659e-01 -5.61632037e-01
-7.79899120e-01 -3.77240092e-01 5.77762246e-01 4.30844575e-01
8.58736873e-01 -1.40105951e+00 -3.51454586e-01 2.86172777e-01
-8.62547278e-01 -1.42809951e+00 -4.92513955e-01 -2.85584509e-01
-5.39670050e-01 -1.82602775e+00 -1.51843905e-01 -1.17979586e+00
5.71735203e-01 1.40781596e-01 1.54110885e+00 5.42666912e-01
1.73205420e-01 8.81777525e-01 -7.98146069e-01 9.04430822e-03
-7.00266287e-02 4.09617305e-01 -5.45381427e-01 -2.64306784e-01
6.79234147e-01 -2.58643329e-01 -7.28717029e-01 3.88017356e-01
-7.74744451e-01 -2.75062829e-01 2.61964679e-01 6.84734762e-01
4.36919421e-01 6.87149167e-02 9.19878185e-01 -8.03612113e-01
1.05258000e+00 -6.88637495e-01 -3.32082868e-01 8.43718350e-01
-5.09709299e-01 2.93982774e-01 5.04056156e-01 -2.31765956e-01
-1.03580916e+00 -4.54944909e-01 -1.73295721e-01 -2.60334238e-02
2.92404622e-01 9.58081722e-01 -4.57536608e-01 -9.03985288e-04
6.31824136e-01 8.20938498e-02 -3.70362788e-01 -2.24535167e-01
8.86745453e-01 2.53214896e-01 6.37516081e-01 -7.29809225e-01
5.74546635e-01 -2.40178704e-02 -5.10806106e-02 -3.78376782e-01
-1.20269263e+00 -5.85113645e-01 -2.08007798e-01 4.43655718e-03
1.10941386e+00 -5.93886852e-01 -1.34690380e+00 1.96738705e-01
-1.50683320e+00 -3.58137935e-01 -2.05782846e-01 2.25130230e-01
-2.60294318e-01 4.55899268e-01 -6.69055402e-01 -4.84685063e-01
-5.05755961e-01 -8.21601212e-01 8.33116770e-01 4.27635401e-01
2.10731328e-02 -1.14713204e+00 4.23167869e-02 8.03246200e-01
1.96797237e-01 -2.12596282e-01 1.49806321e+00 -9.49192405e-01
-8.80220592e-01 5.93841560e-02 -6.76302195e-01 2.82598995e-02
1.72185853e-01 -5.69597781e-01 -5.95793128e-01 2.34930784e-01
-2.15789557e-01 -6.91184223e-01 7.75258541e-01 -2.78832644e-01
9.61138844e-01 -5.65675974e-01 -3.27793568e-01 -8.49657878e-02
1.09343851e+00 1.53318117e-03 7.69645512e-01 3.31045479e-01
1.06203926e+00 7.02188849e-01 6.73547328e-01 -2.06271306e-01
1.44803691e+00 4.10537601e-01 7.33356059e-01 5.02939343e-01
-3.38906348e-01 -6.98308229e-01 -3.93245742e-02 9.52758491e-01
2.39950299e-01 -5.62010169e-01 -1.03813899e+00 7.68069208e-01
-2.07020593e+00 -7.84582496e-01 -4.72572803e-01 1.65828967e+00
1.02520561e+00 -7.62978420e-02 -2.14170933e-01 -1.93455845e-01
5.08213878e-01 1.55019015e-01 -6.76072717e-01 4.85231504e-02
6.91858903e-02 -7.18245208e-02 1.25752091e-01 8.24005544e-01
-7.29007244e-01 1.36114562e+00 5.26089621e+00 7.49948084e-01
-4.03697878e-01 1.55111328e-01 1.48354754e-01 5.13449311e-01
-4.79369104e-01 9.07334015e-02 -6.99547708e-01 4.67299297e-02
8.21378410e-01 -2.55907357e-01 5.31569839e-01 5.50696433e-01
-5.34704268e-01 -1.33768156e-01 -1.15346038e+00 8.13630700e-01
2.14014381e-01 -1.28722978e+00 4.97222453e-01 -5.66873491e-01
2.57121384e-01 -1.56638827e-02 -3.95761192e-01 8.53478193e-01
5.07775843e-01 -9.99299049e-01 2.24839717e-01 6.21494293e-01
1.38423741e-01 -6.07795358e-01 7.96585858e-01 3.32576394e-01
-1.37270081e+00 -1.11243650e-01 -4.50170755e-01 2.58937404e-02
4.17682111e-01 1.82259336e-01 -9.65492904e-01 1.05919230e+00
6.57089293e-01 3.60048205e-01 -8.04756761e-01 7.62020171e-01
-8.29393804e-01 6.20793164e-01 -1.73442453e-01 -4.40171570e-01
1.92183852e-01 5.23666665e-02 2.16881558e-01 6.22330368e-01
-2.58084349e-02 6.29407883e-01 7.25790486e-02 6.68793201e-01
-5.26075304e-01 1.47544533e-01 -1.74083635e-01 -2.62274474e-01
6.38030350e-01 9.74500477e-01 -2.85088897e-01 -4.50303644e-01
-6.13638282e-01 1.01710999e+00 9.01231408e-01 4.65516180e-01
-8.72980535e-01 -7.65396595e-01 2.99858958e-01 -1.64979175e-01
3.52336526e-01 -1.43453538e-01 2.81582415e-01 -1.03796053e+00
2.63666868e-01 -8.08431506e-01 9.87874866e-01 -1.06609964e+00
-1.34893715e+00 7.41472721e-01 -6.63791671e-02 -4.62954879e-01
5.35322644e-04 -6.18302464e-01 -4.89384085e-01 6.55512273e-01
-1.95868385e+00 -1.14196301e+00 -4.96519595e-01 7.67262936e-01
1.38911396e-01 2.20319495e-01 7.36653984e-01 4.52444106e-01
-4.04462188e-01 5.84092736e-01 -6.35365427e-01 3.21013242e-01
4.73127395e-01 -1.16231942e+00 6.18204474e-01 6.22576356e-01
2.00219810e-01 6.22171879e-01 4.07473505e-01 -8.02810967e-01
-1.72936368e+00 -1.13940787e+00 1.23698914e+00 -9.81190860e-01
8.67753804e-01 6.59103543e-02 -1.46298540e+00 8.77621174e-01
3.62633705e-01 -1.98427960e-02 5.07083058e-01 2.73858279e-01
-5.89834094e-01 1.66307434e-01 -8.99975181e-01 5.99900663e-01
1.26100504e+00 -6.43794119e-01 -1.30757403e+00 4.79153574e-01
1.70999277e+00 -5.14377594e-01 -9.80487585e-01 4.24641877e-01
2.86315568e-03 -3.94020110e-01 9.03937638e-01 -1.02116585e+00
2.47275487e-01 -5.69940031e-01 -1.84906840e-01 -1.10895717e+00
-3.02187294e-01 -3.09881061e-01 -8.20484340e-01 1.07869625e+00
7.89048851e-01 -7.45481312e-01 8.09371233e-01 6.20596707e-01
-1.64961830e-01 -8.71517122e-01 -9.63778973e-01 -5.29105425e-01
-1.49107739e-01 -4.83126007e-02 1.02354133e+00 8.81698549e-01
3.36475700e-01 1.16955507e+00 8.17857757e-02 6.01869404e-01
1.82808325e-01 9.92956832e-02 6.69556737e-01 -1.09252799e+00
-2.36724555e-01 -6.26001284e-02 -9.46248397e-02 -1.62445557e+00
3.02006006e-01 -8.63824964e-01 2.62350943e-02 -2.43181992e+00
-2.26032883e-02 -2.06870407e-01 -2.49114543e-01 6.12513363e-01
-8.36728334e-01 -4.06363487e-01 -1.43962592e-01 -1.01822928e-01
-1.15740895e+00 9.42713141e-01 1.80445278e+00 -4.09980893e-01
1.54138347e-02 -3.97314966e-01 -8.53347123e-01 4.69848901e-01
5.84896982e-01 -3.34841460e-01 -1.04924655e+00 -1.00914407e+00
7.59400964e-01 4.76004444e-02 4.44572479e-01 -7.05646038e-01
7.36359060e-01 -1.53813645e-01 -2.21234113e-01 -3.74838144e-01
3.69854152e-01 -9.86596584e-01 -4.97431010e-01 2.63636589e-01
-2.81395555e-01 2.37355232e-01 7.11692274e-02 7.33240843e-01
-3.29303205e-01 -2.93287337e-01 1.34471074e-01 -7.51762018e-02
-1.16000032e+00 5.22337496e-01 1.85541585e-01 7.47838974e-01
6.76717937e-01 2.69254982e-01 -1.02020824e+00 -5.39142430e-01
-4.26888615e-01 1.10696173e+00 -1.88692525e-01 6.69895887e-01
8.03773344e-01 -1.37696958e+00 -5.42334795e-01 -4.20937955e-01
5.73553085e-01 3.43115807e-01 4.43173140e-01 6.87153101e-01
-5.27969480e-01 5.08252800e-01 2.52653301e-01 -7.84641951e-02
-8.30204308e-01 8.35355699e-01 5.83952725e-01 -5.76911390e-01
-3.90476704e-01 1.12981105e+00 -3.60799544e-02 -8.68618608e-01
2.31025159e-01 -5.17965615e-01 -8.48966062e-01 -1.74010396e-01
4.99131978e-01 3.60806525e-01 2.99254600e-02 -4.22589958e-01
-4.97713894e-01 3.12311918e-01 -1.73550591e-01 2.01146930e-01
6.52418733e-01 -2.09685624e-01 -5.58963001e-01 2.12546572e-01
1.00317407e+00 -3.03704053e-01 -5.77037156e-01 -5.45467794e-01
3.82181734e-01 6.82063960e-03 -3.41074914e-01 -9.65154767e-01
-7.41044164e-01 6.37376666e-01 -6.57963902e-02 2.45195359e-01
6.72555327e-01 4.34090734e-01 1.13554490e+00 1.21292329e+00
4.16875660e-01 -8.17648768e-01 4.90493357e-01 1.15478230e+00
1.03400123e+00 -1.11533940e+00 -1.72193766e-01 -6.86109424e-01
-5.87423503e-01 7.61238158e-01 1.27861071e+00 2.30086535e-01
4.63894904e-01 -6.49573386e-01 9.93347540e-02 -6.82589889e-01
-8.58154297e-01 -3.64736617e-01 3.89910907e-01 6.16148949e-01
1.12068363e-01 -2.02197015e-01 -3.05005521e-01 7.29280770e-01
-2.09085241e-01 -6.36221319e-02 2.63869405e-01 8.84036899e-01
-6.12912118e-01 -9.47340012e-01 1.96530968e-01 4.62831408e-01
3.58182117e-02 -3.50598752e-01 -5.51567197e-01 6.27713025e-01
-2.71149099e-01 1.40571141e+00 -3.44836265e-01 -4.55138534e-01
8.02668333e-01 3.67107093e-01 5.62141240e-01 -5.72326362e-01
-5.47412932e-01 -1.13822222e+00 6.14932954e-01 -6.85193539e-01
-1.13898531e-01 -2.65086186e-03 -1.90336454e+00 -6.63262904e-02
-5.16790509e-01 6.96943462e-01 2.57238485e-02 1.06370842e+00
6.76623106e-01 1.01845133e+00 7.13467449e-02 3.24929774e-01
-4.70159352e-01 -9.81643677e-01 -9.33350697e-02 3.85220736e-01
1.70064911e-01 -5.64577401e-01 -3.55719663e-02 -2.69183576e-01] | [10.52880859375, 7.9095892906188965] |
fdc28e7a-f7a0-459f-abde-4ce61f1e1f61 | speech-to-sql-towards-speech-driven-sql-query | 2201.01209 | null | https://arxiv.org/abs/2201.01209v1 | https://arxiv.org/pdf/2201.01209v1.pdf | Speech-to-SQL: Towards Speech-driven SQL Query Generation From Natural Language Question | Speech-based inputs have been gaining significant momentum with the popularity of smartphones and tablets in our daily lives, since voice is the most easiest and efficient way for human-computer interaction. This paper works towards designing more effective speech-based interfaces to query the structured data in relational databases. We first identify a new task named Speech-to-SQL, which aims to understand the information conveyed by human speech and directly translate it into structured query language (SQL) statements. A naive solution to this problem can work in a cascaded manner, that is, an automatic speech recognition (ASR) component followed by a text-to-SQL component. However, it requires a high-quality ASR system and also suffers from the error compounding problem between the two components, resulting in limited performance. To handle these challenges, we further propose a novel end-to-end neural architecture named SpeechSQLNet to directly translate human speech into SQL queries without an external ASR step. SpeechSQLNet has the advantage of making full use of the rich linguistic information presented in speech. To the best of our knowledge, this is the first attempt to directly synthesize SQL based on arbitrary natural language questions, rather than a natural language-based version of SQL or its variants with a limited SQL grammar. To validate the effectiveness of the proposed problem and model, we further construct a dataset named SpeechQL, by piggybacking the widely-used text-to-SQL datasets. Extensive experimental evaluations on this dataset show that SpeechSQLNet can directly synthesize high-quality SQL queries from human speech, outperforming various competitive counterparts as well as the cascaded methods in terms of exact match accuracies. | ['Di Jiang', 'Xuefang Zhao', 'Raymond Chi-Wing Wong', 'Yuanfeng Song'] | 2022-01-04 | null | null | null | null | ['text-to-sql'] | ['computer-code'] | [ 6.86348900e-02 1.61949903e-01 6.54606149e-02 -6.88506424e-01
-1.08659506e+00 -4.45200264e-01 5.24303019e-01 1.38520226e-01
-4.13711399e-01 2.25843996e-01 2.16177702e-01 -7.42451072e-01
3.29214424e-01 -9.07472968e-01 -6.75030351e-01 -9.10424367e-02
5.03185689e-01 6.59989715e-01 3.85139674e-01 -6.04455829e-01
-1.29698411e-01 2.62717515e-01 -1.43543768e+00 5.72164595e-01
1.02580750e+00 1.01750147e+00 4.73405868e-01 6.39065385e-01
-7.43622303e-01 1.08284533e+00 -8.73414636e-01 -6.99537992e-01
-1.66582316e-01 -4.90011185e-01 -8.49342167e-01 -1.19906738e-01
1.89459305e-02 -4.49803293e-01 -3.09457839e-01 7.89214551e-01
6.11326694e-01 7.31321797e-02 -3.14514115e-02 -9.55999374e-01
-6.07169747e-01 8.69987249e-01 2.44373798e-01 -1.81834042e-01
9.75745618e-01 6.15717322e-02 1.11934185e+00 -1.02272511e+00
2.84775704e-01 1.42499816e+00 3.34663749e-01 5.76800108e-01
-9.83416855e-01 -4.55220163e-01 -1.62447952e-02 -1.73548415e-01
-1.43242872e+00 -8.04870963e-01 6.52107477e-01 -4.00148332e-02
1.33543444e+00 5.40207386e-01 2.32535169e-01 1.25788641e+00
-4.23029572e-01 1.02327669e+00 6.47307098e-01 -3.65911931e-01
3.23720634e-01 4.08125430e-01 2.80781910e-02 5.66055596e-01
-3.00318033e-01 -1.10225417e-01 -5.50041020e-01 -8.88847187e-02
3.82835776e-01 1.78238712e-02 -2.65548706e-01 1.96950957e-01
-1.33284402e+00 6.53665423e-01 3.31564933e-01 3.47104788e-01
-3.97478521e-01 -1.95974082e-01 4.67555732e-01 3.90875906e-01
1.92497566e-01 4.62769717e-01 -3.70322317e-01 -4.57484365e-01
-8.39150250e-01 1.69026420e-01 1.22472441e+00 1.23196721e+00
5.16642690e-01 1.88875079e-01 -1.65436059e-01 1.14550543e+00
2.63088584e-01 7.73732841e-01 6.79174006e-01 -5.81344068e-01
8.45469117e-01 7.59542406e-01 7.00586140e-02 -8.07526290e-01
-1.59243047e-01 -1.81506202e-01 -6.51814580e-01 -4.09481555e-01
2.22603068e-01 -1.85630079e-02 -6.01473391e-01 1.51985002e+00
2.81073153e-01 -3.88963312e-01 3.85378271e-01 7.99070299e-01
1.14455831e+00 1.01593828e+00 -1.24615781e-01 1.13981189e-02
1.46774912e+00 -9.65394676e-01 -8.53822589e-01 -3.10489863e-01
5.67494214e-01 -7.24236190e-01 1.71090388e+00 2.42199004e-01
-1.04010284e+00 -6.96191370e-01 -9.00992751e-01 -3.15160275e-01
-4.27257866e-01 3.45496446e-01 4.03997600e-01 7.38532126e-01
-1.07044303e+00 7.84139857e-02 -1.03442359e+00 -5.22751331e-01
-7.06458986e-02 2.06209019e-01 -2.02375814e-01 1.88524753e-01
-1.29894602e+00 5.67913532e-01 2.89884061e-01 1.01298489e-01
-5.23917556e-01 -4.48861659e-01 -1.00760877e+00 2.09440932e-01
8.92870605e-01 -5.52836418e-01 1.72098374e+00 -6.60507679e-01
-2.05394888e+00 6.21641517e-01 -6.06562436e-01 -4.73928452e-01
2.93557078e-01 -1.86430246e-01 -4.91285264e-01 9.94200036e-02
9.04296488e-02 2.56959498e-01 4.69033599e-01 -8.90743852e-01
-4.73729551e-01 -3.71210545e-01 1.10335708e-01 1.01806400e-02
-4.10309285e-01 4.31818724e-01 -7.50630915e-01 -6.66875064e-01
3.99054550e-02 -7.89199412e-01 3.19139324e-02 -3.57936144e-01
-6.90642238e-01 -4.46675420e-01 4.22308385e-01 -7.12626278e-01
1.56180787e+00 -2.09696555e+00 -1.00096159e-01 -1.59467235e-02
-4.75871935e-03 5.64149141e-01 -4.11513783e-02 8.80856156e-01
1.36555374e-01 2.18279526e-01 -2.46038973e-01 -4.55874234e-01
2.41211206e-01 3.32355857e-01 -7.98324645e-01 -3.43393892e-01
4.48337853e-01 1.07231700e+00 -7.53224134e-01 -3.26607168e-01
8.56319740e-02 3.33511680e-01 -4.96379137e-01 7.65060544e-01
-4.74754095e-01 9.53659192e-02 -5.72464466e-01 6.17133617e-01
2.16783404e-01 -2.70994127e-01 6.82460740e-02 -1.50209531e-01
-1.16406744e-02 1.01568508e+00 -9.05013382e-01 1.54178417e+00
-7.95905769e-01 1.24590605e-01 1.53587610e-01 -1.08211172e+00
1.20122623e+00 4.78995621e-01 1.54228136e-01 -9.38206851e-01
-1.31051794e-01 5.47570825e-01 -2.98073947e-01 -6.13030910e-01
5.07408023e-01 -1.63732871e-01 -3.54517430e-01 2.34648585e-01
7.71776140e-02 -4.49770659e-01 -3.83611247e-02 2.96550035e-01
1.10085249e+00 -2.36379132e-01 2.28122607e-01 4.08127338e-01
7.67683029e-01 -1.41396925e-01 1.30645871e-01 8.42507482e-01
1.88560486e-01 6.86695278e-01 4.86329705e-01 3.74842398e-02
-7.41870701e-01 -9.82915998e-01 3.68878931e-01 1.17683029e+00
-1.42841846e-01 -6.16101623e-01 -8.52919936e-01 -6.32970929e-01
-1.47985101e-01 1.02869463e+00 4.49389517e-02 -1.14698485e-01
-6.71062768e-01 -1.82554573e-01 9.94960129e-01 3.84987712e-01
5.25808036e-01 -1.30755866e+00 -2.76081353e-01 3.90017301e-01
-3.99862885e-01 -1.70259714e+00 -6.67935133e-01 -6.13718759e-04
-5.53993642e-01 -6.61928058e-01 -6.58841193e-01 -7.09836781e-01
1.26781613e-01 1.36967242e-01 1.08125556e+00 -3.51324715e-02
1.85199767e-01 1.70508951e-01 -6.19810641e-01 -2.68128753e-01
-8.92366469e-01 3.69036853e-01 -2.37474993e-01 2.94189125e-01
3.01795483e-01 -3.17058504e-01 -1.45946145e-01 4.10382390e-01
-1.14450526e+00 4.76644784e-02 6.68961704e-01 7.06247747e-01
4.42488223e-01 -6.61751404e-02 7.84950614e-01 -7.87207723e-01
9.24050868e-01 -2.42399260e-01 -5.63812792e-01 4.44844991e-01
-4.05878067e-01 1.79966316e-01 1.11137533e+00 -1.71925500e-01
-1.08225763e+00 1.33259743e-01 -8.80997896e-01 -2.22618893e-01
-2.34017387e-01 8.79981816e-01 -5.22489846e-01 3.79647970e-01
6.02098942e-01 6.55418336e-01 1.22156225e-01 -5.71608841e-01
3.83487463e-01 1.31146252e+00 7.84770727e-01 -5.22473037e-01
6.37287557e-01 1.40199021e-01 -6.11505032e-01 -1.19572926e+00
-6.10927045e-01 -4.81386483e-01 8.95023346e-03 2.10925102e-01
8.91943455e-01 -7.89882421e-01 -1.01400924e+00 5.28497696e-01
-1.44688845e+00 -1.13716528e-01 -2.52238959e-01 3.02872449e-01
-5.65047204e-01 3.86312813e-01 -6.84389353e-01 -9.88804638e-01
-4.73794252e-01 -1.31709707e+00 1.39618206e+00 -4.97378558e-02
-2.19812363e-01 -5.64946532e-01 -3.45035851e-01 7.32742131e-01
6.11634851e-01 -3.52621794e-01 8.89249623e-01 -1.24911368e+00
-5.73220611e-01 -3.40942323e-01 -1.04652934e-01 4.84349549e-01
1.87629580e-01 -2.24272206e-01 -7.89585590e-01 1.47997644e-02
2.42844790e-01 -4.88466382e-01 3.84563148e-01 -2.76667804e-01
9.90869761e-01 -5.49818039e-01 8.59406739e-02 4.24105912e-01
1.00024796e+00 5.49576998e-01 5.39524555e-01 -4.26576212e-02
6.01904452e-01 6.12781882e-01 3.92772943e-01 2.59419650e-01
8.19041073e-01 9.54118907e-01 1.91275746e-01 -8.90282020e-02
4.83515002e-02 -7.14428663e-01 6.48303866e-01 1.26140332e+00
7.13495135e-01 -4.17638332e-01 -1.08923984e+00 2.85420895e-01
-1.59261703e+00 -6.81538701e-01 1.45119861e-01 2.22892046e+00
1.02152550e+00 7.99610689e-02 1.40369520e-01 4.34047766e-02
4.70378846e-01 1.71458155e-01 -4.29084122e-01 -3.58585089e-01
4.85195639e-03 4.72077042e-01 -1.16559766e-01 5.63343287e-01
-7.84009278e-01 1.16203749e+00 5.35596514e+00 8.18497181e-01
-1.36471927e+00 -1.38755828e-01 2.73567826e-01 1.90066770e-01
-3.97523493e-01 -1.90858722e-01 -9.11599159e-01 4.54831511e-01
1.26182425e+00 -1.57513544e-01 6.82538271e-01 8.63776863e-01
3.30621183e-01 1.56162009e-01 -1.38342381e+00 1.08034039e+00
2.77031120e-02 -1.19237220e+00 3.70068818e-01 -4.61630464e-01
-1.26558170e-01 -3.01078167e-02 -2.99691021e-01 5.72547376e-01
9.40272734e-02 -9.59346294e-01 7.89968789e-01 3.61206025e-01
7.06997335e-01 -4.99859303e-01 7.03405499e-01 6.56144559e-01
-1.20187926e+00 3.38232778e-02 1.45783238e-02 1.91815302e-01
3.49209219e-01 4.81070638e-01 -1.04169440e+00 6.48004830e-01
5.43493032e-01 3.30421150e-01 -4.98655647e-01 5.27500808e-01
-1.26818374e-01 6.11702800e-01 -4.42059249e-01 -4.95808542e-01
3.35388966e-02 -5.74847572e-02 4.16557252e-01 1.20905626e+00
4.73709255e-01 1.26612425e-01 2.53740430e-01 1.09365606e+00
-1.13517962e-01 3.15014750e-01 -5.35865664e-01 -6.92602277e-01
7.33538687e-01 8.73301744e-01 -3.70455325e-01 -4.97735709e-01
-6.48287296e-01 1.14151037e+00 1.93914682e-01 4.21471417e-01
-6.41085923e-01 -9.01164770e-01 4.40346450e-01 2.47782394e-01
2.81684488e-01 -2.69329011e-01 -9.04323097e-05 -1.35596406e+00
6.56217754e-01 -1.52551031e+00 1.24706805e-01 -9.21449304e-01
-1.17526114e+00 1.14166403e+00 -3.17725033e-01 -9.54457760e-01
-6.32949531e-01 -2.88023412e-01 -2.15020373e-01 1.05734122e+00
-1.34758735e+00 -9.30364192e-01 -2.68618733e-01 6.36575520e-01
6.89377308e-01 -2.82256901e-01 9.58776236e-01 4.94267493e-01
-4.87858921e-01 7.95253515e-01 -3.60246778e-01 3.74710768e-01
4.65141952e-01 -1.06336522e+00 8.08748245e-01 7.41134942e-01
2.45402813e-01 9.12387550e-01 4.41926748e-01 -4.40710664e-01
-1.78283334e+00 -1.02104068e+00 1.26682293e+00 -3.84052098e-01
7.03516364e-01 -7.49132872e-01 -1.16159260e+00 4.91358697e-01
-2.05030125e-02 -2.79635191e-01 4.58248287e-01 -2.82400995e-01
-2.45893374e-01 -4.11356449e-01 -8.20894778e-01 5.45031846e-01
8.75110447e-01 -1.14275050e+00 -7.20033705e-01 1.95380598e-01
1.47500479e+00 -4.79025155e-01 -7.53942609e-01 4.41379398e-01
1.78512543e-01 -9.16433215e-01 8.49515915e-01 -4.26673025e-01
2.49088570e-01 -3.53275329e-01 -4.47432250e-01 -1.04564440e+00
5.61788738e-01 -9.39279854e-01 7.74153173e-02 1.51458836e+00
6.13474250e-01 -7.97719300e-01 7.24200785e-01 5.66287160e-01
-3.84171039e-01 -8.36665452e-01 -8.07783663e-01 -7.29850709e-01
-2.80980855e-01 -7.54383504e-01 8.78649771e-01 5.53501248e-01
1.04354188e-01 6.80654466e-01 -1.16680786e-01 1.88069493e-01
7.08230864e-03 1.23020910e-01 9.88237917e-01 -7.23616183e-01
-5.57014823e-01 -2.97983319e-01 -1.48610935e-01 -1.71225524e+00
5.36073707e-02 -9.77480471e-01 2.80557007e-01 -1.37062931e+00
-2.90917546e-01 -3.42749089e-01 4.25749481e-01 4.80205595e-01
-1.56413481e-01 -5.32028615e-01 2.75665730e-01 8.31428692e-02
-4.35962886e-01 8.74297023e-01 8.95234168e-01 -7.53508955e-02
-4.21478331e-01 2.77064830e-01 -7.82785654e-01 2.74949908e-01
4.06970829e-01 -2.80671030e-01 -5.70628822e-01 -5.42652369e-01
2.51465559e-01 8.38721156e-01 2.09722951e-01 -9.31330621e-01
5.02455652e-01 1.15860002e-02 -4.49532330e-01 -5.60778916e-01
3.88276219e-01 -7.09759414e-01 -4.43176664e-02 1.35665968e-01
-5.56331396e-01 2.00171098e-01 -8.81524384e-02 2.18302146e-01
-8.41057122e-01 -2.00346708e-02 4.22848284e-01 -4.37770784e-02
-3.19174051e-01 6.97581917e-02 -4.15196478e-01 2.03207508e-01
5.23202240e-01 1.50924893e-02 -2.50114948e-01 -8.86634350e-01
-6.16047859e-01 2.66718060e-01 6.36617020e-02 7.34344244e-01
8.88033509e-01 -1.12020063e+00 -4.84006405e-01 5.73038816e-01
3.18885922e-01 1.84238330e-01 -1.77711517e-01 6.49368286e-01
-4.85491186e-01 8.30141306e-01 5.26838124e-01 -5.94523907e-01
-9.29432511e-01 5.31868756e-01 4.09235388e-01 -2.16700867e-01
-3.83878380e-01 7.22149312e-01 -7.37040937e-02 -9.94691133e-01
5.95496774e-01 -7.95749187e-01 5.55304140e-02 -1.56912342e-01
5.33957005e-01 1.01283364e-01 4.55319226e-01 -5.50438762e-01
-3.66116166e-01 2.55364388e-01 9.13574845e-02 -3.49838316e-01
1.04873979e+00 -5.87489009e-02 -1.11554258e-01 4.79952782e-01
1.39254665e+00 1.24074474e-01 -4.03371960e-01 -4.90339577e-01
2.99361736e-01 -1.37061715e-01 -3.24892610e-01 -7.17298448e-01
-9.29061651e-01 9.75634515e-01 9.77113768e-02 5.46364546e-01
1.10559750e+00 1.53005093e-01 1.45075953e+00 9.42050099e-01
4.66181159e-01 -8.38608742e-01 2.09945410e-01 7.15394795e-01
1.03767514e+00 -1.27304947e+00 -7.24027991e-01 -5.54984689e-01
-8.44329178e-01 1.09342635e+00 4.71669942e-01 2.83539802e-01
5.43291986e-01 3.49107176e-01 2.90570736e-01 -1.07919544e-01
-8.58181119e-01 -2.97105134e-01 1.40027285e-01 3.35075527e-01
5.06891251e-01 -9.56075415e-02 2.33998261e-02 9.50014532e-01
-6.55089259e-01 1.56729072e-01 2.70608187e-01 8.01365852e-01
-2.40450412e-01 -1.23505986e+00 -2.31168732e-01 2.95761645e-01
-4.16687340e-01 -2.66645968e-01 -5.95411003e-01 4.64743167e-01
-5.18917859e-01 1.43399596e+00 -1.17159329e-01 -6.33624077e-01
9.22753096e-01 2.77332157e-01 -2.14229867e-01 -8.59187305e-01
-7.35456467e-01 3.83855775e-02 2.77573138e-01 -8.34406972e-01
-1.52331088e-02 -3.05964082e-01 -1.52074873e+00 -1.22049503e-01
-2.26654619e-01 3.23702544e-01 7.42866337e-01 7.97709167e-01
5.34440994e-01 4.29470778e-01 8.32722306e-01 -3.20613027e-01
-1.06902266e+00 -8.05357695e-01 -1.88818023e-01 4.22757000e-01
4.79954600e-01 2.01764349e-02 -2.52972066e-01 -1.97963059e-01] | [14.218802452087402, 6.968374729156494] |
94d7e4e4-24af-4c51-825b-7d21fb02afcb | integrating-heterogeneous-domain-information | 2212.10714 | null | https://arxiv.org/abs/2212.10714v1 | https://arxiv.org/pdf/2212.10714v1.pdf | Integrating Heterogeneous Domain Information into Relation Extraction: A Case Study on Drug-Drug Interaction Extraction | The development of deep neural networks has improved representation learning in various domains, including textual, graph structural, and relational triple representations. This development opened the door to new relation extraction beyond the traditional text-oriented relation extraction. However, research on the effectiveness of considering multiple heterogeneous domain information simultaneously is still under exploration, and if a model can take an advantage of integrating heterogeneous information, it is expected to exhibit a significant contribution to many problems in the world. This thesis works on Drug-Drug Interactions (DDIs) from the literature as a case study and realizes relation extraction utilizing heterogeneous domain information. First, a deep neural relation extraction model is prepared and its attention mechanism is analyzed. Next, a method to combine the drug molecular structure information and drug description information to the input sentence information is proposed, and the effectiveness of utilizing drug molecular structures and drug descriptions for the relation extraction task is shown. Then, in order to further exploit the heterogeneous information, drug-related items, such as protein entries, medical terms and pathways are collected from multiple existing databases and a new data set in the form of a knowledge graph (KG) is constructed. A link prediction task on the constructed data set is conducted to obtain embedding representations of drugs that contain the heterogeneous domain information. Finally, a method that integrates the input sentence information and the heterogeneous KG information is proposed. The proposed model is trained and evaluated on a widely used data set, and as a result, it is shown that utilizing heterogeneous domain information significantly improves the performance of relation extraction from the literature. | ['Masaki Asada'] | 2022-12-21 | null | null | null | null | ['drug-drug-interaction-extraction'] | ['natural-language-processing'] | [ 2.91009128e-01 2.95254439e-01 -8.06515098e-01 -5.15140444e-02
-5.87251246e-01 -2.05730811e-01 2.72376716e-01 7.85533249e-01
-2.17335150e-01 1.24910820e+00 3.93065006e-01 -4.02241290e-01
-5.25147021e-01 -1.15424585e+00 -5.79682767e-01 -6.67003214e-01
-2.90258437e-01 4.64155704e-01 -2.42397159e-01 -2.09480703e-01
-2.38082241e-02 5.59614778e-01 -9.75777149e-01 2.59466112e-01
1.10815513e+00 7.56744981e-01 1.57186687e-01 -2.48675942e-02
-3.03101420e-01 8.79280210e-01 -5.60512662e-01 -3.83124650e-01
-1.53389350e-01 -4.66623873e-01 -1.01258206e+00 -2.56460458e-01
-3.31612289e-01 -1.10630818e-01 -4.76838052e-01 8.63554835e-01
7.89217055e-01 2.16037687e-03 7.02702224e-01 -8.02577734e-01
-6.59130454e-01 6.78381443e-01 -3.10995609e-01 1.08229823e-01
5.33615232e-01 -2.26736769e-01 1.19459105e+00 -9.21101928e-01
1.13577044e+00 9.48964894e-01 3.60699862e-01 5.78130372e-02
-9.68067110e-01 -7.10786283e-01 -1.28767863e-01 4.17978495e-01
-1.59757674e+00 1.22271411e-01 9.54074204e-01 -3.58435541e-01
1.65071487e+00 -1.14740238e-01 6.54562116e-01 1.12257898e+00
5.05173028e-01 5.56289494e-01 7.12020457e-01 -4.22322303e-01
-2.67544221e-02 3.33158553e-01 3.65630627e-01 8.33812773e-01
6.82946265e-01 -7.69668608e-04 -3.65529835e-01 -2.21241698e-01
1.76766858e-01 5.58929928e-02 -3.73080850e-01 -6.67237267e-02
-7.43654966e-01 9.46974337e-01 7.10904002e-01 6.19252682e-01
-7.14908302e-01 -5.99230647e-01 5.99519134e-01 -4.15676460e-02
7.07676649e-01 7.76195109e-01 -5.46594203e-01 3.58239561e-01
-6.16268337e-01 3.53440419e-02 1.03480136e+00 8.04461181e-01
5.96311867e-01 -8.26851130e-02 -3.28505561e-02 7.73527384e-01
2.05043674e-01 1.64301321e-01 5.14048278e-01 -3.18788812e-02
8.17438662e-01 1.20560443e+00 -3.51549864e-01 -1.31891942e+00
-6.76080108e-01 -5.69974899e-01 -1.01282752e+00 -4.12496597e-01
-1.44200295e-01 -3.43637049e-01 -8.32611859e-01 1.61065447e+00
4.95157897e-01 7.53285438e-02 5.70664585e-01 5.20154655e-01
1.60451484e+00 5.25403678e-01 5.35149336e-01 -2.76547164e-01
1.54906654e+00 -5.49300551e-01 -1.14733994e+00 1.04165561e-01
9.66182590e-01 -4.00439948e-01 1.42337054e-01 2.17796177e-01
-7.01218426e-01 -2.86649227e-01 -1.42846537e+00 -1.54238805e-01
-1.05328584e+00 1.06214710e-01 8.13502133e-01 4.54788469e-02
-4.51409876e-01 6.69243038e-01 -4.84157205e-01 -3.74241501e-01
8.33245218e-01 7.47223914e-01 -6.60670102e-01 -3.39540303e-01
-1.83515835e+00 1.37203360e+00 8.45732450e-01 -6.98212236e-02
-4.83595520e-01 -8.08772624e-01 -1.12582421e+00 2.50269264e-01
5.45399725e-01 -9.45803463e-01 3.62050027e-01 -3.80624264e-01
-1.15509748e+00 4.29386288e-01 -4.43640836e-02 -4.78793710e-01
-1.02772728e-01 1.21993735e-01 -6.03537858e-01 3.68828028e-01
3.82372104e-02 2.58732051e-01 1.45152226e-01 -8.68051231e-01
-2.81776458e-01 -6.68377876e-01 1.60350934e-01 3.70928973e-01
-5.89259744e-01 1.40526006e-03 -3.46548438e-01 -4.60627079e-01
-3.06782603e-01 -7.24658430e-01 -2.31763393e-01 -6.66733861e-01
-8.00146043e-01 -4.41214919e-01 4.51473981e-01 -8.76876831e-01
1.52502930e+00 -1.80833113e+00 5.83142042e-01 3.94627154e-01
6.40587270e-01 4.42008018e-01 -1.26236632e-01 8.47031415e-01
-4.68823731e-01 2.89164364e-01 -2.09137365e-01 2.39547387e-01
-3.95429581e-01 2.17230961e-01 2.63373882e-01 1.24199308e-01
5.62100887e-01 9.58605111e-01 -8.40483308e-01 -7.04765618e-01
-4.41114940e-02 7.14697480e-01 -3.71814430e-01 1.51789173e-01
-1.15122408e-01 2.12985590e-01 -8.61239433e-01 7.62042642e-01
4.80716556e-01 -4.80859101e-01 6.91510558e-01 -5.95494926e-01
3.39049369e-01 3.89142126e-01 -6.28324032e-01 1.42534661e+00
-3.91415536e-01 2.16647729e-01 -5.36841333e-01 -1.35031629e+00
9.15416479e-01 7.37704515e-01 1.04157960e+00 -5.80537140e-01
3.22698981e-01 1.71036392e-01 3.41990262e-01 -8.22297156e-01
1.48575991e-01 -2.03169286e-01 3.23384225e-01 -1.71953309e-02
3.53599399e-01 2.71002918e-01 4.01452333e-01 3.08540672e-01
1.34853613e+00 1.48119494e-01 7.85844088e-01 1.22477852e-01
7.19983399e-01 2.04398885e-01 4.40509945e-01 -2.54388005e-02
3.53099376e-01 -2.26057202e-01 5.46853662e-01 -5.55112243e-01
-6.64600134e-01 -4.47279245e-01 -3.64095747e-01 2.88681984e-01
-7.16153085e-02 -8.16232860e-01 -5.20241261e-01 -8.21039498e-01
1.46518528e-01 3.27708423e-01 -7.41301417e-01 -4.30579036e-01
-2.59799778e-01 -1.19955719e+00 4.44153935e-01 3.85723501e-01
5.50246418e-01 -1.12457478e+00 -3.75381112e-02 3.99365306e-01
-1.05281979e-01 -1.25060761e+00 2.27467716e-01 5.60621858e-01
-8.82685661e-01 -1.39181185e+00 -3.50460410e-01 -7.96890259e-01
5.64359009e-01 -8.48479569e-02 9.60587621e-01 -6.30886406e-02
-3.59403461e-01 -2.32509509e-01 -3.08863014e-01 -5.25442541e-01
-3.10624301e-01 2.17890203e-01 -1.02672011e-01 -2.95872271e-01
8.17600012e-01 -3.81414801e-01 -3.04580301e-01 -2.79612690e-01
-1.02555931e+00 -5.09843323e-03 8.31832886e-01 9.58277524e-01
6.58792078e-01 3.24851334e-01 9.21959996e-01 -1.27053595e+00
9.01391685e-01 -9.27359223e-01 -1.96837485e-01 2.92899698e-01
-8.56989801e-01 1.72940761e-01 5.88546038e-01 -3.17275375e-01
-1.06919169e+00 1.21693425e-02 -1.88496038e-01 -9.34119076e-02
-1.14441410e-01 1.58563936e+00 -5.72689474e-01 1.54810756e-01
5.61012447e-01 -7.45522752e-02 7.58418406e-04 -3.89224797e-01
2.15392426e-01 7.46015370e-01 -1.92505851e-01 -4.12602156e-01
3.48734319e-01 -5.53796254e-02 3.65963787e-01 -7.86235034e-01
-6.58441842e-01 -2.55121857e-01 -5.85645974e-01 4.20097709e-01
1.06229377e+00 -7.95933306e-01 -6.67751789e-01 -7.09939897e-02
-1.29885435e+00 4.51486230e-01 3.31038125e-02 9.28861916e-01
-4.77129780e-02 2.48482004e-01 -7.38973081e-01 -3.85705352e-01
-5.90824485e-01 -1.28253484e+00 6.10051453e-01 1.99003696e-01
-3.30390543e-01 -1.18478537e+00 2.46650726e-01 4.09620792e-01
-1.04963154e-01 3.98842067e-01 1.58805490e+00 -1.33227623e+00
-5.04819989e-01 -5.15443981e-01 -2.16773435e-01 1.47993714e-01
6.30632877e-01 -2.10033417e-01 -6.15947843e-01 2.61386558e-02
-2.50785023e-01 -1.59887955e-01 5.75146675e-01 3.78042877e-01
8.24917436e-01 -4.25292134e-01 -7.15931654e-01 5.47613978e-01
1.60448909e+00 7.34156907e-01 6.40464783e-01 3.28664452e-01
9.79992926e-01 7.09571838e-01 3.31630886e-01 2.83128262e-01
3.39682847e-01 6.22571886e-01 2.37515196e-01 -3.71145308e-01
5.84122585e-03 -1.70758307e-01 -3.41497958e-02 9.10051048e-01
-3.52864861e-01 -3.21176052e-01 -7.75216579e-01 2.38505393e-01
-1.50493217e+00 -8.62463117e-01 6.22315109e-02 1.85836637e+00
1.30236840e+00 -5.82231469e-02 -1.84743807e-01 5.98964049e-03
4.81640518e-01 -2.75775611e-01 -7.12406695e-01 -3.99216086e-01
-1.40226737e-01 6.81380928e-01 3.34190428e-01 3.17256033e-01
-1.10498607e+00 9.63822484e-01 5.15569830e+00 8.75107229e-01
-8.97354126e-01 -3.23050350e-01 5.37999511e-01 3.77419323e-01
-2.42126361e-01 -2.57219017e-01 -8.78041625e-01 2.54302531e-01
1.20637035e+00 -4.33645606e-01 8.68995860e-02 6.77904129e-01
1.50733709e-01 1.05975680e-01 -1.19975245e+00 7.90822804e-01
6.87510446e-02 -1.63543284e+00 3.04265350e-01 4.01473701e-01
4.15393949e-01 -6.96686804e-02 -2.03302473e-01 2.95591205e-01
3.37768853e-01 -1.14181399e+00 -4.53024119e-01 4.63140607e-01
5.36991239e-01 -8.94281983e-01 1.25445378e+00 1.02711320e-01
-1.21158743e+00 6.81979805e-02 -2.45591789e-01 1.82789400e-01
-1.23702705e-01 5.95390201e-01 -1.40147424e+00 1.59076548e+00
1.96653247e-01 1.35302615e+00 -6.28953278e-01 8.76548469e-01
-2.18958884e-01 3.21888685e-01 -3.41122784e-02 -1.03206210e-01
9.51283574e-02 -3.69585156e-01 1.10810116e-01 1.10751224e+00
1.84378430e-01 2.96043277e-01 2.08234683e-01 7.51672745e-01
-3.74758810e-01 6.43054426e-01 -1.28094840e+00 -5.71153760e-01
2.42891341e-01 1.26980770e+00 -4.67320532e-01 -5.57988524e-01
-6.18932903e-01 6.12919986e-01 3.81679237e-01 5.17653346e-01
-6.84582055e-01 -5.72822511e-01 4.59928155e-01 -2.99772173e-01
5.91470208e-03 1.44497350e-01 -1.76677585e-01 -1.04804313e+00
-2.70972520e-01 -9.60691571e-01 5.94818115e-01 -5.31324387e-01
-1.24201143e+00 9.26073432e-01 1.62495583e-01 -1.31505966e+00
-2.65200764e-01 -5.79373598e-01 -4.84991558e-02 1.12751019e+00
-1.50217927e+00 -1.14183795e+00 1.76226288e-01 4.61926639e-01
2.87118286e-01 -4.30891156e-01 1.24828076e+00 7.97138393e-01
-9.99960780e-01 5.28222501e-01 -4.43723276e-02 3.12134981e-01
4.53519464e-01 -8.90029311e-01 -2.26559922e-01 2.88183391e-01
1.07642718e-01 1.04081404e+00 2.90328473e-01 -1.13294172e+00
-1.41680086e+00 -1.07497466e+00 1.13752544e+00 -2.85276502e-01
6.64210260e-01 5.10732420e-02 -1.05378938e+00 3.69978875e-01
4.14243877e-01 -1.62097201e-01 1.35574460e+00 7.65510574e-02
-1.64990962e-01 1.73141226e-01 -1.17098713e+00 4.35988575e-01
7.65920520e-01 -5.45412123e-01 -5.96647322e-01 6.16253018e-01
8.69935393e-01 -3.59227806e-01 -1.48633361e+00 5.56948364e-01
2.45883018e-01 -6.71188682e-02 1.16247857e+00 -1.30564666e+00
8.43038797e-01 -1.47585109e-01 -2.08718199e-02 -1.42181313e+00
-2.74594963e-01 -6.66189045e-02 -3.44988972e-01 1.02909720e+00
9.66147304e-01 -5.01222610e-01 5.49589574e-01 4.42193121e-01
-1.92389697e-01 -1.34950364e+00 -7.27383971e-01 -4.71940577e-01
3.22012901e-02 1.44393206e-01 6.29926443e-01 1.22200763e+00
5.41551769e-01 1.04932666e+00 -1.45258725e-01 3.96917462e-02
5.97582236e-02 1.64638996e-01 2.48548090e-01 -1.44414055e+00
-1.92481369e-01 -1.08348645e-01 -4.33370233e-01 -4.00998682e-01
3.73116940e-01 -1.39340127e+00 -5.42082071e-01 -2.04824114e+00
3.87727797e-01 -6.84456388e-03 -5.27687609e-01 6.35103583e-01
-3.58181298e-01 -1.36598930e-01 -1.39570355e-01 -6.24440722e-02
-1.76132530e-01 5.89340210e-01 1.48134923e+00 -5.99569201e-01
-3.65787119e-01 -2.51021534e-01 -8.14718783e-01 5.36082506e-01
7.03026533e-01 -4.38989997e-01 -6.58162117e-01 1.51345357e-01
2.63677627e-01 4.11589384e-01 -1.75017849e-01 -5.34785390e-01
1.94153577e-01 -4.81868710e-06 5.53176761e-01 -5.02201796e-01
3.95367473e-01 -9.46807861e-01 2.21531108e-01 4.80191857e-01
-3.84284914e-01 -6.27736300e-02 3.87240469e-01 6.76053226e-01
-5.74926674e-01 -1.99615598e-01 3.72468889e-01 -1.11331791e-01
-3.98491740e-01 5.25380135e-01 -1.45234749e-01 -1.38157874e-01
9.20032442e-01 -1.37189671e-01 -3.45755130e-01 7.64776543e-02
-9.12801564e-01 3.74047682e-02 -2.84922689e-01 3.14199090e-01
8.40785861e-01 -1.34480572e+00 -5.30982673e-01 3.39039452e-02
2.14072540e-01 -2.02882543e-01 1.74707577e-01 8.80340159e-01
-3.40131581e-01 6.28347993e-01 -7.86351785e-02 -1.20877907e-01
-1.42232358e+00 1.04994059e+00 6.16106689e-02 -7.79941380e-01
-5.64099789e-01 5.11378348e-01 6.09590001e-02 -5.34963347e-02
1.01948641e-01 -2.89295882e-01 -9.91939008e-01 4.28478986e-01
2.94458121e-01 -1.79771379e-01 3.36407512e-01 -6.56260252e-01
-5.12476861e-01 4.65151787e-01 -3.23819250e-01 4.30526793e-01
1.65422237e+00 4.43162799e-01 -3.16393614e-01 -1.27307206e-01
1.47692692e+00 -9.04653072e-02 -2.33356476e-01 -3.52246732e-01
2.18809545e-01 -1.24026109e-02 -1.23583362e-01 -9.44469392e-01
-1.15561938e+00 7.72174537e-01 2.66800135e-01 -2.97739916e-02
1.04232883e+00 -1.68670025e-02 6.41192198e-01 6.12582147e-01
1.18272312e-01 -8.59786570e-01 -1.50348499e-01 4.83465850e-01
7.23898172e-01 -1.19106007e+00 4.83728766e-01 -6.80471957e-01
-7.41467178e-01 1.14285135e+00 4.65464115e-01 1.90924272e-01
8.88680696e-01 2.68285483e-01 -4.45418358e-01 -7.59572387e-01
-5.17914534e-01 -3.36870134e-01 5.55146575e-01 5.76293528e-01
7.77544379e-01 1.47634861e-03 -7.47051895e-01 9.14261281e-01
2.36780807e-01 2.68020391e-01 2.26141945e-01 7.52366602e-01
-7.94079378e-02 -1.53328323e+00 1.12732381e-01 7.15623736e-01
-6.91078603e-01 -5.25331140e-01 -7.66237020e-01 8.75277221e-01
3.36367935e-01 9.30528343e-01 -6.33109629e-01 -6.15597069e-01
2.69005567e-01 3.09357077e-01 3.02116692e-01 -9.21434700e-01
-6.73771322e-01 -7.23936483e-02 5.02588332e-01 -2.10456803e-01
-6.04951262e-01 8.02152455e-02 -1.52288997e+00 -5.55798039e-02
-5.00281513e-01 4.31269109e-01 5.40212691e-01 1.16817009e+00
5.89604855e-01 1.09017622e+00 3.35370570e-01 -2.61665285e-01
-1.63924147e-03 -1.00074732e+00 -3.48541856e-01 3.13389957e-01
2.59990282e-02 -8.96019101e-01 1.19235508e-01 -1.41658977e-01] | [8.462570190429688, 8.652029037475586] |
c83e0854-6b59-406d-ba4d-d71e2fbebaff | unsupervised-domain-expansion-from-multiple | 2005.12544 | null | https://arxiv.org/abs/2005.12544v1 | https://arxiv.org/pdf/2005.12544v1.pdf | Unsupervised Domain Expansion from Multiple Sources | Given an existing system learned from previous source domains, it is desirable to adapt the system to new domains without accessing and forgetting all the previous domains in some applications. This problem is known as domain expansion. Unlike traditional domain adaptation in which the target domain is the domain defined by new data, in domain expansion the target domain is formed jointly by the source domains and the new domain (hence, domain expansion) and the label function to be learned must work for the expanded domain. Specifically, this paper presents a method for unsupervised multi-source domain expansion (UMSDE) where only the pre-learned models of the source domains and unlabelled new domain data are available. We propose to use the predicted class probability of the unlabelled data in the new domain produced by different source models to jointly mitigate the biases among domains, exploit the discriminative information in the new domain, and preserve the performance in the source domains. Experimental results on the VLCS, ImageCLEF_DA and PACS datasets have verified the effectiveness of the proposed method. | ['Lu Sheng', 'Jing Zhang', 'Philip Ogunbona', 'Chang Tang', 'Wanqing Li'] | 2020-05-26 | null | null | null | null | ['unsupervised-domain-expansion'] | ['methodology'] | [ 4.03636426e-01 2.50694543e-01 -3.74659717e-01 -5.98705530e-01
-4.17796820e-01 -7.49935508e-01 3.73216122e-01 -7.58392811e-02
-4.33715612e-01 1.40940881e+00 7.00692460e-02 3.13312978e-01
-3.46977413e-02 -6.80777133e-01 -5.09962618e-01 -7.62804925e-01
3.78957242e-01 1.04619098e+00 5.88053286e-01 -1.39486104e-01
8.97033066e-02 3.07155788e-01 -1.34581864e+00 2.72169173e-01
1.39073396e+00 6.66956544e-01 6.95840597e-01 1.20520428e-01
-4.21367824e-01 4.11479741e-01 -7.40588129e-01 -1.20466284e-01
3.45835805e-01 -5.95831633e-01 -9.39589918e-01 2.66140699e-01
2.83886254e-01 -2.72688687e-01 -2.60732144e-01 1.20749843e+00
3.26596051e-01 2.65746266e-01 5.37982047e-01 -1.12381828e+00
-9.18690085e-01 2.81067848e-01 -4.31275517e-01 8.62093046e-02
1.84565216e-01 -3.46058995e-01 4.06840742e-01 -7.60218263e-01
1.21885657e+00 1.15897143e+00 2.24091589e-01 7.56811142e-01
-1.13992143e+00 -8.14397514e-01 2.98523366e-01 3.72595042e-01
-1.21338654e+00 -2.50256538e-01 8.58825386e-01 -3.02348435e-01
5.95305741e-01 -3.79908115e-01 2.47280046e-01 9.66239274e-01
-9.54260975e-02 6.47760093e-01 1.21801531e+00 -7.03336298e-01
4.32832718e-01 8.97931933e-01 2.74011880e-01 2.20623195e-01
2.18469366e-01 6.95716292e-02 -3.43286574e-01 -4.29427117e-01
4.34544265e-01 -2.26944555e-02 -1.28158003e-01 -9.97485280e-01
-1.02989495e+00 6.73580468e-01 1.41413033e-01 5.24276972e-01
-5.09526432e-01 -7.93693900e-01 4.24066901e-01 4.82749522e-01
6.29913568e-01 2.64634401e-01 -1.05328178e+00 3.22451055e-01
-7.29991257e-01 2.71046221e-01 6.26095831e-01 1.37645483e+00
1.20761168e+00 -2.45802343e-01 2.74755448e-01 1.12775350e+00
2.51235217e-01 7.16853499e-01 1.00681126e+00 -7.37374008e-01
3.75386834e-01 1.03025055e+00 8.90092701e-02 -3.50044996e-01
-1.26956373e-01 -2.07263216e-01 -5.87763131e-01 3.28854173e-02
2.60365158e-01 -8.51434842e-02 -1.18751216e+00 1.91847050e+00
7.90553153e-01 4.50795982e-03 5.27591169e-01 4.48634982e-01
7.00788379e-01 5.85834503e-01 3.64886642e-01 -4.34955776e-01
8.71583879e-01 -9.07282710e-01 -5.96202791e-01 -4.97608244e-01
6.34360850e-01 -5.54773748e-01 7.13990271e-01 3.83377969e-01
-7.03641951e-01 -1.01015639e+00 -1.02045512e+00 1.05108514e-01
-5.07348478e-01 1.19909100e-01 1.25579283e-01 3.17066550e-01
-8.44891131e-01 2.62872308e-01 -2.02941865e-01 -7.97683656e-01
4.29562271e-01 4.27530736e-01 -6.29049420e-01 -5.47309458e-01
-1.43984556e+00 9.69362676e-01 1.29817939e+00 -6.91985130e-01
-1.01288927e+00 -6.40028119e-01 -5.85814476e-01 -4.65853624e-02
3.59350950e-01 -2.52840072e-01 1.13290715e+00 -1.41726804e+00
-1.14369094e+00 8.92056525e-01 -2.80498087e-01 -3.83849293e-01
1.00712739e-01 -6.45674989e-02 -6.94756746e-01 1.17254578e-01
3.47875923e-01 7.40978241e-01 7.95459867e-01 -1.44810867e+00
-1.12674093e+00 -6.42740667e-01 -2.30396256e-01 4.23030645e-01
-5.62479198e-01 -3.56111974e-01 -2.89363325e-01 -3.88868541e-01
1.91490725e-01 -9.66339111e-01 5.75698875e-02 -4.14035052e-01
3.16046566e-01 -4.37615097e-01 1.24587810e+00 -1.01550031e+00
1.03281164e+00 -2.35288739e+00 2.10150212e-01 2.27783933e-01
6.80231899e-02 5.95653892e-01 -3.09373796e-01 7.77980387e-02
-3.09326172e-01 -2.37711206e-01 -4.49697912e-01 1.90156698e-01
-2.10799605e-01 5.21381438e-01 -2.89144486e-01 9.76006091e-02
-6.83190152e-02 2.80109733e-01 -1.04767668e+00 -6.88465118e-01
1.05824441e-01 1.61405593e-01 -4.58644927e-01 3.26933444e-01
-3.83685648e-01 8.68966937e-01 -6.32112801e-01 4.58118647e-01
1.04108489e+00 -8.49654973e-02 5.71199059e-01 3.37322429e-02
3.75886232e-01 3.06551140e-02 -1.14470613e+00 1.91679263e+00
-5.28708696e-02 1.03297383e-01 -1.39807731e-01 -1.13193727e+00
1.31258035e+00 4.01482314e-01 6.57681704e-01 -9.47523713e-01
-3.55534256e-01 5.27779698e-01 -1.06477067e-01 -4.75505769e-01
3.41259509e-01 -3.86636615e-01 -1.16637029e-01 2.34927729e-01
6.72317207e-01 2.07928911e-01 2.14972243e-01 2.73960859e-01
7.89146721e-01 4.24248487e-01 7.59128451e-01 -1.95164289e-02
6.93633437e-01 4.81082946e-01 8.88376355e-01 4.86802906e-01
-4.38201606e-01 9.79865119e-02 1.16864279e-01 -1.69397727e-01
-1.34780502e+00 -1.19259405e+00 -1.69270530e-01 1.39405072e+00
2.47485727e-01 2.07692161e-01 -4.49073344e-01 -1.43517804e+00
6.24122433e-02 7.82128990e-01 -3.51261675e-01 -4.24889565e-01
-6.39451087e-01 -2.89107591e-01 2.94983327e-01 3.59247684e-01
9.54989433e-01 -1.00802445e+00 -3.95160392e-02 3.80988121e-01
-4.06601399e-01 -9.45207059e-01 -1.45803466e-01 2.00357258e-01
-1.19012535e+00 -1.08654523e+00 -1.02465808e+00 -1.31684673e+00
7.87283838e-01 1.06041558e-01 1.08715749e+00 -4.20221388e-01
3.49907130e-01 2.27099791e-01 -6.59892738e-01 -2.00762972e-01
-7.29534388e-01 1.89568058e-01 2.07929686e-01 -1.98385552e-01
9.11599815e-01 -5.53127229e-01 1.19009465e-02 3.81112099e-01
-1.13207352e+00 -2.87392467e-01 7.77071416e-01 1.05302656e+00
7.61441290e-01 4.94036645e-01 1.10114419e+00 -1.42397106e+00
3.54064196e-01 -8.15155447e-01 -4.16784912e-01 7.18725920e-01
-8.27108145e-01 7.91235790e-02 5.55890381e-01 -6.92602873e-01
-1.82971656e+00 2.78395116e-01 2.13587016e-01 -2.28532881e-01
-5.76901734e-01 4.54399705e-01 -7.35875368e-01 6.52275085e-02
9.28349793e-01 4.21912044e-01 -2.12138832e-01 -8.26951265e-01
4.13204134e-01 8.86299431e-01 6.63216949e-01 -6.64973199e-01
5.44319510e-01 3.11965436e-01 -5.58995366e-01 -5.38298011e-01
-6.00528061e-01 -7.25850642e-01 -1.13423240e+00 1.50241017e-01
5.35028696e-01 -9.96492684e-01 3.61565381e-01 3.97168428e-01
-1.16110599e+00 6.31993115e-02 -6.42532945e-01 5.32831252e-01
-3.20338905e-01 6.65383220e-01 1.35709822e-01 -5.31372666e-01
-3.35444137e-02 -7.76300550e-01 7.10688889e-01 6.05866790e-01
-1.14946075e-01 -1.15485597e+00 4.83043462e-01 1.07900739e-01
9.01651382e-02 -1.37048569e-02 1.26871324e+00 -1.40801823e+00
-2.02896684e-01 -9.73448902e-02 -1.87471852e-01 7.21456945e-01
3.77233058e-01 -7.91369855e-01 -9.21179414e-01 -4.93479431e-01
3.64093721e-01 -2.90393561e-01 7.04951346e-01 1.78329542e-01
5.61650217e-01 -1.17741905e-01 -7.18932986e-01 2.55713880e-01
1.54472816e+00 7.77066827e-01 5.86286247e-01 4.35389996e-01
4.97760326e-01 6.54029787e-01 1.00117242e+00 3.99850041e-01
2.28013486e-01 3.39947343e-01 -1.50664628e-01 1.90366656e-02
-2.43436828e-01 -2.52083123e-01 3.99391115e-01 5.91947258e-01
4.05572087e-01 -1.70191199e-01 -8.34950805e-01 9.28845406e-01
-1.67014694e+00 -6.30198121e-01 3.52978110e-01 2.32874870e+00
1.04679632e+00 -1.97224272e-03 4.38123792e-02 -3.99569005e-01
1.01064110e+00 -3.15927237e-01 -1.19392800e+00 -8.77054781e-02
-4.10328448e-01 4.23893005e-01 3.38644326e-01 3.56578618e-01
-1.05143297e+00 1.10007334e+00 6.19379425e+00 7.35278130e-01
-6.97403491e-01 3.16704392e-01 1.30197048e-01 1.85257897e-01
-1.66913584e-01 3.11959565e-01 -1.02690458e+00 4.45777744e-01
7.99661040e-01 -4.13800389e-01 1.94729105e-01 1.20605874e+00
-4.41576451e-01 -3.34876984e-01 -1.09259140e+00 5.74266195e-01
1.63092688e-01 -5.88364184e-01 7.62099549e-02 7.65778944e-02
9.63915944e-01 3.14212702e-02 -1.35359317e-01 8.25031161e-01
6.32651389e-01 -1.98411435e-01 1.38279414e-02 3.01013261e-01
9.51180398e-01 -5.94809651e-01 7.42413521e-01 5.69696903e-01
-8.67057920e-01 -1.66994080e-01 -6.87241793e-01 3.57406825e-01
-1.34220302e-01 5.49396276e-01 -1.32426584e+00 6.46424830e-01
6.98597193e-01 7.32893884e-01 -6.68638587e-01 8.97804558e-01
-2.15421602e-01 3.62285316e-01 -1.14350244e-01 5.02790332e-01
-2.78209746e-01 -1.20679840e-01 5.60006857e-01 7.75279641e-01
4.14524913e-01 5.04110418e-02 3.30861449e-01 5.25444031e-01
-1.19928397e-01 8.80524367e-02 -7.17633605e-01 -1.14308581e-01
7.41770208e-01 9.24738348e-01 -2.67923981e-01 -8.78729880e-01
-5.43363512e-01 1.16552317e+00 3.12782824e-01 6.07176185e-01
-6.27630651e-01 -4.08142447e-01 3.87680560e-01 -4.39506397e-02
3.11209887e-01 6.24705926e-02 -1.20718442e-01 -1.23157012e+00
1.07464753e-01 -1.02859664e+00 8.89339030e-01 -6.57368600e-01
-1.58488631e+00 6.37532115e-01 2.83828259e-01 -1.31888497e+00
-3.69408876e-01 -3.71574759e-01 1.24703869e-01 1.18612611e+00
-1.67544115e+00 -1.00012541e+00 -2.17628539e-01 1.01552320e+00
5.82266927e-01 -7.36635923e-01 9.79803085e-01 4.05183345e-01
4.00622338e-02 4.78108704e-01 8.24011207e-01 -1.14561114e-02
1.36673379e+00 -9.75074172e-01 -6.01645894e-02 6.70125425e-01
-2.28315428e-01 3.92389894e-01 5.22426963e-01 -1.25081551e+00
-4.87459689e-01 -1.03230393e+00 1.13174617e+00 -2.30421290e-01
2.41386920e-01 -8.34322348e-02 -1.36939728e+00 9.64828253e-01
1.59300715e-01 -3.77338678e-01 8.14879715e-01 1.19820703e-02
-4.81870443e-01 -2.07959086e-01 -1.73532057e+00 9.10187513e-03
8.74781370e-01 -2.29118899e-01 -1.17126787e+00 3.61569673e-01
6.70147777e-01 -2.22837627e-01 -7.70526886e-01 4.25744653e-01
2.49749899e-01 -4.96227711e-01 8.38139594e-01 -7.66600072e-01
1.40242726e-01 -2.15386361e-01 -3.16051543e-01 -1.64137506e+00
-5.37028611e-01 2.84405828e-01 -9.36130807e-03 1.51636493e+00
1.81447953e-01 -8.35802078e-01 8.23876381e-01 5.88029802e-01
6.63876254e-03 1.50708675e-01 -9.77393806e-01 -9.65188384e-01
2.68080086e-01 3.47490460e-01 8.61747801e-01 1.30544722e+00
-1.96285713e-02 4.86490279e-01 -2.18874365e-01 1.80710688e-01
6.53928161e-01 1.58126429e-01 7.46342659e-01 -1.61180317e+00
-6.78598881e-02 2.90656626e-01 -9.65365991e-02 -1.13323069e+00
3.54002535e-01 -9.93264437e-01 3.64254341e-02 -1.39752054e+00
3.94931823e-01 -7.49145627e-01 -5.62579274e-01 5.82769632e-01
-1.22078642e-01 -4.32243288e-01 8.95342454e-02 5.92603147e-01
-7.36647844e-01 6.03569746e-01 1.24907386e+00 -1.57131448e-01
-4.90011871e-01 -2.99403161e-01 -9.00505722e-01 7.03734636e-01
6.95640266e-01 -8.69526625e-01 -6.83649600e-01 -4.53627080e-01
-4.53156769e-01 -9.55992118e-02 -6.37111142e-02 -1.11893129e+00
2.36602962e-01 -2.80448198e-01 8.06812286e-01 -6.24123573e-01
1.77318770e-02 -1.22882771e+00 9.21622068e-02 2.53811479e-01
-3.78048420e-01 -4.21992093e-01 2.15878487e-01 7.01702833e-01
-3.70321721e-01 -4.70743775e-01 1.18091917e+00 -3.38317931e-01
-1.53998554e+00 1.52101114e-01 8.92165974e-02 1.57966688e-01
1.30640113e+00 -2.64697433e-01 -3.01640689e-01 5.74165583e-02
-1.14216256e+00 3.25430810e-01 6.28039896e-01 4.43103045e-01
6.08101189e-01 -1.52119565e+00 -8.18164468e-01 4.44481909e-01
4.38219517e-01 1.71253935e-01 4.34434921e-01 2.02107906e-01
-5.54785654e-02 3.43028367e-01 -6.99547768e-01 -5.53204179e-01
-1.28869188e+00 8.88891578e-01 4.67948131e-02 -6.92420065e-01
-2.16643885e-01 7.83748507e-01 5.49211919e-01 -9.50865388e-01
8.33845586e-02 2.12845474e-01 -4.94538814e-01 1.05942669e-03
3.45249176e-01 2.52674550e-01 6.10559918e-02 -6.10472262e-01
-3.58377457e-01 1.88441336e-01 -7.05706000e-01 -1.52640581e-01
1.30479300e+00 -5.29485762e-01 -1.70975596e-01 2.38897800e-01
1.04425991e+00 -2.83681780e-01 -1.08941150e+00 -1.08934867e+00
2.11396590e-01 -5.03199995e-01 -4.32198077e-01 -1.30995822e+00
-7.48639941e-01 6.59836352e-01 1.10553730e+00 -2.62653291e-01
1.59129822e+00 2.07961485e-01 7.44422495e-01 3.30948979e-01
6.06516004e-01 -1.63920295e+00 8.41058209e-04 6.92512393e-01
4.99658763e-01 -1.20095515e+00 1.04877250e-02 -1.38692766e-01
-9.92484450e-01 8.81466985e-01 1.02903926e+00 -2.64894795e-02
5.98677635e-01 -3.30026418e-01 -1.27143309e-01 1.55899256e-01
-4.30898547e-01 -1.31723136e-01 2.16512512e-02 1.14272606e+00
1.24464102e-01 -3.47016118e-02 -2.80599266e-01 7.23358750e-01
1.83741897e-01 4.46457416e-01 2.06152916e-01 1.04420793e+00
-6.98358595e-01 -1.64980376e+00 -5.67936778e-01 3.02062005e-01
-1.53931249e-02 8.28732997e-02 -6.51814342e-01 7.98543215e-01
7.38290429e-01 7.46523857e-01 -1.62546307e-01 -3.19485784e-01
4.32811677e-01 6.67739987e-01 3.91285270e-01 -1.02199256e+00
-1.43688604e-01 2.07525030e-01 -1.92352459e-01 -4.88661714e-02
-6.25545084e-01 -5.85324526e-01 -1.34426522e+00 1.26633689e-01
-3.77165914e-01 4.99729693e-01 4.32401955e-01 9.12965894e-01
5.41919410e-01 2.92491466e-01 5.56949019e-01 -1.33147255e-01
-6.15880549e-01 -1.13615727e+00 -9.50915098e-01 8.08431864e-01
1.01414196e-01 -8.74423563e-01 -9.86220539e-02 5.29029608e-01] | [10.364853858947754, 3.095060348510742] |
7dd4b2af-321b-424f-98d8-57c83a100674 | disentangled-representations-for-domain | 2008.11514 | null | https://arxiv.org/abs/2008.11514v1 | https://arxiv.org/pdf/2008.11514v1.pdf | Disentangled Representations for Domain-generalized Cardiac Segmentation | Robust cardiac image segmentation is still an open challenge due to the inability of the existing methods to achieve satisfactory performance on unseen data of different domains. Since the acquisition and annotation of medical data are costly and time-consuming, recent work focuses on domain adaptation and generalization to bridge the gap between data from different populations and scanners. In this paper, we propose two data augmentation methods that focus on improving the domain adaptation and generalization abilities of state-to-the-art cardiac segmentation models. In particular, our "Resolution Augmentation" method generates more diverse data by rescaling images to different resolutions within a range spanning different scanner protocols. Subsequently, our "Factor-based Augmentation" method generates more diverse data by projecting the original samples onto disentangled latent spaces, and combining the learned anatomy and modality factors from different domains. Our extensive experiments demonstrate the importance of efficient adaptation between seen and unseen domains, as well as model generalization ability, to robust cardiac image segmentation. | ["Alison O'Neil", 'Spyridon Thermos', 'Xiao Liu', 'Agisilaos Chartsias', 'Sotirios A. Tsaftaris'] | 2020-08-26 | null | null | null | null | ['cardiac-segmentation'] | ['medical'] | [ 5.04537940e-01 2.92250644e-02 -8.42569694e-02 -5.90297401e-01
-9.83018219e-01 -6.23096764e-01 2.99700171e-01 -2.62991227e-02
-3.83851945e-01 7.05213964e-01 3.20340425e-01 -1.06451930e-02
-9.98463109e-03 -3.82298261e-01 -3.89872849e-01 -7.82203436e-01
1.62807286e-01 7.45818019e-01 1.88951969e-01 1.43056557e-01
-2.00849652e-01 4.53640312e-01 -9.52573419e-01 2.81392429e-02
1.05572522e+00 6.38063073e-01 -3.83614339e-02 7.18983233e-01
2.56798100e-02 5.53856231e-02 -3.94209713e-01 -1.46851465e-01
3.62161934e-01 -7.20812559e-01 -7.52529562e-01 3.59475940e-01
5.44832587e-01 -4.29277480e-01 -4.04008925e-01 8.85937333e-01
7.24893808e-01 4.05166857e-02 6.56889379e-01 -9.93400872e-01
-9.09448206e-01 5.28634429e-01 -6.33409739e-01 5.32599211e-01
-1.74072817e-01 1.26503840e-01 4.18366224e-01 -4.76959348e-01
7.23342419e-01 8.76917601e-01 5.50127625e-01 1.02484190e+00
-1.79259408e+00 -8.36290598e-01 9.11666751e-02 -2.24856034e-01
-1.06727302e+00 -2.38604039e-01 9.30348516e-01 -8.34313393e-01
2.17296168e-01 1.42177064e-02 4.54792857e-01 1.52665162e+00
-2.21594736e-01 4.12051171e-01 1.26924336e+00 -2.60468781e-01
1.82822958e-01 2.17874631e-01 1.15754180e-01 5.83361387e-01
4.90711570e-01 2.63029169e-02 -4.67569739e-01 -3.22047919e-01
1.22427583e+00 2.31946670e-02 -3.33043337e-01 -7.72766650e-01
-1.75241184e+00 8.44126463e-01 1.90988719e-01 3.98603052e-01
-4.36737210e-01 -2.54256755e-01 4.53546554e-01 -1.30989283e-01
4.19644415e-01 5.95406413e-01 -4.22849000e-01 8.15751478e-02
-1.12279725e+00 3.97349559e-02 4.88204628e-01 8.30568433e-01
4.10877496e-01 5.74436970e-02 -2.01312050e-01 9.15665925e-01
1.29884854e-01 5.10344923e-01 7.44319797e-01 -9.13861930e-01
3.85608554e-01 5.39456069e-01 -2.07370758e-01 -6.56649888e-01
-6.11585915e-01 -6.26406550e-01 -1.14377236e+00 -6.06101304e-02
6.57865107e-01 -4.09000278e-01 -1.32550836e+00 2.11984086e+00
3.60560626e-01 4.34041440e-01 -6.26723794e-03 9.59760666e-01
8.18105638e-01 -3.15539166e-02 5.96395016e-01 -6.64284527e-02
1.45946908e+00 -6.91197336e-01 -7.30546296e-01 -3.95072788e-01
4.68504876e-01 -6.79160774e-01 1.13566327e+00 3.10196728e-02
-8.99837375e-01 -6.75176382e-01 -1.03181303e+00 -4.18376662e-02
-1.72620937e-02 1.19018182e-01 6.42123103e-01 6.86797321e-01
-6.93138540e-01 3.29681516e-01 -1.21473122e+00 -3.12377989e-01
8.66813242e-01 3.51088613e-01 -6.27282739e-01 -8.99194330e-02
-1.03320074e+00 6.39295518e-01 4.08900589e-01 -2.26321802e-01
-7.80582845e-01 -1.19781983e+00 -8.37431431e-01 -2.78542995e-01
3.09898585e-01 -1.12562799e+00 8.66843522e-01 -6.25088513e-01
-1.21698892e+00 1.20282614e+00 5.53312488e-02 -3.62868607e-01
5.62205255e-01 -1.44120520e-02 -4.37750906e-01 4.06408042e-01
2.54746825e-01 7.08130360e-01 9.24311697e-01 -1.26790643e+00
-2.26866931e-01 -7.05230713e-01 -5.03893912e-01 1.29804552e-01
-3.71038139e-01 -2.96690017e-01 -3.90949577e-01 -1.00763118e+00
5.23554087e-01 -1.14647889e+00 -6.19633198e-01 -4.77303676e-02
-3.05374861e-01 3.21628213e-01 5.74765384e-01 -8.81370723e-01
9.96065617e-01 -2.46254420e+00 4.01415408e-01 1.48403555e-01
6.63021922e-01 1.71171755e-01 -1.90002561e-01 -3.32545251e-01
-2.26473033e-01 3.07321221e-01 -5.72509587e-01 -2.80279636e-01
-2.39601016e-01 4.34674054e-01 -5.58755314e-03 4.59651172e-01
4.05271620e-01 8.85589600e-01 -7.78619170e-01 -8.98870409e-01
2.22605631e-01 4.78946954e-01 -7.00558662e-01 3.24208319e-01
8.82735401e-02 1.44727051e+00 -6.82635605e-01 4.69660580e-01
6.87099099e-01 -5.05156517e-01 1.78683296e-01 -3.63557726e-01
3.91699046e-01 -1.97829455e-01 -9.43275869e-01 2.24220490e+00
-3.46977934e-02 2.25175053e-01 1.77887939e-02 -1.08557332e+00
8.15452337e-01 5.10814309e-01 8.98087323e-01 -5.04256248e-01
1.73829973e-01 2.75148600e-01 5.96477129e-02 -6.14085793e-01
2.02674374e-01 -6.95198357e-01 -1.44879371e-01 3.69810820e-01
3.70522261e-01 -1.02196388e-01 1.07125722e-01 1.01217821e-01
8.09923649e-01 5.13383150e-02 9.92776155e-02 -1.93650588e-01
3.87251168e-01 3.11635099e-02 8.09518695e-01 7.23625600e-01
-6.13372862e-01 9.39662158e-01 4.29447532e-01 -2.39360094e-01
-1.31086421e+00 -1.34460211e+00 -4.36717153e-01 8.65361035e-01
-1.51941136e-01 8.25775787e-02 -9.35955405e-01 -8.65285993e-01
-1.14430927e-01 3.15627813e-01 -9.35259104e-01 -2.97662407e-01
-7.65671074e-01 -1.12285042e+00 8.26401234e-01 9.62928355e-01
4.80147600e-01 -7.25243330e-01 -4.24036801e-01 1.89631924e-01
-4.54581439e-01 -1.55844474e+00 -4.56789166e-01 -8.84310827e-02
-1.31130600e+00 -1.01663363e+00 -1.01316500e+00 -7.61035383e-01
7.90870905e-01 -2.17814058e-01 1.19627512e+00 -1.97893471e-01
-5.08364201e-01 3.89929235e-01 -2.31473207e-01 -2.70599157e-01
-5.76003850e-01 3.21395844e-01 6.99012429e-02 7.37378001e-02
2.33117700e-01 -6.67327404e-01 -7.14577734e-01 3.18985343e-01
-1.04223144e+00 6.07490540e-02 7.48396575e-01 9.51678574e-01
8.76955152e-01 -4.21012342e-01 5.24998844e-01 -1.31410909e+00
5.05390584e-01 -3.84261996e-01 -3.17937464e-01 1.65644839e-01
-5.50404787e-01 2.44760334e-01 2.48684525e-01 -8.26739430e-01
-1.13653064e+00 2.34946966e-01 7.77469352e-02 -6.16504967e-01
-5.11248469e-01 2.52136588e-01 5.60502894e-02 -1.70791131e-02
9.64906693e-01 1.63959578e-01 2.55669862e-01 -5.81879437e-01
4.77837712e-01 2.39720210e-01 8.42019439e-01 -8.01819503e-01
9.20119822e-01 7.60399640e-01 1.99433595e-01 -5.97357810e-01
-7.38815546e-01 -3.72719198e-01 -1.18675387e+00 1.40852436e-01
1.27560806e+00 -8.38737965e-01 -2.41540540e-02 2.95148462e-01
-6.03697062e-01 -1.89145163e-01 -5.71479261e-01 7.85240710e-01
-5.44621050e-01 5.86324155e-01 -6.07064486e-01 -1.59435868e-01
-2.77948081e-01 -1.33010423e+00 1.00467360e+00 3.22782874e-01
-2.41317824e-01 -1.10593462e+00 2.97701657e-01 6.15889609e-01
4.84816492e-01 6.84503198e-01 1.16774321e+00 -8.58759165e-01
-3.71730834e-01 -1.14306889e-01 -2.49515235e-01 4.59113210e-01
2.51241416e-01 -3.84301007e-01 -8.60733211e-01 -3.07637453e-01
2.94323657e-02 -2.34868407e-01 7.04680979e-01 6.38300657e-01
1.13556802e+00 2.08582863e-01 -3.54266912e-01 8.72170508e-01
1.12893665e+00 -7.26149380e-02 3.59242469e-01 5.01756556e-02
8.25582445e-01 4.99141246e-01 2.64770299e-01 2.19308525e-01
2.67420679e-01 3.87216538e-01 -8.80432799e-02 -6.49649501e-01
-3.42021942e-01 -1.02460943e-01 -3.88056636e-01 6.96758628e-01
-2.28786975e-01 3.65937293e-01 -1.09812546e+00 5.73391616e-01
-1.57567382e+00 -6.37597203e-01 7.34256282e-02 2.11228299e+00
8.71086776e-01 -1.11554479e-02 2.32212812e-01 -1.59513175e-01
6.85078263e-01 3.91684566e-03 -8.43940079e-01 1.52596533e-01
-5.62645420e-02 3.95813704e-01 4.11063671e-01 1.71756625e-01
-1.27891755e+00 7.46844649e-01 7.08883810e+00 2.77541041e-01
-1.12532473e+00 3.30884457e-01 7.08284974e-01 1.79163720e-02
-7.62178749e-02 -1.74170226e-01 -3.87165099e-01 1.79574147e-01
8.68785977e-01 -4.32460867e-02 2.37423018e-01 6.78135753e-01
-4.74129394e-02 2.56469011e-01 -1.17840457e+00 9.69978392e-01
1.95845306e-01 -1.15243769e+00 -1.18350210e-02 1.09282702e-01
7.55018890e-01 -5.33746742e-02 3.12382549e-01 3.05016041e-01
-1.24312947e-02 -1.00565255e+00 8.02157670e-02 4.75328535e-01
1.00160587e+00 -2.54175842e-01 6.20862067e-01 8.90621692e-02
-8.40549111e-01 2.38595769e-01 -6.81494325e-02 5.44362783e-01
1.86838284e-01 3.95990580e-01 -9.11372006e-01 4.86067623e-01
6.05206430e-01 4.55146164e-01 -7.28590965e-01 7.07186997e-01
-1.18361181e-02 6.46080732e-01 -2.00804010e-01 6.69949472e-01
-9.21492800e-02 -2.45359555e-01 5.35845399e-01 1.14940310e+00
-7.07921281e-04 1.77974999e-01 3.97673786e-01 1.06069076e+00
-1.01569228e-01 7.02804998e-02 -3.35343599e-01 -2.46857911e-01
2.43092820e-01 1.31782293e+00 -8.09052527e-01 -4.43180531e-01
-5.27236044e-01 8.50030780e-01 1.24071769e-01 5.43481410e-01
-8.97127748e-01 1.07914008e-01 5.57303309e-01 2.28208154e-01
-5.34875505e-03 -3.16035241e-01 -5.92216671e-01 -1.42237961e+00
-1.41499445e-01 -8.93081605e-01 7.45536864e-01 -3.31615537e-01
-1.67218113e+00 6.63442552e-01 2.00914741e-01 -1.08792102e+00
-1.17328517e-01 -4.47094709e-01 -1.57318346e-03 1.09740758e+00
-1.42281473e+00 -1.27907825e+00 -4.71363962e-01 7.87539482e-01
3.61172855e-01 -2.02271044e-01 1.04136693e+00 5.09346306e-01
-5.55623889e-01 6.20158732e-01 -9.81027558e-02 5.20611584e-01
9.67718363e-01 -1.23918784e+00 1.64507300e-01 7.79444575e-01
-3.81600601e-03 7.64628708e-01 4.37626719e-01 -6.22879982e-01
-8.54652405e-01 -9.54543412e-01 1.85324386e-01 -5.92833400e-01
5.48602700e-01 -3.25125873e-01 -1.27044761e+00 8.40961099e-01
-1.53602764e-01 4.82845426e-01 1.27458477e+00 2.84957170e-01
-4.44195539e-01 1.53404564e-01 -1.26052225e+00 5.56619346e-01
8.47305775e-01 -4.97859120e-01 -6.53188586e-01 1.59335703e-01
5.82437158e-01 -6.92015171e-01 -1.41401815e+00 6.26075983e-01
4.06235814e-01 -5.53023458e-01 1.19295096e+00 -1.03347623e+00
2.90656269e-01 -1.19063221e-01 -7.34996721e-02 -1.10606062e+00
-4.89693314e-01 -3.74284804e-01 4.19729240e-02 1.10497737e+00
4.66043621e-01 -5.91595948e-01 9.78957176e-01 8.97560179e-01
3.52652110e-02 -4.60814834e-01 -9.42499042e-01 -3.39500070e-01
5.03199100e-01 -1.44501686e-01 4.87764090e-01 1.28421056e+00
-3.44101101e-01 4.49636906e-01 -3.62144232e-01 3.91161412e-01
8.43054473e-01 2.07992747e-01 5.86741090e-01 -1.43469083e+00
-3.62015784e-01 -2.52170324e-01 -3.97323847e-01 -4.53717262e-01
-1.91337839e-02 -9.62311685e-01 -1.82251826e-01 -1.25067687e+00
5.72702289e-01 -5.14956534e-01 -4.97100741e-01 3.60348850e-01
-4.90757525e-01 4.62326527e-01 6.16592914e-02 3.05080593e-01
-2.55018353e-01 2.88303763e-01 1.66490877e+00 1.48316503e-01
-3.21895212e-01 -1.48925930e-01 -9.46931005e-01 7.39480734e-01
7.59668887e-01 -5.05083799e-01 -5.77302933e-01 -4.58782077e-01
-3.26966077e-01 1.98699832e-01 4.02537644e-01 -9.61392879e-01
-9.78421271e-02 -2.66050138e-02 6.51166499e-01 -2.11790696e-01
1.93664968e-01 -6.91476464e-01 -7.37274298e-03 3.78517896e-01
-4.13740009e-01 -3.82757671e-02 3.91311735e-01 6.36317551e-01
-8.81422013e-02 1.15972064e-01 1.09908116e+00 -2.24462569e-01
-3.69253486e-01 5.53690970e-01 -2.40861140e-02 6.17641807e-01
8.72763336e-01 -1.24855652e-01 -5.68219610e-02 8.18696693e-02
-1.40513253e+00 1.45925850e-01 2.82233208e-01 4.02675509e-01
4.23532754e-01 -1.11095643e+00 -1.00626624e+00 4.75574613e-01
1.48616597e-01 1.68847218e-01 6.92771018e-01 1.03445041e+00
-3.03579539e-01 2.25022465e-01 -4.44403350e-01 -9.83160377e-01
-1.19445717e+00 6.18450046e-01 4.10348296e-01 -4.84267652e-01
-7.35649288e-01 5.41588187e-01 3.45554382e-01 -6.18181229e-01
-9.36393365e-02 -2.94677466e-01 -1.56653613e-01 -9.41432267e-02
2.04547822e-01 1.26993343e-01 -8.25465471e-02 -5.95586002e-01
-3.26298535e-01 7.46054113e-01 -2.30436951e-01 -1.44977629e-01
1.21233523e+00 -4.25145738e-02 2.33372346e-01 1.40283197e-01
9.33159888e-01 -1.45379588e-01 -1.26231086e+00 -4.44509029e-01
-1.42843083e-01 -3.68413240e-01 -4.30258736e-02 -9.36353445e-01
-1.12438416e+00 9.81734693e-01 1.01146030e+00 -1.00652859e-01
1.24153817e+00 1.17992580e-01 7.18999803e-01 -1.80643752e-01
2.14143749e-03 -9.54465985e-01 9.66210738e-02 3.21103114e-04
5.69579363e-01 -1.45960605e+00 2.41907407e-03 -5.98357916e-01
-9.57606435e-01 7.64506996e-01 5.66235125e-01 -4.56385463e-02
6.00885630e-01 2.37509087e-01 3.67527336e-01 -1.53441444e-01
-1.97634429e-01 -5.03732041e-02 5.51946223e-01 8.42944741e-01
5.00311732e-01 1.69720322e-01 -2.90669203e-01 7.49075770e-01
-3.37090865e-02 1.09202273e-01 2.81575114e-01 7.68094659e-01
7.35134259e-02 -1.29364681e+00 -3.63044202e-01 4.27620530e-01
-6.15159333e-01 1.01340659e-01 -4.02184613e-02 9.37975228e-01
1.65467471e-01 4.55260456e-01 -1.74978867e-01 -7.76245147e-02
4.52829808e-01 2.78370440e-01 5.65827370e-01 -8.25550973e-01
-2.19629839e-01 3.18100810e-01 -3.53582978e-01 -3.44083965e-01
-5.65164626e-01 -8.29299629e-01 -1.33077157e+00 2.25133717e-01
1.14095043e-02 -1.89864174e-01 5.24159908e-01 9.47672427e-01
4.58649307e-01 8.61193180e-01 2.39546746e-01 -5.43333411e-01
-6.49686098e-01 -9.66264367e-01 -4.95804042e-01 1.08518720e+00
2.72244871e-01 -7.16809630e-01 -5.28462939e-02 4.34354097e-01] | [14.606988906860352, -2.0287351608276367] |
f3c0e481-3393-45c4-a1c4-168b3da6089a | nearest-neighbor-based-out-of-distribution | 2303.16616 | null | https://arxiv.org/abs/2303.16616v1 | https://arxiv.org/pdf/2303.16616v1.pdf | Nearest Neighbor Based Out-of-Distribution Detection in Remote Sensing Scene Classification | Deep learning models for image classification are typically trained under the "closed-world" assumption with a predefined set of image classes. However, when the models are deployed they may be faced with input images not belonging to the classes encountered during training. This type of scenario is common in remote sensing image classification where images come from different geographic areas, sensors, and imaging conditions. In this paper we deal with the problem of detecting remote sensing images coming from a different distribution compared to the training data - out of distribution images. We propose a benchmark for out of distribution detection in remote sensing scene classification and evaluate detectors based on maximum softmax probability and nearest neighbors. The experimental results show convincing advantages of the method based on nearest neighbors. | ['Vladimir Risojević', 'Mitar Simić', 'Dajana Dimitrić'] | 2023-03-29 | null | null | null | null | ['scene-classification', 'remote-sensing-image-classification'] | ['computer-vision', 'miscellaneous'] | [ 6.11131430e-01 -2.87189484e-01 -4.78426218e-02 -7.10941315e-01
-4.62366790e-01 -5.12962401e-01 6.58987343e-01 2.66006887e-01
-7.59293139e-01 5.08529902e-01 -4.09667641e-01 -3.92535388e-01
-4.68748629e-01 -1.26861048e+00 -7.18812704e-01 -8.92059624e-01
-1.67087093e-01 5.87742746e-01 -1.57791659e-01 1.09255306e-01
8.20439234e-02 8.43976080e-01 -1.59091806e+00 2.41738737e-01
3.75413120e-01 9.50142145e-01 5.12801290e-01 7.27508545e-01
-1.23632746e-02 5.87744772e-01 -8.72229457e-01 2.56393433e-01
8.32159340e-01 -6.59180209e-02 -5.50038457e-01 4.32484001e-01
7.78090477e-01 -5.44623792e-01 -2.28570744e-01 1.53760803e+00
5.55143595e-01 2.62801647e-01 8.73360276e-01 -8.85645986e-01
-6.43361926e-01 5.49700558e-01 -6.09765828e-01 6.14051998e-01
-1.36843711e-01 -4.46742633e-03 5.96208572e-01 -7.03645468e-01
3.52188885e-01 1.03597057e+00 5.54192245e-01 3.02984208e-01
-1.09384799e+00 -4.67140466e-01 4.49763596e-01 2.37185508e-01
-1.80733311e+00 -1.72392830e-01 4.01906401e-01 -6.96073771e-01
7.62463808e-01 1.80525124e-01 1.99799702e-01 7.83344209e-01
-1.63778588e-01 5.21068096e-01 1.03972948e+00 -2.92768538e-01
3.63430232e-01 5.27394712e-01 2.96249717e-01 5.72192250e-03
1.26278028e-01 3.13224226e-01 1.75576568e-01 3.38094532e-02
4.42013025e-01 4.10413772e-01 -3.42310160e-01 -1.11780308e-01
-9.34535980e-01 8.93521667e-01 7.78320670e-01 5.45155108e-01
-6.83850825e-01 -1.38672367e-01 -6.07200153e-02 3.65126729e-01
5.36477745e-01 5.65044396e-02 -4.56221074e-01 7.34600186e-01
-1.13442743e+00 -3.24821882e-02 5.88598609e-01 5.77554107e-01
8.79100382e-01 -5.61602637e-02 1.22191176e-01 7.67724037e-01
1.45131424e-01 8.17398906e-01 5.90923011e-01 -1.46445498e-01
3.22223991e-01 4.17588145e-01 1.84656322e-01 -1.02815759e+00
-3.27114046e-01 -7.42594302e-01 -1.01055181e+00 6.40224814e-01
4.35222238e-01 -8.99364725e-02 -1.12335050e+00 1.25055444e+00
2.12080061e-01 1.23160556e-01 3.18542838e-01 9.42373455e-01
7.49992371e-01 9.66415286e-01 2.74425931e-02 1.06060401e-01
8.19027066e-01 -3.29246700e-01 -1.15534566e-01 -3.41051042e-01
1.78202748e-01 -4.11569148e-01 6.88604891e-01 5.06522357e-01
-1.40285432e-01 -7.51682162e-01 -9.35213089e-01 6.17248535e-01
-7.47040629e-01 5.12647271e-01 1.55790165e-01 6.79779708e-01
-8.93912256e-01 5.21556675e-01 -3.15914720e-01 -6.08425617e-01
4.86729532e-01 1.55180261e-01 -3.79799306e-01 1.07014440e-01
-7.29788542e-01 7.78397739e-01 7.06569612e-01 3.39004636e-01
-1.24775600e+00 -2.97399223e-01 -5.10295570e-01 1.61049366e-01
9.35465023e-02 -1.04170188e-01 9.55915987e-01 -1.46626043e+00
-7.31274188e-01 1.11482608e+00 2.49746650e-01 -4.46533889e-01
7.58385181e-01 -1.00465856e-01 -6.12878203e-01 -2.53824741e-01
-1.50596360e-02 4.39839959e-01 9.04097557e-01 -1.11112309e+00
-7.80653536e-01 -4.93707687e-01 6.60383627e-02 1.01125330e-01
-3.81650850e-02 -2.25908402e-02 2.91292429e-01 -2.62889504e-01
2.39177704e-01 -5.00475049e-01 -3.74116927e-01 1.84535645e-02
-5.87062657e-01 4.61740941e-02 1.02221262e+00 -3.80001336e-01
5.02232492e-01 -2.20114613e+00 -4.29800361e-01 6.29947305e-01
-1.06969215e-01 4.11980420e-01 -1.90464541e-01 3.07292372e-01
-3.67607266e-01 2.14629233e-01 -1.61146000e-01 2.77055293e-01
-8.85799993e-03 1.82183623e-01 -6.40356541e-01 7.58302689e-01
4.40438390e-02 4.35567647e-02 -7.19070792e-01 -2.25618929e-01
5.29289901e-01 3.90468150e-01 1.86028600e-01 2.15838596e-01
-2.25894317e-01 5.50049424e-01 -4.44480330e-01 4.79545921e-01
9.91606414e-01 -1.68068215e-01 -1.57985389e-02 -3.90020087e-02
3.75658274e-02 -3.07613879e-01 -1.56011856e+00 1.01426089e+00
-3.12123895e-01 6.84271693e-01 -2.63470918e-01 -1.20389664e+00
1.10838032e+00 3.45802866e-02 2.13310793e-01 -5.08784890e-01
2.73887038e-01 -4.27799970e-02 1.42575626e-03 -6.91408157e-01
4.67171818e-01 -1.14977807e-01 1.97140113e-01 1.46985635e-01
-1.13705970e-01 9.78263095e-02 6.48221374e-02 -4.35279727e-01
6.95809722e-01 -1.64014146e-01 5.13370931e-01 -1.96023509e-01
2.52149969e-01 7.21978694e-02 2.75894731e-01 1.43585861e+00
-1.97126791e-01 5.18624902e-01 -1.47893280e-01 -8.56099784e-01
-1.00272667e+00 -1.22835040e+00 -6.90682113e-01 9.23619330e-01
1.31009147e-01 5.74104786e-01 -3.45064700e-01 -6.65300965e-01
-8.96889269e-02 4.81570482e-01 -5.05770445e-01 3.03364575e-01
8.10923148e-03 -1.13403225e+00 4.31180060e-01 2.57397622e-01
6.30487382e-01 -9.73442972e-01 -1.00051188e+00 -2.25441698e-02
1.39575511e-01 -8.77949834e-01 1.86964720e-01 2.79205501e-01
-6.62337005e-01 -1.25461364e+00 -8.28460395e-01 -6.34905636e-01
6.25703573e-01 3.55298012e-01 1.03416729e+00 -1.50604621e-01
-4.49390084e-01 1.10595673e-01 -3.34726721e-01 -6.51280165e-01
-4.71096605e-01 -4.68966290e-02 -1.38140514e-01 3.78887922e-01
6.74737930e-01 -5.27375042e-01 -6.77472591e-01 1.41182974e-01
-1.17498326e+00 -5.37772834e-01 7.27680922e-01 5.60815990e-01
5.00776470e-01 7.96082675e-01 2.82054633e-01 -9.39503193e-01
1.75220758e-01 -9.15312469e-01 -8.57818842e-01 3.85302246e-01
-2.91966856e-01 -2.91829228e-01 5.61666846e-01 -6.33513987e-01
-8.07746828e-01 4.27279979e-01 -4.28738706e-02 -3.30963880e-01
-1.01301122e+00 6.51714385e-01 -2.25954652e-01 1.62992254e-01
1.16305804e+00 3.52121115e-01 -3.98964345e-01 -5.27002871e-01
3.20526995e-02 1.09867787e+00 4.53819335e-01 -1.60345212e-01
8.07769477e-01 6.75836146e-01 -3.78538370e-01 -1.23418248e+00
-7.92125821e-01 -7.55692542e-01 -5.89105248e-01 -1.41471043e-01
8.12985420e-01 -1.09838903e+00 -1.62306562e-01 6.87940300e-01
-9.03243780e-01 -3.02241892e-01 -2.37901732e-01 7.85151303e-01
-6.98150415e-03 1.69631004e-01 -1.12950847e-01 -1.19357741e+00
-1.41129717e-01 -7.95877278e-01 8.79229903e-01 4.26682472e-01
3.03769976e-01 -9.17947888e-01 -7.72869820e-03 -2.27242246e-01
4.48661655e-01 3.57934624e-01 5.53357422e-01 -9.53933179e-01
-3.61463517e-01 -4.46673065e-01 -2.95439214e-01 7.47165501e-01
2.24013567e-01 5.19999862e-02 -1.31759310e+00 -3.22907358e-01
-1.02154516e-01 -1.46114707e-01 9.86594379e-01 5.35885990e-01
1.15798306e+00 -2.44675547e-01 -3.90419751e-01 4.35619742e-01
2.05010533e+00 3.01701754e-01 5.66427588e-01 2.77000219e-01
3.93343240e-01 7.40120292e-01 3.60330701e-01 5.43160081e-01
-3.82433683e-02 3.20712209e-01 7.19021738e-01 -4.90286201e-01
3.95648360e-01 -6.49135709e-02 4.40382734e-02 -4.03937936e-01
3.51848513e-01 -6.82852447e-01 -1.07676148e+00 8.10062647e-01
-1.66841888e+00 -1.27882457e+00 -7.74551257e-02 2.37131429e+00
2.47426048e-01 -1.14669882e-01 -8.22674334e-02 1.19259536e-01
1.12635779e+00 7.15438649e-02 -7.06200004e-01 6.76082969e-02
-4.20611322e-01 1.25087485e-01 9.24618542e-01 3.78170103e-01
-1.70845592e+00 4.22179818e-01 6.37472677e+00 7.31083453e-01
-1.57999349e+00 3.34492396e-03 8.19901109e-01 2.01040134e-01
2.06978917e-01 -1.23158529e-01 -8.27909172e-01 3.65889579e-01
4.48018670e-01 2.30681106e-01 -2.31789853e-02 9.20430660e-01
1.83644384e-01 -4.00582433e-01 -9.63818789e-01 1.07343137e+00
-5.48225157e-02 -9.45129395e-01 8.74732658e-02 1.98204797e-02
6.61814213e-01 6.88386261e-01 1.94923535e-01 6.11368641e-02
4.70853984e-01 -1.12419748e+00 8.18746448e-01 6.20879173e-01
5.19107461e-01 -4.08766687e-01 7.65867591e-01 9.57838416e-01
-8.56740415e-01 -3.56241465e-01 -7.69335866e-01 -2.22403064e-01
-2.17341915e-01 8.37379694e-01 -1.00261652e+00 2.99633682e-01
9.98424768e-01 6.35894060e-01 -4.77538168e-01 1.30558646e+00
-6.12212345e-02 5.56592047e-01 -6.22615755e-01 -6.16956241e-02
2.78423488e-01 -2.79562980e-01 4.78313357e-01 1.05533886e+00
5.17433107e-01 -1.31313846e-01 5.67610562e-01 1.03535473e+00
2.32780725e-01 -1.07696764e-01 -1.05239677e+00 1.52848870e-01
4.07138258e-01 1.14653528e+00 -7.77684629e-01 -3.84119064e-01
-1.37147367e-01 6.67661130e-01 -1.96750358e-01 5.76421440e-01
-6.82621598e-01 -3.36382002e-01 3.87787133e-01 1.26917839e-01
3.75963509e-01 -6.02959804e-02 1.95153356e-01 -9.66809511e-01
7.87012652e-02 -4.92662638e-01 5.33586144e-01 -7.69936562e-01
-1.57737339e+00 6.90800905e-01 2.48284280e-01 -1.34727597e+00
-9.18855367e-04 -8.32837343e-01 -6.60187900e-01 1.00373721e+00
-1.77518606e+00 -1.02983510e+00 -4.19717640e-01 6.37460530e-01
1.83784649e-01 -1.72838330e-01 7.39041448e-01 4.57046866e-01
-2.10029393e-01 1.35384902e-01 6.14628553e-01 4.60927278e-01
1.36180356e-01 -1.13485944e+00 -3.13264057e-02 1.05329084e+00
4.48649138e-01 3.30526428e-03 6.93170190e-01 -2.64547110e-01
-6.40624821e-01 -1.49131620e+00 7.04056025e-01 1.59720797e-03
4.15009022e-01 -3.14444639e-02 -7.96446681e-01 7.12524533e-01
-7.04268739e-02 1.35900632e-01 6.52039111e-01 -1.72963485e-01
-4.31948721e-01 -2.95788974e-01 -1.48339760e+00 -7.18021765e-02
5.10669053e-01 -5.12131989e-01 -4.65409368e-01 7.02865779e-01
-2.01157406e-01 -1.01254210e-01 -4.81563389e-01 2.96337247e-01
2.43938521e-01 -8.46118093e-01 1.00111032e+00 -5.56629777e-01
1.01153515e-01 -6.38894200e-01 -6.57127380e-01 -1.33091950e+00
-7.22801238e-02 1.71670958e-01 8.87893796e-01 1.05854750e+00
4.10346210e-01 -5.87119281e-01 6.01687193e-01 -9.19459108e-03
3.59851241e-01 1.88835084e-01 -8.82776916e-01 -9.22674358e-01
1.28108129e-01 -2.99137294e-01 6.26922309e-01 1.11266589e+00
-7.19413996e-01 2.60055691e-01 -2.62358934e-01 9.27059174e-01
6.12729490e-01 3.37197781e-01 7.15099633e-01 -1.38721204e+00
-6.13769174e-01 -4.07827914e-01 -6.51491344e-01 -9.75615382e-01
8.58676434e-02 -8.52763414e-01 4.82479751e-01 -1.45610905e+00
3.03904146e-01 -7.29431272e-01 -3.23840529e-01 4.44698721e-01
2.05936015e-01 3.99699062e-01 4.08161096e-02 3.25341523e-01
-3.46112698e-01 2.08240032e-01 5.07647693e-01 -6.25688672e-01
-1.10456254e-02 4.09974396e-01 -2.59779185e-01 7.76384115e-01
1.03751731e+00 -6.60538793e-01 -3.50607485e-01 -7.75638521e-01
2.13091075e-01 -1.98496640e-01 8.08644652e-01 -1.27207899e+00
1.37979105e-01 -2.08868727e-01 6.65669084e-01 -7.37397611e-01
4.81740423e-02 -1.13810170e+00 3.88622224e-01 5.54667532e-01
-2.61017770e-01 -2.75353700e-01 -5.46765104e-02 5.97474396e-01
-2.63817102e-01 -4.81262088e-01 1.02101052e+00 -4.66285706e-01
-1.17688930e+00 3.15348029e-01 -6.62019610e-01 -3.90113771e-01
1.03696537e+00 -5.89268625e-01 -3.43052924e-01 -4.04361129e-01
-1.09466970e+00 -1.41094834e-01 1.45341158e-01 1.75362587e-01
5.65397501e-01 -8.24923813e-01 -1.07012284e+00 3.18886369e-01
4.30014104e-01 -9.29588377e-02 3.00055355e-01 4.16289717e-01
-5.29773712e-01 -1.04283459e-01 -1.76792532e-01 -1.00035799e+00
-1.12337160e+00 4.27400589e-01 9.07954335e-01 1.01716585e-01
-3.36789876e-01 7.89830863e-01 3.74120951e-01 -8.66329312e-01
1.48537502e-01 -2.77397752e-01 -3.23914260e-01 5.51466420e-02
6.06355608e-01 5.72080091e-02 1.99774861e-01 -8.28441322e-01
-2.28929132e-01 3.94116640e-01 1.56798393e-01 1.98958740e-01
1.28417349e+00 3.30492035e-02 5.85149191e-02 5.52141190e-01
1.01601040e+00 -2.71415442e-01 -1.02999079e+00 -6.39723837e-01
-8.61706063e-02 -6.41159296e-01 2.02603415e-01 -9.02668715e-01
-1.03349459e+00 1.04401207e+00 1.56294477e+00 4.75948423e-01
1.13879895e+00 6.15245812e-02 1.36912942e-01 7.63146460e-01
3.16377640e-01 -9.55665112e-01 -5.54972231e-01 4.90801066e-01
5.37856698e-01 -1.47415614e+00 -2.71418631e-01 1.46696374e-01
-1.30327612e-01 1.18358278e+00 3.29857558e-01 -3.41955543e-01
1.18876863e+00 1.11655794e-01 5.53051054e-01 -1.13806427e-01
-1.07667856e-01 -6.66094840e-01 5.51720932e-02 9.70542133e-01
9.29993466e-02 3.97679716e-01 1.37048632e-01 -1.94177032e-02
-2.47348547e-02 -1.16844043e-01 5.36882043e-01 8.39194119e-01
-7.02934265e-01 -7.24556625e-01 -6.43277109e-01 5.81221402e-01
-4.66797084e-01 -2.46477693e-01 -3.65230799e-01 7.28492558e-01
5.32813370e-01 1.07431734e+00 4.95428830e-01 -2.75495559e-01
2.70745307e-01 -2.78668642e-01 2.77872235e-01 -6.61666214e-01
-4.96617973e-01 -7.57914558e-02 -2.71743208e-01 9.59441904e-03
-7.59654641e-01 -4.28832531e-01 -8.82371247e-01 -2.25105599e-01
-4.90021080e-01 -4.65324484e-02 8.10095608e-01 8.93136621e-01
1.09054171e-01 1.36413887e-01 8.01472723e-01 -9.24511194e-01
-9.33728039e-01 -1.05138683e+00 -1.07012498e+00 5.00546753e-01
5.44316232e-01 -2.01061323e-01 -3.91749471e-01 1.51867479e-01] | [9.643939018249512, -1.3386949300765991] |
c545e26f-b233-4ba1-baee-507d74eb0441 | discovering-topics-with-neural-topic-models | null | null | https://openreview.net/forum?id=Skx24yHFDr | https://openreview.net/pdf?id=Skx24yHFDr | Discovering Topics With Neural Topic Models Built From PLSA Loss | In this paper we present a model for unsupervised topic discovery in texts corpora. The proposed model uses documents, words, and topics lookup table embedding as neural network model parameters to build probabilities of words given topics, and probabilities of topics given documents. These probabilities are used to recover by marginalization probabilities of words given documents. For very large corpora where the number of documents can be in the order of billions, using a neural auto-encoder based document embedding is more scalable then using a lookup table embedding as classically done. We thus extended the lookup based document embedding model to continuous auto-encoder based model. Our models are trained using probabilistic latent semantic analysis (PLSA) assumptions. We evaluated our models on six datasets with a rich variety of contents. Conducted experiments demonstrate that the proposed neural topic models are very effective in capturing relevant topics. Furthermore, considering perplexity metric, conducted evaluation benchmarks show that our topic models outperform latent Dirichlet allocation (LDA) model which is classically used to address topic discovery tasks. | ['Sileye Ba'] | 2019-09-25 | null | null | null | null | ['document-embedding', 'topic-models'] | ['methodology', 'natural-language-processing'] | [-1.49874881e-01 5.40895343e-01 -4.94569957e-01 -5.45913696e-01
-8.71790707e-01 -1.64048046e-01 1.18267000e+00 3.26158643e-01
-2.86031634e-01 6.40270829e-01 7.63738573e-01 -1.38971388e-01
-1.18697388e-03 -1.18465054e+00 -6.95328951e-01 -6.30708516e-01
-2.22398683e-01 1.07192600e+00 5.61977662e-02 8.08779970e-02
3.86373103e-01 -7.38818198e-02 -1.42517686e+00 2.11916491e-01
6.65323913e-01 5.01136720e-01 3.71378809e-01 4.28595513e-01
-9.83046293e-01 3.89258355e-01 -6.66740239e-01 -1.15680918e-01
-9.62173268e-02 5.10863028e-02 -8.91087949e-01 4.92147245e-02
6.41362965e-02 -7.13229060e-01 -2.32609242e-01 7.89795518e-01
6.86507300e-02 2.85044998e-01 1.19939172e+00 -1.20271635e+00
-8.82952094e-01 1.21658742e+00 -4.43839997e-01 3.67286056e-02
4.60438710e-03 -5.05040050e-01 1.48852551e+00 -9.81393397e-01
5.60478985e-01 1.72511292e+00 3.43377799e-01 2.94618607e-01
-1.11867917e+00 -6.61760092e-01 3.14276159e-01 8.11423883e-02
-1.28028655e+00 1.40054487e-02 8.37085187e-01 -6.35882676e-01
1.17008817e+00 -3.74908686e-01 4.31922048e-01 1.23920846e+00
3.52724046e-01 8.65249336e-01 6.92167938e-01 -5.97472847e-01
4.11603928e-01 7.28481770e-01 6.67314589e-01 2.21992567e-01
4.75435078e-01 -5.99062741e-01 -5.63248754e-01 -5.24554729e-01
5.65539420e-01 3.48429292e-01 1.49120107e-01 -3.56346697e-01
-1.18956852e+00 1.60847056e+00 3.60031240e-03 3.29203188e-01
-6.97265565e-01 3.71282309e-01 4.39923197e-01 -1.20256841e-01
8.34287465e-01 3.15660864e-01 -3.08639616e-01 -1.04312943e-02
-1.07090521e+00 2.80277580e-01 1.02896833e+00 9.50377047e-01
9.16521728e-01 1.57643221e-02 -2.13081717e-01 9.60581839e-01
7.73447454e-01 3.62251729e-01 1.02737331e+00 -6.15385830e-01
3.09123605e-01 4.68713939e-01 1.00512505e-01 -1.02518129e+00
-1.04982823e-01 -1.27369478e-01 -6.26046777e-01 -4.71310407e-01
-7.57475197e-02 -9.89426970e-02 -9.23096061e-01 1.55286765e+00
1.73488244e-01 -1.47379190e-01 3.17926586e-01 4.40243751e-01
5.35092592e-01 1.47845697e+00 4.10666287e-01 -1.54949605e-01
1.72006619e+00 -8.80237520e-01 -1.07506275e+00 -9.87241510e-03
5.72186768e-01 -6.76730752e-01 1.07184768e+00 3.07752639e-01
-7.67246783e-01 -3.14317793e-01 -8.65656555e-01 -2.79316127e-01
-7.98995495e-01 2.32933760e-01 9.10721123e-01 6.90531611e-01
-1.06506252e+00 2.59055346e-01 -9.80606079e-01 -6.28311694e-01
2.31926382e-01 3.53466421e-01 -7.58240372e-02 1.91105410e-01
-1.48965383e+00 3.38867575e-01 8.55514526e-01 -4.53449339e-01
-1.33008146e+00 -6.00418091e-01 -7.63585389e-01 5.47475576e-01
6.98106736e-02 -4.15524215e-01 1.06017494e+00 -2.70332187e-01
-1.47990477e+00 3.55778664e-01 -2.62874246e-01 -7.85109878e-01
-3.55710506e-01 -4.84671384e-01 -2.13121787e-01 3.23258758e-01
3.75337183e-01 1.10280323e+00 8.96456897e-01 -1.05015910e+00
-5.66290557e-01 -1.39752910e-01 9.85718518e-03 1.34349182e-01
-1.17142010e+00 -4.81528342e-02 -2.58921653e-01 -5.95794141e-01
2.38399431e-01 -6.20639741e-01 -8.90795141e-03 -2.22656861e-01
-4.37488019e-01 -9.27413762e-01 9.96341348e-01 -7.28069842e-01
1.10123754e+00 -1.88852942e+00 -2.38585308e-01 1.39814056e-02
1.94465220e-01 -3.11726838e-01 1.91616282e-01 7.55179465e-01
7.53733469e-03 2.10927248e-01 -1.01872173e-03 -5.35787523e-01
5.10100842e-01 2.41239503e-01 -1.05344605e+00 1.07059173e-01
3.29998098e-02 6.30601168e-01 -4.85552222e-01 -6.40639663e-01
2.96867848e-03 8.19428265e-01 -6.62806809e-01 2.47292861e-01
-5.52495837e-01 -4.61646616e-01 -5.68624973e-01 2.07060814e-01
4.50435519e-01 -4.99783635e-01 1.36321858e-01 6.75602630e-02
2.02262536e-01 6.35290980e-01 -1.00183642e+00 1.69144166e+00
-6.85917675e-01 8.49788189e-01 -5.76984882e-01 -9.69519377e-01
1.30948472e+00 6.80191338e-01 4.92541522e-01 8.26862752e-02
-8.68965760e-02 -2.87973285e-01 -4.52970475e-01 -1.97962701e-01
1.00100958e+00 -1.55810788e-01 -1.34124070e-01 1.04758441e+00
5.71506381e-01 2.32045218e-01 1.23145178e-01 6.85136914e-01
6.71912968e-01 -1.81855962e-01 2.12829173e-01 -4.88991320e-01
1.42763287e-01 6.54548705e-02 1.32398158e-01 5.95334768e-01
2.38085151e-01 1.95924759e-01 8.74016106e-01 -2.87195504e-01
-1.28536832e+00 -1.09654498e+00 -3.52566838e-01 1.29918599e+00
-3.69660288e-01 -7.08285451e-01 -5.62752306e-01 -4.55957890e-01
-2.22250998e-01 1.09170783e+00 -7.58028150e-01 1.88160732e-01
-9.15173739e-02 -1.01864564e+00 3.29169989e-01 5.53842783e-01
2.05800548e-01 -9.45324183e-01 -2.15216801e-01 3.57624352e-01
-1.32867962e-01 -8.79169166e-01 -1.20705821e-01 2.81471252e-01
-1.12389922e+00 -4.09625888e-01 -1.07680666e+00 -7.68430829e-01
5.56260824e-01 1.53132245e-01 8.95440698e-01 -8.24267745e-01
-1.34460106e-01 5.53686976e-01 -3.90777677e-01 -4.99096960e-01
-2.45413512e-01 3.13104302e-01 1.44044995e-01 -1.40710130e-01
1.00706971e+00 -7.14753926e-01 -5.30095100e-01 -9.72451344e-02
-1.16833150e+00 -1.76065803e-01 5.65506279e-01 7.83321679e-01
1.67695120e-01 3.48437279e-01 8.36286008e-01 -1.03831029e+00
9.88905668e-01 -1.04040158e+00 -5.21450996e-01 6.14661239e-02
-1.00263989e+00 3.38008016e-01 2.68361807e-01 -4.86392736e-01
-1.34718955e+00 -4.23965901e-01 2.17853129e-01 -1.34615391e-01
-3.97015154e-01 4.77125257e-01 1.01206079e-01 9.81931210e-01
4.13778841e-01 4.23225552e-01 -2.96296656e-01 -7.12367654e-01
6.86735272e-01 1.03989434e+00 -3.81900668e-02 -6.77318037e-01
3.08303207e-01 6.65516257e-01 -4.59272772e-01 -8.57408285e-01
-6.55579746e-01 -8.56161356e-01 -6.12279832e-01 2.65625089e-01
1.11466002e+00 -1.30277205e+00 -2.94642270e-01 -1.64893523e-01
-1.50991511e+00 2.33715564e-01 -2.74040043e-01 8.31246495e-01
-3.72893155e-01 2.95884728e-01 -6.29684627e-01 -9.22442198e-01
-6.37519538e-01 -9.56320047e-01 1.29862440e+00 -1.27779786e-02
-2.73569226e-01 -1.51938665e+00 4.50548202e-01 -2.90847127e-03
5.45351088e-01 -2.40572140e-01 1.25806952e+00 -1.14509594e+00
-4.18601692e-01 -2.68053591e-01 -3.28109354e-01 1.33585334e-01
1.98038250e-01 -2.03376099e-01 -1.17743492e+00 -2.20863387e-01
-1.12804301e-01 -1.72901973e-01 1.03899312e+00 5.09446144e-01
9.16797161e-01 -5.50923944e-01 -6.57670021e-01 7.40086436e-02
1.51245570e+00 2.68167913e-01 5.73538899e-01 4.07750070e-01
4.74967808e-01 7.87527025e-01 3.02794784e-01 6.19675756e-01
6.23108149e-01 3.77676249e-01 3.66252027e-02 3.33961964e-01
2.31220365e-01 -3.59864354e-01 4.98466343e-01 1.25863135e+00
3.61945957e-01 -4.37052399e-01 -9.98923063e-01 1.04571533e+00
-1.44575202e+00 -7.64681339e-01 1.72369957e-01 1.74626493e+00
9.00158584e-01 1.67140692e-01 -9.20513086e-03 -2.37602353e-01
8.13546658e-01 2.17228815e-01 -1.26542553e-01 -5.35548329e-01
2.82759249e-01 1.85248461e-02 1.51983649e-01 4.17869657e-01
-1.07388246e+00 1.02745914e+00 6.42983055e+00 8.18895102e-01
-7.56266773e-01 4.54677463e-01 5.11967897e-01 1.04711488e-01
-5.68194270e-01 2.04470649e-01 -1.33290029e+00 4.53405946e-01
1.32971489e+00 -4.04683053e-01 -2.71130472e-01 1.35460818e+00
2.86041684e-02 2.50259452e-02 -9.50770497e-01 7.06934154e-01
2.69968122e-01 -1.10503113e+00 5.97654581e-01 3.03855270e-01
8.67678642e-01 -1.30017772e-01 2.08761334e-01 5.78672588e-01
5.56370795e-01 -7.87911415e-01 8.92856345e-02 3.23668122e-01
3.12419981e-01 -6.30397201e-01 8.14851940e-01 1.72761649e-01
-7.53360212e-01 2.34144256e-01 -1.04142332e+00 1.99071974e-01
2.75523990e-01 8.69935632e-01 -1.56738818e+00 9.10727531e-02
7.00968206e-01 5.65205097e-01 -1.32534862e-01 6.40176654e-01
-1.21102154e-01 1.12432396e+00 -3.91128540e-01 -2.70189643e-01
4.98038799e-01 9.84115340e-03 4.28940713e-01 1.33658230e+00
3.12226385e-01 -2.96672493e-01 -1.22635134e-01 1.17032468e+00
-5.15889078e-02 5.33273041e-01 -5.36355793e-01 -4.25015628e-01
3.90763938e-01 1.21486080e+00 -1.05581009e+00 -8.10590565e-01
-2.93178290e-01 7.36311495e-01 4.86084335e-02 4.20301020e-01
-5.65626860e-01 -5.31395793e-01 4.20118183e-01 -2.12789420e-02
5.11440516e-01 -3.20209503e-01 1.71805680e-01 -1.06212640e+00
-2.59496272e-01 -2.71526158e-01 2.64248431e-01 -6.20960176e-01
-1.26762438e+00 6.65529728e-01 7.13269234e-01 -8.77692401e-01
-5.72996795e-01 -4.90107954e-01 -6.64576828e-01 8.19541335e-01
-1.51372361e+00 -1.03157091e+00 -6.85657002e-03 2.37276375e-01
1.08052671e+00 -5.37530184e-01 1.15105748e+00 3.56622078e-02
-2.94615030e-01 1.30842939e-01 6.37649834e-01 -2.46116687e-02
7.50075758e-01 -1.53985763e+00 5.25445998e-01 2.60767817e-01
3.35789263e-01 1.07429898e+00 6.45902216e-01 -7.76848614e-01
-9.85308766e-01 -9.62107837e-01 1.03330290e+00 -1.99990422e-01
7.52480149e-01 -6.90947711e-01 -1.02721012e+00 8.89505744e-01
5.99434078e-01 -7.26330042e-01 1.12783682e+00 5.22535563e-01
-3.65786493e-01 2.33627155e-01 -6.27640963e-01 2.12750673e-01
4.18747813e-02 -4.65113401e-01 -9.99304533e-01 7.82292306e-01
1.38979828e+00 3.82104486e-01 -1.02519977e+00 -2.92528272e-01
3.77321273e-01 -3.41971219e-01 9.08826172e-01 -6.80243313e-01
6.02786183e-01 1.31111518e-01 -3.01096410e-01 -1.09577930e+00
-1.65782183e-01 -3.61726195e-01 -3.62345070e-01 1.54495025e+00
4.89880353e-01 -5.72404921e-01 9.30125952e-01 5.50488710e-01
2.44645789e-01 -4.94163662e-01 -5.93975723e-01 -5.51425278e-01
3.92075628e-01 -3.83463889e-01 6.54928565e-01 1.01271474e+00
2.08508685e-01 4.73172754e-01 -3.00741583e-01 1.38861626e-01
6.79007292e-01 2.24388894e-02 6.75299227e-01 -1.38145900e+00
-3.24448824e-01 -1.15458623e-01 -3.46071810e-01 -1.18761051e+00
2.70803213e-01 -7.34837711e-01 -6.66087046e-02 -1.74675512e+00
4.76918757e-01 -3.85026962e-01 -2.90599257e-01 3.95832151e-01
1.11192375e-01 -2.48772904e-01 -2.21846670e-01 4.39251572e-01
-6.00211561e-01 9.66515481e-01 6.11327231e-01 -2.48934910e-01
-2.26671755e-01 -1.85805559e-01 -6.32239640e-01 7.63992250e-01
9.13725436e-01 -8.40476513e-01 -6.30627215e-01 -4.33625549e-01
3.42554510e-01 -1.13264129e-01 4.37486395e-02 -7.51248956e-01
3.82352501e-01 1.52572051e-01 2.03474849e-01 -1.06657386e+00
6.57990694e-01 -7.29782641e-01 -3.62463087e-01 3.21561508e-02
-8.55108917e-01 -2.22284958e-01 1.22176729e-01 9.20476973e-01
-3.95053416e-01 -4.41242516e-01 6.76514208e-02 -9.24547613e-02
-5.75894892e-01 1.46456197e-01 -6.85902119e-01 -2.91931987e-01
7.97852159e-01 5.36577329e-02 -1.08068056e-01 -7.52375722e-01
-6.87871993e-01 -1.48854302e-02 -1.85762793e-01 5.19876719e-01
5.90138614e-01 -1.25315833e+00 -5.67579329e-01 -6.41626492e-02
1.47438005e-01 -1.45604685e-02 1.59885839e-01 3.01694870e-01
-3.55017483e-01 1.17867005e+00 6.54487498e-03 -6.35911703e-01
-7.27082670e-01 5.80595970e-01 -3.88963997e-01 -4.30524856e-01
-6.53601706e-01 5.90753555e-01 6.93541706e-01 -3.55222583e-01
3.75265241e-01 -4.31022018e-01 -6.11128688e-01 1.74832955e-01
5.78441143e-01 3.72052103e-01 -3.52661520e-01 -3.87220860e-01
1.23116642e-01 2.94627368e-01 -6.79168940e-01 -5.10596633e-01
1.54244542e+00 -2.73328006e-01 -2.10788280e-01 8.29932690e-01
1.49622881e+00 -4.96104658e-01 -8.20811749e-01 -5.36134958e-01
4.10024524e-02 -8.31014886e-02 3.95673186e-01 -2.81003326e-01
-4.29934293e-01 1.35377026e+00 6.02469504e-01 4.26349342e-01
5.52038908e-01 2.60193139e-01 8.21110606e-01 7.65365422e-01
-1.51481584e-03 -1.10208642e+00 1.74178839e-01 3.39292198e-01
6.39131010e-01 -1.20118582e+00 8.21060762e-02 -2.44405970e-01
-4.71469343e-01 1.28716707e+00 4.01610643e-01 -2.52959877e-01
1.03974092e+00 -9.22296494e-02 -3.09207022e-01 -4.44196075e-01
-1.06134403e+00 6.78269938e-02 3.13591480e-01 3.21990490e-01
4.80652392e-01 3.41716222e-02 -2.42947370e-01 7.72342384e-01
-4.13486242e-01 -4.13370609e-01 5.10930538e-01 6.33580685e-01
-9.64544535e-01 -9.94496226e-01 -4.82936859e-01 4.99954611e-01
-6.21775925e-01 -2.91311532e-01 7.33473226e-02 6.78486943e-01
-3.00396383e-01 6.78142250e-01 5.36957085e-01 9.46378112e-02
-6.01057291e-01 5.75045466e-01 -2.84260660e-01 -9.63640273e-01
1.77893698e-01 2.77714640e-01 -2.80785650e-01 -6.98494986e-02
-3.36308807e-01 -5.97209871e-01 -1.08842564e+00 1.77364409e-01
-5.97774267e-01 6.40146911e-01 1.34088838e+00 8.55346382e-01
3.66035342e-01 5.25520682e-01 4.42586452e-01 -4.13670272e-01
-5.30897319e-01 -1.54063952e+00 -7.82517612e-01 4.46668603e-02
-1.29936576e-01 -7.34682739e-01 -4.43497688e-01 4.79231626e-01] | [10.413408279418945, 6.954501628875732] |
3239eba6-6c62-4e1b-92a3-1a76b036ccc9 | shape-preserving-half-projective-warps-for | null | null | http://openaccess.thecvf.com/content_cvpr_2014/html/Chang_Shape-Preserving_Half-Projective_Warps_2014_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2014/papers/Chang_Shape-Preserving_Half-Projective_Warps_2014_CVPR_paper.pdf | Shape-Preserving Half-Projective Warps for Image Stitching | This paper proposes a novel parametric warp which is a spatial combination of a projective transformation and a similarity transformation. Given the projective transformation relating two input images, based on an analysis of the projective transformation, our method smoothly extrapolates the projective transformation of the overlapping regions into the non-overlapping regions and the resultant warp gradually changes from projective to similarity across the image. The proposed warp has the strengths of both projective and similarity warps. It provides good alignment accuracy as projective warps while preserving the perspective of individual image as similarity warps. It can also be combined with more advanced local-warp-based alignment methods such as the as-projective-as-possible warp for better alignment accuracy. With the proposed warp, the field of view can be extended by stitching images with less projective distortion (stretched shapes and enlarged sizes). | ['Yung-Yu Chuang', 'Yoichi Sato', 'Che-Han Chang'] | 2014-06-01 | null | null | null | cvpr-2014-6 | ['image-stitching'] | ['computer-vision'] | [ 3.26820910e-01 -1.76817939e-01 7.66129941e-02 -1.63921997e-01
-4.67605323e-01 -8.56092632e-01 7.14421451e-01 -3.03300530e-01
-2.02840701e-01 5.01136184e-01 1.44905090e-01 1.09722555e-01
-2.14542568e-01 -7.88706183e-01 -5.18055081e-01 -8.51906955e-01
2.22460389e-01 3.87292355e-01 8.73606443e-01 -5.67509830e-01
4.89960372e-01 8.38309288e-01 -7.85844147e-01 -7.69169256e-02
7.16924846e-01 2.68812031e-01 1.04937695e-01 4.40572441e-01
4.24394533e-02 -1.40420154e-01 -4.00525063e-01 -4.36131239e-01
8.46871674e-01 -4.10566390e-01 -6.77950084e-01 2.03692377e-01
6.87980831e-01 -3.72762144e-01 -4.83215451e-01 1.03450489e+00
2.61960536e-01 1.44414961e-01 2.39033222e-01 -1.13059354e+00
-4.67821419e-01 1.39971316e-01 -1.11286235e+00 -4.59119901e-02
5.80434144e-01 4.62913886e-02 5.17650604e-01 -7.70855725e-01
9.37308848e-01 1.23232424e+00 5.60339332e-01 1.96326450e-02
-1.45940661e+00 -8.43272388e-01 -4.23131824e-01 1.31045550e-01
-1.50049615e+00 -7.58242756e-02 1.07548964e+00 -2.09583923e-01
5.85527718e-01 5.42488992e-01 8.30901682e-01 8.07987988e-01
8.20909023e-01 9.02873278e-02 1.13672757e+00 -4.70535278e-01
-2.61459172e-01 -2.48789474e-01 -3.56602162e-01 4.35454756e-01
4.57528047e-02 2.41981640e-01 -1.64590269e-01 -2.81976730e-01
1.42311502e+00 3.22171748e-01 -3.41298252e-01 -5.83748937e-01
-1.90112412e+00 4.78000522e-01 6.55181348e-01 4.63252991e-01
-8.52083787e-02 -2.60298908e-01 5.50153442e-02 2.08012849e-01
2.42908701e-01 9.84837234e-01 1.30091071e-01 1.00721568e-01
-1.31530452e+00 2.51498401e-01 4.22589213e-01 1.07638896e+00
7.64034927e-01 2.59560682e-02 1.63910329e-01 6.86357975e-01
3.90626416e-02 7.50431538e-01 3.93082500e-01 -7.86722898e-01
6.12105429e-01 3.59419197e-01 9.63982791e-02 -1.31989682e+00
-4.83327240e-01 7.48396590e-02 -6.45163417e-01 4.67805147e-01
2.76219070e-01 1.03555419e-01 -6.14168882e-01 1.34040987e+00
5.10125160e-01 1.93385974e-01 -1.84653714e-01 9.01075363e-01
4.45191562e-01 1.03409266e+00 -5.65764368e-01 -3.47188622e-01
1.48961544e+00 -1.07071269e+00 -7.10684836e-01 -2.79139318e-02
9.98901762e-03 -1.49393153e+00 8.30892980e-01 1.75834715e-01
-1.20843565e+00 -6.65592372e-01 -1.45364332e+00 -6.97900057e-02
1.27620265e-01 -4.33451831e-01 -1.72921121e-01 2.76720673e-01
-1.08747923e+00 6.07541561e-01 -8.53663564e-01 -7.11363435e-01
-1.94998339e-01 2.45262995e-01 -7.85625398e-01 7.08410740e-02
-8.36254179e-01 1.01903546e+00 5.51600516e-01 -1.99344218e-01
-2.44161367e-01 -5.57106972e-01 -6.26801431e-01 -6.21825457e-02
1.48505554e-01 -6.25380874e-01 6.03916705e-01 -9.86445308e-01
-1.79230022e+00 7.09645510e-01 1.32750303e-01 6.42579570e-02
6.32813394e-01 -1.14198521e-01 -4.04964238e-01 1.43212482e-01
-2.53376782e-01 7.23283887e-01 1.14913809e+00 -1.24038482e+00
-2.90949911e-01 -2.72779852e-01 -2.17050970e-01 5.57243407e-01
-6.30261078e-02 2.85379440e-01 -4.19759244e-01 -8.62036347e-01
6.03427827e-01 -1.22673452e+00 -1.31953374e-01 1.31731883e-01
-4.94250149e-01 5.19906878e-01 1.45165467e+00 -8.56885016e-01
1.10778570e+00 -2.34166646e+00 5.19575179e-01 3.02896947e-01
2.26929083e-01 2.16884524e-01 -4.57567424e-01 1.03775430e+00
-3.83684367e-01 -2.49219164e-01 -1.26658872e-01 9.04243533e-03
-6.63565040e-01 2.97595590e-01 -4.54003096e-01 7.08983600e-01
1.67477340e-01 7.40509093e-01 -9.29135203e-01 -6.89903677e-01
1.64792925e-01 4.31294024e-01 -2.35540807e-01 2.54908472e-01
5.58489561e-01 6.96108222e-01 -2.28356108e-01 3.33654284e-01
1.19466639e+00 5.45353353e-01 9.46205407e-02 -5.62451482e-01
-5.13364732e-01 -3.87631983e-01 -1.03438246e+00 1.69728673e+00
-3.37214805e-02 8.10963154e-01 -1.52346715e-01 -5.05446434e-01
1.69882762e+00 1.97632536e-01 7.11109161e-01 -1.96078926e-01
1.69736579e-01 9.15217102e-02 3.27997319e-02 -4.10911649e-01
8.79840195e-01 -2.67350733e-01 1.61834985e-01 3.98100585e-01
4.71220277e-02 -7.61655331e-01 -9.30356756e-02 -1.26027048e-01
6.35549366e-01 4.15647060e-01 6.51235819e-01 -5.22940159e-01
6.53002322e-01 -5.42349853e-02 4.19365644e-01 1.66387734e-04
-1.21121734e-01 1.01128483e+00 3.15302938e-01 -6.52184427e-01
-1.92385828e+00 -1.37310958e+00 -2.68411756e-01 2.81903714e-01
7.08450913e-01 -2.36373842e-01 -7.88285732e-01 -4.23466295e-01
-2.77720273e-01 2.54740357e-01 -3.39168996e-01 -8.52827728e-02
-1.00391662e+00 -8.73721093e-02 3.74931157e-01 3.67252648e-01
7.10398793e-01 -7.37518549e-01 -3.21356118e-01 -2.87890658e-02
-1.31363094e-01 -1.06378543e+00 -1.13082433e+00 -4.99091744e-01
-1.07408285e+00 -9.81620073e-01 -8.56468141e-01 -9.83581007e-01
9.87393796e-01 7.53208101e-01 4.55955863e-01 -3.67393076e-01
1.54516593e-01 1.11315630e-01 -5.14769912e-01 -1.70595627e-02
-5.08178949e-01 -2.56876022e-01 1.58534765e-01 1.46671161e-01
-1.68146312e-01 -8.08494568e-01 -7.00554669e-01 1.01168919e+00
-1.08634663e+00 3.91051471e-01 3.86887729e-01 8.98896039e-01
5.41721642e-01 -2.69015282e-02 -1.92303717e-01 -3.37365150e-01
5.32884657e-01 2.51034871e-02 -6.09696329e-01 1.92750663e-01
-2.19640091e-01 1.63984805e-01 6.32667303e-01 -8.06804419e-01
-1.11919641e+00 -8.66357386e-02 1.57565922e-01 -5.78538239e-01
2.88714647e-01 -6.12742715e-02 -2.85085868e-02 -6.68342173e-01
7.44701445e-01 2.51369387e-01 4.96021092e-01 -2.67971814e-01
3.86241913e-01 2.97392666e-01 8.06205511e-01 -4.17950988e-01
1.60727906e+00 7.12357044e-01 4.05075520e-01 -8.02782595e-01
1.32333115e-01 -2.40281999e-01 -1.25560510e+00 -2.31136546e-01
9.05051291e-01 -3.56680483e-01 -2.78633714e-01 3.21858466e-01
-1.21565998e+00 4.12797511e-01 -4.14238513e-01 6.79487765e-01
-4.67188776e-01 9.34075356e-01 -2.91722953e-01 -9.03995335e-02
-3.32042933e-01 -1.31178141e+00 8.03193986e-01 4.07137364e-01
-4.01640177e-01 -7.88912177e-01 4.59423274e-01 2.82675046e-02
3.38355422e-01 5.36107898e-01 6.76857591e-01 -5.66852987e-01
-4.71307546e-01 -3.28151077e-01 -3.14405933e-02 1.01830460e-01
4.21221286e-01 6.80871129e-01 -3.19347113e-01 -6.17854357e-01
2.23451346e-01 3.02758694e-01 2.06069991e-01 2.30246142e-01
3.02244663e-01 -4.61247146e-01 -3.89675230e-01 8.98963690e-01
1.50195277e+00 6.03323340e-01 1.00931156e+00 6.31013989e-01
5.28180778e-01 5.81506014e-01 7.95835197e-01 4.80178520e-02
-1.70745164e-01 1.21943700e+00 3.98034751e-01 -3.96870852e-01
-1.07347174e-02 -3.02911133e-01 3.59447896e-01 1.03244174e+00
-6.17330015e-01 2.81947225e-01 -7.09299386e-01 1.54342428e-01
-1.67267525e+00 -1.09536815e+00 -1.25847654e-02 2.66965461e+00
6.45540535e-01 -3.01956981e-01 5.38746640e-02 -5.36118522e-02
9.55914438e-01 6.52305186e-01 -1.49412483e-01 -7.83101916e-01
-1.29691646e-01 1.48288220e-01 5.99987268e-01 7.16606438e-01
-8.84770811e-01 7.40137577e-01 6.93003130e+00 7.13322461e-01
-1.26403582e+00 4.67732735e-02 -4.16826792e-02 6.56692758e-02
-4.19473678e-01 3.13716471e-01 -4.21090931e-01 3.80609483e-01
1.70396388e-01 -2.29045123e-01 4.36120301e-01 4.69001979e-01
5.11463806e-02 4.34235074e-02 -9.65201557e-01 9.27595794e-01
1.25603601e-01 -1.01984572e+00 1.39916331e-01 3.92359197e-02
9.60050464e-01 -3.50102216e-01 1.68145820e-01 -3.06606263e-01
-1.10620692e-01 -5.64244390e-01 5.21864235e-01 4.34268564e-01
8.52079034e-01 -7.89272606e-01 7.70213783e-01 9.02599990e-02
-1.26456285e+00 2.02856421e-01 -5.22415042e-01 1.70203939e-01
3.96000743e-01 2.60668714e-02 -7.82000482e-01 8.79832149e-01
3.11967134e-01 3.95117134e-01 -3.53874028e-01 8.32865298e-01
-1.17934085e-01 -1.46424592e-01 -2.90488571e-01 4.02540416e-01
2.41942093e-01 -1.09450078e+00 1.09576225e+00 8.53282034e-01
9.17596459e-01 2.32772127e-01 9.03611779e-02 6.28185570e-01
4.06520128e-01 2.75911033e-01 -7.25508332e-01 3.13924342e-01
7.15634525e-01 1.29918325e+00 -6.44030452e-01 -1.79408178e-01
