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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
a619186f-3d2f-4151-89c7-678e2fd771b2 | bidirectional-projection-network-for-cross | 2103.14326 | null | https://arxiv.org/abs/2103.14326v1 | https://arxiv.org/pdf/2103.14326v1.pdf | Bidirectional Projection Network for Cross Dimension Scene Understanding | 2D image representations are in regular grids and can be processed efficiently, whereas 3D point clouds are unordered and scattered in 3D space. The information inside these two visual domains is well complementary, e.g., 2D images have fine-grained texture while 3D point clouds contain plentiful geometry information. However, most current visual recognition systems process them individually. In this paper, we present a \emph{bidirectional projection network (BPNet)} for joint 2D and 3D reasoning in an end-to-end manner. It contains 2D and 3D sub-networks with symmetric architectures, that are connected by our proposed \emph{bidirectional projection module (BPM)}. Via the \emph{BPM}, complementary 2D and 3D information can interact with each other in multiple architectural levels, such that advantages in these two visual domains can be combined for better scene recognition. Extensive quantitative and qualitative experimental evaluations show that joint reasoning over 2D and 3D visual domains can benefit both 2D and 3D scene understanding simultaneously. Our \emph{BPNet} achieves top performance on the ScanNetV2 benchmark for both 2D and 3D semantic segmentation. Code is available at \url{https://github.com/wbhu/BPNet}. | ['Tien-Tsin Wong', 'Jiaya Jia', 'Li Jiang', 'Hengshuang Zhao', 'WenBo Hu'] | 2021-03-26 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Hu_Bidirectional_Projection_Network_for_Cross_Dimension_Scene_Understanding_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Hu_Bidirectional_Projection_Network_for_Cross_Dimension_Scene_Understanding_CVPR_2021_paper.pdf | cvpr-2021-1 | ['scene-recognition', '2d-semantic-segmentation'] | ['computer-vision', 'computer-vision'] | [-7.83015490e-02 2.26489395e-01 -1.55910298e-01 -3.95440131e-01
-1.99279994e-01 -7.22021401e-01 5.77621698e-01 -3.68342996e-02
-1.04404390e-02 -1.24283038e-01 -2.74242103e-01 -5.20073593e-01
-7.83746876e-03 -1.01113415e+00 -6.62484288e-01 -4.12383705e-01
1.90068007e-01 7.86071181e-01 7.32624888e-01 -3.04396391e-01
1.79393366e-01 1.03424335e+00 -1.73091638e+00 3.22873741e-01
6.73805118e-01 1.27158618e+00 4.61037070e-01 4.61147636e-01
-6.85054362e-01 5.21431148e-01 -1.02041908e-01 -1.98237777e-01
5.35678208e-01 2.00307786e-01 -7.23904788e-01 2.95796126e-01
6.75389946e-01 -2.92334378e-01 -4.09213632e-01 1.22993135e+00
3.36862892e-01 -1.41952202e-01 6.24850392e-01 -1.36826146e+00
-6.40358210e-01 3.30437943e-02 -8.91712785e-01 5.10826595e-02
3.01895767e-01 2.92665243e-01 8.68659616e-01 -1.17450988e+00
5.47173083e-01 1.42178154e+00 5.62917531e-01 2.58923620e-01
-1.07994902e+00 -6.99664474e-01 5.12059093e-01 -5.94125222e-03
-1.24504292e+00 2.05284685e-01 1.28115940e+00 -5.38291276e-01
1.12485313e+00 3.60137492e-01 9.75778580e-01 6.51354551e-01
-1.91575557e-01 1.07603455e+00 1.20394123e+00 1.44920079e-02
-1.11213969e-02 -5.75308576e-02 2.38607123e-01 7.36117899e-01
2.31420919e-01 1.59115300e-01 -5.60935855e-01 2.34550983e-01
1.23669577e+00 4.08189416e-01 -6.39220178e-02 -8.55988145e-01
-1.04616868e+00 5.56103110e-01 8.44721437e-01 1.13137849e-01
-2.73718864e-01 1.18107572e-01 1.65175542e-01 -2.34107766e-02
7.28281021e-01 1.75099168e-02 -2.82286644e-01 1.91167906e-01
-6.92202985e-01 9.08887833e-02 5.90158939e-01 1.19046497e+00
9.99215424e-01 5.23704924e-02 1.69687733e-01 1.10530663e+00
6.37655020e-01 9.57309783e-01 -2.15321869e-01 -1.01151395e+00
5.01816571e-01 1.12286973e+00 -1.86965287e-01 -1.28304374e+00
-6.49977684e-01 -3.44903290e-01 -1.03449690e+00 6.27375245e-01
1.85272843e-01 4.64841068e-01 -1.45699668e+00 1.18965650e+00
5.06067872e-01 1.19460188e-01 -2.42487416e-01 1.20326507e+00
1.50726986e+00 5.06462872e-01 -1.23299673e-01 5.41436970e-01
1.39598215e+00 -1.09879982e+00 -1.21840090e-01 -5.56676447e-01
1.51906177e-01 -8.41108084e-01 1.03835690e+00 1.67359248e-01
-1.34449196e+00 -6.87442243e-01 -9.67833459e-01 -4.04117972e-01
-4.73924458e-01 -1.32602602e-01 5.19203305e-01 3.54947716e-01
-1.14104307e+00 1.61269233e-01 -8.46716940e-01 -4.07768548e-01
7.92117655e-01 2.29124352e-01 -3.24185312e-01 -3.79840195e-01
-6.56423330e-01 6.07713878e-01 3.82072508e-01 1.97700545e-01
-6.80561125e-01 -7.46257961e-01 -9.97372985e-01 -1.40416086e-01
3.48183066e-01 -8.21127236e-01 1.05985951e+00 -3.18482101e-01
-1.02102196e+00 1.41438866e+00 8.37525204e-02 -3.24543417e-02
5.65075099e-01 -3.99546996e-02 -1.49587020e-02 3.67336959e-01
2.03048289e-01 9.63746607e-01 5.41752279e-01 -1.69073868e+00
-5.75967491e-01 -7.04346478e-01 2.25319713e-01 5.30844688e-01
9.23933089e-02 -4.91560400e-01 -1.21722281e+00 -5.42871058e-01
8.33823144e-01 -6.74992561e-01 -2.37044856e-01 3.54679734e-01
-6.90434217e-01 -2.70822942e-01 1.15789700e+00 -3.86755198e-01
6.43008232e-01 -2.16599965e+00 1.47014990e-01 4.26916659e-01
5.23701966e-01 3.30041856e-01 -8.68902653e-02 1.46535233e-01
3.71264927e-02 2.57373810e-01 -3.87459666e-01 -4.76694316e-01
1.46153703e-01 4.80250359e-01 -2.07987234e-01 3.46789181e-01
1.85720950e-01 1.04724896e+00 -7.75689960e-01 -4.66670066e-01
8.35824072e-01 4.91993695e-01 -4.18900937e-01 1.69968337e-01
-2.96096921e-01 4.23580229e-01 -7.17458010e-01 1.02740097e+00
1.22704256e+00 -5.35680294e-01 -8.66652951e-02 -4.44490641e-01
-1.64916337e-01 6.14398643e-02 -1.15454900e+00 1.81134021e+00
-2.75739223e-01 3.52208734e-01 2.74531692e-01 -9.00643885e-01
1.23391974e+00 -8.44109505e-02 4.88108039e-01 -1.10536575e+00
1.72563583e-01 2.37223148e-01 -5.44892490e-01 -1.74053282e-01
2.91237831e-01 6.34825230e-02 -2.09658295e-02 2.82280564e-01
-8.62845778e-02 -9.32821155e-01 6.91689700e-02 3.62713151e-02
5.43888927e-01 1.59919560e-01 6.63086846e-02 -1.81038305e-01
3.57304752e-01 4.11904126e-01 4.84454155e-01 7.58528471e-01
-4.02032286e-02 9.10558999e-01 3.99541527e-01 -6.29089236e-01
-1.09609854e+00 -1.39760089e+00 -1.74691930e-01 5.97276807e-01
9.64978695e-01 -1.56892568e-01 -2.31047720e-01 -6.18245244e-01
3.10271651e-01 5.13096511e-01 -4.11767721e-01 1.11749083e-01
-6.01224661e-01 -3.00422668e-01 4.25748706e-01 8.15875053e-01
9.69351947e-01 -8.28812420e-01 -7.12076545e-01 -2.06392348e-01
-3.24945487e-02 -1.23823690e+00 -1.79731190e-01 2.07265452e-01
-9.97526348e-01 -1.14956784e+00 -6.73596382e-01 -8.49776268e-01
6.47442341e-01 8.51086915e-01 1.46538758e+00 2.83223718e-01
-2.29005665e-01 6.13889098e-01 -2.82886833e-01 -5.21558046e-01
9.04728994e-02 -1.38156429e-01 -2.72784859e-01 -4.38423872e-01
4.59641188e-01 -9.26666617e-01 -6.57882035e-01 6.12672389e-01
-9.30566490e-01 6.77138031e-01 5.77008426e-01 4.97337580e-01
1.16676652e+00 -1.98064357e-01 -2.99381942e-01 -8.40398550e-01
1.05735853e-01 -3.39131624e-01 -6.24821126e-01 1.12440564e-01
-2.90520966e-01 -3.65277588e-01 1.56389028e-01 6.63143843e-02
-9.80978787e-01 8.73601660e-02 -3.36677849e-01 -9.34640050e-01
-6.72937870e-01 1.79740027e-01 -2.70001888e-01 -1.38891721e-02
2.16855228e-01 3.08832288e-01 9.02811438e-03 -8.21835041e-01
5.24127841e-01 3.36304843e-01 4.26763117e-01 -5.00555933e-01
9.09146070e-01 8.65752578e-01 1.33558080e-01 -8.57846677e-01
-1.00753391e+00 -7.27280557e-01 -8.93999279e-01 -3.80957186e-01
1.10575402e+00 -9.87003863e-01 -6.34922743e-01 6.64758921e-01
-1.11501825e+00 -4.52085018e-01 -3.66484344e-01 2.09297821e-01
-6.28032506e-01 3.48448217e-01 -3.97068948e-01 -5.85015714e-01
-1.60347730e-01 -1.27723598e+00 1.49833870e+00 2.69780904e-01
1.88232243e-01 -9.67843235e-01 -3.04235041e-01 5.99071920e-01
6.24905787e-02 3.22965771e-01 9.60965514e-01 -3.79330754e-01
-9.12413538e-01 -1.95440021e-03 -9.01335180e-01 4.93404061e-01
-1.13342270e-01 -1.47283986e-01 -9.42979991e-01 4.43267263e-02
-1.99515708e-02 -1.46009296e-01 8.95212293e-01 3.57539743e-01
1.16009116e+00 3.40526909e-01 -3.72399092e-01 9.28822100e-01
1.42798603e+00 -3.31541747e-02 3.50096613e-01 1.55833721e-01
1.29278493e+00 6.20203376e-01 5.41930914e-01 2.12971389e-01
6.61051035e-01 5.88752151e-01 9.60511982e-01 -5.75074315e-01
-4.98903543e-01 -2.12735415e-01 -1.95956439e-01 7.19077289e-01
-2.44553491e-01 -3.76613289e-02 -1.51597440e+00 3.75132024e-01
-1.88838470e+00 -5.44342518e-01 -4.86714959e-01 1.72626066e+00
2.45232090e-01 2.52242178e-01 2.29593515e-02 -4.98853587e-02
5.82763076e-01 3.46683681e-01 -8.82671535e-01 -1.81073427e-01
-4.12616223e-01 1.61703914e-01 6.70072556e-01 2.88183749e-01
-1.09241235e+00 8.72952163e-01 5.10362196e+00 9.33945060e-01
-1.03768110e+00 -2.10340768e-02 7.43707597e-01 -5.26008047e-02
-5.88266432e-01 -1.46524355e-01 -6.46608829e-01 1.31213799e-01
-5.09202071e-02 2.24952385e-01 -4.65322249e-02 9.80604947e-01
-9.29481834e-02 -6.20208085e-02 -1.01734221e+00 1.33738959e+00
-1.10006534e-01 -1.70366132e+00 3.37350428e-01 1.71752781e-01
6.54013574e-01 5.08794487e-01 8.24562386e-02 1.30514935e-01
3.74452561e-01 -9.71906960e-01 1.09813046e+00 3.96509826e-01
8.06913257e-01 -6.19054973e-01 5.17688930e-01 5.07371724e-01
-1.68112659e+00 1.41534969e-01 -2.50246197e-01 1.46519899e-01
3.39513242e-01 7.73874938e-01 -5.51336110e-01 8.08763981e-01
1.21573997e+00 1.05763900e+00 -5.41238427e-01 1.02153993e+00
-3.04246455e-01 2.81765550e-01 -6.10433996e-01 1.96390316e-01
5.19505978e-01 -4.03344840e-01 7.76038826e-01 1.07822907e+00
1.99907467e-01 2.30306938e-01 5.20478785e-01 1.18441164e+00
1.15553327e-01 -2.16669768e-01 -8.04416716e-01 2.41936415e-01
3.82693619e-01 1.14429903e+00 -1.27010739e+00 -3.34497780e-01
-4.74187106e-01 7.72434652e-01 2.93321580e-01 4.42831933e-01
-6.89660370e-01 -4.84168790e-02 8.44891787e-01 9.26910564e-02
5.55719733e-01 -5.99560499e-01 -8.53913546e-01 -1.00445807e+00
1.39448985e-01 -3.35718691e-01 3.96590412e-01 -1.05735552e+00
-1.56226861e+00 5.26978135e-01 1.45885482e-01 -1.38855600e+00
4.26461130e-01 -9.39123929e-01 -3.44449550e-01 9.69105065e-01
-1.72975552e+00 -1.28775156e+00 -6.47970438e-01 8.05931687e-01
5.51989734e-01 2.49815896e-01 3.63177299e-01 3.69010299e-01
-2.85667747e-01 6.96796551e-02 -3.23550582e-01 2.80826867e-01
-1.10097248e-02 -1.37011707e+00 6.18025959e-01 6.70256972e-01
2.78982133e-01 2.29974151e-01 7.35756755e-02 -6.68298662e-01
-1.44667864e+00 -1.23429644e+00 3.32291037e-01 -5.16178071e-01
2.35272422e-01 -4.97511446e-01 -9.25914884e-01 3.87716025e-01
-7.74827376e-02 1.09616980e-01 3.41540724e-01 -4.02031355e-02
-5.76227546e-01 2.51845550e-02 -1.09787238e+00 5.07455051e-01
1.57292354e+00 -6.62914217e-01 -5.24290264e-01 3.90230149e-01
9.08128977e-01 -8.14806402e-01 -7.57246315e-01 7.15017974e-01
1.71831951e-01 -1.35373962e+00 1.47917879e+00 -2.48846516e-01
3.83765370e-01 -6.19981229e-01 -3.97492141e-01 -9.32415843e-01
-1.80549055e-01 -4.99244872e-03 -7.32530802e-02 8.33884597e-01
2.29741052e-01 -4.96095121e-01 9.69000280e-01 4.08113569e-01
-4.33909416e-01 -9.29414630e-01 -8.58540416e-01 -8.40488136e-01
-2.95223445e-02 -1.10628319e+00 7.40434647e-01 7.20703721e-01
-7.02416360e-01 3.17679644e-01 6.38083294e-02 4.69214320e-01
7.90103257e-01 6.97123945e-01 7.53294110e-01 -1.40399039e+00
7.71548599e-02 -9.20480371e-01 -5.87290764e-01 -1.63906133e+00
-2.69448757e-01 -1.24986339e+00 -1.04268588e-01 -2.05484414e+00
4.16382253e-02 -7.68257260e-01 -1.51271045e-01 4.84183133e-01
2.26134315e-01 6.75740063e-01 5.59561789e-01 3.93591374e-01
-4.94871557e-01 5.60651302e-01 1.67265666e+00 -2.76741594e-01
-1.11120440e-01 -1.66247472e-01 -4.54555213e-01 1.07991278e+00
6.43296778e-01 -1.20734371e-01 -4.84966457e-01 -9.71469700e-01
1.13660358e-01 -3.66496518e-02 7.99788177e-01 -9.35394943e-01
3.38126749e-01 -1.97470173e-01 4.58720833e-01 -1.36676431e+00
5.90248466e-01 -9.58512485e-01 -8.80585089e-02 5.27858473e-02
3.05839330e-01 -1.28197381e-02 3.89467508e-01 6.03337586e-01
-4.19282883e-01 3.37786973e-01 7.03247666e-01 -4.07825142e-01
-1.13023746e+00 7.12229490e-01 6.80376440e-02 -3.66189033e-02
1.03137338e+00 -7.94231057e-01 -2.54598826e-01 -5.83627522e-02
-8.73185694e-01 4.89099175e-01 6.47662640e-01 5.32266974e-01
1.00575280e+00 -1.34961283e+00 -4.28275466e-01 3.98523241e-01
1.85647622e-01 9.01397586e-01 7.05703139e-01 9.31999445e-01
-8.03312361e-01 4.03760761e-01 -3.52489084e-01 -1.36741436e+00
-1.12580621e+00 3.33783507e-01 3.85826498e-01 -5.17982990e-02
-8.67221177e-01 1.04284501e+00 6.53316200e-01 -9.59176958e-01
3.62734556e-01 -4.56182003e-01 -8.53301734e-02 -5.80258369e-02
1.59422994e-01 2.42333367e-01 1.49206407e-02 -6.48604870e-01
-5.06362140e-01 9.98157084e-01 8.64927694e-02 1.91063389e-01
1.50778532e+00 -1.90582171e-01 -3.09918016e-01 3.85443985e-01
1.08021593e+00 -4.29805636e-01 -1.42249429e+00 -3.81685346e-01
-3.63838226e-01 -6.82070136e-01 7.42385089e-02 -6.18163407e-01
-1.49285388e+00 1.15998971e+00 5.13382792e-01 4.74514693e-01
1.20748007e+00 4.96195108e-01 7.11452723e-01 2.05909267e-01
4.14063305e-01 -8.12318921e-01 1.45737916e-01 7.44822562e-01
8.41194749e-01 -1.15980375e+00 5.06181940e-02 -9.35646176e-01
-6.07512534e-01 1.11104417e+00 8.59092832e-01 -2.24927053e-01
9.03957427e-01 1.44604042e-01 1.64420456e-01 -8.82457316e-01
-3.05177569e-01 -3.71256828e-01 6.70647442e-01 6.70770109e-01
-7.05640838e-02 1.71187684e-01 2.77190089e-01 3.15360069e-01
-9.81056467e-02 -4.19887424e-01 1.20412558e-02 1.04079080e+00
-3.47593963e-01 -8.89538527e-01 -4.00805175e-01 3.11611176e-01
3.68129492e-01 1.06790900e-01 -5.43591976e-01 1.00431848e+00
4.18260127e-01 6.92912281e-01 4.78094310e-01 -4.74974364e-01
5.55156887e-01 -3.02908927e-01 4.11399096e-01 -6.75891459e-01
-1.98571384e-01 5.23291938e-02 -3.48272957e-02 -9.91328895e-01
-5.27473152e-01 -4.50519919e-01 -1.40796590e+00 -3.03035140e-01
1.99874610e-01 -4.25664574e-01 6.34300709e-01 6.79207504e-01
5.25790453e-01 4.69525337e-01 4.35452193e-01 -1.16116738e+00
1.07243188e-01 -4.32403684e-01 -5.83473444e-01 3.56339395e-01
6.26464859e-02 -9.05692518e-01 -1.09543186e-02 -1.62188575e-01] | [8.006296157836914, -3.26090145111084] |
23754b08-0e83-4bd1-ac5c-f64556da6a16 | federated-nonconvex-sparse-learning | 2101.00052 | null | https://arxiv.org/abs/2101.00052v1 | https://arxiv.org/pdf/2101.00052v1.pdf | Federated Nonconvex Sparse Learning | Nonconvex sparse learning plays an essential role in many areas, such as signal processing and deep network compression. Iterative hard thresholding (IHT) methods are the state-of-the-art for nonconvex sparse learning due to their capability of recovering true support and scalability with large datasets. Theoretical analysis of IHT is currently based on centralized IID data. In realistic large-scale situations, however, data are distributed, hardly IID, and private to local edge computing devices. It is thus necessary to examine the property of IHT in federated settings, which update in parallel on local devices and communicate with a central server only once in a while without sharing local data. In this paper, we propose two IHT methods: Federated Hard Thresholding (Fed-HT) and Federated Iterative Hard Thresholding (FedIter-HT). We prove that both algorithms enjoy a linear convergence rate and have strong guarantees to recover the optimal sparse estimator, similar to traditional IHT methods, but now with decentralized non-IID data. Empirical results demonstrate that the Fed-HT and FedIter-HT outperform their competitor - a distributed IHT, in terms of decreasing the objective values with lower requirements on communication rounds and bandwidth. | ['Jinbo Bi', 'Tan Zhu', 'Guannan Liang', 'Qianqian Tong'] | 2020-12-31 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [ 3.80360223e-02 1.47894949e-01 -1.79834664e-01 -3.63400169e-02
-8.99170518e-01 -2.54119009e-01 -8.17563161e-02 2.29714327e-02
-2.63352185e-01 8.42591345e-01 2.12874219e-01 -6.41936436e-02
-2.95021385e-01 -5.87046444e-01 -1.05850291e+00 -9.53948617e-01
-4.96614486e-01 5.53584874e-01 2.45819278e-02 3.25907290e-01
-2.78395057e-01 -4.81458679e-02 -1.04420257e+00 1.09463140e-01
6.75680041e-01 1.29660976e+00 2.48695746e-01 3.47622067e-01
2.99152285e-01 8.53092313e-01 -1.85699716e-01 -3.27699542e-01
6.50570095e-01 -2.49660090e-01 -3.74381870e-01 1.36459038e-01
3.38108569e-01 -2.79855639e-01 -5.02812266e-01 1.26106071e+00
6.60046339e-01 -6.76046452e-03 2.49453887e-01 -1.32319963e+00
-3.02142560e-01 9.07292247e-01 -9.39887106e-01 9.06010792e-02
1.12732440e-01 -2.29395568e-01 7.85861731e-01 -1.08316970e+00
7.10957944e-01 7.56020486e-01 1.28633904e+00 1.33641407e-01
-1.12789452e+00 -5.91196239e-01 1.24346405e-01 2.16552988e-01
-1.60244083e+00 -7.64925480e-01 5.84541857e-01 -1.11733891e-01
4.33225423e-01 3.82588893e-01 7.80818045e-01 5.46296060e-01
-1.91100806e-01 1.02319753e+00 1.20263588e+00 -2.74107903e-01
6.75921440e-01 -5.62096611e-02 -1.37792736e-01 8.27168941e-01
7.87390053e-01 -2.80673862e-01 -1.04351819e+00 -5.27394116e-01
7.60356128e-01 1.72627628e-01 -5.63641965e-01 -4.87074018e-01
-1.42998040e+00 7.54541159e-01 3.13269138e-01 2.44922236e-01
-8.01221550e-01 4.04377848e-01 6.27483189e-01 6.70991480e-01
8.15055430e-01 -3.24916065e-01 -4.28983003e-01 -1.23851053e-01
-1.46950185e+00 -1.19957328e-01 1.23488688e+00 1.05785203e+00
6.61065280e-01 1.32375300e-01 -1.30745962e-01 5.17073691e-01
9.99829471e-02 8.02814424e-01 3.28753352e-01 -1.02735674e+00
6.10975325e-01 -2.51313020e-02 8.47766548e-02 -1.31466424e+00
-2.75540233e-01 -6.29497468e-01 -1.47806394e+00 -2.58272052e-01
1.39877036e-01 -4.39805269e-01 -3.86096448e-01 1.66824722e+00
5.91072917e-01 7.70444036e-01 -8.42259675e-02 1.25092101e+00
6.21953964e-01 6.75630629e-01 -5.08856535e-01 -6.54336989e-01
9.66724277e-01 -8.13900709e-01 -6.84898078e-01 -6.51328564e-02
7.35759974e-01 -5.70186019e-01 5.70132971e-01 6.16076529e-01
-1.29360151e+00 8.80848095e-02 -8.00196171e-01 1.11860707e-01
1.70572117e-01 2.20012888e-01 6.39367163e-01 6.94853604e-01
-1.25891459e+00 4.49610889e-01 -1.08186376e+00 -1.96835116e-01
7.26207793e-01 4.70074952e-01 -3.38239133e-01 -4.34066534e-01
-6.32458091e-01 2.73866087e-01 -6.52591139e-02 9.38840136e-02
-8.35584581e-01 -7.74054527e-01 -7.11859584e-01 2.75683492e-01
4.54851568e-01 -9.69760180e-01 9.99733508e-01 -8.42839420e-01
-1.30832839e+00 7.01277971e-01 -4.42935407e-01 -7.03662395e-01
8.88047814e-01 -2.61926036e-02 1.20730482e-01 3.63958895e-01
3.42376679e-01 6.79344460e-02 1.13945520e+00 -1.11988997e+00
-5.51109612e-01 -5.80459833e-01 -4.62570935e-01 8.24493319e-02
-7.40795374e-01 -4.78112489e-01 -4.74835962e-01 -7.02199042e-01
4.24420089e-01 -1.00166106e+00 -4.33590949e-01 2.99984306e-01
-4.47009683e-01 1.07676201e-01 1.09045100e+00 -6.23991013e-01
1.01979840e+00 -2.22785449e+00 2.16052428e-01 5.60598612e-01
7.02749372e-01 5.93791157e-02 1.54087869e-02 2.77774155e-01
4.64888066e-01 -2.37814039e-01 -3.11918110e-01 -8.42508078e-01
5.90578057e-02 2.92253703e-01 -8.96035507e-02 1.14669919e+00
-8.03551137e-01 6.88655555e-01 -8.87630701e-01 -5.75361013e-01
-9.38377380e-02 4.62855071e-01 -6.85518980e-01 -2.29645863e-01
2.27102518e-01 3.40097576e-01 -6.40374839e-01 8.45643461e-01
8.14397156e-01 -6.59423828e-01 4.46789324e-01 -3.49548608e-01
8.40972662e-02 -2.48740837e-01 -1.58400559e+00 1.80269396e+00
-5.14820039e-01 6.13054097e-01 1.10470688e+00 -1.50772166e+00
4.57808882e-01 5.70539892e-01 8.21242034e-01 -3.80333990e-01
2.19896451e-01 7.22258151e-01 -7.73281634e-01 -3.51691425e-01
1.81314975e-01 -6.41088039e-02 8.12975466e-02 6.99798346e-01
-1.75549671e-01 1.66184589e-01 2.81779356e-02 4.04381096e-01
1.37238443e+00 -5.87570310e-01 1.39346853e-01 -4.99889046e-01
9.70199332e-02 -2.06649065e-01 8.72997940e-01 8.86499643e-01
-2.32978798e-02 5.68438649e-01 1.59614071e-01 -2.88541257e-01
-9.62750852e-01 -6.37760341e-01 -7.39283487e-03 8.52822661e-01
3.03525478e-01 -3.66337448e-01 -8.07223141e-01 -4.04472411e-01
3.63275588e-01 2.73328298e-03 -3.45532060e-01 1.92359477e-01
-4.29236621e-01 -7.06116438e-01 1.62893623e-01 3.24785918e-01
7.07674086e-01 -4.54010993e-01 -7.34223843e-01 2.42952049e-01
-3.38954508e-01 -1.34715903e+00 -7.98015952e-01 2.49002919e-01
-1.06954336e+00 -8.81211817e-01 -9.87549961e-01 -8.39379549e-01
8.96962106e-01 8.08175147e-01 9.72608864e-01 1.23991586e-01
-2.43612915e-01 6.72470987e-01 -3.13650370e-01 -1.98260516e-01
1.84471652e-01 1.35338351e-01 2.84646064e-01 5.25103450e-01
-3.76409382e-01 -1.00742173e+00 -7.81840503e-01 1.80674672e-01
-8.91138494e-01 -8.84217545e-02 6.16293550e-01 8.75550210e-01
1.12026989e+00 2.66719460e-01 6.25631809e-01 -1.05534875e+00
4.63893622e-01 -8.17657530e-01 -5.90995252e-01 6.03606068e-02
-5.08147955e-01 -3.02615166e-01 7.02170908e-01 -4.43472654e-01
-7.33041763e-01 3.26125741e-01 2.17691585e-01 -8.79117370e-01
5.29444516e-01 7.58990765e-01 1.47241861e-01 -5.63433170e-01
2.19201505e-01 3.79344106e-01 -4.08533700e-02 -4.55257565e-01
2.16507703e-01 6.56580210e-01 4.88644719e-01 -6.63043261e-01
8.65913570e-01 1.14614809e+00 1.86188743e-01 -1.01251733e+00
-9.40609694e-01 -6.79560542e-01 -1.33649662e-01 -1.95619315e-01
2.75483370e-01 -1.30373800e+00 -7.79327989e-01 3.81044447e-01
-6.67018473e-01 -4.81350064e-01 -7.12414145e-01 4.20010567e-01
-5.02141595e-01 5.08285820e-01 -5.87485433e-01 -6.92287803e-01
-7.18886375e-01 -7.71042705e-01 1.22278428e+00 9.56543908e-03
2.76530057e-01 -1.14078856e+00 -1.98562890e-01 3.11161071e-01
7.09570408e-01 3.85470510e-01 1.97783098e-01 -3.68264645e-01
-6.28416657e-01 -4.85293448e-01 -2.70253956e-01 1.42900884e-01
-1.25557125e-01 -7.94323087e-01 -6.54947698e-01 -9.66057360e-01
2.83617347e-01 -4.49035645e-01 7.36318290e-01 8.78473401e-01
1.24317765e+00 -8.91842484e-01 -5.27332187e-01 1.04763913e+00
1.54422617e+00 -6.76239729e-01 2.21350759e-01 3.44433240e-03
5.03714621e-01 -2.07044780e-02 2.56641358e-01 9.90082145e-01
4.79463488e-01 3.32134366e-01 5.77926636e-01 -2.80367106e-01
-2.58656412e-01 -1.29384488e-01 3.15045923e-01 1.25160336e+00
2.38470379e-02 1.84283685e-02 -4.72188056e-01 8.62807512e-01
-2.27395105e+00 -9.19940352e-01 -3.00429791e-01 2.42569447e+00
8.40384483e-01 -1.30524114e-01 3.91966626e-02 2.68161744e-01
7.21709847e-01 5.06686717e-02 -7.26281285e-01 2.37263948e-01
-3.59636158e-01 1.24653980e-01 9.54617739e-01 1.32654980e-01
-9.24917817e-01 4.37134355e-01 5.68736076e+00 8.90960097e-01
-1.11556458e+00 7.94946492e-01 8.57589304e-01 -4.11100566e-01
-1.28914461e-01 -5.59330359e-02 -4.40639466e-01 5.89054883e-01
6.29496276e-01 -4.28874016e-01 7.12309122e-01 1.00620353e+00
3.09420705e-01 -1.12206638e-01 -8.16083372e-01 1.47713447e+00
4.55584750e-02 -1.66865098e+00 -4.41072434e-01 2.76618361e-01
1.39485228e+00 4.87871140e-01 -7.21760020e-02 -1.08510450e-01
2.70384252e-01 -6.15462244e-01 8.23978305e-01 2.53639758e-01
8.50823402e-01 -5.48321962e-01 8.04816842e-01 5.52749217e-01
-1.37064314e+00 -1.91373512e-01 -5.58002293e-01 -8.83690733e-03
3.56573910e-01 1.58494484e+00 -2.60537475e-01 5.04191339e-01
1.10438347e+00 9.20321703e-01 4.22737002e-02 1.34115851e+00
2.25606754e-01 9.72941816e-01 -9.60204720e-01 6.93738833e-02
1.03683762e-01 -3.14583838e-01 8.27455819e-01 1.08658576e+00
6.97710752e-01 2.32919395e-01 5.95309615e-01 4.01240677e-01
-5.77364445e-01 9.25725102e-02 -4.87840831e-01 3.71580303e-01
7.29094863e-01 1.22896099e+00 -6.64623559e-01 -4.32074249e-01
-5.06755352e-01 9.50626612e-01 1.59996063e-01 2.61947811e-01
-5.78231454e-01 -7.52854720e-02 3.87270331e-01 2.83451855e-01
6.19488657e-01 -2.34742939e-01 -4.18704718e-01 -1.36915827e+00
3.97418946e-01 -8.21603656e-01 5.77732205e-01 -2.40217820e-01
-1.43158400e+00 2.01568708e-01 -4.27768439e-01 -1.10698175e+00
3.45763057e-01 -5.81520945e-02 -5.83730459e-01 2.19611660e-01
-1.58099592e+00 -1.02820778e+00 -5.91379941e-01 1.08227074e+00
3.62242460e-01 3.68335471e-02 3.49892139e-01 5.74307799e-01
-4.78377223e-01 6.72004879e-01 7.52911568e-01 -5.54823875e-02
3.77958536e-01 -9.12508965e-01 -2.39786357e-01 8.30682635e-01
2.17135161e-01 1.87259421e-01 6.57254219e-01 -6.12939358e-01
-2.18138051e+00 -1.26963401e+00 5.38718343e-01 3.52680981e-01
6.27009034e-01 -1.96001932e-01 -6.11096144e-01 6.78480804e-01
3.14733312e-02 8.82325947e-01 5.76252818e-01 -2.18728319e-01
-6.57953173e-02 -4.27221417e-01 -1.35805929e+00 1.22557804e-01
1.08276439e+00 -3.23482990e-01 6.01795316e-02 9.85381186e-01
3.73348385e-01 -4.23065066e-01 -9.24635589e-01 2.53130168e-01
2.24070579e-01 -8.76675129e-01 8.82949829e-01 2.26926394e-02
-8.03042799e-02 -2.56416261e-01 -2.88130522e-01 -1.14659035e+00
-1.29482940e-01 -1.40227520e+00 -5.98232806e-01 7.39529550e-01
7.42250308e-02 -9.46590662e-01 1.25351369e+00 2.47411400e-01
-2.30653837e-01 -7.41517901e-01 -1.60155880e+00 -1.07231271e+00
-4.31645662e-01 -2.76839167e-01 3.38206649e-01 1.09948456e+00
6.29730076e-02 -6.87350780e-02 -6.09228253e-01 4.21805173e-01
1.31024146e+00 3.39918733e-01 6.59599245e-01 -1.14951956e+00
-4.08176780e-01 -4.23177332e-02 -3.27716112e-01 -1.21672189e+00
-3.69813181e-02 -1.02456605e+00 -3.62524725e-02 -1.45266187e+00
4.27588820e-01 -7.64415085e-01 -5.98589219e-02 6.29940629e-01
2.15038061e-01 6.17357969e-01 1.75915420e-01 6.04148746e-01
-1.03059435e+00 6.89519107e-01 1.02413285e+00 -1.43846065e-01
-3.61717567e-02 5.15905721e-03 -4.18345660e-01 6.55909240e-01
5.69454849e-01 -7.81537294e-01 -2.58897156e-01 -7.45166898e-01
3.76249015e-01 3.32569599e-01 2.62008220e-01 -1.17172420e+00
7.82151103e-01 1.44246072e-01 2.61738729e-02 -1.82784945e-01
2.67075896e-01 -1.15953648e+00 2.39365935e-01 6.17332160e-01
5.26931994e-02 -1.43101633e-01 -3.05046439e-01 9.56420541e-01
-1.75336629e-01 -8.56698900e-02 6.41679704e-01 -1.25698075e-01
-1.05358928e-01 6.37033582e-01 -1.30224004e-01 2.33110875e-01
1.15661061e+00 -2.36283943e-01 -1.00984916e-01 -7.67322481e-01
-8.05520296e-01 3.49390417e-01 2.74739236e-01 -3.70135367e-01
5.72257936e-01 -1.23745966e+00 -1.01324165e+00 1.92383174e-02
-4.39175934e-01 2.82719016e-01 2.16068566e-01 1.55921125e+00
-2.93245345e-01 2.78221756e-01 4.28242475e-01 -6.30256891e-01
-1.13516331e+00 3.31878781e-01 2.34480668e-02 -2.61582732e-01
-9.83162940e-01 8.46578062e-01 -1.45763159e-01 -1.97215065e-01
5.22625089e-01 -7.23735988e-02 6.75280929e-01 -5.64452782e-02
3.21841121e-01 5.84518075e-01 3.67642790e-01 -2.97186613e-01
-6.35283068e-02 4.75329250e-01 3.38815957e-01 2.02170044e-01
1.70651519e+00 -3.49776715e-01 -3.53164405e-01 -6.05545845e-03
1.26141298e+00 1.16243482e-01 -1.13269758e+00 -5.62211335e-01
-2.53149301e-01 -5.39938986e-01 5.89562714e-01 -2.26236477e-01
-1.84701979e+00 2.04419643e-01 5.03400385e-01 3.32099229e-01
1.29844499e+00 7.52440020e-02 1.45234466e+00 5.51037729e-01
9.82171774e-01 -1.09373677e+00 -1.40548110e-01 7.75652006e-02
5.62492132e-01 -1.12357438e+00 3.29721808e-01 -5.88585556e-01
-2.60125190e-01 7.29469001e-01 1.35653652e-02 -2.45860353e-01
1.02626622e+00 5.60646653e-01 -4.67980027e-01 -1.83575079e-01
-7.69123197e-01 4.68754061e-02 -1.75667033e-01 3.12651038e-01
-1.45138815e-01 1.44871011e-01 -3.00638586e-01 5.72123945e-01
-5.06826714e-02 1.88062057e-01 3.83828998e-01 1.15909672e+00
-3.88566762e-01 -5.99708378e-01 -6.72998667e-01 8.89493346e-01
-3.99229169e-01 -2.31289510e-02 1.10036172e-01 3.82594645e-01
-1.21210456e-01 7.87613094e-01 -2.85623074e-01 -1.01698048e-01
-1.68622993e-02 -4.09396142e-01 3.60113174e-01 -3.33512962e-01
-4.46399480e-01 1.39089108e-01 -1.11785747e-01 -6.75788641e-01
-4.04333740e-01 -8.81197989e-01 -1.07458925e+00 -4.91437644e-01
-3.81299883e-01 3.08417231e-01 7.87924528e-01 7.89399266e-01
7.13587523e-01 8.02226663e-02 9.24691260e-01 -1.07602751e+00
-8.61947656e-01 -5.60446501e-01 -8.49786460e-01 1.95287108e-01
3.94889206e-01 -2.48156577e-01 -6.81126773e-01 1.37012914e-01] | [6.27708625793457, 5.023930549621582] |
b62a2974-3717-4e1d-9b99-6bed5b7d2a35 | effective-cross-lingual-transfer-of-neural | 1905.05475 | null | https://arxiv.org/abs/1905.05475v2 | https://arxiv.org/pdf/1905.05475v2.pdf | Effective Cross-lingual Transfer of Neural Machine Translation Models without Shared Vocabularies | Transfer learning or multilingual model is essential for low-resource neural machine translation (NMT), but the applicability is limited to cognate languages by sharing their vocabularies. This paper shows effective techniques to transfer a pre-trained NMT model to a new, unrelated language without shared vocabularies. We relieve the vocabulary mismatch by using cross-lingual word embedding, train a more language-agnostic encoder by injecting artificial noises, and generate synthetic data easily from the pre-training data without back-translation. Our methods do not require restructuring the vocabulary or retraining the model. We improve plain NMT transfer by up to +5.1% BLEU in five low-resource translation tasks, outperforming multilingual joint training by a large margin. We also provide extensive ablation studies on pre-trained embedding, synthetic data, vocabulary size, and parameter freezing for a better understanding of NMT transfer. | ['Yunsu Kim', 'Yingbo Gao', 'Hermann Ney'] | 2019-05-14 | effective-cross-lingual-transfer-of-neural-1 | https://aclanthology.org/P19-1120 | https://aclanthology.org/P19-1120.pdf | acl-2019-7 | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [ 2.66188651e-01 1.61219507e-01 -4.42925245e-01 -3.75318229e-01
-1.37682498e+00 -8.86450768e-01 6.03289008e-01 -4.67202872e-01
-7.90103495e-01 1.31102216e+00 3.93616170e-01 -7.86256492e-01
6.17500603e-01 -5.89273036e-01 -1.10081446e+00 -3.11793357e-01
4.75997835e-01 8.55762005e-01 -3.95404458e-01 -6.45615220e-01
-4.56047475e-01 2.42930725e-02 -6.93992913e-01 9.14654881e-02
1.18200850e+00 2.60237038e-01 3.25355768e-01 3.73486429e-01
-2.40266368e-01 1.07000291e-01 -5.91499388e-01 -7.38622069e-01
5.83028316e-01 -6.51317179e-01 -6.13790333e-01 -4.43493754e-01
6.93745971e-01 -3.55131596e-01 -3.74586344e-01 7.97285914e-01
7.02503741e-01 -3.51286620e-01 6.01296544e-01 -8.18685949e-01
-1.38026214e+00 9.60668445e-01 -2.99953729e-01 1.14980889e-02
5.03249913e-02 3.21451306e-01 1.00646544e+00 -1.19065630e+00
9.18498337e-01 1.05250895e+00 7.27587879e-01 9.66816962e-01
-1.49589443e+00 -1.21329463e+00 -6.23719543e-02 -3.96993876e-01
-1.41104841e+00 -6.30180359e-01 3.61633748e-01 -2.29728624e-01
1.42311907e+00 1.53864641e-03 5.31846166e-01 1.64287257e+00
2.55229801e-01 4.44452375e-01 1.22125387e+00 -5.57703912e-01
-1.93517596e-01 3.21685374e-01 -3.87697101e-01 5.63736916e-01
4.81945634e-01 5.55047542e-02 -4.98047292e-01 4.54803184e-02
8.93925786e-01 -4.42668080e-01 -2.20048547e-01 -2.55219638e-01
-1.46908164e+00 8.56079757e-01 2.30375051e-01 2.78927475e-01
-1.98553633e-02 3.07590008e-01 6.02056682e-01 1.03590536e+00
7.33210981e-01 8.70021403e-01 -8.27997684e-01 7.19087049e-02
-7.15413570e-01 -1.00268520e-01 5.09438276e-01 1.55017793e+00
1.01534963e+00 4.44259346e-01 -1.40181795e-01 9.80392039e-01
-4.24471274e-02 1.15204608e+00 7.89602339e-01 -4.71553773e-01
9.13097262e-01 7.17029870e-02 -1.55029017e-02 -3.72204185e-02
1.00862756e-01 -4.79945213e-01 -6.62077069e-01 -2.03985840e-01
2.20462680e-01 -6.04335010e-01 -1.06718290e+00 2.16350722e+00
-1.02429733e-01 -1.60335422e-01 5.26845455e-01 6.77846074e-01
5.63647449e-01 8.30271363e-01 3.31017710e-02 -6.43639863e-02
1.12885094e+00 -1.21524405e+00 -7.11888790e-01 -7.27702439e-01
9.82752562e-01 -1.00350630e+00 1.55399048e+00 -2.93856680e-01
-1.14927924e+00 -5.27792096e-01 -1.21845317e+00 -4.16533202e-01
-3.05697918e-01 1.28209397e-01 4.66875464e-01 5.13493240e-01
-1.10339212e+00 3.69330734e-01 -6.77918315e-01 -4.57524419e-01
3.21365356e-01 4.51076567e-01 -7.06207573e-01 -2.28027746e-01
-1.77108324e+00 1.41575837e+00 2.90487945e-01 -3.17748278e-01
-9.49621856e-01 -8.93636942e-01 -1.11529493e+00 -2.66982764e-01
-7.81992823e-02 -9.88872290e-01 1.18578494e+00 -1.09956658e+00
-1.68792892e+00 1.02212250e+00 -2.04736531e-01 -4.77445871e-01
4.62184370e-01 -2.87715405e-01 -4.30625945e-01 -4.40375060e-01
2.86534667e-01 8.76978934e-01 8.98549139e-01 -6.22909546e-01
-3.71670127e-02 5.28712869e-02 -2.21317723e-01 4.11291689e-01
-5.46303511e-01 1.43914714e-01 -4.58568066e-01 -9.22687888e-01
-4.72108781e-01 -1.15531993e+00 -5.29435501e-02 -1.56855062e-01
2.62565352e-02 2.28141490e-02 3.74653548e-01 -8.21163237e-01
8.34147751e-01 -1.94749951e+00 3.61527562e-01 -2.77488619e-01
-9.30526480e-02 3.90673101e-01 -7.63869643e-01 5.55323362e-01
9.59649310e-02 4.64305252e-01 -3.06353927e-01 -4.17987436e-01
3.43573503e-02 4.56696242e-01 -5.46426535e-01 2.63587832e-01
4.76743400e-01 1.53557432e+00 -9.10301685e-01 -3.53766501e-01
-1.32542431e-01 4.70975161e-01 -5.42574286e-01 1.43556088e-01
-2.82447487e-01 5.97589433e-01 -1.06547125e-01 2.77397364e-01
3.64922822e-01 -6.10789172e-02 3.38867068e-01 4.36214395e-02
1.30669802e-01 8.12125683e-01 -3.97475719e-01 2.29125023e+00
-1.15076113e+00 7.81440258e-01 -1.97097823e-01 -5.37581921e-01
9.30697739e-01 4.84479696e-01 -1.35203391e-01 -8.49773526e-01
2.74406914e-02 7.57739544e-01 1.97528467e-01 -8.15072358e-02
3.24646413e-01 -5.22838652e-01 -3.77776772e-01 8.16254795e-01
4.87753898e-01 -4.23557878e-01 1.25165228e-02 -1.06460370e-01
7.59204388e-01 3.61026704e-01 2.04205275e-01 -4.60839063e-01
7.10685775e-02 9.34815630e-02 5.62968910e-01 3.70497018e-01
1.12333789e-01 4.71405119e-01 -9.71619226e-03 -3.92577887e-01
-1.52382493e+00 -1.12997603e+00 -1.20769687e-01 1.19729018e+00
-2.02620178e-01 -3.29649419e-01 -6.36273861e-01 -7.10695386e-01
3.79293784e-02 8.28743041e-01 -4.64000970e-01 -3.14095795e-01
-1.07456875e+00 -5.46557605e-01 9.94792700e-01 5.79008877e-01
2.90657461e-01 -9.65284050e-01 2.71594465e-01 2.29095861e-01
-3.75149459e-01 -1.26523197e+00 -1.14940107e+00 3.29161644e-01
-8.92915547e-01 -2.31230259e-01 -1.00051665e+00 -1.21032000e+00
6.83604419e-01 6.24717474e-02 1.28969681e+00 -2.07469106e-01
2.13808030e-01 -2.03901976e-01 4.09299396e-02 -3.47663581e-01
-8.52656245e-01 7.14532018e-01 4.61790293e-01 -5.49397409e-01
5.25854766e-01 -5.84811270e-01 -1.83519065e-01 3.23577970e-01
-5.54119170e-01 2.57297039e-01 8.43423367e-01 1.10940230e+00
5.40289938e-01 -9.32622433e-01 8.69720638e-01 -8.42004478e-01
8.63453448e-01 -2.97821045e-01 -5.63508451e-01 3.53513390e-01
-7.20109880e-01 5.00544429e-01 8.51748228e-01 -8.21984410e-01
-8.15711617e-01 -3.92580420e-01 -8.40339158e-03 -4.72909540e-01
4.33357120e-01 2.25998908e-01 -2.36211896e-01 6.84623644e-02
9.34471905e-01 2.43126646e-01 -1.48642519e-02 -3.29559416e-01
8.40093136e-01 7.49162972e-01 2.55143642e-01 -6.33558452e-01
1.16486871e+00 -1.07737333e-01 -6.46502435e-01 -3.85009795e-01
-6.50019169e-01 4.84196581e-02 -7.17886269e-01 5.44434130e-01
8.15319479e-01 -1.45476770e+00 1.95603892e-01 7.22440407e-02
-1.24372363e+00 -6.70689762e-01 -3.51891190e-01 7.38323987e-01
-4.60856080e-01 -1.91561982e-01 -8.06301773e-01 9.88728702e-02
-8.20529878e-01 -1.30452216e+00 1.05906224e+00 -3.44370157e-01
-5.39802730e-01 -1.08712721e+00 4.40515906e-01 4.00809795e-01
8.24339330e-01 -3.89882952e-01 1.07051301e+00 -6.64461076e-01
-5.15756607e-01 3.32376622e-02 -2.02852294e-01 5.74552655e-01
4.03320402e-01 -6.15335941e-01 -7.37695217e-01 -6.56663835e-01
-3.68308485e-01 -7.10063100e-01 6.56816721e-01 -1.17018126e-01
3.63852829e-01 -2.82722354e-01 -2.63730109e-01 9.89392221e-01
1.32721436e+00 -1.80611968e-01 3.17980945e-01 2.43190065e-01
7.92138398e-01 2.32392728e-01 2.24348545e-01 -4.30591285e-01
2.57294178e-01 6.52094960e-01 -1.98362112e-01 -3.09094995e-01
-4.86729264e-01 -6.74467444e-01 7.81762540e-01 1.51685357e+00
1.55551404e-01 -1.87000126e-01 -8.82676065e-01 7.76940823e-01
-1.37325025e+00 -4.60416585e-01 4.05779600e-01 2.13373303e+00
1.55762041e+00 8.44776556e-02 -3.46655786e-01 -7.03468800e-01
5.70267379e-01 5.16006500e-02 -6.11039698e-01 -6.96664274e-01
-2.36286044e-01 7.59196401e-01 8.67065132e-01 8.73683393e-01
-6.54351592e-01 1.84565413e+00 6.68188953e+00 7.61295259e-01
-1.34817183e+00 5.66384077e-01 1.25871971e-01 -2.78068781e-01
-7.89458990e-01 2.69459207e-02 -8.64586353e-01 2.11471289e-01
1.20951128e+00 -4.22484696e-01 7.13603854e-01 6.21872902e-01
-1.86748222e-01 8.14863384e-01 -1.24150026e+00 7.79245853e-01
1.84814215e-01 -1.24799752e+00 3.23242158e-01 9.10833105e-02
1.10141373e+00 7.06366718e-01 1.02299258e-01 8.31156850e-01
6.23898566e-01 -1.25917304e+00 3.71169478e-01 -3.19218412e-02
1.73054218e+00 -6.00709558e-01 5.84888577e-01 1.69346869e-01
-7.78260887e-01 5.22461236e-01 -5.85913301e-01 -8.78257975e-02
1.80953145e-01 1.45689026e-01 -1.06502521e+00 3.90795857e-01
1.57242924e-01 5.30518711e-01 -3.52891088e-01 1.03060246e-01
-5.60482800e-01 6.84974730e-01 -3.58057529e-01 -1.25357926e-01
2.65711874e-01 -2.17475399e-01 3.26978505e-01 1.22000420e+00
3.95160168e-01 -5.19774318e-01 -1.50513753e-01 8.44233871e-01
-8.43890369e-01 4.72181767e-01 -9.88076270e-01 -2.84916192e-01
4.79778260e-01 8.24389458e-01 9.74476039e-02 -3.99676412e-01
-6.45768940e-01 1.44738281e+00 7.07367778e-01 5.37201107e-01
-8.29720736e-01 -4.23269093e-01 1.06247520e+00 -7.73300976e-02
2.46858671e-01 -3.92173141e-01 -2.63358861e-01 -1.65744424e+00
2.76000857e-01 -1.20776820e+00 -4.75719512e-01 -6.53435111e-01
-1.34524834e+00 1.04032016e+00 -1.65225655e-01 -1.15562320e+00
-6.30672038e-01 -5.12267470e-01 -3.06010723e-01 1.62696493e+00
-1.56361842e+00 -1.40550709e+00 3.39461386e-01 3.83271456e-01
6.47888243e-01 -5.31707287e-01 1.28100753e+00 4.92040455e-01
-3.79091889e-01 1.13079619e+00 2.95441598e-01 4.38578874e-01
1.42171609e+00 -1.07819235e+00 1.29196703e+00 8.03105652e-01
2.12874383e-01 1.03507054e+00 5.32973766e-01 -7.56731570e-01
-1.21190333e+00 -1.42886209e+00 1.44214177e+00 -8.65768373e-01
8.85743022e-01 -1.09909689e+00 -8.33059132e-01 1.25449502e+00
7.57241905e-01 -2.89438926e-02 7.91094065e-01 3.15853298e-01
-7.37156689e-01 -4.61139567e-02 -8.07366192e-01 1.06302273e+00
1.35258210e+00 -9.21079516e-01 -7.60694206e-01 4.39419657e-01
1.38577127e+00 -3.13134640e-01 -1.02523124e+00 4.08743888e-01
6.04543865e-01 1.03266120e-01 7.77930140e-01 -1.02068663e+00
3.70645523e-01 -6.48933649e-02 -2.09273368e-01 -1.85483098e+00
-1.94514886e-01 -8.09298754e-01 3.63247454e-01 1.00019681e+00
1.10748041e+00 -8.11571181e-01 4.89846587e-01 1.93236604e-01
-1.43811762e-01 -6.05456233e-01 -9.20533299e-01 -1.14326823e+00
1.04113066e+00 -2.43671089e-01 7.80443966e-01 1.21598971e+00
-5.57677522e-02 1.03302038e+00 -6.77205026e-01 -1.49382263e-01
2.36450806e-01 -2.67389193e-02 1.02554846e+00 -6.81430578e-01
-3.27588081e-01 -1.86142951e-01 6.90451637e-02 -1.29822731e+00
5.01955807e-01 -1.60906446e+00 -1.25609949e-01 -1.40503490e+00
-1.39820166e-02 -5.42088449e-01 -1.05902106e-01 6.99480176e-01
-2.68448085e-01 5.46531498e-01 1.11918546e-01 4.12330329e-01
-9.92037728e-02 8.09731066e-01 1.51423275e+00 -3.25759947e-01
-8.04598257e-02 -5.67565620e-01 -8.63690853e-01 2.70046204e-01
9.16184187e-01 -7.42675602e-01 -6.50963008e-01 -1.26557159e+00
3.05707395e-01 -1.27496436e-01 -2.91060030e-01 -5.58868706e-01
-1.13048740e-01 -1.52178869e-01 1.81642964e-01 1.25073522e-01
1.80802271e-01 -6.17820561e-01 -2.43650400e-04 2.96217412e-01
-4.86947030e-01 4.04147804e-01 5.71330845e-01 1.22188874e-01
1.55416783e-02 -5.16328663e-02 6.61749780e-01 -1.34639777e-02
-2.07069114e-01 3.49050105e-01 -3.14297408e-01 4.51867282e-01
6.51817203e-01 8.54300484e-02 -3.41870904e-01 -2.55601913e-01
-5.39536417e-01 2.73787647e-01 8.73286366e-01 8.31044316e-01
1.81888744e-01 -1.65604353e+00 -1.16200984e+00 5.99160433e-01
4.54256773e-01 -2.53431678e-01 -2.37626374e-01 4.60137963e-01
-5.37609816e-01 5.34511983e-01 -3.33926767e-01 -3.28304797e-01
-8.00016522e-01 5.38302600e-01 4.10905272e-01 -3.88586819e-01
-4.07219380e-01 7.06005812e-01 1.59729078e-01 -1.14662492e+00
-1.62023604e-01 -2.79581010e-01 5.09494901e-01 -4.81769800e-01
2.94504166e-01 -2.80152380e-01 2.00525537e-01 -5.26438475e-01
-1.63172841e-01 7.17339754e-01 -4.08050984e-01 -5.39786816e-01
9.01203334e-01 -6.36014566e-02 6.77460283e-02 4.35683519e-01
1.42468488e+00 3.71890157e-01 -8.98107767e-01 -6.02962971e-01
-3.79839242e-01 -8.91066268e-02 -2.56669343e-01 -8.82101119e-01
-6.08266592e-01 9.61631894e-01 2.36404955e-01 -6.83026671e-01
7.62381136e-01 -6.88396394e-02 1.16665673e+00 5.97099304e-01
6.22759879e-01 -8.70320201e-01 -2.57022053e-01 8.45942676e-01
9.79988456e-01 -1.36419153e+00 -2.61196613e-01 -8.71529570e-04
-6.52066767e-01 7.19955385e-01 6.88382208e-01 -1.37324244e-01
3.09151769e-01 3.92110169e-01 6.05632424e-01 4.08922076e-01
-9.48323309e-01 3.42073664e-02 2.63209075e-01 6.39305711e-01
6.72192574e-01 1.49185330e-01 -4.02155668e-01 2.98961163e-01
-6.41885459e-01 -1.48509368e-01 1.93625852e-01 5.72456419e-01
-4.54368703e-02 -1.66537023e+00 3.33771221e-02 2.32223228e-01
-5.14212668e-01 -7.99926102e-01 -3.53401810e-01 9.20603395e-01
-7.39370584e-02 4.89896983e-01 1.66121759e-02 -3.66032183e-01
1.23995349e-01 4.72040117e-01 6.16517365e-01 -9.74034905e-01
-5.49137115e-01 -3.32679716e-03 2.03689739e-01 -1.22809701e-01
2.07882430e-02 -3.61925721e-01 -9.74903047e-01 -2.74408996e-01
-3.83648962e-01 2.18330547e-01 6.41171873e-01 8.34350169e-01
5.77556431e-01 3.48774433e-01 2.93006957e-01 -4.10204142e-01
-6.39186442e-01 -1.42835855e+00 1.38991460e-01 2.70244896e-01
3.55870873e-01 -1.68453932e-01 -2.85896271e-01 -5.21074086e-02] | [11.642223358154297, 10.167377471923828] |
7c742759-5356-4b1c-9429-07058f386211 | towards-interpreting-vulnerability-of-multi | 2211.17071 | null | https://arxiv.org/abs/2211.17071v3 | https://arxiv.org/pdf/2211.17071v3.pdf | Interpreting Vulnerabilities of Multi-Instance Learning to Adversarial Perturbations | Multi-Instance Learning (MIL) is a recent machine learning paradigm which is immensely useful in various real-life applications, like image analysis, video anomaly detection, text classification, etc. It is well known that most of the existing machine learning classifiers are highly vulnerable to adversarial perturbations. Since MIL is a weakly supervised learning, where information is available for a set of instances, called bag and not for every instances, adversarial perturbations can be fatal. In this paper, we have proposed two adversarial perturbation methods to analyze the effect of adversarial perturbations to interpret the vulnerability of MIL methods. Out of the two algorithms, one can be customized for every bag, and the other is a universal one, which can affect all bags in a given data set and thus has some generalizability. Through simulations, we have also shown the effectiveness of the proposed algorithms to fool the state-of-the-art (SOTA) MIL methods. Finally, we have discussed through experiments, about taking care of these kind of adversarial perturbations through a simple strategy. Source codes are available at https://github.com/InkiInki/MI-UAP. | ['Avik Ranjan Adhikary', 'Xue-Mei Cao', 'Mei Yang', 'Zhengchun Zhou', 'Hua Meng', 'Yu-Xuan Zhang'] | 2022-11-30 | null | null | null | null | ['video-anomaly-detection', 'multiple-instance-learning'] | ['computer-vision', 'methodology'] | [ 1.06640540e-01 -9.87173617e-02 5.56669151e-03 3.05201914e-02
-4.27556187e-01 -5.74208856e-01 4.12854761e-01 2.66702771e-01
-1.39881521e-01 1.03415120e+00 -1.66899785e-01 -3.39691520e-01
-8.50193948e-02 -9.15986359e-01 -1.06186044e+00 -1.01843619e+00
-3.76428366e-02 3.09724480e-01 3.16584021e-01 -4.67412263e-01
2.39316121e-01 5.01463473e-01 -1.45343566e+00 3.36558729e-01
8.66745353e-01 9.45990145e-01 -5.07479191e-01 5.49268484e-01
-1.07881136e-01 9.35087562e-01 -6.54915214e-01 -6.14633560e-01
3.55445504e-01 -2.44177505e-01 -7.47790873e-01 -2.16718003e-01
3.13740611e-01 -3.86507064e-02 -4.84463125e-01 1.23639178e+00
5.56463897e-01 1.75606042e-01 7.35706210e-01 -1.75581384e+00
-6.58758163e-01 6.93267882e-01 -5.50416708e-01 4.45573747e-01
1.73091605e-01 3.14600825e-01 4.27577972e-01 -5.17086864e-01
9.10616592e-02 1.36272728e+00 5.37872970e-01 6.95424020e-01
-6.79474533e-01 -1.01495516e+00 3.43147874e-01 3.88829410e-01
-1.19557929e+00 -2.43160814e-01 9.38898444e-01 -3.47424954e-01
3.35048139e-01 6.09603345e-01 2.05280438e-01 1.14600098e+00
3.67371380e-01 8.48280191e-01 1.16166615e+00 -2.70453721e-01
1.33843973e-01 3.70104641e-01 9.66257378e-02 4.34315950e-01
2.29321703e-01 1.36171341e-01 -1.69883650e-02 -5.00335395e-01
2.21042320e-01 2.48899296e-01 -4.53213722e-01 -7.28686079e-02
-9.14356887e-01 6.98540032e-01 4.83634382e-01 2.59181172e-01
-2.42901649e-02 2.16067154e-02 8.44026744e-01 3.16246718e-01
5.89263141e-01 1.94123834e-01 -4.21916872e-01 1.24311544e-01
-2.02673048e-01 3.46487254e-01 6.42428219e-01 8.62689078e-01
5.72035730e-01 1.48695886e-01 -1.96170419e-01 7.08695114e-01
3.66327390e-02 4.21685994e-01 7.44004965e-01 -1.36902586e-01
6.59921885e-01 5.30647874e-01 1.44022420e-01 -1.26787972e+00
-2.80860960e-01 -2.16851458e-01 -1.12540054e+00 1.26810297e-01
3.40327919e-01 -2.98706323e-01 -7.74126828e-01 1.64371097e+00
5.34607768e-01 7.77849078e-01 2.31609926e-01 6.04637265e-01
8.68727624e-01 6.67739213e-01 1.40272945e-01 -3.55193824e-01
8.99757028e-01 -8.45624626e-01 -6.37951136e-01 -2.40344524e-01
4.14957166e-01 -7.98755586e-01 1.10707986e+00 4.88875777e-01
-6.43906415e-01 -3.56727034e-01 -1.00302172e+00 5.63691437e-01
-7.03175962e-01 -5.21613121e-01 6.46770477e-01 8.11272919e-01
-3.19474667e-01 7.11588919e-01 -5.42229652e-01 -2.14473218e-01
5.33919513e-01 2.72637010e-01 -4.48593020e-01 -1.67107433e-01
-1.58571327e+00 8.08261454e-01 6.32007003e-01 1.97442591e-01
-1.00040567e+00 -5.49965918e-01 -6.34752452e-01 -2.95129329e-01
3.51612747e-01 -2.32152149e-01 8.56421411e-01 -1.10537994e+00
-1.07082295e+00 5.76161802e-01 1.03349164e-01 -3.09059203e-01
7.71003842e-01 -2.57979512e-01 -7.08708644e-01 -1.18517704e-01
-2.19283625e-01 -2.45995522e-01 9.07246947e-01 -1.33806694e+00
-3.04088473e-01 -4.20314223e-01 2.08038181e-01 4.52459902e-02
-5.68268955e-01 1.22386761e-01 -1.46520277e-02 -8.19161594e-01
-1.66756332e-01 -8.42726350e-01 -2.23922282e-01 -3.96091729e-01
-8.73100519e-01 -2.92018019e-02 1.09838879e+00 -4.52844203e-01
1.22687984e+00 -2.18835759e+00 7.76752383e-02 2.01823100e-01
-1.07961938e-01 5.61642051e-01 2.00487852e-01 4.23296511e-01
-3.87886941e-01 6.07766747e-01 -3.99332166e-01 -3.68455201e-02
-2.02910498e-01 2.20146596e-01 -6.87885404e-01 7.04980969e-01
4.10047323e-02 5.12579203e-01 -8.75311136e-01 -2.83827782e-01
3.46303850e-01 9.50770304e-02 -2.72340119e-01 4.38264817e-01
-2.26813003e-01 6.18314922e-01 -6.05743110e-01 7.84326017e-01
8.98678422e-01 2.94426858e-01 -2.78784811e-01 -2.58378863e-01
4.53579962e-01 -4.62008983e-01 -1.39999783e+00 8.81678879e-01
-2.25976035e-01 2.33301386e-01 -3.34023356e-01 -1.38284671e+00
7.13483095e-01 4.79521871e-01 3.46888483e-01 -2.18229011e-01
4.76842612e-01 1.03127412e-01 -1.38071580e-02 -6.40162766e-01
8.97144452e-02 -1.94986403e-01 -2.01668769e-01 3.98729950e-01
-1.23333044e-01 1.30151883e-01 1.36973396e-01 6.20905682e-02
9.44789886e-01 -3.56186420e-01 3.51944029e-01 -1.68394506e-01
1.07502902e+00 -3.20313632e-01 6.93530858e-01 7.12021232e-01
-3.89445126e-01 4.82328236e-01 4.29041117e-01 -5.62036335e-01
-9.51476693e-01 -1.07850528e+00 -2.86892146e-01 8.32346201e-01
3.52315515e-01 8.28083977e-02 -6.81774259e-01 -8.99066508e-01
2.04853669e-01 7.53768682e-01 -5.86210787e-01 -6.84486568e-01
-4.47323203e-01 -1.05898249e+00 8.55205417e-01 2.68364102e-01
6.32791877e-01 -1.25552118e+00 1.81608573e-01 -1.53286860e-03
1.05863800e-02 -1.05365205e+00 -2.73461878e-01 -7.33746663e-02
-6.86725438e-01 -1.43628299e+00 -2.63098061e-01 -5.06438911e-01
8.20123494e-01 -4.99086343e-02 8.96114111e-01 3.32770526e-01
-3.11939865e-01 2.40189284e-01 -5.94125032e-01 -8.13488424e-01
-8.64555001e-01 -1.62531212e-01 6.14042401e-01 3.29551429e-01
2.49909535e-01 -7.13963747e-01 -2.20115051e-01 3.81812721e-01
-1.25701165e+00 -2.75689691e-01 3.42391700e-01 8.20419967e-01
5.37705958e-01 4.57219660e-01 7.57132530e-01 -1.30881178e+00
6.90059364e-01 -9.18289781e-01 -3.14137429e-01 3.10355872e-01
-3.21476102e-01 -2.20143482e-01 1.35027099e+00 -7.86237478e-01
-8.74061942e-01 -2.65138417e-01 -2.51585931e-01 -6.87786341e-01
-4.15013582e-01 2.41304100e-01 -6.38751209e-01 -3.33245814e-01
8.12532365e-01 3.21043611e-01 -5.25617748e-02 -3.26572895e-01
1.95700362e-01 8.13186288e-01 3.82061064e-01 -8.30009341e-01
1.17456865e+00 2.83596545e-01 -9.54143051e-03 -6.03115261e-01
-8.88575375e-01 -1.43525213e-01 -2.88440526e-01 -2.38532200e-01
4.78552490e-01 -5.88056862e-01 -6.41610861e-01 1.00020552e+00
-8.20799410e-01 -1.42659232e-01 9.97362509e-02 1.21822581e-01
-3.89317006e-01 5.47739625e-01 -4.80573475e-01 -7.47605443e-01
-3.89413834e-01 -1.16726601e+00 4.75098252e-01 4.56238538e-01
3.90660912e-01 -1.21705878e+00 -2.60714702e-02 3.88324291e-01
1.91530287e-01 8.88173938e-01 8.83451760e-01 -1.17827296e+00
-3.16423297e-01 -4.97230142e-01 2.67382503e-01 7.75184631e-01
2.70015985e-01 4.82623354e-02 -1.05886233e+00 -5.55681169e-01
4.01075333e-02 -4.31934863e-01 5.22051513e-01 -4.16044630e-02
1.76863158e+00 -8.08504462e-01 -2.27172032e-01 7.09838212e-01
1.45842206e+00 2.80375481e-01 8.92681658e-01 4.30926263e-01
9.53426898e-01 2.89585330e-02 7.69631386e-01 4.06607747e-01
-2.06481740e-01 3.51984411e-01 9.20276999e-01 -6.79362789e-02
5.48078179e-01 -1.44070938e-01 3.53591740e-01 5.45565009e-01
-9.58855823e-02 -5.00854492e-01 -7.60204434e-01 3.95137727e-01
-1.94516504e+00 -1.18741882e+00 -3.51789370e-02 2.35111880e+00
8.51135790e-01 3.40444624e-01 -1.18617944e-01 5.36005855e-01
1.04663622e+00 2.43737474e-01 -7.72985041e-01 -6.13262355e-01
-1.12662338e-01 1.32084474e-01 6.16425216e-01 3.21269929e-01
-1.47189903e+00 8.41403186e-01 5.43620586e+00 9.98820186e-01
-1.04984272e+00 1.19929977e-01 8.20046723e-01 -2.95846947e-02
-1.58130646e-01 -2.08787993e-01 -6.19200349e-01 8.45917106e-01
5.75895727e-01 -4.79235947e-01 4.33350056e-01 9.39289331e-01
-2.49151904e-02 1.91503063e-01 -9.59339559e-01 7.20838428e-01
1.76936015e-01 -1.11635232e+00 4.13237542e-01 -3.96608353e-01
7.34271884e-01 -3.72028083e-01 1.15188532e-01 4.29661691e-01
1.45083830e-01 -1.20701551e+00 2.68608272e-01 4.27982062e-01
4.31144714e-01 -1.14949441e+00 1.07130206e+00 7.11177468e-01
-9.22012627e-01 -2.48379797e-01 -5.25927246e-01 -4.27779369e-02
-1.74458042e-01 7.27534115e-01 -5.00849187e-01 7.38669991e-01
6.71733022e-01 5.07220805e-01 -6.08165622e-01 9.47990358e-01
-2.34566167e-01 8.14258516e-01 -2.03601390e-01 3.99147682e-02
4.29482274e-02 -3.27127916e-03 7.92253494e-01 1.00430429e+00
3.00164819e-02 1.81686863e-01 2.56956697e-01 4.16046143e-01
-1.52331218e-01 3.50589186e-01 -8.67140353e-01 2.18527377e-01
5.94897211e-01 1.13670218e+00 -2.87354648e-01 -7.05284774e-02
-1.38890296e-01 8.62224936e-01 1.83945030e-01 3.00438881e-01
-1.09927738e+00 -4.55332339e-01 8.07079554e-01 -4.76271547e-02
-1.39486447e-01 3.33585620e-01 -6.82365522e-02 -1.14967978e+00
2.18185931e-01 -1.08971381e+00 5.83598137e-01 -5.49644470e-01
-1.65327859e+00 4.84928310e-01 2.22730916e-02 -1.44866514e+00
1.49307996e-01 -6.77326858e-01 -1.10150957e+00 8.06931615e-01
-1.18106568e+00 -1.19739473e+00 -4.46367890e-01 9.77305830e-01
2.38511726e-01 -5.61089218e-01 7.69820213e-01 4.45206463e-01
-1.00475347e+00 1.10375237e+00 3.48265350e-01 3.47868323e-01
8.19963694e-01 -1.00733209e+00 8.94839838e-02 1.20030189e+00
-1.40487894e-01 4.24399585e-01 9.26004529e-01 -5.34820199e-01
-1.12909293e+00 -1.50269520e+00 1.14950337e-01 -4.38009918e-01
6.37452304e-01 -2.18099430e-01 -1.09607244e+00 8.96163642e-01
-1.84466496e-01 4.70146805e-01 6.76460266e-01 -1.85808241e-01
-3.43466222e-01 -3.63890380e-01 -1.68063223e+00 7.46542513e-01
6.94095254e-01 -2.28343114e-01 -5.03142893e-01 7.45152891e-01
8.67983758e-01 -6.90003335e-01 -8.94302726e-01 5.70310831e-01
1.47159964e-01 -1.15464842e+00 1.14441228e+00 -1.03735757e+00
5.95171213e-01 -3.92082721e-01 -6.41860440e-02 -1.57495201e+00
-9.71919075e-02 -4.34491605e-01 -3.50079596e-01 1.45337093e+00
1.01355381e-01 -1.06537640e+00 5.06917119e-01 4.23925340e-01
-1.14680901e-01 -9.83399391e-01 -1.00835812e+00 -8.59741986e-01
3.81094813e-01 -4.24288094e-01 8.37663651e-01 1.15823317e+00
-6.83813244e-02 -1.85132056e-01 -5.81334293e-01 6.77827835e-01
6.47890091e-01 -3.59325528e-01 8.79638910e-01 -8.86861384e-01
-3.24143589e-01 -1.82840407e-01 -8.54945362e-01 -1.64252788e-01
3.11762869e-01 -8.15208972e-01 -1.65866718e-01 -1.01006615e+00
1.29857466e-01 -7.67563760e-01 -7.13760197e-01 5.43508410e-01
-7.01494575e-01 2.51141280e-01 7.21466392e-02 1.07795388e-01
-2.69620925e-01 2.60870755e-01 1.14124811e+00 -3.04334074e-01
3.72220054e-02 2.97584891e-01 -8.14517260e-01 8.49330485e-01
1.19841588e+00 -5.17407537e-01 -4.40068722e-01 -7.41146281e-02
-1.87489931e-02 -2.75386214e-01 4.16801125e-01 -1.04043877e+00
-1.10855289e-02 -5.28418362e-01 2.42253840e-01 -1.37334287e-01
-1.46484775e-02 -9.27552521e-01 1.29234821e-01 5.17676890e-01
-6.92856088e-02 -1.62638560e-01 5.17832935e-01 6.22630537e-01
-2.15481192e-01 -3.55216444e-01 1.07381499e+00 -1.97362557e-01
-6.50466681e-01 7.06474841e-01 8.93962905e-02 3.98519516e-01
1.59652114e+00 1.14547379e-01 -6.17452860e-01 -2.55140126e-01
-4.55171525e-01 2.35156536e-01 2.28848651e-01 4.27389741e-01
5.53365827e-01 -1.52459347e+00 -7.87182510e-01 9.88649204e-02
1.87139168e-01 1.83858901e-01 4.93516177e-01 5.82151651e-01
-5.10254085e-01 -7.81078711e-02 -3.29540700e-01 -3.21063608e-01
-1.33048356e+00 1.06471002e+00 4.56813037e-01 -2.60325700e-01
-3.48152310e-01 7.89344788e-01 1.37621105e-01 -4.21759218e-01
2.88223922e-01 2.17124432e-01 -3.21909040e-01 -3.90965253e-01
5.37924051e-01 4.50346023e-01 -8.96185078e-03 -5.96401572e-01
-4.12230313e-01 4.36968327e-01 -2.50259280e-01 5.52549183e-01
9.51484382e-01 2.64325470e-01 -2.82605559e-01 5.59596002e-01
1.12082517e+00 1.29876167e-01 -7.44733572e-01 -1.35922983e-01
-5.05136967e-01 -5.01734197e-01 -3.71469349e-01 -6.20165765e-01
-1.12794411e+00 7.24558532e-01 5.88297725e-01 4.16478336e-01
1.27116549e+00 -3.86808008e-01 8.27409029e-01 3.98359001e-01
3.57524037e-01 -7.98517823e-01 -7.72072449e-02 3.66422385e-01
1.00219178e+00 -1.52402771e+00 1.09626852e-01 -4.36035275e-01
-7.11871326e-01 9.20930326e-01 9.58179116e-01 -3.93515497e-01
8.05600464e-01 2.35724136e-01 -5.14104543e-03 3.11764151e-01
-4.59770143e-01 3.24958831e-01 4.56486940e-02 6.13855124e-01
2.32903361e-02 1.18285552e-01 -2.86524892e-01 8.05514395e-01
-6.28801957e-02 -2.21435159e-01 6.24519706e-01 7.71488905e-01
-2.70922214e-01 -1.05120540e+00 -7.45143890e-01 6.84170783e-01
-7.09998846e-01 -3.95104662e-02 -2.13137329e-01 7.44458020e-01
3.83509010e-01 9.54537809e-01 -1.62103817e-01 -6.55922830e-01
3.52732599e-01 1.13851994e-01 1.47520289e-01 -3.78304034e-01
-6.30234897e-01 -7.84384668e-01 -9.65770707e-02 -3.87424499e-01
-1.60697728e-01 -4.35272783e-01 -1.00242078e+00 -6.51449203e-01
-1.97963476e-01 2.06015319e-01 2.22299963e-01 1.14128411e+00
-3.44645642e-02 6.02833211e-01 1.11270607e+00 -6.22159600e-01
-7.07935095e-01 -9.73806024e-01 -5.77737570e-01 8.10143113e-01
2.71015286e-01 -8.46924484e-01 -9.49280918e-01 -7.41200894e-02] | [5.642446517944336, 7.852972507476807] |
3f992123-8e46-4663-819d-94aeaa5663a7 | negation-detection-in-clinical-reports | null | null | https://aclanthology.org/W16-5113 | https://aclanthology.org/W16-5113.pdf | Negation Detection in Clinical Reports Written in German | An important subtask in clinical text mining tries to identify whether a clinical finding is expressed as present, absent or unsure in a text. This work presents a system for detecting mentions of clinical findings that are negated or just speculated. The system has been applied to two different types of German clinical texts: clinical notes and discharge summaries. Our approach is built on top of NegEx, a well known algorithm for identifying non-factive mentions of medical findings. In this work, we adjust a previous adaptation of NegEx to German and evaluate the system on our data to detect negation and speculation. The results are compared to a baseline algorithm and are analyzed for both types of clinical documents. Our system achieves an F1-Score above 0.9 on both types of reports. | ['Danilo Schmidt', '', 'Rol Roller', 'Hans Uszkoreit', 'Feiyu Xu', 'Klemens Budde', 'Viviana Cotik'] | 2016-12-01 | null | null | null | ws-2016-12 | ['negation-detection'] | ['natural-language-processing'] | [ 1.36152178e-01 5.72116137e-01 -5.72527409e-01 -4.19416845e-01
-5.83565891e-01 -4.03431445e-01 5.44227004e-01 1.16215277e+00
-5.83923519e-01 1.09115243e+00 4.73239630e-01 -7.99419761e-01
-9.28322151e-02 -6.02599084e-01 -2.68462420e-01 -1.99594945e-01
-1.56603143e-01 4.73995388e-01 3.46704900e-01 -2.57384598e-01
3.17244798e-01 3.49777520e-01 -1.08479941e+00 1.10204887e+00
5.91180742e-01 5.68601966e-01 -3.87441456e-01 6.87091708e-01
-1.24514945e-01 1.20497119e+00 -7.66723573e-01 -6.72554135e-01
-2.63838947e-01 -5.81512451e-01 -9.67828929e-01 -2.02920496e-01
7.99425468e-02 4.65423763e-02 1.50715217e-01 1.05849254e+00
3.09657365e-01 -4.84568655e-01 7.32700825e-01 -7.71424174e-01
-1.54287040e-01 9.65029359e-01 -4.14417952e-01 8.08190286e-01
6.44845188e-01 -2.53771275e-01 8.91573906e-01 -8.18466663e-01
1.01591849e+00 9.07537878e-01 7.58227587e-01 5.34939647e-01
-9.48339880e-01 -4.44106668e-01 -7.42266774e-02 3.36631499e-02
-1.31710875e+00 -3.96946609e-01 8.63724947e-02 -5.48383236e-01
1.40975070e+00 5.97096801e-01 5.02586544e-01 6.33455396e-01
7.88439453e-01 4.75662559e-01 9.81109619e-01 -6.71580970e-01
3.19213927e-01 3.18006635e-01 5.23432136e-01 8.24574590e-01
6.69195771e-01 -2.01565191e-01 -4.91041213e-01 -7.55955398e-01
4.16010618e-02 -1.48211926e-01 -3.89417291e-01 5.98154962e-01
-1.22080827e+00 8.13688457e-01 -1.87432826e-01 7.23445475e-01
-6.28662825e-01 -5.06169200e-01 8.23209345e-01 3.30370814e-01
5.64025223e-01 5.03025353e-01 -8.53260517e-01 -3.68176028e-02
-1.10605371e+00 2.60462254e-01 1.29944193e+00 6.44307852e-01
-6.11223467e-02 -5.78888535e-01 -4.37403172e-01 6.22394323e-01
1.02257088e-01 4.90014665e-02 9.44928706e-01 -3.02436024e-01
3.34885210e-01 9.58372295e-01 1.08290784e-01 -1.03305423e+00
-6.99936211e-01 -3.28220427e-01 -5.00138938e-01 -1.46904334e-01
1.67644113e-01 -4.85659331e-01 -9.88075495e-01 1.32815266e+00
2.23772690e-01 5.49159423e-02 5.52643359e-01 4.98815954e-01
1.30676758e+00 3.70366007e-01 3.60730976e-01 -6.77389860e-01
1.66336620e+00 -4.20033157e-01 -1.21175671e+00 -1.68857537e-02
1.07878006e+00 -8.97220075e-01 1.70555279e-01 6.86074197e-01
-1.09834588e+00 1.93226352e-01 -8.46717596e-01 1.12492688e-01
-5.23445070e-01 3.87269884e-01 3.47132742e-01 5.02379298e-01
-7.99255848e-01 4.78108764e-01 -8.91365945e-01 -5.15746713e-01
1.68945342e-01 4.89712417e-01 -3.91695797e-01 3.57736617e-01
-1.38529015e+00 9.85123754e-01 8.15004349e-01 -8.31907541e-02
-7.01549230e-03 -6.37984812e-01 -7.45574594e-01 -7.81382322e-02
5.49410045e-01 -6.58046901e-01 1.45107841e+00 -4.96256232e-01
-7.85059869e-01 1.37444615e+00 -4.26824749e-01 -8.11368346e-01
5.22319853e-01 8.07951316e-02 -1.07548153e+00 2.08296120e-01
2.40406275e-01 3.52019966e-02 1.47344053e-01 -6.07796431e-01
-9.39495564e-01 -2.75526375e-01 -1.74444258e-01 -3.62130225e-01
-4.89645172e-03 3.71957541e-01 -3.54025252e-02 -8.87757480e-01
1.15183815e-01 -6.70695007e-01 -3.99238110e-01 -3.36381495e-01
-7.07024217e-01 -4.36942995e-01 4.50900763e-01 -6.27602279e-01
1.78561521e+00 -1.74740577e+00 -5.02862871e-01 3.52163374e-01
4.66677487e-01 2.80529648e-01 6.16939127e-01 5.54212868e-01
-4.95556921e-01 5.07225573e-01 -1.33291215e-01 1.75833285e-01
-4.97907817e-01 4.29303765e-01 -2.46117845e-01 3.71257901e-01
4.65899497e-01 6.40016258e-01 -1.02780461e+00 -8.87411833e-01
-3.15578431e-01 -6.62542600e-03 -5.09680331e-01 -1.40606523e-01
-2.25955486e-01 -6.77465871e-02 -5.35308719e-01 7.85088837e-01
5.04298091e-01 -4.21082884e-01 7.41336763e-01 -2.03215525e-01
-2.01881170e-01 7.51851141e-01 -9.39305723e-01 9.32283998e-01
1.97483450e-01 1.04428768e-01 -3.56749222e-02 -7.78991938e-01
7.89413393e-01 8.46704066e-01 4.42666292e-01 -2.24275067e-01
2.24139333e-01 3.94739151e-01 2.31485948e-01 -1.01383877e+00
4.80684012e-01 -5.88442266e-01 -7.53609017e-02 4.26253796e-01
-2.07627296e-01 3.42457473e-01 6.37330830e-01 4.77722168e-01
1.57044399e+00 -3.77420008e-01 1.29827154e+00 -5.27840078e-01
9.24094915e-01 4.76530433e-01 7.76814997e-01 5.66839516e-01
-1.68053225e-01 2.95907646e-01 1.07235789e+00 -2.94690937e-01
-3.47966611e-01 -5.96886814e-01 -4.24485922e-01 4.79104370e-01
-4.45991665e-01 -9.85305309e-01 -3.64507645e-01 -1.15081620e+00
7.82555118e-02 6.55682325e-01 -9.31136250e-01 -1.05480244e-03
-5.84486187e-01 -1.18059707e+00 7.23772645e-01 3.49310488e-01
-2.03321785e-01 -1.01029253e+00 -7.92376041e-01 4.42659140e-01
-1.50264040e-01 -1.09329867e+00 -5.85716106e-02 4.36908543e-01
-8.41581881e-01 -1.44834042e+00 -2.97101468e-01 -7.49354541e-01
6.05607152e-01 -5.41457951e-01 1.31828475e+00 1.58994183e-01
-2.18407467e-01 -2.63647698e-02 -5.13650358e-01 -8.22959185e-01
-1.01906931e+00 -1.97019335e-02 -1.20565951e-01 -3.68122846e-01
8.79058957e-01 -4.84257229e-02 -4.51504201e-01 -1.31563723e-01
-1.06176662e+00 -3.72524619e-01 6.35305822e-01 7.90532351e-01
4.89568830e-01 -2.67084688e-01 8.00718069e-01 -1.69577003e+00
9.48271155e-01 -8.74081910e-01 -1.66203991e-01 1.31958470e-01
-9.42921102e-01 1.40004233e-01 4.22693878e-01 -2.89090395e-01
-8.23111355e-01 8.95245150e-02 -4.97030348e-01 1.88742191e-01
-2.42836148e-01 1.08796453e+00 3.16427350e-01 5.73720872e-01
7.97310233e-01 -2.10355848e-01 -7.76089495e-03 -3.39149684e-01
-1.31182492e-01 8.73252451e-01 4.58666921e-01 -6.43579960e-02
-9.21769738e-02 3.96660328e-01 5.53403795e-03 -5.08503973e-01
-9.34674919e-01 -8.16289544e-01 -4.28001642e-01 2.95962483e-01
5.93333662e-01 -5.81782281e-01 -6.71418071e-01 -2.89937377e-01
-1.01311219e+00 1.83331683e-01 -3.08518261e-01 6.92296088e-01
4.39163670e-02 4.04770046e-01 -7.60801136e-01 -4.95424956e-01
-5.18884361e-01 -8.07871521e-01 7.69229114e-01 -2.08896756e-01
-8.67027640e-01 -1.13345015e+00 3.91950160e-01 -2.50831276e-01
-1.52426690e-01 5.11028290e-01 1.11421132e+00 -1.42744076e+00
4.82582897e-01 -3.56476247e-01 1.10567465e-01 -1.88367635e-01
3.17232460e-01 8.14742893e-02 -5.53181946e-01 4.52273060e-03
2.15885833e-01 -2.62608565e-02 1.02141249e+00 2.69955307e-01
7.77035594e-01 -7.00647771e-01 -8.00115466e-01 4.41480689e-02
1.29835534e+00 4.26267326e-01 4.06956613e-01 3.84532034e-01
-1.73533902e-01 5.45446575e-01 6.07910752e-01 6.22747719e-01
1.09822966e-01 1.98175222e-01 1.52796596e-01 -1.65542066e-02
2.13318691e-01 1.44725323e-01 1.52043715e-01 4.66129422e-01
-3.48713402e-05 -4.95581478e-01 -1.16859591e+00 8.47114444e-01
-1.70139873e+00 -8.80297184e-01 -6.00299120e-01 1.67327726e+00
1.37605274e+00 4.43348438e-01 1.99187528e-02 2.56685317e-01
5.93611300e-01 -3.34499627e-01 -1.38949215e-01 -8.60630631e-01
-7.65443593e-02 4.68722552e-01 3.15898478e-01 2.59650469e-01
-1.02904952e+00 5.85241377e-01 7.23066473e+00 3.44677597e-01
-1.30636179e+00 7.05750659e-02 5.95066309e-01 -4.65687737e-02
6.80049211e-02 -2.73093224e-01 -8.61579359e-01 2.76440293e-01
1.34774482e+00 -2.65432477e-01 -8.19621265e-01 6.97546780e-01
4.31033671e-01 -3.85134727e-01 -1.22135901e+00 4.09847945e-01
1.63047940e-01 -1.62017977e+00 -1.76906120e-02 7.09172785e-02
5.02026796e-01 -4.31617238e-02 -3.12539428e-01 5.97620793e-02
2.16572449e-01 -7.51724720e-01 5.48438489e-01 4.00676429e-01
7.37951398e-01 -4.34841722e-01 1.30208290e+00 2.06553787e-01
-6.34671390e-01 1.21959485e-01 -1.53070018e-02 5.64419664e-02
2.82812983e-01 8.42848182e-01 -1.84023190e+00 5.63463986e-01
5.74612916e-01 7.84631670e-01 -4.87574369e-01 9.06788945e-01
-2.84553498e-01 1.01312852e+00 -1.28606468e-01 -1.93244100e-01
1.27223536e-01 3.83078158e-01 5.93860865e-01 2.02015901e+00
1.52607188e-01 7.48282850e-01 1.43512219e-01 6.42002165e-01
3.77960429e-02 5.42020202e-01 -4.47554141e-01 -1.10946126e-01
1.05484478e-01 1.09817624e+00 -9.83721852e-01 -9.68961716e-01
-3.67208123e-01 5.60026824e-01 -9.52085257e-02 -2.07522988e-01
-5.57043314e-01 -3.35109711e-01 8.09105784e-02 3.84894282e-01
3.60431105e-01 5.59723914e-01 -3.56353521e-01 -1.03824854e+00
-1.07755288e-01 -8.73812377e-01 1.15199304e+00 -4.92336333e-01
-1.26765800e+00 8.17162693e-01 -3.37071009e-02 -1.03231096e+00
-5.90294242e-01 -7.67201185e-01 -3.75194490e-01 7.03514516e-01
-1.41133988e+00 -6.09143674e-01 1.10114679e-01 4.19022292e-01
9.40693095e-02 -4.32457551e-02 9.90101755e-01 2.77743548e-01
-4.21873748e-01 5.56185246e-01 -4.32618618e-01 3.75523776e-01
1.01329386e+00 -1.22940528e+00 8.87370482e-02 6.78396583e-01
-2.54424304e-01 1.01969540e+00 9.84294951e-01 -1.04158831e+00
-7.60916650e-01 -1.14218366e+00 1.73281348e+00 -3.41034234e-01
7.92643249e-01 2.26720676e-01 -1.15354061e+00 7.66134560e-01
3.17809373e-01 -1.52410150e-01 1.15204608e+00 4.75855358e-02
-2.26016805e-01 4.27949727e-01 -1.42075086e+00 3.05221856e-01
5.53217649e-01 -8.89640450e-02 -1.33559406e+00 5.39790988e-01
3.61430436e-01 -5.16353369e-01 -1.17176867e+00 5.43268263e-01
4.41371173e-01 -6.92937076e-01 5.76306760e-01 -1.11850321e+00
5.53555489e-01 -2.14343309e-01 9.72190574e-02 -9.30281222e-01
4.92512621e-02 -6.02289736e-01 -8.99479091e-02 7.51209319e-01
1.10418487e+00 -7.18688190e-01 7.44128704e-01 4.24400568e-01
-2.16771498e-01 -9.30346131e-01 -7.40643799e-01 -2.90329337e-01
-5.32866307e-02 -1.54245049e-01 2.92970657e-01 1.29357159e+00
9.58410978e-01 5.89173019e-01 2.92031378e-01 1.25007078e-01
-4.37747687e-02 2.33373195e-01 8.73083714e-03 -1.33446860e+00
-1.57424256e-01 -2.86922634e-01 -4.81320560e-01 -2.80023456e-01
-3.09012253e-02 -1.14337170e+00 -1.83493733e-01 -1.64527178e+00
3.64959240e-01 -2.00101286e-01 -3.67438853e-01 1.06155610e+00
-7.61852190e-02 1.33846566e-01 -4.64162320e-01 8.88012722e-02
-3.27661932e-01 -3.06711733e-01 6.37991548e-01 -1.12309724e-01
-3.81149411e-01 1.82537846e-02 -8.20901573e-01 1.01382160e+00
8.41881931e-01 -8.92959118e-01 1.92191169e-01 3.36604565e-01
6.22373879e-01 2.18678743e-01 1.24517523e-01 -3.92835081e-01
2.31484532e-01 -2.22533390e-01 1.55348316e-01 -9.93427992e-01
-1.73168108e-01 -5.23244679e-01 -1.24940358e-01 1.07632673e+00
-4.91178304e-01 4.87872571e-01 3.72139275e-01 3.95768344e-01
-4.01197940e-01 -4.89725232e-01 5.08648574e-01 -3.86731893e-01
-5.52354634e-01 -4.12869543e-01 -7.84841120e-01 1.61036059e-01
9.91281450e-01 3.05495765e-02 -5.43193519e-01 4.36042100e-02
-1.23573411e+00 -2.74729393e-02 4.00161669e-02 6.57957420e-02
7.73962736e-01 -6.31594598e-01 -1.05217791e+00 -1.02638341e-01
2.62820691e-01 -4.90599900e-01 -1.21886283e-01 1.30385387e+00
-6.30757570e-01 7.21385121e-01 2.24728763e-01 -3.48372012e-01
-1.79120243e+00 6.73404455e-01 2.11213738e-01 -6.64281487e-01
-5.88183701e-01 5.48700035e-01 -2.38879129e-01 -8.16950500e-02
-1.62840173e-01 -1.09313893e+00 -6.09195292e-01 2.67629743e-01
7.74979889e-01 1.20592393e-01 7.15539515e-01 -5.81814051e-01
-8.59854639e-01 -8.93475711e-02 -4.89305079e-01 -1.36782795e-01
1.40342903e+00 3.13835442e-01 -6.91005826e-01 4.30669904e-01
8.87491643e-01 5.66432714e-01 1.64056659e-01 -6.69311434e-02
4.25835282e-01 8.27088505e-02 -4.86164056e-02 -1.32392859e+00
-6.97434604e-01 2.43502080e-01 1.82795569e-01 4.15008485e-01
8.71171892e-01 2.27493986e-01 3.22973520e-01 4.34022874e-01
-1.87161997e-01 -7.83846200e-01 -4.47298020e-01 4.56302136e-01
7.22422957e-01 -1.12256587e+00 3.71396154e-01 -6.49717987e-01
-7.05239236e-01 1.30005670e+00 2.03037575e-01 2.15217739e-01
9.17845905e-01 7.54946530e-01 1.46697268e-01 -7.03170419e-01
-1.26951015e+00 6.39689341e-02 3.20993394e-01 1.58813074e-01
1.07274568e+00 6.31144568e-02 -1.34847856e+00 8.61472726e-01
-1.13594845e-01 5.11832297e-01 9.89428997e-01 1.37805724e+00
-3.00332099e-01 -9.79095161e-01 -4.19382691e-01 8.51835608e-01
-1.49870670e+00 -4.40784484e-01 -8.98719430e-01 1.04452455e+00
4.56046641e-01 1.05875146e+00 -1.37411952e-01 -1.49871618e-01
5.39601862e-01 1.99474186e-01 1.60036489e-01 -9.62271810e-01
-1.15807974e+00 2.43123874e-01 6.54330492e-01 -3.26665640e-01
-7.76466906e-01 -9.65188265e-01 -1.65786529e+00 1.27863646e-01
-4.92605716e-01 6.88427627e-01 2.68828958e-01 8.44154954e-01
4.01885390e-01 6.38326883e-01 -2.23609492e-01 5.03826976e-01
-3.89075071e-01 -9.27114606e-01 -4.91127640e-01 1.80694059e-01
5.46063483e-01 -4.31662887e-01 -2.02634782e-01 4.18383509e-01] | [8.444260597229004, 8.817804336547852] |
5f634655-abb9-4482-bf1f-5518cdf553ad | building-sequential-inference-models-for-end | 1812.00686 | null | http://arxiv.org/abs/1812.00686v1 | http://arxiv.org/pdf/1812.00686v1.pdf | Building Sequential Inference Models for End-to-End Response Selection | This paper presents an end-to-end response selection model for Track 1 of the
7th Dialogue System Technology Challenges (DSTC7). This task focuses on
selecting the correct next utterance from a set of candidates given a partial
conversation. We propose an end-to-end neural network based on enhanced
sequential inference model (ESIM) for this task. Our proposed model differs
from the original ESIM model in the following four aspects. First, a new word
representation method which combines the general pre-trained word embeddings
with those estimated on the task-specific training set is adopted in order to
address the challenge of out-of-vocabulary (OOV) words. Second, an attentive
hierarchical recurrent encoder (AHRE) is designed which is capable to encode
sentences hierarchically and generate more descriptive representations by
aggregation. Third, a new pooling method which combines multi-dimensional
pooling and last-state pooling is used instead of the simple combination of max
pooling and average pooling in the original ESIM. Last, a modification layer is
added before the softmax layer to emphasize the importance of the last
utterance in the context for response selection. In the released evaluation
results of DSTC7, our proposed method ranked second on the Ubuntu dataset and
third on the Advising dataset in subtask 1 of Track 1. | ['Yu-Ping Ruan', 'Zhen-Hua Ling', 'Jia-Chen Gu', 'Quan Liu'] | 2018-12-03 | null | null | null | null | ['conversational-response-selection'] | ['natural-language-processing'] | [ 3.19176883e-01 4.17858630e-01 4.83647585e-02 -7.74476707e-01
-7.17237592e-01 -1.30615115e-01 7.04721212e-01 1.33393453e-02
-8.34960043e-01 6.30256057e-01 8.84286940e-01 -8.19549188e-02
6.52418509e-02 -4.82114464e-01 -1.27473166e-02 -3.75888109e-01
2.01622874e-01 4.18427378e-01 3.18589538e-01 -7.33225346e-01
5.59419930e-01 1.24365725e-01 -1.29699838e+00 7.85647333e-01
7.17290461e-01 9.69276190e-01 6.05517745e-01 9.34417725e-01
-2.05058843e-01 6.88054025e-01 -7.24460781e-01 -2.25185156e-01
-1.29528090e-01 -5.94125748e-01 -1.20785713e+00 7.39987269e-02
2.10683107e-01 -5.62257230e-01 -4.50689435e-01 5.30594051e-01
1.01321054e+00 7.87916541e-01 5.06754637e-01 -7.93345451e-01
-4.10388529e-01 8.57728124e-01 6.65651485e-02 7.78699294e-03
5.38332283e-01 6.33865893e-02 1.18098104e+00 -1.19898403e+00
4.82897341e-01 1.51206505e+00 2.14621976e-01 8.20182741e-01
-8.75996232e-01 -2.30500326e-01 3.11005890e-01 3.08116287e-01
-1.00935686e+00 -7.12366283e-01 5.89975059e-01 -8.71059969e-02
1.67700565e+00 3.38544041e-01 1.79642782e-01 1.09135187e+00
4.43903394e-02 1.13761008e+00 7.32967138e-01 -5.74278355e-01
1.34095728e-01 3.28899622e-01 5.23463786e-01 3.94068092e-01
-6.73892200e-01 -3.55336279e-01 -6.80578172e-01 -3.00455213e-01
2.81708091e-01 -2.58589268e-01 -3.83570045e-01 2.41082385e-01
-9.41208839e-01 1.09635055e+00 1.59715384e-01 3.78120691e-01
-6.08713984e-01 -2.82298803e-01 8.20875823e-01 3.22552562e-01
4.81206357e-01 5.74891746e-01 -5.58565855e-01 -3.27820629e-01
-6.99056149e-01 3.24943483e-01 9.86674428e-01 7.14616537e-01
4.96096969e-01 5.28451130e-02 -8.94020975e-01 1.30558133e+00
2.80848861e-01 -1.63695619e-01 7.81045496e-01 -8.35496724e-01
7.01400578e-01 5.64819038e-01 2.20952719e-01 -6.06207907e-01
-5.95596373e-01 -2.84060210e-01 -7.57615745e-01 -1.49414316e-01
8.09904113e-02 -4.92832333e-01 -5.21325707e-01 1.69462919e+00
1.31150767e-01 -1.07337043e-01 4.85577792e-01 9.10386682e-01
1.24115622e+00 9.82681572e-01 3.64434570e-02 -3.14139187e-01
1.47233248e+00 -1.31781209e+00 -1.05513871e+00 -1.97162583e-01
5.80748618e-01 -6.74506247e-01 1.07075834e+00 2.28679314e-01
-1.13939190e+00 -8.15845430e-01 -1.15685403e+00 -3.28065574e-01
-5.39699137e-01 1.97684333e-01 1.75577879e-01 2.62533486e-01
-1.03214753e+00 3.33810687e-01 -1.98972329e-01 -4.21221942e-01
-4.36936229e-01 3.65789562e-01 -2.39048660e-01 1.85432285e-01
-1.84266806e+00 1.24239624e+00 5.79344571e-01 3.20212424e-01
-2.60019839e-01 -1.79541022e-01 -1.01318479e+00 2.85902619e-01
2.80157179e-01 -3.37394446e-01 1.42166889e+00 -5.49819052e-01
-2.34641719e+00 3.60483408e-01 -3.26779336e-01 -5.89217246e-01
1.40829638e-01 -2.51329303e-01 -2.45281920e-01 1.84180632e-01
-3.47728074e-01 6.54579043e-01 6.01664901e-01 -7.88151264e-01
-7.42166519e-01 -6.63215816e-02 2.30235085e-01 7.03752816e-01
-4.26145524e-01 1.85868517e-01 -4.23895597e-01 -4.07939225e-01
-1.03831716e-01 -7.87258983e-01 -2.38293141e-01 -7.93411076e-01
-2.80709803e-01 -9.74257588e-01 7.75511682e-01 -7.21059740e-01
1.64009202e+00 -2.02165484e+00 4.81444150e-01 -1.52374133e-01
-1.52345210e-01 5.75486541e-01 -1.82298437e-01 9.24576819e-01
2.05950648e-01 -8.33071843e-02 3.37477252e-02 -7.99815714e-01
2.24474087e-01 2.29425319e-02 -2.52094030e-01 4.83272485e-02
2.67031014e-01 5.56448877e-01 -7.72390604e-01 -3.30162078e-01
3.47233891e-01 3.46253812e-01 -5.25686502e-01 6.46241367e-01
-1.40808016e-01 2.51520485e-01 -2.35114023e-01 -5.66206016e-02
4.54430163e-01 1.95542067e-01 1.27699688e-01 -1.16938807e-01
-2.80438125e-01 7.22441018e-01 -1.07167554e+00 1.73275304e+00
-9.01855886e-01 2.70078480e-01 -7.13464012e-03 -7.03183889e-01
1.14610016e+00 7.20734239e-01 9.62603539e-02 -6.49939835e-01
1.04000062e-01 1.23572670e-01 -2.51473505e-02 -6.85942054e-01
1.13307643e+00 2.57598013e-01 -5.67370594e-01 4.79076117e-01
3.94125611e-01 -2.03726798e-01 7.50031695e-02 2.38770425e-01
8.39162230e-01 8.19397792e-02 3.29257637e-01 -1.03019670e-01
9.10260797e-01 -4.84349787e-01 6.07096016e-01 7.82700598e-01
-1.80906236e-01 6.82723284e-01 5.11217713e-01 -2.51642883e-01
-6.70830309e-01 -4.02340651e-01 1.57905579e-01 1.53103948e+00
-9.71458703e-02 -5.38180947e-01 -6.28934741e-01 -7.08959758e-01
-3.34064543e-01 1.08010614e+00 -4.69654888e-01 -1.77737892e-01
-7.02580512e-01 -4.06663090e-01 5.14824152e-01 3.94309193e-01
6.41572654e-01 -1.39804411e+00 -5.74789822e-01 4.73333299e-01
-4.90897715e-01 -1.16231298e+00 -7.95333207e-01 1.79528579e-01
-3.93488586e-01 -5.97135842e-01 -6.43838286e-01 -8.39031816e-01
2.04264104e-01 8.20451677e-02 8.10327888e-01 -1.01355212e-02
3.16378623e-01 8.73689651e-02 -8.26384485e-01 -1.68384641e-01
-3.99419188e-01 3.97898346e-01 -3.05529893e-03 4.00919586e-01
4.69183534e-01 1.16632633e-01 -5.11463642e-01 2.44044870e-01
-5.76391578e-01 1.63731292e-01 3.82554144e-01 1.27047288e+00
-6.19500354e-02 -3.08987647e-01 1.02579761e+00 -7.99496055e-01
1.26242173e+00 -3.99716616e-01 -1.50295019e-01 4.61507171e-01
-3.86367679e-01 1.12380810e-01 6.98169053e-01 -2.77059644e-01
-1.42584229e+00 -9.39724892e-02 -6.96357250e-01 2.34692097e-01
-7.05055073e-02 5.11648834e-01 -1.33626997e-01 5.01836002e-01
2.77892858e-01 2.89998084e-01 5.00650667e-02 -5.96685886e-01
3.63811821e-01 1.41506946e+00 2.28760481e-01 -2.81483501e-01
-5.17494194e-02 -3.72676939e-01 -6.87049747e-01 -1.01791787e+00
-5.55145264e-01 -5.92303932e-01 -6.13043547e-01 -2.04529345e-01
9.92836297e-01 -7.01960683e-01 -9.00378883e-01 6.22599483e-01
-1.54745531e+00 -3.07334512e-01 7.60732517e-02 7.11846948e-01
-5.50187409e-01 4.56916779e-01 -8.16744030e-01 -1.23348439e+00
-7.85156190e-01 -1.33092523e+00 9.28429484e-01 3.96324635e-01
-5.87070882e-01 -7.12509334e-01 -1.42164469e-01 4.14451450e-01
6.49044216e-01 -3.22270423e-01 8.83682966e-01 -1.17780054e+00
1.61961570e-01 -3.44089895e-01 -4.05180715e-02 5.99157989e-01
-1.58427190e-02 -2.67196536e-01 -9.67350900e-01 -4.07155216e-01
1.68586016e-01 -5.75246692e-01 1.01099217e+00 1.21143669e-01
8.26474607e-01 -2.98808604e-01 9.70008001e-02 -6.75066859e-02
1.03302896e+00 4.47473943e-01 5.08951783e-01 8.91922116e-02
3.19218248e-01 1.05914724e+00 7.92952895e-01 5.19525826e-01
6.35847390e-01 1.05720437e+00 1.29796997e-01 1.35186970e-01
6.50949702e-02 -1.09093741e-01 5.44982076e-01 1.03778207e+00
3.12846810e-01 -5.40071249e-01 -4.16782886e-01 5.00644088e-01
-2.07149673e+00 -7.32406497e-01 9.98290703e-02 2.24662614e+00
9.25516248e-01 3.40663046e-01 1.19915850e-01 -1.07855283e-01
7.02450275e-01 3.88752937e-01 -2.69922674e-01 -1.16890061e+00
5.87495863e-02 2.77258873e-01 -3.64325903e-02 1.09653223e+00
-9.31275249e-01 1.01950502e+00 5.16943359e+00 7.82038271e-01
-9.45035338e-01 9.91517976e-02 3.86679322e-01 1.19411215e-01
4.00253804e-03 -1.62129983e-01 -1.12376213e+00 4.68413413e-01
1.18013752e+00 -9.92946178e-02 1.13870785e-01 5.15951276e-01
4.36512411e-01 -2.35041887e-01 -8.56930375e-01 4.71016675e-01
2.77368665e-01 -1.04768527e+00 -1.08044863e-01 -3.03119361e-01
1.77369416e-01 -2.82569408e-01 -1.29286230e-01 8.55588496e-01
5.80054522e-02 -8.34079087e-01 2.35071391e-01 5.59520423e-01
5.15377164e-01 -8.89563560e-01 1.24202895e+00 4.58724558e-01
-9.06735778e-01 -1.93653300e-01 -4.10681099e-01 -1.42945692e-01
3.42070699e-01 7.36757321e-03 -1.17427933e+00 6.38120830e-01
2.44690865e-01 2.00444236e-01 -2.00173169e-01 6.90063238e-01
-2.60602325e-01 4.23647523e-01 -1.41178697e-01 -4.41195220e-01
5.18576920e-01 1.23124972e-01 5.36351860e-01 1.60266685e+00
2.12612282e-02 2.16018066e-01 2.49018893e-01 2.44822145e-01
-2.75944062e-02 2.91979432e-01 -2.62516767e-01 2.45435312e-01
6.10219300e-01 1.28894556e+00 -1.53644383e-01 -4.56108928e-01
-3.32983851e-01 1.14721167e+00 2.58881211e-01 2.93693632e-01
-4.67428744e-01 -8.33200157e-01 3.79107445e-01 -4.94446546e-01
3.36090446e-01 -1.26657844e-01 2.63232030e-02 -8.05735230e-01
-1.16759390e-01 -8.03201497e-01 2.91299790e-01 -4.40565914e-01
-1.05946577e+00 1.21316016e+00 5.84659204e-02 -1.04177129e+00
-7.67281592e-01 -2.85922408e-01 -7.59228885e-01 1.27335906e+00
-1.28049064e+00 -9.42471623e-01 5.41137457e-02 2.61231691e-01
1.39968228e+00 -3.97422642e-01 1.37651646e+00 3.43885385e-02
-7.52130806e-01 7.89931417e-01 -1.25456005e-01 5.59469052e-02
9.74178672e-01 -1.28175437e+00 3.59406203e-01 6.24424756e-01
-5.07955372e-01 6.45125151e-01 6.89849257e-01 -3.35548133e-01
-9.90946531e-01 -7.54950941e-01 1.55761421e+00 -8.35999474e-03
2.90626228e-01 -6.01929188e-01 -1.02309477e+00 3.39575917e-01
9.55575645e-01 -5.66585660e-01 6.33346379e-01 5.02613038e-02
4.93175872e-02 8.73486623e-02 -9.97963488e-01 6.84720695e-01
3.32198977e-01 -5.83885014e-01 -8.79258037e-01 2.16236874e-01
1.09409261e+00 -5.24800539e-01 -9.19847727e-01 3.22081804e-01
5.07566631e-01 -6.66167498e-01 6.39468670e-01 -5.87505162e-01
3.66855323e-01 -4.54582181e-03 -2.75086850e-01 -1.46317697e+00
-2.61980385e-01 -9.45548594e-01 -1.79595053e-01 1.33578992e+00
5.60325503e-01 -4.93717879e-01 4.21698660e-01 6.18276894e-01
-3.46464753e-01 -1.05877757e+00 -9.45743918e-01 -2.69361556e-01
-1.21643096e-01 1.42747499e-02 4.59486306e-01 4.93263274e-01
4.16332662e-01 1.01216316e+00 -7.73924410e-01 -1.69567406e-01
8.46294090e-02 -1.08266808e-01 6.01283491e-01 -8.12272489e-01
-3.27323824e-01 -3.72837007e-01 3.96791473e-02 -1.56431031e+00
3.06272209e-01 -5.18546045e-01 4.00758952e-01 -1.64029884e+00
-2.93714464e-01 1.00002393e-01 -3.54218423e-01 1.33846208e-01
-3.98331136e-01 -1.21153407e-01 4.78210390e-01 -2.40423039e-01
-6.10155702e-01 8.52720559e-01 8.63070488e-01 -1.31032601e-01
-5.23255110e-01 3.97404522e-01 -6.18173420e-01 3.13631684e-01
7.96250701e-01 -8.18954334e-02 -4.12329674e-01 -2.46378422e-01
-3.41252774e-01 6.21640682e-01 -1.88740745e-01 -5.21704555e-01
5.38233340e-01 7.01018125e-02 -4.41773981e-02 -8.68450046e-01
6.98278666e-01 -4.83782738e-01 -6.78930759e-01 2.95971513e-01
-9.22056615e-01 4.69419397e-02 4.70572598e-02 3.02971423e-01
-3.17976743e-01 -6.90186381e-01 5.43343246e-01 -8.76448378e-02
-8.93176496e-01 -2.40868449e-01 -8.45205486e-01 -3.21953535e-01
5.23312986e-01 -5.81175238e-02 -2.52903402e-01 -6.88642919e-01
-8.38272452e-01 6.21981680e-01 -3.24944258e-01 8.82341504e-01
9.31097865e-01 -1.10748243e+00 -1.16197324e+00 3.73753207e-03
2.34512970e-01 -2.30612352e-01 4.91030723e-01 6.82180882e-01
1.87785663e-02 5.55475235e-01 -2.87134089e-02 -5.03187478e-02
-1.37405956e+00 1.68540806e-01 2.28952855e-01 -6.57080054e-01
-3.97613257e-01 9.83921468e-01 -1.71825767e-01 -6.54422879e-01
6.31543875e-01 -1.00775205e-01 -8.62452626e-01 3.65860909e-01
7.20545113e-01 2.47940063e-01 9.32385996e-02 -6.51861370e-01
-2.04390615e-01 -1.58408862e-02 -5.05965650e-01 -3.31104100e-01
1.19239521e+00 -3.82411003e-01 -1.17854670e-01 5.97373962e-01
1.39382195e+00 -1.79913625e-01 -9.85220969e-01 -5.90491891e-01
7.40448684e-02 -8.48819241e-02 6.93082064e-03 -8.82243276e-01
-5.27320147e-01 9.12807822e-01 3.61729503e-01 2.91390181e-01
9.07503247e-01 -4.81253505e-01 9.83904660e-01 5.69875479e-01
2.24970654e-02 -1.55351400e+00 1.09330446e-01 1.22237945e+00
1.35290110e+00 -1.11189771e+00 -2.28745878e-01 -3.06403469e-02
-1.11579585e+00 1.38140845e+00 8.79459620e-01 3.81690674e-02
3.99533182e-01 -3.77897061e-02 -5.52942045e-02 5.97603209e-02
-1.08528709e+00 -1.53785810e-01 1.79531798e-01 1.19420655e-01
7.06445634e-01 -1.25917748e-01 -7.50090837e-01 8.03127348e-01
-8.82763881e-03 -3.08426946e-01 5.52381098e-01 7.54790723e-01
-4.91768539e-01 -1.17228448e+00 1.24982655e-01 2.03443766e-01
-2.64564663e-01 -1.56469300e-01 -5.44317484e-01 2.86426544e-01
-1.79018766e-01 1.48689747e+00 2.36353967e-02 -8.87874961e-01
6.94684625e-01 5.05705237e-01 -4.16220762e-02 -9.23321366e-01
-1.10021091e+00 3.21553871e-02 5.81268013e-01 -2.40018666e-01
-2.15567768e-01 -3.54586691e-01 -1.18297112e+00 7.15200603e-02
-5.07235825e-01 4.57868159e-01 7.01890469e-01 1.03999174e+00
3.96754533e-01 6.51404619e-01 1.02366436e+00 -1.00981569e+00
-9.09540832e-01 -1.69230199e+00 -2.37247601e-01 2.15599403e-01
3.90891701e-01 -4.97117877e-01 -2.46841609e-01 -4.69913304e-01] | [12.637394905090332, 7.813736915588379] |
d645b311-a7e7-403b-b889-c8b82d045cf0 | blind-face-restoration-via-deep-multi-scale | 2008.00418 | null | https://arxiv.org/abs/2008.00418v1 | https://arxiv.org/pdf/2008.00418v1.pdf | Blind Face Restoration via Deep Multi-scale Component Dictionaries | Recent reference-based face restoration methods have received considerable attention due to their great capability in recovering high-frequency details on real low-quality images. However, most of these methods require a high-quality reference image of the same identity, making them only applicable in limited scenes. To address this issue, this paper suggests a deep face dictionary network (termed as DFDNet) to guide the restoration process of degraded observations. To begin with, we use K-means to generate deep dictionaries for perceptually significant face components (\ie, left/right eyes, nose and mouth) from high-quality images. Next, with the degraded input, we match and select the most similar component features from their corresponding dictionaries and transfer the high-quality details to the input via the proposed dictionary feature transfer (DFT) block. In particular, component AdaIN is leveraged to eliminate the style diversity between the input and dictionary features (\eg, illumination), and a confidence score is proposed to adaptively fuse the dictionary feature to the input. Finally, multi-scale dictionaries are adopted in a progressive manner to enable the coarse-to-fine restoration. Experiments show that our proposed method can achieve plausible performance in both quantitative and qualitative evaluation, and more importantly, can generate realistic and promising results on real degraded images without requiring an identity-belonging reference. The source code and models are available at \url{https://github.com/csxmli2016/DFDNet}. | ['WangMeng Zuo', 'Shangchen Zhou', 'Xianhui Lin', 'Lei Zhang', 'Xiaoming Li', 'Chaofeng Chen'] | 2020-08-02 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/806_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123540375.pdf | eccv-2020-8 | ['blind-face-restoration'] | ['computer-vision'] | [ 1.48222670e-01 -3.25119168e-01 5.94855361e-02 -3.81143779e-01
-7.83149898e-01 -2.87239343e-01 4.11000937e-01 -5.34320712e-01
1.35175928e-01 5.38327694e-01 5.03111899e-01 1.90569371e-01
-2.23608688e-01 -7.08509922e-01 -5.22820115e-01 -1.01855302e+00
3.09824228e-01 -2.58740336e-01 -3.73249441e-01 -2.88953990e-01
1.31782234e-01 6.09045208e-01 -1.76471007e+00 1.94180191e-01
1.04360509e+00 1.25486636e+00 2.42346764e-01 8.80028829e-02
2.16564178e-01 2.93676078e-01 -4.19252306e-01 -4.41548258e-01
4.85976577e-01 -6.53411150e-01 -2.74127245e-01 3.09813410e-01
6.19831979e-01 -6.00141883e-01 -7.21937478e-01 1.25946534e+00
8.74487042e-01 2.42903233e-01 5.89661777e-01 -1.02722788e+00
-1.10280550e+00 -7.56539553e-02 -6.47116184e-01 2.48615861e-01
5.82576275e-01 2.43790805e-01 6.44801795e-01 -1.47742772e+00
3.95601511e-01 1.33568919e+00 4.22576904e-01 6.01257145e-01
-1.14226055e+00 -1.04004633e+00 -1.02895340e-02 3.66653413e-01
-1.62873304e+00 -1.12176478e+00 1.08754003e+00 -2.03217030e-01
4.29102808e-01 8.41681361e-02 5.12692630e-01 1.04607618e+00
1.06593080e-01 4.39494222e-01 1.27518392e+00 -2.65606225e-01
2.92877685e-02 -5.63782528e-02 -2.95987815e-01 6.40309393e-01
1.02580823e-01 3.63271147e-01 -6.20563328e-01 1.49037307e-02
1.01429260e+00 1.56592175e-01 -7.97468662e-01 -5.90217933e-02
-8.40035260e-01 7.34508097e-01 5.64125121e-01 3.20531547e-01
-6.47808850e-01 -3.19208801e-01 7.95068964e-02 2.78037161e-01
3.93378109e-01 -1.09730721e-01 -7.70704215e-03 3.36632758e-01
-7.28741467e-01 -7.23129734e-02 3.35146189e-01 8.07635903e-01
7.73593664e-01 3.30596566e-01 -2.20127702e-01 1.33734035e+00
4.64181006e-01 6.82359993e-01 3.62911135e-01 -1.17685080e+00
2.21236780e-01 2.20586196e-01 7.98007473e-02 -1.30343163e+00
1.11813970e-01 -4.77412730e-01 -1.22696757e+00 3.43397230e-01
-1.05591249e-02 2.12333128e-01 -8.85064781e-01 1.66212845e+00
4.96883005e-01 3.43097150e-01 9.67764948e-03 1.22732532e+00
9.76195872e-01 6.32012308e-01 -1.87379882e-01 -5.93559742e-01
1.26948512e+00 -6.05142593e-01 -8.52589846e-01 -1.11455552e-01
-3.54971468e-01 -1.05044913e+00 1.14479923e+00 4.82404947e-01
-1.10247326e+00 -1.02098227e+00 -9.39200759e-01 -1.35035917e-01
2.46665433e-01 3.40780675e-01 2.27806643e-01 4.59235042e-01
-1.16629434e+00 5.23121059e-01 -5.27004063e-01 -2.38948718e-01
6.30160093e-01 2.79622942e-01 -4.96365100e-01 -7.14236915e-01
-1.01371288e+00 5.47116935e-01 -7.63433948e-02 4.09933448e-01
-1.14272749e+00 -4.03661728e-01 -9.56306338e-01 -1.58497885e-01
7.72981569e-02 -6.36676669e-01 6.47808850e-01 -9.47565079e-01
-1.51165748e+00 7.65398920e-01 -2.80085146e-01 2.53905237e-01
9.29441378e-02 -1.54588893e-01 -8.11103404e-01 3.93879831e-01
1.49918795e-01 4.57281619e-01 1.38427770e+00 -1.56502676e+00
-4.71986264e-01 -4.13577080e-01 1.86208449e-02 2.82401025e-01
-5.47917008e-01 1.04726836e-01 -6.27806187e-01 -1.01875794e+00
3.16067368e-01 -4.80491638e-01 2.69485414e-01 3.45520586e-01
-1.68675169e-01 3.13551761e-02 6.77992761e-01 -1.10894012e+00
1.06180310e+00 -2.58807182e+00 2.53167897e-01 8.00840631e-02
1.16276592e-01 2.44768336e-01 -4.20714021e-01 1.70566589e-01
-1.63564160e-01 -2.79986739e-01 -3.41020882e-01 -3.91752332e-01
-1.43085137e-01 1.97393537e-01 -2.88133472e-01 7.36011147e-01
2.04208329e-01 4.77254659e-01 -7.36249864e-01 -3.56079012e-01
3.47070128e-01 9.85138237e-01 -5.76759040e-01 4.24261838e-01
3.37571293e-01 7.62706876e-01 -2.67955691e-01 8.70151341e-01
1.00246930e+00 -4.16002050e-02 -4.82509285e-02 -7.69336104e-01
1.32689595e-01 -1.08672597e-01 -1.37382555e+00 1.52880120e+00
-5.09726346e-01 2.26434350e-01 5.07063806e-01 -8.31338704e-01
1.01774776e+00 3.30358326e-01 4.06842291e-01 -9.39566970e-01
1.18723042e-01 2.30818748e-01 -1.11852191e-01 -3.32025439e-01
5.12914360e-02 -2.59860784e-01 3.93648237e-01 2.20526695e-01
1.19098961e-01 -7.70874089e-03 -1.48939565e-01 -1.18735701e-01
6.38005257e-01 3.72321606e-02 1.98396057e-01 -1.06685698e-01
9.22799170e-01 -6.03446782e-01 7.69418001e-01 8.82227197e-02
-2.12365299e-01 8.71934295e-01 -2.20334351e-01 -6.90429807e-02
-8.45491111e-01 -1.23389077e+00 -3.16542625e-01 8.01883996e-01
3.65227133e-01 -9.86610278e-02 -6.48447692e-01 -3.28330606e-01
-3.14316858e-04 3.49656552e-01 -4.71398592e-01 -3.92399341e-01
-5.37102103e-01 -4.42127794e-01 8.05994645e-02 2.06187546e-01
7.13985443e-01 -1.13420534e+00 9.03766826e-02 9.94953290e-02
-3.39204699e-01 -8.55641842e-01 -8.42284858e-01 -4.44954425e-01
-6.76888406e-01 -1.04096806e+00 -8.25638235e-01 -1.01521659e+00
9.44250524e-01 6.26662135e-01 7.42599666e-01 2.77109236e-01
-2.75302291e-01 3.96127790e-01 -3.02627116e-01 2.81865060e-01
-2.19644949e-01 -7.23816931e-01 4.01414007e-01 5.26279867e-01
1.03928573e-01 -8.80166888e-01 -1.06995487e+00 5.26406109e-01
-9.08603847e-01 -2.05821767e-01 6.31046176e-01 1.07525396e+00
8.54692340e-01 3.28103989e-01 7.20945835e-01 -2.00208545e-01
5.68687677e-01 -4.32951570e-01 -3.39513421e-01 4.09689955e-02
-4.66290265e-01 -4.80086625e-01 8.85288656e-01 -5.54814517e-01
-1.13107061e+00 -2.07084045e-01 -3.31000805e-01 -7.94205725e-01
-1.49180353e-01 2.15076268e-01 -7.30663896e-01 -3.09830576e-01
4.87765372e-01 5.20699620e-01 5.54465353e-02 -6.62351906e-01
3.03551286e-01 7.34357238e-01 7.77093709e-01 -5.23860395e-01
1.06832302e+00 6.24195814e-01 -2.91584313e-01 -6.48787975e-01
-7.19350934e-01 -2.83885241e-01 -4.04006153e-01 -2.92900175e-01
5.52980244e-01 -1.19729280e+00 -5.39644063e-01 6.39530361e-01
-7.61973023e-01 -8.19772705e-02 -2.56090462e-01 5.12413025e-01
-4.02818531e-01 4.72490877e-01 -7.40714312e-01 -5.25626957e-01
-3.99848133e-01 -1.13721442e+00 9.91409421e-01 5.95771611e-01
2.88878888e-01 -6.04802012e-01 -3.25113088e-01 5.32451749e-01
4.67625767e-01 8.37322250e-02 7.46745229e-01 1.32108722e-02
-4.73498434e-01 6.04110472e-02 -3.88728648e-01 8.31293285e-01
7.28109360e-01 -2.93492645e-01 -1.05636370e+00 -8.12165201e-01
3.61180544e-01 -1.99314654e-01 6.23033166e-01 3.39939654e-01
1.11501706e+00 -3.72329146e-01 -7.39176497e-02 9.10323918e-01
1.35248327e+00 2.11833373e-01 7.13421047e-01 -7.13216737e-02
7.56387353e-01 4.97797400e-01 7.83008158e-01 5.27771890e-01
3.38954687e-01 6.86177433e-01 3.56569290e-01 -1.78064317e-01
-7.81757772e-01 -2.43589729e-01 5.18415451e-01 1.02425468e+00
-3.31385508e-02 1.62542313e-02 -3.65344197e-01 6.15971208e-01
-1.20116115e+00 -9.07859802e-01 2.71240771e-01 2.26078343e+00
9.30680633e-01 -1.80373847e-01 -2.24997476e-01 2.13278800e-01
8.64885747e-01 1.72188088e-01 -7.33950615e-01 2.09605172e-01
-3.15911621e-01 3.66389871e-01 -1.76108912e-01 4.53268558e-01
-8.49255085e-01 7.47610390e-01 5.12133169e+00 1.07587492e+00
-1.15863669e+00 3.42057675e-01 5.97219646e-01 -3.88289578e-02
-3.08274478e-01 -1.85379386e-01 -5.44222236e-01 5.50429404e-01
4.95416850e-01 -1.92116126e-01 8.49627316e-01 5.02969503e-01
4.36939538e-01 1.20477550e-01 -7.11995423e-01 1.23509097e+00
3.37559998e-01 -9.03132081e-01 1.95366904e-01 1.42003000e-01
7.59540558e-01 -4.70242411e-01 3.20489079e-01 8.10079947e-02
5.51470332e-02 -1.11361921e+00 5.53928316e-01 6.96635664e-01
1.30603552e+00 -8.84068847e-01 4.76044297e-01 -2.06675585e-02
-1.27730143e+00 -2.31270134e-01 -5.73214293e-01 3.02960545e-01
8.53890255e-02 6.18475497e-01 -1.41522259e-01 6.99121118e-01
9.03720558e-01 8.27469289e-01 -3.45978290e-01 9.09328043e-01
-4.04869199e-01 4.54018682e-01 2.21139416e-02 8.26011837e-01
-4.03870344e-01 -4.21806335e-01 5.11454105e-01 6.99587286e-01
5.69856644e-01 4.62821215e-01 2.41802916e-01 6.08346462e-01
-2.03305885e-01 9.71375927e-02 -3.58710647e-01 3.62757534e-01
6.42447650e-01 1.40363789e+00 -2.55313277e-01 -1.87810913e-01
-4.66453642e-01 1.22497404e+00 5.26480936e-02 4.76633132e-01
-7.08734512e-01 -2.01892748e-01 1.08775985e+00 1.01746969e-01
3.73889387e-01 -2.40857564e-02 1.23036075e-02 -1.10716164e+00
1.87349021e-01 -1.34881818e+00 3.26771647e-01 -8.54034722e-01
-1.63923979e+00 8.31206381e-01 -3.41830492e-01 -1.50801921e+00
3.05734780e-02 -3.19792747e-01 -4.54951733e-01 1.20404935e+00
-1.65437198e+00 -1.29474998e+00 -6.58083558e-01 1.16639376e+00
4.68237430e-01 -2.34858736e-01 6.70364797e-01 7.95934439e-01
-6.45211458e-01 7.89886177e-01 1.67418554e-01 1.36527345e-01
9.99138653e-01 -7.22975254e-01 6.12137243e-02 1.03370988e+00
-1.21539533e-01 8.02383006e-01 5.36809146e-01 -6.40349209e-01
-1.64733851e+00 -1.20411646e+00 3.21809620e-01 -2.28528231e-02
2.49086529e-01 -1.31005570e-01 -1.08665287e+00 2.41386786e-01
2.84428243e-02 4.80001658e-01 5.60751379e-01 -4.70726103e-01
-4.17197198e-01 -4.88180876e-01 -1.38114774e+00 4.82551873e-01
1.26826799e+00 -8.19636285e-01 -4.21211898e-01 2.17524976e-01
4.32639271e-01 -3.10848981e-01 -1.13240516e+00 4.89576906e-01
5.27688980e-01 -1.01563990e+00 1.18637204e+00 1.99824665e-02
3.91097933e-01 -7.25865304e-01 -4.73278940e-01 -1.30023861e+00
-5.35172701e-01 -4.61207002e-01 -2.52352953e-01 1.54699242e+00
-2.87446827e-01 -6.92069590e-01 3.08600068e-01 1.08219713e-01
-2.50245541e-01 -7.30872035e-01 -9.45961773e-01 -6.26324952e-01
-1.89185262e-01 2.72006169e-02 6.98671460e-01 9.48921680e-01
-6.13528371e-01 9.96779874e-02 -5.45319319e-01 4.93665487e-01
9.05369937e-01 3.53649825e-01 6.32824779e-01 -9.95414913e-01
-1.99648738e-01 -2.11597294e-01 -1.86795428e-01 -1.09440303e+00
1.59932539e-01 -8.43316317e-01 8.75267610e-02 -1.54714882e+00
1.06188521e-01 -4.22874242e-01 -3.79836470e-01 3.77316743e-01
-3.59568834e-01 7.71587849e-01 5.44779114e-02 5.07176161e-01
-1.04367070e-01 9.87820566e-01 1.60676920e+00 -1.01066716e-01
4.32963967e-02 -1.88843474e-01 -1.08042872e+00 4.97320920e-01
5.40173471e-01 -2.45173380e-01 -3.61681849e-01 -4.61318761e-01
-4.34086889e-01 2.07399517e-01 5.40311217e-01 -1.02535403e+00
1.10140711e-01 -2.10680783e-01 7.88919032e-01 -1.54400021e-01
6.46880805e-01 -7.61891901e-01 3.70938540e-01 1.94864348e-01
1.27598003e-01 -2.60581881e-01 5.12795150e-02 5.60656369e-01
-4.30511773e-01 1.55410677e-01 1.21986246e+00 1.07694738e-01
-5.76467991e-01 6.94985867e-01 1.85456246e-01 -2.40342796e-01
7.45266795e-01 -3.63888502e-01 -2.70907909e-01 -4.56848383e-01
-6.87272847e-01 -1.32738233e-01 7.73737073e-01 4.78316307e-01
9.84535515e-01 -1.60048282e+00 -9.93668795e-01 6.07151508e-01
1.14278972e-01 -2.25639641e-01 7.11768687e-01 6.92076206e-01
-1.09627627e-01 -1.17253266e-01 -3.94501388e-01 -4.47149575e-01
-1.14199150e+00 7.19434083e-01 4.14478451e-01 3.08307946e-01
-7.38721013e-01 7.43612468e-01 4.20812100e-01 -1.32725179e-01
1.48147807e-01 2.08036140e-01 -3.11043531e-01 7.56636411e-02
8.24828744e-01 1.77421957e-01 3.76658551e-02 -1.20765340e+00
-3.25408071e-01 8.96159232e-01 2.92679314e-02 5.40344156e-02
1.39613986e+00 -5.34689188e-01 -2.52647102e-01 -1.53669029e-01
1.28230679e+00 3.39389026e-01 -1.51430833e+00 -4.37780380e-01
-6.14710629e-01 -1.00493705e+00 1.96454242e-01 -6.38834417e-01
-1.59986591e+00 7.34852612e-01 9.99749422e-01 -2.11195514e-01
1.78792012e+00 -9.20582041e-02 8.45388114e-01 -2.81225711e-01
4.01887983e-01 -6.23123109e-01 4.02484775e-01 1.02313377e-01
1.34610999e+00 -1.12445605e+00 9.64461342e-02 -4.73759085e-01
-3.66535127e-01 8.88659835e-01 7.02212989e-01 -1.27823755e-01
6.30044699e-01 -4.39793244e-02 1.79733530e-01 -1.42339751e-01
-2.79286802e-01 -1.59239516e-01 4.03084457e-01 7.99103975e-01
2.04796270e-01 -1.47070631e-01 -1.03493780e-01 5.79464555e-01
-1.52715549e-01 -4.35373783e-02 2.22060233e-01 5.58855832e-01
-2.03175873e-01 -9.96808589e-01 -6.66347384e-01 4.53620762e-01
-2.95070887e-01 -1.72417477e-01 8.64171684e-02 2.74215251e-01
3.22084665e-01 1.36724126e+00 -2.42887944e-01 -4.60341096e-01
4.83693868e-01 -4.15561527e-01 4.53317881e-01 -5.48197806e-01
-3.17466795e-01 3.37720275e-01 -2.76905149e-01 -6.18589103e-01
-3.88813674e-01 -5.74507952e-01 -1.03225672e+00 -3.99783790e-01
-2.30323464e-01 -1.77841280e-02 2.54332215e-01 6.11737072e-01
4.77405787e-01 3.52954417e-01 1.18230140e+00 -1.06895983e+00
-3.56685340e-01 -9.38740492e-01 -7.54598737e-01 5.70549607e-01
5.03882527e-01 -8.57613385e-01 -6.35924339e-01 1.83208942e-01] | [12.820121765136719, -0.02125396952033043] |
43b236fd-c60c-4faf-b0bf-d8d4c930ce10 | nnunet-raspp-for-retinal-oct-fluid-detection | 2302.13195 | null | https://arxiv.org/abs/2302.13195v1 | https://arxiv.org/pdf/2302.13195v1.pdf | nnUNet RASPP for Retinal OCT Fluid Detection, Segmentation and Generalisation over Variations of Data Sources | Retinal Optical Coherence Tomography (OCT), a noninvasive cross-sectional scan of the eye with qualitative 3D visualization of the retinal anatomy is use to study the retinal structure and the presence of pathogens. The advent of the retinal OCT has transformed ophthalmology and it is currently paramount for the diagnosis, monitoring and treatment of many eye pathogens including Macular Edema which impairs vision severely or Glaucoma that can cause irreversible blindness. However the quality of retinal OCT images varies among device manufacturers. Deep Learning methods have had their success in the medical image segmentation community but it is still not clear if the level of success can be generalised across OCT images collected from different device vendors. In this work we propose two variants of the nnUNet [8]. The standard nnUNet and an enhanced vision call nnUnet_RASPP (nnU-Net with residual and Atrous Spatial Pyramid Pooling) both of which are robust and generalise with consistent high performance across images from multiple device vendors. The algorithm was validated on the MICCAI 2017 RETOUCH challenge dataset [1] acquired from 3 device vendors across 3 medical centers from patients suffering from 2 retinal disease types. Experimental results show that our algorithms outperform the current state-of-the-arts algorithms by a clear margin for segmentation obtaining a mean Dice Score (DS) of 82.3% for the 3 retinal fluids scoring 84.0%, 80.0%, 83.0% for Intraretinal Fluid (IRF), Subretinal Fluid (SRF), and Pigment Epithelium Detachments (PED) respectively on the testing dataset. Also we obtained a perfect Area Under the Curve (AUC) score of 100% for the detection of the presence of fluid for all 3 fluid classes on the testing dataset. | ['Yongmin Li', 'Zidong Wang', 'Alina Miron', 'Nchongmaje Ndipenoch'] | 2023-02-25 | null | null | null | null | ['anatomy'] | ['miscellaneous'] | [ 7.21824169e-02 -2.00290307e-01 3.91358018e-01 7.86707699e-02
-2.74295419e-01 -5.25076509e-01 1.46646649e-01 -1.96385667e-01
-3.65291297e-01 7.58334816e-01 1.03925928e-01 -5.91365814e-01
-2.43781552e-01 -2.67530531e-01 -2.95695394e-01 -5.79789937e-01
-2.56503195e-01 3.45527440e-01 5.30307591e-01 4.48057443e-01
6.00311697e-01 5.83151400e-01 -1.60971057e+00 5.02990246e-01
1.23121893e+00 1.33700943e+00 -1.03181139e-01 9.51041877e-01
1.59194291e-01 3.60841572e-01 -2.92155623e-01 1.49793299e-02
1.00535452e+00 -1.98867843e-01 -6.74829125e-01 2.74249971e-01
1.27135956e+00 -5.44468760e-01 3.17619920e-01 9.58658636e-01
7.91573226e-01 -4.32996750e-01 6.95876658e-01 -4.11782593e-01
-3.78502995e-01 -4.81709033e-01 -8.91431928e-01 6.02690876e-01
-2.19140485e-01 7.39547074e-01 2.15030134e-01 -5.87757170e-01
6.06712401e-01 8.92658293e-01 6.37249351e-01 1.28162324e-01
-9.23185110e-01 -5.50863624e-01 -5.41324437e-01 1.41059905e-01
-8.00895631e-01 -1.66833013e-01 -3.70464206e-01 -1.17864156e+00
1.18012333e+00 2.03122362e-01 1.23458374e+00 1.35704622e-01
5.29530466e-01 2.80584600e-02 2.04888868e+00 -2.22364798e-01
-4.39941324e-03 -2.68967226e-02 1.47599548e-01 5.95192611e-01
6.69032633e-01 4.46938694e-01 1.25487804e-01 -6.84119463e-02
9.73088443e-01 -3.74106079e-01 -5.19943774e-01 1.39761809e-02
-8.09786201e-01 4.51211214e-01 3.85590822e-01 -9.56377983e-02
-4.89304364e-01 -2.47651011e-01 4.33863044e-01 9.86899883e-02
3.70254248e-01 6.51102185e-01 -3.86330575e-01 -1.97325021e-01
-7.49449134e-01 1.29009262e-01 3.12954664e-01 3.20526391e-01
2.75067031e-01 -2.46110961e-01 -2.64434159e-01 8.05907965e-01
3.52263659e-01 3.54309440e-01 4.19450313e-01 -9.98789310e-01
-4.98164892e-02 8.63830328e-01 4.93521273e-01 -2.03666702e-01
-2.56889105e-01 -5.24555922e-01 -6.95558786e-01 9.41791117e-01
4.46449876e-01 -5.19156337e-01 -1.63317251e+00 5.93997002e-01
3.00310165e-01 5.48281074e-01 -3.03910255e-01 1.18578184e+00
7.52223730e-01 1.24013029e-01 -2.27067322e-01 -2.85494655e-01
1.30935037e+00 -6.22104228e-01 -3.21970403e-01 -6.19556531e-02
4.15985465e-01 -1.29829597e+00 5.40896952e-01 7.63120770e-01
-1.12891316e+00 -4.16125000e-01 -9.05989289e-01 6.01056144e-02
2.63951626e-02 4.54619050e-01 4.75930184e-01 7.58946002e-01
-1.47716284e+00 3.39250565e-01 -8.52040052e-01 -5.72377741e-01
1.02397001e+00 7.10700572e-01 -3.84196252e-01 -2.93149531e-01
-4.06146087e-02 9.90530729e-01 -1.37249112e-01 1.57827333e-01
-2.92149603e-01 -9.74524677e-01 -2.20612958e-01 -6.22656703e-01
-1.57173827e-01 -1.13265419e+00 7.49557912e-01 -5.74680150e-01
-1.36292124e+00 1.27481377e+00 -5.36112070e-01 -7.54212618e-01
5.49593329e-01 -4.20109034e-01 -1.48343846e-01 3.20342809e-01
-7.80100329e-03 5.36403954e-01 8.43218386e-01 -9.44626153e-01
-1.08718669e+00 -6.17829740e-01 -2.80147165e-01 -2.53836989e-01
6.46055400e-01 3.87350708e-01 -1.25900224e-01 4.68182862e-02
5.72832115e-02 -1.02905488e+00 -1.24033153e-01 1.32685304e-01
-4.83684719e-01 -7.69117801e-03 5.86263359e-01 -7.59703577e-01
8.27551842e-01 -1.78480625e+00 -4.67322856e-01 4.11303435e-03
7.36833870e-01 1.08430481e+00 2.89479382e-02 -2.79792577e-01
-1.90325603e-01 2.88114607e-01 -1.12034395e-01 1.42107084e-01
-7.80121446e-01 -1.69299588e-01 8.13993365e-02 6.77347600e-01
3.45535100e-01 7.75392711e-01 -6.45040691e-01 -7.84076080e-02
3.88875932e-01 6.37247920e-01 -3.21176529e-01 1.29954554e-02
-1.55263990e-02 6.14086807e-01 -4.40784656e-02 7.19623625e-01
8.89979661e-01 -3.83484751e-01 -1.43883675e-01 -2.52265930e-01
-6.01487994e-01 -2.72318814e-02 -1.05718267e+00 7.46093512e-01
1.50692791e-01 1.03800273e+00 -8.22978392e-02 -2.66050994e-01
6.20011091e-01 1.66489497e-01 3.40963751e-01 -5.23208678e-01
4.31364536e-01 5.42597055e-01 8.07513773e-01 -8.79810691e-01
-9.94671509e-02 -1.69621512e-01 1.05951309e+00 2.23535031e-01
-4.25372422e-01 2.82095522e-01 4.39414568e-02 -3.45021248e-01
9.76430595e-01 -2.90960148e-02 3.23877960e-01 -1.02855444e-01
2.95705944e-01 4.44453359e-02 2.48388574e-01 3.44052464e-01
-6.65196002e-01 9.31014717e-01 5.11945069e-01 -5.27700722e-01
-1.10634828e+00 -1.09658849e+00 -6.96964920e-01 -5.02171040e-01
-8.54152441e-02 5.56610674e-02 -6.10012591e-01 -2.38184571e-01
1.18498720e-01 -3.58242005e-01 -4.79431510e-01 5.76648176e-01
-3.27867299e-01 -9.78711426e-01 -1.97159406e-02 1.37677565e-01
7.53388226e-01 -5.61525285e-01 -1.11269760e+00 -7.52183869e-02
2.70059288e-01 -1.16917408e+00 6.46940991e-02 -7.31792867e-01
-1.09721851e+00 -1.59503508e+00 -7.09340334e-01 -7.35387862e-01
6.19031250e-01 1.30113378e-01 9.21037197e-01 -3.85338478e-02
-1.15466273e+00 1.35884643e-01 -4.65049297e-02 -6.28689051e-01
-8.46747681e-02 -7.11809993e-01 -6.66610375e-02 3.32346141e-01
5.80020368e-01 -4.97248650e-01 -1.27564371e+00 1.84271663e-01
-2.47750491e-01 -1.86748609e-01 9.10875857e-01 5.01024127e-01
7.68500924e-01 -1.29428595e-01 1.43034205e-01 -8.23128402e-01
6.45319343e-01 -1.74504876e-01 -9.73561943e-01 -5.64561151e-02
-8.24555933e-01 -5.94488800e-01 -2.22923726e-01 -1.90160036e-01
-3.80674183e-01 -2.79560477e-01 5.81930220e-01 -5.61173320e-01
-4.58525747e-01 2.06469357e-01 5.81786096e-01 -5.67077875e-01
8.75272214e-01 -1.45113915e-01 4.55286086e-01 -4.14401919e-01
-6.94739074e-02 1.04313087e+00 3.79334629e-01 -7.69455358e-02
2.47531399e-01 8.02535355e-01 3.68404835e-01 -8.28823090e-01
-4.97646689e-01 -9.50052917e-01 -3.00355494e-01 -2.33280510e-01
1.08057404e+00 -7.40368187e-01 -9.74797308e-01 9.43322361e-01
-1.04926085e+00 -2.12505996e-01 -5.31071126e-02 6.40511036e-01
-1.06563263e-01 3.58684629e-01 -4.66630727e-01 -7.40124583e-01
-4.90676850e-01 -1.47031116e+00 5.86475849e-01 5.78797579e-01
-1.25002354e-01 -6.68771982e-01 -1.40092534e-03 6.76631749e-01
6.17842972e-01 6.01411283e-01 1.12172365e+00 -5.56468740e-02
-9.99614596e-01 -3.76879163e-02 -8.42052937e-01 8.20260823e-01
3.04810554e-01 4.91931915e-01 -9.32637215e-01 -1.43814415e-01
-1.40996054e-01 1.51503474e-01 7.40689039e-01 1.36536360e+00
6.85976565e-01 -5.46811186e-02 -2.30487332e-01 6.74906373e-01
1.81101441e+00 3.81571352e-01 1.27936482e+00 3.50430161e-01
3.98352355e-01 7.83213735e-01 3.33373338e-01 2.17812389e-01
-1.16580479e-01 3.05898488e-01 5.12955725e-01 -2.66857117e-01
-7.33712971e-01 7.10857213e-01 -1.85245693e-01 5.25212400e-02
-5.95881224e-01 1.42825544e-01 -1.11516297e+00 9.68119442e-01
-1.37910473e+00 -6.58076406e-01 -8.07786822e-01 2.47747707e+00
6.46741033e-01 -1.66584328e-01 1.89545766e-01 -2.74281263e-01
6.93522453e-01 -7.57692158e-01 -7.05729544e-01 -3.41679990e-01
-1.40418962e-01 8.68248284e-01 6.92563236e-01 5.32850266e-01
-1.04342306e+00 6.77608311e-01 5.85456181e+00 8.93537030e-02
-1.56468344e+00 -4.62624878e-02 5.49162805e-01 -5.75216830e-01
5.11720717e-01 -2.62423139e-03 -8.55619371e-01 5.38506567e-01
8.09245706e-01 2.79273719e-01 2.46540278e-01 -2.00389370e-01
7.22051322e-01 -5.89460313e-01 -6.17852747e-01 1.12873900e+00
-3.01441997e-01 -1.61545396e+00 1.08871289e-01 6.48290277e-01
9.19387579e-01 8.05710077e-01 2.74571240e-01 -6.33731663e-01
-1.53968766e-01 -1.46194017e+00 -3.34713429e-01 8.34154069e-01
1.13159668e+00 -2.69653946e-01 9.70426321e-01 -4.97969925e-01
-4.93434221e-01 8.08015838e-02 -1.38074607e-01 -5.84727004e-02
-9.02488232e-02 5.80338359e-01 -1.26050723e+00 2.08999962e-02
9.08023477e-01 9.67094719e-01 -7.36165583e-01 2.27240896e+00
1.31433785e-01 6.51979029e-01 -2.55340219e-01 4.25213605e-01
2.19330356e-01 -5.47819734e-01 1.10579205e+00 5.76813042e-01
4.38535422e-01 3.40885855e-02 -3.35898101e-01 8.69748056e-01
3.38007361e-01 1.31373718e-01 -3.15534264e-01 1.30975291e-01
3.00283879e-01 1.10802448e+00 -4.08969611e-01 7.14322627e-02
-5.09356856e-01 3.14329207e-01 -3.98806304e-01 6.72352433e-01
-9.15529728e-02 -1.81421474e-01 1.03977537e+00 9.07192647e-01
2.02169314e-01 7.97185078e-02 -6.41534567e-01 -5.07519603e-01
1.97279260e-01 -7.89091229e-01 3.91414016e-02 -9.90933001e-01
-1.29108691e+00 6.07544601e-01 -5.12434900e-01 -1.54424822e+00
1.08922921e-01 -1.17931736e+00 -4.35651481e-01 1.65414762e+00
-1.93417108e+00 -9.84625518e-01 -5.06198227e-01 4.44744349e-01
-8.94123837e-02 -5.64472139e-01 5.80112994e-01 8.05480257e-02
-4.21959996e-01 7.03504533e-02 -1.93130076e-02 2.58425400e-02
8.69054437e-01 -1.24980736e+00 1.27681494e-02 9.42655981e-01
-6.89008832e-01 9.67936754e-01 3.15923631e-01 -8.19572270e-01
-8.84113967e-01 -1.15601420e+00 4.39278930e-01 -5.47796845e-01
4.60705072e-01 7.18032658e-01 -8.35225105e-01 5.05340993e-01
3.52097452e-01 3.50346684e-01 8.17932904e-01 -1.58140615e-01
-1.96572587e-01 -5.32903597e-02 -1.26112020e+00 4.78048772e-01
3.97089899e-01 -4.13528234e-02 -3.11603069e-01 3.02657962e-01
1.47667736e-01 -5.69928586e-01 -1.02510238e+00 6.63929641e-01
7.13884234e-01 -1.59581482e+00 5.44306397e-01 -6.21444345e-01
4.15362209e-01 -5.34705520e-01 1.45707056e-01 -7.18559265e-01
6.97570369e-02 -1.00657153e+00 1.60683095e-01 4.97898340e-01
2.35747203e-01 -1.24755132e+00 7.43178606e-01 4.99759287e-01
-3.09266388e-01 -1.05830395e+00 -8.78521383e-01 -3.74344349e-01
-7.03394506e-03 -1.20503530e-01 7.24689141e-02 5.72871149e-01
-7.74240911e-01 6.57699779e-02 2.95944601e-01 4.43911880e-01
8.10815811e-01 9.51023251e-02 7.50691354e-01 -1.75721145e+00
-3.99396904e-02 -6.16563320e-01 -1.08269691e+00 -4.51953411e-01
-5.92114568e-01 -6.46745086e-01 -6.78963602e-01 -1.90050566e+00
2.14551046e-01 -3.82263958e-01 -1.09353796e-01 1.69486940e-01
8.16391483e-02 4.40494537e-01 -1.71463087e-01 4.73411292e-01
2.63881058e-01 -4.27534848e-01 1.73242450e+00 7.46408477e-02
-5.73528826e-01 1.68616474e-01 -6.59217715e-01 6.19083583e-01
7.28211820e-01 1.42792901e-02 -4.44268793e-01 -3.61869514e-01
1.83593273e-01 -2.63999671e-01 8.69610488e-01 -1.04790413e+00
6.98867366e-02 2.11623520e-01 3.19269657e-01 -4.35969204e-01
-1.83280203e-02 -4.32630897e-01 4.67336513e-02 6.05843425e-01
4.63637531e-01 -5.91999114e-01 5.42667687e-01 4.27853405e-01
-1.51195809e-01 3.38081777e-01 1.37706709e+00 -2.26038665e-01
-4.81161237e-01 5.84388256e-01 -1.55465096e-01 2.17897967e-01
1.00720549e+00 -8.95594180e-01 -1.00329828e+00 1.53705701e-01
-7.94268966e-01 5.58316782e-02 6.44796908e-01 -1.08335149e-02
8.33580494e-01 -4.13527101e-01 -1.13105059e+00 3.74727964e-01
1.38279170e-01 2.17511028e-01 2.99529493e-01 1.54822958e+00
-1.01773393e+00 6.14972413e-01 -5.77582359e-01 -1.08483839e+00
-1.62598693e+00 -4.25019175e-01 1.19162393e+00 2.42154479e-01
-6.64524615e-01 9.13541913e-01 -9.46652889e-02 4.65694278e-01
-1.69342570e-03 -6.71312988e-01 -2.20301539e-01 -1.01023905e-01
6.86630309e-01 6.59007609e-01 3.11907709e-01 -3.12473565e-01
1.20372530e-02 1.26156473e+00 -5.27393818e-01 3.33457321e-01
1.20346773e+00 -1.89612970e-01 -8.62832308e-01 1.22088216e-01
6.78354681e-01 -2.38584559e-02 -1.13392997e+00 -5.01383282e-03
-4.71082985e-01 -7.15464473e-01 4.61840808e-01 -1.50847733e+00
-8.74677300e-01 1.00496447e+00 1.55970395e+00 4.11078371e-02
1.16005039e+00 -2.98654735e-01 9.20498013e-01 -4.97397870e-01
2.52875507e-01 -3.34460109e-01 -4.49362636e-01 1.52976915e-01
6.08307302e-01 -1.31371665e+00 -7.37954723e-03 -7.12011814e-01
-3.19745153e-01 9.68714714e-01 5.27840793e-01 -3.37440223e-01
8.24780881e-01 8.90162680e-03 4.36960191e-01 -4.03426677e-01
-3.90612990e-01 -6.56511545e-01 7.99243808e-01 9.49233770e-01
4.37662870e-01 1.44687623e-01 -4.84672815e-01 7.28874803e-02
-1.07742481e-01 4.02048528e-01 9.01252568e-01 5.07715464e-01
-5.85157514e-01 -7.98902035e-01 -4.23560059e-03 1.14162147e+00
-7.87508428e-01 -2.90837139e-01 -4.25621837e-01 6.90706313e-01
6.62812591e-01 1.00120401e+00 2.09985569e-01 5.31408079e-02
8.51389393e-02 -2.41571099e-01 8.21127236e-01 -9.68578160e-01
-5.86260378e-01 3.42514396e-01 1.40645623e-01 -7.03003883e-01
-6.10965431e-01 -5.93098998e-01 -9.36735392e-01 -3.10796332e-02
1.28713936e-01 -4.83158976e-01 5.99929512e-01 6.00567341e-01
7.91782081e-01 1.99863270e-01 2.60464586e-02 -2.86173493e-01
4.50125001e-02 -1.15551293e+00 -8.76036227e-01 -8.89286548e-02
9.63948190e-01 -6.21993482e-01 -3.93038571e-01 2.69834965e-01] | [15.81814956665039, -3.997189998626709] |
91044228-cf70-4b18-946a-22a2872d5cb0 | cross-modal-ranking-with-soft-consistency-and | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Chenglong_Li_Cross-Modal_Ranking_with_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Chenglong_Li_Cross-Modal_Ranking_with_ECCV_2018_paper.pdf | Cross-Modal Ranking with Soft Consistency and Noisy Labels for Robust RGB-T Tracking | Due to the complementary benefits of visible (RGB) and thermal infrared (T) data, RGB-T object tracking attracts more and more attention recently for boosting the performance under adverse illumination conditions. Existing RGB-T tracking methods usually localize a target object with a bounding box, in which the trackers or detectors is often affected by the inclusion of background clutter. To address this problem, this paper presents a novel approach to suppress background effects for RGB-T tracking. Our approach relies on a novel cross-modal manifold ranking algorithm. First, we integrate the soft cross-modality consistency into the ranking model which allows the sparse inconsistency to account for the different properties between these two modalities. Second, we propose an optimal query learning method to handle label noises of queries. In particular, we introduce an intermediate variable to represent the optimal labels, and formulate it as a $l_1$-optimization based sparse learning problem. Moreover, we propose a single unified optimization algorithm to solve the proposed model with stable and efficient convergence behavior. Finally, the ranking results are incorporated into the patch-based object features to address the background effects, and the structured SVM is then adopted to perform RGB-T tracking. Extensive experiments suggest that the proposed approach performs well against the state-of-the-art methods on large-scale benchmark datasets. | ['Yan Huang', 'Chengli Zhu', 'Liang Wang', 'Jin Tang', 'Chenglong Li'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['rgb-t-tracking'] | ['computer-vision'] | [ 1.09543651e-01 -6.08508170e-01 -2.62029678e-01 -2.69998997e-01
-8.82784307e-01 -2.22918004e-01 1.86118677e-01 -2.19797164e-01
-2.91318804e-01 3.29237819e-01 -1.84369653e-01 1.26646578e-01
-1.14302583e-01 -4.76896346e-01 -6.31538749e-01 -1.06066704e+00
5.22893488e-01 -5.86129874e-02 5.16549349e-01 8.19281414e-02
-3.18249017e-02 2.66756237e-01 -1.54829860e+00 -2.20596977e-02
1.09251451e+00 1.55841088e+00 2.37250045e-01 -6.10427372e-03
-3.17181349e-01 6.34781837e-01 -4.02624607e-01 -2.20210031e-01
4.19252872e-01 -2.64303803e-01 -7.16760680e-02 2.86810994e-01
6.70013309e-01 -1.38344422e-01 -3.31501305e-01 1.25031114e+00
5.54453135e-01 1.65421799e-01 3.53986979e-01 -1.39752626e+00
-4.11678106e-01 -9.40107107e-02 -9.74754512e-01 1.18405312e-01
1.28165588e-01 2.74894476e-01 7.03641057e-01 -9.89234507e-01
3.81044388e-01 1.17963052e+00 6.98950410e-01 2.05524534e-01
-1.12765002e+00 -8.11672568e-01 4.58933771e-01 2.32975662e-01
-1.53174245e+00 -3.59099716e-01 1.04412317e+00 -3.89041632e-01
3.28850925e-01 4.55172032e-01 7.56751060e-01 7.96147108e-01
1.02692209e-02 8.90969932e-01 1.21631956e+00 -2.72775471e-01
7.01785879e-03 4.00142044e-01 8.32033083e-02 8.00967216e-01
3.60301405e-01 8.72577913e-03 -4.42521334e-01 -1.68188766e-01
5.21993697e-01 3.16185564e-01 -2.77217537e-01 -6.76538467e-01
-1.21685886e+00 5.69639027e-01 7.37414956e-01 3.17168266e-01
-1.62591726e-01 1.58599615e-01 2.73508042e-01 -3.54038924e-01
5.21690428e-01 -2.78862745e-01 -1.90323845e-01 2.89449424e-01
-7.61778355e-01 -7.29481280e-02 2.43276268e-01 9.18633878e-01
6.94523752e-01 1.36386529e-01 -3.10179532e-01 8.28535438e-01
8.18019569e-01 9.70361948e-01 1.48098364e-01 -6.56815052e-01
6.33965075e-01 7.64789104e-01 2.13767081e-01 -1.37042904e+00
-1.93084627e-01 -8.33798110e-01 -8.34308803e-01 3.94614413e-02
3.52912784e-01 2.40158081e-01 -6.96229935e-01 1.61423504e+00
8.69152308e-01 3.04491609e-01 -1.70291215e-01 1.27954721e+00
9.81205642e-01 6.29807949e-01 5.48893213e-02 -6.01655245e-01
1.33084202e+00 -8.50590825e-01 -8.69455218e-01 -1.34148583e-01
3.44821155e-01 -8.83546829e-01 8.14891815e-01 1.72416165e-01
-6.55199885e-01 -6.70694888e-01 -8.44509244e-01 1.48987398e-01
-1.77606374e-01 5.80482960e-01 5.11280358e-01 7.28763819e-01
-5.76562762e-01 6.03998937e-02 -7.16253161e-01 -4.01692688e-01
2.56705046e-01 3.46759886e-01 -1.94222212e-01 -1.54553577e-01
-9.73557651e-01 8.00575852e-01 2.29593366e-01 6.93575919e-01
-5.51599920e-01 -4.27082330e-01 -6.92928374e-01 -3.56014699e-01
7.29227662e-01 -5.42172432e-01 7.08779931e-01 -7.75825202e-01
-1.36856318e+00 6.88179553e-01 -4.05038059e-01 2.23100245e-01
3.49380732e-01 -1.19433880e-01 -4.34435129e-01 1.19255736e-01
1.87703893e-01 1.74670205e-01 1.08913541e+00 -1.52808666e+00
-7.21984625e-01 -6.14410043e-01 -2.36790299e-01 2.81984478e-01
-5.41751623e-01 5.31727932e-02 -9.33139503e-01 -6.41181529e-01
6.44150913e-01 -9.35989141e-01 -1.08407661e-01 2.40522981e-01
-4.15669024e-01 -1.86555102e-01 9.77006137e-01 -6.38667047e-01
1.19866037e+00 -2.25538325e+00 2.02810898e-01 2.41480723e-01
7.52239302e-02 1.50955379e-01 1.58117726e-01 -1.89293846e-02
1.13842286e-01 -3.79149020e-01 -4.53846045e-02 -4.66254205e-01
-1.17868967e-01 3.28094736e-02 -7.98327401e-02 8.75810087e-01
-8.28309059e-02 7.10820615e-01 -7.02534258e-01 -1.00679123e+00
3.77001494e-01 5.12501776e-01 -1.67559952e-01 1.16177902e-01
-2.22054720e-01 7.48035431e-01 -8.69514763e-01 9.98886883e-01
8.62884998e-01 -3.13900381e-01 -2.54561633e-01 -8.72121096e-01
-3.76469880e-01 -3.45362544e-01 -1.50589609e+00 1.72182786e+00
-1.10425837e-01 1.05267361e-01 4.24649566e-01 -9.66508210e-01
1.01112533e+00 6.85026944e-02 8.42529655e-01 -7.75581717e-01
2.93968022e-01 4.03501868e-01 -4.13129985e-01 -4.82090026e-01
1.65850535e-01 -2.29123250e-01 1.18640304e-01 6.21245839e-02
-3.31121504e-01 2.08296984e-01 -2.20117807e-01 -3.02281454e-02
6.47444725e-01 3.49807501e-01 -1.16145067e-01 -1.07619999e-05
8.46165121e-01 1.49885803e-01 9.86944437e-01 4.93971527e-01
-2.40580961e-01 3.25831652e-01 -1.80168241e-01 -3.03822458e-01
-3.98664445e-01 -8.22425008e-01 -1.92829773e-01 9.08761621e-01
8.43524218e-01 -1.04647703e-01 -3.37481618e-01 -6.65550947e-01
2.34946430e-01 2.78597265e-01 -4.52538908e-01 -1.54803649e-01
-6.07846797e-01 -7.55542338e-01 4.64369170e-02 3.47971886e-01
7.32092381e-01 -5.97137094e-01 -4.47059482e-01 -2.04011686e-02
-3.07118356e-01 -9.94890153e-01 -5.35071015e-01 1.06241181e-01
-9.60421383e-01 -9.58477437e-01 -7.43476927e-01 -6.45770848e-01
6.03404701e-01 7.22626746e-01 4.80819702e-01 8.85394774e-03
-3.68483603e-01 7.33251810e-01 -3.34200799e-01 -2.69976743e-02
2.23012105e-01 -1.54755443e-01 1.58051997e-01 5.36627829e-01
1.62012264e-01 -9.33060646e-02 -7.02048600e-01 5.77412844e-01
-7.81273425e-01 -1.23142801e-01 6.82253718e-01 7.27876723e-01
1.02059495e+00 1.13459997e-01 9.10863951e-02 -1.88532636e-01
-5.58765195e-02 -2.56046474e-01 -9.13684607e-01 5.34361124e-01
-4.45132405e-01 -1.13893606e-01 4.09986883e-01 -5.66601872e-01
-1.16218758e+00 3.99385899e-01 2.45385706e-01 -8.94685209e-01
1.70782208e-01 3.62383097e-01 -5.47446728e-01 -6.90182447e-01
2.22069830e-01 5.53588390e-01 3.81865352e-03 -5.52843630e-01
3.17200154e-01 4.79859859e-01 2.79347718e-01 -6.04726911e-01
1.35343301e+00 7.43125916e-01 1.79560214e-01 -7.74287224e-01
-1.06850612e+00 -7.77410507e-01 -3.26687992e-01 -5.86293042e-01
8.68909717e-01 -9.29220796e-01 -8.62280130e-01 4.52771246e-01
-9.85239804e-01 2.14533687e-01 -5.74716441e-02 6.72819555e-01
-2.33521312e-01 5.77329040e-01 -4.60849285e-01 -1.12565184e+00
-2.01267138e-01 -1.21801805e+00 1.34125853e+00 5.17533600e-01
7.12574005e-01 -6.77045941e-01 -2.97648385e-02 5.65660059e-01
3.62623811e-01 2.12419316e-01 4.28010583e-01 -2.05087423e-01
-1.11509669e+00 -1.85774922e-01 -4.56219494e-01 2.47830436e-01
1.40442207e-01 -2.19792694e-01 -8.85667503e-01 -4.47694004e-01
2.23059639e-01 -9.38996077e-02 8.11957538e-01 3.31104875e-01
9.78409052e-01 -1.33353677e-02 -6.35960281e-01 8.36649299e-01
1.48411870e+00 1.67769670e-01 1.37989148e-01 5.14121175e-01
9.92371321e-01 4.77870822e-01 1.26068532e+00 4.59083647e-01
2.17708930e-01 9.94201660e-01 5.85851490e-01 -2.89443463e-01
1.81198463e-01 -8.04624185e-02 2.99723566e-01 8.92589569e-01
1.55734971e-01 1.27504811e-01 -6.32576048e-01 1.95255518e-01
-2.00441694e+00 -8.09940219e-01 -2.84637958e-01 2.37339187e+00
6.41712844e-01 2.81006767e-04 1.68871000e-01 -1.11007035e-01
8.50359559e-01 2.44617641e-01 -6.62951708e-01 6.63420260e-01
-1.97350994e-01 -2.78958499e-01 6.09606981e-01 8.23485851e-02
-1.24929404e+00 6.69028044e-01 5.18293858e+00 1.09498751e+00
-1.30841804e+00 2.88572311e-01 3.43072027e-01 -8.74996632e-02
-1.38800517e-01 5.66917509e-02 -1.07767785e+00 6.29331470e-01
1.65270716e-01 1.02340236e-01 1.50528267e-01 7.93627143e-01
2.65207618e-01 -1.64640188e-01 -6.49792135e-01 1.31889188e+00
2.56237268e-01 -7.87625790e-01 -2.43829533e-01 2.14730743e-02
3.81574869e-01 -2.86530644e-01 3.01722050e-01 1.92744687e-01
-3.60659480e-01 -5.43599188e-01 8.89995456e-01 7.49782205e-01
6.23841226e-01 -4.24149930e-01 5.85645020e-01 3.70998770e-01
-1.70975542e+00 -1.30210221e-01 -4.41641241e-01 4.43491757e-01
5.65130636e-02 5.83029151e-01 -1.24258339e-01 9.54036236e-01
8.12020719e-01 9.22404587e-01 -7.49820054e-01 1.36479807e+00
1.36474788e-01 2.66941965e-01 -5.22113264e-01 -4.23468314e-02
-3.06033697e-02 -4.62459177e-01 6.21362925e-01 8.90799701e-01
3.82745683e-01 2.29870796e-01 7.15523601e-01 8.20054948e-01
3.73592973e-01 3.12826633e-01 -3.11971903e-01 2.52649546e-01
4.25009638e-01 1.36702633e+00 -6.79367840e-01 -1.82946101e-01
-6.74509823e-01 7.95989513e-01 6.27294332e-02 4.76995736e-01
-1.14586008e+00 1.90987978e-02 3.07121068e-01 -6.43449184e-03
4.86614257e-01 -1.44155279e-01 3.09015010e-02 -1.38665843e+00
3.92720997e-01 -7.48368323e-01 3.51970583e-01 -8.03162396e-01
-1.33347285e+00 3.51169318e-01 -3.30955237e-02 -1.78075588e+00
5.47395885e-01 -5.42823732e-01 -2.41810739e-01 8.04089129e-01
-1.70368052e+00 -1.33956981e+00 -6.82275414e-01 7.78280377e-01
6.78716972e-02 4.08265628e-02 2.60123163e-01 6.59186900e-01
-9.97564793e-01 4.83718872e-01 2.82202363e-01 -2.53331382e-04
8.30039322e-01 -7.96579599e-01 -6.36288285e-01 9.24237370e-01
-6.12724125e-02 7.96257854e-01 5.94315588e-01 -7.51499891e-01
-1.94010115e+00 -1.16502786e+00 2.27631480e-01 -2.19292223e-01
5.19337475e-01 -9.66184363e-02 -7.68570364e-01 3.80362362e-01
-2.24483714e-01 5.52579522e-01 5.23138523e-01 -1.25168666e-01
-4.02728379e-01 -7.80469298e-01 -9.50766444e-01 4.86071892e-02
8.22644830e-01 -4.07282084e-01 -4.70409125e-01 4.40060377e-01
7.19619751e-01 -6.14566207e-01 -8.38439345e-01 5.54324508e-01
4.87770766e-01 -8.74086857e-01 1.14081347e+00 -6.45293370e-02
-2.33140528e-01 -1.01033258e+00 -2.40121022e-01 -7.54773080e-01
-2.28015438e-01 -2.38368690e-01 -1.33903354e-01 1.38368142e+00
-1.20287418e-01 -6.26057327e-01 8.43782365e-01 5.16634703e-01
-1.05592385e-01 -5.65674543e-01 -1.16407382e+00 -8.98096383e-01
-6.61733270e-01 -2.61666089e-01 1.38038054e-01 8.54096591e-01
-4.32117075e-01 1.15171261e-01 -5.49639344e-01 3.74570757e-01
1.18409860e+00 6.43016934e-01 8.08717191e-01 -1.19916296e+00
-3.17294896e-01 -1.33990750e-01 -2.89707869e-01 -1.28353727e+00
-3.70397232e-02 -8.25446010e-01 2.90035158e-01 -1.33594728e+00
3.60118091e-01 -7.98932672e-01 -5.01401544e-01 2.21883744e-01
-5.15379369e-01 2.09308803e-01 2.94811279e-01 4.43161160e-01
-1.14732373e+00 1.05062580e+00 1.18885338e+00 -2.94603705e-01
-1.09307773e-01 7.49737695e-02 -3.96486729e-01 5.11306226e-01
3.57728720e-01 -5.22987425e-01 -1.00383811e-01 -4.89927083e-01
-6.25649467e-02 1.00300021e-01 5.38701713e-01 -1.02026081e+00
5.32121480e-01 -2.57370561e-01 4.75313962e-01 -9.09901083e-01
5.78956068e-01 -1.43425333e+00 2.54721344e-01 4.44147736e-01
1.17488362e-01 -2.55181074e-01 1.15956709e-01 1.02892435e+00
-3.26677144e-01 1.20682120e-02 8.46301913e-01 1.67190358e-02
-6.48065388e-01 5.27870297e-01 2.49472141e-01 -1.30654126e-01
1.12094736e+00 -3.64639729e-01 -2.90270686e-01 3.44805233e-02
-4.93480593e-01 4.35898840e-01 5.32921672e-01 2.65628844e-01
5.15327752e-01 -1.59150469e+00 -3.29759747e-01 -1.01108644e-02
3.35021675e-01 -9.37662870e-02 3.39607567e-01 1.39856195e+00
-9.10107642e-02 2.89413989e-01 1.86260849e-01 -1.05779028e+00
-1.35911989e+00 6.55623078e-01 4.03637767e-01 -4.69715782e-02
-4.24657613e-01 7.89704978e-01 1.22540183e-01 -1.63107827e-01
5.09241700e-01 -1.70814306e-01 -1.71890810e-01 2.91693211e-02
2.78039634e-01 3.04319441e-01 -1.33302122e-01 -1.05752909e+00
-7.50011802e-01 1.11307693e+00 3.25610161e-01 1.25093639e-01
1.00508451e+00 -2.54537940e-01 -2.27142945e-01 5.60041964e-01
1.15005505e+00 3.40297148e-02 -1.10441124e+00 -5.89331925e-01
-9.44752321e-02 -7.70135045e-01 1.51486620e-01 -3.33732158e-01
-1.32337892e+00 6.77470863e-01 9.50230122e-01 -1.19000152e-02
1.26853931e+00 -1.22405976e-01 7.45759904e-01 2.26384327e-01
5.34512877e-01 -8.31666887e-01 2.48494059e-01 1.22444168e-01
4.93502051e-01 -1.42348325e+00 2.78904706e-01 -7.61443853e-01
-3.79803360e-01 8.39468360e-01 7.76760697e-01 1.03444971e-01
4.45006460e-01 -1.09707646e-01 8.28304142e-02 -6.28916621e-02
-1.99172661e-01 -4.76187676e-01 4.71735775e-01 3.68701339e-01
1.16730683e-01 -3.12017739e-01 -1.88460708e-01 4.75647897e-01
5.44839501e-01 -9.83399004e-02 -3.42467010e-01 9.11068678e-01
-5.73222101e-01 -9.32573140e-01 -9.59625542e-01 3.15498799e-01
-1.46132499e-01 1.72261745e-01 -4.89243492e-02 7.39745677e-01
4.56414878e-01 9.93143439e-01 -4.31734860e-01 -4.17975187e-01
4.98012960e-01 -4.84735146e-02 3.89711440e-01 -3.67175639e-01
-5.07190108e-01 7.17979372e-01 -4.13545877e-01 -8.13273132e-01
-8.79951358e-01 -7.24832892e-01 -9.65857804e-01 1.91300735e-01
-8.72842491e-01 1.87547043e-01 5.83465099e-01 8.00177455e-01
2.63169646e-01 4.46460783e-01 7.19300568e-01 -7.72121251e-01
-5.58961987e-01 -6.28371298e-01 -6.79184556e-01 4.58120525e-01
2.80388176e-01 -1.10015106e+00 -5.84517658e-01 -2.13785753e-01] | [6.378810882568359, -2.184385061264038] |
6c6a86bc-e8b4-4044-a951-d28f6ab4b304 | large-scale-semi-supervised-learning-via | 1912.02233 | null | https://arxiv.org/abs/1912.02233v1 | https://arxiv.org/pdf/1912.02233v1.pdf | Large-Scale Semi-Supervised Learning via Graph Structure Learning over High-Dense Points | We focus on developing a novel scalable graph-based semi-supervised learning (SSL) method for a small number of labeled data and a large amount of unlabeled data. Due to the lack of labeled data and the availability of large-scale unlabeled data, existing SSL methods usually encounter either suboptimal performance because of an improper graph or the high computational complexity of the large-scale optimization problem. In this paper, we propose to address both challenging problems by constructing a proper graph for graph-based SSL methods. Different from existing approaches, we simultaneously learn a small set of vertexes to characterize the high-dense regions of the input data and a graph to depict the relationships among these vertexes. A novel approach is then proposed to construct the graph of the input data from the learned graph of a small number of vertexes with some preferred properties. Without explicitly calculating the constructed graph of inputs, two transductive graph-based SSL approaches are presented with the computational complexity in linear with the number of input data. Extensive experiments on synthetic data and real datasets of varied sizes demonstrate that the proposed method is not only scalable for large-scale data, but also achieve good classification performance, especially for extremely small number of labels. | ['Tieyong Zeng', 'Raymond Chan', 'Zitong Wang', 'Li Wang'] | 2019-12-04 | null | null | null | null | ['graph-structure-learning'] | ['graphs'] | [ 1.36394918e-01 3.27527106e-01 -4.56866086e-01 -4.66912419e-01
-6.62729084e-01 -6.01560235e-01 2.96382844e-01 3.84220988e-01
-6.61560819e-02 7.34418094e-01 -2.11465627e-01 -2.55390465e-01
-2.01316461e-01 -9.20897007e-01 -6.34961784e-01 -7.12475061e-01
-2.01955773e-02 8.07711065e-01 2.79007435e-01 1.48383409e-01
-4.55684662e-02 3.86247039e-01 -1.22031176e+00 -2.16608763e-01
9.39436913e-01 9.38383758e-01 2.05911100e-02 2.73527503e-01
-2.86718756e-01 7.37529814e-01 -2.02605262e-01 -7.75787309e-02
4.20081198e-01 -6.25168622e-01 -8.08331788e-01 8.17874789e-01
3.81588608e-01 1.73356622e-01 -2.32725114e-01 1.29984891e+00
3.19437712e-01 1.24476075e-01 6.47579014e-01 -1.55081713e+00
-4.86077726e-01 6.05691493e-01 -7.36252308e-01 -8.61717984e-02
1.99286073e-01 -1.15818135e-01 1.14986002e+00 -9.24176276e-01
7.13901341e-01 1.12413895e+00 5.07480562e-01 2.82434016e-01
-1.33506262e+00 -4.49622959e-01 2.53033161e-01 -1.10989781e-02
-1.68328083e+00 -8.88017192e-02 1.18527198e+00 -4.38269675e-01
3.15208822e-01 5.59679754e-02 6.51429951e-01 5.05726993e-01
-3.79654318e-01 5.99682450e-01 1.32687390e+00 -4.75226134e-01
3.85040790e-01 2.53998250e-01 4.07206982e-01 1.23826897e+00
4.08216953e-01 -1.82012603e-01 -1.75840080e-01 -3.12631339e-01
7.13440180e-01 1.11641042e-01 -1.96277320e-01 -8.08164358e-01
-1.20675254e+00 9.60731983e-01 6.56581342e-01 -2.29821308e-03
-1.64087668e-01 -3.57146114e-01 2.59285688e-01 2.91440725e-01
6.01457000e-01 1.58124879e-01 -3.64493042e-01 4.99445856e-01
-6.71419144e-01 -2.79769123e-01 9.74639654e-01 1.12618637e+00
1.28116655e+00 7.50027075e-02 1.37846559e-01 7.98169732e-01
4.21822190e-01 4.35550272e-01 1.78334251e-01 -3.29146743e-01
8.20169210e-01 1.36377180e+00 -4.71208505e-02 -1.27574062e+00
-5.37794948e-01 -3.46966982e-01 -1.08474743e+00 -1.03391610e-01
5.86242199e-01 -2.11706564e-01 -9.57201362e-01 1.57301450e+00
7.08549261e-01 1.84095129e-01 2.26072054e-02 8.65269303e-01
8.92738581e-01 7.53730118e-01 -1.65261954e-01 -5.13654172e-01
9.89151776e-01 -1.13573098e+00 -6.01483703e-01 -3.70587051e-01
1.03915906e+00 -3.98432583e-01 1.12430358e+00 6.30819350e-02
-6.02143645e-01 -4.50319558e-01 -1.00740731e+00 9.02039856e-02
-3.41187298e-01 4.47581798e-01 6.77549124e-01 5.40596128e-01
-8.91159534e-01 3.94608200e-01 -4.66238588e-01 -4.32153374e-01
3.56321722e-01 5.46537817e-01 -4.88077015e-01 -4.57377791e-01
-8.07349503e-01 3.77778083e-01 7.43081391e-01 3.22880328e-01
-6.10359132e-01 -2.14469820e-01 -1.00844359e+00 -2.40610284e-03
7.28019595e-01 -2.54977405e-01 4.89617497e-01 -1.05666459e+00
-8.94441187e-01 9.48596716e-01 1.05988421e-01 3.46542336e-02
2.92189419e-01 2.65320271e-01 -2.52207756e-01 1.84616804e-01
1.44074604e-01 2.53749847e-01 7.37341046e-01 -1.36953974e+00
-3.84335786e-01 -5.64837933e-01 -2.56462097e-02 3.36624682e-01
-5.71350574e-01 -3.38392794e-01 -6.13525510e-01 -4.43776041e-01
4.13335115e-01 -1.10085523e+00 -4.67919499e-01 -3.97169292e-02
-6.54277563e-01 -3.35898757e-01 9.18744743e-01 -2.07554027e-01
1.13130105e+00 -2.11447167e+00 1.00762233e-01 4.61613804e-01
5.88401437e-01 4.39950883e-01 -3.61740559e-01 5.78294754e-01
-8.62784684e-02 -8.84341002e-02 -4.91007894e-01 -1.26182839e-01
-3.24566156e-01 3.31700653e-01 6.09224774e-02 7.38508701e-01
1.28519207e-01 9.05139387e-01 -1.28358781e+00 -1.02619755e+00
1.41665995e-01 -5.02319820e-02 -1.27766326e-01 4.56226468e-01
-1.52133211e-01 5.19968033e-01 -7.44356811e-01 8.01138282e-01
6.96809828e-01 -8.42525661e-01 5.15271366e-01 -2.59057730e-01
4.25383449e-01 -1.20393373e-01 -1.48908818e+00 1.39808750e+00
-1.39780417e-01 1.78900555e-01 1.01443857e-01 -1.47762048e+00
1.24276471e+00 1.84457034e-01 5.68011105e-01 -2.37341985e-01
2.19947696e-01 2.04691917e-01 -1.05681255e-01 -3.66532356e-01
-1.33305416e-01 -2.43493214e-01 -8.34403262e-02 7.94098318e-01
1.53910488e-01 -1.41752623e-02 3.29970628e-01 5.10321498e-01
1.05942905e+00 -2.20200419e-01 5.99184036e-01 -3.44251186e-01
6.18745208e-01 1.65755138e-01 7.32057691e-01 3.90966028e-01
-1.67021558e-01 3.90480608e-01 5.74704409e-01 -5.32183230e-01
-8.72442842e-01 -7.81789958e-01 1.54584989e-01 7.97641695e-01
5.13776660e-01 -4.85665917e-01 -6.92987204e-01 -1.18964744e+00
-9.09021348e-02 2.24980917e-02 -5.32444537e-01 -1.55678824e-01
-3.68193835e-01 -8.99496675e-01 9.14111212e-02 3.38956833e-01
4.89117622e-01 -8.56929243e-01 3.78560036e-01 -1.57817937e-02
-5.96994422e-02 -1.40039265e+00 -6.16284847e-01 1.18939847e-01
-1.05549920e+00 -1.46881878e+00 -3.07391167e-01 -1.49649131e+00
1.53687644e+00 5.37047565e-01 9.78745520e-01 4.13353860e-01
-7.78409317e-02 1.16904475e-01 -4.02247459e-01 -2.96741743e-02
-3.76662970e-01 1.67779535e-01 3.45320776e-02 4.46267515e-01
2.54089683e-01 -3.94870579e-01 -2.65831947e-01 5.00904620e-01
-8.80334020e-01 2.43688419e-01 5.86736381e-01 9.62271452e-01
8.99224341e-01 2.46554196e-01 7.28176117e-01 -1.43285775e+00
4.94393408e-01 -6.88484371e-01 -8.09928060e-01 5.65020680e-01
-9.08819139e-01 1.86589792e-01 1.05495930e+00 -5.86874247e-01
-6.19566083e-01 3.93787801e-01 3.30074698e-01 -4.14349914e-01
3.52110565e-02 7.07532227e-01 -3.34902644e-01 -2.76675880e-01
5.66237748e-01 1.60048366e-01 1.13856390e-01 -3.74490231e-01
3.83347899e-01 6.36460185e-01 1.30983785e-01 -5.89846492e-01
1.03508890e+00 4.15505469e-01 3.81210327e-01 -6.06509268e-01
-1.13778365e+00 -6.74798489e-01 -9.82143104e-01 -2.49527529e-01
3.98286223e-01 -8.23979318e-01 -4.28948730e-01 3.34379733e-01
-5.56774735e-01 -2.71099299e-01 -3.15053254e-01 6.17380083e-01
-2.61021733e-01 6.70899510e-01 -5.38683236e-01 -5.79043925e-01
-2.54270941e-01 -1.01432419e+00 9.10124481e-01 1.59594249e-02
3.17323089e-01 -1.31564009e+00 2.41987914e-01 4.87977266e-01
-1.90889090e-01 2.96952963e-01 9.75979149e-01 -8.49669874e-01
-6.01978958e-01 -5.73638320e-01 -5.41476786e-01 3.24080110e-01
5.56491673e-01 -2.01675102e-01 -5.27116179e-01 -6.56739771e-01
-1.32649049e-01 -7.72074163e-01 5.11730969e-01 1.13974018e-02
9.61215496e-01 -3.14937919e-01 -3.79008651e-01 5.80513418e-01
1.53975797e+00 -1.95171699e-01 8.49751979e-02 -3.04361969e-01
1.28201258e+00 6.47106230e-01 8.90991688e-01 3.56839955e-01
3.20389181e-01 2.44864509e-01 3.73470634e-01 -3.85388047e-01
4.78742011e-02 -5.94976008e-01 2.62975842e-02 1.47486818e+00
5.38594462e-02 -3.30952972e-01 -9.07533109e-01 4.32612509e-01
-2.07698131e+00 -3.26930046e-01 -2.94155866e-01 2.28167033e+00
7.51864970e-01 5.94848115e-03 2.24114582e-01 2.75854737e-01
1.06884742e+00 2.24936828e-01 -7.80141771e-01 4.18886468e-02
-9.29811448e-02 -2.16422334e-01 5.08458376e-01 4.87843931e-01
-1.09418964e+00 1.11585760e+00 6.44707632e+00 8.95419896e-01
-1.10849333e+00 -1.72141433e-01 6.02627754e-01 2.25178719e-01
-2.45969370e-01 2.23394036e-01 -7.44252384e-01 3.38632613e-01
6.24807656e-01 -8.39372799e-02 5.28789639e-01 9.10903156e-01
6.56317221e-04 3.29365842e-02 -1.01496816e+00 1.10499668e+00
1.12995796e-01 -1.15127969e+00 1.37878165e-01 4.57571521e-02
8.41207087e-01 1.40144527e-02 -4.89020526e-01 4.99303117e-02
5.09703279e-01 -8.13982785e-01 2.80064285e-01 3.86635177e-02
9.29615617e-01 -5.12111545e-01 5.11666298e-01 7.31932998e-01
-1.53111005e+00 1.08733043e-01 -5.10738134e-01 -1.08215518e-01
-1.49760723e-01 8.25517237e-01 -8.71207714e-01 5.05627394e-01
2.58804351e-01 9.91098821e-01 -6.97907627e-01 7.54177392e-01
-3.89618099e-01 8.71595919e-01 -3.11229914e-01 -2.69661486e-01
2.81198382e-01 -6.76752925e-01 2.17283890e-01 9.17539001e-01
-6.50650635e-02 2.19885290e-01 8.29147220e-01 5.88070393e-01
-3.22453260e-01 6.30699694e-01 -9.82846975e-01 -3.36199015e-01
3.70154440e-01 1.65058708e+00 -1.22181225e+00 -4.26582605e-01
-5.89944482e-01 6.37443304e-01 1.01040387e+00 4.53147948e-01
-5.32533646e-01 -1.37888923e-01 -1.94180608e-01 1.33751169e-01
5.38969189e-02 -2.98882693e-01 -2.64982991e-02 -1.35161638e+00
1.25229701e-01 -7.02120781e-01 7.11768508e-01 -4.57400262e-01
-1.47345102e+00 6.67949677e-01 -6.36308566e-02 -1.49905646e+00
1.34288639e-01 -4.48154122e-01 -4.50484037e-01 5.55315912e-01
-1.37773764e+00 -1.23764431e+00 -4.35718656e-01 8.55769217e-01
2.84964770e-01 -1.25936359e-01 6.71806037e-01 1.98984891e-01
-7.86067665e-01 3.19052219e-01 2.09459022e-01 3.15425575e-01
5.21914601e-01 -1.28570771e+00 6.66234791e-02 7.88640857e-01
4.04771715e-01 3.84570241e-01 1.98810518e-01 -7.37988234e-01
-1.58047783e+00 -1.24869573e+00 7.74353147e-01 -7.16387257e-02
7.25622356e-01 -6.13848269e-01 -9.40446734e-01 7.93815672e-01
-1.98149487e-01 7.42493033e-01 7.57503748e-01 1.08618617e-01
-3.32939923e-01 -2.02465534e-01 -1.20415592e+00 3.40171009e-01
1.09780443e+00 -4.85676050e-01 -3.73507410e-01 7.94882774e-01
5.53609014e-01 -1.33541167e-01 -9.31599796e-01 3.58651757e-01
2.84165293e-02 -4.66400445e-01 6.65436685e-01 -7.29659677e-01
9.07124579e-02 -4.77539390e-01 1.25448287e-01 -1.32568145e+00
-3.14317316e-01 -4.70743120e-01 2.17141490e-02 1.20589745e+00
3.71288329e-01 -8.15491676e-01 1.00848997e+00 4.98191804e-01
3.12621474e-01 -8.45301926e-01 -5.49193978e-01 -8.25078130e-01
-4.20456380e-01 6.75078854e-02 3.15108448e-01 1.35327876e+00
1.10169888e-01 9.12345171e-01 -4.75921124e-01 2.27109984e-01
9.23020661e-01 7.27995694e-01 8.09477985e-01 -1.40837431e+00
-1.75776646e-01 2.04205886e-01 -6.03552103e-01 -8.98659945e-01
2.92841762e-01 -1.20887375e+00 -4.42862064e-02 -1.63621950e+00
4.96008068e-01 -9.24329400e-01 -2.39426941e-01 6.20391846e-01
-3.70537549e-01 1.39699727e-01 -1.17269553e-01 4.24612790e-01
-8.78586054e-01 5.37624955e-01 1.47733212e+00 -1.57686993e-01
-2.36626640e-01 1.95847657e-02 -4.77002829e-01 7.00687408e-01
6.42267704e-01 -5.90644360e-01 -9.11751866e-01 -9.88490209e-02
1.44815952e-01 1.78062022e-01 8.64629149e-02 -5.24247110e-01
1.39380872e-01 -3.92292768e-01 8.15566443e-03 -4.61844325e-01
9.64954961e-03 -8.54611754e-01 -6.50869757e-02 2.15764716e-01
-3.60118955e-01 -1.84147388e-01 -3.06459397e-01 8.53722215e-01
-3.30329567e-01 -2.11743474e-01 8.27776015e-01 -1.21151254e-01
-6.46589279e-01 7.07115889e-01 2.20492423e-01 3.49950582e-01
1.29955935e+00 -1.56994879e-01 -2.16219619e-01 -3.46518278e-01
-7.38362968e-01 4.39894676e-01 5.98191857e-01 1.60036489e-01
6.09380305e-01 -1.50007522e+00 -5.17342687e-01 2.89792120e-01
3.27002913e-01 4.19144362e-01 -1.23173334e-01 5.96031964e-01
-5.40474176e-01 1.75142482e-01 -4.42441814e-02 -5.23204088e-01
-1.17420733e+00 9.98863935e-01 -2.90602595e-02 -4.37564284e-01
-7.16718793e-01 5.84519744e-01 2.82971680e-01 -7.38390267e-01
1.71471208e-01 -1.27580523e-01 -2.08796263e-01 6.30661547e-02
8.34394619e-02 2.54033923e-01 -1.46178588e-01 -7.97077298e-01
-2.39797980e-01 7.38092422e-01 9.75391939e-02 3.72385025e-01
1.29152942e+00 -9.60799679e-02 -2.77535319e-01 4.86405671e-01
1.49134183e+00 -8.95567313e-02 -1.14479315e+00 -8.08278739e-01
-2.72213574e-02 -5.42763829e-01 -3.17921303e-02 -1.58799872e-01
-1.49619937e+00 7.14194894e-01 1.18631773e-01 2.96900570e-01
1.16091728e+00 2.67329782e-01 6.10004008e-01 4.79911178e-01
4.97903973e-01 -1.05803597e+00 1.85664147e-01 4.20237854e-02
5.32524288e-01 -1.57975674e+00 3.56789351e-01 -1.00808394e+00
-7.07895935e-01 9.17498052e-01 6.39840364e-01 -2.82409638e-01
8.76381397e-01 1.92267355e-02 7.42184296e-02 -5.50147355e-01
-5.34637570e-01 -3.18004131e-01 3.39630604e-01 4.26881105e-01
3.26027982e-02 1.16580866e-01 -3.32432598e-01 2.47577637e-01
2.63733476e-01 -1.41798615e-01 3.34098548e-01 1.02092957e+00
-3.88124347e-01 -1.15053940e+00 -2.01090947e-01 7.35142827e-01
4.07889718e-03 8.95589367e-02 -6.41375184e-01 7.28323579e-01
-1.13134235e-01 8.40057731e-01 -2.24214494e-01 -4.63071376e-01
6.14869855e-02 -2.21919954e-01 2.93771267e-01 -9.63332653e-01
-1.80214614e-01 1.08377807e-01 -3.43053392e-03 -3.37817013e-01
-6.90200508e-01 -3.28082889e-01 -1.35422385e+00 1.62974924e-01
-6.38469040e-01 3.02119255e-01 3.53731871e-01 9.99092042e-01
2.59635806e-01 -1.01543650e-01 1.14387465e+00 -4.90525395e-01
-4.27496940e-01 -7.88335323e-01 -1.21465886e+00 7.21775711e-01
-2.48282012e-02 -6.17096364e-01 -5.28332889e-01 5.27413227e-02] | [7.412568092346191, 6.128377914428711] |
99523fec-58f8-4360-aa44-8b13965205e9 | monitoring-of-perception-systems | 2205.10906 | null | https://arxiv.org/abs/2205.10906v1 | https://arxiv.org/pdf/2205.10906v1.pdf | Monitoring of Perception Systems: Deterministic, Probabilistic, and Learning-based Fault Detection and Identification | This paper investigates runtime monitoring of perception systems. Perception is a critical component of high-integrity applications of robotics and autonomous systems, such as self-driving cars. In these applications, failure of perception systems may put human life at risk, and a broad adoption of these technologies requires the development of methodologies to guarantee and monitor safe operation. Despite the paramount importance of perception, currently there is no formal approach for system-level perception monitoring. In this paper, we formalize the problem of runtime fault detection and identification in perception systems and present a framework to model diagnostic information using a diagnostic graph. We then provide a set of deterministic, probabilistic, and learning-based algorithms that use diagnostic graphs to perform fault detection and identification. Moreover, we investigate fundamental limits and provide deterministic and probabilistic guarantees on the fault detection and identification results. We conclude the paper with an extensive experimental evaluation, which recreates several realistic failure modes in the LGSVL open-source autonomous driving simulator, and applies the proposed system monitors to a state-of-the-art autonomous driving software stack (Baidu's Apollo Auto). The results show that the proposed system monitors outperform baselines, have the potential of preventing accidents in realistic autonomous driving scenarios, and incur a negligible computational overhead. | ['Luca Carlone', 'Heath Nilsen', 'Pasquale Antonante'] | 2022-05-22 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [ 4.62386794e-02 2.96544224e-01 8.03442150e-02 -3.88077348e-01
-4.19710129e-01 -2.65181839e-01 7.05422461e-01 3.62218052e-01
-1.64829105e-01 2.67705768e-01 -6.04079664e-01 -7.97363043e-01
-2.09837198e-01 -8.41337144e-01 -9.86389220e-01 -4.31042552e-01
-3.99838895e-01 2.97037095e-01 1.18958950e+00 -4.74347174e-01
3.10016781e-01 6.36927724e-01 -2.76677608e+00 -2.19960541e-01
6.78970516e-01 1.26364362e+00 4.82833087e-02 1.01085651e+00
5.91521204e-01 7.28624940e-01 -7.53531218e-01 1.15998171e-01
6.67493492e-02 1.58526137e-01 -4.73542780e-01 2.19862591e-02
2.16217428e-01 -3.05325598e-01 -2.73294210e-01 9.99931037e-01
7.56322071e-02 -2.00781971e-01 3.11674595e-01 -2.40421677e+00
9.15452465e-02 1.16541624e-01 -9.01310667e-02 1.46582514e-01
2.35164031e-01 4.74298388e-01 3.40811610e-01 -3.74489844e-01
2.19603777e-02 1.21918321e+00 3.38400781e-01 3.65417480e-01
-9.28563356e-01 -3.48504514e-01 4.84128259e-02 6.06138289e-01
-1.32381213e+00 -6.63849354e-01 3.94367665e-01 -5.81103683e-01
1.38010776e+00 2.85163432e-01 8.82869437e-02 6.77781761e-01
1.10507596e+00 3.85164827e-01 1.03936815e+00 -2.78664678e-01
5.75344384e-01 4.21939194e-02 5.58711410e-01 8.57887626e-01
5.27839124e-01 6.85507238e-01 -4.75148201e-01 -2.86015511e-01
-1.63805410e-01 -3.16715807e-01 2.05265686e-01 -5.22033095e-01
-8.36907148e-01 3.56884241e-01 -1.39630750e-01 -2.56342500e-01
-2.12794051e-01 5.68377614e-01 7.43270099e-01 2.68910050e-01
-8.98561329e-02 -4.30250347e-01 -4.13288623e-02 -3.11126828e-01
-4.08718467e-01 1.92174837e-01 9.04534340e-01 1.11741912e+00
8.84930491e-01 7.14285597e-02 1.36698067e-01 1.33959115e-01
5.59436560e-01 6.82706654e-01 1.60882741e-01 -1.19829905e+00
-2.23056525e-01 5.26036143e-01 2.43055373e-01 -7.85305619e-01
-3.93083125e-01 -3.18987332e-02 -3.62288237e-01 1.03422105e+00
-4.63774353e-01 3.82073909e-01 -7.38279104e-01 1.33135676e+00
3.29087138e-01 2.54414082e-01 3.88928294e-01 6.62399948e-01
1.20844640e-01 2.44360596e-01 -2.59176344e-01 -1.54537320e-01
1.25362241e+00 -6.68475211e-01 -7.61231363e-01 -3.90930861e-01
4.44826156e-01 -5.11946976e-01 8.66281688e-01 8.02113175e-01
-8.71043742e-01 -8.97540987e-01 -1.77222574e+00 2.58123964e-01
-2.80265927e-01 -2.36089841e-01 2.18545645e-01 1.07694221e+00
-1.26325631e+00 2.03045979e-01 -1.24219108e+00 -4.35234904e-01
-1.50026411e-01 2.76614875e-01 -1.02974646e-01 -1.50090605e-01
-9.25960422e-01 1.31099117e+00 2.95642674e-01 4.90288399e-02
-1.77409852e+00 -2.46233076e-01 -1.09636378e+00 -3.13834935e-01
5.51936209e-01 -3.38802218e-01 1.61428070e+00 -2.29895506e-02
-1.23478913e+00 4.75934803e-01 -1.51111796e-01 -9.65071797e-01
4.38221604e-01 -3.05516332e-01 -8.90553534e-01 -9.44006722e-03
2.39276573e-01 2.10440248e-01 5.88928699e-01 -1.50457835e+00
-1.13525200e+00 -5.88740259e-02 3.66994590e-01 -3.55525464e-01
2.45758414e-01 -1.24988228e-01 -9.68993828e-02 6.28195941e-01
-1.56870723e-01 -9.60314572e-01 -1.65449589e-01 -2.17028901e-01
-2.81245619e-01 -3.13599974e-01 1.10008836e+00 -3.10447384e-02
8.09056938e-01 -2.48440719e+00 -5.56354821e-01 2.32284665e-01
-3.46743129e-02 -4.16815728e-02 3.04293782e-01 5.59026957e-01
3.27599972e-01 -3.20532590e-01 -3.22073042e-01 -3.86433721e-01
4.15769935e-01 8.27143371e-01 -4.62342769e-01 8.41562986e-01
1.68583423e-01 2.87709832e-01 -8.17350388e-01 -3.81832182e-01
7.39597023e-01 5.49294800e-02 -1.53969407e-01 2.97431439e-01
-3.71801436e-01 -2.20145896e-01 -7.13420287e-02 7.54644752e-01
5.87290645e-01 4.61074203e-01 -1.20813049e-01 1.48052260e-01
-4.67447549e-01 3.48254561e-01 -1.14279664e+00 1.15811121e+00
-4.88547534e-01 4.28578943e-01 2.43756935e-01 -8.77319336e-01
9.52246904e-01 1.50399998e-01 1.49092302e-01 -8.15380633e-01
1.93156973e-01 1.18059911e-01 -6.02258705e-02 -6.42156303e-01
6.26396775e-01 2.37172320e-01 -5.34909546e-01 2.57967323e-01
-3.11421186e-01 -3.33912134e-01 3.16059679e-01 9.89197046e-02
1.65969253e+00 1.74582209e-02 -1.39619140e-02 -2.50270933e-01
6.30596697e-01 2.99303621e-01 4.64136213e-01 7.01402962e-01
-6.83204830e-01 -5.00012673e-02 5.41529834e-01 -7.36904144e-02
-5.77496469e-01 -1.14688504e+00 -2.54924029e-01 7.67613411e-01
6.20344579e-01 -4.33370084e-01 -7.87521005e-01 -4.69658107e-01
3.18883061e-01 1.32269466e+00 -3.22155356e-01 -8.48471045e-01
-1.82230875e-01 -3.12125653e-01 7.47121990e-01 3.65585953e-01
3.36673707e-01 -7.63343692e-01 -1.37600148e+00 1.65609866e-01
8.82093087e-02 -1.19168174e+00 1.95205212e-01 2.82825381e-01
-2.21616283e-01 -1.35576153e+00 7.47569740e-01 -4.53290164e-01
3.70002270e-01 6.21775329e-01 1.01650167e+00 4.10431236e-01
-4.72207457e-01 8.37960362e-01 -1.73082426e-01 -6.51725888e-01
-1.21696997e+00 -6.27759755e-01 5.44402719e-01 -1.96093962e-01
3.07008237e-01 -3.33945066e-01 -3.42525333e-01 6.79949462e-01
-8.96107435e-01 -4.83362764e-01 4.91785526e-01 4.54392046e-01
4.83067364e-01 8.59862208e-01 3.95720780e-01 -3.52735341e-01
6.32478118e-01 -5.15781820e-01 -1.18459618e+00 7.09049031e-02
-1.27969205e+00 -5.11638932e-02 3.96021962e-01 1.39892980e-01
-7.27728367e-01 1.07398517e-01 -1.64381921e-01 -3.39386493e-01
-6.23743773e-01 2.83642799e-01 -4.94817734e-01 -3.41215320e-02
7.16922998e-01 2.19629049e-01 4.47538972e-01 1.33768991e-01
2.72214025e-01 7.04505205e-01 1.06904268e+00 -5.22880912e-01
5.91890037e-01 5.43064415e-01 3.16239893e-01 -6.76815391e-01
-9.34769511e-02 -4.84049171e-01 -1.77916303e-01 -5.78543186e-01
3.96913320e-01 -7.58185148e-01 -1.43571687e+00 5.72892129e-01
-9.09532964e-01 -3.73051375e-01 -9.79805738e-02 -3.90190892e-02
-9.50608909e-01 5.09767592e-01 -2.77316630e-01 -1.20354593e+00
1.65227532e-01 -1.48643076e+00 1.20715761e+00 -1.78330511e-01
-1.97121147e-02 -3.32738519e-01 8.39340910e-02 2.44627640e-01
2.38700807e-01 3.62921983e-01 6.45947635e-01 -2.53448129e-01
-8.88457060e-01 -5.48281670e-01 9.22473520e-02 7.46799111e-01
-3.11670452e-02 3.70561242e-01 -1.18663561e+00 -3.39610666e-01
1.71820357e-01 -1.24121860e-01 4.58840281e-01 -8.94791707e-02
8.03641737e-01 1.26266666e-03 -4.85288501e-01 -7.29324669e-02
1.39669728e+00 2.34880581e-01 6.90121591e-01 4.24579322e-01
1.37150303e-01 7.80478299e-01 1.32173383e+00 4.05778706e-01
7.35160053e-01 5.57993293e-01 1.22963798e+00 4.14254606e-01
6.34985277e-03 4.44789380e-02 9.04817343e-01 3.63039643e-01
4.24229145e-01 -6.78528920e-02 -9.56630468e-01 6.63255274e-01
-1.96213281e+00 -6.81687534e-01 -3.85425597e-01 2.45656037e+00
2.37205967e-01 8.56764317e-01 1.57723725e-01 7.79936194e-01
5.08348584e-01 -4.77028280e-01 -5.95058560e-01 -7.04242349e-01
3.17210048e-01 -2.64416963e-01 7.56101966e-01 6.88281000e-01
-8.20975363e-01 5.37560344e-01 6.69567728e+00 2.89741129e-01
-7.74108410e-01 3.45937997e-01 9.47869942e-03 3.00207227e-01
1.30553037e-01 1.45171702e-01 -8.07792366e-01 4.62479770e-01
1.87466514e+00 -3.42796266e-01 1.18569732e-01 1.22411895e+00
3.82638067e-01 -7.88468480e-01 -1.29350019e+00 4.09010947e-01
3.26717161e-02 -7.97290206e-01 -6.85520053e-01 1.74342006e-01
6.21815324e-02 2.63873845e-01 -1.89329326e-01 2.86407560e-01
5.02319932e-01 -7.34932244e-01 1.31335664e+00 4.92195159e-01
4.92994249e-01 -1.19655931e+00 9.46311176e-01 5.66742301e-01
-8.76470685e-01 -1.88524336e-01 -3.24981093e-01 -3.59678566e-01
4.07721192e-01 7.30674326e-01 -8.97925496e-01 5.98694324e-01
9.49126720e-01 9.88182798e-02 -6.51032746e-01 1.00116944e+00
-3.16389263e-01 4.90388453e-01 -2.62259156e-01 7.92932659e-02
-3.16733211e-01 3.80396247e-01 4.53505397e-01 9.35198486e-01
1.72357500e-01 -5.06165683e-01 4.10081983e-01 5.32737195e-01
8.03038001e-01 -8.17020595e-01 -7.88545132e-01 4.40867275e-01
5.75958252e-01 1.02015638e+00 -5.86474836e-01 -3.26091617e-01
-4.38602209e-01 6.64910793e-01 -1.81973234e-01 -8.84235874e-02
-1.18668807e+00 -2.26199493e-01 1.14575148e+00 2.76967883e-01
-8.97961780e-02 -5.62820196e-01 -2.59804606e-01 -2.01840803e-01
3.35171759e-01 -6.91393673e-01 9.49907675e-02 -6.41840339e-01
-7.62229383e-01 6.76671982e-01 4.00338799e-01 -1.12587273e+00
-5.02254963e-01 -6.32062793e-01 -3.58585626e-01 5.48653960e-01
-1.59759951e+00 -7.91892171e-01 -4.80269611e-01 3.43484759e-01
1.83853000e-01 -3.39091010e-02 7.26320863e-01 1.25340775e-01
-4.96073157e-01 3.79495293e-01 -3.33034992e-01 -7.51272142e-01
5.96977830e-01 -1.10983026e+00 6.93784237e-01 1.34726572e+00
-5.94237447e-01 3.18822742e-01 1.38839269e+00 -8.09687078e-01
-1.91879630e+00 -1.24738908e+00 4.26847309e-01 -5.79274952e-01
7.35352874e-01 -4.59120929e-01 -7.85281003e-01 5.88759720e-01
2.38139644e-01 7.75454715e-02 9.00557414e-02 -2.20510885e-01
-2.45128840e-01 -4.95284021e-01 -1.43809211e+00 2.43079916e-01
8.51021111e-01 -4.98731792e-01 -5.20774066e-01 3.03875983e-01
9.30529773e-01 -2.53890455e-01 -5.91155589e-01 5.87231576e-01
3.21488172e-01 -1.46061289e+00 6.79234743e-01 1.72741830e-01
-3.04447383e-01 -1.01338971e+00 -3.79614621e-01 -9.72824395e-01
-4.70602363e-02 -4.67396110e-01 -1.69681296e-01 1.09856820e+00
3.53144445e-02 -1.10999846e+00 3.03232759e-01 5.30541658e-01
-9.52595413e-01 -3.20660174e-01 -1.29765570e+00 -1.21646070e+00
-4.33199286e-01 -1.08969510e+00 4.73249912e-01 7.75862252e-03
1.50815308e-01 -9.07247141e-02 8.15717503e-02 1.15637875e+00
8.84689033e-01 -2.03512400e-01 9.56573904e-01 -1.01637173e+00
-1.70713082e-01 -2.53401548e-01 -1.15524983e+00 -4.70888793e-01
2.21913904e-01 -3.66185486e-01 1.03532100e+00 -1.46651375e+00
-3.54636729e-01 -3.64090919e-01 -2.56426960e-01 4.80059862e-01
3.07333261e-01 -1.49534315e-01 -4.38874304e-01 9.94003657e-03
-7.46761799e-01 4.79005367e-01 2.11849526e-01 -1.36078820e-01
3.98413986e-01 1.49812475e-01 -1.97732449e-01 5.21135032e-01
7.09569216e-01 -3.32050443e-01 -7.53629327e-01 5.30037507e-02
7.72519410e-02 1.39674321e-02 8.90562177e-01 -1.68125010e+00
5.25956452e-01 -2.36762062e-01 -6.24678791e-01 -7.90022075e-01
3.34015578e-01 -1.02518153e+00 1.91611528e-01 9.63374257e-01
2.80658394e-01 4.49423254e-01 5.13209224e-01 7.86740780e-01
-9.48785245e-02 -1.23278201e-02 7.00376630e-01 4.81551707e-01
-1.16261613e+00 -1.90387011e-01 -1.08451784e+00 -5.29199481e-01
1.65471053e+00 -1.04219548e-01 -7.54685760e-01 8.98259431e-02
-2.88178563e-01 5.08469880e-01 8.84621680e-01 7.04280972e-01
8.68467510e-01 -9.13093925e-01 -4.14687335e-01 3.93541574e-01
7.01524794e-01 -1.56334013e-01 2.17424750e-01 6.59750640e-01
-5.02151251e-01 3.55613858e-01 -2.27114797e-01 -9.77998257e-01
-1.17171490e+00 1.02430272e+00 3.31617057e-01 4.94090945e-01
-4.92623329e-01 3.81261319e-01 -2.57698018e-02 -1.78899333e-01
3.15616161e-01 -5.04806101e-01 3.01140279e-01 -7.09662259e-01
5.82796931e-01 5.83921552e-01 7.15697289e-01 -7.02238798e-01
-7.64437437e-01 -8.73240605e-02 2.35076770e-01 -2.62732267e-01
5.61023355e-01 -4.61666495e-01 -2.37180218e-01 7.93836892e-01
4.54579026e-01 -3.49263340e-01 -1.18096900e+00 2.79146731e-01
1.97797507e-01 -2.56002367e-01 1.43507570e-01 -7.65895128e-01
-4.59435582e-01 7.27258563e-01 1.01890945e+00 6.73128009e-01
9.46504116e-01 5.89902475e-02 5.30896008e-01 2.64142603e-01
1.19132483e+00 -9.56828833e-01 -2.75948912e-01 3.72875929e-01
7.61954844e-01 -1.06695759e+00 -2.65453696e-01 -5.80175519e-01
-2.57864535e-01 5.47865748e-01 9.14167106e-01 -9.65265930e-02
6.52242303e-01 8.43600273e-01 7.48095959e-02 -2.12009385e-01
-1.21921551e+00 -3.30854833e-01 -2.78801471e-01 9.37658787e-01
-3.05581957e-01 3.82739246e-01 -1.48246810e-02 3.53917718e-01
-2.13253900e-01 -2.29513925e-02 1.04501665e+00 1.47448993e+00
-8.89659047e-01 -1.06114030e+00 -6.47342682e-01 1.28091559e-01
9.23137218e-02 3.98306906e-01 -8.60761106e-03 7.12325931e-01
3.28863293e-01 1.80145288e+00 1.82367000e-03 -7.90690064e-01
7.83743620e-01 3.93861942e-02 2.39482880e-01 -5.40519178e-01
3.72090265e-02 -4.72636312e-01 4.19884890e-01 -9.81965244e-01
-7.12338313e-02 -7.81997442e-01 -1.48779619e+00 -3.61426413e-01
-1.05388537e-01 7.81540945e-03 1.14899349e+00 9.56947923e-01
5.65850854e-01 8.94283235e-01 6.71236277e-01 -5.27402520e-01
-7.35291004e-01 -6.22871757e-01 -5.15669286e-01 -1.45778522e-01
5.26511729e-01 -1.16457784e+00 -4.58410710e-01 -2.38918625e-02] | [5.426877021789551, 1.4375098943710327] |
11a4fcdd-a77e-4fe6-a921-491616d229bc | minmax-radon-barcodes-for-medical-image | 1610.00318 | null | http://arxiv.org/abs/1610.00318v1 | http://arxiv.org/pdf/1610.00318v1.pdf | MinMax Radon Barcodes for Medical Image Retrieval | Content-based medical image retrieval can support diagnostic decisions by
clinical experts. Examining similar images may provide clues to the expert to
remove uncertainties in his/her final diagnosis. Beyond conventional feature
descriptors, binary features in different ways have been recently proposed to
encode the image content. A recent proposal is "Radon barcodes" that employ
binarized Radon projections to tag/annotate medical images with content-based
binary vectors, called barcodes. In this paper, MinMax Radon barcodes are
introduced which are superior to "local thresholding" scheme suggested in the
literature. Using IRMA dataset with 14,410 x-ray images from 193 different
classes, the advantage of using MinMax Radon barcodes over \emph{thresholded}
Radon barcodes are demonstrated. The retrieval error for direct search drops by
more than 15\%. As well, SURF, as a well-established non-binary approach, and
BRISK, as a recent binary method are examined to compare their results with
MinMax Radon barcodes when retrieving images from IRMA dataset. The results
demonstrate that MinMax Radon barcodes are faster and more accurate when
applied on IRMA images. | ['Varun Chaudhari', 'Tahmid Mehdi', 'H. R. Tizhoosh', 'Hanson Lo', 'Shujin Zhu'] | 2016-10-02 | null | null | null | null | ['medical-image-retrieval', 'medical-image-retrieval'] | ['computer-vision', 'medical'] | [ 4.11502481e-01 -2.04498265e-02 -4.35804218e-01 -3.83282661e-01
-1.66824758e+00 -3.35245222e-01 4.98167574e-01 7.99483061e-01
-7.97608495e-01 5.76224327e-01 4.14069206e-01 -3.73764604e-01
-8.08037996e-01 -8.00263226e-01 -1.41516224e-01 -9.77619529e-01
1.71246231e-01 4.34533417e-01 2.82064706e-01 5.63383028e-02
5.82022190e-01 7.26791084e-01 -1.52220201e+00 6.09856606e-01
4.78607684e-01 1.30261898e+00 3.48905295e-01 7.78139770e-01
-2.58265197e-01 8.72468948e-01 -6.06137931e-01 -4.24773172e-02
1.77136660e-01 -4.01462764e-01 -9.13266480e-01 -1.22311346e-01
3.36714268e-01 -2.31318131e-01 -5.19402325e-01 1.07652164e+00
5.28307080e-01 7.41146738e-03 1.18005085e+00 -2.22954556e-01
-9.37046349e-01 4.36101496e-01 -5.48874676e-01 6.97638094e-01
6.91765010e-01 -4.45791394e-01 7.38951147e-01 -9.15714383e-01
6.43783689e-01 9.76219833e-01 9.22514617e-01 2.06002340e-01
-8.39069843e-01 -3.55881602e-01 -5.81498146e-01 2.85899013e-01
-1.75615740e+00 -1.49850473e-01 3.69483173e-01 -4.52940911e-01
7.23643899e-01 1.03280175e+00 4.40961123e-01 2.82089561e-01
7.42761850e-01 2.58138239e-01 1.43314826e+00 -7.45246053e-01
4.13448624e-02 1.90766141e-01 4.56741512e-01 1.13532436e+00
5.09405136e-01 -6.37459452e-04 -3.25280160e-01 -5.67669988e-01
4.76464808e-01 4.82009113e-01 -3.51429641e-01 4.83539933e-03
-1.13808703e+00 9.88536775e-01 6.34589612e-01 7.45431066e-01
-5.42020619e-01 1.52562112e-01 3.53187084e-01 -2.94552129e-02
2.71064788e-01 3.81795228e-01 2.82904208e-01 1.63066328e-01
-1.07149732e+00 -6.25067577e-02 2.14719802e-01 3.62750500e-01
2.10980013e-01 -4.64672714e-01 -6.72982574e-01 1.05187368e+00
5.57348728e-01 7.52878785e-01 1.09298277e+00 -5.19202530e-01
1.61938816e-01 6.59643054e-01 -2.03955054e-01 -1.13854849e+00
-4.28450316e-01 -5.17660715e-02 -7.30309248e-01 2.92364091e-01
1.62725657e-01 6.36467576e-01 -1.57912123e+00 6.91878438e-01
-1.43512962e-02 -4.87305492e-01 1.62172511e-01 7.17530429e-01
1.23106968e+00 5.09130895e-01 3.14266086e-02 -6.78486675e-02
1.98370218e+00 -3.12114477e-01 -7.07727194e-01 4.90141094e-01
6.22271478e-01 -1.09442365e+00 6.38599217e-01 5.33357441e-01
-9.24194574e-01 -3.53533566e-01 -1.00247514e+00 6.05086498e-02
-7.12461054e-01 2.99298972e-01 6.05826378e-01 1.05223942e+00
-1.12701821e+00 3.83536637e-01 -7.15331972e-01 -3.50585431e-01
3.68043214e-01 3.92177880e-01 -4.28328484e-01 -3.63617331e-01
-9.12933409e-01 1.05097008e+00 3.24732453e-01 -7.81153142e-02
-3.61998260e-01 -3.11933637e-01 -9.31049526e-01 -6.97077960e-02
7.69575238e-02 -1.48927346e-01 1.03209305e+00 -9.25983787e-02
-9.80870068e-01 1.18104744e+00 -2.59476714e-02 -4.59874332e-01
3.04130405e-01 2.68460363e-01 -5.00451386e-01 1.02000713e+00
2.41347969e-01 6.32261157e-01 5.82943320e-01 -9.84920859e-01
-5.83422661e-01 -3.14836979e-01 -2.56616801e-01 1.48688003e-01
-7.73573220e-02 1.53499529e-01 -2.41093948e-01 -7.87709892e-01
7.69276023e-01 -6.88087404e-01 -3.00378680e-01 2.07015455e-01
-3.43117326e-01 -3.32868308e-01 6.82622254e-01 -5.82616568e-01
1.17801785e+00 -2.02118778e+00 -5.96985221e-01 8.74854982e-01
2.19614133e-01 1.37424514e-01 3.69244158e-01 2.24807993e-01
-1.02456048e-01 2.54966021e-01 -4.40590888e-01 2.74788439e-01
-4.20800000e-01 2.02312008e-01 -5.20717762e-02 7.68077374e-01
-4.94833142e-02 7.17852354e-01 -4.95848387e-01 -1.17317903e+00
6.27712131e-01 7.33027637e-01 -1.63349614e-01 -3.17838013e-01
6.00632310e-01 -3.28337308e-04 -6.22535229e-01 1.05749381e+00
5.45329690e-01 -3.50017160e-01 -2.76560694e-01 -3.28047335e-01
-1.13109464e-03 -2.15701982e-01 -8.23502779e-01 1.32275724e+00
-2.30146289e-01 6.60075784e-01 -4.63156879e-01 -9.61609781e-01
1.07143700e+00 5.40222228e-01 8.85325313e-01 -9.87317324e-01
2.65799314e-01 3.63702357e-01 -3.69800270e-01 -5.99007070e-01
7.61506438e-01 -2.29939550e-01 -1.10952154e-01 1.53362542e-01
-3.78597915e-01 -1.78832427e-01 1.93224967e-01 9.40368026e-02
1.15030110e+00 -3.98614228e-01 6.57940626e-01 -3.36817205e-01
7.93055475e-01 2.75897712e-01 -1.70788299e-02 1.06878603e+00
6.89085349e-02 1.14429402e+00 -4.98863636e-03 -3.07478458e-01
-7.22705126e-01 -1.18163276e+00 -1.11915648e+00 4.60336775e-01
1.56254977e-01 -3.16077441e-01 -4.89628673e-01 -5.41532040e-01
-4.61664386e-02 2.32512847e-01 -8.48291814e-01 4.72890399e-03
-3.86771530e-01 -1.05143929e+00 5.44549286e-01 4.31981653e-01
4.25785840e-01 -9.74121690e-01 -1.04106987e+00 8.67374167e-02
-1.05645746e-01 -5.43704450e-01 -2.96287686e-01 3.30926269e-01
-8.95840883e-01 -1.38211215e+00 -1.42668319e+00 -6.03959858e-01
1.03816330e+00 4.51586694e-01 7.60103881e-01 2.36039102e-01
-1.19118178e+00 6.93084180e-01 -7.73549378e-01 -7.89791420e-02
-4.12492007e-01 -3.71798187e-01 -4.08188194e-01 -4.28497374e-01
2.95377225e-01 1.73822194e-01 -1.07064259e+00 8.72640759e-02
-1.21609962e+00 -5.98794341e-01 9.97663260e-01 8.88074160e-01
9.85469282e-01 -4.68510203e-02 2.75171380e-02 -1.05669391e+00
6.23312831e-01 -3.43947440e-01 -3.21732402e-01 3.13220114e-01
-9.81691480e-01 2.68016994e-01 9.38356817e-02 8.86891969e-03
-8.10074151e-01 -7.84271955e-02 -3.26285511e-01 -1.96740374e-01
1.61826462e-01 6.94160581e-01 8.43360186e-01 -4.83993590e-01
9.46824133e-01 3.25774968e-01 5.12485988e-02 -4.90176797e-01
1.26976237e-01 9.74471629e-01 7.08621860e-01 -2.75819600e-01
4.33965713e-01 8.09025347e-01 1.86964840e-01 -8.30767632e-01
-3.57544035e-01 -1.13120770e+00 -2.29887798e-01 -6.38220385e-02
1.18748963e+00 -6.29486680e-01 -5.49372971e-01 -1.80298999e-01
-6.54780746e-01 5.65962911e-01 -6.04690969e-01 7.67372787e-01
-4.77644950e-01 6.24954164e-01 -7.11331248e-01 -7.73242593e-01
-6.78423405e-01 -1.27433813e+00 1.17876911e+00 3.39448303e-01
-1.80344671e-01 -6.84106648e-01 2.58740216e-01 3.47831815e-01
3.47387910e-01 9.23554823e-02 8.49408627e-01 -7.05772221e-01
-1.85248032e-01 -7.40413368e-01 -5.15932977e-01 1.58453554e-01
2.48068601e-01 -1.60920858e-01 -7.69089520e-01 -1.62216470e-01
1.06488299e-02 4.71716598e-02 1.13307273e+00 6.14737391e-01
1.38531089e+00 -1.77777752e-01 -6.94222808e-01 4.62558627e-01
1.70841849e+00 7.96093106e-01 9.48172331e-01 4.40615207e-01
2.83921510e-01 1.15854427e-01 7.89792597e-01 3.45680952e-01
-8.52096919e-03 3.84504437e-01 2.68720478e-01 -3.67521018e-01
-1.75101817e-01 2.46487305e-01 -1.50464877e-01 6.48660302e-01
-2.04607069e-01 -8.62758141e-03 -9.48378623e-01 3.42448115e-01
-1.14409280e+00 -9.83348489e-01 -2.28804946e-02 2.07347441e+00
6.01392686e-01 3.45779322e-02 -3.44259232e-01 5.09443760e-01
7.30469525e-01 1.43061772e-01 4.06491868e-02 -2.69155234e-01
-1.37193287e-02 8.19840491e-01 9.87153530e-01 3.40769053e-01
-1.23763454e+00 3.35308820e-01 6.59517670e+00 9.70125735e-01
-1.20276010e+00 1.85564011e-01 7.23185897e-01 4.41713214e-01
-2.48016000e-01 -1.62689850e-01 -6.85696244e-01 3.47689271e-01
7.50935376e-01 1.42469108e-01 -4.70830828e-01 8.64282846e-01
-2.30743699e-02 -7.95652747e-01 -4.87121403e-01 1.34599042e+00
4.80062068e-01 -1.43496764e+00 7.68588409e-02 2.29276061e-01
4.50387388e-01 -1.35948494e-01 2.93288201e-01 -2.29979828e-01
-6.15131925e-04 -1.12063253e+00 4.19581711e-01 1.00498462e+00
1.22401261e+00 -6.02254570e-01 1.02658474e+00 -4.44395095e-01
-1.17797244e+00 1.60899073e-01 -5.21223843e-01 6.61237657e-01
-1.18463971e-01 4.21022356e-01 -1.15457070e+00 6.46409154e-01
7.78163552e-01 1.77253246e-01 -9.17215168e-01 1.49795198e+00
4.00941610e-01 2.35909194e-01 -3.15099716e-01 7.98459426e-02
4.25886154e-01 -9.92959961e-02 2.38228425e-01 1.52081954e+00
7.62416959e-01 4.78901714e-01 -1.28647268e-01 1.76801577e-01
3.30960721e-01 5.37993193e-01 -7.25891352e-01 9.28775221e-03
1.90613106e-01 1.13951409e+00 -1.35933745e+00 -5.68084836e-01
-2.46537298e-01 7.79916346e-01 -6.22204959e-01 3.26378085e-03
-5.13060570e-01 -8.12095225e-01 -8.27406570e-02 3.53598058e-01
2.00120494e-01 2.43337169e-01 1.19096767e-02 -5.35311878e-01
-2.31290936e-01 -5.57067990e-01 9.77782667e-01 -8.56992304e-01
-9.95042622e-01 9.36022222e-01 2.82598853e-01 -1.63891757e+00
-2.82908440e-01 -9.22734201e-01 -1.08568087e-01 6.91926479e-01
-1.39313745e+00 -6.90675139e-01 -3.01446259e-01 6.75605416e-01
3.45189869e-01 -2.13080436e-01 1.01793599e+00 1.91116974e-01
1.37797043e-01 3.17539960e-01 7.19267249e-01 2.04825655e-01
6.33055925e-01 -1.21639597e+00 -5.51105618e-01 3.40095371e-01
2.60548949e-01 5.96208870e-01 4.15801495e-01 -5.03173888e-01
-1.04078770e+00 -6.41107082e-01 7.16742694e-01 -2.01563492e-01
1.09135270e-01 4.41323072e-01 -9.11224902e-01 1.03927188e-01
-1.11027760e-02 3.53071481e-01 8.27061117e-01 -6.54974163e-01
-2.59402275e-01 -1.49738014e-01 -1.66392589e+00 1.14976570e-01
2.90956676e-01 -7.29405224e-01 -8.62384498e-01 4.80658889e-01
1.90642461e-01 -4.40269768e-01 -1.14960098e+00 5.84511757e-01
4.89083588e-01 -9.74642694e-01 1.47562706e+00 -2.96507832e-02
1.00172132e-01 -3.06050956e-01 -3.14748138e-01 -6.24418974e-01
9.77219734e-03 9.64632854e-02 5.74318171e-01 5.81129789e-01
2.83047915e-01 -6.24486804e-01 7.19596207e-01 2.88376749e-01
-1.63325176e-01 -7.57936001e-01 -1.26140118e+00 -4.99888182e-01
-2.00934216e-01 -1.77945033e-01 2.03164276e-02 7.53304660e-01
8.20273235e-02 -4.85272974e-01 1.53040573e-01 -2.11332917e-01
3.60058010e-01 -1.55434355e-01 -2.12106720e-01 -1.18918192e+00
-3.96543369e-02 -5.01573741e-01 -9.22142148e-01 -4.40737993e-01
-6.41991019e-01 -1.16427898e+00 2.66781867e-01 -1.83459044e+00
4.55397338e-01 -7.45568395e-01 -8.42738986e-01 4.34862196e-01
1.10446895e-02 1.03195632e+00 -7.59169385e-02 5.39223433e-01
-3.81987512e-01 -5.94098680e-02 8.84166896e-01 -3.74470592e-01
7.53759742e-02 1.21684849e-01 -5.39067090e-01 6.41989350e-01
4.83224303e-01 -8.00252438e-01 -2.57496357e-01 3.20982993e-01
9.31289420e-02 3.28011990e-01 2.36873657e-01 -1.23320794e+00
2.41154656e-01 2.52561271e-01 6.16343737e-01 -1.10070539e+00
5.84380984e-01 -9.15471137e-01 1.38719827e-01 9.24123824e-01
-3.92599642e-01 5.43962300e-01 -6.06147759e-02 5.91649354e-01
-5.09126902e-01 -8.63636315e-01 7.18820870e-01 -4.71344441e-01
-6.13263130e-01 -1.87431201e-02 -7.03374386e-01 -4.88974988e-01
1.01166606e+00 -7.10132778e-01 -3.30910116e-01 -8.55966359e-02
-8.61380458e-01 -5.19868493e-01 2.23356318e-02 -6.62186518e-02
1.15546703e+00 -1.14059925e+00 -4.64961201e-01 -1.53949738e-01
4.43761200e-01 -1.59242928e-01 3.40993971e-01 1.01253736e+00
-1.20862067e+00 8.97956252e-01 -2.67027393e-02 -8.03067386e-01
-1.70318806e+00 5.22046745e-01 7.47343674e-02 -2.70803034e-01
-9.58760798e-01 7.08847344e-01 -8.51346478e-02 7.96925947e-02
-2.06152108e-02 -5.41080892e-01 -6.61286950e-01 4.08647865e-01
5.29538691e-01 1.59239411e-01 4.11594033e-01 -7.37811208e-01
-4.77897018e-01 1.16590834e+00 -5.62494516e-01 -4.19553220e-01
1.04761291e+00 -1.63773820e-02 -1.08761348e-01 2.27616295e-01
1.33012426e+00 1.70890212e-01 -4.01764810e-01 -2.75020480e-01
1.60227388e-01 -5.71764112e-01 3.20979506e-01 -6.94520831e-01
-9.17633772e-01 6.23426020e-01 1.40958416e+00 2.08367124e-01
1.21325815e+00 3.87553155e-01 3.66502911e-01 4.92924124e-01
1.04649857e-01 -9.23385024e-01 4.97020148e-02 -1.52274787e-01
7.62152493e-01 -1.29951048e+00 4.66427952e-01 -3.80608886e-01
-5.73085487e-01 1.23376882e+00 -2.61715919e-01 -9.26050767e-02
6.19095802e-01 9.57105756e-02 3.29173803e-01 -5.53793669e-01
-8.77094194e-02 -2.61238664e-01 4.92317647e-01 3.42864424e-01
6.06227756e-01 2.99246479e-02 -8.84583414e-01 -8.03119168e-02
7.92684441e-04 -1.11554109e-01 2.98352093e-01 1.32700968e+00
-7.02556014e-01 -9.34368491e-01 -1.03051794e+00 1.00877714e+00
-9.25411046e-01 -1.96242869e-01 1.11597620e-01 9.02554154e-01
-5.50430678e-02 7.17696846e-01 1.28582090e-01 -5.96407950e-02
4.79736291e-02 -1.73496515e-01 4.25419897e-01 -4.16129053e-01
-6.02105319e-01 5.58738131e-03 -1.75000295e-01 -3.43490034e-01
-5.25704443e-01 -6.46221042e-01 -1.38032317e+00 2.42869735e-01
-4.53108668e-01 5.10772884e-01 1.03734052e+00 7.14244008e-01
-2.39844576e-01 6.22368097e-01 3.14663619e-01 -2.23765835e-01
-4.54078943e-01 -7.61323452e-01 -4.51988310e-01 2.98640877e-01
3.73689473e-01 -5.90297282e-01 -1.50524348e-01 2.08923146e-02] | [14.291685104370117, -1.4673207998275757] |
aec08f57-9303-4a35-8a4b-f00dda8fc165 | robust-and-efficient-fault-diagnosis-of-mm | 2306.04360 | null | https://arxiv.org/abs/2306.04360v1 | https://arxiv.org/pdf/2306.04360v1.pdf | Robust and Efficient Fault Diagnosis of mm-Wave Active Phased Arrays using Baseband Signal | One key communication block in 5G and 6G radios is the active phased array (APA). To ensure reliable operation, efficient and timely fault diagnosis of APAs on-site is crucial. To date, fault diagnosis has relied on measurement of frequency domain radiation patterns using costly equipment and multiple strictly controlled measurement probes, which are time-consuming, complex, and therefore infeasible for on-site deployment. This paper proposes a novel method exploiting a Deep Neural Network (DNN) tailored to extract the features hidden in the baseband in-phase and quadrature signals for classifying the different faults. It requires only a single probe in one measurement point for fast and accurate diagnosis of the faulty elements and components in APAs. Validation of the proposed method is done using a commercial 28 GHz APA. Accuracies of 99% and 80% have been demonstrated for single- and multi-element failure detection, respectively. Three different test scenarios are investigated: on-off antenna elements, phase variations, and magnitude attenuation variations. In a low signal to noise ratio of 4 dB, stable fault detection accuracy above 90% is maintained. This is all achieved with a detection time of milliseconds (e.g 6~ms), showing a high potential for on-site deployment. | ['Gert F. Pedersen', 'Ming Shen', 'Yingzeng Yin', 'Jian Ren', 'Changbin Xue', 'Yufeng Zhang', 'Martin H. Nielsen'] | 2023-06-07 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [-1.69468850e-01 -9.01512653e-02 1.45780072e-01 7.44381279e-04
-6.00629628e-01 -4.76083666e-01 -9.18553621e-02 2.35605404e-01
2.95175105e-01 6.29272223e-01 -5.54340541e-01 -5.07354081e-01
-8.35635245e-01 -8.51927221e-01 -2.33671695e-01 -1.01357746e+00
-4.28664058e-01 4.14654553e-01 -1.74873948e-01 1.90540746e-01
-1.49501175e-01 9.99145627e-01 -1.26956785e+00 1.90759730e-02
6.34361267e-01 1.80285549e+00 -2.25771680e-01 3.75364900e-01
5.66647530e-01 3.29723984e-01 -1.42653680e+00 1.91499665e-01
1.23569854e-01 -3.55727553e-01 -2.05003768e-01 2.95319594e-02
1.34798000e-02 -4.92284894e-01 -3.29203725e-01 8.24820697e-01
9.88590360e-01 -5.38234949e-01 5.25288045e-01 -1.27571297e+00
2.74824202e-01 3.91145349e-01 -3.37008387e-01 2.71857321e-01
2.94236183e-01 6.56557903e-02 6.92321539e-01 -6.20548069e-01
-1.77183971e-02 1.93607450e-01 8.59001338e-01 -2.77098894e-01
-1.00314629e+00 -8.61747086e-01 -9.00620937e-01 3.09310019e-01
-1.69855034e+00 -5.25713325e-01 9.99525666e-01 -3.27410728e-01
1.06575966e+00 2.24781007e-01 8.25987935e-01 5.83706260e-01
6.63310945e-01 -1.43254071e-01 8.63111496e-01 -6.39869690e-01
4.12453949e-01 -1.50013775e-01 -5.03012761e-02 8.07478547e-01
9.20487761e-01 5.83427995e-02 -2.93020487e-01 -3.37817311e-01
7.72591174e-01 -3.15856099e-01 -5.68109751e-01 -2.20516711e-01
-9.19697583e-01 3.03350121e-01 3.29667151e-01 8.58530760e-01
-8.02377284e-01 3.68631691e-01 -3.64114493e-02 4.41247165e-01
1.61137462e-01 5.44287026e-01 -1.80514351e-01 -2.51498997e-01
-1.06042981e+00 -2.76438713e-01 7.95878112e-01 6.62242055e-01
3.49886268e-01 6.38812900e-01 -4.32107337e-02 4.69184220e-01
5.64976633e-01 8.68014455e-01 1.14854813e-01 -3.44242215e-01
1.55423656e-01 6.89654872e-02 5.45829356e-01 -1.30924582e+00
-1.05038595e+00 -1.67006075e+00 -8.74189496e-01 -1.22745447e-01
3.05391192e-01 -4.75135088e-01 -8.76155734e-01 1.36222005e+00
1.34027153e-01 3.87488455e-01 -2.14493483e-01 6.74892306e-01
4.50686485e-01 6.51078582e-01 -4.16466832e-01 -4.97390091e-01
1.23758650e+00 -3.07506949e-01 -9.09092665e-01 -2.38572523e-01
4.80693191e-01 -8.18140566e-01 3.04081529e-01 9.06667173e-01
-9.55252588e-01 -5.98871648e-01 -1.86199427e+00 8.49030733e-01
2.95714200e-01 4.76575077e-01 2.57825613e-01 1.15089905e+00
-8.83579195e-01 3.99534672e-01 -6.39088929e-01 -3.00649494e-01
3.27958941e-01 6.44161820e-01 -1.57120839e-01 -1.21714920e-01
-1.23362124e+00 6.47926807e-01 -2.81413287e-01 5.03470182e-01
-7.55911052e-01 -8.26977015e-01 -4.32495475e-01 5.49047530e-01
-1.96907744e-01 -6.16783321e-01 1.29068744e+00 -3.42444539e-01
-1.37257230e+00 2.85965800e-01 3.40065211e-01 -4.63669807e-01
-4.81923260e-02 -6.66159252e-03 -1.33419716e+00 4.37367380e-01
-1.30306318e-01 -6.90064669e-01 8.25779021e-01 -9.42288578e-01
-5.92610776e-01 -4.28906322e-01 -2.73198932e-01 -4.68343824e-01
6.17227256e-02 -2.28105560e-01 3.14976782e-01 -4.18041259e-01
8.77665460e-01 -6.36746764e-01 2.16795072e-01 -2.76878744e-01
-6.23548441e-02 2.46439353e-01 6.97113812e-01 -7.40345299e-01
1.25326598e+00 -2.13740659e+00 -5.24668217e-01 8.03919256e-01
3.50251347e-02 2.83236980e-01 3.44494075e-01 7.45483518e-01
-2.58703351e-01 -4.40075189e-01 1.90926701e-01 2.64906794e-01
1.61412314e-01 -1.64523050e-01 1.22170039e-01 9.37590539e-01
4.36296016e-02 3.94638836e-01 -2.85440266e-01 5.22455394e-01
2.86750227e-01 2.38863245e-01 -2.49772653e-01 1.91634595e-01
2.50317216e-01 2.86888123e-01 -3.79893690e-01 9.39149022e-01
8.40032339e-01 -3.17600042e-01 3.49394917e-01 -7.19633937e-01
1.12309447e-02 4.50379938e-01 -1.61132216e+00 1.44749081e+00
-8.52906704e-01 6.48742616e-01 3.50629002e-01 -1.29792488e+00
1.06711662e+00 7.98103273e-01 6.24006510e-01 -1.20196259e+00
4.77056652e-01 5.90043902e-01 4.14320350e-01 -5.66095054e-01
-2.93004364e-01 -3.30660909e-01 -4.18508470e-01 6.57377839e-01
2.49524087e-01 2.16568589e-01 -1.53923392e-01 -2.11803660e-01
1.46988809e+00 -5.45534849e-01 2.96901286e-01 -5.72921097e-01
4.81740862e-01 -2.90378124e-01 8.12306345e-01 6.31357253e-01
-3.66062194e-01 2.77636617e-01 3.93542141e-01 -3.23049515e-01
-5.90667069e-01 -9.41192865e-01 -2.31878981e-01 7.39449412e-02
2.02348784e-01 -5.09525090e-02 -4.10586566e-01 -2.64676541e-01
2.12267607e-01 5.90781987e-01 2.52246857e-03 -3.41190815e-01
-5.20418286e-01 -8.19105625e-01 6.11979008e-01 3.62239420e-01
6.52335107e-01 -3.16381007e-01 -5.73210955e-01 5.20117640e-01
6.21197186e-03 -1.06122422e+00 5.62392116e-01 3.79723549e-01
-3.79790634e-01 -1.01822591e+00 -3.56347680e-01 -7.08729446e-01
5.99767983e-01 1.98979691e-01 1.16859090e+00 5.28009608e-02
-5.42003334e-01 4.28344190e-01 -1.08025335e-01 -6.48538619e-02
-1.16301440e-02 -3.54039341e-01 3.99781764e-01 1.98549077e-01
5.92337176e-02 -9.46370065e-01 -8.21545005e-01 7.59069502e-01
-2.30705202e-01 -6.02210581e-01 1.00570881e+00 5.60927868e-01
1.31399557e-02 7.55643845e-01 1.05593944e+00 -3.62809658e-01
2.65997648e-01 -5.76701820e-01 -8.00831556e-01 -3.23691708e-03
-4.85651404e-01 -5.70294857e-01 6.35430992e-01 1.86386764e-01
-7.90704310e-01 -4.53624368e-01 -5.12245655e-01 1.07455954e-01
-2.49456033e-01 7.79100120e-01 -7.11942911e-01 -4.56994563e-01
6.98942363e-01 -1.69441476e-01 -1.76233545e-01 -1.70736596e-01
-3.72776419e-01 8.62539232e-01 3.74062628e-01 -2.52830863e-01
1.03707623e+00 2.19855577e-01 3.04905802e-01 -9.57931280e-01
-3.53289485e-01 -3.84055674e-01 2.30721966e-03 -4.02013034e-01
1.22797869e-01 -1.07364976e+00 -6.83585644e-01 6.44596040e-01
-7.16375709e-01 -7.20858499e-02 1.91214651e-01 8.63973558e-01
-1.82331100e-01 2.23562047e-01 -5.60789108e-01 -5.57728827e-01
-5.48508763e-01 -8.96586835e-01 6.52945340e-01 2.86610216e-01
-3.01279336e-01 -6.15055263e-01 -4.57941324e-01 3.34224045e-01
9.97988224e-01 2.21107915e-01 9.52079594e-01 -3.30683619e-01
-7.31018662e-01 -8.46018374e-01 9.08424631e-02 4.78100516e-02
3.81754845e-01 -2.50813037e-01 -9.20082510e-01 -6.56042635e-01
3.93340439e-01 2.69876063e-01 -9.51253325e-02 6.97612345e-01
7.41390467e-01 1.47690296e-01 -4.96846527e-01 5.27765930e-01
1.67452705e+00 5.74119925e-01 6.82783961e-01 1.82268713e-02
2.32758150e-02 -1.03151232e-01 7.67493486e-01 6.01826489e-01
-3.22930217e-01 4.52715248e-01 5.00816226e-01 9.27516893e-02
1.25688715e-02 3.56795013e-01 -1.25341967e-01 6.02778196e-01
5.15845239e-01 -9.04200733e-01 -8.63402963e-01 4.12976891e-01
-1.22502315e+00 -5.61652124e-01 -2.44862050e-01 2.10941243e+00
3.65561992e-02 5.02211988e-01 -1.62123427e-01 9.81242239e-01
4.54548150e-01 -1.06537588e-01 -1.71800271e-01 -8.05920735e-02
1.98737070e-01 6.40364528e-01 4.48562205e-01 3.74738544e-01
-9.45251644e-01 -1.21081837e-01 4.99140310e+00 5.23818016e-01
-1.46946454e+00 7.69439414e-02 3.13709587e-01 -4.94155660e-02
-1.11369587e-01 -3.50535214e-01 -3.60196799e-01 4.64044094e-01
1.19211924e+00 2.87831068e-01 -3.44930083e-01 6.17264152e-01
3.49247426e-01 -2.81719297e-01 -8.26617241e-01 1.27674162e+00
-8.47297534e-02 -1.21046293e+00 -6.20047510e-01 -1.43574476e-02
3.87205720e-01 -3.31095129e-01 1.70243345e-02 8.43924209e-02
-6.16965353e-01 -7.05561697e-01 3.04498792e-01 5.24690509e-01
8.88139307e-01 -1.21107721e+00 1.15465343e+00 2.73424476e-01
-8.24066997e-01 -3.80526751e-01 1.50827065e-01 -1.84000716e-01
1.79467052e-01 1.47847688e+00 -1.00785470e+00 9.29490685e-01
6.31048322e-01 1.49456421e-02 1.76611822e-02 1.29345632e+00
-2.96843529e-01 1.12300241e+00 -6.77355647e-01 1.14063621e-01
-1.35613218e-01 1.73288628e-01 4.21401799e-01 6.84991896e-01
1.05370080e+00 -1.36697572e-03 -1.80245921e-01 2.61658162e-01
1.51128992e-01 -3.61209750e-01 -2.18593091e-01 3.19910794e-01
7.70645261e-01 1.26059461e+00 -6.14893317e-01 1.60934061e-01
-3.50870013e-01 5.38853467e-01 -5.29271901e-01 2.65519023e-01
-8.60875845e-01 -6.68905258e-01 6.40723526e-01 6.20587230e-01
4.07531738e-01 -5.32063186e-01 -3.72298360e-01 -4.29122180e-01
1.94681093e-01 -7.45240450e-01 2.13840038e-01 -6.40018344e-01
-8.83557916e-01 4.79162693e-01 -4.95308727e-01 -1.60431051e+00
-5.59313893e-01 -3.67438912e-01 -4.29097563e-01 9.67554390e-01
-1.12124848e+00 -9.42614436e-01 -5.69385946e-01 2.36387298e-01
-1.83030918e-01 -2.86482394e-01 1.03361845e+00 9.62362528e-01
-6.19949162e-01 7.87808657e-01 1.78084940e-01 -1.13876373e-01
4.03859854e-01 -8.15672874e-01 1.14490435e-01 1.22096598e+00
-1.23380378e-01 3.74533594e-01 8.61236513e-01 -4.75751072e-01
-1.75819194e+00 -7.40005851e-01 8.00306857e-01 5.35611868e-01
3.87764007e-01 -3.59037340e-01 -4.83276159e-01 2.82846600e-01
3.96549642e-01 5.10677397e-01 9.04897988e-01 -4.15845066e-02
2.74922162e-01 -8.08843374e-01 -1.51960528e+00 2.19859272e-01
5.95195055e-01 -2.30964363e-01 1.20331399e-01 2.28824496e-01
-7.69668669e-02 -3.54695469e-01 -1.01157594e+00 8.20468068e-01
3.84450585e-01 -1.01778054e+00 6.92737699e-01 1.51356161e-01
-4.86414522e-01 -7.06091940e-01 -2.46673793e-01 -1.41804779e+00
-8.02766800e-01 -6.64187551e-01 -4.35796201e-01 9.21436131e-01
2.34183580e-01 -8.77985060e-01 8.40471506e-01 -1.66151598e-01
-3.59982640e-01 -6.72330856e-01 -1.17599511e+00 -6.36405170e-01
-7.23109066e-01 -4.27089036e-01 6.52019083e-01 8.15226376e-01
-1.42881975e-01 2.42676035e-01 -1.94690093e-01 1.18349600e+00
6.40478313e-01 1.18094116e-01 4.18239266e-01 -1.48509145e+00
-5.84348619e-01 -1.10434957e-01 -9.63204205e-01 -7.10854769e-01
-4.12524402e-01 -3.06584477e-01 -2.04306990e-01 -1.65099955e+00
-7.47893274e-01 -8.20133507e-01 -3.96990210e-01 9.54640582e-02
5.21784902e-01 5.58781564e-01 -6.88007116e-01 -3.03313613e-01
-4.39177826e-02 1.78979710e-01 4.72558558e-01 -1.61977619e-01
3.94617707e-01 6.29739940e-01 -3.04186612e-01 5.32122254e-01
8.15047741e-01 -3.09706450e-01 -3.33690137e-01 -3.02258462e-01
4.14219677e-01 5.14528573e-01 1.84871912e-01 -1.89456713e+00
3.96474630e-01 4.47182894e-01 6.00289643e-01 -7.16976523e-01
2.98450083e-01 -1.25067759e+00 6.33548141e-01 7.39595413e-01
6.98036611e-01 -1.52506500e-01 4.01405275e-01 4.03838307e-01
-3.82697284e-01 -1.10164687e-01 7.16083705e-01 4.75878537e-01
-4.07181770e-01 -8.71968344e-02 -7.18426943e-01 -7.11329460e-01
1.15518737e+00 -1.94754779e-01 -2.32546672e-01 -4.95069623e-01
-7.56096542e-01 -2.27638930e-01 -1.22492850e-01 -2.00916290e-01
4.04498488e-01 -1.32892621e+00 -4.89444762e-01 7.08718121e-01
-8.50544721e-02 -2.21987784e-01 8.65930140e-01 9.64178145e-01
-8.87222946e-01 7.58614838e-01 -1.54873088e-01 -9.36690331e-01
-1.07751524e+00 -1.14647090e-01 9.10179675e-01 8.16014633e-02
-2.28199929e-01 7.55201697e-01 -7.08998799e-01 3.92429322e-01
7.86211118e-02 -3.42538983e-01 4.10581604e-02 1.64793193e-01
3.09742957e-01 3.05068672e-01 1.06820917e+00 -1.14602614e-02
-5.79887152e-01 6.00751221e-01 3.39953125e-01 9.07860994e-02
1.10323250e+00 -4.16503809e-02 9.66976359e-02 -1.76134296e-02
9.41527367e-01 3.94070238e-01 -5.72807610e-01 -1.41784221e-01
-1.56046763e-01 -4.97070700e-01 3.68695468e-01 -1.14559782e+00
-1.12411976e+00 4.92962986e-01 1.01047647e+00 7.03072846e-01
1.41615725e+00 -2.21085563e-01 7.09934175e-01 2.74236917e-01
8.96704793e-01 -8.11204195e-01 -1.13764152e-01 -1.35439858e-01
6.41856551e-01 -4.85553056e-01 1.88476965e-01 -2.97053069e-01
4.43227082e-01 1.01639736e+00 2.42361695e-01 2.67841679e-04
1.08092058e+00 5.34288764e-01 1.95534468e-01 -3.51419151e-01
-4.87101108e-01 2.23760471e-01 1.21711511e-02 5.16747415e-01
2.80796617e-01 3.80077474e-02 -2.17074379e-01 7.43705332e-01
-3.94852869e-02 -2.09092304e-01 4.86958802e-01 1.21805608e+00
-5.56367636e-01 -8.18978131e-01 -4.84621912e-01 6.63412333e-01
-6.80928171e-01 5.15258074e-01 3.72178495e-01 8.29647303e-01
1.34813353e-01 1.21686316e+00 2.79248565e-01 -3.02806169e-01
5.87010443e-01 -9.01719853e-02 6.26146257e-01 -3.04373980e-01
-3.02682161e-01 2.24139407e-01 5.61626613e-01 -6.54596984e-01
3.88939716e-02 -6.19497478e-01 -1.01167428e+00 -1.59563765e-01
-4.65353221e-01 2.74382621e-01 8.55008483e-01 9.15202677e-01
6.20940685e-01 9.87272024e-01 1.03745520e+00 -1.54788718e-01
-5.05100131e-01 -1.04905403e+00 -1.10393012e+00 -3.95178586e-01
3.53147030e-01 -9.13611948e-01 -5.69422424e-01 -8.37789595e-01] | [6.398812770843506, 1.316398024559021] |
90eeeb11-bf58-49a1-81b4-52185e320233 | aligning-sequence-reads-clone-sequences-and | 1303.3997 | null | http://arxiv.org/abs/1303.3997v2 | http://arxiv.org/pdf/1303.3997v2.pdf | Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM | Summary: BWA-MEM is a new alignment algorithm for aligning sequence reads or
long query sequences against a large reference genome such as human. It
automatically chooses between local and end-to-end alignments, supports
paired-end reads and performs chimeric alignment. The algorithm is robust to
sequencing errors and applicable to a wide range of sequence lengths from 70bp
to a few megabases. For mapping 100bp sequences, BWA-MEM shows better
performance than several state-of-art read aligners to date.
