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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
98b331f0-8b0a-4b42-b12b-96920fd12527 | an-interactive-multi-task-learning-network | 1906.06906 | null | https://arxiv.org/abs/1906.06906v1 | https://arxiv.org/pdf/1906.06906v1.pdf | An Interactive Multi-Task Learning Network for End-to-End Aspect-Based Sentiment Analysis | Aspect-based sentiment analysis produces a list of aspect terms and their corresponding sentiments for a natural language sentence. This task is usually done in a pipeline manner, with aspect term extraction performed first, followed by sentiment predictions toward the extracted aspect terms. While easier to develop, such an approach does not fully exploit joint information from the two subtasks and does not use all available sources of training information that might be helpful, such as document-level labeled sentiment corpus. In this paper, we propose an interactive multi-task learning network (IMN) which is able to jointly learn multiple related tasks simultaneously at both the token level as well as the document level. Unlike conventional multi-task learning methods that rely on learning common features for the different tasks, IMN introduces a message passing architecture where information is iteratively passed to different tasks through a shared set of latent variables. Experimental results demonstrate superior performance of the proposed method against multiple baselines on three benchmark datasets. | ['Hwee Tou Ng', 'Daniel Dahlmeier', 'Wee Sun Lee', 'Ruidan He'] | 2019-06-17 | an-interactive-multi-task-learning-network-1 | https://aclanthology.org/P19-1048 | https://aclanthology.org/P19-1048.pdf | acl-2019-7 | ['aspect-term-extraction-and-sentiment'] | ['natural-language-processing'] | [ 2.83094198e-01 1.44092306e-01 -2.69055337e-01 -6.53097570e-01
-1.24851215e+00 -6.90387666e-01 1.00078189e+00 5.66410959e-01
-4.95726228e-01 5.85988700e-01 3.49477679e-01 -3.08206290e-01
2.24842951e-01 -6.46371782e-01 -4.79525715e-01 -6.42934561e-01
2.11275280e-01 5.74819505e-01 -1.96915623e-02 -1.43022671e-01
3.11874598e-01 -2.51430422e-01 -1.12334931e+00 7.83463240e-01
4.70930099e-01 8.46113682e-01 5.53884991e-02 6.62427485e-01
-7.71069467e-01 9.09495056e-01 -6.31428421e-01 -6.94421291e-01
-1.02548398e-01 -2.10055470e-01 -8.74586105e-01 2.02728316e-01
-6.39391243e-02 1.12777881e-01 6.30107939e-01 8.46562564e-01
2.87099004e-01 1.20207421e-01 7.72028446e-01 -1.16134822e+00
-3.81714970e-01 8.67452621e-01 -8.85976970e-01 -1.74183205e-01
2.13855222e-01 -1.64151847e-01 1.61432660e+00 -9.19579506e-01
2.96347350e-01 1.21370065e+00 5.78055084e-01 2.21964955e-01
-9.33041275e-01 -4.87462759e-01 6.80426419e-01 -1.48843929e-01
-6.43886805e-01 -8.30954611e-02 1.00438046e+00 -4.60737556e-01
1.25438583e+00 -1.19856097e-01 4.87725824e-01 1.01755619e+00
4.19381708e-01 1.15034640e+00 1.20346534e+00 -3.95623088e-01
1.36637270e-01 5.88222802e-01 5.94190538e-01 5.56493819e-01
1.77644491e-01 -5.58070600e-01 -8.19322824e-01 -3.32222551e-01
-5.62641136e-02 1.28435781e-02 1.93835452e-01 -2.00175941e-01
-1.13427603e+00 9.15561736e-01 -3.71003002e-02 3.58595997e-01
-5.45577347e-01 1.37206480e-01 8.07273388e-01 4.85563487e-01
1.05567169e+00 3.39521885e-01 -9.94758904e-01 -8.42435136e-02
-9.69796598e-01 1.45874679e-01 9.91109192e-01 8.96860123e-01
1.04314458e+00 2.63916887e-03 -3.69665861e-01 6.39090717e-01
6.64122343e-01 2.89228946e-01 5.79181850e-01 -2.21655294e-01
6.94783151e-01 8.73854935e-01 2.18939800e-02 -7.28764236e-01
-4.27719504e-01 -4.55752850e-01 -5.56628287e-01 1.40845746e-01
9.42356661e-02 -6.14522040e-01 -8.76213729e-01 1.47537172e+00
4.29898173e-01 -1.61268935e-01 3.79427999e-01 3.87726694e-01
9.52220976e-01 7.23124743e-01 3.86363268e-01 -7.65270665e-02
1.72215474e+00 -1.45642269e+00 -8.14625919e-01 -6.97382271e-01
7.61894166e-01 -1.17616725e+00 8.54798317e-01 3.87833685e-01
-8.58433902e-01 -3.26488376e-01 -9.83206272e-01 -1.22273475e-01
-5.79771459e-01 2.50419497e-01 8.00679743e-01 5.84675372e-01
-9.84512806e-01 7.58379325e-02 -9.48584557e-01 -1.07784025e-01
3.37986976e-01 2.87278891e-01 -1.92718759e-01 2.01054499e-01
-8.98615122e-01 8.09800446e-01 2.63037801e-01 7.60006458e-02
-9.01362658e-01 -5.77263355e-01 -1.13237488e+00 1.00945868e-01
4.20328140e-01 -8.44711602e-01 1.42168581e+00 -1.19752061e+00
-1.42411232e+00 6.40131533e-01 -6.25975251e-01 -2.26718411e-01
1.07000984e-01 -7.15828717e-01 -3.80091928e-02 -1.38355657e-01
2.23749995e-01 3.60642076e-01 1.07604253e+00 -1.35634935e+00
-7.25024700e-01 -2.89548367e-01 4.11151916e-01 3.88135791e-01
-5.29154122e-01 4.17657524e-01 -4.91340667e-01 -4.98485208e-01
-2.48493925e-01 -8.47830951e-01 -5.88430822e-01 -3.17660660e-01
-6.23134553e-01 -4.04109985e-01 8.23772609e-01 -3.76098394e-01
8.45653355e-01 -1.76992428e+00 1.31601766e-01 8.94877687e-02
1.62501290e-01 1.15386911e-01 -2.96496123e-01 6.98761761e-01
-7.87368137e-03 4.67149056e-02 -2.08660126e-01 -1.05830193e+00
3.41846906e-02 -2.05040947e-02 -2.31047764e-01 1.51002064e-01
4.13224697e-01 9.08854663e-01 -9.75122273e-01 -4.34105843e-01
-6.35278374e-02 6.00736082e-01 -3.92622501e-01 3.26712281e-01
-5.67576706e-01 3.24111015e-01 -8.92502367e-01 5.20962954e-01
3.77968937e-01 -3.54855627e-01 -5.27339168e-02 -1.26602888e-01
-4.02266681e-02 6.35879099e-01 -1.03492582e+00 1.74911368e+00
-1.08824599e+00 4.13276762e-01 -6.13654554e-02 -1.06680048e+00
9.54268277e-01 7.88718224e-01 4.98464435e-01 -2.72402495e-01
1.30162500e-02 -2.28157290e-03 -2.81196862e-01 -4.61836964e-01
6.11015320e-01 -5.35748780e-01 -5.02260625e-01 1.20356691e+00
5.11376381e-01 -1.50894538e-01 2.68086016e-01 1.94996223e-01
8.92228067e-01 2.98914343e-01 5.94576418e-01 -2.07133099e-01
7.59635031e-01 9.84526202e-02 5.20586967e-01 5.92346966e-01
1.37711927e-01 3.06812316e-01 8.66709173e-01 -5.22695243e-01
-7.92417943e-01 -5.95977783e-01 1.44824252e-01 1.43144858e+00
-1.39962018e-01 -8.93501282e-01 -3.09566110e-01 -9.68428612e-01
-3.87533158e-01 6.96436703e-01 -7.78685570e-01 1.90588996e-01
-2.63010621e-01 -1.05007446e+00 2.41671816e-01 4.03542519e-01
1.97254062e-01 -1.22634220e+00 -5.79826832e-01 2.49433264e-01
-2.25382879e-01 -1.19652700e+00 -2.47960508e-01 5.58950663e-01
-8.24764550e-01 -8.48740757e-01 -4.21898931e-01 -6.87879860e-01
6.91351593e-01 1.38341293e-01 1.44275296e+00 -1.19131960e-01
1.06024928e-01 3.16464931e-01 -6.01053953e-01 -8.58847380e-01
-2.30825067e-01 4.37484115e-01 -4.10063595e-01 3.36420685e-01
7.23659813e-01 -4.50531334e-01 -3.21879417e-01 -2.50134945e-01
-9.84750688e-01 2.12388709e-01 9.26293910e-01 7.87675798e-01
5.21384299e-01 1.69208106e-02 5.81545055e-01 -1.55456007e+00
7.42405653e-01 -5.51472962e-01 -1.89394072e-01 2.85043091e-01
-6.55748427e-01 2.94864088e-01 6.98932409e-01 -1.11720860e-01
-1.50328290e+00 5.54811656e-02 -1.35586709e-01 2.67318517e-01
-1.85082898e-01 9.41960037e-01 -1.44319937e-01 4.91159588e-01
1.60433471e-01 2.37176523e-01 -4.01513129e-01 -4.10379320e-01
4.83898968e-01 5.51284492e-01 -1.57731891e-01 -5.19393384e-01
8.32125247e-01 4.64648217e-01 -2.32340619e-01 -6.73241317e-01
-1.45873260e+00 -8.33220243e-01 -7.56301880e-01 -2.69756112e-02
8.83696556e-01 -1.13003933e+00 -2.50082880e-01 4.72712785e-01
-1.35367954e+00 -2.60380656e-03 -1.47189036e-01 3.23503971e-01
-4.04188246e-01 -2.87157204e-02 -5.65653503e-01 -8.48124802e-01
-8.71543229e-01 -1.17380667e+00 1.47918642e+00 4.15400863e-02
-3.81108403e-01 -1.52217913e+00 3.13459605e-01 5.57650566e-01
4.65042233e-01 3.87090035e-02 1.04011917e+00 -1.16990006e+00
-4.22216773e-01 -4.72996920e-01 -6.55884482e-03 3.33640933e-01
5.31764627e-01 -1.24405809e-01 -1.18305194e+00 -1.25133738e-01
2.83093095e-01 -6.98326528e-01 9.21516716e-01 2.71132171e-01
6.23687685e-01 -3.86606365e-01 -2.57997274e-01 2.30276182e-01
1.60621154e+00 -9.58862826e-02 1.58135593e-01 7.13996172e-01
8.76053631e-01 9.52522933e-01 6.05679691e-01 3.59050870e-01
8.45609009e-01 3.37965697e-01 5.08449934e-02 -3.47824603e-01
2.15735212e-01 8.11955556e-02 6.66913688e-01 1.04493022e+00
1.70199156e-01 -2.26312324e-01 -7.64242887e-01 7.59579718e-01
-1.89727008e+00 -6.72641575e-01 -2.51211852e-01 1.73557901e+00
9.95659411e-01 3.28998953e-01 -1.10890128e-01 -7.71907642e-02
2.30188906e-01 6.84318244e-01 -2.66826808e-01 -8.44951332e-01
8.49537272e-03 1.62764177e-01 1.52006401e-02 5.87666571e-01
-1.27677357e+00 9.66178238e-01 5.54970026e+00 5.66993415e-01
-9.66532171e-01 3.89822006e-01 5.75710595e-01 -1.44358218e-01
-7.43427098e-01 3.75170141e-01 -8.54173481e-01 -1.88719314e-02
9.23769772e-01 -3.65290731e-01 -3.81720811e-01 9.20953751e-01
2.00503156e-01 -2.42735431e-01 -1.04255855e+00 3.28433007e-01
3.43345016e-01 -8.53276551e-01 3.17721456e-01 -2.94779599e-01
9.07825887e-01 1.56634375e-01 -3.54087502e-02 4.81773168e-01
5.19495726e-01 -7.65522301e-01 5.40056169e-01 3.54491442e-01
1.22854322e-01 -7.44469821e-01 1.10329723e+00 3.23302448e-01
-1.37812901e+00 2.56940126e-01 -9.00837779e-02 -1.13276564e-01
4.27975804e-01 8.59201372e-01 -7.50719249e-01 7.28968382e-01
4.85136777e-01 8.64574850e-01 -4.91735131e-01 6.76399171e-01
-5.64801991e-01 6.78801119e-01 1.01236776e-02 -2.01084062e-01
5.55632055e-01 -1.49718150e-01 4.74835992e-01 1.54454172e+00
4.06556064e-03 -4.33751345e-01 2.99717695e-01 3.95434529e-01
3.10757160e-02 5.57570875e-01 -6.97657824e-01 -7.78906122e-02
-2.40861058e-01 1.84946907e+00 -8.14073265e-01 -5.74515343e-01
-9.94327843e-01 8.21260989e-01 3.37183684e-01 4.07970577e-01
-3.51738334e-01 -3.09191227e-01 5.28685510e-01 -4.34238195e-01
5.33340931e-01 -2.30887413e-01 -5.72050035e-01 -1.35045421e+00
2.57646382e-01 -8.85252178e-01 3.57363373e-01 -4.30942148e-01
-1.18945599e+00 9.69869435e-01 -1.91441253e-01 -1.15828979e+00
-4.97670799e-01 -4.31706101e-01 -1.03951919e+00 1.08284223e+00
-2.11776090e+00 -1.70818043e+00 1.87023476e-01 4.77759480e-01
9.32731867e-01 -1.90321326e-01 1.05371618e+00 -5.43392599e-02
-3.48329306e-01 3.57018620e-01 -1.90621853e-01 1.29594430e-01
8.00804436e-01 -1.70082343e+00 4.70707685e-01 8.22715461e-01
3.09959292e-01 7.46124268e-01 7.82662392e-01 -4.85892922e-01
-1.15146279e+00 -1.08559787e+00 1.39534712e+00 -5.39560616e-01
9.47113693e-01 -4.31029826e-01 -6.67506158e-01 9.30783033e-01
6.47119820e-01 -4.15116161e-01 1.33685029e+00 8.19070756e-01
-4.67040926e-01 -6.95408806e-02 -5.29580295e-01 2.98434407e-01
2.18817756e-01 -5.88450491e-01 -7.89871693e-01 4.91135687e-01
7.30712712e-01 -8.02933425e-02 -6.31084502e-01 2.03536823e-01
3.88115972e-01 -6.36304200e-01 5.77724457e-01 -7.10435331e-01
8.22735071e-01 -1.54903829e-01 -5.99349327e-02 -1.57740831e+00
1.55283943e-01 -4.29089636e-01 -1.29786348e-02 1.60485518e+00
1.12820971e+00 -3.25765848e-01 5.61189950e-01 5.52010238e-01
-1.19681679e-01 -9.36069191e-01 -4.78378385e-01 -2.37679437e-01
2.71488968e-02 -8.73778403e-01 3.91216129e-01 6.22260153e-01
5.95353879e-02 1.12164271e+00 -5.98506391e-01 1.33566678e-01
5.76771080e-01 7.54595280e-01 7.47057259e-01 -1.03422582e+00
-4.76257771e-01 -3.14732641e-01 -7.82544434e-04 -7.80435085e-01
3.97123456e-01 -9.64941680e-01 1.38914824e-01 -1.88807070e+00
5.08633316e-01 -3.17587793e-01 -4.52655226e-01 7.32360661e-01
-6.76230907e-01 3.26243043e-03 1.00160148e-02 1.19620003e-02
-1.02174056e+00 8.32380533e-01 1.04958868e+00 -3.03769350e-01
-2.32315302e-01 3.23399752e-01 -1.09013021e+00 1.02253640e+00
6.71034157e-01 -8.65594387e-01 -5.06900489e-01 -7.00226128e-01
6.14720285e-01 -1.47942886e-01 -1.79859444e-01 -5.45318663e-01
4.19206530e-01 5.65251224e-02 2.37040684e-01 -6.26989067e-01
3.70080709e-01 -7.41100132e-01 -4.15988892e-01 4.33123820e-02
-5.35685599e-01 1.94089606e-01 1.83450937e-01 5.82796633e-01
-5.70340693e-01 -3.71178120e-01 2.31060907e-01 -2.93184191e-01
-4.42203909e-01 2.32204393e-01 -4.76891994e-01 -9.08812881e-02
8.77765000e-01 2.45587960e-01 -1.12781525e-01 -3.66189957e-01
-5.63316047e-01 4.66729432e-01 6.08957559e-02 6.01675451e-01
3.74188215e-01 -9.41838443e-01 -7.98909009e-01 2.62183677e-02
3.81091863e-01 1.20687492e-01 1.25211015e-01 6.96994066e-01
1.61485001e-01 5.44908106e-01 1.92882299e-01 -4.78138655e-01
-1.19622374e+00 3.65841538e-01 -2.41608713e-02 -1.00349450e+00
-4.15046602e-01 9.09897625e-01 2.58815050e-01 -7.66246796e-01
-5.36717586e-02 -1.42241687e-01 -8.03338587e-01 6.09453559e-01
5.74482977e-01 -3.63710642e-01 2.65607864e-01 -6.32095754e-01
-2.14018956e-01 6.32520556e-01 -5.35346448e-01 -3.24178308e-01
1.63989866e+00 -3.15931946e-01 -4.35381889e-01 1.03122485e+00
1.22997081e+00 -2.56301407e-02 -9.60046351e-01 -5.99087000e-01