-3.96488905e-01 1.11822057e+00 -2.99847603e-01 3.49565476e-01
-9.15112376e-01 -4.91351873e-01 7.31184661e-01 2.63613701e-01
9.23398957e-02 -6.56903505e-01 -1.23498015e-01 -1.01173246e+00
7.88115896e-03 -7.56366432e-01 2.37921271e-02 -1.02873409e+00
-8.13574970e-01 7.76147008e-01 2.92069346e-01 -1.93850851e+00
-1.69549994e-02 -2.67679214e-01 -9.00401294e-01 9.23219860e-01
-8.80096495e-01 -1.46970713e+00 -5.09962022e-01 6.15699530e-01
5.65215170e-01 -1.34025797e-01 6.47707045e-01 -1.95215464e-01
-1.25814348e-01 4.98152643e-01 3.86064202e-01 -9.42397937e-02
1.32129169e+00 -6.99957669e-01 5.23683906e-01 1.07952130e+00
8.61999393e-02 5.97582102e-01 8.57280970e-01 -5.75298905e-01
-1.10235345e+00 -7.35030234e-01 5.65791070e-01 -3.17942172e-01
6.60473943e-01 5.22978492e-02 -1.00242889e+00 7.92529404e-01
7.41590977e-01 -1.55421332e-01 2.45532662e-01 -3.97504091e-01
-4.93279219e-01 -4.94061798e-01 -1.34302127e+00 7.45776653e-01
7.32736111e-01 -3.61477613e-01 -8.73993874e-01 1.98942404e-02
7.80568182e-01 -8.16190541e-01 -1.06551743e+00 2.13951573e-01
8.89694631e-01 -9.64691937e-01 1.19916487e+00 -1.12358257e-01
5.82450807e-01 -5.19293308e-01 1.26552969e-01 -1.44999254e+00
-8.74215066e-01 -9.36028302e-01 6.94234073e-01 1.22371125e+00
-1.40473202e-01 -9.56612170e-01 3.48990917e-01 1.10936746e-01
-4.31492701e-02 -5.35079062e-01 -1.17031610e+00 -1.03554523e+00
-8.73870105e-02 4.65676755e-01 7.54647434e-01 1.19350898e+00
3.08196396e-01 -7.09516257e-02 -5.98221600e-01 3.23857367e-01
4.90597159e-01 8.37270170e-03 1.00710428e+00 -8.43165755e-01
-1.99642941e-01 -3.04859579e-01 -8.89138222e-01 -7.99343228e-01
-4.01497066e-01 -5.58953583e-01 -2.16777042e-01 -9.22817826e-01
2.43227899e-01 -1.24057166e-01 -6.40352303e-03 3.72186631e-01
-8.86843819e-03 3.24227512e-01 4.80474532e-01 8.12598944e-01
4.58498448e-01 3.22939128e-01 1.90391350e+00 6.26276508e-02
-4.99912232e-01 -2.09172264e-01 3.69505547e-02 5.07473350e-01
5.89187562e-01 -3.48688841e-01 -3.27109963e-01 -2.97808766e-01
-1.20687746e-01 3.02716464e-01 -4.28979360e-02 -9.61105168e-01
1.53157249e-01 -5.37310779e-01 1.13203950e-01 -4.92736280e-01
3.49253386e-01 -1.12145627e+00 7.98347056e-01 5.81041634e-01
5.61593249e-02 6.67193770e-01 7.25338235e-02 8.50236937e-02
-4.37236041e-01 -1.51054785e-01 1.15331388e+00 2.60510981e-01
-4.03395385e-01 2.05148190e-01 2.50168592e-01 -5.91438651e-01
1.54949105e+00 -8.17199767e-01 -5.25213420e-01 -2.46066868e-01
-3.46244454e-01 -1.99019179e-01 1.01923048e+00 4.00462806e-01
6.68426752e-01 -1.71082771e+00 -7.50277817e-01 5.21175086e-01
-2.08708309e-02 -7.09707811e-02 3.05541307e-01 9.87741411e-01
-1.22946227e+00 1.71535924e-01 -1.12118149e+00 -7.19376147e-01
-1.79423153e+00 8.73685300e-01 1.01954952e-01 -2.20874175e-01
-7.62155712e-01 3.64673257e-01 5.07009327e-01 -2.36843631e-01
-5.16382396e-01 -1.02480151e-01 -1.39436945e-02 -2.43689790e-01
5.21977842e-01 4.68225151e-01 -3.10465485e-01 -8.35142136e-01
-2.09802821e-01 1.26431429e+00 2.80063692e-02 -3.25344235e-01
1.27129209e+00 -1.69744760e-01 -4.06043708e-01 8.65979493e-02
1.19201624e+00 5.66895306e-01 -1.43166184e+00 -3.66697013e-01
-6.33108795e-01 -1.14699233e+00 -4.39419985e-01 -2.38928109e-01
-1.04659617e+00 6.80698872e-01 5.05714357e-01 1.21397361e-01
1.29021561e+00 -1.53205305e-01 7.09317684e-01 -1.51426837e-01
3.96663368e-01 -6.80436671e-01 1.88360870e-01 2.68482566e-01
1.33426642e+00 -7.28521943e-01 4.96795475e-01 -5.55434346e-01
-6.40828729e-01 1.59231222e+00 4.31650698e-01 -5.32857895e-01
1.50964305e-01 3.62116069e-01 1.47135453e-02 5.91395572e-02
-1.17622949e-01 5.81622899e-01 5.28677166e-01 8.24323833e-01
7.79698193e-02 -4.02943529e-02 -7.77333558e-01 -2.39642441e-01
-3.42204094e-01 -4.38586891e-01 6.61126971e-01 6.75427258e-01
-3.47963780e-01 -1.37218451e+00 -1.07484436e+00 -3.78956050e-01
2.34475240e-01 8.14680457e-02 -3.15004975e-01 1.13828063e+00
-1.76295247e-02 2.69067496e-01 3.34454149e-01 -7.62728870e-01
4.97871578e-01 -1.03977509e-01 7.25657225e-01 -1.21928640e-01
-5.37046552e-01 3.46407652e-01 -1.57136783e-01 -7.93507636e-01
-4.14514482e-01 -6.02799416e-01 -8.24987292e-01 -7.20574081e-01
-2.06874669e-01 -2.09981829e-01 5.70227265e-01 7.42809892e-01
1.24746531e-01 8.17242786e-02 9.99367476e-01 -1.22627401e+00
-5.22609055e-01 -8.68857026e-01 -7.20139682e-01 5.11295319e-01
1.39819548e-01 -4.16827679e-01 -1.08046629e-01 2.14085415e-01] | [9.394255638122559, -2.3560335636138916] |
d377d624-5b93-4939-87f4-ffd0b0c4c867 | online-dictionary-learning-based-fault-and | 2108.10990 | null | https://arxiv.org/abs/2108.10990v1 | https://arxiv.org/pdf/2108.10990v1.pdf | Online Dictionary Learning Based Fault and Cyber Attack Detection for Power Systems | The emerging wide area monitoring systems (WAMS) have brought significant improvements in electric grids' situational awareness. However, the newly introduced system can potentially increase the risk of cyber-attacks, which may be disguised as normal physical disturbances. This paper deals with the event and intrusion detection problem by leveraging a stream data mining classifier (Hoeffding adaptive tree) with semi-supervised learning techniques to distinguish cyber-attacks from regular system perturbations accurately. First, our proposed approach builds a dictionary by learning higher-level features from unlabeled data. Then, the labeled data are represented as sparse linear combinations of learned dictionary atoms. We capitalize on those sparse codes to train the online classifier along with efficient change detectors. We conduct numerical experiments with industrial control systems cyber-attack datasets. We consider five different scenarios: short-circuit faults, line maintenance, remote tripping command injection, relay setting change, as well as false data injection. The data are generated based on a modified IEEE 9-bus system. Simulation results show that our proposed approach outperforms the state-of-the-art method. | ['Yu Zhang', 'Gabriel Intriago'] | 2021-08-24 | null | null | null | null | ['cyber-attack-detection'] | ['miscellaneous'] | [ 2.71557361e-01 -3.62041503e-01 -1.30165204e-01 -1.69941559e-01
-4.50058877e-01 -4.63985801e-01 3.93388510e-01 7.75361657e-01
2.45760903e-01 7.97765493e-01 -2.80137837e-01 -4.27561402e-01
-3.36525679e-01 -8.24498951e-01 -4.00170356e-01 -9.03544664e-01
-8.12443793e-01 1.64513469e-01 1.36819050e-01 -2.58382738e-01
3.60974133e-01 5.99995017e-01 -1.21497023e+00 2.49542445e-01
8.45773757e-01 1.12831783e+00 -3.22274029e-01 3.09521914e-01
3.28056902e-01 6.37350023e-01 -1.24373138e+00 3.11097026e-01
4.53918070e-01 -4.40346032e-01 -3.30319971e-01 4.09589797e-01
-5.35867512e-01 -1.94284692e-01 -3.61024350e-01 1.29017830e+00
3.86583060e-01 1.12729803e-01 3.15167248e-01 -1.87558901e+00
7.87915140e-02 6.47048354e-01 -8.62264752e-01 8.74608815e-01
4.86656129e-01 2.25988358e-01 4.44014281e-01 -5.81535637e-01
1.70309082e-01 8.56386483e-01 4.44188416e-01 -2.80188650e-01
-9.71018136e-01 -1.01964664e+00 2.55682439e-01 7.46099710e-01
-1.36713898e+00 -1.86828692e-02 9.93798435e-01 -1.69361681e-01
1.09492505e+00 2.26990417e-01 5.82597077e-01 6.98881090e-01
6.71125352e-01 5.57181180e-01 9.21755910e-01 -3.49987805e-01
7.29728997e-01 -5.74804656e-03 4.59226847e-01 2.52375990e-01
5.70623338e-01 3.46830577e-01 -2.76121378e-01 -5.48807204e-01
3.70203406e-01 3.18701178e-01 -5.32405794e-01 -9.26728994e-02
-7.91801512e-01 7.23124266e-01 1.11541294e-01 5.30109107e-01
-5.15848398e-01 -6.47211313e-01 7.55044043e-01 6.60998523e-01
5.37136197e-03 2.08545044e-01 -6.47750735e-01 -2.52642035e-01
-6.84826255e-01 -2.49907643e-01 9.05335367e-01 1.05236065e+00
4.45799112e-01 8.84754360e-01 5.04098594e-01 2.24786401e-01
2.41655950e-02 2.56236911e-01 7.82504559e-01 8.74949545e-02
5.10959327e-01 5.96858025e-01 -1.09112546e-01 -9.80888784e-01
-5.12157440e-01 -6.15851402e-01 -1.05377340e+00 2.12870494e-01
9.59593989e-03 -5.53546429e-01 -5.73190570e-01 9.77449059e-01
3.90551329e-01 7.94542789e-01 1.33041158e-01 4.02331680e-01
-2.10990515e-02 9.43859100e-01 -2.14890257e-01 -9.05911624e-01
1.03365064e+00 -3.39454889e-01 -1.23102379e+00 2.72572309e-01
7.57002473e-01 -4.46745098e-01 3.87912691e-01 1.06290638e+00
-5.92204630e-01 -3.64522904e-01 -1.51361179e+00 1.19459450e+00
-4.96380836e-01 -1.83522999e-01 3.46192867e-01 7.73636997e-01
-2.45713621e-01 5.61706841e-01 -8.38932991e-01 -2.31665969e-01
2.64703363e-01 2.79520005e-01 -1.12374134e-01 -6.50029927e-02
-1.43650317e+00 9.76159811e-01 7.16968894e-01 1.19491249e-01
-1.18412220e+00 -6.84734166e-01 -8.64991426e-01 8.17022566e-03
5.79833090e-01 2.81400234e-01 8.48637521e-01 -5.92005670e-01
-1.09418893e+00 -9.75114629e-02 3.68167311e-01 -8.97681773e-01
-9.64503959e-02 3.34814824e-02 -1.47466052e+00 2.13198751e-01
-1.68674380e-01 -7.80659080e-01 9.75546420e-01 -1.11131227e+00
-8.87843132e-01 -3.63102227e-01 -1.53396115e-01 -9.10294428e-02
-4.00021583e-01 5.90711599e-03 7.38910615e-01 -7.93348134e-01
-4.74557169e-02 -5.21392524e-01 -4.84455109e-01 -7.05261767e-01
-7.09517300e-01 1.57008260e-01 1.52030182e+00 -5.49151838e-01
1.67688489e+00 -2.30522037e+00 -3.34506363e-01 8.52230906e-01
-3.62092882e-01 4.10339653e-01 3.15581709e-01 9.93664801e-01
-7.99029410e-01 -4.27698761e-01 -3.92186642e-01 3.08491439e-01
-2.41213888e-01 3.34791601e-01 -6.30455613e-01 8.87920737e-01
7.07107335e-02 2.38338690e-02 -8.10531318e-01 6.59416467e-02
5.26377499e-01 -1.43254906e-01 -2.62408946e-02 1.88333854e-01
1.69672713e-01 2.62107015e-01 -4.58595216e-01 8.69819999e-01
7.00274408e-01 9.20518339e-02 1.90061122e-01 -3.74543846e-01
9.16420296e-03 -9.45419073e-02 -1.83661914e+00 1.17899263e+00
-2.56073862e-01 2.16937765e-01 -1.09554462e-01 -1.76987028e+00
9.18240190e-01 5.48263252e-01 6.60050750e-01 -4.26262975e-01
3.85062605e-01 -1.54198900e-01 4.12624180e-02 -2.51900226e-01
-2.17065707e-01 1.43808052e-02 -3.00103009e-01 5.09012878e-01
2.47108221e-01 2.94265319e-02 2.27000773e-01 3.50748986e-01
1.35137832e+00 -6.34369314e-01 8.52747142e-01 -4.18451935e-01
8.32968116e-01 1.27715558e-01 9.27059770e-01 3.55850428e-01
5.24642840e-02 -2.52223611e-01 4.69962358e-01 -4.32778537e-01
-3.77972156e-01 -9.74029839e-01 -3.99645925e-01 2.16655239e-01
2.12678611e-01 -4.83843148e-01 -4.05762225e-01 -9.20403421e-01
2.29522288e-01 1.25847208e+00 -2.53246933e-01 -7.53979862e-01
-2.99170792e-01 -1.23618937e+00 5.11046529e-01 4.00260359e-01
3.34522456e-01 -7.20326781e-01 -5.69718599e-01 6.72454834e-01
6.20557070e-01 -9.45171535e-01 -3.92254209e-03 8.48596454e-01
-6.84532464e-01 -1.56150258e+00 1.23003393e-01 -7.04315782e-01
8.52037370e-01 -6.39072387e-03 6.16654575e-01 -5.24709411e-02
-8.13902497e-01 3.13748091e-01 -5.72595835e-01 -4.82788771e-01
-4.23218071e-01 -5.30950308e-01 7.43100226e-01 3.12043726e-01
4.25816625e-01 -7.50578165e-01 -1.59494475e-01 3.29871982e-01
-1.07122314e+00 -8.05391490e-01 3.62216711e-01 8.89611244e-01
3.36748302e-01 1.35313809e+00 1.20605612e+00 -8.83667409e-01
7.12981880e-01 -1.00055838e+00 -7.95111477e-01 1.59660935e-01
-1.03168738e+00 -2.50376850e-01 1.20424747e+00 -5.25447071e-01
-9.73585427e-01 1.68192629e-02 1.95969179e-01 -3.01197737e-01
-5.71053982e-01 6.42648995e-01 -5.64077735e-01 -1.79012164e-01
4.10384834e-01 4.67067868e-01 -1.79716006e-01 -3.79917711e-01
3.85562293e-02 7.94210076e-01 5.20529151e-01 -3.80897790e-01
1.39486682e+00 3.20901632e-01 9.88843758e-03 -9.87041056e-01
-3.17335635e-01 -4.69934285e-01 -4.07512665e-01 -2.54325360e-01
4.83348854e-02 -9.18489635e-01 -6.57642007e-01 5.23586810e-01
-5.66954911e-01 2.36907408e-01 -3.50482792e-01 4.96960670e-01
-1.34329066e-01 6.49182379e-01 -7.10375965e-01 -7.13531852e-01
-1.38099745e-01 -8.78061831e-01 4.37945396e-01 1.42051443e-01
-9.19666737e-02 -1.08322990e+00 7.66284391e-02 -2.83063859e-01
9.86170694e-02 7.03432024e-01 9.97691274e-01 -1.28059268e+00
-2.22413421e-01 -7.18820632e-01 6.10271633e-01 5.41870713e-01
7.06946611e-01 -1.20146707e-01 -4.57585692e-01 -9.32893932e-01
6.53000295e-01 -8.87993872e-02 -1.16488405e-01 3.42411175e-02
1.09720182e+00 -5.00084043e-01 -4.67840493e-01 4.67219472e-01
1.58565974e+00 9.90392089e-01 3.30782235e-01 2.65335768e-01
3.59912395e-01 7.91604072e-03 1.06134152e+00 1.23871458e+00
-9.55952853e-02 1.52492911e-01 4.83298898e-01 -6.64009829e-04
9.42579210e-01 -3.89648345e-03 4.28115845e-01 1.01632535e+00
6.20873868e-01 -7.11078271e-02 -7.49687731e-01 3.50761712e-01
-1.50268483e+00 -9.05464292e-01 1.27179921e-01 2.03794479e+00
6.58745110e-01 6.19823813e-01 -7.51087889e-02 1.24221218e+00
7.90310085e-01 -2.91955601e-02 -8.96923959e-01 -3.12514782e-01
-1.00195162e-01 4.07570750e-01 3.79420042e-01 1.78942531e-01
-9.93697286e-01 2.45329872e-01 4.79450178e+00 7.24090219e-01
-9.40617979e-01 -1.07719809e-01 3.43139708e-01 1.29990116e-01
3.50126773e-01 -4.83724736e-02 -7.64386833e-01 6.05177462e-01
1.42575443e+00 -7.94518650e-01 1.75253958e-01 8.25186729e-01
3.44986677e-01 -1.36792377e-01 -1.21131790e+00 1.05189490e+00
2.73255527e-01 -9.43538845e-01 1.17375977e-01 -2.38402560e-01
6.88508451e-01 -4.08217102e-01 -3.64318371e-01 2.84734279e-01
2.05098823e-01 -5.91546714e-01 2.33145207e-01 1.99413866e-01
3.78228098e-01 -1.26872063e+00 8.21151853e-01 5.48246384e-01
-1.34780085e+00 -6.15995944e-01 5.49328327e-03 -9.79230255e-02
5.46460032e-01 8.38904500e-01 -6.74599648e-01 9.64765370e-01
5.43973088e-01 1.07606852e+00 -2.23163679e-01 1.03546703e+00
4.90780324e-02 1.03497875e+00 -4.95112538e-01 1.27411216e-01
3.94632332e-02 -7.95950294e-02 4.99397337e-01 8.11929107e-01
1.19898535e-01 5.64210415e-01 6.73521757e-01 2.32487366e-01
4.58382010e-01 -1.41789645e-01 -6.69191420e-01 1.80252984e-01
8.70470941e-01 1.02786398e+00 -7.86840618e-01 -4.35702085e-01
-5.39780617e-01 4.63266790e-01 -5.84554791e-01 3.45562607e-01
-8.83189201e-01 -8.98614645e-01 5.24337709e-01 -1.33847758e-01
1.73819944e-01 -6.79135919e-02 -2.28950322e-01 -1.13009834e+00
-1.86682478e-01 -1.17354620e+00 7.28887737e-01 -3.86052310e-01
-1.55733252e+00 4.66114819e-01 1.41788200e-01 -1.81069589e+00
-3.23758483e-01 -9.20342654e-02 -9.18380022e-01 4.38827068e-01
-1.21825063e+00 -2.35332623e-01 -7.67375976e-02 1.30069613e+00
7.09799409e-01 -5.05551696e-01 9.56098735e-01 4.52914387e-01
-9.64765906e-01 5.19499421e-01 3.19268167e-01 2.08355397e-01
1.62402108e-01 -9.71960008e-01 3.00096162e-02 1.15305221e+00
-5.44856936e-02 2.97075093e-01 7.96320498e-01 -7.27229297e-01
-1.48544061e+00 -1.07701075e+00 -4.03891914e-02 3.53954583e-01
1.11470377e+00 -2.43261039e-01 -1.04400754e+00 7.83994496e-01
4.40923899e-01 2.96882480e-01 9.50135887e-01 -4.56067234e-01
1.42860264e-01 -3.14449936e-01 -1.56761754e+00 2.18471110e-01
2.70300418e-01 -3.80392104e-01 -1.01904571e+00 5.56329727e-01
2.68894494e-01 -2.49315023e-01 -1.19898057e+00 3.73891532e-01
-3.85137796e-01 -5.05579412e-01 7.42015004e-01 -7.06207275e-01
-4.16829050e-01 -6.11993134e-01 -5.11106476e-02 -1.82976711e+00
-2.21962705e-01 -9.37330961e-01 -4.02048141e-01 1.05350149e+00
1.49789140e-01 -9.52510238e-01 4.14369583e-01 9.62759182e-03
-2.31263831e-01 -5.58352649e-01 -1.14690197e+00 -9.32779014e-01
-3.89655441e-01 -3.16353768e-01 8.85352254e-01 1.48566151e+00
1.08896530e+00 2.86706299e-01 -1.56265065e-01 8.54278147e-01
1.00911796e+00 -5.05503714e-02 2.12710917e-01 -1.00624454e+00
-1.35537207e-01 1.94883585e-01 -8.09836984e-01 -3.30910981e-01
-4.31429641e-03 -4.56593215e-01 -2.85613090e-01 -8.02275360e-01
-5.09098887e-01 -3.44343573e-01 -7.37120330e-01 5.68516195e-01
7.24757984e-02 -4.73968357e-01 -1.80152565e-01 -2.02406734e-01
-4.33439642e-01 5.82434654e-01 3.50596339e-01 -2.26849571e-01
-6.81363046e-02 3.25770557e-01 -6.16260096e-02 6.55396223e-01
1.02371204e+00 -5.16380310e-01 -6.52969956e-01 2.59001285e-01
-4.08289611e-01 7.23937452e-01 -1.14657886e-01 -1.48534393e+00
4.25554097e-01 -6.52216151e-02 2.88288474e-01 -8.37832391e-01
-1.65275186e-01 -1.43619132e+00 1.28900707e-01 8.84301066e-01
6.63631409e-03 5.24382770e-01 4.19320047e-01 8.71721864e-01
-5.19855738e-01 -1.24279805e-01 8.10114264e-01 3.22899848e-01
-6.50020301e-01 2.86634684e-01 -8.72707963e-01 -1.26272231e-01
1.74151671e+00 -2.01545641e-01 2.29290174e-03 -1.57695740e-01
-8.88001621e-01 4.82783616e-01 -7.13546053e-02 5.38421333e-01
8.65619063e-01 -1.26994288e+00 -3.32160860e-01 6.92750275e-01
5.70990406e-02 -2.40477651e-01 2.32577965e-01 4.55532700e-01
-7.52056614e-02 1.92721725e-01 -2.36387745e-01 -4.84025300e-01
-1.07805204e+00 8.33130419e-01 1.22730754e-01 -6.27684295e-02
-8.15412045e-01 1.95834875e-01 -7.57808506e-01 2.11316645e-01
5.03713548e-01 -5.47140658e-01 -3.60939473e-01 8.10127258e-02
6.14439309e-01 6.00882351e-01 4.01554465e-01 -1.32912233e-01
-5.08359432e-01 -2.02479735e-02 -2.76172131e-01 4.48956668e-01
1.31999516e+00 9.95585620e-02 4.82198708e-02 6.53017223e-01
9.59521115e-01 -3.72543275e-01 -8.65535915e-01 -3.08875233e-01
2.40180060e-01 -4.26856101e-01 8.39112401e-02 -7.98757076e-01
-1.36324131e+00 4.61759180e-01 7.66403615e-01 6.89933896e-01
1.36990881e+00 -6.50031745e-01 7.30182648e-01 4.59783256e-01
1.02553630e+00 -9.96866286e-01 -1.44528657e-01 1.73215061e-01
3.31343561e-01 -8.67803931e-01 1.55645549e-01 -1.56245351e-01
-5.88824034e-01 1.07795143e+00 5.26287019e-01 -4.62778807e-01
1.38185084e+00 1.00792539e+00 -2.82005876e-01 -5.30795977e-02
-9.99955058e-01 4.11742181e-01 -5.34706593e-01 7.49849379e-01
-3.99969012e-01 9.28229541e-02 -1.02806874e-01 7.65375197e-01
6.25957325e-02 -6.78098500e-02 9.72957432e-01 1.50325763e+00
-3.52928102e-01 -8.22014451e-01 -7.15218127e-01 8.18171442e-01
-3.60097021e-01 2.20000237e-01 2.17805475e-01 7.07442939e-01
-8.59812181e-03 1.11940396e+00 3.06402687e-02 -4.72186834e-01
7.21857965e-01 -1.57971814e-01 -5.86787350e-02 -5.98759234e-01
-5.32342255e-01 -1.16605744e-01 -9.16768536e-02 -4.60358143e-01
1.32453606e-01 -9.29044485e-01 -1.51375079e+00 3.25096101e-02
-6.94465101e-01 7.76336133e-01 3.98329437e-01 8.97309721e-01
-3.37905250e-02 8.68094563e-01 1.32409048e+00 -4.74424392e-01
-1.07769525e+00 -9.99405205e-01 -1.33107054e+00 2.91943163e-01
3.33592713e-01 -8.44132662e-01 -9.53576684e-01 3.33739668e-02] | [6.189124584197998, 2.5547306537628174] |
1bbb47af-d740-4417-b598-12d9f33b276f | pk-icr-persona-knowledge-interactive-context | 2302.06674 | null | https://arxiv.org/abs/2302.06674v1 | https://arxiv.org/pdf/2302.06674v1.pdf | PK-ICR: Persona-Knowledge Interactive Context Retrieval for Grounded Dialogue | Identifying relevant Persona or Knowledge for conversational systems is a critical component of grounded dialogue response generation. However, each grounding has been studied in isolation with more practical multi-context tasks only recently introduced. We define Persona and Knowledge Dual Context Identification as the task to identify Persona and Knowledge jointly for a given dialogue, which could be of elevated importance in complex multi-context Dialogue settings. We develop a novel grounding retrieval method that utilizes all contexts of dialogue simultaneously while also requiring limited training via zero-shot inference due to compatibility with neural Q \& A retrieval models. We further analyze the hard-negative behavior of combining Persona and Dialogue via our novel null-positive rank test. | ['Guoyin Wang', 'Jiwei Li', 'Joosung Lee', 'Minsik Oh'] | 2023-02-13 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [ 3.21415931e-01 5.16586721e-01 6.10970110e-02 -2.13650122e-01
-1.24530566e+00 -7.02634811e-01 1.17286742e+00 1.18864991e-01
-4.53519583e-01 1.10060263e+00 6.57515407e-01 -4.83296067e-02
-4.43833917e-01 -6.22612715e-01 -1.35302976e-01 -4.48156059e-01
2.76129425e-01 7.56625175e-01 2.02305242e-01 -6.47579193e-01
2.05858126e-01 -1.94574669e-01 -1.38866425e+00 5.07844269e-01
9.29041862e-01 4.06027615e-01 1.12815171e-01 1.09112489e+00
-2.90325612e-01 9.84258115e-01 -1.00448430e+00 -8.22814941e-01
-6.31834120e-02 -9.52909350e-01 -1.44543171e+00 6.03645481e-03
4.64874536e-01 -3.81410360e-01 -3.27367485e-01 7.84838796e-01
9.91227269e-01 6.19805574e-01 4.71505135e-01 -1.04764009e+00
-3.77611578e-01 7.12963700e-01 6.95261732e-02 3.27498168e-01
1.06591320e+00 8.99737626e-02 1.25941527e+00 -8.34127665e-01
4.97607410e-01 1.47547650e+00 5.09740472e-01 9.17199671e-01
-1.15261900e+00 -1.76379398e-01 1.19765326e-01 6.71532303e-02
-1.19995689e+00 -6.73159897e-01 6.46358609e-01 -2.72384584e-01
1.10048902e+00 3.85822892e-01 5.07863164e-01 1.42783689e+00
-3.18966210e-01 8.64387035e-01 9.82247889e-01 -5.88495493e-01
5.38246669e-02 2.98492402e-01 6.02133334e-01 8.56873095e-01
-3.31661962e-02 -3.94417077e-01 -1.18094110e+00 -6.85496986e-01
4.19119596e-01 -4.20005411e-01 -2.02856153e-01 3.71562004e-01
-1.22708714e+00 9.74465907e-01 -2.35048220e-01 1.20378643e-01
-4.93288308e-01 7.25636855e-02 3.93722683e-01 5.53615391e-01
4.62362409e-01 7.73230374e-01 -8.78528282e-02 -5.93645811e-01
-5.45232058e-01 7.74426639e-01 1.43500400e+00 1.05897868e+00
7.31141925e-01 -3.13920915e-01 -8.36005211e-01 1.14394712e+00
1.05586387e-02 3.86339664e-01 4.34588552e-01 -9.72229898e-01
5.00958145e-01 6.55641258e-01 4.79885936e-01 -7.70881712e-01
-4.51219082e-01 -3.13870125e-02 -4.25719678e-01 -7.20019281e-01
6.01534784e-01 -5.93168736e-01 -2.07674559e-02 1.98016655e+00
5.32540917e-01 -1.78574286e-02 4.91576940e-01 7.82303572e-01
1.13015091e+00 5.91471136e-01 1.87354505e-01 -4.46081340e-01
1.71297812e+00 -7.92528629e-01 -1.03811896e+00 -1.98188856e-01
5.37233531e-01 -6.47155225e-01 1.29039431e+00 1.34140447e-01
-1.32722962e+00 -2.23776802e-01 -8.44039083e-01 -2.78828770e-01
-1.99318036e-01 -1.29320890e-01 6.62114441e-01 7.93105543e-01
-9.70423400e-01 2.39370257e-01 9.76173021e-03 -6.43408835e-01
-2.08274812e-01 7.99588263e-02 7.16666654e-02 7.00872615e-02
-1.91453862e+00 1.16170275e+00 3.61347198e-02 1.73234403e-01
-7.30010569e-01 -2.46819243e-01 -6.91016138e-01 -7.95432553e-02
1.00878441e+00 -1.14485335e+00 1.54381418e+00 -5.05504072e-01
-1.91766179e+00 7.96371698e-01 -3.71303111e-01 -3.46240371e-01
4.29929376e-01 -5.75067401e-01 -3.99627477e-01 4.17376697e-01
2.49279827e-01 3.23946089e-01 6.83397532e-01 -1.01469481e+00
-6.06404543e-01 -2.22148746e-01 6.42969251e-01 8.84431243e-01
-3.22201431e-01 2.45797887e-01 -3.30511421e-01 -2.78096646e-01
-1.86371312e-01 -8.67325604e-01 -4.61347811e-02 -6.30190134e-01
-4.95504707e-01 -6.91678286e-01 3.63300771e-01 -6.50954008e-01
1.29619002e+00 -1.59119058e+00 1.15483098e-01 6.39261901e-02
2.61123955e-01 7.08403438e-02 -6.31171167e-02 8.71228039e-01
6.39653921e-01 1.12444252e-01 3.62976156e-02 -2.03472629e-01
2.09514678e-01 -6.60434887e-02 -4.22963619e-01 9.58701298e-02
2.87233055e-01 9.78912234e-01 -1.32114851e+00 -6.72357142e-01
-1.86301783e-01 1.51480898e-01 -3.22832704e-01 4.74299222e-01
-3.05646896e-01 5.61082922e-02 -5.11035144e-01 4.35113013e-01
-2.46693984e-01 -3.38718832e-01 4.83959764e-01 1.01145633e-01
2.69011647e-01 5.44803023e-01 -9.38736737e-01 1.69372320e+00
-4.66144532e-01 4.90893215e-01 2.02398241e-01 -4.52539951e-01
7.10187495e-01 8.95510077e-01 2.11966738e-01 -5.15385866e-01
-8.53010267e-02 -1.22609563e-01 8.32439885e-02 -6.99693263e-01
1.17088270e+00 -4.97528195e-01 -4.81477350e-01 8.04636300e-01
1.70798689e-01 -1.66731030e-02 1.14856986e-02 6.28530502e-01
1.03895640e+00 -7.62675107e-02 3.79954666e-01 -3.11192930e-01
3.17482203e-01 -5.92089724e-04 3.02335739e-01 1.25468695e+00
-4.33580875e-01 2.18520448e-01 5.73085070e-01 2.43916854e-01
-6.07053280e-01 -8.45273197e-01 8.48873705e-02 1.70518553e+00
2.42780805e-01 -4.78114039e-01 -7.39142895e-01 -6.27462089e-01
-8.77671987e-02 8.45035195e-01 -4.23439503e-01 -2.15096191e-01
-3.87798160e-01 -5.78630328e-01 9.58407581e-01 1.81691036e-01
4.62205380e-01 -1.19458985e+00 -4.85557616e-01 3.04106206e-01
-6.52682781e-01 -1.18980634e+00 -4.97866899e-01 -8.97695124e-02
-4.29406077e-01 -1.18345022e+00 -8.01656187e-01 -6.59868300e-01
6.21086359e-02 4.05841380e-01 1.40615904e+00 2.26784974e-01
1.13559321e-01 1.11200225e+00 -4.49016929e-01 -1.62949115e-01
-5.08776546e-01 1.55065626e-01 2.87913501e-01 3.71749550e-02
4.33404535e-01 -1.39864624e-01 -5.37861228e-01 3.67960989e-01
-4.93924946e-01 9.50925797e-02 1.32515907e-01 1.08356154e+00
-1.14444785e-01 -1.91662505e-01 1.36363423e+00 -1.11130321e+00
1.62345707e+00 -6.31070375e-01 5.66599965e-02 5.94008207e-01
-4.31350261e-01 -2.65212595e-01 2.01917693e-01 -3.23228210e-01
-1.39562058e+00 -5.39874673e-01 6.01452105e-02 2.47495666e-01
6.21495545e-02 5.84320307e-01 -2.35319585e-01 3.14844131e-01
8.52411985e-01 1.46215692e-01 -1.67819895e-02 -1.20685741e-01
6.06260121e-01 9.03425157e-01 3.23083818e-01 -1.19217110e+00
3.41922909e-01 7.45113790e-02 -4.36842263e-01 -1.30625474e+00
-1.04203975e+00 -9.59265769e-01 -3.94839227e-01 -5.82429707e-01
9.13722038e-01 -9.02073622e-01 -1.12892568e+00 1.88114673e-01
-1.42841113e+00 -4.33815777e-01 -1.20412797e-01 1.39767885e-01
-5.38794577e-01 6.99055791e-01 -5.55995107e-01 -1.38673389e+00
-6.12069726e-01 -6.98827624e-01 9.29029405e-01 2.88616598e-01
-7.85535455e-01 -9.59828675e-01 1.82935297e-01 7.23981261e-01
3.68879624e-02 -2.07758173e-01 8.38546216e-01 -1.19849157e+00
-2.22288445e-01 -2.54470378e-01 -1.08783906e-02 5.41522987e-02
4.03553583e-02 -5.77218473e-01 -1.13232195e+00 -2.12299094e-01
1.12397179e-01 -7.75838315e-01 5.51895201e-01 -1.68387339e-01
4.21771944e-01 -6.81373894e-01 3.67391594e-02 -4.09387618e-01
7.70526528e-01 1.21161290e-01 3.30540270e-01 -1.32538319e-01
5.62499821e-01 1.15325618e+00 6.83328927e-01 5.14926910e-01
5.87512732e-01 7.37813890e-01 -4.47150260e-01 2.86943913e-01
1.15156539e-01 -2.69522041e-01 3.56579214e-01 7.72326410e-01
-7.14554787e-02 -4.25915599e-01 -9.73095834e-01 7.21140027e-01
-2.10484791e+00 -1.33665121e+00 1.46651372e-01 2.08827806e+00
1.31924617e+00 -4.53777164e-02 4.70877260e-01 -2.74511099e-01
1.04745936e+00 1.97798654e-01 -3.62731397e-01 -1.41902059e-01
-2.23356947e-01 -3.41063216e-02 -2.70104796e-01 9.17601645e-01
-6.70074999e-01 9.72521424e-01 6.91292095e+00 5.54443717e-01
-2.80598730e-01 3.23013842e-01 3.48214984e-01 -2.42671639e-01
-5.50674915e-01 3.33094969e-02 -8.99186373e-01 1.41554862e-01
1.00970638e+00 -5.54422081e-01 4.11345333e-01 5.34832478e-01
6.26508072e-02 -2.09317818e-01 -1.33933318e+00 7.73610771e-01
1.98575780e-01 -1.04715872e+00 1.74942389e-01 -7.58198127e-02
7.19009757e-01 -5.24445117e-01 -3.82856667e-01 7.41248190e-01
5.86901844e-01 -7.54655123e-01 3.00460666e-01 7.78958380e-01
5.78759193e-01 -5.67854941e-01 5.62261283e-01 4.33618009e-01
-7.00493217e-01 5.10829352e-02 -1.59325540e-01 -2.92238314e-02
1.99209854e-01 1.05548330e-01 -1.32235026e+00 5.62734306e-01
1.43776447e-01 5.02854548e-02 -3.57876658e-01 4.17877674e-01
-1.04451276e-01 5.87988853e-01 -9.29949731e-02 -6.27060473e-01
1.85437873e-01 2.62031313e-02 8.10336232e-01 1.41055942e+00
-1.55616358e-01 5.64639390e-01 4.96303052e-01 7.52401054e-01
-9.08022299e-02 2.54185557e-01 -6.93033755e-01 -1.14546470e-01
8.73188436e-01 1.09912431e+00 -4.75989163e-01 -6.79448664e-01
-3.91986847e-01 1.08687663e+00 3.15775931e-01 5.65005302e-01
-3.55577081e-01 -3.52607995e-01 5.06732702e-01 -1.44843623e-01
-3.34737927e-01 -2.69798171e-02 5.03163263e-02 -1.09051490e+00
-1.27850324e-01 -9.43071067e-01 5.71667910e-01 -5.70395291e-01
-1.64152682e+00 3.51831138e-01 2.19207868e-01 -7.50434995e-01
-7.74436057e-01 -3.16695094e-01 -7.13098109e-01 1.05870974e+00
-1.29722238e+00 -9.82960820e-01 -2.92924792e-01 7.53645241e-01
7.10086465e-01 -3.74764204e-01 1.14001095e+00 -1.70432612e-01
-7.03325808e-01 7.02158511e-01 -1.70572326e-01 1.09674484e-01
9.28499818e-01 -1.44264495e+00 1.40400887e-01 6.03823483e-01
-1.48105755e-01 1.25057685e+00 7.09906816e-01 -7.61668861e-01
-1.54651570e+00 -6.95026517e-01 1.12781060e+00 -8.27615559e-01
5.72724819e-01 -3.76094133e-01 -9.62595284e-01 3.36139947e-01
5.48842192e-01 -8.64150643e-01 1.19559908e+00 9.26805019e-01
-1.09243676e-01 1.64590329e-01 -9.78785276e-01 9.55954909e-01
9.75415587e-01 -1.26394546e+00 -1.02379620e+00 4.75321501e-01
8.34835887e-01 -1.70078784e-01 -8.93319309e-01 -1.17386408e-01
4.23104197e-01 -7.31823206e-01 9.23050880e-01 -7.40048468e-01
3.16837728e-01 -1.31252119e-02 -2.25754216e-01 -1.02438509e+00
-1.18744358e-01 -1.30511665e+00 -4.62430477e-01 1.37148607e+00
2.93799698e-01 -4.81415510e-01 5.71780801e-01 1.02530408e+00
-6.51181936e-02 -3.69742602e-01 -1.05028915e+00 -6.02137983e-01
-7.71037564e-02 -1.65186763e-01 5.14755189e-01 1.15695369e+00
7.97502577e-01 1.20740271e+00 -6.08264685e-01 1.29148792e-02
3.66526574e-01 7.92688578e-02 8.82932723e-01 -1.16922176e+00
-4.31783140e-01 -4.36160505e-01 1.78837478e-01 -1.05658793e+00
3.33867282e-01 -7.66301811e-01 4.55072165e-01 -1.60767055e+00
5.44953167e-01 -9.82886180e-03 -2.33490281e-02 2.93606967e-01
-7.85109222e-01 -1.42761499e-01 2.79190317e-02 2.66194552e-01
-1.08122957e+00 7.74746239e-01 1.02216613e+00 2.41118832e-03
-3.52232903e-01 4.15048823e-02 -9.84129786e-01 4.31689143e-01
5.37979424e-01 -1.28157260e-02 -8.72645855e-01 6.26302809e-02
4.52636033e-01 6.60049736e-01 5.72775066e-01 -5.51626623e-01
4.69341457e-01 -1.82716504e-01 -2.27187619e-01 -3.67458701e-01
5.60719609e-01 -2.04547375e-01 -4.60908175e-01 -1.21157408e-01
-8.33920121e-01 -5.25520444e-02 -3.17792483e-02 9.40159440e-01
-6.50800765e-02 -5.73231161e-01 2.16903180e-01 -4.37481225e-01
-7.42251039e-01 -1.18215166e-01 -6.45686269e-01 5.28890073e-01
6.56079590e-01 -2.10848838e-01 -4.42104071e-01 -9.38465536e-01
-6.98390245e-01 3.25022042e-01 5.52209355e-02 3.71731967e-01
7.28835523e-01 -1.13271081e+00 -9.42661345e-01 -5.11560321e-01
3.84750158e-01 -1.89905480e-01 4.31640625e-01 3.95325452e-01
1.69933781e-01 4.64465797e-01 2.78180122e-01 -3.37043166e-01
-1.23797262e+00 1.91010475e-01 3.45315903e-01 -1.73309386e-01
-3.94987673e-01 7.23267376e-01 7.97288939e-02 -3.35738927e-01
3.54033679e-01 4.22468245e-01 -4.61869538e-01 4.53697145e-01
7.00926244e-01 4.61827427e-01 -1.63718015e-01 -4.89056855e-01
-1.01217419e-01 -1.75042093e-01 -2.17387632e-01 -7.06702590e-01
6.76387310e-01 -4.37406927e-01 -8.49908665e-02 9.19426799e-01
7.93956339e-01 -5.31223491e-02 -9.28138912e-01 -6.31015003e-01
1.18275091e-01 -2.57759273e-01 -1.99710861e-01 -1.00693893e+00
3.95050310e-02 5.17829180e-01 1.92102820e-01 6.34653211e-01
4.82392937e-01 7.87125826e-02 6.53482497e-01 1.07706845e+00
2.93042868e-01 -1.45501351e+00 6.07158661e-01 7.82552660e-01
1.08234406e+00 -1.29235625e+00 5.74062094e-02 -2.28025645e-01
-1.02661800e+00 7.12778389e-01 8.20176780e-01 4.52233970e-01
2.33999640e-01 -3.66111368e-01 4.98940572e-02 -4.85829890e-01
-9.07131851e-01 -4.15158749e-01 2.65751183e-01 3.90953004e-01
6.03880346e-01 -1.41487103e-02 -4.76959974e-01 1.06482220e+00
-2.06365079e-01 -4.61381406e-01 5.09138525e-01 9.04693544e-01
-4.27301943e-01 -8.48937869e-01 -1.46245018e-01 4.80127871e-01
-3.54239285e-01 -4.07865912e-01 -9.10424829e-01 3.39319825e-01
-6.12790048e-01 1.48947632e+00 -3.28293860e-01 -4.89612669e-01
4.07298148e-01 6.49723887e-01 3.68357718e-01 -9.78181779e-01
-1.23546040e+00 -4.68731411e-02 9.53553140e-01 -2.68398106e-01
-5.19950271e-01 -7.57076502e-01 -9.30606484e-01 -3.61209840e-01
-5.96431732e-01 4.41946238e-01 1.40181169e-01 1.14916456e+00
2.22277060e-01 2.64012128e-01 5.53747535e-01 -3.70302796e-01
-7.74372876e-01 -1.24677575e+00 -4.34533179e-01 5.52029848e-01
5.00070266e-02 -5.60650110e-01 -1.38597503e-01 -1.32169247e-01] | [12.526223182678223, 8.10345458984375] |
f4cd8d1d-ff40-41ad-bc07-1fb0722eca9c | a-unified-model-for-arabizi-detection-and | null | null | https://aclanthology.org/2020.wanlp-1.15 | https://aclanthology.org/2020.wanlp-1.15.pdf | A Unified Model for Arabizi Detection and Transliteration using Sequence-to-Sequence Models | While online Arabic is primarily written using the Arabic script, a Roman-script variety called Arabizi is often seen on social media. Although this representation captures the phonology of the language, it is not a one-to-one mapping with the Arabic script version. This issue is exacerbated by the fact that Arabizi on social media is Dialectal Arabic which does not have a standard orthography. Furthermore, Arabizi tends to include a lot of code mixing between Arabic and English (or French). To map Arabizi text to Arabic script in the context of complete utterances, previously published efforts have split Arabizi detection and Arabic script target in two separate tasks. In this paper, we present the first effort on a unified model for Arabizi detection and transliteration into a code-mixed output with consistent Arabic spelling conventions, using a sequence-to-sequence deep learning model. Our best system achieves 80.6% word accuracy and 58.7% BLEU on a blind test set. | ['Nizar Habash', 'Aiza Usman', 'Ali Shazal'] | null | null | null | null | coling-wanlp-2020-12 | ['transliteration'] | ['natural-language-processing'] | [ 2.35780533e-02 -4.18686420e-01 6.82041422e-02 -4.33827907e-01
-8.30553889e-01 -1.16258526e+00 4.26661938e-01 -2.10346162e-01
-1.74056649e-01 2.69360662e-01 3.23223323e-01 -3.50086451e-01
4.57767248e-01 -6.37394905e-01 -4.39700603e-01 -4.28795636e-01
5.04112124e-01 9.18430507e-01 -2.64440496e-02 -9.13185835e-01
2.28308588e-01 3.01539540e-01 -9.83873188e-01 7.03003645e-01
1.06615210e+00 3.24105680e-01 6.29090816e-02 7.72204578e-01
-8.23772922e-02 9.00538146e-01 -8.34950387e-01 -8.01020801e-01
1.34173661e-01 -1.13632703e+00 -7.70335734e-01 1.57325994e-02
7.25675285e-01 -5.52350283e-01 -1.74745128e-01 1.25443721e+00
3.31850380e-01 -4.65008736e-01 9.98853564e-01 -6.86719775e-01
-1.13693893e+00 9.49963212e-01 -4.38425213e-01 -2.81970859e-01
6.79745853e-01 -5.36543503e-02 9.48556006e-01 -1.13408661e+00
7.19408333e-01 1.28627837e+00 8.15828264e-01 2.76466578e-01
-7.06523895e-01 -4.26549911e-01 -3.94145012e-01 -4.80266400e-02
-1.48324466e+00 -3.82737011e-01 1.02509007e-01 -6.42995000e-01
9.77708340e-01 2.20914602e-01 4.10615563e-01 8.76241028e-01
1.21116027e-01 6.06215835e-01 1.19081306e+00 -9.25819814e-01
-2.40436926e-01 -5.11424383e-03 -5.55286035e-02 1.05431533e+00
5.95677905e-02 -5.20720124e-01 -5.63223183e-01 4.52204257e-01
3.66269141e-01 -3.70202512e-01 -2.29012161e-01 1.56791896e-01
-1.19170558e+00 8.67522657e-01 2.68201262e-01 2.28650689e-01
-2.30850819e-02 -1.97547525e-01 5.15426755e-01 5.71153283e-01
-2.37728953e-02 3.22597444e-01 -2.18484834e-01 -3.42225134e-01
-1.08210742e+00 5.48884235e-02 9.84900057e-01 1.15481627e+00
4.56662625e-01 2.61876732e-01 4.37754214e-01 1.16013408e+00
7.88581610e-01 9.93639112e-01 8.05811644e-01 -5.17880678e-01
6.18688881e-01 5.17383277e-01 1.28360748e-01 -8.28703940e-01
-2.21523941e-01 1.47441626e-01 -2.17722639e-01 2.57196397e-01
1.10345352e+00 -3.23435426e-01 -1.02890372e+00 1.14993036e+00
-2.22879991e-01 -1.01392066e+00 5.16894683e-02 1.00357425e+00
4.50372398e-01 1.08493662e+00 -3.34353030e-01 1.55382559e-01
1.39698541e+00 -1.33606446e+00 -8.03142130e-01 -5.08593321e-01
6.26378775e-01 -1.50210166e+00 1.40117157e+00 5.61945796e-01
-1.16478276e+00 -2.78803468e-01 -1.32046950e+00 -2.25678489e-01
-5.55244386e-01 5.94144404e-01 -1.68056250e-01 1.38466072e+00
-1.09190226e+00 -1.42519770e-03 -7.47626245e-01 -9.14605141e-01
-1.71588019e-01 -5.34441210e-02 -3.52780908e-01 -1.36948377e-01
-1.13139141e+00 1.67773938e+00 4.21415150e-01 -7.12498929e-03
-8.25192928e-01 1.05537675e-01 -9.23470378e-01 -1.66325822e-01
-2.85232991e-01 3.86563271e-01 1.38363278e+00 -1.67237508e+00
-1.72003293e+00 1.10016596e+00 -7.17486665e-02 -1.55171588e-01
4.25595313e-01 -3.95648301e-01 -1.05642653e+00 5.41157881e-03
-1.88645229e-01 3.75825673e-01 1.00672507e+00 -1.09063220e+00
-5.64750075e-01 -2.51024902e-01 -2.13972777e-01 4.46425438e-01
-2.90200144e-01 9.09761429e-01 -4.76137102e-01 -9.18020725e-01
1.50651187e-01 -1.17978466e+00 6.01264834e-01 -1.16770379e-01
-1.54823558e-02 2.90113717e-01 5.17709315e-01 -1.60589337e+00
1.59721005e+00 -1.94289339e+00 1.61601886e-01 3.77641410e-01
-3.28417957e-01 5.79504073e-01 -1.91732809e-01 9.10742044e-01
1.14709251e-01 -9.20093805e-02 -5.84164202e-01 2.57044643e-01
1.91966921e-01 1.20252207e-01 -3.62615943e-01 6.83049858e-01
3.56919438e-01 6.11018836e-01 -8.75073910e-01 -1.47967815e-01
-3.30804825e-01 3.77821863e-01 -4.61365044e-01 -7.27998689e-02
-2.28129938e-01 -4.46210057e-02 4.32184726e-01 1.02576029e+00
6.91960633e-01 8.55123773e-02 6.94532096e-01 8.94110426e-02
-2.31952757e-01 5.38967431e-01 -8.52474451e-01 1.49548364e+00
-5.10000527e-01 9.50428903e-01 1.54388592e-01 -3.81507844e-01
8.14743102e-01 2.40412548e-01 -1.77362755e-01 -5.18923998e-01
4.00436729e-01 9.08614695e-01 4.10210431e-01 -2.85859555e-01
9.39017475e-01 1.52236640e-01 -2.23105460e-01 7.61298776e-01
1.80398226e-01 -3.57043684e-01 4.85957325e-01 3.52460891e-02
6.35351956e-01 3.94590080e-01 5.44571161e-01 -3.24221194e-01
4.48495865e-01 3.16127211e-01 -1.62708074e-01 5.66458344e-01
-1.50323316e-01 1.07232010e+00 3.04011047e-01 -1.91389248e-01
-1.02910662e+00 -1.19948661e+00 -6.17204607e-02 1.48328078e+00
-2.77636856e-01 -4.99942690e-01 -1.28702247e+00 -6.97717667e-01
-4.07707483e-01 7.62714863e-01 -2.65173197e-01 1.19293287e-01
-1.02954698e+00 -8.29258442e-01 1.15026546e+00 1.70082644e-01
3.00854057e-01 -1.07931387e+00 -2.70723701e-01 2.98029810e-01
-3.98065209e-01 -6.63580596e-01 -1.05911052e+00 1.04509987e-01
-5.73488735e-02 -9.71163392e-01 -8.61743927e-01 -1.40429461e+00
6.53823197e-01 -8.10840130e-02 1.07643998e+00 -1.01755403e-01
2.23609984e-01 1.41891509e-01 -8.83142531e-01 -3.97532403e-01
-1.31073725e+00 7.39115253e-02 7.10933423e-03 -9.75540504e-02
7.10973680e-01 4.24458653e-01 6.21734820e-02 1.97375327e-01
-9.41834629e-01 -3.77761692e-01 4.22352523e-01 6.86733842e-01
-2.12835208e-01 -5.37817180e-01 2.46049717e-01 -7.00042307e-01
3.82997751e-01 -4.66002464e-01 -5.46087205e-01 4.08188879e-01
-5.08450031e-01 -1.78951696e-01 6.79611862e-01 -2.33000219e-01
-9.67237473e-01 4.33729231e-01 -4.78863090e-01 5.89706540e-01
4.72153053e-02 8.52176845e-01 -1.58246517e-01 -6.31959438e-02
1.17916882e+00 3.29940170e-01 3.62611800e-01 -3.27039033e-01
5.68791091e-01 1.49450445e+00 4.99502599e-01 -2.24311605e-01
5.96762180e-01 5.95029779e-02 -8.76120567e-01 -8.99241149e-01
-4.20527607e-01 -3.28699589e-01 -8.05855215e-01 -2.64338672e-01
9.02639687e-01 -1.09388566e+00 -1.49732202e-01 1.16772866e+00
-1.22958553e+00 -3.45899403e-01 2.42061779e-01 1.79404661e-01
-3.48529488e-01 5.39363980e-01 -9.99451578e-01 -2.49722272e-01
-1.55891433e-01 -1.39863181e+00 6.97059989e-01 -1.18382670e-01
-6.62531853e-01 -6.28255010e-01 2.31534779e-01 2.01775491e-01
3.77154738e-01 -3.48377436e-01 8.88886094e-01 -6.21378839e-01
-1.10043794e-01 -2.62251914e-01 -3.84838969e-01 4.51219350e-01
4.95233446e-01 4.04521912e-01 -6.94177091e-01 -4.23252910e-01
-2.70488173e-01 -5.42656898e-01 4.04936403e-01 -2.78582752e-01
2.07578260e-02 -4.03471559e-01 5.40727019e-01 3.07445914e-01
1.33960736e+00 4.74720240e-01 6.10538960e-01 5.04361987e-01
7.83454418e-01 5.06932199e-01 4.47685450e-01 2.91224927e-01
4.90009487e-01 5.93541563e-01 2.34508052e-01 3.89424622e-01
-4.93435591e-01 -4.10296917e-02 1.39248073e+00 1.49553227e+00
4.16261375e-01 -3.85014296e-01 -1.49864030e+00 5.22798836e-01
-1.20700562e+00 -7.69492984e-01 -4.21762079e-01 2.10854197e+00
1.40561461e+00 -3.45571071e-01 2.96774864e-01 1.34901330e-02
7.31336355e-01 -5.12774028e-02 1.92616343e-01 -1.00457621e+00
-4.04424906e-01 2.06684470e-01 4.84606028e-01 7.93623626e-01
-1.11076307e+00 1.33276856e+00 6.54887199e+00 6.77718759e-01
-1.41744816e+00 1.10749647e-01 3.09203584e-02 3.34092230e-01
4.31511477e-02 -2.20095620e-01 -8.90119433e-01 6.09330535e-01
1.37950623e+00 2.44895339e-01 8.55826020e-01 4.96305972e-01
-2.84814179e-01 1.41038761e-01 -8.50588024e-01 7.69625604e-01
7.22921908e-01 -9.55524147e-01 3.40571195e-01 -4.77426231e-01
9.10719872e-01 4.09704775e-01 4.85132262e-02 1.19949423e-01
5.73507011e-01 -1.19588685e+00 1.54341042e+00 1.57153502e-01
1.19651151e+00 -6.57290101e-01 5.71271658e-01 -1.99196115e-01
-7.56775856e-01 3.32267731e-02 -8.07790682e-02 5.53778037e-02
-4.94941808e-02 -1.92547321e-01 -9.18445766e-01 8.50888863e-02
5.02273619e-01 6.88815057e-01 -9.03941393e-01 8.26617897e-01
-4.54338849e-01 9.53612685e-01 -9.43516344e-02 -2.40207165e-01
5.37749052e-01 -4.68385935e-01 3.94188076e-01 1.81872308e+00
7.18892455e-01 -5.83503783e-01 -6.60211295e-02 9.98091623e-02
-2.56909821e-02 5.41581511e-01 -3.90203834e-01 -6.14554048e-01
3.59914243e-01 8.47427130e-01 -9.87456977e-01 -1.23397477e-01
-7.83085048e-01 1.60735917e+00 8.17523673e-02 1.87324673e-01
-7.51838803e-01 -8.17476988e-01 4.72297549e-01 -1.39202580e-01
3.07397276e-01 -5.70594132e-01 -1.74557865e-01 -9.97167945e-01
-1.43147275e-01 -1.81822419e+00 3.24241340e-01 -9.48311865e-01
-1.18933094e+00 7.36817598e-01 -6.83079422e-01 -1.43791664e+00
-3.58475178e-01 -1.26199830e+00 1.05434181e-02 1.04834270e+00
-9.52834725e-01 -1.44371307e+00 -1.14864051e-01 3.41527492e-01
5.91268778e-01 -7.00253367e-01 1.15621948e+00 4.96498168e-01
-1.81925520e-01 8.08708251e-01 8.87079000e-01 7.35535383e-01
1.44081926e+00 -1.44997823e+00 5.78676045e-01 1.28804219e+00
3.69409710e-01 7.08530307e-01 6.17482722e-01 -5.96631289e-01
-1.31730556e+00 -1.01626754e+00 1.14569330e+00 -8.52857172e-01
1.00111628e+00 -3.74481708e-01 -7.50317812e-01 9.06822383e-01
8.18323791e-01 -7.40775943e-01 6.35062456e-01 -3.37211877e-01
-7.04364777e-01 1.45708397e-01 -9.24449146e-01 9.27559733e-01
3.57731730e-01 -7.98739970e-01 -5.01146257e-01 4.69962597e-01
3.05912137e-01 -6.45739317e-01 -5.51665246e-01 -3.63079548e-01
7.81411469e-01 -6.56000674e-01 5.08342683e-01 -3.19146127e-01
7.16580272e-01 -5.83562315e-01 -4.05531049e-01 -1.56596768e+00
-6.37329742e-02 -6.46776080e-01 2.32911140e-01 1.19083893e+00
5.37609994e-01 -2.49438852e-01 7.67793059e-02 -8.10492709e-02
-3.38386118e-01 3.36928070e-01 -5.40693402e-01 -6.94115102e-01
6.07019067e-01 -5.29482178e-02 4.64052826e-01 1.26808095e+00
2.15452060e-01 2.42423281e-01 -3.79986495e-01 9.58715081e-02
-1.77208990e-01 -2.86075711e-01 1.87464699e-01 -6.62712753e-01
-1.03592105e-01 -6.95435286e-01 -2.43725732e-01 -7.08130240e-01
-7.39441067e-02 -1.32083976e+00 5.50149441e-01 -1.35733795e+00
-4.14005011e-01 -4.36460413e-02 3.12476844e-01 4.74078804e-01
6.18829504e-02 7.38037467e-01 3.01018655e-01 1.91200256e-01
-2.36325234e-01 2.80068591e-02 7.51246691e-01 -3.19271713e-01
-1.05404757e-01 -3.88373464e-01 -4.35975015e-01 6.75578415e-01
1.01927996e+00 -4.78405923e-01 -2.08406765e-02 -9.85502362e-01
7.27574050e-01 -2.49293000e-01 -4.77932066e-01 -1.00586498e+00
4.03888188e-02 -7.39730545e-04 2.43699938e-01 -2.67253846e-01
-3.94564681e-02 -6.64486825e-01 -1.15185745e-01 6.42095208e-01
-1.72920167e-01 5.96357167e-01 1.42119110e-01 -2.30783522e-01
-2.70293415e-01 -6.88462138e-01 1.19409847e+00 -1.31002322e-01
-6.72368169e-01 -3.62857431e-01 -1.39835835e+00 1.45883620e-01
5.47440171e-01 -8.78811181e-02 -5.19680023e-01 -3.71745557e-01
-2.58884996e-01 -3.61265332e-01 9.15172160e-01 6.08039320e-01
3.82994056e-01 -1.12608016e+00 -1.08185995e+00 3.56175303e-01
3.69480640e-01 -7.91492283e-01 -4.91852462e-01 5.71094394e-01
-1.87672627e+00 3.51693243e-01 -7.27358043e-01 -1.00057103e-01
-1.11713171e+00 7.68563151e-02 4.38715756e-01 2.56689906e-01
-4.23363186e-02 8.32262874e-01 -5.60348153e-01 -1.08686602e+00
1.25647351e-01 2.08437547e-01 -3.22264373e-01 5.83537877e-01
7.69183278e-01 3.87130558e-01 4.36046183e-01 -1.38297987e+00
-3.74112606e-01 3.11136156e-01 -1.87502861e-01 -5.70627928e-01
7.18270957e-01 -2.35895768e-01 -8.93623292e-01 4.71407473e-01
9.44720089e-01 8.57717395e-01 -7.11888015e-01 2.11696438e-02
2.39803061e-01 -2.08131686e-01 -4.95900303e-01 -1.37774122e+00
-6.71824932e-01 8.79181325e-01 7.63603985e-01 4.03346010e-02
7.69686162e-01 -4.93255764e-01 7.75313795e-01 6.41934454e-01
1.13009200e-01 -1.43579423e+00 -2.28872895e-02 1.53569007e+00
1.12108827e+00 -1.13184619e+00 -3.81550640e-01 9.54233781e-02
-9.21525359e-01 1.53298759e+00 5.99301159e-01 3.63330990e-02
5.32972097e-01 2.74013907e-01 9.28499401e-01 2.37768084e-01
1.85061350e-01 -2.23303109e-01 2.32334644e-01 7.51355350e-01
9.95574951e-01 4.34637144e-02 -5.22884250e-01 4.48002994e-01
-6.84456944e-01 -4.80299860e-01 1.15535867e+00 9.16891575e-01
-5.20778418e-01 -1.20773208e+00 -6.72614574e-01 2.56610751e-01
-5.48766375e-01 -5.71165323e-01 -6.73838377e-01 5.14946938e-01
-1.14465706e-01 9.80773509e-01 2.00905591e-01 -3.17418694e-01
-1.26736134e-01 4.89877701e-01 4.89475161e-01 -7.13657498e-01
-9.53625619e-01 4.15023938e-02 1.63757190e-01 -1.36899814e-01
-7.02844262e-02 -7.10244656e-01 -1.30737185e+00 -5.04465342e-01
3.18257920e-02 -1.11681737e-01 8.16693544e-01 7.59574354e-01
-2.36807972e-01 1.02204949e-01 2.55102754e-01 -5.89661002e-01
-5.11375010e-01 -1.07238960e+00 -6.32640183e-01 1.73672378e-01
3.32556963e-01 6.52067214e-02 -5.68261482e-02 5.07954836e-01] | [10.392294883728027, 10.562074661254883] |
65a33468-ce85-4937-921d-fa6ef77ae96c | entity-relation-extraction-as-dependency | 2110.09915 | null | https://arxiv.org/abs/2110.09915v1 | https://arxiv.org/pdf/2110.09915v1.pdf | Entity Relation Extraction as Dependency Parsing in Visually Rich Documents | Previous works on key information extraction from visually rich documents (VRDs) mainly focus on labeling the text within each bounding box (i.e., semantic entity), while the relations in-between are largely unexplored. In this paper, we adapt the popular dependency parsing model, the biaffine parser, to this entity relation extraction task. Being different from the original dependency parsing model which recognizes dependency relations between words, we identify relations between groups of words with layout information instead. We have compared different representations of the semantic entity, different VRD encoders, and different relation decoders. The results demonstrate that our proposed model achieves 65.96% F1 score on the FUNSD dataset. As for the real-world application, our model has been applied to the in-house customs data, achieving reliable performance in the production setting. | ['Zuyi Bao', 'Chen Li', 'Junjie Cao', 'Rui Wang', 'Bo Zhang', 'Yue Zhang'] | 2021-10-19 | null | https://aclanthology.org/2021.emnlp-main.218 | https://aclanthology.org/2021.emnlp-main.218.pdf | emnlp-2021-11 | ['key-information-extraction'] | ['natural-language-processing'] | [ 1.10941000e-01 4.80302334e-01 -1.20650068e-01 -2.92616487e-01
-5.73742032e-01 -8.72567892e-01 5.45147538e-01 4.40860122e-01
-3.77800643e-01 4.86824751e-01 6.20557010e-01 -6.94984496e-01
1.73210666e-01 -9.16585445e-01 -7.26619303e-01 -1.64440334e-01
-7.17010424e-02 3.48977834e-01 4.21809524e-01 -1.42804250e-01
2.70359963e-01 3.96147937e-01 -1.23748779e+00 4.49847996e-01
5.83649933e-01 7.92470336e-01 4.04542446e-01 8.72864366e-01
-5.93621790e-01 9.44326520e-01 -9.70844686e-01 -5.46261430e-01
-1.16567343e-01 -7.77195916e-02 -1.21677661e+00 -6.76741898e-02
5.38950682e-01 -2.26566911e-01 -3.55549514e-01 7.46009111e-01
3.21638644e-01 -1.55446991e-01 6.60608172e-01 -1.00924146e+00
-1.04479611e+00 8.64815116e-01 -5.69911540e-01 8.02933425e-02
5.47877312e-01 -6.28972530e-01 1.41694248e+00 -8.77748013e-01
1.18706715e+00 1.46111798e+00 1.68261334e-01 3.34402621e-01
-1.03482497e+00 -4.15290385e-01 7.21716583e-01 2.25937784e-01
-1.32195401e+00 -3.01301837e-01 4.17702824e-01 -5.17040610e-01
1.62687385e+00 1.43412262e-01 3.57582837e-01 1.15546525e+00
1.25935569e-01 9.96029079e-01 7.98418283e-01 -7.42409348e-01
-3.82931679e-02 1.22006014e-01 6.05025172e-01 7.46669769e-01
5.55250108e-01 -3.61508131e-01 -5.14258742e-01 3.54940116e-01
6.03237689e-01 -6.91166103e-01 -1.34487525e-01 -5.31027019e-01
-1.01068616e+00 5.95688403e-01 2.02745244e-01 2.75644004e-01
1.70001984e-01 1.09447703e-01 5.23949742e-01 2.06116643e-02
5.02152085e-01 2.59655744e-01 -7.17129529e-01 4.57622297e-02
-5.09985805e-01 5.83388321e-02 1.08211088e+00 1.58440173e+00
4.54985052e-01 -3.87580782e-01 -3.78795415e-01 5.72927058e-01
5.29881835e-01 2.71805733e-01 8.97534192e-02 -3.00301403e-01
1.09445000e+00 7.71452606e-01 1.58045758e-02 -1.07144284e+00
-6.28641605e-01 -2.20311493e-01 -4.98052508e-01 5.47024608e-02
3.92831326e-01 -7.46666640e-02 -1.18792176e+00 1.45016468e+00
2.63815433e-01 -3.36058438e-01 2.86678314e-01 6.53694868e-01
1.38028753e+00 4.86046582e-01 4.87120330e-01 1.57478482e-01
1.76920903e+00 -1.02241397e+00 -1.02763295e+00 -5.98895311e-01
8.45453501e-01 -8.35769594e-01 1.09354448e+00 2.41022602e-01
-6.59311771e-01 -6.40839338e-01 -1.22292268e+00 -6.90003216e-01
-9.12418962e-01 4.16595817e-01 6.83945954e-01 4.78359759e-01
-9.42868590e-01 3.44761461e-01 -6.11338854e-01 -6.30420685e-01
2.25856692e-01 1.54737577e-01 -6.08816743e-01 -2.41327696e-02
-1.21765566e+00 1.02663028e+00 6.88389063e-01 1.30535349e-01
-4.23300087e-01 -3.53036135e-01 -1.15447164e+00 5.29763661e-02
7.71542132e-01 -5.86112320e-01 1.14802551e+00 -4.14288282e-01
-1.12463009e+00 9.23506081e-01 -2.61999458e-01 -1.88084692e-01
5.80322891e-02 -7.57666051e-01 -5.81372082e-01 -5.84944785e-02
1.42129794e-01 4.66908842e-01 5.08044958e-01 -1.42637563e+00
-8.04210246e-01 -3.18289071e-01 2.70603716e-01 1.90535173e-01
-2.61429042e-01 3.34594458e-01 -9.45146799e-01 -6.96497142e-01
-7.78669640e-02 -7.16451049e-01 1.58762753e-01 -2.64279574e-01
-9.57788050e-01 -3.20128500e-01 9.00923550e-01 -1.06524825e+00
1.66921926e+00 -2.15793872e+00 3.00936133e-01 1.91332832e-01
2.07433209e-01 1.02123149e-01 -1.27490461e-01 5.53682446e-01
-2.37489387e-01 4.08748358e-01 -1.09424412e-01 -2.82229185e-01
1.85435072e-01 3.52791518e-01 -2.70460069e-01 4.48306650e-02
4.40487891e-01 1.06622303e+00 -8.23230684e-01 -7.05000520e-01
1.00922763e-01 4.78445083e-01 -1.17582686e-01 3.61216933e-01
-2.43656293e-01 -2.31943578e-02 -6.13020539e-01 5.76895237e-01
6.64413810e-01 -3.26486796e-01 6.35117531e-01 -3.54653597e-01
-1.51478767e-01 4.73456353e-01 -1.36094964e+00 1.90053010e+00
-5.44931531e-01 8.72706473e-01 -1.47718236e-01 -5.06607473e-01
1.05801439e+00 1.00677967e-01 -4.98677837e-03 -9.30733621e-01
-1.62948947e-02 -2.85249203e-01 -3.20601404e-01 -6.54766858e-01
9.14269567e-01 5.01001477e-01 -4.13216442e-01 1.38265610e-01
2.20757112e-01 9.53750685e-02 4.81588811e-01 5.85302293e-01
1.25843656e+00 6.96657419e-01 5.11294365e-01 -1.22280762e-01
3.95386100e-01 1.40742406e-01 2.06775397e-01 3.59664649e-01
2.01550335e-01 5.33880830e-01 9.48365867e-01 -8.19309056e-02
-8.89244437e-01 -1.03487074e+00 1.76341496e-02 9.84370649e-01
3.50210190e-01 -1.24473107e+00 -6.21896446e-01 -1.11462319e+00
-1.63301274e-01 9.24187958e-01 -5.12303531e-01 3.10375214e-01
-8.43818486e-01 -4.21514153e-01 4.56798255e-01 8.79991591e-01
3.28066260e-01 -1.00426185e+00 -6.10926449e-01 1.48673430e-01
-6.16571419e-02 -1.75134814e+00 -3.80020179e-02 4.39415246e-01
-3.00136387e-01 -1.34355664e+00 -2.55759120e-01 -1.05128729e+00
6.35477960e-01 -8.90275240e-02 1.57867861e+00 -1.90891087e-01
-1.68013394e-01 3.53596210e-01 -7.22314179e-01 -5.01053989e-01
-2.27546692e-01 3.40054303e-01 -6.18053675e-01 -5.49228132e-01
4.56023812e-01 8.69527757e-02 -3.26097906e-01 -6.99078478e-03
-6.33166850e-01 3.17053050e-01 5.77356339e-01 4.97443974e-01
6.43480003e-01 1.45562187e-01 1.59469116e-02 -1.28585911e+00
6.00961745e-01 -1.92847624e-01 -4.94590759e-01 6.00447357e-01
-5.72459221e-01 3.73111755e-01 4.50596392e-01 -7.53331035e-02
-1.46781564e+00 8.48724321e-02 -2.11225092e-01 3.51498902e-01
-4.69793975e-01 4.70106959e-01 -5.74106157e-01 3.85587662e-01
2.14713320e-01 -3.97756279e-01 -7.23497272e-01 -7.70929039e-01
7.87760794e-01 6.10954225e-01 4.42046046e-01 -6.34325325e-01
7.99954712e-01 1.70707658e-01 7.83871710e-02 -7.43195832e-01
-8.10879111e-01 -3.76192659e-01 -1.05112243e+00 1.21003531e-01
1.20957446e+00 -9.41333354e-01 -4.47937906e-01 3.69167291e-02
-1.47301710e+00 -2.49476880e-01 -2.47377127e-01 2.23533392e-01
-1.61198363e-01 3.35080653e-01 -6.72378600e-01 -7.08966434e-01
-2.98782527e-01 -9.65236604e-01 1.39249206e+00 2.39144906e-01
-4.37364966e-01 -8.37116420e-01 7.25278705e-02 7.24964365e-02
-9.59157869e-02 3.35728884e-01 1.39832544e+00 -6.49039626e-01
-5.81990898e-01 1.28786592e-02 -6.35178745e-01 1.95129350e-01
1.54473424e-01 2.89484680e-01 -8.16652894e-01 6.79198578e-02
-6.97737932e-01 -1.04215942e-01 8.33694637e-01 4.61443402e-02
1.02038169e+00 1.99550390e-02 -7.23687887e-01 3.80915701e-01
1.35741031e+00 3.49419504e-01 7.42971182e-01 5.17632961e-01
1.14679873e+00 8.61207485e-01 8.11807215e-01 1.47436991e-01
7.12381482e-01 6.93505466e-01 3.36534977e-01 -2.36014917e-01
-5.81598461e-01 -5.43609798e-01 3.07416350e-01 7.92965710e-01
8.15506745e-03 -9.10586357e-01 -1.02701998e+00 5.03488660e-01
-1.93745685e+00 -3.92050534e-01 -5.82528353e-01 1.62353504e+00
5.71201384e-01 3.71889621e-01 -2.32082814e-01 2.23791525e-02
6.14338338e-01 1.88423067e-01 -2.70952880e-02 -6.64301515e-01
-1.52492419e-01 4.60994691e-01 5.86894035e-01 4.16415513e-01
-1.33131707e+00 1.42741537e+00 6.27872467e+00 5.73772252e-01
-5.71370125e-01 -2.05429211e-01 2.45250344e-01 2.41233930e-01
-2.99559444e-01 2.03793436e-01 -1.29861772e+00 -7.25834519e-02
9.08944428e-01 3.47500324e-01 -1.50666744e-01 7.23951280e-01
-1.22169949e-01 -3.65655452e-01 -1.16947067e+00 6.72070861e-01
1.60265476e-01 -1.12285876e+00 2.78876852e-02 -1.60880938e-01
2.00406343e-01 -5.51844120e-01 -1.88691914e-01 2.61763096e-01
5.38248420e-01 -1.15193546e+00 9.81826842e-01 3.61974418e-01
9.99210417e-01 -7.71855891e-01 5.62459171e-01 -1.29010409e-01
-1.59484172e+00 2.24949643e-01 -5.76814152e-02 7.26360977e-02
3.42402130e-01 4.90530401e-01 -1.01415372e+00 8.56295049e-01
8.67984176e-01 6.51467144e-01 -1.09489322e+00 5.43042183e-01
-1.01458693e+00 3.47892344e-01 1.27535667e-02 -2.14876577e-01
-7.24052936e-02 -5.92855476e-02 2.27041662e-01 1.72122085e+00
7.09170997e-02 -2.50668377e-01 -1.32594824e-01 4.63940591e-01
3.19759473e-02 3.94368559e-01 -6.14108980e-01 -1.44906506e-01
2.34306350e-01 1.35333025e+00 -1.06147015e+00 -1.93626717e-01
-7.53083527e-01 1.10619903e+00 6.41571939e-01 4.49584514e-01
-7.76185513e-01 -7.16207922e-01 5.39076269e-01 -6.41704947e-02
7.48174965e-01 -6.14163339e-01 -3.42546433e-01 -9.64619398e-01
2.85364926e-01 -6.47150815e-01 5.65674722e-01 -9.03751612e-01
-9.48886275e-01 7.82769382e-01 2.65485376e-01 -8.26859474e-01
-9.82890055e-02 -1.06528044e+00 -8.38669669e-03 8.32028389e-01
-1.69465315e+00 -1.44692171e+00 -1.68768749e-01 2.50701934e-01
4.47320461e-01 1.87405318e-01 1.06015742e+00 2.78020680e-01
-8.29199612e-01 3.85237992e-01 -2.86457926e-01 6.04443133e-01
8.38771045e-01 -1.61667204e+00 1.13640547e+00 1.11712801e+00
6.37583613e-01 6.31154478e-01 5.53106546e-01 -9.01417911e-01
-1.31819761e+00 -1.06536651e+00 1.25339210e+00 -6.70830429e-01
7.54626274e-01 -9.91565704e-01 -8.39254260e-01 8.98618639e-01
5.00555873e-01 -1.25276580e-01 6.49525762e-01 3.83417517e-01
-7.65114725e-01 3.73751700e-01 -7.15175688e-01 7.88712263e-01
1.64239502e+00 -5.84680319e-01 -6.92752838e-01 1.95330396e-01
1.03479767e+00 -7.23935843e-01 -8.82613778e-01 3.32755923e-01
4.21200901e-01 -5.56818485e-01 1.12791872e+00 -8.12380791e-01
6.07931256e-01 -5.18390417e-01 -2.58495718e-01 -1.19395781e+00
-3.41249764e-01 -8.58797207e-02 -6.38460934e-01 1.68731618e+00
7.85176456e-01 -1.35342047e-01 5.91013253e-01 4.75548059e-01
-1.37576431e-01 -4.97870117e-01 -6.29575551e-01 -6.64982855e-01
-1.88744918e-01 -6.13913178e-01 4.87090856e-01 7.09579647e-01
-9.61106922e-03 1.00873613e+00 -1.38114750e-01 4.68274325e-01
3.85302395e-01 2.73591310e-01 6.73244238e-01 -1.14185524e+00
-3.63551974e-01 -1.11245424e-01 -2.71958590e-01 -1.20157743e+00
4.50211465e-01 -6.99722111e-01 1.16989754e-01 -2.27660489e+00
1.09061152e-02 -3.04927021e-01 8.77108648e-02 6.44487977e-01
-1.36076018e-01 -1.25983045e-01 1.17127821e-01 -3.62323254e-01
-5.63987970e-01 1.40019450e-02 1.19594836e+00 -3.80016893e-01
-1.35122985e-01 -4.67194498e-01 -9.39971387e-01 6.78000093e-01
6.74912632e-01 -3.65261167e-01 -4.58687186e-01 -6.76014006e-01
6.29570961e-01 -2.91958570e-01 1.47505477e-01 -6.39467359e-01
2.17750683e-01 2.59449817e-02 5.55129528e-01 -7.43602514e-01
1.79884452e-02 -1.01480389e+00 -1.60611346e-01 -1.11212954e-01
-2.98426896e-01 1.15999624e-01 3.74553055e-01 4.33250695e-01
-1.08293623e-01 -3.40962380e-01 6.93258867e-02 8.44545886e-02
-1.41520643e+00 -9.82782319e-02 -2.43062660e-01 2.27880836e-01
1.00061524e+00 -4.94431779e-02 -8.29006910e-01 7.03839511e-02
-7.16293573e-01 1.26702309e-01 1.34760767e-01 7.96687841e-01
6.92259431e-01 -1.09184051e+00 -5.02539814e-01 7.12300465e-03
4.56055492e-01 3.52449805e-01 -7.76127428e-02 4.99650277e-02
-5.28280318e-01 5.17390668e-01 4.73781414e-02 -1.83158994e-01
-1.64551425e+00 8.35758805e-01 -2.42227405e-01 -4.80344892e-01
-8.12929749e-01 8.76720607e-01 1.19758964e-01 -1.11649267e-01
3.16394925e-01 -6.84174716e-01 -6.98815167e-01 3.82804900e-01
4.31426197e-01 1.01698600e-01 3.20898026e-01 -6.61267638e-01
-5.94460189e-01 6.97292387e-01 -2.74763763e-01 -4.63649295e-02
1.18165982e+00 -9.84757096e-02 -3.43269333e-02 3.51558656e-01
1.21340954e+00 2.47734472e-01 -8.77242625e-01 -4.63110171e-02
6.01682723e-01 -2.59666979e-01 1.29068587e-02 -8.58552098e-01
-7.40353763e-01 7.11075604e-01 3.08527350e-01 3.57164502e-01
1.00705075e+00 5.03090501e-01 4.48711187e-01 4.07880247e-01
1.99879140e-01 -1.27777612e+00 -1.41751349e-01 7.09650278e-01
8.79362583e-01 -1.01538515e+00 2.86999732e-01 -1.10761511e+00
-7.76670694e-01 1.22216368e+00 7.00653911e-01 1.74519762e-01
4.82310802e-01 8.81296754e-01 -3.70570607e-02 -3.06679964e-01
-6.08338177e-01 -6.12124801e-01 4.37515467e-01 9.43133056e-01
7.95046270e-01 7.48883560e-02 -1.34656146e-01 6.29964292e-01
-2.65741795e-01 -3.55194569e-01 3.37839335e-01 1.19619942e+00
-8.49883035e-02 -1.35553992e+00 -9.24604461e-02 1.50746197e-01
-3.25824350e-01 -4.46385592e-01 -7.08204687e-01 1.23438942e+00
1.47576824e-01 1.05055368e+00 3.40183139e-01 -3.56726050e-01
9.73599434e-01 1.52174488e-01 6.96258664e-01 -9.13599968e-01
-5.18233299e-01 -4.63055894e-02 7.66151071e-01 -5.97923577e-01
-4.20398265e-01 -4.41355109e-01 -1.59856367e+00 8.60150158e-02
-1.82815701e-01 -1.84487149e-01 6.09692931e-01 8.62839997e-01
2.96877921e-01 1.04178870e+00 5.04580550e-02 -1.11307152e-01
2.17723250e-01 -6.96061373e-01 -4.55430865e-01 2.63059914e-01
9.18483287e-02 -5.49693704e-01 1.27557993e-01 1.29909575e-01] | [9.282936096191406, 8.128338813781738] |
ca4f658c-23ef-4429-92a9-90448cc7f452 | attention-w-net-improved-skip-connections-for | 2110.08811 | null | https://arxiv.org/abs/2110.08811v2 | https://arxiv.org/pdf/2110.08811v2.pdf | Attention W-Net: Improved Skip Connections for better Representations | Segmentation of macro and microvascular structures in fundoscopic retinal images plays a crucial role in the detection of multiple retinal and systemic diseases, yet it is a difficult problem to solve. Most neural network approaches face several issues such as lack of enough parameters, overfitting and/or incompatibility between internal feature-spaces. We propose Attention W-Net, a new U-Net based architecture for retinal vessel segmentation to address these problems. In this architecture, we have two main contributions: Attention Block and regularisation measures. Our Attention Block uses attention between encoder and decoder features, resulting in higher compatibility upon addition. Our regularisation measures include augmentation and modifications to the ResNet Block used, which greatly prevent overfitting. We observe an F1 and AUC of 0.8407 and 0.9833 on the DRIVE and 0.8174 and 0.9865 respectively on the CHASE-DB1 datasets - a sizeable improvement over its backbone as well as competitive performance among contemporary state-of-the-art methods. | ['Sayantari Ghosh', 'Saumik Bhattacharya', 'Shikhar Mohan'] | 2021-10-17 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 1.44193754e-01 2.44478434e-01 -5.69529012e-02 -1.87379941e-01
-4.87220466e-01 -1.77699283e-01 2.86233902e-01 -6.64412156e-02
-5.25286198e-01 6.34515822e-01 3.08077365e-01 -5.41141033e-01
-1.60748512e-01 -5.01569271e-01 -5.00030041e-01 -3.30978751e-01
3.39435823e-02 -2.13452294e-01 4.85100240e-01 4.23242152e-02
2.78673619e-01 4.30171013e-01 -1.58655763e+00 1.00874066e-01
1.13050449e+00 1.20523548e+00 -1.23850755e-01 9.52984154e-01
-1.26064882e-01 9.68858957e-01 -4.02042747e-01 -4.38769311e-01
4.14804786e-01 -4.74657536e-01 -9.15974438e-01 1.88985139e-01
6.40452623e-01 -4.24857348e-01 -3.74713957e-01 1.01869178e+00
9.02598441e-01 -2.50666440e-01 4.10331130e-01 -4.75075871e-01
-7.41530240e-01 2.92304512e-02 -7.84504890e-01 6.49406791e-01
-2.97363400e-01 3.84836525e-01 9.29633081e-01 -4.37395424e-01
4.87620503e-01 9.96869802e-01 6.65476143e-01 6.43536508e-01
-1.18183637e+00 -2.98399985e-01 -1.30277291e-01 1.85762212e-01
-1.00869334e+00 -4.84265298e-01 -7.33458772e-02 -6.20100617e-01
1.07445395e+00 2.53836215e-01 7.89075077e-01 5.26865363e-01
2.33417049e-01 5.43300569e-01 1.18093431e+00 -5.93608558e-01
-1.95719197e-01 1.60303280e-01 1.00965962e-01 6.96615219e-01
1.60526872e-01 7.51407892e-02 3.95810381e-02 2.00651363e-01
1.30232143e+00 -2.88463771e-01 -4.14308339e-01 1.04223110e-01
-8.99642289e-01 8.17556679e-01 7.64627159e-01 -1.25880584e-01
-3.51392567e-01 1.73580766e-01 2.94237345e-01 2.52047509e-01
2.09517777e-01 6.15335584e-01 -2.46585250e-01 -1.57723472e-01
-6.59934580e-01 -1.64954111e-01 2.63978958e-01 7.40045309e-01
2.22252876e-01 -4.75447215e-02 -4.66496915e-01 1.02751434e+00
1.62082464e-01 -1.00423135e-01 6.50170743e-01 -7.77525365e-01
1.30555779e-01 7.46769130e-01 -1.11631654e-01 -4.43526626e-01
-5.07088423e-01 -9.47628438e-01 -9.64984477e-01 4.93729591e-01
3.12879533e-01 -3.88566256e-01 -1.33526051e+00 1.51828313e+00
6.27056435e-02 4.26744103e-01 -1.10166855e-01 7.83614099e-01
1.05415118e+00 1.61133125e-01 5.02777547e-02 -2.18395274e-02
1.41443646e+00 -1.25083721e+00 -5.47268808e-01 -1.72452286e-01
4.58765179e-01 -9.74472702e-01 9.30036008e-01 1.09690376e-01
-1.43117559e+00 -6.22013986e-01 -9.82345521e-01 -4.05063719e-01
1.53544126e-02 4.91607517e-01 6.10485196e-01 5.48143208e-01
-1.41691256e+00 4.60834712e-01 -5.72346628e-01 -3.93740594e-01
9.04134810e-01 6.80781841e-01 -3.76756012e-01 1.56215817e-01
-6.99685454e-01 9.85554636e-01 2.43976917e-02 1.32208973e-01
-2.78823584e-01 -7.17346787e-01 -5.87399721e-01 5.93150221e-02
1.92775413e-01 -1.02983022e+00 1.17646694e+00 -8.08467686e-01
-1.51020026e+00 1.01211238e+00 -1.35329694e-01 -9.11353528e-01
5.21940589e-01 -2.36782074e-01 -3.47529203e-01 1.30611718e-01
-2.57255912e-01 1.04872894e+00 6.13555312e-01 -4.02398914e-01
-9.55255032e-01 -1.81234449e-01 1.60487309e-01 -3.43252197e-02
-1.52727261e-01 1.82209894e-01 -7.26987064e-01 -5.16318917e-01
-1.44726515e-01 -7.72523761e-01 -5.00049412e-01 3.49797398e-01
-5.22281766e-01 -7.52873421e-02 -2.03431747e-03 -8.05501103e-01
1.32699955e+00 -2.01576757e+00 -1.48694053e-01 9.10106748e-02
7.29133964e-01 1.01150906e+00 -2.97641009e-01 -2.56609142e-01
-2.83125997e-01 3.15299809e-01 1.45881651e-02 2.29166728e-02
-5.79927206e-01 -4.88021113e-02 1.82475537e-01 4.59984094e-01
6.54857934e-01 1.00062656e+00 -4.21343207e-01 -1.95469946e-01
2.03297630e-01 6.50412798e-01 -7.46381462e-01 1.47543743e-01
2.88686492e-02 3.47887039e-01 -1.59034401e-01 4.25392866e-01
4.70327616e-01 -4.84224707e-01 -2.12816447e-01 -2.41011664e-01
-3.85379732e-01 4.11467463e-01 -9.18439150e-01 1.48618627e+00
-2.15286106e-01 1.00801826e+00 -3.51841688e-01 -5.87475717e-01
8.43323767e-01 2.13318720e-01 2.10019365e-01 -7.26471543e-01
5.28988898e-01 1.60603046e-01 7.29724467e-01 -7.69603133e-01
1.94920942e-01 1.62386596e-01 9.00111198e-01 6.21965192e-02
1.40343606e-01 5.60110271e-01 2.64545172e-01 -1.86713129e-01
1.13789892e+00 -7.40407184e-02 4.80372667e-01 -1.60381034e-01
5.50633073e-01 -3.81470323e-01 4.76286292e-01 5.79011023e-01
-3.67580473e-01 8.48534942e-01 9.86620665e-01 -5.10609031e-01
-1.12665379e+00 -4.93752390e-01 -5.74502468e-01 2.18797430e-01
-2.36211717e-01 -4.83482927e-01 -7.66539335e-01 -4.62107480e-01
-1.22389406e-01 9.52190682e-02 -8.48836005e-01 -1.42124463e-02
-2.63870597e-01 -8.36659491e-01 4.06622529e-01 6.64753437e-01
6.47845149e-01 -6.52464211e-01 -7.39194453e-01 2.28145316e-01
2.92340636e-01 -9.96101439e-01 -2.72181034e-01 -2.36941040e-01
-9.58628297e-01 -1.33808088e+00 -9.56338346e-01 -7.92321324e-01
6.43202007e-01 1.21512517e-01 9.17909861e-01 1.62999272e-01
-9.16191876e-01 -1.59682170e-01 1.70567213e-03 -4.80965883e-01
-1.89750776e-01 1.08392807e-02 -4.07839894e-01 7.17082620e-02
3.91422153e-01 -3.78714889e-01 -1.13122880e+00 1.49051502e-01
-5.41374922e-01 8.16737581e-03 9.55699027e-01 9.30524886e-01
6.76019609e-01 -3.46134007e-01 4.26387966e-01 -9.52231407e-01
6.09749019e-01 -1.58034578e-01 -8.11993182e-01 1.62810698e-01
-6.74617290e-01 -1.16455927e-01 1.08593367e-01 -2.18597427e-01
-5.79958498e-01 -1.46585748e-01 -3.11861992e-01 -3.49985272e-01
-1.45535385e-02 2.45212778e-01 2.23819405e-01 -5.29543340e-01
7.47307360e-01 -2.14324102e-01 5.89842200e-01 -4.97767687e-01
2.86046177e-01 9.03833091e-01 6.04572475e-01 5.39280176e-02
2.48883545e-01 2.54793167e-01 9.01638195e-02 -8.29037189e-01
-8.35297287e-01 -4.94686186e-01 -4.24323618e-01 1.11068133e-02
8.57365310e-01 -8.95684123e-01 -7.10608184e-01 4.09658432e-01
-1.00036311e+00 1.40315825e-02 -3.95034909e-01 5.12706041e-01
-3.68404478e-01 3.46325994e-01 -5.65935075e-01 -4.50248837e-01
-5.08386850e-01 -1.41027629e+00 3.13881040e-01 7.00790167e-01
-1.19803295e-01 -7.92530954e-01 -1.18181817e-01 3.48027259e-01
6.90082669e-01 3.64118844e-01 9.93598402e-01 -4.09245670e-01
-6.45483673e-01 -6.04711063e-02 -8.16707551e-01 8.05002034e-01
5.35440035e-02 2.17013359e-01 -1.03771150e+00 -2.33967975e-01
-2.39654630e-01 -2.01995924e-01 1.04686344e+00 9.01209295e-01
1.17728639e+00 -1.01779535e-01 -6.38736337e-02 8.90367091e-01
1.43453336e+00 2.27504373e-01 1.24759710e+00 3.88278127e-01
1.82029545e-01 5.11470795e-01 8.33806694e-02 2.21006557e-01
2.23657370e-01 3.66809219e-01 5.67121565e-01 -7.06093907e-01
-6.30762935e-01 1.67056650e-01 -1.86787233e-01 3.36506605e-01
-3.57896388e-01 -8.76494159e-04 -8.46432567e-01 7.80475914e-01
-1.57248950e+00 -5.44619560e-01 -3.86713624e-01 2.45320797e+00
9.95024145e-01 2.46701509e-01 2.88014382e-01 -2.14104027e-01
7.47626305e-01 -2.68398046e-01 -6.17899954e-01 -4.23965663e-01
-1.61636800e-01 5.73072195e-01 6.24409437e-01 4.34571207e-01
-1.08648205e+00 7.94655979e-01 6.21098852e+00 5.09239018e-01
-1.14599347e+00 2.43097674e-02 9.43271458e-01 -2.66575485e-01
1.72969952e-01 -1.66568816e-01 -8.63014460e-01 4.10587519e-01
8.82453859e-01 1.25091955e-01 1.34176418e-01 2.67479837e-01
1.50234208e-01 -5.93381971e-02 -7.54046023e-01 1.04178476e+00
-5.74418269e-02 -1.67712605e+00 -1.04301527e-01 3.48628432e-01
7.08120704e-01 4.38647300e-01 2.44563773e-01 -5.42025492e-02
7.37127140e-02 -1.42930138e+00 -3.62465195e-02 6.46793485e-01
1.17659080e+00 -6.45267963e-01 9.93796170e-01 -3.66341799e-01
-5.79506814e-01 -1.86300161e-03 -6.02044821e-01 4.93735522e-02
-1.75696053e-02 4.24313724e-01 -7.27011979e-01 9.63921174e-02
6.18896306e-01 7.92071819e-01 -7.27790654e-01 1.99880207e+00
-1.89401835e-01 5.96575618e-01 -9.27725807e-02 1.75106481e-01
4.55827951e-01 -1.72367200e-01 4.95133489e-01 9.07392383e-01
2.94173181e-01 -7.40675777e-02 -4.18727756e-01 8.03230941e-01
-1.43230230e-01 2.92048991e-01 -2.35930428e-01 1.40006959e-01
8.87663960e-02 1.20250666e+00 -2.50698775e-01 -1.60067398e-02
-8.38292360e-01 5.08927226e-01 3.21211010e-01 3.09900284e-01
-7.88730800e-01 -6.51873946e-01 8.95241022e-01 3.86134833e-01
3.58825654e-01 2.03432843e-01 -3.09245437e-01 -8.89959335e-01
1.71109997e-02 -7.42561281e-01 2.02859312e-01 -5.99016249e-01
-9.30403173e-01 8.61068189e-01 -8.90568972e-01 -1.33879948e+00
1.57243341e-01 -7.92131245e-01 -6.14024460e-01 1.28538013e+00
-2.11003995e+00 -9.87280488e-01 -3.84331614e-01 4.04843599e-01
2.19408870e-01 -4.79431599e-01 6.59544885e-01 3.25788736e-01
-8.09313595e-01 7.86707580e-01 3.28864455e-02 1.79799825e-01
8.44496727e-01 -1.22036040e+00 5.58018923e-01 9.30223882e-01
-1.99090764e-01 6.88641191e-01 1.70697823e-01 -2.61443526e-01
-7.66562283e-01 -1.30329347e+00 8.15250754e-01 -9.71921757e-02
6.53966844e-01 2.10813716e-01 -8.52686763e-01 6.36644602e-01
4.36094642e-01 3.57974678e-01 8.39708745e-01 8.46126676e-02
-1.78159609e-01 -4.02336493e-02 -9.03376877e-01 6.70855403e-01
8.18817854e-01 -1.02343827e-01 -2.46866956e-01 3.57157856e-01
5.42395234e-01 -6.87774360e-01 -1.02002740e+00 3.23711962e-01
4.41737115e-01 -1.29685628e+00 8.05660188e-01 -1.00348866e+00
7.37636209e-01 -9.58438814e-02 2.91455567e-01 -9.73302424e-01
-5.03838420e-01 -1.10159123e+00 -5.94478287e-02 1.07282436e+00
6.39438212e-01 -9.72245872e-01 5.64732492e-01 4.63407874e-01
-1.80027768e-01 -1.13181138e+00 -8.76561463e-01 -4.36356246e-01
1.32366434e-01 -6.83893189e-02 3.44241470e-01 5.63636124e-01
-4.04445529e-01 4.16297793e-01 -2.78857619e-01 -4.74784896e-02
3.64569306e-01 -2.40745142e-01 6.63982689e-01 -1.29042017e+00
-1.72156543e-01 -6.81123018e-01 -8.46802950e-01 -1.04003406e+00
-4.02619928e-01 -7.70237446e-01 -5.47319055e-01 -1.73903811e+00
1.59519672e-01 -4.13982183e-01 -4.21164840e-01 5.66333652e-01
-1.65153101e-01 3.94070357e-01 -1.00754529e-01 1.63227007e-01
-1.02889739e-01 3.00557017e-01 1.52853882e+00 1.90750197e-01
-3.06598932e-01 1.43225744e-01 -1.10294592e+00 5.89591205e-01
1.03853583e+00 2.44261157e-02 -4.04113770e-01 -6.46715164e-01
2.05236115e-02 -1.31284282e-01 4.80975419e-01 -1.08405435e+00
2.90133923e-01 3.70785981e-01 2.73586184e-01 -1.78197280e-01
-2.32575834e-02 -3.84693086e-01 -3.37630332e-01 5.16481280e-01
-3.52755725e-01 -9.74376276e-02 4.11779404e-01 4.71832484e-01
-3.08574319e-01 -1.69249371e-01 1.22161436e+00 -6.68871403e-02
-5.01681745e-01 4.72712815e-01 -1.53491735e-01 2.03132883e-01
1.03294778e+00 -3.34544867e-01 -6.97883487e-01 -6.43663704e-02
-7.15329528e-01 3.18297714e-01 2.35681668e-01 3.40159416e-01
5.38578153e-01 -8.53089571e-01 -1.00517237e+00 5.14270008e-01
6.08783998e-02 8.11350644e-02 2.80500889e-01 1.40713513e+00
-8.76003981e-01 5.55494964e-01 -4.53255326e-01 -5.52233934e-01
-1.44471347e+00 8.30409005e-02 7.38065422e-01 8.85780528e-02
-8.97814870e-01 1.09616244e+00 6.41335994e-02 2.10352033e-01
3.22370142e-01 -5.69919884e-01 -5.17061472e-01 -2.35607494e-02
7.03363180e-01 3.74919713e-01 1.84698954e-01 -2.27484763e-01
1.50133476e-01 8.01427007e-01 -6.08163595e-01 4.25470114e-01
1.23053944e+00 -5.08172475e-02 -2.99014121e-01 -2.83571184e-01
9.68210936e-01 -2.46924460e-01 -1.28360748e+00 -4.05424982e-01
-2.77720928e-01 -5.30513346e-01 5.28359413e-01 -1.09285581e+00
-1.25000000e+00 1.07232261e+00 1.04006517e+00 1.86668009e-01
1.17121744e+00 -2.34598115e-01 8.16543460e-01 -1.21584728e-01
-4.09906387e-01 -7.90570021e-01 -3.99269879e-01 3.26277286e-01
8.37304413e-01 -1.25743318e+00 -9.58239585e-02 -5.22149980e-01
-5.08507311e-01 1.06133628e+00 6.83152258e-01 -1.76452085e-01
4.83225405e-01 1.01692885e-01 2.98720330e-01 -4.64422032e-02
-7.95321584e-01 -6.71429217e-01 6.42863154e-01 5.79433441e-01
7.04223037e-01 -2.96275616e-01 -4.10708725e-01 3.93059373e-01
-1.07144147e-01 3.04754943e-01 8.16014290e-01 5.23576021e-01
-4.92233008e-01 -1.08197808e+00 2.10698873e-01 9.53355968e-01
-8.96302700e-01 -3.37064415e-01 -1.45337403e-01 9.31071520e-01
6.52984083e-02 6.96887076e-01 3.26703280e-01 -7.78891295e-02
3.43734741e-01 -2.63467371e-01 4.31346834e-01 -6.06266797e-01
-6.79888189e-01 5.51354110e-01 1.60565376e-01 -6.68311715e-01
-3.43553305e-01 -3.11065793e-01 -8.52137566e-01 -1.11453449e-02
-4.24965352e-01 -2.91292071e-01 3.98945868e-01 4.68084723e-01
8.11230123e-01 9.33767200e-01 3.37982923e-01 -1.02629192e-01
-5.64227819e-01 -1.03759456e+00 -4.63256657e-01 -1.73763689e-02
6.84382439e-01 -5.09301245e-01 -1.06566601e-01 -1.48769030e-02] | [15.796382904052734, -3.968369245529175] |
87da1274-c428-44c6-ac40-61564f20c97e | progressive-transformers-for-end-to-end-sign | 2004.14874 | null | https://arxiv.org/abs/2004.14874v2 | https://arxiv.org/pdf/2004.14874v2.pdf | Progressive Transformers for End-to-End Sign Language Production | The goal of automatic Sign Language Production (SLP) is to translate spoken language to a continuous stream of sign language video at a level comparable to a human translator. If this was achievable, then it would revolutionise Deaf hearing communications. Previous work on predominantly isolated SLP has shown the need for architectures that are better suited to the continuous domain of full sign sequences. In this paper, we propose Progressive Transformers, a novel architecture that can translate from discrete spoken language sentences to continuous 3D skeleton pose outputs representing sign language. We present two model configurations, an end-to-end network that produces sign direct from text and a stacked network that utilises a gloss intermediary. Our transformer network architecture introduces a counter that enables continuous sequence generation at training and inference. We also provide several data augmentation processes to overcome the problem of drift and improve the performance of SLP models. We propose a back translation evaluation mechanism for SLP, presenting benchmark quantitative results on the challenging RWTH-PHOENIX-Weather-2014T(PHOENIX14T) dataset and setting baselines for future research. | ['Richard Bowden', 'Necati Cihan Camgoz', 'Ben Saunders'] | 2020-04-30 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1430_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123560664.pdf | eccv-2020-8 | ['sign-language-production'] | ['natural-language-processing'] | [ 7.60435224e-01 3.45161945e-01 1.41947985e-01 -5.73959231e-01
-1.00661421e+00 -5.71189225e-01 8.12066674e-01 -8.99780393e-01
-5.72642148e-01 6.40387297e-01 8.96000385e-01 -5.23986340e-01
2.61240095e-01 -3.84097099e-01 -8.67629588e-01 -3.74540567e-01
8.50204378e-02 7.01698124e-01 3.41125876e-01 -4.60573614e-01
-1.45753443e-01 2.63308465e-01 -1.55899215e+00 7.97227204e-01
4.89708543e-01 6.15451336e-01 1.02539413e-01 1.35124350e+00
-1.47203267e-01 1.14049971e+00 -7.79291391e-01 -3.68692249e-01
3.73829126e-01 -8.61263037e-01 -8.31686735e-01 -1.58274829e-01
8.72889757e-01 -7.27195203e-01 -2.22695097e-01 3.51735353e-01
1.01650465e+00 -1.66376397e-01 6.28352642e-01 -1.27898741e+00
-6.42701149e-01 7.69802272e-01 5.71754910e-02 -3.96258742e-01
6.70516610e-01 5.07835746e-01 1.02142930e+00 -7.53937244e-01
1.06899762e+00 1.40441954e+00 8.18679035e-01 1.06663692e+00
-8.08336556e-01 -6.21820390e-01 1.62286952e-01 -8.16228762e-02
-7.16399789e-01 -8.12160492e-01 2.27911964e-01 -1.32387787e-01
1.29094124e+00 2.74947941e-01 9.29913878e-01 1.59548724e+00
-4.40239757e-01 1.27928030e+00 1.02605212e+00 -7.31464148e-01
-7.71087036e-02 -3.91485900e-01 -4.27533805e-01 6.02982521e-01
-3.82509589e-01 2.40529045e-01 -1.19471848e+00 1.28282413e-01
5.70206881e-01 -6.42895579e-01 -5.74068487e-01 -2.30872452e-01
-1.38033044e+00 2.80054122e-01 2.71957070e-01 2.37403680e-02
-3.36662889e-01 5.39337039e-01 5.01495719e-01 7.50354528e-01
-7.56758526e-02 1.37794361e-01 -4.98159766e-01 -5.72693586e-01
-1.22017002e+00 3.95357788e-01 9.60025430e-01 1.02154040e+00
-3.25416505e-01 2.71471232e-01 -3.75740051e-01 7.40459204e-01
6.16764545e-01 7.70206809e-01 5.60766757e-01 -9.61581767e-01
7.65799403e-01 9.74784717e-02 2.88551417e-03 2.22558141e-01
-1.87805548e-01 -8.18643272e-02 -2.63411433e-01 7.01723933e-01
8.28023791e-01 -3.25579345e-01 -1.68742239e+00 1.69408417e+00
-5.26318438e-02 1.81635931e-01 3.41613948e-01 1.05515456e+00
8.56701910e-01 5.42384624e-01 -3.04870866e-02 3.01564276e-01
1.07961619e+00 -1.07419813e+00 -4.40467924e-01 -1.40101299e-01
5.88200510e-01 -8.26662362e-01 1.42930984e+00 4.73154932e-01
-1.37290442e+00 -9.69160348e-02 -8.67651403e-01 -3.43362927e-01
-1.26930489e-03 1.16540365e-01 2.65971780e-01 6.18435383e-01
-1.47359169e+00 4.13940787e-01 -1.06191051e+00 -6.68779194e-01
3.56319487e-01 3.05117518e-01 -3.55307192e-01 5.01870327e-02
-9.61887836e-01 1.00669444e+00 9.94044021e-02 2.15386569e-01
-6.09226167e-01 -6.52774870e-01 -6.78286016e-01 -3.44842404e-01
-3.37393403e-01 -9.49456930e-01 1.98156583e+00 -1.05929339e+00
-2.15746951e+00 8.77522230e-01 -1.74956337e-01 -6.31506860e-01
1.25826120e+00 -3.68027002e-01 -1.32858425e-01 1.86283648e-01
-2.76633650e-01 1.03661442e+00 8.65514278e-01 -1.03295434e+00
-9.52648580e-01 8.51303861e-02 -4.16445732e-01 3.88126642e-01
1.70035541e-01 3.29507262e-01 -3.33667666e-01 -6.75397754e-01
8.66001025e-02 -1.01588798e+00 1.48854405e-01 3.63300353e-01
-1.93194434e-01 5.82039766e-02 7.05042481e-01 -1.15802717e+00
6.67060971e-01 -1.77099097e+00 2.56616622e-01 2.75609404e-01
-3.23702097e-01 3.82146180e-01 -4.29623276e-01 5.74211776e-01
1.53302565e-01 -1.91763133e-01 -6.37824297e-01 -6.69316471e-01
1.96804628e-01 5.09054899e-01 -5.93602479e-01 1.06790692e-01
2.07293078e-01 1.13787901e+00 -8.61890852e-01 -1.86508983e-01
-8.35765079e-02 7.22257853e-01 -5.84695697e-01 4.79589738e-02
-6.46076679e-01 6.13772213e-01 2.54547149e-01 6.75720215e-01
1.98907673e-01 1.33878320e-01 1.67323202e-01 3.65126938e-01
-1.74028143e-01 7.43580997e-01 -8.26342285e-01 1.90010488e+00
-6.01388097e-01 9.57433820e-01 1.97650507e-01 -4.46211666e-01
6.58484638e-01 5.81348121e-01 2.77478136e-02 -6.91653967e-01
1.02831751e-01 8.04790378e-01 7.71040395e-02 -6.53021634e-01
3.63593578e-01 -4.61124331e-01 7.52742663e-02 6.85089171e-01
5.83334453e-02 -5.99323094e-01 1.91352189e-01 -2.90115066e-02
1.29783797e+00 7.35394239e-01 -2.52787948e-01 3.76129448e-01
1.16704181e-01 -7.43165165e-02 7.94328824e-02 7.47711062e-01
-4.97764498e-02 1.12050962e+00 2.37467840e-01 -5.31094037e-02
-1.14344811e+00 -1.32225704e+00 3.33394527e-01 1.09981132e+00
-5.77278852e-01 -5.46794236e-02 -5.04331350e-01 -5.47530651e-01
-7.80144781e-02 8.18800449e-01 -2.66508669e-01 2.44063497e-01
-1.06700349e+00 -2.98158824e-01 1.25718558e+00 7.66647160e-01
5.51443696e-01 -1.62916863e+00 -8.98861170e-01 2.80956566e-01
-3.29921037e-01 -1.13964415e+00 -5.89774489e-01 5.64320870e-02
-5.21642804e-01 -6.37587607e-01 -1.41087735e+00 -1.29710603e+00
2.67245531e-01 -3.00785244e-01 9.02062595e-01 -1.20413385e-01
4.50205132e-02 3.64940971e-01 -6.60879433e-01 -6.74723506e-01
-1.01576436e+00 1.18959025e-01 -8.10382664e-02 -2.76952535e-01
1.94348156e-01 -6.78548932e-01 -4.21740770e-01 -1.36568770e-01
-7.28897989e-01 4.56359208e-01 7.84061372e-01 9.33047116e-01
9.04194340e-02 -1.04649508e+00 3.84365022e-01 -2.36372054e-01
7.49821067e-01 1.79967791e-01 -3.58084470e-01 3.10764939e-01
-1.81721851e-01 4.46693480e-01 4.14896280e-01 -3.98105919e-01
-1.10755825e+00 2.13222831e-01 -7.34925151e-01 -1.44619751e-03
7.67815858e-02 2.35789835e-01 -1.44264968e-02 8.61360356e-02
7.02528179e-01 6.08688176e-01 4.24369216e-01 -4.86953884e-01
5.98740160e-01 1.11221814e+00 9.51736867e-01 -4.60293144e-01
6.92785263e-01 5.03765643e-01 -1.57249197e-01 -9.09280956e-01
-3.98813576e-01 -1.83139026e-01 -5.34208953e-01 -2.29220003e-01
5.42340040e-01 -8.39103162e-01 -5.68377376e-01 1.01440573e+00
-1.21561980e+00 -1.08549535e+00 -5.83835661e-01 4.79000807e-01
-9.79483604e-01 1.48855463e-01 -5.45293689e-01 -7.40566313e-01
-5.69758534e-01 -7.26461530e-01 1.32082140e+00 -3.39409113e-02
-6.71673238e-01 -5.41475654e-01 3.04699302e-01 3.26210290e-01
4.73651201e-01 1.64414003e-01 4.54012156e-01 -1.88117340e-01
-5.43661535e-01 -6.16431907e-02 -6.99721724e-02 6.78446651e-01
-9.11281258e-02 5.33547923e-02 -9.98350620e-01 -2.52229691e-01
-6.49754286e-01 -6.59331977e-01 1.00270200e+00 1.94819972e-01
2.05418646e-01 -4.35676217e-01 7.81740248e-02 7.79169023e-01
8.16192567e-01 7.73187131e-02 7.57021725e-01 3.98528486e-01
4.89996374e-01 5.86587131e-01 6.87349737e-02 5.33531196e-02
6.59493983e-01 6.07482910e-01 -4.63793389e-02 -3.31719592e-02
-9.83927429e-01 -7.19439626e-01 8.84102404e-01 8.64148438e-01
-2.20662802e-01 -3.86605471e-01 -9.96512830e-01 9.23083365e-01
-1.71144700e+00 -1.09749234e+00 3.03834165e-03 1.82548118e+00
1.05869603e+00 1.09535649e-01 3.14266056e-01 4.05097544e-01
7.31367394e-02 -1.34220151e-02 -4.57518369e-01 -7.18062639e-01
-3.21480095e-01 7.24777639e-01 5.45397758e-01 7.47624159e-01
-6.15740061e-01 1.12894833e+00 6.27601814e+00 5.71285188e-02
-1.40556490e+00 -6.83882162e-02 -2.63563842e-01 -4.87046510e-01
-2.22259954e-01 -1.58499494e-01 -6.68661952e-01 3.14385265e-01
1.06836390e+00 3.05018216e-01 4.34136927e-01 2.98281848e-01
4.46884662e-01 2.91709036e-01 -1.10317409e+00 7.74601042e-01
1.90354809e-01 -1.17883575e+00 2.52332151e-01 -2.10944787e-01
6.05915427e-01 6.35930002e-01 -5.03067635e-02 2.94507861e-01
7.44430244e-01 -1.00802422e+00 1.36176598e+00 5.78540146e-01
1.23905575e+00 -1.62083223e-01 3.45865965e-01 2.66630411e-01
-8.50038946e-01 -5.72785065e-02 6.01790369e-01 -2.03207955e-01
8.34032297e-01 -3.04867744e-01 -1.49796224e+00 1.90657243e-01
4.74771857e-01 4.45750654e-01 -2.15626955e-01 1.13172972e+00
-8.29505205e-01 8.98549616e-01 -8.24968576e-01 -2.44379625e-01
4.12422836e-01 3.38833183e-01 6.67973936e-01 1.49516916e+00
6.14051342e-01 -2.17305392e-01 -3.11683238e-01 2.94327259e-01
-1.40521049e-01 -6.72538802e-02 -5.05960941e-01 1.83029007e-02
2.07466573e-01 2.82956630e-01 -2.93995380e-01 -4.07531887e-01
-3.86583656e-01 1.49852824e+00 -3.39697339e-02 4.56707090e-01
-4.28848475e-01 -4.57161754e-01 6.91497266e-01 2.50461578e-01
5.00457644e-01 -3.27728450e-01 -2.25906551e-01 -1.17950809e+00
4.00036633e-01 -1.12927413e+00 1.75756872e-01 -1.04317439e+00
-9.10520673e-01 5.18750310e-01 -2.82103837e-01 -1.24460459e+00
-8.47705662e-01 -8.42955172e-01 -4.42239463e-01 9.57294703e-01
-1.58906758e+00 -1.67523134e+00 -2.65576363e-01 4.51572716e-01
5.06347537e-01 -3.09174657e-02 7.97201991e-01 2.63616771e-01
3.28495413e-01 8.70380461e-01 -2.80229729e-02 3.67866725e-01
6.96878552e-01 -1.21284533e+00 1.03407109e+00 1.03795743e+00
3.89881253e-01 2.16200709e-01 7.66959369e-01 -4.86151785e-01
-8.51742685e-01 -8.26131463e-01 1.37697637e+00 -6.19426310e-01
4.81338710e-01 -3.88935924e-01 -4.04655635e-01 8.13895404e-01
2.04666868e-01 -3.81619513e-01 1.40770972e-01 -4.41749871e-01
-5.33478916e-01 2.08766952e-01 -9.58930254e-01 9.02297497e-01
1.67720079e+00 -5.87603927e-01 -6.85099125e-01 1.55181572e-01
6.66545630e-01 -7.30898738e-01 -3.90388250e-01 3.26469719e-01
1.33548212e+00 -7.63571739e-01 7.57230878e-01 -6.42982721e-01
4.95055884e-01 -3.51454556e-01 -1.04756936e-01 -1.40432036e+00
4.38950807e-01 -1.09457624e+00 -1.81966528e-01 8.69930744e-01
6.72612607e-01 -5.96487463e-01 9.81736600e-01 4.21830744e-01
-3.44149590e-01 -3.12051892e-01 -1.06081796e+00 -9.51759696e-01
2.73746431e-01 -6.11617446e-01 5.46374559e-01 4.06654745e-01
-8.33709687e-02 2.49163464e-01 -5.63683271e-01 -1.46319985e-01
4.30563599e-01 -2.21989915e-01 1.03454804e+00 -7.33166039e-01
-4.69294816e-01 -6.60229027e-01 -4.46502835e-01 -1.43823791e+00
-1.11443825e-01 -1.18077028e+00 4.46810246e-01 -2.07499409e+00
-5.31088412e-01 -2.86565199e-02 2.71954387e-01 8.19941103e-01
3.54642183e-01 1.93292260e-01 5.39407074e-01 1.21115327e-01
9.15136114e-02 4.84201640e-01 1.47283995e+00 1.83392372e-02
-4.04812217e-01 1.36307776e-01 -3.15182656e-01 4.35159087e-01
7.10157871e-01 -1.83841065e-01 -2.79720366e-01 -9.79214489e-01
4.93141636e-02 -6.26645014e-02 5.47398686e-01 -1.11838186e+00
2.72637069e-01 2.80172557e-01 2.45687496e-02 -5.36836505e-01
4.43504632e-01 -4.75455344e-01 -1.63766712e-01 6.01347923e-01
-5.27404130e-01 -1.64298534e-01 4.05876245e-03 4.44912575e-02
-1.48380280e-01 5.89469038e-02 7.31567740e-01 -1.71534434e-01
-6.14631593e-01 -1.06080502e-01 -4.85803723e-01 1.15783364e-01
4.25030470e-01 -3.17734510e-01 -2.38882318e-01 -8.10849309e-01
-7.23893523e-01 3.26410502e-01 3.17004114e-01 6.82622850e-01
7.11479127e-01 -1.17543423e+00 -1.15325034e+00 3.30097258e-01
4.40074690e-02 8.38117823e-02 -4.21499610e-01 6.05959296e-01
-1.05545664e+00 3.63210201e-01 -2.77776629e-01 -4.54390585e-01
-1.47545516e+00 -4.45379823e-01 4.06459302e-01 -5.19333445e-02
-1.09879208e+00 1.30172658e+00 -7.17381954e-01 -9.33752418e-01
5.52776515e-01 -7.90434062e-01 3.34449947e-01 -5.37680797e-02
5.53809047e-01 6.90390095e-02 3.89516391e-02 -7.54864454e-01
-2.94895798e-01 4.08231020e-01 7.62154683e-02 -1.14855230e+00
1.44988298e+00 1.86922103e-01 2.94141263e-01 3.32229584e-01
9.97888327e-01 -1.33335516e-01 -1.41094732e+00 -1.46712169e-01
8.54630023e-02 -4.84918095e-02 -3.52857023e-01 -1.40179253e+00
-4.79705364e-01 9.10781264e-01 5.63551426e-01 -5.14934719e-01
9.88804996e-01 -1.64867148e-01 1.22856772e+00 5.79057157e-01
2.04489037e-01 -1.08515155e+00 -7.16703907e-02 9.33407426e-01
1.30850899e+00 -9.43751156e-01 -5.34346998e-01 2.26942241e-01
-7.37457395e-01 1.09055924e+00 1.77112833e-01 1.00491866e-01
2.29842007e-01 6.69676304e-01 7.36764908e-01 9.40939933e-02
-7.35737681e-01 -3.82040650e-01 1.73879147e-01 7.41853535e-01
4.67154950e-01 2.87395306e-02 -7.83595666e-02 9.80797336e-02
-9.35259223e-01 6.39148533e-01 4.82780248e-01 1.32348096e+00
-1.69517666e-01 -1.34470487e+00 -4.34181422e-01 3.51827085e-01
-5.50726540e-02 -2.61181742e-01 -6.99600399e-01 7.47404099e-01
-2.61358581e-02 6.86151505e-01 2.15274282e-02 -1.30850703e-01
6.75199151e-01 7.20723271e-01 9.16371226e-01 -4.08787668e-01
-5.92755437e-01 -2.09813938e-01 5.38811922e-01 -2.86077827e-01
-3.93220931e-01 -1.12648022e+00 -1.55728352e+00 -4.93866131e-02
2.29587749e-01 -4.81283605e-01 8.23698342e-01 9.26319957e-01
8.94685313e-02 4.64715987e-01 -8.97933021e-02 -9.38494384e-01
-7.08895862e-01 -1.17655313e+00 -5.82260117e-02 3.21835935e-01
7.80544937e-01 3.19886431e-02 -3.05212528e-01 3.98117542e-01] | [9.212510108947754, -6.5379228591918945] |
11df4460-9827-4601-abfe-f43fe1153af7 | rast-domain-robust-dialogue-rewriting-as | null | null | https://aclanthology.org/2021.emnlp-main.402 | https://aclanthology.org/2021.emnlp-main.402.pdf | RAST: Domain-Robust Dialogue Rewriting as Sequence Tagging | The task of dialogue rewriting aims to reconstruct the latest dialogue utterance by copying the missing content from the dialogue context. Until now, the existing models for this task suffer from the robustness issue, i.e., performances drop dramatically when testing on a different dataset. We address this robustness issue by proposing a novel sequence-tagging-based model so that the search space is significantly reduced, yet the core of this task is still well covered. As a common issue of most tagging models for text generation, the model’s outputs may lack fluency. To alleviate this issue, we inject the loss signal from BLEU or GPT-2 under a REINFORCE framework. Experiments show huge improvements of our model over the current state-of-the-art systems when transferring to another dataset. | ['Dong Yu', 'Zhaopeng Tu', 'Kun Xu', 'LiWei Wang', 'Linfeng Song', 'Jie Hao'] | null | null | null | null | emnlp-2021-11 | ['dialogue-rewriting'] | ['natural-language-processing'] | [ 4.86524850e-01 7.04823554e-01 1.02671809e-01 -4.02179658e-01
-9.31517363e-01 -6.12092078e-01 8.43126833e-01 -2.35807329e-01
-1.96267828e-01 1.18025434e+00 7.65851617e-01 -2.35070884e-01
5.50099671e-01 -5.09199381e-01 -5.71464300e-01 -2.70017356e-01
5.02490819e-01 6.27825439e-01 2.11919755e-01 -7.59476781e-01
2.01786667e-01 -2.68329442e-01 -1.07018054e+00 5.39146602e-01
1.16182518e+00 4.58066523e-01 2.25756913e-01 5.64074695e-01
-3.70266408e-01 9.72173035e-01 -1.07234514e+00 -7.28895664e-01
2.59060524e-02 -1.09716952e+00 -1.13613927e+00 -4.18793187e-02
1.18845724e-01 -3.26288790e-01 -2.07200527e-01 1.07980967e+00
6.23515427e-01 1.55962050e-01 4.43226993e-01 -9.77658808e-01
-6.99680567e-01 1.12997556e+00 -2.31392998e-02 -2.97543883e-01
7.94369161e-01 5.49406670e-02 1.00474250e+00 -8.47736597e-01
8.68482351e-01 1.48624945e+00 6.26504540e-01 1.25861061e+00
-1.17713201e+00 -2.29691178e-01 1.98632568e-01 -6.33748844e-02
-7.64383376e-01 -7.89294600e-01 6.97915733e-01 -6.11445773e-03
1.14093053e+00 4.06334698e-01 3.65538687e-01 1.62958336e+00
1.64004415e-01 8.63999963e-01 1.09490776e+00 -6.38898790e-01
-3.26050608e-03 1.45117462e-01 -1.74997255e-01 6.99326754e-01
-4.05279040e-01 -1.92464456e-01 -6.42211020e-01 -5.23533486e-02
3.26076448e-01 -4.74468410e-01 -4.85338837e-01 2.52913162e-02
-1.11684370e+00 9.00486648e-01 -7.52776163e-03 3.30582500e-01
1.93798766e-01 -2.63673335e-01 5.03983498e-01 6.01953685e-01
5.97274661e-01 8.49023700e-01 -5.18362641e-01 -5.04417300e-01
-7.01508820e-01 4.69272494e-01 1.22824383e+00 1.10149384e+00
4.67615545e-01 -2.19006911e-02 -7.84824848e-01 1.16156089e+00
8.90061855e-02 3.40742967e-03 7.44805217e-01 -9.69644725e-01
8.39232206e-01 6.13000691e-01 1.35606870e-01 -5.19364893e-01
-4.89927739e-01 -2.82237172e-01 -6.72111452e-01 -1.57991573e-01
5.61230719e-01 -5.65882862e-01 -6.28064990e-01 2.19276690e+00
2.20556661e-01 -3.32866050e-02 3.40414494e-01 8.76140416e-01
8.98598075e-01 7.49329567e-01 -4.32128645e-03 -3.20371479e-01
1.28513491e+00 -1.51664519e+00 -9.69345272e-01 -6.31001830e-01
6.39087558e-01 -8.95236254e-01 1.29557359e+00 2.20595121e-01
-1.19373333e+00 -5.35670400e-01 -1.02830839e+00 -2.01260611e-01
-1.77600771e-01 4.54647280e-02 6.10395968e-01 6.31067872e-01
-1.10639060e+00 6.18644893e-01 -2.80395418e-01 -4.18036044e-01
-3.36111426e-01 -1.17985412e-01 -2.03462154e-01 1.15291305e-01
-1.74321032e+00 1.24132013e+00 4.89158094e-01 5.35276681e-02
-4.94039357e-01 -5.55850446e-01 -9.32315111e-01 -3.79754021e-03
4.44159865e-01 -8.28831553e-01 1.84310782e+00 -1.01441443e+00
-2.37793612e+00 4.64865863e-01 -6.70524910e-02 -4.96548682e-01
7.76851773e-01 -3.16983521e-01 -2.48965293e-01 -4.55023646e-01
3.00203487e-02 6.94607675e-01 6.75627589e-01 -1.10618949e+00
-6.08223200e-01 4.06181030e-02 2.54355311e-01 5.12650251e-01
-1.30332187e-01 -1.14217624e-01 -3.53468150e-01 -7.81422973e-01
-2.53652304e-01 -1.21137083e+00 -3.20099950e-01 -7.74869502e-01
-5.55055559e-01 -3.34422708e-01 5.82898736e-01 -8.50854695e-01
1.37728548e+00 -1.61108899e+00 4.97095108e-01 -5.58171391e-01
-3.47436339e-01 5.66779435e-01 -3.51407140e-01 8.05246174e-01
7.96990469e-02 2.01595902e-01 -3.68650168e-01 -6.45946920e-01
4.00149971e-02 1.21913984e-01 -4.39434648e-01 -2.75938720e-01
4.53604877e-01 9.31343973e-01 -1.11318362e+00 -9.06846747e-02
1.00860912e-02 8.51665512e-02 -6.59389913e-01 6.38995111e-01
-6.34575546e-01 7.30809689e-01 -2.52950817e-01 1.45576969e-01
3.69325519e-01 1.02319658e-01 5.75096965e-01 1.58163294e-01
-5.93147837e-02 1.10157478e+00 -7.51540542e-01 2.38279462e+00
-5.79725385e-01 3.23716491e-01 -1.04222655e-01 -7.56431401e-01
8.71801019e-01 5.15121043e-01 3.87871079e-02 -6.68561101e-01
-5.18958941e-02 1.37444705e-01 1.94141433e-01 -3.41646969e-01
1.11838877e+00 -2.76508570e-01 -4.78491813e-01 6.81990266e-01
2.39210695e-01 -1.38907000e-01 8.67898688e-02 1.67035416e-01
1.24955666e+00 4.53636378e-01 1.81677714e-01 4.41053882e-02
4.42634732e-01 3.33259292e-02 6.87609017e-01 7.06686795e-01
1.29008070e-02 7.68600345e-01 6.18252873e-01 4.27666716e-02
-9.83411849e-01 -5.72490990e-01 2.14966550e-01 1.13821590e+00
-1.81087524e-01 -6.50598347e-01 -1.26820052e+00 -1.16664612e+00
-2.70536005e-01 1.03888071e+00 -4.47266936e-01 -3.16473693e-01
-8.91843259e-01 -7.08661675e-01 8.47522318e-01 1.52463287e-01
4.01952177e-01 -1.24808347e+00 -3.00992936e-01 6.28785014e-01
-7.90531158e-01 -1.16207361e+00 -4.91448760e-01 -1.00491077e-01
-7.51861453e-01 -6.84988320e-01 -6.72891319e-01 -6.43981040e-01
3.51127893e-01 3.48887891e-02 1.41544044e+00 1.02689378e-01
2.78124481e-01 -7.84264132e-02 -6.94546163e-01 -2.25122347e-01
-1.19686711e+00 5.78089595e-01 -2.31620356e-01 -1.76869065e-01
1.16877453e-02 -2.38541529e-01 -2.24777713e-01 1.44190967e-01
-7.53275752e-01 4.09756273e-01 3.23888332e-01 1.24741602e+00
3.57811786e-02 -4.16959524e-01 1.15503335e+00 -1.27925432e+00
1.20704889e+00 -3.20773423e-01 -1.19609550e-01 4.94113594e-01
-5.33322275e-01 3.12219411e-01 9.65131998e-01 -1.73855901e-01
-1.49428034e+00 -1.18580930e-01 -3.13365608e-01 1.76215142e-01
-2.11371277e-02 4.32673365e-01 -4.18434918e-01 4.70493585e-01
4.69281822e-01 8.86824578e-02 -5.95852081e-03 -6.20686233e-01
5.81699431e-01 8.27746153e-01 3.90339106e-01 -5.33430636e-01
4.18562382e-01 -2.76063323e-01 -3.93030733e-01 -4.56998020e-01
-1.04825842e+00 -1.53811827e-01 -2.97072053e-01 -1.34196162e-01
4.99579072e-01 -7.87108004e-01 -3.72514248e-01 4.71736938e-01
-1.64900780e+00 -4.85884845e-01 -2.03271255e-01 1.23290017e-01
-5.67018330e-01 7.06811965e-01 -7.64266670e-01 -6.96803391e-01
-5.58883846e-01 -1.14021504e+00 7.42592275e-01 2.41818905e-01
-5.73120534e-01 -9.55211937e-01 4.29294229e-01 5.24765432e-01
4.06141520e-01 7.04077184e-02 1.02681971e+00 -8.01881313e-01
-1.99366868e-01 2.17191860e-01 1.14716284e-01 4.40543115e-01
1.69686824e-01 -2.32597351e-01 -9.48382318e-01 -1.42343089e-01
2.20420793e-01 -6.24090552e-01 8.00372541e-01 -2.48912454e-01
6.74333394e-01 -5.98347187e-01 -7.66693726e-02 2.40099430e-01
8.23528826e-01 4.53741252e-02 6.83388293e-01 1.84555545e-01
4.34457242e-01 9.87385571e-01 8.26247513e-01 3.81917834e-01
5.66338599e-01 9.99815524e-01 2.96185404e-01 1.76645201e-02
-4.71884549e-01 -6.67611182e-01 7.17590809e-01 1.23145521e+00
2.32558504e-01 -7.27701247e-01 -6.76087081e-01 4.49838430e-01
-2.21234059e+00 -8.30008149e-01 8.20853375e-03 2.06008649e+00
1.38821280e+00 1.03127815e-01 -6.15399778e-02 -1.35016248e-01
6.83421969e-01 4.07313973e-01 -3.20290476e-01 -7.95505524e-01
-9.91404951e-02 9.67503786e-02 1.30363284e-02 7.03509152e-01
-8.43491077e-01 1.33565068e+00 6.33440971e+00 6.63284838e-01
-8.60499680e-01 2.08634689e-01 4.20429677e-01 8.13265741e-02
-4.10844505e-01 2.25242198e-01 -9.04863477e-01 6.39311254e-01
1.16223514e+00 -2.52409955e-03 4.42789853e-01 5.67624032e-01
1.95196122e-01 2.71488093e-02 -1.14796937e+00 4.67838943e-01
3.56993288e-01 -1.01266587e+00 1.03886269e-01 -2.84378320e-01
7.12861538e-01 -3.29478651e-01 -2.09239230e-01 7.48727977e-01
5.37504137e-01 -1.05590165e+00 6.94377661e-01 4.61446226e-01
6.85777009e-01 -6.58377945e-01 6.96567774e-01 5.57982743e-01
-5.16557217e-01 2.29424596e-01 -4.02865231e-01 -1.52137130e-01
4.26035315e-01 5.16964495e-01 -1.30627906e+00 6.53566480e-01
4.51770388e-02 4.52114820e-01 -5.32394528e-01 5.75835168e-01
-7.09899902e-01 6.15194261e-01 6.42737597e-02 -8.88952240e-02
2.21706122e-01 -2.28529617e-01 7.07140028e-01 1.40436971e+00
4.26574230e-01 -1.70087174e-01 1.32152006e-01 9.12613213e-01
-4.14474696e-01 7.87092298e-02 -5.44279039e-01 9.73589122e-02
4.94988889e-01 1.25874865e+00 -2.23661050e-01 -1.66608140e-01
-2.98161983e-01 1.42348456e+00 6.89094484e-01 9.83377993e-02
-7.72271454e-01 -2.22458422e-01 5.10831833e-01 -2.34680474e-01
3.19245085e-02 -2.53592655e-02 -1.89775601e-01 -1.26308930e+00
6.54067248e-02 -1.33881974e+00 1.21692210e-01 -5.15189052e-01
-1.22961819e+00 7.22693205e-01 -3.38052839e-01 -1.10752892e+00
-9.25515354e-01 -2.47908190e-01 -4.88229364e-01 8.32484543e-01
-1.54131401e+00 -1.11338723e+00 9.28651690e-02 2.56661057e-01
1.09892404e+00 -2.38940582e-01 1.16886032e+00 6.13271035e-02
-5.79841077e-01 8.68665755e-01 -7.67073408e-02 1.12752050e-01
1.25667703e+00 -1.49794590e+00 8.72733414e-01 8.65090907e-01
2.12305635e-02 5.41315675e-01 7.71549761e-01 -6.49878144e-01
-1.12611806e+00 -1.04718149e+00 1.30021250e+00 -6.23771250e-01
4.67933863e-01 -5.60473204e-01 -9.30831730e-01 4.52340394e-01
6.20695293e-01 -6.72666490e-01 7.00999379e-01 2.60368139e-01
-2.44254768e-01 4.17966664e-01 -9.82167423e-01 8.79749179e-01
1.25830138e+00 -4.39793766e-01 -9.56398427e-01 2.42870554e-01
1.01899731e+00 -8.87885630e-01 -8.41052890e-01 2.40018949e-01
4.39253122e-01 -1.07262611e+00 4.39990938e-01 -7.45605052e-01
6.45072818e-01 -1.12112641e-01 1.92693114e-01 -1.90594208e+00
-7.81664550e-02 -1.25506568e+00 -1.70175567e-01 1.69718277e+00
7.01710582e-01 -4.08969134e-01 5.47891974e-01 5.51880002e-01
-5.08056343e-01 -2.82613665e-01 -9.81078029e-01 -6.42518461e-01
2.30370194e-01 1.39228944e-02 7.18544424e-01 9.89354134e-01
5.93473911e-01 1.09883344e+00 -8.43120098e-01 -4.34682071e-01
9.22171772e-02 -2.05362681e-02 8.78769517e-01 -9.20224130e-01
-5.31219363e-01 -2.55621016e-01 3.42363238e-01 -1.23440599e+00
4.83264506e-01 -8.22700918e-01 5.16030014e-01 -1.58787072e+00
-2.11296510e-02 -4.48137850e-01 1.10142484e-01 5.45017838e-01
-5.80958009e-01 -1.29601225e-01 5.09516478e-01 4.06024307e-02
-4.59699601e-01 8.13010097e-01 1.26299965e+00 1.12184115e-01
-3.41156453e-01 -5.66312931e-02 -7.63632059e-01 4.62948650e-01
9.40140426e-01 -5.14563084e-01 -4.31111366e-01 -6.32026732e-01
2.21184671e-01 5.04972339e-01 -4.69233319e-02 -8.02310705e-01
1.35018393e-01 -7.93168023e-02 -1.52777612e-01 -1.94051802e-01
4.04198825e-01 -3.13371688e-01 6.37901276e-02 3.14383835e-01
-7.50058293e-01 5.46726622e-02 7.28087202e-02 4.02328432e-01
-2.67243892e-01 -5.39410174e-01 4.44328547e-01 -2.30677366e-01
-3.74334216e-01 -3.67498785e-01 -5.48805475e-01 3.62860084e-01
5.63353121e-01 2.38940626e-01 -7.06185818e-01 -6.45717740e-01
-3.24981898e-01 8.56971145e-02 5.21729708e-01 8.42154622e-01
2.77886808e-01 -1.23591614e+00 -9.70233798e-01 -3.34665254e-02
2.21736729e-02 -1.07026674e-01 7.82912225e-02 6.08079374e-01
-9.11359861e-02 5.02489269e-01 -1.60068080e-01 -1.88727051e-01
-1.01204026e+00 1.85313344e-01 3.01274717e-01 -8.70307028e-01
-4.50015068e-01 6.67307258e-01 -2.26681121e-02 -8.85308921e-01
1.13114998e-01 -1.14693597e-01 -2.37101778e-01 3.28201912e-02
3.91474783e-01 2.47050241e-01 2.85189062e-01 -4.83121783e-01
-1.22685172e-02 2.63148621e-02 -3.68342578e-01 -4.35350955e-01
9.99429107e-01 -3.39635998e-01 -2.17608456e-02 3.68535340e-01
7.46825576e-01 1.64281160e-01 -1.24243164e+00 -2.32214868e-01
1.50244921e-01 -1.53160572e-01 -3.76550794e-01 -1.18560398e+00
-6.75757945e-01 7.18290091e-01 6.48866370e-02 3.51963550e-01
7.54933357e-01 -3.73482913e-01 1.10378623e+00 4.21367198e-01
2.73433387e-01 -1.36679733e+00 1.53101021e-02 1.28354740e+00
1.06109881e+00 -1.16153526e+00 -2.38255814e-01 -3.62166226e-01
-1.07783294e+00 1.02078569e+00 6.91261351e-01 1.60833269e-01
7.15841427e-02 1.02203928e-01 2.59625673e-01 1.59865558e-01
-1.08605230e+00 -7.73408711e-02 1.09230235e-01 5.92242301e-01
9.09911454e-01 1.19065955e-01 -7.57458627e-01 8.14069510e-01
-5.83803952e-01 -1.13152124e-01 7.87525177e-01 8.30246389e-01
-2.87651718e-01 -1.75263810e+00 5.94873987e-02 2.15130616e-02
-6.19880557e-01 -2.84166634e-01 -8.98872018e-01 5.30882061e-01
-2.45201498e-01 1.32420397e+00 -2.83477724e-01 -4.57912117e-01
5.76907039e-01 5.54013968e-01 3.06202352e-01 -9.87135172e-01
-1.15538001e+00 -1.31212518e-01 7.84949422e-01 -4.27926332e-01
-3.68009418e-01 -4.32400554e-01 -1.11982346e+00 -1.01833850e-01
-3.89430970e-01 3.78589481e-01 5.40293932e-01 8.94821048e-01
4.37438667e-01 5.80957890e-01 7.22673535e-01 -5.69193542e-01
-9.71839070e-01 -1.42536843e+00 -1.20423652e-01 5.09568810e-01
-3.51494402e-02 -3.78555745e-01 -1.06880344e-01 -1.15476921e-01] | [12.487499237060547, 8.408202171325684] |
9a187593-f51c-4fc9-9535-6f1617436d79 | forecasting-with-deep-learning-s-p-500-index | 2103.14080 | null | https://arxiv.org/abs/2103.14080v1 | https://arxiv.org/pdf/2103.14080v1.pdf | Forecasting with Deep Learning: S&P 500 index | Stock price prediction has been the focus of a large amount of research but an acceptable solution has so far escaped academics. Recent advances in deep learning have motivated researchers to apply neural networks to stock prediction. In this paper, we propose a convolution-based neural network model for predicting the future value of the S&P 500 index. The proposed model is capable of predicting the next-day direction of the index based on the previous values of the index. Experiments show that our model outperforms a number of benchmarks achieving an accuracy rate of over 55%. | ['Ikhlaas Gurrib', 'Linda Smail', 'Firuz Kamalov'] | 2021-03-21 | null | null | null | null | ['stock-price-prediction', 'stock-prediction'] | ['time-series', 'time-series'] | [-6.49397671e-01 -4.52533513e-01 -1.90668255e-01 -5.59803545e-01