Availability and implementation: BWA-MEM is implemented as a component of
BWA, which is available at http://github.com/lh3/bwa.
Contact: hengli@broadinstitute.org | ['Heng Li'] | 2013-03-16 | null | null | null | null | ['3d-object-reconstruction-from-a-single-image'] | ['computer-vision'] | [ 5.80386579e-01 -5.64438164e-01 -4.43085968e-01 -3.47090721e-01
-1.17217636e+00 -1.17189300e+00 2.56697148e-01 3.28325748e-01
-2.92169750e-01 1.13496125e+00 2.78435349e-01 -6.61317110e-01
1.33251594e-02 -8.13162029e-01 -5.89550734e-01 -6.84082925e-01
2.65840411e-01 8.85500014e-01 3.45289499e-01 -5.44774607e-02
5.52296698e-01 3.69374722e-01 -1.09177458e+00 4.12103474e-01
7.42809892e-01 -2.45065745e-02 5.27468920e-01 1.18531215e+00
-3.51219803e-01 -1.92164421e-01 -3.79303128e-01 -2.75251895e-01
1.16449639e-01 -8.05654883e-01 -1.00560522e+00 -8.87807190e-01
2.36807674e-01 -1.23234726e-01 1.48703039e-01 6.42567277e-01
1.08188426e+00 -3.71232539e-01 3.55062693e-01 -8.50653172e-01
-4.87924129e-01 3.78727645e-01 -5.16463757e-01 7.60185719e-01
5.96365333e-01 4.87506151e-01 1.12939060e+00 -9.62530077e-01
1.03075778e+00 9.15835321e-01 8.74659121e-01 3.63736808e-01
-1.45328593e+00 -5.80118835e-01 -7.23954141e-01 4.55301777e-02
-1.41972542e+00 -4.72409666e-01 -3.91486734e-01 -6.50866926e-01
1.59156966e+00 5.59636950e-01 6.74804449e-01 8.79473031e-01
7.80343354e-01 2.01566562e-01 9.69803631e-01 -2.03356326e-01
1.87465146e-01 -1.01095676e+00 4.32705611e-01 3.49719316e-01
4.01380986e-01 -4.22735102e-02 -7.16042221e-01 -8.12467098e-01
4.38421398e-01 -7.34013170e-02 -1.23417489e-01 1.64418280e-01
-1.67444849e+00 6.58209205e-01 -1.59379423e-01 2.26465166e-01
-6.01249993e-01 7.05305710e-02 6.07539058e-01 2.07714483e-01
3.27310801e-01 4.01568890e-01 -3.20092738e-01 -7.50597358e-01
-7.56564379e-01 2.08075136e-01 5.15789747e-01 9.78079379e-01
6.92507505e-01 -5.46479642e-01 -5.23664765e-02 9.11130011e-01
-1.48832679e-01 6.05580747e-01 2.86292553e-01 -6.52296603e-01
-7.24337995e-02 9.34832469e-02 2.70688444e-01 -5.46339333e-01
-6.34940922e-01 -6.04000837e-02 -6.65834725e-01 -2.61145234e-01
4.92512882e-01 1.65924728e-02 -6.70752108e-01 1.50129855e+00
5.77338755e-01 2.43985549e-01 -2.46913750e-02 8.21200728e-01
7.11237013e-01 7.59860218e-01 -2.14133188e-01 -3.05100083e-01
1.61582410e+00 -8.28958929e-01 -5.99453568e-01 -2.76389241e-01
3.18986833e-01 -1.28874719e+00 6.83943510e-01 1.21721357e-01
-1.15741420e+00 -2.81878322e-01 -9.37646210e-01 -2.19059438e-01
-1.40264094e-01 -4.23645228e-01 9.71131474e-02 3.98539215e-01
-1.08461511e+00 3.77728820e-01 -9.78482127e-01 -1.22014260e+00
-6.95784241e-02 4.00331207e-02 -4.74865139e-01 -3.36893767e-01
-9.25148964e-01 9.95806336e-01 6.20280921e-01 -1.79321423e-01
-6.68174267e-01 -5.21167934e-01 -1.72030747e-01 -3.47008288e-01
-1.69660106e-01 -1.27190709e+00 1.03718293e+00 -2.31849477e-01
-1.36389160e+00 1.29065573e+00 -6.55114591e-01 -1.44941539e-01
1.36414751e-01 -3.87632519e-01 -2.82332242e-01 9.51895267e-02
2.26655900e-01 2.45767459e-01 -7.03235418e-02 -3.20070267e-01
-5.33626199e-01 -4.42128688e-01 -6.18041277e-01 -1.07696481e-01
5.50384343e-01 5.79878926e-01 -1.21371239e-01 -5.54682791e-01
7.87649080e-02 -9.60069716e-01 -2.47321486e-01 -2.19342053e-01
-3.01881105e-01 7.64141679e-02 1.11229368e-01 -1.00585043e+00
9.48690355e-01 -1.79888022e+00 3.06063324e-01 7.44599402e-02
-2.49102205e-01 4.49990273e-01 -6.63440883e-01 1.38898659e+00
-2.09144324e-01 9.07132551e-02 -2.91930616e-01 4.90921021e-01
-2.17662185e-01 -3.20870318e-02 -1.25103295e-01 9.27215397e-01
-6.97770938e-02 9.39755797e-01 -1.37581646e+00 -9.16603804e-02
1.21038809e-01 4.66613382e-01 -5.96280061e-02 4.90240663e-01
-6.58928975e-02 6.62792146e-01 -2.34544367e-01 6.39074087e-01
9.00945306e-01 -4.24622625e-01 5.83306849e-01 1.64170265e-01
-5.64902484e-01 6.01373434e-01 -7.26038218e-01 2.09862161e+00
1.20104134e-01 3.93541664e-01 -4.63484861e-02 -4.75271374e-01
1.12259471e+00 1.64151803e-01 1.95332304e-01 -1.43117562e-01
-1.61399655e-02 4.32276666e-01 1.37310028e-01 -4.14709717e-01
2.20170408e-01 -3.11415315e-01 2.37446755e-01 5.97170472e-01
-1.04661293e-01 -1.65645152e-01 3.72434318e-01 -5.43754082e-03
1.42908776e+00 2.09311023e-01 9.99218702e-01 -5.78089595e-01
3.51357847e-01 4.47067857e-01 8.38032424e-01 9.39063847e-01
-6.42119795e-02 6.21508360e-01 -9.38575491e-02 -5.76716244e-01
-1.77180791e+00 -1.34324610e+00 -4.05129820e-01 1.19851542e+00
-9.21853483e-02 -4.81677771e-01 -4.70503956e-01 2.25351349e-01
-2.24718489e-02 7.01728404e-01 -2.04261795e-01 1.78665429e-01
-4.66416061e-01 -1.10080838e+00 1.09616363e+00 3.12893391e-01
-1.47317033e-02 -7.11216092e-01 -4.16911304e-01 4.72307593e-01
-7.32838333e-01 -4.60106105e-01 -7.75874794e-01 2.53686905e-01
-4.85405415e-01 -1.15916312e+00 -6.45363629e-01 -6.04534626e-01
2.03275636e-01 2.55568415e-01 1.16954732e+00 1.56067282e-01
-8.41057062e-01 -1.32509857e-01 -4.33668047e-01 -4.83206570e-01
-5.85150540e-01 3.08082193e-01 1.25854135e-01 -7.92716801e-01
7.08529651e-01 -5.92230201e-01 -7.31376946e-01 5.07958889e-01
-6.07542872e-01 2.84882545e-01 4.55894619e-01 1.13291109e+00
8.79828930e-01 -8.14842582e-01 7.63309956e-01 -6.12800837e-01
4.20372754e-01 -5.64961731e-01 -6.25410914e-01 5.41328967e-01
-3.58752608e-01 -2.48306170e-01 4.34682518e-01 7.89262578e-02
-7.96663046e-01 -1.00591600e-01 -7.42016673e-01 6.90379798e-01
-1.98448300e-01 5.44564128e-01 2.30815727e-02 3.43760610e-01
7.20500648e-01 6.68516517e-01 4.62802052e-01 -3.12246829e-01
2.03533560e-01 1.03313005e+00 6.46543443e-01 -4.64559704e-01
3.87765676e-01 3.84611815e-01 -4.82078455e-02 -7.33115435e-01
-4.78690803e-01 -1.15537679e+00 -7.67609715e-01 1.50763810e-01
8.93553913e-01 -7.00391829e-01 -8.01052451e-01 6.99611366e-01
-8.15871477e-01 -3.28301787e-01 2.86576331e-01 4.66717690e-01
-6.49407506e-01 5.07544637e-01 -1.14436650e+00 -2.41983846e-01
-9.97502744e-01 -9.22225058e-01 9.77950871e-01 8.97554383e-02
-6.25317693e-01 -5.40374339e-01 9.11654413e-01 2.22645834e-01
4.65410233e-01 3.03360611e-01 7.23922253e-01 -8.01068544e-01
-2.33460277e-01 -7.13404119e-02 -1.44158863e-03 -2.55062222e-01
3.88598323e-01 4.70419198e-01 -4.36463445e-01 -5.97708285e-01
-5.47287226e-01 -1.76699415e-01 6.36706829e-01 3.19384962e-01
5.53291082e-01 -8.23528841e-02 -5.48892379e-01 6.76295698e-01
1.46609235e+00 2.74508059e-01 8.49210024e-01 6.25459433e-01
1.65397778e-01 2.23392665e-01 1.06717777e+00 6.69189632e-01
-7.05908760e-02 6.57305062e-01 1.93743318e-01 1.17317863e-01
1.26739055e-01 -3.25391418e-03 -3.81248333e-02 8.93273771e-01
-1.20905831e-01 -5.03604949e-01 -1.38045824e+00 4.48691279e-01
-1.75718939e+00 -1.33107519e+00 -5.14448881e-01 2.40324616e+00
9.17051017e-01 -4.17565495e-01 4.39569026e-01 -5.86158991e-01
1.07442105e+00 2.92709023e-02 -7.15095341e-01 -5.18828452e-01
-4.33906406e-01 3.23465079e-01 6.77159548e-01 6.00094199e-01
-5.30502200e-01 6.82890773e-01 7.60223436e+00 5.89370906e-01
-8.96324813e-01 9.33866890e-04 2.85306543e-01 -2.07056552e-01
-3.86131525e-01 7.81583786e-02 -7.14313865e-01 5.71176291e-01
1.22935271e+00 -7.90949881e-01 4.54381019e-01 4.39199805e-01
2.62564600e-01 8.48696232e-02 -9.27125633e-01 6.04451597e-01
-2.44326979e-01 -1.82955992e+00 -2.92989880e-01 1.31921574e-01
3.86504829e-01 8.90715361e-01 -5.57353079e-01 -4.64068055e-01
7.13203669e-01 -1.13879657e+00 2.63766259e-01 4.45485085e-01
9.66209650e-01 -8.32872391e-01 1.10191000e+00 4.10520405e-01
-9.39921081e-01 5.05384982e-01 -8.92008245e-01 -5.93797266e-02
5.86346328e-01 1.01974785e+00 -1.16016340e+00 7.40826249e-01
7.45351076e-01 4.97152448e-01 -9.13772583e-02 1.20957482e+00
-5.67070395e-02 8.38286459e-01 -3.37057978e-01 -2.92453263e-02
-8.15405473e-02 -5.06379843e-01 4.90124613e-01 1.64903593e+00
6.51723504e-01 4.78492796e-01 1.41799852e-01 2.90778637e-01
2.61779666e-01 2.78844297e-01 -3.92194837e-01 -2.77255356e-01
9.82034743e-01 9.51564431e-01 -7.03229070e-01 -4.56722677e-01
-2.59538203e-01 1.06984842e+00 4.87764448e-01 1.67564660e-01
-6.13408327e-01 -5.25736451e-01 1.20911622e+00 -3.75233926e-02
2.26828426e-01 -7.39267142e-03 1.28812805e-01 -6.76113129e-01
-3.95159811e-01 -1.06272304e+00 8.98890853e-01 -8.06539357e-01
-1.55670643e+00 1.94961309e-01 -4.17141080e-01 -9.52344716e-01
-2.67976969e-01 -3.53389651e-01 -5.31792879e-01 1.31389594e+00
-8.25683713e-01 -9.48450744e-01 -3.03560734e-01 -9.40611809e-02
2.41397053e-01 1.44888133e-01 1.11702490e+00 -4.87629585e-02
-5.14591694e-01 2.83004373e-01 8.85021925e-01 -1.82798728e-01
1.11319745e+00 -8.76226902e-01 1.35723054e+00 7.58658588e-01
-4.99576747e-01 1.10925078e+00 1.01510859e+00 -8.90201569e-01
-1.59851754e+00 -1.00226295e+00 1.07715631e+00 -3.55714351e-01
5.74396491e-01 -2.04232320e-01 -9.02132928e-01 1.05354273e+00
5.27784884e-01 -2.46474370e-01 1.52024078e+00 4.74251769e-02
-5.15294850e-01 2.39539251e-01 -1.18556547e+00 3.72120947e-01
1.08565986e+00 -5.54393172e-01 -6.12403095e-01 5.15867651e-01
3.50350678e-01 -5.27749956e-01 -1.29371607e+00 1.44205034e-01
1.15748441e+00 -9.45348382e-01 1.23636198e+00 -6.41435981e-01
2.65798062e-01 -7.44505525e-01 -7.23768175e-02 -1.34568632e+00
-1.01097560e+00 -6.42047465e-01 7.31425643e-01 1.05905306e+00
2.55858541e-01 -1.02701092e+00 2.04036534e-01 -4.60250407e-01
-3.76869291e-01 -2.41366670e-01 -1.02928066e+00 -1.02100670e+00
2.55379766e-01 3.90188873e-01 9.73742723e-01 1.06061125e+00
5.44056118e-01 2.44695783e-01 -4.38650638e-01 1.59290865e-01
5.72449505e-01 3.62395257e-01 9.06574547e-01 -7.68786490e-01
-4.63733554e-01 -3.28258961e-01 -4.91345584e-01 -8.63755643e-01
-2.40959167e-01 -1.07345510e+00 4.22242701e-01 -1.81193495e+00
7.31804430e-01 -1.27560258e-01 -2.24126771e-01 5.37140965e-01
-4.95479822e-01 4.35628563e-01 -4.01044965e-01 1.58898354e-01
-1.43425688e-01 -2.40877494e-01 6.77798390e-01 6.50995076e-02
3.62748355e-01 -2.24945366e-01 -2.46682033e-01 9.12954658e-02
1.05243683e+00 -5.98560631e-01 1.74674332e-01 -2.36040801e-01
3.35374236e-01 1.18159354e-01 -1.01994254e-01 -4.49390233e-01
-1.08153373e-01 -8.25165987e-01 1.46061540e-01 -8.78270447e-01
-1.85287613e-02 -1.15637839e-01 1.28169405e+00 9.02433336e-01
-2.80480921e-01 7.25340903e-01 1.81392372e-01 2.19191104e-01
1.02655932e-01 -2.28721172e-01 8.00175548e-01 -2.77850688e-01
-5.28091550e-01 -5.37261218e-02 -6.61610425e-01 -2.76760682e-02
9.43727076e-01 -3.29233170e-01 -1.12573087e+00 1.41692206e-01
-2.95910150e-01 1.31426200e-01 1.13761413e+00 2.73686916e-01
2.32121110e-01 -9.15370584e-01 -1.35350645e+00 -7.47869536e-02
6.63282812e-01 -5.20201862e-01 3.71373355e-01 9.75031257e-01
-1.39335704e+00 4.83964503e-01 -7.25246072e-01 -8.84951532e-01
-1.75020909e+00 7.19902217e-01 7.52045140e-02 2.02199891e-01
-6.86582625e-01 9.36904788e-01 -1.02975368e-01 -8.13471973e-01
-3.57133597e-01 3.52155685e-01 5.34918368e-01 -3.50301802e-01
1.11737704e+00 5.55319905e-01 2.47224942e-02 -6.95550025e-01
-8.56175423e-01 5.30108392e-01 1.68760613e-01 2.12257460e-01
1.30983162e+00 -2.66206443e-01 -8.85939360e-01 6.31235957e-01
1.03453588e+00 3.69711965e-02 -3.64442021e-01 1.60315916e-01
1.28885970e-01 -8.01376164e-01 -7.50070810e-01 -7.86525965e-01
-1.36374503e-01 5.57219148e-01 4.51553345e-01 -6.44048989e-01
8.30003202e-01 -2.30078697e-02 7.33733714e-01 2.52858430e-01
4.62250948e-01 -6.46806300e-01 -3.20588857e-01 5.45658469e-01
6.20736241e-01 -7.36486316e-01 4.03802432e-02 -4.24871504e-01
1.24451667e-01 1.05718684e+00 2.30613410e-01 3.80428702e-01
-1.85210913e-01 6.34752631e-01 3.59389067e-01 -7.15956977e-03
-9.57073629e-01 5.44966534e-02 -3.88649225e-01 6.59070611e-01
9.31373775e-01 2.70597607e-01 -1.04835927e+00 -1.48855492e-01
-3.23718458e-01 1.53129905e-01 6.33866429e-01 1.13347292e+00
-7.05438435e-01 -1.47510660e+00 -5.03885925e-01 5.39981008e-01
-5.31080365e-01 -3.72772306e-01 -3.46414864e-01 2.95932949e-01
-3.31797272e-01 9.20409739e-01 1.86409622e-01 -2.13466302e-01
-9.98731777e-02 3.36306363e-01 3.21007937e-01 -4.71625388e-01
-7.81731486e-01 2.61174411e-01 5.99343181e-01 -5.52718043e-01
-2.84104556e-01 -9.66844738e-01 -1.29941070e+00 -8.59026968e-01
-4.72837865e-01 4.58502024e-01 4.89236772e-01 5.05345523e-01
6.46046400e-01 1.48562819e-01 1.91346243e-01 -3.90661329e-01
-3.95864248e-01 -9.34333026e-01 -3.85222942e-01 1.45659596e-01
2.25187019e-01 2.03031655e-02 -2.33341660e-03 -7.74428695e-02] | [4.877599716186523, 5.172602653503418] |
5306a947-2a8f-4cd4-92f5-05fd2f1249b2 | a-new-citation-recommendation-strategy-based | 2009.08948 | null | https://arxiv.org/abs/2009.08948v2 | https://arxiv.org/pdf/2009.08948v2.pdf | A New Citation Recommendation Strategy Based on Term Functions in Related Studies Section | Purpose: Researchers frequently encounter the following problems when writing scientific articles: (1) Selecting appropriate citations to support the research idea is challenging. (2) The literature review is not conducted extensively, which leads to working on a research problem that others have well addressed. This study focuses on citation recommendation in the related studies section by applying the term function of a citation context, potentially improving the efficiency of writing a literature review. Design/methodology/approach: We present nine term functions with three newly created and six identified from existing literature. Using these term functions as labels, we annotate 531 research papers in three topics to evaluate our proposed recommendation strategy. BM25 and Word2vec with VSM are implemented as the baseline models for the recommendation. Then the term function information is applied to enhance the performance. Findings: The experiments show that the term function-based methods outperform the baseline methods regarding the recall, precision, and F1-score measurement, demonstrating that term functions are useful in identifying valuable citations. Research limitations: The dataset is insufficient due to the complexity of annotating citation functions for paragraphs in the related studies section. More recent deep learning models should be performed to future validate the proposed approach. Practical implications: The citation recommendation strategy can be helpful for valuable citation discovery, semantic scientific retrieval, and automatic literature review generation. Originality/value: The proposed citation function-based citation recommendation can generate intuitive explanations of the results for users, improving the transparency, persuasiveness, and effectiveness of recommender systems. | ['Haihua Chen'] | 2020-09-11 | null | null | null | null | ['review-generation'] | ['natural-language-processing'] | [-2.28990152e-01 -1.85869500e-01 -8.05200934e-01 7.14230016e-02
-5.66541672e-01 -5.65728188e-01 5.87792695e-01 1.61135912e-01
-3.19001138e-01 8.18763256e-01 5.12117743e-01 -7.71430910e-01
-6.74450874e-01 -7.49302328e-01 -5.42991340e-01 -4.90503639e-01
6.83046043e-01 1.38705865e-01 -2.41510302e-01 3.46872360e-02
1.42443836e+00 5.49267352e-01 -1.65786874e+00 6.18239045e-02
1.15773618e+00 6.13958716e-01 7.62528598e-01 1.01217136e-01
-8.64826858e-01 1.17302693e-01 -1.08795857e+00 -3.34896863e-01
-1.47087993e-02 -2.38091856e-01 -7.27710962e-01 -3.83397728e-01
2.51338869e-01 3.40588130e-02 -9.55741704e-02 1.09135640e+00
6.00813746e-01 5.37604153e-01 8.83033156e-01 -9.17440355e-01
-1.69566667e+00 8.08460712e-01 -7.39154875e-01 4.35856998e-01
2.65723616e-01 -3.27619433e-01 1.09580445e+00 -1.03219199e+00
6.41257107e-01 1.34990609e+00 2.79228896e-01 5.32118976e-01
-6.40496612e-01 -1.18796003e+00 3.71683657e-01 4.81568754e-01
-1.08604991e+00 -8.44918471e-03 7.48232901e-01 -7.24628627e-01
6.82113409e-01 2.52302140e-01 5.03674150e-01 1.44252050e+00
2.91658103e-01 2.64835268e-01 1.02163482e+00 -3.67915779e-01
1.42622545e-01 1.85369521e-01 7.54302919e-01 1.12856887e-01
8.95326436e-01 -1.25765011e-01 -2.60175079e-01 -7.79695511e-02
6.88528419e-01 3.61098230e-01 -4.24650699e-01 5.84389567e-01
-1.23153281e+00 9.80142713e-01 4.76570874e-01 6.53019071e-01
-5.68897188e-01 -1.74183935e-01 4.32188541e-01 3.98521125e-02
3.10936004e-01 1.12375152e+00 -2.07458615e-01 3.15799303e-02
-8.76737475e-01 3.13759178e-01 5.46241283e-01 9.84261811e-01
2.93801218e-01 2.57771686e-02 -7.02282667e-01 9.20724392e-01
4.87909853e-01 6.31305695e-01 8.43326211e-01 -1.05083930e+00
1.16358891e-01 6.97646916e-01 1.24922521e-01 -1.35748374e+00
-2.56251901e-01 -9.58647192e-01 -5.92531025e-01 -2.09881455e-01
-1.73580348e-01 -1.21178403e-01 -5.78732610e-01 1.19082212e+00
-3.47212851e-01 1.86319530e-01 -6.73356503e-02 1.00949657e+00
1.75385273e+00 7.83461392e-01 4.12491530e-01 -4.53169733e-01
1.37642109e+00 -1.11043453e+00 -1.31930375e+00 1.16592519e-01
4.69945133e-01 -1.03508639e+00 1.11605799e+00 2.37370983e-01
-7.48996854e-01 -6.64376736e-01 -8.99050474e-01 9.19544324e-02
-7.12260842e-01 4.92706209e-01 7.95370340e-01 4.22953784e-01
-7.95193315e-01 4.49669570e-01 -1.58432573e-02 -2.57470101e-01
5.35054982e-01 3.61769572e-02 1.07857063e-01 -1.36642575e-01
-1.34142017e+00 1.17275929e+00 5.97334132e-02 -2.09475737e-02
-2.01383486e-01 -8.67808342e-01 -3.40778738e-01 3.72357875e-01
6.34322315e-02 -9.99507666e-01 8.13349724e-01 -7.18204796e-01
-1.13775289e+00 3.73641163e-01 -1.97217286e-01 1.21306404e-01
-1.95790246e-01 -2.82959938e-01 -6.15752697e-01 -1.03075854e-01
5.34981012e-01 4.24183875e-01 3.95548731e-01 -1.32313299e+00
-7.00046062e-01 -2.29148135e-01 1.00995071e-01 1.86119780e-01
-5.38235128e-01 7.85046220e-02 -6.15662575e-01 -1.08304965e+00
-5.27547538e-01 -6.27612531e-01 -8.40124562e-02 -3.24694902e-01
-8.36334154e-02 -7.37625778e-01 5.64029992e-01 -6.29269361e-01
1.78161979e+00 -1.67118204e+00 -4.08301093e-02 1.82886481e-01
1.80523962e-01 4.99135852e-01 -4.46574897e-01 2.68319100e-01
4.28879298e-02 8.19744587e-01 3.12168270e-01 2.14102313e-01
-1.33690640e-01 -1.19072415e-01 -3.89585853e-01 -1.89358499e-02
-3.48101586e-01 1.03626323e+00 -1.00580382e+00 -2.52706051e-01
5.02336398e-03 4.77466941e-01 -6.38648793e-02 1.77575380e-01
7.33774900e-02 6.79017529e-02 -9.36871409e-01 6.47432148e-01
5.93082845e-01 -4.27625090e-01 -3.23277652e-01 -2.82916933e-01
-2.02028304e-01 4.11377460e-01 -7.84910381e-01 1.39006376e+00
-6.37185931e-01 7.09089696e-01 -5.10986149e-01 -9.42140400e-01
1.21210563e+00 1.36593267e-01 3.00113022e-01 -8.95937145e-01
2.16230512e-01 3.57791692e-01 -3.60881165e-02 -9.27253246e-01
2.98565388e-01 1.46628156e-01 4.24636185e-01 4.72033918e-01
-2.59427279e-01 1.29876852e-01 3.31743300e-01 2.13618517e-01
7.24045932e-01 2.48329490e-01 3.25263999e-02 -3.63112599e-01
5.14106095e-01 -5.22584319e-02 3.35264862e-01 1.02292776e+00
3.11091363e-01 3.29748571e-01 8.35799798e-02 -5.29358052e-02
-4.64172512e-01 -3.82905751e-01 -2.41439447e-01 9.71873581e-01
1.58811226e-01 -2.41941795e-01 -5.09564638e-01 -6.96009040e-01
1.15857214e-01 1.21850908e+00 -7.23095834e-01 -1.67410865e-01
-6.92091733e-02 -5.51290810e-01 2.46995881e-01 6.62710845e-01
4.20294926e-02 -1.14547980e+00 -2.54252732e-01 8.77292156e-02
-2.15019405e-01 -4.22179759e-01 -6.54024720e-01 -3.91057670e-01
-9.08989310e-01 -1.09438813e+00 -1.16422665e+00 -6.69327617e-01
6.69004023e-01 9.11790013e-01 8.81573260e-01 4.76336986e-01
1.04862303e-01 4.68144268e-01 -5.16204298e-01 -6.27875209e-01
9.80109945e-02 1.74547788e-02 -1.63995028e-01 -6.69404507e-01
8.45432997e-01 -8.93934742e-02 -6.93744004e-01 2.24578828e-02
-8.09327483e-01 -2.66706973e-01 7.65515089e-01 8.56924355e-01
3.65750760e-01 -2.35699296e-01 1.18601406e+00 -8.20059419e-01
1.43114209e+00 -8.19370389e-01 -3.76519293e-01 2.73921311e-01
-1.46640372e+00 -8.22119415e-02 3.63091946e-01 -5.75248659e-01
-1.08306527e+00 -1.09876680e+00 4.55722921e-02 -3.75521630e-01
8.73126648e-03 1.03249729e+00 7.31821209e-02 -1.38288885e-01
5.68290293e-01 -2.12623164e-01 3.77023965e-03 -6.19973779e-01
3.12493742e-01 8.73927176e-01 -7.23880157e-02 -7.64289856e-01
3.92157137e-01 -2.14471802e-01 6.06800951e-02 -4.75726038e-01
-9.79493618e-01 -6.05580866e-01 -1.63982417e-02 -2.12522730e-01
5.26759088e-01 -6.30677104e-01 -8.65781486e-01 -5.91100931e-01
-1.55938232e+00 6.70602500e-01 1.82051688e-01 1.00141776e+00
3.39069635e-01 4.09554660e-01 -3.07084024e-01 -4.59491491e-01
-8.14343452e-01 -1.41017079e+00 7.47931838e-01 6.77624822e-01
-3.81800860e-01 -1.02345479e+00 -5.83865903e-02 6.73288763e-01
7.36374736e-01 -3.58051658e-01 1.15536332e+00 -8.56261194e-01
-1.18722506e-01 -5.04958481e-02 -5.10391712e-01 8.38235319e-02
2.59102196e-01 2.07574412e-01 -7.34512031e-01 1.30903274e-01
-2.50078350e-01 4.09960717e-01 1.03307486e+00 6.64094329e-01
1.61366284e+00 -5.68539739e-01 -5.39743125e-01 3.33314657e-01
1.10378456e+00 7.51742899e-01 5.27079582e-01 8.64938974e-01
7.87363112e-01 8.41279268e-01 7.60977864e-01 1.91428438e-01
2.89901823e-01 2.94736207e-01 4.96013574e-02 1.87577173e-01
-2.25323215e-01 1.92171410e-01 2.18428280e-02 9.82864439e-01
-4.72559512e-01 -4.63578433e-01 -7.90812075e-01 5.77058613e-01
-1.71487844e+00 -9.85648096e-01 -3.73253703e-01 2.04492068e+00
6.05548322e-01 -1.21539727e-01 -3.03769797e-01 -3.10649991e-01
7.92664170e-01 -2.12528795e-01 -2.76565164e-01 -6.84490263e-01
-1.62975848e-01 3.49286795e-01 4.10615832e-01 2.52270877e-01
-6.02407098e-01 5.48455238e-01 5.63000774e+00 8.94414186e-01
-1.39043570e+00 -7.00711310e-02 4.54198360e-01 2.18567774e-01
-1.04554689e+00 1.84892304e-02 -7.09685087e-01 7.08816469e-01
7.82275200e-01 -8.17515135e-01 1.40154213e-01 8.53590369e-01
4.59333777e-01 4.24187779e-01 -6.44483030e-01 1.08178830e+00
1.95164576e-01 -1.91857147e+00 6.89947188e-01 -4.71712928e-03
6.77107811e-01 -3.88167083e-01 2.21262008e-01 5.55900931e-01
-9.54402387e-02 -1.06979477e+00 1.12762727e-01 8.43126833e-01
4.98115927e-01 -6.84670806e-01 1.08019018e+00 -2.79735774e-02
-4.83860016e-01 -1.46865454e-02 -5.77616811e-01 -1.64018478e-02
-4.07693624e-01 5.46574533e-01 -5.21458089e-01 7.02984393e-01
5.98377824e-01 1.00610113e+00 -6.58099055e-01 1.24309993e+00
-2.20886603e-01 9.11321640e-01 4.23049510e-01 -5.44228852e-01
2.04074591e-01 -4.88920808e-01 4.78307188e-01 1.49159408e+00
8.32651854e-01 1.02485165e-01 -2.53174722e-01 1.27801275e+00
-3.55034739e-01 5.81746936e-01 -6.47168279e-01 -4.02165860e-01
8.16664100e-01 1.28892028e+00 -6.26327276e-01 -2.88959682e-01
-6.47285700e-01 3.08128744e-01 -2.21263170e-01 6.51131809e-01
-4.74619150e-01 -7.05807328e-01 1.91056937e-01 1.17338039e-02
2.54988939e-01 3.64199817e-01 -8.21482122e-01 -1.03556359e+00
-1.98410809e-01 -7.26204872e-01 2.17755213e-01 -8.99441481e-01
-1.62040317e+00 3.08917135e-01 -4.78951912e-03 -1.20913923e+00
3.29917341e-01 -5.81369996e-01 -8.30754519e-01 1.23130858e+00
-1.44343495e+00 -1.05597019e+00 -2.37904891e-01 -1.76123440e-01
7.87616193e-01 -4.86163437e-01 8.50793421e-01 4.17267352e-01
-7.12085009e-01 4.15034682e-01 2.81624377e-01 -3.98021936e-01
9.74077463e-01 -1.03517723e+00 -1.83924660e-01 6.31349385e-01
-1.07669540e-01 1.57498360e+00 4.25937057e-01 -9.15118217e-01
-1.50518715e+00 -8.33424866e-01 1.07705796e+00 -2.20146552e-01
5.78908861e-01 5.47552049e-01 -1.21989381e+00 1.38630345e-01
5.23355424e-01 -6.89581871e-01 1.22843623e+00 3.69410157e-01
-2.21776068e-01 -8.21170136e-02 -8.86079788e-01 7.96431005e-01
5.82992494e-01 -2.26126164e-02 -1.06531179e+00 1.83728054e-01
8.13018799e-01 1.25010103e-01 -8.86177897e-01 3.05849344e-01
8.04473698e-01 -5.52002117e-02 1.06141007e+00 -6.30710542e-01
8.24489176e-01 -2.32656851e-01 1.66247651e-01 -1.34948158e+00
-9.40641820e-01 -2.64774680e-01 -1.82651982e-01 1.43870997e+00
5.21235287e-01 -5.74795306e-01 1.27596885e-01 6.47118509e-01
-1.51778728e-01 -8.51513445e-01 -4.58040655e-01 -3.81170064e-01
3.31020623e-01 -1.35342240e-01 9.14814949e-01 1.38354039e+00
1.17284907e-02 8.25037181e-01 1.89563766e-01 -2.24758536e-01
3.64167780e-01 4.85190123e-01 3.57749999e-01 -1.60248876e+00
2.15454951e-01 -1.32469916e+00 1.55602798e-01 -9.37155366e-01
5.30970991e-01 -9.93587255e-01 -5.65782368e-01 -2.19214630e+00
3.53098691e-01 -3.59523773e-01 -8.73328269e-01 2.97800958e-01
-7.59411991e-01 -2.79968858e-01 -1.60204574e-01 3.55154872e-01
-2.39139795e-01 4.47742611e-01 1.68890262e+00 -4.27293748e-01
6.55493662e-02 2.63793431e-02 -1.61281514e+00 4.12123799e-01
5.16024649e-01 -4.14501786e-01 -4.67783749e-01 -4.34604466e-01
3.50679040e-01 3.12576210e-03 7.03292415e-02 -2.47605503e-01
3.10905188e-01 -7.28988230e-01 3.78556579e-01 -5.07476330e-01
-3.33233058e-01 -6.21963203e-01 -8.04625824e-02 -3.33092175e-02
-6.84103310e-01 3.09980541e-01 4.64046210e-01 4.57400292e-01
8.04919153e-02 -8.10700595e-01 4.16050851e-02 -6.31824061e-02
-5.86423695e-01 1.00131053e-02 -3.96162510e-01 -2.57325321e-01
5.96168399e-01 -1.44267961e-01 -4.68012244e-01 -2.90196598e-01
-1.98802631e-02 1.47270799e-01 1.77559137e-01 1.10469210e+00
8.55179489e-01 -1.29082751e+00 -7.00587034e-01 -5.13366222e-01
1.90640986e-01 -4.17056561e-01 3.41027342e-02 6.23834789e-01
-1.85357392e-01 7.49520779e-01 -2.96005338e-01 1.48579389e-01
-9.93385136e-01 8.00616503e-01 -3.13214332e-01 1.76917270e-01
-3.89410436e-01 5.90848207e-01 3.05242419e-01 -6.98128194e-02
4.96051759e-01 -1.32263839e-01 -1.26611936e+00 2.34105632e-01
6.45352423e-01 7.89249361e-01 -2.09333897e-02 -5.44495225e-01
-2.93287486e-01 7.02980280e-01 -1.54190585e-01 1.04679897e-01
1.31318140e+00 -5.85164614e-02 -4.42525685e-01 3.86996478e-01
9.70809400e-01 3.46474379e-01 8.97502229e-02 5.24837710e-02
9.42796916e-02 -5.79902351e-01 5.23850322e-01 -1.16854000e+00
-1.15564549e+00 9.12231445e-01 5.50919533e-01 -1.89994276e-01
8.46402287e-01 -2.91858941e-01 6.32305861e-01 5.47226250e-01
-7.20570758e-02 -1.08484387e+00 -3.35297376e-01 3.33615094e-01
1.15923429e+00 -1.04944849e+00 2.50106096e-01 2.84723714e-02
-5.42733252e-01 1.54859567e+00 6.39427483e-01 2.91680336e-01
6.04402542e-01 -2.10605860e-01 1.10096507e-01 -4.80857670e-01
-5.43510377e-01 6.66260943e-02 9.36841607e-01 3.41981322e-01
1.16143620e+00 2.10035108e-02 -1.38532817e+00 1.28107965e+00
2.57920533e-01 9.52845514e-02 4.70560104e-01 5.29213727e-01
-3.41627717e-01 -1.16250610e+00 -4.58540618e-01 1.02935958e+00
-8.46916378e-01 -3.16314340e-01 -4.67313498e-01 3.21734428e-01
2.37565581e-02 1.20175266e+00 -3.89810175e-01 -2.13832483e-01
2.49978617e-01 -2.16381490e-01 3.47676873e-02 -7.59565175e-01
-7.58381665e-01 2.19060868e-01 -7.19344318e-02 2.57723406e-02
-5.53817689e-01 -1.75267458e-01 -1.26702034e+00 -1.04376689e-01
-8.84646356e-01 7.23870933e-01 9.38963234e-01 1.06525505e+00
9.04983342e-01 9.98199761e-01 4.17303950e-01 -3.96182626e-01
-3.08727324e-01 -1.26685393e+00 -1.16897114e-01 4.53896910e-01
-1.54287666e-01 -1.21164560e+00 -4.04758900e-01 -6.53254166e-02] | [9.774740219116211, 8.355619430541992] |
e65d42a4-f1bd-4831-8b51-6624e1c06052 | progressive-temporal-feature-alignment | 2104.03507 | null | https://arxiv.org/abs/2104.03507v1 | https://arxiv.org/pdf/2104.03507v1.pdf | Progressive Temporal Feature Alignment Network for Video Inpainting | Video inpainting aims to fill spatio-temporal "corrupted" regions with plausible content. To achieve this goal, it is necessary to find correspondences from neighbouring frames to faithfully hallucinate the unknown content. Current methods achieve this goal through attention, flow-based warping, or 3D temporal convolution. However, flow-based warping can create artifacts when optical flow is not accurate, while temporal convolution may suffer from spatial misalignment. We propose 'Progressive Temporal Feature Alignment Network', which progressively enriches features extracted from the current frame with the feature warped from neighbouring frames using optical flow. Our approach corrects the spatial misalignment in the temporal feature propagation stage, greatly improving visual quality and temporal consistency of the inpainted videos. Using the proposed architecture, we achieve state-of-the-art performance on the DAVIS and FVI datasets compared to existing deep learning approaches. Code is available at https://github.com/MaureenZOU/TSAM. | ['Yong Jae Lee', 'Ding Liu', 'Linjie Yang', 'Xueyan Zou'] | 2021-04-08 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zou_Progressive_Temporal_Feature_Alignment_Network_for_Video_Inpainting_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zou_Progressive_Temporal_Feature_Alignment_Network_for_Video_Inpainting_CVPR_2021_paper.pdf | cvpr-2021-1 | ['video-inpainting'] | ['computer-vision'] | [ 4.09221239e-02 -3.92683953e-01 -6.01922832e-02 -1.51313573e-01
-5.38708806e-01 -3.60617012e-01 5.13655245e-01 -3.51846516e-01
-1.49474382e-01 9.38436389e-01 6.32397890e-01 1.23386979e-01
5.98848946e-02 -5.80127001e-01 -8.47049892e-01 -4.20564860e-01
-8.98377821e-02 -1.91758528e-01 1.48860380e-01 6.07221760e-02
3.80593240e-01 5.67598701e-01 -1.28808463e+00 4.40593600e-01
7.23923206e-01 1.05113375e+00 2.80577451e-01 6.86721981e-01
-1.49054697e-03 1.19407153e+00 -2.96702504e-01 -1.15117960e-01
4.47987020e-01 -6.00913823e-01 -1.03547263e+00 2.28989154e-01
6.28529608e-01 -8.20262015e-01 -9.80638206e-01 9.85657692e-01
1.27095386e-01 4.23832953e-01 2.03626201e-01 -1.26656127e+00
-9.35732007e-01 4.04568277e-02 -7.03880847e-01 5.36617935e-01
4.63165045e-01 4.67663765e-01 6.85216546e-01 -1.03351080e+00
8.49160373e-01 1.19801295e+00 6.48883760e-01 5.94458938e-01
-1.23015404e+00 -5.64653218e-01 -8.10726285e-02 4.18424159e-01
-1.22834468e+00 -6.75423265e-01 9.01378810e-01 -3.36820006e-01
7.86352038e-01 7.86810145e-02 7.92345762e-01 1.09183300e+00
4.53318268e-01 6.75049484e-01 7.67500401e-01 -1.42946407e-01
-1.38149664e-01 -4.76691455e-01 -5.46482265e-01 6.62382066e-01
-2.23990530e-01 3.82006437e-01 -7.75319457e-01 7.84363002e-02
1.27904141e+00 2.47316837e-01 -6.75378501e-01 -1.99943498e-01
-1.42921042e+00 6.54129803e-01 6.38595760e-01 3.60859245e-01
-5.91516554e-01 4.55297649e-01 2.04772338e-01 3.22586358e-01
6.81913018e-01 4.98310208e-01 -2.69577503e-01 -3.83659214e-01
-1.39411116e+00 3.26355666e-01 2.26655349e-01 7.56448269e-01
7.75639296e-01 3.83199006e-01 -3.21510464e-01 5.49097955e-01
1.99841872e-01 8.72789919e-02 4.57032084e-01 -1.77908385e+00
3.87642354e-01 6.64196685e-02 5.17856419e-01 -1.06787074e+00
-1.75117271e-03 -2.71168590e-01 -8.19527030e-01 4.21056658e-01
4.92946446e-01 -1.28633514e-01 -8.77410054e-01 1.75550544e+00
1.85888931e-01 8.87492120e-01 -1.49529442e-01 1.12358534e+00
5.57253003e-01 8.72854173e-01 -1.06149182e-01 -3.55063945e-01
8.56651545e-01 -1.17576063e+00 -9.88570571e-01 -2.76906520e-01
1.83972389e-01 -1.00790143e+00 7.19649076e-01 9.84908938e-02
-1.43654644e+00 -7.38773465e-01 -8.93210888e-01 -3.69592786e-01
1.71306893e-01 -2.40056217e-01 3.10387641e-01 -7.21079484e-02
-1.17584348e+00 1.07580364e+00 -9.46310997e-01 -1.69278134e-03
6.12102747e-01 9.12605524e-02 -6.32786274e-01 -2.81725317e-01
-1.11481130e+00 8.65951359e-01 1.35252073e-01 1.56496167e-01
-9.92749751e-01 -1.11320829e+00 -9.34503555e-01 -4.93768454e-02
3.00896894e-02 -8.21968615e-01 1.18489289e+00 -1.20486486e+00
-1.31536591e+00 4.40297782e-01 -4.71761733e-01 -5.09475350e-01
8.06312799e-01 -4.59008396e-01 -1.91580668e-01 3.31857622e-01
2.08266288e-01 9.88136828e-01 1.06318736e+00 -1.11516607e+00
-4.41178799e-01 1.23395540e-01 -1.01524383e-01 2.71467924e-01
6.64903373e-02 -4.52179313e-02 -3.92910719e-01 -9.84291911e-01
-9.70947072e-02 -6.59449935e-01 -1.66275710e-01 6.14346445e-01
-2.97426637e-02 2.58169651e-01 1.24365592e+00 -1.11192822e+00
1.06730831e+00 -2.29899096e+00 1.93750024e-01 -3.05956602e-01
3.32660973e-01 5.26314855e-01 -3.18675339e-01 3.92280757e-01
-1.78716078e-01 -1.40364543e-01 -3.59092712e-01 -6.08804584e-01
-5.00461817e-01 2.77847469e-01 -3.82608533e-01 5.94581127e-01
5.65068841e-01 9.88236964e-01 -1.18198991e+00 -3.82438898e-01
5.61690509e-01 9.81465101e-01 -6.83376431e-01 4.44245189e-01
-1.70809925e-01 9.97155130e-01 -2.03773305e-01 4.17676568e-01
7.04230726e-01 -2.08343595e-01 -2.42019102e-01 -3.65120769e-01
-2.57387906e-01 2.45512173e-01 -8.43174875e-01 2.26665974e+00
-3.48121345e-01 1.04685557e+00 8.62015635e-02 -7.73475409e-01
6.50147855e-01 3.78600329e-01 8.66509259e-01 -7.62012005e-01
1.98855564e-01 8.53468254e-02 -2.29889825e-01 -5.84100604e-01
5.55012286e-01 1.97770670e-02 5.52391410e-01 1.58516616e-01
5.34638427e-02 -2.29962990e-02 9.35893953e-02 6.42101169e-02
1.06331038e+00 6.08401477e-01 -1.03468569e-02 9.49474350e-02
3.37429643e-01 -1.91018626e-01 6.76832736e-01 3.04767430e-01
-3.97017628e-01 1.17493796e+00 3.08866024e-01 -7.46591330e-01
-1.37453592e+00 -1.09130788e+00 9.05611217e-02 3.50348979e-01
1.57165259e-01 -2.73269653e-01 -5.32607615e-01 -4.33460891e-01
-1.05747543e-01 5.23375332e-01 -7.37739325e-01 -2.27919251e-01
-7.47389078e-01 -1.62841469e-01 1.62895262e-01 5.43187678e-01
7.73984730e-01 -1.11181569e+00 -7.65319824e-01 5.78393638e-01
-6.80070996e-01 -1.16332090e+00 -8.87387574e-01 -4.48852211e-01
-8.44181120e-01 -8.93550694e-01 -9.82564867e-01 -6.00158215e-01
4.76774395e-01 5.28817892e-01 9.64043975e-01 2.20971435e-01
-3.32708180e-01 -1.62407719e-02 -3.71245086e-01 2.77596354e-01
-3.66151750e-01 -4.71912920e-01 -2.04104960e-01 1.96870849e-01
-1.87495962e-01 -8.59401762e-01 -9.68968391e-01 6.98309094e-02
-1.13549507e+00 3.52716953e-01 1.92632094e-01 9.88757908e-01
3.53970230e-01 -2.02537268e-01 2.70553052e-01 -2.81507611e-01
3.37190002e-01 -4.93248582e-01 -2.90857971e-01 -1.57037869e-01
-1.53762877e-01 4.72588502e-02 6.41706586e-01 -5.18044293e-01
-1.04653811e+00 -7.26904869e-02 -1.24264918e-01 -1.18023360e+00
4.83254641e-02 3.02516639e-01 2.85298973e-01 -1.79117903e-01
5.75804591e-01 3.30264002e-01 3.09911489e-01 -4.03876722e-01
3.32408637e-01 2.28143334e-01 7.84786165e-01 -3.61757815e-01
6.71183228e-01 7.84720421e-01 -5.10568656e-02 -4.21293348e-01
-7.33772814e-01 -2.38190040e-01 -6.25804424e-01 -3.89936388e-01
7.32790828e-01 -8.99460912e-01 -3.23807895e-01 5.21802008e-01
-1.48283398e+00 -4.52478528e-01 -3.80017906e-01 5.82785189e-01
-6.97415292e-01 5.66545486e-01 -8.04340839e-01 -3.38763595e-01
-2.82022029e-01 -1.04554892e+00 8.51005256e-01 2.56595522e-01
-2.96116471e-01 -9.39032197e-01 1.76813215e-01 1.44404098e-01
6.05479002e-01 4.60746676e-01 3.95890564e-01 3.04567724e-01
-7.35916972e-01 8.31243396e-02 -4.50192899e-01 4.01371866e-01
4.04285014e-01 7.76632726e-02 -8.25263321e-01 -4.23880130e-01
3.01272310e-02 -1.07085370e-01 9.02888358e-01 5.48382223e-01
9.98763144e-01 -4.88151073e-01 -4.53029945e-02 9.25961554e-01
1.39597189e+00 2.15980336e-01 9.70309794e-01 2.35604391e-01
7.72301376e-01 4.77208018e-01 5.49691081e-01 5.74355006e-01
1.41847730e-01 6.65178895e-01 5.29911160e-01 -8.95102397e-02
-6.73061013e-01 -3.29431057e-01 3.60250503e-01 4.13526714e-01
-1.42074823e-01 -1.71331882e-01 -5.94205022e-01 8.82581949e-01
-1.96506310e+00 -1.18534100e+00 1.02305584e-01 1.95745492e+00
9.05884087e-01 -1.51381120e-01 -1.84448153e-01 6.05953559e-02
6.08769238e-01 4.55832571e-01 -4.41043854e-01 -1.57409161e-01
8.10524747e-02 2.30868354e-01 1.29820272e-01 8.92306924e-01
-9.92593169e-01 9.22839165e-01 5.47745562e+00 5.11979103e-01
-1.27002454e+00 2.74256289e-01 6.36993051e-01 -3.62879455e-01
-2.44602919e-01 2.67497282e-02 -8.49261507e-02 6.14035368e-01
7.28909612e-01 -8.83119821e-04 6.66260362e-01 2.25598246e-01
6.80677354e-01 -1.06549650e-01 -8.50654542e-01 1.02781880e+00
5.51006803e-03 -1.75898528e+00 -2.47455947e-02 -1.27108067e-01
9.67082381e-01 -5.69205321e-02 1.92534536e-01 -2.46812120e-01
4.81809257e-03 -9.19183910e-01 7.41053045e-01 7.98849642e-01
9.65608120e-01 -7.12092936e-01 3.71517450e-01 2.41078176e-02
-1.08704746e+00 7.26815015e-02 -1.65883452e-01 -1.85843021e-01
3.90232354e-01 5.40842712e-01 -3.78482550e-01 4.06691074e-01
7.96379924e-01 1.14271963e+00 -1.93488166e-01 1.15768933e+00
-8.45954791e-02 2.52570003e-01 -8.86016712e-02 6.77614272e-01
3.63699973e-01 -1.87409505e-01 6.10414326e-01 9.79387879e-01
5.71725368e-01 5.38380966e-02 -1.90206409e-01 1.03754306e+00
-1.23175338e-01 -2.95908064e-01 -6.05718732e-01 9.78405178e-02
3.55978101e-01 9.76329207e-01 -3.46248329e-01 -3.90213877e-01
-6.14106894e-01 1.40571177e+00 3.30920219e-01 5.77180862e-01
-9.37784314e-01 -2.15051517e-01 1.07175815e+00 1.95367411e-01
3.36222798e-01 -4.27980483e-01 -4.63584773e-02 -1.38739657e+00
1.04415305e-01 -5.98994195e-01 2.78732609e-02 -1.14830148e+00
-1.00658631e+00 5.78111649e-01 -2.72756398e-01 -1.50035894e+00
-3.07402104e-01 -1.20517947e-01 -5.92877269e-01 9.75031793e-01
-1.57500613e+00 -1.13669622e+00 -4.94416952e-01 7.17915475e-01
7.74390042e-01 2.13964835e-01 5.12830496e-01 5.17147183e-01
-1.86038330e-01 2.34724462e-01 -9.69355460e-03 1.15888678e-01
8.65866423e-01 -7.45424151e-01 5.86630166e-01 1.21368694e+00
1.13865053e-02 1.73414558e-01 7.66292214e-01 -6.87979817e-01
-1.06939483e+00 -1.44131982e+00 9.93251204e-01 -1.70558125e-01
4.62383866e-01 2.28804648e-01 -1.18388355e+00 7.96654344e-01
5.10392070e-01 6.25132620e-01 3.65063511e-02 -8.47042143e-01
-3.43588352e-01 -5.01460396e-04 -1.20677030e+00 5.91014683e-01
1.05449498e+00 -5.45665920e-01 -3.85684401e-01 2.23474979e-01
7.33732700e-01 -5.44006884e-01 -7.73371935e-01 2.14720964e-01
5.36012232e-01 -1.14220309e+00 9.62942004e-01 -5.09158671e-01
9.54591751e-01 -4.79705930e-01 1.14958398e-01 -1.39058113e+00
-4.62094396e-01 -1.05243671e+00 -3.21050435e-01 9.74047065e-01
-2.12872289e-02 -3.71490628e-01 7.29170918e-01 4.32060987e-01
-2.57885545e-01 -5.79634786e-01 -9.81097281e-01 -5.28292954e-01
-5.09415790e-02 -3.59121412e-01 4.65071648e-01 1.18707645e+00
-2.11226508e-01 -1.73009813e-01 -8.41393888e-01 7.46260583e-02
6.02120459e-01 -1.21476166e-01 4.38265443e-01 -5.89694858e-01
-2.80425578e-01 -2.74703890e-01 -3.40168864e-01 -1.10514677e+00
1.93528324e-01 -5.32907188e-01 -5.21068983e-02 -1.44077730e+00
-6.29068390e-02 1.52822165e-02 -6.80844933e-02 5.12688160e-01
-6.22386262e-02 6.30362511e-01 4.19477403e-01 2.79317051e-01
-2.21084639e-01 6.86798930e-01 1.73601651e+00 6.81263162e-03
-1.81159630e-01 -3.75061333e-01 -3.02674770e-01 4.15406227e-01
8.33573878e-01 -3.17545533e-01 -3.70031297e-01 -8.08633327e-01
-1.07740045e-01 5.40791512e-01 6.41683817e-01 -1.09504688e+00
1.13493219e-01 -3.08001876e-01 7.18105257e-01 -3.65960836e-01
6.38209224e-01 -7.85949230e-01 4.91434485e-01 5.44230044e-01
-2.64040917e-01 3.03315312e-01 2.26180539e-01 4.18166578e-01
-4.70606059e-01 -7.96156097e-03 1.08289957e+00 -1.96359009e-01
-7.49606907e-01 5.20131648e-01 -3.63451540e-01 -5.56562915e-02
8.61558795e-01 -1.56039387e-01 -1.87365711e-01 -6.51070476e-01
-6.41784489e-01 9.30512324e-04 5.87748289e-01 4.67802376e-01
9.96637523e-01 -1.50323558e+00 -7.93551803e-01 4.22865003e-01
-4.03843880e-01 3.68397571e-02 5.34377396e-01 9.89637315e-01
-8.85381341e-01 1.58387557e-01 -6.17719769e-01 -3.93669039e-01
-9.52624500e-01 5.80072224e-01 4.73041266e-01 2.94271931e-02
-7.24559665e-01 8.19583535e-01 -2.66089775e-02 3.07796866e-01
-1.59003735e-02 -2.49235600e-01 6.99098781e-02 -2.39036411e-01
6.73531055e-01 3.35885197e-01 -1.80299848e-01 -7.63752639e-01
-9.72749963e-02 5.66005230e-01 -3.31429206e-02 -2.34867513e-01
1.24578118e+00 -4.39862996e-01 -4.07212321e-03 3.76184806e-02
1.45408916e+00 -1.81568369e-01 -2.06673360e+00 -3.64470184e-01
-4.61074233e-01 -1.02833855e+00 2.72470146e-01 -5.11592329e-01
-1.45945680e+00 9.23271716e-01 4.45915759e-01 -9.68351290e-02
1.29320335e+00 -3.49101007e-01 1.24880469e+00 -3.65147591e-01
4.25987802e-02 -5.60233235e-01 2.87770778e-01 4.22607839e-01
1.11835003e+00 -1.08844209e+00 -8.60311463e-02 -2.86053360e-01
-4.76558864e-01 1.14983654e+00 5.16895235e-01 -4.36451137e-01
5.68219006e-01 1.74294874e-01 -7.04193190e-02 2.10054800e-01
-8.74460101e-01 1.86121278e-02 2.65928060e-01 6.69817984e-01
4.38931733e-01 -3.44136715e-01 -1.96200565e-01 -1.66967154e-01
8.28980356e-02 2.51034677e-01 6.00336790e-01 9.31053221e-01
-1.63542449e-01 -9.75991130e-01 -3.67759466e-01 1.68349385e-01
-5.38992882e-01 -1.36017576e-01 4.13434207e-02 6.21170819e-01
6.26858473e-02 8.27948868e-01 2.93353587e-01 -2.10657597e-01
8.62011462e-02 -1.36644945e-01 6.93470955e-01 -2.48662218e-01
-3.00414652e-01 3.63524407e-01 -3.48592401e-02 -1.10470927e+00
-6.98266506e-01 -6.87469780e-01 -1.11937368e+00 -6.60331666e-01
2.61361599e-01 -2.04273656e-01 1.55197307e-01 8.31927359e-01
5.99664867e-01 5.04532516e-01 6.34478450e-01 -1.24943876e+00
-6.01673275e-02 -8.74873638e-01 -2.50231028e-01 6.80544555e-01
9.38142657e-01 -5.45191824e-01 -3.29367369e-01 4.53290880e-01] | [10.795790672302246, -1.3191653490066528] |
06d7cf2a-1f73-4d4c-aa2e-35c9fd65f63e | assessing-the-performance-of-automated | 2207.01302 | null | https://arxiv.org/abs/2207.01302v1 | https://arxiv.org/pdf/2207.01302v1.pdf | Assessing the Performance of Automated Prediction and Ranking of Patient Age from Chest X-rays Against Clinicians | Understanding the internal physiological changes accompanying the aging process is an important aspect of medical image interpretation, with the expected changes acting as a baseline when reporting abnormal findings. Deep learning has recently been demonstrated to allow the accurate estimation of patient age from chest X-rays, and shows potential as a health indicator and mortality predictor. In this paper we present a novel comparative study of the relative performance of radiologists versus state-of-the-art deep learning models on two tasks: (a) patient age estimation from a single chest X-ray, and (b) ranking of two time-separated images of the same patient by age. We train our models with a heterogeneous database of 1.8M chest X-rays with ground truth patient ages and investigate the limitations on model accuracy imposed by limited training data and image resolution, and demonstrate generalisation performance on public data. To explore the large performance gap between the models and humans on these age-prediction tasks compared with other radiological reporting tasks seen in the literature, we incorporate our age prediction model into a conditional Generative Adversarial Network (cGAN) allowing visualisation of the semantic features identified by the prediction model as significant to age prediction, comparing the identified features with those relied on by clinicians. | ['Giovanni Montana', 'Vicky Goh', 'Charles Hutchinson', 'Ashik Amlani', 'Keerthini Muthuswamy', 'Matthew MacPherson'] | 2022-07-04 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [ 4.44867134e-01 6.41894162e-01 7.30365366e-02 -6.86855614e-01
-1.05366898e+00 -2.18772978e-01 6.57523036e-01 5.86252928e-01
-9.13123548e-01 6.13982260e-01 6.15399003e-01 -4.82382298e-01
-2.72197992e-01 -7.53464282e-01 -6.52127922e-01 -6.45477653e-01
-4.24618274e-01 1.03237355e+00 -2.85101142e-02 3.00873429e-01
-1.18157417e-02 2.85586387e-01 -1.26326823e+00 5.71720421e-01
5.05097926e-01 1.34659910e+00 7.06165880e-02 1.15329337e+00
3.61925155e-01 1.09894741e+00 -6.66434050e-01 -5.63320398e-01
6.84202388e-02 -1.59695581e-01 -1.08529246e+00 -2.48867318e-01
7.27769911e-01 -6.78270340e-01 -6.06661558e-01 4.43200141e-01
1.06002748e+00 -3.48763824e-01 9.33131397e-01 -1.00703764e+00
-6.80430710e-01 4.60928321e-01 -3.12789947e-01 7.15870082e-01
3.52531582e-01 2.35829860e-01 5.43197632e-01 -3.07480007e-01
5.99934757e-01 8.60645235e-01 1.15809298e+00 8.29645574e-01
-8.21466327e-01 -5.36514580e-01 -2.91305125e-01 3.36217314e-01
-8.12925637e-01 -1.44265100e-01 5.46315722e-02 -7.31288970e-01
8.10717404e-01 3.96972418e-01 7.74723947e-01 1.54087424e+00
6.39458835e-01 4.03351694e-01 1.29197133e+00 -1.96161136e-01
1.05097443e-01 -3.43059748e-01 -1.45266518e-01 7.66680360e-01
-1.94471870e-02 3.18508625e-01 -4.01582658e-01 -5.19997776e-01
6.05751753e-01 5.11909463e-02 -2.47330919e-01 -9.95188877e-02
-1.50733888e+00 7.10815072e-01 7.45946467e-01 1.79525584e-01
-6.30642533e-01 4.12796885e-01 5.50592542e-01 2.68098682e-01
6.75205410e-01 6.70324445e-01 -6.23592019e-01 2.12847553e-02
-1.17400742e+00 2.21129343e-01 4.50149477e-01 1.27280161e-01
-2.56381661e-01 -4.12322462e-01 -2.42643967e-01 6.84639275e-01
9.81154293e-02 4.06127363e-01 9.54567373e-01 -8.08082163e-01
2.01966278e-02 3.44879597e-01 -2.40536600e-01 -3.73398632e-01
-8.52802634e-01 -5.36947370e-01 -7.10229099e-01 2.71689177e-01
6.98445618e-01 2.25853398e-02 -1.38409889e+00 1.65427613e+00
1.10490538e-01 -1.62134528e-01 5.57085313e-02 7.17805326e-01
9.87678587e-01 2.50298500e-01 7.36823142e-01 -7.26131946e-02
1.68784463e+00 -4.84043151e-01 -9.50217173e-02 -4.13960278e-01
8.77627134e-01 -4.92095262e-01 6.28659904e-01 3.14622760e-01
-1.35584641e+00 -4.07910049e-01 -8.02509725e-01 -1.90752577e-02
-2.86704838e-01 2.25138396e-01 5.76203883e-01 4.20878530e-01
-1.11393821e+00 9.49262023e-01 -1.15402293e+00 -5.78098714e-01
6.40193284e-01 5.74424088e-01 -4.30574983e-01 -9.53116044e-02
-1.27468729e+00 1.34715378e+00 3.16788077e-01 -2.80011117e-01
-9.18035090e-01 -1.37251794e+00 -6.78235829e-01 -3.19440633e-01
-1.24525808e-01 -1.46564770e+00 1.60419202e+00 -7.83826113e-01
-5.50534368e-01 1.61094904e+00 3.29133809e-01 -9.68805194e-01
1.14224052e+00 -4.07521158e-01 -6.03652477e-01 3.76851588e-01
1.23924069e-01 9.23426926e-01 7.45743930e-01 -8.64919722e-01
-5.04929960e-01 -6.63567483e-01 -1.87464520e-01 1.68253765e-01
-1.92330182e-02 7.45435208e-02 9.34117585e-02 -7.84865499e-01
1.40292257e-01 -9.26005900e-01 -3.83400440e-01 3.70018274e-01
-4.37436439e-03 -2.42305454e-02 3.61888111e-01 -1.38571715e+00
1.11148739e+00 -1.85770202e+00 -4.97095175e-02 1.01532072e-01
7.12415755e-01 -1.29111558e-01 3.75104427e-01 7.95961469e-02
-5.65278709e-01 3.68153453e-02 -3.51893902e-01 -1.59055918e-01
-5.53010702e-01 9.52754766e-02 1.94709256e-01 5.39589524e-01
7.53574073e-02 1.04724622e+00 -9.00805116e-01 -7.08086729e-01
7.12861121e-02 3.71125638e-01 -2.12706074e-01 5.71371913e-01
2.41293028e-01 7.42597520e-01 -1.22600123e-01 5.84750533e-01
2.58840710e-01 -4.08336341e-01 -5.42541742e-02 -4.70575035e-01
3.91603172e-01 1.52423576e-01 -3.55074912e-01 1.70052409e+00
-6.35062814e-01 2.04056695e-01 -2.48737097e-01 -6.21119976e-01
5.90823412e-01 3.76189977e-01 7.27030635e-01 -6.50711417e-01
1.12278052e-01 2.85013288e-01 4.20208305e-01 -5.13848305e-01
-1.70618817e-02 -5.68801165e-01 -2.18569994e-01 6.67878330e-01
1.16279863e-01 -1.82990044e-01 -2.27874532e-01 3.34527314e-01
1.53383434e+00 -2.09208608e-01 4.23171252e-01 -1.17559254e-01
4.35626954e-01 -2.56186724e-01 8.08568373e-02 9.31645572e-01
-3.98981243e-01 1.02816677e+00 4.08357024e-01 -8.39256823e-01
-1.41836500e+00 -1.46013844e+00 -3.94665837e-01 8.71366799e-01
-4.86938596e-01 -9.39602554e-02 -5.88764369e-01 -9.73985076e-01
7.40565360e-02 6.41015410e-01 -1.47056019e+00 -5.22465646e-01
-6.20006025e-01 -1.09666526e+00 6.94783390e-01 1.15791607e+00
1.27167761e-01 -1.05520976e+00 -1.21952748e+00 9.54311788e-02
-6.21370822e-02 -8.15178871e-01 -1.78939104e-01 2.60986686e-01
-9.84508634e-01 -1.22022796e+00 -1.23700547e+00 -4.72363353e-01
6.18436635e-01 -8.30699027e-01 1.72001410e+00 2.72738338e-01
-8.35876703e-01 9.39606547e-01 -1.26349658e-01 -7.60570943e-01
-1.04001081e+00 2.44600967e-01 9.35150217e-03 -5.89568794e-01
1.22059785e-01 -4.58706081e-01 -1.40495193e+00 6.22105002e-02
-7.11257696e-01 1.35969579e-01 1.00689280e+00 9.30654645e-01
4.28705841e-01 -5.45825303e-01 5.84902406e-01 -1.07798195e+00
3.83819371e-01 -7.06411302e-01 2.38961756e-01 1.64842397e-01
-9.03053999e-01 9.87717360e-02 1.14398509e-01 -2.44499803e-01
-9.30333555e-01 -7.58512765e-02 -3.36225122e-01 -5.87556586e-02
-2.08057895e-01 2.28788093e-01 6.54314220e-01 1.34152010e-01
8.84048820e-01 2.69216690e-02 3.47381532e-01 -3.14281732e-01
5.97966574e-02 4.69456464e-01 1.07133615e+00 -3.16256881e-01
4.71676767e-01 4.63780195e-01 2.85515010e-01 -2.37903655e-01
-1.21035421e+00 -3.29059690e-01 -7.68170118e-01 -2.96497107e-01
1.22198069e+00 -9.08684731e-01 -2.20995113e-01 5.39317548e-01
-7.65350997e-01 -2.14438528e-01 -5.24177253e-01 5.72720170e-01
-7.21801758e-01 3.39089453e-01 -8.24609995e-01 -3.35631698e-01
-9.53947246e-01 -1.03158820e+00 1.28171110e+00 -1.73787460e-01
-7.57376552e-01 -1.13469601e+00 1.93535686e-01 3.71327609e-01
5.01923561e-01 6.81256950e-01 1.34996295e+00 -8.60185444e-01
1.73160762e-01 -4.41024154e-01 -2.68770158e-01 3.48659515e-01
-8.32670778e-02 -3.59491408e-01 -1.03996694e+00 -2.35825241e-01
-2.40714457e-02 -4.24600244e-01 9.94512916e-01 7.10235417e-01
1.62580931e+00 9.89563242e-02 -3.73399407e-01 6.38612509e-01
1.16218579e+00 -3.89626697e-02 7.44169772e-01 5.51294327e-01
5.53911686e-01 3.33280742e-01 5.07166922e-01 3.65085900e-01
2.78909236e-01 2.41317749e-02 7.34639108e-01 -4.23560053e-01
-1.38489455e-01 9.49426070e-02 -2.84937054e-01 5.28389096e-01
-3.22784632e-01 1.11421160e-01 -1.35392010e+00 5.01847386e-01
-1.28957129e+00 -5.09386122e-01 -6.59526214e-02 2.12024999e+00
7.03658879e-01 3.39793801e-01 1.90895170e-01 3.03059872e-02
6.34786069e-01 2.27120265e-01 -5.13136804e-01 -3.87822688e-01
9.17179734e-02 5.25712669e-01 8.75219524e-01 -1.43646017e-01
-1.22478426e+00 -1.77756920e-01 7.59168196e+00 8.28870833e-02
-9.85607088e-01 3.78854901e-01 1.17139113e+00 -3.70871007e-01
1.00886576e-01 -4.88145292e-01 -1.95271596e-02 4.20719236e-01
1.49356616e+00 -9.89713892e-02 -1.79472953e-01 7.91663110e-01
-1.95992962e-01 -2.49569878e-01 -1.63506114e+00 8.51415694e-01
2.10533515e-01 -1.04645860e+00 -4.88150343e-02 -8.01381934e-03
4.03981447e-01 1.44325316e-01 3.81400764e-01 1.13138258e-01
1.19554922e-01 -1.34850049e+00 6.43818140e-01 9.00272131e-01
1.25904918e+00 -3.99445385e-01 1.15114439e+00 -2.72756517e-02
-5.83033800e-01 -2.46906832e-01 7.23713264e-02 2.56595433e-01
2.97174528e-02 3.05989325e-01 -1.23913550e+00 4.21227276e-01
1.02098489e+00 4.84094232e-01 -1.14315128e+00 9.79242086e-01
1.06937036e-01 6.13426387e-01 -1.07664637e-01 6.59209371e-01
1.40991658e-01 5.04377306e-01 1.15350455e-01 1.08023953e+00
5.43286860e-01 8.54583979e-02 -2.98863858e-01 3.12366754e-01
-1.30656451e-01 -4.57883766e-03 -2.90507436e-01 1.86584309e-01
4.33463454e-02 1.05685449e+00 -7.93656766e-01 -5.71229279e-01
-3.73075306e-01 9.19430971e-01 1.30490601e-01 -3.49549145e-01
-8.81254256e-01 3.44886035e-01 2.41537437e-01 5.31589389e-01
-5.69997951e-02 2.20192388e-01 -3.49159300e-01 -5.58137953e-01
-2.89369047e-01 -7.13333845e-01 9.18848693e-01 -1.20251477e+00
-1.56896019e+00 5.13661087e-01 -7.29668960e-02 -9.53020096e-01
-4.12140608e-01 -7.00307429e-01 -6.47865951e-01 9.77554202e-01
-9.74476635e-01 -1.21323967e+00 -6.29373729e-01 1.35278612e-01
3.28939795e-01 -5.24896123e-02 1.05082297e+00 1.94601148e-01
2.12398153e-02 5.99595666e-01 -8.35755989e-02 2.80221310e-02
1.08711600e+00 -1.88463235e+00 5.68650663e-01 3.87629420e-01
-4.08544004e-01 2.38810569e-01 7.15450108e-01 -7.50695646e-01
-5.98065138e-01 -1.11915588e+00 5.92554867e-01 -1.20721781e+00
5.35996318e-01 1.51373193e-01 -1.01387393e+00 7.06840813e-01
-5.62927984e-02 3.32279503e-02 8.99731040e-01 -1.30344883e-01
-1.83799326e-01 2.42137566e-01 -1.23083937e+00 9.45483670e-02
9.50937569e-01 -4.40928966e-01 -8.69304180e-01 4.00224894e-01
5.87485313e-01 -6.95809662e-01 -1.61065197e+00 7.84032643e-01
8.70728850e-01 -8.85391355e-01 1.27550960e+00 -8.03609073e-01
9.26422000e-01 3.35319698e-01 2.92124063e-01 -1.31848419e+00
-2.82331437e-01 2.98848636e-02 -2.18571141e-01 6.02020264e-01
4.32668775e-01 -3.37027967e-01 9.24742162e-01 6.60610676e-01
-1.34147003e-01 -1.19597411e+00 -8.69567871e-01 -3.14651042e-01
4.45715189e-01 -4.17998016e-01 4.46513653e-01 6.96841598e-01
-6.77114725e-01 -1.05728015e-01 -5.89089654e-02 1.23869531e-01
4.28594708e-01 -5.39008260e-01 2.45551229e-01 -1.43005764e+00
-5.69207132e-01 -2.85289854e-01 -7.86017954e-01 4.53814156e-02
-2.25520417e-01 -7.89132833e-01 -2.39658579e-01 -1.81007695e+00
3.68691355e-01 -3.08210850e-01 -5.90088904e-01 1.82358906e-01
-3.49851489e-01 4.58768785e-01 -1.38957158e-01 3.82799417e-01
-3.29034895e-01 -5.10267690e-02 1.11417544e+00 -1.11621708e-01
3.81675750e-01 1.99972659e-01 -5.59541225e-01 8.11899304e-01
6.36416852e-01 -6.68007255e-01 -1.72280252e-01 -2.50310659e-01
4.08739895e-01 1.29625455e-01 5.60850561e-01 -1.35112250e+00
-1.85019389e-01 4.23935950e-01 1.10974896e+00 -5.29117405e-01
1.58233613e-01 -6.03393018e-01 1.56449899e-01 1.09389043e+00
-5.88000655e-01 5.20620048e-01 2.47657243e-02 6.10078752e-01
1.27099454e-01 -2.45316491e-01 8.77603292e-01 -5.98989010e-01
-6.75642133e-01 3.91705841e-01 -2.67840266e-01 2.12067038e-01
9.94748116e-01 -1.97983161e-01 -2.73392767e-01 -3.84566069e-01
-1.36029792e+00 4.70514931e-02 4.14717823e-01 3.07279944e-01
3.57604176e-01 -9.71741080e-01 -1.01269674e+00 -2.64012009e-01
4.01388466e-01 3.86642069e-02 4.86949831e-01 1.07780492e+00
-8.63231003e-01 1.57914177e-01 -3.33516181e-01 -7.97760367e-01
-1.43350697e+00 6.90985560e-01 6.08463764e-01 -7.52284706e-01
-6.71968281e-01 1.00434339e+00 4.24596578e-01 -1.94450885e-01
3.53126377e-02 -3.10967475e-01 1.43348183e-02 4.18415144e-02
4.45218503e-01 2.64007986e-01 4.50933307e-01 -4.75011528e-01
-4.46656525e-01 2.95908064e-01 -4.87684727e-01 1.22530684e-01
1.38896835e+00 1.86289120e-02 1.25433981e-01 5.42290866e-01
1.06574845e+00 -5.87410390e-01 -1.02532852e+00 -1.44501105e-01
-1.12155572e-01 -2.89208554e-02 3.45274284e-02 -1.31031299e+00
-9.37358022e-01 8.09110701e-01 1.33483183e+00 1.54710291e-02
1.15310621e+00 6.15849614e-01 8.52083027e-01 -4.78622317e-02
-9.37938020e-02 -1.02476203e+00 2.87659347e-01 -1.14280768e-01
8.74086380e-01 -1.32759130e+00 4.22709018e-01 -1.22597933e-01
-7.20821798e-01 1.04700756e+00 6.35293365e-01 -4.65038717e-02
3.02839696e-01 3.69633324e-02 2.78452188e-01 -3.54793966e-01
-6.71926856e-01 1.50801212e-01 4.24772441e-01 7.24228561e-01
7.58540690e-01 1.35311931e-01 -3.47420692e-01 4.71380770e-01
-4.86371368e-01 -6.42720610e-02 4.04675782e-01 9.81404781e-01
-1.11227073e-01 -8.55580509e-01 -3.61550301e-01 1.27022266e+00
-1.06847417e+00 -1.77402616e-01 -1.62726834e-01 8.42369378e-01
3.44853014e-01 7.17189386e-02 2.63320476e-01 -8.64733160e-02
2.77228564e-01 4.88735765e-01 6.26540542e-01 -5.81130266e-01
-8.29388142e-01 -6.37713075e-01 2.78074294e-01 -4.30805743e-01
-3.56031030e-01 -8.55150282e-01 -9.25918519e-01 2.44292721e-01
8.57196301e-02 -2.56707460e-01 7.29705393e-01 7.93537915e-01
-4.07880591e-03 1.11202204e+00 1.89079151e-01 -5.52506864e-01
-7.10433125e-01 -1.09870565e+00 -3.83012295e-01 8.22270572e-01
2.00259686e-01 -5.42823970e-01 -4.32357848e-01 2.45609075e-01] | [15.20938777923584, -1.961327314376831] |
edf82d86-b320-41c7-b0cc-504aeaefcfad | machine-learned-solutions-for-three-stages-of | null | null | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3168309/ | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3168309/pdf/amiajnl-2011-000150.pdf | Machine-learned solutions for three stages of clinical information extraction: the state of the art at i2b2 2010 | Objective: As clinical text mining continues to mature, its
potential as an enabling technology for innovations in
patient care and clinical research is becoming a reality. A
critical part of that process is rigid benchmark testing of
natural language processing methods on realistic clinical
narrative. In this paper, the authors describe the design
and performance of three state-of-the-art text-mining
applications from the National Research Council of
Canada on evaluations within the 2010 i2b2 challenge.
Design: The three systems perform three key steps in
clinical information extraction: (1) extraction of medical
problems, tests, and treatments, from discharge
summaries and progress notes; (2) classification of
assertions made on the medical problems; (3)
classification of relations between medical concepts.
Machine learning systems performed these tasks using
large-dimensional bags of features, as derived from both
the text itself and from external sources: UMLS, cTAKES,
and Medline.
Measurements: Performance was measured per
subtask, using micro-averaged F-scores, as calculated by
comparing system annotations with ground-truth
annotations on a test set.
Results: The systems ranked high among all submitted
systems in the competition, with the following F-scores:
concept extraction 0.8523 (ranked first); assertion
detection 0.9362 (ranked first); relationship detection
0.7313 (ranked second).
Conclusion: For all tasks, we found that the introduction
of a wide range of features was crucial to success.
Importantly, our choice of machine learning algorithms
allowed us to be versatile in our feature design, and to
introduce a large number of features without overfitting
and without encountering computing-resource
bottlenecks. | ['Svetlana Kiritchenko', 'Xiaodan Zhu', 'Joel Martin', 'Berry de Bruijn', 'Colin Cherry'] | 2011-05-12 | null | null | null | jamia-2011-5 | ['clinical-concept-extraction'] | ['medical'] | [ 4.40100074e-01 2.95129776e-01 -4.96312886e-01 -2.77426004e-01
-1.30414057e+00 -5.17028332e-01 5.89292049e-01 1.14803338e+00
-6.51986361e-01 9.51255679e-01 4.95683879e-01 -6.82765067e-01
-5.23272157e-01 -4.10873055e-01 -1.82843938e-01 -3.32648724e-01
-3.87431145e-01 6.06274724e-01 -8.58802125e-02 8.61473903e-02
3.31445813e-01 3.15837175e-01 -1.09392059e+00 6.56099081e-01
6.39177799e-01 1.04914236e+00 -2.37030402e-01 7.38509238e-01
8.03547800e-02 1.12972689e+00 -5.43461919e-01 -4.50313747e-01
-5.45504130e-02 -3.30588847e-01 -9.78985131e-01 -8.40747878e-02
-2.28362411e-01 -2.88111959e-02 1.56812608e-01 6.39952600e-01
6.19606555e-01 -3.74991655e-01 6.73023939e-01 -8.11617732e-01
-1.45895183e-01 6.30376518e-01 -2.54434288e-01 4.21125948e-01
6.87793672e-01 8.94169733e-02 1.16475224e+00 -8.44709456e-01
9.21350896e-01 6.16849303e-01 8.35751474e-01 4.29476529e-01
-1.23385477e+00 -6.30192518e-01 -2.36716598e-01 -7.07159042e-02
-1.30730700e+00 -6.23317599e-01 -1.27865866e-01 -8.90413761e-01
1.39858103e+00 4.61124271e-01 7.14037359e-01 6.76389933e-01
7.18075097e-01 4.71367627e-01 9.09298718e-01 -5.44989228e-01
3.55875760e-01 4.24329728e-01 2.75514990e-01 7.85838485e-01
5.09899020e-01 -1.87540427e-01 -4.09190267e-01 -8.58877718e-01
1.91626638e-01 -1.13277793e-01 -3.86938751e-01 2.18153253e-01
-1.23316598e+00 7.51280844e-01 2.41781194e-02 2.84230202e-01
-4.97251809e-01 -3.88821900e-01 7.02350318e-01 1.48944929e-01
4.12987322e-01 8.59294593e-01 -9.00534391e-01 -2.42550269e-01
-1.02585590e+00 2.37212151e-01 9.69837427e-01 8.15071583e-01
-1.18394040e-01 -6.15942359e-01 -1.71537608e-01 7.59075642e-01
3.00626814e-01 -5.34255244e-02 8.29332829e-01 -4.68925148e-01
5.58819711e-01 8.11219394e-01 -3.97879481e-02 -7.67791629e-01
-8.97718132e-01 -3.80420834e-01 -6.68268502e-01 -4.03399974e-01
2.70631433e-01 -4.11515564e-01 -9.38970387e-01 1.38366759e+00
1.60829470e-01 -3.80051762e-01 3.55190933e-01 4.72674042e-01
1.07419431e+00 4.63018827e-02 3.88566852e-01 -4.69198942e-01
1.77334058e+00 -3.66494149e-01 -8.46015632e-01 -6.02792390e-02
1.29876125e+00 -9.45983887e-01 5.46879947e-01 5.27392685e-01
-1.21258414e+00 6.35363758e-02 -9.98156607e-01 9.58756357e-02
-3.93889368e-01 2.90131457e-02 6.23479664e-01 5.46249270e-01
-7.73214698e-01 5.17426312e-01 -8.26993644e-01 -4.66186434e-01
7.46637881e-01 4.69114661e-01 -7.28510737e-01 -1.17045112e-01
-1.17599618e+00 1.18605006e+00 3.29092503e-01 -4.81018610e-02
-4.14997786e-01 -1.01297438e+00 -7.83814013e-01 -1.48058543e-02
4.11029875e-01 -8.36734891e-01 1.29874194e+00 -4.38544661e-01
-6.98095620e-01 1.08406365e+00 -7.47307688e-02 -4.89108205e-01
3.61021698e-01 -2.83020914e-01 -6.08304858e-01 1.15466990e-01
3.76834869e-01 2.38747150e-01 4.53490578e-02 -6.10466897e-01
-8.66361499e-01 -4.65546757e-01 -4.37092513e-01 1.00222446e-01
-2.63772666e-01 2.87352234e-01 -3.21006954e-01 -4.73844916e-01
-5.17618917e-02 -6.95515037e-01 -5.78426063e-01 -3.02287817e-01
-5.17077208e-01 -2.67148554e-01 1.91954583e-01 -7.38366425e-01
1.54402661e+00 -1.98606741e+00 -2.35993266e-01 3.62467796e-01
6.25701189e-01 1.52113587e-01 4.83906955e-01 5.69800138e-01
-3.49057496e-01 5.18547952e-01 -1.77952237e-02 -1.50789870e-02
-3.86201650e-01 -7.23955184e-02 1.36590526e-01 4.14556652e-01
4.60369021e-01 9.29914355e-01 -8.77057195e-01 -8.33489478e-01
-8.67015198e-02 1.79183125e-01 -6.04555666e-01 -9.67645422e-02
1.01883672e-01 1.36809081e-01 -5.17780364e-01 6.44234419e-01
-1.94303602e-01 -5.04911900e-01 4.05039310e-01 -9.85602736e-02
-3.31085548e-02 7.46656656e-01 -9.18439090e-01 1.36052561e+00
-2.65638307e-02 2.10366681e-01 -2.91830916e-02 -7.67859399e-01
6.85213089e-01 7.48099267e-01 1.06731391e+00 -4.75958407e-01
1.82131484e-01 2.86302090e-01 2.29956910e-01 -7.60578632e-01
6.97664842e-02 -4.06213880e-01 -1.81972176e-01 3.01771224e-01
-1.66663229e-02 -7.20983893e-02 2.18376845e-01 4.18154925e-01
1.63229191e+00 -4.52584594e-01 1.00363338e+00 -3.67886364e-01
3.79532635e-01 4.19216096e-01 7.46814430e-01 5.48126817e-01
-2.77933348e-02 3.85859877e-01 7.59443104e-01 -4.83743519e-01
-6.66332901e-01 -6.67191744e-01 -6.38118029e-01 4.93159205e-01
-7.37655222e-01 -9.56879377e-01 -3.39026362e-01 -7.14611351e-01
6.72600642e-02 4.86299992e-01 -8.31786096e-01 -1.10307485e-01
-2.25476384e-01 -9.73433554e-01 7.25011528e-01 5.18279850e-01
-1.94922253e-01 -8.32169473e-01 -9.33218360e-01 4.99362797e-01
-8.70676897e-03 -1.23662949e+00 -2.93902338e-01 4.55522418e-01
-8.69446099e-01 -1.50446248e+00 -4.44963247e-01 -5.89083016e-01
4.61460769e-01 -4.11431283e-01 1.14771056e+00 1.16118558e-01
-8.77045810e-01 2.74285138e-01 -2.75945723e-01 -8.67600977e-01
-3.90279472e-01 -6.29332736e-02 -1.06898159e-01 -5.69148421e-01
6.97087049e-01 4.38594483e-02 -5.15607834e-01 7.27368817e-02
-9.53131080e-01 -1.23442365e-02 7.74685800e-01 1.03123176e+00
5.54509521e-01 1.75137874e-02 7.07854927e-01 -1.25116169e+00
8.46500456e-01 -5.91190279e-01 -2.09882736e-01 1.00560509e-01
-1.05153453e+00 -1.73584461e-01 3.55040073e-01 3.41336280e-02
-5.31922698e-01 4.66171466e-02 -1.34036899e-01 2.99860656e-01
-1.09912083e-02 1.12272739e+00 1.97428301e-01 5.37236214e-01
8.53074849e-01 -2.23188534e-01 1.88942403e-01 -2.19416931e-01
5.29892147e-02 8.57847631e-01 1.61500305e-01 -3.13765019e-01
4.46584672e-02 1.40412867e-01 8.52960572e-02 -7.00434148e-01
-9.23446774e-01 -8.41967821e-01 -3.89186591e-01 2.19519794e-01
8.38951230e-01 -9.42255318e-01 -6.58969522e-01 -1.89099446e-01
-7.95872986e-01 -7.83376843e-02 -3.74045640e-01 7.81922698e-01
-2.25442171e-01 1.00882268e-02 -5.96253574e-01 -5.24816751e-01
-6.41524971e-01 -1.22415638e+00 9.18875635e-01 -9.30114388e-02
-9.24770832e-01 -9.53763366e-01 1.84299648e-01 4.23550248e-01
3.25698331e-02 4.79624659e-01 1.26335788e+00 -1.22798181e+00
1.06829025e-01 -6.25691950e-01 -1.99411690e-01 5.56691997e-02
4.92706746e-01 7.49692470e-02 -7.54945576e-01 -1.39223278e-01
-5.49549516e-03 -3.37599993e-01 6.36597753e-01 4.59787220e-01
1.07910573e+00 -3.70140314e-01 -7.24212646e-01 2.24698260e-01
1.23172343e+00 3.84557962e-01 3.70457649e-01 2.45802134e-01
2.72377580e-01 6.48986101e-01 5.31548917e-01 4.12678182e-01
3.37055415e-01 5.14349222e-01 -1.30237475e-01 -1.40011817e-01
2.93002516e-01 1.22788213e-01 -2.61732880e-02 4.98108655e-01
-5.34531176e-02 1.30459711e-01 -1.55950618e+00 6.85312450e-01
-1.78142691e+00 -4.96424347e-01 -1.59690976e-01 2.13867450e+00
1.16787648e+00 6.84245527e-01 5.38568422e-02 3.10874164e-01
1.48089707e-01 -6.27607167e-01 -2.62739748e-01 -3.48939687e-01
2.19979212e-01 5.03757298e-01 4.75642920e-01 1.95136935e-01
-9.31474090e-01 4.19359773e-01 6.26303673e+00 3.56951594e-01
-7.94585824e-01 5.27629405e-02 7.61545599e-01 -3.76342595e-01
5.33173904e-02 -3.19676280e-01 -9.26127613e-01 2.67206788e-01
1.32554054e+00 -3.61156464e-01 -3.76578808e-01 5.24380028e-01
3.39697868e-01 -2.93417364e-01 -1.38090491e+00 6.54966831e-01
-1.48362024e-02 -1.72661471e+00 -2.57001668e-01 2.36601636e-01
5.35170853e-01 1.22481957e-01 -2.57937789e-01 1.46518737e-01
3.89449656e-01 -1.31320727e+00 3.21598053e-01 4.78557736e-01
9.78507280e-01 -3.67988437e-01 1.20791161e+00 1.10148743e-01
-7.36880660e-01 -1.15878828e-01 1.80986822e-01 1.31217703e-01
5.59251383e-02 9.27512288e-01 -1.46524107e+00 7.96490788e-01
7.30332494e-01 6.82403147e-01 -4.00564164e-01 9.77744162e-01
1.25694290e-01 7.60399401e-01 -1.19094931e-01 -1.26535147e-01
1.47861451e-01 5.07457137e-01 2.35115901e-01 1.66127789e+00
-7.58859888e-02 5.36590934e-01 3.11685264e-01 3.19999516e-01
-8.55887830e-02 5.08738220e-01 -5.99288821e-01 -3.59990448e-01
3.13126504e-01 1.37744820e+00 -8.08530152e-01 -4.38439667e-01
-4.21232015e-01 9.36198607e-02 -6.96709678e-02 -7.76686370e-02
-5.06678939e-01 -3.63960505e-01 3.75881523e-01 3.56096029e-01
-3.41276936e-02 2.68002689e-01 -6.77592516e-01 -8.33944499e-01
-9.19887647e-02 -1.14851439e+00 9.39380765e-01 -3.65299284e-01
-1.23905599e+00 7.43467093e-01 -1.74164101e-01 -1.01095951e+00
-3.41781586e-01 -6.35124207e-01 -3.08082043e-03 9.60087597e-01
-1.08853042e+00 -6.48846447e-01 -4.82444465e-02 3.25783521e-01
2.64068961e-01 -2.69501060e-01 1.29424059e+00 4.34136391e-01
-5.09163916e-01 5.90101600e-01 -1.46378696e-01 3.43205601e-01
9.59331810e-01 -1.16457224e+00 1.52936563e-01 1.19842641e-01
-1.15777671e-01 9.73664641e-01 3.74258697e-01 -8.57779741e-01
-1.14530635e+00 -9.62892950e-01 1.23383319e+00 -6.54803932e-01
8.58031690e-01 -1.16452299e-01 -7.81506717e-01 5.81026673e-01
-2.22719908e-01 -2.02278167e-01 1.51949334e+00 3.57711971e-01
-1.18084893e-01 1.56047240e-01 -1.14395881e+00 3.10554504e-01
5.42701066e-01 -3.65001023e-01 -6.22619033e-01 5.81414342e-01
4.44766223e-01 -4.35870707e-01 -1.37690532e+00 5.97331524e-01
7.44550228e-01 -3.04158986e-01 7.68184185e-01 -1.28105962e+00
7.91804314e-01 -4.51404862e-02 -5.52903488e-02 -8.86086404e-01
-2.57099241e-01 -6.57013297e-01 4.88832816e-02 8.58591497e-01
1.15609515e+00 -5.46451509e-01 6.27788067e-01 8.65428865e-01
-1.40373111e-01 -1.42202115e+00 -6.96038783e-01 -1.06985942e-01
-4.38004732e-02 -4.27553654e-01 1.01769149e-01 1.08028388e+00
7.18921185e-01 6.19301140e-01 3.26172799e-01 -5.63085116e-02
2.00829789e-01 -1.93675190e-01 3.74856830e-01 -1.39356339e+00
-3.36780757e-01 -5.18270910e-01 -4.99460846e-01 -8.38327557e-02
-3.77315462e-01 -9.66788590e-01 -6.45816326e-02 -1.88502157e+00
4.37312126e-01 -5.82470715e-01 -5.07708311e-01 8.13990414e-01
-3.90369922e-01 -2.09429432e-02 -2.87082165e-01 3.05363506e-01
-2.66801357e-01 -2.03954726e-01 7.62465119e-01 3.06513254e-02
-2.77270526e-01 3.30397859e-02 -1.14540207e+00 7.86610723e-01
6.38818502e-01 -7.13968754e-01 -2.99924105e-01 -4.97260280e-02
3.23690265e-01 2.49439716e-01 1.04638502e-01 -6.45346463e-01
3.04941446e-01 -3.42903212e-02 5.24931014e-01 -4.78096277e-01
5.48372604e-02 -5.23876369e-01 7.60353431e-02 7.16295719e-01
-5.65213501e-01 2.41330117e-01 5.36661804e-01 4.04657662e-01
-1.04832269e-01 -1.66569829e-01 5.64613044e-01 -1.75471365e-01
-2.15532348e-01 -2.37254545e-01 -4.22726542e-01 2.50601172e-01
1.08492625e+00 -1.02169231e-01 -3.31492364e-01 -1.24550417e-01
-9.29765165e-01 2.77821809e-01 -5.52424416e-03 2.09745333e-01
4.31991309e-01 -7.23379314e-01 -9.52723324e-01 5.54402173e-03
2.60851026e-01 8.65584910e-02 -8.61767158e-02 1.24339318e+00
-3.95183891e-01 8.12228382e-01 1.37577847e-01 -5.88414669e-01
-1.56468141e+00 3.69302720e-01 3.16490307e-02 -8.45572531e-01
-6.22906029e-01 7.66818225e-01 -9.47402045e-02 -1.28956228e-01
2.79481232e-01 -4.76356417e-01 -2.66581893e-01 3.18999171e-01
8.04532230e-01 3.32830139e-02 7.19176710e-01 -1.36826485e-01
-7.82461166e-01 1.01125509e-01 -6.01452291e-01 -1.49183825e-01
1.56578445e+00 4.40990210e-01 -9.70535651e-02 3.42256606e-01
9.07979190e-01 -8.87820348e-02 -2.21768856e-01 -1.12073705e-01
4.89238143e-01 8.74785781e-02 1.37900472e-01 -1.27127957e+00
-5.64426124e-01 3.78150344e-01 2.31628135e-01 3.32563698e-01
9.66638386e-01 2.26918571e-02 3.09545964e-01 2.95529991e-01
5.75907603e-02 -9.49684024e-01 -1.66321144e-01 2.48562038e-01
6.60363376e-01 -1.06267834e+00 5.81182003e-01 -5.28312504e-01
-8.30311894e-01 1.01533461e+00 1.65046006e-01 4.42198843e-01
8.81887794e-01 5.23673654e-01 -1.20682986e-02 -7.02515781e-01
-1.17044604e+00 8.54087695e-02 4.41917241e-01 1.07549414e-01
9.48981047e-01 1.20487474e-01 -5.86679935e-01 8.81483316e-01
-1.40363902e-01 4.27256256e-01 4.83210742e-01 1.24023402e+00
-1.65271670e-01 -1.07697225e+00 -1.38748094e-01 1.17801440e+00
-1.02253890e+00 -4.21176821e-01 -4.98322785e-01 9.35751379e-01
5.10916077e-02 9.63978767e-01 -3.39216769e-01 -2.67094940e-01
5.58748424e-01 2.33772397e-01 2.15851009e-01 -1.10504627e+00
-9.50283945e-01 1.64027318e-01 6.06252074e-01 -5.62516153e-01
-3.70971620e-01 -9.26722705e-01 -1.32309628e+00 1.45069927e-01
-6.32321894e-01 5.29348731e-01 7.49362648e-01 9.79761899e-01
5.97806990e-01 8.86865795e-01 9.14945006e-02 9.14099440e-02
-5.35668671e-01 -9.92212951e-01 -3.05024654e-01 2.72308111e-01
1.72461614e-01 -4.55455542e-01 -2.69379895e-02 2.95646071e-01] | [8.45052433013916, 8.598287582397461] |
255605b1-1e6d-4255-bd22-8a2abbd3202f | mmm-multi-stage-multi-task-learning-for-multi | 1910.00458 | null | https://arxiv.org/abs/1910.00458v2 | https://arxiv.org/pdf/1910.00458v2.pdf | MMM: Multi-stage Multi-task Learning for Multi-choice Reading Comprehension | Machine Reading Comprehension (MRC) for question answering (QA), which aims to answer a question given the relevant context passages, is an important way to test the ability of intelligence systems to understand human language. Multiple-Choice QA (MCQA) is one of the most difficult tasks in MRC because it often requires more advanced reading comprehension skills such as logical reasoning, summarization, and arithmetic operations, compared to the extractive counterpart where answers are usually spans of text within given passages. Moreover, most existing MCQA datasets are small in size, making the learning task even harder. We introduce MMM, a Multi-stage Multi-task learning framework for Multi-choice reading comprehension. Our method involves two sequential stages: coarse-tuning stage using out-of-domain datasets and multi-task learning stage using a larger in-domain dataset to help model generalize better with limited data. Furthermore, we propose a novel multi-step attention network (MAN) as the top-level classifier for this task. We demonstrate MMM significantly advances the state-of-the-art on four representative MCQA datasets. | ['Dilek Hakkani-Tur', 'Jiun-Yu Kao', 'Shuyang Gao', 'Tagyoung Chung', 'Di Jin'] | 2019-10-01 | null | null | null | null | ['multiple-choice-qa'] | ['natural-language-processing'] | [ 5.47683835e-01 5.98083287e-02 3.62390429e-01 -5.36243796e-01
-1.35038340e+00 -5.98724365e-01 3.85496944e-01 5.87408364e-01
-5.53879201e-01 6.76018059e-01 4.41024452e-01 -7.98199415e-01
-1.13841124e-01 -9.75344896e-01 -8.31239283e-01 -9.88869071e-02
3.88726711e-01 6.73646927e-01 5.09590685e-01 -6.67551100e-01
4.91563052e-01 -2.02490017e-01 -1.20831132e+00 8.03393841e-01
1.63037133e+00 1.18480134e+00 5.18997908e-01 9.10769999e-01
-3.93738717e-01 1.29377508e+00 -8.26045275e-01 -5.90257168e-01
-2.66317159e-01 -7.10524321e-01 -1.72324884e+00 -4.43055212e-01
6.89242661e-01 -3.72535735e-01 -1.08622257e-02 8.27722669e-01
4.23613369e-01 5.12162268e-01 7.15689182e-01 -7.96132743e-01
-1.14652443e+00 5.49411118e-01 -9.13973376e-02 5.81014037e-01
9.06545579e-01 2.94834793e-01 1.24086869e+00 -7.09526956e-01
-1.00932103e-02 1.52587867e+00 2.90529490e-01 6.08571291e-01
-7.96516180e-01 -3.38636458e-01 1.68179497e-01 9.04496968e-01
-6.20031655e-01 -2.20671326e-01 5.87417543e-01 -1.80526357e-02
1.14606583e+00 1.64599255e-01 7.30324462e-02 1.00014782e+00
4.52156514e-01 1.08091855e+00 1.15491951e+00 -5.28092861e-01
2.14436755e-01 -6.62182748e-01 6.02864206e-01 4.99614656e-01
-1.23701602e-01 -6.30226731e-01 -2.56388217e-01 7.95846954e-02
2.90215850e-01 -1.43714979e-01 -5.32542348e-01 2.56031066e-01
-1.29077351e+00 9.73323524e-01 4.64417934e-01 1.34333417e-01
-3.10901940e-01 -1.72236308e-01 3.38694960e-01 8.00740361e-01
1.72828913e-01 8.89618039e-01 -9.26603377e-01 -1.78980082e-01
-5.63209713e-01 3.96232545e-01 9.64481235e-01 7.83273578e-01
4.10243630e-01 -3.89951348e-01 -7.77824044e-01 8.86423111e-01
1.69401020e-01 6.16252959e-01 5.46537399e-01 -8.77871871e-01
1.25746584e+00 9.54837561e-01 -1.29181445e-01 -6.48805261e-01
-5.65356970e-01 -4.38207984e-01 -1.02802217e+00 -2.49170989e-01
7.91489840e-01 -1.30841881e-01 -6.71890855e-01 1.68020093e+00
6.99466914e-02 -1.21276639e-01 3.61101419e-01 7.57078826e-01
1.19181967e+00 8.87332916e-01 1.46778032e-01 -1.30288124e-01
1.75506556e+00 -1.48522818e+00 -7.54298210e-01 -4.80030447e-01
5.36907971e-01 -4.02069271e-01 1.72170079e+00 4.34768826e-01
-1.27833307e+00 -7.22995937e-01 -9.22261894e-01 -9.69689071e-01
-3.77806395e-01 -1.78487957e-01 2.24966586e-01 1.08826838e-01
-7.03656673e-01 1.80299118e-01 -1.22803733e-01 5.50929382e-02
5.72066844e-01 -1.43707208e-02 -8.17079917e-02 -6.85165584e-01
-1.66060591e+00 1.13400710e+00 3.95910800e-01 2.62997411e-02
-9.30655003e-01 -7.17014253e-01 -9.90231752e-01 4.74946976e-01
7.64597774e-01 -9.98920798e-01 1.73222458e+00 -6.53255165e-01
-1.58520949e+00 5.87010801e-01 -4.14903641e-01 -4.15543139e-01
1.79824829e-01 -6.34468973e-01 -3.96894366e-01 3.84212822e-01
1.63850069e-01 4.67062205e-01 9.49856877e-01 -7.36998558e-01
-5.83199203e-01 -3.44432533e-01 6.25861943e-01 3.73524785e-01
1.32449165e-01 -2.17738912e-01 -7.72101581e-02 -3.63793880e-01
1.80690233e-02 -1.86471879e-01 3.68640088e-02 -3.89702946e-01
-1.70583233e-01 -1.00635588e+00 2.40624174e-01 -1.14696741e+00
1.39310586e+00 -1.70217526e+00 2.78955579e-01 -3.82359266e-01
4.36224908e-01 2.84492314e-01 -6.32527292e-01 4.08196151e-01
1.58504710e-01 -2.37451419e-02 -4.03622210e-01 -4.74136285e-02
8.35567489e-02 -1.00176416e-01 -5.00720084e-01 -2.67852724e-01
6.32824540e-01 1.33427835e+00 -1.07969546e+00 -3.71156961e-01
-2.75239319e-01 -3.62848222e-01 -4.31685776e-01 5.32586634e-01
-9.24602568e-01 6.35026991e-01 -5.47418475e-01 5.15409946e-01
6.56024456e-01 -5.70812404e-01 -3.36121619e-01 6.52407929e-02
3.59694988e-01 7.91862249e-01 -5.93001306e-01 2.00365448e+00
-7.33156025e-01 4.22627777e-01 -3.32497776e-01 -1.11416566e+00
7.91171789e-01 1.46885112e-01 -4.12317038e-01 -1.35891938e+00
1.93286628e-01 1.05185866e-01 2.86698967e-01 -8.80409002e-01
3.97868812e-01 1.01920903e-01 -3.03609997e-01 6.52249455e-01
9.96102616e-02 -2.50371456e-01 3.36493194e-01 2.56861836e-01
1.20847964e+00 -1.92482665e-01 5.47424138e-01 -1.57984391e-01
1.09990728e+00 2.31250841e-02 2.40368769e-01 9.88466024e-01
-2.12405697e-01 5.56103468e-01 6.31690681e-01 -3.03332716e-01
-6.07420921e-01 -9.73128974e-01 1.52040899e-01 1.48977101e+00
8.00056905e-02 -1.46629170e-01 -9.47398484e-01 -8.08934510e-01
-2.27039322e-01 1.08118498e+00 -5.22052228e-01 -2.62129396e-01
-7.85265744e-01 -3.40405524e-01 4.39540416e-01 6.37337089e-01
1.20446646e+00 -1.43754172e+00 -7.48645306e-01 2.89392829e-01
-7.47479260e-01 -1.12949216e+00 -4.22357827e-01 -9.98552591e-02
-8.43667328e-01 -1.33074689e+00 -6.33731604e-01 -1.03269064e+00
3.60666633e-01 1.69492796e-01 1.76454687e+00 4.03857321e-01
2.07474202e-01 3.13076079e-01 -8.31040740e-01 -5.25220752e-01
-2.94694662e-01 4.83654350e-01 -4.93813694e-01 -2.18237355e-01
6.79974198e-01 -1.67921916e-01 -4.82036740e-01 1.89783499e-02
-9.25671160e-01 2.13138729e-01 7.69281387e-01 9.82700825e-01
4.06930953e-01 -1.21128693e-01 1.11742389e+00 -7.91026831e-01
1.40701115e+00 -5.74303985e-01 -2.04500288e-01 9.24822450e-01
-2.05448449e-01 1.76805794e-01 1.04669809e+00 -3.04050982e-01
-1.15549386e+00 -7.05581844e-01 -4.01756108e-01 2.75619894e-01
-2.56372094e-01 8.61826241e-01 -4.32955682e-01 3.25164855e-01
6.36201978e-01 4.97178704e-01 -5.58103770e-02 -4.83028114e-01
4.46254700e-01 6.95151567e-01 7.28343785e-01 -7.88137853e-01
5.99183202e-01 -2.42167309e-01 -1.50228873e-01 -5.07372737e-01
-1.61582875e+00 -2.95683175e-01 -8.39966774e-01 1.24893241e-01
1.11882913e+00 -7.68777609e-01 -1.09087956e+00 6.64333999e-01
-1.37920165e+00 -5.47889411e-01 2.05148906e-02 3.03499470e-03
-4.66751158e-01 4.25951570e-01 -5.30259907e-01 -5.62974572e-01
-8.02572012e-01 -1.01338243e+00 8.52071881e-01 5.12411356e-01
-1.29893541e-01 -1.13821042e+00 -1.90216437e-01 1.12734330e+00
6.10782444e-01 -9.44233984e-02 1.72536922e+00 -1.00187373e+00
-5.38289309e-01 1.82705551e-01 -2.71317482e-01 3.83438289e-01
-4.08126749e-02 -8.44174802e-01 -6.71932042e-01 -3.09219599e-01
3.51668209e-01 -1.08979607e+00 1.00318980e+00 8.07165056e-02
1.65731871e+00 -3.10816139e-01 3.08125526e-01 1.57734826e-01
1.03175378e+00 6.05149136e-04 7.39339709e-01 4.32194144e-01
5.30772090e-01 5.11725605e-01 6.40798569e-01 -2.78519988e-02
1.05987120e+00 1.20229319e-01 2.39362195e-01 2.52733797e-01
1.06154345e-02 -2.29019165e-01 2.51899511e-01 1.13829565e+00
2.84586072e-01 -3.88103396e-01 -1.19518673e+00 5.56063235e-01
-1.57580209e+00 -8.21401298e-01 -2.37921819e-01 1.90606630e+00
1.16463399e+00 1.43166319e-01 -1.43950105e-01 1.75617650e-01
1.59533903e-01 8.37436244e-02 -8.28357041e-01 -4.92931902e-01
-7.13992268e-02 7.13943124e-01 -3.14187080e-01 5.21721661e-01
-1.00520766e+00 8.37376833e-01 5.56146145e+00 8.61660063e-01
-5.93375385e-01 1.19082376e-01 6.43383324e-01 3.72735888e-01
-3.54654223e-01 -2.21374959e-01 -5.85196137e-01 2.84231544e-01
9.96870041e-01 -1.54650193e-02 4.39464211e-01 4.02737916e-01
-8.96397904e-02 -5.38623333e-01 -1.34698486e+00 7.36797750e-01
4.24216360e-01 -1.12849188e+00 4.98397380e-01 -7.62696803e-01
4.80331421e-01 -2.58709818e-01 -5.74824773e-02 9.45560515e-01
-7.64281955e-03 -1.46201479e+00 3.44758302e-01 6.61706150e-01
6.45332098e-01 -6.47763729e-01 9.52755630e-01 8.44915807e-01
-9.24394011e-01 -4.55465883e-01 -5.47364593e-01 -5.60277104e-01
2.00487629e-01 1.77914664e-01 -3.83356869e-01 7.09635794e-01
6.37653291e-01 4.26961094e-01 -1.15702689e+00 9.20718908e-01
-5.88054299e-01 7.60145247e-01 1.44540831e-01 -4.35473025e-01
2.16683552e-01 -2.05690060e-02 1.72192320e-01 7.58653998e-01
6.82677515e-03 5.65377116e-01 8.49085823e-02 7.53023744e-01
-3.08070868e-01 8.78580883e-02 1.48344800e-01 1.17199615e-01
4.29715753e-01 8.13634396e-01 6.68541109e-03 -3.73643190e-01
-6.32598877e-01 9.77173924e-01 8.26573253e-01 3.32784384e-01
-4.28569496e-01 -6.40131831e-01 1.79540738e-01 -2.80392736e-01
1.98523343e-01 -1.22681081e-01 -4.87638384e-01 -1.37569416e+00
2.18033507e-01 -1.58911383e+00 7.30064452e-01 -8.92484725e-01
-1.50961411e+00 4.38668370e-01 -8.34594369e-02 -9.46298420e-01
-1.96980789e-01 -8.08773041e-01 -9.17382300e-01 1.18164217e+00
-1.86928701e+00 -1.11916685e+00 -5.93070924e-01 7.79668450e-01
1.25266266e+00 -1.83957398e-01 8.61230373e-01 -9.33315456e-02
-4.11495864e-01 8.09835553e-01 6.71743508e-03 2.54703283e-01
6.47348583e-01 -1.52202272e+00 6.49046183e-01 7.91785419e-01
-1.50456250e-01 7.38323212e-01 2.42669627e-01 -3.92458230e-01
-1.39635575e+00 -8.01581442e-01 1.15722358e+00 -7.84248888e-01
4.93965387e-01 -2.71958441e-01 -1.50381899e+00 5.17460585e-01
5.54653645e-01 -5.94006300e-01 6.39215648e-01 2.46051818e-01
-3.83563638e-01 -1.05022311e-01 -9.09699976e-01 5.50350964e-01
7.39924073e-01 -6.42589748e-01 -1.59611309e+00 3.65490705e-01
1.24483693e+00 -5.89179754e-01 -8.90769899e-01 3.33690703e-01
4.91719097e-02 -5.92290699e-01 7.68078029e-01 -8.83565426e-01
8.59756529e-01 -1.90422103e-01 -6.38699997e-03 -1.54196966e+00
-3.40361685e-01 -3.76148373e-01 -4.23084110e-01 9.37338531e-01
6.18717730e-01 -5.03382683e-01 9.18304473e-02 4.91439968e-01
-2.56740451e-01 -9.83642280e-01 -8.60680640e-01 -3.38836998e-01
8.41927171e-01 -3.32685947e-01 8.34925830e-01 6.83178604e-01
-5.72772399e-02 1.11263847e+00 8.34726244e-02 1.06792279e-01
2.99434483e-01 3.28673601e-01 6.19745493e-01 -1.29663086e+00
-1.38447002e-01 -3.33007962e-01 2.68271804e-01 -1.79757297e+00
2.42290854e-01 -8.21267188e-01 5.61438547e-03 -1.95739007e+00
1.47810534e-01 3.80949862e-02 -1.14197344e-01 2.58870542e-01
-8.59371245e-01 -6.00212336e-01 1.82101429e-01 -4.88140546e-02
-1.10762942e+00 7.49850094e-01 1.91971636e+00 -3.92462075e-01
1.01943396e-01 1.07798606e-01 -8.95400345e-01 3.53509665e-01
7.74651229e-01 -6.82321414e-02 -7.15070605e-01 -9.80239272e-01
3.26300353e-01 3.60136569e-01 2.22539142e-01 -9.95052278e-01
4.82336164e-01 -1.76937297e-01 5.33309340e-01 -7.53910244e-01
1.16949819e-01 -3.74894321e-01 -9.17328894e-01 3.02054226e-01
-8.00737858e-01 4.89503264e-01 1.76124573e-01 4.82398331e-01
-4.38978940e-01 -5.13169110e-01 4.86576796e-01 -3.84986043e-01
-8.64946187e-01 1.22350499e-01 -2.35417888e-01 1.02253354e+00
5.79137981e-01 2.01790690e-01 -8.83174002e-01 -6.67890310e-01
-3.60901028e-01 1.07529557e+00 -1.71317339e-01 5.96276104e-01
8.89108956e-01 -9.05340314e-01 -1.17239535e+00 -1.62704200e-01
2.24093601e-01 7.49694049e-01 4.91460890e-01 5.43954134e-01
-6.53954089e-01 6.46416545e-01 -1.47631004e-01 -3.98095429e-01
-9.01224196e-01 5.89274704e-01 3.18154603e-01 -6.98487163e-01
-4.63881075e-01 9.57370758e-01 1.42133400e-01 -6.87175512e-01
2.00139001e-01 -6.82865977e-01 -8.25580835e-01 -1.72225144e-02
1.01741791e+00 3.02096426e-01 2.27373820e-02 -7.77713805e-02
-7.72058358e-03 4.34605837e-01 -4.38133031e-01 1.06531873e-01
9.81732130e-01 -2.10281298e-01 -3.48599225e-01 4.52995270e-01
8.90984654e-01 -4.61788982e-01 -9.94719982e-01 -4.99510616e-01
2.22745821e-01 9.07690078e-03 -2.69334078e-01 -1.38861847e+00
-2.36280262e-01 1.40329456e+00 2.14527994e-02 -5.97311482e-02
1.40024757e+00 1.00872014e-03 1.25562286e+00 1.06912076e+00
2.23845586e-01 -1.02576363e+00 5.30800164e-01 1.22752690e+00
1.38441694e+00 -1.46033251e+00 -1.92538306e-01 -2.31870979e-01
-7.28707194e-01 1.25617325e+00 1.15618253e+00 9.23460722e-02
1.83131739e-01 -6.17770195e-01 1.44304663e-01 -1.38096794e-01
-1.06371891e+00 -6.81703687e-02 5.55802763e-01 5.11225104e-01
3.14332426e-01 -1.28063977e-01 -2.64917940e-01 1.09115314e+00
-4.14066643e-01 -2.04673663e-01 4.71540838e-01 8.35089386e-01
-7.17790008e-01 -7.79193163e-01 -3.11223328e-01 7.62996137e-01
-2.93415040e-01 -4.55161273e-01 -3.51129979e-01 4.90881890e-01
-1.97369426e-01 1.48210633e+00 -4.20412980e-03 -1.35023892e-01
4.59971040e-01 3.68833393e-01 5.45029283e-01 -6.36586666e-01
-8.29155564e-01 -8.61673832e-01 3.21205668e-02 -2.92787075e-01
-2.30337933e-01 -4.63643819e-01 -1.22955143e+00 -1.38237774e-01
-1.09342121e-01 1.92503050e-01 -4.17020312e-03 1.68421888e+00
2.45164916e-01 7.49054074e-01 2.90705681e-01 2.92839669e-02
-1.04434657e+00 -1.47497571e+00 5.88376485e-02 5.35090625e-01
6.32153928e-01 -3.28775972e-01 -7.19784424e-02 -1.00999609e-01] | [11.273195266723633, 8.088088989257812] |
71942676-4794-442e-826d-cf23aab1ed7b | community-detection-via-hashtag-graphs-for | 2111.10401 | null | https://arxiv.org/abs/2111.10401v1 | https://arxiv.org/pdf/2111.10401v1.pdf | Community-Detection via Hashtag-Graphs for Semi-Supervised NMF Topic Models | Extracting topics from large collections of unstructured text-documents has become a central task in current NLP applications and algorithms like NMF, LDA as well as their generalizations are the well-established current state of the art. However, especially when it comes to short text documents like Tweets, these approaches often lead to unsatisfying results due to the sparsity of the document-feature matrices. Even though, several approaches have been proposed to overcome this sparsity by taking additional information into account, these are merely focused on the aggregation of similar documents and the estimation of word-co-occurrences. This ultimately completely neglects the fact that a lot of topical-information can be actually retrieved from so-called hashtag-graphs by applying common community detection algorithms. Therefore, this paper outlines a novel approach on how to integrate topic structures of hashtag graphs into the estimation of topic models by connecting graph-based community detection and semi-supervised NMF. By applying this approach on recently streamed Twitter data it will be seen that this procedure actually leads to more intuitive and humanly interpretable topics. | ['Benjamin Säfken', 'Christoph Weisser', 'Anton Thielmann', 'Mattias Luber'] | 2021-11-17 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [ 6.41335100e-02 3.12180132e-01 3.51358801e-02 -2.17368871e-01
-6.25641763e-01 -5.31742632e-01 1.00347841e+00 9.36918497e-01
-3.95528585e-01 7.40259826e-01 3.38377774e-01 -1.44013032e-01
-2.88617283e-01 -9.72389638e-01 -2.10305944e-01 -7.41525233e-01
-3.79639804e-01 8.44947398e-01 2.41165802e-01 -1.36739969e-01
4.19841558e-01 1.78933561e-01 -1.66165757e+00 -1.97847575e-01
1.00360024e+00 4.10512209e-01 1.24731593e-01 2.50894219e-01
-7.24585652e-01 3.25175613e-01 -5.19695699e-01 -5.00792742e-01
-1.37037784e-01 -3.08773816e-01 -6.97008371e-01 4.11067754e-01
1.63943559e-01 1.57112598e-01 2.89387077e-01 1.07780612e+00
1.58493415e-01 -3.78061496e-02 8.20892751e-01 -1.03710544e+00
2.33810008e-01 6.80382133e-01 -7.41885602e-01 -3.23402137e-02
4.35933679e-01 -4.54171419e-01 1.16169453e+00 -6.70797825e-01
8.59195173e-01 1.19552636e+00 4.45453376e-01 -1.32844746e-01
-1.44771969e+00 -3.04297984e-01 2.12062702e-01 1.38207644e-01
-1.48871839e+00 -7.91371688e-02 9.14243639e-01 -6.52388096e-01
6.24812245e-01 2.84867913e-01 6.98697269e-01 7.10620880e-01
-7.29775801e-02 7.29418278e-01 1.11762691e+00 -4.48305339e-01
4.01665002e-01 5.90430439e-01 4.56492871e-01 3.18160236e-01
6.59650505e-01 -5.15620112e-01 -3.56545091e-01 -5.23038149e-01
8.47251341e-02 -1.03471771e-01 -1.58524096e-01 -6.26264632e-01
-1.06306231e+00 1.14215195e+00 1.77241012e-01 8.65234911e-01
-4.83513504e-01 -3.19726616e-01 5.16934335e-01 1.47609249e-01
1.10969782e+00 2.81991541e-01 5.81354387e-02 8.70186985e-02
-1.26733196e+00 4.51805055e-01 1.08127081e+00 5.78516841e-01
1.10022557e+00 -3.70796829e-01 3.48086119e-01 5.26041090e-01
4.07104850e-01 2.36211777e-01 2.66714931e-01 -1.88187316e-01
4.93832141e-01 9.76616561e-01 4.19902392e-02 -1.57206643e+00
-4.09945488e-01 -6.60758317e-01 -7.60345519e-01 -8.74500945e-02
5.57369769e-01 -1.31560579e-01 -4.65024769e-01 1.32483459e+00
6.08227789e-01 -9.49591026e-02 -1.99961215e-01 7.43154883e-01
4.98648912e-01 5.48228621e-01 2.51498461e-01 -5.62409401e-01
1.36690879e+00 -4.44929183e-01 -8.25258315e-01 1.83003414e-02
5.55728912e-01 -7.18936563e-01 6.57582045e-01 5.45246243e-01
-5.63239515e-01 -2.16838613e-01 -7.83593655e-01 2.14542136e-01
-7.60214925e-01 -1.42082140e-01 6.93624496e-01 7.35032678e-01
-9.48330462e-01 6.40667856e-01 -6.90556467e-01 -7.78232694e-01
1.18450992e-01 3.74403358e-01 -3.58088374e-01 -4.59154844e-02
-1.04929435e+00 7.31978953e-01 6.29397810e-01 7.59504959e-02
-2.57320851e-01 -2.70691425e-01 -5.41885078e-01 1.69178173e-01
6.14174306e-01 -5.35855412e-01 6.48452520e-01 -7.51944959e-01
-1.01520741e+00 8.60043406e-01 -1.85016766e-01 -5.98619521e-01
4.75559026e-01 -1.24003381e-01 -2.45672047e-01 2.18633562e-01
7.03666583e-02 2.46406555e-01 8.55147302e-01 -1.20555091e+00
-4.45844829e-01 -4.86707568e-01 -2.31574938e-01 2.16112986e-01
-7.42493510e-01 3.05795427e-02 -1.99531823e-01 -5.46795368e-01
3.99122626e-01 -8.82246733e-01 -4.34329063e-01 -4.03984785e-01
-6.63189948e-01 -5.45974076e-01 6.77077115e-01 -4.55630422e-01
1.42814076e+00 -1.91983628e+00 2.65979439e-01 5.71191430e-01
5.35602748e-01 1.08163968e-01 3.22870880e-01 9.32428658e-01
2.22446784e-01 -7.82683678e-03 -2.59183526e-01 -4.54542726e-01
5.59452735e-02 -4.00184132e-02 -2.78544903e-01 5.70566237e-01
5.92656210e-02 3.12304050e-01 -1.03154993e+00 -7.93423116e-01
2.87613511e-01 5.76835632e-01 -3.95696133e-01 -1.68542176e-01
-4.11991537e-01 3.56341094e-01 -7.76831150e-01 1.61290154e-01
6.37284935e-01 -2.95097888e-01 5.07134736e-01 1.21376991e-01
-3.11976939e-01 4.67223704e-01 -1.31470263e+00 1.45221007e+00
-5.73561452e-02 6.36031985e-01 -2.00669318e-02 -1.30738735e+00
9.62808549e-01 3.64362776e-01 7.43532121e-01 -3.83645594e-02
1.51399553e-01 2.45610878e-01 -1.44136414e-01 -2.48870313e-01
7.71781802e-01 -1.10955723e-01 7.99576640e-02 6.97953045e-01
-4.13024947e-02 -3.34016941e-02 6.09898031e-01 4.25783873e-01
7.85813808e-01 -7.48078749e-02 6.13252223e-01 -5.51074743e-01
7.55456865e-01 1.95516810e-01 -4.56318520e-02 4.32940811e-01
2.13247433e-01 3.80648226e-01 5.59111416e-01 -2.64622539e-01
-8.27834249e-01 -6.64372623e-01 -1.66679099e-01 8.28481019e-01
-2.15581134e-01 -8.17439675e-01 -8.25813174e-01 -6.44135356e-01
-1.34348879e-02 4.18656915e-01 -5.27037561e-01 3.15083325e-01
-3.09907705e-01 -1.01083803e+00 -1.52780367e-02 -1.98336735e-01
4.85488921e-02 -8.39114726e-01 -2.95562238e-01 3.87681693e-01
-2.81100512e-01 -9.51634824e-01 3.96940745e-02 5.13192937e-02
-1.18407726e+00 -1.01039767e+00 -9.27795827e-01 -3.43708396e-01
7.41472602e-01 4.10423219e-01 1.03274453e+00 1.51194995e-02
-1.61939144e-01 1.76174834e-01 -4.60253894e-01 -3.52950633e-01
-4.00396615e-01 4.93217617e-01 -1.04337901e-01 3.01237464e-01
6.43016338e-01 -8.82809699e-01 -3.63409549e-01 6.01780750e-02
-8.62502933e-01 -2.14025319e-01 5.38973391e-01 3.64375561e-01
2.06269935e-01 1.62562236e-01 5.70835531e-01 -1.26329577e+00
8.15507531e-01 -5.52537441e-01 -5.51435709e-01 8.63042921e-02
-8.90712798e-01 9.59904492e-02 3.87639225e-01 -2.70813584e-01
-9.11029935e-01 -2.40015741e-02 -5.78681193e-03 9.96664539e-02
-3.70525986e-01 8.58034253e-01 1.40330964e-03 1.66636765e-01
5.25768280e-01 3.66810173e-01 -1.27667785e-01 -6.03192508e-01
4.16806936e-01 7.86947310e-01 -9.88535509e-02 -2.76563674e-01
8.50184619e-01 6.45455658e-01 8.73786979e-04 -1.36243296e+00
-6.19653165e-01 -1.11034727e+00 -5.39307535e-01 -2.97381520e-01
6.29710793e-01 -8.36341619e-01 -5.13063669e-01 1.56497866e-01
-1.01958406e+00 3.51427615e-01 -9.14167762e-02 5.37513196e-01
-1.71582609e-01 8.53391528e-01 -3.25706065e-01 -1.16611612e+00
-3.25244635e-01 -7.16188788e-01 1.03723502e+00 -1.28785133e-01
-3.94756436e-01 -1.15681446e+00 3.70917559e-01 2.34713435e-01
3.43312949e-01 1.81489050e-01 9.99929845e-01 -1.06618249e+00
-4.50341403e-01 -4.61777151e-01 -2.44162694e-01 6.71458663e-03
1.24533497e-01 -1.29189808e-03 -9.57923591e-01 -2.66165942e-01
-5.37464879e-02 1.38449922e-01 8.49681735e-01 2.71202266e-01
3.00265074e-01 -2.59111226e-01 -7.08028436e-01 -1.43389091e-01
1.44438350e+00 -3.71677339e-01 4.40990210e-01 2.70708084e-01
4.64159280e-01 1.08525860e+00 5.65302551e-01 5.45389295e-01
3.31816047e-01 7.63699830e-01 3.63341063e-01 4.31638323e-02
3.13842982e-01 -1.79112464e-01 2.06954360e-01 1.11153269e+00
-2.42371917e-01 -3.30676973e-01 -9.17743444e-01 6.00169063e-01
-1.74244618e+00 -1.02915370e+00 -6.73999727e-01 2.33026600e+00
5.69377601e-01 3.39973807e-01 4.44229066e-01 4.67632562e-01
9.32114482e-01 2.93828905e-01 1.92181528e-01 -3.12450062e-02
-3.72076482e-02 -9.53299403e-02 2.33720914e-01 4.28971559e-01
-1.19355094e+00 6.54588580e-01 5.36774111e+00 8.59048605e-01
-9.26024675e-01 8.66590068e-02 2.22131386e-01 3.01054507e-01
-3.60902578e-01 2.64543325e-01 -8.61515820e-01 3.78141195e-01
8.05692315e-01 -2.85395622e-01 -1.44918904e-01 8.35641205e-01
3.70777249e-01 -5.64862728e-01 -7.02304006e-01 6.55131400e-01
2.14361295e-01 -8.99079502e-01 8.94692987e-02 4.61010963e-01
6.26538575e-01 -8.08512866e-02 -3.53180945e-01 5.02830893e-02
5.55005781e-02 -4.47595865e-01 3.73852134e-01 2.79416621e-01
2.07419574e-01 -5.96331358e-01 8.07072520e-01 5.37685812e-01
-1.10560381e+00 1.74052656e-01 -3.94654393e-01 -1.00952663e-01
3.63966763e-01 1.29961908e+00 -1.13734519e+00 7.70024896e-01
3.52736503e-01 7.13605523e-01 -5.22141039e-01 1.27528238e+00
-8.22798088e-02 6.81111276e-01 -6.15180254e-01 -4.56667215e-01
3.20561796e-01 -5.02807915e-01 8.24874818e-01 1.32712245e+00
2.18868509e-01 -4.40550119e-01 1.90372825e-01 6.87212884e-01
3.21553469e-01 7.62909591e-01 -7.06641436e-01 -2.12301314e-01
-9.79113579e-02 1.54954588e+00 -1.43521225e+00 -4.16734904e-01
-4.29647863e-01 4.95887220e-01 2.35028043e-01 7.96955973e-02
-2.73732305e-01 -2.41956010e-01 2.27982581e-01 5.93682408e-01
2.54004806e-01 -5.17597437e-01 1.52402744e-01 -1.10551620e+00
-4.26270962e-02 -5.67424417e-01 3.12853426e-01 -2.49860942e-01
-1.24131858e+00 5.88952601e-01 2.96260864e-01 -1.21693313e+00
-4.68015313e-01 -2.77245671e-01 -5.26217520e-01 5.15310228e-01
-1.34527671e+00 -9.24538076e-01 -1.73802271e-01 5.20783961e-01
4.20870900e-01 1.70654580e-01 8.94220650e-01 4.55910683e-01
-2.54689813e-01 -1.04761474e-01 2.83757716e-01 -1.16966143e-01
5.38648725e-01 -1.27026772e+00 1.82908729e-01 7.05834508e-01
5.17721653e-01 7.07652330e-01 1.19864225e+00 -7.19865322e-01
-9.80910540e-01 -6.93338037e-01 1.41991198e+00 -1.43203244e-01
9.89267766e-01 -6.17002428e-01 -1.02509451e+00 2.57813454e-01
3.76772165e-01 -3.62109363e-01 7.20834970e-01 5.11342645e-01
-2.28710175e-02 6.37575090e-02 -5.50738037e-01 5.30596018e-01
6.18882239e-01 -3.88813406e-01 -5.99294186e-01 6.74348891e-01
1.91972554e-01 2.30960831e-01 -6.03929579e-01 2.47559678e-02
2.42379487e-01 -1.08637786e+00 6.19615555e-01 -2.90087551e-01
-1.03002768e-02 -3.21854979e-01 2.88539857e-01 -1.20868123e+00
1.16781458e-01 -6.36278212e-01 1.27810299e-01 1.37934053e+00
3.87406677e-01 -6.45119786e-01 1.01813126e+00 2.31237963e-01
3.64544511e-01 -2.78831363e-01 -7.63427198e-01 -5.27286291e-01
-2.31287703e-01 -3.13831359e-01 1.11395232e-01 1.05774534e+00
5.07863581e-01 6.19755149e-01 -2.86593586e-01 -1.97582215e-01
7.58735955e-01 4.60475206e-01 9.04799044e-01 -1.98619080e+00
-1.05347103e-02 -3.92245620e-01 -5.35946012e-01 -6.68039560e-01
-6.56530820e-03 -7.45999992e-01 -4.25266065e-02 -1.47292042e+00
3.09919626e-01 -4.65089649e-01 1.40602693e-01 -7.84055740e-02
-5.19791134e-02 1.38419613e-01 1.34408716e-02 3.40448707e-01
-8.53627503e-01 2.78209656e-01 9.36044216e-01 6.69245198e-02
-2.68493026e-01 3.47059309e-01 -2.96501577e-01 7.66859531e-01
6.72650099e-01 -7.84614801e-01 -3.45951706e-01 1.77399248e-01
6.21833026e-01 -8.64129588e-02 2.87662834e-01 -8.28355372e-01
4.30810153e-01 1.71714529e-01 -2.06872016e-01 -7.46592820e-01
1.43178314e-01 -8.62503111e-01 3.06173384e-01 3.74138385e-01
-2.82460123e-01 -1.00603431e-01 -1.67545870e-01 8.49188268e-01
-5.08398473e-01 -3.72754782e-01 1.49696827e-01 -3.40420127e-01
-4.02597100e-01 -5.33827618e-02 -7.01414049e-01 -1.53833032e-01
8.52834046e-01 -1.95707858e-01 -4.71249148e-02 -6.44969642e-01
-9.34058785e-01 -3.01293265e-02 3.19725275e-01 1.57591343e-01
2.54743159e-01 -7.95326233e-01 -7.69155443e-01 -1.42643541e-01
1.23452470e-01 -2.65217181e-02 2.13964447e-01 1.14846420e+00
-2.54371852e-01 7.62798011e-01 2.13187218e-01 -7.38857210e-01
-1.35095882e+00 7.11973965e-01 -3.76989871e-01 -6.89559102e-01
-6.57078683e-01 2.65919238e-01 1.06963605e-01 -8.01174566e-02
2.21829489e-01 -6.56988984e-03 -6.91115677e-01 6.44005299e-01
2.62567580e-01 2.56686836e-01 2.53773987e-01 -6.79165006e-01
-2.30169654e-01 5.62555134e-01 -1.56114966e-01 -7.45609775e-02
1.39387023e+00 -3.91333550e-01 -3.49767089e-01 5.52026033e-01
1.01016796e+00 2.56717831e-01 -6.17984474e-01 -1.62967086e-01
7.20384538e-01 -1.63924113e-01 -1.34476915e-01 -1.17246024e-01
-7.74738371e-01 1.07705271e+00 9.74857509e-02 9.40355599e-01
7.84507632e-01 7.93828443e-02 2.85340816e-01 4.87455547e-01
4.71981317e-01 -9.78925467e-01 -2.48066694e-01 6.41980097e-02
5.60646534e-01 -1.18360066e+00 4.25454408e-01 -8.50789368e-01
-2.25455031e-01 1.19010925e+00 -9.33809951e-02 -1.59942746e-01
8.10961187e-01 -8.79588574e-02 -1.29003242e-01 -3.34754497e-01
-4.48024750e-01 -5.53858459e-01 3.11329246e-01 3.39638382e-01
5.93203485e-01 -6.51997700e-02 -9.54682112e-01 1.06606752e-01
-1.14576302e-01 -1.78919256e-01 4.11391526e-01 7.58866608e-01
-6.23285949e-01 -1.27903700e+00 -4.22627121e-01 4.57769960e-01
-7.15896308e-01 -6.86570629e-02 -4.54054296e-01 1.13783884e+00
-1.30488366e-01 8.82887125e-01 -1.45748958e-01 -4.49199863e-02
-1.72245339e-01 1.54408514e-01 1.51674092e-01 -7.99061537e-01
-5.61690807e-01 4.30098176e-01 2.16715068e-01 -2.34358579e-01
-9.27343726e-01 -9.14961219e-01 -7.96482086e-01 -1.37711391e-01
-7.67015815e-01 7.86613166e-01 9.11148787e-01 1.05865085e+00
1.91851556e-01 1.25182077e-01 3.72913778e-01 -1.00094700e+00
-4.10281062e-01 -1.14087355e+00 -7.83342600e-01 4.55469906e-01
1.51057048e-02 -6.92776561e-01 -5.77994585e-01 1.96305830e-02] | [10.30681324005127, 7.32743501663208] |
e1bed9cd-21fd-4c14-a5b2-61aec5719cbb | panonet3d-combining-semantic-and-geometric | 2012.09418 | null | https://arxiv.org/abs/2012.09418v1 | https://arxiv.org/pdf/2012.09418v1.pdf | PanoNet3D: Combining Semantic and Geometric Understanding for LiDARPoint Cloud Detection | Visual data in autonomous driving perception, such as camera image and LiDAR point cloud, can be interpreted as a mixture of two aspects: semantic feature and geometric structure. Semantics come from the appearance and context of objects to the sensor, while geometric structure is the actual 3D shape of point clouds. Most detectors on LiDAR point clouds focus only on analyzing the geometric structure of objects in real 3D space. Unlike previous works, we propose to learn both semantic feature and geometric structure via a unified multi-view framework. Our method exploits the nature of LiDAR scans -- 2D range images, and applies well-studied 2D convolutions to extract semantic features. By fusing semantic and geometric features, our method outperforms state-of-the-art approaches in all categories by a large margin. The methodology of combining semantic and geometric features provides a unique perspective of looking at the problems in real-world 3D point cloud detection. | ['Martial Hebert', 'David Held', 'Jianren Wang', 'Xia Chen'] | 2020-12-17 | null | null | null | null | ['cloud-detection'] | ['computer-vision'] | [ 1.32353038e-01 -3.23455542e-01 -1.56671777e-01 -7.56399810e-01
-3.26876909e-01 -8.06663752e-01 1.00575185e+00 3.42372656e-01
-3.83307964e-01 -9.21811461e-02 -2.97724903e-01 -2.90184796e-01
-9.37488899e-02 -9.80074048e-01 -8.90273035e-01 -4.02533591e-01
3.59878018e-02 7.91118681e-01 6.76873326e-01 -4.84989166e-01
5.60317159e-01 1.04003215e+00 -2.22850156e+00 -2.99028419e-02
2.81267315e-01 1.04315114e+00 3.88018876e-01 4.69448954e-01
-5.49302280e-01 3.05851046e-02 -2.17795223e-02 -1.59110017e-02
4.56158638e-01 3.65896910e-01 -2.75256127e-01 2.75945276e-01
8.25202584e-01 -1.52098477e-01 -1.03007816e-01 1.17833316e+00
1.35831609e-01 -2.26043105e-01 7.42666245e-01 -1.48128259e+00
-3.73339266e-01 -2.52604127e-01 -6.14480317e-01 1.64604038e-01
4.67054218e-01 -2.72959620e-02 1.00254738e+00 -1.34817839e+00
6.02535963e-01 1.52582967e+00 5.76933980e-01 2.30192676e-01
-9.33464468e-01 -4.62466180e-01 3.33420843e-01 2.56658792e-01
-1.28423131e+00 -2.23273575e-01 9.43869650e-01 -7.74121702e-01
8.91095638e-01 1.24676876e-01 5.08823276e-01 6.82929516e-01
1.88574880e-01 5.82881153e-01 1.15059912e+00 -3.20507407e-01
1.58311501e-01 2.83571601e-01 2.90163755e-01 7.23338127e-01
4.62079704e-01 4.06834334e-01 -4.87746030e-01 -7.96141848e-02
6.32909358e-01 4.73308027e-01 1.45600572e-01 -1.15072155e+00
-1.04761326e+00 7.65900552e-01 4.25398827e-01 9.00133625e-02
-4.72267717e-02 2.47737676e-01 1.27974793e-01 2.22521663e-01
5.28991699e-01 -2.04960629e-01 -4.42731142e-01 1.61290094e-01
-5.59364021e-01 3.40914279e-01 3.90158713e-01 1.39519167e+00
1.36162841e+00 -3.01805556e-01 5.95580935e-01 3.14183444e-01
5.87318659e-01 1.04200029e+00 2.48836987e-02 -9.38885927e-01
4.22233015e-01 8.40904593e-01 -1.07531967e-02 -7.16886878e-01
-3.06803167e-01 -1.29887700e-01 -1.60932913e-01 9.28500354e-01
1.84656158e-01 4.91080821e-01 -1.05535650e+00 1.36389065e+00
4.85106111e-01 -5.39381020e-02 3.85945141e-02 8.36664975e-01
8.75464916e-01 2.59790719e-01 -3.96215677e-01 4.12730306e-01
1.46126437e+00 -3.45771432e-01 -1.93057299e-01 -6.02445185e-01
4.04954672e-01 -6.64525986e-01 6.41641676e-01 4.32247035e-02
-1.09718740e+00 -8.32245529e-01 -1.29067194e+00 -3.75566065e-01
-7.72884667e-01 -9.61890668e-02 2.69180685e-01 4.85777855e-01
-1.07455266e+00 3.85618806e-01 -5.93285441e-01 -6.36199117e-01
3.07352275e-01 2.58028746e-01 -5.38566649e-01 -3.61333251e-01
-4.86414433e-01 1.29145956e+00 2.63040036e-01 -2.47282803e-01
-6.08028889e-01 -6.93445683e-01 -1.13416851e+00 -7.76283816e-02
4.27151740e-01 -1.03883410e+00 1.06922305e+00 -3.41197222e-01
-8.28142881e-01 1.40599751e+00 -5.64054668e-01 -2.13310406e-01
5.45612693e-01 -2.78046668e-01 -1.34174585e-01 1.45169631e-01
1.78209364e-01 6.21629179e-01 9.89891768e-01 -1.78481042e+00
-9.66789663e-01 -1.16844952e+00 4.97833081e-02 3.81835639e-01
4.91837412e-01 -2.68142015e-01 -2.63624042e-01 2.26897702e-01
6.82788968e-01 -8.15065145e-01 -2.63925701e-01 2.92762339e-01
-9.80337989e-03 -2.48087928e-01 1.30113864e+00 1.80941164e-01
1.95784703e-01 -2.34493470e+00 -1.19762994e-01 1.52074382e-01
4.58661795e-01 6.86848685e-02 8.38152096e-02 3.89567226e-01
-5.41166551e-02 -5.98377921e-02 -4.32696119e-02 -4.34032619e-01
1.11893319e-01 4.76858556e-01 -5.69836676e-01 7.29547799e-01
4.48344082e-01 9.59841907e-01 -9.94077802e-01 -3.31099689e-01
8.29307139e-01 4.19699341e-01 -1.11783743e-01 -3.87751758e-02
-2.86297858e-01 1.05595961e-01 -7.72263885e-01 5.89553595e-01
1.04043937e+00 9.65626240e-02 -4.07517642e-01 1.55067533e-01
-4.51249152e-01 2.20281392e-01 -1.04928803e+00 1.85368574e+00
-4.47330564e-01 6.31960452e-01 3.15300077e-02 -8.99553120e-01
1.12146223e+00 -5.96149452e-02 4.10849869e-01 -4.43379968e-01
1.48056273e-03 3.02008510e-01 -3.78359556e-01 -4.18495417e-01
6.00273848e-01 -1.64390087e-01 -1.02296755e-01 1.12524875e-01
-8.89374912e-02 -7.85292089e-01 -2.25738361e-01 2.53160316e-02
7.99095869e-01 3.15079242e-01 3.38024706e-01 -1.63243532e-01
4.64877307e-01 1.52589634e-01 5.76964691e-02 7.39310384e-01
-2.90547386e-02 6.92416608e-01 1.17814153e-01 -5.59005260e-01
-1.19638872e+00 -1.61244869e+00 -3.40723664e-01 6.48933351e-01
6.89637542e-01 -2.05429524e-01 -2.30434746e-01 -4.57485884e-01
5.62629938e-01 6.13074660e-01 -5.17321706e-01 -1.12860158e-01
-4.69006270e-01 -2.98357266e-03 7.94745088e-02 5.50890088e-01
1.06540143e-01 -4.39380944e-01 -1.21724451e+00 -1.18518114e-01
4.07191664e-01 -1.65252507e+00 -1.19193397e-01 2.92694598e-01
-1.19268405e+00 -1.10537267e+00 1.73364319e-02 -3.90906632e-01
4.90108192e-01 1.07379162e+00 1.30219328e+00 -6.01895563e-02
-4.77277398e-01 9.37403858e-01 -2.19922379e-01 -9.79782760e-01
-2.05820635e-01 -4.58343327e-01 -4.54038605e-02 1.04807764e-01
9.36102927e-01 -7.97847629e-01 -3.60922039e-01 1.57730997e-01
-6.34406567e-01 -1.40840188e-01 7.00249135e-01 2.29275331e-01
6.74960315e-01 -2.46887863e-01 -1.65050521e-01 -6.18856549e-01
-5.18379100e-02 -3.22026789e-01 -8.21674943e-01 -1.24747738e-01
-5.58014750e-01 9.63432938e-02 1.52986735e-01 1.09856084e-01
-5.18006623e-01 4.41902995e-01 2.16485068e-01 -9.44106877e-01
-8.41943085e-01 -1.09709119e-02 -3.38537753e-01 -1.19132020e-01
3.65493953e-01 2.94442773e-01 1.11043297e-01 -5.57438374e-01
7.46587634e-01 3.66213262e-01 5.13477266e-01 -4.22264516e-01
1.24966359e+00 1.22008717e+00 6.21900082e-01 -1.09059441e+00
-7.30559587e-01 -1.17733347e+00 -1.20495081e+00 -1.83556959e-01
9.48335230e-01 -1.11542928e+00 -6.20354235e-01 1.79873928e-01
-1.39626193e+00 4.65184540e-01 -5.29873252e-01 4.41931188e-01
-8.70215535e-01 3.70459080e-01 1.98535636e-01 -8.69933426e-01
3.07841510e-01 -1.19417858e+00 1.92766345e+00 -1.80775523e-01
3.27016950e-01 -8.73276114e-01 -6.16867989e-02 2.33765230e-01
2.99256574e-02 3.05558145e-01 1.03882968e+00 -4.67378438e-01
-1.04451418e+00 -1.36879116e-01 -4.73456919e-01 1.22215033e-01
-1.07031576e-01 -3.13962549e-02 -1.38826370e+00 6.87036589e-02
2.56396025e-01 3.91629338e-02 1.09738815e+00 4.21339005e-01
8.05903256e-01 6.14261150e-01 -5.99156022e-01 6.73066437e-01
1.77126980e+00 7.71932155e-02 2.20517576e-01 1.03356473e-01
7.75602102e-01 8.36906850e-01 6.69409573e-01 3.57387841e-01
5.55782497e-01 6.58069849e-01 1.12499988e+00 -4.25607376e-02
5.35077900e-02 -4.42432761e-01 1.95705250e-01 3.67896169e-01
-5.37289754e-02 1.76002920e-01 -1.08461642e+00 5.01037359e-01
-1.73194420e+00 -9.24191654e-01 -5.20298541e-01 2.25824404e+00
-1.21660501e-01 3.81904840e-01 4.58035022e-02 1.02613837e-01
6.59016490e-01 1.10519454e-01 -7.55114675e-01 -3.14967155e-01
-7.23278821e-02 8.83359462e-02 8.22669089e-01 4.41841543e-01
-1.13364625e+00 7.27488339e-01 6.02966976e+00 5.15093923e-01
-9.79510725e-01 5.26191816e-02 -2.00283200e-01 9.38961059e-02
-5.68363249e-01 1.99652642e-01 -1.08514094e+00 -4.18865558e-04
6.04480565e-01 -8.30823332e-02 -1.65582541e-02 1.02818704e+00
-3.87797244e-02 -4.53716256e-02 -1.46476448e+00 1.14382792e+00
1.96838215e-01 -1.27280056e+00 1.25191659e-01 4.75498080e-01
2.78245777e-01 6.58147097e-01 1.66177332e-01 -1.51112705e-01
3.37145984e-01 -8.90925348e-01 1.10542226e+00 4.40198839e-01
6.63731635e-01 -4.45932686e-01 4.19487178e-01 6.40265644e-01
-1.40839148e+00 -1.27006248e-01 -4.15512919e-01 -2.14173958e-01
1.09734826e-01 6.88273251e-01 -8.13067913e-01 7.56991148e-01
7.31468618e-01 1.07483530e+00 -5.74316680e-01 9.00785327e-01
4.20943983e-02 -1.35944247e-01 -4.96206880e-01 1.35470673e-01
4.61430371e-01 -4.72169280e-01 7.95673490e-01 9.50500131e-01
3.36362541e-01 -1.16253778e-01 4.13721710e-01 1.11206949e+00
4.14361328e-01 -3.67606878e-01 -1.58862960e+00 4.03879404e-01
4.28079903e-01 1.05694175e+00 -5.88492274e-01 -1.34121448e-01
-7.62043178e-01 3.52912903e-01 -1.73426624e-02 2.45426714e-01
-4.11744714e-01 -8.32890719e-02 1.06251967e+00 4.00681615e-01
6.31598413e-01 -8.63825619e-01 -5.50399065e-01 -9.41053152e-01
3.45753670e-01 3.29955183e-02 -1.74713567e-01 -1.08578455e+00
-1.12985003e+00 3.97627354e-01 3.58475506e-01 -1.63912892e+00
-1.53122589e-01 -1.07977247e+00 -5.02290070e-01 9.41350818e-01
-2.07489514e+00 -1.37713015e+00 -5.42152226e-01 6.45577133e-01
6.51521504e-01 4.04989906e-02 5.58963716e-01 -1.79268464e-01
4.64562476e-01 -3.42608571e-01 5.93389906e-02 -2.54492521e-01
2.89247990e-01 -1.43558800e+00 7.58153677e-01 6.12649083e-01
4.19433028e-01 3.72607976e-01 7.04415619e-01 -3.63804638e-01
-1.66237485e+00 -9.33057070e-01 8.31529021e-01 -1.13084424e+00
5.14456451e-01 -6.72121763e-01 -6.63016021e-01 6.50714815e-01
-6.03077710e-02 2.32607529e-01 2.58225769e-01 -2.11888049e-02
-7.24114716e-01 -3.13725206e-04 -9.87497628e-01 2.53303617e-01
1.37347364e+00 -6.99842036e-01 -9.06341732e-01 2.08136484e-01
6.93317354e-01 -4.53509182e-01 -4.96102393e-01 6.58170640e-01
3.81280720e-01 -1.18343067e+00 1.43705821e+00 -5.64631522e-01
9.90521535e-02 -5.08999884e-01 -5.66304386e-01 -1.13099790e+00
-1.60591438e-01 2.21200101e-02 2.66659349e-01 6.01348162e-01
7.74254724e-02 -5.97694337e-01 7.08507657e-01 -1.63781061e-03
-4.19424444e-01 -3.84289652e-01 -1.22942591e+00 -8.05650949e-01
5.31206764e-02 -9.05562282e-01 5.95424712e-01 5.18678725e-01
-6.17472470e-01 4.95495051e-01 3.49242806e-01 5.48836112e-01
9.37113523e-01 6.68978691e-01 8.97509336e-01 -1.97073066e+00
2.73127496e-01 -6.03429317e-01 -1.17828250e+00 -1.05768645e+00
3.65080565e-01 -1.02497208e+00 -5.80992177e-02 -1.31126130e+00
-3.57777276e-03 -5.03494322e-01 5.51743209e-02 -5.13439067e-02
2.28276730e-01 8.51321146e-02 3.37976754e-01 1.50730640e-01
-3.42465669e-01 5.07022500e-01 1.00592053e+00 -1.08219169e-01
1.73146158e-01 1.68385327e-01 -1.99797019e-01 1.00966978e+00
4.80560631e-01 -3.50242496e-01 -4.94868547e-01 -6.66753471e-01
1.96649492e-01 -7.11074546e-02 7.91155815e-01 -8.58376443e-01
3.11843395e-01 -2.16274798e-01 2.40721822e-01 -1.32189763e+00
8.73444617e-01 -1.29370642e+00 -1.10185951e-01 2.96676099e-01
1.53766908e-02 2.65852869e-01 1.90755632e-02 8.31639946e-01
-2.21827954e-01 -2.70364344e-01 6.47778332e-01 -5.83940566e-01
-1.08044767e+00 4.76221770e-01 -2.01654121e-01 -1.49156928e-01
1.04376948e+00 -8.00078332e-01 -3.06825578e-01 -1.54799715e-01
-5.84436297e-01 3.03769648e-01 8.06364000e-01 7.82830954e-01
9.80669081e-01 -1.28165424e+00 -5.29110432e-01 6.80456102e-01
6.23802662e-01 4.05251503e-01 -2.03474164e-02 5.28802752e-01
-3.08298111e-01 5.87345481e-01 -2.51802713e-01 -1.32162642e+00
-1.31039166e+00 7.45775938e-01 2.66534805e-01 3.41834188e-01
-7.94183195e-01 4.74013835e-01 5.82328439e-01 -5.76071084e-01
5.65711642e-03 -7.30200827e-01 -7.20523894e-02 -1.68172717e-02
2.17998341e-01 7.95270428e-02 3.01632613e-01 -1.03350818e+00
-5.27362049e-01 1.35139346e+00 1.70134202e-01 -7.93875307e-02
1.22324407e+00 -3.45016837e-01 2.34411433e-02 9.27816808e-01
1.26134050e+00 -9.44747850e-02 -1.08307850e+00 -5.73092818e-01
2.59620309e-01 -7.58028090e-01 1.52411610e-01 -2.07312509e-01
-6.17400467e-01 1.45857322e+00 7.30716109e-01 2.23791823e-01
6.29976451e-01 4.81862873e-01 3.23050290e-01 4.12432313e-01
7.91627765e-01 -6.12246096e-01 -5.00931032e-02 7.22788811e-01
5.88719308e-01 -1.40755975e+00 -2.72411071e-02 -6.56201541e-01
-2.67157733e-01 1.25260174e+00 3.14056158e-01 -3.72746527e-01
9.42427456e-01 2.06141114e-01 5.33755720e-02 -6.77582860e-01
-7.30470061e-01 -6.92556083e-01 4.36417937e-01 9.01655197e-01
-1.61185697e-01 6.19193241e-02 3.75986695e-01 -7.31360093e-02
-1.57965288e-01 -3.52263510e-01 5.30382395e-01 1.15415883e+00
-9.25575197e-01 -1.04675281e+00 -4.59492296e-01 1.17493324e-01
4.62275594e-02 1.82397410e-01 -3.86251062e-01 9.02720571e-01
3.22791070e-01 7.05283880e-01 4.95416343e-01 -3.95627737e-01
6.10255659e-01 1.01014607e-01 6.75592840e-01 -8.34459186e-01
4.48607318e-02 -1.41469479e-01 -2.57510453e-01 -7.68841624e-01
-5.58730602e-01 -9.85292494e-01 -1.12101495e+00 5.68448231e-02
-1.93430051e-01 -2.68189102e-01 1.31434202e+00 1.12016547e+00
3.07758838e-01 7.90277943e-02 6.59431279e-01 -1.26205826e+00
-5.93541563e-01 -4.20510501e-01 -7.54833341e-01 4.36078072e-01
8.19370627e-01 -1.03147399e+00 -3.81471038e-01 -1.20636046e-01] | [7.809700965881348, -2.6803269386291504] |
70f6dca3-74e2-4d5a-a9fc-abb88f4da870 | a-new-k-means-grey-wolf-algorithm-for | 2103.05760 | null | https://arxiv.org/abs/2103.05760v1 | https://arxiv.org/pdf/2103.05760v1.pdf | A New K means Grey Wolf Algorithm for Engineering Problems | Purpose: The development of metaheuristic algorithms has increased by researchers to use them extensively in the field of business, science, and engineering. One of the common metaheuristic optimization algorithms is called Grey Wolf Optimization (GWO). The algorithm works based on imitation of the wolves' searching and the process of attacking grey wolves. The main purpose of this paper to overcome the GWO problem which is trapping into local optima. Design or Methodology or Approach: In this paper, the K-means clustering algorithm is used to enhance the performance of the original Grey Wolf Optimization by dividing the population into different parts. The proposed algorithm is called K-means clustering Grey Wolf Optimization (KMGWO). Findings: Results illustrate the efficiency of KMGWO is superior to GWO. To evaluate the performance of the KMGWO, KMGWO applied to solve 10 CEC2019 benchmark test functions. Results prove that KMGWO is better compared to GWO. KMGWO is also compared to Cat Swarm Optimization (CSO), Whale Optimization Algorithm-Bat Algorithm (WOA-BAT), and WOA, so, KMGWO achieves the first rank in terms of performance. Statistical results proved that KMGWO achieved a higher significant value compared to the compared algorithms. Also, the KMGWO is used to solve a pressure vessel design problem and it has outperformed results. Originality/value: Results prove that KMGWO is superior to GWO. KMGWO is also compared to cat swarm optimization (CSO), whale optimization algorithm-bat algorithm (WOA-BAT), WOA, and GWO so KMGWO achieved the first rank in terms of performance. Also, the KMGWO is used to solve a classical engineering problem and it is superior | ['Nebojsa Bacanin', 'Abeer Alsadoon', 'Tarik A. Rashid', 'Zrar Kh. Abdul', 'Hardi M. Mohammed'] | 2021-02-27 | null | null | null | null | ['metaheuristic-optimization'] | ['methodology'] | [-3.48156840e-01 -3.78290981e-01 2.24704280e-01 5.75457871e-01
2.41025418e-01 -9.26437825e-02 2.05655590e-01 1.60286412e-01
-6.36499345e-01 7.92250812e-01 -2.33080033e-02 -2.47590557e-01
-9.25951660e-01 -9.62156892e-01 -1.71756208e-01 -1.20606089e+00
-2.25920781e-01 3.82428855e-01 1.94770351e-01 -9.23420370e-01
8.61322761e-01 5.06199181e-01 -1.64878118e+00 -4.50686008e-01
9.47779238e-01 7.15628266e-01 6.21443868e-01 5.82360983e-01
-1.85362011e-01 1.20793328e-01 -8.88081193e-01 -1.30638659e-01
5.09382188e-01 -7.68043697e-01 -4.75110859e-01 -1.85282812e-01
-9.30045366e-01 5.02680659e-01 3.09582919e-01 9.98604000e-01
9.83380437e-01 5.59962392e-01 4.52632725e-01 -1.55812311e+00
-6.47365570e-01 4.19686168e-01 -8.81260633e-01 5.51955819e-01
3.17113958e-02 1.64062828e-01 6.55515909e-01 -5.40384948e-01
4.31425363e-01 1.22083962e+00 6.33063674e-01 3.57309550e-01
-5.05190909e-01 -6.14954650e-01 -4.05657619e-01 4.44232166e-01
-1.37564313e+00 4.85598087e-01 5.29820442e-01 9.09523144e-02
9.62851584e-01 7.31003225e-01 1.39108181e+00 7.83855766e-02
5.29802740e-01 6.16483152e-01 1.35712004e+00 -7.84629047e-01
7.11347580e-01 -3.63461047e-01 1.29049599e-01 5.29775679e-01
8.31735730e-01 4.45714444e-01 -4.18633759e-01 -1.20283492e-01
1.48639619e-01 -1.18139707e-01 -2.19436377e-01 6.75740540e-02
-8.97444844e-01 1.19193816e+00 4.36956853e-01 9.51417685e-01
-4.92750406e-01 2.34935656e-02 7.24326894e-02 3.36292237e-01
5.65354014e-03 1.04836321e+00 9.60612968e-02 -5.29964924e-01
-7.52196610e-01 2.64139652e-01 8.38426709e-01 3.99482697e-01
1.22205675e-01 4.49041903e-01 6.92622364e-02 6.04106724e-01
3.81221175e-01 6.99137628e-01 1.11699414e+00 -5.91778517e-01
1.01166926e-01 8.95270348e-01 1.74600899e-01 -1.21844518e+00
-6.61490679e-01 -3.44255745e-01 -6.39971614e-01 7.00712919e-01
-4.79907580e-02 -2.85536170e-01 -7.83531129e-01 1.07077110e+00
2.79294819e-01 -6.14100918e-02 4.74998772e-01 1.00072277e+00
9.50427055e-01 1.14296412e+00 -8.00525546e-02 -3.22699636e-01
1.11535096e+00 -1.19388056e+00 -9.54491973e-01 5.59867686e-03
5.43786108e-01 -8.96181643e-01 7.75491834e-01 3.86675298e-01
-9.89616334e-01 -2.76898772e-01 -1.32261479e+00 1.09831429e+00
-8.41612935e-01 -2.68420577e-01 6.11795604e-01 1.29819548e+00
-8.92696202e-01 5.13021648e-01 -5.47244251e-01 -7.30611026e-01
1.64873272e-01 6.09171033e-01 7.86415339e-02 1.53992653e-01
-8.88462245e-01 1.24287701e+00 8.28052640e-01 2.24881947e-01
-1.88669637e-01 -3.61306518e-01 -5.52060246e-01 1.02926850e-01
3.83286268e-01 -1.75978824e-01 3.87172759e-01 -6.39897287e-01
-1.60374129e+00 3.49282324e-01 -1.00849025e-01 -3.97493720e-01
3.96606028e-01 4.11423653e-01 -4.68579113e-01 1.17865361e-01
9.65254977e-02 6.52990937e-02 2.39016712e-01 -1.45460653e+00
-9.02777016e-01 -2.39593416e-01 -3.60807836e-01 1.56477720e-01
-3.04203629e-01 2.46460289e-01 1.51153117e-01 -8.82956266e-01
8.27077776e-02 -8.86763513e-01 -2.73923665e-01 -8.96295428e-01
4.08888869e-02 -3.94931942e-01 1.02950561e+00 -3.60194653e-01
1.77958691e+00 -1.93905485e+00 1.62001193e-01 7.89328396e-01
1.25105372e-02 7.03547895e-01 -3.86225097e-02 8.17974150e-01
1.64690405e-01 4.49368626e-01 -3.17331880e-01 2.28074148e-01
1.66176721e-01 5.16223431e-01 2.80056924e-01 3.80249768e-01
-2.75001824e-01 8.95155966e-01 -9.04088259e-01 -4.80534613e-01
3.68674010e-01 -1.83889344e-01 -2.11895913e-01 -1.07540965e-01
4.81164783e-01 -5.72062954e-02 -6.59815788e-01 8.83640170e-01
7.50396550e-01 1.96363986e-01 -1.85490683e-01 -5.86219914e-02
-5.36968529e-01 -9.77229357e-01 -1.34051251e+00 8.38264942e-01
-3.10972214e-01 4.03718263e-01 -1.31574586e-01 -1.18189597e+00
1.25212419e+00 1.46920457e-01 6.96914971e-01 -8.22660863e-01
7.35468805e-01 6.63779199e-01 3.84349018e-01 -7.86892653e-01
4.10186529e-01 -1.67363882e-01 2.85156041e-01 5.43408155e-01
-2.25933701e-01 -5.11838973e-01 6.29179478e-01 -2.41795048e-01
7.23605573e-01 -8.08481127e-02 3.58919144e-01 -7.06135511e-01
5.99648297e-01 4.02086526e-01 6.08650506e-01 6.80189490e-01
-1.14739873e-01 2.40478411e-01 -6.13866448e-02 -3.83275926e-01
-7.37819791e-01 -6.29782379e-01 3.20226736e-02 7.36809254e-01
9.33312237e-01 -1.08751893e-01 -6.45338237e-01 3.69567461e-02
1.09233253e-01 1.02243507e+00 -7.39867747e-01 -4.20624614e-01
-6.26308024e-01 -1.58352339e+00 4.13711369e-01 4.65435092e-04
1.22558475e+00 -1.27246356e+00 -1.50704610e+00 3.52046490e-01
2.79791743e-01 -3.94248754e-01 -1.23307735e-01 -9.57064927e-02
-6.51964247e-01 -9.63322043e-01 -9.38382328e-01 -6.29607260e-01
4.22893047e-01 3.06965262e-01 7.19364047e-01 6.98253393e-01
-6.45530641e-01 2.47605368e-01 -1.00763404e+00 -1.05111194e+00
-4.07797128e-01 -2.57886946e-01 -1.72089487e-01 -9.13805813e-02
5.06367445e-01 -1.59442186e-01 -3.86552334e-01 7.13657081e-01
-8.72127831e-01 -4.49227035e-01 5.71445405e-01 9.36365426e-01
2.25134566e-01 7.94598341e-01 6.02416217e-01 8.18555802e-03
1.13462389e+00 -2.55881816e-01 -6.36077285e-01 4.93681282e-01
-1.00501239e+00 -1.43099487e-01 2.83640027e-01 -3.40340227e-01
-5.84048152e-01 -8.03771138e-01 1.56372026e-01 -2.76261806e-01
5.43011963e-01 4.85843301e-01 1.18125439e-01 -4.59951758e-01
1.64072141e-01 9.58157107e-02 2.11006492e-01 -1.52319521e-01
-3.71851772e-01 8.17790210e-01 -4.53203581e-02 -4.39412802e-01
6.26580417e-01 3.09007317e-01 4.06541348e-01 -9.22658682e-01
-5.75234890e-02 -4.93878961e-01 2.34287307e-01 -5.27200818e-01
1.21160316e+00 2.22418636e-01 -1.30834424e+00 4.50104177e-01
-5.03368795e-01 -8.78989324e-02 -3.44838619e-01 7.24266410e-01
-2.56509483e-01 1.77988127e-01 1.26692176e-01 -1.31344175e+00
-5.90741754e-01 -1.31754780e+00 1.80104569e-01 8.66788507e-01
5.21800853e-02 -1.08947432e+00 -2.29897186e-01 3.00064117e-01
6.68760478e-01 8.58713269e-01 6.29835427e-01 -5.90321004e-01
8.01618025e-02 -2.54712641e-01 2.93231785e-01 -1.46561518e-01
3.26986685e-02 3.12022269e-01 1.54091507e-01 -3.77089322e-01
3.75943720e-01 2.98431277e-01 2.12239310e-01 6.02292240e-01
3.59535843e-01 -1.52901458e-02 -3.72476369e-01 6.23721778e-01
1.69495130e+00 1.21585286e+00 6.40095174e-01 1.60980892e+00
-1.45645142e-01 5.54484904e-01 1.18852639e+00 4.24971670e-01
-2.51739323e-02 4.52396095e-01 7.28651524e-01 -1.52790755e-01
4.14933600e-02 2.85024405e-01 8.33338201e-02 8.18038702e-01
-7.35269845e-01 -5.49883544e-01 -1.02521217e+00 3.93041044e-01
-1.84426379e+00 -1.12175512e+00 -3.85147929e-01 1.90670657e+00
-6.45632967e-02 -4.66802455e-02 2.64521837e-01 5.30619979e-01
9.87133563e-01 -6.60511628e-02 -1.68854609e-01 -1.17102957e+00
-4.91355181e-01 5.32717109e-01 9.42863941e-01 2.19858348e-01
-5.46637177e-01 6.09672606e-01 4.93160772e+00 1.20911157e+00
-1.09851241e+00 1.69235140e-01 1.63874805e-01 -1.26159102e-01
-1.18294079e-03 -7.72611573e-02 -4.43631858e-01 9.39790845e-01
5.65841079e-01 -8.10331404e-01 6.33500218e-01 3.15027505e-01
4.22163159e-01 -7.00330377e-01 1.61200836e-01 1.05139792e+00
1.28575727e-01 -1.18587506e+00 -2.21030965e-01 2.32251510e-01
1.08838475e+00 -2.64856756e-01 -1.18458204e-01 8.94770473e-02
5.56998968e-01 -1.29154658e+00 5.95279217e-01 3.70495766e-01
-2.26260945e-01 -1.07903981e+00 1.46518934e+00 3.24642897e-01
-1.04859066e+00 -6.88041151e-01 -5.60404718e-01 1.30479932e-01
3.91236424e-01 9.20705032e-03 -2.31186762e-01 1.15051758e+00
1.36452794e+00 -6.28855154e-02 -4.97073561e-01 1.63307917e+00
1.32243127e-01 6.21594369e-01 -6.03223622e-01 -1.00866437e+00
1.00056374e+00 -8.61829638e-01 9.74940836e-01 6.08744204e-01
9.15801227e-01 3.27199787e-01 -1.03769198e-01 6.31152689e-01
7.01726019e-01 6.05243266e-01 -1.61942169e-01 -2.97736108e-01
5.43644667e-01 9.10728931e-01 -1.17697787e+00 -1.97723016e-01
5.71308248e-02 4.32520717e-01 -7.65073538e-01 6.21854737e-02
-8.88675451e-01 -9.18153882e-01 6.83669839e-03 -1.79835126e-01
2.85089016e-01 3.95919979e-02 -4.61522013e-01 -2.45747358e-01
-4.89428401e-01 -7.12052584e-01 6.17108226e-01 -9.06335890e-01
-8.62644315e-01 6.12780392e-01 2.90189117e-01 -1.31198108e+00
3.39575946e-01 -6.09338164e-01 -1.00203228e+00 9.65289354e-01
-1.11322689e+00 -8.34096193e-01 -6.06361091e-01 3.79654348e-01
4.83342528e-01 -8.75619292e-01 3.51864785e-01 -1.57445326e-01
-5.68301618e-01 1.55276433e-01 6.42809927e-01 -3.72527242e-01
-5.43577671e-02 -1.03373599e+00 -2.68631041e-01 6.70010746e-01
-4.79243457e-01 3.99983525e-01 1.02982831e+00 -7.78934658e-01
-1.36544716e+00 -1.31406814e-01 5.01120985e-01 2.50148356e-01
6.67599916e-01 5.47205210e-01 -4.01879519e-01 -2.67710775e-01
5.08638918e-01 -5.56261837e-01 7.10675776e-01 -5.91751337e-01
9.08119380e-01 4.05892022e-02 -1.61190581e+00 6.25556886e-01
4.77824062e-01 6.54020965e-01 -7.84742594e-01 6.86103702e-02
1.01859525e-01 -3.06902647e-01 -9.21915948e-01 4.09005672e-01
5.25620520e-01 -1.09045005e+00 9.05742228e-01 -3.35585743e-01
1.08427882e-01 -6.58471704e-01 -6.08589202e-02 -1.32442951e+00
-2.44823262e-01 -8.72908354e-01 4.27716106e-01 9.33510542e-01
1.62423536e-01 -1.26368427e+00 6.25556171e-01 4.69403028e-01
-4.93460372e-02 -9.63587523e-01 -1.22692275e+00 -1.20336413e+00
9.11459699e-02 3.90535206e-01 1.13307953e+00 8.44615340e-01
-1.38269648e-01 -2.62150228e-01 -2.31699452e-01 -8.64905715e-02
5.99893630e-01 1.77998379e-01 5.84001303e-01 -9.96603787e-01
-2.05833286e-01 -9.39601183e-01 -5.81166029e-01 1.85510352e-01
-5.34564137e-01 -6.05818272e-01 -3.83225977e-01 -1.97653520e+00
-4.79112476e-01 -2.73035914e-01 -2.43666857e-01 7.05065206e-02
-2.81565756e-01 4.24798280e-01 7.37132728e-01 2.75799453e-01
7.62376264e-02 4.73718524e-01 1.47466004e+00 -2.11028140e-02
-5.57633877e-01 8.09430629e-02 -5.13624012e-01 4.18437183e-01
9.80496526e-01 -4.88586515e-01 -4.69517596e-02 -1.03174470e-01
2.15124220e-01 -1.32386938e-01 6.96246698e-02 -1.11321533e+00
3.29679847e-01 -4.30722594e-01 3.12980786e-02 -4.08503801e-01
6.27946928e-02 -1.04359293e+00 4.72844690e-01 1.30131209e+00
4.07949567e-01 6.40330136e-01 1.99282393e-01 2.02282846e-01
-3.45202446e-01 -1.10729432e+00 7.10198879e-01 -1.45596400e-01
-9.23764944e-01 -3.00920159e-01 -5.27768672e-01 -8.30770954e-02
1.77397048e+00 -1.26144493e+00 -2.98065424e-01 -2.96675831e-01
-7.62004972e-01 5.89531839e-01 3.39620858e-01 3.11644882e-01
3.94480854e-01 -1.10982847e+00 -5.44551730e-01 -2.51723021e-01
-2.83626795e-01 -5.58194637e-01 1.34154126e-01 1.19957983e+00
-1.40380323e+00 6.43706918e-02 -4.95177627e-01 -9.19783935e-02
-1.50105762e+00 4.89804417e-01 3.70560527e-01 -1.76000372e-01
-3.88459623e-01 5.91136217e-01 -6.41406596e-01 -1.02306679e-01
-3.10627550e-01 2.62350082e-01 -6.76675856e-01 1.91890568e-01
9.49203223e-02 1.18515766e+00 -1.70001939e-01 -8.12925637e-01
-3.35166097e-01 9.79850054e-01 7.87177920e-01 -4.36436683e-02
1.74475729e+00 1.99971572e-02 -4.48545456e-01 3.98783498e-02
9.21519458e-01 2.51201212e-01 -2.71279097e-01 7.24887013e-01
8.81641805e-02 -7.69503176e-01 7.36704767e-02 -9.84385192e-01
-1.01030779e+00 4.77811813e-01 9.95730221e-01 4.89991605e-01
1.33048403e+00 -5.51144481e-01 8.07961464e-01 4.37488914e-01
5.20060480e-01 -1.59053564e+00 -1.80629253e-01 2.19131052e-01
7.79423594e-01 -1.02403367e+00 2.34430388e-01 1.67571492e-02
-8.26496422e-01 1.35920358e+00 2.92031258e-01 -3.34458649e-01
7.63840616e-01 1.88760623e-01 1.85938533e-02 -3.66787672e-01
-4.43562716e-02 -4.84959155e-01 2.05148086e-02 4.11767602e-01
-2.84759074e-01 -1.37481838e-01 -1.28189480e+00 5.10282993e-01
-6.19781971e-01 -3.29294242e-02 3.77731562e-01 1.41254163e+00
-7.77131438e-01 -1.06517327e+00 -1.02825093e+00 3.16136092e-01
-4.12455112e-01 1.18429035e-01 -1.71254665e-01 1.21713531e+00
2.62567788e-01 1.38061404e+00 -2.43244186e-01 -2.06656307e-01
3.65304381e-01 -5.00569284e-01 1.58670574e-01 2.46279359e-01
-1.22742069e+00 -1.06720753e-01 -1.85102925e-01 -1.49561793e-01
-8.79923940e-01 -3.12674880e-01 -1.44294679e+00 -7.08754063e-01
-6.81477785e-01 1.18937683e+00 9.51691151e-01 9.15770292e-01
-7.73020387e-02 5.86966693e-01 6.07538104e-01 -6.31181479e-01
-2.49055624e-01 -6.93101048e-01 -7.92618275e-01 2.90953577e-01
-4.80343938e-01 -9.80915606e-01 -5.71594715e-01 -6.72839582e-01] | [5.589442729949951, 3.4477434158325195] |
33fa3623-79f2-4842-8aac-ba00f4339459 | deepasl-enabling-ubiquitous-and-non-intrusive | 1802.07584 | null | http://arxiv.org/abs/1802.07584v3 | http://arxiv.org/pdf/1802.07584v3.pdf | DeepASL: Enabling Ubiquitous and Non-Intrusive Word and Sentence-Level Sign Language Translation | There is an undeniable communication barrier between deaf people and people
with normal hearing ability. Although innovations in sign language translation
technology aim to tear down this communication barrier, the majority of
existing sign language translation systems are either intrusive or constrained
by resolution or ambient lighting conditions. Moreover, these existing systems
can only perform single-sign ASL translation rather than sentence-level
translation, making them much less useful in daily-life communication
scenarios. In this work, we fill this critical gap by presenting DeepASL, a
transformative deep learning-based sign language translation technology that
enables ubiquitous and non-intrusive American Sign Language (ASL) translation
at both word and sentence levels. DeepASL uses infrared light as its sensing
mechanism to non-intrusively capture the ASL signs. It incorporates a novel
hierarchical bidirectional deep recurrent neural network (HB-RNN) and a
probabilistic framework based on Connectionist Temporal Classification (CTC)
for word-level and sentence-level ASL translation respectively. To evaluate its
performance, we have collected 7,306 samples from 11 participants, covering 56
commonly used ASL words and 100 ASL sentences. DeepASL achieves an average
94.5% word-level translation accuracy and an average 8.2% word error rate on
translating unseen ASL sentences. Given its promising performance, we believe
DeepASL represents a significant step towards breaking the communication
barrier between deaf people and hearing majority, and thus has the significant
potential to fundamentally change deaf people's lives. | ['Mi Zhang', 'Biyi Fang', 'Jillian Co'] | 2018-02-21 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 2.91809469e-01 -1.84196949e-01 -2.70366490e-01 -4.14644033e-02
-1.35939670e+00 -4.83176976e-01 2.77368426e-01 -6.11251175e-01
-5.71774304e-01 6.30928636e-01 7.76758671e-01 -4.30092901e-01
3.13691229e-01 -5.40714741e-01 -5.24624228e-01 -6.62876308e-01
5.90165794e-01 -3.12739760e-02 1.62585810e-01 -2.85539687e-01
-5.30886650e-02 4.20381188e-01 -1.54068995e+00 5.09494007e-01
8.93752873e-01 7.28580356e-01 4.06850781e-03 7.60486484e-01
-2.14039147e-01 4.96081501e-01 -5.48401654e-01 -3.52290511e-01
9.75979343e-02 -7.04116046e-01 -4.21666890e-01 -6.09484494e-01
8.06072056e-01 -7.53114223e-01 -4.80216533e-01 7.41973579e-01
1.36338723e+00 -4.65371042e-01 3.90404254e-01 -8.43739271e-01
-1.03431237e+00 3.41015756e-01 -3.15737069e-01 -3.39305371e-01
7.68687189e-01 5.95245361e-01 8.32551122e-01 -9.57780957e-01
4.75728542e-01 1.23784077e+00 8.40059698e-01 1.23538649e+00
-1.08004117e+00 -9.26682293e-01 -1.61722064e-01 2.10678861e-01
-1.17396879e+00 -8.52639258e-01 6.95421994e-01 -3.90649557e-01
1.26164842e+00 1.81590632e-01 1.28723764e+00 1.76269674e+00
3.45962435e-01 9.58263397e-01 1.51337695e+00 -5.69292545e-01
2.10067276e-02 -3.12622666e-01 -1.14714839e-01 4.88867700e-01
2.50340104e-01 2.99993902e-01 -1.11050749e+00 -3.77330021e-03
5.28639793e-01 -3.79516095e-01 -4.30165052e-01 4.19869781e-01
-1.11342156e+00 3.05432469e-01 3.06978822e-01 3.61674905e-01
-1.81777999e-01 3.69931728e-01 3.76680136e-01 4.88900840e-01
2.13659883e-01 -1.28108740e-01 -1.36514947e-01 -5.81697762e-01
-4.78025287e-01 -1.47792205e-01 5.44291675e-01 6.72739208e-01
-8.95990804e-02 6.12761341e-02 -3.73794466e-01 1.23622549e+00
6.67808294e-01 1.25940323e+00 7.07407415e-01 -6.00803196e-01
3.89356077e-01 3.78445774e-01 1.04431473e-02 -4.02949899e-01
-2.48294875e-01 -9.12601352e-02 -5.85566163e-01 5.99571466e-01
2.91390985e-01 3.76533670e-03 -1.45090806e+00 1.92320275e+00
-1.11379676e-01 -2.77457893e-01 -4.35672514e-02 1.01768029e+00
9.43876147e-01 4.27165985e-01 1.76052853e-01 -1.45059809e-01
1.19082355e+00 -5.01477480e-01 -8.60432446e-01 -1.14721447e-01
6.42909288e-01 -1.12493360e+00 1.55744922e+00 2.53837973e-01
-1.06291556e+00 -2.13012341e-02 -9.21542525e-01 -4.90376621e-01
-1.87313810e-01 8.10197517e-02 3.37365448e-01 8.94254982e-01
-1.25810528e+00 -2.07362130e-01 -8.21566761e-01 -8.18795383e-01
4.97819036e-01 2.69089907e-01 -2.94223577e-01 -1.56675279e-01
-1.13592446e+00 1.08630610e+00 -4.66711551e-01 4.86011595e-01
-1.30591646e-01 -3.57751787e-01 -7.47444153e-01 -5.17312646e-01
-4.36531156e-01 -8.34776342e-01 1.35986447e+00 -2.77344048e-01
-2.24541306e+00 1.23693442e+00 -4.61859047e-01 -1.99966162e-01
1.05356181e+00 -6.42902792e-01 -4.86604095e-01 -1.13243490e-01
-2.32156068e-02 5.66761792e-01 7.38754749e-01 -8.51668298e-01
-4.77049768e-01 -3.20066214e-01 -6.64014935e-01 8.09722319e-02
-2.07575276e-01 3.16830337e-01 -7.19360635e-02 -6.49942756e-01
4.94146794e-01 -9.30059910e-01 5.17150402e-01 7.32690930e-01
-4.05445248e-01 -3.50738615e-01 9.46754277e-01 -9.30777311e-01
8.14504206e-01 -1.97482777e+00 -1.06056049e-01 -1.67091519e-01
1.56338647e-01 6.60163164e-01 -2.93518484e-01 5.31618893e-01
5.80280900e-01 -7.77972536e-03 -3.76005262e-01 -4.21809703e-01
3.32451873e-02 1.93704441e-01 -3.98694873e-01 5.91718853e-01
-1.68158144e-01 1.23971462e+00 -8.39419127e-01 -2.99832970e-01
4.55892175e-01 9.96552229e-01 -2.66190886e-01 -9.61529687e-02
3.83560285e-02 4.79718298e-01 1.92192551e-02 1.20212519e+00
4.79379952e-01 5.73661804e-01 -1.88864812e-01 -1.62116468e-01
-2.67176360e-01 5.25000989e-01 -4.46445078e-01 1.81232929e+00
-8.25064838e-01 1.19105184e+00 4.38795201e-02 -1.62486613e-01
8.64853203e-01 3.95141959e-01 3.16304147e-01 -1.20694876e+00
2.40807578e-01 8.61697555e-01 -1.57858014e-01 -8.56391311e-01
-1.23082198e-01 -6.63049102e-01 -1.36685297e-02 3.90639395e-01
-3.20806563e-01 -3.67262900e-01 -4.31142658e-01 -2.92166203e-01
1.21519494e+00 1.13670848e-01 -1.27646729e-01 2.51418501e-01
1.24701679e-01 -5.05551279e-01 3.34664553e-01 6.99680626e-01
-6.20274305e-01 7.25875139e-01 -9.52192545e-02 -1.74976066e-01
-8.06999326e-01 -1.57324708e+00 -1.59990173e-02 7.76326537e-01
1.37997586e-02 1.00956976e-01 -3.81831169e-01 -3.00910976e-02
-7.28313252e-02 5.68477511e-01 -1.59670040e-01 -3.00494552e-01
-5.31285048e-01 -2.74662435e-01 1.07073176e+00 4.49278980e-01
8.74690413e-01 -1.27528298e+00 -5.05596101e-01 2.81020790e-01
-6.15275562e-01 -1.22534096e+00 -4.81209159e-01 -4.43239808e-01
-3.47138792e-01 -6.82663083e-01 -1.07488966e+00 -1.15242755e+00
2.77694464e-01 5.95538244e-02 4.24773157e-01 -3.59809637e-01
-5.58336973e-01 3.31628710e-01 -3.34112167e-01 -5.53156197e-01
-5.60164213e-01 -1.16701543e-01 4.76320028e-01 -3.07053179e-02
7.68891215e-01 -8.75328660e-01 -6.83876693e-01 1.30817756e-01
-3.62631440e-01 8.07501525e-02 8.28890622e-01 8.57084155e-01
9.36049446e-02 -7.75798976e-01 3.76205444e-01 2.67223626e-01
9.45850253e-01 3.43563735e-01 -3.38393867e-01 2.03554824e-01
-3.98518175e-01 -5.26613183e-02 2.29356945e-01 -5.81735134e-01
-8.46941471e-01 -2.66229331e-01 -4.88080770e-01 1.53334185e-01
7.10117118e-03 1.41366869e-01 -1.09948397e-01 -1.18265025e-01
7.65197337e-01 4.73060697e-01 1.49616346e-01 -4.86759394e-01
2.40285292e-01 1.46925271e+00 8.86246443e-01 -2.70655334e-01
7.13382363e-01 6.11267507e-01 -2.34777331e-01 -1.03709972e+00
-4.90967989e-01 -2.29439542e-01 -3.40153426e-01 -5.63633621e-01
7.66234756e-01 -6.56730473e-01 -1.17647994e+00 1.27121961e+00
-1.26005912e+00 -5.91618061e-01 -2.56301761e-01 8.65683734e-01
-5.77937961e-01 4.94742692e-02 -8.03986728e-01 -1.05992198e+00
-6.94170296e-01 -9.39271212e-01 1.32097769e+00 -2.56569386e-02
-5.63773930e-01 -2.02832654e-01 3.24612707e-01 6.81870997e-01
7.06972778e-01 7.23848417e-02 7.69596994e-01 3.06953430e-01
-5.04227459e-01 -4.95497465e-01 -4.19841439e-01 4.15270805e-01
4.50681597e-01 -1.83897048e-01 -1.29765987e+00 -3.93038005e-01
-3.63229036e-01 -5.29704750e-01 1.00622594e+00 4.90352690e-01
2.38990456e-01 -2.06101030e-01 -1.06540129e-01 5.20045161e-01
1.05987751e+00 2.78917104e-01 8.98615181e-01 1.10137388e-02
5.94823658e-01 3.44469726e-01 1.22033060e-02 1.10310003e-01
2.82416403e-01 8.16641867e-01 -1.30362436e-01 -8.05773959e-02
-1.03117061e+00 -7.64893889e-01 8.35655510e-01 1.14828229e+00
-2.21684501e-01 -1.47855282e-01 -9.57645595e-01 5.91437936e-01
-1.47839582e+00 -9.63011205e-01 1.98882185e-02 2.10144353e+00
8.70061100e-01 -4.18754155e-03 3.78295071e-02 4.40874010e-01
5.26355445e-01 4.80896942e-02 -8.00846517e-01 -6.11316085e-01
-4.81928170e-01 3.91220450e-01 5.26829600e-01 8.68370235e-01
-5.65538228e-01 1.11846018e+00 6.30633879e+00 4.30456877e-01
-1.73618555e+00 1.88299790e-01 -1.78715751e-01 -5.31735241e-01
-2.99014121e-01 -5.53626120e-01 -6.90799415e-01 3.97587925e-01
7.04912782e-01 2.49490589e-01 3.90557587e-01 3.40734392e-01
7.49962628e-01 6.16185926e-02 -8.41051400e-01 1.38168049e+00
2.14797527e-01 -9.36228514e-01 7.48652443e-02 1.71270117e-01
3.50899935e-01 5.26042521e-01 2.03834578e-01 1.47774458e-01
6.18008003e-02 -1.07947516e+00 7.37426341e-01 5.66467464e-01
1.51748621e+00 -2.66507626e-01 5.53534269e-01 8.01982544e-03
-1.02295601e+00 -1.74971655e-01 2.64574111e-01 -2.82822996e-01
3.81481975e-01 1.96601108e-01 -6.10504210e-01 -3.43622237e-01
7.61682451e-01 6.21458948e-01 1.03688516e-01 1.00030673e+00
-7.27940917e-01 8.51825297e-01 -7.79304206e-01 -4.45427448e-01
-1.70804579e-02 6.51197210e-02 8.86644542e-01 1.23420906e+00
5.52467525e-01 -4.39750962e-02 -4.35857862e-01 6.42038941e-01
-8.07830319e-02 3.28315422e-02 -8.73777211e-01 6.90979064e-02
5.04177451e-01 3.07097256e-01 -1.79933786e-01 9.48574990e-02
-4.09459829e-01 1.48549676e+00 -2.05537319e-01 6.44397914e-01
-4.65091318e-01 -5.35679996e-01 1.22931767e+00 9.44225118e-02
-2.64864743e-01 -5.37787616e-01 -7.20823050e-01 -1.14757752e+00
8.10591161e-01 -5.96304774e-01 -1.87329322e-01 -9.15297449e-01
-1.22746670e+00 3.38759869e-01 -7.70274818e-01 -1.52761495e+00
1.62861124e-02 -7.82275379e-01 -2.63704687e-01 8.77471924e-01
-1.51033652e+00 -1.65881288e+00 -3.77860576e-01 4.41364080e-01
5.06981015e-01 1.23499110e-02 9.75100875e-01 4.24943686e-01
-3.85445446e-01 1.10735047e+00 1.08229011e-01 2.21119046e-01
8.19092929e-01 -5.65257728e-01 4.64096665e-01 7.32995629e-01
1.24483807e-02 5.66971242e-01 6.59016132e-01 -5.12589931e-01
-1.35604072e+00 -7.53912151e-01 1.46435261e+00 -4.84481454e-01
5.65333307e-01 -5.59592009e-01 -3.85511965e-01 4.23593134e-01
-1.94639012e-01 -2.24587187e-01 5.98876715e-01 -2.98296690e-01
-7.01918304e-01 -2.75684565e-01 -1.22339463e+00 1.07626820e+00
1.75557005e+00 -1.34450984e+00 -5.74617863e-01 3.37973058e-01
5.79473376e-01 -2.56798863e-01 -4.01934415e-01 4.82593089e-01
1.53553092e+00 -7.26050615e-01 7.79304981e-01 -2.35401273e-01
1.76927760e-01 -2.43032277e-01 -3.34621817e-01 -1.13231242e+00
1.60271272e-01 -1.03846252e+00 -3.82683426e-02 9.85715985e-01
3.88602078e-01 -1.18344212e+00 7.02010632e-01 6.98051870e-01
-1.46578774e-01 -6.36121333e-01 -1.59935224e+00 -9.71707821e-01
6.69650808e-02 -8.53578091e-01 1.34240448e-01 4.16543126e-01
1.52816102e-01 1.77370340e-01 -6.25121653e-01 -6.83825761e-02
7.06125140e-01 -1.50776640e-01 6.47893190e-01 -1.00956392e+00
1.30419731e-01 -7.46925414e-01 -7.39251077e-01 -1.15556180e+00
2.92487182e-02 -7.37972975e-01 2.47281834e-01 -1.99896920e+00
-1.39301047e-01 1.01552261e-02 -1.51653796e-01 7.69886613e-01
5.29455245e-01 5.94563127e-01 5.80485985e-02 1.95655391e-01
3.05636972e-01 6.58629596e-01 1.23899949e+00 -4.33148533e-01
-5.59628546e-01 1.10532887e-01 -3.98005188e-01 5.07871807e-01
6.69166148e-01 -1.68325990e-01 -1.18503079e-01 -6.92229927e-01
5.16689904e-02 -2.67183393e-01 5.57084441e-01 -9.83355880e-01
3.04492354e-01 5.49612045e-02 -9.86025482e-02 -4.66304123e-01
6.46219134e-01 -6.06109858e-01 -3.96907896e-01 8.56467724e-01
-3.10339451e-01 -7.01430142e-01 8.64703879e-02 3.18257868e-01
-2.81378683e-02 6.50554717e-01 9.99705672e-01 2.00063854e-01
-3.87173563e-01 -5.43129910e-03 -8.80408943e-01 -1.60248861e-01
5.62690318e-01 -6.01883054e-01 -4.37936693e-01 -6.27942145e-01
-6.24066293e-01 -2.59159654e-01 2.95112222e-01 6.35879397e-01
9.89337802e-01 -1.43274701e+00 -8.51084888e-01 4.84298110e-01
3.81978959e-01 -3.17922741e-01 8.56880918e-02 8.03520083e-01
-5.79850376e-01 5.51780522e-01 -2.05844223e-01 -5.71707904e-01
-1.38408482e+00 -3.62953842e-01 4.16458368e-01 5.37706554e-01
-1.12753522e+00 1.15950811e+00 -5.59746385e-01 -3.45537931e-01
4.98491675e-01 -6.23515606e-01 4.88607854e-01 -1.52164117e-01
5.05751133e-01 3.17394674e-01 -6.14648163e-02 -6.15394533e-01
-4.17685598e-01 1.11359525e+00 5.17656431e-02 -6.02242529e-01
1.03390658e+00 -1.54526934e-01 1.05556980e-01 6.86375380e-01
1.15632617e+00 1.41188219e-01 -8.53685260e-01 -3.04118395e-01
-4.10621554e-01 -3.16647679e-01 -8.73752125e-03 -1.25457501e+00
-4.26579118e-01 1.09386361e+00 1.01814926e+00 -5.58123112e-01
1.14876068e+00 7.49770924e-02 1.64766133e+00 4.79841739e-01
6.38354540e-01 -1.17739165e+00 3.02779451e-02 4.70808744e-01
1.22504365e+00 -1.16795528e+00 -5.45456767e-01 -1.17265172e-01
-6.38925806e-02 8.80953312e-01 2.71342784e-01 3.48285019e-01
3.47666115e-01 4.03175414e-01 8.44954550e-01 1.40334934e-01
-2.36495420e-01 -2.46121630e-01 8.29808936e-02 9.03127253e-01
5.70673764e-01 3.49583328e-01 -5.13043463e-01 -2.79637380e-03
-5.93216717e-01 3.78350168e-01 1.49182916e-01 1.03197360e+00
-4.04722512e-01 -8.95657301e-01 -3.38275075e-01 3.73500258e-01
4.16721664e-02 -1.36706337e-01 -7.67654836e-01 3.52174878e-01
-7.34159797e-02 1.03844512e+00 -3.62870932e-01 -6.08851969e-01
5.40648043e-01 3.98954004e-01 7.26204336e-01 -1.56051233e-01
-2.61724778e-02 -1.26106575e-01 6.49968684e-02 -6.14836931e-01
-1.77437708e-01 -5.99973857e-01 -1.16292858e+00 -2.55592316e-01
-2.70901732e-02 -7.98915684e-01 8.93168390e-01 1.11936486e+00
3.50417405e-01 6.46021366e-02 1.41075999e-01 -6.36766851e-01
-3.39385182e-01 -1.22284925e+00 -3.69434118e-01 8.34656730e-02
7.87728786e-01 -4.70061094e-01 -4.99278426e-01 -2.89555103e-01] | [9.121747016906738, -6.435465335845947] |
32ea0d3e-2772-4de3-aae9-9b22c5b16b47 | decipherment-for-adversarial-offensive | null | null | https://aclanthology.org/W18-5119 | https://aclanthology.org/W18-5119.pdf | Decipherment for Adversarial Offensive Language Detection | Automated filters are commonly used by online services to stop users from sending age-inappropriate, bullying messages, or asking others to expose personal information. Previous work has focused on rules or classifiers to detect and filter offensive messages, but these are vulnerable to cleverly disguised plaintext and unseen expressions especially in an adversarial setting where the users can repeatedly try to bypass the filter. In this paper, we model the disguised messages as if they are produced by encrypting the original message using an invented cipher. We apply automatic decipherment techniques to decode the disguised malicious text, which can be then filtered using rules or classifiers. We provide experimental results on three different datasets and show that decipherment is an effective tool for this task. | ['Nishant Kambhatla', 'Anoop Sarkar', 'Zhelun Wu'] | 2018-10-01 | null | null | null | ws-2018-10 | ['decipherment'] | ['natural-language-processing'] | [ 5.03153145e-01 5.47184683e-02 -2.55753428e-01 -2.70573944e-01
-3.13112140e-01 -1.41498995e+00 9.67537045e-01 2.47542903e-01
-3.36547494e-01 6.94243908e-01 1.24153875e-01 -6.13542616e-01
2.43125185e-01 -8.95465434e-01 -3.35702151e-01 -4.45177406e-01
-3.06040287e-01 -2.36819729e-01 5.51476516e-02 -3.97341311e-01
4.45545316e-01 5.36463380e-01 -1.21389222e+00 6.32228613e-01
8.38808000e-01 7.64898479e-01 -9.46342826e-01 1.25199378e+00
2.22309530e-01 8.25326622e-01 -1.00756443e+00 -1.30632842e+00
5.01942456e-01 -5.12573421e-01 -7.72479296e-01 -4.69980717e-01
4.35688585e-01 -5.56352496e-01 -4.57284242e-01 1.63319421e+00
4.06167597e-01 -4.42324311e-01 3.18249166e-01 -1.25998104e+00
-3.74049336e-01 7.24605918e-01 1.76452734e-02 3.46960425e-01
1.03327131e+00 2.77794451e-01 6.19826078e-01 -3.16255987e-01
3.80446136e-01 1.52363873e+00 3.65544319e-01 8.55053067e-01
-1.33950782e+00 -1.22874379e+00 -1.86701730e-01 1.88359674e-02
-1.22274816e+00 -7.15213656e-01 5.42929113e-01 -3.68995517e-01
9.47255269e-02 1.15567327e+00 3.99399787e-01 1.63624132e+00
5.26118398e-01 6.28800571e-01 1.30437183e+00 -3.47592533e-01
1.00889146e-01 5.73584616e-01 2.03141019e-01 5.56303203e-01
4.20866758e-01 4.43766505e-01 -2.43467420e-01 -9.82556164e-01
-1.49363682e-01 -9.76584852e-02 -6.76338851e-01 3.81302476e-01
-7.26786196e-01 1.43685305e+00 1.55507565e-01 1.47514269e-01
-3.80049762e-03 -2.73472160e-01 6.91815317e-01 8.52722585e-01
4.11505818e-01 8.50010991e-01 -3.10216844e-01 -7.34959766e-02
-5.09447217e-01 3.61191869e-01 1.48561668e+00 4.71634507e-01
4.87841964e-01 -2.59360492e-01 -5.51281217e-03 3.85784000e-01
1.55985773e-01 6.91437840e-01 5.38736343e-01 -5.25108755e-01
2.84359783e-01 5.16394377e-01 7.44554272e-04 -1.61855304e+00
2.32134506e-01 -1.00286104e-01 -6.47307098e-01 1.46924898e-01
4.99167085e-01 -5.76420188e-01 -7.93937266e-01 1.28141868e+00
-3.97487031e-03 5.50541431e-02 -1.47051305e-01 6.07714534e-01
4.27091986e-01 3.41467291e-01 3.43824983e-01 3.47141661e-02
1.45983064e+00 -1.31278098e-01 -7.73507893e-01 -1.33275479e-01
6.19204223e-01 -7.14814723e-01 8.40912819e-01 7.65607476e-01
-6.65516078e-01 -3.42027135e-02 -1.33238208e+00 4.47787195e-01
-9.14914668e-01 -6.12112522e-01 5.14407575e-01 1.84393835e+00
-4.57861543e-01 8.00940812e-01 -4.60863411e-02 -4.34361733e-02
5.86306453e-01 5.81338167e-01 -4.65725154e-01 2.29721606e-01
-2.03968668e+00 7.92666376e-01 2.79908985e-01 -3.31228256e-01
-8.51650000e-01 -5.86968899e-01 -6.89901292e-01 -1.13177232e-01
3.06259304e-01 -1.40661858e-02 1.27315402e+00 -1.27237546e+00
-1.33368993e+00 1.20145929e+00 3.18640023e-01 -7.71664262e-01
7.54352033e-01 -3.23497504e-01 -1.11180258e+00 2.61915654e-01
-2.54852504e-01 -2.46099770e-01 1.77639031e+00 -1.07781148e+00
-3.21681708e-01 -3.92573982e-01 1.95736229e-01 -3.81042957e-01
-7.16807902e-01 4.87787813e-01 3.94594282e-01 -8.43915582e-01
-5.36879301e-01 -8.34284961e-01 -9.09862667e-02 -3.77437145e-01
-8.98592055e-01 2.74000734e-01 1.19819009e+00 -8.83515000e-01
1.81728446e+00 -1.96586680e+00 -4.87042278e-01 8.64488363e-01
4.15383309e-01 1.17001414e+00 3.25606316e-01 6.07648492e-01
-1.23853780e-01 9.75270450e-01 5.29480875e-02 1.87124491e-01
-7.07115978e-02 2.48060793e-01 -8.97639573e-01 7.52929270e-01
-1.00114688e-01 8.43673229e-01 -9.04015005e-01 -2.68759400e-01
6.03016056e-02 2.15400085e-01 -3.25236589e-01 2.80433148e-01
7.48889463e-04 1.89082623e-01 -6.61282897e-01 4.75483656e-01
8.73868346e-01 4.30703610e-01 2.60529727e-01 2.02993378e-01
3.45474571e-01 4.46979403e-01 -9.41454649e-01 4.25965041e-01
-2.63213754e-01 6.52236819e-01 4.78663981e-01 -4.45549637e-01
8.17887425e-01 1.68361664e-01 -3.66431594e-01 -2.26569831e-01
8.33299398e-01 3.72259989e-02 -6.80846050e-02 -6.68500364e-01
3.20078164e-01 -3.91120091e-03 -4.35828447e-01 6.70531094e-01
-2.52476811e-01 1.96902648e-01 -4.38523293e-03 5.38554072e-01
1.33688891e+00 -8.01116347e-01 3.76965374e-01 -1.07193843e-01
1.15395880e+00 -4.80283797e-01 2.14831352e-01 1.13604820e+00
-3.53687167e-01 1.05027705e-01 6.64939523e-01 -4.25319076e-01
-5.81337214e-01 -9.38887477e-01 1.16500802e-01 1.18888533e+00
6.80841655e-02 -9.41927016e-01 -8.62558961e-01 -1.57227373e+00
3.13923776e-01 8.11247945e-01 -5.07318556e-01 -6.82966113e-01
-5.28743088e-01 -5.49649417e-01 1.29135668e+00 -2.61904567e-01
5.24693966e-01 -6.12824917e-01 -1.44238308e-01 1.68842942e-01
-7.30298981e-02 -1.04422998e+00 -5.92704594e-01 -2.42046922e-01
-2.26630002e-01 -1.34009409e+00 -5.81592508e-02 -1.87677637e-01
5.48368931e-01 1.11617669e-02 6.19444907e-01 6.61561131e-01
-2.89800525e-01 1.61467031e-01 -4.17039901e-01 -8.50913286e-01
-1.35153461e+00 2.16899216e-01 1.09435163e-01 7.27844596e-01
7.95349717e-01 -4.00477499e-01 -4.80350226e-01 4.48778480e-01
-1.34032869e+00 -6.25813782e-01 3.26205674e-03 3.09759647e-01
-7.59770691e-01 3.18319947e-01 5.55778801e-01 -1.54747581e+00
1.37040055e+00 -5.45320570e-01 -2.50149757e-01 -1.06012672e-01
-4.19278562e-01 4.76085655e-02 1.15688062e+00 -1.04927778e+00
-7.72901356e-01 -1.36627540e-01 -4.05185193e-01 3.82860564e-02
-3.79200339e-01 -1.68780655e-01 -4.88974512e-01 -7.26424694e-01
9.75303292e-01 2.85000473e-01 2.03269795e-01 -4.26838964e-01
4.11162853e-01 1.17244101e+00 4.12745059e-01 -2.77477324e-01
1.55417740e+00 5.68374753e-01 -3.83524150e-01 -7.89659083e-01
-8.40438545e-01 -2.37947688e-01 -2.55310386e-01 -3.26514721e-01
4.41358149e-01 -4.95864570e-01 -1.24564886e+00 5.99470317e-01
-9.01886404e-01 2.72223860e-01 4.92747515e-01 -2.75717024e-02
3.93378139e-02 1.02887177e+00 -6.71657324e-01 -7.58697927e-01
-5.50860167e-01 -6.25691950e-01 3.21350664e-01 9.90222916e-02
-8.67573738e-01 -9.55178201e-01 2.51724441e-02 2.18832701e-01
6.63189471e-01 6.00366175e-01 8.09521496e-01 -1.42936742e+00
8.27884153e-02 -1.28629744e+00 2.60978669e-01 7.33631492e-01
2.60377616e-01 1.79087937e-01 -9.68955159e-01 -2.81089991e-01
2.99576014e-01 -1.78132087e-01 5.63011825e-01 -8.32672715e-01
1.12903452e+00 -1.28248262e+00 -3.58798802e-01 5.73819995e-01
9.94163632e-01 -4.89788922e-03 9.10797596e-01 2.66441673e-01
3.80515367e-01 5.75229645e-01 1.17130071e-01 6.58023477e-01
-1.65340886e-01 1.92932293e-01 5.61767876e-01 2.37712204e-01
7.06338286e-01 -8.34801733e-01 5.51515043e-01 -6.59583881e-02
2.69166678e-01 -2.52109796e-01 -4.53935087e-01 9.00736265e-03
-1.39734256e+00 -1.26197302e+00 -2.08334714e-01 2.12179303e+00
1.00141549e+00 5.91113508e-01 1.79577604e-01 5.43760538e-01
8.72380137e-01 3.34150314e-01 -7.46772513e-02 -1.24414206e+00
-6.12370819e-02 5.49867511e-01 1.00220776e+00 5.66978037e-01
-1.45059121e+00 1.08078206e+00 6.62365484e+00 8.78149331e-01
-1.07399809e+00 -1.25633478e-01 6.40197158e-01 1.27509832e-01
-7.43749319e-03 6.10447004e-02 -7.31046796e-01 9.30464923e-01
1.16864049e+00 -2.34911561e-01 4.82770056e-01 9.07865882e-01
1.08311310e-01 6.66263774e-02 -8.73548090e-01 8.39701295e-01
3.31210762e-01 -1.05134928e+00 3.96192167e-03 8.54740739e-02
2.80998170e-01 -7.23067820e-01 1.27183810e-01 3.43232721e-01
7.36411750e-01 -1.18560255e+00 3.96785051e-01 1.38660580e-01
3.79841059e-01 -8.64633143e-01 4.98496801e-01 4.75286871e-01
-4.26176369e-01 -2.26065978e-01 -6.77440837e-02 -2.93178439e-01
-1.03358261e-01 3.57989550e-01 -7.57634938e-01 1.08218551e-01
3.64076346e-01 3.35647792e-01 -8.74798119e-01 4.99341041e-01
-7.31562912e-01 1.08558643e+00 -3.59224647e-01 -4.69150275e-01
1.28300354e-01 1.24602810e-01 8.18675637e-01 1.50304413e+00
-3.33211660e-01 2.45497916e-02 -1.61749259e-01 3.95474732e-01
-2.49836594e-01 6.19619824e-02 -9.09288824e-01 -2.80156732e-01
1.97709218e-01 1.46291339e+00 -4.03939545e-01 -3.83807927e-01
1.00219384e-01 1.30889022e+00 -1.14307247e-01 2.82471955e-01
-6.47387385e-01 -8.17999601e-01 1.04983592e+00 2.97661722e-01
1.17530055e-01 -2.97640450e-02 -2.84226499e-02 -1.36762464e+00
-2.71629035e-01 -1.54996049e+00 6.98062420e-01 -9.83853191e-02
-1.40704322e+00 4.57228094e-01 -3.43959093e-01 -9.03490603e-01
-1.44204527e-01 -4.34712589e-01 -9.55486536e-01 5.57477236e-01
-1.04196298e+00 -6.90482497e-01 2.84442961e-01 8.41770172e-01
-1.38533330e-02 -2.20626488e-01 7.80515730e-01 6.85411245e-02
-3.60007167e-01 9.83014822e-01 -1.87329814e-01 8.06900203e-01
6.54421449e-01 -9.17595744e-01 4.03564721e-01 9.80051994e-01
-1.59027815e-01 1.01943386e+00 1.01455069e+00 -9.27051544e-01
-1.34850454e+00 -8.25631320e-01 1.07422531e+00 -7.37179816e-01
9.32665408e-01 -1.19342971e+00 -8.76632035e-01 5.58521509e-01
2.29721427e-01 -3.24797243e-01 9.73907650e-01 -3.53737950e-01
-9.21643317e-01 2.01698229e-01 -1.56092238e+00 8.47305954e-01
7.01332629e-01 -6.75470412e-01 -5.16144812e-01 2.34357312e-01
4.62932616e-01 6.97527379e-02 -4.98623937e-01 8.42290148e-02
7.56667674e-01 -9.73246753e-01 8.51438880e-01 -1.30893838e+00
-8.58058557e-02 -1.89132169e-02 1.54654607e-01 -1.01669812e+00
-3.56529579e-02 -1.70432258e+00 -3.84916782e-01 1.40573847e+00
1.82066858e-01 -8.37777495e-01 7.21946597e-01 6.67623401e-01
9.49554920e-01 -2.09449008e-01 -6.15587890e-01 -5.23036599e-01
-1.23915121e-01 -4.49386358e-01 6.78796828e-01 1.14739180e+00
1.47737294e-01 2.96610117e-01 -7.47311234e-01 3.58165592e-01
7.15542793e-01 -3.88185680e-01 9.47865665e-01 -1.01587272e+00
9.34619084e-02 -3.00136954e-01 -5.57681382e-01 -6.81838334e-01
2.71199495e-01 -8.71053100e-01 -5.06819725e-01 -1.78123057e-01
-4.49281156e-01 5.35196513e-02 1.50635540e-01 2.35466585e-02
-2.15045571e-01 3.63398463e-01 2.96660155e-01 -1.79146379e-01
-4.41435993e-01 8.64384547e-02 5.65118849e-01 -1.96572810e-01
-1.20197549e-01 7.40395427e-01 -1.26814389e+00 9.97439623e-01
1.09921288e+00 -8.63224268e-01 3.19332466e-04 6.10818624e-01
5.22105038e-01 -6.03240669e-01 3.69239956e-01 -5.71057439e-01
-1.75632983e-01 -8.78626406e-02 3.34032863e-01 4.88631614e-02
-1.10958196e-01 -8.45060706e-01 -3.11268657e-01 8.91627610e-01
-5.98092735e-01 -6.32697642e-02 -2.93129206e-01 6.28953278e-01
2.16162860e-01 -4.49482232e-01 1.01145279e+00 -3.47506166e-01
-2.07157955e-01 2.49033540e-01 -1.26145113e+00 1.43449411e-01
1.11458826e+00 -2.30515212e-01 -4.48415428e-01 -9.32519913e-01
-7.45542586e-01 1.10937618e-01 3.55191201e-01 6.38510227e-01
6.63258314e-01 -9.79339778e-01 -7.98724055e-01 6.37775064e-01
-6.92819804e-02 -1.16592252e+00 -2.77517051e-01 1.20386042e-01
-6.01284862e-01 -6.16500452e-02 5.06842360e-02 4.51639444e-02
-1.96842325e+00 8.46388221e-01 2.87797689e-01 -1.56868339e-01
-1.88662827e-01 5.29301465e-01 -4.19634044e-01 -2.12380603e-01
3.15586150e-01 1.12844430e-01 -1.88598871e-01 2.25841522e-01
1.12791264e+00 5.18527985e-01 -5.31136207e-02 -7.87515163e-01
-3.71133059e-01 1.07470285e-02 -4.74786460e-01 2.39002198e-01
6.87303543e-01 -2.06619292e-01 -3.77685249e-01 -2.93099850e-01
1.66439056e+00 5.67225635e-01 -3.73708487e-01 -1.35904461e-01
2.85212368e-01 -1.03561580e+00 -4.44905430e-01 -6.81625485e-01
-5.44175148e-01 5.44667363e-01 2.55298644e-01 1.13567615e+00
8.67193520e-01 -4.02444094e-01 9.62865651e-01 6.13627955e-02
-5.06062200e-03 -8.58278096e-01 -5.48115037e-02 3.66714925e-01
5.52035511e-01 -1.06003380e+00 2.41062678e-02 -7.17674911e-01
-5.98934233e-01 1.22969222e+00 2.91431248e-01 -1.79924384e-01
9.44806695e-01 1.95069879e-01 2.00950935e-01 9.10718367e-02
-6.42891049e-01 2.96223998e-01 1.04383796e-01 7.75133550e-01
-2.52416074e-01 1.84645563e-01 -5.55664062e-01 6.02493584e-01
-6.04850113e-01 -3.92209053e-01 8.94891620e-01 1.06123221e+00
-4.96297061e-01 -1.35459518e+00 -7.49726772e-01 6.62513077e-01
-1.32555187e+00 -5.65264374e-02 -1.37027752e+00 3.60212326e-01
9.17721726e-03 1.33199883e+00 -5.91984332e-01 -9.63424027e-01
4.64559160e-02 1.79317266e-01 -2.49796826e-02 -4.06373203e-01
-1.33443451e+00 -4.94661272e-01 4.86991942e-01 -4.23865795e-01
1.23627886e-01 -2.36505985e-01 -5.19138515e-01 -9.55159545e-01
-3.45263869e-01 2.57471681e-01 4.68798786e-01 7.69841373e-01
3.38360786e-01 -4.38890100e-01 1.27314925e+00 -2.58472800e-01
-9.61548269e-01 -5.55341899e-01 -3.69792879e-01 9.95067239e-01
6.25627875e-01 -2.64467508e-01 -9.48683441e-01 -1.86748415e-01] | [5.888779163360596, 7.869454860687256] |
22e6463b-b366-44a1-a67e-43702aaf9bc6 | learning-physically-realizable-skills-for | 2212.02094 | null | https://arxiv.org/abs/2212.02094v2 | https://arxiv.org/pdf/2212.02094v2.pdf | Learning Physically Realizable Skills for Online Packing of General 3D Shapes | We study the problem of learning online packing skills for irregular 3D shapes, which is arguably the most challenging setting of bin packing problems. The goal is to consecutively move a sequence of 3D objects with arbitrary shapes into a designated container with only partial observations of the object sequence. Meanwhile, we take physical realizability into account, involving physics dynamics and constraints of a placement. The packing policy should understand the 3D geometry of the object to be packed and make effective decisions to accommodate it in the container in a physically realizable way. We propose a Reinforcement Learning (RL) pipeline to learn the policy. The complex irregular geometry and imperfect object placement together lead to huge solution space. Direct training in such space is prohibitively data intensive. We instead propose a theoretically-provable method for candidate action generation to reduce the action space of RL and the learning burden. A parameterized policy is then learned to select the best placement from the candidates. Equipped with an efficient method of asynchronous RL acceleration and a data preparation process of simulation-ready training sequences, a mature packing policy can be trained in a physics-based environment within 48 hours. Through extensive evaluation on a variety of real-life shape datasets and comparisons with state-of-the-art baselines, we demonstrate that our method outperforms the best-performing baseline on all datasets by at least 12.8% in terms of packing utility. | ['Kai Xu', 'Yang Yu', 'Zherong Pan', 'Hang Zhao'] | 2022-12-05 | null | null | null | null | ['action-generation'] | ['computer-vision'] | [-1.99012950e-01 -3.35638933e-02 -1.43909216e-01 5.32773212e-02
-7.18288064e-01 -8.54554415e-01 1.90014720e-01 7.30769634e-01
-5.49727023e-01 6.90719783e-01 -1.51241913e-01 -8.14914167e-01
-1.39098853e-01 -9.19733107e-01 -1.39279997e+00 -8.07854354e-01
-5.07544518e-01 1.46868026e+00 1.15272067e-01 -1.39918178e-01
3.30183834e-01 8.30317616e-01 -1.41593254e+00 -5.05949408e-02
9.66635406e-01 1.11225748e+00 7.64924824e-01 9.23208177e-01
-3.34253721e-02 1.74746007e-01 -4.20003265e-01 -9.56542119e-02
6.71526372e-01 5.51745249e-03 -7.19633043e-01 6.15936220e-01
-9.82917938e-03 -3.95261943e-01 -3.81161958e-01 6.06024563e-01
5.29674649e-01 5.54087639e-01 5.26706517e-01 -8.43643248e-01
-2.82003015e-01 3.52512300e-01 -3.99612606e-01 3.30283493e-01
2.35317513e-01 6.47686183e-01 8.16879392e-01 -3.58982980e-01
4.04478014e-01 1.03521216e+00 1.20671205e-02 2.33165860e-01
-1.21298110e+00 -2.07144041e-02 4.17412162e-01 -9.28402022e-02
-8.24045956e-01 8.66731554e-02 2.50101805e-01 -1.71835095e-01
1.27819228e+00 1.85156256e-01 1.02623332e+00 6.60543263e-01
3.46446931e-01 8.00465763e-01 7.18146622e-01 -2.86270261e-01
8.90358746e-01 -5.49221992e-01 -1.49009019e-01 7.05451310e-01
4.88752127e-01 2.39304766e-01 -2.38760129e-01 -1.78904906e-01
5.81190228e-01 1.92664132e-01 8.88405517e-02 -8.40004027e-01
-7.47312725e-01 7.62594581e-01 4.91084069e-01 -4.24438357e-01
-3.60911101e-01 2.61886679e-02 2.36706525e-01 1.20352827e-01
2.50157028e-01 9.71358955e-01 -7.93973505e-01 -4.18103456e-01
-5.31835973e-01 8.53452682e-01 1.18878245e+00 1.37721920e+00
4.04278398e-01 -1.99470952e-01 -1.90105081e-01 2.77442127e-01
1.14487745e-01 8.96497548e-01 -2.13342816e-01 -6.70038581e-01
9.69120681e-01 3.70602727e-01 8.14512789e-01 -5.04145384e-01
-4.99790847e-01 -2.21092969e-01 -4.09888476e-01 8.91837925e-02
5.89979231e-01 -1.37400582e-01 -1.18109143e+00 1.09912741e+00
7.66835213e-01 2.32916608e-01 -1.51046619e-01 1.10289168e+00
3.14031780e-01 1.03255689e+00 -1.58006757e-01 -2.22087353e-01
1.33016503e+00 -1.24488294e+00 9.08143371e-02 -3.85515332e-01
5.77499866e-01 -6.49940312e-01 7.62936950e-01 4.07307208e-01
-1.53220379e+00 -3.59224647e-01 -1.09437895e+00 2.44044200e-01
-1.46765977e-01 -2.24390373e-01 1.03106511e+00 3.45074236e-01
-6.74482584e-01 1.02795744e+00 -1.35271847e+00 5.25012650e-02
4.07857358e-01 6.81932330e-01 1.07145738e-02 -3.29246700e-01
-4.28944111e-01 8.20846677e-01 4.42912340e-01 1.31853819e-02
-1.08974183e+00 -7.92522788e-01 -9.08221364e-01 2.35264271e-01
7.75659502e-01 -5.64662814e-01 1.63041151e+00 -4.49541695e-02
-1.57431304e+00 4.16027218e-01 -6.35272041e-02 -7.23536611e-01
4.21677679e-01 -1.17285781e-01 1.88506499e-01 1.19096063e-01
-2.05826432e-01 4.39211935e-01 6.93166196e-01 -1.16467357e+00
-7.58266985e-01 -1.70240834e-01 3.53900641e-01 4.90697771e-01
3.94552350e-01 -4.19456929e-01 -5.95254660e-01 -4.06756997e-02
2.70830035e-01 -1.19347525e+00 -7.12204993e-01 -3.82423371e-01
-1.02277644e-01 -2.95470536e-01 2.49406219e-01 -3.01539391e-01
6.19514346e-01 -1.78122628e+00 2.64776677e-01 1.98736042e-01
-1.72796298e-03 2.58935928e-01 -1.36599690e-01 5.75658023e-01
4.26249534e-01 -1.53740183e-01 1.57545105e-01 -5.01798034e-01
2.83290982e-01 5.95589757e-01 -5.94894469e-01 5.38077652e-01
2.05302835e-01 1.07833838e+00 -1.11416101e+00 -9.39048678e-02
3.68000329e-01 -2.53971159e-01 -1.17398930e+00 5.84472656e-01
-9.33404863e-01 4.54553515e-01 -6.75777197e-01 5.10808468e-01
9.54644084e-01 -5.80099523e-01 3.63292128e-01 4.28960294e-01
-3.28496061e-02 5.79750597e-01 -1.21335566e+00 1.65275395e+00
-5.09088635e-01 -2.78110951e-01 -3.36179137e-02 -1.22025847e+00
8.25136006e-01 -2.59795427e-01 4.39133137e-01 -6.59546316e-01
1.62955329e-01 2.94897109e-01 1.95076317e-01 -4.12093818e-01
6.52176142e-01 1.28141314e-01 -2.29232565e-01 5.25403142e-01
-2.33996928e-01 -5.76755583e-01 3.84502411e-01 -5.07460237e-02
1.37953246e+00 2.30822414e-01 6.75810575e-02 -2.18021438e-01
2.58540139e-02 1.37849450e-01 5.01449227e-01 1.08093870e+00
1.36395991e-01 2.94892758e-01 3.80465150e-01 -8.49307716e-01
-1.43584788e+00 -1.14344680e+00 1.24276683e-01 9.15928245e-01
6.52386904e-01 -1.54928878e-01 -4.51775402e-01 -4.76208389e-01
4.68241036e-01 5.57831526e-01 -3.95368814e-01 4.27832156e-02
-9.34973478e-01 -6.21295571e-01 -6.03808641e-01 5.63123524e-01
-3.33046094e-02 -1.23526490e+00 -9.51671124e-01 5.21182001e-01
3.92459154e-01 -1.28814435e+00 -5.67943037e-01 7.18466520e-01
-9.64294195e-01 -1.02884400e+00 -1.92657217e-01 -7.44849622e-01
1.07478380e+00 6.16257846e-01 1.24011993e+00 3.16032171e-01
-4.38296288e-01 3.15395966e-02 -4.58036840e-01 -2.75867790e-01
-4.01347548e-01 2.78594464e-01 4.46353219e-02 -6.57381713e-01
2.81656738e-02 -4.16826308e-01 -7.75558650e-01 1.75736725e-01
-5.49397886e-01 3.06213915e-01 5.78144372e-01 8.42952609e-01
7.81249583e-01 1.91482008e-01 -1.55684259e-02 -4.67169434e-01
2.89900661e-01 -4.71631110e-01 -1.38393116e+00 6.56260028e-02
-2.44315803e-01 4.34736103e-01 1.04445374e+00 -5.13125598e-01
-4.93429303e-01 -8.91783088e-03 4.32322770e-02 -4.00360376e-01
-6.98101521e-02 1.84001058e-01 -6.29057512e-02 1.06916241e-01
1.53953478e-01 2.02691466e-01 -7.38038644e-02 -3.88484180e-01
2.84512676e-02 -1.00828558e-01 1.89844266e-01 -1.44374633e+00
6.49830937e-01 1.54785767e-01 1.96074188e-01 -2.90231794e-01
-1.03223550e+00 -4.82470334e-01 -4.37609285e-01 1.56624749e-01
4.54016924e-01 -5.99029541e-01 -1.36026597e+00 1.29945263e-01
-1.06288505e+00 -9.42001939e-01 -3.91564429e-01 2.80234784e-01
-9.83091295e-01 1.11197576e-01 -5.40745616e-01 -9.05077755e-01
-2.58093745e-01 -1.44675112e+00 1.25018346e+00 4.54457998e-01
2.93998867e-01 -5.91853619e-01 -1.07266225e-01 3.37476343e-01
2.68586785e-01 2.05251183e-02 1.10448635e+00 -6.60681129e-01
-1.15327549e+00 5.48854545e-02 6.35031657e-03 -3.04544657e-01
-1.45839453e-01 -2.00475037e-01 -1.39980540e-01 -7.14609563e-01
-2.06919149e-01 -4.28287446e-01 4.90527779e-01 5.67802191e-01
1.58666039e+00 -4.63999212e-01 -5.57331562e-01 6.15791678e-01
1.42457771e+00 4.79701281e-01 2.73291707e-01 3.84491891e-01
2.11183280e-01 -2.13284441e-03 9.74923313e-01 8.48282456e-01
3.32168967e-01 6.67243659e-01 6.79611862e-01 3.00689101e-01
3.36318105e-01 -5.38557291e-01 -2.23224312e-02 6.98804080e-01
6.93562813e-03 -7.12762415e-01 -9.75611746e-01 2.38459542e-01
-1.95856392e+00 -6.68908656e-01 4.53720748e-01 2.43184805e+00
7.58773863e-01 5.78287959e-01 1.46675855e-01 -2.71510303e-01
4.10400212e-01 -1.33779302e-01 -9.93397236e-01 -5.77638447e-01
3.69389415e-01 4.84582871e-01 1.01625788e+00 5.96181214e-01
-1.19030285e+00 9.83888805e-01 5.76195621e+00 7.50265598e-01
-7.78163254e-01 -3.38630557e-01 7.57000923e-01 -3.58496577e-01
-1.12177685e-01 7.11359233e-02 -1.08884227e+00 6.70151293e-01
9.84665871e-01 1.11006513e-01 1.11311877e+00 9.19335246e-01
2.27564886e-01 -4.72218603e-01 -1.22725821e+00 7.31909633e-01
-4.41850245e-01 -1.64082527e+00 -2.75352985e-01 3.97041321e-01
8.91579449e-01 6.50199428e-02 -1.23835981e-01 4.92934346e-01
7.73113132e-01 -9.37983692e-01 6.62181497e-01 2.22636685e-01
2.54117399e-01 -9.80685115e-01 5.01912653e-01 6.48739636e-01
-1.15668964e+00 -2.58048981e-01 -7.85105348e-01 -3.18074942e-01
3.97860110e-01 3.95951152e-01 -1.12589228e+00 4.87843722e-01
5.28666973e-01 1.01497844e-01 3.42603028e-02 1.43321371e+00
-6.71664104e-02 3.95591468e-01 -7.29887068e-01 -4.54972804e-01
5.46267927e-01 -4.17616159e-01 3.75608385e-01 7.45223939e-01
3.29714149e-01 6.21058762e-01 9.25335169e-01 6.25232756e-01
1.54478438e-02 -1.56228155e-01 -3.95824254e-01 -1.17815055e-01
6.49110198e-01 1.08047974e+00 -1.10744488e+00 -1.27199978e-01
-7.92967379e-02 6.82983100e-01 7.37261057e-01 -1.65758692e-02
-1.09152818e+00 2.15148404e-01 7.34902740e-01 1.13442652e-01
9.30818319e-01 -7.48029768e-01 -1.92040771e-01 -7.28879631e-01
1.39231443e-01 -8.58251393e-01 1.47128582e-01 -3.37590635e-01
-1.24213576e+00 1.40605018e-01 -5.58695383e-02 -8.84544909e-01
-5.54119982e-02 -8.36524367e-01 -5.69812953e-01 6.11412704e-01
-1.50347388e+00 -3.52793962e-01 -6.23315237e-02 -1.26160577e-01
6.33560121e-01 -8.82190689e-02 6.46467447e-01 -1.43490747e-01
-3.72149467e-01 3.92489046e-01 2.05800712e-01 -2.94967175e-01
1.95573829e-02 -1.46892989e+00 1.00905204e+00 5.21368623e-01
-8.68240818e-02 1.46604851e-01 9.83364940e-01 -7.85979867e-01
-2.44616604e+00 -8.61317635e-01 2.98341423e-01 -3.28277230e-01
7.28327513e-01 -7.03733146e-01 -6.54662132e-01 2.74173707e-01
-1.93774402e-01 2.59743303e-01 2.33883709e-01 -3.98021415e-02
2.11005747e-01 4.64779586e-02 -9.30941224e-01 5.73158383e-01
1.40485728e+00 3.31619501e-01 -3.77129942e-01 8.28985274e-01
1.08425725e+00 -1.32811475e+00 -7.56190479e-01 3.26801121e-01
7.87385777e-02 -4.60669070e-01 8.88705909e-01 -1.05275071e+00
5.54190218e-01 -2.30345964e-01 -4.26548049e-02 -1.30140376e+00
-2.44922519e-01 -1.08450985e+00 -6.45758212e-01 4.84917819e-01
4.06137496e-01 -4.92288470e-01 1.12766623e+00 8.60327780e-01
-4.00369763e-01 -1.54735816e+00 -7.64195621e-01 -9.16828156e-01
1.22852780e-01 -2.15686291e-01 8.38375688e-01 7.32818572e-03
3.31438854e-02 2.84209102e-02 -1.51245669e-01 6.15498662e-01
4.26738769e-01 9.76038337e-01 1.00275207e+00 -7.35364616e-01
-7.99014211e-01 -2.24855825e-01 -8.83504525e-02 -1.77256906e+00
1.70242563e-01 -7.17474282e-01 3.97748381e-01 -1.56518114e+00
2.41496146e-01 -1.15335512e+00 1.16145119e-01 2.77156174e-01
-1.52412787e-01 -6.83415890e-01 4.94601995e-01 -9.46006104e-02
-1.02230108e+00 6.28558576e-01 1.71106029e+00 5.66854440e-02
-3.67767513e-01 4.14813519e-01 -4.62835163e-01 1.77865028e-01
8.69888783e-01 -2.47013181e-01 -4.03178841e-01 -5.86688876e-01
2.50337183e-01 5.55386782e-01 -1.18633516e-01 -7.69421458e-01
2.26067737e-01 -5.73320746e-01 2.97373354e-01 -8.42306077e-01
4.41300154e-01 -8.37156475e-01 -2.81994820e-01 6.65953398e-01
-1.83526687e-02 4.05687749e-01 5.62823951e-01 7.50364661e-01
5.93090177e-01 -3.64293545e-01 5.79374909e-01 -1.27749398e-01
-3.45635086e-01 9.57120061e-01 -1.47874489e-01 3.03204268e-01
1.33880067e+00 1.48442779e-02 -4.24574643e-01 3.08222212e-02
-5.92512369e-01 7.23275900e-01 6.15777373e-01 3.43371600e-01
4.52989668e-01 -9.15509522e-01 -4.35497105e-01 4.10031259e-01
-3.52752894e-01 7.12513745e-01 2.49024153e-01 3.38968813e-01
-7.53210306e-01 6.76757395e-01 -1.34514365e-02 -5.88148415e-01
-6.76079690e-01 1.12292135e+00 1.57328323e-01 -8.18290710e-01
-9.08468306e-01 7.78707385e-01 -7.92290419e-02 -2.88446903e-01
2.42534518e-01 -6.46067619e-01 4.42126423e-01 -5.27698278e-01
4.07372802e-01 1.04309738e-01 2.86905408e-01 1.91641703e-01
-1.75857484e-01 1.12221986e-02 -4.13372248e-01 2.62928903e-01
1.68158174e+00 3.70182276e-01 2.65480548e-01 -1.26679808e-01
7.92039633e-01 -2.29363322e-01 -1.84936178e+00 -1.20266505e-01
-1.35899618e-01 -6.30586207e-01 -3.76681387e-01 -6.30600989e-01
-8.05636108e-01 5.85492253e-01 9.42530409e-02 3.12178910e-01
6.16157591e-01 1.59460738e-01 1.01156557e+00 6.46909714e-01
1.02092302e+00 -1.09120142e+00 1.88352302e-01 8.53288829e-01
6.78510129e-01 -1.13577878e+00 1.77666903e-01 -3.43504936e-01
-5.77083707e-01 1.07273614e+00 8.17776918e-01 -6.12548828e-01
4.84152615e-01 5.33893645e-01 -6.85924113e-01 -1.04398057e-01
-1.19657111e+00 -1.59396142e-01 5.22005837e-03 2.53964782e-01
-2.92451262e-01 3.26741368e-01 1.01136200e-01 3.32118034e-01
-5.29319644e-01 -5.27435839e-01 2.72093922e-01 1.14270222e+00
-7.17442513e-01 -1.09564948e+00 -2.40819082e-01 5.68028927e-01
-5.79641899e-03 1.51535556e-01 5.07920802e-01 5.70347965e-01
-1.01224162e-01 6.62298977e-01 5.27992606e-01 1.21550664e-01
3.10525715e-01 -2.05232978e-01 9.56007004e-01 -9.05868471e-01
-2.99269825e-01 -4.22658995e-02 -4.76927944e-02 -6.36353552e-01
1.13231242e-01 -8.24213505e-01 -1.48765337e+00 -6.45725250e-01
-1.76049069e-01 2.51978546e-01 5.98653018e-01 1.17938447e+00
5.86497009e-01 3.44902307e-01 7.89557695e-01 -1.47839379e+00
-1.08502460e+00 -3.59488964e-01 -4.23961133e-01 1.49528593e-01
1.42045200e-01 -9.74927485e-01 2.53960509e-02 -5.30568004e-01] | [4.981522083282471, 2.6872220039367676] |
fae34351-c4db-432c-9c0c-90b689088b9e | glocal-glocalized-curriculum-aided-learning | 2204.06835 | null | https://arxiv.org/abs/2204.06835v1 | https://arxiv.org/pdf/2204.06835v1.pdf | GloCAL: Glocalized Curriculum-Aided Learning of Multiple Tasks with Application to Robotic Grasping | The domain of robotics is challenging to apply deep reinforcement learning due to the need for large amounts of data and for ensuring safety during learning. Curriculum learning has shown good performance in terms of sample- efficient deep learning. In this paper, we propose an algorithm (named GloCAL) that creates a curriculum for an agent to learn multiple discrete tasks, based on clustering tasks according to their evaluation scores. From the highest-performing cluster, a global task representative of the cluster is identified for learning a global policy that transfers to subsequently formed new clusters, while the remaining tasks in the cluster are learned as local policies. The efficacy and efficiency of our GloCAL algorithm are compared with other approaches in the domain of grasp learning for 49 objects with varied object complexity and grasp difficulty from the EGAD! dataset. The results show that GloCAL is able to learn to grasp 100% of the objects, whereas other approaches achieve at most 86% despite being given 1.5 times longer training time. | ['Keng Peng Tee', 'Domenico Campolo', 'Cihan Acar', 'Anil Kurkcu'] | 2022-04-14 | null | null | null | null | ['robotic-grasping'] | ['robots'] | [-1.15198798e-01 6.04212806e-02 -6.91314414e-02 -1.87833428e-01
-7.29802787e-01 -6.88194752e-01 2.58942902e-01 4.34944242e-01
-6.00854456e-01 7.32420027e-01 -3.38369161e-01 1.08569197e-01
-5.52467167e-01 -6.22470140e-01 -1.01541746e+00 -1.14775074e+00
-5.31019330e-01 9.44233537e-01 2.11311519e-01 1.90113634e-01
2.39922732e-01 5.23526430e-01 -1.82293689e+00 2.60418266e-01
1.00421464e+00 8.81751776e-01 9.45819080e-01 5.52226007e-01
2.62974381e-01 6.32519782e-01 -6.99787080e-01 2.33158037e-01
2.75868714e-01 -1.71693146e-01 -8.24549019e-01 1.88948940e-02
3.19400758e-01 -5.05373120e-01 1.01366721e-01 7.99815416e-01
5.96032500e-01 3.11540008e-01 7.31982350e-01 -1.36399603e+00
-7.42916018e-02 7.24894166e-01 -2.08728790e-01 -3.11922312e-01
5.98564334e-02 2.18229607e-01 8.25008214e-01 -5.51463485e-01
4.70508724e-01 1.30266023e+00 1.30101487e-01 8.36375177e-01
-1.30740464e+00 -7.58915782e-01 1.91685170e-01 7.35484213e-02
-8.54311883e-01 -3.40727270e-02 4.30252820e-01 -5.95871270e-01
8.73975813e-01 -3.40501994e-01 7.21031666e-01 1.11888719e+00
3.03256899e-01 1.01124859e+00 1.15538538e+00 -3.04349214e-01
9.01936173e-01 -1.95202958e-02 -5.58964834e-02 7.26646721e-01
1.73265219e-01 3.91086042e-01 -4.06206787e-01 2.84667499e-02
6.03449047e-01 -1.41130894e-01 1.60359278e-01 -1.06324804e+00
-9.93422687e-01 9.48349953e-01 7.47734249e-01 4.19577062e-02
-7.30393469e-01 5.51306427e-01 3.98205400e-01 2.80689955e-01
6.88551664e-02 7.88847446e-01 -6.44941986e-01 -9.14652124e-02
-4.49162453e-01 7.23945558e-01 8.74228120e-01 1.02458048e+00
8.26608598e-01 -2.80857813e-02 -1.47447690e-01 5.42000413e-01
1.56815708e-01 2.38517612e-01 2.50908256e-01 -1.25814867e+00
2.73646176e-01 6.70029461e-01 2.43865848e-01 -5.56156337e-01
-4.49899197e-01 -3.00901294e-01 -2.80657619e-01 9.14762914e-01
3.89404386e-01 -4.35291678e-01 -1.09590745e+00 1.67429829e+00
4.15088892e-01 -1.13414362e-01 2.86043763e-01 8.45513940e-01
3.34374100e-01 6.09946251e-01 4.01317984e-01 1.39268516e-02
8.91458333e-01 -8.85747194e-01 -3.85687768e-01 -1.22730106e-01
5.11401355e-01 -3.28413486e-01 9.51109290e-01 7.40980327e-01
-1.06802511e+00 -7.04858661e-01 -9.38132942e-01 4.54802573e-01
-3.37779343e-01 2.09982291e-01 8.00543964e-01 3.02779287e-01
-1.07100666e+00 8.69344890e-01 -1.10380960e+00 -2.78276414e-01
6.83425903e-01 7.90423155e-01 1.71394367e-02 -2.16075435e-01
-5.37845314e-01 9.95058656e-01 9.37652886e-01 -5.12376189e-01
-2.10577393e+00 -7.34740794e-01 -6.68504298e-01 3.20481211e-01
4.85559255e-01 -2.46423826e-01 1.29470670e+00 -6.12087190e-01
-1.52435672e+00 3.84391189e-01 4.85830933e-01 -5.04096270e-01
5.32092631e-01 -4.04410183e-01 5.21235883e-01 2.48210907e-01
-2.91903894e-02 1.39063835e+00 9.31296289e-01 -1.69859970e+00
-1.00860512e+00 -2.34759003e-01 1.38775721e-01 4.26934987e-01
-3.32470953e-01 -1.82347238e-01 -1.55266374e-01 -2.11623326e-01
-3.29450846e-01 -1.08250952e+00 -2.52338946e-01 -2.56460190e-01
1.99322794e-02 -8.70334566e-01 1.05420315e+00 -2.60606855e-01
5.77213287e-01 -2.01427794e+00 7.91497052e-01 9.92330462e-02
1.74685583e-01 1.51458040e-01 -2.01070294e-01 3.38557094e-01
7.83492774e-02 -2.73905903e-01 -5.42037785e-02 -2.47317702e-01
1.45452425e-01 4.01134014e-01 -8.86738375e-02 2.02388465e-01
1.61930546e-01 6.41027629e-01 -1.27693033e+00 -3.33306342e-01
4.55084741e-01 8.82665738e-02 -6.93314552e-01 5.92345774e-01
-6.97042167e-01 5.39776504e-01 -5.75118899e-01 5.18480241e-01
3.77041399e-01 5.11030033e-02 3.37786704e-01 2.43405715e-01
-9.79024265e-03 -4.87632025e-03 -8.88348579e-01 1.79591930e+00
-3.22612643e-01 2.51694232e-01 3.27850074e-01 -1.20828080e+00
1.01484776e+00 3.11173022e-01 7.88473964e-01 -4.28656131e-01
8.91372114e-02 2.12375209e-01 2.19315320e-01 -5.85437775e-01
1.33023068e-01 1.38631999e-01 -1.99254543e-01 3.65305722e-01
5.55076122e-01 -5.03217280e-01 3.72469753e-01 4.78134938e-02
1.16482306e+00 4.30743247e-01 -2.91199476e-01 -6.35221183e-01
4.00302140e-03 2.58977473e-01 2.44653344e-01 7.10438251e-01
-2.68294036e-01 -8.68164748e-02 4.45366949e-01 -5.71259081e-01
-1.02894568e+00 -1.19107175e+00 2.15211675e-01 1.31712222e+00
1.18224300e-01 -2.03861669e-01 -9.88402367e-01 -8.49650621e-01
3.34937513e-01 8.23752284e-01 -6.53248429e-01 -3.35390121e-01
-6.18694127e-01 -3.69019985e-01 1.49424240e-01 4.98777121e-01
3.54894668e-01 -1.81320751e+00 -1.37506747e+00 3.31975639e-01
2.21739709e-01 -7.07137883e-01 2.25908253e-02 7.50205934e-01
-8.78924787e-01 -1.24269950e+00 -7.17501760e-01 -1.06980884e+00
7.93923676e-01 -4.91812043e-02 8.69473219e-01 8.48795194e-03
-4.24442500e-01 4.98042762e-01 -4.37032908e-01 -7.55592942e-01
-3.89466584e-01 2.34063596e-01 2.78522730e-01 -4.72096503e-01
2.41233245e-01 -3.25121969e-01 -3.91876966e-01 2.06180245e-01
-6.54159546e-01 -2.22522035e-01 6.88732564e-01 8.45735371e-01
3.22511315e-01 3.16045910e-01 8.41685832e-01 -3.44305336e-01
8.22277606e-01 -3.96204382e-01 -9.55443740e-01 2.65511453e-01
-4.63590503e-01 1.19536690e-01 5.35982013e-01 -6.43278837e-01
-8.53541732e-01 4.45548654e-01 4.97500390e-01 -6.82160974e-01
-5.04173696e-01 2.61557549e-01 1.12273023e-01 -3.80818397e-02
5.75792015e-01 -7.01787397e-02 2.15279683e-01 -3.77513498e-01
1.33886233e-01 1.60091832e-01 2.50623107e-01 -1.23086321e+00
5.57227015e-01 2.28161947e-03 -1.94155127e-01 -4.67494816e-01
-7.23006427e-01 -1.85517862e-01 -6.37804449e-01 -6.02992535e-01
1.00460458e+00 -7.67413259e-01 -1.26131010e+00 4.80057776e-01
-8.59841347e-01 -9.92258966e-01 -3.51842582e-01 6.67994261e-01
-1.05614245e+00 -2.23793343e-01 -5.83537936e-01 -8.69374931e-01
-1.18573420e-01 -1.59796822e+00 9.92839754e-01 3.94818902e-01
-5.20270132e-02 -6.80833638e-01 -5.82193732e-02 6.85279295e-02
2.76903868e-01 4.17772502e-01 1.11353314e+00 -4.44957137e-01
-6.84949934e-01 2.69560337e-01 1.56607226e-01 3.27743322e-01
2.15516031e-01 -2.34781533e-01 -7.54915833e-01 -1.00890088e+00
-2.32738137e-01 -1.17170358e+00 6.56105340e-01 5.16003132e-01
1.45305121e+00 -2.55201552e-02 -4.48094904e-01 2.13097617e-01
1.48075318e+00 7.01753557e-01 2.91177571e-01 3.95290256e-01
4.54545468e-01 8.21438432e-01 9.67759609e-01 3.66170526e-01
-6.48323633e-03 3.59215528e-01 9.21606898e-01 1.63518980e-01
3.08796167e-02 -7.19506666e-02 5.19107103e-01 4.33002084e-01
5.12470817e-03 -7.31311366e-02 -9.56523657e-01 7.03559697e-01
-1.99725783e+00 -6.20877028e-01 5.05274713e-01 1.95371592e+00
6.99290931e-01 1.72449172e-01 2.48520404e-01 -3.83703671e-02
5.36177278e-01 -4.60378438e-01 -8.11743498e-01 -3.99132401e-01
5.65630138e-01 5.12468755e-01 2.80184627e-01 3.18223536e-01
-1.19225264e+00 1.30453956e+00 6.44999981e+00 8.11383069e-01
-8.06743205e-01 -2.33075216e-01 3.49946082e-01 -1.55871898e-01
4.00145084e-01 -2.56236732e-01 -7.28142858e-01 3.40185970e-01
7.86127508e-01 1.51441537e-03 8.84910762e-01 1.15626431e+00
-8.22066069e-02 -4.13992286e-01 -1.49574900e+00 5.28186202e-01
-1.65577397e-01 -8.67726207e-01 -7.25105032e-02 2.18294207e-02
9.79623437e-01 7.74545074e-02 6.06130436e-02 6.59122288e-01
1.05189979e+00 -1.14239728e+00 6.74384594e-01 1.80723771e-01
4.22276020e-01 -1.16228044e+00 5.22543609e-01 6.53950572e-01
-8.79071295e-01 -5.75943232e-01 -5.56710005e-01 -7.80676084e-04
-4.40976202e-01 -1.01415440e-01 -1.20513177e+00 4.71383780e-01
1.15118980e+00 4.44694281e-01 -2.86270171e-01 1.10343921e+00
-3.02032560e-01 4.53130484e-01 -1.71610579e-01 -4.06440347e-01
6.66934669e-01 -1.56534076e-01 2.61811644e-01 8.43376875e-01
2.55704582e-01 -6.07047509e-03 8.39966238e-01 8.57419312e-01
4.97372001e-02 -2.67020404e-01 -6.71828151e-01 4.96521182e-02
5.57819664e-01 1.18953192e+00 -6.62174165e-01 -2.90396005e-01
1.83207497e-01 5.58446527e-01 7.55752861e-01 2.03486115e-01
-6.00974441e-01 -2.44396091e-01 5.37112713e-01 -4.07681376e-01
4.79817182e-01 -3.77065450e-01 -1.57917574e-01 -4.18221533e-01
-2.55958200e-01 -1.01437771e+00 4.33111668e-01 -6.33523762e-01
-1.09408045e+00 2.79588580e-01 2.17111081e-01 -1.13425028e+00
-2.18888715e-01 -8.21784019e-01 -4.97723937e-01 5.73196113e-01
-1.23203588e+00 -8.39476109e-01 -5.10014772e-01 6.41736150e-01
6.51025414e-01 -5.61034381e-01 1.02157915e+00 -3.81810367e-02
-3.28566998e-01 2.63647288e-01 1.24444731e-01 -1.54057488e-01
5.35717964e-01 -1.56684148e+00 -3.14535022e-01 1.35684371e-01
-3.53479654e-01 2.72597164e-01 5.61758876e-01 -6.21185124e-01
-1.34408224e+00 -1.06912851e+00 1.17799737e-01 -3.93997580e-01
3.51964861e-01 -4.66146201e-01 -7.36344337e-01 6.09490454e-01
4.36045676e-01 -5.17677724e-01 2.46793181e-01 -4.80349213e-02
1.65662333e-01 -2.19489872e-01 -1.28009176e+00 3.81337404e-01
7.48811066e-01 9.77799967e-02 -6.46096468e-01 3.73339504e-01
7.76106536e-01 -5.38127601e-01 -9.93671596e-01 4.44062531e-01
2.63803244e-01 -6.02990270e-01 8.51610839e-01 -7.91710675e-01
6.06511176e-01 -9.11959261e-02 -5.55281760e-04 -1.83112574e+00
-3.75494033e-01 -3.03662837e-01 -1.48349568e-01 7.70247340e-01
4.86065894e-02 -2.16077030e-01 9.52193499e-01 2.35847086e-01
-3.51246119e-01 -7.20120430e-01 -7.72635460e-01 -8.23800981e-01
4.49788392e-01 1.56469345e-01 3.14393908e-01 9.14853632e-01
1.26477569e-01 7.20069036e-02 -1.35357425e-01 2.02056572e-01
1.00523686e+00 2.99395263e-01 7.43727863e-01 -1.34848583e+00
-1.62589952e-01 -3.14682990e-01 7.57081956e-02 -6.45744920e-01
4.05675143e-01 -9.31605101e-01 5.91985524e-01 -1.58012831e+00
1.83128759e-01 -9.55448151e-01 -3.86302799e-01 8.53999138e-01
-7.41468603e-03 -5.46691298e-01 4.04729247e-01 -7.09746219e-03
-7.52453864e-01 5.74314773e-01 1.51343930e+00 -2.54545063e-01
-3.47264051e-01 -8.19648989e-03 -2.02602029e-01 5.42118669e-01
1.07277799e+00 -5.57240129e-01 -4.81149256e-01 -5.55962861e-01
-3.30756783e-01 7.85519481e-02 2.74505764e-01 -1.34278655e+00
1.55296132e-01 -2.58369595e-01 6.72389030e-01 -6.76428974e-01
3.55021894e-01 -1.09949517e+00 -2.38720626e-01 1.14225781e+00
-5.75208724e-01 -2.79579237e-02 4.00216937e-01 5.25915027e-01
8.48088339e-02 -4.40251708e-01 9.11277175e-01 -2.66848356e-01
-7.42595077e-01 1.71169326e-01 -5.74510276e-01 -3.22328091e-01
1.59944963e+00 2.18265384e-01 -2.62972444e-01 -1.40475899e-01
-9.46664691e-01 8.30104649e-01 3.09111953e-01 6.55076623e-01
6.74221218e-01 -1.13444614e+00 -5.28300762e-01 4.28725854e-02
-1.25759572e-01 4.72769946e-01 1.82675198e-02 1.89992666e-01
-3.50634336e-01 3.33330274e-01 -7.53512800e-01 -9.42759931e-01
-1.11145425e+00 7.84467340e-01 2.79914618e-01 -2.83654362e-01
-5.36837816e-01 6.98413670e-01 2.51570255e-01 -6.16494238e-01
9.70927596e-01 -3.85700256e-01 -1.72169432e-01 -7.46752992e-02
1.73350185e-01 4.48611677e-01 1.33350696e-02 2.75973171e-01
-1.31056324e-01 3.84725213e-01 -2.30458438e-01 1.04938403e-01
1.69493711e+00 6.48285866e-01 4.28345054e-02 -1.17411688e-02
8.63689840e-01 -7.39068866e-01 -2.00832796e+00 9.06952024e-02
2.71859199e-01 -2.51126021e-01 -1.16169855e-01 -1.02137697e+00
-1.06627989e+00 8.46630335e-01 8.29685271e-01 2.81560402e-02
9.51761544e-01 1.93781808e-01 2.28932202e-01 7.91947007e-01
8.11844468e-01 -1.41004276e+00 8.88543129e-01 6.58500195e-01
9.67342138e-01 -1.09541011e+00 -1.86377585e-01 8.37518740e-03
-6.73035264e-01 1.03000796e+00 1.21711051e+00 -5.14269292e-01
2.88190931e-01 3.10369700e-01 -1.73102006e-01 -3.84747446e-01
-8.10530543e-01 -4.02368456e-02 -1.43431246e-01 9.31204259e-01
-1.84359059e-01 2.22857088e-01 5.72080128e-02 4.64047939e-02
-7.43951201e-02 -2.27303326e-01 2.79529363e-01 1.23800826e+00
-8.05560887e-01 -1.03932047e+00 -2.70263314e-01 3.77940506e-01
-5.14731854e-02 5.22969663e-01 -2.97861457e-01 7.96775222e-01
1.75378799e-01 8.20850611e-01 1.99989632e-01 -2.49587357e-01
2.35102907e-01 5.85912913e-03 8.00158620e-01 -7.38147318e-01
-7.16736019e-01 -1.14140911e-02 -9.73594040e-02 -6.70377135e-01
-3.66468638e-01 -6.73798203e-01 -1.55938458e+00 -2.00539157e-02
-9.76286903e-02 2.32555270e-01 8.70425940e-01 7.00590193e-01
1.81735918e-01 8.65669489e-01 6.42175198e-01 -1.32984769e+00
-8.79611731e-01 -8.96416485e-01 -4.18175370e-01 4.09148902e-01
1.40610814e-01 -1.13119984e+00 -8.28572810e-02 -4.81650569e-02] | [4.264570236206055, 1.230695366859436] |
3f07324e-846f-409d-9eed-7474500e35f5 | shas-approaching-optimal-segmentation-for-end | 2202.04774 | null | https://arxiv.org/abs/2202.04774v3 | https://arxiv.org/pdf/2202.04774v3.pdf | SHAS: Approaching optimal Segmentation for End-to-End Speech Translation | Speech translation models are unable to directly process long audios, like TED talks, which have to be split into shorter segments. Speech translation datasets provide manual segmentations of the audios, which are not available in real-world scenarios, and existing segmentation methods usually significantly reduce translation quality at inference time. To bridge the gap between the manual segmentation of training and the automatic one at inference, we propose Supervised Hybrid Audio Segmentation (SHAS), a method that can effectively learn the optimal segmentation from any manually segmented speech corpus. First, we train a classifier to identify the included frames in a segmentation, using speech representations from a pre-trained wav2vec 2.0. The optimal splitting points are then found by a probabilistic Divide-and-Conquer algorithm that progressively splits at the frame of lowest probability until all segments are below a pre-specified length. Experiments on MuST-C and mTEDx show that the translation of the segments produced by our method approaches the quality of the manual segmentation on 5 language pairs. Namely, SHAS retains 95-98% of the manual segmentation's BLEU score, compared to the 87-93% of the best existing methods. Our method is additionally generalizable to different domains and achieves high zero-shot performance in unseen languages. | ['Marta R. Costa-jussà', 'José A. R. Fonollosa', 'Gerard I. Gállego', 'Ioannis Tsiamas'] | 2022-02-09 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 3.77129287e-01 2.93116480e-01 -3.19171190e-01 -4.92935538e-01
-1.60984635e+00 -7.43747294e-01 3.17321956e-01 -9.08883382e-03
-3.60714704e-01 5.53337693e-01 3.01779330e-01 -5.09434044e-01
4.15147156e-01 -4.79627877e-01 -6.58588588e-01 -6.28877938e-01
4.66898918e-01 9.80273902e-01 5.06196260e-01 -9.62752849e-02
-1.44986317e-01 -7.87622184e-02 -1.29633236e+00 4.67566669e-01
9.65060949e-01 9.15127814e-01 2.91234910e-01 7.15706110e-01
-5.46772838e-01 3.76334906e-01 -8.86797845e-01 -4.09476876e-01
2.20267877e-01 -8.98451626e-01 -1.23437047e+00 3.69259268e-01
1.52292311e-01 -1.81479886e-01 -5.98586276e-02 1.16726422e+00
4.93027329e-01 1.11481540e-01 4.85250622e-01 -1.01178443e+00
-2.20176056e-01 1.39596725e+00 -4.35045481e-01 1.79491520e-01
4.31292593e-01 -1.28979430e-01 1.18031967e+00 -8.41767430e-01
6.39665365e-01 1.35531080e+00 3.73837620e-01 5.03926933e-01
-1.40123737e+00 -7.76312590e-01 -1.10710829e-01 1.13304332e-01
-1.58672655e+00 -7.85791159e-01 6.79476619e-01 -4.34641838e-01
8.10198843e-01 3.72193843e-01 5.42225718e-01 1.07755196e+00
-1.94564566e-01 9.47439075e-01 9.34292316e-01 -6.71753168e-01
3.13142866e-01 -4.80577648e-02 6.50958717e-03 3.54287773e-01
-4.85521883e-01 -3.26928765e-01 -7.90187359e-01 8.21055472e-02
3.89705658e-01 -5.65970004e-01 -4.79898930e-01 2.20706537e-01
-1.44030559e+00 6.75102115e-01 -1.13756597e-01 4.61996078e-01
-2.75959373e-01 -4.35250849e-02 4.32342827e-01 3.82685900e-01
5.98618090e-01 2.99620003e-01 -4.82310832e-01 -5.78857183e-01
-1.62354553e+00 -2.93093100e-02 7.83322752e-01 1.23204279e+00
8.57496798e-01 8.98294598e-02 -3.38915646e-01 9.93623614e-01
2.38498658e-01 5.68033516e-01 7.01662242e-01 -1.07774830e+00
6.76510274e-01 2.39913508e-01 -4.41515706e-02 -3.83628935e-01
1.12778872e-01 -3.74852300e-01 -5.99145889e-01 -4.14841950e-01
4.34251577e-01 -3.02542210e-01 -1.05003512e+00 1.58367491e+00
4.12670970e-01 9.73381326e-02 7.65668452e-02 8.50669742e-01
5.74611247e-01 1.21257234e+00 -2.25308165e-01 -5.84170222e-01
1.34759283e+00 -1.17302322e+00 -1.03327060e+00 -3.25060338e-01
5.27321696e-01 -1.25496817e+00 1.30054414e+00 4.55500901e-01
-1.14483738e+00 -6.37378275e-01 -9.24756587e-01 1.24679143e-02
7.31226280e-02 3.09698582e-01 2.88360566e-02 7.10587859e-01
-1.01264620e+00 4.04670000e-01 -9.90489185e-01 -1.72462612e-01
1.37184501e-01 2.14414075e-01 4.04215753e-02 1.77134708e-01
-1.18970001e+00 4.28692251e-01 4.77928042e-01 -8.86656567e-02
-8.71432126e-01 -4.83767867e-01 -8.76523077e-01 1.21419042e-01
6.13219798e-01 -2.07249641e-01 1.85287142e+00 -1.30996430e+00
-1.96232748e+00 7.47652769e-01 -5.31627119e-01 -6.00766003e-01
4.23493773e-01 -2.47619569e-01 -3.06366622e-01 4.41397786e-01
2.32510492e-01 1.01781261e+00 9.11319971e-01 -1.13870978e+00
-8.57855260e-01 -1.75264981e-02 -2.42784962e-01 3.41668129e-01
-1.33872107e-01 1.33901179e-01 -9.92667317e-01 -7.76203394e-01
3.49881500e-01 -1.04197037e+00 -4.78047356e-02 -5.44234812e-01
-6.39098048e-01 -2.14377254e-01 8.61973703e-01 -1.02799594e+00
1.39795625e+00 -2.15241456e+00 2.96851993e-01 8.19425657e-02
-1.21663243e-01 2.37012401e-01 -6.96385428e-02 2.47847050e-01
2.18705356e-01 6.85203075e-02 -2.77945936e-01 -6.46284938e-01
5.02066724e-02 3.70927185e-01 -4.99164522e-01 9.96750444e-02
-8.98693949e-02 6.26972973e-01 -9.25040364e-01 -7.61315763e-01
1.17074400e-02 3.03221315e-01 -5.39225578e-01 4.04210240e-01
-2.19385624e-01 4.12057906e-01 -8.14400911e-02 5.12271702e-01
3.38855982e-01 1.90299377e-01 1.76133454e-01 5.26643707e-04
-7.79165402e-02 9.03373122e-01 -1.12467468e+00 1.87773228e+00
-3.49131435e-01 8.12488019e-01 9.72667113e-02 -1.00881600e+00
9.74411428e-01 7.14400291e-01 3.89579475e-01 -3.83808792e-01
1.74099579e-01 3.57582122e-01 7.83242192e-03 -3.28412682e-01
5.05265117e-01 -1.80764124e-01 -2.64308959e-01 4.37489241e-01
4.01774824e-01 -6.47251546e-01 3.33827645e-01 -3.44120637e-02
7.81202793e-01 4.73990589e-02 1.22124381e-01 -1.68592166e-02
4.43684608e-01 1.10504135e-01 8.71331453e-01 6.05754316e-01
-3.24599445e-01 1.00479078e+00 4.97469872e-01 2.76350360e-02
-9.55461562e-01 -9.75384593e-01 2.06789806e-01 1.24198425e+00
3.92873473e-02 -5.22765994e-01 -1.45266652e+00 -6.82405770e-01
-7.01284707e-01 9.03920233e-01 -5.57062477e-02 -1.22347027e-02
-5.83237052e-01 -2.78711081e-01 7.53857315e-01 2.38844186e-01
3.50517392e-01 -9.97297287e-01 -2.78026998e-01 4.15407270e-01
-1.04920518e+00 -1.38406825e+00 -9.19182181e-01 2.38482207e-01
-5.80906689e-01 -6.07231855e-01 -5.78717589e-01 -9.99057353e-01
3.01779538e-01 1.42648280e-01 1.02621806e+00 -3.10276568e-01
2.52394050e-01 -8.21964517e-02 -6.39792085e-01 -1.95449561e-01
-9.19728816e-01 4.22347426e-01 3.24743614e-03 9.65098366e-02
4.72119629e-01 -3.33848268e-01 -1.92411900e-01 4.71564472e-01
-5.81059754e-01 2.15984493e-01 4.58324790e-01 7.74572253e-01
1.02416873e+00 2.10243165e-01 4.71430272e-01 -5.84175527e-01
4.28743243e-01 -1.29720196e-01 -4.51960683e-01 1.61241800e-01
-3.59636277e-01 -8.06795135e-02 5.70313156e-01 -5.80225945e-01
-9.43161011e-01 2.30095342e-01 -3.05529088e-01 -5.39495289e-01
-3.08558196e-01 4.91845906e-01 -3.66141915e-01 4.80557173e-01
4.37280059e-01 3.55602443e-01 -1.74703717e-01 -5.53387821e-01
3.20526361e-01 1.19681489e+00 7.44191647e-01 -5.63869655e-01
6.80019677e-01 -1.41181245e-01 -7.89154887e-01 -9.09387469e-01
-7.44907379e-01 -5.62808514e-01 -7.15556443e-01 -2.06680641e-01
9.79124069e-01 -8.85415792e-01 -1.15124464e-01 3.07349086e-01
-1.33326685e+00 -4.47473019e-01 -4.16243106e-01 6.74500704e-01
-7.43256688e-01 4.13432300e-01 -8.25526774e-01 -6.97928965e-01
-3.75843376e-01 -1.53301013e+00 1.27243435e+00 6.55827373e-02
-5.95155120e-01 -4.43701446e-01 -5.98487109e-02 7.53086150e-01
1.84797123e-02 -3.06990832e-01 8.08474958e-01 -9.29931283e-01
-4.69816774e-01 6.97485730e-02 3.33359450e-01 4.90011781e-01
1.15775868e-01 1.64055049e-01 -1.01516271e+00 -1.08235039e-01
-1.23997055e-01 -1.71146125e-01 5.98208964e-01 4.60875243e-01
8.77545118e-01 -4.29321349e-01 -1.47405252e-01 2.99191058e-01
7.67694771e-01 5.42580128e-01 5.35382390e-01 -9.69450176e-02
3.24394882e-01 4.87407625e-01 7.61686325e-01 1.80988371e-01
2.40457982e-01 8.30791175e-01 -2.52624694e-02 -7.93018658e-03
-2.72647709e-01 -5.19301653e-01 7.84215868e-01 1.67050612e+00
2.95052618e-01 -3.32257777e-01 -9.58038688e-01 7.17553556e-01
-1.55891097e+00 -8.05460751e-01 7.68719018e-02 2.16758943e+00
1.45155358e+00 4.82069582e-01 1.51721314e-01 5.84243357e-01
9.58519816e-01 1.02557272e-01 -5.97468093e-02 -5.10484755e-01
5.00262491e-02 3.31356704e-01 2.59571314e-01 8.79543781e-01
-1.13715994e+00 1.37986398e+00 6.31658173e+00 1.34274507e+00
-1.01376486e+00 3.71038377e-01 6.27395213e-01 2.01331102e-03
-3.53667110e-01 1.56463802e-01 -7.87844121e-01 4.57716495e-01
1.37826180e+00 -1.53954953e-01 6.49620771e-01 6.47101998e-01
6.12952948e-01 1.62935518e-02 -1.13401186e+00 8.71602237e-01
-1.25265092e-01 -1.14890015e+00 2.29567681e-02 -2.11496145e-01
6.46294892e-01 -3.45836356e-02 -1.22395992e-01 4.71654862e-01
2.56990552e-01 -7.24090993e-01 1.18869138e+00 9.00178403e-03
8.00614476e-01 -8.28935683e-01 7.19746172e-01 5.09295464e-01
-1.13275385e+00 3.03902686e-01 -1.88969344e-01 2.69817054e-01
4.18401688e-01 6.69135928e-01 -1.29732502e+00 4.48136538e-01
5.16136885e-01 1.67716160e-01 -1.54013038e-01 7.80068278e-01
-4.77487832e-01 1.31227529e+00 -3.44317466e-01 9.60685015e-02
3.72665942e-01 -3.66786003e-01 6.50461197e-01 1.44577801e+00
5.27110577e-01 -1.36174500e-01 4.86870617e-01 7.03628123e-01
-1.96695626e-01 2.71621257e-01 -2.40846649e-01 -3.27856481e-01
9.51357484e-01 7.51548052e-01 -1.00199533e+00 -6.03815973e-01
-2.83394754e-01 1.16849220e+00 -1.73955977e-01 4.62767929e-01
-9.59951639e-01 -6.43448949e-01 5.08514106e-01 -7.37688616e-02
4.61304426e-01 -1.97723880e-01 -2.53966540e-01 -9.04581130e-01
-6.20171912e-02 -1.23229158e+00 1.61913335e-01 -6.80393517e-01
-6.21657848e-01 8.40356708e-01 -1.58235461e-01 -1.40090621e+00
-5.50642133e-01 1.99610591e-02 -2.90854633e-01 6.98032677e-01
-1.06205070e+00 -9.23671186e-01 1.73951685e-01 2.93554902e-01
1.33952796e+00 -6.94071576e-02 7.97529817e-01 3.67941648e-01
-4.51164246e-01 7.07630277e-01 -3.52347991e-03 3.47727507e-01
7.68077195e-01 -1.16604853e+00 3.93171698e-01 1.11143589e+00
6.17749751e-01 2.25425333e-01 8.85318458e-01 -5.39453626e-01
-9.59478617e-01 -1.04102969e+00 1.37539530e+00 -1.06118098e-02
5.30936241e-01 -3.32807511e-01 -8.99373531e-01 7.05898583e-01
4.15604919e-01 -3.69606197e-01 7.05990434e-01 -8.38715956e-02
-2.23817885e-01 -1.96004435e-01 -8.45307648e-01 5.17223120e-01
7.10564137e-01 -6.98888421e-01 -9.42271888e-01 1.53888211e-01
1.25169981e+00 -6.45927846e-01 -6.34119213e-01 1.06494449e-01
2.85616130e-01 -7.03520119e-01 6.11837387e-01 -2.41255164e-01
3.34801853e-01 -3.60456586e-01 -3.99588406e-01 -1.42491388e+00
2.00650781e-01 -1.20298040e+00 1.74513534e-01 1.62830782e+00
7.01276779e-01 -6.89738691e-02 4.77021426e-01 2.09537074e-02
-4.58391339e-01 -4.58761036e-01 -1.26561368e+00 -7.65785933e-01
-1.28678799e-01 -6.85980320e-01 6.68504477e-01 7.93038249e-01
8.08668882e-02 5.80418706e-01 -1.72622278e-01 2.77379781e-01
3.98152024e-01 1.69596195e-01 6.43593729e-01 -8.85212243e-01
-3.22504967e-01 -4.37847853e-01 -2.48719022e-01 -1.40798271e+00
3.18472624e-01 -8.32073152e-01 5.68738401e-01 -1.30833566e+00
-7.83437863e-02 -1.67978987e-01 6.21172599e-03 5.57179987e-01
-5.53561002e-02 1.85573816e-01 1.33734718e-01 1.56854093e-01
-4.75504220e-01 6.13501549e-01 1.12467253e+00 -5.43905437e-01
-5.89757919e-01 3.01398188e-01 -3.32545727e-01 7.32861102e-01
6.02497220e-01 -7.05770135e-01 -4.88165647e-01 -4.24351335e-01
-3.76086771e-01 4.00793761e-01 -4.45931077e-01 -9.85392392e-01
2.24440604e-01 -1.95927039e-01 -2.47032866e-01 -9.07044351e-01
2.70204276e-01 -6.21409476e-01 2.01175585e-01 2.06437334e-01
-5.39522946e-01 -1.78939968e-01 -5.54544292e-02 3.27668011e-01
-5.89496136e-01 -3.67362022e-01 7.92820394e-01 1.03413701e-01
-2.80579269e-01 6.39156029e-02 -8.30337346e-01 2.01418042e-01
6.89455688e-01 -2.95311753e-02 1.97302356e-01 -5.78171730e-01
-8.36851954e-01 2.07317490e-02 1.15997285e-01 4.29921150e-01
3.67411852e-01 -1.36308217e+00 -6.16814017e-01 1.91147491e-01
-1.47594362e-01 2.23024055e-01 -2.57027447e-01 8.56036484e-01
-4.46218073e-01 3.07751000e-01 2.80643106e-01 -8.40764999e-01
-1.29390275e+00 1.97850272e-01 1.69710338e-01 -1.33282691e-01
-5.56648314e-01 8.00812364e-01 -8.56584087e-02 -2.72617251e-01
5.60059249e-01 -6.79578602e-01 6.42513856e-02 2.53923297e-01
4.30708736e-01 2.32453063e-01 2.71980107e-01 -1.00124109e+00
-1.43589675e-01 2.72620499e-01 6.99892268e-02 -7.51699269e-01
1.10016286e+00 -5.32680094e-01 1.07347397e-02 7.50271261e-01
1.10416067e+00 1.69986472e-01 -1.10879421e+00 -2.90117711e-01
4.09403816e-02 -3.60728145e-01 9.32477936e-02 -5.40763259e-01
-9.65667546e-01 1.05281675e+00 2.65297383e-01 3.22428554e-01
1.09254456e+00 1.28095239e-01 1.22003508e+00 2.65027612e-01
2.46066839e-01 -1.47954965e+00 -5.86221404e-02 6.77808404e-01
6.88038349e-01 -8.35019886e-01 -5.42119026e-01 -5.49975872e-01
-7.66559958e-01 1.04944158e+00 3.28665346e-01 4.71801996e-01
2.78957397e-01 4.11306858e-01 4.00088608e-01 3.21750790e-01
-6.44901931e-01 -2.51633495e-01 1.79008767e-01 5.83921075e-01
3.66475552e-01 3.58220965e-01 -3.61802608e-01 8.33687305e-01
-7.34654486e-01 -2.61132509e-01 5.75714588e-01 6.66416705e-01
-7.46805966e-01 -1.14481878e+00 -4.96069461e-01 -3.51918228e-02
-5.83886325e-01 -5.71345240e-02 -4.99868840e-01 3.36149067e-01
1.18432514e-01 1.36025393e+00 1.62230805e-01 -4.45283532e-01
2.78097421e-01 6.06243908e-01 2.02304602e-01 -7.25529552e-01
-3.78720939e-01 9.22979593e-01 2.21119508e-01 -4.62095171e-01
-3.55116189e-01 -6.94665968e-01 -1.37361097e+00 -2.83109546e-02
-4.76614088e-01 4.17031646e-01 4.87049490e-01 1.12123370e+00
1.42356986e-02 5.00770807e-01 7.14540660e-01 -8.70743752e-01
-5.26297152e-01 -1.06995666e+00 -3.37906718e-01 2.57801414e-01
3.41723680e-01 -3.27092975e-01 -4.01647121e-01 6.62991285e-01] | [14.589102745056152, 6.841160774230957] |
1c854101-4ad3-473d-8df1-04ee56dea067 | long-term-blood-pressure-prediction-with-deep | 1705.04524 | null | http://arxiv.org/abs/1705.04524v3 | http://arxiv.org/pdf/1705.04524v3.pdf | Long-term Blood Pressure Prediction with Deep Recurrent Neural Networks | Existing methods for arterial blood pressure (BP) estimation directly map the
input physiological signals to output BP values without explicitly modeling the
underlying temporal dependencies in BP dynamics. As a result, these models
suffer from accuracy decay over a long time and thus require frequent
calibration. In this work, we address this issue by formulating BP estimation
as a sequence prediction problem in which both the input and target are
temporal sequences. We propose a novel deep recurrent neural network (RNN)
consisting of multilayered Long Short-Term Memory (LSTM) networks, which are
incorporated with (1) a bidirectional structure to access larger-scale context
information of input sequence, and (2) residual connections to allow gradients
in deep RNN to propagate more effectively. The proposed deep RNN model was
tested on a static BP dataset, and it achieved root mean square error (RMSE) of
3.90 and 2.66 mmHg for systolic BP (SBP) and diastolic BP (DBP) prediction
respectively, surpassing the accuracy of traditional BP prediction models. On a
multi-day BP dataset, the deep RNN achieved RMSE of 3.84, 5.25, 5.80 and 5.81
mmHg for the 1st day, 2nd day, 4th day and 6th month after the 1st day SBP
prediction, and 1.80, 4.78, 5.0, 5.21 mmHg for corresponding DBP prediction,
respectively, which outperforms all previous models with notable improvement.