4.55083460e-01 -9.50285196e-02 8.10076222e-02 -6.15454197e-01
-1.03213465e+00 8.72327566e-01 -1.18118942e-01 1.91817388e-01
1.10996342e+00 1.62108704e-01 8.31980169e-01 5.08848667e-01
1.47598162e-01 -1.00307238e+00 2.84410864e-01 6.49228513e-01
6.28410578e-01 -1.55909526e+00 1.10150732e-01 -1.99600816e-01
-9.79982138e-01 1.03345478e+00 6.18796587e-01 -7.67923892e-03
9.51991737e-01 4.26934123e-01 5.66842973e-01 -5.36554754e-01
-1.49076259e+00 -3.34864378e-01 2.79659629e-01 2.69291490e-01
9.37581718e-01 -1.58671811e-01 -3.95250678e-01 8.30391407e-01
-1.50535777e-01 -3.43710154e-01 4.60180312e-01 1.05226851e+00
-2.63093740e-01 -1.66086960e+00 -6.11163676e-02 6.56994641e-01
-9.53427553e-01 -4.76525545e-01 -2.70438820e-01 2.81787753e-01
-2.02721611e-01 1.05182731e+00 -1.91198185e-01 -1.20708279e-01
1.94487050e-01 3.69223744e-01 -1.92683756e-01 -1.19233668e+00
-1.00249445e+00 2.49823406e-01 3.42020512e-01 -4.01323706e-01
-8.42369258e-01 -8.08313549e-01 -9.33667123e-01 3.38532954e-01
-2.16188297e-01 4.09232497e-01 9.24072087e-01 1.56067884e+00
1.67085215e-01 7.86198616e-01 7.01247752e-01 -6.42118871e-01
-1.78788424e-01 -1.06477797e+00 -3.96999270e-01 2.15081736e-01
4.96948957e-01 -2.71224767e-01 -1.97056711e-01 2.43965149e-01] | [11.393980026245117, 6.684563636779785] |
8802ae0b-d06c-45e0-9f97-1be8d6248b21 | fine-grained-angular-contrastive-learning | 2012.03515 | null | https://arxiv.org/abs/2012.03515v3 | https://arxiv.org/pdf/2012.03515v3.pdf | Fine-grained Angular Contrastive Learning with Coarse Labels | Few-shot learning methods offer pre-training techniques optimized for easier later adaptation of the model to new classes (unseen during training) using one or a few examples. This adaptivity to unseen classes is especially important for many practical applications where the pre-trained label space cannot remain fixed for effective use and the model needs to be "specialized" to support new categories on the fly. One particularly interesting scenario, essentially overlooked by the few-shot literature, is Coarse-to-Fine Few-Shot (C2FS), where the training classes (e.g. animals) are of much `coarser granularity' than the target (test) classes (e.g. breeds). A very practical example of C2FS is when the target classes are sub-classes of the training classes. Intuitively, it is especially challenging as (both regular and few-shot) supervised pre-training tends to learn to ignore intra-class variability which is essential for separating sub-classes. In this paper, we introduce a novel 'Angular normalization' module that allows to effectively combine supervised and self-supervised contrastive pre-training to approach the proposed C2FS task, demonstrating significant gains in a broad study over multiple baselines and datasets. We hope that this work will help to pave the way for future research on this new, challenging, and very practical topic of C2FS classification. | ['Leonid Karlinsky', 'Raja Giryes', 'Rogerio Feris', 'Ori Shahar', 'Kate Saenko', 'Eli Schwartz', 'Guy Bukchin'] | 2020-12-07 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Bukchin_Fine-Grained_Angular_Contrastive_Learning_With_Coarse_Labels_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Bukchin_Fine-Grained_Angular_Contrastive_Learning_With_Coarse_Labels_CVPR_2021_paper.pdf | cvpr-2021-1 | ['learning-with-coarse-labels'] | ['computer-vision'] | [ 5.25423467e-01 8.01666453e-02 -4.06623870e-01 -7.13296890e-01
-4.91045684e-01 -5.23325384e-01 7.84718454e-01 4.69681531e-01
-5.70387304e-01 8.24392974e-01 -5.25996648e-02 2.73816943e-01
-1.98366940e-01 -9.26949918e-01 -6.18518531e-01 -7.73922801e-01
-1.60802707e-01 6.33292556e-01 7.12025225e-01 -4.14550483e-01
-1.03140771e-02 5.08127749e-01 -2.17990851e+00 4.35485303e-01
5.05895972e-01 8.46731603e-01 3.17932636e-01 6.16587996e-01
-1.86213955e-01 3.29968154e-01 -6.79228544e-01 -3.66058022e-01
1.85905933e-01 -5.03386796e-01 -6.66572809e-01 9.48339477e-02
5.47122777e-01 1.99650615e-01 2.33108848e-01 1.09468400e+00
3.83951336e-01 8.46765220e-01 8.51890445e-01 -1.03620481e+00
-3.59743267e-01 2.90789425e-01 -4.14185882e-01 4.90053207e-01
-1.55672505e-01 1.37522928e-02 8.72743011e-01 -6.91474259e-01
7.69614160e-01 1.07578421e+00 6.99067593e-01 8.28538060e-01
-1.39200866e+00 -5.00454545e-01 2.31715247e-01 5.46651304e-01
-1.25299048e+00 -5.46601176e-01 5.60092926e-01 -4.52515662e-01
7.99300790e-01 4.49688464e-01 3.94504547e-01 1.17071724e+00
-7.42834136e-02 5.70162237e-01 1.02727985e+00 -6.44553602e-01
8.27661633e-01 6.48718596e-01 5.89247048e-01 1.28368452e-01
1.59753859e-01 3.18417579e-01 -3.11318964e-01 8.68270248e-02
2.36622065e-01 2.43796930e-01 -6.79951683e-02 -6.75357759e-01
-7.74515688e-01 1.00430214e+00 5.36099851e-01 6.51498199e-01
-1.28063679e-01 -1.51533127e-01 6.06383741e-01 3.38442415e-01
6.91749573e-01 5.97404778e-01 -6.06941581e-01 -1.02344885e-01
-1.11970425e+00 1.19898945e-01 6.90382242e-01 8.61120582e-01
9.45928872e-01 -1.10855438e-01 -3.16417009e-01 1.26093411e+00
-2.17542037e-01 1.20245129e-01 7.78364956e-01 -5.11510611e-01
-1.65255889e-01 3.12271416e-01 3.20106111e-02 -6.04262233e-01
-4.02011007e-01 -6.99436367e-01 -7.14357376e-01 2.34559342e-01
3.82306159e-01 1.77674722e-02 -1.22749174e+00 1.69256103e+00
5.50777614e-01 4.37362611e-01 3.77103314e-02 5.90862632e-01
7.50834823e-01 5.83655477e-01 3.07490647e-01 -2.90347427e-01
1.29258704e+00 -9.92203176e-01 -4.69702393e-01 -5.65013230e-01
6.01762295e-01 -5.30963540e-01 1.04889834e+00 1.03884183e-01
-4.71309543e-01 -7.96604097e-01 -1.14276075e+00 3.42873514e-01
-9.83330309e-01 -2.98783153e-01 7.29333043e-01 5.51798880e-01
-5.09051383e-01 8.30268502e-01 -5.56509614e-01 -8.97887349e-01
6.79810405e-01 2.24009290e-01 -3.63923162e-01 -2.97541529e-01
-1.25159073e+00 9.43498611e-01 7.03917205e-01 -3.44844967e-01
-8.45234215e-01 -8.63368094e-01 -8.01987350e-01 1.75898403e-01
7.78274298e-01 -2.40627289e-01 1.19796669e+00 -1.07362282e+00
-1.22232258e+00 1.11942017e+00 1.13203386e-02 -4.31330353e-01
2.95305282e-01 1.81963611e-02 -5.04427314e-01 -8.28132108e-02
1.15637712e-01 6.60452187e-01 1.09551942e+00 -1.19112861e+00
-7.55379558e-01 -2.71751165e-01 1.41101912e-01 5.66270873e-02
-3.58234972e-01 4.51457798e-02 -5.54538295e-02 -8.16504776e-01
-3.52819234e-01 -8.24613571e-01 -1.12402447e-01 -3.12679261e-02
2.06609014e-02 -3.67146820e-01 9.30564225e-01 -2.79611573e-02
1.12471962e+00 -2.26899171e+00 1.03204288e-01 -1.26807526e-01
4.76073619e-04 6.84536278e-01 -2.02093765e-01 3.13234687e-01
-3.70959163e-01 -2.57862359e-01 -2.53531337e-01 -1.14031591e-01
-6.28592223e-02 4.37177151e-01 -1.26499549e-01 3.53340507e-01
2.70697683e-01 7.14569449e-01 -1.16994750e+00 -2.36864060e-01
3.37595969e-01 3.15923721e-01 -2.46701524e-01 1.95711568e-01
-6.31323978e-02 1.50949106e-01 2.21164036e-03 4.86589134e-01
3.94399494e-01 -1.38336688e-01 -1.79329678e-01 -3.34351920e-02
9.92728025e-02 -2.95168489e-01 -1.30414724e+00 1.36367643e+00
-4.67046499e-01 6.34865880e-01 -3.55998546e-01 -1.47092998e+00
7.72958398e-01 9.21792835e-02 2.48956919e-01 -2.76629806e-01
2.79129863e-01 1.34347258e-02 6.21528812e-02 -3.53096545e-01
1.59801558e-01 -6.58333123e-01 -9.45318863e-02 2.00887188e-01
5.96448362e-01 -1.51136860e-01 5.28631091e-01 -8.48163888e-02
9.21836436e-01 4.16719355e-02 7.86909103e-01 -3.58911723e-01
2.92571962e-01 1.95351634e-02 6.53902292e-01 9.87336278e-01
-4.70722139e-01 7.66236305e-01 9.02495757e-02 -5.18015444e-01
-8.29597116e-01 -9.95521784e-01 -5.35593808e-01 1.67748547e+00
-3.56132276e-02 -1.55557469e-01 -6.41844451e-01 -9.89614546e-01
8.74958467e-03 9.93569195e-01 -1.07029212e+00 -6.04195237e-01
-2.06672102e-01 -1.01361251e+00 -1.64616220e-02 5.31725049e-01
1.62006542e-01 -9.44199920e-01 -7.62215436e-01 4.71882671e-01
2.44290769e-01 -7.90277481e-01 -1.75358355e-01 7.95044482e-01
-7.65097439e-01 -1.03766418e+00 -8.18171442e-01 -8.13808858e-01
6.38129890e-01 4.02888715e-01 1.03104365e+00 -9.74188820e-02
-4.71242607e-01 2.14553535e-01 -7.70472467e-01 -7.61689186e-01
-2.28796199e-01 -2.09067147e-02 1.37517363e-01 2.42491513e-01
8.51365328e-01 -5.00563860e-01 -2.94176549e-01 5.10756254e-01
-1.03639162e+00 -2.49844164e-01 3.43337268e-01 1.03479385e+00
4.88846779e-01 3.10096383e-01 6.63114548e-01 -1.34959900e+00
2.45820001e-01 -5.98553061e-01 -2.44232237e-01 4.15563583e-01
-4.43689317e-01 -1.30252257e-01 7.35801160e-01 -8.94048452e-01
-9.82811809e-01 -1.33441165e-01 8.11594203e-02 -2.72618532e-01
-5.54993868e-01 2.34563693e-01 -7.97211304e-02 -1.24448724e-01
1.15853584e+00 -1.48673013e-01 -2.70508885e-01 -5.98897159e-01
3.23810726e-01 7.46349037e-01 5.05459547e-01 -4.46362555e-01
7.74355471e-01 4.02549773e-01 1.24088945e-02 -1.14546144e+00
-1.15243316e+00 -8.60023797e-01 -1.06748497e+00 -1.91539824e-01
6.86878979e-01 -6.62893474e-01 7.14248717e-02 2.68137395e-01
-6.82309449e-01 -4.01195109e-01 -1.00902140e+00 4.44916159e-01
-3.63105267e-01 1.20987743e-01 -1.86266348e-01 -6.22901261e-01
8.37858859e-03 -6.87873125e-01 9.56671476e-01 6.02847338e-01
-4.23988134e-01 -1.18250024e+00 2.72780031e-01 7.23026469e-02
5.21694303e-01 2.30151787e-01 8.04030538e-01 -1.12752926e+00
1.53138712e-01 -6.23985171e-01 -1.92246474e-02 5.14597476e-01
2.82884508e-01 -1.51507691e-01 -1.27435327e+00 -4.70670015e-01
9.05371755e-02 -5.11030138e-01 9.86988008e-01 2.28061467e-01
1.00920033e+00 1.16256744e-01 -3.73090446e-01 4.69712377e-01
1.35905743e+00 1.70854107e-01 4.46196288e-01 1.81310996e-01
3.68412912e-01 7.73818493e-01 9.46762860e-01 2.93641061e-01
-9.00861546e-02 9.32857454e-01 3.03839166e-02 -5.54986596e-02
-3.12093049e-01 1.80339247e-01 4.40597208e-03 5.60208619e-01
-1.96274683e-01 -8.29901360e-03 -7.46592641e-01 5.78727603e-01
-1.76578581e+00 -1.20133269e+00 3.48855227e-01 2.47001362e+00
8.27543855e-01 2.62225151e-01 7.26949871e-02 1.54043645e-01
9.44595575e-01 1.59296636e-02 -6.47759795e-01 -4.04955894e-01
6.45877793e-02 6.19161308e-01 2.40621433e-01 2.55684555e-01
-1.50715828e+00 8.53357315e-01 5.58072472e+00 1.13912833e+00
-1.08289564e+00 3.14365566e-01 5.56620538e-01 -2.20911503e-01
3.72109711e-01 -5.01293875e-02 -1.04840314e+00 5.12134552e-01
1.02568543e+00 -1.94720432e-01 3.57958078e-01 9.55430031e-01
-1.39421239e-01 -2.06472695e-01 -1.35407352e+00 8.69348288e-01
2.46010900e-01 -1.00380063e+00 -3.57095152e-01 -2.67255425e-01
8.24878573e-01 1.64849702e-02 -1.75825283e-01 9.70305324e-01
1.68780029e-01 -6.57521665e-01 3.41814488e-01 3.89405966e-01
8.81853044e-01 -7.11451650e-01 8.47228467e-01 5.03992319e-01
-1.07468402e+00 -2.74111509e-01 -8.58564436e-01 -7.01966733e-02
-6.11074269e-02 5.56571007e-01 -6.44035995e-01 2.91905165e-01
8.31284642e-01 7.21658289e-01 -8.59835207e-01 1.17702448e+00
-3.62488143e-02 5.03488541e-01 -2.82208830e-01 -1.32527143e-01
3.00701410e-01 1.12825155e-01 4.61830616e-01 1.23398280e+00
1.00999795e-01 1.70382142e-01 8.56247768e-02 3.78527790e-01
1.18053585e-01 2.25556821e-01 -5.31380653e-01 -2.24651881e-02
2.28811502e-01 1.36628640e+00 -1.15520298e+00 -6.52906716e-01
-4.49056268e-01 9.05787289e-01 4.28461969e-01 2.87043244e-01
-7.18943954e-01 -8.02408338e-01 6.07919931e-01 2.68945426e-01
5.87040842e-01 3.38117570e-01 1.57502905e-01 -1.31250453e+00
-4.55810130e-01 -6.75595939e-01 6.67996764e-01 -3.97706062e-01
-1.66715181e+00 4.13979769e-01 3.98908019e-01 -1.25817776e+00
-2.65650868e-01 -6.71456337e-01 -6.35229230e-01 4.82113153e-01
-1.40373385e+00 -1.00271225e+00 -2.66803533e-01 4.10538822e-01
8.19175303e-01 -3.05181742e-01 1.04380178e+00 2.51193076e-01
-3.78728688e-01 5.90588450e-01 4.56801176e-01 -1.03275828e-01
9.41247225e-01 -1.29025936e+00 4.47252356e-02 6.70413852e-01
3.06446552e-01 5.79025805e-01 9.04349983e-01 -5.73622823e-01
-7.64865160e-01 -1.36284983e+00 7.51460373e-01 -4.69392538e-01
5.16044796e-01 -5.60724795e-01 -1.19602346e+00 5.07549047e-01
-1.71853051e-01 5.02419770e-01 1.05360472e+00 3.85136247e-01
-2.34956980e-01 -3.59793276e-01 -1.34647632e+00 2.42699012e-01
8.23048532e-01 -3.70720327e-01 -8.85034084e-01 5.33960700e-01
4.39089715e-01 -2.99014952e-02 -7.03691006e-01 4.87906337e-01
4.02617246e-01 -8.80134583e-01 8.55900168e-01 -9.73206401e-01
1.10463798e-01 -9.14814398e-02 -3.09203446e-01 -1.69244492e+00
-4.67824906e-01 -2.28561834e-01 -2.10482612e-01 1.22652936e+00
5.49269617e-02 -4.94472444e-01 8.12131047e-01 3.66214603e-01
-8.29034597e-02 -6.20559573e-01 -9.11058247e-01 -1.15431988e+00
4.98196036e-02 -3.26602399e-01 4.52409446e-01 1.18504536e+00
-9.09119993e-02 5.33845901e-01 -3.55692953e-01 -2.12094396e-01
5.95067263e-01 3.24133970e-02 7.40231633e-01 -1.62051558e+00