1.15991816e-01 -3.52773428e-01 5.24479568e-01 -1.47649318e-01
-4.93816167e-01 7.66778827e-01 1.66316912e-01 -5.01115978e-01
5.19800298e-02 -1.28401458e+00 -4.28410798e-01 -3.02234977e-01
-1.54475585e-01 1.22831874e-01 3.53988975e-01 -4.38976765e-01
8.24325144e-01 6.64407074e-01 -1.37433612e+00 2.93401510e-01
1.76512390e-01 1.88393807e+00 -2.48954669e-01 1.54739335e-01
-4.40329701e-01 1.38135207e+00 -7.44687259e-01 -7.40916133e-01
7.93770671e-01 -3.04650515e-01 -4.28861618e-01 -3.77097189e-01
5.45542054e-02 -8.30062211e-01 -6.18642926e-01 1.03382647e+00
2.33768001e-01 6.19135983e-02 2.32849360e-01 -1.14285719e+00
-9.11265254e-01 9.70419824e-01 -4.20796812e-01 8.72586250e-01
-4.40398276e-01 -5.78929223e-02 1.35333061e+00 -8.01203430e-01
2.06599340e-01 7.52498448e-01 6.59194171e-01 -3.56627366e-04
-6.86714172e-01 -9.25634801e-01 2.08146840e-01 4.21175748e-01
-1.00833797e+00 1.38452709e-01 9.81206238e-01 -2.61596888e-01
1.30297983e+00 -1.28833145e-01 1.00929987e+00 4.00806010e-01
5.20558894e-01 7.86432803e-01 1.06469560e+00 -2.64870636e-02
3.50461006e-01 -4.52795923e-02 1.95745483e-01 2.08050728e-01
5.12056768e-01 2.54344285e-01 -5.08880198e-01 -1.63722448e-02
8.09013844e-01 3.56030554e-01 6.43205643e-02 3.34437996e-01
-1.15340841e+00 1.19783819e+00 6.80792928e-01 6.84494674e-01
-7.46597946e-01 3.19916308e-01 3.98306906e-01 6.00490570e-01
7.57005095e-01 4.39366281e-01 -7.54699349e-01 -4.14532602e-01
-1.41556561e+00 6.92443848e-01 1.05103624e+00 3.36795598e-01
1.92885235e-01 7.85655975e-01 8.08171779e-02 2.84851402e-01
1.14912391e-01 1.72525182e-01 8.07540596e-01 -6.35502756e-01
3.58634740e-01 6.47321284e-01 2.76453644e-01 -1.09432507e+00
-5.16869187e-01 -7.59329975e-01 -7.52396703e-01 4.86849844e-01
4.72146511e-01 -4.62935448e-01 -5.92105746e-01 9.86055076e-01
-2.74269253e-01 4.79751170e-01 1.30173281e-01 7.00837970e-01
4.46807593e-01 9.73281443e-01 -4.43559110e-01 -7.28455409e-02
7.48332977e-01 -9.27305400e-01 -6.70708954e-01 -1.61928892e-01
1.34873435e-01 -4.84969199e-01 2.36731187e-01 5.16758859e-01
-1.10779011e+00 -4.15539801e-01 -1.25279605e+00 3.99079978e-01
-5.69110453e-01 -2.83562183e-01 6.93657994e-01 4.39033449e-01
-1.01552832e+00 1.13809693e+00 -6.93257868e-01 3.00810069e-01
4.47374851e-01 3.92153531e-01 3.24064255e-01 6.14861190e-01
-1.48465645e+00 9.43712115e-01 7.60505021e-01 4.28848118e-01
-1.83455065e-01 -6.52560771e-01 -2.50604361e-01 4.52135473e-01
-6.64536729e-02 -2.16951236e-01 1.46584296e+00 -8.53338897e-01
-1.51998794e+00 2.31467739e-01 3.95394236e-01 -1.30458951e+00
5.64235151e-01 -1.97934389e-01 -7.78237164e-01 -2.02314332e-01
-3.69731694e-01 3.98622423e-01 7.95120239e-01 -4.52783197e-01
-1.17213452e+00 -1.75321847e-01 -3.02320309e-02 -7.24685714e-02
-6.14555478e-01 5.44824302e-02 6.21040091e-02 -1.08463013e+00
4.42321077e-02 -6.76092267e-01 -1.68834895e-01 -3.03478509e-01
3.20897140e-02 -4.13681179e-01 6.19193733e-01 -6.33733571e-01
1.35592079e+00 -1.83139133e+00 -5.43793023e-01 2.10647136e-01
-4.12895083e-02 3.18265229e-01 2.14526057e-01 3.35426956e-01
-3.48490179e-01 8.28526262e-03 7.07760975e-02 7.10516125e-02
2.21190289e-01 -2.50638295e-02 -9.85083580e-01 1.85640782e-01
1.75239012e-01 1.20112717e+00 -4.17469323e-01 1.39419302e-01
1.11687719e-03 3.05719644e-01 -2.27325588e-01 1.59883752e-01
-4.01824057e-01 -3.24150234e-01 -3.17045510e-01 5.83831251e-01
6.98432326e-01 -3.97937566e-01 -1.81338996e-01 2.15821743e-01
-4.39976960e-01 4.97739345e-01 -1.04929888e+00 7.29179382e-01
1.05575465e-01 7.30636060e-01 -6.67262256e-01 -1.01683390e+00
1.50044763e+00 1.72296837e-01 6.37470901e-01 -8.02618563e-01
3.33242685e-01 6.45670891e-01 3.75942469e-01 1.11929908e-01
6.02568567e-01 -4.64015484e-01 1.67543918e-01 7.64354825e-01
-4.80636150e-01 1.60457477e-01 3.45886916e-01 -3.73486787e-01
7.91273415e-01 -1.81686237e-01 2.82981656e-02 -1.57237738e-01
3.51532936e-01 -4.30430435e-02 4.80096191e-01 4.34992701e-01
-2.71142989e-01 2.00101763e-01 5.58350921e-01 -1.28191650e+00
-9.85812962e-01 -6.55296445e-01 -2.77309269e-01 8.02467465e-01
-3.76197189e-01 3.16035271e-01 -4.45771694e-01 -3.38808805e-01
6.16542995e-01 9.41169202e-01 -7.48594165e-01 6.22738078e-02
-6.56120181e-01 -9.25834358e-01 3.93992633e-01 8.28919947e-01
8.06081891e-01 -1.64327133e+00 -8.01347613e-01 5.49327075e-01
4.01824057e-01 -8.54276001e-01 -1.46734372e-01 3.50166589e-01
-1.15162921e+00 -7.25250363e-01 -1.02238405e+00 -6.89545751e-01
3.57258529e-03 -1.14733629e-01 1.26248848e+00 1.42457709e-01
4.46194172e-01 -3.10268074e-01 -1.42996550e-01 -9.94384706e-01
-8.03780407e-02 2.30562419e-01 2.21499018e-02 3.81652266e-01
8.23094070e-01 -4.87158567e-01 -6.70313001e-01 3.81862894e-02
-8.23604107e-01 -3.38414967e-01 4.69009250e-01 6.09444201e-01
2.87505776e-01 4.38904971e-01 1.14363325e+00 -5.08465827e-01
9.89909470e-01 -5.82236767e-01 -1.25659788e+00 -1.44146785e-01
-1.27716386e+00 3.79120111e-02 5.49166024e-01 -2.48533040e-01
-8.25266004e-01 -3.95155519e-01 -2.48032942e-01 -4.14089233e-01
2.82931238e-01 9.94630933e-01 5.73487282e-01 1.10786311e-01
-7.38970414e-02 5.54845214e-01 -6.39323145e-03 -6.14400327e-01
-2.87414342e-01 4.00815547e-01 4.74625528e-01 7.78861791e-02
6.34467125e-01 3.34198147e-01 -3.51711586e-02 -2.29245692e-01
-1.13417244e+00 -6.36126772e-02 -5.53702414e-01 -1.66253254e-01
4.21569854e-01 -7.78261185e-01 -7.81409264e-01 9.87992585e-01
-9.90270197e-01 -1.08341388e-01 -2.35680938e-01 5.59023798e-01
-2.10842445e-01 -1.48175210e-01 -8.58857512e-01 -7.63159096e-01
-6.02789223e-01 -9.35355306e-01 1.95741862e-01 3.87123346e-01
-9.16892439e-02 -1.06986654e+00 2.31951863e-01 -7.06300661e-02
9.51731384e-01 1.70747608e-01 5.42900085e-01 -1.13582838e+00
-8.56351972e-01 -8.64070714e-01 -4.13412780e-01 5.96296430e-01
-8.03309083e-02 -1.10722244e-01 -7.17823982e-01 -9.07628983e-02
2.55640447e-01 -9.12589803e-02 1.10005152e+00 5.09283960e-01
1.08492804e+00 -2.59637833e-01 1.83411434e-01 6.29761398e-01
1.44366539e+00 6.38381958e-01 8.06998789e-01 9.79566455e-01
1.74866900e-01 1.48649186e-01 3.65681082e-01 8.29073787e-01
2.90761054e-01 1.40360430e-01 5.94679832e-01 2.88430214e-01
6.21470451e-01 -5.99150471e-02 2.16124758e-01 7.68255889e-01
-2.88659424e-01 6.98676705e-03 -7.70056188e-01 4.62630540e-01
-1.60123229e+00 -1.30186105e+00 8.48107710e-02 1.63850272e+00
5.29591501e-01 7.41501927e-01 2.51015693e-01 2.99991608e-01
4.43234771e-01 5.06717384e-01 -9.84319746e-01 -4.66511518e-01
-2.74422206e-02 3.05158138e-01 6.63696170e-01 -9.76442769e-02
-1.23510957e+00 6.72409594e-01 7.19780684e+00 2.81432807e-01
-1.40675461e+00 -6.00997746e-01 1.05567098e+00 -1.48345873e-01
-1.04598783e-01 -3.43504816e-01 -1.17047608e+00 7.79851615e-01
1.25378251e+00 -7.46085703e-01 2.14119941e-01 1.11914790e+00
-1.08155003e-02 2.53683329e-01 -7.23648608e-01 7.74656594e-01
-8.26251134e-02 -1.90221572e+00 1.29934654e-01 3.25343668e-01
9.55770075e-01 3.91933382e-01 4.92445558e-01 4.27110523e-01
1.41199261e-01 -1.00962353e+00 8.29747379e-01 7.93379784e-01
4.39386070e-02 -1.22937310e+00 1.25566542e+00 3.71500313e-01
-1.05577958e+00 -4.38510984e-01 -6.12142026e-01 -6.23502374e-01
8.16799328e-02 6.12630546e-01 -5.03099084e-01 7.68175870e-02
9.20214832e-01 8.17754149e-01 -3.61809522e-01 1.04409802e+00
1.98432371e-01 6.55616164e-01 -1.77080974e-01 -5.11606812e-01
8.10541451e-01 -3.96164894e-01 -2.99530495e-02 7.97327161e-01
6.49971485e-01 1.21152945e-01 -4.69765142e-02 9.04705703e-01
-4.27267283e-01 8.33881572e-02 -3.56510013e-01 -3.44553143e-01
1.56832829e-01 8.62215161e-01 -7.61980534e-01 -3.94825220e-01
-8.48508835e-01 6.21978760e-01 1.98917314e-01 3.87137458e-02
-6.51593685e-01 -3.97433758e-01 7.49268830e-01 5.78761958e-02
9.33836341e-01 -6.74028099e-02 -6.58541203e-01 -1.00939095e+00
1.11603364e-01 -5.87773323e-01 3.60571355e-01 -4.52820629e-01
-1.33644974e+00 7.41004050e-01 -3.78147542e-01 -1.33374178e+00
-5.43355465e-01 -9.89127040e-01 -9.23037350e-01 1.07558131e+00
-1.98831534e+00 -3.92727107e-01 9.57003832e-02 2.59114236e-01
4.07978475e-01 -9.23544228e-01 5.60281932e-01 1.45641118e-02
-5.17266333e-01 2.57185042e-01 6.08346939e-01 8.28908801e-01
7.66931325e-02 -1.30292726e+00 1.17903137e+00 7.46431887e-01
2.17035189e-01 3.45655650e-01 4.91845518e-01 -7.28201985e-01
-8.97054613e-01 -1.07922363e+00 1.13667357e+00 1.23227984e-02
1.31805718e+00 4.43327129e-01 -1.10487294e+00 7.73239195e-01
3.54534686e-01 4.07834612e-02 4.01325852e-01 -3.39118958e-01
-3.41094643e-01 -4.14530128e-01 -9.55262363e-01 3.11740816e-01
1.65838242e-01 -1.26830742e-01 -6.26571596e-01 -1.48530200e-01
5.52637994e-01 -1.66945606e-01 -9.43868935e-01 1.75493509e-01
6.55240953e-01 -1.33558857e+00 9.04964387e-01 -4.21030253e-01
5.35019934e-01 8.09065253e-02 -1.14702784e-01 -1.26844466e+00
-4.00530785e-01 -3.07672322e-01 -5.83823502e-01 6.71683967e-01
4.04758602e-01 -8.66718292e-01 1.13912523e+00 9.31274831e-01
1.20887317e-01 -1.12530196e+00 -9.47301865e-01 -8.11356068e-01
4.97096688e-01 -3.65662277e-01 1.20024788e+00 8.14663470e-01
-3.03193331e-02 -1.20451972e-01 -3.45721632e-01 -2.20589802e-01
4.44986999e-01 7.33428776e-01 2.61416942e-01 -1.46544397e+00
-2.25432187e-01 -1.17152417e+00 -4.32161063e-01 -9.68095899e-01
2.04998463e-01 -7.72809625e-01 -6.22840464e-01 -1.39427340e+00
-2.69141227e-01 -1.92834586e-02 -9.70943034e-01 3.69883031e-01
9.85720083e-02 3.55622619e-01 2.90029943e-01 2.33857483e-01
-9.37401131e-02 4.74602133e-01 1.03891909e+00 -2.34945774e-01
-1.13035545e-01 4.03045505e-01 -5.83294749e-01 8.01036417e-01
1.32126248e+00 -1.67018458e-01 5.12456521e-02 -2.08025977e-01
5.21556735e-01 2.99313553e-02 2.56405137e-02 -1.14177454e+00
9.08770561e-02 -3.08796674e-01 8.44328105e-01 -1.10896242e+00
1.36105165e-01 -7.09830642e-01 5.87612065e-03 8.45355570e-01
-4.81495082e-01 6.34774864e-01 1.51512548e-01 3.61119807e-01
-6.36602819e-01 -4.01565880e-01 6.66538656e-01 -2.62421548e-01
-9.61459696e-01 5.77700913e-01 -3.15725267e-01 -1.35847509e-01
1.01770234e+00 -8.99844244e-02 -1.12396345e-01 -4.79004592e-01
-3.00306559e-01 1.09208442e-01 -7.99801201e-02 6.25693500e-01
8.83558393e-01 -1.47859061e+00 -8.12505782e-01 1.61708012e-01
-3.77339572e-01 -5.08081973e-01 -2.12258264e-01 2.62635618e-01
-8.82457316e-01 8.54176998e-01 -4.39021051e-01 1.71620920e-01
-5.25930703e-01 6.33352458e-01 6.26049936e-01 -4.85984802e-01
-8.92631948e-01 7.70167291e-01 -3.30322206e-01 1.71783417e-01
4.14539367e-01 -6.16387904e-01 -5.96062362e-01 2.71430343e-01
1.05786896e+00 4.98702675e-01 1.94153756e-01 -5.19221187e-01
4.12035398e-02 1.01814985e-01 -2.74648696e-01 9.27478895e-02
2.05888939e+00 4.16094780e-01 -1.88261583e-01 8.49774361e-01
1.12404072e+00 -6.37309849e-01 -1.15495265e+00 -3.19542259e-01
5.69281995e-01 -4.02134866e-01 4.00973320e-01 -7.63816178e-01
-1.63928890e+00 7.87829816e-01 4.34677541e-01 6.29050791e-01
1.03124356e+00 -5.15247643e-01 1.43258607e+00 7.90879130e-01
3.82880479e-01 -1.31530559e+00 -2.65866425e-02 8.30744207e-01
8.62770736e-01 -1.23538184e+00 -1.04744896e-01 4.86432612e-01
-5.29533386e-01 1.52370942e+00 3.12848568e-01 -6.76001132e-01
1.26137364e+00 4.75509912e-01 3.18290383e-01 -2.20002204e-01
-8.76419663e-01 -4.26878221e-02 2.82603383e-01 1.46969007e-02
5.37718773e-01 -1.76349450e-02 -3.37004699e-02 7.46042132e-01
-5.36003172e-01 3.92191976e-01 6.30756497e-01 7.66251743e-01
-6.95389628e-01 -7.90874958e-01 -4.20925617e-01 8.76738966e-01
-9.57848430e-01 -3.18006247e-01 -2.32489660e-01 6.63512051e-01
-3.27016324e-01 4.62841779e-01 6.57177031e-01 -3.11431199e-01
3.34263802e-01 3.63994747e-01 -1.38674006e-01 -9.95759591e-02
-7.41696954e-01 -4.52741124e-02 -3.08151722e-01 -2.31879458e-01
-2.56837338e-01 -7.76943147e-01 -1.15454745e+00 -7.04854965e-01
8.65658298e-02 3.90919410e-02 4.45706815e-01 8.78760040e-01
1.53250560e-01 4.12866831e-01 9.21811283e-01 -7.88679123e-01
-1.14492798e+00 -1.06206346e+00 -1.19264901e+00 2.16424122e-01
3.01670134e-01 -4.43174034e-01 -2.69898683e-01 -1.14339694e-01] | [4.414137840270996, 4.213064670562744] |
7069ac2e-7e81-4577-80cc-800c62c5644c | two-stage-movie-script-summarization-an | null | null | https://aclanthology.org/2022.creativesumm-29.9 | https://aclanthology.org/2022.creativesumm-29.9.pdf | Two-Stage Movie Script Summarization: An Efficient Method For Low-Resource Long Document Summarization | The Creative Summarization Shared Task at COLING 2022 aspires to generate summaries given long-form texts from creative writing. This paper presents the system architecture and the results of our participation in the Scriptbase track that focuses on generating movie plots given movie scripts. The core innovation in our model employs a two-stage hierarchical architecture for movie script summarization. In the first stage, a heuristic extraction method is applied to extract actions and essential dialogues, which reduces the average length of input movie scripts by 66% from about 24K to 8K tokens. In the second stage, a state-of-the-art encoder-decoder model, Longformer-Encoder-Decoder (LED), is trained with effective fine-tuning methods, BitFit and NoisyTune. Evaluations on the unseen test set indicate that our system outperforms both zero-shot LED baselines as well as other participants on various automatic metrics and ranks 1st in the Scriptbase track. | ['Vera Demberg', 'Ernie Chang', 'Pin-Jie Lin', 'Xudong Hong', 'Dongqi Pu'] | null | null | https://aclanthology.org/2022.creativesumm-1.9 | https://aclanthology.org/2022.creativesumm-1.9.pdf | coling-creativesumm-2022-10 | ['document-summarization'] | ['natural-language-processing'] | [ 7.11324811e-01 4.24449503e-01 9.80844647e-02 -3.86047453e-01
-1.39548564e+00 -8.26319039e-01 8.13929796e-01 -2.03047723e-01
-2.43901417e-01 7.70396769e-01 1.20775831e+00 1.36811033e-01
2.88295418e-01 -2.40565553e-01 -6.04414940e-01 -1.31076172e-01
3.67186487e-01 4.98519182e-01 1.10477665e-02 -2.51311153e-01
6.13034368e-01 -1.12630099e-01 -1.03018689e+00 1.25362372e+00
6.88150644e-01 4.37266976e-01 2.41241068e-01 1.45842695e+00
-1.53484955e-01 1.36304843e+00 -1.18229914e+00 -9.35651660e-01
-1.20188043e-01 -1.20106530e+00 -9.29333806e-01 1.56753987e-01
7.61517346e-01 -7.13003278e-01 -5.51876962e-01 7.33939469e-01
8.80422354e-01 2.80116588e-01 7.04153180e-01 -5.27728438e-01
-6.60786152e-01 1.32440031e+00 1.60859078e-02 2.66620755e-01
8.12549770e-01 3.66570503e-01 1.24590671e+00 -1.00915146e+00
1.00966942e+00 1.03149915e+00 7.42060483e-01 8.63839567e-01
-1.19188046e+00 -4.99175936e-02 -2.62527615e-01 -3.77948470e-02
-1.08162546e+00 -1.00076962e+00 5.96700668e-01 -4.16878790e-01
1.49829042e+00 6.09668612e-01 5.31422138e-01 1.55014241e+00
2.80616701e-01 1.06286550e+00 5.90755880e-01 -3.50941986e-01
2.28981286e-01 6.15980383e-03 -2.21925870e-01 4.80200440e-01
-3.11432391e-01 -7.58874834e-01 -1.11090481e+00 -8.41894969e-02
6.82337463e-01 -7.49797702e-01 -1.54512689e-01 5.87967992e-01
-1.33390081e+00 6.32774234e-01 -1.78397119e-01 1.90849781e-01
-3.96514505e-01 2.58776933e-01 8.93776298e-01 1.61518976e-01
6.95508897e-01 1.28734243e+00 6.63747713e-02 -8.44565988e-01
-1.62322259e+00 8.09515715e-01 1.09454072e+00 1.25992346e+00
-3.01512599e-01 2.26297274e-01 -1.09701598e+00 1.02125967e+00
-2.38580912e-01 1.78527713e-01 5.80972910e-01 -1.17032218e+00
1.01159453e+00 2.06895843e-01 2.53660351e-01 -5.52573144e-01
3.33070499e-03 -2.40476459e-01 -5.62305450e-01 -4.50503618e-01
1.34992704e-01 -3.73476893e-01 -6.45203650e-01 1.30201125e+00
-3.68272394e-01 -2.28189319e-01 2.65547577e-02 6.87846243e-01
1.10424840e+00 9.51556623e-01 -2.48445958e-01 -3.95400584e-01
1.14721417e+00 -1.45393562e+00 -1.05262566e+00 -2.68418074e-01
1.65132433e-01 -9.78873253e-01 1.36125612e+00 6.91670239e-01
-1.89677799e+00 -6.40414894e-01 -1.24071980e+00 -4.78105664e-01
2.48155504e-01 4.65738177e-01 3.09051603e-01 1.99645951e-01
-1.04832184e+00 9.59254205e-01 -6.61598563e-01 -5.61285019e-01
4.02051091e-01 -2.61761516e-01 -1.18491640e-02 1.63153768e-01
-8.86371791e-01 7.23997951e-01 1.80191085e-01 -9.56466869e-02
-1.26944089e+00 -6.50581241e-01 -5.23248672e-01 1.68280378e-01
2.13403255e-01 -8.12492669e-01 1.93013918e+00 -7.97421992e-01
-1.95605218e+00 6.24338150e-01 -1.85289338e-01 -4.54931259e-01
8.40934396e-01 -6.84926212e-01 -1.67712495e-01 4.92230982e-01
1.62254632e-01 4.10967618e-01 7.55803585e-01 -7.62735248e-01
-2.72011459e-01 2.32218146e-01 -1.61486998e-01 4.74818915e-01
-2.87757009e-01 4.89416629e-01 -5.20210803e-01 -1.10170841e+00
-3.28071296e-01 -5.51791191e-01 -2.90004164e-01 -8.13487828e-01
-9.80143547e-01 -2.52456486e-01 8.06188062e-02 -1.26355040e+00
1.78015244e+00 -1.74383926e+00 6.21602178e-01 -4.37081009e-01
-6.30595386e-02 -3.98575924e-02 -2.68350273e-01 1.02778518e+00
2.57211953e-01 1.38519734e-01 -2.76378274e-01 -7.95031607e-01
1.40383229e-01 -3.93383712e-01 -7.11673915e-01 -1.28508225e-01
3.08375001e-01 8.96953106e-01 -9.27294314e-01 -5.82325757e-01
-2.10921347e-01 1.77389178e-02 -4.84776378e-01 6.22202516e-01
-7.02879071e-01 2.20599249e-01 -2.43561611e-01 2.31565490e-01
-3.37454639e-02 -9.46050659e-02 1.69705018e-01 -2.19294913e-02
-2.73454487e-01 8.72199357e-01 -7.46354699e-01 2.42239165e+00
-2.81517267e-01 9.18070495e-01 -3.59004557e-01 -1.98908016e-01
7.18304753e-01 7.07350135e-01 5.39028719e-02 -3.01095724e-01
-2.73686443e-02 -4.55827564e-02 -3.54375660e-01 -9.88049626e-01
1.23733008e+00 1.13387592e-01 -5.04965425e-01 7.23145485e-01
2.86005318e-01 -6.48851216e-01 7.11558163e-01 8.39744747e-01
1.51150620e+00 5.20639122e-01 1.90747455e-01 8.53754953e-02
2.44163945e-01 3.11412215e-01 3.52052778e-01 9.73275185e-01
2.21977815e-01 1.17201757e+00 9.44838226e-01 -1.45060390e-01
-1.59705329e+00 -9.20806110e-01 4.09618855e-01 1.01192451e+00
-6.01495266e-01 -1.24579465e+00 -1.32709837e+00 -4.56656426e-01
-5.46290934e-01 1.46108174e+00 -4.91317660e-01 7.28300437e-02
-8.39310229e-01 -3.76884103e-01 1.02726603e+00 6.49530888e-01
3.11277837e-01 -1.31669176e+00 -6.43494964e-01 4.73507464e-01
-5.94577551e-01 -1.02961659e+00 -8.78030002e-01 -2.52198786e-01
-6.65841281e-01 -6.02351904e-01 -7.16325045e-01 -4.64791000e-01
3.60173374e-01 -1.88844040e-01 1.17983425e+00 -5.04271030e-01
-2.74508208e-01 7.84163643e-03 -4.93968874e-01 -2.02202827e-01
-8.66931021e-01 3.83620113e-01 -3.49848866e-01 -1.70419306e-01
-1.26158888e-03 -5.77034712e-01 -2.37740144e-01 -3.04271340e-01
-7.24126399e-01 7.95095921e-01 6.43527150e-01 5.84616303e-01
2.57484406e-01 -4.35117602e-01 6.84287190e-01 -9.34168875e-01
1.18906200e+00 -1.38509616e-01 7.82337189e-02 3.53816032e-01
-1.74686179e-01 -8.77269134e-02 9.56441820e-01 -1.91691667e-01
-1.47102535e+00 8.94728228e-02 -7.65541643e-02 -2.86523532e-02
1.21611513e-01 4.03849155e-01 -1.66680366e-01 1.04335284e+00
9.14671421e-01 4.39332843e-01 -4.60416108e-01 -7.25695193e-01
7.50670314e-01 1.01280951e+00 9.73309457e-01 -3.18373173e-01
3.91551077e-01 -7.71058425e-02 -5.65285623e-01 -9.31994617e-01
-1.06138551e+00 -7.28091449e-02 -5.45569479e-01 -4.07708794e-01
1.07443225e+00 -1.05467808e+00 -1.65058374e-01 4.52299446e-01
-1.64530909e+00 -6.07448697e-01 -5.94520450e-01 1.34846851e-01
-9.50741291e-01 2.06763923e-01 -1.08193171e+00 -6.52613461e-01
-9.44369912e-01 -7.33025312e-01 1.01551938e+00 2.26101667e-01
-1.01336467e+00 -4.73478347e-01 2.72172272e-01 6.91534400e-01
3.36905390e-01 3.25937331e-01 6.37243211e-01 -7.94756114e-01
-4.15961802e-01 -1.97282135e-01 1.14799537e-01 3.91185105e-01
-2.10185692e-01 1.32816240e-01 -7.98033953e-01 7.65172020e-02
-1.96157932e-01 -7.00289369e-01 9.43245888e-01 7.31818303e-02
1.01003087e+00 -7.90288687e-01 1.99801594e-01 4.31847155e-01
1.04972374e+00 -6.11858033e-02 7.10247517e-01 -4.33431081e-02
6.64034069e-01 3.55920970e-01 3.55630845e-01 7.91743219e-01
3.63709152e-01 5.66546321e-01 -1.37283161e-01 4.76740688e-01
-5.06002665e-01 -7.77580321e-01 8.56581688e-01 1.12849057e+00
-1.28220469e-01 -5.74478686e-01 -5.48302352e-01 6.11121416e-01
-1.94602048e+00 -1.48952389e+00 -2.79658049e-01 1.85334647e+00
1.29906631e+00 3.06991845e-01 2.16598153e-01 -2.28737235e-01
4.24290925e-01 5.48271716e-01 -4.61024463e-01 -9.26176906e-01
-2.96978772e-01 1.62050322e-01 -3.52306627e-02 4.70648736e-01
-8.53859484e-01 1.20781434e+00 6.63732576e+00 8.59930754e-01
-6.61074519e-01 2.08114460e-01 4.41138446e-01 -9.57989216e-01
-3.65890115e-01 -1.18639089e-01 -6.94989979e-01 5.08092642e-01
1.36176729e+00 -6.00662231e-01 6.96496725e-01 8.00119877e-01
6.47253990e-01 -4.07513194e-02 -1.39021087e+00 7.00029254e-01
8.70966673e-01 -1.75954008e+00 1.29117653e-01 -4.71998632e-01
1.03968310e+00 -9.70938951e-02 -4.57896858e-01 3.27836156e-01
2.65508801e-01 -1.07740176e+00 1.22225714e+00 8.12730193e-01
1.06884849e+00 -5.78211069e-01 4.14577276e-01 3.21792245e-01
-6.64982557e-01 1.30737379e-01 -2.14582890e-01 -9.75499377e-02
7.49422371e-01 3.76092017e-01 -1.13137460e+00 3.44170600e-01
2.94966429e-01 7.71589875e-01 -7.68025219e-01 7.21627116e-01
-7.44091153e-01 8.43789816e-01 1.84335783e-01 -3.23733211e-01
-2.64581684e-02 2.05678970e-01 9.13485348e-01 1.77457130e+00
1.12289533e-01 2.53355920e-01 -7.97316507e-02 9.77274239e-01
-5.12097895e-01 1.26917079e-01 -3.42493683e-01 -5.12441695e-01
3.35296661e-01 1.38073015e+00 -5.17618179e-01 -8.22492182e-01
1.02026716e-01 1.64625990e+00 2.51928985e-01 2.06003517e-01
-8.35374892e-01 -8.35933030e-01 -4.06049527e-02 2.72012856e-02
2.99114525e-01 -2.82895446e-01 -5.85542083e-01 -1.26120687e+00
3.40369754e-02 -1.12940395e+00 9.62865949e-02 -1.16112709e+00
-9.82851148e-01 9.12433147e-01 -1.39618725e-01 -9.32282448e-01
-6.32500648e-01 1.59085795e-01 -1.07861018e+00 7.85934985e-01
-5.32787442e-01 -1.17767203e+00 -2.38542661e-01 5.37205935e-02
1.49497795e+00 -1.91733614e-01 8.09272170e-01 -1.30178425e-02
-7.36849368e-01 5.77906668e-01 1.02596775e-01 2.08785594e-01
9.30625319e-01 -1.46042609e+00 1.06896424e+00 1.12196159e+00
1.45508468e-01 6.08208358e-01 1.06559968e+00 -8.73579919e-01
-1.25882852e+00 -1.15796638e+00 1.41778755e+00 -7.48885274e-01
5.14487326e-01 -7.32137620e-01 -4.33995187e-01 6.68026865e-01
9.57416058e-01 -9.38805163e-01 7.25005805e-01 -2.20109180e-01
-1.77699670e-01 2.44508252e-01 -6.10683322e-01 7.78255701e-01
1.09882557e+00 -5.86078405e-01 -1.01517498e+00 7.37835169e-01
9.57335711e-01 -5.78868985e-01 -8.75848532e-01 -2.55402088e-01
6.49231136e-01 -7.32913494e-01 2.84584373e-01 -1.00461614e+00
1.60597599e+00 -5.87490536e-02 -1.10439092e-01 -1.42073345e+00
-3.00392240e-01 -1.32973599e+00 -2.93352187e-01 1.69776225e+00
6.17858052e-01 4.95524257e-01 5.97554028e-01 5.98827720e-01
-6.52228653e-01 -6.19262993e-01 -5.73666990e-01 -4.78930235e-01
-2.76323617e-01 -2.93220878e-01 1.82846650e-01 4.12543654e-01
3.93681765e-01 1.12971663e+00 -6.27113163e-01 -5.80967963e-01
3.80536526e-01 -3.46957929e-02 9.83732462e-01 -3.87703598e-01
-5.97294509e-01 -5.02064705e-01 3.33861589e-01 -1.05561852e+00
-1.31279543e-01 -9.02043760e-01 3.85313213e-01 -2.20652914e+00
5.28200150e-01 5.76562166e-01 3.45226556e-01 1.74189761e-01
-2.12734297e-01 1.66496083e-01 5.58434606e-01 2.15080306e-01
-1.00304163e+00 6.78278685e-01 1.00990880e+00 -3.17500606e-02
-4.07538146e-01 -3.85794751e-02 -7.87778258e-01 5.28211415e-01
5.74990392e-01 -3.36892217e-01 -3.04780275e-01 -6.09002352e-01
4.23912942e-01 3.87623250e-01 9.28354338e-02 -1.01383960e+00
3.46066296e-01 1.95169952e-02 4.02641267e-01 -8.26113641e-01
3.63737732e-01 3.45480889e-01 1.25115559e-01 6.36977330e-02
-1.08923948e+00 1.42294437e-01 1.82443217e-03 1.36438072e-01
8.45365450e-02 -4.70774829e-01 3.91362667e-01 -3.10352623e-01
-1.53559133e-01 -3.72420937e-01 -6.77031755e-01 4.43192482e-01
6.18157744e-01 -9.52337384e-02 -5.10609031e-01 -7.71047711e-01
-5.30994177e-01 -1.33765757e-01 3.13105315e-01 4.76573527e-01
5.59946835e-01 -1.14353514e+00 -1.30874574e+00 -3.58186156e-01
-1.36625767e-01 -9.03402269e-02 3.34814817e-01 5.37104189e-01
-7.63364375e-01 4.59266931e-01 -5.61755560e-02 2.24468485e-02
-1.13181853e+00 8.94564614e-02 -1.12256289e-01 -3.87390018e-01
-6.25811040e-01 1.17565966e+00 -4.01041359e-01 2.32483327e-01
1.77668020e-01 -9.05281454e-02 -1.62353635e-01 2.86252081e-01
1.08264601e+00 7.09644437e-01 -3.04598492e-02 -2.36822486e-01
8.44777003e-02 -2.08213879e-03 -3.17336172e-01 -8.40637982e-01
1.42763424e+00 2.74306666e-02 -5.96765503e-02 6.90445185e-01
8.69382083e-01 2.35875875e-01 -1.39300156e+00 1.22699387e-01
2.15491205e-01 -2.11302593e-01 -3.27271283e-01 -1.25188899e+00
-2.54406333e-01 6.89524233e-01 -4.56602335e-01 1.79908022e-01
8.30498815e-01 -3.66077684e-02 1.10510564e+00 3.75674158e-01
-2.80423313e-01 -1.61968279e+00 5.16536117e-01 6.46586180e-01
1.57219052e+00 -6.74161792e-01 3.34356219e-01 2.20245663e-02
-1.25501907e+00 1.20077109e+00 6.46469414e-01 -2.65121460e-01
-4.80698973e-01 1.69032857e-01 -2.45000869e-01 -5.47069386e-02
-1.38874924e+00 4.50380772e-01 2.60424584e-01 -4.41235676e-02
7.95484781e-01 -1.56068141e-02 -5.25110066e-01 1.22800827e+00
-7.86718488e-01 3.15708620e-03 1.10782123e+00 6.52714431e-01
-4.71782714e-01 -6.78106666e-01 -3.71012427e-02 4.91023600e-01
-6.20291173e-01 -3.32371831e-01 -1.00765002e+00 2.01623559e-01
-2.79118538e-01 1.17042875e+00 4.09961864e-02 -4.23407227e-01
6.31220520e-01 3.35403085e-01 6.40329838e-01 -9.70623672e-01
-1.16399336e+00 1.23016983e-01 8.82250905e-01 -6.11444950e-01
1.11646682e-01 -9.88851547e-01 -9.48990405e-01 -1.22393072e-01
1.18288815e-01 1.96869254e-01 8.22479963e-01 6.14629567e-01
4.96344894e-01 1.01113534e+00 5.52263081e-01 -9.17515695e-01
-9.17371213e-01 -1.53468287e+00 -9.56038088e-02 2.08391860e-01
-5.45710474e-02 3.97887319e-01 -9.96781960e-02 6.09630108e-01] | [12.393659591674805, 9.40395736694336] |
03ba238b-aa7c-4486-9041-94cd90d8d73e | explore-propose-and-assemble-an-interpretable | 1906.05210 | null | https://arxiv.org/abs/1906.05210v1 | https://arxiv.org/pdf/1906.05210v1.pdf | Explore, Propose, and Assemble: An Interpretable Model for Multi-Hop Reading Comprehension | Multi-hop reading comprehension requires the model to explore and connect relevant information from multiple sentences/documents in order to answer the question about the context. To achieve this, we propose an interpretable 3-module system called Explore-Propose-Assemble reader (EPAr). First, the Document Explorer iteratively selects relevant documents and represents divergent reasoning chains in a tree structure so as to allow assimilating information from all chains. The Answer Proposer then proposes an answer from every root-to-leaf path in the reasoning tree. Finally, the Evidence Assembler extracts a key sentence containing the proposed answer from every path and combines them to predict the final answer. Intuitively, EPAr approximates the coarse-to-fine-grained comprehension behavior of human readers when facing multiple long documents. We jointly optimize our 3 modules by minimizing the sum of losses from each stage conditioned on the previous stage's output. On two multi-hop reading comprehension datasets WikiHop and MedHop, our EPAr model achieves significant improvements over the baseline and competitive results compared to the state-of-the-art model. We also present multiple reasoning-chain-recovery tests and ablation studies to demonstrate our system's ability to perform interpretable and accurate reasoning. | ['Mohit Bansal', 'Yen-Chun Chen', 'Yichen Jiang', 'Nitish Joshi'] | 2019-06-12 | explore-propose-and-assemble-an-interpretable-1 | https://aclanthology.org/P19-1261 | https://aclanthology.org/P19-1261.pdf | acl-2019-7 | ['multi-hop-reading-comprehension'] | ['natural-language-processing'] | [ 5.77432394e-01 6.37759268e-01 2.66021565e-02 -6.36853695e-01
-1.34139669e+00 -7.37360597e-01 1.17576748e-01 8.59305978e-01
-4.51250046e-01 3.70962709e-01 7.72894681e-01 -7.87551403e-01
-3.34757566e-01 -6.60574198e-01 -8.80814493e-01 7.16388524e-02
2.97110498e-01 7.18553364e-01 3.29290956e-01 -2.96610296e-01
5.50840259e-01 -2.50113130e-01 -1.18751884e+00 9.52836394e-01
1.38089263e+00 6.01537526e-01 6.61895037e-01 1.19500518e+00
-4.92298394e-01 1.24608004e+00 -4.29259658e-01 -8.48706603e-01
-2.51279593e-01 -6.18826270e-01 -1.39405763e+00 -2.79888719e-01
6.76119864e-01 -6.08393848e-01 1.83238667e-02 8.91062677e-01
1.70465052e-01 9.60631669e-02 5.93193293e-01 -6.03548825e-01
-7.69760549e-01 1.46616554e+00 -5.88107526e-01 3.69293571e-01
9.74784672e-01 -1.10039692e-02 1.39404058e+00 -8.67366135e-01
6.17680132e-01 1.41332948e+00 3.15463752e-01 4.11040217e-01
-1.04373717e+00 -4.23218356e-03 8.16561103e-01 4.38669801e-01
-7.37512589e-01 -2.73757130e-01 4.49013352e-01 -4.56751175e-02
1.34955060e+00 5.33687651e-01 5.43100297e-01 9.63537753e-01
5.15134633e-01 1.28977227e+00 8.58884335e-01 -5.17698824e-01
1.81805417e-01 -2.97665894e-01 1.02008832e+00 1.12405634e+00
-2.62348223e-02 -6.04118228e-01 -9.97423530e-01 1.27942532e-01
-5.20059951e-02 -3.57852846e-01 -4.13224190e-01 2.68121213e-01
-1.12036705e+00 5.71292102e-01 4.95491803e-01 -1.09267488e-01
-4.15943623e-01 1.71299413e-01 2.15454653e-01 5.79628885e-01
8.31531882e-02 6.65203512e-01 -5.90334356e-01 -2.45827511e-01
-9.02985990e-01 4.64851171e-01 1.14062202e+00 1.12781894e+00
3.96241486e-01 -8.12216163e-01 -5.74343264e-01 6.92301393e-01
6.23248339e-01 2.72247285e-01 1.20049007e-01 -1.24994302e+00
1.23098731e+00 8.05048227e-01 -1.35259584e-01 -8.01990747e-01
-4.94480282e-01 -7.08848596e-01 -5.78325748e-01 -3.48155022e-01
4.87552136e-01 9.90037471e-02 -6.83712900e-01 1.54691374e+00
2.43128806e-01 -5.18797517e-01 2.95529604e-01 8.20604146e-01
9.67686534e-01 7.57400990e-01 1.24040216e-01 -1.43283665e-01
1.71504343e+00 -1.51843441e+00 -5.73468685e-01 -6.89215660e-01
5.53047121e-01 -7.17614591e-01 1.35025847e+00 6.72847211e-01
-1.70888865e+00 -4.31036264e-01 -9.59064066e-01 -8.94220114e-01
6.55332059e-02 -1.96968447e-02 2.45139346e-01 -1.49430409e-01
-9.84161377e-01 1.59516692e-01 -6.74392939e-01 -2.72815134e-02
2.31483221e-01 -9.82234180e-02 -4.52790819e-02 -5.55100083e-01
-1.15985632e+00 9.32007372e-01 3.30946118e-01 2.31720522e-01
-8.24396014e-01 -7.39944279e-01 -7.73088038e-01 6.22705936e-01
7.03509986e-01 -1.34353089e+00 1.96277761e+00 -3.21653515e-01
-1.23779154e+00 6.49604380e-01 -7.90099025e-01 -4.11414206e-01
5.01118362e-01 -7.93916285e-01 7.16086030e-02 3.27940255e-01
3.18307012e-01 7.00214028e-01 5.62213957e-01 -9.61018085e-01
-7.30384648e-01 -2.40184918e-01 4.97032166e-01 4.52515960e-01
3.48476142e-01 -5.65883145e-02 -8.43393862e-01 -4.24075425e-01
6.38432145e-01 -5.64837337e-01 9.57574137e-03 -7.43683577e-02
-8.67319822e-01 -4.42547530e-01 -5.17892838e-03 -1.16034782e+00
1.45654869e+00 -1.70746601e+00 7.28345275e-01 2.58970894e-02
5.70993423e-01 -3.26994509e-01 -4.70827848e-01 5.87688684e-01
2.96732157e-01 1.44096345e-01 -3.96172553e-01 -5.65401077e-01
1.60803586e-01 1.00612424e-01 -4.44416106e-01 -3.26707810e-01
5.97659126e-02 1.08687258e+00 -1.08599615e+00 -5.10372043e-01
-3.36567491e-01 -6.54504821e-02 -7.14838862e-01 3.39180201e-01
-9.70952928e-01 1.27648547e-01 -6.31300390e-01 5.16213894e-01
4.26171809e-01 -5.42586088e-01 7.28348270e-02 1.85063272e-03
2.96409696e-01 9.12208140e-01 -7.76265204e-01 2.00707817e+00
-6.60319746e-01 6.66014791e-01 4.13158722e-02 -3.57124388e-01
4.35949564e-01 -8.30571949e-02 -5.68625867e-01 -8.92019391e-01
-2.01227322e-01 2.47770786e-01 1.25295490e-01 -7.33801246e-01
7.45171905e-01 3.56184512e-01 -2.29086980e-01 7.30173111e-01
-1.67012006e-01 8.37119296e-03 3.41787785e-01 1.02510965e+00
1.44240630e+00 6.16398156e-02 2.30199486e-01 -6.76751211e-02
7.42646873e-01 -7.48153254e-02 1.01797991e-01 1.20192969e+00
2.17104137e-01 3.59197319e-01 8.25672746e-01 -1.77973390e-01
-4.49997038e-01 -1.40849257e+00 3.58036965e-01 1.45639181e+00
2.15233192e-01 -7.36472189e-01 -8.64534438e-01 -8.93206954e-01
-1.67248160e-01 1.41285396e+00 -4.89604145e-01 3.16682272e-02
-8.17359328e-01 -1.84759069e-02 3.67179841e-01 7.08704948e-01
6.67412221e-01 -9.80312586e-01 -8.00558090e-01 3.80806953e-01
-8.61244440e-01 -8.95286262e-01 -6.06139600e-01 2.60026813e-01
-7.81466782e-01 -1.05584085e+00 -1.53349549e-01 -6.84037745e-01
8.97865653e-01 1.26228467e-01 1.67322874e+00 5.45290768e-01
2.23432288e-01 4.00914967e-01 -6.26794934e-01 -3.21392924e-01
-6.74781799e-01 4.73445147e-01 -6.04790568e-01 -5.17385125e-01
3.06166440e-01 -1.05218843e-01 -6.11176729e-01 4.59632277e-02
-5.27445555e-01 6.43445909e-01 7.48735785e-01 5.51570177e-01
6.53616965e-01 -7.52947479e-02 5.23730457e-01 -9.24823105e-01
1.24427760e+00 -4.74374950e-01 -3.40300173e-01 9.68467057e-01
-4.79026377e-01 6.41877294e-01 5.80040693e-01 4.88629080e-02
-1.44942021e+00 -4.38378245e-01 -3.05939198e-01 4.60651249e-01
1.23766080e-01 9.40488338e-01 -1.33994445e-01 8.27657819e-01
5.04414737e-01 2.79102117e-01 -4.75203186e-01 -5.67981601e-01
7.92962551e-01 4.61407870e-01 9.05465186e-01 -8.73806059e-01
4.80832189e-01 -2.03474760e-01 -3.18424165e-01 -2.80144900e-01
-1.67254126e+00 -2.73501515e-01 -5.64523578e-01 -5.07435873e-02
9.17487800e-01 -7.68395662e-01 -7.45704651e-01 2.52869785e-01
-1.75624871e+00 -3.51057023e-01 -2.24100709e-01 9.91250277e-02
-3.71625543e-01 1.76809445e-01 -8.22738528e-01 -6.97508156e-01
-6.94099844e-01 -1.20880580e+00 1.10580456e+00 4.08531785e-01
-7.46040165e-01 -9.32635844e-01 -1.69534862e-01 8.88973653e-01
1.11479700e-01 -3.08486670e-01 1.45936549e+00 -5.70870221e-01
-1.04727149e+00 2.06899807e-01 -2.88280249e-01 2.46880949e-02
-4.33747053e-01 -1.71368554e-01 -7.32947648e-01 7.23975673e-02
4.64057326e-02 -6.01793587e-01 1.28863513e+00 1.03229731e-01
1.17767751e+00 -6.12934470e-01 -2.72934496e-01 2.05068544e-01
9.62782919e-01 -2.14043677e-01 4.70718473e-01 1.83837429e-01
4.89744127e-01 8.57323587e-01 5.42002916e-01 -6.24777898e-02
1.14599693e+00 1.52351648e-01 2.74827451e-01 3.11545223e-01
-1.31303862e-01 -7.69860625e-01 4.32897508e-01 1.09555614e+00
3.90214771e-01 -7.46953070e-01 -8.02679956e-01 4.02750820e-01
-1.92032027e+00 -8.12757611e-01 -4.86145705e-01 1.62609076e+00
1.17782784e+00 5.21400332e-01 -5.18444955e-01 -6.27046973e-02
4.98608239e-02 1.09260567e-01 -7.38260031e-01 -8.67314935e-01
4.04992029e-02 1.90819830e-01 -3.01795661e-01 1.07128489e+00
-4.46173161e-01 9.36760724e-01 5.71334410e+00 5.40773571e-01
-1.78525701e-01 -6.18221126e-02 5.76192737e-01 -1.04487471e-01
-9.20547605e-01 1.71400160e-01 -9.68834221e-01 8.97864699e-02
7.67811060e-01 1.03888549e-01 5.62360108e-01 3.10032040e-01
9.17304978e-02 -7.87822127e-01 -1.52674532e+00 1.97824314e-01
1.60803929e-01 -1.17678511e+00 1.60816744e-01 -5.49574673e-01
4.64012951e-01 -1.36598080e-01 -6.21900707e-02 3.47354382e-01
4.51764941e-01 -1.02234328e+00 1.08037698e+00 8.80139291e-01
3.03772569e-01 -5.19932985e-01 5.19572377e-01 8.65460455e-01
-1.11740208e+00 -1.55064851e-01 -8.10078606e-02 -2.04185858e-01
5.76950192e-01 5.75033188e-01 -7.57817149e-01 4.84796911e-01
5.34823716e-01 4.34841007e-01 -8.81006896e-01 6.76352322e-01
-1.27353001e+00 6.84504032e-01 -1.80414006e-01 -3.41910064e-01
2.81028926e-01 6.41465485e-02 7.02638626e-01 1.09420383e+00
3.08748055e-02 3.71055245e-01 -1.97330743e-01 1.18383253e+00
-4.88748610e-01 -1.47091001e-01 1.52779251e-01 1.43078089e-01
6.86459601e-01 1.01539993e+00 -4.35758561e-01 -4.33626443e-01
-3.80301684e-01 1.16418576e+00 9.70769286e-01 2.75788814e-01
-5.33382297e-01 -2.92570680e-01 1.25141665e-01 -6.39649332e-02
7.92341754e-02 -1.07136331e-01 -4.94666874e-01 -1.12765014e+00
4.86601502e-01 -1.21518207e+00 7.71094978e-01 -1.29773831e+00
-1.17408788e+00 4.71063823e-01 2.66979244e-02 -4.52061951e-01
-2.43800536e-01 -3.62213999e-01 -7.93076694e-01 1.14861500e+00
-1.46639693e+00 -1.08802021e+00 -3.44753921e-01 6.79087490e-02
1.18162036e+00 2.31314436e-01 6.09788954e-01 -4.30371702e-01
-5.09563088e-01 6.15637958e-01 -3.74032825e-01 -1.92731902e-01
3.32304329e-01 -1.65026879e+00 8.59888196e-01 1.10537517e+00
1.69782281e-01 1.06525493e+00 6.52828276e-01 -7.68939435e-01
-1.40737927e+00 -6.54372692e-01 1.41455686e+00 -8.42719138e-01
5.38933814e-01 -2.60496736e-01 -1.24336934e+00 8.59203875e-01
6.89340234e-01 -9.35070217e-01 4.92126614e-01 5.02241433e-01
-6.24352455e-01 2.56322343e-02 -7.09183455e-01 8.75127792e-01
1.12064672e+00 -5.49052238e-01 -1.47209430e+00 1.82152331e-01
1.13119960e+00 -6.95811510e-01 -5.89631319e-01 -7.05692321e-02
6.78204596e-01 -9.29168761e-01 7.29470670e-01 -6.15521729e-01
1.17860651e+00 -2.65931368e-01 -6.41527995e-02 -1.34816515e+00
-1.59645513e-01 -6.74908280e-01 -5.40624082e-01 1.06637096e+00
1.05579031e+00 -2.46549189e-01 2.35609174e-01 8.23835671e-01
-3.52282703e-01 -1.06718159e+00 -6.64099514e-01 -7.78041705e-02
1.57949060e-01 -6.33248270e-01 5.54772079e-01 6.05546869e-02
3.61691982e-01 9.34934556e-01 1.76961765e-01 2.79873729e-01
6.60333753e-01 3.50622922e-01 5.23207963e-01 -8.43513668e-01
-6.73542500e-01 -4.10968453e-01 7.46640027e-01 -1.91293967e+00
2.34986171e-01 -1.11455238e+00 3.15014094e-01 -2.27943921e+00
5.36354721e-01 -9.43080857e-02 9.15779918e-02 3.90851676e-01
-8.05526972e-01 -6.92963600e-01 3.76481533e-01 9.40731317e-02
-1.10430491e+00 2.52509683e-01 1.59845698e+00 -3.88961256e-01
-1.89872503e-01 2.56492961e-02 -1.29247224e+00 8.01092148e-01
4.87347662e-01 -2.43597806e-01 -7.54599094e-01 -1.19992864e+00
9.47327673e-01 4.80207592e-01 3.18873048e-01 -6.22693837e-01
7.26984859e-01 8.92914012e-02 3.94243479e-01 -1.10674858e+00
1.96785778e-02 -4.69102532e-01 -3.45977098e-01 3.43883216e-01
-1.21592879e+00 3.81772876e-01 -7.77618436e-04 6.58113420e-01
-2.87337881e-02 -5.74337065e-01 2.32111722e-01 -2.39740297e-01
-4.33159798e-01 -3.58962774e-01 -4.95697439e-01 5.23082256e-01
4.33647901e-01 2.10621983e-01 -8.71295214e-01 -5.06815553e-01
-6.59438252e-01 9.41683412e-01 1.07755192e-01 4.13175404e-01
9.62674856e-01 -6.01683855e-01 -9.20742869e-01 -4.18423442e-03
1.23396151e-01 5.90582967e-01 3.25436026e-01 6.83282495e-01
-5.01232862e-01 4.68026370e-01 3.63025904e-01 -4.35152858e-01
-1.26195848e+00 3.33151460e-01 2.92015374e-01 -8.32026839e-01
-7.30343044e-01 1.25316429e+00 1.05588831e-01 -4.61931199e-01
3.90584588e-01 -9.71418321e-01 -1.83476433e-01 -3.75662930e-02
7.93632030e-01 4.48423594e-01 -2.45385207e-02 9.17073786e-02
-2.25204065e-01 5.19741714e-01 -5.67558825e-01 -3.84813100e-01
9.00931180e-01 -5.36390483e-01 -3.30882162e-01 2.57878780e-01
9.14595723e-01 1.77510768e-01 -1.17286205e+00 -2.07726702e-01
1.59776062e-01 -9.31346193e-02 -2.61717260e-01 -1.62384939e+00
-3.98837686e-01 1.02718973e+00 -3.55655700e-01 -4.84473910e-03
1.10038304e+00 3.77237111e-01 9.45310116e-01 7.71341205e-01
1.80733427e-01 -8.53863418e-01 2.93571830e-01 7.53733158e-01
1.33503437e+00 -1.10722017e+00 2.46477500e-02 -4.94339973e-01
-6.95750356e-01 1.15370893e+00 8.18246961e-01 5.69471836e-01
8.93434510e-02 2.49675751e-01 -7.10344687e-02 -3.56873691e-01
-1.36678958e+00 1.48147732e-01 6.11138880e-01 4.62259687e-02
4.57159311e-01 5.38209155e-02 -1.87513947e-01 8.76180887e-01
-5.51232278e-01 -1.09281115e-01 5.15151858e-01 9.16687012e-01
-6.73925936e-01 -9.00488973e-01 -7.57705048e-02 6.05821133e-01
-9.33189318e-02 -4.79261279e-01 -7.08667159e-01 2.92596161e-01
-3.18324596e-01 1.31017828e+00 -1.24589065e-02 -1.24358453e-01
5.15667379e-01 1.58068866e-01 5.88691652e-01 -5.62984347e-01
-8.64286721e-01 -2.71968096e-01 4.52248871e-01 -7.37095237e-01
1.16215378e-01 -4.68270898e-01 -1.70396578e+00 -2.95650512e-01
-1.74685076e-01 1.05552167e-01 2.25144431e-01 1.22244489e+00
4.12565976e-01 9.25391555e-01 9.99798551e-02 6.73198998e-02
-7.89498389e-01 -9.13845241e-01 8.73729736e-02 5.30817173e-02
5.13426900e-01 6.34675622e-02 -1.50966927e-01 8.66906494e-02] | [11.118263244628906, 7.94274377822876] |
42a71d21-c325-43f2-8e40-11c7a8b39152 | learning-video-conditioned-policies-for | 2305.06289 | null | https://arxiv.org/abs/2305.06289v1 | https://arxiv.org/pdf/2305.06289v1.pdf | Learning Video-Conditioned Policies for Unseen Manipulation Tasks | The ability to specify robot commands by a non-expert user is critical for building generalist agents capable of solving a large variety of tasks. One convenient way to specify the intended robot goal is by a video of a person demonstrating the target task. While prior work typically aims to imitate human demonstrations performed in robot environments, here we focus on a more realistic and challenging setup with demonstrations recorded in natural and diverse human environments. We propose Video-conditioned Policy learning (ViP), a data-driven approach that maps human demonstrations of previously unseen tasks to robot manipulation skills. To this end, we learn our policy to generate appropriate actions given current scene observations and a video of the target task. To encourage generalization to new tasks, we avoid particular tasks during training and learn our policy from unlabelled robot trajectories and corresponding robot videos. Both robot and human videos in our framework are represented by video embeddings pre-trained for human action recognition. At test time we first translate human videos to robot videos in the common video embedding space, and then use resulting embeddings to condition our policies. Notably, our approach enables robot control by human demonstrations in a zero-shot manner, i.e., without using robot trajectories paired with human instructions during training. We validate our approach on a set of challenging multi-task robot manipulation environments and outperform state of the art. Our method also demonstrates excellent performance in a new challenging zero-shot setup where no paired data is used during training. | ['Ivan Laptev', 'Cordelia Schmid', 'Elliot Chane-Sane'] | 2023-05-10 | null | null | null | null | ['action-recognition-in-videos', 'action-recognition', 'robot-manipulation'] | ['computer-vision', 'computer-vision', 'robots'] | [ 4.59918529e-01 1.23903016e-02 -4.24882360e-02 -1.75918430e-01
-5.60383499e-01 -7.06107199e-01 1.00623631e+00 -3.97448242e-01
-8.78333390e-01 7.31914878e-01 8.61386955e-02 -8.80016461e-02
1.48531273e-01 -1.60146803e-01 -1.17580175e+00 -6.17246270e-01
-1.30328804e-01 8.87292445e-01 7.27352686e-03 -2.16870219e-01
1.48882419e-01 5.99253476e-01 -1.53534079e+00 1.18452646e-01
6.27720594e-01 3.48015457e-01 7.97714949e-01 1.00545943e+00
5.64064741e-01 1.02490556e+00 -4.32757884e-01 2.42455199e-01
5.92400372e-01 -3.80132407e-01 -7.84506559e-01 5.75927615e-01
2.55132228e-01 -7.16289878e-01 -7.77658343e-01 9.78020966e-01
2.20028207e-01 6.91979527e-01 8.64778459e-01 -1.83804142e+00
-4.80053693e-01 3.00312132e-01 -4.54536490e-02 -2.21927002e-01
8.18559349e-01 9.73995566e-01 6.63044691e-01 -6.72788262e-01
9.36375260e-01 1.34438515e+00 1.57309920e-01 8.01145196e-01
-1.41069067e+00 -4.12833601e-01 2.40761772e-01 3.55482012e-01
-9.83391583e-01 -3.72025669e-01 4.14001554e-01 -7.76442885e-01
1.17140067e+00 -3.66671771e-01 5.85028946e-01 1.76086318e+00
1.36750072e-01 9.21511233e-01 7.27162898e-01 -2.15917960e-01
2.56811738e-01 3.57492827e-02 -3.95921141e-01 6.42569244e-01
-8.28452557e-02 4.35940534e-01 -1.36673793e-01 4.98323776e-02
9.77788091e-01 3.88455272e-01 -6.29292250e-01 -1.09486794e+00
-1.94183266e+00 8.46509635e-01 4.35468853e-01 4.88277748e-02
-8.52126479e-01 4.19012487e-01 4.99576807e-01 7.09876657e-01
-3.97625238e-01 8.25917184e-01 -5.10934293e-01 -2.68290609e-01
-2.50469655e-01 5.42172909e-01 8.98519754e-01 1.34877431e+00
6.45801067e-01 1.10066853e-01 -2.68996596e-01 3.93274367e-01
-6.86034933e-02 6.32387221e-01 5.47762334e-01 -1.49543321e+00
5.00014961e-01 1.93970397e-01 6.45012438e-01 -5.58879912e-01
-2.37654969e-01 2.90642083e-01 -2.18197972e-01 6.65491045e-01
5.15663862e-01 -8.26371163e-02 -1.07582200e+00 1.72630525e+00
2.01120943e-01 2.83881277e-01 5.63804984e-01 1.14419317e+00
1.35937303e-01 8.75469625e-01 1.31791949e-01 -5.65698463e-03
1.10123479e+00 -1.40206230e+00 -3.93406034e-01 -5.27843535e-01
8.05293739e-01 -2.34057948e-01 1.34122539e+00 4.67006713e-01
-5.51115632e-01 -6.74905241e-01 -7.62452722e-01 8.81662127e-03
-3.35855819e-02 2.50920147e-01 1.43144295e-01 -2.32132211e-01
-8.98367047e-01 6.47234797e-01 -1.08242476e+00 -7.51617432e-01
1.59107475e-03 4.10493881e-01 -9.17497039e-01 -4.49646622e-01
-8.04029882e-01 1.17657721e+00 6.35821223e-01 -1.80194318e-01
-1.93721116e+00 -3.03473562e-01 -1.42657983e+00 -6.06759712e-02
8.23742867e-01 -6.46470249e-01 1.66583622e+00 -8.81845593e-01
-1.72643793e+00 5.61545968e-01 1.83099404e-01 -5.69293976e-01
6.04813337e-01 -5.54129779e-01 1.63714558e-01 4.80757862e-01
2.72006392e-01 1.14739180e+00 1.06364894e+00 -1.43961918e+00
-7.39944637e-01 2.46983115e-02 4.49780464e-01 5.39037704e-01
-3.71595621e-02 -2.15100020e-01 -4.41870868e-01 -2.25881130e-01
-5.14697731e-01 -1.56128550e+00 -3.23313743e-01 1.27546906e-01
-1.51291326e-01 -2.08289921e-01 9.21851039e-01 -3.61065567e-01
1.71456337e-01 -2.16916895e+00 8.50015223e-01 -1.19440436e-01
-9.48721468e-02 9.08696130e-02 -5.42665958e-01 6.19190097e-01
-2.57292874e-02 -5.13182461e-01 -3.05091560e-01 -3.18734199e-01
2.91612923e-01 6.35706425e-01 -2.71505356e-01 6.92434549e-01
3.06078106e-01 8.54440570e-01 -1.41026008e+00 -2.86327183e-01
5.77220201e-01 4.24024343e-01 -7.12101758e-01 8.33855033e-01
-5.26501060e-01 1.08867645e+00 -3.96460682e-01 2.30553597e-01
-1.84339389e-01 -1.09049879e-01 2.86714196e-01 4.46987972e-02
2.65982486e-02 -1.07274048e-01 -8.86784494e-01 2.00391650e+00
-5.67020476e-01 5.79989076e-01 1.70890898e-01 -1.28857422e+00
5.97387433e-01 5.48522532e-01 5.21680057e-01 -2.57797599e-01
1.35771886e-01 1.13773547e-01 2.23735943e-01 -1.08713126e+00
2.79962063e-01 3.92011963e-02 -3.05787861e-01 3.37929964e-01
4.01441395e-01 -5.47845125e-01 2.23035261e-01 1.32881388e-01
1.37776315e+00 8.09159994e-01 3.36488664e-01 1.58123419e-01
2.64261335e-01 4.72959876e-01 3.17034572e-01 8.22881997e-01
-4.76515383e-01 3.94593239e-01 3.91986072e-01 -1.61806822e-01
-1.47953737e+00 -9.89538908e-01 5.97304285e-01 1.22628069e+00
9.95102525e-02 3.00461296e-02 -4.07259226e-01 -7.79493809e-01
4.81874943e-02 9.54679251e-01 -5.99035561e-01 -1.11644551e-01
-9.61330950e-01 2.01310858e-01 1.20857701e-01 4.51424867e-01
1.69909164e-01 -1.73973560e+00 -1.27420151e+00 1.89190939e-01
-5.97289801e-02 -1.52597404e+00 -5.78363478e-01 2.66046196e-01
-3.30014527e-01 -1.20398748e+00 -8.84148121e-01 -1.22148633e+00
8.90879810e-01 3.50099146e-01 7.55602896e-01 -3.94253671e-01
-2.03232065e-01 1.20363665e+00 -4.92055297e-01 6.52906597e-02
-6.26444340e-01 -2.61204571e-01 5.14386654e-01 -1.37711808e-01
-1.82861816e-02 -4.46988910e-01 -4.13606793e-01 2.28664219e-01
-6.90523565e-01 3.34091745e-02 8.93110037e-01 1.23791337e+00
1.58829048e-01 -2.82288224e-01 1.80951491e-01 -4.67709213e-01
5.31016946e-01 -6.41706109e-01 -5.43825507e-01 7.75543079e-02
-5.66773005e-02 2.08665714e-01 8.94553483e-01 -9.61354315e-01
-6.49803340e-01 6.33377194e-01 3.20227653e-01 -1.03488743e+00
-5.68421483e-01 1.57635450e-01 4.10520546e-02 2.29825169e-01
6.22796595e-01 2.80348003e-01 2.74563789e-01 5.18389186e-03
5.05927622e-01 4.24462169e-01 1.04384172e+00 -8.60257030e-01
8.91722798e-01 1.62589386e-01 -2.65538454e-01 -7.46495724e-01
-6.27868623e-02 -5.06244004e-01 -7.89282680e-01 -2.60987431e-01
1.08311772e+00 -9.73014355e-01 -1.04160821e+00 2.04544023e-01
-1.19194508e+00 -1.07204998e+00 -1.47641435e-01 8.96350503e-01
-1.28889656e+00 9.74268243e-02 -4.52523082e-01 -6.96160972e-01
2.87829816e-01 -1.65432656e+00 1.20271170e+00 2.11317819e-02
-2.81416953e-01 -7.12582290e-01 -1.35332467e-02 -4.13606167e-02
1.60207339e-02 3.80276084e-01 5.20153403e-01 -8.30967546e-01
-5.05478442e-01 -6.74846247e-02 8.86579305e-02 2.74720430e-01
1.01712026e-01 -4.18727309e-01 -4.97498691e-01 -7.15653062e-01
2.93475259e-02 -1.10431671e+00 4.54707295e-01 8.64750743e-02
7.01642632e-01 -3.01063627e-01 -5.62319577e-01 3.48369360e-01
1.15182710e+00 4.78112906e-01 3.52004200e-01 4.50059831e-01
6.72370017e-01 9.12424088e-01 1.04754162e+00 2.46950909e-01
2.46757686e-01 6.58418655e-01 5.29792249e-01 3.04567516e-01
3.27756703e-01 -4.70458716e-01 9.57959473e-01 2.31556609e-01
2.17652619e-01 -3.38416249e-01 -8.34219635e-01 6.93650723e-01
-2.15783119e+00 -1.01690531e+00 4.46624845e-01 1.91309762e+00
5.07969201e-01 7.65898600e-02 1.91406175e-01 -2.74634540e-01
5.55505037e-01 -1.89996213e-01 -9.20624375e-01 -2.18304247e-02
4.74454939e-01 -1.80212185e-01 4.16954547e-01 5.17781615e-01
-1.19405007e+00 1.08021212e+00 5.13790321e+00 7.89144263e-02
-1.02379036e+00 -3.40615921e-02 -1.00322023e-01 -1.90118447e-01
2.97339171e-01 -8.81692544e-02 -3.52687508e-01 2.66625464e-01
7.91303396e-01 -6.07604794e-02 6.69532716e-01 1.11029112e+00
3.72994810e-01 -9.62061360e-02 -1.80219901e+00 1.03748131e+00
1.39047116e-01 -8.52718234e-01 -2.41771922e-01 -4.65195291e-02
5.05332053e-01 1.70918196e-01 -4.83992603e-03 9.80601072e-01
7.88132727e-01 -1.08917749e+00 6.59933627e-01 2.26439893e-01
7.76200235e-01 -2.18510598e-01 3.66211593e-01 7.63619721e-01
-8.47219884e-01 -5.54817140e-01 -2.06182957e-01 -1.76109105e-01
4.67133582e-01 -6.07161522e-01 -1.32772875e+00 2.28276476e-01
4.79548156e-01 8.85860980e-01 4.70948219e-03 7.57944643e-01
-4.67583358e-01 8.53036940e-02 -3.02741945e-01 -8.81012157e-02
6.09616756e-01 -2.33207494e-01 6.04825556e-01 9.58570242e-01
3.85585010e-01 1.98068187e-01 8.91034842e-01 6.46250010e-01
1.73475921e-01 -3.29235077e-01 -1.36215973e+00 -1.17681362e-01
3.87732387e-01 9.72973704e-01 -4.92218316e-01 -4.04006898e-01
-4.48946595e-01 1.54884815e+00 5.44942439e-01 6.90224826e-01
-8.93093467e-01 -2.64088601e-01 7.41465032e-01 -3.68675053e-01
5.39367080e-01 -7.10586011e-01 6.70679212e-01 -1.15182829e+00
-9.33610499e-02 -1.18068337e+00 -7.52949044e-02 -9.64080513e-01
-8.48456621e-01 5.68377733e-01 3.56014878e-01 -1.55739415e+00
-8.31401825e-01 -9.82992411e-01 -4.77553517e-01 4.67495590e-01
-1.23794866e+00 -9.29460466e-01 -3.85036081e-01 7.29963481e-01
1.11508179e+00 -3.22260529e-01 7.56137431e-01 -4.15414311e-02
-3.18131268e-01 1.51405469e-01 1.28278416e-02 8.63543674e-02
9.06100869e-01 -1.22386146e+00 4.22897041e-01 5.67677915e-01
-3.71069554e-03 5.18513024e-01 1.08112764e+00 -5.82078815e-01
-1.68089855e+00 -1.00144482e+00 8.09336007e-02 -6.13437116e-01
7.76020944e-01 -5.10491788e-01 -7.78644562e-01 1.21373916e+00
3.15543205e-01 1.13541543e-01 -5.42763881e-02 -5.61271012e-01
-6.99209720e-02 4.59902912e-01 -9.44633722e-01 8.53041708e-01
1.08851159e+00 -4.79452848e-01 -9.62635219e-01 6.07148409e-01
9.26163435e-01 -3.56945157e-01 -5.92778265e-01 3.74129444e-01
5.08157909e-01 -3.78799796e-01 8.37961257e-01 -8.64533603e-01
4.80874598e-01 -4.83147025e-01 -2.46684358e-01 -1.90049994e+00
-2.31753349e-01 -6.23025298e-01 8.66986737e-02 4.22688365e-01
2.29124129e-01 -5.05091548e-01 5.69813848e-01 4.37540948e-01
-4.38481182e-01 -4.66942906e-01 -4.64957416e-01 -1.06191921e+00
-1.05566993e-01 -9.67475921e-02 4.60675247e-02 7.50742793e-01
2.61570156e-01 2.92060286e-01 -6.79062605e-01 3.22310418e-01
4.44155693e-01 -1.28564328e-01 1.48684502e+00 -7.44654417e-01
-4.05691743e-01 -1.53716356e-01 -3.44809264e-01 -1.40013111e+00
7.98456013e-01 -7.56875992e-01 8.61006856e-01 -1.37926555e+00
8.15839097e-02 -8.52824226e-02 2.07773864e-01 4.65166390e-01
-3.98496762e-02 -1.95020661e-01 4.87256765e-01 4.53061193e-01
-5.93675494e-01 8.82105291e-01 1.35352969e+00 -2.66397417e-01
-2.22725838e-01 -3.04316312e-01 7.79059827e-02 5.64304233e-01
7.13846624e-01 -3.94969463e-01 -7.78038383e-01 -5.60354888e-01
-4.81609464e-01 2.68894434e-01 6.61868930e-01 -1.14125931e+00
3.02866727e-01 -4.92177099e-01 1.38966292e-01 7.41500705e-02
7.10794628e-01 -1.01604223e+00 -6.87264055e-02 8.69175971e-01
-5.22124708e-01 3.30342591e-01 1.03166802e-02 9.14814770e-01
3.47926170e-02 -2.55772769e-01 5.39017618e-01 -3.62004757e-01
-1.20576346e+00 2.69007474e-01 -5.75176179e-01 -1.49014115e-01
1.59786701e+00 2.18861382e-02 -2.18840376e-01 -6.40012681e-01
-8.93632233e-01 8.03077221e-01 8.64406943e-01 6.04576230e-01
7.52075732e-01 -1.13161457e+00 -6.53536975e-01 1.36790872e-01
3.64858598e-01 -4.84145842e-02 -1.42480209e-01 7.09617555e-01
-5.24065077e-01 4.69916999e-01 -6.26622021e-01 -8.15458179e-01
-9.31327462e-01 1.02257335e+00 1.50080591e-01 -2.21631140e-03
-1.04992127e+00 5.66760898e-01 5.15971601e-01 -8.55619073e-01
4.70402360e-01 -3.66809398e-01 -2.45112300e-01 -5.69028735e-01
2.08484307e-01 -4.21231240e-02 -6.18138671e-01 -7.26223528e-01
-1.29367977e-01 2.09436625e-01 -1.05389796e-01 -5.02000809e-01
1.18582916e+00 1.00498006e-01 3.91773045e-01 6.10226750e-01
1.32591212e+00 -5.66273212e-01 -2.00026894e+00 -1.17650613e-01
-2.22006366e-02 -4.26886111e-01 -6.23708427e-01 -4.66433108e-01
-4.36536342e-01 6.80349052e-01 3.26891184e-01 -2.29602218e-01
6.04175150e-01 2.10828017e-02 6.18508041e-01 1.12689030e+00
7.97032952e-01 -8.97822857e-01 7.58153975e-01 7.08328366e-01
1.18040752e+00 -1.59739804e+00 -3.59717458e-01 4.33290266e-02
-1.26285386e+00 9.60590243e-01 9.64973509e-01 -5.23799419e-01
1.61791354e-01 1.90085254e-03 -8.27139337e-03 -8.85586590e-02
-9.46494877e-01 -1.88687265e-01 -2.03021854e-01 1.00874996e+00
-2.02754572e-01 -1.06699564e-01 3.28041106e-01 1.65884700e-02
-2.40619220e-02 -1.14260744e-02 7.27461874e-01 1.28356147e+00
-4.77190256e-01 -7.72417426e-01 -3.85504752e-01 2.27655545e-01
2.08595067e-01 3.18935126e-01 -1.61609605e-01 1.02471840e+00
-2.46271640e-01 7.26061225e-01 8.79484937e-02 -4.45858628e-01
4.90909666e-01 2.58883059e-01 7.60394633e-01 -1.05818880e+00
-1.85228392e-01 -2.01966524e-01 -3.94389965e-02 -7.62610674e-01
-3.69893998e-01 -9.97070134e-01 -1.33239126e+00 9.26719904e-02
1.86748117e-01 1.55610085e-01 4.90634799e-01 9.59297717e-01
1.79487556e-01 5.49266815e-01 5.05103827e-01 -1.82724977e+00
-9.75875974e-01 -1.09312940e+00 -1.54306933e-01 8.60038519e-01
7.09643662e-01 -1.12471402e+00 -4.07068491e-01 3.48334819e-01] | [4.552096366882324, 0.7894750237464905] |
937580b3-b935-42e2-83a9-b478e59063b9 | evaluation-of-induced-expert-knowledge-in | 2301.01817 | null | https://arxiv.org/abs/2301.01817v1 | https://arxiv.org/pdf/2301.01817v1.pdf | Evaluation of Induced Expert Knowledge in Causal Structure Learning by NOTEARS | Causal modeling provides us with powerful counterfactual reasoning and interventional mechanism to generate predictions and reason under various what-if scenarios. However, causal discovery using observation data remains a nontrivial task due to unobserved confounding factors, finite sampling, and changes in the data distribution. These can lead to spurious cause-effect relationships. To mitigate these challenges in practice, researchers augment causal learning with known causal relations. The goal of the paper is to study the impact of expert knowledge on causal relations in the form of additional constraints used in the formulation of the nonparametric NOTEARS. We provide a comprehensive set of comparative analyses of biasing the model using different types of knowledge. We found that (i) knowledge that corrects the mistakes of the NOTEARS model can lead to statistically significant improvements, (ii) constraints on active edges have a larger positive impact on causal discovery than inactive edges, and surprisingly, (iii) the induced knowledge does not correct on average more incorrect active and/or inactive edges than expected. We also demonstrate the behavior of the model and the effectiveness of domain knowledge on a real-world dataset. | ['Gabriel Terejanu', 'Rezaur Rashid', 'Jawad Chowdhury'] | 2023-01-04 | null | null | null | null | ['causal-discovery'] | ['knowledge-base'] | [ 4.18083876e-01 5.57712555e-01 -7.73362458e-01 -1.87612414e-01
-3.31531912e-01 -5.41485548e-01 5.33280432e-01 3.02718639e-01
-2.08341494e-01 1.34675562e+00 6.10309243e-01 -7.95209169e-01
-7.66364872e-01 -1.00706458e+00 -1.23384416e+00 -3.57653081e-01
-3.10035706e-01 2.20991746e-01 -3.81811932e-02 2.67925054e-01
3.14593911e-01 1.76780298e-01 -1.16347909e+00 1.61974609e-01
1.06798267e+00 3.17937136e-01 -8.09126943e-02 1.58342078e-01
1.29293039e-01 9.75632250e-01 -3.85969609e-01 -4.64262575e-01
3.14010978e-01 -3.68972778e-01 -4.82523829e-01 -1.81232005e-01
3.61022383e-01 -3.84650052e-01 -2.54493803e-01 8.23282480e-01
3.20509166e-01 1.10346012e-01 8.09899926e-01 -1.42117286e+00
-7.58546054e-01 1.27608991e+00 -7.14944065e-01 2.41357520e-01
3.25547367e-01 1.98373452e-01 1.07933736e+00 -3.36659789e-01
7.13844597e-01 1.43518674e+00 6.47478282e-01 8.26305673e-02
-1.39333582e+00 -9.67787445e-01 5.23115754e-01 -7.10727938e-04
-1.10683548e+00 -3.53663534e-01 5.36121428e-01 -5.82225740e-01
3.81530344e-01 2.70683944e-01 4.86741424e-01 1.28899407e+00
2.00771317e-01 1.34531587e-01 1.24553609e+00 -5.29294491e-01
3.40259820e-01 1.17912948e-01 1.41456231e-01 5.42114735e-01
9.82730746e-01 7.07813740e-01 -6.06884480e-01 -5.44265807e-01
9.09787476e-01 -9.66404974e-02 -2.79376209e-01 -2.09662840e-01
-1.04355609e+00 1.00888288e+00 3.71791422e-01 1.79201737e-02
-6.09103262e-01 2.56878465e-01 -5.16881719e-02 2.44578242e-01
3.24631363e-01 9.21979308e-01 -7.31304228e-01 1.54570729e-01
-5.05199313e-01 4.63758111e-01 7.07416058e-01 4.13660228e-01
4.31369066e-01 -1.26921058e-01 -4.47311908e-01 3.52613419e-01
9.56943724e-03 5.02828419e-01 -4.02572453e-02 -1.08582222e+00
4.03589606e-01 5.28100610e-01 5.51331878e-01 -1.08745813e+00
-4.28889751e-01 -5.34909546e-01 -4.52386409e-01 -8.10549036e-02
9.33454156e-01 -6.77991688e-01 -8.15040708e-01 2.04180264e+00
3.06204021e-01 4.26818848e-01 -3.53098154e-01 7.17486918e-01
2.52026409e-01 2.92200893e-01 4.05128837e-01 -5.79319358e-01
1.08657277e+00 -1.87036797e-01 -8.80896568e-01 -1.41087189e-01
7.29078829e-01 -4.42291856e-01 9.89946961e-01 -6.33959994e-02
-9.32222545e-01 -8.03915262e-02 -6.12987161e-01 4.13456738e-01
1.62752811e-02 -3.56842488e-01 1.13073349e+00 6.39216244e-01
-5.60850620e-01 5.70793211e-01 -4.84117985e-01 -5.25861755e-02
4.53182310e-01 2.12438926e-01 -8.46186057e-02 -1.34155065e-01
-1.46097791e+00 7.79888272e-01 2.44709298e-01 -2.03994751e-01
-6.49969041e-01 -1.45260024e+00 -5.76822162e-01 4.28271562e-01
1.12915659e+00 -8.03004742e-01 8.43994558e-01 -9.41767454e-01
-8.32392335e-01 1.90809250e-01 -2.76465774e-01 -4.36523288e-01
7.36384392e-01 3.38671543e-02 -1.33694991e-01 -3.07410717e-01
4.18076962e-01 2.96346277e-01 2.99528301e-01 -1.35648108e+00
-6.48980737e-01 -4.72071975e-01 1.15666129e-01 -1.20378453e-02
3.48181836e-02 -2.78500825e-01 2.71309078e-01 -6.95611835e-01
-2.10012048e-01 -8.66559386e-01 -4.05540764e-01 -4.41656023e-01
-5.51378489e-01 -2.24925339e-01 3.68991703e-01 -4.45446998e-01
1.09554207e+00 -1.97662592e+00 -3.01315725e-01 5.09193659e-01
2.01570272e-01 -3.90855014e-01 6.68656752e-02 2.32583866e-01
-3.41290414e-01 5.44818759e-01 3.55939381e-02 5.85540712e-01
-2.24488154e-01 2.54838876e-02 -4.68309700e-01 3.05181265e-01
-8.14433172e-02 8.09830666e-01 -8.67866993e-01 -2.77152956e-01
-1.40437558e-02 7.53999874e-02 -7.35949695e-01 -2.41759509e-01
-2.33503848e-01 4.36414301e-01 -3.40903670e-01 3.56232494e-01
4.10056859e-01 -4.26097393e-01 7.68761218e-01 6.22952282e-02
-2.26756707e-01 4.85629827e-01 -1.19679320e+00 6.68800533e-01
-2.30117306e-01 5.27470112e-01 -2.65741080e-01 -1.00737488e+00
4.26720798e-01 3.29637498e-01 2.25931928e-01 -6.20510280e-01
-9.21233967e-02 1.21080153e-01 5.17255902e-01 -5.45093596e-01
-4.46684733e-02 -3.25020254e-01 1.82469800e-01 4.11299258e-01
-4.02441740e-01 2.95261174e-01 1.63137212e-01 1.36459142e-01
1.23729694e+00 -3.65744233e-01 4.12802786e-01 -4.57210124e-01
-3.10477018e-01 2.71504074e-01 9.48904812e-01 1.28037250e+00
2.76364505e-01 -1.14383353e-02 1.07653713e+00 -2.85195917e-01
-7.98419058e-01 -1.08329237e+00 -1.06588624e-01 8.36577356e-01
-1.13924168e-01 2.59425968e-01 -1.60919160e-01 -6.86037958e-01
5.52308857e-01 1.36968768e+00 -1.01148772e+00 -2.87999302e-01
-2.06301376e-01 -1.01875579e+00 3.50388348e-01 6.97389364e-01
2.75026470e-01 -7.03834653e-01 -4.76725072e-01 -8.25764388e-02
-2.13548943e-01 -7.90277362e-01 -3.50013316e-01 -8.32987577e-02
-9.98173118e-01 -1.44635701e+00 -1.95529103e-01 -9.65023711e-02
7.88746417e-01 1.23632587e-01 1.00551105e+00 6.65609306e-03
1.16426028e-01 1.49864897e-01 -1.58434972e-01 -8.19552183e-01
-2.40035623e-01 -2.18070671e-01 -1.05927773e-01 -2.81183004e-01
2.97760159e-01 -5.74304104e-01 -4.38786596e-01 1.01190716e-01
-6.13851607e-01 -2.36854881e-01 7.04028964e-01 8.10578942e-01
6.41717985e-02 2.76010245e-01 1.07687497e+00 -1.42657495e+00
8.32126498e-01 -7.27589190e-01 -8.77815545e-01 4.17090356e-01
-1.14218104e+00 2.03195065e-01 2.11754784e-01 -6.63188875e-01
-1.49835289e+00 -2.54395932e-01 7.32953846e-01 -9.03629418e-03
-2.21686531e-02 9.33044970e-01 -1.56457826e-01 3.60029280e-01
9.15401816e-01 -5.74790239e-01 -1.74070328e-01 -2.21600354e-01
3.15593004e-01 1.42677218e-01 7.09385797e-02 -7.27107406e-01
6.70845985e-01 6.44558132e-01 1.40539005e-01 -4.39198583e-01
-8.67118955e-01 1.29085198e-01 -2.92739660e-01 2.07207948e-02
6.29723310e-01 -8.65546465e-01 -8.72672260e-01 -2.25382090e-01
-8.71866107e-01 -6.34588540e-01 -2.94001669e-01 9.47129965e-01
-9.08409357e-02 -1.79987133e-01 -2.41916165e-01 -9.86495137e-01
3.34959745e-01 -7.84266949e-01 2.64600903e-01 1.87665015e-01
-5.99968493e-01 -1.11532927e+00 -3.83495539e-02 3.31619620e-01
-6.39685290e-03 3.75972480e-01 1.44235647e+00 -5.63946784e-01
-7.20895112e-01 -9.94226560e-02 -3.06817740e-01 -4.60700005e-01
3.01000893e-01 1.37881085e-01 -6.10710323e-01 1.46217883e-01
-3.82705480e-01 7.57784769e-02 7.64044046e-01 9.80238497e-01
1.07519007e+00 -6.66356564e-01 -7.07523525e-01 9.09620821e-02
1.15304959e+00 5.94515860e-01 4.58051175e-01 -7.35242153e-03
5.02648652e-01 9.56326663e-01 3.85353088e-01 5.28724492e-01
4.36851561e-01 5.07915735e-01 1.33775458e-01 -9.35146660e-02
2.54830509e-01 -6.37323260e-01 4.21900535e-03 -2.79748529e-01
-1.45852834e-01 -2.43173420e-01 -8.58424127e-01 8.29867303e-01
-1.97317731e+00 -1.17083216e+00 -5.73564470e-01 2.45106816e+00
1.03420484e+00 2.20851615e-01 1.88725486e-01 -8.66159648e-02
7.63576686e-01 -2.95375049e-01 -4.64333564e-01 -1.45822436e-01
-6.29116371e-02 -6.14686571e-02 9.28696871e-01 6.55883789e-01
-4.19664413e-01 4.73801792e-01 7.16050673e+00 2.35786721e-01
-7.63928354e-01 -1.61854345e-02 6.39368534e-01 -2.71938384e-01
-8.50621045e-01 4.21559989e-01 -4.23521340e-01 5.03893197e-01
8.09243202e-01 -5.12068450e-01 2.78551847e-01 4.71208811e-01
6.90578461e-01 -3.76361609e-01 -1.25790429e+00 7.89222941e-02
-4.80635971e-01 -1.51233518e+00 3.43819074e-02 3.93250138e-01
1.13410044e+00 -5.53043544e-01 3.21020867e-04 7.66931325e-02
1.08890116e+00 -1.09756291e+00 6.47888303e-01 5.62898576e-01
5.71542084e-01 -7.08689749e-01 8.22876811e-01 2.38224983e-01
-2.86970109e-01 -3.95046055e-01 -1.77765459e-01 -7.18139589e-01
-9.16612744e-02 1.15665972e+00 -1.05799198e+00 3.53806049e-01
4.03242797e-01 2.02230528e-01 -1.50077999e-01 9.39709067e-01
-4.66742277e-01 1.10628986e+00 -1.24590255e-01 1.39254183e-01
-1.79485559e-01 2.10014544e-02 2.74774313e-01 6.79340541e-01
1.80394918e-01 3.72035414e-01 -9.19198468e-02 1.03642154e+00
-2.94985592e-01 -2.02633142e-01 -8.02542090e-01 -2.72726119e-01
9.67390835e-01 4.58234817e-01 -6.68223441e-01 -2.35075936e-01
-4.59223807e-01 8.76828730e-02 1.08747534e-01 7.25184023e-01
-6.98049784e-01 8.30458403e-02 4.50369924e-01 5.10981977e-01
-9.89119112e-02 2.08507672e-01 -9.27305818e-01 -8.44035745e-01
-2.62158543e-01 -7.37152398e-01 6.95033491e-01 -5.85055947e-01
-1.24432719e+00 -6.97218478e-01 4.55817401e-01 -2.32784301e-01
-9.03534442e-02 -1.44276440e-01 -6.48021877e-01 9.46113050e-01
-1.19696319e+00 -4.63398039e-01 -1.70733444e-02 2.07538992e-01
9.28306952e-02 3.66671652e-01 5.23822248e-01 1.34292141e-01
-4.62333679e-01 4.26298440e-01 -2.19510853e-01 2.19865073e-03
8.25083733e-01 -1.16340816e+00 -1.01417489e-01 6.95354283e-01
-1.49440214e-01 8.80174696e-01 1.00558043e+00 -1.35167158e+00
-8.96992683e-01 -8.48931491e-01 9.08782244e-01 -5.68871319e-01
6.53727651e-01 -5.62215894e-02 -8.00177693e-01 1.13898861e+00
-1.17896914e-01 -4.50844854e-01 6.14471078e-01 8.97146761e-01
-3.54419768e-01 5.37539683e-02 -1.12444723e+00 8.31907630e-01
1.19281018e+00 -2.43567061e-02 -5.52996576e-01 1.91725641e-01
9.34598625e-01 -3.86712560e-03 -5.78040779e-01 4.46101993e-01
6.81297362e-01 -7.05908060e-01 7.28097796e-01 -9.56812501e-01
7.87809670e-01 -6.28292337e-02 8.49467367e-02 -1.41277850e+00
-5.39345562e-01 -2.76178092e-01 3.07307512e-01 1.07604444e+00
1.01206732e+00 -7.38950670e-01 6.59667492e-01 9.91094291e-01
2.56300628e-01 -4.00056541e-01 -6.75389647e-01 -6.08673871e-01
1.30233973e-01 -1.84267744e-01 5.68236828e-01 1.53158879e+00
-1.03181610e-02 3.10526788e-01 -2.58079767e-01 3.60629201e-01
7.20322549e-01 2.92916864e-01 5.63653588e-01 -1.53266251e+00
-3.51430923e-01 -5.97971119e-02 1.49461031e-01 -2.90554732e-01
2.10737228e-01 -4.86886472e-01 -2.78991461e-01 -1.52010822e+00
4.95303482e-01 -6.08213961e-01 -8.33131298e-02 5.90837538e-01
-6.71374559e-01 -5.26862204e-01 -2.12285686e-02 -1.15655050e-01
1.16417214e-01 1.49602711e-01 1.03325129e+00 7.56732235e-03
-4.98662710e-01 6.84924498e-02 -1.16095138e+00 7.43355393e-01
6.56321228e-01 -6.66009367e-01 -7.26130605e-01 -1.77944705e-01
7.22457588e-01 4.56934899e-01 7.36351073e-01 -5.06432950e-02
-6.47220761e-02 -8.15392971e-01 4.61958081e-01 -6.63613155e-02
-3.35062057e-01 -7.34127462e-01 4.68073189e-01 6.36436462e-01
-7.16626942e-01 -1.08018421e-01 2.56901383e-01 7.46654689e-01
2.50159085e-01 5.68672866e-02 2.64745563e-01 -1.47913739e-01
-2.08601594e-01 -3.87710840e-01 -2.90287256e-01 2.03271508e-01
8.83076191e-01 1.28413066e-01 -5.88721395e-01 -6.11063302e-01
-5.23057222e-01 2.70693064e-01 3.20840657e-01 2.81377792e-01
1.47367030e-01 -1.03450549e+00 -7.80068398e-01 -4.18487310e-01
-1.85594216e-01 -4.52317059e-01 2.50841498e-01 9.62316871e-01
2.78373539e-01 5.32150209e-01 2.39902199e-03 8.99033919e-02
-9.90461588e-01 5.96732795e-01 2.38979906e-01 -2.47718289e-01
-2.11196333e-01 6.71019077e-01 6.00459576e-01 -1.18440904e-01
-4.66733286e-03 -1.69159532e-01 1.01816528e-01 1.54179484e-01
2.25415453e-01 9.09318984e-01 -3.28418732e-01 2.95779377e-01
-2.89730668e-01 -1.00686066e-01 -2.11762451e-02 -2.45142832e-01
1.27379346e+00 -7.67279565e-02 6.80126548e-02 4.89529908e-01
3.28284979e-01 4.43809032e-01 -1.25334871e+00 1.12554461e-01
8.31159577e-02 -6.09012246e-01 1.01414725e-01 -1.26806152e+00
-7.25567877e-01 2.41193771e-01 2.26401076e-01 2.91248977e-01
7.06974626e-01 -8.40440169e-02 -5.88592775e-02 5.09145632e-02
2.84837902e-01 -8.18064749e-01 -2.33218282e-01 -1.80790260e-01
6.18201435e-01 -1.18273616e+00 2.37625539e-01 -7.51781821e-01
-4.57733333e-01 4.49914843e-01 5.92148364e-01 1.69871643e-01
4.76617098e-01 1.28098264e-01 -2.48721421e-01 -3.30415905e-01
-1.00574684e+00 -1.34354429e-02 5.31929918e-02 5.49658298e-01
5.89641511e-01 3.99003953e-01 -7.84742117e-01 6.50071979e-01
-1.87464312e-01 2.01858655e-01 8.64628375e-01 4.72799987e-01
-2.87814401e-02 -6.96664453e-01 -6.35159910e-01 9.48947370e-01
-6.39454424e-01 -2.14074582e-01 -5.87706983e-01 1.15150869e+00
6.63656294e-02 1.21664846e+00 2.23091751e-01 2.57835202e-02
4.38548893e-01 8.10436532e-02 3.39712590e-01 -5.87878764e-01
-6.66301325e-02 -1.05210342e-01 3.29492956e-01 -4.41125929e-01
-3.32129657e-01 -8.95838976e-01 -1.15280616e+00 -5.50682545e-01
-5.26070356e-01 2.40420967e-01 1.31859854e-01 1.12211502e+00
5.87889254e-01 7.69869328e-01 3.44326496e-01 7.76369125e-02
-6.55357420e-01 -9.24836516e-01 -5.74705482e-01 3.17384005e-01
2.38763630e-01 -1.07214105e+00 -6.30639911e-01 1.38003181e-03] | [8.170515060424805, 5.371395587921143] |
6c2afada-7cb1-495d-88ac-992dba6a8ce1 | incongruity-detection-between-bangla-news | 2211.07709 | null | https://arxiv.org/abs/2211.07709v1 | https://arxiv.org/pdf/2211.07709v1.pdf | Incongruity Detection between Bangla News Headline and Body Content through Graph Neural Network | Incongruity between news headlines and the body content is a common method of deception used to attract readers. Profitable headlines pique readers' interest and encourage them to visit a specific website. This is usually done by adding an element of dishonesty, using enticements that do not precisely reflect the content being delivered. As a result, automatic detection of incongruent news between headline and body content using language analysis has gained the research community's attention. However, various solutions are primarily being developed for English to address this problem, leaving low-resource languages out of the picture. Bangla is ranked 7th among the top 100 most widely spoken languages, which motivates us to pay special attention to the Bangla language. Furthermore, Bangla has a more complex syntactic structure and fewer natural language processing resources, so it becomes challenging to perform NLP tasks like incongruity detection and stance detection. To tackle this problem, for the Bangla language, we offer a graph-based hierarchical dual encoder (BGHDE) model that learns the content similarity and contradiction between Bangla news headlines and content paragraphs effectively. The experimental results show that the proposed Bangla graph-based neural network model achieves above 90% accuracy on various Bangla news datasets. | ['Ryan Mohammad Bin Shahjahan', 'MD Abdullah Al Nasim', 'Kawsarul Islam', 'Akib Khan', 'Md Aminul Haque Palash'] | 2022-10-26 | null | null | null | null | ['incongruity-detection', 'stance-detection'] | ['natural-language-processing', 'natural-language-processing'] | [-4.37779576e-01 3.20769623e-02 -4.44860995e-01 -2.86274731e-01
-4.78317499e-01 -4.39705223e-01 5.96927643e-01 6.48304045e-01
-4.69927758e-01 5.21962345e-01 8.22585285e-01 -3.45011562e-01
2.75707722e-01 -8.09019446e-01 -5.46581268e-01 -3.97936493e-01
4.27742213e-01 2.49887377e-01 1.93180770e-01 -9.56997633e-01
5.59510767e-01 1.84671208e-01 -9.30109382e-01 5.42404234e-01
9.85706747e-01 6.80954516e-01 -1.54315710e-01 1.47543043e-01
-3.97645563e-01 1.22069001e+00 -7.90902197e-01 -1.23069155e+00
-1.70128584e-01 -7.39272475e-01 -9.67799544e-01 1.06510237e-01
4.40767646e-01 -4.80782509e-01 -6.23223424e-01 1.39051640e+00
5.11226416e-01 -3.93508911e-01 4.80593234e-01 -9.71985161e-01
-8.49899888e-01 9.67397392e-01 -9.11197126e-01 6.17715657e-01
3.51308346e-01 -3.68002772e-01 1.38986838e+00 -5.30793309e-01
3.64099205e-01 1.21137834e+00 4.61973727e-01 3.88481617e-01
-5.66095710e-01 -5.21953404e-01 2.12646365e-01 4.34729308e-01
-1.26735640e+00 -2.26643518e-01 1.23058629e+00 -2.49209210e-01
7.70169497e-01 3.97408932e-01 6.94730520e-01 1.43380785e+00
3.82703006e-01 1.06185710e+00 1.08616424e+00 -2.43914828e-01
-1.55105069e-01 3.82276773e-01 2.91439742e-01 6.29902303e-01
5.70974052e-01 -7.37670183e-01 -5.69866657e-01 3.41335908e-02
5.01563586e-02 -3.52441818e-01 -5.42471170e-01 3.96351218e-01
-8.57159793e-01 1.30952895e+00 3.62824440e-01 5.42048335e-01
-2.65182883e-01 -3.86034727e-01 9.37316179e-01 3.31122071e-01
6.32177830e-01 2.72405416e-01 1.26250535e-01 -5.20504005e-02
-6.93197191e-01 4.09477770e-01 9.33391690e-01 7.77581573e-01
5.65332286e-02 8.16984028e-02 1.94015786e-01 1.09674072e+00
3.65861714e-01 4.43201035e-01 7.47586370e-01 -1.61308974e-01
7.67765760e-01 7.27935314e-01 -2.13912666e-01 -2.14085746e+00
-2.86525697e-01 -6.62635684e-01 -1.04248619e+00 -3.84583116e-01
4.04187858e-01 3.74945477e-02 -1.70474723e-01 1.42743838e+00
1.81473821e-01 -7.85221696e-01 1.66186079e-01 1.22120190e+00
1.08706975e+00 1.08750188e+00 -3.06796312e-01 -3.88787240e-01
1.59386528e+00 -7.91351914e-01 -1.03521013e+00 -3.45368028e-01
7.17432380e-01 -1.18225658e+00 1.06101048e+00 4.84121174e-01
-1.10281539e+00 -1.86357096e-01 -1.30057037e+00 -3.58027101e-01
-4.11567539e-01 -6.60187230e-02 2.21290365e-01 4.09818798e-01
-5.03102481e-01 1.55012369e-01 -2.68822223e-01 -9.69375148e-02
4.26446050e-01 -1.70017824e-01 -2.64582664e-01 -1.45757377e-01
-1.76826870e+00 1.10045147e+00 5.40112257e-01 3.36750537e-01
-2.23283619e-01 -8.21376368e-02 -9.34429288e-01 -2.91855037e-01
3.87909412e-01 -1.84418961e-01 9.88945246e-01 -1.09747684e+00
-1.34087825e+00 1.09833312e+00 1.75113723e-01 -4.01120633e-01
5.93451381e-01 -3.01639080e-01 -8.42757940e-01 2.69033393e-04
2.66382754e-01 -3.23928893e-02 9.51848626e-01 -9.15046871e-01
-5.81469238e-01 -3.71594310e-01 1.35424510e-01 3.38574290e-01
-5.75913846e-01 4.51158315e-01 -2.80550987e-01 -9.06145215e-01
2.30690166e-01 -5.54063976e-01 4.58502203e-01 -5.04637063e-01
-8.73396695e-01 -5.41172028e-01 8.14051807e-01 -1.25106502e+00
1.64253855e+00 -1.98348105e+00 7.73969814e-02 4.49419022e-02
6.27684712e-01 1.15974113e-01 3.06384742e-01 6.42747819e-01
2.31258512e-01 1.89593360e-01 1.83008507e-01 1.46364242e-01
-5.87132126e-02 1.12366661e-01 -2.66623527e-01 7.42743611e-01
-5.66748455e-02 6.29388809e-01 -8.01861405e-01 -7.64070153e-01
-4.26956713e-01 1.98121876e-01 -5.14581740e-01 7.80465007e-02
8.10558349e-03 1.78781971e-01 -5.15579462e-01 4.89826232e-01
6.21923864e-01 -2.19655022e-01 2.02428952e-01 -4.55750763e-01
-2.91798443e-01 8.12025428e-01 -7.25251377e-01 9.48854327e-01
-2.54319936e-01 8.12111676e-01 -1.48178503e-01 -9.50016320e-01
9.07584488e-01 1.99051529e-01 5.53139448e-02 -1.09360564e+00
5.44044614e-01 3.00355017e-01 3.17650825e-01 -8.27166378e-01
6.86161637e-01 -2.37786859e-01 -5.44673562e-01 3.29990536e-01
-4.84002113e-01 -5.26810512e-02 1.88477010e-01 7.18439102e-01
7.02935219e-01 -3.83427054e-01 4.90062773e-01 -2.17219844e-01
7.51644492e-01 1.25994816e-01 5.01314044e-01 3.34643453e-01
-2.83199042e-01 2.45882869e-01 9.02091563e-01 -4.43858624e-01
-1.02670479e+00 -4.94245589e-01 4.91220243e-02 1.02650559e+00
2.56951839e-01 -5.45813084e-01 -7.26851225e-01 -8.08553874e-01
-1.37511790e-01 8.71253729e-01 -5.38189709e-01 -2.52288461e-01
-8.98534834e-01 -6.13579512e-01 6.08877897e-01 2.15466749e-02
8.94946396e-01 -9.20134425e-01 -2.67610610e-01 2.28165925e-01
-8.91772926e-01 -1.34016621e+00 -6.20153189e-01 -1.57502949e-01
-2.88005024e-01 -8.35893691e-01 -5.49699664e-01 -1.03465939e+00
3.71842980e-01 3.34846020e-01 8.53962064e-01 -5.71079552e-02
2.77084231e-01 -2.75313437e-01 -7.13130593e-01 -4.64805275e-01
-8.71575415e-01 -1.15878033e-02 2.06048694e-02 2.59114444e-01
4.72734243e-01 -1.10512525e-01 -4.48547006e-01 6.15388155e-02
-8.59377742e-01 3.73125315e-01 3.96298081e-01 7.47966111e-01
-7.70877153e-02 2.06319019e-01 6.62256539e-01 -9.68965173e-01
1.09553754e+00 -5.84871829e-01 -3.75624657e-01 -1.30061358e-02
-3.00076157e-01 -3.09122473e-01 1.00065875e+00 -1.64115906e-01
-1.00170255e+00 -7.94186473e-01 -3.87595326e-01 2.58814424e-01
1.74641609e-01 1.00495446e+00 -2.98731178e-01 3.41243535e-01
4.83866841e-01 3.28055590e-01 -8.12760741e-02 -1.92205146e-01
1.32421240e-01 1.22486925e+00 4.85493392e-01 5.26157022e-02
4.10804838e-01 2.19362974e-01 -3.85211229e-01 -1.16775954e+00
-1.26323128e+00 -5.55408299e-01 -2.92652220e-01 -5.86812675e-01
7.91850865e-01 -6.84270918e-01 -6.82221293e-01 6.14986777e-01
-1.50825357e+00 3.84727627e-01 3.94373745e-01 4.69359040e-01
-5.00648730e-02 7.67064750e-01 -9.19729888e-01 -3.98918420e-01
-3.70907575e-01 -9.75295901e-01 4.29593891e-01 -1.96145743e-01
-4.38730508e-01 -9.21581089e-01 -5.28729893e-02 8.85780394e-01
2.20436126e-01 2.02276647e-01 1.24256277e+00 -8.84927750e-01
6.88185692e-02 -5.42558968e-01 -2.95825779e-01 4.58016634e-01
2.41125733e-01 -1.62659153e-01 -4.85692948e-01 -2.72301108e-01
3.72007102e-01 -6.58125460e-01 4.52082008e-01 8.81362613e-03
8.14024568e-01 -1.16252387e+00 1.56776249e-01 2.72967806e-03
1.08901870e+00 -9.29751173e-02 5.76388896e-01 5.60335755e-01
9.82037246e-01 7.81496525e-01 4.46209371e-01 6.33591115e-01
6.70019448e-01 5.80027521e-01 5.21760106e-01 2.44172096e-01
-1.57640502e-01 -4.26847816e-01 5.57455242e-01 1.81351554e+00
2.41570145e-01 -7.34354198e-01 -8.41745198e-01 4.91584599e-01
-1.54946601e+00 -9.91505980e-01 -8.01533043e-01 1.80152333e+00
1.10257745e+00 4.99161005e-01 2.68670201e-01 4.76188153e-01
7.92930722e-01 5.24383247e-01 1.26608070e-02 -7.70785451e-01
-3.90196562e-01 -6.05398297e-01 3.09091419e-01 3.43079627e-01
-1.00151503e+00 8.64195108e-01 4.70087719e+00 9.49859321e-01
-1.38673949e+00 3.26601714e-02 5.51693320e-01 3.71998340e-01
-4.13208365e-01 -4.27877814e-01 -8.52616012e-01 8.28026831e-01
7.84778476e-01 -1.94851965e-01 -3.88677195e-02 6.57833993e-01
5.37638247e-01 9.44585428e-02 -6.59579396e-01 1.03428173e+00
5.88578403e-01 -1.19502199e+00 2.82625020e-01 -3.22716348e-02
4.11355674e-01 -1.31583035e-01 1.42133296e-01 2.50757188e-01
7.52888620e-03 -7.76727200e-01 8.49911153e-01 -2.00091258e-01
3.24782103e-01 -9.11630094e-01 1.03229618e+00 5.97092032e-01
-4.58368450e-01 2.40468234e-01 -4.09012139e-01 -2.00023726e-01
2.39604056e-01 6.33261621e-01 -8.04365814e-01 2.10870221e-01
6.17054164e-01 6.41632915e-01 -5.44803977e-01 7.19226003e-01
-3.70447189e-01 8.74634981e-01 -4.73489845e-03 -6.62130892e-01
6.20130837e-01 -1.54789940e-01 6.98917627e-01 1.46556377e+00
-3.85025367e-02 6.47292063e-02 3.41856927e-02 3.92494142e-01
-4.35764819e-01 7.01530218e-01 -5.90471447e-01 -3.24922770e-01
1.02150422e-02 1.04144859e+00 -6.76746547e-01 -1.77002817e-01
-7.89989889e-01 1.17798698e+00 5.76906502e-01 -6.84586912e-02
-9.97441053e-01 -3.73651952e-01 4.93666809e-03 2.89349645e-01
-1.64618960e-03 -9.91237015e-02 -1.33222327e-01 -1.15858853e+00
1.61062423e-02 -1.45940208e+00 5.57599068e-01 -5.08745492e-01
-1.52725804e+00 5.96139014e-01 -2.53755271e-01 -1.18137622e+00
1.12040982e-01 -3.80650431e-01 -3.55989218e-01 5.45347691e-01
-1.49387741e+00 -1.12266517e+00 -1.13949180e-01 5.06295979e-01
8.21682334e-01 1.28281554e-02 4.05751497e-01 4.61059093e-01
-7.71892488e-01 7.11399674e-01 -8.84058326e-02 4.59411979e-01
6.23978913e-01 -1.04309583e+00 1.69910476e-01 8.41918945e-01
-3.17685232e-02 4.26218718e-01 1.11904800e+00 -7.29005158e-01
-1.24612844e+00 -7.97742903e-01 1.58759570e+00 -2.62407046e-02
1.03239334e+00 -1.25841811e-01 -1.15125954e+00 4.75160122e-01
6.15576267e-01 -3.78518432e-01 7.48908281e-01 -1.33262753e-01
-5.18555701e-01 -6.14143126e-02 -7.62261868e-01 8.13692391e-01
5.99723101e-01 -4.29987997e-01 -9.48271871e-01 5.66402316e-01
3.75998765e-01 -3.53097975e-01 -3.86172593e-01 2.13597007e-02
2.61703819e-01 -8.97522688e-01 3.14270139e-01 -6.85831606e-01
1.01037526e+00 -1.09379090e-01 -1.95897482e-02 -1.36416638e+00
-3.74295324e-01 -3.83040607e-01 1.90761238e-02 1.17226362e+00
3.51863176e-01 -4.98601884e-01 4.94988769e-01 1.32016852e-01
-1.53149888e-01 -6.24892056e-01 -6.70946240e-01 -4.94425535e-01
1.09251045e-01 -2.42930889e-01 -7.65900686e-03 1.29980314e+00
4.86508727e-01 9.69750106e-01 -7.69300342e-01 1.90560240e-02
2.93299168e-01 2.49955222e-01 5.52647471e-01 -8.07339907e-01
-1.08574674e-01 -6.22403443e-01 -4.60988581e-01 -1.07926834e+00
1.05243452e-01 -9.89798248e-01 -2.02979892e-01 -1.45332837e+00
2.66654819e-01 1.11706749e-01 2.24445045e-01 1.13963597e-01
5.59227765e-02 3.40534657e-01 -1.26823217e-01 3.10574085e-01
-7.35703468e-01 4.73715812e-01 1.43091547e+00 -3.34531575e-01
5.67282215e-02 -9.65568982e-03 -1.04644501e+00 9.53651726e-01
9.94710147e-01 -4.67453301e-01 -2.09811762e-01 -3.90883952e-01
8.64149213e-01 8.64075422e-02 2.11376082e-02 -3.09184104e-01
2.73452342e-01 -1.78786665e-01 1.11756161e-01 -6.56440496e-01
-1.82986036e-02 -5.88893771e-01 -5.21476507e-01 5.46077967e-01
-4.58858222e-01 4.45317417e-01 -1.46796752e-03 3.73266071e-01
-5.41458189e-01 -3.06085080e-01 8.90210927e-01 -1.52511925e-01
-4.26966280e-01 -1.29444107e-01 -7.69076645e-01 5.25687814e-01
7.72387922e-01 -5.02220169e-02 -5.01280844e-01 -8.35228562e-01
-1.60600126e-01 1.97230037e-02 -7.95452110e-03 5.43704271e-01
9.12562132e-01 -1.16410995e+00 -1.24868739e+00 -8.94092172e-02
2.66200572e-01 -3.63583863e-01 2.91897655e-01 1.00759828e+00
-9.37557459e-01 3.70070755e-01 -1.95059180e-01 -7.23439753e-02
-1.22275984e+00 6.43443465e-01 1.07542640e-02 -8.38256404e-02
-7.14213312e-01 7.23831117e-01 -2.79595982e-03 2.72966418e-02
4.37113903e-02 7.51277283e-02 -7.60955691e-01 4.51319188e-01
5.70478857e-01 4.33249414e-01 1.88477606e-01 -1.33330858e+00
-2.79382646e-01 -1.40238742e-04 -6.13726199e-01 1.34333163e-01
9.34655845e-01 -4.14570361e-01 -4.53930676e-01 4.46574569e-01
1.54968452e+00 5.76312244e-01 -1.79586500e-01 -1.33966506e-01
-7.80947134e-02 -5.06418407e-01 4.21494901e-01 -7.42671967e-01
-8.10899496e-01 7.17815220e-01 -1.93966806e-01 7.23562181e-01
7.35730231e-01 6.87832525e-03 1.25457835e+00 2.17543021e-01
-5.53970505e-03 -1.38909578e+00 3.68018031e-01 7.04824626e-01
1.40535378e+00 -1.29102802e+00 1.40182257e-01 -4.66410488e-01
-8.46404254e-01 1.20480359e+00 7.27554500e-01 9.55711082e-02
2.77413279e-01 -4.62091118e-02 4.10871536e-01 -4.63374555e-01
-2.40180522e-01 1.10713288e-01 2.84652174e-01 1.33355916e-01
7.01061130e-01 -1.32084236e-01 -1.03204274e+00 6.65784180e-01
-7.09857285e-01 -7.21733809e-01 8.85565639e-01 6.31991982e-01
-4.71115649e-01 -6.10213161e-01 -2.89659679e-01 4.18604285e-01
-1.12720370e+00 -1.06699258e-01 -5.89526832e-01 8.34225476e-01
-2.70158380e-01 1.02919412e+00 -1.48808286e-01 -2.69833982e-01
2.69682527e-01 -2.66960531e-01 1.93016827e-01 -1.52312532e-01
-4.49394614e-01 1.26758799e-01 4.05221224e-01 -1.47106782e-01
-2.72053689e-01 -4.59145814e-01 -1.08714950e+00 -8.25834751e-01
-5.38887918e-01 3.07572842e-01 6.08807862e-01 9.50345755e-01
-7.87587315e-02 3.51147652e-01 8.25172067e-01 -1.86695848e-02
-6.07676566e-01 -9.42528248e-01 -5.41418791e-01 6.06993139e-01
2.29193389e-01 -2.63969392e-01 -4.89281327e-01 -2.96380490e-01] | [8.754631042480469, 10.511886596679688] |
efb1749a-507a-4451-89c0-e7abd1117b66 | moocradar-a-fine-grained-and-multi-aspect | 2304.02205 | null | https://arxiv.org/abs/2304.02205v1 | https://arxiv.org/pdf/2304.02205v1.pdf | MoocRadar: A Fine-grained and Multi-aspect Knowledge Repository for Improving Cognitive Student Modeling in MOOCs | Student modeling, the task of inferring a student's learning characteristics through their interactions with coursework, is a fundamental issue in intelligent education. Although the recent attempts from knowledge tracing and cognitive diagnosis propose several promising directions for improving the usability and effectiveness of current models, the existing public datasets are still insufficient to meet the need for these potential solutions due to their ignorance of complete exercising contexts, fine-grained concepts, and cognitive labels. In this paper, we present MoocRadar, a fine-grained, multi-aspect knowledge repository consisting of 2,513 exercise questions, 5,600 knowledge concepts, and over 12 million behavioral records. Specifically, we propose a framework to guarantee a high-quality and comprehensive annotation of fine-grained concepts and cognitive labels. The statistical and experimental results indicate that our dataset provides the basis for the future improvements of existing methods. Moreover, to support the convenient usage for researchers, we release a set of tools for data querying, model adaption, and even the extension of our repository, which are now available at https://github.com/THU-KEG/MOOC-Radar. | ['Jie Tang', 'Juanzi Li', 'Hai-Tao Zheng', 'Lei Hou', 'Manli Li', 'Xiaoya Li', 'Zhengshan Liao', 'Shangqing Tu', 'Zijun Yao', 'Qingyang Zhong', 'Mengying Lu', 'Jifan Yu'] | 2023-04-05 | null | null | null | null | ['knowledge-tracing'] | ['miscellaneous'] | [-3.21153343e-01 -1.32200181e-01 -4.54915464e-01 -4.36651677e-01
-4.33206409e-01 -7.20314264e-01 2.20964804e-01 5.15912652e-01
-8.36060289e-03 7.75769353e-01 1.45578161e-01 -5.34335315e-01
-8.03897262e-01 -1.13231313e+00 -5.12773454e-01 -2.29229018e-01
5.54600477e-01 4.38413918e-01 5.15224338e-01 -2.30781078e-01
3.82673085e-01 4.06262279e-01 -2.26401162e+00 2.84942001e-01
1.29302645e+00 7.60217667e-01 4.44236398e-02 4.42748338e-01
-3.85262966e-01 9.58864450e-01 -4.29169148e-01 -7.22133994e-01
-5.10100365e-01 -2.51515299e-01 -1.07782185e+00 -1.22376010e-01
6.99151933e-01 -7.23143443e-02 -1.84027046e-01 1.10133278e+00
2.76221246e-01 4.02411610e-01 4.07887727e-01 -1.21763027e+00
-1.02513742e+00 6.38607442e-01 4.43977788e-02 7.44506866e-02
5.55371642e-01 -1.87939987e-01 8.25003266e-01 -5.43581963e-01
6.34324908e-01 9.07308459e-01 6.08395100e-01 7.64469564e-01
-6.85737193e-01 -7.21014380e-01 2.07296088e-01 6.49263442e-01
-1.45190299e+00 -2.84297734e-01 6.42045915e-01 -8.38320076e-01
3.33544821e-01 6.22177839e-01 7.44950593e-01 1.10846150e+00
-3.25138211e-01 5.56270838e-01 1.07611752e+00 -6.69281840e-01
9.07968804e-02 5.85093081e-01 9.24784482e-01 1.12874591e+00
4.50882405e-01 -1.56399369e-01 -8.10522795e-01 -2.78882235e-01
4.54847634e-01 1.68138638e-01 -1.43022090e-01 -3.83660823e-01
-8.39961886e-01 4.37621176e-01 -1.55578136e-01 4.63081747e-01
1.88968554e-02 -2.83502936e-01 -8.58304203e-02 2.22811073e-01
2.37344190e-01 4.07864034e-01 -5.32557309e-01 -6.19334102e-01
-5.50351799e-01 5.20330906e-01 8.55736613e-01 1.21708071e+00
7.09326208e-01 -4.03559119e-01 -1.55026510e-01 8.15308273e-01
3.40282977e-01 1.42807037e-01 7.44076490e-01 -1.14251435e+00
1.63668811e-01 1.34262228e+00 -8.01556036e-02 -8.57815146e-01
-3.45378697e-01 -4.69892561e-01 -2.04645276e-01 -2.72849947e-01
5.66611290e-01 1.25195175e-01 -3.98432553e-01 1.49643624e+00
5.35185814e-01 3.93023849e-01 -1.48196533e-01 5.23376942e-01
1.30710912e+00 2.99656484e-02 2.51195163e-01 1.50186837e-01
1.85585284e+00 -7.65398800e-01 -9.30729449e-01 8.71368945e-02
7.49658644e-01 -7.40373552e-01 1.24615407e+00 4.85456049e-01
-1.18337464e+00 -7.59065092e-01 -6.65251672e-01 -1.52942032e-01
-8.54072809e-01 -1.04214177e-02 7.28415072e-01 1.04894125e+00
-5.90604246e-01 2.11062968e-01 -5.84073961e-01 -2.73888141e-01
4.01695490e-01 -2.22391278e-01 -9.53579769e-02 -2.72100508e-01
-1.11986601e+00 7.20376968e-01 2.43380368e-01 -2.97296643e-01
-5.51288545e-01 -1.09261000e+00 -5.66105664e-01 3.22145343e-01
5.12674749e-01 -6.10660791e-01 1.15226030e+00 -1.58196464e-01
-1.34833682e+00 7.81173110e-01 -6.65649250e-02 1.98496252e-01
2.77227163e-01 -2.37362951e-01 -5.10370791e-01 -1.14779584e-01
-7.27700591e-02 6.03247900e-03 -1.50722325e-01 -9.90265548e-01
-8.03103685e-01 -6.12989545e-01 1.71172947e-01 1.50043726e-01
-9.13060069e-01 -1.04599416e-01 -7.21530914e-01 -3.27351332e-01
6.33583218e-02 -6.47060752e-01 -3.98980500e-03 -4.46892321e-01
-1.77794695e-01 -7.53763855e-01 2.57793307e-01 -5.54223001e-01
1.88674343e+00 -1.94532037e+00 -1.61007449e-01 1.13036737e-01
3.10874701e-01 2.86450416e-01 1.98519871e-01 2.81141102e-01
1.00381084e-01 2.19249383e-01 1.18482761e-01 -1.46043450e-02
2.91864336e-01 -3.12907882e-02 -9.67994556e-02 -7.47609977e-03
-3.60807717e-01 6.88359678e-01 -9.51586545e-01 -6.28262997e-01
2.83595413e-01 2.20255136e-01 -5.76091409e-01 3.60408753e-01
-2.08437696e-01 2.63291508e-01 -8.94590616e-01 8.13644052e-01
2.96737701e-01 -1.94509089e-01 2.28872627e-01 3.18551540e-01
-2.09781602e-01 4.04660553e-01 -1.32880437e+00 1.68511271e+00
-1.88859612e-01 3.82036477e-01 -3.18262607e-01 -7.82985628e-01
1.10863388e+00 3.64065707e-01 5.02395868e-01 -6.35737538e-01
-9.44478437e-02 1.07780553e-01 -1.82790503e-01 -8.90049994e-01
6.13422334e-01 1.59237400e-01 -1.02793522e-01 4.41271245e-01
7.00763240e-02 2.10188553e-01 4.05862540e-01 5.63531779e-02
1.08391643e+00 2.27640569e-01 2.28295416e-01 -2.44419470e-01
5.91741562e-01 1.58671886e-01 6.05189204e-01 6.60970032e-01
-1.89740762e-01 1.49584427e-01 2.43646562e-01 -3.19000900e-01
-3.68361503e-01 -8.84021461e-01 -3.50094944e-01 1.56058836e+00
-1.50187463e-01 -7.12140441e-01 -8.94739807e-01 -7.01518774e-01
1.12711370e-01 8.89132798e-01 -3.41227353e-01 -2.22775877e-01
-3.19826096e-01 -4.68615949e-01 7.57215619e-01 3.80482346e-01
3.88699174e-01 -9.00536716e-01 -2.68173665e-01 8.42992738e-02
-6.61524773e-01 -9.09394979e-01 -9.29046795e-02 -2.10916743e-01
-8.50954592e-01 -1.54686010e+00 -5.11999577e-02 -7.65847087e-01
5.79623222e-01 2.36224532e-01 1.42673695e+00 6.28804564e-01
-2.40870059e-01 6.33900583e-01 -3.69521439e-01 -5.87700486e-01
-2.43658200e-01 1.69130445e-01 -5.61449863e-02 -3.90841663e-01
1.01576591e+00 -5.66190541e-01 -2.65642047e-01 3.01570445e-01
-7.89620697e-01 -1.17097668e-01 1.84207678e-01 3.04781437e-01
4.61375266e-01 3.37631166e-01 7.53352582e-01 -1.27976286e+00
7.53342032e-01 -6.01613760e-01 -6.10026360e-01 5.76150656e-01
-1.14159465e+00 -2.35524416e-01 1.79152802e-01 -2.52010167e-01
-1.38071835e+00 -2.87344635e-01 -3.35363537e-01 -6.84417188e-02
-8.15082669e-01 6.66197717e-01 -3.44074756e-01 7.83490837e-02
7.19122410e-01 2.66448230e-01 -5.11257052e-01 -9.53018546e-01
3.10442716e-01 8.44510496e-01 4.34995294e-01 -1.08846438e+00
5.42405546e-01 -2.64429338e-02 -2.98237443e-01 -5.54259419e-01
-1.19087887e+00 -5.93267143e-01 -6.01778209e-01 -4.12740648e-01
6.34308934e-01 -9.76855934e-01 -1.31493342e+00 2.00587213e-01
-5.49398839e-01 -2.07039729e-01 -2.19882488e-01 4.73077178e-01
-2.67389834e-01 3.36004972e-01 -5.56544125e-01 -7.01697648e-01
9.70064253e-02 -9.49436605e-01 2.67267913e-01 6.35182440e-01
-2.63632923e-01 -1.21512854e+00 2.02204451e-01 1.43067920e+00
3.38387340e-01 -2.66505688e-01 1.37190580e+00 -8.16571057e-01
-5.99689186e-01 -9.74882320e-02 2.27411106e-01 2.04770099e-02
-7.28776753e-02 2.11608738e-01 -9.95098889e-01 1.79114908e-01
-2.09002778e-01 -5.10504723e-01 4.88346338e-01 -6.61927611e-02
1.71688068e+00 -6.17358349e-02 -1.77977681e-01 2.79880494e-01
1.09462917e+00 2.15815917e-01 4.00653064e-01 5.27139008e-01
7.53036141e-01 7.26208508e-01 6.42771721e-01 4.42800552e-01
1.02913809e+00 5.53176939e-01 2.18677908e-01 4.02027398e-01
-2.59278923e-01 -4.81527001e-01 4.73089516e-02 1.25483727e+00
-3.64426553e-01 -6.05204701e-02 -1.17048025e+00 7.13899016e-01
-1.74787569e+00 -9.24908578e-01 -3.83530051e-01 2.08869648e+00
1.19862819e+00 -9.27894861e-02 -1.02632999e-01 -5.17426655e-02
4.52678233e-01 -1.84307918e-01 -2.84649372e-01 -2.01728031e-01
3.93552601e-01 1.58659920e-01 2.89294999e-02 4.48250979e-01
-6.10746741e-01 7.85597146e-01 6.04694748e+00 9.10903454e-01
-3.01733315e-01 2.48054177e-01 1.97490156e-01 1.45806242e-02
-6.07757211e-01 -1.38897955e-01 -1.34718525e+00 4.00142401e-01
1.26945448e+00 -2.13312611e-01 3.78588766e-01 8.96755457e-01
2.61194352e-02 1.80743895e-02 -8.87303531e-01 5.84280133e-01
2.83259768e-02 -1.45346141e+00 -7.92245939e-02 7.44434148e-02
8.60010564e-01 -5.17820776e-01 -4.48080935e-02 7.89446831e-01
4.93574500e-01 -8.57945025e-01 5.08340716e-01 1.05672872e+00
3.57517987e-01 -7.68753946e-01 4.74817634e-01 7.59178400e-01
-9.54113543e-01 -5.78906350e-02 -2.67053276e-01 -2.98272431e-01
-5.82846463e-01 5.80062926e-01 -4.85319495e-01 5.71030378e-01
8.56688023e-01 5.26237547e-01 -9.34751570e-01 8.83645296e-01
-1.99223727e-01 9.44988251e-01 2.45600402e-01 -1.01799332e-01
-3.32356632e-01 -2.00581953e-01 -2.12538950e-02 9.11459982e-01
4.09566045e-01 3.13451022e-01 1.82595134e-01 8.03035021e-01
-2.30717152e-01 1.70234501e-01 -3.24753642e-01 -1.06068090e-01
1.03179431e+00 1.33484316e+00 -3.65290195e-01 -4.26325351e-01
-6.43972456e-01 1.69041947e-01 5.79908848e-01 2.50162959e-01
-6.97560191e-01 -2.80369788e-01 8.28578770e-01 2.76388884e-01
-2.11414650e-01 -2.14219838e-02 -4.55278665e-01 -1.17317247e+00
5.67882471e-02 -1.05002785e+00 7.22331107e-01 -6.69137418e-01
-1.21623468e+00 -4.60122414e-02 2.52000336e-02 -9.03449953e-01
-7.10697919e-02 -6.12253189e-01 -4.58116502e-01 7.52172709e-01
-1.43744826e+00 -9.31610405e-01 -6.77371860e-01 7.35983014e-01
3.68934900e-01 -2.76003152e-01 1.16254163e+00 6.41634345e-01
-8.44532847e-01 5.37399888e-01 6.13987409e-02 1.53108135e-01
8.45365822e-01 -1.35183573e+00 -2.00380564e-01 4.93286997e-01
1.94160014e-01 8.16848278e-01 5.05850494e-01 -6.67337477e-01
-1.38790619e+00 -9.80087817e-01 1.18552375e+00 -1.04105270e+00
6.43093467e-01 -7.00660571e-02 -1.28615344e+00 9.28839147e-01
-1.49146346e-02 -4.05475020e-01 1.43471944e+00 7.37757683e-01
-1.81096584e-01 -1.08287245e-01 -9.59120214e-01 4.51774746e-01
1.06084132e+00 -4.07171130e-01 -8.93646479e-01 2.72204757e-01
4.61982757e-01 -5.11454582e-01 -1.57480383e+00 1.89099237e-01
4.88644689e-01 -1.12219131e+00 9.05153990e-01 -9.04510617e-01
4.22093391e-01 -1.55126825e-01 1.61519516e-02 -9.23959076e-01
-3.98594648e-01 1.13510527e-02 -4.56574649e-01 1.66343069e+00
1.70926154e-01 -2.88882732e-01 1.11468554e+00 9.13671672e-01
-4.06921953e-01 -7.99704313e-01 -6.17324591e-01 -4.55688089e-01
6.78693131e-02 -6.54690206e-01 1.08343494e+00 1.47422886e+00
3.03714544e-01 8.75211731e-02 6.44850656e-02 2.41010159e-01
5.79040706e-01 4.95914817e-01 6.16235793e-01 -1.71445000e+00
-3.01828552e-02 -5.59863746e-01 -9.42122564e-02 -8.34010363e-01
1.75325558e-01 -9.83647525e-01 -3.18272531e-01 -1.56959116e+00
4.61762965e-01 -5.86664736e-01 -4.09646720e-01 6.35056019e-01
-3.71540815e-01 -2.42812205e-02 -2.66650021e-01 -6.94212411e-03
-8.71903718e-01 3.82432252e-01 1.45873511e+00 3.19115430e-01
-1.37852607e-02 1.84903532e-01 -1.02763486e+00 1.00847495e+00
8.97675991e-01 -5.14903545e-01 -4.99452859e-01 -2.12122664e-01
3.12352329e-01 -5.87556064e-02 2.66075671e-01 -1.00812113e+00
5.25254369e-01 -7.15818882e-01 4.05718297e-01 -5.56821346e-01
2.69697189e-01 -7.53252506e-01 1.21365041e-01 3.26634288e-01
-4.85008538e-01 -1.48294583e-01 8.07671472e-02 2.83547312e-01
-2.46681735e-01 -6.86972141e-01 3.68293226e-01 -2.87866145e-01
-8.59692097e-01 1.91166580e-01 -3.96780431e-01 2.60220736e-01
9.61682796e-01 3.15026939e-02 -7.99701571e-01 1.02705143e-01
-7.45444715e-01 3.73323649e-01 3.17022920e-01 7.35050738e-01
5.55104434e-01 -1.29349625e+00 -4.50210959e-01 2.62541413e-01
4.21076208e-01 -5.52753583e-02 5.85090876e-01 4.93796438e-01
-2.50400424e-01 6.70792401e-01 -3.48432958e-01 -1.13672398e-01
-1.46052337e+00 4.06352341e-01 1.31352812e-01 -2.66671151e-01
-3.22776705e-01 5.67545950e-01 -2.65293419e-01 -8.88007045e-01
5.21557808e-01 -1.74311712e-01 -8.21886003e-01 1.54155567e-01
6.45537972e-01 7.82568932e-01 1.66457415e-01 -1.63002402e-01
-9.03880745e-02 2.73763090e-01 1.70227170e-01 3.30271333e-01
1.20658755e+00 -2.09333524e-01 -6.22423552e-02 7.42645025e-01
3.25890064e-01 3.92058998e-01 -7.46631861e-01 -3.51079494e-01
3.14947575e-01 -4.71295387e-01 -2.33181007e-02 -1.05974126e+00
-8.02552581e-01 8.11009645e-01 5.16128659e-01 2.91789979e-01
9.81745839e-01 9.70600471e-02 3.29263061e-01 5.20227313e-01
2.31125042e-01 -1.02554750e+00 -3.43845896e-02 6.70695066e-01
2.57299006e-01 -1.02404857e+00 3.10885347e-02 -5.28159916e-01
-2.37961859e-01 1.00610685e+00 1.09544659e+00 6.55750155e-01
7.00499952e-01 -1.09466545e-01 2.08785757e-01 -4.30763096e-01
-1.01026428e+00 -2.86748379e-01 5.78131557e-01 4.69552845e-01
9.56956565e-01 3.91480565e-01 -3.76769960e-01 1.22677183e+00
-4.54301149e-01 2.27868825e-01 5.38083375e-01 8.33679855e-01
-8.95682573e-01 -1.33364642e+00 -5.40318251e-01 4.05088425e-01
-4.50538009e-01 1.07791305e-01 -2.71637589e-01 7.89163411e-01
4.44544345e-01 1.06137908e+00 -2.28679031e-01 -2.67277181e-01
7.10932612e-01 6.65038466e-01 3.90649885e-01 -8.81082356e-01
-7.28511870e-01 -4.71678913e-01 1.37069538e-01 -4.28023577e-01
-3.85992616e-01 -5.40111661e-01 -1.20609260e+00 -6.13354027e-01
-1.81552559e-01 4.59210783e-01 6.00582957e-01 9.52736795e-01
3.27863306e-01 8.72290313e-01 4.53244569e-03 3.18450928e-01
-4.59771842e-01 -1.00171387e+00 -6.50509357e-01 4.61639285e-01
-3.37042898e-01 -8.56272519e-01 4.97754812e-02 2.79584825e-01] | [10.118603706359863, 7.2196173667907715] |
e2bc3e1b-04d7-4a3f-a769-0aa95dbd1017 | non-invasive-blood-pressure-estimation-from | null | null | https://doi.org/10.3390/s18041160 | https://www.mdpi.com/280924 | Non-Invasive Blood Pressure Estimation from ECG Using Machine Learning Techniques | Background: Blood pressure (BP) measurements have been used widely in clinical and private environments. Recently, the use of ECG monitors has proliferated; however, they are not enabled with BP estimation. We have developed a method for BP estimation using only electrocardiogram (ECG) signals. Methods: Raw ECG data are filtered and segmented, and, following this, a complexity analysis is performed for feature extraction. Then, a machine-learning method is applied, combining a stacking-based classification module and a regression module for building systolic BP (SBP), diastolic BP (DBP), and mean arterial pressure (MAP) predictive models. In addition, the method allows a probability distribution-based calibration to adapt the models to a particular user. Results: Using ECG recordings from 51 different subjects, 3129 30-s ECG segments are constructed, and seven features are extracted. Using a train-validation-test evaluation, the method achieves a mean absolute error (MAE) of 8.64 mmHg for SBP, 18.20 mmHg for DBP, and 13.52 mmHg for the MAP prediction. When models are calibrated, the MAE decreases to 7.72 mmHg for SBP, 9.45 mmHg for DBP and 8.13 mmHg for MAP. Conclusion: The experimental results indicate that, when a probability distribution-based calibration is used, the proposed method can achieve results close to those of a certified medical device for BP estimation. | ['Ana Madevska Bogdanova', 'Matjaž Gams', 'Martin Gjoreski', 'Monika Simjanoska'] | 2018-04-18 | null | null | null | sensors-2018-4 | ['blood-pressure-estimation'] | ['medical'] | [ 1.15721911e-01 -3.28007750e-02 9.61108580e-02 -6.90441668e-01
-4.27794516e-01 -1.59550563e-01 -2.43378773e-01 5.38709283e-01
-3.82915378e-01 1.04895592e+00 -2.21748158e-01 -5.21501899e-01
-8.44162628e-02 -9.24525857e-01 -2.04030499e-01 -5.75868905e-01
-3.37431610e-01 1.75896838e-01 1.70741722e-01 1.58096790e-01
1.54386610e-01 4.99524355e-01 -1.11997807e+00 -3.76196764e-02
1.27856803e+00 1.17042363e+00 -4.00659382e-01 7.77233720e-01
1.06052972e-01 3.92672032e-01 -9.46648240e-01 -2.11693749e-01
3.85236770e-01 -7.09223151e-01 -2.51094162e-01 -1.87774271e-01
3.51258308e-01 -4.10249978e-01 3.24973166e-02 7.36936212e-01
9.20064032e-01 -1.91841051e-01 4.08006102e-01 -6.75126314e-01
5.62487962e-03 6.06393993e-01 -3.29613924e-01 7.49293938e-02
1.94724038e-01 6.42541721e-02 2.09442928e-01 -4.06288087e-01
-9.82188135e-02 6.63640678e-01 9.34503615e-01 2.33070850e-01
-1.68272448e+00 -5.79934180e-01 -3.77028435e-01 -7.72935972e-02
-1.63968050e+00 -2.50363320e-01 5.50900698e-01 -6.06546044e-01
3.77307236e-01 6.44261956e-01 1.16457927e+00 3.38273376e-01
6.16279364e-01 -1.63925007e-01 1.47449243e+00 -5.25297344e-01
3.51882398e-01 7.23739207e-01 4.22198296e-01 2.78705567e-01
6.44406021e-01 -4.46529128e-02 6.03223071e-02 -3.85536522e-01
8.89583588e-01 1.55809605e-02 -3.31493407e-01 5.39884642e-02
-6.67095065e-01 3.98866028e-01 8.24437290e-02 3.43220323e-01
-5.37436545e-01 -4.41494524e-01 3.51100206e-01 2.75584221e-01
2.26455793e-01 5.01612961e-01 -2.97651500e-01 -2.60176659e-01
-1.23060215e+00 3.68235081e-01 1.32648897e+00 7.25448608e-01
5.44588327e-01 -9.16313455e-02 -4.99533176e-01 7.59950817e-01
4.34564054e-01 8.94896150e-01 3.14360827e-01 -8.48978221e-01
2.49298096e-01 5.88933885e-01 4.48989540e-01 -1.32722044e+00
-7.01065660e-01 -5.02343118e-01 -1.09469235e+00 1.18700778e-02
5.66936672e-01 -4.42328691e-01 -6.60074353e-01 1.24491096e+00
8.40202048e-02 -1.63462888e-02 -2.15653688e-01 8.42820227e-01
6.53065205e-01 2.10413814e-01 3.85995656e-01 -6.50081813e-01
1.27912188e+00 -1.99536726e-01 -6.91994190e-01 8.92291293e-02
1.32514760e-01 -4.08843011e-01 7.52159476e-01 6.12498701e-01
-9.04473901e-01 -9.49317575e-01 -9.25510645e-01 5.47981024e-01
-5.47060482e-02 2.01734528e-01 8.10287334e-03 1.34116864e+00
-6.24926686e-01 9.38486993e-01 -7.37272263e-01 -2.58464307e-01
8.54623169e-02 2.00507775e-01 1.55475184e-01 4.61217433e-01
-1.40499687e+00 1.03081119e+00 2.01179102e-01 4.71640408e-01
-4.85842042e-02 -8.76912594e-01 -4.85306740e-01 6.12678118e-02
-4.76240665e-01 -6.20491385e-01 6.78899765e-01 -3.07113707e-01
-1.79651523e+00 5.90046585e-01 -1.43675953e-01 -4.88232434e-01