The experimental results suggest that modeling the temporal dependencies in BP
dynamics significantly improves the long-term BP prediction accuracy. | ['Yuan-Ting Zhang', 'Xiao-Rong Ding', 'Ni Zhao', 'Fen Miao', 'Peng Su', 'Jing Liu'] | 2017-05-12 | null | null | null | null | ['blood-pressure-estimation', 'photoplethysmography-ppg', 'electrocardiography-ecg'] | ['medical', 'medical', 'methodology'] | [-8.10487345e-02 -3.07954222e-01 6.54007792e-02 -6.92002177e-01
-5.08476608e-02 -1.79977745e-01 -2.34978385e-02 1.30129397e-01
-2.13012651e-01 1.28490090e+00 2.00258017e-01 -6.84090316e-01
-1.46284059e-01 -8.96678448e-01 -5.28887570e-01 -5.39931297e-01
-5.45683444e-01 1.64636280e-02 1.35921270e-01 -2.48433098e-01
2.05849394e-01 3.75283539e-01 -9.05786455e-01 9.41067114e-02
1.13452590e+00 1.07555497e+00 -1.70239851e-01 9.49157119e-01
-1.81489006e-01 1.04983246e+00 -4.70176518e-01 -1.32357761e-01
2.37188160e-01 -6.54973567e-01 -4.23355043e-01 -4.27682936e-01
3.76132399e-01 -5.37616968e-01 -5.22170722e-01 8.66115332e-01
9.35670495e-01 1.95028648e-01 7.66546652e-02 -6.45288885e-01
-8.09434712e-01 6.33472681e-01 -4.37669188e-01 6.03430986e-01
-3.83823425e-01 3.15151364e-01 4.32321995e-01 -5.14132738e-01
-1.22639112e-01 9.72555280e-01 8.32515121e-01 4.25154775e-01
-1.43575549e+00 -6.08613908e-01 2.25030571e-01 1.02875186e-02
-1.27776611e+00 -3.64322662e-01 4.72656965e-01 -6.87821269e-01
6.44448161e-01 4.07025427e-01 9.09269512e-01 8.86716425e-01
5.45100689e-01 2.75623295e-02 1.23760772e+00 -1.14545740e-01
-9.94000733e-02 2.66812265e-01 6.16990983e-01 4.36601847e-01
5.19666588e-03 4.08301771e-01 1.99959278e-01 -1.09037295e-01
1.11201680e+00 3.99956793e-01 -3.57355356e-01 3.49147886e-01
-8.63835156e-01 5.31371057e-01 7.62002885e-01 1.85051978e-01
-7.21275151e-01 -2.96440776e-02 4.64291930e-01 3.47036839e-01
4.23709899e-01 8.21274519e-02 -6.29768491e-01 -2.40410313e-01
-8.14339042e-01 2.76315659e-01 9.56091940e-01 5.11165321e-01
4.34594601e-01 2.96811342e-01 -6.76442385e-01 1.16883612e+00
3.58036458e-01 6.67574525e-01 5.10570109e-01 -7.61966765e-01
5.94985485e-01 3.12021136e-01 4.48009938e-01 -1.28920221e+00
-7.61892855e-01 -8.53066087e-01 -1.39243662e+00 -1.73513323e-01
4.20658141e-01 -5.42387843e-01 -1.16723061e+00 1.62725103e+00
-5.76721989e-02 4.51544523e-01 7.49846473e-02 9.79537368e-01
9.77804899e-01 8.58044684e-01 4.80626047e-01 -5.55288851e-01
1.14174879e+00 -9.78336573e-01 -6.90639734e-01 -3.23299468e-01
1.78858325e-01 -4.02356237e-01 7.78923273e-01 -4.26029712e-02
-9.14277077e-01 -1.04842389e+00 -8.67524147e-01 1.43723041e-01
-1.98067307e-01 1.07609816e-01 9.22353938e-02 4.87017989e-01
-1.00097501e+00 1.05512071e+00 -7.00417399e-01 -7.20924139e-02
1.12297080e-01 1.22231752e-01 2.98796326e-01 3.57952774e-01
-1.81033361e+00 9.89704132e-01 2.45353267e-01 9.85772491e-01
-4.07196105e-01 -1.24608815e+00 -4.48543817e-01 9.97180194e-02
-3.48101497e-01 -7.45893776e-01 1.11693728e+00 -3.35262030e-01
-1.85147262e+00 4.54819530e-01 -3.85519356e-01 -6.67407513e-01
6.62216365e-01 -5.09451926e-01 -8.64181340e-01 -2.04800978e-01
-2.45421499e-01 2.84540683e-01 3.98283899e-01 -6.22267306e-01
-3.70045781e-01 -5.55716872e-01 -4.25424933e-01 -6.71608746e-02
7.49578103e-02 -1.01882406e-01 8.61728936e-02 -3.86521161e-01
1.65136799e-01 -8.20863724e-01 -8.07893872e-01 -1.34366751e-01
-3.96572560e-01 3.01008731e-01 2.07434654e-01 -1.27544856e+00
1.62926936e+00 -1.73605847e+00 -1.52922601e-01 4.04304922e-01
4.43973571e-01 7.19801009e-01 2.92287290e-01 2.29388595e-01
-3.09008032e-01 2.42790535e-01 -9.14328545e-02 3.52880895e-01
-2.95161098e-01 2.29937136e-01 -4.24589008e-01 1.44238159e-01
5.37453778e-02 1.10542226e+00 -7.24903107e-01 -1.16357110e-01
2.97191918e-01 4.51272339e-01 -1.25263304e-01 2.74680018e-01
1.05607063e-01 8.64068925e-01 -2.70171314e-01 -3.10521312e-02
7.85321116e-01 -3.45051646e-01 7.90888742e-02 -1.24110490e-01
-5.10322928e-01 1.97652400e-01 -9.23744261e-01 1.00605190e+00
-3.24370146e-01 5.36711514e-01 -5.24170995e-01 -9.79325771e-01
1.73812985e+00 5.53744256e-01 5.16000688e-01 -1.08457661e+00
-1.43243641e-01 2.73596317e-01 5.56650877e-01 -7.63545871e-01
-1.56384051e-01 -3.61843228e-01 5.93595862e-01 8.20786580e-02
-6.21218026e-01 7.07577765e-01 3.08887102e-02 -4.06496644e-01
6.01570845e-01 -1.62596852e-01 1.27253041e-01 -3.53532314e-01
1.02485538e+00 -4.97137189e-01 8.08259130e-01 9.62188542e-01
-6.48694217e-01 3.16723257e-01 5.96313298e-01 -1.13628721e+00
-1.10341740e+00 -9.55404401e-01 -6.43343210e-01 6.70999765e-01
-1.31357089e-01 -9.89598483e-02 3.09931617e-02 7.33406842e-02
4.10968721e-01 4.69124943e-01 -4.62346524e-01 -2.73246855e-01
-1.06981528e+00 -1.09532738e+00 4.97309566e-01 7.51527786e-01
9.00200605e-01 -9.53730285e-01 -2.86949694e-01 8.86675835e-01
-3.42218727e-02 -8.22235584e-01 -3.66179533e-02 -1.32222176e-01
-1.32614017e+00 -6.37817681e-01 -9.12617445e-01 -5.40648341e-01
1.66339606e-01 -4.14226115e-01 1.04865980e+00 -1.69037700e-01
-2.27076113e-01 -4.71439183e-01 3.02725941e-01 -3.67214501e-01
-2.47391298e-01 -1.81558073e-01 7.33795688e-02 -7.36132357e-03
4.19803917e-01 -1.09035730e+00 -1.19285035e+00 3.00018728e-01
3.83850597e-02 1.12370327e-01 3.77178550e-01 7.66048372e-01
3.92146498e-01 -5.48135042e-01 1.01850665e+00 -6.84681952e-01
9.06540215e-01 -4.38445479e-01 -7.12191880e-01 1.65809155e-01
-1.00757992e+00 -2.45112672e-01 8.77322793e-01 -2.71003097e-01
-7.88511813e-01 -3.97921920e-01 -3.07531238e-01 -1.59906924e-01
8.17892775e-02 6.65643156e-01 6.34711325e-01 2.26790607e-01
9.03957844e-01 6.85288310e-01 1.24097332e-01 -6.55715168e-01
-1.64486486e-02 6.59965932e-01 4.58600163e-01 -5.69054544e-01
1.54451787e-01 -2.43773147e-01 2.17699736e-01 -7.44326055e-01
-9.44339514e-01 3.18339020e-02 -7.42049098e-01 -1.64466217e-01
6.92260385e-01 -9.16923642e-01 -1.11358368e+00 4.03409749e-01
-9.46644902e-01 -2.98712492e-01 8.03247020e-02 6.76193714e-01
-1.76091611e-01 2.27137923e-01 -1.06707263e+00 -8.84648919e-01
-1.03233218e+00 -6.94877505e-01 2.10398529e-02 3.83987188e-01
-5.13423234e-02 -1.00810814e+00 1.52652919e-01 -1.18188053e-01
1.07175148e+00 5.23578107e-01 9.34187710e-01 -3.92367154e-01
-1.85623959e-01 -2.63339996e-01 -6.33170366e-01 7.55583465e-01
3.33556831e-01 4.09176126e-02 -6.50631547e-01 -1.54808629e-02
9.13975760e-02 2.37187698e-01 5.54665208e-01 8.66410136e-01
1.21919715e+00 -4.87911493e-01 -8.01090598e-02 6.17211282e-01
1.40180266e+00 6.27654910e-01 9.99824286e-01 1.38427690e-01
5.74714303e-01 8.57162103e-02 -7.79649913e-02 5.44559002e-01
2.71119475e-01 2.88114190e-01 -9.40708444e-02 -2.05861598e-01
1.28183559e-01 -4.64537442e-02 -4.34616674e-03 9.72314417e-01
-5.06883740e-01 3.50938261e-01 -1.09746063e+00 1.36989087e-01
-1.84007251e+00 -9.97584403e-01 -6.28996789e-01 2.29026794e+00
1.13514173e+00 2.08594128e-01 1.74001470e-01 -2.51002163e-01
8.11134577e-01 2.00772658e-01 -7.31129587e-01 -7.62241662e-01
3.36884499e-01 2.44783178e-01 3.76334190e-01 6.96667075e-01
-1.14702797e+00 4.86943126e-01 5.96305418e+00 -4.05043513e-01
-1.44988537e+00 -3.24585974e-01 1.10603511e+00 1.54305443e-01
5.09000123e-01 -2.25978225e-01 -9.22740221e-01 1.03958082e+00
1.63318026e+00 -3.09787959e-01 3.24033618e-01 6.29852414e-01
9.12811756e-01 3.85764122e-01 -9.02968884e-01 1.00350726e+00
-6.82449996e-01 -1.17228436e+00 -1.58483833e-01 -2.61924207e-01
5.01383543e-01 1.79716825e-01 -1.32019922e-01 6.40190125e-01
-2.43031550e-02 -1.23434949e+00 -2.06492558e-01 1.20294011e+00
5.90685844e-01 -4.05017734e-01 8.88239920e-01 3.09233189e-01
-1.02857268e+00 -2.42904589e-01 -6.14918828e-01 -4.72508371e-01
2.02830926e-01 9.21836793e-01 -4.99187648e-01 2.61523128e-01
5.80482125e-01 9.14376140e-01 -4.23307121e-01 1.21891761e+00
-9.13641229e-02 7.52657771e-01 -1.25535160e-01 -1.67684764e-01
2.89348811e-01 -4.71583754e-01 2.71597773e-01 1.21354496e+00
2.15921655e-01 2.25443527e-01 3.41438502e-01 1.13694692e+00
2.23073035e-01 2.07231417e-01 -8.70132968e-02 3.02350253e-01
3.87853712e-01 7.25532115e-01 1.32046476e-01 -6.38986945e-01
-2.50337154e-01 4.82826084e-01 2.22752988e-01 6.98932409e-01
-1.11312497e+00 -6.91110075e-01 6.74477220e-01 5.03471075e-03
3.06686182e-02 2.49294210e-02 -6.15615547e-01 -1.03009045e+00
3.71622324e-01 -3.89521152e-01 3.73352855e-01 -4.58016694e-01
-1.37793601e+00 5.73240340e-01 -4.37081724e-01 -9.86864209e-01
-9.37977210e-02 -3.83575946e-01 -7.27890790e-01 1.80955863e+00
-1.65817392e+00 -3.84815216e-01 -5.60837746e-01 2.82446027e-01
1.63905382e-01 2.35789657e-01 9.60507393e-01 6.69617355e-01
-9.99886155e-01 3.64515424e-01 6.08142987e-02 4.55921292e-01
5.96601784e-01 -1.07063663e+00 2.97847748e-01 6.27477527e-01
-8.68418396e-01 1.07407594e+00 6.32095695e-01 -6.18208170e-01
-8.32213402e-01 -1.47768307e+00 1.10467649e+00 -4.32772711e-02
5.23949385e-01 7.70679116e-02 -1.09503996e+00 7.31887519e-01
-2.47018829e-01 5.17098606e-01 5.65632880e-01 1.90320849e-01
-1.78465843e-02 -6.54397428e-01 -8.93640459e-01 5.70148110e-01
7.21859336e-01 -2.05006346e-01 -4.22314912e-01 2.75387734e-01
6.27342761e-01 -7.84981966e-01 -1.52803469e+00 7.38175869e-01
9.31149483e-01 -7.49858141e-01 1.19724131e+00 -1.06496334e+00
7.86806881e-01 -2.23930076e-01 1.28233731e-01 -1.11084580e+00
-8.36850941e-01 -4.37212288e-01 -3.51095527e-01 7.51829743e-01
6.90780520e-01 -1.05663812e+00 4.93746400e-01 1.17866290e+00
-2.52216309e-01 -1.09255898e+00 -6.19488001e-01 -5.95939875e-01
2.12196440e-01 1.39147878e-01 5.36493003e-01 8.26154888e-01
-1.04959570e-01 5.39119303e-01 -9.29932594e-01 2.37773940e-01
4.58736986e-01 3.88444990e-01 5.19575655e-01 -1.36491084e+00
-1.88918471e-01 -4.47576284e-01 -4.84009355e-01 -1.27682316e+00
-5.06112993e-01 -7.28844404e-01 -1.95311442e-01 -1.60333514e+00
2.70979125e-02 -3.89722228e-01 -9.60771680e-01 4.60760921e-01
-5.34199059e-01 -2.54781336e-01 -1.55144155e-01 1.27851039e-01
3.27670515e-01 4.78562117e-01 1.48959291e+00 1.01437680e-01
-9.57365096e-01 5.30353069e-01 -5.84211230e-01 2.58792460e-01
1.49079955e+00 -2.06079707e-01 -4.29224312e-01 -3.24361533e-01
-1.25410736e-01 4.80071157e-01 4.30195063e-01 -1.04487109e+00
7.67268091e-02 -2.60084659e-01 1.09518182e+00 -4.40518945e-01
1.38940245e-01 -4.62894082e-01 8.94167572e-02 1.00554395e+00
-6.35898054e-01 -2.00150143e-02 3.63377929e-01 4.56842303e-01
-8.44904408e-02 1.89074159e-01 8.95055056e-01 -4.47559297e-01
-4.90147471e-01 5.81291854e-01 -2.65192658e-01 -2.09640525e-02
6.57179356e-01 -7.01032486e-03 -3.42541665e-01 -1.69073775e-01
-1.33225107e+00 3.74806911e-01 -6.57280326e-01 3.86381894e-01
6.29486978e-01 -1.37166810e+00 -1.05711579e+00 3.64627779e-01
-3.28708291e-01 -3.02441537e-01 5.80042720e-01 1.16703701e+00
-6.42674506e-01 7.00304210e-01 -3.58544260e-01 -7.35991657e-01
-8.66668105e-01 3.34583968e-01 9.96680498e-01 -1.36716470e-01
-1.25992680e+00 5.79302907e-01 -2.42025584e-01 -3.79406035e-01
3.39763016e-01 -6.26014948e-01 -2.65101969e-01 -4.28241551e-01
7.10542142e-01 5.82431614e-01 -5.78671098e-02 -1.02168331e-02
-1.93800226e-01 4.33545947e-01 -1.71412677e-01 4.02747184e-01
1.41518068e+00 -2.30377391e-01 -5.35608888e-01 6.97926462e-01
1.12546492e+00 -5.85348547e-01 -8.10447812e-01 -4.87922251e-01
3.19041684e-02 -3.58737022e-01 -1.23154499e-01 -1.15788901e+00
-1.09942055e+00 9.60019946e-01 1.08553970e+00 1.52241811e-02
8.58908772e-01 -7.30181158e-01 1.07288754e+00 3.63952309e-01
-2.92916864e-01 -7.79011369e-01 -4.67002869e-01 7.16126859e-01
9.80420113e-01 -9.95345414e-01 3.56564596e-02 -7.95927793e-02
-3.60739738e-01 1.62881219e+00 8.19085479e-01 -2.43558779e-01
1.05898666e+00 -3.48241031e-01 3.41387749e-01 -3.02711949e-02
-7.60400057e-01 4.65762049e-01 2.40010962e-01 3.72716151e-02
6.85057282e-01 3.20074886e-01 -4.94510382e-01 5.77264786e-01
-1.63937181e-01 4.49781030e-01 3.01784039e-01 6.57091916e-01
-6.99166656e-01 -7.19218194e-01 1.05783246e-01 8.32191348e-01
-4.04481381e-01 -2.59512126e-01 3.65664780e-01 3.53264660e-01
-1.86011553e-01 7.15745866e-01 -1.67970620e-02 -3.23623568e-01
6.67584717e-01 5.34751356e-01 4.65356521e-02 -2.24236682e-01
-4.47273731e-01 -9.94065329e-02 -2.81443689e-02 -4.60558951e-01
-2.33169235e-02 -1.21022232e-01 -1.10571945e+00 -3.84524256e-01
2.08223835e-01 4.05782945e-02 4.15421963e-01 7.15848982e-01
6.21533990e-01 7.04031527e-01 7.59395123e-01 -2.72355348e-01
-9.29402590e-01 -1.27785146e+00 -6.35848641e-01 1.62613809e-01
7.14041293e-01 -1.87317610e-01 -1.75528169e-01 1.95241515e-02] | [14.106114387512207, 3.0116355419158936] |
f200cb83-14e9-44cb-8831-c617306e375d | flock-defending-malicious-behaviors-in | 2211.04344 | null | https://arxiv.org/abs/2211.04344v1 | https://arxiv.org/pdf/2211.04344v1.pdf | FLock: Defending Malicious Behaviors in Federated Learning with Blockchain | Federated learning (FL) is a promising way to allow multiple data owners (clients) to collaboratively train machine learning models without compromising data privacy. Yet, existing FL solutions usually rely on a centralized aggregator for model weight aggregation, while assuming clients are honest. Even if data privacy can still be preserved, the problem of single-point failure and data poisoning attack from malicious clients remains unresolved. To tackle this challenge, we propose to use distributed ledger technology (DLT) to achieve FLock, a secure and reliable decentralized Federated Learning system built on blockchain. To guarantee model quality, we design a novel peer-to-peer (P2P) review and reward/slash mechanism to detect and deter malicious clients, powered by on-chain smart contracts. The reward/slash mechanism, in addition, serves as incentives for participants to honestly upload and review model parameters in the FLock system. FLock thus improves the performance and the robustness of FL systems in a fully P2P manner. | ['Shuhao Zheng', 'Shuoying Zhang', 'Zhipeng Wang', 'Jiahao Sun', 'Nanqing Dong'] | 2022-11-05 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [-6.01826668e-01 3.29336703e-01 -3.00777227e-01 -3.33789676e-01
-8.76226544e-01 -1.09568012e+00 6.12065732e-01 1.68577015e-01
-1.72588378e-01 8.20675015e-01 -2.36375540e-01 -2.41616398e-01
4.53198142e-02 -7.01666415e-01 -8.52754653e-01 -1.12044740e+00
-7.25026429e-02 6.43276095e-01 1.25271708e-01 3.44079167e-01
8.00173432e-02 3.89534831e-01 -1.05630481e+00 1.85034201e-01
6.87496483e-01 1.02209747e+00 -5.23398817e-01 3.46533239e-01
3.04526687e-01 1.18157566e+00 -4.02302951e-01 -7.23645091e-01
8.52672935e-01 -6.12341017e-02 -6.15640342e-01 -5.71077466e-01
1.36231869e-01 -9.09564257e-01 -3.52620147e-02 1.05876040e+00
2.38135234e-01 -7.05911756e-01 -9.49856825e-03 -2.30613828e+00
-4.98031169e-01 1.05819488e+00 -3.89219880e-01 -6.32913291e-01
-7.01137707e-02 5.81107378e-01 9.27715182e-01 -3.82093221e-01
6.45421386e-01 8.38906944e-01 6.86137497e-01 6.85024321e-01
-1.24138093e+00 -1.13177204e+00 -3.13618302e-01 3.58994782e-01
-1.10422039e+00 -5.02818286e-01 8.41630578e-01 -1.14422999e-01
2.81311899e-01 5.12062788e-01 5.96361995e-01 7.60555685e-01
3.80678803e-01 8.71613622e-01 1.20616674e+00 -1.16261549e-01
6.36758506e-01 4.59498107e-01 -1.31809339e-01 1.62720978e-01
6.67219698e-01 9.36974771e-03 -7.88715065e-01 -1.15409863e+00
5.16048789e-01 1.95909888e-01 -2.59235978e-01 -8.76178026e-01
-8.83025408e-01 8.74650657e-01 1.96839392e-01 -5.05552739e-02
-4.99698520e-01 7.27370858e-01 5.51774383e-01 7.60297656e-01
2.62592763e-01 5.55397496e-02 -8.44456971e-01 8.07706416e-02
-7.58885086e-01 4.56310451e-01 1.34990442e+00 5.76325715e-01
7.48594820e-01 -2.59781480e-01 1.41841704e-02 -2.72725672e-01
8.58087242e-01 7.12615609e-01 4.22781184e-02 -1.66805577e+00
1.67932689e-01 6.54738069e-01 5.40174603e-01 -8.29665124e-01
2.83754915e-01 7.50052780e-02 -5.59337378e-01 6.03630304e-01
5.28116941e-01 -3.04157048e-01 7.49981105e-02 1.68857074e+00
9.27620947e-01 -2.16830105e-01 2.98171639e-01 8.60676110e-01
6.81458041e-02 5.43828607e-01 -1.10259220e-01 -3.91951144e-01
1.06092560e+00 -7.96639979e-01 -5.77141821e-01 7.31558740e-01
8.91070664e-01 -1.72523871e-01 4.83622670e-01 6.62689328e-01
-1.20037091e+00 6.60651565e-01 -6.69888854e-01 4.77893986e-02
7.75821730e-02 -5.61478615e-01 3.62363964e-01 6.57368660e-01
-9.55184579e-01 6.06241643e-01 -1.13065732e+00 1.10140964e-01
8.02525461e-01 7.32898414e-01 -7.86634982e-01 1.42515168e-01
-1.05624843e+00 5.82480967e-01 -3.43031064e-02 -3.60379905e-01
-1.46691334e+00 -8.95738184e-01 -3.61624360e-01 2.96154886e-01
2.82988131e-01 -7.90112615e-01 1.51584101e+00 -9.19117391e-01
-1.59720480e+00 5.89277148e-01 4.26051348e-01 -1.06684673e+00
1.02414203e+00 1.53761648e-04 1.43544897e-01 4.15618718e-01
-2.28456810e-01 1.63997635e-01 7.99238920e-01 -1.46166337e+00
-6.59962118e-01 -5.73689759e-01 -2.65124977e-01 -2.03084871e-01
-4.37779933e-01 4.38821286e-01 4.89584148e-01 1.30124956e-01
-2.28279680e-01 -8.77618074e-01 -1.01346545e-01 5.14467835e-01
-2.52559453e-01 -5.65490663e-01 1.56864107e+00 -6.14958525e-01
6.68134809e-01 -1.98157442e+00 -7.22643912e-01 3.15749109e-01
4.73411202e-01 4.80210602e-01 2.10890040e-01 8.03918898e-01
6.36844039e-01 1.60605684e-01 4.04811315e-02 -5.36021590e-01
4.94197696e-01 2.70806313e-01 -4.83021557e-01 8.90336037e-01
-3.50687981e-01 9.51828420e-01 -8.09808135e-01 -5.48593283e-01
-1.54373646e-01 3.80547076e-01 -5.01845717e-01 8.74276310e-02
-5.18170714e-01 7.54764259e-01 -5.49966812e-01 6.09933615e-01
8.74372721e-01 -3.32644165e-01 5.83245814e-01 4.35645550e-01
-6.30481020e-02 1.37099430e-01 -1.15430081e+00 1.21200097e+00
-1.52765065e-01 1.80615827e-01 9.92736578e-01 -4.49444652e-01
9.50122833e-01 9.26641822e-01 8.61517847e-01 -2.86954284e-01
3.13815176e-01 4.93484139e-01 -5.12900114e-01 -3.10446531e-01
8.73515233e-02 8.19042549e-02 -7.23495558e-02 1.26085603e+00
-3.70518386e-01 2.23707825e-01 -5.62356949e-01 3.04536462e-01
1.26124692e+00 8.50323439e-02 3.64943705e-02 -1.69797525e-01
4.27985787e-01 8.26157406e-02 1.19059741e+00 5.62089920e-01
-5.58286369e-01 -1.43486515e-01 8.49802017e-01 -7.56830573e-01
-1.15551901e+00 -7.06908762e-01 4.65888619e-01 9.03387785e-01
2.99108252e-02 -3.06897223e-01 -8.17072868e-01 -9.54739511e-01
6.24931395e-01 5.16823649e-01 -1.06305234e-01 1.45823851e-01
-5.40666521e-01 -1.78607199e-02 8.20181191e-01 1.04514904e-01
6.84143186e-01 -8.46259773e-01 -6.78849876e-01 3.04619938e-01
-4.21968699e-02 -6.00542426e-01 -3.58159840e-01 -1.36154741e-01
-6.01766109e-01 -1.40620148e+00 -8.56063887e-02 -3.63838315e-01
6.82629704e-01 1.51068777e-01 5.82753539e-01 1.44798219e-01
3.28551501e-01 2.80723780e-01 -1.32840335e-01 -5.47970295e-01
-9.22541618e-01 -1.86284155e-01 3.01111966e-01 3.33094597e-01
2.97210604e-01 -9.33027208e-01 -1.13976038e+00 4.65864450e-01
-8.65066409e-01 -2.14869574e-01 1.43399134e-01 5.17541647e-01
3.44501317e-01 -3.08098346e-01 9.70079184e-01 -9.24818575e-01
4.48316246e-01 -4.59020913e-01 -1.22592759e+00 3.96957248e-01
-1.29662681e+00 -1.02243662e-01 9.93075430e-01 -3.64152253e-01
-7.75425971e-01 3.03389549e-01 5.29322863e-01 -6.87844396e-01
8.44647735e-02 -4.27787565e-02 -4.20673013e-01 -5.52292883e-01
4.54436332e-01 1.91428110e-01 4.65709627e-01 -6.13407135e-01
6.06357455e-01 1.01155174e+00 3.91688555e-01 -5.67989230e-01
9.10758376e-01 7.37234712e-01 1.87041938e-01 2.69631416e-01
3.65330800e-02 3.84178534e-02 -6.02224320e-02 6.26293421e-02
2.10248321e-01 -1.02419782e+00 -1.74824500e+00 7.92475939e-01
-1.30519080e+00 -2.67850995e-01 -4.11162794e-01 2.86592036e-01
-4.84496057e-01 4.33636665e-01 -8.38266909e-01 -9.10116792e-01
-1.14316928e+00 -7.93034613e-01 2.97260314e-01 1.64278686e-01
-5.55752590e-02 -5.97645700e-01 3.71753693e-01 7.24266946e-01
8.07231188e-01 4.52014863e-01 4.62735236e-01 -1.11195350e+00
-8.88112009e-01 -6.45075679e-01 -7.83406422e-02 5.04554808e-01
-1.69968158e-01 -6.12325855e-02 -1.02356660e+00 -6.56710207e-01
2.84101993e-01 -5.20563424e-01 -9.13642272e-02 -2.60215908e-01
7.64581561e-01 -1.53328562e+00 -1.14143141e-01 3.15960348e-01
1.34228957e+00 -3.81458789e-01 2.43079990e-01 6.00569397e-02
5.03255248e-01 4.74804461e-01 1.03475705e-01 8.19243789e-01
8.96985173e-01 2.86284029e-01 8.09017479e-01 4.44127023e-01
2.68714190e-01 -7.00431466e-01 4.90024418e-01 4.46433216e-01
2.79814452e-01 2.84758002e-01 -7.55809247e-01 6.74224138e-01
-1.98523092e+00 -7.17612207e-01 -3.12074512e-01 2.16300535e+00
1.19577348e+00 -5.26428998e-01 1.09751850e-01 4.90771942e-02
5.30902684e-01 -2.89181441e-01 -7.22496092e-01 -4.25457507e-01
8.44559669e-02 -3.03557605e-01 9.51282203e-01 1.52623430e-01
-5.02902865e-01 5.00847697e-01 4.86481810e+00 4.48403776e-01
-1.31462693e+00 7.55461931e-01 4.34606045e-01 -2.11428791e-01
-3.01504105e-01 7.59345889e-01 -4.38023776e-01 6.21380150e-01
1.09411216e+00 -6.65666997e-01 7.15378404e-01 1.27750778e+00
4.10276473e-01 7.21459836e-02 -1.18684649e+00 7.72512555e-01
-4.67625022e-01 -1.57139719e+00 -6.24647796e-01 4.00214940e-01
9.93269265e-01 3.30790311e-01 -2.74398118e-01 -1.02608480e-01
1.07522571e+00 -6.34275496e-01 6.70644045e-01 3.90066326e-01
4.37647164e-01 -7.67370343e-01 4.29317623e-01 7.50507772e-01
-6.07326210e-01 -5.89235961e-01 -4.21154462e-02 -4.77405824e-03
2.53277142e-02 4.45598006e-01 -7.97845960e-01 4.78614748e-01
7.21765101e-01 7.33395740e-02 -9.05193537e-02 9.76164460e-01
-5.28694749e-01 8.76946688e-01 -5.86572051e-01 2.59286821e-01
-3.22768055e-02 -5.77771664e-02 4.89957690e-01 4.39067572e-01
2.85323448e-02 -1.21468119e-01 5.35633340e-02 9.25889254e-01
-7.63422549e-01 -9.72560272e-02 -4.33121532e-01 2.79294074e-01
1.05792093e+00 1.46529174e+00 -2.67214254e-02 -2.26973787e-01
-1.16949067e-01 5.38100958e-01 9.00357664e-02 9.92437154e-02
-2.31879205e-01 -1.48679733e-01 6.32227659e-01 2.43960679e-01
3.01319122e-01 -1.70625106e-01 -4.59434003e-01 -1.06589019e+00
2.99120784e-01 -1.00379252e+00 6.08415663e-01 -4.50994492e-01
-1.35313463e+00 -1.72022343e-01 -4.56072330e-01 -1.00970209e+00
-1.83178186e-01 4.27154243e-01 -9.08540606e-01 4.65128541e-01
-1.81027019e+00 -1.68500304e+00 2.10150018e-01 1.02401209e+00
-4.45000678e-01 9.46690291e-02 8.02980006e-01 -7.13757128e-02
-2.61692643e-01 7.34235227e-01 5.82055151e-01 7.92379528e-02
8.78918767e-01 -8.62151086e-01 -1.50429283e-03 6.34233415e-01
-1.32553086e-01 4.70986217e-01 7.02084959e-01 -7.22313344e-01
-2.10560727e+00 -1.18913007e+00 1.13850188e+00 -4.58855480e-01
7.08956659e-01 -5.01921892e-01 -1.08277571e+00 7.94644952e-01
2.52213538e-01 5.47598779e-01 5.41769326e-01 -5.37034273e-01
-7.85617948e-01 -7.33240247e-01 -2.09681153e+00 2.00212374e-02
1.15544252e-01 -5.87365150e-01 -5.20525612e-02 5.04006445e-01
8.22250962e-01 1.28320098e-01 -1.05143702e+00 -6.65919483e-02
5.88222563e-01 -7.14535832e-01 2.24237815e-01 -5.17872036e-01
-2.27482736e-01 -4.67956245e-01 -1.41259115e-02 -8.48429143e-01
3.52671564e-01 -1.70149732e+00 -7.50571549e-01 1.61320412e+00
-3.55491005e-02 -1.21572042e+00 1.12894404e+00 1.42967689e+00
6.36403680e-01 -3.53444159e-01 -1.57384050e+00 -8.37439895e-01
4.76383865e-01 -5.27687296e-02 1.14465058e+00 1.06265438e+00
4.16904569e-01 -4.89900738e-01 -5.86314738e-01 3.63893956e-01
1.30480850e+00 8.48970339e-02 1.14613032e+00 -1.36926901e+00
-9.83716920e-02 1.21530257e-01 2.41214596e-02 -2.17552260e-01
1.13333412e-01 -8.15168083e-01 -1.26056373e-01 -1.11059356e+00
1.99348554e-01 -8.51072609e-01 -2.81670660e-01 1.11412072e+00
5.38122296e-01 2.04403885e-02 6.47711337e-01 9.43365455e-01
-8.83739054e-01 4.96464789e-01 7.93493927e-01 8.82967860e-02
-6.85185865e-02 8.38030726e-02 -7.14414358e-01 3.01470578e-01
9.46996748e-01 -1.08291352e+00 -7.56677389e-02 -3.77720952e-01
2.79680997e-01 3.23923051e-01 8.87590408e-01 -4.83506054e-01
9.97635543e-01 -4.32989478e-01 -3.03562194e-01 -2.96265095e-01
-2.87621528e-01 -1.29333174e+00 7.53666282e-01 8.50205004e-01
-2.33452901e-01 -2.92120039e-01 -7.29334712e-01 5.99376261e-01
2.01357916e-01 1.04967482e-01 7.62874603e-01 -1.59546241e-01
3.78798038e-01 4.23543274e-01 -9.23814401e-02 -4.89683002e-01
1.52058136e+00 4.17639762e-01 -7.53919601e-01 -5.61241210e-01
-2.04843268e-01 1.00509441e+00 1.13067174e+00 3.29396464e-02
1.06072672e-01 -1.10051107e+00 -9.09912527e-01 2.15055779e-01
-1.95800304e-01 3.36584985e-01 -2.90587544e-04 1.13790369e+00
-3.41603249e-01 3.27667028e-01 -3.94951552e-02 -2.27733538e-01
-1.35006773e+00 6.26933396e-01 2.62029827e-01 -3.27224404e-01
-6.32842898e-01 1.72418296e-01 -5.85409522e-01 -5.58936656e-01
5.43634951e-01 1.77830115e-01 6.51354730e-01 -1.79607242e-01
6.35074914e-01 5.92025399e-01 -2.32477952e-02 -2.14171156e-01
-2.60473549e-01 -2.93554246e-01 8.14091489e-02 -2.65545938e-02
1.54680073e+00 -2.15379894e-01 -6.34005249e-01 -8.72774050e-02
8.88737619e-01 -9.04468354e-03 -1.59920454e+00 -5.51917493e-01
9.31211635e-02 -5.52456558e-01 -1.00114852e-01 -8.41510117e-01
-1.17035723e+00 3.69353801e-01 3.03049862e-01 2.55670190e-01
5.94120264e-01 -5.29223122e-02 1.25692368e+00 3.81941646e-01
8.36455643e-01 -1.02175045e+00 -4.49092180e-01 -1.73738793e-01
4.57626641e-01 -1.00483048e+00 -1.13869041e-01 1.52991995e-01
-6.78522408e-01 8.73469591e-01 3.26155245e-01 -2.37780716e-02
5.06621420e-01 2.42259756e-01 2.20830843e-01 5.16642146e-02
-1.29280519e+00 9.41678405e-01 -8.41111422e-01 6.34247422e-01
-6.33499742e-01 2.79762477e-01 -3.81414860e-01 9.75543380e-01
1.59245193e-01 6.34168506e-01 6.83692515e-01 1.16156185e+00
-3.54853630e-01 -1.57320237e+00 -5.91601253e-01 -3.40274833e-02
-7.40213692e-01 5.27449310e-01 -4.27014947e-01 1.85706168e-01
-3.17971781e-02 1.08978367e+00 -4.31883991e-01 2.38661971e-02
-1.77817881e-01 2.07782567e-01 -2.61423379e-01 6.48364574e-02
-1.26013613e+00 -2.46478304e-01 -1.80567712e-01 -6.78750336e-01
-7.28310719e-02 -6.12749040e-01 -1.29367256e+00 -8.74047160e-01
-4.77666646e-01 6.10085011e-01 1.17818761e+00 5.48756182e-01
8.24669003e-01 -8.19631279e-01 1.60417676e+00 -2.88440585e-01
-1.25045574e+00 -5.40407419e-01 -8.72564197e-01 3.12973648e-01
4.39903110e-01 1.65446207e-01 -6.85545385e-01 4.87110801e-02] | [5.8483405113220215, 6.6025390625] |
abb4c0c1-be41-4dfe-81b1-6464bcbc0f96 | augment-and-criticize-exploring-informative | 2303.11243 | null | https://arxiv.org/abs/2303.11243v1 | https://arxiv.org/pdf/2303.11243v1.pdf | Augment and Criticize: Exploring Informative Samples for Semi-Supervised Monocular 3D Object Detection | In this paper, we improve the challenging monocular 3D object detection problem with a general semi-supervised framework. Specifically, having observed that the bottleneck of this task lies in lacking reliable and informative samples to train the detector, we introduce a novel, simple, yet effective `Augment and Criticize' framework that explores abundant informative samples from unlabeled data for learning more robust detection models. In the `Augment' stage, we present the Augmentation-based Prediction aGgregation (APG), which aggregates detections from various automatically learned augmented views to improve the robustness of pseudo label generation. Since not all pseudo labels from APG are beneficially informative, the subsequent `Criticize' phase is presented. In particular, we introduce the Critical Retraining Strategy (CRS) that, unlike simply filtering pseudo labels using a fixed threshold (e.g., classification score) as in 2D semi-supervised tasks, leverages a learnable network to evaluate the contribution of unlabeled images at different training timestamps. This way, the noisy samples prohibitive to model evolution could be effectively suppressed. To validate our framework, we apply it to MonoDLE and MonoFlex. The two new detectors, dubbed 3DSeMo_DLE and 3DSeMo_FLEX, achieve state-of-the-art results with remarkable improvements for over 3.5% AP_3D/BEV (Easy) on KITTI, showing its effectiveness and generality. Code and models will be released. | ['Junjun Jiang', 'Xianming Liu', 'Ke Wang', 'Yuan He', 'Heng Fan', 'Zhipeng Zhang', 'Zhenyu Li'] | 2023-03-20 | null | null | null | null | ['monocular-3d-object-detection', 'pseudo-label'] | ['computer-vision', 'miscellaneous'] | [ 1.79922506e-01 5.94249554e-02 -1.04159005e-01 -2.60778844e-01
-8.64381194e-01 -8.10064793e-01 7.27531672e-01 -2.60406554e-01
-3.59460890e-01 5.13247311e-01 -3.01235288e-01 -1.73677444e-01
3.23512316e-01 -2.33741865e-01 -8.69188786e-01 -9.49177384e-01
8.96751136e-02 4.63513047e-01 5.70576310e-01 2.26893008e-01
-2.15334422e-03 6.81228638e-01 -1.88772082e+00 7.25479890e-03
5.06703973e-01 1.20230556e+00 2.02577263e-01 6.14064991e-01
1.86742961e-01 7.33584762e-01 -5.37753701e-01 -1.50727272e-01
7.15475500e-01 -7.23747462e-02 -2.86714643e-01 5.63794553e-01
6.38396204e-01 -6.93745971e-01 -2.51597524e-01 9.43759620e-01
5.48130691e-01 -1.59189835e-01 7.77107835e-01 -1.30146897e+00
-3.34278584e-01 1.59108981e-01 -7.73065448e-01 1.27574399e-01
1.10514231e-01 5.17254889e-01 8.51228774e-01 -1.38287401e+00
7.65787721e-01 1.11576462e+00 5.95739365e-01 6.05338573e-01
-1.18444514e+00 -6.78089678e-01 4.56951439e-01 -5.46461828e-02
-1.19693053e+00 -5.14953375e-01 6.74188554e-01 -6.11650705e-01
7.06714809e-01 3.10164853e-03 4.62116450e-01 1.32052243e+00
-2.62418181e-01 1.19628608e+00 1.13054216e+00 -4.41445023e-01
3.44375312e-01 4.77394372e-01 1.23557478e-01 7.32385755e-01
4.48144555e-01 6.17444813e-01 -6.09406233e-01 -1.56139198e-04
5.79112768e-01 -7.53454417e-02 -1.25163067e-02 -9.02075469e-01
-1.07611024e+00 6.49546266e-01 4.56296265e-01 -3.40681642e-01
-3.03683192e-01 -8.03297535e-02 2.08389804e-01 1.44719034e-01
7.01992750e-01 3.19347650e-01 -5.01213610e-01 2.27191359e-01
-7.97739685e-01 4.71017733e-02 4.94924039e-01 1.02195680e+00
8.35499763e-01 -1.94969960e-02 -2.25596815e-01 6.32204175e-01
3.64080161e-01 6.99150264e-01 -5.77026047e-02 -1.03403521e+00
2.35001668e-01 7.47519732e-01 3.18487078e-01 -3.76478285e-01
-3.51523519e-01 -7.36660480e-01 -4.80094433e-01 6.07311130e-01
2.68819600e-01 -2.26822078e-01 -1.22354615e+00 1.60429883e+00
7.73226380e-01 1.88792169e-01 -8.65716785e-02 9.87003624e-01
6.96292341e-01 2.43328556e-01 -1.42630652e-01 -2.93355733e-01
1.06332600e+00 -1.09948647e+00 -1.93484038e-01 -5.78708887e-01
5.77723145e-01 -7.83976674e-01 8.10477316e-01 3.70319933e-01
-8.21367502e-01 -6.20929241e-01 -1.31969118e+00 2.81983584e-01
-8.20046961e-02 4.70108539e-01 4.71946418e-01 5.83420753e-01
-9.25240517e-01 3.68686408e-01 -9.55268741e-01 -4.16779041e-01
5.63945174e-01 1.55619472e-01 -3.04630220e-01 -2.12927431e-01
-6.43424153e-01 8.18910539e-01 2.51720011e-01 -1.13650374e-01
-1.30866730e+00 -5.00478923e-01 -6.49055958e-01 -4.26091075e-01
8.08593571e-01 -6.26313925e-01 1.49261725e+00 -6.70201778e-01
-1.37091863e+00 1.27196026e+00 -2.23688781e-02 -6.64129198e-01
7.13296711e-01 -4.94693607e-01 1.94008593e-02 2.19633773e-01
2.41896600e-01 1.02666724e+00 1.21751666e+00 -1.57494855e+00
-9.65801358e-01 -3.87472212e-01 -3.42652686e-02 3.23122323e-01
-2.38331500e-02 -1.01265356e-01 -6.97477877e-01 -4.46337163e-01
3.63133997e-01 -1.17138612e+00 -1.97556496e-01 3.01673472e-01
-4.39013362e-01 -2.12338403e-01 8.24482322e-01 -2.54337609e-01
7.60639369e-01 -2.26290154e+00 -4.88519855e-02 2.78714229e-03
4.23601657e-01 4.24676269e-01 -8.73917341e-02 1.44978642e-01
1.48761466e-01 -1.69633582e-01 -2.29545131e-01 -7.89605916e-01
-1.07978538e-01 2.30088696e-01 -3.43212396e-01 7.02898443e-01
5.99078715e-01 7.44655609e-01 -1.08372557e+00 -2.55238086e-01
5.06270409e-01 1.47559106e-01 -5.11923373e-01 3.07620794e-01
-4.14757311e-01 5.56528449e-01 -3.97455066e-01 1.15530062e+00
8.04568052e-01 -4.28589612e-01 -1.51825577e-01 -7.48535246e-02
-1.84360832e-01 2.33741447e-01 -1.31428158e+00 1.37159979e+00
-6.88178763e-02 5.20714045e-01 1.39159754e-01 -7.69386888e-01
9.74030375e-01 -1.34557649e-01 4.46147829e-01 -3.68018448e-01
1.75791830e-01 2.68743604e-01 -4.12301660e-01 -2.94819355e-01
4.44046676e-01 1.90341249e-01 6.56824484e-02 5.36566556e-01
2.66589642e-01 -1.44452617e-01 1.28865615e-01 3.68789732e-01
1.09018254e+00 6.48898482e-01 2.95602083e-01 1.39442712e-01
3.53355139e-01 1.81138664e-01 6.25745237e-01 1.10644221e+00
-5.95234752e-01 8.01460326e-01 2.80120224e-01 -1.55080721e-01
-1.08798075e+00 -1.17127454e+00 -2.98324138e-01 9.22302127e-01
3.50508511e-01 -1.79044649e-01 -4.53409970e-01 -1.11169243e+00
1.82128057e-01 5.27984321e-01 -5.81351638e-01 -1.27110794e-01
-3.76255184e-01 -8.42454791e-01 2.69901842e-01 5.80995202e-01
3.78864437e-01 -8.37198436e-01 -8.03266704e-01 9.93994102e-02
1.59122556e-01 -1.18405116e+00 -1.23643473e-01 6.27441406e-01
-6.75583243e-01 -1.22568774e+00 -4.74402457e-01 -4.44499701e-01
6.68369591e-01 8.73323083e-01 1.11988831e+00 -7.16940314e-02
-2.65573859e-01 5.31878710e-01 -5.58483899e-01 -4.29696649e-01
-4.09307629e-01 3.38211432e-02 3.45549464e-01 2.48783547e-02
4.53826249e-01 -3.16793114e-01 -6.16663933e-01 4.58132267e-01
-6.39208674e-01 7.24327192e-02 8.22448134e-01 9.51572537e-01
7.56843925e-01 -4.68365967e-01 3.25582922e-01 -8.01348925e-01
-8.80727246e-02 -3.74195039e-01 -9.92172956e-01 7.63083175e-02
-7.29580879e-01 6.72282651e-02 2.61641532e-01 -4.60028768e-01
-9.37290490e-01 5.57588756e-01 1.05211750e-01 -8.58429968e-01
-2.57184058e-01 -1.94926448e-02 9.70504209e-02 -2.36535281e-01
9.83290195e-01 1.33516937e-01 3.84250209e-02 -4.74400908e-01
5.31165183e-01 6.84105337e-01 6.94886804e-01 -3.71691883e-01
1.27810407e+00 8.05432618e-01 -4.28983271e-02 -5.81323862e-01
-1.28685248e+00 -9.10263479e-01 -5.97782433e-01 -3.66529971e-01
5.23086905e-01 -1.32221568e+00 -4.11658943e-01 5.15716016e-01
-1.02477837e+00 -2.55815208e-01 -5.18373251e-01 4.41031694e-01
-4.70776945e-01 4.20694113e-01 -4.52096641e-01 -1.17980933e+00
-1.24321818e-01 -9.66211677e-01 1.36615908e+00 1.78046301e-01
2.09566861e-01 -4.25790042e-01 -4.82723787e-02 1.98614255e-01
2.89965924e-02 2.73191929e-01 2.78357089e-01 -7.31994033e-01
-9.73064542e-01 -3.85646254e-01 -4.56822693e-01 6.90657973e-01
-2.59636901e-02 5.63600510e-02 -1.35578275e+00 -5.14272392e-01
1.36557117e-01 -6.72138929e-01 1.02319622e+00 2.15519339e-01
7.23124444e-01 7.31080696e-02 -4.91359919e-01 6.08664036e-01
1.22359073e+00 -8.76061469e-02 1.63271442e-01 3.89123052e-01
6.47800624e-01 5.22476077e-01 9.53678668e-01 5.66368699e-01
1.99646145e-01 6.29843354e-01 7.06222534e-01 -4.45132367e-02
-3.70319545e-01 -2.85452336e-01 5.08242249e-01 4.42962557e-01
7.13605657e-02 -1.74816892e-01 -6.18235528e-01 3.58457267e-01
-1.84421766e+00 -7.78181016e-01 -4.94086631e-02 2.36390471e+00
5.61015844e-01 4.70662296e-01 2.30401471e-01 1.38766080e-01
7.11813450e-01 1.19806677e-01 -8.86605322e-01 3.46951365e-01
-4.02850807e-01 -9.70345438e-02 7.87121952e-01 3.64191741e-01
-1.41633010e+00 9.69874322e-01 5.63671780e+00 6.02775037e-01
-8.72744560e-01 1.21759363e-01 5.46462297e-01 -2.75379598e-01
2.07931638e-01 5.93305789e-02 -1.31942976e+00 2.42153794e-01
5.89405060e-01 2.44270369e-01 3.53028595e-01 1.13925529e+00
1.09294534e-01 -3.71259123e-01 -1.28426170e+00 9.39940691e-01
1.61712930e-01 -1.14566422e+00 -1.77978203e-01 -2.50298660e-02
8.40428114e-01 4.19785887e-01 2.42760524e-01 3.64513487e-01
5.05735338e-01 -2.55318671e-01 9.81709778e-01 1.39787540e-01
7.31687844e-01 -4.13557053e-01 5.68203509e-01 5.47588468e-01
-1.02092040e+00 -2.52352029e-01 -4.23674881e-01 7.10704103e-02
2.32318744e-01 7.80885398e-01 -1.20539737e+00 5.88613153e-01
6.63588881e-01 9.23597813e-01 -8.80143225e-01 1.15982056e+00
-6.48640275e-01 4.80948269e-01 -5.67443132e-01 2.20551819e-01
3.09024930e-01 1.38441235e-01 8.58500183e-01 8.59180570e-01
2.47923151e-01 -1.83496907e-01 3.12231988e-01 6.55285418e-01
-4.55368608e-02 -3.41650575e-01 -6.11139297e-01 3.32853407e-01
7.22738922e-01 1.27371418e+00 -6.55295610e-01 -2.63940603e-01
-4.10923928e-01 8.04627597e-01 3.38450909e-01 1.94552585e-01
-8.30904186e-01 3.04993153e-01 3.68376255e-01 1.21576466e-01
4.60110903e-01 -1.63440287e-01 -2.19526649e-01 -1.16097832e+00
8.66859034e-02 -5.97567320e-01 3.78589332e-01 -7.45080709e-01
-1.51659977e+00 4.32753682e-01 -2.68172324e-02 -1.61380744e+00
-3.01906496e-01 -7.51095772e-01 -1.96980029e-01 5.17274678e-01
-1.77786589e+00 -1.12792170e+00 -5.04965603e-01 3.20528030e-01
5.81444979e-01 -3.69200520e-02 2.83918679e-01 2.93594450e-01
-6.69010878e-01 4.59140867e-01 7.59795457e-02 -1.89211950e-01
9.14508939e-01 -1.30216467e+00 4.76535171e-01 1.21426332e+00
2.34645739e-01 8.67948532e-02 5.95408142e-01 -6.02352738e-01
-1.22466218e+00 -1.31824791e+00 5.21742642e-01 -8.32944810e-01
5.75399995e-01 -4.46219891e-01 -7.09323049e-01 7.07657814e-01
-2.91798085e-01 3.23264003e-01 1.53370053e-01 -1.48597762e-01
-3.74550551e-01 -6.19452968e-02 -1.03009057e+00 5.37782192e-01
1.41912043e+00 -4.14040327e-01 -5.70360065e-01 5.73928654e-01
8.05607378e-01 -5.05752981e-01 -2.77600676e-01 6.75936103e-01
2.11424559e-01 -1.20464039e+00 1.17000580e+00 -2.42437154e-01
8.67293328e-02 -6.87702715e-01 -2.52500892e-01 -9.09970284e-01
-1.43552244e-01 -6.38329268e-01 -5.37862539e-01 1.10553241e+00
2.15581968e-01 -5.57101667e-01 1.05781245e+00 1.92838520e-01
-3.29278648e-01 -5.84205925e-01 -9.16214287e-01 -9.42742527e-01
-4.15408105e-01 -4.52134103e-01 1.74550474e-01 6.31008148e-01
-6.22999668e-01 3.99706751e-01 -5.40654063e-01 4.70232666e-01
9.65768456e-01 4.29530665e-02 1.10340619e+00 -1.32121074e+00
-5.18751740e-01 -1.28745958e-01 -1.85674146e-01 -1.40837824e+00
-1.75833568e-01 -7.50479937e-01 3.38706017e-01 -9.71774101e-01
2.31738582e-01 -7.31119215e-01 -2.39768416e-01 4.07761365e-01
-3.39352965e-01 4.76634562e-01 2.34698161e-01 5.73591411e-01
-8.87894511e-01 5.14567912e-01 1.06259215e+00 -3.46545577e-02
-3.36984694e-01 2.11051732e-01 -3.61900866e-01 6.94384456e-01
4.95487839e-01 -5.54584503e-01 -2.11709648e-01 -2.12879777e-01
-8.01099539e-02 -3.01278234e-01 6.28561378e-01 -1.09219706e+00
2.21000925e-01 1.91773951e-01 4.26999360e-01 -1.09620345e+00
4.03079391e-01 -5.84432006e-01 -5.71399815e-02 4.71636534e-01
8.99979100e-02 -2.42431208e-01 5.18468488e-03 8.01605225e-01
-3.10457498e-02 -1.35811776e-01 9.50164080e-01 -1.46571010e-01
-8.96698833e-01 3.56174886e-01 -9.39230546e-02 -6.42493069e-02
1.15797925e+00 -3.87953222e-01 -4.79580909e-01 -3.18420194e-02
-5.75255454e-01 3.77071589e-01 6.85463786e-01 2.96081990e-01
4.88251865e-01 -1.18314230e+00 -6.20054483e-01 3.04977298e-01
5.36641479e-01 3.89434993e-01 1.49763077e-01 8.06540847e-01
-2.71955073e-01 2.34628424e-01 1.06342137e-01 -1.14726877e+00
-9.29106414e-01 7.22725153e-01 2.32344314e-01 -1.75822094e-01
-6.04736745e-01 9.45304275e-01 2.67955720e-01 -3.74806464e-01
4.58824158e-01 -1.89467538e-02 7.16215819e-02 6.77340999e-02
3.52222592e-01 3.19518626e-01 1.41869053e-01 -3.94585133e-01
-3.98835719e-01 3.07631373e-01 -2.53840148e-01 5.00337258e-02
1.32571924e+00 -3.20660502e-01 3.90818685e-01 3.95273864e-01
7.89914250e-01 -1.39647275e-01 -2.13898993e+00 -5.27153611e-01
3.51960771e-02 -4.46107149e-01 -3.78709808e-02 -7.95290649e-01
-7.45292485e-01 7.15296745e-01 8.58982205e-01 9.38864499e-02
9.09134328e-01 3.51338744e-01 3.19402128e-01 3.58564407e-01
4.62797225e-01 -9.99803066e-01 3.99329841e-01 4.74276990e-01
5.48611462e-01 -1.59014702e+00 1.13417648e-01 -4.42505032e-01
-5.18914044e-01 7.27809846e-01 9.04942095e-01 -7.11016431e-02
2.99897850e-01 1.46520659e-01 2.20659211e-01 -1.29412502e-01
-7.46621609e-01 -4.81954634e-01 9.60998386e-02 5.76154351e-01
-2.78693646e-01 -1.29728392e-01 2.83167601e-01 2.20898598e-01
2.65978009e-01 -1.95456550e-01 4.52485681e-01 1.03584373e+00
-6.54882491e-01 -9.60767806e-01 -5.37530243e-01 4.12694365e-01
6.08016364e-02 1.62197635e-01 -4.09400493e-01 8.74282122e-01
3.12210172e-01 8.65514696e-01 -3.67590711e-02 -4.57453251e-01
3.15805107e-01 4.06046733e-02 4.76254433e-01 -8.88810039e-01
-2.47809559e-01 2.63238996e-01 1.55103467e-02 -5.41638255e-01
-7.13461518e-01 -9.23256755e-01 -7.33260453e-01 2.07286812e-02
-6.98865235e-01 -3.71411979e-01 5.28940201e-01 8.39482784e-01
3.91448468e-01 3.23250532e-01 1.00265491e+00 -1.32389641e+00
-9.50373292e-01 -8.77792776e-01 -4.16572362e-01 2.61907339e-01
4.65999097e-01 -1.16801453e+00 -7.67090857e-01 -7.38741383e-02] | [9.1622953414917, 1.2399746179580688] |
9498628c-d370-40a2-aab6-662a8a7f3b78 | improving-face-recognition-by-clustering | 2007.06995 | null | https://arxiv.org/abs/2007.06995v2 | https://arxiv.org/pdf/2007.06995v2.pdf | Improving Face Recognition by Clustering Unlabeled Faces in the Wild | While deep face recognition has benefited significantly from large-scale labeled data, current research is focused on leveraging unlabeled data to further boost performance, reducing the cost of human annotation. Prior work has mostly been in controlled settings, where the labeled and unlabeled data sets have no overlapping identities by construction. This is not realistic in large-scale face recognition, where one must contend with such overlaps, the frequency of which increases with the volume of data. Ignoring identity overlap leads to significant labeling noise, as data from the same identity is split into multiple clusters. To address this, we propose a novel identity separation method based on extreme value theory. It is formulated as an out-of-distribution detection algorithm, and greatly reduces the problems caused by overlapping-identity label noise. Considering cluster assignments as pseudo-labels, we must also overcome the labeling noise from clustering errors. We propose a modulation of the cosine loss, where the modulation weights correspond to an estimate of clustering uncertainty. Extensive experiments on both controlled and real settings demonstrate our method's consistent improvements over supervised baselines, e.g., 11.6% improvement on IJB-A verification. | ['Manmohan Chandraker', 'Erik Learned-Miller', 'Kihyuk Sohn', 'Aruni RoyChowdhury', 'Xiang Yu'] | 2020-07-14 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/4531_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123690120.pdf | eccv-2020-8 | ['face-clustering'] | ['computer-vision'] | [ 2.43294522e-01 3.68001238e-02 -1.73575401e-01 -8.58743370e-01
-9.48372602e-01 -5.97733200e-01 3.51571411e-01 -8.52643847e-02
-3.63745898e-01 5.89096367e-01 3.10033914e-02 -2.69997921e-02
9.78690013e-03 -3.06948215e-01 -5.08772731e-01 -9.41752374e-01
1.35660961e-01 4.77052093e-01 -4.25458431e-01 2.80054569e-01
4.38661054e-02 3.30729604e-01 -1.50274944e+00 1.31823644e-01
7.76869357e-01 1.12633538e+00 -3.42994541e-01 -7.09112883e-02
-2.09704489e-01 4.03151572e-01 -7.83726454e-01 -6.97936833e-01
5.88782549e-01 -5.52547872e-01 -5.70845306e-01 2.98322022e-01
8.42788994e-01 -2.38619000e-01 -5.20887896e-02 1.37347364e+00
6.27211034e-01 -1.07905149e-01 7.40713418e-01 -1.77959311e+00
-7.17807710e-01 5.01184464e-01 -1.10670674e+00 -2.44199961e-01
2.87306868e-02 -1.86258048e-01 9.32977796e-01 -1.01420140e+00
4.74592507e-01 1.23313522e+00 9.76324618e-01 7.89931297e-01
-1.43669963e+00 -1.09893692e+00 9.14621949e-02 1.55295506e-01
-1.81269491e+00 -7.58468032e-01 5.22813439e-01 -4.48018700e-01
3.64023089e-01 2.87140626e-02 4.52604964e-02 9.53806400e-01
-4.41609651e-01 6.02747560e-01 8.54171574e-01 -5.65987229e-01
3.59842449e-01 3.66321683e-01 1.66637823e-01 3.55359882e-01
4.73815292e-01 3.21513414e-02 -4.71461952e-01 -2.31447622e-01
2.70505458e-01 -5.92282340e-02 -2.82858193e-01 -5.04916608e-01
-8.16657603e-01 7.72052050e-01 7.49006793e-02 3.17504741e-02
-1.20125636e-01 -1.21495932e-01 2.50923634e-01 3.37826721e-02
4.86257523e-01 9.59635377e-02 -3.63217086e-01 7.40748420e-02
-1.24197519e+00 -8.58503655e-02 7.65250146e-01 1.07854259e+00
9.25287187e-01 -1.15098603e-01 -5.66868111e-02 1.05865622e+00
4.18564051e-01 5.52731097e-01 3.65025163e-01 -1.17558587e+00
3.19207489e-01 4.80057955e-01 1.29579306e-01 -9.22132492e-01
-6.81186914e-02 -3.09354663e-01 -7.97331274e-01 2.34268308e-01
8.45780909e-01 -2.89055675e-01 -9.06645179e-01 2.22264004e+00
4.06909853e-01 3.81315738e-01 -5.35501949e-02 9.00837779e-01
3.47780854e-01 1.58391148e-01 7.35541955e-02 -3.27570975e-01
1.42136192e+00 -5.07397354e-01 -9.36942935e-01 -1.48522735e-01
5.33243060e-01 -7.86929429e-01 7.69166887e-01 2.69435436e-01
-7.15165555e-01 -3.10225397e-01 -1.06678033e+00 1.42912671e-01
-8.34414288e-02 1.13829076e-01 5.27839839e-01 1.12611508e+00
-1.02106750e+00 3.87580067e-01 -5.83339930e-01 -2.58153439e-01
8.04643869e-01 5.73580146e-01 -6.03160262e-01 -1.64625362e-01
-1.03387368e+00 6.73903108e-01 1.11540347e-01 1.89285576e-01
-4.77076888e-01 -6.99155509e-01 -8.92334640e-01 1.07697710e-01
4.19611663e-01 -6.35787174e-02 1.19189060e+00 -1.20646584e+00
-1.09333229e+00 9.20346916e-01 -5.34056664e-01 -4.62882146e-02
3.04972887e-01 2.00349540e-02 -4.59611565e-01 -9.47535932e-02
1.68467090e-01 7.34678626e-01 8.02570105e-01 -1.51311278e+00
-5.60625613e-01 -6.59557819e-01 -4.07652557e-01 1.74800720e-04
-6.05778396e-01 1.51620314e-01 -6.14736259e-01 -5.29395640e-01
2.55592227e-01 -1.13672268e+00 -3.86165306e-02 1.08517349e-01
-1.97810784e-01 -2.02024698e-01 8.51205111e-01 -5.11795342e-01
8.82600248e-01 -2.37460113e+00 -5.43151259e-01 4.56593812e-01
1.70606583e-01 3.01922709e-01 -1.55336276e-01 3.31884553e-03
-1.54455230e-01 1.86173946e-01 -2.76189417e-01 -5.72654426e-01
1.36857480e-01 1.31693613e-02 -1.27110615e-01 5.74321747e-01
2.78340876e-01 5.48552692e-01 -6.63723111e-01 -5.55997908e-01
-2.46051967e-01 4.67218518e-01 -4.87166554e-01 -6.22412823e-02
1.05803847e-01 3.22820783e-01 -2.99909580e-02 6.00948453e-01
1.18828905e+00 -3.67229015e-01 4.75143820e-01 -2.69915730e-01
2.88529813e-01 -7.17411712e-02 -1.46300423e+00 1.44757032e+00
-1.71927929e-01 5.36441922e-01 3.86849403e-01 -1.01279426e+00
8.14415514e-01 3.45291913e-01 6.79300785e-01 -3.00508767e-01
2.42628127e-01 3.56810950e-02 7.51265138e-02 -1.56666636e-01
2.50768125e-01 -3.02121073e-01 1.03985898e-01 6.36563122e-01
7.43887294e-03 4.85083848e-01 -8.51609781e-02 2.55930215e-01
7.78554738e-01 -1.29217505e-01 9.16850045e-02 -2.29891792e-01
2.67082065e-01 -2.53406733e-01 1.03869760e+00 6.14246011e-01
-6.90518141e-01 7.12736607e-01 4.76554275e-01 7.07412064e-02
-7.87620842e-01 -9.86452699e-01 -3.81047904e-01 7.74169803e-01
2.92825907e-01 -2.67570555e-01 -1.07631433e+00 -9.11719501e-01
2.90445954e-01 3.87927234e-01 -5.67878902e-01 -7.34740496e-02
-3.33160728e-01 -9.76026297e-01 6.98068559e-01 4.72517610e-01
3.45876813e-01 -6.65919244e-01 9.49834008e-03 -4.58050258e-02
-3.13674271e-01 -1.22944927e+00 -8.85964990e-01 4.28627245e-02
-4.32204485e-01 -1.13413525e+00 -3.62124562e-01 -9.25824463e-01
9.61404562e-01 3.55385840e-01 8.27038586e-01 -4.05572057e-02
-4.55571860e-01 3.18142414e-01 -1.87214404e-01 -4.51511890e-01
-3.29590887e-01 -2.81011313e-01 4.30584818e-01 4.93171602e-01
1.05399573e+00 -4.02863592e-01 -5.52397251e-01 6.15712702e-01
-6.95245028e-01 -4.59718108e-01 4.12225544e-01 1.17921543e+00
4.34686601e-01 7.35225007e-02 9.62801814e-01 -8.03687513e-01
3.47526222e-01 -4.58274275e-01 -4.67797726e-01 4.83993590e-01
-8.50841224e-01 -1.59533527e-02 4.32798684e-01 -6.53033495e-01
-1.18947661e+00 2.13991061e-01 1.83052555e-01 -4.68952030e-01
-2.04770774e-01 1.26063257e-01 -5.84600747e-01 -2.87096292e-01
5.15382826e-01 -1.59813315e-01 3.46950471e-01 -3.40589851e-01
3.35479081e-01 1.16458261e+00 3.99977058e-01 -6.80630624e-01
7.09162474e-01 6.04661286e-01 -1.80619240e-01 -7.06870258e-01
-7.69978106e-01 -5.50906718e-01 -5.71953535e-01 -9.29811820e-02
4.97494102e-01 -9.83317018e-01 -1.01138210e+00 4.84906256e-01
-8.88573885e-01 1.01737462e-01 -2.55014092e-01 6.04724765e-01
-5.24046347e-02 6.45518303e-01 -5.38154602e-01 -8.86404395e-01
-1.97557896e-01 -1.20783746e+00 9.42037284e-01 3.34155679e-01
-2.40722492e-01 -6.97959781e-01 -2.99375862e-01 4.53681260e-01
2.35640556e-01 -2.39537638e-02 5.81126451e-01 -1.01514709e+00
-3.83996725e-01 -3.44691813e-01 -3.82452399e-01 5.88206053e-01
3.79573584e-01 -1.28367662e-01 -1.25860703e+00 -4.09427464e-01
2.07473993e-01 -5.61159730e-01 5.56747198e-01 2.26809755e-01
1.13845992e+00 -1.08183689e-01 -4.42761719e-01 4.50645745e-01
1.10732532e+00 1.71821147e-01 3.88007641e-01 -2.01095134e-01
6.74863875e-01 9.17559445e-01 5.66117287e-01 6.06481433e-01
2.71487296e-01 8.40764165e-01 5.42429164e-02 -9.72657129e-02
-1.35950565e-01 -8.53793696e-02 1.59625202e-01 4.44154710e-01
4.02297050e-01 -6.06376007e-02 -8.56536031e-01 4.25579011e-01
-1.63397920e+00 -1.10943556e+00 1.12067901e-01 2.44918633e+00
1.09650421e+00 -1.66016787e-01 5.28562714e-06 9.92657468e-02
1.25051904e+00 -2.64080524e-01 -7.02148616e-01 7.68732131e-02
-9.64245424e-02 1.15990706e-01 5.17351449e-01 3.83081198e-01
-1.08649647e+00 6.99096024e-01 6.01712084e+00 1.09952652e+00
-1.02666140e+00 8.30629468e-02 8.34228933e-01 -2.04994097e-01
5.12594618e-02 -2.95545161e-02 -1.12799680e+00 6.80587649e-01
7.71845102e-01 -8.13476369e-02 3.58377934e-01 8.29523802e-01
1.93291306e-02 -3.62730622e-02 -1.30838633e+00 1.14727807e+00
3.65640610e-01 -7.58556008e-01 -2.77926177e-01 3.67838055e-01
7.63284147e-01 -1.93374962e-01 1.40134618e-01 1.50032595e-01
4.05512929e-01 -9.58679259e-01 6.65333509e-01 -6.75865337e-02
9.87145901e-01 -7.21452773e-01 7.34525621e-01 1.05884954e-01
-1.08686626e+00 -2.76271068e-02 -2.89774925e-01 2.36351147e-01
-1.44661054e-01 7.25113392e-01 -8.39305937e-01 2.30139881e-01
4.56521809e-01 3.38223428e-01 -3.24706584e-01 8.22078049e-01
-8.73032287e-02 6.46357298e-01 -5.21811724e-01 2.74828225e-01
-2.40902558e-01 -3.24419141e-01 -3.01634502e-02 9.49145615e-01
2.63482988e-01 6.43323138e-02 2.75791794e-01 7.29311705e-01
-5.26121140e-01 1.09975055e-01 -4.66614097e-01 6.17862828e-02
1.00899899e+00 1.32897973e+00 -6.91232324e-01 -2.61956155e-01
-5.84300280e-01 9.23164964e-01 3.03558081e-01 3.36759448e-01
-8.67529213e-01 -3.53527755e-01 9.29583132e-01 -1.05754271e-01
2.33986571e-01 5.91983311e-02 -4.52013016e-01 -1.06161296e+00
2.26365954e-01 -9.54207063e-01 3.81582052e-01 -1.41511679e-01
-1.58491838e+00 3.69948715e-01 -2.03228116e-01 -1.09502769e+00
-7.65718594e-02 -3.34260345e-01 -2.62307316e-01 9.56314206e-01
-1.25875211e+00 -1.02016377e+00 -2.14878693e-01 4.77433056e-01
1.41988277e-01 -8.93626362e-02 8.13530087e-01 7.63840973e-01
-7.18268216e-01 1.44134879e+00 3.83060008e-01 4.25808579e-01
1.23734868e+00 -9.98870611e-01 1.42774850e-01 7.53396690e-01
2.91746795e-01 8.61536145e-01 3.20095867e-01 -4.91308242e-01
-1.15008199e+00 -1.00737476e+00 8.35233629e-01 -4.63083446e-01
3.51896077e-01 -5.96751869e-01 -1.01262200e+00 5.18714786e-01
-1.46969080e-01 2.77370661e-01 1.20175886e+00 3.51752162e-01
-8.12661767e-01 -3.79371166e-01 -1.57823384e+00 3.91432196e-01
1.13667214e+00 -5.58891833e-01 -1.81178838e-01 2.47669503e-01
2.98355341e-01 4.72535789e-02 -8.26524317e-01 4.92886841e-01
7.05582440e-01 -7.53191233e-01 7.93145418e-01 -4.48158413e-01
-2.43535507e-02 -4.26240981e-01 -1.36153787e-01 -1.14343023e+00
-1.48390234e-01 -4.55527902e-01 1.94322124e-01 1.72121215e+00
3.20106357e-01 -5.87743998e-01 1.11452472e+00 1.20631242e+00
2.43290067e-01 -4.25254852e-01 -1.00100911e+00 -1.02742100e+00
-9.16654095e-02 -2.06379995e-01 7.13991523e-01 1.32695866e+00
1.87528819e-01 2.31807753e-01 -2.95951694e-01 2.67493337e-01
9.23046589e-01 5.18786395e-03 6.11403286e-01 -1.29628181e+00
-6.65633008e-02 -3.14560473e-01 -4.92194831e-01 -7.67231703e-01
5.18020332e-01 -8.42718422e-01 3.21056008e-01 -9.12780344e-01
4.26439881e-01 -7.36120284e-01 -2.98721105e-01 7.31072724e-01
-3.59937727e-01 5.97476482e-01 1.18339442e-01 3.65955472e-01
-5.34270585e-01 4.45911050e-01 6.84728146e-01 -1.58867389e-01
2.63451450e-02 -1.72941506e-01 -8.81792605e-01 6.56316638e-01
8.07529986e-01 -5.42422473e-01 -4.99122322e-01 -2.33776957e-01
-4.10084784e-01 -1.96051285e-01 -1.07463859e-01 -9.06944335e-01
3.32476526e-01 1.99182853e-02 4.91863757e-01 -1.08236618e-01
3.73329312e-01 -1.04813528e+00 1.69958547e-01 8.05983171e-02
-3.92971277e-01 -3.54558319e-01 1.19716659e-01 5.93531787e-01
-2.76109129e-01 -2.58344173e-01 1.09581470e+00 2.77954757e-01
-3.53876084e-01 3.98338944e-01 1.47582740e-02 2.37915203e-01
1.21375513e+00 -2.92730182e-01 -3.31159979e-01 -3.89097542e-01
-5.75586617e-01 2.35856354e-01 6.76314592e-01 3.05510849e-01
2.47748941e-01 -1.46782672e+00 -6.44024014e-01 4.29605275e-01
2.03605711e-01 -2.15565726e-01 1.68671861e-01 6.98973536e-01
-3.35199423e-02 7.73959681e-02 1.47067264e-01 -5.74466288e-01
-1.56851757e+00 5.58852017e-01 2.03818649e-01 1.20453805e-01
-1.42033756e-01 1.00256908e+00 1.82093143e-01 -3.38991225e-01
5.85260749e-01 3.17214340e-01 7.60723874e-02 2.10955471e-01
6.67697012e-01 2.27187589e-01 1.93468288e-01 -7.46449947e-01
-5.54974198e-01 6.04823291e-01 -3.55691075e-01 -1.70478135e-01
8.68595362e-01 -1.18016005e-01 -1.35831982e-01 8.27191249e-02
1.34293246e+00 1.54919311e-01 -1.26808500e+00 -2.98145235e-01
3.07613313e-01 -7.07193315e-01 -2.80298918e-01 -7.78423667e-01
-1.15591466e+00 7.74645746e-01 8.16173911e-01 -2.55382080e-02
1.05180764e+00 -1.33547053e-01 7.43182242e-01 2.65790522e-01
4.86838251e-01 -1.11385882e+00 -4.78144251e-02 2.20325351e-01
3.21596980e-01 -1.64431953e+00 -3.94803919e-02 -4.64083463e-01
-6.92348778e-01 6.98897064e-01 7.56798685e-01 4.46853280e-01
7.64323592e-01 4.53206867e-01 3.77559930e-01 2.11796407e-02
-3.21034044e-01 -2.21047960e-02 -3.26628536e-02 6.30799770e-01
3.93957108e-01 1.11800671e-01 -1.93057045e-01 7.12728202e-01
1.25752702e-01 -5.72874546e-02 3.25924486e-01 8.19227755e-01
-1.52874172e-01 -1.44352210e+00 -6.58904374e-01 5.23828447e-01
-5.96202672e-01 -1.53425306e-01 -3.61029059e-01 5.03861547e-01
2.93586850e-01 1.26876020e+00 1.44338652e-01 -4.86721337e-01
1.15926377e-03 3.60928595e-01 4.76281554e-01 -5.55528104e-01
-2.33360156e-01 -3.64664719e-02 -7.70623758e-02 -2.94213921e-01
-4.04009461e-01 -7.91945159e-01 -1.30564857e+00 -4.29450303e-01
-6.75957143e-01 4.59965825e-01 8.97540092e-01 7.05975711e-01
6.84073687e-01 -2.94784885e-02 8.90496671e-01 -5.65440595e-01
-9.90001082e-01 -9.78725255e-01 -7.64253318e-01 7.75394678e-01
1.00671135e-01 -6.98036313e-01 -6.19135559e-01 2.39823908e-01] | [13.467323303222656, 1.0567272901535034] |
1c621c78-b72a-420f-9026-c6d3c221d8e4 | irit-at-trac-2018 | null | null | https://aclanthology.org/W18-4403 | https://aclanthology.org/W18-4403.pdf | IRIT at TRAC 2018 | This paper describes the participation of the IRIT team to the TRAC 2018 shared task on Aggression Identification and more precisely to the shared task in English language. The three following methods have been used: a) a combination of machine learning techniques that relies on a set of features and document/text vectorization, b) Convolutional Neural Network (CNN) and c) a combination of Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM). Best results were obtained when using the method (a) on the English test data from Facebook which ranked our method sixteenth out of thirty teams, and the method (c) on the English test data from other social media, where we obtained the fifteenth rank out of thirty. | ['Josiane Mothe', 'Rami', 'Faneva risoa'] | 2018-08-01 | null | null | null | coling-2018-8 | ['aggression-identification'] | ['natural-language-processing'] | [-3.20235103e-01 -2.08287790e-01 -2.06602402e-02 -1.61020875e-01
-6.02100134e-01 -1.11914642e-01 9.49546576e-01 1.26346081e-01
-1.16064394e+00 7.90394247e-01 5.79420745e-01 -2.74533778e-01
-5.72541356e-01 -5.70581913e-01 -2.00818330e-01 -2.97197580e-01
-1.16669804e-01 5.09593785e-01 -7.45549947e-02 -5.02287209e-01
4.52051610e-01 5.06175101e-01 -1.18660808e+00 5.99525750e-01
4.86615062e-01 9.96096849e-01 -4.35780317e-01 7.72018313e-01
-2.69335330e-01 1.20041013e+00 -5.42726874e-01 -8.81859899e-01
1.07857391e-01 7.50284642e-02 -1.41436684e+00 -6.15583241e-01
5.84623992e-01 -5.17441034e-01 -5.82198918e-01 7.46865809e-01
6.58726811e-01 4.10741776e-01 6.51405811e-01 -8.21112990e-01
-6.49562955e-01 9.36350346e-01 -3.44851166e-01 7.53311157e-01
4.16837335e-01 2.63938960e-02 9.54252541e-01 -9.73244905e-01
5.06011367e-01 1.24008811e+00 9.52487826e-01 6.52109981e-01
-9.41074133e-01 -8.08846772e-01 -3.02186579e-01 3.96155387e-01
-1.48989344e+00 -3.46197754e-01 4.85529661e-01 -8.25134695e-01
1.49105382e+00 8.49420391e-03 4.45139825e-01 1.89502215e+00
-2.67053127e-01 5.85091352e-01 8.84479821e-01 -2.45986015e-01
-9.24527869e-02 3.85660082e-01 8.22853982e-01 6.63435876e-01
-3.86002511e-02 4.60452914e-01 -7.23203361e-01 -5.20847499e-01
1.92990884e-01 -3.57497275e-01 3.05723041e-01 5.77636123e-01
-5.79322398e-01 1.28303123e+00 2.12308496e-01 8.32856297e-01
-2.88373232e-01 3.33494134e-02 9.90469515e-01 3.75536680e-01
6.11892402e-01 6.58701837e-01 -4.08812612e-01 -6.40876234e-01
-9.53630686e-01 4.83271092e-01 9.02596593e-01 1.15970381e-01
3.38200837e-01 1.15640424e-01 -4.06217188e-01 1.34401512e+00
-1.26330899e-02 -6.04004189e-02 8.67203474e-01 -6.09959662e-01
8.99823606e-01 4.53309953e-01 -2.13855192e-01 -9.16834891e-01
-7.01160133e-01 -3.97715747e-01 -4.80808794e-01 7.16596618e-02
5.90906799e-01 -4.47930366e-01 -4.70072389e-01 1.51005232e+00
-3.86967331e-01 2.30635260e-03 -1.49697646e-01 6.01574481e-01
1.19365537e+00 2.28910610e-01 1.97430104e-01 3.29436451e-01
9.53496754e-01 -1.10610044e+00 -3.81985307e-01 5.32807522e-02
1.07787883e+00 -3.58228236e-01 4.89343882e-01 7.22136915e-01
-1.11326015e+00 -5.63334167e-01 -1.03068066e+00 -1.85058549e-01
-1.13837349e+00 -7.81347007e-02 6.82692170e-01 9.84548867e-01
-1.37062228e+00 7.07427740e-01 -2.87134975e-01 -5.86173356e-01
3.50701481e-01 5.18871188e-01 -6.47978842e-01 3.66328537e-01
-1.26107478e+00 1.16054213e+00 2.53477514e-01 -6.96258433e-03
-7.33202517e-01 -5.69537044e-01 -5.84134400e-01 5.61839482e-03
9.55713615e-02 -5.42052209e-01 1.01903892e+00 -9.50960517e-01
-1.67245734e+00 1.10232139e+00 3.79163682e-01 -7.55049407e-01
6.64759994e-01 -5.19510031e-01 -3.97613615e-01 -1.06819920e-01
-7.17027932e-02 2.89133728e-01 4.76112098e-01 -5.22876620e-01
-6.08446658e-01 -4.73007202e-01 2.09753308e-03 -2.93408513e-01
-9.34122980e-01 8.88081193e-01 1.54335067e-01 -3.41652751e-01
-5.66917181e-01 -7.90785193e-01 1.35846958e-01 -6.39596522e-01
-5.67708552e-01 -9.35847342e-01 6.66248620e-01 -1.09644127e+00
1.52619386e+00 -1.97535348e+00 4.72345650e-01 1.52882233e-01
3.05668056e-01 5.98160625e-01 -1.30457163e-01 6.31136775e-01
-7.14326560e-01 5.81607997e-01 2.37644628e-01 -8.00153196e-01
1.45930067e-01 -7.63334855e-02 -2.58248508e-01 3.29751581e-01
-5.49410693e-02 6.26993835e-01 -6.72775805e-01 -1.28170028e-01
1.26248837e-01 1.87065050e-01 -3.86891037e-01 1.24325547e-02
4.82363820e-01 3.02702397e-01 -1.88171268e-01 2.55587339e-01
3.03820252e-01 5.23985803e-01 -5.86282134e-01 2.17952773e-01
-4.56596315e-01 4.49409127e-01 -6.63226008e-01 1.41208565e+00
-5.37351370e-01 7.47208893e-01 9.17407125e-02 -6.37301445e-01
8.46376359e-01 5.24025381e-01 5.01363754e-01 -4.33702767e-01
4.25889075e-01 2.67072320e-01 2.71788269e-01 -8.57615232e-01
3.79880905e-01 9.84634534e-02 -2.34973915e-02 6.49909019e-01
3.54612708e-01 3.28260243e-01 3.47619176e-01 2.55419195e-01
1.41905451e+00 -3.75549436e-01 1.01672105e-01 -2.02407807e-01
6.40283227e-01 -6.70048952e-01 5.40918186e-02 9.87079620e-01
-4.26641732e-01 4.28161412e-01 7.67755687e-01 -8.25377285e-01
-9.44119751e-01 -6.61642313e-01 -2.47961313e-01 1.33844602e+00
-1.07205665e+00 -6.61041379e-01 -7.15570927e-01 -7.76693881e-01
8.24674219e-02 7.45628893e-01 -1.06113648e+00 -4.80219841e-01
-4.55348879e-01 -5.22826612e-01 1.21833479e+00 5.36418736e-01
4.80959296e-01 -1.16269529e+00 -1.93747357e-01 -4.12081406e-02
2.00790279e-02 -9.00616765e-01 -6.95505217e-02 1.55365765e-01
-2.68752515e-01 -9.92130756e-01 -4.54303861e-01 -1.77016199e-01
-3.38759869e-01 -4.91680592e-01 7.29978740e-01 1.27742156e-01
-2.09742382e-01 3.27580035e-01 -4.85088974e-01 -5.72626352e-01
-7.84181058e-02 5.01404941e-01 2.42071167e-01 1.02706067e-01
6.49834931e-01 -5.64906776e-01 1.18405595e-01 -3.57938617e-01
-5.68933070e-01 -3.27778757e-01 3.54835004e-01 7.92444229e-01
-6.55061424e-01 -4.00348574e-01 5.49708843e-01 -7.50485182e-01
1.12256181e+00 -7.86829650e-01 -7.82734305e-02 -1.97877824e-01
-2.51346529e-01 -3.65474969e-01 3.90786231e-01 -3.44444007e-01
-7.20993102e-01 -3.49638045e-01 -7.18394876e-01 -4.09174740e-01
-4.00256783e-01 8.04656386e-01 5.80944419e-01 -6.67587444e-02
7.04338312e-01 -1.50545731e-01 -1.81844398e-01 -7.99504280e-01
-1.03185708e-02 1.04164362e+00 2.67863274e-01 -5.08565128e-01
4.96208400e-01 2.60415617e-02 -9.98811498e-02 -1.13711488e+00
-9.41741526e-01 -3.92218262e-01 -1.00518882e+00 -2.16324002e-01
1.42506909e+00 -4.58204031e-01 -8.43490660e-01 7.81101227e-01
-1.19548917e+00 -4.23182338e-01 3.99045236e-02 6.70149565e-01
-2.10235089e-01 1.35872290e-01 -1.14915776e+00 -9.26270783e-01
-5.53885341e-01 -8.63928795e-01 8.64121735e-01 -1.50288446e-02
-7.03431427e-01 -1.13519859e+00 4.45131481e-01 6.60838664e-01
7.00249493e-01 2.17988133e-01 1.02541196e+00 -1.33600807e+00
5.77590466e-01 -3.65696430e-01 -3.03077996e-01 6.03941739e-01
-3.10449392e-01 2.99904853e-01 -1.11572325e+00 -2.22726420e-01
-3.01604599e-01 -6.80591166e-01 1.17979908e+00 9.58181769e-02
1.39937901e+00 -2.60743499e-01 3.59801129e-02 4.08264905e-01
8.28167081e-01 -4.59052715e-03 6.07469916e-01 6.45056486e-01
7.79393733e-01 6.53021097e-01 -2.30142862e-01 4.51585412e-01
2.88875461e-01 9.99881387e-01 2.69353807e-01 7.82481879e-02
6.55024350e-02 -1.51026726e-01 3.11388820e-01 6.83565140e-01
-3.31647873e-01 9.56731066e-02 -1.16674507e+00 4.54258412e-01
-1.79474652e+00 -1.11830425e+00 -4.31407243e-01 2.12257934e+00
3.16785812e-01 1.43531799e-01 8.21401417e-01 1.99889481e-01
4.68610823e-01 2.57995073e-02 -1.48795033e-03 -1.39463687e+00
1.52815104e-04 7.00560272e-01 1.27654880e-01 3.53039026e-01
-1.29929340e+00 8.46178532e-01 6.91123295e+00 9.21532750e-01
-1.11422336e+00 3.57872665e-01 4.54165578e-01 -4.13015902e-01
5.57370603e-01 -5.07216275e-01 -7.55883336e-01 6.20822549e-01
1.59107935e+00 -2.61996657e-01 5.17262578e-01 5.67720890e-01
-4.13625240e-02 -3.42576504e-02 -1.04066110e+00 9.92145300e-01
2.55727172e-01 -9.30622876e-01 9.41954777e-02 4.74481225e-01
4.60552961e-01 4.11016822e-01 3.06062520e-01 7.81747937e-01
3.15345645e-01 -1.53004062e+00 5.83066463e-01 8.91911149e-01
3.02889317e-01 -8.23780239e-01 1.06038094e+00 4.30117100e-01
-5.37102401e-01 -7.40395606e-01 -2.32082859e-01 -4.71833050e-01
-3.13757509e-01 1.58316538e-01 -5.92736304e-01 5.53939581e-01
1.13499689e+00 8.35738361e-01 -8.30559313e-01 9.62787509e-01
-1.93894431e-02 8.87995243e-01 -2.11749852e-01 -2.38432169e-01
6.50437653e-01 -1.92650020e-01 8.64627898e-01 1.61366594e+00
2.87352409e-02 -3.26503575e-01 -1.02501348e-01 5.95652640e-01
-1.93488032e-01 2.80341119e-01 -7.74616539e-01 -1.08972460e-01
-1.07865043e-01 1.43986297e+00 5.68561889e-02 -1.33952960e-01
-4.51326311e-01 7.12139904e-01 1.04311490e+00 1.78941458e-01
-6.80342197e-01 -3.08054864e-01 4.00434613e-01 -1.62029549e-01
4.43837084e-02 -1.05352871e-01 -2.89977163e-01 -5.94994545e-01
-4.16268736e-01 -6.12764537e-01 8.36934090e-01 -6.72942817e-01
-1.77140939e+00 8.37377965e-01 2.58150119e-02 -5.77557206e-01
-4.15973365e-01 -8.54116321e-01 -9.14838970e-01 8.45876873e-01
-6.76945329e-01 -1.10347712e+00 9.55761001e-02 7.24287689e-01
3.26020420e-01 -7.30692387e-01 9.86917794e-01 8.61985683e-01
-1.07629955e+00 7.18859255e-01 -1.87542289e-03 4.69192952e-01
6.34858191e-01 -1.26222503e+00 -1.48207575e-01 3.14039350e-01
-1.61602872e-03 5.47407925e-01 4.26834375e-01 -2.48359188e-01
-6.70052767e-01 -9.80046332e-01 1.14962256e+00 -6.72062039e-01
1.12170863e+00 -3.22384834e-01 -6.97171807e-01 8.95177126e-01
4.28204983e-01 -4.82498586e-01 9.40283477e-01 5.21463871e-01
-4.54379141e-01 -2.47112960e-02 -1.04486406e+00 3.13361913e-01
9.30872798e-01 -9.10922885e-01 -6.55822635e-01 5.47332168e-01
7.52457604e-02 -1.57467961e-01 -8.77673864e-01 2.21215263e-01
6.82721019e-01 -1.10292268e+00 7.51876295e-01 -1.24181724e+00
9.24815357e-01 5.35019457e-01 -7.21998662e-02 -1.37340546e+00
-4.28109765e-01 -2.82469124e-01 8.06480870e-02 1.39252555e+00
5.94473600e-01 -8.40843797e-01 3.03994149e-01 6.76566839e-01
-1.72683835e-01 -7.91511834e-01 -1.45214868e+00 -6.93300307e-01
7.11188257e-01 -4.85964179e-01 1.23514161e-01 1.08651555e+00
1.76181436e-01 4.16987538e-01 -5.53202033e-01 -5.97463131e-01
2.25166511e-02 -6.99325860e-01 7.85491765e-01 -1.31773293e+00
4.39209938e-02 -1.03317153e+00 -7.03992784e-01 -1.32205024e-01
7.09551513e-01 -1.13371301e+00 -6.57928765e-01 -1.16307831e+00
5.14157593e-01 9.09767672e-02 -4.07221407e-01 6.48949742e-01
1.84497938e-01 3.77113491e-01 2.66426243e-02 -3.01217645e-01
-4.89482194e-01 2.92760432e-01 3.19118083e-01 -1.10192329e-01
-1.00254724e-02 7.48514235e-02 -9.11237597e-01 8.65337372e-01
8.47066760e-01 -4.65312630e-01 3.12870711e-01 -4.48553145e-01
3.24347854e-01 -1.66685820e-01 4.34336871e-01 -1.12065983e+00
2.71584272e-01 3.94709617e-01 3.77153397e-01 -3.60349983e-01
6.71324253e-01 -2.48847246e-01 -3.10033977e-01 5.37452459e-01
-6.39172733e-01 1.00075744e-01 2.94853061e-01 -1.53167294e-02
-1.26033604e-01 -6.39807761e-01 6.08419716e-01 2.78921667e-02
-1.85249642e-01 3.33221555e-01 -7.84264207e-01 -2.67618477e-01
7.07355499e-01 -5.63565036e-03 -5.17499924e-01 -5.97167134e-01
-1.08527613e+00 9.40560736e-03 -4.05199826e-01 7.58537591e-01
3.33871245e-01 -1.08652472e+00 -1.10909033e+00 7.29368702e-02
-1.42511591e-01 -1.07619810e+00 2.51472592e-01 1.10437584e+00
-2.28423044e-01 7.58635879e-01 -3.18177819e-01 -1.20413482e-01
-1.48116016e+00 2.08803728e-01 6.27700388e-01 -6.77226663e-01
-1.74654782e-01 1.05820203e+00 -4.15426493e-01 -7.37067401e-01
1.58808783e-01 4.05599445e-01 -8.37287247e-01 2.60824203e-01
7.67284095e-01 8.52787197e-01 2.85296649e-01 -1.17097747e+00
-1.74233973e-01 2.37782344e-01 -3.54733467e-01 -4.20887619e-01
1.67386174e+00 3.12681615e-01 -5.73504448e-01 5.94187200e-01
1.65018415e+00 1.63112134e-01 -2.51884405e-02 -9.76470560e-02
3.20436209e-01 -3.50535601e-01 3.75185251e-01 -8.62684846e-01
-9.34205115e-01 9.91674840e-01 5.29999852e-01 6.10399723e-01
4.58369613e-01 -1.56075299e-01 6.41501904e-01 5.15870988e-01
1.26346275e-01 -1.21174574e+00 7.85075650e-02 1.08245826e+00
1.22346330e+00 -8.14749599e-01 -2.29309186e-01 1.09086581e-01
-4.79936332e-01 1.51164353e+00 7.55575180e-01 -2.90166944e-01
9.10140276e-01 -5.54877007e-03 -2.77369589e-01 -3.72639447e-01
-9.47543561e-01 -2.01143503e-01 4.46645379e-01 3.35182846e-01
4.73582834e-01 -9.15334672e-02 -6.51014984e-01 1.14023852e+00
-7.17900097e-01 -2.13729609e-02 2.91875958e-01 3.42438102e-01
7.95577392e-02 -9.01605725e-01 -7.04251900e-02 9.65135217e-01
-9.59050953e-01 -1.30079791e-01 -1.15028644e+00 8.59648049e-01
3.18299085e-01 1.25394475e+00 1.54883474e-01 -8.25334609e-01
3.18178892e-01 3.44726354e-01 2.92501718e-01 -6.46460950e-01
-1.39385164e+00 -5.18283904e-01 6.73731804e-01 -6.17429614e-01
3.45641654e-03 -9.37970042e-01 -5.95697463e-01 -6.47380114e-01
-2.61125296e-01 6.24184720e-02 8.51816416e-01 1.30139565e+00
-4.30557951e-02 4.28338289e-01 5.07643402e-01 -8.24893594e-01
-5.24390638e-01 -1.61112976e+00 -7.81199276e-01 4.83242720e-01
1.53102845e-01 -7.58606255e-01 -5.92257142e-01 -7.11953044e-01] | [8.806180000305176, 10.729721069335938] |
bdfcb860-079b-45dc-a4ac-2a8360bc1f9f | penet-a-joint-panoptic-edge-detection-network | 2303.08848 | null | https://arxiv.org/abs/2303.08848v1 | https://arxiv.org/pdf/2303.08848v1.pdf | PENet: A Joint Panoptic Edge Detection Network | In recent years, compact and efficient scene understanding representations have gained popularity in increasing situational awareness and autonomy of robotic systems. In this work, we illustrate the concept of a panoptic edge segmentation and propose PENet, a novel detection network called that combines semantic edge detection and instance-level perception into a compact panoptic edge representation. This is obtained through a joint network by multi-task learning that concurrently predicts semantic edges, instance centers and offset flow map without bounding box predictions exploiting the cross-task correlations among the tasks. The proposed approach allows extending semantic edge detection to panoptic edge detection which encapsulates both category-aware and instance-aware segmentation. We validate the proposed panoptic edge segmentation method and demonstrate its effectiveness on the real-world Cityscapes dataset. | ['Giuseppe Loianno', 'Yang Zhou'] | 2023-03-15 | null | null | null | null | ['edge-detection'] | ['computer-vision'] | [ 5.59513494e-02 1.36482060e-01 -2.80982435e-01 -4.28351760e-01
-1.27102062e-01 -2.88203239e-01 8.29906821e-01 2.77944624e-01
-3.41158032e-01 4.99929637e-01 -1.20988078e-02 5.53908795e-02
-4.54419613e-01 -8.97640884e-01 -5.80773115e-01 -2.54204839e-01
-3.89465898e-01 2.92104900e-01 5.25899410e-01 -1.57161906e-01
3.95976186e-01 4.21314985e-01 -1.75077665e+00 3.04219425e-01
1.17410839e+00 1.20233583e+00 5.00503719e-01 6.20552242e-01
2.63579413e-02 3.77822518e-01 -3.49998027e-01 2.65149534e-01
6.92223310e-01 2.83023089e-01 -6.02170706e-01 3.11081111e-01
5.65983474e-01 3.66436131e-02 -2.27747634e-01 1.02366877e+00
2.01985464e-02 5.99118710e-01 7.08850920e-01 -1.74957120e+00
-4.64512914e-01 1.90343782e-01 -7.90917754e-01 3.02781105e-01
-2.26442199e-02 2.40321066e-02 1.25120878e+00 -7.31339514e-01
6.99571848e-01 1.20900393e+00 8.09632838e-01 -1.67237520e-01
-9.72515166e-01 -3.54595423e-01 5.08603454e-01 4.58668590e-01
-1.37093449e+00 -9.52720046e-02 7.68041432e-01 -6.36157811e-01
1.38851249e+00 -1.47783577e-01 7.02332258e-01 6.33364379e-01
3.95448685e-01 1.05978787e+00 9.35359597e-01 -8.53296369e-03
2.57818997e-01 -3.17641795e-02 3.21666121e-01 1.01881254e+00
5.12996435e-01 4.34145093e-01 -2.03987762e-01 2.71278143e-01
8.88729751e-01 2.82128721e-01 -6.75758794e-02 -9.24681723e-01
-9.37056124e-01 7.37508774e-01 9.06680703e-01 1.72760606e-01
-3.92748386e-01 2.78149098e-01 6.03510261e-01 2.13727150e-02
6.46156609e-01 1.90402135e-01 -4.34362739e-01 3.20579559e-01
-1.00483227e+00 -2.24089902e-03 7.92731762e-01 1.19482231e+00
1.07225299e+00 1.28558770e-01 9.01465490e-02 7.98324585e-01
2.08041698e-01 2.78227717e-01 3.50209951e-01 -8.15593183e-01
4.00195539e-01 8.93370986e-01 2.21499532e-01 -1.28113687e+00
-1.06614339e+00 -4.42497641e-01 -7.24216580e-01 5.17793357e-01
1.93876430e-01 -2.48636812e-01 -1.08687162e+00 1.56918871e+00
3.98309171e-01 6.27889633e-01 1.17034651e-01 1.08847368e+00
6.28688037e-01 5.66600204e-01 3.84076893e-01 4.16877657e-01
1.54166222e+00 -1.47699010e+00 -3.97617519e-01 -7.27849662e-01
6.99846148e-01 -3.27557862e-01 5.72628736e-01 2.62409210e-01
-5.36927283e-01 -9.45690036e-01 -1.31975114e+00 -6.98020086e-02
-9.71361816e-01 5.60619354e-01 1.09805858e+00 4.48644280e-01
-1.04196918e+00 3.18105370e-01 -8.29948127e-01 -5.82835436e-01
4.36626226e-01 1.56561360e-01 -4.73711133e-01 2.09167928e-01
-1.02530456e+00 9.24723566e-01 1.19891942e+00 -4.28865440e-02
-7.64114678e-01 -4.53923017e-01 -1.38454437e+00 2.79525280e-01
8.03564906e-01 -8.14109683e-01 8.85235488e-01 -7.67340541e-01
-1.44485068e+00 6.49766922e-01 2.57164747e-01 -7.37542987e-01
4.88232702e-01 -5.64624190e-01 -4.88498390e-01 2.54145771e-01
3.35244030e-01 1.06200457e+00 7.15915024e-01 -1.08429825e+00
-1.32592094e+00 -4.65937734e-01 -2.17413120e-02 4.62399364e-01
1.51646942e-01 -5.42132676e-01 -3.91638279e-01 -4.27753001e-01
2.42261738e-01 -7.90717006e-01 -4.76927966e-01 3.55645530e-02
-3.54339600e-01 -4.55397815e-01 1.11129582e+00 -5.40523350e-01
8.41028988e-01 -2.05812716e+00 -1.31832987e-01 2.18964756e-01
1.84605911e-01 1.23916231e-01 -1.23780489e-01 2.90513821e-02
-8.44340213e-03 -2.56575733e-01 -3.31783593e-01 -3.86022508e-01
5.91657795e-02 2.20673382e-01 4.70867567e-02 4.22252864e-01
3.75444740e-02 8.35398078e-01 -1.06781125e+00 -5.09002984e-01
7.76024282e-01 1.82498857e-01 -6.32612526e-01 -5.97563805e-03
-3.17062885e-01 1.86324522e-01 -6.21161878e-01 4.68226135e-01
8.23925436e-01 -2.25087598e-01 -8.32567923e-03 -2.32675821e-01
-3.03770006e-01 -2.04881549e-01 -1.49320722e+00 2.15588903e+00
-5.85584283e-01 8.03744137e-01 1.01763777e-01 -1.32710767e+00
1.09056592e+00 -2.50684600e-02 5.90776801e-01 -6.04320765e-01
1.37612164e-01 5.71921878e-02 -4.92923826e-01 -5.77161431e-01
1.05436671e+00 2.35880271e-01 -5.20476878e-01 -3.25256400e-02
2.47643948e-01 -3.60433400e-01 4.09701049e-01 5.82398623e-02
6.64541781e-01 5.31583786e-01 6.45614624e-01 -5.33574760e-01
5.73478699e-01 3.72608721e-01 5.60587525e-01 7.57124066e-01
-4.79868710e-01 3.72861594e-01 9.06230286e-02 -4.81094748e-01
-8.46720934e-01 -1.18688810e+00 -5.89583628e-02 1.16255546e+00
7.47768939e-01 -2.40674347e-01 -5.71313620e-01 -7.98988938e-01
3.20359230e-01 7.67697453e-01 -5.76074958e-01 -8.28897431e-02
-3.97124916e-01 -6.75480723e-01 2.72283524e-01 8.17066073e-01
9.99192476e-01 -8.37192655e-01 -1.12177420e+00 3.45550686e-01
-1.00272872e-01 -1.45103872e+00 -3.19083810e-01 4.01630938e-01
-6.42563820e-01 -1.17755973e+00 -4.03135717e-01 -9.35510099e-01
4.58151579e-01 6.03268087e-01 8.25550616e-01 -4.46102500e-01
-6.70712113e-01 5.69669425e-01 -1.69498995e-01 -1.47724405e-01
2.63931602e-02 4.48702974e-03 -2.15241969e-01 5.17730601e-03
2.77362704e-01 -5.61854780e-01 -7.42254257e-01 2.20923737e-01
-6.62205279e-01 3.03457201e-01 6.89605415e-01 6.64624870e-01
5.62987328e-01 -1.06099442e-01 5.32366991e-01 -7.79966295e-01
4.54654187e-01 -6.51978970e-01 -1.07507372e+00 4.03417766e-01
-6.01943374e-01 1.42840713e-01 4.32821006e-01 5.71585074e-02
-1.37263429e+00 2.22517982e-01 2.45757863e-01 -4.48555082e-01
-4.88943607e-01 4.26547319e-01 1.89966541e-02 -6.98653236e-02
6.53084338e-01 -4.74081412e-02 -3.37548584e-01 -1.86977968e-01
9.20062780e-01 5.74293077e-01 8.84587288e-01 -3.68757874e-01
2.27469549e-01 7.33910024e-01 2.47747432e-02 -1.01061988e+00
-8.21010649e-01 -1.05540907e+00 -8.01218927e-01 -2.68802494e-01
1.14546108e+00 -1.02901685e+00 -6.20578349e-01 3.86657566e-01
-9.56493914e-01 -4.17249799e-01 -2.17863858e-01 4.87411559e-01
-8.72232616e-01 4.99143571e-01 -3.13815475e-01 -6.81864858e-01
-1.81652337e-01 -9.57253098e-01 1.36840940e+00 4.24931020e-01
9.77619290e-02 -1.25950170e+00 6.16844334e-02 1.02825888e-01
1.91202998e-01 3.94815296e-01 3.92215908e-01 -7.53729761e-01
-8.60665441e-01 -7.60521740e-02 -8.68649364e-01 -5.66721596e-02
1.65605292e-01 -3.06080252e-01 -7.31728911e-01 -1.00624047e-01
-3.35813463e-01 -1.50783835e-02 1.22405016e+00 6.66005433e-01
9.38282073e-01 1.21222258e-01 -8.27125847e-01 8.19685161e-01
1.83320522e+00 2.50142276e-01 1.42651424e-01 4.14238423e-01
7.29198515e-01 4.84376878e-01 6.24367774e-01 6.05292380e-01
6.28038764e-01 5.06143749e-01 4.97298986e-01 -6.78144023e-02
-2.06214786e-01 -2.30536133e-01 -1.11320309e-01 1.60550967e-01
1.91733569e-01 -3.79773587e-01 -8.29767227e-01 8.05377245e-01
-2.15694070e+00 -7.31166482e-01 -7.64758959e-02 1.75800455e+00
-2.12300584e-01 2.69495577e-01 -1.19778456e-03 -3.58248860e-01
1.00805509e+00 3.22486073e-01 -6.82358623e-01 -3.42976451e-01
-5.54556102e-02 -1.58182144e-01 8.58811319e-01 5.23325324e-01
-1.70664477e+00 1.38269830e+00 6.25146627e+00 6.15097940e-01
-9.04441357e-01 1.64668173e-01 3.90750229e-01 5.04923880e-01
2.08356678e-01 2.06350777e-02 -7.65177131e-01 3.88830081e-02
3.84738386e-01 -1.02936693e-01 1.10824026e-01 1.37591040e+00
5.47410809e-02 -4.97768760e-01 -7.65608311e-01 1.04597771e+00
1.64716020e-01 -1.23632574e+00 6.70019351e-03 -2.47332096e-01
8.36292326e-01 4.17417586e-01 -1.95048690e-01 3.45581055e-01
7.88925529e-01 -5.25017977e-01 5.35601616e-01 4.36153203e-01
3.49387079e-01 -4.92168784e-01 3.45709860e-01 3.10827196e-01
-1.71438384e+00 -5.74529707e-01 -2.28180423e-01 -6.73133880e-02
4.32866693e-01 3.49841893e-01 -9.43365932e-01 9.31466222e-01
5.95675945e-01 1.01635146e+00 -4.16694254e-01 1.56041884e+00
-2.46411532e-01 -1.57648310e-01 -6.45622253e-01 1.10776752e-01
7.90811300e-01 -5.86503565e-01 7.51626730e-01 1.50871193e+00
3.80951732e-01 -1.23951308e-01 7.73221076e-01 9.10315454e-01
2.98522115e-01 -1.09667495e-01 -7.50131309e-01 4.22836363e-01
2.22611219e-01 1.58593190e+00 -1.35141218e+00 -7.64632404e-01
-4.94642586e-01 9.88520741e-01 3.49438846e-01 4.96556610e-01
-7.69522369e-01 -5.15496790e-01 6.70603037e-01 -5.69293737e-01
5.15023291e-01 -4.90978688e-01 -5.26568651e-01 -1.04683495e+00
-4.10592347e-01 9.76533741e-02 6.96390629e-01 -9.15713608e-01
-8.32776248e-01 4.87033397e-01 3.79516512e-01 -1.10920739e+00
-1.10016003e-01 -8.39137614e-01 -7.19929755e-01 2.10800126e-01
-1.73446620e+00 -1.22939253e+00 -7.85471141e-01 4.55750942e-01
9.05253708e-01 6.73061311e-02 3.75787765e-01 1.60587817e-01
-5.16148865e-01 1.88811287e-01 7.54406974e-02 -3.98868024e-02
2.78870851e-01 -1.45563841e+00 4.75430816e-01 1.00132632e+00
2.60826856e-01 -6.49437457e-02 6.40045226e-01 -7.77967274e-01
-5.22795737e-01 -1.40313137e+00 3.58147204e-01 5.19457534e-02
7.11807489e-01 -2.06263214e-01 -5.74338913e-01 9.52416360e-01
1.61717236e-01 -1.84329879e-02 9.02413055e-02 -2.66474839e-02
-3.00644130e-01 -1.35949790e-01 -1.19940913e+00 5.97159863e-01
1.47385132e+00 -2.83326387e-01 -9.54198241e-01 3.42232078e-01
6.46461070e-01 -2.33459249e-01 -5.91225684e-01 6.22284055e-01
2.46075466e-01 -8.79660666e-01 1.06918669e+00 -1.69491202e-01
-2.79247969e-01 -2.63957143e-01 4.32368517e-02 -1.65218556e+00
-4.93507087e-01 -4.06548053e-01 1.73387170e-01 4.39997792e-01
2.34543785e-01 -7.60848224e-01 9.45226490e-01 2.20985934e-01
-9.20829058e-01 -1.92194581e-01 -8.55953097e-01 -8.78198206e-01
-4.58862096e-01 -5.06143510e-01 4.08993244e-01 7.62516737e-01
1.97722867e-01 2.62130946e-01 -1.57207549e-01 3.21614653e-01
7.20328867e-01 5.69882631e-01 6.10307992e-01 -1.56766856e+00
4.26711589e-02 -6.72914207e-01 -7.90921748e-01 -1.37983632e+00
2.30321407e-01 -1.09511435e+00 2.20774859e-01 -1.89545929e+00
-1.41033903e-01 -3.65654230e-01 -3.17003995e-01 4.50288445e-01
-6.06068261e-02 -6.99218782e-03 8.52166787e-02 -1.28685564e-01
-1.05213678e+00 6.60531282e-01 1.12779152e+00 -1.70486823e-01
-5.20978868e-01 -2.86026616e-02 -2.20850557e-01 8.99086773e-01
8.05282474e-01 -1.61102474e-01 -6.56679749e-01 -3.02688688e-01
-1.25769183e-01 -2.30804794e-02 4.38131899e-01 -1.51865828e+00
6.01934254e-01 2.50398051e-02 1.13688208e-01 -7.16141164e-01
4.67084318e-01 -9.69783723e-01 -1.36006176e-01 4.73697722e-01
-8.40272605e-02 -4.40144353e-02 4.11993235e-01 9.61425364e-01
-1.59110084e-01 -1.86887402e-02 5.87271631e-01 -2.71281153e-01
-1.87538779e+00 3.24490964e-01 -4.26477700e-01 -1.53618678e-01
1.41484976e+00 -6.27949119e-01 -1.89971119e-01 1.07355021e-01
-1.04097569e+00 6.24156415e-01 3.01323161e-02 7.62126088e-01
5.91482520e-01 -9.62895989e-01 -4.00366783e-01 2.98768044e-01
3.70319515e-01 6.80797920e-03 4.22044635e-01 6.49208248e-01
-5.23907185e-01 5.80757022e-01 -6.38841391e-01 -8.38745892e-01
-8.68066788e-01 6.95753992e-01 4.26302880e-01 -8.35381821e-02
-9.11205530e-01 5.72015226e-01 6.13316119e-01 -5.33616304e-01
6.49027377e-02 -7.27891922e-01 -5.26497543e-01 9.60721225e-02
2.60071605e-02 6.19233191e-01 -1.80318713e-01 -6.70109630e-01
-9.54274610e-02 5.15456796e-01 1.67742908e-01 1.05504461e-01
1.21745217e+00 -4.43169743e-01 3.30042154e-01 2.99519211e-01
1.10134733e+00 -6.67919338e-01 -1.74167764e+00 -1.09708436e-01
4.15039718e-01 -2.10541517e-01 2.38550261e-01 -6.74258411e-01
-1.04104340e+00 6.41486943e-01 8.50279570e-01 2.32064784e-01
1.02729380e+00 -1.43749854e-02 7.83401668e-01 6.17204547e-01
5.55761337e-01 -1.55732048e+00 8.91938508e-02 5.70226908e-01
6.14889562e-01 -1.29876208e+00 -1.14844136e-01 -7.21708596e-01
-5.47011137e-01 9.54320729e-01 9.54091251e-01 -5.80753684e-01
8.68025243e-01 3.61753330e-02 -3.70952845e-01 -4.99172568e-01
-3.54295075e-01 -7.34373391e-01 2.75766313e-01 7.17377126e-01
-3.80718529e-01 3.32722545e-01 -3.96923274e-02 2.81618387e-01
1.10504769e-01 4.38115299e-02 2.23940954e-01 6.23029947e-01
-9.10383642e-01 -4.65629190e-01 -1.50270492e-01 3.42634529e-01
1.68830916e-01 1.44718438e-01 8.73916522e-02 1.11139238e+00
2.11879835e-01 8.62074018e-01 5.31289637e-01 -2.27086455e-01
3.37059826e-01 2.05960460e-02 1.78564176e-01 -5.09443581e-01
-3.51899385e-01 -7.36225918e-02 2.94099599e-01 -7.07725465e-01
-3.70127559e-01 -4.55924183e-01 -1.47451699e+00 3.94518822e-01
-4.67559069e-01 1.66860208e-01 7.84916759e-01 1.03583765e+00
7.11923361e-01 6.33679748e-01 3.99053335e-01 -1.20039678e+00
-8.42570662e-02 -9.71941054e-01 -5.87476194e-01 4.40556914e-01
1.11279577e-01 -1.15719032e+00 -2.71373510e-01 -1.25094473e-01] | [8.138388633728027, -2.658522367477417] |
3a744a3b-a4a6-4df2-b6db-2f3d64d15c2d | planning-automated-driving-with-accident | 2301.10892 | null | https://arxiv.org/abs/2301.10892v1 | https://arxiv.org/pdf/2301.10892v1.pdf | Planning Automated Driving with Accident Experience Referencing and Common-sense Inferencing | Although a typical autopilot system far surpasses humans in term of sensing accuracy, performance stability and response agility, such a system is still far behind humans in the wisdom of understanding an unfamiliar environment with creativity, adaptivity and resiliency. Current AD brains are basically expert systems featuring logical computations, which resemble the thinking flow of a left brain working at tactical level. A right brain is needed to upgrade the safety of automated driving vehicle onto next generation by making intuitive strategical judgements that can supervise the tactical action planning. In this work, we present the concept of an Automated Driving Strategical Brain (ADSB): a framework of a scene perception and scene safety evaluation system that works at a higher abstraction level, incorporating experience referencing, common-sense inferring and goal-and-value judging capabilities, to provide a contextual perspective for decision making within automated driving planning. The ADSB brain architecture is made up of the Experience Referencing Engine (ERE), the Common-sense Referencing Engine (CIE) and the Goal and Value Keeper (GVK). 1,614,748 cases from FARS/CRSS database of NHTSA in the period 1975 to 2018 are used for the training of ERE model. The kernel of CIE is a trained model, COMET-BART by ATOMIC, which can be used to provide directional advice when tactical-level environmental perception conclusions are ambiguous; it can also use future scenario models to remind tactical-level decision systems to plan ahead of a perceived hazard scene. GVK can take in any additional expert-hand-written rules that are of qualitative nature. Moreover, we believe that with good scalability, the ADSB approach provides a potential solution to the problem of long-tail corner cases encountered in the validation of a rule-based planning algorithm. | ['Boqi Li', 'Hong Wang', 'Guoxi Chen', 'Ji Li', 'Shaobo Qiu'] | 2023-01-26 | null | null | null | null | ['common-sense-reasoning'] | ['reasoning'] | [-1.25866815e-01 4.66728568e-01 1.46337897e-01 -6.58505619e-01
-8.73766169e-02 -5.55337965e-01 7.46144593e-01 1.31448343e-01
-3.55822831e-01 4.45165455e-01 4.32987094e-01 -9.89730358e-01
-7.75485694e-01 -8.62871408e-01 -4.07648832e-01 -3.16868246e-01
3.50331631e-03 4.96054024e-01 1.28074870e-01 -9.95308995e-01
5.20995677e-01 7.51037598e-01 -1.80427730e+00 2.33336136e-01
7.11613417e-01 1.23264217e+00 2.75813192e-01 4.86044228e-01
3.25502843e-01 1.27089810e+00 -2.86351144e-01 -1.19124249e-01
4.17923510e-01 -8.65746439e-02 -7.12815285e-01 -3.91947746e-01
-1.73941359e-01 -8.68084952e-02 1.75392143e-02 6.75715327e-01
3.57548773e-01 3.75667751e-01 5.79504788e-01 -1.47110689e+00
-3.72534662e-01 3.86727512e-01 3.47033530e-01 3.08714300e-01
3.29008520e-01 8.46130311e-01 4.36382860e-01 -6.03809297e-01
5.77724099e-01 1.27714229e+00 6.40010118e-01 3.09031546e-01
-8.11289251e-01 -6.31965756e-01 2.07318291e-01 7.58543611e-01
-1.37920237e+00 -4.91449535e-01 5.26453435e-01 -8.31028342e-01
1.41943717e+00 1.70118868e-01 1.01700580e+00 8.75751436e-01
6.92852378e-01 1.01153120e-01 1.42980468e+00 -7.43928924e-02
7.60429382e-01 3.17644894e-01 9.50502232e-02 2.72202253e-01
2.35340461e-01 8.62502575e-01 -7.28517592e-01 2.29707986e-01
1.96749955e-01 -3.17860574e-01 2.35971622e-02 -3.56647559e-02
-1.00282311e+00 6.31932318e-01 7.11837947e-01 3.07081461e-01
-9.14007902e-01 1.96346436e-02 3.41060758e-01 5.03182352e-01
-2.56184399e-01 9.49807286e-01 -2.35318169e-01 -1.58067226e-01
-7.86613047e-01 4.68061358e-01 7.38256931e-01 6.39323413e-01
5.98114252e-01 3.11556131e-01 -2.79634148e-01 1.73910130e-02
3.41787755e-01 5.01570642e-01 4.06985790e-01 -1.22458875e+00
-1.17423408e-01 7.81086206e-01 1.56465620e-01 -1.22073317e+00
-7.63129711e-01 -5.19856513e-01 -3.04809153e-01 9.87648010e-01
1.86198503e-02 -1.29032984e-01 -7.91231871e-01 1.48877919e+00
1.48922592e-01 -4.06077653e-02 5.83502769e-01 1.09655154e+00
3.86679977e-01 4.67383742e-01 3.88023525e-01 2.26740360e-01
1.56703031e+00 -5.50969303e-01 -5.86632967e-01 -5.81002295e-01
5.24272263e-01 -1.35846034e-01 9.05505061e-01 8.68256330e-01
-9.78693783e-01 -8.47384453e-01 -1.56992388e+00 6.51570410e-02
-7.66991198e-01 -4.09212112e-01 6.47206128e-01 4.78907585e-01
-1.21061194e+00 1.99315101e-01 -2.77834147e-01 -4.77458328e-01
2.87010908e-01 1.45099382e-03 -2.65673459e-01 1.65211916e-01
-1.58391798e+00 1.82070768e+00 8.17435205e-01 1.85299769e-01
-1.26519752e+00 -6.26713455e-01 -8.06895018e-01 7.06478655e-02
5.91479063e-01 -7.04857230e-01 1.15113497e+00 -8.68757486e-01
-1.36167002e+00 6.37941957e-01 2.38796264e-01 -9.93307889e-01
4.15538371e-01 1.22946866e-01 -9.58595395e-01 -6.49782792e-02
2.49521136e-01 6.97679579e-01 5.15596509e-01 -1.23786366e+00
-7.44523764e-01 -2.38088191e-01 1.83402747e-01 2.71864623e-01
5.32197416e-01 -2.22470373e-01 4.96223480e-01 -1.48243010e-01
-6.72422908e-03 -8.26427221e-01 -4.34390306e-01 -3.82361412e-01
5.21618761e-02 -2.59943306e-01 4.88350183e-01 -5.22344589e-01
1.22456038e+00 -2.36263537e+00 -2.07709834e-01 4.21470881e-01
1.43089488e-01 8.43541250e-02 1.25520095e-01 5.84476411e-01
-1.98326632e-01 -9.66527909e-02 1.70424789e-01 3.69418055e-01
3.36804926e-01 1.48786798e-01 -7.36546218e-01 -1.81094147e-02
2.25817278e-01 6.76941872e-01 -1.03541994e+00 -6.20675348e-02
4.51415688e-01 -3.61127332e-02 -4.71738368e-01 9.30522010e-03
-4.62088659e-02 3.35431427e-01 -4.57994133e-01 3.56229484e-01
3.25586885e-01 3.93318474e-01 1.05206296e-01 4.97001000e-02
-7.38566816e-01 3.43138278e-01 -1.08112812e+00 1.47108483e+00
-5.49673021e-01 7.49812007e-01 -8.10702890e-02 -7.10363746e-01
1.19650614e+00 1.98049158e-01 -3.50889452e-02 -1.19475353e+00
3.96866053e-01 2.30402768e-01 4.16666329e-01 -7.17061520e-01
5.47925770e-01 -5.19842267e-01 -3.54465246e-01 5.89847937e-02
-1.56002358e-01 -6.21151507e-01 -5.63959666e-02 8.10873210e-02
1.18877316e+00 -1.65961701e-02 4.83195394e-01 -5.97543478e-01
6.08039081e-01 5.29314876e-01 6.65170550e-01 5.11637747e-01
-4.74921435e-01 -1.52564555e-01 5.17235577e-01 -7.36917555e-01
-7.20043480e-01 -1.01036143e+00 -1.18697815e-01 8.50676775e-01
1.90082803e-01 -2.64455408e-01 -6.99701428e-01 1.65855363e-02
-1.57797471e-01 2.02243900e+00 -5.85620522e-01 -6.15757346e-01
-3.88544090e-02 1.01234220e-01 4.80058640e-01 5.89625180e-01
8.20927680e-01 -1.06567752e+00 -1.57786942e+00 2.47671470e-01
9.38689187e-02 -8.81962299e-01 3.13638419e-01 3.57246131e-01
-2.64236450e-01 -8.67036104e-01 3.47407579e-01 -2.73622900e-01
1.37175053e-01 1.09025083e-01 8.98647606e-01 6.27401620e-02
1.23548627e-01 3.34780931e-01 -2.60148406e-01 -1.03708375e+00
-6.39397204e-01 -3.95582646e-01 3.38041514e-01 -2.10440472e-01
6.49021745e-01 -5.11507213e-01 -6.77682459e-01 5.59279621e-01
-6.60294950e-01 3.47262084e-01 6.74265265e-01 4.75625068e-01
3.05141687e-01 4.65736896e-01 8.98536026e-01 -1.08417146e-01
7.88001120e-01 -7.36925364e-01 -4.65197176e-01 4.48562130e-02
-1.01048160e+00 3.21055129e-02 3.30691695e-01 4.28991355e-02
-1.40569746e+00 -1.14690997e-01 1.20156454e-02 -1.77236527e-01
-4.87782747e-01 7.45464742e-01 -2.36174434e-01 1.47940949e-01
1.10127842e+00 2.61839805e-03 1.27089620e-01 1.39842793e-01
4.19616580e-01 7.01057136e-01 6.90847576e-01 -4.19019938e-01
5.40297389e-01 3.23050857e-01 6.36344329e-02 -4.41536367e-01
-4.05921876e-01 -1.16667531e-01 -5.04891872e-01 -9.13630366e-01
1.05240262e+00 -9.43821847e-01 -9.01747823e-01 1.72478184e-01
-8.59207511e-01 -6.73689783e-01 -6.05663478e-01 4.22950029e-01
-9.24039006e-01 -3.81748199e-01 2.72497118e-01 -9.30186927e-01
1.61306053e-01 -1.11192524e+00 5.21155953e-01 3.52715164e-01
-5.17112494e-01 -6.22231841e-01 -8.88973400e-02 5.20314038e-01
6.91567898e-01 3.94685596e-01 9.15104806e-01 -8.77371848e-01
-5.80974340e-01 -1.90933719e-01 9.21519399e-02 2.72004664e-01
-3.77142102e-01 -3.77285480e-01 -1.02482116e+00 2.67166615e-01
3.75344485e-01 -1.16758607e-01 4.17786330e-01 1.30450487e-01
6.46039486e-01 -1.92218825e-01 -1.48900852e-01 2.08874628e-01
1.26086640e+00 9.37428236e-01 1.01029480e+00 5.93785465e-01
-1.48182049e-01 1.16873813e+00 1.18281686e+00 3.78411651e-01
7.34966159e-01 4.84653294e-01 5.64041317e-01 2.23652318e-01
1.69094151e-03 -3.32091838e-01 5.91276348e-01 5.34879342e-02
-3.08259964e-01 7.14848340e-02 -1.35707712e+00 5.20671964e-01
-1.87717676e+00 -1.32150114e+00 -1.01008654e-01 2.05550694e+00
2.77789742e-01 6.36591434e-01 -1.12074241e-01 2.40147039e-01
1.65843651e-01 -1.41152784e-01 -6.69512033e-01 -1.04771423e+00
3.24980058e-02 -1.62776858e-01 6.30096376e-01 3.74150634e-01
-3.54290932e-01 1.06577718e+00 5.97920179e+00 7.51599908e-01
-1.04590321e+00 2.02391803e-01 4.82961744e-01 -4.28721105e-04
-3.45026404e-01 1.93856522e-01 -4.45474535e-01 9.32208449e-02
1.25260663e+00 -2.94448078e-01 7.43849993e-01 9.66872215e-01
6.49938703e-01 -6.96096301e-01 -9.99878824e-01 7.39618003e-01
3.38107944e-02 -1.37167382e+00 -3.45991462e-01 -1.02350324e-01
2.00340569e-01 -1.00352257e-01 2.21272632e-02 6.53800964e-01
4.14476186e-01 -1.29725969e+00 1.52412128e+00 1.12009704e+00
3.91741365e-01 -7.88305819e-01 7.72803366e-01 5.04069507e-01
-8.73779953e-01 -5.94278395e-01 -1.17079511e-01 -6.94977880e-01
4.09534752e-01 1.50648311e-01 -1.15174043e+00 4.63556230e-01
6.37983501e-01 1.72626644e-01 -5.77361166e-01 5.16641498e-01
-2.35736534e-01 3.40870947e-01 -2.19234005e-02 -1.40115947e-01
5.13114929e-01 -7.41114914e-02 7.97477186e-01 8.24486792e-01
3.50614250e-01 4.38197702e-01 -9.65897217e-02 9.51694191e-01
9.50392723e-01 -3.07481289e-01 -9.40638781e-01 7.62019828e-02
5.41653812e-01 1.15292215e+00 -7.36215830e-01 -4.23382252e-01
-2.90845484e-02 2.64794290e-01 -2.44758978e-01 4.02248204e-01
-8.66039693e-01 -1.18249983e-01 8.26342762e-01 4.25173908e-01
-7.78064355e-02 -2.20600992e-01 -6.41633034e-01 -2.16554090e-01
-3.39391530e-01 -9.61840808e-01 1.30867362e-01 -1.60960221e+00
-8.53564918e-01 7.60440409e-01 3.75721961e-01 -1.06710494e+00
-3.89426112e-01 -9.16701674e-01 -5.76746941e-01 8.80484045e-01
-1.17930377e+00 -1.03636968e+00 -6.11757994e-01 4.97262627e-01
4.14869785e-01 -4.22057182e-01 5.27024627e-01 -2.32727543e-01
1.23348031e-02 -1.59947470e-01 -7.45994329e-01 -3.57097328e-01
3.76722693e-01 -9.75461245e-01 1.08384542e-01 6.46112859e-01
-6.29028022e-01 5.02001822e-01 1.08257949e+00 -7.83556163e-01
-1.19562507e+00 -8.24229717e-01 7.49404788e-01 -6.55005395e-01
7.37365246e-01 -7.17398748e-02 -6.02136374e-01 5.96069992e-01
3.24044198e-01 -7.61685371e-01 7.31102347e-01 -1.26654297e-01
4.60416451e-02 -3.38810533e-01 -1.26594138e+00 8.97545755e-01
1.09682882e+00 -3.55390787e-01 -1.19391859e+00 -3.70201492e-03
5.90992033e-01 -1.97816402e-01 -5.48371792e-01 2.55546391e-01
4.40930814e-01 -1.24824381e+00 7.98993647e-01 -6.69023395e-01
1.93556994e-02 -7.37981677e-01 -4.48688895e-01 -1.34666777e+00
-5.24952590e-01 -5.85615337e-01 6.81966841e-01 8.17468405e-01
3.64321053e-01 -1.01229262e+00 -5.50516881e-02 9.49200690e-01
-8.47495139e-01 -5.85115910e-01 -1.24393237e+00 -7.62330711e-01
-1.68121964e-01 -1.18209231e+00 1.07618582e+00 6.74652278e-01
2.64306366e-01 3.14334959e-01 2.10718870e-01 2.08204135e-01
4.90265712e-02 -4.34348434e-01 6.29145086e-01 -1.29260564e+00
4.21012938e-02 -6.06749892e-01 -8.13390791e-01 -2.61773914e-01
6.29072040e-02 -9.05958831e-01 3.12117189e-01 -1.60965073e+00
-4.39574152e-01 -4.42952305e-01 -2.12514237e-01 5.35905898e-01
6.37908220e-01 -3.98030549e-01 3.19607824e-01 -1.46510273e-01
-1.53904751e-01 1.94982603e-01 1.02963758e+00 7.92353302e-02
-2.31974810e-01 -2.14696631e-01 -1.06485891e+00 1.02027428e+00
8.39081645e-01 -2.13759556e-01 -7.89279401e-01 7.51703978e-02
7.92710960e-01 2.27351829e-01 9.10842538e-01 -1.45107615e+00
5.73020279e-01 -6.15761042e-01 1.03051312e-01 -5.24900615e-01
2.70508051e-01 -1.06211197e+00 7.43465483e-01 6.95893884e-01
-2.16969475e-01 1.88645229e-01 4.23957020e-01 3.47779900e-01
-1.81919001e-02 -2.41532803e-01 4.86672133e-01 6.97818175e-02
-1.38645709e+00 -2.72813231e-01 -1.06030416e+00 -3.54500085e-01
1.46227348e+00 -6.82872891e-01 -4.18561190e-01 -3.74471992e-01
-6.83929443e-01 5.19761086e-01 3.70034128e-01 5.74893892e-01
6.12192810e-01 -1.18784118e+00 -5.29586852e-01 2.85905808e-01
2.06813276e-01 -1.78154886e-01 6.35489702e-01 1.07207263e+00
-5.33450782e-01 5.58875382e-01 -7.31406391e-01 -2.35514298e-01
-3.53540778e-01 7.84288347e-01 6.05037749e-01 9.45834517e-02
-4.20235068e-01 5.08155584e-01 2.21540123e-01 -6.40818700e-02
-2.63747185e-01 -4.00769532e-01 -4.98557866e-01 3.43104929e-01
6.19533598e-01 5.19952297e-01 2.35847592e-01 -8.17064226e-01
-5.84279120e-01 8.20757523e-02 6.54956281e-01 -5.81994891e-01
9.88575220e-01 -1.36351824e-01 1.84501380e-01 6.05634749e-01
3.98259573e-02 -2.24485591e-01 -1.15915453e+00 4.35431808e-01
5.63439094e-02 -2.19170734e-01 4.10794914e-01 -1.53531766e+00
-3.76749635e-01 8.24518085e-01 6.99215829e-01 2.44014606e-01
1.29881310e+00 -1.98880598e-01 2.63963670e-01 4.60686564e-01
6.79825127e-01 -1.61853516e+00 -1.36726007e-01 5.34940362e-01
1.53282833e+00 -9.42450702e-01 -3.01371753e-01 1.93354085e-01
-1.31389618e+00 9.58840966e-01 6.56699479e-01 -1.01677757e-02
5.81923366e-01 1.49447635e-01 8.01126212e-02 -5.57380140e-01
-1.05541670e+00 -3.85390550e-01 3.40718150e-01 7.87791312e-01
-2.26150692e-01 4.50364739e-01 -2.21628621e-01 1.13455045e+00
-6.95810080e-01 2.02403888e-01 3.55719030e-01 5.94907284e-01
-8.10096860e-01 -3.56850535e-01 -3.92321527e-01 1.53993517e-01
4.33329254e-01 4.25045975e-02 -3.74340415e-01 9.07766104e-01
7.74693668e-01 1.31905794e+00 2.73523688e-01 -7.69898772e-01
6.65728986e-01 3.27819109e-01 -6.60773227e-03 -3.35727602e-01
-7.05269516e-01 -5.21851480e-01 4.93307054e-01 -7.81748295e-01
-1.03467613e-01 -8.87450039e-01 -1.64401639e+00 -4.60154057e-01
2.50336885e-01 -1.18737696e-02 8.08284938e-01 1.15439868e+00
5.67619979e-01 8.24147284e-01 3.35489780e-01 -7.51305401e-01
-3.70335281e-01 -7.92167306e-01 -3.66322517e-01 -9.96472761e-02
4.94771376e-02 -9.57273304e-01 -2.77269572e-01 -1.83690861e-01] | [5.178920745849609, 1.258935809135437] |
338fecf0-5784-4e3b-acf5-ae8ee6336bac | depth-estimation-maps-of-lidar-and-stereo | 2212.11741 | null | https://arxiv.org/abs/2212.11741v1 | https://arxiv.org/pdf/2212.11741v1.pdf | Depth Estimation maps of lidar and stereo images | This paper as technology report is focusing on evaluation and performance about depth estimations based on lidar data and stereo images(front left and front right). The lidar 3d cloud data and stereo images are provided by ford. In addition, this paper also will explain some details about optimization for depth estimation performance. And some reasons why not use machine learning to do depth estimation, replaced by pure mathmatics to do stereo depth estimation. The structure of this paper is made of by following:(1) Performance: to discuss and evaluate about depth maps created from stereo images and 3D cloud points, and relationships analysis for alignment and errors;(2) Depth estimation by stereo images: to explain the methods about how to use stereo images to estimate depth;(3)Depth estimation by lidar: to explain the methods about how to use 3d cloud datas to estimate depth;In summary, this report is mainly to show the performance of depth maps and their approaches, analysis for them. | ['Luoyu Chen', 'Fei Wu'] | 2022-12-22 | null | null | null | null | ['stereo-depth-estimation'] | ['computer-vision'] | [-2.79758759e-02 -1.05403617e-01 -4.38559204e-02 -6.22277081e-01
-3.47516507e-01 -2.43589029e-01 4.59531367e-01 9.95714664e-02
-4.11290318e-01 8.37648630e-01 -6.69461265e-02 -4.51439828e-01
-1.67301297e-01 -1.14210951e+00 -2.68120319e-01 -5.76084733e-01
-1.70246035e-01 9.11576927e-01 1.56096548e-01 -6.05429187e-02
1.05531621e+00 1.10387051e+00 -2.04404807e+00 -3.57003994e-02
8.28974783e-01 1.10055614e+00 5.64556777e-01 8.57977867e-01
-5.96165240e-01 2.69249886e-01 -4.09134060e-01 -5.78710921e-02
4.71115977e-01 1.13561675e-01 -6.78404152e-01 4.71015424e-02
8.42864692e-01 -1.00644374e+00 1.31712526e-01 8.23256969e-01
7.02870607e-01 -2.91464299e-01 1.14208817e+00 -1.24453723e+00
2.42647246e-01 -1.31715074e-01 -5.69529891e-01 1.32833049e-01
7.07987964e-01 -9.93089229e-02 3.51326555e-01 -9.98447180e-01
3.82277191e-01 1.18323743e+00 6.50893927e-01 3.44122112e-01
-7.82164156e-01 -6.09376967e-01 -1.59327880e-01 2.08667666e-01
-1.29667497e+00 -1.79019812e-02 4.10231233e-01 -7.54264712e-01
1.39749837e+00 1.04997024e-01 1.20245326e+00 2.36965552e-01
3.40087354e-01 5.65638065e-01 1.24429798e+00 -6.46430135e-01
3.86924833e-01 3.46202105e-01 5.35188392e-02 5.46665609e-01
3.18702161e-01 5.29054046e-01 -3.67676705e-01 2.25014061e-01
9.68772829e-01 -4.77097750e-01 2.52147734e-01 -4.76739049e-01
-6.40519679e-01 1.07979631e+00 2.20174119e-01 4.84643728e-02
-6.65758625e-02 -7.82875195e-02 1.69416770e-01 1.57377198e-01
6.07556164e-01 1.13605142e-01 -6.69282079e-01 -3.38392705e-01
-9.80769873e-01 4.43206668e-01 5.18631041e-01 1.23185301e+00
1.20733058e+00 2.41216585e-01 8.64505529e-01 7.08809316e-01
3.09443414e-01 8.12609673e-01 2.98423201e-01 -1.33208358e+00
6.58395231e-01 3.92538279e-01 -7.98116550e-02 -6.57023191e-01
-5.84714651e-01 3.00747991e-01 -4.51476544e-01 1.04134917e+00
1.60391644e-01 -1.33276418e-01 -7.93274879e-01 4.74473953e-01
2.08923012e-01 -9.55418497e-02 1.06009953e-01 5.73429942e-01
1.28051376e+00 6.10788584e-01 -5.38147688e-01 -1.67987511e-01
8.07661057e-01 -5.54817915e-01 -5.30592382e-01 -3.86249095e-01
8.81125569e-01 -7.44093835e-01 7.41639912e-01 6.28857315e-01
-1.26248622e+00 -7.26858377e-01 -1.25738478e+00 -2.49947071e-01
-6.00855649e-01 -1.16873980e-01 5.58485210e-01 7.66695917e-01
-1.05227470e+00 8.40838730e-01 -7.20207691e-01 -4.96229768e-01
2.11381435e-01 4.46922064e-01 -2.44976372e-01 5.51024787e-02
-5.94526291e-01 1.31571352e+00 2.94809490e-01 -1.20679997e-01
-3.36328954e-01 -6.75824940e-01 -1.08933806e+00 -7.17625022e-01
-2.96625197e-01 -7.55295932e-01 1.28231609e+00 -1.29601836e-01
-1.39164460e+00 1.38335752e+00 -5.59615314e-01 -2.69838572e-01
5.64464152e-01 -4.35799837e-01 1.27414033e-01 1.66001260e-01
4.27342027e-01 1.10961688e+00 2.16294587e-01 -1.28252435e+00
-1.21422529e+00 -7.46703327e-01 -1.01096109e-02 4.63476628e-01
3.67710948e-01 -3.67254436e-01 -3.35601121e-01 2.56861299e-01
9.74353552e-01 -4.81091917e-01 -2.23300099e-01 1.17461860e-01
-1.84680417e-01 -1.40426233e-01 8.37067246e-01 -3.95957142e-01
7.89158165e-01 -1.80644989e+00 -2.76537776e-01 1.34895757e-01
3.24360542e-02 -1.39165193e-01 3.39365184e-01 4.80790257e-01
2.96636987e-02 1.11726396e-01 -1.62130669e-01 -6.04571581e-01
-2.59903193e-01 3.35858792e-01 -6.55630678e-02 3.99552792e-01
-1.44425452e-01 3.11387986e-01 -6.41460598e-01 -8.54797959e-01
1.17894566e+00 3.17096233e-01 -3.32146823e-01 -1.04005365e-02
3.51231456e-01 5.63462675e-01 -2.71263689e-01 1.00724447e+00
1.35215056e+00 7.16100752e-01 -3.44424337e-01 -2.94128805e-01
-8.66789043e-01 4.16095406e-01 -1.23672342e+00 1.45031357e+00
-5.77465594e-01 8.13337982e-01 -3.33448872e-02 -7.37792492e-01
1.44706750e+00 2.60232259e-02 7.03181565e-01 -8.77991319e-01
8.54182523e-03 6.53963089e-01 -5.70761800e-01 -6.78110600e-01
5.53553581e-01 -3.53540719e-01 2.76529849e-01 1.19064271e-01
-2.48025194e-01 -1.39239037e+00 1.00179933e-01 -1.87886775e-01
4.44913268e-01 4.34032768e-01 3.83782715e-01 -1.89276114e-01
6.45505190e-01 2.80320078e-01 1.99625030e-01 2.69476295e-01
-1.18660524e-01 6.94770515e-01 2.01339647e-01 -7.94474781e-01
-1.12731659e+00 -1.20933616e+00 -8.55822086e-01 1.01157211e-01
4.72645342e-01 -1.52376264e-01 -4.88068879e-01 -6.38971552e-02
4.95451719e-01 7.09076166e-01 -3.25951099e-01 6.72615945e-01
-2.36829266e-01 -4.04422164e-01 3.85081582e-02 8.10574591e-01
8.25335026e-01 -6.32714510e-01 -7.13006556e-01 -9.53939632e-02
-2.03646068e-02 -1.10034811e+00 4.35031295e-01 2.64629811e-01
-1.88567150e+00 -9.91341829e-01 -4.16942835e-01 -5.89029908e-01
3.56724143e-01 3.69774610e-01 1.10738528e+00 4.77728881e-02
-1.83402553e-01 5.25922179e-01 -2.54821591e-02 -9.29668427e-01
2.48027205e-01 -2.00133515e-03 2.83301412e-03 -1.02350795e+00
8.33650053e-01 -1.03081977e+00 -3.82400244e-01 3.82941693e-01
-3.41367394e-01 -6.61049038e-03 2.09735736e-01 1.01015426e-01
6.49876773e-01 1.16175085e-01 -3.19711447e-01 -6.20548129e-01
2.06041411e-01 -6.52231276e-02 -1.22169375e+00 -5.87770939e-01
-8.23663712e-01 -2.42398545e-01 -3.34594846e-01 5.43423057e-01
-7.97784328e-01 2.92497367e-01 -5.42875707e-01 -1.94899872e-01
-4.93849456e-01 -7.62416944e-02 -1.71352565e-01 -1.20659404e-01
6.16823018e-01 -3.53028253e-02 1.97138518e-01 -6.09597623e-01
3.74167152e-02 8.34476769e-01 4.66288924e-01 -4.10485834e-01
8.66445839e-01 9.48591292e-01 4.73572314e-01 -1.22968161e+00
-5.43997228e-01 -8.08800697e-01 -1.57931972e+00 -4.56081629e-01
1.01471114e+00 -1.10057616e+00 -7.54062295e-01 5.06951571e-01
-1.20114982e+00 -3.27519655e-01 -2.40154341e-01 7.55759358e-01
-9.84011590e-01 1.82138860e-01 -4.62728441e-01 -1.23701668e+00
-1.09480202e-01 -1.25778878e+00 1.48462903e+00 2.32350886e-01
-1.23678930e-01 -1.44495559e+00 2.04148695e-01 5.37498951e-01
-5.48746157e-03 4.69116092e-01 6.78268611e-01 4.92363870e-01
-8.17680955e-01 -2.17581660e-01 -2.63971895e-01 2.67338395e-01
3.37388329e-02 4.11674082e-01 -1.37436295e+00 5.57673499e-02
-6.72175363e-02 7.69735649e-02 5.49845815e-01 8.63464177e-01
1.00459063e+00 3.89783919e-01 -5.87593675e-01 9.91977155e-01
1.93370736e+00 4.14218336e-01 1.01690257e+00 9.57203507e-01
3.83270890e-01 1.09078002e+00 1.13856685e+00 4.32257801e-01
4.19719934e-01 6.06308818e-01 9.14962709e-01 -5.61443456e-02
4.75783534e-02 -3.27199161e-01 5.36379591e-02 8.86230469e-01
-6.20115936e-01 4.68035579e-01 -8.50285113e-01 3.40599239e-01
-1.29079938e+00 -7.27522552e-01 -8.17351222e-01 2.39417982e+00
-4.30354103e-02 2.32392415e-01 3.08712482e-01 6.13372028e-01
4.32398677e-01 -1.74122915e-01 -2.57690787e-01 -8.95476818e-01
-1.95657030e-01 3.06580156e-01 8.20815563e-01 1.11259472e+00
-9.94579077e-01 7.94608176e-01 7.47667408e+00 3.04958016e-01
-1.03289700e+00 -1.43925354e-01 5.99028692e-02 -1.86029654e-02
-1.99785024e-01 2.06466988e-01 -1.18554175e+00 2.87682146e-01
5.05870938e-01 -8.34754780e-02 -7.77236372e-02 1.07278895e+00
3.92557949e-01 -1.03756845e+00 -9.35676992e-01 1.42413330e+00
-7.37828240e-02 -1.23153782e+00 -9.14868265e-02 4.14151102e-01
8.15810025e-01 7.54161403e-02 -2.56347626e-01 9.11724754e-03
-7.93129653e-02 -9.08214211e-01 6.56584859e-01 4.09195036e-01
9.61442769e-01 -7.41418362e-01 1.09353888e+00 5.12399495e-01
-1.36949635e+00 8.17148313e-02 -9.71040308e-01 -7.55598962e-01
2.12373778e-01 1.01195705e+00 -9.34635997e-01 7.54830301e-01
1.09859467e+00 9.72842932e-01 -3.86585683e-01 1.31221306e+00
-4.38618541e-01 -3.18040460e-01 -3.39434087e-01 -1.02352954e-01
1.98724315e-01 -7.30684221e-01 2.05076590e-01 1.03688586e+00
5.52471042e-01 -1.96690053e-01 -2.28201658e-01 5.83453178e-01
8.53453338e-01 6.37125522e-02 -1.34152532e+00 6.06201887e-01
5.56893110e-01 9.59736764e-01 -4.09917384e-01 -2.30450332e-01
-4.32348251e-01 5.97201169e-01 -3.31214279e-01 -4.85322215e-02
-4.12643582e-01 -4.99235660e-01 8.16272736e-01 6.99194789e-01
-1.47159413e-01 -3.99274290e-01 -1.01388538e+00 -6.13199234e-01
-1.39483348e-01 7.10450485e-02 8.28575492e-02 -1.50784373e+00
-9.66914296e-01 1.63716674e-01 5.31740546e-01 -1.58767927e+00
-4.28263396e-01 -1.12330246e+00 -5.21976292e-01 1.22190630e+00
-1.71100640e+00 -5.47172606e-01 -7.26019084e-01 2.65634090e-01
5.52525222e-01 -6.74374476e-02 6.50418043e-01 3.83373350e-01
9.64905471e-02 -1.70856103e-01 -2.33052205e-02 -1.64401338e-01
3.14706504e-01 -1.52931821e+00 4.68196183e-01 7.62194470e-02
-4.65954632e-01 9.66338143e-02 6.11452699e-01 -7.88194537e-01
-9.34296489e-01 -5.72229564e-01 1.10776675e+00 -7.26584971e-01
1.90771781e-02 2.46317405e-03 -3.27301890e-01 6.25640154e-01
-1.02088280e-01 -5.06382346e-01 2.83225238e-01 1.16456486e-02
3.62286180e-01 -3.68399113e-01 -1.41589868e+00 2.89967448e-01
1.20103312e+00 -5.78823864e-01 -5.52139699e-01 3.27607214e-01
2.48479858e-01 -8.43632936e-01 -6.91569984e-01 6.53133273e-01
8.64821374e-01 -1.78780746e+00 1.13225329e+00 1.07561782e-01
7.19764054e-01 -3.38485330e-01 -3.52450520e-01 -7.63046443e-01
4.15545218e-02 3.37418988e-02 3.65412682e-01 7.13805199e-01
2.34253943e-01 -8.49903226e-01 1.29267848e+00 1.88436165e-01
-4.43909466e-01 -5.94392240e-01 -1.12667584e+00 -7.58604884e-01
4.18255001e-01 -9.24196959e-01 5.22802889e-01 7.03365862e-01
-1.57504171e-01 2.93894887e-01 2.13693917e-01 3.18230599e-01
8.17358077e-01 7.85355866e-02 1.21921670e+00 -1.54145849e+00
3.24495882e-01 -3.20989281e-01 -9.92861688e-01 -1.23071098e+00
-2.41421342e-01 -5.28429925e-01 -2.19887763e-01 -2.14134383e+00
-2.12853253e-01 -5.09783030e-01 8.31013620e-01 -5.25925420e-02
4.56668377e-01 1.93336681e-01 -5.23029044e-02 2.41423443e-01
4.47936624e-01 1.55293629e-01 1.34148300e+00 5.31432554e-02
-1.23179749e-01 4.36955422e-01 -9.13468972e-02 1.05487812e+00
7.42942393e-01 -2.83865035e-01 -5.03619373e-01 -5.99495769e-01
2.66514868e-01 2.78908700e-01 1.48672923e-01 -1.38783169e+00
7.52374083e-02 -2.06776530e-01 5.78522980e-01 -1.79749727e+00
9.72490549e-01 -9.19003308e-01 -2.02461898e-01 5.41866004e-01
5.93384326e-01 3.70540261e-01 4.32185799e-01 -1.34578630e-01
-3.44051540e-01 -8.11480463e-01 9.28288341e-01 -3.97753745e-01
-8.99716377e-01 1.67138889e-01 -4.53506470e-01 -5.97242951e-01
1.07719243e+00 -1.37214518e+00 2.06099427e-03 -6.55815601e-01
-8.34706545e-01 1.74608380e-01 6.62121594e-01 -1.92414507e-01
8.28655243e-01 -1.35350883e+00 -4.86218452e-01 3.22815597e-01
-1.00435071e-01 4.05793518e-01 1.12527646e-01 6.07753992e-01
-1.27421045e+00 7.93980300e-01 -3.44355136e-01 -1.15464103e+00
-1.40780449e+00 2.02751234e-01 6.62548184e-01 2.58543611e-01
-4.12643582e-01 9.07883048e-01 2.02989593e-01 -8.62288892e-01
8.37172717e-02 -5.84365785e-01 -2.82014430e-01 2.14675255e-02
3.45622152e-01 7.81039178e-01 4.47679281e-01 -2.61838347e-01
-3.34187925e-01 1.66283965e+00 5.91336846e-01 -4.55715984e-01
1.08828425e+00 -3.79625350e-01 -1.00527093e-01 7.44086027e-01
1.09051609e+00 -1.70980394e-02 -1.02054000e+00 4.09073174e-01
-2.33427480e-01 -9.14628804e-01 2.19643295e-01 -4.77943242e-01
-9.41566169e-01 1.69945455e+00 9.16674793e-01 -3.14447731e-02
1.07217181e+00 -1.87018394e-01 2.77959436e-01 2.49459133e-01
6.93351626e-01 -1.23904371e+00 -9.59944651e-02 6.73393309e-01
7.84331918e-01 -1.22755814e+00 4.49014336e-01 -8.75410616e-01
-2.20681772e-01 1.56071961e+00 5.78501284e-01 -3.46081741e-02
8.74546826e-01 6.25378251e-01 2.25979879e-01 -2.97270596e-01
-2.03063235e-01 -3.97615850e-01 -1.94800377e-01 1.30095255e+00
3.84013206e-01 -1.67139128e-01 -1.15729868e-01 -4.63663548e-01
-8.58259022e-01 -7.90229663e-02 4.20946300e-01 1.09436131e+00
-1.01438451e+00 -1.20663714e+00 -8.71715128e-01 4.48320448e-01
2.72891939e-01 7.37855909e-04 -4.95049089e-01 1.24070108e+00
6.19922221e-01 9.05308425e-01 7.24081218e-01 -5.66095650e-01
6.24458373e-01 -1.91620842e-01 8.70427072e-01 -7.01958001e-01
-9.76080894e-02 -1.59139559e-01 1.64984345e-01 -5.89327991e-01
-5.47148526e-01 -8.34951341e-01 -1.18949974e+00 -7.57494628e-01
-2.97516793e-01 -4.33591269e-02 1.45110893e+00 7.76595533e-01
-3.10970753e-01 2.07487959e-02 7.07306981e-01 -1.19795549e+00
3.81139815e-02 -8.88710856e-01 -1.00922728e+00 -3.37227643e-01
8.42389241e-02 -8.97283435e-01 -7.89925694e-01 -2.18500406e-01] | [7.898592948913574, -2.539938449859619] |
c744022b-f9e7-48b0-b86a-031474c114de | liquid-theory-analogy-of-direct-coupling | 1510.06794 | null | http://arxiv.org/abs/1510.06794v2 | http://arxiv.org/pdf/1510.06794v2.pdf | Liquid-theory analogy of direct-coupling analysis of multiple-sequence alignment and its implications for protein structure prediction | The direct-coupling analysis is a powerful method for protein contact
prediction, and enables us to extract "direct" correlations between distant
sites that are latent in "indirect" correlations observed in a protein
multiple-sequence alignment. I show that the direct correlation can be obtained
by using a formulation analogous to the Ornstein-Zernike integral equation in
liquid theory. This formulation intuitively illustrates how the indirect or
apparent correlation arises from an infinite series of direct correlations, and
provides interesting insights into protein structure prediction. | [] | 2015-11-10 | null | null | null | null | ['multiple-sequence-alignment'] | ['medical'] | [ 4.62307781e-01 -1.23065757e-02 -4.88701254e-01 -4.07927901e-01
-4.08482254e-01 -6.26951396e-01 4.55096871e-01 2.86821157e-01
-2.16337532e-01 1.23526120e+00 3.43219563e-02 -6.47079825e-01
-4.57508117e-01 -5.09043813e-01 -7.66525745e-01 -1.28623319e+00
-4.75037217e-01 6.17275894e-01 4.16805446e-01 -4.20859993e-01
3.68085831e-01 5.63429594e-01 -9.52258885e-01 2.66270250e-01
9.55996752e-01 3.89966995e-01 2.34566376e-01 7.00385869e-01
-6.47521168e-02 4.50447798e-01 -3.63166511e-01 -3.47110152e-01
-2.93109834e-01 -7.32784986e-01 -1.16736364e+00 -5.98197281e-01
-1.05133988e-01 1.43220618e-01 1.04788411e-02 9.12625074e-01
8.23438093e-02 -3.31679098e-02 9.78287518e-01 -9.15048897e-01
-6.17609143e-01 3.44824791e-03 -4.95127499e-01 2.82357931e-01
9.96992528e-01 -1.77847564e-01 1.56146657e+00 -7.86035240e-01
9.88046467e-01 1.13006735e+00 7.14897752e-01 2.37223566e-01
-1.76122808e+00 -1.13076091e-01 -3.05908382e-01 1.15964666e-01
-1.10971761e+00 1.78559393e-01 4.11249042e-01 -3.54921401e-01
1.59967041e+00 6.01906717e-01 8.07822645e-01 8.74278903e-01
9.98532534e-01 6.31092787e-01 1.11950982e+00 -6.50542676e-01
-8.11193734e-02 -2.25401372e-01 5.72401524e-01 4.96706516e-01
1.12227187e-01 2.82373309e-01 -6.96079671e-01 -1.21867371e+00
7.15616822e-01 1.42481849e-01 -4.58755732e-01 -6.52545273e-01
-1.13027012e+00 8.01054716e-01 7.47066513e-02 3.40906203e-01
-1.85438588e-01 -7.26147518e-02 9.02848542e-02 4.29287821e-01
4.07361060e-01 5.87205708e-01 -8.12487483e-01 -1.63482785e-01
-4.68419701e-01 4.72192526e-01 1.26730216e+00 6.51347697e-01
8.94693077e-01 -8.02812636e-01 4.94525790e-01 5.20906627e-01
3.05908680e-01 4.63875711e-01 -6.82603642e-02 -6.93076134e-01
6.76103607e-02 1.55222714e-01 4.60282028e-01 -6.64341390e-01
-3.11306417e-01 1.92018136e-01 -5.33025861e-01 1.26040936e-01
5.59685409e-01 2.54053444e-01 -3.63745272e-01 1.54967701e+00
1.69606864e-01 -7.67882541e-02 5.61608039e-02 5.33390343e-01
2.19559088e-01 5.86843729e-01 1.22582518e-01 -8.89576435e-01
1.06370497e+00 -7.42116213e-01 -1.02710140e+00 3.02124143e-01
7.90584624e-01 -9.94487524e-01 4.66124892e-01 2.35657007e-01
-7.85362899e-01 -9.05296858e-03 -8.46193433e-01 -2.50291210e-02
-3.14531714e-01 -5.10499299e-01 1.15634418e+00 -3.43440287e-02
-8.67620587e-01 1.31532514e+00 -8.46295953e-01 -4.78535801e-01
-4.76408541e-01 7.28716791e-01 -6.21576607e-01 4.51807350e-01
-1.19481468e+00 1.06617582e+00 6.07649647e-02 -7.48272091e-02
9.25671384e-02 -5.04381657e-01 -1.62499264e-01 -2.67369330e-01
8.73445570e-02 -3.61228704e-01 9.19725835e-01 -7.55800903e-01
-1.12283587e+00 9.44257557e-01 -9.00380671e-01 -1.43957645e-01
-2.16538414e-01 -3.29755187e-01 -4.03018922e-01 3.70455980e-01
-7.27174804e-02 -2.07751207e-02 1.58791274e-01 -9.51126099e-01
1.10595115e-01 -3.86066169e-01 -1.69523939e-01 1.61367416e-01
4.38111037e-01 3.38393569e-01 3.61170292e-01 -5.66679418e-01
4.51478273e-01 -1.12271559e+00 -5.60148001e-01 -5.92138208e-02
-2.96099216e-01 -2.88252085e-01 5.83172381e-01 -5.02285063e-01
8.88141751e-01 -1.43108523e+00 4.37639356e-01 6.77017748e-01
3.56145829e-01 9.02008712e-02 4.04332504e-02 1.20081854e+00
-7.42231309e-01 5.64406291e-02 -7.86865950e-02 6.40526056e-01
-3.50562155e-01 4.07962382e-01 -2.97810167e-01 4.52055603e-01
2.86203265e-01 1.00542605e+00 -1.00969839e+00 -2.35746667e-01
-1.02718353e-01 4.01699781e-01 -5.90109408e-01 3.61617446e-01
-9.39007476e-02 5.70311546e-01 -6.84809446e-01 2.90621221e-01
8.58513772e-01 -6.47142529e-01 9.59961474e-01 -1.85175985e-01
-2.13398263e-01 6.48352742e-01 -3.62816751e-01 1.15932906e+00
4.69980210e-01 4.97789055e-01 -8.34087059e-02 -1.07523704e+00
9.61130559e-01 3.99103969e-01 9.14887369e-01 -2.67117918e-01
-2.58842885e-01 1.98090702e-01 1.81397900e-01 -2.92876273e-01
1.38360009e-01 -3.28949362e-01 3.45803201e-01 4.41244453e-01
3.78841981e-02 2.90409476e-01 -1.67620510e-01 2.54758090e-01
1.13391602e+00 4.65436548e-01 7.62012303e-01 -8.87883842e-01
5.38782716e-01 1.41846821e-01 4.06559318e-01 4.96827841e-01
-1.35875400e-02 1.52238250e-01 8.34883809e-01 -6.52006090e-01
-1.04802561e+00 -1.25424492e+00 -4.34753180e-01 7.41360962e-01
4.24665481e-01 -8.29412937e-01 -7.51109719e-01 -1.82543039e-01
5.75245433e-02 -1.68753453e-02 -4.62392211e-01 -1.89932451e-01
-6.07472479e-01 -1.10103583e+00 3.58881116e-01 3.51592630e-01
-1.45734921e-01 -9.44555581e-01 2.20737476e-02 3.39931250e-01
-3.82757157e-01 -5.95758617e-01 -3.39684993e-01 9.74543214e-01
-1.12714767e+00 -1.33057976e+00 -5.21298528e-01 -4.87693965e-01
5.35092354e-01 3.26414138e-01 1.09671533e+00 3.30273092e-01
-4.70223010e-01 -5.63323461e-02 -1.54172286e-01 1.95261091e-01
-4.00685519e-01 -2.54985362e-01 4.24613416e-01 -6.93097413e-01
1.31555688e+00 -8.48616481e-01 -4.90853190e-01 4.94722664e-01
-6.06956780e-01 -2.35121429e-01 2.35817388e-01 1.23345780e+00
7.45946586e-01 -4.63216275e-01 2.15564802e-01 -7.48745978e-01
7.55220175e-01 -2.21552595e-01 -5.12004077e-01 5.17525971e-01
-7.31769800e-01 6.99375749e-01 4.17060256e-01 -3.84701997e-01
-9.27345335e-01 2.66088918e-02 -2.31310621e-01 3.55306178e-01
-6.42872974e-02 4.88997251e-01 -1.14909373e-01 -5.11041701e-01
4.56631511e-01 3.51012498e-01 8.64425898e-02 -6.84479654e-01
5.77422902e-02 4.24702406e-01 2.03597873e-01 -9.12609458e-01
3.71627986e-01 3.04253012e-01 2.18938559e-01 -1.08367765e+00
-5.06077051e-01 -7.70354927e-01 -1.20643187e+00 2.58459181e-01
9.77092624e-01 -3.57102960e-01 -1.35489345e+00 1.40006125e-01
-1.29433095e+00 -2.77318861e-02 1.00186586e-01 7.60627449e-01
-8.73269320e-01 8.24518204e-01 -9.40179586e-01 -7.40142465e-01
3.85752842e-02 -1.00453281e+00 8.32985759e-01 1.05677925e-01
-6.36862218e-01 -1.33915532e+00 6.30116165e-01 1.42294362e-01
-5.41093834e-02 1.27604157e-01 1.14527321e+00 -5.59969127e-01
-5.22386491e-01 1.00119635e-01 -1.84490848e-02 -1.26969293e-01
9.70996469e-02 2.77572900e-01 -3.84402275e-01 -7.57920891e-02
-6.22712746e-02 1.26864295e-02 6.83344483e-01 3.99841726e-01
5.16095340e-01 -1.49706274e-01 -8.35668683e-01 3.74546230e-01
1.33506250e+00 3.27972502e-01 6.74729228e-01 1.72653839e-01
4.66561675e-01 4.70366389e-01 7.84432054e-01 5.02432734e-02
-2.79675752e-01 8.35936427e-01 -2.74973691e-01 -8.94574821e-02
6.88320816e-01 -3.96011680e-01 1.82979941e-01 9.07770216e-01
-8.50547016e-01 2.31999546e-01 -7.81114280e-01 -4.63232845e-02
-2.23378420e+00 -1.35960793e+00 -7.40020514e-01 2.18394852e+00
1.20942783e+00 5.53198974e-04 7.78317750e-02 -3.01450551e-01
4.28311408e-01 -6.47112280e-02 -4.85313237e-01 -8.10961127e-01
-4.96555209e-01 5.45798779e-01 7.89549708e-01 8.48456740e-01
-6.58464491e-01 9.04338896e-01 9.53297901e+00 7.12326825e-01
-5.41414559e-01 -1.76297218e-01 1.27970725e-01 5.59800088e-01
-5.24914682e-01 5.47022283e-01 -7.20620096e-01 7.67180920e-02
1.03728974e+00 -1.78962290e-01 2.43575752e-01 6.34837210e-01
1.75026923e-01 -4.58716035e-01 -1.21506071e+00 4.78807569e-01
-6.05836570e-01 -1.39185286e+00 -2.93281097e-02 5.48374653e-01
6.85606837e-01 -6.39697760e-02 -2.39055291e-01 -6.77825749e-01
2.84438431e-01 -1.02332628e+00 -3.08461815e-01 7.72873104e-01
4.40161526e-01 -7.84850299e-01 5.71656048e-01 3.59716594e-01
-1.30212080e+00 4.75655288e-01 -7.67331123e-01 -3.54699224e-01
2.58820772e-01 7.74151623e-01 -4.23824281e-01 3.96679401e-01
3.69711399e-01 6.72296047e-01 1.16387509e-01 6.14486516e-01
-7.22953603e-02 3.16345870e-01 -3.10609221e-01 -3.20736796e-01
5.30585535e-02 -1.27205634e+00 6.34991467e-01 1.13291156e+00
-2.84698725e-01 5.83896518e-01 -2.13824615e-01 8.48216414e-01
5.44916272e-01 1.60923839e-01 -7.46720731e-01 -3.23117554e-01
-6.11428283e-02 8.14840496e-01 -7.29635835e-01 -1.66156918e-01
-5.35251856e-01 1.11117005e+00 4.32558924e-01 6.52059734e-01
-5.19898772e-01 -4.07514036e-01 1.07024693e+00 5.81329949e-02
1.22932330e-01 -4.94983703e-01 2.91681617e-01 -1.39430630e+00
-2.47201398e-02 -8.38115573e-01 -1.20737053e-01 -7.10062206e-01
-1.31645763e+00 3.10818046e-01 1.15239240e-01 -8.05892467e-01
-2.98021108e-01 -1.01465905e+00 -4.16530430e-01 1.09827054e+00
-9.23820555e-01 -6.02962434e-01 6.09400153e-01 3.78202558e-01
-2.37600982e-01 1.13896325e-01 1.49727881e+00 -5.95089383e-02
1.91552215e-03 3.63790393e-01 7.81901717e-01 -3.79957587e-01
7.46765971e-01 -1.35105574e+00 4.18251634e-01 5.17370142e-02
-1.50072187e-01 1.38060284e+00 1.15705347e+00 -9.55113232e-01
-1.27301872e+00 -5.88610135e-02 1.31568921e+00 -5.31480372e-01
1.08569098e+00 -2.13998556e-01 -1.21558428e+00 6.20119512e-01
1.59560859e-01 -2.01382622e-01 1.25849223e+00 4.69756097e-01
-3.94410133e-01 5.14434040e-01 -9.28376198e-01 5.30487716e-01
1.30294609e+00 -9.90982056e-01 -6.95478678e-01 6.94861412e-01
6.26785338e-01 2.62133162e-02 -1.21707845e+00 2.59819090e-01
9.71482277e-01 -1.07393873e+00 1.23854244e+00 -1.20016551e+00
2.84559757e-01 -4.44439054e-02 -1.93475708e-01 -8.17001402e-01
-5.36817551e-01 -8.71227443e-01 2.81482637e-02 2.71428525e-01
6.50049806e-01 -8.87150824e-01 7.44904160e-01 7.36861944e-01
4.22178447e-01 -9.99212444e-01 -8.16057920e-01 -7.35450983e-01
3.23304117e-01 3.82264912e-01 1.73977479e-01 1.04071450e+00
9.89253104e-01 4.24320102e-01 -4.02678192e-01 -1.12269253e-01
6.99031174e-01 2.72244960e-01 3.33227128e-01 -1.27446604e+00
-7.09279001e-01 3.24566476e-02 -4.99913216e-01 -1.69876003e+00
2.10775957e-01 -5.37357509e-01 -2.42673099e-01 -8.22386622e-01
7.35948324e-01 -2.13406399e-01 -3.42178166e-01 1.20606281e-01
-5.44327833e-02 -1.60553768e-01 -4.32431877e-01 5.85806966e-01
-3.71780813e-01 3.02507222e-01 1.26469481e+00 2.12654158e-01
-3.59327830e-02 1.43927664e-01 -2.26997793e-01 8.37148249e-01
4.86680746e-01 -5.70227921e-01 -5.64780533e-02 4.49024945e-01
7.70966530e-01 4.28529650e-01 2.31720492e-01 -2.46117413e-01
5.52865537e-03 -5.50230980e-01 1.98899671e-01 -9.76681709e-01
4.33742344e-01 -6.24742568e-01 5.04586995e-01 6.38598323e-01
-4.36796516e-01 2.27277771e-01 -2.54780263e-01 7.50792146e-01
-2.40723789e-01 -3.39270979e-01 5.55566549e-01 -9.00063068e-02
-1.24433547e-01 -5.47715649e-02 -6.61075830e-01 -4.24939871e-01
6.51175141e-01 -2.10054189e-01 -2.57935941e-01 -2.11192161e-01
-1.24924839e+00 -1.83812737e-01 7.55612373e-01 -1.46689087e-01
5.56395352e-01 -1.31544781e+00 5.06406650e-02 4.09001708e-02
-5.92663176e-02 -8.55255902e-01 -3.02864999e-01 1.06980002e+00
-7.36244440e-01 8.64782929e-01 -2.42626309e-01 -6.19862914e-01
-1.82228613e+00 5.28130352e-01 3.84130299e-01 -5.82537353e-01
-4.33696717e-01 6.64091051e-01 3.14284831e-01 -4.43719953e-01
-4.74710912e-01 1.40290996e-02 2.07553376e-02 -3.98408860e-01
5.51855624e-01 9.75631475e-02 -3.58885825e-01 -8.22893143e-01
-6.03849530e-01 9.56098974e-01 -3.15328240e-01 7.65110692e-03
1.32106519e+00 -1.12373427e-01 -8.82935405e-01 5.44990063e-01
1.50709546e+00 -1.51379108e-02 -7.22140670e-01 -2.55072802e-01
3.68142992e-01 -2.39545345e-01 -6.97889805e-01 -4.31748182e-01
-1.03916638e-01 7.45508015e-01 2.19606519e-01 2.56257117e-01
4.62976754e-01 2.97287285e-01 7.36771524e-01 1.01720643e+00
6.15327597e-01 -9.14655566e-01 -7.21999258e-02 5.76700866e-01
5.37830472e-01 -1.08603323e+00 2.94231802e-01 -8.68035257e-01
-3.21194410e-01 1.30064356e+00 1.51689142e-01 -1.98577806e-01
9.35885489e-01 5.64860404e-01 -1.68038666e-01 -4.71146524e-01
-9.66626763e-01 8.52155387e-02 2.71337003e-01 6.77164912e-01
1.06428039e+00 2.24011362e-01 -1.03211570e+00 1.85278550e-01
4.12709936e-02 -2.33361706e-01 2.11680710e-01 1.00856245e+00
-6.12180173e-01 -1.90396023e+00 -2.13348329e-01 2.95348972e-01
-6.36222899e-01 -3.41760188e-01 -8.97770405e-01 4.14762199e-01
-3.78742099e-01 7.19405293e-01 -1.47359639e-01 -2.10955292e-01
-1.47828251e-01 2.79729217e-01 7.84431100e-01 -1.87297300e-01
-1.67063192e-01 3.93426657e-01 2.00255916e-01 -7.68862605e-01
-9.84714150e-01 -4.74624723e-01 -1.54902661e+00 -7.12329984e-01
-8.79354894e-01 9.66251493e-01 3.02385896e-01 1.15435708e+00
1.01050399e-01 -6.50798678e-02 3.79936665e-01 -6.62679374e-01
-5.00702977e-01 -6.69632733e-01 -7.96487272e-01 3.24612141e-01
3.10948998e-01 -7.04180896e-01 -4.65404898e-01 9.16192830e-02] | [4.744943618774414, 5.308642387390137] |
abea32d8-91d0-4e02-8b9b-65da10e54dd4 | analyzing-bias-in-diffusion-based-face | 2305.06402 | null | https://arxiv.org/abs/2305.06402v1 | https://arxiv.org/pdf/2305.06402v1.pdf | Analyzing Bias in Diffusion-based Face Generation Models | Diffusion models are becoming increasingly popular in synthetic data generation and image editing applications. However, these models can amplify existing biases and propagate them to downstream applications. Therefore, it is crucial to understand the sources of bias in their outputs. In this paper, we investigate the presence of bias in diffusion-based face generation models with respect to attributes such as gender, race, and age. Moreover, we examine how dataset size affects the attribute composition and perceptual quality of both diffusion and Generative Adversarial Network (GAN) based face generation models across various attribute classes. Our findings suggest that diffusion models tend to worsen distribution bias in the training data for various attributes, which is heavily influenced by the size of the dataset. Conversely, GAN models trained on balanced datasets with a larger number of samples show less bias across different attributes. | ['Vishal M. Patel', 'Malsha V. Perera'] | 2023-05-10 | null | null | null | null | ['face-generation', 'synthetic-data-generation', 'synthetic-data-generation'] | ['computer-vision', 'medical', 'miscellaneous'] | [ 2.26555496e-01 2.63811141e-01 -2.86314130e-01 -2.95489579e-01
-1.33489579e-01 -6.16319597e-01 9.69524920e-01 -1.44987628e-01
-8.99428278e-02 9.53833997e-01 5.72751701e-01 -1.13378651e-01
3.57389867e-01 -1.13485682e+00 -5.82388639e-01 -6.10146761e-01
5.66671193e-01 1.10572159e-01 -5.51577151e-01 -2.16102898e-01
-7.98654631e-02 2.37528965e-01 -1.26933920e+00 6.41564429e-02
1.04781091e+00 7.90470064e-01 -5.03504455e-01 3.19183022e-01
7.55874068e-02 7.55195498e-01 -1.15703487e+00 -9.39723849e-01
3.31762493e-01 -8.78012240e-01 -1.30103737e-01 -1.46222055e-01
7.01484978e-01 -4.10241425e-01 -2.84697413e-01 9.90625858e-01
8.90921772e-01 -1.62701666e-01 1.04034781e+00 -1.56928909e+00
-1.31154835e+00 7.43265808e-01 -5.48169076e-01 8.82510282e-03
4.70498437e-03 6.05743885e-01 6.32770777e-01 -6.51525676e-01
8.51568282e-01 1.33963823e+00 5.75670123e-01 9.80623007e-01
-1.44844389e+00 -1.23474872e+00 6.72237203e-02 -2.85688549e-01
-9.19344783e-01 -7.56473362e-01 9.29569244e-01 -5.37477255e-01
1.22522414e-01 1.47117093e-01 7.32557416e-01 1.86599338e+00
2.11710781e-01 2.59841919e-01 1.45378828e+00 -1.48769751e-01
1.87542811e-01 4.25885201e-01 -4.78152096e-01 2.52201289e-01
6.02186203e-01 3.27212602e-01 -6.38866544e-01 -1.88326567e-01
9.10479784e-01 -3.72289419e-01 4.58165854e-02 -9.29363444e-02
-1.05545175e+00 9.49215591e-01 2.77228504e-01 -8.52141646e-04
-4.74860102e-01 2.17786491e-01 2.39651218e-01 3.20711344e-01
7.83284903e-01 8.29642355e-01 -1.80505291e-01 7.00614182e-04
-8.10286045e-01 3.50622356e-01 3.15856844e-01 8.15395415e-01
4.94549155e-01 6.66184723e-01 -6.43974781e-01 1.02801847e+00
1.23235201e-02 8.91939878e-01 4.30389971e-01 -9.87020373e-01
3.24110955e-01 4.24789339e-01 -3.60062905e-02 -1.06791949e+00
7.16648251e-02 -4.26156461e-01 -9.19298232e-01 2.98078984e-01
5.59444129e-01 -6.70818567e-01 -1.06209576e+00 2.37586498e+00
2.52312630e-01 -2.93978840e-01 -9.81751084e-02 6.45307600e-01
7.08052933e-01 2.94066429e-01 6.27675235e-01 -4.16060584e-03
9.94243622e-01 -5.04178643e-01 -6.69067085e-01 -4.51796353e-01
2.68993974e-01 -8.72641563e-01 1.19122517e+00 -2.21801400e-01
-1.26588535e+00 -7.31551826e-01 -8.01662147e-01 1.97848290e-01
-2.36602589e-01 1.94711745e-01 5.14616787e-01 1.21633816e+00
-9.91873145e-01 3.93350720e-01 -7.45687336e-02 -3.05016249e-01
9.51350749e-01 1.79439381e-01 -9.15853977e-02 -9.74118263e-02
-1.17577899e+00 7.23222792e-01 -2.76386052e-01 -2.55367160e-01
-1.01300335e+00 -9.65380132e-01 -5.92457592e-01 -1.50544479e-01
-2.32700095e-01 -1.01938879e+00 8.05056870e-01 -1.64729714e+00
-1.25969827e+00 7.74758518e-01 -6.45262608e-03 -1.49499983e-01
7.99822748e-01 1.36987984e-01 -4.71989721e-01 -4.92287725e-01
1.93081185e-01 1.08187056e+00 1.31750190e+00 -1.28453636e+00
-1.17574409e-01 -5.12377501e-01 -3.90866473e-02 4.78564724e-02
-4.81477529e-01 -1.72875196e-01 2.66467363e-01 -1.11501539e+00
-4.84069258e-01 -1.08624518e+00 -2.91360226e-02 1.85208637e-02
-3.80288363e-01 1.96653217e-01 5.54634750e-01 -5.83225965e-01
9.47545886e-01 -2.37765574e+00 5.48669323e-02 2.27684438e-01
2.48131171e-01 5.44450842e-02 -3.00219119e-01 1.63930491e-01
-1.10956974e-01 5.18465042e-01 7.76937082e-02 -2.57127970e-01
-1.32898599e-01 -1.73927858e-01 -1.92572415e-01 2.08536342e-01
4.42052931e-01 9.32543993e-01 -6.93097293e-01 -4.33008969e-01
-2.40411192e-01 5.68709970e-01 -6.36879921e-01 1.28642589e-01
-1.21355280e-01 7.68730760e-01 -2.56492347e-01 7.37119555e-01
8.44657481e-01 2.87717372e-01 2.45389383e-04 4.53378074e-02
2.56151408e-01 -8.07268247e-02 -5.30450106e-01 9.84153450e-01
-5.03446996e-01 9.79302704e-01 5.81552424e-02 -1.70184299e-01
1.12608159e+00 1.20170966e-01 1.05647080e-01 -8.48792255e-01
2.63162971e-01 1.47754967e-01 5.72341561e-01 4.75302339e-02
7.09810972e-01 -5.03754616e-01 6.70981482e-02 6.74237609e-01
-2.21607953e-01 -3.08148563e-01 3.96370105e-02 1.83115557e-01
7.07040191e-01 -3.30617368e-01 -3.39806974e-01 -1.59168661e-01
-3.81982327e-02 -1.57858580e-01 7.41024017e-01 6.30865514e-01
-1.85853839e-01 5.90223491e-01 8.26747298e-01 -3.13500054e-02
-1.42252088e+00 -1.04219759e+00 -2.19051801e-02 9.51243103e-01
-2.57333577e-01 -6.90469369e-02 -8.74176204e-01 -6.71503544e-01
3.04494321e-01 9.80148613e-01 -1.11940539e+00 -7.44437397e-01
-9.38489363e-02 -9.83258605e-01 9.21775401e-01 5.85067630e-01
4.30478156e-01 -1.01556587e+00 -1.93352059e-01 -2.72945285e-01
-3.22271883e-02 -6.99582458e-01 -5.79181790e-01 -6.17386639e-01
-8.68372798e-01 -7.98824728e-01 -1.03386664e+00 -2.95771629e-01
9.93748009e-01 -3.94031167e-01 1.26882386e+00 -2.88808206e-03
1.04658432e-01 3.19639385e-01 -1.40468806e-01 -9.52074945e-01
-7.77353406e-01 2.95172185e-01 -2.96787508e-02 1.98921904e-01
-3.25201489e-02 -4.22273189e-01 -8.84821773e-01 4.10337120e-01
-8.67988646e-01 2.92917853e-03 5.66995382e-01 5.67063510e-01
9.19490233e-02 -2.29732275e-01 1.14830530e+00 -1.33989275e+00
1.13395321e+00 -6.58825934e-01 -8.67642909e-02 9.02612880e-03
-1.02844334e+00 -2.39936858e-01 4.45911676e-01 -9.01129782e-01
-1.45025063e+00 -5.21635532e-01 1.91445842e-01 -2.85306603e-01
1.08687077e-02 5.00220098e-02 -3.32699299e-01 1.38466835e-01
9.15569484e-01 -1.74925148e-01 2.22858265e-01 -9.17352084e-03
3.98566455e-01 5.35505772e-01 8.74963477e-02 -6.84790671e-01
7.67833889e-01 5.07945776e-01 -2.45520007e-02 -5.32304347e-01
-3.36173654e-01 6.70676768e-01 -2.12425321e-01 -4.32674378e-01
6.89160466e-01 -1.06650627e+00 3.64552140e-02 8.77358019e-01
-7.03005016e-01 -8.22209060e-01 -6.61114752e-01 2.70814985e-01
-2.51643956e-01 -3.78254741e-01 -4.75747854e-01 -5.96139908e-01
-3.17655355e-01 -1.12207496e+00 6.41845822e-01 4.13749367e-01
-6.70085132e-01 -8.10635626e-01 -8.16799700e-02 3.80019665e-01
8.47462595e-01 4.59354132e-01 1.21388328e+00 -2.58939177e-01
-2.97070801e-01 -1.59640110e-03 -2.67267108e-01 3.96610320e-01
4.46307033e-01 4.15524781e-01 -1.00721467e+00 -1.45936906e-01
-2.88302273e-01 -2.22866759e-01 6.57180607e-01 6.17551923e-01
1.06705391e+00 -2.55564064e-01 -1.82468861e-01 6.11045957e-01
9.77286518e-01 2.41043881e-01 7.59356797e-01 -8.84398539e-03
6.95142508e-01 8.40009630e-01 4.13204908e-01 3.26740801e-01
1.06614396e-01 1.26310557e-01 1.51609674e-01 -4.14162815e-01
-4.68934983e-01 -6.39593184e-01 2.69958019e-01 1.70909196e-01
-1.06099054e-01 -3.07601005e-01 -6.98890865e-01 5.61932266e-01
-1.09920585e+00 -8.07957947e-01 5.45806661e-02 2.16838980e+00
1.00992763e+00 -9.97211412e-03 3.50225776e-01 -1.67302608e-01
8.56989920e-01 3.00119728e-01 -8.30564916e-01 -5.98258555e-01
-5.61996579e-01 1.53817579e-01 5.81344485e-01 2.07932456e-03
-3.84951115e-01 8.55323493e-01 7.03498554e+00 5.23907959e-01
-1.29016697e+00 4.88759726e-02 1.61378598e+00 -5.26527107e-01
-9.14602160e-01 -2.70521402e-01 -3.41652840e-01 7.35729158e-01
7.84636915e-01 -3.49906117e-01 4.95753706e-01 6.51116788e-01
1.84570625e-01 -2.34577153e-02 -1.03963685e+00 5.00890017e-01
1.45070970e-01 -1.08805799e+00 2.42294222e-01 3.47873390e-01
1.29097593e+00 -4.12442178e-01 9.07398224e-01 1.97323620e-01
5.34016788e-01 -1.41804159e+00 9.58481193e-01 3.34215403e-01
1.33266664e+00 -7.52681732e-01 5.27694523e-01 -1.36756927e-01
-3.21266979e-01 -2.60805935e-01 -1.83528870e-01 -3.19295414e-02
-5.78343682e-02 7.16218233e-01 -6.42342985e-01 -2.31777877e-01
4.94083613e-01 1.96390122e-01 -8.61612141e-01 3.31541389e-01
-3.93522292e-01 8.79378915e-01 2.03330293e-01 2.23532364e-01
-2.31810167e-01 -2.49055833e-01 2.73402244e-01 5.71253836e-01
4.52048689e-01 -1.40309930e-01 -7.87372887e-01 1.11226213e+00
-6.57607496e-01 3.67057137e-02 -7.95101523e-01 -4.05679345e-01
7.34028041e-01 9.39228058e-01 -4.84398782e-01 -3.02136064e-01
-1.88180834e-01 7.40714073e-01 1.01648785e-01 6.13666952e-01
-8.94355178e-01 4.38822582e-02 1.01565099e+00 4.53053296e-01
-7.87751675e-02 7.78840631e-02 -8.12622726e-01 -7.05999196e-01
-2.90552855e-01 -1.32254779e+00 6.12631775e-02 -8.56243968e-01
-1.39288485e+00 4.10758257e-01 -2.95998335e-01 -5.76335013e-01
-8.92395452e-02 -1.07309982e-01 -7.34433413e-01 1.11411893e+00
-1.17473388e+00 -1.28846681e+00 -3.06331873e-01 4.89048690e-01
2.89707333e-01 -3.94070148e-01 4.54532236e-01 1.18461736e-01
-6.80865526e-01 1.03424656e+00 -1.59621369e-02 2.85250902e-01
1.23045981e+00 -8.59911382e-01 5.45154393e-01 5.77418149e-01
-2.04492435e-01 8.35748494e-01 5.59883296e-01 -9.88663733e-01
-8.77970397e-01 -1.19952762e+00 5.52906632e-01 -6.21292710e-01
1.13961726e-01 -2.47290909e-01 -4.25963223e-01 7.40207911e-01
4.26214635e-01 -4.03279781e-01 1.01060724e+00 2.29764700e-01
-4.62236702e-01 -1.50421679e-01 -1.49205041e+00 9.62543786e-01
1.29495478e+00 -4.93158340e-01 2.51493424e-01 -1.03093468e-01
5.59030235e-01 -1.07969709e-01 -9.56255198e-01 4.20338154e-01
6.57795429e-01 -9.83350337e-01 8.45087409e-01 -6.49837494e-01
1.02699220e+00 2.17825010e-01 1.00684479e-01 -1.80369258e+00
-4.12777871e-01 -4.42443103e-01 3.44598472e-01 1.73556972e+00
8.27765167e-01 -8.59726250e-01 8.47342849e-01 8.99383366e-01
3.23182076e-01 -5.39492726e-01 -5.43037593e-01 -3.84097368e-01
5.99965990e-01 2.26450916e-02 1.28117907e+00 1.13350713e+00
-6.10237002e-01 3.54146153e-01 -3.69786948e-01 -3.43148053e-01
5.32496631e-01 -6.46917745e-02 1.00109208e+00 -1.08291471e+00
2.83035517e-01 -5.86003006e-01 -1.50975110e-02 -2.17749983e-01
3.67520064e-01 -5.52233815e-01 -4.31845814e-01 -1.15436912e+00
8.07688236e-02 -9.75792527e-01 -7.55838975e-02 1.69185281e-01
-5.17556489e-01 3.89111191e-01 4.16906536e-01 9.25789401e-02
2.92006731e-01 7.26490200e-01 1.53386617e+00 -2.31572270e-01
-6.94152117e-02 -1.63829029e-02 -1.24434054e+00 4.35179532e-01
1.11964095e+00 -4.63008016e-01 -6.59372270e-01 -6.91879570e-01
4.01844054e-01 -2.99605817e-01 1.55942783e-01 -7.72290170e-01
-3.66805136e-01 -4.80438471e-01 1.02039075e+00 2.35936165e-01
3.42123151e-01 -5.02403200e-01 4.03592169e-01 2.51714736e-01
-5.88408947e-01 3.79155457e-01 2.16071784e-01 3.14017087e-01
7.03817010e-02 1.99407563e-01 7.53638625e-01 -2.61706486e-02
5.34035601e-02 3.25322032e-01 -4.43213284e-01 3.90362471e-01
1.00379193e+00 -2.49413654e-01 -4.15782541e-01 -7.73596466e-01
-3.23168337e-01 -2.07836509e-01 1.03858447e+00 5.57206511e-01
2.51580656e-01 -1.45705390e+00 -9.91600096e-01 4.05066550e-01
-9.26437695e-03 -3.18325251e-01 2.55115867e-01 4.91969705e-01
-1.33779764e-01 -1.79544508e-01 -5.97913861e-01 -1.82873551e-02
-1.19257081e+00 2.98618525e-01 2.29076341e-01 2.73109585e-01
3.20033908e-01 9.83182490e-01 3.61396879e-01 -3.56276870e-01
-2.59041637e-01 6.88757077e-02 4.41863462e-02 3.57269615e-01
-6.56139627e-02 7.07575858e-01 -4.40351486e-01 -7.61968017e-01
-5.02805458e-03 1.97867602e-01 1.29874691e-01 -2.85880148e-01
8.76966774e-01 1.82465427e-02 7.93736428e-03 9.00091976e-02
6.14397585e-01 4.23506767e-01 -1.34640348e+00 1.82980165e-01
-7.17078984e-01 -7.82710433e-01 -1.81977510e-01 -8.60637426e-01
-1.57814670e+00 6.77614272e-01 5.51478028e-01 -8.31036270e-02
1.08086526e+00 -7.24798515e-02 5.95533431e-01 -6.21867776e-01
4.97730933e-02 -1.18664038e+00 2.02109188e-01 -4.75885645e-02
8.52690995e-01 -1.15093803e+00 -5.25930673e-02 -3.27835172e-01
-9.50359523e-01 3.42727989e-01 9.80211854e-01 2.36509725e-01
4.00347799e-01 1.30598411e-01 2.99004018e-01 8.39459151e-02
-7.77101994e-01 2.15153098e-01 -3.96413691e-02 1.10551178e+00
5.43721974e-01 2.21828863e-01 -3.97968650e-01 4.59878385e-01
-6.08188570e-01 -2.60582566e-01 5.35429955e-01 4.88865018e-01
4.65074927e-01 -1.21908569e+00 -4.38471317e-01 7.83963859e-01
-6.48224235e-01 -1.71224773e-01 -1.00791311e+00 5.96167326e-01
6.00127280e-01 8.82629931e-01 3.17308635e-01 -4.24253494e-01
1.71736866e-01 1.77740782e-01 4.44276601e-01 -4.82487530e-01
-8.39256823e-01 -5.34023285e-01 2.68023998e-01 -9.22836438e-02
-1.60422370e-01 -8.73908222e-01 -5.09556592e-01 -6.98616683e-01
-3.15646440e-01 -2.01356262e-01 6.91211998e-01 3.05677742e-01
6.76941454e-01 5.16716719e-01 6.07307732e-01 -1.95008546e-01
-3.00382733e-01 -1.21108353e+00 -4.70402062e-01 7.48772442e-01
-7.50020519e-02 -6.10967875e-01 -4.13128525e-01 2.96487100e-02] | [12.765375137329102, 0.9297983646392822] |
7b9adc22-5605-4983-9b24-755dcb22dc13 | gate-time-extraction-of-temporal-expressions | null | null | https://aclanthology.org/L16-1587 | https://aclanthology.org/L16-1587.pdf | GATE-Time: Extraction of Temporal Expressions and Events | GATE is a widely used open-source solution for text processing with a large user community. It contains components for several natural language processing tasks. However, temporal information extraction functionality within GATE has been rather limited so far, despite being a prerequisite for many application scenarios in the areas of natural language processing and information retrieval. This paper presents an integrated approach to temporal information processing. We take state-of-the-art tools in temporal expression and event recognition and bring them together to form an openly-available resource within the GATE infrastructure. GATE-Time provides annotation in the form of TimeML events and temporal expressions complying with this mature ISO standard for temporal semantic annotation of documents. Major advantages of GATE-Time are (i) that it relies on HeidelTime for temporal tagging, so that temporal expressions can be extracted and normalized in multiple languages and across different domains, (ii) it includes a modern, fast event recognition and classification tool, and (iii) that it can be combined with different linguistic pre-processing annotations, and is thus not bound to license restricted preprocessing components. | ['Jannik Str{\\"o}tgen', 'Mark A. Greenwood', 'Manuel Jung', 'Leon Derczynski', 'Diana Maynard'] | 2016-05-01 | gate-time-extraction-of-temporal-expressions-1 | https://aclanthology.org/L16-1587 | https://aclanthology.org/L16-1587.pdf | lrec-2016-5 | ['temporal-information-extraction', 'temporal-tagging'] | ['natural-language-processing', 'natural-language-processing'] | [-9.21355784e-02 -1.62636265e-01 -2.36890495e-01 -3.33725184e-01
-4.82016116e-01 -8.40300620e-01 1.00317872e+00 7.75258958e-01
-8.59296441e-01 5.71746886e-01 2.89558113e-01 -2.41127118e-01
-3.89679104e-01 -7.78731346e-01 6.14421591e-02 -3.87262762e-01
-3.98122281e-01 4.92813170e-01 7.83306718e-01 -3.64706546e-01
-9.32052135e-02 7.99165726e-01 -1.72272611e+00 3.19522977e-01
3.73245239e-01 1.15198469e+00 -7.97434151e-02 6.25113666e-01
-5.51882803e-01 1.11440933e+00 -4.27641898e-01 -1.26320124e-01
1.47166625e-02 -3.56359452e-01 -1.17604375e+00 -3.27629924e-01
-5.12878656e-01 2.32786506e-01 -1.26843572e-01 5.26459813e-01
1.47539318e-01 4.76572156e-01 3.68151873e-01 -1.23280311e+00
5.36332838e-02 6.54554009e-01 -2.10426450e-02 3.91609401e-01
7.07116485e-01 -1.48981586e-01 7.10232019e-01 -3.77267122e-01
1.29082608e+00 9.30097580e-01 4.41685051e-01 1.71165571e-01
-9.06995833e-01 -3.68954062e-01 4.73907143e-02 4.06310588e-01
-1.34163654e+00 -4.35015827e-01 5.46960354e-01 -5.47325373e-01
1.55472827e+00 4.05621707e-01 7.33130097e-01 9.66139674e-01
3.38732809e-01 4.45040405e-01 9.50764000e-01 -8.28575909e-01
3.20962757e-01 -2.77590230e-02 3.77142847e-01 3.13391030e-01
-3.42213243e-01 -2.37177581e-01 -7.01686144e-01 -3.59764062e-02
4.20413733e-01 -2.04206362e-01 1.52732387e-01 -6.80034310e-02
-1.30270493e+00 3.45663160e-01 -1.64607659e-01 1.34467590e+00
-4.75327730e-01 -3.61412130e-02 1.15923369e+00 3.41877371e-01
6.71261013e-01 3.04781049e-02 -5.79537332e-01 -7.58635640e-01
-8.03201377e-01 4.07221943e-01 9.04494166e-01 7.55976796e-01
3.51060778e-01 -2.31486157e-01 -2.35155776e-01 5.67428052e-01
3.46271843e-01 4.30075452e-02 5.26662171e-01 -6.00597680e-01
3.63628447e-01 8.64480674e-01 2.51567006e-01 -8.02334845e-01
-6.41938448e-01 1.47539124e-01 -3.63903821e-01 7.30773360e-02
5.60061157e-01 1.13943234e-01 -5.39346933e-01 1.68817782e+00
4.00872737e-01 -1.18853435e-01 4.17358689e-02 3.93585801e-01
6.46744490e-01 7.97906578e-01 5.09114623e-01 -8.49454284e-01
1.59714317e+00 -3.67605120e-01 -1.43443203e+00 -3.16506252e-02
9.52663183e-01 -8.00365746e-01 7.84603417e-01 4.15986985e-01
-1.02885342e+00 -2.11907830e-02 -9.72928405e-01 -2.41479889e-01
-1.18541837e+00 -3.06049228e-01 8.61514866e-01 4.65948641e-01
-8.92204583e-01 5.53253531e-01 -1.16077971e+00 -8.83188367e-01
-6.25929460e-02 1.87408552e-01 -5.94104111e-01 3.88999939e-01
-1.47046292e+00 1.08650684e+00 9.45007324e-01 -1.53170720e-01
-3.11884403e-01 -3.66392165e-01 -1.07946193e+00 -2.18346581e-01
4.80213135e-01 4.72823018e-03 1.54933214e+00 -6.30704284e-01
-1.54300404e+00 9.91374135e-01 -3.05398256e-01 -4.71121192e-01
5.98259687e-01 -4.16442528e-02 -1.14697087e+00 2.06424102e-01
2.92141289e-01 1.68732870e-02 1.45163998e-01 -3.56366217e-01
-6.07347250e-01 -3.86781842e-01 -2.13302746e-01 -1.66758266e-03
-4.81411785e-01 1.04974639e+00 -7.13096619e-01 -7.59318948e-01
-1.77689418e-01 -5.57947576e-01 -2.04277068e-01 -2.06794322e-01
2.25464851e-01 -6.26161933e-01 8.53826821e-01 -6.65626764e-01
1.92001450e+00 -2.14558387e+00 -2.54804790e-01 2.10628837e-01
-3.29784960e-01 2.03689888e-01 3.29080641e-01 9.78826463e-01
-2.98619926e-01 1.35572925e-01 -1.60620227e-01 -1.43647939e-01
3.60518575e-01 4.21305269e-01 -2.31964350e-01 5.69639564e-01
-5.08799031e-03 6.52099133e-01 -1.04599869e+00 -7.68460333e-01
7.34884977e-01 3.26898307e-01 1.18773632e-01 -1.24963425e-01
-5.04633009e-01 4.03138787e-01 -5.97140551e-01 3.10759246e-01
1.91638246e-01 2.30272382e-01 3.34389389e-01 2.10624829e-01
-8.99876177e-01 6.15908086e-01 -1.60331655e+00 1.94916892e+00
-5.23084819e-01 4.75038022e-01 -1.88212842e-01 -8.27475429e-01
7.97333300e-01 1.08165419e+00 9.71581399e-01 -7.71971405e-01
3.84248465e-01 3.47998023e-01 -5.51141083e-01 -5.86725712e-01
6.88985229e-01 -2.38772072e-02 -5.43564916e-01 7.16023028e-01
3.86248261e-01 -4.05139588e-02 1.21860528e+00 -2.56108716e-02
1.11714911e+00 6.12028658e-01 9.24910367e-01 -2.22115055e-01
6.91107512e-01 6.60532340e-02 7.38705635e-01 4.72342744e-02
-2.40989089e-01 -1.55156657e-01 4.54352915e-01 -6.46986306e-01
-8.89702559e-01 -7.72215545e-01 -3.34675282e-01 1.28930831e+00
-3.82293284e-01 -1.02482665e+00 -3.96416634e-01 -3.47718358e-01
-6.32102609e-01 7.99187064e-01 -4.06371236e-01 3.92408729e-01
-6.53025627e-01 -3.86518180e-01 8.53641331e-01 6.32463276e-01
3.71007770e-01 -1.20877182e+00 -9.33077455e-01 5.87643683e-01
-3.56875092e-01 -1.38303423e+00 -2.25136369e-01 3.82347316e-01
-5.99216342e-01 -1.07386851e+00 -1.80383876e-01 -6.25072539e-01
-3.66245918e-02 -3.32616180e-01 1.05107749e+00 -4.68490839e-01
-2.23158985e-01 4.44390118e-01 -6.76984012e-01 -7.48038173e-01
-3.71291578e-01 -1.45533428e-01 -4.82149608e-02 -3.81375216e-02
7.28150725e-01 -5.93965530e-01 2.64131390e-02 3.12025577e-01
-1.12434542e+00 -3.35910738e-01 -2.22699121e-01 -3.45315896e-02
4.72366929e-01 4.29361463e-01 3.48395526e-01 -5.42589009e-01
5.43543696e-01 -2.83712298e-01 -8.70402277e-01 3.12739938e-01
-4.93460208e-01 -6.33828044e-02 4.82635617e-01 -1.52985364e-01
-1.43017948e+00 2.55797785e-02 -1.32767111e-01 2.44883761e-01
-4.69956696e-01 8.08317363e-01 -1.70275345e-01 4.10184413e-01
6.61440670e-01 -6.67967573e-02 -5.81545115e-01 -4.26839978e-01
4.81801927e-01 5.03897548e-01 5.79907775e-01 -6.64900780e-01
2.78279096e-01 6.63884282e-01 7.22373873e-02 -8.64140511e-01
-4.71735120e-01 -9.61513281e-01 -9.99102592e-01 -5.49096644e-01
1.15607989e+00 -7.07002580e-01 -5.45829594e-01 4.54125226e-01
-1.21100414e+00 -3.47355753e-01 -5.34325957e-01 5.72841108e-01
-5.95981717e-01 4.47495997e-01 -6.71284795e-01 -1.11538589e+00
-1.71516538e-01 -3.86035889e-01 6.80357933e-01 1.48521021e-01
-8.21543574e-01 -1.24373960e+00 1.66650712e-01 -6.69707879e-02
7.99959078e-02 5.63112080e-01 5.25344193e-01 -6.76623702e-01
5.40503003e-02 -5.36955059e-01 1.69671074e-01 -7.94085935e-02
1.16619077e-02 4.45927203e-01 -8.07399631e-01 2.32613519e-01
4.85353321e-02 -5.15945926e-02 1.91278398e-01 1.15210295e-01
6.99064076e-01 -5.34402169e-02 -5.08533359e-01 6.32040948e-02
1.17051518e+00 7.03320086e-01 7.44230211e-01 6.93941951e-01
-7.39111751e-03 8.91428649e-01 1.13806629e+00 7.50976861e-01
1.65533662e-01 9.45403934e-01 -8.36855173e-02 2.56692320e-01
3.68049949e-01 1.03919216e-01 4.39298153e-01 7.33929992e-01
-3.02576095e-01 -2.50648528e-01 -1.18991876e+00 8.85549009e-01
-2.24340272e+00 -1.26214409e+00 -6.50307238e-01 2.02994728e+00
8.27567756e-01 2.07091659e-01 2.23149478e-01 5.55985570e-01
4.22464132e-01 2.85220981e-01 2.32287452e-01 -6.35981262e-01
-1.06840536e-01 4.45778936e-01 3.33918601e-01 2.03602076e-01
-1.33128369e+00 9.08119440e-01 5.73199081e+00 7.62890697e-01
-9.76028323e-01 3.65484804e-01 -1.74633831e-01 2.07550591e-03
1.77100614e-01 1.32484958e-01 -5.74043572e-01 2.65030950e-01
1.47841477e+00 -6.84592247e-01 1.53210267e-01 4.84464020e-01
7.59706140e-01 -2.10017681e-01 -1.03957224e+00 8.57879341e-01
-2.69566119e-01 -1.22980106e+00 -4.33735758e-01 -2.31482521e-01
2.33288631e-01 -5.38586229e-02 -6.68564260e-01 2.01571316e-01
1.13911361e-01 -5.56944549e-01 1.18716705e+00 4.30882692e-01
5.71005523e-01 -6.78358912e-01 8.59475672e-01 2.80182809e-01
-1.84560907e+00 1.51166648e-01 3.33211541e-01 -1.74367577e-01
8.85932922e-01 6.35911405e-01 -4.22203273e-01 1.12053692e+00
9.25674796e-01 8.34369063e-01 -1.07211553e-01 8.64175797e-01
-3.93679768e-01 3.39777976e-01 -6.36094987e-01 2.59902384e-02
1.90796912e-01 -1.64453879e-01 4.75668252e-01 1.65045083e+00
3.89162391e-01 9.29222628e-02 2.55803913e-01 3.55919927e-01
5.28787374e-01 5.46176732e-01 -5.84436119e-01 -4.66390163e-01
3.85807633e-01 1.03878236e+00 -1.27827358e+00 -3.14757705e-01
-4.48422700e-01 7.93438017e-01 -2.09718943e-01 -3.50108892e-02
-8.74805570e-01 -5.38639724e-01 3.11117083e-01 1.31615162e-01
8.25736579e-03 -5.74074745e-01 3.86861749e-02 -9.81570899e-01
2.18825758e-01 -4.71987337e-01 9.96149182e-01 -7.91850924e-01
-7.72616684e-01 6.61518455e-01 5.93253732e-01 -1.38118303e+00
-5.36835790e-01 -6.90865815e-01 -3.38487118e-01 7.74607956e-01
-1.00757968e+00 -1.31275642e+00 -1.65578462e-02 9.83671248e-01
3.03210825e-01 3.29164267e-01 1.15444469e+00 7.43064940e-01
-3.62841785e-01 -1.60725057e-01 -4.61129844e-01 1.17319457e-01
8.11383665e-01 -1.19742441e+00 1.38768151e-01 1.26069891e+00
3.60962637e-02 6.41530395e-01 7.30578780e-01 -5.95155954e-01
-9.86633241e-01 -1.08238459e+00 1.91918480e+00 -3.32099080e-01
1.21532619e+00 -3.28019798e-01 -7.29087055e-01 7.86529541e-01
3.33535105e-01 -8.63812789e-02 8.31691921e-01 1.23897061e-01
-3.31475735e-01 -1.07919574e-01 -9.01369095e-01 5.88659644e-01
9.26253974e-01 -8.05157304e-01 -8.28279197e-01 5.65145671e-01
4.80650783e-01 -4.34095621e-01 -1.30966437e+00 1.26020193e-01
1.96168840e-01 -6.67363584e-01 5.11080682e-01 -4.07366425e-01
-1.03613406e-01 -6.15698874e-01 1.12593949e-01 -5.33208191e-01
-7.04736216e-03 -9.53462303e-01 -8.83724391e-02 1.84409750e+00
3.05486858e-01 -6.38146102e-01 1.71101108e-01 7.17932165e-01
-3.43179554e-01 1.01367161e-01 -1.25653160e+00 -1.08619642e+00
-5.16524374e-01 -1.25193524e+00 3.89320791e-01 1.09590244e+00
7.56416261e-01 2.29294866e-01 -1.75319850e-01 -1.76353678e-01
-7.52307102e-02 -1.72790647e-01 3.30142230e-01 -1.43803167e+00
5.97283319e-02 -4.97741848e-01 -4.65528756e-01 -3.18460345e-01
1.97581440e-01 -9.33897853e-01 -5.12730889e-02 -1.61307073e+00
-4.54974502e-01 -1.64884269e-01 -1.91584438e-01 9.41086531e-01
5.82379460e-01 3.20972055e-02 -1.99994177e-01 1.47025332e-01
-7.45740294e-01 1.30417168e-01 4.91466761e-01 1.83141902e-01
-4.45193827e-01 -1.56339720e-01 -3.94216292e-02 8.17122042e-01
6.16275191e-01 -5.87210476e-01 -3.77960145e-01 1.57290734e-02
6.03541493e-01 -7.26261688e-03 1.87911108e-01 -1.03443432e+00
3.19435745e-01 -3.94063413e-01 -1.51761681e-01 -7.85649300e-01
5.09777479e-02 -1.07920206e+00 5.94465196e-01 3.53674382e-01
3.23623009e-02 2.67781347e-01 5.95022798e-01 3.04557621e-01
-5.05835176e-01 -2.92715013e-01 7.10459948e-01 -1.85910285e-01
-1.29226375e+00 8.23513493e-02 -9.81087565e-01 -1.75294057e-01
1.57291186e+00 -2.39614055e-01 -2.23832339e-01 2.23457608e-02
-8.10298443e-01 3.28136720e-02 3.90658379e-01 5.59219003e-01
-1.86439127e-01 -1.12024033e+00 -2.25749955e-01 -2.26385519e-01
4.36962783e-01 -3.08666795e-01 1.98295176e-01 8.51500630e-01
-6.56912863e-01 8.07842016e-01 -7.97386020e-02 -4.09041017e-01
-1.28584266e+00 7.70922899e-01 -1.22699607e-02 -7.43622065e-01
-7.64863431e-01 1.84019104e-01 -5.77921510e-01 -1.04198225e-01
1.33362427e-01 -5.52934885e-01 -5.87644815e-01 6.74138069e-01
8.69850636e-01 3.39283943e-01 5.03273189e-01 -7.77107716e-01
-7.12663352e-01 2.52734333e-01 3.57014060e-01 -6.71136618e-01
1.54645705e+00 -8.56544226e-02 -6.93677843e-01 9.75153387e-01
6.10579789e-01 9.80414674e-02 -4.05676246e-01 -1.81291364e-02
8.02901268e-01 5.36982082e-02 -1.66317169e-02 -5.95460951e-01
-2.65611947e-01 3.82781237e-01 4.96219605e-01 6.55909121e-01
1.25843382e+00 -7.56071582e-02 4.72620875e-01 1.37548730e-01
5.75977266e-01 -1.49638999e+00 -4.74629521e-01 8.22095513e-01
7.62319148e-01 -4.94965166e-01 -1.23566493e-01 -5.55522263e-01
-3.57596785e-01 1.19474483e+00 2.69546341e-02 4.37558949e-01
7.51307845e-01 5.64919651e-01 2.01468930e-01 -2.78103024e-01
-9.22909677e-01 -4.95874673e-01 1.77041143e-01 4.93872970e-01
9.91195738e-01 -1.37222201e-01 -1.08684564e+00 6.82973027e-01
2.24622451e-02 6.45958602e-01 1.29184142e-01 1.41785038e+00
-8.02395493e-02 -1.55052340e+00 -3.30514342e-01 -9.64242220e-02
-8.08350027e-01 1.73415557e-01 -2.19658911e-01 9.85490441e-01
2.93645173e-01 1.27491808e+00 -2.96600983e-02 2.11781278e-01
5.98858595e-01 4.07211483e-01 3.61904263e-01 -7.13579118e-01
-9.62390304e-01 6.48195669e-02 5.47982216e-01 -7.90261507e-01
-8.48282158e-01 -9.63796258e-01 -1.55500925e+00 -5.73497312e-03
1.06997423e-01 4.95217919e-01 6.98148549e-01 1.10918570e+00
1.75800189e-01 7.25954473e-01 -2.35121902e-02 -7.66299307e-01
2.89608032e-01 -9.18948948e-01 -6.76364183e-01 4.33557987e-01
-3.86547208e-01 -6.51713371e-01 -1.40839994e-01 4.30447191e-01] | [9.228060722351074, 9.252634048461914] |
3de0ac32-c0d6-4629-8351-42fe73686cb3 | deep-single-image-deraining-via-estimating | 1906.09433 | null | https://arxiv.org/abs/1906.09433v1 | https://arxiv.org/pdf/1906.09433v1.pdf | Deep Single Image Deraining Via Estimating Transmission and Atmospheric Light in rainy Scenes | Rain removal in images/videos is still an important task in computer vision field and attracting attentions of more and more people. Traditional methods always utilize some incomplete priors or filters (e.g. guided filter) to remove rain effect. Deep learning gives more probabilities to better solve this task. However, they remove rain either by evaluating background from rainy image directly or learning a rain residual first then subtracting the residual to obtain a clear background. No other models are used in deep learning based de-raining methods to remove rain and obtain other information about rainy scenes. In this paper, we utilize an extensively-used image degradation model which is derived from atmospheric scattering principles to model the formation of rainy images and try to learn the transmission, atmospheric light in rainy scenes and remove rain further. To reach this goal, we propose a robust evaluation method of global atmospheric light in a rainy scene. Instead of using the estimated atmospheric light directly to learn a network to calculate transmission, we utilize it as ground truth and design a simple but novel triangle-shaped network structure to learn atmospheric light for every rainy image, then fine-tune the network to obtain a better estimation of atmospheric light during the training of transmission network. Furthermore, more efficient ShuffleNet Units are utilized in transmission network to learn transmission map and the de-raining image is then obtained by the image degradation model. By subjective and objective comparisons, our method outperforms the selected state-of-the-art works. | ['Qinfeng Shi', 'Chao Ma', 'Yinglong Wang', 'Ehsan Abbasnejad', 'Bing Zeng', 'Xiaoping Ma'] | 2019-06-22 | null | null | null | null | ['single-image-deraining'] | ['computer-vision'] | [ 2.21306175e-01 -5.63005030e-01 4.59018558e-01 -6.98246181e-01