-4.82436717e-01 -3.48846942e-01 -8.61813307e-01 -4.89523649e-01
-8.86172876e-02 -9.76418495e-01 2.80080795e-01 -1.07797945e+00
3.07772994e-01 -3.50698650e-01 -6.13686681e-01 5.59745073e-01
-3.54418218e-01 4.71095443e-01 1.63298234e-01 -4.32224758e-02
-6.35814011e-01 3.96724910e-01 9.90053654e-01 -1.69646025e-01
-3.27848852e-01 3.96067023e-01 -6.30019486e-01 5.04460454e-01
7.17997611e-01 -6.79483294e-01 -5.19829035e-01 1.84259415e-01
-2.78420448e-01 -5.71804702e-01 1.75877213e-01 -1.17079413e+00
1.28078070e-02 -3.75767142e-01 4.45551753e-01 -1.47603840e-01
3.13553870e-01 -7.56529033e-01 6.92614764e-02 2.57408619e-01
-2.74311841e-01 -6.50274456e-01 2.65706599e-01 7.50269234e-01
-7.88775161e-02 -7.27276266e-01 1.33718705e+00 -3.36003840e-01
-1.09444714e+00 2.61901617e-01 -1.19986989e-01 2.85075933e-01
1.44321716e+00 -3.75005364e-01 -2.23073050e-01 -1.03636488e-01
-1.12226677e+00 1.34396916e-02 4.28153336e-01 5.14358759e-01
2.80437827e-01 -1.23949206e+00 -5.12239099e-01 2.32888907e-01
6.03996933e-01 -7.40054250e-02 2.98818976e-01 6.15258098e-01
-1.31344900e-01 2.28583053e-01 -4.16751981e-01 -6.07836843e-01
-1.29334593e+00 8.71140718e-01 2.27970093e-01 -1.04728751e-01
-5.42698741e-01 1.04881048e+00 3.80813122e-01 -3.78378510e-01
3.44876677e-01 -8.65116343e-03 -4.66571957e-01 5.55410683e-01
8.28750670e-01 4.69847500e-01 1.59452379e-01 -6.60143137e-01
-1.60808757e-01 4.01920110e-01 -3.49961936e-01 3.79669785e-01
1.69669592e+00 3.77733335e-02 2.21715972e-01 9.87852514e-01
1.21496153e+00 -3.73160571e-01 -1.26010966e+00 -4.51129764e-01
1.30665228e-01 -4.93935704e-01 -1.06876463e-01 -7.43551016e-01
-6.83964372e-01 1.08032823e+00 8.94113183e-01 2.63189137e-01
9.83300865e-01 1.00142762e-01 4.45586115e-01 4.17385250e-01
4.11214709e-01 -1.48086727e+00 3.51018971e-03 3.88727814e-01
5.06104469e-01 -1.45692825e+00 1.26159742e-01 -1.38618544e-01
-5.16346157e-01 1.19039059e+00 5.54692149e-01 -8.59037861e-02
7.97587037e-01 -1.58564776e-01 -6.27124235e-02 -2.10768491e-01
-6.47584200e-01 -4.74253058e-01 6.21888936e-01 8.57329905e-01
1.30809009e-01 1.64476126e-01 -6.48175329e-02 5.69172263e-01
4.41158041e-02 -1.40304063e-02 3.75719786e-01 1.09612954e+00
-7.98006356e-01 -1.05769598e+00 -1.54100955e-01 9.27215815e-01
-2.36892864e-01 1.27790153e-01 -2.09231421e-01 8.02107930e-01
6.04613423e-01 7.67953455e-01 4.35527749e-02 -1.74636945e-01
4.02114451e-01 4.37038273e-01 5.14271736e-01 -1.41825581e+00
-4.09900486e-01 -1.37726098e-01 -1.49437487e-01 -3.50938827e-01
-4.74941790e-01 -6.44451439e-01 -8.12062562e-01 2.81306449e-02
-5.19251943e-01 1.15111314e-01 4.50841457e-01 1.06082952e+00
-3.99884991e-02 5.38036942e-01 6.10019982e-01 -1.05246341e+00
-5.53785086e-01 -1.01319563e+00 -8.17052424e-01 7.26473153e-01
2.42848247e-01 -1.17562377e+00 -5.88548124e-01 8.80927667e-02] | [9.966852188110352, 2.975789785385132] |
29dff4d1-6737-4dea-bf57-d40f45a6a9f1 | adaptive-risk-sensitive-model-predictive | 2009.01090 | null | https://arxiv.org/abs/2009.01090v2 | https://arxiv.org/pdf/2009.01090v2.pdf | Adaptive Risk Sensitive Model Predictive Control with Stochastic Search | We present a general framework for optimizing the Conditional Value-at-Risk for dynamical systems using stochastic search. The framework is capable of handling the uncertainty from the initial condition, stochastic dynamics, and uncertain parameters in the model. The algorithm is compared against a risk-sensitive distributional reinforcement learning framework and demonstrates outperformance on a pendulum and cartpole with stochastic dynamics. We also showcase the applicability of the framework to robotics as an adaptive risk-sensitive controller by optimizing with respect to the fully nonlinear belief provided by a particle filter on a pendulum, cartpole, and quadcopter in simulation. | ['Evangelos A. Theodorou', 'Camilo A. Duarte', 'Keuntaek Lee', 'Oswin So', 'Ziyi Wang'] | 2020-09-02 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-3.41051698e-01 5.73496930e-02 -2.72819940e-02 -5.39154410e-02
-8.31954777e-01 -4.94253099e-01 7.17010140e-01 -1.40534073e-01
-6.76362753e-01 1.36589015e+00 -1.99474931e-01 -2.40758806e-01
-7.52107978e-01 -5.88194132e-01 -6.31960809e-01 -9.26825047e-01
-3.44999611e-01 6.45298243e-01 1.29313126e-01 -4.73603874e-01
3.33849579e-01 3.83617640e-01 -1.42381513e+00 -9.56990600e-01
7.39399195e-01 9.22934651e-01 5.39240897e-01 8.23769450e-01
7.15250254e-01 5.06914437e-01 -6.46985412e-01 9.98152420e-02
2.52894431e-01 2.28102654e-01 4.68213819e-02 -3.38799596e-01
-1.32749379e-01 -1.02078244e-01 1.24451205e-01 1.33447099e+00
4.82618839e-01 7.15827167e-01 9.17611957e-01 -1.07892966e+00
-5.70938140e-02 2.79853165e-01 -1.14355668e-01 3.20409596e-01
1.03843056e-01 4.50547487e-01 7.54511476e-01 -5.25957286e-01
2.39982933e-01 1.45433009e+00 7.67141700e-01 3.40602249e-01
-8.75670612e-01 -2.76172400e-01 3.40613931e-01 -2.60647386e-01
-8.06300521e-01 1.46250293e-01 2.26269647e-01 -6.96681499e-01
7.25790501e-01 -1.34470910e-01 5.45549273e-01 7.88394749e-01
1.05309474e+00 3.67701322e-01 1.01251256e+00 -1.08474225e-01
9.00995433e-01 -7.90878981e-02 -1.42959490e-01 7.54866600e-01
2.79739827e-01 1.17101598e+00 -1.39408082e-01 -5.68877697e-01
4.12808806e-01 -2.44610906e-01 1.09250993e-01 -6.93752944e-01
-7.90086627e-01 1.00081301e+00 4.79720570e-02 -3.12333643e-01
-6.28891468e-01 5.10406435e-01 8.59296247e-02 1.89566582e-01
3.99713874e-01 7.21569479e-01 -5.97632229e-01 -2.66021341e-01
-5.72793543e-01 8.49459708e-01 1.17084372e+00 7.68494904e-01
1.17633112e-01 6.61666989e-01 -2.11960286e-01 6.06268048e-02
7.35844076e-01 1.09465969e+00 4.20833170e-01 -1.19239235e+00
1.36059359e-01 -3.03422749e-01 9.54586089e-01 -4.82575297e-01
-5.64954937e-01 -4.25670445e-01 -2.44959936e-01 1.05122125e+00
1.48480937e-01 -9.12688971e-01 -7.45765209e-01 1.74102938e+00
4.85694885e-01 1.56307638e-01 1.83394149e-01 7.66867042e-01
-3.62139910e-01 6.06241763e-01 -7.68177509e-02 -4.89780843e-01
8.06123316e-01 -4.88478601e-01 -8.44428062e-01 -1.59689412e-03
-9.09777805e-02 -5.01374245e-01 6.76912844e-01 4.13152784e-01
-1.14388454e+00 -3.97161424e-01 -1.31388676e+00 6.18968487e-01
-2.42558166e-01 -1.99807540e-01 1.14571406e-02 7.06604719e-01
-8.00028265e-01 1.05855918e+00 -1.14423108e+00 2.35126615e-01
-3.10651600e-01 2.45172501e-01 4.02520984e-01 7.46708691e-01
-1.31884181e+00 1.66552913e+00 4.53338653e-01 -1.20467491e-01
-1.45295203e+00 -8.55484247e-01 -7.95622289e-01 -1.41109094e-01
4.99426335e-01 -7.48153269e-01 1.57458496e+00 6.63291141e-02
-2.35449648e+00 -3.35665941e-01 4.68120158e-01 -9.71235216e-01
7.58538246e-01 -3.95943075e-01 3.41966301e-02 2.76514161e-02
7.12507367e-02 3.19681726e-02 1.33129132e+00 -7.31223226e-01
-7.50410557e-01 -1.88867018e-01 6.71816394e-02 5.90771019e-01
4.50807452e-01 -2.98320293e-01 7.15718806e-01 -6.10909283e-01
-2.52886325e-01 -1.37175667e+00 -6.62251711e-01 -1.37817830e-01
-2.49537349e-01 -1.49603516e-01 7.85076439e-01 -2.94583678e-01
5.86242557e-01 -1.86585557e+00 3.53930920e-01 2.59645015e-01
-5.60721397e-01 -1.26905486e-01 4.06030059e-01 4.64521140e-01
4.57045101e-02 -1.78416267e-01 -1.95297837e-01 -6.86077774e-02
5.56855142e-01 1.45891264e-01 -7.04604387e-01 6.61999285e-01
3.16795200e-01 4.68145072e-01 -1.12911963e+00 -2.77343355e-02
5.22984445e-01 2.72371888e-01 -2.78343201e-01 2.73350719e-02
-6.08440936e-01 3.26883882e-01 -8.04566264e-01 3.67960781e-01
2.70523220e-01 2.95265377e-01 -8.39274228e-02 3.90123188e-01
-4.87930328e-01 -2.31214270e-01 -1.33993387e+00 9.76806879e-01
-5.35347283e-01 1.05646841e-01 5.13197899e-01 -6.66867316e-01
8.92145216e-01 8.01861137e-02 5.17479181e-01 1.79389670e-01
3.31056565e-01 8.13285261e-02 -2.13414282e-01 -1.35300249e-01
5.05476415e-01 -7.13270962e-01 -3.40960026e-01 3.45164597e-01
2.52645284e-01 -1.32023597e+00 2.52865646e-02 -1.60283297e-01
8.35703671e-01 4.58321691e-01 2.91870356e-01 -8.14282477e-01
3.03237081e-01 -3.12759657e-03 4.46417630e-01 7.15289593e-01
-5.43741047e-01 1.09491512e-01 2.79694051e-01 -1.38314113e-01
-7.20167518e-01 -1.17064154e+00 -2.58058310e-01 8.01733911e-01
2.15659589e-01 1.50513709e-01 -4.14727867e-01 -2.23494008e-01
6.80092156e-01 1.19427204e+00 -6.38347864e-01 -3.38478029e-01
-8.30470622e-02 -9.43144739e-01 -1.68729395e-01 3.21302801e-01
-2.17159130e-02 -4.89026546e-01 -9.50960636e-01 4.35777605e-01
5.92100263e-01 -6.69745564e-01 -2.35187620e-01 4.68360722e-01
-6.26019180e-01 -1.09127736e+00 -5.25359273e-01 -1.52033463e-01
1.62790611e-01 -6.19206607e-01 7.97496438e-01 -8.32706749e-01
-1.67132318e-01 9.72877443e-01 4.11379665e-01 -9.86059904e-01
-3.46487284e-01 -5.42465985e-01 6.58593655e-01 -4.80974108e-01
-6.24941468e-01 -2.12426469e-01 -3.79740268e-01 2.82339364e-01
-3.81542683e-01 -7.68875897e-01 -4.68882099e-02 1.04094791e+00
6.93882704e-01 7.49910951e-01 4.16801661e-01 -2.84711748e-01
9.91990268e-01 -6.19178295e-01 -1.49758005e+00 -1.17216125e-01
-6.48882389e-01 3.11843008e-01 4.10889924e-01 -5.27134538e-01
-1.30199039e+00 3.02823633e-01 2.05493703e-01 -2.88395256e-01
4.85325158e-01 4.20667887e-01 4.05209273e-01 -2.63967961e-01
5.00118256e-01 -2.07348526e-01 4.88136292e-01 -1.28024966e-01
3.86938363e-01 -5.33908866e-02 4.04280722e-01 -8.64414573e-01
9.02890742e-01 3.98521423e-01 6.28234208e-01 -5.19600272e-01
-6.28347874e-01 -4.29742299e-02 -1.91065192e-01 -2.24146128e-01
5.37371814e-01 -8.54244411e-01 -1.07026589e+00 4.84828293e-01
-5.27848959e-01 -3.58415097e-01 -8.60005438e-01 8.27505767e-01
-1.48689747e+00 1.50727769e-02 -4.06817466e-01 -1.60191810e+00
-1.69956386e-01 -1.10854948e+00 9.66839552e-01 3.76724601e-01
1.43214524e-01 -1.23792875e+00 4.22573745e-01 -4.68097717e-01
3.99505496e-01 4.81487870e-01 6.51396990e-01 -3.98741990e-01
-2.44534194e-01 -5.07667184e-01 6.62315965e-01 4.10163641e-01
-3.08414549e-01 2.81429261e-01 -3.80837739e-01 -4.63746995e-01
7.21632183e-01 -4.53196764e-01 5.59876561e-01 8.93823922e-01
3.00299346e-01 -1.67579427e-01 -1.08530499e-01 1.71690702e-01
1.42691743e+00 2.64797062e-01 -3.75728965e-01 4.19762492e-01
4.23694402e-02 4.72093821e-01 1.17558122e+00 7.60320902e-01
2.39105642e-01 2.77928680e-01 9.34678614e-01 8.59549046e-01
6.73373938e-01 -2.21441761e-01 5.84843814e-01 1.34636492e-01
4.13682669e-01 1.18116207e-01 -7.42805541e-01 5.36335588e-01
-1.88847816e+00 -8.58357787e-01 4.07848001e-01 2.24645543e+00
5.70889235e-01 2.44276777e-01 1.08641095e-01 -4.31178331e-01
9.40795541e-01 -5.13287634e-02 -9.88880634e-01 -3.59452635e-01
5.53214960e-02 -1.20983504e-01 9.65540349e-01 1.10325634e+00
-1.21852970e+00 3.56406242e-01 7.77748585e+00 8.52197230e-01
-7.69110858e-01 -1.82416990e-01 2.46211290e-01 -1.29328325e-01
1.41415268e-01 -4.70061898e-02 -1.20659292e+00 8.71745706e-01
1.07930720e+00 -7.33304620e-01 4.12162870e-01 1.16297424e+00
4.35050905e-01 -3.83404911e-01 -6.39043391e-01 1.98332950e-01
-4.73846942e-01 -8.84774089e-01 -5.80433786e-01 -1.72392681e-01
9.47372913e-01 3.32916200e-01 7.03446031e-01 5.64518929e-01
1.08802223e+00 -5.22644997e-01 8.79060864e-01 9.07666028e-01
1.34228349e-01 -9.62708771e-01 6.19099796e-01 7.81838179e-01
-6.02149129e-01 -5.80153406e-01 -4.10171896e-01 -1.82600066e-01
4.41098124e-01 5.28755128e-01 -7.80686200e-01 1.95110664e-01
3.65082800e-01 2.41456091e-01 3.09474230e-01 1.13583112e+00
-2.85067290e-01 1.34546608e-01 -9.96387601e-01 -6.34446084e-01
4.32891965e-01 -3.69446248e-01 1.29261649e+00 4.81273860e-01
4.89001036e-01 -1.32106707e-01 5.97359419e-01 6.47251844e-01
7.44850218e-01 -3.68216932e-01 -5.78051269e-01 2.42465660e-01
2.91226387e-01 8.68633687e-01 -2.39058316e-01 -1.19431570e-01
2.85639763e-01 7.96633214e-02 -6.56435639e-02 2.11518526e-01
-7.21066952e-01 -3.78207177e-01 7.87111998e-01 -3.50710660e-01
5.28408170e-01 -3.94154519e-01 4.05038260e-02 -7.54082620e-01
-1.91989139e-01 -3.37506950e-01 3.46381456e-01 -5.88449419e-01
-1.37849379e+00 2.70950407e-01 3.47974986e-01 -1.06768548e+00
-1.04284966e+00 -8.13285410e-01 -5.67175090e-01 1.02364933e+00
-1.36067951e+00 -4.51513499e-01 6.12778008e-01 3.76693040e-01
2.97760159e-01 -4.08575386e-01 6.22156143e-01 -7.14900136e-01
-2.85981476e-01 -3.43532413e-01 8.13827395e-01 -7.62400150e-01
4.16906446e-01 -1.69481504e+00 2.58242935e-01 5.67746758e-01
-5.86367965e-01 4.39740568e-01 1.51733792e+00 -8.91275942e-01
-1.44181585e+00 -1.05747247e+00 2.35546902e-02 -6.69071138e-01
1.34098697e+00 2.48130202e-01 -3.27245533e-01 2.81854510e-01
1.13211222e-01 -5.40901758e-02 -8.45681652e-02 -1.32264897e-01
4.04433638e-01 9.79625583e-02 -1.62220645e+00 4.40555274e-01
1.87167302e-01 -3.13585967e-01 -8.43172848e-01 6.59494758e-01