7.25604415e-01 -4.50410753e-01 -8.18115234e-01 3.98567200e-01
-8.14379007e-02 -1.43060625e-01 6.47502363e-01 -9.18079495e-01
-3.69059950e-01 -6.61401868e-01 -3.53464752e-01 -1.87994152e-01
3.80126350e-02 -3.01938597e-02 -8.25111568e-02 -1.06825925e-01
4.78350222e-01 -6.50185347e-01 -6.48515165e-01 -2.33129114e-01
-3.31639051e-01 5.28650641e-01 -4.77907285e-02 -1.23313475e+00
1.82274532e+00 -2.01345634e+00 -2.90163398e-01 1.04963124e+00
4.04715508e-01 5.32761037e-01 8.00548673e-01 1.63561791e-01
8.68107453e-02 7.32987747e-02 -4.31800425e-01 1.71814442e-01
-1.78457439e-01 -3.46665606e-02 1.62406843e-02 4.43300426e-01
-9.57451239e-02 5.17170191e-01 -5.97532988e-01 -5.79748631e-01
6.58808291e-01 3.42424572e-01 -3.22290599e-01 2.51686275e-01
4.30251926e-01 1.02562380e+00 -2.82176703e-01 3.25669795e-01
8.54041159e-01 9.59068015e-02 5.74994087e-01 -2.94692874e-01
-3.37886155e-01 -1.19725205e-01 -1.41820323e+00 1.03199625e+00
-2.97197998e-01 2.47294396e-01 -3.15023363e-01 -7.60297120e-01
1.79603457e+00 4.66436386e-01 5.95906258e-01 -6.66635334e-01
2.07527101e-01 4.75707799e-01 3.23953331e-01 -8.55631888e-01
-3.11494470e-01 -1.59301594e-01 2.99567431e-01 -1.23015465e-02
-4.07196045e-01 -1.65883794e-01 1.84041485e-01 -2.97622800e-01
7.56300688e-01 -9.54981819e-02 4.05575991e-01 -5.11218488e-01
1.00063396e+00 -3.09475422e-01 5.81979334e-01 8.33998978e-01
-4.14815128e-01 4.26622719e-01 4.94273573e-01 -8.50786090e-01
-8.77617002e-01 -1.13721168e+00 -9.62940335e-01 1.34824961e-01
-2.11513769e-02 -3.43295842e-01 -6.61143720e-01 2.07528565e-02
3.61009449e-01 4.82871950e-01 -1.67087585e-01 -1.06979504e-01
-6.85897768e-01 -1.01804745e+00 5.23228884e-01 4.22307312e-01
6.13802850e-01 -6.13987803e-01 -8.06322336e-01 5.96526682e-01
-8.58862400e-02 -7.32573628e-01 2.59251028e-01 -9.32020843e-02
-1.24198020e+00 -9.24462616e-01 -7.62266219e-01 -1.84346855e-01
4.70071793e-01 -6.08277857e-01 9.66964841e-01 -1.45838141e-01
-4.05468822e-01 1.47189379e-01 -1.12282291e-01 -6.89547122e-01
-4.35924828e-01 -2.46600389e-01 1.69771776e-01 2.09583387e-01
6.47034347e-01 -9.84118402e-01 -1.06046391e+00 5.13291061e-01
-3.66614084e-03 -3.41642588e-01 4.77064222e-01 2.84256786e-01
6.00979745e-01 -5.11555791e-01 8.48735154e-01 -8.23870778e-01
7.19129443e-01 -1.18674979e-01 -7.07707703e-01 1.34974688e-01
-1.14060390e+00 -5.52852273e-01 7.08756268e-01 -1.64422303e-01
-6.34895861e-01 -9.88653600e-02 -2.23369688e-01 5.98243177e-02
-4.04884338e-01 3.22675020e-01 -2.15059370e-02 5.23901656e-02
1.11016476e+00 2.10241646e-01 3.25141519e-01 -5.44408500e-01
-1.90824151e-01 1.04583156e+00 4.66356725e-01 -6.52436495e-01
3.84444147e-01 -1.18274555e-01 2.80778199e-01 -1.05868328e+00
-5.54322541e-01 -2.47453704e-01 -8.49800289e-01 -4.67115760e-01
7.79170632e-01 -7.72315264e-01 -9.74788427e-01 3.78320426e-01
-6.00715876e-01 2.08867006e-02 -2.41381779e-01 9.53178346e-01
-4.76375133e-01 4.69207227e-01 -4.54366893e-01 -1.05356610e+00
-8.26413035e-01 -6.93204761e-01 2.90219247e-01 3.67842525e-01
-5.01699328e-01 -6.86734259e-01 2.82744635e-02 1.89676985e-01
8.26415896e-01 6.96935594e-01 6.06489837e-01 -6.19617522e-01
9.35800432e-04 -6.38646126e-01 -1.63898125e-01 6.76090777e-01
3.75816882e-01 -6.87543601e-02 -8.43769670e-01 -8.86121467e-02
2.72311598e-01 3.85315895e-01 1.77218810e-01 6.64189041e-01
1.14444017e+00 -1.32031158e-01 -2.12481841e-01 5.08425832e-01
1.42614102e+00 5.57436943e-01 1.04938340e+00 2.18951613e-01
2.89246470e-01 5.48047237e-02 3.25192869e-01 7.37997472e-01
2.41519421e-01 4.03075308e-01 -6.27568960e-02 -1.37924001e-01
3.32515061e-01 1.36086583e-01 -1.29180312e-01 4.67559129e-01
-7.90147662e-01 5.26300073e-01 -9.12973344e-01 -1.02635361e-01
-1.59647727e+00 -8.24769437e-01 -5.85411370e-01 2.69901252e+00
1.05508590e+00 2.34742060e-01 3.92861426e-01 4.45149302e-01
7.79956222e-01 -4.29121763e-01 -4.06266540e-01 -5.98492742e-01
3.46820652e-01 5.23538470e-01 4.65547562e-01 3.54016483e-01
-1.08359897e+00 8.24830383e-02 6.37392712e+00 -3.34926188e-01
-1.32786596e+00 -2.85222024e-01 7.61455595e-01 4.22351092e-01
4.61743861e-01 -1.87429741e-01 -7.16163814e-01 8.30199957e-01
1.32475436e+00 -3.59733731e-01 4.19242755e-02 8.32413793e-01
6.57186747e-01 -2.32594416e-01 -8.36166918e-01 1.17499673e+00
-1.78102180e-01 -8.64093065e-01 -4.79643881e-01 -2.06296235e-01
2.28558540e-01 -5.57533681e-01 -5.25739551e-01 3.27824712e-01
-6.75558686e-01 -8.42238128e-01 1.30264059e-01 1.13759482e+00
8.20612013e-01 -4.81097788e-01 1.17230034e+00 3.12848479e-01
-8.37858379e-01 -8.59919861e-02 -4.36295629e-01 -3.53607297e-01
1.45475194e-01 1.26872265e+00 -6.68795824e-01 5.96065640e-01
6.12987757e-01 3.20219338e-01 -5.62796712e-01 1.44518554e+00
-2.17581049e-01 7.44116187e-01 -3.94453853e-01 -6.05575778e-02
-5.23510695e-01 -4.93501991e-01 3.30313802e-01 1.19613862e+00
2.55174071e-01 2.50880361e-01 2.68381625e-01 1.09608471e+00
6.30534947e-01 5.83612382e-01 -5.23515046e-02 6.04792535e-01
4.57049608e-01 9.68034625e-01 -3.06120425e-01 -5.25109470e-01
-1.69742271e-01 3.95983249e-01 -2.29388043e-01 2.64250964e-01
-1.01043427e+00 -8.20439577e-01 1.44695014e-01 4.15647119e-01
-2.78691173e-01 1.02108546e-01 -8.51466298e-01 -8.60243678e-01
3.07723731e-01 -6.16808057e-01 2.17883319e-01 -1.76992416e-01
-1.07796144e+00 4.86274034e-01 1.53092474e-01 -1.29870987e+00
-1.62180007e-01 -4.18428183e-01 -6.67658210e-01 1.73601890e+00
-1.15357876e+00 -3.37585926e-01 -5.93933046e-01 2.14115277e-01
-2.61013150e-01 8.08117092e-02 1.17854524e+00 7.33255565e-01
-6.18054152e-01 5.18575490e-01 -2.01291814e-01 2.73777902e-01
6.85264409e-01 -1.27425742e+00 -1.48399189e-01 5.67811012e-01
-7.33329535e-01 8.78971219e-01 6.92855954e-01 -5.39047837e-01
-1.01754820e+00 -1.11419547e+00 1.12622249e+00 -2.89778054e-01
1.57628953e-01 2.80406997e-02 -8.18311572e-01 2.70964861e-01
-3.88335526e-01 2.80843288e-01 9.43819940e-01 2.05611531e-02
1.65517628e-01 -6.83168292e-01 -1.50361323e+00 3.11005265e-01
2.54206777e-01 -3.07256263e-02 -5.42838395e-01 -3.66108827e-02
2.86284331e-02 -6.89800322e-01 -1.85066235e+00 6.07448816e-01
1.22429955e+00 -1.09906900e+00 8.69080126e-01 -3.65516335e-01
5.31285778e-02 -4.06700879e-01 1.06859878e-01 -1.01165855e+00
-4.00173426e-01 -5.26378572e-01 -1.13770574e-01 1.03992462e+00
6.18276119e-01 -1.08008158e+00 5.02706647e-01 1.37092304e+00
4.61713485e-02 -8.61586511e-01 -7.29042590e-01 -6.63094044e-01
-1.57630280e-01 -6.18981663e-03 4.60317761e-01 5.85695565e-01
3.05502355e-01 2.46085316e-01 -3.83557886e-01 2.43421912e-01
7.32031882e-01 2.05070794e-01 8.42839062e-01 -1.57267833e+00
-3.26416641e-01 -1.52878180e-01 -7.02884078e-01 -6.65714979e-01
-6.90218627e-01 -3.85488510e-01 -2.50394762e-01 -1.51229036e+00
-9.59570557e-02 -5.16520917e-01 -6.60699964e-01 1.88584000e-01
-3.14722061e-01 7.98083916e-02 4.73014340e-02 -3.91033664e-02
2.22231001e-01 -4.80546840e-02 1.07154751e+00 2.27643475e-01
-1.03577900e+00 7.16713130e-01 -3.78570825e-01 4.83518630e-01
1.28145635e+00 -3.76403093e-01 -1.68727264e-01 3.85621876e-01
-5.73833920e-02 2.98494458e-01 2.41137847e-01 -1.47769082e+00
-2.22392872e-01 -3.36184651e-02 1.01965058e+00 -4.25460875e-01
-1.81436371e-02 -9.38215375e-01 4.19151157e-01 1.01740420e+00
-7.60491341e-02 -3.65068018e-01 1.50414884e-01 1.50219873e-01
-9.88275781e-02 -1.35540947e-01 9.79001343e-01 -2.25750163e-01
-8.00282508e-02 1.05288737e-01 -3.35679114e-01 -1.82751819e-01
9.37267363e-01 -4.69970852e-01 1.74956850e-03 -1.28356829e-01
-1.30255842e+00 -5.54578262e-04 -1.36237174e-01 -1.39189586e-01
5.11924028e-01 -1.07796729e+00 -8.42384100e-01 4.44675624e-01
3.46203074e-02 -1.99867636e-01 3.54012936e-01 1.39810371e+00
-7.98243821e-01 1.52663529e-01 -2.87992030e-01 -8.44333589e-01
-1.12179506e+00 1.50061965e-01 7.18780518e-01 5.40953204e-02
-9.20815825e-01 2.33904332e-01 -7.48491049e-01 -4.90875356e-03
2.59232342e-01 -5.47116876e-01 -4.46803749e-01 -1.30455285e-01
8.07540357e-01 7.27325320e-01 1.40567780e-01 -6.27931673e-03
-2.78947502e-01 6.08369172e-01 4.99289036e-01 2.47715533e-01
9.97237623e-01 -2.34948322e-01 -3.69058043e-01 6.36091530e-01
5.61812162e-01 9.01154205e-02 -4.47442323e-01 -3.85359414e-02
-1.58269048e-01 -4.58208531e-01 -3.90735179e-01 -9.29358006e-01
-7.19866395e-01 8.75690460e-01 1.08593357e+00 5.33710599e-01
1.18219745e+00 -5.49379349e-01 4.05131906e-01 1.20839566e-01
3.71672153e-01 -1.01828277e+00 -7.69105375e-01 2.31378883e-01
7.57753730e-01 -7.60812819e-01 1.97823927e-01 -5.47457993e-01
-4.28188950e-01 1.37216210e+00 5.12327313e-01 -1.13289960e-01
1.18275905e+00 -7.47837648e-02 3.03099036e-01 1.90781251e-01
-4.13111523e-02 2.06800461e-01 1.21614151e-01 5.18990576e-01
6.52992308e-01 4.73631203e-01 -8.89598727e-01 8.84527862e-01
-2.46152386e-01 4.64304745e-01 3.77470016e-01 7.04219878e-01
-8.06613505e-01 -8.44040155e-01 -3.92749608e-01 8.49175394e-01
-5.16175926e-01 4.72191237e-02 1.92761347e-01 5.07490396e-01
2.06890866e-01 1.10499573e+00 -2.55086012e-02 -2.39492834e-01
8.63397121e-01 5.59114814e-01 4.61309165e-01 -4.78092819e-01
-5.51288605e-01 6.77300468e-02 1.00062221e-01 -3.44701231e-01
-3.69887680e-01 -2.92212278e-01 -1.09924173e+00 -2.22844064e-01
-1.05996624e-01 3.74006063e-01 6.76186800e-01 6.51844025e-01
4.63748306e-01 5.21324635e-01 8.12366068e-01 -3.16954762e-01
-8.41373682e-01 -1.13539052e+00 -9.52357471e-01 2.48252153e-01
1.27401292e-01 -3.02366257e-01 -3.28386962e-01 1.04159191e-01] | [14.087732315063477, 2.9890785217285156] |
0dc0a347-7058-42e1-8e23-4a47130d3104 | master-thesis-neural-sign-language | 2011.09289 | null | https://arxiv.org/abs/2011.09289v1 | https://arxiv.org/pdf/2011.09289v1.pdf | Master Thesis: Neural Sign Language Translation by Learning Tokenization | In this thesis, we propose a multitask learning based method to improve Neural Sign Language Translation (NSLT) consisting of two parts, a tokenization layer and Neural Machine Translation (NMT). The tokenization part focuses on how Sign Language (SL) videos should be represented to be fed into the other part. It has not been studied elaborately whereas NMT research has attracted several researchers contributing enormous advancements. Up to now, there are two main input tokenization levels, namely frame-level and gloss-level tokenization. Glosses are world-like intermediate presentation and unique to SLs. Therefore, we aim to develop a generic sign-level tokenization layer so that it is applicable to other domains without further effort. We begin with investigating current tokenization approaches and explain their weaknesses with several experiments. To provide a solution, we adapt Transfer Learning, Multitask Learning and Unsupervised Domain Adaptation into this research to leverage additional supervision. We succeed in enabling knowledge transfer between SLs and improve translation quality by 5 points in BLEU-4 and 8 points in ROUGE scores. Secondly, we show the effects of body parts by extensive experiments in all the tokenization approaches. Apart from these, we adopt 3D-CNNs to improve efficiency in terms of time and space. Lastly, we discuss the advantages of sign-level tokenization over gloss-level tokenization. To sum up, our proposed method eliminates the need for gloss level annotation to obtain higher scores by providing additional supervision by utilizing weak supervision sources. | ['Alptekin Orbay'] | 2020-11-18 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 2.83262819e-01 -6.58015683e-02 -4.62415874e-01 -3.66664886e-01
-9.69700515e-01 -4.80770051e-01 6.16050601e-01 -4.02308792e-01
-6.82889044e-01 7.22100317e-01 4.52313662e-01 -2.51852334e-01
2.58779734e-01 -3.32488656e-01 -7.61675656e-01 -4.80865985e-01
3.60943437e-01 3.13154459e-01 2.62995869e-01 -2.34205961e-01
1.03037477e-01 1.37995288e-01 -1.37343526e+00 6.61924899e-01
8.52367282e-01 5.64559340e-01 1.47980556e-01 4.69666123e-01
-2.76712805e-01 5.41488111e-01 -7.82220483e-01 -4.73679096e-01
2.46967778e-01 -6.37800455e-01 -1.02853954e+00 -9.14611965e-02
8.23260427e-01 -5.51010370e-01 6.22992255e-02 7.36246586e-01
8.94694448e-01 9.43555310e-02 5.47020078e-01 -1.21178877e+00
-7.98161209e-01 4.71612394e-01 -3.51681411e-01 -2.68434674e-01
4.12322909e-01 2.48711929e-01 9.11327839e-01 -8.25239420e-01
8.45561981e-01 1.34417152e+00 7.17337668e-01 9.85493839e-01
-5.97283006e-01 -7.93372810e-01 1.27013296e-01 2.88563222e-01
-8.67897451e-01 -2.44750157e-01 6.43447876e-01 -3.23643237e-01
1.04718077e+00 -1.08379917e-02 6.62889600e-01 1.41546810e+00
6.95526525e-02 1.23587000e+00 1.33491611e+00 -7.18083382e-01
-1.04471825e-01 -4.65288460e-02 -1.65774465e-01 8.47467065e-01
1.17870703e-01 -4.47481200e-02 -7.90681541e-01 4.30257827e-01
8.57635379e-01 -3.30465198e-01 1.35887936e-01 -1.26542300e-01
-1.50836647e+00 5.14489233e-01 1.02055237e-01 6.17898226e-01
-1.24617860e-01 4.03269380e-01 7.87394106e-01 4.58703756e-01
2.92012006e-01 1.12084270e-01 -4.95432109e-01 -3.86467516e-01
-1.05750608e+00 6.75751418e-02 4.18149978e-01 1.09331191e+00
5.18197536e-01 2.02023372e-01 -4.78532732e-01 8.48321497e-01
3.84243846e-01 7.08433867e-01 7.85121560e-01 -8.16407502e-01
7.03997016e-01 4.71268654e-01 -1.84206720e-02 -3.56508732e-01
-1.91350490e-01 -1.98069111e-01 -3.52044880e-01 3.73339534e-01
6.51845455e-01 -3.33578557e-01 -1.38068080e+00 1.81539810e+00
9.73392837e-03 1.15774892e-01 2.45299488e-02 1.06076336e+00
9.70546842e-01 2.80022740e-01 4.43906754e-01 1.69785053e-01
1.51241994e+00 -1.24439275e+00 -8.40953290e-01 8.97789747e-02
1.07870996e+00 -1.06012166e+00 1.52934361e+00 4.02745694e-01
-1.02078152e+00 -6.13357604e-01 -9.14646983e-01 -3.61848801e-01
-5.81231654e-01 6.68522537e-01 6.32249296e-01 7.59776533e-01
-9.62719440e-01 4.69774663e-01 -9.96843278e-01 -8.36205721e-01
2.00704738e-01 4.48125333e-01 -3.70554805e-01 8.17984119e-02
-1.25529742e+00 1.29101121e+00 2.18942925e-01 -1.70162711e-02
-6.81765020e-01 -3.36309373e-01 -8.67244840e-01 -4.18559432e-01
9.80063155e-02 -8.96164179e-01 1.39410245e+00 -1.29671705e+00
-1.92461443e+00 1.03199697e+00 -3.53461176e-01 -2.02791274e-01
7.32651293e-01 -3.51052970e-01 -3.12385887e-01 1.10589616e-01
1.14309505e-01 9.25253630e-01 8.38483274e-01 -9.55672145e-01
-8.85474563e-01 -2.52512265e-02 1.35570288e-01 4.54394698e-01
-4.14560586e-01 4.14505392e-01 -5.26943803e-01 -9.12021577e-01
-7.05037415e-02 -9.26361382e-01 2.78620362e-01 1.17532695e-02
-2.08936390e-02 -5.04735172e-01 9.87258255e-01 -8.03369284e-01
8.27616811e-01 -1.89746451e+00 1.67539597e-01 -2.01156408e-01
-7.77597502e-02 5.72512746e-01 -4.21035707e-01 2.91085869e-01
4.26044948e-02 1.11598805e-01 3.28549109e-02 -6.44087315e-01
1.45474732e-01 5.31179428e-01 7.50782862e-02 2.38012314e-01
3.03826720e-01 1.17937887e+00 -9.74767923e-01 -6.86159909e-01
2.92883337e-01 4.89111483e-01 -5.11473119e-01 -1.32450461e-01
-1.44982278e-01 5.63604474e-01 -2.51182675e-01 1.07316530e+00
3.56862366e-01 1.45484999e-01 -3.49632800e-02 -3.72450233e-01
-2.37595096e-01 3.08171898e-01 -9.82959151e-01 2.09281206e+00
-5.90399146e-01 7.79458821e-01 -7.90523291e-02 -1.06018984e+00
6.48602724e-01 5.87769330e-01 5.65199614e-01 -7.29729474e-01
2.75067270e-01 4.72854227e-01 -1.83678225e-01 -7.88913310e-01
6.27516866e-01 -3.39367986e-01 -1.81671828e-01 4.35061783e-01
3.77060622e-01 9.71616656e-02 2.76255608e-01 -2.39420980e-01
7.29143500e-01 8.41887116e-01 4.03862558e-02 1.02976345e-01
3.49113762e-01 1.54731378e-01 6.59480274e-01 4.18188542e-01
-4.88662213e-01 4.95440513e-01 1.07957996e-01 -1.59026071e-01
-1.00904357e+00 -9.26989615e-01 3.58194470e-01 1.53766310e+00
-6.71967044e-02 -3.74020219e-01 -7.43378103e-01 -8.74115288e-01
-1.49258152e-01 4.39451844e-01 -4.11282748e-01 1.58063591e-01
-9.70764279e-01 -3.51762056e-01 1.08536458e+00 7.76166677e-01
7.22312808e-01 -1.13055563e+00 -6.51099443e-01 -1.14521116e-01
-5.22162497e-01 -1.16654706e+00 -5.99445760e-01 2.87487637e-02
-9.24077511e-01 -7.61178553e-01 -1.26707971e+00 -1.16230345e+00
5.64382732e-01 7.71586373e-02 6.35111153e-01 -1.73638776e-01
-3.04445531e-03 4.78012681e-01 -6.99566543e-01 -6.33235991e-01
-4.27890927e-01 1.74528569e-01 1.55452147e-01 -2.67482400e-01
4.70512867e-01 -3.66423279e-01 -2.15317294e-01 3.82434458e-01
-8.38725746e-01 2.05381572e-01 1.11154079e+00 8.18853319e-01
3.75757724e-01 -7.16492176e-01 3.61523747e-01 -4.00256246e-01
7.10941553e-01 1.64694816e-01 -1.41174108e-01 4.98210907e-01
-4.42339331e-01 1.91849962e-01 3.67972672e-01 -6.82760537e-01
-1.06747496e+00 1.68972790e-01 -7.66561329e-02 -3.57496887e-01
-3.02900136e-01 3.38380635e-01 -7.47888386e-02 -2.51131475e-01
4.23310906e-01 2.58980781e-01 9.52299163e-02 -5.03565907e-01
5.64231336e-01 8.38188469e-01 5.00539780e-01 -8.71830344e-01
8.32578540e-01 3.49292219e-01 -1.23886503e-01 -7.39772260e-01
-6.12897217e-01 -4.59991246e-01 -9.99823868e-01 -5.32907009e-01
1.05772400e+00 -8.22429180e-01 -6.56087816e-01 5.86151600e-01
-1.12710726e+00 -4.82525468e-01 -4.19440299e-01 8.40173066e-01
-7.74780691e-01 6.71045303e-01 -6.14620745e-01 -4.31444347e-01
-2.72185177e-01 -1.25255477e+00 1.24271834e+00 1.66678727e-01
-2.53385127e-01 -9.44241524e-01 6.59172460e-02 7.36962914e-01
4.83424097e-01 1.26301646e-01 6.99166775e-01 -5.22981048e-01
-4.42576021e-01 1.79460317e-01 -2.60116965e-01 5.17571092e-01
3.12030852e-01 -1.53227895e-01 -8.00653577e-01 -2.58881867e-01
-5.07625639e-01 -6.12437963e-01 7.95808613e-01 3.18737477e-01
4.67500657e-01 -4.68698293e-02 -1.01654984e-01 5.61302960e-01
9.99809325e-01 4.82708439e-02 5.96213102e-01 6.33563876e-01
7.05061316e-01 4.94333267e-01 8.12299788e-01 -6.10519014e-02
5.65107584e-01 8.83927822e-01 1.37110520e-02 -4.25589889e-01
-9.48238790e-01 -3.56697440e-01 8.30363989e-01 1.05457461e+00
-7.02540517e-01 6.76812381e-02 -6.85658157e-01 6.59884512e-01
-1.86081541e+00 -9.64712560e-01 5.35901710e-02 1.92076087e+00
9.37819183e-01 -1.07265234e-01 3.39296311e-01 -1.35258529e-02
5.56748033e-01 -6.29581288e-02 -3.77420932e-01 -6.40468240e-01
-1.51393175e-01 2.48732880e-01 6.90995455e-01 4.18405443e-01
-1.12780213e+00 1.55145991e+00 6.26041317e+00 6.99746966e-01
-1.49372613e+00 3.58445346e-01 -1.37638032e-01 -1.05644837e-01
9.33115196e-04 -9.81749892e-02 -8.95026982e-01 2.57605582e-01
7.46731043e-01 2.04173043e-01 1.37723476e-01 7.34152257e-01
5.64825594e-01 -1.85122844e-02 -9.02260125e-01 8.84606779e-01
3.46731573e-01 -1.01327300e+00 3.61619353e-01 6.93958299e-03
7.28948116e-01 8.09513777e-02 5.16051687e-02 5.67138672e-01
1.58805326e-01 -6.79817498e-01 8.61396909e-01 3.55899155e-01
1.05105281e+00 -3.10064077e-01 7.78940618e-01 5.25201671e-02
-1.27275121e+00 1.80967256e-01 -6.81612194e-02 -8.10031220e-02
3.46277148e-01 -2.37503108e-02 -8.88597906e-01 7.99126148e-01
4.93416548e-01 5.24579763e-01 -2.36106157e-01 9.83719528e-01
-5.62318385e-01 6.11906350e-01 -3.46208096e-01 -1.44449115e-01
5.15761554e-01 8.60887207e-03 4.32216078e-01 1.62165451e+00
3.86790633e-01 -3.18386614e-01 1.87179208e-01 4.14157301e-01
3.73655115e-03 2.65156686e-01 -4.53657180e-01 1.18924864e-01
1.73440315e-02 8.76899838e-01 -5.93932033e-01 -5.35183966e-01
-6.23233438e-01 1.39394414e+00 -1.18289188e-01 4.47442681e-01
-9.52194870e-01 -3.99493545e-01 6.59107149e-01 -3.95524725e-02
2.74661541e-01 -4.65899944e-01 -3.66189450e-01 -1.35504985e+00
1.35800868e-01 -9.05686557e-01 1.97333142e-01 -6.32871389e-01
-9.27037358e-01 2.57215858e-01 1.87803239e-01 -1.59830594e+00
-3.07437509e-01 -8.28838825e-01 -4.07750636e-01 8.06918561e-01
-1.91586018e+00 -2.01199460e+00 -2.44756103e-01 7.47815847e-01
7.31801271e-01 -1.34851828e-01 8.11880708e-01 6.09635055e-01
-4.57724094e-01 9.95792568e-01 -1.13369077e-01 4.85899895e-01
1.42341876e+00 -1.15378153e+00 2.78242081e-01 9.88044679e-01
1.45379260e-01 6.12550139e-01 4.23376620e-01 -7.60191977e-01
-1.05457973e+00 -9.81073141e-01 1.41435337e+00 -3.78464252e-01
5.65156341e-01 -1.26947522e-01 -3.71644020e-01 7.35490859e-01
2.82061994e-01 -1.95342481e-01 5.46760976e-01 -1.79650467e-02
-3.81582886e-01 4.55428697e-02 -1.13674653e+00 6.29819214e-01
1.15593493e+00 -5.94296157e-01 -9.12238300e-01 2.98762560e-01
3.99012655e-01 -5.43305695e-01 -8.75822067e-01 3.82422864e-01
9.77098107e-01 -6.00813389e-01 6.79013908e-01 -6.76784277e-01
3.51449937e-01 -4.39285219e-01 3.46847698e-02 -1.24583554e+00
-7.51461536e-02 -6.06778562e-01 1.23425141e-01 1.22161794e+00
4.17845696e-01 -5.25497913e-01 9.40026283e-01 2.55389035e-01
-5.60439110e-01 -4.78206277e-01 -1.06530440e+00 -1.14978325e+00
3.18716884e-01 -5.48676312e-01 2.84956098e-01 1.02687001e+00
1.55931100e-01 1.53669134e-01 -6.08121872e-01 -1.54606178e-01
3.81199628e-01 -1.51913920e-02 9.10255373e-01 -8.09305906e-01
-7.43239596e-02 -5.95028996e-01 -2.57494628e-01 -1.26702583e+00
4.43997160e-02 -9.24765348e-01 3.55008827e-03 -1.84226906e+00
-3.42319198e-02 -4.44071218e-02 -2.33364940e-01 1.14622986e+00
8.68762359e-02 5.07867813e-01 4.70152408e-01 2.68803328e-01
-5.73319674e-01 3.91801089e-01 1.65399683e+00 -1.99550003e-01
-2.50921726e-01 -9.02874842e-02 -2.31285393e-01 7.54081190e-01
8.40237916e-01 -2.97848284e-01 -1.93972975e-01 -8.68454754e-01
-2.33242020e-01 -3.71238559e-01 1.66197911e-01 -1.01727593e+00
1.88606828e-01 -1.19898804e-01 1.35014042e-01 -5.75660586e-01
3.29128921e-01 -5.78285754e-01 -4.14925456e-01 4.18862402e-01
-2.71229595e-01 9.39958841e-02 1.89743921e-01 5.60900662e-03
-2.53109664e-01 -6.47177920e-02 7.82890201e-01 -1.98673397e-01
-1.09886146e+00 -9.35009792e-02 -5.85730433e-01 -2.62156069e-01
9.32062328e-01 -6.25316501e-01 -1.30836666e-01 -2.92936206e-01
-7.52817571e-01 1.66184470e-01 2.83199966e-01 5.85362136e-01
4.58151937e-01 -1.41418970e+00 -6.74777150e-01 1.02559805e-01
2.24303797e-01 -4.22144830e-01 -1.31071463e-01 1.04345417e+00
-4.95159328e-01 8.08349013e-01 -6.41864538e-01 -4.22056407e-01
-1.58542585e+00 -4.67896871e-02 1.43019646e-01 -3.63119304e-01
-5.29421151e-01 7.06584692e-01 -3.62437785e-01 -5.80358744e-01
5.41172445e-01 -9.01188433e-01 -1.34472415e-01 1.40951738e-01
2.29138970e-01 3.38847876e-01 3.78357209e-02 -7.39906132e-01
-3.61129165e-01 1.00491619e+00 2.59175133e-02 -3.57692182e-01
1.04274225e+00 1.52063807e-02 2.63893545e-01 2.79306442e-01
8.77221107e-01 -2.51147915e-02 -1.05868626e+00 -9.53076482e-02
1.27448395e-01 -1.57978803e-01 -2.53119439e-01 -1.13126779e+00
-8.34916234e-01 8.91364753e-01 7.09172547e-01 -5.38765252e-01
1.16893482e+00 -6.88210502e-02 1.14475417e+00 5.35945952e-01
2.45987386e-01 -1.45049810e+00 1.11374922e-01 6.60057604e-01
7.36850083e-01 -1.19091535e+00 -2.07990870e-01 -6.15166947e-02
-6.11259460e-01 1.18840861e+00 5.76855361e-01 2.25772373e-02
6.02036528e-02 2.44683534e-01 4.87067610e-01 7.46711940e-02
-2.34647050e-01 -5.39298594e-01 4.91855174e-01 8.83660257e-01
6.58823490e-01 -3.25203165e-02 -8.23094785e-01 2.65335321e-01
-1.51725128e-01 3.94024044e-01 1.67224094e-01 1.01501369e+00
-4.25207317e-01 -1.68424785e+00 -4.76585358e-01 4.29601371e-02
-4.80495840e-01 -1.39809493e-02 -5.19221961e-01 1.06671810e+00
5.92365503e-01 7.65502572e-01 -4.01703686e-01 -5.91589928e-01
4.98574585e-01 3.47952127e-01 8.32750976e-01 -6.00040495e-01
-7.12485433e-01 5.12160026e-02 2.70761609e-01 -4.08965260e-01
-9.71008003e-01 -7.30020344e-01 -1.22907543e+00 -1.15125999e-01
-2.16008440e-01 -2.05066025e-01 9.32790041e-01 1.38309598e+00
9.12132040e-02 4.57958996e-01 -1.45592913e-01 -9.03876185e-01
-4.23355758e-01 -1.23882377e+00 -1.05799027e-01 3.30928385e-01
1.10533156e-01 -6.91010535e-01 5.02522290e-02 4.24223453e-01] | [9.182079315185547, -6.502620220184326] |
d5b89ab7-fd1d-4f2b-b90d-ec3b11c5dafe | cmta-covid-19-misinformation-multilingual | null | null | https://aclanthology.org/2021.acl-srw.28 | https://aclanthology.org/2021.acl-srw.28.pdf | CMTA: COVID-19 Misinformation Multilingual Analysis on Twitter | The internet has actually come to be an essential resource of health knowledge for individuals around the world in the present situation of the coronavirus condition pandemic(COVID-19). During pandemic situations, myths, sensationalism, rumours and misinformation, generated intentionally or unintentionally, spread rapidly through social networks. Twitter is one of these popular social networks people use to share COVID-19 related news, information, and thoughts that reflect their perception and opinion about the pandemic. Evaluation of tweets for recognizing misinformation can create beneficial understanding to review the top quality and also the readability of online information concerning the COVID-19. This paper presents a multilingual COVID-19 related tweet analysis method, CMTA, that uses BERT, a deep learning model for multilingual tweet misinformation detection and classification. CMTA extracts features from multilingual textual data, which is then categorized into specific information classes. Classification is done by a Dense-CNN model trained on tweets manually annotated into information classes (i.e., {`}false{'}, {`}partly false{'}, {`}misleading{'}). The paper presents an analysis of multilingual tweets from February to June, showing the distribution type of information spread across different languages. To access the performance of the CMTA multilingual model, we performed a comparative analysis of 8 monolingual model and CMTA for the misinformation detection task. The results show that our proposed CMTA model has surpassed various monolingual models which consolidated the fact that through transfer learning a multilingual framework could be developed. | ['Genoveva Vargas-Solar', 'Ambesh Shekhar', 'Mehrdad Farokhenajd', 'Raj Pranesh'] | 2021-08-01 | null | null | null | acl-2021-5 | ['rumour-detection'] | ['natural-language-processing'] | [-2.53329575e-01 1.26153678e-01 -1.68238744e-01 3.16458195e-02
-6.39841855e-01 -5.71879268e-01 1.10596585e+00 7.78923154e-01
-5.83376706e-01 9.23893094e-01 5.74061275e-01 -5.12926817e-01
3.63101780e-01 -9.45344627e-01 -6.64304733e-01 -4.75889266e-01
-1.01254627e-01 9.39105690e-01 -2.50942379e-01 -7.27381945e-01
1.83407545e-01 1.66724056e-01 -1.04018366e+00 7.10472167e-01
6.34672284e-01 6.83483303e-01 -3.68622988e-02 5.89980543e-01
-4.94559497e-01 1.10001945e+00 -9.44840670e-01 -4.33297694e-01
-3.77024114e-01 -1.37993202e-01 -9.26315308e-01 -5.54352105e-01
9.88108441e-02 -2.84625828e-01 8.16718414e-02 1.09326363e+00
4.99833733e-01 -6.73973918e-01 8.05129528e-01 -1.08984852e+00
-5.58991671e-01 8.33938777e-01 -3.66914481e-01 6.76501453e-01
6.55692577e-01 -1.09515995e-01 3.72542650e-01 -7.41017103e-01
1.10482264e+00 1.38130748e+00 1.17716968e+00 1.02970168e-01
-5.48203051e-01 -9.64491844e-01 -2.22754613e-01 9.81072709e-02
-1.36615109e+00 1.79218158e-01 4.74192560e-01 -1.15934861e+00
1.00472021e+00 2.93427587e-01 7.44177580e-01 1.55877006e+00
7.74858594e-01 5.49343228e-01 1.33088458e+00 1.21950291e-01
-4.38511729e-01 7.53147840e-01 1.11789428e-01 7.01610446e-01
3.39334339e-01 2.13434137e-02 -2.22908109e-01 -4.00329500e-01
-1.03547558e-01 2.72529963e-02 -2.25613475e-01 6.70482993e-01
-1.35071337e+00 1.34728384e+00 5.78376234e-01 7.54834533e-01
-5.28526664e-01 -5.71141005e-01 7.90432990e-01 4.45215821e-01
1.12879777e+00 2.07441211e-01 -4.29223597e-01 1.64362296e-01
-8.23376536e-01 3.06908339e-01 9.86939609e-01 4.48577702e-01
7.57151723e-01 -1.64147347e-01 -6.22463860e-02 5.36396146e-01
3.24915200e-01 1.25507653e+00 4.59259599e-01 5.26741780e-02
5.51130652e-01 5.15336931e-01 1.45918727e-01 -1.92420197e+00
-1.19139338e+00 -6.76963925e-01 -1.13238764e+00 -5.12916923e-01
2.33396944e-02 -6.50670052e-01 -4.57615554e-01 1.44612908e+00
2.73830146e-01 -2.28682440e-02 6.11730441e-02 6.58328474e-01
1.37879324e+00 1.03851163e+00 3.01336814e-02 -2.85964608e-01
1.53363335e+00 -5.48904955e-01 -1.08323920e+00 1.58217609e-01
6.25168324e-01 -1.19298768e+00 4.29228008e-01 1.92663401e-01
-5.07254839e-01 -2.83237547e-01 -9.26486671e-01 8.45209062e-02
-1.21727777e+00 -3.19491416e-01 9.54066366e-02 6.80888116e-01
-1.04765987e+00 1.59258857e-01 -2.50589699e-01 -5.98496318e-01
2.64446110e-01 -8.96508470e-02 -2.52980173e-01 1.35653704e-01
-1.86759651e+00 1.19922590e+00 6.48524821e-01 -1.46390069e-02
-9.83903170e-01 -5.73352337e-01 -7.16379941e-01 -6.90714657e-01
-1.75126076e-01 -5.01631081e-01 9.53072548e-01 -8.90818536e-01
-8.31905246e-01 1.24756515e+00 7.37961233e-02 -5.73621631e-01
5.34238517e-01 -1.59425408e-01 -1.10812008e+00 4.47273180e-02
5.86952865e-01 3.52174550e-01 6.65348768e-01 -1.12662613e+00
-6.42636538e-01 -2.74015993e-01 -1.04762405e-01 -2.58223355e-01
-6.85412660e-02 4.74241406e-01 2.59003103e-01 -8.24095547e-01
-5.87906182e-01 -8.67440879e-01 2.66729295e-01 -7.53629923e-01
-7.63599634e-01 -1.78786770e-01 9.19960499e-01 -1.12496376e+00
1.20419288e+00 -1.71708822e+00 -3.86784852e-01 1.09470136e-01
4.99426574e-01 5.73107660e-01 1.57858506e-01 9.95620728e-01
2.73942322e-01 5.44568777e-01 -5.90873510e-02 8.57254118e-02
-1.96023196e-01 1.05049528e-01 -3.43674809e-01 6.79132700e-01
8.18322748e-02 8.16618085e-01 -1.24845517e+00 -2.17634976e-01
-9.92332548e-02 8.62277687e-01 -4.13086504e-01 1.09279275e-01
-3.42285752e-01 8.18778634e-01 -5.70785224e-01 2.30958849e-01
1.02266085e+00 -4.30399090e-01 8.67684111e-02 -3.25417280e-01
-5.66759527e-01 3.73485208e-01 -3.06751519e-01 8.95656288e-01
-5.26240289e-01 9.46667731e-01 1.91056039e-02 -5.37014842e-01
6.76535606e-01 5.80843508e-01 4.45854664e-01 -7.75642693e-01
6.98252320e-01 2.08589733e-01 -4.13034290e-01 -8.32554162e-01
5.38236856e-01 -9.49032307e-02 -3.90933722e-01 6.23089969e-01
-3.86111468e-01 3.58005047e-01 -1.15228198e-01 3.39750618e-01
4.72648889e-01 -6.68362379e-01 4.60308403e-01 -5.05925715e-01
7.09865093e-01 5.00428081e-01 4.17671241e-02 6.55460119e-01
-7.57928789e-02 2.21348837e-01 1.69780180e-01 -7.90318668e-01
-8.41266453e-01 -8.37169945e-01 -3.51164222e-01 1.02534723e+00
-2.69621938e-01 -7.53331780e-02 -8.08077157e-01 -5.10556936e-01
-1.59964696e-01 6.36132240e-01 -8.30855191e-01 3.24687630e-01
-4.78622615e-01 -1.20471978e+00 9.82420504e-01 -2.61135340e-01
7.83018887e-01 -1.17824697e+00 -4.63807374e-01 1.54141381e-01
-9.67597961e-01 -9.91745353e-01 -3.23128074e-01 -3.19371492e-01
5.16794138e-02 -1.24120331e+00 -6.72898650e-01 -8.33408654e-01
3.41973394e-01 1.47500366e-01 1.13094366e+00 -3.54791135e-02
-2.53556557e-02 2.13276565e-01 -4.48475361e-01 -7.76422679e-01
-1.01101327e+00 2.50961315e-02 6.63574487e-02 4.06129099e-02
5.05571246e-01 1.32860604e-03 -4.78151917e-01 -1.25749439e-01
-1.01111495e+00 -2.43204050e-02 3.53334814e-01 5.44528484e-01
8.00479352e-02 -5.48392273e-02 8.18876684e-01 -1.42357576e+00
9.06666636e-01 -1.43821716e+00 -3.57055590e-02 -1.89078718e-01
-3.40495378e-01 -3.26953948e-01 4.71150279e-01 -1.04437523e-01
-8.90913963e-01 -9.41196322e-01 -5.79775453e-01 2.84942031e-01
-3.27803135e-01 1.17929876e+00 5.12282372e-01 3.30901861e-01
7.69817650e-01 2.03627259e-01 -1.14693761e-01 -4.23863679e-01
2.56480277e-01 1.32987404e+00 3.05858165e-01 2.76273131e-01
4.90408212e-01 7.44299233e-01 -6.56059742e-01 -1.10331368e+00
-1.06697118e+00 -6.65172577e-01 -2.89841205e-01 -3.60653102e-01
1.21159387e+00 -1.10501671e+00 -7.18075037e-01 7.88023233e-01
-1.64755023e+00 2.12960944e-01 5.55716813e-01 2.22205073e-01
8.79449472e-02 1.30423844e-01 -9.87848818e-01 -5.33583224e-01
-6.98771954e-01 -8.88467193e-01 7.45663285e-01 -2.68162996e-01
-4.14625406e-01 -1.45750535e+00 6.11179531e-01 6.05541587e-01
9.12959099e-01 6.85361326e-01 8.85440290e-01 -1.32695556e+00
1.40124381e-01 -2.04404503e-01 -5.76211214e-01 -9.94760245e-02
3.42032433e-01 -3.25440198e-01 -1.07419527e+00 -4.51549679e-01
4.59833257e-02 -4.74059910e-01 6.87996924e-01 5.89836463e-02
3.05836409e-01 -1.20525563e+00 -4.55346376e-01 7.88567811e-02
1.29516840e+00 8.45624581e-02 3.10017943e-01 4.34529901e-01
6.29680336e-01 5.83073676e-01 1.45536423e-01 6.85409367e-01
8.90319824e-01 4.02401239e-01 4.26673055e-01 -2.02594802e-01
6.57663643e-02 -3.43133688e-01 5.89512706e-01 1.38459325e+00
1.91188768e-01 -4.50879306e-01 -1.08964622e+00 5.21405101e-01
-1.30527639e+00 -1.13241744e+00 -3.34925205e-01 1.79345846e+00
9.78675485e-01 -6.27687201e-02 1.96016282e-01 -2.43211746e-01
8.30566943e-01 3.39805603e-01 -2.83419788e-02 -7.43719220e-01
-3.10993195e-01 -2.16456294e-01 4.75686431e-01 7.60570824e-01
-1.24414170e+00 6.60055697e-01 5.93028831e+00 6.93292499e-01
-1.64668572e+00 5.96325636e-01 4.53655064e-01 6.55832052e-01
-5.13821065e-01 -7.90342093e-01 -8.25927019e-01 5.42672217e-01
1.22782874e+00 -9.86118913e-02 9.88953486e-02 3.16739738e-01
5.26184440e-01 2.35716283e-01 -2.30160296e-01 6.85714424e-01
6.25232995e-01 -1.66430295e+00 4.00398433e-01 -1.72487691e-01
9.20699716e-01 8.93290639e-01 1.38670847e-01 1.94013700e-01
3.06677192e-01 -8.90697777e-01 6.27441525e-01 5.64899325e-01
6.21171832e-01 -7.50166655e-01 1.36985874e+00 5.31367123e-01
-8.26550364e-01 3.18964928e-01 -1.19127840e-01 1.60122216e-01
2.67090231e-01 7.65522122e-01 -1.54341757e+00 6.37224495e-01
6.46691799e-01 9.34891045e-01 -4.91425037e-01 5.78313947e-01
5.76209910e-02 5.78137159e-01 -8.93426389e-02 -3.27584416e-01
5.74048936e-01 1.80169240e-01 8.33434105e-01 2.07405639e+00
2.48673037e-01 -4.91797388e-01 3.80870044e-01 6.76762819e-01
1.33859396e-01 6.17487133e-01 -1.09547877e+00 -4.06744480e-01
1.78200617e-01 9.61047828e-01 -5.48689604e-01 -4.71261919e-01
-2.41256088e-01 6.56768143e-01 6.77176341e-02 1.67914718e-01
-8.74374926e-01 -9.58334208e-02 2.65306123e-02 3.45210314e-01
-2.77202539e-02 2.25479290e-01 2.10634395e-01 -9.49499726e-01
-5.54123878e-01 -1.06744909e+00 5.61281502e-01 -5.63108385e-01
-1.50119793e+00 1.16966009e+00 -2.55224612e-02 -1.07915604e+00
-4.46497232e-01 -4.27300066e-01 -3.29681903e-01 6.33315742e-01
-1.57413507e+00 -1.38160658e+00 -6.75505325e-02 6.72724128e-01
4.14550841e-01 -2.94450074e-01 8.33358109e-01 6.71528041e-01
-2.62291014e-01 3.33646715e-01 6.72885850e-02 3.41091067e-01
5.29550314e-01 -7.39015698e-01 -3.34622525e-02 1.58978641e-01
-3.44993651e-01 5.11280596e-01 8.83336306e-01 -1.04187226e+00
-6.93689227e-01 -1.59474409e+00 1.72191978e+00 -4.13256943e-01
1.07014585e+00 -1.09510958e-01 -6.37149155e-01 7.80156970e-01
8.78429949e-01 -9.55048084e-01 9.60002601e-01 -1.52282700e-01
-5.26924968e-01 2.84955978e-01 -1.42816269e+00 3.65829170e-01
3.13155979e-01 -6.52976215e-01 -6.58970237e-01 9.59839344e-01
8.54381442e-01 -1.16020270e-01 -7.04368234e-01 2.21956432e-01
4.31704909e-01 -9.32780206e-01 7.82117963e-01 -7.29648292e-01
3.44450295e-01 -2.99044028e-02 2.47311797e-02 -1.48676133e+00
-1.70792416e-01 -4.66290027e-01 5.98900437e-01 9.19601798e-01
6.75028563e-01 -1.00666702e+00 -1.01182975e-01 -5.65076113e-01
-1.05716750e-01 -1.63790032e-01 -8.44417214e-01 -1.99182227e-01
3.67029667e-01 -5.11002719e-01 4.51327175e-01 1.48677754e+00
1.37689918e-01 4.23075885e-01 -6.99405193e-01 1.94252551e-01
2.63591290e-01 -9.46909115e-02 4.05155003e-01 -1.11475444e+00
3.41421396e-01 -4.83536333e-01 -1.51954710e-01 -5.13334811e-01
-1.05323672e-01 -1.19829118e+00 -1.77734166e-01 -1.35634601e+00
3.34442765e-01 -3.47750932e-01 -8.56462494e-02 1.74004182e-01
2.96576977e-01 5.89481533e-01 -6.37211800e-02 2.28708312e-01
-5.24557889e-01 -7.35663176e-02 1.12794960e+00 -5.33261001e-01
9.92708877e-02 -7.67527297e-02 -4.92059886e-01 8.41199338e-01
9.21771526e-01 -5.58384717e-01 -1.07552804e-01 -2.38321230e-01
1.09719169e+00 -4.56069261e-02 2.34368667e-01 -7.07966983e-01
-2.81361174e-02 1.86177224e-01 2.63295043e-03 -1.06137288e+00
-7.77270459e-03 -5.85535765e-01 3.50630730e-01 9.74527061e-01
-3.52424264e-01 4.58545357e-01 2.94223905e-01 4.93587673e-01
-3.60842228e-01 1.95888147e-01 6.00939870e-01 -2.34033167e-01
-1.87144727e-01 1.53466523e-01 -1.25964177e+00 3.57180178e-01
5.31969190e-01 5.32964289e-01 -8.87525976e-01 -5.99641383e-01
-7.83816993e-01 5.45221195e-02 -8.38328078e-02 6.96095705e-01
4.77910250e-01 -1.16993856e+00 -1.21619248e+00 2.13601246e-01
2.45405942e-01 -5.69159925e-01 2.53578007e-01 1.19589484e+00
-9.21655059e-01 1.05690444e+00 -2.90010184e-01 -4.59005207e-01
-7.91331768e-01 8.48022640e-01 1.92148656e-01 -5.01873672e-01
-2.21015856e-01 4.12915111e-01 5.11348993e-02 -8.49438190e-01
-8.56314152e-02 -2.64556706e-01 -9.05747592e-01 6.36507809e-01
8.86268079e-01 4.74672914e-01 8.39117765e-02 -1.45003879e+00
-4.01917070e-01 3.57217729e-01 -1.34203464e-01 2.43490309e-01
9.52226341e-01 -4.20852393e-01 -6.66497707e-01 6.47031665e-01
1.57627130e+00 2.51971841e-01 4.45972346e-02 -1.17607407e-01
-1.41966134e-01 8.27875137e-02 -1.14361830e-01 -1.23009670e+00
-9.12600756e-01 7.43422151e-01 6.36091113e-01 6.47892356e-01
5.89954495e-01 -1.02715157e-02 1.20966196e+00 1.94752082e-01
3.24286729e-01 -6.96434379e-01 -1.86348498e-01 7.81360984e-01
1.27318966e+00 -1.36848319e+00 -4.10851538e-01 -1.63424000e-01
-6.37961388e-01 9.14208770e-01 -1.21506281e-01 1.82742670e-01
1.30227029e+00 9.63560492e-02 6.55254722e-01 -6.66663647e-01
-3.73256505e-01 -6.52454747e-03 2.76345849e-01 7.43228436e-01
4.99506146e-01 4.56706733e-01 -5.61186790e-01 4.29074079e-01
-4.07129526e-01 -1.07734241e-01 4.90283221e-01 4.87584084e-01
-5.43150663e-01 -5.66735148e-01 -5.04966378e-01 2.98332155e-01
-9.99068677e-01 -2.29219899e-01 -4.97534782e-01 6.99501872e-01
7.15214074e-01 1.23246312e+00 3.51609178e-02 -5.06340742e-01
-2.03579038e-01 -1.88230589e-01 -3.20671767e-01 -3.04563820e-01
-9.96590853e-01 -7.85360336e-02 4.94906455e-01 -8.26083720e-02
-7.43624568e-01 -4.02564734e-01 -1.05813825e+00 -6.97877467e-01
1.54690400e-01 2.87593931e-01 6.17672980e-01 1.25793242e+00
2.31046587e-01 2.45686576e-01 6.32471323e-01 -5.17915666e-01
1.28232926e-01 -1.08018100e+00 -7.15515316e-02 5.42202652e-01
8.21746349e-01 -3.06257367e-01 -3.50635588e-01 -1.18366264e-01] | [8.431573867797852, 9.856377601623535] |
2c48db47-cbbc-4ba9-8cd2-fcbf443915d3 | multi-microphone-complex-spectral-mapping-for | 2010.01703 | null | https://arxiv.org/abs/2010.01703v2 | https://arxiv.org/pdf/2010.01703v2.pdf | Multi-microphone Complex Spectral Mapping for Utterance-wise and Continuous Speech Separation | We propose multi-microphone complex spectral mapping, a simple way of applying deep learning for time-varying non-linear beamforming, for speaker separation in reverberant conditions. We aim at both speaker separation and dereverberation. Our study first investigates offline utterance-wise speaker separation and then extends to block-online continuous speech separation (CSS). Assuming a fixed array geometry between training and testing, we train deep neural networks (DNN) to predict the real and imaginary (RI) components of target speech at a reference microphone from the RI components of multiple microphones. We then integrate multi-microphone complex spectral mapping with minimum variance distortionless response (MVDR) beamforming and post-filtering to further improve separation, and combine it with frame-level speaker counting for block-online CSS. Although our system is trained on simulated room impulse responses (RIR) based on a fixed number of microphones arranged in a given geometry, it generalizes well to a real array with the same geometry. State-of-the-art separation performance is obtained on the simulated two-talker SMS-WSJ corpus and the real-recorded LibriCSS dataset. | ['DeLiang Wang', 'Peidong Wang', 'Zhong-Qiu Wang'] | 2020-10-04 | null | null | null | null | ['speaker-separation'] | ['speech'] | [ 2.40862086e-01 -5.32989800e-01 7.46646941e-01 -2.25697979e-01
-1.46913195e+00 -7.62203813e-01 2.09992364e-01 -3.33783895e-01
-2.89688706e-01 3.56387258e-01 4.80888098e-01 -6.09856009e-01
-2.18902841e-01 -3.12202740e-02 -6.87375247e-01 -1.07028103e+00
-3.23597789e-01 -6.44548014e-02 -2.34691307e-01 -2.01432124e-01
-3.89453918e-01 5.94652712e-01 -1.51878178e+00 3.66671562e-01
6.31771266e-01 9.96861994e-01 2.38549486e-01 1.53722548e+00
3.45164388e-01 5.05571961e-01 -1.09005392e+00 3.28141004e-02
3.72370750e-01 -3.74095052e-01 -2.23616064e-01 -2.10371718e-01
7.10089087e-01 -2.60547459e-01 -3.78438413e-01 8.52849841e-01
1.30418217e+00 4.68218356e-01 5.49925983e-01 -5.50903082e-01
-2.52897143e-01 7.16652036e-01 -2.55373299e-01 5.23098230e-01
4.40415978e-01 -1.20686740e-01 6.06049597e-01 -1.18366408e+00
-4.80965525e-01 1.23829782e+00 8.26854408e-01 5.52403569e-01
-1.07178259e+00 -8.84579599e-01 1.15362048e-01 1.29866347e-01
-1.51171124e+00 -1.38857472e+00 7.45076537e-01 -2.27554083e-01
1.17096961e+00 8.42413902e-01 2.22835734e-01 1.12500083e+00
-3.61147672e-01 5.26983738e-01 5.96036673e-01 -6.20581031e-01
1.91745281e-01 -2.34920010e-01 1.00532971e-01 3.22213471e-02
-4.54195380e-01 1.63556173e-01 -7.49893248e-01 -2.62488779e-02
4.09865171e-01 -2.69213200e-01 -9.01498079e-01 2.25668952e-01
-1.31037211e+00 3.38180274e-01 1.82461262e-01 5.36820710e-01
-1.75854489e-01 -7.71229044e-02 1.57122731e-01 4.56149012e-01
5.03678203e-01 3.22208107e-01 -5.52495956e-01 -9.57053807e-03
-1.28969908e+00 -7.33809918e-02 1.04440355e+00 7.16563582e-01
1.03561558e-01 7.56388426e-01 -2.69884169e-01 1.38243198e+00
3.27808648e-01 1.06993425e+00 5.81865013e-01 -8.65867019e-01
6.25709176e-01 -7.32608616e-01 2.25750506e-01 -7.21956968e-01
-3.16049099e-01 -1.05218256e+00 -9.86002326e-01 6.39972985e-02
3.68526906e-01 -5.61340451e-01 -8.11675191e-01 1.80838311e+00
3.58373702e-01 7.35469878e-01 3.31157148e-01 1.10464668e+00
9.40915465e-01 9.11538064e-01 -7.98511207e-01 -5.12534738e-01
9.62353110e-01 -1.24046981e+00 -8.09287608e-01 -3.46704960e-01
2.05003843e-02 -9.43321526e-01 6.77710891e-01 7.45270193e-01
-1.30085957e+00 -8.20620418e-01 -1.21150470e+00 1.67388558e-01
-8.53096396e-02 1.09660633e-01 -3.15480858e-01 1.12277937e+00
-1.31809664e+00 1.21818170e-01 -8.74635935e-01 3.31584901e-01
-2.89071262e-01 3.60665143e-01 -1.44210488e-01 3.27030309e-02
-1.18788207e+00 3.85318190e-01 -5.88593543e-01 6.02520764e-01
-1.07644928e+00 -8.45571399e-01 -8.76281798e-01 3.67052704e-01
2.11439803e-01 -4.90003794e-01 1.50705922e+00 -1.05780315e+00
-1.80813360e+00 2.53199130e-01 -7.56244004e-01 -4.66852188e-01
1.85561463e-01 -5.17971694e-01 -1.18111908e+00 -9.18337051e-03
-1.77237228e-01 -1.13569915e-01 1.34113336e+00 -1.41129768e+00
-3.07573020e-01 -1.77552670e-01 -2.89287329e-01 4.62542564e-01
-2.84201801e-01 2.30197534e-01 -3.59037071e-02 -6.92481220e-01
2.41385743e-01 -4.39885706e-01 5.09601012e-02 -7.21016049e-01
-5.47232687e-01 3.74142528e-01 2.94968873e-01 -1.35683155e+00
1.20440364e+00 -2.62614846e+00 2.29252428e-01 1.19736291e-01
-6.24833889e-02 5.67635238e-01 -3.67583513e-01 2.72499382e-01
-4.82384473e-01 -3.78022939e-01 -1.14480913e-01 -1.02121353e+00
2.19840482e-02 -5.51175892e-01 -5.00669956e-01 6.63373470e-01
-4.18488756e-02 1.80583999e-01 -6.43428445e-01 5.20219803e-02
3.07326168e-01 7.92590201e-01 -4.84049708e-01 5.16213715e-01
5.86878359e-01 5.61549187e-01 4.70465094e-01 4.11190420e-01
1.09228015e+00 4.36760753e-01 -1.40852138e-01 -1.17725566e-01
-2.46067435e-01 6.04765415e-01 -1.59114599e+00 1.51002049e+00
-9.50368285e-01 9.48380947e-01 1.24130213e+00 -1.04722404e+00
7.68405557e-01 7.12149024e-01 1.78649768e-01 -5.71255803e-01
-6.39194250e-02 4.81377125e-01 3.90315801e-01 -3.21034044e-01
2.33780608e-01 -1.23062022e-01 3.27923775e-01 6.36750385e-02
1.89460680e-01 5.41899465e-02 -3.12440187e-01 -2.36252770e-01
1.19490266e+00 -6.21063173e-01 -4.06392142e-02 -6.27207905e-02
9.15723979e-01 -1.01040268e+00 1.95253566e-01 8.39757264e-01
-6.42832508e-03 1.00508606e+00 -4.17777538e-01 2.69872785e-01
-6.59532845e-01 -1.51981354e+00 1.84187666e-02 1.35974717e+00
-2.40736902e-01 2.35776026e-02 -8.59409392e-01 8.51991475e-02
-3.13401192e-01 1.07264197e+00 2.91821659e-02 -1.70073211e-02
-9.42252815e-01 -3.34881335e-01 6.94472253e-01 2.79149294e-01
1.79649070e-01 -5.70230186e-01 -1.21445991e-01 3.32397968e-01
-3.41418386e-01 -9.98102903e-01 -9.75329220e-01 6.05043352e-01
-2.01077625e-01 -5.64785719e-01 -1.10220611e+00 -9.75364029e-01
3.09082448e-01 8.18481266e-01 8.24235380e-01 -4.16798502e-01
-3.16666998e-02 6.91200793e-01 -1.11437485e-01 -3.51332396e-01
-5.51073313e-01 -3.80539536e-01 5.65891862e-01 3.84793460e-01
-9.23187062e-02 -8.29548836e-01 -7.80758381e-01 4.98092920e-01
-6.11143291e-01 -3.15647274e-01 2.17308760e-01 7.70079315e-01
5.80566674e-02 3.54185462e-01 7.10495710e-01 1.70980439e-01
7.43719637e-01 -2.27048412e-01 -3.44569117e-01 1.58277497e-01
1.68734223e-01 -4.21820074e-01 9.24982965e-01 -7.68576682e-01
-1.25113106e+00 -3.30905139e-01 -6.47461355e-01 -5.81194520e-01
-2.50817567e-01 2.87886202e-01 -4.99976695e-01 2.60677069e-01
7.98773587e-01 4.72444177e-01 -2.60862231e-01 -6.55349255e-01
2.99336761e-01 1.27042079e+00 8.41760278e-01 -1.09915905e-01
6.87081456e-01 1.63461730e-01 -4.21706825e-01 -1.23136246e+00
-4.37153339e-01 -8.92401159e-01 -3.71494800e-01 -1.38860662e-02
5.39932907e-01 -1.26774371e+00 -8.58726859e-01 7.74356484e-01
-1.28041887e+00 -5.97386658e-01 -8.80833566e-02 8.25841486e-01
-1.78759038e-01 2.14004844e-01 -4.60776478e-01 -1.45571494e+00
-4.23163682e-01 -1.00419235e+00 1.21575356e+00 -2.06966251e-02
7.04860911e-02 -8.47949505e-01 1.55194864e-01 4.05416071e-01
6.14875734e-01 -4.96020496e-01 2.59579241e-01 -7.08273232e-01
-1.23159856e-01 -2.11064547e-01 3.37198138e-01 9.99086916e-01
3.91227305e-01 -4.30995256e-01 -1.51488400e+00 -5.88247836e-01
8.31378162e-01 2.40630805e-01 8.22844148e-01 1.01365840e+00
9.08527315e-01 -5.09759247e-01 -1.61423981e-01 7.71919549e-01
9.25264001e-01 5.13988554e-01 3.28254670e-01 -2.56059587e-01
6.43812180e-01 4.37281787e-01 1.06960967e-01 1.64429232e-01
6.12782985e-02 6.18348181e-01 1.84294209e-01 -3.44352275e-01
-5.00529766e-01 3.13648641e-01 6.78708792e-01 1.46445382e+00
3.11789989e-01 -7.06685185e-01 -8.08749855e-01 6.64197922e-01
-1.25832260e+00 -8.93995225e-01 -4.59536277e-02 2.50856519e+00
6.73968554e-01 -1.95427313e-01 1.11234322e-01 5.32587826e-01
8.52000415e-01 1.14888489e-01 -5.72147608e-01 -3.83262932e-01
-3.04147542e-01 5.21325707e-01 4.25774872e-01 1.10228920e+00
-9.72511590e-01 1.25319421e-01 5.66328859e+00 8.91274154e-01
-1.28440332e+00 3.01777422e-01 5.43146670e-01 -6.07754230e-01
-6.15733936e-02 -8.64031851e-01 -6.23153210e-01 2.06021667e-01
1.52664614e+00 2.10937172e-01 1.16532791e+00 2.92321354e-01
3.46597701e-01 1.32560179e-01 -1.44197893e+00 1.27838433e+00
3.48033369e-01 -8.05218339e-01 -7.03491986e-01 -2.39258692e-01
3.98294389e-01 1.01030342e-01 4.61507976e-01 2.70887315e-01
-5.92070520e-02 -1.22396553e+00 9.74070847e-01 4.45884109e-01
7.59363949e-01 -5.94474733e-01 5.16915202e-01 5.85112631e-01
-1.23193622e+00 -3.53154957e-01 3.54595087e-03 4.44027185e-02
8.57503340e-02 7.03773260e-01 -1.14096129e+00 5.79482436e-01
6.76316857e-01 1.59325361e-01 1.79678332e-02 1.00667548e+00
4.87134084e-02 9.49206889e-01 -4.13700968e-01 2.05786020e-01
-2.74437904e-01 9.61973295e-02 9.51811194e-01 1.65830374e+00
6.81294084e-01 5.79842627e-02 -2.94555873e-01 4.91210103e-01
1.12647541e-01 -2.59695739e-01 -1.66121006e-01 2.93827534e-01
7.09423482e-01 9.87158835e-01 -2.44454294e-01 -1.15271010e-01
-2.21354708e-01 9.34635997e-01 -1.88308135e-01 1.01998842e+00
-7.83857107e-01 -5.50602674e-01 9.41311955e-01 -5.90379052e-02
7.01788306e-01 -5.43409765e-01 -1.14039674e-01 -9.79894996e-01
1.11071169e-01 -1.19247258e+00 -1.45121962e-01 -7.90827513e-01
-1.21589506e+00 6.75495684e-01 -2.29330674e-01 -1.14787233e+00
-3.81557703e-01 -6.03035510e-01 -7.07818210e-01 1.54907548e+00
-1.35717070e+00 -7.01583147e-01 8.72214511e-02 7.35893548e-01
7.98208833e-01 -1.72124714e-01 1.03386986e+00 7.06468940e-01
-4.87954706e-01 8.84093761e-01 7.32798457e-01 2.07281038e-01
7.96758831e-01 -1.30878532e+00 6.16273522e-01 1.03129673e+00
1.92196131e-01 7.62413263e-01 9.35004950e-01 -5.21165654e-02
-1.35493124e+00 -8.26261222e-01 4.74784344e-01 -1.71258584e-01
3.95068318e-01 -1.04062462e+00 -8.82979393e-01 3.22695494e-01
3.17609280e-01 -4.95972559e-02 9.36585844e-01 7.79431760e-02
-2.94884175e-01 -4.65772212e-01 -8.66895854e-01 5.26122928e-01
9.72331285e-01 -6.26085758e-01 -6.02095962e-01 3.18334103e-01
9.92910743e-01 -6.75835609e-01 -3.36528003e-01 1.21321321e-01
4.16993558e-01 -9.94160235e-01 1.53272474e+00 -2.17694476e-01
-5.74030131e-02 -3.63724321e-01 -6.14512086e-01 -1.85032487e+00
-3.39754701e-01 -9.68372881e-01 -2.72132486e-01 1.38809836e+00
5.02022326e-01 -7.92820215e-01 1.99870437e-01 2.27344915e-01
-7.30983317e-01 -2.06295222e-01 -1.36142290e+00 -9.58501279e-01
6.41856808e-03 -6.65569246e-01 5.93175232e-01 5.82052648e-01
-3.10827941e-01 4.22890693e-01 -3.84433717e-01 9.98690248e-01
5.48795283e-01 -2.29675621e-01 6.91907108e-01 -7.33204961e-01
-7.94354916e-01 -4.68479484e-01 2.27603182e-01 -1.57574117e+00
1.57701656e-01 -4.91563112e-01 5.48434496e-01 -1.35688698e+00
-7.16270804e-01 -1.05463423e-01 -3.69289875e-01 -3.10648263e-01
-2.78204173e-01 -1.77982897e-01 8.83954018e-02 -3.13573271e-01
-2.93134838e-01 5.58549821e-01 8.52893174e-01 -3.60282421e-01
-3.94281596e-01 5.98035574e-01 -5.46703637e-01 5.19554377e-01
3.52087259e-01 -1.85866401e-01 -3.03924233e-01 -6.21106505e-01
-1.41923115e-01 5.06229103e-01 2.62813777e-01 -1.35048604e+00
5.04178464e-01 4.49498057e-01 4.04945284e-01 -6.25019729e-01
9.42289770e-01 -9.40734625e-01 2.46009856e-01 1.57506838e-01
-4.88124609e-01 -5.04142880e-01 4.61850166e-01 4.74087983e-01
-2.25422412e-01 -1.15405194e-01 5.54882526e-01 1.60034254e-01
2.11230859e-01 -1.86037168e-01 -5.11910677e-01 -1.21745415e-01
2.46197596e-01 -1.41397387e-01 -9.85310525e-02 -7.44446456e-01
-9.46985126e-01 -4.69875596e-02 -3.11170191e-01 1.46343544e-01
5.73345959e-01 -1.22197676e+00 -9.78791595e-01 6.03707135e-01
-5.76847613e-01 2.36224338e-01 6.41196132e-01 8.20663929e-01
-9.18778181e-02 5.54443955e-01 5.30153573e-01 -6.90624952e-01
-1.59787893e+00 4.58327860e-01 8.49711955e-01 1.79591388e-01
-8.84447396e-02 1.54185748e+00 4.75221217e-01 -5.41535676e-01
7.45401740e-01 -4.98135954e-01 -2.23560259e-02 -2.03032538e-01
8.70279789e-01 7.07503438e-01 5.16960859e-01 -7.51454413e-01
-4.38927472e-01 4.17381555e-01 3.22603494e-01 -6.38394117e-01
1.16408539e+00 -5.16320586e-01 2.84863561e-02 8.17680836e-01
1.40294015e+00 8.55150580e-01 -1.15000618e+00 -2.98520297e-01
-6.47306323e-01 -5.05858123e-01 1.65467992e-01 -9.82214868e-01
-7.10102618e-01 1.20003510e+00 8.90937448e-01 6.80169940e-01
1.46039557e+00 -2.77866900e-01 6.54147327e-01 3.87547165e-01
-6.16774298e-02 -6.45600379e-01 1.41036257e-01 6.59539521e-01
1.01615524e+00 -9.20992076e-01 -6.06903851e-01 -1.08822756e-01
-5.83359972e-02 1.05189633e+00 2.40139723e-01 1.35094374e-01
9.03092563e-01 5.33123136e-01 3.43205750e-01 4.45406467e-01
-3.97544265e-01 1.10593706e-01 5.00958860e-01 7.91303098e-01
5.92009366e-01 1.77276924e-01 6.63668513e-01 8.20005357e-01
-5.96401393e-01 -7.08542526e-01 2.07810968e-01 3.24178517e-01
-4.39995110e-01 -6.05938971e-01 -1.20901239e+00 1.38581365e-01
-4.79684263e-01 -5.72936296e-01 -1.55237526e-01 -2.48026419e-02
-1.52839273e-01 1.75641644e+00 1.68183386e-01 -2.97832429e-01
5.58274508e-01 1.96739420e-01 2.77025759e-01 -5.09412110e-01
-7.35451221e-01 7.96726286e-01 1.05274312e-01 1.99921783e-02
-2.78539985e-01 -8.00209343e-01 -9.65376198e-01 -2.14103207e-01
-4.62014496e-01 1.79546535e-01 9.62206721e-01 7.78847694e-01
2.59546250e-01 1.25971532e+00 9.84309673e-01 -1.22430682e+00
-6.61635876e-01 -1.21888852e+00 -7.20663369e-01 -2.87230387e-02
1.52890885e+00 -6.83007948e-03 -8.94174457e-01 -8.89830105e-03] | [15.01480484008789, 5.931805610656738] |
d834f819-cc29-462b-88c2-1d8afc73d3fa | low-cost-relevance-generation-and-evaluation | 2205.10298 | null | https://arxiv.org/abs/2205.10298v1 | https://arxiv.org/pdf/2205.10298v1.pdf | Low-cost Relevance Generation and Evaluation Metrics for Entity Resolution in AI | Entity Resolution (ER) in voice assistants is a prime component during run time that resolves entities in users request to real world entities. ER involves two major functionalities 1. Relevance generation and 2. Ranking. In this paper we propose a low cost relevance generation framework by generating features using customer implicit and explicit feedback signals. The generated relevance datasets can serve as test sets to measure ER performance. We also introduce a set of metrics that accurately measures the performance of ER systems in various dimensions. They provide great interpretability to deep dive and identifying root cause of ER issues, whether the problem is in relevance generation or ranking. | ['Kurtis Voris', 'Haotian Jiang', 'Jitesh Mehta', 'Mina Ghashami', 'Venkat Varada'] | 2022-05-20 | null | null | null | null | ['entity-resolution'] | ['natural-language-processing'] | [-1.99728861e-01 5.02111793e-01 8.46958235e-02 -5.72233796e-01
-1.01061285e+00 -5.39386868e-01 5.01383901e-01 2.53774405e-01
-4.17149276e-01 1.13593066e+00 4.17766809e-01 -8.26491639e-02
-5.88816822e-01 -8.62488627e-01 -1.20679304e-01 1.11715861e-01
-6.85101897e-02 1.08910251e+00 2.91514546e-01 -7.88787127e-01
3.28579515e-01 6.74327970e-01 -1.63816094e+00 4.47386950e-01
1.14289904e+00 8.94934595e-01 2.23092452e-01 9.32094991e-01
-6.25600755e-01 9.59032834e-01 -1.13552082e+00 -4.87055153e-01
-1.10815261e-02 4.07626867e-01 -9.32717919e-01 -5.29958189e-01
4.69098911e-02 -2.13827312e-01 1.31956451e-02 1.02215803e+00
9.95202541e-01 2.77054071e-01 4.20473158e-01 -1.47614908e+00
-7.29954362e-01 9.54108179e-01 -2.76525989e-02 5.83152175e-01
9.19766009e-01 -2.93241084e-01 1.45368481e+00 -1.09480190e+00
7.61428714e-01 1.42284226e+00 4.53722119e-01 4.84487474e-01
-8.24716985e-01 -4.95410949e-01 1.02727406e-01 4.05930132e-01
-1.18278050e+00 -7.68332779e-01 3.05984437e-01 -2.96361744e-01
1.26675749e+00 7.80427873e-01 7.01873675e-02 6.67514026e-01
-1.64847434e-01 7.80229628e-01 6.96350336e-01 -3.26851606e-01
1.63741797e-01 7.33260632e-01 5.71916342e-01 2.32825771e-01
4.53131020e-01 -2.38215730e-01 -5.20820022e-01 -5.85225284e-01
4.44981009e-01 -4.01854753e-01 -2.64581263e-01 4.05640244e-01
-5.52963734e-01 6.54264390e-01 1.80521592e-01 1.39674276e-01
-7.76893020e-01 -3.85859787e-01 2.43998781e-01 4.52054113e-01
-1.99430585e-01 9.13659811e-01 -7.97585309e-01 -2.94717044e-01
-3.33775073e-01 4.84620869e-01 1.14145350e+00 1.43588722e+00
3.26777309e-01 -1.20590374e-01 -5.45543313e-01 7.71614969e-01
4.00582373e-01 3.57304484e-01 7.28342414e-01 -6.76900208e-01
6.41154647e-01 8.40622962e-01 6.57808959e-01 -8.94451261e-01
-4.65446740e-01 -3.57417852e-01 -1.54709637e-01 -2.81954885e-01
-2.79502988e-01 -2.17784613e-01 -4.80950773e-01 1.47850943e+00
3.56794626e-01 -2.44224638e-01 4.35235053e-01 8.08830142e-01
1.51404238e+00 5.09473860e-01 2.41356075e-01 -2.69928217e-01
1.59757590e+00 -6.59203172e-01 -1.20608950e+00 -1.27946481e-01
3.88055831e-01 -8.65877628e-01 7.99610734e-01 4.55167592e-01
-1.14176571e+00 -2.84290373e-01 -5.31856239e-01 -8.38883892e-02
-5.85075498e-01 3.11738968e-01 8.95989835e-01 4.64997321e-01
-1.12320685e+00 4.18555707e-01 1.63801134e-01 -4.17612970e-01
-3.45081508e-01 7.39115894e-01 -2.86165535e-01 6.22862220e-01
-1.83346617e+00 8.41848135e-01 8.64092335e-02 6.49144575e-02
-8.23240280e-02 -2.15238884e-01 -4.35634673e-01 2.57597864e-01
4.04907584e-01 -7.29782403e-01 1.80527782e+00 -2.62690037e-01
-1.09911311e+00 5.16605020e-01 -1.89537793e-01 -2.91119814e-01
3.15384448e-01 -4.93589044e-01 -1.13405538e+00 -2.05943331e-01
4.79469925e-01 9.94225442e-02 3.61269444e-01 -1.00118804e+00
-1.38664865e+00 -4.68199819e-01 -6.37110248e-02 4.23521161e-01
1.99949071e-02 4.58107561e-01 -4.19915020e-01 -3.31257671e-01
-8.91943723e-02 -4.71762985e-01 -1.76012814e-01 -7.66157329e-01
-4.87223625e-01 -6.22605443e-01 5.16603470e-01 -7.78090417e-01
1.72247469e+00 -1.58335972e+00 -3.34441960e-01 3.22682768e-01
4.26448077e-01 4.57945615e-01 1.08112618e-02 4.91626501e-01
-7.55411908e-02 4.56755370e-01 6.24306500e-01 4.27569821e-02
1.82015002e-01 -7.69462436e-02 -5.07579207e-01 -5.91707528e-01
4.02505726e-01 8.43566537e-01 -9.17384744e-01 -3.97670507e-01
-2.90839404e-01 1.15169451e-01 -2.23711774e-01 3.97837996e-01
4.60157245e-02 -4.72145453e-02 -8.18960786e-01 6.77387953e-01
3.69095713e-01 -1.35318324e-01 2.86414206e-01 -2.52177328e-01
-7.16379136e-02 8.78518343e-01 -1.66312122e+00 7.01820254e-01
-1.58270046e-01 3.26726794e-01 2.54437476e-01 -4.93885607e-01
1.27397132e+00 5.21978438e-01 1.90413028e-01 -9.52088416e-01
1.80482328e-01 3.72565478e-01 -5.03833175e-01 -8.16307187e-01
1.36993825e+00 4.62633371e-01 -3.07388484e-01 1.60430476e-01
-1.75387636e-01 5.16581893e-01 2.52423078e-01 4.38602984e-01
1.39511383e+00 -4.90269333e-01 5.26187897e-01 1.63002640e-01
5.48156083e-01 5.01135737e-02 8.43383551e-01 7.98476815e-01
-3.07183355e-01 1.40603289e-01 3.24846119e-01 2.18476169e-02
-6.30681753e-01 -9.37701046e-01 -4.28086109e-02 1.21785831e+00
2.23582581e-01 -4.54670221e-01 -3.33135486e-01 -7.61496782e-01
2.12219700e-01 9.13266599e-01 -2.69870251e-01 -1.07877769e-01
-3.89810562e-01 -3.92411500e-01 3.12716872e-01 6.65626407e-01
1.93915546e-01 -1.35714924e+00 -1.94827482e-01 5.18686533e-01
-5.16190946e-01 -1.19463122e+00 -2.36326739e-01 3.57595950e-01
-7.52270103e-01 -1.00443733e+00 -7.70819262e-02 -7.53539681e-01
3.05212349e-01 2.72020489e-01 1.33393860e+00 -9.24144387e-02
-3.44099343e-01 3.08905959e-01 -5.16751826e-01 -4.16967273e-01
-1.31450281e-01 1.12298951e-01 3.62092793e-01 -4.40198362e-01
8.45924139e-01 -3.07075411e-01 -4.17208672e-01 5.49422860e-01
-3.23771328e-01 -5.52288353e-01 6.69064879e-01 5.08713901e-01
3.72337967e-01 1.94517106e-01 1.27379441e+00 -1.20208049e+00
1.73042619e+00 -8.38127494e-01 -3.29160035e-01 5.41107416e-01
-1.18451297e+00 3.08311224e-01 4.07308042e-01 -1.04272649e-01
-1.01364648e+00 -3.00807327e-01 -9.02227387e-02 1.28279418e-01
-4.99508008e-02 4.84161496e-01 -2.90588945e-01 5.46643496e-01
8.39124858e-01 -3.19651306e-01 -6.77994668e-01 -8.43747795e-01
3.59937459e-01 1.24093699e+00 7.02077985e-01 -4.90428180e-01
5.13674915e-01 -2.45914668e-01 -5.53936183e-01 -3.29283416e-01
-4.03773636e-01 -9.52849805e-01 -1.58436432e-01 1.95546001e-02
3.43963444e-01 -9.02427554e-01 -8.28497767e-01 -3.91910851e-01
-1.20151246e+00 4.59999025e-01 -5.04556835e-01 1.57824829e-01
-2.11631283e-01 7.25791529e-02 -5.94925761e-01 -1.12359858e+00
-9.03183579e-01 -8.03044260e-01 1.08691168e+00 6.76904500e-01
-6.16270006e-01 -4.43968058e-01 -6.25027195e-02 4.57424372e-01
5.15716732e-01 -2.10831508e-01 6.34931743e-01 -1.12812901e+00
-4.73861247e-01 -3.79013866e-01 -3.13622147e-01 -3.07891279e-01
1.47872388e-01 1.49292886e-01 -9.33462799e-01 1.48012295e-01
-4.91921693e-01 1.66406199e-01 3.15399349e-01 -5.84958643e-02
6.16045475e-01 -5.40713608e-01 -4.90202338e-01 2.77770031e-02
9.90901411e-01 5.15034020e-01 6.92847073e-01 6.33297443e-01
1.31996840e-01 9.27007854e-01 1.36649299e+00 5.86881936e-01
5.66986740e-01 9.93289709e-01 7.04989908e-03 9.07953009e-02
1.90087691e-01 -2.75162250e-01 2.37897024e-01 6.10366344e-01
-3.45872492e-02 -5.24468243e-01 -6.32561028e-01 5.00873089e-01
-2.26626825e+00 -1.17551541e+00 -4.90526766e-01 2.08513546e+00
9.48318601e-01 1.13175020e-01 2.11226314e-01 1.43229350e-01
1.04258168e+00 -6.19236410e-01 -3.17093909e-01 -6.55387998e-01
2.88816951e-02 1.65203243e-01 4.22578812e-01 5.23747027e-01
-7.40328252e-01 9.78896618e-01 6.62226343e+00 2.96298683e-01
-8.38603199e-01 -1.62297741e-01 1.06589869e-01 2.00059280e-01
-4.28947955e-01 -2.26099521e-01 -1.51460409e+00 3.41836065e-01
9.36385870e-01 -7.37861395e-01 1.15723863e-01 1.31590021e+00
2.80508429e-01 2.56405562e-01 -9.80272830e-01 1.00989664e+00
-4.00207251e-01 -1.03210473e+00 6.10629208e-02 -3.15056648e-03
2.00444818e-01 -1.99683100e-01 -1.45305082e-01 7.85823524e-01
6.98866665e-01 -8.12967718e-01 5.35384297e-01 7.92905033e-01
7.68414378e-01 -6.33762598e-01 8.85446787e-01 1.19084358e-01
-1.32749820e+00 -3.40285689e-01 -4.61254805e-01 3.56954336e-01
4.58893836e-01 5.31346560e-01 -1.66140234e+00 4.33285445e-01
8.60879004e-01 -1.69756085e-01 -5.03125489e-01 1.12488210e+00
-2.09575132e-01 1.23146154e-01 -3.72251093e-01 8.33328366e-02
-4.06329930e-01 1.57458082e-01 8.23404849e-01 1.30499291e+00
3.52823824e-01 2.80279845e-01 -1.11122638e-01 7.48430610e-01
-1.90202743e-01 3.77337784e-01 -5.21041214e-01 -1.88198343e-01
1.20039141e+00 1.61289728e+00 -6.37837127e-02 -2.74197996e-01
-2.96572521e-02 7.95211196e-01 2.87067741e-01 2.60225534e-01
-6.08816087e-01 -8.30758512e-01 6.12296879e-01 2.68577307e-01
1.53586775e-01 1.39058784e-01 1.73187256e-02 -8.16930830e-01
1.62293032e-01 -9.47968125e-01 5.69676161e-01 -7.69461870e-01
-1.25437295e+00 9.66418982e-01 -3.03661704e-01 -1.05395651e+00
-6.26688540e-01 -4.10461932e-01 -7.57552981e-01 1.02756584e+00
-1.66813564e+00 -4.16924804e-01 -7.92144299e-01 3.69765908e-01
4.11040962e-01 -4.42481756e-01 9.94538486e-01 7.86521673e-01
-8.07228327e-01 8.94255877e-01 -1.72842890e-01 -1.29678890e-01
8.60471666e-01 -1.66837096e+00 7.41222978e-01 4.55173701e-01
-1.44952685e-01 1.09159839e+00 8.46845508e-01 -8.49694908e-01
-1.14379001e+00 -1.03145337e+00 1.42722225e+00 -4.47858632e-01
4.24606264e-01 -2.01601745e-03 -7.73652196e-01 4.59554285e-01
-3.85011107e-01 -3.16066742e-01 8.01273823e-01 4.07975048e-01
3.40192541e-02 -2.06044331e-01 -1.51965773e+00 4.13880914e-01
1.01124465e+00 -4.41587448e-01 -8.54437351e-01 1.32298976e-01
8.88930202e-01 -7.69762844e-02 -1.00774455e+00 5.07588685e-01
4.61073965e-01 -6.44947290e-01 9.15269196e-01 -9.01004493e-01
-2.17109308e-01 -3.66417438e-01 -2.12404504e-01 -9.77459252e-01
-5.35509944e-01 -1.00364196e+00 -5.44242799e-01 1.78948772e+00
9.37402368e-01 -6.44438028e-01 5.64234853e-01 1.29904270e+00
1.12662591e-01 -3.71485740e-01 -4.24157768e-01 -4.68072593e-01
-9.27775502e-01 -6.38568103e-02 1.07622492e+00 9.04037237e-01
1.20681994e-01 9.44817126e-01 3.43557522e-02 3.44751179e-01
2.12132543e-01 -3.08832139e-01 7.65178502e-01 -1.89892995e+00
-1.91271454e-01 -3.60785931e-01 -2.59417981e-01 -9.64328647e-01
-3.57122183e-01 -5.65907240e-01 -2.38129824e-01 -1.83993995e+00
-2.56678313e-01 -8.98410201e-01 -3.43986243e-01 2.01081559e-01
-3.13992321e-01 -4.28124726e-01 -1.78474575e-01 3.28430206e-01
-7.84017026e-01 3.91277850e-01 5.47161102e-01 3.00090551e-01
-7.46898651e-01 2.99755871e-01 -1.20157456e+00 3.33661079e-01
9.34958816e-01 -3.95838469e-01 -4.03743625e-01 -1.19449012e-01
5.88494241e-01 2.35482186e-01 -1.00807205e-01 -5.13154447e-01
4.77067828e-01 -2.72974938e-01 1.82962362e-02 -6.85325980e-01
1.82812214e-01 -8.35001469e-01 1.60436973e-01 -2.52939135e-01
-5.40930986e-01 6.11948073e-01 -6.16840273e-02 6.14450395e-01
-1.98910698e-01 -5.72056651e-01 8.54143873e-02 2.20654607e-02
-1.13549900e+00 8.74881223e-02 -1.60811797e-01 7.02332333e-02
7.89901376e-01 -1.75583869e-01 -6.60140872e-01 -6.16356015e-01
-9.00996566e-01 7.81495988e-01 -1.04047611e-01 7.63795316e-01
6.55624986e-01 -1.29780972e+00 -5.47826171e-01 -4.43670750e-02
2.77587473e-01 -3.92797366e-02 -2.61541367e-01 4.05935347e-01
-1.54232979e-01 7.48225808e-01 1.17130116e-01 -1.36538357e-01
-1.61587071e+00 3.39221433e-02 3.11079681e-01 -4.83704776e-01
-3.16599309e-01 8.78751695e-01 -2.81282425e-01 -6.23007953e-01
5.35949588e-01 7.40518868e-02 -1.17470598e+00 2.65722841e-01
8.32328379e-01 6.05922520e-01 3.87977362e-01 -5.17648935e-01
-4.41830665e-01 -2.37536971e-02 -5.01096904e-01 -1.00215279e-01
1.22943413e+00 -4.59156960e-01 1.55644745e-01 1.14726670e-01
5.64919531e-01 3.69049996e-01 -2.24910766e-01 -4.35315967e-01
7.95887351e-01 -4.45303202e-01 9.67959315e-02 -1.04103673e+00
-7.31821537e-01 2.70729363e-01 8.04766774e-01 5.87806106e-01
9.42653894e-01 -5.73947243e-02 8.26429307e-01 7.29039013e-01
5.75942516e-01 -1.66483200e+00 -2.60709941e-01 4.58432466e-01
1.15349138e+00 -1.19995630e+00 -2.12942451e-01 -5.47855556e-01
-1.02737820e+00 1.01977313e+00 8.69555116e-01 3.97619307e-01
3.61511171e-01 2.22881526e-01 1.59004942e-01 -5.67025542e-01
-1.22003937e+00 -4.75910544e-01 2.70651877e-01 8.46435070e-01
7.63382196e-01 3.02348733e-01 -6.26695573e-01 1.27659464e+00
-4.48951840e-01 -8.33036436e-04 4.34022903e-01 7.54575729e-01
-9.58254516e-01 -1.35442221e+00 -1.74546674e-01 5.67353010e-01
-4.53753233e-01 -1.03322983e-01 -8.22490335e-01 3.76803190e-01
-1.84062377e-01 1.25884783e+00 -3.06120545e-01 -7.82220900e-01
1.14106488e+00 6.50517866e-02 -2.84434766e-01 -8.79688382e-01
-8.05004239e-01 -3.62471044e-01 9.59132552e-01 -5.48658431e-01
2.36423209e-01 -5.36202013e-01 -1.62559652e+00 -2.15983048e-01
-9.99926507e-01 8.10978413e-01 7.77229369e-01 3.37674320e-01
8.76489043e-01 4.75872755e-01 9.21000063e-01 -4.42764834e-02
-9.61451888e-01 -1.02343535e+00 -6.49525404e-01 4.29033577e-01
-8.49704742e-02 -6.94768906e-01 -3.30790550e-01 -4.61438805e-01] | [9.478676795959473, 8.869369506835938] |
c09280ce-ae55-4c2d-ac88-472b09adc4c4 | spatio-temporal-crop-aggregation-for-video | 2211.17042 | null | https://arxiv.org/abs/2211.17042v2 | https://arxiv.org/pdf/2211.17042v2.pdf | Spatio-Temporal Crop Aggregation for Video Representation Learning | We propose Spatio-temporal Crop Aggregation for video representation LEarning (SCALE), a novel method that enjoys high scalability at both training and inference time. Our model builds long-range video features by learning from sets of video clip-level features extracted with a pre-trained backbone. To train the model, we propose a self-supervised objective consisting of masked clip feature prediction. We apply sparsity to both the input, by extracting a random set of video clips, and to the loss function, by only reconstructing the sparse inputs. Moreover, we use dimensionality reduction by working in the latent space of a pre-trained backbone applied to single video clips. These techniques make our method not only extremely efficient to train but also highly effective in transfer learning. We demonstrate that our video representation yields state-of-the-art performance with linear, non-linear, and KNN probing on common action classification and video understanding datasets. | ['Paolo Favaro', 'Simon Jenni', 'Sepehr Sameni'] | 2022-11-30 | null | null | null | null | ['action-classification', 'video-understanding'] | ['computer-vision', 'computer-vision'] | [ 5.12894511e-01 -5.66118099e-02 -4.44440931e-01 -3.39019984e-01
-1.09505785e+00 -4.48557615e-01 5.17585754e-01 -2.40270883e-01
-1.96530163e-01 4.48682278e-01 5.09909868e-01 2.50558615e-01
-4.73010167e-02 -6.04142308e-01 -1.34329557e+00 -6.23663306e-01
-4.93877918e-01 2.63497233e-01 5.16208448e-03 1.80394873e-01
9.25596505e-02 2.09136188e-01 -1.83391118e+00 1.01717126e+00
4.48716372e-01 1.31098986e+00 2.14573234e-01 6.83449388e-01
2.33750209e-01 1.60232210e+00 -4.60089184e-02 -1.07807547e-01
4.02747869e-01 -4.69407350e-01 -6.49565279e-01 5.27523100e-01
6.90687299e-01 -6.39082849e-01 -6.44148767e-01 6.31975412e-01
8.36931169e-02 1.76350787e-01 5.08514345e-01 -1.30554247e+00
-5.90564847e-01 5.68898737e-01 -4.73877400e-01 4.21982929e-02
6.75343156e-01 1.54693812e-01 1.20647728e+00 -9.07766581e-01
8.66342366e-01 1.04898012e+00 7.03044593e-01 5.77551782e-01
-1.28113437e+00 -5.26164114e-01 3.35561365e-01 5.17749310e-01
-1.28195667e+00 -7.37588882e-01 6.74031138e-01 -6.36644065e-01
1.08799398e+00 8.78891908e-03 7.12801933e-01 1.21648932e+00
-1.20427512e-01 1.20199513e+00 6.88460886e-01 -3.53122830e-01
3.01177263e-01 -1.57814428e-01 -9.79269296e-02 9.50201571e-01
-4.63377506e-01 -8.31552818e-02 -1.03037250e+00 -1.99399844e-01
7.15026259e-01 3.42332363e-01 -3.72944653e-01 -7.17845142e-01
-1.01478386e+00 8.08684647e-01 1.94471374e-01 1.80782422e-01
-5.84633708e-01 4.67858106e-01 4.64644104e-01 5.78485072e-01
4.06449854e-01 8.86640102e-02 -6.73523426e-01 -4.13251817e-01
-1.22535837e+00 2.60255754e-01 5.50613761e-01 9.55561876e-01
8.33465934e-01 -6.08680286e-02 -1.60233527e-01 5.09767592e-01
-4.23000604e-02 3.90661180e-01 5.58471322e-01 -1.34464800e+00
5.15005350e-01 2.78872013e-01 -3.40101421e-02 -9.20876682e-01
-9.31092650e-02 -8.51656273e-02 -7.16437459e-01 6.09525405e-02
2.43337408e-01 1.94610775e-01 -9.51047182e-01 1.98827076e+00
3.62892039e-02 8.10010135e-01 3.76048335e-03 6.29688442e-01
2.86464542e-01 8.91993046e-01 -1.37131009e-02 -3.83496433e-01
9.38235521e-01 -1.33853006e+00 -3.88276994e-01 3.75988446e-02
7.76458979e-01 -2.26182938e-01 9.50078964e-01 5.82174957e-01
-1.17878556e+00 -7.09606111e-01 -8.39869618e-01 -2.37333283e-01
-2.61603408e-02 2.98782796e-01 5.32964706e-01 1.56380355e-01
-1.07250631e+00 8.77002835e-01 -1.19285119e+00 -1.66160971e-01
6.96777999e-01 2.25044101e-01 -7.99881637e-01 -5.97383142e-01
-6.99129999e-01 4.26697195e-01 1.09023303e-01 -3.61119330e-01
-1.31484580e+00 -8.90832365e-01 -1.00670767e+00 1.53377682e-01
5.19294918e-01 -5.67887485e-01 8.59408498e-01 -1.34884214e+00
-1.45649242e+00 7.50030994e-01 -3.39085609e-01 -7.08969712e-01
3.62600356e-01 -7.51684070e-01 -2.66342983e-02 6.77520335e-01
1.55297011e-01 8.61456931e-01 1.37036324e+00 -8.76088381e-01
-5.53562820e-01 -2.67580390e-01 1.20769426e-01 4.78461906e-02
-5.26050091e-01 -1.03910044e-01 -5.99877059e-01 -7.35842943e-01
-9.36977640e-02 -1.01236069e+00 -1.49837777e-01 3.17176521e-01
8.01864266e-02 -5.91368824e-02 8.42632174e-01 -1.01957011e+00
1.09178174e+00 -2.41539216e+00 6.71485305e-01 7.23391250e-02
6.03554435e-02 4.06178907e-02 -5.99770367e-01 4.00607437e-01
-7.13024214e-02 -2.17538357e-01 -2.58874089e-01 -5.03671944e-01
-1.58759236e-01 2.23416120e-01 -5.88099003e-01 3.34493905e-01
3.75393540e-01 8.60279381e-01 -9.09385264e-01 -4.17165846e-01
2.86944270e-01 4.30040389e-01 -1.05210233e+00 6.54784322e-01
-4.23086196e-01 3.28921378e-01 -3.97042155e-01 6.26963437e-01
3.69356245e-01 -4.34559822e-01 3.40482771e-01 -2.64574647e-01
1.84018329e-01 5.30709997e-02 -1.01098406e+00 2.35781932e+00
-4.48621452e-01 4.96557295e-01 -1.38288155e-01 -1.37056839e+00
3.31421822e-01 4.03801382e-01 8.72637689e-01 -4.48451787e-01
-3.39654416e-01 -1.47385091e-01 -6.87366784e-01 -8.77570271e-01
1.49741322e-01 9.07018483e-02 -7.68412044e-03 5.07153451e-01
6.44147098e-01 4.22942162e-01 2.12444514e-01 2.83719510e-01
1.51379168e+00 6.06699049e-01 1.84444577e-01 2.19522819e-01
3.78446162e-01 -2.64627278e-01 6.07139111e-01 7.14509368e-01
2.07164943e-01 6.34750187e-01 6.69624090e-01 -6.58140540e-01
-1.04717612e+00 -9.87135828e-01 2.10563406e-01 1.39191210e+00
-3.16092610e-01 -8.02829623e-01 -6.21548951e-01 -9.80448723e-01
9.89960060e-02 3.87507111e-01 -8.30276132e-01 -1.31970033e-01
-6.54204726e-01 -1.88462123e-01 2.53729641e-01 7.99378097e-01
1.97778448e-01 -1.02667713e+00 -5.69133461e-01 2.34266996e-01
-2.43928358e-01 -1.41950190e+00 -4.12372619e-01 7.38105699e-02
-1.01593220e+00 -9.98980284e-01 -5.25965393e-01 -6.03550434e-01
6.10892415e-01 3.49395543e-01 1.12092888e+00 1.49944648e-01
-2.19740257e-01 6.24759972e-01 -6.51526392e-01 2.71773607e-01
-1.25969633e-01 -3.88193168e-02 4.04666550e-03 3.24350387e-01
3.51241231e-01 -8.88541281e-01 -3.18738222e-01 4.53053713e-02
-9.77851927e-01 1.06298551e-01 5.67653477e-01 8.23210955e-01
7.67495275e-01 -2.18295887e-01 3.78734082e-01 -6.81351423e-01
-1.66740343e-01 -6.22689784e-01 -3.69053572e-01 3.90102506e-01
-5.38043566e-02 1.29211247e-01 7.14326441e-01 -4.58966523e-01
-7.48388946e-01 5.26040375e-01 1.49617374e-01 -1.08643246e+00
-2.16336444e-01 4.41988677e-01 -1.38699010e-01 3.99521142e-02
5.26714504e-01 3.38406861e-01 5.37666716e-02 -5.72154522e-01
6.54029667e-01 3.83560270e-01 5.16610742e-01 -5.29020071e-01
8.31195951e-01 6.69207275e-01 -9.93009806e-02 -8.93180668e-01
-1.03181207e+00 -4.51641053e-01 -9.20758426e-01 -6.74169734e-02
1.00059497e+00 -1.38647795e+00 -2.98278064e-01 3.57969850e-01
-7.87466884e-01 -4.69020933e-01 -5.21038532e-01 4.84861881e-01
-9.42357898e-01 3.49992901e-01 -6.82175040e-01 -4.42141742e-01
1.69479512e-02 -8.80761564e-01 1.23426366e+00 -2.97118247e-01
-1.24830268e-01 -7.26242721e-01 1.11200966e-01 4.29977596e-01
1.86915100e-01 3.52277279e-01 7.52700627e-01 -3.41503471e-01
-1.05810750e+00 -5.74561395e-02 -1.24369755e-01 6.30569696e-01
5.05454578e-02 -1.68406591e-01 -1.00148702e+00 -5.63619077e-01
-1.27779931e-01 -1.01946068e+00 1.29714537e+00 3.65763873e-01
1.50035262e+00 -4.17435884e-01 -3.40575278e-01 9.51170564e-01
1.41044140e+00 -2.05040857e-01 6.19688988e-01 1.13750890e-01
9.18309212e-01 4.78481263e-01 4.41531688e-01 6.54982388e-01
3.10896993e-01 7.00377285e-01 1.36112437e-01 3.04212987e-01
-3.25492509e-02 -5.09753227e-01 7.36524522e-01 6.86133027e-01
-2.31799439e-01 6.18769415e-02 -4.36666012e-01 5.23068905e-01
-2.17320466e+00 -1.38274491e+00 6.26153827e-01 2.01533961e+00
7.02920020e-01 2.44811792e-02 2.10492045e-01 3.95724317e-03
2.64030546e-01 3.98014605e-01 -5.63277602e-01 1.34194523e-01
4.97353412e-02 3.97889405e-01 1.75076574e-01 2.14886293e-01
-1.42653775e+00 1.00360680e+00 6.03436089e+00 6.84032619e-01
-1.05541027e+00 1.02347262e-01 4.57359731e-01 -4.53547418e-01
-2.47984484e-01 1.91148631e-02 -3.99660259e-01 4.16343689e-01
1.02211070e+00 5.03605530e-02 6.98533833e-01 9.41790521e-01
-2.87000206e-03 1.68987229e-01 -1.57510257e+00 1.08909488e+00
4.62425321e-01 -1.61295736e+00 2.95191407e-01 -5.59637621e-02
8.00397635e-01 1.04952656e-01 -5.97658381e-02 3.88098925e-01
-2.96200742e-03 -1.04694819e+00 8.01255763e-01 5.90127349e-01
8.60009789e-01 -4.64732021e-01 2.50415295e-01 2.14156076e-01
-1.28907013e+00 -4.73701239e-01 -3.99625480e-01 -2.00852513e-01
2.12311402e-01 3.08952004e-01 -2.34645098e-01 4.78970528e-01
6.96166754e-01 1.43665469e+00 -3.90921712e-01 7.97268867e-01
-1.67378619e-01 6.78275347e-01 -3.14358771e-01 6.68340921e-01
2.17537418e-01 -5.93063198e-02 2.17541024e-01 1.18280423e+00
3.81468236e-01 1.32228345e-01 5.81972897e-01 4.22019720e-01
-2.83463299e-01 -8.94914865e-02 -8.56682241e-01 -1.24953479e-01
2.63855070e-01 1.00881267e+00 -3.43464434e-01 -5.04734755e-01
-7.29474247e-01 1.19772542e+00 7.30978191e-01 3.91536176e-01
-7.68286586e-01 -2.59195510e-02 7.73547113e-01 1.43310666e-01
8.94916713e-01 -2.14241356e-01 2.92492390e-01 -1.64767861e+00
3.14814627e-01 -9.59230185e-01 5.05680263e-01 -5.78232944e-01
-1.14801717e+00 3.94203007e-01 3.84765305e-02 -1.48326313e+00
-7.30923057e-01 -5.44971168e-01 -3.20036292e-01 1.83499739e-01
-1.55683696e+00 -1.39205205e+00 -2.56635904e-01 8.40507507e-01
6.86684251e-01 -2.84749895e-01 9.14557815e-01 3.17263275e-01
-3.00424725e-01 5.97052515e-01 1.54029533e-01 1.85018361e-01
6.17723584e-01 -9.44711506e-01 1.89283103e-01 6.71345294e-01
4.75631624e-01 2.04973564e-01 3.48599166e-01 -2.84929544e-01
-1.57595098e+00 -1.15818822e+00 5.51498592e-01 -4.19300437e-01
7.35507846e-01 -4.18868035e-01 -8.82206082e-01 1.09523523e+00
-8.39555040e-02 3.59187394e-01 8.30206752e-01 7.72210211e-02
-7.59008288e-01 -1.57011911e-01 -8.43069434e-01 1.48249969e-01
1.38961864e+00 -9.70599949e-01 -6.26080155e-01 4.51805860e-01
7.05828249e-01 -2.71047354e-01 -8.36073816e-01 2.89636552e-01
8.48361015e-01 -9.86404061e-01 1.02175498e+00 -1.06539524e+00
8.51813436e-01 -1.00677468e-01 -5.20739794e-01 -1.10468924e+00
-4.44598615e-01 -5.65881073e-01 -6.55862808e-01 9.58327115e-01
2.62100548e-01 3.93484645e-02 9.66768682e-01 4.18847889e-01
1.83708277e-02 -8.28468919e-01 -8.50738883e-01 -6.91258907e-01
-2.24437386e-01 -4.91342694e-01 2.61696815e-01 8.83687675e-01
1.64648853e-02 1.16876975e-01 -8.80458236e-01 1.48888320e-01
7.44324744e-01 3.86764884e-01 9.09999073e-01 -9.90386367e-01
-7.41379797e-01 9.58529785e-02 -6.76163971e-01 -1.25963163e+00
5.59358180e-01 -7.30966866e-01 -1.56638101e-01 -1.13713145e+00
4.23043042e-01 -4.37723026e-02 -5.80561101e-01 6.56559169e-01
3.94765250e-02 3.08135122e-01 4.49267596e-01 3.17181051e-01
-1.09705889e+00 5.34845769e-01 8.61871123e-01 -1.57692492e-01
-9.29907337e-02 -1.65704086e-01 -3.89154583e-01 7.47946560e-01
3.46139163e-01 -5.37537217e-01 -5.39018452e-01 -6.42790854e-01
8.13939124e-02 3.03741157e-01 4.22334731e-01 -1.24534404e+00
1.09288897e-02 -1.13833748e-01 4.87265021e-01 -3.31069559e-01
6.46253526e-01 -7.52266824e-01 -4.96453047e-02 2.06940174e-01
-6.73239648e-01 -1.69643134e-01 -1.70110583e-01 7.35534191e-01
-3.43568116e-01 1.49871586e-02 6.13660336e-01 -2.93797582e-01
-8.51232648e-01 5.22565007e-01 -1.42445385e-01 7.89442882e-02
1.06531489e+00 -6.43034503e-02 -3.06158699e-02 -5.43212652e-01
-9.32594359e-01 2.04560682e-01 6.05585515e-01 4.63905990e-01
7.32835531e-01 -1.45982325e+00 -6.37311399e-01 4.01234895e-01
1.93724915e-01 -2.75464773e-01 4.51408952e-01 6.34162962e-01
-1.50226697e-01 2.69480765e-01 -5.17502189e-01 -6.72936916e-01
-1.18860292e+00 8.10271740e-01 -7.99264461e-02 -3.84367675e-01
-8.42114329e-01 9.40120101e-01 1.78085536e-01 -1.27932681e-02
5.39208293e-01 -3.11713099e-01 -7.78305903e-02 1.89198285e-01
8.51887047e-01 1.93057463e-01 -2.43771344e-01 -6.72633767e-01
-2.52157956e-01 7.50145376e-01 -1.01680703e-01 1.11497223e-01
1.72923136e+00 5.50404713e-02 8.23911056e-02 4.44541484e-01
1.67888474e+00 -2.37206087e-01 -1.84255505e+00 -4.13391650e-01
-1.14827782e-01 -5.76642990e-01 -9.86874700e-02 -3.33449662e-01
-1.09316742e+00 7.23305345e-01 4.08412099e-01 -8.95465240e-02
1.22625530e+00 7.53231123e-02 9.46229994e-01 6.95851386e-01
4.86509025e-01 -1.04088020e+00 4.95735615e-01 4.37828511e-01
8.32438171e-01 -1.19332719e+00 7.17187077e-02 -1.78839386e-01
-5.83739579e-01 1.04820514e+00 4.36873823e-01 -4.91713583e-01
8.32817435e-01 2.47557715e-01 -1.42498702e-01 -3.44439223e-03
-1.22922313e+00 -5.98639660e-02 2.74058133e-01 4.91761476e-01
1.74982294e-01 -1.17608748e-01 2.91744977e-01 3.22705984e-01
1.44724607e-01 5.46347618e-01 2.62891680e-01 1.18821192e+00
-4.02906805e-01 -1.10527563e+00 1.21679790e-01 6.19018376e-01
-3.98876727e-01 -4.27720100e-02 -5.20874746e-02 4.63436633e-01
1.61775604e-01 5.46653390e-01 1.58963606e-01 -5.66808879e-01
1.35985285e-01 3.08617204e-01 6.25562608e-01 -5.45759618e-01
-6.94611073e-02 -1.03370212e-01 1.70721970e-02 -1.27768254e+00
-7.80617476e-01 -8.26320827e-01 -9.02832031e-01 5.82714528e-02
5.46601526e-02 1.35341957e-01 3.02448958e-01 1.12195086e+00
5.78681886e-01 1.97352767e-01 9.02327180e-01 -1.22073567e+00
-6.60283804e-01 -8.09423625e-01 -5.26550233e-01 7.58755028e-01
4.65752959e-01 -6.01317227e-01 -2.55632252e-01 6.54614866e-01] | [8.744112968444824, 0.7508679032325745] |
df4d0208-1387-4f39-995f-970dfa0a2206 | two-shot-video-object-segmentation | 2303.12078 | null | https://arxiv.org/abs/2303.12078v1 | https://arxiv.org/pdf/2303.12078v1.pdf | Two-shot Video Object Segmentation | Previous works on video object segmentation (VOS) are trained on densely annotated videos. Nevertheless, acquiring annotations in pixel level is expensive and time-consuming. In this work, we demonstrate the feasibility of training a satisfactory VOS model on sparsely annotated videos-we merely require two labeled frames per training video while the performance is sustained. We term this novel training paradigm as two-shot video object segmentation, or two-shot VOS for short. The underlying idea is to generate pseudo labels for unlabeled frames during training and to optimize the model on the combination of labeled and pseudo-labeled data. Our approach is extremely simple and can be applied to a majority of existing frameworks. We first pre-train a VOS model on sparsely annotated videos in a semi-supervised manner, with the first frame always being a labeled one. Then, we adopt the pre-trained VOS model to generate pseudo labels for all unlabeled frames, which are subsequently stored in a pseudo-label bank. Finally, we retrain a VOS model on both labeled and pseudo-labeled data without any restrictions on the first frame. For the first time, we present a general way to train VOS models on two-shot VOS datasets. By using 7.3% and 2.9% labeled data of YouTube-VOS and DAVIS benchmarks, our approach achieves comparable results in contrast to the counterparts trained on fully labeled set. Code and models are available at https://github.com/yk-pku/Two-shot-Video-Object-Segmentation. | ['Yan Lu', 'Ping Wang', 'Chenbin Zhang', 'Jinglu Wang', 'Fangyun Wei', 'Xiao Li', 'Kun Yan'] | 2023-03-21 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Yan_Two-Shot_Video_Object_Segmentation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yan_Two-Shot_Video_Object_Segmentation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-object-segmentation', 'video-semantic-segmentation', 'pseudo-label'] | ['computer-vision', 'computer-vision', 'miscellaneous'] | [ 2.79776394e-01 6.88246489e-02 -5.14542699e-01 -5.22603393e-01
-8.98593664e-01 -5.65079927e-01 1.32977247e-01 -3.04531842e-01
-4.79476869e-01 5.95836639e-01 -1.47633761e-01 -1.11613683e-02
6.16188288e-01 -4.25332397e-01 -9.79932606e-01 -5.65737963e-01
3.58040422e-01 4.11920398e-01 6.65000558e-01 2.03053489e-01
8.08705688e-02 1.66048616e-01 -1.55499125e+00 2.79358268e-01
5.48819125e-01 1.05024481e+00 3.97176743e-01 7.54675448e-01
-2.73025513e-01 8.54173839e-01 -4.12169725e-01 -1.42751694e-01
6.50481403e-01 -5.89819372e-01 -1.12493777e+00 7.94158876e-01
4.70064819e-01 -5.99536180e-01 -3.33743036e-01 9.14896190e-01
1.15776710e-01 2.55640209e-01 5.77449083e-01 -1.24658418e+00
-3.74160320e-01 3.11644733e-01 -5.79917431e-01 2.99406350e-01
2.85040379e-01 3.83455426e-01 9.14556086e-01 -9.54198241e-01
1.02636623e+00 7.35555172e-01 4.60515559e-01 9.44952428e-01
-1.09502470e+00 -4.04476523e-01 2.19692245e-01 1.37216181e-01
-1.32745981e+00 -6.39367044e-01 6.31602705e-01 -7.32790291e-01
5.87227702e-01 7.82503411e-02 9.24795270e-01 9.90161180e-01
-2.74719179e-01 1.26208639e+00 8.44286144e-01 -3.66551876e-01
3.45021337e-01 2.81210616e-02 2.98270524e-01 6.80804491e-01
-1.09172016e-01 -1.31759599e-01 -3.51287454e-01 1.90086916e-01
7.11684823e-01 2.74054974e-01 -3.18998307e-01 -4.47266698e-01
-1.13223886e+00 6.62238002e-01 1.19351335e-01 2.76084036e-01
-4.44012761e-01 2.20400766e-01 4.98459399e-01 1.48155794e-01
5.09490490e-01 1.23529598e-01 -6.01693928e-01 -2.28409633e-01
-1.55236888e+00 -3.25204581e-02 6.65742636e-01 1.24727440e+00
9.62552547e-01 1.88111559e-01 -2.95052141e-01 7.63868153e-01
3.80331278e-01 2.07414746e-01 4.75159317e-01 -1.33585167e+00
1.97292045e-01 2.99126327e-01 2.53979862e-01 -4.12709177e-01
5.02530001e-02 7.56804720e-02 -4.06486243e-01 -3.95108722e-02
6.02756619e-01 -3.75659436e-01 -1.52301562e+00 1.52810585e+00
5.56927502e-01 7.11109221e-01 8.27291608e-02 1.06106639e+00
1.01851690e+00 8.39899182e-01 2.56107420e-01 -3.85885209e-01
1.11433578e+00 -1.62385464e+00 -7.32564747e-01 -3.30770016e-01
5.99001288e-01 -6.92771137e-01 1.09033179e+00 5.49599677e-02
-1.22670233e+00 -6.47130251e-01 -8.52018118e-01 -8.50154161e-02
-1.85886070e-01 9.93770361e-02 2.90869057e-01 3.81532073e-01
-1.15228868e+00 5.53678572e-01 -1.03101683e+00 -3.05426240e-01
7.66467750e-01 3.26823324e-01 -2.28484049e-01 -2.99077213e-01
-8.55152011e-01 3.46481800e-01 5.38235068e-01 -6.17131740e-02
-1.51631296e+00 -6.16238952e-01 -1.06244779e+00 -6.64520264e-02
7.94656277e-01 -5.03416777e-01 1.51416779e+00 -1.36897922e+00
-1.46489394e+00 9.26242769e-01 -4.34826523e-01 -3.79784435e-01
4.45905417e-01 -1.02232829e-01 -1.13032199e-01 5.10489345e-01
4.13967282e-01 1.12263560e+00 1.02829909e+00 -1.36628580e+00
-5.90252340e-01 1.26129031e-01 1.60767704e-01 1.44054070e-01
-7.73724392e-02 4.78018373e-02 -1.06754720e+00 -5.11892498e-01
-1.33528123e-02 -9.81934369e-01 -3.80957842e-01 -1.18435830e-01
-5.58565557e-01 -1.14419222e-01 8.68082047e-01 -6.17402256e-01
9.53339696e-01 -2.32925344e+00 1.98340058e-01 -2.64414251e-01
2.24490594e-02 6.24409020e-01 -3.38962048e-01 5.68657517e-02
-7.79306665e-02 1.23120859e-01 -5.11477888e-01 -6.46497071e-01
-2.92187542e-01 3.81184876e-01 -6.23424910e-02 4.28867847e-01
3.96989107e-01 1.12193048e+00 -1.00440419e+00 -8.75602067e-01
3.13837826e-01 1.88530698e-01 -5.97854257e-01 5.96756577e-01
-5.41027844e-01 8.17459702e-01 -3.15912247e-01 9.34426308e-01
4.08866048e-01 -4.61442918e-01 -7.18187215e-03 -9.71665308e-02
-3.44699956e-02 -8.03968608e-02 -1.18694484e+00 1.81821382e+00
-9.53506902e-02 6.47778273e-01 -3.59973796e-02 -1.10761189e+00
4.68909144e-01 4.95918244e-01 9.10807729e-01 -3.12460303e-01
3.90511543e-01 2.68631041e-01 -5.24840832e-01 -7.78541088e-01
3.72223705e-01 -1.00374207e-01 8.84826332e-02 3.75278085e-01
6.38507545e-01 -9.97970626e-02 7.69720495e-01 2.88589388e-01
7.04444110e-01 5.60068548e-01 2.02210560e-01 -4.10842523e-02
5.46197832e-01 3.22111577e-01 7.31873155e-01 5.09757042e-01
-3.93699199e-01 9.06764090e-01 3.25112373e-01 -4.07193691e-01
-1.14113355e+00 -6.87147200e-01 -6.87270015e-02 1.02895558e+00
3.48339498e-01 -5.04312098e-01 -1.05094719e+00 -9.44060385e-01
-3.18980783e-01 4.71001327e-01 -4.51910943e-01 2.47798830e-01
-3.61202270e-01 -2.78674155e-01 2.70326763e-01 6.22535765e-01
3.39151382e-01 -1.29664505e+00 -5.98934710e-01 1.28072292e-01
-2.32792541e-01 -1.36166275e+00 -7.35261142e-01 1.26671016e-01
-8.00127864e-01 -1.23395360e+00 -9.82484818e-01 -9.88625824e-01
7.96714425e-01 4.60391641e-01 1.00187767e+00 2.72529751e-01
-2.15673834e-01 6.43426478e-01 -4.28496152e-01 -9.07923132e-02
-2.70038277e-01 4.95785549e-02 -1.74326926e-01 3.07947814e-01
3.17188323e-01 -2.02025652e-01 -6.00193560e-01 3.83479744e-01
-1.03438914e+00 2.38240331e-01 2.56772846e-01 7.97719598e-01
1.09850073e+00 -4.25101310e-01 3.60701740e-01 -1.19618464e+00
-3.47618043e-01 -6.27443016e-01 -5.80267370e-01 2.00243503e-01
-1.52866185e-01 -3.03301305e-01 4.11348999e-01 -4.73310888e-01
-8.84319603e-01 6.80930257e-01 -3.03128392e-01 -9.40940201e-01
-5.00405192e-01 3.45565617e-01 7.63952807e-02 -1.13714680e-01
2.88013250e-01 1.55046552e-01 -2.60486305e-01 -3.12363863e-01
4.23323333e-01 7.30929136e-01 4.78618354e-01 -4.07154828e-01
6.46595299e-01 5.74467599e-01 -4.86291736e-01 -8.58213186e-01
-1.21983433e+00 -1.01044631e+00 -7.93274105e-01 -4.45821524e-01
1.18735695e+00 -1.08190894e+00 -2.11202919e-01 4.32766020e-01
-1.07842016e+00 -9.05935585e-01 -7.58990288e-01 4.21272218e-01
-7.68531501e-01 4.53441054e-01 -7.53569961e-01 -6.32736266e-01
-9.64428708e-02 -1.26078308e+00 1.22525692e+00 4.05211061e-01
1.52896708e-02 -1.02649248e+00 -2.83903256e-02 6.41717374e-01
-1.63671263e-02 2.77431130e-01 4.41688113e-02 -6.74504399e-01
-6.95999026e-01 -8.15213174e-02 -1.68523341e-01 5.59685946e-01
3.87365222e-02 8.77699703e-02 -8.89818072e-01 -1.53267294e-01
-5.33578843e-02 -6.81122005e-01 9.14229929e-01 5.28280139e-01
1.06457949e+00 1.01960123e-01 -1.67278409e-01 7.82008111e-01
1.47363281e+00 2.61488914e-01 5.13662040e-01 1.91888914e-01
1.02654052e+00 4.10125643e-01 1.01601887e+00 2.59632200e-01
3.13351005e-01 4.61057365e-01 3.96864653e-01 -1.79181933e-01
-2.17132062e-01 -1.63923562e-01 4.71686423e-01 7.79748082e-01
-9.47099328e-02 -3.36708874e-01 -7.48646736e-01 8.43989730e-01
-1.91571736e+00 -9.00313973e-01 -2.42176458e-01 1.90312898e+00
8.77214313e-01 -1.02594204e-01 2.77966529e-01 -4.74315472e-02
9.02406335e-01 3.29576761e-01 -5.98393500e-01 -3.47264186e-02
1.42961249e-01 2.34930813e-02 5.32257199e-01 4.50069308e-01
-1.31644773e+00 1.35336876e+00 5.99803162e+00 5.82408249e-01
-1.12217391e+00 4.54462588e-01 8.52596045e-01 -3.38940054e-01
-2.15547364e-02 1.12417564e-01 -9.24808741e-01 5.86060941e-01
1.01462567e+00 3.09019238e-02 3.61420959e-01 1.08277547e+00
2.51098067e-01 -2.32451767e-01 -1.09255373e+00 9.52514648e-01
1.30886570e-01 -1.42151284e+00 -1.54187232e-01 -1.76764727e-01
1.28278160e+00 2.20389619e-01 -3.10465753e-01 2.75383174e-01
7.19564557e-02 -5.96501470e-01 1.01246798e+00 1.90192744e-01
1.13777101e+00 -1.64822698e-01 7.84172297e-01 2.49434486e-01
-1.27522194e+00 1.53577641e-01 -2.94791251e-01 1.80133715e-01
5.67560554e-01 2.11650133e-01 -6.13251030e-01 5.53995311e-01
7.43890882e-01 1.12387526e+00 -4.45803583e-01 1.10200572e+00
-3.39185804e-01 8.28323483e-01 -3.35165918e-01 4.38696295e-01
5.48924029e-01 -2.96757132e-01 1.96928486e-01 1.24244070e+00
1.73635513e-01 2.73741603e-01 6.26286864e-01 5.31237304e-01
-1.41436026e-01 1.49299160e-01 -3.69333535e-01 -2.28775069e-01
1.69690132e-01 1.24799109e+00 -1.11993587e+00 -9.22649860e-01
-7.91050732e-01 1.18355584e+00 1.28978044e-01 5.57875454e-01
-9.51255620e-01 4.34749387e-02 1.98110774e-01 1.37443855e-01
5.84473133e-01 -1.31130621e-01 3.61498892e-02 -1.44071949e+00
-2.00903431e-01 -5.78164160e-01 4.13611561e-01 -7.47222483e-01
-9.40100431e-01 6.38321936e-01 4.13144901e-02 -1.30636120e+00
-3.88053089e-01 -3.71767461e-01 -5.02413273e-01 4.31675673e-01
-1.78466105e+00 -1.04499793e+00 -5.61756313e-01 8.42104554e-01
1.18833816e+00 7.07780644e-02 5.11637092e-01 5.94470799e-01
-5.99011362e-01 1.07255079e-01 -1.11783832e-01 3.81824017e-01
6.08822227e-01 -1.13292885e+00 3.51983905e-01 1.00959611e+00
5.11858702e-01 1.85442537e-01 5.16387820e-01 -6.34905338e-01
-1.01162744e+00 -1.30175960e+00 5.75675309e-01 -2.52434015e-01
4.87354755e-01 -1.34243116e-01 -9.55473781e-01 1.08176303e+00
3.06645393e-01 7.16450572e-01 7.32694328e-01 -4.79817778e-01
9.34374332e-02 3.46473038e-01 -9.26466525e-01 4.75170434e-01
1.08176887e+00 -3.09941560e-01 -5.67756653e-01 6.16466105e-01
7.61658251e-01 -6.61269307e-01 -7.30793357e-01 2.37744644e-01
2.25527287e-01 -8.65153134e-01 7.88485169e-01 -6.29073024e-01
4.06533033e-01 -5.59934199e-01 -1.98308587e-01 -8.74547541e-01
3.58512178e-02 -6.26111627e-01 -1.15418755e-01 1.13764465e+00
2.68854350e-01 -1.75140232e-01 1.07919788e+00 5.62184155e-01
-4.04231220e-01 -8.42576861e-01 -6.44266427e-01 -6.93910539e-01
-4.33531940e-01 -4.20486391e-01 1.31705388e-01 9.43141520e-01
-3.93186897e-01 1.85762480e-01 -5.59404254e-01 6.17941432e-02
5.10752082e-01 1.54068038e-01 8.87454271e-01 -9.02093351e-01
-2.85011590e-01 1.43986002e-01 -2.83343315e-01 -1.23200727e+00
2.81720281e-01 -7.73154378e-01 3.94374043e-01 -1.60187411e+00
1.58199370e-01 -3.42265755e-01 -2.01367423e-01 6.29230082e-01
-2.62976289e-01 9.07891035e-01 3.82264525e-01 4.56709325e-01
-1.11429656e+00 4.22607809e-01 1.26018512e+00 -9.08180326e-02
-3.65084022e-01 -5.13534956e-02 -1.50103673e-01 8.59423697e-01
8.38446081e-01 -5.53148806e-01 -4.68489796e-01 -4.03963685e-01
-5.07082164e-01 1.85944974e-01 2.62921244e-01 -9.10414279e-01
2.36445531e-01 -3.70530814e-01 1.62379190e-01 -4.54980791e-01
4.34833527e-01 -8.25061083e-01 2.83218235e-01 3.08654845e-01
-6.20164946e-02 -4.74103272e-01 6.52113464e-03 5.73886991e-01
-3.59338552e-01 -6.56346619e-01 9.57244992e-01 -3.64523798e-01
-1.26728523e+00 6.94904208e-01 -2.81468451e-01 2.78535277e-01
1.58973134e+00 -4.31132525e-01 1.96976140e-01 -1.56931669e-01
-1.09237504e+00 3.77657175e-01 6.95206285e-01 2.76224405e-01
5.81411779e-01 -1.11876965e+00 -4.20916438e-01 1.83428958e-01
9.51119065e-02 4.45838004e-01 4.88473356e-01 7.46886551e-01
-7.05847740e-01 2.35640854e-01 -2.65967995e-01 -9.25844014e-01
-1.20976818e+00 8.43649387e-01 5.20521700e-02 1.00264050e-01
-6.07048512e-01 9.30934012e-01 1.86028734e-01 3.11121205e-03
1.74551576e-01 -1.67767420e-01 -2.04557985e-01 1.01951458e-01
5.12596607e-01 3.10302880e-02 -3.43109339e-01 -9.21864092e-01
-1.76392823e-01 7.33779550e-01 2.11708955e-02 -1.00662760e-01
1.29777169e+00 -2.09642932e-01 1.61934748e-01 7.53774464e-01
1.25476515e+00 -1.54688761e-01 -1.88156021e+00 -2.09143251e-01
-1.82810647e-04 -5.98097861e-01 -1.83064401e-01 -1.51995778e-01
-1.54203868e+00 7.25291133e-01 3.04399729e-01 1.25347599e-01
1.03328300e+00 2.62804747e-01 9.56873000e-01 3.03803757e-02
2.02502131e-01 -1.10568810e+00 3.22369993e-01 2.07621247e-01
2.73542732e-01 -1.51203370e+00 -1.82364911e-01 -7.06060052e-01
-1.02917504e+00 9.40852702e-01 6.19776666e-01 -4.08171684e-01
5.27209640e-01 1.05734855e-01 1.65607214e-01 -1.70599923e-01
-4.97320026e-01 -4.84772891e-01 2.01432273e-01 4.16757286e-01
2.39558265e-01 -2.40371928e-01 -1.91467598e-01 3.86963040e-01
3.60599130e-01 4.00644064e-01 6.24449492e-01 9.86517370e-01
-2.91679978e-01 -1.01200294e+00 -1.34941384e-01 3.62596452e-01
-5.66724420e-01 -3.02396044e-02 -4.87410054e-02 5.90305448e-01
2.02875376e-01 8.79801929e-01 5.18010929e-02 -2.19939277e-01
1.59308210e-01 3.09609890e-01 2.45825350e-01 -1.17309570e+00
-3.61528426e-01 3.51381153e-01 -1.47350356e-01 -7.51932621e-01
-9.27237034e-01 -8.62097144e-01 -1.35109341e+00 -8.96878541e-03
-3.88388336e-01 2.29195356e-01 4.90466565e-01 9.91846681e-01
2.62821674e-01 5.17226934e-01 5.17333508e-01 -1.30186546e+00
-1.07973814e-01 -7.15075850e-01 -5.14616013e-01 5.97882926e-01
2.94351399e-01 -6.34855986e-01 -3.39623541e-01 7.89337456e-01] | [9.18018627166748, 0.06681721657514572] |
5f2c7d44-25ec-458e-aa0f-eff5f3a11e3c | knowledge-enhanced-attentive-learning-for | 1912.07915 | null | https://arxiv.org/abs/1912.07915v1 | https://arxiv.org/pdf/1912.07915v1.pdf | Knowledge-Enhanced Attentive Learning for Answer Selection in Community Question Answering Systems | In the community question answering (CQA) system, the answer selection task aims to identify the best answer for a specific question, and thus is playing a key role in enhancing the service quality through recommending appropriate answers for new questions. Recent advances in CQA answer selection focus on enhancing the performance by incorporating the community information, particularly the expertise (previous answers) and authority (position in the social network) of an answerer. However, existing approaches for incorporating such information are limited in (a) only considering either the expertise or the authority, but not both; (b) ignoring the domain knowledge to differentiate topics of previous answers; and (c) simply using the authority information to adjust the similarity score, instead of fully utilizing it in the process of measuring the similarity between segments of the question and the answer. We propose the Knowledge-enhanced Attentive Answer Selection (KAAS) model, which enhances the performance through (a) considering both the expertise and the authority of the answerer; (b) utilizing the human-labeled tags, the taxonomy of the tags, and the votes as the domain knowledge to infer the expertise of the answer; (c) using matrix decomposition of the social network (formed by following-relationship) to infer the authority of the answerer and incorporating such information in the process of evaluating the similarity between segments. Besides, for vertical community, we incorporate an external knowledge graph to capture more professional information for vertical CQA systems. Then we adopt the attention mechanism to integrate the analysis of the text of questions and answers and the aforementioned community information. Experiments with both vertical and general CQA sites demonstrate the superior performance of the proposed KAAS model. | ['Fengshi Jing', 'Qingpeng Zhang'] | 2019-12-17 | null | null | null | null | ['answer-selection'] | ['natural-language-processing'] | [-3.68252367e-01 1.23750299e-01 2.62737311e-02 -1.43715203e-01
-6.01324141e-01 -5.18994868e-01 3.89435530e-01 5.12053072e-01
-3.80885988e-01 2.10807964e-01 7.40116239e-01 -2.33646885e-01
-5.56500137e-01 -9.06741202e-01 -1.39352724e-01 -6.57726943e-01
3.68561953e-01 5.80844879e-01 7.78974652e-01 -4.73843038e-01
4.35280919e-01 -1.78912491e-01 -1.23194063e+00 2.68238991e-01
1.54386485e+00 1.17654395e+00 2.16317475e-01 4.25499529e-01
-6.64696634e-01 1.08122075e+00 -4.36193585e-01 -4.97590095e-01
6.64929627e-03 -6.65524960e-01 -1.24555731e+00 1.65764913e-01
-9.71055999e-02 -7.36397430e-02 -1.11127935e-01 8.63363266e-01
3.32954913e-01 2.18805492e-01 4.82990921e-01 -1.02984929e+00
-6.15206540e-01 5.53907335e-01 -3.67482424e-01 3.63775700e-01
6.61998510e-01 -1.46110147e-01 1.43025815e+00 -6.42275810e-01
2.69891769e-01 1.10314465e+00 3.63969356e-01 6.75957352e-02
-4.82487440e-01 -3.68628591e-01 4.27281082e-01 6.47457063e-01
-1.31035984e+00 -1.50458232e-01 9.10229623e-01 -6.11506224e-01
3.94603997e-01 1.40341371e-01 6.24229610e-01 2.56913900e-01
-3.83037090e-01 6.61949635e-01 8.45986128e-01 -4.01522040e-01
3.48265290e-01 2.46972278e-01 6.07902527e-01 5.94757318e-01
-3.49235207e-01 -7.52925992e-01 -2.43386656e-01 -6.85470641e-01
2.25661322e-01 1.94494367e-01 -4.11987454e-01 -3.49670760e-02
-9.75080967e-01 9.67223406e-01 7.50041246e-01 6.42397940e-01
-7.18622208e-01 -4.45958167e-01 9.35832709e-02 2.64262319e-01
2.91977614e-01 5.58632076e-01 -2.11333215e-01 1.67269632e-01
-5.67751884e-01 8.92066360e-02 1.00755155e+00 5.43104291e-01
1.03456867e+00 -5.90583801e-01 -5.62230706e-01 6.64234757e-01
7.28136122e-01 2.14247793e-01 4.07402545e-01 -1.29779971e+00
4.27071601e-01 1.35951662e+00 4.15047944e-01 -1.25373971e+00
-1.24461211e-01 -6.07219100e-01 -3.67088348e-01 -4.67768997e-01
5.18164098e-01 -2.55517483e-01 -3.46731663e-01 1.62007332e+00
7.23758519e-01 6.22806326e-02 -2.08990902e-01 9.01560009e-01
1.08025694e+00 4.69447196e-01 -1.19363531e-01 -3.56858641e-01
1.54852319e+00 -1.15957594e+00 -8.37166548e-01 1.11119315e-01
5.34216464e-01 -7.83062696e-01 8.08536351e-01 -6.10725656e-02
-6.12653375e-01 -3.98614496e-01 -6.93083465e-01 -2.89416104e-03
-1.20654292e-01 -2.17497908e-02 1.46066546e-01 2.24381223e-01
-1.06024659e+00 1.59370437e-01 -6.41551167e-02 -4.16794956e-01
1.03175588e-01 2.69648552e-01 2.12387159e-01 -2.99923509e-01
-1.58591962e+00 5.96867442e-01 -1.17066950e-01 9.13456455e-02
-4.28792298e-01 -2.68575281e-01 -3.55325401e-01 4.29587990e-01
7.02410936e-01 -1.02077663e+00 9.40979540e-01 -1.17998040e+00
-1.07552183e+00 1.84715912e-01 -2.80690283e-01 5.47945201e-02
2.46499076e-01 3.17640081e-02 -3.80420297e-01 4.84900922e-01
3.61593127e-01 2.04965025e-01 5.09478748e-01 -1.04843056e+00
-1.00141394e+00 -5.50483167e-01 6.99642360e-01 6.56206071e-01
-5.30393481e-01 3.48374918e-02 -7.46795774e-01 3.63941826e-02
2.00099364e-01 -5.80136657e-01 -2.12621987e-01 -1.47096604e-01
-5.49862348e-02 -1.04529786e+00 5.52925527e-01 -1.04302633e+00
1.59878790e+00 -1.91821349e+00 5.71315885e-02 5.01718044e-01
6.17129922e-01 2.68378884e-01 -1.79751143e-01 6.10813260e-01
6.47178769e-01 4.21390533e-02 1.45619676e-01 2.10416660e-01
-1.51034564e-01 1.01900116e-01 -2.34542377e-02 -2.94559598e-02
-2.38891825e-01 5.42111993e-01 -1.18372810e+00 -8.83342505e-01
-4.77144778e-01 1.10634506e-01 -6.87242746e-01 3.90061468e-01
-2.52550989e-01 5.49529314e-01 -1.08113253e+00 2.40127966e-01
2.90669531e-01 -6.57598794e-01 1.79763585e-01 -3.49464059e-01
7.81238005e-02 4.28358793e-01 -1.21921945e+00 1.10750353e+00
-3.68260682e-01 4.66308072e-02 4.86188799e-01 -7.22450972e-01
8.82687509e-01 4.99274045e-01 6.24177217e-01 -6.49700761e-01
-3.14493701e-02 2.03435972e-01 2.45981321e-01 -9.34703052e-01
1.71845615e-01 3.82792473e-01 1.96729898e-01 7.37582803e-01
-9.59657505e-02 4.16121840e-01 2.83745199e-01 9.44318533e-01
1.21151412e+00 -1.96608409e-01 1.04018867e-01 -1.68809637e-01
1.00404787e+00 -3.71011347e-02 4.15126085e-01 6.04408622e-01
-3.45380664e-01 2.20581025e-01 4.34893966e-01 -1.34524638e-02
-5.88667810e-01 -5.15117228e-01 3.68430197e-01 1.29753828e+00
3.44412923e-01 -3.94993305e-01 -7.31584609e-01 -8.95112991e-01
3.43956240e-02 4.13997263e-01 -6.29462183e-01 -8.92770588e-02
-4.58346188e-01 -1.97493732e-01 2.24728044e-02 4.09674168e-01
6.38942957e-01 -8.85123134e-01 2.07077473e-01 1.82361752e-01
-1.02497768e+00 -8.07880044e-01 -8.67162526e-01 -4.43583369e-01
-5.99781275e-01 -1.39619148e+00 -5.43833673e-01 -8.11880112e-01
6.58806324e-01 5.43165326e-01 9.22135174e-01 7.52006471e-01
6.20655298e-01 8.01116109e-01 -4.92317617e-01 1.30395234e-01
2.85637478e-04 1.27765268e-01 -1.04793303e-01 6.37400210e-01
4.86675054e-01 -6.47098720e-01 -1.09557104e+00 7.80544579e-01
-6.60811245e-01 -2.87373990e-01 3.17110896e-01 5.06588519e-01
7.04653934e-02 7.69241201e-03 1.13231683e+00 -8.22612643e-01
1.02857852e+00 -8.82781506e-01 6.98664486e-02 5.36222219e-01
-9.78216767e-01 -2.15297341e-01 4.00935590e-01 -1.80430144e-01
-1.33249187e+00 -5.56212723e-01 -3.19598585e-01 7.75708407e-02
3.08694959e-01 7.99579561e-01 -1.01175167e-01 7.07950965e-02
4.69880015e-01 -1.07103556e-01 -1.06533334e-01 -5.48616469e-01
3.49817425e-01 1.09607756e+00 1.09327234e-01 -4.57112402e-01
6.24834180e-01 4.46860641e-01 -4.17553335e-01 -2.92198569e-01
-1.16644073e+00 -1.33709919e+00 -4.10870910e-01 -5.26241183e-01