-3.96725163e-02 -1.49880096e-01 -5.22574969e-02 -5.60791433e-01
-5.12736499e-01 1.05548847e+00 -2.39097983e-01 -1.78360835e-01
3.01191628e-01 -1.15031493e+00 -6.64760530e-01 -1.26990306e+00
1.30408555e-01 -2.85987146e-02 2.76127994e-01 -3.08377385e-01
8.82793069e-02 4.68285680e-01 -1.44696772e+00 3.44839576e-03
1.29576063e+00 4.68582243e-01 8.46024990e-01 7.30233848e-01
-3.07814538e-01 9.50810730e-01 -6.97908044e-01 2.10135415e-01
2.22752944e-01 -6.99928999e-01 -2.05153763e-01 1.02858387e-01
5.73481143e-01 -7.73049712e-01 -2.90873110e-01 1.24711609e+00
5.23393571e-01 5.88110723e-02 4.04228300e-01 -5.54942906e-01
-6.02344871e-01 2.67773092e-01 -7.00470805e-01 6.24947488e-01
-3.69842276e-02 4.36112463e-01 4.26872343e-01 -8.69575381e-01
2.47177839e-01 1.27123606e+00 4.77062494e-01 4.47588146e-01
-7.01858163e-01 -9.82088745e-01 2.55545616e-01 4.67078567e-01
-1.10680747e+00 -3.34891856e-01 6.12407923e-01 2.53197327e-02
2.48961896e-01 4.17648226e-01 9.45254087e-01 7.05274582e-01
3.34501088e-01 6.02687418e-01 1.52303064e+00 -4.17643577e-01
-1.42100556e-02 1.35357127e-01 3.93905789e-01 7.61618793e-01
6.29653931e-01 2.11531192e-01 -8.70341882e-02 5.33322692e-01
7.52393663e-01 3.74077350e-01 -8.28222454e-01 4.35698897e-01
-6.91994607e-01 6.39722824e-01 8.65865290e-01 2.36366555e-01
-4.92760420e-01 4.47056442e-02 -3.35312694e-01 4.90032047e-01
5.30393064e-01 5.22479229e-02 -3.02620530e-01 6.03323996e-01
-1.12331653e+00 3.79094690e-01 6.59548461e-01 3.54523838e-01
1.25181901e+00 5.11150539e-01 -2.40329698e-01 6.68130100e-01
6.18069828e-01 1.36506331e+00 -4.82817516e-02 -9.09626067e-01
4.13085163e-01 8.92040581e-02 4.42702234e-01 -8.79777014e-01
-3.86048287e-01 -6.88254118e-01 -1.22010744e+00 8.29536021e-01
1.62668362e-01 -2.04208329e-01 -1.34139204e+00 1.27827287e+00
2.47030392e-01 5.56237459e-01 2.35266745e-01 1.44167566e+00
1.04141915e+00 1.10814428e+00 -3.32994819e-01 -7.74060488e-01
1.05928421e+00 -9.85110879e-01 -9.64773357e-01 -3.59453857e-01
-1.02019385e-01 -8.73827100e-01 9.09471214e-01 7.01583743e-01
-9.71616507e-01 -7.28377402e-01 -1.20181727e+00 1.25136256e-01
-2.41835471e-02 -1.14248861e-02 4.50565904e-01 6.73388243e-01
-8.87699604e-01 5.42835057e-01 -7.16849208e-01 -7.64931589e-02
2.90724128e-01 9.98062342e-02 2.44676620e-01 -7.05572844e-01
-1.48526561e+00 9.42843020e-01 1.48887962e-01 8.55528176e-01
-1.34904337e+00 -3.81513655e-01 -4.21319336e-01 -1.27433345e-01
1.56358257e-01 -8.61567557e-01 7.13820755e-01 -1.34760642e+00
-1.57962060e+00 4.05755550e-01 -1.84278518e-01 -3.43884587e-01
1.88867792e-01 -6.04213893e-01 -4.87676948e-01 1.51749775e-01
-4.16924030e-01 2.27307454e-01 1.39084518e+00 -1.77557576e+00
-6.52526021e-01 -7.79070631e-02 2.32068181e-01 5.21327317e-01
9.49102491e-02 -1.74037352e-01 -4.41314608e-01 -4.75605965e-01
7.10809603e-02 -8.03969622e-01 -3.36402386e-01 -4.79538515e-02
-1.73940137e-01 3.92645329e-01 9.50531006e-01 -1.04879320e+00
1.10150063e+00 -1.79252219e+00 -7.40433410e-02 1.05498411e-01
4.34164435e-01 7.69984961e-01 -1.69957817e-01 -1.87702075e-01
1.00925945e-01 -3.05593371e-01 -6.26699805e-01 -2.47789204e-01
-6.62111461e-01 5.62043369e-01 -3.12305927e-01 5.93189061e-01
-6.69663912e-03 3.76037687e-01 -6.82129204e-01 -4.99669909e-01
3.72463316e-01 8.33253562e-01 -2.86191553e-01 6.12883687e-01
-1.93146840e-01 7.87216008e-01 -3.86993915e-01 5.26818693e-01
1.49965668e+00 1.92495868e-01 -1.76559523e-01 -3.20286512e-01
-3.00611377e-01 -1.99016616e-01 -1.19545448e+00 1.08877456e+00
-7.57449627e-01 5.37303925e-01 5.58351159e-01 -7.54683614e-01
1.08539963e+00 4.23744842e-02 -1.30115509e-01 -6.73657298e-01
1.90956697e-01 9.16574970e-02 -4.15085256e-02 -1.06495285e+00
2.24439368e-01 -3.99103194e-01 8.83478820e-01 1.62303105e-01
-3.87800336e-01 -2.47127384e-01 -1.41811296e-01 -6.68384507e-02
7.87838817e-01 1.72630534e-01 -2.87987053e-01 5.56444936e-02
6.75465703e-01 -2.62898773e-01 8.11688662e-01 7.73090303e-01
2.24477034e-02 7.39497602e-01 -3.91418666e-01 -6.30450845e-01
-7.10055649e-01 -1.07534754e+00 -6.37939572e-02 7.76920736e-01
4.97427464e-01 3.37119281e-01 -7.51975119e-01 -2.98815250e-01
-2.71397084e-01 7.60145366e-01 -4.32823330e-01 4.08983976e-02
-8.82365108e-01 -1.33155656e+00 6.08029179e-02 -1.52958319e-01
1.11086583e+00 -1.34515595e+00 -3.49371731e-01 6.63630813e-02
-4.41483855e-01 -8.67028713e-01 -2.20311992e-02 1.88442335e-01
-9.96051252e-01 -1.02926099e+00 -7.01381028e-01 -6.26101553e-01
6.37363553e-01 8.48204434e-01 1.14052618e+00 5.46864331e-01
-3.46073508e-01 -1.75510630e-01 -4.95317727e-01 -5.69925487e-01
-1.84836507e-01 -4.63555694e-01 -1.86969221e-01 1.23559453e-01
1.54264703e-01 -6.44745767e-01 -1.01637244e+00 -1.59589965e-02
-1.05199194e+00 -8.26207474e-02 8.35768580e-01 5.21270931e-01
3.58022392e-01 5.76749563e-01 1.31169513e-01 -9.29883838e-01
3.72423798e-01 -3.96168113e-01 -8.07911158e-01 1.15083545e-01
-5.51827312e-01 -1.14211805e-01 6.03186429e-01 -1.50804520e-02
-1.63206506e+00 -1.86222807e-01 -2.15573698e-01 -5.89471817e-01
-9.75667983e-02 3.40095937e-01 -1.92814142e-01 -2.39330336e-01
4.56533700e-01 3.86909872e-01 -1.81297630e-01 -4.24220353e-01
2.59033144e-01 4.55210328e-01 4.78034973e-01 -7.54605457e-02
1.23561561e+00 6.85861230e-01 -1.74130667e-02 -8.82894516e-01
-1.22375107e+00 -2.18073279e-01 -3.26013952e-01 -3.94666463e-01
1.05718195e+00 -1.18353343e+00 -6.46297634e-01 7.62019455e-01
-1.27279305e+00 -3.83613855e-01 2.38077819e-01 6.50873601e-01
1.35151923e-01 3.88160348e-01 -6.69010222e-01 -1.16051495e+00
-7.36267805e-01 -9.88703787e-01 6.02567255e-01 5.44067919e-01
7.38023698e-01 -8.64325464e-01 1.68743864e-01 2.39792526e-01
6.00309074e-01 2.73411274e-02 5.14486551e-01 6.20215833e-01
-1.15367091e+00 2.50457436e-01 -5.29054284e-01 8.24999332e-01
3.85469109e-01 1.57913789e-01 -1.13609886e+00 -3.44574034e-01
5.55351019e-01 1.22544199e-01 1.60552323e+00 7.25080431e-01
9.59553421e-01 -3.12035054e-01 -5.34074903e-02 1.02146566e+00
1.92266178e+00 1.37298286e-01 1.11977494e+00 4.73140717e-01
8.53560209e-01 4.66020107e-01 7.34416068e-01 3.70838016e-01
1.73382267e-01 1.38328850e-01 8.94074082e-01 -4.45191562e-01
-4.28150833e-01 5.00913739e-01 5.48412383e-01 1.03032911e+00
-5.93286037e-01 -4.39230740e-01 -4.22134489e-01 2.89396375e-01
-1.45943987e+00 -1.13063490e+00 -4.39227968e-01 2.22253633e+00
1.14550173e+00 2.25402787e-01 -5.95633686e-01 -8.26387405e-02
6.30939305e-01 3.21145415e-01 -4.50335592e-01 -5.20993136e-02
-3.98432374e-01 4.94768590e-01 7.72377908e-01 9.77904260e-01
-9.12366748e-01 1.01346886e+00 5.63413000e+00 5.04380405e-01
-1.35320389e+00 6.31769523e-02 3.82134259e-01 -4.33859639e-02
-3.34475547e-01 6.19776659e-02 -9.75630522e-01 5.78165412e-01
4.85615730e-01 6.11535430e-01 8.20091546e-01 1.14851534e-01
8.82419825e-01 -4.50499535e-01 -2.63116032e-01 9.99664485e-01
-1.21667422e-03 -7.78272688e-01 7.15838745e-02 -4.73468214e-01
7.56039739e-01 1.58056244e-01 -6.23838156e-02 1.59133181e-01
5.67555666e-01 -9.85682070e-01 2.33984724e-01 1.28103399e+00
4.94827181e-01 -4.43193346e-01 1.01290667e+00 2.28408664e-01
-7.98478246e-01 4.16060351e-03 -7.96875119e-01 -1.48380026e-01
8.97656158e-02 1.47319865e+00 -3.97891790e-01 6.61881089e-01
1.01499021e+00 7.10207820e-01 -4.56953019e-01 1.18399441e+00
-6.88999951e-01 1.06202412e+00 -3.35982919e-01 4.00550097e-01
1.06374502e-01 -6.77337110e-01 5.63549399e-01 1.22337031e+00
3.40512097e-01 3.57578218e-01 3.68545465e-02 7.74971545e-01
6.52502626e-02 -2.58898765e-01 -2.65439898e-01 6.20743454e-01
1.30207673e-01 1.42640817e+00 -4.07273531e-01 -4.19304043e-01
-2.66919702e-01 1.02614546e+00 -1.23842061e-01 7.93421626e-01
-9.80009556e-01 -4.94235337e-01 5.36650419e-01 -1.32919639e-01
2.97316879e-01 -2.70967245e-01 -8.33499804e-02 -1.18165076e+00
-1.28077164e-01 -8.38986993e-01 -2.59321690e-01 -1.26120102e+00
-1.05085993e+00 6.78862870e-01 -2.92408079e-01 -1.15347040e+00
4.27743226e-01 -3.95969152e-01 -9.83862519e-01 1.25547147e+00
-2.47488332e+00 -7.70764649e-01 -1.22011185e+00 8.47382069e-01
5.62873840e-01 1.77579999e-01 3.24778378e-01 4.43927407e-01
-6.72129691e-01 -6.41775280e-02 3.82121980e-01 -2.82304138e-01
1.14146733e+00 -1.12927043e+00 -1.78047821e-01 1.02586448e+00
-2.65142053e-01 3.47644538e-01 1.20383775e+00 -7.00439632e-01
-1.23170650e+00 -1.33897400e+00 3.41924280e-01 7.89324474e-03
1.76744327e-01 8.82573426e-03 -1.19759870e+00 4.36507910e-01
4.81245309e-01 1.70872375e-01 -1.06860556e-01 -1.23013929e-01
-1.04111083e-01 -8.10593188e-01 -1.08099484e+00 3.80406082e-01
6.65844619e-01 3.24312970e-02 -5.04415989e-01 4.60979521e-01
9.18567300e-01 -3.14349353e-01 -2.10070699e-01 5.09105027e-01
3.25722098e-01 -1.29436564e+00 8.55449736e-01 1.08761430e-01
4.28194165e-01 -7.34466732e-01 -7.58286193e-02 -1.63092148e+00
-4.44673300e-01 -2.39841864e-01 5.99286593e-02 1.03291976e+00
5.68257384e-02 -7.09829688e-01 7.58575976e-01 -2.16550499e-01
-2.53821373e-01 -3.84617448e-01 -2.65574366e-01 -3.51605654e-01
-2.04328179e-01 2.19151601e-02 3.84093076e-01 1.00935781e+00
-1.12520754e+00 2.87073553e-01 -7.60257542e-01 8.14252853e-01
1.10959470e+00 3.18010926e-01 7.08001375e-01 -1.17047226e+00
-9.26222652e-02 6.76733479e-02 3.00008655e-01 -9.64097679e-01
-1.12197243e-01 -3.36797088e-01 5.56925833e-01 -1.84132671e+00
9.76235196e-02 -5.76679289e-01 -4.98149306e-01 2.84406453e-01
-6.02365851e-01 3.95567924e-01 1.92076743e-01 5.71876228e-01
-4.08271968e-01 6.26646042e-01 1.80881679e+00 -2.84829319e-01
-3.09401006e-01 2.00516105e-01 -2.56568313e-01 7.60216832e-01
9.86355901e-01 -4.08756256e-01 -2.68963665e-01 -9.24088120e-01
2.36756548e-01 2.00709030e-02 4.48926896e-01 -1.30087650e+00
1.47458138e-02 -3.55961353e-01 6.64169788e-01 -6.64592266e-01
3.68714690e-01 -9.54525471e-01 -1.36947725e-02 4.73374277e-01
1.38807625e-01 -4.99935001e-01 -9.38861594e-02 5.09678423e-01
-3.58279526e-01 -4.51027840e-01 1.07963550e+00 -4.48348492e-01
-6.08814895e-01 3.30357969e-01 -3.94033790e-01 -5.62569320e-01
4.91061896e-01 -9.48600769e-02 -5.86141765e-01 -4.01825458e-01
-7.29074836e-01 3.28190684e-01 1.18575387e-01 -5.83186969e-02
8.67853403e-01 -7.53109634e-01 -1.09318662e+00 1.81350306e-01
-4.57913667e-01 1.36262447e-01 5.40294409e-01 8.79750192e-01
-1.00292182e+00 -4.90351915e-02 -2.02611446e-01 -3.79434407e-01
-1.40467370e+00 4.04672712e-01 5.50201297e-01 -2.26390228e-01
-6.47461474e-01 7.33644128e-01 6.04125381e-01 -2.12178603e-01
-1.11549105e-02 -3.97702903e-01 -4.77282524e-01 -2.25735679e-01
7.18459904e-01 2.57827431e-01 -7.61499181e-02 -1.11574531e-01
6.82210252e-02 6.72058105e-01 -1.36891240e-02 3.60741466e-01
1.45661545e+00 -5.74030399e-01 -5.49091935e-01 3.95364851e-01
8.17748964e-01 1.07735343e-01 -1.32297611e+00 -3.12602907e-01
-8.60284448e-01 -5.24741232e-01 4.34790283e-01 -8.81990254e-01
-1.74158001e+00 9.96275246e-01 1.14846528e+00 8.80627185e-02
1.59710240e+00 -5.77308714e-01 9.42263603e-01 6.68659449e-01
5.81977591e-02 -6.90369427e-01 -8.75134990e-02 7.23542154e-01
7.68762410e-01 -1.34653807e+00 3.49940240e-01 -4.50372159e-01
-1.77090421e-01 1.27764106e+00 6.78870261e-01 -3.96553636e-01
8.17491055e-01 3.10344756e-01 4.87093091e-01 -1.58825591e-01
-2.43271664e-01 -4.30580854e-01 -3.54056031e-01 4.89687145e-01
1.37280270e-01 -1.35176182e-01 -2.77922720e-01 1.39847159e-01
1.17329918e-01 2.73210347e-01 7.34847128e-01 6.92394614e-01
-1.34146893e+00 -7.30292499e-01 -9.89981592e-01 4.05804306e-01
-3.03094417e-01 -5.61858237e-01 3.40604872e-01 3.69431853e-01
4.62642938e-01 1.06306422e+00 -1.40221030e-01 -2.03972161e-01
3.50675695e-02 -4.90612388e-01 5.49613774e-01 -3.86988729e-01
-3.32404494e-01 9.13385451e-02 -2.61139721e-01 -2.73969650e-01
-1.17608070e+00 -2.11939856e-01 -1.03204119e+00 -3.84200841e-01
-5.56297004e-01 1.25503838e-01 6.53489351e-01 8.82962227e-01
-2.36488491e-01 8.21970820e-01 9.28817928e-01 -1.06715488e+00
8.75903293e-02 -1.03238833e+00 -9.77817774e-01 8.98278952e-02
8.24534118e-01 -4.99931842e-01 -7.07624435e-01 2.29153186e-01] | [10.9114408493042, -3.2586185932159424] |
4d4ff259-4978-4b8e-b900-de99a9a54545 | lmbot-distilling-graph-knowledge-into | 2306.17408 | null | https://arxiv.org/abs/2306.17408v2 | https://arxiv.org/pdf/2306.17408v2.pdf | LMBot: Distilling Graph Knowledge into Language Model for Graph-less Deployment in Twitter Bot Detection | As malicious actors employ increasingly advanced and widespread bots to disseminate misinformation and manipulate public opinion, the detection of Twitter bots has become a crucial task. Though graph-based Twitter bot detection methods achieve state-of-the-art performance, we find that their inference depends on the neighbor users multi-hop away from the targets, and fetching neighbors is time-consuming and may introduce bias. At the same time, we find that after finetuning on Twitter bot detection, pretrained language models achieve competitive performance and do not require a graph structure during deployment. Inspired by this finding, we propose a novel bot detection framework LMBot that distills the knowledge of graph neural networks (GNNs) into language models (LMs) for graph-less deployment in Twitter bot detection to combat the challenge of data dependency. Moreover, LMBot is compatible with graph-based and graph-less datasets. Specifically, we first represent each user as a textual sequence and feed them into the LM for domain adaptation. For graph-based datasets, the output of LMs provides input features for the GNN, enabling it to optimize for bot detection and distill knowledge back to the LM in an iterative, mutually enhancing process. Armed with the LM, we can perform graph-less inference, which resolves the graph data dependency and sampling bias issues. For datasets without graph structure, we simply replace the GNN with an MLP, which has also shown strong performance. Our experiments demonstrate that LMBot achieves state-of-the-art performance on four Twitter bot detection benchmarks. Extensive studies also show that LMBot is more robust, versatile, and efficient compared to graph-based Twitter bot detection methods. | ['Minnan Luo', 'Qinghua Zheng', 'Hongrui Wang', 'Zifeng Zhu', 'Zhenyu Lei', 'Zhaoxuan Tan', 'Zijian Cai'] | 2023-06-30 | null | null | null | null | ['twitter-bot-detection', 'misinformation'] | ['miscellaneous', 'miscellaneous'] | [-3.11618716e-01 -1.16813526e-01 -5.47190845e-01 2.98216581e-01
-1.31431624e-01 -8.97081077e-01 5.99851191e-01 1.39953271e-01
-5.83700180e-01 3.23909342e-01 -3.37820888e-01 -6.49711251e-01
3.55101317e-01 -1.16016769e+00 -4.36914146e-01 -4.49315786e-01
-7.97824040e-02 6.28879011e-01 7.89486349e-01 -4.24153596e-01
2.12033123e-01 6.75368190e-01 -8.53479028e-01 1.33899003e-01
6.37245834e-01 4.47359830e-01 -6.83538895e-03 8.51214051e-01
-2.01496705e-01 1.27210152e+00 -9.07332718e-01 -7.80588865e-01
2.74163008e-01 -1.28614381e-01 -7.28839993e-01 -2.94283479e-01
1.32397234e-01 -4.67263430e-01 -8.44257414e-01 1.25519824e+00
4.61530447e-01 -1.68975666e-01 4.74216312e-01 -1.56079316e+00
-6.00625157e-01 1.03816569e+00 -6.97122216e-01 4.41653430e-01
-4.62774709e-02 8.30693841e-01 1.06111407e+00 -5.30697048e-01
6.56562030e-01 1.44350195e+00 9.71576095e-01 4.87951070e-01
-1.00152695e+00 -8.90437603e-01 3.93676221e-01 1.36807367e-01
-1.12657952e+00 -6.02632016e-02 6.57276988e-01 -6.31646574e-01
9.14928496e-01 4.57243882e-02 5.01140773e-01 1.50071609e+00
-9.59495008e-02 7.82678783e-01 6.34090185e-01 1.19001031e-01
4.42749597e-02 1.58973217e-01 1.45775825e-01 1.09113944e+00
6.02212548e-01 -1.82926998e-01 -1.31908864e-01 -4.32150930e-01
4.91473556e-01 2.05244601e-01 2.25110665e-01 3.56520504e-01
-7.48595536e-01 1.34173322e+00 9.71031070e-01 3.80133629e-01
-4.01396215e-01 3.81543964e-01 7.10759044e-01 2.34183639e-01
6.04724884e-01 6.35243237e-01 -1.29140466e-01 -2.82505266e-02
-5.45707285e-01 -4.43227552e-02 9.45551276e-01 7.91247547e-01
1.15353596e+00 1.22356996e-01 -3.66548240e-01 4.12722141e-01
1.72234878e-01 8.05408239e-01 2.48343393e-01 -4.36273366e-01
5.89605987e-01 1.11031640e+00 -3.97554278e-01 -1.45783031e+00
-6.11107886e-01 -4.08238947e-01 -7.25626707e-01 -2.10617274e-01
5.92374861e-01 -4.60750729e-01 -6.64085090e-01 1.59924066e+00
4.18533921e-01 4.32291180e-01 -4.70791221e-01 4.92015123e-01
7.96067178e-01 3.42784554e-01 1.50032833e-01 4.74751629e-02
1.19160140e+00 -9.72962499e-01 -2.91818231e-01 -6.08893335e-01
9.63805139e-01 -1.58646509e-01 1.20904052e+00 -2.15235166e-03
-3.38947296e-01 -5.93667813e-02 -5.04171550e-01 1.78479806e-01
-8.13948333e-01 -2.38552764e-01 5.77676833e-01 7.39091575e-01
-9.30959702e-01 4.77815956e-01 -7.49805927e-01 -6.67998433e-01
7.43786812e-01 2.90380895e-01 -1.30841305e-04 1.51368916e-01
-1.26081252e+00 6.48985922e-01 3.89862150e-01 -1.71433061e-01
-1.05793524e+00 -3.81960064e-01 -6.08247280e-01 -2.37988457e-01
7.23790884e-01 -3.62786770e-01 1.09338653e+00 -3.89032930e-01
-1.40110934e+00 6.55143082e-01 1.61264203e-02 -5.98541975e-01
5.59604526e-01 3.53239626e-01 -1.40883297e-01 1.52781472e-01
4.32416886e-01 1.15984321e-01 1.30015385e+00 -1.06535435e+00
-6.28736079e-01 -1.28525406e-01 5.34603536e-01 -4.40839738e-01
-7.55563080e-01 1.56892478e-01 -5.04992485e-01 -3.20196271e-01
-4.92323488e-01 -9.37296629e-01 -2.97663420e-01 -5.66097558e-01
-8.93690765e-01 -5.12414217e-01 1.44841003e+00 -4.85963702e-01
1.49536860e+00 -1.92678523e+00 -1.81444660e-01 4.06125009e-01
9.76165414e-01 7.60011375e-01 -4.94654208e-01 6.66365743e-01
3.43547016e-01 5.35954654e-01 6.54150993e-02 -5.53127289e-01
-5.31302616e-02 1.46616966e-01 -2.72394449e-01 6.13279998e-01
3.40908140e-01 1.31513369e+00 -1.28952980e+00 -5.18031657e-01
3.42456885e-02 1.37474716e-01 -8.05196762e-01 2.58535415e-01
-4.96650279e-01 6.25618100e-01 -6.49984419e-01 8.16367924e-01
5.04055917e-01 -7.43833840e-01 1.96608514e-01 8.70431811e-02
1.50828660e-01 3.08757454e-01 -5.57151318e-01 7.47174323e-01
-3.74827832e-01 7.03983784e-01 1.82899043e-01 -8.45973670e-01
8.37075472e-01 -1.79851621e-01 3.78950447e-01 -3.95572156e-01
5.12321055e-01 1.01095356e-01 -4.07959037e-02 -5.51588237e-01
3.48038286e-01 4.01632011e-01 -4.82899398e-02 8.03228378e-01
-1.11366510e-01 -3.20391469e-02 5.87099493e-01 4.91311699e-01
1.75082970e+00 -7.31589019e-01 9.28863510e-02 3.16237122e-01
5.79428196e-01 7.55538344e-02 8.76426324e-02 1.27660310e+00
-3.52245122e-01 -1.63740546e-01 1.00720143e+00 -3.54304314e-01
-6.40105724e-01 -7.11200476e-01 5.62816978e-01 1.57276201e+00
3.31641436e-02 -7.55307376e-01 -9.48602319e-01 -1.58697689e+00
1.88407674e-01 4.31806684e-01 -5.67310274e-01 -1.65477440e-01
-9.40653920e-01 -8.75311553e-01 1.11896682e+00 5.12336148e-03
4.84117299e-01 -1.17317939e+00 -4.45070192e-02 2.83800572e-01
-2.07171559e-01 -1.50278211e+00 -3.96252185e-01 -1.36673123e-01
-3.18346232e-01 -1.47252393e+00 -1.19368844e-01 -3.83946508e-01
6.36561632e-01 5.42800903e-01 9.78820264e-01 6.29093051e-01
-5.26433066e-02 2.64271498e-01 -5.26282966e-01 -4.11185235e-01
-9.07196522e-01 7.79046476e-01 -1.10550527e-03 5.43317981e-02
3.47891480e-01 -7.12472737e-01 -3.50568354e-01 4.30598795e-01
-7.40054131e-01 -5.79973936e-01 4.35804486e-01 5.84637642e-01
-1.42033517e-01 2.11051628e-01 5.44847488e-01 -1.16868138e+00
1.03832841e+00 -9.39146519e-01 -8.13149035e-01 -1.55659378e-01
-3.89023095e-01 -1.98544309e-01 1.09909248e+00 -8.42962980e-01
-1.89892352e-01 -1.78941265e-01 -2.80195847e-02 -5.67695737e-01
2.21740812e-01 4.99169886e-01 2.30132625e-01 -4.01043624e-01
1.18514669e+00 3.30008194e-03 1.61547288e-01 -1.87407568e-01
5.37456989e-01 8.63497138e-01 7.02688470e-02 -2.31800169e-01
1.35527992e+00 5.51751912e-01 9.99692362e-03 -8.98581445e-01
-8.86741936e-01 -6.40797853e-01 -5.18670797e-01 -3.56842935e-01
7.30605781e-01 -6.10890031e-01 -1.36317885e+00 6.92158580e-01
-1.47975588e+00 -7.34230697e-01 -4.44090962e-02 -8.40508565e-02
-1.94460955e-02 4.91468191e-01 -9.35206592e-01 -8.04100871e-01
-3.22459608e-01 -1.24885142e+00 9.45479810e-01 -1.43499756e-02
1.37886358e-02 -1.26789606e+00 -1.09013110e-01 9.73839462e-02
6.19907916e-01 1.80506825e-01 7.82196522e-01 -1.09997439e+00
-6.39308631e-01 -6.61489427e-01 -7.46621013e-01 1.54485613e-01
2.08951503e-01 -7.25580053e-03 -9.81597781e-01 -4.54783171e-01
-2.97639906e-01 -1.36221826e-01 9.11671102e-01 -6.52321726e-02
9.93151307e-01 -8.27663898e-01 -6.01013184e-01 4.47926521e-01
1.04396725e+00 -3.66085231e-01 1.07575998e-01 2.92294890e-01
1.37898147e+00 3.87783825e-01 6.15487956e-02 4.73197043e-01
5.32080770e-01 3.16923201e-01 1.04297125e+00 -4.54291292e-02
-8.74852240e-02 -6.85514271e-01 6.68947279e-01 6.03062034e-01
1.16604581e-01 -6.54390693e-01 -9.69064534e-01 2.17662007e-01
-1.93318844e+00 -1.08400452e+00 -4.82150763e-01 1.69444978e+00
5.66754520e-01 1.73819467e-01 9.86654401e-01 -1.27876014e-01
1.00850022e+00 4.88612890e-01 -6.68768227e-01 -9.87799391e-02
1.16620101e-01 -3.91435176e-02 1.04797566e+00 7.27978766e-01
-1.17053533e+00 1.69108152e+00 6.20733213e+00 1.13383651e+00
-1.46997929e+00 5.21893382e-01 1.98942259e-01 2.88341939e-01
1.47382006e-01 -5.32456301e-02 -9.87641454e-01 5.83729625e-01
8.54913175e-01 -1.07992128e-01 1.08164918e+00 9.69566643e-01
1.98354974e-01 3.58873785e-01 -7.87107289e-01 7.82391965e-01
-2.34935939e-01 -1.34506464e+00 3.85664441e-02 3.30750048e-01
4.28353518e-01 7.08929300e-01 4.32793377e-03 6.05967283e-01
8.46252799e-01 -9.66138124e-01 4.63348329e-01 4.72348258e-02
6.40175402e-01 -4.29394454e-01 6.67249501e-01 9.06076670e-01
-1.28945792e+00 -5.75442255e-01 -5.08725941e-01 -6.47067577e-02
-2.47306265e-02 5.91681957e-01 -1.53037572e+00 2.75953770e-01
3.17588568e-01 1.02558112e+00 -9.07000303e-01 5.86955547e-01
-7.27835953e-01 9.86572981e-01 -3.67227435e-01 -6.26200140e-01
5.18778265e-01 8.85435641e-02 9.19381261e-01 1.50197518e+00
-1.19424529e-01 -4.30777252e-01 5.65219462e-01 1.14306891e+00
-3.47819418e-01 -3.93543802e-02 -9.79284227e-01 -5.71432710e-01
6.59119666e-01 1.48813570e+00 -7.40891159e-01 -3.49207878e-01
-1.02446750e-01 7.43504882e-01 8.35606277e-01 2.83184320e-01
-9.42320764e-01 -1.99917778e-01 6.31340683e-01 4.11395222e-01
1.47230819e-01 -2.95137674e-01 -4.25389893e-02 -9.59241271e-01
-4.81749266e-01 -1.08460879e+00 6.51020825e-01 -4.60105911e-02
-1.68026972e+00 6.57612562e-01 -1.40565708e-01 -8.27798724e-01
-3.30374986e-01 -8.06792855e-01 -8.59775245e-01 5.58738530e-01
-1.33707166e+00 -1.24928713e+00 -2.73899257e-01 7.20455647e-01
5.99601418e-02 -3.25370729e-01 4.03750747e-01 3.12461913e-01
-9.49661374e-01 8.02590251e-01 -3.44822854e-01 7.44136930e-01
3.52950096e-01 -9.86092448e-01 8.28748167e-01 9.18145657e-01
-7.05090351e-03 5.16099334e-01 4.26342607e-01 -1.01896799e+00
-1.52530944e+00 -1.62195861e+00 5.50979137e-01 -8.65275323e-01
1.45778680e+00 -7.00539589e-01 -6.52211249e-01 8.26765716e-01
-1.80334315e-01 1.42145887e-01 6.16882695e-03 2.28822492e-02
-7.51363814e-01 1.14241466e-01 -1.17052686e+00 7.33746350e-01
1.24470413e+00 -7.87518859e-01 2.25642264e-01 9.05446470e-01
9.04790878e-01 -1.76813394e-01 -2.95110643e-01 3.64332460e-02
6.15024343e-02 -8.83397162e-01 7.88795590e-01 -8.45278382e-01
-1.48925483e-01 -1.97547868e-01 3.08102965e-01 -1.17312801e+00
-4.25545782e-01 -1.02971125e+00 -3.06990594e-01 1.09004188e+00
4.79993224e-01 -1.20191145e+00 9.64316607e-01 -1.24602057e-01
3.44869375e-01 -3.89760464e-01 -6.42475367e-01 -1.05191183e+00
-3.08845332e-03 -4.36077267e-01 5.84205687e-01 1.00916064e+00
-1.46615701e-02 5.52361906e-01 -3.67636204e-01 4.19523299e-01
5.90750635e-01 -3.69439125e-01 1.37537801e+00 -1.27477419e+00
-3.20245981e-01 -7.08251774e-01 -3.76268685e-01 -1.06530523e+00
6.03052735e-01 -1.19405746e+00 -1.86815977e-01 -1.41977847e+00
4.52448502e-02 -5.39914370e-01 8.85732174e-02 8.44012737e-01
-2.46427611e-01 4.94107544e-01 3.51097465e-01 3.87254238e-01
-7.04669476e-01 2.23483369e-02 1.22353852e+00 -5.01748860e-01
-3.68289918e-01 1.97076142e-01 -7.14639902e-01 8.76018822e-01
1.01142502e+00 -9.30091679e-01 -2.13715941e-01 -2.75285482e-01
3.73723775e-01 -5.36490858e-01 3.73079658e-01 -6.41416728e-01
4.11738217e-01 -1.72283307e-01 -3.95145088e-01 -3.02794725e-01
8.81194621e-02 -5.67298889e-01 -4.48230803e-01 7.79163599e-01
2.33290300e-01 1.10904865e-01 1.54199069e-02 6.53411031e-01
2.83690542e-01 -1.96955875e-01 8.47195327e-01 -2.57533610e-01
-2.67640054e-01 7.95269966e-01 -5.94113052e-01 5.05850375e-01
8.23823094e-01 4.61673550e-02 -9.23880100e-01 -6.41382992e-01
-1.88384965e-01 1.47643521e-01 3.19449663e-01 2.37303078e-01
1.26862153e-01 -8.48356545e-01 -7.21980929e-01 1.05372056e-01
9.16935280e-02 -2.74710029e-01 -4.38583642e-01 1.10513413e+00
-4.00786281e-01 5.64089678e-02 4.77402002e-01 -7.82652915e-01
-9.87944841e-01 8.11196566e-01 3.31015140e-01 -6.96191430e-01
-3.60453933e-01 8.23995709e-01 -4.44581769e-02 -7.89761543e-01
1.66022167e-01 -2.66471468e-02 -1.45182982e-01 1.36874050e-01
4.14892763e-01 5.39439917e-01 -1.55118868e-01 -5.84109068e-01
-4.08923924e-01 2.15028882e-01 -1.77437011e-02 3.01843315e-01
1.11213851e+00 1.87578455e-01 -4.90383476e-01 1.10487500e-02
1.09990335e+00 4.73761350e-01 -7.71778762e-01 -3.52942824e-01
1.20247394e-01 -4.49789077e-01 -4.32691336e-01 -4.35389251e-01
-1.13313580e+00 6.84577942e-01 -1.82247236e-01 1.01824284e+00
5.69299996e-01 1.71312109e-01 8.66519153e-01 7.64378607e-01
2.61460274e-01 -6.64942026e-01 5.81512332e-01 1.18093777e+00
4.40631628e-01 -1.24324071e+00 -2.76944816e-01 -6.61052644e-01
-2.50231385e-01 9.17784393e-01 7.14285791e-01 -2.98465669e-01
6.51013792e-01 2.52731144e-01 1.24022458e-02 -5.19450128e-01
-3.92926663e-01 -4.97654945e-01 -1.17887676e-01 8.34475100e-01
-2.20749423e-01 4.62252833e-02 1.08040571e-01 2.83747822e-01
-2.43965730e-01 -5.64675093e-01 3.36129457e-01 4.65866089e-01
-5.18790126e-01 -1.06857836e+00 -1.40905052e-01 4.01603699e-01
-2.82323509e-01 -2.20255286e-01 -1.02299559e+00 9.58396792e-01
-2.93329470e-02 1.38427639e+00 -3.36677313e-01 -9.08912003e-01
2.96252787e-01 -3.25164884e-01 -1.40197668e-03 -7.72260487e-01
-9.94236648e-01 -5.83813190e-01 2.25771740e-01 -6.81783676e-01
5.93986325e-02 -8.11618716e-02 -1.08819258e+00 -1.00889015e+00
-6.43465161e-01 1.18960366e-01 4.82718259e-01 9.27263141e-01
4.86497372e-01 2.18765035e-01 9.72986460e-01 -8.72016072e-01
-6.21155262e-01 -1.13695145e+00 -2.54177868e-01 2.95286685e-01
3.68383706e-01 -4.14101154e-01 -7.97977746e-01 -6.59672618e-01] | [8.072781562805176, 10.110623359680176] |
a17bf800-b330-4bb6-ae5d-2597b34888f6 | heuristic-dropout-an-efficient-regularization | null | null | https://ieeexplore.ieee.org/abstract/document/9747409 | https://ieeexplore.ieee.org/abstract/document/9747409 | Heuristic Dropout: An Efficient Regularization Method for Medical Image Segmentation Models | For medical image segmentation in a real scenario, the amount of accurate annotation data at the pixel level is typically small, which tends to cause an overfitting problem. This manuscript goes deep into the research of the Dropout algorithm, which is commonly used in neural networks to alleviate the overfitting problem. From the perspective of solving the co-adaptation problem, this manuscript explains the basic principles of the Dropout algorithm and discusses the existing limitations of its derivative methods. Furthermore, we propose a novel Heuristic Dropout algorithm to address these limitations. The proposed algorithm takes information entropy and variance as heuristic rules. It guides our algorithm to drop features suffering from co-adaptation problem more efficiently and thus can better alleviate the overfitting problem of small-scale medical image segmentation datasets. Experiments on medical image segmentation datasets and models show that the proposed algorithm significantly improves the performance of these models. | ['Chun Yuan', 'Linmi Tao', 'Ruiyang Liu', 'Dachuan Shi'] | 2022-05-23 | null | null | null | international-conference-on-acoustics-speech | ['medical-image-segmentation'] | ['medical'] | [ 3.38661492e-01 1.69608489e-01 -4.23392683e-01 -4.80715841e-01
-6.68775380e-01 1.60983086e-01 -4.79155779e-01 -7.30104297e-02
-8.78754735e-01 7.38570571e-01 -1.14286125e-01 -1.24991789e-01
3.40625234e-02 -5.34724295e-01 -6.14928424e-01 -9.16575015e-01
4.03690904e-01 -1.06466182e-01 1.32082552e-01 3.53650331e-01
1.62017465e-01 -1.25279957e-02 -1.03433561e+00 3.96233834e-02
1.37017274e+00 7.81955659e-01 2.57321596e-01 2.38261178e-01
-1.18204899e-01 7.48294055e-01 -6.82654500e-01 -4.65452820e-01
1.29384458e-01 -6.49209976e-01 -6.71348989e-01 3.12222987e-01
3.53944421e-01 -5.58422983e-01 -5.25488317e-01 1.38793898e+00
6.90031350e-01 2.27678582e-01 4.86985713e-01 -1.04742324e+00
-6.74918175e-01 7.51220644e-01 -6.30545259e-01 4.18349653e-01
-5.59562624e-01 3.76113020e-02 5.04932702e-01 -3.33833814e-01
4.02002752e-01 9.43192661e-01 1.02872026e+00 7.65135646e-01
-8.43257964e-01 -5.82936645e-01 1.60759062e-01 6.48081154e-02
-1.39684689e+00 -2.00776421e-02 8.14319432e-01 -1.45413920e-01
7.23059237e-01 1.46588357e-02 6.55386567e-01 1.00402331e+00
3.70283186e-01 1.26163089e+00 7.81920314e-01 -1.50154606e-01
2.68654913e-01 2.99454004e-01 5.62009335e-01 7.75099099e-01
4.15901661e-01 -2.07470134e-01 -8.96473676e-02 -9.90754291e-02
1.06194282e+00 7.27603361e-02 -3.18937153e-01 5.10222353e-02
-4.79096204e-01 8.96151781e-01 4.54805195e-01 3.42983812e-01
-1.76564112e-01 3.07546183e-02 4.74824876e-01 1.12801209e-01
4.27353978e-01 1.77830562e-01 -4.86199707e-01 -8.93750973e-03
-1.20581830e+00 -1.17994674e-01 3.59088749e-01 8.95432055e-01
3.90250355e-01 2.47785553e-01 -2.76361734e-01 1.05665290e+00
2.49012589e-01 4.31053862e-02 8.30636561e-01 -1.01051915e+00
3.03092003e-01 6.31865144e-01 -2.48657450e-01 -7.73582757e-01
-5.68474233e-01 -9.23521280e-01 -1.12894142e+00 -1.26974747e-01
4.27298665e-01 -3.15792561e-01 -1.13646388e+00 1.81103778e+00
1.83164313e-01 4.10739243e-01 -1.49918318e-01 9.38218117e-01
9.13623929e-01 3.66030693e-01 1.72851846e-01 -5.32984793e-01
1.05765951e+00 -1.08380020e+00 -1.33814085e+00 -1.40861511e-01
8.04294348e-01 -4.23474193e-01 1.32534516e+00 3.76323402e-01
-1.08713174e+00 -4.82713342e-01 -1.17256320e+00 -1.72037378e-01
2.58819815e-02 2.69132346e-01 7.75548935e-01 8.51819694e-01
-7.11853743e-01 7.64484525e-01 -1.22301805e+00 -9.98936445e-02
8.81176353e-01 5.65135300e-01 6.45490587e-02 4.56822179e-02
-1.18352985e+00 6.86655462e-01 5.10799587e-01 3.89367402e-01
-3.71384472e-01 -7.25405753e-01 -6.13461554e-01 -9.83563438e-02
3.44377756e-01 -7.97013044e-01 1.09522378e+00 -1.25117540e+00
-1.37420154e+00 6.31111562e-01 -1.86321095e-01 -4.83126193e-01
7.33449399e-01 -4.29809183e-01 -2.15075277e-02 -3.90636316e-03
-2.11856201e-01 6.36711538e-01 6.46761060e-01 -8.90521049e-01
-4.10513818e-01 -4.71174181e-01 -3.94593269e-01 2.38069236e-01
-6.72169983e-01 -2.44524330e-01 -7.06680894e-01 -7.45919347e-01
1.82844251e-01 -8.03430200e-01 -6.23457789e-01 -1.58214703e-01
-4.66265768e-01 -7.26913940e-03 6.68437183e-01 -5.87204158e-01
1.48784959e+00 -2.32413888e+00 -2.41503462e-01 -2.08668821e-02
2.31134206e-01 3.73239338e-01 2.37304475e-02 -3.55140328e-01
2.50800028e-02 2.49779880e-01 -7.12634087e-01 -3.47655684e-01
-4.58282560e-01 4.20344919e-01 2.47187704e-01 4.69794154e-01
9.35101733e-02 7.36610889e-01 -7.19511986e-01 -7.63648331e-01
-9.94156674e-03 7.38812685e-01 -6.35510564e-01 7.83319548e-02
2.65738100e-01 3.57901871e-01 -7.27813423e-01 4.08535659e-01
1.01183605e+00 -2.18067437e-01 -1.77775621e-01 -2.03080446e-01
1.48099065e-01 -2.64603272e-02 -1.12591302e+00 1.68701994e+00
-1.49635956e-01 5.00789523e-01 -8.29301104e-02 -1.05520225e+00
8.41304958e-01 3.32631081e-01 7.11329639e-01 -2.87281096e-01
6.22844875e-01 1.71804667e-01 5.52506484e-02 -8.48044574e-01
2.56168485e-01 -1.93499386e-01 3.74799758e-01 -2.75473539e-02
6.86268508e-02 1.63822815e-01 3.66198905e-02 -3.44623476e-02
9.87216532e-01 -6.62337709e-03 1.23445652e-01 -1.49869591e-01
2.87678897e-01 -1.64554000e-01 1.21399152e+00 9.14989889e-01
-7.13604331e-01 9.07715261e-01 4.92815256e-01 -3.68582904e-01
-8.67871404e-01 -8.17270398e-01 -3.50851595e-01 5.28300941e-01
-1.39745492e-02 -2.13718303e-02 -1.32835329e+00 -7.33309329e-01
-2.36993849e-01 5.69649339e-01 -7.82341480e-01 -5.34477592e-01
-5.30800700e-01 -1.63032329e+00 6.84377909e-01 6.77472472e-01
9.45530236e-01 -9.49204206e-01 -6.24178350e-01 2.50286222e-01
-3.88417393e-01 -9.55568016e-01 -5.66446781e-01 2.54609257e-01
-1.66481602e+00 -8.53606284e-01 -8.93041492e-01 -9.91511285e-01
1.06697989e+00 8.12844187e-02 6.94290578e-01 3.73710006e-01
-4.16731179e-01 -1.51862815e-01 -5.25823794e-02 -3.58524114e-01
-2.12228090e-01 2.69933462e-01 -2.04268873e-01 -1.52286842e-01
5.99208832e-01 -4.13958818e-01 -5.79798937e-01 -9.38448124e-03
-1.10024405e+00 -3.44762765e-02 7.88852274e-01 1.13679767e+00
8.58162344e-01 3.50237310e-01 5.33314764e-01 -1.32172680e+00
7.68754184e-01 -2.37921327e-01 -4.81588989e-01 2.31337205e-01
-9.37757552e-01 -7.93889910e-02 6.16253972e-01 -6.10058784e-01
-1.12832332e+00 2.63781488e-01 -2.40114823e-01 -4.43797916e-01
-6.67982846e-02 4.06434536e-01 -2.44581342e-01 -1.07116587e-01
3.99280310e-01 1.39636710e-01 -3.64587940e-02 -6.00473940e-01
-7.97552094e-02 6.79860473e-01 4.60403264e-01 -9.74910557e-02
2.43386656e-01 2.59332359e-01 -1.38569027e-01 -9.14359629e-01
-9.50784981e-01 -2.60733634e-01 -6.15254581e-01 2.52840202e-02
7.88001180e-01 -7.53139198e-01 -6.22883081e-01 7.35523045e-01
-9.19268787e-01 -3.21320921e-01 -2.87943691e-01 7.25256085e-01
-5.64228535e-01 4.81786102e-01 -1.10634053e+00 -7.75743246e-01
-4.49523389e-01 -1.39469731e+00 3.30431283e-01 7.55310416e-01
5.49892429e-03 -1.05061114e+00 -1.93808928e-01 4.78299141e-01
2.34586075e-01 1.70182765e-01 9.98174012e-01 -4.95589852e-01
-4.15234208e-01 -1.54294312e-01 -2.26004645e-02 7.02651739e-01
-3.55082266e-02 2.58633681e-02 -8.91621709e-01 -2.21491724e-01
6.66081607e-01 -1.25855580e-01 1.08357036e+00 1.09636974e+00
1.42074156e+00 -6.94645867e-02 -2.30412170e-01 9.49917257e-01
1.57100606e+00 2.94315010e-01 7.49212980e-01 3.24110925e-01
6.08651042e-01 3.57150704e-01 4.97453153e-01 2.87018180e-01
1.08805977e-01 1.57120228e-01 1.59891173e-01 -4.82069641e-01
-1.15894161e-01 -1.20583341e-01 7.09624216e-02 1.27424705e+00
1.30059466e-01 -1.72202751e-01 -6.23132885e-01 5.44216692e-01
-1.91533744e+00 -6.37283742e-01 -4.47475016e-01 2.00965738e+00
1.04077578e+00 2.53553599e-01 -1.60157159e-01 1.23448804e-01
7.51092732e-01 -2.41125748e-01 -8.72814059e-01 -4.03331190e-01
-2.71884769e-01 -9.35728922e-02 7.88343310e-01 4.39424664e-01
-1.22028637e+00 9.35797155e-01 7.34573317e+00 9.40619826e-01
-9.31035697e-01 3.15514535e-01 1.00164807e+00 -1.54277205e-01
5.38082942e-02 -2.31904864e-01 -8.33747923e-01 5.50843775e-01
8.25856507e-01 8.98338258e-02 2.19463289e-01 7.32394457e-01
5.47092795e-01 -2.85708755e-01 -7.47740030e-01 9.79105830e-01
-4.70245369e-02 -9.87823844e-01 5.58956352e-04 -2.35959589e-01
8.23884368e-01 -6.47771219e-03 4.09229577e-01 1.99088812e-01
-6.48377314e-02 -8.79552126e-01 1.37292966e-01 3.75118673e-01
2.84388840e-01 -9.04639184e-01 1.10480988e+00 4.49912995e-01
-6.30973518e-01 -1.63470045e-01 -7.31880605e-01 1.31442830e-01
6.56909049e-02 9.49183583e-01 -4.02186692e-01 2.13562369e-01
6.41908407e-01 6.39537692e-01 -5.13423979e-01 1.50698841e+00
-1.37262493e-01 8.80248845e-01 -3.68222952e-01 8.69653746e-02
3.04870456e-01 -4.87994373e-01 2.29433969e-01 1.12090838e+00
1.96938097e-01 8.97804722e-02 -1.00797899e-01 9.87483621e-01
-1.97686866e-01 2.70181656e-01 -2.34229133e-01 1.12474442e-01
2.44576082e-01 9.28683996e-01 -9.67302978e-01 -2.02038497e-01
-5.40377498e-01 9.56083953e-01 2.16588631e-01 5.00314653e-01
-1.07178414e+00 -5.95004439e-01 1.50139123e-01 -1.91754565e-01
2.67468300e-02 1.97485611e-01 -8.43091309e-01 -9.80727911e-01
8.52298215e-02 -7.52509117e-01 4.48402107e-01 -5.28968573e-01
-9.98874426e-01 4.69140351e-01 -2.81830192e-01 -9.69368637e-01
4.78431731e-01 -2.60491103e-01 -4.67545450e-01 5.18189073e-01
-1.50370622e+00 -6.63642108e-01 -1.61236733e-01 4.37105507e-01
6.14369690e-01 1.59060344e-01 5.18991351e-01 7.60612428e-01
-1.24441934e+00 1.02784562e+00 4.88121688e-01 3.41371715e-01
5.89250505e-01 -8.86137128e-01 -6.81132972e-02 8.47172618e-01
-3.04024130e-01 7.08011270e-01 6.64344192e-01 -7.39734828e-01
-7.87323356e-01 -1.12010777e+00 6.18925631e-01 1.52024060e-01
1.69809461e-01 1.14590436e-01 -1.28373301e+00 5.59761882e-01
-9.37370677e-03 -6.69679120e-02 9.26307917e-01 -7.82031044e-02
3.07358772e-01 -6.07484579e-02 -1.37020767e+00 5.86008966e-01
9.01988268e-01 -1.16645880e-01 -4.70990807e-01 3.28015327e-01
5.64941823e-01 -5.48798025e-01 -9.02049482e-01 5.18927574e-01
2.74081945e-01 -8.33667934e-01 6.28359616e-01 -5.12919426e-01
3.11731517e-01 7.54720643e-02 3.42001975e-01 -1.08763397e+00
-2.60919303e-01 -7.09631205e-01 -1.27139352e-02 1.16632915e+00
4.94631499e-01 -3.51595670e-01 1.13424206e+00 8.96113276e-01
-1.07498333e-01 -9.06579912e-01 -1.06769848e+00 -7.01815486e-01
3.45248401e-01 -2.43435144e-01 1.58740133e-01 9.29235101e-01
-1.26284033e-01 8.86937827e-02 -3.38885158e-01 -1.33563861e-01
6.78524017e-01 -4.20449644e-01 1.95541114e-01 -9.85064149e-01
-1.03057571e-01 -5.27054429e-01 -2.34079733e-01 -1.11093760e+00
3.84230912e-02 -7.94816375e-01 1.33516863e-01 -1.42078185e+00
5.50377965e-01 -2.05810279e-01 -6.42582238e-01 3.80475402e-01
-6.06926382e-01 1.41012162e-01 -6.96081519e-02 1.31965056e-01
-3.69061261e-01 6.42542541e-01 1.43607759e+00 -1.50317952e-01
-4.20971155e-01 2.50250936e-01 -6.11569107e-01 9.90672350e-01
9.18142915e-01 -9.67753887e-01 -5.85552871e-01 -7.83252537e-01
-1.48834378e-01 -8.81368816e-02 -1.76578406e-02 -9.62570608e-01
1.33260831e-01 8.52681417e-03 3.97808760e-01 -4.22679961e-01
1.37147233e-02 -9.04285252e-01 -1.65446430e-01 7.08662450e-01
-5.96152604e-01 -6.76209852e-02 1.78422928e-01 4.41824406e-01
-2.57483155e-01 -6.71359599e-01 1.12700450e+00 -1.89043745e-01
-2.24772096e-01 4.45492446e-01 -4.53525126e-01 2.77276188e-01
7.94912219e-01 -3.83485168e-01 6.34431466e-02 -1.35111243e-01
-8.38187695e-01 3.11911434e-01 2.65208691e-01 1.09717429e-01
7.31870830e-01 -1.11947930e+00 -4.43480998e-01 1.97560027e-01
-4.44337279e-01 5.94816431e-02 5.31205595e-01 1.23390615e+00
-5.01155734e-01 1.52319208e-01 -1.14362642e-01 -4.15117711e-01
-1.24257445e+00 3.77202183e-01 5.94560444e-01 -2.16474429e-01
-7.92061687e-01 1.05623865e+00 -3.04983985e-02 -1.35345070e-03
6.95799053e-01 -5.83874226e-01 -2.88226068e-01 -2.39257962e-01
4.21332568e-01 4.52019185e-01 9.18038376e-03 -3.16843629e-01
-1.47545502e-01 5.81918895e-01 -3.14076930e-01 1.40623197e-01
1.20211208e+00 -2.07747608e-01 3.86469848e-02 4.89383280e-01
1.36856449e+00 -4.89455611e-01 -1.54098105e+00 -1.66217029e-01
-5.95345870e-02 -2.92682856e-01 5.17224073e-01 -6.26694858e-01
-1.54916680e+00 1.06566691e+00 9.83006716e-01 -2.23079190e-01
1.32355630e+00 -5.41988969e-01 1.24766064e+00 3.21271867e-01
-3.15298550e-02 -1.61476552e+00 -2.87555128e-01 3.61385792e-01
1.23139724e-01 -1.32003808e+00 6.55046552e-02 -4.27173018e-01
-6.55836940e-01 1.02466238e+00 7.68917501e-01 -1.82666168e-01
6.87528491e-01 5.07785618e-01 2.82499254e-01 1.92044154e-01
-4.31195527e-01 1.12574406e-01 -5.33161946e-02 5.31403482e-01
4.92148250e-01 -2.14433253e-01 -8.69660139e-01 9.45168138e-01
9.57984030e-02 4.60644543e-01 5.92811167e-01 8.13575506e-01
-3.92865002e-01 -9.07247365e-01 -2.20859438e-01 5.44255018e-01
-9.48503613e-01 -7.83063024e-02 -1.47841603e-01 7.71535397e-01
2.72543311e-01 7.65683651e-01 -4.70397286e-02 -1.82821110e-01
2.26511508e-01 5.56376437e-03 4.35258210e-01 -2.03059033e-01
-7.70295858e-01 3.91420901e-01 -4.76304621e-01 -3.62465441e-01
-4.04363096e-01 -5.34745872e-01 -1.63345158e+00 -1.17765196e-01
-6.18381798e-01 1.68972835e-01 3.98903728e-01 8.48030150e-01
1.56956986e-01 8.89519930e-01 3.27327579e-01 -4.29356620e-02
-6.87002063e-01 -9.44478512e-01 -6.87782884e-01 5.06018877e-01
2.97068447e-01 -4.35044527e-01 -4.12016481e-01 4.96102832e-02] | [14.625158309936523, -2.3288018703460693] |
982b0907-349a-4828-ab2b-98d95284fd00 | on-the-effectiveness-of-regularization | 2006.05336 | null | https://arxiv.org/abs/2006.05336v1 | https://arxiv.org/pdf/2006.05336v1.pdf | On the Effectiveness of Regularization Against Membership Inference Attacks | Deep learning models often raise privacy concerns as they leak information about their training data. This enables an adversary to determine whether a data point was in a model's training set by conducting a membership inference attack (MIA). Prior work has conjectured that regularization techniques, which combat overfitting, may also mitigate the leakage. While many regularization mechanisms exist, their effectiveness against MIAs has not been studied systematically, and the resulting privacy properties are not well understood. We explore the lower bound for information leakage that practical attacks can achieve. First, we evaluate the effectiveness of 8 mechanisms in mitigating two recent MIAs, on three standard image classification tasks. We find that certain mechanisms, such as label smoothing, may inadvertently help MIAs. Second, we investigate the potential of improving the resilience to MIAs by combining complementary mechanisms. Finally, we quantify the opportunity of future MIAs to compromise privacy by designing a white-box `distance-to-confident' (DtC) metric, based on adversarial sample crafting. Our metric reveals that, even when existing MIAs fail, the training samples may remain distinguishable from test samples. This suggests that regularization mechanisms can provide a false sense of privacy, even when they appear effective against existing MIAs. | ['Tudor Dumitras', 'Sanghyun Hong', 'Yigitcan Kaya'] | 2020-06-09 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [ 4.58703667e-01 4.18987274e-01 -5.24114072e-02 -3.97149980e-01
-1.12570679e+00 -1.16834104e+00 5.34281850e-01 2.47171775e-01
-5.25423110e-01 6.59591138e-01 6.42837510e-02 -9.03780937e-01
5.40818423e-02 -7.87817717e-01 -1.03910875e+00 -8.16659510e-01
-2.69418210e-01 -3.18456650e-01 -6.51242882e-02 2.16264799e-01
2.51299262e-01 6.66069567e-01 -9.26110029e-01 3.93035799e-01
5.36665201e-01 8.72110546e-01 -8.16911876e-01 5.74625969e-01
2.91937590e-01 7.73597598e-01 -1.01428986e+00 -1.03780293e+00
8.62230897e-01 -1.63296103e-01 -8.09386075e-01 -2.05846146e-01
6.80582404e-01 -6.51012361e-01 -3.75433862e-01 1.30704033e+00
2.03837574e-01 -2.64619231e-01 3.78185332e-01 -1.66032076e+00
-6.05855107e-01 5.11041403e-01 -3.99632752e-01 4.33484092e-02
3.25426571e-02 3.17933053e-01 8.18191051e-01 -2.77901590e-01
4.17032778e-01 1.04708374e+00 8.63892794e-01 8.48722398e-01
-1.56838059e+00 -1.05240655e+00 -1.32859334e-01 -2.27042049e-01
-1.33083522e+00 -7.34653711e-01 5.88942051e-01 -3.34013581e-01
4.72694665e-01 7.98698485e-01 1.25293300e-01 1.22237825e+00
9.93731320e-02 6.19022727e-01 1.36609721e+00 -2.63968885e-01
3.38997513e-01 5.67183793e-01 2.22992212e-01 5.42874336e-01
7.48632073e-01 3.60529333e-01 -3.19534838e-01 -1.05758214e+00
4.85415697e-01 -1.97404355e-01 -5.23410141e-01 -3.68187428e-01
-4.83045816e-01 1.04054606e+00 2.94512659e-01 -8.65817294e-02
1.86797947e-01 2.37098172e-01 5.31780481e-01 5.73761880e-01
3.30120802e-01 7.69308209e-01 -5.72055936e-01 2.89057136e-01
-6.19738877e-01 3.66067410e-01 1.00359023e+00 8.15214992e-01
7.50260472e-01 -2.46668398e-01 2.20975373e-03 7.21779317e-02
1.96096927e-01 4.27095070e-02 2.04349365e-02 -1.08591282e+00
5.08861184e-01 2.85719752e-01 2.07874462e-01 -1.26301777e+00
2.90266126e-02 -3.31480742e-01 -5.15929878e-01 6.66517198e-01
7.36147881e-01 -4.29254651e-01 -5.30637503e-01 2.06706238e+00
1.18019976e-01 7.82724991e-02 2.65080154e-01 6.51093006e-01
7.59218335e-02 2.20224142e-01 2.34100387e-01 -9.18354467e-03
1.11354005e+00 -3.07347208e-01 -3.66719842e-01 -2.59487063e-01
9.24654782e-01 -3.94678950e-01 9.02139783e-01 4.10713553e-01
-9.90962803e-01 1.04331471e-01 -1.36570692e+00 -6.46986021e-03
-4.54319298e-01 -4.31527883e-01 8.64642084e-01 1.44840717e+00
-8.37042928e-01 6.11444235e-01 -9.02790129e-01 1.11998897e-03
9.99634922e-01 3.96514118e-01 -7.25706577e-01 5.68780527e-02
-1.24245155e+00 6.50336266e-01 1.03957400e-01 -5.11275418e-03
-8.42929959e-01 -9.07029748e-01 -7.25424826e-01 5.98262809e-02
3.20800751e-01 -3.35931897e-01 9.75348830e-01 -9.57235336e-01
-8.54708433e-01 1.05172181e+00 1.58780992e-01 -1.07962394e+00
8.67648005e-01 -2.02516448e-02 -3.94717336e-01 1.56817615e-01
-3.33573282e-01 4.53929484e-01 9.01226223e-01 -1.52930605e+00
-3.61343443e-01 -6.09951317e-01 4.03267622e-01 -2.29817092e-01
-6.83796525e-01 -4.27142484e-03 1.31103873e-01 -6.89777851e-01
-2.26812467e-01 -1.05034089e+00 -3.94390881e-01 5.66654205e-01
-7.16107726e-01 2.99193203e-01 1.09148312e+00 -4.14758503e-01
1.15777600e+00 -2.43855047e+00 -7.55655408e-01 5.31951666e-01
4.49609041e-01 6.18679702e-01 5.72717711e-02 1.02754109e-01
1.18943520e-01 8.57530415e-01 -5.15071630e-01 -5.16938925e-01
1.38957247e-01 2.48819932e-01 -6.50739074e-01 9.25214410e-01
9.76640284e-02 7.90968359e-01 -4.39824879e-01 -2.59871744e-02
-1.85740083e-01 4.14475262e-01 -7.79568434e-01 1.63989738e-02
-2.37390310e-01 2.14089453e-01 -2.22996294e-01 4.77002829e-01
1.25398374e+00 -1.25502825e-01 1.59613729e-01 1.99166089e-02
3.70100439e-01 3.06468487e-01 -9.51028824e-01 9.65533972e-01
-3.78354453e-02 8.25872183e-01 4.15033519e-01 -8.58661652e-01
6.38652742e-01 1.76375136e-01 1.03762865e-01 -2.83475876e-01
8.63123238e-02 6.82230890e-02 -8.38509798e-02 -2.47554556e-01
2.80688941e-01 -1.76032931e-01 -2.01736480e-01 6.42946959e-01
-4.68323052e-01 3.94786239e-01 -6.20116472e-01 7.47774765e-02
1.40291214e+00 -3.61336648e-01 -1.14861466e-01 -3.21479857e-01
2.71480441e-01 -1.33941174e-01 7.25702643e-01 1.15589046e+00
-6.26853585e-01 4.89171028e-01 6.67628169e-01 -2.82479137e-01
-9.47793901e-01 -9.51521337e-01 -3.51590931e-01 7.83709526e-01
3.44581269e-02 -4.29479778e-01 -1.04443145e+00 -1.17326951e+00
4.32040572e-01 6.37421548e-01 -6.69346690e-01 -5.25928378e-01
-2.53326863e-01 -7.35137761e-01 1.30873764e+00 2.37292975e-01
7.30332375e-01 -4.48952258e-01 -5.94906628e-01 -1.18744388e-01
3.67018580e-02 -1.03454936e+00 -4.70016479e-01 1.83580756e-01
-7.33734906e-01 -1.21412849e+00 -8.42608064e-02 -2.15526760e-01
8.09649587e-01 2.66906440e-01 6.79385900e-01 3.63386750e-01
-1.62028641e-01 3.84397209e-01 -1.84716415e-02 -5.35485089e-01
-8.40335310e-01 -1.38546556e-01 -8.52156579e-02 -6.42541274e-02
7.36337900e-01 -5.25896370e-01 -6.61461830e-01 2.64055997e-01
-1.14272928e+00 -5.49802899e-01 2.39698291e-01 4.40236270e-01
1.51591778e-01 3.13502878e-01 4.89000618e-01 -1.51544595e+00
8.13037932e-01 -4.90366697e-01 -6.14709020e-01 2.31105357e-01
-7.51840949e-01 8.96983966e-02 7.28168845e-01 -5.86144149e-01
-8.20996821e-01 -1.05756223e-01 -1.88047469e-01 -4.96354669e-01
-3.28747064e-01 2.06317320e-01 -4.44381207e-01 -7.30808914e-01
9.67853308e-01 2.00887956e-03 1.43165559e-01 -3.35670918e-01
2.93592304e-01 7.23449469e-01 4.79120404e-01 -7.48192549e-01
8.90647352e-01 6.33300602e-01 2.01890454e-01 -4.66258734e-01
-7.90595710e-01 -1.86716653e-02 -6.98315576e-02 2.55264193e-01
6.37985647e-01 -8.58268619e-01 -1.03830838e+00 5.62392116e-01
-9.96973634e-01 -2.38043740e-01 -3.66693214e-02 1.58207223e-01
-2.90303230e-01 6.82961226e-01 -7.15494514e-01 -8.17115545e-01
-2.32142895e-01 -1.15695739e+00 4.72543895e-01 -5.93712851e-02
-3.21103394e-01 -9.69245553e-01 -3.88757586e-01 7.11472869e-01
3.74731362e-01 6.22544289e-01 9.64418173e-01 -1.06056595e+00
-6.78265691e-01 -6.16057277e-01 -6.77056238e-02 6.53178573e-01
7.26756603e-02 -2.59374559e-01 -1.36904848e+00 -7.09725022e-01
5.03253937e-01 -3.81830186e-01 7.60976255e-01 -6.05401397e-02
1.56268692e+00 -1.21931577e+00 -1.43633708e-01 7.85682440e-01
1.42189789e+00 8.72906744e-02 9.34379578e-01 3.40254307e-01
5.58327019e-01 5.33888459e-01 1.69459969e-01 3.75041485e-01
4.64509567e-03 2.95756340e-01 5.62527180e-01 -7.01122060e-02
4.77958888e-01 -2.47462571e-01 3.40129018e-01 -2.13840038e-01
4.53298748e-01 -2.27933571e-01 -5.06702781e-01 1.85473710e-01
-1.45590985e+00 -9.61858749e-01 -9.22038257e-02 2.59914255e+00
1.01020348e+00 5.12638211e-01 5.07897511e-02 1.59980208e-01
4.04919654e-01 2.42434353e-01 -6.96777344e-01 -7.58249938e-01
-1.82434291e-01 1.28637552e-01 1.20963991e+00 4.82473671e-01
-1.26792872e+00 6.96711957e-01 6.32384968e+00 5.40521324e-01
-9.64582622e-01 7.98123851e-02 1.06809139e+00 2.58070603e-03
-5.91133952e-01 2.33908117e-01 -6.37848318e-01 4.14916754e-01
9.51443493e-01 -2.45037451e-01 4.04384673e-01 9.12833989e-01
-1.80329382e-01 1.17916331e-01 -1.23862517e+00 4.86308306e-01
1.00082774e-02 -1.48716950e+00 -1.05525434e-01 6.44965053e-01
4.66638505e-01 -3.73774141e-01 3.42352629e-01 1.58939838e-01
5.86118937e-01 -1.03890252e+00 4.60086942e-01 2.01301754e-01
6.16106987e-01 -1.16840756e+00 3.73165429e-01 4.58414465e-01
-5.53390503e-01 -1.58797488e-01 -3.87553692e-01 1.01245016e-01
-4.01708186e-01 4.23210740e-01 -8.91224504e-01 1.30674616e-01
3.98655295e-01 -6.29681349e-02 -7.58063078e-01 7.01569915e-01
-2.16227740e-01 8.53041828e-01 -3.47815454e-01 2.93692321e-01
3.17981631e-01 8.08097869e-02 5.97640872e-01 1.09845138e+00
-4.99663018e-02 6.82525635e-02 -1.69983611e-01 1.07852364e+00
-4.98039186e-01 -3.58328491e-01 -8.51390481e-01 -2.35746294e-01
7.56545067e-01 8.47173452e-01 -2.64805764e-01 1.16737664e-01
-3.74623895e-01 1.15536880e+00 2.33528629e-01 3.44671369e-01
-6.21278167e-01 -2.39720166e-01 1.24031520e+00 1.75894216e-01
-1.36587694e-01 -1.10260412e-01 -6.15704358e-01 -1.04443932e+00
2.33544588e-01 -1.19500327e+00 5.19677222e-01 -2.08719820e-01
-1.49368310e+00 7.65989423e-02 -3.89452338e-01 -8.78461003e-01
1.47662818e-01 -3.91116053e-01 -5.60615361e-01 8.84247124e-01
-1.17128694e+00 -8.63964975e-01 3.24470580e-01 6.60152674e-01
-1.91346496e-01 6.60121441e-05 1.02642131e+00 1.81329191e-01
-3.67565334e-01 1.53041077e+00 7.63977915e-02 4.82608855e-01
6.05116308e-01 -1.14596057e+00 5.40477455e-01 1.16470468e+00
2.44876340e-01 1.14974475e+00 7.11654067e-01 -6.18053794e-01
-1.58003700e+00 -1.17820084e+00 5.49973369e-01 -5.72446167e-01
4.39663589e-01 -6.77443504e-01 -1.42840385e+00 9.27517056e-01
-1.94213912e-01 2.00829104e-01 1.03335285e+00 -5.01479860e-03
-9.49260533e-01 7.87554160e-02 -1.86909914e+00 7.29608536e-01
7.36644030e-01 -9.71963167e-01 -6.41174838e-02 1.72265321e-01
9.01691794e-01 -1.48192048e-01 -8.15615833e-01 3.11751306e-01
5.41190326e-01 -1.01632714e+00 1.13214028e+00 -8.52881670e-01
1.20407589e-01 -3.13907638e-02 -3.36280078e-01 -7.85950840e-01
3.37631814e-02 -9.75767076e-01 -2.35508353e-01 1.42631483e+00
4.68575060e-01 -9.12905633e-01 1.31810570e+00 1.58798909e+00
4.67867374e-01 -5.77898026e-01 -9.65041220e-01 -8.38883519e-01
5.66735744e-01 -5.80524206e-01 7.75875568e-01 1.38512063e+00
-1.12713456e-01 -3.29394698e-01 -5.27075648e-01 6.72568202e-01
1.01750803e+00 -4.50414062e-01 8.27312589e-01 -9.41357970e-01
-3.80840212e-01 -3.28160942e-01 -4.94155049e-01 -4.97182757e-01
3.13315600e-01 -9.01884139e-01 -3.61251622e-01 -6.88120127e-01
-6.74149767e-03 -6.53187335e-01 -3.66653711e-01 6.81606650e-01
-1.78525954e-01 1.23026296e-01 2.55200326e-01 1.55282021e-01
-9.08172503e-03 -1.20105827e-02 7.83658504e-01 -5.64132743e-02
4.29187715e-02 3.21090817e-01 -1.26756299e+00 6.42525434e-01
1.00834632e+00 -7.70002961e-01 -4.73233312e-01 -1.46162659e-01
2.06123531e-01 -1.68906435e-01 7.76558280e-01 -8.91413808e-01
3.00777495e-01 -9.18206871e-02 2.66607195e-01 -1.32130325e-01
3.47731486e-02 -1.03712463e+00 2.96082407e-01 5.00866413e-01
-7.62665987e-01 -3.82448465e-01 3.62553149e-01 8.49579275e-01
1.76711798e-01 -2.73893476e-01 9.74957824e-01 -7.78627247e-02
-2.26646155e-01 3.68873954e-01 -3.39277774e-01 7.18181878e-02
1.05531752e+00 -1.91633552e-01 -6.40355408e-01 -2.75861114e-01
-5.08076489e-01 5.06443046e-02 8.52859914e-01 4.97536957e-02
4.55857515e-01 -9.91862893e-01 -4.47830677e-01 4.03538615e-01
7.97638968e-02 -2.76420176e-01 -4.88368198e-02 3.02966595e-01
-2.82743394e-01 -6.10179901e-02 1.52021796e-01 -6.14032187e-02
-1.59345889e+00 8.55927885e-01 4.62836981e-01 6.44207969e-02
-6.86833858e-01 9.78031993e-01 1.85244337e-01 -2.94729650e-01
6.90088511e-01 -8.68951008e-02 5.05253077e-01 -2.39320457e-01
6.48153186e-01 2.26480335e-01 9.01882946e-02 -3.22953105e-01
-3.61643255e-01 -1.29300550e-01 -3.13098848e-01 -6.24654591e-02
9.85598922e-01 -5.89154847e-02 -5.71234450e-02 -1.46994844e-01
1.78606915e+00 2.94625849e-01 -1.36230588e+00 -2.32319266e-01
1.63724869e-01 -8.94895792e-01 -9.41062272e-02 -8.18722486e-01
-1.03755164e+00 9.10189986e-01 5.83397150e-01 5.38310885e-01
1.09341836e+00 -3.43279570e-01 9.13176239e-01 5.16213298e-01
3.79428715e-01 -7.79364705e-01 -3.28534514e-01 -1.41307544e-02
4.98767734e-01 -1.23380268e+00 -4.84252051e-02 -3.77269387e-01
-5.86567283e-01 7.61075497e-01 4.34993654e-01 4.19979393e-02
6.93189323e-01 6.13069177e-01 1.45568579e-01 2.85648778e-02
-5.72319150e-01 4.86233652e-01 -1.45863205e-01 5.90340972e-01
-6.95215911e-02 2.13625640e-01 -6.83254153e-02 7.21928060e-01
-2.74207950e-01 -1.86614662e-01 7.16204047e-01 1.09837222e+00
-1.35241359e-01 -1.17085314e+00 -4.36542749e-01 3.71060550e-01
-1.12442183e+00 7.94943795e-02 -6.55715883e-01 5.97668409e-01
1.09313637e-01 9.75206494e-01 -1.71644092e-01 -5.07802010e-01
1.21959493e-01 -5.54702953e-02 1.92802846e-01 -3.14059913e-01
-7.80779362e-01 -3.39678615e-01 2.22522467e-01 -5.27639329e-01
-8.37778300e-02 -5.69329143e-01 -9.40633178e-01 -8.33431363e-01
-3.00161541e-01 7.08346143e-02 5.10564506e-01 7.50475883e-01
4.73049998e-01 -2.32828140e-01 9.60938632e-01 -8.00233334e-02
-1.02021015e+00 -2.74755299e-01 -5.37198603e-01 5.73072553e-01
6.74148321e-01 -1.18357450e-01 -8.07149827e-01 -1.13297701e-01] | [5.953433036804199, 7.099343776702881] |
5f1ed51f-8a5c-4abe-a7ac-f6afd3d31dcf | investigating-certain-choices-of-cnn | 2212.01235 | null | https://arxiv.org/abs/2212.01235v1 | https://arxiv.org/pdf/2212.01235v1.pdf | Investigating certain choices of CNN configurations for brain lesion segmentation | Brain tumor imaging has been part of the clinical routine for many years to perform non-invasive detection and grading of tumors. Tumor segmentation is a crucial step for managing primary brain tumors because it allows a volumetric analysis to have a longitudinal follow-up of tumor growth or shrinkage to monitor disease progression and therapy response. In addition, it facilitates further quantitative analysis such as radiomics. Deep learning models, in particular CNNs, have been a methodology of choice in many applications of medical image analysis including brain tumor segmentation. In this study, we investigated the main design aspects of CNN models for the specific task of MRI-based brain tumor segmentation. Two commonly used CNN architectures (i.e. DeepMedic and U-Net) were used to evaluate the impact of the essential parameters such as learning rate, batch size, loss function, and optimizer. The performance of CNN models using different configurations was assessed with the BraTS 2018 dataset to determine the most performant model. Then, the generalization ability of the model was assessed using our in-house dataset. For all experiments, U-Net achieved a higher DSC compared to the DeepMedic. However, the difference was only statistically significant for whole tumor segmentation using FLAIR sequence data and tumor core segmentation using T1w sequence data. Adam and SGD both with the initial learning rate set to 0.001 provided the highest segmentation DSC when training the CNN model using U-Net and DeepMedic architectures, respectively. No significant difference was observed when using different normalization approaches. In terms of loss functions, a weighted combination of soft Dice and cross-entropy loss with the weighting term set to 0.5 resulted in an improved segmentation performance and training stability for both DeepMedic and U-Net models. | ['Michel Koole', 'Karolien Goffin', 'Stefan Sunaert', 'Henri Vandermeulen', 'Ahmed Radwan', 'Masoomeh Rahimpour'] | 2022-12-02 | null | null | null | null | ['tumor-segmentation', 'brain-tumor-segmentation'] | ['computer-vision', 'medical'] | [-5.40134273e-02 1.32962540e-02 -7.80700073e-02 -3.07467431e-01
-4.42248106e-01 -2.20303223e-01 4.49265003e-01 4.04840380e-01
-9.71668661e-01 5.97393215e-01 -1.02658391e-01 -4.21935767e-01
-6.40826076e-02 -7.21840262e-01 -2.33123392e-01 -9.75233376e-01
-9.39996019e-02 6.06437325e-01 2.78653294e-01 -6.50590211e-02
2.39733905e-02 7.59600163e-01 -9.68217850e-01 -5.91521263e-02
1.03350008e+00 1.21290040e+00 4.55275923e-01 3.57860744e-01
-1.64660498e-01 4.89732087e-01 -4.14187849e-01 -1.83342904e-01
1.63446426e-01 -1.88144863e-01 -8.47797036e-01 -8.63786861e-02
1.61251158e-01 -1.77937806e-01 -1.33244842e-01 1.10235894e+00
6.91680372e-01 7.47186020e-02 5.66841722e-01 -8.76920700e-01
4.95123975e-02 6.08749866e-01 -2.12633327e-01 4.30281162e-01
-2.49479592e-01 3.77596229e-01 3.57368976e-01 -4.13151264e-01
5.72161853e-01 5.16562641e-01 5.67423105e-01 5.90746522e-01
-1.10117829e+00 -6.51194155e-01 -2.76145279e-01 2.13844433e-01
-1.16085887e+00 -4.57038134e-02 3.06114525e-01 -7.06764817e-01
8.24380875e-01 2.63542831e-01 8.38038743e-01 7.39614725e-01
7.29185283e-01 3.76260340e-01 1.25305152e+00 -2.11543337e-01
3.97298813e-01 1.60387624e-02 2.36902073e-01 6.43690109e-01
1.94330528e-01 3.78292687e-02 1.23212077e-01 6.41120076e-02
5.49594223e-01 -5.05246446e-02 -5.45867980e-01 -1.82011351e-01
-1.06279922e+00 9.66005504e-01 7.51896560e-01 7.94770896e-01
-4.00891960e-01 1.88713819e-01 6.30881786e-01 3.50178182e-02
5.89761317e-01 5.58894694e-01 -3.14925075e-01 -1.13080949e-01
-1.11069322e+00 -7.30431750e-02 5.06691635e-01 2.53000766e-01
5.36373407e-02 5.22190519e-02 -3.79106373e-01 8.65848660e-01
9.43543911e-02 1.05994470e-01 1.19913995e+00 -4.52959865e-01
2.71096200e-01 4.93677408e-01 -4.30478513e-01 -5.22428215e-01
-1.06350350e+00 -8.41981709e-01 -9.95172381e-01 2.89136350e-01
6.17271602e-01 -7.77368993e-02 -1.40587473e+00 1.54534602e+00
1.22386083e-01 -1.37824759e-01 -3.22516598e-02 9.82207835e-01
9.48157847e-01 4.72695380e-02 1.85186207e-01 -1.16618179e-01
1.37077236e+00 -8.74640644e-01 -6.58315420e-01 9.09999013e-02
9.46952105e-01 -5.98219991e-01 8.57394695e-01 1.03697978e-01
-1.05203259e+00 -1.04848802e-01 -1.02410233e+00 2.42414922e-01
-4.43726152e-01 9.16487575e-02 5.09955406e-01 9.11922991e-01
-1.13615692e+00 7.95436203e-01 -1.33501112e+00 -4.25430506e-01
6.03792191e-01 7.10302353e-01 -4.63399112e-01 1.47862479e-01
-1.18077326e+00 1.35666800e+00 6.23744488e-01 3.95397022e-02
-9.55854416e-01 -9.85669792e-01 -5.66942096e-01 1.04186155e-01
2.75928527e-02 -4.50854927e-01 9.78551328e-01 -8.19499254e-01
-1.62494123e+00 9.91568625e-01 1.35875329e-01 -9.97859478e-01
8.48168373e-01 3.09938520e-01 -2.25333884e-01 8.82564932e-02
-1.38891190e-01 7.98786879e-01 4.58833843e-01 -7.15826392e-01
-2.51978040e-01 -6.40373170e-01 -3.74007314e-01 1.67155698e-01
-9.19245258e-02 -8.42664540e-02 -2.57790357e-01 -3.86835754e-01
1.98533475e-01 -1.04024673e+00 -2.72007465e-01 -2.08684191e-01
-2.24455222e-01 2.63610959e-01 7.72830606e-01 -9.11544502e-01
8.33002329e-01 -1.88084185e+00 8.80887657e-02 4.31116849e-01
4.78627980e-01 2.66126215e-01 2.20151782e-01 -3.97233427e-01
-4.19669598e-01 2.11753994e-01 -5.08937538e-01 -1.64042279e-01
-4.62512821e-01 -1.99497174e-02 5.62557518e-01 6.44487381e-01
-1.28760040e-01 9.86711264e-01 -5.43570220e-01 -3.45064074e-01
4.41170692e-01 6.46202922e-01 -3.80427837e-01 -2.61854324e-02
4.31035310e-02 7.34086692e-01 -6.88455254e-02 4.90746200e-01
5.19951701e-01 -6.25762939e-02 -3.87634076e-02 -1.80237263e-01
-9.36797559e-02 -2.78433919e-01 -4.75947320e-01 1.63966644e+00
-4.69341785e-01 7.49469161e-01 2.31822699e-01 -9.39767778e-01
8.75654995e-01 4.79724467e-01 9.07465219e-01 -9.34109569e-01
5.87295294e-01 4.55920935e-01 6.80336475e-01 -4.29656625e-01
1.52621344e-01 -3.43013614e-01 4.97543991e-01 -6.28555380e-03
3.46808098e-02 -2.51736194e-01 1.90159708e-01 -1.82200775e-01
1.08183050e+00 -5.04769981e-01 6.34808168e-02 -6.14249408e-01
6.58745766e-01 8.42158347e-02 2.47033000e-01 2.60210574e-01
-5.64979911e-01 7.81992376e-01 5.93530655e-01 -3.37546438e-01
-7.36409783e-01 -8.61090004e-01 -6.50146723e-01 4.13076341e-01
-1.60333633e-01 1.11722127e-02 -1.03907263e+00 -4.95528609e-01
-2.53768176e-01 7.76152074e-01 -9.00776625e-01 -1.76681936e-01
-4.13875788e-01 -1.25509667e+00 5.38214386e-01 3.27024132e-01
7.07625031e-01 -1.05591571e+00 -9.87372279e-01 2.66169518e-01
-1.53956264e-01 -1.04877520e+00 -2.11466685e-01 5.77726662e-01
-1.24416661e+00 -1.23340619e+00 -1.15656400e+00 -4.31575418e-01
6.71650052e-01 -4.07327563e-01 7.71490157e-01 1.42039075e-01
-4.72232282e-01 1.91522717e-01 -1.66678891e-01 -4.13613200e-01
-5.36582232e-01 5.10997057e-01 -2.10813642e-01 -2.27520704e-01
6.22440837e-02 -3.33117753e-01 -6.57660186e-01 2.61591047e-01
-1.00588322e+00 6.12636209e-02 5.37635386e-01 8.46584976e-01
6.35500193e-01 9.93102416e-03 2.60496642e-02 -6.32863998e-01
8.99062574e-01 -2.39067093e-01 -6.05394006e-01 1.08861886e-01
-7.20951021e-01 -5.28253317e-02 5.63059807e-01 -2.71023721e-01
-5.91251433e-01 -1.37297213e-01 -3.24187279e-01 -4.29228991e-01
-2.48446818e-02 7.50896811e-01 2.12691158e-01 -3.80736738e-01
5.89535594e-01 4.97576967e-02 5.68147659e-01 -5.99872470e-02
-3.26942176e-01 2.43204504e-01 2.02449232e-01 -2.87491739e-01
1.96680829e-01 4.85644519e-01 3.07801276e-01 -6.92552447e-01
-3.40401262e-01 -2.41547227e-01 -6.30968094e-01 -3.87604952e-01
1.27731991e+00 -3.34346384e-01 -5.79611361e-01 7.02999830e-01
-7.84357607e-01 -6.83303893e-01 -2.29411900e-01 8.32716048e-01
-3.92198086e-01 7.71757960e-02 -5.12035608e-01 -8.87043849e-02
-8.02019835e-01 -1.98373675e+00 4.15732235e-01 3.26517314e-01
-1.07203461e-01 -1.19493532e+00 -8.03845003e-02 2.02375948e-01
1.03396297e+00 5.73069572e-01 9.44266498e-01 -8.80859733e-01
1.10411607e-02 -3.02670658e-01 -1.08887278e-01 3.80267590e-01
1.37291282e-01 -3.71232815e-02 -8.61252546e-01 -3.54429007e-01
2.47641653e-02 -1.52783992e-04 8.96512508e-01 8.52853775e-01
1.43278515e+00 2.75026768e-01 -1.95935324e-01 8.13598454e-01
1.50694287e+00 5.39286971e-01 7.58745134e-01 5.71198225e-01
6.19747043e-01 3.37282926e-01 -3.99666093e-02 7.23876655e-02
-1.06820978e-01 6.18351698e-01 8.11952412e-01 -2.00264707e-01
-1.00468509e-01 3.93745929e-01 -9.37072858e-02 4.52423185e-01
-2.09534094e-01 8.58707428e-02 -1.35354877e+00 2.86784083e-01
-1.27406037e+00 -7.08198130e-01 -1.94810227e-01 2.36244392e+00
5.21838367e-01 3.06696385e-01 -6.10177368e-02 1.35675475e-01
5.85026801e-01 -1.42042473e-01 -5.08818269e-01 -4.20297295e-01
1.08127389e-02 3.37075204e-01 9.12865639e-01 4.07855749e-01
-9.43030477e-01 5.92507362e-01 5.74087572e+00 7.84458101e-01
-1.97293103e+00 3.74961495e-01 1.03855240e+00 -3.13078105e-01
1.76708594e-01 -3.06695461e-01 -4.12838697e-01 6.55118287e-01
1.02498865e+00 1.52254760e-01 2.25150317e-01 6.38287842e-01
4.01091158e-01 -4.75332171e-01 -7.36404121e-01 8.69351268e-01
-1.94014326e-01 -1.32244098e+00 -4.00389254e-01 3.12690943e-01
4.85070616e-01 5.89620769e-01 -1.48831224e-02 2.06498861e-01
-2.43273795e-01 -1.36063933e+00 5.51880836e-01 3.44194829e-01
7.49181986e-01 -6.35167241e-01 1.20067143e+00 9.49336812e-02
-7.43866503e-01 1.06826507e-01 -1.62858665e-02 4.31669593e-01
8.01418200e-02 6.69501245e-01 -9.73060548e-01 3.95337909e-01
5.87375581e-01 2.63172269e-01 -6.19090676e-01 1.22084844e+00
1.91263184e-01 6.85473919e-01 -4.66524839e-01 1.36358604e-01
4.52376813e-01 -2.84414083e-01 3.98324162e-01 1.10073411e+00
3.18669617e-01 -2.85808034e-02 -7.32412189e-02 8.29466164e-01
3.54958177e-02 2.35730365e-01 9.00144223e-03 1.48104861e-01
2.38235742e-02 1.44776821e+00 -1.28682530e+00 -1.19969390e-01
-9.11642388e-02 5.97336709e-01 1.33452043e-01 9.23448130e-02
-9.93755162e-01 -1.37132913e-01 4.16243702e-01 5.10832250e-01
-7.30227772e-03 -9.56988707e-02 -5.73776960e-01 -9.32119906e-01
-2.40882859e-01 -5.14428258e-01 3.03101391e-01 -4.86389846e-01
-6.66797161e-01 7.82561839e-01 9.00308415e-02 -7.98263490e-01
-4.06256272e-03 -7.52486646e-01 -1.02717435e+00 9.75023925e-01
-1.41059816e+00 -5.71081996e-01 -7.22341776e-01 5.71762979e-01
1.67761028e-01 -8.49614516e-02 7.10986078e-01 2.62291610e-01
-7.72827387e-01 4.87916917e-01 5.99314608e-02 2.35022441e-01
3.78752172e-01 -1.20075750e+00 -2.84924299e-01 6.10171616e-01
-5.40160179e-01 3.00348639e-01 6.06391907e-01 -4.31224376e-01
-6.96902692e-01 -1.02241027e+00 2.48871669e-01 1.90588966e-01
5.11409879e-01 7.41388649e-02 -8.87639761e-01 4.13702995e-01
2.04017445e-01 9.00062472e-02 6.38779521e-01 -4.48108435e-01
3.08288753e-01 -2.43185814e-02 -1.55776215e+00 3.78427029e-01
2.78327882e-01 -4.64193635e-02 2.31816843e-02 2.68197954e-01
5.07667303e-01 -8.24233890e-01 -1.36028910e+00 6.96567833e-01
4.67653334e-01 -9.74843740e-01 7.12264001e-01 -2.71295726e-01
4.67501521e-01 8.32357630e-02 1.44592509e-01 -1.55671751e+00
-2.41684407e-01 8.56113061e-02 2.47605830e-01 6.75087929e-01
6.91407382e-01 -9.25803006e-01 9.41137016e-01 9.66946959e-01
-3.77043605e-01 -1.35671926e+00 -9.64594305e-01 -5.88199675e-01
4.47443306e-01 -4.24708754e-01 4.42518950e-01 6.88330948e-01
-3.07282060e-01 -3.39388698e-01 4.34293061e-01 -2.97732681e-01
3.95166695e-01 -5.25712490e-01 2.03116387e-01 -1.04819381e+00
1.74788982e-01 -1.12792313e+00 -6.44262135e-01 3.58242840e-02
9.95496511e-02 -1.22852933e+00 -2.35594213e-01 -1.58167231e+00
5.05691469e-02 -4.08060163e-01 -4.98992622e-01 4.02784228e-01
3.86996344e-02 4.52094376e-02 2.77006090e-01 2.37931460e-01
2.48633415e-01 3.97097439e-01 1.43846846e+00 -3.52551103e-01
-3.60808045e-01 1.36607125e-01 -2.01348141e-01 5.25018036e-01
1.23513317e+00 -3.56960386e-01 -2.77381957e-01 -3.01194668e-01
-4.65520978e-01 4.67085354e-02 3.52939427e-01 -1.27878869e+00
2.10437268e-01 9.61065143e-02 5.54747641e-01 -4.54815388e-01
2.39767760e-01 -9.04097617e-01 1.59087315e-01 9.62152600e-01
-2.48803601e-01 1.14140145e-01 4.21186507e-01 -1.68914199e-01
-2.79515177e-01 -4.07481790e-01 1.28344917e+00 -1.24953516e-01
-4.33030248e-01 5.23682475e-01 -4.96642858e-01 -9.01018754e-02
1.22333825e+00 -4.04410809e-01 -1.52761517e-02 -1.78582951e-01
-1.04972649e+00 1.24876983e-01 3.15711409e-01 1.48507208e-01
4.95520264e-01 -9.83783364e-01 -6.53584957e-01 1.47802755e-02
-9.58170295e-02 1.32653698e-01 3.32158834e-01 1.63014650e+00
-9.62140739e-01 5.59872389e-01 -5.22026539e-01 -7.78446317e-01
-1.29002547e+00 1.47884533e-01 1.09604967e+00 -7.52270997e-01
-5.29929042e-01 7.54126549e-01 -8.77189189e-02 -3.90847147e-01
3.34363788e-01 -7.66388178e-01 -4.00723040e-01 -1.97573900e-02
3.11433554e-01 3.06502253e-01 6.50383949e-01 -5.37529826e-01
-4.97219920e-01 2.41148189e-01 -1.91868141e-01 -1.66534677e-01
1.29687119e+00 2.55996972e-01 -2.47202531e-01 1.39740612e-02
1.39546359e+00 -6.37125909e-01 -9.92032588e-01 1.67024300e-01
-1.13100179e-01 2.27603856e-02 6.75509334e-01 -9.84888375e-01
-1.84386432e+00 8.12035084e-01 1.20591247e+00 -9.26534981e-02
1.12282526e+00 -2.39803836e-01 5.31978190e-01 -1.01890191e-01
1.86230361e-01 -9.28289771e-01 -2.96007693e-01 4.94939148e-01
7.82668591e-01 -1.34189963e+00 -8.51529613e-02 -1.48215607e-01
-6.08860433e-01 1.30860853e+00 6.83726370e-01 9.08715501e-02
8.44192863e-01 3.49800289e-01 9.01657343e-02 -4.14328307e-01
-1.95782468e-01 -1.29127368e-01 3.03839594e-01 2.20494494e-01
7.06539512e-01 3.27190042e-01 -5.32337546e-01 3.82170260e-01
-3.70702952e-01 1.78988114e-01 4.33382481e-01 7.46914029e-01
-2.43513420e-01 -7.95045316e-01 -3.41899335e-01 8.40350747e-01
-6.56937659e-01 2.06497591e-02 -9.93648767e-02 1.02270985e+00
5.54666743e-02 5.32650828e-01 1.13600925e-01 -1.73354372e-01
2.19722524e-01 6.75622597e-02 5.07847548e-01 -2.82130182e-01
-1.06622088e+00 -2.56665289e-01 -3.08564544e-01 -4.38848287e-01
-2.93467939e-01 -5.40139616e-01 -1.41255260e+00 -2.91825622e-01
-3.95641059e-01 9.70683694e-02 1.24997735e+00 1.00496745e+00
-1.04701638e-01 8.92644465e-01 2.09781662e-01 -7.08251774e-01
-3.57136786e-01 -1.20274150e+00 -4.71434176e-01 1.96457550e-01
-1.37712717e-01 -7.41405904e-01 -2.15528026e-01 -4.10487443e-01] | [14.610426902770996, -2.5136966705322266] |
725f3318-5342-4b03-bde6-93075f074227 | data-efficient-contrastive-self-supervised | 2302.09195 | null | https://arxiv.org/abs/2302.09195v4 | https://arxiv.org/pdf/2302.09195v4.pdf | Data-Efficient Contrastive Self-supervised Learning: Most Beneficial Examples for Supervised Learning Contribute the Least | Self-supervised learning (SSL) learns high-quality representations from large pools of unlabeled training data. As datasets grow larger, it becomes crucial to identify the examples that contribute the most to learning such representations. This enables efficient SSL by reducing the volume of data required. Nevertheless, quantifying the value of examples for SSL has remained an open question. In this work, we address this problem for the first time, by proving that examples that contribute the most to contrastive SSL are those that have the most similar augmentations to other examples, in expectation. We provide rigorous guarantees for the generalization performance of contrastive learning on such subsets. Through extensive experiments, we show that we can safely exclude 20% of examples from CIFAR100 and 40% from STL10 and TinyImageNet, without affecting downstream task performance. In general, subsets selected by our method outperform random subsets by over 3% across these datasets. Interestingly, we also discover the subsets that contribute the most to contrastive learning are those that contribute the least to supervised learning. | ['Baharan Mirzasoleiman', 'Siddharth Joshi'] | 2023-02-18 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 1.42587647e-01 2.99974293e-01 -2.53807873e-01 -4.17529911e-01
-9.17315006e-01 -7.39535034e-01 5.76456606e-01 2.94823796e-01
-6.43906116e-01 1.03602219e+00 1.55383479e-02 -6.09347299e-02
-1.40764996e-01 -5.40554643e-01 -1.11123776e+00 -5.97839952e-01
-2.18369678e-01 4.39114243e-01 2.28423655e-01 -8.59173760e-02
-1.04286067e-01 3.75228077e-01 -1.54901099e+00 2.73382932e-01
8.42121482e-01 1.00500190e+00 6.02093898e-02 2.22002640e-01
-6.41009137e-02 9.48900461e-01 -5.37714601e-01 -2.21070841e-01
4.27672267e-01 -5.82733035e-01 -8.62331867e-01 -3.69301662e-02
7.96839595e-01 -7.00935796e-02 4.44432311e-02 8.60102952e-01
3.37521583e-01 2.53962636e-01 7.92445362e-01 -1.34909773e+00
-3.61215264e-01 9.24207449e-01 -4.15852994e-01 4.61571306e-01
-2.93893695e-01 1.43560484e-01 1.10934114e+00 -1.08504856e+00
7.43744254e-01 8.74035835e-01 6.51824534e-01 8.08708072e-01
-1.38852847e+00 -8.00652325e-01 3.25423688e-01 -1.88621894e-01
-1.23539710e+00 -5.72156072e-01 6.63499177e-01 -2.50050426e-01
7.05891609e-01 2.32216999e-01 3.33120674e-01 8.59853506e-01
-7.11513281e-01 1.05564976e+00 1.14631581e+00 -5.23303509e-01
3.60956132e-01 4.15209651e-01 5.93540311e-01 5.54605901e-01
4.78536755e-01 -1.22485697e-01 -5.34135878e-01 -9.76936072e-02
5.44127584e-01 4.58109751e-02 -3.78575563e-01 -4.04406160e-01
-1.07864594e+00 8.25266957e-01 5.80390096e-01 2.77817190e-01
-1.39560819e-01 8.61283392e-02 1.84182510e-01 3.69775027e-01
6.05446160e-01 9.58335578e-01 -8.91954601e-01 1.81882098e-01
-8.62409294e-01 1.65670648e-01 5.83761811e-01 9.55191970e-01
1.02230811e+00 1.91516299e-02 8.78241360e-02 9.47544813e-01
-1.79556921e-01 3.50769043e-01 6.68744206e-01 -9.39354420e-01
5.60700059e-01 7.76049495e-01 -1.35538243e-02 -2.48447731e-01
-3.97558242e-01 -8.90970528e-01 -5.43111384e-01 -2.88165007e-02
6.93987429e-01 -2.60815740e-01 -7.17578232e-01 2.04766536e+00
3.23227607e-02 1.74586266e-01 1.43678069e-01 6.28941834e-01
7.46918142e-01 3.62569988e-01 2.19451368e-01 -2.33562529e-01
1.01889265e+00 -7.85918355e-01 -2.91874319e-01 -4.36148375e-01
1.11500895e+00 -3.76786083e-01 1.34468853e+00 1.17727660e-01
-8.81133676e-01 -4.57903624e-01 -1.12864745e+00 1.77126795e-01
-2.38964871e-01 1.59087420e-01 6.98844194e-01 5.51171124e-01
-7.35243917e-01 8.90782773e-01 -4.92482513e-01 -7.58230537e-02
8.47408533e-01 3.94954383e-01 -3.86265814e-01 -1.05011910e-01
-1.08944809e+00 7.03903079e-01 4.34051901e-01 -3.03772897e-01
-7.80677438e-01 -1.04344535e+00 -7.23453999e-01 1.96280822e-01
3.12447876e-01 -1.75189018e-01 1.24024248e+00 -1.34991658e+00
-6.59933686e-01 9.96120930e-01 -2.17272267e-01 -9.40543413e-01
4.67859447e-01 -3.33854377e-01 -7.30815902e-02 2.16366231e-01
2.42079228e-01 9.13106978e-01 5.06506741e-01 -1.46479571e+00
-6.16148651e-01 -3.35505515e-01 3.16788368e-02 2.14646891e-01
-7.18900859e-01 -2.33001515e-01 -3.75297368e-01 -5.24686277e-01
1.46728620e-01 -1.06025910e+00 -4.35745537e-01 -2.13029504e-01
-2.56046861e-01 -3.78724217e-01 5.57749867e-01 -1.30276188e-01
7.83638716e-01 -2.31502557e+00 -2.93552577e-01 1.29256800e-01
5.34350514e-01 3.97480428e-01 -2.56273955e-01 2.12417450e-03
-1.53701201e-01 3.37064207e-01 -2.46402413e-01 -5.04685879e-01
-1.81678489e-01 1.63959250e-01 -5.10189116e-01 3.30024153e-01
5.81949294e-01 1.04705441e+00 -1.01797879e+00 -5.83343863e-01
-3.35823670e-02 3.92933339e-02 -5.03528655e-01 -1.90540887e-02
-3.58806670e-01 4.05502826e-01 -3.32483977e-01 2.55890518e-01
4.56010938e-01 -6.95902944e-01 6.59961402e-02 -1.01103164e-01
2.80272365e-01 4.75436807e-01 -9.93677199e-01 1.18048716e+00
-3.79400581e-01 6.62097692e-01 -2.56450355e-01 -1.15992916e+00
8.66289794e-01 1.63294882e-01 4.66718912e-01 -6.13287032e-01
3.34469019e-03 3.59453529e-01 5.48918024e-02 -1.44432351e-01
2.59822696e-01 -2.73700088e-01 6.25836849e-02 6.74437881e-01
3.44397426e-01 2.82492876e-01 3.33927214e-01 4.83000010e-01
9.56692100e-01 -1.81792095e-01 2.18433410e-01 -2.66455680e-01
4.67922054e-02 3.63385566e-02 6.23450518e-01 9.42324460e-01
-2.96697646e-01 6.31730258e-01 5.06860912e-01 -3.03872526e-01
-9.39865828e-01 -1.10399675e+00 -1.20152041e-01 9.94585335e-01
-2.82143503e-02 -1.88730523e-01 -6.91160679e-01 -1.05588126e+00
-4.22205105e-02 5.91061354e-01 -7.00330794e-01 -2.47579917e-01
-4.64394689e-01 -8.63740742e-01 4.30770695e-01 6.62190199e-01
4.07430887e-01 -1.13357842e+00 -2.89824694e-01 -4.49177227e-04
-3.59803326e-02 -1.30389977e+00 -2.25220561e-01 6.46706700e-01
-1.33870256e+00 -1.26437497e+00 -7.10924804e-01 -9.26750839e-01
1.07538867e+00 3.82065386e-01 1.32697356e+00 2.81468093e-01
-1.27554476e-01 7.29660615e-02 -4.97772902e-01 -4.27913874e-01
-3.48848879e-01 3.76568764e-01 -1.19102255e-01 -1.12460054e-01
4.86378640e-01 -5.06309450e-01 -3.40744823e-01 2.50953317e-01
-5.78330636e-01 -1.13088667e-01 5.51451802e-01 7.89895236e-01
7.63857722e-01 5.68939336e-02 9.41776752e-01 -1.48341238e+00
3.31068426e-01 -7.05669880e-01 -4.50214237e-01 6.68280423e-02
-5.95956147e-01 3.14709693e-01 9.84556675e-01 -6.23566091e-01
-6.32467747e-01 2.95013785e-01 1.79638803e-01 -5.82422793e-01
-1.98473409e-01 3.04115862e-01 -9.28638652e-02 1.56748757e-01
1.12647033e+00 2.59584367e-01 -4.24776897e-02 -6.05315089e-01
4.72868085e-01 3.32368672e-01 2.40104362e-01 -5.66878796e-01
7.71521807e-01 4.94267076e-01 -9.62540284e-02 -9.02039886e-01
-1.35881114e+00 -4.24358577e-01 -5.36399066e-01 -1.85764190e-02
2.91838348e-01 -9.72418010e-01 -3.34055126e-01 3.09393201e-02
-6.44240260e-01 -4.58465159e-01 -8.53671551e-01 5.56681633e-01
-4.77409542e-01 1.14457451e-01 -4.10134226e-01 -7.92238712e-01
-2.09775582e-01 -9.78568077e-01 7.02031553e-01 4.00775149e-02
-4.05154437e-01 -7.27672637e-01 -1.66439310e-01 3.24872941e-01
9.89100188e-02 3.23168859e-02 8.62550437e-01 -1.23035860e+00
-5.17083168e-01 -1.97130784e-01 -1.06970295e-01 6.11287653e-01
1.45029992e-01 -3.90589416e-01 -1.23547506e+00 -2.11461470e-01
-1.27376705e-01 -7.32257485e-01 1.37693846e+00 5.11500001e-01
1.28851724e+00 -1.49760440e-01 -3.21704865e-01 4.91175413e-01
1.33802879e+00 -1.17549084e-01 3.43807846e-01 1.64517894e-01
6.13169372e-01 7.60967493e-01 5.34852147e-01 2.84402054e-02
-2.99514849e-02 3.61696482e-01 2.35701576e-01 -1.84987381e-01
-2.70142466e-01 -2.94210792e-01 1.92646235e-01 4.70441133e-01
1.24761440e-01 4.48761657e-02 -7.98985600e-01 7.77252913e-01
-1.57793820e+00 -9.78977144e-01 -1.39862821e-01 2.35737991e+00
1.05708718e+00 5.47616303e-01 2.21081644e-01 4.26556557e-01
6.68420494e-01 -1.24609426e-01 -6.63841844e-01 1.21030606e-01
-3.78857464e-01 4.82344568e-01 6.72502458e-01 3.52576703e-01
-1.24341249e+00 1.03180265e+00 6.38971710e+00 9.09022510e-01
-8.21257949e-01 1.24587476e-01 1.06787586e+00 -4.23550814e-01
-3.37037057e-01 -1.37310205e-02 -1.05652368e+00 3.40169579e-01
1.01328886e+00 -2.32966170e-02 2.86173254e-01 1.15676820e+00
2.34228224e-02 8.18994865e-02 -1.26712036e+00 7.45116949e-01
-8.17311630e-02 -1.39390278e+00 -3.49381268e-02 -1.14416130e-01
1.11171675e+00 3.45834851e-01 5.55620342e-02 4.55239683e-01
5.09081304e-01 -9.82580245e-01 5.13209224e-01 -7.87159577e-02
6.11256123e-01 -7.22392440e-01 5.91641605e-01 4.51408803e-01
-7.78349817e-01 -9.47918817e-02 -4.76322502e-01 5.08562699e-02
-3.57132494e-01 9.34300423e-01 -8.57419729e-01 -4.68042828e-02
5.03650665e-01 7.16170132e-01 -8.22312415e-01 9.41076338e-01
-2.51337826e-01 1.25517106e+00 -3.51241708e-01 -1.82049066e-01
2.16701955e-01 2.11851045e-01 1.15716159e-01 1.01953018e+00
-1.03778765e-01 -8.69751126e-02 9.77871493e-02 8.46838951e-01
-8.09674740e-01 2.07698315e-01 -7.26204693e-01 -1.74538612e-01
7.10070968e-01 1.03641415e+00 -8.00267339e-01 -5.29443383e-01
-3.80369663e-01 5.88044763e-01 7.04463482e-01 3.47324133e-01
-4.35504109e-01 -1.16902620e-01 4.38155562e-01 2.22136572e-01
2.17678383e-01 9.36153010e-02 -3.86958033e-01 -1.11391282e+00
1.33548886e-01 -8.31717312e-01 4.85957593e-01 -4.54849750e-01
-1.53599715e+00 6.37483001e-01 -2.13922501e-01 -1.31095076e+00
-2.23387823e-01 -5.37696004e-01 -3.46683145e-01 7.46461809e-01
-1.49994016e+00 -7.78293967e-01 -2.15577736e-01 4.08505827e-01
5.56526184e-01 -1.87441707e-01 7.34972537e-01 1.46429837e-01
-3.71333331e-01 6.77439153e-01 2.29373083e-01 3.63013208e-01
5.45203805e-01 -1.35948646e+00 4.29526776e-01 7.33346045e-01
5.91553867e-01 7.66998649e-01 5.04749715e-01 -4.34521705e-01
-7.93713570e-01 -1.23470926e+00 9.20378923e-01 -5.65193772e-01
4.77295816e-01 -2.44916305e-01 -9.57534492e-01 8.28476369e-01
-3.22673231e-01 4.79299933e-01 8.26704919e-01 4.67435181e-01
-7.04004407e-01 -1.36849612e-01 -1.03857863e+00 5.99488318e-01
1.20078158e+00 -4.69480127e-01 -5.94485283e-01 4.17323440e-01
7.84269273e-01 8.27276185e-02 -6.21596932e-01 2.60050714e-01
1.33643001e-01 -8.14610243e-01 9.35412705e-01 -9.37892437e-01
7.04660714e-01 -6.91875443e-02 -1.03927709e-01 -1.37042224e+00
6.83012092e-03 -3.13193053e-01 -1.54273048e-01 1.23752058e+00
8.29324067e-01 -5.48930287e-01 1.17502820e+00 4.75236893e-01
-5.74396960e-02 -8.48297238e-01 -6.51373982e-01 -1.27736795e+00
4.70022023e-01 -4.29267257e-01 3.82231593e-01 1.09760892e+00
-7.29696080e-02 6.28626704e-01 -2.54068762e-01 -1.41500458e-01
8.96057010e-01 1.39012024e-01 6.24751747e-01 -1.38979661e+00
-3.19166809e-01 -4.15411025e-01 -2.93875247e-01 -1.01479959e+00
2.75509357e-01 -1.23140121e+00 -5.87756112e-02 -1.25833070e+00
4.42522466e-01 -8.45521748e-01 -5.77512443e-01 7.11100101e-01
-5.36859393e-01 5.10162473e-01 2.38074586e-01 4.03200388e-01
-7.53157258e-01 3.06982219e-01 1.03479826e+00 3.27538922e-02
-1.37648150e-01 1.51563366e-03 -1.05183399e+00 6.89512789e-01
1.06143177e+00 -6.12474859e-01 -6.98776603e-01 -2.97482193e-01
1.21153615e-01 -4.96198535e-01 2.34631255e-01 -9.04840589e-01
-1.54587984e-01 9.43701938e-02 5.85370064e-01 -3.06239337e-01
2.19959944e-01 -6.73594713e-01 -4.43011940e-01 4.70962465e-01
-9.58370209e-01 -4.59151924e-01 1.63823664e-01 3.89316857e-01
-1.50914475e-01 -5.50646603e-01 1.02391100e+00 -1.98300779e-01
-5.60997248e-01 3.37968886e-01 -1.77477244e-02 6.91707730e-01
8.51111054e-01 7.06494376e-02 -3.31129879e-01 -3.05276692e-01
-7.87829697e-01 2.01111391e-01 3.27548653e-01 1.46837696e-01
3.37970912e-01 -1.18021560e+00 -7.33163953e-01 1.54976353e-01
2.88598984e-01 5.29003106e-02 -4.79120761e-02 5.92070401e-01
-4.90801372e-02 3.12291861e-01 9.35682058e-02 -3.51143986e-01
-1.16503131e+00 5.25021732e-01 3.42406809e-01 -2.65899628e-01
-5.73214054e-01 1.05971646e+00 2.69700795e-01 -3.14636290e-01
2.52222419e-01 -6.59653544e-02 -1.96179926e-01 8.28126520e-02
6.89669311e-01 2.10245535e-01 1.84040785e-01 -4.46186930e-01
-2.78058976e-01 1.44981027e-01 -3.89576435e-01 -5.42969629e-02
1.52316606e+00 3.16859335e-01 3.92238677e-01 4.89229739e-01
1.51958334e+00 -7.77644366e-02 -1.41969573e+00 -4.34734017e-01
1.54444858e-01 -2.74124444e-01 -2.77276542e-02 -7.51091123e-01
-1.20223129e+00 9.16232765e-01 4.43090796e-01 1.56325892e-01
9.16477442e-01 2.85380900e-01 5.40184438e-01 5.55829763e-01
3.57913613e-01 -9.98292506e-01 1.94684148e-01 3.01234692e-01
4.95225191e-01 -1.53305554e+00 -7.05171674e-02 -4.76583660e-01
-6.44956172e-01 6.49847507e-01 7.29075551e-01 -3.88224870e-01
5.07608175e-01 1.21555120e-01 -3.60260233e-02 -1.46353379e-01
-8.93118024e-01 -4.26892608e-01 1.58119708e-01 6.19052231e-01
4.60217416e-01 3.01931035e-02 -1.60472423e-01 7.45472610e-01
-1.64043307e-01 -2.51675192e-02 3.24770123e-01 8.59152496e-01
-6.58294916e-01 -9.48825121e-01 -1.03880577e-01 9.92245376e-01
-3.81730676e-01 -4.29263681e-01 -5.32155931e-01 8.49608541e-01
-5.00032380e-02 8.12715352e-01 1.37237132e-01 -1.61746711e-01
2.29124531e-01 3.66161913e-01 5.16045034e-01 -7.89670348e-01
-5.30160248e-01 -2.34543532e-01 1.11977257e-01 -1.89068735e-01
-2.95106143e-01 -5.26971877e-01 -1.49943066e+00 -4.73971628e-02
-4.21100408e-01 2.37252325e-01 3.79869968e-01 1.08256519e+00
2.01911673e-01 3.19871366e-01 7.86710620e-01 -4.77149814e-01
-8.00278068e-01 -9.87743437e-01 -6.11848533e-01 7.56175220e-01
2.21991524e-01 -5.51706970e-01 -8.66418064e-01 8.53670016e-02] | [9.47859001159668, 3.029827356338501] |
244bee68-da71-45d4-bcf4-9095542aa2d7 | model-based-constrained-mdp-for-budget | 2303.01049 | null | https://arxiv.org/abs/2303.01049v1 | https://arxiv.org/pdf/2303.01049v1.pdf | Model-based Constrained MDP for Budget Allocation in Sequential Incentive Marketing | Sequential incentive marketing is an important approach for online businesses to acquire customers, increase loyalty and boost sales. How to effectively allocate the incentives so as to maximize the return (e.g., business objectives) under the budget constraint, however, is less studied in the literature. This problem is technically challenging due to the facts that 1) the allocation strategy has to be learned using historically logged data, which is counterfactual in nature, and 2) both the optimality and feasibility (i.e., that cost cannot exceed budget) needs to be assessed before being deployed to online systems. In this paper, we formulate the problem as a constrained Markov decision process (CMDP). To solve the CMDP problem with logged counterfactual data, we propose an efficient learning algorithm which combines bisection search and model-based planning. First, the CMDP is converted into its dual using Lagrangian relaxation, which is proved to be monotonic with respect to the dual variable. Furthermore, we show that the dual problem can be solved by policy learning, with the optimal dual variable being found efficiently via bisection search (i.e., by taking advantage of the monotonicity). Lastly, we show that model-based planing can be used to effectively accelerate the joint optimization process without retraining the policy for every dual variable. Empirical results on synthetic and real marketing datasets confirm the effectiveness of our methods. | ['Shuang Yang', 'Jun Zhu', 'Yuanbo Chen', 'Lei Lv', 'Zaifan Jiang', 'Le Guo', 'Shuai Xiao'] | 2023-03-02 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [ 2.33830363e-01 3.43800068e-01 -9.34217632e-01 -1.28547028e-01
-5.59829831e-01 -4.58070129e-01 -4.99750338e-02 6.63092658e-02
-4.47276413e-01 1.01484823e+00 -1.11521212e-02 -6.44053519e-01
-5.69379568e-01 -8.69747639e-01 -9.50450599e-01 -8.75800610e-01
-1.55927718e-01 6.80861354e-01 -4.26544219e-01 3.75274345e-02
3.35083693e-01 1.91123083e-01 -1.22204673e+00 -1.76044613e-01
1.03478932e+00 1.03986228e+00 4.81146008e-01 3.00216168e-01
-1.11892350e-01 5.03453553e-01 -1.98254794e-01 -2.34200597e-01
5.69176316e-01 -2.24777058e-01 -8.06691945e-01 6.09131992e-01
-2.05694959e-01 -7.74380624e-01 1.16807655e-01 9.62045372e-01
2.34008893e-01 8.08196440e-02 1.12496205e-01 -1.51602769e+00
-8.74472558e-02 7.42031455e-01 -9.82676089e-01 -8.82001035e-03
1.13747366e-01 1.22628413e-01 1.41188979e+00 -1.43631428e-01
5.13902128e-01 1.21618259e+00 -7.33460411e-02 3.44046116e-01
-1.55587780e+00 -6.15161657e-01 6.88204348e-01 1.58359528e-01
-7.51205146e-01 7.55588636e-02 7.99589574e-01 -2.51160622e-01
5.35060883e-01 1.02662742e-01 9.31040704e-01 6.80721760e-01
1.41774893e-01 1.07748771e+00 1.26684964e+00 -5.17347395e-01
4.65694815e-01 6.86705783e-02 -1.36716902e-01 3.18958133e-01
4.85709488e-01 2.95363426e-01 -1.66791767e-01 -1.25577718e-01
5.65292120e-01 1.46178022e-01 -1.00300632e-01 -5.94912112e-01
-7.79003620e-01 1.08816099e+00 1.00038372e-01 -1.94880709e-01
-7.37956166e-01 1.90387830e-01 2.05310419e-01 3.92139375e-01
2.03999519e-01 4.18873429e-01 -4.62924272e-01 -9.27307382e-02
-6.01738393e-01 5.10891438e-01 1.10270643e+00 8.87522936e-01
5.87958872e-01 -1.76362172e-01 6.89232051e-02 6.15656376e-01
4.54701751e-01 5.86598873e-01 4.55907779e-03 -1.21462452e+00
8.32766950e-01 2.22535715e-01 6.61945283e-01 -7.85207808e-01
-1.99279413e-01 -4.75705862e-01 -4.20303702e-01 -2.56074569e-03
6.21445596e-01 -4.43931341e-01 -4.56407517e-01 1.79933083e+00
2.85329491e-01 1.82382166e-01 -2.59815659e-02 8.75219882e-01
-2.50637650e-01 7.93730140e-01 -1.03235617e-01 -9.82835472e-01
1.08415222e+00 -7.43742526e-01 -8.24720860e-01 -4.73912954e-01
6.05274379e-01 -3.56261969e-01 8.29280019e-01 5.46975553e-01
-1.12626481e+00 1.00696914e-01 -8.95232201e-01 8.23150098e-01
3.97380352e-01 -3.06501091e-01 8.85012150e-01 5.84739923e-01
-4.99994427e-01 3.84880483e-01 -7.01505184e-01 4.34888117e-02
4.14584279e-01 6.69990242e-01 2.52547652e-01 -3.40736181e-01
-1.10671163e+00 5.65883279e-01 6.91304922e-01 1.15082242e-01
-9.35745478e-01 -5.00999868e-01 -5.17548978e-01 2.39564925e-01
1.34000325e+00 -5.03745675e-01 1.65616918e+00 -9.69229877e-01
-1.54197133e+00 4.01542515e-01 -1.45626031e-02 -5.87426543e-01
6.43702984e-01 5.22816703e-02 1.91867892e-02 1.03470683e-01
1.02473125e-01 2.16136336e-01 6.92181706e-01 -1.52129388e+00
-1.34037507e+00 -5.26046574e-01 7.71629512e-01 3.31769496e-01
-4.00777847e-01 -2.95678645e-01 -2.92746007e-01 -2.24073395e-01
2.38277331e-01 -1.09501374e+00 -8.72844994e-01 -6.61451638e-01
-2.50447989e-01 -1.39385864e-01 4.85646516e-01 -6.77806020e-01
1.41307127e+00 -1.86915421e+00 1.58871576e-01 4.42817748e-01
-6.60023689e-02 -2.51405746e-01 1.46712614e-02 3.83820891e-01
1.86391592e-01 2.21834108e-01 -1.77840814e-01 -3.61915715e-02
6.84670731e-02 6.54265583e-01 -4.11780864e-01 3.43780458e-01
-2.73547441e-01 7.90106416e-01 -9.11784947e-01 -4.89918441e-01
2.56945435e-02 -5.85246682e-01 -8.16211224e-01 8.17391872e-02
-4.18502688e-01 2.97672361e-01 -7.78314471e-01 7.72010565e-01
7.93223023e-01 -7.37176538e-02 8.62437069e-01 2.51239985e-01
-2.52645373e-01 1.41781503e-02 -1.54801154e+00 9.93559062e-01
-9.05647933e-01 1.01749913e-03 4.96034861e-01 -1.51146924e+00
4.62513357e-01 5.11677153e-02 9.12489533e-01 -9.85872805e-01
-6.10913038e-02 3.31097066e-01 -3.32069909e-03 -5.49865246e-01
2.93000221e-01 -5.24528205e-01 -2.83112079e-01 8.18144262e-01
-4.32850540e-01 1.23494834e-01 3.15461308e-01 -7.60539621e-02
5.63662469e-01 1.12321943e-01 3.71724516e-01 2.22942187e-03
3.38281602e-01 2.31643245e-01 1.17171872e+00 8.30566287e-01
-2.92036626e-02 -3.80656123e-01 1.11842179e+00 -3.63479406e-02
-9.00424123e-01 -6.28430009e-01 1.11202665e-01 9.08933938e-01
2.93713778e-01 3.19586217e-01 -5.55828989e-01 -6.63568437e-01
3.73364240e-01 9.50210869e-01 -2.39818841e-01 5.35585992e-02
-7.83338964e-01 -8.68831336e-01 -5.02063274e-01 4.68229711e-01
4.79739189e-01 -6.95832074e-01 -7.53371894e-01 7.10148692e-01
-2.45741829e-01 -9.81315196e-01 -4.39249158e-01 9.57340300e-02
-1.11910105e+00 -9.99914467e-01 -4.36052382e-01 -5.44348955e-01
6.88149214e-01 5.00217915e-01 7.92884052e-01 -2.76028544e-01
3.05296361e-01 3.76469940e-01 -6.41607726e-03 -4.63604122e-01
-1.85871199e-01 2.28407513e-03 -1.34332523e-01 1.63824126e-01
3.51438392e-03 -3.59760672e-01 -4.91893709e-01 4.76926118e-01
-8.85562837e-01 1.57716259e-01 7.92430878e-01 1.08095407e+00
8.51765156e-01 6.60727441e-01 8.15976024e-01 -1.09851706e+00
7.67568886e-01 -4.57995713e-01 -1.26326823e+00 3.97008449e-01
-8.39882970e-01 8.65514129e-02 4.03367847e-01 -5.10149181e-01
-1.21730852e+00 1.54063001e-01 3.40842128e-01 -1.26005515e-01
3.63799125e-01 9.56320345e-01 -5.22524953e-01 1.93378523e-01
-1.79026574e-01 1.37049764e-01 1.67728469e-01 -2.95194089e-01
2.42498770e-01 7.46155739e-01 3.44691165e-02 -7.43956089e-01
4.73567277e-01 6.62631631e-01 2.59029120e-01 -3.06504399e-01
-1.07473946e+00 -4.69645858e-01 -3.20636004e-01 -3.59199256e-01
4.02395219e-01 -6.54171824e-01 -1.39822876e+00 6.22777604e-02
-7.35360682e-01 -2.79599875e-01 -1.24477088e-01 7.00591743e-01
-1.09236407e+00 2.14876920e-01 -2.87406504e-01 -1.32653189e+00
1.49681255e-01 -9.79011238e-01 4.36584592e-01 2.55897492e-01
3.27427715e-01 -9.27294791e-01 -2.04131246e-01 8.07196259e-01
-3.91467065e-02 4.58217412e-01 9.48901296e-01 -7.09089190e-02
-9.72024322e-01 -1.75609425e-01 -7.66098080e-03 1.99162662e-01
9.27430578e-03 -3.00518155e-01 -2.66670644e-01 -5.50416589e-01
3.88769776e-01 -1.17972411e-01 6.17603540e-01 6.49790108e-01
1.10238945e+00 -9.76466119e-01 -5.28840184e-01 2.44419798e-01
1.65820789e+00 6.23863101e-01 2.17145160e-01 6.98728383e-01
3.03507537e-01 9.33370888e-01 1.35999465e+00 9.24266517e-01
4.90118653e-01 8.14922273e-01 8.41463745e-01 2.75557339e-01
8.07062149e-01 -3.14404696e-01 3.10069621e-01 2.97335029e-01
-1.20219752e-01 -4.01073843e-02 -6.89057350e-01 4.80773598e-01
-2.31577992e+00 -1.09559727e+00 1.17518827e-01 2.55945206e+00
7.45546699e-01 4.09968793e-01 4.01258945e-01 3.99487205e-02
7.60510623e-01 -1.56439617e-01 -9.97549772e-01 -6.55636549e-01
1.18907496e-01 -2.00406611e-01 1.08037949e+00 5.25767803e-01
-7.88192689e-01 6.03654444e-01 5.36584234e+00 5.52577794e-01
-8.82066667e-01 8.40626284e-02 7.32838571e-01 -2.94613749e-01
-5.12553096e-01 4.56658363e-01 -9.17999685e-01 5.56660414e-01
8.22198331e-01 -6.07815266e-01 7.88367689e-01 9.12721038e-01
8.24993432e-01 -3.44232619e-01 -9.97901499e-01 6.24468029e-01
-3.81788969e-01 -1.16198349e+00 -4.13980514e-01 7.17816412e-01
9.86591518e-01 -6.31940305e-01 -1.28714833e-02 5.61937869e-01
4.86801684e-01 -5.86827159e-01 8.33719194e-01 1.98161662e-01
3.34370077e-01 -1.32368743e+00 7.70818174e-01 7.39057839e-01
-8.76882195e-01 -9.55065429e-01 -2.20505893e-01 -1.88063413e-01
5.70846736e-01 5.92607021e-01 -8.35506141e-01 5.82737267e-01
2.66991138e-01 3.26191545e-01 4.46663380e-01 1.05221462e+00
-4.31048423e-01 5.33634245e-01 -2.98556715e-01 -7.32141584e-02
4.23123032e-01 -6.73625469e-01 4.15603906e-01 5.49345851e-01
9.16716307e-02 1.19680380e-02 6.41055167e-01 7.44889855e-01
7.34351128e-02 1.12703837e-01 -2.25994885e-01 -2.54176944e-01
3.89388233e-01 9.53880310e-01 -5.74312687e-01 -7.49619585e-03
-4.41793799e-01 3.67554098e-01 -8.97843949e-03 4.48591739e-01
-8.01578283e-01 -9.71049145e-02 5.09646535e-01 1.76400363e-01
3.87404442e-01 -1.37278005e-01 -3.92088890e-01 -7.49199390e-01
2.50610828e-01 -8.06910992e-01 5.34264922e-01 -1.65069904e-02
-9.25100386e-01 -2.77997613e-01 7.72759318e-02 -1.06148136e+00
-5.90929508e-01 -3.24196279e-01 -3.59252661e-01 6.98394775e-01
-1.94204903e+00 -7.66718209e-01 2.10959360e-01 2.17934623e-01
3.55857700e-01 3.89358252e-01 2.04574630e-01 2.29773343e-01
-7.78293312e-01 2.28829294e-01 4.63139594e-01 -3.93627524e-01
6.93650246e-02 -1.28752542e+00 -4.50094461e-01 5.31764686e-01
-4.81782198e-01 2.36905321e-01 6.13069654e-01 -6.64817095e-01
-1.75944626e+00 -8.29241574e-01 6.68641090e-01 2.94529945e-01
8.73945177e-01 -2.37471730e-01 -5.51229656e-01 6.16797626e-01
-1.27598748e-01 -3.98806661e-01 4.84708160e-01 2.05592558e-01
3.89050156e-01 -5.21064997e-01 -1.07196748e+00 5.72931767e-01
8.26026618e-01 1.25428289e-01 -2.31292829e-01 5.17295480e-01
8.25689673e-01 -4.07982796e-01 -8.08679700e-01 3.05368602e-01
5.14786720e-01 -4.80676234e-01 7.01523483e-01 -8.59259784e-01
3.27684611e-01 7.85863847e-02 -1.75046548e-01 -1.16285789e+00
-1.83877870e-01 -7.87672579e-01 -4.60541725e-01 1.13006532e+00
4.80422646e-01 -8.38013887e-01 9.14378822e-01 6.37579262e-01
2.81992197e-01 -1.27927661e+00 -1.10786569e+00 -1.08694780e+00
-4.96312380e-02 -2.91837543e-01 6.78510129e-01 7.67633855e-01
3.12159657e-02 9.54976603e-02 -5.67298353e-01 1.70717314e-01
8.29883099e-01 7.92598307e-01 6.52460992e-01 -9.70433176e-01
-7.52466977e-01 -3.34341407e-01 3.88342321e-01 -1.25109673e+00
2.61493593e-01 -6.01549685e-01 8.20808113e-02 -1.63354254e+00
2.75722891e-01 -8.34968328e-01 -3.32004160e-01 2.86031872e-01
2.36439742e-02 -7.41224349e-01 4.57894295e-01 3.27853143e-01
-4.30422813e-01 4.82300669e-01 1.70508301e+00 -1.86613008e-01
-6.46368861e-01 6.82432055e-01 -1.10348821e+00 5.12167573e-01
8.62155855e-01 -4.93526042e-01 -3.76871377e-01 -3.20275477e-03
4.33761239e-01 9.24674749e-01 8.34918022e-02 -9.50469673e-02
-1.01690196e-01 -1.09112370e+00 -4.56395209e-01 -6.20900571e-01
5.77939413e-02 -9.51740444e-01 4.08407720e-03 9.84970868e-01
-2.63460070e-01 -2.09011957e-01 -1.72298312e-01 9.62797582e-01
-3.15456577e-02 -6.90421224e-01 5.55451751e-01 -1.25663415e-01
-5.48758686e-01 3.47757310e-01 -1.92409784e-01 -3.88502479e-02
1.39203012e+00 -9.52405781e-02 9.88141224e-02 -5.12711823e-01
-6.75290167e-01 9.82565224e-01 7.43054301e-02 1.31172925e-01
3.00979763e-01 -1.15080261e+00 -5.39997160e-01 -4.86749560e-01
-1.68725386e-01 1.92174744e-02 2.97842652e-01 9.61958528e-01
-5.22898100e-02 6.92744493e-01 -1.78337656e-02 -4.32260573e-01
-8.95577312e-01 7.94552147e-01 1.97381973e-01 -9.32517350e-01
-1.81128040e-01 2.74706930e-01 -1.31377876e-01 -1.19863644e-01
2.15061352e-01 -1.04037344e-01 -2.05619887e-01 4.18919683e-01
1.56267092e-01 5.19363403e-01 -2.31092393e-01 -1.52597427e-02
-9.28399116e-02 1.36550158e-01 -2.23722398e-01 -4.07132685e-01
1.67745554e+00 -4.82181758e-01 -5.49585652e-03 1.52939912e-02
8.22548866e-01 -2.11032540e-01 -1.60215783e+00 -4.32799995e-01
3.72564673e-01 -8.58201385e-01 2.96297699e-01 -4.62728590e-01
-1.16577137e+00 4.15867120e-01 1.94267064e-01 4.24528092e-01
1.28256977e+00 -2.75173485e-01 7.59843528e-01 2.82758117e-01
5.73460937e-01 -1.67979586e+00 8.13536048e-02 2.33876891e-02
5.83312154e-01 -1.12562025e+00 -1.41024925e-02 -4.22173947e-01
-7.40837395e-01 7.81522751e-01 5.45228779e-01 9.25557241e-02
5.76515198e-01 5.48258908e-02 -4.20617819e-01 2.13552684e-01
-8.02735806e-01 -1.54750839e-01 -2.93560147e-01 2.25443050e-01
-8.25231299e-02 5.56033850e-01 -8.92923951e-01 6.61653697e-01
-1.49992220e-02 8.83610174e-02 7.52693653e-01 1.30249429e+00
-4.49237883e-01 -1.30929160e+00 -4.72406536e-01 6.53501689e-01
-3.93380195e-01 2.75311172e-01 1.56094328e-01 7.25980580e-01
1.51240053e-02 1.03876984e+00 -1.09069392e-01 1.65098697e-01
5.30787706e-01 -3.99456710e-01 5.06226838e-01 -6.13947928e-01
4.42974716e-02 4.55461919e-01 7.71011487e-02 -4.70165998e-01
-4.13076729e-01 -9.08759773e-01 -1.19759429e+00 -1.75673559e-01
-6.52701735e-01 1.76423699e-01 8.22334051e-01 1.12616682e+00
-2.04768442e-02 5.04712641e-01 1.31781960e+00 -4.18551803e-01
-1.36365259e+00 -4.42426980e-01 -8.13220680e-01 5.88882109e-03
1.35753909e-02 -7.35124767e-01 -3.70971203e-01 -2.86140561e-01] | [4.52748966217041, 2.9324777126312256] |
bb910c4a-a1dc-4f7d-9c00-3a5a9add7cf9 | optimal-spatial-signal-design-for-mmwave | 2105.07664 | null | https://arxiv.org/abs/2105.07664v4 | https://arxiv.org/pdf/2105.07664v4.pdf | Optimal Spatial Signal Design for mmWave Positioning under Imperfect Synchronization | We consider the problem of spatial signal design for multipath-assisted mmWave positioning under limited prior knowledge on the user's location and clock bias. We propose an optimal robust design and, based on the low-dimensional precoder structure under perfect prior knowledge, a codebook-based heuristic design with optimized beam power allocation. Through numerical results, we characterize different position-error-bound (PEB) regimes with respect to clock bias uncertainty and show that the proposed low-complexity codebook-based designs outperform the conventional directional beam codebook and achieve near-optimal PEB performance for both analog and digital architectures. | ['Henk Wymeersch', 'Gonzalo Seco-Granados', 'Florent Munier', 'Fan Jiang', 'Musa Furkan Keskin'] | 2021-05-17 | null | null | null | null | ['robust-design'] | ['miscellaneous'] | [-1.87866375e-01 9.13032517e-02 -1.96130082e-01 -3.14116597e-01
-9.91732538e-01 -6.56976104e-01 2.08048239e-01 -2.27386370e-01
-2.03151405e-01 7.95110106e-01 2.59475797e-01 -8.91654015e-01
-7.37261295e-01 -3.98239464e-01 -2.17019126e-01 -1.08883965e+00
-2.83856660e-01 2.27891296e-01 -3.00429285e-01 -5.19289263e-02
3.78203303e-01 5.60368896e-01 -3.04949611e-01 -6.94383025e-01
4.31371957e-01 1.29570055e+00 2.04033881e-01 8.84600639e-01
3.71195793e-01 2.21671268e-01 -6.58587396e-01 -2.19557881e-01
3.27031106e-01 -4.38274562e-01 2.10566118e-01 -2.76876420e-01
-8.21566284e-02 -2.10782036e-01 -1.03130329e+00 7.68533885e-01
1.22462404e+00 -3.85042250e-01 5.68365037e-01 -8.46597254e-01
-6.46338686e-02 6.28810942e-01 -2.40407765e-01 4.02074337e-01
6.71028718e-02 -1.99911848e-01 6.29875243e-01 -7.79373050e-01
4.46330875e-01 6.89569294e-01 1.07977569e+00 1.13710396e-01
-9.12777364e-01 -8.10622811e-01 -4.99421149e-01 1.54687926e-01
-1.83532310e+00 -8.93488467e-01 5.78148484e-01 -1.83979511e-01
3.64862561e-01 3.05494994e-01 4.18311507e-01 4.05167818e-01
6.62237525e-01 1.92105785e-01 8.09530318e-01 -7.07796216e-01
5.42936504e-01 -6.74597248e-02 -2.99440622e-01 6.37589157e-01
8.91365111e-01 4.18422163e-01 -6.14331305e-01 -4.67478991e-01
8.81449997e-01 -4.55854684e-01 -6.81986153e-01 -5.51018596e-01
-1.10385430e+00 3.24488401e-01 3.25254709e-01 6.46858811e-01
-4.35276121e-01 9.43233132e-01 -9.68257308e-01 3.80168498e-01
-7.56160393e-02 6.11240208e-01 -3.00526232e-01 -5.46458922e-02
-1.26858854e+00 1.42950201e-02 8.54537725e-01 1.88267684e+00
4.60134685e-01 2.45906383e-01 -7.53107250e-01 3.32142383e-01
9.83903944e-01 1.19428301e+00 -3.38003099e-01 -6.27848446e-01
6.76784992e-01 -5.55399597e-01 7.14864492e-01 -9.41510260e-01
-9.56282675e-01 -1.74056256e+00 -9.88466084e-01 -5.56337118e-01
3.01326036e-01 -8.33701909e-01 -9.21111941e-01 1.52938128e+00
-2.11335402e-02 2.52811581e-01 2.46621594e-01 5.70526600e-01
5.61730146e-01 4.58734006e-01 -6.72453880e-01 -7.11999714e-01
8.48410606e-01 -2.07584694e-01 -1.17294681e+00 -4.13335234e-01
4.80715185e-01 -7.50981092e-01 -3.82705718e-01 3.71826321e-01
-1.11279953e+00 2.89299250e-01 -1.67485189e+00 6.27086043e-01
4.73808587e-01 2.36433372e-01 2.22394317e-01 1.60567975e+00
-1.18470347e+00 6.98779747e-02 -4.02301192e-01 2.34845672e-02
4.37984943e-01 3.86121184e-01 5.82954884e-01 -2.87456691e-01
-7.83532381e-01 5.88158786e-01 -2.61370361e-01 4.41381276e-01
-2.48173550e-01 -1.21467757e+00 -3.36341858e-01 1.35456309e-01
-1.67511106e-01 -7.65752077e-01 1.57451868e+00 5.29883862e-01
-1.52189720e+00 -1.82698652e-01 -5.19277811e-01 -6.68066502e-01
2.60687292e-01 2.89498389e-01 -6.10714734e-01 1.78751707e-01
-2.48197347e-01 -2.36728117e-01 7.09921360e-01 -1.29379475e+00
-8.55183363e-01 -1.20404035e-01 -3.52581501e-01 -1.31917521e-01
9.33630913e-02 -5.26336253e-01 -4.72814500e-01 -7.70095825e-01
1.16017151e+00 -9.65489388e-01 -8.92292142e-01 -2.69627273e-01
-4.47493553e-01 5.96948266e-01 3.33507329e-01 -1.73434466e-01
1.43573630e+00 -2.10190344e+00 -9.70078856e-02 9.04673159e-01
6.07177913e-02 -3.46079707e-01 1.39486015e-01 8.24067950e-01
4.60473686e-01 -9.70717520e-02 1.53978080e-01 -3.46581578e-01
5.62870950e-02 -1.04731508e-01 -7.76246712e-02 9.22524929e-01
-7.21126020e-01 4.47472870e-01 -7.88253069e-01 1.89817529e-02
-3.22897807e-02 9.55979005e-02 -7.40639985e-01 5.11892065e-02
4.42584902e-01 6.68685794e-01 -5.96101344e-01 7.93371677e-01
1.19684219e+00 6.50083497e-02 4.55253184e-01 -2.36245051e-01
-2.84166574e-01 1.48291275e-01 -1.29892695e+00 1.49033821e+00
-9.27203953e-01 9.80396390e-01 7.17445314e-01 -3.50345463e-01
6.89597487e-01 6.09960496e-01 4.34088886e-01 -7.10043550e-01
2.90008932e-01 4.33525532e-01 3.73845622e-02 -2.05627188e-01
4.44507957e-01 -4.07090545e-01 -3.26235205e-01 3.48541260e-01
2.03803509e-01 -3.13136488e-01 -4.22145069e-01 2.70229727e-01
1.35817194e+00 -7.69696474e-01 2.43942827e-01 -6.95812941e-01
2.86185324e-01 -5.45324802e-01 6.17473483e-01 1.29097605e+00
9.62352753e-02 3.32889974e-01 9.33151245e-02 4.03805912e-01
-8.63690197e-01 -8.02464306e-01 -7.49702513e-01 2.32086658e-01
5.65705478e-01 -4.87884998e-01 -2.26720884e-01 2.34211311e-01
3.13663572e-01 9.04133797e-01 1.12334624e-01 2.46573344e-01
-3.50374550e-01 -9.37853456e-01 4.03097272e-01 2.93027580e-01
4.78962988e-01 5.19553065e-01 -3.04048449e-01 5.60777903e-01
1.78711131e-01 -1.22738147e+00 -2.21076146e-01 4.41428930e-01
-4.50093091e-01 -4.66941088e-01 -7.38345087e-01 -6.08008623e-01
8.14191759e-01 4.66289043e-01 5.62781572e-01 -2.57449150e-01
2.22754717e-01 8.39488566e-01 -3.25153440e-01 -4.42442119e-01
3.97392139e-02 -4.39930521e-02 2.72784412e-01 1.74989719e-02
-4.20351386e-01 -7.63807178e-01 -1.01934612e+00 6.05145097e-01
1.07713252e-01 -3.01795512e-01 9.81093705e-01 6.07921898e-01
2.61010248e-02 2.52135903e-01 6.20116055e-01 -2.07063720e-01
3.02006006e-01 -5.62262177e-01 -1.01534963e+00 -1.06835529e-01
-6.21536255e-01 1.62228063e-01 5.33106923e-02 3.13039601e-01
-9.18502510e-01 2.26884522e-02 -4.66786593e-01 2.90119559e-01
6.56194866e-01 4.55663770e-01 -3.52753103e-01 -1.06737792e+00
4.40158993e-01 3.77140015e-01 -8.35936487e-01 -2.70222753e-01
5.57553947e-01 1.11773062e+00 5.80473661e-01 -2.96576500e-01
1.26318538e+00 4.65882659e-01 6.38890922e-01 -1.15405500e+00
-3.95915836e-01 -5.93825042e-01 -4.18848187e-01 -2.63833910e-01
6.04652911e-02 -1.10086739e+00 -6.58298671e-01 -1.18033662e-01
-1.20281613e+00 5.70431612e-02 5.47573149e-01 8.72966588e-01
-6.23734534e-01 1.94607973e-01 -1.74550563e-02 -1.39046407e+00
-7.89155215e-02 -6.16274357e-01 5.99234104e-01 -5.25741465e-02
-3.52239683e-02 -6.17241979e-01 -1.44561604e-01 -2.77421594e-01
1.03626525e+00 7.87271187e-02 6.74670398e-01 9.80817154e-02
-1.19685233e+00 -5.21574318e-01 5.14071919e-02 -6.68853402e-01
-1.18982837e-01 -1.03299677e+00 -6.33876920e-01 -6.09661639e-01
7.75845796e-02 7.59041786e-01 2.46227220e-01 9.26002562e-01
5.83609283e-01 -5.45601785e-01 -1.03550768e+00 1.00515616e+00
1.73345900e+00 3.78349483e-01 5.66869199e-01 -1.10276960e-01
5.07027209e-02 -4.38460261e-01 5.41539848e-01 8.98932815e-01
2.27813363e-01 7.37247527e-01 2.32949406e-01 5.28329372e-01
-1.86808351e-02 7.82420672e-03 -5.62403917e-01 5.78744054e-01
3.73758554e-01 -6.68304741e-01 -6.04861259e-01 6.62685812e-01
-1.69957793e+00 -7.48751998e-01 -4.15349483e-01 2.10160756e+00
4.57554966e-01 3.32357511e-02 -4.28080082e-01 1.33377418e-01
4.65766430e-01 1.44378319e-01 -1.75960168e-01 1.43484503e-01
-1.58632994e-01 3.73723805e-02 1.72664320e+00 6.86215043e-01
-7.92253077e-01 1.31950721e-01 7.54975128e+00 7.84758270e-01
-6.44381225e-01 4.62751150e-01 -1.69796590e-03 -2.25385502e-01
-5.88313758e-01 -6.86253756e-02 -1.01683605e+00 4.41286236e-01
9.80173767e-01 -4.13198620e-01 1.50293941e-02 4.85032439e-01
2.39183694e-01 -1.82342365e-01 -9.99191523e-01 1.57921743e+00
-3.31294149e-01 -1.90548408e+00 -8.61972272e-01 3.12031806e-01
9.77651060e-01 -2.44812101e-01 2.16224000e-01 -2.17606381e-01
-4.21565510e-02 -4.04016554e-01 8.64996850e-01 7.54625201e-01
7.52252460e-01 -8.02462101e-01 7.72791207e-01 2.72538364e-01
-1.20692861e+00 -6.76321507e-01 -1.65327430e-01 1.54872960e-03
5.16587138e-01 1.20576537e+00 -8.54123831e-01 8.12166214e-01
3.19253951e-01 1.23524450e-01 2.34493107e-01 1.88145721e+00
-3.70901555e-01 8.81487429e-01 -6.76057935e-01 -6.24800205e-01
2.83441126e-01 5.62607348e-02 8.28421116e-01 1.26403081e+00
1.21180809e+00 7.72281468e-01 -4.66432631e-01 1.41148239e-01
3.57958414e-02 -2.35287473e-01 -3.65880162e-01 5.93625247e-01
1.51192153e+00 8.00552368e-01 -2.50892401e-01 2.16557786e-01
-1.51471183e-01 4.01802748e-01 -7.06688583e-01 6.82099283e-01
-4.41473097e-01 -9.05391097e-01 7.49282837e-01 2.41681665e-01
8.40418279e-01 -1.08314168e+00 -1.11096597e+00 -3.11614752e-01
-1.28498793e-01 -6.58183321e-02 -1.74263939e-01 -4.61673796e-01
-4.87059474e-01 4.55852710e-02 -2.75852978e-01 -1.61460078e+00
-1.93104237e-01 -2.25558430e-01 -4.16931093e-01 9.45921481e-01
-1.35840094e+00 -6.20306194e-01 4.66663279e-02 -4.98020910e-02
-2.91900516e-01 -2.88270742e-01 5.04132152e-01 6.30041063e-01
-2.64038835e-02 1.04775202e+00 1.21823597e+00 -9.14128050e-02
1.55733839e-01 -9.07220244e-01 -3.47134471e-02 9.86651599e-01
-1.33756399e-01 7.21012354e-01 1.25330424e+00 -2.92169869e-01
-2.09967875e+00 -1.03575563e+00 8.89850914e-01 -2.32130483e-01
3.24916750e-01 -6.45825207e-01 2.63747960e-01 5.21027267e-01
3.78909111e-01 -1.95253715e-01 9.20961380e-01 7.94793665e-02
3.86008710e-01 -2.24210829e-01 -1.45893025e+00 6.34583712e-01
1.31250381e+00 -5.74609749e-02 -2.74073128e-02 5.63966632e-01
1.86395705e-01 -9.06471968e-01 -7.53546953e-01 4.50637370e-01
6.64132714e-01 -3.75440747e-01 1.04333651e+00 2.49599785e-01
-6.07137620e-01 -5.75318515e-01 -6.70396030e-01 -1.47580111e+00
-9.99277472e-01 -1.49087715e+00 -1.70583516e-01 9.13435519e-01
6.86471820e-01 -6.44705296e-01 1.05906534e+00 -3.04753035e-02
-1.96488589e-01 -7.76472986e-01 -1.62855899e+00 -1.05571115e+00
-3.32281828e-01 -7.59508669e-01 5.50220549e-01 1.21950544e-01
3.04224730e-01 1.42827258e-01 -6.13751054e-01 1.36607301e+00
1.04433036e+00 -1.46774560e-01 6.43262804e-01 -8.00784051e-01
-3.90635699e-01 -3.45891595e-01 -6.68655217e-01 -1.97193766e+00
-6.70795918e-01 -5.09571671e-01 5.27339280e-01 -1.80160248e+00
-5.49476683e-01 -9.81415689e-01 -4.82770056e-03 -8.62309411e-02
4.22993422e-01 4.96277094e-01 -1.34115338e-01 -2.24916190e-01
-4.27096456e-01 6.49692357e-01 6.71725929e-01 -4.47125025e-02
-3.56418602e-02 7.79139519e-01 -8.79804254e-01 3.12543988e-01
4.78933841e-01 -5.16615510e-01 -2.41015375e-01 -4.52729970e-01
5.88078022e-01 7.03036368e-01 -1.29755780e-01 -1.58964097e+00
7.55075216e-01 -4.22092006e-02 4.90498424e-01 -1.11845565e+00
4.69349504e-01 -9.99175549e-01 2.76105762e-01 7.70015597e-01
1.20880313e-01 -5.97373784e-01 -3.55300307e-02 1.09736097e+00
5.02912581e-01 -2.00433224e-01 8.02555799e-01 7.16236532e-01
-1.16551943e-01 2.27907732e-01 -1.06221759e+00 -4.55882311e-01
9.31666613e-01 -1.78009972e-01 -2.40175545e-01 -1.19292688e+00
-6.01647496e-01 6.06222630e-01 -2.76658416e-01 -9.31181312e-02
4.33116287e-01 -1.60173750e+00 -6.19703114e-01 -4.12993133e-02
-3.72593030e-02 -5.52154779e-01 2.93212265e-01 8.02936733e-01
-3.58502626e-01 1.37515628e+00 3.91873330e-01 -7.80879021e-01
-8.87002766e-01 -2.30746910e-01 6.70314729e-01 4.88114417e-01
-2.65300572e-01 1.18858945e+00 -4.96331483e-01 1.85391501e-01
3.48802626e-01 -3.84626299e-01 3.61698180e-01 -4.75874811e-01
4.07804787e-01 8.48946646e-02 4.22585100e-01 -1.77570179e-01
-4.93767351e-01 8.96800280e-01 7.50887513e-01 -5.40237665e-01
8.74688923e-01 -5.86893201e-01 1.98919877e-01 -2.18803555e-01
8.54143858e-01 6.91498995e-01 -9.21744227e-01 -4.62487519e-01
3.84226665e-02 -1.07182622e+00 4.48230267e-01 -7.19610929e-01
-7.23877728e-01 2.18682438e-01 8.45516920e-01 -6.30002376e-03
9.57530022e-01 -5.49312793e-02 4.94538426e-01 8.24340999e-01
1.25098562e+00 -8.42085123e-01 -2.69093245e-01 4.19991851e-01
6.23313963e-01 -3.50955814e-01 4.11933720e-01 -3.88637155e-01
4.16247845e-01 8.17564905e-01 -2.24959806e-01 2.77587831e-01
1.40267241e+00 5.09997845e-01 -4.13235370e-03 -1.82237569e-02
-4.12521362e-01 -1.22329481e-01 9.62644368e-02 7.74800062e-01
1.30080059e-01 9.64604132e-03 -7.21665740e-01 8.86488497e-01
-6.56096935e-01 -3.95540863e-01 6.13551378e-01 1.30058122e+00
-9.81753826e-01 -1.07172942e+00 -8.09020162e-01 5.18942595e-01
-1.47141472e-01 -4.37455364e-02 3.20124120e-01 3.16500455e-01
4.53192405e-02 1.39277554e+00 -4.31726798e-02 -3.92027169e-01
4.60960954e-01 -4.90591496e-01 8.46165657e-01 -3.62803519e-01
3.61727178e-01 1.57860965e-01 4.25254405e-01 -1.75191402e-01
9.93157327e-02 -6.17356241e-01 -8.92744958e-01 -2.14842305e-01
-4.24765468e-01 4.99260515e-01 9.70175207e-01 1.08597684e+00
5.78004360e-01 5.29998004e-01 1.13839018e+00 -7.19486356e-01
-4.36287850e-01 -7.00204909e-01 -7.48217523e-01 -1.08587110e+00
6.93723500e-01 -4.30784255e-01 -1.77152738e-01 -8.23645711e-01] | [6.271188259124756, 1.2144765853881836] |
9c4812df-82ef-4fa7-b84d-05dd1afb72a8 | jrdb-pose-a-large-scale-dataset-for-multi | 2210.11940 | null | https://arxiv.org/abs/2210.11940v2 | https://arxiv.org/pdf/2210.11940v2.pdf | JRDB-Pose: A Large-scale Dataset for Multi-Person Pose Estimation and Tracking | Autonomous robotic systems operating in human environments must understand their surroundings to make accurate and safe decisions. In crowded human scenes with close-up human-robot interaction and robot navigation, a deep understanding requires reasoning about human motion and body dynamics over time with human body pose estimation and tracking. However, existing datasets either do not provide pose annotations or include scene types unrelated to robotic applications. Many datasets also lack the diversity of poses and occlusions found in crowded human scenes. To address this limitation we introduce JRDB-Pose, a large-scale dataset and benchmark for multi-person pose estimation and tracking using videos captured from a social navigation robot. The dataset contains challenge scenes with crowded indoor and outdoor locations and a diverse range of scales and occlusion types. JRDB-Pose provides human pose annotations with per-keypoint occlusion labels and track IDs consistent across the scene. A public evaluation server is made available for fair evaluation on a held-out test set. JRDB-Pose is available at https://jrdb.erc.monash.edu/ . | ['Hamid Rezatofighi', 'Jianfei Cai', 'Duy Tho Le', 'Edward Vendrow'] | 2022-10-20 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Vendrow_JRDB-Pose_A_Large-Scale_Dataset_for_Multi-Person_Pose_Estimation_and_Tracking_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Vendrow_JRDB-Pose_A_Large-Scale_Dataset_for_Multi-Person_Pose_Estimation_and_Tracking_CVPR_2023_paper.pdf | cvpr-2023-1 | ['multi-person-pose-estimation', 'multi-person-pose-estimation-and-tracking', 'social-navigation'] | ['computer-vision', 'computer-vision', 'robots'] | [-4.89957124e-01 -9.56488773e-02 7.40966648e-02 -3.93205762e-01
-5.06253839e-01 -5.45773625e-01 3.84158522e-01 -3.81930172e-01
-6.32595062e-01 7.10663736e-01 3.67075533e-01 3.47205788e-01
1.46839499e-01 -2.79859304e-01 -6.79501951e-01 -3.06823879e-01
-3.70334685e-01 1.06332541e+00 5.00535548e-01 -3.48569989e-01
-2.91747063e-01 3.53672057e-01 -1.52177024e+00 1.12130061e-01
1.45571649e-01 6.00979388e-01 1.90109998e-01 1.02448034e+00
8.18998814e-01 9.23875451e-01 -4.67980146e-01 -3.55338126e-01
7.26414025e-01 6.42076693e-03 -8.07210863e-01 8.16437751e-02
9.35876787e-01 -6.13252282e-01 -8.65821481e-01 6.87233269e-01
1.00093246e+00 4.33669835e-01 4.90643114e-01 -1.80099940e+00
-1.56874523e-01 -8.48564208e-02 -5.02600491e-01 1.53830171e-01
1.03732836e+00 6.56404853e-01 6.73960626e-01 -6.83113873e-01
1.12450957e+00 1.80516732e+00 1.21761239e+00 7.31135070e-01
-8.79019439e-01 -5.44321537e-01 4.40368652e-02 3.46377820e-01
-1.36476672e+00 -6.16545141e-01 1.63665071e-01 -7.16884732e-01
1.05519640e+00 -1.20597966e-01 7.85020471e-01 1.60027254e+00
1.55941740e-01 9.26149845e-01 6.16700470e-01 3.59449685e-02
5.02619334e-02 -3.08962137e-01 1.05836719e-01 8.54334652e-01
5.88177919e-01 1.58929490e-02 -8.45788062e-01 -2.18134299e-01
6.52496755e-01 8.20549205e-02 -6.91526160e-02 -1.03045952e+00
-1.50208592e+00 5.22559822e-01 5.38572371e-01 -3.97815704e-01
-4.02117103e-01 4.26700801e-01 5.20798266e-01 6.76880777e-02
-2.45630294e-01 1.79794610e-01 -5.36394477e-01 -2.96192497e-01
-6.70155361e-02 9.07102585e-01 9.91944849e-01 1.49176967e+00
3.91140819e-01 -4.24809963e-01 -9.83010232e-02 6.99707925e-01
3.33063871e-01 1.12768304e+00 2.31268808e-01 -1.52279341e+00
6.00007951e-01 3.13266516e-01 6.86131239e-01 -1.14311683e+00
-1.07762372e+00 2.35934611e-02 -2.32149482e-01 2.91524142e-01
8.72472227e-01 -3.51095051e-01 -7.05866516e-01 1.68200767e+00
8.67603362e-01 -4.23297614e-01 -1.08436421e-02 1.35130489e+00
9.73682940e-01 -9.72990601e-05 2.19814524e-01 5.19851327e-01
1.76309371e+00 -1.24544013e+00 -6.11081421e-01 -8.25323224e-01
6.19920313e-01 -5.16305745e-01 8.04970086e-01 4.02990520e-01
-5.50010204e-01 -5.46977162e-01 -9.09461319e-01 -4.50073242e-01
-2.84417331e-01 1.37640178e-01 5.62323809e-01 3.70537966e-01
-8.45934689e-01 6.37310445e-02 -1.06918061e+00 -1.12658656e+00
1.70929313e-01 2.38101602e-01 -7.74680793e-01 -3.28980893e-01
-1.03794229e+00 1.18973064e+00 1.03490286e-01 1.47478998e-01
-9.48759913e-01 -2.31476605e-01 -1.16600740e+00 -7.02466130e-01
4.74808156e-01 -9.93916392e-01 1.74543512e+00 -1.47062242e-01
-1.02351308e+00 1.04447722e+00 6.97537363e-02 -5.79044580e-01
1.10680878e+00 -9.30107534e-01 -2.27547616e-01 1.81901440e-01
5.82783639e-01 9.31001425e-01 4.80604708e-01 -9.99130666e-01
-7.23592281e-01 -7.65297651e-01 -1.72821488e-02 6.29321337e-01
2.97249913e-01 -7.86947310e-02 -8.89598906e-01 -2.21415132e-01
2.39101514e-01 -1.52277684e+00 -2.83419251e-01 4.58881170e-01
-4.64039981e-01 6.06765039e-02 8.99894476e-01 -7.35950172e-01
3.00469428e-01 -1.81845582e+00 2.31811583e-01 -1.70284674e-01
1.32070944e-01 -1.83150753e-01 1.04498252e-01 3.06889206e-01
4.39992458e-01 -7.16117024e-01 3.32759231e-01 -7.26479292e-01
3.86921853e-01 2.39288732e-01 1.70736730e-01 1.04683554e+00
-4.58853841e-01 7.69994974e-01 -9.33306873e-01 -7.19935894e-01
2.88203567e-01 5.04663885e-01 -5.32939970e-01 9.43334252e-02
-9.62650701e-02 7.98647761e-01 -3.50288123e-01 9.49482024e-01
3.28377694e-01 -1.23468891e-01 7.09438473e-02 -1.61771491e-01
2.93866336e-01 -1.78487167e-01 -1.46779156e+00 2.08606195e+00
2.01558009e-01 6.96007967e-01 3.07735026e-01 -1.44216716e-01
4.47597533e-01 7.38291815e-02 6.24478102e-01 -4.87068534e-01
2.65366971e-01 -6.68312833e-02 -2.44449586e-01 -6.38726473e-01
9.13618147e-01 3.65791947e-01 -4.84083116e-01 3.15218642e-02
-1.00519851e-01 7.29112932e-03 2.46471852e-01 1.23962872e-01
1.55004179e+00 5.17327487e-01 3.54660362e-01 -6.32955208e-02
1.12126440e-01 5.11175513e-01 7.71880567e-01 9.36571360e-01
-9.89993930e-01 6.55938804e-01 -2.97200847e-02 -8.68674695e-01
-1.12730479e+00 -1.32423198e+00 -3.36096244e-04 1.40638113e+00
4.86693352e-01 -3.73950332e-01 -6.69420481e-01 -2.06952676e-01
3.69700193e-01 2.52996441e-02 -5.14871359e-01 1.53181061e-01
-7.21134603e-01 -4.16802377e-01 7.18975365e-01 5.76806545e-01
6.33007050e-01 -1.19979203e+00 -1.20411503e+00 3.13824937e-02
-6.85669959e-01 -1.70731497e+00 -4.27828908e-01 -2.81436414e-01
-2.38087177e-01 -1.58689165e+00 -7.82189608e-01 -7.08207965e-01
3.66993099e-01 3.14227223e-01 1.09615493e+00 -3.06558251e-01
-8.50304425e-01 1.00750756e+00 -4.03217643e-01 -5.38633287e-01
1.97610501e-02 -2.70997398e-02 7.20017672e-01 -5.15834808e-01
5.31216502e-01 -1.67518541e-01 -9.18648303e-01 7.91398346e-01
1.40223041e-01 -2.14736998e-01 2.95321703e-01 4.62273270e-01
3.89262170e-01 -4.22609091e-01 2.05975026e-02 -3.35663587e-01
6.81097806e-02 -4.22294289e-01 -5.04406810e-01 5.32359183e-02
1.80727765e-01 -3.79737228e-01 -8.44277292e-02 -2.97526598e-01
-9.47313666e-01 5.70969820e-01 2.24500239e-01 -2.51816541e-01
-4.49857295e-01 -3.50886226e-01 -2.76228666e-01 8.99073780e-02
1.13835776e+00 -2.30716527e-01 6.32416606e-02 -3.41777951e-01
3.07422638e-01 5.42852521e-01 1.22088397e+00 -7.78712511e-01
5.48100650e-01 8.00655484e-01 -1.20773487e-01 -7.83566654e-01
-4.99332845e-01 -7.95179129e-01 -9.65162635e-01 -5.40197074e-01
1.20674264e+00 -1.48454440e+00 -1.14507115e+00 7.87858188e-01
-1.08129871e+00 -6.84758306e-01 5.72378375e-02 7.00731516e-01
-9.68284070e-01 3.00698608e-01 -7.20390439e-01 -8.55674326e-01
-2.14441299e-01 -1.07529366e+00 1.47437263e+00 1.51863992e-02
-7.68699944e-01 -3.83993208e-01 1.43230751e-01 5.77107131e-01
-1.17752112e-01 6.46713138e-01 -2.97351003e-01 -3.45611632e-01
-5.39526284e-01 -4.92684990e-01 9.34168845e-02 -4.50335115e-01
-2.02880129e-01 -5.64329326e-01 -7.85291433e-01 -5.79476118e-01
-4.69239980e-01 -7.39995539e-01 3.53778034e-01 4.91351932e-01
2.90814281e-01 -6.51093349e-02 -6.27781987e-01 4.43176001e-01
7.45756686e-01 -1.84543222e-01 3.29305053e-01 8.50431204e-01
8.30606580e-01 8.19763064e-01 9.82312500e-01 7.05275238e-01
9.56073105e-01 1.14333403e+00 3.07080090e-01 2.33093873e-01
-1.83992103e-01 -1.75720870e-01 4.72602695e-01 -1.36287794e-01
-5.78054711e-02 -1.07258141e-01 -1.39997947e+00 4.00093526e-01
-2.15459347e+00 -1.02695358e+00 -2.54926413e-01 1.90120077e+00
3.90428573e-01 8.45777467e-02 6.11503184e-01 -3.60279679e-01
7.51730859e-01 -1.25098214e-01 -9.03391898e-01 5.12562633e-01
-1.49388850e-01 -8.91789198e-01 7.93284178e-01 4.90913242e-01
-1.49652731e+00 9.54687178e-01 6.02748966e+00 5.97181395e-02
-4.82042491e-01 2.14913189e-01 -1.35019377e-01 -4.58841562e-01
8.14382613e-01 -5.05511642e-01 -1.19773328e+00 1.69448763e-01
5.25188982e-01 3.71393472e-01 2.03173295e-01 1.30905545e+00
-7.92353451e-02 -3.18477958e-01 -1.24773991e+00 1.38460064e+00
5.09487605e-03 -5.29528856e-01 -8.01824033e-01 6.20144196e-02
4.99669343e-01 5.18169224e-01 -1.47169098e-01 5.01709938e-01
8.27233613e-01 -8.28376889e-01 1.07189155e+00 5.73257387e-01
4.43459481e-01 -4.18612987e-01 5.87913513e-01 5.53993106e-01
-1.43463087e+00 -3.37897092e-01 -3.37915272e-01 -3.83409142e-01
3.73806387e-01 -1.60743400e-01 -7.52559543e-01 8.20019320e-02
1.46305144e+00 7.40588665e-01 -7.85409093e-01 1.07009482e+00
-6.05464764e-02 -3.36458981e-01 -5.85314572e-01 -2.82341279e-02
-1.86152011e-01 3.00025344e-01 7.25288272e-01 1.07277381e+00
1.16042525e-01 7.64457211e-02 6.51861668e-01 9.54452530e-02
3.20821762e-01 -2.22304761e-01 -7.65170336e-01 6.48556769e-01
6.79096580e-01 1.06494713e+00 -7.17557013e-01 -2.10406706e-01
-1.33910000e-01 1.08865964e+00 2.66040832e-01 2.96887338e-01
-9.60371673e-01 1.16357757e-02 1.19565499e+00 1.53171331e-01
6.63213059e-02 -5.97480476e-01 9.10987183e-02 -1.04120338e+00
2.29201138e-01 -9.64859366e-01 6.38292074e-01 -9.07463014e-01
-1.31971908e+00 4.35918123e-01 3.08502346e-01 -1.34851611e+00
-5.66274941e-01 -7.96472013e-01 1.75990716e-01 4.18885559e-01
-5.36944509e-01 -1.29580033e+00 -8.33276391e-01 8.34157884e-01
6.20758533e-01 -1.81237042e-01 7.13404775e-01 2.82249570e-01
-4.33563173e-01 5.22430480e-01 -1.98035896e-01 4.85639006e-01
1.17667973e+00 -1.13413668e+00 6.65013015e-01 6.27484977e-01
-4.63831097e-01 5.61362326e-01 1.25975204e+00 -9.52392519e-01
-1.64101636e+00 -1.05620944e+00 5.05710185e-01 -1.39438450e+00
3.19310725e-01 -7.09738255e-01 -4.04244661e-01 1.13559520e+00
-1.01481140e-01 2.47987777e-01 3.05695206e-01 1.77399248e-01
-3.85246098e-01 1.55999362e-01 -1.28747451e+00 7.01965451e-01
1.70953190e+00 -1.99852064e-01 -7.51620948e-01 7.19887137e-01
5.77768326e-01 -1.09195232e+00 -6.46707058e-01 4.38260317e-01
9.97640729e-01 -7.84564197e-01 1.36212134e+00 -3.56462121e-01
-1.05194665e-01 -5.93797386e-01 -5.06931007e-01 -8.22120547e-01
-3.08220059e-01 -3.95908087e-01 -3.04688215e-01 5.92566669e-01
-1.02412635e-02 -3.67052644e-01 1.04628074e+00 1.02365959e+00
1.11866236e-01 -1.00831777e-01 -8.73070836e-01 -1.03741109e+00
-3.89102727e-01 -3.39988351e-01 2.84401119e-01 3.91151845e-01
-3.49661410e-02 2.58372068e-01 -6.99252784e-01 4.36401576e-01
1.07067430e+00 -3.31764638e-01 1.73338306e+00 -1.16896927e+00
-2.00539201e-01 2.98681781e-02 -8.44097912e-01 -9.27565873e-01
1.48892850e-01 -1.76378787e-01 6.45938396e-01 -1.57515967e+00
1.50135055e-01 -1.90165699e-01 6.22772336e-01 4.95126367e-01
4.36486043e-02 4.05977041e-01 3.24206740e-01 6.24831021e-01
-1.31393266e+00 4.25479531e-01 8.70731831e-01 -2.60483585e-02
6.15511984e-02 -1.46044478e-01 -8.56424645e-02 1.03127015e+00
7.04883575e-01 -4.26931739e-01 -2.58309931e-01 -4.05003756e-01
-8.22036713e-02 3.85120846e-02 9.34422553e-01 -1.81695247e+00
6.89168990e-01 1.03591211e-01 8.19939852e-01 -8.11566472e-01
9.85424101e-01 -7.18775570e-01 4.86094922e-01 8.62985194e-01
-6.11258782e-02 2.96786159e-01 1.17710508e-01 8.14874291e-01
4.19907719e-01 3.41118366e-01 5.91869414e-01 -3.64811510e-01
-1.31814146e+00 2.61108905e-01 -2.61350393e-01 2.73862213e-01
1.11531508e+00 -3.07435364e-01 -4.73646998e-01 -5.53573966e-01
-9.24683392e-01 6.83271170e-01 7.70412862e-01 9.46909964e-01
2.81232774e-01 -1.28591490e+00 -5.89600146e-01 -1.29201382e-01
4.85396504e-01 9.24191177e-02 4.10536230e-01 8.07278097e-01
-9.51701105e-01 3.96955132e-01 -5.31876445e-01 -9.43924963e-01
-1.43576586e+00 2.54942089e-01 4.87830281e-01 1.90167695e-01
-8.91777933e-01 8.08988571e-01 1.26384467e-01 -1.22680807e+00
6.00852132e-01 5.01056574e-02 1.30360961e-01 -2.75072217e-01
7.57682920e-01 7.50733912e-01 -3.38861287e-01 -1.31715226e+00
-6.67856574e-01 5.82709253e-01 3.67693156e-01 -2.48964757e-01
8.85508060e-01 -6.05693460e-01 2.95425087e-01 2.07340211e-01
1.03658915e+00 -3.27399790e-01 -1.66435289e+00 -3.46439630e-01
5.43944817e-03 -4.05865490e-01 -7.30619490e-01 -6.12107158e-01
-4.68670309e-01 3.33567411e-01 8.89660358e-01 -5.90587914e-01
4.05715704e-01 2.04006270e-01 7.55992174e-01 1.22960556e+00
1.09835660e+00 -1.40001249e+00 3.18005502e-01 9.17125821e-01
1.12456453e+00 -1.61713040e+00 1.35604218e-01 -3.37590098e-01
-9.00066912e-01 6.12098813e-01 1.19401836e+00 1.40982985e-01
3.74371797e-01 2.80313343e-01 4.35931385e-01 -3.07040602e-01
-2.37025127e-01 -3.95737112e-01 -2.02213675e-02 1.07445002e+00
-3.50473002e-02 1.94334731e-01 2.88774133e-01 5.43583333e-01
-7.89851010e-01 -1.28623173e-01 1.78478271e-01 1.29315031e+00
-5.35718083e-01 -4.38876480e-01 -1.04841721e+00 -2.41693519e-02
-1.77911699e-01 6.08135641e-01 -4.34330344e-01 9.95734870e-01
9.68420655e-02 1.12541997e+00 -1.04784429e-01 -4.92687285e-01
7.08202779e-01 -8.99976678e-03 5.93626678e-01 -4.18130249e-01
-1.60478458e-01 -2.45759070e-01 3.51142466e-01 -1.06592929e+00
-2.69163549e-01 -1.21708381e+00 -1.49641430e+00 -4.81625289e-01
2.10202143e-01 -2.93786168e-01 6.08143985e-01 6.10769153e-01
3.25759053e-01 2.94303656e-01 -4.04852897e-01 -1.61842203e+00
-5.04708946e-01 -1.08670270e+00 -5.02379656e-01 7.09747434e-01
3.91934931e-01 -1.12922800e+00 5.60818017e-02 1.71068296e-01] | [7.033779621124268, -0.8377019762992859] |
dae2ea5a-fc13-476f-beb4-150aef405324 | context-modulated-dynamic-networks-for-actor | null | null | https://ojs.aaai.org//index.php/AAAI/article/view/6895 | https://ojs.aaai.org/index.php/AAAI/article/view/6895/6749 | Context Modulated Dynamic Networks for Actor and Action Video Segmentation with Language Queries | Actor and action video segmentation with language queries aims to segment out the expression referred objects in the video. This process requires comprehensive language reasoning and fine-grained video understanding. Previous methods mainly leverage dynamic convolutional networks to match visual and semantic representations. However, the dynamic convolution neglects spatial context when processing each region in the frame and is thus challenging to segment similar objects in the complex scenarios. To address such limitation, we construct a context modulated dynamic convolutional network. Specifically, we propose a context modulated dynamic convolutional operation in the proposed framework. The kernels for the specific region are generated from both language sentences and surrounding context features. Moreover, we devise a temporal encoder to incorporate motions into the visual features to further match the query descriptions. Extensive experiments on two benchmark datasets, Actor-Action Dataset Sentences (A2D Sentences) and J-HMDB Sentences, demonstrate that our proposed approach notably outperforms state-of-the-art methods. | ['Yi Yang', 'Fan Ma', 'Cheng Deng', 'Hao Wang'] | 2020-04-03 | null | null | null | null | ['referring-expression-segmentation'] | ['computer-vision'] | [ 1.83290645e-01 -3.58684182e-01 -4.55142379e-01 -5.03048956e-01
-6.13330901e-01 -4.71242249e-01 6.03722572e-01 -2.88025558e-01
-5.21779597e-01 4.65958863e-01 4.92732763e-01 -1.52144611e-01
2.46469840e-01 -5.18711567e-01 -8.94547999e-01 -4.31436419e-01
2.55003214e-01 -4.43837382e-02 7.34901011e-01 -8.70790258e-02
2.64175504e-01 3.96510929e-01 -1.14445496e+00 7.79280066e-01
5.11722088e-01 1.13243651e+00 5.77080488e-01 6.32399440e-01
-4.87988561e-01 1.38513732e+00 -5.14205217e-01 -2.79394686e-01
5.09440154e-02 -4.54549134e-01 -1.09968913e+00 4.53892857e-01
3.71218264e-01 -6.61695540e-01 -6.88065767e-01 9.86932218e-01
2.81336993e-01 4.64209855e-01 2.43221238e-01 -1.08223271e+00
-8.12462151e-01 5.89494407e-01 -5.57079613e-01 6.78715348e-01
7.63795733e-01 4.54423130e-01 7.88104415e-01 -8.99142146e-01
9.29274380e-01 1.56700730e+00 2.52899706e-01 5.73031425e-01
-7.01818585e-01 -4.59148318e-01 8.54083121e-01 5.06371796e-01
-1.40164363e+00 -5.16922057e-01 9.45640564e-01 -4.06297475e-01
1.06946778e+00 8.68550092e-02 8.17271054e-01 1.38372159e+00
-1.18090972e-01 1.43153107e+00 6.05782688e-01 -4.66346368e-03
2.35625803e-02 -2.14923427e-01 -1.86534926e-01 7.90096760e-01
-4.22873914e-01 -1.84692711e-01 -4.18639839e-01 3.68960917e-01
9.37327921e-01 4.06028271e-01 -2.61113465e-01 -2.66706020e-01
-1.49991596e+00 4.98925149e-01 3.22727233e-01 4.44659948e-01
-4.83713120e-01 6.78743422e-01 8.02535415e-01 -1.61317308e-02
1.60248935e-01 -2.27655411e-01 -4.47732627e-01 -2.37956151e-01
-9.88024712e-01 2.34135330e-01 5.07272184e-01 1.32424629e+00
6.32605612e-01 1.13372039e-02 -7.45434701e-01 5.61751842e-01
3.37780118e-01 3.47779691e-01 2.63535291e-01 -1.21294224e+00
8.01983535e-01 6.06866956e-01 1.20318361e-01 -1.05325985e+00
-8.42641294e-02 -5.05259708e-02 -5.87757587e-01 -5.22283494e-01
2.67743498e-01 2.74885055e-02 -1.04761982e+00 1.50557005e+00
4.49347913e-01 5.63172340e-01 2.41378218e-01 1.21023595e+00
1.12174261e+00 9.23447609e-01 6.06367707e-01 -2.10646614e-01
1.44814289e+00 -1.46189475e+00 -9.30369139e-01 -3.60615253e-01
4.25445706e-01 -6.93240404e-01 1.07150030e+00 -1.76609114e-01
-1.32098413e+00 -8.58574092e-01 -7.84302950e-01 -4.06772614e-01
-2.73182452e-01 8.57482478e-02 6.19295716e-01 7.83923045e-02
-8.26624215e-01 1.53026626e-01 -8.49261642e-01 -4.25562501e-01
6.22504890e-01 8.51639882e-02 -1.96688160e-01 -2.27427095e-01
-1.32345641e+00 3.69426578e-01 6.64217293e-01 3.31320733e-01
-1.24940109e+00 -4.40715462e-01 -1.03212655e+00 1.27513096e-01
8.66231859e-01 -7.28116930e-01 1.34423125e+00 -1.29827380e+00
-1.42288291e+00 8.19670498e-01 -3.94493192e-01 -4.73491907e-01
6.35362208e-01 -2.83584177e-01 -3.57861131e-01 7.18582630e-01
1.55844450e-01 1.02506125e+00 8.92025709e-01 -1.05031419e+00
-7.74095356e-01 -5.17400280e-02 6.82001293e-01 3.29459578e-01
2.37066746e-02 3.63094538e-01 -1.40915692e+00 -9.21212077e-01
-1.25856861e-01 -6.67838752e-01 -2.97377944e-01 1.05061226e-01
-3.69649202e-01 -2.45393902e-01 1.15250695e+00 -7.03957081e-01
1.31937861e+00 -2.14765334e+00 2.46263549e-01 -3.15746993e-01
2.10811030e-02 2.35667497e-01 -1.22610256e-01 2.13047460e-01
1.27318814e-01 -6.40136823e-02 -9.67378691e-02 -3.66338700e-01
2.41700001e-02 3.11413676e-01 -3.56159687e-01 2.88255215e-01
3.72800052e-01 1.23635066e+00 -8.79547715e-01 -9.38212156e-01
3.81548941e-01 5.19257963e-01 -6.13917232e-01 3.34613442e-01
-7.02966571e-01 5.07686555e-01 -9.82042134e-01 9.55007195e-01
4.37067807e-01 -4.18947011e-01 -8.75262022e-02 -4.66980666e-01
-3.03328037e-02 -1.66776597e-01 -9.38223660e-01 2.36998725e+00
-3.17798316e-01 6.67151332e-01 -2.73012705e-02 -1.23487103e+00
5.99903703e-01 2.48235211e-01 6.08282685e-01 -8.13941240e-01
1.33170292e-01 -1.68656707e-01 -4.08229738e-01 -1.16821122e+00
5.64841926e-01 2.92129010e-01 -2.46064574e-01 1.25715524e-01
-2.66477875e-02 2.32843608e-01 2.44533777e-01 2.84711868e-01
7.93142438e-01 5.53252637e-01 1.62429616e-01 1.26965448e-01
9.80611861e-01 -3.27060074e-02 7.54481554e-01 6.66789293e-01
-6.93148911e-01 4.66295660e-01 6.01302266e-01 -6.27876699e-01
-7.18275666e-01 -8.11640382e-01 3.40081155e-01 1.26739597e+00
6.79624319e-01 -2.15937778e-01 -8.14009666e-01 -8.41992021e-01
-3.07611763e-01 5.53358197e-01 -6.41773820e-01 -4.15656157e-02
-8.56104255e-01 -2.30473176e-01 4.81616497e-01 9.88388896e-01
9.64490473e-01 -1.25691605e+00 -7.87186980e-01 1.84612319e-01
-4.42373306e-01 -1.66698444e+00 -9.19984281e-01 -3.02134871e-01
-5.51039875e-01 -1.11751926e+00 -7.47795403e-01 -1.08174551e+00
2.79283047e-01 3.05954635e-01 9.97967958e-01 6.36484474e-02
-2.58804023e-01 7.47033477e-01 -5.24887145e-01 1.54598132e-01
-2.05428123e-01 -8.21478441e-02 -4.48855698e-01 2.02933684e-01
6.09216094e-01 -1.43660650e-01 -1.00539076e+00 3.18353415e-01
-1.09562898e+00 3.23099256e-01 5.93786120e-01 6.27180517e-01
7.88433373e-01 -2.79231519e-01 3.96037728e-01 -6.21940076e-01
2.24244624e-01 -5.20926595e-01 -4.12195414e-01 6.75727844e-01
3.15309241e-02 -1.08363703e-01 4.29587185e-01 -5.07321477e-01
-1.31486380e+00 3.94268453e-01 -4.93986197e-02 -8.46348405e-01
-3.67195278e-01 2.18460470e-01 -3.18411946e-01 2.42716640e-01
3.13391164e-02 4.89741296e-01 -4.00223762e-01 -2.49110162e-01
6.91056192e-01 4.31494534e-01 7.62600005e-01 -7.31456816e-01
3.40758473e-01 6.62685394e-01 -3.46653789e-01 -5.43002367e-01
-8.62246394e-01 -5.51361918e-01 -7.47097731e-01 -3.42971563e-01
1.70171845e+00 -1.09947503e+00 -7.94078231e-01 1.58152252e-01
-1.51197231e+00 -3.10995132e-01 7.50286039e-03 3.14060211e-01
-8.47280324e-01 4.25371259e-01 -6.10593617e-01 -5.83045006e-01
-1.99876726e-01 -1.58049738e+00 1.41776824e+00 3.45722854e-01
2.13812530e-01 -9.52580392e-01 -3.89478087e-01 5.26796222e-01
1.80003271e-01 3.43051910e-01 5.41732371e-01 -5.21767855e-01
-1.16018152e+00 5.94673343e-02 -5.56826234e-01 1.10943682e-01
-1.57958141e-03 9.95095670e-02 -7.28993952e-01 -5.25022149e-02
-1.33800328e-01 -3.59411925e-01 9.66324806e-01 3.52053225e-01
1.65567946e+00 -2.18269616e-01 -4.04781103e-01 7.92318046e-01
1.22454870e+00 4.62581187e-01 6.06380999e-01 2.67088234e-01
8.86608839e-01 4.71144170e-01 9.19739842e-01 4.85045046e-01
4.60488170e-01 6.07183278e-01 3.52830201e-01 -1.08929873e-02
-1.76286325e-01 -3.57295245e-01 2.78806686e-01 5.97845137e-01
-4.10711616e-02 -3.04266900e-01 -7.59564281e-01 5.67012012e-01
-2.11732364e+00 -1.23855352e+00 3.16523999e-01 1.35939765e+00
5.53462565e-01 1.12193212e-01 -7.37905279e-02 -4.66884136e-01
8.11830103e-01 7.04179943e-01 -7.79425263e-01 -5.25944680e-02
-2.11501464e-01 -3.11362416e-01 2.59643227e-01 2.53283560e-01
-1.32341397e+00 1.42335999e+00 5.73935795e+00 8.82644176e-01
-1.02211976e+00 2.17442691e-01 7.86672056e-01 -1.80815712e-01
-1.84902981e-01 1.06445841e-01 -6.75825059e-01 3.75748694e-01
4.68054742e-01 1.83646958e-02 2.86833167e-01 8.69526327e-01
6.34349167e-01 -1.38273180e-01 -1.14748788e+00 1.19991076e+00
1.97276950e-01 -1.62122226e+00 2.79192179e-01 -5.06242335e-01
7.28208423e-01 -2.67647445e-01 -1.16072915e-01 5.18858910e-01
-2.29593158e-01 -9.06174541e-01 9.62078691e-01 9.80271518e-01
7.01693594e-01 -6.28793418e-01 3.86956483e-01 3.77886668e-02
-1.58249438e+00 -2.29435861e-01 -1.70188099e-01 2.48110995e-01
5.27175069e-01 -1.61994025e-01 -4.34480786e-01 4.30762112e-01
9.16835368e-01 1.11644721e+00 -6.00716054e-01 5.74830890e-01
4.62252386e-02 3.73013884e-01 2.02744812e-01 1.51379975e-02
6.99786544e-01 -6.11179322e-02 3.55517536e-01 1.43574536e+00
-1.82835776e-02 2.55698204e-01 5.29248655e-01 8.79433811e-01
-4.92149927e-02 1.80396978e-02 -4.03281510e-01 -2.76641101e-01
3.61993760e-01 8.65452528e-01 -8.25191200e-01 -6.99601352e-01
-8.99632573e-01 1.44646728e+00 1.91467836e-01 8.50103915e-01
-1.18026173e+00 -2.40943700e-01 6.61370158e-01 -1.39215544e-01
6.50591910e-01 -2.98179030e-01 2.13550612e-01 -1.29646468e+00
2.04997826e-02 -8.31870735e-01 5.49597263e-01 -9.95694757e-01
-9.27105188e-01 4.84465331e-01 2.93007016e-01 -1.13611567e+00
-2.49282092e-01 -4.55347508e-01 -4.66920882e-01 4.58533317e-01
-1.53116298e+00 -1.29670453e+00 -5.45061827e-01 1.00350690e+00
1.28562307e+00 2.32292488e-02 1.56272605e-01 4.58728462e-01
-6.90058291e-01 3.20825428e-01 -3.02990317e-01 4.32415664e-01
5.06045640e-01 -7.44786024e-01 2.79827148e-01 7.99695432e-01
-3.47987749e-02 5.72902977e-01 3.86208802e-01 -6.50557399e-01
-1.47176147e+00 -1.29088652e+00 4.58181620e-01 -1.45372942e-01
6.49578273e-01 -1.82337537e-01 -8.40772510e-01 8.17032337e-01
4.12214875e-01 3.93985003e-01 3.33100289e-01 -7.39046693e-01
-1.44823432e-01 5.55079505e-02 -8.51234496e-01 7.19189405e-01
1.43828893e+00 -8.21460307e-01 -7.16068506e-01 4.06843305e-01
1.36163747e+00 -6.49779081e-01 -7.22540975e-01 3.12356949e-01
4.88423526e-01 -9.50973809e-01 1.22405326e+00 -8.89059424e-01
6.36217594e-01 -5.10641813e-01 -4.31783557e-01 -4.14493561e-01
3.77561264e-02 -5.31318426e-01 -3.38062972e-01 1.25025022e+00
-5.87414838e-02 8.37606490e-02 7.60645270e-01 6.09560192e-01
-2.52032071e-01 -9.62984383e-01 -6.84793830e-01 -4.65192050e-01
-2.30277807e-01 -6.40010297e-01 5.77201128e-01 7.65097260e-01
-4.74404484e-01 9.11856890e-02 -3.37984771e-01 6.20492958e-02
2.03763187e-01 3.16518664e-01 7.31202304e-01 -4.02382463e-01
-2.74846494e-01 -3.85984749e-01 -5.09901404e-01 -1.74310601e+00
5.66738307e-01 -5.75711429e-01 -2.25369609e-03 -1.56575632e+00
3.38360906e-01 -7.61680529e-02 -1.97412878e-01 2.90658593e-01
-3.37947309e-01 8.58862698e-03 3.60189974e-01 4.80744950e-02
-1.30524790e+00 6.77528799e-01 1.55990660e+00 -4.46662128e-01
-2.08872139e-01 -2.45372325e-01 -2.78059453e-01 7.40594029e-01
4.52711523e-01 -3.20193470e-02 -6.51374400e-01 -7.44630516e-01
-9.79008675e-02 4.30707455e-01 6.60920024e-01 -9.28580403e-01
3.51702511e-01 -5.84364891e-01 5.24958253e-01 -9.86029923e-01
4.23917532e-01 -9.18251991e-01 -8.77799653e-03 3.15011948e-01
-6.17689013e-01 1.89100027e-01 8.61217007e-02 9.78084803e-01
-4.63852584e-01 -3.71357650e-02 4.88583982e-01 -4.64487165e-01
-1.50308776e+00 6.75705373e-01 -2.55926490e-01 2.73200095e-01
1.35373366e+00 -2.51417369e-01 -1.69902116e-01 -3.29755783e-01
-7.65711367e-01 5.96470416e-01 3.68864506e-01 7.72824049e-01
9.31274116e-01 -1.30382717e+00 -2.97229409e-01 -1.16730966e-01
1.57648250e-01 1.81542784e-01 7.19040811e-01 9.51203048e-01
-7.42360651e-01 5.35155654e-01 5.30931912e-03 -7.99367607e-01
-1.09059095e+00 1.03160846e+00 3.99476856e-01 1.59976915e-01
-7.14699447e-01 8.17070484e-01 5.08213937e-01 2.13700086e-01
5.09691715e-01 -6.64542496e-01 -3.27728748e-01 5.76133747e-03
5.18204033e-01 1.97069235e-02 -6.65026426e-01 -9.42941368e-01
-4.05000895e-01 6.85150445e-01 -1.26979247e-01 9.33811814e-02
1.00427771e+00 -5.87039769e-01 3.54047194e-02 3.84533048e-01
1.44610715e+00 -3.54982615e-01 -1.74162257e+00 -4.19918209e-01
-3.18745486e-02 -6.31207824e-01 -2.85948217e-01 -3.17456841e-01
-1.25860095e+00 9.05750513e-01 4.31003332e-01 -1.19636618e-01
1.32731593e+00 2.68425643e-01 1.04987931e+00 5.51883996e-01
1.72174871e-02 -1.16451132e+00 5.28823197e-01 4.71798420e-01
8.08354795e-01 -1.31462085e+00 -8.92624259e-02 -4.38128859e-01
-9.46758568e-01 1.17515349e+00 1.00441611e+00 -1.45018939e-02
4.36792970e-01 -8.46137628e-02 1.68055922e-01 -2.16473028e-01
-7.02386916e-01 -4.18019652e-01 2.32449308e-01 3.34585875e-01
2.37609297e-01 -3.66836786e-01 -3.13312739e-01 5.64149201e-01
4.75515932e-01 6.75410405e-02 1.80580944e-01 1.07858288e+00
-3.71983171e-01 -6.45788491e-01 -1.86463982e-01 -1.32821165e-02
-5.36609828e-01 -3.70015651e-02 -1.99544787e-01 8.23103309e-01
2.04395860e-01 7.46252656e-01 1.69407710e-01 -1.44881457e-01
2.66562700e-01 1.20221831e-01 2.31943876e-01 -3.29666436e-01
-3.80144596e-01 2.91702241e-01 -1.36450812e-01 -1.06345201e+00
-1.07238042e+00 -6.13760293e-01 -1.59617317e+00 1.94782332e-01
1.89815685e-01 -2.46714920e-01 3.19707662e-01 1.08573568e+00
2.86990136e-01 8.52308691e-01 3.55256498e-01 -8.09413493e-01
-1.20876797e-01 -6.32278979e-01 -1.73183605e-01 6.31129920e-01
4.63871241e-01 -6.71494842e-01 4.17797118e-02 4.75609958e-01] | [9.827776908874512, 0.6541498899459839] |
1b7a4dca-1ad8-44bd-af8e-57a1c6e684aa | remax-relaxing-for-better-training-on | 2306.17319 | null | https://arxiv.org/abs/2306.17319v1 | https://arxiv.org/pdf/2306.17319v1.pdf | ReMaX: Relaxing for Better Training on Efficient Panoptic Segmentation | This paper presents a new mechanism to facilitate the training of mask transformers for efficient panoptic segmentation, democratizing its deployment. We observe that due to its high complexity, the training objective of panoptic segmentation will inevitably lead to much higher false positive penalization. Such unbalanced loss makes the training process of the end-to-end mask-transformer based architectures difficult, especially for efficient models. In this paper, we present ReMaX that adds relaxation to mask predictions and class predictions during training for panoptic segmentation. We demonstrate that via these simple relaxation techniques during training, our model can be consistently improved by a clear margin \textbf{without} any extra computational cost on inference. By combining our method with efficient backbones like MobileNetV3-Small, our method achieves new state-of-the-art results for efficient panoptic segmentation on COCO, ADE20K and Cityscapes. Code and pre-trained checkpoints will be available at \url{https://github.com/google-research/deeplab2}. | ['Liang-Chieh Chen', 'Philip Torr', 'Andrew Howard', 'Qihang Yu', 'Weijun Wang', 'Shuyang Sun'] | 2023-06-29 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [-6.03816565e-03 1.72872826e-01 -4.13158149e-01 -5.09386122e-01
-8.74307334e-01 -7.56403625e-01 3.28466326e-01 -2.07320184e-01
-2.80278146e-01 5.65222502e-01 -2.16284409e-01 -7.69006252e-01
1.45978883e-01 -6.00478709e-01 -5.89462101e-01 -6.23228729e-01
-3.70985448e-01 9.27111506e-01 3.41288269e-01 2.07094941e-02
-2.15858798e-02 4.84440118e-01 -1.25456953e+00 6.64830729e-02
1.09596264e+00 9.42252874e-01 1.66818857e-01 9.59431410e-01
2.28323802e-01 6.64076984e-01 -5.04758596e-01 -3.82379711e-01
8.81828129e-01 -8.58110189e-02 -9.17322516e-01 1.77221924e-01
9.47677255e-01 -6.36133313e-01 -2.97462702e-01 1.01334500e+00
3.99640977e-01 2.45480895e-01 3.14409494e-01 -9.31889951e-01
2.62497842e-01 9.30830002e-01 -8.45013022e-01 5.15102506e-01
-6.90618157e-01 6.09454036e-01 1.29281235e+00 -7.05376685e-01
3.17194581e-01 1.11312568e+00 7.64394581e-01 2.35027507e-01
-1.27901983e+00 -8.33958864e-01 3.63368869e-01 -1.94574565e-01
-1.35543346e+00 -6.39710665e-01 2.61966139e-01 -2.33240545e-01
1.06017137e+00 5.69895208e-01 8.15715373e-01 4.50147927e-01
-1.73205942e-01 9.89025533e-01 9.72685039e-01 1.25590842e-02
-6.37189299e-02 2.41321735e-02 1.88280761e-01 8.73270214e-01
1.09334402e-01 1.96077779e-01 1.62193224e-01 -2.83564534e-02
5.16970575e-01 -2.15137318e-01 -1.85272112e-01 1.74234629e-01
-7.56786764e-01 8.99234474e-01 7.26767182e-01 -5.22019602e-02
-8.91810134e-02 4.68000442e-01 2.07556874e-01 1.45837024e-01
1.08546031e+00 6.03147566e-01 -8.47563982e-01 -9.81779024e-02
-1.56501234e+00 3.47519130e-01 8.28476369e-01 6.83348894e-01
9.88904953e-01 2.02104256e-01 -1.09337457e-01 8.61029267e-01
3.10133398e-01 8.78078401e-01 -2.13146925e-01 -1.32453537e+00
6.78467512e-01 3.20535511e-01 -3.40589397e-02 -5.47692418e-01
-5.76791763e-01 -9.39987600e-01 -7.38130331e-01 8.89669880e-02
3.97423655e-01 -5.77165723e-01 -1.47486520e+00 1.49268699e+00
6.61738038e-01 5.82496464e-01 -2.81136423e-01 9.03882861e-01
2.90693134e-01 9.65037525e-01 -4.14086357e-02 5.83873019e-02
1.14362323e+00 -1.39704871e+00 -2.21171662e-01 -5.05666792e-01
7.32230186e-01 -8.53665471e-01 1.02869368e+00 1.76948965e-01
-1.08645332e+00 -3.33851278e-01 -7.09901035e-01 4.74954024e-02
-1.23499341e-01 1.08876295e-01 7.81535327e-01 4.98783529e-01
-1.16977453e+00 6.72020197e-01 -1.27656412e+00 -2.59175807e-01
6.01919174e-01 5.19277275e-01 3.98391664e-01 3.66015524e-01
-9.35533643e-01 6.72334731e-01 3.38240445e-01 2.45317683e-01
-1.05826795e+00 -9.82711315e-01 -6.13056540e-01 2.70172149e-01
4.85357314e-01 -4.99012619e-01 1.73404276e+00 -7.58858562e-01
-1.40683246e+00 8.95474911e-01 1.25572337e-02 -9.52549696e-01
5.78554690e-01 -4.06336337e-01 -3.68869677e-02 2.72076577e-01
2.18968838e-02 1.38418245e+00 8.71559441e-01 -1.08351541e+00
-9.38754976e-01 -1.85671419e-01 9.09089819e-02 2.36922860e-01
2.24456415e-02 -1.23663761e-01 -7.64472783e-01 -6.26927555e-01
-4.96752597e-02 -1.25028384e+00 -7.39697278e-01 -1.11600548e-01
-6.09496236e-01 5.85179888e-02 7.24957049e-01 -7.17482865e-01
1.24178004e+00 -1.91332948e+00 -2.62781233e-01 3.56471568e-01
4.74545330e-01 5.81681609e-01 -1.25160873e-01 -3.64125660e-03
3.17518830e-01 3.86467308e-01 -8.83910835e-01 -6.79797173e-01
-8.34590867e-02 3.81836355e-01 -5.66382647e-01 5.70017099e-01
1.03261292e-01 7.44117141e-01 -6.92837059e-01 -5.07820308e-01
4.54503059e-01 2.41666481e-01 -7.66515791e-01 1.31755009e-01
-7.99659550e-01 4.79409218e-01 -3.46629977e-01 7.53788531e-01
1.06901872e+00 -5.99699676e-01 6.38279468e-02 2.76711911e-01
-2.93385595e-01 6.40385211e-01 -8.09866190e-01 1.50906515e+00
-5.04880905e-01 7.79059768e-01 4.43069726e-01 -8.12347412e-01
3.98790091e-01 1.84171461e-02 3.22281152e-01 -5.86660326e-01
1.04735039e-01 1.72065556e-01 5.94383627e-02 2.28732228e-02
6.53150737e-01 3.09546590e-01 3.79000515e-01 3.71408880e-01
-2.05093130e-01 -4.86736953e-01 3.38874072e-01 8.43286291e-02
6.16229951e-01 -2.61680409e-02 -2.21202478e-01 -5.69744527e-01
4.22757119e-02 2.84098893e-01 5.72704077e-01 8.78607035e-01
-1.34903580e-01 4.48466718e-01 3.87978047e-01 -5.48147559e-01
-1.07754350e+00 -9.33097720e-01 -5.54402351e-01 1.09036267e+00
-1.24535590e-01 -3.05314600e-01 -8.41320693e-01 -7.29364932e-01
-5.65963238e-02 6.66536331e-01 -5.25746346e-01 5.04560530e-01
-8.39033961e-01 -1.38566637e+00 8.43675196e-01 2.13212222e-01
7.22590446e-01 -7.22188771e-01 -5.98671615e-01 2.11787075e-01
-9.99888927e-02 -1.12470841e+00 -5.10032952e-01 3.40724766e-01
-1.32226980e+00 -8.15022051e-01 -7.48711944e-01 -4.62804765e-01
5.33577859e-01 3.23301524e-01 1.37696040e+00 3.13589215e-01
-9.72622558e-02 -2.82309771e-01 -8.51192772e-02 -2.08447084e-01
-2.65155047e-01 7.44162202e-01 -4.34828758e-01 -3.88595074e-01
-6.94372132e-03 -7.42908061e-01 -1.00869095e+00 3.09679717e-01
-6.53673291e-01 3.99828225e-01 4.03435498e-01 4.58661258e-01
6.99168444e-01 -2.74442077e-01 -1.51369825e-01 -1.10372114e+00
-5.77236116e-02 -5.39256155e-01 -1.27692294e+00 3.25114354e-02
-8.53936136e-01 -3.00395396e-02 4.51254368e-01 -5.23677766e-02
-8.17430675e-01 -3.15442532e-02 -4.61875409e-01 -5.00499666e-01
3.21646705e-02 3.33758354e-01 4.04751956e-01 6.71728253e-02
4.99611914e-01 -5.27004488e-02 -3.63370419e-01 -7.12217689e-01
4.86663938e-01 5.12801290e-01 3.57422382e-01 -4.56219554e-01
7.88538992e-01 6.68506205e-01 -2.11378947e-01 -8.44909787e-01
-1.25031447e+00 -5.26947856e-01 -1.91775098e-01 -1.15981167e-02
6.42007828e-01 -1.18138945e+00 -4.02915776e-01 5.53989768e-01
-8.63154173e-01 -1.14772558e+00 -3.06027591e-01 2.64772862e-01
-1.59551248e-01 1.66927278e-01 -8.57173681e-01 -5.91382861e-01
-8.75210822e-01 -1.22819781e+00 1.11211598e+00 2.71794468e-01
1.03305094e-01 -1.01797748e+00 3.02672237e-01 4.37485367e-01
4.80988681e-01 -3.98285175e-03 4.94141936e-01 -3.95974427e-01
-9.66944933e-01 2.38107949e-01 -4.54035670e-01 4.40472513e-01
-1.53369576e-01 4.09464478e-01 -1.29671741e+00 -4.41993177e-01
-1.24510437e-01 -2.37104550e-01 1.43855512e+00 7.70224571e-01
1.40502119e+00 -6.59559846e-01 -3.27798516e-01 1.35580039e+00
1.38728166e+00 -6.49824739e-02 5.27186513e-01 1.05007350e-01
7.86220491e-01 3.19481492e-01 4.69292223e-01 4.38994199e-01
4.15335894e-01 5.70511401e-01 6.00712061e-01 -5.25053680e-01
-1.13412865e-01 -5.11428118e-02 1.75443247e-01 7.19234407e-01
1.01229679e-02 -5.94886184e-01 -1.26208782e+00 7.69410729e-01
-1.74905968e+00 -6.79675102e-01 -2.19431207e-01 2.02940965e+00
9.79716718e-01 2.11322948e-01 7.91611150e-03 -3.40480536e-01
5.37624180e-01 7.52414942e-01 -6.04128480e-01 -2.64302015e-01
8.59261397e-03 4.73837018e-01 1.11241961e+00 1.12447739e+00
-1.41079724e+00 1.59318769e+00 5.63369417e+00 1.13013101e+00
-1.30636919e+00 3.45772535e-01 1.22434068e+00 -2.26022467e-01
-2.11789042e-01 2.29028821e-01 -1.12444949e+00 4.17643845e-01
1.16571450e+00 4.16284651e-01 5.54485381e-01 7.49056220e-01
3.93931299e-01 -1.19768307e-01 -6.00394487e-01 5.17353356e-01
-6.34588838e-01 -1.62271047e+00 -2.75808852e-02 1.12274021e-01
1.01038277e+00 1.12857747e+00 2.46961251e-01 1.98587462e-01
7.53125250e-01 -8.08461845e-01 5.40442109e-01 -1.47137865e-01
8.15889299e-01 -6.84737504e-01 4.07918751e-01 9.83000249e-02
-1.21417367e+00 2.45184407e-01 -4.21977043e-01 -1.99326817e-02
1.81504801e-01 9.09825623e-01 -1.04432297e+00 2.23911911e-01
7.29377985e-01 7.36807942e-01 -4.26247090e-01 1.13058341e+00
-3.79395157e-01 1.29956341e+00 -9.36573803e-01 4.29687709e-01
8.81930172e-01 -2.47020036e-01 7.27138519e-01 1.46962547e+00
4.09121700e-02 -3.69401202e-02 4.36028540e-01 8.10211182e-01
-3.49984258e-01 -8.22930411e-02 -3.45852464e-01 9.65374187e-02
4.23944950e-01 1.27849400e+00 -1.07120347e+00 -6.62554979e-01
-1.30723283e-01 5.89831471e-01 2.66815573e-01 3.69797587e-01
-1.16883647e+00 1.94199923e-02 9.21710193e-01 1.39388084e-01
4.78177905e-01 -2.03575656e-01 -5.99125445e-01 -1.24923635e+00
-2.82783717e-01 -8.48488748e-01 3.83129239e-01 -3.00414026e-01
-6.87015772e-01 6.84974790e-01 1.01620860e-01 -8.44510257e-01
-5.80631196e-03 -2.68681586e-01 -7.60582626e-01 5.97916126e-01
-1.76786816e+00 -1.07136321e+00 -2.97739860e-02 1.48389682e-01
8.48322034e-01 3.03660065e-01 3.80232751e-01 5.77439249e-01
-7.35184968e-01 5.58759511e-01 1.33125693e-01 9.90910083e-02
3.34368110e-01 -1.35267544e+00 1.04973078e+00 1.23453844e+00
3.09264094e-01 1.58546552e-01 6.65765107e-01 -4.59295750e-01
-6.61971688e-01 -1.38980758e+00 7.28384614e-01 -2.76983857e-01
8.83188605e-01 -3.31804127e-01 -7.61981666e-01 9.95153606e-01
2.85273761e-01 -1.09204650e-01 2.61672169e-01 2.07615525e-01
-1.14430852e-01 -2.85865903e-01 -8.70503902e-01 6.07643783e-01
8.46077442e-01 -1.25495702e-01 -1.32304132e-01 7.82512486e-01
1.10252512e+00 -7.64883876e-01 -4.93166476e-01 6.83797300e-01
3.01516294e-01 -8.17364275e-01 7.92513072e-01 -3.27475309e-01
2.43601680e-01 -2.37854272e-01 -3.43966261e-02 -1.03820574e+00
-2.45569255e-02 -1.03307354e+00 -2.11834013e-01 8.70008469e-01
9.33247626e-01 -7.05251098e-01 9.33166087e-01 2.28573427e-01
-4.61473823e-01 -7.91221678e-01 -1.00233948e+00 -5.96442819e-01
2.45074183e-01 -5.92834592e-01 5.31537831e-01 8.90070498e-01
-7.49994159e-01 1.39731616e-01 -5.06029189e-01 6.61955595e-01
6.44713223e-01 3.76510024e-01 7.46344030e-01 -9.96156752e-01
-5.14250576e-01 -6.17584646e-01 3.39882433e-01 -1.68435097e+00
-1.79197509e-02 -8.72355700e-01 1.33172438e-01 -1.25103366e+00
3.90823884e-03 -9.83580351e-01 -8.07355717e-02 6.94541216e-01
-2.39748731e-01 3.96498859e-01 3.46787721e-01 4.33417022e-01
-5.08119404e-01 4.87946123e-01 1.38225865e+00 -2.39407435e-01
-4.72787112e-01 3.28322202e-01 -3.00359845e-01 7.62572587e-01
1.32991898e+00 -9.27052021e-01 -4.42814291e-01 -1.02924263e+00
1.26791626e-01 -3.47159915e-02 3.59967500e-01 -1.00224197e+00
-6.75058961e-02 -1.28859058e-01 -1.49934202e-01 -8.13534200e-01
3.85930568e-01 -5.94598889e-01 -8.16048309e-02 7.19496489e-01
-6.61613271e-02 2.86819011e-01 6.02892280e-01 1.30821094e-01
-4.55984473e-02 3.18987705e-02 1.10305262e+00 -2.13218376e-01
-4.90602911e-01 6.98372781e-01 -3.06431085e-01 4.97120798e-01
5.67769110e-01 1.45554870e-01 -4.75990325e-01 -1.13170825e-01
-6.27415180e-01 8.66018534e-01 4.07532513e-01 -8.90320539e-02
8.54203030e-02 -4.69307780e-01 -8.71165872e-01 1.02471076e-02
-4.00782973e-01 5.03332555e-01 2.73789257e-01 1.23639905e+00
-1.13629138e+00 3.58124971e-01 1.97167143e-01 -8.81382227e-01
-9.74558175e-01 1.27119077e-02 7.01163650e-01 -8.05992305e-01
-8.48669350e-01 1.15036237e+00 4.20189917e-01 -6.35366499e-01
2.53739059e-01 -8.13054264e-01 3.56851757e-01 2.81742383e-02
2.54163176e-01 2.99916178e-01 -5.74754402e-02 -8.51197764e-02
-1.93795696e-01 3.32028180e-01 -3.06483686e-01 -2.44595394e-01
1.38873398e+00 -9.51212719e-02 -4.50446941e-02 4.15381007e-02
9.39487576e-01 -1.19633704e-01 -1.78277028e+00 -1.08255148e-01
-1.96967527e-01 -3.38186800e-01 3.88634533e-01 -8.88385415e-01
-1.67693305e+00 1.11586571e+00 6.08907759e-01 -9.10071284e-03
1.09466326e+00 -1.11267842e-01 1.10793579e+00 5.07876933e-01
-3.36748362e-02 -1.01441741e+00 -6.39828563e-01 7.10610092e-01
3.71175766e-01 -1.34291923e+00 8.65195990e-02 -3.16606730e-01
-3.95148069e-01 6.46772861e-01 5.87772071e-01 -2.49369577e-01
8.61841857e-01 3.28388393e-01 2.96806395e-01 -3.61434937e-01
-8.35793376e-01 -4.38512146e-01 5.21365292e-02 -3.63407261e-03
1.66870002e-02 3.62220556e-01 1.17038809e-01 -2.98038185e-01
-3.08424711e-01 -3.94254804e-01 2.06363127e-01 6.01886094e-01
-6.73143446e-01 -7.86252797e-01 -1.10275514e-01 5.51766157e-01
-8.05416703e-01 -7.85503924e-01 -2.64588576e-02 9.29081202e-01
7.29105920e-02 7.00511754e-01 4.97514248e-01 -5.97038791e-02
-2.08848417e-01 -2.62040824e-01 1.25695124e-01 -5.22328138e-01
-7.38330007e-01 3.10417920e-01 3.12074333e-01 -5.97863853e-01
-3.66749525e-01 -3.99954110e-01 -1.26174831e+00 -8.03728819e-01
-2.81676233e-01 2.98098236e-01 3.95496488e-01 7.09794343e-01
3.43957037e-01 2.48974606e-01 6.31720901e-01 -9.78725374e-01
-4.98282254e-01 -1.06951880e+00 -3.21292847e-01 -2.12297007e-01
5.22679210e-01 -1.28515795e-01 -4.81218189e-01 -1.02144621e-01] | [9.465005874633789, 0.11435875296592712] |