6.37558043e-01 -8.09856892e-01 -8.25596213e-01 6.71765149e-01
3.97251338e-01 -3.29360545e-01 8.23273897e-01 -5.13872564e-01
-3.16423476e-01 -3.19195032e-01 -1.55598801e-02 -1.89346874e+00
-2.14847893e-01 -1.51024127e+00 -3.72951716e-01 3.34381729e-01
1.28297448e-01 -7.74454296e-01 6.12101078e-01 5.85825980e-01
-1.42915204e-01 -7.32049823e-01 -1.46261525e+00 -1.13222587e+00
4.19672817e-01 -1.98941350e-01 2.89243519e-01 1.04895204e-01
7.00526014e-02 -8.09768364e-02 -2.27912173e-01 5.54015875e-01
1.05508530e+00 -6.53509283e-04 2.02414498e-01 -1.07081783e+00
-2.52744526e-01 -4.17254984e-01 -2.71458507e-01 -4.81912822e-01
4.48578537e-01 -2.15373114e-01 7.41283953e-01 -9.82673049e-01
-3.87336582e-01 -1.49426982e-01 -4.58678007e-01 -1.15721464e-01
-3.76959801e-01 -3.42703909e-01 1.10110767e-01 -1.27802223e-01
-4.09803331e-01 9.81290936e-01 9.78541017e-01 -4.06986922e-02
-2.38801658e-01 8.37992370e-01 -2.24256128e-01 1.09120703e+00
9.38855886e-01 -4.05101359e-01 -6.95573092e-01 1.65139839e-01
1.36954129e-01 3.87623072e-01 2.80019462e-01 -1.30220068e+00
3.66863400e-01 -3.82403046e-01 1.25512108e-02 -7.41283357e-01
5.86006463e-01 -6.89845741e-01 1.80410761e-02 9.26714659e-01
-2.25501984e-01 3.62434000e-01 3.57017189e-01 1.24464130e+00
-6.26913160e-02 -5.94608545e-01 1.09579742e+00 -1.28443539e-01
-5.90549529e-01 2.59258121e-01 -7.28856623e-01 9.80252177e-02
1.19665802e+00 3.34998220e-01 -1.76064983e-01 -6.78846538e-01
-9.70030367e-01 6.94450080e-01 1.98563442e-01 1.96912929e-01
4.01426017e-01 -8.24170351e-01 -4.47534472e-01 1.00435309e-01
-5.39366841e-01 -4.33236629e-01 5.66764474e-02 2.92347729e-01
-2.74881482e-01 3.64684045e-01 -2.58645535e-01 -4.16485399e-01
-6.13005459e-01 6.18432403e-01 5.76536894e-01 -1.80248007e-01
-1.34199977e-01 7.91180074e-01 -2.88498938e-01 -5.08996964e-01
2.65891284e-01 -5.77568233e-01 1.00187205e-01 -2.78660711e-02
2.94723570e-01 5.19470274e-01 -2.00528994e-01 -3.36312681e-01
-9.21848416e-02 4.54943120e-01 4.35441315e-01 -7.08899438e-01
1.00681984e+00 -2.12891981e-01 3.91230285e-01 7.76942730e-01
3.87914002e-01 -1.04961313e-01 -2.08077526e+00 1.29502133e-01
1.36072323e-01 -1.20934341e-02 2.82366365e-01 -8.47108364e-01
-3.25693935e-01 4.69426125e-01 6.92460358e-01 2.98355430e-01
3.43576491e-01 -7.24563599e-01 2.22030863e-01 7.92309701e-01
7.62984097e-01 -1.39020479e+00 -2.08658114e-01 1.01629555e+00
9.72629607e-01 -9.78623033e-01 3.12225163e-01 1.31966397e-01
-9.27046716e-01 9.83000338e-01 2.46502399e-01 -9.81963515e-01
1.29794490e+00 7.08055675e-01 -1.46172494e-01 2.74805933e-01
-1.00628591e+00 -8.89489949e-02 9.89131629e-02 7.51923621e-01
-4.24019367e-01 1.96527138e-01 -8.30492601e-02 1.36957437e-01
-1.61440879e-01 -1.71941683e-01 4.63506043e-01 1.18398213e+00
-8.15326512e-01 -6.43021524e-01 -4.50681329e-01 1.56500787e-01
-5.70284247e-01 1.46316782e-01 3.67074400e-01 8.44110012e-01
-2.89882272e-01 8.30424547e-01 -1.51437730e-01 2.25704953e-01
4.26484019e-01 -4.09813151e-02 3.77952129e-01 -6.55601144e-01
-4.71912831e-01 -5.05126407e-03 -3.87705714e-02 -5.87350905e-01
7.98601210e-02 -9.69431102e-01 -1.13131762e+00 2.03473926e-01
-5.03286242e-01 4.86794174e-01 1.03220618e+00 7.30837584e-01
1.97936580e-01 3.70386690e-01 7.11825132e-01 -1.14581907e+00
-1.99919522e+00 -7.97959745e-01 -8.63336980e-01 -4.88849789e-01
7.74555147e-01 -1.27145481e+00 -7.38022327e-01 -4.94526207e-01] | [4.782378673553467, 2.414876937866211] |
ede714e7-47b4-4be7-8339-4a37c73041f3 | hierarchical-transfer-learning-with | 2111.08512 | null | https://arxiv.org/abs/2111.08512v3 | https://arxiv.org/pdf/2111.08512v3.pdf | Hierarchical transfer learning with applications for electricity load forecasting | The recent abundance of data on electricity consumption at different scales opens new challenges and highlights the need for new techniques to leverage information present at finer scales in order to improve forecasts at wider scales. In this work, we take advantage of the similarity between this hierarchical prediction problem and multi-scale transfer learning. We develop two methods for hierarchical transfer learning, based respectively on the stacking of generalized additive models and random forests, and on the use of aggregation of experts. We apply these methods to two problems of electricity load forecasting at national scale, using smart meter data in the first case, and regional data in the second case. For these two usecases, we compare the performances of our methods to that of benchmark algorithms, and we investigate their behaviour using variable importance analysis. Our results demonstrate the interest of both methods, which lead to a significant improvement of the predictions. | ['Solenne Gaucher', 'Anestis Antoniadis', 'Yannig Goude'] | 2021-11-16 | null | null | null | null | ['additive-models'] | ['methodology'] | [ 1.04228653e-01 1.19357221e-02 1.26394376e-01 -3.18145156e-01
-8.10645282e-01 -5.78458846e-01 8.69502425e-01 4.09996927e-01
-1.62766233e-01 1.18919563e+00 3.88298929e-01 -4.57604378e-01
-4.45044845e-01 -1.27486265e+00 -5.28976023e-01 -9.50436890e-01
-2.52953202e-01 5.47723949e-01 2.72877067e-01 -2.79125005e-01
2.21202180e-01 4.06998664e-01 -1.24341297e+00 2.46604785e-01
1.16584611e+00 1.16749418e+00 7.01508299e-02 4.06032115e-01
1.74412802e-02 7.89894998e-01 -2.76864618e-01 -1.56484663e-01
4.72602963e-01 -1.51079372e-01 -8.28181326e-01 -9.91930217e-02
-3.47925387e-02 -1.73983529e-01 3.78086984e-01 7.40291178e-01
4.45116341e-01 2.26671979e-01 8.83621275e-01 -1.29873645e+00
-1.48193449e-01 8.86838257e-01 -4.53882515e-01 1.21764272e-01
1.49231881e-01 -1.68489128e-01 1.02062547e+00 -3.08574021e-01
1.49278477e-01 9.90261316e-01 9.95077610e-01 -2.38972217e-01
-1.71681070e+00 -6.16187334e-01 1.42518893e-01 5.38318396e-01
-1.15823829e+00 -7.86528662e-02 6.21348560e-01 -9.28807557e-01
1.05613697e+00 -5.51938862e-02 5.30729175e-01 7.53258288e-01
9.42999795e-02 4.98798013e-01 1.68773508e+00 -5.42392135e-01
6.32989466e-01 4.54104394e-01 9.53790545e-02 -5.82437739e-02
-1.07259061e-02 2.30174914e-01 -2.28695020e-01 -2.95185834e-01
3.00263882e-01 -6.51730746e-02 -7.15028718e-02 -3.63206625e-01
-9.28979456e-01 1.23291564e+00 5.56170285e-01 7.60156333e-01
-6.71319783e-01 -2.94353515e-01 3.41765255e-01 3.12214613e-01
9.41906393e-01 2.40554318e-01 -8.65533173e-01 -1.06765088e-02
-1.28869843e+00 8.18379670e-02 1.01325667e+00 2.81718165e-01
8.94866168e-01 -2.01360397e-02 8.31029639e-02 6.22550786e-01
2.92777680e-02 3.02986294e-01 5.14981806e-01 -4.33861434e-01
4.79462385e-01 5.03524661e-01 1.58802867e-01 -6.41565919e-01
-7.12262630e-01 -4.14199620e-01 -1.00942671e+00 3.76375943e-01
3.24530661e-01 -2.74848759e-01 -5.34378529e-01 1.53321469e+00
3.17640573e-01 1.47344396e-01 5.68663850e-02 2.65323758e-01
-9.57356859e-03 7.30301857e-01 2.12830603e-01 -2.33909339e-01
8.44335437e-01 -9.05205548e-01 -2.98330903e-01 4.23115581e-01
7.54643798e-01 -2.96671897e-01 5.15771985e-01 5.87166846e-01
-6.73724473e-01 -6.24905407e-01 -8.35135818e-01 4.56386358e-01
-7.85277605e-01 -1.46461621e-01 3.89738202e-01 6.56635642e-01
-1.11494267e+00 1.20433438e+00 -6.86299980e-01 -4.40235704e-01
2.16398820e-01 1.98249057e-01 3.50353234e-02 2.61640579e-01
-1.40388155e+00 1.18264723e+00 5.02052724e-01 -1.43621996e-01
-2.45757788e-01 -9.43778396e-01 -4.98530596e-01 3.78217667e-01
-9.56302881e-02 -5.63768148e-01 1.07384038e+00 -9.92610216e-01
-1.43799365e+00 7.02203438e-02 3.07856977e-01 -8.07089150e-01
9.15299892e-01 -6.80042505e-02 -1.86910093e-01 -3.50361884e-01
8.22752342e-03 3.72477144e-01 5.59484780e-01 -9.99169648e-01
-9.69638765e-01 -4.38973188e-01 -1.38816893e-01 -1.97607711e-01
-4.83479798e-01 -2.41626784e-01 5.96549392e-01 -5.66076100e-01
-4.15471911e-01 -7.87692249e-01 -3.55264872e-01 -7.99557328e-01
-1.72352508e-01 -5.11511266e-01 5.66516578e-01 -1.04130876e+00
1.13209927e+00 -1.88934135e+00 3.64098698e-01 5.19772649e-01
-2.03863695e-01 -7.93478191e-02 1.74198195e-01 7.47731924e-01
-2.75319219e-01 6.70753866e-02 -4.29699570e-01 -2.92453676e-01
3.04642409e-01 2.53411412e-01 -4.22080904e-01 1.21755458e-01
5.42537719e-02 9.20737267e-01 -6.62812412e-01 -4.57028300e-02
4.72966075e-01 3.17457289e-01 -1.77865699e-01 1.72423065e-01
-4.37988825e-02 5.79349458e-01 -3.67241472e-01 1.05513096e-01
5.95842123e-01 -2.14105666e-01 3.68267536e-01 -8.02118480e-02
-4.13701922e-01 3.99137318e-01 -1.25088799e+00 1.01780331e+00
-7.66265333e-01 2.13871047e-01 -2.10825771e-01 -1.43436491e+00
6.96398437e-01 4.40158933e-01 6.14545703e-01 -7.21532404e-01
-2.59538203e-01 2.40737647e-01 -1.18064970e-01 -1.26564294e-01
9.82000530e-02 -3.71799737e-01 5.96435852e-02 6.26441717e-01
-1.09302521e-01 -2.21305132e-01 3.26655626e-01 -2.64353335e-01
8.61883819e-01 1.38558045e-01 7.25872636e-01 -6.90910459e-01
5.19317210e-01 -3.06406766e-01 3.13883603e-01 6.52108550e-01
3.35772187e-01 1.19131766e-01 4.31305826e-01 -6.58211648e-01
-9.56774890e-01 -8.91206324e-01 -3.53033334e-01 1.13876081e+00
-6.53956115e-01 -5.97536713e-02 -5.86469054e-01 -7.74918735e-01
5.15453517e-01 1.02440393e+00 -8.92786980e-01 3.43292326e-01
-2.68595368e-01 -1.17644250e+00 3.08230170e-03 7.00952768e-01
2.81237632e-01 -1.09722435e+00 -6.99502051e-01 3.45581621e-01
7.91037828e-02 -9.28919137e-01 3.74433815e-01 5.39604068e-01
-1.00397408e+00 -7.74591327e-01 -7.64643729e-01 -1.82278529e-01
1.20608941e-01 -2.98488975e-01 1.42414629e+00 -2.98730284e-01
3.84469517e-02 1.77495167e-01 -5.71632028e-01 -5.10483682e-01
-4.83007401e-01 7.28100896e-01 8.50125700e-02 8.64959508e-02
1.38461471e-01 -1.24822986e+00 -2.94015497e-01 2.09799320e-01
-8.56032968e-01 7.90691301e-02 6.92942262e-01 6.96633995e-01
-1.95713323e-02 2.43297681e-01 9.63459015e-01 -9.58645999e-01
5.41922092e-01 -7.96372056e-01 -1.11086118e+00 2.56371737e-01
-1.15390468e+00 2.31788546e-01 8.26303601e-01 -6.77585974e-02
-7.69296885e-01 1.92258969e-01 -7.64825046e-02 5.23992144e-02
-3.30118060e-01 7.01636493e-01 4.05387469e-02 -3.60745825e-02
3.06965142e-01 2.56848127e-01 -4.48690206e-01 -8.36306632e-01
3.63003612e-01 6.70260251e-01 -7.86694959e-02 -3.53569716e-01
9.90648508e-01 1.07041866e-01 1.05575472e-01 -7.32648075e-01
-4.82961923e-01 -3.55963975e-01 -1.11670458e+00 3.62768769e-02
7.02148855e-01 -8.25582802e-01 -3.61702621e-01 4.32572961e-01
-7.95745313e-01 -8.38155150e-01 -5.37165999e-01 3.99481356e-01
-6.03313506e-01 3.10232341e-01 -5.33423960e-01 -9.38888013e-01
-2.90018201e-01 -9.94806945e-01 7.88963139e-01 -2.10594404e-02
1.18802562e-02 -1.41116905e+00 5.67936182e-01 1.30615294e-01
7.61219442e-01 3.06541860e-01 1.27601874e+00 -9.65963602e-01
-3.84317935e-01 6.16702549e-02 -2.33981952e-01 3.75339597e-01
1.80249706e-01 -4.04061735e-01 -1.02194512e+00 -4.54800278e-01
-1.99606940e-01 -4.30927545e-01 9.80838776e-01 3.34870785e-01
8.33333194e-01 -2.28692040e-01 -2.34636441e-01 1.05072796e-01
1.55426705e+00 5.67818359e-02 6.23547196e-01 5.26404083e-01
3.37078363e-01 8.61315191e-01 1.61062181e-01 7.01933742e-01
6.72850311e-01 9.05560493e-01 9.74755287e-02 -3.61727595e-01
4.65898335e-01 -5.64881638e-02 -3.84135847e-03 8.77661049e-01
-4.24694031e-01 8.47987831e-02 -9.66062486e-01 6.27392173e-01
-2.09666872e+00 -8.39316964e-01 6.40271232e-02 2.22938752e+00
5.37355661e-01 1.90505316e-03 5.89187026e-01 2.97391623e-01
4.35678035e-01 2.87168194e-02 -3.30094546e-01 -3.98147404e-01
1.21200033e-01 3.87209594e-01 5.51900685e-01 4.62857455e-01
-1.24908257e+00 3.93903852e-01 6.76932907e+00 7.19559848e-01
-9.91884530e-01 3.05480272e-01 7.75573552e-01 1.91152930e-01
-8.35587606e-02 -2.92682089e-03 -5.84505856e-01 6.40487611e-01
1.36539805e+00 -1.75004795e-01 6.95462406e-01 6.96768701e-01
2.75216162e-01 -1.89522475e-01 -1.16906857e+00 2.94517905e-01
-2.84515679e-01 -1.06292057e+00 -2.25595921e-01 3.49714756e-01
1.08548057e+00 3.83603513e-01 -2.28044465e-01 4.77331728e-01
7.67776251e-01 -9.54019785e-01 4.22087640e-01 5.52236974e-01
8.28064084e-02 -6.53651595e-01 9.50075030e-01 4.78004187e-01
-1.32227802e+00 -3.89987022e-01 -8.42338577e-02 -4.19842035e-01
2.78422236e-01 8.52717638e-01 -7.32970297e-01 1.11450863e+00
8.30095887e-01 7.91531146e-01 -4.39512461e-01 9.56766903e-01
-1.20653622e-01 7.47037947e-01 -5.18245101e-01 7.11258128e-02
1.81453601e-01 -4.52676892e-01 -2.23429017e-02 1.27493715e+00
4.78023797e-01 -1.92552194e-01 3.44280779e-01 6.76455677e-01
2.78258353e-01 5.08333743e-01 -4.78311509e-01 3.96207064e-01
5.19401766e-02 1.44646835e+00 -6.05939090e-01 -4.15608823e-01
-7.10552633e-01 6.85241640e-01 3.55538845e-01 2.73535848e-01
-6.24103308e-01 -1.23274051e-01 2.00396523e-01 -1.30431065e-02