9.19560909e-01 -7.80120373e-01 -9.80706453e-01 1.73194960e-01
-1.05723655e+00 4.98336941e-01 -1.70871019e-02 4.64225709e-01
2.07639977e-01 6.78106189e-01 -6.40091181e-01 -9.33647335e-01
-4.49100792e-01 -8.61975968e-01 4.84721750e-01 5.63989997e-01
-5.81805520e-02 -1.13336992e+00 1.90609321e-02 1.39064455e+00
5.13257623e-01 -3.48274052e-01 1.12489021e+00 -1.03564847e+00
-7.29395807e-01 -2.88879901e-01 -4.13057536e-01 1.81368276e-01
1.24537855e-01 -2.14239940e-01 -6.35224819e-01 -1.08316064e-01
2.17693612e-01 -4.53097522e-02 7.30960429e-01 9.30114985e-02
3.48417103e-01 -4.70904440e-01 -2.51579136e-01 -3.58326435e-01
9.76509809e-01 8.86248574e-02 2.99983323e-01 3.01070306e-02
7.71113753e-01 1.02346730e+00 6.32216573e-01 4.02505159e-01
1.03780055e+00 5.85469604e-01 4.02658671e-01 1.48138478e-01
4.65565287e-02 -3.68579745e-01 2.80880302e-01 1.70077312e+00
-2.44118705e-01 -1.25911534e-01 -7.79198170e-01 8.63538623e-01
-2.07701874e+00 -8.69542718e-01 -5.43658257e-01 1.96581018e+00
7.46967971e-01 -1.58861771e-01 1.54750153e-01 3.04437093e-02
9.74987984e-01 -1.13293240e-02 -5.11671305e-01 5.71108758e-02
8.16678926e-02 -2.77439415e-01 -2.06987873e-01 5.83296120e-01
-3.43806952e-01 3.38266075e-01 4.70310926e+00 1.00228226e+00
-2.93280840e-01 3.40999335e-01 3.54662687e-01 3.09642076e-01
-8.73770416e-01 4.76707846e-01 -5.82795203e-01 4.70217556e-01
4.23267186e-01 -2.69870758e-01 5.41154325e-01 5.28606534e-01
2.81511664e-01 -2.85615951e-01 -6.93677425e-01 1.96486607e-01
1.90136373e-01 -9.35634434e-01 -5.57650551e-02 -1.65504813e-01
7.12658823e-01 -1.27140984e-01 -4.57798570e-01 5.96307516e-01
3.74141514e-01 -2.72539914e-01 2.76687592e-01 7.95971870e-01
-1.10817932e-01 -3.75722110e-01 1.15666580e+00 7.31385648e-01
-1.23504210e+00 -4.34923530e-01 -2.19358221e-01 -1.07851237e-01
4.05185409e-02 5.08346140e-01 -5.12775302e-01 7.39466846e-01
8.51733565e-01 5.25429189e-01 -6.88325942e-01 1.04048479e+00
-4.39641386e-01 9.42362249e-01 -1.28449306e-01 -2.09144667e-01
2.45492324e-01 -4.18393642e-01 3.31157148e-01 6.39331043e-01
1.09565645e-01 3.54813606e-01 4.09948409e-01 6.20106041e-01
-2.44711444e-01 6.91307187e-01 1.07974268e-01 9.07925665e-02
6.66445613e-01 1.29756546e+00 -4.79066581e-01 -2.90306717e-01
-6.76227689e-01 5.36986053e-01 3.60373378e-01 6.48306906e-01
-2.69264698e-01 -4.88572836e-01 1.58619553e-01 4.20506179e-01
3.54179591e-01 2.77731478e-01 6.50164261e-02 -8.49236310e-01
3.69552486e-02 -5.98693192e-01 8.74048948e-01 -9.94344354e-01
-1.47286618e+00 3.51759553e-01 -4.37059671e-01 -9.21692967e-01
8.74222890e-02 9.83453095e-02 -8.39874446e-01 1.20851040e+00
-1.60288322e+00 -9.02951837e-01 -4.73136753e-01 7.28117406e-01
6.40752241e-02 3.60515825e-02 3.50881934e-01 5.35187960e-01
-3.63904834e-01 1.62244022e-01 -3.70910726e-02 2.08333611e-01
7.47046113e-01 -9.89424825e-01 -4.72693741e-01 4.38301325e-01
-1.41832307e-01 6.81909382e-01 3.15781504e-01 -5.47747135e-01
-8.70417297e-01 -5.81950545e-01 1.46696496e+00 -3.51001471e-01
6.76016748e-01 9.88950655e-02 -1.17189813e+00 2.44677633e-01
3.52301091e-01 -4.14631933e-01 9.06039715e-01 5.42874634e-01
-1.72609657e-01 -3.55323076e-01 -1.09941554e+00 2.98783511e-01
7.40372837e-01 -7.23478496e-01 -6.82628989e-01 2.77382553e-01
9.91424203e-01 3.33750844e-01 -9.43921924e-01 2.85200655e-01
3.09853673e-01 -8.13199103e-01 6.14891827e-01 -3.71825933e-01
3.07478487e-01 -7.10689366e-01 1.48442835e-01 -9.22536433e-01
-6.89601898e-01 -8.38341787e-02 -1.55000493e-01 1.52757537e+00
3.66086185e-01 -6.25781655e-01 5.50481796e-01 6.13164783e-01
1.41615709e-02 -9.28422630e-01 -7.39794374e-01 7.47744814e-02
-3.57693851e-01 1.93782046e-01 6.68940902e-01 1.10491776e+00
1.66511461e-01 7.62545526e-01 -6.04541562e-02 2.11142048e-01
5.58647290e-02 4.67970669e-01 3.70072961e-01 -1.54313445e+00
-2.29132548e-01 -3.12741935e-01 2.03276232e-01 -1.32256925e+00
-5.20903021e-02 -8.07157874e-01 -1.29666343e-01 -2.08228183e+00
4.98401135e-01 -4.73300666e-01 -5.74041903e-01 7.79448748e-02
-7.73648560e-01 -1.93084359e-01 -3.57801691e-02 8.27531159e-01
-1.29884636e+00 4.43835348e-01 1.58959854e+00 2.36635841e-02
-1.18470676e-01 3.99611473e-01 -1.05519331e+00 5.85573375e-01
3.74472678e-01 -4.85289246e-01 -5.19881070e-01 -4.34500396e-01
8.00483227e-01 3.49073142e-01 1.12163350e-02 -3.98734421e-01
8.59538734e-01 -1.43690452e-01 -1.23718046e-01 -3.70663971e-01
-7.26595968e-02 -7.52899528e-01 -3.27099591e-01 2.59323984e-01
-5.09788871e-01 1.04422239e-03 -6.32104158e-01 1.04012287e+00
-4.22991097e-01 -5.86601436e-01 1.73134416e-01 -2.41593674e-01
-4.38369691e-01 3.01316351e-01 -3.59827518e-01 3.18710059e-01
5.19666731e-01 -2.29121987e-02 -4.37665999e-01 -9.29226398e-01
-7.05894828e-01 9.94323194e-01 2.13043392e-01 3.97481859e-01
2.43676320e-01 -1.29392827e+00 -8.13274086e-01 -3.02314430e-01
2.48544216e-01 -1.22422986e-01 5.50386250e-01 1.09001362e+00
-1.18080482e-01 2.09455729e-01 2.86746830e-01 -3.88550878e-01
-1.17114031e+00 4.27972674e-01 3.04927289e-01 -7.05477834e-01
2.71346569e-01 6.00063384e-01 1.80798963e-01 -4.87383097e-01
2.07849219e-01 3.15898687e-01 -1.21504307e+00 4.44661796e-01
3.25715303e-01 5.31030893e-01 -1.18522421e-01 -5.92491090e-01
-3.00932795e-01 4.23451722e-01 4.05458435e-02 3.41928704e-03
9.76139247e-01 -5.83240092e-01 -6.06393993e-01 2.59436280e-01
7.13834405e-01 3.19930702e-01 -7.26524711e-01 -1.01947713e+00
1.03945658e-02 -1.13616817e-01 8.83565396e-02 -8.08387101e-01
-1.03877652e+00 7.17992306e-01 1.64822161e-01 4.76221025e-01
1.08021569e+00 1.78053781e-01 8.32906365e-01 2.91545093e-01
1.86853677e-01 -1.13688087e+00 5.13336718e-01 3.65202576e-01
5.77139676e-01 -1.02083743e+00 -3.33469361e-02 -4.35169071e-01
-6.11527264e-01 8.53050530e-01 6.84723914e-01 2.92545885e-01
9.74789917e-01 -6.14384890e-01 3.08900744e-01 -4.18038487e-01
-7.52896786e-01 -4.72710133e-01 5.05750120e-01 1.43494904e-01
2.50789911e-01 -1.69385165e-01 -7.36577213e-01 8.37347806e-01
2.84774452e-01 -2.01650903e-01 2.11540699e-01 7.70723820e-01
-8.17671895e-01 -9.69460666e-01 -2.19855607e-01 5.54774880e-01
-4.11349326e-01 -6.94039240e-02 -5.04115641e-01 -9.97233614e-02
3.69366616e-01 1.72429538e+00 -2.32788682e-01 -4.22866464e-01
2.00329334e-01 1.25045374e-01 -1.49952173e-01 -5.84565580e-01
-1.16316867e+00 -4.03654240e-02 1.73054993e-01 -2.12168656e-02
-7.67347932e-01 -5.07804692e-01 -1.03298795e+00 -2.67919958e-01
-9.41850543e-01 9.69122648e-01 3.93508315e-01 1.31843233e+00
6.02558196e-01 4.05926615e-01 9.79969263e-01 1.30982369e-01
-4.14106160e-01 -1.15546882e+00 -4.50086683e-01 7.11618781e-01
8.04696418e-03 -3.43219072e-01 -6.29845619e-01 -3.55983794e-01] | [11.419778823852539, 7.976271152496338] |
f35fe6b2-a6a2-49dc-a8f7-5ae4307ce7c4 | novel-deep-learning-model-for-traffic-sign | 1805.04424 | null | http://arxiv.org/abs/1805.04424v1 | http://arxiv.org/pdf/1805.04424v1.pdf | Novel Deep Learning Model for Traffic Sign Detection Using Capsule Networks | Convolutional neural networks are the most widely used deep learning
algorithms for traffic signal classification till date but they fail to capture
pose, view, orientation of the images because of the intrinsic inability of max
pooling layer.This paper proposes a novel method for Traffic sign detection
using deep learning architecture called capsule networks that achieves
outstanding performance on the German traffic sign dataset.Capsule network
consists of capsules which are a group of neurons representing the
instantiating parameters of an object like the pose and orientation by using
the dynamic routing and route by agreement algorithms.unlike the previous
approaches of manual feature extraction,multiple deep neural networks with many
parameters,our method eliminates the manual effort and provides resistance to
the spatial variances.CNNs can be fooled easily using various adversary attacks
and capsule networks can overcome such attacks from the intruders and can offer
more reliability in traffic sign detection for autonomous vehicles.Capsule
network have achieved the state-of-the-art accuracy of 97.6% on German Traffic
Sign Recognition Benchmark dataset (GTSRB). | ['Amara Dinesh Kumar'] | 2018-05-11 | null | null | null | null | ['traffic-sign-recognition', 'traffic-sign-detection'] | ['computer-vision', 'computer-vision'] | [-3.74642581e-01 -3.45624536e-01 4.98527549e-02 -4.47474986e-01
-2.68685043e-01 -8.27446759e-01 3.50907832e-01 -1.04501128e+00
-4.42088515e-01 2.77315438e-01 -3.51416916e-01 -3.62958133e-01
-7.16973469e-02 -5.98048151e-01 -8.06328356e-01 -9.96986210e-01
-3.30895066e-01 2.57630318e-01 8.06555867e-01 3.03339306e-02
1.95882276e-01 9.67435002e-01 -1.56620455e+00 2.77775496e-01
3.84154677e-01 1.30768883e+00 -3.41625899e-01 8.12030137e-01
4.64046337e-02 8.80537450e-01 -1.08143592e+00 -1.99781299e-01
4.40362751e-01 6.60050437e-02 -1.46347716e-01 -1.88320175e-01
9.18820977e-01 -5.18465996e-01 -9.78638351e-01 1.03503060e+00
6.64219856e-01 -3.68164092e-01 4.17971730e-01 -1.71185052e+00
-3.29689324e-01 1.83096573e-01 -1.85597807e-01 1.69146448e-01
-2.07884625e-01 5.07349074e-01 2.05027714e-01 -3.44685435e-01
7.81759202e-01 9.67519641e-01 8.88557374e-01 8.11267614e-01
-4.80243713e-01 -1.34632719e+00 -1.36907041e-01 5.38595855e-01
-1.25971568e+00 -3.93161088e-01 8.33965540e-01 -3.33164126e-01
7.40150452e-01 1.41934514e-01 3.60769749e-01 1.32016766e+00
2.09945247e-01 8.26628566e-01 1.04890072e+00 2.01375917e-01
5.83711825e-02 -3.45570147e-02 4.04344261e-01 9.32037532e-01
6.37130141e-01 5.06219625e-01 -2.13868350e-01 -1.35143474e-01
5.56570828e-01 -1.68949798e-01 -1.52114123e-01 -4.11539555e-01
-8.67490888e-01 7.15738177e-01 6.27583385e-01 -8.73638242e-02
-1.03486732e-01 7.52864242e-01 5.79985142e-01 4.68321472e-01
-7.73747146e-01 2.58639678e-02 -6.17689312e-01 9.84485913e-03
-5.09319544e-01 2.32734904e-01 7.89364159e-01 1.04959464e+00
8.94338042e-02 5.48729539e-01 -9.31969434e-02 9.95486975e-02
5.71398914e-01 1.00458634e+00 4.09192562e-01 -6.42652154e-01
4.69803840e-01 3.37785453e-01 -5.54106124e-02 -1.24439609e+00
-6.84005916e-01 -4.40752357e-01 -4.88151640e-01 8.27504635e-01
6.93893731e-01 -3.88111293e-01 -1.48014259e+00 1.26120985e+00
1.47680402e-01 3.48084241e-01 1.21517763e-01 9.15468872e-01
1.33202648e+00 1.50002539e-01 2.31507849e-02 7.10106194e-01
1.48257279e+00 -8.83410752e-01 -6.22258186e-01 -3.01947564e-01
4.86703306e-01 -6.67823792e-01 8.82817246e-03 4.72601801e-01
-4.11438376e-01 -5.16687930e-01 -1.45347810e+00 3.22723687e-01
-9.24481690e-01 3.91807705e-01 9.48324978e-01 1.43073177e+00
-8.67609441e-01 1.89994290e-01 -7.94768155e-01 -1.73250884e-01
1.05456865e+00 9.49160397e-01 -4.60770935e-01 1.71887234e-01
-8.69258583e-01 9.41374242e-01 -1.06205419e-01 7.54619241e-01
-1.00733948e+00 -3.36290926e-01 -7.05498636e-01 -2.71245718e-01
4.03165489e-01 -1.26066595e-01 9.13153350e-01 -7.74975061e-01
-1.56443787e+00 7.67757237e-01 9.50489342e-02 -5.82719147e-01
7.72888958e-01 9.52537581e-02 -7.70451128e-01 1.82171583e-01
-3.27711791e-01 6.37838602e-01 1.09210598e+00 -9.62078869e-01
-3.43469858e-01 -1.53091505e-01 -2.72192389e-01 -7.89345622e-01
2.43771553e-01 3.16649795e-01 -3.10028195e-01 -1.69350833e-01
2.31724724e-01 -9.96999562e-01 2.06636637e-02 6.31774426e-01
-3.59682709e-01 -7.69828632e-02 1.91323805e+00 -5.36684096e-01
5.62515736e-01 -2.14498210e+00 -7.37846851e-01 5.58563888e-01
2.35190853e-01 9.73046541e-01 -2.95436323e-01 -2.01001227e-01
8.39628056e-02 -3.31481807e-02 3.74216050e-01 6.57541305e-02
2.61113644e-01 2.87856430e-01 -3.56300086e-01 9.75985050e-01
1.20299205e-01 1.32310843e+00 -4.71051842e-01 -3.03710669e-01
3.30513775e-01 6.79296672e-01 -1.41144678e-01 -1.60624549e-01
2.32820049e-01 1.23219319e-01 -8.81602645e-01 1.13380086e+00
1.06690228e+00 1.86647937e-01 -2.50457764e-01 -4.84720916e-01
2.16680169e-01 -1.24632627e-01 -1.05170071e+00 8.25571895e-01
1.11557513e-01 1.28571475e+00 2.13890120e-01 -8.42841804e-01
1.21317244e+00 2.49186978e-01 2.22566217e-01 -7.35553443e-01
7.93634355e-01 4.01938230e-01 2.97528267e-01 -8.84357452e-01
1.83440419e-03 3.60910833e-01 -3.09639663e-01 1.22566856e-01
-1.84353471e-01 7.83866793e-02 -1.30590692e-01 -1.22606449e-01
1.20274460e+00 -7.92542472e-02 -3.89180809e-01 2.00132072e-01
5.79690397e-01 -1.37486875e-01 4.96040732e-01 9.19990063e-01
-9.82770443e-01 3.74182642e-01 5.50704181e-01 -7.51763165e-01
-7.24255383e-01 -1.06390417e+00 -2.40112141e-01 2.20398396e-01
1.60993129e-01 2.81119317e-01 -4.25743639e-01 -9.40629721e-01
5.06754100e-01 1.36669025e-01 -8.21551919e-01 -2.24246934e-01
-8.87003005e-01 -4.06102717e-01 1.56878769e+00 9.90143597e-01
1.10974908e+00 -1.08223271e+00 -7.82329798e-01 -2.19407156e-02
3.78968805e-01 -1.61868477e+00 -3.52803797e-01 5.12759164e-02
-4.02548522e-01 -1.51499987e+00 -3.03324789e-01 -9.33198452e-01
8.19364607e-01 5.57404310e-02 3.08959126e-01 3.20818722e-01
-6.38689816e-01 3.79306227e-02 -1.86324611e-01 -5.52741826e-01
-2.38098592e-01 -3.99313211e-01 -8.72184262e-02 2.74434447e-01
6.26050889e-01 -2.52691537e-01 -6.17830694e-01 8.26780438e-01
-5.06167829e-01 -6.77888393e-01 4.66318160e-01 6.07919514e-01
1.75411869e-02 -1.37005121e-01 4.52583313e-01 -2.88260490e-01
1.86488584e-01 1.90240219e-02 -1.00772035e+00 5.00037968e-02
2.21146140e-02 -9.12534669e-02 3.19995254e-01 -6.58611059e-01
-5.11157095e-01 3.70908737e-01 1.88953921e-01 -3.46231192e-01
-3.67837638e-01 -2.67659009e-01 -1.44065320e-01 -1.21901596e+00
4.10640091e-01 1.59116149e-01 3.05683106e-01 -9.36460271e-02
1.29158691e-01 8.28764677e-01 6.24627411e-01 -1.65497348e-01
1.15446806e+00 8.90727222e-01 3.72981042e-01 -7.45163620e-01
-2.38514185e-01 -1.61465272e-01 -5.19559920e-01 -5.69079220e-01
8.04176509e-01 -6.09716475e-01 -1.40519381e+00 1.15143287e+00
-1.14387846e+00 -3.49210650e-02 4.20454025e-01 4.08097029e-01
-2.86227912e-01 4.68943268e-01 -3.61552566e-01 -5.85744441e-01
-2.26052642e-01 -1.22629261e+00 7.68606842e-01 4.73943383e-01
7.91746229e-02 -5.56842148e-01 -5.64488828e-01 4.62528765e-01
1.03123140e+00 6.88578725e-01 3.37004423e-01 -7.74937212e-01
-1.14366090e+00 -1.00324428e+00 -3.50252122e-01 2.24243551e-01
-1.33870706e-01 2.19952896e-01 -1.13810349e+00 -1.87447160e-01
-4.18562472e-01 -2.14456066e-01 1.01303697e+00 5.06651044e-01
1.00167835e+00 -4.17208783e-02 -6.28500044e-01 1.08206820e+00
1.20284367e+00 4.97773409e-01 9.72709179e-01 5.34001470e-01
6.11313701e-01 2.08014995e-01 -7.22147478e-03 -9.21584293e-02
5.48354685e-02 6.60548210e-01 6.43225670e-01 7.95314088e-02
-3.74529928e-01 6.66460441e-03 3.44263792e-01 -3.24881375e-02
3.28138210e-02 -1.09108344e-01 -5.89180171e-01 2.08952248e-01
-1.66739845e+00 -1.17363322e+00 -3.80505472e-01 1.81955314e+00
-3.45299207e-02 2.97741503e-01 2.58111581e-02 1.90333172e-03
5.06978333e-01 1.29486606e-01 -5.53430200e-01 -3.98420244e-01
-2.74256021e-01 1.20153897e-01 1.32659435e+00 8.62727612e-02
-1.40813398e+00 1.13949156e+00 6.64900541e+00 4.34154958e-01
-1.63840604e+00 -6.07553646e-02 8.06220621e-02 2.90567894e-02
5.53403676e-01 -3.97272676e-01 -9.92095649e-01 4.82449472e-01
7.09731281e-01 6.33367479e-01 8.21241289e-02 9.87663448e-01
7.59967491e-02 1.14981040e-01 -5.55814207e-01 9.36940074e-01
3.76095802e-01 -1.33775222e+00 -4.04144198e-01 1.45142168e-01
5.37712693e-01 4.75830227e-01 2.65145779e-01 2.96894491e-01
3.13937366e-01 -1.18973470e+00 7.13101923e-01 4.20888752e-01
5.60267746e-01 -5.67154467e-01 1.06235027e+00 -1.72054410e-01
-1.26712906e+00 -3.66983414e-01 -2.40973979e-01 3.81128907e-01
2.36974359e-01 -2.82543421e-01 -6.05683744e-01 6.43863752e-02
6.61180675e-01 2.42960021e-01 -7.41792858e-01 1.59478188e+00
-3.26374322e-01 6.02125227e-01 -7.72965252e-01 -5.26082695e-01
5.19990981e-01 4.47093219e-01 5.95897019e-01 1.20510602e+00
1.03262760e-01 -5.63395381e-01 3.70172132e-03 7.62810647e-01
2.51852751e-01 -5.36336958e-01 -6.66785419e-01 -2.48630941e-02
3.17360520e-01 1.04325879e+00 -8.17744076e-01 -1.32705227e-01
-2.43043065e-01 6.49419427e-01 -2.37178236e-01 5.99646389e-01
-1.26143754e+00 -7.52681196e-01 7.89019167e-01 -2.36949548e-01
1.25461686e+00 -4.34960246e-01 -3.46743494e-01 -7.92834580e-01
3.26149583e-01 -7.75906682e-01 2.06034869e-01 -6.71090245e-01
-7.97981858e-01 6.51383579e-01 -3.87165815e-01 -1.44711459e+00
1.15690589e-01 -1.43320155e+00 -9.67464924e-01 3.57402951e-01
-1.65278792e+00 -1.51164067e+00 -5.94539702e-01 6.01444900e-01
7.01617971e-02 -6.67735815e-01 5.20250559e-01 5.26328802e-01
-5.56075692e-01 1.00371122e+00 -8.11182261e-02 9.79546964e-01
5.95684707e-01 -5.41049480e-01 2.61583418e-01 8.13114822e-01
-1.02995664e-01 3.26162517e-01 4.76091266e-01 -5.77191889e-01
-1.52660561e+00 -8.71869862e-01 7.64151454e-01 -6.13197625e-01
6.77700698e-01 -4.06569391e-01 -3.71685028e-01 6.68911457e-01
-2.08641693e-01 5.62392771e-01 2.96236575e-01 -5.33929050e-01
-8.97769392e-01 -4.81883556e-01 -1.39028180e+00 3.66191953e-01
6.89032018e-01 -3.85639489e-01 -3.52531821e-01 1.50959834e-01
3.74902338e-02 -5.81157744e-01 -1.98424846e-01 3.49290162e-01
1.00310087e+00 -6.60560131e-01 1.01209044e+00 -7.67779768e-01
-3.70133519e-01 -8.81518722e-01 6.41350523e-02 -5.47786534e-01
-7.14347810e-02 -7.99103677e-01 -2.98135076e-02 7.64184833e-01
3.58638197e-01 -1.02491653e+00 1.11764395e+00 6.20218813e-01
-1.88542143e-01 -3.82483751e-01 -1.40544665e+00 -1.08294785e+00
-2.55428076e-01 -5.88706732e-01 7.42058575e-01 6.03775978e-01
-3.82892907e-01 -5.44535339e-01 -5.18539608e-01 6.93844199e-01
9.27408993e-01 -2.21757367e-01 1.07458460e+00 -9.46466506e-01
9.85777229e-02 -6.62393212e-01 -1.65753531e+00 -7.86181450e-01
9.81349200e-02 -6.23864293e-01 1.24043142e-02 -9.92281497e-01
-3.43175560e-01 -1.16141953e-01 -1.71091884e-01 7.12947607e-01
6.03659749e-01 4.28249657e-01 2.19900265e-01 -3.04564625e-01
-6.29767478e-01 1.13275416e-01 1.22020102e+00 -4.66000587e-01
9.29022431e-02 1.05543382e-01 -1.96027011e-01 7.10089684e-01
7.86407292e-01 -5.42162478e-01 2.12306634e-01 -3.76743913e-01
-2.96263397e-01 -5.49176872e-01 1.03191185e+00 -1.22635388e+00
8.91695678e-01 2.00503066e-01 5.35170257e-01 -9.09200907e-01
2.06598550e-01 -1.23030317e+00 8.41064453e-02 8.41467261e-01
2.31458277e-01 -3.87351960e-01 5.48406482e-01 5.75727284e-01
-1.14869408e-01 -5.76328002e-02 7.92376697e-01 2.11390883e-01
-9.83747900e-01 3.84995997e-01 -6.03499770e-01 -5.17471910e-01
1.13297737e+00 -8.20479751e-01 -8.04471493e-01 -2.65915364e-01
-4.33044314e-01 4.58034813e-01 -1.85898915e-01 7.77830899e-01
6.35451019e-01 -1.34375834e+00 -3.55369866e-01 6.51512325e-01
4.25232537e-02 -4.01836634e-01 1.97243452e-01 8.74080300e-01
-9.04207587e-01 6.55702651e-01 -3.92721593e-01 -5.60295641e-01
-1.66990757e+00 1.93252623e-01 9.87889171e-01 3.36765289e-01
-6.03987873e-01 8.36202741e-01 -4.35256720e-01 -3.46457571e-01
7.05682933e-01 -4.63431686e-01 -1.27217129e-01 -3.07742119e-01
5.80802977e-01 3.04646969e-01 -1.43851742e-01 -8.77207577e-01
-7.43943751e-01 1.10005224e+00 -1.95692599e-01 4.00479138e-01
1.07448459e+00 4.53663915e-01 3.53971750e-01 -5.46359718e-01
1.30662906e+00 -1.72675401e-01 -1.36081469e+00 1.45154670e-01
-1.32050887e-01 -5.17880797e-01 1.43441528e-01 -9.05963242e-01
-1.66649437e+00 6.68670952e-01 9.97117341e-01 -1.82985261e-01
4.17447448e-01 -2.12635458e-01 9.43523228e-01 8.61255884e-01
3.19053560e-01 -1.13868356e+00 -1.53514579e-01 5.26177108e-01
7.16476798e-01 -1.26395845e+00 -2.94961900e-01 -1.80436224e-01
-3.32952321e-01 1.61701870e+00 7.32937217e-01 -4.03894901e-01
1.00223923e+00 7.45434701e-01 7.61518598e-01 -5.23278654e-01
-2.28932306e-01 -2.11453170e-01 4.08026904e-01 9.85014915e-01
-3.02457094e-01 1.05698809e-01 1.38264805e-01 5.06509602e-01
-4.28842455e-02 7.07189217e-02 2.22159505e-01 9.28638697e-01
-5.94234467e-01 -7.94342697e-01 -4.66233015e-01 2.12500513e-01
-2.57650346e-01 5.54490566e-01 -6.22886777e-01 1.04624426e+00
3.43581736e-01 7.73894191e-01 -7.18426928e-02 -5.85955739e-01
4.73897070e-01 3.91084999e-02 3.34129244e-01 2.66792744e-01
-5.77227235e-01 -3.68941128e-01 2.71634072e-01 -9.10888672e-01
-1.71573907e-01 -6.56100690e-01 -1.27596629e+00 -1.80587545e-01
-4.61993635e-01 -4.34672713e-01 9.00943279e-01 1.02830470e+00
2.48320520e-01 3.71808887e-01 4.12668973e-01 -6.00859940e-01
-6.53034210e-01 -6.31565452e-01 -3.91103745e-01 2.99158961e-01
6.27094924e-01 -7.72078931e-01 -6.67390108e-01 -3.11417311e-01] | [8.01069164276123, -0.8027552962303162] |
ebcc982b-1863-483d-ac97-a7e3c371f1cc | discomat-distantly-supervised-composition | 2207.01079 | null | https://arxiv.org/abs/2207.01079v3 | https://arxiv.org/pdf/2207.01079v3.pdf | DiSCoMaT: Distantly Supervised Composition Extraction from Tables in Materials Science Articles | A crucial component in the curation of KB for a scientific domain is information extraction from tables in the domain's published articles -- tables carry important information (often numeric), which must be adequately extracted for a comprehensive machine understanding of an article. Existing table extractors assume prior knowledge of table structure and format, which may not be known in scientific tables. We study a specific and challenging table extraction problem: extracting compositions of materials (e.g., glasses, alloys). We first observe that materials science researchers organize similar compositions in a wide variety of table styles, necessitating an intelligent model for table understanding and composition extraction. Consequently, we define this novel task as a challenge for the ML community and create a training dataset comprising 4,408 distantly supervised tables, along with 1,475 manually annotated dev and test tables. We also present DiSCoMaT, a strong baseline geared towards this specific task, which combines multiple graph neural networks with several task-specific regular expressions, features, and constraints. We show that DiSCoMaT outperforms recent table processing architectures by significant margins. | ['Mausam', 'N. M. Anoop Krishnan', 'Mohd Zaki', 'Tanishq Gupta'] | 2022-07-03 | null | null | null | null | ['table-extraction'] | ['miscellaneous'] | [ 2.35470966e-01 3.22436512e-01 -3.60793710e-01 -2.78832883e-01
-1.04728639e+00 -1.11026490e+00 3.89989257e-01 1.06059492e+00
-1.56844586e-01 9.22911763e-01 4.14232284e-01 -6.35585010e-01
1.16528600e-01 -9.28351104e-01 -1.22418523e+00 -1.24970347e-01
1.41123995e-01 9.24255073e-01 -8.18646327e-02 -2.57077247e-01
4.58235592e-01 5.82341015e-01 -1.13651919e+00 6.58199728e-01
6.87857211e-01 1.27534294e+00 -2.50412852e-01 6.54464424e-01
-5.33199430e-01 1.14227247e+00 -6.45653069e-01 -8.70362282e-01
2.45667212e-02 -1.33924201e-01 -1.03976560e+00 -3.50354835e-02
6.37936234e-01 1.06115870e-01 -2.31819689e-01 9.77741539e-01
5.57146259e-02 -1.24641597e-01 9.68864441e-01 -1.16975713e+00
-6.41792655e-01 1.54374337e+00 -2.83329666e-01 2.36116126e-01
4.97370869e-01 -1.50373951e-01 1.61724293e+00 -7.86672652e-01
1.06304085e+00 9.01934147e-01 6.61113381e-01 1.11604705e-01
-1.09412611e+00 -4.20372784e-01 2.29795337e-01 2.17139736e-01
-1.28243911e+00 -7.46546507e-01 7.77912199e-01 -5.72263837e-01
1.31132579e+00 2.93931156e-01 5.68967700e-01 9.57534492e-01
2.87214339e-01 6.26635969e-01 6.64014697e-01 -3.65752876e-01
3.06331992e-01 4.15340602e-01 3.86751205e-01 6.28533363e-01
1.10269809e+00 -1.02263117e+00 -7.75056958e-01 9.45869163e-02
3.58417571e-01 -5.33134997e-01 -1.86157040e-02 -3.91465276e-01
-1.46116924e+00 5.56721330e-01 1.24098986e-01 1.18448265e-01
-1.77689996e-02 -1.64926603e-01 6.41969740e-01 2.96202779e-01
1.98881015e-01 1.12790990e+00 -9.15350616e-01 -9.97556597e-02
-3.90790850e-01 5.13595045e-01 1.43969035e+00 1.56695402e+00
6.60155892e-01 -1.27938703e-01 -6.42971881e-03 4.81454551e-01
1.23710878e-01 2.94356287e-01 -1.66072875e-01 -4.59639311e-01
1.17110682e+00 9.30282056e-01 -1.60581358e-02 -1.01979053e+00
-4.78391677e-01 -3.09743375e-01 -7.75695801e-01 -3.25211674e-01
6.64920986e-01 -3.59729305e-02 -8.79478395e-01 1.22958720e+00
3.23450953e-01 -8.08784544e-01 1.28483996e-01 3.85297060e-01
1.41077757e+00 5.27262151e-01 -6.35221880e-03 -1.06916338e-01
1.58365786e+00 -6.32488430e-01 -8.75139832e-01 -3.35798770e-01
6.89028263e-01 -7.59659767e-01 9.00005519e-01 7.74678051e-01
-1.05281746e+00 -1.42799690e-01 -1.33626330e+00 -8.46851528e-01
-1.18176043e+00 6.34474605e-02 9.63096738e-01 3.61037910e-01
-6.06132805e-01 8.03108633e-01 -4.84725684e-01 -1.06427558e-01
6.52327061e-01 4.15985703e-01 -5.68364978e-01 -5.82494475e-02
-9.80814099e-01 9.50671017e-01 4.71709549e-01 1.38925463e-01
-2.07568705e-01 -1.16608596e+00 -1.18426943e+00 1.11308523e-01
1.02584434e+00 -5.22722125e-01 1.11226678e+00 8.54991972e-02
-9.91589367e-01 9.93888140e-01 -1.10525444e-01 -3.18036765e-01
2.59499699e-01 -1.71974286e-01 -4.27920073e-01 -7.50702545e-02
2.71507025e-01 1.64395243e-01 3.58139694e-01 -1.20576167e+00
-3.53344440e-01 -2.95099169e-01 1.97922662e-01 -1.60768814e-02
-2.13060737e-01 -6.44080937e-02 -6.58387005e-01 -6.27500892e-01
1.23509057e-01 -6.62432015e-01 -4.53869738e-02 -1.51257768e-01
-1.23122048e+00 -3.14126998e-01 2.02085391e-01 -8.37125897e-01
1.32971942e+00 -1.64412594e+00 2.36231089e-01 4.29428875e-01
8.05455565e-01 -4.62143928e-01 3.79388750e-01 5.18751740e-01
-1.59053758e-01 5.29274881e-01 -2.31775403e-01 -1.31006792e-01
4.59953070e-01 -5.82844578e-02 -2.75628507e-01 3.96603383e-02
4.11093056e-01 1.02434719e+00 -5.62864542e-01 -8.46051514e-01
-2.63549238e-01 3.35036963e-02 -6.15673006e-01 9.88243222e-02
-9.61737037e-01 -5.97403608e-02 -5.50467730e-01 1.11890996e+00
5.06042957e-01 -6.84328675e-01 3.90198976e-01 -5.78140140e-01
6.97685704e-02 9.52633262e-01 -1.23238826e+00 1.65858006e+00
-1.85690075e-01 5.55556536e-01 -1.87660947e-01 -6.52949333e-01
6.81214154e-01 -1.35874629e-01 5.49582720e-01 -3.62309545e-01
3.33149731e-01 1.47947133e-01 6.22768290e-02 -3.75234842e-01
7.24110544e-01 1.53072745e-01 -5.04269719e-01 3.54598075e-01
1.61411688e-01 -4.46680278e-01 1.02905297e+00 5.23924530e-01
1.25263464e+00 8.75426605e-02 3.90099496e-01 -4.37196136e-01
4.80835527e-01 4.49827641e-01 4.64449793e-01 5.36874712e-01
3.70580018e-01 4.10845309e-01 1.21112549e+00 -6.47579372e-01
-1.04691744e+00 -9.01064217e-01 -1.85025424e-01 7.51017690e-01
-3.31784934e-01 -9.89348233e-01 -5.26528478e-01 -8.34880352e-01
4.09475833e-01 6.36067390e-01 -8.32944989e-01 4.95488085e-02
-6.87529325e-01 -7.09342003e-01 2.35820219e-01 6.25111938e-01
-5.37582152e-02 -8.01899433e-01 2.02316716e-02 1.22525975e-01
-1.11306243e-01 -1.58830261e+00 -3.92679662e-01 7.32975364e-01
-4.36022878e-01 -1.30618393e+00 6.74934015e-02 -7.58279979e-01
6.67514980e-01 -4.37948734e-01 1.90769100e+00 -1.87593877e-01
-2.33755544e-01 -1.13713846e-01 -1.30421340e-01 -9.94424284e-01
-6.58399463e-01 5.93998313e-01 -1.61878228e-01 -4.18741882e-01
5.95118165e-01 -4.26979244e-01 1.23498708e-01 -2.07150668e-01
-7.71922469e-01 2.06867844e-01 6.41239464e-01 2.66158193e-01
7.67586052e-01 2.25395173e-01 1.33461773e-01 -1.59649575e+00
4.23812240e-01 -4.48512375e-01 -7.72814393e-01 5.56980968e-01
-6.23642087e-01 5.17925322e-01 1.06494212e+00 -8.27125013e-02
-5.44309497e-01 -1.04082897e-01 -6.54618293e-02 1.27264590e-04
-2.34397594e-02 8.98345649e-01 -1.01878870e+00 9.58219990e-02
6.17884398e-01 -1.46517769e-01 -3.64914536e-01 -5.83318651e-01
4.59311217e-01 4.53673810e-01 5.99733055e-01 -9.77691233e-01
1.06883430e+00 2.86362078e-02 3.37830991e-01 -5.84124506e-01
-1.16239643e+00 -2.34570742e-01 -1.00361800e+00 2.23417670e-01
7.89059043e-01 -9.33128655e-01 -9.19970095e-01 1.37889534e-01
-1.06182623e+00 -2.16961086e-01 -3.03498924e-01 1.02508761e-01
-2.12895036e-01 7.10663944e-02 -8.20467293e-01 -3.12091231e-01
-3.24453861e-01 -9.51926827e-01 1.02240264e+00 -9.46977437e-02
-6.29860878e-01 -8.96886528e-01 -3.60445648e-01 4.99842525e-01
-1.13118328e-01 5.49829543e-01 1.65935612e+00 -1.02510500e+00
-1.03364325e+00 -1.47755057e-01 -3.88223708e-01 -1.59177363e-01
4.01678175e-01 2.53910750e-01 -7.96788931e-01 2.29509681e-01
-3.05859178e-01 -4.94651705e-01 8.59154463e-01 6.18759776e-03
1.42692363e+00 -5.35336971e-01 -3.40121031e-01 6.06002390e-01
1.28776634e+00 2.05152586e-01 5.00295818e-01 5.08939564e-01
1.41638923e+00 6.72965944e-01 1.55740425e-01 2.97485620e-01
5.73525369e-01 2.36075908e-01 1.19261481e-01 1.29647925e-01
7.38839954e-02 -3.66543233e-01 -1.67279735e-01 1.16928077e+00
1.58438161e-01 -4.91232574e-01 -1.08129954e+00 3.96996588e-01
-1.23360860e+00 -4.85967666e-01 -1.56635106e-01 1.71252370e+00
1.45652115e+00 7.67302990e-01 -5.48860393e-02 5.20704865e-01
1.04310602e-01 8.54059234e-02 -6.50243640e-01 -2.75328130e-01
-1.87298238e-01 2.32582510e-01 7.54052997e-01 2.39388540e-01
-1.30369925e+00 8.35213840e-01 6.33384228e+00 6.80316150e-01
-7.81317711e-01 -6.74435437e-01 7.67260969e-01 7.02205598e-02
-7.95767188e-01 -1.39712363e-01 -1.17746305e+00 2.32833236e-01
1.05444157e+00 -3.21445644e-01 3.68290365e-01 7.63663411e-01
-3.29980642e-01 -2.98010763e-02 -1.73714185e+00 1.02960789e+00
2.33200073e-01 -1.84066463e+00 5.83813965e-01 -2.17898865e-03
4.33016598e-01 -4.13667858e-01 -1.22397833e-01 3.05513829e-01
5.82995772e-01 -1.21910286e+00 8.74301016e-01 4.72260922e-01
9.45182383e-01 -7.44586468e-01 5.62276065e-01 -4.03822392e-01
-1.12890780e+00 3.68919969e-01 -2.41250187e-01 1.59378022e-01
-3.00597727e-01 8.97762477e-01 -8.60553920e-01 8.16402137e-01
6.44338131e-01 1.04286063e+00 -1.06443167e+00 4.48363215e-01
-1.20712072e-01 3.84951770e-01 -4.31410640e-01 -3.41237009e-01
-7.50633031e-02 -1.74002558e-01 1.52951017e-01 1.23738873e+00
-7.62584135e-02 -6.84924498e-02 -7.14170039e-02 1.02407598e+00
-9.48890805e-01 2.12073192e-01 -7.05908716e-01 -7.82269299e-01
2.51679540e-01 1.32280004e+00 -1.16321301e+00 -5.03133893e-01
-4.52828914e-01 2.62703508e-01 5.21022499e-01 1.18562177e-01
-5.39301038e-01 -7.44789064e-01 6.21330619e-01 1.00346997e-01
3.95612150e-01 -2.80689627e-01 -1.05914438e+00 -1.37927413e+00
4.39818561e-01 -1.19869924e+00 5.55926561e-01 -7.36944318e-01
-1.31380618e+00 5.34604788e-01 -1.12054199e-01 -1.00378025e+00
-2.99662441e-01 -1.15567434e+00 -5.73522411e-02 6.39144599e-01
-1.29929948e+00 -1.07801747e+00 -9.29610357e-02 2.11224943e-01
4.07414734e-01 -2.20338076e-01 6.13664091e-01 3.14139068e-01
-7.57569611e-01 6.75208271e-01 -1.06673986e-01 6.73116386e-01
7.88443565e-01 -1.81770802e+00 8.76641035e-01 7.13189542e-01
4.27853048e-01 8.80760491e-01 7.49118090e-01 -9.68768895e-01
-2.44251776e+00 -1.20089400e+00 1.30112803e+00 -1.30956459e+00
1.17492902e+00 -1.13701451e+00 -9.75455701e-01 9.00835514e-01
2.64649168e-02 -1.98091298e-01 7.52555668e-01 5.04310668e-01
-7.91447699e-01 -2.47600064e-01 -7.39564657e-01 7.06072927e-01
1.10478115e+00 -7.37551272e-01 -4.43832189e-01 7.12399721e-01
9.98247445e-01 -9.68301833e-01 -1.34922469e+00 1.61146373e-01
4.62472200e-01 -1.73245132e-01 9.26978946e-01 -9.24767077e-01
1.05298007e+00 -1.75912648e-01 -9.57168341e-02 -9.62936521e-01
-8.60748664e-02 -6.92016900e-01 -5.21531940e-01 1.31556368e+00
1.10417736e+00 -2.35155877e-02 9.75332499e-01 9.40655112e-01
-2.63656914e-01 -7.15746641e-01 -2.79573083e-01 -5.56586146e-01
3.65594000e-01 -6.03254139e-01 8.48005116e-01 1.02854288e+00
2.43270487e-01 8.46769452e-01 1.62341597e-03 6.29379530e-04
5.48032284e-01 2.80635744e-01 8.19139123e-01 -1.39304888e+00
2.99833901e-02 -4.52044338e-01 -1.40839279e-01 -5.92932343e-01
6.56429306e-02 -1.17285883e+00 -5.28629720e-02 -1.88165462e+00
7.35723749e-02 -2.98520118e-01 -1.81962520e-01 5.87035418e-01
-7.57028759e-02 -1.41686112e-01 -3.64522822e-02 -4.81910043e-04
-6.26435697e-01 1.30193859e-01 1.23905110e+00 -6.08820260e-01
-2.05447581e-02 -4.32419837e-01 -1.35817409e+00 5.29979706e-01
4.98163730e-01 -5.44299185e-01 -3.71336073e-01 -3.47397983e-01
9.65383053e-01 -1.79303035e-01 -2.39419177e-01 -9.67663586e-01
3.68206918e-01 -3.83810908e-01 7.85863578e-01 -8.54304373e-01
-3.27303596e-02 -6.68710470e-01 -1.21485859e-01 -1.86573222e-01
-5.21825731e-01 2.67854840e-01 4.62715775e-01 3.26784343e-01
-4.47309874e-02 8.56173486e-02 9.12780613e-02 -2.70580620e-01
-4.33841556e-01 4.07074988e-01 -1.80434614e-01 5.96280575e-01
4.37738180e-01 1.65932626e-01 -5.54206550e-01 3.08740009e-02
-4.88103002e-01 2.74871647e-01 3.88101816e-01 4.47176009e-01
3.91339868e-01 -1.05955184e+00 -5.94343364e-01 1.51157305e-01
4.63172436e-01 4.20011759e-01 -2.85205424e-01 3.71715367e-01
-7.61649370e-01 5.31815350e-01 -9.13964435e-02 -7.29663149e-02
-8.28954995e-01 8.78243804e-01 -2.79781938e-01 -5.58613360e-01
-5.21305263e-01 6.95099652e-01 -2.61075765e-01 -3.72411489e-01
1.98222145e-01 -1.20264506e+00 -3.66076022e-01 3.34672630e-01
3.48525524e-01 -2.03163743e-01 6.89774096e-01 -2.09423348e-01
-4.75313574e-01 3.68099898e-01 -5.41540265e-01 2.80016959e-01
1.50759566e+00 1.89210236e-01 -3.66322786e-01 9.11727309e-01
1.12581909e+00 4.65903938e-01 -5.30938804e-01 -1.92299888e-01
4.84504908e-01 1.78357959e-02 -3.31684530e-01 -9.91180778e-01
-9.04680014e-01 5.27186453e-01 -5.40429175e-01 2.67654806e-01
6.61629677e-01 1.73408613e-01 7.66767859e-01 1.00698864e+00
1.16404787e-01 -1.11202633e+00 -3.03260863e-01 5.68497896e-01
9.91170287e-01 -1.38033330e+00 6.38469696e-01 -9.84047651e-01
-3.97817165e-01 1.31784332e+00 6.15309954e-01 3.81880403e-01
6.35142386e-01 7.87511349e-01 1.17277447e-02 -4.84487325e-01
-9.30262923e-01 -9.14050266e-02 6.72978222e-01 4.13139790e-01
6.16054177e-01 -2.74251491e-01 2.33215079e-01 8.98181319e-01
-8.20291638e-01 -2.77661771e-01 6.90055847e-01 1.21675479e+00
-8.82427469e-02 -1.08792627e+00 -1.86644346e-01 1.06728470e+00
-5.55741787e-01 -3.57985407e-01 -8.35098982e-01 9.11154389e-01
1.52157536e-02 6.74476922e-01 3.17905545e-02 -3.07427496e-01
4.83396798e-01 1.39008760e-01 4.49237555e-01 -6.80105746e-01
-4.52304900e-01 -2.11382687e-01 7.30069339e-01 -2.65234679e-01
-1.26231432e-01 -6.25015676e-01 -1.29114878e+00 -4.83172536e-01
2.33921602e-01 3.73356968e-01 5.93561888e-01 1.01857769e+00
9.00127739e-02 9.27133620e-01 6.74955100e-02 -1.86526150e-01
-1.39328346e-01 -7.85727859e-01 -6.34127975e-01 4.00654018e-01
3.67465883e-01 -4.83503938e-01 -2.20449597e-01 4.11093473e-01] | [9.608098030090332, 7.907163619995117] |
f859104f-c974-4aa0-9700-53eed032f935 | interpretable-simultaneous-localization-of | 2306.00473 | null | https://arxiv.org/abs/2306.00473v1 | https://arxiv.org/pdf/2306.00473v1.pdf | Interpretable simultaneous localization of MRI corpus callosum and classification of atypical Parkinsonian disorders using YOLOv5 | Structural MRI(S-MRI) is one of the most versatile imaging modality that revolutionized the anatomical study of brain in past decades. The corpus callosum (CC) is the principal white matter fibre tract, enabling all kinds of inter-hemispheric communication. Thus, subtle changes in CC might be associated with various neurological disorders. The present work proposes the potential of YOLOv5-based CC detection framework to differentiate atypical Parkinsonian disorders (PD) from healthy controls (HC). With 3 rounds of hold-out validation, mean classification accuracy of 92% is obtained using the proposed method on a proprietary dataset consisting of 20 healthy subjects and 20 cases of APDs, with an improvement of 5% over SOTA methods (CC morphometry and visual texture analysis) that used the same dataset. Subsequently, in order to incorporate the explainability of YOLO predictions, Eigen CAM based heatmap is generated for identifying the most important sub-region in CC that leads to the classification. The result of Eigen CAM showed CC mid-body as the most distinguishable sub-region in classifying APDs and HC, which is in-line with SOTA methodologies and the current prevalent understanding in medicine. | ['Sandhya M', 'Pramod Kumar Pal', 'Jitender Saini', 'Neelam Sinha', 'Debanjali Bhattacharya', 'Vamshi Krishna Kancharla'] | 2023-06-01 | null | null | null | null | ['texture-classification'] | ['computer-vision'] | [-1.61205113e-01 2.37559676e-01 -9.66382548e-02 9.31675360e-03
-8.67695436e-02 -1.31069615e-01 6.09731257e-01 -1.95004195e-01
-2.58056760e-01 7.98158586e-01 4.69377041e-01 -6.20441139e-03
-2.32534319e-01 -4.30839479e-01 7.78451649e-05 -7.86898434e-01
-6.42272234e-01 5.27019083e-01 1.63627341e-01 1.25991497e-02
4.88376766e-01 5.43587744e-01 -1.18108296e+00 2.97159523e-01
1.00044751e+00 9.44241166e-01 4.69911456e-01 1.10362217e-01
5.27488650e-04 2.18741536e-01 -2.60217160e-01 -1.57432139e-01
1.42893568e-01 -4.84466940e-01 -8.70340407e-01 6.68619797e-02
2.20714048e-01 -1.66525006e-01 6.93689734e-02 1.03153896e+00
5.97092032e-01 -2.27664903e-01 9.56833065e-01 -6.35696709e-01
-7.21960843e-01 4.88653898e-01 -7.53516853e-01 6.47728801e-01
2.83162832e-01 3.32133770e-01 7.21742511e-01 -5.66169202e-01
1.18413126e+00 1.29160583e+00 5.12185395e-01 5.43896437e-01
-1.23444033e+00 -4.05094713e-01 -1.22056991e-01 8.74786139e-01
-8.75909209e-01 -1.47995949e-01 6.20883048e-01 -8.45406055e-01
9.19633925e-01 2.78012633e-01 1.34402251e+00 9.11509573e-01
8.29378128e-01 4.84062195e-01 1.93087256e+00 -1.60027504e-01
3.34702373e-01 -3.42661381e-01 1.89454153e-01 3.72412086e-01
3.08761090e-01 5.28667159e-02 -9.11173001e-02 -2.23045170e-01
4.78833288e-01 -2.33983144e-01 -3.47867787e-01 -3.39868903e-01
-1.48054159e+00 9.13961411e-01 3.86504292e-01 7.08966613e-01
-3.77677321e-01 -2.59424657e-01 7.08278656e-01 3.78278270e-02
3.98976743e-01 1.22990355e-01 -1.90976575e-01 5.34145385e-02
-1.22085071e+00 1.07692271e-01 1.07787095e-01 1.04930222e-01
-2.23800510e-01 1.82433605e-01 -2.56770909e-01 9.21363294e-01
5.19578457e-01 5.49448907e-01 1.10186660e+00 -8.56003404e-01
2.96972692e-01 8.16315591e-01 -5.18884540e-01 -8.06943476e-01
-8.09009612e-01 -6.17018700e-01 -9.81644392e-01 5.22153795e-01
2.36136869e-01 -7.81258196e-03 -7.63366818e-01 1.30438542e+00
3.65241587e-01 -4.82045501e-01 -3.66739422e-01 1.19182169e+00
3.92427176e-01 -1.83433533e-01 1.55663490e-01 -1.89718932e-01
1.63982129e+00 -5.42104185e-01 -5.36991417e-01 -1.87448293e-01
4.13533956e-01 -3.50704908e-01 8.47428441e-01 2.21974865e-01
-7.14795947e-01 -2.17156142e-01 -8.61154974e-01 3.00496131e-01
-2.24051759e-01 3.56677234e-01 5.47251403e-01 5.75537801e-01
-1.02360582e+00 6.30429089e-01 -9.16242421e-01 -7.42635906e-01
5.86101234e-01 2.04631343e-01 -7.72402942e-01 3.00810426e-01
-8.74651134e-01 1.47559416e+00 3.81243557e-01 2.87352622e-01
-7.40083039e-01 -5.62885165e-01 -1.64229065e-01 -4.65275764e-01
-1.11593381e-01 -6.60006583e-01 2.74694264e-01 -6.77081883e-01
-1.19629061e+00 1.07349098e+00 -3.09130624e-02 -5.05610049e-01
7.77502358e-01 -5.65356500e-02 -5.95312774e-01 7.43250370e-01
2.83123195e-01 6.98004007e-01 8.71914625e-01 -1.02833450e+00
-1.39510944e-01 -1.15179551e+00 -4.41361547e-01 -9.74193141e-02
2.37787962e-01 1.06488504e-01 3.53140414e-01 -8.09664905e-01
4.63676363e-01 -1.06455195e+00 1.15240468e-02 1.55599713e-01
-6.97073221e-01 1.28200063e-02 9.56454754e-01 -1.49379480e+00
9.33071136e-01 -1.90809059e+00 2.95271605e-01 3.25332016e-01
8.12295139e-01 2.90110218e-03 3.68554682e-01 7.95104951e-02
-3.84025484e-01 5.35592809e-02 -6.53581023e-01 1.65679634e-01
-7.48024508e-02 -1.56641200e-01 2.85287112e-01 1.00824392e+00
8.06928352e-02 8.45164418e-01 -4.97060061e-01 -4.10162449e-01
1.64225906e-01 4.12527621e-01 -2.16641903e-01 -2.96116978e-01
1.94139495e-01 1.03617203e+00 -2.91084349e-01 7.66885102e-01
6.42981112e-01 2.00307280e-01 4.05661941e-01 -3.09984535e-01
-9.67173129e-02 -2.12433532e-01 -7.05594242e-01 1.46300399e+00
1.69759035e-01 5.78474641e-01 6.87872618e-02 -1.06642580e+00
8.19344878e-01 3.07836711e-01 5.34441233e-01 -1.02670288e+00
2.00281546e-01 2.83866286e-01 6.43042386e-01 -6.88079894e-01
-4.16725963e-01 -1.29887491e-01 5.12311161e-01 3.27478588e-01
-1.68092594e-01 2.42005065e-01 2.16410682e-01 -1.80368766e-01
7.67721236e-01 2.72149771e-01 3.21195841e-01 -7.53214419e-01
7.30955362e-01 1.21447191e-01 3.96223575e-01 -3.49470763e-03
-7.97215939e-01 4.85331208e-01 6.74139798e-01 -3.57459068e-01
-1.17149270e+00 -1.09221566e+00 -6.44178987e-01 1.81850418e-01
-2.84906358e-01 2.23679677e-01 -1.08868480e+00 -3.78453523e-01
1.05656080e-01 5.20359516e-01 -9.03999448e-01 -6.55397177e-02
-3.74237388e-01 -1.18076456e+00 4.29374546e-01 2.63703138e-01
7.49940276e-01 -1.08288991e+00 -1.00794983e+00 6.67614639e-02
-4.92108464e-02 -7.36183286e-01 -9.52768102e-02 -1.02669083e-01
-1.43125439e+00 -1.46234024e+00 -1.22706437e+00 -4.76085871e-01
6.23056412e-01 -1.61115959e-01 5.79457164e-01 -2.63026863e-01
-7.53439426e-01 2.91711539e-02 -2.55832165e-01 2.64786631e-01
-3.00668448e-01 -3.32350492e-01 2.19838187e-01 -1.19682550e-01
2.92170882e-01 -7.62529075e-01 -1.07959473e+00 1.28916934e-01
-5.83709121e-01 -2.97344737e-02 8.06410730e-01 5.40538967e-01
6.06297076e-01 -3.31894040e-01 5.90865493e-01 -4.09345120e-01
7.08267629e-01 -6.43262088e-01 -1.07582502e-01 1.15723409e-01
-7.26858079e-01 5.45931235e-02 2.33582050e-01 -2.62369633e-01
-7.45327234e-01 -2.35540763e-01 1.16404906e-01 9.92112886e-03
-3.93775910e-01 4.85143840e-01 -1.79267339e-02 -7.57049099e-02
4.33154970e-01 2.21413240e-01 4.92828846e-01 -6.11568511e-01
2.71372914e-01 4.75879997e-01 3.52187544e-01 -2.10945383e-01
1.46809757e-01 6.57081783e-01 7.74512067e-02 -8.47080827e-01
-2.07537804e-02 -2.57812142e-01 -1.24828076e+00 -4.76647735e-01
1.46050775e+00 -4.71049219e-01 -5.14019072e-01 2.89935380e-01
-7.53120959e-01 -1.93823814e-01 7.60421604e-02 7.73433805e-01
-4.69839841e-01 6.92639172e-01 -3.76593828e-01 -6.11640990e-01
-8.60574782e-01 -1.32921159e+00 7.73749292e-01 -2.73144394e-01
-4.79438484e-01 -9.05896008e-01 3.66000086e-01 5.34193754e-01
4.72697824e-01 6.15656793e-01 1.48832750e+00 -4.06525493e-01
1.78265721e-01 7.45418817e-02 -2.78931975e-01 1.19851165e-01
3.53654549e-02 -9.79858041e-02 -7.33231902e-01 -1.16004959e-01
2.84326911e-01 1.01238519e-01 6.08610272e-01 7.03349233e-01
6.13658190e-01 -4.33700755e-02 -2.65630722e-01 1.21669032e-01
1.42945516e+00 4.73075777e-01 7.52301872e-01 7.32262135e-01
5.68143606e-01 7.23776817e-01 1.15834951e-01 1.75042376e-01
3.47536981e-01 7.97301888e-01 5.33804357e-01 1.05748877e-01
-4.83641118e-01 5.25818467e-01 5.70136428e-01 8.59607399e-01
-6.92018569e-01 7.50596762e-01 -9.40117300e-01 4.38605070e-01
-1.34354186e+00 -1.03210795e+00 -5.71240544e-01 2.06991482e+00
4.25416678e-01 3.64795923e-02 6.34388804e-01 2.01921701e-01
8.88498962e-01 -9.82025489e-02 -7.19236314e-01 -3.43673199e-01
-2.44207636e-01 1.28111809e-01 3.68186414e-01 5.96054271e-02
-7.48070359e-01 4.08348382e-01 6.78593159e+00 5.60672402e-01
-1.41315973e+00 3.86521190e-01 5.92407823e-01 -6.14559762e-02
-1.08162751e-02 -2.53600925e-01 -1.73640579e-01 7.51547515e-01
9.35252845e-01 1.99672267e-01 5.05061328e-01 5.57729065e-01
8.08116436e-01 -5.14255702e-01 -5.05084455e-01 6.84758127e-01
2.50488687e-02 -8.69316697e-01 -1.47222847e-01 4.23139244e-01
3.46536100e-01 3.11507612e-01 -1.70500442e-01 -2.81111717e-01
-4.55264658e-01 -9.29113865e-01 8.73928547e-01 1.00920022e+00
9.69957054e-01 -3.59798491e-01 8.09137046e-01 1.40790995e-02
-1.02862000e+00 -1.03226893e-01 -2.48909965e-01 2.62445122e-01
1.82618678e-01 7.41584003e-01 -6.66259646e-01 5.07020295e-01
8.22762907e-01 5.97526133e-01 -7.78927743e-01 1.23407531e+00
-3.59364673e-02 5.28590620e-01 4.11258414e-02 3.58750552e-01
2.35588793e-02 -6.81673706e-01 8.95082235e-01 9.90559578e-01
4.72180426e-01 -2.36021906e-01 -3.09042931e-01 1.25821221e+00
6.13155186e-01 3.51670086e-01 -4.33058470e-01 -1.19661108e-01
7.19963834e-02 1.14243460e+00 -1.05421972e+00 -2.98895538e-01
-2.52552509e-01 1.00367475e+00 6.11896515e-02 6.40774220e-02
-4.82199192e-01 6.05694205e-02 3.43857199e-01 2.93629795e-01
6.98046535e-02 -3.52151006e-01 -5.75242519e-01 -1.06302106e+00
7.41087645e-02 -7.74285674e-01 1.10948205e-01 -6.95354581e-01
-1.27445388e+00 6.92841232e-01 7.06866458e-02 -1.27758563e+00
9.03909355e-02 -6.28709674e-01 -6.81043446e-01 9.92277980e-01
-1.01528978e+00 -1.11931610e+00 -1.75236557e-02 3.88956696e-01
2.79404640e-01 -2.31726184e-01 9.66915727e-01 1.24680258e-01
-6.78780794e-01 -9.18128341e-02 2.80990064e-01 -1.42328367e-01
3.63810301e-01 -1.38749671e+00 -6.53969795e-02 6.30454183e-01
-5.41262984e-01 6.41984761e-01 6.20384455e-01 -1.10178161e+00
-9.72318172e-01 -7.73300350e-01 6.30981743e-01 -4.62695286e-02
9.20011759e-01 -4.81905006e-02 -6.96699619e-01 1.62669167e-01
2.28713915e-01 -2.36656100e-01 7.32393444e-01 -4.75266129e-01
-1.05410265e-02 2.26168200e-01 -1.58613420e+00 4.38353151e-01
8.44272017e-01 -4.55159068e-01 -1.12719703e+00 2.77518332e-01
-1.19200602e-01 1.33929566e-01 -1.35183537e+00 1.62135035e-01
8.95547807e-01 -1.22731185e+00 9.07125652e-01 -2.78426737e-01
3.98670018e-01 -2.89650261e-01 5.18215187e-02 -1.33595181e+00
-3.25418323e-01 1.49555147e-01 -1.11706831e-01 6.88968241e-01
1.45062849e-01 -7.21490681e-01 5.32578170e-01 4.43702817e-01
-2.51810074e-01 -7.97687054e-01 -1.08011258e+00 -7.80158401e-01
2.12576956e-01 -3.01748812e-01 2.44220793e-01 8.23617458e-01
2.08086371e-01 -2.56432086e-01 7.18685314e-02 -3.36566806e-01
7.67017424e-01 -1.07003585e-01 -3.95992398e-02 -1.51073444e+00
3.07157659e-03 -8.60321403e-01 -8.25107992e-01 1.74614683e-01
1.07014664e-02 -1.28035975e+00 -5.71848810e-01 -1.85853994e+00
4.17948425e-01 -1.76125273e-01 5.53492866e-02 3.45865637e-01
1.55848116e-01 4.94171441e-01 2.45636597e-01 6.57175004e-01
1.43608317e-01 4.41738546e-01 1.63933372e+00 -1.71630457e-01
-1.40028358e-01 -3.97702277e-01 -4.05848265e-01 6.60289049e-01
9.37027633e-01 -2.41092280e-01 -3.06169927e-01 -4.37239707e-02
-2.96619684e-01 -2.46966347e-01 5.29469788e-01 -1.17033374e+00
-5.36677837e-01 8.23848173e-02 7.26312757e-01 -4.42308128e-01
6.45839721e-02 -7.89560795e-01 4.03819829e-01 1.07268286e+00
5.52439019e-02 1.98081657e-01 2.01420067e-03 2.52751857e-01
5.01364842e-02 -1.16050236e-01 9.51741457e-01 -3.73308122e-01
-5.63638210e-01 1.58065166e-02 -6.12355173e-01 -1.78769067e-01
1.15078568e+00 -6.56250775e-01 -4.19366628e-01 1.95609540e-01
-1.34841681e+00 -2.92176455e-01 2.63551891e-01 1.26249492e-01
4.28246200e-01 -1.23254526e+00 -6.29275382e-01 -1.07281059e-01
-2.04806954e-01 -7.39831150e-01 4.98759359e-01 1.88934886e+00
-7.86735415e-01 7.16555119e-01 -1.09777653e+00 -6.00137472e-01
-1.04915679e+00 1.79332763e-01 3.22302133e-01 -1.66062206e-01
-1.31257284e+00 1.89757377e-01 -1.80955186e-01 -1.02428690e-01
-1.94607154e-01 -4.54909325e-01 -8.64702404e-01 3.75033736e-01
4.07837927e-01 8.77618372e-01 1.52126491e-01 -1.04979634e+00
-5.22082210e-01 6.93157494e-01 1.83912858e-01 -1.69613943e-01
1.45806301e+00 -4.21894789e-01 -6.25091016e-01 5.22217929e-01
9.40622091e-01 3.76459211e-02 -9.16928709e-01 5.67038178e-01
3.63104314e-01 -4.51019071e-02 1.77172512e-01 -1.17501128e+00
-1.45262790e+00 9.55031037e-01 1.55977547e+00 1.53842926e-01
8.00588071e-01 -4.91474345e-02 7.05474496e-01 -3.12847257e-01
5.50182402e-01 -1.11084545e+00 -5.39322019e-01 -5.83418123e-02
1.25284028e+00 -9.24051106e-01 9.71260220e-02 -1.15077734e-01
-9.36845124e-01 1.41163003e+00 2.24745646e-01 -4.33201522e-01
5.53965151e-01 -3.90357167e-01 -2.09407806e-02 -7.18400359e-01
-1.40631855e-01 -1.03785910e-01 6.08689249e-01 8.83674502e-01
5.36670387e-01 4.62662518e-01 -1.14617920e+00 5.86724102e-01
-2.50731528e-01 -8.09198916e-02 1.00197650e-01 6.67492867e-01
-5.62231719e-01 -9.77347672e-01 -5.90516269e-01 1.04700625e+00
-5.41525781e-01 7.85788521e-02 -5.65183938e-01 9.09036815e-01
4.39083666e-01 4.83152568e-01 -1.05249554e-01 -2.26292029e-01
-2.08403096e-02 3.64928931e-01 4.92130220e-01 -3.26882958e-01
-4.63845372e-01 1.88566715e-01 1.20173067e-01 -6.56294465e-01
-6.26161337e-01 -9.61620152e-01 -1.55640352e+00 -1.38528481e-01
-6.62224041e-03 -2.00480372e-01 7.48138726e-01 1.42340231e+00
3.35591674e-01 4.67814744e-01 1.74696773e-01 -8.43246758e-01
-5.07606601e-04 -1.36500156e+00 -8.60497117e-01 4.79402721e-01
5.68514084e-03 -1.20853102e+00 -3.27506870e-01 7.55642802e-02] | [14.100139617919922, -1.9926878213882446] |
b85d7caf-75f2-49d7-9a40-958e3f227452 | ladi-vton-latent-diffusion-textual-inversion | 2305.13501 | null | https://arxiv.org/abs/2305.13501v2 | https://arxiv.org/pdf/2305.13501v2.pdf | LaDI-VTON: Latent Diffusion Textual-Inversion Enhanced Virtual Try-On | The rapidly evolving fields of e-commerce and metaverse continue to seek innovative approaches to enhance the consumer experience. At the same time, recent advancements in the development of diffusion models have enabled generative networks to create remarkably realistic images. In this context, image-based virtual try-on, which consists in generating a novel image of a target model wearing a given in-shop garment, has yet to capitalize on the potential of these powerful generative solutions. This work introduces LaDI-VTON, the first Latent Diffusion textual Inversion-enhanced model for the Virtual Try-ON task. The proposed architecture relies on a latent diffusion model extended with a novel additional autoencoder module that exploits learnable skip connections to enhance the generation process preserving the model's characteristics. To effectively maintain the texture and details of the in-shop garment, we propose a textual inversion component that can map the visual features of the garment to the CLIP token embedding space and thus generate a set of pseudo-word token embeddings capable of conditioning the generation process. Experimental results on Dress Code and VITON-HD datasets demonstrate that our approach outperforms the competitors by a consistent margin, achieving a significant milestone for the task. Source code and trained models will be publicly released at: https://github.com/miccunifi/ladi-vton. | ['Rita Cucchiara', 'Marco Bertini', 'Marcella Cornia', 'Giuseppe Cartella', 'Alberto Baldrati', 'Davide Morelli'] | 2023-05-22 | null | null | null | null | ['virtual-try-on'] | ['computer-vision'] | [ 2.14740455e-01 1.96917892e-01 -1.32372573e-01 -2.69242257e-01
-5.88055372e-01 -3.25978279e-01 8.71212602e-01 -4.14615512e-01
4.84371148e-02 2.94464141e-01 2.35466331e-01 7.85742104e-02
5.28786518e-02 -9.53203738e-01 -9.34365451e-01 -8.76234412e-01
1.58554778e-01 3.23267609e-01 -3.73450398e-01 -4.35234129e-01
-3.65607627e-02 -2.24307994e-03 -1.40164590e+00 5.32565594e-01
6.46282494e-01 1.15160489e+00 5.14343798e-01 5.81374526e-01
1.81123540e-01 6.80115759e-01 -2.33196229e-01 -8.29524100e-01
2.71412164e-01 -8.78714800e-01 -2.32465386e-01 5.06165206e-01
2.13871956e-01 -3.36831480e-01 -6.06892705e-01 9.71859038e-01
5.10940075e-01 2.69170292e-02 5.28998494e-01 -1.14298415e+00
-1.74759555e+00 7.78613031e-01 -4.53985304e-01 -2.21226811e-01
2.99490303e-01 3.16742003e-01 9.86077547e-01 -1.01196980e+00
9.22521710e-01 1.20390272e+00 4.10583675e-01 6.95038140e-01
-1.34651494e+00 -5.24572074e-01 9.05414000e-02 3.12844276e-01
-1.48552871e+00 -1.89909369e-01 1.18277884e+00 -1.18682556e-01
6.07101440e-01 7.46524036e-02 9.74882722e-01 1.66953814e+00
3.67948532e-01 9.67477083e-01 1.13221979e+00 -4.36198145e-01
1.82344049e-01 3.73705208e-01 -5.65145612e-01 7.55404294e-01
-1.75696582e-01 2.22410396e-01 -6.36415899e-01 3.36526364e-01
9.91522729e-01 1.58063695e-01 -8.91299322e-02 -4.77292478e-01
-9.55860496e-01 1.18530560e+00 7.09332764e-01 3.10736567e-01
-6.98321640e-01 3.31949860e-01 7.28838071e-02 1.98099673e-01
6.88250005e-01 7.47918487e-02 7.54193440e-02 8.88750404e-02
-9.15390611e-01 3.14375162e-01 6.14852905e-01 1.02669871e+00
5.06204128e-01 1.89568713e-01 -2.16553584e-01 7.69310892e-01
4.83712584e-01 4.71947819e-01 5.07000268e-01 -6.03874147e-01
2.18995333e-01 5.25811136e-01 -1.23340026e-01 -8.50039721e-01
2.02784345e-01 -6.59562469e-01 -8.87821019e-01 1.40705436e-01
-2.02823132e-01 9.79431421e-02 -1.24467778e+00 1.52236176e+00
2.71335334e-01 1.42808437e-01 -9.82097983e-02 1.01715553e+00
6.88822985e-01 9.58532155e-01 7.76297525e-02 2.62888879e-01
1.19072378e+00 -1.20930123e+00 -7.09395349e-01 -2.60297000e-01
1.03226520e-01 -9.67387974e-01 1.11598158e+00 3.06944847e-01
-1.19934702e+00 -8.89316916e-01 -1.28759718e+00 -1.36267573e-01
-3.58901978e-01 2.64917817e-02 7.17880368e-01 6.41592383e-01
-1.00510502e+00 5.06727397e-01 -6.19668961e-01 -3.09205920e-01
5.56834459e-01 3.77840549e-02 -1.64427385e-01 -3.95325452e-01
-1.23791277e+00 6.36701643e-01 3.10348630e-01 4.12717819e-01
-1.22340465e+00 -6.02367282e-01 -1.06640804e+00 -7.96388239e-02
2.24370047e-01 -9.63946581e-01 9.27185416e-01 -1.27360976e+00
-1.75043488e+00 7.95336604e-01 3.15605015e-01 -4.40479279e-01
6.35746360e-01 -1.78740755e-01 -4.50053692e-01 3.73473763e-02
2.60415971e-02 9.45489287e-01 1.26951587e+00 -1.55673134e+00
-2.61316329e-01 -2.10938334e-01 -3.06656410e-04 1.93259940e-01
-3.40405613e-01 -3.58621299e-01 -6.80763781e-01 -9.24408078e-01
-5.22759706e-02 -1.16563761e+00 -2.20346078e-01 -2.37367302e-02
-2.36537874e-01 1.87828932e-02 7.73149431e-01 -8.22681427e-01
9.15288746e-01 -2.34287453e+00 3.77639860e-01 1.50049433e-01
1.31660372e-01 2.50974119e-01 -4.28850979e-01 6.59988999e-01
-3.34146060e-02 -7.18156770e-02 -2.08592355e-01 -6.40833378e-01
2.33899981e-01 3.06865066e-01 -2.77888596e-01 2.79503912e-01
2.53706634e-01 1.44164717e+00 -8.01696777e-01 -1.94759890e-01
4.98247892e-01 1.14627898e+00 -6.76703990e-01 1.03232495e-01
-3.24762702e-01 3.56809705e-01 -2.33011141e-01 4.96431530e-01
6.98320091e-01 -2.84307271e-01 2.74079740e-01 -2.48603180e-01
1.01522051e-01 -1.59511387e-01 -9.76729691e-01 2.12766123e+00
-7.31020689e-01 3.64578307e-01 -1.59520686e-01 -7.19381928e-01
9.78221178e-01 1.70683697e-01 3.38255882e-01 -1.13492417e+00
4.21400070e-01 1.47166193e-01 -1.63717940e-01 -4.54666704e-01
5.74301004e-01 -3.09137344e-01 -2.28204988e-02 2.52480000e-01
3.68448406e-01 -1.16706543e-01 1.90040097e-01 3.44230205e-01
3.89698148e-01 6.22813106e-01 -1.41297266e-01 2.86777541e-02
9.28146392e-02 -3.41760486e-01 1.23852827e-01 5.28403819e-01
7.55724162e-02 8.52253318e-01 1.66219361e-02 -3.08403581e-01
-1.40954518e+00 -1.32051551e+00 1.15309566e-01 8.37541103e-01
2.90298700e-01 -4.17606175e-01 -8.41721177e-01 -4.47090596e-01
-1.05020553e-01 9.64562118e-01 -9.64562237e-01 -3.57806742e-01
-3.18226427e-01 -6.96705997e-01 1.07564993e-01 4.72732425e-01
6.99878931e-01 -1.14942372e+00 -2.40420625e-01 3.09278578e-01
-3.34188104e-01 -9.53673780e-01 -5.28151333e-01 -8.60308185e-02
-5.62675536e-01 -4.04103547e-01 -1.04960096e+00 -1.01560175e+00
6.37670994e-01 1.52645335e-01 8.64890516e-01 -2.34555095e-01
-5.01905262e-01 4.16208595e-01 -6.40730560e-01 -2.52234578e-01
-6.23425663e-01 9.43585411e-02 -3.73179942e-01 5.59217930e-01
2.26553187e-01 -6.71219230e-01 -9.21947479e-01 1.42119765e-01
-1.23199594e+00 3.99400562e-01 7.89646029e-01 9.20010805e-01
6.06494129e-01 1.33990124e-02 4.49284822e-01 -6.28111184e-01
5.99805832e-01 -5.86169958e-01 -1.29127786e-01 9.52919275e-02
-5.78238845e-01 -4.96405549e-02 3.39163184e-01 -6.56130552e-01
-1.21203327e+00 -5.72791621e-02 -4.22296375e-01 -5.01937091e-01
1.50661469e-01 5.42981267e-01 -1.51628971e-01 1.73248872e-01
5.19801140e-01 4.57055062e-01 4.23204787e-02 -4.74554867e-01
8.84680569e-01 4.82439786e-01 3.59168619e-01 -2.39359096e-01
8.97334218e-01 5.93389630e-01 -4.18583483e-01 -6.66987598e-01
-6.26368403e-01 -1.09475017e-01 -3.87156337e-01 -4.32516903e-01
9.10613120e-01 -9.88615930e-01 -5.26900709e-01 5.80250382e-01
-8.24180782e-01 -5.05798757e-01 -6.46874487e-01 3.11019421e-01
-6.40038252e-01 2.18541235e-01 -6.79436505e-01 -6.34570599e-01
-3.05856645e-01 -9.49965298e-01 1.06150603e+00 3.63574594e-01
3.68002653e-02 -1.05471134e+00 6.72974288e-02 5.10804892e-01
5.60007751e-01 2.66127467e-01 6.30725801e-01 -3.09239049e-02
-8.28504801e-01 -3.20483536e-01 -4.23808582e-02 7.39391804e-01
-1.26142558e-02 -2.60402232e-01 -9.54415917e-01 -3.31886470e-01
8.54245871e-02 -3.64090562e-01 1.08203256e+00 4.96476263e-01
8.63232195e-01 -3.62287104e-01 -8.37076828e-02 6.35406196e-01
1.56924331e+00 9.48165059e-02 1.09738255e+00 2.16143981e-01
6.18450880e-01 4.06167030e-01 3.12502205e-01 5.29317081e-01
4.81922328e-01 6.62267268e-01 4.57105637e-01 -4.80608910e-01
-6.75852954e-01 -8.63520682e-01 4.85513866e-01 9.41646099e-01
-2.53934562e-01 -4.91891176e-01 -2.86543697e-01 6.51385784e-01
-1.69174480e+00 -1.04494107e+00 1.58186957e-01 1.74968266e+00
6.56057775e-01 -3.91948270e-03 -1.31102800e-01 -8.37395266e-02
6.55527651e-01 3.09555024e-01 -5.43822646e-01 -4.15531486e-01
-2.89147317e-01 3.65666121e-01 2.16612667e-01 4.25019532e-01
-8.71847272e-01 9.84218419e-01 5.30989456e+00 9.63322401e-01
-1.14342988e+00 2.86453784e-01 7.57450700e-01 -1.05897874e-01
-4.70516980e-01 -2.05739811e-01 -6.10820413e-01 3.78595978e-01
6.96932316e-01 -9.96992961e-02 6.21939063e-01 9.44185793e-01
9.48943496e-02 2.09676772e-01 -7.95916915e-01 1.01927757e+00
6.23870671e-01 -1.35358989e+00 2.98865706e-01 3.03052157e-01
1.04572737e+00 -2.50523359e-01 9.45994973e-01 3.05987149e-01
1.18091196e-01 -9.26735282e-01 1.15335023e+00 5.13918519e-01
8.66197050e-01 -7.53715754e-01 5.69468975e-01 3.55641730e-03
-9.82549787e-01 -5.91787249e-02 -2.21955165e-01 1.00119449e-01
4.60411668e-01 4.37484592e-01 -7.41253376e-01 5.62906146e-01
5.71834743e-01 7.11896837e-01 -4.50337529e-01 6.00772381e-01
-4.47026730e-01 4.17074949e-01 -3.39814974e-03 6.82286918e-02
3.74190062e-01 -2.72454530e-01 3.83678615e-01 1.05635870e+00
5.44321477e-01 -1.71112627e-01 -9.81009826e-02 1.27451539e+00
-2.32585624e-01 1.72601417e-01 -5.74381411e-01 -2.93157846e-01
-2.94837028e-01 1.42497861e+00 -6.31943107e-01 -2.27436543e-01
-3.03166389e-01 1.68714058e+00 1.86667442e-02 3.77398193e-01
-1.30752635e+00 -1.85584962e-01 2.98425496e-01 3.22962761e-01
7.83585727e-01 -2.26116791e-01 5.18013500e-02 -1.17028105e+00
1.00190923e-01 -7.96489000e-01 -7.79592320e-02 -8.49667668e-01
-1.42671561e+00 8.43829036e-01 -2.13858619e-01 -9.68334019e-01
-1.88941374e-01 -4.06715840e-01 -2.62309670e-01 7.49653041e-01
-1.47601426e+00 -1.96236825e+00 -3.24931145e-01 6.34258091e-01
7.13939846e-01 1.17521491e-02 7.58033216e-01 4.37153518e-01
-2.93670624e-01 5.44172823e-01 2.11876214e-01 -1.70129612e-01
5.41393220e-01 -8.58524680e-01 6.04559779e-01 6.67846262e-01
3.84248167e-01 5.43390036e-01 7.12942362e-01 -6.56847000e-01
-1.38097453e+00 -1.22317529e+00 7.37613738e-01 -3.16799611e-01
5.12177169e-01 -8.02466393e-01 -5.25112927e-01 8.71820450e-01
7.37189054e-01 -3.79045129e-01 6.95595443e-01 -1.70321345e-01
-2.84288555e-01 -1.73352033e-01 -8.64123762e-01 6.62216008e-01
1.03503203e+00 -5.17727315e-01 -1.89732090e-01 2.53904909e-01
4.42817509e-01 -2.40336269e-01 -7.75026143e-01 6.71424195e-02
7.32143760e-01 -9.74098325e-01 1.01899648e+00 -1.68354332e-01
8.94959390e-01 4.09985781e-02 -2.53541678e-01 -1.42345738e+00
-4.88402337e-01 -7.40468383e-01 -1.62171945e-02 1.27738202e+00
2.88360447e-01 -3.77044946e-01 7.12408900e-01 4.02101576e-01
-1.34978920e-01 -8.03266823e-01 -6.20269597e-01 -5.89872837e-01
-1.09675109e-01 -4.27366495e-01 4.64759767e-01 7.17712522e-01
-3.22145283e-01 3.56890082e-01 -8.36246312e-01 -1.81566849e-01
7.84649670e-01 5.50418831e-02 6.84417844e-01 -6.14599466e-01
-5.81289887e-01 -1.74752116e-01 -3.88912916e-01 -1.22357047e+00
-1.95094928e-01 -1.11920929e+00 -1.95365325e-02 -1.68123531e+00
2.93708920e-01 -1.76272467e-01 -2.55869836e-01 1.82052091e-01
-9.29423841e-04 8.60820413e-01 4.20269728e-01 -9.36069191e-02
-3.82018447e-01 1.03709435e+00 1.64345682e+00 -3.00612330e-01
-1.00033358e-01 -2.92697191e-01 -8.08217347e-01 3.41549039e-01
7.34226823e-01 -2.99301326e-01 -5.84204435e-01 -4.20294434e-01
3.03048342e-01 -2.50553399e-01 6.67370856e-01 -8.43947232e-01
-6.86152354e-02 1.88660204e-01 5.81829727e-01 -2.67884940e-01
6.72934234e-01 -5.70609391e-01 4.42436725e-01 4.55420136e-01
-3.21752518e-01 -8.11687112e-02 7.39204884e-02 6.49792075e-01
-1.77577540e-01 3.43357995e-02 7.84386098e-01 -1.96889773e-01
-8.41670156e-01 4.19804662e-01 -2.69789755e-01 -4.20546293e-01
1.16157770e+00 -2.02640980e-01 3.73929515e-02 -5.06408334e-01
-1.01524520e+00 -2.01478541e-01 4.15700942e-01 8.97847652e-01
9.56398427e-01 -1.61841214e+00 -9.52626526e-01 4.56479639e-01
1.57277375e-01 -5.28805077e-01 8.67672682e-01 4.51943666e-01
-4.69777882e-01 2.32690230e-01 -4.20711994e-01 -3.36734146e-01
-9.44858909e-01 8.13884020e-01 3.22959311e-02 -2.79817879e-01
-7.08629727e-01 9.27227318e-01 5.47404230e-01 -9.67249274e-02
-1.57125458e-01 -4.18403484e-02 2.04157740e-01 -1.07295744e-01
4.29098248e-01 5.44764288e-02 -1.67525664e-01 -6.86643302e-01
1.84735522e-01 3.54656875e-01 -2.88525850e-01 -2.93485463e-01
1.49071097e+00 -3.97459447e-01 2.48009607e-01 3.37435067e-01
1.18731511e+00 -2.49087960e-01 -1.44549644e+00 -1.09136321e-01
-6.81854904e-01 -6.63659334e-01 2.35812098e-01 -8.96463633e-01
-1.22589362e+00 8.24304819e-01 8.83380532e-01 -1.53604534e-03
1.15893936e+00 2.06820086e-01 1.12309444e+00 -3.33779365e-01
2.27353960e-01 -1.04063559e+00 5.92977226e-01 -8.60137641e-02
1.31368995e+00 -9.85360861e-01 -2.33256906e-01 -3.44486177e-01
-9.75693405e-01 7.12530255e-01 3.00427884e-01 -3.86921197e-01
6.23051405e-01 -1.73120701e-03 1.19277775e-01 -2.62871385e-01
-5.72370231e-01 -1.19389988e-01 2.39935726e-01 6.93927824e-01
2.58021802e-01 2.01096669e-01 -2.22588494e-01 5.03940880e-01
-1.57119274e-01 1.56397343e-01 2.53822237e-01 7.20502973e-01
3.86263616e-02 -1.23323452e+00 -1.09996632e-01 -2.30236501e-02
-1.60143599e-01 -2.50277579e-01 -7.48535544e-02 7.88321733e-01
4.22996938e-01 7.27472305e-01 -5.92904426e-02 -4.69300151e-01
2.48646945e-01 1.33979442e-02 8.32349777e-01 -3.03722173e-01
-5.19088507e-01 4.25391763e-01 -2.09579661e-01 -3.83458912e-01
-4.61346507e-01 -4.81158048e-01 -7.35529482e-01 -4.22782063e-01
-2.74128497e-01 -8.36749673e-02 1.00981009e+00 4.72288996e-01
4.36635405e-01 5.50922215e-01 6.44848228e-01 -1.01329732e+00
-4.27950263e-01 -9.23627198e-01 -8.64985466e-01 7.13915408e-01
-5.96591234e-02 -5.56191564e-01 -4.66867685e-02 4.73685265e-01] | [11.490551948547363, -0.5384437441825867] |
0f93e15f-d358-40ed-a612-8ede345e55ae | language-models-in-word-sense-disambiguation | 2111.13982 | null | https://arxiv.org/abs/2111.13982v1 | https://arxiv.org/pdf/2111.13982v1.pdf | Language models in word sense disambiguation for Polish | In the paper, we test two different approaches to the {unsupervised} word sense disambiguation task for Polish. In both methods, we use neural language models to predict words similar to those being disambiguated and, on the basis of these words, we predict the partition of word senses in different ways. In the first method, we cluster selected similar words, while in the second, we cluster vectors representing their subsets. The evaluation was carried out on texts annotated with plWordNet senses and provided a relatively good result (F1=0.68 for all ambiguous words). The results are significantly better than those obtained for the neural model-based unsupervised method proposed in \cite{waw:myk:17:Sense} and are at the level of the supervised method presented there. The proposed method may be a way of solving word sense disambiguation problem for languages that lack sense annotated data. | ['Piotr Rychlik', 'Agnieszka A. Mykowiecka', 'Agnieszka Mykowiecka'] | 2021-11-27 | null | null | null | null | ['word-sense-disambiguation'] | ['natural-language-processing'] | [ 1.96110755e-01 3.20577562e-01 -1.79983392e-01 -1.11106612e-01
-2.94977307e-01 -6.30681336e-01 6.21730983e-01 8.52166653e-01
-1.04743147e+00 9.73172605e-01 4.02116328e-01 -3.88928562e-01
-4.09902632e-01 -9.64103460e-01 7.64818192e-02 -7.25369811e-01
4.17737067e-01 7.90188491e-01 9.31978971e-02 -8.04542601e-01
5.13458729e-01 1.99800223e-01 -1.53255355e+00 1.30682260e-01
8.68717492e-01 5.00743508e-01 7.28395879e-01 1.62488103e-01
-8.16907585e-01 3.37250382e-01 -6.48312688e-01 -9.75907519e-02
-5.26869446e-02 -2.06324190e-01 -1.26444232e+00 -1.07622221e-01
-1.03826389e-01 7.35100448e-01 2.25646153e-01 1.32523036e+00
4.08610910e-01 5.03946006e-01 7.02706754e-01 -6.49149418e-01
-4.10113156e-01 1.22497344e+00 -3.34181964e-01 2.07883269e-01
4.12222296e-01 -4.95737344e-01 1.30315924e+00 -6.86167181e-01
7.66664803e-01 1.21827412e+00 3.20477456e-01 5.94440758e-01
-1.34733570e+00 -6.52697325e-01 2.04469457e-01 2.14052603e-01
-1.61939061e+00 -1.43748447e-01 6.19477093e-01 -6.09370232e-01
1.23857665e+00 8.31970200e-02 5.45912623e-01 8.28188062e-01
-1.08256228e-01 5.85491538e-01 1.16220474e+00 -9.15919185e-01
3.62238258e-01 2.19478145e-01 7.87306666e-01 2.60715723e-01
6.69446230e-01 -1.63013905e-01 -2.24716559e-01 -2.57544637e-01
2.93595910e-01 -3.09110969e-01 -2.54508585e-01 9.39128995e-02
-1.24330807e+00 1.01415169e+00 1.36572063e-01 1.47737646e+00
-5.65659821e-01 -4.79954898e-01 4.91369992e-01 1.44958019e-01
7.17673182e-01 1.07967436e+00 -8.18960011e-01 2.84238964e-01
-1.09610760e+00 2.77262062e-01 8.54714334e-01 3.82751793e-01
9.66147780e-01 9.34665576e-02 -6.72310665e-02 9.72173810e-01
3.48072499e-01 2.87212908e-01 1.21873116e+00 -3.24962109e-01
1.72643617e-01 7.12411761e-01 5.88156730e-02 -1.21341193e+00
-6.28092289e-01 -2.02112421e-01 -6.11000240e-01 -6.33226857e-02
2.88764864e-01 -2.97535241e-01 -1.06911373e+00 1.68792450e+00
-3.16529945e-02 5.92835015e-03 6.37090623e-01 6.33038104e-01
1.07482302e+00 6.90314412e-01 5.93890369e-01 -3.87939364e-01
1.48388517e+00 -4.20629054e-01 -9.17478621e-01 -2.52419949e-01
6.52715743e-01 -8.63240004e-01 9.20575738e-01 5.48450351e-01
-6.55072153e-01 -5.71589053e-01 -1.22300708e+00 3.51010412e-01
-1.01374722e+00 1.01838559e-01 4.94432330e-01 6.38922274e-01
-1.31441736e+00 5.00954807e-01 -3.07974696e-01 -7.73038030e-01
-3.65474820e-01 3.00411433e-01 -3.10968995e-01 4.07796085e-01
-1.73921895e+00 1.07380700e+00 1.44262564e+00 -1.41601473e-01
-9.11340415e-02 -7.16056228e-02 -1.06469691e+00 9.35977921e-02
2.85357803e-01 -3.17897111e-01 7.49020278e-01 -1.11876070e+00
-9.15884137e-01 1.35312176e+00 -2.81971872e-01 -5.15679836e-01
-2.18692780e-01 -6.01135753e-03 -7.34832585e-01 -1.58845708e-01
4.30041850e-01 5.78173399e-01 1.97470367e-01 -1.48893785e+00
-8.08514595e-01 -3.56649041e-01 -1.82140786e-02 -7.15841306e-03
-4.62828130e-01 -5.83022609e-02 9.37208161e-02 -8.45253170e-01
6.02666736e-02 -6.92309082e-01 -5.45748413e-01 -1.07078302e+00
-3.48462909e-01 -8.81465852e-01 1.38415933e-01 -5.10983884e-01
1.39469779e+00 -1.90805387e+00 -8.21671728e-03 5.46991289e-01
2.82665491e-01 6.03977501e-01 -1.36649475e-01 5.43980718e-01
-4.55837965e-01 3.37872595e-01 -3.08784008e-01 -1.16957255e-01
-1.29874581e-02 4.11341488e-01 -3.26716959e-01 -1.56903952e-01
-2.41051931e-02 3.79450917e-01 -1.11699665e+00 -5.05591154e-01
1.97062120e-01 2.04564646e-01 -6.52661249e-02 -1.71006590e-01
-2.64641851e-01 -1.91042796e-02 -5.15299141e-01 1.87482491e-01
3.76363873e-01 2.64376074e-01 6.87134266e-01 -3.69084626e-02
-3.47408503e-02 3.78781319e-01 -1.35077298e+00 1.52914798e+00
-5.56579113e-01 5.88531554e-01 -3.98231804e-01 -1.25567722e+00
1.31929338e+00 7.08619595e-01 5.18367827e-01 -3.40799451e-01
2.31506556e-01 3.33019048e-01 8.86666328e-02 -5.13197422e-01
1.13848329e+00 -4.29997116e-01 -2.40625024e-01 3.03510219e-01
3.94828707e-01 -2.16252282e-02 6.33273304e-01 -2.00220048e-01
6.23273432e-01 -2.19483417e-03 9.73202467e-01 -7.77488649e-01
7.20901430e-01 2.96461910e-01 4.56337571e-01 6.82914615e-01
1.22292951e-01 1.60024732e-01 2.36179769e-01 -5.00138998e-01
-6.40496671e-01 -8.75175476e-01 -9.26687345e-02 1.09779763e+00
8.20785314e-02 -6.88890755e-01 -8.24063480e-01 -3.81862611e-01
-2.90123552e-01 1.07148981e+00 -3.45417410e-01 2.13286594e-01
-3.29402477e-01 -7.02818692e-01 6.51023924e-01 2.16212094e-01
1.78914800e-01 -1.50758696e+00 -4.68899667e-01 5.31567931e-01
-3.34311903e-01 -9.31108832e-01 2.22393647e-01 4.58579421e-01
-4.76837784e-01 -9.59366441e-01 -4.59424019e-01 -1.07565546e+00
2.77843624e-01 -4.90100458e-02 1.36835802e+00 2.42020622e-01
1.31135536e-02 7.10496828e-02 -8.06826770e-01 -8.11916590e-01
-3.45391929e-01 5.15564680e-01 4.08699334e-01 -1.92089558e-01
1.04215622e+00 -4.69369918e-01 7.94604868e-02 -2.02936575e-01
-1.29619169e+00 -6.36736929e-01 2.48209447e-01 7.97966063e-01
4.95152444e-01 2.99543589e-01 4.44021612e-01 -9.90938306e-01
1.02532220e+00 -5.00780225e-01 -3.75676006e-01 7.71631002e-02
-9.53745186e-01 3.97864789e-01 4.11568075e-01 -4.84152585e-01
-9.29654062e-01 1.84469417e-01 -5.57095110e-01 1.97087824e-01
-6.65617049e-01 7.76030898e-01 -2.39932597e-01 3.40326041e-01
7.97083497e-01 1.03919588e-01 -3.61625969e-01 -5.82027793e-01
2.99961269e-01 7.88925588e-01 1.91379458e-01 -5.39792836e-01
4.00219262e-01 -5.76096699e-02 -4.27547663e-01 -1.10828018e+00
-7.71382689e-01 -7.52578735e-01 -6.31783247e-01 5.44229597e-02
1.35322702e+00 -5.67093372e-01 -3.49069744e-01 6.23009820e-03
-1.32620978e+00 1.33798867e-01 -2.87115216e-01 6.57246470e-01
-3.57835948e-01 4.78206933e-01 -9.70899537e-02 -9.86473560e-01
-5.47495008e-01 -8.65412474e-01 7.09381700e-01 2.50827283e-01
-9.89240706e-01 -1.30598402e+00 3.26527238e-01 -6.37063980e-02
1.83623195e-01 2.01758504e-01 1.13603115e+00 -1.58260751e+00
5.36480606e-01 8.67889449e-02 2.31155589e-01 1.98148623e-01
3.75168592e-01 -3.93290579e-01 -9.12169933e-01 -4.16138284e-02
-9.71883386e-02 -2.71909032e-02 1.13469183e+00 3.25541139e-01
6.56517267e-01 -1.54868439e-01 -4.81795877e-01 -1.38646454e-01
1.75118065e+00 4.56030637e-01 4.71267700e-01 5.70922077e-01
3.76794994e-01 6.98316574e-01 6.97343528e-01 1.59942627e-01
1.64755434e-01 3.94762158e-01 6.28115237e-02 -2.61457682e-01
3.92187804e-01 7.49607608e-02 6.72924006e-03 6.88053429e-01
-1.39481470e-01 -5.37842274e-01 -1.13789785e+00 1.01755869e+00
-1.83241820e+00 -9.67053652e-01 -1.64368048e-01 1.99423671e+00
6.78432941e-01 3.42702121e-01 1.46915466e-01 4.97804672e-01
9.51688468e-01 5.01352191e-01 3.22628826e-01 -7.52952516e-01
-2.88739413e-01 8.14431071e-01 2.68696785e-01 1.06286907e+00
-9.83282804e-01 1.54105890e+00 6.16058397e+00 1.00634682e+00
-9.39876676e-01 5.38673550e-02 2.47416839e-01 3.55074048e-01
-4.81821924e-01 -7.29316205e-04 -6.48053825e-01 3.26008528e-01
6.60522640e-01 -1.81307986e-01 1.55328199e-01 5.05576849e-01
9.81627330e-02 -2.27883771e-01 -6.08729243e-01 9.05069888e-01
1.60834610e-01 -1.09966505e+00 3.97540838e-01 -1.23154722e-01
3.99688482e-01 -3.59116159e-02 -6.37641072e-01 2.45665416e-01
5.76657951e-01 -1.11174917e+00 4.11065072e-01 6.48859620e-01
2.73915291e-01 -1.03604054e+00 1.16753125e+00 3.28634530e-01
-1.18730700e+00 1.70132324e-01 -7.11071134e-01 -2.77192801e-01
4.31477018e-02 7.80368865e-01 -9.68234897e-01 7.43618071e-01
4.76717442e-01 3.88044178e-01 -3.78573656e-01 6.03871167e-01
-3.54027241e-01 7.06068933e-01 -1.63349569e-01 -6.04244351e-01
3.65851730e-01 -1.76442757e-01 8.68556499e-01 1.60842526e+00
3.37433964e-01 1.17245473e-01 4.86581326e-01 6.62617683e-01
4.36117530e-01 6.88090503e-01 -7.52529621e-01 -8.48324448e-02
5.02002358e-01 1.14439535e+00 -9.90461886e-01 -5.58769524e-01
2.23659426e-01 7.99660146e-01 8.56339708e-02 2.77979076e-01
-6.69636298e-03 -6.83192849e-01 7.36639678e-01 -7.67186582e-02
5.23906536e-02 -1.26765430e-01 -3.33535403e-01 -1.04085124e+00
-1.80936083e-01 -4.71407592e-01 6.79567695e-01 -7.16949940e-01
-1.36615956e+00 1.20630586e+00 2.66823053e-01 -8.46238434e-01
-6.90165937e-01 -1.05528462e+00 -3.41505080e-01 1.07378376e+00
-1.21849930e+00 -8.04647446e-01 3.56159508e-01 4.61833030e-01
5.77486098e-01 -5.06146431e-01 1.36090577e+00 -6.32491931e-02
9.56605971e-02 2.03216150e-01 -1.17997348e-01 1.95364982e-01
5.20464063e-01 -1.41940665e+00 1.78011253e-01 6.84208333e-01
5.36449492e-01 7.11739182e-01 1.04041839e+00 -6.81474805e-01
-3.65950614e-01 -6.73992515e-01 1.80466771e+00 -1.55019134e-01
8.18388879e-01 1.33231580e-01 -8.38385284e-01 2.67450929e-01
6.47007346e-01 -7.54862666e-01 1.06954348e+00 2.44234174e-01
-9.49568897e-02 1.79664969e-01 -1.18980336e+00 4.66456652e-01
6.16355836e-01 -2.88842618e-01 -1.42292297e+00 1.13086127e-01
7.69018829e-01 2.04496011e-01 -7.42481887e-01 3.77614051e-02
2.69937396e-01 -6.97247624e-01 9.03644621e-01 -8.00177872e-01
7.98334628e-02 -3.37760568e-01 -3.92313898e-01 -1.82439148e+00
-3.65303546e-01 -2.00757638e-01 7.45881975e-01 1.45651293e+00
7.48394728e-01 -8.70363653e-01 2.55350620e-01 -1.80850640e-01
2.21749321e-01 -2.52311587e-01 -9.06801522e-01 -5.00135005e-01
2.69264281e-01 -5.73325157e-01 6.93739653e-01 1.40834379e+00
4.37321246e-01 5.85182250e-01 9.43743140e-02 -5.77224568e-02
1.75390631e-01 -2.42679399e-02 -5.07816747e-02 -1.71925473e+00
-1.22186616e-01 -7.38578320e-01 -6.79901838e-01 -3.44578415e-01
8.10160220e-01 -9.85868335e-01 2.03938320e-01 -1.63173974e+00
-1.56999588e-01 -2.39802644e-01 -7.90060341e-01 8.34710777e-01
-3.77740115e-01 1.52071655e-01 2.84559160e-01 -1.24135450e-01
-1.16154894e-01 -6.69386014e-02 5.11342525e-01 -3.10807496e-01
-2.92743742e-01 -1.49718121e-01 -8.86780262e-01 9.76370335e-01
8.89654815e-01 -5.24083972e-01 -2.05160230e-01 -3.33376884e-01
5.08434415e-01 -1.84689745e-01 -9.81050655e-02 -8.01107883e-01
-4.93200645e-02 -3.30598950e-01 2.07394108e-01 -4.71418142e-01
9.93396249e-03 -7.82638192e-01 -2.12560996e-01 4.88520831e-01
-3.66043270e-01 4.46475983e-01 2.14565694e-01 9.51447785e-02
-3.48104656e-01 -6.27230406e-01 5.92713773e-01 -4.47320074e-01
-1.12703967e+00 -3.70842278e-01 -1.03719914e+00 9.00811609e-03
5.14845610e-01 -2.31578037e-01 3.54860067e-01 -2.49663651e-01
-1.12808466e+00 1.91524811e-02 2.33349726e-01 5.97366929e-01
3.30820233e-01 -1.26450932e+00 -7.34711766e-01 -1.85356304e-01
4.44831431e-01 -3.66929233e-01 -2.56864280e-01 8.14114586e-02
-2.73150086e-01 4.04907256e-01 -1.80608541e-01 -1.61737561e-01
-1.21247423e+00 7.01883852e-01 1.80876762e-01 -5.88860452e-01
-1.30376458e-01 4.75128770e-01 -1.04677439e-01 -6.17457688e-01
-1.12431265e-01 -2.23412946e-01 -1.24755692e+00 5.88201165e-01
4.88472313e-01 2.66644564e-02 1.85210347e-01 -1.01788032e+00
-3.98810178e-01 4.98366296e-01 2.09845483e-01 -4.33133721e-01
1.30556250e+00 -3.91951017e-02 -3.97539079e-01 7.34030426e-01
6.53575122e-01 2.96668708e-01 9.68111232e-02 -2.72529751e-01
6.37784898e-01 1.45532340e-01 -8.44109729e-02 -7.53265858e-01
-7.61087596e-01 5.29245794e-01 8.52246106e-01 5.80218375e-01
1.07935143e+00 2.99926382e-02 4.68062133e-01 7.03543544e-01
2.98168987e-01 -1.22015870e+00 -7.15102732e-01 1.03576374e+00
5.01312315e-01 -8.99464190e-01 -2.60194600e-01 -2.59349793e-01
-6.02566004e-01 1.26747918e+00 3.06486696e-01 -2.22677335e-01
6.96138382e-01 1.33378640e-01 3.78544658e-01 -2.91238964e-01
-3.47984999e-01 -8.97623777e-01 2.89418399e-01 7.75431275e-01
8.08603168e-01 3.82411271e-01 -1.48098683e+00 7.36418784e-01
-5.91894686e-01 -2.65191287e-01 4.97365683e-01 6.23061478e-01
-6.98493183e-01 -1.41835523e+00 -6.13632679e-01 1.92345336e-01
-5.37027240e-01 -4.10083681e-01 -7.70139933e-01 9.00854170e-01
4.47906911e-01 1.45844853e+00 1.49422526e-01 -5.91044068e-01
2.54724741e-01 4.14160579e-01 5.33899516e-02 -7.93345869e-01
-6.61018789e-01 8.29297975e-02 3.37508649e-01 -1.52376756e-01
-7.63313890e-01 -3.86310339e-01 -1.38834035e+00 2.33888522e-01
-2.86752194e-01 7.87454545e-01 4.87509161e-01 1.61264336e+00
-2.30207548e-01 3.28774750e-01 3.06934446e-01 -8.33091378e-01
-1.16889209e-01 -1.25700259e+00 -8.13360035e-01 4.43721861e-01
4.80764657e-02 -6.62217915e-01 -1.09287597e-01 3.68441991e-03] | [10.193241119384766, 9.201297760009766] |
f773823d-a146-4ea1-b6e7-1a9587c6997a | korc-knowledge-oriented-reading-comprehension | 2307.03115 | null | https://arxiv.org/abs/2307.03115v1 | https://arxiv.org/pdf/2307.03115v1.pdf | KoRC: Knowledge oriented Reading Comprehension Benchmark for Deep Text Understanding | Deep text understanding, which requires the connections between a given document and prior knowledge beyond its text, has been highlighted by many benchmarks in recent years. However, these benchmarks have encountered two major limitations. On the one hand, most of them require human annotation of knowledge, which leads to limited knowledge coverage. On the other hand, they usually use choices or spans in the texts as the answers, which results in narrow answer space. To overcome these limitations, we build a new challenging benchmark named KoRc in this paper. Compared with previous benchmarks, KoRC has two advantages, i.e., broad knowledge coverage and flexible answer format. Specifically, we utilize massive knowledge bases to guide annotators or large language models (LLMs) to construct knowledgable questions. Moreover, we use labels in knowledge bases rather than spans or choices as the final answers. We test state-of-the-art models on KoRC and the experimental results show that the strongest baseline only achieves 68.3% and 30.0% F1 measure in the in-distribution and out-of-distribution test set, respectively. These results indicate that deep text understanding is still an unsolved challenge. The benchmark dataset, leaderboard, and baseline methods are released in https://github.com/THU-KEG/KoRC. | ['Juanzi Li', 'Lei Hou', 'Jifan Yu', 'Shulin Cao', 'Xin Lv', 'Yantao Liu', 'Zijun Yao'] | 2023-07-06 | null | null | null | null | ['reading-comprehension'] | ['natural-language-processing'] | [-2.56686926e-01 6.06457628e-02 -3.36828232e-01 -3.32717180e-01
-9.23840106e-01 -8.22523057e-01 3.25740516e-01 -9.88811031e-02
-4.83091921e-01 9.49291885e-01 3.03448558e-01 -2.89107561e-01
-2.41443634e-01 -7.83638239e-01 -6.05005324e-01 -3.23908091e-01
7.25220084e-01 7.42933810e-01 4.57259983e-01 -4.38158035e-01
1.47986218e-01 -3.50810081e-01 -1.06537092e+00 6.29727244e-01
1.38252842e+00 9.53629553e-01 1.83081940e-01 3.65087867e-01
-7.11024821e-01 8.32598209e-01 -6.39700711e-01 -8.30545425e-01
-2.35316873e-01 -2.84322023e-01 -1.30280221e+00 -2.96124786e-01
2.47153163e-01 -3.83352339e-01 -4.27811712e-01 9.12929058e-01
3.49080592e-01 -3.72680500e-02 5.56377470e-01 -1.00778532e+00
-1.08150709e+00 9.17036831e-01 -3.02170157e-01 -2.13534385e-01
5.58846951e-01 1.11352913e-02 1.09679890e+00 -9.69258726e-01
4.23901141e-01 1.33076429e+00 6.71963990e-01 6.77288949e-01
-8.35594475e-01 -6.79555714e-01 3.12343597e-01 5.43623447e-01
-1.32149816e+00 -3.30117613e-01 5.71082234e-01 -4.05696988e-01
8.16573143e-01 3.15559834e-01 1.19463444e-01 1.04085517e+00
-3.64901304e-01 1.11399460e+00 1.02367032e+00 -4.80166107e-01
-1.64038494e-01 2.34839320e-01 5.66725373e-01 6.75768256e-01
1.01575933e-01 -4.70077574e-01 -4.88301545e-01 1.02701327e-02
3.20450902e-01 -1.75733432e-01 -7.92245686e-01 2.05293730e-01
-1.12046790e+00 8.47733259e-01 2.83896685e-01 1.17443785e-01
7.50183240e-02 -3.29888053e-02 5.23229718e-01 2.30131090e-01
3.64918649e-01 2.87254035e-01 -8.66962969e-01 -1.93542331e-01
-3.76306802e-01 8.92036855e-02 1.11038387e+00 1.05398810e+00
8.40229094e-01 -4.69983339e-01 -2.58213371e-01 1.12240970e+00
2.21740678e-01 6.07894421e-01 4.97590005e-01 -8.46012294e-01
1.07870722e+00 8.89189243e-01 1.89728454e-01 -1.02524674e+00
-3.78579855e-01 -4.11825359e-01 -8.52753580e-01 -7.96526134e-01
4.68685508e-01 -3.21990728e-01 -7.35819995e-01 1.67430627e+00
3.23480725e-01 1.53208852e-01 3.44866723e-01 6.96577430e-01
1.56663609e+00 8.24273884e-01 -8.85505974e-02 -8.01888332e-02
1.46410167e+00 -1.26466393e+00 -1.02532148e+00 -2.07357019e-01
8.93334746e-01 -8.77462983e-01 1.47244120e+00 3.06930214e-01
-6.07665479e-01 -5.49038589e-01 -6.85857415e-01 -3.78747672e-01
-6.06112301e-01 4.31654006e-01 4.88270313e-01 5.47851920e-01
-7.71783710e-01 -6.65010232e-03 -3.65527868e-01 -2.67749548e-01
2.27554619e-01 1.05638541e-02 -3.00876945e-02 -4.51223850e-01
-1.72671056e+00 7.04938531e-01 7.39015520e-01 8.19530115e-02
-6.09761953e-01 -6.18233979e-01 -6.77637339e-01 1.44289210e-01
8.00777137e-01 -5.68589389e-01 1.45193279e+00 -2.98299670e-01
-1.55781531e+00 6.60228670e-01 -1.56441331e-01 -1.25375614e-01
4.32645142e-01 -6.03791714e-01 -4.48423654e-01 -5.37658036e-02
1.38570085e-01 5.56006491e-01 1.76046088e-01 -1.17824674e+00
-6.11171246e-01 -2.33368203e-01 4.49181437e-01 2.03000292e-01
-5.07649362e-01 -2.25414157e-01 -1.13621414e+00 -4.57476526e-01
2.27305051e-02 -8.08232486e-01 2.77791739e-01 -6.04195476e-01
-6.65704250e-01 -6.80967331e-01 8.86605322e-01 -7.45848715e-01
1.77936828e+00 -1.74426687e+00 -6.93272129e-02 -3.12681884e-01
1.24348558e-01 4.81460333e-01 -2.27293432e-01 6.60454035e-01
2.93461621e-01 3.67207289e-01 4.83404705e-03 -1.20225094e-01
2.31245577e-01 2.98200071e-01 -6.87112927e-01 -2.12292105e-01
-5.67065366e-02 1.08991039e+00 -8.76485348e-01 -5.23388684e-01
-5.03015593e-02 5.05360588e-02 -3.44651967e-01 8.76384974e-02
-6.54916286e-01 3.84682715e-01 -8.44477355e-01 4.05686140e-01
6.88880265e-01 -5.09248734e-01 1.30030587e-02 -2.55508423e-01
1.68402821e-01 5.34158349e-01 -9.96079147e-01 1.72199452e+00
-5.65959275e-01 3.31356823e-01 -1.97178438e-01 -8.44661832e-01
9.59811032e-01 4.64721441e-01 -8.09004530e-02 -6.05236888e-01
3.41939107e-02 5.56550324e-01 -1.30358145e-01 -7.01485455e-01
3.93549383e-01 1.78951532e-01 -9.35180634e-02 4.33468908e-01
-5.24297878e-02 -2.08023980e-01 3.78560036e-01 3.33595395e-01
9.40087378e-01 1.13162491e-02 9.28998291e-02 -8.82097110e-02
8.25092435e-01 1.12211049e-01 7.95532405e-01 6.27909660e-01
-1.48687642e-02 3.75389427e-01 6.46543980e-01 -3.35191667e-01
-4.88195419e-01 -7.17203617e-01 -2.44530439e-02 1.07195127e+00
2.56193697e-01 -6.84005618e-01 -7.84239471e-01 -1.10409415e+00
-8.64120349e-02 8.69147778e-01 -4.64042842e-01 -7.18763694e-02
-6.08945668e-01 -5.45518637e-01 6.82160079e-01 6.39291286e-01
9.03290451e-01 -9.59062219e-01 5.35489395e-02 1.62762731e-01
-8.01154315e-01 -1.39102125e+00 -6.65847123e-01 -1.90884456e-01
-7.08968461e-01 -1.35713720e+00 -6.64009094e-01 -8.43152225e-01
4.54530597e-01 2.47016490e-01 1.47310734e+00 2.05620527e-01
2.01783374e-01 4.42358583e-01 -7.94056773e-01 -4.27273035e-01
-3.29478681e-02 4.59250659e-01 -3.35224628e-01 -2.04746887e-01
6.12730443e-01 -1.76874891e-01 -6.89413071e-01 7.69133508e-01
-7.88795233e-01 3.00174475e-01 5.04293263e-01 9.57560599e-01
6.21268690e-01 1.08282864e-01 1.17840779e+00 -1.19175065e+00
7.22767770e-01 -5.22623420e-01 -4.71334726e-01 7.58307278e-01
-3.57941240e-01 -1.17195465e-01 7.88162529e-01 -2.37266853e-01
-1.33932674e+00 -1.96693093e-01 -3.05213422e-01 6.50787354e-02
5.39558940e-02 8.90486002e-01 -3.80600691e-01 3.61100525e-01
5.56162894e-01 3.25235695e-01 -3.74768674e-01 -6.93373084e-01
4.07861263e-01 1.03285885e+00 4.75144655e-01 -9.86512125e-01
5.16374409e-01 2.00905785e-01 -6.69935524e-01 -4.66328681e-01
-1.71546221e+00 -5.46416283e-01 -4.22444165e-01 -3.75231095e-02
7.31179714e-01 -8.40011716e-01 -7.73814023e-01 5.66908121e-01
-1.46592367e+00 -4.30418193e-01 1.37141898e-01 3.83829117e-01
-2.19302535e-01 3.67524385e-01 -6.53067529e-01 -4.94597375e-01
-5.87193310e-01 -1.04983747e+00 8.00853908e-01 4.40091044e-01
-1.21388450e-01 -1.10750639e+00 -1.78429589e-01 1.06300664e+00
2.64910132e-01 -6.86748549e-02 1.18768978e+00 -8.48981321e-01
-5.72480857e-01 -1.34258553e-01 -4.24919397e-01 4.45621938e-01
1.71661124e-01 -2.27753729e-01 -8.72045159e-01 -4.61032130e-02
-9.57615525e-02 -8.77868056e-01 1.00674736e+00 -6.23139367e-02
1.55787623e+00 -2.38990903e-01 -2.40808994e-01 2.57584572e-01
1.09998333e+00 1.49527460e-01 6.68093503e-01 3.20275217e-01
7.16659546e-01 6.96463227e-01 7.33085871e-01 2.26485193e-01
8.22556794e-01 5.50264716e-01 1.51321143e-01 3.13176334e-01
-2.83349991e-01 -4.63045597e-01 1.59315616e-01 1.40247858e+00
2.34175295e-01 -5.79499006e-01 -1.22238183e+00 6.26995564e-01
-2.12021899e+00 -4.78993535e-01 -4.40121293e-01 1.92222238e+00
1.33361077e+00 7.26659596e-02 -3.94724071e-01 -9.03686509e-02
7.41654873e-01 4.50761467e-02 -7.48005450e-01 1.59967139e-01
-1.79791570e-01 -2.90398281e-02 1.54487520e-01 6.10261202e-01
-8.82715225e-01 1.07851970e+00 5.02472353e+00 1.33078456e+00
-5.99766552e-01 1.77408397e-01 6.08754754e-01 2.56470591e-01
-4.63995665e-01 3.15875486e-02 -1.31673455e+00 6.61535561e-01
5.97490668e-01 -1.76929593e-01 1.39689684e-01 5.97114563e-01
-1.62139729e-01 -7.53240101e-03 -1.01517260e+00 9.65948582e-01
9.34542343e-02 -1.18889797e+00 1.31042644e-01 -3.28918219e-01
9.21883404e-01 -6.95100278e-02 -6.96813017e-02 9.17404711e-01
4.57920760e-01 -1.12743306e+00 3.81836593e-01 5.68319857e-01
8.04469109e-01 -7.26741135e-01 1.08618760e+00 7.90320575e-01
-1.10567474e+00 1.64054736e-01 -5.02235532e-01 1.04555755e-03
1.04105100e-01 7.42141068e-01 -5.54260492e-01 7.09699214e-01
6.41946316e-01 5.61552167e-01 -4.77707475e-01 7.17770159e-01
-7.58272946e-01 9.72448885e-01 -3.74221623e-01 -1.81070745e-01
3.28061730e-01 -3.38986106e-02 -9.27597359e-02 1.12849855e+00
2.91111410e-01 4.11394596e-01 4.99358177e-01 8.27342749e-01
-6.36684537e-01 2.72776365e-01 -7.98182189e-02 -2.80194376e-02
8.45905662e-01 1.12948966e+00 -2.77198672e-01 -3.75897944e-01
-6.51540518e-01 7.47408271e-01 5.01320601e-01 4.80671614e-01
-8.19712162e-01 -6.27597451e-01 1.21646754e-01 -2.98692167e-01
4.86817211e-02 6.54355288e-02 -2.59079069e-01 -1.38625312e+00
2.49973357e-01 -1.05500937e+00 6.99172080e-01 -7.12420523e-01
-1.49502850e+00 4.23802763e-01 -1.65831551e-01 -8.06341887e-01
1.64393075e-02 -6.57951593e-01 -5.27616978e-01 9.76015210e-01
-1.75926995e+00 -1.14193857e+00 -6.21375382e-01 5.11130035e-01
6.71888590e-01 -1.33411452e-01 7.79654860e-01 3.60995531e-01
-7.59687304e-01 7.10569263e-01 3.71328443e-01 4.74236667e-01
9.28712308e-01 -1.26230454e+00 2.84786165e-01 2.70769358e-01
3.28815356e-02 6.66874468e-01 2.49889150e-01 -6.88341677e-01
-1.26268494e+00 -9.94219959e-01 9.41243470e-01 -8.22182357e-01
8.01605761e-01 -1.86856300e-01 -1.35990238e+00 7.41909742e-01
2.54519284e-01 -2.37452447e-01 8.36516798e-01 5.52133441e-01
-5.29975772e-01 -1.18800737e-01 -6.32796586e-01 5.51137924e-01
7.74181843e-01 -5.86518764e-01 -7.37619877e-01 5.13387442e-01
1.07839954e+00 -6.10707462e-01 -8.24452817e-01 6.11681163e-01
2.68302053e-01 -8.17957819e-01 6.09357655e-01 -4.72563863e-01
5.26179731e-01 -3.45867336e-01 -7.83418566e-02 -1.23148417e+00
6.43483251e-02 -3.75725478e-01 -3.55712920e-01 1.76905322e+00
7.15518177e-01 -7.13636041e-01 8.74000371e-01 5.47245264e-01
-3.01591724e-01 -1.23167467e+00 -6.30888820e-01 -7.40349352e-01
4.14970368e-01 -4.90999997e-01 7.25274146e-01 9.57937241e-01
-9.59287286e-02 7.57972300e-01 -2.95646667e-01 -4.88502495e-02
2.31518358e-01 3.18037868e-01 8.60460699e-01 -1.08376825e+00
-1.54405788e-01 -3.13960105e-01 3.77468675e-01 -1.69861114e+00
3.24818343e-01 -8.92358124e-01 -5.79603128e-02 -1.85444558e+00
3.98454040e-01 -6.73194706e-01 -1.30246669e-01 6.73639774e-01
-6.05076313e-01 -2.49663189e-01 -4.49284278e-02 2.60319084e-01
-1.00555134e+00 8.94948006e-01 1.45522285e+00 -1.76420629e-01
5.91999441e-02 -4.52495515e-02 -9.15364981e-01 9.61782098e-01
9.97308254e-01 -3.26885283e-01 -6.21637821e-01 -9.28157628e-01
4.92912859e-01 -1.47256274e-02 1.33812174e-01 -5.37175655e-01
4.26807374e-01 -9.90468636e-02 3.40461619e-02 -8.64615142e-01
1.30244583e-01 -5.09665430e-01 -4.16402757e-01 3.05066347e-01
-3.55589539e-01 -1.49756908e-01 1.89744085e-01 5.40267229e-01
-5.54296672e-01 -6.41801000e-01 4.07714039e-01 -1.12433694e-01
-8.50656331e-01 1.42171353e-01 2.28658587e-01 7.60718644e-01
5.54174364e-01 1.84203520e-01 -8.81490588e-01 -5.34010828e-01
-2.62276053e-01 1.01452851e+00 3.92869823e-02 4.96705770e-01
4.47499156e-01 -1.29332018e+00 -9.77536142e-01 -4.26293612e-01
3.46091241e-01 3.85851979e-01 6.88838422e-01 7.69395530e-01
-5.03519952e-01 9.15343761e-01 4.97770548e-01 -3.54982376e-01
-1.01246917e+00 2.51095861e-01 2.02469468e-01 -6.22955382e-01
-2.23870054e-01 9.75183487e-01 4.31703120e-01 -1.05223858e+00
3.87044430e-01 -1.56369120e-01 -3.42424572e-01 5.63678779e-02
5.96150100e-01 3.30374897e-01 8.95762164e-03 -1.10848680e-01
-2.37714887e-01 6.74256265e-01 -3.72366935e-01 2.94502705e-01
9.03790593e-01 -2.66246229e-01 -1.65209800e-01 2.49242783e-01
9.92698193e-01 4.88637760e-02 -9.32727933e-01 -7.08604574e-01
2.30631724e-01 -2.59787470e-01 -1.58006072e-01 -1.43465245e+00
-9.12663758e-01 1.02936983e+00 2.34460905e-02 9.53952819e-02
1.01371872e+00 2.26777047e-01 1.22906935e+00 9.35153127e-01
1.91659108e-01 -1.25514114e+00 2.33917087e-01 1.27246225e+00
1.10467851e+00 -1.41731942e+00 -1.49109825e-01 -7.03281224e-01
-6.41913474e-01 1.03235269e+00 1.04647672e+00 6.23756766e-01
4.57085162e-01 -6.03640638e-02 3.49748373e-01 -1.55072331e-01
-9.58438873e-01 -3.39378744e-01 4.35127318e-01 2.84607351e-01
6.99027896e-01 -2.30184421e-02 -4.15489584e-01 1.10192358e+00
-2.95227915e-01 -6.30319938e-02 2.45388702e-01 4.98259217e-01
-6.04374290e-01 -1.08299363e+00 -2.19016910e-01 4.25567746e-01
-4.92524594e-01 -2.07131922e-01 -5.59752941e-01 8.72770965e-01
1.73495650e-01 1.31923008e+00 -5.49366891e-01 -3.82789850e-01
4.57804322e-01 2.70238638e-01 9.54246372e-02 -8.36135685e-01
-3.34469616e-01 -2.84910262e-01 3.25819910e-01 -1.59579918e-01
-1.53127447e-01 -1.18731193e-01 -1.37469983e+00 -2.62736112e-01
-6.90501094e-01 5.94617069e-01 2.79574156e-01 1.00565672e+00
4.35090572e-01 4.85888720e-01 1.93459094e-01 1.42341107e-01
-6.67178631e-01 -1.24367630e+00 -2.33140722e-01 3.05117995e-01
-1.94818527e-01 -5.56271195e-01 -1.66004345e-01 1.44978017e-02] | [10.840435028076172, 8.153508186340332] |
3b4fb46a-70fc-4683-b713-d46195ce3fa7 | equitable-multi-task-learning | 2306.09373 | null | https://arxiv.org/abs/2306.09373v2 | https://arxiv.org/pdf/2306.09373v2.pdf | Equitable Multi-task Learning | Multi-task learning (MTL) has achieved great success in various research domains, such as CV, NLP and IR etc. Due to the complex and competing task correlation, naive training all tasks may lead to inequitable learning, i.e. some tasks are learned well while others are overlooked. Multi-task optimization (MTO) aims to improve all tasks at same time, but conventional methods often perform poor when tasks with large loss scale or gradient norm magnitude difference. To solve the issue, we in-depth investigate the equity problem for MTL and find that regularizing relative contribution of different tasks (i.e. value of task-specific loss divides its raw gradient norm) in updating shared parameter can improve generalization performance of MTL. Based on our theoretical analysis, we propose a novel multi-task optimization method, named EMTL, to achieve equitable MTL. Specifically, we efficiently add variance regularization to make different tasks' relative contribution closer. Extensive experiments have been conduct to evaluate EMTL, our method stably outperforms state-of-the-art methods on the public benchmark datasets of two different research domains. Furthermore, offline and online A/B test on multi-task recommendation are conducted too. EMTL improves multi-task recommendation significantly, demonstrating the superiority and practicability of our method in industrial landscape. | ['Rui Zhang', 'Jun Yuan'] | 2023-06-15 | null | null | null | null | ['multi-task-learning'] | ['methodology'] | [ 1.24053277e-01 -6.80200100e-01 -3.43867987e-01 -3.51003677e-01
-8.98763955e-01 -3.75231206e-01 9.76032242e-02 -2.10313872e-01
-4.10515577e-01 8.72513175e-01 1.16966985e-01 -1.07037142e-01
-6.68942034e-01 -1.58707276e-01 -5.99835992e-01 -8.13980639e-01
2.88850248e-01 2.49892846e-01 5.54683320e-02 -2.30268016e-01
4.50445592e-01 -2.48331621e-01 -1.17288268e+00 2.70746768e-01
1.37181497e+00 1.04939961e+00 6.29530430e-01 1.17640407e-03
6.05996810e-02 4.49214280e-01 -6.35792851e-01 -5.76291144e-01
3.02805871e-01 1.73620209e-01 -7.29975164e-01 -7.39567578e-02
3.73752683e-01 2.41716996e-01 2.23325506e-01 1.06637418e+00
8.13691795e-01 3.25965405e-01 7.80197859e-01 -1.46007121e+00
-8.52900326e-01 3.57853144e-01 -1.00529897e+00 2.13615820e-01
-2.21331358e-01 -1.43184751e-01 1.18007863e+00 -9.53227162e-01
-9.78519842e-02 1.25459909e+00 8.09395254e-01 2.89886743e-01
-1.19024265e+00 -9.74464238e-01 5.20232558e-01 1.22844584e-01
-1.20901132e+00 -1.65301561e-02 7.49363661e-01 -5.29587805e-01
7.56995320e-01 1.15044855e-01 -2.20173955e-01 1.09274900e+00
6.38871729e-01 8.35577190e-01 1.17051637e+00 -7.17358589e-02
-2.94573933e-01 5.91899991e-01 2.84416974e-01 3.32911044e-01
3.69373053e-01 -1.99160501e-01 -4.29068118e-01 -1.69801578e-01
4.67654258e-01 1.86610550e-01 -5.67135327e-02 -1.26168475e-01
-1.37537563e+00 7.16235399e-01 1.53081819e-01 4.17072088e-01
-2.63014793e-01 8.58533233e-02 7.64531255e-01 6.06483161e-01
9.28361237e-01 3.36570859e-01 -8.19942534e-01 2.35515058e-01
-5.74055791e-01 3.44534129e-01 3.09136808e-01 6.70863748e-01
6.07591987e-01 1.83079168e-02 -6.11424029e-01 1.43179893e+00
2.96557516e-01 3.41550618e-01 7.94582486e-01 -5.92146993e-01
9.51815665e-01 4.05261695e-01 -3.94711867e-02 -1.15417612e+00
-4.49807525e-01 -8.81046891e-01 -1.15677953e+00 -1.26341254e-01
3.49710315e-01 -4.15765285e-01 -3.04231197e-01 1.76614606e+00
2.04048559e-01 1.17232993e-01 1.06683914e-02 9.04145002e-01
6.41871452e-01 5.94248593e-01 3.37139845e-01 -3.59336317e-01
1.32402658e+00 -1.14117718e+00 -7.61578262e-01 -4.64000225e-01
8.36569488e-01 -1.14720142e+00 1.40542710e+00 5.58231175e-01
-6.76795483e-01 -9.15355086e-01 -8.01742554e-01 1.44391164e-01
-1.48249283e-01 5.96891940e-01 7.47239411e-01 7.84399390e-01
-4.57823277e-01 5.57524145e-01 -1.62621558e-01 1.69086561e-01
4.79659140e-01 3.16397607e-01 -1.37411982e-01 -4.19226475e-02
-1.40060031e+00 7.54674911e-01 4.61484164e-01 1.93377361e-01
-8.72055292e-01 -1.05149329e+00 -4.07752454e-01 -7.95853510e-02
6.78690016e-01 -6.02203786e-01 1.04914713e+00 -9.72036064e-01
-1.15959144e+00 4.90234137e-01 1.19555313e-02 -1.39837593e-01
4.37409639e-01 -6.20155394e-01 -4.84720200e-01 -7.01972663e-01
2.50363380e-01 2.38043383e-01 8.98240626e-01 -1.18938875e+00
-8.13911617e-01 -4.68937129e-01 -2.56942343e-02 4.88176346e-01
-8.76919389e-01 5.52200340e-02 -1.35949388e-01 -9.87987816e-01
-2.48642311e-01 -8.01853657e-01 -2.53367484e-01 -3.03054154e-01
-3.29386145e-01 -5.96711099e-01 7.60473371e-01 -6.30114615e-01
1.59087884e+00 -2.23380589e+00 1.50458351e-01 -3.74851786e-02
3.37472737e-01 4.17954475e-01 -1.84256777e-01 9.59661826e-02
-2.94445343e-02 2.95014560e-01 -1.57791879e-02 -3.82373542e-01
-2.34647486e-02 9.76121724e-02 -4.14831862e-02 3.17434549e-01
-8.97357836e-02 7.98569918e-01 -6.52272463e-01 -5.66818476e-01
2.28262581e-02 3.66731524e-01 -4.40689534e-01 -2.30724812e-02
-2.17524663e-01 5.50400972e-01 -9.27473307e-01 4.56413090e-01
7.61568904e-01 -5.52183807e-01 -1.52909651e-01 -4.60444868e-01
3.49016627e-03 2.18474075e-01 -1.33636498e+00 1.70082021e+00
-9.25452352e-01 2.37660840e-01 -1.62777215e-01 -1.27519727e+00
9.51220691e-01 3.79284322e-01 5.72613478e-01 -7.04906464e-01
-1.25456646e-01 3.64914030e-01 1.13929451e-01 -5.62656760e-01
2.90934294e-01 -1.75053388e-01 2.16302704e-02 3.78278196e-01
-2.19101638e-01 3.40557456e-01 -2.03646600e-01 -2.04124257e-01
5.38963914e-01 7.87499249e-02 1.85612947e-01 -5.81791520e-01
7.61359453e-01 -2.13138804e-01 7.29828358e-01 6.33991897e-01
-2.92104930e-01 1.45271108e-01 2.40226969e-01 -2.91776240e-01
-8.24964941e-01 -6.42255545e-01 -3.29532325e-01 1.75108445e+00
1.77351743e-01 -2.05053762e-03 -2.78746158e-01 -8.68836820e-01
3.03724825e-01 5.14330745e-01 -4.94899958e-01 -2.03839034e-01
-5.14729261e-01 -1.44245601e+00 3.15947652e-01 2.42038563e-01
7.12218404e-01 -8.81369829e-01 -7.56383091e-02 3.25263768e-01
-3.44436377e-01 -9.89169896e-01 -8.77241433e-01 6.37466982e-02
-9.08860087e-01 -7.99148023e-01 -1.00970984e+00 -9.28688407e-01
3.38842422e-01 7.00519562e-01 1.08246398e+00 -1.42362550e-01
-6.63371524e-03 -2.83120442e-02 -3.47815603e-01 -5.29802442e-01
1.56489104e-01 3.99984032e-01 1.52021736e-01 2.54346430e-01
1.85187161e-02 -5.88580012e-01 -5.69620073e-01 8.78574610e-01
-6.61508977e-01 -9.91335884e-02 6.95632279e-01 9.47218359e-01
6.89294755e-01 3.60539526e-01 1.20656967e+00 -1.17408860e+00
1.08505166e+00 -6.39552712e-01 -3.78582895e-01 7.38683701e-01
-1.11718392e+00 6.40977696e-02 7.32823133e-01 -8.31726313e-01
-1.35719121e+00 -4.73160595e-01 8.81229043e-02 -4.46560562e-01
3.52434278e-01 5.72427869e-01 -5.48902564e-02 -1.82944164e-01
5.77681482e-01 2.03971162e-01 -2.02074915e-01 -5.87630033e-01
5.00929542e-03 7.56832540e-01 -9.77505818e-02 -8.74409437e-01
5.75277448e-01 8.39602649e-02 -2.27549989e-02 -4.03022498e-01
-1.37977612e+00 -5.23548841e-01 -1.71643525e-01 -1.67054638e-01
6.40023232e-01 -1.02608526e+00 -8.72526944e-01 4.17098492e-01
-1.12177372e+00 -1.50852337e-01 3.64016086e-01 7.39292383e-01
-1.60980806e-01 2.93191016e-01 -2.17768863e-01 -8.01744401e-01
-7.32812941e-01 -1.31592858e+00 1.04269457e+00 1.05369724e-02
3.52878064e-01 -1.26345193e+00 -5.17456532e-02 6.15759790e-01
6.20568275e-01 -1.26614630e-01 9.33932543e-01 -7.66449988e-01
-1.09204486e-01 2.31561378e-01 -4.99656558e-01 6.40768766e-01
1.70836106e-01 -4.44081068e-01 -9.57748771e-01 -4.88968849e-01
2.38779649e-01 -4.10363704e-01 8.30640554e-01 6.67633533e-01
1.51524234e+00 -2.92551070e-01 -3.62470716e-01 4.18172389e-01
1.46466446e+00 1.89120337e-01 3.25906068e-01 4.63038355e-01
9.08461869e-01 6.87217116e-01 8.94852698e-01 4.80397940e-01
4.07007188e-01 8.00520241e-01 1.90767556e-01 -3.75215919e-03
4.96250540e-02 1.76418900e-01 4.55578446e-01 1.06674850e+00
-1.13882773e-01 -1.86679572e-01 -6.52283251e-01 2.30307281e-01
-1.94857228e+00 -6.63651884e-01 -2.37176791e-01 2.30371428e+00
8.23711693e-01 2.63910085e-01 1.41353697e-01 -1.04710102e-01
7.96012163e-01 1.22575462e-01 -6.26424015e-01 -2.55339026e-01
-1.18412599e-01 -2.20239356e-01 5.68046033e-01 1.73931912e-01
-1.18138993e+00 6.47768855e-01 5.22713423e+00 1.65636456e+00
-1.09366572e+00 7.17419863e-01 9.67131197e-01 -1.51218325e-01
-3.65249306e-01 -3.01195264e-01 -1.07937872e+00 6.81722105e-01
4.16081667e-01 -5.33495843e-01 4.61396307e-01 8.49622726e-01
2.76778847e-01 1.34644791e-01 -7.23132491e-01 1.11455142e+00
-3.18095870e-02 -8.59689057e-01 4.57299575e-02 2.90468764e-02
8.54182363e-01 -2.59562433e-02 3.75488579e-01 8.11668396e-01
7.31415674e-02 -9.05534863e-01 5.06389797e-01 2.31062844e-01
6.00349367e-01 -5.65569758e-01 8.03718507e-01 5.65390348e-01
-1.41122353e+00 -3.32694739e-01 -6.83116555e-01 1.22644186e-01
-6.34836257e-02 1.00811064e+00 -2.92662978e-01 9.87328649e-01
6.53264225e-01 8.69153142e-01 -5.30281246e-01 9.14786935e-01
2.44788781e-01 6.01490736e-01 4.44800183e-02 -1.08329974e-01
2.22276732e-01 -4.10505384e-01 4.95079160e-01 1.05649865e+00
2.76712596e-01 -2.50388086e-01 5.28727114e-01 5.23031354e-01
-5.53523861e-02 5.69170177e-01 -3.76262307e-01 2.53169239e-01
4.22865242e-01 1.31769121e+00 -1.21158943e-01 -6.75229430e-02
-5.52029133e-01 7.13080049e-01 3.14563841e-01 3.91338706e-01
-1.02257144e+00 -2.01211989e-01 5.47476768e-01 -2.97206402e-01
5.00135077e-03 -1.70604736e-02 -5.60586393e-01 -1.01352727e+00
2.94780433e-01 -9.65931296e-01 5.00127017e-01 -3.29819858e-01
-1.71955550e+00 5.23512185e-01 -6.78847730e-02 -1.45725143e+00
3.64925146e-01 -4.35883403e-01 -5.86596131e-01 1.06202149e+00
-1.76121807e+00 -1.04607129e+00 -1.02992259e-01 6.56147242e-01
9.42067325e-01 -5.12244642e-01 3.07905287e-01 1.02384126e+00
-1.01154661e+00 9.69689250e-01 4.18027699e-01 -4.86924797e-01
1.09085834e+00 -1.01124549e+00 8.03746283e-03 2.78274417e-01
-7.65385479e-02 5.31800926e-01 4.60599989e-01 -5.59163690e-01
-1.25993001e+00 -1.23909640e+00 6.97825015e-01 -3.32207531e-01
6.76386714e-01 -1.86002731e-01 -8.62405241e-01 5.99689186e-01
-8.97785351e-02 -1.79493681e-01 6.13093376e-01 5.56827009e-01
-3.38443965e-01 -5.45440912e-01 -8.91934633e-01 3.02383721e-01
8.39011848e-01 -3.24533343e-01 -3.04165870e-01 8.34605634e-01
9.46399868e-01 -1.07016712e-01 -1.06807411e+00 6.68102980e-01
6.83942914e-01 -5.97182214e-01 1.18459129e+00 -6.51151955e-01
4.51123148e-01 -4.63648401e-02 -2.91775197e-01 -1.57123208e+00
-4.75115120e-01 -3.94066036e-01 3.08933437e-01 1.35800874e+00
6.06369376e-01 -9.76983666e-01 3.73953313e-01 2.96207011e-01
-2.52171993e-01 -1.05814254e+00 -8.42670500e-01 -1.10219407e+00
3.09311718e-01 -4.36951429e-01 3.65792543e-01 1.15563178e+00
-4.47106093e-01 6.61270499e-01 -8.82056773e-01 -5.39433956e-02