613aba6c-5cc6-4655-8211-ed635da4cd0d | towards-single-integrated-spoofing-aware | 2305.19051 | null | https://arxiv.org/abs/2305.19051v2 | https://arxiv.org/pdf/2305.19051v2.pdf | Towards single integrated spoofing-aware speaker verification embeddings | This study aims to develop a single integrated spoofing-aware speaker verification (SASV) embeddings that satisfy two aspects. First, rejecting non-target speakers' input as well as target speakers' spoofed inputs should be addressed. Second, competitive performance should be demonstrated compared to the fusion of automatic speaker verification (ASV) and countermeasure (CM) embeddings, which outperformed single embedding solutions by a large margin in the SASV2022 challenge. We analyze that the inferior performance of single SASV embeddings comes from insufficient amount of training data and distinct nature of ASV and CM tasks. To this end, we propose a novel framework that includes multi-stage training and a combination of loss functions. Copy synthesis, combined with several vocoders, is also exploited to address the lack of spoofed data. Experimental results show dramatic improvements, achieving a SASV-EER of 1.06% on the evaluation protocol of the SASV2022 challenge. | ['Jee-weon Jung', 'Nam Soo Kim', 'Tomi Kinnunen', 'Nicholas Evans', 'Junichi Yamagishi', 'Kong Aik Lee', 'Massimiliano Todisco', 'Min Hyun Han', 'Myeonghun Jeong', 'Md Sahidullah', 'Xuechen Liu', 'Xin Wang', 'Hemlata Tak', 'Hye-jin Shim', 'Sung Hwan Mun'] | 2023-05-30 | null | null | null | null | ['speaker-verification'] | ['speech'] | [ 3.19254369e-01 8.66810232e-03 -1.16924606e-01 -4.23190266e-01
-9.66400623e-01 -7.20461011e-01 8.45148265e-01 -4.85575013e-02
-3.97967011e-01 2.26451173e-01 4.57510054e-01 -7.56634593e-01
3.51547420e-01 -1.75080135e-01 -4.67814058e-01 -5.49960673e-01
2.09300909e-02 6.70696720e-02 1.65961444e-01 -3.02285403e-01
-1.21469364e-01 3.94260496e-01 -1.44799447e+00 9.78503004e-02
7.23817527e-01 8.93260837e-01 -1.61549479e-01 7.82234907e-01
-3.37167867e-02 2.88148195e-01 -8.52638483e-01 -7.88815141e-01
1.67478383e-01 -2.28138432e-01 -4.42833334e-01 -3.18968803e-01
6.83575273e-01 -4.35111374e-01 -5.09313941e-01 1.11707306e+00
8.95759225e-01 -7.03776553e-02 3.64550054e-01 -1.51677823e+00
-7.11698592e-01 1.04784238e+00 -2.08699778e-01 1.86452210e-01
2.98185319e-01 1.84827790e-01 1.01777494e+00 -1.06676733e+00
2.96460748e-01 1.50418150e+00 6.40342236e-01 1.09153366e+00
-1.12583435e+00 -1.09761560e+00 1.44410297e-01 1.98865175e-01
-1.46909404e+00 -1.00513685e+00 9.79200602e-01 -3.05316895e-01
8.16890061e-01 5.43078542e-01 9.66393799e-02 1.75920177e+00
-2.62546033e-01 7.77470529e-01 8.45232189e-01 -3.28652829e-01
1.34036511e-01 7.68950820e-01 4.37528431e-01 4.08246785e-01
2.78934836e-01 5.61096489e-01 -6.59732521e-01 -3.71412963e-01
-4.88175303e-02 -4.10715044e-01 -5.29559314e-01 -1.45502180e-01
-9.44937110e-01 9.68153656e-01 1.97872877e-01 2.41630927e-01
-6.48982078e-02 -2.12428495e-02 6.86590493e-01 4.60408002e-01
2.27271691e-01 7.91084319e-02 -2.81298697e-01 6.28236681e-02
-1.03140593e+00 -1.70908812e-02 6.02494597e-01 7.73960471e-01
3.45895410e-01 6.44641817e-01 -2.37670586e-01 8.92647088e-01
6.41167819e-01 9.00536895e-01 6.68783784e-01 -4.92383003e-01
7.04519391e-01 9.51714590e-02 1.44790068e-01 -9.58261192e-01
1.51783451e-01 -6.17698312e-01 -7.61660516e-01 1.93674222e-01
-7.72355497e-02 -1.80288460e-02 -9.96974528e-01 1.95436108e+00
3.22837532e-01 3.78128946e-01 4.44578409e-01 8.36191714e-01
8.25619459e-01 5.34281552e-01 -1.68019906e-01 -4.63400856e-02
1.54148543e+00 -9.57977116e-01 -1.04843497e+00 -1.53001696e-01
4.54038680e-01 -8.23730767e-01 9.52302396e-01 1.72244638e-01
-6.07531488e-01 -6.98893368e-01 -1.35216546e+00 1.23367235e-01
-4.63604867e-01 -5.75952046e-02 1.53904900e-01 1.59164989e+00
-1.23603559e+00 1.33339062e-01 -4.73585337e-01 2.47851592e-02
3.15542579e-01 4.54219162e-01 -4.85822350e-01 1.16464652e-01
-1.61041117e+00 6.50948584e-01 1.23095542e-01 1.28597215e-01
-1.38902593e+00 -6.42839134e-01 -1.08060098e+00 2.36299723e-01
-1.72820762e-01 -1.15784504e-01 1.05800259e+00 -7.03740358e-01
-1.65280104e+00 7.38461733e-01 -2.88045466e-01 -7.71332264e-01
6.92021310e-01 -2.72425771e-01 -1.10200500e+00 8.93871784e-02
-1.19121164e-01 2.70004034e-01 1.23692322e+00 -1.38215435e+00
-3.86828452e-01 -2.42583230e-01 -4.52894956e-01 -2.85525113e-01
-8.36212695e-01 2.88656473e-01 -1.41970783e-01 -4.83444959e-01
-8.32339469e-03 -8.94596159e-01 2.24815980e-01 -3.41044724e-01
-5.26157260e-01 -2.93617100e-01 1.17775404e+00 -9.41566646e-01
1.35280848e+00 -2.61142612e+00 -2.88327664e-01 1.59970447e-01
-2.31311563e-02 9.97833312e-01 -4.93101954e-01 4.19414282e-01
-1.91809088e-01 4.57712203e-01 -2.60188311e-01 -9.80348229e-01
3.38732928e-01 -1.65883154e-01 -7.60612249e-01 6.20123923e-01
3.14177662e-01 5.56135237e-01 -6.62717879e-01 -3.37192804e-01
2.32448325e-01 9.43973482e-01 -6.01940632e-01 2.39372909e-01
2.98903137e-01 6.40273541e-02 4.77767512e-02 6.60970032e-01
1.30707729e+00 3.13922495e-01 7.91445598e-02 1.05148815e-01
1.27154663e-01 7.38272309e-01 -1.30352974e+00 1.22737134e+00
-3.86715353e-01 6.43906891e-01 6.01464272e-01 -6.93072259e-01
1.00655305e+00 7.88711250e-01 5.31048179e-02 -4.17888254e-01
2.46251389e-01 4.43192571e-01 -5.73528074e-02 -2.21765891e-01
3.44372451e-01 -1.03163265e-01 -2.36334233e-03 2.98034757e-01
2.89729655e-01 1.16066270e-01 -5.00801802e-01 3.44519466e-01
9.37729657e-01 -5.49240410e-01 -1.52529955e-01 -2.36549005e-01
1.00342572e+00 -5.27817428e-01 7.45840371e-01 8.52586448e-01
-1.00189805e+00 4.23069537e-01 1.83029518e-01 4.79730256e-02
-7.19114840e-01 -1.20777118e+00 -3.29836488e-01 8.89577687e-01
1.96231574e-01 -4.40566361e-01 -6.53335035e-01 -8.73038113e-01
2.50893176e-01 8.45564544e-01 -6.38414681e-01 -2.94946760e-01
-6.57202363e-01 -4.04671937e-01 1.11007810e+00 2.25480720e-01
3.02547157e-01 -7.67412484e-01 5.50111122e-02 1.43536851e-01
-1.88153625e-01 -1.29864550e+00 -7.76629984e-01 1.38369620e-01
-3.67701828e-01 -8.17531466e-01 -5.47468483e-01 -9.27696228e-01
2.19804674e-01 6.30504787e-01 5.48773766e-01 -3.81561145e-02
1.62386924e-01 3.37025002e-02 -3.99478167e-01 -4.10188705e-01
-7.78933048e-01 -3.87265049e-02 5.98547041e-01 3.69832665e-01
5.42216599e-01 -3.38282734e-01 -2.71848261e-01 4.40927595e-01
-7.51475632e-01 -5.69786847e-01 3.36749226e-01 1.07608187e+00
-8.96867514e-02 -8.78241733e-02 7.23374188e-01 -3.63446862e-01
5.15441060e-01 -2.83784926e-01 -4.07505602e-01 2.05125138e-01
-7.02131808e-01 -9.78487283e-02 4.82329220e-01 -6.61570132e-01
-8.52505207e-01 -2.37950400e-01 -5.66120982e-01 -6.10051274e-01
-3.31274532e-02 -1.78658798e-01 -5.40852785e-01 -8.18291306e-02
3.97986203e-01 5.65370321e-01 2.13355348e-01 -6.17259741e-01
2.50554651e-01 1.33075225e+00 4.64038968e-01 -2.75246333e-02
1.16415668e+00 2.09821790e-01 -6.65268123e-01 -1.07398891e+00
-1.84269100e-01 -6.77282393e-01 -1.67435467e-01 2.11882576e-01
6.05442703e-01 -1.09375906e+00 -8.04329932e-01 5.78264952e-01
-1.27474117e+00 1.91756099e-01 7.60980323e-02 7.14681387e-01
1.64109707e-01 7.57172644e-01 -6.28383696e-01 -1.20456362e+00
-5.79668283e-01 -1.40352869e+00 9.89469588e-01 -3.67270738e-01
-1.19276568e-01 -6.42760396e-01 1.04517914e-01 7.03015625e-01
6.69619262e-01 -1.98892877e-01 4.53015149e-01 -1.20601881e+00
-2.69515812e-01 -4.26521927e-01 -4.81498986e-02 9.58098114e-01
1.88925669e-01 -6.35483935e-02 -1.59458649e+00 -8.27336490e-01
2.61681408e-01 -1.87422708e-01 1.16288817e+00 -4.07777317e-02
7.95685172e-01 -4.39502686e-01 -2.60670573e-01 8.36527407e-01
1.19165063e+00 7.88889155e-02 3.55822951e-01 -7.96845276e-03
5.79325497e-01 7.33578026e-01 -4.71074060e-02 1.13289826e-01
1.53677925e-01 8.43087018e-01 6.36512578e-01 2.42191300e-01
-4.56976384e-01 -5.30914187e-01 1.02701437e+00 1.29646873e+00
6.54124081e-01 -4.73111391e-01 -6.69105768e-01 7.77976274e-01
-1.09164262e+00 -1.10023832e+00 -2.07971975e-01 2.29337883e+00
4.70050067e-01 1.89921752e-01 1.12730563e-01 7.12619603e-01
9.22235727e-01 5.32042265e-01 -2.36885682e-01 -7.61003911e-01
-3.84879649e-01 8.43051970e-02 3.66862357e-01 1.14514101e+00
-1.23122811e+00 1.01411140e+00 6.30302668e+00 7.79772699e-01
-1.34866393e+00 5.63528836e-01 3.17638107e-02 1.94268242e-01
-5.73437452e-01 -3.20604444e-01 -1.16930485e+00 7.54578114e-01
1.27168512e+00 1.27083272e-01 2.50278443e-01 8.44557762e-01
-1.16517164e-01 9.69745815e-01 -6.69457316e-01 1.04873621e+00
4.37358439e-01 -9.14376616e-01 7.98943564e-02 2.06109539e-01
4.20270115e-01 2.64062107e-01 3.57836932e-01 4.92251456e-01
2.56452680e-01 -9.11311924e-01 8.98439884e-01 -4.50488269e-01
8.24690223e-01 -5.90963781e-01 8.53454769e-01 8.22886676e-02
-1.19360232e+00 -3.11081201e-01 -1.99097648e-01 4.08522636e-01
2.59915411e-01 4.53780949e-01 -1.18193507e+00 6.71485186e-01
4.66914833e-01 2.81074107e-01 -3.91350478e-01 5.73123217e-01
-3.37924063e-01 1.05026746e+00 -3.33850503e-01 1.18180811e-02
1.26068324e-01 3.64228368e-01 8.56500030e-01 1.44557571e+00
1.93840891e-01 -5.67465663e-01 -2.55793452e-01 5.85880339e-01
-2.37719238e-01 6.62332550e-02 -6.74854636e-01 -1.35667250e-01
9.18567955e-01 9.01572049e-01 1.69733182e-01 -2.44247168e-01
-2.24473909e-01 1.17346287e+00 8.55384618e-02 3.66110593e-01
-9.81735706e-01 -4.52851921e-01 1.22187114e+00 -6.05694242e-02
6.27179444e-01 -1.19762689e-01 -1.81255266e-01 -1.39881825e+00
7.85396919e-02 -9.93345797e-01 1.37020782e-01 1.27495363e-01
-1.06502068e+00 9.18891847e-01 -5.87778032e-01 -1.12450898e+00
-4.68805134e-02 -3.94593984e-01 -7.19597399e-01 9.50620413e-01
-1.94658160e+00 -1.21706903e+00 -1.62210595e-02 5.13648927e-01
2.53862530e-01 -4.81038034e-01 9.53107595e-01 5.70493221e-01
-7.41797566e-01 1.47560215e+00 1.92879498e-01 3.68449688e-01
6.91524684e-01 -9.69524562e-01 7.85151958e-01 1.19276738e+00
1.12335414e-01 7.41787016e-01 5.96661329e-01 -3.27833593e-01
-1.49899292e+00 -1.00566280e+00 1.47420490e+00 -5.39093554e-01
4.75272298e-01 -9.16725755e-01 -9.21880007e-01 5.93818963e-01
2.48424590e-01 1.09964691e-01 8.94581974e-01 6.39122650e-02
-1.02061772e+00 -2.93409318e-01 -1.33781230e+00 2.43387297e-01
8.49298894e-01 -1.01233160e+00 -6.54890597e-01 -1.08081670e-02
1.15855944e+00 1.55354246e-01 -5.35491884e-01 4.18662399e-01
5.22241771e-01 -8.39280009e-01 1.18882000e+00 -7.02242434e-01
-1.40914306e-01 -3.51276428e-01 -6.58646762e-01 -1.09341216e+00
-2.18775630e-01 -9.41531420e-01 -2.92001277e-01 1.85228693e+00
3.87577027e-01 -9.19928968e-01 5.86659193e-01 -2.44928505e-02
-2.06456006e-01 -1.84635013e-01 -1.35169625e+00 -1.14976442e+00
1.78129196e-01 -5.86372316e-01 9.62268412e-01 1.14567983e+00
-2.11663261e-01 2.38529205e-01 -8.06793749e-01 7.52191722e-01
9.20974910e-01 -3.60899776e-01 8.29801977e-01 -9.16655660e-01
-1.55633166e-01 -4.41753209e-01 -5.91291189e-01 -1.05244863e+00
3.70656788e-01 -9.29370224e-01 -1.82421468e-02 -9.71393883e-01
-2.34338254e-01 -2.94106692e-01 -7.66542554e-01 2.33914018e-01
-3.76160890e-01 2.59656668e-01 2.56127238e-01 4.17912230e-02
-3.00285555e-02 8.36260855e-01 5.82752824e-01 -5.47897339e-01
-1.46209568e-01 7.43021630e-03 -7.29361475e-01 7.41126388e-02
7.17571855e-01 -5.16556263e-01 -3.34128410e-01 -5.05925596e-01
-5.69074750e-01 -1.47495270e-01 2.25272477e-01 -1.06591654e+00
8.46577436e-02 3.20994437e-01 -3.54724601e-02 -4.74251986e-01
5.69556296e-01 -7.34756649e-01 -1.07199967e-01 8.39884520e-01
-4.76531357e-01 -2.76524603e-01 1.36389285e-01 5.91787219e-01
-3.82667899e-01 9.30633917e-02 9.50944304e-01 5.79562545e-01
-3.53518546e-01 1.28595933e-01 -1.79398507e-01 -1.24946244e-01
7.70224810e-01 -1.09862268e-01 -5.13218462e-01 -3.00236434e-01
-3.95088732e-01 1.65757090e-01 2.88530111e-01 7.89052188e-01
7.88376451e-01 -1.48524952e+00 -1.11828113e+00 7.41570830e-01
2.65036106e-01 -7.20985174e-01 3.72365773e-01 6.28000140e-01
-1.74419448e-01 6.53165698e-01 1.45138532e-01 -4.82107341e-01
-1.53501821e+00 7.48815954e-01 1.60451174e-01 -3.35794687e-02
-4.06216770e-01 1.24905133e+00 4.43966500e-02 -9.50920284e-01
5.76524436e-01 1.08996592e-02 7.71214515e-02 2.14314694e-03
1.00155234e+00 1.99871987e-01 2.11560696e-01 -1.19443798e+00
-9.62436020e-01 1.59866005e-01 -1.78187713e-01 -2.48761073e-01
9.83838141e-01 -2.27370948e-01 3.92807245e-01 1.70548214e-03
1.68864238e+00 4.70733881e-01 -8.56900752e-01 -3.51012230e-01
-1.77982420e-01 -5.94175935e-01 8.09810311e-02 -7.32341409e-01
-9.35425818e-01 1.30677938e+00 9.70764756e-01 2.51538068e-01
7.10857809e-01 -2.53580779e-01 1.07842267e+00 -2.19568893e-01
4.17671688e-02 -7.50179589e-01 -1.14539899e-01 3.74450058e-01
8.09945524e-01 -1.43797863e+00 -5.65141320e-01 -4.68414932e-01
-7.01898158e-01 7.24540532e-01 2.61581838e-01 2.33296946e-01
8.06291759e-01 7.93089792e-02 4.70845073e-01 1.75069064e-01
-5.81928551e-01 1.51724890e-01 1.68016702e-01 7.90665925e-01
1.59108356e-01 3.23879540e-01 -2.71998849e-02 6.05448544e-01
-3.60292792e-01 -6.12792611e-01 2.29644522e-01 7.57255912e-01
-3.26350749e-01 -1.23896456e+00 -5.14926255e-01 -5.48294820e-02
-6.21184230e-01 -1.81955978e-01 -4.69163746e-01 3.24921072e-01
1.67978659e-01 1.37530482e+00 -1.78535536e-01 -1.08754194e+00
1.82742611e-01 1.91928983e-01 -1.12244770e-01 -3.23383749e-01
-8.11698556e-01 -1.13121167e-01 1.47964895e-01 -3.98949504e-01
-1.42052576e-01 -4.39631671e-01 -5.38447380e-01 -5.21781445e-01
-6.27178013e-01 4.60016608e-01 1.11338580e+00 5.23435235e-01
4.67854440e-01 2.87336439e-01 1.19015312e+00 -4.88966197e-01
-1.27006912e+00 -1.04718184e+00 -5.19998848e-01 6.58664107e-01
1.11066461e+00 -4.72145855e-01 -9.98521447e-01 -3.40389431e-01] | [14.095584869384766, 5.8846001625061035] |
c04e2080-945b-4308-8254-e83930743185 | attentive-neural-network-for-named-entity | 1810.13097 | null | https://arxiv.org/abs/1810.13097v2 | https://arxiv.org/pdf/1810.13097v2.pdf | Attentive Neural Network for Named Entity Recognition in Vietnamese | We propose an attentive neural network for the task of named entity recognition in Vietnamese. The proposed attentive neural model makes use of character-based language models and word embeddings to encode words as vector representations. A neural network architecture of encoder, attention, and decoder layers is then utilized to encode knowledge of input sentences and to label entity tags. The experimental results show that the proposed attentive neural network achieves the state-of-the-art results on the benchmark named entity recognition datasets in Vietnamese in comparison to both hand-crafted features based models and neural models. | ['Cam-Tu Nguyen', 'Kim Anh Nguyen', 'Ngan Dong'] | 2018-10-31 | null | null | null | null | ['named-entity-recognition-in-vietnamese'] | ['natural-language-processing'] | [-1.82377413e-01 1.01129532e-01 -1.51781529e-01 -7.57028222e-01
-2.62779653e-01 -1.17398545e-01 5.80761969e-01 6.51111752e-02
-1.19261110e+00 7.12276161e-01 6.02095127e-01 -1.91888392e-01
4.25087839e-01 -8.94427001e-01 -4.63179916e-01 -2.70175159e-01
-3.01707357e-01 3.70654553e-01 -5.51665016e-02 -2.27550834e-01
2.77906120e-01 6.11579895e-01 -9.43646550e-01 1.01655960e-01
7.53469706e-01 7.98866391e-01 -4.92822714e-02 5.76562285e-01
-5.71880162e-01 1.18007624e+00 -7.34510481e-01 -7.09125698e-01
-1.31302863e-01 -1.07884251e-01 -9.16352630e-01 -2.44755536e-01
5.52264601e-02 -3.00002843e-01 -7.70637274e-01 9.38327730e-01
6.67228103e-01 6.07781768e-01 6.69220090e-01 -6.68375731e-01
-1.82824457e+00 7.87068784e-01 7.61087835e-02 6.53345525e-01
-2.86758356e-02 -2.23291039e-01 1.12264836e+00 -1.43651390e+00
6.13385856e-01 9.77527618e-01 5.49750507e-01 6.41713738e-01
-5.37167549e-01 -4.84761447e-01 2.23012209e-01 5.89560628e-01
-1.74667990e+00 -4.20477152e-01 4.42151934e-01 -2.95569122e-01
1.74083841e+00 -4.01177824e-01 4.57566530e-01 1.11195469e+00
3.77067357e-01 9.55448151e-01 3.34712446e-01 -3.08957249e-01
2.53529042e-01 1.99827507e-01 6.50340855e-01 7.42281437e-01
3.35062563e-01 2.18590945e-01 -3.08940440e-01 9.84247774e-02
5.29991031e-01 5.41439652e-02 8.18731040e-02 1.11621417e-01
-9.84166384e-01 1.08040607e+00 1.02368784e+00 5.41919708e-01
-8.88893545e-01 1.14978351e-01 7.14411259e-01 -9.41375736e-03
5.32309830e-01 6.70665145e-01 -5.64964533e-01 -1.86515287e-01
-6.36042297e-01 -4.46357608e-01 7.93839335e-01 1.13262963e+00
5.89274764e-01 8.56540203e-01 -4.93700236e-01 9.34118092e-01
2.86672205e-01 5.70411921e-01 8.48150134e-01 -1.55766532e-01
4.54651266e-01 7.06960559e-01 -3.90515365e-02 -8.47974896e-01
-2.28170544e-01 -3.21002811e-01 -7.81664789e-01 -3.93142432e-01
-5.42152584e-01 -6.10816419e-01 -1.28970182e+00 1.41852117e+00
-7.92518631e-02 1.87265754e-01 5.32506824e-01 8.02180529e-01
1.29948759e+00 1.26329172e+00 7.31563747e-01 1.67291127e-02
1.28889072e+00 -1.20899498e+00 -1.11221576e+00 -4.41818178e-01
4.80504900e-01 -3.33677411e-01 3.99867117e-01 -4.06213701e-01
-7.96044886e-01 -7.76011944e-01 -9.81178820e-01 -3.97773892e-01
-9.97558355e-01 2.60770530e-01 4.37706083e-01 8.32938254e-01
-7.43567407e-01 1.45978123e-01 -8.94663215e-01 -4.29301977e-01
5.14892161e-01 4.65533972e-01 -4.12944674e-01 9.40298960e-02
-1.64196610e+00 1.15649021e+00 1.13074815e+00 4.23400879e-01
-9.78755951e-01 -3.50311071e-01 -1.41906464e+00 5.86823821e-01
-3.27457115e-02 -3.57529283e-01 1.05601037e+00 -7.06736147e-01
-1.57654381e+00 6.32229149e-01 -1.71178460e-01 -4.91156548e-01
-4.14350480e-01 -4.51856494e-01 -8.84911597e-01 -8.92723352e-02
-2.41492569e-01 7.56480873e-01 2.33342826e-01 -7.73328960e-01
-5.56706131e-01 -1.22793213e-01 -1.68519676e-01 1.06636308e-01
-9.05645311e-01 2.09043264e-01 -2.67482638e-01 -6.36455536e-01
-5.03548682e-01 -4.64903742e-01 -3.08747172e-01 -5.01528800e-01
-2.27718800e-01 -5.90839446e-01 7.78833687e-01 -7.41192758e-01
1.34576678e+00 -2.11958575e+00 -1.23493612e-01 -1.06442600e-01
1.10623457e-01 8.47787023e-01 -5.25832891e-01 4.93329525e-01
-3.31422985e-02 1.13675907e-01 4.57084775e-02 -1.72697991e-01
1.25844300e-01 3.87631863e-01 -1.59601793e-01 4.49629486e-01
6.00096226e-01 1.24352300e+00 -9.26215947e-01 -5.57852387e-01
2.06450030e-01 9.01459754e-01 -3.68034124e-01 8.31854343e-01
1.57609686e-01 -3.21584463e-01 -4.52929139e-01 3.96262139e-01
5.82224250e-01 -1.57402381e-02 3.19393486e-01 2.02539470e-02
-2.04438061e-01 5.60983181e-01 -6.91818297e-01 1.54555452e+00
-5.80940723e-01 8.50793421e-01 -3.89199138e-01 -8.81984055e-01
1.38571727e+00 7.29468226e-01 -1.91299960e-01 -6.15810454e-01
6.27229214e-01 1.93045698e-02 -4.17370014e-02 -6.02587819e-01
8.71249974e-01 -2.40432918e-01 -3.03010464e-01 1.93524763e-01
8.28179955e-01 4.48196143e-01 9.75373983e-02 -4.27602157e-02
1.02393281e+00 -2.05894500e-01 8.38886440e-01 -1.36838391e-01
6.85236156e-01 -4.82214503e-02 7.98967242e-01 4.96035159e-01
-4.74105597e-01 2.21824065e-01 1.41519055e-01 -6.38017535e-01
-1.13273597e+00 -9.24930990e-01 -2.03573778e-02 1.63122582e+00
-1.65046081e-01 -2.10251376e-01 -5.95852554e-01 -8.39245915e-01
-2.86377758e-01 1.15261066e+00 -9.67637122e-01 -2.88731843e-01
-7.11043775e-01 -3.45464408e-01 8.55164468e-01 1.16600907e+00
6.22084796e-01 -1.65930200e+00 -3.59439790e-01 4.53458101e-01
3.74831498e-01 -1.12763250e+00 -4.90662545e-01 5.07494211e-01
-5.17342567e-01 -5.76348007e-01 -7.68712819e-01 -1.33427870e+00
6.45584106e-01 -2.41334349e-01 9.42395329e-01 -2.92279512e-01
-2.54236050e-02 1.00080006e-01 -6.82342768e-01 -4.14435983e-01
-1.34176165e-01 5.15566826e-01 -1.45350456e-01 1.40756637e-01
8.74683082e-01 -1.38993517e-01 -3.60913873e-01 -1.89389259e-01
-9.07995164e-01 -6.79401875e-01 8.55964780e-01 1.10211825e+00
1.87606826e-01 -4.73296434e-01 6.24731660e-01 -1.14183545e+00
8.04652095e-01 -8.75199616e-01 -4.83863711e-01 4.73653883e-01
-6.26203716e-01 3.60839605e-01 8.11829984e-01 -4.24216092e-01
-1.26139367e+00 -8.55585486e-02 -6.25776768e-01 -1.10699840e-01
-3.48040432e-01 7.30812609e-01 -1.86383784e-01 1.77413136e-01
2.89583296e-01 4.32352334e-01 -5.80953419e-01 -5.04414856e-01
7.14060009e-01 1.19977820e+00 5.35787165e-01 -8.97970498e-02
2.69776851e-01 -2.24796429e-01 -7.60104537e-01 -9.12453771e-01
-8.22422683e-01 -4.20556754e-01 -1.00444710e+00 5.54846898e-02
1.37482154e+00 -8.38111937e-01 -3.85698229e-01 1.82004571e-01
-1.58185601e+00 1.39206752e-01 -3.50016266e-01 6.01893425e-01
-1.68969974e-01 5.78707531e-02 -9.95564878e-01 -8.05979311e-01
-6.82220161e-01 -8.11450779e-01 6.31918550e-01 5.52808106e-01
-2.62762401e-02 -1.42004859e+00 2.94056535e-01 -2.04444546e-02
6.35960698e-01 -2.14186877e-01 1.05698037e+00 -1.59007704e+00
-8.49560648e-02 -5.61375320e-01 -2.26618826e-01 5.23921430e-01
-3.77481766e-02 -2.65278518e-01 -9.25241411e-01 -2.00547367e-01
-3.57533872e-01 -3.55200887e-01 8.22052717e-01 -1.66217640e-01
7.55260527e-01 -4.27315474e-01 -1.93026945e-01 4.70553219e-01
1.55726194e+00 7.42233276e-01 7.78686762e-01 2.04290450e-01
6.69425547e-01 2.00402975e-01 1.91980004e-01 5.24366438e-01
4.68568772e-01 1.52244031e-01 2.95027256e-01 -2.18940094e-01
1.65479168e-01 -3.25354159e-01 2.19240487e-01 1.19320846e+00
3.38298827e-02 -7.42166936e-01 -1.11943257e+00 9.92154241e-01
-1.47806752e+00 -8.98647904e-01 2.56665111e-01 1.73174918e+00
4.90979433e-01 -1.03930473e-01 -2.69444257e-01 -4.34529752e-01
1.02111173e+00 5.06852686e-01 -3.86886984e-01 -1.11257315e+00
9.25803930e-02 6.17102265e-01 4.72316533e-01 2.97442287e-01
-1.31786692e+00 1.39217556e+00 7.02557421e+00 4.47043508e-01
-1.23981154e+00 5.44941485e-01 2.01333866e-01 3.94309640e-01
9.29244906e-02 -5.52732944e-01 -8.89430821e-01 1.69796631e-01
1.66227388e+00 -3.03788722e-01 -1.51598871e-01 1.17991340e+00
-2.55318671e-01 7.52950728e-01 -8.60927761e-01 5.99261284e-01
4.59282577e-01 -1.23918045e+00 2.29456365e-01 -1.63244396e-01
7.71487296e-01 4.21910912e-01 -9.88778919e-02 8.39407980e-01
4.80404615e-01 -1.09565556e+00 4.96427447e-01 5.67935288e-01
5.67588210e-01 -8.93657804e-01 1.41459548e+00 5.61085995e-03
-1.15120018e+00 -1.05681255e-01 -8.23964477e-01 5.54255582e-02
3.57090503e-01 1.89194940e-02 -1.13117313e+00 2.17951760e-01
2.06603870e-01 6.30456328e-01 -4.15477872e-01 1.05397058e+00
-4.90500808e-01 9.08990264e-01 1.95494682e-01 -6.80153966e-01
6.42942131e-01 3.65375638e-01 -9.87664098e-04 1.79214144e+00
8.52238759e-02 4.25655127e-01 1.96397100e-02 7.21586466e-01
-4.41790760e-01 4.64823782e-01 -7.16592729e-01 -7.13208497e-01
6.44776940e-01 8.77521992e-01 -3.44975263e-01 -6.88294649e-01
-6.80296302e-01 1.22221804e+00 9.70223665e-01 3.92336249e-01
-6.42395616e-01 -1.23740327e+00 7.24319339e-01 -4.13854897e-01
1.13595366e+00 -5.33852994e-01 -5.10361493e-02 -1.14007688e+00
-5.88801622e-01 -2.93822974e-01 3.82686287e-01 -5.59834182e-01
-1.36832309e+00 1.28640139e+00 -3.18620443e-01 -7.91987181e-01
-3.64376843e-01 -1.09446502e+00 -7.08655596e-01 6.55688822e-01
-1.45130813e+00 -8.73148382e-01 -3.33339199e-02 3.87114674e-01
6.61781192e-01 -5.52561700e-01 1.40320754e+00 3.75657678e-01
-9.00340557e-01 7.81907737e-01 3.68023425e-01 1.13158584e+00
3.13290566e-01 -1.03094840e+00 8.21618736e-01 9.24865186e-01
2.91689396e-01 9.14412677e-01 -2.70103458e-02 -6.04018688e-01
-1.37071204e+00 -1.38320720e+00 1.71751165e+00 -1.56218156e-01
5.93213439e-01 -4.22956616e-01 -8.84598553e-01 9.39992070e-01
8.33859026e-01 3.80073190e-01 1.22258413e+00 2.30204821e-01
-4.47066128e-01 7.13682547e-02 -9.51245308e-01 8.23438242e-02
7.11096048e-01 -6.57304406e-01 -1.25199592e+00 4.04331833e-02
9.15463746e-01 -1.70574766e-02 -1.10912955e+00 3.05446666e-02
3.57973546e-01 -3.84647876e-01 7.91307628e-01 -1.26039922e+00
1.75366476e-01 1.38070449e-01 -1.56585887e-01 -1.45434105e+00
-8.49732459e-01 7.33029991e-02 -7.12498277e-02 1.40654898e+00
8.16916525e-01 -4.07883406e-01 4.55686897e-01 4.51062620e-01
-2.41083249e-01 -5.29266477e-01 -1.02123070e+00 -6.74328327e-01
1.86748356e-01 -1.67114377e-01 5.89658558e-01 1.09269404e+00
-6.50927648e-02 8.75203729e-01 -3.47624242e-01 1.88242093e-01
3.46662328e-02 -3.42251182e-01 1.98375434e-02 -8.28501582e-01
4.43246782e-01 -3.45788896e-02 -9.24769104e-01 -1.03796899e+00
7.94342160e-01 -9.43923295e-01 3.56769770e-01 -1.73137832e+00
1.62579522e-01 -1.76048037e-02 -8.41165304e-01 7.19762623e-01
-4.28543925e-01 1.11350007e-01 2.74587214e-01 -1.74749643e-01
-8.38862658e-01 1.01057458e+00 6.45360708e-01 -3.28934520e-01
5.69998547e-02 -4.66136664e-01 -5.14105797e-01 2.37175301e-01
6.87033653e-01 -5.61873972e-01 -1.95172727e-02 -9.68659878e-01
-1.07302889e-01 -2.05805883e-01 -1.45287722e-01 -1.00532818e+00
2.83551365e-01 -4.87674773e-02 4.56651270e-01 -4.29230571e-01
2.37458751e-01 -7.51009643e-01 -4.80510980e-01 5.77515185e-01
-7.88033426e-01 3.67202014e-01 7.35189989e-02 3.43583763e-01
-4.84080821e-01 -4.50925469e-01 5.63722908e-01 -9.75647345e-02
-1.45802128e+00 3.66507709e-01 -8.89287114e-01 1.30734265e-01
8.40386748e-01 1.25143439e-01 -3.75596076e-01 -1.62328854e-01
-6.96348667e-01 5.51824085e-02 -2.88826108e-01 7.15095103e-01
1.01180005e+00 -1.53780520e+00 -8.82625401e-01 3.74945730e-01
2.99764633e-01 -7.46130109e-01 7.42548704e-02 1.10937275e-01
-6.85998797e-01 8.67577970e-01 -5.46609640e-01 1.00666530e-01
-7.12702692e-01 8.88917267e-01 1.95409507e-01 -2.80031115e-01
-1.95088118e-01 1.11047101e+00 -5.08970469e-02 -6.59463704e-01
2.45304957e-01 -1.76315263e-01 -7.05192566e-01 3.97463851e-02
6.81042671e-01 3.86984289e-01 -3.27993445e-02 -9.22104955e-01
-6.86559319e-01 1.10291861e-01 -2.73465961e-01 1.96807474e-01
1.39488924e+00 1.18906081e-01 1.20530643e-01 3.78799438e-01
1.72668660e+00 -3.95954490e-01 -5.34602523e-01 -4.71211165e-01
9.25111696e-02 -1.10189952e-01 2.59583682e-01 -6.65514827e-01
-1.07543862e+00 1.28672588e+00 8.29961479e-01 -1.08964190e-01
7.70749629e-01 -2.89824754e-01 1.09094632e+00 9.58039522e-01
8.15947726e-02 -1.12360132e+00 -1.38297543e-01 1.24527299e+00
6.30223095e-01 -1.29170728e+00 -2.83808857e-01 -7.03884289e-02
-8.81833315e-01 1.26190758e+00 8.55407059e-01 -4.97250080e-01
8.29874516e-01 2.22222015e-01 3.60200852e-01 -1.04840621e-01
-9.22869265e-01 -4.84178960e-01 4.32257116e-01 7.46368289e-01
7.91981637e-01 1.20901398e-01 -5.12463033e-01 8.58180165e-01
-2.76434217e-02 -1.31767109e-01 2.27914795e-01 1.11719716e+00
-6.88255310e-01 -8.82403672e-01 5.73654380e-03 1.79098904e-01
-5.24626970e-01 -4.83068109e-01 -4.42451984e-01 7.16828406e-01
1.01996824e-01 8.51456225e-01 4.21921432e-01 -5.21432698e-01
4.76732552e-01 3.98921937e-01 -8.43591988e-02 -1.03838944e+00
-9.90717590e-01 -5.16656697e-01 2.72998095e-01 -1.96496129e-01
-2.32035205e-01 -6.41654804e-02 -1.36133051e+00 -1.07060678e-01
-4.24902380e-01 6.39836311e-01 7.60841548e-01 9.01893735e-01
5.99258602e-01 5.91015816e-01 6.24674797e-01 -5.25419891e-01
-5.02456427e-01 -1.31918228e+00 -6.36491835e-01 2.36729220e-01
1.60524815e-01 -4.59023356e-01 7.25310519e-02 -3.37857865e-02] | [9.784268379211426, 9.607739448547363] |
e49b33a3-1c3e-4c41-ac31-a370a36e6f6e | enriching-existing-conversational-emotion | 1912.00819 | null | https://arxiv.org/abs/1912.00819v3 | https://arxiv.org/pdf/1912.00819v3.pdf | EDA: Enriching Emotional Dialogue Acts using an Ensemble of Neural Annotators | The recognition of emotion and dialogue acts enriches conversational analysis and help to build natural dialogue systems. Emotion interpretation makes us understand feelings and dialogue acts reflect the intentions and performative functions in the utterances. However, most of the textual and multi-modal conversational emotion corpora contain only emotion labels but not dialogue acts. To address this problem, we propose to use a pool of various recurrent neural models trained on a dialogue act corpus, with and without context. These neural models annotate the emotion corpora with dialogue act labels, and an ensemble annotator extracts the final dialogue act label. We annotated two accessible multi-modal emotion corpora: IEMOCAP and MELD. We analyzed the co-occurrence of emotion and dialogue act labels and discovered specific relations. For example, Accept/Agree dialogue acts often occur with the Joy emotion, Apology with Sadness, and Thanking with Joy. We make the Emotional Dialogue Acts (EDA) corpus publicly available to the research community for further study and analysis. | ['Chandrakant Bothe', 'Cornelius Weber', 'Sven Magg', 'Stefan Wermter'] | 2019-12-02 | eda-enriching-emotional-dialogue-acts-using | https://aclanthology.org/2020.lrec-1.78 | https://aclanthology.org/2020.lrec-1.78.pdf | lrec-2020-5 | ['emotional-dialogue-acts', 'dialogue-act-classification'] | ['natural-language-processing', 'natural-language-processing'] | [ 6.18295260e-02 5.57715535e-01 1.00077894e-02 -8.27758253e-01
-6.25640675e-02 -6.83087885e-01 8.68548334e-01 1.72042698e-01
-4.00626808e-01 8.76030684e-01 9.59465981e-01 1.75380707e-01
4.02244389e-01 -5.77473402e-01 2.84525841e-01 -4.03977871e-01
1.30136624e-01 6.19256258e-01 -5.16456306e-01 -7.16215253e-01
2.63768464e-01 4.59117778e-02 -1.25617838e+00 8.28422666e-01
5.36836028e-01 1.01924729e+00 -3.70077312e-01 7.96971381e-01
-6.24150634e-01 1.58215058e+00 -8.20304096e-01 -6.93153203e-01
-3.18851411e-01 -9.02397573e-01 -1.38825452e+00 1.41944543e-01
-6.95050776e-01 -1.64006948e-01 7.77578652e-02 7.77161539e-01
3.74792308e-01 4.44496483e-01 6.83404803e-01 -1.34430170e+00
-2.20506445e-01 8.40013146e-01 -1.60177355e-04 -1.29187465e-01
7.40016162e-01 -7.68042281e-02 1.10589457e+00 -5.39557755e-01
8.41364682e-01 1.20624149e+00 3.96420717e-01 1.14824212e+00
-8.29147875e-01 -3.68100315e-01 -1.03760235e-01 8.77105668e-02
-6.61862373e-01 -6.94434702e-01 1.01261711e+00 -3.21558535e-01
1.45683217e+00 3.60541344e-01 8.18906903e-01 1.52311063e+00
-1.77593291e-01 1.06458449e+00 1.07442880e+00 -5.03925443e-01
6.04860894e-02 3.39265555e-01 6.12953126e-01 4.96517658e-01
-1.08017361e+00 -3.49108905e-01 -4.99057561e-01 -3.90860260e-01
1.84642091e-01 -1.21187091e-01 -2.09337011e-01 4.49682653e-01
-9.76047635e-01 1.04972792e+00 2.25437852e-03 6.86258674e-01
-6.81624830e-01 -3.58812064e-01 1.13715971e+00 6.94307804e-01
8.09397578e-01 7.15789497e-01 -6.32557154e-01 -1.03057611e+00
-1.76404327e-01 3.50335725e-02 1.45024657e+00 3.76134038e-01
8.24955106e-01 -1.92836255e-01 -5.67905046e-03 1.54429412e+00
2.11222656e-02 -3.18498239e-02 6.56332254e-01 -1.04152191e+00
6.38160631e-02 1.01105952e+00 5.72484918e-02 -1.03520811e+00
-9.61635411e-01 5.61943829e-01 -8.53474438e-01 -2.48650104e-01
2.76634961e-01 -7.36675918e-01 -8.15809444e-02 1.64991462e+00
3.83727729e-01 -2.52708882e-01 7.12689638e-01 7.37186372e-01
1.29867518e+00 7.86190093e-01 2.73328662e-01 -6.11207962e-01
1.56600273e+00 -8.30756664e-01 -1.42403769e+00 -1.98880315e-01
9.98959124e-01 -6.49818838e-01 9.75015700e-01 2.36807242e-01
-9.21353519e-01 -2.49113485e-01 -4.99551624e-01 -1.36866467e-02
-5.72766066e-01 5.46390451e-02 6.38429224e-01 3.88069689e-01
-5.07715464e-01 3.38626742e-01 -4.33638334e-01 -3.41254175e-01
-1.46522090e-01 8.28335658e-02 -4.22023386e-01 5.84690571e-01
-1.62210393e+00 1.31833112e+00 4.42862749e-01 1.76052943e-01
-1.69718891e-01 -1.78105626e-02 -1.13565016e+00 -1.83457479e-01
2.78683484e-01 3.82837318e-02 1.45867383e+00 -1.49355435e+00
-2.08365726e+00 1.19050705e+00 -2.62040228e-01 -2.72799134e-01
-8.33088085e-02 -2.20699161e-01 -6.85989738e-01 -8.44141021e-02
-3.98893446e-01 5.13262391e-01 3.19811165e-01 -8.53688359e-01
-3.96988600e-01 -4.09085065e-01 1.76909328e-01 5.20568490e-01
-5.08330107e-01 5.79707742e-01 -1.04857478e-02 -7.78262764e-02
-3.50535959e-01 -7.95053840e-01 -2.84909289e-02 -7.99688637e-01
-4.71833020e-01 -8.89417887e-01 7.49653935e-01 -4.66584980e-01
1.25144064e+00 -2.25471830e+00 3.05428535e-01 -1.36941969e-01
2.96482354e-01 -1.13078784e-02 1.26875779e-02 7.13242829e-01
-2.86682788e-02 6.66486099e-03 4.58555035e-02 -3.32962632e-01
3.61398637e-01 5.25207460e-01 -2.30754763e-01 1.37345031e-01
1.97764680e-01 8.03091645e-01 -8.36047709e-01 -5.12862504e-01
3.50552857e-01 3.75736833e-01 -4.03654665e-01 8.66843820e-01
-3.34361374e-01 6.08391345e-01 -5.81910372e-01 3.02462071e-01
-1.23263389e-01 -1.13214709e-01 5.11260808e-01 -2.45828852e-01
-1.91235095e-01 7.37440765e-01 -5.72746038e-01 1.32679069e+00
-7.47716844e-01 7.93922186e-01 2.01707095e-01 -8.18155348e-01
1.29877067e+00 8.82323325e-01 5.29062390e-01 -5.48075616e-01
6.16216302e-01 -4.32871701e-03 1.69408754e-01 -9.35400963e-01
7.31264293e-01 -5.12345731e-01 -6.98329747e-01 1.06242704e+00
2.07040235e-01 -2.41911680e-01 1.55623674e-01 8.39227960e-02
1.00092232e+00 -1.35581926e-01 6.89028263e-01 3.26043993e-01
6.68321729e-01 -1.27913341e-01 5.28423309e-01 7.46010467e-02
-4.90974247e-01 -1.34819448e-01 9.34201121e-01 -8.34304512e-01
-6.25499010e-01 -2.64420599e-01 8.56536701e-02 1.59977567e+00
-2.10143849e-01 -6.25222623e-01 -4.58887041e-01 -6.42190993e-01
-7.15360820e-01 9.21895981e-01 -6.39582872e-01 -7.02525303e-02
-4.27033812e-01 -4.67315972e-01 8.52668822e-01 8.64704847e-02
5.67898035e-01 -1.79315281e+00 -5.72822690e-01 4.26331133e-01
-7.45897770e-01 -1.14360845e+00 2.23773755e-02 4.57241327e-01
-4.59884182e-02 -1.00532043e+00 -1.52308956e-01 -5.57900429e-01
2.00155288e-01 -5.81167102e-01 1.30227876e+00 -1.55276502e-03
8.61060712e-03 4.04295385e-01 -8.90391171e-01 -4.57048982e-01
-9.18578625e-01 -5.77326901e-02 -2.80500781e-02 1.03116833e-01
8.13119471e-01 -4.09740120e-01 3.50249931e-03 2.70073473e-01
-6.63985074e-01 1.90276876e-01 -2.51620412e-01 9.53479052e-01
-3.20696682e-01 -2.71294415e-01 6.92693472e-01 -9.95746017e-01
1.39651954e+00 -6.58127010e-01 2.46459410e-01 9.93341357e-02
-5.23343012e-02 -1.27358422e-01 5.89614093e-01 -3.36036861e-01
-1.44767785e+00 -3.29188742e-02 -5.82809210e-01 5.81458695e-02
-5.96498966e-01 5.74668348e-01 -4.09684628e-02 5.71559668e-01
5.45615196e-01 -7.02307522e-02 -4.25664447e-02 -2.63768137e-02
4.29443747e-01 1.16551137e+00 2.88390398e-01 -7.25244701e-01
-2.49543533e-01 -7.46402144e-03 -6.80229425e-01 -1.07158887e+00
-8.68937254e-01 -5.34586549e-01 -3.85219365e-01 -8.25188398e-01
1.11468959e+00 -5.90596437e-01 -1.32605541e+00 4.40537333e-01
-1.44251060e+00 -3.94535810e-01 -9.41434875e-02 3.87908190e-01
-7.54746616e-01 2.89103478e-01 -1.14136326e+00 -1.37676561e+00
-5.15089512e-01 -8.83269966e-01 7.13588297e-01 3.36246669e-01
-1.19748545e+00 -1.12042165e+00 3.17130327e-01 4.06598032e-01
3.87061745e-01 3.52900684e-01 8.59208941e-01 -1.30518806e+00
7.68615961e-01 -1.52296558e-01 1.80528700e-01 1.83227733e-01
2.76411444e-01 -4.41074967e-02 -1.02549207e+00 4.90797311e-01
2.65449435e-01 -1.33326530e+00 2.77665228e-01 -1.18857414e-01
6.91908419e-01 -6.86438978e-01 1.30516723e-01 -3.82704958e-02
5.69177628e-01 5.27336061e-01 5.50380290e-01 -1.80094726e-02
1.78343117e-01 1.31730068e+00 5.18167794e-01 8.53664458e-01
4.28242952e-01 6.38520062e-01 1.49959043e-01 7.42379799e-02
7.10059464e-01 1.68564618e-01 5.35670877e-01 1.04927611e+00
-2.37674162e-01 -4.33804482e-01 -9.34904099e-01 3.75686616e-01
-1.96340752e+00 -1.16272032e+00 -1.14868298e-01 1.18008018e+00
1.32385242e+00 -1.52423799e-01 1.13595754e-01 -7.25566447e-02
4.98553872e-01 6.34444237e-01 -3.43871385e-01 -1.12293470e+00
-1.08591020e-01 2.56593022e-02 -4.09085959e-01 6.22803569e-01
-1.08029854e+00 1.15321469e+00 5.71026659e+00 3.16716611e-01
-9.19254839e-01 5.03957905e-02 7.33341038e-01 -1.01744449e-02
-1.18656112e-02 -3.70187312e-01 -4.50665146e-01 9.71105248e-02
1.29257381e+00 -1.23143382e-01 5.50168574e-01 9.17780221e-01
7.07480535e-02 7.16950446e-02 -1.14630234e+00 8.23498189e-01
1.46592930e-01 -1.07863641e+00 -6.21585190e-01 -2.40687251e-01
1.96907640e-01 -9.75470319e-02 -6.55870080e-01 8.98835599e-01
6.31129920e-01 -9.65173721e-01 -3.02055571e-02 6.49975419e-01
3.18143487e-01 -7.52093434e-01 9.18782115e-01 5.13611495e-01
-7.22262502e-01 2.24899352e-01 1.82548463e-02 -5.99277377e-01
4.91549820e-01 1.99595675e-01 -9.86491621e-01 1.72651276e-01
3.91633123e-01 8.72216940e-01 1.07565865e-01 -4.95468527e-02
-3.46737146e-01 5.31078398e-01 -2.44621411e-01 -6.43399477e-01
2.12049082e-01 -4.40099716e-01 5.14860392e-01 1.46369135e+00
-3.31378996e-01 6.22618377e-01 2.86351621e-01 6.00948811e-01
-2.12019876e-01 4.29309100e-01 -6.42683685e-01 -7.58619308e-01
2.78508991e-01 1.59824598e+00 -5.69033921e-01 -4.28719521e-01
-3.38406712e-01 1.13278568e+00 4.42322314e-01 1.56058833e-01
-5.87742269e-01 -2.07541406e-01 8.81058097e-01 -7.66874433e-01
-4.82799709e-01 1.35777876e-01 8.40421095e-02 -1.18106127e+00
-4.66719478e-01 -1.01875758e+00 4.02295828e-01 -7.80664921e-01
-1.58836544e+00 1.02808237e+00 -2.86110878e-01 -6.75620079e-01
-9.53371763e-01 -5.22928119e-01 -8.98149908e-01 6.63362563e-01
-8.43734086e-01 -9.21868980e-01 -9.39016789e-02 5.62299430e-01
7.75522292e-01 -1.55854836e-01 1.59413040e+00 1.38951525e-01
-6.02230489e-01 1.56141371e-01 -5.67508996e-01 6.68990433e-01
8.03780377e-01 -1.08316171e+00 -1.29794165e-01 -1.37030169e-01
-1.96677685e-01 4.69477385e-01 9.10075903e-01 -3.15416157e-01
-9.63191271e-01 -4.89611477e-01 1.22527492e+00 -3.07714343e-01
8.84299278e-01 -2.38663033e-01 -9.32735622e-01 6.71772838e-01
9.01100397e-01 -5.98902285e-01 1.48894334e+00 7.33765602e-01
-1.38733640e-01 6.42909467e-01 -9.05026019e-01 6.12297952e-01
5.85468709e-01 -8.55635762e-01 -1.15433669e+00 3.66456389e-01
6.52803183e-01 -3.99871737e-01 -1.00698805e+00 2.16450155e-01
5.57462931e-01 -1.02221727e+00 4.95737463e-01 -1.03357553e+00
8.25435221e-01 2.79712915e-01 -1.19333811e-01 -1.44014680e+00
3.93506140e-01 -6.98389530e-01 5.48000634e-02 1.43493891e+00
4.59021956e-01 -3.79065633e-01 3.13475400e-01 1.19333410e+00
-2.32523620e-01 -4.69594419e-01 -7.98247516e-01 9.80760828e-02
-1.45447195e-01 -5.27477801e-01 4.59555477e-01 1.62843847e+00
1.21984589e+00 1.00837946e+00 -7.55398512e-01 -6.57570541e-01
-2.80201942e-01 1.84990302e-01 8.12970817e-01 -1.24789214e+00
-1.31306753e-01 -6.20713294e-01 -2.32410431e-02 -7.95028746e-01
8.32723856e-01 -6.69646680e-01 2.71238238e-01 -1.29586983e+00
-1.24233380e-01 -3.42673026e-02 1.93358213e-01 8.97380650e-01
-6.58559725e-02 -9.52281505e-02 -5.70158698e-02 1.85744930e-02
-8.22598159e-01 9.36808348e-01 8.96584570e-01 -4.28768955e-02
-5.03233850e-01 -1.72660783e-01 -3.55979800e-01 1.11575580e+00
9.92208779e-01 -1.74610719e-01 6.26303703e-02 3.03851724e-01
4.86835837e-01 4.69406217e-01 -1.73888519e-01 -3.06488663e-01
-1.03689292e-02 -4.23190564e-01 2.21866556e-02 -3.56282651e-01
8.07648182e-01 -7.12030411e-01 -1.81173131e-01 1.48317039e-01
-1.06929541e+00 -1.49766862e-01 3.70508507e-02 4.66615185e-02
-5.61255991e-01 -3.12884510e-01 6.72844291e-01 -2.19634145e-01
-6.58677101e-01 -3.08284074e-01 -1.08025742e+00 2.01613046e-02
9.34371293e-01 1.16253555e-01 -4.37405854e-01 -9.41911221e-01
-1.12257767e+00 3.88113350e-01 8.21029991e-02 7.00677574e-01
5.19893348e-01 -1.17057574e+00 -7.47934818e-01 -3.82650346e-02
2.30238259e-01 -4.27884191e-01 3.07530791e-01 5.87818146e-01
5.44049479e-02 8.50913525e-02 -4.49689031e-01 -8.53442475e-02
-1.49255323e+00 9.50260088e-02 4.69961911e-01 -3.97126734e-01
-2.97267199e-01 7.75734484e-01 -1.23458132e-01 -1.04935193e+00
2.68986076e-01 1.20925583e-01 -9.11918581e-01 5.78261197e-01
6.38214886e-01 8.20925236e-02 -4.86978263e-01 -1.02307463e+00
-1.31075129e-01 -1.49407670e-01 4.36943881e-02 -2.48648435e-01
1.36472929e+00 -1.43105894e-01 -6.11330211e-01 1.08442569e+00
1.23290265e+00 -1.10915951e-01 -5.51268458e-01 -1.38768375e-01
1.30662426e-01 8.31344649e-02 -2.36648768e-01 -9.22433317e-01
-6.22570574e-01 6.86335325e-01 -3.23762074e-02 9.73121226e-01
9.59906995e-01 5.14878519e-02 7.90906608e-01 6.51648819e-01
3.82349566e-02 -1.48324001e+00 2.74255395e-01 1.29958761e+00
1.03512073e+00 -1.36820674e+00 -4.45740283e-01 -1.08104303e-01
-1.40294588e+00 1.38297391e+00 1.01419377e+00 2.02490121e-01
5.06205499e-01 2.18858123e-01 6.36433780e-01 -6.76595747e-01
-1.21535981e+00 -1.64116532e-01 -4.94023636e-02 1.09554552e-01
1.07740188e+00 1.32290646e-01 -6.88239396e-01 8.48228693e-01
-4.34273362e-01 -2.40833327e-01 4.16020811e-01 6.76580429e-01
-3.85853261e-01 -1.13591599e+00 -2.28773504e-02 3.12014133e-01
-4.11249012e-01 -7.97912925e-02 -1.37028468e+00 3.75144690e-01
-2.43481472e-01 1.44822419e+00 2.29726180e-01 -5.23478985e-01
1.74902588e-01 8.90512705e-01 -1.26598716e-01 -5.77650607e-01
-1.36368299e+00 -2.02242415e-02 9.73626375e-01 -5.44573605e-01
-1.03557956e+00 -3.42008531e-01 -1.60374951e+00 -2.25564197e-01
-3.50784153e-01 6.31592274e-01 5.87489665e-01 1.34426296e+00
2.56711971e-02 4.10451978e-01 7.30014622e-01 -5.86552083e-01
3.95965539e-02 -1.48859346e+00 -3.53663415e-01 6.99100912e-01
1.70559958e-02 -2.90963650e-01 -3.52571309e-01 1.74896382e-02] | [12.983922958374023, 6.352423191070557] |
6947c9d1-7068-4a58-bab4-f1e9dbdb5ad9 | exploring-the-application-of-large-scale-pre | 2306.09008 | null | https://arxiv.org/abs/2306.09008v1 | https://arxiv.org/pdf/2306.09008v1.pdf | Exploring the Application of Large-scale Pre-trained Models on Adverse Weather Removal | Image restoration under adverse weather conditions (e.g., rain, snow and haze) is a fundamental computer vision problem and has important indications for various downstream applications. Different from early methods that are specially designed for specific type of weather, most recent works tend to remove various adverse weather effects simultaneously through either spatial feature representation learning or semantic information embedding. Inspired by the various successful applications of large-scale pre-trained models (e.g, CLIP), in this paper, we explore the potential benefits of them for this task through both spatial feature representation learning and semantic information embedding aspects: 1) for spatial feature representation learning, we design a Spatially-Adaptive Residual (\textbf{SAR}) Encoder to extract degraded areas adaptively. To facilitate its training, we propose a Soft Residual Distillation (\textbf{CLIP-SRD}) strategy to transfer the spatial knowledge from CLIP between clean and adverse weather images; 2) for semantic information embedding, we propose a CLIP Weather Prior (\textbf{CWP}) embedding module to make the network handle different weather conditions adaptively. This module integrates the sample specific weather prior extracted by CLIP image encoder together with the distribution specific information learned by a set of parameters, and embeds them through a cross attention mechanism. Extensive experiments demonstrate that our proposed method can achieve state-of-the-art performance under different and challenging adverse weather conditions. Code will be made available. | ['Nenghai Yu', 'Jieping Ye', 'Le Lu', 'Qi Chu', 'Qiankun Liu', 'Yue Wu', 'Zhentao Tan'] | 2023-06-15 | null | null | null | null | ['image-restoration'] | ['computer-vision'] | [ 2.82373190e-01 -3.11993420e-01 1.10060588e-01 -5.77368915e-01
-6.17253780e-01 -2.88217098e-01 4.71069127e-01 -3.18484247e-01
-1.44621432e-01 6.69349194e-01 5.92708707e-01 -2.56691277e-01
-1.57606855e-01 -8.39739859e-01 -9.01590407e-01 -1.11919069e+00
-4.33020443e-02 -4.33579206e-01 2.68988550e-01 -4.15175796e-01
2.71659583e-01 4.55456287e-01 -1.56436169e+00 2.15459645e-01
1.34395790e+00 9.63659883e-01 8.31255734e-01 6.88244939e-01
-2.36035623e-02 6.54907942e-01 -7.47899234e-01 1.40800208e-01
3.96970063e-01 -5.23011088e-01 -2.58141905e-01 9.40438583e-02
3.97536576e-01 -5.07386446e-01 -6.61767483e-01 1.04806209e+00
7.14780509e-01 3.91811907e-01 6.13927245e-01 -8.53872895e-01
-1.11919820e+00 -7.52436519e-02 -7.33971894e-01 6.33248150e-01
1.66753232e-01 3.82240295e-01 5.81011832e-01 -9.72932816e-01
1.95669264e-01 1.10459280e+00 6.06119394e-01 2.73084253e-01
-8.03890705e-01 -6.55385256e-01 6.06286287e-01 5.50881445e-01
-1.21067894e+00 -3.97520453e-01 6.91474855e-01 -2.13605687e-01
8.09578896e-01 3.31801474e-01 4.06485349e-01 7.91331768e-01
4.01521832e-01 8.11636209e-01 1.30257380e+00 -2.37362877e-01
1.45799756e-01 1.86479345e-01 -4.89794686e-02 2.66743481e-01
-4.69777957e-02 3.91614318e-01 -3.53604853e-01 2.94608653e-01
5.73672771e-01 3.04589957e-01 -1.00665343e+00 1.79862622e-02
-8.05910528e-01 7.06307471e-01 1.00792789e+00 6.41209707e-02
-4.19389188e-01 -3.59782465e-02 -3.46587226e-02 4.43346441e-01
7.77876318e-01 5.60531020e-02 -4.88195062e-01 5.08208394e-01
-9.18752968e-01 2.96635449e-01 4.10966724e-01 7.83394337e-01
1.01234162e+00 3.24712992e-01 -3.09778363e-01 1.06120896e+00
4.13302600e-01 1.00313210e+00 4.09619361e-01 -5.21163344e-01
5.75118601e-01 1.79821074e-01 1.68653607e-01 -1.05442715e+00
-2.34746546e-01 -6.00621104e-01 -1.04898477e+00 4.40876812e-01
-2.18622625e-01 -2.05019280e-01 -1.34050333e+00 1.36376083e+00
3.66969407e-01 7.26974607e-01 3.02023470e-01 1.18356013e+00
8.53578746e-01 1.30657399e+00 -1.54932320e-01 -2.99542487e-01
1.18393528e+00 -1.22050297e+00 -8.91015649e-01 -5.44160962e-01
1.44674420e-01 -7.55466104e-01 1.05064702e+00 1.97248399e-01
-6.15175545e-01 -6.07130706e-01 -1.05442762e+00 7.57946298e-02
-6.22513533e-01 1.56617180e-01 1.94181591e-01 4.28283006e-01
-1.20147145e+00 3.74349236e-01 -3.40690553e-01 -3.43562305e-01
2.79313833e-01 3.02821659e-02 -3.10193360e-01 -7.20004320e-01
-1.38977623e+00 9.51528430e-01 1.29482701e-01 6.41096532e-01
-1.14011300e+00 -7.08131909e-01 -1.20771122e+00 4.42845374e-02
2.43012160e-01 -6.19000137e-01 6.75994992e-01 -1.01498508e+00
-1.36653042e+00 2.42202491e-01 -4.01503205e-01 -2.42472470e-01
-4.58115339e-02 -6.16677284e-01 -7.01622903e-01 3.99588853e-01
1.38760611e-01 2.97360510e-01 1.25954163e+00 -1.59485853e+00
-6.35073304e-01 -2.89295286e-01 1.22557893e-01 6.02669477e-01
-4.45921749e-01 1.99215457e-01 -3.78361434e-01 -1.11867702e+00
-1.14487037e-01 -6.00275517e-01 -3.13992649e-01 1.52883247e-01
-9.21468139e-02 2.38000587e-01 1.04455340e+00 -1.22637606e+00
1.14333165e+00 -2.24635124e+00 1.92668542e-01 -2.24124119e-02
-1.15534395e-01 4.89870518e-01 -3.05558681e-01 4.05288219e-01
-1.76684454e-01 -1.57458499e-01 -8.38602543e-01 -2.28049606e-01
-1.90226853e-01 2.72263110e-01 -6.17464602e-01 7.55699098e-01
3.67580712e-01 6.69911146e-01 -7.25139678e-01 -1.55760989e-01
5.81883729e-01 8.16415012e-01 -3.32641363e-01 6.16076767e-01
-5.68223894e-02 5.10749638e-01 -4.95026797e-01 6.10240877e-01
1.23905349e+00 2.41219118e-01 -4.91681367e-01 -3.35001796e-01
-3.00873399e-01 -1.67334199e-01 -1.11318135e+00 1.47938931e+00
-7.09028780e-01 4.89509165e-01 3.37660670e-01 -1.11148119e+00
9.84735966e-01 2.27565825e-01 -1.05259337e-01 -9.57697034e-01
-2.04311267e-01 1.35035470e-01 -3.67870510e-01 -9.21226025e-01
3.58597487e-01 -2.92934984e-01 1.78245723e-01 1.06749050e-01
-3.40156890e-02 -1.71356648e-01 -3.81333351e-01 5.27230091e-02
8.81449282e-01 3.14746022e-01 1.23228866e-03 -1.72719657e-01
7.43652999e-01 -2.24589407e-01 7.37804770e-01 5.00028014e-01
-2.82922983e-01 1.11139953e+00 -1.45487055e-01 -4.64268655e-01
-7.77016580e-01 -1.03129685e+00 -1.70127943e-01 1.00107658e+00
6.43763065e-01 1.17358647e-01 -4.56725925e-01 -5.65548599e-01
-1.61695808e-01 5.94693005e-01 -7.15810239e-01 -4.05463010e-01
-6.71047032e-01 -1.11173511e+00 1.95799619e-01 5.89346349e-01
8.53475749e-01 -1.15836442e+00 -1.85912192e-01 2.24941373e-02
-2.00656503e-01 -9.11682129e-01 -4.77519810e-01 1.45597145e-01
-5.38000822e-01 -7.87623286e-01 -1.11648285e+00 -8.02285314e-01
5.47307014e-01 9.28524911e-01 7.51204610e-01 7.56778866e-02
-3.37795019e-01 3.26860279e-01 -9.08306479e-01 -3.38655949e-01
2.92030603e-01 -2.95122892e-01 -1.15045220e-01 3.96205425e-01
-5.64532652e-02 -7.96661615e-01 -1.06463981e+00 2.34929055e-01
-1.38390613e+00 -1.91275671e-01 7.91201413e-01 8.56063187e-01
3.27319831e-01 1.84765354e-01 5.10378599e-01 -5.39606512e-01
5.01136422e-01 -8.74807239e-01 -2.76753038e-01 1.12232909e-01
-3.90604883e-01 -2.04504400e-01 7.52882421e-01 -9.25940648e-02
-1.56038785e+00 -1.46407038e-01 -3.02552521e-01 -5.03597796e-01
-4.98527855e-01 6.14440501e-01 -3.93802315e-01 -1.49567172e-01
4.61761892e-01 8.73052061e-01 -1.71293452e-01 -6.56077087e-01
4.41807181e-01 9.16432798e-01 6.19355500e-01 -2.65403569e-01
1.25582266e+00 5.83209753e-01 -3.85498315e-01 -8.90096545e-01
-9.74306226e-01 -4.98406887e-01 -2.73428857e-01 -2.53967494e-01
1.14420927e+00 -1.34190488e+00 -7.64169917e-02 6.13126993e-01
-8.54131758e-01 -4.76801693e-01 2.21108906e-02 4.99039322e-01
-2.16952309e-01 4.88145351e-01 -5.01719236e-01 -8.59409571e-01
-3.18949819e-01 -9.44922268e-01 1.17318892e+00 7.42372036e-01
6.30478024e-01 -9.90244150e-01 6.61847293e-02 2.54637092e-01
8.02916825e-01 1.98984027e-01 4.92136627e-01 6.43847361e-02
-7.00815856e-01 -2.99716201e-02 -5.13140261e-01 6.77648902e-01
3.11059624e-01 -5.12621939e-01 -1.13970947e+00 -5.03754318e-01
3.04796904e-01 -1.21458016e-01 1.35344386e+00 5.00850677e-01
1.10088336e+00 -3.92071486e-01 -2.54346967e-01 1.09032500e+00
1.68115783e+00 1.70912459e-01 1.04457605e+00 6.21020794e-01
7.28949487e-01 5.67360222e-01 7.57925749e-01 3.17228436e-01
4.44873303e-01 5.33401489e-01 7.88723230e-01 -4.46067125e-01
-3.58533353e-01 9.63695422e-02 5.48149526e-01 4.90270257e-01
4.46361909e-03 -4.52617615e-01 -3.89984936e-01 8.39870632e-01
-1.74433577e+00 -9.49682117e-01 1.81230590e-01 2.00648785e+00
6.21574342e-01 -2.70753384e-01 -5.31872511e-01 5.88494055e-02
7.03543484e-01 8.01761568e-01 -5.81548929e-01 -1.87326238e-01
-4.06626582e-01 2.49842778e-01 5.36323011e-01 5.87186933e-01
-1.13777530e+00 8.78408909e-01 5.05725431e+00 8.23356986e-01
-1.15774214e+00 8.57500061e-02 5.11208892e-01 4.76553738e-02
-5.11476099e-01 5.63201727e-03 -5.85080266e-01 6.99969172e-01
7.21834898e-01 3.10855746e-01 4.17167604e-01 4.28668976e-01
4.54626679e-01 -6.17356077e-02 -2.99193561e-01 6.44775391e-01
3.81388128e-01 -1.11677182e+00 -7.27170184e-02 -2.01817498e-01
7.46551096e-01 6.23242930e-02 8.28894377e-02 4.03001815e-01
4.99151833e-02 -1.04134047e+00 5.81925988e-01 7.62698770e-01
8.14341784e-01 -5.86887121e-01 8.72718871e-01 3.10740650e-01
-1.37014568e+00 -4.66213197e-01 -4.67342079e-01 -1.17243342e-01
4.05530661e-01 8.29217315e-01 -1.44175708e-01 1.17422295e+00
1.18542802e+00 1.10413766e+00 -4.62562710e-01 1.32557583e+00
-5.80986977e-01 7.65619338e-01 -8.41422155e-02 4.39249724e-01
1.70543283e-01 -3.20935071e-01 6.31956339e-01 1.26254046e+00
4.16745692e-01 3.91678810e-01 6.24011010e-02 4.72921699e-01
3.33816111e-01 -1.53591946e-01 -6.00940168e-01 5.58092892e-01
2.41868272e-01 1.17030573e+00 -3.17586005e-01 -2.81819910e-01
-4.59637254e-01 1.33954823e+00 5.04649058e-02 7.69420505e-01
-8.13089490e-01 -7.05121696e-01 8.81118894e-01 -4.47879992e-02
8.03371668e-01 -1.53476611e-01 -1.20043114e-01 -1.21585310e+00
3.76414806e-01 -7.43755281e-01 1.69731662e-01 -1.16086674e+00
-1.30864334e+00 6.43689692e-01 -1.21074267e-01 -1.48336077e+00
4.12188411e-01 -5.05979538e-01 -1.08838654e+00 1.39341497e+00
-2.52060199e+00 -1.20959103e+00 -7.09281087e-01 7.24117935e-01
8.27117622e-01 7.39687979e-02 4.85754758e-01 5.03323495e-01
-7.04784513e-01 3.20395678e-01 3.80347997e-01 -1.35850608e-01
9.21897531e-01 -1.19676721e+00 2.52882659e-01 1.29209769e+00
-2.77650952e-01 3.85765105e-01 8.54220331e-01 -5.12305737e-01
-1.43238354e+00 -1.60981607e+00 5.63432038e-01 2.59279571e-02
2.33799160e-01 -1.00104190e-01 -1.15790761e+00 5.89198589e-01
4.55936402e-01 3.74011040e-01 4.26829070e-01 -4.33521837e-01
-4.25118744e-01 -5.07747233e-01 -1.14905989e+00 5.12537122e-01
8.84338260e-01 -3.70437950e-01 -6.47767484e-01 4.26569968e-01
1.17253518e+00 -3.87170076e-01 -5.49282253e-01 6.91962779e-01
1.22794226e-01 -1.01384556e+00 1.08544147e+00 -1.08907305e-01
4.47364241e-01 -8.83894622e-01 -4.04126972e-01 -1.69840682e+00
-5.77086747e-01 -2.49532849e-01 -1.84905216e-01 1.18919146e+00
1.76006779e-01 -8.66601646e-01 3.13948482e-01 1.92032512e-02
-7.87068665e-01 -7.21927762e-01 -7.24005222e-01 -5.62677145e-01
3.90375145e-02 -3.56930308e-02 5.93384624e-01 9.03092384e-01
-6.11886919e-01 -3.45220976e-02 -7.26536632e-01 9.76874650e-01
4.90072638e-01 2.27286845e-01 5.06436288e-01 -6.94476545e-01
-2.86417902e-01 -1.81540787e-01 -2.16050953e-01 -1.21949446e+00
5.07530198e-02 -4.59724993e-01 5.63537598e-01 -1.91144979e+00
-4.70054485e-02 -3.36545646e-01 -5.00660300e-01 3.68755043e-01
-7.03325570e-01 4.54486340e-01 2.11806968e-02 2.12677822e-01
-2.88544774e-01 1.08663344e+00 1.22155821e+00 -2.41461873e-01
-3.64017561e-02 -8.83720517e-02 -9.47272360e-01 3.41684341e-01
6.72679484e-01 -3.00727099e-01 -4.17406321e-01 -8.60456586e-01
5.83974458e-02 1.88436713e-02 4.52880532e-01 -1.14205861e+00
1.43673077e-01 -4.26388562e-01 6.16132200e-01 -3.78469944e-01
4.12119269e-01 -7.95926929e-01 -2.45631829e-01 1.58631772e-01
7.54655227e-02 -2.19730198e-01 2.36095354e-01 9.60324943e-01
-5.63730299e-01 -1.43494308e-01 7.16429174e-01 -1.10441141e-01
-1.05474007e+00 4.23070014e-01 -3.41227859e-01 -2.46200040e-01
9.16827679e-01 -2.18097061e-01 -6.25154972e-01 -4.95537907e-01
-4.70685661e-01 5.02913177e-01 2.93266088e-01 4.64852899e-01
8.71758938e-01 -1.03039002e+00 -8.53369534e-01 3.01876634e-01
1.22794420e-01 1.81917399e-01 7.81471014e-01 7.89819360e-01
-4.71668631e-01 -5.21765687e-02 1.33152986e-02 -3.46296757e-01
-9.01773989e-01 5.92298746e-01 4.09096330e-01 4.29912098e-02
-9.65470493e-01 8.76075923e-01 6.66903794e-01 -4.54690397e-01
1.22721195e-02 -1.26698479e-01 -4.41216677e-01 -2.46021986e-01
1.08943689e+00 9.85873565e-02 1.10225029e-01 -6.03302836e-01
-1.50972947e-01 8.59064639e-01 2.04628576e-02 5.44726551e-02
1.67144108e+00 -4.81273413e-01 -9.27401334e-02 1.16959713e-01
1.12681639e+00 -5.04915975e-03 -1.72478890e+00 -3.38354290e-01
-5.95412970e-01 -7.56974518e-01 2.36928329e-01 -8.97792518e-01
-1.35236824e+00 1.05114186e+00 9.61616278e-01 1.05363622e-01
1.76668274e+00 -1.73356324e-01 9.65236247e-01 -2.50414182e-02
-6.94708526e-02 -7.36983597e-01 -7.09267259e-02 5.56971967e-01
1.12618756e+00 -1.25021422e+00 3.24787349e-02 -3.87005508e-01
-5.98127186e-01 8.74622822e-01 5.57746112e-01 -3.26250136e-01
8.39419305e-01 1.22177295e-01 3.69683266e-01 2.49548815e-02
-4.16132629e-01 -3.38058054e-01 2.26325467e-01 8.83576095e-01
2.71439254e-02 -8.24092105e-02 -6.71535283e-02 5.27617037e-01
1.49160460e-01 -1.72692925e-01 5.32251954e-01 1.00413024e+00
-8.57808053e-01 -6.78138137e-01 -6.42110348e-01 3.17056388e-01
-1.60969317e-01 -3.28148395e-01 1.16602115e-01 2.13353574e-01
4.03081745e-01 1.11249554e+00 -1.30409718e-01 -3.96681845e-01
3.89486104e-01 -2.80260324e-01 6.82260990e-02 -5.49854875e-01
-4.43299085e-01 -7.38655478e-02 -2.24649265e-01 -5.48658192e-01
-5.19285977e-01 -4.90234047e-01 -9.51021194e-01 2.03472972e-01
-3.17787290e-01 4.46210615e-02 6.13561153e-01 1.11277759e+00
4.86320347e-01 7.51064003e-01 1.03546560e+00 -1.31382740e+00
-2.66824871e-01 -1.15299368e+00 -7.18591988e-01 1.04482099e-01
8.76350641e-01 -6.29421592e-01 -7.36691892e-01 1.29082784e-01] | [10.993401527404785, -3.0468790531158447] |
366ada70-6e3f-4dce-ae8a-ce63298ffcdf | deletion-robust-submodular-maximization-data | null | null | https://icml.cc/Conferences/2017/Schedule?showEvent=698 | http://proceedings.mlr.press/v70/mirzasoleiman17a/mirzasoleiman17a.pdf | Deletion-Robust Submodular Maximization: Data Summarization with "the Right to be Forgotten" |
How can we summarize a dynamic data stream when elements selected for the summary can be deleted at any time? This is an important challenge in online services, where the users generating the data may decide to exercise their right to restrict the service provider from using (part of) their data due to privacy concerns. Motivated by this challenge, we introduce the dynamic deletion-robust submodular maximization problem. We develop the first resilient streaming algorithm, called ROBUST-STREAMING, with a constant factor approximation guarantee to the optimum solution. We evaluate the effectiveness of our approach on several real-world applica tions, including summarizing (1) streams of geo-coordinates (2); streams of images; and (3) click-stream log data, consisting of 45 million feature vectors from a news recommendation task.
| ['Baharan Mirzasoleiman', 'Andreas Krause', 'Amin Karbasi'] | 2017-08-01 | null | null | null | icml-2017-8 | ['data-summarization'] | ['miscellaneous'] | [ 4.45647240e-01 2.53687263e-03 -3.58323097e-01 -5.01111209e-01
-1.16998172e+00 -8.30965817e-01 4.02149111e-02 5.27871847e-01
-3.86251181e-01 5.69390714e-01 5.97012043e-01 -2.23976910e-01
-3.00333530e-01 -4.15257812e-01 -1.00727534e+00 -6.63248718e-01
-7.80855596e-01 3.58056694e-01 1.22137286e-01 -1.51413843e-01
2.76348710e-01 4.78387713e-01 -1.19836318e+00 5.68279266e-01
6.26164973e-01 1.27233148e+00 5.89408576e-02 6.59169257e-01
6.68056533e-02 2.43698806e-01 -4.03552383e-01 -4.30291086e-01
8.40918124e-01 -1.10247619e-01 -3.20212632e-01 4.07695323e-01
3.99188548e-01 -7.78807998e-01 -4.36316222e-01 1.14654899e+00
4.82399613e-01 3.09193999e-01 2.78524518e-01 -1.74599040e+00
-2.28931680e-01 8.29858243e-01 -8.15757811e-01 4.07381624e-01
6.22401953e-01 -3.20161700e-01 1.28479099e+00 -6.60041213e-01
8.48394632e-01 8.95107210e-01 3.56819421e-01 2.18941450e-01
-1.10305297e+00 -1.77630693e-01 5.68995357e-01 2.58965176e-02
-1.23595810e+00 -8.25711071e-01 4.25139755e-01 3.52718718e-02
6.74773991e-01 1.04643607e+00 2.75009722e-01 6.23426855e-01
-5.21222837e-02 1.21310580e+00 3.06061089e-01 2.89562821e-01
4.74374712e-01 3.09438676e-01 6.84370548e-02 9.92102176e-02
6.40737593e-01 -5.33990562e-01 -9.86037910e-01 -1.17976463e+00
2.82277521e-02 3.04942578e-01 -5.15728891e-01 -4.41373497e-01
-7.89871633e-01 7.18038797e-01 -1.65914625e-01 -3.60203296e-01
-7.28857815e-01 1.80455208e-01 3.30417126e-01 6.13695383e-01
6.11464024e-01 2.25892350e-01 -5.29992342e-01 -6.77451044e-02
-1.03859305e+00 8.77605915e-01 1.21780527e+00 1.41278279e+00
3.03170592e-01 -1.16995305e-01 -3.31490263e-02 6.00842595e-01
2.24488601e-01 4.25833136e-01 1.40599698e-01 -8.63019288e-01
1.02622175e+00 1.73464254e-01 4.20364559e-01 -1.21553063e+00
-8.62954855e-02 -1.91381305e-01 -6.33075297e-01 -4.43227649e-01
1.46816298e-01 -5.07363141e-01 -2.35450506e-01 1.50998855e+00
5.56988358e-01 1.74394399e-01 -8.42575729e-02 9.43773150e-01
5.23189187e-01 9.65189695e-01 -5.59634447e-01 -8.33024859e-01
9.74220693e-01 -4.94619429e-01 -6.35678351e-01 2.50990912e-02
4.32222009e-01 -3.34815234e-01 3.04553121e-01 7.75054157e-01
-1.44305658e+00 4.62086260e-01 -9.53919947e-01 1.01267186e-03
7.22731501e-02 -6.83513582e-01 6.28246248e-01 5.95871925e-01
-9.58519340e-01 5.37491202e-01 -6.58425272e-01 -2.17501953e-01
6.98421121e-01 3.29247713e-01 -3.99357378e-01 -2.88548440e-01
-8.19966733e-01 -2.70789534e-01 -8.52116868e-02 -2.32715562e-01
-8.34391594e-01 -1.11511993e+00 -5.89094758e-01 3.91712010e-01
8.45253706e-01 -6.34739161e-01 1.06607258e+00 -6.59219086e-01
-6.98954999e-01 4.35556352e-01 -5.20990610e-01 -5.73047698e-01
8.53233576e-01 -1.96213946e-01 -2.60217667e-01 2.71130472e-01
1.12498194e-01 -1.03769600e-01 1.02921534e+00 -1.23141444e+00
-1.19487607e+00 -7.25347638e-01 -6.37314245e-02 2.34612510e-01
-6.12572074e-01 3.94618928e-01 -5.22076786e-01 -7.49153614e-01
1.98704049e-01 -8.70551646e-01 -6.57041430e-01 1.95917040e-01
-6.19640768e-01 1.26784444e-01 7.11099029e-01 -7.30171263e-01
1.47426879e+00 -2.18324828e+00 1.21527184e-02 5.01528800e-01
1.93929702e-01 -1.30357206e-01 -2.06065580e-01 8.08763981e-01
3.10114473e-01 3.71565729e-01 -2.25416332e-01 -5.77525020e-01
-8.05291012e-02 6.71550781e-02 -8.47384810e-01 7.56380439e-01
-5.02288878e-01 3.52374285e-01 -8.21805596e-01 5.54779433e-02
-7.78557241e-01 -2.01847270e-01 -9.04503584e-01 1.19565576e-01
-4.33889747e-01 -2.17958152e-01 -7.63997972e-01 8.22996736e-01
1.16992700e+00 -2.11255252e-01 1.00126833e-01 3.27728420e-01
1.54980868e-01 1.48494691e-01 -1.70041633e+00 1.57888925e+00
9.58408117e-02 3.18728447e-01 6.78153336e-01 -6.98005736e-01
3.96923125e-01 2.78158724e-01 1.09063029e+00 -1.42526254e-01
-2.10772634e-01 3.96575294e-02 -7.07621694e-01 -5.99482000e-01
1.08645916e+00 2.33084902e-01 -2.17118964e-01 9.22891915e-01
-4.97334361e-01 4.02168661e-01 1.41931437e-02 7.71646082e-01
1.35484290e+00 -6.96833372e-01 9.74122137e-02 9.70711641e-04
1.75073631e-02 -1.01021051e-01 7.12154329e-01 1.21914458e+00
-2.35377513e-02 1.03807974e+00 9.91875529e-01 -2.30370522e-01
-9.84292626e-01 -8.92381966e-01 1.25000954e-01 1.32149100e+00
1.04711756e-01 -6.48007572e-01 -3.93154472e-01 -8.75214756e-01
6.13159955e-01 7.23112226e-01 -3.44125539e-01 2.79577881e-01
-3.31502378e-01 -8.11051190e-01 1.26925990e-01 5.44506088e-02
-1.40448987e-01 -2.38452792e-01 -2.73218244e-01 3.16436261e-01
-3.55507851e-01 -1.10521436e+00 -1.20981348e+00 -3.69439930e-01
-8.62528563e-01 -7.15188444e-01 -5.07199466e-01 -1.98040679e-01
7.27408111e-01 9.54284310e-01 6.84458971e-01 -5.08731939e-02
-4.09932323e-02 6.49415195e-01 -2.58394241e-01 -5.41601360e-01
-2.79939305e-02 1.22412808e-01 1.04674511e-01 6.81856751e-01
-4.62987125e-02 -7.62537301e-01 -5.35872996e-01 6.34050369e-02
-1.38035560e+00 -4.99680132e-01 -2.14297414e-01 2.33712807e-01
9.31655705e-01 1.20537400e-01 1.02794325e+00 -1.27096796e+00
9.55769122e-01 -1.39241052e+00 -5.94942093e-01 1.81446865e-01
-5.42339802e-01 -3.77538383e-01 6.75948203e-01 -2.42700472e-01
-7.80402064e-01 3.50814983e-02 1.20417863e-01 -1.57459021e-01
1.90124482e-01 5.85846424e-01 -4.06369328e-01 1.31389424e-01
4.58754361e-01 3.88947219e-01 -9.66786072e-02 -6.96201861e-01
4.29475665e-01 9.26254749e-01 2.65880883e-01 -2.19058186e-01
5.67195415e-01 1.09504211e+00 4.86857481e-02 -1.03021419e+00
-5.94468653e-01 -8.83063018e-01 9.30939242e-02 8.96129757e-02
7.58044645e-02 -1.04627848e+00 -6.67211473e-01 3.52013439e-01
-9.32363093e-01 3.47008467e-01 -4.90874946e-01 1.63979650e-01
-5.98176360e-01 6.97252631e-01 -2.96836406e-01 -8.62731934e-01
-4.46336925e-01 -5.54654181e-01 8.46212089e-01 -2.44278274e-02
2.33593225e-01 -4.63975519e-01 -1.37100101e-01 2.10005239e-01
4.00301456e-01 4.82504338e-01 4.81391817e-01 -1.23942769e+00
-9.47405517e-01 -7.38296866e-01 1.02721229e-01 2.85408437e-01
2.99315602e-02 -2.68134922e-01 -5.44370949e-01 -6.81395292e-01
1.92765206e-01 5.03148437e-02 7.84818172e-01 2.61269480e-01
1.58666277e+00 -1.20394576e+00 -1.61020502e-01 1.01407719e+00
1.27804315e+00 -9.58820134e-02 2.89943725e-01 -2.03658715e-02
4.52668667e-01 4.29113209e-01 8.26008499e-01 1.50233984e+00
5.75125217e-01 4.34982866e-01 5.33205926e-01 6.50463343e-01
6.41162574e-01 -4.82223660e-01 5.74641049e-01 6.49379134e-01
1.96312010e-01 -8.34488571e-01 -1.87023252e-01 7.09185839e-01
-2.13687992e+00 -1.05085278e+00 -9.70836580e-02 2.46368027e+00
5.51395655e-01 -1.77577436e-01 3.75578403e-01 -1.87604904e-01
5.63165724e-01 6.12664640e-01 -1.00319576e+00 -4.85119134e-01
-3.86898547e-01 -5.51343858e-01 1.09667194e+00 1.59324348e-01
-1.04579997e+00 3.16785723e-01 6.19256115e+00 8.19095314e-01
-8.01206589e-01 3.85519862e-02 7.67406583e-01 -9.89217222e-01
-1.10268092e+00 -1.61275163e-01 -9.18588281e-01 6.68283701e-01
1.00818586e+00 -8.97052526e-01 7.23202646e-01 1.00151408e+00
4.05222833e-01 -1.10895000e-01 -1.13562083e+00 9.89290059e-01
3.76571268e-01 -1.66217804e+00 1.07878894e-01 3.75954211e-01
1.07530165e+00 1.75034001e-01 2.29538530e-01 -1.87825590e-01
9.85241681e-02 -5.98858595e-01 5.73310494e-01 5.05711317e-01
7.73377895e-01 -8.79172564e-01 1.58758506e-01 5.79171479e-01
-6.99684262e-01 -3.28396261e-01 -5.87574720e-01 2.93534607e-01
7.26836801e-01 7.09678590e-01 -8.51908505e-01 5.80577135e-01
8.94107997e-01 5.52045643e-01 -1.34121463e-01 1.38458323e+00
2.78379709e-01 9.42549646e-01 -8.74680996e-01 -7.21024424e-02
-2.68194601e-02 -2.84255981e-01 1.34172869e+00 1.09170473e+00
7.25829065e-01 3.22556645e-01 -8.44740570e-02 3.20333689e-01
-5.57534397e-01 4.06219900e-01 -7.36870646e-01 -2.08373694e-03
6.04806900e-01 9.88351643e-01 -2.87385464e-01 -1.62189662e-01
-4.89613801e-01 8.44764650e-01 -1.35484770e-01 5.67567170e-01
-5.90606272e-01 -3.04229528e-01 1.21015024e+00 5.36780775e-01
6.28739536e-01 -2.09992409e-01 -3.99407953e-01 -1.27711511e+00
4.99122113e-01 -9.25345957e-01 7.80819118e-01 -3.58523428e-01
-1.21097863e+00 8.83269608e-02 1.06134918e-02 -1.06868219e+00
-3.11614275e-01 2.28371590e-01 -4.35483456e-01 6.18419945e-01
-1.29121506e+00 -4.09214795e-01 4.12366986e-02 6.63137197e-01
2.77112782e-01 -3.12591046e-01 2.38156840e-01 4.79203612e-01
-3.54284674e-01 9.57824588e-01 7.46055722e-01 -4.71201211e-01
4.51290160e-01 -1.00140190e+00 5.56457043e-01 1.07985938e+00
5.11722453e-02 4.21649247e-01 8.56517911e-01 -6.84264660e-01
-2.24998713e+00 -1.19941461e+00 1.04350328e+00 -1.87442243e-01
4.68651682e-01 -6.39865041e-01 -9.30433273e-01 8.13910723e-01
-2.46738881e-01 2.20924333e-01 9.77684081e-01 -3.18993092e-01
-7.17655569e-02 -3.94986063e-01 -1.67122412e+00 5.65361440e-01
1.25293446e+00 -3.65588889e-02 -1.08842567e-01 6.95828199e-01
1.16938627e+00 -3.14350873e-01 -6.69437408e-01 2.23225504e-02
4.89059865e-01 -7.05100656e-01 7.84033954e-01 -1.25635993e+00
2.25112945e-01 3.17597622e-03 -4.68702942e-01 -1.14910400e+00
-5.53028509e-02 -1.60565734e+00 -5.64765990e-01 1.24281621e+00
4.23755616e-01 -7.05211401e-01 1.12517047e+00 1.04391408e+00
1.97983772e-01 -6.43926322e-01 -1.23081374e+00 -6.58176184e-01
-4.29810435e-01 -5.97872734e-01 1.00769162e+00 6.80039048e-01
-1.08650103e-01 -3.35087866e-01 -7.72589028e-01 6.10995114e-01
8.48626256e-01 2.27422923e-01 9.70149636e-01 -7.67345309e-01
-4.71892178e-01 3.03184669e-02 -1.44080684e-01 -1.55766547e+00
-2.94687718e-01 -9.72301245e-01 -2.05223486e-01 -1.22406268e+00
1.63606733e-01 -5.12393594e-01 -2.40873918e-01 6.04705513e-02
-3.95512879e-02 -4.17132556e-01 4.83909041e-01 1.76028252e-01
-1.05846965e+00 5.32269478e-01 8.29565108e-01 1.63777530e-01
-3.31291527e-01 8.19132626e-01 -1.40954208e+00 3.78547668e-01
4.23059553e-01 -9.09401059e-01 -4.16190863e-01 -4.14060175e-01
7.48493910e-01 5.87981582e-01 -1.25698328e-01 -2.31594399e-01
5.28353572e-01 -6.13159895e-01 -2.73537248e-01 -8.17359090e-01
2.12494865e-01 -1.00832951e+00 1.80059493e-01 3.34574543e-02
-6.98274612e-01 5.96827678e-02 -2.89668739e-01 1.19210851e+00
-4.53374386e-02 -3.96651119e-01 2.01111555e-01 1.05926469e-01
-2.42630765e-01 1.11041546e+00 -3.10534000e-01 4.22831506e-01
1.07882774e+00 -2.79011875e-02 -2.98386097e-01 -1.11831021e+00
-4.89582598e-01 8.00706923e-01 4.00200874e-01 3.86998087e-01
7.91774035e-01 -1.15119696e+00 -1.17653477e+00 1.38637483e-01
6.22279011e-02 1.49281040e-01 4.90427136e-01 6.29796207e-01
-2.05624297e-01 1.45946190e-01 3.93589765e-01 -2.44990528e-01
-1.30293763e+00 8.74926805e-01 -3.03581089e-01 9.88548324e-02
-4.91965264e-01 9.25358772e-01 -6.65333942e-02 1.70262843e-01
5.46340942e-01 -2.11841211e-01 7.52383024e-02 3.29917699e-01
1.08830702e+00 9.25582767e-01 1.74350753e-01 -3.35278124e-01
-2.29691386e-01 -5.60162365e-01 -3.33858818e-01 -2.75248587e-01
1.78566480e+00 -5.34716427e-01 -1.68115541e-01 2.14983881e-01
1.46225655e+00 4.93266135e-01 -1.29410207e+00 -4.88036543e-01
-1.47240102e-01 -1.12439787e+00 -2.69397289e-01 -2.25621149e-01
-1.30060554e+00 5.90148643e-02 -5.81639819e-02 7.42426693e-01
1.06188726e+00 -7.97180235e-02 1.27166998e+00 3.23399007e-01
5.32117665e-01 -1.03948963e+00 -4.26356524e-01 4.80037816e-02
1.06500661e+00 -9.57148552e-01 3.18268597e-01 -4.15587097e-01
-8.18960428e-01 7.85445213e-01 -1.63003027e-01 -2.39648849e-01
1.12228537e+00 2.86270559e-01 -4.61091220e-01 1.10206582e-01
-1.33747101e+00 4.38455135e-01 1.22109270e-02 4.22613680e-01
-3.74066025e-01 1.91383839e-01 -5.05372465e-01 1.15609980e+00
1.38491103e-02 -1.23512313e-01 9.76236165e-01 1.18448913e+00
-5.62680006e-01 -8.42984200e-01 -3.20276499e-01 1.07681596e+00
-9.49494541e-01 -3.60331987e-03 -1.61931768e-01 -1.98837847e-01
-4.37815607e-01 1.13797271e+00 1.06707886e-02 -1.71967477e-01
3.90482366e-01 -3.72375339e-01 -1.94392964e-01 -5.12961626e-01
-5.94618738e-01 -1.77841093e-02 3.55138600e-01 -7.40782917e-01
2.78815836e-01 -1.36414146e+00 -1.02732956e+00 -4.21315342e-01
1.02510311e-01 1.65134877e-01 9.57943201e-01 5.05660176e-01
7.81950295e-01 -1.03821449e-01 1.31060553e+00 -4.46650356e-01
-1.25773132e+00 -2.48163283e-01 -1.05314434e+00 4.65855956e-01
5.90009034e-01 2.11236343e-01 -6.85388803e-01 -1.67676553e-01] | [6.578424453735352, 5.006245136260986] |
6d752dfe-97c5-48cc-91ca-1c33f8045dd3 | the-dota-2-bot-competition | 2103.02943 | null | https://arxiv.org/abs/2103.02943v1 | https://arxiv.org/pdf/2103.02943v1.pdf | The Dota 2 Bot Competition | Multiplayer Online Battle Area (MOBA) games are a recent huge success both in the video game industry and the international eSports scene. These games encourage team coordination and cooperation, short and long-term planning, within a real-time combined action and strategy gameplay. Artificial Intelligence and Computational Intelligence in Games research competitions offer a wide variety of challenges regarding the study and application of AI techniques to different game genres. These events are widely accepted by the AI/CI community as a sort of AI benchmarking that strongly influences many other research areas in the field. This paper presents and describes in detail the Dota 2 Bot competition and the Dota 2 AI framework that supports it. This challenge aims to join both, MOBAs and AI/CI game competitions, inviting participants to submit AI controllers for the successful MOBA \textit{Defense of the Ancients 2} (Dota 2) to play in 1v1 matches, which aims for fostering research on AI techniques for real-time games. The Dota 2 AI framework makes use of the actual Dota 2 game modding capabilities to enable to connect external AI controllers to actual Dota 2 game matches using the original Free-to-Play game.se of the actual Dota 2 game modding capabilities to enable to connect external AI controllers to actual Dota 2 game matches using the original Free-to-Play game. | ['Tobias Mahlmann', 'Jose M. Font'] | 2021-03-04 | null | null | null | null | ['dota-2'] | ['playing-games'] | [-2.41563454e-01 2.56691426e-02 1.61943734e-01 1.87127993e-01
-2.27195203e-01 -6.67394102e-01 5.99138916e-01 -2.58032262e-01
-8.00387263e-01 6.71897590e-01 -2.80466139e-01 -5.29589236e-01
-6.05768740e-01 -1.08553326e+00 -1.82925284e-01 -2.52428949e-01
-2.34763101e-01 1.09397864e+00 7.11830437e-01 -1.26183379e+00
2.50940025e-01 3.50180060e-01 -1.64222693e+00 2.80999690e-01
2.05554113e-01 9.15865421e-01 2.64216512e-01 9.69660997e-01
6.72198534e-02 1.17431986e+00 -7.25037754e-01 -3.85647595e-01
8.32553387e-01 -6.66170895e-01 -8.53448331e-01 -5.75507343e-01
-4.53144461e-01 -3.51313204e-02 -1.58129781e-01 6.66060328e-01
5.69176793e-01 4.08796161e-01 2.54913449e-01 -1.71274078e+00
5.94236553e-01 7.01504052e-01 -3.18199337e-01 5.57877779e-01
4.91170645e-01 4.80454564e-01 9.78998542e-01 -9.61799696e-02
8.96951258e-01 8.16829205e-01 5.70972741e-01 2.74348557e-01
-5.43552697e-01 -8.65106165e-01 -1.99882716e-01 4.56860662e-01
-1.35947692e+00 6.48892522e-02 5.13495505e-01 -4.79558110e-01
1.12771356e+00 4.99396056e-01 1.35518646e+00 7.00706184e-01
3.28830928e-01 5.87156057e-01 1.02673209e+00 -4.47656870e-01
6.42606199e-01 -3.40153039e-01 -3.00018489e-01 2.10076854e-01
7.98364133e-02 3.82190168e-01 -4.80263829e-01 -8.32783207e-02
9.18637216e-01 -4.77024019e-01 3.98798108e-01 -2.08716784e-02
-9.48720574e-01 1.09321463e+00 9.30537209e-02 6.50567830e-01
-8.31628919e-01 3.14999282e-01 6.61100924e-01 5.12932122e-01
-7.51617923e-02 1.17207503e+00 5.22062816e-02 -1.25918519e+00
-6.43571854e-01 1.22248137e+00 7.60310769e-01 3.06293160e-01
4.65373307e-01 2.27092907e-01 2.90878952e-01 5.87542355e-01
1.41256303e-01 1.06354833e-04 3.53035629e-01 -1.21740925e+00
1.55223802e-01 7.76383936e-01 8.14955533e-02 -1.27443469e+00
-8.41631711e-01 -1.02390841e-01 -2.44141221e-01 1.11676216e+00
6.30355239e-01 -3.99343520e-01 -3.25800776e-01 1.23981702e+00
3.31428379e-01 -8.43936503e-02 -2.56696325e-02 1.02978516e+00
6.16882205e-01 7.21293032e-01 -1.45898014e-01 1.47183865e-01
1.43367076e+00 -6.96070254e-01 -2.91731328e-01 -4.74553436e-01
6.92070663e-01 -6.21805966e-01 7.72247732e-01 7.01560318e-01
-1.46314621e+00 -4.20775503e-01 -1.06219983e+00 5.41787565e-01
-2.64191329e-01 -6.98519409e-01 9.26361322e-01 7.43359149e-01
-8.30846012e-01 2.59136468e-01 -8.04372847e-01 -4.82183486e-01
1.59375578e-01 7.79310465e-01 -4.59405392e-01 4.38196175e-02
-1.39104438e+00 1.05024850e+00 8.53626728e-01 -2.12171525e-01
-7.66744077e-01 -4.54706758e-01 -6.24479711e-01 -1.39849514e-01
9.83853221e-01 -4.38074350e-01 1.11023426e+00 -7.88513005e-01
-1.76164269e+00 1.26273572e+00 9.73377585e-01 -8.01252127e-01
5.67045212e-01 1.14015877e-01 -2.26814806e-01 -3.24066699e-01
2.92735368e-01 5.82277775e-01 -1.35949150e-01 -5.11034489e-01
-9.36376214e-01 -1.92802131e-01 9.01526332e-01 6.22935236e-01
2.88566858e-01 5.37990212e-01 -9.65176001e-02 -5.14185846e-01
-3.18060607e-01 -9.74561155e-01 -3.94690484e-01 -6.72271609e-01
-8.53004213e-03 -2.76279002e-01 4.33152318e-01 5.56101315e-02
1.33093965e+00 -1.77005184e+00 3.65254521e-01 3.61434132e-01
3.33850324e-01 4.31367725e-01 5.60304634e-02 9.80169833e-01
1.30133286e-01 -1.75277635e-01 4.70249563e-01 3.57271343e-01
2.82617480e-01 3.88452679e-01 1.56310748e-03 3.01582878e-03
-4.25101578e-01 7.78495967e-01 -8.51118207e-01 -1.27754107e-01
1.56667441e-01 -3.94527018e-01 -6.79568887e-01 1.14624426e-01
-1.66738361e-01 2.03418985e-01 -6.50112629e-01 2.04571113e-01
2.51262993e-01 4.15745169e-01 2.52321243e-01 4.62788016e-01
-7.67534196e-01 2.40506560e-01 -1.48332369e+00 1.64821017e+00
6.91993907e-02 5.54730296e-01 3.93617034e-01 -1.01982319e+00
1.05881643e+00 2.89329320e-01 7.13229954e-01 -8.98734093e-01
8.04163218e-01 4.13309723e-01 8.14607084e-01 -8.81202295e-02
7.82805204e-01 -3.76630932e-01 -6.47226334e-01 8.03716421e-01
-1.78851798e-01 -7.78035581e-01 6.74184680e-01 2.40907803e-01
1.39243746e+00 3.41658324e-01 4.66901541e-01 -3.70545655e-01
3.28445554e-01 7.75676608e-01 5.20479560e-01 1.01240706e+00
-4.63723779e-01 1.43728882e-01 5.53762496e-01 -9.75560367e-01
-9.71036434e-01 -5.56205571e-01 6.06606722e-01 1.35486662e+00
2.88986474e-01 -9.35109317e-01 -6.74119771e-01 -1.25997886e-01
-4.65350419e-01 4.93500769e-01 -5.78899324e-01 -1.53101072e-01
-6.51970863e-01 -2.50426650e-01 9.31064665e-01 2.78904885e-01
8.02887082e-01 -1.45892596e+00 -1.37797868e+00 6.10135913e-01
-2.03135848e-01 -9.32310700e-01 -4.50766981e-02 1.29683375e-01
-4.22417000e-02 -1.11875463e+00 -3.78732443e-01 -4.68624115e-01
-6.68286622e-01 1.48925111e-01 1.02777791e+00 1.91518627e-02
-5.73822856e-01 2.97945708e-01 -6.44827425e-01 -1.01401556e+00
-3.50534946e-01 2.07386374e-01 2.20662981e-01 -5.14866412e-01
5.48680305e-01 -8.05794477e-01 -1.37295738e-01 7.57698834e-01
-7.03263521e-01 5.41665435e-01 -1.71717312e-02 5.33526301e-01
5.76982349e-02 1.56586096e-01 3.23922068e-01 -3.73847216e-01
8.86017084e-01 -3.31609547e-01 -7.11748064e-01 -3.00010234e-01
-6.98802471e-02 -4.87327158e-01 3.30944538e-01 -2.80891120e-01
-3.63176107e-01 -5.29173434e-01 -2.48018771e-01 -8.66698921e-02
-1.31067216e-01 7.37828195e-01 -7.79955164e-02 -2.72598118e-01
1.01218164e+00 -1.31094724e-01 2.24662229e-01 1.91272140e-01
-1.74182821e-02 6.09869957e-01 5.07582784e-01 -6.67787671e-01
7.94185281e-01 1.99739456e-01 -1.98346395e-02 -5.51102936e-01
-1.00150898e-01 -3.56715143e-01 -2.20222861e-01 -9.70232010e-01
1.01847553e+00 -6.24060214e-01 -1.30432844e+00 6.92142367e-01
-7.92507231e-01 -7.11446404e-01 -6.07491195e-01 7.12541938e-01
-8.71575713e-01 -2.99553931e-01 -3.64426613e-01 -8.72689962e-01
-8.15502927e-02 -1.34272540e+00 3.25815499e-01 7.45112181e-01
-8.09094191e-01 -6.05938554e-01 5.47882378e-01 7.17672348e-01
6.01942718e-01 3.79902810e-01 3.49684328e-01 -7.77021527e-01
-3.08762491e-01 -3.87372136e-01 1.71745881e-01 -1.56859294e-01
-1.65686786e-01 -3.38327974e-01 -1.73005641e-01 9.53557864e-02
-1.63030684e-01 -5.67619622e-01 -1.57807365e-01 3.69125783e-01
-1.54156890e-02 4.01411653e-01 -1.26164928e-01 1.27737343e-01
1.01398003e+00 1.14491189e+00 9.95031238e-01 1.21542251e+00
1.74552843e-01 4.58658069e-01 9.34569478e-01 6.46143436e-01
2.52703816e-01 1.30318618e+00 6.73666596e-01 1.80173069e-01
7.35552236e-02 1.40281394e-01 3.89784366e-01 6.02665842e-01
-9.81151223e-01 -2.32470990e-03 -1.17329681e+00 4.30909187e-01
-1.92853856e+00 -1.25356948e+00 -2.05411345e-01 1.91088176e+00
3.13300699e-01 6.62711501e-01 9.56100464e-01 3.25358391e-01
4.48879659e-01 9.14717913e-02 3.27565297e-02 -1.11004138e+00
-3.71284261e-02 6.82669222e-01 3.39239359e-01 4.28847611e-01
-7.25426972e-01 1.32529163e+00 5.51389647e+00 1.61327720e+00
-7.70880044e-01 7.93208256e-02 9.11377147e-02 -2.97456443e-01
6.79398999e-02 2.72216767e-01 -4.51539308e-01 2.19739228e-01
8.28612804e-01 -7.42266655e-01 8.28229904e-01 7.99857438e-01
4.29227412e-01 -4.42208558e-01 -4.15393144e-01 8.79682958e-01
-2.03341827e-01 -1.82881677e+00 -4.47981924e-01 5.25279284e-01
4.10036445e-01 -2.09268238e-02 -3.56283039e-01 6.62499905e-01
7.90156543e-01 -1.18372250e+00 9.29922819e-01 -5.18814474e-03
4.21043217e-01 -9.21446919e-01 8.57187212e-01 4.17116493e-01
-1.34624755e+00 -2.80866653e-01 -6.13985546e-02 -1.33984721e+00
4.35134411e-01 -5.33230305e-01 -6.28552079e-01 8.24193418e-01
8.49610865e-01 1.11372039e-01 1.05905540e-01 1.15430689e+00
1.48432449e-01 4.41447794e-01 -5.96572280e-01 -3.56873512e-01
9.57743764e-01 -4.92725164e-01 1.08448279e+00 3.98485452e-01
5.78579586e-03 6.41913831e-01 3.20567012e-01 6.00023508e-01
5.10434866e-01 1.82110786e-01 -6.03380144e-01 -1.83040380e-01
1.98788717e-01 8.92038703e-01 -1.08084452e+00 1.58859700e-01
5.41747920e-02 5.78713238e-01 -1.17566392e-01 -3.40177089e-01
-9.11154091e-01 -6.43493891e-01 9.76118267e-01 5.23354053e-01
8.66983309e-02 -3.11614215e-01 -1.11563295e-01 -5.48124611e-01
-4.87168759e-01 -1.43615448e+00 4.69415367e-01 -1.02922881e+00
-7.66666234e-01 9.09242570e-01 4.95536596e-01 -1.09401715e+00
-7.94098675e-01 -5.89740038e-01 -1.04102659e+00 5.54368436e-01
-3.32382470e-01 -1.18364382e+00 -3.33986640e-01 6.22217000e-01
5.50307453e-01 -7.69746065e-01 7.62791991e-01 1.30680010e-01
-2.25149140e-01 2.97205567e-01 -2.52730876e-01 2.75487150e-03
1.59837797e-01 -6.94169462e-01 5.21055937e-01 5.21299481e-01
7.20215216e-02 -3.88386771e-02 7.81432986e-01 -5.78883648e-01
-1.32370818e+00 6.41428903e-02 4.14896727e-01 -2.22749203e-01
7.76504517e-01 -3.05621386e-01 -1.46604404e-01 5.29098868e-01
2.56269842e-01 -7.43872106e-01 5.10626972e-01 1.18202694e-01
2.54505098e-01 1.99381318e-02 -8.37461114e-01 8.44814301e-01
9.53821719e-01 -1.52571574e-01 -5.25157869e-01 1.44805044e-01
2.78954685e-01 -5.54263413e-01 -5.39168000e-01 4.86502945e-02
6.16516650e-01 -1.21294272e+00 7.45776832e-01 -7.26520598e-01
2.66415596e-01 -4.20118749e-01 -1.63917825e-01 -1.00299013e+00
-3.52528363e-01 -9.97807920e-01 9.00700390e-01 8.15140426e-01
5.34527376e-02 -6.00736201e-01 1.21937072e+00 7.04948545e-01
-1.89581394e-01 -5.07576704e-01 -1.32061613e+00 -9.17923629e-01
2.31947973e-01 -1.05184340e+00 3.90906334e-01 7.88650453e-01
5.23091376e-01 1.24671720e-01 -5.01804709e-01 -5.85548937e-01
9.03885141e-02 -3.25948924e-01 1.31162357e+00 -1.31367564e+00
-9.40798461e-01 -9.14351881e-01 -1.20453501e+00 -5.61761975e-01
-5.57873189e-01 -5.97839594e-01 -1.73072949e-01 -1.44491899e+00
-2.80137092e-01 -5.58611691e-01 2.52159774e-01 3.67079884e-01
3.88169557e-01 6.06273055e-01 8.39420140e-01 2.16352791e-01
-6.42292738e-01 1.43373916e-02 9.36270893e-01 8.37399215e-02
-4.60713506e-01 1.87256470e-01 -5.11145175e-01 6.48968041e-01
6.69655144e-01 -3.96599889e-01 -3.50136071e-01 -6.13930114e-02
8.92476380e-01 1.69553190e-01 8.22558627e-02 -1.60007751e+00
4.81142491e-01 -5.95397532e-01 -6.16514862e-01 3.92662473e-02
7.15024233e-01 -6.11496627e-01 6.47871435e-01 6.58844829e-01
1.60066366e-01 2.36364067e-01 6.96983635e-01 -3.22285830e-03
-4.36928809e-01 -3.78081352e-01 3.60459507e-01 -4.24389243e-01
-8.86958420e-01 -1.05686270e-01 -1.36190045e+00 3.17298025e-01
1.70159161e+00 -1.01946425e+00 -4.14464101e-02 -8.89331996e-01
-6.37196124e-01 2.73730278e-01 3.36872369e-01 1.61744967e-01
2.23454554e-02 -1.08160985e+00 -9.33320582e-01 -8.12753290e-02
-2.66774625e-01 -3.01319122e-01 7.10196495e-01 6.93969607e-01
-1.33734429e+00 2.14499384e-01 -1.08195829e+00 -4.70425710e-02
-1.40214753e+00 2.40636796e-01 4.45544362e-01 -8.55012536e-01
-6.14244580e-01 9.95874763e-01 1.40058771e-01 -2.80801624e-01
-1.51086330e-01 4.70598668e-01 -4.20111686e-01 -2.04959005e-01
7.62811005e-01 4.77730215e-01 -2.26885349e-01 -5.01147270e-01
-5.71912766e-01 1.85475796e-01 1.40764713e-01 -7.74959922e-01
1.33732975e+00 3.12946826e-01 4.12080586e-02 4.43095773e-01
2.33954579e-01 3.47277559e-02 -6.33903086e-01 3.83750439e-01
-1.45283118e-01 -4.42253053e-01 -1.50011525e-01 -9.44490671e-01
-8.08441341e-01 6.20224476e-01 3.43718588e-01 6.00245595e-01
1.02877378e+00 -2.75237381e-01 6.47175431e-01 3.97628546e-02
9.48112249e-01 -1.06002510e+00 4.98077422e-02 9.54832077e-01
8.46858501e-01 -4.62546349e-01 -1.61966354e-01 -2.26576388e-01
-1.24644136e+00 1.05531418e+00 8.16515625e-01 -4.93934691e-01
2.07967415e-01 5.72344482e-01 2.07431450e-01 -4.16924506e-01
-6.22966826e-01 -3.43860209e-01 -2.55182445e-01 9.30510938e-01
-3.24488878e-02 5.07881790e-02 -8.22390497e-01 1.11512661e+00
-6.67071104e-01 2.33929545e-01 7.03702927e-01 1.18918276e+00
-2.99664736e-01 -1.52889931e+00 -6.06002510e-01 1.04895495e-01
-2.15137944e-01 9.04056877e-02 -6.55373573e-01 1.28557742e+00
4.26183730e-01 9.15977776e-01 1.76675260e-01 -8.95747662e-01
5.69637477e-01 -3.28371078e-01 3.81198406e-01 -3.15515578e-01
-1.65545118e+00 -1.45388916e-01 5.86831093e-01 -6.39840901e-01
-1.62940565e-02 -6.05801582e-01 -1.04148483e+00 -9.58359599e-01
-1.45335004e-01 6.48898959e-01 6.08582258e-01 9.22629356e-01
2.50834882e-01 4.58887994e-01 7.15526119e-02 -1.14028835e+00
1.55928522e-01 -8.49469125e-01 -8.19992721e-01 6.06976636e-02
-1.00508583e+00 -9.12007272e-01 2.37567961e-01 -7.51429677e-01] | [3.462144136428833, 1.4778019189834595] |
7c75f70a-3b46-4da3-84d4-90d38c855865 | novel-modelling-strategies-for-high-frequency | 2212.00148 | null | https://arxiv.org/abs/2212.00148v1 | https://arxiv.org/pdf/2212.00148v1.pdf | Novel Modelling Strategies for High-frequency Stock Trading Data | Full electronic automation in stock exchanges has recently become popular, generating high-frequency intraday data and motivating the development of near real-time price forecasting methods. Machine learning algorithms are widely applied to mid-price stock predictions. Processing raw data as inputs for prediction models (e.g., data thinning and feature engineering) can primarily affect the performance of the prediction methods. However, researchers rarely discuss this topic. This motivated us to propose three novel modelling strategies for processing raw data. We illustrate how our novel modelling strategies improve forecasting performance by analyzing high-frequency data of the Dow Jones 30 component stocks. In these experiments, our strategies often lead to statistically significant improvement in predictions. The three strategies improve the F1 scores of the SVM models by 0.056, 0.087, and 0.016, respectively. | ['Li Xing', 'Ke Xu', 'Yuying Huang', 'Xuekui Zhang'] | 2022-11-30 | null | null | null | null | ['feature-engineering'] | ['methodology'] | [-4.65666920e-01 -2.25718454e-01 -1.42052963e-01 -4.13361102e-01
-2.77802378e-01 -6.10707700e-01 6.57206953e-01 1.75483927e-01
-2.08975911e-01 1.00527442e+00 -1.78736031e-01 -6.02820814e-01
2.27788538e-02 -1.14383113e+00 -5.52825689e-01 -4.70403135e-01
-3.04904848e-01 1.57835931e-01 1.30035460e-01 -3.25040489e-01
7.35663474e-01 4.99927282e-01 -1.43093204e+00 2.71617025e-01
8.21643293e-01 1.55499017e+00 -2.26739287e-01 1.44672036e-01
-3.08495134e-01 6.49121761e-01 -8.14319313e-01 -6.93559289e-01
7.54639745e-01 -1.48271397e-01 -7.74414837e-02 -3.57562378e-02
-2.65572011e-01 -4.56141710e-01 1.05873168e-01 1.03864181e+00
4.96057197e-02 -1.40465247e-02 4.15428340e-01 -1.32747126e+00
-3.86376053e-01 1.02400005e+00 -7.70715237e-01 3.56988698e-01
-1.32978067e-01 -5.92656918e-02 1.25800776e+00 -9.94550407e-01
5.59494831e-02 6.99485958e-01 6.39614642e-01 2.59401798e-02
-1.25440192e+00 -8.36708426e-01 1.30135402e-01 1.48126855e-01
-1.02474427e+00 -5.45638725e-02 9.06443655e-01 -4.42408800e-01
8.79391968e-01 3.46903294e-01 8.83138359e-01 4.09630716e-01
6.11221671e-01 7.71037757e-01 1.21542835e+00 -2.60637373e-01
4.17694718e-01 3.78968567e-01 3.57143253e-01 -9.50345993e-02
7.16444612e-01 3.80204260e-01 -4.83513743e-01 -3.95969123e-01
7.31827259e-01 1.88342065e-01 1.19567737e-01 1.48165166e-01
-9.64901865e-01 1.07722080e+00 3.55397128e-02 7.34958425e-02
-7.15002000e-01 -3.48872453e-01 2.52111912e-01 7.06973135e-01
8.08644772e-01 6.26216352e-01 -9.77502704e-01 -3.17549348e-01
-9.50722098e-01 6.74575806e-01 9.96707022e-01 7.81451285e-01
3.42109770e-01 4.41458255e-01 3.88248175e-01 5.79593718e-01
1.37765810e-01 3.16818535e-01 6.49272084e-01 -4.38316882e-01
4.20660913e-01 5.51876545e-01 4.24086511e-01 -8.18887055e-01
-5.49651027e-01 -5.85254133e-01 -6.66726232e-01 2.38114536e-01
5.54218769e-01 -4.38517332e-01 -4.61423099e-01 9.89831448e-01
2.42291689e-02 9.50272474e-03 3.17561597e-01 5.72970867e-01
1.03497468e-01 8.00826371e-01 -4.57965061e-02 -5.70280254e-01
1.05554688e+00 -8.12191546e-01 -8.15277755e-01 1.89493716e-01
5.90912044e-01 -1.01655114e+00 6.67511523e-01 8.37892354e-01
-8.73530447e-01 -6.27960503e-01 -9.87926960e-01 6.43849373e-01
-4.45505857e-01 -1.08327992e-01 6.37334168e-01 5.43191195e-01
-2.95588076e-01 9.86138105e-01 -6.70021534e-01 5.12079418e-01
6.91756606e-02 3.88676792e-01 1.80035144e-01 7.32320964e-01
-1.42745280e+00 9.55579579e-01 5.82395554e-01 1.00907804e-02
2.13461921e-01 -9.51987207e-01 -2.94065535e-01 1.11940362e-01
1.40402153e-01 -1.20207764e-01 1.45402396e+00 -7.86450505e-01
-1.49676406e+00 1.35409579e-01 3.04601014e-01 -1.12088990e+00
6.50376618e-01 -4.41080719e-01 -1.02163982e+00 -4.18633908e-01
-3.54523987e-01 2.39068434e-01 8.47825527e-01 -5.78220665e-01
-1.19475305e+00 -1.86033770e-01 -4.26410943e-01 -2.92173982e-01
-1.36824399e-01 -8.28634873e-02 4.79390293e-01 -1.29380214e+00
2.96229184e-01 -6.75867379e-01 -2.26150781e-01 -5.52656054e-01
-2.85762809e-02 -2.21453547e-01 5.42801380e-01 -8.52435350e-01
1.26882100e+00 -2.01530886e+00 -7.23164678e-01 5.81050158e-01
-1.83154792e-01 7.78550729e-02 3.80820274e-01 3.82437468e-01
-3.82606775e-01 7.11177066e-02 4.62607630e-02 2.30604380e-01
2.16187611e-01 -6.00695796e-02 -1.06671464e+00 1.57317236e-01
4.16345775e-01 9.29064631e-01 -2.63644367e-01 1.47981808e-01
2.99995661e-01 5.71833849e-02 -3.15778881e-01 1.92871958e-01
-3.47954601e-01 -2.42614597e-02 -2.67894238e-01 4.89420652e-01
7.86458969e-01 -1.31630644e-01 5.23764465e-04 -3.20953801e-02
-4.13377225e-01 4.47158992e-01 -1.26894474e+00 4.61791396e-01
-1.36974499e-01 3.50911707e-01 -5.95798135e-01 -8.82383525e-01
1.57213652e+00 1.06240623e-01 5.58615088e-01 -8.76762033e-01
-4.71582897e-02 5.47648311e-01 2.10498929e-01 1.10493772e-01
7.04283714e-01 -4.86897916e-01 9.24288481e-03 4.57924992e-01
-5.92557073e-01 7.52622262e-02 2.40867183e-01 -4.50533718e-01
2.97435105e-01 -2.23143503e-01 4.57591385e-01 -4.38258767e-01
3.29128921e-01 2.32906207e-01 7.74251103e-01 3.34221780e-01
-8.34010728e-03 2.57504910e-01 5.43165684e-01 -1.00318766e+00
-9.60616231e-01 -7.44393051e-01 -4.50641870e-01 7.06786811e-01
-2.16539383e-01 -2.64232665e-01 -4.23192441e-01 -4.03239578e-01
7.58049548e-01 1.20824265e+00 -4.11062777e-01 -4.51265611e-02
-3.34324360e-01 -1.03546476e+00 -6.83581233e-02 7.10916042e-01
3.95406276e-01 -1.02010822e+00 -5.49662828e-01 6.90363407e-01
4.28195715e-01 -9.13027287e-01 -7.75998980e-02 3.23375106e-01
-1.24619091e+00 -8.36182892e-01 -6.79422855e-01 -2.92749405e-01
3.79792124e-01 -7.94765204e-02 1.00005734e+00 -1.57433171e-02
2.16997117e-01 -4.18760389e-01 -4.15248603e-01 -9.15936708e-01
-1.78851411e-01 1.53269678e-01 3.17624032e-01 2.66599804e-01
7.82663167e-01 -3.24192852e-01 -2.16381475e-01 5.46493351e-01
-6.20530367e-01 -4.96111251e-02 5.20479679e-01 8.96997750e-01
5.37864983e-01 2.85226285e-01 1.04016864e+00 -9.42949653e-01
8.60026658e-01 -4.32029724e-01 -1.53142989e+00 1.83998659e-01
-1.38863432e+00 1.58730179e-01 8.34683895e-01 -4.12185937e-01
-8.66922021e-01 -7.51665384e-02 -4.17741202e-02 -2.33069018e-01
9.83249471e-02 7.93247402e-01 2.78245717e-01 8.32283571e-02
2.59342670e-01 3.94539565e-01 2.62123734e-01 -7.72300720e-01
-8.47896039e-02 5.65026224e-01 1.81226432e-01 -1.65446937e-01
9.64483142e-01 -7.64400586e-02 -4.77876887e-02 -5.99887669e-01
-5.74371457e-01 -2.26107333e-02 -6.13320053e-01 1.42725870e-01
2.74093002e-01 -7.93109536e-01 -6.40967786e-01 7.09258676e-01
-6.84308112e-01 -7.63196126e-02 -3.73237461e-01 7.15042293e-01
-3.74568135e-01 1.25267580e-02 -7.18621194e-01 -1.00933421e+00
-4.54401106e-01 -7.30938137e-01 4.14759576e-01 1.92498073e-01
-5.46523750e-01 -1.04682100e+00 -1.47576690e-01 -3.26749273e-02
6.55032456e-01 1.33895546e-01 8.40433657e-01 -1.31058562e+00
-4.74561512e-01 -5.52074134e-01 6.17551524e-03 4.09059525e-01
3.10078561e-01 2.06353277e-01 -8.10784280e-01 -6.79837838e-02
3.28063011e-01 2.49175891e-01 5.00120819e-01 3.58670682e-01
9.93415534e-01 -4.17424381e-01 1.52742453e-02 3.19625437e-01
9.43137765e-01 8.22899163e-01 4.23181713e-01 7.20557630e-01
6.10695332e-02 7.00518250e-01 1.09740543e+00 1.03315759e+00
-8.76021385e-02 4.64060456e-01 -1.11814626e-01 2.57730991e-01
8.19268703e-01 -3.38472843e-01 2.87636071e-01 8.31779838e-01
-2.15395130e-02 4.81576383e-01 -8.58638048e-01 6.49904087e-02
-1.70444238e+00 -9.42699790e-01 -5.59170060e-02 2.11002088e+00
8.91491115e-01 7.44628251e-01 3.26748967e-01 4.51127946e-01
5.65203965e-01 -2.61977520e-02 -5.89347959e-01 -2.58025348e-01
-8.13881457e-02 1.49320751e-01 6.45757139e-01 3.12461168e-01
-1.19514382e+00 5.71562111e-01 6.05439615e+00 6.14815593e-01
-1.29405570e+00 -6.05465055e-01 1.03222835e+00 -6.75735436e-03
-3.45983863e-01 -1.71807408e-01 -1.26705992e+00 9.17134225e-01
1.22360969e+00 -9.20361638e-01 1.53677493e-01 1.27257252e+00
3.34158778e-01 1.15972735e-01 -9.95806456e-01 1.07690668e+00
-4.40687358e-01 -1.70993209e+00 1.89315662e-01 3.80693436e-01
7.00885773e-01 -4.17363256e-01 9.76988450e-02 4.91817147e-01
5.94551861e-02 -7.69000888e-01 7.51585186e-01 5.63546717e-01
7.68770427e-02 -1.17149329e+00 1.11948490e+00 4.69645917e-01
-9.84723747e-01 -1.60468146e-01 -3.81645828e-01 -5.79890847e-01
2.79502928e-01 1.05098319e+00 -6.44616067e-01 4.71678972e-01
5.13055325e-01 6.61416173e-01 -3.85695696e-01 1.10997617e+00
8.87521952e-02 7.46074080e-01 -4.58131254e-01 -3.62459034e-01
6.27690852e-02 -5.12172401e-01 -2.21940465e-02 4.82362032e-01
3.47301841e-01 3.11882973e-01 2.03013942e-02 7.36807048e-01
2.17122689e-01 1.67056754e-01 -2.56666303e-01 -1.01297580e-01
3.11744452e-01 8.05982292e-01 -6.03374183e-01 -4.29823339e-01
-7.16850102e-01 2.85182714e-01 -3.27247441e-01 1.68018937e-01
-7.50971496e-01 -5.44120252e-01 7.11069405e-01 4.18762356e-01
4.03490335e-01 -2.16791853e-01 -1.01073694e+00 -1.07567894e+00
2.00934798e-01 -9.25193906e-01 3.14926744e-01 -1.79529324e-01
-1.68493688e+00 5.39166749e-01 -7.30422363e-02 -1.69783497e+00
-6.90065682e-01 -7.69478858e-01 -6.75305724e-01 1.00630724e+00
-1.53757071e+00 -1.97725520e-01 2.15280175e-01 3.14176023e-01
5.00849247e-01 -7.40936935e-01 6.20585740e-01 9.61991921e-02
-4.57322717e-01 4.78395283e-01 3.15528244e-01 2.87450284e-01
3.03268284e-01 -1.09768140e+00 1.15512609e+00 8.01911652e-01
3.24538529e-01 7.15446651e-01 6.52746558e-01 -7.99101710e-01
-1.15079057e+00 -9.46007490e-01 1.11768448e+00 6.35870993e-02
1.10283303e+00 -7.28321895e-02 -1.20209837e+00 5.83467126e-01
-2.91670784e-02 -1.49062708e-01 7.25809276e-01 -2.00922694e-02
-1.03508748e-01 -5.43773532e-01 -1.13922048e+00 4.22266454e-01
1.04527451e-01 -1.36510253e-01 -9.95342076e-01 -1.51426708e-02
4.37449306e-01 -1.44048557e-01 -1.10149169e+00 4.08831656e-01
5.18554389e-01 -8.35936248e-01 6.15349650e-01 -7.04946160e-01
1.31584287e-01 -5.69379888e-02 -2.00302564e-02 -1.31920791e+00
-2.53768682e-01 -8.67628098e-01 -4.39642310e-01 8.88526380e-01
8.55471134e-01 -1.30193913e+00 8.59140515e-01 1.18154550e+00
1.84697032e-01 -7.67485023e-01 -6.77374780e-01 -9.94684875e-01
2.29755178e-01 -4.87116128e-01 1.11667848e+00 9.72326815e-01
2.17672706e-01 -1.51126245e-02 -2.78109819e-01 -6.13403395e-02
6.03534400e-01 8.03435385e-01 8.41076434e-01 -1.61988127e+00
-4.00921494e-01 -6.57153428e-01 -2.64613479e-01 -1.04612768e+00
-1.16470493e-01 -3.37689579e-01 -5.92786968e-01 -7.04784751e-01
-5.21610022e-01 -3.51104975e-01 -4.60571885e-01 3.45976353e-01
-1.03921548e-01 -1.41295597e-01 2.83361673e-01 4.00304288e-01
3.76328349e-01 4.88539189e-01 7.63135016e-01 1.38933554e-01
-3.33047092e-01 7.75932133e-01 -5.87967813e-01 6.33403480e-01
1.50954461e+00 -2.18293399e-01 -2.53914922e-01 2.50952661e-01
1.58747986e-01 -3.66835445e-02 -5.24222739e-02 -7.49830365e-01
-2.29046890e-03 -3.63883913e-01 7.19726920e-01 -1.00086188e+00
-4.12987284e-02 -9.07213509e-01 3.49568993e-01 8.38511944e-01
-1.86736122e-01 6.06045842e-01 3.48609567e-01 2.74426997e-01
-6.79599106e-01 -1.95109725e-01 5.79882920e-01 3.59047949e-02
-8.03777397e-01 5.20520657e-02 -2.00944349e-01 -1.70242861e-01
1.25030077e+00 -3.19717348e-01 -1.08984210e-01 -2.96450257e-01
-7.67409384e-01 1.47861734e-01 -1.32131791e-02 6.12771928e-01
4.24148917e-01 -1.29580486e+00 -6.83524013e-01 8.68643105e-01
-2.00169832e-01 -4.41845536e-01 -3.69316727e-01 6.04354143e-01
-5.10495603e-01 6.82934105e-01 -2.99062192e-01 -2.04659864e-01
-9.53917205e-01 4.78601456e-01 1.24466673e-01 -1.80919752e-01
-6.37240827e-01 5.04366398e-01 -3.00250888e-01 3.23601067e-02
8.87741670e-02 -8.87928724e-01 -1.39159903e-01 4.42511559e-01
7.65587866e-01 5.35623968e-01 4.36863899e-01 -1.19691445e-02
-1.34242401e-01 4.14445877e-01 -4.26762909e-01 -7.91806653e-02
1.64911246e+00 3.27191859e-01 7.29541779e-02 7.38775551e-01
8.09424162e-01 -7.28628188e-02 -1.26784432e+00 -3.03098019e-02
7.97932744e-01 -6.52165651e-01 -2.04112738e-01 -5.56174159e-01
-1.22619104e+00 6.66554928e-01 2.78253585e-01 1.02463007e+00
1.02166390e+00 -4.50994134e-01 1.01615763e+00 4.87489343e-01
6.93364978e-01 -1.20875764e+00 -5.46441555e-01 4.81304973e-01
8.54377747e-01 -1.10585928e+00 5.81859536e-02 -3.89293253e-01
-7.80717194e-01 1.34965837e+00 2.22168192e-01 -2.84872293e-01
1.06982112e+00 5.65770030e-01 3.32574517e-01 2.88761199e-01
-1.07954919e+00 4.01661932e-01 3.01922351e-01 6.33800849e-02
3.61430317e-01 2.87772030e-01 -6.44345522e-01 1.23127425e+00
-7.84797847e-01 1.03804499e-01 5.37477970e-01 8.32461417e-01
-6.45604730e-01 -1.23491156e+00 -4.71809834e-01 9.52437460e-01
-6.69770479e-01 -1.92991033e-01 -1.94016844e-01 1.02973914e+00
-5.45729339e-01 6.15080595e-01 5.01141608e-01 -4.91647959e-01
5.94118714e-01 5.28012216e-01 -3.03119838e-01 -2.67927974e-01
-6.89466536e-01 3.21205139e-01 -3.22449133e-02 -2.07394466e-01
3.56445089e-02 -1.01985109e+00 -1.26660681e+00 -6.13668978e-01
-3.23434919e-01 4.80361313e-01 4.38223898e-01 5.98336518e-01
3.38549972e-01 7.47096911e-02 1.05117738e+00 -5.17944098e-01
-1.47457850e+00 -9.04446840e-01 -1.11443245e+00 3.59406471e-01
1.11895405e-01 -6.36274815e-01 -5.11674702e-01 1.47419900e-01] | [4.56086540222168, 4.176708221435547] |
b91e6a00-35ae-459f-9404-de0b6337ca8d | attention-based-residual-autoencoder-for | null | null | https://link.springer.com/article/10.1007/s10489-022-03613-1 | https://link.springer.com/content/pdf/10.1007/s10489-022-03613-1.pdf | Attention-based residual autoencoder for video anomaly detection | Automatic anomaly detection is a crucial task in video surveillance system intensively used for public safety and others. The present system adopts a spatial branch and a temporal branch in a unified network that exploits both spatial and temporal information effectively. The network has a residual autoencoder architecture, consisting of a deep convolutional neural network-based encoder and a multi-stage channel attention-based decoder, trained in an unsupervised manner. The temporal shift method is used for exploiting the temporal feature, whereas the contextual dependency is extracted by channel attention modules. System performance is evaluated using three standard benchmark datasets. Result suggests that our network outperforms the state-of-the-art methods, achieving 97.4% for UCSD Ped2, 86.7% for CUHK Avenue, and 73.6% for ShanghaiTech dataset in term of Area Under Curve, respectively. | ['Yong-Guk Kim', 'Viet-Tuan Le'] | 2022-05-25 | null | null | null | applied-intelligence-2022-5 | ['anomaly-detection-in-surveillance-videos', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'methodology'] | [-1.28465250e-01 -3.09392333e-01 -9.83524993e-02 -1.99896842e-01
-3.25606346e-01 7.30885640e-02 4.75912958e-01 -2.51870304e-01
-7.17683613e-01 4.19387996e-01 2.76345581e-01 -3.22881460e-01
3.21776628e-01 -5.09164393e-01 -7.25826979e-01 -9.83865440e-01
-2.77485490e-01 -4.68319505e-01 6.10091448e-01 -1.20037302e-01
-3.38056870e-03 2.63540775e-01 -1.19833565e+00 1.61936909e-01
5.06255388e-01 1.51740766e+00 -1.97141811e-01 8.27122808e-01
2.72778243e-01 1.12817025e+00 -4.51675415e-01 -2.92046666e-01
4.04508114e-01 -3.03394794e-01 -3.50695550e-01 7.90289566e-02
1.09411523e-01 -9.05785382e-01 -1.03012323e+00 1.09376025e+00
3.49667698e-01 -5.59230782e-02 5.69606245e-01 -1.06016862e+00
-6.52756274e-01 6.02435507e-03 -7.51248240e-01 7.98790932e-01
-2.06614792e-01 3.38900089e-01 6.59249187e-01 -4.99196619e-01
1.43438935e-01 7.94597745e-01 6.81982934e-01 4.56261009e-01
-6.27434254e-01 -6.30256772e-01 3.32364500e-01 5.99207699e-01
-1.30613303e+00 -1.62145808e-01 6.91172779e-01 -6.44266605e-01
1.15124834e+00 -1.89762041e-01 7.13031232e-01 1.27276552e+00
6.71042800e-01 7.73708522e-01 5.06575882e-01 -3.59296128e-02
-3.07875816e-02 -1.74528897e-01 1.64073467e-01 8.53465259e-01
8.48770514e-02 4.13607806e-01 -7.64482329e-03 -3.70400958e-02
7.75390267e-01 4.16653961e-01 -2.86107540e-01 1.36232197e-01
-9.26298440e-01 9.16293740e-01 5.03639460e-01 2.23199591e-01
-6.07798994e-01 8.05047974e-02 8.08007061e-01 4.61429745e-01
4.99052703e-01 -5.59643900e-04 -6.81389153e-01 -1.32827507e-02
-7.24187076e-01 7.11681275e-03 3.19982171e-01 1.06058359e+00
2.88464844e-01 5.04434645e-01 -4.20258373e-01 5.58491886e-01
5.54189563e-01 3.61509502e-01 7.33769178e-01 -6.54584765e-01
1.30549192e-01 5.76767087e-01 -2.05334410e-01 -9.94311154e-01
-4.32419181e-01 -5.24128377e-01 -1.28814840e+00 8.60832110e-02
1.19707145e-01 -5.87380648e-01 -1.01836252e+00 1.56966472e+00
6.62733689e-02 6.52236104e-01 3.01403731e-01 8.78675282e-01
7.98751175e-01 9.74403858e-01 2.98810899e-01 -2.91184515e-01
1.37328744e+00 -1.34656930e+00 -8.23012471e-01 -5.49263693e-02
5.68737924e-01 -4.89980042e-01 2.28520572e-01 4.58716571e-01
-7.50950813e-01 -7.24325180e-01 -1.15178895e+00 4.71906960e-02
-4.44386125e-01 5.63553333e-01 2.99794227e-01 2.73775667e-01
-9.19721007e-01 1.21349886e-01 -7.29925275e-01 -3.58857244e-01
6.10212088e-01 1.92771241e-01 -4.70647156e-01 1.73439011e-01
-1.27033317e+00 5.28322399e-01 5.05799174e-01 4.14769411e-01
-1.08730936e+00 -2.51046211e-01 -9.70758736e-01 3.01090591e-02
1.79640099e-01 -4.08555716e-01 1.25248110e+00 -1.18550313e+00
-1.50538695e+00 6.46104515e-01 -5.65853678e-02 -8.83784711e-01
3.05304050e-01 -4.59342808e-01 -9.78934050e-01 2.43843168e-01
-1.26323784e-02 4.17898059e-01 9.72863734e-01 -4.47717696e-01
-1.12013006e+00 -3.32990587e-01 -1.72726810e-01 6.02376531e-04
-3.12627226e-01 1.72860980e-01 -7.52822280e-01 -8.54445040e-01
-1.34084135e-01 -7.31754005e-01 -3.69363070e-01 3.74805629e-02
-3.30336571e-01 -1.47731245e-01 1.45120442e+00 -1.19833553e+00
1.56286895e+00 -2.36573482e+00 -9.77202412e-03 5.84748946e-03
2.13343352e-01 6.33385420e-01 1.35896087e-01 -1.69681255e-02
-1.99434817e-01 -1.77510768e-01 -3.02544147e-01 -3.64781879e-02
-4.79308456e-01 -3.34197730e-02 -8.71540383e-02 5.66036701e-01
4.49301720e-01 6.89612329e-01 -6.81862712e-01 -3.02973330e-01
1.69438332e-01 3.66951078e-01 -5.26358187e-01 5.27638614e-01
3.52268219e-02 5.13211548e-01 -5.44103205e-01 6.54628456e-01
5.72341084e-01 -7.22245574e-02 -3.43860567e-01 -1.03450298e-01
-3.90143991e-01 -2.22666070e-01 -7.26801157e-01 1.54276979e+00
4.67579812e-02 1.01803911e+00 -1.41464975e-02 -1.02330601e+00
8.39088678e-01 6.91819906e-01 4.39950436e-01 -8.24911833e-01
4.03439224e-01 1.20469294e-01 8.29322115e-02 -1.02256513e+00
3.91385764e-01 3.72397393e-01 6.64771572e-02 -1.86109737e-01
2.32981354e-01 9.51350272e-01 -1.07381016e-01 -8.23685376e-04
1.36104119e+00 -6.88035488e-02 2.95424104e-01 -3.68971199e-01
8.47513616e-01 -2.23509923e-01 8.22203159e-01 4.82874840e-01
-7.09900618e-01 4.61891323e-01 6.95559442e-01 -8.54305446e-01
-1.12285388e+00 -7.07091868e-01 1.43487426e-03 7.56198049e-01
1.19332612e-01 -1.53973639e-01 -6.64086640e-01 -1.00869286e+00
-2.55651236e-01 3.32920164e-01 -8.79855156e-01 -5.11738241e-01
-5.93079865e-01 -7.65828133e-01 6.19802415e-01 8.77243161e-01
1.26084316e+00 -1.12580740e+00 -7.11953461e-01 1.96834728e-01
8.03702418e-03 -1.22556138e+00 -4.08732206e-01 -1.45566434e-01
-6.90605998e-01 -1.28296101e+00 -8.19901526e-01 -7.28940666e-01
4.85794008e-01 1.64905250e-01 7.88043499e-01 1.57735869e-02
-2.88735181e-01 2.78313935e-01 -4.16485846e-01 -5.07864892e-01
1.22079119e-01 6.69532269e-02 -3.04046541e-01 4.41707134e-01
6.88524365e-01 -2.69034177e-01 -8.65315735e-01 1.98234782e-01
-8.49800050e-01 -2.20724121e-01 8.31889570e-01 1.09277880e+00
2.45148152e-01 -6.50801184e-03 4.15205151e-01 -5.30383050e-01
2.44767070e-01 -8.31480324e-01 -8.59201670e-01 -1.24503888e-01
-2.90760875e-01 -1.35794535e-01 6.80988848e-01 -8.34895521e-02
-1.06793213e+00 4.87611182e-02 -3.15516829e-01 -6.23160303e-01
-4.02839243e-01 1.33524284e-01 -1.17308393e-01 1.54815510e-01
2.97589272e-01 5.64313710e-01 -6.50067627e-02 -2.92343050e-01
-3.15975875e-01 8.32310319e-01 7.17754960e-01 -1.69767402e-02
4.60997224e-01 3.81027997e-01 -2.19766781e-01 -1.03626728e+00
-6.10843003e-01 -5.55101752e-01 -5.23530781e-01 -2.54302025e-01
1.67177057e+00 -1.24707222e+00 -4.78033394e-01 9.93153751e-01
-1.35837996e+00 -6.70957789e-02 4.38373685e-02 7.82770991e-01
-1.89914078e-01 1.68567315e-01 -7.32150316e-01 -5.75735867e-01
-4.66108471e-01 -1.37576401e+00 1.04059923e+00 4.47204441e-01
5.29100657e-01 -8.46875370e-01 2.09432226e-02 -6.42307922e-02
5.89406490e-01 3.64284277e-01 5.49261928e-01 -7.06021428e-01
-6.73083663e-01 -4.23191309e-01 -4.89338160e-01 7.53360629e-01
-1.73570842e-01 1.09807007e-01 -1.11311328e+00 -2.56997555e-01
-1.24919459e-01 8.37562755e-02 1.21053624e+00 7.01477528e-01
1.64555383e+00 -4.10819232e-01 -3.17930013e-01 8.43406558e-01
1.37400174e+00 7.88711727e-01 1.04456937e+00 3.28838080e-01
6.84204042e-01 8.52530971e-02 2.68300354e-01 4.01010960e-01
3.12879831e-01 4.45810884e-01 5.79943657e-01 -9.72268134e-02
3.02284330e-01 1.53910920e-01 4.65069979e-01 5.77737570e-01
-2.29487225e-01 -4.03523088e-01 -8.12387705e-01 5.94074368e-01
-2.14341426e+00 -1.19756603e+00 -2.10117206e-01 1.98659754e+00
2.39486650e-01 1.80525988e-01 1.05689064e-01 5.75407734e-03
6.50896549e-01 3.61785144e-01 -3.11930001e-01 -3.74019772e-01
-6.62426874e-02 -2.47062340e-01 7.08998799e-01 1.21676564e-01
-1.81276476e+00 8.03538859e-01 5.98178101e+00 4.69032705e-01
-1.17627835e+00 1.93698239e-03 8.56999874e-01 2.12149113e-01
6.28236532e-01 -3.93328965e-01 -5.57387292e-01 8.52499604e-01
1.06741464e+00 4.51928616e-01 -8.27584788e-03 9.72530484e-01
8.79101083e-02 3.02908979e-02 -7.11697042e-01 1.04453838e+00
4.91962768e-02 -1.08077538e+00 -1.45539269e-01 6.06479961e-03
5.26147306e-01 2.39641652e-01 1.04109563e-01 4.73256588e-01
-1.48425296e-01 -1.01941609e+00 4.23955709e-01 7.95142531e-01
6.10221207e-01 -8.40800643e-01 1.29379618e+00 1.49863422e-01
-1.35737884e+00 -2.98241913e-01 -2.02916652e-01 -4.66484902e-03
-7.77668059e-02 2.98878521e-01 -2.30192780e-01 5.01077354e-01
1.15492654e+00 1.12186587e+00 -4.61130798e-01 1.13091505e+00
-1.55129194e-01 8.31588447e-01 -5.45507111e-02 2.65631914e-01
6.93401158e-01 -1.05905429e-01 5.78958035e-01 1.52408087e+00
2.25866392e-01 2.88414598e-01 9.41906571e-02 2.47053966e-01
2.88422708e-03 -1.57315761e-01 -8.07369709e-01 2.77122438e-01
8.09255838e-02 1.10196841e+00 -7.69556239e-02 -4.58401620e-01
-8.56828511e-01 1.22544146e+00 -3.28332968e-02 4.51203525e-01
-1.23495209e+00 -6.73036873e-01 6.25368714e-01 -4.12237495e-01
6.43940926e-01 -8.08829069e-02 1.24691255e-01 -1.18376172e+00
1.08778790e-01 -7.90866911e-01 5.88499784e-01 -6.04981840e-01
-1.03447676e+00 7.71941960e-01 -1.46550894e-01 -1.49754095e+00
-1.34814933e-01 -8.28597128e-01 -9.01240885e-01 6.82380557e-01
-1.67353106e+00 -1.01103199e+00 -6.38286412e-01 7.54274070e-01
7.84198403e-01 -7.73994386e-01 6.20865881e-01 6.09623313e-01
-1.29013658e+00 7.45295227e-01 1.71997517e-01 6.87965631e-01
4.51801628e-01 -1.02594423e+00 1.92855060e-01 1.30654383e+00
-3.85344565e-01 8.92308578e-02 2.17529863e-01 -4.14175898e-01
-1.07633877e+00 -1.51040959e+00 6.26196682e-01 -2.66400054e-02
3.99673015e-01 1.58990100e-01 -9.50426519e-01 1.00800097e+00
6.54454887e-01 4.40720767e-01 4.47125047e-01 -3.10367316e-01
-3.64560962e-01 -1.66603059e-01 -1.07762933e+00 4.61465687e-01
8.06484580e-01 -2.88051188e-01 -3.25188369e-01 -7.12708384e-02
6.59154117e-01 -4.54425603e-01 -7.86319017e-01 5.27154386e-01
4.96122092e-01 -1.08702826e+00 6.01613104e-01 -7.34586835e-01
6.60187125e-01 -4.05933172e-01 -1.32097840e-01 -1.08381474e+00
-6.27556682e-01 -4.26020205e-01 -4.72132951e-01 7.42827535e-01
4.00193810e-01 -7.43814111e-01 4.88269389e-01 2.12859705e-01
-5.10388970e-01 -9.86094534e-01 -7.69515872e-01 -5.45027018e-01
-3.50551844e-01 -3.25917274e-01 3.79529059e-01 7.34209478e-01
-5.50366759e-01 2.85758346e-01 -7.93483913e-01 4.98340189e-01
4.23503071e-01 -6.07518435e-01 5.23201287e-01 -1.03119278e+00
-1.00144133e-01 -4.85099018e-01 -8.57802689e-01 -1.18472183e+00
-2.39737816e-02 -2.11585656e-01 -1.55656084e-01 -9.00741577e-01
8.12277123e-02 1.55686051e-01 -6.49653971e-01 3.36914986e-01
-1.56584784e-01 3.13851982e-02 -2.18566760e-01 -4.05412577e-02
-7.94887722e-01 8.01953852e-01 8.02147925e-01 -1.71058625e-01
-9.55167264e-02 1.23426141e-02 -3.87966812e-01 1.01908481e+00
1.07821488e+00 -1.62354827e-01 -1.55123055e-01 -6.86798155e-01
-4.64472204e-01 5.27411653e-03 5.64284623e-01 -1.35957885e+00
2.66656458e-01 7.24554360e-02 7.47955203e-01 -6.59121752e-01
8.12321156e-02 -1.13514066e+00 -3.87497753e-01 6.52374089e-01
-2.25293517e-01 4.32567745e-01 4.36635315e-01 9.63321507e-01
-4.80559498e-01 2.34507978e-01 8.47385108e-01 7.75577128e-02
-1.22524679e+00 7.65631318e-01 -5.20673156e-01 -1.07836671e-01
1.40613854e+00 -9.21460465e-02 -2.75737524e-01 -2.45232239e-01
-4.71351117e-01 3.98901314e-01 -1.24067947e-01 5.59131742e-01
7.53314257e-01 -1.50126898e+00 -8.39292467e-01 5.92218578e-01
2.64905661e-01 -1.53332353e-01 5.19113779e-01 9.98451352e-01
-6.71639502e-01 5.74980974e-01 -4.31245476e-01 -6.36968136e-01
-9.84977603e-01 5.12229085e-01 6.28816605e-01 -3.44835579e-01
-7.64817119e-01 7.73899972e-01 3.40933204e-01 3.02281946e-01
5.37681460e-01 -3.04004699e-01 -5.53358734e-01 -3.70834887e-01
7.72653043e-01 4.40624624e-01 -1.09134085e-01 -6.96736157e-01
-3.00045699e-01 4.64476466e-01 -2.19258279e-01 2.00998753e-01
1.05505896e+00 9.02514607e-02 1.17942899e-01 1.38838261e-01
1.42138541e+00 -4.92182583e-01 -1.46293437e+00 -3.33042294e-01
-1.72856808e-01 -1.56829312e-01 3.51338923e-01 -5.72297275e-01
-1.47597075e+00 8.69521260e-01 9.42976713e-01 2.49608189e-01
1.63395333e+00 -4.05942649e-01 1.11836255e+00 3.41629922e-01
-1.86476737e-01 -9.24493551e-01 -8.47054645e-04 8.03841650e-01
8.11884582e-01 -1.66098189e+00 -2.80319184e-01 9.20581967e-02
-7.38426685e-01 1.11098552e+00 1.02192116e+00 -4.51179057e-01
1.12615609e+00 6.19900711e-02 -1.38758942e-01 -1.76646248e-01
-9.52250183e-01 -1.65889964e-01 5.25600910e-01 3.65754545e-01
3.81385386e-01 -1.84982479e-01 -1.65473983e-01 7.01742947e-01
4.41835225e-01 -1.59265310e-01 9.40228999e-02 9.08463001e-01
-4.78686631e-01 -3.51705730e-01 -8.71335417e-02 5.18107593e-01
-7.80279577e-01 2.94879880e-02 7.70476535e-02 7.94571757e-01
4.38819021e-01 7.41915464e-01 4.83856112e-01 -7.31695890e-01
3.03607702e-01 -1.41284280e-02 -1.70707345e-01 -1.31770633e-02
-5.47115088e-01 1.17520504e-01 -6.10700697e-02 -8.48749578e-01
-3.11089694e-01 -3.98455918e-01 -8.68470788e-01 -1.94480926e-01
-3.93684730e-02 2.28725141e-03 3.50721449e-01 9.38720524e-01
5.27577519e-01 9.60053146e-01 6.96459651e-01 -4.95818287e-01
-1.33851752e-01 -1.07216811e+00 -2.12474301e-01 1.49000108e-01
9.49249744e-01 -4.70882177e-01 -1.32043019e-01 1.71672761e-01] | [7.891859531402588, 1.5032962560653687] |
aeacff06-9abe-4aa6-81f9-08621301eb48 | cctc-a-cross-sentence-chinese-text-correction | null | null | https://aclanthology.org/2022.coling-1.294 | https://aclanthology.org/2022.coling-1.294.pdf | CCTC: A Cross-Sentence Chinese Text Correction Dataset for Native Speakers | The Chinese text correction (CTC) focuses on detecting and correcting Chinese spelling errors and grammatical errors. Most existing datasets of Chinese spelling check (CSC) and Chinese grammatical error correction (GEC) are focused on a single sentence written by Chinese-as-a-second-language (CSL) learners. We find that errors caused by native speakers differ significantly from those produced by non-native speakers. These differences make it inappropriate to use the existing test sets directly to evaluate text correction systems for native speakers. Some errors also require the cross-sentence information to be identified and corrected. In this paper, we propose a cross-sentence Chinese text correction dataset for native speakers. Concretely, we manually annotated 1,500 texts written by native speakers. The dataset consists of 30,811 sentences and more than 1,000,000 Chinese characters. It contains four types of errors: spelling errors, redundant words, missing words, and word ordering errors. We also test some state-of-the-art models on the dataset. The experimental results show that even the model with the best performance is 20 points lower than humans, which indicates that there is still much room for improvement. We hope that the new dataset can fill the gap in cross-sentence text correction for native Chinese speakers. | ['Guoping Hu', 'Zhigang Chen', 'Wanxiang Che', 'Dayong Wu', 'Xingyi Duan', 'Baoxin Wang'] | null | null | null | null | coling-2022-10 | ['grammatical-error-correction'] | ['natural-language-processing'] | [ 1.96047336e-01 -3.10769588e-01 2.01446995e-01 -4.65151280e-01
-9.40754354e-01 -4.20291185e-01 2.37132698e-01 5.19102156e-01
-7.78862596e-01 8.42664480e-01 4.68858570e-01 -6.68117106e-01
5.92559457e-01 -3.81225526e-01 -6.41456068e-01 -1.04127973e-01
4.83129203e-01 3.47498834e-01 4.42503780e-01 -5.50879955e-01
8.36202979e-01 -3.02950591e-01 -1.20373917e+00 5.95929682e-01
2.03790045e+00 -6.79137781e-02 7.76225328e-01 8.28201115e-01
-4.76491213e-01 7.77223706e-01 -1.04590201e+00 -7.69461095e-01
-4.11120445e-01 -7.60864615e-01 -8.50450099e-01 -4.68086481e-01
5.18911600e-01 1.11540034e-01 2.89283156e-01 1.46111691e+00
4.86768037e-01 -1.89365700e-01 4.78300452e-01 -5.51716983e-01
-1.35773265e+00 9.09415424e-01 -9.42116827e-02 2.69477695e-01
5.61786354e-01 -3.46059829e-01 6.29789770e-01 -1.29022825e+00
6.45678580e-01 1.13526952e+00 9.49939728e-01 1.19912958e+00
-4.04307783e-01 -7.58403599e-01 2.44532511e-01 1.10663779e-01
-1.26965904e+00 -2.99944550e-01 1.89392525e-03 -2.20732033e-01
1.17522204e+00 5.82728744e-01 3.19308907e-01 8.79974365e-01
3.55225474e-01 8.42721641e-01 9.54247057e-01 -1.19985950e+00
-6.80655837e-02 2.68603563e-01 4.27621365e-01 3.55257273e-01
5.79276145e-01 -1.24752261e-01 -4.34480786e-01 3.15287709e-01
-1.14035398e-01 -4.93329987e-02 -4.10141349e-01 8.06902468e-01
-1.02974534e+00 6.71622932e-01 -1.79317564e-01 6.33820951e-01
3.40455025e-01 -9.65741575e-02 4.70921934e-01 4.94136870e-01
6.15636051e-01 5.69839716e-01 -8.63207817e-01 -4.89969611e-01
-7.78576553e-01 2.20918640e-01 6.11597955e-01 1.62177598e+00
3.01862627e-01 1.01482965e-01 -2.20977232e-01 8.98090422e-01
3.12913120e-01 9.70031023e-01 9.90657151e-01 -1.83632672e-01
9.91644382e-01 5.43157697e-01 1.40528560e-01 -8.69499803e-01
-1.55672953e-01 -2.82076299e-01 -7.18649030e-01 -3.13236237e-01
3.14190269e-01 -2.70992011e-01 -7.80813634e-01 1.26624835e+00
-3.95336181e-01 -3.44139427e-01 -5.79600362e-03 6.45889103e-01
7.65540838e-01 5.65600276e-01 1.78723261e-01 -2.44503036e-01
1.05900717e+00 -1.24544036e+00 -1.14956117e+00 -4.79704738e-01
1.37281072e+00 -1.37089825e+00 1.57543790e+00 3.04764777e-01
-9.84975696e-01 -5.75658143e-01 -8.36394548e-01 -1.96240291e-01
-5.08157611e-01 4.74568069e-01 1.46724626e-01 1.03678894e+00
-7.46138692e-01 5.77930868e-01 -6.06218278e-01 -4.50905085e-01
-2.39352822e-01 -2.33060762e-01 -1.34952545e-01 -2.15740010e-01
-1.30860877e+00 1.11867857e+00 2.84589946e-01 -1.41235009e-01
-2.73150027e-01 -7.24476159e-01 -7.77057111e-01 -5.24053127e-02
-5.96018322e-02 1.55407295e-01 1.55593121e+00 -1.02160048e+00
-1.07544172e+00 9.00299788e-01 -6.80066824e-01 -7.37732798e-02
5.57501316e-01 -5.63565195e-01 -1.05298293e+00 -4.78039652e-01
4.90869612e-01 3.68098691e-02 3.37707192e-01 -9.41223800e-01
-1.11457860e+00 -2.86479950e-01 -8.23816597e-01 9.36814174e-02
-2.76161551e-01 9.00913835e-01 -5.45889258e-01 -1.16670239e+00
3.52743268e-02 -9.05897737e-01 9.23617855e-02 -6.01959050e-01
-2.92454600e-01 -3.67911875e-01 5.17809808e-01 -1.18702877e+00
2.12548137e+00 -1.92589498e+00 -2.87970960e-01 -1.36242419e-01
-3.19070667e-01 8.57312143e-01 -2.84668118e-01 4.81697351e-01
-3.98372263e-02 8.24979484e-01 -3.80074561e-01 -8.26330781e-01
-1.91487700e-01 -1.49703389e-02 -1.71473786e-01 1.26533076e-01
4.22687940e-02 8.85132074e-01 -1.10769081e+00 -3.85399789e-01
-2.89524585e-01 8.51796195e-03 -5.78398824e-01 4.66557546e-03
-1.03758462e-03 3.53291094e-01 -1.50710449e-01 6.31989002e-01
9.95397389e-01 3.53369772e-01 5.01251034e-02 8.29477549e-01
-5.47125578e-01 9.31307733e-01 -1.02353537e+00 1.39881158e+00
-2.87318051e-01 7.79838562e-01 -4.79637980e-01 -1.61175385e-01
9.89004135e-01 2.39853114e-01 -6.04372442e-01 -9.52580333e-01
-1.02600222e-02 8.13557446e-01 6.97987676e-02 -5.50790370e-01
1.01001418e+00 1.66444018e-01 -3.27667177e-01 5.90807199e-01
-2.19935521e-01 -1.97791621e-01 5.44129431e-01 2.96905756e-01
8.30266178e-01 -3.32874328e-01 2.43894160e-01 -4.63128835e-01
6.22455001e-01 1.96213618e-01 8.17577481e-01 9.67860281e-01
-3.66732389e-01 7.91260898e-01 4.89389524e-02 -1.44488782e-01
-9.68564153e-01 -7.20325172e-01 5.63781001e-02 1.19255459e+00
-1.20749921e-01 -9.83136773e-01 -1.16338515e+00 -7.04475105e-01
-2.73762971e-01 1.10577786e+00 -3.50067258e-01 -9.83191952e-02
-9.27803576e-01 -4.47185129e-01 8.35553050e-01 5.82284391e-01
4.96060342e-01 -1.21546125e+00 -2.10548341e-02 1.81086421e-01
-5.50847471e-01 -8.78718913e-01 -1.27758241e+00 -1.02678977e-01
-7.28673875e-01 -1.03555906e+00 -5.40953279e-01 -1.30117607e+00
9.36075389e-01 3.85318100e-01 1.05881882e+00 1.08312225e+00
1.97245836e-01 -1.25068843e-01 -1.18032992e+00 -9.88400340e-01
-1.00162947e+00 1.61981970e-01 2.56724983e-01 -7.06727445e-01
9.22094941e-01 4.35665131e-01 1.55793220e-01 9.91549492e-02
-4.44035769e-01 3.95197049e-02 1.92303374e-01 6.96269512e-01
1.87159672e-01 -1.87969133e-01 3.45028609e-01 -1.31266999e+00
8.55388820e-01 -5.24601862e-02 -4.89974916e-01 7.67252088e-01
-8.14936578e-01 -9.86971706e-02 8.52284670e-01 -2.30170101e-01
-1.41030443e+00 -2.21850857e-01 -3.33520591e-01 6.05833769e-01
-8.91651511e-02 6.69495106e-01 -8.32494795e-02 6.89765662e-02
7.51592338e-01 3.93736303e-01 -5.49423873e-01 -9.20861840e-01
-9.52914730e-02 1.20298731e+00 6.37396872e-01 -3.81620169e-01
5.90961754e-01 -4.54064697e-01 -9.21798527e-01 -7.53170550e-01
-8.59974861e-01 -4.26674545e-01 -9.27043438e-01 -9.07135457e-02
7.29350448e-01 -9.29350078e-01 -2.88137764e-01 9.83800590e-01
-1.56586933e+00 -2.28163421e-01 2.74006724e-01 5.59593558e-01
2.63222069e-01 7.77376711e-01 -9.05179143e-01 -7.22872913e-01
-5.58078289e-01 -9.05622959e-01 8.04293275e-01 3.23055267e-01
-3.55540186e-01 -9.74634230e-01 2.03619413e-02 2.40339369e-01
5.09288967e-01 -5.81498802e-01 8.58328760e-01 -5.36726058e-01
-1.51947588e-01 -1.43995345e-01 5.72280586e-02 5.44544458e-01
9.18327495e-02 2.83639461e-01 -5.18883467e-01 -2.52019286e-01
-1.71300694e-01 5.56491055e-02 8.00453842e-01 1.25990557e-02
1.09744334e+00 -2.67249495e-01 -1.11123614e-01 4.61032867e-01
1.15070069e+00 3.35606009e-01 8.61957729e-01 2.80428976e-01
8.32142770e-01 3.30968827e-01 8.66975129e-01 2.62077570e-01
6.42361164e-01 2.51060694e-01 -1.93926916e-02 3.24776828e-01
-1.91052109e-01 -6.97698474e-01 6.07462406e-01 1.79182196e+00
-2.54792850e-02 -5.07574797e-01 -1.25524712e+00 6.10036969e-01
-1.54931962e+00 -7.91104615e-01 -1.21102750e+00 2.33014369e+00
1.33112931e+00 4.87561710e-02 -5.91597080e-01 1.23014368e-01
9.85573351e-01 -4.52555716e-01 2.06370801e-01 -8.69866967e-01
-3.43096465e-01 1.93158939e-01 5.70484281e-01 8.39343727e-01
-1.03184664e+00 1.53880012e+00 6.38527250e+00 7.83844233e-01
-9.28492904e-01 1.36413604e-01 2.35798270e-01 5.71386874e-01
-6.14419162e-01 1.06501929e-01 -1.42833674e+00 9.28218067e-01
1.02266800e+00 -1.80391774e-01 2.88647234e-01 6.26765847e-01
3.41502696e-01 -2.63921082e-01 -6.22743785e-01 8.06789875e-01
6.05103254e-01 -1.06330276e+00 2.84721721e-02 -5.79163492e-01
1.19940042e+00 4.96445820e-02 -3.40112388e-01 6.22386515e-01
3.34166884e-01 -9.41336632e-01 1.05915868e+00 5.26077092e-01
8.98064315e-01 -5.66169739e-01 8.87058496e-01 7.31653512e-01
-7.72060513e-01 9.77843776e-02 -7.64578462e-01 -4.71274197e-01
-5.97083643e-02 4.86936927e-01 -4.22844559e-01 1.53079823e-01
8.01555693e-01 8.43995750e-01 -1.34557939e+00 1.16910243e+00
-8.19421709e-01 9.49801683e-01 2.59124100e-01 -6.73104465e-01
-1.83030590e-01 3.94508727e-02 2.03151867e-01 1.74924505e+00
7.59730160e-01 8.06561559e-02 -2.42290393e-01 4.22639310e-01
-1.27066135e-01 4.61609662e-01 -1.06733389e-01 2.48720925e-02
8.34397018e-01 5.82337022e-01 -3.54746789e-01 -5.05475342e-01
-4.98936713e-01 1.39055562e+00 5.06782174e-01 1.06905594e-01
-6.48968160e-01 -7.94549882e-01 4.56296235e-01 -2.36889452e-01
7.92016387e-02 -2.26912767e-01 -8.42906713e-01 -1.55226040e+00
2.24562407e-01 -1.07015383e+00 1.18597858e-01 -7.09932387e-01
-1.24940121e+00 4.43083525e-01 -6.23467624e-01 -1.29498482e+00
-2.90581081e-02 -6.83729768e-01 -8.43681037e-01 1.13795054e+00
-1.62998533e+00 -6.94266617e-01 -2.93364108e-01 3.41311693e-01
9.18880999e-01 -2.59862751e-01 9.35261607e-01 6.61462545e-01
-6.65337145e-01 1.22745180e+00 5.41962981e-01 4.63816285e-01
1.19463265e+00 -1.47159588e+00 8.41531515e-01 1.49200654e+00
-1.68270677e-01 8.67802858e-01 6.41969383e-01 -1.25046408e+00
-6.87587619e-01 -1.35942197e+00 2.39691973e+00 -9.01887834e-01
2.01110750e-01 -4.27245080e-01 -1.18202460e+00 6.93240345e-01
2.57697642e-01 -3.44873041e-01 6.18717849e-01 2.42503081e-02
-2.24563312e-02 2.41814181e-01 -7.46845603e-01 6.67095244e-01
9.98651564e-01 -4.28333163e-01 -8.74854445e-01 3.41717482e-01
1.03032255e+00 -6.26078844e-01 -1.70850754e-01 -7.43347406e-02
3.83063778e-02 -6.26901448e-01 1.12620369e-01 -7.70212173e-01
6.70822203e-01 -4.19022977e-01 1.20272644e-01 -1.80344379e+00
-4.21457082e-01 -4.57470268e-01 5.51796794e-01 1.56880403e+00
7.50698864e-01 -2.68937826e-01 1.80851281e-01 7.09674954e-01
-9.04554486e-01 8.71598721e-02 -5.19348502e-01 -7.69972920e-01
6.92241609e-01 -6.67269766e-01 7.98880994e-01 1.19600785e+00
3.23112041e-01 5.18800020e-02 -3.82895589e-01 -4.53244373e-02
9.18357894e-02 -5.07941961e-01 5.23384631e-01 -1.05939901e+00
2.22431719e-01 -2.45843336e-01 2.65900940e-01 -9.17188942e-01
1.25300452e-01 -9.69045520e-01 4.47192103e-01 -1.21315491e+00
9.68931168e-02 -5.92431903e-01 2.49126792e-01 4.14955318e-01
-9.52371597e-01 -2.88694311e-04 4.33927536e-01 1.28184304e-01
-5.38530707e-01 4.66828644e-01 1.23190868e+00 3.96072753e-02
-7.94947073e-02 1.22061849e-01 -8.29871416e-01 7.54800558e-01
9.35676455e-01 -8.72313261e-01 2.90235847e-01 -1.27124858e+00
5.10107279e-01 -3.79071265e-01 -2.71855086e-01 -9.55228984e-01
5.82115471e-01 -2.56339788e-01 2.75778443e-01 -6.13858342e-01
-7.76991963e-01 -3.63592625e-01 -4.30348605e-01 6.81377828e-01
-4.67538685e-01 6.25274479e-01 6.81834817e-02 4.64258939e-02
-2.26215035e-01 -9.49120760e-01 7.71921039e-01 -2.99184471e-01
-7.65036106e-01 -2.36562490e-01 -8.72608900e-01 4.17499721e-01
6.15100682e-01 -1.30368352e-01 -5.63241720e-01 -1.30854756e-01
-6.28949404e-02 2.19784409e-01 4.12919462e-01 7.50392675e-01
5.59381545e-01 -1.20681989e+00 -1.16274023e+00 4.62771505e-01
4.99167383e-01 -3.94291162e-01 9.05421525e-02 4.38072771e-01
-1.01341069e+00 6.95216477e-01 6.09587543e-02 -4.61416580e-02
-1.54741371e+00 2.43979692e-01 2.18436569e-01 -8.49516094e-02
-2.43780434e-01 1.06599391e+00 -6.11145496e-01 -1.00601006e+00
3.30934048e-01 -6.02849066e-01 -3.09838235e-01 -2.87565649e-01
1.09926105e+00 5.02304494e-01 5.01265824e-01 -5.94815314e-01
-4.03901309e-01 4.92931038e-01 -3.79900962e-01 2.39382297e-01
8.04081380e-01 -3.83463174e-01 -3.89588952e-01 2.57546008e-01
8.54449213e-01 7.80284882e-01 -4.30427313e-01 -2.64150381e-01
4.76779103e-01 -7.99800456e-01 -4.14898157e-01 -1.20001042e+00
-5.76741576e-01 1.11248040e+00 3.98587912e-01 -3.25101376e-01
8.51281166e-01 -4.29877877e-01 9.21617508e-01 4.63678598e-01
3.77702773e-01 -1.85016322e+00 -2.44934738e-01 1.61017287e+00
8.07296813e-01 -1.58389091e+00 -4.02443111e-01 -4.32540447e-01
-8.44639182e-01 1.23776901e+00 9.95243847e-01 8.23465437e-02
5.34077287e-01 1.07991658e-01 4.48269814e-01 3.74870181e-01
-6.66021585e-01 -1.61791276e-02 2.71062762e-01 6.90099478e-01
1.25873649e+00 4.57819194e-01 -1.17216778e+00 9.83259797e-01
-5.90434730e-01 -2.48266950e-01 1.12817252e+00 8.33933592e-01
-6.86673284e-01 -1.45744622e+00 -4.61195052e-01 3.71502250e-01
-5.04869580e-01 -7.62472749e-01 -4.66220796e-01 5.79803526e-01
2.94245362e-01 1.35850346e+00 -6.37893379e-03 -3.14231664e-01
3.76429528e-01 3.94028910e-02 8.02068859e-02 -1.05073357e+00
-1.02628136e+00 -3.12062711e-01 -2.92865019e-02 -1.37465373e-01
4.47767749e-02 -9.30088758e-01 -1.55611300e+00 -6.56834602e-01
-5.31510711e-01 2.73842186e-01 6.19717896e-01 1.01352668e+00
1.11920632e-01 2.68490553e-01 2.74308413e-01 -1.07523904e-03
-5.35059094e-01 -1.38803613e+00 -5.47095478e-01 4.16583240e-01
1.94080591e-01 8.27909410e-02 -3.51476371e-01 1.21910296e-01] | [11.022405624389648, 10.742475509643555] |
2340fa81-cfcd-4069-bd51-5042c70dc9cf | partially-does-it-towards-scene-level-fg-sbir | 2203.14804 | null | https://arxiv.org/abs/2203.14804v1 | https://arxiv.org/pdf/2203.14804v1.pdf | Partially Does It: Towards Scene-Level FG-SBIR with Partial Input | We scrutinise an important observation plaguing scene-level sketch research -- that a significant portion of scene sketches are "partial". A quick pilot study reveals: (i) a scene sketch does not necessarily contain all objects in the corresponding photo, due to the subjective holistic interpretation of scenes, (ii) there exists significant empty (white) regions as a result of object-level abstraction, and as a result, (iii) existing scene-level fine-grained sketch-based image retrieval methods collapse as scene sketches become more partial. To solve this "partial" problem, we advocate for a simple set-based approach using optimal transport (OT) to model cross-modal region associativity in a partially-aware fashion. Importantly, we improve upon OT to further account for holistic partialness by comparing intra-modal adjacency matrices. Our proposed method is not only robust to partial scene-sketches but also yields state-of-the-art performance on existing datasets. | ['Yi-Zhe Song', 'Tao Xiang', 'Aneeshan Sain', 'Viswanatha Reddy Gajjala', 'Ayan Kumar Bhunia', 'Pinaki Nath Chowdhury'] | 2022-03-28 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Chowdhury_Partially_Does_It_Towards_Scene-Level_FG-SBIR_With_Partial_Input_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Chowdhury_Partially_Does_It_Towards_Scene-Level_FG-SBIR_With_Partial_Input_CVPR_2022_paper.pdf | cvpr-2022-1 | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 4.19639319e-01 -3.40170711e-01 -8.58693868e-02 -1.43593967e-01
-5.51655591e-01 -8.43330145e-01 9.38162327e-01 1.40293851e-01
1.40473843e-01 2.83339828e-01 4.94063735e-01 -2.76430044e-02
-2.95080006e-01 -7.68948555e-01 -7.55666614e-01 -4.71704930e-01
1.19204886e-01 3.50041509e-01 4.19226438e-01 -2.48582542e-01
4.30362582e-01 9.67592478e-01 -1.67309690e+00 6.86314046e-01
4.35324967e-01 8.05172801e-01 7.19675198e-02 4.68741149e-01
-5.05396545e-01 4.91803139e-01 -3.44506145e-01 -5.51265419e-01
4.03565794e-01 -3.05154532e-01 -8.12201500e-01 5.19399524e-01
1.16449153e+00 -4.33626771e-01 -4.44379956e-01 1.05652940e+00
3.15999649e-02 1.64654061e-01 7.62279093e-01 -1.30187988e+00
-7.74731994e-01 3.51912230e-01 -6.91438735e-01 -2.84362197e-01
5.24584949e-01 2.33767495e-01 1.38676918e+00 -1.20903361e+00
1.02714288e+00 1.42972362e+00 5.40016592e-01 1.73029259e-01
-1.56514955e+00 -4.23815131e-01 4.45161730e-01 -2.18103379e-01
-1.82610023e+00 -4.44929063e-01 1.21613371e+00 -3.08060408e-01
7.35884786e-01 6.79648340e-01 7.75278568e-01 8.39240074e-01
-3.47483605e-02 7.74009347e-01 1.09898520e+00 -2.68396109e-01
1.70568794e-01 2.17290059e-01 -1.63321599e-01 6.99389935e-01
2.91515559e-01 -2.14155465e-01 -6.43014312e-01 -4.02761936e-01
9.53082085e-01 2.19486251e-01 -3.15277837e-02 -8.94924819e-01
-1.36066020e+00 3.93386155e-01 3.94275844e-01 4.16999251e-01
-4.38524544e-01 2.44891629e-01 2.92516112e-01 1.78078488e-01
2.94771254e-01 3.27682495e-01 1.21999886e-02 1.18161067e-01
-1.28605831e+00 5.85594952e-01 6.82448268e-01 1.06737685e+00
1.13219512e+00 -8.54814798e-02 -1.53544962e-01 8.76160324e-01
3.92146185e-02 2.75719434e-01 -2.34075323e-01 -1.00114489e+00
3.53249729e-01 8.66711438e-01 6.68631867e-02 -1.42895007e+00
6.09876104e-02 -2.80702800e-01 -9.94893789e-01 1.00742050e-01
2.77239591e-01 7.36470997e-01 -8.46465170e-01 1.60064793e+00
2.13882163e-01 6.16309717e-02 -5.16232610e-01 1.07281411e+00
7.03484654e-01 4.13983017e-01 2.28536457e-01 4.03152704e-02
1.53496706e+00 -5.75550258e-01 -6.54563785e-01 -8.79321694e-02
1.30631924e-01 -9.25616682e-01 1.37969148e+00 4.28331435e-01
-1.09443665e+00 -4.78918582e-01 -9.64377940e-01 -2.26261392e-01
-7.11540997e-01 5.43264784e-02 7.17406034e-01 6.57289743e-01
-1.20976520e+00 3.58883768e-01 -2.30342612e-01 -7.49536753e-01
5.59674025e-01 1.14817573e-02 -5.96027136e-01 -5.28278708e-01
-5.59723377e-01 5.17772377e-01 1.65853694e-01 -8.74252096e-02
-6.51895702e-01 -9.41212893e-01 -6.20831966e-01 1.69400379e-01
5.26219964e-01 -7.56593645e-01 4.62949991e-01 -7.74307311e-01
-1.01762629e+00 9.16593790e-01 -4.60368156e-01 2.21486643e-01
5.39072216e-01 -1.06425984e-02 -2.61247635e-01 4.46255893e-01
5.04048467e-02 8.28936577e-01 9.76737022e-01 -2.16652608e+00
-2.16029346e-01 -3.13153476e-01 2.44762331e-01 2.54167557e-01
-3.05891246e-01 -2.22175673e-01 -8.16389382e-01 -9.16747570e-01
4.04742479e-01 -8.96811187e-01 -6.30365685e-02 5.06591320e-01
-5.13726592e-01 -8.13793093e-02 1.02694464e+00 -3.47666353e-01
1.42521119e+00 -2.24564958e+00 3.96680422e-02 4.90754247e-01
3.04123849e-01 1.26638161e-02 -4.49605167e-01 9.82369483e-01
-1.06981061e-01 2.88526744e-01 -4.28232551e-01 -5.19125104e-01
1.25113174e-01 3.25321287e-01 -6.41948044e-01 3.64489973e-01
3.59648407e-01 9.46909964e-01 -9.19627547e-01 -7.68065810e-01
5.85051060e-01 3.95128846e-01 -6.18796825e-01 -2.48908415e-01
-2.28158265e-01 2.04523876e-01 -4.67124432e-01 1.08436620e+00
1.11979055e+00 -2.74221569e-01 2.24785075e-01 -7.49478340e-01
-2.18678445e-01 -1.16043642e-01 -1.36774397e+00 2.16986322e+00
-4.34531458e-02 4.92225140e-01 5.15420847e-02 -6.85296416e-01
5.98033011e-01 1.94037613e-02 7.20584869e-01 -5.40719867e-01
-3.30550730e-01 -1.16118237e-01 -3.77524376e-01 -8.74423534e-02
1.01943040e+00 -1.71264991e-01 -1.65788874e-01 4.98283565e-01
-1.17209420e-01 -3.45855951e-01 1.07521333e-01 7.94133306e-01
1.19468319e+00 1.93494603e-01 8.45881253e-02 -5.67517817e-01
4.63258803e-01 7.22024813e-02 1.91444475e-02 8.24615777e-01
-8.61209631e-02 9.36201572e-01 5.00608861e-01 -3.55576694e-01
-1.15311611e+00 -1.33728456e+00 -3.62358242e-02 8.04083228e-01
4.69711542e-01 -6.51435077e-01 -5.45597911e-01 -3.05150390e-01
2.16972992e-01 5.42313993e-01 -6.99536979e-01 1.55089498e-01
-2.92441279e-01 -2.11589813e-01 4.48127508e-01 2.51816213e-01
3.25459033e-01 -8.55437040e-01 -2.77563840e-01 2.39702710e-03
-3.18569958e-01 -1.26608670e+00 -5.96800268e-01 -4.95364100e-01
-7.38862574e-01 -9.58047152e-01 -8.07641566e-01 -3.24201316e-01
9.06158388e-01 8.23622525e-01 1.22725999e+00 4.75051880e-01
-5.14516890e-01 8.81344259e-01 -2.45705709e-01 2.10402012e-01
-8.96134675e-02 -2.69033551e-01 -3.57755907e-02 3.41413647e-01
5.05993553e-02 -6.46159470e-01 -7.75749266e-01 2.73901254e-01
-1.17046607e+00 8.31055716e-02 7.35993564e-01 5.11505485e-01
6.52045190e-01 2.78448641e-01 1.00531653e-01 -7.98216581e-01
6.80995286e-01 -1.59209982e-01 -2.78349817e-01 5.85778773e-01
-3.18622530e-01 -3.85989575e-03 4.21252966e-01 -5.37386358e-01
-1.13171697e+00 1.53152691e-02 3.67242545e-01 -8.37539434e-01
-2.52389669e-01 1.63772017e-01 -2.00964972e-01 -2.64893800e-01
3.72324616e-01 2.99951196e-01 -2.01296747e-01 -3.84192973e-01
6.91629410e-01 2.44049415e-01 3.63579124e-01 -9.59918261e-01
1.04787481e+00 9.93699968e-01 2.90631056e-01 -1.32949674e+00
-3.26733768e-01 -5.92407584e-01 -8.26398432e-01 -4.73375499e-01
6.71180844e-01 -8.68629813e-01 -7.82313108e-01 1.20183155e-01
-1.12886643e+00 -1.74787566e-01 -2.40084484e-01 -5.17123006e-02
-3.79308909e-01 7.86764145e-01 -4.19849694e-01 -1.01861989e+00
9.06207711e-02 -9.86341238e-01 1.56752706e+00 -2.29112267e-01
-3.37798834e-01 -7.82350004e-01 -1.73047006e-01 1.72950998e-01
2.27235630e-01 2.40104973e-01 1.11997914e+00 1.31499067e-01
-1.17558920e+00 -1.41330764e-01 -7.67040670e-01 -6.09866269e-02
8.90256837e-02 1.86853379e-01 -9.44239199e-01 -1.53153732e-01
-5.97883284e-01 3.64424884e-02 6.04868293e-01 3.10937166e-02
1.08400178e+00 -3.56924713e-01 -2.89064437e-01 2.62397707e-01
1.85826135e+00 -3.79713207e-01 7.74444580e-01 -2.45612547e-01
8.95784140e-01 8.33074450e-01 5.88977933e-01 6.03472054e-01
3.96565259e-01 7.62856543e-01 3.35375816e-01 -1.67135477e-01
-6.09016180e-01 -6.80094302e-01 5.97851127e-02 6.24544084e-01
3.99486441e-03 -2.21939042e-01 -6.42532110e-01 7.83375919e-01
-1.83162129e+00 -1.15198851e+00 -3.98613542e-01 2.28421879e+00
4.10696387e-01 -2.50779331e-01 1.79870203e-01 -1.60074513e-02
4.23023850e-01 4.57585126e-01 -2.90195733e-01 -2.86980182e-01
-4.28847939e-01 1.34837419e-01 4.85539973e-01 3.86551350e-01
-7.61054337e-01 1.18798757e+00 6.56121349e+00 1.13165104e+00
-7.86108017e-01 -6.09193146e-02 3.11292380e-01 -3.13132145e-02
-8.93477023e-01 3.68775666e-01 -1.90910846e-01 2.56039858e-01
-1.17414370e-01 1.85124665e-01 7.19296038e-01 4.90105391e-01
-1.06486589e-01 -3.30116779e-01 -1.06346691e+00 1.02319801e+00
2.18735710e-01 -1.40001237e+00 7.72869885e-01 3.40470999e-01
6.76648796e-01 -3.91843319e-01 1.06332645e-01 -4.13759947e-02
2.28034649e-02 -8.25053632e-01 9.52573180e-01 6.96386099e-01
9.75472093e-01 -5.20804822e-01 6.16833754e-02 1.25375167e-01
-1.71269870e+00 2.15853378e-01 -4.25785422e-01 1.40461966e-01
1.16077535e-01 4.68193352e-01 -3.46674114e-01 7.20809579e-01
6.36746883e-01 4.71650094e-01 -8.03788543e-01 7.78923988e-01
2.45714173e-01 5.66163100e-02 -4.56250727e-01 1.22305185e-01
2.99915761e-01 -3.32682371e-01 5.97716451e-01 1.24693418e+00
1.89605668e-01 2.76056826e-01 3.87305319e-02 1.25208747e+00
8.88585523e-02 -3.50714847e-02 -9.03546870e-01 -1.39028862e-01
4.76305693e-01 1.26737571e+00 -1.05558455e+00 -3.94579053e-01
-4.39471662e-01 1.31038892e+00 1.98548406e-01 7.43019819e-01
-5.00376105e-01 9.66180339e-02 7.76653945e-01 2.96221137e-01
3.19292128e-01 -3.97542924e-01 -6.23693347e-01 -1.18029523e+00
9.24839675e-02 -5.57213485e-01 5.67856207e-02 -1.04984903e+00
-1.40060580e+00 1.73166528e-01 4.20712493e-02 -1.36068928e+00
2.97110230e-01 -4.38342094e-01 -4.11833078e-01 6.43758953e-01
-1.30609739e+00 -1.65665412e+00 -3.04172605e-01 7.73475289e-01
5.89759946e-01 3.27834249e-01 7.72909760e-01 5.09410620e-01
-1.38534561e-01 4.70496833e-01 -3.64907950e-01 -9.01751369e-02
6.16207004e-01 -1.00700665e+00 2.35689193e-01 9.20970857e-01
4.38427836e-01 1.01157176e+00 7.38252580e-01 -7.40619123e-01
-1.91202986e+00 -8.09400856e-01 8.45194221e-01 -6.84081256e-01
5.97527623e-01 -6.09213650e-01 -8.29339921e-01 3.79925281e-01
6.95060864e-02 5.16116247e-02 3.90138924e-01 2.44320676e-01
-7.40868568e-01 -2.64695678e-02 -1.16525400e+00 1.02816701e+00
1.40828693e+00 -1.02638650e+00 -3.69598985e-01 1.28260225e-01
3.00137162e-01 5.65742282e-03 -8.60584676e-01 1.84918299e-01
1.06412745e+00 -1.17521036e+00 1.35866237e+00 -2.04897717e-01
3.96926552e-01 -4.21575665e-01 -3.82461637e-01 -7.88484812e-01
-4.44646060e-01 -4.51107651e-01 -2.79345568e-02 1.25847244e+00
-6.87382072e-02 -2.32052818e-01 7.86059797e-01 7.67994165e-01
9.02775601e-02 -4.94903684e-01 -9.08223748e-01 -7.05283105e-01
-2.63491452e-01 -5.81531882e-01 6.78240418e-01 1.12847972e+00
-3.30929160e-02 8.01132172e-02 -4.56023782e-01 1.89179003e-01
8.55802774e-01 2.56282002e-01 1.03596461e+00 -1.13243639e+00
1.84949905e-01 -8.47851872e-01 -2.99231112e-01 -1.15680492e+00
-8.84096324e-02 -5.45883536e-01 -9.09407660e-02 -1.49055839e+00
3.24228913e-01 -7.05888569e-01 -2.68536121e-01 4.37984645e-01
5.43315932e-02 6.69402838e-01 4.22802597e-01 3.13003063e-01
-8.03671837e-01 3.39896828e-01 1.40551591e+00 -1.95997819e-01
1.14536799e-01 -4.53760594e-01 -6.74256742e-01 3.83315891e-01
2.31267348e-01 -1.74373627e-01 -4.16289538e-01 -2.85094768e-01
3.51174116e-01 -1.59957993e-03 7.37995386e-01 -7.26039767e-01
1.98092625e-01 -3.80750746e-01 2.19643816e-01 -9.44262624e-01
8.33645940e-01 -1.26917183e+00 5.11470556e-01 1.22092307e-01
-9.09396335e-02 -1.68447971e-01 2.17588678e-01 7.08829820e-01
-1.13333084e-01 1.81287646e-01 3.90286803e-01 -2.70684302e-01
-8.60715628e-01 3.36116493e-01 -1.27547652e-01 -4.52838302e-01
6.85257256e-01 -6.53819144e-01 -2.89047062e-01 -4.21905339e-01
-2.16710642e-01 -1.45826727e-01 9.82608914e-01 6.04799747e-01
7.33398736e-01 -1.51781321e+00 -4.13917363e-01 2.88338572e-01
5.24674118e-01 -2.35041618e-01 7.65788794e-01 7.79817820e-01
-5.05354524e-01 4.66232538e-01 -4.08972353e-02 -5.99497855e-01
-1.23877454e+00 6.00671887e-01 -1.99098617e-01 6.84864894e-02
-7.50976086e-01 6.62569404e-01 8.13138664e-01 -3.37140292e-01
2.67696492e-02 -3.74069601e-01 4.45947945e-01 1.25924386e-02
1.63014218e-01 2.11428136e-01 -1.26163870e-01 -9.35483515e-01
-5.84820688e-01 8.80011797e-01 1.51229322e-01 -3.29141498e-01
9.42461431e-01 -2.60935694e-01 -2.08260298e-01 4.36992347e-01
1.11747885e+00 2.85665721e-01 -1.00842249e+00 -2.24089473e-01
-2.72799790e-01 -9.16827917e-01 -8.83091763e-02 -7.90774345e-01
-6.85975492e-01 7.68580854e-01 2.23054647e-01 2.71665901e-01
1.07211840e+00 1.25704244e-01 6.13415539e-01 1.80894077e-01
5.33087671e-01 -8.95504653e-01 1.24108367e-01 7.89453760e-02
1.03518581e+00 -9.09520686e-01 5.95374942e-01 -6.33836985e-01
-5.62716544e-01 7.89444208e-01 1.89493582e-01 -3.30437571e-01
5.57169497e-01 -1.54361390e-02 -3.58770043e-01 -5.49952745e-01
-4.69082505e-01 -1.67707667e-01 6.15493894e-01 3.17877769e-01
2.09440544e-01 1.77904576e-01 -1.42626375e-01 8.67048651e-02
1.24089353e-01 -1.94877774e-01 8.64471868e-02 8.74534965e-01
-1.07481666e-01 -1.10484445e+00 -4.56845701e-01 4.15888399e-01
-3.28472368e-02 -1.53629303e-01 -7.91718602e-01 9.37798798e-01
-5.56078460e-03 6.95954502e-01 2.32482001e-01 -2.96066552e-01
3.65661711e-01 -8.62491131e-02 7.96554089e-01 -3.34245563e-01
-3.83133501e-01 9.28191021e-02 -1.14064269e-01 -8.09952319e-01
-5.33335268e-01 -5.76150715e-01 -7.71870196e-01 -4.98697996e-01
-2.27943793e-01 -3.92244101e-01 6.42864525e-01 7.63179958e-01
5.43462396e-01 4.01956409e-01 2.45711118e-01 -1.03150189e+00
1.23392463e-01 -5.79455435e-01 -6.80160284e-01 6.92053318e-01
3.34264547e-01 -7.44545877e-01 -2.54330784e-01 5.60615696e-02] | [11.642879486083984, 0.6004000306129456] |
f33657ff-c309-4e24-8515-27a9e42c6331 | neca-network-embedded-deep-representation | 2205.12752 | null | https://arxiv.org/abs/2205.12752v1 | https://arxiv.org/pdf/2205.12752v1.pdf | NECA: Network-Embedded Deep Representation Learning for Categorical Data | We propose NECA, a deep representation learning method for categorical data. Built upon the foundations of network embedding and deep unsupervised representation learning, NECA deeply embeds the intrinsic relationship among attribute values and explicitly expresses data objects with numeric vector representations. Designed specifically for categorical data, NECA can support important downstream data mining tasks, such as clustering. Extensive experimental analysis demonstrated the effectiveness of NECA. | ['Wenjun Zhou', 'Sen Wu', 'Xiaonan Gao'] | 2022-05-25 | null | null | null | null | ['network-embedding'] | ['methodology'] | [-2.40178913e-01 2.93953985e-01 -6.46844029e-01 -9.57543492e-01
-4.16596308e-02 -2.03115627e-01 3.07911426e-01 7.75700688e-01
-5.52444980e-02 3.68851602e-01 7.88668811e-01 -4.94556129e-01
-6.07524872e-01 -1.35480237e+00 -2.95910597e-01 -6.40044332e-01
-6.18249714e-01 3.58050168e-01 -8.41317177e-01 -7.43259415e-02
-1.94247544e-01 5.11190653e-01 -1.12281299e+00 5.86981118e-01
1.51707098e-01 1.13232124e+00 -8.81758869e-01 1.45416319e-01
-4.34292614e-01 1.20789206e+00 -4.72006977e-01 -2.77089924e-01
-1.10100314e-01 -5.82980923e-02 -7.38020182e-01 -4.64884043e-01
-1.85354516e-01 -2.39678532e-01 -1.14609480e+00 8.06374729e-01
1.46354347e-01 1.95638940e-01 1.07712865e+00 -1.58857262e+00
-1.24849868e+00 1.29136181e+00 -4.35958087e-01 1.19528152e-01
-1.22594096e-01 1.12708293e-01 1.43807197e+00 -8.61586392e-01
2.43195534e-01 1.55315185e+00 8.66823018e-01 4.06779587e-01
-1.47140145e+00 -5.45791626e-01 -1.73709728e-02 -1.20594604e-02
-1.46452641e+00 -1.09859325e-01 6.75604820e-01 -3.33463073e-01
1.00600135e+00 1.74287528e-01 5.48054516e-01 1.22177434e+00
-7.75198191e-02 6.88060224e-01 4.19066340e-01 6.49417937e-02
3.02186489e-01 -4.66107130e-01 6.10618770e-01 5.95018923e-01
4.31555361e-01 1.95834786e-01 -7.79833347e-02 -3.43388021e-01
7.84646213e-01 6.67908192e-01 3.05673301e-01 -2.76957929e-01
-1.49852383e+00 1.14867806e+00 1.06968820e+00 1.90837860e-01
-5.03749073e-01 8.70802104e-01 9.92946506e-01 5.66114724e-01
1.18933953e-01 4.98213947e-01 -4.39411074e-01 -1.30656913e-01
-1.74485087e-01 -2.46173590e-01 4.04709011e-01 7.96296179e-01
4.81750667e-01 3.53734821e-01 -5.84048629e-01 9.29403186e-01
2.86098331e-01 3.09910774e-02 4.22805965e-01 -9.45450783e-01
2.41904110e-01 1.24868977e+00 -6.78408265e-01 -1.34782577e+00
-5.18571079e-01 -1.26954526e-01 -1.46323144e+00 -1.98217109e-01
1.26235083e-01 -5.64327128e-02 -7.15451598e-01 1.72850347e+00
-6.06836565e-02 -5.47299087e-02 3.54418457e-01 5.85174739e-01
1.16311610e+00 8.04757297e-01 3.60210419e-01 7.10492507e-02
1.14496684e+00 -5.14384389e-01 -7.56211579e-01 4.40457255e-01
9.69237447e-01 2.88395971e-01 7.66766965e-01 1.31679639e-01
-6.30938232e-01 -3.97844255e-01 -8.17208886e-01 -4.31015432e-01
-8.34494710e-01 -5.75672463e-02 1.50749993e+00 2.83114970e-01
-9.50277925e-01 3.29592109e-01 -6.03982151e-01 -1.75220132e-01
1.03347123e+00 6.79429829e-01 -5.83434880e-01 -7.23872781e-02
-1.26092124e+00 3.41329351e-02 5.12036681e-01 2.77024806e-01
-3.38851452e-01 -7.87660122e-01 -1.53005981e+00 7.81147838e-01
-1.45999342e-01 -6.68917060e-01 6.92082882e-01 -2.43395239e-01
-1.05809307e+00 6.36839926e-01 7.83107132e-02 -6.56297863e-01
-9.84378308e-02 1.58364072e-01 -6.94182992e-01 2.80109998e-02
-1.50565192e-01 6.84463143e-01 2.08178952e-01 -6.57239914e-01
-8.36114436e-02 -4.24026310e-01 -5.21970280e-02 -3.87304157e-01
-8.65917385e-01 -4.27352995e-01 9.67340171e-02 -8.76307011e-01
3.09559256e-02 -3.08203727e-01 -6.53084636e-01 4.02727991e-01
-6.07740760e-01 -8.57757211e-01 8.00716043e-01 -3.07561792e-02
1.11090600e+00 -2.74144864e+00 1.44830257e-01 8.23997855e-01
9.05029893e-01 -2.33622938e-01 -1.84529260e-01 5.59038162e-01
-4.38672453e-01 2.93343306e-01 -2.18640238e-01 3.33480611e-02
2.58627474e-01 5.52678466e-01 -5.97268701e-01 6.06406391e-01
3.56452882e-01 1.47434711e+00 -1.15120196e+00 -3.74371350e-01
3.63144994e-01 6.60410523e-01 -6.58327937e-01 1.48354679e-01
-2.11794704e-01 -2.23097518e-01 -5.21806121e-01 7.92142808e-01
3.79316121e-01 -4.72562492e-01 4.49040592e-01 -4.75338072e-01
3.62927020e-01 2.50922978e-01 -4.53090906e-01 1.68975902e+00
-2.86077917e-01 8.03455710e-01 -3.69936138e-01 -1.60840666e+00
1.25565541e+00 9.46122557e-02 8.38130593e-01 -6.02724552e-01
5.52483320e-01 -2.10254669e-01 1.44162297e-01 -2.13910341e-01
3.07299078e-01 1.00144051e-01 -6.50312543e-01 1.97306558e-01
6.82187453e-02 4.27353203e-01 -2.14095145e-01 5.90256214e-01
1.16346729e+00 -6.75266564e-01 3.75734508e-01 -2.79927552e-01
2.18103319e-01 -1.97142571e-01 5.54554343e-01 3.87727082e-01
-9.25627202e-02 2.39692539e-01 1.17514110e+00 -7.72970498e-01
-6.77655697e-01 -1.15555394e+00 -3.52321625e-01 1.02539384e+00
-1.81887984e-01 -8.34734738e-01 5.24460152e-02 -8.27623725e-01
8.44976068e-01 3.78376633e-01 -1.41392159e+00 -6.60105526e-01
-2.16761634e-01 -5.95145822e-01 7.47487247e-01 1.15713918e+00
1.09384060e-02 -1.00894260e+00 8.65770727e-02 7.64750093e-02
2.39656165e-01 -7.49237537e-01 -6.76873177e-02 6.28325939e-01
-8.44502628e-01 -1.16445732e+00 -4.00725417e-02 -1.04066563e+00
8.06250870e-01 -1.11680582e-01 1.10009742e+00 2.56466776e-01
-5.52209437e-01 2.44086072e-01 -3.69300157e-01 -3.61037329e-02
-2.53243387e-01 1.71505157e-02 2.69056499e-01 1.46056503e-01
7.45438457e-01 -9.03165638e-01 -6.30611539e-01 -4.95730452e-02
-8.36613476e-01 -4.30211842e-01 6.27169549e-01 1.03406012e+00
7.11418748e-01 8.27170983e-02 1.05425406e+00 -1.14657664e+00
6.35093033e-01 -1.10943151e+00 -2.09483892e-01 -2.16758158e-02
-5.14398694e-01 2.77059019e-01 8.74122441e-01 -2.72105157e-01
-2.88026869e-01 -2.29418203e-01 -1.88191250e-01 -5.68352461e-01
-1.81570411e-01 1.03377342e+00 -2.86711454e-01 5.19188046e-01
3.71952921e-01 9.31437910e-02 3.12690437e-01 -4.09235001e-01
6.84887767e-01 7.12065101e-01 8.34183991e-01 -5.52191138e-01
4.20860916e-01 4.05594736e-01 2.41703838e-01 -5.49210012e-01
-5.07356226e-01 -1.44835502e-01 -5.39187431e-01 3.24010998e-01
8.62690151e-01 -9.19495404e-01 -1.31291449e+00 5.71364574e-02
-6.74634635e-01 -2.95433700e-01 -6.43338919e-01 4.22349602e-01
-3.49565178e-01 3.17474641e-02 -9.02912140e-01 -1.90282911e-01
-3.53239149e-01 -6.02167964e-01 6.18559241e-01 -1.76218316e-01
-4.15315211e-01 -1.30416942e+00 -1.46552637e-01 -1.88302770e-01
3.15440953e-01 7.14604378e-01 1.66914058e+00 -8.11312556e-01
1.51418105e-01 -3.64914954e-01 -5.57703316e-01 1.27058938e-01
5.47270000e-01 1.56155124e-01 -6.24903798e-01 -1.46548465e-01
-7.74439991e-01 -6.44496083e-01 1.15963244e+00 3.79704624e-01
2.43207359e+00 -4.96813118e-01 -3.19964796e-01 1.13291526e+00
1.19167709e+00 -8.73129368e-02 6.04578853e-01 -1.70955677e-02
9.76106048e-01 4.80795085e-01 1.33787155e-01 6.75381958e-01
4.95961368e-01 1.92765072e-01 8.96251798e-01 -3.14690173e-01
1.42622069e-02 -3.80642086e-01 -9.16132778e-02 9.38119888e-01
3.16539854e-01 -7.72867631e-03 -7.00519919e-01 7.76886284e-01
-1.83383060e+00 -8.17525268e-01 -2.19560057e-01 1.34278035e+00
7.32383966e-01 1.25373350e-02 1.65659353e-01 4.37399000e-01
4.34386790e-01 2.62072861e-01 -8.15655112e-01 -8.69582534e-01
-2.80231088e-01 1.87198669e-01 2.99115837e-01 -1.83221027e-01
-1.07117677e+00 7.07953155e-01 7.06396580e+00 3.57544750e-01
-8.32159877e-01 -4.15237188e-01 8.65457892e-01 -8.15398693e-02
-6.60766900e-01 -4.88690674e-01 1.37379589e-02 5.27275681e-01
9.05859351e-01 -3.78195465e-01 7.74706751e-02 8.87550056e-01
-2.34157801e-01 7.71492004e-01 -1.60851002e+00 1.11929345e+00
-5.11075974e-01 -1.82728100e+00 5.61942101e-01 3.51220161e-01
4.02613699e-01 -9.81821865e-02 5.18883526e-01 7.23847210e-01
9.38392818e-01 -1.84738028e+00 -4.86410223e-02 2.47423887e-01
1.28777623e+00 -1.13951647e+00 8.07917595e-01 -4.21362400e-01
-1.31611657e+00 -4.59516764e-01 -6.69146895e-01 -3.33403260e-01
-3.63884866e-01 5.72377563e-01 -5.04950345e-01 5.05317867e-01
5.48225820e-01 1.76550865e+00 -6.43197536e-01 4.65859324e-01
-9.90734920e-02 8.38093162e-01 6.14609122e-02 3.07743512e-02
5.46617150e-01 -1.43529087e-01 -1.38218910e-01 1.46573985e+00
-1.03463769e-01 3.63609552e-01 -3.63244973e-02 1.05750501e+00
-6.45033717e-01 -1.23869359e-01 -1.01052070e+00 -5.91998458e-01
7.27328598e-01 1.20725191e+00 -4.92654800e-01 -2.75316626e-01
-3.48579079e-01 6.03698730e-01 6.14629686e-01 3.95603597e-01
-6.92251027e-01 -8.24090600e-01 1.26604331e+00 -2.24093467e-01
4.25039589e-01 -3.69788148e-02 -5.51916301e-01 -9.21943784e-01
-5.02295136e-01 -6.15903080e-01 8.97078216e-01 -3.96642089e-01
-1.87686288e+00 3.83488327e-01 -1.93260759e-01 -1.11450601e+00
-2.87992269e-01 -9.32599425e-01 -8.37500572e-01 4.25938219e-01
-1.08442879e+00 -1.06208313e+00 -1.00037478e-01 9.37339425e-01
5.29175401e-02 -6.14371300e-01 1.30031943e+00 1.76778898e-01
-8.80153775e-01 1.22115219e+00 3.55924606e-01 1.10230720e+00
2.42811397e-01 -1.32947278e+00 5.66120327e-01 6.16328008e-02
1.21255040e-01 1.08843577e+00 3.33728716e-02 -1.27406791e-01
-1.80960882e+00 -1.50473750e+00 5.00113368e-01 -2.33619049e-01
9.79944527e-01 -6.28531098e-01 -1.12689376e+00 7.80799091e-01
-7.46060237e-02 7.15922654e-01 1.45875382e+00 5.83175719e-01
-1.04820120e+00 -3.28103244e-01 -1.19058871e+00 2.53560156e-01
9.63857651e-01 -7.97948241e-01 -6.80469871e-01 3.01207900e-02
1.14217210e+00 3.06557745e-01 -1.43299818e+00 4.23230559e-01
5.43210208e-01 -3.24577272e-01 1.36655736e+00 -1.42058754e+00
9.34461951e-01 1.89666897e-02 -4.81021106e-01 -1.21622586e+00
-6.71684921e-01 -1.44100547e-01 -7.71246552e-01 1.25568640e+00
1.69585168e-01 -6.01770878e-01 6.88803792e-01 4.82640684e-01
4.93095443e-02 -1.01748013e+00 -7.52164721e-01 -6.47987962e-01
2.46729642e-01 -3.73869419e-01 1.03977966e+00 1.58120990e+00
3.72991800e-01 4.24343109e-01 -9.43696722e-02 4.16217409e-02
6.30127072e-01 3.36218089e-01 3.26877892e-01 -1.46227717e+00
9.34659615e-02 -7.27965534e-01 -1.06143594e+00 -5.95181108e-01
6.45062089e-01 -1.52712882e+00 -5.37969589e-01 -1.67110801e+00
2.22323071e-02 -7.15503275e-01 -1.12865102e+00 1.02278531e+00
1.62647337e-01 3.26384485e-01 -2.28961319e-01 -9.12000611e-02
-6.52500391e-01 8.62385869e-01 6.04217052e-01 -6.16527438e-01
6.74530566e-02 -4.55940008e-01 -1.27963519e+00 3.86913747e-01
8.68559182e-01 -3.32721889e-01 -4.73341286e-01 -6.21709526e-01
5.56395173e-01 -3.40286233e-02 3.12088192e-01 -5.09216666e-01
4.35082391e-02 -2.07363814e-01 8.33422482e-01 -7.30189383e-01
1.64186388e-01 -9.07792866e-01 -3.66605103e-01 4.74875212e-01
-9.36201215e-01 1.23255037e-01 1.57581076e-01 8.85310769e-01
-4.58821863e-01 4.71516848e-01 2.76755154e-01 3.95720869e-01
-8.95761847e-01 6.87378168e-01 -3.14081937e-01 -1.17652252e-01
8.30984414e-01 2.56061857e-03 -3.82794678e-01 -2.57790536e-01
-7.09206522e-01 7.01302111e-01 4.23975773e-02 5.61618924e-01
1.11559594e+00 -1.95719123e+00 -6.94514513e-01 5.27956784e-01
4.75950241e-01 1.67910710e-01 -9.62622836e-02 3.63885015e-01
-2.94333875e-01 2.67919481e-01 -3.54678094e-01 -4.31598186e-01
-6.92628384e-01 1.04226267e+00 -1.66716091e-02 2.15699315e-01
-8.50158513e-01 7.68836737e-01 3.61100495e-01 -6.06960177e-01
5.78929305e-01 -6.32146537e-01 -5.63201189e-01 1.94510058e-01
5.81754208e-01 2.93596447e-01 -2.64372915e-01 -2.62160808e-01
-6.34339690e-01 3.04577593e-02 -9.55497772e-02 4.91360664e-01
1.78530920e+00 3.16960245e-01 -6.46490574e-01 6.12930954e-01
1.76312411e+00 -4.24369901e-01 -8.42295825e-01 -2.29297921e-01
-7.74157059e-04 -3.61149013e-01 4.35462482e-02 -6.37005687e-01
-1.52000976e+00 1.05373573e+00 9.70964208e-02 2.20389664e-02
9.70429599e-01 2.13297367e-01 5.47457218e-01 8.56890678e-01
1.61123462e-02 -4.58618551e-01 7.71280304e-02 5.23327172e-01
6.85550511e-01 -1.20510674e+00 -1.38179094e-01 -1.82003543e-01
-4.22629178e-01 1.31736779e+00 5.64601064e-01 -2.03310445e-01
9.41734672e-01 2.60300934e-01 -4.42970023e-02 -5.43477774e-01
-9.57581699e-01 6.67743832e-02 1.61525562e-01 7.33190358e-01
4.71891969e-01 3.64628792e-01 6.73500150e-02 1.12332296e+00
-1.32798746e-01 -2.56187111e-01 3.77920359e-01 5.22543609e-01
1.83538929e-01 -8.14049363e-01 3.08579236e-01 9.45669174e-01
-2.47279316e-01 -1.37545988e-01 -4.82980877e-01 5.86872578e-01
-1.18642949e-01 6.76913619e-01 6.94564700e-01 -7.17185378e-01
1.54067323e-01 -6.72154576e-02 -9.94940624e-02 -6.09461665e-01
-6.80003345e-01 -5.77974975e-01 -2.69321144e-01 -5.69786429e-01
-3.84105630e-02 -4.30346012e-01 -1.72270465e+00 -6.86257303e-01
4.60748136e-01 2.23732263e-01 2.05085665e-01 3.20192635e-01
6.34351850e-01 9.30362880e-01 9.72592473e-01 -9.62137580e-02
-4.26507086e-01 -7.76899576e-01 -7.71222234e-01 6.09607816e-01
5.31949520e-01 -5.52340329e-01 -2.00150251e-01 -3.11800331e-01] | [7.122446060180664, 6.311608791351318] |
1f267fb1-b5c7-4e10-8ad4-33e7c59ddda5 | visual-text-correction | 1801.01967 | null | http://arxiv.org/abs/1801.01967v3 | http://arxiv.org/pdf/1801.01967v3.pdf | Visual Text Correction | Videos, images, and sentences are mediums that can express the same
semantics. One can imagine a picture by reading a sentence or can describe a
scene with some words. However, even small changes in a sentence can cause a
significant semantic inconsistency with the corresponding video/image. For
example, by changing the verb of a sentence, the meaning may drastically
change. There have been many efforts to encode a video/sentence and decode it
as a sentence/video. In this research, we study a new scenario in which both
the sentence and the video are given, but the sentence is inaccurate. A
semantic inconsistency between the sentence and the video or between the words
of a sentence can result in an inaccurate description. This paper introduces a
new problem, called Visual Text Correction (VTC), i.e., finding and replacing
an inaccurate word in the textual description of a video. We propose a deep
network that can simultaneously detect an inaccuracy in a sentence, and fix it
by replacing the inaccurate word(s). Our method leverages the semantic
interdependence of videos and words, as well as the short-term and long-term
relations of the words in a sentence. In our formulation, part of a visual
feature vector for every single word is dynamically selected through a gating
process. Furthermore, to train and evaluate our model, we propose an approach
to automatically construct a large dataset for VTC problem. Our experiments and
performance analysis demonstrates that the proposed method provides very good
results and also highlights the general challenges in solving the VTC problem.