8.61899078e-01 -3.92893814e-02 -2.21937180e-01 -1.00705397e+00
-9.71582979e-02 -6.64124250e-01 5.01182020e-01 -5.73496878e-01
-1.73242581e+00 3.71729642e-01 1.66587144e-01 -1.05501437e+00
-6.82813883e-01 -4.95173782e-01 -8.43518972e-01 1.04945946e+00
-1.90779161e+00 -1.16977763e+00 -5.52113205e-02 3.43516737e-01
3.60672742e-01 1.15993999e-01 9.38560843e-01 1.39997259e-01
-2.32099444e-01 1.35901615e-01 7.95524240e-01 -8.01976025e-02
2.93044657e-01 -1.69087255e+00 4.57926959e-01 3.95058572e-01
1.43777564e-01 -1.77666709e-01 4.86634463e-01 -4.62932408e-01
-5.89858592e-01 -1.12279284e+00 8.34975243e-01 -4.06044185e-01
9.69225347e-01 -3.83044004e-01 -9.56142843e-01 6.81046605e-01
4.60125864e-01 -3.46725762e-01 8.46725464e-01 5.48445046e-01
-2.97094166e-01 -2.97408402e-01 -1.12455714e+00 -1.46309480e-01
2.73350507e-01 -2.36114457e-01 -5.86804688e-01 3.08721662e-01
2.43458614e-01 4.17144775e-01 -1.48358762e+00 6.22594416e-01
6.95232332e-01 -1.09004295e+00 6.98761284e-01 -4.18677390e-01
3.61083776e-01 -2.56268140e-02 -3.74706239e-01 -1.94412947e+00
-5.02517283e-01 -2.31062829e-01 1.60182863e-01 1.36163521e+00
6.87399566e-01 -1.08096015e+00 4.29604113e-01 4.76859689e-01
3.41060042e-01 -5.36352873e-01 -1.07123208e+00 -7.28046417e-01
7.28177607e-01 -2.95732412e-02 8.17001283e-01 1.04964757e+00
1.51354685e-01 3.75695825e-01 -2.46957958e-01 -2.03177542e-03
4.93156552e-01 6.45691097e-01 6.21191680e-01 -1.74706352e+00
-6.03787065e-01 -3.97781193e-01 -3.70472610e-01 -5.03152907e-01
2.57303894e-01 -6.75233245e-01 -1.27416372e-01 -1.37443864e+00
2.87687093e-01 -4.65480536e-01 -6.92305028e-01 5.35658181e-01
-3.39394286e-02 3.28491777e-01 3.37345481e-01 -2.46520285e-02
-2.28999078e-01 6.03180528e-01 2.80437350e-01 -9.35131535e-02
-2.01767370e-01 2.40237236e-01 -2.93006629e-01 6.69084609e-01
1.09331667e+00 -1.90140963e-01 -1.06210522e-01 8.96104332e-03
2.83235461e-01 -7.97092170e-02 1.82548761e-01 -1.01899540e+00
-8.74170288e-02 -1.03557752e-02 6.63276911e-01 -5.37549019e-01
-3.41151208e-02 -8.59647095e-01 3.37004542e-01 5.86330295e-01
-2.23049968e-01 3.14861596e-01 1.42074630e-01 3.82601142e-01
-5.78232966e-02 -1.33571476e-01 6.39526725e-01 -1.57815650e-01
-4.94978875e-01 -2.00840965e-01 -4.87382591e-01 -3.20793808e-01
9.63684976e-01 2.17099950e-01 3.36746499e-02 -4.39235121e-01
-8.62033069e-01 2.25126758e-01 2.41810039e-01 3.28179747e-01
-2.51365870e-01 -1.07200146e+00 -1.07490027e+00 9.26245004e-02
-1.19778164e-01 -5.33660352e-01 8.94251391e-02 1.10234058e+00
1.04004681e-01 7.23614633e-01 -2.58092672e-01 -4.74302679e-01
-9.74037588e-01 7.61623383e-01 3.03166807e-01 -1.04460943e+00
-3.26132208e-01 1.05003469e-01 1.33274138e-01 -6.52310491e-01
-1.39369607e-01 -4.97308433e-01 -5.59981763e-01 5.18390596e-01
3.49765480e-01 6.52248442e-01 4.18147027e-01 -4.20854568e-01
-8.63222107e-02 5.77563822e-01 2.49604374e-01 2.56234463e-02
1.70390797e+00 -1.04471341e-01 -2.49869481e-01 7.85951138e-01
8.17887068e-01 -6.47905543e-02 -1.20181561e+00 -1.40318483e-01
4.54317987e-01 -1.71879992e-01 1.01813249e-01 -1.03413057e+00
-8.69102538e-01 9.80789006e-01 7.02483356e-01 1.08715713e+00
1.36860812e+00 -1.43862039e-01 4.41605598e-01 1.21794879e-01
5.69790721e-01 -1.00776970e+00 -6.04417324e-01 2.11105809e-01
6.38739407e-01 -1.14095926e+00 3.83149162e-02 -1.75224498e-01
-3.42887133e-01 1.02840722e+00 -3.18783894e-02 -8.13661367e-02
7.89045393e-01 2.89245576e-01 -9.11792964e-02 4.45028603e-01
-8.13975751e-01 -4.27273154e-01 9.64338183e-02 4.88866150e-01
3.02634269e-01 4.36078548e-01 -2.91239291e-01 4.90802050e-01
-1.99676022e-01 2.42752939e-01 1.97968572e-01 6.57452345e-01
-2.44998664e-01 -1.36380720e+00 -3.22469980e-01 5.66200137e-01
-3.47225308e-01 -2.52070397e-01 -1.48697957e-01 8.21829855e-01
8.45322311e-02 8.91716778e-01 4.56712097e-02 -2.46262312e-01
2.74965316e-01 5.17239571e-01 2.78660655e-01 -2.43778735e-01
-7.88239062e-01 2.19572037e-02 -1.84037894e-01 -2.53295690e-01
-5.38652301e-01 -1.00974667e+00 -4.68804389e-01 -7.98530653e-02
-3.55057597e-01 4.94921088e-01 7.19553113e-01 1.16735208e+00
1.78611547e-01 4.45696890e-01 1.05463707e+00 -1.33591473e+00
-9.04892206e-01 -1.39459920e+00 -8.40552926e-01 4.15126801e-01
1.61191568e-01 -5.55859923e-01 -4.76159245e-01 -7.71695822e-02] | [6.254243850708008, 2.961545467376709] |
50b064d4-9a79-4372-ac1b-c934a72da0b2 | lexical-normalization-for-code-switched-data-1 | null | null | https://aclanthology.org/2021.eacl-main.200 | https://aclanthology.org/2021.eacl-main.200.pdf | Lexical Normalization for Code-switched Data and its Effect on POS Tagging | Lexical normalization, the translation of non-canonical data to standard language, has shown to improve the performance of many natural language processing tasks on social media. Yet, using multiple languages in one utterance, also called code-switching (CS), is frequently overlooked by these normalization systems, despite its common use in social media. In this paper, we propose three normalization models specifically designed to handle code-switched data which we evaluate for two language pairs: Indonesian-English and Turkish-German. For the latter, we introduce novel normalization layers and their corresponding language ID and POS tags for the dataset, and evaluate the downstream effect of normalization on POS tagging. Results show that our CS-tailored normalization models significantly outperform monolingual ones, and lead to 5.4{\%} relative performance increase for POS tagging as compared to unnormalized input. | ['{\\"O}zlem {\\c{C}}etino{\\u{g}}lu', 'Rob van der Goot'] | 2021-04-01 | null | null | null | eacl-2021-2 | ['lexical-normalization'] | ['natural-language-processing'] | [ 9.02659595e-02 -3.97260115e-02 -1.83694452e-01 -4.81703460e-01
-6.69278026e-01 -7.61518300e-01 6.84785485e-01 6.15105629e-01
-8.50650907e-01 6.33366764e-01 3.64906311e-01 -5.03967464e-01
3.30801040e-01 -3.59957069e-01 -3.11882287e-01 -3.08546215e-01
3.28007638e-01 3.77369255e-01 9.86486971e-02 -5.54267406e-01
-8.05353466e-03 2.22380474e-01 -1.13137448e+00 1.72218502e-01
8.40012133e-01 1.74345151e-01 9.11388472e-02 4.32861745e-01
-5.50579727e-01 6.34220958e-01 -4.95339572e-01 -9.73139703e-01
1.43728420e-01 -3.04839045e-01 -6.83660865e-01 -4.68387932e-01
4.12915349e-01 3.45781833e-01 1.61737964e-01 1.54765642e+00
3.67230058e-01 4.31834720e-02 5.46438336e-01 -8.37373257e-01
-8.06591809e-01 1.30023813e+00 -5.07513225e-01 2.56754339e-01
5.46726465e-01 -4.62992847e-01 1.10224986e+00 -7.89428532e-01
9.20402348e-01 1.04277456e+00 7.76340783e-01 7.10738599e-01
-1.33875299e+00 -6.30491972e-01 -1.07422158e-01 -2.00741127e-01
-1.43754029e+00 -4.40819681e-01 3.01670820e-01 -7.57729828e-01
1.21965730e+00 7.42128119e-02 1.07058585e-01 1.20896018e+00
3.00825357e-01 3.42447758e-01 8.99593771e-01 -9.37561512e-01
-1.97151184e-01 1.16111867e-01 2.21133262e-01 3.56064051e-01
3.17536563e-01 -4.50203091e-01 -3.05576831e-01 1.36526361e-01
2.12353826e-01 -2.30235159e-01 7.90788829e-02 1.40015073e-02
-1.36676037e+00 6.94921851e-01 -8.58328119e-02 7.89297462e-01
-8.82433206e-02 -2.43690178e-01 7.26820886e-01 3.26740801e-01
5.66303492e-01 4.43227351e-01 -6.19702578e-01 -5.47110498e-01
-7.22895741e-01 9.24810022e-02 9.39594924e-01 1.31794667e+00
7.34523356e-01 -1.78938746e-01 -2.50135362e-01 1.26492941e+00
5.59948087e-02 8.28390300e-01 6.83331907e-01 -4.33925480e-01
6.76804841e-01 6.07972920e-01 -1.55619949e-01 -7.80185997e-01
-7.70196021e-01 -3.86876225e-01 -7.13785589e-01 -5.50102115e-01
4.31110978e-01 -3.29774052e-01 -6.55902743e-01 1.95034254e+00
-1.56563316e-02 -3.69777501e-01 3.23257506e-01 2.66775072e-01
5.65856040e-01 5.15751898e-01 4.17320251e-01 -3.06304067e-01
1.29330444e+00 -1.01816547e+00 -8.76926899e-01 -3.52668256e-01
1.11539686e+00 -1.32722390e+00 1.23518622e+00 -8.86145607e-03
-7.00031996e-01 -3.07548255e-01 -7.83150434e-01 -1.45977050e-01
-4.94561583e-01 1.26454189e-01 3.82297933e-01 9.31985855e-01
-8.91242325e-01 4.02996123e-01 -8.46022487e-01 -9.25990641e-01
-1.64588466e-01 2.36579403e-01 -5.35995781e-01 -6.75447658e-02
-1.19838297e+00 9.77881908e-01 4.89078283e-01 -5.51737666e-01
-2.30463482e-02 -5.15714645e-01 -9.79895711e-01 4.63229008e-02
3.87893580e-02 -1.13959730e-01 1.47817254e+00 -9.83058810e-01
-1.51909900e+00 1.21039450e+00 -1.70572251e-01 -2.95370787e-01
3.09933841e-01 -3.20093751e-01 -8.50855410e-01 -5.29841244e-01
4.02497411e-01 4.31168854e-01 3.07141125e-01 -7.28051603e-01
-6.30143046e-01 1.25544041e-01 3.40886670e-03 1.29083887e-01
-6.91776693e-01 7.99574614e-01 -7.25890219e-01 -9.24348712e-01
-1.98262800e-02 -1.09426522e+00 -3.19182217e-01 -8.94927800e-01
-3.61835808e-01 -2.10391656e-01 1.55528724e-01 -9.49236691e-01
1.69525385e+00 -2.33730292e+00 9.86990333e-02 3.23770940e-01
-1.79341465e-01 3.95423800e-01 -3.73299778e-01 6.00352049e-01
-1.88197061e-01 4.01668549e-01 -4.11359042e-01 -7.34335542e-01
4.74222079e-02 5.97196281e-01 9.02171209e-02 2.51031369e-01
3.04637581e-01 7.20332980e-01 -9.15535450e-01 -3.61739069e-01
-2.04905584e-01 4.36983705e-01 -9.76929367e-01 4.27765325e-02
9.87655669e-02 3.99403691e-01 1.51683018e-01 4.63461339e-01
5.83878219e-01 2.02530846e-01 5.53833544e-01 2.27471918e-01
-4.23135877e-01 4.52552646e-01 -8.67909849e-01 1.69919741e+00
-6.12582088e-01 5.16718447e-01 -2.61298455e-02 -4.42040443e-01
8.97330701e-01 2.11169347e-01 2.05695570e-01 -7.36808896e-01
3.40404093e-01 4.49151516e-01 3.55834574e-01 -4.27882969e-01
8.12565327e-01 -1.14626758e-01 -6.01974368e-01 3.48881900e-01
4.22918439e-01 8.76020789e-02 9.47118580e-01 4.49206419e-02
1.07331300e+00 -2.65015326e-02 7.78894484e-01 -5.84846020e-01
8.18974912e-01 -2.25711286e-01 7.55117595e-01 5.63424408e-01
-1.55756563e-01 6.06806278e-01 7.39937186e-01 -3.12488206e-04
-1.03259718e+00 -7.43951857e-01 -8.66060928e-02 1.61545050e+00
-2.88419574e-01 -7.74282753e-01 -1.06959283e+00 -7.12174118e-01
-2.20124528e-01 7.62736797e-01 -3.79177243e-01 1.95890069e-02
-8.28529835e-01 -8.09207618e-01 8.70408833e-01 2.78042078e-01
7.38582760e-02 -9.09618258e-01 8.89492854e-02 3.53790581e-01
-4.15992379e-01 -1.56546128e+00 -7.09278047e-01 4.54568446e-01
-4.49014992e-01 -8.75869453e-01 -3.59653682e-01 -9.68973696e-01
4.91188228e-01 -6.80328384e-02 1.09274495e+00 1.34266600e-01
4.21675116e-01 1.91630960e-01 -9.12648380e-01 -4.81341630e-01
-1.08875740e+00 8.39402854e-01 3.47064465e-01 -6.27152482e-03
7.36363411e-01 -5.38089097e-01 2.21274167e-01 1.16169780e-01
-1.10473752e+00 -8.03158209e-02 6.35618031e-01 6.19094849e-01
1.85161099e-01 -7.86398113e-01 2.82383770e-01 -1.25419509e+00
7.54607499e-01 -5.46196222e-01 -5.24167836e-01 1.94703162e-01
-4.65344518e-01 2.72475392e-01 7.42695510e-01 -4.28101748e-01
-9.72796261e-01 1.44089162e-01 -7.33927310e-01 2.48891965e-01
-2.59130776e-01 9.79296446e-01 -2.57023603e-01 -8.01270008e-02
8.50874603e-01 -1.38452485e-01 -3.31062824e-01 -6.94932938e-01
2.90268004e-01 7.72583723e-01 7.14041829e-01 -6.90943718e-01
8.07958186e-01 -4.12717387e-02 -4.00299340e-01 -7.96374679e-01
-8.73111665e-01 -8.10522079e-01 -1.12808204e+00 6.63018525e-02
1.09702384e+00 -8.16790044e-01 -7.21200481e-02 3.43463629e-01
-1.60426366e+00 -1.07387587e-01 -5.85334338e-02 6.56081736e-01
-1.00880586e-01 3.71383399e-01 -6.90605402e-01 -2.38061205e-01
-1.42142013e-01 -1.14199197e+00 5.76027632e-01 1.70190156e-01
-4.42396939e-01 -1.16802049e+00 5.30336797e-01 1.29489794e-01
5.47472417e-01 -1.33427326e-02 1.05840206e+00 -1.23515499e+00
2.39315599e-01 -3.33384424e-01 -1.88705310e-01 7.09504545e-01
9.65392366e-02 2.03174725e-01 -8.50965261e-01 -2.51880139e-01
-4.43469316e-01 1.07505381e-01 3.67646784e-01 -2.77173162e-01
4.08896416e-01 -6.55999482e-02 4.14359793e-02 5.28296411e-01
1.22967255e+00 1.36812702e-01 4.59981859e-01 3.78118753e-01
7.62100756e-01 3.31131101e-01 2.67923564e-01 3.03211510e-01
4.11158830e-01 6.09956682e-01 1.13602772e-01 3.09977755e-02
-9.47723240e-02 -1.71014264e-01 7.66658485e-01 1.98401630e+00
-9.48205292e-02 -1.72518075e-01 -1.33875751e+00 5.18512130e-01
-1.59859192e+00 -5.40251493e-01 -4.45113778e-01 2.15736270e+00
1.05769312e+00 2.26113126e-01 -2.43736401e-01 -1.97187871e-01
8.83812428e-01 -6.29038513e-02 2.57365227e-01 -8.36121440e-01
-4.82864738e-01 5.28116047e-01 5.92143774e-01 6.96004272e-01
-1.28073585e+00 1.24119139e+00 5.94517994e+00 7.02617526e-01
-1.18109560e+00 6.60285771e-01 -3.07007376e-02 3.60403806e-01
-8.31764862e-02 1.18000008e-01 -1.07383966e+00 2.60286003e-01
1.30262148e+00 -3.57919127e-01 5.21457911e-01 8.97340834e-01
-1.40213836e-02 1.96708634e-01 -1.22553921e+00 9.54522252e-01