7.25430667e-01 1.05710672e-02 6.06427550e-01 -1.45554149e+00
-3.68002772e-01 -5.60573518e-01 2.89498508e-01 -1.00567532e+00
2.50194520e-01 -1.01552129e+00 -1.76294401e-01 -1.26793194e+00
4.18213993e-01 -9.81931329e-01 -6.96864486e-01 5.26134789e-01
-7.31046557e-01 -1.03855997e-01 1.01400442e-01 5.44681668e-01
-7.58130908e-01 6.32932127e-01 1.70670581e+00 -8.19139257e-02
-2.25044265e-01 2.67957211e-01 -8.45743239e-01 4.52683330e-01
8.02802384e-01 -7.52234638e-01 -5.71853101e-01 -6.49718344e-01
2.89120376e-01 -7.81521797e-02 2.14594975e-02 -8.05875063e-01
-5.40575385e-02 -2.78948009e-01 1.87579691e-01 -1.83648914e-01
2.52665877e-02 -7.54558980e-01 1.54703572e-01 2.91299284e-01
-4.74000126e-01 9.13974345e-02 1.80641726e-01 6.16786957e-01
-1.65086269e-01 -3.21329176e-01 5.44674516e-01 5.75349815e-02
-5.95182717e-01 4.55096245e-01 2.62780666e-01 2.79214263e-01
9.60259676e-01 1.42531291e-01 -4.18006808e-01 1.31511301e-01
-3.48455310e-01 6.02901399e-01 -4.31155890e-01 5.99858701e-01
2.58135110e-01 -1.47685409e+00 -1.06750846e+00 -2.40370944e-01
9.83543620e-02 -2.40055174e-01 6.04512215e-01 1.18515432e+00
1.99625790e-01 5.15326262e-01 -2.70325895e-02 -4.39191520e-01
-1.26392233e+00 4.46037561e-01 3.59678954e-01 -8.39632511e-01
-2.00380966e-01 8.11972737e-01 6.62668526e-01 -6.74231529e-01
3.45840335e-01 5.10688946e-02 -3.82933080e-01 1.76751003e-01
1.90551817e-01 5.95029593e-01 1.82037488e-01 -2.49819949e-01
-3.38349581e-01 7.40756392e-01 -3.86131614e-01 3.13481241e-01
1.09156489e+00 -2.11639822e-01 -1.73092913e-02 6.27108037e-01
1.24990475e+00 -2.05142707e-01 -9.64537084e-01 -4.87720013e-01
4.63451445e-02 -5.46966314e-01 2.49473065e-01 -8.18196237e-01
-1.16502309e+00 8.40537906e-01 6.45460427e-01 1.92177519e-01
1.15569818e+00 -3.12116832e-01 9.13524747e-01 4.51029271e-01
3.89183491e-01 -1.21318698e+00 4.79605377e-01 4.30652082e-01
8.59931648e-01 -1.72807455e+00 2.03224197e-01 -4.04914737e-01
-1.08086896e+00 7.37059057e-01 7.92500019e-01 6.19204119e-02
7.62137890e-01 -1.77331895e-01 -2.15267465e-01 -6.54432401e-02
-8.13092768e-01 1.58125103e-01 6.39939487e-01 1.70630530e-01
6.41189396e-01 3.31460834e-02 -7.86165833e-01 8.11544836e-01
3.25854957e-01 -3.04143168e-02 -1.94479883e-01 4.39156204e-01
-4.97909874e-01 -1.14955962e+00 -3.71212453e-01 8.29646587e-01
-7.91499913e-01 -2.28446215e-01 2.64217675e-01 6.23600125e-01
2.79551297e-01 9.99781847e-01 -3.93402427e-01 -4.96105164e-01
4.51240063e-01 -1.12075180e-01 2.16241196e-01 -3.21110219e-01
-1.00086284e+00 2.13522092e-01 3.63010690e-02 -2.65651137e-01
-4.34027821e-01 -5.44521630e-01 -7.13245690e-01 -7.96555132e-02
-5.86153507e-01 1.87104151e-01 6.46549404e-01 1.13372183e+00
2.76199907e-01 8.20693970e-01 9.78708923e-01 -4.46816534e-01
-1.07143986e+00 -1.11537659e+00 -6.20174289e-01 4.74673301e-01
1.06879763e-01 -1.02030456e+00 -3.73076767e-01 -1.90660030e-01] | [9.441699981689453, 4.093064785003662] |
28761f50-9fde-4b82-a7b7-cd9c2fd60195 | beyond-learned-metadata-based-raw-image | 2306.12058 | null | https://arxiv.org/abs/2306.12058v1 | https://arxiv.org/pdf/2306.12058v1.pdf | Beyond Learned Metadata-based Raw Image Reconstruction | While raw images have distinct advantages over sRGB images, e.g., linearity and fine-grained quantization levels, they are not widely adopted by general users due to their substantial storage requirements. Very recent studies propose to compress raw images by designing sampling masks within the pixel space of the raw image. However, these approaches often leave space for pursuing more effective image representations and compact metadata. In this work, we propose a novel framework that learns a compact representation in the latent space, serving as metadata, in an end-to-end manner. Compared with lossy image compression, we analyze the intrinsic difference of the raw image reconstruction task caused by rich information from the sRGB image. Based on the analysis, a novel backbone design with asymmetric and hybrid spatial feature resolutions is proposed, which significantly improves the rate-distortion performance. Besides, we propose a novel design of the context model, which can better predict the order masks of encoding/decoding based on both the sRGB image and the masks of already processed features. Benefited from the better modeling of the correlation between order masks, the already processed information can be better utilized. Moreover, a novel sRGB-guided adaptive quantization precision strategy, which dynamically assigns varying levels of quantization precision to different regions, further enhances the representation ability of the model. Finally, based on the iterative properties of the proposed context model, we propose a novel strategy to achieve variable bit rates using a single model. This strategy allows for the continuous convergence of a wide range of bit rates. Extensive experimental results demonstrate that the proposed method can achieve better reconstruction quality with a smaller metadata size. | ['Bihan Wen', 'Alex C. Kot', 'Lap-Pui Chau', 'Lanqing Guo', 'Wenhan Yang', 'Yi Yu', 'YuFei Wang'] | 2023-06-21 | null | null | null | null | ['image-reconstruction', 'image-compression', 'quantization'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 5.59181988e-01 -3.71470302e-01 -4.38278139e-01 -3.11008066e-01
-4.32863355e-01 2.85665272e-03 2.32130423e-01 -3.30888145e-02
-2.43782356e-01 4.76276338e-01 3.39597315e-01 5.26974536e-02
-4.50109988e-01 -1.05713558e+00 -5.13864458e-01 -1.03330481e+00
1.89514130e-01 -2.57825702e-01 2.41646141e-01 -6.16257638e-02
4.72349524e-01 3.10803592e-01 -1.70373464e+00 3.53576452e-01
1.11023116e+00 1.48878455e+00 1.01097739e+00 2.46208593e-01
-3.39907974e-01 6.72200918e-01 -3.70898008e-01 3.78743820e-02
4.52667624e-01 -4.08307821e-01 -2.42996737e-01 3.69419515e-01
2.14371249e-01 -6.99292660e-01 -6.11594558e-01 1.20360541e+00
3.83204103e-01 3.21080349e-02 2.36031473e-01 -6.23550773e-01
-6.13652647e-01 5.48329473e-01 -4.69870210e-01 5.45374714e-02
5.81531636e-02 2.34867096e-01 8.57237101e-01 -5.56910336e-01
4.56376255e-01 1.22228098e+00 2.91468829e-01 2.38774762e-01
-1.15944433e+00 -8.09357524e-01 6.24665618e-02 5.49595654e-01
-1.63422370e+00 -5.64112842e-01 8.49914014e-01 -1.57514494e-02
5.30405700e-01 3.58000726e-01 6.10242486e-01 4.56496477e-01
7.45278001e-02 4.53186363e-01 1.16762424e+00 -4.56965327e-01
1.53704688e-01 6.04941100e-02 -1.15139559e-01 5.32850504e-01
3.08532238e-01 -6.07041717e-02 -5.97309768e-01 1.49781898e-01
1.07708216e+00 2.81938940e-01 -7.18116224e-01 -5.20137846e-01
-1.13555753e+00 5.12138724e-01 4.54497576e-01 4.09407169e-01
-4.71184045e-01 1.57740548e-01 1.91589028e-01 1.79099925e-02
1.09770939e-01 -3.01280543e-02 -1.16116829e-01 -5.30907065e-02
-1.18728960e+00 -5.40775023e-02 1.46021023e-01 1.00345802e+00
1.00287938e+00 1.28365457e-02 -4.00619775e-01 8.87763143e-01
3.34523350e-01 4.79677647e-01 7.32922733e-01 -9.14254248e-01
7.22142041e-01 6.02541447e-01 2.53371242e-02 -1.36211324e+00
-7.94618726e-02 -6.42066658e-01 -1.19643831e+00 -2.95482911e-02
1.78417359e-02 6.27145767e-01 -7.01547384e-01 1.64003742e+00
1.82226107e-01 2.33857423e-01 9.19649079e-02 1.01466787e+00
3.51750731e-01 1.01415598e+00 -5.77957183e-02 -6.15466595e-01
1.38108623e+00 -6.54420853e-01 -1.00713301e+00 1.55969858e-01
3.52234453e-01 -6.08732581e-01 1.16810548e+00 5.42547703e-01
-1.10538864e+00 -8.74426365e-01 -1.28329122e+00 -2.25006551e-01
9.36992392e-02 4.15054649e-01 3.61102194e-01 5.95855355e-01
-9.54014063e-01 6.70805871e-01 -7.50681698e-01 8.34830701e-02
3.96389276e-01 4.11092132e-01 2.40448732e-02 -3.50403070e-01
-9.55740273e-01 3.70127559e-01 6.59605920e-01 9.10836831e-02
-4.20149803e-01 -5.20074546e-01 -5.33040345e-01 5.12085736e-01
3.14811319e-01 -3.97695482e-01 7.08999336e-01 -8.25726151e-01
-1.49590898e+00 3.19100738e-01 -3.26016158e-01 -4.40259635e-01
3.36744279e-01 2.87306253e-02 -4.10343587e-01 5.27578354e-01
-1.49295703e-01 5.08630991e-01 1.08829165e+00 -1.11066997e+00
-7.55691648e-01 -5.31781375e-01 -1.37638003e-01 2.23887309e-01
-8.99787068e-01 -4.23670083e-01 -6.79173112e-01 -8.37392807e-01
4.92628783e-01 -5.23030579e-01 -2.66989738e-01 2.41901323e-01
-4.74395752e-02 1.16647050e-01 8.36174130e-01 -6.20226085e-01
1.72771657e+00 -2.56411958e+00 8.66570473e-02 2.71913797e-01
2.24080637e-01 2.95675576e-01 -3.52909230e-03 1.99777529e-01
2.29048997e-01 8.45030099e-02 -2.47625485e-01 -2.02946693e-01
-1.81136876e-01 2.37453580e-01 -3.97214174e-01 3.32485467e-01
-6.26179054e-02 4.83731836e-01 -6.65879488e-01 -5.49972057e-01
3.05558205e-01 5.12246668e-01 -6.32474244e-01 2.51609564e-01
-1.21318121e-02 3.77727449e-01 -6.06878579e-01 3.12222719e-01
1.08925354e+00 -4.15563226e-01 3.70646358e-01 -6.83221698e-01
-2.17961013e-01 2.32756749e-01 -1.40934563e+00 1.72391129e+00
-6.02893651e-01 1.91550836e-01 1.83579564e-01 -1.06092680e+00
1.30466175e+00 6.19399101e-02 3.27171326e-01 -1.08243251e+00
-1.11860462e-01 2.66833514e-01 -3.16781163e-01 -3.09793174e-01
7.23402679e-01 2.31590681e-03 1.83207288e-01 9.67635736e-02
-3.29347432e-01 8.85803923e-02 1.07142717e-01 -1.80640016e-02
7.04530239e-01 -5.35202436e-02 3.46805751e-01 -1.56675369e-01
8.48769367e-01 -4.19806868e-01 7.58142054e-01 5.16985774e-01
-7.80254006e-02 3.89757931e-01 4.04103845e-02 -3.77643585e-01
-1.13709950e+00 -8.60168099e-01 -3.69272053e-01 5.47469974e-01
7.50907600e-01 -6.65758431e-01 -6.06831491e-01 -1.32044986e-01
-2.71737665e-01 4.70451355e-01 -6.35709167e-02 -2.99344331e-01
-8.07876050e-01 -5.78704178e-01 2.00030223e-01 1.24949522e-01
1.04396617e+00 -6.41364574e-01 -7.64210999e-01 2.80081421e-01
-3.29441845e-01 -1.06970477e+00 -3.45834345e-01 -2.67126560e-01
-1.27231646e+00 -6.25204325e-01 -5.20732701e-01 -6.67119503e-01
5.22583425e-01 6.96510434e-01 5.86468995e-01 3.57100904e-01
-1.03085086e-01 4.66756634e-02 -5.15741169e-01 3.56818885e-01
-4.15425301e-01 -4.66895215e-02 -9.37468838e-03 3.39012295e-01
8.88060182e-02 -7.37353861e-01 -1.03868616e+00 2.21174330e-01
-1.25917280e+00 4.86879408e-01 9.46875215e-01 8.73501599e-01
9.95046854e-01 4.27421182e-01 4.65285689e-01 -4.64808345e-01
3.08266193e-01 -3.34353685e-01 -7.15618491e-01 1.95482850e-01
-9.42289650e-01 2.85165519e-01 9.70593691e-01 -2.21308962e-01
-1.08490837e+00 -1.54312864e-01 -1.15515426e-01 -4.26167041e-01
3.93641219e-02 1.26834780e-01 -4.20290381e-01 -9.42895561e-02
1.24559380e-01 8.35461855e-01 2.78451383e-01 -6.97663128e-01
4.24067736e-01 8.91003847e-01 4.58740324e-01 -4.34879541e-01
6.29381597e-01 5.62143326e-01 1.13715477e-01 -7.63601303e-01
-1.75755084e-01 -2.75780618e-01 -3.55478376e-01 4.40246388e-02
5.89541614e-01 -1.07837653e+00 -5.63375473e-01 3.01339328e-01
-8.86488914e-01 8.87707695e-02 -2.00748235e-01 5.34567714e-01
-6.75179660e-01 8.86441469e-01 -6.86073720e-01 -6.31848156e-01
-3.36191684e-01 -1.49663103e+00 9.05505717e-01 2.19303325e-01
3.66533816e-01 -4.06183124e-01 -5.52248478e-01 2.72076368e-01
6.38926029e-01 -2.58013934e-01 1.09185374e+00 8.68523121e-02
-1.33165109e+00 2.74633855e-01 -6.04044855e-01 4.64575976e-01
3.41819584e-01 -3.85994136e-01 -5.88053286e-01 -4.27006036e-01
2.51950502e-01 -8.61307420e-03 9.42871273e-01 2.98358828e-01
1.79066539e+00 -5.56643426e-01 -2.60712594e-01 9.77777958e-01
1.87832975e+00 3.01796734e-01 1.10423172e+00 4.57726926e-01
5.50205588e-01 2.44323030e-01 9.00983274e-01 8.92637253e-01
2.85394847e-01 1.00183952e+00 6.21800780e-01 2.31288210e-01
-3.05391669e-01 -3.04537922e-01 2.56246179e-01 1.09769726e+00
-2.98992172e-03 -1.11804947e-01 -2.55250007e-01 2.87191451e-01
-1.50495291e+00 -8.98673236e-01 2.18852714e-01 2.43573904e+00
9.18450892e-01 5.73928505e-02 -4.15725708e-01 3.99598211e-01
6.45892978e-01 5.16016901e-01 -4.18409854e-01 -1.35193482e-01
-1.64434686e-01 1.39750302e-01 6.75970733e-01 1.87725708e-01
-7.12952912e-01 6.52171612e-01 5.37765169e+00 1.60236955e+00
-1.40799510e+00 -1.27474219e-02 6.56434894e-01 -1.09643161e-01
-5.31280041e-01 -3.16167809e-02 -9.90726590e-01 9.14221287e-01
7.60359704e-01 -1.04255579e-01 6.70663357e-01 6.89444959e-01
4.74111259e-01 -2.12515704e-02 -6.15194261e-01 1.33230817e+00
-9.62041542e-02 -1.52717221e+00 5.33524573e-01 2.26467267e-01
5.38757503e-01 -5.17774761e-01 2.60991395e-01 -6.25844523e-02
-4.51845527e-01 -7.09242642e-01 9.33112621e-01 5.55950582e-01
1.13002396e+00 -7.00364947e-01 4.19716865e-01 4.38553810e-01
-1.28668988e+00 -4.76758897e-01 -7.16655076e-01 -2.62821186e-02
5.76136746e-02 8.13212454e-01 -4.47844714e-01 7.07751513e-01
7.07641840e-01 7.39574134e-01 -4.73894656e-01 7.83962369e-01
1.34505015e-02 5.29033124e-01 -1.53967157e-01 1.82340741e-01
5.75868823e-02 -5.37607670e-01 4.08681691e-01 9.96933281e-01
8.14992666e-01 3.06247532e-01 1.15159778e-02 8.49916637e-01
6.15463071e-02 3.51022303e-01 -4.46861267e-01 1.63077489e-01
7.66917169e-01 1.05064344e+00 -4.37135309e-01 -3.62450838e-01
-3.91735017e-01 8.62960160e-01 1.35537118e-01 3.29036653e-01
-6.12835705e-01 -2.09339991e-01 5.22578776e-01 3.24601829e-01
7.53253400e-01 -3.51368964e-01 -4.35479432e-01 -1.08418894e+00
2.92010099e-01 -9.03707027e-01 -3.45611275e-04 -5.15976548e-01
-7.63927579e-01 3.99522662e-01 -8.56670216e-02 -1.66323066e+00
-1.68392546e-02 -2.48312846e-01 -7.31225386e-02 8.08626413e-01
-1.82498658e+00 -9.25936460e-01 -5.25230944e-01 6.47883356e-01
5.17655492e-01 -7.65026780e-03 5.44019461e-01 5.30294657e-01
-4.61277515e-01 6.99631691e-01 4.72866684e-01 -3.46484423e-01
4.72564220e-01 -5.50450206e-01 -1.68608621e-01 8.81435156e-01
-6.36795908e-02 6.42635107e-01 4.06283259e-01 -5.00587463e-01
-1.66044104e+00 -1.11573541e+00 3.86440486e-01 4.66857433e-01
1.43810675e-01 -7.45469406e-02 -1.12220335e+00 1.75778523e-01
-1.36354193e-01 -3.81714036e-03 4.46406126e-01 -4.83492732e-01
-2.41931215e-01 -7.98626006e-01 -1.21436512e+00 5.06798804e-01
1.00799263e+00 -3.45274597e-01 -1.91703469e-01 -4.27117618e-03
9.41485643e-01 -1.80342659e-01 -9.44490433e-01 5.65048039e-01
4.85095799e-01 -1.16632950e+00 1.11547458e+00 2.55756527e-01
5.83362222e-01 -5.85184455e-01 -5.32818556e-01 -8.95258665e-01
-5.47473431e-01 -2.68908709e-01 -3.01513731e-01 1.28548896e+00
-2.45580882e-01 -5.55551767e-01 6.20974064e-01 2.28140399e-01
-7.03271329e-02 -9.62040305e-01 -1.11163497e+00 -8.00649941e-01
-4.87944812e-01 -1.90688238e-01 9.23267782e-01 4.87900257e-01
-3.08998197e-01 -2.46506438e-01 -4.94187057e-01 2.83243179e-01
7.68120527e-01 3.84679228e-01 5.87430894e-01 -8.73524308e-01
-4.90752041e-01 -4.54147518e-01 -6.43647909e-01 -1.83591378e+00
-2.95719624e-01 -6.59430087e-01 -5.22252247e-02 -1.27760971e+00
2.71834582e-01 -1.01311946e+00 -5.62841058e-01 9.28112045e-02
-1.46234870e-01 3.36367667e-01 3.44390303e-01 6.70267880e-01
-4.91519034e-01 7.89966941e-01 1.43019533e+00 -1.15181297e-01
-1.28653124e-01 -2.91139722e-01 -5.86344302e-01 2.94375330e-01
6.83752060e-01 -1.36102974e-01 -5.66744268e-01 -6.23325586e-01
3.53180245e-02 2.04320088e-01 2.14353681e-01 -1.14924812e+00
3.76673386e-04 -1.92083642e-01 3.70286405e-01 -6.22822583e-01
3.73949230e-01 -9.41936135e-01 3.21220517e-01 3.70974660e-01
-1.72744229e-01 -3.01098853e-01 -4.36689220e-02 7.74817586e-01
-3.93072188e-01 -2.48409405e-01 8.65954578e-01 -8.29808265e-02
-9.08892214e-01 3.61432195e-01 -9.97919962e-03 -5.79168856e-01
8.87090802e-01 -5.87704539e-01 -1.79395989e-01 -2.62669355e-01
-2.59286821e-01 -8.27448815e-02 8.08127880e-01 2.10765973e-01
8.81672144e-01 -1.44460964e+00 -4.60974813e-01 5.24087429e-01
-5.36922775e-02 2.62135752e-02 6.64525270e-01 6.55555308e-01
-5.64814150e-01 4.34213817e-01 -3.59180033e-01 -7.40019977e-01
-1.11530840e+00 7.51402855e-01 -1.83421522e-01 -4.64539856e-01
-7.98601866e-01 3.60606581e-01 3.83941591e-01 3.52517903e-01
-1.89273413e-02 -4.66867268e-01 -2.12249845e-01 -2.53220052e-01
9.02894080e-01 3.40009212e-01 -1.80396624e-02 -7.49579847e-01
9.02752206e-02 7.40879238e-01 -5.25490455e-02 2.02157080e-01
1.16265905e+00 -8.25833738e-01 -2.38903180e-01 9.40358043e-02
1.15838516e+00 1.56961307e-01 -1.39996052e+00 -4.04523402e-01
-2.87800729e-01 -1.12165070e+00 3.49588633e-01 -2.39042491e-01
-1.21479106e+00 8.90265405e-01 9.52558756e-01 6.57567307e-02
1.85178554e+00 -3.30894142e-01 1.11754405e+00 -1.01539873e-01
7.28903592e-01 -8.87645483e-01 1.85630433e-02 -3.66388150e-02
6.13240778e-01 -8.51825356e-01 1.76496983e-01 -6.30939901e-01
-1.08297758e-01 1.21376884e+00 3.34652901e-01 1.19936697e-01
2.51167476e-01 -5.08793741e-02 -2.72156388e-01 1.46856919e-01
-4.83544916e-01 8.61964747e-02 3.23840789e-02 3.95976394e-01
6.85969889e-02 1.42985761e-01 -6.40163004e-01 6.26833022e-01
-8.03950615e-03 2.37244248e-01 4.76046056e-01 6.35764360e-01
-7.36779690e-01 -1.24325633e+00 -5.13568342e-01 4.75981146e-01
-3.25864613e-01 -1.77408502e-01 6.36939228e-01 2.81413406e-01
3.98365647e-01 7.86355019e-01 7.58323297e-02 -4.04381424e-01
8.80075395e-02 -3.29633534e-01 4.81093615e-01 -3.03754807e-01
9.79384780e-02 1.55428335e-01 -4.65939373e-01 -6.92999959e-01
-4.73508000e-01 -2.51646340e-01 -1.22341955e+00 -2.88218945e-01
-3.58224809e-01 1.00664221e-01 6.01793945e-01 7.00602770e-01
6.48010969e-01 3.62669259e-01 8.89844716e-01 -7.58619249e-01
-7.84795225e-01 -7.10707605e-01 -8.10626447e-01 4.58275676e-01
2.78101206e-01 -5.87167263e-01 -1.63166344e-01 3.47938351e-02] | [11.250225067138672, -1.6791279315948486] |
29c1b3a0-a18c-4b54-bd48-d692c7cf9db9 | concentrated-document-topic-model | 2102.04449 | null | https://arxiv.org/abs/2102.04449v1 | https://arxiv.org/pdf/2102.04449v1.pdf | Concentrated Document Topic Model | We propose a Concentrated Document Topic Model(CDTM) for unsupervised text classification, which is able to produce a concentrated and sparse document topic distribution. In particular, an exponential entropy penalty is imposed on the document topic distribution. Documents that have diverse topic distributions are penalized more, while those having concentrated topics are penalized less. We apply the model to the benchmark NIPS dataset and observe more coherent topics and more concentrated and sparse document-topic distributions than Latent Dirichlet Allocation(LDA). | ['Ying Chen', 'Hao Lei'] | 2021-02-06 | null | null | null | null | ['unsupervised-text-classification'] | ['natural-language-processing'] | [-1.03897631e-01 4.66445923e-01 -6.79389715e-01 -6.32397294e-01
-5.85699320e-01 -2.59844270e-02 8.65531564e-01 4.14912671e-01
-8.72403458e-02 7.01727867e-01 5.72319508e-01 1.06222175e-01
-3.65060978e-02 -8.61816108e-01 -3.61835212e-01 -1.05377793e+00
-1.61547828e-02 1.06939399e+00 -4.08524610e-02 4.33536649e-01
3.33828062e-01 -3.85314673e-01 -1.37048221e+00 1.76702321e-01
1.07784963e+00 5.00209689e-01 5.23320854e-01 3.50141734e-01
-4.17302281e-01 3.65615964e-01 -8.71234179e-01 3.91290300e-02
-4.55179781e-01 -2.48331398e-01 -6.06481314e-01 2.71365762e-01
2.49546513e-01 -4.43824142e-01 1.27639920e-01 9.43084538e-01
1.89474553e-01 5.31105340e-01 1.50115871e+00 -1.28030753e+00
-5.85837305e-01 1.13769329e+00 -8.62221897e-01 -1.19525567e-02
-1.18816726e-01 -5.68328381e-01 1.26759589e+00 -8.22682559e-01
7.19850957e-01 1.54289341e+00 1.35489389e-01 3.44639242e-01
-1.44567275e+00 -5.86082399e-01 6.33683264e-01 -2.64290661e-01
-1.20657909e+00 1.20703526e-01 1.12976718e+00 -4.98558581e-01
7.30184019e-01 -2.89282799e-01 5.36821187e-01 1.64602387e+00
6.41333878e-01 1.27068579e+00 5.61423838e-01 -3.17152917e-01
7.90519953e-01 4.89118665e-01 6.50187731e-01 -1.41146123e-01
5.33938646e-01 -5.72890043e-01 -5.69520593e-01 -5.35698652e-01
2.18619242e-01 4.88753110e-01 -1.71907514e-01 -6.19149804e-01
-8.71148527e-01 1.53008330e+00 -1.98079851e-02 4.75248754e-01
-6.55934513e-01 2.79297344e-02 5.33555210e-01 9.86554995e-02
1.40473366e+00 1.38770521e-01 -2.11119205e-01 -1.67059928e-01
-1.35366988e+00 6.36007965e-01 9.33042347e-01 9.82178032e-01
8.67381394e-01 -8.71273875e-02 -3.83366525e-01 1.19952643e+00
6.31821454e-01 4.98922318e-01 9.04428363e-01 -6.18826032e-01
3.75051260e-01 4.50488389e-01 -4.59110476e-02 -1.14116633e+00
-1.57736287e-01 -2.88185745e-01 -9.47746992e-01 -2.84061074e-01
-2.09770679e-01 -2.63989300e-01 -9.57625806e-01 1.54526865e+00
9.27654803e-02 -2.67175168e-01 9.83164012e-02 5.81987262e-01
6.27438009e-01 1.41052175e+00 5.00464797e-01 -2.94102460e-01
1.21835685e+00 -6.75216079e-01 -1.14557743e+00 -3.71421240e-02
5.30890048e-01 -3.60809863e-01 9.08058941e-01 4.08184350e-01
-9.12807584e-01 -1.38107881e-01 -6.03257060e-01 -8.99594128e-02
-3.85776281e-01 -9.44950990e-03 4.97347116e-01 4.07099754e-01
-8.49119246e-01 3.05963755e-01 -9.85665083e-01 -5.03350198e-01
6.33322895e-01 -1.60455648e-02 -2.46052071e-03 1.56342790e-01
-1.18968391e+00 3.38384390e-01 7.42228627e-01 -6.27276659e-01
-1.02014124e+00 -8.47259700e-01 -9.52699125e-01 5.88969469e-01
-1.24187529e-01 -4.84179944e-01 1.14773750e+00 -3.86998236e-01
-1.51201761e+00 5.08979261e-01 -4.61329252e-01 -8.64657164e-01
5.98057881e-02 -3.82703304e-01 1.46521911e-01 2.86306709e-01
2.79081706e-02 1.01288402e+00 1.11114824e+00 -1.28139400e+00
-6.93496764e-01 -2.90290415e-01 -7.04841256e-01 4.34882015e-01
-9.72854018e-01 -1.11864515e-01 -7.63445124e-02 -9.88698304e-01
1.92770839e-01 -7.95670629e-01 -5.82023300e-02 -4.46350843e-01
-7.72777855e-01 -9.40526068e-01 1.43423462e+00 -3.63765180e-01
1.00241351e+00 -2.13977027e+00 2.45729998e-01 3.27451080e-02
5.26180983e-01 -4.93500054e-01 3.37797999e-01 3.33892792e-01
2.04405770e-01 4.18601669e-02 -8.93665552e-02 -8.14815700e-01
3.66653025e-01 2.31197044e-01 -9.58751976e-01 2.79302716e-01
-8.92311655e-05 2.84194440e-01 -7.44801819e-01 -5.84560215e-01
-2.23048851e-01 4.57178086e-01 -8.07423115e-01 2.28867427e-01
-7.08719492e-01 4.85358015e-02 -6.67835593e-01 3.27730030e-01
8.31520438e-01 -3.65086347e-01 2.16348127e-01 3.94924402e-01
-3.36206635e-04 4.42923725e-01 -7.00802326e-01 1.34701872e+00
-3.21530342e-01 1.03576601e+00 -1.40094683e-01 -1.00606132e+00
1.40171540e+00 5.75314760e-01 6.07642114e-01 2.28415467e-02
2.37219974e-01 -2.94915885e-01 -3.59861672e-01 8.14405233e-02
1.19166338e+00 -8.72486308e-02 -1.16746284e-01 1.04047894e+00
3.56688350e-01 -3.13364983e-01 1.04648896e-01 7.32259572e-01
5.03623009e-01 -3.13802928e-01 -1.21164285e-01 -7.54106998e-01
-3.83620799e-01 -1.00823246e-01 2.29133219e-01 8.49432409e-01
1.54610753e-01 5.22110224e-01 9.94136453e-01 1.58228818e-02
-1.14465928e+00 -9.87225771e-01 -4.08026576e-01 1.61581087e+00
-8.80495310e-02 -4.39338773e-01 -4.63912487e-01 -6.41598880e-01
-3.54206678e-03 1.02863955e+00 -6.28869832e-01 -1.27509162e-01
-8.72290730e-02 -7.53053963e-01 5.14825173e-02 3.82671326e-01
8.43967125e-02 -9.72214222e-01 -3.85568649e-01 1.14148930e-01
-2.57446319e-01 -5.02611876e-01 -4.22949970e-01 6.16756916e-01
-1.13954484e+00 -3.23843956e-01 -1.53658438e+00 -8.85406315e-01
7.92030931e-01 -5.48467785e-02 1.01949799e+00 -8.66474748e-01
1.80436835e-01 2.43606612e-01 -4.80817884e-01 -5.95076025e-01
-2.26001397e-01 5.04852891e-01 9.06712115e-02 -3.62402648e-01
6.57698095e-01 -3.95509899e-01 -3.79922271e-01 1.82553783e-01
-9.01273608e-01 -1.53055489e-01 1.77580908e-01 1.07181013e+00
1.85970917e-01 5.33337235e-01 8.59312892e-01 -1.09682369e+00
1.02387691e+00 -1.01423025e+00 -3.05144578e-01 -2.85431802e-01
-9.55159843e-01 -5.43312505e-02 3.40211630e-01 -8.10587704e-01
-1.52792668e+00 -4.83325392e-01 3.60543311e-01 -2.84691274e-01
-4.99523401e-01 6.23076141e-01 1.48750067e-01 8.73764515e-01
3.77007216e-01 3.36238116e-01 -2.91105628e-01 -7.29959786e-01
3.20096493e-01 9.43902612e-01 -6.18578214e-03 -9.04896498e-01
2.12258652e-01 6.32468224e-01 -6.28610373e-01 -1.13522410e+00
-6.17592633e-01 -7.28099346e-01 -3.32179755e-01 8.39243680e-02
7.07731068e-01 -8.80546212e-01 -4.95087728e-02 4.77852821e-01
-1.11088204e+00 -2.40161717e-01 -4.99452651e-01 7.84503579e-01
-5.43134928e-01 1.83459073e-01 -5.95080256e-01 -1.03970218e+00
-6.63717389e-01 -7.41842687e-01 1.24314153e+00 2.79624045e-01
-5.73640823e-01 -1.51123738e+00 4.33542669e-01 -3.83092284e-01
5.29375196e-01 -1.68470100e-01 1.18544734e+00 -1.19899845e+00
3.12626958e-02 -1.60737738e-01 -1.20023258e-01 2.28312746e-01
3.29254806e-01 1.22485153e-01 -8.17171991e-01 -4.42697883e-01
1.04006091e-02 -3.47262412e-01 1.29955041e+00 9.43213284e-01
8.99467826e-01 -4.27925080e-01 -8.02017689e-01 1.59658208e-01
9.04510677e-01 2.55067110e-01 2.91853994e-01 1.82565138e-01
3.70501786e-01 7.00229049e-01 5.40387154e-01 6.65999889e-01
5.43102562e-01 3.97338033e-01 -7.13924170e-02 1.97423398e-01
3.64721119e-01 -4.06280607e-01 3.15508425e-01 7.85900593e-01
5.73865771e-01 -9.13904369e-01 -9.38736737e-01 1.06172299e+00
-1.67253780e+00 -7.25299656e-01 -5.10426722e-02 1.77256453e+00
1.08672392e+00 1.52325422e-01 1.56205386e-01 1.66569427e-02
1.08973122e+00 4.60056961e-01 -5.18511355e-01 -2.91696191e-01
3.81656960e-02 -1.98389098e-01 -1.85381487e-01 2.16769904e-01
-1.26234782e+00 1.06788731e+00 7.34341192e+00 8.28535140e-01
-1.04523039e+00 -1.06206335e-01 8.09614837e-01 -2.33297527e-01
-4.23057735e-01 -1.27689362e-01 -1.14846122e+00 9.17566419e-01
9.08299506e-01 -7.39086330e-01 -6.21005654e-01 1.45189893e+00
9.62230414e-02 -3.23213160e-01 -9.22621429e-01 5.59675515e-01
6.69269636e-02 -8.85381043e-01 4.62572664e-01 3.01899076e-01
1.08406293e+00 -7.01593533e-02 5.47300994e-01 4.77186024e-01
7.36258626e-01 -5.03568530e-01 3.34257275e-01 3.35436821e-01
3.59260857e-01 -7.80401528e-01 7.75175154e-01 4.95108664e-01
-5.73547006e-01 9.92474779e-02 -9.91682947e-01 3.11951548e-01
2.51198739e-01 1.24632263e+00 -9.98230696e-01 9.24484730e-02
8.92318487e-01 9.55965936e-01 1.77656874e-01 8.97820354e-01
-3.44878435e-02 1.13063562e+00 -4.22167063e-01 -2.82076389e-01
3.29987526e-01 -3.35872114e-01 8.04393589e-01 1.42968798e+00
4.55251276e-01 -3.70038956e-01 3.41609150e-01 1.02547693e+00
-2.33317584e-01 2.84432322e-01 -4.62368429e-01 -2.72429407e-01
3.73417944e-01 9.62874055e-01 -8.49644840e-01 -7.71559119e-01
7.92274922e-02 7.47649431e-01 -5.10849506e-02 6.27689064e-01
-3.63203794e-01 -4.63739187e-01 4.00240511e-01 -1.20254129e-01
5.01680255e-01 -1.65318236e-01 -3.75265718e-01 -1.14352405e+00
-5.28950691e-01 -2.76675344e-01 6.11133993e-01 -5.61615765e-01
-1.49994361e+00 6.06671929e-01 5.43437660e-01 -9.16123688e-01
-5.48549891e-01 -1.24383062e-01 -9.95416820e-01 8.22456896e-01
-1.32888639e+00 -6.61972582e-01 -1.19261228e-01 2.30449945e-01
9.95546579e-01 -4.41480994e-01 7.17261314e-01 -7.56200254e-02
-2.09235221e-01 1.52404502e-01 6.75801694e-01 -2.60350198e-01
7.59942889e-01 -1.59587610e+00 1.67033240e-01 -1.82450339e-02
-1.76133245e-01 8.24006438e-01 9.97177184e-01 -8.22156668e-01
-7.09862471e-01 -1.27473128e+00 9.86083269e-01 -1.44398108e-01
5.67107856e-01 -6.05990767e-01 -1.09253776e+00 7.68320501e-01
4.62093830e-01 -8.54301155e-01 9.43861842e-01 6.26500607e-01
-1.64440766e-01 4.37753618e-01 -1.07061934e+00 2.21791819e-01
-5.40508889e-02 -1.45597741e-01 -7.73006737e-01 7.13057697e-01
9.57539797e-01 8.04391652e-02 -6.97398782e-01 -1.30919695e-01
2.69386470e-01 -3.17843825e-01 5.07022500e-01 -5.11456609e-01
6.69119537e-01 3.56442273e-01 5.21579571e-02 -1.82522464e+00
-2.71800518e-01 -3.33186865e-01 -4.51283455e-01 1.39213550e+00
3.69754136e-01 -5.88317394e-01 1.09631050e+00 3.06986898e-01
1.29282534e-01 -5.68325162e-01 -6.85559452e-01 -7.42485881e-01
5.67463338e-01 -2.13521183e-01 2.26063818e-01 1.01451385e+00
4.62130159e-01 3.37468892e-01 -3.77151012e-01 -2.01710016e-01
8.20758164e-01 2.07997963e-01 4.75765705e-01 -1.62078965e+00
1.28471717e-01 -4.55891013e-01 2.84790876e-03 -1.35616314e+00
4.16769773e-01 -6.15021825e-01 3.90985131e-01 -1.65088844e+00
6.65550947e-01 -4.12422359e-01 -2.10755602e-01 3.36934060e-01
4.94536236e-02 -1.89174339e-01 -2.83020705e-01 5.25820136e-01
-7.91289806e-01 1.23720491e+00 6.94376767e-01 -3.74701440e-01
-5.43199658e-01 1.35612682e-01 -5.96204817e-01 8.11894417e-01
6.23707414e-01 -9.30755258e-01 -3.78597260e-01 -2.80357480e-01
6.43773004e-02 -3.77718687e-01 -2.12453559e-01 -5.24050951e-01
3.79808754e-01 1.15936555e-01 1.26831517e-01 -1.33612657e+00
3.50963086e-01 -4.59693283e-01 -5.03075838e-01 2.81652421e-01
-9.90510046e-01 -7.96316206e-01 -1.23172127e-01 6.97060108e-01
-4.36552376e-01 -2.60828525e-01 7.18924165e-01 1.51376009e-01
8.72058645e-02 2.86187649e-01 -9.95205224e-01 -2.13883802e-01
7.29486942e-01 1.60139933e-01 -2.22151503e-01 -7.83333898e-01
-8.13387215e-01 3.26013386e-01 2.02402130e-01 4.88653660e-01
4.84732509e-01 -1.21414697e+00 -8.45653236e-01 6.13544844e-02
-2.09532361e-02 2.83967763e-01 3.08051258e-01 2.35552281e-01
2.19802707e-01 9.20413375e-01 2.28415087e-01 -9.07428622e-01
-9.46617961e-01 2.88805753e-01 -3.62229466e-01 -2.36473352e-01
-9.74530280e-01 8.39816689e-01 5.80383003e-01 -3.94389570e-01
5.32427788e-01 -4.51267362e-01 -6.20219529e-01 3.96915555e-01
4.23245877e-01 4.49388355e-01 -4.15799648e-01 -2.15860918e-01
-1.68731716e-03 1.95641398e-01 -8.40963185e-01 -1.44562140e-01
1.60280895e+00 -1.50959194e-01 -1.07912160e-01 7.44284451e-01
1.14864922e+00 -4.14528817e-01 -1.25818336e+00 -3.74844939e-01
1.32798806e-01 -2.72332251e-01 4.77799922e-01 -3.50864798e-01
-6.68677449e-01 9.35472727e-01 2.64426887e-01 9.64748085e-01
5.50375879e-01 5.38133144e-01 6.46526456e-01 4.66939121e-01
-3.11213117e-02 -1.17562175e+00 3.36445540e-01 8.40865850e-01
6.60130978e-01 -8.55565369e-01 1.03833787e-01 -1.20729506e-01
-9.36710417e-01 1.00311589e+00 3.46889764e-01 -2.37293646e-01
1.02223468e+00 2.15441734e-01 -4.28339094e-01 -5.04332483e-01
-1.13199866e+00 2.80792505e-01 3.24458957e-01 6.04280055e-01
5.63415885e-01 -1.16901390e-01 -3.19901735e-01 8.06761861e-01
-3.12115252e-01 -3.53138328e-01 4.43320841e-01 6.50866628e-01
-8.10958564e-01 -5.76313019e-01 -4.03719783e-01 5.64640522e-01
-4.56223190e-01 -2.90688742e-02 -2.06951976e-01 2.21524090e-01
-4.66988415e-01 8.10087562e-01 8.18506002e-01 2.29854494e-01
-5.77705383e-01 4.10771519e-01 -4.17548746e-01 -9.97556686e-01
1.43734798e-01 6.78268492e-01 -3.73369455e-01 8.22298229e-02
5.32696806e-02 -7.37983763e-01 -1.14777982e+00 2.21829973e-02
-7.23415554e-01 7.30805039e-01 1.14389527e+00 7.93237686e-01
3.61562699e-01 6.87106073e-01 6.88430429e-01 -8.17306876e-01
-4.33400631e-01 -1.55190289e+00 -1.06133461e+00 1.64169028e-01
1.00303844e-01 -6.79591179e-01 -8.10927272e-01 3.06040764e-01] | [10.392657279968262, 6.947778701782227] |
47a27570-4467-4690-8a3d-7daecfa55c22 | bert-lid-leveraging-bert-to-improve-spoken | 2203.00328 | null | https://arxiv.org/abs/2203.00328v3 | https://arxiv.org/pdf/2203.00328v3.pdf | BERT-LID: Leveraging BERT to Improve Spoken Language Identification | Language identification is the task of automatically determining the identity of a language conveyed by a spoken segment. It has a profound impact on the multilingual interoperability of an intelligent speech system. Despite language identification attaining high accuracy on medium or long utterances(>3s), the performance on short utterances (<=1s) is still far from satisfactory. We propose a BERT-based language identification system (BERT-LID) to improve language identification performance, especially on short-duration speech segments. We extend the original BERT model by taking the phonetic posteriorgrams (PPG) derived from the front-end phone recognizer as input. Then we deployed the optimal deep classifier followed by it for language identification. Our BERT-LID model can improve the baseline accuracy by about 6.5% on long-segment identification and 19.9% on short-segment identification, demonstrating our BERT-LID's effectiveness to language identification. | ['Jinfeng Bai', 'Wei-Qiang Zhang', 'Junhong Zhao', 'Yuting Nie'] | 2022-03-01 | null | null | null | null | ['spoken-language-identification'] | ['speech'] | [-1.77562431e-01 -1.52068079e-01 -2.77959973e-01 -5.46864748e-01
-1.21738136e+00 -1.01075327e+00 5.67835927e-01 -2.10698709e-01
-6.96091175e-01 4.04311478e-01 1.28183410e-01 -7.96568990e-01
3.95210028e-01 -8.21748897e-02 -5.04363418e-01 -4.15509254e-01
1.63988158e-01 6.76207602e-01 3.47314551e-02 -1.66578859e-01
-3.95693891e-02 6.02483332e-01 -1.17952156e+00 2.45123073e-01
9.03301656e-01 9.08253431e-01 2.98924595e-01 8.46814334e-01
-4.90416825e-01 4.74728614e-01 -7.64909327e-01 -1.75285786e-01
-7.98197016e-02 -1.91492960e-01 -1.07063472e+00 -2.24724427e-01
9.29417834e-02 -3.76083225e-01 -1.94366634e-01 1.03353477e+00
6.50197387e-01 -3.12432963e-02 5.61596394e-01 -1.02458501e+00
-4.97663707e-01 1.16732419e+00 -1.59806520e-01 2.44435698e-01
5.42768180e-01 -5.93143562e-03 8.17987859e-01 -1.07552505e+00
1.16692357e-01 1.64982557e+00 6.94669306e-01 7.42315710e-01
-1.10193312e+00 -7.96689868e-01 1.38338536e-01 3.69000174e-02
-1.81038666e+00 -1.07977724e+00 3.91089499e-01 -6.42138362e-01
1.12098861e+00 3.22973609e-01 1.09939553e-01 9.91176188e-01
-2.25192219e-01 1.09611440e+00 8.30510080e-01 -5.91083109e-01
-7.42183551e-02 3.24775666e-01 4.46975172e-01 3.34458441e-01
-4.74443465e-01 -1.40107304e-01 -3.86105627e-01 1.57464091e-02
3.11390042e-01 -6.17686749e-01 -1.18621744e-01 5.75928807e-01
-1.21482909e+00 6.94485426e-01 -2.76471358e-02 4.95674282e-01
-2.25303277e-01 -2.53556669e-01 6.09281003e-01 3.81457239e-01
4.94672477e-01 4.48851734e-01 -4.38783288e-01 -5.66052973e-01
-8.97686720e-01 -2.28757977e-01 8.11118424e-01 9.13863301e-01
4.43092316e-01 2.47940525e-01 -9.82984751e-02 1.47940123e+00
1.60839200e-01 7.35015392e-01 7.39991844e-01 -4.94571567e-01
3.03949386e-01 3.63850117e-01 6.37252256e-02 -2.43075356e-01
-4.16568667e-01 -3.67809713e-01 -4.67125088e-01 -4.23016399e-01
5.64086497e-01 -2.12211609e-01 -7.24716902e-01 1.76707244e+00
-6.06791936e-02 -1.67659342e-01 2.34753057e-01 6.79921806e-01
7.86698580e-01 9.13017511e-01 8.34192485e-02 -8.57307017e-03
1.41555834e+00 -9.88651037e-01 -7.65556395e-01 -2.14125782e-01
7.05169499e-01 -1.08081388e+00 1.40769148e+00 6.82480857e-02
-7.62797594e-01 -8.15702379e-01 -7.07141876e-01 -7.04158377e-03
-3.74037772e-01 4.73429054e-01 1.11927994e-01 8.36839080e-01
-1.13888967e+00 -1.64120406e-01 -4.78321731e-01 -3.96828175e-01
-1.10251419e-01 6.58785880e-01 -4.53884542e-01 1.38173908e-01
-1.39606822e+00 5.84159672e-01 3.72858584e-01 5.05406894e-02
-7.45163202e-01 -5.11633337e-01 -7.94175327e-01 -1.35469168e-01
5.79946600e-02 -1.32277772e-01 1.66435051e+00 -8.03068221e-01
-1.86302102e+00 1.18026233e+00 -5.94590604e-01 -5.08426011e-01
4.60824996e-01 -1.50677279e-01 -8.96739781e-01 2.95468550e-02
2.33554438e-01 7.55147994e-01 6.32135570e-01 -9.75956678e-01
-7.21457124e-01 -2.30229586e-01 -3.16756725e-01 2.87399709e-01
-4.27143335e-01 6.28558517e-01 -5.56747437e-01 -4.47286576e-01
4.66367044e-02 -1.09515929e+00 2.98481882e-01 -7.10138500e-01
-7.06277072e-01 -5.82207263e-01 7.55481660e-01 -1.14456201e+00
1.38193727e+00 -2.38709617e+00 -3.46882850e-01 -6.03359416e-02
-2.09249556e-01 7.52877891e-01 -1.82865024e-01 3.76233369e-01
2.12940231e-01 1.54905245e-01 -1.10011734e-02 -8.17604721e-01
1.20491497e-01 6.71832934e-02 -4.41557050e-01 2.88518339e-01
-5.92367537e-02 8.77101302e-01 -7.57325828e-01 -2.00854704e-01
2.22358406e-01 3.88522446e-01 -1.40256330e-01 4.38146055e-01
6.05050549e-02 6.16273999e-01 4.02590372e-02 5.89859724e-01
6.00346744e-01 2.14022085e-01 -2.97368504e-02 8.80272314e-02
-4.07476068e-01 7.13469386e-01 -7.70161808e-01 1.25383949e+00
-7.24404216e-01 7.95620024e-01 3.69274557e-01 -5.47190428e-01
1.22662520e+00 5.24160683e-01 2.15827271e-01 -5.08016288e-01
7.46470839e-02 6.08235896e-01 3.61369103e-01 -3.14673215e-01
5.42417228e-01 9.99037698e-02 -4.06476557e-01 4.94073004e-01
6.99879900e-02 7.47765154e-02 -1.57305449e-01 1.63106509e-02
5.55524111e-01 -6.14726603e-01 1.33649886e-01 -4.54432547e-01
1.10749638e+00 -2.53490359e-01 3.33930105e-01 8.42782199e-01
-3.89963448e-01 5.07776320e-01 1.24450680e-02 -6.97252378e-02
-8.29932213e-01 -1.08428788e+00 -3.48328769e-01 1.53177953e+00
-8.30048025e-02 -1.80198669e-01 -1.11479318e+00 -5.08441925e-01
6.51601627e-02 8.26529324e-01 1.02095857e-01 -1.33479610e-01
-5.01775980e-01 -3.93121690e-01 1.41921055e+00 4.27171707e-01
4.49328393e-01 -1.06567657e+00 4.44753945e-01 3.94867390e-01
-4.34368223e-01 -1.52148342e+00 -1.07170939e+00 3.61741632e-02
-2.93512305e-04 -4.61125284e-01 -8.00979197e-01 -1.28598285e+00
2.72171915e-01 4.04149257e-02 7.74885535e-01 -2.77671516e-01
3.27792823e-01 2.10511327e-01 -1.78075343e-01 -3.45478356e-01
-1.07556069e+00 5.83990157e-01 6.63047969e-01 3.59876305e-01
7.35882759e-01 5.12709208e-02 -2.30928153e-01 4.28430438e-01
-2.99376994e-01 -8.89060646e-02 1.63770765e-01 7.95287907e-01
2.35644445e-01 -2.23372906e-01 1.00484002e+00 -3.74667019e-01
7.17749834e-01 -2.55428523e-01 -5.39829433e-01 2.66973704e-01
-3.32154274e-01 -1.56211019e-01 9.00862336e-01 -6.88567162e-01
-7.94942021e-01 1.15395322e-01 -9.02885437e-01 -1.35487288e-01
-3.34702343e-01 2.45780721e-01 -4.88473862e-01 -1.13219731e-02
1.94029227e-01 5.07073343e-01 2.49688774e-01 -8.96883070e-01
2.15066105e-01 1.79595780e+00 1.01469862e+00 -3.77246529e-01
2.45201573e-01 6.10812334e-03 -9.91879106e-01 -1.25226760e+00
-5.05931556e-01 -8.42923582e-01 -7.56372094e-01 -7.28546306e-02
7.30243623e-01 -1.15924394e+00 -1.01727653e+00 1.09237754e+00
-1.18215370e+00 -2.87235230e-01 1.65987238e-01 6.21831179e-01
-2.01654956e-01 3.20070297e-01 -8.55520785e-01 -1.28353572e+00
-6.48758590e-01 -1.49120021e+00 1.16627562e+00 -7.87117854e-02
-5.32231688e-01 -9.81224060e-01 -2.27742940e-01 3.68902683e-01
4.85060990e-01 -7.31662691e-01 6.97182000e-01 -1.12201691e+00
1.79041252e-02 -2.46067226e-01 -1.68472394e-01 6.64572179e-01
2.00924188e-01 -1.84245050e-01 -1.37864828e+00 -3.43528748e-01
-2.43305996e-01 -9.47161764e-02 5.63325942e-01 3.28786641e-01
8.38321924e-01 -2.79640555e-01 -2.07653090e-01 6.52430058e-01
7.69950390e-01 7.07438827e-01 1.32284820e-01 1.58358529e-01
8.87496352e-01 6.75100148e-01 3.66395086e-01 1.01770505e-01
6.41202867e-01 9.64252532e-01 -1.90429673e-01 -7.81848066e-05
-3.13915610e-01 -4.18005675e-01 8.29397738e-01 1.43901742e+00
7.83707976e-01 -2.38300204e-01 -1.36608243e+00 7.34823525e-01
-1.38121784e+00 -6.05710685e-01 -1.32853566e-02 2.23484206e+00
1.02810729e+00 3.56851667e-02 4.77719575e-01 2.21195430e-01
1.13152027e+00 -8.46325904e-02 -5.86487651e-01 -8.33793640e-01
-2.99173206e-01 -2.89017081e-01 4.43781465e-01 1.06468236e+00
-1.14641631e+00 1.45892072e+00 6.70502377e+00 1.05165911e+00
-1.50905621e+00 5.49693368e-02 6.58954203e-01 3.49668354e-01
-1.35582462e-01 -3.52174550e-01 -1.57993615e+00 7.96840191e-01
1.41435075e+00 -4.29558039e-01 5.69385946e-01 7.55272210e-01
3.20106655e-01 1.54120266e-01 -1.16303575e+00 1.30255663e+00
3.09908688e-02 -8.05315971e-01 4.52452265e-02 5.81561960e-02
4.37028140e-01 3.09769988e-01 1.33611217e-01 3.96063149e-01
2.18694672e-01 -1.09948456e+00 9.32601631e-01 1.08309232e-01
1.25068891e+00 -8.76133859e-01 6.47018611e-01 6.94156051e-01
-1.22688079e+00 -1.93974562e-02 -5.23252301e-02 5.65765053e-02
2.71365792e-01 2.52903521e-01 -1.52688301e+00 5.69510348e-02
2.13612780e-01 4.68162507e-01 -4.35274571e-01 8.23019624e-01
6.04411997e-02 1.12445223e+00 -4.64031994e-01 -1.13543250e-01
4.04566020e-01 -6.65962175e-02 5.43777108e-01 1.68290746e+00
2.41099089e-01 -4.10601437e-01 3.30703646e-01 4.35155332e-01
-9.89630297e-02 2.46932924e-01 -5.70589662e-01 -3.63485307e-01
1.05586839e+00 8.51303577e-01 -3.43451858e-01 -4.35515314e-01
-2.43175179e-01 1.06865656e+00 1.99191377e-01 2.72155792e-01
-4.44451123e-01 -4.62211519e-01 9.27441835e-01 -2.90192634e-01
-3.26014496e-02 -2.24484116e-01 -2.13380665e-01 -7.89102912e-01
-1.75623015e-01 -1.07629228e+00 2.05657780e-01 -3.03941816e-01
-1.22216249e+00 9.91780519e-01 -4.20920521e-01 -1.05776250e+00
-9.01980639e-01 -5.77826560e-01 -2.50695437e-01 1.30894554e+00
-1.22714865e+00 -1.24605596e+00 2.31250018e-01 3.16949159e-01
7.75089741e-01 -6.35266542e-01 1.13177025e+00 5.01246989e-01
-8.04478109e-01 1.22869503e+00 4.28271681e-01 5.98503053e-01
6.70476735e-01 -1.20449209e+00 9.35824394e-01 8.24403644e-01
2.31603548e-01 5.95621645e-01 3.32307965e-01 -2.19392166e-01
-1.35438955e+00 -9.78830040e-01 1.36721444e+00 -3.99719447e-01
6.89808965e-01 -7.02062309e-01 -9.51385200e-01 6.52544916e-01
6.97423667e-02 -3.84969413e-01 6.93242490e-01 1.56295404e-01
-4.03450489e-01 -1.49676517e-01 -9.43764627e-01 5.80447733e-01
8.43760550e-01 -1.37417305e+00 -4.54660833e-01 1.69964418e-01
8.60977113e-01 -1.31546393e-01 -8.14475536e-01 1.65872052e-01
7.64307439e-01 -4.58806127e-01 9.47292387e-01 -2.47289538e-01
-3.76905233e-01 -2.17465788e-01 -2.29554802e-01 -1.32680500e+00
-1.68806583e-01 -7.60287642e-01 2.37696990e-01 1.65254176e+00
4.66573268e-01 -9.49111342e-01 2.41594389e-01 1.90685377e-01
-2.89341450e-01 -2.26605862e-01 -1.01328540e+00 -1.15320742e+00
2.46689007e-01 -6.48602486e-01 9.52835500e-01 8.36911619e-01
4.31172410e-03 5.70456624e-01 -2.34012276e-01 3.92497212e-01
2.16945410e-01 -2.01727808e-01 7.10154653e-01 -8.90512347e-01
-1.10060014e-01 -7.05540299e-01 -2.45621860e-01 -1.49492824e+00
7.50350654e-01 -1.11260951e+00 2.97468185e-01 -9.78373170e-01
-1.38175115e-01 -5.31480551e-01 -2.21189320e-01 3.60518098e-01
-1.89089835e-01 2.21900001e-01 1.60868213e-01 3.14609528e-01
-4.71993238e-01 3.70711923e-01 7.14837015e-01 -3.28656405e-01
-2.81557262e-01 1.90212488e-01 -4.37705904e-01 5.95530033e-01
8.16867888e-01 -7.26000220e-02 -2.23512024e-01 -5.67403495e-01
-6.18428886e-01 3.87299582e-02 -2.83698291e-01 -9.96817052e-01
3.25009882e-01 4.45570350e-01 -1.85687527e-01 -7.21821904e-01
4.39848512e-01 -3.47177982e-01 3.04271169e-02 4.31436241e-01
-5.31383097e-01 2.55230591e-02 2.78787851e-01 -2.15059042e-01
-5.70272624e-01 -1.81538433e-01 7.86672473e-01 1.53260067e-01
-1.01194918e+00 1.54351190e-01 -7.72865951e-01 -1.31063676e-02
6.02118492e-01 -1.79391488e-01 -4.08845432e-02 -4.74161595e-01
-7.33841538e-01 2.67816126e-01 3.55708301e-01 7.71549821e-01
2.36233130e-01 -1.28183305e+00 -8.09320807e-01 7.48372734e-01
3.24018955e-01 -3.78850818e-01 -7.47883394e-02 6.42255902e-01
-3.33134025e-01 9.00525510e-01 2.78712720e-01 -8.58606696e-01
-1.50507534e+00 2.15292662e-01 6.04184806e-01 7.89640397e-02
-2.90478170e-01 1.19699442e+00 1.95212975e-01 -8.85638773e-01
6.00671709e-01 -6.63937181e-02 -3.19156766e-01 5.41441850e-02
8.26768577e-01 2.12764516e-01 1.19762018e-01 -1.36759531e+00
-7.31480062e-01 4.77368087e-01 -3.47627193e-01 -4.35613036e-01
6.00839734e-01 -5.38622618e-01 -7.87053481e-02 9.51375425e-01
1.64932477e+00 2.05739394e-01 -9.14880633e-01 -3.29959124e-01
2.84473926e-01 -8.41112658e-02 1.88094042e-02 -7.00567842e-01
-5.17987549e-01 9.28994715e-01 5.20483315e-01 3.40059966e-01
6.82197869e-01 9.12970752e-02 1.12144232e+00 2.59588093e-01
3.67713004e-01 -1.12849164e+00 -4.12926167e-01 1.12519991e+00
7.97280908e-01 -1.44007432e+00 -8.76888573e-01 -2.95868188e-01
-5.92906892e-01 9.19164598e-01 3.31178367e-01 4.42113280e-01
5.92497885e-01 3.81604910e-01 6.49242461e-01 3.49910498e-01
-3.43634099e-01 -2.50148058e-01 4.72738475e-01 4.96916741e-01
6.19411647e-01 5.86885631e-01 -7.83398226e-02 5.99204957e-01
-6.47594810e-01 -5.52994490e-01 4.74477410e-02 2.79067785e-01
-3.08773190e-01 -1.19488597e+00 -6.15776777e-01 1.81504130e-01
-5.39501250e-01 -2.16747448e-01 -4.67132598e-01 2.36466393e-01
-1.56204969e-01 1.35681403e+00 3.74124408e-01 -6.61923528e-01
1.49478480e-01 4.69267190e-01 -1.01672597e-01 -6.57224476e-01
-6.51926219e-01 1.91013545e-01 3.72887135e-01 -1.41406849e-01
1.15096398e-01 -7.57132530e-01 -1.29219472e+00 -4.47274119e-01
-1.93420410e-01 3.94580364e-01 8.64285111e-01 9.42251027e-01
3.23127955e-01 1.98089272e-01 1.01930428e+00 -5.54832101e-01
-8.55758905e-01 -1.12194276e+00 -7.25610316e-01 2.35938296e-01
6.74527407e-01 -2.13679373e-01 -3.57198417e-01 -1.05074242e-01] | [14.180026054382324, 6.717308521270752] |
37c97c17-b0ed-439c-9c2f-4cc80e6be323 | end-to-end-clinical-event-extraction-from | 2208.09354 | null | https://arxiv.org/abs/2208.09354v1 | https://arxiv.org/pdf/2208.09354v1.pdf | End-to-end Clinical Event Extraction from Chinese Electronic Health Record | Event extraction is an important work of medical text processing. According to the complex characteristics of medical text annotation, we use the end-to-end event extraction model to enhance the output formatting information of events. Through pre training and fine-tuning, we can extract the attributes of the four dimensions of medical text: anatomical position, subject word, description word and occurrence state. On the test set, the accuracy rate was 0.4511, the recall rate was 0.3928, and the F1 value was 0.42. The method of this model is simple, and it has won the second place in the task of mining clinical discovery events (task2) in the Chinese electronic medical record of the seventh China health information processing Conference (chip2021). | ['Yun Liu', 'Huiting Sun', 'Yun Yu', 'Ruochen Huang', 'Wei Feng'] | 2022-08-19 | null | null | null | null | ['event-extraction', 'text-annotation'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.58877397e-01 2.38234118e-01 -3.05950373e-01 -3.22762400e-01
-7.42025018e-01 -2.06135556e-01 2.21877232e-01 6.99473977e-01
-8.30351532e-01 9.58366036e-01 5.33507526e-01 -4.09972072e-01
-9.58380699e-02 -8.48987162e-01 -1.35510638e-01 -4.17732537e-01
-2.36908108e-01 2.91943580e-01 2.19171032e-01 1.28237680e-01
9.99434292e-02 2.73267031e-02 -1.19299293e+00 8.24762940e-01
8.12467933e-01 9.96380627e-01 7.89652467e-02 6.06779337e-01
-3.45482379e-01 8.16986382e-01 -7.32026637e-01 -3.57499152e-01
-4.22813743e-01 -4.93270278e-01 -9.17540729e-01 -3.46043825e-01
-7.67721891e-01 4.84936498e-02 -1.86377719e-01 1.00262249e+00
9.32419717e-01 -1.92966640e-01 5.56861758e-01 -7.64434576e-01
-1.11913353e-01 1.02857888e+00 -4.55156446e-01 3.90517116e-01
2.85527319e-01 -3.08000028e-01 7.71779656e-01 -8.66168201e-01
1.17367101e+00 7.58936286e-01 6.27711713e-01 4.75871861e-01
-4.47325349e-01 -6.68186367e-01 -2.23813787e-01 3.45646329e-02
-1.39756989e+00 -2.06149206e-01 1.94977865e-01 -4.52042758e-01
1.08567727e+00 4.28370088e-01 6.03383362e-01 9.01728749e-01
7.87751555e-01 8.86690736e-01 5.67800462e-01 -5.13988554e-01
1.38352901e-01 2.88171768e-02 2.39861578e-01 8.09518456e-01
2.52591610e-01 -1.65560156e-01 -3.39354545e-01 -5.63720107e-01
2.46967465e-01 2.00158834e-01 -4.58482802e-02 8.72217357e-01
-1.52360845e+00 7.62675881e-01 -9.94570553e-02 5.16496301e-01
-5.44361651e-01 -4.18245584e-01 8.72591615e-01 5.34327924e-02
4.07567710e-01 2.80340135e-01 -9.54373717e-01 -1.89546809e-01
-7.04330385e-01 4.80382256e-02 7.42565572e-01 1.03486180e+00
-1.38418838e-01 -6.02729857e-01 -7.09286273e-01 6.18188381e-01
2.00069651e-01 1.48258418e-01 9.00513589e-01 -1.79626256e-01
4.19860929e-01 7.70071507e-01 -4.56641130e-02 -6.28621995e-01
-1.04790127e+00 -3.11649740e-01 -1.07162714e+00 -7.38993347e-01
4.69708703e-02 -7.49958813e-01 -1.09446537e+00 1.42743969e+00
3.87301266e-01 -2.37543240e-01 6.60353601e-02 3.86563212e-01
1.31153238e+00 7.58040488e-01 7.37910092e-01 -7.57337213e-01
2.06852770e+00 -4.54190880e-01 -1.51165116e+00 6.24451078e-02
9.27020609e-01 -9.55733776e-01 4.64276701e-01 4.57843035e-01
-9.39327776e-01 -2.40345880e-01 -9.19262707e-01 7.60120749e-02
-5.05811453e-01 4.67294693e-01 8.03411961e-01 3.55340779e-01
-2.04421848e-01 3.23477000e-01 -7.93397069e-01 -3.80408376e-01
5.64880073e-01 3.09591204e-01 -1.69031352e-01 2.02967316e-01
-1.69337118e+00 7.79144943e-01 9.49746490e-01 -2.20705390e-01
-3.43037903e-01 -8.34241033e-01 -5.49730778e-01 -4.78787497e-02
3.92670751e-01 -7.46502459e-01 1.10307038e+00 -2.36219317e-01
-1.07075489e+00 1.01560342e+00 -7.46518672e-02 -2.36488342e-01
1.41829595e-01 -1.12310886e-01 -1.09817111e+00 7.08471462e-02
1.31922722e-01 3.21446687e-01 7.10633844e-02 -2.62091249e-01
-1.18442714e+00 -5.97224355e-01 -7.04356670e-01 -8.45000520e-02
-3.82211685e-01 4.54321682e-01 -6.71816528e-01 -9.08048689e-01
-9.02703255e-02 -6.83142781e-01 -6.43195689e-01 -4.73749161e-01
-6.16696119e-01 -4.53995615e-01 2.52054095e-01 -9.15164053e-01
2.00950956e+00 -2.60235786e+00 -4.74050820e-01 3.26294810e-01
2.62368917e-01 1.30460143e-01 3.53720695e-01 3.77488971e-01
-3.26812118e-01 3.19001377e-01 -6.62005926e-03 3.47709149e-01
-3.86427760e-01 -1.10198975e-01 -4.45613434e-04 2.41051987e-01
2.99703568e-01 9.39416587e-01 -9.38854635e-01 -1.05898535e+00
-1.90647438e-01 2.66137809e-01 -4.73219752e-01 2.37963781e-01
4.82500792e-02 8.27648044e-02 -8.84413123e-01 4.80835289e-01
2.53714532e-01 -4.01004523e-01 2.21003801e-01 -3.67659956e-01
-1.00396611e-01 5.33213794e-01 -1.22991598e+00 1.81613958e+00
-2.98319142e-02 1.76580340e-01 -4.94559795e-01 -5.10046244e-01
7.73346186e-01 8.25898945e-01 1.07747960e+00 -5.37315428e-01
4.83161211e-01 2.43244562e-02 1.14799522e-01 -1.15091550e+00
3.07742715e-01 -1.46616966e-01 -4.65991169e-01 1.72924042e-01
9.70548242e-02 3.58445853e-01 4.26923662e-01 2.60893434e-01
1.31019258e+00 -3.86918783e-01 9.74292040e-01 -2.31395647e-01
3.65720183e-01 1.46330923e-01 8.11388969e-01 5.09981573e-01
-2.76735332e-03 5.44910312e-01 5.71440458e-01 -4.68959838e-01
-6.88670576e-01 -7.45855451e-01 -6.16879582e-01 6.75594807e-01
-3.64768982e-01 -1.01286638e+00 -5.41461647e-01 -7.92703688e-01
-3.22694391e-01 6.01941526e-01 -7.28908479e-01 -3.37905347e-01
-4.80770916e-01 -1.47060704e+00 8.41774046e-01 4.13889408e-01
2.66901672e-01 -1.39511240e+00 -5.96091568e-01 7.02876687e-01
-5.41522384e-01 -1.21748292e+00 -5.77255368e-01 4.30912763e-01
-7.11329758e-01 -1.07632852e+00 -4.33408707e-01 -8.54366243e-01
4.29111063e-01 -7.64283121e-01 1.02246106e+00 -3.56462635e-02
-7.79119909e-01 -3.89359385e-01 -5.09607911e-01 -1.10260534e+00
-2.02718109e-01 3.05720568e-01 -3.30298543e-01 -1.94185600e-01
6.57095850e-01 -6.43475819e-03 -5.43676019e-01 -1.33687019e-01
-1.08647943e+00 1.45954520e-01 8.02766919e-01 8.84622574e-01
6.85203850e-01 2.13840261e-01 8.07234347e-01 -1.30168498e+00
7.78613508e-01 -5.07465899e-01 -2.16828316e-01 2.00146079e-01
-8.12873602e-01 -9.86958593e-02 3.70677114e-01 -4.45643961e-01
-1.06247044e+00 5.02004147e-01 -6.74880564e-01 5.70858181e-01
-1.24859393e-01 1.00209439e+00 -2.03010455e-01 8.78523529e-01
6.01210535e-01 9.93730128e-02 -3.06666315e-01 -3.30119759e-01
3.85785848e-02 9.66019750e-01 5.45952737e-01 -3.80771533e-02
-8.66028816e-02 1.55706242e-01 -1.03407137e-01 -6.33480608e-01
-1.00576019e+00 -5.19077837e-01 -3.32063705e-01 2.97294587e-01
1.18532622e+00 -8.39034915e-01 -9.22516644e-01 1.94681644e-01
-1.15632081e+00 2.86363006e-01 -2.51920164e-01 8.43315542e-01
-6.92810044e-02 1.14873365e-01 -7.60304987e-01 -5.66712379e-01
-8.41878891e-01 -7.82192767e-01 8.17105114e-01 2.21110493e-01
-5.38984120e-01 -6.27615154e-01 -3.59594561e-02 -1.54817998e-01
-1.11779764e-01 4.56402689e-01 1.27981925e+00 -1.09198260e+00
3.90305012e-01 -5.22367954e-01 1.43178254e-02 -2.38272697e-01
2.21669987e-01 9.10105631e-02 -5.86994171e-01 1.89070761e-01
9.37631726e-02 1.50144458e-01 8.95146072e-01 6.39178991e-01
1.61253536e+00 -2.31368914e-01 -7.45465398e-01 3.49768788e-01
1.20232069e+00 7.07194030e-01 9.00019526e-01 -2.89421622e-02
3.66434753e-01 3.97732377e-01 6.79101765e-01 7.21580446e-01
3.81226718e-01 3.36866289e-01 -1.78910226e-01 -2.34546587e-01
8.76380429e-02 -2.28158817e-01 5.67849651e-02 8.21684241e-01
-2.57293340e-02 -3.64056230e-01 -1.00724077e+00 5.15707493e-01
-1.66130233e+00 -8.79838109e-01 -6.46661639e-01 1.82065368e+00
1.42320073e+00 4.35901433e-01 7.32731894e-02 2.31535017e-01
5.82086027e-01 -4.24888909e-01 -1.91054583e-01 -2.26381168e-01
2.43321732e-02 3.41512024e-01 5.68123937e-01 3.17537785e-02
-1.23311627e+00 6.68294668e-01 6.44741058e+00 9.40775752e-01
-7.28281081e-01 5.82847223e-02 7.58736074e-01 4.01220843e-02
1.76678941e-01 -5.46409965e-01 -1.20148599e+00 7.88122594e-01
1.34878194e+00 -3.33251894e-01 -3.04098666e-01 4.59178060e-01
2.22084090e-01 -1.73576921e-01 -9.35465991e-01 9.28218603e-01
-1.17614456e-01 -1.41999578e+00 4.80249114e-02 -1.50397047e-02
3.94164711e-01 -2.33955815e-01 -2.90392488e-01 4.18373942e-01
1.23776793e-01 -1.04512739e+00 2.99604088e-01 5.33041596e-01
1.00284600e+00 -7.81199455e-01 1.19799376e+00 4.48549211e-01
-1.06463408e+00 -1.24752797e-01 -5.34141213e-02 2.34853715e-01
3.41676444e-01 1.07564330e+00 -1.10631406e+00 7.52101362e-01
6.93118513e-01 5.47133625e-01 -3.43419224e-01 1.04976237e+00
-1.72502607e-01 7.47922599e-01 -2.81286597e-01 -2.73766816e-01
-1.46487698e-01 4.96281475e-01 2.11297005e-01 1.77562189e+00
3.37591261e-01 6.25429213e-01 1.53530926e-01 1.91030249e-01
-2.89744586e-01 5.02786160e-01 -9.37217548e-02 -1.41672686e-01
3.35063070e-01 1.27723360e+00 -9.48779941e-01 -6.55138791e-01
-9.25730541e-02 5.23562014e-01 -2.67630875e-01 -1.79753333e-01
-9.62911308e-01 -9.39588785e-01 3.30596380e-02 -2.12343425e-01
1.46616906e-01 3.87024522e-01 -5.86450517e-01 -9.41430926e-01
-1.42541453e-01 -8.43049526e-01 1.13778508e+00 -4.62166756e-01
-1.27348328e+00 7.80567527e-01 -1.59498513e-01 -1.04814589e+00
-1.81797937e-01 -5.28225720e-01 -2.20072120e-01 7.51320779e-01
-9.22484577e-01 -6.71674490e-01 9.55856144e-02 6.45379126e-01
4.31845844e-01 -1.87596172e-01 1.22560441e+00 9.01877701e-01
-7.24106371e-01 8.05252969e-01 -1.94052473e-01 5.95388293e-01
7.63655543e-01 -9.56390262e-01 1.76755086e-01 4.71986413e-01
-1.65845752e-01 5.51399291e-01 4.80667025e-01 -9.34448659e-01
-1.11023128e+00 -1.27743030e+00 1.74968314e+00 -3.10891449e-01
3.68464828e-01 -1.82492003e-01 -6.75400019e-01 3.31262618e-01
2.85292864e-02 -2.31783852e-01 1.24228561e+00 3.08358315e-02
1.99848935e-01 1.25004902e-01 -1.11859524e+00 3.56017798e-01
7.22345412e-01 -1.77312061e-01 -6.87587380e-01 4.84190822e-01
8.84011626e-01 -6.45950377e-01 -1.28393126e+00 6.10960305e-01
4.82866943e-01 8.96282718e-02 9.93444562e-01 -1.05163729e+00
5.82021654e-01 -2.30723217e-01 1.74967960e-01 -8.74505579e-01
-3.97227049e-01 -3.87285054e-01 -2.67629381e-02 1.26141763e+00
1.06195712e+00 -2.17738986e-01 3.54137450e-01 2.87667662e-01
-6.37209341e-02 -9.84095097e-01 -7.25705266e-01 -1.67427525e-01
-3.75839710e-01 -4.25329119e-01 5.58351994e-01 1.02964866e+00
4.88872856e-01 7.74113238e-01 -2.38059461e-01 -4.84312065e-02
1.01190530e-01 -8.18729028e-02 -1.61328744e-02 -1.33541918e+00
1.90453734e-02 -1.48851499e-01 -4.63403948e-02 -4.53544557e-01
-3.67364019e-01 -1.00300241e+00 1.25588346e-02 -1.67279983e+00
6.01828694e-01 -1.74766883e-01 -7.43679643e-01 6.95167720e-01
-6.77525342e-01 -7.30681196e-02 -3.72585237e-01 -5.70563301e-02
-5.65702796e-01 5.85470796e-02 1.18234539e+00 -1.15046836e-02
-3.60060394e-01 2.02395007e-01 -8.14356029e-01 6.45689964e-01
8.12349141e-01 -9.35174525e-01 -3.48468944e-02 -5.28782653e-03
5.66218555e-01 3.54591370e-01 -3.83720487e-01 -4.68212336e-01
2.94449329e-01 -1.93038151e-01 7.51990616e-01 -9.75642979e-01
-2.12032318e-01 -6.65080488e-01 1.70606911e-01 8.90958369e-01
-5.45714259e-01 1.74718916e-01 3.03347707e-01 3.66203070e-01
-1.30978584e-01 -1.35999233e-01 4.03334469e-01 -1.43643707e-01
-5.31125963e-01 2.82137036e-01 -7.62732446e-01 3.23245645e-01
9.51045156e-01 2.84246743e-01 -1.70586750e-01 3.84704381e-01
-1.00192916e+00 2.58347541e-01 -4.90376115e-01 4.35957074e-01
6.74353659e-01 -1.22195780e+00 -1.10575366e+00 1.23703904e-01
2.35171318e-01 -1.48505688e-01 9.33066234e-02 9.15205240e-01
-3.38637352e-01 2.96156824e-01 -8.06878358e-02 -2.55073577e-01
-1.33844209e+00 6.65132344e-01 -1.40109256e-01 -8.14690828e-01
-7.31429815e-01 6.78139865e-01 -3.13493758e-02 4.68786573e-03
2.54810601e-01 -3.53864193e-01 -7.27710187e-01 3.06752741e-01
1.00370502e+00 4.09206115e-02 4.91038144e-01 2.72687525e-03
-6.77538514e-01 -7.89972320e-02 -2.51344174e-01 -1.25035375e-01
1.59441042e+00 1.64713964e-01 -2.75265396e-01 3.66999775e-01
1.07616639e+00 -5.58460727e-02 -1.69348404e-01 -4.91709150e-02
4.43231702e-01 2.03116059e-01 7.83063769e-02 -1.17004061e+00
-8.08460593e-01 5.34103394e-01 5.93112230e-01 1.03716746e-01
1.23447573e+00 2.00617686e-02 8.38080168e-01 2.42029324e-01
-1.23476408e-01 -1.18951106e+00 -4.51826155e-01 7.14499414e-01
4.89676833e-01 -8.69984448e-01 1.79706007e-01 -5.54320514e-01
-6.80284917e-01 1.06893158e+00 2.75485814e-01 3.83481771e-01
1.02435124e+00 8.64019871e-01 -1.28883004e-01 -3.60045522e-01
-9.54902589e-01 6.38962612e-02 4.68318820e-01 8.00325200e-02
7.56778896e-01 2.20497429e-01 -1.04654145e+00 1.44089162e+00
-2.05586329e-01 3.78943086e-01 1.07755125e-01 1.03320944e+00
-2.48830229e-01 -9.49941039e-01 -1.51525512e-01 7.99719453e-01
-1.43380618e+00 -4.27508593e-01 -1.99918464e-01 5.23729026e-01
2.98622936e-01 9.31783855e-01 -1.08573504e-01 -4.17773843e-01
6.03064477e-01 3.03583711e-01 1.81724399e-01 -5.61067820e-01
-9.21001554e-01 2.66527981e-01 3.07985693e-01 -3.89223278e-01
-2.28082031e-01 -4.44586486e-01 -1.87127805e+00 3.78667824e-02
-3.57999712e-01 3.88889253e-01 4.86119598e-01 1.03886545e+00
5.54626405e-01 1.14183354e+00 3.49474281e-01 3.57774377e-01
-6.32401332e-02 -1.09953439e+00 -4.53912437e-01 3.68813276e-01
6.79790750e-02 -1.54791355e-01 2.83097386e-01 6.04218125e-01] | [8.439409255981445, 8.738224983215332] |
fc97be20-31b5-463c-b7e1-54563276de43 | weak-shot-object-detection-through-mutual | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Du_Weak-Shot_Object_Detection_Through_Mutual_Knowledge_Transfer_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Du_Weak-Shot_Object_Detection_Through_Mutual_Knowledge_Transfer_CVPR_2023_paper.pdf | Weak-Shot Object Detection Through Mutual Knowledge Transfer | Weak-shot Object Detection methods exploit a fully-annotated source dataset to facilitate the detection performance on the target dataset which only contains image-level labels for novel categories. To bridge the gap between these two datasets, we aim to transfer the object knowledge between the source (S) and target (T) datasets in a bi-directional manner. We propose a novel Knowledge Transfer (KT) loss which simultaneously distills the knowledge of objectness and class entropy from a proposal generator trained on the S dataset to optimize a multiple instance learning module on the T dataset. By jointly optimizing the classification loss and the proposed KT loss, the multiple instance learning module effectively learns to classify object proposals into novel categories in the T dataset with the transferred knowledge from base categories in the S dataset. Noticing the predicted boxes on the T dataset can be regarded as an extension for the original annotations on the S dataset to refine the proposal generator in return, we further propose a novel Consistency Filtering (CF) method to reliably remove inaccurate pseudo labels by evaluating the stability of the multiple instance learning module upon noise injections. Via mutually transferring knowledge between the S and T datasets in an iterative manner, the detection performance on the target dataset is significantly improved. Extensive experiments on public benchmarks validate that the proposed method performs favourably against the state-of-the-art methods without increasing the model parameters or inference computational complexity. | ['Chen Li', 'Chong Sun', 'Weitao Wan', 'Xuanyi Du'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['multiple-instance-learning'] | ['methodology'] | [ 5.26507080e-01 3.84605438e-01 -2.30958834e-01 -3.67165089e-01
-1.07863188e+00 -4.10512447e-01 5.74785411e-01 1.34507805e-01
-4.60622400e-01 5.72000742e-01 -5.51373124e-01 2.69768357e-01
-7.20892623e-02 -8.22528839e-01 -9.13164258e-01 -9.56456184e-01
3.15424442e-01 2.77628511e-01 7.26047158e-01 2.23353729e-01
2.72962809e-01 5.48759811e-02 -1.76195848e+00 4.05308664e-01
9.57914710e-01 1.68903017e+00 2.97204316e-01 3.15536171e-01
-6.92070574e-02 7.23598003e-01 -4.68143404e-01 -6.93381846e-01
6.18224978e-01 -3.71091515e-01 -5.18999338e-01 3.12571794e-01
5.45577645e-01 -2.12982893e-01 -1.13381691e-01 1.41166162e+00
3.66348922e-01 5.95033206e-02 5.95441699e-01 -1.27165818e+00
-5.78854203e-01 6.36754990e-01 -7.56918907e-01 1.30686164e-01
-1.95785776e-01 4.23931241e-01 8.81369293e-01 -1.33264434e+00
6.72907054e-01 9.91425335e-01 7.18302011e-01 3.61211002e-01
-1.25852215e+00 -7.78602719e-01 3.40146989e-01 5.22383690e-01
-1.70350266e+00 -3.59386116e-01 7.55095065e-01 -3.98686111e-01
3.57537597e-01 9.10832807e-02 4.27017361e-01 8.88986349e-01
-1.99923500e-01 1.04388511e+00 9.32354271e-01 -2.75054306e-01
4.27925289e-01 7.34413385e-01 2.96403319e-01 6.99456394e-01
3.11655134e-01 9.21641886e-02 -5.26980102e-01 5.24181612e-02
3.02581578e-01 1.24065569e-02 -2.29980588e-01 -7.54398882e-01
-1.07547653e+00 6.88848138e-01 8.04048955e-01 1.24282949e-01
-2.44663134e-01 -1.31611764e-01 5.69837511e-01 9.12512690e-02
5.80099702e-01 2.45756179e-01 -4.97381777e-01 4.02279615e-01
-8.32975805e-01 -4.91798520e-02 5.28055847e-01 1.13644469e+00
1.04104722e+00 -3.40211451e-01 -7.23980665e-01 7.74526536e-01
2.10673019e-01 3.65773737e-01 2.47204766e-01 -6.25897825e-01
4.50713515e-01 9.93318737e-01 1.49078250e-01 -6.70079589e-01
-9.15717613e-03 -8.38477194e-01 -6.79139316e-01 5.85271865e-02
4.67998952e-01 1.54810667e-01 -1.01907492e+00 1.71984410e+00
7.96173394e-01 4.93590534e-01 2.48131603e-02 9.32112873e-01
8.74593198e-01 5.27741849e-01 1.06976524e-01 -2.48669341e-01
1.25662208e+00 -9.48962450e-01 -3.48247349e-01 -2.99438357e-01
4.92321014e-01 -4.48618501e-01 7.96809018e-01 1.85445145e-01
-8.58518362e-01 -8.72201443e-01 -1.16071546e+00 1.26722232e-01
-3.73158336e-01 5.96901774e-01 2.55883813e-01 4.87163365e-01
-4.97102916e-01 4.26085413e-01 -5.57067990e-01 -2.70770580e-01
1.06343305e+00 2.27498487e-01 -1.54522240e-01 -1.07501529e-01
-1.06877327e+00 8.83737803e-01 9.83782589e-01 1.32834598e-01
-1.21389639e+00 -9.59867716e-01 -7.38497734e-01 1.46524921e-01
8.02999377e-01 -4.63345736e-01 1.05081189e+00 -1.13704693e+00
-1.11504042e+00 9.32078242e-01 1.55014902e-01 -4.69666749e-01
8.70146573e-01 1.10348910e-01 -3.13290805e-02 9.66377929e-02
3.18133742e-01 8.48211646e-01 1.20149922e+00 -1.46903598e+00
-1.24500132e+00 -4.66189831e-01 -7.09012076e-02 -2.77888831e-02
-3.11476618e-01 -5.43645382e-01 -5.67192495e-01 -5.29539108e-01
1.80906087e-01 -7.31159985e-01 6.81258887e-02 4.67345536e-01
-4.79237318e-01 -3.68797302e-01 9.85606313e-01 -4.17325854e-01
9.71228540e-01 -2.20085025e+00 1.40897974e-01 1.62632734e-01
7.13409036e-02 4.52165127e-01 -2.33417615e-01 -3.21487367e-01
1.34564057e-01 -2.86653459e-01 -5.53843617e-01 -3.59823763e-01
-2.23337620e-01 1.70706615e-01 -4.48381722e-01 5.84445655e-01
6.12594962e-01 1.02305603e+00 -1.00888467e+00 -5.30623436e-01
1.90503687e-01 2.40723491e-01 -3.71129334e-01 3.89899731e-01
-1.87723875e-01 3.96618724e-01 -3.74272496e-01 6.88391984e-01
7.69982040e-01 -2.53679663e-01 -1.02151774e-01 -3.72157007e-01
1.10991314e-01 -6.37584031e-02 -1.33561313e+00 1.57998848e+00
-1.49916574e-01 2.07061186e-01 -6.34706020e-02 -1.25902641e+00
9.94589567e-01 3.46301794e-02 2.13724762e-01 -7.08362818e-01
1.49910450e-01 1.66683480e-01 -1.17506273e-01 -4.57987607e-01
1.81741327e-01 -1.94602132e-01 -4.88765948e-02 1.32563084e-01
3.86872709e-01 5.88542558e-02 1.84032053e-01 7.39555135e-02
7.68296421e-01 1.35657355e-01 2.19119936e-01 -5.39159216e-02
6.54546201e-01 -1.11377478e-01 8.40639532e-01 1.07640827e+00
-4.44901019e-01 3.17876995e-01 3.29263568e-01 -2.38797471e-01
-1.02009749e+00 -1.18039727e+00 -5.13185680e-01 1.16667569e+00
5.14531195e-01 1.67469606e-01 -6.66789412e-01 -1.28847826e+00
1.84220687e-01 7.32777119e-01 -9.65213180e-01 -6.77354455e-01
-2.00847730e-01 -7.23377824e-01 3.89124632e-01 5.37465572e-01
6.13176227e-01 -8.17393661e-01 -6.12867117e-01 1.16123118e-01
-1.12343512e-01 -1.06549609e+00 -3.86516005e-01 4.51830834e-01
-7.20226407e-01 -1.17794192e+00 -6.15077615e-01 -7.48571396e-01
9.35422897e-01 2.34132856e-01 6.14513755e-01 2.04541883e-03
-5.59297740e-01 2.08593473e-01 -4.32766914e-01 -4.02592748e-01
-4.36773479e-01 -8.64967704e-02 1.88342799e-02 5.37376225e-01
3.99402499e-01 -9.97765064e-02 -5.77779233e-01 4.57206339e-01
-9.42125082e-01 3.38427573e-02 6.58364654e-01 1.00202894e+00
7.09790945e-01 2.33537614e-01 8.18642795e-01 -8.38583291e-01
-1.68104455e-01 -5.20490766e-01 -8.27188909e-01 4.49094564e-01
-6.85520232e-01 3.34519558e-02 3.34030181e-01 -6.75061285e-01
-1.41046500e+00 3.05467188e-01 3.78065199e-01 -6.00090623e-01
1.99486557e-02 8.54424015e-02 -2.98880190e-01 -1.81768671e-01
6.69950843e-01 3.32242608e-01 -1.44127384e-01 -3.65158558e-01
4.91350919e-01 6.63797319e-01 6.95701003e-01 -2.56422013e-01
1.03024364e+00 6.65143073e-01 -2.01102331e-01 -3.82121772e-01
-1.26089120e+00 -6.89655006e-01 -8.70400965e-01 -2.45194703e-01
7.45152473e-01 -9.87386525e-01 -3.95686984e-01 5.28110981e-01
-1.13868558e+00 -1.36673050e-02 -6.14376247e-01 2.30196431e-01
-4.43203241e-01 1.92984551e-01 -2.90666252e-01 -9.49894011e-01
-3.03124517e-01 -1.03608346e+00 1.34869432e+00 3.07963878e-01
2.64818460e-01 -6.65741384e-01 -2.85143077e-01 3.74936104e-01
5.40276766e-02 -9.08934046e-03 8.11083674e-01 -8.05207729e-01
-7.90412366e-01 -5.28850734e-01 -6.18637741e-01 6.70966983e-01
-4.84400615e-02 -2.81532079e-01 -1.18323159e+00 -3.54394138e-01
1.37923673e-01 -5.37928581e-01 1.44913256e+00 1.72440603e-01
1.22983181e+00 -2.70622373e-01 -4.01230782e-01 3.78689140e-01
1.46872091e+00 2.49123238e-02 4.50718910e-01 1.32520989e-01
5.99642932e-01 7.47135878e-01 1.05045223e+00 6.10120118e-01
2.15758026e-01 6.19547725e-01 4.67961848e-01 1.22326724e-01
-2.23026142e-01 -3.12441438e-01 4.54004973e-01 2.72356868e-01
3.19009513e-01 1.30145445e-01 -5.63083708e-01 5.61196923e-01
-1.94333529e+00 -1.03228390e+00 3.82524543e-02 2.47210002e+00
1.09816980e+00 2.62637913e-01 -1.76994354e-02 1.45608678e-01
1.10052681e+00 -1.52827159e-01 -9.42718506e-01 3.40530485e-01
1.70822054e-01 -2.16416866e-01 4.62060750e-01 9.71927941e-02
-1.35527539e+00 7.55958200e-01 5.18235445e+00 1.09192801e+00
-9.72626388e-01 3.83319527e-01 7.19808578e-01 -1.74958512e-01
3.70139591e-02 7.55040944e-02 -1.21006584e+00 6.14176750e-01
4.75795686e-01 -2.38582283e-01 8.22133496e-02 1.17744517e+00
-9.73997563e-02 -1.90122277e-01 -1.41173959e+00 8.15477371e-01
1.71027035e-01 -1.29338539e+00 1.34027585e-01 -1.84306800e-01
8.18155348e-01 -1.55592397e-01 1.59200177e-01 7.12647438e-01
-1.59195408e-01 -4.09204632e-01 1.00380099e+00 6.50420904e-01
7.75114059e-01 -5.51333308e-01 8.16771328e-01 6.24566734e-01
-1.33652902e+00 -4.79198068e-01 -5.73986590e-01 2.20440269e-01
-1.00846633e-01 7.05866039e-01 -8.91257763e-01 5.03219604e-01
7.36599565e-01 7.76742101e-01 -8.72419834e-01 1.31129539e+00
-2.74420649e-01 4.79118913e-01 -1.57739773e-01 2.80522496e-01
2.79520731e-02 -5.24293780e-02 6.97255015e-01 1.02434838e+00
5.81570156e-02 2.58090571e-02 3.91103268e-01 1.30693197e+00
-3.12361568e-01 -2.20429942e-01 -2.67164052e-01 2.50956595e-01
4.56740379e-01 1.44496083e+00 -8.59787703e-01 -6.56210482e-01
-2.43711919e-01 7.94648468e-01 5.29335797e-01 1.85659587e-01
-1.02118325e+00 -3.33152562e-01 1.82408378e-01 -1.04844265e-01
6.80451870e-01 5.63641310e-01 -4.29196119e-01 -1.08212042e+00
2.77825832e-01 -4.30069357e-01 5.63833356e-01 -4.57414150e-01
-1.60616422e+00 2.30630800e-01 -1.04219019e-01 -1.43568468e+00
1.87604785e-01 -4.94415492e-01 -4.81407762e-01 5.95020354e-01
-1.70331800e+00 -1.26653171e+00 -5.43671012e-01 3.60374510e-01
6.74118936e-01 -5.26116118e-02 3.37118387e-01 3.81014049e-01
-7.60787845e-01 8.24548483e-01 -3.18113416e-02 1.06698960e-01
6.90128088e-01 -1.18381751e+00 -3.00999414e-02 8.69793296e-01
-9.92445424e-02 3.01060796e-01 4.39333498e-01 -6.55674160e-01
-1.11242104e+00 -1.64791799e+00 1.47823423e-01 -5.03357589e-01
5.57791889e-01 -3.67607296e-01 -1.24720502e+00 4.93646383e-01
-4.71095085e-01 4.16944504e-01 3.67749989e-01 -3.41308147e-01
-6.56304538e-01 -4.75782692e-01 -1.28011966e+00 -3.90161350e-02
9.58627760e-01 -4.67326909e-01 -7.44844973e-01 2.96416998e-01
7.17457950e-01 -2.19460532e-01 -7.02714443e-01 6.60848081e-01
3.59147549e-01 -6.34308636e-01 8.39217901e-01 -3.51393372e-01
3.26690346e-01 -5.97929597e-01 -9.47451666e-02 -9.09813881e-01
-2.72616744e-01 -5.10304682e-02 -3.54660481e-01 1.40985548e+00
3.34484279e-01 -4.32543367e-01 6.11086428e-01 5.37923396e-01
-5.16365767e-02 -7.54713476e-01 -1.34843671e+00 -9.44106519e-01
-2.20036790e-01 -3.18802148e-01 2.44389802e-01 6.87955081e-01
-3.05644423e-01 2.63376534e-01 -1.87242866e-01 3.29084307e-01
1.03773546e+00 2.40724653e-01 7.15643764e-01 -1.38259387e+00
-2.96473175e-01 -3.39592189e-01 -5.98638356e-01 -8.24202895e-01
1.35446325e-01 -1.18458331e+00 4.66881126e-01 -1.16750085e+00
7.82394886e-01 -6.39213920e-01 -6.16015911e-01 6.21300280e-01
-5.80349147e-01 5.21876097e-01 2.09019825e-01 3.20850372e-01
-1.01597083e+00 8.05203617e-01 1.03043091e+00 -4.29824829e-01
-1.85554281e-01 7.46119022e-02 -6.25815332e-01 6.28181219e-01
2.59547740e-01 -7.31340647e-01 -2.41411537e-01 -4.72203176e-03
-1.02802336e-01 -4.05063391e-01 7.38280952e-01 -9.45675731e-01
3.56386751e-01 1.46366626e-01 5.62181354e-01 -5.39988816e-01
5.89018911e-02 -8.29164684e-01 -2.14871883e-01 5.73696434e-01
-5.03092170e-01 -1.02946651e+00 1.23359971e-01 1.09992146e+00
-1.25547079e-02 -2.98870653e-01 1.36464524e+00 1.02282360e-01
-9.18971300e-01 4.09588069e-01 3.26885164e-01 1.77753214e-02
1.51574218e+00 -3.65719825e-01 -3.15044522e-01 3.05578768e-01
-6.79565072e-01 4.14880276e-01 3.31565976e-01 6.12022042e-01
5.52838087e-01 -1.24228632e+00 -6.35803998e-01 2.77001470e-01
6.29408777e-01 1.73045903e-01 3.57674748e-01 1.05205369e+00
3.14264566e-01 1.14904597e-01 -9.54056084e-02 -9.52192247e-01
-1.08803177e+00 8.35266471e-01 4.81090814e-01 -7.08561093e-02
-5.59960008e-01 1.03997731e+00 5.38172305e-01 -3.86732489e-01
4.76058394e-01 -9.49219242e-02 -7.65816867e-02 4.15028960e-01
6.95696115e-01 5.35737276e-01 4.38544787e-02 -4.68036592e-01
-4.32871819e-01 3.54920387e-01 -3.73255730e-01 3.20308775e-01
1.22779560e+00 -2.49371096e-01 -4.15246859e-02 5.27421892e-01
1.15212035e+00 -5.68064928e-01 -1.70959473e+00 -5.90735078e-01
1.46995738e-01 -5.29002905e-01 5.48350019e-03 -8.11793149e-01
-1.08128071e+00 6.52314246e-01 9.41391230e-01 -7.07076639e-02
1.05379140e+00 3.59494299e-01 5.27337909e-01 3.15667897e-01
4.68761176e-01 -1.31300282e+00 3.47863108e-01 1.75841063e-01
7.60525823e-01 -1.63244402e+00 -1.07257441e-01 -4.75100040e-01
-6.00307822e-01 8.70262682e-01 8.67325962e-01 6.68310970e-02
3.49003762e-01 -8.21921378e-02 -3.56139541e-01 -3.76282968e-02
-6.12281203e-01 -3.86592805e-01 4.87826794e-01 4.61117297e-01
-2.92216033e-01 -1.09954007e-01 -6.92825615e-02 9.11554098e-01
4.24415857e-01 5.79238944e-02 2.75293976e-01 6.02285385e-01
-7.85181999e-01 -6.18230522e-01 -3.80202532e-01 6.90721393e-01
6.44812807e-02 1.62989423e-01 -2.90122181e-01 3.89053851e-01
6.75822318e-01 7.83861578e-01 5.66362403e-02 -2.41618559e-01
4.04856741e-01 2.36797005e-01 2.97817230e-01 -7.79905438e-01
-4.93949324e-01 -2.22917259e-01 -3.69254410e-01 -4.72436666e-01
-2.82512724e-01 -4.07404095e-01 -1.12185824e+00 2.82740504e-01
-8.54564488e-01 8.68228916e-03 3.76911104e-01 1.03744328e+00
4.54881966e-01 4.82395172e-01 7.10560679e-01 -8.37627351e-01
-9.75830078e-01 -9.79306221e-01 -5.90673447e-01 5.54169297e-01
3.58534932e-01 -9.33201134e-01 -5.20455301e-01 8.70036036e-02] | [9.354456901550293, 1.404369831085205] |
1b734fa6-bd43-4137-a979-85d54674c44c | rm-prt-realistic-robotic-manipulation | 2306.11335 | null | https://arxiv.org/abs/2306.11335v2 | https://arxiv.org/pdf/2306.11335v2.pdf | RM-PRT: Realistic Robotic Manipulation Simulator and Benchmark with Progressive Reasoning Tasks | Recently, the advent of pre-trained large-scale language models (LLMs) like ChatGPT and GPT-4 have significantly advanced the machine's natural language understanding capabilities. This breakthrough has allowed us to seamlessly integrate these open-source LLMs into a unified robot simulator environment to help robots accurately understand and execute human natural language instructions. To this end, in this work, we introduce a realistic robotic manipulation simulator and build a Robotic Manipulation with Progressive Reasoning Tasks (RM-PRT) benchmark on this basis. Specifically, the RM-PRT benchmark builds a new high-fidelity digital twin scene based on Unreal Engine 5, which includes 782 categories, 2023 objects, and 15K natural language instructions generated by ChatGPT for a detailed evaluation of robot manipulation. We propose a general pipeline for the RM-PRT benchmark that takes as input multimodal prompts containing natural language instructions and automatically outputs actions containing the movement and position transitions. We set four natural language understanding tasks with progressive reasoning levels and evaluate the robot's ability to understand natural language instructions in two modes of adsorption and grasping. In addition, we also conduct a comprehensive analysis and comparison of the differences and advantages of 10 different LLMs in instruction understanding and generation quality. We hope the new simulator and benchmark will facilitate future research on language-guided robotic manipulation. Project website: https://necolizer.github.io/RM-PRT/ . | ['Xiaodan Liang', 'Mas Ma', 'Fengda Zhu', 'Yuhang Wen', 'Zixuan Li', 'Hetao Zheng', 'Kaidong Zhang', 'Pengzhen Ren'] | 2023-06-20 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [-2.37600878e-01 2.16860086e-01 4.05169129e-02 -4.97151792e-01
-6.67154193e-01 -7.76524842e-01 5.11522174e-01 -1.79276407e-01
-8.07899088e-02 2.92887896e-01 2.06599265e-01 -4.96524841e-01
-2.08495557e-02 -8.00779700e-01 -1.15097880e+00 3.21431225e-03
-1.92123830e-01 9.14451003e-01 1.54869124e-01 -8.81870389e-01
3.73296410e-01 3.81838232e-01 -1.67189527e+00 7.18987048e-01
8.88037205e-01 5.45893371e-01 9.68498409e-01 1.06071818e+00
-1.33301377e-01 1.31417286e+00 -4.57396358e-01 -2.55650710e-02
2.43227378e-01 -1.58655569e-02 -1.24334991e+00 -3.58613044e-01
-6.47516316e-03 -7.85627544e-01 -5.77428460e-01 7.40712762e-01
1.91594154e-01 3.39243561e-01 5.30475616e-01 -1.61627269e+00
-5.84163725e-01 1.00426853e+00 -3.50178257e-02 -6.27895772e-01
1.06851363e+00 9.01222885e-01 6.58367932e-01 -5.32202780e-01
7.74632692e-01 2.15403152e+00 2.98376113e-01 6.87862337e-01
-6.72872126e-01 -5.41358769e-01 1.56489238e-01 3.20349813e-01
-1.00909781e+00 2.76859431e-03 1.82490095e-01 -3.61916542e-01
1.35408878e+00 -8.75970125e-02 3.54527950e-01 1.39512753e+00
6.20588005e-01 1.17132235e+00 5.87381065e-01 -2.54797757e-01
2.42773667e-02 -3.39272261e-01 2.88000137e-01 1.02500200e+00
-1.66572466e-01 2.92264551e-01 -2.83273071e-01 1.22790225e-01
1.11708021e+00 -2.95757711e-01 -5.60486913e-02 -5.58838844e-01
-1.66973197e+00 5.30402064e-01 6.40853941e-01 5.04621081e-02
-3.72864544e-01 6.16805911e-01 7.65586495e-01 5.01906097e-01
-4.57533240e-01 7.43673742e-01 -6.93103790e-01 -5.41630208e-01
4.54314679e-01 7.58760035e-01 1.05985594e+00 1.66476023e+00
5.58327675e-01 -1.46225065e-01 -1.90929264e-01 8.58185649e-01
3.94956708e-01 6.19641960e-01 3.87512416e-01 -1.67406523e+00
6.44106686e-01 6.76122665e-01 3.27665806e-01 -9.94980276e-01
-6.69020534e-01 3.96456689e-01 -2.04453349e-01 3.80855441e-01
3.81617188e-01 5.35330400e-02 -8.74630690e-01 1.74751079e+00
2.20734492e-01 -4.14863944e-01 6.75874054e-01 1.07451999e+00
1.06786966e+00 9.64342713e-01 2.35646829e-01 7.43109822e-01
1.49068153e+00 -1.32951474e+00 -5.81244886e-01 -3.73634636e-01
1.11927748e+00 -5.40225089e-01 1.74497557e+00 7.45340228e-01
-9.27363575e-01 -9.26164806e-01 -9.59816575e-01 -4.34292406e-01
-4.43312973e-01 3.67042124e-01 9.70300972e-01 -1.53011888e-01
-1.00368416e+00 4.89341587e-01 -1.14619029e+00 -6.21582747e-01
1.07669113e-02 3.23844284e-01 -4.13647920e-01 -6.15706980e-01
-1.01430368e+00 1.34068000e+00 8.12413752e-01 5.30563965e-02
-1.46057606e+00 -5.00731766e-01 -1.35651207e+00 -2.21645594e-01
4.79708344e-01 -7.23561406e-01 2.12583041e+00 -2.90812731e-01
-2.08379912e+00 7.33577251e-01 2.01624498e-01 -2.50502974e-01
2.95282423e-01 -6.38336778e-01 8.78938809e-02 3.30006599e-01
2.79666573e-01 1.37533677e+00 1.76244855e-01 -1.34137714e+00
-4.99830812e-01 -6.05397187e-02 7.80593991e-01 2.22086534e-01
5.43507576e-01 -1.32257976e-02 -4.61728662e-01 -3.85802805e-01
-1.10509701e-01 -1.15299249e+00 -2.59730279e-01 7.90427849e-02
-2.55875468e-01 -5.24616063e-01 6.47798479e-01 -5.85397243e-01
1.52490392e-01 -2.00887275e+00 5.65211058e-01 -3.69267374e-01
-1.36253983e-01 -3.47808637e-02 -7.91052520e-01 7.10648596e-01
1.10138863e-01 -3.22507620e-01 -3.75984274e-02 -5.80777265e-02
6.24760270e-01 6.35370314e-01 -5.14055312e-01 -4.55311611e-02
3.75960231e-01 1.29033756e+00 -1.33663642e+00 -2.11625442e-01
6.62679076e-01 1.96037635e-01 -8.49546552e-01 5.59103489e-01
-9.95296121e-01 6.64446592e-01 -5.70452452e-01 5.76138973e-01
4.10771281e-01 5.47522977e-02 7.41744787e-02 -1.02430992e-01
3.91905807e-04 3.84495944e-01 -6.54883683e-01 2.54561687e+00
-9.33403671e-01 4.66782451e-01 1.99579641e-01 -5.12338877e-01
7.92564154e-01 1.76144704e-01 1.02914408e-01 -5.01933336e-01
3.62988681e-01 2.35251501e-01 1.29685923e-01 -1.05581975e+00
6.77903891e-01 2.17321247e-01 -6.54250562e-01 4.16869819e-01
1.97536692e-01 -1.23570800e+00 2.87736088e-01 4.58233684e-01
1.26504779e+00 8.19249570e-01 1.39288619e-01 -1.06877707e-01
2.05847189e-01 5.53623319e-01 -1.03758991e-01 7.55474150e-01
-1.39094353e-01 3.09604079e-01 3.85730982e-01 -3.39023560e-01
-8.16912115e-01 -1.22064734e+00 4.31852072e-01 1.33649266e+00
4.44444925e-01 -5.57199061e-01 -6.02911949e-01 -3.24051559e-01
1.52272824e-02 1.34144127e+00 -2.22254589e-01 -3.91484648e-01
-6.60847485e-01 -5.16986735e-02 7.42662430e-01 5.18181801e-01
6.84351027e-01 -1.67144835e+00 -9.81826186e-01 2.88250204e-02
-5.54101944e-01 -1.40077758e+00 -1.46954078e-02 4.16093282e-02
-6.06143773e-01 -1.19122350e+00 -2.05036238e-01 -1.11856878e+00
5.02625883e-01 2.82234669e-01 1.15277159e+00 -6.65273890e-02
-3.40225965e-01 9.51429069e-01 -9.10823584e-01 -3.76713663e-01
-1.07954168e+00 -8.00236389e-02 2.18969341e-02 -1.08765745e+00
4.45901603e-02 -1.85257062e-01 -9.48042423e-02 4.31609362e-01
-8.53350401e-01 4.96714383e-01 7.67626584e-01 5.79547882e-01
1.50160317e-03 -1.65110394e-01 2.88440228e-01 -6.54924959e-02
8.86124253e-01 -3.79909962e-01 -5.47846079e-01 1.99885100e-01
3.36124063e-01 2.70089239e-01 4.58200157e-01 -4.76751119e-01
-1.27258766e+00 7.54793808e-02 -2.04805225e-01 -2.31150925e-01
-4.88208383e-01 5.68616152e-01 -1.59997284e-01 4.67578806e-02
6.71315849e-01 1.03951380e-01 5.83182350e-02 -1.21875398e-01
8.53989363e-01 7.50519395e-01 8.92956078e-01 -1.41081738e+00
4.27226841e-01 6.28251955e-02 -3.52872044e-01 -6.53782010e-01
-4.02279824e-01 -1.24875411e-01 -5.77785254e-01 -1.17797278e-01
7.93555200e-01 -9.03549910e-01 -1.38708293e+00 6.50579572e-01
-1.66667473e+00 -1.24952996e+00 -4.34120856e-02 5.36025643e-01
-1.34360743e+00 3.15024912e-01 -1.09614778e+00 -6.21145844e-01
-8.27547610e-02 -1.81093407e+00 1.52436221e+00 2.99596731e-02
-7.25367546e-01 -3.91693711e-01 -1.54720306e-01 2.38181397e-01
2.06727296e-01 1.30255595e-01 1.04866767e+00 -2.87343293e-01
-6.67433202e-01 -1.41243950e-01 -3.42544466e-01 1.32825688e-01
1.40459493e-01 -1.64833486e-01 -3.93713742e-01 -9.97280180e-02
-1.52043372e-01 -1.10722780e+00 1.06255636e-01 1.45080909e-01
1.06262743e+00 5.38505837e-02 -2.82686919e-01 2.60187685e-01
9.25343215e-01 4.28120553e-01 8.08973968e-01 3.67547393e-01
5.71546793e-01 8.02683413e-01 1.36735463e+00 1.27599061e-01
7.44791985e-01 5.87467968e-01 6.65674031e-01 5.03730118e-01
9.22743138e-03 -4.16689217e-01 8.12089682e-01 7.42143095e-01
1.16564475e-01 -2.59410411e-01 -1.19111478e+00 3.32176626e-01
-1.87996352e+00 -5.18741548e-01 -2.98974048e-02 1.48559678e+00
8.18217039e-01 -1.65892262e-02 -3.88521254e-01 -3.92434508e-01
3.25513572e-01 -3.30814451e-01 -6.87250018e-01 -7.27533042e-01
3.02202374e-01 8.12299550e-02 3.59883867e-02 6.73297524e-01
-7.16066301e-01 1.52071357e+00 5.76660490e+00 6.10701680e-01
-8.02852213e-01 -3.40282649e-01 1.64766517e-02 2.44849592e-01
2.02094883e-01 -2.18756437e-01 -4.30128068e-01 -1.19869895e-01
7.99144804e-01 -1.25787333e-01 6.55291080e-01 1.16760647e+00
2.73681164e-01 -4.34703320e-01 -1.53733802e+00 9.37942386e-01
-7.99328908e-02 -8.43701839e-01 3.99215370e-01 -5.95572412e-01
1.48473784e-01 3.81491840e-01 -2.58185208e-01 1.14057803e+00
1.08279943e+00 -1.03334510e+00 9.64044273e-01 3.68931055e-01
6.06763482e-01 -4.34807122e-01 5.03855884e-01 5.89705110e-01
-1.10614562e+00 -3.44044387e-01 -3.01001519e-01 -4.03649896e-01
4.62372750e-01 -3.13405991e-01 -1.02257550e+00 7.10501134e-01
7.35436797e-01 7.49095678e-01 1.82756558e-02 4.92550403e-01
-7.27623761e-01 -1.42489672e-01 -1.97285265e-01 -1.08475722e-01
3.39569598e-01 2.25240532e-02 5.84474087e-01 8.86371613e-01
1.41036555e-01 5.17555833e-01 4.77542788e-01 1.06589890e+00
2.84733087e-01 -3.05080652e-01 -6.86028123e-01 -2.78924018e-01
2.59742260e-01 9.70806539e-01 -5.63520908e-01 -5.36959589e-01
-1.70705363e-01 1.12117720e+00 3.35151166e-01 2.28733405e-01
-1.12730873e+00 -5.12460530e-01 9.24231052e-01 -4.85929072e-01
-1.67226136e-01 -8.78283739e-01 1.23889923e-01 -1.02635884e+00
-1.96042329e-01 -1.36264002e+00 -7.87813067e-02 -1.60951328e+00
-9.93076444e-01 5.62540472e-01 5.46930254e-01 -9.34554875e-01
-3.71227026e-01 -1.33020687e+00 -2.82709509e-01 4.62297976e-01
-1.20457780e+00 -1.21655631e+00 -7.40594625e-01 2.80163825e-01
1.18249655e+00 3.31744812e-02 1.08705878e+00 -1.07413888e-01
-2.27653548e-01 5.77823855e-02 -5.78072190e-01 8.74844193e-03
6.99609995e-01 -9.90206122e-01 7.87706673e-01 4.00584079e-02
-5.44561863e-01 8.51539969e-01 7.99745262e-01 -8.27131987e-01
-1.90329647e+00 -8.57381523e-01 -1.38147380e-02 -8.58071804e-01
8.26087773e-01 -4.38342243e-01 -6.50320947e-01 1.14711237e+00
1.63533270e-01 -6.46621406e-01 -6.11159578e-02 -3.84804755e-01
-1.80332929e-01 5.93854845e-01 -1.13013327e+00 1.06802189e+00
1.38067329e+00 -3.37229341e-01 -1.14097309e+00 6.43544614e-01
1.45517361e+00 -9.72604752e-01 -8.54892015e-01 6.97906673e-01
6.72223389e-01 -5.01445353e-01 9.63128924e-01 -5.84447205e-01
1.02300107e+00 -3.06058556e-01 -4.05044287e-01 -1.50226247e+00
-7.17205703e-02 -4.28863466e-01 2.59394437e-01 6.27298594e-01
1.15296386e-01 -6.20897949e-01 2.80557275e-01 6.18181765e-01
-7.28090644e-01 -5.40683329e-01 -3.16715837e-01 -8.69989276e-01
2.44643018e-01 -8.94321144e-01 6.18367076e-01 4.57630605e-01
7.35215485e-01 2.47939304e-01 1.54397517e-01 2.91022003e-01
7.04111308e-02 1.25617623e-01 1.37727296e+00 -6.75943732e-01
-2.97670156e-01 -3.77998352e-01 -1.96209133e-01 -1.64952123e+00
5.92836559e-01 -9.73491013e-01 8.68553340e-01 -1.73130441e+00
-1.13904779e-03 -3.78916621e-01 5.03125548e-01 8.46492350e-01
3.21327686e-01 -4.68701512e-01 3.08982879e-01 -2.49628425e-02
-6.61287665e-01 9.65214014e-01 1.66632795e+00 -3.15285206e-01
-1.99123070e-01 -4.98442680e-01 -2.78356701e-01 7.86156416e-01
9.90180790e-01 4.74196076e-02 -4.89644676e-01 -9.69537437e-01
1.18383463e-03 2.08273992e-01 5.61132610e-01 -1.03478134e+00
9.60187986e-02 -3.41958791e-01 -2.39087522e-01 -5.54276288e-01
4.79278922e-01 -7.25608349e-01 -1.46665081e-01 8.34455371e-01
-5.02835929e-01 1.50169626e-01 8.16722870e-01 2.19818264e-01
-1.01662226e-01 -7.79444203e-02 3.64206910e-01 -5.19264221e-01
-1.28035235e+00 -2.18942136e-01 -8.22638869e-01 -3.86397421e-01
1.17483640e+00 1.65011764e-01 -5.68657696e-01 -5.71645617e-01
-6.03039384e-01 8.83360147e-01 5.53658843e-01 1.09618139e+00
7.60546863e-01 -1.13714528e+00 -5.24778605e-01 -9.15182605e-02
4.55499053e-01 3.07841837e-01 2.22680494e-01 4.82874930e-01
-1.15727675e+00 7.45102882e-01 -5.49782395e-01 -6.49912357e-01
-8.80590558e-01 5.39817810e-01 2.22956285e-01 -5.21510355e-02
-6.93486631e-01 7.47563064e-01 4.78104204e-01 -1.55457091e+00
2.74439067e-01 -1.15725625e+00 1.25234857e-01 -9.40039575e-01
4.21222359e-01 2.02961564e-01 -4.78896797e-01 -4.40831929e-01
-1.89843968e-01 4.24643338e-01 8.39461908e-02 -1.10357165e-01
1.16433096e+00 5.92322052e-02 -3.74608815e-01 4.86037791e-01
8.54572713e-01 -5.79821050e-01 -1.11140060e+00 2.57068813e-01
-3.36944163e-02 -9.31143239e-02 -6.25990450e-01 -9.76207733e-01
-5.24434328e-01 7.03735709e-01 5.90167791e-02 -4.57412720e-01
5.85024774e-01 2.18853906e-01 9.32545781e-01 1.28543341e+00
1.15361404e+00 -7.71678627e-01 5.11598408e-01 1.28775632e+00
1.68745375e+00 -1.28173029e+00 -4.46256846e-01 -6.34848535e-01
-6.54869854e-01 1.25366759e+00 1.14811826e+00 -1.73634738e-01
5.54145779e-03 3.46952438e-01 1.04020044e-01 -4.28033799e-01
-7.69107044e-01 -1.30800428e-02 -2.24081129e-01 7.03863859e-01
1.93808392e-01 1.01439916e-01 1.60438687e-01 7.11679101e-01
-6.66947782e-01 3.06420296e-01 5.39679050e-01 1.36355388e+00
-4.33419168e-01 -9.22193050e-01 -3.93932730e-01 -7.57028833e-02
4.56536114e-01 2.20992759e-01 -4.50784683e-01 9.85458255e-01
-2.55217940e-01 1.07394826e+00 -2.60995656e-01 -5.95927775e-01
6.82015240e-01 -8.83685723e-02 9.20194030e-01 -9.40080404e-01
-3.55074823e-01 -7.51914084e-01 2.73531288e-01 -1.13328850e+00
5.07816486e-02 -3.44696134e-01 -2.11249352e+00 -1.91042259e-01
1.01766452e-01 -5.12698628e-02 7.84134924e-01 8.53919566e-01
3.57218742e-01 7.80601680e-01 -1.69388071e-01 -1.63040781e+00
-6.53084874e-01 -1.18580425e+00 1.01761945e-01 2.79121965e-01
-4.03985418e-02 -8.83600354e-01 -1.49528608e-01 4.62518353e-03] | [4.455013275146484, 0.7481858134269714] |
7d1a46e1-9155-4619-a4c8-f967b7a72304 | abusive-language-detection-with-graph | 1904.04073 | null | http://arxiv.org/abs/1904.04073v1 | http://arxiv.org/pdf/1904.04073v1.pdf | Abusive Language Detection with Graph Convolutional Networks | Abuse on the Internet represents a significant societal problem of our time.
Previous research on automated abusive language detection in Twitter has shown
that community-based profiling of users is a promising technique for this task.
However, existing approaches only capture shallow properties of online
communities by modeling follower-following relationships. In contrast, working
with graph convolutional networks (GCNs), we present the first approach that
captures not only the structure of online communities but also the linguistic
behavior of the users within them. We show that such a heterogeneous
graph-structured modeling of communities significantly advances the current
state of the art in abusive language detection. | ['Ekaterina Shutova', 'Helen Yannakoudakis', 'Marco del Tredici', 'Pushkar Mishra'] | 2019-04-05 | abusive-language-detection-with-graph-1 | https://aclanthology.org/N19-1221 | https://aclanthology.org/N19-1221.pdf | naacl-2019-6 | ['abuse-detection'] | ['natural-language-processing'] | [-7.79590458e-02 7.21769556e-02 -6.05540276e-01 -1.18538633e-01
-4.67855856e-03 -8.77588868e-01 8.18723023e-01 9.00063097e-01
-2.36450374e-01 2.88029611e-01 3.78012508e-01 -4.85751331e-01
3.02916795e-01 -1.13773906e+00 -1.70897543e-01 -5.39891832e-02
-4.62722540e-01 5.43625295e-01 3.76020998e-01 -5.19854546e-01
3.11955273e-01 6.79712772e-01 -7.34344125e-01 2.27785394e-01
8.32135260e-01 2.35161066e-01 -7.52798975e-01 5.31052172e-01
-6.01090848e-01 1.28699625e+00 -5.03083467e-01 -1.01601994e+00
-1.27409786e-01 -2.69941539e-01 -8.82564306e-01 1.30411014e-01
6.02102399e-01 -3.72340620e-01 -8.79017591e-01 1.37893593e+00
2.56879646e-02 -2.79394597e-01 5.15592158e-01 -1.38146973e+00
-4.23974335e-01 1.24319506e+00 -9.21592772e-01 5.64038217e-01
5.42568088e-01 -5.80430664e-02 1.22278368e+00 -2.27559894e-01
9.63220000e-01 1.30186987e+00 9.51863348e-01 4.10754323e-01
-1.21847856e+00 -7.62369156e-01 2.02986121e-01 -9.08898190e-02
-1.24955416e+00 -1.94105238e-01 1.06371248e+00 -9.35480118e-01
7.84309208e-01 8.74858871e-02 9.71860707e-01 1.43217039e+00
-3.77340496e-01 6.26241505e-01 6.54249966e-01 -1.81421712e-01
-1.63523495e-01 2.06345975e-01 8.26419532e-01 1.04518437e+00
8.01673651e-01 -6.40857935e-01 -3.17070901e-01 -9.90548849e-01
3.32186222e-01 1.04488418e-01 9.15291458e-02 -3.32144409e-01
-3.67086649e-01 1.44577122e+00 6.15248621e-01 8.81053448e-01
-8.86572897e-02 3.33371848e-01 8.04427266e-01 1.58036664e-01
9.08900201e-01 3.74974251e-01 4.47348803e-01 8.20911527e-02
-1.21731591e+00 3.25304836e-01 1.28642797e+00 5.82771659e-01
7.33930290e-01 -8.72560069e-02 2.97940135e-01 7.35558569e-01
2.80206054e-01 6.29556179e-02 2.72055060e-01 -5.27185917e-01
5.10289848e-01 1.03620207e+00 -3.10178190e-01 -1.64043891e+00
-3.93783480e-01 -3.88789833e-01 -6.30312562e-01 -3.70128185e-01
4.95644510e-01 -2.34666258e-01 -4.24367607e-01 1.48113132e+00
2.86310017e-01 1.29437223e-01 -7.57569075e-01 2.82293051e-01
7.18707979e-01 1.93457067e-01 6.23434670e-02 -1.22273304e-01
9.32339311e-01 -6.93986058e-01 -7.22143292e-01 -3.39974046e-01
8.93335044e-01 -1.66817471e-01 4.44391519e-01 1.02760591e-01
-8.79682243e-01 3.71853590e-01 -9.07899737e-01 -4.51341271e-02
-7.81559348e-01 -8.06840062e-01 8.60567868e-01 1.29157758e+00
-1.13175035e+00 6.40331328e-01 -5.70014536e-01 -5.61513782e-01
8.22562337e-01 1.33590937e-01 -3.37359667e-01 -7.50927031e-02
-1.23411405e+00 7.58248806e-01 -1.61908697e-02 -3.74183714e-01
-7.06869721e-01 -3.98706049e-01 -9.83017087e-01 -9.40105766e-02
4.53023493e-01 -2.31263265e-01 1.03180051e+00 -1.09240401e+00
-8.15380871e-01 1.24538910e+00 -1.46864414e-01 -7.71450222e-01
5.18680930e-01 6.28839955e-02 -4.03849632e-01 3.57094318e-01
4.58790660e-02 -6.66899011e-02 5.78923583e-01 -1.12865472e+00
-1.43028006e-01 -4.95972842e-01 2.75004447e-01 -3.34268779e-01
-1.09353161e+00 6.79089606e-01 -1.21582523e-02 -3.37435186e-01
-3.38804394e-01 -7.72836208e-01 -2.73732007e-01 -1.38893232e-01
-7.51633644e-01 -3.53573382e-01 9.08146858e-01 -8.03543329e-01
1.90999949e+00 -1.87661600e+00 5.42158186e-02 7.51006603e-01
1.18025851e+00 4.02587771e-01 1.57170370e-01 1.02063513e+00
2.84879386e-01 9.54701185e-01 -2.98611522e-01 -8.47852468e-01
-9.82251614e-02 1.91483181e-02 -2.02713266e-01 9.69857037e-01
-1.12442262e-01 7.78383613e-01 -1.19557667e+00 -5.48867762e-01
-1.41971409e-01 3.85077178e-01 -6.28297746e-01 -2.07154214e-01
-1.20503023e-01 7.71703050e-02 -4.58621264e-01 4.02015865e-01
6.49720550e-01 -4.44332570e-01 7.07819998e-01 5.45471489e-01
6.53300136e-02 5.67646623e-01 -5.93467653e-01 9.38425541e-01
-1.34901583e-01 9.58259881e-01 5.94052255e-01 -7.75789142e-01
5.16290426e-01 -1.73782352e-02 6.33831620e-01 -9.91403684e-02
3.80933523e-01 9.40005928e-02 2.46672049e-01 -4.21071708e-01
4.34844315e-01 1.14252046e-01 -1.04507141e-01 7.17559874e-01
2.86185736e-04 1.91910729e-01 4.83518243e-01 9.17481065e-01
1.45434666e+00 -6.65079355e-01 5.46290517e-01 -1.24163374e-01
4.55502003e-01 8.89534783e-03 8.23633373e-02 7.60246456e-01
-5.93897939e-01 3.00083309e-01 1.00697291e+00 -2.77024031e-01
-1.03162408e+00 -6.06766462e-01 2.12791815e-01 1.19079602e+00
-2.49592379e-01 -8.68120015e-01 -1.05070269e+00 -8.06951582e-01
3.93449664e-01 3.71259958e-01 -7.77423322e-01 6.84921592e-02
-9.00981724e-01 -7.62294412e-01 1.03014326e+00 -1.08641237e-01
2.31844485e-01 -8.42664778e-01 9.20114890e-02 2.11127847e-01
-3.25343996e-01 -1.48959064e+00 -4.89521265e-01 -5.54606318e-01
-6.52210951e-01 -1.51139772e+00 -1.67736009e-01 -4.66714114e-01
3.90228391e-01 3.88531834e-01 1.32949340e+00 8.97720814e-01
-2.33387232e-01 2.69534796e-01 -2.69199789e-01 -1.72060296e-01
-7.64584601e-01 4.63158190e-01 -1.22921821e-02 4.77053821e-01
5.57762980e-01 -1.11121762e+00 -1.68228000e-01 -3.49839538e-01
-7.24890292e-01 -5.47738135e-01 -1.76958650e-01 8.50846544e-02
-4.70527053e-01 -2.33654469e-01 1.46583989e-01 -1.53864515e+00
1.16077614e+00 -1.20362914e+00 -2.47916400e-01 -9.85542908e-02
-4.18355167e-01 -6.05354190e-01 5.51836252e-01 -5.84791243e-01
-3.32102239e-01 -4.94394362e-01 -7.43821636e-02 -2.01823667e-01
-1.54747188e-01 5.90119123e-01 2.69818991e-01 -3.16525161e-01
7.74068117e-01 4.83877100e-02 1.75281521e-03 -4.16757882e-01
2.25519717e-01 8.12453687e-01 -1.31757841e-01 -1.38802379e-01
8.43861938e-01 8.51036489e-01 5.22448570e-02 -1.28635335e+00
-7.52673626e-01 -9.44675684e-01 -7.55249023e-01 -3.48789901e-01
7.29277372e-01 -6.82422459e-01 -8.68046284e-01 5.79802334e-01
-1.20770895e+00 -5.61875626e-02 3.77836645e-01 -3.07060599e-01
1.57130495e-01 1.01015103e+00 -1.24293721e+00 -1.28080428e+00
-4.02938098e-01 -6.43500149e-01 6.37435675e-01 -3.44504893e-01
-6.23944223e-01 -1.41087008e+00 4.23316360e-01 4.99287397e-01
4.21317071e-01 8.62553596e-01 8.46232414e-01 -1.13540769e+00
-2.22086579e-01 -4.69186276e-01 -2.22851679e-01 -4.83175144e-02
-2.76669078e-02 2.59145916e-01 -7.43392289e-01 -1.59122169e-01
-3.83926153e-01 -1.13468170e-01 8.10230792e-01 -7.81468004e-02
7.26099730e-01 -7.66203523e-01 -7.43289411e-01 1.79921076e-01
1.37245047e+00 -4.98603851e-01 3.70424926e-01 2.17212096e-01
1.26876080e+00 7.25413322e-01 -4.95276839e-01 4.17406619e-01
4.43309337e-01 5.45030057e-01 6.08581185e-01 1.70041218e-01
3.31656896e-02 -5.32298446e-01 3.90891641e-01 6.34807348e-01
1.33782834e-01 -4.64198709e-01 -1.54940116e+00 8.51837277e-01
-1.79561257e+00 -1.25473630e+00 -9.52706754e-01 1.60044801e+00
6.98878109e-01 3.32833618e-01 1.00189424e+00 1.55774787e-01
1.15124893e+00 4.55142945e-01 -5.98314181e-02 -5.77394903e-01
-6.86187446e-02 3.99222635e-02 5.65434217e-01 9.28635299e-01
-1.08248317e+00 9.29209113e-01 7.04477119e+00 6.05764687e-01
-8.40604186e-01 3.34337890e-01 2.13430256e-01 -9.20692906e-02
-3.83026212e-01 -4.75671999e-02 -6.19760454e-01 6.15637898e-01
1.12143314e+00 -9.12233889e-02 6.66578591e-01 7.03015387e-01
1.75454468e-01 1.90057561e-01 -8.28988016e-01 6.18095756e-01
2.97050208e-01 -1.47082019e+00 1.75044104e-01 7.05111802e-01
6.05395973e-01 4.13927495e-01 -1.63762644e-01 -9.31958184e-02
7.45940387e-01 -1.06423831e+00 5.96769273e-01 1.89941436e-01
2.50497282e-01 -7.13067830e-01 4.90988910e-01 7.38690257e-01
-7.94439018e-01 -3.21551144e-01 -9.76040680e-03 -3.52772415e-01
2.10246965e-01 8.27844024e-01 -9.48469758e-01 1.86436534e-01
3.68710130e-01 1.06845665e+00 -8.28618765e-01 9.09137785e-01
-2.16337457e-01 1.18688011e+00 -4.49251294e-01 -2.89988786e-01
5.05339921e-01 -4.30121064e-01 9.21634018e-01 1.68332195e+00
-2.90672630e-01 -3.34081918e-01 4.52026516e-01 8.65878701e-01
-6.04428709e-01 1.86242715e-01 -1.29008079e+00 -7.12218761e-01
3.65294099e-01 1.31992126e+00 -6.34625673e-01 -3.04964602e-01
-4.21691328e-01 5.86625040e-01 1.03143752e+00 -1.69902071e-01
-3.44805777e-01 -6.00394346e-02 7.08636999e-01 9.97843623e-01
-5.04527502e-02 -7.16606081e-01 -1.59231752e-01 -1.26660597e+00
-3.89990360e-01 -8.81294310e-01 6.44631088e-01 -1.70609161e-01
-1.72064960e+00 3.22045237e-01 -1.23849317e-01 -5.74054062e-01
-3.46270174e-01 -3.49594146e-01 -8.74518454e-01 5.38832128e-01
-1.21483958e+00 -1.45162916e+00 -8.65936130e-02 5.06681859e-01
9.17462166e-03 6.15494289e-02 6.33481622e-01 3.61277401e-01
-7.54767299e-01 4.25839663e-01 -7.70293549e-02 9.07029390e-01
-2.75573647e-03 -1.09294021e+00 6.05840683e-01 8.55549037e-01
3.21911126e-01 9.02271807e-01 7.10459530e-01 -1.07229626e+00
-9.02244389e-01 -8.59541833e-01 1.31318462e+00 -6.73575759e-01
1.58475614e+00 -9.91499543e-01 -9.90073860e-01 9.72521126e-01
3.96923423e-01 -1.11151412e-01 9.52048719e-01 4.47412312e-01
-9.15683389e-01 4.77085859e-01 -1.19444144e+00 6.86186016e-01
1.44867575e+00 -9.78954732e-01 -1.05695643e-01 6.63076639e-01
3.60226333e-01 1.57685146e-01 -5.63063860e-01 -2.23798826e-01
3.11599135e-01 -1.11116695e+00 6.30639017e-01 -1.03057849e+00
5.17379344e-01 3.32195073e-01 2.85407126e-01 -9.25398409e-01
-2.82952696e-01 -8.72017205e-01 -5.72606444e-01 1.27714503e+00
2.77506113e-01 -7.26601541e-01 8.91877711e-01 3.54364693e-01
4.25194442e-01 -3.00350338e-01 -6.33730233e-01 -4.23594922e-01
4.91772085e-01 -1.22324117e-01 1.76469088e-01 1.58698726e+00
3.56448829e-01 4.27281171e-01 -3.13891590e-01 3.55884694e-02
8.36561382e-01 -1.34855524e-01 5.62463403e-01 -1.71263742e+00
1.80859044e-02 -8.55041862e-01 -5.40074170e-01 -4.75928456e-01
6.12997055e-01 -1.05173147e+00 -7.17861831e-01 -1.51882935e+00
6.09068692e-01 -1.17690332e-01 3.30478013e-01 6.73296750e-02
1.62454277e-01 4.76675570e-01 1.73748448e-01 1.59877077e-01
-6.61665916e-01 -2.00193048e-01 7.20359027e-01 -2.77484685e-01
2.21974570e-02 -1.14044055e-01 -6.74914896e-01 9.67372477e-01
7.80222237e-01 -6.69344127e-01 9.43638384e-03 -1.29483595e-01
8.22156191e-01 -2.83636540e-01 3.02687556e-01 -6.79597557e-01
1.78478047e-01 -8.24370012e-02 -1.97358459e-01 -1.62393019e-01
1.95695505e-01 -5.10851026e-01 -2.61203080e-01 5.84846437e-01
-2.91370004e-01 6.77699596e-02 -2.06243619e-01 8.76639843e-01
-9.32299942e-02 -5.17307580e-01 7.00982928e-01 -5.94528973e-01
-5.14339395e-02 5.47667921e-01 -8.96700680e-01 2.92147249e-01
7.16440022e-01 5.66675374e-03 -3.95972610e-01 -9.88521457e-01
-7.75178909e-01 3.82325426e-02 6.34604275e-01 1.96182340e-01
1.96153477e-01 -8.59620512e-01 -7.78428495e-01 -3.45751315e-01
-9.02632028e-02 -6.88706815e-01 -3.61362785e-01 7.39479125e-01
-5.44233501e-01 -2.61848215e-02 1.71257421e-01 -1.65951833e-01
-1.53962135e+00 6.69218600e-01 6.29072249e-01 -6.96380615e-01
-3.16849321e-01 5.24935842e-01 -4.53284353e-01 -3.50296259e-01
7.07634464e-02 4.27596211e-01 -7.94314921e-01 5.00051320e-01
3.87840956e-01 5.74687183e-01 -1.44661725e-01 -1.13242650e+00
-3.94399166e-01 1.36367008e-02 -2.30325744e-01 2.65634395e-02
1.36213493e+00 -8.25315267e-02 -7.62695551e-01 5.71319103e-01
1.37812817e+00 7.95851469e-01 -2.02523261e-01 -2.82544523e-01
5.23619592e-01 -4.68407124e-01 -1.48130253e-01 -1.40962601e-01
-8.97970319e-01 8.76769543e-01 -3.91005576e-01 1.15585804e+00
2.16274172e-01 -5.76893501e-02 9.31910217e-01 2.93309540e-01
4.80842173e-01 -7.20362186e-01 2.06639275e-01 6.70971513e-01
5.75103402e-01 -1.02222276e+00 2.38844067e-01 -1.14997780e+00
-2.31196880e-01 1.07046580e+00 4.38900173e-01 -5.30479610e-01
1.03821862e+00 3.38945985e-02 -2.96281368e-01 -6.20063305e-01
-1.07231461e-01 -2.79882789e-01 -1.06508275e-02 5.17556667e-01
4.79750723e-01 6.66661412e-02 -3.61985177e-01 2.36259490e-01
-1.78850859e-01 -4.79505807e-01 1.14723265e+00 9.59355950e-01
-4.09274042e-01 -1.18359923e+00 7.70771354e-02 4.81227219e-01
-1.03658950e+00 -3.15113515e-01 -1.53810668e+00 9.03921485e-01
-8.07477757e-02 1.22894824e+00 -1.18135465e-02 -4.78369206e-01
-1.07578553e-01 9.56087187e-02 3.76722723e-01 -1.02076781e+00
-1.28415155e+00 -4.83997554e-01 6.83123887e-01 -3.68265212e-01
-4.05577034e-01 -6.70154274e-01 -8.72702479e-01 -1.07690978e+00
-2.88263023e-01 1.22993186e-01 3.44580770e-01 7.82653570e-01
1.73964947e-01 -8.63501355e-02 6.46434009e-01 -5.43394208e-01
-1.73831239e-01 -1.14752233e+00 -6.92558587e-01 8.21254492e-01
7.03368068e-01 -3.37915331e-01 -6.11807287e-01 -5.13075411e-01] | [8.446728706359863, 10.339388847351074] |
c9b4050b-471b-490d-8e7a-975b7695cb2b | playing-fps-games-with-deep-reinforcement | 1609.05521 | null | http://arxiv.org/abs/1609.05521v2 | http://arxiv.org/pdf/1609.05521v2.pdf | Playing FPS Games with Deep Reinforcement Learning | Advances in deep reinforcement learning have allowed autonomous agents to
perform well on Atari games, often outperforming humans, using only raw pixels
to make their decisions. However, most of these games take place in 2D
environments that are fully observable to the agent. In this paper, we present
the first architecture to tackle 3D environments in first-person shooter games,
that involve partially observable states. Typically, deep reinforcement
learning methods only utilize visual input for training. We present a method to
augment these models to exploit game feature information such as the presence
of enemies or items, during the training phase. Our model is trained to
simultaneously learn these features along with minimizing a Q-learning
objective, which is shown to dramatically improve the training speed and
performance of our agent. Our architecture is also modularized to allow
different models to be independently trained for different phases of the game.