To the best of our knowledge, this work is the first of its kind for the Visual
Text Correction task. | ['Amir Mazaheri', 'Mubarak Shah'] | 2018-01-06 | visual-text-correction-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Amir_Mazaheri_Visual_Text_Correction_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Amir_Mazaheri_Visual_Text_Correction_ECCV_2018_paper.pdf | eccv-2018-9 | ['visual-text-correction'] | ['computer-vision'] | [ 3.96942735e-01 -2.93286771e-01 2.20106430e-02 -5.29600263e-01
-4.77115393e-01 -3.86601865e-01 3.97341162e-01 2.25169510e-01
-2.05899492e-01 5.28866708e-01 3.91299784e-01 1.63271856e-02
4.92862016e-01 -3.39112759e-01 -1.17582142e+00 -4.77360308e-01
4.96205658e-01 -1.90918744e-01 2.66225517e-01 -3.71854715e-02
5.22283733e-01 -4.07416548e-04 -1.37201583e+00 6.85601473e-01
4.29509521e-01 9.03385401e-01 7.12049663e-01 4.27317470e-01
-3.20313603e-01 1.06951308e+00 -7.98533857e-01 -6.33603275e-01
-1.11203477e-01 -6.97291911e-01 -7.99043834e-01 5.56841493e-01
8.04630399e-01 -6.04616106e-01 -4.75245565e-01 1.48302603e+00
2.08327278e-01 3.45382802e-02 2.55449712e-01 -1.27552831e+00
-1.04811537e+00 5.46388090e-01 -6.22107267e-01 2.59324521e-01
7.03507185e-01 2.22745970e-01 7.46846437e-01 -1.12885916e+00
8.12563241e-01 1.24325097e+00 5.84260762e-01 4.58593428e-01
-9.10883248e-01 -5.41234612e-01 5.09015977e-01 5.81961393e-01
-1.47717464e+00 -6.27586484e-01 7.73712456e-01 -5.20490289e-01
1.02212024e+00 1.83695480e-01 9.23307419e-01 1.32637787e+00
5.55060625e-01 6.87575996e-01 6.07339621e-01 -2.66636699e-01
1.26935065e-01 1.22860268e-01 -1.91663578e-01 7.08944917e-01
2.37694457e-01 -3.28042626e-01 -7.87328601e-01 3.47275108e-01
4.44204152e-01 2.96737671e-01 -4.95157331e-01 -1.70835346e-01
-1.18669832e+00 6.10272288e-01 2.55601436e-01 4.35546726e-01
-1.89775974e-01 6.10868096e-01 4.17070955e-01 3.84347975e-01
3.53609800e-01 3.10840607e-01 -1.89134434e-01 -1.92806736e-01
-9.72648025e-01 1.49388000e-01 5.18300176e-01 9.84456301e-01
6.33462548e-01 5.42931370e-02 -3.28071892e-01 3.89916956e-01
2.79464334e-01 5.65300167e-01 6.58131063e-01 -8.57264519e-01
6.25866055e-01 2.35916898e-01 2.36455128e-01 -1.70354307e+00
-1.28857926e-01 -1.54533893e-01 -5.70349991e-01 -1.94024488e-01
6.79446012e-02 3.98121178e-01 -8.51747632e-01 1.91463900e+00
-4.06378284e-02 5.76968014e-01 -1.73202619e-01 1.13841105e+00
9.86053169e-01 8.00683379e-01 -4.20973264e-02 -6.31941736e-01
1.24417090e+00 -1.00435591e+00 -1.25172734e+00 -5.39464295e-01
4.00108904e-01 -8.91104937e-01 1.07003653e+00 1.26525804e-01
-1.12181878e+00 -7.23775864e-01 -1.10625815e+00 -3.76576573e-01
-7.67858177e-02 -1.29151449e-01 3.58999483e-02 7.85787404e-02
-1.14504576e+00 5.17658889e-01 -5.98903477e-01 -5.34278035e-01
2.36350581e-01 -1.30686000e-01 -4.05901939e-01 -4.27815616e-01
-1.24116600e+00 8.77838492e-01 3.01815420e-01 1.52265891e-01
-7.70384073e-01 -3.90913785e-01 -1.00779963e+00 9.55712572e-02
5.96550763e-01 -7.46147096e-01 1.24740303e+00 -1.76007593e+00
-9.52446997e-01 8.77680421e-01 -5.45970917e-01 -3.58764231e-01
4.55072612e-01 -6.63512945e-02 -6.01530969e-01 2.45204687e-01
3.99957150e-01 6.62429512e-01 1.37155211e+00 -1.32264543e+00
-5.82249343e-01 -3.19642365e-01 1.62703827e-01 2.79567033e-01
-2.67249882e-01 -1.59267168e-02 -7.70672143e-01 -9.21766520e-01
1.79542065e-01 -5.24500370e-01 2.97970802e-01 2.56407827e-01
-3.08209211e-01 1.00483663e-01 8.43263924e-01 -9.22165990e-01
1.49375236e+00 -2.46406007e+00 2.39740431e-01 -1.59377843e-01
3.44239444e-01 8.04596320e-02 -2.50716478e-01 4.11812544e-01
-7.15903193e-02 4.25553709e-01 -1.72598854e-01 -5.65490603e-01
-2.84746140e-01 3.62838745e-01 -5.02200544e-01 5.69160581e-01
1.66116461e-01 1.01645744e+00 -1.13507247e+00 -5.46735764e-01
1.44888967e-01 5.85646093e-01 -4.71090794e-01 1.24818988e-01
-1.30783111e-01 2.99513340e-01 -2.95833290e-01 3.17897111e-01
6.31288886e-01 -3.44719082e-01 1.60785258e-01 -7.43587673e-01
-9.75973010e-02 1.36497691e-01 -8.79538059e-01 1.94271255e+00
-2.53247589e-01 1.00701928e+00 -3.52733612e-01 -1.05022645e+00
5.59179366e-01 3.02959532e-01 2.59814233e-01 -9.78018522e-01
1.36087820e-01 1.08706638e-01 -3.66053313e-01 -1.05666268e+00
7.72356987e-01 -1.44962385e-01 -3.21434885e-02 3.47220421e-01
5.97906783e-02 -1.96547493e-01 2.21678078e-01 4.51181233e-01
9.28523600e-01 6.29429892e-02 5.36414325e-01 3.71150315e-01
2.55474240e-01 -1.89452887e-01 5.04888833e-01 9.84685659e-01
-5.10542929e-01 8.71086419e-01 5.04179597e-01 -5.84689677e-01
-1.18480349e+00 -8.25622201e-01 2.02986509e-01 7.71005511e-01
7.01375008e-01 -6.61652446e-01 -5.68572402e-01 -4.42248642e-01
-6.50073215e-02 7.75869489e-01 -5.53170145e-01 -5.47240138e-01
-4.89575714e-01 -1.52419880e-01 2.62646794e-01 4.04970586e-01
6.17738664e-01 -9.01278377e-01 -5.82591176e-01 8.84853825e-02
-6.60202622e-01 -1.49979591e+00 -9.54339683e-01 -4.08935696e-01
-5.66664755e-01 -9.02472675e-01 -1.68424815e-01 -8.57494712e-01
7.58798242e-01 7.66862512e-01 1.00523305e+00 4.45142239e-01
1.06198683e-01 3.42320621e-01 -5.74050009e-01 1.22204483e-01
-4.75392252e-01 -5.75478077e-01 -5.43829426e-03 3.01112890e-01
1.65635556e-01 -9.37568322e-02 -4.99963880e-01 -5.15989587e-02
-1.22470045e+00 4.35601532e-01 2.76068091e-01 7.48244882e-01
6.67424917e-01 6.26879409e-02 2.88598746e-01 -7.44748771e-01
4.84082371e-01 -4.77048218e-01 -5.29877357e-02 3.66233110e-01
-4.93671685e-01 -5.38695157e-02 5.97561896e-01 -3.23632360e-01
-9.35073376e-01 7.29110613e-02 -1.29630998e-01 -7.06323028e-01
1.22706875e-01 6.36862934e-01 -5.64592592e-02 1.32171586e-01
2.14398727e-01 3.60038817e-01 -2.62532264e-01 -2.20888525e-01
3.21184367e-01 5.61139226e-01 6.27265334e-01 4.01702337e-02
5.94281793e-01 6.61536634e-01 -3.48559499e-01 -5.95602930e-01
-1.23226714e+00 -3.92386556e-01 -5.33984542e-01 -5.41361570e-01
9.18530345e-01 -1.03582227e+00 -2.94273525e-01 4.66870636e-01
-1.66262734e+00 1.00138851e-01 -1.57738283e-01 2.31066495e-01
-4.39521313e-01 7.86845446e-01 -3.60037118e-01 -3.55376184e-01
6.59952015e-02 -1.23405325e+00 1.13489389e+00 4.94850762e-02
-2.63149440e-01 -8.22676003e-01 -3.01482618e-01 3.01012397e-01
2.96609938e-01 8.70966092e-02 5.50851285e-01 -2.65748680e-01
-5.55781305e-01 -9.08925533e-02 -3.18609506e-01 2.93358713e-01
2.87447482e-01 1.32320106e-01 -7.19116330e-01 -5.13613105e-01
3.67443770e-01 -2.13895410e-01 1.03534377e+00 2.63061911e-01
1.36326659e+00 -5.59473932e-01 -1.58310011e-01 6.78281128e-01
1.48807287e+00 2.72300154e-01 6.75119758e-01 2.77553886e-01
9.36352253e-01 2.18641311e-01 6.31687462e-01 5.05257607e-01
4.17841077e-01 8.35533917e-01 5.97892046e-01 1.69798225e-01
-3.93362612e-01 -5.33374131e-01 5.09461820e-01 9.69262838e-01
2.76831746e-01 -6.68059409e-01 -6.72079027e-01 4.71678942e-01
-1.95651066e+00 -1.22752130e+00 -1.42985210e-01 2.01120877e+00
7.86500931e-01 -4.13510725e-02 -5.92259943e-01 -1.01939149e-01
9.98868167e-01 4.65296507e-01 -4.91544992e-01 -3.09732556e-01
-2.07579032e-01 -2.90448725e-01 1.30367383e-01 5.88495135e-01
-1.00163949e+00 1.05985117e+00 5.78572321e+00 5.61722934e-01
-1.38451326e+00 2.98441976e-01 5.29594004e-01 -1.81748360e-01
-4.36457008e-01 1.33385239e-02 -3.23706001e-01 7.51109719e-01
6.33734703e-01 -2.48809859e-01 4.60896254e-01 3.90765131e-01
6.34708285e-01 -1.67064011e-01 -1.19936645e+00 1.31192780e+00
9.69634831e-01 -1.35237598e+00 5.50828934e-01 -5.48717141e-01
7.33211696e-01 -3.87177646e-01 1.06012695e-01 4.48075309e-02
-5.01439691e-01 -9.16669607e-01 1.28986752e+00 6.86395586e-01
9.30303276e-01 -4.00325984e-01 8.63215268e-01 5.93475699e-02
-1.12538910e+00 6.18858784e-02 -3.90164614e-01 -1.67627051e-01
3.13603133e-01 4.18073952e-01 -3.89557034e-01 2.70098031e-01
8.49666476e-01 1.28851044e+00 -8.15747440e-01 8.02991152e-01
-3.49277645e-01 2.39076123e-01 2.83923537e-01 1.34001836e-01
2.67858505e-02 1.06255881e-01 6.36844516e-01 1.10918701e+00
5.51380455e-01 4.92739608e-04 5.28493896e-02 8.92388403e-01
-3.61871183e-01 -4.25378140e-03 -7.02048182e-01 -1.69613063e-01
4.77038950e-01 7.72798181e-01 -5.24585783e-01 -5.02645671e-01
-7.79220641e-01 1.66009688e+00 3.36875081e-01 5.84912181e-01
-1.19745922e+00 -8.15199316e-02 5.57531953e-01 -2.71872524e-02
4.65748936e-01 8.45751837e-02 -1.62242219e-01 -1.58552289e+00
4.32494074e-01 -6.63917601e-01 7.51923248e-02 -1.52706528e+00
-1.07117116e+00 4.63913202e-01 -3.81900698e-01 -1.30086172e+00
-1.47497479e-03 -2.16671467e-01 -2.78683513e-01 4.07103747e-01
-1.54272866e+00 -8.63778174e-01 -6.81720376e-01 5.70651770e-01
9.53069985e-01 1.10346936e-01 4.36750293e-01 3.06957304e-01
-4.78905380e-01 3.63332868e-01 -1.38010653e-02 5.09756394e-02
9.34548855e-01 -7.69505143e-01 2.84328163e-01 1.20988059e+00
2.18562424e-01 3.93715799e-01 1.02974141e+00 -8.40629876e-01
-1.36874962e+00 -1.37040508e+00 1.19548714e+00 -2.10887790e-01
7.95396507e-01 -2.16577202e-01 -1.08042157e+00 9.43550885e-01
3.52622539e-01 -1.42630329e-02 2.56522387e-01 -7.66759694e-01
-4.48250473e-01 -1.12892315e-01 -9.22030985e-01 5.87950766e-01
1.14617884e+00 -8.37988496e-01 -6.88907683e-01 4.34410125e-01
1.06797135e+00 -6.41825616e-01 -2.16527030e-01 5.03422618e-02
5.33994138e-01 -1.13645852e+00 8.21008265e-01 -5.60691118e-01
9.10112441e-01 -4.60644603e-01 -2.64109164e-01 -1.42503309e+00
-2.96793073e-01 -2.34331742e-01 -2.20589533e-01 1.12109435e+00
2.40502860e-02 -8.33295882e-02 1.56398192e-01 3.88755262e-01
-2.60774821e-01 -2.92914122e-01 -1.05035782e+00 -5.67348540e-01
-3.65561843e-01 -6.06172562e-01 5.26451290e-01 1.08837533e+00
-6.93655163e-02 2.69492626e-01 -8.15732479e-01 1.67122304e-01
2.24746376e-01 9.95061398e-02 3.42205286e-01 -6.20934308e-01
-9.31693763e-02 -2.13324606e-01 -7.26704478e-01 -1.16598487e+00
3.06419134e-01 -7.90734291e-01 2.92535514e-01 -1.60857749e+00
6.68542147e-01 3.60238463e-01 -1.24073341e-01 3.10815483e-01
-3.02239895e-01 3.28460783e-01 4.51155692e-01 3.76649082e-01
-9.47689116e-01 4.74086463e-01 1.28343010e+00 -4.57373857e-01
1.48772776e-01 -5.61195850e-01 -6.35434568e-01 6.61971390e-01
6.69480026e-01 -6.05847776e-01 -2.81289220e-01 -9.99220550e-01
6.00083470e-01 1.52348056e-01 6.31226659e-01 -6.76537395e-01
3.60813826e-01 -2.05501243e-01 3.17742616e-01 -3.44731569e-01
2.75219083e-01 -1.02355719e+00 1.88482881e-01 4.17134225e-01
-3.55044335e-01 4.49889272e-01 -6.19206801e-02 8.91658008e-01
-5.57789743e-01 -3.08441907e-01 7.36024439e-01 -1.94975615e-01
-1.15012813e+00 8.80560055e-02 -4.22357649e-01 2.03212234e-03
1.11762607e+00 -3.10487866e-01 -5.68070710e-01 -7.89980054e-01
-6.63141370e-01 1.31962642e-01 6.93216681e-01 6.74637973e-01
1.20866323e+00 -1.44340742e+00 -5.70477188e-01 1.63176194e-01
4.06755865e-01 -3.69377375e-01 3.98261189e-01 8.54138792e-01
-5.49533963e-01 2.02555060e-01 -1.13438651e-01 -5.31570733e-01
-1.36203015e+00 7.59508550e-01 4.27137673e-01 2.71327078e-01
-5.35582185e-01 6.31252646e-01 3.06402385e-01 3.82000715e-01
2.14789033e-01 -5.90175390e-01 -1.44189715e-01 3.74756157e-02
7.97471702e-01 -1.65220588e-01 -1.24518618e-01 -1.03116930e+00
-3.93935740e-01 5.84740639e-01 -9.04507097e-03 1.37598306e-01
1.09445775e+00 -8.38312566e-01 -2.46096402e-01 7.39532351e-01
1.38682103e+00 -1.85268387e-01 -1.10359573e+00 -2.18172640e-01
-3.77708346e-01 -9.03794646e-01 1.41932011e-01 -5.70446610e-01
-1.28128695e+00 6.98568344e-01 5.13423383e-01 5.84945455e-02
1.16813743e+00 1.67344257e-01 8.66181970e-01 3.02896559e-01
1.71950534e-01 -1.18381441e+00 6.00733459e-01 4.18187141e-01
1.14805388e+00 -1.52654171e+00 6.35490268e-02 -3.39512914e-01
-8.53296936e-01 1.26947594e+00 6.36828005e-01 1.18582271e-01
2.31539041e-01 -2.12649833e-02 -4.44556922e-02 -2.19343990e-01
-7.59938121e-01 6.37321500e-03 9.56370980e-02 3.66067290e-01
2.84312606e-01 -1.02126986e-01 -3.27278793e-01 4.43581611e-01
8.12266991e-02 6.97226450e-02 9.56577778e-01 8.84297311e-01
-4.98815358e-01 -5.55090725e-01 -2.59426117e-01 2.88425714e-01
-2.52867043e-01 -3.42245191e-01 -4.85500395e-01 4.95734423e-01
3.26787204e-01 1.09374523e+00 4.07041579e-01 -5.41568816e-01
2.81448066e-01 -2.85481475e-03 3.73114318e-01 -5.04096270e-01
-3.38267177e-01 -1.11309856e-01 -1.96902350e-01 -8.37137282e-01
-8.57853711e-01 -5.53507030e-01 -1.38792181e+00 -5.17958522e-01
2.75619328e-03 -2.70708501e-01 5.77498376e-01 1.13961506e+00
2.73937732e-01 6.19153500e-01 5.51946938e-01 -5.07467210e-01
-1.32519051e-01 -6.30442500e-01 -5.48956633e-01 1.01928806e+00
6.01670980e-01 -5.80712855e-01 -5.61667383e-01 5.43086350e-01] | [10.269360542297363, 0.8659043312072754] |
5152230a-731a-460d-baad-0388c8871e11 | structural-imbalance-aware-graph-augmentation | 2303.13757 | null | https://arxiv.org/abs/2303.13757v1 | https://arxiv.org/pdf/2303.13757v1.pdf | Structural Imbalance Aware Graph Augmentation Learning | Graph machine learning (GML) has made great progress in node classification, link prediction, graph classification and so on. However, graphs in reality are often structurally imbalanced, that is, only a few hub nodes have a denser local structure and higher influence. The imbalance may compromise the robustness of existing GML models, especially in learning tail nodes. This paper proposes a selective graph augmentation method (SAug) to solve this problem. Firstly, a Pagerank-based sampling strategy is designed to identify hub nodes and tail nodes in the graph. Secondly, a selective augmentation strategy is proposed, which drops the noisy neighbors of hub nodes on one side, and discovers the latent neighbors and generates pseudo neighbors for tail nodes on the other side. It can also alleviate the structural imbalance between two types of nodes. Finally, a GNN model will be retrained on the augmented graph. Extensive experiments demonstrate that SAug can significantly improve the backbone GNNs and achieve superior performance to its competitors of graph augmentation methods and hub/tail aware methods. | ['Zheng Liu', 'Kejia-Chen', 'Zulong Liu'] | 2023-03-24 | null | null | null | null | ['graph-classification'] | ['graphs'] | [-1.08658098e-01 6.10267997e-01 -6.94952250e-01 1.03182912e-01
-1.09563150e-01 -2.32399791e-01 4.34772462e-01 4.23740566e-01
2.26832017e-01 8.60000670e-01 1.46247506e-01 -4.04161602e-01
-1.83645487e-02 -1.16610992e+00 -2.99956679e-01 -8.93292367e-01
-5.02740204e-01 7.51937270e-01 5.07603884e-01 -9.22376513e-02
-9.80945230e-02 3.08864117e-01 -7.07121193e-01 -3.61744404e-01
8.97923052e-01 4.24507082e-01 -2.81019974e-02 2.86026359e-01
-4.05091554e-01 1.04548252e+00 -2.84092069e-01 -4.35977429e-01
2.11649433e-01 -3.46507847e-01 -7.59264410e-01 9.95968059e-02
2.52418995e-01 -2.12758277e-02 -8.19540739e-01 1.22135305e+00
4.86294687e-01 -4.04185615e-02 3.80764991e-01 -1.70967543e+00
-3.73553395e-01 1.11596441e+00 -1.21611226e+00 1.50042593e-01
1.74813569e-01 -1.09003857e-01 1.29971862e+00 -6.32370770e-01
6.83095753e-01 1.26521432e+00 8.07843029e-01 2.62647063e-01
-1.10200787e+00 -6.17281675e-01 5.03112137e-01 1.77362785e-01
-1.51992059e+00 3.49870883e-02 1.30136180e+00 -3.97732630e-02
2.93610901e-01 2.73598194e-01 7.89317012e-01 7.82268465e-01
-1.33879978e-04 8.06363106e-01 7.51950145e-01 -2.02078506e-01
-1.33591043e-02 -1.46611974e-01 1.04221165e-01 8.53866518e-01
5.80331683e-01 -3.35563868e-01 -3.51868778e-01 -3.19524646e-01
7.48835564e-01 2.01141089e-01 -2.04692274e-01 -6.07124329e-01
-1.01793945e+00 9.39703941e-01 1.01676488e+00 2.64205992e-01
-3.19074184e-01 2.69757230e-02 4.61794198e-01 2.47972354e-01
6.89673364e-01 1.98608533e-01 -3.94660443e-01 4.19068545e-01
-8.94567728e-01 -2.51546562e-01 9.09807980e-01 9.26403344e-01
9.97087777e-01 -6.16572052e-02 -2.15267047e-01 6.56393588e-01
4.38435078e-01 1.39838830e-01 1.74562812e-01 -3.41766000e-01
4.55677241e-01 1.25879717e+00 -5.34530878e-01 -1.43387699e+00
-5.04486918e-01 -9.06652927e-01 -1.49492431e+00 -4.40184653e-01
-8.17767307e-02 -1.00099221e-02 -1.00662494e+00 1.59393549e+00
6.78638756e-01 3.17561984e-01 -5.19533813e-01 7.61065841e-01
9.46465850e-01 4.98589516e-01 2.35731192e-02 -2.65070677e-01
1.00134122e+00 -1.26731169e+00 -6.77499115e-01 -3.60540926e-01
8.14814806e-01 -5.10738730e-01 9.14586186e-01 -1.58777073e-01
-7.74155676e-01 -6.85111657e-02 -7.91493535e-01 2.99020052e-01
-2.06733778e-01 -1.87030643e-01 9.36235905e-01 3.98914844e-01
-1.11592686e+00 3.73605937e-01 -6.38264000e-01 -3.22554529e-01
4.92053121e-01 2.84883618e-01 -2.62888670e-01 -2.86971211e-01
-1.25755930e+00 2.39160374e-01 5.60601294e-01 -2.39047408e-02
-7.20700085e-01 -6.09240472e-01 -1.06513035e+00 1.50392354e-01
6.32682741e-01 -5.39950550e-01 6.47068262e-01 -7.80619442e-01
-7.22213030e-01 5.52956879e-01 -9.90674719e-02 -2.74347812e-01
4.21301484e-01 4.11124885e-01 -5.04668236e-01 5.08052157e-03
4.01542813e-01 3.78226727e-01 6.39529884e-01 -1.19378209e+00
-3.54984879e-01 -4.87716019e-01 -7.54135549e-02 1.16471484e-01
-6.58039570e-01 -5.36249757e-01 -8.96540701e-01 -7.46363163e-01
6.51254058e-01 -8.35357785e-01 -6.61836505e-01 -3.57317746e-01
-1.06028223e+00 -2.76449054e-01 9.92830157e-01 -3.71721566e-01
1.62641191e+00 -1.87945223e+00 9.96301300e-04 9.96001780e-01
1.04491246e+00 2.67210066e-01 -4.25456613e-01 6.41849935e-01
-2.24303395e-01 2.79988766e-01 1.32150650e-01 -1.69048816e-01
-3.52880061e-01 2.72438109e-01 -1.71046630e-02 4.56414640e-01
-1.63129613e-01 9.82839942e-01 -1.11549485e+00 -7.82371163e-01
-6.58782795e-02 1.71927765e-01 -4.99721080e-01 1.77228469e-02
-1.20522171e-01 3.04684013e-01 -5.41947365e-01 7.49049544e-01
9.23433125e-01 -8.14770222e-01 7.32677042e-01 -1.19076349e-01
3.40111643e-01 3.98998141e-01 -1.17087257e+00 1.08157194e+00
-2.22606417e-02 2.37661198e-01 8.97545740e-02 -9.01872694e-01
1.08515680e+00 5.36813736e-02 5.10722756e-01 -4.07482833e-01
1.50475521e-02 -5.55638708e-02 1.56019881e-01 -1.68293238e-01
1.17241926e-01 2.98396111e-01 2.75054455e-01 3.78974617e-01
-2.79089838e-01 4.66732562e-01 4.72129971e-01 9.38544333e-01
1.54867554e+00 -4.57500726e-01 8.00549090e-02 -1.69638395e-01
6.33259356e-01 -1.70054987e-01 8.21061075e-01 5.48707247e-01
-1.55328244e-01 3.02929699e-01 1.07392395e+00 -4.59683299e-01
-8.83645773e-01 -8.10843050e-01 5.56456625e-01 1.06370175e+00
2.66083807e-01 -8.74878287e-01 -6.07916236e-01 -9.79232073e-01
8.06183666e-02 1.15022957e-01 -5.92205584e-01 -5.36414981e-01
-4.67533588e-01 -9.04266000e-01 2.52380848e-01 3.50705177e-01
5.22379875e-01 -8.06968927e-01 6.25967264e-01 1.32442236e-01
-2.20221654e-01 -1.03677344e+00 -5.32160997e-01 -1.55195758e-01
-1.09401190e+00 -1.33825815e+00 -3.89392018e-01 -9.77984905e-01
1.32055783e+00 6.50442064e-01 1.28579903e+00 7.36637592e-01
4.60179932e-02 -2.07440004e-01 -5.12869537e-01 5.48521094e-02
-2.08547175e-01 7.33485639e-01 -3.25260721e-02 -2.53139678e-02
8.88049006e-02 -8.99862468e-01 -6.09719574e-01 3.15736383e-01
-7.29138136e-01 1.85060725e-01 9.44188774e-01 8.57282102e-01
5.22866547e-01 4.56930220e-01 6.10689819e-01 -1.54603624e+00
6.83711886e-01 -6.34714484e-01 -4.48657602e-01 1.73553169e-01
-1.06668150e+00 -5.83321974e-02 6.29919946e-01 -2.91005492e-01
-5.41849971e-01 -7.87044764e-02 5.52446805e-02 -3.26303601e-01
5.17464995e-01 9.04694796e-01 -5.37075698e-01 -3.64994168e-01
2.38894030e-01 6.45151958e-02 7.02566728e-02 -5.89223027e-01
1.82086647e-01 1.30534038e-01 1.88108012e-01 -9.05293897e-02
1.34661710e+00 2.97825456e-01 4.75319207e-01 -6.43860579e-01
-7.98926055e-01 -4.52343524e-01 -5.33228397e-01 -1.88668683e-01
4.72458489e-02 -8.47653329e-01 -4.31875378e-01 3.79428536e-01
-7.28196383e-01 -2.63907909e-01 -2.58188434e-02 2.43225098e-01
2.55568594e-01 5.95484257e-01 -8.24940562e-01 -6.13216460e-01
-4.63615477e-01 -6.49127364e-01 7.05452383e-01 3.19722414e-01
5.88052571e-02 -1.25859404e+00 1.55223116e-01 2.53390461e-01
2.29505047e-01 2.14723513e-01 1.18334389e+00 -8.33210170e-01
-8.14296007e-01 -5.83736956e-01 -5.67106426e-01 1.42751560e-02
2.65815049e-01 5.16399033e-02 -3.42949390e-01 -6.56890452e-01
-7.88971901e-01 2.43173330e-03 1.00115860e+00 1.04798719e-01
1.02082849e+00 -6.50789976e-01 -8.79846871e-01 5.02033353e-01
1.15171349e+00 -3.77011389e-01 5.65480590e-01 1.00143813e-01
1.17974544e+00 2.92144239e-01 4.53144461e-01 1.73931375e-01
6.03792012e-01 2.30038911e-01 8.18773985e-01 -5.65402925e-01
-3.26493979e-01 -6.92138255e-01 1.46241903e-01 1.10273492e+00
6.47837818e-02 -3.73750001e-01 -9.32141483e-01 5.46712816e-01
-1.87040973e+00 -4.58517283e-01 -5.41311741e-01 2.09223366e+00
6.98086262e-01 3.23404223e-01 2.08408937e-01 3.04689288e-01
1.05891490e+00 5.01589119e-01 -5.85509658e-01 3.14342529e-01
-1.84791967e-01 -1.46438688e-01 6.24135494e-01 3.97353679e-01
-9.84197140e-01 1.06370258e+00 5.82674837e+00 9.60040987e-01
-6.40253544e-01 3.35104950e-02 7.55513728e-01 3.78767252e-01
-5.65520108e-01 2.85939068e-01 -5.49938142e-01 5.01893282e-01
3.63307685e-01 -1.25864372e-01 2.98512369e-01 1.04788661e+00
4.06607985e-02 1.95947990e-01 -4.88454014e-01 5.84654570e-01
-8.02662373e-02 -1.24652159e+00 3.47367138e-01 3.78051817e-01
1.00200510e+00 1.06510617e-01 -3.20756167e-01 2.51590282e-01
5.75235486e-01 -8.80125761e-01 -1.56201243e-01 2.95095593e-01
6.21208787e-01 -1.01251030e+00 9.94790137e-01 5.02515197e-01
-1.49503517e+00 9.72835869e-02 -4.83339250e-01 6.44982383e-02
-1.46227226e-01 1.17435360e+00 -1.04896629e+00 7.40064204e-01
4.35084105e-01 7.07973957e-01 -8.35662067e-01 1.23553061e+00
-4.97254640e-01 8.48136365e-01 -3.09435517e-01 -1.24059297e-01
2.09239304e-01 -4.63324457e-01 8.00981998e-01 6.93550527e-01
-1.11319967e-01 -2.65495002e-01 6.38684869e-01 6.15197539e-01
-5.93491852e-01 2.59097904e-01 -5.36930323e-01 -1.60621807e-01
7.64216185e-01 1.74588013e+00 -9.60200548e-01 -1.16259694e-01
-3.46430451e-01 7.48196602e-01 5.75841784e-01 4.50012356e-01
-6.53114200e-01 -5.33298969e-01 3.81517857e-01 5.38233817e-01
-3.78271677e-02 -9.12324637e-02 1.90617710e-01 -1.03219736e+00
-9.35487673e-02 -7.31900394e-01 7.03283250e-01 -4.06015515e-01
-1.48354256e+00 5.31519473e-01 -5.62347531e-01 -9.13929880e-01
2.02996060e-01 -1.28537834e-01 -1.07427227e+00 5.95352054e-01
-1.42331660e+00 -1.25496829e+00 -5.24585545e-01 5.74019372e-01
-8.51441249e-02 -8.58723447e-02 4.67426687e-01 5.43594360e-01
-8.78874660e-01 7.30524421e-01 -6.96865842e-02 4.82614875e-01
3.48434091e-01 -1.34352553e+00 4.11064088e-01 8.76864910e-01
3.13348956e-02 5.96815050e-01 4.05561924e-01 -1.20506692e+00
-1.23362541e+00 -1.44398081e+00 8.62331688e-01 1.48416862e-01
7.57129371e-01 -3.06547254e-01 -9.63502407e-01 7.73428440e-01
-2.15362936e-01 2.43258193e-01 6.02209091e-01 3.87578577e-01
-3.63522679e-01 -2.80938029e-01 -9.94800746e-01 8.82028043e-01
1.13562715e+00 -3.06138873e-01 1.76744774e-01 7.27499127e-01
7.99868643e-01 -1.54875398e-01 -8.09928536e-01 6.53517246e-01
1.51851639e-01 -6.95033610e-01 8.94649327e-01 -7.33200371e-01
1.11217320e-01 -4.51406747e-01 5.02760231e-01 -1.30271661e+00
-5.69796324e-01 -7.73035109e-01 -5.45476615e-01 1.57835364e+00
5.19616723e-01 -6.88297868e-01 1.40820825e+00 2.33657975e-02
2.79243916e-01 -9.48778272e-01 -6.88622832e-01 -6.13253355e-01
-4.50562686e-01 2.26892233e-01 7.69334912e-01 1.19137049e+00
-1.63299829e-01 8.80526721e-01 -4.89405781e-01 1.63478598e-01
7.82775581e-01 1.19306907e-01 1.01501656e+00 -1.63525772e+00
1.81276396e-01 -4.22762305e-01 -5.35729706e-01 -9.48769689e-01
8.56289566e-02 -1.21337867e+00 -4.33960468e-01 -1.85432458e+00
5.21793306e-01 -7.22597420e-01 -3.05241674e-01 6.19887114e-01
-5.36204994e-01 3.83410513e-01 -1.66702300e-01 3.45215201e-01
-8.98120821e-01 5.27867317e-01 1.52122605e+00 -2.17366561e-01
-2.83566326e-01 2.77733326e-01 -7.59678841e-01 7.89270461e-01
9.55997407e-01 -4.42618787e-01 -5.60075164e-01 9.55614001e-02
6.24060571e-01 -1.87753007e-01 3.47828209e-01 -7.52370536e-01
3.24824333e-01 9.66273323e-02 3.12940329e-01 -7.57854700e-01
-1.21526480e-01 -8.13010395e-01 7.22770691e-02 5.40004194e-01
-9.53531563e-02 -6.13191165e-03 -3.67571741e-01 8.21418405e-01
-1.26104383e-03 1.11642875e-01 6.10771596e-01 -9.52836201e-02
-3.70363504e-01 8.63590717e-01 7.49400184e-02 7.44030103e-02
8.37778568e-01 2.06259061e-02 -4.79101658e-01 -5.76132536e-01
-5.51079214e-01 8.06647182e-01 3.85056764e-01 2.33398601e-01
5.73162615e-01 -1.47024179e+00 -6.25781536e-01 2.51516044e-01
1.13792300e-01 2.07171991e-01 4.23391834e-02 1.14281404e+00
-4.15608883e-01 1.07267881e-02 2.34287113e-01 -2.03444302e-01
-1.35949385e+00 5.86342812e-01 1.23733737e-01 -8.97789359e-01
-5.95915556e-01 9.31620121e-01 3.50196399e-02 -5.73140025e-01
2.67817676e-01 4.36446697e-01 -2.63742805e-01 7.64739560e-03
1.03950858e-01 5.12312293e-01 -3.27950716e-02 -5.20795107e-01
-3.09607744e-01 8.42365846e-02 -5.22487819e-01 6.00834131e-01
1.20934367e+00 -2.72669822e-01 -8.68347883e-01 1.72642916e-01
9.39125061e-01 2.80132651e-01 -6.78379178e-01 -5.95033109e-01
1.21221177e-01 -4.36508894e-01 1.43109366e-01 -4.60382372e-01
-1.70012939e+00 5.37786782e-01 -8.05008113e-02 6.83184505e-01
1.09608972e+00 1.86862499e-01 9.67290461e-01 2.59834886e-01
2.10600674e-01 -8.59296739e-01 -1.77420042e-02 4.01414752e-01
4.19407934e-01 -1.02975380e+00 2.61503100e-01 -1.08679819e+00
-2.28855863e-01 7.78537333e-01 9.37746584e-01 -3.73198502e-02
7.44559705e-01 5.47567047e-02 -3.43094677e-01 -4.35762137e-01
-7.07145095e-01 -2.88665056e-01 1.81943953e-01 6.20406687e-01
1.99262306e-01 1.55238554e-01 -2.67478585e-01 3.06186080e-01
-5.36714979e-02 -5.89490175e-01 3.00724089e-01 5.31411767e-01
-4.35992867e-01 -1.25966001e+00 -2.93667540e-02 8.57281625e-01
-2.49202579e-01 -4.54215497e-01 -8.25403869e-01 8.37500334e-01
-1.35773480e-01 7.03151286e-01 -9.48319957e-02 -7.30149508e-01
-3.70613821e-02 -3.33733112e-01 -1.06290251e-01 -5.47974586e-01
-2.72208512e-01 1.15415953e-01 8.36560354e-02 -3.99139345e-01
-8.10948238e-02 -2.04468250e-01 -1.23524606e+00 -7.42534518e-01
-6.55761838e-01 6.25511765e-01 2.33897105e-01 5.71899593e-01
4.75015759e-01 7.42222369e-01 9.97343779e-01 -3.98792207e-01
-2.96431452e-01 -9.60194767e-01 -9.80575681e-01 1.52399555e-01
1.99508741e-01 -5.21066666e-01 -6.12162292e-01 -6.12768173e-01] | [7.197645664215088, 6.141395568847656] |
7c74530b-695c-4468-bb98-223c1c7db0a0 | look-listen-and-learn | 1705.08168 | null | http://arxiv.org/abs/1705.08168v2 | http://arxiv.org/pdf/1705.08168v2.pdf | Look, Listen and Learn | We consider the question: what can be learnt by looking at and listening to a
large number of unlabelled videos? There is a valuable, but so far untapped,
source of information contained in the video itself -- the correspondence
between the visual and the audio streams, and we introduce a novel
"Audio-Visual Correspondence" learning task that makes use of this. Training
visual and audio networks from scratch, without any additional supervision
other than the raw unconstrained videos themselves, is shown to successfully
solve this task, and, more interestingly, result in good visual and audio
representations. These features set the new state-of-the-art on two sound
classification benchmarks, and perform on par with the state-of-the-art
self-supervised approaches on ImageNet classification. We also demonstrate that
the network is able to localize objects in both modalities, as well as perform
fine-grained recognition tasks. | ['Relja Arandjelović', 'Andrew Zisserman'] | 2017-05-23 | look-listen-and-learn-1 | http://openaccess.thecvf.com/content_iccv_2017/html/Arandjelovic_Look_Listen_and_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Arandjelovic_Look_Listen_and_ICCV_2017_paper.pdf | iccv-2017-10 | ['sound-classification'] | ['audio'] | [ 4.23551083e-01 -1.39476344e-01 -9.39619541e-02 -3.54435742e-01
-1.04200399e+00 -6.63420200e-01 7.24777460e-01 2.45099328e-02
-2.68133491e-01 4.12629664e-01 3.39860260e-01 2.03582674e-01
4.58298549e-02 -3.45949292e-01 -1.05620265e+00 -6.36838257e-01
-3.60187441e-01 3.51577222e-01 2.56284654e-01 4.91645634e-02
5.74061275e-02 2.13599026e-01 -2.21485209e+00 9.13685977e-01
-7.01686591e-02 1.53488505e+00 1.04672186e-01 1.03629899e+00
1.15866706e-01 1.21554148e+00 -2.59939045e-01 -1.22717373e-01
5.08414768e-02 -4.02767122e-01 -1.17782390e+00 2.33953759e-01
1.00630438e+00 -2.92169511e-01 -4.58965600e-01 9.38677847e-01
2.84444809e-01 2.06257164e-01 5.60643077e-01 -1.26086211e+00
-7.08380759e-01 4.96635228e-01 -3.22795600e-01 5.43224454e-01
8.02934647e-01 6.29684851e-02 1.35903597e+00 -9.11767185e-01
7.20668793e-01 1.24086022e+00 7.81130612e-01 4.74210054e-01
-1.36340332e+00 -5.23030341e-01 1.47688359e-01 4.58597600e-01
-1.20376837e+00 -9.72733438e-01 6.88278973e-01 -6.87572062e-01
1.07179439e+00 2.35882029e-01 7.63403714e-01 1.49590051e+00
-2.37668291e-01 7.89384544e-01 1.07452595e+00 -6.59746528e-01
-1.93829718e-03 1.86381310e-01 -1.54425964e-01 7.94827580e-01
-4.00210112e-01 2.71949589e-01 -1.09312928e+00 7.24581331e-02
6.94158018e-01 -1.85299292e-01 -5.59387624e-01 -5.07124126e-01
-1.45375133e+00 5.32952547e-01 3.01501989e-01 5.19615531e-01
-4.95215245e-02 3.94815594e-01 6.48409128e-01 6.81774259e-01
4.33200538e-01 3.33428591e-01 -3.93035978e-01 -2.07133457e-01
-1.08116269e+00 -1.42123953e-01 7.84137011e-01 7.03402519e-01
7.38900781e-01 3.87414098e-01 1.73278868e-01 5.17589390e-01
2.88915366e-01 4.36310470e-01 5.38944602e-01 -1.11468422e+00
3.28425705e-01 -5.23181669e-02 -1.68317452e-01 -9.24328923e-01
-3.34075123e-01 -3.96729022e-01 -7.31059253e-01 3.15109760e-01
5.51599324e-01 2.50440240e-01 -8.34102631e-01 1.84153867e+00
3.24694514e-02 7.09092438e-01 -1.78924143e-01 8.40690911e-01
1.13946736e+00 6.21181428e-01 -1.39945775e-01 -2.19562888e-01
1.10640025e+00 -1.06343937e+00 -4.60839450e-01 -1.17530279e-01
3.11520528e-02 -6.09890521e-01 1.05790424e+00 4.93207157e-01
-1.28792548e+00 -1.01610708e+00 -1.15262401e+00 9.18587148e-02
-3.74636799e-01 5.65031879e-02 4.77335453e-01 2.48272523e-01
-1.31689322e+00 8.42652977e-01 -7.48029292e-01 -5.33360124e-01
5.14816225e-01 2.68890649e-01 -9.15992856e-01 -3.48929577e-02
-8.06617439e-01 5.73071599e-01 2.17306390e-01 1.77353974e-02
-1.61641204e+00 -7.86739409e-01 -9.93083835e-01 1.14142060e-01
4.19508219e-01 -5.68352342e-01 1.34881878e+00 -1.55889595e+00
-1.23739636e+00 1.28986537e+00 -6.72523156e-02 -3.97420824e-01
3.01206172e-01 -8.21133256e-02 -5.10984719e-01 7.82007277e-01
8.87474343e-02 8.62290740e-01 1.34248447e+00 -1.44573307e+00
-6.64849937e-01 -3.56336236e-01 1.18469112e-01 -2.35653426e-02
-3.78702730e-01 1.46051258e-01 -4.04822499e-01 -6.98097348e-01
-1.65338859e-01 -7.05497861e-01 3.09427202e-01 1.19833209e-01
-1.90746576e-01 -7.00379461e-02 6.86238289e-01 -6.01556242e-01
6.65574014e-01 -2.32006645e+00 3.60458463e-01 6.76812530e-02
3.02228779e-01 1.03255831e-01 -4.67134356e-01 3.13323349e-01
-4.01711136e-01 3.00232098e-02 1.74363106e-01 -3.99268597e-01
-6.16070442e-02 1.52374566e-01 -4.26581532e-01 5.47439635e-01
2.72813380e-01 7.83621013e-01 -9.60957348e-01 -4.52771336e-01
1.19967707e-01 4.57757086e-01 -3.67765963e-01 4.66198415e-01
-1.94189847e-01 6.46185279e-01 -2.60255486e-02 7.17618287e-01
1.39267340e-01 -4.16340649e-01 2.69318558e-02 -4.32669729e-01
2.35224277e-01 1.98278621e-01 -1.15086687e+00 1.97736394e+00
-2.93827325e-01 1.03837061e+00 3.65923971e-01 -1.50688195e+00
5.13211668e-01 6.92603290e-01 4.89673644e-01 -6.95489585e-01
8.02213028e-02 1.14985555e-01 -2.80157864e-01 -8.70850623e-01
-5.02972528e-02 -3.87300998e-01 1.96981624e-01 3.28505933e-01
1.01457393e+00 8.70824382e-02 6.85656592e-02 3.05896644e-02
1.13545763e+00 2.29461864e-01 3.40779692e-01 -1.77820958e-02
4.78047967e-01 -1.15348794e-01 1.95095286e-01 8.45529675e-01
-1.49255201e-01 7.77878582e-01 3.59376580e-01 -4.18991745e-01
-1.03855288e+00 -1.08976817e+00 -7.21370522e-03 1.55569279e+00
-4.37624529e-02 -4.91013318e-01 -5.38411081e-01 -6.01191342e-01
4.23084684e-02 -2.23347172e-03 -9.59864974e-01 6.20403774e-02
-2.05227792e-01 3.76431718e-02 6.14654601e-01 8.43036354e-01
1.00007951e-01 -1.34980488e+00 -3.07350546e-01 2.99960021e-02
-2.37102449e-01 -1.38675177e+00 -1.98484972e-01 4.41959143e-01
-6.22289956e-01 -1.15330076e+00 -5.83603680e-01 -9.77036715e-01
2.21536770e-01 2.22120732e-01 1.50777745e+00 1.36422083e-01
-4.92962122e-01 1.09533262e+00 -3.71228755e-01 -2.31721371e-01
-4.26347166e-01 -2.61435121e-01 -1.58367492e-02 3.58694077e-01
1.75787628e-01 -1.00593472e+00 -1.46297783e-01 1.68966308e-01
-7.67346740e-01 -3.30912828e-01 3.48747998e-01 9.62957144e-01
6.60772204e-01 -2.88143400e-02 5.97791076e-01 -5.61596811e-01
7.49087930e-02 -3.60067189e-01 -2.65686542e-01 2.17893049e-01
8.87017772e-02 -9.07038376e-02 6.65077269e-01 -5.39766669e-01
-6.69871628e-01 3.02586436e-01 -1.96915179e-01 -1.00539839e+00
-6.80664122e-01 2.91155219e-01 -4.12736014e-02 -4.01171118e-01
6.44172788e-01 1.13185249e-01 1.71033859e-01 -6.85633361e-01
4.54433084e-01 5.25551856e-01 1.04465604e+00 -3.45182359e-01
7.26585209e-01 7.60122180e-01 4.22037113e-03 -9.19465125e-01
-1.04650843e+00 -7.99447477e-01 -9.98774230e-01 -3.53677124e-01
1.05474353e+00 -1.15829217e+00 -8.19641888e-01 1.77388340e-01
-1.03198028e+00 -2.39167556e-01 -5.85791707e-01 4.92674619e-01
-9.25527513e-01 2.21570879e-01 -4.82766569e-01 -5.90377390e-01
1.27215937e-01 -8.35604966e-01 1.14062333e+00 -1.19900666e-01
-1.15875080e-01 -9.87488270e-01 2.11548179e-01 4.48960572e-01
1.19268008e-01 1.87843785e-01 8.11780274e-01 -7.68780649e-01
-6.07976139e-01 -1.23384155e-01 -2.34824926e-01 6.21363640e-01
-2.16302410e-01 -9.86540094e-02 -1.62177896e+00 -4.15456831e-01
1.03089456e-02 -9.82600987e-01 1.35098588e+00 1.04637973e-01
1.10548520e+00 -2.93914527e-01 -1.52969360e-01 6.43938482e-01
1.35222042e+00 -2.59195536e-01 3.19506794e-01 1.04337998e-01
7.47687340e-01 7.31079340e-01 1.05099447e-01 2.45179728e-01
2.31337264e-01 8.17763031e-01 7.62518227e-01 -7.55279809e-02
-5.81845224e-01 -4.00944471e-01 4.56094623e-01 8.53646517e-01
-2.41634056e-01 4.13863845e-02 -7.53148377e-01 6.25338793e-01
-1.79730046e+00 -1.33376169e+00 2.05563307e-01 2.00991511e+00
6.97203398e-01 -4.64677326e-02 3.23130906e-01 5.40063381e-01
4.90136981e-01 3.41660023e-01 -2.89258450e-01 3.72208096e-02
-1.70293450e-01 5.04497230e-01 -1.56142022e-02 2.58963883e-01
-1.58595645e+00 6.03386819e-01 7.34313154e+00 7.58937001e-01
-1.28901827e+00 3.19942445e-01 3.42564017e-01 -1.26601800e-01
-7.39839347e-03 -2.07958326e-01 -3.31836194e-01 1.41025290e-01
1.07037008e+00 3.54579419e-01 7.49408007e-01 7.63954341e-01
-2.73503006e-01 9.82675180e-02 -1.72724950e+00 1.34049785e+00
5.36607146e-01 -1.59144223e+00 3.60005051e-02 -2.34921321e-01
5.10008335e-01 1.51534840e-01 5.77612594e-02 1.65036291e-01
-4.54652086e-02 -1.33438492e+00 1.11081028e+00 6.06635869e-01
8.46916020e-01 -3.60477537e-01 5.82617223e-01 2.18550682e-01
-1.48286629e+00 -3.50818574e-01 -2.62615651e-01 -2.47351415e-02
4.15770011e-03 2.45725334e-01 -4.61548030e-01 4.38176334e-01
1.23456252e+00 1.14882934e+00 -8.05209041e-01 1.04053843e+00
-9.02834386e-02 6.11719787e-01 -1.02381974e-01 3.56758744e-01
2.07969129e-01 5.19217312e-01 4.96496320e-01 1.47423434e+00
9.57506299e-02 -2.29387745e-01 3.75811130e-01 6.52028263e-01
-2.54351556e-01 -4.13834713e-02 -9.02809024e-01 -2.32102990e-01
1.45640343e-01 1.22848141e+00 -5.46065629e-01 -3.85250002e-01
-6.29545689e-01 6.85064077e-01 4.26193982e-01 2.96767592e-01
-4.26673800e-01 -6.61740750e-02 4.13943768e-01 4.06979211e-02
6.38141811e-01 -9.95195843e-03 2.20572367e-01 -1.26854730e+00
1.06141105e-01 -1.10247433e+00 4.26552624e-01 -1.11130416e+00
-1.52165830e+00 6.15386546e-01 -3.46074641e-01 -1.34754622e+00
-4.62711185e-01 -6.67387426e-01 -6.14392340e-01 2.54741281e-01
-1.56885529e+00 -1.28276813e+00 -4.88251239e-01 1.03771639e+00
4.17935938e-01 -4.12719071e-01 1.05333924e+00 3.74935061e-01
4.14941758e-02 5.32105684e-01 1.49396127e-02 2.67440379e-01
6.14739716e-01 -1.11775804e+00 -9.94540974e-02 3.59424323e-01
1.16396248e+00 1.08481469e-02 5.48663020e-01 1.43496301e-02
-1.49060059e+00 -7.94772267e-01 5.02755344e-01 -6.12623274e-01
9.34987068e-01 -6.00666821e-01 -9.41107094e-01 8.09692800e-01
4.12738293e-01 6.36249483e-01 7.96329200e-01 1.72998786e-01
-9.47059929e-01 -2.07286373e-01 -7.35374928e-01 -8.37014839e-02
1.15352881e+00 -1.31979644e+00 -8.28338981e-01 4.49425906e-01
7.71396160e-01 -1.97101474e-01 -6.86414421e-01 2.61433095e-01
8.07939291e-01 -1.15734267e+00 1.42161453e+00 -1.09485173e+00
5.57121217e-01 -3.79482992e-02 -5.50739944e-01 -1.17891502e+00
-1.78447500e-01 -5.90027452e-01 -1.37438461e-01 1.37990427e+00
2.16082856e-01 -5.98354116e-02 5.94827950e-01 -2.43008509e-01
-1.98353026e-02 -4.60900396e-01 -1.06444192e+00 -8.78786206e-01
-3.35410893e-01 -7.22642243e-01 3.20408605e-02 9.83952284e-01
1.36638924e-01 5.87075889e-01 -5.70214748e-01 1.42520651e-01
6.79404020e-01 1.54478535e-01 6.37381792e-01 -1.40065777e+00
-6.52651668e-01 -2.93986350e-01 -9.60200965e-01 -7.77532220e-01
5.59464514e-01 -9.98125017e-01 1.37257725e-01 -1.13914347e+00
1.99814945e-01 1.96293760e-02 -4.58616406e-01 5.98355055e-01
4.25773650e-01 8.61459196e-01 4.12169158e-01 4.05350178e-01
-1.10997081e+00 3.32668036e-01 9.92827535e-01 -3.21040601e-01
2.28926897e-01 -7.98073038e-02 -5.25592327e-01 8.95796001e-01
2.09416062e-01 -3.28187913e-01 -2.03055352e-01 -3.82041126e-01
1.67959169e-01 3.07866871e-01 8.97444069e-01 -1.27097094e+00
2.54345924e-01 1.91038430e-01 5.84951282e-01 -3.76842469e-01
5.98604798e-01 -8.37824166e-01 -4.97514084e-02 -4.50615957e-02
-7.21455991e-01 -1.37607485e-01 2.03096926e-01 9.11931098e-01
-7.24811494e-01 -2.00283110e-01 7.66114473e-01 -3.94822419e-01
-7.14781046e-01 3.31518054e-02 -3.31527948e-01 3.06000024e-01
6.65601492e-01 -1.40894204e-01 -3.44784290e-01 -6.67346895e-01
-1.30356658e+00 -1.85518518e-01 1.68433428e-01 4.14649427e-01
4.85459954e-01 -1.56994963e+00 -4.71291244e-01 3.20441455e-01
2.43378997e-01 -3.78485709e-01 2.55050838e-01 9.74868000e-01
-4.35575955e-02 4.68345761e-01 -4.68880355e-01 -1.04672742e+00
-1.43405724e+00 8.71368170e-01 2.51989722e-01 1.30907139e-02
-6.49231434e-01 1.00463021e+00 2.23874137e-01 -1.34236544e-01
8.20175767e-01 -2.19542146e-01 -3.86752427e-01 4.09521341e-01
7.25107253e-01 1.29137203e-01 1.51879489e-01 -8.90411794e-01
-3.50975960e-01 8.37085068e-01 2.16064990e-01 -6.14953600e-02
1.47754395e+00 -1.41705751e-01 -3.81497778e-02 8.75923276e-01
1.52468693e+00 -1.92111194e-01 -1.20934379e+00 -5.23457885e-01
-1.40252322e-01 -6.07156694e-01 6.25582635e-02 -6.64241314e-01
-1.14503801e+00 1.42384768e+00 6.53048456e-01 6.99247122e-01
1.11040580e+00 4.62896794e-01 3.49592328e-01 6.34189188e-01
1.38217956e-01 -9.91709232e-01 6.55899644e-01 3.67906570e-01
9.33583975e-01 -1.48849499e+00 -2.46317133e-01 -1.55639753e-01
-5.65379083e-01 1.27202749e+00 3.08466345e-01 -1.91660255e-01
8.27183962e-01 2.26818442e-01 8.55529159e-02 -2.56742209e-01
-1.03744245e+00 -3.65827709e-01 5.94019294e-01 9.94676292e-01
2.96696156e-01 -3.13274026e-01 9.39336300e-01 5.69956779e-01
-2.81466283e-02 9.28426012e-02 2.46163622e-01 8.86228442e-01
-5.20847738e-01 -6.68310881e-01 -3.13206077e-01 2.97449261e-01
-6.55191302e-01 2.68938777e-04 -4.91296083e-01 7.51795053e-01
2.18232453e-01 9.53788638e-01 2.28156045e-01 -5.29636919e-01
2.23160431e-01 2.04364002e-01 7.84193635e-01 -6.43368900e-01
-6.32480979e-01 3.13976735e-01 9.81740057e-02 -9.21236277e-01
-1.02505910e+00 -6.01428032e-01 -9.12668347e-01 4.90652546e-02
-1.38104483e-01 -1.33202570e-02 4.16433573e-01 1.15291870e+00
9.34869349e-02 5.43730915e-01 6.75135076e-01 -1.59736919e+00
-5.29870689e-01 -7.42201686e-01 -7.43990719e-01 6.20612919e-01
9.09408987e-01 -7.72480130e-01 -6.56186104e-01 4.82147545e-01] | [9.947213172912598, 1.1065850257873535] |
216c02d5-3fca-471e-9008-363442485f27 | robust-automatic-monocular-vehicle-speed | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Revaud_Robust_Automatic_Monocular_Vehicle_Speed_Estimation_for_Traffic_Surveillance_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Revaud_Robust_Automatic_Monocular_Vehicle_Speed_Estimation_for_Traffic_Surveillance_ICCV_2021_paper.pdf | Robust Automatic Monocular Vehicle Speed Estimation for Traffic Surveillance | Even though CCTV cameras are widely deployed for traffic surveillance and have therefore the potential of becoming cheap automated sensors for traffic speed analysis, their large-scale usage toward this goal has not been reported yet. A key difficulty lies in fact in the camera calibration phase. Existing state-of-the-art methods perform the calibration using image processing or keypoint detection techniques that require high-quality video streams, yet typical CCTV footage is low-resolution and noisy. As a result, these methods largely fail in real-world conditions. In contrast, we propose two novel calibration techniques whose only inputs come from an off-the-shelf object detector. Both methods consider multiple detections jointly, leveraging the fact that cars have similar and well-known 3D shapes with normalized dimensions. The first one is based on minimizing an energy function corresponding to a 3D reprojection error, the second one instead learns from synthetic training data to predict the scene geometry directly. Noticing the lack of speed estimation benchmarks faithfully reflecting the actual quality of surveillance cameras, we introduce a novel dataset collected from public CCTV streams. Experimental results conducted on three diverse benchmarks demonstrate excellent speed estimation accuracy that could enable the wide use of CCTV cameras for traffic analysis, even in challenging conditions where state-of-the-art methods completely fail. Additional information can be found on our project web page: https://rebrand.ly/nle-cctv | ['Martin Humenberger', 'Jerome Revaud'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['vehicle-speed-estimation'] | ['computer-vision'] | [ 2.23494977e-01 -3.61310512e-01 -3.13925266e-01 -3.13497543e-01
-8.39065790e-01 -6.20091379e-01 4.33154851e-01 -3.67027104e-01
-3.99807692e-01 4.47964281e-01 -3.54782432e-01 -4.96200353e-01
2.81460404e-01 -6.14327729e-01 -7.82038391e-01 -6.68221295e-01
5.09726286e-01 5.80797613e-01 7.54335821e-01 -2.94375625e-02
9.82543528e-02 5.57943225e-01 -1.64782274e+00 -8.48524570e-02
8.42163205e-01 1.12800467e+00 5.97529374e-02 9.35554743e-01
1.97223127e-01 4.57547516e-01 -3.87961537e-01 -7.67839134e-01
6.10027313e-01 -3.51779163e-02 -1.75453693e-01 3.84044349e-01
8.38357210e-01 -6.95002139e-01 -5.27073860e-01 1.02355504e+00
2.04138458e-01 -2.06063554e-01 3.71906161e-01 -1.50406981e+00
-6.81908578e-02 -4.62502748e-01 -3.92002136e-01 3.79265964e-01
4.14611965e-01 5.76759398e-01 6.97642803e-01 -5.12368262e-01
4.18001413e-01 8.76540482e-01 7.19608486e-01 3.32051218e-01
-1.12212431e+00 -7.10153222e-01 -1.25119522e-01 3.86886626e-01
-1.38247323e+00 -6.62458837e-01 9.88121688e-01 -5.71710348e-01
5.94030261e-01 1.24107599e-01 8.06296587e-01 1.24768257e+00
1.16803730e-02 4.77554202e-01 9.96289909e-01 -7.79607072e-02
1.56339884e-01 4.39350039e-01 -1.62581041e-01 7.83205688e-01
5.63477039e-01 3.84410560e-01 -1.30582109e-01 -3.47772473e-03
7.86961973e-01 1.92982703e-01 -2.80486047e-01 -6.90641701e-01
-1.23694026e+00 6.69486940e-01 2.24993318e-01 -2.78111338e-03
-1.40170425e-01 2.59590149e-01 3.17529589e-01 1.69762865e-01
3.62512767e-01 -7.20455572e-02 -4.06391233e-01 -4.76722836e-01
-8.01009655e-01 2.67921060e-01 3.71581018e-01 1.22618246e+00
7.50955224e-01 -8.13946575e-02 1.65600434e-01 4.39122528e-01
1.41174704e-01 1.08997273e+00 -5.87563403e-02 -1.06556797e+00
8.12823296e-01 6.16293490e-01 3.22268158e-01 -9.69334066e-01
-2.13274449e-01 -2.51666695e-01 -5.07520378e-01 2.54232228e-01
8.71198237e-01 -1.33020142e-02 -5.58708370e-01 1.28228784e+00
5.22609413e-01 4.55071807e-01 -1.09646238e-01 1.07344007e+00
4.39389884e-01 5.36620438e-01 -2.15207905e-01 -1.93812981e-01
1.33841062e+00 -6.36465371e-01 -3.17128241e-01 -1.70222551e-01
4.37723190e-01 -7.56240606e-01 9.04570282e-01 4.67376769e-01
-7.86982596e-01 -6.77544236e-01 -9.52938974e-01 2.31392905e-01
-3.22089255e-01 3.02355379e-01 4.47711438e-01 1.06586266e+00
-7.37058997e-01 3.22128609e-02 -7.34206498e-01 -5.03237069e-01
4.20192003e-01 2.69787386e-02 -2.67384738e-01 -2.99853772e-01
-7.72016883e-01 8.42630684e-01 -8.73047411e-02 5.68841547e-02
-9.96510983e-01 -7.40438521e-01 -5.48111200e-01 -2.04976469e-01
8.20239902e-01 -5.94381928e-01 1.32048714e+00 -7.57038116e-01
-1.54044878e+00 8.42916131e-01 -2.09251434e-01 -3.27898264e-01
9.63373244e-01 -2.74253160e-01 -5.68343401e-01 4.55626428e-01
1.15296565e-01 4.46280330e-01 1.12102723e+00 -1.27863503e+00
-8.25156033e-01 -2.15683714e-01 7.35596474e-03 -3.00550312e-01
-1.77509755e-01 2.90067270e-02 -6.61216676e-01 -3.02192360e-01
-1.84337035e-01 -1.03430080e+00 -4.26981635e-02 2.81802386e-01
-2.98274279e-01 4.86536883e-02 1.17296898e+00 -4.35522139e-01
8.82883310e-01 -1.87200880e+00 -3.43393356e-01 -4.21203934e-02
1.44417420e-01 5.93731701e-01 3.68615836e-02 1.74147487e-01
1.04574978e-01 -1.92101672e-01 -8.23359787e-02 -2.70986050e-01
-2.07594320e-01 -6.75071869e-03 -2.36017287e-01 7.70269990e-01
2.78632104e-01 7.32153118e-01 -1.02918780e+00 -4.33545947e-01
7.68110096e-01 6.06891632e-01 -3.51822525e-01 2.41248667e-01
-1.06565185e-01 5.67004621e-01 -4.76085842e-01 8.04605186e-01
8.13021123e-01 -8.13466311e-02 -5.22739850e-02 -4.92231160e-01
-1.96472183e-01 -1.12542078e-01 -1.08930850e+00 1.24544168e+00
-2.75076330e-01 9.25512671e-01 -1.19900536e-02 -1.03545296e+00
6.81676209e-01 3.67726177e-01 7.04461277e-01 -7.19127476e-01
3.49713564e-01 1.84696823e-01 -3.67370993e-01 -7.21157551e-01
4.61794347e-01 8.36881548e-02 7.69852847e-02 1.17696829e-01
-2.18590483e-01 -4.32910055e-01 3.38250428e-01 1.23477779e-01
1.15482688e+00 1.05417166e-02 1.43634036e-01 2.01497734e-01
6.06198728e-01 1.90788984e-01 6.03932083e-01 4.67338562e-01
-4.89576131e-01 7.24952638e-01 4.20295745e-01 -4.91251618e-01
-1.46454155e+00 -1.20522273e+00 -1.69298574e-01 3.57971311e-01
2.58515984e-01 -2.28424489e-01 -8.69558096e-01 -7.81750083e-01
-6.36493042e-02 5.24659574e-01 -3.81917119e-01 5.83065674e-02
-6.85915887e-01 -4.01213944e-01 4.13129002e-01 4.93024409e-01
6.02697730e-01 -3.10189247e-01 -1.07050502e+00 9.73460525e-02
-2.16299519e-01 -1.93765914e+00 -5.60814917e-01 -5.07494628e-01
-7.82958746e-01 -1.52129471e+00 -6.07898474e-01 -8.84427801e-02
7.29671478e-01 9.67733026e-01 1.10117638e+00 1.58272177e-01
-3.72218966e-01 7.71406114e-01 -3.57672423e-01 -3.86423796e-01
-4.18259680e-01 -1.97029561e-01 2.10227326e-01 3.04143459e-01
4.74157512e-01 -3.87166053e-01 -7.22192049e-01 7.02835739e-01
-6.47108436e-01 2.67306883e-02 5.43285251e-01 4.48370159e-01
4.20551002e-01 -1.77067928e-02 2.81110972e-01 -4.79165912e-01
2.83486992e-02 -1.81198224e-01 -1.31213105e+00 5.30068949e-03
-5.50172448e-01 -4.17286903e-01 5.88936508e-01 -3.59690964e-01
-8.51559162e-01 2.49293447e-01 -2.09093087e-05 -7.80614972e-01
-5.28338194e-01 -3.19077522e-01 -1.47714570e-01 -8.45770687e-02
4.20613676e-01 1.74118116e-01 -1.47354200e-01 -1.16423585e-01
1.58854261e-01 5.15248060e-01 6.64803803e-01 -3.72698516e-01
1.15132368e+00 8.48184407e-01 8.86853188e-02 -1.05100060e+00
-7.19453216e-01 -7.36268878e-01 -5.89820921e-01 -7.18281746e-01
1.05475163e+00 -8.73789907e-01 -8.97951245e-01 4.74998415e-01
-1.32987630e+00 -1.60561036e-02 -1.44531086e-01 6.68900609e-01
-5.71184814e-01 4.83492613e-01 -1.78725928e-01 -9.23632205e-01
3.42177823e-02 -1.37376475e+00 1.26964998e+00 1.80343002e-01
2.59300292e-01 -7.47105122e-01 -3.40331160e-02 8.29914451e-01
3.90412688e-01 1.54144734e-01 3.96815985e-01 -8.54797885e-02
-1.10864282e+00 -5.88350415e-01 -3.86967272e-01 5.37968874e-01
-1.28264040e-01 2.07319170e-01 -1.13571334e+00 -1.86757207e-01
-1.60680667e-01 -2.77466737e-02 6.04279518e-01 5.13771713e-01
9.19104218e-01 8.65162238e-02 -3.70336592e-01 5.43779552e-01
1.45714366e+00 1.23928033e-01 6.62650943e-01 1.05411738e-01
7.46243775e-01 5.09043396e-01 6.61886156e-01 2.33504087e-01
4.89526063e-01 1.08966529e+00 5.80511868e-01 8.77880398e-03
-1.40008539e-01 -2.36399740e-01 4.17594641e-01 5.54872334e-01
-4.16816890e-01 -2.18825221e-01 -1.04820788e+00 4.44098413e-01
-1.79882872e+00 -1.23915446e+00 -4.88372684e-01 2.52966595e+00
2.07128376e-01 5.12588382e-01 4.32058811e-01 2.42949918e-01
5.84678471e-01 -1.22115679e-01 -4.47651297e-01 9.80500057e-02
4.40664999e-02 -3.12368453e-01 1.01838076e+00 2.15655193e-01
-1.01467979e+00 6.68480575e-01 5.49424791e+00 5.78235269e-01
-1.06083035e+00 3.79965991e-01 5.75331151e-01 -1.16013154e-01
1.11489497e-01 6.82330951e-02 -1.01566315e+00 6.02250636e-01
9.84986305e-01 2.72124708e-01 2.95350075e-01 8.24346542e-01
5.70368588e-01 -3.36950332e-01 -9.57437038e-01 1.32230163e+00
-4.31867726e-02 -1.24231827e+00 -3.28055173e-01 2.14197934e-01
5.54925561e-01 1.80461302e-01 -7.12872744e-02 -2.31425781e-02
3.62675898e-02 -5.23355305e-01 9.60136652e-01 5.64106166e-01
9.08601105e-01 -5.30255675e-01 7.13586211e-01 3.46832305e-01
-1.49731445e+00 -1.11837819e-01 -2.81421423e-01 2.17036575e-01
6.29754305e-01 7.55740583e-01 -7.63899982e-01 4.20407563e-01
6.96602821e-01 6.77687764e-01 -7.72398531e-01 1.26369536e+00
-1.51813343e-01 7.82643437e-01 -3.76627415e-01 1.32275909e-01
-1.83967818e-02 -3.83505791e-01 8.23133826e-01 1.20406485e+00
4.49798018e-01 -2.33116932e-02 2.64235884e-01 6.93984032e-01
6.84767514e-02 -3.85757118e-01 -7.76052535e-01 3.74200225e-01
4.36221510e-01 1.24352252e+00 -7.14182019e-01 -3.49152654e-01
-7.94571280e-01 6.31617606e-01 -6.51933253e-02 3.89538944e-01
-1.37769353e+00 -7.42275864e-02 8.07637036e-01 6.12221539e-01
4.89284128e-01 -4.61760312e-01 -5.69671169e-02 -1.30552542e+00
3.92920673e-01 -6.15224183e-01 -1.01409495e-01 -7.03774810e-01
-1.04092371e+00 4.61586624e-01 3.50058466e-01 -1.67594194e+00
-1.59730524e-01 -7.83898056e-01 -6.09531283e-01 2.54152387e-01
-1.64150691e+00 -1.17568564e+00 -6.11565173e-01 6.17248535e-01
6.50014341e-01 -1.57837778e-01 2.41372362e-01 5.92687964e-01
-6.61359847e-01 3.90206367e-01 4.18168902e-02 4.12345380e-02
5.62259316e-01 -7.61758506e-01 3.51348132e-01 9.78073239e-01
-7.63487145e-02 5.75102493e-02 7.75293589e-01 -4.34777617e-01
-1.65737307e+00 -1.05652797e+00 5.20989597e-01 -8.85820389e-01
6.30606174e-01 -5.60514033e-01 -5.60924768e-01 2.95353740e-01
-1.52326420e-01 3.21568519e-01 1.17877930e-01 -4.84784305e-01
-3.64632219e-01 -5.65658927e-01 -1.02938688e+00 2.39703417e-01
1.03525090e+00 -3.89628053e-01 -1.50282785e-01 3.89276087e-01
4.51023489e-01 -2.02162430e-01 -4.81571645e-01 1.83997095e-01
5.53632855e-01 -1.18311107e+00 1.14221275e+00 -2.32353583e-01
1.03098698e-01 -5.72247922e-01 -2.45082512e-01 -9.03348923e-01
1.64344311e-01 -4.99011874e-01 -1.18822217e-01 1.21039784e+00
1.27500236e-01 -7.40867019e-01 9.68043029e-01 5.35428762e-01
-1.03891939e-01 -5.31155109e-01 -1.22421277e+00 -9.81178939e-01
-4.25958514e-01 -9.64248955e-01 4.11649287e-01 6.15653753e-01
-5.03834903e-01 2.88223594e-01 -4.62325901e-01 2.04227731e-01
9.75483358e-01 -8.14796090e-02 1.27459192e+00 -1.36157668e+00
-4.98683006e-02 -3.63023520e-01 -7.57903814e-01 -1.23666549e+00
-4.67428640e-02 -2.17723265e-01 -7.36601576e-02 -1.03343427e+00
1.75891146e-01 -5.39103091e-01 4.07036096e-01 -1.23139150e-01
3.48468348e-02 2.97155976e-01 2.90560186e-01 1.21529363e-01
-6.80300176e-01 3.74006242e-01 1.01232600e+00 -1.64026409e-01
1.14205487e-01 2.22311750e-01 -2.75781780e-01 7.79572666e-01
7.98767447e-01 -5.05430043e-01 -4.94355351e-01 -4.61658150e-01
1.08684421e-01 7.43239447e-02 9.54813063e-01 -1.29364288e+00
3.76576662e-01 -2.99260467e-01 2.12064147e-01 -6.24245167e-01
5.31260073e-01 -1.35525763e+00 2.18175799e-01 3.10475379e-01
3.67899954e-01 2.53554434e-01 1.00792944e-02 7.68456519e-01
-1.34229874e-02 -5.33868112e-02 8.56159687e-01 -2.25894451e-02
-6.80388868e-01 3.83235723e-01 -4.44946170e-01 1.38595238e-01
1.30578887e+00 -6.22075677e-01 -4.60650533e-01 -4.44251686e-01
4.42382954e-02 -9.50277373e-02 7.83688784e-01 3.46317649e-01
6.67172551e-01 -1.29600227e+00 -6.66828454e-01 2.22752452e-01
2.36312047e-01 5.14577469e-03 9.94475782e-02 1.00054276e+00
-5.31043351e-01 7.27450669e-01 -3.11469436e-02 -1.08831716e+00
-1.16663575e+00 7.00847745e-01 3.59695405e-01 -2.09792294e-02
-6.27413809e-01 2.15095982e-01 -2.75425483e-02 -1.53741404e-01
4.59407717e-02 -4.79608357e-01 1.77333146e-01 -2.03842044e-01
5.34889221e-01 6.27749026e-01 3.04381579e-01 -9.76647973e-01
-4.22568709e-01 9.62261438e-01 1.95812091e-01 9.91404355e-02
1.17520452e+00 -2.73678809e-01 5.47094405e-01 1.11969359e-01
1.22586501e+00 -1.96581148e-02 -1.67520130e+00 -4.85098250e-02
-2.74838567e-01 -8.90636563e-01 1.03461184e-01 -1.56338692e-01
-1.33430707e+00 8.96576583e-01 9.17928755e-01 1.38735458e-01
1.10808170e+00 7.74199283e-03 9.46079016e-01 4.44890618e-01
5.33044934e-01 -1.04567051e+00 2.50103977e-02 1.51501253e-01
3.94800723e-01 -1.63296366e+00 6.23138882e-02 -6.64943218e-01
-4.68747199e-01 1.13594818e+00 5.91726959e-01 1.20731935e-01
3.72319072e-01 1.48025453e-01 -1.47384023e-02 -1.09763533e-01
-6.52969003e-01 -2.56549269e-01 1.13399766e-01 8.57868254e-01
3.00081968e-02 7.46563310e-03 2.37012058e-01 3.11539881e-02
1.76412776e-01 5.80334440e-02 6.59387767e-01 7.47374833e-01
-3.37954402e-01 -8.85860562e-01 -7.36580372e-01 3.27341974e-01
-1.45823404e-01 3.50497872e-01 3.93318161e-02 7.68582344e-01
1.70691863e-01 1.19859660e+00 8.03988129e-02 -3.79263401e-01
7.62625515e-01 -4.87156689e-01 3.89783442e-01 -1.22035481e-01
-1.63461700e-01 -1.96603373e-01 1.10666759e-01 -8.46531093e-01
-5.15143692e-01 -9.33811188e-01 -5.70642114e-01 -5.68395376e-01
-2.83199757e-01 -1.44850165e-01 8.58555317e-01 7.16701090e-01
1.54905722e-01 1.89342529e-01 5.65848529e-01 -1.14285839e+00
-4.82290030e-01 -5.60826242e-01 -1.46071777e-01 4.92265940e-01
5.43480098e-01 -1.00249660e+00 -3.22283298e-01 3.80417109e-01] | [8.058159828186035, -1.1637208461761475] |
a5f4d128-800a-47b5-826c-72d5b6f2f6cf | the-statistical-benefits-of-quantile-temporal | 2305.18388 | null | https://arxiv.org/abs/2305.18388v1 | https://arxiv.org/pdf/2305.18388v1.pdf | The Statistical Benefits of Quantile Temporal-Difference Learning for Value Estimation | We study the problem of temporal-difference-based policy evaluation in reinforcement learning. In particular, we analyse the use of a distributional reinforcement learning algorithm, quantile temporal-difference learning (QTD), for this task. We reach the surprising conclusion that even if a practitioner has no interest in the return distribution beyond the mean, QTD (which learns predictions about the full distribution of returns) may offer performance superior to approaches such as classical TD learning, which predict only the mean return, even in the tabular setting. | ['Will Dabney', 'Marc G. Bellemare', 'Rémi Munos', 'Clare Lyle', 'Yunhao Tang', 'Mark Rowland'] | 2023-05-28 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-3.70735347e-01 2.59972721e-01 -6.08623445e-01 -1.91799477e-01
-1.16668522e+00 -6.47305846e-01 6.56371117e-01 3.93941134e-01
-8.82676780e-01 1.45031679e+00 2.24406257e-01 -9.62021470e-01
-5.20121753e-01 -8.26198936e-01 -6.79285586e-01 -7.63938904e-01
-4.94751185e-01 7.07785845e-01 9.71794128e-04 -1.67815834e-01
5.83175063e-01 3.47600907e-01 -1.38123107e+00 1.65459160e-02
9.60749805e-01 1.26743364e+00 -2.15905204e-01 6.29342377e-01
-1.76279381e-01 9.25162315e-01 -8.98565054e-01 -3.90142649e-01
4.24714625e-01 -3.69943678e-01 -5.56346595e-01 -5.12137949e-01
-5.49088866e-02 -8.95689785e-01 -1.89361915e-01 1.02371156e+00
5.14672935e-01 3.53696436e-01 1.14902484e+00 -1.14852166e+00
-1.21660799e-01 6.41670942e-01 -5.39655805e-01 3.44035298e-01
3.09521586e-01 4.30379689e-01 1.08084166e+00 -8.28879476e-02
1.18908867e-01 1.40514159e+00 4.32132661e-01 4.62165773e-01
-1.41846073e+00 -3.36726427e-01 1.28527582e-01 -9.26240459e-02
-5.26614308e-01 -1.87320802e-02 2.70244747e-01 -5.58160067e-01
9.22261238e-01 -2.12071404e-01 5.99422932e-01 9.30953205e-01
7.73432910e-01 9.00060236e-01 1.66216493e+00 -3.52262855e-01
6.22853994e-01 -7.31764408e-03 -5.54370999e-01 7.62429237e-02
1.83604136e-01 1.08612013e+00 -1.65170640e-01 -4.82416123e-01
6.35630310e-01 -6.54667616e-02 2.10390031e-01 -4.20832574e-01
-8.42314422e-01 1.18570244e+00 8.53350945e-03 -1.91932932e-01
-8.87203157e-01 3.33962560e-01 7.94478059e-01 8.63330543e-01
7.49317884e-01 5.01916409e-01 -7.69421756e-01 -5.14685512e-01
-8.20855021e-01 9.32684720e-01 1.00394130e+00 4.55767661e-01
4.89149034e-01 4.52180743e-01 -6.95731521e-01 3.39229733e-01
8.42694864e-02 6.88965976e-01 3.95914495e-01 -1.29694414e+00
5.23420453e-01 -7.14253187e-02 8.39139998e-01 -3.64786945e-02
-1.77420527e-01 -2.59517014e-01 -1.94994688e-01 9.02844071e-01
7.91444480e-01 -1.00461638e+00 -4.87776995e-01 1.60035276e+00
2.16952920e-01 -7.95267224e-02 3.64114225e-01 4.26038712e-01
-2.15735883e-01 5.03600776e-01 2.16041073e-01 -6.45125628e-01
4.57684636e-01 -2.34646782e-01 -3.54249984e-01 3.66068274e-01
7.26822615e-01 -4.89673048e-01 8.62189233e-01 6.50901377e-01
-9.87637937e-01 -3.86011302e-01 -6.60069466e-01 8.77554119e-01
-3.33533250e-02 -5.25465071e-01 4.33868408e-01 6.83420300e-01
-8.84137213e-01 1.25569642e+00 -5.01494586e-01 1.78203046e-01
3.55363101e-01 2.42698163e-01 5.26686847e-01 2.04843059e-01
-1.27318180e+00 1.02110231e+00 6.43492997e-01 -4.26374286e-01
-1.22228241e+00 -7.23193824e-01 -3.94329429e-01 -1.08438665e-02
7.66332924e-01 -3.28587085e-01 1.94391012e+00 -1.10510969e+00
-1.73013425e+00 2.43408352e-01 3.88266146e-01 -1.07425761e+00
9.55880642e-01 -2.60360181e-01 -1.82604775e-01 -1.92809120e-01
-1.41790032e-01 1.96592614e-01 1.11294281e+00 -8.21449339e-01
-1.10768759e+00 -3.51725698e-01 2.18018338e-01 2.13236257e-01
4.22496647e-01 -3.48850638e-01 9.34144139e-01 -4.98031318e-01
-7.83347607e-01 -6.29214525e-01 -4.75554705e-01 -5.84203243e-01
-4.73389439e-02 -9.19413507e-01 3.43926877e-01 -4.31794792e-01
9.98181820e-01 -1.88326108e+00 -4.73471016e-01 3.99141520e-01
-1.88785121e-01 1.59417205e-02 4.74862531e-02 7.50017107e-01
-1.02325901e-01 -1.38100475e-01 1.21326521e-01 3.78542572e-01
5.54344773e-01 3.32900763e-01 -7.89723516e-01 4.70902532e-01
1.48989007e-01 6.99818730e-01 -9.56471145e-01 -2.15146706e-01
-1.81367658e-02 -3.53001446e-01 -5.48357844e-01 3.23409617e-01
-9.04267311e-01 2.74156839e-01 -7.26181090e-01 2.62578309e-01
4.59331989e-01 3.01387787e-01 2.03053206e-01 6.18559122e-01
-4.88104552e-01 4.31508809e-01 -9.93834853e-01 1.00890696e+00
-2.81242609e-01 3.07411492e-01 -3.51811528e-01 -1.43926537e+00
8.92113328e-01 2.43098140e-01 7.12747812e-01 -1.06203067e+00
-6.53555244e-02 4.27674085e-01 2.11742476e-01 -3.50816309e-01
3.76884669e-01 -8.54731739e-01 -1.84671834e-01 9.25828099e-01
-1.34153128e-01 -1.03362702e-01 7.50395581e-02 -1.05000287e-01
9.52292264e-01 6.91274703e-01 2.01932982e-01 -5.44432640e-01
-1.29607812e-01 1.00532100e-01 4.17454988e-01 9.14825857e-01
-3.12654734e-01 -3.84803981e-01 1.40031099e+00 -3.02658051e-01
-1.23023629e+00 -1.44990766e+00 -1.92413926e-01 1.20344341e+00
-3.54667366e-01 1.84045300e-01 -4.85357404e-01 -1.04007220e+00
7.71547496e-01 1.19582903e+00 -6.49113476e-01 -3.12486198e-02
-2.20138296e-01 -5.73388517e-01 1.37404203e-01 4.11320537e-01
1.03732057e-01 -1.04134631e+00 -9.66466963e-01 4.73397523e-01
4.56302643e-01 -1.55458257e-01 -3.36599588e-01 5.67744553e-01
-1.10279059e+00 -1.06028795e+00 -8.70428681e-01 2.45985538e-02
-1.21349767e-01 -5.91804981e-01 1.15603209e+00 -6.19427085e-01
1.62371144e-01 8.77163231e-01 -1.90086365e-01 -8.96048903e-01
-3.56725156e-01 -2.08505288e-01 2.86364425e-02 -4.00547653e-01
4.41523761e-01 -3.17426354e-01 -7.96739101e-01 -5.69271781e-02
-7.37249494e-01 -9.83749092e-01 4.30119991e-01 1.08609796e+00
3.27060699e-01 2.47566432e-01 1.23444271e+00 -1.07209647e+00
1.33896351e+00 -5.63006163e-01 -1.02565372e+00 3.54357392e-01
-1.11562037e+00 6.02728307e-01 6.44974470e-01 -4.27504092e-01
-1.20259690e+00 -5.57134032e-01 1.12090468e-01 -1.08329371e-01
-1.07759029e-01 6.36925578e-01 4.88316387e-01 5.56251943e-01
7.60158777e-01 3.06101382e-01 3.82559031e-01 -3.66114289e-01
7.60547817e-02 4.70582217e-01 3.02168399e-01 -1.13791430e+00
3.75296831e-01 -2.35512018e-01 1.90034181e-01 -3.59557986e-01
-7.01793909e-01 -4.53123040e-02 -2.55051911e-01 -2.68929422e-01
4.28128958e-01 -7.45795786e-01 -1.19668949e+00 4.13558260e-02
-6.71017706e-01 -9.51051414e-01 -9.32535827e-01 7.39850104e-01
-1.49467480e+00 -1.44554079e-01 -3.02124023e-01 -1.50116920e+00
3.42592709e-02 -6.13606334e-01 5.97594261e-01 2.16932788e-01
8.03105533e-02 -1.29732645e+00 5.27678788e-01 -5.55287182e-01
2.24644139e-01 6.29501760e-01 1.24674296e+00 -9.97572243e-01
-2.19231114e-01 1.54791772e-01 1.50424406e-01 4.24794436e-01
-2.60277428e-02 -7.65257850e-02 -5.64446628e-01 -4.86311436e-01
1.35803223e-03 -8.48578274e-01 7.11564243e-01 8.69352877e-01
1.03061461e+00 -4.78761822e-01 1.66602045e-01 7.79386284e-03
1.52903378e+00 7.84743905e-01 4.86923784e-01 5.82842648e-01
-2.63599873e-01 8.43001723e-01 1.21884131e+00 1.22194874e+00
-1.31733017e-03 2.28404716e-01 4.44037110e-01 3.53521317e-01
6.66673064e-01 -6.74094558e-01 4.94160414e-01 -2.58128136e-01
-7.37172067e-02 4.45713103e-02 -7.57302165e-01 2.99110234e-01
-1.83359277e+00 -1.40490460e+00 5.29081762e-01 2.79925418e+00
9.45332825e-01 5.37648916e-01 1.04744256e+00 -1.86812133e-01
3.89636427e-01 8.05673450e-02 -1.22759867e+00 -1.00219512e+00
3.39910120e-01 4.70002651e-01 8.51602137e-01 3.99235189e-01
-6.99582458e-01 5.04140735e-01 7.71327400e+00 1.14505160e+00
-9.08291638e-01 -3.35470229e-01 9.35654521e-01 2.16037361e-03
-4.57479447e-01 -5.17391637e-02 -6.98393285e-01 5.09753406e-01
1.46467793e+00 -6.00916684e-01 4.98333722e-01 9.53909993e-01
4.05712008e-01 -5.86010635e-01 -1.17010880e+00 5.75514793e-01
-6.94882393e-01 -8.61031651e-01 -3.31603795e-01 2.63025165e-01
8.80533040e-01 -1.14735350e-01 3.91716063e-01 9.41580594e-01
9.97760832e-01 -1.13800633e+00 6.40178800e-01 7.95875669e-01
7.14181483e-01 -1.59792900e+00 8.07780981e-01 7.88259983e-01
-5.20804107e-01 -3.66320223e-01 -5.79228222e-01 -2.23480731e-01
-2.79610127e-01 3.54046434e-01 -8.95235956e-01 4.04742479e-01
4.54097897e-01 4.98709828e-01 4.09658775e-02 1.07102656e+00
4.45898958e-02 8.36276293e-01 -1.19473323e-01 -3.29479009e-01
7.15983212e-01 -4.28464264e-01 1.85145259e-01 7.64049172e-01
4.11305726e-01 -2.89368853e-02 3.43171954e-01 5.47039866e-01
3.57682645e-01 -7.35748410e-02 -7.86722064e-01 -2.12553129e-01
1.56613007e-01 3.78756821e-01 -2.25111619e-01 -3.50515008e-01
-2.62588263e-01 2.90849864e-01 2.44502321e-01 5.14534295e-01
-5.84347248e-01 -4.23637092e-01 6.50226533e-01 6.10027723e-02
6.19436562e-01 2.51921117e-02 -1.53553151e-02 -4.89053756e-01
-2.30138898e-01 -8.93627942e-01 7.05343187e-01 -1.75369754e-01
-1.52485228e+00 1.63453934e-03 3.79184961e-01 -1.26546037e+00
-1.13372421e+00 -8.41686904e-01 -6.31487727e-01 1.01525223e+00
-1.58811581e+00 -5.54509424e-02 1.06968057e+00 6.32339895e-01
3.38835686e-01 -2.44173512e-01 4.82392430e-01 -5.07136643e-01
-2.72361822e-02 3.73270452e-01 1.12139034e+00 -1.60603613e-01
6.71277344e-01 -1.94387186e+00 -1.68856159e-01 -8.78036097e-02
-2.85637587e-01 7.44355544e-02 1.10459220e+00 -6.21938527e-01
-1.16777992e+00 -8.35470438e-01 2.99758732e-01 -2.65227735e-01
8.23403537e-01 2.43999660e-01 -5.80644250e-01 5.54295003e-01
3.24237347e-01 -3.52954507e-01 4.04784113e-01 2.16496751e-01
-1.33279696e-01 -2.27192387e-01 -1.24214041e+00 3.31292778e-01
3.19781542e-01 -4.97844905e-01 -7.55342066e-01 1.58263907e-01
5.10279298e-01 -2.23031715e-01 -1.03605270e+00 1.74327865e-01
6.47727251e-01 -1.15929985e+00 6.44229531e-01 -1.19729507e+00
4.44655567e-01 3.71748209e-01 -9.86051708e-02 -1.92851794e+00
8.67505521e-02 -8.83306324e-01 -7.98471719e-02 8.16571057e-01
3.25312763e-01 -1.04590547e+00 8.32870781e-01 4.97310787e-01
2.50607967e-01 -9.94106948e-01 -1.19721067e+00 -1.27539349e+00
8.89724612e-01 -3.73387039e-01 6.28983021e-01 5.13727546e-01
1.83518857e-01 -1.17750037e-02 -3.87956977e-01 -4.35112745e-01
1.02003491e+00 4.94641274e-01 5.29153585e-01 -1.14532614e+00
-5.60108244e-01 -5.77534080e-01 2.89971251e-02 -1.05737638e+00
4.16229635e-01 -3.90328974e-01 5.46664298e-02 -1.09370708e+00
-5.94056658e-02 -2.94592649e-01 -6.69803917e-01 5.39821200e-02
1.54575348e-01 -7.25569427e-01 -3.49737816e-02 -2.44614184e-01
-5.95290542e-01 9.09666955e-01 1.51572204e+00 5.23258140e-03
-2.14661583e-01 7.12100565e-01 -5.87305188e-01 4.40443397e-01
9.59494114e-01 -7.81020105e-01 -5.48817873e-01 3.96593988e-01
1.71176448e-01 9.80119944e-01 1.44315392e-01 -4.39521283e-01
-1.65100425e-01 -1.04999375e+00 5.78181446e-01 -7.74366796e-01
-2.58144170e-01 -5.49701095e-01 -2.65951008e-01 9.35684204e-01
-8.27574670e-01 2.33128533e-01 2.66380191e-01 8.62122893e-01
-1.83000416e-01 -4.95864987e-01 5.37530720e-01 -2.51947790e-01
-3.88720453e-01 3.90015572e-01 -5.54244220e-01 6.34969473e-01
9.37879324e-01 2.99466532e-02 -3.07418942e-01 -6.95829332e-01
-5.31595409e-01 3.92653406e-01 1.00592971e-01 -2.24297419e-01
5.21269917e-01 -1.30202436e+00 -8.57929111e-01 -3.19918782e-01
-1.48724273e-01 -3.81026506e-01 7.58016482e-02 6.42009258e-01
5.47349043e-02 5.67221522e-01 -2.13721380e-01 -1.79643273e-01
-3.31333876e-01 6.75357163e-01 5.30650079e-01 -5.88743627e-01
-3.20407659e-01 2.33487502e-01 -2.75958981e-02 -3.66194963e-01
1.98233247e-01 -1.70860752e-01 -7.65801221e-02 3.84002537e-01
6.11337543e-01 5.78406632e-01 -2.67983049e-01 1.66747898e-01
2.14870632e-01 1.55828986e-02 9.03083757e-02 -5.31739593e-01
1.40318036e+00 1.71851903e-01 4.13079679e-01 8.13301682e-01
7.39926457e-01 -4.05891210e-01 -2.10645103e+00 -3.19527507e-01
5.63869715e-01 -7.63036489e-01 2.91634388e-02 -1.04674029e+00
-5.03149092e-01 8.75423491e-01 6.86637759e-01 6.11024380e-01
8.68402302e-01 -3.35396230e-01 2.48741582e-01 5.10003448e-01
5.56901276e-01 -1.59428811e+00 1.15768254e-01 5.89595497e-01
6.65332854e-01 -1.17451692e+00 8.12343583e-02 9.27631319e-01
-9.71179128e-01 1.20747995e+00 4.05239761e-01 -6.31561697e-01
6.84116602e-01 7.88358450e-02 -1.81409866e-01 2.69407630e-01
-1.41286170e+00 -4.37355667e-01 3.76929715e-02 9.30718243e-01
3.93142313e-01 3.82988274e-01 -3.29936087e-01 7.99183249e-02
-1.58086270e-01 -4.21745293e-02 3.88734818e-01 9.05017972e-01
-8.05979371e-01 -1.28212380e+00 -3.70575935e-01 9.88947451e-01
-7.03878164e-01 1.33504286e-01 5.45542408e-03 1.03534400e+00
-3.29845607e-01 7.94437408e-01 3.15033942e-01 3.78377177e-02
4.63533282e-01 2.19161481e-01 6.50832057e-01 -3.93739343e-01
-4.04080272e-01 4.16530758e-01 4.85952571e-02 -4.94101822e-01
-2.85280019e-01 -1.01694822e+00 -1.06361866e+00 -5.93771517e-01
2.52668500e-01 5.82632124e-01 3.53572488e-01 1.04275298e+00
-3.00460100e-01 2.34251082e-01 1.14556527e+00 -4.37650502e-01
-1.95387900e+00 -9.22153115e-01 -1.35712361e+00 5.97539470e-02
6.94808662e-01 -7.48942077e-01 -4.63997930e-01 -6.15127563e-01] | [4.187106132507324, 2.6112005710601807] |
a2b67e79-3418-4579-a56a-dcdb673f34f2 | posegu-3d-human-pose-estimation-with-novel | 2207.03618 | null | https://arxiv.org/abs/2207.03618v1 | https://arxiv.org/pdf/2207.03618v1.pdf | PoseGU: 3D Human Pose Estimation with Novel Human Pose Generator and Unbiased Learning | 3D pose estimation has recently gained substantial interests in computer vision domain. Existing 3D pose estimation methods have a strong reliance on large size well-annotated 3D pose datasets, and they suffer poor model generalization on unseen poses due to limited diversity of 3D poses in training sets. In this work, we propose PoseGU, a novel human pose generator that generates diverse poses with access only to a small size of seed samples, while equipping the Counterfactual Risk Minimization to pursue an unbiased evaluation objective. Extensive experiments demonstrate PoseGU outforms almost all the state-of-the-art 3D human pose methods under consideration over three popular benchmark datasets. Empirical analysis also proves PoseGU generates 3D poses with improved data diversity and better generalization ability. | ['Gengfa Fang', 'Linchao Zhu', 'Haiyan Lu', 'Shannan Guan'] | 2022-07-07 | null | null | null | null | ['3d-pose-estimation'] | ['computer-vision'] | [ 5.08320518e-02 2.76167989e-01 -2.90224761e-01 -2.64260709e-01
-1.03083110e+00 -7.15403318e-01 6.39997065e-01 -4.10363942e-01
-3.81484747e-01 1.04766536e+00 2.57446647e-01 2.14203432e-01
4.63268906e-02 -3.88580918e-01 -8.73896360e-01 -4.63354379e-01
-2.18836322e-01 9.21807408e-01 -2.38791792e-04 -8.94988477e-02
-1.16465114e-01 3.79289925e-01 -1.36151862e+00 -4.71317708e-01
7.44351625e-01 8.74044001e-01 -2.99494892e-01 3.98175418e-01
7.31350839e-01 -2.32240811e-01 -7.22891927e-01 -6.20607555e-01
6.39634848e-01 -3.23214710e-01 -3.01301241e-01 1.79157123e-01
7.10622728e-01 -4.64171618e-01 -2.48232991e-01 8.63325834e-01
1.10036838e+00 2.44514391e-01 9.04190540e-01 -1.38669229e+00
-2.39389330e-01 6.20965026e-02 -6.41578972e-01 5.44866174e-02
8.32062304e-01 2.60378361e-01 8.38509083e-01 -9.03456628e-01
9.04524088e-01 1.40043795e+00 7.21890271e-01 8.14046741e-01
-1.31759608e+00 -8.34725797e-01 3.09819579e-02 -2.82900393e-01
-1.44266844e+00 4.06776480e-02 8.44871163e-01 -3.18654865e-01
5.87041259e-01 3.55922967e-01 9.63246644e-01 1.91468716e+00
2.91788936e-01 1.06394446e+00 1.18035555e+00 -9.69278067e-02
2.92465419e-01 -1.04833491e-01 -3.87009233e-01 5.25427699e-01
6.74956441e-01 4.73462582e-01 -7.00342894e-01 -3.31543922e-01
1.05884719e+00 -2.97517091e-01 -1.75394163e-01 -9.81502831e-01
-1.17506361e+00 7.57707953e-01 5.00110030e-01 -5.12486815e-01
-4.18230921e-01 1.53009221e-01 2.87493140e-01 -2.27248743e-01
6.08774185e-01 5.64471781e-01 -5.37632883e-01 -5.50476275e-02
-4.59744483e-01 9.80839849e-01 6.68898940e-01 1.23836327e+00
2.20204517e-01 -1.64595515e-01 -3.26648027e-01 5.09397507e-01
1.60249516e-01 8.27340901e-01 4.11966473e-01 -6.00757539e-01
7.25913286e-01 5.06204247e-01 2.71631718e-01 -1.05210364e+00
-4.50983644e-01 -9.07030880e-01 -5.72887301e-01 -8.97539705e-02
3.64574075e-01 -2.34088868e-01 -1.06144178e+00 1.96706474e+00
7.76247561e-01 8.02344307e-02 -2.22705439e-01 1.07898688e+00
5.97951829e-01 1.80953413e-01 -7.72909150e-02 5.04164994e-02
9.55768168e-01 -4.95271921e-01 -2.76433408e-01 -4.53750700e-01
2.00817704e-01 -5.52262664e-01 9.78207052e-01 3.75982255e-01
-9.24038172e-01 -4.70162690e-01 -1.02708495e+00 2.27278218e-01
3.57925519e-02 9.33007598e-02 7.61821747e-01 9.88253832e-01
-3.65773231e-01 3.86233896e-01 -6.52325094e-01 -2.96222717e-01
6.57348633e-01 4.51995939e-01 -6.06693625e-01 -1.89249720e-02
-1.03779972e+00 1.07784998e+00 5.38540959e-01 1.59806505e-01
-1.07777071e+00 -6.01737916e-01 -9.77994442e-01 -7.70782053e-01
7.90231824e-01 -9.94509459e-01 1.13678527e+00 -1.33872882e-01
-1.43854284e+00 1.06042910e+00 2.79049009e-01 -4.18167770e-01
1.27986336e+00 -8.50980997e-01 1.22089550e-01 -1.26074329e-01
2.73238808e-01 9.82167780e-01 8.55627418e-01 -1.50769842e+00
-2.09428430e-01 -6.76149607e-01 -1.05459802e-01 6.04410529e-01
-5.13736308e-02 -6.34510994e-01 -6.53516471e-01 -8.43605518e-01
3.60812336e-01 -1.22428882e+00 -5.27530670e-01 -1.36315912e-01
-9.13368404e-01 -5.00599369e-02 3.62526774e-01 -3.40688258e-01
9.77142513e-01 -1.48619688e+00 5.76212764e-01 2.16325000e-01
9.49294716e-02 9.09750909e-02 1.14693763e-02 2.58720160e-01
1.13141999e-01 5.22921309e-02 -3.35524827e-02 -3.33111495e-01
1.99076176e-01 8.05502236e-02 -1.98678881e-01 6.80309355e-01
2.42420375e-01 1.08555603e+00 -1.02460802e+00 -6.38526559e-01
2.83336788e-01 2.56185114e-01 -9.25226331e-01 2.48731673e-01
-2.01311588e-01 6.24982119e-01 -8.00266683e-01 8.01351905e-01
6.30939901e-01 -1.56384617e-01 7.68563077e-02 -1.80013567e-01
6.45688415e-01 -1.60569906e-01 -1.13965058e+00 2.01629734e+00
6.87761465e-03 -5.55645376e-02 -6.53253078e-01 -4.76415604e-01
8.59066606e-01 1.62312359e-01 5.07857680e-01 -8.84807557e-02
6.15761459e-01 2.73990542e-01 -7.91309476e-02 -2.89270967e-01
2.83962965e-01 -3.00633848e-01 -6.33844137e-01 2.10598066e-01
1.30350977e-01 -6.00517571e-01 -1.01276249e-01 -1.12301573e-01
8.40222836e-01 7.37390757e-01 6.03979349e-01 -2.19419338e-02
2.26050839e-01 -1.24182455e-01 6.09945476e-01 7.02663422e-01
-5.07547200e-01 9.52494323e-01 2.08118245e-01 -3.16163927e-01
-1.14994085e+00 -1.71994662e+00 -1.35537922e-01 4.27592307e-01
3.55937570e-01 -2.57304847e-01 -6.64913595e-01 -1.09790981e+00
3.42613429e-01 4.17867839e-01 -9.29508567e-01 -3.00352186e-01
-6.32370949e-01 -8.32785130e-01 6.81226313e-01 5.19392133e-01
2.90446550e-01 -8.41341078e-01 -7.94903100e-01 -5.06494530e-02
-2.08533064e-01 -1.06491792e+00 -4.71304893e-01 -1.16039753e-01
-8.54395568e-01 -1.25083220e+00 -1.11423790e+00 -5.14395833e-01
7.94439912e-01 -1.80966482e-01 1.20355606e+00 -5.04968405e-01
-4.88664806e-01 2.48753622e-01 -4.19838697e-01 -8.21816981e-01
-9.51414332e-02 2.50237495e-01 5.54576874e-01 -3.41459990e-01
4.40540850e-01 -3.83757919e-01 -6.00741506e-01 5.57608485e-01
-3.76224279e-01 -1.93125736e-02 6.63618624e-01 1.00304449e+00
7.02802598e-01 -1.87011257e-01 6.46500230e-01 -7.00857103e-01
5.43175876e-01 -3.02031994e-01 -3.27178895e-01 -1.58303585e-02
-1.72245264e-01 -6.73545003e-02 1.01469703e-01 -6.66024148e-01
-9.19485867e-01 3.10116351e-01 7.43780732e-02 -5.51768720e-01
-1.25437349e-01 6.56331703e-02 -4.73468006e-01 1.18393742e-01
1.02896023e+00 8.25838223e-02 2.78808355e-01 -3.86540174e-01
4.42569703e-01 1.71366915e-01 5.26507735e-01 -9.31989670e-01
1.31535757e+00 3.97281170e-01 2.65225828e-01 -4.53752697e-01
-9.73860025e-01 -3.23104054e-01 -8.63857865e-01 -4.35644001e-01
7.56793916e-01 -1.05188620e+00 -5.87934196e-01 3.81741345e-01
-8.06872785e-01 1.45090818e-01 -2.01810539e-01 6.56170368e-01
-9.78469074e-01 2.00342581e-01 -3.58510278e-02 -9.41092670e-01
-1.99689746e-01 -1.11940265e+00 1.55593956e+00 -4.21925709e-02
-7.07484543e-01 -4.93111551e-01 -8.50709714e-03 5.36383271e-01
-2.79764205e-01 1.18387341e+00 3.45335424e-01 -5.25953531e-01
-5.87024152e-01 -8.76473844e-01 2.82769084e-01 2.38342330e-01
1.09858818e-01 -6.64067507e-01 -8.29231024e-01 -5.16620338e-01
-2.99693823e-01 -8.10775340e-01 4.90325570e-01 5.52802563e-01
9.52225745e-01 -9.26173180e-02 -5.33886552e-01 4.52161461e-01
1.05159187e+00 -1.81920767e-01 2.55115747e-01 1.10755660e-01
6.42881334e-01 6.29724860e-01 1.01731777e+00 5.88971496e-01
4.20739204e-02 7.55229533e-01 5.35424531e-01 2.73481607e-01
2.01119274e-01 -8.77997518e-01 -7.00863674e-02 2.16141894e-01
-1.88466534e-01 -3.17197233e-01 -7.94253886e-01 3.95693541e-01
-1.75280285e+00 -6.22459650e-01 4.07326937e-01 2.42160487e+00
9.08891678e-01 4.56444234e-01 4.23980951e-01 1.77798718e-01
6.48108006e-01 8.74756575e-02 -8.17544699e-01 5.16787946e-01
4.57300059e-02 2.35248983e-01 6.24868333e-01 3.24364640e-02
-1.26474535e+00 8.68580461e-01 6.22336769e+00 8.75637054e-01
-5.39075196e-01 -2.11763665e-01 5.32725632e-01 -5.73637426e-01
-1.24546185e-01 -3.81802589e-01 -9.27266359e-01 2.64090091e-01
1.48312807e-01 -2.62996465e-01 -7.05938712e-02 1.12568855e+00
-1.16891172e-02 -9.42810252e-02 -1.45725799e+00 1.31793213e+00
2.21296728e-01 -9.14328396e-01 3.25214565e-01 3.45018417e-01
8.91578615e-01 -4.97993022e-01 2.47000054e-01 2.67381191e-01
2.02254102e-01 -1.25720870e+00 9.69709396e-01 2.60717392e-01
8.74455452e-01 -9.91246760e-01 7.90773451e-01 5.93096673e-01
-8.52325141e-01 9.64167714e-02 -1.47910520e-01 -1.05009787e-01
4.25077051e-01 2.91274130e-01 -1.10826182e+00 6.75569296e-01
5.79314411e-01 4.96589512e-01 -5.35501659e-01 1.12656462e+00
-3.83495271e-01 2.39947855e-01 -5.75311720e-01 -3.94415170e-01
4.82446998e-02 2.21239522e-01 9.56868589e-01 4.92817312e-01
2.23868281e-01 -4.03343104e-02 5.76529503e-01 6.35565758e-01
2.29637548e-02 1.64791495e-02 -7.78279781e-01 1.85323179e-01
7.10692227e-01 7.35251009e-01 -5.61153710e-01 5.89635000e-02
2.89497226e-01 9.31453526e-01 1.11701719e-01 8.61108974e-02
-9.09236908e-01 6.94886521e-02 6.44179225e-01 2.38730937e-01
-7.72504434e-02 -1.93145081e-01 -4.53618169e-01 -1.06281948e+00
2.45829016e-01 -9.67279494e-01 3.23073536e-01 -5.35557270e-01
-1.44634593e+00 3.31024706e-01 5.92324674e-01 -1.62648511e+00
-5.79268277e-01 -6.50180340e-01 -2.91995984e-03 7.42973566e-01
-8.16345394e-01 -1.29780090e+00 -2.04551637e-01 2.87624806e-01
5.51859617e-01 -1.65123671e-01 6.29608989e-01 -3.07067893e-02
-3.97160709e-01 9.45195436e-01 -6.79631770e-01 7.37047270e-02
6.86855316e-01 -1.18179607e+00 6.37847722e-01 5.50470769e-01
-9.45905000e-02 6.99679434e-01 1.04874873e+00 -1.02542937e+00
-1.40640903e+00 -9.69311059e-01 5.01956999e-01 -1.12645650e+00
1.40450194e-01 -6.03456795e-01 -1.64988950e-01 6.10975981e-01
-4.64955747e-01 -4.92043756e-02 5.25490582e-01 1.05597399e-01
-1.26373634e-01 3.24060887e-01 -1.35686302e+00 8.77797365e-01
1.61559391e+00 -5.64596206e-02 -8.44194412e-01 2.55056530e-01
5.65931439e-01 -9.59295809e-01 -7.79977381e-01 8.52230489e-01
8.45916152e-01 -6.70750201e-01 1.32627344e+00 -8.48320663e-01
3.12609643e-01 -1.55460924e-01 -3.25819403e-01 -1.29186046e+00
-1.29604833e-02 -6.99159622e-01 -1.73363298e-01 7.23499894e-01
3.51880282e-01 -3.31423521e-01 1.48008680e+00 4.10643339e-01
1.88116342e-01 -1.03625083e+00 -1.09874666e+00 -1.08177054e+00
4.99450266e-02 -4.72983360e-01 6.38201237e-01 3.84006083e-01
-4.35440004e-01 4.42507863e-01 -7.84691453e-01 1.72013268e-01
1.14354217e+00 -1.52837977e-01 1.48108244e+00 -1.31635845e+00
-3.80186945e-01 -1.22490786e-01 -7.93697298e-01 -1.19928050e+00
1.38855398e-01 -7.03227460e-01 1.82041422e-01 -1.13080907e+00
3.41601372e-01 -2.50527114e-01 1.30815074e-01 1.02198929e-01
-4.13561285e-01 4.84340996e-01 1.19515203e-01 -1.01980619e-01
-3.75213385e-01 9.63937402e-01 1.78324580e+00 1.31326213e-01
-1.72019973e-01 3.01224828e-01 -4.26646411e-01 8.01256776e-01
5.73942125e-01 -2.54521370e-01 -8.61386538e-01 2.97991987e-02
8.83785784e-02 -6.03455827e-02 6.47632897e-01 -1.06054592e+00
-4.39869612e-01 -1.59370765e-01 9.18276191e-01 -9.74531591e-01
7.80404150e-01 -6.25215709e-01 2.36668676e-01 5.55246413e-01
-1.27425835e-01 -7.40593225e-02 1.06387615e-01 1.05519116e+00
1.76514432e-01 2.27141038e-01 7.00458229e-01 -4.79668707e-01
-5.33620656e-01 6.29444718e-01 3.93568784e-01 5.50734103e-01
1.28442776e+00 -5.16613722e-01 2.23067164e-01 -3.04341167e-01
-6.84509873e-01 1.39194667e-01 4.66633976e-01 7.55354583e-01
6.52395785e-01 -1.76066697e+00 -6.83012128e-01 1.80494189e-01
3.63588333e-01 3.69400680e-01 5.65992668e-02 3.28375816e-01
-3.67376000e-01 3.98440689e-01 -2.25963265e-01 -7.94227660e-01
-1.07598174e+00 2.63020545e-01 2.14955822e-01 -2.61805326e-01
-5.31866312e-01 1.22130346e+00 1.76659390e-01 -8.03769767e-01
3.18186849e-01 -1.25371516e-01 2.32880488e-01 -2.99515314e-02
4.53436933e-02 4.64645892e-01 -2.36038014e-01 -7.85921097e-01
-5.96351683e-01 5.64238429e-01 4.82298173e-02 -2.14221209e-01
1.06896448e+00 1.29479155e-01 7.56475806e-01 1.72772780e-01
1.17919338e+00 -1.73717946e-01 -1.56009662e+00 -4.52219769e-02
-2.39147529e-01 -9.37365174e-01 -4.04868275e-01 -8.21265697e-01
-5.71687996e-01 3.63921314e-01 4.98154938e-01 -4.40633386e-01
6.97847545e-01 7.61654302e-02 9.02244329e-01 2.76039511e-01
1.10281754e+00 -1.14636409e+00 5.48491657e-01 1.89809918e-01
1.15963471e+00 -1.42904377e+00 4.24078405e-01 -7.01116443e-01
-6.91775143e-01 4.79073912e-01 9.20995176e-01 -4.00775522e-01
5.67277372e-01 -1.58370018e-01 5.84037378e-02 -9.97977257e-02
-3.30426216e-01 -9.29093733e-02 8.15249443e-01 8.61634612e-01
1.61236197e-01 2.84948379e-01 -4.25992042e-01 7.31817424e-01
-7.06544459e-01 -2.10116571e-03 3.90422344e-02 1.03342676e+00
-2.83839405e-01 -1.07073879e+00 -5.61695635e-01 3.52987140e-01
-2.09322557e-01 3.71581495e-01 -5.12078106e-01 1.21312177e+00
1.10263862e-02 5.33010721e-01 -3.53689164e-01 -5.23112059e-01
6.08077884e-01 -1.05144851e-01 1.01208401e+00 -7.09711850e-01
-2.26850271e-01 1.31581992e-01 1.11985087e-01 -6.17972255e-01
-1.56992316e-01 -7.75446177e-01 -8.32202017e-01 7.79014677e-02
-4.84878331e-01 -2.21563905e-01 3.28105092e-01 7.25563586e-01
2.36862078e-01 1.37655273e-01 3.75763237e-01 -1.48355532e+00
-1.20750546e+00 -9.92149413e-01 -4.90149051e-01 7.08875895e-01
1.25426978e-01 -1.36341786e+00 -2.07575709e-01 -3.76112491e-01] | [6.993582725524902, -1.012322187423706] |
4a7abdaa-3a8d-4e9d-8170-c8c7063e1717 | deep-reinforcement-learning-for-programming | 1801.10467 | null | http://arxiv.org/abs/1801.10467v1 | http://arxiv.org/pdf/1801.10467v1.pdf | Deep Reinforcement Learning for Programming Language Correction | Novice programmers often struggle with the formal syntax of programming
languages. To assist them, we design a novel programming language correction
framework amenable to reinforcement learning. The framework allows an agent to
mimic human actions for text navigation and editing. We demonstrate that the
agent can be trained through self-exploration directly from the raw input, that
is, program text itself, without any knowledge of the formal syntax of the
programming language. We leverage expert demonstrations for one tenth of the
training data to accelerate training. The proposed technique is evaluated on
6975 erroneous C programs with typographic errors, written by students during
an introductory programming course. Our technique fixes 14% more programs and
29% more compiler error messages relative to those fixed by a state-of-the-art
tool, DeepFix, which uses a fully supervised neural machine translation
approach. | ['Rahul Gupta', 'Aditya Kanade', 'Shirish Shevade'] | 2018-01-31 | null | null | null | null | ['program-repair', 'program-repair'] | ['computer-code', 'reasoning'] | [ 2.81663626e-01 2.14495867e-01 -4.28264648e-01 -2.98017174e-01
-3.96076351e-01 -9.08964813e-01 2.35232338e-01 5.02917469e-01
-6.14956200e-01 2.85115600e-01 -3.91965508e-01 -1.18478251e+00
3.62959415e-01 -8.14288616e-01 -1.35179007e+00 3.59381020e-01
1.78086892e-01 2.75050044e-01 1.26753435e-01 -6.50359154e-01
6.83264017e-01 1.09155618e-01 -1.38208580e+00 3.09941262e-01
1.50797760e+00 2.46203952e-02 4.69043851e-01 1.22074938e+00
-3.16492975e-01 1.23864424e+00 -7.49949098e-01 -4.85240787e-01
-2.18750183e-02 -2.49059170e-01 -8.59654784e-01 -7.06441224e-01
6.40319824e-01 -4.32552725e-01 -2.23560333e-01 1.47108996e+00
4.63951565e-02 -1.46853864e-01 1.65270060e-01 -9.65095997e-01
-9.21559811e-01 1.35944963e+00 -2.65686929e-01 4.81926128e-02
5.95963955e-01 4.81848598e-01 7.96950102e-01 -6.53071105e-01
5.93560576e-01 8.97715926e-01 8.44430804e-01 8.39710474e-01
-1.16096556e+00 -6.87905550e-01 5.48963957e-02 -1.13142096e-01
-8.88863087e-01 1.66247308e-01 6.89797819e-01 -8.59567344e-01
1.35677862e+00 -4.83937077e-02 8.34285855e-01 1.02003241e+00
4.46794271e-01 8.09060335e-01 9.63642716e-01 -1.02725255e+00
4.08073254e-02 5.36779344e-01 6.06091261e-01 1.45423663e+00
1.57440275e-01 4.10898954e-01 -3.83778960e-01 -1.09251996e-03
4.26071525e-01 -9.62684453e-02 -3.71643566e-02 -5.02722681e-01
-1.25103807e+00 6.30772412e-01 1.19754858e-02 1.01095863e-01
1.95891842e-01 6.03588939e-01 4.72983897e-01 7.83821046e-01
-5.54126203e-01 1.05459523e+00 -5.79964995e-01 -8.63210380e-01
-6.26002312e-01 2.62399197e-01 1.02171707e+00 1.53902173e+00
5.63705742e-01 3.83598387e-01 5.01085997e-01 4.48697329e-01
3.25547129e-01 6.39153957e-01 7.15279996e-01 -8.85426164e-01
8.14445734e-01 1.05864346e+00 -1.20349340e-01 -5.29474914e-01
-1.52031779e-02 -2.76389360e-01 -9.65806842e-02 7.40607202e-01
4.89485055e-01 -3.87208939e-01 -4.88262236e-01 1.56998241e+00
-1.25741318e-01 -2.44546652e-01 1.18607469e-01 4.30202246e-01
3.60036552e-01 2.84103334e-01 5.27302995e-02 3.04829895e-01
9.09776688e-01 -1.16698694e+00 -3.37271154e-01 -2.34468669e-01
1.07597530e+00 -7.18689859e-01 1.71366096e+00 8.18229556e-01
-1.36658657e+00 -7.36389697e-01 -1.39862037e+00 1.50964595e-03
-4.37277257e-01 2.90697068e-01 6.43342495e-01 8.73835444e-01
-8.82595062e-01 9.86039579e-01 -1.09119499e+00 -8.33572894e-02
3.13665457e-02 3.74431670e-01 5.66437468e-02 2.38305107e-01
-7.53414869e-01 8.35707366e-01 3.46083045e-01 -4.03285474e-01
-1.05983484e+00 -1.37484324e+00 -8.66232693e-01 3.53031367e-01
2.10456729e-01 -5.03000438e-01 2.11079311e+00 -1.15330231e+00
-2.11080503e+00 7.73247778e-01 4.87162441e-01 -2.68187225e-01
8.92621100e-01 -6.75643206e-01 -9.05462950e-02 -3.11157495e-01
-1.50799096e-01 2.40901828e-01 7.24394560e-01 -8.26937437e-01
-6.02285981e-01 2.99256533e-01 2.55647331e-01 -1.29492268e-01
-3.67250264e-01 1.68918505e-01 1.64293870e-01 -6.34789824e-01
-4.23351943e-01 -9.98478174e-01 -1.20401643e-01 7.18690753e-02
-1.59276426e-01 -2.52038687e-01 2.44028434e-01 -4.57795918e-01
1.23914778e+00 -1.98139548e+00 3.34497899e-01 3.46136183e-01
2.91329235e-01 2.77105361e-01 -2.61899948e-01 3.97902191e-01
-4.61591899e-01 4.00891960e-01 -3.20526250e-02 -7.72424266e-02
3.42686355e-01 -3.55252922e-02 -6.44432545e-01 3.28328162e-01
-1.99990943e-01 7.03413486e-01 -1.34663820e+00 -2.10060656e-01
2.18123913e-01 -2.10160226e-01 -1.13358903e+00 4.01827842e-01
-5.38234293e-01 3.48566398e-02 -3.74296755e-01 5.99954069e-01
3.17241788e-01 -3.53958942e-02 3.60445619e-01 6.87291145e-01
-3.81991178e-01 5.22892058e-01 -9.98302221e-01 2.07475901e+00
-9.82089281e-01 7.01963365e-01 -3.13673198e-01 -7.60119736e-01
8.21854830e-01 7.47854486e-02 -1.80230558e-01 -6.60062730e-01
-8.39514285e-02 3.77025634e-01 3.56439531e-01 -4.77880567e-01
5.80749989e-01 4.79454547e-01 -2.24027693e-01 1.04033482e+00
1.37318969e-01 -4.65726465e-01 2.78569579e-01 3.36877853e-01
1.33030200e+00 4.23825502e-01 3.87228966e-01 -2.05366462e-01
4.59166437e-01 2.56450713e-01 5.61142080e-02 1.37504053e+00
-4.13786806e-02 -9.05142352e-02 6.07490420e-01 -7.02997148e-01
-1.08438385e+00 -9.97454226e-01 3.82915258e-01 1.60578454e+00
-3.27960402e-01 -7.56581604e-01 -1.01990199e+00 -9.21325386e-01
3.31920385e-03 1.10469377e+00 -3.79127175e-01 -5.85289299e-01
-1.12387717e+00 6.75300732e-02 8.42793465e-01 7.48281717e-01
1.90029144e-01 -1.18347478e+00 -1.13907039e+00 3.28333139e-01
3.81566703e-01 -7.94316590e-01 -6.56920731e-01 4.11970049e-01
-8.76114547e-01 -1.04121304e+00 2.69775447e-02 -1.00321424e+00
8.92799854e-01 -4.92714271e-02 1.32286680e+00 9.61533844e-01
-3.52701217e-01 5.80591977e-01 -1.86058059e-01 -5.03323317e-01
-1.37885237e+00 2.81054705e-01 8.79369527e-02 -1.09898734e+00
2.61203438e-01 -5.72292626e-01 -5.90328760e-02 -2.74203420e-01
-4.61145937e-01 2.58285344e-01 6.06267810e-01 9.86622036e-01
-1.06150798e-01 -1.76162481e-01 -1.25442386e-01 -1.34608555e+00
8.61654878e-01 -8.85864049e-02 -1.19471085e+00 4.50404137e-01
-1.02576029e+00 2.57974267e-01 1.15010810e+00 -6.37488544e-01
-8.06105375e-01 3.58236730e-01 1.77545287e-02 -1.04901694e-01
-2.08761185e-01 6.54279172e-01 4.16401893e-01 -5.24764001e-01
1.17862070e+00 3.47408056e-01 -3.14891249e-01 -2.39242733e-01
4.24126178e-01 3.83618295e-01 8.06605756e-01 -1.49241436e+00
1.03307045e+00 -6.48447931e-01 -4.34003651e-01 -3.51394624e-01
-3.53922576e-01 9.21655353e-03 -6.84583127e-01 2.38924902e-02
3.35376143e-01 -7.78364182e-01 -1.16293335e+00 4.91514683e-01
-1.54993796e+00 -9.70150471e-01 -8.94454494e-02 2.45311305e-01
-6.02393270e-01 2.02737257e-01 -8.26142967e-01 -4.14136350e-01
-1.82999551e-01 -1.55633724e+00 4.94752049e-01 1.85808301e-01
-5.80356240e-01 -8.76018167e-01 3.74296129e-01 1.68521971e-01
4.53238040e-01 -2.69407094e-01 1.54638994e+00 -6.23336613e-01
-7.59195387e-01 -1.91882297e-01 2.72165239e-01 3.04351866e-01
-1.66093379e-01 5.48444748e-01 -5.97718239e-01 -2.26931065e-01
-3.58045697e-01 -4.70024526e-01 2.43305907e-01 -4.55058992e-01
1.32126880e+00 -4.37526822e-01 -9.21620652e-02 7.15761423e-01
1.29924071e+00 1.44033626e-01 1.17859371e-01 8.35989594e-01
8.20783973e-01 1.10745728e-01 2.35951513e-01 3.77447516e-01
2.77759612e-01 3.33189756e-01 2.70111710e-01 4.43223476e-01
2.66842525e-02 -7.71660507e-01 6.12910807e-01 8.62599194e-01
2.53596846e-02 2.29824036e-01 -1.43976367e+00 5.24160147e-01
-1.48763478e+00 -7.28169858e-01 3.36966328e-02 1.92055249e+00
1.37857902e+00 4.76339906e-01 -1.61998272e-01 -8.77143256e-03
3.64584744e-01 -4.52500194e-01 -3.16910297e-01 -1.05911839e+00
8.49721134e-01 7.76090920e-01 3.97195667e-01 8.32027435e-01
-6.75311387e-01 1.22341299e+00 6.29278183e+00 1.71078399e-01
-1.34162927e+00 -1.42610267e-01 -2.34102592e-01 3.23481917e-01
-3.36238891e-01 3.04293148e-02 -6.94291353e-01 4.27898169e-01
1.18259728e+00 -3.18947077e-01 9.49006736e-01 1.56791735e+00
-1.25780910e-01 1.12701200e-01 -1.71062911e+00 4.82551157e-01
-2.31708050e-01 -1.43894911e+00 -1.32367045e-01 -5.76419294e-01
8.72481644e-01 -2.01283202e-01 2.51335382e-01 1.04596615e+00
1.06407952e+00 -1.03624690e+00 1.11109793e+00 1.93184987e-01
6.97591543e-01 -6.82945549e-01 3.28237474e-01 7.05413938e-01
-5.28014660e-01 -1.95248082e-01 -2.07513750e-01 -4.16340679e-01
-7.33691990e-01 -1.00570366e-01 -1.32806337e+00 -2.07527652e-01
5.10052085e-01 5.16813219e-01 -1.02687478e+00 6.87761843e-01
-6.74782515e-01 6.90909326e-01 5.37413880e-02 -5.96725702e-01
-2.68715154e-02 1.94591686e-01 3.14248145e-01 1.47208941e+00
3.24889570e-01 4.28630151e-02 3.76517653e-01 1.42505109e+00
-2.79458761e-01 1.66986138e-02 -5.69080651e-01 -4.50918108e-01
5.34820676e-01 9.40018177e-01 -2.52842873e-01 -2.76703864e-01
-3.96118015e-01 9.58901465e-01 7.83497453e-01 2.68675119e-01
-8.59259069e-01 -7.77418792e-01 4.43002194e-01 -1.75385401e-01
1.56952754e-01 -3.47968489e-01 -6.33476853e-01 -1.08842957e+00
1.41325861e-01 -1.54599774e+00 -1.69273913e-01 -6.93776011e-01
-6.78445935e-01 3.78223091e-01 -2.76867002e-01 -1.01669967e+00
-3.48822862e-01 -8.65635455e-01 -1.11228073e+00 8.06908131e-01
-1.41829717e+00 -6.70726061e-01 -3.18847805e-01 2.53137857e-01
4.92460907e-01 -7.23372877e-01 9.67582405e-01 1.88272983e-01
-5.77737689e-01 1.14129806e+00 -1.57569572e-01 2.36382112e-01
6.34718895e-01 -1.92118406e+00 7.65461206e-01 1.02767253e+00
-4.72768173e-02 1.52580845e+00 8.38157952e-01 -5.25865793e-01
-1.88344204e+00 -8.32898140e-01 7.83497095e-01 -5.97692013e-01
1.07534099e+00 -3.00149500e-01 -8.07906806e-01 9.51298654e-01
3.07769328e-01 -3.28405946e-02 5.77419341e-01 1.50742456e-02
-6.09978318e-01 1.99689567e-01 -7.45694339e-01 8.55821967e-01
6.75734460e-01 -8.50287080e-01 -1.01872659e+00 4.80404735e-01
7.76839852e-01 -1.24958503e+00 -7.43491709e-01 -1.86318949e-01
5.02464533e-01 -5.67981422e-01 7.23399758e-01 -9.57321644e-01
1.12725592e+00 -1.86484694e-01 1.39137000e-01 -1.53964627e+00
2.36916803e-02 -1.07376933e+00 -2.27039695e-01 8.74611676e-01
5.26412189e-01 -9.76339728e-02 8.15317988e-01 6.65853083e-01
-5.96927226e-01 -4.43915367e-01 -2.24737719e-01 -3.69999439e-01
6.65564477e-01 -5.42700350e-01 5.02299190e-01 1.21629131e+00
7.59972036e-01 9.01799873e-02 7.23154172e-02 2.27686644e-01
3.33781451e-01 2.13659704e-01 1.09638321e+00 -1.00836468e+00
-7.83707142e-01 -7.75184214e-01 1.27862871e-01 -1.17499387e+00
6.59468234e-01 -1.15489638e+00 1.65715888e-01 -4.58293408e-01
-5.87164424e-02 -4.20897990e-01 5.83250523e-02 6.68013036e-01
-7.09711760e-02 -4.85029995e-01 2.10868612e-01 -2.44708449e-01
-5.94103098e-01 1.22229673e-01 9.77766871e-01 -3.53095680e-01
-4.14943457e-01 3.72582152e-02 -5.44664502e-01 1.13675451e+00
5.27729094e-01 -6.75876856e-01 -2.47323707e-01 -7.40005851e-01
6.98238790e-01 3.63901556e-01 3.13668281e-01 -1.28166175e+00
6.54227197e-01 -3.46546859e-01 1.57000452e-01 -5.09899370e-02
-7.50313401e-01 -9.26144004e-01 -5.28596520e-01 8.15627038e-01
-1.02025592e+00 6.57570183e-01 5.87195814e-01 2.02197194e-01
1.13188542e-01 -8.90382230e-01 8.21150124e-01 -2.75120556e-01
-8.17940056e-01 -4.55043554e-01 -6.82409525e-01 3.81909162e-02
8.66403699e-01 1.39958337e-01 -5.66374838e-01 1.98562846e-01
-5.77851355e-01 9.55832750e-02 4.81526703e-01 4.21497285e-01
5.49581647e-01 -9.26465631e-01 -2.38890335e-01 3.88158321e-01
2.44542137e-01 -2.97925591e-01 -3.49938869e-01 4.64373112e-01
-1.27077448e+00 3.15763474e-01 -5.40863931e-01 -5.69696844e-01
-1.24775338e+00 6.99622273e-01 5.17668784e-01 -3.53679836e-01
-6.88272476e-01 6.95797086e-01 -4.68502551e-01 -1.18930554e+00
4.42246288e-01 -1.05926573e+00 4.01020288e-01 -7.81691313e-01
6.28373206e-01 4.51668799e-02 2.14225858e-01 6.08787239e-01
-1.46830529e-01 4.16653246e-01 -5.26389062e-01 2.12061912e-01
1.24776232e+00 4.87249255e-01 5.00969291e-02 2.95625687e-01
7.07651794e-01 4.36173975e-01 -8.64847243e-01 -1.44888386e-01
2.83049941e-01 -3.21931392e-01 -1.45849839e-01 -1.27636766e+00
-5.28410316e-01 1.13040292e+00 5.42383969e-01 -2.71059781e-01
4.03107554e-01 -7.20683277e-01 4.04618293e-01 1.37225342e+00
4.49580729e-01 -1.02074182e+00 4.40924644e-01 9.10932541e-01
6.06137812e-01 -1.39363158e+00 -1.42519817e-01 -2.43983343e-01
-3.55990380e-01 1.80268288e+00 1.20740175e+00 -3.78782541e-01
1.85081929e-01 5.94330728e-01 -5.80790043e-02 1.21116571e-01
-8.06041181e-01 7.38950491e-01 -1.38180926e-01 3.69638085e-01
8.40508878e-01 -4.17116508e-02 1.84680089e-01 7.07258463e-01
-5.54764926e-01 2.59798974e-01 1.42813921e+00 1.31499565e+00
-2.71132946e-01 -1.21824741e+00 -2.18091026e-01 3.77691314e-02
-2.45766550e-01 -5.00131547e-01 -2.63435066e-01 9.62269723e-01
-2.36351807e-02 2.49862626e-01 -2.87086487e-01 -3.48206162e-01
5.84119916e-01 3.16022605e-01 8.26743066e-01 -1.07530987e+00
-1.22093844e+00 -8.35332215e-01 -2.10993841e-01 -5.33308446e-01
3.65986615e-01 -6.20324552e-01 -1.52539527e+00 -6.71399474e-01
2.15498749e-02 3.02650351e-02 6.40507281e-01 7.39729941e-01
-1.85047686e-02 8.68700922e-01 2.31539398e-01 -2.87523836e-01
-1.29147100e+00 -6.21836126e-01 1.10021256e-01 2.82502770e-02
6.63968086e-01 -3.06642562e-01 -8.32156762e-02 3.32394332e-01] | [8.17969036102295, 7.627690315246582] |
4714530d-3762-44b2-a328-1ab1b73bc98c | discovering-playing-patterns-time-series | 1710.02268 | null | http://arxiv.org/abs/1710.02268v1 | http://arxiv.org/pdf/1710.02268v1.pdf | Discovering Playing Patterns: Time Series Clustering of Free-To-Play Game Data | The classification of time series data is a challenge common to all
data-driven fields. However, there is no agreement about which are the most
efficient techniques to group unlabeled time-ordered data. This is because a
successful classification of time series patterns depends on the goal and the
domain of interest, i.e. it is application-dependent.
In this article, we study free-to-play game data. In this domain, clustering
similar time series information is increasingly important due to the large
amount of data collected by current mobile and web applications. We evaluate
which methods cluster accurately time series of mobile games, focusing on
player behavior data. We identify and validate several aspects of the
clustering: the similarity measures and the representation techniques to reduce
the high dimensionality of time series. As a robustness test, we compare
various temporal datasets of player activity from two free-to-play video-games.
With these techniques we extract temporal patterns of player behavior
relevant for the evaluation of game events and game-business diagnosis. Our
experiments provide intuitive visualizations to validate the results of the
clustering and to determine the optimal number of clusters. Additionally, we
assess the common characteristics of the players belonging to the same group.
This study allows us to improve the understanding of player dynamics and churn
behavior. | ['África Periáñez', 'Anna Guitart', 'Alain Saas'] | 2017-10-06 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [-3.96828294e-01 -6.57197297e-01 -1.09479874e-01 2.51855962e-02
-1.64731786e-01 -9.01632309e-01 3.07637215e-01 5.42775750e-01
-3.69368643e-01 1.83548465e-01 3.17030512e-02 -3.19115102e-01
-8.28385830e-01 -8.91422510e-01 2.63110735e-02 -6.83047295e-01
-6.62077546e-01 6.07529759e-01 6.00241780e-01 -5.10429740e-01
4.18078005e-01 3.73255908e-01 -2.11069894e+00 4.51564431e-01
7.50638247e-01 8.39783192e-01 1.83954746e-01 6.15422845e-01
-2.54923165e-01 8.69854450e-01 -7.67202735e-01 6.64801151e-02
2.66300797e-01 -7.49234796e-01 -5.61618209e-01 5.76919615e-02
-6.78016782e-01 2.99116910e-01 -4.93549705e-02 7.34139502e-01
2.23041892e-01 3.73948306e-01 5.86465538e-01 -1.60169220e+00
1.76805243e-01 5.35662949e-01 -3.21810722e-01 7.06562400e-01
4.75853026e-01 -1.81596309e-01 8.18698406e-01 -8.61403644e-02
8.53583574e-01 6.70421124e-01 4.81732965e-01 3.69998813e-02
-8.84988904e-01 -4.28326368e-01 -4.05786969e-02 7.65110850e-01
-1.41173506e+00 7.75085837e-02 8.58299255e-01 -7.74359345e-01
6.05068207e-01 5.45018911e-01 1.08502483e+00 6.64061964e-01
-2.06741840e-01 3.53774011e-01 9.92157042e-01 -4.14284468e-01
6.00154400e-01 8.73151273e-02 3.66997659e-01 7.14680776e-02
5.20646572e-05 -1.95311666e-01 -3.06550831e-01 -1.17751144e-01
4.54867005e-01 3.46339166e-01 -4.44956236e-02 -5.28763473e-01
-7.39755154e-01 8.78448725e-01 -1.53883457e-01 9.54269230e-01
-3.09966147e-01 -3.42475355e-01 7.26303816e-01 6.63495541e-01
4.72131908e-01 3.99198711e-01 -2.94659793e-01 -1.13245928e+00
-8.58206511e-01 3.32304060e-01 1.01336861e+00 4.71392453e-01
4.18019056e-01 -1.75630301e-01 2.45976552e-01 7.20122099e-01
-1.32503033e-01 -1.89347193e-03 8.71753573e-01 -7.55885482e-01
2.33226374e-01 9.69016492e-01 -7.41150975e-02 -1.25183010e+00
-4.98170882e-01 -3.26289833e-01 -4.65218306e-01 1.64862543e-01
8.37554276e-01 1.21141747e-01 -4.29062307e-01 1.39978051e+00
1.16659001e-01 2.14681447e-01 -1.42450601e-01 8.09581220e-01
5.02874553e-01 4.66037035e-01 -1.25683904e-01 -6.29000783e-01
1.25848699e+00 -2.76837289e-01 -7.26453364e-01 3.47577453e-01
9.34126377e-01 -4.75555688e-01 9.66419697e-01 4.63924170e-01
-8.16579700e-01 -3.92915308e-01 -6.68938518e-01 6.77907467e-01
-7.21389174e-01 -2.65411258e-01 3.90816122e-01 7.39346743e-01
-6.37660623e-01 7.98535109e-01 -1.00819027e+00 -6.57979965e-01
-3.64088826e-02 3.40427458e-01 -1.78912997e-01 4.50486720e-01
-1.06287336e+00 6.42536163e-01 5.06739259e-01 -5.03755867e-01
-2.44416699e-01 -6.56215906e-01 -3.26193243e-01 1.21323466e-01
4.10762668e-01 1.40434816e-01 9.73092198e-01 -9.53323126e-01
-1.01077890e+00 8.77886653e-01 1.30132988e-01 -6.04048669e-01
6.08265877e-01 4.00260180e-01 -7.04704762e-01 5.07230833e-02
2.56122947e-01 -3.78844291e-01 2.83839732e-01 -8.24059308e-01
-9.29934382e-01 -3.92340690e-01 -1.28209889e-01 1.40553266e-01
-6.26209021e-01 2.42813900e-01 -5.79014659e-01 -4.68598783e-01
1.29696578e-01 -7.96709478e-01 -1.43773407e-01 -8.49769711e-01
1.82243407e-01 -2.24544108e-01 9.78062451e-01 -4.34741348e-01
1.93470001e+00 -2.48035383e+00 9.98336300e-02 6.47307336e-01
3.53592187e-01 -1.53611422e-01 3.21918875e-01 8.53128195e-01
-3.16650867e-01 2.53690243e-01 1.35257974e-01 1.59800410e-01
-7.55557343e-02 2.62544632e-01 -1.47766173e-01 3.65633160e-01
-4.54231352e-01 4.66268837e-01 -8.95833194e-01 -4.32434022e-01
3.49570990e-01 -1.10704139e-01 -3.75644326e-01 2.06813827e-01
1.08421765e-01 4.77912933e-01 -4.00777996e-01 2.97709644e-01
2.09230214e-01 -3.05582941e-01 3.97224784e-01 1.09680042e-01
-4.86899465e-01 2.41470516e-01 -1.23431396e+00 1.29562378e+00
5.33842593e-02 6.28682852e-01 -4.82803524e-01 -1.32081246e+00
9.81106579e-01 2.72180319e-01 1.34817290e+00 -9.74061131e-01
2.94657886e-01 2.71290630e-01 4.51826096e-01 -6.79521859e-01
5.79873264e-01 1.08214349e-01 -2.78651536e-01 6.91905379e-01
-1.70026869e-01 1.33468777e-01 9.17572618e-01 6.85428679e-02
1.34533024e+00 -3.48413080e-01 4.91420448e-01 -2.44530082e-01
1.65359020e-01 3.56876493e-01 5.57363331e-01 4.04390633e-01
-1.92658454e-01 5.48196137e-01 8.93271267e-01 -5.43933272e-01
-9.92548406e-01 -8.04466426e-01 1.95855051e-01 1.12555599e+00
1.66447610e-01 -8.24329376e-01 -5.97075641e-01 -2.94789016e-01
-3.06135058e-01 2.61354864e-01 -6.97133422e-01 6.07634485e-02
-4.36937183e-01 -8.32451880e-01 2.13153750e-01 9.47893262e-02
5.87426797e-02 -9.80538130e-01 -9.87378359e-01 4.00557518e-01
-2.86666632e-01 -8.28337073e-01 -3.09818517e-02 3.44388455e-01
-9.65789795e-01 -1.47515845e+00 -1.47226870e-01 -5.82699299e-01
1.60796940e-01 2.69966960e-01 1.23033023e+00 2.67950557e-02
-2.03125566e-01 4.59944785e-01 -9.49130535e-01 -4.62996364e-01
-4.46362436e-01 2.63307631e-01 -6.53136447e-02 -1.63713638e-02
8.74890804e-01 -8.70204568e-01 -3.65487248e-01 6.90930188e-01
-8.83062065e-01 -2.70676196e-01 -1.02073245e-01 1.57978475e-01
4.23612297e-01 8.19604874e-01 2.89237648e-02 -6.38767004e-01
1.08839047e+00 -8.88181508e-01 -4.05027479e-01 1.69914186e-01
-5.28462768e-01 -4.25883830e-01 6.60140038e-01 -6.84999049e-01
-2.37498507e-01 -1.44495368e-01 2.74326086e-01 -5.36843777e-01
-1.67667001e-01 7.80949712e-01 1.53078392e-01 3.83275628e-01
8.89955759e-01 2.09637508e-01 1.21671304e-01 -4.95936573e-01
-2.65485328e-02 6.43797874e-01 2.63996542e-01 -3.85638952e-01
5.23587763e-01 5.86908996e-01 -2.73312807e-01 -9.61934149e-01
1.37470700e-02 -1.08547437e+00 -5.89621425e-01 -7.79842019e-01
7.78396606e-01 -4.12076265e-01 -9.72626686e-01 3.59555125e-01
-6.42745078e-01 -2.42717981e-01 -7.22335696e-01 5.44470310e-01
-5.55921614e-01 1.95254594e-01 -3.35635394e-01 -9.66825843e-01
2.42969841e-01 -8.43626976e-01 3.68197620e-01 1.55359775e-01
-5.74342251e-01 -1.01253796e+00 3.49897981e-01 3.76955159e-02
1.22650452e-01 4.95857000e-01 7.94879436e-01 -9.37542140e-01
-1.72890753e-01 -2.18409032e-01 3.57565582e-01 -2.14554578e-01
2.80098528e-01 2.80759454e-01 -5.28182268e-01 -9.82747972e-02
2.30776817e-01 2.95563161e-01 2.41449028e-01 5.45206547e-01
1.12159598e+00 3.70528325e-02 -3.55096936e-01 2.16739774e-01
1.20711255e+00 8.21307123e-01 6.90777659e-01 5.78110158e-01
4.52907026e-01 1.04433954e+00 9.61210132e-01 8.39411616e-01
2.24919572e-01 9.17413235e-01 3.64580870e-01 1.24302253e-01
5.01618624e-01 -9.75166261e-03 1.30691707e-01 1.13803864e+00
-6.30370915e-01 -1.98873043e-01 -1.17804778e+00 6.43918633e-01
-2.17606139e+00 -1.42289186e+00 -5.36853909e-01 2.18904424e+00
3.26372534e-01 2.05950245e-01 1.01814365e+00 9.65347648e-01
7.60479808e-01 -8.32092166e-02 -1.55259758e-01 -2.94722497e-01
1.14460960e-01 2.66145058e-02 2.64049739e-01 8.02623555e-02
-9.72424388e-01 4.90426183e-01 6.00447893e+00 9.59887028e-01
-1.18807423e+00 1.34726167e-01 4.72512603e-01 -3.68698418e-01
-4.15309854e-02 -2.04803109e-01 -1.69521838e-01 8.81973743e-01
1.04766703e+00 -6.33920193e-01 4.58101869e-01 8.17681849e-01
7.73390949e-01 -6.15580156e-02 -8.98082137e-01 1.23746181e+00
-1.74110264e-01 -1.15316260e+00 -5.66041410e-01 4.47889179e-01
3.18916738e-01 -1.65264219e-01 -1.69349074e-01 2.50461996e-01
1.82742774e-01 -8.89912128e-01 6.67261302e-01 4.24743652e-01
3.74470860e-01 -7.17999339e-01 5.85483551e-01 4.82659936e-01
-1.45321369e+00 -3.12825024e-01 1.02066599e-01 -5.06389976e-01
8.89179483e-03 4.75629717e-01 -5.52097619e-01 4.33227956e-01
1.10765433e+00 8.53015363e-01 -6.02167010e-01 1.13521445e+00
4.43893313e-01 5.96296191e-01 -5.06007671e-01 -3.11488271e-01
3.38241383e-02 -5.95031142e-01 4.19547588e-01 8.74310434e-01
5.89585185e-01 2.47478157e-01 1.65772438e-01 5.60334921e-01
6.22590303e-01 3.89156222e-01 -6.44090176e-01 -3.34466875e-01
5.39363027e-01 1.02713335e+00 -1.30871022e+00 -5.65095879e-02
-3.18284363e-01 5.89630842e-01 -4.64065485e-02 8.43470544e-02
-7.05322027e-01 -2.30323523e-01 8.38324428e-01 5.99649429e-01
2.27155417e-01 -4.94457811e-01 -1.81220695e-01 -1.01970017e+00
1.38498694e-01 -9.62046742e-01 7.98523128e-01 -3.50777477e-01
-1.26921153e+00 5.09665787e-01 2.06873357e-01 -2.07212496e+00
-7.38032699e-01 -3.07368904e-01 -6.39636517e-01 3.98274869e-01
-5.83372533e-01 -4.82000887e-01 -3.88199508e-01 8.97506595e-01
2.92259514e-01 -3.46474111e-01 6.40174150e-01 4.83375996e-01
-2.40099549e-01 -1.86275728e-02 3.23906928e-01 1.80810586e-01
2.58981466e-01 -1.24213243e+00 1.09183580e-01 7.62467861e-01
3.06452245e-01 3.47481906e-01 8.16374063e-01 -5.61504841e-01
-9.09373164e-01 -4.85685140e-01 6.52349710e-01 -4.60845113e-01
9.17707801e-01 -3.65658224e-01 -9.88568604e-01 6.34296387e-02
-1.76017940e-01 -3.56773853e-01 1.14113748e+00 3.17671925e-01
1.26698658e-01 -1.70878932e-01 -7.78679371e-01 5.13991952e-01
9.34048653e-01 -5.15184522e-01 -4.87489223e-01 2.98761260e-02
1.00439668e-01 -1.44696921e-01 -1.01789713e+00 -3.53589840e-03
4.66673136e-01 -1.40671444e+00 6.74521446e-01 -5.62538385e-01
2.47075200e-01 -3.27339619e-01 5.90829179e-02 -1.27462649e+00
-3.47595155e-01 -4.90135014e-01 4.41583544e-02 1.09983265e+00
9.20898318e-02 -2.41584256e-01 1.09464526e+00 4.46328193e-01
3.58489066e-01 -4.16777492e-01 -8.93155098e-01 -9.14978683e-01
-1.97253406e-01 -8.34339261e-01 6.47297144e-01 1.19570017e+00
6.15635753e-01 3.10615413e-02 -1.80966899e-01 -2.38441050e-01
2.89286345e-01 3.05753469e-01 8.46691191e-01 -1.63815379e+00
-2.76756763e-01 -8.60406399e-01 -1.01181424e+00 -4.19231176e-01
-4.04038429e-01 -5.27853847e-01 -3.74833047e-01 -1.13106525e+00
-1.49083346e-01 -4.86705750e-01 -1.95831314e-01 4.05615643e-02
2.13615924e-01 1.52298391e-01 2.87434220e-01 7.12338507e-01
-7.26122856e-01 2.62009539e-03 6.99475586e-01 1.92847848e-01
-7.69802332e-01 2.56187826e-01 -2.69261718e-01 4.74424571e-01
9.59941030e-01 -5.61102748e-01 -7.61405110e-01 2.30021238e-01
3.80496979e-01 1.59175515e-01 5.94280213e-02 -1.13719511e+00
3.32964987e-01 -2.83959925e-01 -2.56834924e-01 -4.71915960e-01
1.76841505e-02 -1.06675315e+00 5.27641296e-01 5.39457679e-01
-2.79287379e-02 2.97591448e-01 3.58122140e-02 5.37683904e-01
-5.76339900e-01 -1.57089412e-01 4.18735772e-01 -1.96455151e-01
-9.87072587e-01 1.20941356e-01 -8.90979171e-01 2.00214013e-01
1.33550882e+00 -6.77903175e-01 8.73417705e-02 -8.43826890e-01
-1.01404607e+00 7.81640410e-02 4.88152653e-01 5.82873940e-01
1.44236237e-01 -1.23941720e+00 -4.14403170e-01 4.50436473e-02
2.87685573e-01 -5.22902608e-01 2.44186237e-01 8.75988603e-01
-6.41107798e-01 7.60398805e-02 -5.49633861e-01 -7.98221231e-01
-1.63364434e+00 7.52616048e-01 2.62622625e-01 -3.86443079e-01
-3.73365015e-01 2.42143702e-02 -3.04427564e-01 -6.87360168e-02
1.43485978e-01 -3.06375265e-01 -8.32203209e-01 5.75378895e-01
4.04555202e-01 7.22630620e-01 2.93127716e-01 -4.42016184e-01
-4.08592999e-01 4.37053710e-01 2.03645676e-01 -2.28564128e-01
1.50639439e+00 -1.48067445e-01 -5.43748811e-02 1.00742817e+00
9.82513249e-01 -1.31412670e-01 -7.44902670e-01 1.89530879e-01
5.22909760e-01 -5.08388400e-01 -4.63001758e-01 -1.78534046e-01
-9.10164237e-01 5.94224632e-01 7.42155313e-01 1.43618643e+00
1.31136131e+00 -8.58971700e-02 2.11006403e-01 -5.90957329e-02
4.97833729e-01 -1.25967145e+00 -2.17738748e-02 4.23564404e-01
3.36791456e-01 -9.18562651e-01 -2.54606783e-01 -4.58829582e-01
-6.84157729e-01 1.04690599e+00 3.08898866e-01 -3.42769437e-02
9.96549428e-01 4.34053600e-01 1.97871268e-01 -4.11389053e-01
-5.74741721e-01 -5.54158449e-01 6.02458045e-02 6.89482868e-01
4.46855128e-01 1.08162276e-01 -9.00775194e-01 6.68237805e-01
-5.63081503e-01 2.44966522e-02 4.82074916e-01 8.33222270e-01
-2.69582748e-01 -1.28113627e+00 -4.22625154e-01 5.07482529e-01
-4.29557323e-01 4.29455638e-01 -5.68310916e-01 9.34735954e-01
1.87704429e-01 1.18951845e+00 3.41706038e-01 -6.88627303e-01
4.31964785e-01 -3.04478351e-02 7.80279189e-02 -4.56930161e-01
-9.08564925e-01 1.59280509e-01 -1.29746377e-01 -5.51367581e-01
-5.34004927e-01 -8.90510321e-01 -1.09937584e+00 -7.09003091e-01
3.96632329e-02 6.78741813e-01 6.08556867e-01 7.82778919e-01
3.09876621e-01 4.66140270e-01 6.91037893e-01 -4.22170639e-01
1.43858239e-01 -7.81792879e-01 -9.76669133e-01 1.05266738e+00
-3.97668421e-01 -5.53259015e-01 -3.51356864e-01 1.21851377e-01] | [7.235499382019043, 3.3399460315704346] |
d9fcbfa3-332a-4ffc-ac57-a9ee6bc4404a | flow-guided-transformer-for-video-inpainting | 2208.06768 | null | https://arxiv.org/abs/2208.06768v1 | https://arxiv.org/pdf/2208.06768v1.pdf | Flow-Guided Transformer for Video Inpainting | We propose a flow-guided transformer, which innovatively leverage the motion discrepancy exposed by optical flows to instruct the attention retrieval in transformer for high fidelity video inpainting. More specially, we design a novel flow completion network to complete the corrupted flows by exploiting the relevant flow features in a local temporal window. With the completed flows, we propagate the content across video frames, and adopt the flow-guided transformer to synthesize the rest corrupted regions. We decouple transformers along temporal and spatial dimension, so that we can easily integrate the locally relevant completed flows to instruct spatial attention only. Furthermore, we design a flow-reweight module to precisely control the impact of completed flows on each spatial transformer. For the sake of efficiency, we introduce window partition strategy to both spatial and temporal transformers. Especially in spatial transformer, we design a dual perspective spatial MHSA, which integrates the global tokens to the window-based attention. Extensive experiments demonstrate the effectiveness of the proposed method qualitatively and quantitatively. Codes are available at https://github.com/hitachinsk/FGT. | ['Dong Liu', 'Jingjing Fu', 'Kaidong Zhang'] | 2022-08-14 | null | null | null | null | ['video-inpainting'] | ['computer-vision'] | [-2.26883620e-01 -3.03109318e-01 -2.05836505e-01 -1.42698184e-01
-5.27965188e-01 -5.16497850e-01 4.01337475e-01 -3.61065626e-01
-2.16852248e-01 6.20659053e-01 7.43880033e-01 -9.95746627e-02
-1.81140110e-01 -8.06285203e-01 -6.72287464e-01 -6.09705806e-01
1.24475271e-01 -3.30335259e-01 4.34975177e-01 9.74010825e-02
2.81857044e-01 3.81621450e-01 -1.03045201e+00 1.72520354e-01
9.88641679e-01 9.74938214e-01 3.97413343e-01 4.89648908e-01
-2.68647343e-01 1.26106191e+00 -2.07548156e-01 -1.14788979e-01
3.10287893e-01 -6.11375034e-01 -7.26751864e-01 1.36163697e-01
3.09437692e-01 -7.91418016e-01 -9.68215644e-01 8.07334781e-01
3.07539523e-01 4.57006335e-01 2.25009724e-01 -1.25782299e+00
-7.08055794e-01 3.92132103e-01 -7.88945258e-01 6.55729532e-01
3.38266879e-01 6.60226583e-01 9.27428484e-01 -9.30112958e-01
6.29213512e-01 1.28589475e+00 2.25279272e-01 4.17443067e-01
-8.57041180e-01 -7.47824430e-01 6.56254888e-01 3.57327133e-01
-1.07974088e+00 -6.37715220e-01 1.26681805e+00 -5.16674936e-01
4.03380066e-01 1.96898907e-01 7.31681108e-01 6.90612912e-01
-7.22345188e-02 9.26695108e-01 6.10853374e-01 1.93282813e-02
-7.76173770e-02 -4.51602370e-01 -1.66101635e-01 8.46824288e-01
-2.46529102e-01 8.73837546e-02 -5.32979310e-01 2.19074264e-01
1.09251380e+00 4.71656948e-01 -6.73594475e-01 -1.72491014e-01
-1.39225495e+00 6.17274880e-01 6.64443254e-01 1.26315221e-01
-3.95677924e-01 3.16928238e-01 4.22418803e-01 6.37099147e-02
6.84108257e-01 1.27906218e-01 -2.30309650e-01 -1.52435377e-01
-1.12338102e+00 8.72959346e-02 2.03095376e-01 1.03079045e+00
9.53837454e-01 8.12215433e-02 -7.76176810e-01 6.84390128e-01
2.62269914e-01 2.24222139e-01 1.82500437e-01 -1.40998411e+00
8.54134381e-01 4.28097069e-01 1.79568738e-01 -1.22080135e+00
1.08446941e-01 -2.95507789e-01 -9.22141433e-01 -7.42729008e-02
2.56161660e-01 -1.85363874e-01 -6.97622836e-01 1.69804418e+00
4.48498964e-01 7.50186741e-01 -3.94078523e-01 1.28559661e+00
5.77787459e-01 1.00302839e+00 -1.17469747e-02 -3.79822254e-01
1.11970460e+00 -1.43961501e+00 -7.47138143e-01 -1.65886153e-02
3.68879020e-01 -7.04414189e-01 1.08924878e+00 -4.52242419e-02
-1.34497046e+00 -5.05602717e-01 -8.99610341e-01 -4.10639763e-01
2.54739374e-01 -1.87523700e-02 2.28763282e-01 -1.17187172e-01
-7.65169382e-01 6.26854181e-01 -1.06916142e+00 5.62892146e-02
7.67051816e-01 6.08727057e-03 -6.49116039e-02 -3.46070439e-01
-1.27693892e+00 3.24534684e-01 1.84633471e-02 3.05305034e-01
-9.98277843e-01 -1.09044647e+00 -9.98089910e-01 1.39045537e-01
4.07725960e-01 -7.15604663e-01 1.07262337e+00 -7.81009018e-01
-1.34381020e+00 2.50895262e-01 -4.84934121e-01 -8.32836702e-02
6.93971217e-01 -1.36712700e-01 -1.71113275e-02 5.30132711e-01
3.23170692e-01 6.90146685e-01 7.90032685e-01 -1.01579320e+00
-7.59241104e-01 1.19901963e-01 3.02567780e-01 3.64484280e-01
-4.95151788e-01 9.67993587e-03 -8.69907737e-01 -9.46814895e-01
-7.15376511e-02 -4.57002223e-01 -1.54764026e-01 4.72677737e-01
-3.91394287e-01 9.23406854e-02 8.94592583e-01 -8.66597056e-01
1.64522815e+00 -2.21471334e+00 3.85717034e-01 -1.19145490e-01
6.27100766e-01 1.74535468e-01 -2.01703548e-01 2.93466926e-01
7.52856433e-02 7.32682198e-02 -4.65404034e-01 -5.11066616e-01
-3.11659724e-01 2.77183622e-01 -4.92643952e-01 3.61092746e-01
3.62213075e-01 9.02404368e-01 -1.17431581e+00 -5.36132932e-01
5.02182543e-01 6.94839060e-01 -8.16636443e-01 4.21886116e-01
-1.06686115e-01 7.50191689e-01 -9.66569602e-01 2.58196056e-01
9.44400132e-01 -2.50415087e-01 -3.61929983e-01 -4.53809232e-01
-4.43185508e-01 4.36632335e-01 -9.47392583e-01 1.86828268e+00
-5.86924911e-01 5.30413747e-01 2.57017881e-01 -7.40760088e-01
5.73725522e-01 1.39768019e-01 5.55482626e-01 -7.23224282e-01
-5.49707972e-02 -4.48618736e-03 -4.31991845e-01 -7.51069605e-01
3.70847672e-01 -4.84995395e-02 2.98787743e-01 4.30339426e-01
-1.67783778e-02 3.49572629e-01 2.05941215e-01 3.69820476e-01
1.00080967e+00 3.78169894e-01 -1.29957378e-01 -6.89887581e-03
8.62276912e-01 -3.31639409e-01 9.12721574e-01 2.46045113e-01
-4.67489123e-01 9.24815297e-01 7.33949840e-01 -5.17071068e-01
-9.75948513e-01 -9.81514752e-01 4.51414645e-01 1.00171542e+00
5.65661728e-01 -3.54899764e-01 -5.46762049e-01 -6.13646984e-01
-2.14270711e-01 1.31739035e-01 -6.10942483e-01 -1.11911409e-01
-9.72252548e-01 -2.50823259e-01 1.56454951e-01 5.12428999e-01
8.51987422e-01 -1.17545807e+00 -3.92073244e-01 3.44334841e-01
-5.05558312e-01 -1.09566879e+00 -1.37897027e+00 -5.51712096e-01
-8.09347153e-01 -9.39227223e-01 -1.19088125e+00 -8.13632786e-01
4.93749589e-01 6.01339757e-01 7.32105196e-01 3.28543007e-01
2.89670397e-02 1.19811468e-01 -4.05944616e-01 5.85317135e-01
6.00785203e-02 1.73163816e-01 -4.16166693e-01 4.50283796e-01
-3.42396587e-01 -9.21827018e-01 -1.24750054e+00 2.78130054e-01
-1.05413747e+00 3.52593839e-01 3.33495647e-01 7.40487635e-01
3.39178711e-01 -1.40459016e-01 3.43229830e-01 -5.36784470e-01
4.97366190e-01 -6.36343360e-01 -3.63130569e-01 -1.07714087e-02
-3.86360399e-02 5.52984886e-02 7.57578731e-01 -3.61549348e-01
-1.13641548e+00 -1.71850219e-01 -9.13666785e-02 -1.02181482e+00
1.13423936e-01 2.52730399e-01 -3.11411917e-01 1.24819227e-01
2.42966432e-02 3.73851031e-01 -7.28143752e-02 -4.87635761e-01
3.60019922e-01 2.30967626e-01 3.81548345e-01 -6.53328419e-01
8.49837959e-01 6.87053978e-01 -2.64756203e-01 -3.51499468e-01
-7.95677781e-01 -3.34703892e-01 -2.79964358e-01 -3.20159227e-01
1.00864971e+00 -8.38205755e-01 -6.96450472e-01 4.86219645e-01
-1.37626874e+00 -4.77498382e-01 -2.72966295e-01 3.73736650e-01
-3.94879639e-01 5.02179921e-01 -9.36810732e-01 -4.73673046e-01
-2.00952262e-01 -1.21635413e+00 1.09175754e+00 3.68366301e-01
2.60909468e-01 -1.05262637e+00 1.12255923e-01 1.60100326e-01
3.13126236e-01 1.65389508e-01 4.81543779e-01 1.11473344e-01
-1.08305323e+00 4.08942610e-01 -6.67196214e-01 2.61776716e-01
4.14644331e-01 4.64754216e-02 -7.95122683e-01 -2.73601234e-01
-7.71423951e-02 9.73203704e-02 1.18665338e+00 3.93376976e-01
1.18752062e+00 -5.71398616e-01 -2.31500164e-01 8.84670556e-01
1.24023843e+00 1.19654842e-01 6.28118396e-01 1.65134683e-01
1.13637102e+00 6.14412904e-01 5.66115618e-01 4.65265393e-01
4.91043955e-01 5.06499410e-01 2.65092432e-01 -1.37225449e-01
-4.24943417e-01 -6.09301746e-01 3.16860020e-01 9.60038185e-01
-8.18147063e-02 -1.81649014e-01 -5.50235808e-01 7.47763038e-01
-1.88717914e+00 -1.12452090e+00 2.00935081e-01 2.00644946e+00
7.28063524e-01 1.75603181e-02 -1.32856602e-02 1.28450822e-02
9.38810587e-01 6.57822013e-01 -3.91836822e-01 1.41074806e-01
1.81692764e-01 7.56346956e-02 2.34443009e-01 1.01518369e+00
-9.39352810e-01 8.42959583e-01 4.55204296e+00 1.07219887e+00
-1.13021171e+00 2.77233392e-01 7.41695702e-01 -3.58869851e-01
-7.53966510e-01 2.20059350e-01 -2.95444548e-01 8.77175212e-01
3.24035227e-01 -2.16841549e-01 4.18322831e-01 2.21055403e-01
9.34895217e-01 9.72650275e-02 -6.34443879e-01 7.11119592e-01
-2.39013299e-01 -1.63999212e+00 2.64805406e-01 -4.29504067e-02
5.80158472e-01 -3.07387471e-01 1.11965284e-01 5.40752672e-02
-9.56652164e-02 -5.52552104e-01 7.68117249e-01 6.25569463e-01
6.80280149e-01 -8.20576131e-01 1.94420412e-01 2.14159396e-02
-1.70734680e+00 -1.69270471e-01 -1.66872829e-01 -7.14527071e-02
5.01752198e-01 6.62383735e-01 1.94081347e-02 7.24851489e-01
5.85322320e-01 1.30742788e+00 -2.74143517e-01 1.10350263e+00
-2.60398239e-01 4.59563792e-01 1.66312177e-02 5.22972167e-01
3.77472222e-01 -3.84809375e-01 5.90340436e-01 1.02381885e+00
4.40949589e-01 2.82438815e-01 2.06550092e-01 9.18828368e-01
-1.25070840e-01 -7.18275607e-02 -4.17941213e-01 2.73956895e-01
6.88273549e-01 1.25127625e+00 -4.97382879e-01 -3.20046753e-01
-5.49158633e-01 1.22084188e+00 2.52093345e-01 6.83081329e-01
-1.01298380e+00 -4.87935275e-01 8.72050405e-01 2.73523360e-01
4.89502221e-01 -2.08143756e-01 2.20053308e-02 -1.40134001e+00
3.84010762e-01 -3.42361867e-01 2.74155796e-01 -9.45110738e-01
-1.20917428e+00 5.95564604e-01 -1.20117933e-01 -1.49587369e+00
2.54747212e-01 -5.60372025e-02 -1.13600218e+00 9.04169858e-01
-1.88788474e+00 -1.17681849e+00 -5.20447373e-01 8.05067956e-01
6.70196474e-01 3.69893968e-01 -1.07601047e-01 7.17724562e-01
-8.19956601e-01 4.67615008e-01 -2.76902944e-01 3.03911239e-01
8.75116527e-01 -7.52715588e-01 3.00595939e-01 1.08106673e+00
-3.78357351e-01 4.85852003e-01 3.46013784e-01 -4.80538815e-01
-1.08692074e+00 -1.43941104e+00 6.58724427e-01 2.90346202e-02
6.75701559e-01 -1.18935443e-01 -1.06271720e+00 6.68586552e-01
2.25508973e-01 4.80099440e-01 8.06345493e-02 -6.25300765e-01
-3.85466278e-01 -4.17690516e-01 -9.49624062e-01 6.69243157e-01
1.16653514e+00 -5.46703339e-01 -3.37487847e-01 4.68551815e-02
1.30478883e+00 -3.43639910e-01 -6.27299607e-01 2.37578511e-01
4.06495512e-01 -1.04420376e+00 9.26219881e-01 -2.47399226e-01
8.91818523e-01 -6.13973916e-01 2.05279455e-01 -1.19489992e+00
-4.65205967e-01 -9.76760268e-01 -2.00934380e-01 1.47637558e+00
1.12297691e-01 -6.44952297e-01 8.18108678e-01 3.83232772e-01
-4.61188197e-01 -7.92079329e-01 -8.92333567e-01 -4.54131663e-01
2.37534475e-02 -2.39968255e-01 6.86428130e-01 9.96853113e-01
-1.26954421e-01 1.23110361e-01 -7.05354571e-01 4.17064205e-02
4.88926321e-01 1.38594791e-01 3.96650136e-01 -5.81422329e-01
-3.85855824e-01 -5.18186629e-01 -1.47487938e-01 -1.59115934e+00
1.53994337e-01 -6.42324567e-01 -3.59783657e-02 -1.42460406e+00
1.89715430e-01 -3.77337992e-01 -3.27254355e-01 3.20759833e-01
-5.33572674e-01 2.55027026e-01 4.86236155e-01 3.53012025e-01
-4.89169657e-01 9.65892971e-01 1.92189133e+00 -1.99349802e-02
-2.71414429e-01 -2.97501981e-01 -6.01837635e-01 4.23899859e-01
7.86464214e-01 -3.58806282e-01 -7.21097648e-01 -8.45664501e-01
-1.08556576e-01 5.81101656e-01 5.18778265e-01 -9.05510604e-01
4.04953957e-01 -3.85362178e-01 2.26023361e-01 -2.77852565e-01
2.79429227e-01 -5.11608779e-01 -1.09510496e-01 3.42057884e-01
-4.50494826e-01 5.94735183e-02 3.68377157e-02 7.06145287e-01
-3.22697073e-01 2.32571736e-01 7.07464695e-01 1.71780065e-02
-5.07057488e-01 1.07180429e+00 -2.32344747e-01 3.53836417e-01
9.44681108e-01 -3.62326875e-02 -2.37067595e-01 -4.11771864e-01
-6.37896478e-01 5.58881879e-01 4.08814996e-01 4.16427881e-01
7.85761237e-01 -1.56469619e+00 -5.55246592e-01 3.36741596e-01
-7.18089342e-02 1.90613240e-01 6.63445354e-01 1.05817318e+00
-6.69562995e-01 1.44167840e-01 -2.38287345e-01 -4.23039854e-01
-7.39055097e-01 7.03236997e-01 3.84161621e-01 -2.39827082e-01
-7.70183861e-01 7.81586289e-01 7.18181372e-01 -2.34285537e-02
1.93096790e-02 -3.97467315e-01 -9.86699015e-02 -6.15725406e-02
7.18395412e-01 4.61568207e-01 -4.42566305e-01 -5.26803195e-01
-2.51859754e-01 6.86858654e-01 -7.55140046e-03 -3.38202715e-01
1.12768519e+00 -5.37195921e-01 -7.00739622e-02 7.97445923e-02
1.38666868e+00 1.31748065e-01 -1.93908107e+00 -2.64256656e-01
-5.68823993e-01 -8.32324743e-01 5.64019941e-02 -1.82209820e-01
-1.80826187e+00 1.08047020e+00 5.19645363e-02 -5.14385328e-02
1.48212039e+00 -2.55624980e-01 1.35433877e+00 -3.94869596e-01
-1.61548425e-02 -4.07823116e-01 1.63244396e-01 3.08636427e-01
7.52985358e-01 -8.16939533e-01 -1.27019361e-01 -5.66804290e-01
-4.71210241e-01 9.39651966e-01 7.20008373e-01 -4.13033545e-01
6.69188678e-01 9.80373174e-02 -2.46481851e-01 9.62199569e-02
-8.16962242e-01 2.48914547e-02 1.27713457e-01 3.62937510e-01
3.17657948e-01 -2.59695351e-01 -2.98117846e-01 3.14208776e-01
4.03938264e-01 2.61931628e-01 5.46641946e-01 7.60191202e-01
-3.52414221e-01 -1.00281930e+00 -1.44429922e-01 1.88917026e-01
-2.17455089e-01 -2.55783767e-01 2.16207340e-01 4.37631041e-01
2.02792153e-01 7.79388607e-01 1.66641608e-01 -3.30994010e-01
3.15957904e-01 -2.80788541e-01 2.57171243e-01 -2.38815621e-01
-3.97864789e-01 3.27783018e-01 -1.72006011e-01 -9.19209659e-01
-4.33188319e-01 -5.08398831e-01 -1.21308482e+00 -6.39452159e-01
1.96201846e-01 2.54835904e-01 -2.59611070e-01 7.87310660e-01
5.88268876e-01 7.36294270e-01 1.08602726e+00 -1.00813580e+00
1.53999314e-01 -7.80077398e-01 -3.05544883e-01 2.77246177e-01
8.89299631e-01 -5.42862058e-01 -5.99756896e-01 1.17918178e-01] | [10.72584342956543, -1.4096647500991821] |
83e1214f-b0a9-428e-8509-b4dff1556633 | youtube-vos-sequence-to-sequence-video-object | 1809.00461 | null | http://arxiv.org/abs/1809.00461v1 | http://arxiv.org/pdf/1809.00461v1.pdf | YouTube-VOS: Sequence-to-Sequence Video Object Segmentation | Learning long-term spatial-temporal features are critical for many video
analysis tasks. However, existing video segmentation methods predominantly rely
on static image segmentation techniques, and methods capturing temporal
dependency for segmentation have to depend on pretrained optical flow models,
leading to suboptimal solutions for the problem. End-to-end sequential learning
to explore spatial-temporal features for video segmentation is largely limited
by the scale of available video segmentation datasets, i.e., even the largest
video segmentation dataset only contains 90 short video clips. To solve this
problem, we build a new large-scale video object segmentation dataset called
YouTube Video Object Segmentation dataset (YouTube-VOS). Our dataset contains
3,252 YouTube video clips and 78 categories including common objects and human
activities. This is by far the largest video object segmentation dataset to our
knowledge and we have released it at https://youtube-vos.org. Based on this
dataset, we propose a novel sequence-to-sequence network to fully exploit
long-term spatial-temporal information in videos for segmentation. We
demonstrate that our method is able to achieve the best results on our
YouTube-VOS test set and comparable results on DAVIS 2016 compared to the
current state-of-the-art methods. Experiments show that the large scale dataset
is indeed a key factor to the success of our model. | ['Yuchen Liang', 'Jianchao Yang', 'Brian Price', 'Linjie Yang', 'Dingcheng Yue', 'Scott Cohen', 'Ning Xu', 'Yuchen Fan', 'Thomas Huang'] | 2018-09-03 | youtube-vos-sequence-to-sequence-video-object-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Ning_Xu_YouTube-VOS_Sequence-to-Sequence_Video_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Ning_Xu_YouTube-VOS_Sequence-to-Sequence_Video_ECCV_2018_paper.pdf | eccv-2018-9 | ['one-shot-visual-object-segmentation'] | ['computer-vision'] | [ 1.80854984e-02 -4.71314818e-01 -7.60351002e-01 -1.23899877e-01
-6.47893667e-01 -7.08297968e-01 2.55903602e-01 -4.05032635e-01
-6.42590880e-01 4.26965237e-01 6.32853247e-03 -1.97042763e-01
1.65996999e-01 -2.80932456e-01 -9.05481577e-01 -4.07779902e-01
-3.06205124e-01 7.09171370e-02 9.10781741e-01 1.11593716e-01
3.32122266e-01 2.37485647e-01 -1.48382723e+00 2.35494122e-01
7.49244988e-01 1.29502773e+00 4.20784831e-01 9.22159970e-01
-8.30560923e-02 8.54762793e-01 -1.98877871e-01 -2.03105025e-02
6.32231355e-01 -4.57511216e-01 -1.02372599e+00 2.63960779e-01
7.10160971e-01 -6.69176400e-01 -6.65981650e-01 8.24181318e-01
2.31938988e-01 2.47321025e-01 3.66021544e-01 -1.41393578e+00
-3.43084008e-01 5.15884817e-01 -5.37843108e-01 7.37681925e-01
4.46176410e-01 3.93775314e-01 1.05278075e+00 -6.20814860e-01
9.89240229e-01 7.89747834e-01 6.63583815e-01 6.32229984e-01
-7.38672316e-01 -6.46806121e-01 4.01235491e-01 5.95056593e-01
-1.24681997e+00 -3.14047277e-01 6.59344375e-01 -6.31199718e-01
8.08866322e-01 6.75085187e-02 1.04160988e+00 1.21594834e+00
-1.44202158e-01 1.31952131e+00 6.22144282e-01 2.03532297e-02
1.35496110e-01 -3.96795958e-01 1.37555212e-01 7.08502471e-01
-2.72214502e-01 -1.17036112e-01 -5.29797435e-01 3.59039694e-01
8.66135061e-01 1.58413380e-01 -3.56732726e-01 -4.35055405e-01
-1.27034569e+00 6.99193716e-01 4.00815815e-01 3.65790695e-01
-2.70875782e-01 2.58876950e-01 6.38252974e-01 1.93528339e-01
3.55975240e-01 1.94987833e-01 -6.21402800e-01 -7.44826853e-01
-1.50130427e+00 2.30708525e-01 7.59275675e-01 9.05520380e-01
7.70672202e-01 -6.32571464e-04 -1.76817909e-01 6.08928144e-01
1.46063000e-01 4.87913251e-01 4.43680018e-01 -1.44232976e+00
5.13873339e-01 3.51877570e-01 -9.19164941e-02 -8.15487444e-01
-2.57954866e-01 2.30963588e-01 -4.40675557e-01 -2.41796508e-01
6.17060661e-01 -2.14538768e-01 -1.15826952e+00 1.51861882e+00
2.59431928e-01 8.10740232e-01 -1.93928093e-01 1.17868185e+00
1.06492007e+00 7.99433589e-01 1.36176288e-01 -2.11196020e-01
1.01876450e+00 -1.43480122e+00 -4.54322994e-01 -7.77738839e-02
5.97261429e-01 -5.52654088e-01 9.88785148e-01 3.59811544e-01
-9.97099519e-01 -6.48135126e-01 -8.12107384e-01 -1.00402050e-01
-2.09388047e-01 -1.23752147e-01 6.39483809e-01 5.05318105e-01
-8.40276361e-01 5.84699333e-01 -9.89653289e-01 -4.74748939e-01
8.48314404e-01 2.31415257e-01 -2.52430975e-01 -1.29450560e-01
-1.10440040e+00 1.98829591e-01 4.57259029e-01 1.76440656e-01
-1.13927794e+00 -7.58802116e-01 -7.70681441e-01 -9.96861756e-02
9.66118276e-01 -3.28470260e-01 1.19665670e+00 -1.32966316e+00
-1.31564283e+00 5.35799563e-01 -2.95425802e-01 -6.58185184e-01
7.60281444e-01 -3.52286518e-01 -6.21525571e-02 8.05105209e-01
2.50375897e-01 1.06459773e+00 8.85304332e-01 -1.00950062e+00
-7.94155777e-01 1.66787561e-02 1.86520845e-01 1.46229323e-02
-4.81570840e-01 1.42634153e-01 -1.17422259e+00 -5.86176395e-01
-2.24477276e-01 -9.71597910e-01 -2.60447115e-01 -8.64837542e-02
-3.16588461e-01 -2.52181232e-01 1.26222467e+00 -7.08650231e-01
1.24794447e+00 -2.10394883e+00 2.48478100e-01 -5.21236844e-02
2.82505721e-01 6.01750076e-01 -3.82096201e-01 7.76975453e-02
1.67061105e-01 3.33918035e-01 -1.74300551e-01 -2.28257328e-01
-3.00216228e-01 1.29169002e-01 -2.92304367e-01 4.33304042e-01
-8.85185972e-02 1.17408502e+00 -9.27823484e-01 -7.49766946e-01
3.92605871e-01 3.37213010e-01 -6.65556431e-01 2.44270653e-01
-4.30284053e-01 5.51949859e-01 -5.78982711e-01 9.00946856e-01
4.37637150e-01 -2.14892834e-01 -1.56909198e-01 -6.14383370e-02
-1.26749396e-01 -1.52982518e-01 -8.58979940e-01 2.00495338e+00
-8.93845856e-02 9.15651441e-01 -8.14844817e-02 -1.12831545e+00
2.89826423e-01 1.87118337e-01 1.17686081e+00 -5.60896277e-01
1.32703736e-01 1.99329302e-01 -2.31334731e-01 -8.28388989e-01
4.44976181e-01 4.26186413e-01 1.17746349e-02 6.29433244e-02
1.59390688e-01 6.07105456e-02 8.49830627e-01 4.24113363e-01
1.18316448e+00 4.40617681e-01 -9.45646316e-02 -6.59243017e-02
5.01387119e-01 1.21862173e-01 8.37913215e-01 7.55074918e-01
-7.14158356e-01 9.42526281e-01 7.77288258e-01 -5.80350339e-01
-1.01684904e+00 -9.22982395e-01 9.23589915e-02 9.85221684e-01
4.08881336e-01 -7.39293456e-01 -9.94068563e-01 -9.78463709e-01
-2.31235027e-01 9.25079808e-02 -5.46148419e-01 3.41158509e-01
-8.34683359e-01 -2.65377045e-01 5.93827724e-01 6.92564905e-01
7.14101553e-01 -1.35689020e+00 -7.16812670e-01 1.74977988e-01
-3.57623875e-01 -1.83484066e+00 -8.61204565e-01 -3.44305336e-01
-9.27717447e-01 -1.54686666e+00 -9.38458502e-01 -6.98554456e-01
2.31326520e-01 4.92053509e-01 1.05024397e+00 5.55660799e-02
-4.85871077e-01 5.57749569e-01 -6.26045585e-01 1.45849988e-01
-1.18482955e-01 3.69644225e-01 -2.53783613e-01 -8.51247758e-02
3.57671648e-01 -3.42699379e-01 -8.54949415e-01 5.36659598e-01
-9.59792972e-01 1.44435102e-02 3.29454750e-01 5.05781293e-01
5.98559797e-01 -1.70830682e-01 3.47768694e-01 -7.21108854e-01
-2.03288738e-02 -5.94043732e-01 -6.59147024e-01 6.54773936e-02
-2.13485628e-01 -5.34074724e-01 6.00125670e-01 -4.49665189e-01
-6.01014674e-01 2.00240865e-01 -1.84381217e-01 -9.76276040e-01
-2.69173175e-01 5.87445259e-01 6.63527921e-02 -5.98162003e-02
1.09332405e-01 3.04920405e-01 -1.47003248e-01 -3.44129652e-01
2.17395529e-01 4.91751730e-01 5.34216762e-01 -3.08219463e-01
4.44639295e-01 4.00553912e-01 -1.28847614e-01 -1.16765916e+00
-8.07225823e-01 -9.61332798e-01 -8.88480365e-01 -4.83326823e-01
1.38643479e+00 -1.07101488e+00 -6.05643272e-01 6.37932122e-01
-9.56452072e-01 -8.42036784e-01 -1.77046284e-01 3.81410599e-01
-6.75827384e-01 6.60030961e-01 -7.54255652e-01 -3.71366054e-01
-1.03081129e-01 -1.38863575e+00 1.05690324e+00 1.55580610e-01
4.51526865e-02 -9.50918138e-01 2.98128892e-02 6.12846911e-01
1.29445344e-01 2.79981971e-01 1.64078221e-01 -2.80935705e-01
-9.69217896e-01 4.17978317e-02 -4.10359740e-01 5.10675669e-01
-1.20141581e-01 3.47695112e-01 -4.86317545e-01 -2.25935027e-01
-2.65324682e-01 -4.99892175e-01 1.40484679e+00 6.25760496e-01
1.49188507e+00 1.12255616e-02 -2.46048138e-01 8.87238622e-01
1.32823718e+00 8.54153112e-02 7.21145749e-01 3.25155020e-01
1.33606064e+00 3.28412026e-01 8.22006404e-01 3.45982075e-01
3.74427944e-01 7.34196484e-01 3.39272171e-01 7.00241402e-02
-2.80359864e-01 -1.27799749e-01 5.86725414e-01 7.28079379e-01
-2.72584856e-01 -3.36104244e-01 -1.07781756e+00 7.30374038e-01
-1.99178302e+00 -1.13564265e+00 -6.32988513e-02 1.79257572e+00
5.86899221e-01 1.24025218e-01 4.05888200e-01 -1.10079385e-01
6.27592504e-01 5.48916638e-01 -6.60470426e-01 7.34308586e-02
3.09517756e-02 -4.66568395e-02 6.34372354e-01 2.72571474e-01
-1.62295604e+00 1.32498729e+00 5.84281063e+00 9.97486949e-01
-1.24691617e+00 2.33984679e-01 5.40968239e-01 -4.29404467e-01
9.84172672e-02 -1.30062610e-01 -8.00728381e-01 6.94968998e-01
9.68194008e-01 1.25765502e-01 3.20174932e-01 8.66586208e-01
3.43510240e-01 -2.16795236e-01 -9.28084016e-01 1.15706205e+00
2.25309297e-01 -1.39575577e+00 -1.03045262e-01 3.41095552e-02
9.94127691e-01 4.87812281e-01 -1.12457648e-01 3.18698257e-01
-2.18494639e-01 -8.63696754e-01 8.02232683e-01 2.16851905e-01
7.40706980e-01 -6.35178387e-01 5.52707434e-01 2.22238023e-02
-1.44741929e+00 -2.25909576e-02 -1.04911901e-01 2.16989443e-01
4.98522013e-01 4.28063989e-01 -3.22831869e-01 2.32624680e-01
9.73265707e-01 1.41606843e+00 -7.00350344e-01 1.32942295e+00
-5.94766662e-02 9.05213833e-01 -3.46667260e-01 2.17207581e-01
7.37189293e-01 -1.25900820e-01 5.77708125e-01 1.24127781e+00
2.75860578e-01 3.32984179e-02 7.22270668e-01 4.18415099e-01
-3.90237808e-01 2.51420792e-02 -4.15108562e-01 -4.14638072e-01
5.22206351e-02 1.19715858e+00 -1.18982708e+00 -5.22510290e-01
-7.63007522e-01 1.05260384e+00 4.48284410e-02 5.66007972e-01
-1.16456449e+00 -6.06228225e-02 7.06759214e-01 1.19750284e-01
9.11217272e-01 -4.72674847e-01 7.16257617e-02 -1.50661278e+00
-4.85247709e-02 -8.13013911e-01 4.27369267e-01 -4.29506004e-01
-9.92054284e-01 4.49387550e-01 5.95232397e-02 -1.19985676e+00
-2.25369766e-01 -4.55248356e-01 -4.28215921e-01 1.66565046e-01
-1.59965551e+00 -1.11789227e+00 -4.65623885e-01 8.73900473e-01
1.13033259e+00 -2.05008805e-01 2.29480848e-01 3.92369300e-01
-6.04452789e-01 2.99496055e-01 8.20551291e-02 5.11346340e-01
6.41374707e-01 -1.04543686e+00 3.91662031e-01 1.07570612e+00
3.91633511e-01 2.21059486e-01 3.88234079e-01 -7.05179334e-01
-1.48501027e+00 -1.08619714e+00 3.72740179e-01 -4.24104095e-01
7.06076860e-01 -4.65243012e-01 -7.82359600e-01 7.83938527e-01
2.81081945e-01 5.21277547e-01 5.66029370e-01 -1.92873582e-01
-2.44893268e-01 -1.01938033e-02 -7.39859641e-01 5.48732042e-01
1.27989376e+00 -4.13476527e-01 -2.81287879e-01 2.78619766e-01
7.29605675e-01 -5.31403959e-01 -7.44870365e-01 4.47378129e-01
6.61878467e-01 -1.02942431e+00 9.48602378e-01 -5.07923961e-01
7.92264044e-01 -2.53699392e-01 -2.91018281e-02 -8.29360545e-01
9.46989954e-02 -7.18797565e-01 -3.56939167e-01 1.06558430e+00
1.30621150e-01 -1.51932329e-01 1.23203897e+00 5.33635020e-01
-1.59193084e-01 -8.53494823e-01 -9.40530360e-01 -8.64068389e-01
-7.09728524e-02 -7.63327718e-01 1.19560912e-01 7.63993084e-01
-3.98438960e-01 -9.02925357e-02 -5.03404081e-01 -2.59068787e-01
5.59035242e-01 3.55027020e-01 8.82658601e-01 -9.14262414e-01
-1.79131925e-01 -5.92997491e-01 -5.96968055e-01 -1.40962219e+00
4.64706779e-01 -7.40766466e-01 1.45352855e-01 -1.37133586e+00
3.69884342e-01 -2.71549821e-01 -2.41411477e-01 3.04545909e-01
-1.80593029e-01 6.96200252e-01 5.60128033e-01 3.46396446e-01
-1.28645742e+00 4.38577801e-01 1.34179199e+00 -1.18167557e-01
-3.92968953e-01 -1.60580963e-01 -2.10781291e-01 7.43623078e-01
7.20776737e-01 -3.04366738e-01 -4.01718497e-01 -3.09302837e-01
-1.98027760e-01 2.28580832e-02 3.88627440e-01 -9.43713903e-01
2.03508377e-01 -3.67738664e-01 2.73243129e-01 -6.19576573e-01
1.51497424e-01 -5.88622153e-01 -2.07664803e-01 4.08578604e-01
-1.80884391e-01 -2.67275095e-01 1.06619179e-01 6.26563668e-01
-4.54603493e-01 -1.11565262e-01 6.13774657e-01 -2.36948833e-01
-1.56941438e+00 8.56496394e-01 -5.43732882e-01 4.91275996e-01
1.27702129e+00 -3.46514910e-01 2.10934654e-02 -1.74248934e-01
-6.03236318e-01 4.17654991e-01 4.49217469e-01 7.20663249e-01
6.92496777e-01 -1.12079620e+00 -5.33634126e-01 6.69912845e-02
-5.23537397e-02 5.36309406e-02 4.33791071e-01 9.74767089e-01
-7.99830556e-01 5.26583731e-01 -2.84633785e-01 -9.45499837e-01
-1.40132344e+00 6.12311304e-01 -9.15891603e-02 -1.96572803e-02
-7.52640426e-01 8.16905975e-01 1.49047717e-01 1.10326596e-01
2.78875798e-01 -4.00377989e-01 -2.74428040e-01 2.66044348e-01
4.65933740e-01 4.09455538e-01 -3.97140056e-01 -9.02632892e-01
-3.10533941e-01 8.33713591e-01 -1.26266986e-01 8.81557390e-02
1.20152330e+00 -3.54339302e-01 1.07766443e-03 4.54632759e-01
1.62883079e+00 -2.23158568e-01 -1.72607589e+00 5.18014245e-02
-1.80995375e-01 -9.14154172e-01 3.37045565e-02 -3.29492450e-01
-1.66588128e+00 8.23089182e-01 3.98540676e-01 8.22883248e-02
1.07879090e+00 -1.52221192e-02 1.50696778e+00 3.80158693e-01
3.68266195e-01 -1.18475664e+00 3.52663070e-01 6.48497760e-01
4.99078780e-01 -1.58369172e+00 6.61412813e-03 -5.97113192e-01
-7.81915605e-01 1.15892220e+00 6.14748120e-01 -2.70868719e-01
6.42759085e-01 3.92264202e-02 1.25851169e-01 3.42019573e-02
-4.83087689e-01 -3.61931801e-01 3.42306584e-01 3.09437364e-01
2.28551179e-01 -2.47786477e-01 -2.98124671e-01 3.73383105e-01
8.40296894e-02 3.71328980e-01 4.94112283e-01 7.80034840e-01
-1.47704452e-01 -9.79283690e-01 2.23203674e-02 3.42172444e-01
-7.14370489e-01 5.87683544e-02 -1.27687380e-01 7.53236830e-01
1.12743720e-01 8.50190043e-01 -1.33709470e-02 -2.67535925e-01
-2.44225934e-02 -1.69717744e-01 4.34420794e-01 -3.11895698e-01
-3.85110915e-01 1.92941234e-01 -1.69271782e-01 -1.22440064e+00
-8.41485977e-01 -9.03790236e-01 -1.34022319e+00 -2.09336102e-01
5.12551777e-02 -6.83109835e-02 5.02159595e-01 8.60651374e-01
2.10083336e-01 4.37031865e-01 4.62588519e-01 -1.33386493e+00
9.18786228e-02 -5.87315023e-01 -4.22595024e-01 6.52142167e-01
3.63682330e-01 -5.50281823e-01 -3.30866575e-01 3.74465823e-01] | [9.184981346130371, 0.06069455295801163] |
07594955-7fbc-4e66-ab65-9df45f2279de | framm-fair-ranking-with-missing-modalities | 2305.19407 | null | https://arxiv.org/abs/2305.19407v1 | https://arxiv.org/pdf/2305.19407v1.pdf | FRAMM: Fair Ranking with Missing Modalities for Clinical Trial Site Selection | Despite many efforts to address the disparities, the underrepresentation of gender, racial, and ethnic minorities in clinical trials remains a problem and undermines the efficacy of treatments on minorities. This paper focuses on the trial site selection task and proposes FRAMM, a deep reinforcement learning framework for fair trial site selection. We focus on addressing two real-world challenges that affect fair trial sites selection: the data modalities are often not complete for many potential trial sites, and the site selection needs to simultaneously optimize for both enrollment and diversity since the problem is necessarily a trade-off between the two with the only possible way to increase diversity post-selection being through limiting enrollment via caps. To address the missing data challenge, FRAMM has a modality encoder with a masked cross-attention mechanism for handling missing data, bypassing data imputation and the need for complete data in training. To handle the need for making efficient trade-offs, FRAMM uses deep reinforcement learning with a specifically designed reward function that simultaneously optimizes for both enrollment and fairness. We evaluate FRAMM using 4,392 real-world clinical trials ranging from 2016 to 2021 and show that FRAMM outperforms the leading baseline in enrollment-only settings while also achieving large gains in diversity. Specifically, it is able to produce a 9% improvement in diversity with similar enrollment levels over the leading baselines. That improved diversity is further manifested in achieving up to a 14% increase in Hispanic enrollment, 27% increase in Black enrollment, and 60% increase in Asian enrollment compared to selecting sites with an enrollment-only model. | ['Jimeng Sun', 'Cao Xiao', 'Lucas Glass', 'Brandon Theodorou'] | 2023-05-30 | null | null | null | null | ['imputation', 'imputation', 'imputation'] | ['computer-vision', 'miscellaneous', 'time-series'] | [ 5.67076430e-02 3.33458424e-01 -1.13642478e+00 -5.78776896e-01
-1.40790915e+00 -3.53395313e-01 3.17221314e-01 5.39504945e-01
-7.29192972e-01 1.02878714e+00 9.45941746e-01 -8.71738195e-01
-1.22373961e-01 -6.73746288e-01 -8.36731374e-01 -3.94602895e-01
1.41078457e-01 5.74269056e-01 -9.20641482e-01 1.10446267e-01
1.74223602e-01 1.71279520e-01 -9.97778535e-01 3.22039127e-01
1.17211080e+00 5.76406598e-01 -4.88321841e-01 8.77435952e-02
1.08370297e-01 6.56931937e-01 -4.42642361e-01 -2.94541121e-01
5.55213153e-01 -5.36470234e-01 -4.12673861e-01 -1.34614632e-01
7.54666269e-01 -8.29900622e-01 -2.18913823e-01 5.67836404e-01
1.14412391e+00 -4.52873521e-02 6.35243058e-01 -1.18633795e+00
-4.97865468e-01 9.58872795e-01 -8.72380316e-01 -2.49460563e-01
7.32715577e-02 5.79353094e-01 9.97890592e-01 -4.61229801e-01
7.08323836e-01 1.20433402e+00 7.88680613e-01 7.38531470e-01
-1.51495397e+00 -9.45057094e-01 1.44187128e-02 -5.49872994e-01
-1.01455688e+00 -6.30140245e-01 8.47814754e-02 -6.34159982e-01
5.33601463e-01 3.02623153e-01 5.97969830e-01 1.15171504e+00
1.51979685e-01 4.09226298e-01 1.15662992e+00 -6.96245432e-02
6.37340605e-01 -1.04129635e-01 2.09263518e-01 1.79528818e-01
3.38440150e-01 4.19209599e-01 -2.90677011e-01 -6.72379434e-01
9.73172545e-01 2.79059350e-01 -1.59010455e-01 -2.85501182e-01
-1.08797407e+00 1.16148114e+00 5.10718405e-01 -2.51433074e-01
-5.77050209e-01 8.45877603e-02 3.23485643e-01 1.51483178e-01
-1.14149630e-01 6.34307683e-01 -2.23308012e-01 -2.80452874e-02
-1.26260662e+00 4.13658410e-01 5.33169031e-01 5.34327209e-01
2.45900303e-01 2.48986762e-02 -8.32575500e-01 5.83941698e-01
-7.43461773e-03 5.95134139e-01 2.55613387e-01 -1.24846864e+00
6.27656996e-01 7.08107173e-01 2.29973108e-01 -3.23122770e-01
-6.22230411e-01 -6.09417737e-01 -9.37892079e-01 5.19705832e-01
6.34735525e-01 -7.52105236e-01 -1.17151880e+00 2.30529881e+00
2.81739563e-01 -3.03361088e-01 -6.81649223e-02 6.49911106e-01
6.21859908e-01 1.48670465e-01 8.32595944e-01 -2.77709812e-01
1.23595822e+00 -4.56033558e-01 -6.67226255e-01 -1.37305334e-01
7.42491245e-01 -4.92691875e-01 9.08467889e-01 -1.26784056e-01
-1.36831272e+00 4.61775661e-02 -6.81572139e-01 9.32278261e-02
7.75357783e-02 -8.45757276e-02 4.76675957e-01 8.22635710e-01
-7.64804006e-01 4.54805106e-01 -6.16342604e-01 -1.58043176e-01
1.24766505e+00 6.44176781e-01 -3.62473249e-01 -1.59261093e-01
-9.51020658e-01 7.23335564e-01 -1.28713742e-01 -1.04805067e-01
-8.39205325e-01 -1.59242272e+00 -7.95168340e-01 4.07904088e-01
3.34072500e-01 -1.10417724e+00 8.56315494e-01 -9.54956830e-01
-9.39435422e-01 5.60023844e-01 5.73253743e-02 -3.04857999e-01
9.08193827e-01 1.45160988e-01 -1.27645850e-01 -4.78016973e-01
4.59075928e-01 9.02034700e-01 2.11602449e-01 -7.77294219e-01
-5.02471507e-01 -5.94022691e-01 -2.54504114e-01 6.13151252e-01
-2.62347817e-01 -6.55972660e-02 2.17426240e-01 -6.78587914e-01
-5.15174627e-01 -8.72634768e-01 -7.13823795e-01 4.72407648e-03
-3.25798035e-01 3.67820151e-02 3.01302201e-03 -7.92195857e-01
9.73820984e-01 -2.14574647e+00 -2.65327245e-01 2.50290751e-01
3.95128727e-01 -8.46873410e-03 -4.65809375e-01 1.39278412e-01
-2.86124349e-01 5.06524920e-01 -9.01282877e-02 -1.03768036e-01
-1.59054965e-01 -2.11094454e-01 -5.85844256e-02 3.90295029e-01
1.10833742e-01 9.48143840e-01 -5.16453743e-01 -1.73100561e-01
2.05172859e-02 3.22561920e-01 -1.09919477e+00 -1.18000709e-01
1.42203674e-01 5.56425273e-01 -2.13330954e-01 8.07195961e-01
1.02406073e+00 -2.47067437e-01 5.02921164e-01 3.30450058e-01
-5.80196083e-02 2.00401619e-01 -1.23119581e+00 1.35921168e+00
-1.83827966e-01 9.65630338e-02 1.06476061e-01 -5.62145233e-01
5.19959927e-01 1.78663835e-01 7.83067286e-01 -8.59875321e-01
7.59334713e-02 6.41300380e-01 3.02664518e-01 -7.78932199e-02
2.05468968e-01 -5.27414978e-01 -1.84366807e-01 4.97896045e-01
-3.86032939e-01 4.05503690e-01 -2.80670643e-01 8.71698782e-02
1.01501966e+00 -3.41649264e-01 7.07770362e-02 -5.22134066e-01
-1.76607579e-01 1.90556183e-01 1.18458951e+00 1.03658676e+00
-4.11411196e-01 6.87096179e-01 9.45820332e-01 -4.57942277e-01
-1.08249962e+00 -1.07965577e+00 -3.35269034e-01 8.43114674e-01
-4.16131675e-01 2.88774908e-01 -3.97611558e-01 -7.41745532e-01
4.82730627e-01 7.16105402e-01 -7.98651040e-01 -1.24822140e-01
-5.14893234e-01 -1.25964546e+00 5.40726721e-01 5.21237850e-01
2.24950686e-01 -7.87075877e-01 -5.17320216e-01 1.58539444e-01
-6.78136498e-02 -6.09182417e-01 -9.07846689e-01 1.59702465e-01
-9.70198274e-01 -1.11517096e+00 -1.12221026e+00 -5.30802965e-01
6.43587410e-01 -2.49006316e-01 1.07800817e+00 -1.65617779e-01
-3.22514743e-01 -1.27060488e-01 1.66613296e-01 -5.50427079e-01
-3.25763375e-01 2.06894845e-01 -8.31421986e-02 -4.04286206e-01
2.83783495e-01 -3.26138586e-01 -1.12177539e+00 3.50808129e-02
-7.80153811e-01 -3.15629333e-01 4.64665681e-01 1.15889192e+00
3.44876081e-01 -6.18959844e-01 1.15259540e+00 -1.13767767e+00
4.27331060e-01 -5.80557287e-01 -5.46547115e-01 1.61534607e-01
-9.25606787e-01 4.13665362e-02 2.39500448e-01 -3.94822150e-01
-7.60587633e-01 -2.35045198e-02 5.72773293e-02 3.42693017e-03
1.75344080e-01 5.51992834e-01 -1.52204290e-01 1.39765337e-01
8.01265001e-01 -4.21578407e-01 5.78246653e-01 -1.62111700e-01
6.17490150e-02 6.93478525e-01 -1.40580267e-01 -4.49523628e-01
1.96144819e-01 3.02621365e-01 -1.65440202e-01 -1.55127212e-01
-3.43504608e-01 1.09801419e-01 2.82299042e-01 3.44483793e-01
6.27676249e-01 -1.46115589e+00 -1.16497648e+00 2.42427230e-01
-4.29081470e-01 -8.03528547e-01 -6.81135297e-01 8.15232813e-01
-3.20755243e-01 -1.18395150e-01 -5.29232383e-01 -5.11413515e-01
-5.08453965e-01 -1.47622621e+00 6.99373364e-01 3.23846132e-01
-3.75420421e-01 -6.61356807e-01 1.67537764e-01 5.60396969e-01
7.11172342e-01 5.02295434e-01 1.41028011e+00 -6.07030869e-01
-5.34482658e-01 -2.24832282e-03 -2.17498675e-01 -1.86925590e-01
2.64981478e-01 -2.07637876e-01 -5.71233630e-01 -5.05926907e-01
-3.65028888e-01 -4.40786779e-01 7.38619149e-01 1.29142678e+00
1.08388805e+00 -1.05918244e-01 -1.13441043e-01 7.01400399e-01
1.37701392e+00 3.02891672e-01 6.11147225e-01 2.69779861e-01
4.88604724e-01 5.44783175e-01 2.27659732e-01 8.09511125e-01
6.61285162e-01 5.25485218e-01 4.09520715e-01 -5.81198633e-01
-1.54259741e-01 -3.22977841e-01 6.43877760e-02 -5.79669997e-02
3.14966053e-01 1.10805318e-01 -9.08989966e-01 5.54753006e-01
-1.80659604e+00 -8.76245916e-01 4.88174111e-02 2.90384316e+00
1.15855587e+00 -1.75193369e-01 5.27292907e-01 -5.64937294e-01
7.99299479e-01 -1.22118682e-01 -1.04342163e+00 -6.81703389e-01
-2.54540235e-01 2.54583985e-01 8.84497046e-01 5.94012678e-01
-8.50957990e-01 3.81894022e-01 7.11609077e+00 5.55881202e-01
-1.13753057e+00 -1.72043875e-01 1.46493804e+00 -6.16125941e-01
-5.96939445e-01 2.57234871e-02 -6.30557895e-01 4.15603012e-01
8.86795163e-01 -1.05702110e-01 3.45690131e-01 2.96986878e-01
5.31893611e-01 -7.89651647e-02 -1.15306580e+00 7.54820526e-01
-3.78348947e-01 -1.63930643e+00 -1.76716410e-02 5.13580561e-01
1.10030115e+00 -7.62716606e-02 4.55917209e-01 5.13582647e-01
7.20085740e-01 -1.67501903e+00 3.70043933e-01 2.34286457e-01
1.12505555e+00 -8.43904614e-01 8.13577116e-01 6.00810274e-02
-2.73296386e-01 -3.09685498e-01 1.40437894e-02 -3.70132737e-03
-6.43192530e-02 8.09317827e-01 -5.80236614e-01 3.89739603e-01
3.91732782e-01 3.28801751e-01 -1.99830577e-01 1.21279728e+00
3.50043654e-01 4.54364926e-01 -3.27797264e-01 4.06083405e-01
2.23102197e-01 5.35753332e-02 8.23329017e-02 5.50047636e-01
4.01290685e-01 -1.57359526e-01 2.23718360e-01 7.78831363e-01
-6.31003022e-01 2.14600489e-01 -3.39257181e-01 -8.23934004e-02
6.53915167e-01 8.90065193e-01 -7.25329593e-02 -1.47651166e-01
-1.89995214e-01 2.44846836e-01 1.56971768e-01 4.57684845e-01
-8.07177782e-01 -1.13254003e-01 7.96560407e-01 4.14673626e-01
-1.51897773e-01 4.09621954e-01 -9.77124989e-01 -9.65658009e-01
-1.88776448e-01 -1.49655414e+00 8.09839129e-01 -1.77413836e-01
-1.28390026e+00 1.51083142e-01 -3.95788431e-01 -9.97271419e-01
-4.23164777e-02 -1.71071127e-01 -2.20094934e-01 1.54648674e+00
-1.42165959e+00 -1.06282365e+00 1.46601483e-01 4.65478420e-01
6.27758279e-02 -1.55374199e-01 7.30754137e-01 6.87129259e-01
-7.89977968e-01 1.14944005e+00 1.08270481e-01 9.38570797e-02
1.23353291e+00 -1.04159939e+00 -6.53770939e-02 2.41061136e-01
-6.42967284e-01 7.11354256e-01 1.48047507e-01 -8.90815675e-01
-1.13469708e+00 -8.43331695e-01 1.02363014e+00 -4.74520743e-01
8.10883790e-02 -9.00577158e-02 -4.97781515e-01 7.03290641e-01
1.37035817e-01 -2.82313287e-01 9.68618155e-01 6.17671371e-01
-4.30310100e-01 -1.29100129e-01 -1.39071262e+00 9.59015012e-01
6.34542406e-01 -2.32040182e-01 4.55548242e-02 2.38860190e-01
4.45672423e-01 -6.68788135e-01 -1.07523561e+00 4.63112086e-01
5.45746207e-01 -5.24928689e-01 8.82573903e-01 -9.61258590e-01
8.45212340e-01 1.72476679e-01 -7.89916888e-02 -1.12450683e+00
-2.83602208e-01 -6.05867445e-01 5.98138332e-01 9.86248314e-01
7.94569075e-01 -5.34546554e-01 1.30426061e+00 1.46102417e+00
1.50322124e-01 -8.56739998e-01 -1.08226371e+00 -3.17915499e-01
8.48314762e-01 1.93544164e-01 8.25002730e-01 1.30025899e+00
-6.92420453e-02 1.19976223e-01 -5.09922266e-01 -5.21862023e-02
6.16068065e-01 1.89196989e-01 4.42790776e-01 -7.96408057e-01
-2.28723049e-01 -7.44566441e-01 1.63037926e-01 -3.40094477e-01
-1.40888005e-01 -9.47720468e-01 -3.39569777e-01 -1.57962513e+00
9.27557886e-01 -9.53635573e-01 -3.31833065e-01 9.37484980e-01
-5.07251740e-01 -1.55647501e-01 3.83491397e-01 5.83666861e-02
-4.08500852e-03 4.38985348e-01 1.17974675e+00 -2.24112958e-01
-6.95328772e-01 3.90172489e-02 -1.45854807e+00 -6.71460293e-03
7.04778910e-01 -6.34893477e-01 -1.78285435e-01 -5.38056016e-01
9.85352397e-02 5.56618273e-01 2.81375766e-01 -6.32110655e-01
-3.67085226e-02 -5.28325260e-01 5.86120903e-01 -1.90506920e-01
-2.08299711e-01 -7.38653958e-01 3.71503741e-01 9.19434965e-01
-7.94269443e-01 1.83748662e-01 3.37791950e-01 2.96440333e-01
1.04332820e-01 4.16752845e-02 7.08302498e-01 1.29218996e-01
1.35686159e-01 4.02945817e-01 -2.51639277e-01 4.47720617e-01
8.00857663e-01 -4.76766936e-03 -6.00076735e-01 -3.88352334e-01
-7.09637880e-01 7.37764895e-01 5.35897851e-01 1.54225618e-01
3.19593512e-02 -1.42884481e+00 -1.13630319e+00 1.45764977e-01
-1.08866811e-01 -1.63745314e-01 6.97178543e-01 8.30643892e-01
-3.63737762e-01 -6.68032244e-02 -4.89177257e-01 -2.11708039e-01
-9.44236338e-01 2.75070339e-01 4.48372483e-01 -5.73655903e-01
-2.02825442e-01 3.73854101e-01 2.44334191e-01 -7.23686755e-01
3.41843575e-01 8.03624019e-02 1.25660554e-01 1.17708080e-01
4.39721882e-01 4.80603427e-01 -5.45700686e-03 -4.34393995e-02
-5.11990011e-01 2.63243645e-01 -1.80806264e-01 -2.01919138e-01
1.25665402e+00 1.37604460e-01 1.70441598e-01 -8.13842863e-02
1.07878423e+00 1.07824340e-01 -1.39375699e+00 8.63937289e-02
-3.59923542e-01 -3.93843532e-01 1.70892745e-01 -1.35606086e+00
-1.19855642e+00 7.04347253e-01 9.02652204e-01 -5.57180941e-01
9.55735743e-01 -5.61547518e-01 5.39411902e-01 -3.66772592e-01
7.29875732e-03 -1.04843950e+00 -2.47173205e-01 -2.62489859e-02
5.16509235e-01 -1.67919266e+00 2.24709481e-01 9.10128951e-02
-9.49682891e-01 5.97800970e-01 4.95108068e-01 1.24837942e-01
5.07306993e-01 1.55481368e-01 1.30268723e-01 -9.17569548e-03
-5.01329601e-01 2.91383207e-01 -1.46980867e-01 4.15813297e-01
5.43732703e-01 4.82473910e-01 -5.98244429e-01 5.84333718e-01
2.54116625e-01 3.34012389e-01 6.30364239e-01 7.97735393e-01
1.45543560e-01 -1.51578104e+00 -4.40963596e-01 1.00299513e+00
-6.56070471e-01 -3.68121117e-01 2.06763167e-02 7.54141271e-01
7.51975253e-02 7.03492403e-01 2.25049451e-01 1.51505128e-01
4.05030370e-01 -2.68517155e-02 9.32066962e-02 -4.40150827e-01
-1.08589125e+00 3.22685987e-01 1.12386949e-01 -4.47547346e-01
-1.13175988e-01 -8.73412132e-01 -1.04978395e+00 -7.45237648e-01
8.36093798e-02 -8.17020461e-02 4.84631151e-01 4.12437499e-01
9.25040603e-01 7.02846229e-01 7.64919102e-01 3.95069048e-02
-1.05967271e+00 -3.77126485e-01 -6.38008893e-01 4.58047539e-01
5.15968800e-01 -2.89307714e-01 -1.01884037e-01 -7.02903450e-01] | [15.001489639282227, -2.056020736694336] |
2e4ab606-e4ca-459f-8d5a-b43b3ee04cee | a-hierarchical-approach-for-visual | 1909.12401 | null | https://arxiv.org/abs/1909.12401v1 | https://arxiv.org/pdf/1909.12401v1.pdf | A Hierarchical Approach for Visual Storytelling Using Image Description | One of the primary challenges of visual storytelling is developing techniques that can maintain the context of the story over long event sequences to generate human-like stories. In this paper, we propose a hierarchical deep learning architecture based on encoder-decoder networks to address this problem. To better help our network maintain this context while also generating long and diverse sentences, we incorporate natural language image descriptions along with the images themselves to generate each story sentence. We evaluate our system on the Visual Storytelling (VIST) dataset and show that our method outperforms state-of-the-art techniques on a suite of different automatic evaluation metrics. The empirical results from this evaluation demonstrate the necessities of different components of our proposed architecture and shows the effectiveness of the architecture for visual storytelling. | ['Ryan Gaines', 'Sagar Gandhi', 'Md Sultan Al Nahian', 'Brent Harrison', 'Tasmia Tasrin'] | 2019-09-26 | null | null | null | null | ['visual-storytelling'] | ['natural-language-processing'] | [-6.67359727e-03 2.26842090e-01 1.43540785e-01 -3.34254682e-01
-6.73068464e-01 -4.47714537e-01 1.10766518e+00 -1.30631775e-01
4.36106212e-02 8.08051229e-01 8.63183200e-01 -4.72902358e-02
3.91894400e-01 -8.22753489e-01 -1.01060760e+00 -2.63847500e-01
8.70284513e-02 2.21234411e-01 3.80529016e-01 -3.28364432e-01
3.66758347e-01 1.10881247e-01 -1.55136812e+00 1.03707683e+00
4.24544603e-01 5.14849067e-01 5.31005859e-01 9.30947959e-01
-2.78867155e-01 1.65312946e+00 -8.33824039e-01 -3.13987404e-01
-3.08192186e-02 -8.25845897e-01 -7.55588770e-01 2.95485079e-01
5.02837062e-01 -6.43843234e-01 -5.97361565e-01 4.75544244e-01
5.62221706e-01 1.28706411e-01 8.22105944e-01 -1.26332939e+00
-9.37964737e-01 9.63876307e-01 -5.09671569e-01 -8.18031654e-02
5.41185975e-01 4.00942326e-01 9.06718850e-01 -9.70500886e-01
1.10179865e+00 1.26027703e+00 4.80085135e-01 5.53380132e-01
-1.02897501e+00 -3.82705927e-01 -6.67940825e-02 3.79697323e-01
-1.33703732e+00 -5.32245994e-01 6.82804644e-01 -5.21941125e-01
1.24244189e+00 -3.50697711e-02 7.03663766e-01 1.43547595e+00
3.26022148e-01 1.05930531e+00 8.00815880e-01 -4.43518579e-01
2.45051637e-01 1.27236709e-01 -2.75140852e-01 6.13999844e-01
-1.20328888e-01 -6.42171726e-02 -8.44672322e-01 3.58774066e-01
9.35069978e-01 -3.96059573e-01 1.11529090e-01 -4.39919829e-01
-1.26032782e+00 8.18933308e-01 4.16819006e-01 3.18965822e-01
-4.75529909e-01 7.00859368e-01 6.46928310e-01 -4.52119224e-02
2.44862452e-01 4.11631972e-01 4.52106267e-01 -2.34924361e-01
-1.27751529e+00 5.94397843e-01 5.03610790e-01 1.13489377e+00
1.84671879e-01 3.46500605e-01 -8.38399112e-01 7.38937497e-01
2.04484388e-01 2.88527422e-02 3.02066237e-01 -9.59633291e-01
5.70905983e-01 4.80055451e-01 1.90941334e-01 -9.55300808e-01
-8.77406374e-02 -3.58780622e-01 -7.51181781e-01 4.40060318e-01
1.00254446e-01 -1.79095909e-01 -9.85629141e-01 1.78321981e+00
-2.07867250e-02 1.66358873e-01 2.31447667e-01 7.05320597e-01
1.19686711e+00 1.29270422e+00 1.24827370e-01 9.67262313e-02
1.01691389e+00 -1.23196232e+00 -6.81932867e-01 -2.73342073e-01
3.11942369e-01 -5.31888723e-01 1.31842542e+00 1.12111136e-01
-1.46807516e+00 -6.58714533e-01 -1.29213464e+00 -2.78178871e-01
-6.84813038e-02 9.59941000e-02 3.34088415e-01 1.27230898e-01
-1.32291591e+00 2.68975675e-01 -5.99321187e-01 -5.30636251e-01
4.39025283e-01 -3.83117676e-01 -2.75679588e-01 -2.15075575e-02
-1.03468955e+00 9.34524059e-01 6.65369928e-01 -3.25311422e-01
-1.51784635e+00 -5.47406912e-01 -9.96198118e-01 2.32448891e-01
-3.04619316e-02 -8.91429365e-01 1.50644207e+00 -8.34093809e-01
-1.13790679e+00 6.80671811e-01 -1.57647654e-01 -7.65598297e-01
7.00930119e-01 -3.60871218e-02 -4.98068966e-02 4.54496443e-01
2.44964302e-01 1.25199449e+00 5.47934234e-01 -1.58398414e+00
-5.24785817e-01 3.72747153e-01 2.70539612e-01 2.79770464e-01
-2.48093620e-01 -1.00431323e-01 -3.88026744e-01 -7.37687111e-01
-5.40262222e-01 -5.56744397e-01 -1.71377286e-01 -1.72198817e-01
-5.45921087e-01 -7.98504129e-02 9.22207475e-01 -7.13003099e-01
1.08943117e+00 -2.01557612e+00 1.58658892e-01 -3.16665798e-01
2.02401072e-01 -4.62088138e-02 -3.02980900e-01 9.91111636e-01
-7.46756746e-03 3.35538946e-02 -2.56354570e-01 -5.30554295e-01
-7.13270158e-02 -1.15378574e-02 -4.49387372e-01 -1.62819520e-01
2.61039972e-01 1.13161576e+00 -7.49343455e-01 -6.95489705e-01
3.40434700e-01 6.13331556e-01 -3.59475195e-01 3.97231758e-01
-6.76348209e-01 2.64349461e-01 -2.46174373e-02 1.12134859e-01
1.04794495e-01 -3.72199208e-01 1.09273545e-01 6.46809265e-02
-5.90232685e-02 3.79529387e-01 -8.88170362e-01 2.06651950e+00
-4.63128507e-01 1.14904344e+00 -5.72711885e-01 -5.02345264e-01
7.75925100e-01 6.02702856e-01 7.07573593e-02 -7.75158286e-01
4.88290228e-02 -3.26870441e-01 -3.95889997e-01 -6.92017376e-01
8.00904632e-01 -1.25677198e-01 -1.26737401e-01 8.19356918e-01
1.46690026e-01 -1.92258313e-01 4.96999264e-01 7.13730216e-01
1.17382944e+00 3.43571872e-01 3.56091052e-01 1.38126075e-01
1.83844015e-01 3.11671406e-01 1.67966224e-02 9.42538023e-01
7.27161020e-02 1.03245902e+00 6.72404349e-01 -4.80231225e-01
-1.61647356e+00 -1.14263213e+00 3.82391363e-01 8.71311843e-01
-1.05863191e-01 -6.15905344e-01 -8.30687046e-01 -5.64947426e-01
-4.46556091e-01 1.39770353e+00 -7.18786597e-01 4.25150171e-02
-5.10934591e-01 -3.53226155e-01 6.29092276e-01 7.06537247e-01
6.63119137e-01 -1.48477888e+00 -1.01910686e+00 1.93067476e-01
-4.19251174e-01 -1.15075207e+00 -2.14969754e-01 -4.51057374e-01
-3.18648010e-01 -6.61505222e-01 -7.67413855e-01 -9.19873059e-01
4.25969034e-01 4.13482279e-01 1.31379008e+00 -1.63791969e-01
-8.56509730e-02 2.74821281e-01 -5.51388562e-01 -3.43772858e-01
-9.09476697e-01 7.13198781e-02 -5.61174631e-01 -2.19411567e-01
-2.60124028e-01 -7.98422277e-01 -4.60070580e-01 -9.86431465e-02
-1.25650811e+00 1.03504276e+00 5.33471584e-01 6.14140511e-01
3.17862064e-01 1.36210918e-01 6.59536600e-01 -6.66864038e-01
8.15938652e-01 -5.03292441e-01 -2.36656114e-01 3.07902277e-01
6.21367581e-02 8.74804333e-02 4.98115212e-01 -3.56571227e-01
-1.26873946e+00 -1.76060735e-03 -5.96356653e-02 -3.73793423e-01
-6.30768910e-02 4.76964027e-01 -5.79742454e-02 4.84427273e-01
7.47392178e-01 4.59941596e-01 -3.29558104e-01 4.15714364e-03
7.83913076e-01 4.47275639e-01 7.77955294e-01 -3.64739984e-01
6.46826446e-01 4.39097613e-01 -1.91743150e-01 -5.63172042e-01
-6.15188062e-01 1.55789375e-01 -2.06337556e-01 -7.84050167e-01
1.00162840e+00 -1.13263929e+00 -3.56022388e-01 3.25342059e-01
-1.54364169e+00 -4.61055934e-01 -4.85702842e-01 -6.97094575e-02
-8.66223693e-01 8.40503201e-02 -7.26345003e-01 -7.66541600e-01
-4.25117940e-01 -9.97037172e-01 1.00465071e+00 1.83050409e-01
-6.24044895e-01 -7.91943312e-01 3.09166968e-01 2.24770948e-01
4.30395812e-01 7.35000134e-01 8.89472902e-01 -1.63893521e-01
-7.98274517e-01 -6.04728274e-02 -3.72591197e-01 4.38235141e-03
-1.62685230e-01 5.77656925e-03 -9.37727630e-01 1.76166501e-02
-2.95749336e-01 -7.07919538e-01 8.87786448e-01 3.34287852e-01
8.34332824e-01 -3.28321129e-01 -8.26473627e-03 2.59932101e-01
1.50401437e+00 2.97115028e-01 1.12100661e+00 4.19810295e-01
8.63407314e-01 6.46054804e-01 4.14999753e-01 5.13538420e-01
7.63619125e-01 4.85253543e-01 4.57623124e-01 -1.02323361e-01
-4.91827130e-01 -9.24159169e-01 5.48254788e-01 5.02874970e-01
1.72728837e-01 -6.27664566e-01 -9.22617435e-01 1.01873457e+00
-2.14166808e+00 -1.43245828e+00 -1.10542014e-01 1.59851491e+00
7.21795917e-01 1.95558578e-01 1.43700466e-01 -1.43152457e-02
5.66554844e-01 6.26425683e-01 -2.91106611e-01 -6.16602659e-01
-2.43266746e-01 -2.62688976e-02 7.90999271e-03 4.12157297e-01
-7.93802679e-01 1.28655350e+00 7.32383251e+00 8.49908829e-01
-9.61137414e-01 6.72479719e-02 6.66722476e-01 -4.80264932e-01
-4.47739094e-01 -1.74606234e-01 -3.95338535e-01 2.03604504e-01
8.87439489e-01 -3.50728691e-01 2.55809218e-01 8.52752864e-01
5.53002596e-01 -1.79805145e-01 -1.16125298e+00 9.06492949e-01
3.96535546e-01 -2.07820010e+00 4.67631549e-01 -3.61589432e-01
9.86137271e-01 -2.76946247e-01 -7.41550922e-02 2.40843117e-01
7.17031300e-01 -1.22102427e+00 1.44329977e+00 4.19961631e-01
8.01747680e-01 -8.03468764e-01 3.90645653e-01 2.35397175e-01
-1.14997709e+00 1.95357352e-01 -2.00871125e-01 -2.61750907e-01
6.05775654e-01 4.03136969e-01 -1.26899672e+00 3.33003432e-01
4.07554954e-01 9.07005727e-01 -6.21842623e-01 9.20031011e-01
-5.69050968e-01 4.55364048e-01 2.34604344e-01 -7.52116740e-02
3.62042606e-01 4.19514507e-01 5.30256391e-01 1.45661879e+00
4.51429039e-01 -5.49511500e-02 -1.34247556e-01 1.23145962e+00
-2.20210955e-01 -3.57294157e-02 -1.01711524e+00 -3.57437730e-01
3.02231044e-01 9.93491411e-01 -6.76429331e-01 -6.26919508e-01
-1.88454926e-01 9.12314177e-01 4.08571869e-01 3.01999569e-01
-1.04312348e+00 -2.82998651e-01 2.62035549e-01 2.10381672e-01
3.31762969e-01 -3.39821130e-01 -5.76936662e-01 -9.86347556e-01
4.67867330e-02 -9.09735620e-01 1.71450749e-01 -1.65164959e+00
-8.55107963e-01 7.63866365e-01 2.45618150e-01 -1.11405706e+00
-6.52611315e-01 1.72973555e-02 -8.29327047e-01 5.57495296e-01
-9.87568259e-01 -1.54854572e+00 -3.56381506e-01 4.76508647e-01
1.07195961e+00 -2.36752257e-01 5.46102822e-01 -6.28405735e-02
-2.66839981e-01 2.37230480e-01 -2.95894265e-01 8.06121230e-02
4.14916009e-01 -1.09263003e+00 9.19374883e-01 1.23764384e+00
4.08529967e-01 1.00312896e-01 1.00296938e+00 -8.05026889e-01
-1.04456961e+00 -1.13815188e+00 9.85426605e-01 -3.50160688e-01
4.50755179e-01 -6.63565397e-01 -6.76714957e-01 7.86671996e-01
1.02317071e+00 -8.39315236e-01 4.10490572e-01 -3.78755063e-01
-4.78294969e-01 2.26177976e-01 -9.20494199e-01 9.49286401e-01
1.09706593e+00 -3.69148910e-01 -7.54489601e-01 2.17731059e-01
9.29529250e-01 -3.65653247e-01 -2.09585622e-01 -6.67940825e-02
5.32758832e-01 -1.15907133e+00 8.55894625e-01 -5.00673413e-01
1.59940445e+00 -3.60114396e-01 -3.29669118e-02 -1.35066259e+00
-3.31796408e-01 -4.51098442e-01 2.36237496e-02 1.37767828e+00
4.32333142e-01 1.26065716e-01 7.84288943e-01 2.85605371e-01
-1.36032939e-01 -4.17575389e-01 -5.46944141e-01 -4.09047604e-01
-8.00000057e-02 -5.84300458e-01 5.66401184e-01 7.27687001e-01
7.47114196e-02 7.04665065e-01 -8.38488758e-01 -1.05624415e-01
4.39756066e-01 -3.93272564e-02 9.91595984e-01 -4.04357910e-01
-3.51262122e-01 -1.77940875e-01 -2.88646668e-01 -7.08862185e-01
7.07491934e-02 -7.43606150e-01 1.86276570e-01 -2.34389377e+00
6.47682309e-01 2.89577782e-01 9.55614522e-02 5.18243372e-01
7.08426982e-02 2.68799037e-01 5.28759301e-01 1.15802353e-02
-8.47594738e-01 6.90028310e-01 1.27002943e+00 -2.22932056e-01
-1.33137837e-01 -7.06881702e-01 -7.84162521e-01 3.88520658e-01
7.17138648e-01 -3.94378304e-01 -8.20440412e-01 -8.19118023e-01
3.48725021e-01 2.89468169e-01 4.87790346e-01 -1.27677500e+00
1.59466982e-01 -2.01710269e-01 5.32968342e-01 -8.98856997e-01
4.02853131e-01 -4.31635469e-01 4.51057583e-01 2.42611200e-01
-6.73541427e-01 1.91922978e-01 2.98115820e-01 2.09752947e-01
-2.39633158e-01 4.85207327e-02 7.01734543e-01 -2.11520106e-01
-7.00472653e-01 -1.53367639e-01 -6.65121078e-01 3.49792726e-02
1.29289961e+00 -1.82215184e-01 -4.79129970e-01 -9.33799326e-01
-3.81670535e-01 3.09094757e-01 5.27961791e-01 6.51697576e-01
9.74945486e-01 -1.62614143e+00 -1.10408092e+00 -2.55147278e-01
2.26921439e-01 -1.18463919e-01 3.38048995e-01 4.83876504e-02
-9.38502908e-01 3.91709894e-01 -6.07075691e-01 -4.97887045e-01
-1.23023522e+00 4.69774097e-01 2.11451486e-01 -4.10172194e-01
-9.14184988e-01 7.06126809e-01 4.76223320e-01 2.12149754e-01
2.12065831e-01 -4.00715917e-02 -3.75378013e-01 -1.07993595e-01
9.08532619e-01 6.87814951e-02 -3.78145546e-01 -4.86245036e-01
1.13170028e-01 9.31952149e-02 -1.47691175e-01 -7.18556404e-01
1.53551221e+00 -1.18856877e-01 2.29874536e-01 5.75630307e-01
7.36016393e-01 -3.72365028e-01 -1.52628565e+00 5.53822666e-02
-1.68959886e-01 -3.40232641e-01 -1.57093853e-01 -9.04358566e-01
-8.38210642e-01 9.78601336e-01 2.59838849e-01 1.97763532e-01
9.84561205e-01 1.56887285e-02 9.00400162e-01 1.22563735e-01
3.50253135e-01 -7.48355150e-01 4.17117268e-01 4.38671052e-01
1.38923824e+00 -9.64302182e-01 -2.09900051e-01 -6.00720011e-02
-1.15000272e+00 8.80947828e-01 5.88092566e-01 -2.77520776e-01
-1.19348876e-01 3.27670693e-01 2.97437906e-02 -2.07138941e-01
-1.20463550e+00 -3.85793336e-02 5.73557094e-02 5.38696170e-01
4.50947016e-01 -8.45030844e-02 -2.07347915e-01 3.09080511e-01
-3.98003697e-01 2.60859996e-01 8.17639947e-01 8.81203353e-01
-4.12340403e-01 -9.26023960e-01 -3.79521012e-01 1.44263301e-02
-2.02599153e-01 -5.00256158e-02 -7.35963762e-01 8.64816844e-01
-1.44009918e-01 9.92588222e-01 -1.05166528e-02 -3.79675508e-01
2.39899725e-01 8.75562429e-02 5.84954500e-01 -6.26724660e-01
-6.51235402e-01 -4.31137979e-02 4.44212317e-01 -4.78815466e-01
-3.03404361e-01 -5.32257199e-01 -1.18393028e+00 -4.62982088e-01
5.05682707e-01 -1.63072571e-01 5.51603079e-01 7.94716656e-01
3.33117425e-01 8.94929051e-01 3.41650546e-01 -9.61315572e-01
8.66741613e-02 -9.72779214e-01 -1.87426060e-01 5.36229730e-01
9.01491866e-02 -2.60130435e-01 1.91163585e-01 4.84822214e-01] | [11.180729866027832, 0.7322661876678467] |
691b4687-4230-4621-8711-add5c7c7323d | is-ai-model-interpretable-to-combat-with | 2010.02006 | null | https://arxiv.org/abs/2010.02006v7 | https://arxiv.org/pdf/2010.02006v7.pdf | Interpretable Machine Learning for COVID-19: An Empirical Study on Severity Prediction Task | The black-box nature of machine learning models hinders the deployment of some high-accuracy models in medical diagnosis. It is risky to put one's life in the hands of models that medical researchers do not fully understand. However, through model interpretation, black-box models can promptly reveal significant biomarkers that medical practitioners may have overlooked due to the surge of infected patients in the COVID-19 pandemic. This research leverages a database of 92 patients with confirmed SARS-CoV-2 laboratory tests between 18th Jan. 2020 and 5th Mar. 2020, in Zhuhai, China, to identify biomarkers indicative of severity prediction. Through the interpretation of four machine learning models, decision tree, random forests, gradient boosted trees, and neural networks using permutation feature importance, Partial Dependence Plot (PDP), Individual Conditional Expectation (ICE), Accumulated Local Effects (ALE), Local Interpretable Model-agnostic Explanations (LIME), and Shapley Additive Explanation (SHAP), we identify an increase in N-Terminal pro-Brain Natriuretic Peptide (NTproBNP), C-Reaction Protein (CRP), and lactic dehydrogenase (LDH), a decrease in lymphocyte (LYM) is associated with severe infection and an increased risk of death, which is consistent with recent medical research on COVID-19 and other research using dedicated models. We further validate our methods on a large open dataset with 5644 confirmed patients from the Hospital Israelita Albert Einstein, at S\~ao Paulo, Brazil from Kaggle, and unveil leukocytes, eosinophils, and platelets as three indicative biomarkers for COVID-19. | ['Xiangfei Chai', 'Sumi Helal', 'Kunwei Li', 'Jian Chen', 'Bei Liu', 'Yayuan Gen', 'Shaolin Li', 'Dingchang Zheng', 'Jiangtao Wang', 'Wenjie Ruan', 'Han Wu'] | 2020-09-30 | null | null | null | null | ['severity-prediction'] | ['computer-vision'] | [ 3.21645200e-01 -1.32013634e-01 -4.42613631e-01 -2.51766354e-01
-3.18241984e-01 -3.64225090e-01 2.23374873e-01 6.40955031e-01
-2.52032518e-01 9.35017645e-01 4.10000086e-01 -9.41223323e-01
-4.41434920e-01 -5.71469128e-01 -4.20602620e-01 -5.04407287e-01
-5.89619696e-01 9.20526743e-01 -5.89410841e-01 2.37304017e-01
5.84888160e-02 5.45953810e-01 -5.66339970e-01 4.49328721e-01
9.85181451e-01 5.20126045e-01 -2.72217155e-01 7.54028499e-01
4.41598058e-01 8.47622156e-01 -2.28804335e-01 -1.06274143e-01
2.22837403e-01 -5.12969136e-01 -3.09633076e-01 -4.92959619e-01
-3.79536986e-01 -5.36273360e-01 1.95933208e-01 3.27134222e-01
3.64555418e-01 -3.79294276e-01 1.02050304e+00 -1.33396709e+00
-3.69388044e-01 4.65536982e-01 -4.28613096e-01 4.33005989e-01
-1.91200525e-01 6.81744993e-01 7.67942429e-01 -8.15708220e-01
5.99106729e-01 1.25597322e+00 1.15190721e+00 3.85230362e-01
-1.24706149e+00 -6.30376935e-01 -1.98669508e-01 -4.27683853e-02
-1.03145969e+00 2.37828344e-01 -6.11995943e-02 -9.56447840e-01
1.07378948e+00 4.83458638e-01 7.20605433e-01 9.15249646e-01
6.05064392e-01 1.54782804e-02 1.26005447e+00 1.43592551e-01
5.41818701e-02 -3.77219723e-04 3.60574216e-01 6.63210034e-01
9.04636443e-01 3.85231584e-01 -2.52906173e-01 -1.03993809e+00
6.52007580e-01 8.03397894e-01 -1.10319197e-01 2.53246389e-02
-1.19581914e+00 9.46117520e-01 -1.25147566e-01 -1.93532974e-01
-6.46449029e-01 -2.29347702e-02 2.51438230e-01 -1.16113335e-01
4.68351245e-01 4.19330895e-01 -1.11577916e+00 1.98695272e-01
-7.50942111e-01 1.11893080e-01 4.85804826e-01 -2.21668422e-01
3.67211968e-01 9.05303732e-02 -2.44723529e-01 3.86909932e-01
3.65817726e-01 1.00444984e+00 1.04274213e-01 -5.60708463e-01
2.99164597e-02 7.55456030e-01 2.84469366e-01 -7.78377652e-01
-8.28300655e-01 -8.50975215e-01 -1.03915215e+00 -2.94580221e-01
1.54089421e-01 -5.78339159e-01 -7.21804023e-01 1.69237363e+00
2.09881272e-02 4.07648861e-01 -1.25239551e-01 8.33444774e-01
1.49361938e-01 3.46340477e-01 3.98512870e-01 -3.95384014e-01
1.49309790e+00 -4.17005748e-01 -1.50020495e-01 -1.97835371e-01
7.22979188e-01 -1.58038422e-01 5.91594696e-01 4.63870257e-01
-6.17024481e-01 -1.50251478e-01 -5.65781951e-01 6.71002924e-01
-1.59472510e-01 -8.95717889e-02 6.63830817e-01 5.87665796e-01
-5.50508201e-01 5.48058689e-01 -7.67057896e-01 -7.67553210e-01
3.55796248e-01 3.62513870e-01 -1.66455612e-01 1.67972539e-02
-1.13081646e+00 1.00308204e+00 -1.15897981e-02 1.26868233e-01
-7.84643471e-01 -1.21483338e+00 -3.61089379e-01 1.44338265e-01
-1.60081163e-01 -1.40013802e+00 2.92841077e-01 -2.74615377e-01
-3.03046435e-01 5.04695475e-01 -4.26549286e-01 -6.39076412e-01
1.10081472e-01 -2.16143399e-01 -3.62660348e-01 2.20063314e-01
5.42105315e-03 3.42531651e-01 3.20746899e-01 -8.99611652e-01
-3.07484716e-01 -8.82061958e-01 -5.21349490e-01 -2.25764647e-01
1.16530135e-01 2.77147740e-01 8.20658267e-01 -6.22011542e-01
-1.65391877e-01 -8.16047847e-01 -6.60673738e-01 -4.25028116e-01
-3.88756037e-01 1.95111483e-01 4.49943334e-01 -7.76181340e-01
1.12415516e+00 -1.77090883e+00 -4.87881541e-01 4.23705697e-01
5.17903447e-01 1.44637927e-01 1.28316522e-01 3.19164127e-01
-1.57752305e-01 6.80741787e-01 -4.35947269e-01 3.81096601e-01
-2.47521847e-01 7.20318258e-02 -5.35889626e-01 4.89845514e-01
5.30868649e-01 1.11023498e+00 -8.35571945e-01 -2.59695500e-01
3.06954741e-01 5.31132042e-01 -9.10461426e-01 -3.65628133e-04
-1.71115175e-01 6.34083629e-01 -4.38477933e-01 6.10078692e-01
6.43937111e-01 -7.68307507e-01 6.67382240e-01 2.41933599e-01
1.52957857e-01 9.43960026e-02 -4.18059021e-01 2.58579940e-01
1.97835878e-01 3.25405985e-01 -2.10447520e-01 -7.30750680e-01
8.99260998e-01 3.61990482e-01 5.95597923e-01 -1.61175802e-02
9.28793550e-02 4.34668325e-02 3.90087783e-01 -3.42741579e-01
-3.55067700e-01 -5.55469155e-01 3.19070995e-01 7.64132082e-01
-6.93754017e-01 2.81051785e-01 -1.04850337e-01 7.27481907e-03
1.25178206e+00 -3.18142742e-01 5.65138042e-01 -2.70703703e-01
2.33236954e-01 2.39032775e-01 8.37690115e-01 7.74095118e-01
-9.97818932e-02 3.90010208e-01 7.85799026e-01 -5.54420054e-01
-7.71718144e-01 -1.27817261e+00 -4.15819645e-01 3.89436603e-01
-6.10689938e-01 -6.78455234e-02 -2.37271965e-01 -1.41140103e-01
3.63469899e-01 7.76056707e-01 -8.14285576e-01 -6.41539097e-02
-3.96812886e-01 -1.51537156e+00 7.62818396e-01 5.38394392e-01
3.64523754e-02 -4.94282156e-01 -9.04924750e-01 1.21941909e-01
-1.55816078e-01 -6.05270088e-01 4.83133234e-02 1.50375888e-01
-1.28886032e+00 -1.47864377e+00 -3.89534801e-01 2.54837796e-03
6.59450948e-01 -1.25090048e-01 8.11177552e-01 3.59140724e-01
-4.59127218e-01 -1.15972636e-02 -2.34385692e-02 -7.18392372e-01
-2.20721424e-01 -4.57043290e-01 4.67487007e-01 -3.18742961e-01
6.48979187e-01 -3.52978945e-01 -9.97803748e-01 1.89485177e-01
-4.40043509e-01 1.66756883e-02 7.93293178e-01 7.96819806e-01
4.92007732e-01 -5.27730584e-01 7.06182301e-01 -1.01016331e+00
2.76175290e-01 -1.14124238e+00 -8.16102102e-02 2.04884157e-01
-1.26856577e+00 -3.69466662e-01 4.94079590e-01 -1.50575489e-01
-6.13072932e-01 -5.99484026e-01 4.79878664e-01 -1.60146073e-01
2.32774708e-02 8.64821255e-01 5.64348280e-01 7.03058481e-01
5.74703991e-01 7.83403218e-02 3.11232090e-01 -3.49032938e-01
-2.47716948e-01 6.78511977e-01 7.51730278e-02 -2.98645616e-01
6.78826571e-01 6.34199381e-01 2.82671571e-01 -7.21604943e-01
-6.02311671e-01 -3.18856925e-01 -2.82385349e-01 5.53140827e-02
8.71972501e-01 -9.46783781e-01 -1.09041667e+00 1.21852919e-01
-8.16458583e-01 -4.39658880e-01 1.39011830e-01 1.01982820e+00
-1.07012063e-01 1.12678416e-01 -6.76423430e-01 -9.24959362e-01
-5.38097143e-01 -8.88641477e-01 4.56898868e-01 -7.45610744e-02
-1.01134706e+00 -1.25467420e+00 4.68335301e-01 4.85154629e-01
4.67293739e-01 6.76563859e-01 1.88735223e+00 -1.21553636e+00
-4.43310380e-01 -2.08516344e-01 -4.89942402e-01 1.22965999e-01
5.82985282e-02 9.22764391e-02 -5.49226701e-01 -4.09707427e-02
1.78501476e-02 1.96927600e-02 5.73480189e-01 7.95959830e-01
7.75484264e-01 -5.49700439e-01 -6.40230119e-01 5.21735370e-01
1.17826641e+00 3.47443461e-01 3.12938601e-01 -9.75931063e-02
2.28044510e-01 5.28114021e-01 2.88842201e-01 8.14733386e-01
3.39797556e-01 2.47029915e-01 1.69873744e-01 -2.54171193e-01
3.98131132e-01 -2.56462336e-01 2.50713408e-01 2.82439739e-01
-2.83519804e-01 -1.08219162e-01 -1.37779510e+00 4.27077502e-01
-1.41245747e+00 -8.38534713e-01 -5.36436737e-01 2.11430073e+00
5.51450014e-01 -2.43835766e-02 1.75762363e-02 -4.27983522e-01
4.97207791e-01 -2.95692831e-01 -6.54965699e-01 -6.09040201e-01
-2.37437069e-01 2.01168075e-01 3.34633827e-01 3.80606383e-01
-4.46338266e-01 1.13980792e-01 6.97980022e+00 4.07627027e-04
-1.05941796e+00 8.97336751e-02 1.24230027e+00 -1.56804636e-01
-5.78554332e-01 2.69029230e-01 -2.50769347e-01 4.65590656e-01
1.20981646e+00 3.84117402e-02 2.08104298e-01 4.63013440e-01
9.02559638e-01 -2.01232493e-01 -1.01831746e+00 4.61595595e-01
3.62009406e-02 -1.50149453e+00 -1.73603948e-02 2.70693928e-01
5.35071194e-01 2.60449052e-01 1.74021959e-01 1.60115793e-01
2.96249241e-01 -1.28023279e+00 5.18407449e-02 8.33815098e-01
8.79391253e-01 -4.43651438e-01 1.10180414e+00 1.28552258e-01
-5.79674363e-01 -4.31565978e-02 9.99512966e-04 -3.95207167e-01
1.61261246e-01 1.01252329e+00 -1.70974779e+00 1.43301561e-01
3.39181483e-01 3.51025403e-01 -2.55061924e-01 6.18237197e-01
-4.35896292e-02 1.23190081e+00 -2.49426976e-01 1.03947617e-01
2.65869219e-02 -1.03473194e-01 5.80745399e-01 1.10039735e+00
1.29998416e-01 4.99437809e-01 -1.75138339e-01 1.09396267e+00
3.76546174e-01 9.64222923e-02 -4.08711195e-01 -3.56233358e-01
2.40510002e-01 9.44901228e-01 -7.71034241e-01 -4.29885447e-01
-1.47507250e-01 2.59699464e-01 -4.76203024e-01 4.18275952e-01
-6.95334136e-01 1.64169788e-01 7.67235577e-01 4.21188116e-01
-2.62190461e-01 7.90974125e-02 -1.07249987e+00 -8.96823645e-01
-6.31459713e-01 -9.11115289e-01 6.28694177e-01 -8.67910504e-01
-1.29544604e+00 1.76883921e-01 -1.20113887e-01 -7.60640979e-01
-5.35143852e-01 -5.35017908e-01 -6.54493928e-01 1.13058579e+00
-1.09017551e+00 -8.66839886e-01 1.07870363e-01 -1.11924848e-02
6.84284195e-02 -4.71934341e-02 1.02624512e+00 -1.45369112e-01
-4.94521260e-01 1.61357135e-01 2.55181432e-01 4.70315516e-02
5.41432440e-01 -7.79297411e-01 2.00159803e-01 4.57933217e-01
-5.93623042e-01 1.28726757e+00 5.69316626e-01 -1.33140087e+00
-9.45735872e-01 -1.11878228e+00 1.29805195e+00 -7.24012196e-01
6.32051170e-01 1.25228956e-01 -5.56580722e-01 8.19278955e-01
-3.57503891e-01 -5.17437935e-01 1.24850023e+00 1.56541571e-01
-3.24374825e-01 2.24807054e-01 -1.43159854e+00 6.46483839e-01
4.60822344e-01 -3.40513617e-01 -5.78794301e-01 3.89222115e-01
4.37241852e-01 3.30320746e-01 -8.78676593e-01 8.35122049e-01
8.99933398e-01 -8.94637108e-01 1.01274884e+00 -1.34803081e+00
3.46928388e-01 -2.62093037e-01 -1.26213476e-01 -8.93786371e-01
-2.59327203e-01 -3.64744544e-01 3.46991330e-01 3.13951313e-01
6.76044822e-01 -9.29410756e-01 6.73901141e-01 6.91484153e-01
2.61018723e-01 -1.10947156e+00 -5.35907388e-01 -5.26988983e-01
1.99337434e-02 -3.56416911e-01 6.76603377e-01 1.07043576e+00
1.68522641e-01 1.36811242e-01 -2.27416515e-01 2.86669701e-01
6.63708746e-01 1.41394064e-01 4.91951555e-01 -1.46635318e+00
-5.15863180e-01 -1.58232152e-01 2.89377570e-02 -8.15793872e-02
-2.00924888e-01 -9.08911705e-01 -6.01566255e-01 -1.66435325e+00
1.02518642e+00 -6.23228014e-01 -4.88397866e-01 6.69338524e-01
-3.80452812e-01 7.18686804e-02 1.31900776e-02 4.08054024e-01
7.75309727e-02 -6.63426742e-02 5.67137599e-01 -5.15299216e-02
-1.16048418e-01 -1.00279925e-02 -7.39536583e-01 8.89221728e-01
9.15438294e-01 -6.26892805e-01 -3.65650952e-01 -1.04770765e-01
3.49376231e-01 3.50931734e-01 9.16760683e-01 -4.25236255e-01
-3.07556152e-01 -7.39570200e-01 7.98956037e-01 -8.62730443e-01
1.50129870e-01 -4.75254089e-01 5.96788168e-01 1.53141737e+00
-1.58832312e-01 3.07755649e-01 2.10054293e-01 3.04106742e-01
4.85399395e-01 3.00622135e-01 4.21956509e-01 -1.04688890e-01
1.68415651e-01 9.46407169e-02 -7.84751832e-01 2.35315278e-01
6.77430093e-01 -2.05611974e-01 -7.01620936e-01 -2.81979114e-01
-5.68018436e-01 1.84843391e-01 2.82628149e-01 -1.22329772e-01
5.57173967e-01 -5.05932331e-01 -1.17013943e+00 1.08390018e-01
-1.43668443e-01 -5.95321298e-01 5.34628093e-01 1.50729907e+00
-7.27917850e-01 8.94924521e-01 -3.80319878e-02 -5.56338668e-01
-9.56030667e-01 4.75302100e-01 2.41933867e-01 -5.02898276e-01
-1.31981105e-01 4.93522286e-01 3.42053145e-01 -3.24790329e-01
-4.05989558e-01 -4.88457307e-02 1.93539783e-01 -1.30943581e-01
4.96993661e-01 8.44082773e-01 -3.17336082e-01 -1.13906585e-01
-7.74879813e-01 1.39782459e-01 1.69291094e-01 1.59726113e-01
1.35709918e+00 6.89466521e-02 -6.12468541e-01 4.01032418e-01
7.32271731e-01 4.78515914e-03 -6.66516244e-01 2.94654518e-01
-1.75466031e-01 -8.83794874e-02 -4.71133798e-01 -1.29598498e+00
-3.84167910e-01 8.28506052e-01 7.84434855e-01 -2.26615340e-01
5.98262727e-01 -1.69491336e-01 5.87280869e-01 2.31207594e-01
7.93426409e-02 -4.33778971e-01 -3.16999793e-01 2.92851180e-01
3.63899022e-01 -8.01889777e-01 1.51521400e-01 2.53442880e-02
-6.23677194e-01 7.54749238e-01 2.87097514e-01 1.81926250e-01
6.02852762e-01 1.02211379e-01 8.51822495e-02 -2.55371869e-01
-1.36086023e+00 1.79500669e-01 -1.46169588e-01 7.21040666e-01
2.61959225e-01 4.73668963e-01 -4.38445032e-01 1.06739855e+00
4.11013626e-02 2.87268698e-01 4.51096833e-01 2.71513194e-01
-4.89764601e-01 -8.12185466e-01 -6.38115585e-01 1.18845594e+00
-6.04378760e-01 -6.29454792e-01 -4.97375399e-01 8.20853174e-01
4.26985562e-01 9.65707660e-01 3.85511249e-01 -1.67508215e-01
-1.82552621e-01 5.76683521e-01 4.26583588e-02 -4.94317770e-01
-6.29857063e-01 2.21219342e-02 2.93495413e-02 -4.26143967e-02
-2.31750906e-01 -7.45374799e-01 -1.42846870e+00 -4.69784260e-01
8.69842172e-02 1.95852399e-01 3.44191283e-01 1.01113832e+00
7.11433172e-01 2.14190960e-01 4.85843509e-01 6.47697598e-02
-5.08436501e-01 -9.35694814e-01 -4.76315796e-01 -1.34593463e-02
4.03544784e-01 -3.71997029e-01 -6.50115073e-01 -9.05841812e-02] | [7.950652599334717, 5.931516170501709] |
471558ec-8ddc-4140-b005-6adc8e2008fd | improving-semantic-dependency-parsing-with | null | null | https://aclanthology.org/W19-6202 | https://aclanthology.org/W19-6202.pdf | Improving Semantic Dependency Parsing with Syntactic Features | We extend a state-of-the-art deep neural architecture for semantic dependency parsing with features defined over syntactic dependency trees. Our empirical results show that only gold-standard syntactic information leads to consistent improvements in semantic parsing accuracy, and that the magnitude of these improvements varies with the specific combination of the syntactic and the semantic representation used. In contrast, automatically predicted syntax does not seem to help semantic parsing. Our error analysis suggests that there is a significant overlap between syntactic and semantic representations. | ['Robin Kurtz', 'Marco Kuhlmann', 'Daniel Roxbo'] | 2019-09-01 | null | null | null | ws-2019-9 | ['semantic-dependency-parsing'] | ['natural-language-processing'] | [-1.20076761e-01 5.17225742e-01 -1.55918181e-01 -9.66864109e-01
-7.44844735e-01 -6.17725790e-01 4.66223001e-01 2.76418716e-01
-4.88630444e-01 6.62789822e-01 6.10222399e-01 -5.07304847e-01
6.08765893e-03 -8.81214619e-01 -5.82727551e-01 -2.18604341e-01
5.86687997e-02 4.48512495e-01 2.59837121e-01 -4.68095660e-01
1.73187181e-01 2.08910450e-01 -1.26569688e+00 5.31944096e-01
5.36780357e-01 8.64941955e-01 2.05925882e-01 5.76528430e-01
-8.65250707e-01 7.31223881e-01 -7.80011833e-01 -3.97141367e-01
-1.33415729e-01 -5.30814946e-01 -1.33470190e+00 -3.99185032e-01
2.71013737e-01 -1.77734688e-01 -5.65983448e-03 9.81088758e-01
9.30160210e-02 -2.29560003e-01 2.62512505e-01 -5.03551126e-01
-1.08555615e+00 8.63587856e-01 -1.84505936e-02 5.23927987e-01
3.72412235e-01 -2.33992845e-01 1.49752080e+00 -5.65875292e-01
7.77901947e-01 1.55662799e+00 8.77189040e-01 6.35518134e-01
-1.04161310e+00 -5.00007451e-01 4.15099233e-01 -3.28658909e-01
-4.98336881e-01 -2.24420056e-01 6.60774469e-01 -2.41215646e-01
1.49844241e+00 -3.60206634e-01 4.22500521e-01 1.00628626e+00
2.75980115e-01 3.02851439e-01 1.20377672e+00 -6.34287179e-01
2.57121623e-01 -1.67674258e-01 9.31536436e-01 6.79832458e-01
4.54465359e-01 3.96416038e-02 -3.04539353e-01 -1.57174140e-01
5.10314643e-01 -5.36785245e-01 1.74020499e-01 -2.38946173e-02
-5.55366695e-01 1.30806613e+00 3.89109254e-01 7.22574413e-01
-2.70065725e-01 3.49216610e-01 8.08391631e-01 3.76266092e-01
5.36760926e-01 6.75695777e-01 -1.22431898e+00 -1.71282187e-01
-4.52555209e-01 2.17866004e-01 1.03324437e+00 4.71113205e-01
6.41953826e-01 9.76355970e-02 3.33921194e-01 9.96851385e-01
2.83735126e-01 1.51924819e-01 4.89237756e-01 -1.18570387e+00
5.49518704e-01 6.71502233e-01 -1.73535235e-02 -6.60594404e-01
-6.63319170e-01 3.35561559e-02 2.52702475e-01 1.34874970e-01
7.89886892e-01 -3.98185760e-01 -9.51723695e-01 2.11469388e+00
-1.31751001e-01 -4.85639155e-01 3.32070887e-01 7.73959398e-01
9.48538303e-01 2.39066020e-01 8.98557782e-01 1.60043657e-01
1.50581861e+00 -5.30254543e-01 -5.30336738e-01 -7.94416249e-01
1.04236078e+00 -5.78535974e-01 9.28453088e-01 -2.05315411e-01
-1.00015759e+00 -3.03303361e-01 -9.14242029e-01 -1.86148182e-01
-5.20864904e-01 -2.07775474e-01 1.10687113e+00 8.01166117e-01
-1.15852654e+00 7.85228968e-01 -7.90946186e-01 -6.65920198e-01
4.37735766e-01 4.11946535e-01 -5.34483552e-01 1.54485515e-05
-1.27669561e+00 1.17545772e+00 7.61218965e-01 -2.86758512e-01
-2.55138248e-01 -3.46797407e-01 -1.08645463e+00 2.10815623e-01
-9.41054076e-02 -6.55427277e-01 1.45553553e+00 -1.58687472e+00
-1.21094346e+00 1.16887057e+00 -2.55654842e-01 -3.32190931e-01
-2.42642298e-01 -3.01903874e-01 -1.54786706e-01 1.53114989e-01
2.64847428e-01 7.70771623e-01 2.96295255e-01 -1.00576138e+00
-5.36915362e-01 -6.00941181e-01 4.69016254e-01 2.12034181e-01
2.52736155e-02 7.03299880e-01 4.26741038e-03 -5.00024557e-01
2.34776050e-01 -7.06933796e-01 -3.18400025e-01 -5.39835155e-01
-7.22663999e-02 -7.67406046e-01 6.41796947e-01 -9.23030138e-01
8.32054436e-01 -2.04719687e+00 -1.33855015e-01 -1.55477002e-01
-1.23076029e-01 4.40925322e-02 -1.08505055e-01 1.58161044e-01
-3.62820327e-01 4.02875394e-01 -2.15818405e-01 -2.47480541e-01
-5.13253808e-02 6.14843071e-01 -1.55037969e-01 5.95488176e-02
4.20628220e-01 1.13656902e+00 -8.29679072e-01 -2.18376204e-01
7.14576012e-03 2.74593025e-01 -5.59840679e-01 1.41989384e-02
-3.10569525e-01 2.24641845e-01 -7.67971873e-01 5.82544923e-01
2.71811634e-01 -3.47758621e-01 6.38147414e-01 3.33927006e-01
4.74256463e-02 1.36697423e+00 -5.16874731e-01 1.91549206e+00
-4.22375590e-01 4.25802022e-01 2.08821684e-01 -1.31437802e+00
9.42224801e-01 2.00810790e-01 -4.86877142e-03 -7.84375548e-01
3.37784618e-01 3.34061921e-01 3.77611905e-01 -1.62241876e-01
2.04112157e-01 -3.44064116e-01 -3.98031235e-01 3.75272363e-01
4.15873617e-01 4.53637075e-03 1.40161499e-01 -8.92652720e-02
1.24355376e+00 3.64944607e-01 3.24961931e-01 -7.64085829e-01
1.93367824e-01 3.35708112e-01 6.48033619e-01 5.47113597e-01
-1.56288743e-01 5.08412004e-01 8.46912146e-01 -6.90811336e-01
-8.89327407e-01 -8.67277265e-01 -3.68263483e-01 1.46197510e+00
-1.25117958e-01 -3.94455314e-01 -1.09263217e+00 -1.14741886e+00
-1.66502237e-01 8.79419148e-01 -6.73197746e-01 1.78984582e-01
-8.98122191e-01 -9.42665100e-01 4.86326516e-01 1.04659855e+00
2.30710939e-01 -1.45456707e+00 -8.13700140e-01 4.05537546e-01
2.29669422e-01 -1.31547046e+00 2.96841711e-01 7.04953790e-01
-1.33637846e+00 -1.16621196e+00 -3.74716073e-02 -9.36793089e-01
4.73935038e-01 -2.55591124e-02 1.58302450e+00 3.44439298e-01
1.26312390e-01 2.50880778e-01 -5.67079961e-01 -1.65187478e-01
-7.15152025e-01 1.18641064e-01 -5.41041136e-01 -1.03773510e+00
8.67815316e-01 -4.09978777e-01 -3.20946366e-01 -2.42420852e-01
-4.87770587e-01 -4.02470559e-01 2.28721321e-01 8.82471681e-01
2.25531515e-02 -3.88229489e-01 6.76728725e-01 -1.20950913e+00
6.41346931e-01 -5.11841059e-01 -3.43480587e-01 1.00223012e-02
-5.78513801e-01 4.61710095e-01 5.24569273e-01 2.73033887e-01
-1.19274449e+00 -2.14221515e-03 -6.51167512e-01 1.58137023e-01
-7.01109588e-01 5.18546760e-01 4.48656678e-02 2.52686679e-01
5.33674598e-01 -5.07923543e-01 -2.07235962e-01 -6.64561212e-01
2.94563711e-01 1.70658380e-01 2.81098574e-01 -8.36532593e-01
1.93406940e-01 2.02967077e-01 -1.73463315e-01 -5.17149568e-01
-1.21402192e+00 1.53218672e-01 -7.37284899e-01 6.45798266e-01
1.52802730e+00 -7.16313183e-01 -1.49706289e-01 1.39136806e-01
-1.55444956e+00 -4.25182104e-01 -2.18993545e-01 2.82707423e-01
-3.71383280e-01 3.77676487e-01 -1.01398027e+00 -3.13833147e-01
-3.81957144e-01 -1.22611117e+00 1.09338927e+00 -2.50409842e-02
-6.14783764e-01 -1.22149563e+00 6.00309996e-03 3.72412741e-01
2.82492101e-01 -1.32669797e-02 1.31396961e+00 -1.13737452e+00
-1.89421251e-01 -1.53908208e-01 -4.32813764e-01 3.98945868e-01
5.87634221e-02 -3.60319257e-01 -1.01583529e+00 2.71687090e-01
1.87945869e-02 -4.06519234e-01 9.69098568e-01 5.79974353e-01
7.66476691e-01 1.26420349e-01 -1.56316444e-01 4.51078773e-01
1.42216706e+00 1.31845608e-01 4.48551655e-01 5.99811614e-01
6.14474475e-01 9.49255347e-01 4.11924362e-01 -2.00580835e-01
5.48737526e-01 2.71240473e-01 1.89115986e-01 1.74020693e-01
-1.44134030e-01 -9.68950987e-02 3.02917838e-01 3.62397373e-01
9.76228714e-02 1.07370824e-01 -1.10224152e+00 4.45089608e-01
-1.57965994e+00 -5.59457183e-01 -2.64601737e-01 1.69486332e+00
6.07118785e-01 3.71490389e-01 -2.32667267e-01 -1.19036540e-01
6.73681736e-01 2.52491713e-01 -1.40421212e-01 -1.11473250e+00
-8.84841979e-02 7.86197841e-01 6.12018824e-01 6.03522420e-01
-1.06030774e+00 1.53829300e+00 8.41839886e+00 1.31702632e-01
-8.39577317e-01 3.96227092e-01 5.96261144e-01 4.97863740e-01
-3.33591312e-01 2.52256274e-01 -8.80399287e-01 3.39572281e-01
1.39549303e+00 2.77899832e-01 1.26028016e-01 1.02349341e+00
-3.15631270e-01 -1.84353277e-01 -1.09173548e+00 2.34269410e-01
-1.88017026e-01 -1.27125657e+00 -1.65421247e-01 -2.07613587e-01
1.57828763e-01 3.12406570e-01 -5.78825831e-01 3.49636793e-01
8.78364503e-01 -1.02904856e+00 6.23521328e-01 -2.25807190e-01
2.89223522e-01 -5.24649560e-01 8.76020610e-01 3.51709984e-02
-8.87092292e-01 -1.65551707e-01 -3.68469208e-01 -4.78911310e-01
2.09692046e-01 3.36285442e-01 -2.90773422e-01 1.93669647e-01
9.89253461e-01 3.67638022e-01 -6.59626782e-01 -3.76697397e-03
-8.87734234e-01 7.89120018e-01 -1.87931016e-01 1.03825592e-01
4.72929806e-01 8.95061195e-02 1.92892849e-01 1.50336301e+00
-3.89216207e-02 2.22053453e-01 6.31971881e-02 8.03053498e-01
3.64531390e-02 2.26571575e-01 -7.22406685e-01 -1.27817616e-01
4.69589382e-01 8.59994829e-01 -9.47059810e-01 -4.40676987e-01
-9.35943604e-01 6.55500889e-01 8.01115692e-01 1.09562397e-01
-4.67687637e-01 2.68656332e-02 8.94973278e-01 -1.93642557e-01
5.64812660e-01 -3.45726162e-01 -9.20938134e-01 -1.00139189e+00
-4.91328631e-03 -4.56921130e-01 7.07851231e-01 -7.82707512e-01
-1.24981856e+00 4.91383523e-01 2.62840614e-02 -6.23397343e-02
-4.34008449e-01 -9.44649935e-01 -7.89393961e-01 1.06312680e+00
-1.59057987e+00 -9.58995044e-01 4.61975038e-02 1.20847315e-01
6.49671674e-01 -4.21089642e-02 1.37026203e+00 -3.07274342e-01
-2.43594125e-01 3.31628799e-01 -4.14643466e-01 5.16388476e-01
3.57137591e-01 -1.42056644e+00 9.53032374e-01 6.62690938e-01
-7.04312185e-03 7.06961811e-01 7.28629887e-01 -5.88102162e-01
-1.18422115e+00 -6.74117744e-01 1.10902977e+00 -6.92272127e-01
7.95027912e-01 -1.68117374e-01 -1.02649355e+00 9.39087689e-01
3.83750916e-01 -9.25791413e-02 5.70047319e-01 6.43242717e-01
-9.13646698e-01 4.46778595e-01 -1.09804142e+00 4.83278893e-02
1.19521570e+00 -5.29764354e-01 -1.30374587e+00 2.71363277e-02
9.70184863e-01 -2.07747757e-01 -7.40485847e-01 4.18756515e-01
3.25877845e-01 -1.06941020e+00 8.44775736e-01 -8.11538875e-01
7.45051622e-01 3.60160351e-01 -4.15492922e-01 -1.17073965e+00
-4.16942745e-01 2.16892511e-01 3.68860960e-01 1.12533665e+00
6.86508238e-01 -7.68964350e-01 9.58742559e-01 9.58275259e-01
-3.65642428e-01 -4.63590264e-01 -8.73533428e-01 -5.13510585e-01
8.20571899e-01 -5.96131861e-01 3.35258454e-01 1.06776237e+00
2.28655681e-01 7.55532265e-01 4.98910010e-01 2.90223192e-02
2.74419457e-01 2.00032920e-01 1.58518329e-01 -1.53240502e+00
-3.87837201e-01 -5.87281346e-01 -3.67262512e-01 -6.29514575e-01
9.23832119e-01 -1.01518250e+00 3.26014571e-02 -1.79856122e+00
4.99770939e-02 -4.94801879e-01 -3.51936698e-01 6.96395874e-01
-3.01889271e-01 -1.82228640e-01 1.67933255e-01 -5.34714982e-02
-3.46019208e-01 6.82666376e-02 6.86145842e-01 3.65566283e-01
-5.60206249e-02 -4.60898697e-01 -1.16121054e+00 1.06210279e+00
1.09132802e+00 -6.15095794e-01 -6.80185631e-02 -9.04363692e-01
1.41989201e-01 1.68530941e-02 5.33631779e-02 -5.62921226e-01
-4.34716970e-01 -1.00108320e-02 3.96282196e-01 2.72663999e-02
1.69225305e-01 -3.94258201e-01 -4.71958786e-01 3.91606867e-01
-3.86139899e-01 3.57478172e-01 3.92006963e-01 3.50089729e-01
-1.63612321e-01 -5.38184285e-01 8.45287383e-01 -6.06554449e-01
-8.46494257e-01 -4.58941013e-01 -3.86552542e-01 3.53733301e-01
4.90978241e-01 -4.37816344e-02 -4.71547157e-01 5.27440719e-02
-7.95628548e-01 -1.40273049e-01 4.80781406e-01 6.53206468e-01
5.81267886e-02 -8.59420955e-01 -4.59472448e-01 -8.81429240e-02
-1.19764894e-01 -2.69465595e-01 -3.07371229e-01 2.29765296e-01
-4.23172146e-01 5.70411265e-01 -1.32261276e-01 -3.04719418e-01
-1.07866144e+00 2.52111465e-01 2.78009921e-01 -3.57350200e-01
-7.14907169e-01 9.51546192e-01 3.52197140e-01 -5.64622581e-01
-1.82674639e-02 -5.30476093e-01 -2.46290013e-01 -5.14383987e-02
2.68124789e-01 -1.61832318e-01 9.33620483e-02 -5.89513838e-01
-4.62382972e-01 3.77031863e-01 -2.17206459e-02 6.59969747e-02
1.47196805e+00 1.35795265e-01 -2.72100836e-01 4.50298309e-01
9.28076088e-01 -3.37636054e-01 -9.21615779e-01 7.68733695e-02
5.30219436e-01 -1.69007406e-01 4.14280258e-02 -9.14460301e-01
-1.02966845e+00 8.42328429e-01 3.07174772e-01 2.40309417e-01
6.70218945e-01 5.32724679e-01 8.11841786e-01 4.66643572e-01
5.96075296e-01 -1.07500565e+00 -3.40539962e-01 1.09945595e+00
4.46287423e-01 -1.35724163e+00 -1.27982154e-01 -5.16237676e-01
-6.20307744e-01 1.12655306e+00 6.21493220e-01 -6.42037213e-01
7.15063930e-01 4.80906576e-01 1.68668509e-01 -4.88538772e-01
-6.79214001e-01 -2.98238248e-01 -2.37382889e-01 5.79985380e-01
1.08055127e+00 1.23118222e-01 -7.96494305e-01 8.84773433e-01
-5.28690040e-01 -3.16364050e-01 2.77788311e-01 1.05468106e+00
-7.06636548e-01 -1.46968937e+00 -2.28917018e-01 3.11554432e-01
-9.01290357e-01 -4.29835379e-01 -6.18695736e-01 1.01733851e+00
-1.61918476e-01 9.42439198e-01 2.85023510e-01 3.11791170e-02
2.21137732e-01 8.16341460e-01 5.23841798e-01 -1.10087216e+00
-8.51859272e-01 -8.68330374e-02 8.23190868e-01 -7.66652644e-01
-4.79704440e-01 -6.81968749e-01 -1.83533800e+00 3.69592234e-02
-1.60889447e-01 6.98158219e-02 9.43198085e-01 1.20177126e+00
3.15058917e-01 3.34599555e-01 4.44325395e-02 -5.80404103e-01
-4.36294824e-01 -1.01197016e+00 -2.42746934e-01 6.30904913e-01
4.30864468e-02 -6.77990496e-01 -3.29316944e-01 -1.10431619e-01] | [10.482059478759766, 9.34813117980957] |
ae36aa6d-1433-4b10-be40-34e1bbb0852d | learning-semi-supervised-gaussian-mixture | 2305.06144 | null | https://arxiv.org/abs/2305.06144v1 | https://arxiv.org/pdf/2305.06144v1.pdf | Learning Semi-supervised Gaussian Mixture Models for Generalized Category Discovery | In this paper, we address the problem of generalized category discovery (GCD), \ie, given a set of images where part of them are labelled and the rest are not, the task is to automatically cluster the images in the unlabelled data, leveraging the information from the labelled data, while the unlabelled data contain images from the labelled classes and also new ones. GCD is similar to semi-supervised learning (SSL) but is more realistic and challenging, as SSL assumes all the unlabelled images are from the same classes as the labelled ones. We also do not assume the class number in the unlabelled data is known a-priori, making the GCD problem even harder. To tackle the problem of GCD without knowing the class number, we propose an EM-like framework that alternates between representation learning and class number estimation. We propose a semi-supervised variant of the Gaussian Mixture Model (GMM) with a stochastic splitting and merging mechanism to dynamically determine the prototypes by examining the cluster compactness and separability. With these prototypes, we leverage prototypical contrastive learning for representation learning on the partially labelled data subject to the constraints imposed by the labelled data. Our framework alternates between these two steps until convergence. The cluster assignment for an unlabelled instance can then be retrieved by identifying its nearest prototype. We comprehensively evaluate our framework on both generic image classification datasets and challenging fine-grained object recognition datasets, achieving state-of-the-art performance. | ['Kai Han', 'Xin Wen', 'Bingchen Zhao'] | 2023-05-10 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [ 5.69848716e-01 2.92003721e-01 -3.03805709e-01 -4.52636868e-01
-6.38231695e-01 -6.54028654e-01 7.44315743e-01 2.46371686e-01
-4.69309449e-01 5.15782833e-01 -3.41967046e-01 -1.00169197e-01
-3.16626489e-01 -6.16029501e-01 -5.87165236e-01 -1.27949107e+00
2.19694585e-01 1.15266037e+00 2.24340603e-01 3.37490350e-01
2.87558138e-01 5.01445532e-01 -2.21892166e+00 5.11098742e-01
6.98686004e-01 1.14051068e+00 3.83440644e-01 6.15655065e-01
-2.35142067e-01 5.57010233e-01 -4.11997557e-01 -7.96097368e-02
2.89497346e-01 -4.68854517e-01 -1.17919075e+00 8.87677789e-01
3.39763880e-01 2.85852432e-01 9.85516757e-02 1.08179188e+00
1.61495999e-01 1.14357099e-01 1.08182216e+00 -1.41590524e+00
-5.81631437e-02 3.39386195e-01 -7.08221853e-01 3.35829109e-02
-6.13528639e-02 -2.24364921e-01 8.81442070e-01 -7.96192288e-01
5.42462826e-01 1.16377699e+00 3.68179053e-01 4.83753592e-01
-1.41611540e+00 -5.22995174e-01 3.57291579e-01 4.04996395e-01
-1.55138445e+00 -5.86942494e-01 6.54295444e-01 -6.19145572e-01
4.06535953e-01 2.93235689e-01 2.98547775e-01 5.57528198e-01
-5.26268482e-01 6.28207326e-01 1.21998119e+00 -8.94553483e-01
6.44174755e-01 4.33738142e-01 3.14221770e-01 3.87389779e-01
9.76606309e-02 -5.87036274e-02 -3.71337868e-02 -2.32145011e-01
2.73181140e-01 8.76172855e-02 -1.51941210e-01 -9.12873328e-01
-1.14690471e+00 9.85477746e-01 1.68600142e-01 2.93718874e-01
-3.90092015e-01 -4.31935310e-01 2.49145254e-01 4.68744636e-02
3.63646090e-01 2.05970049e-01 -6.43801272e-01 2.90644109e-01
-1.13903368e+00 -4.36666794e-02 8.51872623e-01 8.56793165e-01
1.10514724e+00 -4.98559326e-01 1.23403370e-01 1.02392256e+00
4.11266536e-01 1.46587700e-01 7.42550969e-01 -1.03238368e+00
2.26964533e-01 8.80393505e-01 -6.55211229e-03 -6.46345079e-01
-3.49426925e-01 -4.53653008e-01 -8.44031334e-01 1.98947996e-01
5.64234853e-01 3.06638986e-01 -1.22360051e+00 1.51418614e+00
6.32752061e-01 3.18948925e-01 3.64517458e-02 6.83529854e-01
7.22398460e-01 4.60749418e-01 -1.41575158e-01 -5.76227188e-01
1.18497908e+00 -7.45124996e-01 -3.30651462e-01 -1.70028046e-01
5.98359466e-01 -5.66346765e-01 5.50362408e-01 5.40542781e-01
-5.91401458e-01 -7.06119895e-01 -9.79066551e-01 4.88228977e-01
-4.36466664e-01 2.69166648e-01 3.93835247e-01 5.69901288e-01
-8.58074307e-01 4.42158550e-01 -6.79351211e-01 -3.84344310e-01
4.36882168e-01 5.35725772e-01 -4.16635811e-01 -3.36996466e-01
-6.00076437e-01 5.34498155e-01 9.25548971e-01 3.55390720e-02
-7.34497666e-01 -1.47016108e-01 -7.77144730e-01 -1.72572553e-01
5.50493181e-01 -3.45178962e-01 1.07469261e+00 -1.20557213e+00
-1.15013826e+00 1.28104329e+00 -1.85694605e-01 -2.85829484e-01
2.82601178e-01 4.19346392e-01 -3.13559026e-01 2.06860736e-01
2.09884360e-01 7.68053174e-01 1.04099762e+00 -1.64655793e+00
-9.29177999e-01 -5.24232268e-01 -1.52893379e-01 1.19485870e-01
-4.49263826e-02 -1.42325312e-01 -5.81642568e-01 -3.53505313e-01
5.50607145e-01 -1.19280612e+00 -2.34757721e-01 -2.07018301e-01
-5.49590528e-01 -3.62499356e-01 1.01992953e+00 -1.14434399e-01
9.56806183e-01 -2.25251341e+00 1.61430582e-01 4.05424118e-01
2.60912865e-01 3.64050835e-01 -8.79037678e-02 1.07189626e-01
-2.84373999e-01 -1.59089416e-01 -4.86836076e-01 -3.10264468e-01
-9.05191153e-02 6.28926039e-01 -1.52857602e-01 5.71294427e-01
1.75065920e-01 4.85035688e-01 -9.13677216e-01 -7.23427474e-01
3.14379990e-01 1.22450605e-01 -4.50614870e-01 3.10594559e-01
-1.29461408e-01 6.40300035e-01 -1.12261690e-01 4.28580970e-01
7.98158586e-01 -4.06009644e-01 3.91462684e-01 -1.31081179e-01
8.82178470e-02 -2.51354855e-02 -1.83202553e+00 1.28102124e+00
-2.58855611e-01 1.61954805e-01 -2.90904716e-02 -1.53136754e+00
8.48320484e-01 2.09745601e-01 4.69836354e-01 -2.72330225e-01
7.29636252e-02 2.21342385e-01 5.81193268e-02 -4.66926724e-01
1.43477172e-01 -3.12679470e-01 2.61405986e-02 5.00050485e-01
3.45616996e-01 2.32527088e-02 4.61305082e-01 7.36725926e-02
7.39020765e-01 -5.22814281e-02 5.34651875e-01 -1.32345974e-01
5.37506282e-01 -2.67402232e-02 5.29357076e-01 9.03763354e-01
1.68963820e-02 7.39982247e-01 3.32284957e-01 -2.31540889e-01
-8.60075057e-01 -1.00918818e+00 -4.94861007e-01 1.10149896e+00
5.06386347e-02 -1.98954031e-01 -7.02304602e-01 -1.00904977e+00
-1.21987976e-01 4.44003612e-01 -7.34247744e-01 -1.25627860e-01
-1.44327641e-01 -9.44371283e-01 4.10165079e-02 3.00754219e-01
3.45108449e-01 -9.52918530e-01 -5.21966457e-01 9.44602862e-02
-2.31432617e-01 -9.49741662e-01 -4.83926795e-02 5.87011576e-01
-7.97432125e-01 -1.41968167e+00 -4.89253044e-01 -1.03360474e+00
1.07189763e+00 2.81403720e-01 9.27131593e-01 4.71056625e-02
-4.02833015e-01 2.91148394e-01 -4.83148575e-01 -1.43456191e-01
-5.51041901e-01 -9.81639996e-02 6.12227619e-02 4.36840951e-01
3.05410028e-01 -3.44606847e-01 -2.94729412e-01 5.87679386e-01
-9.98169541e-01 -1.09854460e-01 7.00381935e-01 9.61768031e-01
9.60279226e-01 6.67433262e-01 6.22356474e-01 -1.19442022e+00
-8.95909667e-02 -6.71080828e-01 -3.64635855e-01 1.93625867e-01
-5.98242462e-01 7.84009919e-02 5.85736752e-01 -7.37200558e-01
-7.83989549e-01 4.57807899e-01 3.11102599e-01 -6.11774981e-01
-7.50129521e-01 3.85635287e-01 -5.18195510e-01 8.64622369e-02
5.61827719e-01 1.45112261e-01 4.06094939e-02 -6.19821131e-01
5.03652573e-01 1.04685152e+00 7.15736628e-01 -5.30974329e-01
5.31438589e-01 6.34776354e-01 -1.23670988e-01 -8.34982872e-01
-8.84686589e-01 -1.04886544e+00 -1.27195215e+00 -3.03619187e-02
5.46994984e-01 -7.22741663e-01 -4.17621046e-01 3.58118057e-01
-7.92775989e-01 -2.05870375e-01 -5.84867060e-01 4.54255372e-01
-6.26743078e-01 6.06830955e-01 -1.89190671e-01 -7.95226455e-01
1.63919672e-01 -1.20425296e+00 1.12073040e+00 -2.78568435e-02
-8.84761140e-02 -8.43839169e-01 5.35791144e-02 3.04218173e-01
-1.21375039e-01 1.80427089e-01 1.05088985e+00 -1.09403741e+00
-2.42359698e-01 -4.52918291e-01 -1.69764787e-01 4.54944581e-01
2.05256984e-01 -2.18876213e-01 -1.06120348e+00 -4.83865738e-01
8.67305696e-02 -2.66527265e-01 9.44312513e-01 3.10771793e-01
1.17625070e+00 -3.40084195e-01 -5.56413770e-01 4.09326643e-01
1.18590796e+00 2.20635861e-01 3.47534925e-01 2.45819345e-01
6.92580342e-01 9.13902104e-01 6.65423214e-01 5.05443692e-01
4.01205033e-01 7.46503174e-01 4.03880894e-01 -2.71749627e-02
-1.67189203e-02 1.12099349e-01 -8.92455578e-02 6.02811992e-01
1.92103669e-01 -1.37165174e-01 -8.03670764e-01 5.03065765e-01
-1.93736970e+00 -7.71533966e-01 -1.91702589e-01 2.42528629e+00
7.33551025e-01 -2.48464639e-03 3.08105737e-01 6.15142047e-01
1.13815784e+00 -2.20387682e-01 -5.91663182e-01 -1.41514465e-01
2.90525034e-02 2.33016759e-01 2.84105301e-01 1.65024489e-01
-1.46739614e+00 6.38656020e-01 5.67146111e+00 8.41082931e-01
-8.41672659e-01 -1.04415663e-01 7.47795522e-01 1.05789438e-01
1.95008799e-01 2.87158698e-01 -7.66981423e-01 4.85442609e-01
8.61429930e-01 2.10127264e-01 3.18358421e-01 8.07578146e-01
-6.58527613e-02 -3.59983236e-01 -1.40987790e+00 9.52359200e-01
3.13081026e-01 -8.72850060e-01 5.80195375e-02 3.50784808e-01
7.30737448e-01 -1.34170815e-01 -2.01934166e-02 3.44798952e-01
2.78401822e-01 -8.10507119e-01 6.58925831e-01 3.67135257e-01
8.39507222e-01 -8.01310062e-01 7.49916732e-01 7.61088192e-01
-1.12519586e+00 -2.76637256e-01 -3.65995675e-01 5.71805015e-02
-3.31570119e-01 7.13492632e-01 -9.55581248e-01 5.19432306e-01
6.15095735e-01 6.29778266e-01 -8.47376049e-01 1.27644598e+00
-2.82372050e-02 5.93886077e-01 -3.53774875e-01 5.23237824e-01
1.00455239e-01 -2.57679105e-01 2.65762627e-01 1.14703023e+00
-8.26766528e-03 4.63150349e-03 5.78205407e-01 5.94111919e-01
7.30795413e-02 1.48774758e-01 -3.86823624e-01 2.58276910e-01
4.73813862e-01 1.33085597e+00 -1.27255666e+00 -5.02026856e-01
-2.02440888e-01 8.94569576e-01 3.28866631e-01 2.56654590e-01
-3.50292355e-01 -1.61612079e-01 1.00227378e-01 -2.59688925e-02
6.90106571e-01 1.75782084e-01 2.63858009e-02 -1.01925957e+00
-2.47061774e-02 -7.20420957e-01 7.49695003e-01 -3.60978812e-01
-1.38170016e+00 5.09229124e-01 2.97841936e-01 -1.36845243e+00
-5.45785785e-01 -5.83850324e-01 -3.49804670e-01 5.16966879e-01
-1.41974843e+00 -1.07853746e+00 -1.70294583e-01 4.54613537e-01
4.39607561e-01 -7.69481808e-02 8.18503320e-01 1.46568611e-01
-5.13900161e-01 4.12518233e-01 3.17028284e-01 8.98847058e-02
5.83461285e-01 -1.44553101e+00 -1.88851789e-01 6.63019598e-01
2.87546873e-01 3.42186391e-01 4.78609800e-01 -3.77263546e-01
-8.08739781e-01 -1.34270155e+00 7.43720114e-01 -3.00146967e-01
4.87039864e-01 -4.71173704e-01 -1.05211282e+00 4.68584925e-01
-3.27583015e-01 3.48933131e-01 8.34065974e-01 -7.13771135e-02
-3.83695304e-01 1.19443715e-01 -1.19987798e+00 1.20360609e-02
7.92863190e-01 -3.30943316e-01 -6.12139940e-01 4.13803101e-01
3.16705793e-01 -2.35519763e-02 -6.97963834e-01 4.62818325e-01
2.36409023e-01 -7.98234582e-01 8.68473649e-01 -4.99058157e-01
9.29358080e-02 -6.49325609e-01 -3.80114824e-01 -1.25185776e+00
-2.82043159e-01 2.23192479e-02 -1.78995654e-01 1.38119280e+00
8.86698887e-02 -5.77628911e-01 8.42440486e-01 3.44199508e-01
-2.69174073e-02 -5.99039257e-01 -1.11368561e+00 -8.85481477e-01
-5.75600117e-02 -3.59705120e-01 4.71300721e-01 1.07846642e+00
-1.34154096e-01 3.71204615e-01 -8.28351825e-02 2.28356019e-01
7.88177311e-01 6.10850871e-01 6.89884901e-01 -1.67545211e+00
-4.64966446e-01 -2.93978393e-01 -7.42338538e-01 -7.66981483e-01
4.12520021e-01 -1.19407988e+00 3.03559572e-01 -1.25019145e+00
5.87365210e-01 -7.05863476e-01 -2.58828372e-01 6.87415719e-01
-1.72992840e-01 3.77532899e-01 1.51514888e-01 5.97235560e-01
-1.02635157e+00 2.34827057e-01 7.82749236e-01 -1.73414782e-01
-2.86783606e-01 2.93445587e-01 -6.80853546e-01 7.94753432e-01
5.70031226e-01 -6.04468346e-01 -3.63377810e-01 2.16127664e-01
-4.17652756e-01 -1.10815853e-01 3.89891475e-01 -9.72859442e-01
2.51706541e-01 -2.28844769e-02 3.92650992e-01 -6.46169186e-01
1.35624826e-01 -8.74349117e-01 2.36526638e-01 2.26891384e-01
-4.00700629e-01 -5.22536635e-01 -1.26436457e-01 7.40162075e-01
-8.97188932e-02 -4.36195940e-01 1.04402852e+00 -5.67098148e-02
-7.93831766e-01 2.72035241e-01 -3.09399247e-01 4.86426391e-02
1.17799127e+00 -4.70628679e-01 -1.29170597e-01 -3.47394459e-02
-1.13170791e+00 5.53753674e-02 6.32318377e-01 2.47805148e-01
5.28594315e-01 -1.17276216e+00 -6.26896918e-01 3.60995680e-01
6.10896885e-01 4.55425084e-01 6.74778596e-02 8.36726725e-01
1.25558928e-01 2.77332067e-01 1.21550173e-01 -1.00835359e+00
-1.22238100e+00 1.07412469e+00 2.44390801e-01 -1.79632664e-01
-3.48588824e-01 5.13140261e-01 5.30316293e-01 -7.95961201e-01
2.20122322e-01 2.59910105e-03 -5.21552086e-01 3.09747577e-01
5.24977505e-01 3.25342834e-01 3.45992118e-01 -1.05355680e+00
-2.06401005e-01 4.70843822e-01 -3.43680620e-01 1.50565699e-01
1.22800827e+00 -1.74142882e-01 -2.94257492e-01 7.16700792e-01
1.48836160e+00 -5.32186031e-01 -1.12399817e+00 -5.90937734e-01
3.29336911e-01 -3.97631586e-01 -3.05418707e-02 -6.28837407e-01
-9.93671000e-01 6.32160485e-01 8.26033294e-01 1.19511686e-01
1.12293971e+00 5.86800754e-01 1.46957099e-01 2.67993748e-01
4.20484871e-01 -1.11836052e+00 1.85445435e-02 3.06487590e-01
3.62717807e-01 -1.27324581e+00 -2.78338641e-02 -4.01909024e-01
-5.85328162e-01 1.05655885e+00 3.06368589e-01 -1.06493253e-02
7.25616574e-01 2.59739254e-02 1.81985740e-02 -3.30903411e-01
-6.72113299e-01 -4.71491575e-01 4.21627700e-01 8.01099896e-01
4.45747701e-03 2.56622255e-01 1.16877332e-01 4.40704077e-01
-6.23346642e-02 -2.29337826e-01 4.24239188e-01 9.24334049e-01
-5.42977035e-01 -1.18687260e+00 -6.79329574e-01 8.31324756e-01
-6.40119612e-02 1.99575052e-01 -4.02985811e-01 6.25630975e-01
5.90146780e-01 1.19981194e+00 2.97064245e-01 -3.65357518e-01
1.17249332e-01 1.36662886e-01 4.82902378e-01 -1.05460644e+00
3.40043713e-04 2.60733247e-01 -1.81206748e-01 -1.34300873e-01
-8.21552992e-01 -8.95887971e-01 -1.23885489e+00 1.82992861e-01
-7.20955431e-01 3.85247439e-01 6.70306921e-01 1.20415890e+00
1.68587625e-01 1.68368340e-01 1.09184170e+00 -1.10744047e+00
-4.09957260e-01 -7.74691403e-01 -8.55477452e-01 5.22290170e-01
2.57657647e-01 -9.13927138e-01 -6.37813270e-01 3.29651505e-01] | [9.53947925567627, 2.929332971572876] |
dba54e19-38bf-4c0a-a42d-8b1d199c7630 | variational-context-deformable-convnets-for | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Xiong_Variational_Context-Deformable_ConvNets_for_Indoor_Scene_Parsing_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Xiong_Variational_Context-Deformable_ConvNets_for_Indoor_Scene_Parsing_CVPR_2020_paper.pdf | Variational Context-Deformable ConvNets for Indoor Scene Parsing | Context information is critical for image semantic segmentation. Especially in indoor scenes, the large variation of object scales makes spatial-context an important factor for improving the segmentation performance. Thus, in this paper, we propose a novel variational context-deformable (VCD) module to learn adaptive receptive-field in a structured fashion. Different from standard ConvNets, which share fixed-size spatial context for all pixels, the VCD module learns a deformable spatial-context with the guidance of depth information: depth information provides clues for identifying real local neighborhoods. Specifically, adaptive Gaussian kernels are learned with the guidance of multimodal information. By multiplying the learned Gaussian kernel with standard convolution filters, the VCD module can aggregate flexible spatial context for each pixel during convolution. The main contributions of this work are as follows: 1) a novel VCD module is proposed, which exploits learnable Gaussian kernels to enable feature learning with structured adaptive-context; 2) variational Bayesian probabilistic modeling is introduced for the training of VCD module, which can make it continuous and more stable; 3) a perspective-aware guidance module is designed to take advantage of multi-modal information for RGB-D segmentation. We evaluate the proposed approach on three widely-used datasets, and the performance improvement has shown the effectiveness of the proposed method.