2.45985642e-01 -8.88919950e-01 1.36945844e-01 -1.43207982e-01
1.05244720e+00 5.78352749e-01 -4.66831446e-01 7.07459450e-01
4.31863397e-01 -5.69532692e-01 1.02261066e+00 1.54589623e-01
7.70038366e-01 -7.77674854e-01 1.21496379e+00 2.75592744e-01
-1.13660669e+00 1.25167787e-01 -3.38120401e-01 -2.75861412e-01
1.99875459e-01 5.49930632e-01 -6.10460818e-01 5.78720212e-01
5.41196942e-01 7.03902662e-01 -8.94932806e-01 8.78501415e-01
-4.90925372e-01 7.61422455e-01 -1.61308154e-01 2.98302770e-02
1.63890302e-01 -2.33630702e-01 5.78395903e-01 1.95701003e+00
4.80482936e-01 -3.64364892e-01 1.67616084e-01 3.80439222e-01
-4.86585289e-01 8.91284108e-01 -3.94444585e-01 -2.14252651e-01
4.05567706e-01 1.43552423e+00 -9.38379169e-01 -1.97098762e-01
-8.09053063e-01 8.98668647e-01 5.06800056e-01 1.69515848e-01
-8.28226447e-01 -7.31900096e-01 8.26528549e-01 -4.23584953e-02
2.97829900e-02 -3.56303960e-01 -2.71134060e-02 -1.52359188e+00
-6.37302399e-02 -8.01891208e-01 3.02467287e-01 -3.00643682e-01
-1.29091990e+00 7.76505411e-01 -5.46891876e-02 -1.15778100e+00
-1.30263686e-01 -7.88368881e-01 -2.96280861e-01 7.66894162e-01
-1.47261941e+00 -1.04393840e+00 2.73299254e-02 3.20937514e-01
1.09741122e-01 -2.48652920e-01 1.09986675e+00 1.14297044e+00
-7.10816443e-01 1.01005805e+00 4.20530289e-01 6.15277112e-01
1.23938298e+00 -1.15386379e+00 7.06400335e-01 1.30677831e+00
2.91853786e-01 1.04576898e+00 6.59282565e-01 -5.66312909e-01
-7.98674166e-01 -1.14650738e+00 1.80866861e+00 -6.09090209e-01
1.00344360e+00 -6.68587565e-01 -8.15925896e-01 8.03903580e-01
4.63453174e-01 -1.41647875e-01 1.02932560e+00 2.65610784e-01
-6.87194943e-01 5.32754101e-02 -8.28274369e-01 7.33111560e-01
1.01300621e+00 -6.05410755e-01 -6.71947420e-01 1.08840674e-01
8.47774267e-01 -3.73101205e-01 -1.02181721e+00 9.88399312e-02
2.95969516e-01 -7.97882557e-01 4.48966175e-01 -6.95072591e-01
4.20055389e-01 -2.82178491e-01 -4.45316166e-01 -1.40350938e+00
-4.10235137e-01 -6.28853321e-01 7.38426864e-01 1.72458196e+00
7.40927696e-01 -5.60378850e-01 2.71636754e-01 1.80055097e-01
-4.81409907e-01 6.24700375e-02 -9.61087763e-01 -8.02397609e-01
4.25279945e-01 -6.58633053e-01 5.72770953e-01 1.43142080e+00
2.91656822e-01 4.92452383e-01 -2.13280797e-01 2.37621777e-02
6.96328282e-02 -4.64474142e-01 6.92376018e-01 -1.41339922e+00
-2.46983860e-02 -5.61302006e-01 -5.12570679e-01 -7.09805846e-01
4.85987037e-01 -1.45080817e+00 9.15362835e-02 -1.18893731e+00
1.93708077e-01 -3.44439387e-01 -2.54929632e-01 7.38500774e-01
-1.14692792e-01 4.61520821e-01 3.88005793e-01 7.41813378e-03
-6.69651985e-01 1.69810459e-01 5.90095341e-01 5.84267564e-02
-1.55845493e-01 -2.76699126e-01 -6.09518051e-01 8.26201141e-01
7.55516946e-01 -9.02153492e-01 2.02779680e-01 -6.87044442e-01
7.39530802e-01 -4.77836877e-01 -4.66032147e-01 -1.15414250e+00
7.54655525e-02 -2.11139411e-01 -2.96571583e-01 -2.77399290e-02
-3.84248972e-01 -7.11163044e-01 -4.68486771e-02 3.06517512e-01
-2.23628640e-01 4.49903697e-01 4.31288600e-01 -2.43946612e-02
-2.96738654e-01 -5.62729537e-01 9.02399302e-01 1.42939493e-01
-4.74910468e-01 -1.62343413e-01 -6.74726188e-01 4.97487009e-01
6.84542477e-01 2.36185879e-01 -1.83626279e-01 1.66472290e-02
-6.12134218e-01 -1.31816879e-01 5.42589605e-01 8.00536990e-01
-2.01570347e-01 -1.36287463e+00 -6.80407286e-01 2.59989262e-01
5.56326270e-01 -4.55209255e-01 -9.12284851e-02 1.04701555e+00
-7.84418523e-01 3.89228463e-01 -3.46730858e-01 -3.38999391e-01
-1.24489427e+00 2.07061097e-01 8.89647976e-02 -5.19782901e-01
4.50408012e-02 6.96914196e-01 -3.23997319e-01 -1.25973535e+00
-3.81942987e-02 -6.67912126e-01 -4.49778557e-01 3.43425661e-01
2.88831353e-01 2.14454815e-01 6.92112982e-01 -1.14884770e+00
-4.18655157e-01 4.29464042e-01 -8.17888454e-02 1.16801839e-02
1.27490211e+00 -1.21298768e-01 -7.74808288e-01 4.70292538e-01
1.17069006e+00 8.36959541e-01 -4.39586699e-01 -4.19333875e-01
6.73311055e-01 -7.49693438e-02 -5.13969779e-01 -5.24684131e-01
-6.81631982e-01 7.17899263e-01 2.98723698e-01 1.52596474e-01
6.55848444e-01 -9.81056020e-02 5.53685606e-01 5.90779305e-01
2.15740785e-01 -1.19381654e+00 -5.64313650e-01 1.54876220e+00
5.47707677e-01 -1.05526793e+00 -3.02492082e-01 -4.58562076e-01
-4.63787168e-01 1.18970823e+00 5.21324337e-01 2.09810615e-01
6.39648557e-01 4.91684198e-01 5.43710589e-01 3.68876070e-01
-2.89323658e-01 -2.67228693e-01 1.28832921e-01 5.42987525e-01
1.07258880e+00 1.74615860e-01 -7.03103364e-01 8.48035812e-01
-6.00760877e-01 -3.88568461e-01 7.95529544e-01 9.90843654e-01
-1.00350395e-01 -1.79272568e+00 -4.01396990e-01 4.32524562e-01
-9.09743011e-01 -6.00817323e-01 -1.25929311e-01 7.26811707e-01
3.46045405e-01 7.16908693e-01 1.32389188e-01 -4.00494784e-01
6.06068969e-01 4.40465689e-01 7.92171657e-02 -1.16696584e+00
-1.13290644e+00 4.09405632e-03 7.24475831e-02 -2.09640458e-01
-5.60873330e-01 -9.20190096e-01 -1.34829772e+00 -3.63134027e-01
-1.82434484e-01 2.80007213e-01 6.79168582e-01 9.40920591e-01
1.32010028e-01 4.89051729e-01 1.45953193e-01 -7.66449690e-01
-3.58697116e-01 -1.15353096e+00 -3.39019120e-01 7.00364530e-01
2.91348733e-02 -5.60309291e-01 -3.68668497e-01 3.66017878e-01] | [10.240554809570312, 9.989635467529297] |
269e1711-fea4-454e-aeee-dcaa2e0aa58a | confidence-intervals-and-hypothesis-testing-1 | null | null | http://papers.nips.cc/paper/4931-confidence-intervals-and-hypothesis-testing-for-high-dimensional-statistical-models | http://papers.nips.cc/paper/4931-confidence-intervals-and-hypothesis-testing-for-high-dimensional-statistical-models.pdf | Confidence Intervals and Hypothesis Testing for High-Dimensional Statistical Models | Fitting high-dimensional statistical models often requires the use of non-linear parameter estimation procedures. As a consequence, it is generally impossible to obtain an exact characterization of the probability distribution of the parameter estimates. This in turn implies that it is extremely challenging to quantify the `uncertainty' associated with a certain parameter estimate. Concretely, no commonly accepted procedure exists for computing classical measures of uncertainty and statistical significance as confidence intervals or p-values. We consider here a broad class of regression problems, and propose an efficient algorithm for constructing confidence intervals and p-values. The resulting confidence intervals have nearly optimal size. When testing for the null hypothesis that a certain parameter is vanishing, our method has nearly optimal power. Our approach is based on constructing a `de-biased' version of regularized M-estimators. The new construction improves over recent work in the field in that it does not assume a special structure on the design matrix. Furthermore, proofs are remarkably simple. We test our method on a diabetes prediction problem. | ['Andrea Montanari', 'Adel Javanmard'] | 2013-12-01 | null | null | null | neurips-2013-12 | ['diabetes-prediction'] | ['medical'] | [ 3.52089047e-01 1.87459230e-01 -3.25210452e-01 -2.90678293e-01
-9.51019108e-01 -5.22342324e-01 2.45489493e-01 5.28443754e-01
-4.08823013e-01 1.19267499e+00 -1.82379857e-01 -4.89548773e-01
-3.93375099e-01 -7.29425311e-01 -8.38109791e-01 -9.07270789e-01
-3.14342380e-01 4.32475626e-01 -1.30261462e-02 3.20285112e-01
5.45004070e-01 3.41622084e-01 -1.36396790e+00 -4.96910214e-01
1.04331207e+00 1.14913309e+00 -1.51162431e-01 7.04654455e-01
1.98832348e-01 3.55612002e-02 -5.54124773e-01 -4.09920335e-01
1.00380398e-01 -6.25087440e-01 -4.17299151e-01 7.14797005e-02
1.28799275e-01 -1.72498614e-01 5.40832996e-01 1.26457179e+00
4.88300323e-01 -9.23782215e-02 1.03078854e+00 -1.26464331e+00
-3.16951424e-01 4.95297849e-01 -6.94487393e-01 -9.98645648e-02
2.25636616e-01 -3.66045028e-01 8.79083395e-01 -6.03442073e-01
2.59109616e-01 9.23674166e-01 9.65976655e-01 -1.06572555e-02
-1.68366969e+00 -5.44250488e-01 -3.90832365e-01 -2.15042114e-01
-1.60945380e+00 -4.30206895e-01 2.95707226e-01 -5.95628858e-01
1.81113303e-01 3.79576504e-01 3.54007095e-01 7.62130916e-01
4.97095883e-01 2.09130824e-01 1.20307517e+00 -6.43548429e-01
6.08690321e-01 3.27482522e-01 1.97832420e-01 5.64237058e-01
1.07712674e+00 2.80101508e-01 -4.86772656e-02 -5.00451505e-01
9.53144073e-01 -1.79971471e-01 -2.02710614e-01 -8.13659072e-01
-1.30051303e+00 1.03852797e+00 -2.74592906e-01 2.92010397e-01
-2.93115854e-01 1.55731067e-01 3.32456112e-01 3.37432712e-01
4.55885798e-01 2.22564191e-01 -4.63824689e-01 4.75532524e-02
-8.20438445e-01 3.66897970e-01 1.03992879e+00 7.34093666e-01
4.66571808e-01 -2.37623021e-01 -2.41602510e-02 6.47439301e-01
1.83711767e-01 6.46655500e-01 -1.74621120e-01 -1.10407937e+00
1.17718831e-01 -4.87042516e-02 6.05176091e-01 -1.03072000e+00
-5.00010848e-01 -4.72118556e-01 -1.10933518e+00 1.13201663e-01
9.57829058e-01 -4.34940338e-01 -2.96485513e-01 1.92224193e+00
3.11905026e-01 -2.30145194e-02 -1.07571505e-01 5.09106159e-01
-2.00069863e-02 4.20053810e-01 -1.06669493e-01 -8.40854466e-01
1.18722296e+00 -3.67156938e-02 -7.81768501e-01 1.99474469e-01
5.55356920e-01 -5.56372702e-01 6.59909666e-01 6.25743270e-01
-9.70582783e-01 -1.45882219e-01 -1.05679584e+00 3.36910248e-01
-2.19706818e-02 8.82158577e-02 6.49894536e-01 9.45761800e-01
-6.44860446e-01 7.08656788e-01 -5.63289702e-01 -9.72777233e-02
1.31677210e-01 4.30234998e-01 -3.02612811e-01 1.47582022e-02
-8.69806230e-01 7.93507218e-01 3.47314686e-01 1.06565721e-01
-2.86444962e-01 -8.28166127e-01 -7.49617398e-01 -4.65223752e-02
3.58910203e-01 -6.61738396e-01 1.19650817e+00 -6.64114714e-01
-1.22524929e+00 5.69623590e-01 -4.46886830e-02 -2.68774867e-01
5.14999390e-01 4.52015623e-02 -1.83214545e-01 -8.81352741e-03
2.08741397e-01 4.09375690e-02 8.92173767e-01 -8.41732264e-01
-4.66854662e-01 -5.14282882e-01 -2.07734838e-01 -2.70265847e-01
1.76095143e-01 1.06068281e-02 -1.04823321e-01 -6.70801342e-01
2.32486412e-01 -8.98145556e-01 -5.73164999e-01 1.21215008e-01
-6.66185200e-01 -7.39932433e-02 -2.76724339e-01 -3.02795291e-01
1.40267134e+00 -2.10950136e+00 -6.63609430e-02 7.18639731e-01
-1.47000579e-02 -2.45545536e-01 2.92791456e-01 4.05047625e-01
-2.14770406e-01 8.15746337e-02 -5.72152078e-01 3.83301467e-01
1.16845421e-01 -1.14254805e-03 -1.39469668e-01 8.62141013e-01
5.75872958e-02 2.96813339e-01 -5.84963441e-01 -6.12422407e-01
7.51156881e-02 2.67405450e-01 -8.36641431e-01 -6.44014925e-02
7.22515881e-02 2.90293723e-01 -5.25055289e-01 2.64391363e-01
8.07644546e-01 -2.46993214e-01 4.68689442e-01 -1.21117234e-02
-2.87025005e-01 1.30368611e-02 -1.63529491e+00 1.21303928e+00
-2.17903480e-01 2.83285499e-01 3.28250900e-02 -1.42155766e+00
8.08217585e-01 2.84476399e-01 4.95549977e-01 -1.75814569e-01
3.16603482e-01 4.05363083e-01 -1.46143556e-01 -3.65946591e-01
-6.84092492e-02 -6.46483183e-01 -3.02291632e-01 3.42727900e-01
-1.75778329e-01 7.84486532e-02 1.84530824e-01 -3.67383838e-01
9.36914623e-01 -3.04675430e-01 8.39351296e-01 -7.33909249e-01
4.23834592e-01 -2.24408060e-01 5.80240190e-01 8.57798874e-01
1.60845533e-01 4.79214251e-01 1.19534552e+00 -1.87308788e-02
-1.09297955e+00 -1.21981347e+00 -9.23134148e-01 4.67032641e-01
-2.88517952e-01 -9.23503786e-02 -7.92908609e-01 -3.47555041e-01
3.02961379e-01 6.76879406e-01 -8.51863384e-01 9.98307616e-02
-3.81114148e-02 -1.05494606e+00 1.96164414e-01 3.02317083e-01
6.72069788e-02 -2.37934679e-01 -3.90530050e-01 2.35065311e-01
2.63403505e-02 -7.52968907e-01 -4.07650739e-01 2.97736555e-01
-1.14386654e+00 -1.26831758e+00 -8.71034980e-01 -4.65709150e-01
7.99479187e-01 -1.51851829e-02 9.03844774e-01 -3.31236362e-01
-9.04664025e-02 1.31693259e-01 -1.43300556e-02 -5.19458771e-01
-3.97685260e-01 -4.07539129e-01 1.41719982e-01 -7.16910809e-02
2.47551337e-01 -5.49130857e-01 -5.77974200e-01 5.02853215e-01
-6.89076722e-01 -5.51440835e-01 7.00342119e-01 8.73596489e-01
6.88393474e-01 2.78216749e-01 8.95308852e-01 -9.61470306e-01
5.88399947e-01 -6.58272266e-01 -1.02125132e+00 1.99028909e-01
-7.68482268e-01 4.47765797e-01 3.84299040e-01 -3.84142399e-01
-5.82760572e-01 2.68858764e-02 7.71871954e-02 1.00633144e-01
-7.99330845e-02 6.14829242e-01 -5.37541248e-02 1.68802485e-01
6.10904038e-01 -2.16698259e-01 1.58411175e-01 -4.48584139e-01
9.35304686e-02 5.08555233e-01 4.93375957e-01 -6.28219903e-01
5.17544627e-01 1.64184123e-01 5.51059246e-01 -9.22437429e-01
-7.39031792e-01 -3.00535321e-01 -3.40213299e-01 1.94701880e-01
6.43413603e-01 -5.53058028e-01 -1.07809782e+00 -6.59702644e-02
-8.38096857e-01 4.61801998e-02 -1.71858177e-01 9.69242454e-01
-9.06197369e-01 4.52477932e-01 -1.12063745e-02 -1.06989920e+00
9.65908468e-02 -8.75136375e-01 7.43337154e-01 -2.41310552e-01