We show that the proposed architecture substantially outperforms built-in AI
agents of the game as well as humans in deathmatch scenarios. | ['Devendra Singh Chaplot', 'Guillaume Lample'] | 2016-09-18 | null | null | null | null | ['game-of-doom', 'fps-games'] | ['playing-games', 'playing-games'] | [-7.59050250e-02 1.01403192e-01 3.00061852e-01 8.67447928e-02
-2.64402926e-01 -7.37565398e-01 6.19940758e-01 6.90350356e-03
-9.69269872e-01 7.09025264e-01 -4.23313886e-01 -1.30380571e-01
1.91997856e-01 -8.33963573e-01 -7.47692645e-01 -5.82846820e-01
-4.43466723e-01 7.97425807e-01 2.75345951e-01 -7.07767904e-01
1.73580587e-01 3.57899368e-01 -1.43829441e+00 3.52395922e-02
4.91779655e-01 8.01477313e-01 4.64639455e-01 1.02034676e+00
4.48996782e-01 1.31356645e+00 -8.71883392e-01 1.17915841e-02
7.64705658e-01 -2.92040557e-01 -4.58741874e-01 2.80012786e-01
-6.03212183e-03 -8.53012443e-01 -6.04522824e-01 7.56453276e-01
5.28251946e-01 3.12937856e-01 4.10766333e-01 -1.11022735e+00
-2.41858616e-01 3.43895644e-01 -4.71894324e-01 1.37663901e-01
2.19432816e-01 9.18062270e-01 9.10058022e-01 -3.66008252e-01
4.88023758e-01 1.05631721e+00 3.46805513e-01 6.91596210e-01
-1.02404153e+00 -4.57775295e-01 3.67219687e-01 2.33227193e-01
-9.58446085e-01 -5.79652824e-02 5.81677198e-01 -3.18639189e-01
1.40575993e+00 -2.29527071e-01 1.14169049e+00 1.16796398e+00
1.40170574e-01 1.02617836e+00 1.24793410e+00 -3.26830059e-01
6.67123437e-01 -3.03611219e-01 -3.87540489e-01 9.69891071e-01
9.45111960e-02 6.77380264e-01 -4.32105571e-01 1.34621859e-01
1.13281906e+00 1.58746261e-02 2.24926502e-01 -7.50868917e-01
-8.86368871e-01 9.89777505e-01 6.92272305e-01 -3.94031703e-02
-7.81425118e-01 5.40491521e-01 1.29647762e-01 4.48291540e-01
-1.42623847e-02 8.34797800e-01 -2.14330450e-01 -3.03768009e-01
-4.76567507e-01 7.08774269e-01 6.29964650e-01 6.17264926e-01
6.87238574e-01 3.99084866e-01 2.64966022e-02 4.14076507e-01
7.41802230e-02 5.91476679e-01 2.27429986e-01 -1.09527254e+00
3.48614454e-01 6.21272922e-01 4.16305155e-01 -5.81233740e-01
-7.29094267e-01 -5.31921327e-01 -3.46521139e-01 1.17797804e+00
3.70753109e-01 -6.33741975e-01 -1.21303749e+00 1.74753129e+00
4.29274380e-01 8.94333348e-02 2.11151466e-01 1.15003538e+00
3.75662744e-01 4.30994034e-01 -3.94298509e-02 3.83612663e-01
1.09191465e+00 -1.14090264e+00 -2.37301350e-01 -9.24838245e-01
4.94171530e-01 8.75236932e-03 8.81338239e-01 6.15482092e-01
-1.24122822e+00 -4.10248727e-01 -1.33129108e+00 1.46629035e-01
-1.74665883e-01 -1.27710372e-01 7.55007744e-01 4.74888295e-01
-9.83472526e-01 6.58526361e-01 -1.16151047e+00 -2.56922036e-01
4.97173786e-01 7.81835735e-01 -3.27188373e-01 2.40866363e-01
-1.13130701e+00 1.29150331e+00 4.75157708e-01 7.65044689e-02
-1.79684186e+00 -9.96513441e-02 -9.69412744e-01 8.86335969e-02
8.68925035e-01 -7.15104878e-01 1.59697115e+00 -9.80735898e-01
-1.83761477e+00 7.89910674e-01 5.44493139e-01 -9.03662860e-01
6.11304104e-01 -3.38192791e-01 3.92601818e-01 1.50703743e-01
-5.44642881e-02 8.30402136e-01 7.42551267e-01 -1.13398027e+00
-7.45864928e-01 -5.22746146e-01 1.04510653e+00 8.56253266e-01
1.00686371e-01 -2.93164164e-01 -2.67677844e-01 -2.93576032e-01
-5.33553839e-01 -1.09040570e+00 -8.61953974e-01 2.07650531e-02
-5.11559695e-02 1.47417918e-01 5.26189625e-01 -2.45964661e-01
4.30573583e-01 -1.79892349e+00 4.32442397e-01 1.30056292e-02
3.31506401e-01 2.16890812e-01 -3.44898194e-01 4.57762182e-01
4.38017219e-01 -3.67635995e-01 -1.31243154e-01 -2.91667610e-01
1.62895575e-01 5.51856518e-01 4.30731848e-02 2.86743313e-01
2.65203685e-01 1.11230600e+00 -1.02640593e+00 3.59894037e-02
2.77489245e-01 8.91636163e-02 -7.35111296e-01 4.91226345e-01
-5.13246059e-01 5.63772023e-01 -5.36323607e-01 2.79084176e-01
2.62249380e-01 2.33792104e-02 3.92712682e-01 7.60164559e-01
-4.49149869e-02 2.22297743e-01 -1.02357054e+00 1.86073208e+00
-5.95908701e-01 3.45191091e-01 3.96924794e-01 -1.01460803e+00
6.88447118e-01 1.27149493e-01 4.32949573e-01 -8.90934885e-01
3.81282508e-01 -3.76150012e-02 3.66868019e-01 -2.50541657e-01
4.17739034e-01 -3.11273783e-01 -3.53428274e-01 4.73447263e-01
2.30450720e-01 -4.54691052e-01 4.59220558e-01 -9.30177495e-02
1.46495867e+00 4.28023368e-01 3.64451230e-01 1.09506160e-01
5.82424216e-02 5.15513241e-01 4.20119166e-01 1.30061460e+00
-2.64419317e-01 1.71017170e-01 5.78574419e-01 -6.25796139e-01
-1.07418275e+00 -1.03582108e+00 6.10400617e-01 1.09123182e+00
4.83509839e-01 -1.96819857e-01 -8.28363240e-01 -7.45527864e-01
-1.81872636e-01 5.81035852e-01 -8.36390495e-01 -2.72584200e-01
-6.84683263e-01 -5.05261719e-01 4.06904459e-01 6.64111435e-01
5.51711261e-01 -1.48750877e+00 -1.69875491e+00 2.85236865e-01
3.64705026e-01 -9.90347922e-01 -6.40900340e-03 7.16474950e-01
-5.82461059e-01 -1.15118980e+00 -6.08847797e-01 -5.25249243e-01
4.92570937e-01 1.22722164e-01 9.52354789e-01 2.62545109e-01
-5.27948916e-01 3.83161783e-01 -4.40103948e-01 -5.04831135e-01
-1.15631558e-01 -4.23736349e-02 1.81743041e-01 -5.13678670e-01
1.77597061e-01 -5.41176796e-01 -5.10306954e-01 1.02318510e-01
-7.71917760e-01 2.59187490e-01 7.90540278e-01 1.12872577e+00
2.05211729e-01 8.14550295e-02 1.35956453e-02 -6.05807543e-01
6.47973359e-01 -2.48071209e-01 -8.71196508e-01 -9.71447229e-02
-2.03367826e-02 2.32825741e-01 7.25718558e-01 -4.98648882e-01
-8.12858343e-01 1.95576727e-01 -4.13379446e-02 -4.23671365e-01
-2.96630502e-01 3.77625376e-01 1.05468102e-01 -1.78789362e-01
9.27496910e-01 4.70106453e-02 3.09379585e-02 -2.23862365e-01
1.95655808e-01 2.43635982e-01 2.96608776e-01 -5.37934840e-01
9.01878834e-01 3.73055220e-01 1.41731054e-01 -5.25364876e-01
-6.43126667e-01 -2.61800140e-01 -6.03566051e-01 -2.58740753e-01
8.98472250e-01 -9.60948527e-01 -1.21011460e+00 6.43785298e-01
-8.18004727e-01 -1.01852942e+00 -5.11629462e-01 4.20079231e-01
-1.00379050e+00 -7.97147453e-02 -6.35356605e-01 -9.61772025e-01
6.09744154e-02 -1.10962200e+00 8.33319664e-01 5.10326028e-01
2.08749995e-01 -7.33459294e-01 3.83668959e-01 2.16214746e-01
3.20603549e-01 2.08251730e-01 4.34432596e-01 -4.94197071e-01
-4.79902893e-01 -8.03938881e-02 2.50170380e-01 -1.15779795e-01
-1.83520988e-02 -6.29798174e-01 -7.68555284e-01 -4.98806626e-01
-1.50921166e-01 -9.03269649e-01 8.35117161e-01 2.85278410e-01
7.25501120e-01 -1.61113724e-01 -7.93190300e-02 3.92515272e-01
1.13036978e+00 4.50368971e-01 4.50113624e-01 6.62229121e-01
4.63598788e-01 4.94679928e-01 6.46450281e-01 7.38335907e-01
3.17536771e-01 7.85443604e-01 1.07665873e+00 -2.35856235e-01
1.36589333e-01 -3.26259822e-01 5.35810173e-01 -1.44175708e-01
-4.22806323e-01 -3.20154816e-01 -8.25790346e-01 3.12438309e-01
-2.06895781e+00 -1.13924539e+00 4.74982113e-01 1.99844348e+00
3.21414769e-01 3.78935337e-01 4.99992877e-01 -4.20626029e-02
2.90966898e-01 1.02317803e-01 -9.56690848e-01 -3.64881188e-01
5.47495969e-02 5.52634776e-01 5.50311923e-01 5.94981611e-01
-1.29492319e+00 1.34998763e+00 6.63790178e+00 4.26413655e-01
-9.39253449e-01 -1.35687515e-01 2.65669465e-01 -5.61503649e-01
1.95282593e-01 5.91041595e-02 -2.58075088e-01 -9.28280130e-02
6.44804716e-01 1.51325732e-01 8.60645115e-01 9.35938537e-01
2.54560798e-01 -3.91838104e-01 -1.04006112e+00 7.08261490e-01
-1.04365587e-01 -1.02737427e+00 -4.06858832e-01 4.48924690e-01
5.38185179e-01 1.17519870e-01 2.99253106e-01 7.48636901e-01
1.08094501e+00 -1.19355428e+00 8.72401893e-01 1.02182418e-01
2.44904965e-01 -8.64107668e-01 6.44219697e-01 7.69379258e-01
-8.25666726e-01 -5.31757712e-01 -3.75785083e-01 -8.62772346e-01
7.70614967e-02 -3.78877401e-01 -1.02775407e+00 2.59359688e-01
5.59386790e-01 4.20012563e-01 -4.35778677e-01 1.01302600e+00
-5.21534681e-01 1.44366041e-01 -4.25986648e-01 -3.68839592e-01
8.54396582e-01 -1.35661662e-01 4.69838321e-01 4.58231986e-01
1.43996477e-01 4.15802985e-01 7.11821318e-01 6.89686239e-01
2.20064312e-01 -3.06428015e-01 -6.80846155e-01 -2.97838356e-03
-9.31996293e-03 1.03504491e+00 -6.32868767e-01 -1.06103942e-01
-2.25200370e-01 1.19395947e+00 8.38105619e-01 2.09067762e-01
-8.80395710e-01 -1.51818722e-01 7.93463469e-01 1.60121545e-01
5.13691604e-01 -5.91251791e-01 -5.13425693e-02 -1.16487277e+00
-1.74501449e-01 -1.01461017e+00 2.54218400e-01 -6.67736292e-01
-9.27790880e-01 6.25737190e-01 -2.28185356e-01 -1.02533841e+00
-6.05327606e-01 -1.00047839e+00 -6.52142823e-01 5.19986808e-01
-1.33333015e+00 -1.07799911e+00 -1.41464949e-01 6.43883228e-01
5.85046470e-01 -3.72389764e-01 7.76366055e-01 -1.62409157e-01
-4.65402097e-01 2.24295884e-01 -1.51638880e-01 2.30835438e-01
1.57025516e-01 -1.51716423e+00 5.00705302e-01 8.34750652e-01
3.70396554e-01 1.49343476e-01 7.51724422e-01 -5.36825001e-01
-1.38585246e+00 -4.83326018e-01 -2.36855164e-01 -2.75827855e-01
5.74045956e-01 -5.99085510e-01 -2.71565527e-01 7.08562911e-01
3.03391457e-01 -1.41902164e-01 3.35424662e-01 2.46901497e-01
-5.65800108e-02 2.56441504e-01 -1.04480648e+00 8.47728968e-01
1.10780275e+00 -2.48224288e-01 -6.50915265e-01 1.07358620e-01
1.98651180e-01 -8.02171946e-01 -1.72896653e-01 1.56226251e-02
2.98139781e-01 -1.08297360e+00 9.24440145e-01 -1.03578198e+00
3.88025463e-01 -3.81571710e-01 7.55579993e-02 -1.59294260e+00
-4.54275101e-01 -5.39540172e-01 -1.95147157e-01 8.80126208e-02
3.12374979e-01 -2.49618068e-01 1.20710516e+00 4.67486709e-01
-4.72369231e-02 -4.88560855e-01 -9.80871260e-01 -6.09536231e-01
1.41551182e-01 -9.68687534e-02 2.59368181e-01 3.38434637e-01
1.85421124e-01 3.27491909e-01 -7.46648610e-01 3.19336355e-01
5.82176924e-01 1.59870729e-01 1.01713836e+00 -1.05704641e+00
-9.40961838e-01 -3.32703978e-01 -5.51476717e-01 -1.25759828e+00
2.12116875e-02 -4.92029130e-01 3.78894418e-01 -1.45664597e+00
1.30342364e-01 -3.69582206e-01 -1.52320713e-01 8.08758736e-01
-1.78722799e-01 2.73826510e-01 5.51731408e-01 -8.16702694e-02
-7.93977976e-01 6.78716838e-01 1.38123894e+00 -3.31017882e-01
-2.71589339e-01 2.44275667e-02 -4.61978287e-01 8.09488654e-01
9.55459952e-01 -3.02349716e-01 -5.06741822e-01 -6.98472261e-01
2.40166530e-01 2.02704281e-01 6.40994251e-01 -1.34265316e+00
2.39553347e-01 -4.91709858e-01 6.59701526e-01 -6.01418577e-02
8.70396376e-01 -8.07790756e-01 -1.51013374e-01 8.57900918e-01
-2.33443335e-01 1.32261440e-01 2.83172876e-01 5.43323934e-01
1.60541739e-02 -4.06126142e-01 6.05228841e-01 -6.51355028e-01
-9.28070843e-01 3.55002373e-01 -8.53060842e-01 3.15476917e-02
1.43999803e+00 -2.09741905e-01 -1.00267507e-01 -6.31555557e-01
-9.12934721e-01 5.69219589e-01 5.75281739e-01 1.45760745e-01
5.55000365e-01 -9.40329909e-01 -6.50375664e-01 -5.57902921e-03
-1.77702054e-01 1.32990461e-02 4.68403220e-01 2.74611443e-01
-9.12830949e-01 1.18994214e-01 -8.61226857e-01 -2.64147699e-01
-9.93686557e-01 7.61135101e-01 7.21484900e-01 -7.65557110e-01
-6.64612293e-01 7.83819258e-01 3.91859412e-01 -5.80489695e-01
1.92338780e-01 2.89409794e-03 -2.42561221e-01 -4.14722145e-01
4.71112102e-01 -1.13248833e-01 -8.60875621e-02 -2.64930487e-01
-2.18910292e-01 2.22046793e-01 -1.79390371e-01 -5.71550310e-01
1.48950291e+00 2.00101763e-01 6.05159402e-01 5.01154251e-02
2.68713355e-01 -4.00296748e-01 -2.04412866e+00 -1.11837253e-01
-3.03492963e-01 -4.47295636e-01 1.41962796e-01 -9.51116920e-01
-9.94459629e-01 1.00910044e+00 5.10031939e-01 1.43345837e-02
9.51176763e-01 -2.89338589e-01 5.86302578e-01 8.27371597e-01
9.13221598e-01 -1.21105587e+00 5.57175994e-01 8.02527189e-01
5.47672212e-01 -1.34946442e+00 -1.82059631e-01 2.88285047e-01
-1.01767755e+00 9.28391337e-01 1.03376055e+00 -7.14857399e-01
6.84354231e-02 4.50699151e-01 1.58063978e-01 -3.44691783e-01
-1.02237093e+00 -7.61200726e-01 -1.24246158e-01 9.28864539e-01
-2.23752365e-01 2.18471363e-02 2.71270007e-01 2.85049528e-01
-3.54103506e-01 -3.27703357e-01 6.69939160e-01 1.33480203e+00
-6.67628288e-01 -1.10665321e+00 -2.83484280e-01 1.22417979e-01
-2.63996512e-01 1.55749127e-01 -5.91901779e-01 8.45792234e-01
1.37929603e-01 1.03188086e+00 1.35625020e-01 -2.37649471e-01
4.09502625e-01 -4.11068261e-01 1.01990044e+00 -8.08721125e-01
-9.58514631e-01 -2.20088333e-01 -2.01815944e-02 -6.82348609e-01
-3.48418728e-02 -4.11364645e-01 -1.21912086e+00 -2.55458534e-01
-2.54238229e-02 5.37629724e-02 4.05815393e-01 9.12376761e-01
2.28319373e-02 6.07276976e-01 5.18451214e-01 -1.23323989e+00
-5.71735919e-01 -6.51752532e-01 -6.49094760e-01 2.10829735e-01
3.54912370e-01 -9.87489939e-01 7.09195286e-02 -4.55782086e-01] | [3.8398807048797607, 1.3932883739471436] |
651b322f-a917-46b8-afea-c1882c6f42c7 | closing-the-loop-fast-interactive-semi | null | null | https://aclanthology.org/D11-1136/ | https://aclanthology.org/D11-1136.pdf | Closing the Loop: Fast, Interactive Semi-Supervised Annotation With Queries on Features and Instances | This paper describes DUALIST, an active learning annotation paradigm which solicits and learns from labels on both features (e.g., words) and instances (e.g., documents). We present a novel semi-supervised training algorithm developed for this setting, which is (1) fast enough to support real-time interactive speeds, and (2) at least as accurate as preexisting methods for learning with mixed feature and instance labels. Human annotators in user studies were able to produce near-state-of-the-art classifiers—on several corpora in a variety of application domains—with only a few minutes of effort. | ['Burr Settles'] | 2011-07-01 | null | null | null | proceedings-of-the-2011-conference-on | ['information-extraction'] | ['natural-language-processing'] | [ 2.46623561e-01 5.70615232e-01 -6.40339136e-01 -6.56011462e-01
-1.32063043e+00 -1.01014841e+00 8.77936959e-01 6.76782072e-01
-5.83669543e-01 8.83472204e-01 -1.73489437e-01 -3.85227472e-01
1.35515286e-02 -3.58456552e-01 -3.76914740e-01 -4.38014597e-01
-3.07403803e-01 9.30411935e-01 2.01292634e-01 2.29685791e-02
2.87679672e-01 2.95270205e-01 -1.75548315e+00 5.31484008e-01
6.61735237e-01 7.43349969e-01 1.84074670e-01 5.78920841e-01
-5.67603767e-01 1.19237876e+00 -5.98999202e-01 -4.47685063e-01
7.48099759e-02 1.03010893e-01 -1.41813076e+00 2.74838388e-01
6.36946440e-01 -2.58447304e-02 -1.68664642e-02 5.80534041e-01
2.67351270e-01 2.52318591e-01 6.25908911e-01 -1.25219083e+00
-5.71013331e-01 7.61309803e-01 -2.34604493e-01 2.11853549e-01
5.63955963e-01 8.01496729e-02 1.32984471e+00 -1.17444611e+00
6.42770648e-01 8.95665824e-01 5.68736494e-01 6.27823114e-01
-1.41798079e+00 -7.49569535e-01 2.80485451e-01 -8.03352669e-02
-1.25854480e+00 -9.30080056e-01 3.39557886e-01 -5.34513533e-01
1.11643934e+00 3.51203322e-01 6.16484106e-01 1.04881334e+00
-2.16015533e-01 1.30757117e+00 1.15109134e+00 -1.07352757e+00
2.25783423e-01 8.85838687e-01 5.77152252e-01 9.85693574e-01
1.34126917e-01 -3.42480332e-01 -7.90365458e-01 -6.89643025e-01
4.82307434e-01 -3.06020379e-01 4.98763621e-02 -3.24498236e-01
-1.00011742e+00 8.71445894e-01 -1.22420974e-01 4.10316229e-01
2.99142965e-04 -2.09558845e-01 5.63438118e-01 4.40268725e-01
1.01971102e+00 8.76139224e-01 -9.44459379e-01 -2.74468422e-01
-6.98157370e-01 4.57130314e-04 1.22654724e+00 1.36398065e+00
7.51086414e-01 -3.15346003e-01 1.06495693e-01 9.34792817e-01
3.73948097e-01 1.01605698e-01 3.96418631e-01 -7.36988902e-01
5.16647637e-01 7.53118038e-01 5.29835224e-01 -2.77354598e-01
-3.28435212e-01 -2.67223835e-01 -2.98990220e-01 4.39453006e-01
4.64244217e-01 -1.76174387e-01 -8.35966468e-01 1.38692975e+00
2.35693425e-01 -8.00510868e-02 -4.40346561e-02 3.65078419e-01
1.20224774e+00 4.04834241e-01 4.29078609e-01 -6.92555666e-01
1.24620175e+00 -1.02312148e+00 -7.29553699e-01 -4.21218753e-01
1.11865354e+00 -8.44489753e-01 1.17114508e+00 5.48423827e-01
-1.13067412e+00 -5.06352544e-01 -9.38179970e-01 4.14438695e-02
-6.46566629e-01 5.09622544e-02 1.00238574e+00 7.84850478e-01
-9.92396355e-01 5.21520615e-01 -8.13587189e-01 -4.33973312e-01
2.89546132e-01 5.60283422e-01 -4.93625313e-01 2.35386923e-01
-1.10209060e+00 1.08586836e+00 4.51481521e-01 -3.79173815e-01
-6.92776024e-01 -3.45291167e-01 -8.04022014e-01 -1.49793431e-01
5.16231477e-01 -6.72469288e-02 1.79101205e+00 -1.35660720e+00
-1.05608797e+00 1.29050922e+00 -1.68487400e-01 -7.76676461e-02
1.81754410e-01 -4.30873603e-01 -5.65029263e-01 -1.67494178e-01
1.32164866e-01 4.42474991e-01 5.03598511e-01 -1.36316907e+00
-9.03198898e-01 -2.14980319e-01 3.51441890e-01 6.89036846e-01
-9.21282709e-01 3.86153936e-01 -8.55657980e-02 -4.16731238e-01
-6.07978813e-02 -7.97163725e-01 -1.88833013e-01 1.12297118e-01
-1.28958404e-01 -7.70351171e-01 9.83948231e-01 -4.95284110e-01
1.05078769e+00 -2.20706415e+00 -2.45040134e-01 8.99769738e-02
4.18445289e-01 4.40509528e-01 8.18735957e-02 5.17557263e-01
-1.64949521e-01 3.12271476e-01 -6.57202527e-02 -7.07512498e-01
-1.26000077e-01 2.35221148e-01 -2.29603961e-01 4.58212823e-01
3.49333659e-02 6.14073873e-01 -1.32052028e+00 -8.29327047e-01
-8.56012106e-02 -5.91416564e-03 5.84728532e-02 5.46493828e-01
-2.20322341e-01 3.97454828e-01 -1.66121006e-01 5.68950415e-01
-2.88334582e-02 -4.72916335e-01 2.24683344e-01 3.07376653e-01
-2.88581103e-01 5.69896102e-01 -1.25207615e+00 1.60704422e+00
-7.05328465e-01 9.27402973e-01 1.26605362e-01 -1.05331957e+00
9.69711363e-01 1.01681912e+00 5.56713402e-01 -2.02364013e-01
-1.89355180e-01 5.08387506e-01 -3.48989874e-01 -5.87230742e-01
3.95693988e-01 2.52985835e-01 -1.39469132e-01 1.03486347e+00
6.27892733e-01 3.16365361e-02 4.45442051e-01 4.08804744e-01
1.05484807e+00 5.56268990e-02 8.37258518e-01 -3.87941092e-01
2.89873570e-01 2.51707733e-01 2.61490136e-01 8.96369874e-01
-7.22111240e-02 1.33258447e-01 1.92362964e-01 -6.25974476e-01
-9.72746551e-01 -6.37823701e-01 -1.88488826e-01 1.74040389e+00
-2.27585938e-02 -6.44445539e-01 -4.87417489e-01 -1.16363966e+00
-2.25565881e-01 7.72391677e-01 -5.85076690e-01 3.15484315e-01
-3.94038260e-01 -5.76274097e-01 3.85819525e-01 7.13308513e-01
1.24880388e-01 -1.16983271e+00 -3.44452441e-01 2.88107783e-01
1.16152383e-01 -8.84075522e-01 -1.79285616e-01 1.00384986e+00
-7.42328584e-01 -1.17247951e+00 -1.18323222e-01 -1.16241848e+00
7.87494481e-01 -1.29170511e-02 1.56538212e+00 2.91602284e-01
-1.65767401e-01 4.79817003e-01 -7.41988540e-01 -8.15255880e-01
-4.08044368e-01 2.84557372e-01 -7.14038536e-02 -2.61334091e-01
5.27303398e-01 -6.57561868e-02 3.87520306e-02 3.41765881e-01
-7.32151151e-01 -2.78474521e-02 3.71334344e-01 9.57253158e-01
2.23781273e-01 5.45326173e-02 3.86067808e-01 -1.66630638e+00
6.09388173e-01 -5.13263047e-01 -4.33484495e-01 5.66525042e-01
-9.21686053e-01 -2.02679560e-01 5.24350464e-01 -6.07177615e-01
-1.11116552e+00 5.56417048e-01 8.67531300e-02 2.91763872e-01
-4.42212611e-01 4.32252347e-01 7.76797766e-03 -2.04154834e-01
1.14881051e+00 -1.84748814e-01 -3.07904661e-01 -4.70533013e-01
3.45259875e-01 1.30111551e+00 1.72082409e-01 -4.50026363e-01
6.21804714e-01 9.31768045e-02 -6.15852773e-01 -5.59512854e-01
-1.36909103e+00 -1.01929569e+00 -1.16138697e+00 -2.05471307e-01
2.70219147e-01 -9.22114909e-01 -8.34181085e-02 2.51244634e-01
-1.11300397e+00 -5.43437600e-01 -3.61296594e-01 3.84774029e-01
-6.00305498e-01 -4.18828987e-03 -5.23843408e-01 -1.08131289e+00
-4.55172539e-01 -7.81875372e-01 8.05230618e-01 2.01323956e-01
-8.64566147e-01 -1.39466453e+00 1.87577665e-01 3.83101702e-01
1.42390221e-01 1.48187773e-02 7.27558851e-01 -1.45330667e+00
-7.86608434e-05 -4.26159680e-01 5.20269722e-02 5.52352630e-02
2.11113870e-01 6.68466017e-02 -1.23910880e+00 -4.93128479e-01
-1.13828599e-01 -8.98429930e-01 4.20076549e-01 -3.15270394e-01
1.07571089e+00 -4.31954771e-01 -6.56159163e-01 -2.06672758e-01
1.09065115e+00 3.66811395e-01 2.12217286e-01 2.44853020e-01
4.80776370e-01 5.39813817e-01 9.08578515e-01 3.04042608e-01
-9.79204625e-02 8.59046161e-01 -1.97756570e-02 -4.81209636e-01
1.04540333e-01 6.19967505e-02 3.14901508e-02 5.44732928e-01
-6.17263727e-02 -3.62736702e-01 -1.33425927e+00 6.42112195e-01
-2.13434887e+00 -8.08100462e-01 -2.83706989e-02 2.16869283e+00
1.39912665e+00 5.45698822e-01 1.69812351e-01 3.86284500e-01
7.53892601e-01 -1.49313197e-01 -3.65674585e-01 -3.87174875e-01
1.49589509e-01 4.28932428e-01 3.39630604e-01 4.30541635e-01
-1.49635816e+00 8.47369194e-01 7.26257801e+00 8.01934123e-01
-7.25533903e-01 5.43879747e-01 5.51921546e-01 1.85425758e-01
8.13415051e-02 1.09870233e-01 -8.60786021e-01 2.62811780e-01
9.99677956e-01 -1.01232141e-01 1.75632417e-01 1.24131203e+00
-3.51165622e-01 -1.16472915e-01 -1.33500218e+00 7.76848733e-01
1.73819646e-01 -1.23967469e+00 -4.48289484e-01 -2.02770859e-01
6.86390400e-01 -6.90480247e-02 -2.58220464e-01 4.27643836e-01
7.24525273e-01 -9.31330323e-01 9.62231815e-01 1.21072449e-01
1.11046612e+00 -5.34066498e-01 9.57501352e-01 8.16224992e-01
-1.03264570e+00 -3.10930699e-01 5.77961206e-02 -1.67680800e-01
2.99061742e-02 3.72727185e-01 -1.20081794e+00 2.45010674e-01
6.62960052e-01 5.78500628e-01 -6.43328607e-01 1.14369047e+00
-3.17155421e-01 9.96790767e-01 -2.07574591e-01 -2.83938557e-01
1.69585735e-01 4.33634758e-01 1.96411982e-01 1.58787644e+00
-3.69161308e-01 2.42274687e-01 6.21048450e-01 1.35582149e-01
4.07656282e-02 5.14417350e-01 -7.65572667e-01 -2.88308352e-01
7.90131688e-01 1.47188950e+00 -1.16283584e+00 -6.60698771e-01
-5.28128743e-01 6.53471172e-01 7.62840688e-01 1.87625542e-01
-2.56984591e-01 -6.20239973e-01 -6.62856968e-03 1.83559731e-01
-4.78162616e-02 -2.70544082e-01 -4.61621910e-01 -9.22180295e-01
-2.46644989e-01 -8.17545474e-01 4.96729493e-01 -4.79212284e-01
-1.43049598e+00 7.90446341e-01 1.31233007e-01 -1.23675931e+00
-4.67606187e-01 -6.57038689e-01 -4.21266407e-01 7.26392686e-01
-7.80965924e-01 -1.13829780e+00 -2.61801124e-01 4.87055302e-01
6.15924060e-01 -5.65586925e-01 1.66406369e+00 3.08415830e-01
-2.35829577e-01 6.55971348e-01 9.43906158e-02 4.75149900e-01
8.96905005e-01 -1.63565826e+00 7.67868042e-01 2.62950271e-01
8.14923704e-01 7.98656881e-01 5.56858003e-01 -4.68913406e-01
-1.01700461e+00 -7.30598509e-01 1.31092644e+00 -9.36569571e-01
5.32799184e-01 -7.78828800e-01 -7.93166637e-01 8.94201517e-01
2.52775937e-01 1.29827827e-01 1.13916814e+00 8.03626597e-01
-4.51702148e-01 1.54425815e-01 -1.27343893e+00 1.25296816e-01
1.01429570e+00 -8.34163845e-01 -5.33487201e-01 9.17632818e-01
4.36856598e-01 -6.19084179e-01 -6.16782010e-01 2.84428746e-01
4.70304787e-01 -4.17963564e-01 4.65521008e-01 -9.24920559e-01
-8.39336663e-02 1.73147500e-01 9.58516151e-02 -1.14232957e+00
-5.02455831e-01 -8.05771351e-01 -4.90606934e-01 1.35562527e+00
1.04997003e+00 -5.26308835e-01 5.92209518e-01 7.67301381e-01
-1.90558851e-01 -9.01115954e-01 -6.29169822e-01 -7.43885398e-01
-4.32568103e-01 -3.96842718e-01 9.39514786e-02 1.35055172e+00
6.59985960e-01 7.51905560e-01 -7.55093321e-02 -4.92553949e-01
4.46239412e-01 -1.29693210e-01 4.26863700e-01 -1.70130908e+00
-8.25844258e-02 -4.81949449e-02 -2.62012273e-01 -9.47605550e-01
3.56646419e-01 -6.89216495e-01 2.36249760e-01 -1.49658060e+00
3.08549494e-01 -1.03115737e+00 -2.90497273e-01 1.15728056e+00
-8.23625177e-02 5.20776272e-01 -3.64197046e-01 5.42869627e-01
-9.36815500e-01 -2.44791284e-01 5.99760115e-01 -1.98494792e-01
-2.07066730e-01 2.06235692e-01 -7.17082322e-01 6.98208749e-01
7.29099572e-01 -6.85968995e-01 -5.94776154e-01 -2.68411517e-01
3.03432703e-01 -2.14015007e-01 -1.81970760e-01 -6.44943595e-01
4.33990926e-01 -2.11046413e-01 3.55430931e-01 -2.72078842e-01
2.83678144e-01 -7.45221913e-01 -9.26446840e-02 -2.14929804e-02
-9.97752786e-01 -1.70817927e-01 8.81393924e-02 3.28128189e-01
-1.04887031e-01 -7.82875836e-01 6.75742686e-01 -4.58191484e-01
-9.52806950e-01 1.28451347e-01 -4.54750687e-01 8.12201574e-02
9.96671379e-01 -2.04779848e-01 -2.54244864e-01 -1.78394556e-01
-7.49655664e-01 -3.00844330e-02 3.52938354e-01 4.30528343e-01
1.09583758e-01 -1.10505378e+00 -6.77811027e-01 8.13577697e-03
4.89024043e-01 -5.61867055e-05 -5.19775152e-01 4.10806239e-01
-3.46473068e-01 5.24170756e-01 1.43597767e-01 -5.39462149e-01
-1.65284896e+00 2.71569341e-01 1.07513927e-02 -2.70403445e-01
-5.04253566e-01 9.80097055e-01 -3.40796143e-01 -5.23464084e-01
6.32585943e-01 5.34694612e-01 -5.38392603e-01 4.10107523e-01
5.86158335e-01 9.85581428e-02 4.53754604e-01 -4.94500071e-01
-2.87118822e-01 -1.00912169e-01 -4.82299864e-01 -3.31119567e-01
1.24857819e+00 3.33941504e-02 6.41753823e-02 1.01447451e+00
9.70864713e-01 -1.05054364e-01 -8.53550673e-01 -6.45041287e-01
4.23989564e-01 -4.48087573e-01 1.67925045e-01 -9.29391205e-01
-5.48145294e-01 3.05036306e-01 5.56741357e-01 7.78303802e-01
6.49787366e-01 3.18046927e-01 3.58004212e-01 8.81144524e-01
6.76183164e-01 -1.46252716e+00 2.08859131e-01 3.16074759e-01
5.51410913e-01 -1.44654584e+00 3.72716457e-01 -5.94621718e-01
-6.22344971e-01 1.19555581e+00 6.19722486e-01 3.83521348e-01
6.95505321e-01 6.76820457e-01 5.14487922e-01 -4.89308715e-01
-1.15003371e+00 -1.74211208e-02 4.00217950e-01 4.60305572e-01
1.21059966e+00 2.10753344e-02 -2.62087524e-01 2.39816889e-01
1.39899999e-01 -2.65853643e-01 1.96367845e-01 1.61063099e+00
-6.73702002e-01 -1.08427930e+00 -1.17230922e-01 6.44996405e-01
-3.85313362e-01 2.99695749e-02 -7.57940650e-01 8.60669434e-01
3.49555045e-01 1.21898186e+00 1.37729332e-01 -2.22040460e-01
2.78186947e-01 3.91331851e-01 4.50133920e-01 -1.42670429e+00
-9.74561214e-01 -2.52464920e-01 6.98421478e-01 -1.67489842e-01
-7.59842694e-01 -7.88771927e-01 -1.14650810e+00 1.41033322e-01
-9.29911137e-01 6.39949262e-01 6.22936308e-01 1.20224428e+00
-1.32089972e-01 -2.01548159e-01 7.43537009e-01 -6.92747712e-01
-3.08171868e-01 -1.40527046e+00 -7.62142897e-01 5.78890741e-01
-1.96788311e-02 -6.89304054e-01 -5.98525286e-01 3.62638503e-01] | [9.665583610534668, 4.468758583068848] |
a3d39a74-ba08-4897-88a1-b1ddef5b1004 | arb-sen-at-semeval-2018-task1-a-new-set-of | null | null | https://aclanthology.org/S18-1055 | https://aclanthology.org/S18-1055.pdf | ARB-SEN at SemEval-2018 Task1: A New Set of Features for Enhancing the Sentiment Intensity Prediction in Arabic Tweets | This article describes our proposed Arabic Sentiment Analysis system named ARB-SEN. This system is designed for the International Workshop on Semantic Evaluation 2018 (SemEval-2018), Task1: Affect in Tweets. ARB-SEN proposes two supervised models to estimate the sentiment intensity in Arabic tweets. Both models use a set of features including sentiment lexicon, negation, word embedding and emotion symbols features. Our system combines these features to assist the sentiment analysis task. ARB-SEN system achieves a correlation score of 0.720, ranking 6th among all participants in the valence intensity regression (V-reg) for the Arabic sub-task organized within the SemEval 2018 evaluation campaign. | ['El Moatez Billah Nagoudi'] | 2018-06-01 | null | null | null | semeval-2018-6 | ['arabic-sentiment-analysis'] | ['natural-language-processing'] | [-3.79553348e-01 1.15290940e-01 -1.63245544e-01 -9.25102711e-01
-6.41400099e-01 -6.68677509e-01 5.73750973e-01 6.20665252e-01
-7.61811554e-01 5.78761935e-01 5.67314208e-01 3.68533373e-01
4.07287419e-01 -6.48271203e-01 2.22276594e-03 -1.09726191e-01
8.76263976e-02 4.91458744e-01 -4.94089842e-01 -1.08412075e+00
5.23625910e-01 -5.64854816e-02 -1.19004154e+00 9.07692730e-01
7.79534876e-01 1.25357258e+00 -6.95111573e-01 6.14259064e-01
-2.58813113e-01 1.07345319e+00 -7.06774056e-01 -9.67561841e-01
-2.60113984e-01 -2.95664996e-01 -1.02772355e+00 -3.03890437e-01
-5.92952780e-02 1.53294727e-01 3.97453517e-01 9.64502037e-01
5.29143751e-01 5.86773120e-02 8.85894358e-01 -1.39610100e+00
-6.08778656e-01 1.09876418e+00 -5.66785276e-01 -3.25989537e-02
6.14324629e-01 -5.89885890e-01 1.30446613e+00 -1.36259413e+00
5.30460060e-01 1.35377753e+00 6.68512404e-01 4.37024921e-01
-5.49132884e-01 -6.66277170e-01 -1.21000864e-01 2.70485073e-01
-9.43688273e-01 -1.06054626e-01 7.47330070e-01 -2.58212805e-01
1.16529632e+00 2.15626568e-01 6.48969650e-01 8.29743207e-01
5.97346723e-02 1.01560521e+00 1.64274764e+00 -5.29782176e-01
2.33890116e-01 6.87043250e-01 8.74880672e-01 4.42922950e-01
-2.30552748e-01 -7.13906407e-01 -9.77780879e-01 7.17811510e-02
-3.80160689e-01 -7.69164622e-01 2.59600669e-01 6.94185793e-01
-8.32452178e-01 1.29858720e+00 5.13376415e-01 1.53713167e-01
-4.05103385e-01 -1.99372470e-01 9.83700037e-01 6.65903628e-01
1.12686658e+00 6.16494894e-01 -8.72685015e-01 -3.50302279e-01
-4.21989888e-01 1.18913494e-01 9.31803942e-01 4.78597373e-01
4.47364479e-01 -1.46598578e-03 -1.51559627e-02 1.15992463e+00
5.50951064e-01 7.61889517e-01 6.48593068e-01 -4.61838424e-01
1.86273977e-01 8.39461923e-01 1.66792452e-01 -1.22696126e+00
-8.78112674e-01 1.17895834e-01 -2.18934476e-01 -2.24169288e-02
5.83541319e-02 -7.24644303e-01 -7.11602390e-01 1.48274446e+00
2.37485230e-01 -7.84094393e-01 3.25462967e-01 9.63978767e-01
1.36488152e+00 9.78249013e-01 5.58074057e-01 -2.40422323e-01
1.65629256e+00 -1.04119027e+00 -1.10052490e+00 -4.11650151e-01
8.63635659e-01 -1.13271844e+00 1.32517946e+00 6.37982726e-01
-1.02750301e+00 -2.34220162e-01 -1.31596589e+00 3.33508626e-02
-1.09178627e+00 2.18707502e-01 6.58801019e-01 9.22908604e-01
-1.08603930e+00 -3.97621430e-02 -1.63094714e-01 -3.80643427e-01
1.01480857e-01 5.87704480e-01 -4.17292565e-01 3.31220925e-01
-1.77373195e+00 1.36407447e+00 2.06528038e-01 2.85522453e-02
-2.49885663e-01 -2.79223055e-01 -1.10682797e+00 -4.26363915e-01
-4.34204519e-01 1.33044556e-01 1.16839421e+00 -1.76514769e+00
-1.74425113e+00 1.27630126e+00 -3.69058758e-01 -1.59588486e-01
-4.40078795e-01 -6.03217781e-01 -7.36529648e-01 -2.71360110e-02
1.38809115e-01 7.56311774e-01 5.81467748e-01 -9.95507002e-01
-2.60790080e-01 -5.02008677e-01 3.88623849e-02 4.83364344e-01
-7.79026926e-01 7.16842592e-01 -2.90075634e-02 -6.20948195e-01
-2.52268910e-01 -8.69216919e-01 7.85761699e-03 -1.03892720e+00
-1.11534588e-01 -6.10307395e-01 8.97960186e-01 -6.18757069e-01
1.29082060e+00 -1.73963916e+00 2.91211382e-02 5.48439980e-01
-2.44574219e-01 9.79515091e-02 -4.23407793e-01 3.72857869e-01
-4.19232935e-01 6.84585869e-02 4.78252918e-02 -4.22263592e-01
1.29712805e-01 -2.79152185e-01 -2.53841162e-01 2.73979396e-01
4.79874730e-01 8.89918745e-01 -9.25185382e-01 -3.36792558e-01
-8.00229833e-02 4.51226175e-01 -3.14079762e-01 5.78097813e-02
-2.50297278e-01 1.45327849e-02 -3.62648398e-01 8.48995566e-01
6.65656626e-01 1.43571541e-01 1.64688930e-01 -4.90772784e-01
-6.95949048e-02 4.33553249e-01 -6.16805971e-01 1.29454744e+00
-5.91046751e-01 5.75411737e-01 -1.42695963e-01 -6.08795226e-01
1.15973556e+00 2.54074723e-01 3.45493853e-01 -8.36301386e-01
7.81958580e-01 2.55558580e-01 -3.13572437e-01 -3.17236394e-01
8.66202772e-01 -4.36585873e-01 -7.36014485e-01 8.55578005e-01
1.54634044e-01 -4.94700521e-01 4.69150841e-01 5.44979453e-01
5.54128766e-01 -9.38107148e-02 3.36344838e-01 -5.28073967e-01
9.05654490e-01 1.40584767e-01 -8.49654991e-03 -1.42354472e-02
-5.71402431e-01 1.22399949e-01 7.74585187e-01 -4.52716649e-01
-6.07201815e-01 -5.54394126e-01 -1.22896865e-01 1.73026681e+00
-3.46794352e-02 -9.25273180e-01 -1.03456724e+00 -9.91858482e-01
-3.98137361e-01 7.97581971e-01 -1.08792877e+00 -1.33577570e-01
-8.79210755e-02 -1.11699247e+00 5.82988322e-01 3.16383362e-01
3.02314043e-01 -1.47984111e+00 -1.93966806e-01 -3.88704129e-02
-6.70636892e-01 -1.09753823e+00 -9.59390551e-02 2.90639102e-01
-2.29887143e-01 -7.47759223e-01 -1.19108409e-01 -8.96120727e-01
3.98955017e-01 -4.85159367e-01 1.33949161e+00 -2.49774709e-01
7.87458122e-02 4.58763599e-01 -9.85501826e-01 -1.21152020e+00
-2.60330111e-01 3.57972458e-02 -1.84107274e-02 -3.25341076e-02
1.15236890e+00 1.94965675e-01 -4.05226350e-01 3.82782929e-02
-6.69075072e-01 -3.89768481e-01 1.78375602e-01 6.15755200e-01
2.87684768e-01 -4.74733114e-01 1.25813830e+00 -8.58650804e-01
1.05829620e+00 -7.50285864e-01 -5.31858616e-02 9.21238586e-03
-6.50165856e-01 -4.00078982e-01 3.89576286e-01 5.51264919e-02
-1.03755009e+00 -1.03629678e-01 -5.86530089e-01 7.74410903e-01
2.37888917e-01 1.07178140e+00 5.37496759e-03 2.40335166e-01
6.85779631e-01 -5.22711694e-01 -7.22261444e-02 2.67955571e-01
4.84608769e-01 1.24143076e+00 -1.63947195e-02 -2.11128935e-01
-5.05851991e-02 2.39792421e-01 -3.22165132e-01 -7.53766775e-01
-1.71724474e+00 -4.76813197e-01 -3.86977702e-01 -6.92981720e-01
1.04522419e+00 -1.24551153e+00 -7.17741311e-01 7.60913849e-01
-1.19917130e+00 -2.05140814e-01 3.70111838e-02 2.20507547e-01
-2.10001886e-01 -1.66060865e-01 -8.28656316e-01 -7.44877994e-01
-9.73910809e-01 -9.54929411e-01 1.03223252e+00 1.13371804e-01
-1.03334272e+00 -1.11558950e+00 4.13057029e-01 5.48434854e-01
4.60849404e-01 3.58936369e-01 5.80848455e-01 -9.07007575e-01
1.06957924e+00 -4.52949852e-01 -1.96297050e-01 6.34730875e-01
-2.01245081e-02 5.21921590e-02 -1.25604558e+00 2.01232389e-01
6.78861747e-03 -1.27134073e+00 6.93487108e-01 2.73063511e-01
8.38391006e-01 -1.65972292e-01 3.98959637e-01 -3.71359825e-01
1.12541962e+00 -9.80734453e-02 7.10838735e-01 6.11061931e-01
3.18556726e-01 8.29741120e-01 1.28379238e+00 7.17555702e-01
7.42210567e-01 3.39782268e-01 5.34353554e-01 1.31875360e-02
4.08154309e-01 3.63277197e-01 1.09111011e+00 1.24215937e+00
7.32884556e-02 -2.88507998e-01 -9.97791469e-01 5.81851721e-01
-1.61732495e+00 -5.22532523e-01 -7.33016074e-01 1.45794022e+00
1.20032418e+00 -4.77516986e-02 3.08470100e-01 2.56266117e-01
3.89445215e-01 3.08646888e-01 1.89668000e-01 -1.53380489e+00
-6.19634748e-01 7.85959601e-01 -6.42854348e-02 8.04008365e-01
-1.49562013e+00 1.44252312e+00 6.58786201e+00 5.74407518e-01
-1.00891197e+00 2.50276506e-01 9.21628714e-01 5.26576154e-02
-3.02494913e-01 -3.66703659e-01 -5.23849666e-01 1.83675721e-01
1.51639879e+00 -1.72361154e-02 7.62434676e-02 7.63303638e-01
1.72261864e-01 -2.92652100e-01 -4.82687742e-01 5.46035111e-01
6.09294176e-01 -7.11309850e-01 8.15591812e-02 -7.12201178e-01
9.33149219e-01 1.82940692e-01 3.74162883e-01 5.51171958e-01
1.10598698e-01 -1.43397248e+00 7.31164336e-01 2.26080239e-01
6.98826909e-01 -1.39631736e+00 1.49817050e+00 -3.81289810e-01
-6.40438080e-01 1.22748829e-01 -1.02907903e-01 -3.35368872e-01
1.36698201e-01 7.10350752e-01 -5.50420821e-01 -3.63966376e-02
9.08202469e-01 1.08832359e+00 -5.54876745e-01 -8.96754190e-02
-4.89853978e-01 9.21783209e-01 -5.31114452e-02 -6.72048867e-01
5.08462131e-01 -2.67263830e-01 2.51365781e-01 1.79573894e+00
-1.08083084e-01 1.67343840e-01 -2.76454061e-01 2.14516390e-02
-3.37151945e-01 8.81317198e-01 -1.93275496e-01 -3.52789909e-01
7.21774697e-02 1.79320359e+00 -7.93389738e-01 -3.50626141e-01
-1.82401389e-01 1.00719011e+00 3.51231217e-01 -1.44885704e-01
-9.32217360e-01 -7.59102285e-01 3.96943063e-01 -5.09822011e-01
-6.08852841e-02 1.84686035e-01 -8.15666914e-01 -7.53823161e-01
-3.01268816e-01 -1.19802094e+00 4.53005046e-01 -1.07440925e+00
-1.44695294e+00 9.62684810e-01 -4.45382863e-01 -6.80582762e-01
1.32053187e-02 -1.01165009e+00 -4.54287499e-01 6.26184046e-01
-1.51927972e+00 -1.47005761e+00 -4.03760523e-01 6.38943613e-01
5.05575359e-01 -2.36590371e-01 1.22717798e+00 1.89976573e-01
-3.40364844e-01 6.49597406e-01 -4.97465044e-01 1.78460419e-01
1.33817017e+00 -1.43971550e+00 -2.45678082e-01 2.30624124e-01
-3.27427179e-01 3.03111881e-01 7.29706347e-01 -5.54518342e-01
-1.05119705e+00 -1.03963637e+00 1.49059308e+00 -7.79697180e-01
1.07400048e+00 -8.37864056e-02 -2.72581965e-01 4.95629191e-01
8.50405216e-01 -5.88509977e-01 1.32277131e+00 2.50702292e-01
-5.42370439e-01 2.58210730e-02 -1.16225791e+00 2.74977922e-01
-1.37276009e-01 -5.43662727e-01 -5.16432703e-01 4.19399738e-01
6.18610799e-01 -1.90617263e-01 -1.10983026e+00 5.27794778e-01
6.90756857e-01 -6.99404359e-01 5.82167029e-01 -7.76350021e-01
1.36931014e+00 -8.00590962e-02 -3.93262208e-01 -1.62027371e+00
2.54145354e-01 -3.14370751e-01 1.55584231e-01 1.13529265e+00
9.24931526e-01 -2.34096691e-01 3.57943773e-01 3.05963874e-01
-1.84194464e-02 -6.85875475e-01 -4.07800943e-01 1.47144735e-01
3.34379047e-01 -9.28185999e-01 2.52303243e-01 1.23423803e+00
7.84823895e-01 1.03788674e+00 -1.51273653e-01 -1.91578001e-01
-3.95099334e-02 -3.01983953e-01 3.43332767e-01 -8.53484690e-01
4.61148024e-01 -4.85655725e-01 -3.40209335e-01 -1.61827236e-01
5.61295867e-01 -8.41964960e-01 -4.37947027e-02 -1.45524585e+00
1.92617044e-01 -1.94265261e-01 -4.72831517e-01 7.45724618e-01
-2.30142787e-01 9.94573414e-01 -1.51068307e-02 -3.66360605e-01
-9.52388108e-01 5.94194651e-01 7.52105355e-01 -2.20095858e-01
-1.12207681e-01 -4.18573320e-01 -1.10217917e+00 9.99228835e-01
1.21435261e+00 -3.32303047e-01 -6.62896559e-02 -1.69308126e-01
1.17987835e+00 -6.99362636e-01 -1.93569884e-01 -2.81226605e-01
-3.43138754e-01 -3.30570936e-02 3.71202528e-01 -7.51510859e-01
4.41082627e-01 -5.23655772e-01 -7.51494110e-01 1.63347378e-01
-4.13567007e-01 4.23899114e-01 5.07157207e-01 -2.30157331e-01
-7.10779011e-01 -2.49217927e-01 8.49443614e-01 2.15737551e-01
-7.81378448e-01 -2.86692649e-01 -7.40478158e-01 9.39499214e-02
9.33377802e-01 3.51799279e-01 -3.90114456e-01 -6.69892311e-01
-8.70152116e-01 3.13756555e-01 -1.35908589e-01 5.59699893e-01
7.08253324e-01 -1.35335791e+00 -1.11752474e+00 -2.79765427e-01
4.86686498e-01 -6.06382251e-01 -1.12062901e-01 1.08074689e+00
-6.82275057e-01 2.01510504e-01 -2.31918991e-01 -4.72558104e-02
-1.38387322e+00 -1.10901229e-01 3.81429717e-02 -5.31667709e-01
4.88673955e-01 1.04771674e+00 -4.45264906e-01 -8.24716747e-01
-2.06673890e-01 3.65125656e-01 -9.90950763e-01 8.07950079e-01
8.39812994e-01 4.18765455e-01 2.85460651e-01 -1.22284830e+00
-5.03054738e-01 4.55126345e-01 -5.60188852e-02 -5.32633960e-01
1.50268114e+00 -7.99200758e-02 -1.05617356e+00 7.75108278e-01
1.28940690e+00 3.21300715e-01 4.53164950e-02 2.14078680e-01
2.82918245e-01 1.87957227e-01 3.12988698e-01 -1.51088643e+00
-7.41469562e-01 6.46317542e-01 5.80924273e-01 1.60168096e-01
1.13027930e+00 -6.94438592e-02 8.49080026e-01 5.16797245e-01
-1.99408993e-01 -1.81631780e+00 4.37046081e-01 1.36394894e+00
1.24330008e+00 -1.49578047e+00 -8.54782108e-03 -3.19545895e-01
-1.48401797e+00 1.32459271e+00 8.22883010e-01 -1.97740436e-01
9.68669832e-01 2.66344666e-01 7.27182806e-01 -6.32784009e-01
-5.82112610e-01 -1.16033882e-01 5.58043540e-01 4.29968119e-01
1.30497646e+00 2.55306095e-01 -8.98349106e-01 1.24623120e+00
-6.18353605e-01 -4.49712843e-01 7.05515206e-01 7.17453420e-01
-4.52771544e-01 -9.51707184e-01 -1.83248907e-01 2.13581085e-01
-8.21290910e-01 -2.66108781e-01 -1.20117247e+00 6.88393891e-01
-8.87605846e-02 1.46402550e+00 1.24992624e-01 -5.98875523e-01
3.09564739e-01 1.74757406e-01 2.23506033e-01 -4.11239356e-01
-1.04446590e+00 -8.21799412e-02 7.06887782e-01 -6.15199804e-01
-1.16193223e+00 -5.01874149e-01 -1.60955787e+00 -2.28379026e-01
-3.01798910e-01 4.45178151e-01 1.27542698e+00 1.03663373e+00
4.81747016e-02 3.10924739e-01 9.98317063e-01 -5.16930044e-01
-1.76919401e-02 -1.43188119e+00 -5.63960850e-01 4.52463478e-01
-6.18795566e-02 -3.22924942e-01 -3.63579690e-01 5.38473241e-02] | [11.27458667755127, 6.881308555603027] |
0d547d52-13cd-4c84-9cdf-aced18ae27fe | grafenne-learning-on-graphs-with | 2306.03447 | null | https://arxiv.org/abs/2306.03447v1 | https://arxiv.org/pdf/2306.03447v1.pdf | GRAFENNE: Learning on Graphs with Heterogeneous and Dynamic Feature Sets | Graph neural networks (GNNs), in general, are built on the assumption of a static set of features characterizing each node in a graph. This assumption is often violated in practice. Existing methods partly address this issue through feature imputation. However, these techniques (i) assume uniformity of feature set across nodes, (ii) are transductive by nature, and (iii) fail to work when features are added or removed over time. In this work, we address these limitations through a novel GNN framework called GRAFENNE. GRAFENNE performs a novel allotropic transformation on the original graph, wherein the nodes and features are decoupled through a bipartite encoding. Through a carefully chosen message passing framework on the allotropic transformation, we make the model parameter size independent of the number of features and thereby inductive to both unseen nodes and features. We prove that GRAFENNE is at least as expressive as any of the existing message-passing GNNs in terms of Weisfeiler-Leman tests, and therefore, the additional inductivity to unseen features does not come at the cost of expressivity. In addition, as demonstrated over four real-world graphs, GRAFENNE empowers the underlying GNN with high empirical efficacy and the ability to learn in continual fashion over streaming feature sets. | ['Srikanta Bedathur', 'Sayan Ranu', 'Sahil Manchanda', 'Shubham Gupta'] | 2023-06-06 | null | null | null | null | ['imputation', 'imputation', 'imputation'] | ['computer-vision', 'miscellaneous', 'time-series'] | [ 3.60600829e-01 2.37479344e-01 -5.75059932e-03 -2.46048853e-01
-2.29870692e-01 -6.78255498e-01 6.22472405e-01 2.97309548e-01
-3.59815806e-01 8.46763194e-01 -2.75027126e-01 -3.43939841e-01
-5.54288566e-01 -1.31068063e+00 -1.01318824e+00 -6.94470406e-01
-3.93474251e-01 4.06857133e-01 5.95707782e-02 -3.45947146e-01
-2.36260325e-01 4.83275384e-01 -1.39664626e+00 -1.75013810e-01
6.47300959e-01 7.34213173e-01 -3.43518883e-01 7.46407092e-01
-3.93341109e-02 7.44203627e-01 -3.19377899e-01 -5.92456698e-01
5.37138700e-01 -4.69325036e-01 -7.73267865e-01 1.45299911e-01
2.39407048e-01 -1.29095376e-01 -7.06414461e-01 9.62401271e-01
1.87595621e-01 1.10374756e-01 3.49017322e-01 -1.73335278e+00
-7.25082219e-01 9.45201874e-01 -3.71527314e-01 -1.78164601e-01
8.43523070e-02 3.31477374e-02 1.23343837e+00 -6.56277120e-01
5.88939667e-01 9.15104449e-01 1.00047684e+00 4.08842057e-01
-1.65796375e+00 -5.24207473e-01 2.29270816e-01 -1.93726018e-01
-1.39663398e+00 -3.45808238e-01 9.94418323e-01 -2.39523247e-01
6.72071099e-01 5.31769395e-01 7.93874621e-01 1.05113661e+00
9.42596644e-02 6.28129125e-01 7.67162085e-01 -3.36343408e-01
4.11653370e-01 1.07420040e-02 2.70849764e-01 8.60823870e-01
5.55892527e-01 7.65392184e-02 -3.35842431e-01 -4.10292476e-01
5.78085244e-01 1.97959647e-01 -2.54899651e-01 -7.34470665e-01
-9.48353589e-01 1.13367093e+00 5.24197161e-01 2.91311115e-01
-1.38353050e-01 3.81394058e-01 4.04282093e-01 6.97315693e-01
2.74205714e-01 3.76252681e-01 -3.86905134e-01 2.59885937e-01
-6.46986127e-01 3.51485908e-02 1.03737283e+00 8.70241582e-01
1.11008620e+00 8.32845718e-02 1.63371474e-01 5.10622442e-01
1.93981797e-01 3.34887862e-01 3.93747389e-01 -5.60186267e-01
2.74650484e-01 7.77673662e-01 -4.05504614e-01 -1.38418043e+00
-5.73964536e-01 -7.43393421e-01 -1.22747862e+00 6.43285736e-02
3.68235618e-01 -3.34070027e-01 -9.38591361e-01 2.28962803e+00
2.97831237e-01 1.83555149e-02 -8.44088346e-02 4.94906366e-01
6.78329408e-01 4.62612510e-01 -2.08826631e-01 -1.12445420e-02
9.07927632e-01 -4.65701818e-01 -2.97776490e-01 -3.25463116e-01
8.51491570e-01 -1.59493744e-01 8.84329379e-01 2.21607104e-01
-8.67977679e-01 -1.35654345e-01 -1.02950168e+00 1.41857952e-01
-3.91364962e-01 -4.84987885e-01 1.11063266e+00 8.83308053e-01
-1.39662242e+00 6.44780457e-01 -7.94290125e-01 -2.76596874e-01
2.60609835e-01 8.39503944e-01 -7.88112640e-01 -1.27655432e-01
-1.26877642e+00 4.32813644e-01 4.58133727e-01 3.43357503e-01
-4.48475033e-01 -3.94458771e-01 -9.40706253e-01 2.15860367e-01
4.73176628e-01 -8.88605595e-01 7.53264427e-01 -1.39590657e+00
-1.29834521e+00 4.21364516e-01 1.86300337e-01 -4.08246309e-01
5.21906495e-01 4.03512269e-01 -3.01382899e-01 2.38205250e-02
-2.18574926e-01 4.94381785e-01 7.36070931e-01 -1.24710429e+00
-2.59306461e-01 -3.08988214e-01 3.68292391e-01 -5.86544797e-02
-5.46045184e-01 -5.57147980e-01 -3.19109827e-01 -5.60612381e-01
1.79150358e-01 -1.03324568e+00 -2.99862444e-01 7.63521492e-02
-6.81417942e-01 2.30354033e-02 6.90182507e-01 -3.56790237e-02
9.15292740e-01 -2.08430576e+00 1.59591094e-01 7.88680613e-01
7.94594228e-01 9.63814482e-02 -4.59234715e-01 7.19412863e-01
1.35483278e-03 3.00385237e-01 -4.06768233e-01 -2.29118288e-01
1.39181927e-01 5.51579595e-01 1.07982587e-02 7.38111615e-01
2.22769707e-01 1.06818092e+00 -8.73070359e-01 -2.79670954e-01
-1.94140319e-02 4.31945503e-01 -8.95562291e-01 -2.18228683e-01
-6.93699941e-02 6.90886825e-02 -4.88247454e-01 3.20530325e-01
5.84983468e-01 -5.34829915e-01 4.55391377e-01 1.10955499e-01
4.28985894e-01 2.74303672e-03 -1.16979051e+00 1.37160754e+00
-2.93818623e-01 4.34362024e-01 -2.27977205e-02 -1.04135334e+00
8.27347338e-01 1.51658207e-01 5.12974322e-01 -4.07359749e-01
2.04257980e-01 -6.26577362e-02 2.32374370e-01 -5.38276285e-02
3.68378282e-01 -1.95549175e-01 -1.00999326e-01 5.94007134e-01
2.06031308e-01 3.48903835e-01 3.00019890e-01 3.71386647e-01
1.69911790e+00 -3.18413287e-01 1.74964920e-01 -1.62424251e-01
2.70465046e-01 -3.03396016e-01 6.41033649e-01 1.06694758e+00
1.20761938e-01 4.13668871e-01 8.55288446e-01 -4.03539449e-01
-9.41952109e-01 -1.19297814e+00 1.50368772e-02 1.06636059e+00
3.72808799e-02 -4.91968393e-01 -5.80192208e-01 -8.18765402e-01
5.33547141e-02 3.77668887e-01 -9.09468353e-01 -4.32820082e-01
-2.60532111e-01 -8.65872920e-01 7.63457775e-01 2.90520489e-01
1.68072402e-01 -8.79691005e-01 -3.53157669e-02 1.66165441e-01
2.16906726e-01 -8.86814833e-01 -2.19968662e-01 4.15536880e-01
-8.35667372e-01 -1.03676605e+00 -2.17554256e-01 -7.09311008e-01
1.07363462e+00 1.48207769e-02 1.02138722e+00 4.34828877e-01
3.61856930e-02 3.99982452e-01 -3.64979059e-01 5.97057305e-02
-4.97711897e-01 4.58178997e-01 -1.24220908e-01 2.58369416e-01
2.34533027e-01 -1.09461856e+00 -2.93904036e-01 7.29670227e-02
-1.30216134e+00 1.40385628e-01 6.22413337e-01 1.07045972e+00
2.41477042e-01 2.06435874e-01 8.01897287e-01 -1.33357656e+00
5.10479271e-01 -7.14973867e-01 -6.57521009e-01 2.23327011e-01
-7.06191540e-01 2.50173181e-01 9.41279292e-01 -5.17916203e-01
-4.87727553e-01 1.98575646e-01 -1.73353180e-02 -1.71796799e-01
2.04245280e-02 6.93558514e-01 -1.32754177e-01 -4.79747206e-01
7.13146925e-01 2.48287573e-01 2.22044557e-01 -8.98580849e-02
5.28422773e-01 2.40579978e-01 3.02143961e-01 -5.79712749e-01
1.21041441e+00 3.58537555e-01 3.69176686e-01 -8.19532156e-01
-4.23855126e-01 2.17370708e-02 -5.11374474e-01 -1.24118879e-01
1.49993509e-01 -5.12728691e-01 -9.69240010e-01 4.91395622e-01
-6.89734817e-01 -3.39886159e-01 -4.34583217e-01 3.26674491e-01
-3.13858211e-01 4.85353619e-01 -7.42029488e-01 -6.24367654e-01
-2.19155610e-01 -8.61325800e-01 4.52188760e-01 -1.49576172e-01
-2.15104878e-01 -1.25623798e+00 -1.50444418e-01 -1.23740785e-01
4.58193868e-01 4.10998464e-01 1.22594082e+00 -9.10634100e-01
-5.88256776e-01 -4.62418675e-01 -1.94085687e-01 2.01863989e-01
2.13273764e-01 -1.33327646e-02 -7.79375076e-01 -5.68015158e-01
-1.64582789e-01 -4.09538269e-01 8.67178023e-01 6.95741475e-02
9.06240404e-01 -5.94605744e-01 -7.58539513e-02 6.81041837e-01
1.53663373e+00 -2.33058140e-01 4.60271508e-01 7.49983415e-02
7.93636918e-01 4.08217877e-01 -4.35193479e-01 3.71483326e-01
3.23548943e-01 2.47630417e-01 6.54085338e-01 -1.51707903e-01
9.54991654e-02 -3.73876512e-01 2.77461946e-01 1.04672372e+00
2.08871931e-01 -5.86130559e-01 -7.67939210e-01 5.13700604e-01
-1.80507636e+00 -7.35484242e-01 -2.68901661e-02 2.14104486e+00
8.71636212e-01 2.25371167e-01 5.88493049e-02 3.46993327e-01
6.66440427e-01 3.71000133e-02 -5.08528709e-01 -3.47570926e-01
-3.34566921e-01 2.01937065e-01 7.03419745e-01 3.67430866e-01
-8.31311703e-01 6.23541594e-01 6.21260309e+00 5.92117786e-01
-1.04298401e+00 -1.00359485e-01 4.34660584e-01 5.12925424e-02
-7.38792241e-01 2.30060890e-01 -3.77471745e-01 2.66139358e-01
7.63901293e-01 -2.94468135e-01 7.56804287e-01 6.51426792e-01
-4.25413936e-01 2.47885540e-01 -1.32343411e+00 5.17816722e-01
5.19741094e-03 -1.08162951e+00 8.81804824e-02 1.60566702e-01
5.79683661e-01 1.56355947e-01 -2.18540872e-03 3.77153248e-01
6.41504765e-01 -1.14325082e+00 4.53105807e-01 3.00538480e-01
6.49421215e-01 -8.42687190e-01 5.50206065e-01 3.01816255e-01
-8.91172707e-01 -1.98249951e-01 -3.26597124e-01 -1.12182543e-01
-2.00391427e-01 7.71004736e-01 -6.85633838e-01 6.52316630e-01
2.26018488e-01 3.90546232e-01 -6.58426464e-01 9.46640670e-01
-6.46329969e-02 8.04060340e-01 -7.43707418e-01 -2.64078304e-02
3.34394097e-01 -5.54739460e-02 5.14270067e-01 8.93142700e-01
1.60743356e-01 -2.83137560e-01 2.31637850e-01 7.56838024e-01
-4.51878071e-01 3.51313017e-02 -7.81726122e-01 -1.11745834e-01
4.64044243e-01 1.24009287e+00 -8.47459078e-01 1.21849865e-01
-5.33845603e-01 6.22921407e-01 5.25728106e-01 5.12024522e-01
-5.41658044e-01 -4.56019670e-01 3.93827796e-01 8.34058598e-02
4.10007417e-01 -8.44454914e-02 -7.28729218e-02 -1.16466916e+00
1.43880770e-01 -7.16922879e-01 3.19508940e-01 -3.32093328e-01
-1.60701823e+00 6.28373563e-01 -2.81773716e-01 -8.68411720e-01
-2.37840369e-01 -5.09103894e-01 -4.18355197e-01 4.51446980e-01
-1.25957966e+00 -1.31001163e+00 -6.80460304e-04 7.34712303e-01
-1.43096715e-01 8.64032432e-02 9.12448585e-01 3.83163482e-01
-5.48177004e-01 7.73172021e-01 1.69416681e-01 3.89471442e-01
3.18033606e-01 -1.14963782e+00 4.53272253e-01 8.09263289e-01
3.27806652e-01 7.29182065e-01 5.95771551e-01 -6.41568244e-01
-1.76106894e+00 -1.24665356e+00 7.01531172e-01 -8.83499756e-02
8.43446255e-01 -7.95034111e-01 -9.06747222e-01 9.90337014e-01
-2.79892713e-01 3.74662787e-01 6.82039559e-01 5.70113182e-01
-5.21910071e-01 -2.35760376e-01 -1.07030272e+00 6.28611147e-01
1.25267756e+00 -3.77163231e-01 -2.53000110e-01 2.72138208e-01
7.83070326e-01 -1.42695457e-01 -8.73923838e-01 4.19452846e-01
4.83468324e-01 -8.12357008e-01 7.49266505e-01 -8.23213398e-01
1.98202640e-01 -1.59841284e-01 -1.73766151e-01 -1.17769992e+00
-4.77903455e-01 -7.59693921e-01 -5.34118079e-02 1.24223447e+00
6.69011354e-01 -1.05807698e+00 9.19967115e-01 5.71061075e-01
1.71576053e-01 -8.28612626e-01 -1.01756644e+00 -9.09541607e-01
1.23875625e-02 -4.25875247e-01 7.39818215e-01 1.10760880e+00
-6.51631877e-02 4.61811066e-01 -5.52565157e-01 2.62879170e-02
5.73319137e-01 6.13373742e-02 8.22893858e-01 -1.55009878e+00
-5.87440670e-01 -4.02285278e-01 -7.77046084e-01 -5.27046740e-01
2.39583656e-01 -1.30884266e+00 -9.23702270e-02 -1.27102435e+00
3.17217141e-01 -8.08148205e-01 -4.20659691e-01 8.98763597e-01
1.44154191e-01 2.32527196e-01 2.10749414e-02 -7.11888596e-02
-5.09507895e-01 5.93869686e-01 9.75961685e-01 -5.23634739e-02
-1.96616456e-01 5.91169260e-02 -9.09662485e-01 5.37510395e-01
8.74002278e-01 -6.70748889e-01 -6.88408136e-01 -3.02441359e-01
7.08875239e-01 -7.36304298e-02 5.96782684e-01 -7.32474267e-01
3.93131286e-01 -6.82028010e-02 4.12336916e-01 -1.71369985e-02
2.69484222e-01 -8.45500588e-01 4.78070825e-01 3.92933607e-01
-3.80117506e-01 -6.19107112e-02 -7.63581544e-02 7.87480891e-01
1.13175631e-01 -1.77203521e-01 4.60057467e-01 2.27432117e-01
-2.71925300e-01 5.96419394e-01 -2.30578557e-01 1.62444606e-01
7.68948913e-01 -3.59349042e-01 -4.49726611e-01 -5.89649618e-01
-4.44295347e-01 1.13370806e-01 6.25823140e-01 2.01411709e-01
4.42431808e-01 -1.21142936e+00 -5.89976490e-01 5.56506276e-01
9.04105678e-02 5.57301082e-02 1.29371330e-01 8.95634472e-01
-2.31792316e-01 -1.33333102e-01 -5.03034405e-02 -4.34565037e-01
-8.80386174e-01 7.34542608e-01 1.78491965e-01 -2.70939052e-01
-7.91832447e-01 5.33275068e-01 2.27827609e-01 -6.51606560e-01
1.49647132e-01 1.97890121e-02 1.90706745e-01 -1.54239520e-01
-1.29558123e-03 7.73800090e-02 1.76707879e-01 -5.70107400e-01
-1.73306346e-01 2.31783584e-01 -1.10828884e-01 9.38037932e-02
1.44499147e+00 -1.19137159e-02 -2.33797342e-01 4.20237273e-01
1.26329887e+00 1.25641599e-01 -1.04654264e+00 -3.70706588e-01
-1.13135107e-01 -2.28184015e-01 -1.32808119e-01 -5.21121144e-01
-1.31825256e+00 4.41943765e-01 3.99246737e-02 6.20734215e-01
1.19080937e+00 -1.55348539e-01 7.40310609e-01 7.24335074e-01
2.11865544e-01 -6.81921840e-01 -4.38467145e-01 4.31813478e-01
4.41751301e-01 -9.17646468e-01 -9.45264772e-02 -3.79919171e-01
-1.30374074e-01 1.07803547e+00 2.39311606e-01 -3.02724659e-01
6.23457968e-01 4.40098464e-01 -3.23103935e-01 -1.54565349e-01
-9.83653426e-01 -8.39384049e-02 1.26941115e-01 6.29897773e-01
1.22801594e-01 -4.41458970e-02 -1.92969605e-01 6.05752409e-01
-2.84879386e-01 -5.50602898e-02 5.76967835e-01 8.61778915e-01
-2.31990829e-01 -1.24393487e+00 8.21260810e-02 8.16756308e-01
-3.79705518e-01 -1.53748274e-01 -5.26359797e-01 1.11681914e+00
-1.31533027e-01 8.56705368e-01 -2.34601106e-02 -5.31803846e-01
6.57160357e-02 -4.50484045e-02 4.46227700e-01 -3.32357794e-01
-7.06380785e-01 -3.18852991e-01 6.31794557e-02 -2.87874937e-01
-1.23181544e-01 -3.06002736e-01 -1.05561769e+00 -7.09866345e-01
-5.73135316e-01 1.59927770e-01 4.65818256e-01 9.47950959e-01
4.55268323e-01 4.88467306e-01 7.87902474e-01 -4.44468349e-01
-6.69630945e-01 -5.57336330e-01 -7.99598396e-01 3.92199576e-01
3.25399756e-01 -4.16342646e-01 -6.83050096e-01 -2.99083948e-01] | [6.9033331871032715, 6.032680511474609] |
6c8a0254-f1fc-48c6-83c2-1ef147d99d04 | cost-sensitive-bert-for-generalisable | null | null | https://aclanthology.org/D19-5018 | https://aclanthology.org/D19-5018.pdf | Cost-Sensitive BERT for Generalisable Sentence Classification on Imbalanced Data | The automatic identification of propaganda has gained significance in recent years due to technological and social changes in the way news is generated and consumed. That this task can be addressed effectively using BERT, a powerful new architecture which can be fine-tuned for text classification tasks, is not surprising. However, propaganda detection, like other tasks that deal with news documents and other forms of decontextualized social communication (e.g. sentiment analysis), inherently deals with data whose categories are simultaneously imbalanced and dissimilar. We show that BERT, while capable of handling imbalanced classes with no additional data augmentation, does not generalise well when the training and test data are sufficiently dissimilar (as is often the case with news sources, whose topics evolve over time). We show how to address this problem by providing a statistical measure of similarity between datasets and a method of incorporating cost-weighting into BERT when the training and test sets are dissimilar. We test these methods on the Propaganda Techniques Corpus (PTC) and achieve the second highest score on sentence-level propaganda classification. | ['Michael Castelle', 'Harish Tayyar Madabushi', 'Elena Kochkina'] | 2019-11-01 | null | null | null | ws-2019-11 | ['propaganda-detection'] | ['natural-language-processing'] | [ 2.23506987e-01 -1.09217197e-01 -2.18801185e-01 -5.29217005e-01
-5.77184439e-01 -6.27090693e-01 1.17309070e+00 7.81000495e-01
-5.78446090e-01 6.99913323e-01 5.78099012e-01 -4.69650447e-01
-7.92667270e-02 -7.51954854e-01 -1.70138285e-01 -5.82104027e-01
-8.81263614e-02 5.83524227e-01 3.55877392e-02 -7.42988646e-01
6.29557431e-01 1.82200253e-01 -1.40988171e+00 6.04123056e-01
8.46988320e-01 6.48227155e-01 -4.39259291e-01 7.11368382e-01
-3.10285926e-01 1.13458097e+00 -1.37524807e+00 -6.22220516e-01
-1.13260753e-01 -6.02208078e-01 -1.06026244e+00 -1.11675829e-01
4.72305596e-01 5.12490906e-02 1.33830860e-01 9.34891820e-01
5.40544450e-01 -2.35599607e-01 9.43221390e-01 -8.57573628e-01
-2.98099279e-01 7.80665934e-01 -6.41143024e-01 6.73514545e-01
4.27094102e-01 -4.97217298e-01 8.77818286e-01 -4.45057690e-01
6.87392712e-01 1.35465896e+00 7.91095138e-01 4.08074319e-01
-1.26519644e+00 -4.23864454e-01 2.88930982e-02 2.64423549e-01
-6.07925892e-01 -3.29270154e-01 9.75882590e-01 -6.91198170e-01
8.14013481e-01 5.22840023e-01 6.78560734e-01 1.34365857e+00
3.54163378e-01 8.02410126e-01 1.04552841e+00 -5.95855594e-01
2.54737556e-01 3.45553696e-01 3.67162019e-01 2.68983126e-01
1.95594817e-01 -4.86384720e-01 -5.16062975e-01 -5.00597417e-01
-1.75408259e-01 -4.84397858e-01 -7.12141320e-02 6.30512554e-03
-1.17684805e+00 1.34309077e+00 9.72817540e-02 8.18156779e-01
-1.76317543e-01 -4.79299054e-02 9.65183914e-01 8.22457790e-01
1.12108767e+00 9.61997092e-01 -4.19712156e-01 -3.47125351e-01
-9.96404111e-01 4.43232834e-01 1.04348767e+00 2.37583444e-01
1.05815478e-01 1.04458086e-01 -7.39868581e-02 9.04070973e-01
-1.62118345e-01 3.73608530e-01 8.39764118e-01 -2.64868021e-01
6.43101871e-01 5.66419899e-01 -1.20641306e-01 -1.39205337e+00
-7.54454851e-01 -6.40721858e-01 -9.20842588e-01 1.38186872e-01
5.04235566e-01 -2.17047811e-01 -6.81016862e-01 1.72552347e+00
4.06444222e-01 -6.18345320e-01 3.41507606e-02 5.07238448e-01
7.84168661e-01 6.37522936e-01 -1.07338307e-02 -6.29865825e-01
1.31172490e+00 -5.05202055e-01 -8.87747824e-01 -3.33718151e-01
1.00767243e+00 -9.58563805e-01 8.76657307e-01 4.73285198e-01
-8.12302709e-01 -3.28886434e-02 -1.09363663e+00 1.97280899e-01
-6.02954268e-01 -4.85509962e-01 6.58886492e-01 8.86407197e-01
-4.77290630e-01 5.40240288e-01 -4.72072333e-01 -3.10317308e-01
5.47292769e-01 1.57282595e-02 -2.83479095e-01 4.00420785e-01
-1.40231848e+00 1.29869986e+00 2.86058396e-01 -2.47947618e-01
-3.94172072e-01 -4.54624444e-01 -5.70292652e-01 -1.38839275e-01
1.95389718e-01 -2.08044678e-01 1.13193500e+00 -1.35387802e+00
-1.10534251e+00 1.01584482e+00 3.31095427e-01 -4.88224447e-01
5.75682878e-01 -4.92754439e-03 -6.23044729e-01 -1.36970833e-01
1.39355645e-01 -8.46868455e-02 1.14180541e+00 -7.88469315e-01
-4.34351504e-01 -4.45221514e-01 -1.24295540e-01 1.52963057e-01
-7.08169818e-01 5.99240899e-01 6.31164134e-01 -8.45847905e-01
-4.17721942e-02 -4.92353737e-01 -4.98206727e-02 -6.08447492e-01
-4.73664105e-01 -2.50128776e-01 8.65804315e-01 -6.15882635e-01
1.16067672e+00 -1.88068843e+00 2.93939382e-01 2.56820470e-01
3.62891763e-01 4.52885538e-01 2.26924792e-02 6.33093536e-01
-9.69655737e-02 3.09207857e-01 1.43703707e-02 -5.07441573e-02
-1.68292853e-03 1.45222070e-02 -3.89362663e-01 6.98233783e-01
1.57076418e-01 3.37893099e-01 -8.89134824e-01 -3.51102382e-01
-2.22810686e-01 8.72590765e-02 -5.22381127e-01 -2.34689608e-01
-2.22043663e-01 1.71400845e-01 -4.59396362e-01 1.87803507e-01
3.75505596e-01 -1.21527940e-01 1.99531272e-01 1.92477584e-01
-1.71677306e-01 7.22498715e-01 -8.62308264e-01 1.17086375e+00
-3.23702931e-01 1.10427105e+00 1.36681730e-02 -1.30998588e+00
8.58661473e-01 2.70398229e-01 3.34111661e-01 -8.26352298e-01
5.66698909e-01 1.71373621e-01 2.36751035e-01 -6.09292686e-01
5.51973641e-01 -4.42252398e-01 -5.05482316e-01 8.56466234e-01
-1.68404266e-01 -3.19748312e-01 5.10749400e-01 4.27042335e-01
1.38784015e+00 -8.53668094e-01 4.63233739e-01 -4.47799742e-01
3.76312882e-01 4.39727932e-01 3.37950408e-01 8.08227241e-01
-1.34246752e-01 2.78857529e-01 8.73376608e-01 -7.73217916e-01
-1.07940769e+00 -7.67710209e-01 -3.79586399e-01 1.22128725e+00
-1.47409469e-01 -5.95677853e-01 -5.41736543e-01 -9.84525740e-01
-1.11428082e-01 6.26879334e-01 -7.21490741e-01 -8.76227915e-02
-6.13866031e-01 -1.36324501e+00 4.87481058e-01 -7.17567839e-03
1.10554099e-01 -9.69894469e-01 -8.40520799e-01 4.62506682e-01
-2.32120693e-01 -6.80814266e-01 1.58157080e-01 5.03764153e-01
-5.54893374e-01 -1.03709686e+00 -2.97228605e-01 -6.92576945e-01
3.98260295e-01 2.97909230e-03 1.31702650e+00 1.06348254e-01
-4.03681070e-01 1.20707221e-01 -7.79234409e-01 -7.62622237e-01
-1.05802798e+00 3.13610405e-01 4.25550677e-02 -1.74345523e-01
2.63620943e-01 -5.63137472e-01 -1.17128700e-01 6.97827991e-03
-1.00035596e+00 2.08130196e-01 2.29897469e-01 1.07438600e+00
-3.27206135e-01 2.94448972e-01 1.08688474e+00 -1.16787565e+00
1.13242817e+00 -5.36071122e-01 -9.39944983e-02 -8.43349099e-02
-3.90714735e-01 -2.27926262e-02 6.88474476e-01 -4.21992540e-01
-8.41083288e-01 -6.33676171e-01 -4.09219027e-01 5.96499979e-01
-1.46695405e-01 8.49435210e-01 3.81209821e-01 2.41126835e-01
1.25530088e+00 -2.12739870e-01 1.96581662e-01 -4.52124298e-01
2.58531302e-01 1.11677897e+00 1.86258793e-01 1.33661227e-03
8.48161697e-01 5.19731700e-01 -3.11136812e-01 -8.12194228e-01
-1.24929190e+00 -5.46764791e-01 -4.11250532e-01 -7.53681734e-02
4.91713166e-01 -6.00605547e-01 -2.02393293e-01 5.45219898e-01
-1.24479926e+00 -3.62496786e-02 -3.08334976e-01 3.37411106e-01
-3.57434034e-01 2.38464713e-01 -6.32997215e-01 -8.32966268e-01
-3.71151865e-01 -4.97069180e-01 6.00379348e-01 -2.19023913e-01
-6.56905890e-01 -1.01712692e+00 3.40467006e-01 4.73192155e-01
6.28600955e-01 4.75752622e-01 1.35456944e+00 -1.22009254e+00
5.14844239e-01 -5.26474833e-01 1.12095624e-01 4.72840279e-01
2.04010963e-01 -1.33698374e-01 -7.81915784e-01 -2.18648300e-01
5.72299480e-01 -6.08195245e-01 8.87514532e-01 1.21277727e-01
6.72448039e-01 -8.06007087e-01 7.33926473e-03 -1.60166472e-01
8.99255276e-01 4.70820591e-02 5.42918146e-01 6.69500291e-01
3.70287865e-01 9.95785713e-01 4.68218774e-01 4.53053653e-01
1.63709000e-01 7.34601498e-01 4.53941315e-01 -2.27799743e-01
8.38604420e-02 3.13058615e-01 3.43151301e-01 9.78170872e-01
1.35443076e-01 -3.26036185e-01 -9.68384683e-01 4.48949158e-01
-1.70001161e+00 -1.60444248e+00 -4.21388805e-01 1.88213086e+00
1.15117466e+00 3.15637201e-01 2.22718060e-01 8.24842453e-01
5.33958554e-01 4.92848635e-01 -4.96944077e-02 -1.04283178e+00
-3.66064519e-01 8.75034705e-02 -1.97963174e-02 4.18671072e-01
-1.44410455e+00 4.76379097e-01 6.32174683e+00 9.72609103e-01
-1.13319421e+00 2.05687270e-01 6.74320936e-01 -2.85907507e-01
-3.10162038e-01 -4.16931868e-01 -3.63367379e-01 6.13593817e-01
8.65070999e-01 -1.56756014e-01 1.81221459e-02 6.12206876e-01
2.15061769e-01 -6.28370568e-02 -9.19422388e-01 6.44801915e-01
4.61699575e-01 -1.45608938e+00 -1.95978180e-01 -1.23560108e-01
7.54506946e-01 1.30425066e-01 1.61022656e-02 2.91327119e-01
9.94747281e-02 -9.42083478e-01 8.49307299e-01 -2.22603694e-01
1.90390691e-01 -6.95958316e-01 9.92639303e-01 6.96652591e-01
-2.28698790e-01 -2.41203770e-01 -1.70345753e-01 -5.01476347e-01
1.80211201e-01 1.07543480e+00 -8.71992886e-01 3.31491202e-01
5.33287287e-01 5.78020692e-01 -7.08213747e-01 7.60052323e-01
-4.54256982e-02 7.66207933e-01 -2.35500738e-01 -4.48790193e-01
3.11118960e-01 2.68660963e-01 8.06561589e-01 1.37614858e+00
-1.14942594e-02 -4.34879601e-01 -9.62117910e-02 2.13918626e-01
1.75692290e-01 5.78697979e-01 -5.50047994e-01 -1.06017053e-01
-1.22909062e-02 1.17577231e+00 -6.76499903e-01 -4.14234698e-01
1.32687101e-02 4.53498542e-01 3.34093064e-01 -1.63176015e-01
-5.83612323e-01 -3.33837122e-01 3.12881887e-01 2.36761823e-01
-4.56671603e-02 -2.53644101e-02 -1.72518358e-01 -1.23892534e+00
1.49576943e-02 -1.15709865e+00 4.57325697e-01 -2.52496302e-01
-1.54313183e+00 6.77273273e-01 -1.03159308e-01 -1.05520880e+00
-5.31347632e-01 -4.13281173e-01 -6.66344345e-01 3.11307043e-01
-1.20790756e+00 -7.87474990e-01 -4.31143157e-02 2.60292441e-01
5.32684743e-01 -1.60515085e-01 8.67408335e-01 3.34448725e-01
-1.82567045e-01 2.31254295e-01 3.46406817e-01 1.90017805e-01
7.38247335e-01 -1.23928261e+00 1.59025431e-01 5.75846612e-01
1.72375172e-01 1.58829302e-01 1.30596614e+00 -3.93954217e-01
-9.26728725e-01 -8.42864275e-01 1.31418622e+00 -5.33717215e-01
1.06109917e+00 -7.87573099e-01 -7.80721545e-01 1.37428671e-01
2.54547119e-01 -5.61976314e-01 9.15833414e-01 5.06375074e-01
-7.19287038e-01 1.13114931e-01 -1.14936805e+00 4.86812830e-01
8.30844760e-01 -2.26926416e-01 -9.52608168e-01 9.49494123e-01
3.92855227e-01 -2.40225405e-01 -4.66229677e-01 2.31279850e-01
3.80734831e-01 -1.03695297e+00 6.03177726e-01 -1.10468316e+00
6.83519542e-01 7.45215192e-02 9.32972953e-02 -1.61415815e+00
-1.23848796e-01 -4.29052651e-01 1.44371480e-01 1.14081478e+00
5.08846402e-01 -6.06508613e-01 6.54186487e-01 -1.07337132e-01
-3.48944403e-02 -5.40908396e-01 -1.23941135e+00 -7.68664122e-01
2.65893370e-01 -3.84595484e-01 2.45061800e-01 1.42573786e+00
5.37988722e-01 8.25993717e-01 -5.64345717e-01 -5.17844677e-01
2.98371673e-01 2.60620892e-01 6.14739716e-01 -1.60786128e+00
-1.21776119e-01 -7.23276734e-01 -8.40603650e-01 -4.15986240e-01
1.84678644e-01 -8.89131248e-01 -7.13988766e-02 -1.40655351e+00
2.12060586e-01 -4.44227278e-01 1.49650142e-01 3.13514143e-01
7.42805377e-02 3.61154675e-01 1.59496106e-02 1.47047520e-01
-4.31067824e-01 4.40539718e-01 9.37790394e-01 -3.54766488e-01
-1.48593336e-01 9.28145200e-02 -8.49277079e-01 8.69525135e-01
1.03913605e+00 -7.71383286e-01 -1.60394311e-01 -2.79268414e-01
8.28670442e-01 -3.19590271e-01 1.68551177e-01 -8.68344069e-01
3.79352383e-02 -9.10190642e-02 1.64697111e-01 -5.12122095e-01
2.76208073e-01 -4.79340106e-01 -1.02322817e-01 6.73846781e-01
-6.44541800e-01 1.56089976e-01 -5.95671423e-02 5.19239426e-01
-2.63728380e-01 -4.60495025e-01 7.68159926e-01 7.24708885e-02
1.52165387e-02 -3.07931513e-01 -9.50085819e-01 2.63704062e-01
8.67299139e-01 6.61615208e-02 -1.10687256e+00 -4.89994138e-01
-5.52352786e-01 -6.36857674e-02 2.12510154e-01 5.15963137e-01
1.20698527e-01 -1.14845991e+00 -1.12813985e+00 6.74675629e-02
2.44476013e-02 -3.31347167e-01 1.24073952e-01 1.13289249e+00
-5.19708633e-01 9.30916741e-02 -9.92581993e-02 -3.70235711e-01
-1.57381928e+00 3.03609103e-01 4.45514508e-02 -5.37251770e-01
-4.74948287e-01 6.20341659e-01 -1.87177300e-01 -1.82534933e-01
6.83335736e-02 6.05139602e-03 -6.09626591e-01 7.75980651e-01
7.80082405e-01 2.00812384e-01 4.56757724e-01 -7.22288489e-01
-2.79801130e-01 5.28740212e-02 -4.56663966e-01 5.31939380e-02
1.68316960e+00 1.50367454e-01 -5.04162490e-01 6.12967074e-01
1.11945164e+00 2.61710435e-01 -1.33203208e-01 -9.22638327e-02
1.91531852e-01 -3.71738493e-01 -1.18823741e-02 -9.13903117e-01
-6.69010580e-01 7.84673035e-01 2.20972016e-01 1.23308444e+00
7.41066933e-01 9.82707888e-02 3.74399364e-01 4.05666590e-01
7.50949383e-02 -1.44015408e+00 3.54613721e-01 8.49269152e-01
1.03043091e+00 -1.08181107e+00 2.81019390e-01 -2.17116669e-01
-4.26518500e-01 9.71360445e-01 6.08478710e-02 -6.12901524e-03
3.65779936e-01 3.12504202e-01 1.61930948e-01 -5.23685813e-01
-9.47594225e-01 1.66853309e-01 9.79737639e-02 4.35868055e-01
4.75813776e-01 2.06338577e-02 -8.88286054e-01 2.40223944e-01
-5.27812004e-01 -7.05647469e-01 8.60877872e-01 1.11797130e+00
-7.64932752e-01 -9.36848640e-01 -5.08565485e-01 1.00016475e+00
-8.23457360e-01 -5.31582050e-02 -8.27846348e-01 7.01522470e-01
-3.96388248e-02 1.16849303e+00 1.70515761e-01 -3.77201468e-01
1.72654480e-01 6.21625409e-02 3.03859770e-01 -7.21852422e-01
-1.20139086e+00 -8.24212879e-02 8.08109641e-01 -1.16272666e-01
-6.87370360e-01 -6.02974832e-01 -6.92507982e-01 -6.00333154e-01
-5.68140924e-01 6.23716772e-01 8.74325693e-01 1.10141647e+00
7.86254276e-03 4.34660703e-01 7.79406548e-01 -7.57848263e-01
-9.40417647e-01 -1.14210093e+00 -5.06300688e-01 6.56115055e-01
4.92379963e-01 -4.96187359e-01 -8.41209948e-01 -1.32194355e-01] | [8.583970069885254, 10.378092765808105] |
9a4918b4-4523-4720-9372-3e9f8a7507af | d4ft-a-deep-learning-approach-to-kohn-sham | 2303.00399 | null | https://arxiv.org/abs/2303.00399v1 | https://arxiv.org/pdf/2303.00399v1.pdf | D4FT: A Deep Learning Approach to Kohn-Sham Density Functional Theory | Kohn-Sham Density Functional Theory (KS-DFT) has been traditionally solved by the Self-Consistent Field (SCF) method. Behind the SCF loop is the physics intuition of solving a system of non-interactive single-electron wave functions under an effective potential. In this work, we propose a deep learning approach to KS-DFT. First, in contrast to the conventional SCF loop, we propose to directly minimize the total energy by reparameterizing the orthogonal constraint as a feed-forward computation. We prove that such an approach has the same expressivity as the SCF method, yet reduces the computational complexity from O(N^4) to O(N^3). Second, the numerical integration which involves a summation over the quadrature grids can be amortized to the optimization steps. At each step, stochastic gradient descent (SGD) is performed with a sampled minibatch of the grids. Extensive experiments are carried out to demonstrate the advantage of our approach in terms of efficiency and stability. In addition, we show that our approach enables us to explore more complex neural-based wave functions. | ['Shuicheng Yan', 'Kostya S. Novoselov', 'A. H. Castro Neto', 'Kenji Kawaguchi', 'Giovanni Vignale', 'Kunhao Zheng', 'Zheyuan Hu', 'Min Lin', 'Tianbo Li'] | 2023-03-01 | null | null | null | null | ['numerical-integration', 'total-energy'] | ['miscellaneous', 'miscellaneous'] | [ 8.68143514e-02 -3.69265825e-01 1.59698695e-01 -3.42699468e-01
-9.76938128e-01 -1.72730893e-01 3.58277321e-01 1.00885583e-02
-7.31931150e-01 1.31500280e+00 -2.55205482e-01 -4.04792398e-01
-3.48358974e-02 -9.01984632e-01 -8.21788251e-01 -1.15800393e+00
-6.63544312e-02 3.28551352e-01 -6.87793344e-02 -1.69906721e-01
4.01269168e-01 6.12626135e-01 -1.42559338e+00 2.29648992e-01
1.14560318e+00 1.23256361e+00 1.07118770e-01 3.93861771e-01
-1.64711140e-02 9.37268913e-01 3.63087445e-03 -1.92152575e-01
1.86812028e-01 -4.30569142e-01 -9.85019267e-01 -2.45877579e-01
3.35514724e-01 -1.02472976e-01 -3.53009075e-01 1.25506341e+00
7.04503894e-01 6.89521611e-01 6.17892265e-01 -8.69927287e-01
-4.60491627e-01 1.50623411e-01 -2.66378671e-01 -8.57509524e-02
-1.66189838e-02 1.74580708e-01 9.54198956e-01 -1.23278236e+00
4.80096757e-01 8.89612556e-01 9.30523872e-01 4.67642456e-01
-1.69833434e+00 -5.41575730e-01 -2.75960982e-01 3.54685724e-01
-1.80758500e+00 -4.24929112e-01 8.72898638e-01 -2.20644340e-01
1.22109520e+00 1.71019047e-01 9.84039009e-01 5.74111044e-01
3.88483524e-01 4.20156509e-01 1.07696319e+00 -7.77164161e-01
8.74030769e-01 -5.98513708e-02 7.79529363e-02 7.74196684e-01
2.99076140e-01 6.14201948e-02 -6.28658831e-01 -3.72377366e-01
4.58222151e-01 -2.24111248e-02 -3.34681302e-01 -6.04151845e-01
-6.60049021e-01 1.34702897e+00 6.27692223e-01 1.45311147e-01
-4.19064939e-01 5.29066384e-01 5.27376533e-01 1.15579687e-01
6.80944979e-01 5.45773506e-01 -3.32869053e-01 1.18905254e-01
-1.19026458e+00 5.65746725e-01 9.37462687e-01 1.99584842e-01
1.03618896e+00 2.40976453e-01 1.78910747e-01 4.67029780e-01
2.23697573e-01 2.84228832e-01 1.79731742e-01 -1.18480325e+00
-6.44512614e-03 1.89088538e-01 3.63287777e-01 -7.69200385e-01
-2.16575623e-01 -3.27806562e-01 -1.11178410e+00 4.94527608e-01
1.95500135e-01 -3.43941748e-01 -5.84835708e-01 1.49836504e+00
3.16353559e-01 -4.37227711e-02 -8.15983191e-02 9.10996556e-01
4.60796624e-01 8.43299747e-01 -2.54300207e-01 -6.46190345e-01
7.47817397e-01 -9.22420025e-01 -6.36537910e-01 -4.17183973e-02
8.19678426e-01 -3.35403651e-01 9.29367602e-01 5.11403680e-01
-1.29870796e+00 -1.32716492e-01 -9.28804815e-01 -3.80859554e-01
-2.19426870e-01 -6.53373599e-02 1.02566707e+00 2.45829269e-01
-1.30627549e+00 1.26180184e+00 -9.29540694e-01 1.97542429e-01
4.64098960e-01 7.59334683e-01 -3.93461019e-01 1.27779916e-01
-1.08135343e+00 8.54821682e-01 3.77667129e-01 1.10319220e-01
-5.76775134e-01 -7.07839668e-01 -8.91696990e-01 1.78383559e-01
3.28588188e-01 -5.69228232e-01 1.23149407e+00 -9.08095479e-01
-1.66790140e+00 5.31865299e-01 -7.30569839e-01 -4.97202665e-01
2.08041936e-01 -6.24313718e-03 1.49256080e-01 -4.26436169e-03
3.46688018e-03 3.96897227e-01 4.16984141e-01 -9.78285253e-01
3.23053859e-02 -2.68487930e-01 5.49307503e-02 1.61693230e-01
-1.42848834e-01 -2.25818977e-01 -7.30357412e-03 -2.31799990e-01
1.83516711e-01 -9.54829872e-01 -5.53158939e-01 1.40390128e-01
-2.33416483e-01 -1.81549147e-01 1.55855611e-01 -3.50791901e-01
1.27421653e+00 -1.94740796e+00 3.50744992e-01 3.29879820e-01
2.71697372e-01 1.54393703e-01 3.49425733e-01 7.03024864e-01
-1.27453238e-01 -8.77846330e-02 -5.96192777e-01 -4.82098639e-01
-1.41049046e-02 1.66684464e-01 -1.57151595e-01 7.09506512e-01
-8.40701237e-02 9.75743949e-01 -5.75131297e-01 -3.54866654e-01
1.64457671e-02 7.43778110e-01 -9.97155726e-01 -1.93352088e-01
-1.25125125e-01 3.99894893e-01 -2.56763369e-01 1.49135083e-01
9.07797933e-01 -6.48184121e-01 3.21620286e-01 -3.81296605e-01
-5.94493032e-01 7.06096292e-01 -1.21410429e+00 1.82727325e+00
-3.50035399e-01 3.26358795e-01 2.62722403e-01 -1.53847086e+00
4.15183485e-01 1.40558586e-01 5.16408861e-01 -6.45282447e-01
8.22599456e-02 5.52543700e-01 -2.67385156e-03 -9.76641029e-02
2.20372528e-01 -6.76135480e-01 3.76408473e-02 3.35445732e-01
1.96196839e-01 -2.19016790e-01 1.63661569e-01 -1.88556779e-02
7.83774853e-01 1.84110389e-03 5.83400905e-01 -6.95243001e-01
6.81664646e-01 1.52247623e-01 3.92467588e-01 7.58578420e-01
1.34348735e-01 1.66883022e-01 3.56289178e-01 -8.50391924e-01
-9.55499649e-01 -7.77124524e-01 -3.19475681e-01 6.67867124e-01
-4.14302237e-02 -4.74250048e-01 -8.70984018e-01 -2.16788262e-01
2.08860580e-02 8.07253003e-01 -5.10934532e-01 -1.94830924e-01
-6.10372663e-01 -8.64136875e-01 1.30284667e-01 3.21563780e-01
4.27690804e-01 -1.16758561e+00 -5.88085830e-01 2.60003299e-01
1.80763274e-01 -4.45168942e-01 -3.33649457e-01 6.96994483e-01
-8.49256098e-01 -6.60100281e-01 -5.31588137e-01 -6.62821293e-01
4.83276427e-01 1.39133558e-01 8.90311658e-01 -6.94776699e-02
-3.38793665e-01 -1.28375456e-01 2.64016539e-01 1.86428465e-02
4.68892092e-03 1.83583554e-02 3.39696139e-01 -1.47005886e-01
3.99811953e-01 -9.17114794e-01 -6.14200294e-01 -4.46842045e-01
-7.11541295e-01 1.33961454e-01 3.05060685e-01 1.06520069e+00
9.30409849e-01 1.67730048e-01 2.81804651e-01 -6.93539202e-01
6.01498485e-01 -3.10770810e-01 -9.38842475e-01 1.69927943e-02
-6.51368618e-01 4.34765965e-01 7.86151052e-01 -2.88789690e-01
-9.04970109e-01 2.78912246e-01 -3.90899658e-01 -3.68616104e-01
2.54405886e-01 7.29473531e-01 1.47863999e-01 -8.63806307e-01
6.30187690e-01 5.32305360e-01 -1.74831264e-02 -5.42871594e-01
1.70823470e-01 3.64559174e-01 2.89620429e-01 -5.92997074e-01
6.05146885e-01 6.01907253e-01 3.82866919e-01 -7.11919248e-01
-9.63164747e-01 -7.61890411e-02 -5.40105999e-01 1.25712782e-01
8.02878797e-01 -7.95621455e-01 -1.41201174e+00 3.90780389e-01
-1.34044814e+00 -3.28658581e-01 -3.72656673e-01 5.65072060e-01
-5.82799494e-01 4.36286777e-01 -6.83726192e-01 -9.85354483e-01
-5.73436439e-01 -1.14052188e+00 8.11657846e-01 -1.44804046e-02
-3.70034911e-02 -1.08734536e+00 3.59075010e-01 5.61250448e-02
5.95667183e-01 2.54187845e-02 9.00606692e-01 -3.50139201e-01
-3.73576105e-01 -4.20719720e-02 -1.09026246e-01 3.29936683e-01
-7.30421096e-02 -5.86817324e-01 -7.79262066e-01 -4.52539980e-01
4.58807528e-01 -5.83122075e-01 1.03705037e+00 5.20719469e-01
1.53753257e+00 -4.38286692e-01 -8.10971409e-02 9.22730446e-01
1.77790904e+00 2.21824095e-01 4.81117934e-01 -2.73631904e-02
7.15595365e-01 1.78751543e-01 4.96087596e-02 5.30944586e-01
2.25171655e-01 6.83108985e-01 2.08468407e-01 -5.15484214e-02
4.35723513e-01 -4.33603823e-02 1.89625069e-01 7.33242869e-01
-2.24814847e-01 -2.82645486e-02 -8.76162529e-01 1.50486916e-01
-1.79364812e+00 -1.10035837e+00 -1.71349376e-01 2.20857215e+00
1.15829778e+00 -1.03179432e-01 -1.23036623e-01 2.24361137e-01
3.34675521e-01 6.52693585e-02 -8.38929594e-01 -4.22776222e-01
1.40188292e-01 7.66629100e-01 5.23727357e-01 9.46759880e-01
-8.93839538e-01 9.52296138e-01 6.76518679e+00 1.08510804e+00
-1.56480217e+00 2.26777688e-01 4.78101879e-01 -5.16594946e-01
-2.45761618e-01 6.49682358e-02 -6.11552715e-01 5.71521342e-01
8.81735682e-01 -1.70260057e-01 1.18948293e+00 8.79794300e-01
1.09986663e-01 -2.28733420e-01 -9.23197389e-01 1.14403045e+00
-2.00623825e-01 -1.90623629e+00 -1.38222218e-01 1.86019212e-01
7.20532358e-01 2.04717785e-01 -2.52351493e-01 2.68344700e-01
1.14115752e-01 -1.22171557e+00 6.93915665e-01 3.07020098e-01
9.27957833e-01 -1.04307485e+00 5.48589349e-01 4.93465364e-01
-1.00592315e+00 1.18661955e-01 -3.62040132e-01 -5.11666954e-01
2.28248015e-01 8.03936958e-01 -1.86713174e-01 2.52373129e-01
5.68025708e-01 3.93866748e-01 1.00111939e-01 6.96729064e-01
1.69673547e-01 3.47953826e-01 -3.69857311e-01 -1.70571804e-01
3.60745907e-01 -6.81128085e-01 2.11701035e-01 9.12242234e-01
3.47889990e-01 2.85317272e-01 1.59896389e-01 1.18709433e+00
-1.25478789e-01 -5.98318642e-03 -5.23714840e-01 -1.90992758e-01
2.51295626e-01 1.05102420e+00 -4.57930356e-01 -2.76069164e-01
-3.34888011e-01 1.14537561e+00 7.19626129e-01 3.90943468e-01
-5.86765945e-01 -4.42750692e-01 5.29809654e-01 2.02569008e-01
6.52505755e-01 -3.58600050e-01 -2.26724789e-01 -1.41907060e+00
5.17893620e-02 -7.21995771e-01 -1.40711412e-01 -5.69365323e-01
-1.36958826e+00 3.31700414e-01 -1.53668284e-01 -6.04066312e-01
-7.85186514e-02 -7.82783568e-01 -6.60336912e-01 1.11249959e+00
-1.50902259e+00 -5.80617785e-01 2.04975214e-02 5.88963389e-01
4.52688895e-02 1.59464747e-01 1.01302850e+00 3.81783545e-01
-4.69168097e-01 3.48604053e-01 4.90104258e-01 -1.87611714e-01
1.72377959e-01 -1.22774446e+00 2.47481897e-01 5.42310834e-01
-1.97136804e-01 8.77100766e-01 6.71726227e-01 -5.40928245e-01
-1.78459191e+00 -7.02645421e-01 1.09988201e+00 2.52131701e-01
6.43990636e-01 -5.05894661e-01 -1.20258629e+00 3.54275137e-01
1.48679465e-01 4.75728571e-01 7.70913601e-01 -3.10868118e-02
3.98306586e-02 -6.32974878e-02 -1.21075463e+00 4.45411205e-01
1.03963435e+00 -6.88433051e-01 -1.76019445e-01 5.91117084e-01
4.78907853e-01 -3.86447638e-01 -6.45865142e-01 2.93635577e-01
3.95684183e-01 -1.09056723e+00 9.36684072e-01 -5.53596139e-01
3.66568655e-01 -1.90414980e-01 -3.46812695e-01 -1.15882766e+00
-4.23681796e-01 -8.62529457e-01 -3.72393459e-01 2.78221130e-01
1.45930558e-01 -7.61199296e-01 7.88379550e-01 5.17047763e-01
-2.07136735e-01 -1.35655677e+00 -1.26963258e+00 -5.29731095e-01
4.47312891e-01 -4.50164795e-01 2.27961093e-01 8.26199710e-01
1.26020074e-01 3.86420339e-01 -4.21747833e-01 -5.86943142e-02
7.37840295e-01 2.04191372e-01 2.26629496e-01 -1.34068704e+00
-6.55207932e-01 -1.74033776e-01 8.43500420e-02 -9.81221855e-01
4.32323426e-01 -1.10803187e+00 6.47502989e-02 -1.33820820e+00
2.97819436e-01 -2.16199175e-01 -3.10568005e-01 4.05771941e-01
1.36881739e-01 2.77494371e-01 -9.80570093e-02 1.07665099e-01
-6.07955575e-01 8.63936543e-01 1.06437314e+00 6.85021877e-02
-2.55845994e-01 -2.52769321e-01 -5.92113137e-01 8.95309567e-01
6.66681170e-01 -7.22487152e-01 -1.80569783e-01 -2.37557247e-01
6.57840669e-01 2.50134170e-01 5.87153554e-01 -9.21696961e-01
4.14248973e-01 -2.46914148e-01 2.45379582e-01 -5.37373424e-01
6.43567502e-01 -3.86792392e-01 6.96811229e-02 7.27860987e-01
-2.37317055e-01 -1.69135451e-01 2.28932098e-01 2.70258278e-01
-1.31860882e-01 -5.15407681e-01 1.01707780e+00 -4.02826011e-01
-1.25941381e-01 2.87572861e-01 -4.93946403e-01 -7.16860071e-02
6.32466197e-01 -1.05870962e-02 3.42214286e-01 -2.67747521e-01
-5.89661479e-01 -4.33401540e-02 5.67267537e-01 -7.55423009e-01
5.78593135e-01 -1.41526246e+00 -3.18066448e-01 2.88574219e-01
-4.60593253e-01 5.81242144e-02 3.02160800e-01 9.84241903e-01
-7.92460382e-01 5.52756369e-01 4.42453288e-02 -3.14963132e-01
-8.24351370e-01 4.21678603e-01 6.95063591e-01 -4.22389239e-01
-5.79229057e-01 8.83227766e-01 1.34007588e-01 -3.93015534e-01
2.63431370e-02 -7.89725035e-02 2.69304901e-01 -8.72445926e-02
5.24746895e-01 4.64899540e-01 1.36358753e-01 -4.32768613e-01
-4.63880211e-01 4.93029833e-01 -5.21435998e-02 -2.97919184e-01
1.65288341e+00 3.75806957e-01 -4.85325187e-01 3.47752899e-01
1.54591393e+00 4.36478928e-02 -1.20053315e+00 -1.63805768e-01
-2.57983774e-01 -2.20785774e-02 5.86874843e-01 -4.47216332e-01
-7.72502363e-01 1.09827721e+00 4.56796348e-01 1.24771334e-01
7.94731140e-01 -3.46247524e-01 9.56541121e-01 9.82917964e-01
4.76444930e-01 -9.52620089e-01 -3.20940286e-01 6.63312793e-01
6.44714415e-01 -1.20256639e+00 1.47681236e-01 -1.53189942e-01
-3.34172696e-01 1.17181325e+00 2.67059922e-01 -4.36521024e-01
6.89112484e-01 2.43241593e-01 -6.27083659e-01 -3.53406578e-01
-6.91713035e-01 1.52069911e-01 3.05245727e-01 -6.55586869e-02
5.30816555e-01 -2.45009735e-01 -5.08998454e-01 5.02775311e-01
-3.57381292e-02 4.42118108e-01 2.20086813e-01 9.09833610e-01
-5.76360822e-01 -1.16295266e+00 9.89651904e-02 2.88480073e-01
-4.32763129e-01 -4.41732943e-01 -1.81038514e-01 5.91804445e-01
1.39483541e-01 5.48430920e-01 -3.38383541e-02 -3.77000608e-02
-1.36662319e-01 3.58980447e-01 6.92572892e-01 -6.17766798e-01
-3.94843251e-01 -1.35927677e-01 -2.03876227e-01 -6.40116811e-01
-3.55667055e-01 -3.57360989e-01 -1.57017505e+00 -5.81020713e-01
-4.24384862e-01 6.11585498e-01 5.90500951e-01 1.22302055e+00
4.62912768e-01 2.18221799e-01 4.01855201e-01 -1.29029000e+00
-8.41673732e-01 -6.03954971e-01 -5.96553266e-01 -3.73684689e-02
4.63891715e-01 -6.95949554e-01 -7.70487010e-01 -4.58826125e-01] | [5.419877052307129, 5.066863536834717] |
f4ca912b-ab9a-4f37-bddb-a65a8cf872ce | alignot-an-optimal-transport-based-algorithm | 2210.09361 | null | https://arxiv.org/abs/2210.09361v1 | https://arxiv.org/pdf/2210.09361v1.pdf | AlignOT: An optimal transport based algorithm for fast 3D alignment with applications to cryogenic electron microscopy density maps | Aligning electron density maps from Cryogenic electron microscopy (cryo-EM) is a first key step for studying multiple conformations of a biomolecule. As this step remains costly and challenging, with standard alignment tools being potentially stuck in local minima, we propose here a new procedure, called AlignOT, which relies on the use of computational optimal transport (OT) to align EM maps in 3D space. By embedding a fast estimation of OT maps within a stochastic gradient descent algorithm, our method searches for a rotation that minimizes the Wasserstein distance between two maps, represented as point clouds. We quantify the impact of various parameters on the precision and accuracy of the alignment, and show that AlignOT can outperform the standard local alignment methods, with an increased range of rotation angles leading to proper alignment. We further benchmark AlignOT on various pairs of experimental maps, which account for different types of conformational heterogeneities and geometric properties. As our experiments show good performance, we anticipate that our method can be broadly applied to align 3D EM maps. | ['K. Dao Duc', 'A. Condon', 'F. Poitevin', 'G. Woollard', 'A. Tajmir Riahi'] | 2022-10-17 | null | null | null | null | ['cryogenic-electron-microscopy-cryo-em'] | ['computer-vision'] | [ 5.55555075e-02 -4.04540390e-01 9.24213827e-02 -3.63323510e-01
-8.26308489e-01 -7.13862836e-01 6.74419224e-01 2.78655171e-01
-6.57911062e-01 8.13578606e-01 -1.49481595e-01 -4.45280224e-01
-4.90037017e-02 -2.96912163e-01 -6.74465179e-01 -8.54775906e-01
-1.12121820e-01 9.13757026e-01 3.49809527e-01 9.88623966e-03
6.73371911e-01 7.87933767e-01 -9.19393957e-01 -3.55456211e-02
6.84510589e-01 4.71523315e-01 6.63263619e-01 5.24819970e-01
2.39754081e-01 -8.94753635e-02 -1.96022421e-01 -2.29134813e-01
9.47614238e-02 -3.74732584e-01 -1.14972377e+00 -2.26795599e-01
2.80174524e-01 8.46960098e-02 4.88112926e-01 9.91475821e-01
6.09645903e-01 1.16071500e-01 9.88505960e-01 -5.91345251e-01
-2.46252984e-01 -1.99018389e-01 -2.47183442e-01 2.41759121e-01
4.16689724e-01 7.84881562e-02 8.56045127e-01 -9.75979686e-01
1.27762318e+00 1.02791929e+00 6.48910880e-01 2.64355451e-01
-1.83635962e+00 -1.96515560e-01 -1.26015708e-01 3.06205958e-01
-1.34920084e+00 -4.17678446e-01 6.20088577e-01 -5.27532756e-01
1.28601015e+00 6.68037683e-02 5.94265878e-01 8.01748037e-01
8.33949745e-01 6.46019280e-02 1.28626597e+00 -4.28406447e-01
5.93986630e-01 -2.64743865e-01 -7.45413676e-02 4.04971689e-01
7.36095905e-02 -1.29155830e-01 -3.73418629e-01 -6.17439330e-01
5.34483850e-01 -6.39679655e-02 -1.54385254e-01 -1.06881738e+00
-1.40695643e+00 7.03679264e-01 4.01132107e-01 5.81671111e-02
-4.39939439e-01 -1.92274034e-01 1.33667722e-01 1.26518309e-01
2.96756357e-01 7.72365868e-01 -3.43571424e-01 -2.53722757e-01
-8.08235288e-01 5.30574560e-01 5.59406817e-01 5.27238607e-01
9.34318364e-01 -8.07296455e-01 5.20874918e-01 6.77717268e-01
3.74532491e-01 2.83699065e-01 1.49818361e-01 -9.45117116e-01
3.11761945e-01 4.59110886e-01 5.18740475e-01 -7.35644698e-01
-5.89930773e-01 3.97786707e-01 -5.26681483e-01 2.57732928e-01
3.78928274e-01 -4.89861146e-02 -7.18137741e-01 1.50675309e+00
6.40463889e-01 -2.92445630e-01 -1.51436001e-01 9.49947238e-01
7.96120986e-02 5.13834476e-01 3.92886586e-02 -5.38946807e-01
1.04948223e+00 -4.55383033e-01 -3.77120644e-01 -1.77204274e-02
7.02299356e-01 -9.69474256e-01 5.66393197e-01 1.98023051e-01
-1.17602849e+00 -1.25524268e-01 -1.17076576e+00 -1.87243316e-02
-2.56555796e-01 -2.29388684e-01 2.27303103e-01 2.04076409e-01
-9.74945486e-01 1.27527094e+00 -1.36882365e+00 -5.81827104e-01
2.15626545e-02 6.20867014e-01 -6.79277360e-01 2.34802678e-01
-6.00788534e-01 1.36535859e+00 3.18871081e-01 3.21209840e-02
-4.56994295e-01 -4.62633491e-01 -4.93950665e-01 -4.41910088e-01
-1.06899247e-01 -6.26795471e-01 9.22422349e-01 -1.75099388e-01
-1.59660959e+00 9.81419981e-01 -6.98540270e-01 -3.78539890e-01
4.53025758e-01 7.64626861e-02 5.37540838e-02 1.70343027e-01
-6.28498942e-02 6.44381702e-01 3.90968144e-01 -9.51990604e-01
1.02308188e-02 -5.34609258e-01 -5.19733429e-01 2.30534166e-01
2.89619833e-01 3.41367006e-01 1.53772578e-01 1.11296751e-01
6.55660689e-01 -1.18625867e+00 -5.56005657e-01 -2.55313277e-01
-3.33738744e-01 -6.35262206e-02 6.81412160e-01 -3.11570674e-01
7.36725628e-01 -1.62261748e+00 8.14638853e-01 4.29203272e-01
2.17945635e-01 5.29090576e-02 2.48130441e-01 7.83553958e-01
-5.41802831e-02 9.76505801e-02 -2.88682520e-01 -2.80626148e-01
-1.04320213e-01 8.54843184e-02 -6.70430586e-02 8.67116511e-01
2.18188018e-01 7.92962492e-01 -8.00750017e-01 -3.84944946e-01
4.11161155e-01 5.17144978e-01 -4.99829590e-01 1.60951257e-01
-1.93130314e-01 8.12277913e-01 -3.75979751e-01 2.92703778e-01
8.76089275e-01 -5.17275453e-01 7.88317084e-01 -1.98033690e-01
-5.72757900e-01 6.36527419e-01 -9.59182978e-01 1.54473495e+00
8.37663189e-02 2.90552020e-01 1.18453301e-01 -7.90038645e-01
1.10954154e+00 6.96650445e-02 6.37762189e-01 -4.91659462e-01
-6.15570955e-02 4.35694814e-01 1.44182488e-01 -1.96328893e-01
3.52219939e-01 -3.89048517e-01 3.32603335e-01 5.95213592e-01
7.72397146e-02 -2.25606620e-01 3.00310850e-02 -3.97428200e-02
9.45610523e-01 3.63110840e-01 4.03612047e-01 -8.23842108e-01
4.49673593e-01 1.40553966e-01 3.86748642e-01 2.63504863e-01
-2.42082298e-01 7.16289520e-01 3.62905174e-01 -8.51197362e-01
-1.69544733e+00 -1.07128036e+00 -6.20823026e-01 5.08450329e-01
3.64853710e-01 -6.17477119e-01 -8.41017902e-01 -1.81980446e-01
-1.08333277e-02 5.60868233e-02 -3.70376766e-01 7.88621604e-02
-7.60566831e-01 -1.30647814e+00 -9.18302014e-02 1.57855809e-01
-1.61005229e-01 -1.01962888e+00 -8.05131257e-01 3.75929505e-01
2.92457659e-02 -9.85588312e-01 -2.14982569e-01 5.55486619e-01
-1.10282743e+00 -1.12734270e+00 -5.90375125e-01 -5.31658113e-01
8.24331343e-01 3.24642897e-01 1.03963685e+00 1.02140345e-02
-1.86110213e-01 -1.78850204e-01 -9.25989673e-02 6.10083044e-02
-5.01021981e-01 3.05146694e-01 4.48732615e-01 -3.51310849e-01
6.01958871e-01 -8.79778564e-01 -8.42641175e-01 7.85951138e-01
-4.07543123e-01 1.68347701e-01 2.86058336e-01 8.22861969e-01
1.03379333e+00 -6.34120643e-01 1.84210703e-01 -5.26353002e-01
6.47926569e-01 -3.39331090e-01 -8.86818647e-01 1.25744298e-01
-6.70136511e-01 4.51678544e-01 6.47892535e-01 -1.02807678e-01
-5.75611889e-01 2.70316780e-01 -1.74408540e-01 -1.38487652e-01
-2.06574395e-01 2.04855978e-01 -4.49682064e-02 -4.74638641e-01
5.57876170e-01 2.91261047e-01 2.53720701e-01 -6.27292097e-01
4.62004207e-02 4.43721563e-01 1.89902306e-01 -8.62706482e-01
3.58312875e-01 6.58635855e-01 4.55891162e-01 -7.95919895e-01
-2.78153002e-01 -5.04907370e-01 -1.17851877e+00 -5.83427250e-02
9.36548769e-01 -3.17126811e-01 -1.04632998e+00 1.07992075e-01
-1.17284489e+00 -2.18207657e-01 4.73237097e-01 5.73034346e-01
-9.02688324e-01 7.82376170e-01 -4.67591465e-01 -4.40459341e-01
-2.28761569e-01 -1.86084747e+00 1.24696326e+00 -1.52449101e-01
-5.76492369e-01 -9.67518806e-01 7.89356709e-01 8.43831673e-02
2.89655656e-01 2.94901192e-01 1.10182023e+00 -4.68601167e-01
-6.30120516e-01 2.12238207e-02 1.05308123e-01 -1.92006528e-01
8.44358802e-02 3.18824798e-01 -5.32958388e-01 -4.91872638e-01
-7.27466941e-02 -2.00407445e-01 5.79847395e-01 4.15490925e-01
4.81352895e-01 -1.05832368e-02 -6.94649875e-01 7.51422346e-01
1.46690321e+00 1.87610224e-01 7.28674352e-01 7.41875172e-01
4.48333263e-01 5.67675114e-01 5.20095468e-01 1.95369452e-01
9.12818536e-02 1.09004569e+00 3.96953464e-01 2.24746108e-01
6.08671725e-01 -4.56811376e-02 2.30017558e-01 9.87925529e-01
-5.41835129e-01 2.01605380e-01 -1.05990434e+00 2.97933996e-01
-1.87387800e+00 -9.45190549e-01 -4.93784584e-02 2.56458735e+00
8.00810039e-01 -1.37495324e-01 2.40456983e-01 -3.67897809e-01
6.11680150e-01 2.20149104e-02 -7.09766924e-01 -4.65921402e-01
5.59254847e-02 2.81078666e-01 5.42118192e-01 6.58328176e-01
-8.37761283e-01 7.26613045e-01 7.88018894e+00 2.45452046e-01
-1.17255640e+00 -1.67541325e-01 3.19126099e-01 -1.45799011e-01
-7.99127147e-02 3.30742806e-01 -7.91117549e-01 7.03241110e-01
1.10443008e+00 -1.18388437e-01 4.34604585e-01 6.67090893e-01
4.57845956e-01 -1.20162085e-01 -1.14914906e+00 7.85504401e-01
-4.82040465e-01 -1.66541398e+00 -8.61495435e-02 4.33365971e-01
6.88389778e-01 4.59757715e-01 -2.98451960e-01 -5.44432580e-01
3.18124533e-01 -9.90290582e-01 3.95467728e-01 4.44459945e-01
3.95544976e-01 -9.04862761e-01 6.24657273e-01 3.20201963e-01
-1.02771008e+00 6.22313023e-01 -8.50472927e-01 1.52646050e-01
5.02677739e-01 4.47700620e-01 -1.04519093e+00 3.42850447e-01
5.85793614e-01 5.83748817e-01 -1.21617936e-01 8.16932917e-01
1.37752831e-01 -8.54797289e-02 -5.08592725e-01 -1.01330750e-01
8.84564891e-02 -9.87265348e-01 5.48678577e-01 9.13630068e-01
2.14818090e-01 -1.57210499e-01 9.29379016e-02 1.00712955e+00
1.05992205e-01 8.52883980e-02 -4.92392570e-01 8.45181197e-02
6.12780333e-01 1.17091119e+00 -8.47985387e-01 7.70136714e-02
-1.86250545e-03 8.64921510e-01 5.77595174e-01 1.28267869e-01
-4.87089992e-01 -5.78794405e-02 9.48790908e-01 2.13504896e-01
4.55489278e-01 -6.94648027e-01 8.24294388e-02 -1.03567946e+00
1.97618783e-01 -5.94796479e-01 -1.17686629e-01 -7.06964791e-01
-9.74078834e-01 4.23872292e-01 3.25228274e-02 -1.02007639e+00
-4.06640679e-01 -9.03835773e-01 -5.35854161e-01 8.94861460e-01
-1.18146622e+00 -5.31638205e-01 1.19817704e-01 -1.49839539e-02
6.60227612e-02 2.33733103e-01 8.51106942e-01 -1.22244120e-01
-4.34865147e-01 6.51072711e-02 6.44197106e-01 -5.69980800e-01
8.76282632e-01 -1.32505763e+00 7.73777723e-01 4.85713720e-01
-1.54093981e-01 1.09557605e+00 1.00350487e+00 -7.07039237e-01
-1.52992713e+00 -6.91647172e-01 8.54785621e-01 -7.83115089e-01
6.86196506e-01 -4.34717923e-01 -1.03496587e+00 7.17822731e-01
1.25727147e-01 -1.87287405e-01 6.68418109e-01 9.68501903e-03
1.61162112e-02 4.48024035e-01 -1.20316017e+00 6.28749013e-01
9.33301210e-01 -5.06803513e-01 -5.49595535e-01 5.47602355e-01
2.31530935e-01 -4.01157469e-01 -1.31975996e+00 2.89160788e-01
6.00775659e-01 -1.08095884e+00 1.06026614e+00 -8.34238887e-01
1.44897282e-01 -5.37528396e-01 -9.54166204e-02 -1.30133843e+00
-3.67727607e-01 -8.62159610e-01 1.16447285e-01 5.12824774e-01
4.11922187e-01 -7.19413579e-01 8.32972527e-01 5.19152582e-01
-7.23421946e-02 -8.44647408e-01 -1.21692848e+00 -7.24974692e-01
3.53630245e-01 2.22679190e-02 6.11016095e-01 8.42475474e-01
4.56772447e-01 2.59144336e-01 -2.14272574e-01 1.25190839e-01
7.65048206e-01 2.27908432e-01 7.69117415e-01 -1.25867081e+00
8.92294198e-02 -1.63101166e-01 -6.36556387e-01 -1.01620030e+00
1.59247860e-01 -8.37276459e-01 7.44875371e-02 -1.16386485e+00
5.00252008e-01 -2.71505356e-01 -5.16880527e-02 1.13029063e-01
4.43342142e-02 1.05986051e-01 -5.62386662e-02 6.01314366e-01
-7.21844256e-01 5.63109756e-01 1.12410641e+00 2.66996443e-01
-1.93796739e-01 -3.74503493e-01 -1.29515320e-01 5.56077480e-01
6.31947398e-01 -6.71743870e-01 1.70549303e-01 -6.02595881e-02
4.40458238e-01 -1.58167034e-01 1.90139607e-01 -8.84239674e-01
2.84938544e-01 -2.95692384e-01 2.08391637e-01 -7.67604411e-01
3.67643684e-01 -6.71006680e-01 4.85917985e-01 3.26453388e-01
9.72594041e-03 5.18301368e-01 -3.32929939e-02 3.17910075e-01
7.68126771e-02 -2.57533491e-01 1.01967061e+00 -8.93363953e-02
-2.72301912e-01 1.62223458e-01 -4.33382213e-01 -2.60641694e-01
8.20635438e-01 -3.34403068e-01 -2.90579736e-01 1.71428517e-01
-7.87926197e-01 1.17029198e-01 1.35710776e+00 -1.59733683e-01
4.95282561e-01 -1.18189526e+00 -3.28564584e-01 2.34104246e-01
1.63275097e-02 4.15015742e-02 -5.49024437e-03 1.16714847e+00
-1.02290905e+00 6.53346598e-01 -4.76365894e-01 -1.07106888e+00
-1.37915778e+00 4.06094551e-01 6.35376036e-01 -1.75239056e-01
-6.33182287e-01 2.09428892e-01 -1.26455605e-01 -6.95519328e-01
-4.76033270e-01 -2.84611970e-01 2.20829651e-01 -4.47228789e-01
4.95252728e-01 3.57612997e-01 4.41220433e-01 -7.06176877e-01
-6.15500689e-01 9.54369485e-01 -4.86845344e-01 -3.29770194e-03
1.63052392e+00 -1.39218852e-01 -3.49603713e-01 1.16828844e-01
1.26791906e+00 -2.12579012e-01 -1.45956600e+00 1.77090883e-01
1.89979047e-01 -2.47465685e-01 -2.95547485e-01 -1.60190597e-01
-2.90268660e-01 6.98312581e-01 5.76312482e-01 -2.59504002e-02
4.31300849e-01 1.51457250e-01 5.53166270e-01 7.10032344e-01
7.45278537e-01 -1.00070977e+00 -3.84783983e-01 4.36134040e-01
4.37237054e-01 -1.18367124e+00 3.75465721e-01 -6.96791932e-02
-2.77214259e-01 1.22784615e+00 9.99190882e-02 -2.79911011e-01
1.77680507e-01 1.94427997e-01 2.72172596e-02 -3.84609818e-01
-7.77663410e-01 2.61202604e-01 1.56133939e-02 5.05242944e-01
4.09573138e-01 -9.98518914e-02 -3.71628642e-01 -2.61000276e-01
-2.14980245e-01 -2.15112388e-01 3.30352336e-01 1.21824086e+00
-6.34405971e-01 -1.52613127e+00 -4.16429877e-01 5.60107790e-02
-4.15584385e-01 1.86471894e-01 -7.54130661e-01 7.13621020e-01
-4.18399453e-01 4.39232826e-01 1.18185110e-01 -2.35486045e-01
1.03751920e-01 2.23022535e-01 7.33539343e-01 -2.36916676e-01
-1.67456925e-01 -7.47064352e-02 -1.08875640e-01 -6.54493868e-01
-6.64033830e-01 -8.12343597e-01 -1.24791336e+00 -6.07352734e-01
-4.34894919e-01 4.88673598e-01 9.17114556e-01 1.12793815e+00
6.90911412e-01 -2.84921080e-01 5.26944458e-01 -1.31309986e+00
-4.95278388e-01 -8.12812090e-01 -3.69439006e-01 4.75762010e-01
2.74470091e-01 -8.31865668e-01 -4.66081381e-01 -7.29883164e-02] | [13.300750732421875, -3.065548896789551] |
57e51737-2919-4ca2-bcdc-01f408c0e0dd | aiatrack-attention-in-attention-for | 2207.09603 | null | https://arxiv.org/abs/2207.09603v2 | https://arxiv.org/pdf/2207.09603v2.pdf | AiATrack: Attention in Attention for Transformer Visual Tracking | Transformer trackers have achieved impressive advancements recently, where the attention mechanism plays an important role. However, the independent correlation computation in the attention mechanism could result in noisy and ambiguous attention weights, which inhibits further performance improvement. To address this issue, we propose an attention in attention (AiA) module, which enhances appropriate correlations and suppresses erroneous ones by seeking consensus among all correlation vectors. Our AiA module can be readily applied to both self-attention blocks and cross-attention blocks to facilitate feature aggregation and information propagation for visual tracking. Moreover, we propose a streamlined Transformer tracking framework, dubbed AiATrack, by introducing efficient feature reuse and target-background embeddings to make full use of temporal references. Experiments show that our tracker achieves state-of-the-art performance on six tracking benchmarks while running at a real-time speed. | ['Junsong Yuan', 'Xinggang Wang', 'Chao Ma', 'Chunluan Zhou', 'Shenyuan Gao'] | 2022-07-20 | null | null | null | null | ['visual-tracking', 'visual-object-tracking'] | ['computer-vision', 'computer-vision'] | [-2.38563359e-01 -5.78316450e-01 -2.37890244e-01 -1.51414782e-01
-5.77618241e-01 -2.58501172e-01 7.16503799e-01 -1.29804075e-01
-2.85333425e-01 2.32330456e-01 3.78609061e-01 -7.86759555e-02
7.50666335e-02 -5.51849961e-01 -5.29171169e-01 -5.95996797e-01
2.77518528e-03 8.54658112e-02 6.79366410e-01 -1.41345650e-01
-9.98239219e-02 3.59977752e-01 -1.46202779e+00 -1.47498339e-01
8.23936284e-01 1.09127164e+00 1.33734411e-02 3.80302340e-01
-2.96627402e-01 6.89525247e-01 -5.37809074e-01 -5.10071576e-01
4.22720701e-01 -1.28128514e-01 -6.94302842e-02 -8.17156360e-02
6.82083070e-01 -2.64990270e-01 -6.14347041e-01 1.24601388e+00
4.92653191e-01 -1.00488476e-01 5.74543476e-02 -1.43534279e+00
-9.75457668e-01 4.28212553e-01 -9.76443172e-01 6.33521616e-01
5.67283342e-03 5.75755298e-01 1.13705051e+00 -1.14166081e+00
1.96811631e-01 1.43681347e+00 7.73742497e-01 3.85824680e-01
-1.17324257e+00 -1.11965358e+00 5.79250872e-01 3.24758053e-01
-1.28126788e+00 -3.63466948e-01 5.75493991e-01 -3.33632678e-01
6.81579590e-01 1.59806803e-01 9.04243469e-01 8.17789674e-01
1.81622416e-01 1.08465040e+00 7.35649824e-01 4.01594117e-02
-1.45816475e-01 -2.20273033e-01 2.41449490e-01 5.42569935e-01
4.79258865e-01 3.80007982e-01 -6.63932502e-01 -1.33077130e-02
7.91493773e-01 2.36332804e-01 -1.99721843e-01 -3.30310345e-01
-1.33951271e+00 6.85428023e-01 8.74199986e-01 1.21058516e-01
-3.48559409e-01 3.54944497e-01 4.08433259e-01 1.67020947e-01
3.81747842e-01 1.60217494e-01 -1.42436221e-01 -8.56406987e-02
-6.91903949e-01 9.94067639e-02 1.07176490e-01 9.82914567e-01
5.76022029e-01 2.20973387e-01 -7.88477540e-01 5.93944907e-01
5.52093089e-01 8.20069790e-01 3.89099300e-01 -6.32911623e-01
3.60383362e-01 8.34892273e-01 6.92465901e-02 -1.08679152e+00
-2.10741624e-01 -8.34886432e-01 -6.48667216e-01 2.58119702e-01
1.90802768e-01 3.21364552e-02 -9.62657154e-01 1.85712433e+00
6.24401748e-01 4.88138735e-01 -3.57287288e-01 1.22887683e+00
6.47182226e-01 3.65734279e-01 2.24206701e-01 -1.31195754e-01
1.52938092e+00 -1.16805696e+00 -9.24541533e-01 -3.07944387e-01
1.59755990e-01 -8.09515595e-01 8.99113595e-01 -6.96237907e-02
-8.86088908e-01 -8.61176789e-01 -1.01202548e+00 -1.03762917e-01
-1.71701118e-01 1.03253789e-01 6.79579616e-01 4.45070565e-01
-7.82709718e-01 3.65795523e-01 -1.15555561e+00 -2.41486087e-01
6.39189124e-01 4.61669385e-01 -9.74434242e-02 2.75633305e-01
-1.00714338e+00 6.98650360e-01 -2.33498253e-02 3.93681288e-01
-7.59519994e-01 -8.09219539e-01 -7.54102826e-01 1.85724258e-01
5.46101570e-01 -5.92964351e-01 1.11591768e+00 -7.35761225e-01
-1.29516447e+00 3.09701562e-01 -2.16887802e-01 -5.14991164e-01
3.76040518e-01 -6.97238684e-01 -4.82314616e-01 -3.26213539e-01
1.24390692e-01 6.87290430e-01 9.20994222e-01 -8.85280669e-01
-7.82868505e-01 -3.54329258e-01 -9.84048173e-02 -4.94841300e-02
-6.38091326e-01 1.55002683e-01 -1.06597984e+00 -9.87199664e-01
-5.38231619e-02 -9.27611768e-01 -1.58295915e-01 4.31007743e-01
-2.73660898e-01 -2.39074439e-01 1.12377644e+00 -4.39182252e-01
1.53052020e+00 -2.23347640e+00 1.26463538e-02 5.76655492e-02
4.43104327e-01 6.07735813e-01 -1.63859561e-01 1.09170415e-02
8.75546783e-02 -2.14024097e-01 1.61881313e-01 -4.85474259e-01
1.66340768e-01 1.71937063e-01 -3.18406075e-01 4.69494671e-01
5.06812453e-01 1.05471647e+00 -1.10301888e+00 -6.19373918e-01
2.38421723e-01 6.53730571e-01 -6.98400736e-01 1.43701017e-01
-1.86147332e-01 4.02615219e-01 -7.95385599e-01 7.20108032e-01
7.53827989e-01 -4.30195570e-01 -1.26387728e-02 -5.08433223e-01
-4.45574969e-01 8.65263268e-02 -9.28665340e-01 1.44830620e+00
-8.42830241e-02 7.81868160e-01 -1.15756191e-01 -4.07313406e-01
8.78773868e-01 -7.36916736e-02 6.63099527e-01 -7.52240181e-01
1.74975038e-01 -1.47513792e-01 3.87198776e-01 2.42367759e-02
7.40954459e-01 4.00463551e-01 1.91239983e-01 2.85318196e-01
-1.45720750e-01 7.70714581e-01 9.81372371e-02 3.05760026e-01
1.12774229e+00 3.21540028e-01 8.19743648e-02 -2.72443116e-01
6.31917000e-01 -2.88580079e-02 1.02119434e+00 6.37252390e-01
-4.83266920e-01 3.36178303e-01 2.17153460e-01 -5.14677644e-01
-7.37202823e-01 -9.61544037e-01 1.23271033e-01 1.31147158e+00
4.35175568e-01 -8.09943080e-01 -3.45318198e-01 -7.35751927e-01
2.25683674e-01 8.14435780e-02 -8.25859070e-01 -3.03414583e-01
-6.20276511e-01 -5.73404908e-01 2.86900908e-01 9.67162549e-01
5.35266280e-01 -8.74597847e-01 -5.78833997e-01 2.76955754e-01
8.07339028e-02 -9.88834918e-01 -1.12719285e+00 -1.14086442e-01
-6.07468665e-01 -1.04154563e+00 -5.42599499e-01 -2.86338806e-01
6.24775469e-01 7.26823568e-01 8.56816292e-01 2.51044720e-01
-1.22920580e-01 2.62210190e-01 -2.91824996e-01 -3.60618532e-01
2.03496784e-01 -4.63420972e-02 7.68867508e-02 2.97749370e-01
6.01950526e-01 -3.79643857e-01 -7.87350476e-01 4.74605948e-01
-6.13141179e-01 -6.50450364e-02 6.51319444e-01 9.76149797e-01
4.95665789e-01 -4.27800387e-01 2.69228011e-01 -2.95249820e-01
3.38780791e-01 -2.31283337e-01 -9.47847962e-01 3.20881069e-01
-5.73344111e-01 3.52444410e-01 4.96250331e-01 -6.15283966e-01
-9.23092782e-01 1.10053495e-01 5.04994430e-02 -1.07323122e+00
5.53616881e-01 1.08782902e-01 -2.51183450e-01 -3.42279732e-01
1.99090511e-01 2.30565935e-01 -2.59399395e-02 -5.49723685e-01
3.16237748e-01 3.07332546e-01 6.96478546e-01 -3.91521633e-01
1.24667239e+00 5.42847455e-01 -2.16290757e-01 -3.86776149e-01
-8.92990112e-01 -5.54260612e-01 -3.50154132e-01 -3.71392220e-01
6.73687756e-01 -1.01262116e+00 -7.62482584e-01 3.55873942e-01
-1.02181613e+00 1.43665805e-01 -2.42710829e-01 5.41358292e-01
8.51820260e-02 3.03752959e-01 -4.80448574e-01 -7.89672852e-01
-5.00649929e-01 -1.36435926e+00 1.18450677e+00 5.62537611e-01
2.16016397e-02 -5.53923070e-01 9.71910879e-02 -2.36393027e-02
6.41779304e-01 -1.08394012e-01 1.25589356e-01 -5.03004313e-01
-1.11799896e+00 1.90553684e-02 -6.06841862e-01 3.55577730e-02
1.71471491e-01 1.21725217e-01 -9.78186131e-01 -4.79799241e-01
-5.17050564e-01 -4.95435903e-03 1.04767799e+00 2.35545158e-01
9.17918205e-01 -6.17427053e-03 -5.70087373e-01 7.38240659e-01
1.06067085e+00 -1.11705810e-02 5.17013311e-01 3.72701079e-01
9.09548223e-01 -3.60236205e-02 7.83345699e-01 4.62291002e-01
4.08354729e-01 1.07948947e+00 4.54313070e-01 -2.07083419e-01
-3.93664688e-01 -2.69931138e-01 5.07239759e-01 8.54336500e-01
6.04547467e-03 5.07529713e-02 -5.97396851e-01 5.96682906e-01
-2.04020452e+00 -9.18431997e-01 -1.13956198e-01 2.30552697e+00
5.64679623e-01 4.05247808e-01 2.84088135e-01 -2.98582613e-01
8.57251167e-01 3.38619500e-01 -6.89573526e-01 2.12848037e-01
1.41405955e-01 2.42995992e-02 5.50904810e-01 3.01816225e-01
-1.18659043e+00 9.60877359e-01 6.00319815e+00 8.29767168e-01
-1.24492311e+00 1.83811337e-01 1.20739274e-01 -1.39625490e-01
-1.99533150e-01 -8.25744960e-03 -1.09514022e+00 7.67262638e-01
4.83760655e-01 -2.30760261e-01 4.64420989e-02 1.00250089e+00
9.60613266e-02 2.90126234e-01 -7.97991276e-01 9.60532606e-01
-3.02049611e-02 -1.20027709e+00 1.57133359e-02 9.55381915e-02
4.06013042e-01 2.74389535e-01 1.35143176e-01 5.45555472e-01
4.81838852e-01 -4.47406203e-01 6.82454228e-01 3.75456929e-01
4.59579736e-01 -6.04400754e-01 5.81804335e-01 -1.73916489e-01
-1.94413614e+00 6.30870014e-02 -3.31173509e-01 2.38698512e-01
1.68344632e-01 4.04020011e-01 -3.85114551e-01 6.83826387e-01
8.36689353e-01 9.41836894e-01 -8.05127680e-01 1.49317718e+00
-1.42460376e-01 4.40595925e-01 -4.52525645e-01 -4.15131561e-02
2.97417790e-01 4.06249091e-02 8.08522701e-01 1.14666975e+00
2.76504815e-01 -1.56264663e-01 4.21673030e-01 6.63237393e-01
-1.32097059e-03 -1.42702341e-01 -1.59087464e-01 5.11126965e-02
6.52398407e-01 1.42355359e+00 -6.96971893e-01 -3.56023878e-01
-6.78406537e-01 6.55195296e-01 3.46424490e-01 9.31506231e-02
-1.40820360e+00 -2.37614483e-01 1.22768033e+00 -4.22739983e-02
8.40202749e-01 -1.73199028e-01 4.41859961e-02 -1.26047087e+00
4.20526937e-02 -6.90111578e-01 5.12534142e-01 -6.51257992e-01
-1.27947104e+00 6.96846604e-01 -2.46346340e-01 -1.64951396e+00
3.86518270e-01 -4.13495928e-01 -7.17219114e-01 6.52126908e-01
-1.54167747e+00 -1.19321680e+00 -6.29732251e-01 6.02930903e-01
4.79633540e-01 -7.16685951e-02 1.66913033e-01 7.41258204e-01
-8.87652397e-01 8.88107836e-01 -1.12048835e-01 2.08349138e-01
9.45410073e-01 -1.01458132e+00 7.59870827e-01 1.09782994e+00
2.36809596e-01 9.04034257e-01 5.66123962e-01 -7.53446341e-01
-1.68531585e+00 -1.45224428e+00 6.57031238e-01 -3.29391688e-01
9.78229940e-01 -2.95851111e-01 -1.00232923e+00 7.48728096e-01
2.62630224e-01 4.48212296e-01 3.65680158e-01 2.79868066e-01
-6.19098246e-01 -3.95783007e-01 -5.43711841e-01 7.57253170e-01
1.38976288e+00 -3.43350232e-01 -5.33459425e-01 3.55119854e-02
7.24182367e-01 -5.27161956e-01 -7.84167826e-01 3.80289882e-01
6.76067352e-01 -8.13560665e-01 1.03027534e+00 -4.02594358e-01
-1.40524045e-01 -8.28372180e-01 2.40955576e-01 -1.02914441e+00
-7.72655904e-01 -8.63737702e-01 -4.16282386e-01 1.39394844e+00
1.27862558e-01 -6.56173944e-01 6.54601038e-01 2.92730153e-01
-2.13324770e-01 -5.94445944e-01 -7.88257539e-01 -9.51409876e-01
-4.77782667e-01 -3.77952605e-01 7.70231187e-01 7.31435895e-01
-3.19318652e-01 3.25969577e-01 -6.47881150e-01 2.52429396e-01
8.59261632e-01 3.56754929e-01 9.77867186e-01 -1.19043362e+00
-2.99191862e-01 -4.95419025e-01 -5.81520438e-01 -1.45537412e+00
-1.48926422e-01 -5.68470836e-01 1.14197567e-01 -1.02699232e+00
2.74216264e-01 -5.42346776e-01 -6.41647756e-01 5.54455400e-01
-6.14877462e-01 3.97069216e-01 5.02993286e-01 3.61744940e-01
-1.22141659e+00 9.11182642e-01 1.14145434e+00 -1.58236429e-01
2.89439596e-02 -2.54928231e-01 -7.94636548e-01 5.45618296e-01
4.52857137e-01 -4.61426765e-01 -1.45227879e-01 -5.61430514e-01
-2.42885455e-01 -4.39120978e-01 3.88339520e-01 -1.24674904e+00
6.70217752e-01 6.92045093e-02 4.69993621e-01 -8.92451465e-01
3.50183696e-01 -8.69173348e-01 2.42252126e-01 4.57256615e-01
-6.69225529e-02 3.45741302e-01 4.96223271e-01 7.25868106e-01
-2.74649620e-01 5.21808088e-01 6.18525803e-01 3.25990558e-01
-7.66741753e-01 7.38358319e-01 2.65446934e-03 6.86127916e-02
8.96068931e-01 -1.23228937e-01 -5.32723248e-01 -1.32423386e-01
-3.39281112e-01 7.18115449e-01 5.13397932e-01 7.45704830e-01
4.15800363e-01 -1.87785590e+00 -6.05197668e-01 3.73123169e-01
2.50202119e-01 -4.42978442e-01 2.73492664e-01 1.08787274e+00
6.42126650e-02 3.86175990e-01 -1.95222259e-01 -8.69123697e-01
-1.43149412e+00 6.95807457e-01 1.44438431e-01 -3.38636011e-01
-7.71400630e-01 6.97892904e-01 4.76073682e-01 1.24459088e-01
3.92425716e-01 -4.41981375e-01 -1.43732533e-01 -1.98654756e-01
9.27279353e-01 1.13666221e-01 -2.89588720e-01 -7.06930935e-01
-6.65256083e-01 8.21507633e-01 -3.08887839e-01 3.22717726e-01
9.22327459e-01 -4.97166403e-02 2.63578445e-01 -7.57750943e-02
7.68530548e-01 2.54496455e-01 -1.63356280e+00 -5.15859842e-01
4.24161293e-02 -8.10002327e-01 2.46708497e-01 -4.33403909e-01
-1.49135149e+00 6.69602394e-01 6.18246019e-01 1.40843362e-01
1.16765225e+00 -1.55272037e-01 9.92285371e-01 8.25812221e-02
1.37804791e-01 -6.99247241e-01 9.58912000e-02 4.62831974e-01
7.33667672e-01 -1.33793080e+00 1.07713409e-01 -3.55690360e-01
-5.85660517e-01 8.48521829e-01 1.04279482e+00 -1.67664304e-01
5.07186592e-01 5.32773554e-01 4.73850481e-02 -1.44630164e-01
-8.85182917e-01 -5.78037977e-01 5.28109252e-01 5.23847163e-01
3.31080377e-01 -3.55691433e-01 -2.10537717e-01 4.82741058e-01
3.48396480e-01 -5.14027737e-02 -2.14702338e-01 8.28087389e-01
-4.35543329e-01 -1.17626727e+00 -5.23780763e-01 2.24139065e-01
-3.96393985e-01 -1.08276159e-01 -3.59004200e-01 9.33762193e-01
7.34555796e-02 6.15346611e-01 1.14957653e-01 -6.56035125e-01
4.57213253e-01 -3.81363988e-01 2.96973854e-01 -1.28246114e-01
-9.06047463e-01 3.46112818e-01 -9.52984318e-02 -9.18572724e-01
-5.05533099e-01 -6.36759400e-01 -9.77043152e-01 -3.56866449e-01
-6.00032747e-01 2.10046187e-01 1.71941027e-01 7.13232517e-01
6.97376430e-01 8.67697060e-01 4.50749904e-01 -7.89011598e-01
-5.19445896e-01 -8.95595610e-01 -4.98402789e-02 2.99380094e-01
4.18664038e-01 -1.17834878e+00 -7.91939795e-02 -2.35152051e-01] | [6.275995254516602, -2.128560781478882] |
49cabfc7-1150-422e-ad18-cb3ef4ef489d | towards-robust-text-prompted-semantic | 2304.14672 | null | https://arxiv.org/abs/2304.14672v1 | https://arxiv.org/pdf/2304.14672v1.pdf | Towards Robust Text-Prompted Semantic Criterion for In-the-Wild Video Quality Assessment | The proliferation of videos collected during in-the-wild natural settings has pushed the development of effective Video Quality Assessment (VQA) methodologies. Contemporary supervised opinion-driven VQA strategies predominantly hinge on training from expensive human annotations for quality scores, which limited the scale and distribution of VQA datasets and consequently led to unsatisfactory generalization capacity of methods driven by these data. On the other hand, although several handcrafted zero-shot quality indices do not require training from human opinions, they are unable to account for the semantics of videos, rendering them ineffective in comprehending complex authentic distortions (e.g., white balance, exposure) and assessing the quality of semantic content within videos. To address these challenges, we introduce the text-prompted Semantic Affinity Quality Index (SAQI) and its localized version (SAQI-Local) using Contrastive Language-Image Pre-training (CLIP) to ascertain the affinity between textual prompts and visual features, facilitating a comprehensive examination of semantic quality concerns without the reliance on human quality annotations. By amalgamating SAQI with existing low-level metrics, we propose the unified Blind Video Quality Index (BVQI) and its improved version, BVQI-Local, which demonstrates unprecedented performance, surpassing existing zero-shot indices by at least 24\% on all datasets. Moreover, we devise an efficient fine-tuning scheme for BVQI-Local that jointly optimizes text prompts and final fusion weights, resulting in state-of-the-art performance and superior generalization ability in comparison to prevalent opinion-driven VQA methods. We conduct comprehensive analyses to investigate different quality concerns of distinct indices, demonstrating the effectiveness and rationality of our design. | ['Weisi Lin', 'Qiong Yan', 'Wenxiu Sun', 'Jingwen Hou', 'Chaofeng Chen', 'Annan Wang', 'Liang Liao', 'HaoNing Wu'] | 2023-04-28 | null | null | null | null | ['video-quality-assessment', 'video-quality-assessment'] | ['computer-vision', 'time-series'] | [ 2.63885617e-01 -4.70140874e-01 7.35540986e-02 -4.16323513e-01