| [' Qi Wang', ' Nianhui Guo', ' Yuan Yuan', 'Zhitong Xiong'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['scene-parsing'] | ['computer-vision'] | [ 7.87780210e-02 -7.67026693e-02 -1.14697509e-01 -6.57118499e-01
-3.30332667e-01 -5.06073236e-01 4.27564234e-01 5.92336915e-02
-5.28315067e-01 2.84475863e-01 3.83490287e-02 5.10618463e-02
-3.35549027e-01 -1.04199171e+00 -6.97587550e-01 -9.30040896e-01
1.26718596e-01 -4.09845822e-02 8.74641836e-01 -2.95354966e-02
3.60909790e-01 5.55715501e-01 -1.51699138e+00 2.57459253e-01
1.25008380e+00 1.21400595e+00 8.96858871e-01 2.74826467e-01
-3.99924576e-01 3.29691917e-01 -1.85323164e-01 2.31964871e-01
2.19164208e-01 -4.92627732e-02 -6.51620984e-01 2.84117848e-01
3.87040824e-01 -3.07711035e-01 -2.29497835e-01 1.02200627e+00
2.53311813e-01 6.76622033e-01 8.80339503e-01 -6.96583688e-01
-7.68330395e-01 2.16672748e-01 -3.97158623e-01 2.06039816e-01
-1.25674024e-01 1.66271567e-01 7.84847021e-01 -8.34183991e-01
4.30635840e-01 1.36667466e+00 3.61793637e-01 6.50619209e-01
-1.03211713e+00 -3.33256394e-01 7.23507226e-01 1.26169935e-01
-1.30127347e+00 2.07647726e-01 1.04713476e+00 -4.47252423e-01
6.09052062e-01 1.26987115e-01 8.26900184e-01 7.67068207e-01
2.23789647e-01 8.76240671e-01 1.11277211e+00 -1.85911179e-01
4.47646141e-01 1.33027345e-01 2.52843261e-01 7.25745082e-01
-2.79394276e-02 -8.52963403e-02 -2.38606468e-01 1.67090714e-01
1.16519988e+00 4.45661634e-01 -1.95740104e-01 -6.04200065e-01
-1.01242685e+00 7.52402604e-01 9.24742937e-01 4.39041495e-01
-1.66058391e-01 2.34424114e-01 4.92214747e-02 -3.66022348e-01
4.63469476e-01 -8.14592242e-02 -4.30574834e-01 2.74979681e-01
-8.37664604e-01 9.90245864e-02 3.02148342e-01 8.40494752e-01
1.10704482e+00 -9.02083740e-02 -5.22422791e-01 8.45422983e-01
6.94098949e-01 6.71046853e-01 4.23048705e-01 -1.01930213e+00
3.16441000e-01 8.38282943e-01 -7.85809755e-02 -9.77261066e-01
-4.47012186e-01 -2.57677794e-01 -7.59043694e-01 3.10840696e-01
2.38102660e-01 1.43825918e-01 -1.17837071e+00 1.55340922e+00
6.67316079e-01 4.42615569e-01 -1.64608384e-04 1.02268541e+00
1.05295491e+00 5.14205992e-01 6.11187443e-02 1.06220566e-01
1.22501302e+00 -7.25493312e-01 -4.38148946e-01 -1.36828780e-01
1.33780345e-01 -5.56204259e-01 1.26203144e+00 1.23748079e-01
-6.27098560e-01 -8.10029149e-01 -9.46810186e-01 -2.39140034e-01
-3.40236843e-01 7.22434372e-02 7.68258154e-01 5.59571981e-01
-9.58641827e-01 4.52019483e-01 -1.07676113e+00 -3.01293850e-01
5.46505392e-01 2.40016192e-01 -1.20089255e-01 -1.65574729e-01
-8.64179373e-01 2.87480384e-01 7.35351264e-01 2.12188959e-01
-8.16997588e-01 -5.39293766e-01 -9.89763021e-01 -7.36495331e-02
2.98954189e-01 -7.00996339e-01 7.44601130e-01 -7.79159725e-01
-1.59459472e+00 5.48398554e-01 -2.89846420e-01 -1.45346925e-01
4.12960470e-01 -1.52298883e-01 7.66620040e-02 4.33802009e-01
3.40188369e-02 7.78860152e-01 9.98157501e-01 -1.41770899e+00
-7.01018095e-01 -5.63819647e-01 3.29507619e-01 3.79834801e-01
-3.54725868e-01 -4.91401404e-01 -8.93298745e-01 -7.93498933e-01
4.55937475e-01 -7.15169966e-01 -3.89567405e-01 6.75105825e-02
-2.80988663e-01 -3.89050424e-01 1.00145018e+00 -4.17759448e-01
1.09976006e+00 -2.22345948e+00 2.78182000e-01 3.46684128e-01
1.48218796e-01 1.82110935e-01 2.03911886e-01 -1.47165626e-01
4.56206411e-01 -7.64095336e-02 -6.51363909e-01 -3.27622175e-01
9.29193851e-03 5.45863092e-01 -9.48817506e-02 2.00404957e-01
2.50219762e-01 9.32850957e-01 -8.16558123e-01 -5.86243570e-01
7.47562289e-01 7.51432121e-01 -6.49074554e-01 1.33376002e-01
-4.77746427e-01 7.40530729e-01 -8.57474267e-01 5.83319783e-01
9.25016165e-01 -1.33028463e-01 -1.95316002e-01 -4.56842303e-01
-2.57679820e-01 -4.01197135e-01 -1.55218506e+00 1.94037962e+00
-4.74962145e-01 2.54976988e-01 1.04138821e-01 -8.92196655e-01
9.44790781e-01 -2.04361245e-01 2.28221878e-01 -5.86356521e-01
4.78727259e-02 9.85518396e-02 -5.59083343e-01 -6.28465295e-01
4.55851704e-01 2.17145383e-01 1.41078666e-01 -6.13137968e-02
3.79284546e-02 -4.65267420e-01 -9.70003158e-02 1.67402416e-01
5.85060656e-01 3.29366505e-01 -1.77150100e-01 -4.28166837e-01
7.95434117e-01 -2.64691114e-01 8.32254350e-01 5.45918107e-01
-7.74033293e-02 5.82790434e-01 3.95066030e-02 -4.26589161e-01
-5.16221106e-01 -1.19066846e+00 -4.17347729e-01 6.86932385e-01
7.94492781e-01 -1.32827774e-01 -8.80154312e-01 -7.23332465e-01
9.77954343e-02 4.71550226e-01 -7.32747316e-01 3.03278528e-02
-4.75112051e-01 -6.87552214e-01 -6.97780773e-02 7.77296603e-01
1.12857461e+00 -7.82803178e-01 -5.96297681e-01 2.01048926e-02
-2.96707079e-02 -1.05362248e+00 -6.25070393e-01 1.08861081e-01
-1.22337461e+00 -9.56310034e-01 -7.10947931e-01 -8.73120070e-01
7.17957616e-01 4.97919947e-01 5.31110525e-01 -1.47163957e-01
-4.71450657e-01 8.10738385e-01 -2.91655183e-01 -1.36344537e-01
2.05899417e-01 -5.79541549e-02 -1.79759234e-01 2.96239048e-01
1.43161073e-01 -6.57265127e-01 -9.73176241e-01 4.25697118e-01
-1.16669190e+00 -3.74547876e-02 6.46301329e-01 6.73809886e-01
1.27624750e+00 3.08006629e-02 1.93595812e-01 -7.79218376e-01
1.47875160e-01 -3.06501806e-01 -6.80646002e-01 1.86017230e-01
-3.38942140e-01 1.14886597e-01 3.70318234e-01 -4.32100862e-01
-1.36276841e+00 8.98173228e-02 -1.88087091e-01 -3.83664906e-01
-3.66706014e-01 1.51302382e-01 -4.61819530e-01 -3.10611725e-01
3.81130099e-01 3.11029106e-01 -1.64368317e-01 -5.73652387e-01
5.51287055e-01 4.69026387e-01 3.98486257e-01 -7.95850933e-01
6.77354097e-01 8.01561773e-01 1.39674693e-01 -9.72707808e-01
-7.14134336e-01 -4.99581069e-01 -8.78201544e-01 -2.09120423e-01
1.41540837e+00 -8.29339862e-01 -7.03423858e-01 6.75800800e-01
-8.02251339e-01 -6.55646801e-01 -2.79944658e-01 6.50765181e-01
-5.69163918e-01 4.56407636e-01 -5.44454217e-01 -6.19731665e-01
-9.45081487e-02 -1.29379499e+00 1.09070122e+00 8.43112230e-01
4.69994158e-01 -1.29603779e+00 -1.48496628e-01 3.62908393e-01
3.03708076e-01 3.54968250e-01 7.96622038e-01 -2.03473553e-01
-1.15317738e+00 3.44700485e-01 -3.80750388e-01 6.40596330e-01
1.44416809e-01 1.30087463e-02 -9.92391169e-01 -1.65835153e-02
-1.75179671e-02 -9.81979743e-02 1.13253772e+00 7.54240394e-01
1.56652296e+00 1.66500464e-01 -4.54150170e-01 1.04203379e+00
1.55537450e+00 3.26791666e-02 4.59091842e-01 1.79396704e-01
1.01693082e+00 4.98912066e-01 7.97497392e-01 2.41637424e-01
5.28939128e-01 5.07650077e-01 6.97969079e-01 -1.56048805e-01
-1.12989716e-01 -1.21754542e-01 5.94942681e-02 5.51452875e-01
-2.81662166e-01 1.06264800e-01 -5.84660351e-01 4.48632926e-01
-1.92741656e+00 -7.69273877e-01 -2.12600902e-02 1.97633731e+00
5.90914428e-01 1.22963101e-01 -1.42426774e-01 -2.55445510e-01
5.86206436e-01 2.02802658e-01 -7.19441235e-01 -3.80516425e-02
-1.28088355e-01 1.84309557e-01 3.86355042e-01 7.33421326e-01
-1.31596458e+00 1.02017415e+00 4.83125877e+00 1.29392922e+00
-1.14473879e+00 1.46508992e-01 4.58874941e-01 1.34497434e-01
-5.87473333e-01 -1.93946958e-01 -9.60881114e-01 5.00548244e-01
8.16625729e-02 4.17773008e-01 1.27594545e-01 1.01756704e+00
2.23041281e-01 -3.80891234e-01 -6.49522543e-01 1.02752423e+00
-1.39823332e-01 -1.25859368e+00 1.09677166e-01 6.22882172e-02
8.05168509e-01 -1.25769377e-02 3.22237134e-01 3.75515446e-02
1.16338268e-01 -7.06186414e-01 6.97003782e-01 8.43573093e-01
4.84628052e-01 -9.10389125e-01 4.67806488e-01 2.71199554e-01
-1.53888083e+00 -2.50312418e-01 -6.84723556e-01 2.89470643e-01
-2.60630753e-02 8.16632509e-01 -3.61689925e-01 7.05057442e-01
9.50094521e-01 9.97933924e-01 -5.46430409e-01 1.04293776e+00
-3.53503019e-01 4.99355227e-01 -3.08745742e-01 -1.49301370e-03
4.32390749e-01 -4.90019679e-01 5.20372450e-01 1.25683570e+00
1.35873392e-01 2.60500997e-01 3.95007402e-01 1.03939033e+00
2.98102260e-01 9.51631069e-02 -3.18179071e-01 4.20829326e-01
3.22154343e-01 1.37313163e+00 -1.13332427e+00 -1.36888579e-01
-2.90657192e-01 1.13580537e+00 2.67456174e-01 6.14586949e-01
-6.66822195e-01 -3.05702657e-01 6.00392997e-01 -8.44607577e-02
6.57355785e-01 -5.60629606e-01 -2.75475442e-01 -1.07713914e+00
3.80179249e-02 -5.52689992e-02 3.33790302e-01 -5.56639016e-01
-1.07241738e+00 4.12024766e-01 1.91887870e-01 -9.90010798e-01
2.93112695e-01 -8.56720686e-01 -6.27221406e-01 1.01271248e+00
-1.73855734e+00 -1.44123161e+00 -6.42217875e-01 9.06066298e-01
6.46859288e-01 1.27222493e-01 5.03470004e-01 6.15920722e-02
-4.40219969e-01 3.43456000e-01 2.88049042e-01 7.81944543e-02
4.88743454e-01 -1.44963455e+00 -8.72441903e-02 8.17984819e-01
-1.04876538e-03 7.23513186e-01 1.75025731e-01 -6.44714832e-01
-1.07830143e+00 -1.31050813e+00 1.23616256e-01 -3.85719448e-01
1.61941677e-01 -3.37674290e-01 -9.51045513e-01 3.90882075e-01
-2.14508802e-01 3.31134319e-01 5.80382049e-01 -1.95581645e-01
-1.46094427e-01 -4.22587454e-01 -1.26023138e+00 3.45713645e-01
1.00534427e+00 -5.86765826e-01 -5.86270034e-01 7.68963844e-02
1.07873368e+00 -5.95560610e-01 -8.17453623e-01 4.46263134e-01
3.58061224e-01 -1.08753085e+00 1.09838367e+00 1.07860956e-02
1.10533558e-01 -7.60117114e-01 -4.40879405e-01 -1.03569436e+00
-2.30056360e-01 -8.12154487e-02 -1.59150869e-01 1.33361816e+00
-7.05898032e-02 -5.73529124e-01 7.59677231e-01 6.03895307e-01
-4.67791528e-01 -9.95707333e-01 -8.98242772e-01 -5.68689287e-01
-2.73300875e-02 -6.27182066e-01 4.74905491e-01 6.44258201e-01
-6.84578359e-01 -2.15762869e-01 2.13715687e-01 4.08404797e-01
7.23045170e-01 2.72960573e-01 5.10214984e-01 -1.10505807e+00
-4.01343465e-01 -3.21535915e-01 -5.98458171e-01 -1.50320256e+00
-8.32884610e-02 -8.97942781e-01 -3.25255767e-02 -1.66654539e+00
1.69137046e-01 -6.37460947e-01 -3.83134097e-01 2.68025845e-01
-4.63995397e-01 1.91055387e-01 7.34581575e-02 7.57206157e-02
-5.94920754e-01 7.25149691e-01 1.51012671e+00 -2.10859761e-01
-4.29283619e-01 2.25551531e-01 -3.93119961e-01 7.55201459e-01
5.85695982e-01 -4.34994232e-03 -7.81059384e-01 -4.57687855e-01
-1.12283394e-01 -4.17980433e-01 6.15292013e-01 -9.71351027e-01
1.25782669e-01 -2.83878863e-01 6.03325963e-01 -7.71662891e-01
2.97064155e-01 -8.21406543e-01 -3.76927316e-01 1.98499814e-01
1.89522337e-02 -5.39127529e-01 2.19472587e-01 9.30543065e-01
-1.70080155e-01 -1.63043082e-01 8.32809806e-01 -2.39692912e-01
-1.21411324e+00 5.44680655e-01 -1.90971196e-01 5.23869395e-02
1.07404482e+00 -4.70946878e-01 1.49733275e-01 1.98503658e-01
-8.11802447e-01 3.26094419e-01 5.60597599e-01 3.17171484e-01
8.64789248e-01 -1.22462201e+00 -1.35770395e-01 2.19761878e-01
1.26790866e-01 7.56974459e-01 7.65144706e-01 6.64133847e-01
-4.97374624e-01 6.24743439e-02 -7.92732090e-02 -1.05892217e+00
-1.02823353e+00 5.25279224e-01 4.68351245e-01 2.54788637e-01
-5.91253638e-01 1.18914855e+00 7.01204240e-01 -4.87396926e-01
2.73861885e-01 -8.77096355e-01 -4.59770501e-01 1.85032914e-04
5.57906806e-01 1.76086634e-01 -2.05224961e-01 -5.71694255e-01
-2.33411089e-01 1.14761603e+00 2.99770311e-02 -5.02831601e-02
1.22828519e+00 -5.17257154e-01 -1.69152424e-01 5.37820220e-01
9.90632653e-01 -3.26757208e-02 -1.80644774e+00 -3.70063961e-01
-3.93939227e-01 -5.49322248e-01 2.17884079e-01 -4.75351036e-01
-1.16011429e+00 1.17078769e+00 9.19107378e-01 -1.32883534e-01
1.17789924e+00 1.50248095e-01 7.19391465e-01 1.27101392e-01
4.60573345e-01 -1.20463192e+00 3.60443622e-01 5.20191431e-01
5.24251044e-01 -1.34060943e+00 -2.86366791e-01 -6.09541416e-01
-5.37176847e-01 1.25859833e+00 7.09005415e-01 -1.04163624e-01
8.56408715e-01 -1.69830129e-01 -6.39150068e-02 -1.36312723e-01
7.96207041e-02 -3.75313878e-01 6.18615389e-01 7.79289246e-01
4.72655743e-02 5.95692471e-02 -6.30420297e-02 7.97421396e-01
3.19299459e-01 -2.31191233e-01 -1.59343593e-02 7.65905261e-01
-7.10685730e-01 -9.79372561e-01 -3.38441521e-01 1.45496696e-01
-2.40642261e-02 9.07397717e-02 9.26599354e-02 5.94539881e-01
6.84848070e-01 8.21413815e-01 1.97501764e-01 -3.18227381e-01
2.13846937e-01 -1.84248835e-01 5.54063439e-01 -7.38299847e-01
-1.64463222e-01 1.46834999e-01 -7.35665798e-01 -8.07531774e-01
-6.99379027e-01 -5.43330431e-01 -1.67535460e+00 1.56187132e-01
-1.88328862e-01 -9.00496016e-05 6.38563514e-01 1.04902291e+00
2.54843891e-01 6.12228811e-01 5.40409505e-01 -1.09355259e+00
-6.38772845e-02 -6.22204900e-01 -6.72625303e-01 2.97378868e-01
2.81453967e-01 -8.88920188e-01 -2.47363791e-01 8.21744353e-02] | [9.356664657592773, -0.9767367839813232] |
531bb5a0-1bf0-4eb7-9e25-9b61f984a64f | lifelong-change-detection-continuous-domain | 2306.16086 | null | https://arxiv.org/abs/2306.16086v1 | https://arxiv.org/pdf/2306.16086v1.pdf | Lifelong Change Detection: Continuous Domain Adaptation for Small Object Change Detection in Every Robot Navigation | The recently emerging research area in robotics, ground view change detection, suffers from its ill-posed-ness because of visual uncertainty combined with complex nonlinear perspective projection. To regularize the ill-posed-ness, the commonly applied supervised learning methods (e.g., CSCD-Net) rely on manually annotated high-quality object-class-specific priors. In this work, we consider general application domains where no manual annotation is available and present a fully self-supervised approach. The present approach adopts the powerful and versatile idea that object changes detected during everyday robot navigation can be reused as additional priors to improve future change detection tasks. Furthermore, a robustified framework is implemented and verified experimentally in a new challenging practical application scenario: ground-view small object change detection. | ['Yoshimasa Nakamura', 'Kanji Tanaka', 'Koji Takeda'] | 2023-06-28 | null | null | null | null | ['change-detection', 'robot-navigation'] | ['computer-vision', 'robots'] | [ 4.50938165e-01 2.36120254e-01 6.61130846e-02 -2.59249806e-01
-3.13254833e-01 -4.64201093e-01 6.96576834e-01 -1.51513368e-01
-4.92278099e-01 8.15770090e-01 -3.25215787e-01 1.77489698e-01
-1.76544175e-01 -3.49250793e-01 -7.77292490e-01 -6.13601327e-01
1.18233375e-01 4.77435827e-01 6.84942901e-01 -2.77558148e-01
3.08472991e-01 5.62179923e-01 -1.67837739e+00 -3.29294235e-01
9.26122665e-01 6.15242660e-01 7.03913450e-01 3.97396684e-01
5.29746652e-01 4.66893554e-01 -1.04763480e-02 -3.23411170e-03
5.58617055e-01 -8.47213492e-02 -2.34620318e-01 4.89639968e-01
4.89433348e-01 -1.26233399e-01 -1.22532785e-01 1.45834255e+00
2.96329737e-01 3.12140197e-01 6.85158968e-01 -1.24087548e+00
1.85613669e-02 1.16222808e-02 -4.88796055e-01 4.75262441e-02
6.14183605e-01 3.17350358e-01 7.03959703e-01 -1.26775563e+00
1.15530622e+00 1.12078178e+00 8.00769091e-01 2.88177550e-01
-1.19953513e+00 -3.98651399e-02 5.81268013e-01 4.34754252e-01
-1.04299545e+00 -2.97235519e-01 1.17939353e+00 -6.87019706e-01
6.30665660e-01 4.33347858e-02 8.22542429e-01 9.74428415e-01
4.53295320e-01 8.27518106e-01 1.23979628e+00 -4.50537533e-01
3.97466451e-01 2.28604093e-01 -1.10639753e-02 8.15163553e-01
4.84251618e-01 7.60788843e-02 -3.20984066e-01 9.23831984e-02
8.40810657e-01 2.25475460e-01 -6.00416064e-01 -1.41147661e+00
-1.45218372e+00 3.89030874e-01 4.73486215e-01 1.25570446e-01
-3.45183790e-01 2.96281930e-02 1.20543756e-01 1.51064843e-01
3.89761508e-01 2.82901317e-01 -5.75430512e-01 -3.69676724e-02
-5.38377702e-01 1.05010599e-01 6.64434075e-01 1.33728719e+00
9.29263532e-01 3.64240408e-02 4.03739870e-01 5.62424123e-01
6.83211625e-01 5.95993102e-01 4.37592566e-01 -9.35777426e-01
3.67036700e-01 5.14837980e-01 4.69863653e-01 -1.09603608e+00
-6.50376916e-01 -4.81409490e-01 -7.74582326e-01 8.41528177e-01
4.79231626e-01 5.36730140e-02 -9.57644224e-01 1.40988350e+00
5.58056831e-01 -4.26374050e-03 2.46856939e-02 9.93348718e-01
2.89599627e-01 1.91631481e-01 -4.32345808e-01 -2.09648803e-01
1.05411839e+00 -9.30540085e-01 -6.76445186e-01 -5.93132496e-01
1.51952520e-01 -6.40132785e-01 1.12462258e+00 7.38298893e-01
-7.75932133e-01 -4.99503225e-01 -1.36249721e+00 6.88693300e-02
-3.42655092e-01 2.00028226e-01 2.55947560e-01 3.03340077e-01
-7.09613979e-01 4.67930406e-01 -9.53372240e-01 -6.97855532e-01
1.89156592e-01 2.34284610e-01 -6.51029348e-01 -3.22618842e-01
-6.29558623e-01 1.18879628e+00 4.57129866e-01 3.81814539e-01
-7.40507662e-01 -4.11094099e-01 -9.72182751e-01 -4.24440205e-01
9.19960976e-01 -9.60681379e-01 1.17744744e+00 -8.36141586e-01
-2.08241987e+00 9.31271136e-01 1.43065616e-01 -3.46739113e-01
1.12568378e+00 -4.79627490e-01 3.81698757e-02 2.32364219e-02
-1.02944784e-01 4.50885445e-01 1.15577698e+00 -1.68720138e+00
-6.54391766e-01 -5.36649585e-01 3.44439954e-01 6.30651236e-01
1.68282226e-01 -5.65177023e-01 -3.13334197e-01 -5.27227640e-01
8.33543420e-01 -1.31477189e+00 -5.93333781e-01 3.88653845e-01
-1.91682622e-01 2.25540400e-01 9.24642026e-01 -4.37939018e-01
3.30236644e-01 -1.90362084e+00 4.26682770e-01 1.36354789e-01
2.44008414e-02 -1.38861435e-02 2.94299036e-01 2.63222605e-01
2.37227157e-01 -6.26339078e-01 -5.79198599e-01 -2.72043735e-01
5.91201186e-02 3.06187898e-01 5.92445247e-02 9.93606925e-01
1.05054408e-01 4.60756510e-01 -1.13790989e+00 -2.17226729e-01
7.46063888e-01 2.76053131e-01 -5.61357498e-01 1.79985225e-01
-3.97427320e-01 8.05101991e-01 -8.30295086e-02 5.04181206e-01
7.34118223e-01 8.13837349e-02 -2.87528392e-02 -2.04144880e-01
-3.70130122e-01 -1.69602722e-01 -1.63285387e+00 1.99254310e+00
-5.97780228e-01 5.93159080e-01 2.37079903e-01 -9.61918235e-01
7.83671975e-01 -1.57260045e-01 4.35108721e-01 -8.92371535e-02
-2.92984601e-02 3.07705790e-01 -1.63626999e-01 -7.02744663e-01
5.64821303e-01 -1.21542566e-01 7.25484341e-02 -1.52630165e-01
-3.38319060e-03 -6.85492992e-01 -8.78741816e-02 -1.23626709e-01
1.03889716e+00 8.11500192e-01 9.91235197e-01 -2.82902360e-01
6.90645099e-01 6.43457919e-02 6.71198428e-01 4.94297951e-01
-3.19953024e-01 8.36249232e-01 -8.73878300e-02 -2.67973155e-01
-7.11352468e-01 -9.55624759e-01 -1.68990076e-01 5.82155228e-01
7.31514335e-01 1.56472158e-02 -3.68782043e-01 -7.85428107e-01
-3.47437221e-03 4.99584258e-01 -3.01470369e-01 -5.37361912e-02
-6.42608941e-01 -5.14084756e-01 -1.67025730e-01 2.99796909e-01
5.17806888e-01 -8.68894339e-01 -1.19693017e+00 2.83526421e-01
-1.37847394e-01 -1.25638187e+00 -1.43526390e-01 2.92471230e-01
-9.74819064e-01 -1.19750905e+00 -8.17568064e-01 -6.32603824e-01
8.05729389e-01 5.79988539e-01 7.99408197e-01 -5.39877892e-01
-2.08843276e-01 1.00807106e+00 -2.84119517e-01 -3.70887309e-01
-2.99607486e-01 -3.23984265e-01 2.49702647e-01 3.56671885e-02
-1.56328440e-01 -7.35167027e-01 -3.65783364e-01 4.14823711e-01
-6.51527286e-01 -7.04125082e-03 8.11872661e-01 9.16694760e-01
8.28952909e-01 -8.72508213e-02 2.42847398e-01 -8.07505965e-01
-9.20375362e-02 -2.69771159e-01 -9.58128214e-01 1.29600555e-01
-7.39750445e-01 -1.18208967e-01 3.29147577e-01 -5.50024748e-01
-1.32443297e+00 6.19326711e-01 5.42762019e-02 -3.72510195e-01
-4.87066776e-01 6.01554513e-01 -5.05986392e-01 -3.52945447e-01
6.91040874e-01 6.02378175e-02 -1.04409508e-01 -2.89784998e-01
5.67464292e-01 2.74225533e-01 6.97157919e-01 -1.75048679e-01
1.02095044e+00 7.75438130e-01 1.12123974e-01 -9.56785560e-01
-4.34112847e-01 -7.98987687e-01 -1.08117783e+00 -6.43729091e-01
5.49459398e-01 -1.11369264e+00 -2.63199419e-01 5.67549825e-01
-1.18651247e+00 -3.44578683e-01 -4.47445542e-01 7.45156765e-01
-1.18999255e+00 7.67000914e-01 -1.37896791e-01 -9.29223955e-01
1.37528822e-01 -8.91965330e-01 1.06591523e+00 4.87240702e-02
1.75281882e-01 -8.21597040e-01 1.23853035e-01 8.05535614e-02
-1.12732172e-01 4.33679938e-01 2.71369010e-01 -2.43580490e-01
-8.63091886e-01 -1.24256581e-01 -3.39515284e-02 3.04598421e-01
2.18859538e-01 -1.92677677e-01 -9.24130201e-01 -3.08565319e-01
4.08002049e-01 4.93012331e-02 6.71287000e-01 3.94575924e-01
3.17004383e-01 9.29095000e-02 -3.35135102e-01 4.22005653e-01
1.61513352e+00 8.02171528e-02 3.05734813e-01 5.50056815e-01
7.36214161e-01 7.36226380e-01 1.08114243e+00 5.89421868e-01
4.02905643e-01 8.94035995e-01 8.27575803e-01 3.91376346e-01
-8.30567554e-02 -2.46570427e-02 5.61033130e-01 7.89124846e-01
-2.01848999e-01 -2.43616372e-01 -1.01289773e+00 3.36526155e-01
-2.25407267e+00 -6.81437910e-01 -2.40618184e-01 2.15735412e+00
3.64705741e-01 4.01510507e-01 -2.71021456e-01 3.51155192e-01
6.46044075e-01 2.90742945e-02 -7.15074539e-01 5.23279309e-01
-4.33629490e-02 -6.62821889e-01 4.78085190e-01 6.10928953e-01
-1.17038035e+00 7.64120638e-01 5.48537970e+00 2.75467008e-01
-9.17500377e-01 1.62825868e-01 -3.45015794e-01 3.22138399e-01
6.28857538e-02 -9.08721462e-02 -6.88813329e-01 1.05859965e-01
1.07738122e-01 1.71530575e-01 2.18069777e-01 9.29321647e-01
1.50555789e-01 -6.31499171e-01 -1.11324930e+00 1.16699433e+00
5.50210178e-01 -8.36029410e-01 -5.43108821e-01 -1.45792633e-01
7.79662192e-01 1.60903662e-01 -4.06008177e-02 -4.61885966e-02
2.48958781e-01 -1.11381225e-02 9.00560796e-01 7.64875770e-01
4.27188009e-01 -2.45088071e-01 7.31300175e-01 6.35365665e-01
-1.02269983e+00 -2.12134302e-01 -3.18505824e-01 -1.70121923e-01
4.82151896e-01 8.92414808e-01 -1.04097033e+00 9.17570651e-01
6.84915364e-01 1.34167659e+00 -4.86605287e-01 1.28696871e+00
-3.51201177e-01 1.33434862e-01 -6.69446766e-01 1.70459434e-01
-7.62111926e-03 -3.79233688e-01 1.23239207e+00 1.00021863e+00
3.26957077e-01 -5.67352548e-02 3.33089888e-01 6.66801333e-01
4.37801927e-01 -3.93449850e-02 -9.76558268e-01 6.48227930e-01
-3.46568972e-02 1.23731554e+00 -1.08494723e+00 -6.47506863e-02
-4.11061972e-01 1.07636046e+00 9.13057774e-02 4.30228800e-01
-5.54343700e-01 -7.31071979e-02 3.24850291e-01 8.12027529e-02
5.18138111e-01 -5.35525620e-01 -2.37974793e-01 -1.53085268e+00
3.68471473e-01 -4.60944831e-01 4.23386171e-02 -1.03267610e+00
-1.25418162e+00 2.86538959e-01 1.62702084e-01 -1.82432926e+00
-3.11321139e-01 -7.59046376e-01 -2.04440221e-01 2.95272499e-01
-1.67737937e+00 -1.20985436e+00 -6.57591820e-01 6.82867765e-01
9.02099431e-01 -4.37583253e-02 3.91545534e-01 1.43856719e-01
-2.49364302e-01 -1.90089680e-02 2.10551977e-01 -5.57245970e-01
8.34674299e-01 -1.45661056e+00 -1.79222390e-01 1.15842772e+00
8.97724628e-02 2.35463157e-01 1.15948176e+00 -8.10932577e-01
-1.42947495e+00 -1.07279074e+00 2.21937016e-01 -4.29834247e-01
6.49701536e-01 -3.15449089e-01 -7.62461722e-01 8.76867175e-01
-1.14627041e-01 3.46338600e-01 2.97440751e-03 -1.83241367e-01
4.28651413e-03 -3.15594301e-02 -1.27861857e+00 5.76151252e-01
1.31394470e+00 -2.70933986e-01 -8.90236259e-01 3.61924797e-01
4.18273062e-01 -7.09711254e-01 -5.24972916e-01 6.87108457e-01
4.24903005e-01 -8.98442924e-01 9.02087152e-01 -1.61635038e-02
-1.48627818e-01 -6.27354026e-01 -3.03671241e-01 -1.48954391e+00
-1.61650792e-01 -6.39179170e-01 -2.27609366e-01 9.07877862e-01
-1.67191908e-01 -7.11178541e-01 7.18702495e-01 2.21347153e-01
-5.08858621e-01 -2.93215126e-01 -1.07899261e+00 -9.38833296e-01
-5.12504756e-01 -4.10981685e-01 -1.84953034e-01 8.27521384e-01
1.57420877e-02 2.39413485e-01 -3.63731712e-01 5.94352365e-01
6.90577328e-01 -8.55389014e-02 1.00643289e+00 -1.67992580e+00
-1.80055931e-01 -7.43826628e-02 -8.83017123e-01 -1.26130497e+00
1.51463285e-01 -5.49986362e-01 8.39401126e-01 -1.43914461e+00
-2.79708415e-01 -3.39552939e-01 -1.08342402e-01 -8.38391259e-02
1.97439436e-02 1.65469930e-01 2.89320320e-01 1.79541811e-01
-8.10967863e-01 8.51083398e-01 1.04263198e+00 -8.73522907e-02
-3.61314982e-01 3.62913936e-01 9.20252800e-02 1.37750185e+00
5.61685383e-01 -3.91543835e-01 -6.03798568e-01 -9.73957628e-02
3.89172912e-01 -1.77143544e-01 4.96288449e-01 -1.36881506e+00
4.31989253e-01 -1.49898604e-01 1.14259838e-05 -8.08913827e-01
6.08696282e-01 -1.48437035e+00 2.02993944e-01 5.52839339e-01
1.14536650e-01 1.88732166e-02 -3.62077989e-02 1.24524033e+00
-9.43147466e-02 -6.00724876e-01 8.08545291e-01 -3.46537054e-01
-9.81967986e-01 9.78639424e-02 -6.58438385e-01 -1.31947264e-01
1.24409783e+00 -4.13375914e-01 1.56892136e-01 -2.84462929e-01
-8.44066083e-01 -6.08971342e-02 7.85502732e-01 3.29870731e-01
7.31460273e-01 -1.12265372e+00 -2.09966689e-01 2.23350689e-01
5.36161900e-01 3.80533665e-01 1.30860120e-01 1.03089595e+00
-4.77197319e-01 7.27298260e-02 -3.57711852e-01 -1.05919409e+00
-1.12978470e+00 7.10438073e-01 2.66165435e-01 -4.94214669e-02
-1.12483680e+00 4.97569799e-01 8.03989470e-02 -7.49626756e-01
2.78533846e-01 -7.46107936e-01 -2.20962599e-01 -3.45750414e-02
1.32872492e-01 6.11068130e-01 3.37705798e-02 -6.80533111e-01
-3.41888487e-01 8.60646486e-01 3.37428063e-01 -3.19836974e-01
1.31716287e+00 -7.00629890e-01 1.34873867e-01 9.66622531e-01
7.00115144e-01 -1.15745135e-01 -1.77311611e+00 -5.40994465e-01
3.76603246e-01 -4.66855556e-01 -1.70103610e-01 -5.51697314e-01
-6.32281125e-01 6.91582382e-01 8.74287307e-01 -9.67383906e-02
8.48698020e-01 -1.66041493e-01 1.61005482e-01 1.02472270e+00
9.94002163e-01 -1.13946116e+00 1.55973703e-01 6.43627107e-01
1.19707108e+00 -1.65189373e+00 3.08811426e-01 -8.42788875e-01
-4.72381145e-01 1.14169407e+00 7.53573716e-01 -3.46269131e-01
8.45701754e-01 6.07821308e-02 1.40466867e-03 2.57644057e-02
-2.75983930e-01 -3.04823816e-01 1.91722780e-01 9.52441454e-01
-3.59988153e-01 -2.89396584e-01 -2.07117617e-01 2.45650977e-01
1.39555678e-01 6.56211674e-02 8.27306628e-01 1.29391146e+00
-4.66003239e-01 -6.54543161e-01 -4.22610909e-01 9.25692916e-02
1.46836460e-01 4.01298344e-01 -1.26946196e-01 9.83831108e-01
2.57154286e-01 6.25729442e-01 -3.47423315e-01 -2.18544394e-01
5.76934695e-01 -5.92455082e-02 8.73485684e-01 -7.60585189e-01
-1.03233643e-01 1.55641198e-01 -1.45599648e-01 -6.38917625e-01
-1.18922615e+00 -1.06635582e+00 -1.10082066e+00 5.70996404e-01
-4.29396898e-01 -4.69050527e-01 7.51280665e-01 1.03006768e+00
-3.11194872e-03 3.97806227e-01 3.87726784e-01 -1.51069224e+00
-4.61340129e-01 -9.58633602e-01 -5.09245455e-01 1.22263283e-01
5.58664024e-01 -1.30062079e+00 -4.95321453e-01 2.97419041e-01] | [7.6480631828308105, -2.423809766769409] |
0c986cdd-0ddc-43cc-9392-a1680f969714 | a-challenging-benchmark-of-anime-style | 2204.14034 | null | https://arxiv.org/abs/2204.14034v1 | https://arxiv.org/pdf/2204.14034v1.pdf | A Challenging Benchmark of Anime Style Recognition | Given two images of different anime roles, anime style recognition (ASR) aims to learn abstract painting style to determine whether the two images are from the same work, which is an interesting but challenging problem. Unlike biometric recognition, such as face recognition, iris recognition, and person re-identification, ASR suffers from a much larger semantic gap but receives less attention. In this paper, we propose a challenging ASR benchmark. Firstly, we collect a large-scale ASR dataset (LSASRD), which contains 20,937 images of 190 anime works and each work at least has ten different roles. In addition to the large-scale, LSASRD contains a list of challenging factors, such as complex illuminations, various poses, theatrical colors and exaggerated compositions. Secondly, we design a cross-role protocol to evaluate ASR performance, in which query and gallery images must come from different roles to validate an ASR model is to learn abstract painting style rather than learn discriminative features of roles. Finally, we apply two powerful person re-identification methods, namely, AGW and TransReID, to construct the baseline performance on LSASRD. Surprisingly, the recent transformer model (i.e., TransReID) only acquires a 42.24% mAP on LSASRD. Therefore, we believe that the ASR task of a huge semantic gap deserves deep and long-term research. We will open our dataset and code at https://github.com/nkjcqvcpi/ASR. | ['Huanqiang Zeng', 'Jianqing Zhu', 'Tianchen Chen', 'Xiao Yang', 'Kailin Lyu', 'Shengtao Guo', 'Haotang Li'] | 2022-04-29 | null | null | null | null | ['art-analysis'] | ['computer-vision'] | [ 0.1969417 -0.5608254 0.1942494 -0.4957109 -0.5241314 -0.6595017
0.6804162 -0.6340627 -0.25762466 0.559555 0.2113811 0.333817
0.05849309 -0.6960617 -0.6553064 -0.5520225 0.43603775 0.5267313
-0.19815005 -0.39986387 0.17303157 0.49172196 -1.5304401 0.40061748
0.51427066 1.0598004 -0.1978903 0.43427143 0.18476234 0.4801709
-1.0998152 -1.1938951 0.51530004 -0.5583242 -0.68003446 0.27691948
1.0221636 -0.38207307 -0.7090135 1.1072863 0.843736 0.01451882
0.43218842 -1.5349562 -1.2268841 0.2955474 -0.83003545 0.09930986
0.6308801 0.571287 0.68674064 -1.0479933 0.5729771 1.7251822
0.6778982 0.91554856 -1.1617682 -1.2070595 0.01094058 0.24282031
-1.6320977 -0.5184198 0.89012873 -0.22709908 0.02715573 0.55016273
0.68324614 1.6727253 -0.469412 0.9338887 1.6332557 -0.03174334
-0.20425563 0.31698 0.09353455 0.5401433 0.19936138 -0.17455955
-0.66160893 0.07159246 0.879931 0.42779437 -0.46030593 0.12819965
-1.3185639 0.4631984 0.41140658 0.20060119 0.0960766 -0.01328618
0.2264464 0.5143244 0.07757177 0.5177309 -0.06010386 -0.07565336
-0.6236924 0.20478643 0.44703835 1.0191171 0.5993989 -0.2256689
-0.42737642 1.327221 0.02216926 0.89744025 0.40477023 -0.8690102
0.5323663 0.8479969 0.06959949 -1.1952488 0.11526051 -0.29542333
-1.1577424 -0.02551682 0.64332956 0.1259699 -0.85798275 1.6329185
0.06902953 0.24267437 -0.10631118 1.1559006 1.0973195 0.4916631
-0.1841658 0.19638324 1.7375062 -1.035439 -0.5512935 -0.06119447
0.0323946 -0.8956543 1.3998804 0.43897375 -0.99170035 -0.91644686
-0.8156619 -0.14330208 -0.46341142 0.61944103 0.29907563 0.7588554
-0.9521733 0.32698724 -0.05587029 -0.4844406 0.6022141 0.10126545
-0.5112422 -0.48938116 -1.22908 0.5587303 -0.0753006 0.330408
-0.91631985 -0.713595 -0.6722705 -0.36730003 0.42681113 -0.5193159
0.6669079 -0.9081239 -1.3110415 1.515047 -0.1838611 0.03421003
0.8934674 -0.26616365 -0.81620705 0.11612931 0.1083108 0.48976812
0.9826408 -1.4108223 -0.28743392 -0.82990223 0.21498847 0.06971273
-0.6243922 0.2248536 -0.9487062 -1.056873 -0.1245683 -1.1496557
0.36927083 0.20588078 -0.45035872 -0.41010728 0.6405633 -0.92578655
1.0614784 -2.211454 0.26719627 0.0830254 0.26348385 0.24021229
-0.3261585 0.08900431 -0.19814733 0.25164366 -0.13273264 -0.44209746
0.05436941 -0.1088907 -0.4091936 0.34971347 0.17915629 1.0620387
-0.63544714 -0.4060184 0.06499974 0.37018868 -0.02663369 0.40939602
0.314858 0.40979317 -0.26331675 1.0836054 0.8041221 -0.3009384
-0.11512925 -0.5778632 0.37748325 -0.42393452 -1.2117952 1.7288706
-0.21953121 0.6097715 -0.12333191 -0.62558836 1.146936 -0.02463679
0.22797547 -0.9374576 -0.03437445 0.17840728 -0.40589583 -0.34970158
0.4695448 0.05754544 -0.30494365 0.36844558 -0.18984501 0.11540692
0.12242509 0.2529859 0.78568834 0.03926466 -0.22263293 -0.05414369
0.71394867 -0.34877303 0.48956162 0.7739182 -0.39286754 1.0319877
0.2173353 -0.55724543 -0.9559554 -1.1424419 -0.18864653 0.97018546
0.70575887 -0.31956652 -0.68707544 -0.75071967 0.14375493 0.3192041
-0.76795465 -0.11160975 -0.76188254 -0.8606249 0.96133536 0.32855862
1.0192206 -0.9751818 0.07076864 -0.40760964 -0.3230521 -1.1484314
-1.043423 -0.9435678 -0.26759484 -1.3261259 -1.272682 -0.6913564
0.57065225 0.48547554 1.3276652 0.35415778 -0.5053222 0.55930686
-0.29842606 -0.15485041 -0.12366793 -0.19501074 0.18630216 0.5278837
0.6623305 -0.2930264 -0.94364786 1.0180373 -0.6668631 -0.02931767
0.5421752 0.7724267 0.49828666 0.07933746 0.42744437 -0.800296
0.56623006 -0.2593171 -0.23301151 0.70989144 -0.31134745 -0.43962145
0.5114602 -0.5972735 -1.033515 -0.3443674 0.09180976 -0.7449171
-0.3183874 -0.4085707 -0.56886226 0.02771823 0.51134133 0.52855676
0.0413413 -0.6408929 0.09103125 0.8911398 0.62849486 -0.8637278
1.1431926 0.4760066 -0.35828978 -0.7217221 -0.776511 -0.3349111
-0.3715555 -0.38943115 0.7508413 -1.2836491 -0.9009708 1.0664135
-1.0730418 -0.15805717 -0.15109545 0.05391097 -0.10153795 0.46162638
-0.617435 -0.7190257 -0.30408445 -1.0752501 1.2729499 0.5083719
-0.14076422 -0.45995364 -0.37956685 1.0872186 0.15414125 0.29378048
0.3614883 -0.30444312 -0.5946757 -0.12005623 -0.548973 0.3057016
0.31374618 -0.14209062 -1.0476496 -0.5837202 -0.08358542 -0.630501
0.7984041 -0.3071054 1.6471694 -0.24330127 -0.08163295 0.763283
1.0965277 0.20186397 0.9358196 0.14494745 1.0404085 0.77497154
0.6124949 0.2541967 0.32379037 0.8492754 0.03572505 -0.23476739
-0.45596224 -0.37493452 0.41011238 0.5103411 -0.4731732 -0.18893297
-0.69627094 0.3095562 -1.5096989 -1.0040207 -0.02031275 2.1870773
0.9360935 -0.16824418 0.46442133 0.01672725 1.0643114 0.19911456
-0.8205688 0.27411702 -0.5238024 0.01144762 0.27191064 -0.06410862
-0.86519706 0.83815545 5.0925045 1.1323001 -0.8648075 0.06331538
0.9718972 -0.24277098 -0.19342768 -0.4616362 -0.8944358 1.0475041
0.2616624 0.21558271 0.84046346 0.5979791 -0.21206869 0.31237245
-1.0332124 1.7868855 0.6791913 -0.9214636 0.32551718 -0.04004413
0.5635526 -0.5869275 0.36568317 0.36076325 0.21408772 -1.298988
0.5657702 0.45937026 1.2072369 -0.61541563 0.5708186 -0.17071824
-1.2856952 -0.13155936 -0.35085648 0.29248518 -0.11382083 0.27733982
-0.07886826 0.5850073 1.1550463 1.0981214 -1.1648551 0.69498146
-0.20349096 0.32133886 -0.05052906 0.22440267 -0.34891745 -0.277038
0.5725436 0.90177745 0.24207495 0.06067714 -0.01053639 1.0207908
-0.47251645 -0.2630744 -0.43854564 -0.06847427 0.5793222 1.1640382
-0.37908295 -0.31502864 -0.3263954 1.5494303 0.10214937 0.51601624
-1.0158962 -0.21344174 1.0136601 0.17893966 -0.11431155 0.08252278
-0.0084752 -1.6472032 0.3155179 -1.3159734 0.609035 -0.98562264
-1.8556329 0.57230693 -0.19330887 -1.2428429 0.3321064 -0.7584799
-0.46414867 0.91330564 -1.3208847 -1.4960141 -0.8807261 0.9578884
0.6801225 -0.5860942 0.4956177 0.65342516 -0.9137069 1.1737834
-0.17909354 0.7605368 1.1103724 -1.1212839 0.499182 0.75600374
0.12495928 0.80156523 0.3948426 -0.5446004 -1.6376551 -1.2017535
0.28067067 -1.0279005 0.48012158 -0.62275696 -0.86418766 0.5806154
-0.03134065 0.03299757 0.54448134 0.05488119 -0.70578986 -0.48513615
-1.1108285 0.63087404 1.6236 -0.68871623 -0.50535285 0.2672629
0.5982772 -0.27696657 -1.1292195 0.33156398 0.77972376 -0.9319524
1.3622587 -0.4784264 0.6509587 -0.38901514 -0.22012673 -1.0028603
-0.18657511 -0.6011536 0.07300434 1.7039349 -0.04704787 -0.76858264
0.8470737 0.72381943 0.42565775 -0.68141764 -0.5355375 -0.9125137
-0.25959304 0.13211131 1.0180372 1.1711544 -0.9121181 0.3627223
-0.8153278 0.01665401 0.9058989 0.51793355 1.204695 -1.1458614
-0.2894783 -0.38242045 -0.4240047 -0.9272336 0.06532769 -0.7338507
-0.4294896 -1.0392998 0.5824805 -0.6400688 -0.38130537 0.44356683
-0.37599817 0.9317089 0.42912742 0.5744925 -0.70367026 0.7292464
1.6212213 -0.58798915 0.09802615 -0.0430893 -0.90478647 0.5378353
0.70988894 -0.09682569 -0.29711637 -0.52816343 -0.02945082 -0.27611685
0.8590942 -0.80501276 0.02284234 -0.04821161 0.9200157 -0.45184222
0.5485938 -0.67005676 0.43427148 0.23655634 -0.29539636 0.22588517
-0.1326375 0.5917315 -0.17379226 0.0864925 0.80955404 -0.25608793
-0.93608 0.5519343 0.47416568 0.37405184 0.9569275 -0.5523261
-0.5774583 -0.17089841 -0.39958817 0.33063212 0.8335176 0.89438754
0.7705939 -1.5576566 -1.0754607 0.18362956 0.42316437 -0.04088052
0.55537707 0.4741574 -0.13104343 -0.14502089 -0.32578713 -0.49472785
-1.62932 0.51693183 0.3061008 0.12487616 -0.5843559 0.89736766
0.3616102 -0.52265304 0.15227975 0.47881478 -0.28629413 0.21418834
0.7802435 0.4913016 -0.34101442 -0.8815672 -0.33681607 0.81555414
-0.20588769 0.25993773 1.0526736 -0.09671699 -0.29286936 0.20928755
1.2518044 0.02559259 -1.1552955 -0.35188302 -0.3788646 -1.0838169
-0.55978036 -0.9927497 -1.3356946 0.7508262 0.93487257 -0.10426919
1.179122 0.12994643 1.0262448 0.17427146 0.55131614 -1.1264247
0.5945343 0.0635386 1.2772446 -1.4545888 0.02565931 -0.33156794
-0.7578362 0.7292839 1.0665698 0.07159463 0.17398313 -0.33750117
0.05125921 -0.13522527 -0.08853127 -0.03480701 0.39538246 0.4661053
0.08913945 0.10237174 0.10504769 0.69588363 -0.50411814 -0.15759441
0.20721969 0.40325302 0.3146746 -1.0245322 -0.6263586 0.3928255
-0.23726039 0.17713326 -0.82107705 0.7592031 0.28739277 0.8761864
0.03925643 -0.68171996 0.5200317 -0.28392556 0.5818864 -0.37232053
-0.5663574 -0.50655967 -0.0580945 -0.58568525 -0.5199945 -0.4571573
-0.5680416 -0.6185723 0.14301999 -0.07070257 0.3032738 0.5792166
0.26670402 0.31949022 0.81808376 -0.5375324 -0.5408821 -0.7773545
-0.6797381 1.2327102 -0.03103963 -0.74156874 -0.23717262 -0.0489229 ] | [14.622374534606934, 0.9457252621650696] |
fc560f6a-8260-4454-831d-f057791a8d77 | do-ssl-models-have-deja-vu-a-case-of | 2304.13850 | null | https://arxiv.org/abs/2304.13850v2 | https://arxiv.org/pdf/2304.13850v2.pdf | Do SSL Models Have Déjà Vu? A Case of Unintended Memorization in Self-supervised Learning | Self-supervised learning (SSL) algorithms can produce useful image representations by learning to associate different parts of natural images with one another. However, when taken to the extreme, SSL models can unintendedly memorize specific parts in individual training samples rather than learning semantically meaningful associations. In this work, we perform a systematic study of the unintended memorization of image-specific information in SSL models -- which we refer to as d\'ej\`a vu memorization. Concretely, we show that given the trained model and a crop of a training image containing only the background (e.g., water, sky, grass), it is possible to infer the foreground object with high accuracy or even visually reconstruct it. Furthermore, we show that d\'ej\`a vu memorization is common to different SSL algorithms, is exacerbated by certain design choices, and cannot be detected by conventional techniques for evaluating representation quality. Our study of d\'ej\`a vu memorization reveals previously unknown privacy risks in SSL models, as well as suggests potential practical mitigation strategies. Code is available at https://github.com/facebookresearch/DejaVu. | ['Casey Meehan', 'Kamalika Chaudhuri', 'Pascal Vincent', 'Florian Bordes', 'Chuan Guo'] | 2023-04-26 | null | null | null | null | ['memorization'] | ['natural-language-processing'] | [ 8.07562113e-01 5.52765906e-01 -1.40212610e-01 -2.43972048e-01
-5.78518331e-01 -7.07915246e-01 5.21013021e-01 5.18164814e-01
-5.72713390e-02 7.36208141e-01 8.57244655e-02 -4.49772954e-01
-1.65514082e-01 -8.62611055e-01 -1.35536361e+00 -6.52677476e-01
-1.16989061e-01 -1.53684646e-01 -9.47829112e-02 3.51836294e-01
2.72166193e-01 4.25029308e-01 -1.65254033e+00 3.33704680e-01
6.74472511e-01 1.10219169e+00 1.23272300e-01 7.47806430e-01
-7.40380213e-02 7.94328928e-01 -7.16102898e-01 -4.74766672e-01
4.07586843e-01 -4.91815299e-01 -8.26022983e-01 2.26992965e-01
1.12762868e+00 -3.98599684e-01 -2.67354459e-01 1.16259706e+00
2.74992973e-01 -1.63866103e-01 6.67103648e-01 -1.30317521e+00
-8.59568357e-01 5.20260394e-01 -5.76859713e-01 1.23931793e-03
5.70571780e-01 2.52601385e-01 8.29635561e-01 -5.95680654e-01
7.23858953e-01 1.12934816e+00 7.18889534e-01 3.69116873e-01
-1.74089205e+00 -1.06776488e+00 1.00179859e-01 1.23427063e-01
-1.64468455e+00 -6.02621078e-01 5.32644749e-01 -3.24536026e-01
4.63276356e-01 5.10901809e-01 3.28106999e-01 1.25163162e+00
2.86703408e-01 1.05801249e+00 1.30390322e+00 -6.00807846e-01
2.70204574e-01 5.00036240e-01 4.12853152e-01 9.34546471e-01
6.62846208e-01 2.01533943e-01 -8.17767620e-01 -3.74118119e-01
4.62663382e-01 -1.37064278e-01 -3.50749195e-01 -4.76455450e-01
-8.03773999e-01 6.60249233e-01 4.12761033e-01 -9.52722505e-03
-6.59953952e-02 2.27886438e-01 7.78943449e-02 3.37688237e-01
4.95317221e-01 5.05266011e-01 -3.63164246e-01 5.07746816e-01
-9.28554177e-01 -1.40280336e-01 6.63665175e-01 1.17951095e+00
1.11381543e+00 -9.14446414e-02 -1.25373527e-01 4.35911506e-01
9.37049091e-02 6.61583424e-01 1.76785827e-01 -9.71358240e-01
6.59466162e-02 1.82960972e-01 8.10237825e-02 -1.16180110e+00
-1.30876619e-02 -4.37855154e-01 -9.60290432e-01 3.55152905e-01
3.78104240e-01 -9.52169299e-03 -7.64666975e-01 1.86521626e+00
-1.38741136e-01 3.15213501e-01 1.19022802e-01 3.38675976e-01
8.52754176e-01 4.81048614e-01 1.42715946e-01 -3.18100393e-01
1.23447502e+00 -5.12975991e-01 -5.75394213e-01 -4.02334541e-01
4.48381394e-01 -7.43626237e-01 9.44665670e-01 4.09985185e-01
-1.07168210e+00 -6.55420899e-01 -1.10732651e+00 1.15638770e-01
-3.85224253e-01 8.76796767e-02 5.87025762e-01 1.09909523e+00
-1.19868326e+00 5.32058120e-01 -4.00233507e-01 -5.52747130e-01
8.71668816e-01 1.54405236e-01 -4.78675783e-01 -3.43401521e-01
-9.45780754e-01 8.02184224e-01 2.99380183e-01 -3.39290828e-01
-8.54380131e-01 -7.87170529e-01 -1.10772026e+00 1.18138250e-02
4.82680470e-01 -4.65557873e-01 9.43320572e-01 -1.09284973e+00
-7.60270298e-01 1.09965718e+00 -5.00787854e-01 -7.38651991e-01
5.00677407e-01 -4.58240569e-01 -4.89999384e-01 1.66973531e-01
2.80525554e-02 6.23365343e-01 1.33058834e+00 -1.83714163e+00
-3.29277694e-01 -3.31876546e-01 -8.79768189e-03 -1.45876437e-01
-2.92655349e-01 -2.65901923e-01 -2.34538596e-02 -7.14428008e-01
1.18297227e-01 -7.32400954e-01 -1.32254884e-01 5.54760993e-01
-7.68760741e-01 3.58812749e-01 8.90981793e-01 -6.07415617e-01
1.14498210e+00 -2.49711084e+00 -5.24974227e-01 4.92435068e-01
4.22241032e-01 6.49888813e-01 -1.94312155e-01 3.69663566e-01
-1.59414366e-01 3.63342166e-01 -4.04147595e-01 -4.92894471e-01
-1.97329834e-01 2.82389641e-01 -5.32444239e-01 5.79974651e-01
1.58118024e-01 6.93242311e-01 -1.00341189e+00 -4.99494761e-01
2.45822787e-01 3.44654560e-01 -1.68496892e-01 1.20799214e-01
-2.62393594e-01 1.56666815e-01 -1.69939369e-01 5.34375966e-01
1.18039584e+00 -3.03294748e-01 1.10781014e-01 -1.69892609e-01
1.62679136e-01 1.04487367e-01 -1.19789529e+00 1.55245316e+00
-3.49533200e-01 1.01764727e+00 -9.65028107e-02 -7.86000371e-01
8.92105758e-01 1.28839657e-01 -6.34014532e-02 -5.96217155e-01
-1.39721036e-01 -1.18970796e-01 -5.65846443e-01 -4.25962210e-01
5.76626122e-01 1.79504961e-01 1.86361700e-01 8.27029407e-01
9.60478038e-02 5.95619045e-02 -3.86979431e-01 4.93978351e-01
1.08048904e+00 -2.01284796e-01 5.75587869e-01 -1.99937105e-01
1.14655390e-01 -3.18182856e-01 2.53195703e-01 1.46238470e+00
-8.75056908e-02 7.01672375e-01 2.13562369e-01 -1.78795218e-01
-7.36538351e-01 -1.46990919e+00 -1.94957390e-01 7.58180439e-01
3.54479373e-01 -6.10710740e-01 -6.71098650e-01 -7.49275744e-01
1.71484828e-01 1.08687949e+00 -6.54659927e-01 -5.10630012e-01
9.24164876e-02 -4.34293091e-01 8.03631961e-01 2.70521432e-01
6.75615311e-01 -8.73388290e-01 -7.41615891e-01 -2.29384094e-01
-6.75603151e-02 -9.19367194e-01 -3.49470109e-01 2.42327228e-01
-6.25907838e-01 -1.07234597e+00 -3.71760219e-01 -4.51139778e-01
1.05310750e+00 6.46675289e-01 1.22952437e+00 4.09192562e-01
-6.64049804e-01 7.38326609e-01 -3.51976067e-01 -6.27690554e-01
-5.32290041e-01 -4.64822680e-01 -1.22737534e-01 1.92947954e-01
3.33037049e-01 -3.73154879e-01 -5.75148046e-01 2.83522129e-01
-1.10674512e+00 -3.28802951e-02 6.54061019e-01 5.91012061e-01
4.42909241e-01 2.97223091e-01 2.14460149e-01 -1.34829593e+00
4.93130058e-01 -4.99858886e-01 -2.78022438e-01 4.87315118e-01
-7.77926326e-01 2.11054653e-01 2.36042321e-01 -4.52782601e-01
-1.00466752e+00 2.26404905e-01 2.34853745e-01 -3.71906251e-01
-5.51354349e-01 2.84672022e-01 -2.57773489e-01 -1.59574926e-01
1.04652464e+00 5.00850677e-01 1.04656130e-01 -3.40137362e-01
5.16533434e-01 5.21553934e-01 5.00254512e-01 -4.82339293e-01
8.88524532e-01 5.71833372e-01 -1.53619960e-01 -1.15188658e+00
-8.27820897e-01 -1.58710137e-01 -4.38514769e-01 -2.02936545e-01
4.21367586e-01 -8.50609004e-01 -3.92811149e-01 4.25560832e-01
-1.03386188e+00 -2.19186857e-01 -6.32075429e-01 1.52084410e-01
-4.10621643e-01 5.19285262e-01 -2.19711721e-01 -1.00574410e+00
-1.58354014e-01 -7.48143673e-01 7.70857573e-01 1.96266875e-01
-6.07646108e-01 -8.19420815e-01 -3.45040113e-01 2.47338921e-01
3.03482354e-01 1.87744811e-01 8.72013211e-01 -5.56908071e-01
-7.70704091e-01 -2.29546741e-01 -4.92949039e-01 6.09882951e-01
2.77571589e-01 -1.28626645e-01 -1.43501592e+00 -4.58491027e-01
-4.22826558e-02 -3.81599605e-01 1.09758425e+00 4.16903853e-01
1.35872018e+00 -7.59452164e-01 -4.11226988e-01 4.98810053e-01
1.31843114e+00 -1.10697672e-02 8.86199951e-01 2.49044783e-02
3.12545717e-01 5.54596066e-01 5.47827125e-01 5.77749193e-01
1.01657376e-01 1.60910651e-01 4.30889100e-01 -1.06462203e-01
-3.23968798e-01 -5.23812830e-01 2.95531750e-01 1.10852048e-01
2.60781556e-01 -4.46220994e-01 -5.05822957e-01 4.04230475e-01
-1.60040319e+00 -9.04954195e-01 -2.48329360e-02 2.48091555e+00
7.50706732e-01 1.97158068e-01 -3.60161066e-01 1.60987943e-01
7.01621771e-01 3.78736883e-01 -6.55529082e-01 -3.53325248e-01
-3.54794860e-01 4.58371758e-01 9.60873485e-01 3.89908284e-01
-1.12130558e+00 7.86456466e-01 6.36055183e+00 9.30676043e-01
-9.03283060e-01 -6.43887445e-02 7.39054799e-01 4.76262793e-02
-5.48747301e-01 1.22796878e-01 -6.55088961e-01 3.34981143e-01
5.66838861e-01 -2.46566743e-01 3.29562426e-01 6.79381192e-01
-1.01030841e-01 -5.33127904e-01 -1.21255851e+00 8.95927966e-01
4.02537733e-01 -1.42008054e+00 2.45833263e-01 -8.70277584e-02
5.75232089e-01 -5.54259181e-01 4.18776870e-01 -1.94757745e-01
4.98507619e-01 -1.21815312e+00 6.92532778e-01 5.02968252e-01
1.03133023e+00 -4.85444427e-01 3.35803449e-01 4.07899469e-01
-8.69696140e-01 8.38402435e-02 -4.18085307e-01 -1.45127833e-01
-3.27355266e-01 6.44164264e-01 -8.25026572e-01 5.36527574e-01
7.31764853e-01 6.80179536e-01 -9.54789937e-01 1.09121859e+00
-3.53330195e-01 9.25694942e-01 -2.29120731e-01 1.27418548e-01
-2.47153774e-01 1.88148409e-01 4.83208030e-01 1.22849369e+00
4.42070752e-01 6.57055303e-02 -1.71097964e-01 9.46578741e-01
-1.40674382e-01 -9.59390178e-02 -1.04678810e+00 -4.97952625e-02
6.59436524e-01 6.34142399e-01 -5.37948251e-01 -1.95582911e-01
-1.78327486e-01 1.16586173e+00 2.25363851e-01 4.70442265e-01
-4.30098027e-01 -1.11999914e-01 7.48199999e-01 2.43942887e-01
1.24214515e-01 -9.97093171e-02 -5.48446059e-01 -9.56453145e-01
-5.06993905e-02 -7.63230085e-01 4.40525115e-01 -1.09999156e+00
-1.24404430e+00 3.33275765e-01 6.78129122e-03 -1.24182177e+00
1.51974574e-01 -3.06628466e-01 -5.78561306e-01 7.11960852e-01
-1.22188950e+00 -1.10709739e+00 -3.14598948e-01 5.45939565e-01
2.64479578e-01 -1.82739809e-01 1.25134671e+00 -1.70045778e-01
-2.90114284e-01 9.27998126e-01 1.86775282e-01 -8.17612633e-02
8.17702413e-01 -9.70834613e-01 2.72236437e-01 9.73478854e-01
4.78805959e-01 6.98646605e-01 9.49831545e-01 -6.27329767e-01
-1.18167353e+00 -1.30165231e+00 9.66519475e-01 -3.37579399e-01
3.26638818e-01 -3.48443747e-01 -1.10054517e+00 9.84699309e-01
2.96916187e-01 -3.29308808e-02 9.36563849e-01 -1.43253431e-01
-7.25994647e-01 2.36642715e-02 -1.50971675e+00 4.95150119e-01
1.04740107e+00 -7.99890518e-01 -3.90276223e-01 3.36596012e-01
6.25305414e-01 -1.02997452e-01 -3.72485340e-01 6.71272427e-02
4.73427415e-01 -1.09106827e+00 1.23804188e+00 -4.86409187e-01
4.72601712e-01 -1.16697326e-01 -1.65627062e-01 -1.15932333e+00
-1.66704282e-01 -6.24189258e-01 -1.16780601e-01 1.38046539e+00
2.88895398e-01 -6.63338721e-01 8.65760028e-01 8.01831603e-01
3.68919551e-01 -1.45798489e-01 -7.48949230e-01 -9.61974740e-01
-2.72031367e-01 -6.77607238e-01 4.71149176e-01 1.17383671e+00
-2.14897722e-01 -1.11872427e-01 -7.37546384e-01 6.16520882e-01
1.14525652e+00 1.20253257e-01 8.35540414e-01 -9.59466159e-01
-1.53796405e-01 -1.03689775e-01 -4.36094016e-01 -1.02168536e+00
2.31988624e-01 -8.33362937e-01 -6.99593276e-02 -1.30419016e+00
1.94664940e-01 -5.36013663e-01 -3.63195240e-01 5.43255389e-01
-8.93364847e-02 3.01530212e-01 1.83314949e-01 5.69992289e-02
-4.86408710e-01 1.24333739e-01 9.30633664e-01 -3.05108279e-01
1.35653675e-01 2.45405182e-01 -1.01887071e+00 6.43235624e-01
9.73574400e-01 -5.28873563e-01 -6.40344799e-01 -3.69343847e-01
1.38223097e-01 -1.54239073e-01 8.21255386e-01 -8.71562362e-01
2.78622836e-01 -6.42891526e-02 5.99085510e-01 -3.82273674e-01
3.37651610e-01 -9.22963381e-01 4.01884884e-01 4.76859242e-01
-7.50561118e-01 -6.01727426e-01 3.81651670e-01 7.57920802e-01
1.74555570e-01 -5.11511326e-01 6.37077153e-01 -3.43639374e-01
-8.09536755e-01 2.29841340e-02 -5.13733268e-01 -8.21090639e-02
1.01960862e+00 -3.66939813e-01 -4.68942106e-01 -7.21783459e-01
-6.21250629e-01 1.11511961e-01 6.32528424e-01 1.20537408e-01
8.99212182e-01 -9.67384279e-01 -5.18247902e-01 5.37971854e-01
3.92593026e-01 -9.16433781e-02 3.24193120e-01 1.28802001e-01
-1.76058263e-01 -7.03363791e-02 -1.67001665e-01 -3.96355152e-01
-1.57720804e+00 6.44057930e-01 -9.99451280e-02 1.34443715e-01
-5.06239414e-01 1.23594439e+00 3.29139024e-01 -2.68549025e-02
4.43184167e-01 -1.24875009e-01 2.51752764e-01 6.75012693e-02
5.91818631e-01 1.57474771e-01 -2.94483274e-01 -5.20925999e-01
-2.01419920e-01 2.81937629e-01 -1.79561198e-01 1.99545085e-01
9.95778501e-01 -3.08209121e-01 8.38842243e-02 3.65476936e-01
1.23491347e+00 9.95345861e-02 -1.23164964e+00 -6.93451524e-01
-1.43111914e-01 -1.07820308e+00 -1.93461120e-01 -7.55616546e-01
-9.21293378e-01 7.98353910e-01 7.09846795e-01 2.73867011e-01
1.10469925e+00 1.35448754e-01 5.49526513e-01 4.33085829e-01
5.15723050e-01 -7.97873020e-01 2.99378708e-02 -8.40931293e-03
8.71778309e-01 -1.29432583e+00 3.59618753e-01 -6.62640929e-01
-6.13228619e-01 7.81524897e-01 2.22880647e-01 2.71193702e-02
8.26299906e-01 2.35124916e-01 2.73441207e-02 -4.83842008e-02
-5.23257077e-01 -4.37097587e-02 1.25690669e-01 9.37870979e-01
1.82191692e-02 1.40099958e-01 2.02865139e-01 3.44869554e-01
-1.99001238e-01 1.60160989e-01 6.24990225e-01 1.22043717e+00
-3.38521719e-01 -9.47684824e-01 -5.47265112e-01 7.29857683e-01
-1.76971436e-01 -1.84140041e-01 -5.69501460e-01 5.03489017e-01
2.09466770e-01 9.87334907e-01 8.06908384e-02 -5.78366756e-01
9.82971713e-02 1.34095252e-01 4.69145000e-01 -5.25770128e-01
-3.79143625e-01 -3.15776438e-01 1.17261246e-01 -5.41178524e-01
-4.34953600e-01 -7.86314547e-01 -8.42699111e-01 -4.88357186e-01
-2.78687567e-01 -1.95814386e-01 3.66263330e-01 8.29061031e-01
4.34449881e-01 1.62794799e-01 4.88916725e-01 -6.53809428e-01
-3.27076435e-01 -4.30147409e-01 -7.25856423e-01 5.15407205e-01
6.61956131e-01 -4.39999342e-01 -5.35334826e-01 3.39564621e-01] | [10.035548210144043, 2.2451930046081543] |
ab5be222-ae90-4bb2-9f22-b05884b0e301 | architecture-and-evolution-of-semantic | 1908.04911 | null | https://arxiv.org/abs/1908.04911v2 | https://arxiv.org/pdf/1908.04911v2.pdf | Architecture and evolution of semantic networks in mathematics texts | Knowledge is a network of interconnected concepts. Yet, precisely how the topological structure of knowledge constrains its acquisition remains unknown, hampering the development of learning enhancement strategies. Here we study the topological structure of semantic networks reflecting mathematical concepts and their relations in college-level linear algebra texts. We hypothesize that these networks will exhibit structural order, reflecting the logical sequence of topics that ensures accessibility. We find that the networks exhibit strong core-periphery architecture, where a dense core of concepts presented early is complemented with a sparse periphery presented evenly throughout the exposition; the latter is composed of many small modules each reflecting more narrow domains. Using tools from applied topology, we find that the expositional evolution of the semantic networks produces and subsequently fills knowledge gaps, and that the density of these gaps tracks negatively with community ratings of each textbook. Broadly, our study lays the groundwork for future efforts developing optimal design principles for textbook exposition and teaching in a classroom setting. | ['Nicolas H. Christianson', 'Ann Sizemore Blevins', 'Danielle S. Bassett'] | 2019-08-14 | null | null | null | null | ['logical-sequence'] | ['reasoning'] | [-2.24504158e-01 3.15071404e-01 -1.60268500e-01 -2.67051551e-02
5.06013691e-01 -1.17754447e+00 5.69306970e-01 6.21649504e-01
1.43485382e-01 2.75820583e-01 6.47363245e-01 -8.40169430e-01
-1.15948665e+00 -1.22930098e+00 -8.03446949e-01 -1.21245362e-01
-3.21072549e-01 2.06188902e-01 5.27668536e-01 -6.08040750e-01
4.66414541e-01 4.40904617e-01 -1.67254603e+00 1.18252793e-02
1.27202070e+00 2.98926622e-01 2.66992718e-01 3.80416244e-01
-6.39466584e-01 9.46686089e-01 -5.30165970e-01 -4.54717875e-01
-6.97921291e-02 -5.11038065e-01 -9.76861477e-01 -1.29873514e-01
9.42567050e-01 1.24746650e-01 -6.01248741e-01 9.83846486e-01
2.31252853e-02 5.09631693e-01 6.32116377e-01 -9.64778721e-01
-1.01883793e+00 1.02348506e+00 -2.81242102e-01 5.73664308e-01
2.58688062e-01 -1.43514127e-01 1.12313092e+00 -4.39154059e-01
8.81603718e-01 1.08210194e+00 6.68886483e-01 8.92229304e-02
-1.34139645e+00 -3.80827188e-01 4.28458303e-01 6.86072186e-02
-1.35358584e+00 2.23223609e-03 7.29856014e-01 -8.86773825e-01
4.02196378e-01 -4.53113280e-02 1.35308516e+00 5.95243394e-01
8.95537511e-02 2.26894677e-01 1.20939875e+00 -4.72710133e-01
7.42779374e-02 4.20052737e-01 4.35698271e-01 8.91921520e-01
6.99246407e-01 -2.11304411e-01 -9.53483760e-01 2.53783137e-01
1.07173336e+00 -2.27583870e-01 -2.86524892e-01 -8.76576245e-01
-8.77970755e-01 4.78188366e-01 5.94374239e-01 7.32336700e-01
8.82547628e-03 -1.78790525e-01 -4.33262587e-02 7.89900780e-01
-1.37965515e-01 9.26624477e-01 -2.48053789e-01 -1.76714465e-01
-1.04615819e+00 5.72708212e-02 9.56626236e-01 1.14323580e+00
5.49366415e-01 -3.76052231e-01 5.10071576e-01 7.06811607e-01
2.51204878e-01 4.50084209e-02 6.65566698e-02 -1.10067582e+00
2.50897706e-01 9.31964219e-01 -5.46303868e-01 -1.31587076e+00
-1.72984958e-01 -9.13236320e-01 -9.23018381e-02 1.70311686e-02
7.93317497e-01 -1.12141920e-02 -2.51940072e-01 2.01064014e+00
1.90067768e-01 -1.38664013e-02 -4.00789648e-01 5.17044246e-01
5.13184309e-01 3.06543738e-01 3.24391603e-01 2.45262697e-01
1.31587410e+00 -5.08862734e-01 -6.74950123e-01 6.15422167e-02
5.55892408e-01 -8.68773639e-01 9.98780489e-01 2.32324168e-01
-1.51019037e+00 -5.27840078e-01 -1.20268977e+00 -9.58956704e-02
-7.11322546e-01 -6.35528386e-01 5.81690729e-01 8.74484360e-01
-1.49551046e+00 7.78115332e-01 -3.33544105e-01 -6.40589774e-01
6.28464758e-01 3.71723883e-02 -2.61479408e-01 -1.26236573e-01
-1.04137826e+00 1.08928192e+00 1.27451599e-01 -3.66420299e-01
-5.90078115e-01 -1.18915796e+00 -4.63804632e-01 4.14951593e-01
2.10197017e-01 -8.59784603e-01 7.92381942e-01 -5.90050697e-01
-1.23449755e+00 6.99900150e-01 5.52184999e-01 2.14635491e-01
2.19464853e-01 9.53014866e-02 -1.65968239e-01 2.50937819e-01
-1.85589287e-02 5.46964109e-01 3.39480877e-01 -1.25640452e+00
-4.27975625e-01 -3.09071213e-01 5.31776428e-01 4.75116879e-01
-9.09470797e-01 -4.74997193e-01 -6.41457736e-02 -5.24409950e-01
5.52401960e-01 -5.35664260e-01 1.55435041e-01 5.26525937e-02
1.85591038e-02 -4.13840860e-01 5.43629169e-01 -2.46948317e-01
1.51521599e+00 -2.08746338e+00 1.52316436e-01 9.59004104e-01
6.84233189e-01 -3.52212101e-01 -2.33306959e-01 8.28087270e-01
-1.82262182e-01 3.55292559e-01 5.76104939e-01 5.99056244e-01
2.37793043e-01 -1.50740191e-01 -2.62883723e-01 3.62510711e-01
-4.06626314e-01 7.63958871e-01 -1.30242908e+00 -6.02609217e-01
4.43101339e-02 2.78536469e-01 -6.90517843e-01 -1.45683572e-01
-1.61452755e-01 1.71047762e-01 -3.82832199e-01 4.34129387e-01
3.53630096e-01 -5.60915470e-01 5.85854709e-01 3.88979763e-01
-4.24721867e-01 6.64572537e-01 -1.06425512e+00 1.64574242e+00
-3.51197451e-01 1.12472236e+00 3.09071392e-01 -9.61395562e-01
7.74695218e-01 -1.19151652e-01 3.10195237e-01 -5.37475348e-01
-5.42035401e-02 -2.06236795e-01 5.32970309e-01 -3.84536415e-01
4.80864048e-01 8.31009746e-02 2.52909929e-01 7.57273853e-01
3.33072305e-01 -5.13102114e-01 3.77913088e-01 9.34410334e-01
1.14446390e+00 1.48648641e-03 -4.09449100e-01 -1.05582035e+00
8.23692381e-02 1.45806044e-01 5.91068389e-03 7.94301987e-01
-2.67676115e-01 -2.64496773e-01 7.82480776e-01 -9.07763764e-02
-1.17655599e+00 -1.61468971e+00 -3.59971166e-01 1.41910470e+00
3.42209846e-01 -6.07138276e-01 -8.57806623e-01 -1.83694199e-01
1.57824054e-01 6.16427243e-01 -3.44314098e-01 -1.36100650e-01
-3.90595645e-01 1.44906774e-01 1.38631240e-01 2.98450440e-01
2.42091238e-01 -8.67413521e-01 -4.37686563e-01 -1.03818569e-02
-1.54522285e-01 -7.13420212e-01 -4.83674616e-01 4.75962274e-02
-1.20832205e+00 -1.18991673e+00 -2.69218236e-01 -1.08036053e+00
9.51824665e-01 4.70960706e-01 1.46888006e+00 8.18401754e-01
-2.15902209e-01 9.62020934e-01 -8.43235254e-02 -2.94361144e-01
-4.20561666e-03 1.38728961e-01 2.31662542e-02 -8.51439059e-01
3.14505011e-01 -1.07446492e+00 -7.29046345e-01 2.27839753e-01
-8.06650341e-01 -5.84362932e-02 4.53110814e-01 3.87232810e-01
-6.08920343e-02 7.07367420e-01 4.48857903e-01 -5.83393872e-01
8.88387680e-01 -6.41933441e-01 -6.12981319e-01 3.51995975e-01
-7.08609402e-01 -3.39552797e-02 3.16565365e-01 -2.55702108e-01
-1.05831993e+00 -6.69354379e-01 5.76119423e-01 1.48917869e-01
5.01153506e-02 6.48362458e-01 1.01067744e-01 -4.04421180e-01
7.77494013e-01 -4.80565093e-02 3.43934484e-02 -2.42688835e-01
6.12654805e-01 6.09161854e-02 2.13863373e-01 -1.27174807e+00
1.09121275e+00 1.53332829e-01 2.63040420e-02 -1.25976717e+00
-7.80027092e-01 -1.61873147e-01 -8.45687687e-01 -5.43929160e-01
5.20428836e-01 -9.63227093e-01 -1.00986278e+00 -2.13325068e-01
-6.42034233e-01 -6.04244769e-01 -4.52470869e-01 5.09516299e-01
-2.06994995e-01 1.69879928e-01 -7.52085388e-01 -2.70107985e-01
5.91628730e-01 -8.48650753e-01 4.50052023e-02 5.17937124e-01
-7.01882541e-01 -1.48988068e+00 1.84825480e-01 7.49849617e-01
5.63339949e-01 -2.62593538e-01 1.55820882e+00 -3.69623035e-01
-1.20211446e+00 4.90430117e-01 1.94123406e-02 -2.58987427e-01
-1.96644977e-01 1.18885063e-01 -4.24151987e-01 6.26728125e-03
-3.37642282e-01 -2.38428608e-01 5.72510183e-01 4.23155665e-01
9.38753188e-01 -1.71412632e-01 -2.56717801e-01 2.70807624e-01
1.52625465e+00 -1.08493492e-03 3.52790833e-01 5.21005809e-01
3.76055151e-01 1.06859457e+00 -1.36688426e-01 -1.86363474e-01
4.61331129e-01 -3.07668280e-03 -3.06276888e-01 3.02515209e-01
-2.45567724e-01 -6.38324857e-01 7.66624436e-02 1.39876807e+00
9.23634470e-02 2.06862167e-01 -1.25409913e+00 1.04583681e+00
-1.32262957e+00 -1.06443131e+00 -1.05656318e-01 1.92162132e+00
1.25209284e+00 2.69900113e-01 -1.78025272e-02 -2.33388782e-01
4.94272143e-01 -8.27046111e-02 5.11973118e-03 -4.35155004e-01
5.76472543e-02 5.91016412e-01 1.85712293e-01 6.73162699e-01
-3.26875329e-01 9.16480362e-01 7.45472956e+00 7.12262154e-01
-4.55596775e-01 -1.76165536e-01 2.45941088e-01 -7.30551928e-02
-8.33623409e-01 6.83761910e-02 -5.31283617e-01 3.35983366e-01
8.97406578e-01 -3.66193652e-01 6.05701745e-01 5.01023054e-01
3.59631591e-02 -2.32313380e-01 -9.27244186e-01 1.65405586e-01
-3.42773825e-01 -1.45261824e+00 -6.73323870e-02 8.04426968e-02
1.32810485e+00 -4.49279517e-01 3.02607268e-01 6.96421042e-02
9.64596272e-01 -1.04564273e+00 5.98079622e-01 4.95056242e-01
4.57933813e-01 -6.56890213e-01 5.61136752e-03 7.39039555e-02
-1.12368286e+00 -1.25076041e-01 -2.88454443e-02 -4.49794501e-01
-2.92728513e-01 4.39175516e-01 -4.40233320e-01 2.86959801e-02
6.82003915e-01 3.06647450e-01 -8.02705765e-01 1.00685096e+00
-3.34705114e-01 5.89078128e-01 -8.73072539e-03 -1.16375349e-01
3.51106599e-02 -5.67319632e-01 3.85633826e-01 8.32244933e-01
2.34564289e-01 4.08904791e-01 1.22846581e-01 1.22522557e+00
-1.76582709e-01 1.30882964e-01 -8.89722288e-01 -3.94739628e-01
1.15137577e+00 1.15318000e+00 -1.30041134e+00 -1.92487285e-01
-4.55105931e-01 2.44367406e-01 4.61518019e-01 5.13623893e-01
-2.97014624e-01 -4.54975218e-01 7.75159657e-01 6.36408031e-01
9.14182439e-02 -6.14695311e-01 -9.21369374e-01 -8.08955967e-01
-3.91595691e-01 -6.52302742e-01 8.67682099e-02 -6.59322619e-01
-1.52097166e+00 -2.31804609e-01 -7.67513439e-02 -3.23806763e-01
7.07796872e-01 -4.82376307e-01 -8.50620925e-01 8.17024231e-01
-9.39918697e-01 -5.56992292e-01 -4.78881806e-01 5.84676564e-01
1.26434462e-02 -1.20415859e-01 4.02394444e-01 -3.12919319e-02
-3.46473813e-01 3.92781883e-01 2.11208761e-01 1.09129526e-01
5.30072689e-01 -1.44310951e+00 2.94709448e-02 5.18353403e-01
2.79218316e-01 1.43070066e+00 5.50213754e-01 -8.53508651e-01
-1.34737241e+00 -3.69796664e-01 1.01914704e+00 -8.31514835e-01
1.35964024e+00 -3.17935556e-01 -9.16180789e-01 5.38333416e-01
5.56270182e-01 -7.08731353e-01 9.47377026e-01 7.98828602e-01
-4.73673314e-01 1.40952930e-01 -7.71488667e-01 7.04797268e-01
1.59338796e+00 -7.43570507e-01 -1.02322066e+00 3.25114667e-01
7.30202377e-01 -1.66352734e-01 -1.35692823e+00 -2.06864953e-01
4.63733047e-01 -9.92762864e-01 1.07902992e+00 -3.40112835e-01
8.41178119e-01 -5.17839678e-02 3.24760258e-01 -1.40300429e+00
-6.92344487e-01 -4.54922795e-01 9.63491425e-02 1.04219866e+00
3.49311352e-01 -3.15635949e-01 1.08621383e+00 5.38821340e-01
-2.45975479e-01 -5.18089831e-01 -3.74221474e-01 -6.15730286e-01
6.85687840e-01 1.93851277e-01 4.22055155e-01 1.61339438e+00
7.31716335e-01 4.37099218e-01 8.97531688e-01 -2.29069754e-01
8.62530887e-01 -1.98705234e-02 1.89292386e-01 -1.67611778e+00
1.77745700e-01 -1.09545708e+00 -4.07019466e-01 -9.32338417e-01
8.72942284e-02 -9.49428558e-01 -5.33900023e-01 -1.45652449e+00
1.84750482e-01 -7.59851694e-01 -1.87912688e-01 1.21544339e-01
2.27390394e-01 -3.86114120e-01 6.28887191e-02 -9.47402492e-02
-6.39477015e-01 1.27327606e-01 1.52975452e+00 1.62410989e-01
-3.91491383e-01 -7.36356199e-01 -1.05669534e+00 7.21277416e-01
7.37989247e-01 5.51249087e-02 -1.04014611e+00 -4.96238142e-01
6.46130621e-01 -3.15266699e-01 1.23786829e-01 -1.05211127e+00
7.56226838e-01 -4.75770146e-01 7.68378258e-01 -4.29446846e-01
-2.88057208e-01 -1.04179370e+00 -2.34748181e-02 3.72927278e-01
-5.85326731e-01 1.14114039e-01 2.39885136e-01 4.14039850e-01
1.08365081e-01 -1.90946415e-01 7.08052814e-01 -1.06255956e-01
-3.76920432e-01 -1.54538840e-01 -5.15655458e-01 4.49417233e-01
1.07560647e+00 -3.85904700e-01 -6.10682368e-01 -2.96964645e-01
-7.07939684e-01 3.82104248e-01 6.21131122e-01 4.32458401e-01
4.21256393e-01 -1.26316893e+00 -1.67539880e-01 3.17662992e-02
-2.60157973e-01 -2.59871602e-01 2.58664280e-01 6.00600183e-01
-6.48572147e-01 5.84179759e-01 -5.45552015e-01 -3.61671805e-01
-1.02337682e+00 4.02835757e-01 1.50706574e-01 1.31011561e-01
-5.32208622e-01 1.20782554e+00 2.35503271e-01 -2.83460766e-01
4.71911281e-01 -7.29021132e-02 -4.10517931e-01 3.18209529e-01
2.93040454e-01 7.02612579e-01 -4.61541653e-01 -8.52895901e-02
2.04402462e-01 5.13597727e-01 -5.91443479e-02 -6.95364550e-02
1.30764711e+00 -3.50391060e-01 -4.78864551e-01 3.23626041e-01
6.45328462e-01 2.66388834e-01 -8.72803211e-01 -3.15628201e-01
3.39706749e-01 -3.74463171e-01 -5.35729863e-02 -8.76747668e-01
-7.39722908e-01 6.35336876e-01 4.01705839e-02 4.30737585e-01
7.50970602e-01 1.82776287e-01 1.85384795e-01 2.63636976e-01
1.97252676e-01 -1.37741244e+00 6.42052233e-01 6.83778286e-01
4.10848469e-01 -4.64949608e-01 2.44697139e-01 -6.11978948e-01
7.78391361e-02 1.16406715e+00 9.74771976e-01 9.49711800e-02
1.00563419e+00 6.61549047e-02 -5.40439725e-01 -8.92123401e-01
-7.41820276e-01 -1.28453784e-02 2.93863207e-01 4.34005737e-01
7.69493282e-01 1.20104395e-01 -2.53727168e-01 3.23609680e-01
-7.95553029e-01 -3.19911391e-01 6.94022655e-01 8.73294294e-01
-1.02757883e+00 -9.74451721e-01 -4.85182285e-01 3.76564831e-01
-1.86320141e-01 -3.94834667e-01 -6.78414643e-01 8.89504135e-01
3.83737415e-01 7.98613250e-01 4.04020399e-01 -1.39861748e-01
2.60625005e-01 3.37523013e-01 9.71747875e-01 -6.13194168e-01
-6.55193806e-01 -3.20537925e-01 -1.35552898e-01 -1.75315574e-01
7.84608908e-03 -5.99847794e-01 -1.11101353e+00 -1.02824903e+00
-3.47758740e-01 4.78839636e-01 4.99796242e-01 6.13137424e-01
1.94601730e-01 8.88172686e-01 1.09189367e-02 1.32944241e-01
-3.89823437e-01 -6.85693026e-01 -7.45572567e-01 3.58703166e-01
8.07211772e-02 -7.58930385e-01 -2.59420365e-01 1.34704947e-01] | [9.543160438537598, 7.167332172393799] |
6c092d45-b1db-4153-ae62-1ca2e09b4acd | backtracking-regression-forests-for-accurate | 1710.07965 | null | http://arxiv.org/abs/1710.07965v1 | http://arxiv.org/pdf/1710.07965v1.pdf | Backtracking Regression Forests for Accurate Camera Relocalization | Camera relocalization plays a vital role in many robotics and computer vision
tasks, such as global localization, recovery from tracking failure, and loop
closure detection. Recent random forests based methods directly predict 3D
world locations for 2D image locations to guide the camera pose optimization.
During training, each tree greedily splits the samples to minimize the spatial
variance. However, these greedy splits often produce uneven sub-trees in
training or incorrect 2D-3D correspondences in testing. To address these
problems, we propose a sample-balanced objective to encourage equal numbers of
samples in the left and right sub-trees, and a novel backtracking scheme to
remedy the incorrect 2D-3D correspondence predictions. Furthermore, we extend
the regression forests based methods to use local features in both training and
testing stages for outdoor RGB-only applications. Experimental results on
publicly available indoor and outdoor datasets demonstrate the efficacy of our
approach, which shows superior or on-par accuracy with several state-of-the-art
methods. | ['Clarence W. de Silva', 'Julien Valentin', 'James J. Little', 'Lili Meng', 'Frederick Tung', 'Jianhui Chen'] | 2017-10-22 | null | null | null | null | ['camera-relocalization', 'loop-closure-detection'] | ['computer-vision', 'computer-vision'] | [ 5.80801927e-02 -3.52911443e-01 -3.63456398e-01 -4.06206548e-01
-5.89635849e-01 -5.51743746e-01 3.67815256e-01 -1.24436013e-01
-4.39068317e-01 8.95947039e-01 -1.50101900e-01 -4.20440286e-01
2.93728244e-02 -6.63241684e-01 -8.66625845e-01 -7.07834959e-01
3.58025879e-02 4.16417778e-01 6.50740385e-01 1.51460826e-01
4.10697788e-01 6.75446451e-01 -1.50637865e+00 -4.21315223e-01
8.53645384e-01 1.17560279e+00 3.20659012e-01 5.24528563e-01
6.46669194e-02 5.74731112e-01 -1.47585928e-01 5.49986809e-02
4.50451344e-01 -8.00826326e-02 -3.15922469e-01 3.03691953e-01
5.83427966e-01 -2.84003764e-01 -1.08109780e-01 1.06779635e+00
2.85365075e-01 4.16493155e-02 3.76552492e-01 -1.28588510e+00
-1.60374455e-02 -1.76318642e-03 -1.04604709e+00 -1.20116048e-01
5.99586785e-01 1.90506801e-01 5.14515817e-01 -1.01817453e+00
4.79182661e-01 9.74263608e-01 1.18064082e+00 3.17310005e-01
-1.26088262e+00 -9.19742584e-01 4.81113642e-01 8.38219747e-03
-1.44977748e+00 -3.27201366e-01 6.68137431e-01 -4.12466109e-01
5.80505013e-01 1.56778574e-01 8.66482913e-01 7.75568008e-01
2.05947652e-01 5.76734066e-01 1.32479477e+00 -3.43763232e-01
1.61922097e-01 -5.68080246e-02 -1.42038479e-01 9.99946356e-01
3.21012199e-01 1.91692591e-01 -7.32921124e-01 -9.98402312e-02
1.02939570e+00 2.44941667e-01 -3.84931356e-01 -1.00047219e+00
-1.53054523e+00 6.11364722e-01 8.12750757e-01 -2.81666338e-01
-2.92604655e-01 1.89391136e-01 -1.79383755e-01 -1.22044399e-01
2.85256207e-01 2.61097431e-01 -4.83145237e-01 9.23005491e-02
-8.74810874e-01 2.43893519e-01 4.15585965e-01 1.30648291e+00
1.12573588e+00 -4.26978886e-01 1.65669903e-01 6.39551044e-01
4.97243732e-01 6.84033692e-01 3.37067127e-01 -9.15008366e-01
6.41894877e-01 6.53540254e-01 4.12781209e-01 -1.01702833e+00
-3.16903025e-01 -2.58030504e-01 -8.20812047e-01 3.27132493e-01
4.05067742e-01 2.19765939e-02 -1.22082281e+00 1.42107809e+00
8.07937920e-01 2.31840834e-01 -2.29448929e-01 1.08498275e+00
3.87892038e-01 4.45198029e-01 -2.72592336e-01 -1.14642382e-01
8.09764683e-01 -1.09977710e+00 -2.49881387e-01 -7.44546056e-01
4.03859377e-01 -7.84856617e-01 8.43189120e-01 1.46023974e-01
-5.04187465e-01 -4.16621476e-01 -9.95857716e-01 1.49732634e-01
-1.08539678e-01 2.64266133e-01 5.85320890e-01 4.42868412e-01
-8.56923342e-01 4.27695692e-01 -9.41245139e-01 -5.83121359e-01
3.08976799e-01 5.48179388e-01 -5.48725843e-01 -3.82502466e-01
-5.47164142e-01 7.48318970e-01 2.00121731e-01 3.35662931e-01
-6.24366641e-01 -3.23430330e-01 -8.25050890e-01 -5.15248120e-01
5.11090934e-01 -6.19406939e-01 1.04526985e+00 -5.00415683e-01
-1.50579476e+00 8.62835109e-01 -4.30084139e-01 -4.09590691e-01
7.05113590e-01 -3.84278059e-01 1.87839374e-01 -3.18211675e-01
4.66626614e-01 9.07834291e-01 7.49273241e-01 -1.30987072e+00
-1.02045429e+00 -6.60980284e-01 -1.97564691e-01 4.82640535e-01
1.62402019e-01 -5.27260721e-01 -8.23624730e-01 -4.90114272e-01
1.02267170e+00 -1.30567002e+00 -7.20727205e-01 3.89204115e-01
-6.42676771e-01 1.30243495e-01 7.85728216e-01 -3.00810575e-01
7.79642045e-01 -1.88189793e+00 1.51893860e-02 2.39380255e-01
1.12650193e-01 -2.99746722e-01 1.13153107e-01 -1.05254240e-02
3.19908768e-01 -1.69418484e-01 -2.57952455e-02 -4.99303848e-01
-4.35314447e-01 2.51762658e-01 -3.09916079e-01 8.51047754e-01
8.48030373e-02 3.64346474e-01 -9.60121095e-01 -5.00716209e-01
5.34263611e-01 1.83629036e-01 -4.12694991e-01 1.51215449e-01
-6.81582689e-02 7.98892617e-01 -7.20614254e-01 8.85279715e-01
9.58899438e-01 -1.05290942e-01 -1.28673660e-02 -5.02846576e-02
-3.81669343e-01 3.75416875e-01 -1.21512294e+00 1.77835846e+00
-3.17274243e-01 5.44240654e-01 -4.90103811e-02 -5.75849175e-01
9.97304082e-01 -3.26644808e-01 3.97392005e-01 -5.14477849e-01
-8.37087408e-02 3.21588606e-01 -5.70206404e-01 4.76137139e-02
6.71225309e-01 1.58917189e-01 -6.25886843e-02 -5.69540635e-02
-3.94712776e-01 -3.14149231e-01 -2.27006018e-01 -2.80108571e-01
8.78324449e-01 6.89230978e-01 4.36156601e-01 -3.32360566e-02
3.53033870e-01 4.51038837e-01 8.39302540e-01 8.09597552e-01
-2.97535330e-01 7.99236000e-01 1.29522949e-01 -4.73852217e-01
-8.71332467e-01 -7.90929437e-01 1.12871818e-01 6.43801391e-01
9.31898355e-01 -1.89156666e-01 -4.03009266e-01 -8.02395046e-01
1.77389890e-01 3.68680716e-01 -4.16511625e-01 1.57751277e-01
-5.79393744e-01 -5.48488438e-01 1.53925925e-01 5.64766824e-01
7.41837800e-01 -5.36862493e-01 -1.01135552e+00 1.05198614e-01
-2.21254185e-01 -1.16982532e+00 -3.54963392e-01 6.51337326e-01
-1.08372903e+00 -1.26475477e+00 -6.01790130e-01 -8.72170985e-01
9.88856435e-01 8.27556193e-01 7.28008986e-01 -9.45670232e-02
-4.14781757e-02 -1.18493237e-01 -2.38648057e-01 -1.01326570e-01
8.84014219e-02 1.47499681e-01 2.75392950e-01 -3.44727457e-01
1.70387104e-01 -3.98956209e-01 -5.79540431e-01 9.02125418e-01
-7.11138994e-02 3.04749489e-01 4.88143474e-01 9.47369874e-01
1.08755875e+00 -1.28285453e-01 -2.54094571e-01 -5.52182794e-01
-2.34512202e-02 -2.32194349e-01 -1.17349660e+00 7.56725520e-02
-6.92925453e-01 7.36311674e-02 4.25761104e-01 -5.62239528e-01
-5.30069530e-01 8.11834991e-01 3.23858708e-01 -8.36167336e-01
-1.04549162e-01 2.31481791e-01 4.55769561e-02 -4.56159234e-01
4.40664858e-01 2.71452904e-01 -1.05617158e-01 -3.19671005e-01
1.54326692e-01 4.89097118e-01 5.39228141e-01 -3.18635076e-01
9.92532969e-01 4.57628161e-01 1.34573683e-01 -5.33193588e-01
-7.24277198e-01 -6.62477553e-01 -7.02959359e-01 -3.23923260e-01
6.18026257e-01 -1.15151119e+00 -5.10827124e-01 5.60379326e-01
-9.32298660e-01 -2.87443429e-01 -6.17047101e-02 5.07164121e-01
-5.25690019e-01 8.01161379e-02 7.30747133e-02 -8.97678673e-01
3.09628453e-02 -1.39204252e+00 1.34586024e+00 4.99976635e-01
1.19813763e-01 -5.05351484e-01 7.86106214e-02 9.77940708e-02
8.96271970e-03 3.10348928e-01 3.60369265e-01 -8.11850950e-02
-1.02562022e+00 -4.32929128e-01 -1.27624437e-01 -2.53977209e-01
6.30878434e-02 -8.94780830e-02 -7.92268217e-01 -3.66375506e-01
-3.81664097e-01 -4.01286334e-01 8.55094492e-01 4.01032448e-01
9.02157187e-01 -4.94100936e-02 -8.76342118e-01 7.98172235e-01
1.39787769e+00 -2.91429888e-02 3.45663726e-01 6.73462808e-01
6.57058716e-01 3.97117019e-01 1.21611607e+00 3.41326416e-01
5.50725162e-01 8.38382423e-01 7.50352085e-01 2.98698917e-02
1.22661330e-02 -7.98427105e-01 2.32838973e-01 1.95258647e-01
1.16119102e-01 1.80199812e-03 -9.32996273e-01 5.06832838e-01
-2.07153010e+00 -5.21430314e-01 -2.32842788e-01 2.71441579e+00
5.54028153e-01 2.85603613e-01 3.34944725e-02 -6.45452067e-02
9.07010019e-01 1.41545907e-01 -6.79278195e-01 4.26985383e-01
5.18644080e-02 -3.14100474e-01 1.31352603e+00 5.47588229e-01
-1.36162066e+00 1.16294789e+00 5.95179987e+00 5.44988036e-01
-1.30020726e+00 -1.93427533e-01 4.70297605e-01 1.06413923e-01
5.79108000e-02 4.42116559e-01 -1.23995590e+00 3.10820192e-01
1.66817352e-01 4.30973232e-01 4.00001824e-01 1.04895914e+00
1.16296045e-01 -7.14249909e-01 -8.91482055e-01 1.09547842e+00
-2.40204141e-01 -1.24369407e+00 -4.35175329e-01 1.59410670e-01
7.78660357e-01 3.61541063e-01 -1.44638360e-01 -3.83452699e-02
5.73536932e-01 -6.34459376e-01 1.16352344e+00 2.73036331e-01
9.38639879e-01 -5.50973594e-01 4.94720608e-01 7.47342408e-01
-1.37297499e+00 -9.10463408e-02 -3.70121896e-01 -4.41619195e-02
-4.79286686e-02 5.79015374e-01 -1.13934910e+00 3.90568912e-01
1.03095400e+00 9.40861881e-01 -6.31063700e-01 1.35945082e+00
-4.34456706e-01 1.79057494e-01 -8.50115716e-01 -3.35345566e-01
1.88254252e-01 -1.86013952e-01 6.62617326e-01 6.67352438e-01
3.96160483e-01 -2.57344395e-01 5.36084116e-01 5.89326739e-01
1.87750190e-01 -2.48773336e-01 -8.16215575e-01 4.79620606e-01
8.89866412e-01 9.37894344e-01 -9.16499078e-01 1.29242718e-01
-2.05390573e-01 9.59605694e-01 3.48042488e-01 2.57286757e-01
-9.17562783e-01 -2.53026158e-01 5.86822569e-01 2.91911840e-01
4.15369332e-01 -5.26092827e-01 -3.61524194e-01 -1.11800921e+00
2.42902890e-01 -4.40771133e-01 -2.58903690e-02 -7.20094860e-01
-9.20577347e-01 4.67755824e-01 -8.03316161e-02 -1.78668606e+00
-3.10595393e-01 -3.92154485e-01 -2.10064232e-01 7.56384492e-01
-1.60868800e+00 -1.22326303e+00 -6.37305260e-01 5.22234023e-01
5.55889368e-01 2.49043971e-01 5.12781501e-01 -4.00497988e-02
-5.11618853e-01 2.23108262e-01 1.44274071e-01 -3.69446427e-02
8.10655117e-01 -1.04264808e+00 3.82668942e-01 1.02420461e+00
2.71766558e-02 5.06092906e-01 7.28738964e-01 -7.96055973e-01
-1.39929533e+00 -1.22267842e+00 6.78046167e-01 -3.17870349e-01
3.79231036e-01 -3.81191641e-01 -2.15704933e-01 7.19501793e-01
-4.39419568e-01 3.89155358e-01 1.02796063e-01 7.17425868e-02
-9.43078697e-02 -3.21743518e-01 -1.17453945e+00 5.09447455e-01
1.18511164e+00 -1.20120980e-01 -5.47539108e-02 3.82955194e-01
4.07366455e-01 -9.06077743e-01 -3.60882312e-01 5.57874799e-01
7.31808603e-01 -1.12372482e+00 9.53448296e-01 7.97938481e-02
1.05655141e-01 -8.89355421e-01 -3.67189825e-01 -1.11765480e+00
-1.56515971e-01 -6.42445087e-01 9.67126489e-02 9.44803476e-01
2.93944478e-01 -6.02666497e-01 1.26637232e+00 2.36294389e-01
3.09895687e-02 -7.85901427e-01 -1.09931898e+00 -6.30736649e-01
-5.07468700e-01 -3.55742127e-01 3.36839229e-01 5.54907799e-01
-5.30392289e-01 2.48479858e-01 -4.13455218e-01 5.83968043e-01
7.95970321e-01 4.13496286e-01 1.39330208e+00 -1.20844483e+00
1.17867321e-01 -1.33138433e-01 -6.10730588e-01 -1.79511189e+00
-1.17973678e-01 -2.52240062e-01 6.87057137e-01 -1.50936639e+00
1.99893005e-02 -1.07989001e+00 2.10626006e-01 5.15450478e-01
-2.38370046e-01 2.58754760e-01 8.03939924e-02 6.21631205e-01
-6.89130902e-01 4.42200154e-01 1.00889909e+00 5.04368506e-02
-4.06674176e-01 4.39199537e-01 -2.97175020e-01 8.65732074e-01
4.66478854e-01 -6.15609229e-01 -2.27919146e-01 -4.68419194e-01
-3.17320600e-02 3.13596837e-02 5.53123951e-01 -1.40196085e+00
4.71261948e-01 -4.12259966e-01 7.76511788e-01 -1.13588727e+00
3.78626823e-01 -1.13995659e+00 1.92412615e-01 5.29754996e-01
5.74439624e-03 1.30125642e-01 -4.19584438e-02 9.39984381e-01
-6.76688328e-02 1.01416476e-01 8.79306734e-01 -1.53797209e-01
-8.68888080e-01 3.95752996e-01 -1.35895804e-01 -2.99502075e-01
1.31502330e+00 -8.19265425e-01 3.80619653e-02 -3.16785753e-01
-3.01754266e-01 6.38032079e-01 1.04631233e+00 3.36062580e-01
7.78649688e-01 -1.25761104e+00 -2.55468398e-01 4.90277410e-01
2.49908000e-01 7.42577612e-01 -3.84859741e-01 1.07961512e+00
-8.14600468e-01 2.71413594e-01 7.03474432e-02 -1.22655129e+00
-1.34380519e+00 2.01656416e-01 3.94051880e-01 -2.48124853e-01
-4.95378226e-01 9.82962608e-01 -2.55203433e-02 -7.42117286e-01
5.03228545e-01 -5.16433418e-01 1.37612805e-01 -3.73267800e-01
9.36141908e-02 2.69829899e-01 -5.09801954e-02 -5.16835034e-01
-6.59702599e-01 9.72132087e-01 1.67701878e-02 -6.47670403e-02
1.21043968e+00 -5.03012419e-01 2.20896788e-02 4.28352475e-01
7.15011418e-01 1.30681902e-01 -1.82470918e+00 -1.15896948e-01
-2.12191902e-02 -8.19619179e-01 7.83379525e-02 -5.54559350e-01
-8.78821611e-01 6.07264876e-01 7.94596791e-01 -7.11797029e-02
1.07033718e+00 -1.35003135e-01 5.09765327e-01 4.01437014e-01
1.02447438e+00 -8.04668784e-01 -2.40041167e-01 5.67868829e-01
4.43413943e-01 -1.35374498e+00 3.09335798e-01 -5.89295805e-01
-4.37586665e-01 1.14320242e+00 8.32635760e-01 -1.65404975e-01
3.75675827e-01 2.80090481e-01 -1.92014650e-02 2.53343076e-01
-2.71613479e-01 -2.41236482e-02 -1.64237037e-01 7.46477783e-01
-2.88488716e-03 -7.76105598e-02 3.29175562e-01 -2.81597953e-03
-5.96542731e-02 7.02869445e-02 1.22687615e-01 1.33923650e+00
-6.60267055e-01 -9.35573459e-01 -8.03612113e-01 1.57124430e-01
1.27524376e-01 1.28402606e-01 -3.97597879e-01 8.78938437e-01
1.10356286e-01 7.36108363e-01 -3.09825502e-02 -6.72903419e-01
3.03610235e-01 -3.69511276e-01 5.62247276e-01 -3.71811211e-01
-1.54265493e-01 2.34517023e-01 -1.29746974e-01 -7.90238857e-01
-5.52144110e-01 -6.43072844e-01 -1.13146007e+00 -2.13355139e-01
-8.61388505e-01 -7.31885806e-02 8.03987801e-01 7.34243989e-01
3.70395392e-01 5.26023246e-02 9.04906392e-01 -1.36023164e+00
-5.60964286e-01 -8.30608428e-01 -2.84851402e-01 -2.26288334e-01
7.29555905e-01 -8.82753432e-01 -2.49008521e-01 -7.06803352e-02] | [7.548652172088623, -2.3766424655914307] |
7cd12a08-ff4b-45e5-beb9-604480865b1b | divide-conquer-and-combine-mixture-of | 2306.00434 | null | https://arxiv.org/abs/2306.00434v1 | https://arxiv.org/pdf/2306.00434v1.pdf | Divide, Conquer, and Combine: Mixture of Semantic-Independent Experts for Zero-Shot Dialogue State Tracking | Zero-shot transfer learning for Dialogue State Tracking (DST) helps to handle a variety of task-oriented dialogue domains without the cost of collecting in-domain data. Existing works mainly study common data- or model-level augmentation methods to enhance the generalization but fail to effectively decouple the semantics of samples, limiting the zero-shot performance of DST. In this paper, we present a simple and effective "divide, conquer and combine" solution, which explicitly disentangles the semantics of seen data, and leverages the performance and robustness with the mixture-of-experts mechanism. Specifically, we divide the seen data into semantically independent subsets and train corresponding experts, the newly unseen samples are mapped and inferred with mixture-of-experts with our designed ensemble inference. Extensive experiments on MultiWOZ2.1 upon the T5-Adapter show our schema significantly and consistently improves the zero-shot performance, achieving the SOTA on settings without external knowledge, with only 10M trainable parameters1. | ['Li Guo', 'DaCheng Tao', 'Shi Wang', 'Zheng Lin', 'Yibing Zhan', 'Yanan Cao', 'Liang Ding', 'Qingyue Wang'] | 2023-06-01 | null | null | null | null | ['dialogue-state-tracking'] | ['natural-language-processing'] | [ 1.82751077e-03 4.22904998e-01 -3.62677574e-01 -5.62622607e-01
-8.91907275e-01 -6.65761411e-01 7.31099546e-01 -1.77264571e-01
-3.86254609e-01 7.74792135e-01 5.41774988e-01 -3.61367464e-01
1.01197250e-01 -4.66053277e-01 -2.30104387e-01 -5.48115432e-01
3.04290950e-01 1.00451350e+00 3.83959442e-01 -6.44738376e-01
-2.10816011e-01 -1.74169362e-01 -1.26624024e+00 4.73190159e-01
1.00936449e+00 7.37963140e-01 1.57991052e-02 8.50090265e-01
-4.03945178e-01 9.41906989e-01 -7.40579307e-01 -5.44926047e-01
9.97905582e-02 -5.06279230e-01 -1.03179884e+00 2.60095328e-01
2.99173445e-01 -7.39076495e-01 -5.21746457e-01 7.36905754e-01
6.72024310e-01 5.23111343e-01 6.86566055e-01 -1.30302012e+00
-5.57011187e-01 5.90204656e-01 -4.48835582e-01 -5.56447655e-02
2.62953013e-01 3.98753822e-01 9.68311191e-01 -8.59270394e-01
6.47253096e-01 1.45308578e+00 4.83871311e-01 1.19553244e+00
-1.25221944e+00 -4.98368323e-01 2.07619607e-01 1.22774411e-02
-7.98058450e-01 -8.83812070e-01 5.65038800e-01 -2.23842472e-01
9.73759055e-01 1.40642181e-01 2.10389256e-01 1.67095888e+00
-4.87676591e-01 1.34199011e+00 1.06490052e+00 -4.15112406e-01
3.30813438e-01 3.11449379e-01 5.05131543e-01 6.36781514e-01
-3.74607354e-01 -2.32726157e-01 -7.08894134e-01 -3.18822533e-01
5.15206337e-01 -2.33729873e-02 -1.48256794e-01 -5.17010868e-01
-8.81081462e-01 9.62518930e-01 -8.40920210e-02 -3.97506095e-02
1.71338603e-01 -3.56635392e-01 7.26781726e-01 4.88908648e-01
7.45374024e-01 3.32144409e-01 -8.20787966e-01 -5.03048122e-01
-4.83600646e-01 1.84163660e-01 1.06294191e+00 1.20835161e+00
5.52861273e-01 4.39270809e-02 -5.94594896e-01 1.18848193e+00
1.63227916e-02 3.65041375e-01 5.46295464e-01 -1.12857461e+00
6.93066776e-01 7.54135966e-01 2.50814348e-01 -9.29751322e-02
-1.74380109e-01 -8.98142233e-02 -6.65240288e-01 8.39051902e-02
7.44136274e-01 -6.61132514e-01 -8.15779388e-01 1.99363053e+00
5.74610531e-01 2.72646576e-01 5.23005486e-01 7.98617065e-01
8.10146153e-01 5.16350150e-01 2.05313489e-01 9.00194049e-03
1.51160347e+00 -1.21411324e+00 -9.53204155e-01 -5.50261736e-01
8.41419578e-01 -2.41051391e-01 1.47202635e+00 2.49209896e-01
-9.08615530e-01 -4.08684075e-01 -9.02196229e-01 -2.85896927e-01
-4.32064086e-01 -9.16992128e-02 4.36443210e-01 6.12046182e-01
-7.43290424e-01 4.94414508e-01 -7.20903993e-01 -3.75979424e-01
2.74992704e-01 1.52744111e-02 -3.19615275e-01 1.28407041e-02
-1.55025232e+00 1.21737659e+00 5.19326389e-01 -3.33358467e-01
-1.13691080e+00 -7.54099071e-01 -9.81233120e-01 2.31808230e-01
8.12669337e-01 -6.04646385e-01 1.82675052e+00 -4.63135183e-01
-2.09065199e+00 7.81338573e-01 -2.04768091e-01 -5.33450127e-01
6.43938959e-01 -4.14098769e-01 -1.37278512e-01 -1.34541869e-01
-1.22859009e-01 5.46592474e-01 5.68056762e-01 -1.21355271e+00
-6.70026898e-01 -4.64270771e-01 2.69710720e-01 4.12733734e-01
-6.40399694e-01 -1.33575186e-01 -4.24402773e-01 -2.61353880e-01
-3.07901055e-01 -7.33897209e-01 -1.08416565e-01 -1.49771377e-01
-2.69059330e-01 -4.30774659e-01 9.95410144e-01 -6.62625968e-01
1.14290059e+00 -2.10994554e+00 2.31901467e-01 -4.71171886e-01
2.79778630e-01 5.03139734e-01 -2.73263484e-01 5.95866144e-01
3.37860256e-01 -2.67721444e-01 -2.53372163e-01 -7.38076985e-01
2.49734163e-01 4.76875097e-01 -3.59907240e-01 9.70278084e-02
3.84503484e-01 9.15589094e-01 -1.28875589e+00 -4.05579120e-01
3.71393919e-01 4.34770323e-02 -4.79056269e-01 7.30704129e-01
-5.53285420e-01 5.61035633e-01 -3.56134206e-01 1.78538561e-01
3.76203299e-01 -3.09717357e-01 4.95045364e-01 5.74960932e-02
3.76711309e-01 5.38000882e-01 -9.84454155e-01 2.09868670e+00
-5.07894933e-01 3.02583367e-01 3.14741552e-01 -7.50816226e-01
9.69543457e-01 5.62385798e-01 1.20162614e-01 -3.18435818e-01
1.58619031e-01 -7.24121556e-02 -1.25778571e-01 -6.05384767e-01
5.61603606e-01 -2.98006415e-01 -5.34895658e-01 6.35944247e-01
7.40533054e-01 2.37544999e-01 -1.55746102e-01 3.12759995e-01
9.83017206e-01 2.29308009e-01 4.48944598e-01 -4.96560298e-02
1.38274342e-01 -3.03555969e-02 6.48806274e-01 7.72491872e-01
-4.26205039e-01 3.24295640e-01 5.79809546e-01 -1.44409403e-01
-9.91847873e-01 -1.11000419e+00 9.89772007e-02 1.76310301e+00
7.43208230e-02 -2.27877066e-01 -8.93095016e-01 -1.14236760e+00
-4.09611538e-02 1.23639345e+00 -6.35082006e-01 -4.77957308e-01
-2.14227542e-01 -5.81527829e-01 8.13411295e-01 5.72345495e-01
5.79250753e-01 -7.78666198e-01 -4.14493591e-01 1.80648938e-01
-4.15239245e-01 -1.02315176e+00 -4.27948654e-01 1.55434772e-01
-6.39273942e-01 -9.82841909e-01 -5.69038510e-01 -3.79338473e-01
2.33957052e-01 2.18125939e-01 1.08653522e+00 -3.93564612e-01
1.28327653e-01 2.98593998e-01 -4.95698869e-01 -2.22224563e-01
-5.90759993e-01 2.42129132e-01 1.83135599e-01 9.35955048e-02
6.36994600e-01 -4.92215097e-01 -1.36264071e-01 2.09370703e-01
-4.58202034e-01 1.27434552e-01 1.45627350e-01 1.32645094e+00
-3.73247832e-01 -4.64305758e-01 8.34945917e-01 -1.25156534e+00
8.96183789e-01 -6.36159241e-01 -5.53527512e-02 5.06373107e-01
-4.34813440e-01 3.04869384e-01 5.20589709e-01 -6.35877788e-01
-1.62734938e+00 -2.52558082e-01 1.59294531e-01 -7.05720603e-01
-4.08051640e-01 1.50973633e-01 -5.60888052e-01 5.19112289e-01
6.77390993e-01 2.59672582e-01 2.25356504e-01 -6.03622198e-01
9.09688532e-01 1.09420800e+00 6.54178381e-01 -9.65462863e-01
4.42069590e-01 2.57949799e-01 -7.47362018e-01 -6.60465419e-01
-1.09777141e+00 -6.07431531e-01 -6.85065687e-01 -1.09787077e-01
8.39915454e-01 -9.40072536e-01 -6.56653345e-01 4.97317493e-01
-1.21649063e+00 -6.77241743e-01 -4.70346928e-01 1.94709614e-01
-5.88448882e-01 3.41588557e-01 -9.31680918e-01 -1.26403439e+00
-3.89904797e-01 -8.57157350e-01 1.01074779e+00 2.32454449e-01
-4.28619236e-01 -1.14273274e+00 1.93149596e-01 5.77502608e-01
2.92533278e-01 -2.03299001e-01 9.62065458e-01 -1.43117559e+00
8.23572744e-03 -1.62256919e-02 -8.35514963e-02 2.24855632e-01
7.14261606e-02 -3.72631490e-01 -1.57053912e+00 -3.75133395e-01
1.11779138e-01 -8.53333414e-01 5.41028023e-01 1.78248044e-02
8.10956776e-01 -3.68051171e-01 -1.60885155e-01 2.05931112e-01
8.40569198e-01 1.14439473e-01 2.95458198e-01 -3.52494307e-02
5.43357849e-01 9.57522810e-01 6.30528927e-01 6.40766680e-01
5.21180511e-01 5.72374523e-01 1.04039699e-01 -5.74301854e-02
6.96055964e-03 -4.89856839e-01 3.69938105e-01 8.45802248e-01
1.95045009e-01 -4.44192082e-01 -7.60319889e-01 5.52534461e-01
-2.20719457e+00 -8.68346453e-01 4.26887453e-01 2.03399539e+00
1.16345453e+00 -7.05068000e-04 2.80812025e-01 -2.91610420e-01
7.99302340e-01 3.22538197e-01 -7.34208167e-01 -2.23581746e-01
9.15631801e-02 6.14937171e-02 1.53027862e-01 7.68805444e-01
-9.70296562e-01 1.17940974e+00 6.02598810e+00 1.01997423e+00
-5.84762335e-01 4.93544817e-01 5.11732101e-01 -1.51203796e-01
-1.27891377e-01 1.13860786e-01 -9.58284020e-01 3.51607174e-01
1.05223954e+00 -1.12550013e-01 3.36657137e-01 9.69985604e-01
-1.41640395e-01 2.38294572e-01 -1.19997919e+00 5.05805254e-01
4.52612452e-02 -9.52966869e-01 -1.31575793e-01 -5.76788969e-02
4.88081962e-01 -1.55806825e-01 -1.20825112e-01 1.03886950e+00
9.65532780e-01 -6.90992475e-01 3.49118859e-01 4.63967443e-01
8.00231218e-01 -4.55851585e-01 5.75962245e-01 8.63833904e-01
-5.96159577e-01 -1.68497100e-01 -1.42658442e-01 -1.93036437e-01
1.87333003e-01 -8.22921395e-02 -1.21175063e+00 5.08814633e-01
4.00174350e-01 4.78544503e-01 -1.98786944e-01 3.74654680e-01
-1.71624467e-01 5.85589349e-01 -1.18411370e-01 -1.08459197e-01
2.23839357e-01 -1.55479178e-01 5.55195510e-01 1.16042578e+00
-1.12313375e-01 2.72265643e-01 4.34593260e-01 8.99829626e-01
-1.12354189e-01 -2.40887880e-01 -6.07815862e-01 4.12005037e-02
9.08247709e-01 1.15319276e+00 -1.17609203e-02 -8.36229622e-01
-4.71756101e-01 1.11779833e+00 7.05762684e-01 3.97306502e-01
-6.42578244e-01 -2.96730518e-01 7.89740622e-01 -1.96303844e-01
7.69413263e-02 -1.51226791e-02 -1.92003921e-01 -1.50280523e+00
-2.58892417e-01 -1.04067314e+00 5.64030051e-01 -5.11627078e-01
-1.75413585e+00 4.61701483e-01 2.17661411e-01 -1.02026272e+00
-5.50468862e-01 -6.15613520e-01 -7.46852517e-01 8.36795390e-01
-1.07713866e+00 -1.33052683e+00 -2.14676902e-01 5.02279460e-01
1.08144164e+00 -4.16510582e-01 1.26407981e+00 -5.12546748e-02
-8.05626750e-01 7.06220746e-01 3.30405831e-01 5.03180504e-01
1.05887759e+00 -1.53124940e+00 5.00095904e-01 5.31113505e-01
-2.34432131e-01 5.60137987e-01 6.64386928e-01 -6.10536754e-01
-1.13641059e+00 -8.03169906e-01 5.48482537e-01 -8.79859924e-01
8.06936741e-01 -7.89610684e-01 -1.26402175e+00 8.20813835e-01
4.14631844e-01 -2.18450472e-01 8.99669886e-01 5.86624742e-01
-6.35845602e-01 2.21284285e-01 -1.24019194e+00 6.85609758e-01
9.70922947e-01 -7.42979467e-01 -1.09850562e+00 6.28133584e-03
9.91106927e-01 -4.30783540e-01 -9.46794450e-01 3.19354743e-01
3.09092283e-01 -7.59169996e-01 8.58267307e-01 -1.32209647e+00
3.91515166e-01 1.03564844e-01 -7.63818845e-02 -1.46120715e+00
-8.90392661e-02 -7.03898311e-01 -6.31155372e-01 1.46560991e+00
4.14333701e-01 -5.78705072e-01 8.18244219e-01 9.80689168e-01
-2.00272828e-01 -5.32060504e-01 -9.21655655e-01 -6.61644340e-01
1.08382411e-01 -5.65885231e-02 6.26386762e-01 1.37278199e+00
5.78732312e-01 1.05921328e+00 -7.48620987e-01 -2.30164200e-01
5.59237957e-01 -5.90623543e-02 1.06076002e+00 -1.11543202e+00
-6.38090253e-01 -1.82808399e-01 2.22020164e-01 -1.42755544e+00
3.62935334e-01 -5.77046931e-01 1.78894863e-01 -1.05994081e+00
2.41388500e-01 -4.16280210e-01 -1.20544344e-01 7.39616454e-01
-5.93794942e-01 -4.91723359e-01 1.32050082e-01 2.98490264e-02
-9.71878767e-01 9.59803283e-01 1.11640799e+00 -1.38238177e-01
-2.15380460e-01 -9.35124531e-02 -5.43789864e-01 6.24477267e-01
6.69590950e-01 -1.62391365e-01 -8.28337729e-01 -3.73281837e-01
-4.68645632e-01 4.33957666e-01 7.07271546e-02 -5.87637782e-01
1.35188594e-01 -2.12198466e-01 1.00952862e-02 -2.16082782e-01
8.27071726e-01 -5.35755157e-01 -4.64900255e-01 1.64754003e-01
-6.92351103e-01 -4.55263883e-01 2.19141588e-01 8.25691164e-01
5.92890158e-02 -4.99875993e-01 6.58974528e-01 -2.14154363e-01
-8.04921448e-01 1.37956768e-01 -2.86916077e-01 5.19981980e-01
9.01538372e-01 5.44047207e-02 -7.37375200e-01 -5.44668615e-01
-7.68451571e-01 6.28924608e-01 3.62025410e-01 4.99775529e-01
2.40379333e-01 -1.15904760e+00 -7.29204953e-01 8.61149728e-02
3.92090142e-01 1.81353331e-01 4.86423969e-01 5.88219345e-01
1.88515455e-01 1.29012438e-02 -2.20452502e-01 -4.34892535e-01
-1.05395043e+00 6.66656375e-01 3.17457378e-01 -4.65231508e-01
-6.49230659e-01 7.29574323e-01 1.70289129e-01 -1.10382998e+00
3.50474656e-01 1.40245214e-01 -2.52437383e-01 1.66782275e-01
5.33286631e-01 4.88511473e-01 -2.99594551e-01 -1.38146833e-01
3.16761471e-02 -1.48606196e-01 -3.50197732e-01 -3.74489039e-01
1.16012311e+00 -2.11315334e-01 3.74147028e-01 1.00965941e+00
8.69659960e-01 -4.15486306e-01 -1.70067525e+00 -6.40924096e-01
9.71931741e-02 -2.93156713e-01 -3.37105006e-01 -8.81039023e-01
-4.97147769e-01 1.33112478e+00 2.09720030e-01 3.08291823e-01
6.17372036e-01 -6.35557175e-02 9.43092883e-01 4.25017506e-01
3.65008444e-01 -1.19312584e+00 2.15909600e-01 8.38336825e-01
3.89466971e-01 -1.40854943e+00 -5.44025481e-01 -2.49163106e-01
-1.32846546e+00 7.56517649e-01 1.09022141e+00 9.69165862e-02
1.39130414e-01 3.68161350e-01 1.49730667e-01 8.23115930e-03
-1.23648274e+00 -1.98132083e-01 -8.26039538e-02 6.88126445e-01
2.97831237e-01 1.14468165e-01 2.34390408e-01 1.06171703e+00
1.49084553e-01 -1.71133474e-01 2.37732977e-01 9.16137576e-01
-5.91989756e-01 -1.09172988e+00 -5.60516790e-02 2.50503898e-01
4.31523547e-02 -9.69851241e-02 -4.84635055e-01 7.59606540e-01
-3.47097516e-01 9.28941011e-01 9.31617990e-02 -5.74071109e-01
4.10847187e-01 7.71483064e-01 3.47813725e-01 -7.51169741e-01
-5.97983539e-01 -8.28970969e-02 6.48121834e-01 -1.91652581e-01
-3.86066958e-02 -2.91776925e-01 -1.02971697e+00 -3.16658974e-01
-4.88753170e-01 3.91851217e-01 6.51140511e-02 1.14865029e+00
5.72526157e-01 4.76266772e-01 5.21494031e-01 -4.70806748e-01
-1.29393494e+00 -1.55836856e+00 -3.95789444e-01 6.61164045e-01
2.00008929e-01 -7.58255124e-01 -3.58165443e-01 -1.37803480e-01] | [12.760828971862793, 7.916084289550781] |
2b895baa-cd7f-4e2d-ae11-06dc2a4a7bb2 | risk-averse-model-uncertainty-for | 2301.12593 | null | https://arxiv.org/abs/2301.12593v1 | https://arxiv.org/pdf/2301.12593v1.pdf | Risk-Averse Model Uncertainty for Distributionally Robust Safe Reinforcement Learning | Many real-world domains require safe decision making in the presence of uncertainty. In this work, we propose a deep reinforcement learning framework for approaching this important problem. We consider a risk-averse perspective towards model uncertainty through the use of coherent distortion risk measures, and we show that our formulation is equivalent to a distributionally robust safe reinforcement learning problem with robustness guarantees on performance and safety. We propose an efficient implementation that only requires access to a single training environment, and we demonstrate that our framework produces robust, safe performance on a variety of continuous control tasks with safety constraints in the Real-World Reinforcement Learning Suite. | ['Mouhacine Benosman', 'James Queeney'] | 2023-01-30 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-6.76561818e-02 5.55851579e-01 -1.27335802e-01 -2.43215114e-01
-1.35064280e+00 -6.13237381e-01 5.83706498e-01 2.58415490e-01
-7.09579945e-01 1.11053860e+00 -1.17604606e-01 -4.79230374e-01
-5.77254236e-01 -7.54225254e-01 -1.02098489e+00 -6.04665220e-01
-4.41928804e-01 3.90515924e-01 -3.64413075e-02 -8.38427246e-02
3.00216019e-01 2.24819213e-01 -1.09516704e+00 -3.67228389e-01
8.69965792e-01 1.16676736e+00 -4.89818811e-01 6.35950267e-01
9.13472354e-01 8.00931454e-01 -5.24196863e-01 -2.70923972e-01
6.38093352e-01 1.60000026e-01 -7.40584731e-01 -4.82320823e-02
1.97470650e-01 -7.78401434e-01 -1.06679156e-01 1.35375452e+00
4.36453521e-01 5.30793548e-01 7.26477206e-01 -1.55835283e+00
-3.28822047e-01 5.03001511e-01 -3.66213769e-01 -6.81402460e-02
6.97636977e-02 5.47797322e-01 9.71667647e-01 -4.05079991e-01
2.68101096e-01 1.27778745e+00 5.09822309e-01 7.62089610e-01
-1.20932651e+00 -4.47858512e-01 3.59024048e-01 -2.60938942e-01
-1.08232260e+00 -2.99418837e-01 1.73686355e-01 -5.47512770e-01
1.01108849e+00 -2.00627312e-01 2.92669863e-01 1.18076921e+00
7.23056734e-01 5.88675618e-01 1.10676503e+00 -1.90804183e-01
1.02413380e+00 -1.30482256e-01 -2.90917337e-01 4.31040436e-01
4.09727514e-01 1.10728014e+00 3.80235463e-02 -3.50975245e-01
4.78446454e-01 -2.45364338e-01 7.35751465e-02 -7.77325153e-01
-8.37661505e-01 1.15936410e+00 2.30163693e-01 -2.81775802e-01
-2.74722993e-01 7.11095512e-01 4.73486304e-01 5.81161261e-01
3.51527542e-01 6.91187382e-01 -2.72487760e-01 -1.29247576e-01
-3.11328769e-01 7.37457693e-01 8.19967747e-01 9.66787040e-01
1.72052893e-03 4.86834615e-01 -1.57745332e-01 2.43258476e-01
6.31003618e-01 5.76522291e-01 1.65132396e-02 -1.50970447e+00
6.49125874e-01 -2.72407711e-01 8.45698893e-01 -6.26318812e-01
-3.43778729e-01 -3.24597538e-01 -2.69275755e-01 1.03261220e+00
2.58137256e-01 -6.91964924e-01 -6.32250309e-01 2.11870742e+00
3.11853439e-01 1.83940485e-01 3.66951227e-01 5.00769734e-01
-4.79305089e-01 6.43609583e-01 1.63416952e-01 -4.66411263e-01
5.62042475e-01 -5.20388901e-01 -7.29428530e-01 -1.64811790e-01
3.45650941e-01 -6.96156397e-02 9.50444639e-01 6.99237466e-01
-1.23953569e+00 -1.14035338e-01 -1.38473535e+00 3.48300010e-01
4.10465943e-03 -6.92776322e-01 2.80661345e-01 6.28643632e-01
-9.65400577e-01 9.18165028e-01 -9.37016249e-01 1.36006817e-01
4.94928777e-01 2.96263397e-01 -2.16998041e-01 2.16877371e-01
-1.32572281e+00 1.34035611e+00 6.48201585e-01 -8.12291913e-03
-1.89647377e+00 -7.23199904e-01 -1.13842332e+00 1.56914610e-02
1.04516137e+00 -5.07296026e-01 1.91841972e+00 -4.03899848e-01
-1.66377115e+00 1.87372521e-01 8.48773837e-01 -9.83620226e-01
1.13939893e+00 -6.49253845e-01 -1.30420670e-01 1.14207920e-02
1.69714424e-03 2.99065977e-01 1.11904550e+00 -1.10916233e+00
-8.46423686e-01 -1.06447853e-01 3.84342879e-01 2.36418575e-01
2.20353566e-02 -1.52347665e-02 5.33451676e-01 -6.61107838e-01
-9.23101127e-01 -1.04152954e+00 -6.33735538e-01 5.24610616e-02
-1.41105875e-01 -2.20740706e-01 4.38589543e-01 -3.16817313e-01
9.33742940e-01 -2.03773808e+00 8.33523944e-02 3.20641845e-01
-6.82613179e-02 1.22143790e-01 -3.27260196e-02 3.03373843e-01
-2.84595024e-02 2.76476234e-01 -7.94714987e-01 -1.25409618e-01
7.23413527e-01 2.18671098e-01 -7.83024848e-01 8.60666990e-01
3.76174808e-01 5.40001690e-01 -9.51883733e-01 -1.64590985e-01
1.98324159e-01 4.80071362e-03 -7.40043283e-01 4.29178685e-01
-6.00313962e-01 2.16205388e-01 -5.75457156e-01 1.58499584e-01
4.48482573e-01 5.95682800e-01 1.43032279e-02 7.36083686e-01
1.17331348e-01 -1.61228441e-02 -1.27975118e+00 1.32959819e+00
-5.92181563e-01 1.78691614e-02 2.38513798e-01 -8.47858310e-01
5.81368864e-01 2.46071860e-01 3.00800234e-01 -4.19111103e-01
2.43676916e-01 6.84317127e-02 -1.24151297e-01 -2.41946623e-01
9.04710516e-02 -7.67023385e-01 -6.53060794e-01 7.23735809e-01
-1.79311391e-02 -7.11629689e-01 -7.22799450e-02 -1.34414822e-01
1.00141859e+00 1.83449283e-01 5.41766882e-01 -5.36194086e-01
3.97023372e-02 -3.95660281e-01 8.66766453e-01 8.90400171e-01
-6.67311132e-01 2.60835495e-02 8.46113265e-01 -2.89319903e-01
-9.50988770e-01 -1.49197435e+00 -1.49491712e-01 7.47737050e-01
-1.42326415e-01 1.58871263e-02 -7.95991063e-01 -9.80584562e-01
4.17283446e-01 1.38306653e+00 -6.95422351e-01 -6.74298823e-01
-4.08228844e-01 -3.91732603e-01 4.18334752e-01 6.90815210e-01
1.40877873e-01 -8.39443445e-01 -1.02346885e+00 1.66857585e-01
6.32684886e-01 -6.83747590e-01 -5.86446762e-01 3.83286864e-01
-5.93636870e-01 -1.14902854e+00 -2.72846580e-01 -3.01903784e-01
2.57651836e-01 -5.36980808e-01 8.92551839e-01 -5.15324056e-01
-1.74104691e-01 6.74907684e-01 2.05297798e-01 -7.64673948e-01
-5.30448675e-01 -5.92263877e-01 5.00278592e-01 -3.41669708e-01
-1.05852470e-01 -1.83375269e-01 -2.64919847e-01 5.66606820e-02
-1.00269711e+00 -6.56417966e-01 -4.83538546e-02 8.46829832e-01
5.94772816e-01 5.24912357e-01 8.95829380e-01 -6.26172960e-01
1.09464276e+00 -6.55960023e-01 -1.42751181e+00 1.63285434e-01
-6.91757679e-01 3.44822764e-01 7.19897211e-01 -1.33940816e-01
-1.10095930e+00 3.02825626e-02 7.67746717e-02 -3.61332476e-01
-1.21222980e-01 3.73703271e-01 -3.46632212e-01 1.54279634e-01
6.59792364e-01 -2.05866262e-01 1.65772483e-01 -3.90997268e-02
3.76106590e-01 3.50756139e-01 3.52045864e-01 -1.15036285e+00
8.42580855e-01 1.20822817e-01 1.81403726e-01 -2.53501177e-01
-7.32623875e-01 3.75102878e-01 -2.00298309e-01 -5.96068762e-02
7.16728866e-01 -8.66053343e-01 -9.98940170e-01 2.63296366e-01
-6.44318342e-01 -6.41770482e-01 -6.28102601e-01 4.45252597e-01
-1.58154047e+00 2.60779560e-01 -3.84242505e-01 -1.18109691e+00
-1.86241195e-01 -1.20191050e+00 8.47264409e-01 -8.33229423e-02
-2.84594577e-03 -1.01096642e+00 4.29050267e-01 -1.95643321e-01
2.12048933e-01 6.62798345e-01 8.32245886e-01 -6.24325812e-01
-3.43754023e-01 -4.68457229e-02 2.75676489e-01 7.45122373e-01
-9.87810642e-02 -1.18803475e-02 -8.29985261e-01 -6.38214886e-01
5.22664726e-01 -8.55066776e-01 6.99108422e-01 3.96675020e-01
1.36422932e+00 -6.98000610e-01 3.09414566e-01 4.25993562e-01
1.45642185e+00 5.44277668e-01 1.55693695e-01 3.80323559e-01
2.69933760e-01 7.06140041e-01 1.21024013e+00 7.88781345e-01
2.30716735e-01 3.53892833e-01 8.62811744e-01 4.39078003e-01
8.66356373e-01 -3.43387008e-01 6.22591496e-01 -1.53284907e-01
3.44031036e-01 -1.57253474e-01 -8.77038479e-01 5.25504053e-01
-2.19259644e+00 -1.08286750e+00 9.02917802e-01 2.49482203e+00
9.49321628e-01 3.86791587e-01 3.44752133e-01 -7.26856515e-02
5.79999804e-01 -2.97724269e-02 -1.12350190e+00 -9.38260138e-01
4.58562911e-01 1.19852990e-01 6.30803168e-01 9.12303209e-01
-1.51495087e+00 5.83923697e-01 7.86905813e+00 6.16948664e-01
-5.59876740e-01 -2.22137183e-01 7.66195595e-01 -4.19423789e-01
-3.33697259e-01 -4.13861424e-01 -4.32974368e-01 3.09869796e-01
1.23730826e+00 -8.04523766e-01 4.28053528e-01 1.31682920e+00
2.55426466e-01 -1.79624632e-02 -1.55039167e+00 5.75856388e-01
-5.34946442e-01 -8.97788227e-01 -4.03517753e-01 -5.04595339e-02
6.91496611e-01 -1.40023842e-01 5.79305708e-01 5.09990513e-01
1.21608412e+00 -1.37942779e+00 9.38511908e-01 3.04332495e-01
5.48489332e-01 -1.52157080e+00 5.66445768e-01 4.17665988e-01
-5.73636770e-01 -5.73455215e-01 -3.05716902e-01 -2.08301935e-02
7.58348554e-02 2.91570902e-01 -6.88254297e-01 3.38128716e-01
3.32465470e-01 6.86747372e-01 1.36387246e-02 9.98304844e-01
-3.41955036e-01 3.51852953e-01 -3.24383199e-01 2.27231637e-01
5.57498574e-01 6.00547083e-02 4.22405869e-01 9.29807901e-01
3.34111989e-01 -1.40020058e-01 5.68677664e-01 9.11388874e-01
1.13260940e-01 -4.53513384e-01 -1.16727877e+00 1.04270965e-01
3.02961588e-01 6.85103834e-01 -1.80735245e-01 1.50330337e-02
4.76408377e-03 3.62669051e-01 4.30016428e-01 2.34208971e-01
-1.12259793e+00 -4.36038673e-01 9.96528268e-01 -5.78122675e-01
2.60677576e-01 -2.59845316e-01 -1.87222853e-01 -9.28423822e-01
7.28949625e-03 -1.07601953e+00 8.30612540e-01 -1.92720324e-01
-1.64728272e+00 1.79340482e-01 2.62354136e-01 -1.02796960e+00
-6.93284750e-01 -7.94668734e-01 -6.03669524e-01 7.61082530e-01
-1.39714587e+00 -3.62892658e-01 6.96512580e-01 7.10252583e-01
3.27836961e-01 -2.03912139e-01 6.70267105e-01 -3.00493300e-01
-6.28086567e-01 4.68355209e-01 3.12014014e-01 -3.66547465e-01
6.57456815e-01 -1.80989742e+00 3.94339293e-01 9.28190172e-01
-3.82999122e-01 3.60251069e-01 9.55417395e-01 -4.35341984e-01
-1.37594676e+00 -1.43554783e+00 3.46836261e-02 -5.38578451e-01
9.86136019e-01 -2.59019345e-01 -5.89004755e-01 9.23373938e-01
2.04419419e-01 1.44508868e-01 3.95272344e-01 -1.37712270e-01
-4.48741049e-01 -1.60429403e-01 -1.73784351e+00 5.94382763e-01
7.89923489e-01 -4.96262580e-01 -9.91204560e-01 3.02984923e-01
1.09942937e+00 -4.56549555e-01 -9.40488160e-01 4.96199191e-01
1.15762979e-01 -4.98524010e-01 8.38359833e-01 -1.03810143e+00
2.71788627e-01 -2.99853206e-01 -2.55279869e-01 -1.76100755e+00
-1.85423598e-01 -9.11266983e-01 -2.30381936e-01 7.26161540e-01
3.10635597e-01 -6.96831346e-01 2.47397035e-01 9.78743613e-01
-2.00513050e-01 -7.05537379e-01 -1.36999190e+00 -1.19642103e+00
9.44301486e-01 -5.85064650e-01 4.68849182e-01 5.65128088e-01
3.32550347e-01 -3.33127171e-01 -5.48190355e-01 2.34123036e-01
9.82890368e-01 -1.01791583e-01 1.89410746e-01 -6.60016239e-01
-4.82865214e-01 -3.69800508e-01 -1.18883185e-01 -1.94342241e-01
6.29518449e-01 -4.84862119e-01 5.97875178e-01 -1.08954668e+00
-1.26134992e-01 -2.82397270e-01 -6.41970992e-01 5.13473868e-01
-1.37322741e-02 -4.91256773e-01 2.46123523e-01 -5.56154191e-01
-9.52041745e-01 1.01031816e+00 8.67573977e-01 -2.51128614e-01
1.69788107e-01 2.89155006e-01 -9.16833103e-01 8.52117538e-01
1.01029897e+00 -3.07190657e-01 -9.17912960e-01 -3.76916528e-01
2.91556746e-01 1.53807521e-01 2.14564219e-01 -8.96802008e-01
-1.69943139e-01 -8.95909488e-01 1.10910378e-01 1.94909833e-02
-2.53114291e-03 -9.96387362e-01 -2.21827820e-01 9.00709927e-01
-8.42832863e-01 1.83238283e-01 4.65753496e-01 8.99717391e-01
6.97503537e-02 -2.54095763e-01 1.22498274e+00 7.55914003e-02
-5.28059721e-01 4.94395226e-01 -4.68251377e-01 4.95717406e-01
1.66647601e+00 6.00720942e-01 -2.50247180e-01 -5.38079917e-01
-7.17703164e-01 7.54147589e-01 3.34799230e-01 4.04079795e-01
7.04276502e-01 -1.20661807e+00 -6.65688038e-01 -8.81342739e-02
6.89772069e-02 -2.40401807e-03 -1.76827721e-02 3.93197909e-02
-1.55687436e-01 2.94613749e-01 -3.77748072e-01 -2.07778409e-01
-6.26055360e-01 1.09652090e+00 6.71880543e-01 -2.48821750e-01
-4.42777544e-01 6.60993040e-01 3.36510569e-01 -4.18560743e-01
6.30548120e-01 -4.86421764e-01 2.50286847e-01 -1.96773320e-01
7.42977738e-01 4.60207552e-01 -8.13508034e-02 -1.23393713e-02
-1.88225195e-01 8.60375538e-02 -1.09769896e-01 -5.41512549e-01
1.23528934e+00 -2.11744700e-02 4.72827166e-01 4.01184201e-01
8.55569065e-01 -5.03195167e-01 -2.02982378e+00 1.46339819e-01
3.44330251e-01 -5.53708851e-01 -2.65793819e-02 -9.41449046e-01
-6.99092925e-01 7.97394514e-01 5.45177102e-01 7.18344674e-02
9.05678809e-01 -4.59044486e-01 3.23406637e-01 6.97280288e-01
7.26313055e-01 -1.56684160e+00 -1.95574816e-02 6.30092621e-01
1.22584116e+00 -1.17421699e+00 -1.97474971e-01 2.53532380e-01
-1.05243993e+00 8.10897708e-01 8.18697274e-01 -5.95323980e-01
8.12878668e-01 5.65838337e-01 -2.80423224e-01 3.00103337e-01
-1.18045533e+00 1.05542824e-01 -1.92346111e-01 9.38674092e-01
-2.27473855e-01 2.44354144e-01 -4.94254287e-03 5.31999826e-01
-7.28545189e-02 -5.04588448e-02 7.54297078e-01 1.12973773e+00
-6.61922574e-01 -8.81632507e-01 -2.83043027e-01 8.33054483e-02
-5.99162459e-01 2.73860186e-01 8.07637721e-03 7.05160856e-01
-2.80003130e-01 1.05527794e+00 -1.71741620e-02 -6.08386062e-02
3.30423981e-01 -4.43231761e-02 5.90286434e-01 -6.48353994e-01
-2.44111970e-01 -1.31076887e-01 1.11048639e-01 -9.61395025e-01
-9.92371235e-04 -5.97916007e-01 -1.21141064e+00 -3.39938074e-01
2.11140886e-01 -2.91215591e-02 3.03136945e-01 9.44785237e-01
1.63940772e-01 4.50774312e-01 9.91251469e-01 -3.73656988e-01
-1.88832295e+00 -4.50499654e-01 -8.51561010e-01 8.66354927e-02
8.24217141e-01 -8.66580367e-01 -4.20362383e-01 -4.30284381e-01] | [4.472442150115967, 2.3362348079681396] |
eae68543-7524-4ffd-a498-fcc611945083 | learning-structural-graph-layouts-and-3d | 1907.03387 | null | https://arxiv.org/abs/1907.03387v2 | https://arxiv.org/pdf/1907.03387v2.pdf | Learning Structural Graph Layouts and 3D Shapes for Long Span Bridges 3D Reconstruction | A learning-based 3D reconstruction method for long-span bridges is proposed in this paper. 3D reconstruction generates a 3D computer model of a real object or scene from images, it involves many stages and open problems. Existing point-based methods focus on generating 3D point clouds and their reconstructed polygonal mesh or fitting-based geometrical models in urban scenes civil structures reconstruction within Manhattan world constrains and have made great achievements. Difficulties arise when an attempt is made to transfer these systems to structures with complex topology and part relations like steel trusses and long-span bridges, this could be attributed to point clouds are often unevenly distributed with noise and suffer from occlusions and incompletion, recovering a satisfactory 3D model from these highly unstructured point clouds in a bottom-up pattern while preserving the geometrical and topological properties makes enormous challenge to existing algorithms. Considering the prior human knowledge that these structures are in conformity to regular spatial layouts in terms of components, a learning-based topology-aware 3D reconstruction method which can obtain high-level structural graph layouts and low-level 3D shapes from images is proposed in this paper. We demonstrate the feasibility of this method by testing on two real long-span steel truss cable-stayed bridges. | ['Yong Huang', 'Fangqiao Hu', 'Jin Zhao', 'Hui Li'] | 2019-07-08 | null | null | null | null | ['generating-3d-point-clouds'] | ['computer-vision'] | [-4.39377911e-02 3.28540444e-01 3.81718099e-01 -1.00445282e-02
-4.94560122e-01 -3.39045376e-01 2.99503595e-01 2.30938762e-01
3.21565956e-01 6.82476103e-01 -3.17010075e-01 -3.28481823e-01
-6.89357102e-01 -1.15938044e+00 -8.48360956e-01 -2.21723244e-01
-5.09137213e-01 1.25490904e+00 6.30766988e-01 -5.04566491e-01
3.80038559e-01 1.29108775e+00 -1.50741100e+00 -6.60892129e-02
7.66090512e-01 3.63495827e-01 2.71132231e-01 3.40921909e-01
-5.62876344e-01 3.91860038e-01 -2.61130184e-01 -3.29721421e-01
4.96547073e-01 1.36425374e-02 -8.27738404e-01 8.24755609e-01
5.17424822e-01 4.37089950e-02 -4.67482567e-01 7.73305476e-01
2.06777558e-01 -1.02734186e-01 9.94486630e-01 -1.12728012e+00
-1.59644708e-01 2.03163072e-01 -7.22417891e-01 -5.99322379e-01
2.11354375e-01 -9.65114385e-02 6.81254923e-01 -1.11357355e+00
7.75456667e-01 1.53356433e+00 1.01572597e+00 5.33788167e-02
-1.42481446e+00 -1.99549630e-01 -1.22876965e-01 2.50043929e-01
-1.54656029e+00 6.55912934e-03 1.40036428e+00 -6.13581300e-01
5.13035297e-01 3.48769605e-01 9.22209680e-01 2.22417578e-01
2.40741715e-01 3.94178927e-01 1.08628333e+00 -4.97676075e-01
2.99186021e-01 -3.94470334e-01 -5.11904433e-02 7.44309723e-01
3.22425187e-01 -1.05007827e-01 -2.37662166e-01 -6.02028631e-02
1.34248233e+00 3.12981941e-02 -1.63425729e-01 -8.89731646e-01
-9.69019234e-01 6.64606690e-01 8.07041168e-01 6.59077018e-02
-3.32214475e-01 -1.48562416e-02 1.79114208e-01 3.81800920e-01
4.95517224e-01 9.84156430e-02 -2.94156969e-01 2.92804480e-01
-9.14886177e-01 4.34094369e-01 6.93176091e-01 1.16533279e+00
1.31822407e+00 4.36355501e-01 7.15153217e-01 8.85154426e-01
3.38665187e-01 5.08906841e-01 -4.07602996e-01 -1.04625607e+00
4.13182884e-01 1.09550929e+00 -2.53140986e-01 -1.35140109e+00
-5.90801418e-01 -4.08981323e-01 -9.82302189e-01 8.51310194e-01
1.18511967e-01 3.75401467e-01 -8.88616562e-01 8.14317584e-01
7.65395820e-01 3.80622484e-02 -2.36076415e-01 7.44573593e-01
8.04272950e-01 4.00839120e-01 -3.77235413e-01 -3.93359773e-02
1.05591834e+00 -4.47008669e-01 -2.68408626e-01 -8.33990052e-02
5.03450036e-01 -6.71164572e-01 1.01576424e+00 1.65403292e-01
-1.14778173e+00 -4.75862503e-01 -9.90400076e-01 8.96772370e-02
-8.91218334e-03 -2.78082162e-01 2.24770337e-01 2.61227280e-01
-8.79635513e-01 8.09235573e-01 -5.54798067e-01 -6.24132864e-02
6.52860880e-01 1.83095798e-01 -6.44289553e-01 -4.63315010e-01
-5.26639104e-01 9.94839132e-01 2.97716945e-01 3.13082755e-01
-6.05910242e-01 -7.60274351e-01 -6.93998754e-01 -9.95277837e-02
4.45315003e-01 -8.54952633e-01 4.68560487e-01 -2.20683947e-01
-1.06498682e+00 9.72016752e-01 2.27270365e-01 -2.48286240e-02
8.19413126e-01 1.49880633e-01 8.32332019e-03 1.25533283e-01
2.69532613e-02 1.72681957e-01 6.20502412e-01 -1.91916621e+00
-1.53118238e-01 -6.64924264e-01 -1.02614008e-01 -5.21795414e-02
2.96757787e-01 -7.02393055e-01 -1.31873131e-01 -4.45540279e-01
9.59917963e-01 -7.53979266e-01 -6.64186478e-01 3.82472843e-01
-5.83540082e-01 -1.67708158e-01 1.45116067e+00 -5.99822640e-01
6.13336325e-01 -1.83030474e+00 1.21859759e-01 7.17070937e-01
2.87036031e-01 9.02684703e-02 -4.72936518e-02 1.00635564e+00
-1.30908117e-01 5.16262352e-02 -7.09069133e-01 -1.85313225e-01
-5.38875945e-02 5.59411764e-01 1.73893809e-01 6.61861718e-01
-7.00585097e-02 5.35438120e-01 -5.47337294e-01 -9.52299058e-01
4.40328717e-01 4.69320953e-01 -4.48083371e-01 9.42004174e-02
-1.15678899e-01 2.97052383e-01 -6.35964930e-01 8.69358063e-01
1.05639410e+00 2.02256620e-01 5.17961755e-02 -5.35865664e-01
-2.93999463e-01 -5.05253553e-01 -1.45990229e+00 1.81188750e+00
-4.88053262e-01 3.32883269e-01 4.43650544e-01 -1.44383597e+00
1.35163713e+00 2.14341372e-01 7.32615352e-01 -4.98915851e-01
-1.34971231e-01 2.90219903e-01 -3.36640328e-01 -5.79714060e-01
1.71734348e-01 -4.31554765e-01 1.86680201e-02 1.46923780e-01
-4.87634987e-01 -9.59252954e-01 -1.57211125e-01 3.57235581e-01
1.03853095e+00 3.43710870e-01 -1.78138509e-01 -5.16629815e-01
4.50410306e-01 5.61114788e-01 5.34780443e-01 1.23455822e-01
6.90632820e-01 8.33403111e-01 2.24824160e-01 -8.88963997e-01
-1.45371354e+00 -1.35243034e+00 -1.62048876e-01 -1.34294361e-01
2.58392036e-01 -1.84235424e-01 -5.11763036e-01 -3.83542776e-01
1.21937692e-01 3.44084501e-01 -2.36837596e-01 1.60277277e-01
-9.82150674e-01 -2.37468541e-01 -3.32723698e-03 3.56891364e-01
3.47491205e-01 -6.87656462e-01 -6.11068666e-01 5.21060705e-01
-4.91798185e-02 -1.07904387e+00 1.11279674e-01 -1.27725527e-01
-1.40093398e+00 -1.50327897e+00 -4.87891048e-01 -1.06775951e+00
1.05745113e+00 3.03925663e-01 1.38008463e+00 7.19647467e-01
-5.06297529e-01 3.42758834e-01 -3.34140778e-01 -2.87491083e-01
-2.88288057e-01 -2.34126836e-01 -2.67970920e-01 -1.09524988e-01
-3.64601433e-01 -1.12197399e+00 -4.31633323e-01 4.54823524e-01
-8.83599281e-01 1.94087878e-01 5.69698393e-01 6.90650105e-01
9.62394357e-01 4.97363240e-01 4.87299889e-01 -7.50264764e-01
2.42422909e-01 -3.01721066e-01 -6.34546220e-01 1.56020597e-01
-1.91051647e-01 -2.82310218e-01 4.39748973e-01 -7.83798546e-02
-7.03386843e-01 3.94686460e-01 -2.76133776e-01 -7.22879410e-01
-1.93448201e-01 4.57308590e-01 -3.67105812e-01 -1.46743730e-01
5.26512623e-01 9.93988737e-02 4.05721813e-01 -7.38837719e-01
2.68301845e-01 3.28657866e-01 5.70627391e-01 -8.02449644e-01
1.54693794e+00 7.10794091e-01 6.58775091e-01 -1.40981281e+00
-4.86789852e-01 -1.81305751e-01 -1.23432660e+00 -8.26779008e-01
5.27518094e-01 -6.91894114e-01 -4.79764819e-01 2.65755236e-01
-1.12664294e+00 -2.71888465e-01 -5.19321561e-01 2.76812553e-01
-7.92917669e-01 6.12074435e-01 -3.46908391e-01 -8.57722700e-01
-2.34770507e-01 -9.57195163e-01 1.07508898e+00 -3.61254871e-01
-1.69016179e-02 -1.05695367e+00 5.84162725e-03 5.37603974e-01
1.20068848e-01 8.90550256e-01 1.48380601e+00 1.98302656e-01
-1.01134396e+00 -2.26155549e-01 1.57516539e-01 9.82231274e-02
9.15415734e-02 7.86426142e-02 -3.08779866e-01 -2.00514406e-01
-1.48507997e-01 -4.51531410e-02 2.20396236e-01 2.55002201e-01
8.48822236e-01 -1.05405971e-01 -4.11360919e-01 2.47809604e-01
1.79083753e+00 -3.31824094e-01 8.69499683e-01 3.01276952e-01
8.93568099e-01 9.89395976e-01 4.63464230e-01 3.60314101e-01
4.26799566e-01 7.30377018e-01 1.00148129e+00 -3.04476202e-01
-4.89712834e-01 -5.64066052e-01 -3.06578159e-01 8.62892985e-01
-3.08611691e-01 1.92361563e-01 -1.10191655e+00 7.57958353e-01
-1.65233648e+00 -8.83915663e-01 -1.09056652e+00 2.08701396e+00
3.68096799e-01 2.32954234e-01 5.18897176e-02 7.71252811e-01
5.86148620e-01 -2.44938686e-01 -5.35464287e-01 -6.53257668e-02
-2.80905485e-01 1.71117261e-01 3.34692925e-01 5.92235267e-01
-4.28805977e-01 9.12803113e-01 5.54121208e+00 1.04270840e+00
-7.39851475e-01 -1.75921947e-01 2.61093497e-01 4.24721748e-01
-6.95776343e-01 3.47664624e-01 -3.46691072e-01 6.01621903e-02
2.75397241e-01 2.76674747e-01 5.48839793e-02 7.85952985e-01
4.43675011e-01 1.32227644e-01 -6.33312821e-01 1.14935219e+00
-1.92344129e-01 -1.81933951e+00 1.47618040e-01 3.85157198e-01
6.00961626e-01 -1.95021987e-01 -5.53054690e-01 -5.99837191e-02
2.91158885e-01 -9.79491293e-01 1.03178990e+00 6.57415330e-01
7.14798987e-01 -8.20655525e-01 3.80863190e-01 7.87454665e-01
-1.43113971e+00 6.65432513e-02 -5.36434293e-01 -8.85499418e-02
7.42274940e-01 1.13194370e+00 -7.50941575e-01 1.08268142e+00
7.46422768e-01 6.09038711e-01 -4.05400842e-01 1.38964248e+00
8.94475430e-02 3.71282935e-01 -5.76000988e-01 3.60885382e-01
1.07365429e-01 -7.03402758e-01 6.63118303e-01 6.80259943e-01
5.02125919e-01 3.41893762e-01 4.49875444e-01 9.03620780e-01
1.61538154e-01 2.49756485e-01 -9.38050032e-01 7.20337749e-01
4.90421653e-01 1.11006927e+00 -1.00313294e+00 -7.00687915e-02
-1.45496249e-01 8.73909742e-02 3.14277738e-01 1.11288391e-01
-5.31262755e-01 2.73193084e-02 2.72134274e-01 1.09814560e+00
2.06117392e-01 -7.80729294e-01 -7.17191696e-01 -6.55022025e-01
7.18042627e-02 -5.10875404e-01 1.51526928e-01 -9.77085948e-01
-1.36268675e+00 6.16512060e-01 3.37467283e-01 -1.62234592e+00
3.51654351e-01 -4.16580200e-01 -7.64838576e-01 4.22509164e-01
-1.46505308e+00 -1.56532490e+00 -2.24765450e-01 6.75588250e-01
9.46821988e-01 -5.61987422e-02 4.91924673e-01 1.77588850e-01
-9.13921595e-02 -2.27055594e-01 -7.10220039e-02 -3.47355641e-02
-7.91563839e-02 -9.48188305e-01 2.00609323e-02 4.62859631e-01
1.20535173e-01 -1.00560756e-02 7.09792316e-01 -7.75742292e-01
-1.82252336e+00 -1.14746356e+00 5.79797387e-01 -8.68489295e-02
1.90821961e-01 -2.54921079e-01 -1.26969624e+00 4.43783075e-01
-1.64192431e-02 -6.33706301e-02 1.13933884e-01 -1.59950644e-01
-2.16291863e-02 -9.47842896e-02 -1.29057050e+00 5.29643297e-01
1.56992996e+00 7.30634760e-03 -6.86193943e-01 5.38550496e-01
2.51914039e-02 -3.55864167e-01 -1.10135436e+00 6.87747598e-01
3.15905154e-01 -9.21459138e-01 1.35836542e+00 -3.39526355e-01
3.03792626e-01 -6.43170714e-01 -9.95667651e-02 -1.35372734e+00
-2.20418021e-01 -4.51346189e-01 4.20330316e-01 9.78501141e-01
3.44360434e-02 -3.81694257e-01 1.09353626e+00 3.01920265e-01
-7.15273976e-01 -7.35003293e-01 -1.23697686e+00 -8.34459066e-01
5.20740330e-01 -3.71007383e-01 7.10979044e-01 8.88172090e-01
-5.61923623e-01 2.92000622e-01 -1.01386070e-01 4.23219681e-01
1.28126693e+00 4.51815814e-01 1.24618447e+00 -1.93481612e+00
4.97385651e-01 -7.33827055e-02 -7.51998603e-01 -1.08537376e+00
2.31519848e-01 -5.99790156e-01 -5.59598906e-03 -1.90951574e+00
-3.92846853e-01 -1.10129988e+00 6.85636878e-01 2.46080443e-01
5.35653114e-01 2.39203364e-01 -9.67322960e-02 4.31117088e-01
4.79613617e-02 8.00081372e-01 1.51303399e+00 -3.30455869e-01
7.83032626e-02 9.39565618e-03 -7.80999213e-02 9.36675191e-01
6.12656415e-01 -4.29550767e-01 -5.13934672e-01 -6.00740850e-01
8.44154507e-02 4.34765935e-01 5.35225689e-01 -1.13963461e+00
3.56171757e-01 -2.00333685e-01 7.83856660e-02 -1.29531538e+00
6.43574893e-01 -1.58264899e+00 8.96211386e-01 3.73602360e-01
2.47627825e-01 3.77347767e-01 -8.99140462e-02 6.87523603e-01
-1.98514879e-01 -2.24655777e-01 8.87195408e-01 -6.95237398e-01
-5.86476326e-01 4.21834767e-01 -1.78690061e-01 -2.13205129e-01
1.13811851e+00 -1.01457596e+00 1.56291604e-01 -9.94153619e-02
-9.05481517e-01 2.28347361e-01 7.37798631e-01 -3.48217855e-03
1.19133532e+00 -1.37960398e+00 -1.21299136e+00 3.97367001e-01
-1.78207934e-01 7.41375446e-01 7.15038419e-01 3.40898007e-01
-1.18405688e+00 -9.25300121e-02 -3.05298358e-01 -1.01436067e+00
-1.25798857e+00 2.30746999e-01 3.27014446e-01 5.23796268e-02
-1.31214690e+00 3.50256115e-01 -2.10961103e-01 -6.12513721e-01
-1.73940688e-01 -1.17783286e-01 -2.04066977e-01 -2.87211329e-01
-2.48540044e-01 5.71485043e-01 2.58976609e-01 -1.07493174e+00
-3.61268558e-02 1.40558720e+00 5.52643120e-01 2.64331907e-01
1.79104054e+00 -1.27588406e-01 -1.65681019e-01 9.26489532e-02
9.46477413e-01 -1.27108395e-01 -1.13679266e+00 -2.30466232e-01
5.35095707e-02 -8.37036133e-01 -1.31109599e-02 -3.00329886e-02
-1.55203927e+00 8.90593648e-01 2.75509596e-01 3.55206251e-01
6.61796749e-01 2.93534040e-01 8.02038312e-01 1.61339402e-01
8.97577405e-01 -1.00270402e+00 1.81482717e-01 2.76535511e-01
1.36624944e+00 -7.51934230e-01 4.60925817e-01 -1.05936241e+00
-1.54884934e-01 1.43614388e+00 4.79434967e-01 -5.19135833e-01
1.01740050e+00 2.20472604e-01 2.97368895e-02 -8.60416770e-01
-3.51837665e-01 -1.47880763e-01 1.07358973e-02 9.42826092e-01
-1.88981175e-01 -2.23520920e-01 -2.75208235e-01 -2.93453932e-01
-4.97004300e-01 -5.11322856e-01 5.01187682e-01 1.28350568e+00
-7.18699872e-01 -1.19409418e+00 -7.98020065e-01 2.42904484e-01
2.24586621e-01 6.36118293e-01 5.13695972e-03 1.02917850e+00
1.53112551e-02 7.71018982e-01 1.46814004e-01 -3.30093443e-01
1.08555305e+00 -2.67426103e-01 2.95598865e-01 -5.92781246e-01
-6.90945312e-02 7.82169476e-02 1.73169389e-01 -1.39339432e-01
-5.61043859e-01 -6.00122750e-01 -1.31557393e+00 -2.31567979e-01
-2.83084333e-01 2.91828841e-01 6.62205338e-01 8.66732180e-01
2.06668839e-01 1.81953520e-01 8.05617630e-01 -1.21312881e+00
-2.09573612e-01 -5.59695601e-01 -8.69970202e-01 5.71774960e-01
-2.47204438e-01 -9.75199878e-01 -1.89518467e-01 2.82697231e-01] | [8.59327507019043, -3.1007919311523438] |
db2b923d-e013-4a56-b7ef-849af8f5fd81 | a-survey-on-complex-knowledge-base-question | 2105.11644 | null | https://arxiv.org/abs/2105.11644v1 | https://arxiv.org/pdf/2105.11644v1.pdf | A Survey on Complex Knowledge Base Question Answering: Methods, Challenges and Solutions | Knowledge base question answering (KBQA) aims to answer a question over a knowledge base (KB). Recently, a large number of studies focus on semantically or syntactically complicated questions. In this paper, we elaborately summarize the typical challenges and solutions for complex KBQA. We begin with introducing the background about the KBQA task. Next, we present the two mainstream categories of methods for complex KBQA, namely semantic parsing-based (SP-based) methods and information retrieval-based (IR-based) methods. We then review the advanced methods comprehensively from the perspective of the two categories. Specifically, we explicate their solutions to the typical challenges. Finally, we conclude and discuss some promising directions for future research. | ['Ji-Rong Wen', 'Wayne Xin Zhao', 'Jing Jiang', 'Jinhao Jiang', 'Gaole He', 'Yunshi Lan'] | 2021-05-25 | null | null | null | null | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-3.16170424e-01 5.16983747e-01 -4.07648832e-02 -4.74842548e-01
-1.61299765e+00 -8.15262437e-01 5.72950915e-02 2.69751787e-01
-2.82099038e-01 1.14854240e+00 4.58591402e-01 -4.62607831e-01
-2.16989413e-01 -1.14088929e+00 -5.28633177e-01 -1.52325898e-01
4.13562864e-01 8.41950953e-01 8.42131853e-01 -8.47586632e-01
3.38829011e-01 2.11377710e-01 -1.41549802e+00 6.43766284e-01
1.18985629e+00 1.00633800e+00 1.94682285e-01 7.04794526e-01
-9.94035900e-01 1.36666501e+00 -9.33643878e-01 -9.52056289e-01
-3.63795310e-01 -3.51222277e-01 -1.98847508e+00 -6.39874458e-01
5.66768646e-01 -2.82574594e-01 -2.82542527e-01 1.08765864e+00
4.52574074e-01 3.21647793e-01 3.60357702e-01 -1.14003515e+00
-9.74109232e-01 5.04870594e-01 2.88297027e-01 5.51143348e-01
9.67707038e-01 -4.51719135e-01 1.47531867e+00 -7.74124205e-01
6.81056917e-01 1.38812673e+00 5.17710388e-01 8.26903701e-01
-2.65365630e-01 -1.61497056e-01 2.19532967e-01 1.14279389e+00
-1.30173206e+00 -2.68912762e-01 6.63052559e-01 -6.62031919e-02
1.27224994e+00 5.26154101e-01 3.28654110e-01 2.75526196e-01
-2.67213255e-01 9.80089784e-01 1.00238013e+00 -7.55922854e-01
1.87988952e-01 -2.99358182e-02 9.17951405e-01 7.00960338e-01
2.83275366e-01 -7.25638926e-01 -3.94514978e-01 -5.50113559e-01
3.96905810e-01 -7.79582977e-01 -2.42221445e-01 -1.54355511e-01
-7.51000404e-01 1.12270701e+00 1.64631039e-01 2.84000248e-01
-2.13069528e-01 1.40550643e-01 3.43995482e-01 2.91284531e-01
1.81463316e-01 8.84531796e-01 -9.32704747e-01 -1.44704819e-01
-2.50090212e-01 7.62146831e-01 1.30936694e+00 1.08263826e+00
8.70110571e-01 -3.30148786e-01 -2.24268138e-01 8.35516155e-01
4.92928058e-01 6.37585998e-01 9.70817357e-02 -1.23161256e+00
5.79406559e-01 4.89075541e-01 4.43089813e-01 -9.52287614e-01
-3.52368772e-01 2.13878393e-01 1.55497212e-02 -7.83834279e-01
3.27776223e-01 1.79194137e-02 -6.15237892e-01 1.12757373e+00
7.07473576e-01 -2.08472580e-01 6.90527320e-01 8.13894629e-01
2.03593683e+00 8.77936721e-01 4.94371504e-01 -1.62286833e-01
2.16550875e+00 -1.32313192e+00 -1.40803897e+00 -2.78021216e-01
6.84252560e-01 -6.78544402e-01 1.20381188e+00 -1.55157015e-01
-1.17003918e+00 -1.71201497e-01 -5.02412975e-01 -8.83229852e-01
-7.98450589e-01 -1.20416908e-02 6.66671634e-01 7.00448215e-01
-1.27228463e+00 -2.36781791e-01 -3.33478272e-01 -4.67269331e-01
1.13448404e-01 9.56783071e-02 1.93810381e-03 -5.37314653e-01
-1.89858866e+00 1.35552967e+00 5.28900325e-01 7.09089637e-02
-6.80552483e-01 -6.37892544e-01 -1.00841820e+00 3.72178890e-02
9.91111755e-01 -1.29842091e+00 1.91325879e+00 -1.86267599e-01
-1.59501684e+00 8.20524633e-01 -6.34705484e-01 -2.28335947e-01
-4.53014374e-01 -5.37366033e-01 -6.85337007e-01 7.13710248e-01
2.82267869e-01 4.87839818e-01 3.02162468e-01 -1.25839567e+00
-8.27965081e-01 -4.00079012e-01 1.05704355e+00 5.19155920e-01
-1.09097157e-02 6.81815565e-01 -6.90809369e-01 -3.82556856e-01
-7.07782060e-02 -3.33444953e-01 -1.43416762e-01 -4.83349204e-01
-2.24007279e-01 -8.00161839e-01 5.33897221e-01 -1.10829616e+00
1.66962504e+00 -1.42096674e+00 6.74187988e-02 -3.21559906e-01
2.02507406e-01 3.64300489e-01 -6.41148165e-02 8.58170688e-01
3.40370476e-01 2.14688197e-01 -2.14078546e-01 2.27005005e-01
4.28828821e-02 5.50784707e-01 -8.77181828e-01 -4.70217615e-01
2.84144312e-01 1.63435149e+00 -1.21266234e+00 -8.75377059e-01
-3.35887045e-01 6.88978471e-03 -3.18877518e-01 3.11532140e-01
-8.06077302e-01 1.88357487e-01 -1.15640914e+00 1.08676350e+00
3.27980340e-01 -2.87811756e-01 1.60841733e-01 -1.57667309e-01
2.67878443e-01 8.73070180e-01 -7.41292477e-01 1.55677378e+00
-3.46162796e-01 2.54829317e-01 1.38184905e-01 -9.08776164e-01
6.37941122e-01 4.97932643e-01 -2.19926789e-01 -6.17942631e-01
-2.58042574e-01 4.85649168e-01 -5.16447723e-01 -9.46841061e-01
8.41120362e-01 -4.69484597e-01 -4.34663683e-01 2.18110412e-01
-2.77970731e-02 -5.91736555e-01 2.23493010e-01 3.44636649e-01
1.23850429e+00 -6.96914643e-02 7.85315812e-01 -1.67465925e-01
1.00169528e+00 6.68638229e-01 1.70367554e-01 9.35956180e-01
-3.46920520e-01 3.21001895e-02 2.20686063e-01 -5.43448985e-01
-4.68199253e-01 -1.06515825e+00 1.36939719e-01 1.25797248e+00
4.19095725e-01 -6.46021068e-01 -8.40712190e-01 -1.08938920e+00
-1.01441391e-01 7.89844394e-01 -4.40636933e-01 7.48567358e-02
-7.76165605e-01 -5.44543207e-01 6.99529469e-01 5.59860289e-01
6.83055937e-01 -1.46015680e+00 -1.19907930e-01 8.87888446e-02
-1.02454245e+00 -1.19410217e+00 3.89038742e-01 -3.89970869e-01
-9.75811720e-01 -1.14752793e+00 -4.47322220e-01 -9.25123453e-01
3.49732816e-01 4.19936836e-01 1.65313077e+00 2.21394718e-01
7.55030811e-02 1.21061456e+00 -9.56102610e-01 -5.13366878e-01
-7.28042573e-02 1.10531069e-01 -4.66060340e-01 -7.61914253e-01
7.79487014e-01 1.28008813e-01 -4.35033083e-01 -8.12190026e-02
-1.01774561e+00 -4.60061401e-01 2.90352851e-01 3.85682285e-01
5.59384525e-01 -8.04622099e-03 1.06068099e+00 -9.75323260e-01
1.18009055e+00 -5.00154495e-01 -2.52017409e-01 9.76585805e-01
-2.16825485e-01 1.40575720e-02 2.58438438e-01 2.88488090e-01
-1.47307110e+00 -5.82353830e-01 -7.02980578e-01 3.78373235e-01
-2.08076805e-01 8.29315305e-01 -4.13137645e-01 -2.51578659e-01
6.89043105e-01 1.43064186e-01 -5.31173527e-01 -5.80588639e-01
8.34205091e-01 5.44127703e-01 3.27601820e-01 -1.02864218e+00
6.42921150e-01 4.97600049e-01 -5.09789705e-01 -7.61856794e-01
-1.66349542e+00 -8.70067060e-01 -3.89062822e-01 -1.31294832e-01
1.04668677e+00 -7.02893615e-01 -6.37451887e-01 1.82444140e-01
-1.43157542e+00 -9.76699144e-02 -5.47030687e-01 1.54688749e-02
-6.61097229e-01 6.98058367e-01 -8.12361240e-01 -7.74642527e-01
-6.10802829e-01 -7.49522805e-01 1.18167150e+00 5.30212283e-01
-7.38794580e-02 -1.21109128e+00 3.04965228e-01 1.47904599e+00
3.45714092e-01 -2.33094394e-01 1.18573427e+00 -7.88671434e-01
-5.97511888e-01 -5.79626821e-02 -2.99457639e-01 2.39951000e-01
1.19426034e-01 -6.25678480e-01 -8.56020331e-01 2.72602826e-01
1.93948701e-01 -6.57567799e-01 6.60226166e-01 9.87225994e-02
9.89590406e-01 -3.62376392e-01 -1.88888520e-01 -2.06532046e-01
1.36676681e+00 3.49381417e-01 8.45711052e-01 6.78285778e-01
5.84373474e-01 9.03865099e-01 1.12105060e+00 -1.71098739e-01
1.26852012e+00 3.77951145e-01 1.51564956e-01 4.41482544e-01
-1.07246697e-01 -1.40328005e-01 1.03416666e-01 1.16827643e+00
-1.44035786e-01 -2.69154847e-01 -1.10269320e+00 6.64781690e-01
-1.99561954e+00 -6.10025346e-01 -1.10772029e-01 1.40311575e+00
1.08862531e+00 -4.88740832e-01 -1.27961114e-01 -2.20424980e-01
5.93412995e-01 1.18626080e-01 -4.22936141e-01 -5.40440321e-01
-2.80934900e-01 5.27066827e-01 -2.09792495e-01 7.17430294e-01
-9.78780806e-01 1.59930956e+00 7.52843237e+00 8.47894549e-01
-1.10687241e-01 2.89572060e-01 1.60206303e-01 6.02732360e-01
-5.47462165e-01 2.51077354e-01 -1.17075551e+00 -2.62862116e-01
9.00780678e-01 -1.71220362e-01 3.23876560e-01 8.15015733e-01
-5.45899689e-01 -3.26456934e-01 -5.76249063e-01 7.32704103e-01
3.63267243e-01 -1.59286976e+00 7.12029517e-01 -6.57741368e-01
3.11994314e-01 -2.72471607e-01 -3.22544664e-01 8.20454240e-01
3.80590022e-01 -8.24520767e-01 2.78241843e-01 6.15559459e-01
2.88403064e-01 -5.49342155e-01 9.58638549e-01 3.46287847e-01
-1.10124183e+00 -1.13184854e-01 -6.50571167e-01 -1.77614406e-01
5.00588179e-01 4.17327523e-01 -5.24384797e-01 1.29831612e+00
1.16470480e+00 1.07369401e-01 -3.89016509e-01 8.13579381e-01
-6.10710561e-01 5.02974212e-01 -1.08373845e-02 -2.89888680e-01
2.74379700e-01 -9.75344777e-02 2.16786236e-01 1.03582108e+00
4.46924381e-02 1.01196575e+00 -1.27471492e-01 4.50146586e-01
9.64786932e-02 3.51816803e-01 -1.93539187e-01 -1.55869722e-01
4.62271512e-01 1.02243483e+00 -2.72292018e-01 -7.03606606e-01
-7.74402261e-01 7.25148737e-01 6.78015351e-01 4.60594267e-01
-3.94783586e-01 -5.45105100e-01 6.03178620e-01 -2.95946479e-01
2.85446197e-01 -6.25339523e-02 7.53676444e-02 -1.29599643e+00
1.49222896e-01 -1.08920312e+00 1.22457266e+00 -1.24653590e+00
-1.57906461e+00 4.81318235e-01 3.77150446e-01 -3.32295030e-01
-1.74972177e-01 -6.93683088e-01 -1.74164236e-01 7.21041322e-01
-2.29153156e+00 -1.00107384e+00 -1.65435791e-01 5.48581600e-01
4.08741772e-01 3.82813662e-01 1.17128730e+00 2.70659745e-01
-1.65227786e-01 -7.29699135e-02 -1.96782723e-01 2.46140406e-01
6.14610672e-01 -1.32111979e+00 2.46225402e-01 2.92260736e-01
4.97823209e-02 8.32780778e-01 4.72004443e-01 -8.80256534e-01
-1.56629765e+00 -8.84236217e-01 1.42112589e+00 -1.01947522e+00
7.83003330e-01 9.57426131e-02 -1.27491879e+00 7.97785401e-01
2.15115309e-01 -1.33831650e-01 1.06988561e+00 1.71442568e-01
-5.04241765e-01 -8.90341178e-02 -1.33961511e+00 4.58553225e-01
8.81245196e-01 -7.94759750e-01 -1.52506673e+00 6.14440382e-01
1.24803662e+00 -4.19222921e-01 -1.07354915e+00 6.34048939e-01
1.87907249e-01 -5.96269429e-01 1.11353552e+00 -1.10360181e+00
1.20007232e-01 -5.96114576e-01 -2.78155118e-01 -9.24093306e-01
-2.32058197e-01 -2.00062200e-01 -5.45223057e-01 8.99159610e-01
3.69823068e-01 -6.06257200e-01 7.09275424e-01 8.45966876e-01
-3.46097946e-01 -8.56026649e-01 -1.09699810e+00 -5.22026300e-01
3.29235464e-01 -5.76726139e-01 6.21329427e-01 9.78904843e-01
3.53480875e-01 6.84957027e-01 1.57527596e-01 4.72180843e-01
5.73835187e-02 3.53797466e-01 5.54217339e-01 -9.62230563e-01
-1.08673843e-02 -7.95636252e-02 -1.44670352e-01 -1.38146615e+00
2.46860370e-01 -5.28322101e-01 -1.38741182e-02 -2.31751657e+00
2.49121562e-01 -2.71476090e-01 3.36106634e-03 5.08621216e-01
-6.25590682e-01 2.48209629e-02 -9.63754430e-02 1.09897636e-01
-1.18285120e+00 4.52144057e-01 1.52039742e+00 -1.01395929e-02
1.82589531e-01 -2.87249565e-01 -1.18544877e+00 6.26207709e-01
8.90037119e-01 -1.31186664e-01 -6.25720799e-01 -6.75573528e-01
9.62669790e-01 2.04485580e-01 5.52222654e-02 -3.01855683e-01
3.77116054e-01 -4.97854292e-01 -4.27864164e-01 -7.14108884e-01
4.81860816e-01 -6.10502601e-01 -5.56993663e-01 8.41994733e-02
-1.10851750e-01 1.20889857e-01 2.75639862e-01 5.18895686e-01
-7.89133728e-01 -7.87737370e-01 2.30253071e-01 -5.13041139e-01
-1.30811989e+00 1.70052052e-04 -4.24132288e-01 6.48835659e-01
9.49750543e-01 1.65156543e-01 -7.20950365e-01 -6.98165715e-01
-7.42217481e-01 7.88133025e-01 -2.49780133e-01 3.55774343e-01
8.07697892e-01 -1.10724604e+00 -5.30580163e-01 -7.67887771e-01
5.38126767e-01 3.03780854e-01 5.26052237e-01 2.47616097e-01
-8.21521759e-01 1.21603739e+00 2.40333796e-01 1.12588421e-01
-1.11751199e+00 6.93454325e-01 3.38499784e-01 -4.76870865e-01
-1.84697568e-01 9.58456695e-01 2.55957037e-01 -8.51551890e-01
1.23585083e-01 -3.08340460e-01 -9.29197252e-01 -1.52872890e-01
7.67282009e-01 4.68338311e-01 2.57419497e-01 -4.83569801e-01
-5.75213373e-01 5.50585628e-01 -6.70838580e-02 -4.69582248e-03
9.99061465e-01 -3.15958500e-01 -8.47533822e-01 3.59164923e-01
6.98943436e-01 -1.48878455e-01 -1.71344988e-02 -5.54037094e-01
5.73110223e-01 -4.83440831e-02 -2.05982551e-01 -1.14603508e+00
-5.45004129e-01 6.78704858e-01 -1.34451449e-01 4.33552265e-01
1.16969895e+00 8.54068458e-01 1.24648726e+00 1.17669106e+00
6.33049130e-01 -1.18041360e+00 8.51744786e-02 9.11575019e-01
1.05230272e+00 -9.15365458e-01 -1.44536614e-01 -9.57889438e-01
-5.95860720e-01 1.01383162e+00 7.41812408e-01 2.99103200e-01
7.19257891e-01 -2.49141440e-01 3.78448635e-01 -7.16796577e-01
-7.09597290e-01 -5.93076587e-01 3.81548762e-01 7.89549768e-01
1.61806747e-01 -2.10360125e-01 -3.52835953e-01 1.00070989e+00
-4.95937258e-01 -1.13033734e-01 2.96493739e-01 1.40729129e+00
-9.46981668e-01 -1.19660044e+00 -5.84406376e-01 2.81101137e-01
-5.17713606e-01 -4.30656791e-01 -8.98802996e-01 7.93743014e-01
-3.64821762e-01 1.42206824e+00 -1.46985546e-01 1.57537639e-01
6.24989092e-01 5.19154608e-01 6.79540694e-01 -1.09961164e+00
-4.22621906e-01 -7.44187951e-01 7.56051600e-01 -3.82904798e-01
-7.80266404e-01 -2.21771762e-01 -1.43597639e+00 1.73355956e-02
-5.36909461e-01 8.19097698e-01 3.67231190e-01 1.18613601e+00
4.49354321e-01 1.92121044e-02 -1.66486770e-01 2.92298675e-01
-4.31193322e-01 -7.30047047e-01 -2.94773549e-01 -8.10575262e-02
1.29528016e-01 -4.50734615e-01 -2.05364451e-01 1.27731031e-02] | [10.637533187866211, 7.952310085296631] |
fdcb3eba-3663-49a9-8ffa-bd9af61cc1ad | semi-supervised-medical-image-segmentation | 1911.01218 | null | https://arxiv.org/abs/1911.01218v1 | https://arxiv.org/pdf/1911.01218v1.pdf | Semi-Supervised Medical Image Segmentation via Learning Consistency under Transformations | The scarcity of labeled data often limits the application of supervised deep learning techniques for medical image segmentation. This has motivated the development of semi-supervised techniques that learn from a mixture of labeled and unlabeled images. In this paper, we propose a novel semi-supervised method that, in addition to supervised learning on labeled training images, learns to predict segmentations consistent under a given class of transformations on both labeled and unlabeled images. More specifically, in this work we explore learning equivariance to elastic deformations. We implement this through: 1) a Siamese architecture with two identical branches, each of which receives a differently transformed image, and 2) a composite loss function with a supervised segmentation loss term and an unsupervised term that encourages segmentation consistency between the predictions of the two branches. We evaluate the method on a public dataset of chest radiographs with segmentations of anatomical structures using 5-fold cross-validation. The proposed method reaches significantly higher segmentation accuracy compared to supervised learning. This is due to learning transformation consistency on both labeled and unlabeled images, with the latter contributing the most. We achieve the performance comparable to state-of-the-art chest X-ray segmentation methods while using substantially fewer labeled images. | ['Gerda Bortsova', 'Florian Dubost', 'Marleen de Bruijne', 'Laurens Hogeweg', 'Ioannis Katramados'] | 2019-11-04 | null | null | null | null | ['semi-supervised-medical-image-segmentation'] | ['computer-vision'] | [ 4.55595851e-01 6.47335291e-01 -4.11681950e-01 -1.03127444e+00
-1.25164986e+00 -4.51531023e-01 2.60586053e-01 1.74257115e-01
-7.25507081e-01 6.55364037e-01 5.22105359e-02 -1.86028674e-01
1.50733665e-01 -5.13395429e-01 -8.02274287e-01 -6.89787507e-01
3.11892390e-01 9.18135703e-01 2.09386215e-01 1.86496258e-01
-1.61876619e-01 1.81727946e-01 -8.49849105e-01 3.00644606e-01
9.67389345e-01 9.90842819e-01 2.12972201e-02 5.08194804e-01
-1.07735179e-01 9.10302997e-01 5.36799692e-02 -4.07324284e-01
3.72122765e-01 -6.89958036e-01 -1.35712588e+00 6.00173175e-01
7.13481545e-01 -5.06237522e-02 2.68183146e-02 9.80709195e-01
3.99384677e-01 2.52222344e-02 8.64129901e-01 -7.05374599e-01
-4.54003364e-01 7.05467582e-01 -6.40607476e-01 1.54478520e-01
-1.81491122e-01 3.75848403e-03 1.05911493e+00 -6.58057034e-01
8.00925434e-01 7.37750411e-01 8.02073181e-01 6.48018003e-01
-1.55858994e+00 -2.44646132e-01 -1.04764856e-01 -3.39552134e-01
-1.04006803e+00 -7.30959848e-02 7.70630896e-01 -6.43756509e-01
4.81991112e-01 -1.11321025e-02 3.99059564e-01 8.06923807e-01
1.49289608e-01 8.75955820e-01 1.33318675e+00 -4.35382575e-01
3.90780598e-01 2.01666221e-01 3.17158192e-01 1.00039876e+00
4.71008867e-02 6.11745752e-02 -2.28182189e-02 1.24994647e-02
5.31571567e-01 -7.29350001e-02 -7.81751499e-02 -7.90206671e-01
-1.17442429e+00 8.66720378e-01 6.74060464e-01 3.46619785e-01
-4.70896065e-01 -1.70549855e-03 3.35214794e-01 -4.69482541e-02
6.63936853e-01 4.09000844e-01 -3.51150990e-01 4.21567351e-01
-1.45169985e+00 -1.19981080e-01 6.52497590e-01 5.30445874e-01
7.61853933e-01 -1.15775198e-01 -9.82692167e-02 9.07623649e-01
4.64812428e-01 2.38811240e-01 8.67779136e-01 -6.88372016e-01
3.25177163e-01 6.30931497e-01 -3.94248158e-01 -4.04997498e-01
-6.62918806e-01 -4.85298187e-01 -8.17110598e-01 2.82407314e-01
6.83573425e-01 -1.01188086e-01 -1.30240321e+00 1.76368535e+00
3.79072279e-01 5.17963879e-02 2.32583210e-02 8.80143225e-01
7.70483911e-01 2.40420789e-01 1.82529703e-01 -1.37891620e-01
1.00632048e+00 -1.26573384e+00 -5.47130227e-01 -2.08646044e-01
7.58320987e-01 -7.38260865e-01 1.07003975e+00 3.21572304e-01
-1.28957379e+00 -5.86162269e-01 -9.74911869e-01 1.61717869e-02
-2.96333153e-02 2.58485109e-01 3.60220790e-01 7.90575564e-01
-1.02465785e+00 7.98641086e-01 -1.39108098e+00 -2.00269535e-01
6.58457398e-01 6.86692834e-01 -2.66106039e-01 1.29221171e-01
-7.75915027e-01 8.10192406e-01 4.33683187e-01 -2.01846391e-01
-4.90171403e-01 -7.20103502e-01 -8.53287995e-01 -2.60031164e-01
2.56720871e-01 -5.42433023e-01 1.35801470e+00 -1.30315149e+00
-1.48520911e+00 1.54554105e+00 2.52104431e-01 -7.00033545e-01
9.08046603e-01 -3.68429460e-02 9.70485434e-03 3.94271225e-01
3.15659195e-01 8.75412881e-01 9.25252676e-01 -1.37153316e+00
-4.39013660e-01 -4.54610139e-01 -4.52662736e-01 2.65936330e-02
-4.70060930e-02 -2.62708217e-01 -4.96746331e-01 -9.22234595e-01
3.61586571e-01 -1.30803883e+00 -4.06884462e-01 8.26761201e-02
-7.10170031e-01 8.77354201e-03 6.64175689e-01 -6.74844801e-01
7.30922818e-01 -1.95134330e+00 2.60670662e-01 5.02032697e-01
2.89274365e-01 1.67965472e-01 7.70078525e-02 -2.39810809e-01
-3.38412791e-01 7.96254352e-02 -9.23038363e-01 -6.52555108e-01
-2.95623094e-01 3.99049342e-01 -7.00426772e-02 6.92104161e-01
2.67913431e-01 1.03410983e+00 -7.19062209e-01 -8.54865849e-01
2.19702348e-01 2.82367408e-01 -6.05470061e-01 2.69018114e-01
-2.10159555e-01 1.10012412e+00 -3.76914293e-01 5.08559525e-01
4.92030531e-01 -6.61155760e-01 2.90945053e-01 -1.81043684e-01
3.14383984e-01 -2.49321423e-02 -9.18451309e-01 1.95433760e+00
-3.90777290e-01 1.90513074e-01 -5.61881214e-02 -1.32154512e+00
7.14410603e-01 3.77523422e-01 9.83512342e-01 -4.35268015e-01
2.65085191e-01 4.90017891e-01 -7.82294944e-02 -4.99980539e-01
-1.54877126e-01 -5.22682667e-01 1.00016050e-01 7.73728073e-01
4.92330164e-01 -4.23538029e-01 2.24486113e-01 1.56290844e-01
8.08431625e-01 2.36593053e-01 1.19791739e-01 -3.56782109e-01
5.89341164e-01 -1.95333324e-02 4.02010113e-01 6.26652420e-01
-1.86183915e-01 9.44226086e-01 1.92217156e-01 -5.46184659e-01
-9.47448432e-01 -1.15829992e+00 -4.31420505e-01 8.38917434e-01
8.20341930e-02 2.20059812e-01 -1.02651787e+00 -1.30610085e+00
-1.03934489e-01 4.84547406e-01 -7.49159515e-01 -6.37224913e-02
-8.56072724e-01 -8.63762140e-01 5.17893374e-01 7.54599452e-01
4.50129718e-01 -1.05258667e+00 -6.26496911e-01 5.76037243e-02
-1.95240632e-01 -1.09523714e+00 -5.85754335e-01 4.69903618e-01
-1.22177446e+00 -1.11331129e+00 -8.66035461e-01 -1.08796847e+00
1.16280353e+00 -4.33041632e-01 1.18774843e+00 7.39855319e-02
-3.11506093e-01 4.37455297e-01 -1.83606476e-01 -9.31191146e-02
-7.40129888e-01 3.28463376e-01 -1.59905344e-01 2.24102244e-01
-1.48040846e-01 -2.72013515e-01 -5.71765006e-01 3.50384265e-01
-1.06073308e+00 8.41255337e-02 6.54458463e-01 1.06425560e+00
1.19710886e+00 -3.18687052e-01 3.21998864e-01 -1.54089808e+00
8.02580416e-02 -3.23981911e-01 -4.60945457e-01 3.80706638e-01
-8.62428427e-01 3.18523407e-01 6.83512330e-01 -2.85375088e-01
-1.23444641e+00 7.64526546e-01 -2.19582513e-01 -1.96111336e-01
-3.09869915e-01 5.27756691e-01 3.11305225e-01 -2.68408895e-01
7.63916016e-01 -1.24658309e-01 2.23976135e-01 -3.85236531e-01
4.19044465e-01 2.82156169e-01 7.54762053e-01 -4.74240720e-01
7.20116675e-01 7.75194287e-01 6.41604587e-02 -5.23386359e-01
-1.17942309e+00 -5.31797051e-01 -1.00187230e+00 -1.33870602e-01
1.36142015e+00 -4.36396807e-01 7.22888336e-02 3.93980205e-01
-6.88466489e-01 -6.75467134e-01 -7.63090730e-01 8.96272242e-01
-8.30871046e-01 5.22461653e-01 -7.78935075e-01 -1.36597261e-01
-3.78247887e-01 -1.47930467e+00 1.11944425e+00 1.59050912e-01
-2.50175118e-01 -1.27915001e+00 4.82747823e-01 7.21919775e-01
1.83063522e-01 4.35415417e-01 8.10020447e-01 -1.11840618e+00
-2.14675456e-01 -2.58054137e-01 -2.64637154e-02 7.61648417e-01
3.35839629e-01 -1.10467002e-01 -9.13376510e-01 -4.05142128e-01
2.32918799e-01 -7.60002077e-01 1.02458537e+00 7.37324774e-01
1.30083954e+00 -1.69650000e-02 -1.81422293e-01 8.41443896e-01
1.23884380e+00 -7.42788613e-02 2.75673479e-01 8.35367665e-02
7.96422839e-01 6.09911859e-01 3.97175163e-01 -4.33333591e-02
8.73692259e-02 4.19362664e-01 2.07618982e-01 -7.69339740e-01
-3.03392708e-01 -6.26607388e-02 -2.02035740e-01 9.21788871e-01
-2.36946400e-02 -2.78213266e-02 -1.16968155e+00 3.17180187e-01
-1.69296217e+00 -4.90291148e-01 -3.34615111e-02 2.17468405e+00
1.09387410e+00 2.21669331e-01 1.79337561e-01 1.01183869e-01
6.52784050e-01 2.02702545e-02 -7.18833804e-01 1.45847714e-02
1.49173886e-01 7.62623608e-01 6.19081080e-01 5.18246770e-01
-1.61473238e+00 6.67392492e-01 6.21212912e+00 5.65293252e-01
-1.48547721e+00 2.05800757e-01 1.08538353e+00 2.49250159e-01
5.84137850e-02 -2.20558003e-01 -3.42769504e-01 3.09029102e-01
7.32271016e-01 3.70899349e-01 2.62119481e-03 8.60665560e-01
-5.23203351e-02 -7.82773085e-03 -1.16823912e+00 5.15502036e-01
1.37293562e-01 -1.37553346e+00 -9.39812958e-02 -1.23398848e-01
1.12794375e+00 1.02289483e-01 2.50278801e-01 -1.48935439e-02
2.30951861e-01 -1.00835836e+00 6.15955353e-01 3.79666954e-01
7.98009992e-01 -3.87911707e-01 7.37055779e-01 2.34135374e-01
-8.80145669e-01 2.55131364e-01 1.17986113e-01 5.72416425e-01
1.52732775e-01 2.66083002e-01 -9.04570758e-01 5.34337342e-01
4.91223067e-01 6.72571838e-01 -5.95054030e-01 8.13188612e-01
-2.85262376e-01 9.91117537e-01 -3.41223329e-01 5.49304187e-01
4.89104241e-01 -2.98913270e-01 8.14102814e-02 1.07721865e+00
-1.61660299e-01 -3.63645563e-03 5.32131493e-01 8.95426035e-01
-2.84649342e-01 4.19605583e-01 -1.59366548e-01 1.87052786e-01
-1.57380864e-01 1.20834541e+00 -1.26158404e+00 -4.54035431e-01
-4.15560156e-01 9.10697639e-01 2.63317943e-01 1.70630619e-01
-7.98236847e-01 1.64573967e-01 -5.12842298e-01 1.77109748e-01
1.50932655e-01 1.13006748e-01 -5.95936835e-01 -1.11980820e+00
-1.30510554e-01 -5.68586588e-01 7.18521893e-01 -3.62906337e-01
-1.50360668e+00 6.64397061e-01 1.06426768e-01 -1.12223256e+00
-2.63352185e-01 -4.87819880e-01 -6.31470382e-01 6.29655898e-01
-1.49452925e+00 -1.30780184e+00 -2.29481667e-01 4.55182463e-01
5.51735818e-01 -1.09711871e-01 6.97456360e-01 3.21777910e-01
-4.52703983e-01 5.87172210e-01 9.14150253e-02 3.78169060e-01
8.44590485e-01 -1.65851104e+00 2.24305570e-01 7.22063184e-01
4.06592578e-01 3.11622560e-01 3.67029011e-01 -6.16675735e-01
-7.52450168e-01 -1.21254504e+00 6.32621109e-01 -2.07252920e-01
3.70044202e-01 1.44917101e-01 -1.04332626e+00 8.32141578e-01
7.34907016e-02 5.52968442e-01 1.03236973e+00 -3.30532968e-01
-1.99746609e-01 9.38355178e-02 -1.37114680e+00 1.82229176e-01
4.79278237e-01 -3.10282081e-01 -7.56565869e-01 7.00813532e-01
3.05837959e-01 -6.35227442e-01 -9.12596881e-01 6.10881150e-01
3.62823904e-01 -8.79873514e-01 9.25937831e-01 -6.38550162e-01
4.95180249e-01 -4.46250513e-02 1.85401782e-01 -1.08606315e+00
5.32916859e-02 -3.11490655e-01 3.06274891e-01 6.96283698e-01
6.14585876e-01 -4.46711361e-01 1.14866197e+00 8.18817854e-01
-1.63944855e-01 -1.02972794e+00 -8.88382137e-01 -6.24760091e-01
4.67443347e-01 -1.36418641e-01 -1.19024925e-01 1.08895648e+00
-2.62436450e-01 2.77730942e-01 -1.14451654e-01 -1.69792488e-01
8.14342618e-01 2.27712646e-01 3.67610872e-01 -1.10756445e+00
-5.06528318e-01 -4.16420758e-01 -2.47662708e-01 -6.98218346e-01
3.71403694e-01 -1.41581368e+00 1.49859831e-01 -1.38979876e+00
3.15627933e-01 -5.19009411e-01 -2.19183058e-01 7.42936850e-01
-1.35133177e-01 6.28375411e-01 -1.47024363e-01 4.73752677e-01
-5.86694658e-01 2.55651295e-01 1.29901636e+00 -2.25897729e-01
-2.30776817e-01 3.29315186e-01 -1.94367602e-01 1.09055293e+00
9.13904250e-01 -7.39277005e-01 -4.22938883e-01 -3.16661358e-01
-1.92459166e-01 3.22000235e-02 2.28336543e-01 -9.36963677e-01
1.17428608e-01 1.38175458e-01 3.32314223e-01 -4.20754999e-01
-5.70107475e-02 -8.10288727e-01 -1.39702633e-01 8.26342523e-01
-8.56996298e-01 -2.20206037e-01 -5.81870303e-02 5.26012480e-01
-3.10770482e-01 -3.75320762e-01 1.35572350e+00 -2.64478415e-01
-3.64265829e-01 3.37160170e-01 -2.00580761e-01 3.26363891e-01
9.87131357e-01 -1.47665694e-01 1.93288222e-01 -1.39691964e-01
-1.24581122e+00 1.16614848e-01 4.13863361e-01 -6.30538389e-02
4.67946500e-01 -1.08348894e+00 -6.95395887e-01 1.49823189e-01
-3.86543013e-02 2.66548127e-01 8.01268443e-02 1.07299578e+00
-7.92739987e-01 1.25016153e-01 -2.60292888e-01 -9.58983600e-01
-1.14220810e+00 3.75195771e-01 5.53299367e-01 -8.03281426e-01
-5.78080714e-01 9.02521908e-01 1.39868811e-01 -8.55874002e-01
1.85535356e-01 -4.63438630e-01 -2.40026917e-02 -1.77501991e-01
-6.80297688e-02 1.41482741e-01 3.04981023e-01 -9.30811048e-01
-2.62785882e-01 7.88965225e-01 -2.80767113e-01 -1.02059387e-01
1.05740702e+00 1.70831069e-01 3.21785361e-03 3.59571517e-01
1.44581401e+00 -1.78841591e-01 -1.17654669e+00 -4.39396292e-01
1.87192351e-01 -1.48075551e-01 9.96285528e-02 -7.85380125e-01
-1.55741048e+00 7.84004033e-01 9.07161772e-01 -6.85636178e-02
1.00415599e+00 2.35660493e-01 8.70958209e-01 2.24544659e-01
5.96626885e-02 -1.13141096e+00 3.51157099e-01 4.58755977e-02
4.42583025e-01 -1.67837036e+00 1.23354718e-01 -4.98118430e-01
-8.64531100e-01 9.84172463e-01 4.18214530e-01 -4.31424707e-01
7.31017411e-01 2.09689006e-01 4.27056432e-01 -3.47732812e-01
-1.34571612e-01 -1.30044177e-01 7.64376998e-01 3.37949783e-01
6.45510614e-01 -8.43825936e-03 -4.87350345e-01 4.03801233e-01
4.33098413e-02 1.62224267e-02 5.34018725e-02 1.04603767e+00
-9.16076750e-02 -1.28946865e+00 -2.15911224e-01 6.75685525e-01
-8.20246935e-01 2.43725494e-01 -3.54539424e-01 7.77386963e-01
5.25555126e-02 6.01992667e-01 -7.21446648e-02 4.02198033e-03
1.88217968e-01 4.66720723e-02 4.32873160e-01 -8.82050097e-01
-9.02621865e-01 3.17663014e-01 -3.39339912e-01 -4.95421141e-01
-7.64025927e-01 -6.11647964e-01 -1.60949469e+00 4.49031502e-01
-4.06647950e-01 -1.41029786e-02 4.56751794e-01 1.00800467e+00
-1.42941400e-01 4.88170177e-01 6.75662458e-01 -8.27241540e-01
-8.97985101e-01 -6.60360992e-01 -5.52574813e-01 9.12734151e-01
9.51110348e-02 -5.10845006e-01 -1.95949987e-01 5.09530723e-01] | [14.648211479187012, -2.2642641067504883] |
d739795e-a90d-49bd-b51c-c1802cb98337 | multi-view-representation-learning-from | 2210.15429 | null | https://arxiv.org/abs/2210.15429v1 | https://arxiv.org/pdf/2210.15429v1.pdf | Multi-view Representation Learning from Malware to Defend Against Adversarial Variants | Deep learning-based adversarial malware detectors have yielded promising results in detecting never-before-seen malware executables without relying on expensive dynamic behavior analysis and sandbox. Despite their abilities, these detectors have been shown to be vulnerable to adversarial malware variants - meticulously modified, functionality-preserving versions of original malware executables generated by machine learning. Due to the nature of these adversarial modifications, these adversarial methods often use a \textit{single view} of malware executables (i.e., the binary/hexadecimal view) to generate adversarial malware variants. This provides an opportunity for the defenders (i.e., malware detectors) to detect the adversarial variants by utilizing more than one view of a malware file (e.g., source code view in addition to the binary view). The rationale behind this idea is that while the adversary focuses on the binary view, certain characteristics of the malware file in the source code view remain untouched which leads to the detection of the adversarial malware variants. To capitalize on this opportunity, we propose Adversarially Robust Multiview Malware Defense (ARMD), a novel multi-view learning framework to improve the robustness of DL-based malware detectors against adversarial variants. Our experiments on three renowned open-source deep learning-based malware detectors across six common malware categories show that ARMD is able to improve the adversarial robustness by up to seven times on these malware detectors. | ['Hsinchun Chen', 'Xin Li', 'Weifeng Li', 'MohammadReza Ebrahimi', 'James Lee Hu'] | 2022-10-25 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [ 1.31282717e-01 -4.41000789e-01 -2.33737320e-01 1.61766022e-01
-2.86337078e-01 -1.49443948e+00 7.93965101e-01 -5.25802635e-02
1.41655132e-01 2.68349677e-01 -2.85181016e-01 -8.80755544e-01
5.00656009e-01 -9.35785353e-01 -8.50854635e-01 -7.30176032e-01
-3.38036090e-01 2.64028192e-01 4.15188164e-01 -4.03673142e-01
1.89388767e-01 7.30077505e-01 -1.11872828e+00 6.30521238e-01
3.54851961e-01 5.91394186e-01 -7.18904138e-02 1.03303134e+00
-1.23079069e-01 7.24187970e-01 -8.44949901e-01 -4.66029525e-01
6.03178799e-01 -3.63673875e-03 -3.75913709e-01 -2.60559529e-01
4.10214484e-01 -8.17262948e-01 -7.51968920e-01 1.35315633e+00
6.43392801e-02 -4.52958882e-01 5.29419601e-01 -1.51731026e+00
-6.33911312e-01 5.17402947e-01 -5.81323028e-01 3.98190230e-01
3.92639697e-01 7.14848638e-01 7.57160485e-01 -6.29281580e-01
6.59248710e-01 1.46134174e+00 6.34158015e-01 7.69570589e-01
-1.11478913e+00 -8.18441987e-01 1.95168331e-01 4.89118174e-02
-1.02879000e+00 -2.68099099e-01 7.89404154e-01 -7.23850906e-01
1.00424385e+00 4.06666040e-01 2.61646688e-01 1.79602981e+00
7.80133247e-01 5.55727899e-01 1.08947992e+00 1.49629906e-01
1.75060460e-03 2.34455153e-01 1.52467981e-01 8.07494223e-01
5.58440983e-01 6.19289875e-01 1.60627425e-01 -6.39002204e-01
1.00776657e-01 5.28090239e-01 -1.46515623e-01 -4.16901290e-01
-6.33252323e-01 1.21560514e+00 5.18340051e-01 1.85778499e-01
-3.26801762e-02 2.11970694e-03 1.15767562e+00 4.89350289e-01
2.00258285e-01 5.55880010e-01 -6.02740109e-01 1.84939906e-01
-5.04692614e-01 7.73963630e-02 7.45434046e-01 5.03048897e-01
5.94602883e-01 6.37122273e-01 -1.09811254e-01 1.58396587e-01
2.13441357e-01 9.36794519e-01 3.66073906e-01 -1.05884753e-01
4.33136195e-01 6.46595061e-01 -4.99287307e-01 -1.08796060e+00
-5.10357954e-02 -2.03599647e-01 -6.43246591e-01 5.80759227e-01
1.54824555e-01 -1.01490512e-01 -9.08312619e-01 1.65174019e+00
2.21038029e-01 4.79125604e-02 8.28476250e-02 1.95180804e-01
5.79375744e-01 5.26379764e-01 -1.92539930e-01 5.96658178e-02
1.20807660e+00 -6.84530377e-01 -2.35709965e-01 -2.10044578e-01
5.64917147e-01 -4.00584161e-01 1.06305242e+00 5.14051691e-02
-3.29977244e-01 -2.62339056e-01 -1.38639402e+00 8.05064261e-01
-8.92405570e-01 -5.42238772e-01 2.61394918e-01 1.09143674e+00
-7.77855992e-01 4.87458736e-01 -1.00250912e+00 9.11241546e-02
5.65902710e-01 3.31735492e-01 -5.24309039e-01 8.59070849e-03
-1.04492831e+00 5.77122509e-01 4.63420182e-01 -6.23961866e-01
-1.73550785e+00 -5.73273540e-01 -6.86027348e-01 1.33572042e-01
5.28181136e-01 -1.75971687e-01 9.66558635e-01 -1.09227443e+00
-9.30236995e-01 7.07215071e-01 1.44642070e-01 -6.94533587e-01
6.82964444e-01 6.71032742e-02 -5.94681442e-01 2.73694634e-01
-1.23959489e-01 -1.19156793e-01 1.77189159e+00 -1.44707716e+00
-3.23182166e-01 -5.89772761e-01 5.25946856e-01 -5.20832360e-01
-5.91035485e-01 1.40822858e-01 1.09512761e-01 -8.40120673e-01
-6.23953402e-01 -1.23483443e+00 2.64906913e-01 -4.62269008e-01
-8.18460345e-01 2.72980750e-01 1.99650502e+00 -6.98943436e-01
9.83252823e-01 -2.21248603e+00 -2.88624819e-02 1.36962488e-01
6.63631737e-01 9.08410251e-01 -1.89184934e-01 3.40343118e-01
-3.83023381e-01 6.08682513e-01 -5.28423786e-02 1.25849649e-01
-2.23303080e-01 8.06154311e-02 -1.00544047e+00 6.81571543e-01
2.20412597e-01 1.03991556e+00 -1.00090468e+00 6.00177981e-02
3.32351357e-01 2.67227769e-01 -4.91776794e-01 3.36728752e-01
-4.78567600e-01 5.26669919e-01 -4.47005272e-01 1.06527567e+00
7.86890507e-01 3.62335965e-02 1.59188896e-01 1.27180412e-01
2.95292228e-01 4.08614129e-02 -4.93189365e-01 4.27603483e-01
-4.46586221e-01 6.36188269e-01 2.33763248e-01 -7.66116440e-01
4.97705996e-01 2.19306797e-01 2.05530375e-01 -1.96911708e-01
1.71188071e-01 1.02856167e-01 2.42320195e-01 -2.87653834e-01
1.34511217e-01 2.17638388e-01 -7.14774579e-02 6.55993104e-01
-2.00825125e-01 2.77126104e-01 -1.15370564e-01 3.70387226e-01
1.64634025e+00 -4.23324168e-01 5.17560780e-01 1.50524497e-01
8.48234117e-01 -5.53729907e-02 4.34645236e-01 9.10978794e-01
-4.40585405e-01 -6.23756945e-02 7.45757103e-01 -5.27529001e-01
-1.13271558e+00 -1.49681783e+00 1.54401660e-01 1.29360521e+00
-1.64945960e-01 -4.32401776e-01 -6.82955503e-01 -1.65383875e+00
2.02616408e-01 4.57099378e-01 -6.43739343e-01 -6.37118876e-01
-8.85238171e-01 -4.09161508e-01 1.18313265e+00 5.29086709e-01
3.69135350e-01 -7.18261838e-01 -5.71350634e-01 5.24909701e-03
3.62299681e-01 -1.03491998e+00 -4.23810065e-01 2.63219953e-01
-5.67229271e-01 -1.44272518e+00 -5.77417063e-03 -2.98599601e-01
4.20689672e-01 3.62175584e-01 8.82193923e-01 1.87225208e-01
-2.85773486e-01 4.33220327e-01 -5.07027745e-01 -4.10709023e-01
-1.44854712e+00 -1.07075702e-02 4.38730448e-01 -1.66211814e-01
1.78889006e-01 -4.37613845e-01 1.26084257e-02 2.12883785e-01
-1.15857649e+00 -6.64217651e-01 4.12030131e-01 8.59736264e-01
2.91926205e-01 1.72234595e-01 4.08858478e-01 -9.92461860e-01
4.84249324e-01 -8.56193781e-01 -8.56245875e-01 1.87206864e-01
-4.17840749e-01 -6.17377572e-02 1.66921175e+00 -1.00258291e+00
-5.57837903e-01 -4.26087491e-02 -1.48815453e-01 -1.10184157e+00
-8.42833370e-02 7.77425170e-02 -4.97508615e-01 -3.22701603e-01
9.10501599e-01 5.05905509e-01 9.52198207e-02 -7.79784173e-02
2.79310197e-01 5.29215872e-01 5.13857424e-01 -1.28875807e-01
1.55316734e+00 5.23815334e-01 1.18568875e-01 -6.46399975e-01
-2.40440503e-01 -1.89530700e-01 -4.31562096e-01 -1.35202080e-01
7.64117002e-01 -6.67125702e-01 -7.57069707e-01 7.62988329e-01
-1.14475274e+00 3.46743315e-02 2.41513729e-01 -1.67065009e-01
-2.23192379e-01 6.39274597e-01 -6.91915751e-01 -3.99780869e-01
-4.49039608e-01 -1.75306427e+00 8.92247498e-01 -1.28478110e-01
-6.92357123e-02 -1.05329645e+00 4.61939126e-02 3.37786563e-02
4.93381232e-01 6.99116647e-01 1.35020697e+00 -1.50268233e+00
-5.35441756e-01 -6.76456571e-01 1.23872779e-01 5.84014058e-01
3.81659210e-01 3.42443407e-01 -1.10065746e+00 -1.04051709e+00
4.02107984e-01 -1.74292475e-01 7.89973497e-01 -2.92750269e-01
1.07023585e+00 -6.96652055e-01 -5.95941305e-01 9.28111970e-01
1.41431129e+00 4.83725727e-01 2.54729897e-01 2.22167417e-01
1.12462342e+00 2.02031463e-01 5.33985868e-02 4.16412443e-01
-2.80786514e-01 5.31543493e-01 1.15840614e+00 2.45419517e-01
1.56376317e-01 -2.99970359e-01 9.85282123e-01 2.86541432e-01
4.45299059e-01 -3.84181947e-01 -8.80407512e-01 7.65030831e-02
-1.13469648e+00 -1.37543440e+00 8.76891911e-02 2.13522172e+00
3.07111055e-01 3.98876697e-01 1.08339444e-01 7.38237947e-02
8.27526629e-01 7.37210989e-01 -1.00683701e+00 -9.66155827e-01
-7.15682209e-02 1.33467540e-01 6.65186465e-01 2.73539156e-01
-1.55719805e+00 9.50816214e-01 5.59781933e+00 8.80941093e-01
-1.48607683e+00 3.19928855e-01 2.26986647e-01 9.35397819e-02
-1.02302067e-01 -1.67275786e-01 -8.11741233e-01 7.01463938e-01
1.00224340e+00 -1.69775173e-01 7.99784064e-01 1.40280306e+00
-1.74170107e-01 4.69876319e-01 -1.14264750e+00 5.47727644e-01
1.37509599e-01 -1.43387461e+00 3.13658863e-01 4.78793532e-01
6.22157216e-01 2.91371375e-01 5.47521710e-01 5.62204838e-01
5.27575493e-01 -1.19170225e+00 4.30653423e-01 -1.98109418e-01
9.45464194e-01 -8.65125358e-01 7.60949850e-01 4.41242367e-01
-1.51166558e+00 -5.12212396e-01 -8.88682902e-02 2.53592461e-01
-2.40349755e-01 1.44897804e-01 -1.08132577e+00 4.10209000e-01
5.35936058e-01 4.65763927e-01 -7.84319818e-01 3.95469852e-02
-2.07855865e-01 5.73517680e-01 1.88937306e-01 3.21073234e-01
4.26290423e-01 3.32485348e-01 1.06470227e+00 1.13065600e+00
-1.11225106e-01 -5.12936890e-01 3.96747530e-01 9.51704144e-01
-1.81955963e-01 -3.91149729e-01 -1.55388284e+00 -6.51263297e-01
5.57713330e-01 1.22246087e+00 -5.69633901e-01 -2.50135064e-01
-4.14560914e-01 8.99298489e-01 2.11108804e-01 1.60757639e-02
-1.11136544e+00 -2.43292972e-01 1.14450693e+00 3.50811295e-02
6.15302682e-01 -2.10285276e-01 6.04756884e-02 -1.10844493e+00
-1.50003776e-01 -1.63134646e+00 3.00130755e-01 2.97050606e-02
-1.46428025e+00 7.56236732e-01 -1.12672560e-01 -1.28011429e+00
-6.21103466e-01 -9.09358561e-01 -1.18750751e+00 7.17789710e-01
-8.37637365e-01 -1.35080040e+00 1.47963865e-02 7.33441591e-01
5.24667621e-01 -9.54927266e-01 7.13449419e-01 -1.66141748e-01
-5.34154117e-01 9.41814005e-01 4.03573155e-01 4.31289226e-01
3.66863787e-01 -1.06551802e+00 9.40053523e-01 1.16620851e+00
-9.77913961e-02 7.11804390e-01 6.27209842e-01 -1.05067515e+00
-1.92238963e+00 -1.53846371e+00 -7.36539736e-02 -7.77015865e-01
1.02415168e+00 -6.03117466e-01 -1.05181921e+00 9.22618866e-01
-1.23297505e-01 7.91669041e-02 5.73488057e-01 -5.68521023e-01
-1.17423368e+00 2.03408554e-01 -1.44746220e+00 6.43717647e-01
6.83421552e-01 -1.01709878e+00 -4.28996444e-01 4.06548142e-01
9.69357491e-01 -2.65459925e-01 -4.47362155e-01 6.20411336e-01
4.05422062e-01 -1.18004704e+00 1.35273635e+00 -8.78219366e-01
3.50388110e-01 -2.57550687e-01 -4.46102202e-01 -1.24667549e+00
-4.67193797e-02 -5.21744370e-01 -8.52235496e-01 1.02550256e+00
-1.76441655e-01 -9.92459238e-01 3.10435295e-01 -3.73892426e-01
-7.03258291e-02 -8.28808784e-01 -8.24548841e-01 -1.17956972e+00
2.44084045e-01 -2.44639426e-01 6.92443550e-01 1.02643561e+00
-5.90189338e-01 5.61816245e-02 -2.52828091e-01 5.08083642e-01
4.48159665e-01 -6.98170625e-03 9.44800198e-01 -1.00920618e+00
-6.00827515e-01 -5.09325325e-01 -5.58671355e-01 -2.97933936e-01
7.22139120e-01 -1.06588995e+00 -3.78830135e-01 -4.77427244e-01
1.64208353e-01 -1.51614668e-02 -4.68442217e-02 3.78070503e-01
-2.33875617e-01 1.79705828e-01 3.74058604e-01 2.39923388e-01
-1.96229555e-02 9.47846547e-02 7.30328321e-01 -6.91932261e-01
2.74815951e-02 1.42442688e-01 -4.50355560e-01 9.42610443e-01
9.34953153e-01 -5.74467897e-01 -3.80777299e-01 -7.67425820e-02
-2.45501295e-01 -4.12971020e-01 4.96506095e-01 -7.75491416e-01
-3.85059685e-01 -6.87188432e-02 3.17037076e-01 -6.39026821e-01
2.23385960e-01 -6.43727958e-01 -3.15407217e-02 1.18166018e+00
2.00494692e-01 6.19449615e-01 3.55473846e-01 8.04221690e-01
1.75804183e-01 -8.69365484e-02 1.07539439e+00 -4.14011359e-01
-7.46553540e-01 5.08054614e-01 -6.98397517e-01 1.36355028e-01
1.45068252e+00 -1.34893328e-01 -8.42185378e-01 -6.50217310e-02
-2.23924547e-01 -1.49098650e-01 1.03819704e+00 5.80531538e-01
5.16372025e-01 -9.50414717e-01 -5.23754060e-01 4.31018710e-01
2.03104168e-01 -8.15491617e-01 1.18623376e-01 4.40901786e-01
-4.07964796e-01 3.90305579e-01 -4.60284323e-01 -6.59676492e-01
-1.84078062e+00 1.41530335e+00 3.71353924e-01 -6.62286758e-01
-4.20688540e-01 5.95481813e-01 4.22553778e-01 -4.89503235e-01
4.23627980e-02 2.56662905e-01 -5.01097692e-03 -1.25679359e-01
6.99572861e-01 4.46410626e-01 5.79294227e-02 -8.84459257e-01
-6.67506516e-01 1.48729831e-01 -5.53808451e-01 5.20222545e-01
9.09004569e-01 5.44189334e-01 -3.33910823e-01 2.13914528e-01
1.55337226e+00 4.98162001e-01 -1.02374494e+00 5.71281724e-02
-1.43604994e-01 -6.33605957e-01 -4.34300929e-01 -6.28687143e-01
-1.00341880e+00 8.40223134e-01 7.00232744e-01 6.10397220e-01
8.68763864e-01 -1.47576332e-01 9.26968873e-01 5.62517822e-01
3.42441529e-01 -3.09635609e-01 5.06712675e-01 9.52997863e-01
5.97333610e-01 -1.22860074e+00 -2.07195461e-01 -1.10658921e-01
-4.19914216e-01 1.15009177e+00 7.89136589e-01 -2.18796685e-01
2.89912730e-01 6.35345161e-01 -8.06243867e-02 -7.99002126e-02
-6.16751075e-01 3.33466053e-01 5.07855825e-02 9.95977044e-01
-1.00387886e-01 4.31341290e-01 3.87748599e-01 2.66890049e-01
-2.98704170e-02 -1.02999413e+00 5.81582546e-01 1.01842940e+00
-3.63406301e-01 -1.13354695e+00 -8.32336843e-01 6.03516221e-01
-6.61718726e-01 -1.85166851e-01 -1.06266236e+00 7.21427858e-01
1.19816244e-01 9.00047719e-01 -1.70428589e-01 -1.07197905e+00
-1.00614280e-01 1.41046837e-01 1.41102403e-01 -6.70626760e-01
-1.03793466e+00 -4.87148046e-01 -2.84758747e-01 -5.53318679e-01
4.12491262e-01 -3.30457628e-01 -1.07243431e+00 -8.65583777e-01
-6.06237948e-02 -4.97823715e-01 4.72618073e-01 8.71729672e-01
4.27403897e-01 5.75385273e-01 1.30473614e+00 -9.00137067e-01
-1.18096077e+00 -6.16737545e-01 -6.34452924e-02 5.89034677e-01
7.21619666e-01 -6.51518047e-01 -6.94380105e-01 -2.46384472e-01] | [14.385711669921875, 9.658913612365723] |
5ab3ea30-1855-445d-aedc-cd673b166082 | global-and-local-consistent-age-generative | 1801.08390 | null | http://arxiv.org/abs/1801.08390v1 | http://arxiv.org/pdf/1801.08390v1.pdf | Global and Local Consistent Age Generative Adversarial Networks | Age progression/regression is a challenging task due to the complicated and
non-linear transformation in human aging process. Many researches have shown
that both global and local facial features are essential for face
representation, but previous GAN based methods mainly focused on the global
feature in age synthesis. To utilize both global and local facial information,
we propose a Global and Local Consistent Age Generative Adversarial Network
(GLCA-GAN). In our generator, a global network learns the whole facial
structure and simulates the aging trend of the whole face, while three crucial
facial patches are progressed or regressed by three local networks aiming at
imitating subtle changes of crucial facial subregions. To preserve most of the
details in age-attribute-irrelevant areas, our generator learns the residual
face. Moreover, we employ an identity preserving loss to better preserve the
identity information, as well as age preserving loss to enhance the accuracy of
age synthesis. A pixel loss is also adopted to preserve detailed facial
information of the input face. Our proposed method is evaluated on three face
aging datasets, i.e., CACD dataset, Morph dataset and FG-NET dataset.