-3.68789583e-01 -9.42488551e-01 3.04926187e-01 -1.19566374e-01
7.82922134e-02 4.73027289e-01 1.08401740e+00 -5.27996838e-01
-1.14747256e-01 -4.82769996e-01 -3.90300274e-01 3.59697610e-01
1.61008053e-02 -3.41280289e-02 -5.55053592e-01 -1.76904410e-01
1.15442924e-01 -3.53509225e-02 4.42941338e-01 1.04999125e+00
1.37454855e+00 -4.10775065e-01 -3.09471846e-01 2.44456857e-01
1.35183704e+00 9.91885960e-02 4.81289834e-01 -1.04641691e-01
-7.65147153e-04 6.90999508e-01 7.44724631e-01 7.74604321e-01
7.98010901e-02 6.99500561e-01 -2.71333009e-02 3.66963655e-01
7.51142919e-01 -1.02805495e-02 9.97134596e-02 4.59030479e-01
-1.80701092e-01 3.80932875e-02 -6.02057397e-01 3.71347696e-01
-1.63433623e+00 -9.25764918e-01 -2.04653531e-01 3.02981687e+00
1.10280538e+00 2.53189858e-02 4.13116544e-01 3.61915261e-01
8.31940472e-01 -4.47942466e-01 -3.20501626e-01 -4.13809359e-01
1.45663723e-01 2.47844964e-01 7.60233998e-01 4.83422607e-01
-1.00216687e+00 -8.60402137e-02 7.56760168e+00 6.90859914e-01
-5.64897478e-01 -2.02007383e-01 5.72950840e-01 1.97583482e-01
-2.75764048e-01 1.12036252e-02 -7.20320463e-01 4.67093974e-01
1.14757645e+00 -5.05530536e-01 -6.26558140e-02 8.41727912e-01
3.21501553e-01 -6.06148899e-01 -1.23413754e+00 9.84138191e-01
-2.65084833e-01 -9.41034794e-01 -4.49341118e-01 3.71517479e-01
4.48969066e-01 -6.44134998e-01 3.49983089e-02 -5.39215840e-02
9.02924985e-02 -8.68722498e-01 3.26853216e-01 6.16760314e-01
9.24408615e-01 -1.09664524e+00 7.95805037e-01 3.77953619e-01
-5.25510311e-01 1.05024077e-01 -7.20195413e-01 5.09541249e-03
-4.83637564e-02 1.25765157e+00 -7.82482088e-01 2.50931978e-01
2.92004913e-01 2.12097317e-01 -1.09379962e-01 1.41759717e+00
8.06520581e-02 6.83908224e-01 -7.22260714e-01 -1.00264341e-01
-1.49605811e-01 -6.29073501e-01 3.47651511e-01 1.10577095e+00
7.95502841e-01 1.10314347e-01 -2.31484041e-01 6.86389327e-01
2.23737344e-01 4.17294294e-01 -7.97215164e-01 -9.53257903e-02
4.14188236e-01 7.47670174e-01 -7.51916766e-01 -1.01996828e-02
-5.18020391e-01 4.84299541e-01 -6.95850104e-02 1.09071225e-01
-6.43053234e-01 -4.08164114e-01 5.35954356e-01 1.11987283e-02
4.57563937e-01 -2.10159525e-01 -4.06026512e-01 -1.06235981e+00
5.44633381e-02 -9.05980587e-01 4.53813881e-01 -2.96011746e-01
-1.21937752e+00 -3.92322503e-02 2.88290143e-01 -1.21520758e+00
-6.53720379e-01 -6.62671089e-01 -1.74226910e-01 8.27259302e-01
-9.32104528e-01 -2.39293694e-01 3.00707430e-01 3.21990222e-01
-1.94439024e-01 3.09570998e-01 1.03461766e+00 1.72393888e-01
-6.24786556e-01 6.11420512e-01 4.77620602e-01 -2.66765803e-01
5.90352952e-01 -1.31834817e+00 -3.30702245e-01 6.41103983e-01
-3.53720576e-01 6.47521079e-01 1.29974639e+00 -6.61403656e-01
-1.35649800e+00 -5.64956367e-01 9.55735385e-01 -3.27768773e-01
9.69373345e-01 -2.20531002e-01 -7.00655520e-01 5.53853273e-01
-3.71093839e-01 -1.26516789e-01 9.34489846e-01 5.38721263e-01
-2.00692162e-01 -2.25453511e-01 -1.16397142e+00 3.55998248e-01
6.69287264e-01 1.26800751e-02 -4.56259161e-01 3.95770401e-01
2.33578891e-01 7.54257664e-02 -1.28598547e+00 4.98681307e-01
8.25837314e-01 -9.25695360e-01 9.78276432e-01 -5.82336724e-01
2.28532642e-01 -1.24749795e-01 -4.98089433e-01 -1.05413973e+00
-5.69734462e-02 -6.73113227e-01 2.16738105e-01 1.02684236e+00
5.08156180e-01 -6.83721364e-01 5.15095055e-01 6.61825418e-01
3.12324375e-01 -7.36264408e-01 -9.95918512e-01 -9.33302641e-01
2.39659801e-01 -6.72913492e-01 2.55324662e-01 8.58371139e-01
2.77442008e-01 2.15725467e-01 -3.27223390e-01 2.78950959e-01
1.24181128e+00 1.94033638e-01 5.87799668e-01 -1.55070388e+00
-4.91991967e-01 -3.50259602e-01 -5.87888062e-01 -6.99962795e-01
-3.76933590e-02 -4.52858716e-01 2.39551775e-02 -9.84173000e-01
3.36539328e-01 -5.89570284e-01 -2.03596756e-01 6.89107552e-02
8.83697066e-03 9.22578722e-02 -4.89125431e-01 -2.16047406e-01
-3.23337168e-01 2.50795841e-01 9.87484813e-01 3.87464195e-01
-1.57009870e-01 6.81367636e-01 -8.48898590e-01 8.06263030e-01
9.56819892e-01 -5.65273643e-01 -4.46890622e-01 3.22391957e-01
7.00927913e-01 5.77850282e-01 3.56904417e-01 -7.05793023e-01
-5.58224134e-02 -2.34766677e-01 4.82686967e-01 -6.82398558e-01
-2.61631191e-01 -7.01305985e-01 4.32024240e-01 4.17985737e-01
-2.56238103e-01 -6.53455555e-02 1.79129809e-01 6.37477398e-01
1.13066524e-01 -6.59720778e-01 9.01711404e-01 1.86988592e-01
-6.51109368e-02 -2.39026975e-02 -4.88192916e-01 6.22816496e-02
8.59950483e-01 -3.49504203e-02 8.44524726e-02 -4.00023103e-01
-9.05196071e-01 1.16176605e-01 4.05613065e-01 -2.12799817e-01
4.03285950e-01 -1.29561377e+00 -6.64935410e-01 1.23365961e-01
1.04865767e-01 -3.79351228e-01 1.69861227e-01 1.20520580e+00
-3.95900190e-01 5.24604738e-01 -2.35173367e-02 -5.19627571e-01
-1.03787875e+00 8.26198757e-01 3.71135436e-02 4.53078514e-03
-3.80434424e-01 2.75267094e-01 3.96368921e-01 -1.50051996e-01
9.50684324e-02 -4.35497820e-01 1.37136713e-01 1.28664434e-01
7.46805370e-01 6.50310457e-01 -1.50656432e-01 -1.39138490e-01
-2.11407438e-01 6.60185695e-01 2.44761452e-01 -3.26718807e-01
1.35176516e+00 -3.09683561e-01 -3.09670925e-01 7.04113543e-01
1.31472063e+00 1.88587144e-01 -9.70853329e-01 -7.80771598e-02
1.85854852e-01 -2.96681404e-01 -6.92170858e-02 -5.73795855e-01
-5.75515449e-01 5.39284408e-01 3.50662410e-01 3.90338182e-01
1.10904241e+00 -1.03079667e-02 1.04636187e-02 5.61089635e-01
4.81786400e-01 -9.22426701e-01 -4.61733788e-01 -2.20160242e-02
9.80684042e-01 -1.21791589e+00 2.85024256e-01 -5.99618733e-01
-2.82548428e-01 9.91699994e-01 -1.44601136e-01 -1.18387669e-01
9.06541109e-01 2.45160118e-01 -2.79894263e-01 1.03613645e-01
-5.16256154e-01 -6.55898824e-02 3.40198517e-01 4.90260303e-01
5.66349685e-01 1.97356015e-01 -8.89756799e-01 5.96649706e-01
-3.28734159e-01 1.84265494e-01 6.02148890e-01 5.62174022e-01
-5.71873546e-01 -1.01020122e+00 -5.56833744e-01 6.26619399e-01
-8.18693876e-01 6.57207370e-02 9.30410400e-02 9.83812451e-01
-2.61938542e-01 9.95863020e-01 -1.53114125e-02 7.70143121e-02
1.54629931e-01 1.17475241e-01 7.40186632e-01 -3.42303097e-01
4.66344148e-01 2.26030856e-01 1.98278651e-01 -5.55447459e-01
-3.74923259e-01 -1.05365849e+00 -9.20999110e-01 -6.67336881e-01
-5.01901388e-01 4.87250090e-01 7.22029746e-01 7.29124904e-01
-3.23736630e-02 5.71925528e-02 6.93415523e-01 -4.43188846e-01
-1.00486827e+00 -6.36173904e-01 -1.06512165e+00 8.66532326e-04
4.46809322e-01 -8.35811019e-01 -4.52233732e-01 -3.00811857e-01] | [7.506474018096924, 4.493092060089111] |
e0c04221-94bd-4f11-8e2b-6c5ffcd10ec9 | evaluating-the-consistency-of-word-embeddings | null | null | https://aclanthology.org/R19-1016 | https://aclanthology.org/R19-1016.pdf | Evaluating the Consistency of Word Embeddings from Small Data | In this work, we address the evaluation of distributional semantic models trained on smaller, domain-specific texts, specifically, philosophical text. Specifically, we inspect the behaviour of models using a pre-trained background space in learning. We propose a measure of consistency which can be used as an evaluation metric when no in-domain gold-standard data is available. This measure simply computes the ability of a model to learn similar embeddings from different parts of some homogeneous data. We show that in spite of being a simple evaluation, consistency actually depends on various combinations of factors, including the nature of the data itself, the model used to train the semantic space, and the frequency of the learnt terms, both in the background space and in the in-domain data of interest. | ["Aur{\\'e}lie Herbelot", 'Antske Fokkens', 'Jelke Bloem'] | 2019-09-01 | null | null | null | ranlp-2019-9 | ['small-data'] | ['computer-vision'] | [ 4.84687537e-02 -3.49365128e-03 -1.69218093e-01 -7.31796980e-01
-6.10915720e-01 -6.18925512e-01 9.90170658e-01 7.04877615e-01
-9.37619984e-01 5.02314031e-01 4.57513630e-01 -1.12789981e-01
-9.95995775e-02 -7.27230549e-01 -5.14246762e-01 -7.39797890e-01
3.80469918e-01 7.42514074e-01 3.57739300e-01 -3.43610168e-01
2.50575542e-01 1.83628753e-01 -1.19612646e+00 1.55160472e-01
6.45878553e-01 7.24812508e-01 2.81678200e-01 4.07762110e-01
-4.01564628e-01 5.12321889e-01 -7.14601517e-01 -4.01339322e-01
-1.59034625e-01 -6.09297276e-01 -8.98441255e-01 -2.69166797e-01
5.60821533e-01 3.37698758e-01 -2.76093513e-01 1.23693585e+00
3.28936577e-01 3.17990631e-01 1.07370210e+00 -8.00170004e-01
-7.02735543e-01 6.69895530e-01 1.17846854e-01 3.88917774e-01
1.80523768e-01 -2.54868597e-01 1.26413953e+00 -8.29394758e-01
9.19131577e-01 1.04883814e+00 6.18477702e-01 6.49419725e-01
-1.32088125e+00 -2.76725680e-01 -4.19216082e-02 3.56897771e-01
-1.12837660e+00 -1.80751041e-01 8.41243029e-01 -4.43746448e-01
5.17007232e-01 1.55482711e-02 3.05136323e-01 1.24930441e+00
9.66489762e-02 5.61320424e-01 1.09230244e+00 -7.44781077e-01
7.15800107e-01 6.07434273e-01 6.21373117e-01 2.48460740e-01
2.67018437e-01 -1.27197996e-01 -4.06449288e-01 -4.07971203e-01
3.23801428e-01 -3.56012970e-01 -2.97178119e-01 -8.87128472e-01
-7.71231651e-01 1.00888145e+00 2.50842184e-01 9.38784719e-01
-2.27530986e-01 -6.11434691e-02 5.04888773e-01 2.36170739e-01
6.49386406e-01 5.31514347e-01 -7.92522490e-01 -2.05493733e-01
-8.19607973e-01 2.49542236e-01 8.73114288e-01 5.01836002e-01
7.46873319e-01 -3.93089592e-01 2.16299489e-01 7.74320841e-01
1.24721505e-01 1.79252490e-01 8.95671427e-01 -5.33596456e-01
3.80275518e-01 5.09256065e-01 -9.76954624e-02 -7.39782095e-01
-2.20343098e-01 -3.56413275e-01 -1.46309227e-01 -4.97925170e-02
6.66680992e-01 -8.94376114e-02 -5.56098163e-01 2.09005523e+00
3.76835138e-01 -2.65431870e-03 3.79985213e-01 6.38074160e-01
5.35471201e-01 4.74039823e-01 2.59445041e-01 1.62301913e-01
9.42849636e-01 -7.54871070e-01 -5.22148192e-01 -2.69430876e-01
8.98181379e-01 -5.44476569e-01 1.32219732e+00 -1.08097054e-01
-6.41996443e-01 -5.91025054e-01 -9.98815417e-01 -1.84303507e-01
-8.43986630e-01 -2.17113346e-01 -6.64438233e-02 4.36985642e-01
-8.48452449e-01 7.87143171e-01 -3.96171123e-01 -7.50784576e-01
9.61116701e-02 -1.80161834e-01 -5.29587924e-01 -2.35228747e-01
-1.36092973e+00 1.47660291e+00 6.31505072e-01 -6.09269202e-01
-8.25147629e-01 -8.02835107e-01 -8.42802405e-01 1.32102117e-01
1.48388445e-01 -3.58960152e-01 1.00124443e+00 -1.06307650e+00
-9.71107244e-01 1.13816774e+00 -1.41165212e-01 -4.60656196e-01
4.57887024e-01 -2.02223331e-01 -4.88727242e-01 -7.18286857e-02
1.56666428e-01 -4.55436260e-02 7.32966721e-01 -1.13274658e+00
-3.40351045e-01 -6.49837255e-01 3.83705236e-02 1.62050575e-01
-4.25044835e-01 -5.67993429e-03 -1.16269916e-01 -6.21095121e-01
-2.00187370e-01 -8.86873484e-01 -5.78848571e-02 -1.18852995e-01
1.36719137e-01 -5.81052542e-01 6.36416554e-01 -6.02680624e-01
1.11785591e+00 -2.41821408e+00 2.18867496e-01 3.02643955e-01
-1.05623133e-01 2.01157495e-01 -6.70599788e-02 4.99398023e-01
-1.39790982e-01 -4.54790778e-02 -3.28300804e-01 -2.58497119e-01
2.02148750e-01 2.97944605e-01 -3.03186327e-01 5.40154994e-01
1.52964294e-01 6.62328780e-01 -1.07479310e+00 -4.22190130e-01
2.30336338e-01 4.03211087e-01 -3.38602096e-01 1.76037773e-01
-2.04091102e-01 8.91027749e-02 -3.40046853e-01 -8.44209120e-02
1.95861444e-01 -6.72375113e-02 3.05854708e-01 -1.34734794e-01
9.29722711e-02 7.58070290e-01 -1.00651240e+00 1.85530066e+00
-5.78185320e-01 7.58432209e-01 -2.95584112e-01 -1.28211617e+00
7.31141627e-01 1.43726349e-01 9.45914239e-02 -6.93300605e-01
3.30502063e-01 4.08966780e-01 4.59280573e-02 -3.31633508e-01
4.88809496e-01 -4.80576068e-01 -1.48455173e-01 5.21881938e-01
5.44950843e-01 -1.69758439e-01 1.81964412e-01 6.49377285e-03
1.04563427e+00 6.16657250e-02 2.64438421e-01 -7.83195674e-01
3.37911308e-01 7.30160177e-02 1.07391506e-01 5.52723050e-01
-9.70716178e-02 5.58983564e-01 3.56710434e-01 -3.24397117e-01
-1.07557642e+00 -1.15392864e+00 -7.76182234e-01 1.30131447e+00
3.16771775e-01 -6.51026785e-01 -6.70012355e-01 -1.02119875e+00
-1.64813478e-03 1.15647995e+00 -1.00017166e+00 -3.87576491e-01
-3.85054559e-01 -7.30054200e-01 3.66791844e-01 4.57217515e-01
-1.42505001e-02 -6.90812290e-01 -4.49529916e-01 1.34907126e-01
1.35741234e-01 -9.68219221e-01 -2.74828523e-01 4.97217238e-01
-9.63842571e-01 -1.10515428e+00 -4.73501354e-01 -6.49718583e-01
5.57669520e-01 -9.23863202e-02 1.41114151e+00 -2.24304765e-01
-5.23603633e-02 4.01465058e-01 -3.06393892e-01 -3.65227371e-01
-4.68022943e-01 3.82187814e-02 -1.30186379e-01 -1.58049971e-01
7.85944104e-01 -1.96667820e-01 -2.11445279e-02 -7.95233622e-02
-1.11075246e+00 -4.79146600e-01 8.20993111e-02 9.08448458e-01
3.74375015e-01 8.28540400e-02 7.97226727e-02 -1.08927381e+00
7.06128418e-01 -6.86096013e-01 -1.53175473e-01 4.98978168e-01