-1.11262536e+00 -7.01861441e-01 5.13472736e-01 1.02506116e-01
-4.23745304e-01 3.48884851e-01 5.16201913e-01 -1.09699339e-01
-4.32237685e-01 -4.62728590e-01 -4.69709963e-01 -5.72319508e-01
6.62726611e-02 -7.60260373e-02 1.85080752e-01 -3.83527607e-01
3.11899453e-01 9.64631215e-02 -1.83470607e+00 3.13425958e-01
1.25640023e+00 1.36131239e+00 1.99030694e-02 5.73747635e-01
8.33849534e-02 7.48843193e-01 -8.52290988e-01 -9.56042886e-01
2.25988060e-01 -5.46007097e-01 -5.06532967e-01 2.79428482e-01
7.14585066e-01 -5.61603367e-01 -2.68930674e-01 1.14985013e+00
6.53094649e-01 4.00812961e-02 6.58065617e-01 -1.39125907e+00
-9.13208187e-01 1.36180013e-01 -2.44255975e-01 4.19314206e-01
7.34333575e-01 5.43345749e-01 1.31722164e+00 -1.04429877e+00
6.30831182e-01 1.09371483e+00 6.22076035e-01 4.79654938e-01
-1.06555092e+00 -3.92693371e-01 2.90681440e-02 4.69342053e-01
-1.12799823e+00 -5.75658023e-01 8.03564310e-01 -5.96883774e-01
7.87192345e-01 1.68338388e-01 6.32229567e-01 1.22429729e+00
-2.00111456e-02 7.63728797e-01 1.21531236e+00 -2.04069257e-01
4.26378310e-01 -5.01484331e-03 -3.20025653e-01 6.80713773e-01
9.27568600e-02 1.58499673e-01 -8.65800381e-01 -1.63374445e-03
5.41421890e-01 -2.73753464e-01 -5.32511592e-01 -5.76266527e-01
-1.33512747e+00 5.91643929e-01 3.00047487e-01 1.68260962e-01
-3.13247472e-01 -2.25968495e-01 4.27614093e-01 4.03235555e-01
3.61158133e-01 4.90331203e-01 -2.31728911e-01 -4.12628353e-01
-1.21331465e+00 1.84925839e-01 3.69378835e-01 8.57026756e-01
6.18107438e-01 2.00044021e-01 -5.38536310e-01 7.59569645e-01
4.42900322e-02 8.60928535e-01 2.83272207e-01 -1.30265784e+00
6.35785043e-01 6.91096008e-01 2.69227862e-01 -1.04966617e+00
-2.19613910e-01 -3.44695121e-01 -6.37502670e-01 1.94891542e-01
4.70273465e-01 1.23634905e-01 -8.66436601e-01 1.65328038e+00
3.14542204e-02 -4.13585782e-01 -4.49545979e-02 1.29022133e+00
7.25047112e-01 4.74471599e-01 1.74894437e-01 -3.12132895e-01
1.36560905e+00 -7.03713059e-01 -7.54139483e-01 9.25187469e-02
2.82952636e-01 -6.73636615e-01 1.56051683e+00 7.10463762e-01
-1.25723445e+00 -8.58948827e-01 -1.09571528e+00 7.83148035e-03
-2.48218477e-01 -8.94353017e-02 4.06207889e-01 8.06051850e-01
-1.25714397e+00 4.80575562e-01 -3.90110224e-01 -1.64640069e-01
4.94428575e-01 1.75975174e-01 -4.97373164e-01 -1.72260776e-01
-1.15815258e+00 6.82887375e-01 5.34057133e-02 -7.93130603e-03
-1.17922425e+00 -7.02048838e-01 -7.75292933e-01 4.77535501e-02
5.74283063e-01 -6.81007087e-01 1.06956184e+00 -1.28216386e+00
-1.46820557e+00 7.07098901e-01 -8.49356279e-02 -1.44631013e-01
5.51448464e-01 -1.87764347e-01 -6.57492042e-01 6.31682575e-01
8.15744027e-02 6.83851600e-01 1.27181005e+00 -1.34433150e+00
-6.10835671e-01 -1.49251118e-01 4.40034389e-01 1.57219261e-01
-7.34827161e-01 1.16126701e-01 -6.80000782e-01 -9.44141746e-01
-1.49172887e-01 -4.02500808e-01 2.76303053e-01 3.82933050e-01
9.54388827e-02 -7.46135786e-02 4.22502249e-01 -8.01286399e-01
1.44723046e+00 -2.21925116e+00 3.20614994e-01 8.00212920e-02
3.45891058e-01 4.90968674e-01 -5.26126146e-01 2.74170816e-01
2.45820120e-01 1.16810940e-01 -3.36180627e-01 -1.42406091e-01
2.24173471e-01 5.61800115e-02 -5.51640876e-02 3.89885932e-01
6.35096252e-01 1.05373883e+00 -1.17600417e+00 -7.10707009e-01
2.73620784e-01 5.98566771e-01 -9.70926583e-01 4.57593113e-01
-5.29008098e-02 3.29994738e-01 -1.22440860e-01 9.74748194e-01
3.89423758e-01 -3.08023989e-01 -1.71302482e-01 -6.24907672e-01
1.66739061e-01 -1.40278056e-01 -1.02025151e+00 1.78214824e+00
-3.56623381e-01 5.37912488e-01 -5.71181886e-02 -7.23467827e-01
6.79183364e-01 4.64120865e-01 6.90141916e-01 -1.29663312e+00
1.49206236e-01 2.27500454e-01 -1.73894763e-01 -7.44674742e-01
5.01650333e-01 -1.49392262e-01 6.92668036e-02 1.65378943e-01
5.88040292e-01 -1.49750307e-01 4.33883846e-01 2.70725638e-01
9.98815894e-01 2.24431530e-01 7.48785958e-02 -1.61105450e-02
7.43275940e-01 -2.52021670e-01 5.25270045e-01 6.36098921e-01
-7.13378310e-01 9.83770967e-01 4.86063451e-01 -1.22213729e-01
-1.21027648e+00 -1.36076033e+00 7.88698718e-02 1.20990300e+00
3.21335793e-01 -5.78817368e-01 -9.15550172e-01 -7.93752372e-01
-2.00975090e-01 3.01290303e-01 -4.97351587e-01 -1.65581256e-01
-2.15098843e-01 -4.23484534e-01 4.44741189e-01 4.19459194e-01
6.08767569e-01 -9.33703482e-01 -5.23420513e-01 7.86316171e-02
-6.33509934e-01 -1.34407234e+00 -5.36631346e-01 -3.17952603e-01
-5.79924941e-01 -1.17163086e+00 -9.57174540e-01 -2.36338302e-01
4.83475357e-01 3.66522491e-01 1.35868073e+00 1.60046536e-02
-4.77688201e-02 7.12543309e-01 -7.19235718e-01 6.17443351e-03
-2.67774701e-01 -3.19848090e-01 2.20977798e-01 2.03937888e-01
2.33617499e-01 -3.47971290e-01 -9.92490351e-01 5.28745770e-01
-1.20705593e+00 -3.23201239e-01 6.00113153e-01 7.90606916e-01
3.64959508e-01 4.35680412e-02 5.28286040e-01 -2.67767519e-01
6.22741282e-01 -1.68329969e-01 -4.61607993e-01 3.41411293e-01
-7.22443998e-01 -6.85602501e-02 5.24954259e-01 -4.54755992e-01
-1.04024887e+00 -5.46915531e-01 -1.51001543e-01 -5.29090226e-01
1.57069668e-01 2.48462901e-01 -4.41156715e-01 -1.92181274e-01
7.42214978e-01 1.27335638e-01 -7.43216872e-02 -1.63613096e-01
5.42402089e-01 6.06249750e-01 7.91135550e-01 -4.72620338e-01
8.49135935e-01 4.33681101e-01 -3.74987513e-01 -6.10863149e-01
-8.37580442e-01 -4.84445930e-01 -4.10205603e-01 -5.48336625e-01
9.61811066e-01 -1.10083449e+00 -5.85577905e-01 4.89878654e-01
-8.37075293e-01 1.73482415e-03 -2.86006689e-01 2.78572202e-01
-7.23107934e-01 7.43667066e-01 -4.63038743e-01 -6.54720306e-01
-2.93680608e-01 -1.34227836e+00 1.08218408e+00 -7.35634267e-02
-1.99916646e-01 -7.71834552e-01 -1.20806836e-01 8.70377183e-01
5.71456671e-01 -1.58012658e-02 8.51750433e-01 -1.75420612e-01
-3.85311812e-01 -5.09563871e-02 -5.16678452e-01 7.70955563e-01
4.53690626e-03 2.91260495e-03 -9.88251090e-01 -3.43401879e-01
-1.32954225e-01 -4.71606731e-01 6.24916732e-01 2.22605973e-01
9.17498112e-01 -2.67143786e-01 5.00208735e-01 6.59007132e-01
1.44128227e+00 1.24046486e-02 6.87100112e-01 3.78228694e-01
5.64857721e-01 4.41405535e-01 6.38863921e-01 5.48976183e-01
3.57564688e-01 8.09987903e-01 5.13620436e-01 -1.31631747e-01
-2.88646400e-01 -2.83419371e-01 6.96403623e-01 7.36269474e-01
-3.48558396e-01 -4.32925880e-01 -6.08847082e-01 6.06511414e-01
-1.41756952e+00 -9.11173403e-01 2.79945076e-01 2.05702758e+00
8.33989382e-01 1.68222919e-01 2.96595335e-01 5.08086026e-01
3.49309653e-01 3.80397588e-01 -3.94439429e-01 -2.08317563e-01
-3.58218312e-01 1.44263193e-01 2.53014177e-01 2.48335898e-01
-1.07454574e+00 7.34884322e-01 6.34928083e+00 8.36159885e-01
-9.08292472e-01 2.97799230e-01 3.26553345e-01 -1.74788028e-01
-5.46934009e-01 -1.73117116e-01 -1.97973102e-01 5.10549426e-01
8.38353932e-01 1.53704166e-01 5.98929882e-01 5.61419904e-01
4.23923105e-01 2.06181649e-02 -1.02623785e+00 1.22826278e+00
3.53348881e-01 -9.74718750e-01 4.74030524e-01 -1.44522637e-01
8.35699797e-01 -2.24979505e-01 3.16583455e-01 1.44685358e-01
-1.43261865e-01 -8.34105372e-01 1.12242961e+00 5.17052293e-01
1.06000185e+00 -4.79781926e-01 7.39230096e-01 -2.61398286e-01
-9.01708961e-01 -3.39501351e-01 -1.39141575e-01 1.66175604e-01
2.71549791e-01 5.75010061e-01 -1.24001699e-02 5.60880601e-01
8.40814352e-01 6.27654135e-01 -8.53313863e-01 9.82320964e-01
-8.47742558e-02 5.29141366e-01 2.28369310e-01 3.03178161e-01
1.52074262e-01 2.89899986e-02 5.66433430e-01 1.05192506e+00
3.78865480e-01 1.17348887e-01 -1.72047168e-01 7.26702452e-01
-4.39177528e-02 2.02996597e-01 -2.22902089e-01 -2.35690475e-01
2.23795056e-01 1.03203261e+00 -2.70588517e-01 -3.52108300e-01
-6.93183661e-01 1.04476666e+00 -4.34829528e-03 6.04656041e-01
-9.04429197e-01 -2.65769601e-01 8.22159708e-01 1.43140629e-01
4.14612919e-01 -1.44818604e-01 -7.30112568e-02 -1.48170161e+00
3.55093569e-01 -1.39810586e+00 2.19469711e-01 -9.26424444e-01
-1.36995220e+00 7.03351617e-01 -1.72100797e-01 -1.63580465e+00
-1.29673526e-01 -7.03657210e-01 -1.50327757e-01 4.73179609e-01
-1.62556553e+00 -1.02819788e+00 -5.35425663e-01 8.94207001e-01
5.46172321e-01 -2.67153740e-01 4.76667434e-01 5.85301697e-01
-3.00328165e-01 8.36683512e-01 -1.31070942e-01 -1.25158057e-02
9.56973493e-01 -1.03791285e+00 5.26080765e-02 1.03289390e+00
9.08249244e-02 3.16630363e-01 7.46897578e-01 -2.59726644e-01
-1.52154553e+00 -7.19243944e-01 6.15319967e-01 -5.56011677e-01
8.02858472e-01 -1.81632549e-01 -8.80258679e-01 -7.66850486e-02
-3.07505056e-02 -5.66251809e-04 4.64146256e-01 -2.65199631e-01
-7.77465582e-01 -3.06131452e-01 -1.01314688e+00 4.47417915e-01
1.22244179e+00 -9.42373157e-01 -4.88306135e-01 -2.67663482e-03
6.41089678e-01 9.36596170e-02 -1.06495321e+00 5.66935420e-01
4.96363819e-01 -1.44435024e+00 1.05621457e+00 -2.90487558e-01
6.86776876e-01 -4.54374254e-01 -4.34194326e-01 -1.08993280e+00
-3.88335109e-01 -4.58244890e-01 -2.72239298e-01 1.28105748e+00
2.10897565e-01 -2.37185985e-01 3.82610887e-01 4.78027165e-01
-1.88853275e-02 -4.95348573e-01 -9.20337141e-01 -6.34960532e-01
-2.85595059e-01 -6.57213032e-01 5.17600596e-01 6.71586156e-01
-3.81409600e-02 -1.32449372e-02 -5.53009570e-01 5.66413999e-02
6.48592293e-01 -2.52600253e-01 6.12572193e-01 -9.27384138e-01
-1.97277054e-01 -7.23036826e-01 -7.83153176e-01 -7.86694288e-01
-1.49855807e-01 -4.91852164e-01 8.62832833e-03 -1.40925694e+00
2.84513921e-01 -9.37617645e-02 -5.35209000e-01 2.62225807e-01
-4.76794958e-01 6.72931194e-01 3.11853766e-01 2.52691805e-01
-1.07273054e+00 7.68608510e-01 1.32183290e+00 -3.93074483e-01
3.69611084e-02 -4.42874163e-01 -6.98013723e-01 4.73986924e-01
3.19312125e-01 5.27174994e-02 -6.18530929e-01 -6.35796785e-01
5.54121673e-01 7.79566541e-02 5.60463428e-01 -1.27097821e+00
-5.37528992e-02 -4.78218608e-02 2.56406993e-01 -1.23764470e-01
2.80956596e-01 -7.60424137e-01 -1.55291319e-01 2.60701962e-02
-2.61130542e-01 1.38943449e-01 -1.71730712e-01 5.48180401e-01
-6.54614270e-01 1.03952006e-01 5.61508179e-01 6.39883280e-02
-1.01952946e+00 3.31002891e-01 -9.54624340e-02 3.02592456e-01
6.89083636e-01 -4.32707012e-01 -3.40151310e-01 -5.95655203e-01
-4.61589247e-01 1.25406742e-01 7.75305390e-01 5.98559260e-01
9.17951286e-01 -1.29026484e+00 -7.84358084e-01 3.37280005e-01
5.85295498e-01 -5.32160163e-01 4.89462167e-01 8.89864802e-01
-4.78261054e-01 1.84309527e-01 -3.68843347e-01 -6.45447254e-01
-9.59577501e-01 7.35693216e-01 2.83762753e-01 1.90693177e-02
-3.71588022e-01 6.75525069e-01 2.00040266e-01 5.17054275e-02
3.83270979e-01 -1.99272916e-01 -2.46280894e-01 2.07844555e-01
6.89043701e-01 3.93453866e-01 1.43485099e-01 -8.72367799e-01
-2.91681468e-01 6.54730380e-01 3.96598428e-01 -1.65875286e-01
1.07465649e+00 -4.11673665e-01 1.86835378e-01 1.61631808e-01
1.13210332e+00 -9.06287208e-02 -1.53027463e+00 -3.07566136e-01
-1.34955645e-01 -7.03233302e-01 9.63514000e-02 -9.61896062e-01
-1.23607242e+00 8.75374675e-01 8.88325751e-01 -6.57282919e-02
1.55819821e+00 -1.61551580e-01 8.22271943e-01 1.87915806e-02
4.77309823e-01 -1.09283578e+00 5.59892237e-01 1.30088881e-01
7.82560468e-01 -1.55846703e+00 -2.95557845e-02 -6.90002516e-02
-9.30022120e-01 7.73853123e-01 2.84325868e-01 2.84352541e-01
2.73407340e-01 -1.98718384e-01 3.77387643e-01 -2.02152535e-01
-5.81688404e-01 -4.13917243e-01 8.14118743e-01 6.95826411e-01
2.03063458e-01 -1.12365834e-01 -1.68911815e-01 4.83537287e-01
-1.21203840e-01 6.99150562e-02 2.43861482e-01 5.98850548e-01
-3.25615704e-01 -7.97109663e-01 -2.42026284e-01 2.32431561e-01
-5.29376268e-01 -1.24756910e-01 -1.42605454e-01 5.27602911e-01
2.95472741e-01 1.33598590e+00 -1.74091235e-01 -6.43408716e-01
5.25733948e-01 -1.23055875e-01 5.23878276e-01 -3.53947282e-02
-4.65868503e-01 -5.06368876e-02 -8.15984011e-02 -9.44459975e-01
-7.73557544e-01 -4.05532837e-01 -6.32274985e-01 -2.58116573e-01
-1.97669789e-01 -6.39099777e-02 5.32920897e-01 9.56037581e-01
1.90501407e-01 3.33520085e-01 6.38067305e-01 -8.16663086e-01
-5.45933843e-01 -7.33449638e-01 -3.77621293e-01 1.03644443e+00
4.90242124e-01 -8.34275961e-01 -3.92673075e-01 2.44698077e-01] | [11.815016746520996, -1.8229495286941528] |
17a5d936-5f8e-4fd5-a8d5-8f269ac728fc | fully-self-supervised-learning-for-semantic | 2202.11981 | null | https://arxiv.org/abs/2202.11981v1 | https://arxiv.org/pdf/2202.11981v1.pdf | Fully Self-Supervised Learning for Semantic Segmentation | In this work, we present a fully self-supervised framework for semantic segmentation(FS^4). A fully bootstrapped strategy for semantic segmentation, which saves efforts for the huge amount of annotation, is crucial for building customized models from end-to-end for open-world domains. This application is eagerly needed in realistic scenarios. Even though recent self-supervised semantic segmentation methods have gained great progress, these works however heavily depend on the fully-supervised pretrained model and make it impossible a fully self-supervised pipeline. To solve this problem, we proposed a bootstrapped training scheme for semantic segmentation, which fully leveraged the global semantic knowledge for self-supervision with our proposed PGG strategy and CAE module. In particular, we perform pixel clustering and assignments for segmentation supervision. Preventing it from clustering a mess, we proposed 1) a pyramid-global-guided (PGG) training strategy to supervise the learning with pyramid image/patch-level pseudo labels, which are generated by grouping the unsupervised features. The stable global and pyramid semantic pseudo labels can prevent the segmentation from learning too many clutter regions or degrading to one background region; 2) in addition, we proposed context-aware embedding (CAE) module to generate global feature embedding in view of its neighbors close both in space and appearance in a non-trivial way. We evaluate our method on the large-scale COCO-Stuff dataset and achieved 7.19 mIoU improvements on both things and stuff objects | ['Wenwu Zhu', 'Qi Ju', 'Zhi Wang', 'Yucong Li', 'Wei Zhuo', 'YuAn Wang'] | 2022-02-24 | null | null | null | null | ['unsupervised-semantic-segmentation'] | ['computer-vision'] | [ 3.61191392e-01 4.45444942e-01 5.64470403e-02 -5.44950604e-01
-5.55523634e-01 -5.25886536e-01 3.28843832e-01 6.45133555e-02
-3.13299537e-01 4.48350817e-01 -1.06255271e-01 1.03293180e-01
1.26619980e-01 -9.35726762e-01 -7.74031520e-01 -7.01711595e-01
1.47097364e-01 6.22624874e-01 9.62157190e-01 -1.23610824e-01
1.88810349e-01 2.67958671e-01 -1.84392440e+00 2.62573481e-01
1.23067033e+00 8.77887189e-01 7.60021567e-01 5.00177681e-01
-5.22396147e-01 3.96930665e-01 -1.89825773e-01 -2.67496072e-02
4.79528129e-01 -4.27827150e-01 -1.04173636e+00 4.83805478e-01
4.37662512e-01 1.27703026e-01 4.11550999e-01 1.10358667e+00
3.80451262e-01 1.46114333e-02 5.96828163e-01 -1.04858303e+00
-4.19684470e-01 5.01716614e-01 -7.42969990e-01 -2.54268110e-01
-1.60418063e-01 2.39340395e-01 1.01443601e+00 -5.94062090e-01
7.08873570e-01 1.07629097e+00 7.16803014e-01 4.70544189e-01
-9.69536781e-01 -3.76467824e-01 2.21443295e-01 1.23014316e-01
-1.12101030e+00 -2.01497689e-01 9.31542337e-01 -4.55365777e-01
6.33213878e-01 8.42226744e-02 7.35861659e-01 7.35764503e-01
-2.00503558e-01 9.36967790e-01 1.22064388e+00 -4.85011995e-01
3.17066610e-01 3.96377265e-01 2.10990176e-01 8.60188425e-01
2.57599168e-02 -2.95029074e-01 -1.84945196e-01 3.50070566e-01
7.61139452e-01 -8.27814713e-02 6.16530478e-02 -7.35805273e-01
-1.06913817e+00 6.57670200e-01 5.97480893e-01 4.31969225e-01
-2.44462445e-01 1.84997078e-03 2.37540260e-01 -3.35515998e-02
4.37259287e-01 3.11874807e-01 -7.83609509e-01 2.54736632e-01
-1.19732165e+00 -1.75819844e-01 6.15741670e-01 9.05200720e-01
1.25469291e+00 -8.19150582e-02 -3.76303382e-02 9.73182380e-01
3.09189558e-01 2.52494544e-01 4.26525146e-01 -8.22694659e-01
2.07826555e-01 1.09657288e+00 -5.75507060e-02 -9.01951551e-01
-4.98747766e-01 -7.45622694e-01 -6.41426027e-01 2.57828265e-01
3.67912889e-01 7.45565072e-02 -1.37180424e+00 1.47546399e+00
7.22932816e-01 3.17372978e-01 3.51000763e-02 7.86967993e-01
7.76170492e-01 4.31800365e-01 2.45316312e-01 1.41454354e-01
1.35343003e+00 -1.67577958e+00 -4.43312734e-01 -3.81047636e-01
7.58012772e-01 -9.08574164e-01 1.30751729e+00 1.67773440e-01
-8.01855624e-01 -7.83210754e-01 -9.94844556e-01 -1.13866903e-01
-6.15660012e-01 2.28589907e-01 7.51661241e-01 6.82633936e-01
-1.15801203e+00 7.72355258e-01 -8.70436847e-01 -6.95261359e-01
6.75714016e-01 3.89497787e-01 -3.97027582e-01 2.10788790e-02
-8.04074705e-01 4.52399313e-01 7.31194913e-01 9.23425034e-02
-8.57193649e-01 -3.13483506e-01 -8.13118696e-01 9.79846865e-02
6.39335811e-01 -6.01216018e-01 6.64659023e-01 -1.36349761e+00
-1.37324309e+00 9.80463326e-01 4.29994054e-02 -4.18037981e-01
4.77506220e-01 -2.10035816e-01 -9.35484469e-02 3.37482065e-01
5.10956109e-01 1.13597202e+00 1.03955925e+00 -1.55517697e+00
-6.26385033e-01 -4.53421324e-01 -1.81007981e-01 3.54379743e-01
-2.50030667e-01 -2.49716148e-01 -7.50927031e-01 -5.12690902e-01
4.68052119e-01 -8.58300745e-01 -4.75904912e-01 -1.58436418e-01
-4.07126695e-01 -5.35929203e-02 1.16759455e+00 -6.85753047e-01
8.90502930e-01 -2.14315796e+00 -1.28591299e-01 2.36902162e-01
-1.02359086e-01 3.62545460e-01 -7.92594030e-02 6.73234388e-02
5.01274504e-03 6.76808283e-02 -6.23462856e-01 -5.12965858e-01
-1.20420754e-01 3.38069409e-01 8.34957957e-02 2.40245908e-01
2.29349345e-01 9.60653961e-01 -8.40723753e-01 -9.22570586e-01
4.93084282e-01 3.28004062e-01 -7.36161590e-01 2.97598958e-01
-5.97206056e-01 6.34034991e-01 -3.85152727e-01 8.36425304e-01
8.83084416e-01 -3.69953662e-01 -8.95208865e-02 -3.44343543e-01
-1.92200691e-02 -2.72268355e-01 -1.48369873e+00 2.01948524e+00
-3.63090843e-01 -1.16465122e-01 2.22186595e-01 -1.21136260e+00
8.40499043e-01 -1.86803807e-02 4.59821850e-01 -4.64864343e-01
2.61227041e-01 1.97979972e-01 -2.96243668e-01 -4.64090943e-01
2.38627940e-01 1.06318787e-01 5.02856262e-02 1.63436010e-01
2.83054918e-01 -2.07962334e-01 1.77799776e-01 1.58502325e-01
9.27615106e-01 7.66090512e-01 -3.54177952e-02 -4.82569486e-01
7.25880623e-01 4.66346979e-01 6.50177181e-01 5.18088102e-01
-2.07403108e-01 1.01641846e+00 2.50247359e-01 -2.81194448e-01
-8.19218218e-01 -9.80772316e-01 -4.81227264e-02 1.09774363e+00
3.99802327e-01 -3.14946592e-01 -1.21231365e+00 -1.05201674e+00
-3.40262949e-01 4.87820148e-01 -4.55992281e-01 2.10170820e-01
-2.95538932e-01 -8.27117920e-01 2.29812875e-01 5.25455415e-01
9.41461623e-01 -1.17867362e+00 -6.67779207e-01 2.65528947e-01
3.20821628e-02 -1.09942925e+00 -3.97968799e-01 5.17482221e-01
-9.16800797e-01 -1.12691391e+00 -6.35264397e-01 -1.07437432e+00
8.99574995e-01 3.14987540e-01 8.97479355e-01 -5.27558569e-03
-4.60927427e-01 1.92082435e-01 -5.80049336e-01 -1.64750859e-01
-1.98884740e-01 2.14465454e-01 -3.74212295e-01 2.04476684e-01
6.85928613e-02 -5.92079878e-01 -8.92166674e-01 5.06287158e-01
-8.67236674e-01 3.54659230e-01 7.20502794e-01 6.96117818e-01
9.34532762e-01 8.10965896e-02 6.19954526e-01 -1.30719995e+00
-5.75894900e-02 -2.86875457e-01 -5.31516373e-01 2.84724385e-01
-6.23412490e-01 -2.69676633e-02 6.92742467e-01 1.56666187e-03
-1.18735158e+00 6.21099174e-01 -3.68460685e-01 -2.66280264e-01
-4.86766756e-01 -2.49130726e-02 -4.92203802e-01 7.37832161e-03
5.55803537e-01 2.20667452e-01 -1.20785251e-01 -5.78555763e-01
7.03934371e-01 7.82953203e-01 4.61002767e-01 -5.44009984e-01
7.53979981e-01 6.90693974e-01 -2.79205978e-01 -7.83323884e-01
-9.76564944e-01 -7.39301503e-01 -9.55529034e-01 3.92389670e-02
1.39968860e+00 -1.00301504e+00 -1.04205340e-01 4.99896973e-01
-7.56191969e-01 -4.17713970e-01 -3.90956461e-01 -2.53965445e-02
-5.00493050e-01 6.22793853e-01 -2.60199010e-01 -6.20873630e-01
-3.58403862e-01 -1.16355443e+00 1.20253718e+00 5.00907004e-01
6.14767037e-02 -9.97748375e-01 -1.91493869e-01 8.58968079e-01
4.37324345e-01 2.95301229e-01 6.12697780e-01 -6.85859442e-01
-8.39275539e-01 8.45611989e-02 -5.36355913e-01 7.61854947e-01
1.25148505e-01 -2.33750448e-01 -1.14949846e+00 -2.00261623e-02
-1.45607591e-01 -3.21584255e-01 9.49771702e-01 1.79157019e-01
1.28155422e+00 1.75530866e-01 -4.71636534e-01 8.34335506e-01
1.55553722e+00 -1.45392597e-01 3.83622587e-01 3.59214216e-01
1.13872159e+00 6.30863547e-01 7.71234274e-01 1.42277673e-01
3.70572358e-01 4.35441226e-01 4.92740154e-01 -5.72536349e-01
-5.48449099e-01 -3.78837347e-01 8.55515301e-02 6.91667497e-01
1.12569347e-01 2.55507976e-02 -7.42380381e-01 8.74177754e-01
-1.92197669e+00 -4.70120370e-01 -3.35278869e-01 1.89855754e+00
7.37109601e-01 2.28975609e-01 -4.07133847e-02 4.20748331e-02
7.13980019e-01 1.03228562e-01 -5.52749932e-01 -2.55424112e-01
-2.74148732e-01 4.77728248e-01 7.04273582e-01 3.98985088e-01
-1.32676506e+00 1.59646249e+00 5.06713057e+00 1.12783229e+00
-1.00680649e+00 4.21916813e-01 7.37223685e-01 5.46688795e-01
-2.27824450e-01 2.77910680e-01 -8.61600876e-01 4.74503934e-01
4.27936405e-01 5.43261528e-01 -1.46868965e-02 1.17151892e+00
5.94970919e-02 -4.08837140e-01 -6.11227810e-01 7.54407525e-01
-7.39960372e-02 -1.12056518e+00 -2.04744637e-02 -2.90370315e-01
9.41407681e-01 7.99135491e-02 -3.16092253e-01 1.84718877e-01
3.54088217e-01 -8.16499472e-01 5.53578258e-01 2.77287923e-02
6.46257639e-01 -4.45982069e-01 7.28119075e-01 3.62935930e-01
-1.22685432e+00 4.92265895e-02 -3.19976002e-01 2.59899229e-01
9.49390158e-02 7.68047750e-01 -7.82466769e-01 7.01394260e-01
8.60753000e-01 6.96870208e-01 -8.28607738e-01 7.81483293e-01
-2.79678464e-01 7.86677599e-01 -4.57131058e-01 4.01106238e-01
4.85924840e-01 -3.68592322e-01 1.76203385e-01 1.26475906e+00
1.26048578e-02 -2.36545831e-01 3.83013457e-01 7.54474699e-01
2.30797946e-01 4.28589970e-01 -2.51162142e-01 8.38023648e-02
3.84123065e-02 1.58806777e+00 -1.55047667e+00 -2.68065989e-01
-3.05949152e-01 1.44970477e+00 1.89071476e-01 1.72083661e-01
-8.05850089e-01 -2.91439891e-01 1.02422900e-01 3.50536019e-01
5.18836319e-01 -3.57692204e-02 -4.95580196e-01 -1.16435468e+00
-7.49192983e-02 -3.92262161e-01 3.91058326e-01 -6.05984449e-01
-1.18788862e+00 4.64875132e-01 -2.45852381e-01 -1.00883019e+00
2.01354608e-01 -2.40656480e-01 -6.68012857e-01 5.85142553e-01
-1.65877879e+00 -1.63188183e+00 -4.24299628e-01 6.66061461e-01
6.92192435e-01 -1.00304997e-02 7.25551367e-01 3.57473403e-01
-5.15157342e-01 3.32907498e-01 -5.28444983e-02 -3.95229198e-02
6.53302610e-01 -1.46750498e+00 1.26086175e-01 8.42795134e-01
2.20710680e-01 3.59692454e-01 5.09872615e-01 -7.19499171e-01
-9.03386772e-01 -1.26596689e+00 5.10517359e-01 -2.70535499e-01
3.76574755e-01 -4.31015760e-01 -8.25090647e-01 4.80307966e-01
8.69048387e-02 2.01508388e-01 4.91907716e-01 1.12549020e-02
-6.49789795e-02 -2.29552925e-01 -1.31380975e+00 3.88971150e-01
1.22179556e+00 -4.13874388e-02 -6.18193626e-01 4.14259434e-01
8.08092237e-01 -2.15409487e-01 -5.01135468e-01 5.14670551e-01
1.53654665e-01 -1.24728942e+00 8.92572284e-01 2.79329065e-02
3.35292161e-01 -7.85248578e-01 -1.17331654e-01 -9.27718461e-01
-9.55009386e-02 -3.13311070e-01 3.87952924e-01 1.46982992e+00
2.49280065e-01 -3.49051505e-01 1.14870524e+00 1.39010817e-01
-4.48032528e-01 -7.81204283e-01 -7.21924722e-01 -7.05814958e-01
-1.16751567e-01 -2.96592504e-01 3.00278008e-01 9.34783101e-01
-5.33913493e-01 3.58905137e-01 -1.72938034e-01 2.60616690e-01
7.09026754e-01 2.60253608e-01 8.56164694e-01 -1.03504300e+00
-4.46378559e-01 -1.58047080e-01 -5.28498948e-01 -9.77054596e-01
9.45929531e-03 -1.15359485e+00 7.26978034e-02 -1.81404459e+00
2.15222627e-01 -7.35243976e-01 -3.01544070e-01 5.77273548e-01
-2.03159779e-01 4.37581092e-01 2.17154175e-02 5.90520166e-02
-9.29654658e-01 4.70705956e-01 1.25158858e+00 -4.33833599e-02
-2.42710173e-01 -2.31734276e-01 -7.68580258e-01 8.28169525e-01
7.26192176e-01 -4.15584773e-01 -6.70781732e-01 -3.19917232e-01
-1.39057368e-01 -4.14173961e-01 3.41446102e-01 -1.24330854e+00
1.13933854e-01 6.24078475e-02 3.20782721e-01 -7.99480677e-01
1.36896715e-01 -9.85854626e-01 9.18012708e-02 2.26434082e-01
2.07885727e-01 -6.39848292e-01 5.99664748e-02 5.38502097e-01
-3.42894047e-01 -3.04093719e-01 1.11973333e+00 -5.73097527e-01
-1.03196716e+00 1.37556866e-01 1.90097138e-01 1.35246823e-02
1.32469726e+00 -4.92006719e-01 1.21813580e-01 1.27359971e-01
-1.02516758e+00 5.34143448e-01 7.74958968e-01 1.72789499e-01
3.16042066e-01 -7.89862394e-01 -2.34351680e-01 3.51369083e-01
1.19956464e-01 6.40646636e-01 5.39357662e-01 6.53508782e-01
-7.30275154e-01 -4.55628149e-03 -3.20167661e-01 -9.00237322e-01
-1.07476282e+00 4.94344592e-01 1.36111870e-01 -2.77926415e-01
-6.48046553e-01 9.59280670e-01 3.59320849e-01 -7.07892299e-01
1.32586926e-01 -2.20880136e-01 -2.52804130e-01 1.37583420e-01
1.15969002e-01 1.04781538e-01 1.33711146e-02 -6.15882814e-01
-3.03515255e-01 1.03647864e+00 -1.20654568e-01 1.09490111e-01
1.44284809e+00 -3.54088485e-01 -1.19557872e-01 2.53275096e-01
1.06929028e+00 5.42438924e-02 -1.34623861e+00 1.49967866e-02
6.15022294e-02 -3.06346536e-01 1.22456655e-01 -9.29676950e-01
-1.23686993e+00 8.87019098e-01 7.95700669e-01 -3.11915204e-02
1.13584054e+00 1.76003858e-01 1.07512295e+00 7.55505636e-02
4.31750417e-01 -1.36838067e+00 9.89193544e-02 3.01561177e-01
2.97183722e-01 -1.47542763e+00 -1.08931512e-01 -8.80476236e-01
-8.24624658e-01 8.76047373e-01 7.30781794e-01 -2.63830513e-01
7.09788918e-01 2.57750563e-02 3.14693838e-01 -2.57840753e-01
-1.52071342e-01 -5.61206818e-01 7.45677724e-02 6.26626194e-01
1.53422296e-01 5.88459484e-02 -2.62080073e-01 7.57803619e-01
2.90913843e-02 3.50106247e-02 1.93239823e-01 9.01785195e-01
-7.39505827e-01 -1.28065574e+00 -2.72411466e-01 3.09455693e-01
-2.32162729e-01 -4.37455960e-02 -1.68206871e-01 5.02838850e-01
7.66860306e-01 7.64583468e-01 -2.46988297e-01 -2.37430289e-01
1.53934941e-01 2.68361181e-01 2.26068676e-01 -9.49281096e-01
-5.50898373e-01 3.45366389e-01 -1.12810232e-01 -6.11732483e-01
-5.98698854e-01 -5.38266301e-01 -1.60198998e+00 3.96530509e-01
-3.92978370e-01 -1.34273261e-01 6.97743833e-01 8.12599480e-01
4.53262836e-01 5.22552967e-01 4.94526029e-01 -7.78597414e-01
-1.44060940e-01 -8.19144428e-01 -6.00413084e-01 7.10426748e-01
-2.44292110e-01 -6.31492317e-01 -4.00157988e-01 2.75181323e-01] | [9.632708549499512, 0.5910286903381348] |
c8c9660a-4bae-42f4-af5e-570ed3b6ab64 | neural-monocular-3d-human-motion-capture-with | 2105.01057 | null | https://arxiv.org/abs/2105.01057v1 | https://arxiv.org/pdf/2105.01057v1.pdf | Neural Monocular 3D Human Motion Capture with Physical Awareness | We present a new trainable system for physically plausible markerless 3D human motion capture, which achieves state-of-the-art results in a broad range of challenging scenarios. Unlike most neural methods for human motion capture, our approach, which we dub physionical, is aware of physical and environmental constraints. It combines in a fully differentiable way several key innovations, i.e., 1. a proportional-derivative controller, with gains predicted by a neural network, that reduces delays even in the presence of fast motions, 2. an explicit rigid body dynamics model and 3. a novel optimisation layer that prevents physically implausible foot-floor penetration as a hard constraint. The inputs to our system are 2D joint keypoints, which are canonicalised in a novel way so as to reduce the dependency on intrinsic camera parameters -- both at train and test time. This enables more accurate global translation estimation without generalisability loss. Our model can be finetuned only with 2D annotations when the 3D annotations are not available. It produces smooth and physically principled 3D motions in an interactive frame rate in a wide variety of challenging scenes, including newly recorded ones. Its advantages are especially noticeable on in-the-wild sequences that significantly differ from common 3D pose estimation benchmarks such as Human 3.6M and MPI-INF-3DHP. Qualitative results are available at http://gvv.mpi-inf.mpg.de/projects/PhysAware/ | ['Christian Theobalt', 'Patrick Pérez', 'Weipeng Xu', 'Vladislav Golyanik', 'Soshi Shimada'] | 2021-05-03 | null | null | null | null | ['3d-pose-estimation'] | ['computer-vision'] | [-1.04078546e-01 2.30953291e-01 -3.92084010e-02 1.53147861e-01
-6.92027211e-01 -5.83061695e-01 5.47467470e-01 -3.85048062e-01
-5.97803712e-01 6.83926702e-01 1.19953915e-01 -8.64241272e-02
-1.01411760e-01 -3.03547442e-01 -9.79341030e-01 -6.61964893e-01
-1.82642445e-01 6.79431558e-01 4.75460827e-01 -5.11806250e-01
-1.57094181e-01 7.61956155e-01 -1.20656335e+00 -2.45602563e-01
4.39506084e-01 6.26310766e-01 2.49890506e-01 9.69259620e-01
7.71649718e-01 5.45938075e-01 -2.14375988e-01 -1.45012334e-01
5.27997851e-01 -2.39081353e-01 -6.39715552e-01 7.95086399e-02
4.69240516e-01 -2.98252195e-01 -4.42630023e-01 6.02584243e-01
9.39849019e-01 5.03651321e-01 4.61921901e-01 -8.25245798e-01
-5.17462231e-02 -1.58079609e-01 -3.91619980e-01 -6.52819499e-02
6.01755977e-01 6.97162986e-01 4.25526410e-01 -7.99551785e-01
9.95962977e-01 1.23557913e+00 1.09453321e+00 8.67978215e-01
-1.24639523e+00 -9.42293778e-02 -7.76903555e-02 8.55289325e-02
-1.30296171e+00 -4.55021530e-01 7.65265882e-01 -5.81759274e-01
1.14054489e+00 5.05818725e-01 9.58482742e-01 1.52786613e+00
4.80326056e-01 6.23160779e-01 5.48330724e-01 -3.37695599e-01
1.55369267e-01 -3.49283367e-01 -3.88604671e-01 6.80403233e-01
-2.14361791e-02 3.88933152e-01 -3.18953544e-01 6.56041354e-02
1.32608068e+00 -2.23692939e-01 -4.58875388e-01 -8.40436518e-01
-1.59561980e+00 4.62348193e-01 3.80407214e-01 -4.38982621e-02
-3.87170941e-01 6.00584149e-01 2.47466490e-01 -2.93933284e-02
3.11448306e-01 3.72253478e-01 -6.15849495e-01 -3.65151405e-01
-7.65672028e-01 7.58132041e-01 6.14501119e-01 8.94817710e-01
3.48665446e-01 2.88269162e-01 7.01700523e-02 5.38353741e-01
2.76362032e-01 7.18439341e-01 2.58676887e-01 -1.41167581e+00
5.27008593e-01 1.30995572e-01 4.01588649e-01 -9.44221973e-01
-7.93913066e-01 -3.44398260e-01 -9.36082244e-01 5.79172194e-01
3.89935315e-01 -2.20879287e-01 -9.16641712e-01 1.67211437e+00
7.26922035e-01 3.29312801e-01 -1.91147909e-01 1.40705395e+00
4.78995174e-01 4.00804996e-01 -2.56079733e-01 9.81462225e-02
1.07865810e+00 -9.18995917e-01 -4.89964277e-01 -4.28536832e-01
5.22421002e-01 -5.89680314e-01 1.02588856e+00 3.62884223e-01
-1.43988931e+00 -6.99755192e-01 -8.72154772e-01 -2.21039429e-01
-3.94067876e-02 -6.81976452e-02 3.61555994e-01 4.20032263e-01
-1.08189523e+00 8.73058379e-01 -1.24666917e+00 -3.94201428e-01
9.20683593e-02 6.05137110e-01 -6.98149025e-01 3.03664535e-01
-1.14191020e+00 1.12993491e+00 2.47329608e-01 4.61231649e-01
-7.57156670e-01 -6.36527359e-01 -1.03362167e+00 -2.52796710e-01
6.40096605e-01 -1.35765803e+00 1.25453019e+00 -5.88741899e-01
-2.06890416e+00 9.08644438e-01 2.05921680e-01 -3.87411356e-01
1.30875266e+00 -7.14279890e-01 1.29574269e-01 7.28336349e-02
-2.70752102e-01 9.06166732e-01 7.55444586e-01 -1.14294171e+00
-1.20142169e-01 -5.39020337e-02 -1.16965316e-01 4.70928431e-01
4.29742992e-01 -3.39088887e-01 -8.62638354e-01 -8.46013069e-01
-1.43421784e-01 -1.47488427e+00 -6.92699313e-01 2.31662855e-01
-5.59883356e-01 4.52199966e-01 5.76328278e-01 -6.51670456e-01
7.46971250e-01 -1.76645744e+00 8.22188854e-01 1.04355529e-01
-2.78580133e-02 3.00885648e-01 1.27475187e-02 2.89564908e-01
-9.19675902e-02 -1.63232267e-01 -3.08717400e-01 -5.99149406e-01
8.23350027e-02 2.14227691e-01 2.00716872e-02 6.93835795e-01
1.84247807e-01 1.06112528e+00 -8.27885509e-01 -2.44166061e-01
6.52950585e-01 7.87262142e-01 -6.60973728e-01 2.71348715e-01
-2.53615379e-01 9.41956282e-01 -3.35004330e-01 3.49731594e-01
3.28474849e-01 2.08124854e-02 -1.54741168e-01 -3.14116012e-03
-1.36073202e-01 5.17906211e-02 -1.53764248e+00 2.17717671e+00
-1.53617740e-01 4.34254527e-01 2.42958978e-01 -6.52091622e-01
6.34423792e-01 4.82182443e-01 5.44514537e-01 -3.79040390e-01
2.64058024e-01 1.40368000e-01 -2.45985016e-01 -4.97911572e-01
5.65449357e-01 2.29730643e-02 -1.78220555e-01 -1.82018951e-01
-8.97633657e-02 -3.62035722e-01 -1.88485101e-01 -1.57460049e-01
1.09385300e+00 9.36393201e-01 2.11508542e-01 -3.24217081e-01
4.25669372e-01 1.46856189e-01 6.99517667e-01 4.47343826e-01
-9.32525322e-02 1.14519274e+00 2.46311158e-01 -5.02135336e-01
-1.25090945e+00 -1.17482769e+00 3.07244718e-01 4.47121739e-01
2.84440011e-01 -2.33638227e-01 -7.42091417e-01 -2.44504362e-01
-5.57238422e-02 3.09780419e-01 -5.10043085e-01 -2.29075372e-01
-1.09713340e+00 -3.78264844e-01 4.94459152e-01 5.90412140e-01
2.31476128e-01 -1.03897524e+00 -1.15694606e+00 2.96049446e-01
-5.46889827e-02 -1.27529168e+00 -4.11797136e-01 1.40393049e-01
-9.73393321e-01 -9.87125039e-01 -9.77063715e-01 -3.98072600e-01
2.94016600e-01 -1.26050055e-01 1.09547353e+00 -4.59385738e-02
-3.02046984e-01 4.04434800e-01 -4.28222418e-02 -1.51695833e-01
-3.42277437e-01 6.22325167e-02 3.57723922e-01 -3.93411219e-01
-4.51049656e-01 -5.61913669e-01 -6.86533034e-01 5.78751087e-01
-4.82693195e-01 3.77366573e-01 2.62221247e-01 8.04349303e-01
8.27165127e-01 -4.73461479e-01 -1.48655862e-01 -5.34977257e-01
1.46989182e-01 -1.13653906e-01 -6.50645494e-01 -2.27970749e-01
7.42785186e-02 -5.88106588e-02 4.52519745e-01 -6.87654018e-01
-8.94867778e-01 5.16186953e-01 -3.59075278e-01 -8.14982593e-01
-4.06158954e-01 1.30875155e-01 -2.62781501e-01 -1.19849302e-01
7.75453627e-01 -1.03043377e-01 -5.66034615e-02 -5.92261612e-01
3.44217151e-01 -1.31745666e-01 1.03154111e+00 -5.57817400e-01
9.74407852e-01 4.70594794e-01 4.92821813e-01 -9.50720370e-01
-2.19341457e-01 -3.57765108e-01 -1.07327712e+00 -3.99558723e-01
1.02077127e+00 -9.22684610e-01 -9.03589904e-01 7.48374283e-01
-1.21979499e+00 -1.05712259e+00 -5.26593447e-01 8.03937316e-01
-1.14582896e+00 3.70619267e-01 -7.07859993e-01 -6.54655516e-01
-2.34904334e-01 -1.23378468e+00 1.22749722e+00 3.03316470e-02
-5.16054571e-01 -1.03066921e+00 3.40421379e-01 1.51038826e-01
8.19078833e-02 9.30727899e-01 3.32674295e-01 -2.75393482e-02
-5.16210198e-01 -1.24408089e-01 4.19645578e-01 1.13759235e-01
-1.96189672e-01 8.29713196e-02 -7.93763220e-01 -3.88603717e-01
-2.04967275e-01 -4.17619906e-02 5.25475979e-01 6.46723688e-01
5.95078409e-01 -1.81495622e-01 -3.80980134e-01 8.64883423e-01
1.17774689e+00 -1.85789555e-01 7.91755378e-01 5.24068356e-01
1.13773835e+00 5.38299084e-01 5.23035765e-01 2.36388758e-01
3.78012389e-01 1.31254542e+00 6.98556840e-01 -2.66060859e-01
-1.21917874e-01 -3.96779552e-02 5.15661001e-01 6.87523246e-01
-6.89388335e-01 -2.85999656e-01 -1.03651965e+00 4.58102912e-01
-2.18625093e+00 -8.45524192e-01 -4.43271875e-01 2.27664852e+00
5.77222705e-01 2.93968529e-01 3.57816279e-01 5.51502071e-02
4.34669644e-01 4.82392237e-02 -4.46612328e-01 -1.94445655e-01
8.60112458e-02 1.93785384e-01 6.21788681e-01 7.50989795e-01
-9.95814264e-01 9.37260032e-01 5.44743299e+00 5.26749253e-01
-1.06323993e+00 4.26309370e-02 1.09237216e-01 -4.71610159e-01
1.41412452e-01 -9.81827080e-02 -5.57433546e-01 3.82041425e-01
8.51518869e-01 3.61437678e-01 2.15563491e-01 6.71631157e-01
6.32018983e-01 -1.04550498e-04 -1.13501608e+00 8.99071872e-01
-2.09432960e-01 -1.38433743e+00 -3.82754207e-01 1.89739257e-01
5.15903413e-01 1.40288547e-01 -1.36837006e-01 -4.01714481e-02
1.08719870e-01 -8.93832386e-01 1.28236520e+00 8.49345028e-01
7.30769694e-01 -8.16365302e-01 4.83392268e-01 7.31504619e-01
-1.02806330e+00 3.11996400e-01 -8.56343135e-02 -1.62828133e-01
6.60142004e-01 4.12513494e-01 -3.64207536e-01 8.08931589e-01
6.66133881e-01 6.38110816e-01 -2.60760069e-01 1.14219499e+00
-3.60226065e-01 2.24199161e-01 -6.39108777e-01 2.75008947e-01
2.89899498e-01 -5.91029762e-04 9.30807650e-01 1.22131550e+00
4.02626693e-01 6.67017475e-02 1.52841911e-01 6.26313806e-01
4.11926478e-01 -2.66509354e-01 -5.69864690e-01 6.69708073e-01
-1.87154189e-02 1.02466786e+00 -5.51926851e-01 -3.18953916e-02
1.56595826e-01 1.22669327e+00 1.56871840e-01 2.60661602e-01
-9.91602898e-01 -2.65119015e-03 7.35827386e-01 3.17497015e-01
3.02061975e-01 -8.35314035e-01 8.99006277e-02 -1.22756779e+00
1.91935807e-01 -6.27008379e-01 1.58434331e-01 -9.73719835e-01
-7.69204021e-01 4.13990587e-01 3.27734321e-01 -1.26194596e+00
-7.59042680e-01 -7.50175297e-01 -4.10751909e-01 7.99893856e-01
-1.16795444e+00 -1.19450784e+00 -3.08045685e-01 7.00465262e-01
4.27468807e-01 3.48774344e-01 7.77596295e-01 3.35431159e-01
-4.80309844e-01 3.03850859e-01 -1.73108771e-01 -4.12985962e-03
7.50841439e-01 -1.21249735e+00 9.22806680e-01 7.90139496e-01
1.02785779e-02 1.35283381e-01 9.44000423e-01 -5.92201471e-01
-1.51036680e+00 -8.77151966e-01 6.38903260e-01 -8.72980475e-01
3.37322503e-01 -5.08502007e-01 -9.13936913e-01 7.01842904e-01
-2.05433220e-01 2.78961629e-01 -6.44049197e-02 -3.62168163e-01
2.11667672e-01 2.78011471e-01 -8.11533451e-01 8.06797922e-01
1.38399732e+00 -1.46546692e-01 -4.96589214e-01 2.70916373e-01
5.70779741e-01 -1.29307592e+00 -7.74053752e-01 4.32247251e-01
7.33354688e-01 -8.99062395e-01 1.38168633e+00 -5.10288537e-01
1.72882229e-01 -3.91073167e-01 6.64710775e-02 -1.22561610e+00
-4.04115677e-01 -1.20481706e+00 -5.04558206e-01 5.97637177e-01
6.40822649e-02 -3.66300255e-01 9.36024725e-01 4.83297288e-01
-2.80509502e-01 -6.26030445e-01 -1.24470878e+00 -1.05835044e+00
2.22789925e-02 -6.73213720e-01 1.43399075e-01 7.86509931e-01
-3.62175822e-01 2.31616616e-01 -8.39529693e-01 3.21741641e-01
5.96598327e-01 -4.68996674e-01 1.26886153e+00 -1.07977378e+00
-6.14218891e-01 -3.01288068e-01 -6.22061670e-01 -1.19988918e+00
6.73504621e-02 -3.59856099e-01 2.62037367e-01 -1.26450610e+00
-4.07363176e-01 -2.26308510e-01 1.98483646e-01 3.74511480e-01
-4.46442701e-02 1.94842592e-01 3.77691150e-01 2.13509724e-01
-3.03085089e-01 5.35472631e-01 1.40718985e+00 2.28321701e-01
-4.05574054e-01 1.27564564e-01 1.29525289e-01 9.93937314e-01
6.16586685e-01 -4.56127852e-01 -2.87365854e-01 -6.19450271e-01
1.17657341e-01 3.31821352e-01 1.03924966e+00 -1.36594844e+00
1.22101322e-01 -2.72338331e-01 4.49438125e-01 -4.77590799e-01
7.60885060e-01 -8.56316984e-01 7.91019917e-01 6.26016200e-01
6.83525205e-03 2.43293211e-01 4.76091892e-01 4.78603333e-01
2.36676499e-01 7.10795820e-02 7.90181160e-01 -2.58479029e-01
-7.87736058e-01 2.95260608e-01 -2.77317852e-01 5.75485313e-03
9.16521668e-01 -5.72732985e-01 3.90445702e-02 -3.57026100e-01
-9.95051622e-01 9.14969444e-02 8.16670477e-01 4.45001960e-01
4.19765949e-01 -1.38707340e+00 -6.88503861e-01 8.85694027e-02
-1.66118160e-01 4.67749476e-01 3.70259702e-01 9.87249494e-01
-8.56840551e-01 1.77778542e-01 -3.08452308e-01 -9.02387500e-01
-1.14978778e+00 3.92194659e-01 6.42243981e-01 -3.97536993e-01
-1.09175944e+00 5.39003074e-01 7.19019398e-02 -6.16632938e-01
2.08506826e-02 -5.62614858e-01 1.68503150e-01 -5.33172131e-01
1.50886491e-01 4.91575062e-01 1.12298906e-01 -9.76534188e-01
-5.05664706e-01 1.01041055e+00 5.28696656e-01 -3.58044416e-01
1.40087199e+00 -8.27926695e-02 5.14867544e-01 3.02507222e-01
9.18488026e-01 1.33026326e-02 -1.99396193e+00 2.56170005e-01
-2.09188923e-01 -3.59455109e-01 -1.22251906e-01 -6.41980886e-01
-8.67444098e-01 7.80105472e-01 6.07662261e-01 -2.69296408e-01
7.64618814e-01 -2.25612417e-01 7.61352897e-01 4.12581950e-01
3.40900451e-01 -1.02204847e+00 9.78807267e-03 8.02010834e-01
1.05536926e+00 -8.52473199e-01 8.71616825e-02 -3.71512920e-01
-4.78628486e-01 1.03167343e+00 4.81501907e-01 -4.04073238e-01
2.87425637e-01 3.99608403e-01 9.43341479e-02 -1.21375091e-01
-5.35621285e-01 -5.57427704e-02 4.86875534e-01 8.09588253e-01
6.59149662e-02 -1.00797661e-01 1.48249596e-01 2.75128186e-01
-2.05506369e-01 4.54650000e-02 4.53504860e-01 1.06949186e+00
-5.47824763e-02 -9.79557872e-01 -5.60421586e-01 -4.41382796e-01
-3.16020131e-01 3.11109602e-01 -1.63959235e-01 1.30203021e+00
6.01847246e-02 3.26991469e-01 -9.54834372e-02 -3.78731489e-01
8.16688299e-01 -1.51753962e-01 7.21332908e-01 -4.64240968e-01
-6.34109318e-01 2.67580867e-01 2.64286280e-01 -1.01403606e+00
-5.16005576e-01 -6.15508080e-01 -1.34791231e+00 -4.50601250e-01
-2.28236929e-01 -3.02578896e-01 4.98078763e-01 8.10324967e-01
4.53481674e-01 5.54705918e-01 -2.02250276e-02 -1.87652290e+00
-3.49931329e-01 -5.72278023e-01 -2.16530070e-01 3.45675379e-01
5.72891891e-01 -7.72497237e-01 -5.22918850e-02 3.29429567e-01] | [7.12580680847168, -0.7799766063690186] |
cef59fac-18aa-4efc-8c0e-a1bb0d145b7b | exploiting-hybrid-models-of-tensor-train | 2201.10609 | null | https://arxiv.org/abs/2201.10609v1 | https://arxiv.org/pdf/2201.10609v1.pdf | Exploiting Hybrid Models of Tensor-Train Networks for Spoken Command Recognition | This work aims to design a low complexity spoken command recognition (SCR) system by considering different trade-offs between the number of model parameters and classification accuracy. More specifically, we exploit a deep hybrid architecture of a tensor-train (TT) network to build an end-to-end SRC pipeline. Our command recognition system, namely CNN+(TT-DNN), is composed of convolutional layers at the bottom for spectral feature extraction and TT layers at the top for command classification. Compared with a traditional end-to-end CNN baseline for SCR, our proposed CNN+(TT-DNN) model replaces fully connected (FC) layers with TT ones and it can substantially reduce the number of model parameters while maintaining the baseline performance of the CNN model. We initialize the CNN+(TT-DNN) model in a randomized manner or based on a well-trained CNN+DNN, and assess the CNN+(TT-DNN) models on the Google Speech Command Dataset. Our experimental results show that the proposed CNN+(TT-DNN) model attains a competitive accuracy of 96.31% with 4 times fewer model parameters than the CNN model. Furthermore, the CNN+(TT-DNN) model can obtain a 97.2% accuracy when the number of parameters is increased. | ['Javier Tejedor', 'Jun Qi'] | 2022-01-11 | null | null | null | null | ['spoken-command-recognition'] | ['speech'] | [-1.07968085e-01 -1.50053412e-01 -5.27498201e-02 -6.28802896e-01
-8.42869997e-01 -5.22948086e-01 2.85770327e-01 -6.41786516e-01
-6.10325158e-01 -7.68371820e-02 2.85503358e-01 -5.61081588e-01
6.01654291e-01 -3.02774400e-01 -6.22092009e-01 -6.21500552e-01
4.50980477e-02 1.33105993e-01 1.08506233e-01 -1.98754951e-01
6.13825507e-02 4.60628599e-01 -1.47751820e+00 4.10601526e-01
8.81223857e-01 1.64313424e+00 5.02754711e-02 1.13249576e+00
-3.57258767e-02 1.16497076e+00 -8.59395683e-01 -2.75963426e-01
1.23676471e-01 -7.60041699e-02 -8.11436772e-01 -3.28743070e-01
2.93913692e-01 -5.86078227e-01 -6.35136306e-01 8.01027477e-01
7.12233961e-01 1.93601608e-01 2.33155027e-01 -1.11292350e+00
-5.96347630e-01 7.94649184e-01 -1.28506953e-02 1.66514851e-02
1.19332731e-01 4.91871476e-01 1.06211197e+00 -1.20570743e+00
3.62062901e-02 1.16613781e+00 5.81074238e-01 5.35379231e-01
-1.11729884e+00 -6.96937561e-01 2.74547935e-01 1.25946224e-01
-1.61239791e+00 -6.43325090e-01 6.34013891e-01 -1.89002633e-01
1.32390070e+00 2.96457857e-01 4.92491543e-01 1.30478871e+00
-1.57411098e-01 1.14205277e+00 7.11952925e-01 -1.53873742e-01
4.15870994e-01 -2.29008019e-01 4.46392566e-01 6.46046698e-01
-6.74676657e-01 -1.12232141e-01 -7.60267138e-01 -9.83467400e-02
6.62244618e-01 -2.03768745e-01 -4.20988798e-01 4.26673263e-01
-1.05606937e+00 7.57971942e-01 6.15535557e-01 2.44420081e-01
-3.96085352e-01 3.60462546e-01 5.84319413e-01 3.12826365e-01
3.98280412e-01 2.45380044e-01 -8.15262139e-01 -4.96247023e-01
-9.84049916e-01 9.12027061e-02 1.02789390e+00 1.21237111e+00
4.36988354e-01 7.94643283e-01 -2.59415209e-01 1.21620655e+00
1.16819851e-02 2.79014260e-01 6.99565887e-01 -6.61893725e-01
4.81908917e-01 5.24475694e-01 -3.69281352e-01 -3.11010242e-01
-2.29428843e-01 -8.32544684e-01 -9.70434964e-01 -1.38191029e-01
3.74783166e-02 -4.29852366e-01 -1.22004378e+00 1.37563121e+00
-6.65578917e-02 7.69182444e-02 2.42932096e-01 1.15412295e+00
1.03622329e+00 8.07009578e-01 -2.19577719e-02 3.52976590e-01
1.27017295e+00 -1.53742707e+00 -7.37867415e-01 -1.98214501e-01
7.79371858e-01 -8.08068037e-01 1.42866170e+00 5.05014360e-01
-8.31963778e-01 -7.78889358e-01 -8.28703821e-01 -2.58648425e-01
-2.85578877e-01 8.66887212e-01 4.39532787e-01 3.11687857e-01
-1.30088508e+00 2.54481345e-01 -7.50698566e-01 -3.13103139e-01
1.60992831e-01 4.43388790e-01 -1.52496129e-01 7.12428614e-02
-9.38917518e-01 7.08212912e-01 4.25905645e-01 6.86997175e-01
-1.38465977e+00 -6.24911845e-01 -6.33569837e-01 1.27491951e-01
2.09653482e-01 -3.20952326e-01 1.73121226e+00 -1.05130315e+00
-2.19556618e+00 2.78287113e-01 -1.89115539e-01 -4.58223850e-01
3.61214727e-01 -3.29374820e-01 -3.56477827e-01 -7.66266044e-03
-5.41818380e-01 5.51953495e-01 7.76233971e-01 -8.78077447e-01
-5.64809859e-01 -9.79635715e-02 7.75110945e-02 1.31165013e-01
-2.14316845e-01 4.27540928e-01 -5.07119477e-01 -7.25204289e-01
2.25446552e-01 -8.61000478e-01 9.00930017e-02 -1.49374083e-01
-7.29194820e-01 -4.90375936e-01 1.00033092e+00 -8.19370925e-01
1.19677091e+00 -2.38279057e+00 2.98831671e-01 -1.42524794e-01
9.46252123e-02 6.32632554e-01 -3.30679059e-01 4.72497374e-01
-2.79783905e-02 8.13449621e-02 -1.77799895e-01 -7.20300555e-01
-1.40348598e-01 3.18303198e-01 -2.91836977e-01 2.08644882e-01
4.50026900e-01 7.04311848e-01 -4.27990168e-01 6.91754976e-04
5.73806316e-02 5.51662803e-01 -6.00125790e-01 8.26888621e-01
-3.94961804e-01 2.39327028e-01 -2.33713135e-01 9.08935130e-01
7.39090800e-01 4.37912047e-02 -5.74678853e-02 -3.41904134e-01
-3.85677516e-01 7.61065543e-01 -7.74567366e-01 1.56082261e+00
-5.68674266e-01 7.28299797e-01 3.82239252e-01 -7.75619209e-01
1.23465097e+00 3.45842540e-01 -1.08420387e-01 -4.20567960e-01
3.21625561e-01 4.41985995e-01 6.19729571e-02 -5.49800098e-01
4.39670056e-01 3.03610712e-01 -2.96174157e-02 1.11519352e-01
4.31891501e-01 -1.46111041e-01 -4.86478269e-01 -3.54136596e-03
9.19457495e-01 -1.44833401e-01 -4.26083565e-01 -1.82788149e-01
5.14883697e-01 -1.03556514e-01 4.71485794e-01 5.43732405e-01
-1.40930608e-01 7.78266370e-01 5.98318696e-01 -4.97716844e-01
-9.64179337e-01 -6.73629344e-01 -2.47697420e-02 1.43467057e+00
-4.89830226e-01 -3.80959809e-01 -8.91659379e-01 -4.28701490e-01
-2.44612083e-01 7.26949811e-01 -4.60590780e-01 -1.67246014e-01
-5.79833865e-01 -5.58394670e-01 1.22873163e+00 6.25572264e-01
1.10286546e+00 -7.80176044e-01 -1.67523652e-01 1.47368327e-01
-2.49061093e-01 -1.40491998e+00 -7.10461795e-01 6.16341889e-01
-6.06436312e-01 -6.11459792e-01 -5.85132599e-01 -9.40929294e-01
2.61236757e-01 -5.86746074e-02 6.88537419e-01 2.63582498e-01
3.36286366e-01 -2.37147585e-02 -7.99733102e-01 -3.44448537e-01
-2.39771292e-01 6.90214097e-01 1.75683238e-02 4.46969539e-01
3.09817314e-01 -1.79873526e-01 -4.32742923e-01 1.47823825e-01
-8.45465362e-01 1.05697177e-01 4.55787152e-01 1.03934026e+00
2.90770918e-01 -4.31254148e-01 2.86164761e-01 -3.27510893e-01
6.50142312e-01 -1.65808260e-01 -5.08548498e-01 8.85804296e-02
-4.33441132e-01 5.24818636e-02 9.32201147e-01 -7.28476644e-01
-6.70702219e-01 3.17289144e-01 -6.16141140e-01 -8.97915900e-01
-1.85992252e-02 7.76183367e-01 -9.34294537e-02 -1.86289996e-01
3.32281291e-01 6.68534279e-01 8.64628032e-02 -8.29024017e-01
1.54900983e-01 1.17738736e+00 7.43224621e-01 -2.27683544e-01
5.72475135e-01 -2.29547992e-01 -7.94303954e-01 -9.97066319e-01
-7.69825697e-01 -3.36919487e-01 -8.15639257e-01 -2.13965788e-01
8.97573590e-01 -1.22201693e+00 -8.46572876e-01 1.13953257e+00
-1.39702523e+00 -5.96644700e-01 1.71835184e-01 6.20278776e-01
-1.15332998e-01 -6.15437105e-02 -8.27957988e-01 -9.77262914e-01
-6.65267229e-01 -1.32105350e+00 1.16749573e+00 1.49990156e-01
1.92255244e-01 -6.73306823e-01 -3.57163846e-01 4.51048732e-01
8.78806710e-01 -2.20674053e-01 7.36444592e-01 -1.02310085e+00
-5.55621207e-01 -2.43468702e-01 -2.06652611e-01 9.49218273e-01
-3.71281803e-01 3.82960290e-01 -1.57807791e+00 -2.61260003e-01
-7.05402419e-02 -5.12136757e-01 9.32064652e-01 9.75150168e-02
1.38589048e+00 -5.74612379e-01 4.20132071e-01 1.00059664e+00
1.08588707e+00 1.50280789e-01 6.84243560e-01 1.79448240e-02
9.02080953e-01 1.98206007e-01 1.10678457e-01 3.57484668e-01
6.74393892e-01 7.89196014e-01 5.48351943e-01 -2.75890499e-01
-1.07679799e-01 -2.26128712e-01 7.09885955e-01 1.39568698e+00
7.91807249e-02 -2.33556584e-01 -9.62399125e-01 2.95174539e-01
-1.58576417e+00 -3.66820842e-01 -1.54715031e-01 1.99496269e+00
9.11476791e-01 -5.78316115e-02 2.06889912e-01 1.27286077e-01
3.41861725e-01 2.88188398e-01 -5.19886971e-01 -7.67406821e-01
-2.77957976e-01 2.86904991e-01 3.53158385e-01 5.13456106e-01
-9.45589364e-01 1.47301579e+00 6.14329386e+00 8.05428565e-01
-1.60754275e+00 4.23514172e-02 4.70036864e-01 -1.00493282e-01
2.09474452e-02 -6.51747435e-02 -9.16803718e-01 3.19500893e-01
1.41961896e+00 3.98149401e-01 7.97592461e-01 1.13651943e+00
2.47178525e-01 1.72037497e-01 -1.15036726e+00 1.03646541e+00
5.09212576e-02 -1.22284079e+00 1.79904863e-01 -1.47135481e-01
1.83157086e-01 4.45255637e-01 -9.37513039e-02 7.61295021e-01
3.30351919e-01 -1.20707381e+00 1.09319305e+00 4.16634172e-01
9.34788048e-01 -7.92923748e-01 1.05213714e+00 3.65321577e-01
-1.22648478e+00 -7.53365681e-02 -2.64187694e-01 -1.02997929e-01
-2.17656702e-01 4.85651493e-01 -9.38621700e-01 3.95345002e-01
9.81708169e-01 3.79063457e-01 -4.26501721e-01 7.33194172e-01
-1.52534917e-01 1.11755109e+00 -4.02721822e-01 -3.69289249e-01
5.22759140e-01 7.69050270e-02 3.43915932e-02 1.48137057e+00
-1.11687802e-01 1.92864269e-01 2.51917243e-01 9.25495684e-01
-3.99168253e-01 -1.12566121e-01 -1.01236060e-01 -1.13440655e-01
7.57946670e-01 1.22242296e+00 -5.24300002e-02 -3.94124031e-01
-2.63121814e-01 1.05114436e+00 5.06855071e-01 5.29303730e-01
-7.98579991e-01 -6.25018597e-01 9.75560427e-01 -4.43750054e-01
4.40442383e-01 -4.84987646e-01 -2.27855057e-01 -1.37855053e+00
1.68321475e-01 -1.14792526e+00 -1.03014432e-01 -7.66240180e-01
-1.08656597e+00 1.28293180e+00 -4.72423226e-01 -1.04026520e+00
-2.21134678e-01 -6.49580002e-01 -6.71587884e-01 1.13705778e+00
-1.58304238e+00 -1.34709644e+00 -5.75694919e-01 7.53295243e-01
7.88087308e-01 -1.80811882e-01 9.95864034e-01 2.60524869e-01
-1.04815519e+00 1.04313958e+00 -7.44015425e-02 8.00181150e-01
2.30639279e-01 -1.03482890e+00 5.24277031e-01 8.69734704e-01
-3.10762227e-01 5.26003599e-01 2.77790666e-01 -2.04770073e-01
-1.75250590e+00 -1.28754580e+00 8.41357827e-01 -8.22044611e-02
4.79172409e-01 -9.27714169e-01 -1.01917827e+00 6.95294559e-01
2.21076131e-01 1.41013280e-01 4.64397818e-01 -1.32076263e-01
-8.23211551e-01 -2.41810396e-01 -8.63300145e-01 4.44736302e-01
7.18515217e-01 -1.01293898e+00 -1.82975739e-01 2.61429071e-01
1.22465599e+00 -5.94728708e-01 -1.03987908e+00 2.51744717e-01
6.06002688e-01 -7.90735006e-01 4.89215970e-01 -6.61502421e-01
2.17691988e-01 -7.71823004e-02 -5.60184300e-01 -1.32097173e+00
-4.62548109e-03 -7.25926161e-01 1.14343077e-01 1.24315310e+00
6.57071114e-01 -5.99386513e-01 4.74807501e-01 5.71686268e-01
-7.58336127e-01 -9.28601027e-01 -1.20153332e+00 -7.35508621e-01
4.67590727e-02 -6.88564479e-01 6.60245478e-01 7.40551889e-01
-4.16628331e-01 4.59276468e-01 -2.70626664e-01 3.56138140e-01
1.20536707e-01 -2.33672410e-01 8.49574208e-01 -9.55947340e-01
-7.32807741e-02 -3.10564309e-01 -9.80562121e-02 -1.48302698e+00
1.19927652e-01 -7.29446828e-01 2.03891575e-01 -1.13985693e+00
-1.49080649e-01 -4.27851081e-01 -8.21786746e-02 8.29268456e-01
6.95188269e-02 -2.07850456e-01 3.28723222e-01 3.70079279e-01
-1.91647246e-01 9.60429788e-01 8.64318371e-01 -1.03101738e-01
-2.41377234e-01 -9.20171216e-02 -1.40041918e-01 4.66928035e-01
6.65061891e-01 -1.82191879e-01 8.21985975e-02 -7.55062699e-01
-4.36239064e-01 4.30429168e-02 4.30691689e-01 -9.86038506e-01
5.16884804e-01 1.85948446e-01 1.55096874e-01 -7.31000662e-01
4.81555492e-01 -5.93795836e-01 -2.18969151e-01 2.53021985e-01
-5.21853089e-01 1.53508708e-01 9.81473997e-02 -3.85858142e-03
-4.73685592e-01 -1.10787302e-01 8.33757102e-01 1.48860097e-01
-4.20476913e-01 1.39479831e-01 -6.36314690e-01 -3.19480181e-01
3.98381770e-01 6.84168413e-02 -4.97789025e-01 -5.61687946e-01
-5.66852927e-01 2.64737010e-01 -7.84644932e-02 4.93886322e-01
8.24650228e-01 -1.06762981e+00 -6.51982903e-01 5.07164776e-01
1.61798503e-02 3.38320106e-01 1.15331843e-01 8.64552379e-01
-6.80188596e-01 5.34898221e-01 1.91861272e-01 -7.33919501e-01
-9.69842255e-01 3.91936079e-02 8.13028634e-01 6.81970119e-02
-2.71686614e-01 1.21328843e+00 -2.72779614e-01 -1.06321597e+00
1.02060974e+00 -8.01418662e-01 2.86009256e-02 -2.06571236e-01
3.63116652e-01 2.78194129e-01 4.36594576e-01 -4.90726203e-01
-4.75585878e-01 3.00620139e-01 -2.66185105e-01 -1.02619700e-01
1.48502970e+00 3.43551278e-01 -1.47952139e-02 4.49346483e-01
1.44367707e+00 -4.86318171e-01 -1.24189806e+00 -2.01521203e-01
-1.93370879e-01 -3.36649716e-02 5.78357041e-01 -9.61830854e-01
-1.54346097e+00 1.07122183e+00 5.56239724e-01 -7.42936730e-02
1.34881198e+00 -1.60444960e-01 9.25182700e-01 5.51095247e-01
2.26158023e-01 -1.00546777e+00 1.78775638e-01 1.02834618e+00
1.17189860e+00 -9.81234074e-01 -5.97344100e-01 -3.66627604e-01
-9.40217972e-01 1.33035612e+00 7.30433822e-01 -2.05935270e-01
6.13874257e-01 4.94370311e-01 2.87440777e-01 -8.57843235e-02
-1.12116945e+00 -2.00686768e-01 1.49356514e-01 2.94465512e-01
4.66191322e-01 1.01730816e-01 3.60887378e-01 7.34856904e-01
-5.34679830e-01 8.93327221e-02 5.37987053e-01 7.40109563e-01
-1.10946245e-01 -6.71377361e-01 -2.00439338e-02 4.25804079e-01
-2.65788347e-01 -4.62260664e-01 -5.89033008e-01 4.95928407e-01
-2.61045963e-01 1.23739183e+00 1.77071065e-01 -1.31864989e+00
6.20333552e-01 3.47190589e-01 -2.09045798e-01 -5.59662223e-01
-1.39840913e+00 3.63737732e-01 2.60305613e-01 -6.36091530e-01
4.74323854e-02 -3.38396102e-01 -1.20301592e+00 -5.83645940e-01
-6.09743118e-01 -3.28858681e-02 1.00227404e+00 7.30173290e-01
6.08878613e-01 6.15899801e-01 9.46088731e-01 -7.31228173e-01
-9.98552322e-01 -1.52681386e+00 -4.99623448e-01 -3.03254783e-01
6.32468402e-01 -2.22218826e-01 -4.98348624e-01 -7.92779922e-02] | [14.271072387695312, 6.313569068908691] |
cc7d47f9-72b8-4aed-832e-12d970a878c5 | deep-imitator-handwriting-calligraphy | null | null | https://www.sciencedirect.com/science/article/pii/S0031320319303814?via%3Dihub | https://www.sciencedirect.com/science/article/pii/S0031320319303814?via%3Dihub | Deep imitator: Handwriting calligraphy imitation via deep attention networks | Calligraphy imitation (CI) from a handful of target handwriting samples is such a challenging task that most of the existing writing style analysis or handwriting generation methods do not exhibit satisfactory performance. In this paper, we propose a novel multi-module framework to address the problem of CI. Firstly, we utilized a deep convolution neural network (CNN) to extract personalized calligraphical features. Then we built a calligraphy-clustering attention module and a mata-style matrix (msM) to compute an embedding of calligraphy. The structure of conditional gated recurrent unit (cGRU) is then improved to predict the probabilistic density of pen tip movement displacement by dual condition inputs. Finally, we generated personalized handwriting stroke sequences through iterative sampling with Gaussian mixture model (GMM). Experiments on public online handwriting databases verify that the proposed method could achieve satisfactory performance; the generated samples achieved high similarities with original handwriting examples. | ['Ye Bai', 'Cunhang Fan', 'Zhengkun Tian', 'Minghao Yang', 'JianHua Tao', 'Bocheng Zhao'] | 2020-08-01 | null | null | null | pattern-recognition-2020-8 | ['deep-attention', 'handwriting-generation', 'deep-attention'] | ['computer-vision', 'computer-vision', 'natural-language-processing'] | [ 2.37320676e-01 -3.43471378e-01 -2.62686431e-01 -2.34386846e-01
-5.79498827e-01 -7.18167245e-01 8.44444811e-01 -8.03869843e-01
7.34721264e-03 6.19453609e-01 3.85524929e-01 -3.17006022e-01
-7.90470243e-02 -6.57817960e-01 -7.23505199e-01 -7.52610147e-01
7.59092510e-01 4.97533917e-01 -2.14038074e-01 7.07281055e-03
8.36306334e-01 8.38057458e-01 -1.10795474e+00 3.69521409e-01
1.04680169e+00 3.24044764e-01 7.59535074e-01 1.15049958e+00
-2.85857379e-01 7.89861262e-01 -8.65905523e-01 -4.06911314e-01
-2.06070095e-02 -6.22142375e-01 -5.85380018e-01 5.04161716e-01
1.23023771e-01 -5.99776387e-01 -6.93130553e-01 8.18913937e-01
6.92353547e-01 2.74857014e-01 1.26220989e+00 -8.11951160e-01
-1.34544933e+00 7.56307960e-01 -4.91305351e-01 -5.05018421e-02
-9.18900874e-03 2.59098023e-01 6.34644687e-01 -9.86519217e-01
7.72388697e-01 1.24760568e+00 5.07012010e-01 5.84673464e-01
-7.90547252e-01 -3.77164364e-01 9.95347574e-02 2.17152104e-01
-1.01691115e+00 5.32721207e-02 9.86073732e-01 -5.45677245e-01
9.63022947e-01 1.38658002e-01 3.97604972e-01 1.58649790e+00
2.90546745e-01 1.41477942e+00 1.03557515e+00 -4.26579595e-01
-6.66267052e-02 1.02116965e-01 3.64654213e-01 5.77715933e-01
-4.05577153e-01 -1.12145104e-01 -2.03000754e-01 1.07887365e-01
1.36408198e+00 2.52669692e-01 -2.11468935e-02 2.03349441e-01
-1.06574988e+00 6.42356575e-01 1.07211284e-01 4.37636137e-01
-1.75636649e-01 3.51904659e-03 3.73138666e-01 3.79287265e-02
-1.09722257e-01 3.70193958e-01 3.24753784e-02 -5.01438856e-01
-1.00379562e+00 2.49139339e-01 6.65884614e-01 1.46638918e+00
2.61134356e-01 3.11455041e-01 -7.05088556e-01 1.12887049e+00
1.79536656e-01 6.46534145e-01 9.44765151e-01 -5.67954361e-01
7.55850136e-01 5.37285805e-01 3.41950241e-03 -1.04954255e+00
-9.34218690e-02 -1.67145327e-01 -9.95816469e-01 3.09347734e-02
3.22458237e-01 -1.12118006e-01 -9.77607131e-01 1.11769783e+00
-4.96741802e-01 4.48239073e-02 -6.46565855e-02 1.06368828e+00
3.90109211e-01 8.98440838e-01 -3.04911196e-01 1.85412154e-01
9.43270028e-01 -1.14350855e+00 -9.56946313e-01 1.88094497e-01
1.13019086e-01 -1.12847018e+00 1.52134323e+00 4.57280129e-01
-9.69844937e-01 -9.42126811e-01 -1.07455516e+00 -2.43039623e-01
-6.05864637e-02 1.14045250e+00 4.77192134e-01 6.42902851e-01
-4.52553988e-01 9.26920116e-01 -9.43483829e-01 -2.71951947e-02
5.40529609e-01 6.15272038e-02 -1.15377277e-01 5.40281311e-02
-5.09322107e-01 5.99279284e-01 3.08340907e-01 3.56115282e-01
-9.78706717e-01 -2.71388263e-01 -6.14378035e-01 -2.97173765e-02
-1.01284340e-01 -1.94466427e-01 1.02284610e+00 -7.33542740e-01
-2.30906081e+00 4.13549870e-01 -2.99879640e-01 3.58035117e-02
7.00847268e-01 -3.83780748e-01 -2.43862405e-01 -1.56186789e-01
-9.16218460e-02 1.91722527e-01 1.31391263e+00 -1.06515205e+00
-3.81289303e-01 -3.60320717e-01 -6.02922976e-01 3.20731789e-01
-3.33213329e-01 9.09709036e-02 -7.02439070e-01 -1.27438581e+00
7.63041005e-02 -8.37757587e-01 7.68341720e-02 -5.67416906e-01
-9.37921107e-01 -3.28701794e-01 8.32478106e-01 -1.21589732e+00
1.23864388e+00 -2.00562620e+00 5.76547205e-01 2.79848009e-01
-1.63810268e-01 4.13942158e-01 -1.98771685e-01 5.79590917e-01
3.67286474e-01 -1.20844513e-01 -3.12284529e-01 -4.95869219e-01
2.38065034e-01 3.48865800e-02 -8.15668583e-01 2.50914603e-01
3.61884564e-01 1.01954830e+00 -9.29810703e-01 -2.49794975e-01
3.83806854e-01 3.62877250e-01 -4.13267732e-01 5.66827834e-01
-3.70179951e-01 3.54852378e-01 -4.00033653e-01 7.03588247e-01
4.57438976e-01 -6.68592378e-02 2.22238824e-01 -1.95623264e-02
4.83923964e-02 -1.77341938e-01 -9.88957882e-01 1.50724542e+00
-4.75670695e-01 7.64408886e-01 -5.32122731e-01 -4.50667948e-01
1.43577492e+00 -5.60103469e-02 -1.48623109e-01 -3.06056350e-01
1.08103439e-01 2.38438457e-01 -3.16029601e-02 -7.89011419e-01
7.95898914e-01 3.64270151e-01 2.33285248e-01 8.83148730e-01
2.34418407e-01 -2.31783301e-01 -1.39694452e-01 2.90999133e-02
8.36817622e-01 5.85797727e-01 -3.49876434e-01 6.45733019e-03
4.61948127e-01 -1.77650928e-01 3.01690340e-01 8.37170362e-01
1.36605710e-01 1.12119305e+00 5.10303736e-01 -1.25833765e-01
-1.44034815e+00 -1.06901884e+00 1.74024150e-01 7.44811833e-01
-5.78712262e-02 -2.19350800e-01 -9.11174536e-01 -4.29657221e-01
-8.55912119e-02 7.60885000e-01 -4.55897719e-01 -9.62816551e-02
-8.03640664e-01 -7.51485884e-01 8.53894472e-01 1.06016481e+00
6.84044719e-01 -1.52809620e+00 -5.67591116e-02 4.40467387e-01
7.16589317e-02 -6.87127054e-01 -9.96747434e-01 -2.24908307e-01
-7.17233777e-01 -8.54075432e-01 -1.35219932e+00 -1.08690250e+00
7.61473179e-01 4.49116807e-03 2.26489916e-01 -2.93066263e-01
-6.04963243e-01 -8.47680587e-03 -2.92180598e-01 -2.61244416e-01
-5.30289710e-01 2.93097377e-01 1.25310868e-01 1.59877300e-01
4.47371989e-01 -4.84288424e-01 -3.24281573e-01 1.72913700e-01
-6.23556972e-01 1.48555622e-01 7.72479534e-01 1.20148468e+00
4.50633973e-01 -2.37184748e-01 5.23921013e-01 -6.62907839e-01
1.23514366e+00 -1.48097128e-01 -4.12124664e-01 4.15834367e-01
-4.29793835e-01 -2.91842949e-02 1.06435394e+00 -7.24324703e-01
-1.42487800e+00 3.27569619e-02 -1.08509891e-01 -6.95524156e-01
-1.80610463e-01 1.15037553e-01 -2.46214077e-01 4.01618212e-01
3.57176632e-01 9.71721411e-01 2.06932619e-01 -7.68170118e-01
5.23839891e-01 1.27549672e+00 8.67679834e-01 -7.04453409e-01
2.39774898e-01 7.32212439e-02 -4.66520071e-01 -1.00138915e+00
-9.51612294e-02 6.94540888e-03 -8.39352369e-01 -1.10854730e-01
9.03939545e-01 -4.30714339e-01 -9.29746807e-01 1.11718452e+00
-1.22734010e+00 -6.21136904e-01 1.59734756e-01 4.28297281e-01
-9.75752532e-01 7.68749237e-01 -1.16036034e+00 -9.82990384e-01
-2.85306811e-01 -1.33219123e+00 1.07458246e+00 4.76015806e-01
-6.72612190e-02 -7.29553878e-01 9.23771411e-03 2.93298960e-01
3.39994878e-01 -1.13030963e-01 1.09799755e+00 -3.54589701e-01
-6.05594456e-01 -2.92138875e-01 -3.11662316e-01 5.31820893e-01
4.60291654e-01 4.09649521e-01 -9.61884797e-01 -6.36289790e-02
-2.50525504e-01 -1.74181193e-01 8.15656960e-01 4.09737319e-01
1.48346674e+00 -2.69721776e-01 -9.35062319e-02 7.93152511e-01
9.88667488e-01 4.17316467e-01 8.01893294e-01 1.71105698e-01
1.07077241e+00 2.09242314e-01 5.68780482e-01 7.32274175e-01
2.23660283e-03 6.10943794e-01 -2.31050953e-01 2.88028002e-01
-3.07484984e-01 -7.42674768e-01 5.14755785e-01 1.12334609e+00
-4.51108545e-01 -2.02457294e-01 -6.43248379e-01 4.40831006e-01
-1.92125237e+00 -1.06417549e+00 6.43224642e-02 1.71351779e+00
9.21434939e-01 1.46309985e-02 2.46977597e-01 3.79107781e-02
8.35319161e-01 -1.38851386e-02 -7.37914741e-01 -5.06438732e-01
-3.86568218e-01 4.84216511e-02 4.26568687e-01 4.47099686e-01
-9.22925889e-01 1.37711155e+00 5.74924850e+00 1.12324548e+00
-1.06577539e+00 -2.32516646e-01 5.92392266e-01 2.82054603e-01
-2.36933216e-01 -3.69936049e-01 -1.11979067e+00 8.51135433e-01
3.52103919e-01 -3.57000157e-02 6.19643748e-01 8.57585847e-01
1.09676562e-01 4.14569557e-01 -8.62325847e-01 1.17459595e+00
3.35023403e-01 -1.42736995e+00 3.07710081e-01 -1.66342761e-02
8.39867890e-01 -3.67357850e-01 2.65418142e-01 2.34635010e-01
3.71848792e-01 -1.24446559e+00 5.26727498e-01 1.24647009e+00
9.27267194e-01 -7.54354835e-01 4.92183417e-01 5.46435177e-01
-7.77657330e-01 -2.19997421e-01 -7.48465240e-01 1.72036022e-01
8.92512500e-02 1.25616118e-01 -8.65444899e-01 4.30387497e-01
1.58432275e-01 8.50053370e-01 -6.49045587e-01 7.31869340e-01
-2.53976971e-01 7.34102547e-01 5.70980422e-02 -6.70649946e-01
4.71204370e-01 -6.12316430e-01 4.56854969e-01 1.48160243e+00
3.36045384e-01 -1.08625643e-01 -3.46350938e-01 1.30795908e+00
7.96611831e-02 1.46882450e-02 -5.45144618e-01 -5.97869217e-01
5.85000336e-01 1.15639186e+00 -6.65662229e-01 -3.58936816e-01
-3.25733684e-02 1.73155689e+00 2.74184734e-01 4.98695135e-01
-6.52244687e-01 -8.81662309e-01 1.79389790e-01 -5.03220737e-01
2.84489959e-01 -3.45230937e-01 -4.66997772e-01 -1.25803518e+00
1.83491096e-01 -1.02961433e+00 -2.81044811e-01 -7.54819870e-01
-1.58804333e+00 5.32412887e-01 -5.25114298e-01 -1.15625405e+00
-3.46567988e-01 -1.00774348e+00 -9.84823048e-01 1.33015752e+00
-7.72697628e-01 -1.35928237e+00 -3.28629375e-01 5.66737533e-01
1.08736885e+00 -1.02508044e+00 8.07779431e-01 6.34124801e-02
-8.21607351e-01 9.21293259e-01 6.21064723e-01 4.72792119e-01
6.11099780e-01 -1.43042481e+00 6.97790265e-01 6.69824421e-01
4.28192224e-03 9.49507952e-01 1.75543875e-01 -9.12730813e-01
-1.65863407e+00 -1.13994396e+00 4.57040578e-01 -5.12742460e-01
4.55024123e-01 -3.40101182e-01 -7.17930794e-01 8.45987260e-01
8.94299969e-02 -7.05521524e-01 3.87631506e-01 -2.46617630e-01
-6.51532859e-02 4.59990740e-01 -6.75254643e-01 9.64918673e-01
9.68107879e-01 -7.29432583e-01 -6.66450143e-01 3.27897906e-01
3.09248030e-01 -4.52664614e-01 -8.27657878e-01 -7.95724615e-02
7.67258167e-01 -5.97150803e-01 7.26527333e-01 -8.66082788e-01
9.26542699e-01 -1.79555938e-01 -1.63484305e-01 -1.21687162e+00
-5.59030116e-01 -7.47540653e-01 -4.28622186e-01 1.48305726e+00
1.05138727e-01 -2.62488514e-01 9.19860363e-01 4.55346018e-01
-4.58137840e-01 -7.23024487e-01 -3.18753004e-01 -9.05867040e-01
2.35883847e-01 -4.97299843e-02 6.24555647e-01 6.92571521e-01
-1.58808827e-02 2.02814206e-01 -7.30932593e-01 -1.62034795e-01
5.35823464e-01 2.73702949e-01 9.77208614e-01 -4.70985562e-01
-7.07638919e-01 -7.55024195e-01 -3.05636197e-01 -1.35277343e+00
2.98496187e-01 -1.11777985e+00 -5.06092533e-02 -1.25401831e+00
3.31091017e-01 -2.33848527e-01 1.73377499e-01 4.27254364e-02
-4.54101652e-01 -1.53221473e-01 3.38086605e-01 3.82094860e-01
-3.20211165e-02 8.51248384e-01 1.41714025e+00 -3.30474526e-01
-4.14186299e-01 2.96827406e-01 -2.30150014e-01 3.85803789e-01
8.66868496e-01 1.72069773e-01 -3.13158303e-01 -4.14708376e-01
-3.33985627e-01 1.08875297e-01 1.63150653e-01 -7.54062712e-01
2.15481967e-01 -9.25991684e-02 7.26439238e-01 -9.50721085e-01
2.18317017e-01 -2.36766949e-01 -9.98887196e-02 2.30269641e-01
-5.46880841e-01 -5.25440276e-03 4.60981503e-02 4.36340332e-01
-2.54087541e-02 -5.53907812e-01 4.38573569e-01 1.17452079e-02
-6.05300128e-01 2.46530995e-01 -7.24735022e-01 -5.17372787e-01
7.65012443e-01 -2.49059215e-01 -1.68171600e-01 -1.72842860e-01
-7.06565261e-01 -6.48278818e-02 3.68843526e-01 7.16440439e-01
1.02886224e+00 -1.53926444e+00 -6.96914732e-01 4.65268493e-01
-1.02231614e-01 -3.65661234e-01 2.77028590e-01 3.72392207e-01
-8.04620683e-01 4.56970006e-01 -4.65608984e-01 -4.91070598e-01
-1.05607653e+00 4.25935924e-01 2.70152718e-01 -3.56873334e-03
-9.57289636e-01 8.08237672e-01 -3.12132478e-01 -7.71035790e-01
3.80415320e-01 -1.22942217e-01 -2.02849001e-01 -2.61770338e-01
6.74579382e-01 5.35989165e-01 -3.31901640e-01 -4.72566545e-01
2.28592604e-01 6.68471277e-01 -3.43625486e-01 -3.55644912e-01
1.27477312e+00 8.22824687e-02 1.10242866e-01 5.24760902e-01
1.13558590e+00 -1.69671863e-01 -1.56886852e+00 9.77008976e-03
-1.48263006e-02 -5.55551648e-01 -4.43813115e-01 -7.06312418e-01
-7.83430636e-01 1.17168713e+00 4.03688371e-01 -4.54484344e-01
6.55576587e-01 -5.15050054e-01 7.80492544e-01 7.66576588e-01
1.54133439e-01 -1.54392946e+00 2.48560265e-01 6.70842052e-01
1.15371180e+00 -1.03658175e+00 -4.43906844e-01 9.16402880e-03
-1.04423869e+00 1.67860568e+00 6.76175833e-01 -5.61869621e-01
4.38471109e-01 1.30356401e-01 6.95542768e-02 1.04536422e-01
-2.44826764e-01 3.76111507e-01 1.83745682e-01 5.53679883e-01
4.82149243e-01 2.01126054e-01 -2.74561346e-01 1.11915016e+00
-3.06908876e-01 1.75291464e-01 5.70010483e-01 9.38878655e-01
-1.68691292e-01 -1.05156791e+00 -2.45352551e-01 7.55069733e-01
-9.49529633e-02 -2.76215952e-02 -5.77105701e-01 3.54898334e-01
-4.07765061e-01 5.30131698e-01 9.07761604e-02 -6.24092400e-01
1.12656012e-01 1.20397933e-01 7.71720767e-01 -5.47357976e-01
-4.41899359e-01 1.97809532e-01 -3.44866574e-01 -3.32510561e-01
1.76102772e-01 -6.58458352e-01 -1.09392405e+00 -2.29037225e-01
-1.83985770e-01 -3.56519222e-01 6.62533104e-01 7.99527645e-01
3.26369882e-01 6.33269966e-01 8.59300375e-01 -1.16469455e+00
-9.36629355e-01 -1.32440376e+00 -9.91625190e-01 4.89361316e-01
-6.81727231e-02 -2.64073282e-01 7.05078766e-02 3.89288068e-01] | [11.911980628967285, 2.252321720123291] |
9a73e483-8678-46c1-bbdf-8dd2ebd67e4e | danish-stance-classification-and-rumour | 1907.01304 | null | https://arxiv.org/abs/1907.01304v1 | https://arxiv.org/pdf/1907.01304v1.pdf | Danish Stance Classification and Rumour Resolution | The Internet is rife with flourishing rumours that spread through microblogs and social media. Recent work has shown that analysing the stance of the crowd towards a rumour is a good indicator for its veracity. One state-of-the-art system uses an LSTM neural network to automatically classify stance for posts on Twitter by considering the context of a whole branch, while another, more simple Decision Tree classifier, performs at least as well by performing careful feature engineering. One approach to predict the veracity of a rumour is to use stance as the only feature for a Hidden Markov Model (HMM). This thesis generates a stance-annotated Reddit dataset for the Danish language, and implements various models for stance classification. Out of these, a Linear Support Vector Machine provides the best results with an accuracy of 0.76 and macro F1 score of 0.42. Furthermore, experiments show that stance labels can be used across languages and platforms with a HMM to predict the veracity of rumours, achieving an accuracy of 0.82 and F1 score of 0.67. Even higher scores are achieved by relying only on the Danish dataset. In this case veracity prediction scores an accuracy of 0.83 and an F1 of 0.68. Finally, when using automatic stance labels for the HMM, only a small drop in performance is observed, showing that the implemented system can have practical applications. | ['Anders Edelbo Lillie', 'Emil Refsgaard Middelboe'] | 2019-07-02 | null | null | null | null | ['rumour-detection'] | ['natural-language-processing'] | [-4.43091035e-01 3.81478459e-01 -3.61842722e-01 -1.68622226e-01
-2.41400018e-01 -2.62791425e-01 9.85668838e-01 5.09746850e-01
-2.32925236e-01 8.53188872e-01 2.08858147e-01 -3.96474212e-01
3.57258111e-01 -1.02967906e+00 -2.43967667e-01 -6.08915210e-01
1.07277632e-01 5.46491146e-01 3.80478978e-01 -6.18038177e-01
5.39274335e-01 2.13338479e-01 -1.81387579e+00 6.01078331e-01
4.75724161e-01 9.94374573e-01 -3.07361454e-01 6.61515057e-01
-4.16861385e-01 1.45560324e+00 -1.16846132e+00 -4.75683331e-01
-2.62980819e-01 -4.63564873e-01 -8.55014026e-01 -1.12846226e-01
-7.18979985e-02 -1.28948614e-01 -5.30631188e-03 6.94145918e-01
2.25548595e-01 -3.00051510e-01 5.15670300e-01 -1.15306735e+00
-2.37781331e-01 7.81386197e-01 -1.41977534e-01 2.33166903e-01
7.90201306e-01 -6.15068153e-02 7.81401515e-01 -5.25249839e-01
6.44823670e-01 8.43819559e-01 8.32470953e-01 -7.60073513e-02
-8.77350807e-01 -5.04945755e-01 -5.12762427e-01 3.45336705e-01
-1.03424728e+00 -2.75745690e-01 5.94322920e-01 -9.50576186e-01
1.03366089e+00 3.65700454e-01 8.26590717e-01 1.19158721e+00
3.71545464e-01 4.36314911e-01 1.61675155e+00 -3.92039299e-01
1.98146567e-01 7.75580525e-01 3.00710440e-01 5.84126115e-01
2.06950739e-01 -2.12097019e-01 -5.28161526e-01 -3.55173856e-01
2.49430642e-01 -2.33397111e-01 1.27546221e-01 4.23060536e-01
-9.34732556e-01 1.26054847e+00 1.43309504e-01 7.95831144e-01
-5.99423170e-01 -4.73408192e-01 9.12479520e-01 7.96582878e-01
9.14353430e-01 3.09952468e-01 -7.00973198e-02 -4.27952230e-01
-1.52421188e+00 1.75570518e-01 1.50905168e+00 1.68688491e-01
5.74689925e-01 1.63037062e-01 9.96476039e-02 7.80728698e-01
1.80541828e-01 2.45715156e-01 7.81763375e-01 -7.25275755e-01
-4.90346458e-03 6.89524353e-01 1.69843376e-01 -1.36948323e+00
-6.95242226e-01 -6.57008886e-01 -5.54233670e-01 2.74897724e-01
4.11616057e-01 -1.92689031e-01 -4.18976843e-01 1.08552003e+00
2.00042993e-01 -2.25110948e-01 7.80587923e-03 5.48375368e-01
9.94169533e-01 7.38124609e-01 -1.88877776e-01 -5.78135967e-01
1.10683942e+00 -5.99230111e-01 -9.17466342e-01 -1.13212556e-01
8.07406306e-01 -9.48296070e-01 5.65971375e-01 7.25045562e-01
-9.80190039e-01 -1.41469836e-01 -1.19937003e+00 3.28285307e-01
-5.98822236e-01 -3.21652770e-01 2.65499383e-01 9.30388689e-01
-8.98967981e-01 8.31851244e-01 -4.36161757e-01 -4.12860692e-01
-5.29097728e-02 9.90110487e-02 -1.14443548e-01 3.89942110e-01
-1.43740022e+00 1.40371299e+00 4.87737566e-01 -2.24383146e-01
-3.59838158e-01 -2.60549933e-02 -5.72531044e-01 -3.31240773e-01
-6.36115521e-02 -2.22964302e-01 1.07506478e+00 -1.10786569e+00
-1.66106534e+00 1.05504453e+00 7.94356689e-03 -8.15419912e-01
9.37803686e-01 1.53296500e-01 -6.61385059e-01 7.86043257e-02
3.25290859e-01 -1.17869414e-01 6.50919259e-01 -1.04322326e+00
-7.15720177e-01 -1.90229654e-01 -1.74424976e-01 -3.62629116e-01
-2.06775993e-01 4.60762799e-01 4.39401299e-01 -1.91231281e-01
3.52613404e-02 -7.71692455e-01 3.58577877e-01 -9.09518123e-01
-2.11735755e-01 -4.67860043e-01 7.06800222e-01 -1.14397514e+00
1.66541457e+00 -1.62577641e+00 -4.85752225e-01 1.59995824e-01
2.90487379e-01 3.31679195e-01 6.07999504e-01 7.31680751e-01
1.85626060e-01 2.52972424e-01 1.58151880e-01 -1.73263848e-01
6.09974861e-02 1.74302712e-01 -1.99512199e-01 8.13128114e-01
-1.98964536e-01 4.82863963e-01 -6.80226088e-01 -4.69625711e-01
1.29337544e-02 4.09876347e-01 -1.10110946e-01 -3.42995278e-03
-7.94646665e-02 5.10679521e-02 -3.02511938e-02 3.25092435e-01
2.46691719e-01 -2.83811957e-01 2.05043286e-01 4.22283679e-01
-4.94132876e-01 5.78758538e-01 -7.03267395e-01 5.15825629e-01
-5.79408228e-01 1.15975749e+00 -1.55544564e-01 -9.83188748e-01
1.54015350e+00 7.03185081e-01 4.38183606e-01 -6.54464304e-01
4.81219560e-01 6.86689019e-01 -7.34402016e-02 -7.81528652e-01
5.95111072e-01 -4.45362896e-01 -1.03203006e-01 6.50030375e-01
-1.93064407e-01 2.23471433e-01 2.48709276e-01 -8.08494091e-02
8.76872659e-01 -2.42704540e-01 5.45317173e-01 -2.37064045e-02
6.70876563e-01 1.40112042e-01 1.87469900e-01 5.22626579e-01
-2.16625258e-01 3.19050729e-01 9.50284183e-01 -5.42589009e-01
-1.18955457e+00 -2.83730417e-01 -1.46674305e-01 1.15743077e+00
-4.47106540e-01 -1.19669899e-01 -8.26181948e-01 -4.35289234e-01
-1.63112059e-01 1.06877816e+00 -7.67842710e-01 2.91471004e-01
-4.97493207e-01 -7.45094299e-01 6.79731965e-01 -1.65990025e-01
6.11383021e-01 -1.15537918e+00 -5.78804910e-01 4.45678294e-01
-5.71435809e-01 -1.02386701e+00 5.07246792e-01 -8.16091076e-02
-6.71310246e-01 -7.89127707e-01 -6.64119363e-01 -4.77128923e-01
-2.31532380e-01 -1.91098064e-01 9.65435386e-01 2.84259021e-01
5.49643099e-01 -3.14059645e-01 -5.00739217e-01 -3.50866228e-01
-1.19268489e+00 2.09586173e-01 4.44027409e-02 -2.75752544e-02
4.59969342e-01 -5.75277746e-01 -1.43222110e-02 4.36005622e-01
-5.50513327e-01 -2.82564133e-01 2.95011371e-01 6.67312562e-01
-2.43343905e-01 -7.48347715e-02 9.04665411e-01 -9.99479413e-01
8.52982163e-01 -9.31073368e-01 -8.32137167e-02 -3.22797239e-01
-8.11827242e-01 -4.81757760e-01 5.74954033e-01 -2.25251466e-01
-6.23314381e-01 -6.31196141e-01 -4.20465976e-01 1.81856453e-01
-3.10659111e-01 8.12372088e-01 5.79090476e-01 1.28798172e-01
1.14857340e+00 2.23642007e-01 4.19893444e-01 -3.47865999e-01
-1.77784756e-01 1.29066861e+00 1.48505792e-01 3.17235023e-01
2.39817768e-01 3.90591681e-01 -2.29229912e-01 -1.33517659e+00
-1.03691518e+00 -7.24694073e-01 -4.92729574e-01 -5.63152671e-01
5.64341664e-01 -7.20799029e-01 -9.16333258e-01 6.10782027e-01
-1.16906047e+00 -2.20839918e-01 2.15083271e-01 3.83193433e-01
-5.50249696e-01 4.09922749e-01 -9.36868787e-01 -1.09957242e+00
-3.10550392e-01 -6.79071963e-01 1.80613950e-01 -1.33519188e-01
-8.51607740e-01 -1.19365203e+00 2.58504540e-01 7.03234434e-01
6.55235052e-01 7.44109273e-01 4.04845655e-01 -1.30399740e+00
2.71874815e-01 -6.80339456e-01 1.47835799e-02 3.37395102e-01
-1.84743986e-01 1.00014627e-01 -1.00226760e+00 -1.44783363e-01
4.11959022e-01 -1.71825051e-01 6.62011862e-01 1.30971983e-01
4.87007014e-02 -8.56377780e-01 -2.52914522e-02 -2.08409041e-01
1.20176399e+00 -1.02193348e-01 6.34030342e-01 1.21682584e+00
2.52047241e-01 8.80726576e-01 6.16795421e-01 6.25602067e-01
4.18948710e-01 5.66610098e-01 5.02129972e-01 2.45899722e-01
1.32814452e-01 -5.96600445e-03 6.89208746e-01 9.29062068e-01
-1.18354149e-01 -5.82310222e-02 -1.18891406e+00 4.26841557e-01
-1.69604063e+00 -1.33694994e+00 -9.28925514e-01 2.06080174e+00
7.01331258e-01 7.65532553e-01 7.74282634e-01 7.18697906e-01
8.24480712e-01 2.95935482e-01 2.53508985e-01 -1.11663115e+00
-1.04556546e-01 -3.05332780e-01 4.05979812e-01 7.90286064e-01
-9.28559959e-01 5.84704041e-01 6.39113903e+00 5.33098876e-01
-1.47961664e+00 2.56989390e-01 2.80636728e-01 4.94686142e-02
4.78392914e-02 -1.47415802e-01 -7.00820208e-01 8.47170174e-01
1.57500684e+00 -2.05488801e-01 -1.18472269e-02 7.65763104e-01
7.65137970e-01 -4.48317051e-01 -3.38358819e-01 6.20204568e-01
3.79828036e-01 -1.44976950e+00 -3.45493019e-01 3.46469373e-01
6.97900832e-01 2.74076313e-01 -6.87894002e-02 3.60647917e-01
9.85915437e-02 -1.01874948e+00 8.91636729e-01 5.74376106e-01
3.06873083e-01 -8.08259070e-01 1.40729606e+00 8.82908702e-01
-2.76707292e-01 -1.46398127e-01 -6.70168400e-02 -7.54587591e-01
2.82239318e-01 9.62928474e-01 -1.49593139e+00 2.13375255e-01
4.52024817e-01 5.12562037e-01 -4.18497711e-01 9.39888895e-01
-2.60223508e-01 9.99047339e-01 -2.97929734e-01 -5.27783751e-01
5.47813594e-01 7.93955177e-02 7.03996480e-01 1.66788578e+00
1.74698219e-01 -3.07275504e-01 -8.45809504e-02 4.71295238e-01
4.25656110e-01 3.59843671e-01 -5.97218573e-01 1.29766643e-01
3.11928719e-01 9.83520746e-01 -5.95251858e-01 -5.64759493e-01
-1.02480642e-01 5.97034395e-01 -5.51876687e-02 -1.31844386e-01
-7.98976302e-01 -1.37768805e-01 9.83876586e-02 6.93102837e-01
5.48439212e-02 -8.60799029e-02 -6.16368532e-01 -6.08539641e-01
-2.14312464e-01 -9.59141970e-01 1.32817745e-01 -4.86712843e-01
-1.04368770e+00 7.81105697e-01 -3.22216630e-01 -1.17055833e+00
-7.09751129e-01 -3.23419482e-01 -6.72247231e-01 8.43976617e-01
-1.24740016e+00 -1.14095914e+00 -2.02258959e-01 2.58779917e-02
2.42506281e-01 -2.71332651e-01 6.37611151e-01 8.99370834e-02
-4.48971480e-01 1.08762763e-01 -4.45446856e-02 1.59386501e-01
5.34594357e-01 -1.13362479e+00 1.03892572e-01 3.82847458e-01
-3.21882784e-01 2.09219605e-01 1.48342288e+00 -7.08177507e-01
-3.90810937e-01 -7.76035488e-01 1.77073574e+00 -4.06179219e-01
1.14149797e+00 1.59701392e-01 -1.16120934e+00 4.75608528e-01
3.45250636e-01 -7.36059308e-01 8.37071419e-01 1.97650075e-01
-1.96030334e-01 4.65152591e-01 -1.14402366e+00 1.26792997e-01
1.46427020e-01 -4.62693065e-01 -7.96170950e-01 5.68735063e-01
2.26355255e-01 -1.75823897e-01 -1.10714996e+00 -1.60464644e-01
7.03920186e-01 -1.70735085e+00 4.23526198e-01 -2.13336185e-01
7.20767379e-01 1.78276459e-04 5.93268909e-02 -1.00280118e+00
-2.09181353e-01 -5.20404220e-01 -1.31312013e-01 1.08374321e+00
6.46993279e-01 -9.72353399e-01 8.86355698e-01 1.03635624e-01
2.28964105e-01 -6.64970934e-01 -8.20577562e-01 -7.49082088e-01
2.70714372e-01 -2.86231935e-01 2.41107047e-01 1.27674913e+00
4.33142573e-01 5.30791938e-01 -4.97517914e-01 -2.50173539e-01
3.52583945e-01 -1.17912613e-01 7.90566921e-01 -1.61502683e+00
5.47978543e-02 -7.74477422e-01 -6.17487609e-01 -3.86166126e-01
1.82177201e-01 -7.19189346e-01 -3.17065299e-01 -1.62538815e+00
-7.37003833e-02 -2.95140415e-01 2.55764991e-01 2.53742188e-01
4.43582952e-01 4.97303814e-01 3.04541439e-02 6.06710613e-01
-2.34126866e-01 -9.91934016e-02 6.99960530e-01 1.41494140e-01
-3.60612601e-01 3.46387327e-01 -2.06946775e-01 1.03580225e+00
1.34947264e+00 -6.23586893e-01 3.53851914e-01 3.62580031e-01
7.57540822e-01 2.64003783e-01 3.11316729e-01 -9.49759126e-01
2.73825496e-01 5.60301952e-02 2.77730655e-02 -6.90782428e-01
3.35222244e-01 -2.56218165e-01 2.29603007e-01 7.51402795e-01
-1.46603256e-01 3.04114446e-02 -1.56690598e-01 3.06071699e-01
-4.42736834e-01 -5.50926626e-01 8.72984052e-01 -2.72386491e-01
-4.01003271e-01 -4.35930163e-01 -8.92432451e-01 -1.37765378e-01
8.11257005e-01 -6.52180195e-01 -4.37095851e-01 -8.46561909e-01
-9.36829329e-01 -1.95932314e-01 6.46948814e-01 2.29244962e-01
1.56186044e-01 -8.03308547e-01 -1.16158283e+00 -9.45108011e-02
-5.95478378e-02 -9.44522738e-01 -6.34618849e-02 1.33952045e+00
-8.26865613e-01 4.98813421e-01 -3.11789304e-01 -5.94620526e-01
-1.40740716e+00 3.17109495e-01 3.48628819e-01 -3.47815961e-01
-4.30660903e-01 4.06501353e-01 -8.31882894e-01 -2.60054588e-01
-2.81700969e-01 1.96579963e-01 -8.77164066e-01 9.46353734e-01
5.59790194e-01 7.80920565e-01 5.07981241e-01 -1.25361240e+00
-1.61890715e-01 8.77945796e-02 1.38923094e-01 -3.14876854e-01
1.31423199e+00 -1.53497383e-01 -4.14015502e-01 1.07444847e+00
1.24509728e+00 2.67437398e-01 -3.85812521e-01 -8.30093026e-03
3.39566827e-01 -2.05854863e-01 1.36823893e-01 -7.15470254e-01
-4.44595635e-01 6.42297506e-01 -2.56779827e-02 1.32242310e+00
4.13868010e-01 -2.27324009e-01 6.67206466e-01 9.60203260e-02
1.99681386e-01 -1.30607367e+00 -9.90490764e-02 9.51732874e-01
8.75960290e-01 -1.15989399e+00 1.08560972e-01 -1.12857632e-01
-9.04981315e-01 1.28840911e+00 -2.50034869e-01 -1.38048977e-01
6.00596607e-01 6.65866211e-02 2.52367973e-01 -3.05922359e-01
-6.48146152e-01 1.62962172e-02 -2.01160342e-01 4.02612507e-01
7.18138754e-01 3.36546600e-01 -9.14424479e-01 3.22626293e-01
-9.24754739e-01 1.73957407e-01 1.14062762e+00 6.38525546e-01
-1.09205782e+00 -6.74223304e-01 -6.07302248e-01 5.10717452e-01
-9.64058816e-01 2.03046069e-01 -5.95028639e-01 8.38803053e-01
-1.15636820e-02 1.39265621e+00 -7.95430318e-02 -6.24373674e-01
7.15845823e-02 3.47985625e-01 -1.23035565e-01 -4.58231539e-01
-1.05986524e+00 -2.10594088e-01 8.31803620e-01 -2.48971924e-01
-4.96983320e-01 -7.88764715e-01 -9.26639795e-01 -1.01616585e+00
-3.57380718e-01 3.77141327e-01 8.90185654e-01 1.12691283e+00
-1.62671104e-01 2.19113693e-01 9.16571319e-01 -6.02311790e-01
-5.35330713e-01 -1.30020487e+00 -7.03441620e-01 1.10959545e-01
5.03938496e-01 -5.37003696e-01 -7.33352721e-01 -1.61454752e-01] | [8.24662971496582, 10.120420455932617] |
c57e27a5-78ad-43d7-9614-ff00cc308922 | leveraging-code-generation-to-improve-code | 2002.10198 | null | https://arxiv.org/abs/2002.10198v2 | https://arxiv.org/pdf/2002.10198v2.pdf | Leveraging Code Generation to Improve Code Retrieval and Summarization via Dual Learning | Code summarization generates brief natural language description given a source code snippet, while code retrieval fetches relevant source code given a natural language query. Since both tasks aim to model the association between natural language and programming language, recent studies have combined these two tasks to improve their performance. However, researchers have yet been able to effectively leverage the intrinsic connection between the two tasks as they train these tasks in a separate or pipeline manner, which means their performance can not be well balanced. In this paper, we propose a novel end-to-end model for the two tasks by introducing an additional code generation task. More specifically, we explicitly exploit the probabilistic correlation between code summarization and code generation with dual learning, and utilize the two encoders for code summarization and code generation to train the code retrieval task via multi-task learning. We have carried out extensive experiments on an existing dataset of SQL and Python, and results show that our model can significantly improve the results of the code retrieval task over the-state-of-art models, as well as achieve competitive performance in terms of BLEU score for the code summarization task. | ['Xiaoyin Wang', 'Tianxiang Hu', 'Shikun Zhang', 'Wei Ye', 'Rui Xie', 'Jinglei Zhang'] | 2020-02-24 | null | null | null | null | ['code-summarization'] | ['computer-code'] | [ 2.24975899e-01 -9.46651474e-02 -3.65313202e-01 -2.83555597e-01
-1.40988696e+00 -5.64445257e-01 5.20102143e-01 4.17991728e-01
-1.38152584e-01 9.07992572e-02 4.88013923e-01 -4.40768600e-01
4.03781652e-01 -4.77736503e-01 -6.91734433e-01 -7.86153376e-02
-7.86547177e-03 1.87345482e-02 2.52219409e-01 -1.99415293e-02
5.80770910e-01 -4.44946885e-01 -1.47106886e+00 6.78935647e-01
1.30685115e+00 3.19028199e-01 5.45632660e-01 8.35693300e-01
-6.03308141e-01 1.22326183e+00 -6.09899879e-01 -4.54062849e-01
-1.75979480e-01 -4.01426971e-01 -1.00124896e+00 -3.29077303e-01
3.10479224e-01 -2.09661573e-01 -2.00600103e-01 1.07783055e+00
5.60277820e-01 -2.23114550e-01 4.41248208e-01 -9.75833833e-01
-9.06548500e-01 9.22567189e-01 -7.80934274e-01 -1.17131531e-01
7.39838302e-01 5.86030632e-02 1.31569731e+00 -8.61576319e-01
3.99129421e-01 1.04172218e+00 6.06494725e-01 6.12237096e-01
-1.36776960e+00 -5.72369576e-01 -4.26238775e-02 -1.55527979e-01
-1.04850113e+00 -4.57607090e-01 7.37441361e-01 -6.54731274e-01
1.44139981e+00 -6.01957403e-02 2.44836971e-01 7.97210515e-01
4.33772266e-01 1.27133811e+00 3.16235900e-01 -3.93292218e-01
-6.88851625e-02 1.44036442e-01 3.90051037e-01 9.79793608e-01
2.03684583e-01 -3.31787914e-01 -1.82866111e-01 -4.27401274e-01
5.02890488e-03 2.20611900e-01 -1.41146183e-01 -5.42413473e-01
-1.22309411e+00 9.16590393e-01 3.80751163e-01 2.54181296e-01
-7.30798393e-02 4.09049809e-01 9.63978708e-01 3.05858761e-01
4.21647489e-01 6.00696385e-01 -2.96142846e-01 -2.74026424e-01
-1.17455244e+00 5.12824595e-01 1.07446778e+00 1.30440402e+00
8.69948268e-01 -2.30447561e-01 -5.49665451e-01 1.04016614e+00
5.48709035e-01 4.56689954e-01 7.19788373e-01 -5.75383067e-01
9.06805038e-01 1.00244844e+00 -2.22793907e-01 -8.75617385e-01
-8.84681419e-02 -2.87354857e-01 -4.88514662e-01 -1.06428243e-01
-1.47668064e-01 -8.11169222e-02 -6.33450866e-01 1.61969626e+00
-3.32115382e-01 -3.58912438e-01 3.14665586e-01 3.38763475e-01
1.08229649e+00 8.27249944e-01 1.19083874e-01 1.72727197e-01
1.31082821e+00 -1.55220616e+00 -4.87533212e-01 -6.46309018e-01
9.76334214e-01 -9.70017612e-01 1.36002183e+00 3.11078317e-02
-1.19894528e+00 -6.42101884e-01 -1.12801075e+00 -4.17699695e-01
1.68288574e-02 5.86765885e-01 7.55388260e-01 2.71300405e-01
-1.01069009e+00 1.70848995e-01 -1.06823921e+00 -3.53692323e-01
3.95205766e-01 -1.45629421e-01 3.32029387e-02 -1.27558917e-01
-7.41166711e-01 6.18772686e-01 4.55626965e-01 -3.94883811e-01
-9.09884393e-01 -7.15331614e-01 -1.07386482e+00 4.56082076e-01
3.05423915e-01 -8.06859016e-01 1.70005286e+00 -7.70837784e-01
-1.13870323e+00 8.53847921e-01 -3.57754678e-01 -3.95236313e-01
1.84842139e-01 -4.10292834e-01 -5.79460524e-02 -1.58763602e-01
5.02038360e-01 4.85284477e-01 3.74304503e-01 -1.16729808e+00
-5.44235885e-01 1.48552861e-02 2.60757208e-01 -1.06581226e-01
-3.50034684e-01 1.93121284e-01 -6.45421684e-01 -7.10699379e-01
-3.17709118e-01 -7.71940887e-01 -2.05244154e-01 -3.67086649e-01
-3.63797218e-01 -4.87182826e-01 5.90219975e-01 -7.99414396e-01
1.70644557e+00 -2.37918186e+00 2.14242324e-01 -3.96100879e-01
2.63637424e-01 3.42596054e-01 -4.93769228e-01 6.64117992e-01
-2.44736951e-02 2.88568556e-01 -5.81800818e-01 -4.71333474e-01
1.00439608e-01 -1.41551301e-01 -7.30985224e-01 -3.49377445e-03
3.29499096e-01 1.10586202e+00 -1.09527659e+00 -5.35539508e-01
-3.21147203e-01 1.22672804e-01 -1.14814246e+00 6.39259160e-01
-5.52496314e-01 -1.98772848e-01 -7.43570209e-01 4.65285778e-01
3.78224164e-01 -4.09083098e-01 -5.23022488e-02 2.85191685e-01
-5.10106571e-02 5.92136145e-01 -6.65551484e-01 2.30538893e+00
-9.88641262e-01 6.60354018e-01 -2.61904120e-01 -8.68093193e-01
8.22128832e-01 3.13786477e-01 2.11197227e-01 -6.82675302e-01
-3.90856892e-01 3.90773594e-01 -1.86983183e-01 -9.44564223e-01
6.01882994e-01 1.79239169e-01 -6.55446947e-01 8.21035922e-01
1.11062042e-01 -4.21436101e-01 4.46439922e-01 5.75418890e-01
1.43137586e+00 3.15466732e-01 4.18951452e-01 -1.52355343e-01
6.76875830e-01 1.14959121e-01 2.30520964e-01 9.67867136e-01
1.98793739e-01 5.20139515e-01 8.89750540e-01 -2.68271178e-01
-8.58105242e-01 -9.43717360e-01 3.22695464e-01 1.46571946e+00
2.00410653e-02 -8.19645107e-01 -7.83787370e-01 -9.45852101e-01
-1.52284086e-01 8.54415655e-01 -2.89474219e-01 -3.59326571e-01
-6.66745365e-01 -5.57738364e-01 7.79635310e-01 5.04856765e-01
4.67994004e-01 -9.98909652e-01 -7.11233735e-01 2.92492062e-01
-3.84629220e-01 -8.70126426e-01 -8.01777840e-01 9.15550888e-02
-5.99585891e-01 -1.04735970e+00 -5.67847550e-01 -1.05629587e+00
4.89344299e-01 4.60479409e-01 1.50536215e+00 4.29993242e-01
-1.77710131e-01 3.75239730e-01 -6.10547841e-01 -3.56045425e-01
-9.27092016e-01 6.98699176e-01 -7.28101850e-01 -5.75614691e-01
4.70939606e-01 -4.04461265e-01 -2.66619176e-01 -2.50501513e-01
-1.30531144e+00 2.67107666e-01 9.67999697e-01 8.38700235e-01
-5.71321957e-02 -3.49837989e-01 6.91370547e-01 -9.70421553e-01
8.73941422e-01 -6.03672028e-01 -5.47163546e-01 6.24857545e-01
-4.81314301e-01 6.66010678e-01 5.86154342e-01 -3.00228335e-02
-1.17333722e+00 1.98995292e-01 -2.80694246e-01 1.91751018e-01
1.77665547e-01 1.14052546e+00 2.36987006e-02 2.35887855e-01
6.59120977e-01 4.78729725e-01 -1.15721405e-01 -4.92736012e-01
3.66824955e-01 1.02116978e+00 4.64075327e-01 -7.12093115e-01
7.83475518e-01 1.16317131e-01 -6.34698153e-01 -4.10330504e-01
-8.69651735e-01 -6.44700468e-01 -4.24634367e-01 3.48008126e-01
8.47571731e-01 -1.19337428e+00 -1.48640126e-01 3.68133157e-01
-1.65053689e+00 -1.56932890e-01 8.54046941e-02 1.30900085e-01
-5.57927489e-01 6.48259401e-01 -4.85274762e-01 -4.05456722e-01
-5.90582192e-01 -1.51027644e+00 1.40916562e+00 4.93225828e-02
-2.79346019e-01 -8.87661755e-01 4.32970375e-01 2.08571270e-01
7.11960435e-01 4.61533293e-02 1.36129320e+00 -8.50807607e-01
-7.64403641e-01 -3.25616628e-01 -3.44275594e-01 2.53067642e-01
1.48084715e-01 1.97768494e-01 -7.99633205e-01 -4.37638193e-01
3.44899856e-02 -6.68065012e-01 1.27081215e+00 -9.42641273e-02
1.22352028e+00 -4.20551896e-01 -3.60487074e-01 4.82470781e-01
1.54314840e+00 3.06117386e-02 4.99899238e-01 1.44544274e-01
7.29967833e-01 5.13530970e-01 2.22466886e-01 4.09198552e-01
7.79521286e-01 5.79608977e-01 3.50452811e-01 1.36916116e-01
-1.54496670e-01 -3.63885134e-01 6.62203610e-01 1.13295484e+00
6.37790442e-01 -1.63849909e-02 -1.25347257e+00 7.16624856e-01
-1.94844759e+00 -9.74547803e-01 -2.80056950e-02 1.99949551e+00
1.20438218e+00 -8.93860832e-02 -5.80596179e-02 -4.32546735e-01
4.48872924e-01 4.01803792e-01 -4.46588367e-01 -4.57428515e-01
4.13257927e-01 -2.83901971e-02 1.17073059e-01 3.69896293e-01
-1.06644273e+00 8.15709293e-01 6.02854967e+00 7.08940208e-01
-1.06176949e+00 5.04144840e-02 1.74988076e-01 -1.26791432e-01
-6.55873001e-01 2.86792815e-01 -6.27048254e-01 4.24376488e-01
9.69209075e-01 -7.71795154e-01 4.25815910e-01 1.19088197e+00
-2.40889058e-01 -3.66233401e-02 -1.54419494e+00 7.84328699e-01
4.10619825e-01 -1.07417524e+00 2.27061674e-01 -3.30633640e-01
8.60745370e-01 2.78311551e-01 -1.85370222e-01 9.69386458e-01
3.99498105e-01 -8.38106275e-01 7.09298253e-01 3.18256289e-01
5.05192935e-01 -4.18240517e-01 6.85156167e-01 4.76880014e-01
-1.28356683e+00 -1.46165550e-01 -2.08560929e-01 1.15284882e-01
-5.90304583e-02 5.75227857e-01 -6.81214988e-01 6.17434680e-01
3.29590499e-01 8.83937180e-01 -1.12011135e+00 1.18032396e+00
-2.79926270e-01 3.09365362e-01 2.62856126e-01 -1.88800931e-01
2.65390635e-01 3.72422934e-01 4.23957139e-01 1.62272918e+00
3.83506745e-01 -5.51696658e-01 4.83914047e-01 1.41166294e+00
-3.25416684e-01 9.29326937e-02 -6.55639410e-01 -4.52675700e-01
1.52384356e-01 1.14113259e+00 -3.42009187e-01 -4.73306090e-01
-9.42112803e-01 8.30837429e-01 3.44682842e-01 3.14325869e-01
-9.31967735e-01 -1.10915601e+00 4.26570445e-01 -2.93448597e-01
2.80866742e-01 -1.53255016e-01 -2.66506702e-01 -1.53746915e+00
4.82072711e-01 -1.00380027e+00 1.81856915e-01 -5.62072635e-01
-9.41914678e-01 5.07386804e-01 1.90673824e-02 -1.12598872e+00
-5.43052435e-01 -1.67542398e-01 -7.64276385e-01 8.91610742e-01
-1.73815250e+00 -1.09863949e+00 -3.05030078e-01 1.92530900e-02
8.84243071e-01 -2.88708091e-01 7.43754387e-01 4.15644020e-01
-4.75588828e-01 5.92172146e-01 2.02580497e-01 3.41351986e-01
7.86245048e-01 -1.37634277e+00 8.25153053e-01 1.14246964e+00
2.89608892e-02 1.01350594e+00 4.55453724e-01 -5.21498442e-01
-1.50370038e+00 -1.22592390e+00 1.03618777e+00 -6.40170574e-01
5.74541628e-01 -5.77198267e-01 -1.07657576e+00 5.64933419e-01
4.13610011e-01 -2.54606217e-01 7.12178767e-01 7.50003308e-02
-6.85320020e-01 1.10681415e-01 -5.59063494e-01 5.17525136e-01
5.78520656e-01 -8.95274878e-01 -1.00208795e+00 3.34464848e-01
1.06571603e+00 -3.76973182e-01 -5.24188101e-01 2.67120421e-01
4.51823235e-01 -9.75278258e-01 7.41545975e-01 -4.47815210e-01
1.27499247e+00 -2.23387808e-01 -7.33690485e-02 -1.25454462e+00
-6.02936223e-02 -4.53641742e-01 -1.28627747e-01 1.58664322e+00
4.25619721e-01 -2.00108796e-01 3.80539298e-01 3.96773756e-01
-2.99000293e-01 -6.26961529e-01 -3.83025467e-01 -6.02884233e-01
2.02461481e-01 -2.08251163e-01 5.71530998e-01 5.53009689e-01
4.59046423e-01 6.16781056e-01 1.05439359e-02 -1.33965626e-01
2.82934487e-01 6.33116126e-01 1.01970351e+00 -1.02525139e+00
-5.26676416e-01 -7.74570644e-01 7.14626908e-02 -1.37806618e+00
6.62084043e-01 -1.33952606e+00 4.38739240e-01 -1.81767845e+00
9.91557717e-01 -2.90314034e-02 7.99527913e-02 6.19110644e-01
-6.76671267e-01 -4.24658597e-01 1.39895454e-01 4.46033508e-01
-9.56661701e-01 5.30552864e-01 6.65769994e-01 -4.55039948e-01
-2.67760843e-01 6.11224733e-02 -1.13512301e+00 5.11347413e-01
5.11942089e-01 -6.00957096e-01 -5.47109365e-01 -1.08167338e+00
4.59838986e-01 2.99129397e-01 1.50247797e-01 -9.74576652e-01
4.16311860e-01 2.06603304e-01 -3.56263280e-01 -4.24559742e-01
-4.72985566e-01 -4.97784942e-01 -4.54706728e-01 5.88030875e-01
-7.95421839e-01 2.67677844e-01 4.28328335e-01 5.45156777e-01
-4.68094349e-01 -6.84038043e-01 6.55759454e-01 -2.74817199e-01
-6.20327592e-01 1.52676292e-02 -3.59988004e-01 4.24067140e-01
7.79663742e-01 1.04872286e-01 -4.72130001e-01 -8.95528197e-02
6.44045472e-02 4.45616603e-01 5.80439687e-01 7.54693866e-01
5.23074508e-01 -1.11356807e+00 -9.15457249e-01 3.07330787e-01
6.64004505e-01 4.51214463e-02 4.47921343e-02 5.13547719e-01
-5.88202775e-01 6.60899758e-01 5.23871696e-03 -5.44954896e-01
-1.19202971e+00 6.73840344e-01 1.12890102e-01 -6.21177197e-01
-2.48316273e-01 6.35420620e-01 2.87130624e-01 -7.26562858e-01
3.21746081e-01 -4.74343449e-01 -1.78388864e-01 -9.90715176e-02
6.55977488e-01 9.89037529e-02 8.65894333e-02 -1.89033270e-01
-2.67981231e-01 4.01826560e-01 -5.10361552e-01 2.44789213e-01
1.50314605e+00 5.54423295e-02 -5.42722762e-01 3.20580035e-01
1.71456015e+00 8.08648989e-02 -9.84388709e-01 -4.33641821e-01
5.42742848e-01 -3.98355454e-01 -2.32899800e-01 -7.36916363e-01
-8.20122063e-01 1.05069196e+00 1.63235188e-01 2.86057711e-01
1.03028905e+00 1.85013711e-01 8.78950477e-01 7.55402386e-01
2.35079959e-01 -6.86994135e-01 4.68014389e-01 9.07740891e-01
8.13954711e-01 -1.38026392e+00 -6.63920790e-02 -9.82982367e-02
-5.29417634e-01 1.15249062e+00 7.07765520e-01 -1.10946842e-01
1.73437133e-01 2.41567284e-01 -1.64405450e-01 -2.50235081e-01
-1.08678329e+00 3.41324285e-02 2.81753212e-01 2.07809016e-01
1.15012658e+00 -3.46972793e-01 -2.68749177e-01 5.50973177e-01
4.71401475e-02 6.17485307e-02 7.40189135e-01 1.12884367e+00
-4.84406978e-01 -1.21283889e+00 9.83640626e-02 5.18768072e-01
-6.19693816e-01 -4.00416732e-01 -2.84159511e-01 3.57194960e-01
-4.85635310e-01 8.22417557e-01 -1.62899703e-01 -3.04011941e-01
4.70152289e-01 1.68659434e-01 1.34849161e-01 -1.32538450e+00
-8.18080544e-01 -2.41755694e-01 -1.93217650e-01 -5.47067761e-01
-3.43895793e-01 -4.87290561e-01 -1.21113610e+00 1.29817566e-03
-2.30736315e-01 4.18392867e-01 4.82434034e-01 7.58229971e-01
6.13475680e-01 7.56167769e-01 5.55315733e-01 -6.82069480e-01
-8.20031822e-01 -8.61405969e-01 -5.88855781e-02 3.14793497e-01
5.57056904e-01 -9.48296934e-02 -2.53085941e-01 2.80152023e-01] | [7.626534938812256, 7.94273042678833] |
bebbf2d5-8e51-4910-80e9-c11cc45fea54 | bodiffusion-diffusing-sparse-observations-for | 2304.11118 | null | https://arxiv.org/abs/2304.11118v1 | https://arxiv.org/pdf/2304.11118v1.pdf | BoDiffusion: Diffusing Sparse Observations for Full-Body Human Motion Synthesis | Mixed reality applications require tracking the user's full-body motion to enable an immersive experience. However, typical head-mounted devices can only track head and hand movements, leading to a limited reconstruction of full-body motion due to variability in lower body configurations. We propose BoDiffusion -- a generative diffusion model for motion synthesis to tackle this under-constrained reconstruction problem. We present a time and space conditioning scheme that allows BoDiffusion to leverage sparse tracking inputs while generating smooth and realistic full-body motion sequences. To the best of our knowledge, this is the first approach that uses the reverse diffusion process to model full-body tracking as a conditional sequence generation task. We conduct experiments on the large-scale motion-capture dataset AMASS and show that our approach outperforms the state-of-the-art approaches by a significant margin in terms of full-body motion realism and joint reconstruction error. | ['Artsiom Sanakoyeu', 'Ali Thabet', 'Pablo Arbeláez', 'Albert Pumarola', 'Guillaume Jeanneret', 'Maria Escobar', 'Angela Castillo'] | 2023-04-21 | null | null | null | null | ['mixed-reality'] | ['computer-vision'] | [-1.31367013e-01 -1.63758837e-03 -2.16590628e-01 -4.13596779e-02
-8.35304022e-01 -3.79679054e-01 6.14120364e-01 -8.86367321e-01
-1.29987046e-01 7.91489780e-01 5.75721502e-01 -5.70647009e-02
2.87389427e-01 -3.70375663e-01 -7.28570998e-01 -4.31987762e-01
-6.05451781e-03 3.69611233e-01 1.30922750e-01 -1.05539158e-01
-5.38644671e-01 2.16108903e-01 -1.31778467e+00 7.03520083e-04
5.80804884e-01 4.70935076e-01 2.75399864e-01 1.14409816e+00
5.60008407e-01 8.11560273e-01 -1.69655010e-01 -2.27909654e-01
2.16051728e-01 -6.85568333e-01 -2.43427426e-01 1.36993006e-01
5.74678063e-01 -6.43277764e-01 -8.39365840e-01 6.16834283e-01
1.03040385e+00 3.80249202e-01 4.44848686e-01 -1.18607891e+00
-3.94895107e-01 4.59471457e-02 -4.62326705e-01 8.23768526e-02
1.02063954e+00 5.21580219e-01 5.77642500e-01 -8.03523004e-01
1.10443807e+00 1.08767629e+00 8.75434458e-01 1.06434810e+00
-1.22529173e+00 -7.31726289e-01 1.64697081e-01 -2.94896364e-01
-1.41898811e+00 -7.14478135e-01 5.71798146e-01 -6.74251735e-01
7.28140473e-01 3.79939258e-01 1.09353888e+00 1.57036769e+00
3.07230979e-01 9.64708686e-01 7.33214438e-01 -1.07619138e-02
2.46080726e-01 -3.82842094e-01 -2.54566312e-01 7.62578964e-01
2.40460962e-01 2.78607398e-01 -7.18433738e-01 -1.00880988e-01
1.10795844e+00 -1.32014394e-01 -6.45900548e-01 -6.03962302e-01
-1.39252841e+00 5.39505363e-01 1.34235442e-01 1.41415134e-01
-5.86114466e-01 6.55793905e-01 -1.07127102e-02 -2.56432295e-01
2.14848951e-01 -1.26086086e-01 -1.31016955e-01 -5.69176078e-01
-1.34776020e+00 8.27765942e-01 8.48128378e-01 1.11648870e+00
1.63509902e-02 4.89436358e-01 -3.24713379e-01 2.77471006e-01
5.61332881e-01 8.10015798e-01 4.70071167e-01 -1.17855644e+00
4.64042902e-01 -1.72273397e-01 4.10293072e-01 -7.07813919e-01
-4.90532994e-01 -3.41592342e-01 -6.94817841e-01 3.62586409e-01
4.67635691e-01 -6.01660550e-01 -1.05258489e+00 2.10166025e+00
7.11091936e-01 7.79293060e-01 -2.06363052e-01 1.40754604e+00
1.05723357e+00 4.82384145e-01 2.05858171e-01 -1.43129438e-01
1.08382607e+00 -7.78961658e-01 -1.08568323e+00 -4.66104716e-01
4.51773852e-01 -5.27012050e-01 8.79866242e-01 3.56478304e-01
-1.42717183e+00 -5.39029598e-01 -9.57422256e-01 8.94977823e-02
5.71874857e-01 -3.95358838e-02 6.15787983e-01 8.95510793e-01
-8.72091532e-01 5.80491662e-01 -1.32917750e+00 -2.78265208e-01
1.63640410e-01 1.65809602e-01 -4.91843313e-01 -1.62964076e-01
-9.59775686e-01 8.01189423e-01 -1.65539339e-01 -1.25690522e-02
-9.58059907e-01 -9.61910009e-01 -1.16248906e+00 -3.95049214e-01
2.11230025e-01 -1.57461214e+00 1.52762866e+00 -3.26546967e-01
-1.81473804e+00 6.74210012e-01 -3.08437139e-01 -3.26006323e-01
1.05566680e+00 -6.14629686e-01 -3.54075313e-01 4.80519943e-02
-1.43075615e-01 7.23345041e-01 8.05669188e-01 -1.14428401e+00
-1.15361772e-01 -2.07868889e-01 -3.68215889e-01 6.21480346e-01
3.17290545e-01 -2.76475489e-01 -9.02697682e-01 -9.26563680e-01
1.17990114e-01 -1.51325226e+00 -3.26443046e-01 2.71359205e-01
-3.12501401e-01 7.03613460e-01 7.57340133e-01 -9.07509089e-01
1.30348003e+00 -1.86464763e+00 4.80156720e-01 -1.53598100e-01
1.59188971e-01 1.23190442e-02 2.85131216e-01 3.40671912e-02
6.58894032e-02 -2.75449991e-01 -1.27540424e-01 -1.04155433e+00
7.73388073e-02 3.28205943e-01 -2.47987360e-01 7.39800930e-01
-4.15809304e-01 1.24780500e+00 -1.01590466e+00 -5.53657174e-01
4.81575549e-01 9.46641088e-01 -8.02280247e-01 1.23519257e-01
-3.90794352e-02 1.23222077e+00 -2.78338999e-01 5.27846038e-01
5.33664584e-01 -5.73479608e-02 1.21407852e-01 -4.60264832e-02
1.22812994e-01 4.52752486e-02 -1.39984798e+00 2.39754081e+00
-3.96805793e-01 5.87822437e-01 2.60677367e-01 -8.32293089e-03
3.20281982e-01 5.83031356e-01 7.72226334e-01 -4.51072365e-01
1.58770829e-01 1.03416517e-01 -8.24038237e-02 -4.51983750e-01
6.91635549e-01 -6.41506970e-01 -1.53288782e-01 2.61432618e-01
-1.20573923e-01 -2.81299770e-01 -3.35021585e-01 2.07107425e-01
1.01706564e+00 7.18353689e-01 9.98198390e-02 8.82794857e-02
-5.61949275e-02 -2.77518481e-01 5.72770476e-01 7.05429733e-01
-2.95864016e-01 1.25568855e+00 -2.41864577e-01 2.75239889e-02
-8.26507747e-01 -1.27061367e+00 2.61984825e-01 6.59640491e-01
7.26954192e-02 -3.70008171e-01 -7.07924545e-01 -2.05155835e-01
5.10665178e-02 7.69145906e-01 -5.91285169e-01 -9.04401392e-03
-8.34963679e-01 -6.44179761e-01 5.31484663e-01 7.46617854e-01
1.79534748e-01 -9.63495016e-01 -8.57281148e-01 2.56294250e-01
-3.47265542e-01 -1.28111517e+00 -8.39046955e-01 -5.21707058e-01
-9.39070344e-01 -6.05239093e-01 -1.36840165e+00 -2.16090843e-01
2.77595758e-01 1.96297988e-02 1.04693031e+00 -2.90434211e-01
-3.76970619e-01 6.01172268e-01 7.74915447e-04 9.57304388e-02
-2.23795786e-01 -4.20736343e-01 2.05760583e-01 -1.40176579e-01
-4.94911760e-01 -5.74433148e-01 -9.49635804e-01 2.55188137e-01
-5.74617445e-01 2.35795707e-01 1.06223762e-01 5.91518462e-01
5.92532754e-01 -4.66657460e-01 -1.43815745e-02 -4.92538363e-01
3.66355985e-01 -5.48751712e-01 -1.75919041e-01 -2.58712351e-01
-8.62692147e-02 -1.71978503e-01 3.42667587e-02 -9.36221719e-01
-1.19523692e+00 3.58779073e-01 -3.12853485e-01 -8.49666774e-01
6.05089637e-03 2.51810431e-01 -3.65713030e-01 2.78870184e-02
5.29624164e-01 1.73590273e-01 8.35406687e-03 -3.40781957e-01
6.90813601e-01 9.16005224e-02 9.93263841e-01 -3.80119532e-01
5.30789733e-01 8.17628622e-01 2.03824744e-01 -7.86261201e-01
-3.66238385e-01 -3.75091314e-01 -4.70520586e-01 -5.14534652e-01
1.09504628e+00 -1.28622341e+00 -7.23297596e-01 5.62420726e-01
-8.63276601e-01 -8.70274425e-01 -3.32113266e-01 9.00195599e-01
-1.06020176e+00 4.96299505e-01 -7.51930714e-01 -9.30433691e-01
-2.38927200e-01 -1.11908472e+00 1.26936841e+00 1.57552287e-02
-8.01528096e-01 -9.20968592e-01 5.10595322e-01 2.02602386e-01
1.51733041e-01 6.90752745e-01 -2.35988870e-01 2.41165325e-01
-5.91175497e-01 -2.07434237e-01 4.27015007e-01 -2.90069669e-01
-1.01952814e-01 -2.22526297e-01 -8.21016073e-01 -4.46610898e-01
-4.81703505e-02 -1.91142429e-02 5.20575166e-01 1.05839503e+00
4.86866653e-01 -1.52335450e-01 -4.99978900e-01 9.81292069e-01
1.00054085e+00 -1.07749172e-01 8.66553426e-01 -2.26004794e-02
1.03298056e+00 1.77289769e-01 6.47215605e-01 7.19066978e-01
5.10138452e-01 1.10646892e+00 1.70582339e-01 2.27867216e-02
-6.81398392e-01 -6.70493186e-01 4.38881993e-01 5.74823976e-01
-1.16958141e-01 -4.38895613e-01 -7.42166460e-01 5.58433235e-01
-1.90573490e+00 -1.27365589e+00 -3.58901232e-01 2.19009328e+00
4.68030095e-01 -7.40056634e-02 4.01938826e-01 -1.07614182e-01
4.54063654e-01 2.35742718e-01 -6.06631994e-01 2.53572643e-01
7.48532340e-02 2.23430600e-02 3.63338292e-01 7.38161981e-01
-9.24301326e-01 7.50611365e-01 6.97612476e+00 4.33442652e-01
-1.02117670e+00 3.07800114e-01 5.94398454e-02 -6.30509079e-01
-1.89415246e-01 -1.30376816e-01 -8.43185008e-01 6.50397062e-01
7.69058287e-01 8.10981765e-02 1.95211828e-01 7.23853648e-01
4.33977574e-01 -9.45556238e-02 -9.24649417e-01 1.32665634e+00
1.72053680e-01 -9.68390167e-01 -4.66713160e-01 3.38388056e-01
8.13504934e-01 -2.74085756e-02 2.50825167e-01 1.84862390e-01
4.32332069e-01 -9.28645730e-01 1.15130961e+00 8.56932759e-01
1.02337015e+00 -4.21268284e-01 -1.63774099e-02 6.38993323e-01
-1.24585688e+00 2.73725599e-01 1.87787637e-01 -3.94210853e-02
1.27730191e+00 2.88286328e-01 -3.99099588e-01 3.77481639e-01
4.75947678e-01 5.76186299e-01 5.60866632e-02 1.12547755e+00
-1.52470902e-01 5.35748243e-01 -4.48610485e-01 4.35774237e-01
-1.40053079e-01 2.72858739e-02 9.86225605e-01 1.18800044e+00
8.09142888e-01 3.94701213e-01 1.16462357e-01 7.60816514e-01
2.95893550e-01 -3.01585346e-01 -7.15574026e-01 3.35360587e-01
2.03503117e-01 8.87489319e-01 -3.49171191e-01 -3.60489219e-01
-3.05489361e-01 1.39678669e+00 -1.48373514e-01 3.38266343e-01
-1.16565824e+00 2.77629316e-01 8.12673509e-01 5.40352404e-01
3.58361781e-01 -6.35952592e-01 -2.90991277e-01 -1.48461568e+00
-9.30239037e-02 -6.98579609e-01 3.42402071e-01 -1.16646707e+00
-7.92508841e-01 3.71448398e-01 2.05125928e-01 -1.47606885e+00
-7.93935776e-01 -1.12025835e-01 -3.69929522e-01 8.94837976e-01
-8.04683030e-01 -1.26381588e+00 -5.02765715e-01 7.07309783e-01
4.18840110e-01 2.47339249e-01 5.47464430e-01 4.84010696e-01
-3.79931211e-01 5.20190418e-01 -2.32563928e-01 -1.98855489e-01
6.08174682e-01 -9.16532457e-01 8.03082347e-01 9.63500500e-01
7.13985488e-02 6.64023042e-01 1.22539043e+00 -1.14199829e+00
-1.67903101e+00 -9.05601919e-01 4.39879954e-01 -8.31240773e-01
3.48312199e-01 -2.81657547e-01 -7.24184930e-01 9.88481045e-01
-4.01738808e-02 5.24198353e-01 5.57498276e-01 -4.02141988e-01
1.52796283e-01 5.75872302e-01 -1.15950644e+00 9.11715627e-01
1.50994265e+00 -3.32790703e-01 -5.08342624e-01 -1.74432501e-01
4.11359876e-01 -1.19498754e+00 -7.79509366e-01 3.54710281e-01
1.03868747e+00 -8.16800654e-01 1.33812320e+00 -3.77629101e-01
1.63379177e-01 -4.13613528e-01 -1.20995581e-01 -1.17517734e+00
-3.91333133e-01 -1.10063016e+00 -9.09735143e-01 8.09782803e-01
-4.99377958e-02 -1.94736049e-01 1.12471676e+00 9.81561601e-01
-2.24088854e-03 -3.20665896e-01 -9.47354913e-01 -7.95339406e-01
-6.32549077e-02 -8.35275710e-01 3.45441431e-01 7.75327623e-01
-9.13871974e-02 7.21699372e-03 -1.09489357e+00 1.09745815e-01
8.00906777e-01 -5.81471808e-02 1.21079504e+00 -6.88133180e-01
-8.42818379e-01 -1.45757556e-01 -3.37239385e-01 -1.59729040e+00
8.31116661e-02 -4.69480425e-01 3.07264835e-01 -1.50236511e+00
1.03898175e-01 -1.99295849e-01 3.47665906e-01 -1.17419347e-01
-2.92741209e-01 4.28387463e-01 4.72682744e-01 1.04828335e-01
-4.32892412e-01 7.03075886e-01 1.54639888e+00 3.16967756e-01
-4.38675255e-01 4.02231924e-02 -2.81978846e-01 7.90316880e-01
3.76884997e-01 -4.13392633e-01 -6.23981118e-01 -3.47981483e-01
-1.91595219e-02 8.44142258e-01 6.05055749e-01 -1.34483922e+00
1.95401296e-01 -1.57106578e-01 5.00935972e-01 -6.53273880e-01
8.88721108e-01 -6.06351495e-01 1.15127957e+00 6.09283745e-01
3.48844193e-02 1.77984759e-02 4.13081676e-01 7.08427668e-01
2.27605030e-01 2.38438815e-01 6.16605937e-01 -1.06593028e-01
-4.41701561e-01 5.60324073e-01 -5.69891572e-01 2.44699597e-01
9.05412018e-01 -4.51409549e-01 1.25068069e-01 -1.01956224e+00
-1.15717530e+00 -1.28915042e-01 7.99104095e-01 4.54082906e-01
6.91822350e-01 -1.61732781e+00 -6.06732011e-01 1.57881886e-01
-2.19402149e-01 -7.73345008e-02 7.42506087e-01 8.80885065e-01
-5.15353322e-01 2.80190051e-01 -2.72669457e-02 -7.00910807e-01
-1.42322469e+00 5.31912744e-01 4.06428248e-01 -2.35926956e-01
-1.19250524e+00 6.12940192e-01 3.00305396e-01 -2.77038187e-01
-1.18660204e-01 -1.96251005e-01 2.03995451e-01 -4.31699991e-01
4.98459846e-01 4.97827768e-01 -2.81042218e-01 -1.29310679e+00
-3.58185023e-01 7.87234008e-01 5.82111776e-01 -9.22243893e-01
8.23579371e-01 -3.89624208e-01 8.21390212e-01 2.77380168e-01
9.45405722e-01 2.63640493e-01 -1.72473240e+00 1.06773749e-01
-6.61805093e-01 -7.50012994e-01 1.13862827e-01 -6.52621806e-01
-9.93511021e-01 6.62631154e-01 5.20660639e-01 -6.44237697e-01
7.46812761e-01 -1.96251616e-01 1.11836147e+00 -2.38139540e-01
7.07126558e-01 -6.09034598e-01 5.30217066e-02 2.01546103e-01
9.08621669e-01 -8.72486949e-01 9.82929245e-02 -4.53502774e-01
-1.06942499e+00 4.05581176e-01 5.23819327e-01 -1.16586044e-01
6.93846762e-01 5.98107517e-01 -3.60656492e-02 -2.50529628e-02
-5.62610030e-01 -2.05678985e-01 6.54666662e-01 7.58876979e-01
4.61004257e-01 1.65027842e-01 -1.05614163e-01 5.84097266e-01
-3.87444884e-01 5.31875312e-01 2.34164193e-01 1.14207935e+00
3.02856117e-02 -8.24392557e-01 -7.13553250e-01 -9.73533168e-02
-5.25424600e-01 2.06237569e-01 1.22697674e-01 8.62427890e-01
-1.13479406e-01 7.52564788e-01 -1.16583012e-01 -3.01271528e-01
4.04372096e-01 -7.53778070e-02 1.03569448e+00 -4.05434221e-01
-2.89266974e-01 3.95903796e-01 1.38912424e-01 -9.42896545e-01
-2.15335533e-01 -9.22324896e-01 -1.30945718e+00 -4.83421743e-01
-8.86043981e-02 -4.17969406e-01 4.49314445e-01 4.82548267e-01
3.57599020e-01 5.85303187e-01 -1.60217822e-01 -1.51919723e+00
-2.07905889e-01 -8.12084854e-01 -5.80806792e-01 7.46807933e-01
5.61696172e-01 -7.84059107e-01 -1.19040407e-01 2.64226735e-01] | [7.192646026611328, -0.5321860313415527] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.