Experimental results show appealing performance of the proposed method by
comparing with the state-of-the-art. | ['Pei-Pei Li', 'Zhenan Sun', 'Yibo Hu', 'Qi Li', 'Ran He'] | 2018-01-25 | null | null | null | null | ['human-aging'] | ['miscellaneous'] | [-2.39090435e-02 -6.71008974e-02 7.84779787e-02 -3.65419835e-01
-2.20853955e-01 6.13784194e-02 2.49376893e-01 -5.23497522e-01
-1.79582328e-01 8.98471594e-01 2.95035481e-01 5.21734953e-01
3.08614314e-01 -9.40573573e-01 -5.30352235e-01 -1.10647976e+00
2.47638166e-01 -7.78766423e-02 -1.55610397e-01 -1.69232577e-01
-1.59283161e-01 3.57607126e-01 -1.54049802e+00 -1.89258128e-01
1.11258185e+00 1.37155044e+00 -3.81095380e-01 -3.81065486e-03
7.47617409e-02 7.31670976e-01 -5.70805967e-01 -5.57059646e-01
2.71524221e-01 -7.07015812e-01 -2.23226190e-01 4.93042842e-02
6.72895253e-01 -6.91176176e-01 -5.89706838e-01 1.14124238e+00
9.17509019e-01 -1.52852505e-01 7.57405400e-01 -1.50771940e+00
-8.21689010e-01 1.83084011e-01 -1.01843572e+00 -6.19970486e-02
8.37542117e-02 2.64476031e-01 3.77957791e-01 -7.05669701e-01
3.31810713e-01 1.52697778e+00 7.01768875e-01 9.13443744e-01
-7.65487731e-01 -1.17293215e+00 2.59729803e-01 3.23814541e-01
-1.43707502e+00 -4.50779408e-01 1.19980347e+00 -1.47313014e-01
-2.74670869e-02 7.41179883e-02 7.28466690e-01 1.12498116e+00
5.82794070e-01 5.31058908e-01 1.40783620e+00 -7.12671950e-02
-4.24975343e-02 -2.99301326e-01 -3.88794065e-01 1.00927711e+00
1.13856062e-01 1.46266952e-01 -4.59108740e-01 1.23733073e-01
1.02370739e+00 1.85924992e-01 -3.26595485e-01 -1.16264418e-01
-7.17493713e-01 3.87365371e-01 5.95466554e-01 2.70412695e-02
-5.90675831e-01 2.25684315e-01 3.77808571e-01 3.41454387e-01
7.68943727e-01 -3.57238561e-01 -2.48713285e-01 2.85113007e-01
-8.42114627e-01 2.64104962e-01 7.64110461e-02 6.84068084e-01
9.66388285e-01 4.15731966e-01 -3.37600142e-01 7.42070019e-01
2.77725726e-01 5.68131030e-01 5.45391262e-01 -8.40165675e-01
2.55080163e-01 7.27156341e-01 -3.13925862e-01 -9.52245176e-01
-1.31129801e-01 -3.77773911e-01 -1.28008616e+00 5.76241672e-01
3.25510591e-01 -1.04150437e-01 -1.20685256e+00 2.16396475e+00
5.45788229e-01 2.36112699e-01 -2.37276509e-01 7.01443851e-01
8.22082877e-01 4.97311354e-01 4.49284375e-01 -6.25090063e-01
1.31279266e+00 -8.15704346e-01 -8.67781043e-01 -3.08404595e-01
2.08089817e-02 -6.21956050e-01 8.98369908e-01 9.45182294e-02
-1.30419934e+00 -8.14149618e-01 -1.03349710e+00 -1.21782772e-01
-1.04794860e-01 2.72721440e-01 4.77389634e-01 5.67751646e-01
-9.25566018e-01 3.69797289e-01 -7.55652368e-01 -1.09219737e-01
9.73456740e-01 3.87661785e-01 -4.71920252e-01 -2.76538193e-01
-1.14793360e+00 6.27434254e-01 -6.91207796e-02 3.81673843e-01
-1.00447214e+00 -1.07096410e+00 -8.85074377e-01 -9.36103910e-02
-8.87305737e-02 -1.12696874e+00 7.00125039e-01 -1.34541285e+00
-1.27234137e+00 7.97444224e-01 -7.22728968e-02 -1.03656913e-03
7.14917898e-01 -1.95662603e-01 -4.83174086e-01 1.25162318e-01
5.88649465e-03 7.36131907e-01 1.26196897e+00 -1.06272435e+00
-2.03888565e-01 -9.32619631e-01 -2.86128730e-01 2.05883905e-01
-6.41190469e-01 6.94520026e-02 -2.38991678e-01 -1.10156369e+00
-8.34453478e-02 -5.98222375e-01 4.29216437e-02 6.64352953e-01
2.79900245e-02 5.18071726e-02 1.07688475e+00 -1.51609731e+00
1.43269575e+00 -2.07984781e+00 3.17717671e-01 6.80644065e-02
3.28302234e-01 -2.44080671e-03 -2.17132658e-01 -5.68075068e-02
-2.10135475e-01 -1.05304562e-01 -4.19469386e-01 -5.35931468e-01
-3.66665274e-01 1.17171519e-01 8.58265460e-02 6.10632539e-01
1.49951085e-01 9.28075969e-01 -5.83765984e-01 -7.92832971e-01
-9.85549092e-02 8.97134781e-01 -3.31734776e-01 2.82539904e-01
-2.04729680e-02 7.05598056e-01 -5.14513433e-01 9.47614789e-01
1.16480923e+00 4.52407181e-01 -3.43337864e-01 -3.77942473e-01
2.77696252e-01 -4.55605477e-01 -8.47709835e-01 1.73600614e+00
-4.59613949e-01 -2.26168483e-01 2.22067848e-01 -6.90837920e-01
1.11116540e+00 2.49896392e-01 4.50516969e-01 -7.91557908e-01
4.13462460e-01 2.25826070e-01 -1.60680637e-01 -2.23951980e-01
-1.96552590e-01 -2.13579774e-01 2.07857817e-01 3.68854344e-01
6.73617944e-02 2.54044056e-01 -2.64866054e-01 -6.73708469e-02
7.66149580e-01 2.39428446e-01 -8.54878407e-03 -2.01253325e-01
9.46357369e-01 -8.90959442e-01 9.59076941e-01 -3.14366341e-01
-4.55275267e-01 8.12057436e-01 5.00615120e-01 -5.44562936e-01
-1.05207276e+00 -9.79853332e-01 2.15416342e-01 6.62177801e-01
8.00990686e-02 -2.06405729e-01 -1.16374385e+00 -9.05153692e-01
-1.85437575e-01 1.05984524e-01 -1.15002632e+00 -7.19726443e-01
-6.73309624e-01 -6.68675184e-01 3.51319671e-01 7.13029146e-01
1.16005361e+00 -1.33285284e+00 5.13200723e-02 -1.10544385e-02
-1.21115558e-02 -6.52140677e-01 -9.85620558e-01 -7.34086871e-01
-1.01067626e+00 -9.28978443e-01 -1.18184662e+00 -8.39278162e-01
1.18767548e+00 -1.80927724e-01 8.23165715e-01 3.30333501e-01
-4.05650020e-01 1.25555128e-01 3.75763699e-02 -3.04108202e-01
-2.67635375e-01 -5.23125008e-03 1.34805590e-01 5.18556654e-01
-7.23989755e-02 -1.13979387e+00 -1.16387641e+00 3.11041147e-01
-6.75524890e-01 -1.98601425e-01 7.65375972e-01 7.94423401e-01
5.79830110e-01 2.81666130e-01 7.29218245e-01 -6.70074165e-01
3.37406248e-01 -2.48917103e-01 2.88322964e-03 2.31893986e-01
-7.78604448e-01 -1.03075653e-01 5.78918636e-01 -5.22939086e-01
-1.42703724e+00 -2.91202236e-02 -1.88452244e-01 -7.04224408e-01
1.46819904e-01 3.32710557e-02 -9.52967405e-01 -2.32615486e-01
3.69293511e-01 4.29047942e-01 5.24631143e-01 -3.40895474e-01
2.04468116e-01 1.75914526e-01 5.74364543e-01 -7.78976023e-01
1.07574117e+00 3.69313061e-01 2.45101035e-01 -3.27007830e-01
-4.84317809e-01 3.26406568e-01 -3.50289106e-01 -4.63378161e-01
7.90905654e-01 -1.11522603e+00 -4.71034676e-01 1.32741225e+00
-7.90478051e-01 -1.35628417e-01 -3.30490112e-01 6.97357068e-03
-5.10249019e-01 3.24800879e-01 -6.54905081e-01 -5.62909603e-01
-8.88461709e-01 -8.57080698e-01 1.05007827e+00 6.89310670e-01
9.15489048e-02 -7.57843077e-01 -1.13192335e-01 2.32686087e-01
4.54930902e-01 7.27766991e-01 9.34075832e-01 3.50876302e-01
-2.53493994e-01 -1.24632187e-01 -1.40396491e-01 7.16268957e-01
5.60260475e-01 4.83906418e-02 -1.01509762e+00 -4.22054142e-01
3.92846316e-02 -3.18553239e-01 9.15794492e-01 3.01497579e-01
1.39081156e+00 -5.20560086e-01 -1.29906252e-01 6.29069805e-01
1.21122932e+00 6.66253418e-02 1.18863130e+00 -1.34859964e-01
9.03629363e-01 6.46506011e-01 6.29161000e-01 3.31064939e-01
2.99588025e-01 2.37952068e-01 3.87149394e-01 -2.63752759e-01
-5.53926349e-01 -5.22416711e-01 4.36969250e-01 7.64100969e-01
-5.37688076e-01 2.72325009e-01 -3.34317446e-01 3.17648590e-01
-1.56115305e+00 -8.77513349e-01 3.74661177e-01 2.05943322e+00
9.91261780e-01 -1.98417306e-02 1.47289231e-01 2.94788867e-01
8.71728420e-01 3.28542799e-01 -7.56467223e-01 3.05209141e-02
-2.78986692e-01 6.43329561e-01 1.49242595e-01 3.86679396e-02
-8.21963966e-01 6.75690889e-01 4.99892902e+00 1.03846967e+00
-1.09214830e+00 1.28188074e-01 1.21351790e+00 2.07737442e-02
-2.98794508e-01 -3.00174683e-01 -4.03355092e-01 7.15400159e-01
3.40493143e-01 -2.47181699e-01 4.34997737e-01 7.92305410e-01
1.40542999e-01 2.81200320e-01 -5.70409358e-01 9.99672413e-01
2.40157098e-01 -6.36305034e-01 2.38838345e-01 9.19385105e-02
8.04533064e-01 -1.00878584e+00 5.21142900e-01 2.41703495e-01
-2.27301255e-01 -1.02115488e+00 5.38913727e-01 9.49309945e-01
1.24615693e+00 -1.18035269e+00 6.24981284e-01 -8.76454040e-02
-1.46011913e+00 -1.78805634e-01 -1.78932369e-01 8.92898589e-02
-1.49646461e-01 5.08269370e-01 2.39194691e-01 6.03363931e-01
8.19562376e-01 7.82508492e-01 -9.39383268e-01 6.77621067e-01
-4.67187643e-01 3.20332050e-01 -2.25445232e-03 5.89973688e-01
-2.88342535e-01 -3.67730528e-01 1.33224949e-01 3.35145026e-01
5.85548043e-01 2.39955902e-01 -4.39472228e-01 7.57595778e-01
-4.15717006e-01 2.05488846e-01 -3.06198746e-01 2.11713910e-01
3.64626557e-01 1.37726235e+00 -2.81353921e-01 -4.39423844e-02
-2.29821518e-01 1.08303523e+00 1.68110341e-01 2.62797952e-01
-7.69732296e-01 -2.28654474e-01 8.92760277e-01 3.98304671e-01
5.52175902e-02 -1.23040099e-02 -1.22006334e-01 -8.99872422e-01
3.02069724e-01 -1.03859425e+00 3.25758904e-01 -7.83880293e-01
-1.35080433e+00 4.55235600e-01 -3.47677737e-01 -1.10814321e+00
1.50017768e-01 -1.22477509e-01 -1.22675323e+00 1.12115717e+00
-1.22214007e+00 -1.87713563e+00 -8.28789115e-01 7.85842121e-01
2.90421695e-01 -2.38988906e-01 5.10246158e-01 5.66282451e-01
-8.10960472e-01 1.03620756e+00 -4.96440202e-01 3.78776602e-02
1.15147412e+00 -8.20758939e-01 1.24260396e-01 8.59671652e-01
-6.74178898e-01 5.80766916e-01 3.41653973e-01 -8.22170019e-01
-1.07800043e+00 -1.35110807e+00 5.32708943e-01 -6.02270514e-02
1.72889620e-01 -1.65550455e-01 -9.51045632e-01 6.26733780e-01
2.34181777e-01 2.76689410e-01 1.61107525e-01 -3.88018608e-01
-4.22925681e-01 -6.81915045e-01 -1.48653817e+00 5.61074018e-01
1.25866699e+00 -3.10025692e-01 -5.71463332e-02 -2.76202828e-01
6.86503172e-01 -2.44436190e-01 -1.04252458e+00 7.37214625e-01
7.69553602e-01 -7.81827033e-01 9.15967405e-01 -1.24357231e-01
6.73907459e-01 -4.50102717e-01 2.92946368e-01 -1.24097466e+00
-2.99240857e-01 -4.91642892e-01 -3.39412183e-01 1.87133312e+00
-2.59646744e-01 -6.25977576e-01 9.63654280e-01 3.63126308e-01
-1.14983052e-01 -8.86240184e-01 -8.67907643e-01 -4.58610982e-01
2.42798969e-01 4.43992227e-01 1.02898824e+00 7.23499119e-01
-8.11109483e-01 9.71338525e-02 -5.28484404e-01 3.18063609e-03
8.84543180e-01 -3.32626641e-01 6.66349351e-01 -1.02903342e+00
3.11947972e-01 -4.26787317e-01 -6.12599730e-01 -3.33220065e-01
3.87943029e-01 -3.17664623e-01 -3.15460324e-01 -1.17926443e+00
1.83728516e-01 -3.66646886e-01 -4.85622883e-01 4.65421677e-01
-5.92768371e-01 5.01177013e-01 -2.31710762e-01 -7.73220807e-02
5.24419360e-02 1.17533672e+00 1.84577477e+00 -3.75041932e-01
5.15389256e-02 -3.27201188e-02 -8.04841459e-01 4.24280196e-01
8.36070418e-01 -1.62759140e-01 -5.73294997e-01 -1.29451081e-01
-5.13248667e-02 -2.05055356e-01 3.81006747e-01 -1.32176602e+00
7.33314529e-02 -1.07022054e-01 9.81206477e-01 -2.74081707e-01
4.33922946e-01 -6.53421998e-01 3.31243545e-01 5.38166106e-01
1.83501542e-01 4.68304195e-02 1.38298064e-01 3.73570681e-01
-3.19713265e-01 3.28045040e-01 1.12840903e+00 -1.26962960e-02
-5.03097177e-01 1.14777505e+00 3.72184157e-01 -1.32496953e-01
1.12888753e+00 -1.00970030e-01 -3.16394567e-01 -3.05920541e-01
-6.74904168e-01 1.82156414e-01 6.66416049e-01 4.78826970e-01
7.17324674e-01 -1.79089260e+00 -9.46660519e-01 5.03115416e-01
-2.08835620e-02 -1.06913177e-02 7.84894705e-01 7.64837980e-01
-4.72774059e-01 -4.06871319e-01 -9.03082371e-01 -1.37152329e-01
-1.37386012e+00 7.27104545e-01 4.36512619e-01 -9.07785520e-02
-2.72193164e-01 7.42145538e-01 6.49267018e-01 -1.29294008e-01
8.50105807e-02 1.33063808e-01 -2.68380284e-01 5.28224669e-02
6.43276989e-01 4.34005857e-01 -2.50951231e-01 -6.63672626e-01
-2.23324418e-01 9.98480201e-01 -1.82858169e-01 2.57968009e-01
1.16096675e+00 -2.15187848e-01 -5.27574301e-01 -8.83282349e-02
1.00622404e+00 4.10796143e-03 -1.51801968e+00 -1.72746971e-01
-9.47234988e-01 -3.59626293e-01 1.91780436e-03 -5.66313028e-01
-1.88882363e+00 9.41487312e-01 1.06120837e+00 -5.64043283e-01
1.82432044e+00 -2.70364195e-01 1.03046441e+00 -6.64692521e-01
2.83638686e-01 -7.98715949e-01 3.80657881e-01 -1.40192106e-01
1.13474333e+00 -8.72140169e-01 2.43278652e-01 -4.68652219e-01
-3.87313634e-01 9.03787494e-01 1.20480442e+00 -7.69280568e-02
6.94670737e-01 -1.18942395e-01 3.93349715e-02 1.12826079e-01
-2.88312405e-01 2.24663213e-01 1.88470155e-01 7.07994401e-01
1.40578330e-01 -2.54265994e-01 -5.81809938e-01 8.70670915e-01
-1.55195162e-01 1.41500190e-01 6.00763783e-02 7.41735518e-01
-1.30695149e-01 -1.17305791e+00 -3.01177979e-01 4.54901516e-01
-4.01992977e-01 9.13664550e-02 -1.53405681e-01 6.94440603e-01
6.94839180e-01 3.72045666e-01 1.70479808e-03 -4.10464197e-01
1.75658897e-01 -3.29465047e-03 8.13178718e-01 -1.26248628e-01
-3.75890195e-01 -1.17483258e-01 -3.12221348e-01 -5.80881059e-01
-3.90271097e-01 -7.05696762e-01 -1.00043011e+00 -6.27417803e-01
6.19106591e-02 -1.58706576e-01 4.30751473e-01 4.65452999e-01
1.67274296e-01 6.85339034e-01 9.88826394e-01 -5.77302039e-01
-2.42755130e-01 -1.13313675e+00 -7.86301672e-01 3.42591584e-01
2.55600631e-01 -7.60265172e-01 -4.10968393e-01 3.14433903e-01] | [13.16911506652832, 0.4411361813545227] |
d65c289f-6f64-4454-9f5e-4283870ae385 | a-deep-behavior-path-matching-network-for | 2302.00302 | null | https://arxiv.org/abs/2302.00302v1 | https://arxiv.org/pdf/2302.00302v1.pdf | A Deep Behavior Path Matching Network for Click-Through Rate Prediction | User behaviors on an e-commerce app not only contain different kinds of feedback on items but also sometimes imply the cognitive clue of the user's decision-making. For understanding the psychological procedure behind user decisions, we present the behavior path and propose to match the user's current behavior path with historical behavior paths to predict user behaviors on the app. Further, we design a deep neural network for behavior path matching and solve three difficulties in modeling behavior paths: sparsity, noise interference, and accurate matching of behavior paths. In particular, we leverage contrastive learning to augment user behavior paths, provide behavior path self-activation to alleviate the effect of noise, and adopt a two-level matching mechanism to identify the most appropriate candidate. Our model shows excellent performance on two real-world datasets, outperforming the state-of-the-art CTR model. Moreover, our model has been deployed on the Meituan food delivery platform and has accumulated 1.6% improvement in CTR and 1.8% improvement in advertising revenue. | ['Dong Wang', 'Xingxing Wang', 'Yongkang Wang', 'Beihong Jin', 'Shuli Wang', 'Yimin Lv', 'Yapeng Zhang', 'Yisong Yu', 'Jian Dong'] | 2023-02-01 | null | null | null | null | ['click-through-rate-prediction'] | ['miscellaneous'] | [-1.29225478e-01 -4.08573359e-01 -8.60957980e-01 -7.50938296e-01
-1.89181328e-01 -2.88283467e-01 -2.67248660e-01 3.10405433e-01
-3.34651142e-01 -1.75895289e-01 4.24472153e-01 -6.25750244e-01
-1.36693537e-01 -8.83748114e-01 -6.44469321e-01 4.31009382e-02
-2.04680189e-01 1.71488211e-01 -1.15336098e-01 -4.33012515e-01
1.03904814e-01 -2.73721278e-01 -1.24154782e+00 5.78045547e-01
8.59084785e-01 1.49544406e+00 1.16044983e-01 1.42371789e-01
-1.76735315e-02 7.84656167e-01 -8.57788399e-02 -8.16245496e-01
3.40133220e-01 -3.63912523e-01 -3.44838589e-01 -3.33648883e-02
1.73414856e-01 -8.45195532e-01 -5.39132833e-01 1.10331523e+00
1.13924250e-01 2.44885504e-01 2.69014537e-01 -9.40079629e-01
-1.51583230e+00 1.21438992e+00 -6.95804894e-01 3.19513112e-01
5.63905776e-01 5.04008634e-03 1.24445355e+00 -4.61800724e-01
2.62653142e-01 1.11455429e+00 8.03688943e-01 3.93666685e-01
-1.18106902e+00 -9.04721200e-01 6.63676322e-01 3.83532971e-01
-8.74420762e-01 -8.06252956e-02 8.75001252e-01 -3.06319505e-01
4.88571167e-01 -4.56967130e-02 7.97623336e-01 1.56911266e+00
5.55159450e-01 1.30211806e+00 8.90226245e-01 4.24571596e-02
2.24895790e-01 3.25698316e-01 9.17741358e-01 5.20423949e-01
9.62320417e-02 4.01604027e-02 -2.38088682e-01 -3.46629679e-01
7.30982244e-01 5.75309992e-01 2.72144049e-01 9.89443250e-03
-5.98942995e-01 1.02537203e+00 6.57021642e-01 1.36419177e-01
-6.43598437e-01 -1.85663223e-01 9.16260928e-02 5.90969503e-01
2.16436312e-01 3.96643192e-01 -8.11358392e-01 2.03818455e-02
-2.82156765e-01 4.47764605e-01 8.99388075e-01 9.24203455e-01
3.83634686e-01 1.15229236e-02 -5.01929402e-01 9.49227273e-01
4.24704194e-01 3.55670273e-01 8.09768736e-01 -8.27989817e-01
1.93178862e-01 7.19074190e-01 1.52799830e-01 -1.27143848e+00
-4.41286981e-01 -6.36793733e-01 -5.35204530e-01 -6.93321109e-01
4.68497872e-01 -1.49637595e-01 -7.23947883e-01 1.88666320e+00
-8.08572322e-02 4.56024632e-02 -3.88916314e-01 7.81255603e-01
6.47347569e-01 4.30895805e-01 5.34224629e-01 -2.80306548e-01
1.48911607e+00 -7.42967904e-01 -6.43677294e-01 -3.52973044e-01
4.74012971e-01 -3.64318907e-01 1.51892304e+00 6.78155780e-01
-1.14998639e+00 -7.68864989e-01 -9.16058600e-01 1.19979344e-01
-1.53970256e-01 2.89414853e-01 1.26533139e+00 6.57080293e-01
-4.49017555e-01 5.56805849e-01 -6.74399972e-01 -1.37195334e-01
4.76036668e-01 5.38393557e-01 2.50619948e-01 -1.46194458e-01
-1.36239707e+00 4.12738442e-01 -1.21185727e-01 -7.53856599e-02
-5.54580152e-01 -9.66162026e-01 -5.88434935e-01 4.37581837e-01
5.32430410e-01 -3.27595055e-01 1.75554705e+00 -1.02882123e+00
-1.50977194e+00 3.16773206e-01 1.21579890e-03 -4.80013847e-01
8.91507417e-02 -3.52581531e-01 -1.09807730e+00 -6.21466398e-01
6.31764308e-02 2.64378339e-01 4.80541438e-01 -6.85100675e-01
-5.82894504e-01 -6.14083946e-01 7.93124512e-02 -5.54928146e-02
-7.77092636e-01 -4.42097057e-03 -6.01777434e-01 -4.49163288e-01
1.80571735e-01 -7.92180359e-01 -5.54970682e-01 -4.09615457e-01
3.65558192e-02 -3.94462705e-01 4.05529857e-01 -8.11722159e-01
1.71826410e+00 -2.16853690e+00 -6.26530588e-01 2.50859648e-01
2.12927014e-01 5.38649745e-02 -3.29552531e-01 -5.78357726e-02
2.28119299e-01 -2.41083708e-02 5.41221082e-01 2.41303220e-02
3.08423955e-02 -2.13219840e-02 -2.66991973e-01 6.63732737e-02
-3.17014217e-01 1.23752475e+00 -8.17752421e-01 2.08434477e-01
-2.33119354e-01 7.94601627e-03 -1.06069613e+00 1.38385549e-01
-5.70005834e-01 1.68504894e-01 -8.21043670e-01 7.84237862e-01
7.25777686e-01 -6.69269264e-01 5.12957871e-01 -3.04147214e-01
8.53154734e-02 6.29324198e-01 -7.30463326e-01 1.49086642e+00
-2.83186316e-01 -2.11182490e-01 -1.44226581e-01 -7.54158378e-01
9.65936542e-01 -2.99294710e-01 9.56656218e-01 -1.59771323e+00
4.77783263e-01 -1.99450523e-01 4.18513805e-01 -5.46475172e-01
4.95720029e-01 3.90565664e-01 -4.20749128e-01 8.10489237e-01
-4.00673300e-02 1.15572679e+00 4.99082990e-02 -5.94400801e-03
1.03639925e+00 -1.24957547e-01 4.28105563e-01 -2.24280760e-01
3.89862359e-01 -1.61094934e-01 8.27579200e-01 8.56599391e-01
-3.73783469e-01 -2.47492582e-01 4.47363228e-01 -6.48997843e-01
-5.84352195e-01 -8.59032869e-01 5.75779676e-02 1.73599565e+00
2.93522328e-01 -3.05335820e-01 -3.84055614e-01 -9.63541567e-01
2.86148816e-01 8.67678702e-01 -8.16660523e-01 -5.41072607e-01
-6.61161005e-01 -1.01855290e+00 4.92918119e-02 8.24893236e-01
6.67069793e-01 -1.02368104e+00 2.67270319e-02 5.64997792e-01
-1.45047873e-01 -9.43440557e-01 -8.54444087e-01 -9.60031077e-02
-8.88703108e-01 -8.43938231e-01 -1.91388026e-01 -6.09995484e-01
2.04490557e-01 6.45474792e-01 1.22813344e+00 -1.04224592e-01
1.74956068e-01 1.27332792e-01 -2.59587765e-01 -4.44148362e-01
-1.42083898e-01 1.59387887e-01 2.65943497e-01 7.52909034e-02
1.36463165e+00 -6.02088809e-01 -1.02397895e+00 7.42978454e-01
-3.71927202e-01 -5.76067388e-01 6.87719047e-01 4.64997292e-01
4.00542915e-01 1.68221012e-01 6.36648297e-01 -1.07540572e+00
1.03430188e+00 -1.23239541e+00 -5.51655710e-01 4.16182354e-02
-1.28807437e+00 -2.96131015e-01 6.32931054e-01 -1.03150558e+00
-1.21457016e+00 -5.81764942e-03 -2.90435612e-01 -1.27046287e-01
2.22364198e-02 7.03001320e-01 2.43792348e-02 2.96064794e-01
6.35925591e-01 9.77834873e-03 -8.84795338e-02 -8.27645957e-01
3.73495311e-01 6.34518981e-01 3.74023110e-01 -3.95245701e-01
1.03726164e-01 1.23813577e-01 -7.59339273e-01 -1.93195507e-01
-1.29018188e+00 -5.75543106e-01 -6.74266070e-02 -6.68230504e-02
5.33044577e-01 -8.85946155e-01 -1.39712822e+00 2.94902980e-01
-6.50141060e-01 -1.56390354e-01 1.40875444e-01 4.74874765e-01
-1.72555864e-01 4.94960099e-02 -1.05917954e+00 -5.99002302e-01
-3.83140117e-01 -8.53291512e-01 3.65356237e-01 3.12526911e-01
-4.15364891e-01 -7.79024959e-01 -9.57080200e-02 5.72555363e-01
5.27496278e-01 -4.89684314e-01 1.05768323e+00 -8.52267504e-01
-3.59061867e-01 -1.87028080e-01 -2.45889246e-01 8.85767639e-02
1.12417147e-01 -5.44021845e-01 -5.48992753e-01 -1.07126184e-01
3.55757862e-01 -5.74736781e-02 6.09759688e-01 7.23120332e-01
1.45877016e+00 -5.24832428e-01 -3.09009552e-01 4.21094388e-01
1.08767641e+00 7.54048228e-01 4.96727735e-01 7.39688575e-02
4.67916131e-01 4.53800857e-01 5.77557683e-01 4.40638036e-01
8.72568011e-01 6.73887253e-01 2.14648187e-01 7.21779242e-02
3.14967155e-01 -7.83638000e-01 7.03306913e-01 7.06430674e-01
2.01133817e-01 -2.06343740e-01 -2.75169820e-01 1.87061988e-02
-2.09689236e+00 -8.72020125e-01 -2.81409889e-01 2.05331206e+00
4.90998238e-01 2.71261454e-01 6.64066195e-01 -4.11072582e-01
2.98375398e-01 -1.99301571e-01 -1.05527449e+00 -4.51793969e-01
3.18792999e-01 -1.19045146e-01 5.81497788e-01 1.48361370e-01
-9.70253050e-01 9.26576614e-01 7.27452517e+00 2.84424901e-01
-1.09842217e+00 3.03891361e-01 9.13396478e-01 -1.15686014e-01
-2.52827436e-01 -5.49694300e-01 -9.39512968e-01 7.07346857e-01
1.05148232e+00 -1.42681837e-01 6.61984086e-01 1.19885254e+00
3.40840757e-01 2.25132316e-01 -1.13736749e+00 8.64494205e-01
8.21493659e-03 -9.80359614e-01 -9.21589434e-02 3.70564759e-01
5.12738407e-01 -4.59033996e-02 4.30600315e-01 1.01007247e+00
7.35767961e-01 -6.00050271e-01 4.83530343e-01 3.65215778e-01
2.19630092e-01 -4.30810392e-01 5.33214331e-01 3.21650177e-01
-9.57191885e-01 -9.11387444e-01 -4.10278797e-01 -3.82912129e-01
5.36379740e-02 3.50299180e-01 -4.30809170e-01 1.71806235e-02
9.06651735e-01 9.93169487e-01 -5.82179785e-01 6.29107714e-01
2.64370710e-01 9.51340139e-01 1.82291046e-02 -3.22495103e-01
1.86440766e-01 -5.45067549e-01 -1.21870331e-01 7.77262270e-01
4.23383772e-01 2.30358347e-01 4.56893742e-01 1.11019814e+00
-2.27561250e-01 4.93381053e-01 -2.56803274e-01 -2.48954445e-01
2.39703521e-01 1.17267704e+00 -2.83350855e-01 -9.08506364e-02
-8.71305346e-01 7.45395243e-01 3.22917759e-01 3.16213399e-01
-8.70868802e-01 3.96821707e-01 9.77083087e-01 3.97279054e-01
3.28603208e-01 -1.13901813e-02 -3.68576348e-01 -1.06021607e+00
1.19653251e-02 -9.55671728e-01 6.57847166e-01 -4.56958771e-01
-1.70546472e+00 1.77009583e-01 -2.76002020e-01 -1.15030372e+00
-5.96386530e-02 -7.60579467e-01 -7.05406189e-01 5.28599858e-01
-9.79005814e-01 -8.73049378e-01 -2.01533154e-01 6.30628347e-01
5.70626497e-01 -1.83315456e-01 6.28789783e-01 8.17131877e-01
-5.77647865e-01 9.49289441e-01 -1.79950148e-01 8.31400678e-02
3.85882258e-01 -7.47293055e-01 5.64286888e-01 3.39383870e-01
-4.78850864e-02 1.18725121e+00 4.27700669e-01 -8.29392374e-01
-1.65786183e+00 -8.79737616e-01 7.40392983e-01 -5.43537736e-01
8.29079866e-01 -4.71283913e-01 -1.00350428e+00 8.87055516e-01
-1.83719411e-01 -5.69255531e-01 1.12612617e+00 7.97690034e-01
-4.76292014e-01 -4.35011685e-01 -1.06686437e+00 7.90880919e-01
1.33718026e+00 -7.71334991e-02 -2.55559981e-01 2.21823350e-01
8.42136085e-01 -1.17857367e-01 -8.87032747e-01 1.73708096e-01
1.13728690e+00 -7.98884988e-01 9.62334573e-01 -1.03764045e+00
4.96942818e-01 3.92672867e-01 -2.40722284e-01 -9.03150558e-01
-1.25922370e+00 -5.90751946e-01 -5.23455918e-01 8.27257574e-01
6.44021332e-01 -3.28591496e-01 1.13477600e+00 9.50778246e-01
1.36126563e-01 -9.12333727e-01 -3.74616385e-02 -4.61475968e-01
-1.48075908e-01 -7.00959563e-01 1.00033200e+00 9.63175476e-01
3.23150456e-01 6.63241744e-01 -8.11015427e-01 -1.49225369e-01
4.30119038e-01 2.68888623e-01 5.32943070e-01 -1.32126284e+00
-7.25146055e-01 -5.41209459e-01 2.65490890e-01 -1.74655831e+00
-4.96404357e-02 -7.58590817e-01 -4.12625521e-01 -9.50397909e-01
4.23834860e-01 -4.48343337e-01 -9.16610420e-01 3.52453679e-01
-8.10802355e-02 -1.15830101e-01 -1.31003678e-01 1.32789284e-01
-6.53158486e-01 3.07454228e-01 1.33223617e+00 -1.50597826e-01
-7.30165184e-01 6.88961685e-01 -1.75195527e+00 6.19991899e-01
7.62831628e-01 -2.14303598e-01 -7.10996389e-01 -3.04293811e-01
4.83071923e-01 1.98499143e-01 -1.03160873e-01 -4.74878192e-01
1.17624536e-01 -3.10620904e-01 5.33397555e-01 -5.67402124e-01
1.86959267e-01 -8.56137991e-01 -8.66233557e-02 4.49961394e-01
-8.77206683e-01 3.07646573e-01 -3.13325189e-02 8.55991304e-01
3.31950396e-01 6.10762201e-02 3.02474797e-01 -2.32107893e-01
-8.58527184e-01 8.08715224e-01 -4.79944527e-01 -4.74621028e-01
6.25430346e-01 3.64225730e-02 -1.71833411e-01 -5.72792768e-01
-7.75023818e-01 4.05273139e-01 -1.71074957e-01 1.07030034e+00
2.87781626e-01 -1.51699364e+00 -2.73381382e-01 4.71198291e-01
1.51441589e-01 -1.08310473e+00 4.22552615e-01 7.15330720e-01
4.15171266e-01 4.15753037e-01 -2.70292819e-01 -2.06282720e-01
-9.89799917e-01 8.85895967e-01 -2.31840294e-02 -2.91110545e-01
-3.32695276e-01 5.44616938e-01 6.25393763e-02 -4.07763958e-01
2.81584352e-01 -5.62568665e-01 -5.98647714e-01 -8.38962942e-02
7.75280893e-01 4.91610825e-01 -2.81557702e-02 -2.53523648e-01
-7.66038224e-02 5.07695414e-02 -8.08273017e-01 3.28125656e-01
1.15296876e+00 -2.01604620e-01 2.72825718e-01 2.31160343e-01
1.13993955e+00 -2.53990859e-01 -9.64042425e-01 -7.03034699e-01
2.36323550e-02 -5.56482196e-01 1.81934565e-01 -1.19167542e+00
-1.21929431e+00 3.62231910e-01 6.88478112e-01 4.61376339e-01
1.10820222e+00 -7.24004060e-02 1.39528108e+00 6.52455389e-01
3.98840964e-01 -1.19407296e+00 2.49436632e-01 4.18135852e-01
4.40884948e-01 -1.54986167e+00 -6.96040988e-02 -4.27803487e-01
-8.47493112e-01 6.46044672e-01 9.95499551e-01 -1.38204202e-01
1.11086607e+00 -1.20117115e-02 2.24546671e-01 -2.28680611e-01
-9.73222315e-01 6.42859787e-02 3.66041183e-01 4.55786705e-01
7.40683019e-01 3.58367145e-01 -5.80023348e-01 1.50180566e+00
-1.19410023e-01 3.34900081e-01 8.41139033e-02 4.88314271e-01
-3.75310153e-01 -1.15363705e+00 2.82248676e-01 9.41254079e-01
-6.90020978e-01 -1.52232036e-01 -2.08082452e-01 4.18262988e-01
2.28195414e-02 9.72176194e-01 2.38190796e-02 -9.98075068e-01
7.16315210e-01 -1.74731478e-01 3.19986671e-01 -3.55972171e-01
-9.21096504e-01 2.94842184e-01 3.85770872e-02 -1.07822084e+00
1.23967171e-01 -8.31087708e-01 -8.34274292e-01 -5.86000562e-01
-4.52945232e-02 -1.30631238e-01 4.68177766e-01 6.65887117e-01
5.87892830e-01 4.14624184e-01 8.83428872e-01 -1.65243134e-01
-1.12228560e+00 -9.34589326e-01 -6.01636827e-01 7.48166621e-01
-2.03708470e-01 -5.87413073e-01 1.99781150e-01 -2.23902613e-01] | [10.097901344299316, 5.603891372680664] |
d652a418-80e9-4570-b01c-aac088598a4f | electricity-an-efficient-transformer-for-non | null | null | https://www.mdpi.com/1424-8220/22/8/2926 | https://www.mdpi.com/1424-8220/22/8/2926 | ELECTRIcity: An Efficient Transformer for Non-Intrusive Load Monitoring | Non-Intrusive Load Monitoring (NILM) describes the process of inferring the consumption pattern of appliances by only having access to the aggregated household signal. Sequence-to-sequence deep learning models have been firmly established as state-of-the-art approaches for NILM, in an attempt to identify the pattern of the appliance power consumption signal into the aggregated power signal. Exceeding the limitations of recurrent models that have been widely used in sequential modeling, this paper proposes a transformer-based architecture for NILM. Our approach, called ELECTRIcity, utilizes transformer layers to accurately estimate the power signal of domestic appliances by relying entirely on attention mechanisms to extract global dependencies between the aggregate and the domestic appliance signals. Another additive value of the proposed model is that ELECTRIcity works with minimal dataset pre-processing and without requiring data balancing. Furthermore, ELECTRIcity introduces an efficient training routine compared to other traditional transformer-based architectures. According to this routine, ELECTRIcity splits model training into unsupervised pre-training and downstream task fine-tuning, which yields performance increases in both predictive accuracy and training time decrease. Experimental results indicate ELECTRIcity’s superiority compared to several state-of-the-art methods. | ['Nikolaos Doulamis', 'Anastasios Doulamis', 'Maria Kaselimi', 'Stavros Sykiotis'] | 2022-04-11 | electricity-an-efficient-transformer-for-non-1 | https://www.mdpi.com/1424-8220/22/8/2926/htm | https://www.mdpi.com/1424-8220/22/8/2926/pdf?version=1649762674 | mdpi-sensors-2022-4 | ['non-intrusive-load-monitoring', 'unsupervised-pre-training', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'methodology', 'miscellaneous', 'time-series'] | [ 2.26670057e-01 -1.14752017e-02 -4.98847514e-02 -3.86448115e-01
-6.29438877e-01 -3.71289521e-01 6.54337645e-01 -1.78681031e-01
5.76404184e-02 3.52704406e-01 1.69627979e-01 -1.45731226e-01
-9.01196525e-02 -7.39827573e-01 -4.42278624e-01 -7.66890168e-01
3.11570559e-02 3.78990114e-01 -3.94806504e-01 -1.27531260e-01
-1.36401176e-01 1.87150210e-01 -1.59451652e+00 1.57248065e-01
8.52444291e-01 1.23665559e+00 2.81424642e-01 5.22141397e-01
-8.14372227e-02 1.40667808e+00 -6.99857533e-01 -3.44173491e-01
1.09290034e-02 -6.03603840e-01 -7.26162016e-01 8.32613930e-02
-1.95433721e-01 -5.76057613e-01 -3.89266759e-01 9.28628206e-01
5.74422598e-01 -2.61152595e-01 4.46723670e-01 -1.33292472e+00
-5.21737933e-01 1.26814210e+00 -4.61509585e-01 2.64653593e-01
4.11908686e-01 2.72596747e-01 1.04652143e+00 -2.25277141e-01
-6.52872324e-02 7.74442911e-01 8.58324230e-01 1.69985756e-01
-1.58114028e+00 -7.53975034e-01 2.38858953e-01 6.32781327e-01
-1.26774287e+00 -4.80510175e-01 1.27352905e+00 -3.27061474e-01
1.61713326e+00 5.54047942e-01 7.62439251e-01 1.57192862e+00
5.62347844e-02 9.95706201e-01 9.79200065e-01 -3.81156713e-01
2.23406002e-01 2.53197759e-01 1.50797531e-01 2.16254830e-01
-1.15900107e-01 -1.49788642e-02 -4.24503714e-01 -4.08372842e-02
1.67987213e-01 6.84747398e-02 -2.25922763e-01 5.15859984e-02
-7.75614202e-01 7.01139271e-01 2.09816977e-01 7.00893581e-01
-8.33925784e-01 1.97803706e-01 7.65833855e-01 2.42978036e-01
7.45476604e-01 1.27839223e-01 -5.01057208e-01 -4.40129757e-01
-1.28840792e+00 -2.56580770e-01 9.16989982e-01 7.37603724e-01
5.23918629e-01 5.35340071e-01 -8.54630470e-02 5.63161671e-01
2.67469287e-01 4.61876631e-01 8.68161500e-01 -5.44174016e-01
5.10945678e-01 7.17295408e-01 -2.04457179e-01 -7.54807174e-01
-5.87257266e-01 -5.21585643e-01 -1.18715811e+00 -1.94491252e-01
-7.65580237e-02 -1.09034270e-01 -4.46364671e-01 1.80628979e+00
-6.90707704e-03 3.52702439e-01 -3.83387916e-02 3.72356027e-01
4.83649999e-01 8.37756813e-01 1.63528249e-01 -5.06669819e-01
1.25412846e+00 -6.11578345e-01 -9.70037580e-01 8.28661844e-02
3.93564194e-01 -4.88230407e-01 9.20031607e-01 3.57373714e-01
-8.39979291e-01 -5.72294831e-01 -9.54174817e-01 1.04615025e-01
-5.13375998e-01 2.14982659e-01 3.99419367e-01 6.19443357e-01
-8.96190345e-01 7.72387266e-01 -9.73278344e-01 -3.67064416e-01
4.45198387e-01 4.97939855e-01 8.64664316e-02 6.01585090e-01
-1.08411384e+00 1.02032018e+00 4.21926141e-01 4.59429562e-01
-8.86018038e-01 -9.32882607e-01 -6.93977535e-01 4.74016994e-01
1.53588086e-01 -5.39260805e-01 1.36778545e+00 -1.19387925e+00
-2.02328897e+00 4.17230368e-01 -1.58311784e-01 -8.71657610e-01
5.57576120e-01 -3.33567649e-01 -6.86606348e-01 -5.84220327e-02
-6.83684275e-02 -7.47517273e-02 9.12826002e-01 -9.23441529e-01
-6.57576084e-01 -2.28916377e-01 -3.08261663e-01 -2.62962192e-01
-5.57983339e-01 -1.46615297e-01 9.04782955e-03 -5.59984624e-01
-2.94588536e-01 -6.08832419e-01 4.14963886e-02 -9.99640405e-01
-6.88001871e-01 -6.12786710e-01 1.17826533e+00 -8.34082186e-01
1.47885382e+00 -2.08519959e+00 -2.84596860e-01 2.53885597e-01
-3.96589935e-02 3.29294741e-01 -3.26245874e-02 7.47422874e-01
-3.60603660e-01 -2.86510140e-01 -2.78934781e-02 -7.21156955e-01
5.54962456e-01 6.57806844e-02 -7.27254689e-01 5.96771240e-01
1.01975359e-01 1.31881917e+00 -7.60910630e-01 -1.81396738e-01
8.16633701e-01 7.63679504e-01 -5.58174103e-02 3.79086524e-01
-2.48032972e-01 3.84705454e-01 -1.65375113e-01 2.29124963e-01
4.21264946e-01 -3.46695453e-01 5.08848786e-01 -5.25085866e-01
-1.96890622e-01 7.36841798e-01 -7.60698855e-01 1.51902151e+00
-8.47011328e-01 6.07110739e-01 -2.10911632e-01 -1.46229768e+00
7.41796255e-01 7.83525527e-01 8.91962707e-01 -9.27618623e-01
4.08969492e-01 9.09376293e-02 -2.85449594e-01 -4.95834589e-01
1.89903975e-01 1.29474625e-01 -1.86547771e-01 5.57950914e-01
2.30551794e-01 2.06074581e-01 1.76125839e-01 -2.55752087e-01
9.96587455e-01 2.84169346e-01 4.94875252e-01 -2.70013839e-01
6.55873239e-01 -4.58529234e-01 5.38237095e-01 5.47973335e-01
2.33709767e-01 2.06111167e-02 2.25417912e-01 -4.98817176e-01
-1.00492358e+00 -7.76542246e-01 1.42019764e-01 1.03315735e+00
-3.22569251e-01 -4.71188694e-01 -9.24972534e-01 -3.83504540e-01
-1.54187769e-01 1.15387642e+00 -5.75688481e-01 -1.36865675e-01
-7.38542974e-01 -1.05190074e+00 5.01719832e-01 5.16624749e-01
6.68800235e-01 -1.32193720e+00 -1.02470863e+00 7.15751171e-01
-4.19954002e-01 -1.07907665e+00 -3.99041504e-01 7.12876558e-01
-4.34651107e-01 -9.81563747e-01 -3.09653908e-01 -5.35356700e-01
1.86521515e-01 -1.34746671e-01 1.40474176e+00 -4.14956421e-01
-3.18124294e-01 4.58306372e-01 -2.40706757e-01 -5.18488765e-01
-4.81112331e-01 5.43720722e-01 2.03920454e-02 4.79100436e-01
8.72311592e-01 -1.39287055e+00 -4.31464940e-01 4.02604304e-02
-6.36627197e-01 -9.84359980e-02 6.40634656e-01 5.12793779e-01
3.55764806e-01 4.92989093e-01 7.55627215e-01 -7.01898277e-01
5.41011453e-01 -6.52517557e-01 -9.45712924e-01 1.03683874e-01
-8.93082261e-01 -2.99161207e-02 1.14870977e+00 -4.99080896e-01
-9.60665405e-01 2.02500805e-01 -8.46874863e-02 -5.85188270e-01
-1.82959288e-01 1.34774506e-01 -3.20672631e-01 6.49927199e-01
-4.72680852e-02 7.61458814e-01 -2.59736657e-01 -7.38601267e-01
2.50244766e-01 7.41190493e-01 6.93041146e-01 -1.24890506e-01
7.55242288e-01 2.65103489e-01 -7.00735524e-02 -8.92508924e-01
-6.58753872e-01 -3.96678835e-01 -7.85957336e-01 1.51381479e-03
9.13728356e-01 -9.35008585e-01 -1.43082952e+00 7.13049352e-01
-1.12887323e+00 -3.67141217e-01 -6.60658360e-01 1.72945991e-01
-6.27531350e-01 2.42357969e-01 -6.11318827e-01 -1.33485937e+00
-8.35115910e-01 -8.94650578e-01 1.01340044e+00 -6.49777576e-02
-7.43607104e-01 -9.77304757e-01 -1.80699695e-02 1.27831012e-01
7.16990411e-01 2.14491263e-01 9.51514184e-01 -7.29124784e-01
-4.68937665e-01 -1.65867358e-01 1.62326053e-01 4.11020398e-01
4.44756567e-01 -2.58311152e-01 -1.40473962e+00 -3.37184638e-01
5.48319817e-01 -2.80053425e-03 4.10931766e-01 4.22300577e-01
1.23901725e+00 -6.23744845e-01 -2.75356919e-01 5.41147411e-01
1.48106909e+00 3.21429104e-01 4.96741951e-01 2.49535903e-01
6.81142151e-01 3.03454578e-01 -1.12072177e-01 6.07576072e-01
5.99488199e-01 7.22193062e-01 4.05682445e-01 -1.63705125e-01
1.36735186e-01 -5.70297956e-01 6.66123629e-01 1.31637609e+00
3.46722960e-01 -2.68999547e-01 -4.45096225e-01 5.61588109e-01
-1.97496271e+00 -1.32428658e+00 1.53300852e-01 1.83900428e+00
6.31648123e-01 9.23443660e-02 3.48697335e-01 7.61599004e-01
2.96344697e-01 2.63918817e-01 -8.24322701e-01 -3.44760895e-01
1.59263005e-03 2.76740313e-01 4.79849547e-01 3.08979303e-01
-1.16755772e+00 3.36184293e-01 6.15097761e+00 7.43349612e-01
-1.25470364e+00 2.81365752e-01 4.30044264e-01 -3.31103057e-01
-1.83759913e-01 -5.72281420e-01 -6.33617640e-01 1.06054091e+00
1.52421546e+00 -1.34861499e-01 7.36907184e-01 7.17849076e-01
7.61095166e-01 5.73611893e-02 -1.62969041e+00 1.19733107e+00
-6.64310157e-02 -8.13227534e-01 -1.56895101e-01 1.45251095e-01
3.27222675e-01 2.04084843e-01 -1.13156694e-03 3.05503130e-01
4.48330373e-01 -9.27873790e-01 8.86201024e-01 5.95669389e-01
4.81893182e-01 -7.82239854e-01 9.78061080e-01 5.00042379e-01
-1.58683836e+00 -4.02960360e-01 1.18551329e-01 -1.57793701e-01
2.86322564e-01 7.96102762e-01 -7.72080243e-01 8.75795007e-01
1.02526963e+00 8.49896669e-01 -3.34215432e-01 2.80930430e-01
-1.63651541e-01 1.03636348e+00 -5.88898122e-01 7.21568763e-02
1.37312070e-01 -3.45780730e-01 2.58650035e-01 1.30088961e+00
2.46123955e-01 -2.06673205e-01 2.12145388e-01 1.09531856e+00
4.74518724e-02 -1.57862216e-01 -6.58685923e-01 -2.68343315e-02
5.36533475e-01 1.10179603e+00 -3.63734096e-01 -3.43104601e-01
-4.42972183e-01 1.15084136e+00 1.45755941e-02 1.83339179e-01
-9.29331183e-01 -3.15471202e-01 6.69406950e-01 -4.41821292e-02
7.39817619e-01 2.05297068e-01 -2.25352541e-01 -1.00041270e+00
1.83163300e-01 -8.40155542e-01 2.49505118e-01 -5.51282227e-01
-1.45091510e+00 5.56854486e-01 -1.08901560e-01 -1.25657976e+00
-8.83885145e-01 4.97759245e-02 -8.31610799e-01 9.52461302e-01
-1.50416124e+00 -1.48967612e+00 -1.14453763e-01 6.14295542e-01
6.72726393e-01 -1.40285820e-01 1.10341597e+00 4.16310668e-01
-7.89400935e-01 5.73244512e-01 2.59713143e-01 2.33278245e-01
-2.16439515e-01 -1.32001865e+00 5.25041282e-01 9.28017199e-01
2.96620607e-01 1.64664730e-01 7.47191072e-01 -1.17900066e-01
-1.48221600e+00 -1.17569768e+00 9.41326976e-01 -5.43890595e-01
7.76042104e-01 -6.39681280e-01 -7.05596387e-01 8.69098723e-01
6.55483961e-01 -4.50529456e-01 8.14421415e-01 -7.79577345e-03
-2.14042261e-01 -6.53569996e-01 -9.90306973e-01 1.17358886e-01
6.72830045e-01 -9.15563703e-01 -6.75268888e-01 -3.81253846e-02
4.47586775e-01 1.81188375e-01 -9.68036473e-01 1.75810859e-01
5.05113244e-01 -9.79122818e-01 7.83316314e-01 -5.39510511e-03
6.71459287e-02 -3.11952978e-01 -3.35623592e-01 -1.32607746e+00
-5.21439731e-01 -1.02832782e+00 -6.77690327e-01 1.90150499e+00
8.64900351e-02 -8.87039423e-01 4.59480435e-01 3.09101224e-01
1.06356464e-01 -5.07606030e-01 -1.05024242e+00 -6.46438718e-01
-3.71512830e-01 -6.34573638e-01 1.15337479e+00 7.73936152e-01
1.84795633e-01 6.78490579e-01 -6.32467031e-01 2.98043460e-01
6.47853434e-01 3.48096967e-01 5.40748298e-01 -1.16356516e+00
-4.10038888e-01 -6.29372120e-01 -2.65622735e-01 -7.87743986e-01
3.99722934e-01 -8.33943784e-01 -8.68457854e-02 -1.22197068e+00
1.61975045e-02 4.20535468e-02 -6.95644200e-01 6.42311633e-01
1.92334875e-01 3.00650269e-01 3.27981114e-01 2.01663032e-01
-3.63154382e-01 6.68513715e-01 2.78716117e-01 -4.62507188e-01
-2.90448993e-01 2.08932340e-01 -5.71965635e-01 5.78223288e-01
9.82105315e-01 -2.67391413e-01 -6.82751298e-01 -2.57064104e-01
-1.14752673e-01 -3.18206668e-01 2.14840144e-01 -9.24020052e-01
1.05339591e-03 4.73101705e-01 4.36450362e-01 -1.04943001e+00
3.45568568e-01 -1.33715796e+00 6.41940594e-01 6.96304917e-01
-1.44881606e-01 1.91352263e-01 1.30577028e-01 3.38970631e-01
-1.04091257e-01 8.14985111e-02 6.08990550e-01 -1.44927010e-01
-5.36264718e-01 -6.44437177e-03 -6.19619489e-01 -4.78323370e-01
9.43806648e-01 -6.48556103e-04 3.37912925e-02 -4.44855630e-01
-5.48321486e-01 2.50455379e-01 -9.19859707e-02 4.91698205e-01
-5.81642753e-03 -1.30936933e+00 -5.40319741e-01 4.34883058e-01
-3.35002422e-01 -2.93786138e-01 1.44034877e-01 9.39328015e-01
1.46485299e-01 9.37166452e-01 2.47506499e-01 -7.78397024e-01
-1.03062928e+00 8.17184567e-01 3.73693883e-01 -6.99207246e-01
-8.71783257e-01 2.11316258e-01 -7.90008809e-03 -1.75023124e-01
3.25150192e-01 -6.31243587e-01 -2.26421520e-01 4.84922677e-01
5.35026252e-01 6.92203999e-01 3.53965342e-01 -6.05485737e-01
-3.73113215e-01 4.55677032e-01 8.00457150e-02 3.45579982e-01
1.63501644e+00 -4.16268975e-01 -2.72527158e-01 1.03126562e+00
1.33386230e+00 -4.55057234e-01 -9.11743164e-01 -1.12615474e-01
2.45717809e-01 2.00633362e-01 2.49484077e-01 -9.23249602e-01
-1.21646440e+00 6.97709680e-01 8.86526227e-01 8.45924795e-01
1.53757727e+00 -2.20524132e-01 9.91876364e-01 5.17014451e-02
4.58535224e-01 -1.17462492e+00 -4.86153096e-01 9.26482230e-02
5.12988389e-01 -1.05360162e+00 -3.05528522e-01 5.52492477e-02
-8.41928422e-02 8.65893602e-01 1.39550284e-01 7.29050040e-02
7.37134397e-01 6.24878049e-01 7.73881674e-02 -1.93973966e-02
-9.10138428e-01 -1.54113635e-01 -1.40891923e-02 6.33667469e-01
2.49320015e-01 2.13051036e-01 1.14414342e-01 7.61249244e-01
-5.74635386e-01 3.34697187e-01 8.81312415e-02 5.50455391e-01
1.85151756e-01 -8.31338465e-01 -1.88778073e-01 6.34006321e-01
-6.16530478e-01 -6.53555170e-02 5.74774817e-02 5.64483225e-01
1.76829081e-02 1.19478369e+00 3.11026424e-01 -4.23378766e-01
4.96688455e-01 4.02444333e-01 2.25352898e-01 -7.73689449e-02
-8.57742071e-01 -3.43226385e-03 -1.40942499e-01 -6.81437194e-01
-5.09120584e-01 -5.01408458e-01 -6.91702008e-01 -4.45681870e-01
-2.28247181e-01 1.69299781e-01 3.63128781e-01 1.08662570e+00
3.86309505e-01 9.17119324e-01 1.21452141e+00 -1.19944429e+00
-7.38521278e-01 -1.35450172e+00 -6.48994863e-01 4.13253337e-01
5.97728908e-01 -1.94252178e-01 -3.25528026e-01 7.06798732e-02] | [16.070533752441406, 7.584036350250244] |
bdb072c4-43c7-4e5f-b27f-a5d3039f9c99 | automatic-cell-counting-in-flourescent | 2103.01141 | null | https://arxiv.org/abs/2103.01141v1 | https://arxiv.org/pdf/2103.01141v1.pdf | Automatic Cell Counting in Flourescent Microscopy Using Deep Learning | Counting cells in fluorescent microscopy is a tedious, time-consuming task that researchers have to accomplish to assess the effects of different experimental conditions on biological structures of interest. Although such objects are generally easy to identify, the process of manually annotating cells is sometimes subject to arbitrariness due to the operator's interpretation of the borderline cases. We propose a Machine Learning approach that exploits a fully-convolutional network in a binary segmentation fashion to localize the objects of interest. Counts are then retrieved as the number of detected items. Specifically, we adopt a UNet-like architecture leveraging residual units and an extended bottleneck for enlarging the field-of-view. In addition, we make use of weighted maps that penalize the errors on cells boundaries increasingly with overcrowding. These changes provide more context and force the model to focus on relevant features during pixel-wise classification. As a result, the model performance is enhanced, especially in presence of clumping cells, artifacts and confounding biological structures. Posterior assessment of the results with domain experts confirms that the model detects cells of interest correctly. The model demonstrates a human-level ability inasmuch even erroneous predictions seem to fall within the limits of operator interpretation. This qualitative assessment is also corroborated by quantitative metrics as an ${F_1}$ score of 0.87. Despite some difficulties in interpretation, results are also satisfactory with respect to the counting task, as testified by mean and median absolute error of, respectively, 0.8 and 1. | ['A. Zoccoli', 'L. Rinaldi', 'M. Luppi', 'M. Dalla', 'L. Clissa', 'R. Morelli'] | 2021-02-24 | null | null | null | null | ['automatic-cell-counting'] | ['miscellaneous'] | [ 3.83265078e-01 7.35601410e-02 4.19004768e-01 -2.15580419e-01
-5.80025315e-01 -7.19909906e-01 3.94224524e-01 6.79194808e-01
-1.11551404e+00 1.02282107e+00 -4.40938950e-01 -1.42196015e-01
8.60973969e-02 -4.83261615e-01 -7.82663226e-01 -1.05052614e+00
1.05625808e-01 3.30271155e-01 3.66872638e-01 3.01944017e-01
5.44840336e-01 7.88160145e-01 -1.51599586e+00 1.93938032e-01
7.35440612e-01 9.98019993e-01 3.00591916e-01 7.07216144e-01
-4.34744284e-02 4.84004885e-01 -7.68844008e-01 -2.28613302e-01
1.96280181e-02 -9.42109004e-02 -6.23897076e-01 1.76190019e-01
4.88128752e-01 -2.36517176e-01 4.05514628e-01 1.06666243e+00
4.40739781e-01 -1.45939976e-01 8.01181018e-01 -7.62385845e-01
-2.50157237e-01 1.15711085e-01 -5.59484065e-01 6.65784299e-01
1.55826062e-01 6.22572422e-01 7.24983394e-01 -9.05810297e-01
6.78588390e-01 8.09773386e-01 7.10548043e-01 2.00959235e-01
-1.55849159e+00 -3.51624399e-01 1.46343917e-01 -1.99084327e-01
-1.43045676e+00 -2.32338905e-01 2.75159121e-01 -7.97688544e-01
8.82172048e-01 2.30642498e-01 5.61809838e-01 7.53947437e-01
3.34442466e-01 2.35390842e-01 1.26743388e+00 -3.06683630e-01
4.45261747e-01 4.64360088e-01 -9.72351506e-02 5.16774297e-01
5.73974907e-01 -2.50288695e-01 -1.47936597e-01 1.28787071e-01
9.08341885e-01 4.52972539e-02 -1.94248781e-01 -3.10657948e-01
-1.29155099e+00 4.51552033e-01 2.92320848e-01 5.48066258e-01
-2.80264735e-01 6.71209320e-02 4.08018976e-01 -2.61211932e-01
4.17799085e-01 6.37183309e-01 -4.27609593e-01 2.29819193e-01
-1.07036734e+00 8.80914181e-02 3.31559479e-01 3.43239903e-01
6.06924832e-01 -2.82559633e-01 -1.94173738e-01 7.00629711e-01
8.42054635e-02 2.33086765e-01 2.97413081e-01 -1.00183642e+00
-7.07830787e-02 1.04228890e+00 3.38816762e-01 -8.51833522e-01
-6.39656305e-01 -7.38756657e-01 -8.38087022e-01 7.76020765e-01
1.06345904e+00 7.80544654e-02 -8.41335595e-01 1.46509778e+00
2.43625462e-01 -2.32489944e-01 -3.96725714e-01 9.87412930e-01
3.82028490e-01 2.28995577e-01 4.50218797e-01 -1.04461052e-01
1.50255203e+00 -4.92442787e-01 -5.79516470e-01 -9.87153947e-02
7.02005625e-01 -5.34847558e-01 1.22764087e+00 4.03268903e-01
-1.00620568e+00 -5.24414659e-01 -9.89653230e-01 -5.28993718e-02
-6.83910847e-01 3.87425065e-01 5.09830475e-01 4.53255862e-01
-1.04311323e+00 7.07462549e-01 -7.39745617e-01 -4.54361677e-01
9.44995284e-01 5.96664429e-01 -5.19674778e-01 3.07765871e-01
-5.46229482e-01 9.22020853e-01 3.56314510e-01 3.29359442e-01
-5.92692196e-01 -7.49097943e-01 -5.14004529e-01 2.26660103e-01
8.76110196e-02 -5.92837095e-01 6.99248970e-01 -7.93949783e-01
-1.05506158e+00 1.33548474e+00 -2.45714694e-01 -3.97877097e-01
8.23898792e-01 2.05691785e-01 5.09576797e-02 2.82750934e-01
1.59834236e-01 8.42896283e-01 3.32361013e-01 -1.20646167e+00
-8.37906182e-01 -5.32823503e-01 1.31301349e-02 -7.29725659e-02
-1.35168388e-01 -6.19888119e-02 -1.51591659e-01 -3.94145250e-01
-6.83631375e-02 -5.81112623e-01 -3.52562815e-01 5.61088264e-01
-1.27434224e-01 -1.93883236e-02 5.86612403e-01 -4.12938923e-01
1.00503743e+00 -2.06306052e+00 -3.12585533e-01 1.42389506e-01
4.88451898e-01 1.00632519e-01 2.22764418e-01 -2.90820468e-02
8.76682252e-03 2.75356442e-01 -3.85316223e-01 -3.25411350e-01
-2.68265754e-01 -1.04298227e-01 1.03444636e-01 7.24720478e-01
5.61808407e-01 6.98131382e-01 -8.85042310e-01 -6.60481930e-01
3.69801104e-01 6.19013727e-01 -3.47048849e-01 -1.58997066e-02
5.27222045e-02 6.38519764e-01 -1.55471578e-01 6.51330650e-01
6.34480000e-01 -6.37642026e-01 8.42014235e-03 -2.25705519e-01
-4.62273538e-01 -1.00205295e-01 -1.09776759e+00 1.22801006e+00
-2.04593495e-01 5.75151742e-01 2.60561138e-01 -7.93901026e-01
7.62871265e-01 7.64977857e-02 3.36695492e-01 -4.22625542e-01
3.75265479e-01 5.10345638e-01 1.80054858e-01 -3.71226370e-01
1.89937145e-01 -2.23241881e-01 3.69101793e-01 2.92997509e-01
4.01209854e-02 -8.44153538e-02 4.31733519e-01 4.97506484e-02
1.12008882e+00 6.90112188e-02 4.20664936e-01 -6.10817254e-01
6.40478909e-01 -9.64119881e-02 2.88012475e-01 5.99876404e-01
-3.81739795e-01 7.83639729e-01 7.01188385e-01 -5.67073047e-01
-1.01458359e+00 -8.10781240e-01 -4.56224829e-01 7.65100479e-01
7.06954300e-03 2.45690942e-01 -8.48199248e-01 -5.74655533e-01
-6.30532280e-02 4.18357491e-01 -9.65597034e-01 2.28681207e-01
-2.80619383e-01 -9.64606225e-01 2.99887955e-01 4.40864861e-01
2.44217068e-01 -1.13879037e+00 -1.36452103e+00 2.26654857e-01
7.10871071e-02 -1.06249857e+00 -4.50288281e-02 6.45530581e-01
-8.45885515e-01 -1.08494437e+00 -9.51467395e-01 -6.08393192e-01
1.09684849e+00 2.95831803e-02 8.94680202e-01 4.26316261e-01
-5.39690554e-01 -6.75344244e-02 -1.02763705e-01 -4.24090087e-01
-2.01667935e-01 -6.08875267e-02 -2.89616823e-01 -1.69859454e-01
4.71528441e-01 -3.29637587e-01 -8.96268427e-01 3.88478309e-01
-1.00496030e+00 -2.61990190e-01 4.48854148e-01 9.55073535e-01
8.00859034e-01 -1.21128961e-01 3.92503947e-01 -9.44333553e-01
3.08292031e-01 -1.03227049e-01 -7.64918089e-01 2.77666585e-03
-4.20466870e-01 -1.22195922e-01 6.90953076e-01 -4.03454393e-01
-6.44183338e-01 1.52232766e-01 2.14005768e-01 -2.11436525e-01
-5.29370010e-01 1.47625402e-01 1.60449937e-01 -2.63859361e-01
6.93611741e-01 -1.46293208e-01 1.21909268e-01 -4.95461561e-03
-4.08837676e-01 4.49541479e-01 3.79748285e-01 -2.77871817e-01
3.48340929e-01 8.71216118e-01 1.28192395e-01 -5.18400669e-01
-6.10917270e-01 -5.28475106e-01 -6.37583911e-01 -4.25328702e-01
8.09586346e-01 -6.05035245e-01 -1.04053795e+00 4.81691062e-01
-1.16167688e+00 -4.98009115e-01 -2.56637186e-01 3.34380209e-01
-3.18614364e-01 4.39563960e-01 -5.08918226e-01 -9.24923360e-01
-1.19877838e-01 -1.30405951e+00 1.09082592e+00 3.52612734e-01
-5.30464292e-01 -1.03456986e+00 -3.41622859e-01 2.97445506e-01
3.64853114e-01 5.34215808e-01 9.80065823e-01 -6.63647413e-01
-4.72561985e-01 -3.26241285e-01 -4.16581750e-01 1.98972419e-01
7.88648278e-02 3.87920201e-01 -1.39279830e+00 -1.38864666e-01
-1.66096464e-01 -9.89862382e-02 7.73336411e-01 6.34939373e-01
1.48097527e+00 1.46333233e-01 -4.84709829e-01 3.08699369e-01
1.49566531e+00 1.20457672e-01 5.81228673e-01 3.95604253e-01
1.84871420e-01 9.29790080e-01 3.95937532e-01 4.02803093e-01
-4.33511510e-02 4.29485887e-01 7.08099425e-01 -5.77537537e-01
4.21336889e-02 2.19172671e-01 -1.13672353e-01 -3.94479483e-02
-1.89681634e-01 -3.39818865e-01 -9.39080477e-01 6.52930737e-01
-1.41713226e+00 -6.96256757e-01 -3.59299362e-01 2.47701073e+00
5.96599579e-01 4.61678296e-01 7.68229812e-02 3.13558668e-01
7.95443773e-01 -2.70854205e-01 -5.16283035e-01 -3.58958393e-01
-2.22186461e-01 1.03856817e-01 6.37014508e-01 4.11799133e-01
-9.82725859e-01 7.00034916e-01 6.10756493e+00 6.37712777e-01
-1.24531174e+00 -2.65394628e-01 1.33276939e+00 -8.47404823e-02
3.12079135e-02 -2.19776377e-01 -6.91457570e-01 8.10332775e-01
5.24058878e-01 3.40587229e-01 1.60084918e-01 3.76163512e-01
5.26923597e-01 -7.79419541e-01 -1.17120457e+00 7.01713860e-01
-2.19260067e-01 -1.34383130e+00 -1.11494638e-01 2.56412566e-01
4.10430819e-01 -1.85951591e-01 8.01230371e-02 -2.80594882e-02
-1.17235277e-02 -1.21721661e+00 7.22146511e-01 5.85351050e-01
1.02035427e+00 -3.97506505e-01 1.17763746e+00 2.02952564e-01
-7.97888696e-01 -4.67820205e-02 -4.04772729e-01 -1.74502656e-01
9.81919616e-02 8.16364467e-01 -9.42635596e-01 -2.17253685e-01
8.38742256e-01 6.56744689e-02 -7.35321164e-01 1.12052238e+00
1.19304381e-01 2.80790836e-01 -4.72824901e-01 -1.50258347e-01
1.37137055e-01 -7.90814683e-02 2.71166980e-01 1.47696149e+00
3.42106909e-01 -1.13149822e-01 -3.37895632e-01 1.17581415e+00
5.09584323e-02 9.91521850e-02 -4.59139585e-01 1.04096353e-01
3.31446797e-01 1.60433960e+00 -1.48730063e+00 -3.02414447e-01
-1.98966667e-01 6.96116447e-01 3.79063994e-01 4.04751480e-01
-7.38389552e-01 -3.29347551e-01 2.22117648e-01 5.41872978e-01
3.23854059e-01 1.25745893e-01 -7.21129656e-01 -6.05400264e-01
1.64499566e-01 -4.52547818e-01 1.52584717e-01 -7.87004054e-01
-1.01412165e+00 3.85010779e-01 -3.99828672e-01 -1.06354606e+00
1.78061724e-01 -1.06277812e+00 -4.99555200e-01 8.26948047e-01
-1.34152114e+00 -6.67815328e-01 -4.05736148e-01 2.42181774e-02
2.45780662e-01 2.89741307e-01 7.43482888e-01 2.55062401e-01
-4.48688149e-01 4.54937726e-01 -1.20045088e-01 8.67428854e-02
5.76564074e-01 -1.33084810e+00 -1.43606931e-01 6.57188773e-01
-2.17842296e-01 8.52338791e-01 9.26381290e-01 -2.98975289e-01
-5.50235331e-01 -7.23008931e-01 9.19757843e-01 -4.64266121e-01
6.06945693e-01 -3.48763853e-01 -1.02385986e+00 2.35926196e-01
-1.00759320e-01 3.88227075e-01 7.05648005e-01 -1.28149733e-01
-1.37483239e-01 1.74576901e-02 -1.44523335e+00 5.71901202e-01
7.15167224e-01 -1.99180245e-01 -8.18726793e-02 2.20232531e-01
2.75368877e-02 -1.58998370e-01 -7.83664525e-01 3.18145305e-01
6.23669744e-01 -1.19461703e+00 6.90854490e-01 -2.59186894e-01
4.04726386e-01 -5.11824906e-01 2.42774308e-01 -8.83868396e-01
-3.33250195e-01 -4.58791628e-02 5.46867430e-01 9.77802277e-01
6.23591900e-01 -5.64741135e-01 9.29112971e-01 5.77893913e-01
5.41904569e-03 -8.81226063e-01 -1.11292231e+00 -5.02986491e-01
5.08774212e-03 1.87502410e-02 1.76857531e-01 7.50028253e-01
-5.81950024e-02 1.72854260e-01 3.41396034e-01 -7.72008672e-03
4.44058508e-01 -1.93002135e-01 5.69493055e-01 -1.39900875e+00
-9.83105227e-02 -7.67202199e-01 -6.44112408e-01 -6.18430197e-01
-1.69308662e-01 -5.99949062e-01 6.82462305e-02 -1.30045700e+00
3.20995212e-01 -3.38239044e-01 -4.99615103e-01 3.08571517e-01
-2.52000004e-01 5.16437054e-01 -5.07252701e-02 6.48931414e-02
-5.91125011e-01 -1.77320406e-01 1.18557847e+00 -1.07795417e-01
9.46188420e-02 -2.14675531e-01 -5.28638959e-01 7.94650376e-01
8.36993337e-01 -4.21001703e-01 1.07929118e-01 -4.07603025e-01
3.45707864e-01 -3.79436523e-01 8.00183117e-01 -1.24088383e+00
1.10487111e-01 9.98373702e-02 8.43018353e-01 -3.37181658e-01
1.82436585e-01 -9.62388277e-01 1.23072602e-01 6.12209439e-01
-2.98317164e-01 -1.35231346e-01 3.07764769e-01 3.94340158e-01
-5.63278943e-02 -3.59839231e-01 1.09128988e+00 -3.21375132e-01
-2.53210336e-01 -1.41428038e-01 -7.18370616e-01 -2.94858277e-01
9.82427597e-01 -9.02812183e-01 -3.67851913e-01 -7.19350725e-02
-7.58796871e-01 5.63971885e-02 8.50856781e-01 -4.18024153e-01
2.08710089e-01 -6.18570507e-01 -4.55795079e-01 3.08637060e-02
3.11781019e-01 2.41011426e-01 3.64233077e-01 1.24556172e+00
-8.12035441e-01 3.64978403e-01 -2.62592047e-01 -7.89103508e-01
-1.23019612e+00 4.16045636e-01 4.47706282e-01 -5.25109529e-01
-1.71092376e-01 9.70572829e-01 4.08750683e-01 -7.50135854e-02
1.58150196e-01 -7.42322803e-01 -3.30574125e-01 1.30375266e-01
4.13429886e-01 4.31562632e-01 1.89720154e-01 -3.64901483e-01
-4.02586281e-01 5.66312850e-01 -1.45307526e-01 1.68713957e-01
1.17202139e+00 -1.39112666e-01 -1.26058295e-01 5.06859004e-01
7.93964326e-01 9.89978388e-02 -1.67376494e+00 2.82516390e-01
1.87400095e-02 -2.55713671e-01 -1.66742235e-01 -1.12962282e+00
-8.30240428e-01 1.07825506e+00 7.74679601e-01 1.70070127e-01
8.28077793e-01 -1.16136588e-01 1.26390964e-01 1.23359278e-01
2.39403650e-01 -1.16685569e+00 -1.53067231e-01 2.76473016e-01
5.41389287e-01 -1.35921836e+00 7.54100680e-02 -3.52746397e-01
-1.63033813e-01 1.11493409e+00 6.27357364e-01 -1.97761524e-02
1.18132919e-01 4.74361122e-01 2.15685740e-01 -2.47973949e-01
-5.47457337e-01 -6.66363612e-02 -2.37236321e-01 5.77506602e-01
5.92737377e-01 -1.64848953e-01 -5.28564036e-01 3.43778461e-01
1.57413006e-01 1.89619064e-01 5.21776676e-01 6.29495859e-01
-6.38587892e-01 -4.92264599e-01 -3.96906674e-01 5.06196201e-01
-7.43492365e-01 6.07085675e-02 -3.34033608e-01 9.23459291e-01
3.43330175e-01 6.78829670e-01 4.25148815e-01 1.49300188e-01
5.21888882e-02 -1.63684800e-01 2.89229691e-01 -5.26560545e-01
-7.47422278e-01 2.47519448e-01 -3.89352709e-01 -3.37142915e-01
-4.93448824e-01 -5.09544790e-01 -1.47283340e+00 -8.83668941e-03
-3.01055551e-01 -7.35310242e-02 4.54114825e-01 7.75207639e-01
2.25822598e-01 6.20436251e-01 1.10603236e-01 -8.04989874e-01
-1.40291736e-01 -9.49162722e-01 -6.29907310e-01 3.97247970e-01
2.43686423e-01 -7.02988744e-01 -6.10251665e-01 3.74308825e-01] | [14.670273780822754, -3.1581003665924072] |
191f5348-1f6b-404f-83d4-aab416adac75 | deep-view-sensitive-pedestrian-attribute | 1707.06089 | null | http://arxiv.org/abs/1707.06089v1 | http://arxiv.org/pdf/1707.06089v1.pdf | Deep View-Sensitive Pedestrian Attribute Inference in an end-to-end Model | Pedestrian attribute inference is a demanding problem in visual surveillance
that can facilitate person retrieval, search and indexing. To exploit semantic
relations between attributes, recent research treats it as a multi-label image
classification task. The visual cues hinting at attributes can be strongly
localized and inference of person attributes such as hair, backpack, shorts,
etc., are highly dependent on the acquired view of the pedestrian. In this
paper we assert this dependence in an end-to-end learning framework and show
that a view-sensitive attribute inference is able to learn better attribute
predictions. Our proposed model jointly predicts the coarse pose (view) of the
pedestrian and learns specialized view-specific multi-label attribute
predictions. We show in an extensive evaluation on three challenging datasets
(PETA, RAP and WIDER) that our proposed end-to-end view-aware attribute
prediction model provides competitive performance and improves on the published
state-of-the-art on these datasets. | ['Yan Wang', 'Rainer Stiefelhagen', 'M. Saquib Sarfraz', 'Arne Schumann'] | 2017-07-19 | null | null | null | null | ['person-retrieval', 'multi-label-image-classification'] | ['computer-vision', 'computer-vision'] | [ 1.26690581e-01 -1.10842278e-02 -3.52953792e-01 -9.55432951e-01
-7.63849974e-01 -5.87701321e-01 1.03467786e+00 2.62901366e-01
-3.01994652e-01 7.86489964e-01 4.11231846e-01 3.60980660e-01
-6.40879497e-02 -5.50066769e-01 -8.12461793e-01 -6.89842641e-01
1.21728063e-01 1.14775956e+00 2.43763149e-01 4.07640301e-02
3.40078101e-02 4.88190502e-01 -2.11300516e+00 6.03626609e-01
3.18135500e-01 1.38585472e+00 -1.31599709e-01 6.12410069e-01
3.82604122e-01 6.89723909e-01 -1.56006917e-01 -9.47260439e-01
2.41406605e-01 3.01779300e-01 -8.76182079e-01 2.92800993e-01
1.32775974e+00 -2.80050725e-01 -2.20635489e-01 6.58707023e-01
5.44731379e-01 -3.85893434e-02 1.27666461e+00 -1.48776543e+00
-4.98254448e-01 -1.81931809e-01 -5.77224255e-01 -3.24288718e-02
6.74845219e-01 4.15042639e-02 1.16660547e+00 -1.03364050e+00
7.75094986e-01 1.50266159e+00 8.43168736e-01 5.84994674e-01
-1.12241971e+00 -4.25062418e-01 4.76708740e-01 6.81199074e-01
-1.51621616e+00 -3.94832939e-01 7.03136265e-01 -6.04406357e-01
5.46348631e-01 5.38075507e-01 8.41348946e-01 1.79788148e+00
-1.86700728e-02 1.01760018e+00 1.39551532e+00 -3.96754183e-02
-8.76740515e-02 4.00582045e-01 -1.28849400e-02 9.63655710e-01
7.25496411e-02 1.55973464e-01 -8.02484810e-01 -3.16920429e-01
2.46657252e-01 2.39081487e-01 4.66976501e-02 -8.95668030e-01
-1.28592622e+00 6.44115686e-01 3.58837485e-01 -5.79870820e-01
-4.96383429e-01 1.03116699e-01 5.70596159e-01 -6.52101561e-02
6.27570629e-01 -7.92865679e-02 -5.28184950e-01 3.00132155e-01
-7.59246647e-01 3.21219593e-01 7.57973492e-01 1.27016354e+00
8.43446672e-01 -5.16432941e-01 -6.78698063e-01 7.97049463e-01
4.03259993e-01 8.57185781e-01 -3.81875157e-01 -9.51528370e-01
2.94467896e-01 5.50500095e-01 1.50818139e-01 -6.81165695e-01
-5.10831177e-01 -3.87270600e-01 -8.32185566e-01 1.23387337e-01
5.00323057e-01 1.43451214e-01 -8.69612932e-01 1.50900269e+00
7.59464741e-01 6.69291914e-02 -2.96027571e-01 1.07926655e+00
9.74718153e-01 2.97967255e-01 7.70241559e-01 6.86161518e-02
1.81284559e+00 -1.18312371e+00 -3.97300094e-01 -4.86540318e-01
1.17581755e-01 -6.93530619e-01 7.69127309e-01 2.61858821e-01
-8.30813885e-01 -7.78860509e-01 -7.24312127e-01 -1.53461859e-01
-6.04803026e-01 4.15068716e-01 7.07775593e-01 6.10030651e-01
-9.78540599e-01 2.18459129e-01 -2.48955488e-01 -9.57556248e-01
8.25144947e-01 2.15034485e-01 -8.10779750e-01 -2.30371177e-01
-9.87331569e-01 9.36536372e-01 1.66829094e-01 -2.48753309e-01
-1.25861919e+00 -6.68463409e-01 -8.18371713e-01 -2.85696834e-01
6.48646653e-01 -1.60043430e+00 7.35332012e-01 -6.59677923e-01
-8.78615558e-01 1.67908013e+00 -3.58011872e-01 -2.37664402e-01
8.11670303e-01 -4.85135108e-01 -3.08513582e-01 4.08686399e-01
3.17076981e-01 9.58494008e-01 1.15659833e+00 -1.48767054e+00
-9.12941694e-01 -9.48015630e-01 4.40728432e-03 4.19388533e-01
-1.56563461e-01 -5.03014289e-02 -4.36923742e-01 -4.80177969e-01
-3.07931989e-01 -1.13022828e+00 4.51340228e-02 4.98028725e-01
-5.63935220e-01 -4.89000648e-01 7.05211639e-01 -5.70182621e-01
5.53695560e-01 -1.75260258e+00 1.92607567e-01 1.27561972e-01
1.74760178e-01 -2.01814994e-02 2.14553311e-01 2.37672970e-01
1.55721858e-01 -2.97921777e-01 1.47202060e-01 -5.93845308e-01
2.01713040e-01 2.61819035e-01 3.27528566e-02 9.14105654e-01
7.47304335e-02 8.05110931e-01 -6.64262831e-01 -1.10743606e+00
3.75524312e-01 7.02480972e-01 -3.78509879e-01 5.17524719e-01
-1.75272420e-01 5.81233025e-01 -6.55456126e-01 1.11660659e+00
5.75804532e-01 -2.67823696e-01 -2.76827842e-01 -7.36903667e-01
2.48801913e-02 -4.14306730e-01 -8.16439331e-01 1.74371874e+00
-2.93948293e-01 3.33231360e-01 1.12087779e-04 -8.19281518e-01
8.02945673e-01 2.30148628e-01 4.44195837e-01 -2.43671894e-01
-1.36192769e-01 -2.12565184e-01 -9.38453972e-01 -6.57463729e-01
2.39205480e-01 1.89245448e-01 -2.91197091e-01 -9.20633078e-02
1.80328771e-01 4.37877804e-01 -7.77614415e-02 1.21051945e-01
3.77515703e-01 5.02997994e-01 6.55360401e-01 -4.94207174e-01
9.20181751e-01 -1.08758032e-01 2.08150074e-01 7.79489279e-01
-3.90378952e-01 5.71417093e-01 2.08651587e-01 -1.13471711e+00
-1.26699269e+00 -1.28137589e+00 -2.97422379e-01 1.86028993e+00
2.76024848e-01 -2.77156085e-01 -7.07022667e-01 -1.03616726e+00
1.39177755e-01 4.76696372e-01 -8.67326915e-01 1.74731717e-01
-2.98263222e-01 -5.40530205e-01 4.31569815e-01 7.38453865e-01
5.78248858e-01 -8.20123434e-01 -6.00226343e-01 -2.18210325e-01
-4.62420940e-01 -1.61329210e+00 -4.79086071e-01 -1.05594896e-01
-3.03860962e-01 -1.18620503e+00 -6.76976740e-01 -5.22565365e-01
4.97869104e-01 -6.06880002e-02 1.62447107e+00 1.07240947e-02
-3.21578294e-01 1.06650543e+00 -1.71656445e-01 -5.36303937e-01
1.18195690e-01 1.22116633e-01 1.86802909e-01 5.37257969e-01
5.58566988e-01 -3.97202760e-01 -9.57077682e-01 5.49939930e-01
-7.36576691e-02 7.29716048e-02 8.48433673e-01 8.01779926e-01
1.04420781e+00 -8.85043323e-01 2.60437459e-01 -1.08858871e+00
-1.65593609e-01 -2.94993252e-01 -3.16490054e-01 6.76618278e-01
-4.49192673e-01 1.81984186e-01 2.33521461e-01 -9.70189199e-02
-1.17654240e+00 6.37549341e-01 -1.14714146e-01 -3.80094200e-01
-1.03655100e+00 -3.03824872e-01 -5.43238163e-01 -7.01299533e-02
4.51533765e-01 1.92412779e-01 -1.18750595e-01 -3.48723322e-01
4.84358281e-01 3.84851187e-01 7.41235554e-01 -8.27914059e-01
5.45537174e-01 8.61174762e-01 6.53722584e-01 -7.16495335e-01
-1.47666943e+00 -8.41815233e-01 -1.13113081e+00 -7.35107839e-01
1.42454994e+00 -1.26615047e+00 -1.11883593e+00 2.79323190e-01
-1.06073630e+00 1.01617195e-01 8.01213384e-02 1.45339087e-01
-1.01554191e+00 1.14371262e-01 -1.87177405e-01 -7.01332390e-01
-4.25415993e-01 -9.21194851e-01 1.78362238e+00 1.10141031e-01
-1.22580029e-01 -9.11637127e-01 -2.96920478e-01 7.81142712e-01
2.00036705e-01 5.72885692e-01 6.70615375e-01 -8.54681492e-01
-7.12910354e-01 -2.57697225e-01 -4.41675782e-01 -2.44585514e-01
-4.19707775e-01 -4.13377136e-01 -1.53257620e+00 -3.59792620e-01
-5.75305045e-01 -5.43623388e-01 1.10230827e+00 2.73866087e-01
1.44828701e+00 -2.86635071e-01 -8.41046154e-01 7.50257194e-01
1.21618330e+00 -6.76542819e-01 3.99721056e-01 2.80046672e-01
9.56764042e-01 8.71568859e-01 1.02556336e+00 6.79611981e-01
7.34008133e-01 1.09839153e+00 7.50007689e-01 -5.16534150e-02
-3.60926628e-01 -1.26574874e-01 6.57907203e-02 -1.43027410e-01
-5.55998147e-01 -1.84270069e-01 -6.74444973e-01 3.63356501e-01
-1.95315015e+00 -1.13583040e+00 -1.94427505e-01 2.11570191e+00
4.02845860e-01 -8.50389376e-02 5.31074584e-01 -2.70926297e-01
7.89521873e-01 2.60245711e-01 -4.53226358e-01 -2.31973287e-02
-9.12985727e-02 -4.52891797e-01 5.29718280e-01 3.03819478e-01
-1.95042980e+00 8.22975576e-01 5.79360819e+00 7.70393550e-01
-3.65055948e-01 1.18292354e-01 1.00371218e+00 3.88372391e-02
1.29636362e-01 -4.80770707e-01 -1.31341708e+00 2.20196009e-01
5.59633613e-01 4.12735730e-01 8.25995579e-02 1.09145355e+00
-4.52264786e-01 2.54304260e-02 -1.68717623e+00 1.15347517e+00
4.03789967e-01 -9.22555923e-01 4.06005710e-01 -3.58759868e-03
3.67171615e-01 -3.90871674e-01 3.66416663e-01 1.89591840e-01
3.96645209e-03 -1.05557930e+00 7.24825561e-01 1.00432527e+00
9.60009754e-01 -8.68320465e-01 8.48138869e-01 1.52038857e-01
-1.51704395e+00 -4.51892503e-02 -4.32774633e-01 3.79799843e-01
6.76596463e-02 1.49371058e-01 -8.23213637e-01 5.29545963e-01
9.80501950e-01 8.77518415e-01 -1.08111155e+00 8.68929446e-01
-1.65904418e-01 2.91622311e-01 8.65052361e-03 1.61038682e-01
1.55832335e-01 1.05312960e-02 5.42877316e-01 1.49693644e+00
8.06227550e-02 1.96986258e-01 5.69695592e-01 4.43301201e-01
1.43071786e-01 -6.53518960e-02 -7.52166092e-01 7.28689492e-01
3.73416692e-01 1.45784199e+00 -2.75628209e-01 -3.65105748e-01
-5.19978464e-01 1.18157732e+00 3.93130273e-01 3.15523267e-01
-8.57193232e-01 4.67684895e-01 9.28281128e-01 3.23148400e-01
5.62804639e-01 2.85777360e-01 -3.59153748e-02 -7.31517613e-01
-8.64406228e-02 -6.13049805e-01 9.43101466e-01 -7.85533190e-01
-1.72784984e+00 5.39769411e-01 2.97207087e-01 -1.29605865e+00
-5.42738795e-01 -8.75811338e-01 -5.47842868e-03 8.03576052e-01
-1.44059980e+00 -2.21498442e+00 -7.95114815e-01 1.01385391e+00
6.97660208e-01 -5.27023613e-01 1.12342465e+00 3.04970354e-01
-3.09206750e-02 7.26286232e-01 -4.24234062e-01 6.73793023e-03
1.12403023e+00 -1.37059057e+00 7.87565038e-02 3.11091393e-01
1.81154564e-01 7.95429479e-03 9.60817337e-01 -4.61187214e-01
-1.17278373e+00 -1.38360775e+00 8.22703063e-01 -1.23470867e+00
1.84373319e-01 -4.14571673e-01 -4.80862260e-01 6.56656623e-01
3.72052819e-01 5.38453996e-01 7.46366739e-01 2.54489928e-01
-6.89056098e-01 -3.39949936e-01 -1.14689362e+00 1.07742630e-01
1.44485140e+00 -5.15491009e-01 -2.02507719e-01 5.25242746e-01
2.54063696e-01 -2.51481861e-01 -1.01802969e+00 5.21418631e-01
7.91780293e-01 -1.19604409e+00 1.91311741e+00 -7.85872281e-01
2.78908640e-01 -1.07011579e-01 -4.27310139e-01 -1.16960347e+00
-5.06795585e-01 2.26342659e-02 -1.91790506e-01 1.28459990e+00
2.00023815e-01 -1.04878046e-01 7.45416343e-01 6.31309569e-01
-5.79558648e-02 -9.06659544e-01 -8.90134037e-01 -5.52341998e-01
-3.49810034e-01 5.51434718e-02 4.10525531e-01 6.62136376e-01
-7.25860894e-01 4.42084938e-01 -8.23828697e-01 5.17290115e-01
1.43215311e+00 2.47554585e-01 8.31038773e-01 -1.77697456e+00
2.76538134e-02 -1.42788768e-01 -8.07225704e-01 -7.59060383e-01
4.40832287e-01 -6.42620742e-01 -3.70853186e-01 -1.46830368e+00
7.27417409e-01 -2.83551544e-01 -1.02987759e-01 3.20960522e-01
-4.40709442e-01 4.35469300e-01 1.05692707e-01 1.41511545e-01
-1.29465210e+00 4.16500241e-01 1.05244231e+00 -4.88771826e-01
7.09827483e-01 2.82425135e-01 -3.25042576e-01 9.22011435e-01
4.90623385e-01 -4.03860241e-01 -2.12438747e-01 -6.67559803e-02
1.39038801e-01 2.67944578e-02 1.02747333e+00 -1.08317518e+00
3.52763325e-01 5.87214716e-03 1.08073854e+00 -9.56342280e-01
9.79887426e-01 -1.10803998e+00 -2.22217124e-02 1.84088901e-01
-4.78522748e-01 1.42165601e-01 -1.28071189e-01 1.08781457e+00
4.11808155e-02 1.71777055e-01 6.52281642e-01 -2.57226825e-01
-1.14895856e+00 7.33239949e-01 4.21498120e-02 1.79246709e-01
1.23460829e+00 -2.39160106e-01 -2.23959506e-01 -4.62291837e-01
-9.87032831e-01 2.78697670e-01 4.40691471e-01 4.04949307e-01
7.21711218e-01 -1.61827481e+00 -1.05729723e+00 2.60229036e-02
7.17693806e-01 -4.45841461e-01 2.27861404e-01 7.31492281e-01
-6.70396164e-02 3.90499353e-01 -5.65882862e-01 -9.71185863e-01
-1.74460089e+00 1.03492808e+00 1.49905741e-01 1.90448817e-02
-3.80741984e-01 9.05760705e-01 5.05346596e-01 -2.46776626e-01
4.00534064e-01 5.31764448e-01 -5.97017705e-01 2.87414193e-01
4.59086388e-01 4.69218343e-01 -2.28325143e-01 -1.08987045e+00
-6.09692991e-01 9.24593925e-01 -1.29766449e-01 3.10953468e-01
1.03707159e+00 -3.67215276e-01 8.97212252e-02 2.16366127e-01
1.05073714e+00 -4.35072094e-01 -1.50809050e+00 -3.05898160e-01
-2.10410953e-01 -5.64733088e-01 -3.41844350e-01 -8.54694545e-01
-9.41610336e-01 9.58350003e-01 9.17004108e-01 -1.05152093e-01
9.30767775e-01 6.14735544e-01 4.11826223e-01 4.97413248e-01
5.71343720e-01 -1.16375172e+00 2.27223992e-01 5.63933291e-02
8.90165150e-01 -1.74563980e+00 3.17408293e-01 -7.95789003e-01
-9.80664551e-01 9.26007569e-01 8.20578575e-01 2.10778072e-01
4.41278696e-01 -1.51673511e-01 -1.16611362e-01 -3.36466223e-01
-5.62390327e-01 -3.77494484e-01 9.01478589e-01 1.02570093e+00
2.23483756e-01 2.38256112e-01 2.58694082e-01 3.57828945e-01
7.30836838e-02 -5.35583436e-01 -1.67315587e-01 2.93124318e-01
-4.18052524e-01 -9.13909912e-01 -5.62430441e-01 4.51504141e-01
-3.36284906e-01 2.15307493e-02 -5.13621986e-01 6.97878063e-01
4.01170254e-01 7.36867428e-01 -4.32434678e-03 -1.68959811e-01
3.50140721e-01 2.62666523e-01 6.61192894e-01 -4.16338220e-02
-4.39143091e-01 -3.07633847e-01 4.82730716e-01 -9.68828917e-01
-8.98170173e-01 -9.81710434e-01 -4.70524281e-01 -2.07893923e-01
2.77038306e-01 -3.05436105e-01 2.35948846e-01 9.27194118e-01
2.37147227e-01 2.09356472e-01 2.83957332e-01 -9.19460952e-01
-4.77386713e-01 -7.38155186e-01 -2.41787180e-01 1.03028548e+00
3.77163053e-01 -1.16121352e+00 1.42644763e-01 4.40860599e-01] | [14.360862731933594, 0.9878052473068237] |
149d5ff3-612b-49e6-9233-6cac0edc1652 | comparison-of-short-text-sentiment-analysis | null | null | https://aclanthology.org/W17-1411 | https://aclanthology.org/W17-1411.pdf | Comparison of Short-Text Sentiment Analysis Methods for Croatian | We focus on the task of supervised sentiment classification of short and informal texts in Croatian, using two simple yet effective methods: word embeddings and string kernels. We investigate whether word embeddings offer any advantage over corpus- and preprocessing-free string kernels, and how these compare to bag-of-words baselines. We conduct a comparison on three different datasets, using different preprocessing methods and kernel functions. Results show that, on two out of three datasets, word embeddings outperform string kernels, which in turn outperform word and n-gram bag-of-words baselines. | ['Jan {\\v{S}}najder', 'Leon Rotim'] | 2017-04-01 | null | null | null | ws-2017-4 | ['stock-price-prediction'] | ['time-series'] | [-1.90408528e-01 -3.97706389e-01 -2.96122640e-01 -3.16707641e-01
-8.89685333e-01 -8.30611527e-01 9.63448346e-01 8.70952785e-01
-1.22421896e+00 4.68684107e-01 7.41893649e-01 -7.09397554e-01
-8.54159892e-02 -8.75436246e-01 -3.72751392e-02 -6.81415915e-01
-3.54767106e-02 1.13838494e-01 -3.30777131e-02 -4.28125054e-01
5.77329576e-01 2.77086377e-01 -1.25796402e+00 1.00989625e-01
3.88127297e-01 7.54424334e-01 -4.47222084e-01 8.52764785e-01
-4.71568614e-01 3.62759918e-01 -6.87945247e-01 -8.25174510e-01
-7.51272887e-02 1.89429551e-01 -8.00848365e-01 -4.54915136e-01
4.21941817e-01 1.83591187e-01 -2.97306448e-01 8.82907450e-01
7.16345429e-01 4.22464699e-01 9.72821057e-01 -5.76960683e-01
-1.27934635e+00 5.81540883e-01 -1.61829650e-01 2.87278861e-01
5.33770740e-01 -1.31130904e-01 1.63605881e+00 -1.16688812e+00
3.51516515e-01 1.12883222e+00 1.14349210e+00 2.78339684e-01
-1.17914677e+00 -3.43214571e-01 5.41248545e-02 -1.19183801e-01
-8.61296415e-01 -9.49882343e-02 4.52030718e-01 -5.33108771e-01
1.59209585e+00 6.02740236e-03 3.55395138e-01 1.20154679e+00
4.61516798e-01 5.74413955e-01 1.15524554e+00 -8.48240376e-01
2.20196694e-01 5.55702567e-01 9.12204325e-01 4.48801160e-01
7.40676999e-01 -3.62223059e-01 -5.39891064e-01 -7.47496724e-01
1.48192078e-01 9.45568606e-02 -6.27105907e-02 -6.64239936e-03
-1.11944234e+00 1.44488394e+00 -2.19454825e-01 6.82108402e-01
-1.99385419e-01 2.65350133e-01 1.06082225e+00 4.26298827e-01
1.07157886e+00 7.44347036e-01 -1.09233844e+00 -5.07238388e-01
-5.96766531e-01 2.38127291e-01 8.79550397e-01 5.40761888e-01
6.05378389e-01 8.77702311e-02 -1.79900810e-01 1.08059072e+00
-2.10011285e-02 4.50257868e-01 1.03565240e+00 1.62884220e-01
3.87984365e-01 6.45526052e-01 -5.57623655e-02 -1.03242826e+00
-4.75009918e-01 2.14715570e-01 -3.37439179e-01 -5.44173550e-03
4.38692540e-01 -4.07952756e-01 -8.96354556e-01 1.22704089e+00
-1.47066802e-01 -1.61775231e-01 2.08647862e-01 3.82751524e-01
1.12480819e+00 7.48405337e-01 2.07438648e-01 8.15685652e-03
1.87052250e+00 -1.00721407e+00 -8.46904755e-01 -2.84665793e-01
1.31964898e+00 -9.92153287e-01 1.29641700e+00 2.91112155e-01
-7.22501278e-01 -3.92317474e-01 -8.42861116e-01 -1.64778218e-01
-1.42582047e+00 7.72487670e-02 9.89164948e-01 1.32232201e+00
-9.30605352e-01 5.04202425e-01 -4.69469279e-01 -4.68993157e-01
2.77376115e-01 1.73327148e-01 -6.84437811e-01 4.35766913e-02
-1.39640629e+00 1.16942096e+00 2.23722652e-01 -4.26043123e-01
1.15020769e-02 -9.66286361e-01 -1.39807153e+00 -8.69984552e-02
-3.73813212e-01 -7.35905841e-02 1.08038199e+00 -5.40502489e-01
-1.33575523e+00 1.06716430e+00 1.32345948e-02 -3.47378045e-01
-2.33768493e-01 -3.97294998e-01 -4.91203666e-01 -6.66523650e-02
-5.35018928e-02 -5.07494882e-02 6.43474758e-01 -6.41738892e-01
-4.95616376e-01 -1.81233406e-01 1.09724917e-01 8.68049450e-03
-1.14742160e+00 5.43310285e-01 -1.02955259e-01 -9.17405725e-01
-5.38761735e-01 -7.48321772e-01 -3.88475150e-01 -5.46411216e-01
6.80046752e-02 -7.82749593e-01 6.09336317e-01 -7.73950219e-01
1.55524373e+00 -2.19724226e+00 -1.63858831e-01 -4.74830111e-03
-8.26385319e-02 5.40467203e-01 -3.32988143e-01 9.04960215e-01
-2.79452413e-01 2.41240203e-01 -1.61964610e-01 -6.42212689e-01
1.71394482e-01 3.19300920e-01 -3.45196098e-01 6.73934937e-01
3.10347438e-01 1.04250848e+00 -9.79011118e-01 -3.46639901e-01
2.86338061e-01 7.20238805e-01 -3.21478605e-01 -2.11840332e-01
2.88741648e-01 -4.33272690e-01 -3.21602702e-01 5.64131618e-01
5.17193377e-01 1.85494944e-01 1.92090049e-01 5.91455474e-02
-1.52349904e-01 5.84245324e-01 -7.79775620e-01 1.49729419e+00
-8.40779245e-01 9.46537495e-01 -4.50536698e-01 -1.18937659e+00
8.77023041e-01 3.53216439e-01 9.91701782e-02 -4.21699852e-01
3.37037683e-01 2.59578049e-01 -3.10815483e-01 -4.24491435e-01
8.88738871e-01 -3.78659666e-01 -6.39780521e-01 4.83871192e-01
5.42369545e-01 -3.38246763e-01 2.84796149e-01 1.49056733e-01
1.29791927e+00 -4.09472704e-01 4.96830761e-01 -6.63327456e-01
5.62602580e-01 -3.22040170e-01 -1.45339966e-03 6.36950374e-01
-9.14546847e-02 5.61483324e-01 8.55804086e-01 -6.48115695e-01
-8.12542498e-01 -8.62990916e-01 -3.72730166e-01 1.42182457e+00
-4.06151354e-01 -9.22240555e-01 -4.00621831e-01 -8.47635746e-01
4.31267023e-01 8.32209647e-01 -1.12539291e+00 -7.23368153e-02
-4.66871262e-01 -1.15093338e+00 6.87604845e-01 8.97526979e-01
-3.73210549e-01 -1.21947110e+00 -4.63415533e-01 2.08418712e-01
4.31565464e-01 -1.06511641e+00 -4.57690746e-01 6.74071193e-01
-8.85803282e-01 -9.57092166e-01 -7.61733949e-01 -1.05349970e+00
3.83581489e-01 1.26233846e-01 1.21454251e+00 -8.40739533e-02
-3.67546469e-01 6.47569716e-01 -9.78078842e-01 -4.89281088e-01
-1.19851250e-02 2.48928979e-01 3.32880437e-01 -1.64386600e-01
1.16601253e+00 -2.42669702e-01 -2.20496148e-01 -2.81987876e-01
-1.02821684e+00 -9.27473426e-01 3.67489129e-01 1.04669344e+00
-1.07206412e-01 -2.35958055e-01 4.31382895e-01 -1.30092895e+00
1.49328065e+00 -3.13003480e-01 -3.39714557e-01 1.11584373e-01
-8.06806087e-01 -9.34336111e-02 7.77529478e-01 -7.59735703e-01
-5.15826523e-01 -2.97701538e-01 -2.18996048e-01 4.82164249e-02
5.48928641e-02 6.88576519e-01 2.41294011e-01 -2.03076638e-02
7.20823705e-01 8.34583342e-02 -2.17497468e-01 -3.47519159e-01
8.57962072e-01 9.50084031e-01 -1.60246283e-01 -5.11551380e-01
6.76065326e-01 2.17445806e-01 -5.48546910e-01 -1.18733454e+00
-7.42226064e-01 -1.09661698e+00 -6.06867015e-01 4.19518709e-01
1.04482019e+00 -7.24138618e-01 -3.09392154e-01 4.87524360e-01
-1.25918972e+00 -5.17411605e-02 -5.37392020e-01 6.43985331e-01
-2.88831443e-01 5.21096051e-01 -8.69046152e-01 -8.69253099e-01
-4.14574862e-01 -8.42811346e-01 9.81084645e-01 1.39903918e-01
-5.61777532e-01 -1.85974574e+00 5.24564266e-01 3.94995511e-02
4.29376304e-01 -8.00425410e-02 1.02079821e+00 -1.28624487e+00
7.86594033e-01 -9.01249349e-01 -2.84905165e-01 9.43358898e-01
3.82977664e-01 -1.31016046e-01 -1.12554646e+00 -3.24503213e-01
-1.48008436e-01 -4.49166358e-01 1.15033305e+00 2.61279196e-01
7.62942195e-01 -1.18675023e-01 -4.19801548e-02 3.83428693e-01
1.64993525e+00 -1.08609185e-01 6.86428607e-01 4.65970427e-01
7.06696987e-01 5.40018082e-01 2.38815814e-01 3.09929430e-01
6.32064641e-02 2.18996838e-01 -1.97861776e-01 1.18472404e-03
3.50870699e-01 4.87710945e-02 6.37211382e-01 1.28713071e+00
-8.32327008e-02 -2.56834179e-01 -9.34147537e-01 1.01634181e+00
-1.73508990e+00 -6.34805799e-01 -3.74648124e-01 1.90209758e+00
7.97842264e-01 2.95350939e-01 1.17627323e-01 3.10031116e-01
2.99721479e-01 6.22537017e-01 2.12640196e-01 -1.26278853e+00
-3.20270985e-01 1.05797410e+00 8.81585896e-01 5.14457166e-01
-1.35255527e+00 1.14838278e+00 7.55984592e+00 9.64330733e-01
-7.82747865e-01 3.55568469e-01 1.72218889e-01 1.86829537e-01
-4.58521336e-01 -9.59705487e-02 -8.86585414e-01 2.82840729e-01
1.31993997e+00 -1.49747580e-01 -1.84048533e-01 8.95149112e-01
-2.92210162e-01 -1.49130197e-02 -9.12747622e-01 1.02045035e+00
5.30708373e-01 -1.41071236e+00 5.72916567e-02 -2.06402987e-02
7.79588103e-01 -9.88111496e-02 4.75927293e-02 5.62751532e-01
2.16429219e-01 -1.11427546e+00 2.34324425e-01 1.86876357e-01
6.68038309e-01 -8.27361703e-01 1.42976379e+00 -3.32971573e-01
-9.22841847e-01 -3.99225345e-03 -5.82942605e-01 -4.54202443e-01
8.37080032e-02 9.25820589e-01 -1.45392910e-01 4.86733496e-01
7.07404435e-01 7.65620470e-01 -6.66096151e-01 4.83704150e-01
-1.95767373e-01 1.03878403e+00 6.80893753e-03 -6.34614408e-01
6.07640445e-01 -2.82427430e-01 -3.89144905e-02 2.09528828e+00
-1.47504970e-01 -7.23481327e-02 -2.17022896e-01 8.35661218e-02
6.31327555e-03 5.72138071e-01 -7.93674827e-01 -4.62943166e-01
2.96436511e-02 1.45159602e+00 -7.50776708e-01 -4.99855369e-01
-1.00777173e+00 1.00636899e+00 4.57729727e-01 1.92356244e-01
-5.68465352e-01 -1.20208216e+00 1.21258128e+00 -2.03294724e-01
5.41122496e-01 -5.03367722e-01 -3.41800630e-01 -1.13305891e+00
-3.18237435e-05 -7.77210236e-01 4.10190046e-01 -2.56170720e-01
-1.90016639e+00 3.58110785e-01 -1.31152734e-01 -6.97997510e-01
1.84349790e-01 -1.63978648e+00 -9.23925281e-01 9.43574488e-01
-1.57496512e+00 -1.01108336e+00 3.11168730e-01 1.56959504e-01
5.52606940e-01 -3.16578656e-01 1.55852759e+00 1.36912465e-01
-2.57064283e-01 8.02766442e-01 6.13404214e-01 4.94661659e-01
8.93589854e-01 -1.46318126e+00 7.64067411e-01 3.08981597e-01
4.06289011e-01 9.24116850e-01 4.62228179e-01 -3.38708967e-01
-1.39067090e+00 -6.98665619e-01 1.41021013e+00 -9.45781112e-01
1.39900553e+00 -8.07748914e-01 -7.69675434e-01 5.29637694e-01
3.69036853e-01 1.76020309e-01 1.37288642e+00 6.95044935e-01
-8.20297241e-01 2.71289527e-01 -8.63860428e-01 3.02139342e-01
4.56280082e-01 -8.85055661e-01 -1.05506313e+00 5.36282182e-01
8.48692298e-01 3.48054677e-01 -8.78535032e-01 1.26890883e-01
7.63950884e-01 -6.04333937e-01 8.62200320e-01 -1.24925411e+00
4.57050651e-01 3.38274211e-01 -4.26417112e-01 -1.60141242e+00
-2.92879969e-01 -3.86065066e-01 2.52296507e-01 1.31374729e+00
6.66076422e-01 -1.07820356e+00 6.43391132e-01 9.78007466e-02
1.69407234e-01 -9.63155687e-01 -8.17044199e-01 -1.01737070e+00
8.75554919e-01 -6.71479642e-01 2.06274137e-01 1.07392728e+00
5.18436491e-01 4.85003650e-01 5.91358431e-02 -4.36615855e-01
-8.30771681e-03 -3.68995339e-01 5.49375594e-01 -1.04907584e+00
6.54429272e-02 -7.14047313e-01 -1.11345565e+00 -4.97467637e-01
5.48377573e-01 -6.38123989e-01 -1.69232592e-01 -1.71521389e+00
-1.60007969e-01 -1.55547455e-01 -6.00293100e-01 4.01504099e-01
-4.51793641e-01 4.31929320e-01 -1.55433267e-01 -3.29840153e-01
-3.01477253e-01 5.04304469e-01 5.22356331e-01 -1.42048463e-01
3.35643589e-02 -3.13604236e-01 -7.15402424e-01 7.62218058e-01
9.12265182e-01 -3.86240780e-01 -1.61097556e-01 -2.45133922e-01
4.93632883e-01 -8.00516307e-01 -1.67307049e-01 -4.87829000e-01
-9.39207599e-02 -1.24981264e-02 3.61918211e-02 -1.86610386e-01
2.11728424e-01 -4.47164595e-01 -9.48464334e-01 2.21316218e-01
-2.31456757e-01 5.84942520e-01 5.91664195e-01 5.42389095e-01
-1.86693028e-01 -6.61696970e-01 4.90972281e-01 1.72792953e-02
-5.93983889e-01 -2.05363750e-01 -7.77998686e-01 2.51436859e-01
9.01665986e-01 -3.19505036e-01 -1.67832062e-01 -2.75826696e-02
-3.47167701e-01 -3.87735032e-02 2.70096362e-01 6.98065937e-01
4.87373352e-01 -1.23915863e+00 -7.04644561e-01 2.43562311e-01
6.70571268e-01 -6.95826232e-01 -2.62152582e-01 7.10851908e-01
-7.77971983e-01 6.19515896e-01 2.16505483e-01 7.07482174e-02
-1.26924753e+00 5.34430504e-01 -1.01778701e-01 -4.96590048e-01
-3.60725135e-01 1.11959100e+00 -7.86435679e-02 -9.91990447e-01
2.11778611e-01 -6.66326702e-01 -5.34757853e-01 8.30038190e-01
6.84512675e-01 1.85673982e-01 2.37897098e-01 -6.01350367e-01
-7.30293274e-01 9.74699616e-01 -2.00567305e-01 -1.46821275e-01
1.50815761e+00 4.20109242e-01 -2.49556273e-01 9.92063582e-01
1.59220850e+00 3.34933996e-01 -2.01251552e-01 -1.24656230e-01
2.45501027e-01 -3.73752862e-01 1.55862607e-02 -3.59481931e-01
-4.66320276e-01 8.93166006e-01 3.61115336e-01 6.96757495e-01
6.53844476e-01 -4.64526936e-02 7.97032595e-01 3.90064687e-01
-2.38439552e-02 -1.31389070e+00 -7.49103352e-02 8.58280063e-01
3.07789832e-01 -1.21554077e+00 2.20376641e-01 -1.37567148e-01
-6.46394670e-01 1.35647798e+00 6.70459960e-03 -5.68206012e-01
1.24097025e+00 2.66248018e-01 4.34650034e-01 -2.04582423e-01
-5.51336825e-01 -3.89820963e-01 3.90887022e-01 7.77022779e-01
8.86083007e-01 1.80963099e-01 -8.48252535e-01 6.13899887e-01
-3.46415490e-01 -4.41251338e-01 4.55919653e-01 1.04344904e+00
-3.93867731e-01 -1.32488072e+00 -8.80944952e-02 8.59069288e-01
-7.95061171e-01 -5.10772288e-01 -3.85707974e-01 7.99337149e-01
-1.81902885e-01 9.72007275e-01 8.85804445e-02 -3.65663320e-01
5.05292714e-01 5.29292703e-01 3.31606597e-01 -1.04497194e+00
-1.05599344e+00 -4.94219095e-01 4.32639718e-01 -9.64821428e-02
-4.30581778e-01 -5.62392652e-01 -9.84322667e-01 -3.98161471e-01
-9.24890876e-01 4.11358863e-01 9.06227291e-01 9.45119083e-01
-4.29010764e-02 2.11019188e-01 4.48083997e-01 -5.99830508e-01
-7.99890399e-01 -1.24257183e+00 -7.85209894e-01 6.45080507e-01
3.21590364e-01 -2.79853761e-01 -8.41416061e-01 -2.61593983e-02] | [10.484956741333008, 8.72552490234375] |
0bb67ba9-e245-4d1c-bb26-fa7a586ee782 | evaluation-phonemic-transcription-of-low | null | null | https://aclanthology.org/L18-1530 | https://aclanthology.org/L18-1530.pdf | Evaluation Phonemic Transcription of Low-Resource Tonal Languages for Language Documentation | null | ['Alexis Michaud', 'Steven Bird', 'Hilaria Cruz', 'Trevor Cohn', 'Oliver Adams', 'Graham Neubig'] | 2018-05-01 | evaluation-phonemic-transcription-of-low-1 | https://aclanthology.org/L18-1530 | https://aclanthology.org/L18-1530.pdf | lrec-2018-5 | ['acoustic-modelling'] | ['speech'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.327335357666016, 3.7193233966827393] |
fe5ba9f3-5f0f-4ad1-a5af-b631736dc58a | non-linear-prediction-with-lstm-recurrent | null | null | https://ieeexplore.ieee.org/abstract/document/7280757 | https://ieeexplore.ieee.org/abstract/document/7280757 | Non-linear prediction with LSTM recurrent neural networks for acoustic novelty detection | Acoustic novelty detection aims at identifying abnormal/novel acoustic signals which differ from the reference/normal data that the system was trained with. In this paper we present a novel approach based on non-linear predictive denoising autoencoders. In our approach, auditory spectral features of the next short-term frame are predicted from the previous frames by means of Long-Short Term Memory (LSTM) recurrent denoising autoencoders. We show that this yields an effective generative model for audio. The reconstruction error between the input and the output of the autoencoder is used as activation signal to detect novel events. The autoencoder is trained on a public database which contains recordings of typical in-home situations such as talking, watching television, playing and eating. The evaluation was performed on more than 260 different abnormal events. We compare results with state-of-the-art methods and we conclude that our novel approach significantly outperforms existing methods by achieving up to 94.4% F-Measure. | ['Erik Marchi ; Fabio Vesperini ; Felix Weninger ; Florian Eyben ; Stefano Squartini ; Björn Schuller'] | 2015-10-01 | null | null | null | 2015-international-joint-conference-on-neural | ['acoustic-novelty-detection'] | ['audio'] | [ 3.41564745e-01 3.12529095e-02 7.40362465e-01 -2.57167280e-01
-1.01319826e+00 -4.98357825e-02 3.25803399e-01 1.63290009e-01
-4.11930352e-01 5.19646108e-01 4.10121650e-01 3.05236578e-01
1.71038643e-01 -6.23095155e-01 -9.61188793e-01 -6.62507892e-01
-3.24328631e-01 -5.55502139e-02 3.72491241e-01 -1.34923220e-01
-7.02717975e-02 8.82546008e-02 -1.97006905e+00 8.93171549e-01
1.53543890e-01 1.10552382e+00 2.80443311e-01 1.18384385e+00
2.01411426e-01 7.31454253e-01 -8.69377196e-01 -1.66047081e-01
-8.36251378e-02 -7.25730717e-01 -5.20171046e-01 -2.16148645e-01
3.72501798e-02 -3.16862136e-01 -4.05593127e-01 9.79111731e-01
8.64143431e-01 5.88783801e-01 3.69865000e-01 -6.85259759e-01
-4.47815776e-01 6.72174454e-01 2.58123249e-01 7.19730079e-01
6.51589751e-01 -1.07777126e-01 7.06285357e-01 -1.07721221e+00
2.75554150e-01 8.80245507e-01 9.69725728e-01 5.32792032e-01
-1.13307464e+00 -5.05092740e-01 -2.14107856e-01 7.25420415e-01
-1.47070277e+00 -7.69126654e-01 1.01325667e+00 -2.09056124e-01
1.48178136e+00 2.10814044e-01 6.20020866e-01 1.59762907e+00
4.47265208e-01 5.16240299e-01 6.21795654e-01 -5.81679583e-01
4.48659390e-01 1.59530416e-02 -2.72643149e-01 4.25532192e-01
-5.24983168e-01 3.39034408e-01 -9.27985370e-01 -4.42110687e-01
2.36032739e-01 -1.15225799e-02 -1.88049123e-01 4.57700163e-01
-9.64576364e-01 6.28492057e-01 -8.79853368e-02 8.24146628e-01
-1.06951940e+00 1.49425551e-01 7.56831169e-01 5.65322220e-01
6.29496217e-01 2.27719042e-02 -4.07612205e-01 -4.97822315e-01
-1.02986968e+00 -7.82996230e-03 7.33747542e-01 2.63448298e-01
3.30392390e-01 7.58226693e-01 7.77008235e-02 9.17178094e-01
1.35914430e-01 2.31701866e-01 1.16919339e+00 -8.65015209e-01
-1.47275388e-01 -1.68486521e-01 -2.11729392e-01 -1.08380878e+00
-3.02238286e-01 -5.19519866e-01 -6.69154763e-01 -5.68555593e-02
-1.75631374e-01 -3.34155798e-01 -6.75530970e-01 1.49336314e+00
1.65726289e-01 8.78484011e-01 3.67650270e-01 3.13871175e-01
8.87453794e-01 1.17177904e+00 -1.19519278e-01 -5.26629508e-01
8.74532998e-01 -6.02059662e-01 -1.02525234e+00 3.41621153e-02
1.17508397e-01 -9.34840798e-01 8.18585157e-01 8.89158368e-01
-1.13219380e+00 -9.57646310e-01 -8.92166615e-01 4.38917607e-01
-3.39219153e-01 7.72170052e-02 -1.31933063e-01 5.88940501e-01
-1.02634490e+00 9.05353367e-01 -8.94137621e-01 -3.48266304e-01
-1.01077512e-01 5.68010546e-02 -2.40445822e-01 4.33508992e-01
-1.19729364e+00 5.77593148e-01 6.59096062e-01 -1.66790579e-02
-1.23194146e+00 -5.63493013e-01 -6.61638021e-01 2.39522696e-01
1.02754012e-01 -3.68316829e-01 1.61106408e+00 -1.10503733e+00
-1.75843406e+00 3.28022897e-01 -2.83012211e-01 -9.17727411e-01
-8.62699151e-02 -4.39780653e-01 -1.20736361e+00 3.10807884e-01
-2.86223590e-01 1.10740438e-01 1.36694407e+00 -8.89092565e-01
-6.25286043e-01 2.81987488e-02 -7.24160612e-01 -1.35194436e-01
-4.48454797e-01 1.60032675e-01 1.71550080e-01 -1.01330054e+00
5.99819086e-02 -6.31203711e-01 1.81396022e-01 -7.28598952e-01
-2.06574976e-01 -2.08339274e-01 9.65877414e-01 -1.01458025e+00
1.41701078e+00 -2.61241746e+00 -1.62422597e-01 1.86338469e-01
-2.71228045e-01 1.39555171e-01 1.33339927e-01 6.39544129e-01
-4.46528435e-01 -2.94783175e-01 -4.91216369e-02 -4.64778781e-01
-1.29184976e-01 3.57908070e-01 -6.40795231e-01 3.26562136e-01
9.65889078e-03 3.33382696e-01 -7.54687309e-01 -1.53476894e-01
2.78787404e-01 8.31431866e-01 -3.37846160e-01 4.05712098e-01
2.22057644e-02 2.88411140e-01 1.51542217e-01 2.88141370e-01
-4.09476869e-02 3.57730985e-01 -3.57547194e-01 -1.58386543e-01
-3.42372395e-02 5.40415525e-01 -1.14805043e+00 1.67036140e+00
-3.32121313e-01 6.62792146e-01 -4.22813654e-01 -1.03752875e+00
9.64268744e-01 9.83931124e-01 2.59265065e-01 -3.90218258e-01
1.82891577e-01 1.26338214e-01 -9.65220556e-02 -7.59613097e-01
2.94546366e-01 -4.14219975e-01 1.46699041e-01 1.48926914e-01
5.83470702e-01 1.93220079e-01 -7.51236156e-02 -2.69972682e-01
1.35891318e+00 -2.58977383e-01 2.00163916e-01 2.20040798e-01
5.55989444e-01 -5.39523900e-01 5.45887351e-01 9.00050104e-01
-7.08772019e-02 6.38624907e-01 4.92740124e-02 -4.97702092e-01
-9.44647908e-01 -1.32185841e+00 1.32482857e-01 1.16559112e+00
-7.87022889e-01 -5.79625428e-01 -6.93552196e-01 -2.45056927e-01
-3.95047784e-01 1.19682264e+00 -5.23873866e-01 -6.20374084e-01
-5.20421505e-01 -4.65286255e-01 9.15794134e-01 5.51500022e-01
1.18584797e-01 -1.36576545e+00 -5.78801215e-01 7.29679942e-01
-2.16238305e-01 -9.67290342e-01 -6.49909601e-02 4.45043743e-01
-8.94411027e-01 -4.72806603e-01 -4.56530243e-01 -7.62801647e-01
3.51017676e-02 -3.71441990e-01 9.07481968e-01 -5.15320659e-01
-2.55160540e-01 5.83543897e-01 -6.59237206e-01 -6.38405025e-01
-8.49427044e-01 -3.70674789e-01 4.35981274e-01 2.79898345e-01
4.60577756e-01 -1.03429437e+00 -3.76251370e-01 -1.30908191e-01
-9.42717552e-01 -6.19767904e-01 4.19970542e-01 8.66768599e-01
9.01132524e-01 6.39022887e-01 8.30842674e-01 -4.38988775e-01
7.95513868e-01 -5.87442696e-01 -6.51217774e-02 -8.46336633e-02
-3.11742902e-01 -1.85663134e-01 8.64257336e-01 -7.65451610e-01
-1.10250187e+00 1.52084723e-01 -7.03598857e-01 -7.44121969e-01
-6.06957018e-01 4.19932574e-01 2.51721919e-01 3.52717280e-01
8.09084237e-01 6.17249727e-01 -3.27811241e-01 -7.00619340e-01
-6.78951740e-02 7.05829322e-01 1.04682708e+00 -6.98030069e-02
4.13792491e-01 2.31388137e-01 -4.91969228e-01 -1.09852278e+00
-3.73529673e-01 -3.43280911e-01 -2.64716744e-01 -4.40520465e-01
7.35020280e-01 -7.24269569e-01 -5.27945340e-01 4.59343344e-01
-1.18306077e+00 9.15804207e-02 -7.35037923e-01 9.62794662e-01
-5.78488171e-01 3.61453742e-01 -8.17730665e-01 -1.13897324e+00
-5.47808886e-01 -7.20081389e-01 8.39061558e-01 -7.03714862e-02
-4.20255482e-01 -7.06999958e-01 5.10187030e-01 -1.97748587e-01
4.94841605e-01 1.90572530e-01 7.29215205e-01 -1.20193744e+00
2.62882859e-01 -4.96644139e-01 7.85669029e-01 1.03469622e+00
3.34636383e-02 -1.69304702e-02 -1.46378124e+00 -1.13848552e-01
6.84332192e-01 2.08053011e-02 8.67354631e-01 4.90921855e-01
1.24085140e+00 -4.39300567e-01 1.20292388e-01 2.03288451e-01
1.10801971e+00 5.91032445e-01 6.72013879e-01 -7.08837528e-03
2.19618559e-01 2.88166940e-01 1.86326534e-01 6.49945676e-01
-5.32766208e-02 2.76033074e-01 1.85128421e-01 3.80913079e-01
1.94371656e-01 -2.27879778e-01 8.96626174e-01 1.32986927e+00
-4.26767245e-02 -1.75691545e-01 -6.69934750e-01 7.05830634e-01
-1.53343880e+00 -1.41925883e+00 1.66440129e-01 2.17620969e+00
5.96637547e-01 3.60676318e-01 1.35239750e-01 7.28838503e-01
7.08390355e-01 -3.89936417e-02 -2.84258991e-01 -8.25724304e-01
4.30944040e-02 6.64954782e-01 -2.32348785e-01 2.54351765e-01
-1.11132109e+00 5.05989790e-01 6.69699907e+00 7.85249174e-01
-1.15940166e+00 4.42475855e-01 2.36150458e-01 -3.48142922e-01
1.84182286e-01 -5.08760571e-01 -2.55690426e-01 5.36239326e-01
1.92248380e+00 1.10228229e-02 4.37621981e-01 9.01022732e-01
2.96407849e-01 1.44195482e-02 -1.07300591e+00 1.17221761e+00
4.23741579e-01 -8.97136629e-01 -1.39096379e-01 -4.70379919e-01
3.88442427e-01 2.00585783e-01 2.62288213e-01 4.24058467e-01
-1.93594113e-01 -7.65950084e-01 6.30147278e-01 9.90943432e-01
3.65760446e-01 -1.03390729e+00 8.00143600e-01 3.70781213e-01
-9.24344122e-01 -1.55650198e-01 -3.63605261e-01 -4.45198081e-02
2.17002019e-01 8.24618578e-01 -1.07221997e+00 3.05549443e-01
1.24375093e+00 4.99635100e-01 -2.05740884e-01 9.05433178e-01
-2.42976216e-03 1.20300233e+00 -4.09553021e-01 2.09969729e-01
-9.98623436e-04 3.57824832e-01 9.09323335e-01 1.44560933e+00
6.58451259e-01 9.49815672e-04 -1.28717244e-01 6.29845440e-01
8.43165144e-02 1.49492711e-01 -8.12558711e-01 -1.40201107e-01
1.03430279e-01 7.51032114e-01 -3.32539678e-01 -4.64322865e-01
-4.04527426e-01 1.14398801e+00 -2.95934260e-01 2.26962239e-01
-6.39669001e-01 -4.87236798e-01 3.11841100e-01 -2.52460003e-01
6.56190097e-01 1.48100644e-01 5.11440516e-01 -8.74757528e-01
6.14565052e-02 -8.50188255e-01 4.62199330e-01 -9.59376931e-01
-1.45055211e+00 1.07610476e+00 -7.41528794e-02 -1.09726942e+00
-9.91531372e-01 -1.40406787e-01 -7.79107153e-01 6.60398543e-01
-8.87580633e-01 -5.37781119e-01 1.19118616e-01 6.51969433e-01
8.60159278e-01 -5.19964278e-01 1.37960768e+00 4.40932691e-01
-2.66215473e-01 3.99538338e-01 3.41417104e-01 -8.95234868e-02
5.89842081e-01 -1.04348111e+00 3.12871248e-01 9.79965568e-01
4.61041182e-01 2.99409449e-01 1.03438437e+00 -5.70343494e-01
-8.83442879e-01 -1.02876794e+00 1.06780815e+00 -8.98350477e-02
5.69400430e-01 -2.44625077e-01 -1.21971726e+00 6.74056828e-01
2.37948582e-01 3.02150734e-02 1.06062198e+00 -2.19460860e-01
-3.07976967e-03 -2.98869848e-01 -1.14683580e+00 1.03990979e-01
5.24810135e-01 -8.95874977e-01 -1.15394282e+00 1.35903610e-02
7.85090923e-01 -2.64918715e-01 -8.91713679e-01 3.72480989e-01
5.42715907e-01 -1.07457924e+00 9.17627275e-01 -4.69734460e-01
1.37463540e-01 -1.62247047e-01 -5.79908609e-01 -1.43189466e+00
-1.68305412e-01 -7.63456225e-01 -5.07084668e-01 1.07819879e+00
2.53632247e-01 -4.82662112e-01 4.41662222e-01 -1.25647724e-01
-4.34775472e-01 -5.22988379e-01 -1.20717657e+00 -7.72719741e-01
-4.71229792e-01 -1.04025006e+00 3.70612800e-01 5.48547208e-01
-2.22935379e-02 1.65390164e-01 -5.82034230e-01 3.46337706e-01
1.81816846e-01 -4.72114474e-01 2.60723025e-01 -1.05877757e+00
-7.72189140e-01 2.41952520e-02 -7.83035338e-01 -4.48045105e-01
1.79089904e-01 -5.83608747e-01 1.48083091e-01 -8.48151922e-01
-3.14392447e-01 6.40086174e-01 -6.35834992e-01 3.90367150e-01
2.39512220e-01 2.08205700e-01 -1.97381258e-01 -2.51521051e-01
-4.13954526e-01 6.65274560e-01 2.49631941e-01 -4.08802666e-02
-3.97625804e-01 2.48870462e-01 -1.02500692e-01 1.11638439e+00
9.15503919e-01 -7.93225348e-01 -2.15830058e-01 -3.69488895e-02
1.42344102e-01 9.19408575e-02 4.77909565e-01 -1.55887854e+00
3.23062539e-01 5.01241326e-01 6.92001641e-01 -7.48160303e-01
5.78690648e-01 -8.78413498e-01 4.97810245e-01 5.61078668e-01
-4.04526412e-01 3.01748425e-01 2.74635136e-01 8.80575061e-01
-5.37735939e-01 -4.91266608e-01 5.02951086e-01 -1.16266586e-01
-6.40895247e-01 -3.73577535e-01 -1.01390302e+00 -3.01138729e-01
6.35576010e-01 2.65180338e-02 1.79443479e-01 -6.82037532e-01
-1.35578191e+00 -7.37319171e-01 -2.08084047e-01 4.33567077e-01
1.04195642e+00 -1.19353580e+00 -7.51623929e-01 4.82521027e-01
-6.12677373e-02 -5.31109869e-01 4.82412398e-01 6.27350926e-01
-1.50460392e-01 1.44535705e-01 -3.51345576e-02 -4.61880147e-01
-1.38817000e+00 5.53785205e-01 4.01501507e-01 1.68198109e-01
-8.79915714e-01 7.56015480e-01 -2.65744835e-01 1.60166308e-01
4.96002585e-01 -4.74218339e-01 -1.80109397e-01 -1.51757389e-01
7.70261645e-01 6.76806211e-01 5.23787618e-01 -7.57753551e-01
-2.59550512e-01 -6.31106123e-02 3.54332780e-03 -3.09284300e-01
1.68393111e+00 -1.51023669e-02 1.07114939e-02 1.26499104e+00
1.38946807e+00 -2.33119689e-02 -5.87698162e-01 -2.73025811e-01
-7.05487207e-02 -5.27923480e-02 2.72978753e-01 -6.24809265e-01
-7.05052257e-01 7.19058275e-01 1.35410190e+00 4.94829983e-01
1.47543240e+00 1.23276301e-02 1.11479199e+00 6.48105264e-01
5.24548143e-02 -1.36235559e+00 1.90507904e-01 4.35621172e-01
9.06719089e-01 -7.90729880e-01 -5.33333898e-01 2.75114655e-01
-4.20254707e-01 1.26629734e+00 -3.90343484e-03 -4.27032560e-01
7.90525436e-01 3.09172362e-01 -1.35024995e-01 -5.20902127e-02
-9.26037133e-01 -1.43575519e-02 3.86647105e-01 5.66425741e-01
2.84380406e-01 -1.25500038e-01 5.77174723e-02 9.65517759e-01
-4.95480418e-01 -1.03261828e-01 3.75624329e-01 8.36189985e-01
-5.94556689e-01 -8.70070636e-01 -6.63615704e-01 6.25474513e-01
-9.48041975e-01 -1.24076962e-01 -1.98843449e-01 1.72765464e-01
2.52755105e-01 1.19026482e+00 8.94089937e-02 -7.14762747e-01
5.19339859e-01 8.11636865e-01 1.71010107e-01 -5.37517488e-01
-7.46818304e-01 4.65437829e-01 -6.09807447e-02 -5.53265572e-01
-3.61731261e-01 -8.50461066e-01 -1.12001920e+00 1.66101038e-01
-1.57648697e-01 1.80606455e-01 6.61457121e-01 8.86044919e-01
3.19389641e-01 9.41026628e-01 6.55015171e-01 -7.94544399e-01
-4.59165722e-01 -1.20256877e+00 -6.16656840e-01 3.00267339e-01
6.51353955e-01 -2.51584381e-01 -6.28160238e-01 5.44126213e-01] | [15.307180404663086, 5.584121227264404] |
c0ef951c-5fb4-4564-972a-a31dc89299d2 | the-exploitation-of-multiple-feature | 2112.07940 | null | https://arxiv.org/abs/2112.07940v1 | https://arxiv.org/pdf/2112.07940v1.pdf | The exploitation of Multiple Feature Extraction Techniques for Speaker Identification in Emotional States under Disguised Voices | Due to improvements in artificial intelligence, speaker identification (SI) technologies have brought a great direction and are now widely used in a variety of sectors. One of the most important components of SI is feature extraction, which has a substantial impact on the SI process and performance. As a result, numerous feature extraction strategies are thoroughly investigated, contrasted, and analyzed. This article exploits five distinct feature extraction methods for speaker identification in disguised voices under emotional environments. To evaluate this work significantly, three effects are used: high-pitched, low-pitched, and Electronic Voice Conversion (EVC). Experimental results reported that the concatenated Mel-Frequency Cepstral Coefficients (MFCCs), MFCCs-delta, and MFCCs-delta-delta is the best feature extraction method. | ['Ali Bou Nassif', 'Ismail Shahin', 'Noor Ahmad Al Hindawi'] | 2021-12-15 | null | null | null | null | ['speaker-identification'] | ['speech'] | [ 4.76749055e-02 -4.83793139e-01 2.45601614e-03 -1.36770919e-01
-4.27973986e-01 -5.63351274e-01 5.34905553e-01 -7.71863712e-03
-2.46597633e-01 6.22592092e-01 4.84822154e-01 2.72888411e-02
-1.28960848e-01 -2.39003390e-01 3.19451213e-01 -8.75917673e-01
-2.10009560e-01 -4.10084546e-01 -1.21793382e-01 -3.59351039e-01
4.09616113e-01 7.68048584e-01 -1.97342348e+00 1.33779412e-02
6.65892422e-01 1.01273143e+00 5.15182056e-02 7.37079620e-01
-1.82218328e-01 1.59303546e-01 -9.12082613e-01 -4.56962019e-01
-2.18206882e-01 -6.75376534e-01 -5.70596457e-01 -9.52142254e-02
-4.78993565e-01 1.59771442e-01 1.24777518e-01 1.12005568e+00
8.21713150e-01 3.26848030e-01 7.99115598e-01 -1.02742589e+00
-2.41700277e-01 6.03389382e-01 -1.14681892e-01 4.56019133e-01
7.09490836e-01 -1.62951946e-01 7.28395462e-01 -9.17049348e-01
1.86123818e-01 1.22240722e+00 6.02489352e-01 4.36578482e-01
-8.35270643e-01 -7.02089846e-01 -2.80190140e-01 4.47802484e-01
-1.49962032e+00 -5.85189641e-01 1.11358058e+00 -2.94117153e-01
1.17672300e+00 7.35162973e-01 7.09747255e-01 9.19408262e-01
2.19643578e-01 4.51968312e-01 1.24344969e+00 -7.45446563e-01
6.61124438e-02 5.46539962e-01 1.30363852e-01 1.42693341e-01
-2.47642905e-01 2.89299190e-01 -4.43955719e-01 -1.13039173e-01
3.51586044e-01 -3.41596782e-01 -3.97986889e-01 4.93732899e-01
-9.35502172e-01 6.87457144e-01 -1.98370144e-01 1.00738978e+00
-3.91616881e-01 -4.81950819e-01 4.99481678e-01 3.30797106e-01
3.19679439e-01 3.61523390e-01 -2.45185807e-01 -7.13766932e-01
-7.81354129e-01 -1.13229148e-01 8.54035079e-01 4.26973462e-01
1.33540064e-01 5.71429610e-01 5.31643257e-02 1.11689603e+00
1.97691008e-01 6.47596896e-01 9.12800491e-01 -5.17566621e-01
1.42806068e-01 2.12839589e-01 -3.03841718e-02 -1.07410252e+00
-2.14215294e-01 -4.66734618e-01 -7.30016708e-01 3.02694421e-02
-6.03315197e-02 -2.48619556e-01 -2.71220297e-01 1.43456066e+00
2.57669836e-01 2.28699427e-02 2.59088576e-01 7.31143057e-01
7.28507638e-01 7.73166239e-01 1.14482515e-01 -6.70634627e-01
1.55636752e+00 -3.96333933e-01 -1.33451462e+00 1.68880373e-01
-2.22986832e-01 -1.26734459e+00 1.08881044e+00 5.93566179e-01
-9.13501024e-01 -8.55308414e-01 -1.09469473e+00 3.45919400e-01
-6.39234006e-01 2.27507800e-01 4.00544792e-01 1.20917439e+00
-7.64737844e-01 4.41396683e-01 -5.28675616e-01 -1.52950391e-01
-3.39072078e-01 3.52636367e-01 -3.41299504e-01 7.51743019e-01
-1.33109665e+00 7.83190727e-01 1.95177391e-01 1.47546291e-01
-2.44199201e-01 -1.47201732e-01 -6.42421961e-01 2.27204934e-01
-1.45553663e-01 2.56647281e-02 1.21148574e+00 -1.12420619e+00
-1.97939372e+00 4.49236274e-01 -4.64120895e-01 -1.75188228e-01
5.91796190e-02 -1.61206976e-01 -1.24790001e+00 2.77926505e-01
-4.00940210e-01 1.09270744e-01 1.15571010e+00 -9.88956511e-01
-6.25286937e-01 -3.13495159e-01 -6.40441775e-01 3.30221593e-01
-6.27630532e-01 7.56274283e-01 1.53086692e-01 -6.90054953e-01
1.39349565e-01 -6.39744699e-01 2.84352630e-01 -9.36759472e-01
-3.94256979e-01 -3.88630748e-01 8.10667396e-01 -9.40399826e-01
1.68539083e+00 -2.60554957e+00 9.56731662e-02 1.75913557e-01
-3.53256613e-01 6.83708727e-01 4.96429443e-01 5.16846001e-01
-4.43707071e-02 2.42740944e-01 -1.24954525e-02 -2.64170766e-01
3.78677994e-02 -3.47556993e-02 -2.37477273e-01 4.92837876e-02
2.41808593e-01 2.62925655e-01 -5.05838394e-01 -6.28513575e-01
4.64326680e-01 9.52980936e-01 -1.65495381e-01 1.83343977e-01
5.28665125e-01 3.70288670e-01 -1.56602830e-01 5.44917643e-01
4.77294683e-01 8.36900055e-01 -7.29081780e-02 -3.11120570e-01
-6.38976455e-01 3.71632934e-01 -1.25241840e+00 7.18674004e-01
-4.62578118e-01 6.07863724e-01 2.25456625e-01 -6.52375638e-01
1.26058829e+00 9.49837089e-01 2.12393627e-01 -2.95704782e-01
5.04311919e-01 3.60926718e-01 2.62136787e-01 -7.50367939e-01
5.51568568e-01 -4.44211930e-01 4.16306593e-02 -1.41778678e-01
8.67178589e-02 -3.59296143e-01 6.49376214e-02 -3.08256090e-01
3.13475579e-01 -5.17542303e-01 8.36893022e-01 -1.27015635e-01
1.19469380e+00 -5.69538653e-01 6.07573032e-01 -4.54636328e-02
-6.85679317e-01 4.10963267e-01 1.78513691e-01 2.43140638e-01
-5.90908706e-01 -8.99257600e-01 -2.30924919e-01 8.50009382e-01
-2.99553692e-01 -1.99113980e-01 -9.06007409e-01 -1.20082587e-01
-3.12171489e-01 8.04440200e-01 -1.13765672e-01 -1.37410939e-01
-4.63896543e-01 -2.50743419e-01 8.20301116e-01 4.26298350e-01
4.20017570e-01 -1.34708345e+00 -4.38392550e-01 3.68432015e-01
-1.86895341e-01 -8.44995260e-01 -5.11304617e-01 3.88082922e-01
-7.42941380e-01 -5.72820902e-01 -6.42014921e-01 -9.41110134e-01
-3.69871710e-03 1.94611713e-01 6.23899043e-01 -3.69062603e-01
-2.84068763e-01 3.30843955e-01 -5.60486436e-01 -7.17858374e-01
-6.55915797e-01 -3.97453131e-03 4.02561665e-01 1.66741684e-01
6.11582339e-01 -6.49957597e-01 -2.97895998e-01 2.84257770e-01
-5.86037755e-01 -5.49673438e-01 3.87787402e-01 6.97280347e-01
3.42837334e-01 3.22919101e-01 1.08357525e+00 -3.39865237e-01
1.07120490e+00 -2.21312359e-01 -1.89198896e-01 1.15558356e-01
-3.18732232e-01 -2.82081723e-01 9.57048059e-01 -5.16345322e-01
-1.56330442e+00 -5.80324419e-02 -7.38023520e-01 -7.53810033e-02
-6.85734570e-01 4.54122037e-01 -5.60534239e-01 -4.23413515e-02
5.25333583e-01 4.74872917e-01 6.80068210e-02 -5.61797500e-01
2.34502628e-01 1.36246848e+00 5.83051324e-01 -1.56811148e-01
5.67616105e-01 4.18778919e-02 -4.68752116e-01 -1.47011745e+00
-2.32092559e-01 -6.69219971e-01 -5.99939764e-01 -4.50729430e-01
8.15564811e-01 -5.00754893e-01 -7.13871360e-01 6.73119903e-01
-8.13154697e-01 5.17165422e-01 -1.69515342e-01 1.04901433e+00
-2.16328532e-01 4.67495352e-01 -6.13919139e-01 -1.40533447e+00
-6.44228756e-01 -1.11137235e+00 4.30893511e-01 7.93386459e-01
-3.05330843e-01 -8.40618551e-01 -4.48297113e-02 1.90314963e-01
7.50238419e-01 1.21294700e-01 7.80439079e-01 -7.08370328e-01
4.71249551e-01 -2.18428642e-01 3.61402333e-01 8.36313844e-01
5.57297647e-01 3.39914620e-01 -1.23213506e+00 4.25980426e-02
4.12671685e-01 1.04850516e-01 2.44291425e-01 3.46646041e-01
7.92036712e-01 -3.12942833e-01 2.23782547e-02 4.36155915e-01
9.67878878e-01 1.02630246e+00 4.76640254e-01 -1.16687037e-01
1.66274622e-01 5.07519901e-01 5.64588308e-01 4.63655204e-01
-1.51948854e-01 4.31609243e-01 -7.06798956e-03 2.53207326e-01
-1.72973976e-01 -1.06379762e-01 6.24108672e-01 1.56561506e+00
-3.62942189e-01 -1.00229755e-01 -5.68263948e-01 4.35166985e-01
-8.89624655e-01 -1.20445347e+00 1.33904628e-03 2.10631585e+00
7.74231315e-01 1.17612407e-01 1.75817326e-01 1.13121533e+00
8.67612839e-01 1.61668494e-01 -1.90267101e-01 -1.00376964e+00
-1.87607855e-01 5.51773727e-01 -4.40464169e-02 5.10144174e-01
-1.07371247e+00 6.60788059e-01 6.27714014e+00 8.80533338e-01
-1.52569818e+00 -7.22278133e-02 3.96538168e-01 3.08992386e-01
-5.84409013e-02 -4.81404811e-01 -6.85623050e-01 6.07888401e-01
1.15171385e+00 -3.32859188e-01 6.02342546e-01 9.31568861e-01
3.34217548e-01 -2.04242975e-01 -5.77326059e-01 1.21062136e+00
4.48694304e-02 -5.08028626e-01 -1.52814925e-01 -6.70293048e-02
3.62214983e-01 -5.03438056e-01 3.14702392e-01 3.65181357e-01
-4.31786448e-01 -9.33310986e-01 6.76029444e-01 2.94195026e-01
7.88537741e-01 -1.18515551e+00 9.03332114e-01 1.11586526e-01
-1.48379433e+00 -5.15080877e-02 -1.81199193e-01 -1.51984662e-01
1.94505408e-01 5.01018584e-01 -9.78822947e-01 6.70910776e-01
4.40739214e-01 2.06944332e-01 -1.24792784e-01 1.05925357e+00
-2.82735676e-01 1.11102521e+00 -2.44677633e-01 -4.74244177e-01
-1.00518391e-01 -2.28489548e-01 7.88609743e-01 1.60107851e+00
3.97906870e-01 2.33713418e-01 -2.93928355e-01 3.36979598e-01
3.59507412e-01 4.61955577e-01 -3.03911656e-01 -3.56572539e-01
6.79277182e-01 1.24196267e+00 -7.20423996e-01 -1.03460267e-01
-2.23520353e-01 8.52344930e-01 -3.69118541e-01 2.20935583e-01
-7.67130733e-01 -8.61838043e-01 9.20772672e-01 -3.82792860e-01
6.21226840e-02 -3.09677064e-01 -1.89944550e-01 -6.72376394e-01
-2.02621892e-01 -9.24627781e-01 1.86336815e-01 -2.43005633e-01
-1.12673509e+00 8.72377813e-01 -6.39118627e-02 -1.19993031e+00
-5.59105992e-01 -4.48178262e-01 -7.39188612e-01 1.14261746e+00
-1.43867803e+00 -5.50059855e-01 -9.16067660e-02 6.20179772e-01
7.48567164e-01 -3.23797554e-01 1.13294792e+00 5.36140263e-01
-6.45309091e-01 6.18594766e-01 2.08311126e-01 9.42875072e-02
4.35459197e-01 -1.04712474e+00 2.09783137e-01 8.07107925e-01
1.59470588e-01 6.39041543e-01 7.11793840e-01 -1.78641528e-01
-1.30828512e+00 -4.77123886e-01 1.22517049e+00 2.85416096e-01
3.28887850e-01 -1.59792989e-01 -7.48872757e-01 1.47891462e-01
2.29419425e-01 -3.76383781e-01 9.74869728e-01 -9.57853273e-02
-8.17422569e-02 -2.01051876e-01 -1.31551409e+00 4.29972589e-01
4.85568613e-01 -8.58356059e-01 -9.38741565e-01 -4.33869034e-01
3.72426420e-01 -3.71086411e-02 -9.23403561e-01 1.27714857e-01
5.97341776e-01 -1.18420935e+00 9.28973138e-01 -2.27239415e-01
-3.39448333e-01 -2.33185440e-01 -1.26450703e-01 -1.44524884e+00
-2.66914457e-01 -7.41670430e-01 1.66088138e-02 1.72403181e+00
2.63514698e-01 -9.36018527e-01 1.96527004e-01 2.32503727e-01
-1.34059042e-01 -3.28079045e-01 -9.51076090e-01 -8.40429187e-01
-3.31925988e-01 -4.47131425e-01 6.14931703e-01 7.62768090e-01
5.35217166e-01 3.53825301e-01 -3.18998545e-01 1.14317108e-02
1.87607065e-01 -1.75392196e-01 2.11988717e-01 -1.27670670e+00
-4.77307253e-02 -7.51266420e-01 -5.13882339e-01 -5.00852823e-01
-1.10896446e-01 -5.33890963e-01 -1.14171421e-02 -1.07359028e+00
-2.89094210e-01 6.71871156e-02 -5.20695090e-01 3.16082989e-03
-3.60774934e-01 4.11279537e-02 2.66045719e-01 -3.52580287e-02
1.85321137e-01 6.17253602e-01 7.43335664e-01 5.86353615e-02
-5.05629897e-01 5.02937615e-01 -3.06236356e-01 7.15068281e-01
1.22967291e+00 -1.81222141e-01 -3.83335471e-01 3.93806189e-01
-3.69929224e-01 9.55017880e-02 -3.29945862e-01 -1.18543375e+00
1.88795663e-02 -3.26110679e-03 2.71819383e-01 -6.32241070e-01
6.52952194e-01 -7.30448008e-01 3.32406193e-01 4.94871199e-01
-1.96617797e-01 -1.41053079e-02 1.45318642e-01 1.07697643e-01
-7.17935205e-01 -4.25456166e-01 8.94718707e-01 8.90916288e-02
-6.19080186e-01 -4.09366548e-01 -8.38660538e-01 -3.24677229e-01
8.90171409e-01 -3.76710445e-01 1.04924686e-01 -3.95870835e-01
-7.36432612e-01 -5.37586927e-01 -7.08567798e-02 3.79534394e-01
6.25741184e-01 -1.07335532e+00 -4.85186189e-01 5.04795730e-01
-2.39565283e-01 -7.26081312e-01 3.95405650e-01 8.89072299e-01
-2.90612489e-01 6.09830737e-01 -3.85310978e-01 -2.04715118e-01
-1.81870162e+00 3.21442634e-01 1.53062075e-01 8.32342207e-02
-7.26584643e-02 7.76337802e-01 -3.80304903e-01 6.19739592e-02
5.34876168e-01 -4.00108278e-01 -8.29447210e-01 3.92479300e-01
7.52234995e-01 7.46000290e-01 1.82635292e-01 -1.05483639e+00
-4.12569314e-01 5.55376053e-01 2.34442919e-01 -4.10347760e-01
9.18244362e-01 -3.22237611e-01 -1.76459476e-02 7.00901866e-01
1.13862181e+00 4.97222960e-01 -6.88903749e-01 1.52490556e-01
5.95565811e-02 -3.63925457e-01 1.28173918e-01 -8.83969367e-01
-6.60232246e-01 1.03793025e+00 7.48185754e-01 8.24604094e-01
1.42147744e+00 -4.45250601e-01 8.06310833e-01 7.36206071e-03
3.40396404e-01 -1.30203474e+00 -2.16296479e-01 4.30004120e-01
8.78078818e-01 -8.87499452e-01 -3.84533286e-01 -5.99998295e-01
-8.88246834e-01 1.28794861e+00 9.46398228e-02 2.83921510e-01
9.28448856e-01 4.39619720e-01 2.69015372e-01 4.30327594e-01
-2.68794179e-01 -4.13806051e-01 3.98717284e-01 7.21461475e-01
8.88598979e-01 3.86784762e-01 -7.17331588e-01 1.06759775e+00
-6.97106361e-01 -4.28683311e-01 1.49167418e-01 7.92832077e-01
-6.97850943e-01 -1.13219690e+00 -9.07093465e-01 1.82016924e-01
-8.18530798e-01 6.08450845e-02 -5.89550912e-01 5.20567596e-01
5.90324700e-02 1.64167976e+00 -3.04715246e-01 -7.47790575e-01
3.66522819e-01 6.13878369e-01 1.18297771e-01 -1.73101231e-01
-1.21580684e+00 3.15277129e-01 -5.60036721e-03 -5.77683076e-02
-4.23880070e-01 -8.33296955e-01 -1.40027547e+00 -1.51045769e-01
-6.44183755e-01 6.12766743e-01 1.33262801e+00 6.25716507e-01
2.13335276e-01 6.13782525e-01 1.10591757e+00 -7.54923165e-01
-5.84519625e-01 -1.30061030e+00 -6.75299823e-01 2.87453622e-01
1.82314456e-01 -5.90538085e-01 -7.01523960e-01 1.32843956e-01] | [14.703463554382324, 5.909375190734863] |
01f21440-5c1d-4fac-9e5f-7144f1f94e10 | rnn-sm-fast-steganalysis-of-voip-streams | null | null | https://github.com/fjxmlzn/RNN-SM | http://www.andrew.cmu.edu/user/zinanl/publications/rnn-sm.pdf | RNN-SM: Fast Steganalysis of VoIP Streams Using Recurrent Neural Network | Quantization index modulation (QIM) steganography makes it possible to hide secret information in voice-over IP (VoIP) streams, which could be utilized by unauthorized entities to set up covert channels for malicious purposes. Detecting short QIM steganography samples, as is required by real circumstances, remains an unsolved challenge. In this paper, we propose an effective online steganalysis method to detect QIM steganography. We find four strong codeword correlation patterns in VoIP streams, which will be distorted after embedding with hidden data. To extract those correlation features, we propose the codeword correlation model, which is based on recurrent neural network (RNN). Furthermore, we propose the feature classification model to classify those correlation features into cover speech and stego speech categories. The whole RNN-based steganalysis model (RNN-SM) is trained in a supervised learning framework. Experiments show that on full embedding rate samples, RNN-SM is of high detection accuracy, which remains over 90% even when the sample is as short as 0.1 s, and is significantly higher than other state-of-the-art methods. For the challenging task of conducting steganalysis towards low embedding rate samples, RNN-SM also achieves a high accuracy. The average testing time for each sample is below 0.15% of sample length. These clues show that RNN-SM meets the short sample detection demand and is a state-of-the-art algorithm for online VoIP steganalysis. | ['Yongfeng Huang', 'Jilong Wang', 'Zinan Lin'] | 2018-02-15 | null | null | null | ieee-transactions-on-information-forensics | ['steganalysis'] | ['computer-vision'] | [ 6.73633099e-01 8.36603194e-02 -3.95406723e-01 3.71294141e-01
-7.20044434e-01 -1.90516040e-01 2.52567798e-01 -7.64182031e-01
2.56181248e-02 2.80078679e-01 -1.17888898e-01 -7.66431034e-01
3.52304727e-01 -6.16941690e-01 -3.58917445e-01 -8.35414052e-01
-3.63514572e-01 6.24190420e-02 2.42991909e-01 -2.42018521e-01
3.00901145e-01 2.02102914e-01 -1.36946189e+00 2.92640150e-01
5.29822052e-01 9.14649010e-01 1.82054266e-01 1.09117079e+00
-3.98859642e-02 7.05812097e-01 -9.73503649e-01 -1.24714367e-01
2.25444913e-01 -1.10770369e+00 -4.44216073e-01 2.62963682e-01
-3.05016398e-01 -2.60694236e-01 -7.79829085e-01 1.30019438e+00
5.22212863e-01 -3.91152859e-01 3.54716599e-01 -1.39377952e+00
-1.11793220e-01 7.17852473e-01 -1.79809228e-01 2.89430350e-01
-6.06286619e-03 5.16173601e-01 8.87032807e-01 -6.28014266e-01
4.08279836e-01 1.20934296e+00 4.21187669e-01 7.04776883e-01
-5.97513080e-01 -1.16306734e+00 -4.33429688e-01 4.33250874e-01
-1.46372843e+00 -7.23888099e-01 1.07399559e+00 -7.46064484e-02
7.54144132e-01 5.82587481e-01 7.94387221e-01 8.77914131e-01
2.62047082e-01 7.53614664e-01 7.93733239e-01 -4.03154463e-01
-4.08396646e-02 2.05030501e-01 -7.04924822e-01 6.39377475e-01
2.66712844e-01 3.28033656e-01 -3.73600185e-01 -3.77058908e-02
5.00117660e-01 5.74955018e-03 -5.40750325e-01 2.05786869e-01
-1.27956331e+00 1.10858119e+00 -1.70474440e-01 5.43429792e-01
-5.45288771e-02 3.95838916e-01 4.95071560e-01 8.54403794e-01
2.76823699e-01 1.69226885e-01 -1.09378889e-01 -3.42219949e-01
-1.04661405e+00 -3.41776252e-01 9.28880036e-01 8.52950871e-01
4.27502632e-01 6.18989587e-01 2.81330228e-01 5.05692899e-01
6.18293524e-01 1.01449013e+00 6.30645633e-01 -6.12559557e-01
7.35677898e-01 7.06312954e-02 -4.63521987e-01 -1.26704860e+00
8.26047435e-02 -5.38511336e-01 -1.04898381e+00 4.69065867e-02
1.29803821e-01 -1.40418708e-01 -6.36501431e-01 1.20869064e+00
1.07402653e-01 5.75388253e-01 2.99833268e-01 5.95891356e-01
6.67590618e-01 9.22568619e-01 -6.26739502e-01 -5.41450858e-01
1.35152888e+00 -7.72609413e-01 -1.00046301e+00 -1.24837741e-01
9.77047145e-01 -8.34926963e-01 5.73763371e-01 3.80850405e-01
-7.26151407e-01 -2.33134240e-01 -1.34673500e+00 7.23874807e-01
-3.01237721e-02 -3.09508562e-01 1.45846158e-01 1.32911909e+00
-8.08704495e-01 4.93278146e-01 -2.57466167e-01 1.74286425e-01
4.86984164e-01 5.00872016e-01 -1.10357657e-01 -8.85687545e-02
-1.68093848e+00 3.15018326e-01 3.86123449e-01 -1.08891157e-02
-1.01711059e+00 -1.42231852e-01 -9.45999146e-01 9.36336443e-02
2.46988207e-01 8.99130318e-05 8.60258758e-01 -8.20969343e-01
-1.61915362e+00 5.66184878e-01 -2.68028021e-01 -4.76776332e-01
3.16587627e-01 7.94448733e-01 -1.25740254e+00 5.96615374e-01
-2.49265850e-01 1.83543548e-01 1.88353229e+00 -1.00713098e+00
-6.38111055e-01 6.65500090e-02 -4.50435758e-01 -2.44838193e-01
-3.90379339e-01 1.72898784e-01 -2.51968473e-01 -5.61176538e-01
3.97201627e-01 -1.13538265e+00 -5.89679033e-02 -5.57116985e-01
-4.47149515e-01 1.26496255e-01 1.52841163e+00 -6.77071214e-01
1.62729156e+00 -2.28072476e+00 -2.32202500e-01 7.48471200e-01
3.32086831e-01 8.94308388e-01 -1.01401523e-01 4.87879217e-01
-1.02673983e-02 6.41144872e-01 -2.02334389e-01 -1.17316239e-01
-1.64041430e-01 1.07035749e-02 -3.47756267e-01 6.40602112e-01
-1.17806591e-01 8.71726155e-01 -8.74015272e-01 -4.52338099e-01
2.91386604e-01 4.23169136e-01 -3.88469905e-01 2.94021107e-02
2.11152896e-01 5.44585764e-01 -1.63926154e-01 7.16355145e-01
7.79982746e-01 -3.54815692e-01 5.59248984e-01 1.10221677e-01
2.60788321e-01 5.54553688e-01 -1.04189980e+00 8.66438806e-01
-4.13027763e-01 1.09126568e+00 -1.81993634e-01 -9.03695405e-01
1.13934290e+00 8.28415036e-01 3.61394882e-01 -6.16380513e-01
3.61280680e-01 6.55169010e-01 3.37507278e-01 -6.61632061e-01
2.44934678e-01 -1.35652229e-01 -6.72159046e-02 5.23646772e-01
-1.87217012e-01 9.40677598e-02 -2.24161431e-01 -2.04226106e-01
1.16634536e+00 -7.90789723e-01 2.22839624e-01 2.43506134e-01
9.28502023e-01 -5.98970890e-01 5.77037096e-01 5.94893873e-01
-2.48376682e-01 3.71311128e-01 6.01856828e-01 -4.58265580e-02
-1.25375891e+00 -1.98135272e-01 2.10862339e-01 2.74957597e-01
2.83740371e-01 -4.62052763e-01 -7.88868666e-01 -6.73271835e-01
-3.12403232e-01 2.89447814e-01 1.50797209e-02 -3.98201495e-01
-7.51560688e-01 -3.56468618e-01 1.15529931e+00 -3.00276190e-01
8.35759997e-01 -1.33686221e+00 -2.51246899e-01 4.43442255e-01
-4.11426693e-01 -1.45218647e+00 -7.21024215e-01 -4.07070339e-01
-9.00054514e-01 -9.81951892e-01 -5.16769528e-01 -1.00349927e+00
4.15380269e-01 5.67994177e-01 5.22446871e-01 4.41758096e-01
1.00538410e-01 -9.25422385e-02 -7.07739532e-01 -9.89445448e-02
-1.24254429e+00 2.04218805e-01 -6.76885247e-03 3.62748623e-01
5.31244814e-01 -5.66622615e-01 -5.25460541e-01 4.96018738e-01
-9.08799291e-01 -2.78893799e-01 6.91609561e-01 8.29729855e-01
9.19240341e-02 5.26993692e-01 5.42008400e-01 -7.31071055e-01
2.81815886e-01 -6.34033084e-01 -4.30452257e-01 -1.86568841e-01
-7.38322675e-01 -1.06302768e-01 6.48277760e-01 -6.32842541e-01
1.80329885e-02 -4.16665047e-01 -5.24908721e-01 -6.57594383e-01
2.56564766e-01 3.19888264e-01 -4.49211568e-01 -5.21852016e-01
1.08840205e-01 8.81187558e-01 4.93735224e-01 -1.49979398e-01
-1.96750343e-01 1.37190306e+00 -1.48804009e-01 5.15330970e-01
1.26385021e+00 5.20919859e-01 4.97215837e-02 -1.33990157e+00
7.39768893e-02 -5.25622189e-01 -3.75238955e-02 -2.43658200e-01
3.23383957e-01 -8.16718400e-01 -1.04420352e+00 6.49941504e-01
-1.15903640e+00 -5.96285798e-02 4.47194166e-02 6.08235478e-01
-3.24022025e-01 9.33196604e-01 -5.25606394e-01 -1.10275233e+00
-4.36978936e-01 -1.33449817e+00 7.25100696e-01 -4.72483277e-01
-1.34240892e-02 -1.01692486e+00 -4.10345793e-01 4.50435221e-01
4.18824524e-01 1.58952326e-01 5.00073910e-01 -8.24056506e-01
-7.72648454e-01 -3.94358277e-01 1.62862554e-01 6.31552994e-01
1.33816928e-01 -3.99100929e-01 -8.64257216e-01 -6.67887688e-01
4.45382386e-01 3.05294901e-01 8.97755384e-01 1.24415897e-01
1.07487464e+00 -8.90685737e-01 -2.16615543e-01 8.61758828e-01
1.10880399e+00 5.88127792e-01 1.29217279e+00 1.24348067e-01
8.05645525e-01 3.76726896e-01 6.44895017e-01 5.46469927e-01
3.30462568e-02 4.17969793e-01 6.59761965e-01 1.35954112e-01
-4.57498580e-02 -3.60639423e-01 7.67876506e-01 1.52184010e+00
2.29634166e-01 -7.04751194e-01 -5.12836277e-01 5.22401690e-01
-1.36423600e+00 -1.34903955e+00 -3.37124139e-01 2.13103056e+00
5.15781045e-01 3.95333260e-01 2.70736441e-02 1.12830031e+00
1.08247256e+00 8.23796749e-01 -2.81969726e-01 -6.14101112e-01
3.18995528e-02 8.50571021e-02 9.02520180e-01 6.10727251e-01
-9.64618504e-01 8.51598322e-01 5.52695799e+00 1.46121418e+00
-1.37099719e+00 3.23855251e-01 3.49626511e-01 4.34975356e-01
-5.04945576e-01 7.95841590e-02 -8.52221370e-01 1.03212631e+00
1.40274227e+00 3.21051359e-01 6.27214491e-01 5.38784087e-01
3.88718992e-01 5.34817457e-01 -3.76679301e-01 1.41650212e+00
3.55625749e-01 -1.28431392e+00 -1.52373314e-02 6.45544708e-01
4.78716850e-01 -1.67860538e-01 2.65868902e-01 8.34534690e-02
-3.91241610e-01 -1.11036611e+00 1.21360779e-01 -4.07724492e-02
1.42366588e+00 -9.07473087e-01 8.79540801e-01 3.93204749e-01
-1.30985487e+00 -1.77027792e-01 -3.39230895e-01 2.89024711e-01
2.61290580e-01 7.21870601e-01 -1.11304140e+00 1.89643487e-01
1.93317324e-01 7.48522460e-01 1.84841082e-01 5.87724745e-01
-2.78873503e-01 1.19001853e+00 -1.59064069e-01 -2.74723977e-01
3.92455995e-01 5.68945557e-02 1.17887080e+00 1.21861804e+00
7.96241343e-01 -2.26807773e-01 -2.15424120e-01 3.56568187e-01
-2.54226923e-01 6.52961656e-02 -9.18163180e-01 -3.90956998e-01
7.04700410e-01 7.72290409e-01 -3.57158720e-01 -4.34398443e-01
-6.62953258e-02 1.10770929e+00 -7.36034393e-01 1.68463081e-01
-6.65897906e-01 -8.34938765e-01 7.17684150e-01 1.86867684e-01
8.53309751e-01 -2.11930767e-01 3.71551700e-02 -1.19818413e+00
-4.25349325e-02 -1.41653299e+00 -1.55262306e-01 1.37744043e-02
-4.79003012e-01 3.17662925e-01 -7.44370878e-01 -2.07611275e+00
-5.43933988e-01 -2.66490906e-01 -5.42257905e-01 4.46639121e-01
-1.94660819e+00 -7.34205782e-01 -4.60324110e-03 5.46423018e-01
5.81848383e-01 -7.39459455e-01 7.74120748e-01 4.23171610e-01
-3.10254335e-01 9.69651818e-01 4.25784886e-01 4.37954903e-01
1.11205019e-01 -4.48184788e-01 7.63053536e-01 8.25210392e-01
1.07598878e-01 1.63280368e-01 6.19774520e-01 -6.92560017e-01
-1.52926886e+00 -1.14294112e+00 1.26468611e+00 1.80813998e-01
3.42123777e-01 -3.14312339e-01 -7.50902891e-01 2.78735697e-01
-1.59879863e-01 5.95325604e-02 8.03585708e-01 -7.71709800e-01
-2.95903444e-01 -3.52077372e-02 -1.36891484e+00 3.67267549e-01
8.77503157e-01 -7.94467270e-01 -7.29586184e-02 2.12798283e-01
9.05681193e-01 -1.67155772e-01 -7.33989716e-01 1.82662159e-01
6.31531239e-01 -8.79407108e-01 7.53448188e-01 -9.79395211e-02
1.78143218e-01 -3.90233010e-01 -2.21773997e-01 -9.43406641e-01
1.17570333e-01 -1.44202280e+00 -6.62828922e-01 9.52080309e-01
3.44763428e-01 -9.64721084e-01 9.31681395e-01 -6.01763427e-01
2.38826975e-01 -4.49836880e-01 -1.26729405e+00 -1.18013489e+00
-4.70561773e-01 -7.97392607e-01 8.47120941e-01 9.58217502e-01
1.27683893e-01 -2.17760541e-02 -1.08528590e+00 3.44428599e-01
9.32707727e-01 -3.92869115e-01 8.28336537e-01 -9.49611723e-01
-3.57965201e-01 -2.63975143e-01 -9.31941152e-01 -1.37922132e+00
1.38366163e-01 -8.69441271e-01 -2.34585270e-01 -7.64613330e-01
-4.34074044e-01 -4.41959888e-01 -2.34453291e-01 -5.47982231e-02
1.59676149e-01 6.47565901e-01 3.14683080e-01 5.80417275e-01
-2.83517689e-01 3.61796170e-01 1.35570598e+00 -3.23544055e-01
-3.04564357e-01 4.08366740e-01 -4.13252205e-01 4.96799886e-01
1.15910149e+00 -8.78426135e-01 -2.90830880e-01 2.55997688e-01
3.40360999e-02 4.68697518e-01 6.17554076e-02 -1.14347315e+00
3.65373977e-02 5.43650612e-02 -2.63787508e-01 -6.14258289e-01
3.73033434e-01 -9.98985231e-01 1.52969331e-01 1.18822312e+00
-1.39861356e-03 -5.31923294e-01 -3.49553704e-01 5.19124508e-01
-3.03967506e-01 -3.83704245e-01 7.45002985e-01 4.34779469e-03
-5.88639021e-01 4.43347424e-01 -6.86302841e-01 1.25312582e-01
8.01627398e-01 -7.13867307e-01 1.62498444e-01 -9.33468401e-01
-3.05314302e-01 -1.30470246e-01 1.31484672e-01 2.65137851e-01
1.20176649e+00 -1.25447786e+00 -8.13473403e-01 7.36507654e-01
-1.12539474e-02 -6.52834713e-01 1.81543946e-01 8.27077568e-01
-6.32786453e-01 5.60924828e-01 3.41056049e-01 -8.23879540e-01
-1.60645199e+00 4.90788311e-01 7.08733574e-02 -3.03569317e-01
-6.87015057e-01 3.70341480e-01 -4.10850763e-01 -8.87813196e-02
-3.06259040e-02 -2.54772939e-02 -3.69923770e-01 -8.91050249e-02
8.25386107e-01 4.82559800e-01 -8.80750194e-02 -9.21511471e-01
-1.26875445e-01 6.65573835e-01 2.20935330e-01 -1.04572184e-01
8.35218847e-01 -4.56730515e-01 4.15152684e-02 1.93825588e-01
1.83636868e+00 1.54861376e-01 -7.48675048e-01 -2.25612983e-01
-3.04395109e-01 -8.00687551e-01 1.40952663e-02 1.00209743e-01
-1.19111884e+00 1.08296800e+00 7.06537724e-01 4.75204676e-01
7.58284450e-01 -2.82463223e-01 1.64432883e+00 2.19443530e-01
3.95078003e-01 -8.08806241e-01 1.04905479e-01 5.08211672e-01
2.81270206e-01 -1.21322429e+00 -3.97736788e-01 -4.82948840e-01
-5.54008365e-01 1.26398873e+00 -2.58797228e-01 -1.63840484e-02
7.59699583e-01 -5.83046935e-02 -7.41575658e-02 -8.69956836e-02
-5.88692129e-01 -1.84927151e-01 4.62146588e-02 6.19426012e-01
-2.48523831e-01 1.50588170e-01 -2.74806708e-01 -1.67701408e-01
-3.50878835e-01 -2.73239583e-01 7.08648264e-01 6.62048459e-01
-8.18819225e-01 -1.11590183e+00 -5.70274651e-01 3.97161514e-01
-8.02287757e-01 -2.63184369e-01 2.32968241e-01 5.22394657e-01
-4.06097434e-02 1.38706350e+00 -7.35305026e-02 -1.09338498e+00
-1.83410272e-01 -7.27701113e-02 -1.28957361e-01 -2.44913980e-01
-2.54369467e-01 1.40110284e-01 5.91012277e-03 -2.60210186e-01
-2.94369668e-01 -4.33365375e-01 -1.19229329e+00 -7.83858240e-01
-6.62577093e-01 2.52437770e-01 8.08109403e-01 1.03172207e+00
2.62566984e-01 4.85907376e-01 1.62453735e+00 -5.18648922e-01
-4.59908724e-01 -8.94964695e-01 -8.56607676e-01 1.30616322e-01
8.34126711e-01 9.98589258e-06 -1.08358335e+00 -2.32036024e-01] | [4.299074172973633, 8.055582046508789] |
39d51d4c-3750-4133-a029-ac0ecf260751 | jointly-modeling-motion-and-appearance-cues | 2007.02041 | null | https://arxiv.org/abs/2007.02041v1 | https://arxiv.org/pdf/2007.02041v1.pdf | Jointly Modeling Motion and Appearance Cues for Robust RGB-T Tracking | In this study, we propose a novel RGB-T tracking framework by jointly modeling both appearance and motion cues. First, to obtain a robust appearance model, we develop a novel late fusion method to infer the fusion weight maps of both RGB and thermal (T) modalities. The fusion weights are determined by using offline-trained global and local multimodal fusion networks, and then adopted to linearly combine the response maps of RGB and T modalities. Second, when the appearance cue is unreliable, we comprehensively take motion cues, i.e., target and camera motions, into account to make the tracker robust. We further propose a tracker switcher to switch the appearance and motion trackers flexibly. Numerous results on three recent RGB-T tracking datasets show that the proposed tracker performs significantly better than other state-of-the-art algorithms. | ['Dong Wang', 'Xiaoyun Yang', 'Pengyu Zhang', 'Jie Zhao', 'Huchuan Lu'] | 2020-07-04 | null | null | null | null | ['rgb-t-tracking'] | ['computer-vision'] | [-3.02392486e-02 -6.29387498e-01 -4.54173833e-02 -3.06853354e-01
-7.20589697e-01 -7.25219131e-01 5.08240998e-01 -6.16793096e-01
-4.50904518e-01 2.98239499e-01 -2.53340125e-01 1.24459714e-01
2.74687618e-01 -8.60632956e-02 -5.77821434e-01 -9.95848894e-01
5.44175267e-01 -1.70624942e-01 4.57106918e-01 2.05837652e-01
-1.79324403e-01 4.22280043e-01 -1.50092924e+00 -2.83109844e-01
8.20497870e-01 1.35596526e+00 1.38250282e-02 6.59822524e-01
2.18575761e-01 4.66734886e-01 -3.78551811e-01 -2.77627796e-01
3.59808832e-01 -1.82037845e-01 -6.97222501e-02 3.47303957e-01
8.57819796e-01 -4.18606550e-01 -4.90266919e-01 1.04891384e+00
3.99919599e-01 1.64475635e-01 2.46774659e-01 -1.39006937e+00
-5.34964859e-01 -9.85515490e-02 -8.90167236e-01 9.29834247e-02
4.00921166e-01 6.22921288e-01 3.86073828e-01 -8.98712993e-01
4.50381398e-01 1.18999064e+00 6.96478724e-01 5.29430270e-01
-1.10850918e+00 -9.41409349e-01 4.66828942e-01 1.03057139e-01
-1.19170010e+00 -6.07980609e-01 8.70545506e-01 -3.77608418e-01
3.86129349e-01 1.95602372e-01 7.59068072e-01 1.17183435e+00
3.09063286e-01 8.58402252e-01 1.01889074e+00 -1.79017887e-01
-1.08079404e-01 -1.54569730e-01 -1.98671088e-01 5.98202884e-01
3.33727330e-01 3.21451008e-01 -6.27914310e-01 -8.39854106e-02
9.17419791e-01 2.52742261e-01 -1.86644629e-01 -4.52222645e-01
-1.52394307e+00 1.21347211e-01 4.74478275e-01 2.27821931e-01
-3.96524638e-01 5.51115572e-01 3.36077102e-02 -3.40686366e-02
3.22737426e-01 -2.16153696e-01 -2.39973754e-01 -9.31850895e-02
-9.27396894e-01 -1.44320711e-01 5.82583658e-02 9.87702310e-01
7.30599284e-01 2.85826772e-01 -3.93972725e-01 5.42992592e-01
9.84201193e-01 1.16425753e+00 1.47824258e-01 -1.04730952e+00
3.45493793e-01 5.01787305e-01 4.18046534e-01 -8.89172971e-01
-3.43413711e-01 -1.44976512e-01 -6.33728862e-01 2.77466267e-01
5.88201940e-01 -8.16878900e-02 -1.20933020e+00 1.79456449e+00
7.18939900e-01 5.62892258e-01 -9.52110514e-02 1.26934850e+00
8.57039928e-01 2.39645720e-01 2.09648907e-01 -2.99524754e-01
1.35842574e+00 -7.81573951e-01 -1.00169671e+00 -3.16334635e-01
5.23275323e-02 -1.01822329e+00 3.76179844e-01 2.33210281e-01
-1.01258028e+00 -8.62163901e-01 -1.07834756e+00 1.27981216e-01
-2.05479413e-01 6.03169739e-01 6.03980899e-01 6.93989754e-01
-9.80679393e-01 2.27047265e-01 -1.29231679e+00 -4.11180675e-01
-9.43662226e-02 4.93039191e-01 -3.13891351e-01 -1.60970896e-01
-1.05662811e+00 8.80854249e-01 1.95401385e-01 7.08298206e-01
-8.16276133e-01 -1.57987699e-01 -6.79095149e-01 -5.32114029e-01
4.42911088e-01 -8.77752006e-01 1.10296226e+00 -8.17923546e-01
-1.75564921e+00 5.11107862e-01 -3.56526673e-01 2.41019666e-01
4.28530812e-01 -1.59544811e-01 -5.26895583e-01 1.10168025e-01
-2.41308495e-01 6.06973886e-01 1.19273961e+00 -1.42272389e+00
-7.19341755e-01 -3.34280789e-01 -1.33625746e-01 1.33561984e-01
-2.98006833e-01 7.67559111e-02 -1.19027925e+00 -4.51637596e-01
4.08713937e-01 -1.24563324e+00 -6.80222064e-02 4.08317626e-01
-2.89077431e-01 -1.46795496e-01 9.07359481e-01 -7.32355893e-01
9.82181966e-01 -2.26935649e+00 2.33126134e-01 2.12266237e-01
3.21069062e-01 1.93428963e-01 -8.58893096e-02 -2.58849949e-01
1.78809404e-01 -3.84232104e-01 1.63924545e-01 -6.52554214e-01
1.74911097e-02 2.54222572e-01 8.60757008e-02 7.64628947e-01
-4.83569736e-03 9.51342344e-01 -9.19008315e-01 -6.27804458e-01
5.60045123e-01 6.93951964e-01 7.24687725e-02 2.43402556e-01
1.00909032e-01 9.06851888e-01 -7.18541622e-01 1.04746664e+00
8.77184093e-01 -2.42422998e-01 5.99209927e-02 -7.80730903e-01
-2.81357974e-01 -2.08100542e-01 -1.13820910e+00 1.86303890e+00
-1.23729318e-01 4.77911860e-01 2.02956185e-01 -2.11304292e-01
6.99364066e-01 4.02288318e-01 8.24863076e-01 -5.84175229e-01
5.01875699e-01 1.52086496e-01 -1.71216637e-01 -2.57938415e-01
5.52919924e-01 1.16384521e-01 -3.63074429e-02 2.70703703e-01
1.83592796e-01 1.77509144e-01 -1.26984775e-01 -9.05274302e-02
9.18556154e-01 8.34871531e-01 -1.14832319e-01 4.61326659e-01
4.61424053e-01 -2.13521555e-01 1.01343668e+00 6.43268287e-01
-6.22457385e-01 6.61807477e-01 -2.66763270e-01 -3.45276922e-01
-6.34156823e-01 -1.17366791e+00 1.29360318e-01 9.57878709e-01
7.73835301e-01 -8.89589116e-02 -4.78918552e-01 -6.60509408e-01
2.22646564e-01 1.22392192e-01 -6.21556163e-01 -2.53687918e-01
-5.91842294e-01 -5.78867614e-01 3.89438212e-01 6.82290792e-01
4.69714731e-01 -5.69394886e-01 -6.21658981e-01 6.96306080e-02
-3.51463258e-01 -1.42551863e+00 -8.43988836e-01 -2.20660679e-02
-9.90228951e-01 -8.93488586e-01 -6.73195183e-01 -2.08961308e-01
6.31653607e-01 6.23293102e-01 3.87815803e-01 1.07443251e-01
-4.43586819e-02 9.04100239e-01 -3.12550575e-01 7.92455375e-02
5.87008223e-02 -2.87757784e-01 4.09036845e-01 3.54350597e-01
7.74404556e-02 -2.66316056e-01 -7.69275904e-01 5.94504654e-01
-8.01803708e-01 9.79088433e-03 7.22722232e-01 4.19316679e-01
6.30091250e-01 -3.92347306e-01 -1.03923865e-01 1.22979805e-02
-1.09624341e-01 2.86649056e-02 -8.09965789e-01 6.68520689e-01
-4.20241863e-01 -7.29673877e-02 9.81835574e-02 -8.97239089e-01
-1.36058366e+00 5.76117694e-01 3.27962667e-01 -1.16278660e+00
6.54035807e-02 4.61815298e-02 -2.86048084e-01 -7.27907896e-01
1.03045918e-01 1.80677891e-01 -1.84405088e-01 -3.63300890e-01
6.42028630e-01 3.83719414e-01 9.13364470e-01 -5.71278572e-01
1.34957385e+00 6.46168530e-01 -1.32253572e-01 -3.79146695e-01
-4.74758178e-01 -6.08969033e-01 -7.93043852e-01 -8.88370872e-01
1.02779353e+00 -1.06671453e+00 -1.02137148e+00 8.74716878e-01
-1.16791964e+00 -4.88957344e-03 1.52606979e-01 7.83835590e-01
-3.80484998e-01 5.33511162e-01 -5.95484138e-01 -1.04928911e+00
-1.51827976e-01 -1.22671282e+00 1.40430856e+00 7.70294905e-01
3.19954962e-01 -8.15135777e-01 1.71899006e-01 1.50483415e-01
3.85986477e-01 3.57724935e-01 -2.17250466e-01 -2.07824614e-02
-1.02212846e+00 -3.09896201e-01 -3.05356711e-01 -8.00831467e-02
2.72911638e-01 2.98192292e-01 -1.07166660e+00 -4.52280849e-01
-1.78320333e-01 -4.48237807e-02 8.47870409e-01 4.64393288e-01
6.18032455e-01 2.71785796e-01 -6.52496159e-01 9.64560807e-01
1.17570519e+00 2.56078035e-01 4.65147763e-01 3.66840482e-01
1.10838866e+00 3.51746790e-02 8.97906065e-01 4.06256229e-01
5.52069485e-01 1.05463636e+00 2.95375228e-01 -1.76608175e-01
-2.54873127e-01 -8.20028558e-02 7.00739503e-01 8.73058617e-01
-3.44295442e-01 -2.66716350e-02 -5.48372269e-01 1.00548387e-01
-2.16954923e+00 -7.23854363e-01 -7.97978044e-02 2.32689667e+00
7.29205787e-01 -1.30573064e-02 -1.16412686e-02 -6.24522448e-01
8.39976132e-01 7.09314495e-02 -8.41618359e-01 5.01011848e-01
-1.96208343e-01 -2.09392816e-01 8.73167396e-01 1.31897017e-01
-1.25317585e+00 8.96513462e-01 6.46506262e+00 5.48584938e-01
-1.29853344e+00 2.27160528e-01 2.75878534e-02 -3.46035600e-01
-1.10806867e-01 2.20907182e-01 -7.58732378e-01 5.58754146e-01
6.49361312e-01 2.64268875e-01 3.34842831e-01 4.11927432e-01
1.57318935e-01 -2.04043522e-01 -9.31960940e-01 1.11801422e+00
1.70476049e-01 -6.43063009e-01 -5.41913211e-01 8.04096311e-02
5.51185429e-01 5.50358668e-02 2.50564039e-01 1.71723083e-01
4.71158355e-01 -3.80931109e-01 9.73591208e-01 1.13428199e+00
5.90836287e-01 -1.77665502e-01 5.15674174e-01 -5.49831502e-02
-1.77081323e+00 1.49042547e-01 -1.44914284e-01 5.59854567e-01
4.06890094e-01 3.56193125e-01 -5.57050034e-02 9.74431515e-01
8.31823230e-01 9.77088332e-01 -7.64399767e-01 1.25276554e+00
-2.69332498e-01 2.03241512e-01 -6.06583118e-01 4.69120324e-01
-1.09990917e-01 -1.07262835e-01 5.55176437e-01 7.37551093e-01
4.64349329e-01 1.13042615e-01 5.08871198e-01 7.94179678e-01
2.76660919e-01 -5.36325634e-01 -2.28359297e-01 1.95184916e-01
5.79909921e-01 1.68269122e+00 -6.66220486e-01 -2.67313123e-01
-5.91907322e-01 1.03235316e+00 4.69692983e-02 7.89136171e-01
-1.21129179e+00 1.40254483e-01 5.77043951e-01 -5.15266180e-01
5.66742897e-01 -4.41224724e-01 1.97138369e-01 -1.53833854e+00
9.12019461e-02 -4.07960415e-01 2.03183681e-01 -1.19632697e+00
-1.22414088e+00 3.89862776e-01 -1.45627907e-03 -1.82842922e+00
-9.99329910e-02 -5.95701158e-01 -5.01282215e-01 8.01272810e-01
-1.46232522e+00 -1.74615085e+00 -4.56897050e-01 6.21004343e-01
-1.26062900e-01 1.91875115e-01 1.31563470e-01 3.74566048e-01
-8.87066901e-01 6.86888278e-01 5.36559410e-02 7.32972547e-02
1.24552572e+00 -9.08127904e-01 1.88552946e-01 1.13523221e+00
-6.29441738e-02 8.45550358e-01 5.27689397e-01 -7.90968478e-01
-1.94465697e+00 -8.98243904e-01 1.67378917e-01 -6.05903387e-01
5.53498805e-01 -1.40007421e-01 -6.30095601e-01 7.20906615e-01
1.59960628e-01 4.40254062e-01 3.87753546e-01 -7.65930489e-02
-4.14179683e-01 -3.30208600e-01 -9.07045424e-01 2.77866393e-01
7.74019480e-01 -5.23429096e-01 -4.27818894e-01 -1.38325453e-01
5.95487654e-01 -7.87387550e-01 -1.10061181e+00 5.29833734e-01
1.12045443e+00 -7.00430691e-01 1.17584133e+00 1.33095786e-01
-5.40702462e-01 -1.07693267e+00 -1.48282379e-01 -9.42478597e-01
-2.90246397e-01 -5.15950501e-01 -5.04594862e-01 1.33643770e+00
-8.12126696e-02 -6.09564602e-01 5.49223423e-01 8.54699314e-01
-1.09629989e-01 -1.14073627e-01 -1.32712960e+00 -7.26413608e-01
-6.27198100e-01 -3.52536261e-01 3.10293525e-01 7.92183399e-01
-2.71541476e-01 -9.72362384e-02 -8.06218445e-01 3.95872265e-01
1.02422190e+00 2.11227074e-01 8.45196784e-01 -1.06499922e+00
-1.17609456e-01 -3.49031687e-01 -3.53713065e-01 -1.20264506e+00
-9.80182290e-02 -3.76575887e-01 4.90664184e-01 -1.14654291e+00
2.32710481e-01 -2.76576251e-01 -7.60791659e-01 6.55465961e-01
-4.84523296e-01 4.39942896e-01 4.69316125e-01 5.40869951e-01
-1.11790240e+00 8.02369773e-01 1.37500024e+00 -1.85041428e-01
-1.55128106e-01 -1.14115328e-01 -3.11895728e-01 5.35350442e-01
3.21054459e-01 -3.46880168e-01 7.74067342e-02 -5.20314872e-01
-7.79749528e-02 1.83562949e-01 6.14054620e-01 -1.03255105e+00
7.55561054e-01 -2.16314465e-01 9.29892302e-01 -8.62458885e-01
5.29469848e-01 -1.11514258e+00 3.75399321e-01 1.72746211e-01
1.80800959e-01 3.12008038e-02 3.55396241e-01 7.97942579e-01
-4.68878001e-02 4.34631348e-01 6.25542700e-01 3.27687711e-01
-7.29038894e-01 6.40817761e-01 -9.65590626e-02 -5.26929259e-01
9.65581298e-01 -5.59859335e-01 -3.87672603e-01 -2.61432588e-01
-7.68641472e-01 4.85820979e-01 7.84256697e-01 7.26085782e-01
6.22974157e-01 -1.90622580e+00 -2.39655763e-01 1.77936301e-01
3.51889312e-01 -3.99325907e-01 5.01166523e-01 1.58361089e+00
7.04723895e-02 1.54283479e-01 -1.21940859e-01 -1.06636584e+00
-1.19896138e+00 5.18342972e-01 4.80015874e-01 1.16726466e-01
-3.88349265e-01 4.72291946e-01 1.49150401e-01 -2.83203661e-01
2.35580415e-01 -5.56936502e-01 9.52522922e-03 -1.86854824e-01
4.85771239e-01 2.02714726e-01 -2.99477577e-01 -1.04009771e+00
-7.94214010e-01 1.13856709e+00 1.93060741e-01 -1.69156626e-01
8.08901429e-01 -5.73937833e-01 3.45297791e-02 4.79029626e-01
7.81394899e-01 -2.26694554e-01 -1.78543544e+00 -3.84988099e-01
-2.43903667e-01 -6.88837826e-01 1.52564302e-01 -6.29983604e-01
-1.53670764e+00 4.79647905e-01 1.02976584e+00 -1.31644130e-01
1.36795068e+00 -8.49424824e-02 7.29620397e-01 8.39440152e-02
3.94077271e-01 -9.47710216e-01 7.05646072e-03 3.32852930e-01
2.93103814e-01 -1.30089629e+00 2.21679375e-01 -1.86719254e-01
-3.84851456e-01 1.10459673e+00 8.40672493e-01 1.17930263e-01
3.31765264e-01 2.08851293e-01 3.22387666e-01 2.61704829e-02
-5.94333827e-01 -4.86394167e-01 7.48150170e-01 5.64115822e-01
2.80253649e-01 -1.61595166e-01 1.76945806e-01 3.71543884e-01
6.23485923e-01 9.70504805e-02 -5.45022115e-02 8.49428773e-01
-3.03228021e-01 -1.17107618e+00 -8.98423135e-01 -1.16343778e-02
-4.52231653e-02 2.63352841e-01 -3.54949623e-01 7.44808197e-01
2.18982533e-01 1.01979601e+00 -2.70980895e-01 -7.98380017e-01
2.93240279e-01 -1.12077653e-01 8.45608115e-01 4.26882245e-02
-5.90884447e-01 7.08110988e-01 -2.45965406e-01 -9.47403610e-01
-1.09804308e+00 -8.16733181e-01 -9.65637743e-01 -2.62848228e-01
-7.75850534e-01 -3.39363635e-01 6.60334289e-01 1.14667559e+00
1.56388879e-01 6.48651481e-01 5.80972731e-01 -1.40851998e+00
-1.57230720e-01 -8.93028140e-01 -4.67722327e-01 2.26785600e-01
5.21550715e-01 -1.18083870e+00 -3.48781615e-01 1.48498848e-01] | [6.389822006225586, -2.1567201614379883] |
5e61bf67-18c8-4b7d-9534-5e4414696882 | exploring-intermediate-representation-for | 2011.08464 | null | https://arxiv.org/abs/2011.08464v5 | https://arxiv.org/pdf/2011.08464v5.pdf | Exploring intermediate representation for monocular vehicle pose estimation | We present a new learning-based framework to recover vehicle pose in SO(3) from a single RGB image. In contrast to previous works that map from local appearance to observation angles, we explore a progressive approach by extracting meaningful Intermediate Geometrical Representations (IGRs) to estimate egocentric vehicle orientation. This approach features a deep model that transforms perceived intensities to IGRs, which are mapped to a 3D representation encoding object orientation in the camera coordinate system. Core problems are what IGRs to use and how to learn them more effectively. We answer the former question by designing IGRs based on an interpolated cuboid that derives from primitive 3D annotation readily. The latter question motivates us to incorporate geometry knowledge with a new loss function based on a projective invariant. This loss function allows unlabeled data to be used in the training stage to improve representation learning. Without additional labels, our system outperforms previous monocular RGB-based methods for joint vehicle detection and pose estimation on the KITTI benchmark, achieving performance even comparable to stereo methods. Code and pre-trained models are available at this https URL. | ['Kwang-Ting Cheng', 'Hongyang Li', 'Zengqiang Yan', 'Shichao Li'] | 2020-11-17 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Li_Exploring_intermediate_representation_for_monocular_vehicle_pose_estimation_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Li_Exploring_intermediate_representation_for_monocular_vehicle_pose_estimation_CVPR_2021_paper.pdf | cvpr-2021-1 | ['vehicle-pose-estimation'] | ['computer-vision'] | [ 1.66297168e-01 2.58595735e-01 -3.19044948e-01 -6.72270954e-01
-9.00902748e-01 -8.02934647e-01 8.37801218e-01 -4.50174332e-01
-4.13452923e-01 4.90210652e-01 3.58614214e-02 -3.63772959e-01
4.60794330e-01 -7.21803844e-01 -1.11223924e+00 -5.85694671e-01
3.27303439e-01 6.41599357e-01 1.70239553e-01 -1.38920829e-01
3.40938896e-01 8.49999666e-01 -1.57582533e+00 -1.59067318e-01
4.64445442e-01 8.96582782e-01 -5.29595204e-02 6.10690534e-01
1.31853849e-01 7.72963166e-01 2.87354607e-02 -2.31432363e-01
6.21399224e-01 -1.40823632e-01 -8.12375426e-01 3.56986195e-01
8.73458207e-01 -6.47200227e-01 -3.92818153e-01 8.67453039e-01
6.17759600e-02 6.36795908e-02 9.27748859e-01 -1.35483301e+00
-2.93039143e-01 -2.86630869e-01 -4.74795491e-01 -2.14092135e-01
6.14501476e-01 1.34431824e-01 9.13163304e-01 -9.49826300e-01
9.56679642e-01 1.33389831e+00 7.39886582e-01 3.56092244e-01
-1.26343811e+00 -3.24260861e-01 1.69688150e-01 4.20893610e-01
-1.59865248e+00 -4.28473890e-01 7.95264125e-01 -5.20458162e-01
9.41326678e-01 2.08175883e-01 6.34209216e-01 9.70131636e-01
-2.77750015e-01 7.91559935e-01 1.08462298e+00 -4.36392099e-01
1.87777311e-01 3.17573935e-01 -1.01352707e-01 6.99828565e-01
1.48415685e-01 1.60774812e-01 -2.59422362e-01 1.08079128e-01
1.02961814e+00 6.24157712e-02 -1.49208456e-01 -1.19611371e+00
-1.01223266e+00 9.70499814e-01 7.80118227e-01 -3.12803388e-01
-8.85850936e-02 3.21700871e-01 1.46576941e-01 1.87348619e-01
4.81622696e-01 1.71811759e-01 -3.62935096e-01 8.60147551e-02
-5.25871158e-01 3.43121737e-01 5.70627272e-01 1.02781618e+00
1.34087777e+00 -8.19658637e-02 3.16399425e-01 5.35971105e-01
5.86584866e-01 8.24190855e-01 -2.06861109e-01 -1.52746165e+00
4.67933327e-01 6.02415442e-01 2.32011199e-01 -6.77276254e-01
-3.47657859e-01 -1.96639881e-01 -2.63971090e-01 5.35666227e-01
4.88326967e-01 1.35021716e-01 -9.35619473e-01 1.53672659e+00
6.32617950e-01 2.58933187e-01 6.71327533e-03 1.22174048e+00
5.88639259e-01 4.53133166e-01 -2.08684549e-01 3.47581685e-01
1.04548347e+00 -8.05820882e-01 -2.46368274e-01 -3.63567263e-01
8.83722425e-01 -5.95455408e-01 7.24019289e-01 3.44696194e-01
-6.64097428e-01 -4.64576900e-01 -1.11366093e+00 -5.02018094e-01
-5.33115864e-01 1.55987784e-01 5.77026844e-01 7.79467463e-01
-1.37461841e+00 2.89685071e-01 -8.15128267e-01 -6.16490781e-01
4.21473719e-02 3.54654133e-01 -6.07199490e-01 -1.47886813e-01
-9.64299440e-01 1.21645260e+00 9.55861881e-02 5.70383668e-02
-7.63406873e-01 -3.55389327e-01 -1.29420424e+00 -5.68584144e-01
3.29861403e-01 -8.14882338e-01 1.15284038e+00 -8.32962275e-01
-1.66537786e+00 1.11204064e+00 -3.15110207e-01 -4.54699129e-01
6.82689786e-01 -1.62730247e-01 2.00515881e-01 3.05360198e-01
1.35205582e-01 1.17617476e+00 8.28934789e-01 -1.65108013e+00
-4.66236055e-01 -5.60727596e-01 4.04617578e-01 5.73115408e-01
1.81509346e-01 -3.21605951e-01 -8.59222710e-01 -8.09065402e-02
4.89150733e-01 -1.29971385e+00 -2.73584157e-01 2.70577103e-01
-2.83553123e-01 -8.54422450e-02 8.01558316e-01 -6.58727348e-01
3.44897836e-01 -1.85658157e+00 3.87851685e-01 2.85507381e-01
6.26721978e-02 -2.03980524e-02 -9.88802612e-02 7.21281022e-02
8.18945002e-03 -1.82751283e-01 -9.96703282e-02 -6.44021630e-01
1.31113619e-01 2.14549825e-01 -2.74412483e-01 9.86336887e-01
2.83646196e-01 8.66429150e-01 -8.56416166e-01 -3.12486023e-01
8.38599086e-01 8.50150466e-01 -6.45787776e-01 1.94926009e-01
-1.35045061e-02 6.81514502e-01 -4.05035228e-01 5.44478059e-01
1.08898616e+00 9.24278423e-02 -7.92764574e-02 -3.70734185e-01
-1.56925783e-01 2.99709499e-01 -1.06416702e+00 1.69827592e+00
-5.78961313e-01 6.60108268e-01 -4.58468273e-02 -1.04881656e+00
1.03905118e+00 -4.22186079e-03 4.18821365e-01 -7.81996191e-01
1.29562765e-01 1.72033161e-01 -7.68293738e-01 -2.89949626e-01
5.11588752e-01 1.11496545e-01 -1.84939519e-01 6.66565374e-02
2.80806664e-02 -5.50623655e-01 -9.36463401e-02 1.46996364e-01
7.38187253e-01 8.58656943e-01 1.51812822e-01 -6.26407266e-02
7.74911106e-01 9.42440107e-02 4.07824874e-01 3.88136148e-01
-1.36240572e-03 9.88229454e-01 3.99024010e-01 -7.42347479e-01
-1.32705367e+00 -1.17543077e+00 -2.48056248e-01 6.81142926e-01
4.33111697e-01 -8.82393643e-02 -6.66415393e-01 -7.22327471e-01
1.44188166e-01 5.84638059e-01 -7.39595890e-01 7.46184513e-02
-6.60491228e-01 -2.86022246e-01 2.43706569e-01 6.81601882e-01
4.18379962e-01 -5.46002805e-01 -6.48793519e-01 -1.86015338e-01
-2.83169776e-01 -1.45517683e+00 -1.98622331e-01 1.84112549e-01
-8.41040790e-01 -9.45904970e-01 -6.79193377e-01 -4.85540867e-01
8.34757924e-01 4.88757938e-01 9.20174837e-01 -2.34915033e-01
-1.27822801e-01 6.81314707e-01 -2.50379175e-01 -1.36913776e-01
-2.17747211e-01 1.20311752e-01 3.05604115e-02 -2.24579908e-02
5.60245752e-01 -2.78148293e-01 -7.15653718e-01 3.65880340e-01
-4.61155027e-01 2.79877335e-01 5.02070904e-01 3.42850298e-01
5.77254832e-01 -8.60291600e-01 -3.15108486e-02 -5.45742571e-01
-3.51001769e-01 -2.57491589e-01 -9.77253854e-01 -2.21067131e-01
-3.26802820e-01 1.68356851e-01 4.51876640e-01 1.56164676e-01
-9.10765350e-01 7.25540936e-01 -3.14992547e-01 -5.88659644e-01
-3.72796476e-01 -6.60290569e-02 -1.68621078e-01 -3.27345282e-01
5.27333856e-01 3.24749574e-02 1.39565125e-01 -3.31393659e-01
6.66343331e-01 5.82603931e-01 4.66319531e-01 -3.60538721e-01
9.51199889e-01 8.84663463e-01 2.82460451e-01 -9.06540990e-01
-6.91036403e-01 -8.83490026e-01 -1.08587384e+00 -3.01949531e-01
9.75884676e-01 -1.38355911e+00 -7.69054532e-01 2.10202083e-01
-1.20144832e+00 -4.12231922e-01 -7.74792805e-02 6.29202724e-01
-1.00065517e+00 4.90195453e-01 -4.15213525e-01 -8.46960485e-01
1.32768974e-01 -1.21047914e+00 1.85816514e+00 -1.27738208e-01
-3.70146856e-02 -9.60412145e-01 2.00345814e-01 5.49442410e-01
1.72784224e-01 3.12215239e-01 2.80248225e-01 3.48399356e-02
-1.11070466e+00 -3.01360309e-01 -4.78213251e-01 3.00858229e-01
-2.95002848e-01 -3.04905713e-01 -1.13944662e+00 -2.77589798e-01
-1.07375439e-02 -4.27888274e-01 1.06326771e+00 2.69456774e-01
7.72906482e-01 1.00035621e-02 -4.42525148e-01 1.02727032e+00
1.57227767e+00 -2.03598499e-01 7.40077913e-01 6.44025624e-01
1.11133575e+00 6.85011327e-01 5.48994422e-01 1.02358133e-01
8.64016712e-01 1.14353192e+00 6.83129072e-01 -1.50360942e-01
-1.50396168e-01 -3.16620499e-01 5.15106559e-01 4.59268987e-01
-2.91448504e-01 2.02215388e-01 -8.28477919e-01 3.78792256e-01
-1.81323850e+00 -7.37289488e-01 -1.81954458e-01 2.32176113e+00
3.88907135e-01 2.69338451e-02 1.40627906e-01 8.54546502e-02
2.58981287e-01 1.14611246e-01 -5.02057016e-01 -3.62479568e-01
3.27307805e-02 -3.89470011e-01 1.07325733e+00 9.17937815e-01
-1.28158104e+00 1.04558110e+00 5.77371407e+00 2.63150305e-01
-1.30970526e+00 8.97805095e-02 6.25542760e-01 2.42395461e-01
-4.58809972e-01 2.68088102e-01 -1.02051365e+00 -4.23754938e-02
7.79546678e-01 4.54702228e-01 4.21910286e-01 9.30569887e-01
1.67270824e-01 -1.10336110e-01 -1.08774078e+00 1.08710730e+00
4.62391585e-01 -1.06559718e+00 1.00689586e-02 2.13389903e-01
7.64963806e-01 4.76020992e-01 7.87660405e-02 2.44675815e-01
2.57218808e-01 -8.27937365e-01 9.40592527e-01 4.56906825e-01
8.69477808e-01 -5.39133191e-01 5.74742734e-01 9.30453762e-02
-1.42045915e+00 2.19176829e-01 -3.97995949e-01 -6.03806693e-03
1.22504435e-01 -3.05989385e-02 -1.15828943e+00 6.64361119e-01
5.29762447e-01 8.78870130e-01 -7.50597835e-01 7.80827224e-01
-3.48129839e-01 5.59178330e-02 -5.16458154e-01 2.00143471e-01
3.25423211e-01 -4.67242539e-01 4.13058639e-01 1.07621419e+00
3.33300859e-01 -3.04614902e-01 1.45926520e-01 8.85078549e-01
1.86770961e-01 -6.60322458e-02 -9.65840518e-01 6.35291934e-01
1.26535863e-01 1.38279963e+00 -7.35811293e-01 -2.37381741e-01
-4.95517671e-01 1.06805468e+00 3.70830864e-01 6.25175297e-01
-8.90154362e-01 -1.17845282e-01 5.44983149e-01 3.51558417e-01
5.64254701e-01 -4.06826377e-01 -1.40034094e-01 -1.26031876e+00
-1.01443119e-02 -4.32303935e-01 -1.65490821e-01 -8.63269329e-01
-8.14439356e-01 5.68447351e-01 2.80593067e-01 -1.60210621e+00
-7.06500888e-01 -8.95741105e-01 -2.06236497e-01 7.06113756e-01
-1.80616057e+00 -1.51056743e+00 -4.63083357e-01 4.33730394e-01
2.75355011e-01 2.99062282e-01 5.69553614e-01 1.70712784e-01
-1.59361854e-01 3.25636297e-01 1.48510160e-02 1.55228183e-01
5.32750666e-01 -1.37667763e+00 6.77186430e-01 6.44731700e-01
1.75303191e-01 4.46048200e-01 7.37958491e-01 -1.99624106e-01
-1.65752852e+00 -1.21490371e+00 7.80013502e-01 -1.02706587e+00
2.20443279e-01 -6.75734818e-01 -5.53624392e-01 9.66264486e-01
-6.46154508e-02 3.54936332e-01 1.07533604e-01 -7.17392564e-02
-6.73464715e-01 -1.26918808e-01 -1.02674150e+00 5.55962086e-01
1.16727531e+00 -8.69393051e-01 -3.61404836e-01 1.91232800e-01
3.50487739e-01 -5.74589431e-01 -4.06021923e-01 5.53884923e-01
5.22119164e-01 -9.95177925e-01 1.21038318e+00 -1.11893505e-01
-6.37770295e-02 -7.45839715e-01 -2.80057847e-01 -1.13198137e+00
9.96806175e-02 -3.07218164e-01 1.26678601e-01 6.20448172e-01
1.94690600e-01 -6.49647355e-01 1.03810048e+00 3.62318754e-01
-1.78366259e-01 -5.04345000e-01 -1.00443733e+00 -7.53095806e-01
7.55395219e-02 -5.90587020e-01 4.24025625e-01 5.84420621e-01
-2.10986257e-01 3.51811111e-01 -3.18129063e-01 3.13317508e-01
9.37439382e-01 1.65400624e-01 1.22192073e+00 -1.16495609e+00
-5.65739535e-03 -1.69035658e-01 -9.01517689e-01 -1.57453430e+00
2.88347691e-01 -9.36988056e-01 1.18580006e-01 -1.51925755e+00
1.07733950e-01 -3.74728769e-01 9.21064913e-02 2.98125952e-01
1.11973628e-01 8.55534017e-01 1.72691450e-01 8.08157846e-02
-9.20039475e-01 6.22315586e-01 9.79748011e-01 4.46348228e-02
9.90314502e-03 -2.61062056e-01 -2.49312341e-01 8.78380060e-01
6.53857231e-01 -1.27778396e-01 -2.26336300e-01 -4.28304374e-01
2.40360260e-01 -1.47661477e-01 8.13410342e-01 -9.37876999e-01
1.24877401e-01 1.08555190e-01 5.05481362e-01 -8.38112056e-01
8.11090171e-01 -8.03542733e-01 1.47141159e-01 2.05508888e-01
-2.30835658e-02 -2.13494390e-01 -4.91097346e-02 5.05066812e-01
-1.63387179e-01 -8.16467479e-02 5.99790812e-01 -8.04995000e-02
-9.70507979e-01 2.44515926e-01 -2.36004204e-01 -2.67457932e-01
9.96745825e-01 -5.88409185e-01 3.39223407e-02 -6.99840128e-01
-5.72027266e-01 6.93213493e-02 1.01087391e+00 5.88883638e-01
6.09523773e-01 -1.47791088e+00 -5.48567832e-01 4.18159038e-01
4.05115008e-01 1.40326694e-01 -2.67996155e-02 7.74576604e-01
-9.55907404e-01 8.51459980e-01 -1.68239281e-01 -1.06551576e+00
-9.52419460e-01 4.11297858e-01 4.86732423e-01 8.22430030e-02
-4.54784244e-01 6.17244959e-01 3.73635024e-01 -1.00442839e+00
2.67311335e-01 -3.55365008e-01 -1.67423457e-01 -1.95831388e-01
2.95035988e-01 2.28390455e-01 2.53267646e-01 -1.30159783e+00
-3.92811537e-01 1.05692589e+00 1.04274333e-01 -2.27712244e-01
1.15623677e+00 -5.34481227e-01 3.09878200e-01 2.83864230e-01
1.70705295e+00 -1.80109903e-01 -1.96991658e+00 -1.29724503e-01
-5.06063737e-02 -5.92606544e-01 1.24580629e-01 -1.62226677e-01
-9.69493151e-01 8.47681344e-01 7.67917395e-01 -2.35063255e-01
6.46311879e-01 2.93685228e-01 3.66698176e-01 5.51101267e-01
5.43374062e-01 -9.16384757e-01 -2.18090555e-03 7.67421007e-01
8.22897375e-01 -1.57977021e+00 6.75129443e-02 -5.04514575e-01
-5.78118205e-01 1.19497859e+00 4.64014411e-01 -3.33103448e-01
4.21836466e-01 1.32003929e-02 3.36849779e-01 -2.41809487e-02
-2.33149379e-01 -5.08421361e-01 3.45160574e-01 7.02317595e-01
3.23410869e-01 -1.63620174e-01 2.12660417e-01 -3.15919012e-01
-1.45979002e-01 -1.24455325e-01 3.14408451e-01 7.13897824e-01
-3.57252359e-01 -1.07458854e+00 -5.14087319e-01 -6.54941872e-02
-6.97141467e-03 3.09350640e-01 -3.25546205e-01 7.76445806e-01
1.85660049e-01 6.76283717e-01 3.19391459e-01 -3.16550463e-01
3.25472891e-01 -8.91492516e-02 7.52785861e-01 -5.73529363e-01
1.88832238e-01 -1.73780881e-02 8.82326961e-02 -8.07576895e-01
-6.78327799e-01 -8.47523332e-01 -1.10303617e+00 -1.20409369e-01
-2.23495692e-01 -1.89721227e-01 9.36073363e-01 9.08770561e-01
1.24249078e-01 -3.54188122e-02 9.32747900e-01 -1.63675439e+00
-3.36228609e-01 -6.01509154e-01 -3.15160245e-01 5.54328859e-01
6.53655529e-01 -8.14885616e-01 -4.54981744e-01 1.45293355e-01] | [7.836111068725586, -2.411095142364502] |
248d349c-5ebb-4213-bcfc-66cabbe41a28 | finding-the-smallest-or-largest-element-of-a | 2210.11413 | null | https://arxiv.org/abs/2210.11413v2 | https://arxiv.org/pdf/2210.11413v2.pdf | Finding the smallest or largest element of a tensor from its low-rank factors | We consider the problem of finding the smallest or largest entry of a tensor of order $N$ that is specified via its rank decomposition. Stated in a different way, we are given $N$ sets of $R$-dimensional vectors and we wish to select one vector from each set such that the sum of the Hadamard product of the selected vectors is minimized or maximized. This is a fundamental tensor problem with numerous applications in embedding similarity search, recommender systems, graph mining, multivariate probability, and statistics. We show that this discrete optimization problem is NP-hard for any tensor rank higher than one, but also provide an equivalent continuous problem reformulation which is amenable to disciplined non-convex optimization. We propose a suite of gradient-based approximation algorithms whose performance in preliminary experiments appears to be promising. | ['Aritra Konar', 'Paris Karakasis', 'Nicholas D. Sidiropoulos'] | 2022-10-16 | null | null | null | null | ['graph-mining'] | ['graphs'] | [-7.20187724e-02 -6.33323342e-02 -2.81037629e-01 -2.49437302e-01
-6.55354798e-01 -7.05267251e-01 2.54549146e-01 2.22775936e-01
-3.10382307e-01 3.73066813e-01 2.81136841e-01 -4.68973935e-01
-7.49773979e-01 -6.64475024e-01 -4.96190548e-01 -5.66553473e-01
-9.80891287e-01 7.79459298e-01 -3.29385996e-01 -3.56531620e-01
4.15130794e-01 4.84596908e-01 -1.31463170e+00 -4.32382934e-02
5.52805722e-01 1.32881057e+00 -1.89478219e-01 7.35991955e-01
5.37044890e-02 4.09493566e-01 -7.39589855e-02 -6.07346892e-01
7.51545608e-01 -5.91896567e-03 -1.13330400e+00 2.94061899e-01
5.05914152e-01 1.04476884e-01 -3.70629936e-01 1.38737667e+00
2.28869244e-01 3.92565310e-01 5.87305248e-01 -1.21755087e+00
-4.36275959e-01 7.88303494e-01 -7.75571465e-01 3.74712437e-01
5.54926157e-01 -1.93643212e-01 1.87169826e+00 -9.51484621e-01
5.41682422e-01 1.04060709e+00 3.65320891e-01 5.96311726e-02
-1.67168760e+00 -3.06591183e-01 1.20949239e-01 -2.00246330e-02
-1.46786511e+00 -2.87447780e-01 9.07573044e-01 -4.93070215e-01
8.57717574e-01 6.08607173e-01 6.29966736e-01 2.22834483e-01
-2.25068644e-01 4.95101839e-01 9.39756155e-01 -2.34446049e-01
2.74694979e-01 5.95571548e-02 2.53173888e-01 1.08404541e+00
5.13439357e-01 -2.89614677e-01 -5.40627718e-01 -7.93843269e-01
2.11910054e-01 -9.04931054e-02 -2.38017961e-01 -6.00131452e-01
-1.51655722e+00 1.04506814e+00 2.51914173e-01 3.70810986e-01
-3.95890504e-01 4.54930216e-01 2.66535789e-01 5.05061328e-01
4.10086989e-01 6.43206835e-01 -3.38684976e-01 -3.81775796e-02
-6.99160218e-01 4.12464261e-01 1.09037495e+00 7.76679695e-01
8.67722511e-01 5.94993830e-02 3.09862524e-01 7.79090583e-01
2.68044144e-01 5.46074688e-01 -7.07798377e-02 -1.00074089e+00
6.79408908e-01 3.43863457e-01 3.27758640e-01 -1.51070356e+00
-9.82021764e-02 -3.40819836e-01 -7.11806774e-01 -1.43864304e-01
3.73591900e-01 5.30247502e-02 -3.19639891e-01 1.70123136e+00
3.00087512e-01 3.65379453e-02 -3.77532482e-01 9.41658318e-01
1.61507070e-01 7.63857126e-01 -3.11707586e-01 -4.43396062e-01
1.16671109e+00 -4.50652510e-01 -3.84093314e-01 -2.25512292e-02
7.30439067e-01 -7.95117497e-01 7.88160741e-01 6.25468731e-01
-1.06773770e+00 2.42308989e-01 -1.03140879e+00 1.23032562e-01
-1.46806806e-01 -1.37428194e-02 1.14905095e+00 6.32732153e-01
-1.26230121e+00 7.60830164e-01 -5.44379711e-01 1.60629563e-02
-1.98503181e-01 7.50985265e-01 -5.10346115e-01 -3.09269518e-01
-8.32921326e-01 4.23333019e-01 8.55029840e-03 1.66049525e-01
-4.42982912e-01 -4.54845577e-01 -6.79461360e-01 -1.19388886e-01
4.33432281e-01 -4.55300689e-01 7.82516778e-01 -3.11437190e-01
-9.77798343e-01 7.73350954e-01 -1.13376111e-01 -3.42188746e-01
-1.90116793e-01 4.07357156e-01 -2.02044979e-01 1.48553297e-01
2.70981580e-01 -1.80146053e-01 8.89227927e-01 -8.01580906e-01
-3.94550055e-01 -5.59081078e-01 2.60018528e-01 1.82326376e-01
-4.05399442e-01 4.78219353e-02 -2.37297907e-01 -4.55947220e-01
5.93826115e-01 -9.96051431e-01 -7.42328763e-01 -3.61463964e-01
-5.48603237e-01 -4.88886982e-01 3.49962652e-01 -5.37056267e-01
1.38754725e+00 -2.06624866e+00 7.40149677e-01 9.40719306e-01
6.06714487e-01 -4.35011119e-01 -3.15682113e-01 6.94891036e-01
-3.45048457e-01 3.86604130e-01 -1.62014708e-01 -1.79450989e-01
3.09365243e-01 1.39751434e-02 -2.64698952e-01 9.01444495e-01
-1.70423537e-01 2.52203405e-01 -8.61388206e-01 -2.33871982e-01
-2.50101596e-01 -1.21516868e-01 -7.98417747e-01 -1.39112016e-02
-2.39471465e-01 -3.19660217e-01 -6.62451804e-01 6.55119956e-01
6.51499808e-01 -4.88645911e-01 4.66693014e-01 -3.80980194e-01
4.21709865e-02 1.75567061e-01 -1.82023609e+00 1.32984948e+00
-1.31189838e-01 4.19748515e-01 5.31008303e-01 -1.35317278e+00
6.44844294e-01 1.46429464e-01 1.18307579e+00 -1.80088580e-01
1.28193706e-01 3.78278136e-01 -3.97032611e-02 -2.52836645e-01
8.22290719e-01 -2.62052417e-01 -2.63850480e-01 8.94106984e-01
-2.73209274e-01 -8.20021033e-02 5.83225906e-01 7.34516561e-01
1.37161088e+00 -9.62455213e-01 -1.83144316e-01 -3.93502086e-01
4.40190971e-01 -7.36725405e-02 4.32423562e-01 5.29767394e-01
8.51246342e-02 1.71460658e-01 7.66039968e-01 -2.42956176e-01
-1.13915968e+00 -9.12720680e-01 4.96026175e-03 1.07135868e+00
3.08494046e-02 -7.07013071e-01 -3.59525353e-01 -4.93130922e-01
3.58842313e-01 1.90227434e-01 -5.48012376e-01 2.26602376e-01
-4.77506310e-01 -8.08746874e-01 -8.11113566e-02 -2.71971002e-02
-2.04774924e-02 -2.06851140e-01 3.20968360e-01 2.44289726e-01
-1.51915357e-01 -9.43564415e-01 -1.14196086e+00 9.66159031e-02
-9.66556668e-01 -1.26249540e+00 -4.73023295e-01 -6.45994842e-01
8.29565883e-01 5.08745909e-01 9.44204032e-01 -1.63335744e-02
-3.30278218e-01 7.68259883e-01 -1.54005647e-01 3.42447490e-01
1.61087170e-01 -1.25574857e-01 4.78643030e-01 4.05500829e-01
1.54570505e-01 -4.88537550e-01 -6.27270639e-01 2.07040817e-01
-9.28777754e-01 -7.60238230e-01 2.52992660e-01 7.69818664e-01
7.84996152e-01 3.63665730e-01 8.10989086e-03 -5.65316498e-01
1.14382648e+00 -7.60117352e-01 -8.32820594e-01 2.15166017e-01
-7.24934697e-01 5.36142290e-01 6.80003464e-01 -3.36812973e-01
7.12252827e-03 2.74203047e-02 3.72499019e-01 -2.34999537e-01
6.99888587e-01 8.34755719e-01 -2.45856657e-03 -3.55854571e-01
4.88124162e-01 1.87720597e-01 4.91351448e-02 -5.52687287e-01
4.49310869e-01 4.11927015e-01 1.52788386e-01 -9.35744762e-01
8.87903154e-01 3.49838585e-01 3.93572479e-01 -9.09320474e-01
-6.13399029e-01 -9.62706029e-01 -1.22083485e-01 9.82148722e-02
4.42728251e-01 -5.85961819e-01 -1.06821263e+00 -1.80110559e-01
-9.15480614e-01 2.27035061e-01 -2.61913627e-01 5.36242902e-01
-5.97338498e-01 6.31898105e-01 -4.11710352e-01 -8.05643082e-01
-1.95453182e-01 -1.07524431e+00 5.74032366e-01 -4.53600645e-01
9.50159505e-02 -7.71794438e-01 3.74430686e-01 3.47808868e-01
2.32953787e-01 2.22279616e-02 1.12281299e+00 -7.29519546e-01
-7.96103537e-01 -5.78925669e-01 -1.72710344e-01 3.81164819e-01
-1.95577472e-01 -1.72736719e-01 5.12240734e-03 -5.75199246e-01
-1.58321872e-01 -1.65080622e-01 5.63634932e-01 1.23438455e-01
1.06640589e+00 -8.63952458e-01 -1.93293825e-01 5.53655028e-01
1.54631281e+00 -1.47314861e-01 -6.79171458e-03 -2.18486134e-02
4.17779714e-01 4.64606851e-01 4.59374785e-01 6.73682690e-01
4.08813566e-01 4.73596036e-01 3.54312181e-01 3.34912926e-01
5.01446962e-01 4.29561138e-02 2.00593561e-01 1.15893722e+00
-1.63064957e-01 -1.17320791e-02 -8.08175087e-01 6.50241315e-01
-1.73357856e+00 -1.17138731e+00 -4.42138501e-03 2.44612002e+00
6.80961490e-01 -1.81312531e-01 4.64538723e-01 1.53573394e-01
6.18991673e-01 3.28468382e-01 -3.48344445e-01 -6.49697781e-01
-1.25025153e-01 6.10620141e-01 9.33119297e-01 7.05455959e-01
-8.12765479e-01 5.20692050e-01 7.37874460e+00 5.37984490e-01
-7.21040428e-01 2.32709502e-03 4.30121809e-01 -1.69954926e-01
-9.02145386e-01 6.97871596e-02 -5.41734755e-01 1.55532762e-01
9.20324683e-01 -4.52665269e-01 1.10980463e+00 8.72168064e-01
1.40140876e-02 -4.46130335e-02 -1.19801939e+00 1.21721220e+00
-1.64951846e-01 -1.45217896e+00 -1.29110768e-01 6.58943951e-01
9.51297402e-01 1.73693284e-01 3.62213880e-01 -1.89156547e-01
4.77222174e-01 -9.54834938e-01 2.31234848e-01 2.00080916e-01
6.72417462e-01 -9.58391547e-01 1.55001551e-01 -2.40458585e-02
-1.33797669e+00 -1.98386610e-01 -3.21485221e-01 8.33408814e-03
9.59364418e-03 9.21447754e-01 -7.55653381e-01 1.53972581e-01
3.83464456e-01 4.44370151e-01 -3.12839262e-02 1.18682110e+00
5.14432728e-01 5.10174453e-01 -7.32535720e-01 -4.75493163e-01
3.40051323e-01 -6.98465943e-01 7.08077431e-01 9.62911427e-01
4.35770422e-01 5.23426950e-01 4.02516335e-01 4.94303793e-01
-3.10680419e-01 5.36192060e-01 -8.69069934e-01 -6.13061666e-01
5.03089607e-01 1.17238331e+00 -5.64728737e-01 -1.78155571e-01
-3.37750077e-01 7.98676491e-01 4.06091332e-01 3.97940785e-01
-3.09307367e-01 -7.86227286e-01 1.21468902e+00 2.14088351e-01
5.40487349e-01 -7.93312192e-01 -1.77089915e-01 -1.42368805e+00
2.05268860e-01 -8.06173682e-01 5.44943690e-01 -2.62006789e-01
-1.50969076e+00 6.48549736e-01 -1.83672011e-01 -9.48227048e-01
-3.22405636e-01 -8.04657340e-01 -2.46441901e-01 8.92428696e-01
-9.16079462e-01 -3.06723207e-01 5.38870394e-01 7.45955944e-01
-1.20527059e-01 -1.13853358e-01 6.88668251e-01 2.79301286e-01
-5.56557953e-01 5.17864525e-01 3.61164033e-01 5.38642853e-02
-7.28288619e-03 -1.41216743e+00 -7.09948838e-02 8.78377438e-01
3.03233415e-01 8.69628131e-01 1.02420652e+00 -4.34039444e-01
-2.34186578e+00 -6.54795885e-01 1.02132428e+00 -1.39484070e-02
1.26327288e+00 -1.64589003e-01 -2.56516963e-01 5.46857595e-01
-9.16742310e-02 2.40185156e-01 9.24082577e-01 5.76462090e-01
-6.69652760e-01 -3.71559143e-01 -1.22243607e+00 3.99638444e-01
8.74358952e-01 -6.83100343e-01 -1.29330412e-01 8.27124298e-01
6.14640415e-01 -6.06305040e-02 -1.19850266e+00 1.40639767e-01
2.35227272e-01 -5.85015893e-01 1.11207187e+00 -9.71449375e-01
4.10963520e-02 -3.49444538e-01 -8.15488100e-01 -1.22863007e+00
-5.48007011e-01 -9.88205791e-01 -1.96453497e-01 4.54860896e-01
6.44198000e-01 -5.32524645e-01 1.10297024e+00 7.65981197e-01
2.51362801e-01 -9.66941953e-01 -1.11696768e+00 -8.16997945e-01
-2.72705376e-01 -4.73176122e-01 5.06713748e-01 9.69442189e-01
6.01295590e-01 4.46150422e-01 -4.23968166e-01 2.31450856e-01
1.00399864e+00 3.55519623e-01 4.33247298e-01 -1.03246057e+00
-6.19651675e-01 -4.65301901e-01 -5.34380734e-01 -1.25926518e+00
5.68362623e-02 -1.20853877e+00 -1.52837440e-01 -1.38208187e+00
9.75204557e-02 -1.01365936e+00 -2.84982443e-01 4.28043455e-02
3.12517822e-01 8.90074745e-02 2.86069270e-02 -1.17955729e-02
-6.88549101e-01 3.31997722e-01 9.86446261e-01 -3.79654735e-01
1.91722773e-02 1.86348319e-01 -9.17072296e-01 3.41487706e-01
4.58497465e-01 -3.60862076e-01 -2.80385643e-01 -3.37947249e-01
5.80036581e-01 5.76827168e-01 -2.37891287e-01 -3.76924098e-01
1.79132268e-01 -4.59638357e-01 -3.67120057e-01 -6.35741234e-01
4.85280871e-01 -7.17480242e-01 1.99807175e-02 3.92646402e-01
-3.44648868e-01 5.20769417e-01 -4.17169958e-01 7.11492836e-01
-9.55996215e-02 -3.09311628e-01 4.15089279e-01 -1.15592174e-01
-3.36539179e-01 1.02599263e+00 -3.07485580e-01 2.73048669e-01
6.63051009e-01 3.65439206e-02 -4.31062616e-02 -5.08377373e-01
-7.51239538e-01 1.07065611e-01 1.26432344e-01 -3.93929295e-02
7.84764230e-01 -1.43463588e+00 -6.53428078e-01 -1.62029102e-01
1.11505799e-01 -3.95940542e-01 -5.53962365e-02 8.37503254e-01
-4.11379158e-01 6.09005988e-01 4.01217878e-01 -2.52620101e-01
-1.09470320e+00 6.35788441e-01 8.79253000e-02 -2.96387047e-01
-1.23589076e-01 1.05824172e+00 -5.15365899e-01 -3.01872909e-01
2.61106938e-01 -1.52565494e-01 2.50180990e-01 -1.18007764e-01
5.09306133e-01 5.56790709e-01 7.92245381e-03 -6.85589671e-01
-4.09020364e-01 4.78659123e-01 -8.38049278e-02 -5.24308383e-01
1.44913435e+00 6.27980903e-02 -8.50299299e-01 2.07150932e-02
1.80275738e+00 4.14890081e-01 -4.15072471e-01 -5.40182412e-01
9.40301195e-02 -6.78779304e-01 1.91676635e-02 -2.22897962e-01
-1.24581099e+00 2.60467142e-01 2.36688763e-01 7.21654594e-01
9.53246117e-01 3.14271808e-01 7.61009753e-01 8.26572239e-01
5.44865787e-01 -1.00241685e+00 2.20484123e-01 5.11974335e-01
7.40218759e-01 -8.98569763e-01 1.98503733e-01 -3.90519112e-01
-2.71156698e-01 1.05520606e+00 2.99535710e-02 -4.39375609e-01
1.14641738e+00 4.33466621e-02 -5.57612062e-01 -5.11860192e-01
-8.68882656e-01 -2.51345634e-01 4.98134643e-01 2.35518098e-01
4.08218563e-01 3.67942631e-01 -7.35256612e-01 2.23744333e-01
-3.87943327e-01 -5.16595960e-01 5.73812842e-01 8.17445695e-01
-6.56705439e-01 -1.37717390e+00 -3.38190943e-01 9.23082411e-01
-4.71335471e-01 -2.92084336e-01 -2.05744341e-01 1.05039828e-01
-1.77469060e-01 9.85928774e-01 -9.13856849e-02 -8.31999779e-01
1.05680138e-01 -2.75070757e-01 5.19614100e-01 -6.84074521e-01
-1.78325742e-01 -1.17338710e-01 4.34610099e-01 -6.97744846e-01
-8.33860636e-02 -9.14382815e-01 -9.47661221e-01 -8.42394710e-01
-3.07826459e-01 8.15512419e-01 9.84209418e-01 6.87376142e-01
2.77346462e-01 -3.49743336e-01 1.34871066e+00 -3.85006040e-01
-1.03986609e+00 -5.44758320e-01 -1.11246061e+00 1.67212322e-01
3.64321887e-01 -3.91198754e-01 -5.99469244e-01 -3.23249876e-01] | [6.926605224609375, 4.669559478759766] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.