-4.89831746e-01 4.29705262e-01 5.15036285e-01 -4.58001852e-01
-7.63357759e-01 -3.89731646e-01 -6.66103605e-03 -1.49196401e-01
-1.10083573e-01 5.66090226e-01 -2.25094095e-01 3.90363485e-01
8.43792796e-01 8.49968940e-02 -1.77342460e-01 -6.90514624e-01
4.73303527e-01 5.00247419e-01 2.39657030e-01 -7.70735681e-01
6.70660019e-01 2.56670535e-01 -2.44320869e-01 -9.45154965e-01
-9.63733494e-01 -5.89568079e-01 -6.95466638e-01 2.47096345e-01
8.01013947e-01 -5.82287312e-01 3.33029091e-01 1.22224316e-01
-1.01411927e+00 -3.66636366e-01 -5.17098606e-01 5.19444883e-01
-3.72095942e-01 1.21842317e-01 -7.99900591e-02 -4.98885602e-01
1.11464828e-01 -5.93800128e-01 8.01223338e-01 1.24080945e-03
-6.89999342e-01 -1.69315910e+00 6.00963295e-01 -1.73156321e-01
4.29339290e-01 5.91116324e-02 1.40530825e+00 -1.44654906e+00
1.40151501e-01 -3.47631007e-01 2.83384379e-02 7.33947635e-01
2.92039454e-01 -7.52939433e-02 -1.14464188e+00 -9.28522050e-02
4.33580667e-01 -3.32591742e-01 8.01987946e-01 1.35154858e-01
9.03950036e-01 -1.68984696e-01 -2.77639240e-01 2.26187080e-01
1.62020445e+00 -1.64735347e-01 2.92211920e-01 5.27405739e-01
5.56189954e-01 8.95656765e-01 4.26030457e-01 -6.68486208e-02
2.60146588e-01 7.26035476e-01 -3.41161005e-02 -1.37349609e-02
-5.12863360e-02 -3.63152385e-01 1.64909989e-01 8.80086958e-01
2.60403872e-01 -5.65813854e-02 -1.10547531e+00 7.84866214e-01
-1.55545926e+00 -8.25749874e-01 2.38827705e-01 2.27896428e+00
1.09654403e+00 1.41494483e-01 -1.22079132e-02 2.99814297e-03
7.91354835e-01 9.88167301e-02 -2.80718893e-01 -3.74062926e-01
-3.28697443e-01 6.37361407e-01 2.66981781e-01 6.25486910e-01
-1.00933659e+00 6.83384955e-01 6.58561087e+00 8.84296417e-01
-1.08115196e+00 3.69733959e-01 2.96766043e-01 5.92345931e-02
-5.87879002e-01 -4.23634574e-02 -5.54879069e-01 6.82955980e-01
1.16165495e+00 -3.93919230e-01 -1.37154758e-01 1.10411203e+00
-4.51874524e-01 7.35479444e-02 -1.45021594e+00 5.78201115e-01
3.32415462e-01 -1.05190039e+00 1.70199707e-01 -6.29596179e-03
5.40099680e-01 1.90423116e-01 -2.05420349e-02 2.92199492e-01
2.81146139e-01 -9.01803792e-01 8.44263315e-01 1.17788464e-01
3.21642816e-01 -4.99726921e-01 9.31159735e-01 3.53090674e-01
-5.36620796e-01 2.57795811e-01 -4.33135390e-01 1.83081597e-01
-2.11260766e-01 6.02352321e-01 -7.02736318e-01 3.83162320e-01
3.88350844e-01 5.48046470e-01 -6.07527673e-01 6.99742496e-01
-2.11645782e-01 5.84243000e-01 -1.80680364e-01 -2.35660434e-01
3.61583233e-01 -7.28739575e-02 3.57621908e-01 1.31434345e+00
1.22654907e-01 -2.47678742e-01 -8.01826343e-02 7.79819846e-01
-1.26293348e-02 3.24520320e-01 -6.28967166e-01 -1.30678430e-01
4.07482475e-01 9.49492633e-01 -2.11081013e-01 -3.85442853e-01
-5.39375663e-01 7.61909664e-01 6.04060709e-01 2.58903116e-01
-7.22548127e-01 -2.62029082e-01 7.92814195e-01 1.46447107e-01
2.96234190e-01 -2.58075356e-01 -1.38125241e-01 -1.22741854e+00
1.23480782e-01 -5.64794600e-01 3.18043232e-01 -4.37090576e-01
-1.65093732e+00 6.16607130e-01 1.38562202e-01 -7.12151766e-01
-2.24634528e-01 -9.20515418e-01 -6.22200072e-01 9.59193051e-01
-1.44131243e+00 -6.61768198e-01 -1.18869748e-02 4.46142077e-01
3.47518295e-01 -2.14174032e-01 1.02470660e+00 -5.21817617e-02
-2.34847337e-01 3.55621248e-01 5.94773889e-01 3.47008348e-01
7.97162533e-01 -1.40074658e+00 2.71984756e-01 5.81019819e-01
6.03869140e-01 7.28194833e-01 1.08551109e+00 -3.17955792e-01
-8.67550433e-01 -8.16938519e-01 1.19720435e+00 -8.16457927e-01
1.04996395e+00 -3.23540211e-01 -1.18205547e+00 5.74688494e-01
9.16564912e-02 -3.36174369e-02 9.90153849e-01 4.42294806e-01
-6.45459414e-01 6.70768395e-02 -1.01813352e+00 4.47796524e-01
7.69098878e-01 -1.04801607e+00 -1.38648677e+00 1.01202913e-01
4.37791497e-01 -1.33483857e-01 -7.04626143e-01 2.91511446e-01
2.99583077e-01 -7.00419247e-01 7.70795763e-01 -9.80266333e-01
4.11876470e-01 7.99770206e-02 -5.94783843e-01 -1.50629044e+00
-2.42056921e-01 1.01914421e-01 1.37267724e-01 1.22296321e+00
4.01250988e-01 -4.51702684e-01 5.34518540e-01 8.38834524e-01
1.80436090e-01 -5.22326648e-01 -9.59187746e-01 -8.51666868e-01
5.82199097e-01 -5.01639366e-01 3.70533139e-01 1.17291319e+00
1.68339297e-01 6.66136205e-01 3.51886243e-01 -2.33633131e-01
4.19273227e-01 2.95000635e-02 4.38160896e-01 -1.42184544e+00
-7.20883831e-02 -3.15311551e-01 -5.47326982e-01 -6.35270417e-01
5.51217854e-01 -1.02877367e+00 2.32550390e-02 -1.15561736e+00
1.32387176e-01 -5.83927691e-01 -6.34235322e-01 6.00101948e-02
-2.96139568e-01 -8.62788185e-02 2.98336763e-02 1.74006656e-01
-4.59278226e-01 5.44190526e-01 6.55499279e-01 -6.13319464e-02
1.70989469e-01 -4.39505398e-01 -5.08891344e-01 8.06362808e-01
7.85805106e-01 -7.72853613e-01 -3.37380558e-01 -5.01818955e-01
2.05729492e-02 -7.14351118e-01 3.76159310e-01 -8.97110105e-01
2.37172134e-02 -5.61074428e-02 3.36266637e-01 1.66169405e-02
2.08295286e-01 -8.85288537e-01 -1.58464968e-01 2.93410957e-01
-8.25060666e-01 8.09827745e-02 1.56029508e-01 5.24848521e-01
-4.08603728e-01 -8.43613267e-01 8.17545056e-01 -1.13835707e-01
-8.22948694e-01 -4.62461486e-02 1.53392588e-03 6.93491399e-01
8.27440441e-01 3.39074596e-03 -3.16106498e-01 4.85740229e-02
-6.04722321e-01 -2.73549438e-01 9.26555216e-01 4.93421495e-01
1.35513142e-01 -1.32791734e+00 -6.93947196e-01 -6.07958296e-03
4.32807297e-01 -3.07127684e-01 -1.49535462e-01 3.45722824e-01
-2.65478134e-01 3.60379368e-01 -5.79538639e-04 -4.42586929e-01
-1.06677127e+00 4.82641220e-01 2.50560552e-01 -1.16730683e-01
-3.17045927e-01 6.03790283e-01 3.56808215e-01 -4.47681308e-01
2.23307952e-01 -3.07566226e-01 -5.87691888e-02 2.40514219e-01
4.27903205e-01 8.26203376e-02 1.65528998e-01 -7.46164262e-01
-4.82280254e-01 4.77916002e-01 -8.06568470e-03 -1.79903850e-01
1.32800674e+00 -1.85794793e-02 -1.29193634e-01 1.13314617e+00
1.51013851e+00 -1.35247165e-03 -8.34863067e-01 -7.47996747e-01
4.01522249e-01 -3.62137735e-01 7.62062594e-02 -5.83090425e-01
-5.41857839e-01 9.39114451e-01 5.26638091e-01 4.40775633e-01
5.07047415e-01 3.28397512e-01 3.83087516e-01 2.18762457e-01
2.39451304e-01 -1.51939011e+00 1.26597080e-02 6.63722694e-01
6.37100399e-01 -9.96527672e-01 -2.34640062e-01 1.30782783e-01
-6.66225314e-01 1.10573280e+00 2.81945825e-01 -3.36356133e-01
8.81418347e-01 2.43849978e-02 2.43796363e-01 -1.72735721e-01
-6.18251026e-01 -1.43777639e-01 4.73997951e-01 6.39201880e-01
6.07731879e-01 -9.53770205e-02 -2.90991545e-01 5.36190212e-01
-2.79309630e-01 -2.70416468e-01 1.27931461e-01 8.17550182e-01
-5.07554948e-01 -1.17190254e+00 8.39288905e-03 1.29447445e-01
-5.06286144e-01 -2.27086052e-01 -5.89299738e-01 8.36418033e-01
1.96285173e-01 4.87061739e-01 2.31058627e-01 -8.79822373e-02
2.86904305e-01 5.33373177e-01 5.62431514e-01 -8.98683310e-01
-3.07561874e-01 -5.46992779e-01 1.01533517e-01 -3.18325460e-01
-6.53309107e-01 -6.44937992e-01 -9.98301864e-01 -5.80792949e-02
-1.74499080e-01 4.48961705e-01 8.03395212e-01 1.33013535e+00
3.02083790e-02 1.63335413e-01 3.21704209e-01 -3.72540712e-01
-8.10908437e-01 -1.08016300e+00 -6.61679626e-01 1.00754774e+00
2.85668105e-01 -6.69137180e-01 -6.05147541e-01 -1.95129499e-01] | [10.487313270568848, 8.975203514099121] |
ffcb3ca5-2861-454b-a3c5-096beeca1619 | isolation-distributional-kernel-a-new-tool | 2009.12196 | null | https://arxiv.org/abs/2009.12196v1 | https://arxiv.org/pdf/2009.12196v1.pdf | Isolation Distributional Kernel: A New Tool for Point & Group Anomaly Detection | We introduce Isolation Distributional Kernel as a new way to measure the similarity between two distributions. Existing approaches based on kernel mean embedding, which convert a point kernel to a distributional kernel, have two key issues: the point kernel employed has a feature map with intractable dimensionality; and it is {\em data independent}. This paper shows that Isolation Distributional Kernel (IDK), which is based on a {\em data dependent} point kernel, addresses both key issues. We demonstrate IDK's efficacy and efficiency as a new tool for kernel based anomaly detection for both point and group anomalies. Without explicit learning, using IDK alone outperforms existing kernel based point anomaly detector OCSVM and other kernel mean embedding methods that rely on Gaussian kernel. For group anomaly detection,we introduce an IDK based detector called IDK$^2$. It reformulates the problem of group anomaly detection in input space into the problem of point anomaly detection in Hilbert space, without the need for learning. IDK$^2$ runs orders of magnitude faster than group anomaly detector OCSMM.We reveal for the first time that an effective kernel based anomaly detector based on kernel mean embedding must employ a characteristic kernel which is data dependent. | ['Zhi-Hua Zhou', 'Bi-Cun Xu', 'Takashi Washio', 'Kai Ming Ting'] | 2020-09-24 | null | null | null | null | ['group-anomaly-detection'] | ['methodology'] | [-3.89903843e-01 -2.00351492e-01 1.49670407e-01 -3.59451234e-01
-7.82475114e-01 -6.65796459e-01 5.91881454e-01 4.87210572e-01
-3.79439443e-01 5.08834198e-02 -4.72108066e-01 -8.09055030e-01
-5.76111436e-01 -7.31618822e-01 -5.27230024e-01 -9.51536238e-01
-5.36010504e-01 2.92896003e-01 4.28972334e-01 -2.12105528e-01
6.90338731e-01 7.45614171e-01 -1.28417039e+00 -1.20396234e-01
8.36450160e-01 1.12967265e+00 -5.59335232e-01 9.36604023e-01
-4.94354546e-01 3.10300112e-01 -5.16813695e-01 -4.38330285e-02
3.57277304e-01 -2.77028978e-01 -7.22457230e-01 -3.55271399e-01
5.62462866e-01 -1.35399923e-01 -2.51895875e-01 1.23457754e+00
3.29983652e-01 2.49605551e-01 1.44427812e+00 -1.96896601e+00
-1.12415051e+00 2.82104481e-02 -6.16649628e-01 5.04710197e-01
1.97561353e-01 -2.57229716e-01 8.48121881e-01 -1.10749519e+00
-2.78701991e-01 9.23719287e-01 1.08130717e+00 3.19407403e-01
-1.67144597e+00 -4.46001142e-01 -2.28021488e-01 2.35997736e-01
-1.49577785e+00 3.91581543e-02 8.54231596e-01 -8.02085459e-01
1.11898363e+00 3.87791216e-01 3.83448124e-01 6.47780478e-01
3.57664317e-01 6.44049585e-01 9.69515741e-01 -6.32021666e-01
5.74697316e-01 2.01521829e-01 7.67740548e-01 7.58434236e-01
3.18935513e-01 1.65721387e-01 -1.31362990e-01 -9.88544047e-01
6.00198805e-01 3.29290301e-01 -9.27118883e-02 -7.68317401e-01
-1.00801373e+00 1.21232414e+00 -7.37157986e-02 3.12493145e-01
1.13819018e-01 3.31472158e-01 6.49455309e-01 8.08181345e-01
5.16943336e-01 5.51355958e-01 -5.59846699e-01 -2.14917973e-01
-6.13040864e-01 8.19599256e-02 7.63380289e-01 8.33835304e-01
8.90348375e-01 1.32836148e-01 1.65209934e-01 7.19368875e-01
4.16800916e-01 8.21519256e-01 8.09771717e-01 -4.36007142e-01
-1.15564592e-01 9.04542387e-01 -8.08390230e-02 -1.05343664e+00
-2.33656734e-01 4.10869807e-01 -7.38453507e-01 6.26013458e-01
4.43408489e-01 1.87495455e-01 -6.68079138e-01 1.44558799e+00
3.09928298e-01 3.66598636e-01 1.98748469e-01 1.87766626e-01
2.19555721e-01 5.10989249e-01 -3.28794301e-01 7.06023052e-02
9.79970276e-01 -4.85184848e-01 -6.22727275e-01 5.33711433e-01
1.41000855e+00 -5.21374822e-01 1.21051133e+00 2.56384104e-01
-4.98555541e-01 -2.06281394e-01 -1.22788954e+00 3.72252345e-01
-1.14996684e+00 -1.01949453e-01 6.65247440e-01 8.49188089e-01
-1.28233695e+00 7.06179023e-01 -7.88010240e-01 -5.16699076e-01
8.97685140e-02 5.95826387e-01 -7.25003898e-01 4.31550175e-01
-9.84466195e-01 8.34806502e-01 4.30708617e-01 -4.40679938e-01
-5.65233408e-03 -9.71694946e-01 -1.13606632e+00 -1.43876851e-01
-6.82791397e-02 1.85374156e-01 1.02843571e+00 -4.82770115e-01
-1.34348118e+00 7.90899158e-01 1.32812977e-01 -4.98020142e-01
7.63545781e-02 -1.77840918e-01 -7.36508667e-01 2.19398737e-02
9.30441469e-02 -2.67863661e-01 1.37836289e+00 -4.85108405e-01
-5.73862016e-01 -4.35745150e-01 -6.27520502e-01 -4.85854477e-01
-5.05256116e-01 -4.81156297e-02 4.63264555e-01 -5.29690623e-01
1.31623462e-01 -8.16800356e-01 7.70322382e-02 -6.09454401e-02
-2.38397330e-01 -8.45748365e-01 1.51625276e+00 -3.09426755e-01
1.32075953e+00 -2.61454988e+00 -2.32606649e-01 7.15757251e-01
3.01339597e-01 4.82038438e-01 1.50825918e-01 6.19273007e-01
-6.91243827e-01 8.75408947e-03 -5.79622984e-01 5.21312393e-02
2.86942273e-01 1.87253058e-01 -6.97242081e-01 1.11839676e+00
3.33804011e-01 6.41613781e-01 -8.46010566e-01 -2.81151235e-01
4.08923060e-01 2.37371907e-01 -5.04989505e-01 1.36816874e-01
4.68203247e-01 -2.46567681e-01 -1.61205962e-01 5.24895132e-01
8.31033885e-01 3.41192968e-02 -7.13266671e-01 -1.30600482e-01
-2.09997654e-01 -5.67901373e-01 -1.29061234e+00 1.36748099e+00
3.34984846e-02 7.83994198e-01 -3.78910542e-01 -1.51039290e+00
1.17022502e+00 2.06201032e-01 6.27506196e-01 -1.32985175e-01
-1.90802768e-01 4.98103142e-01 -1.54264808e-01 -1.90622285e-01
1.85442433e-01 1.33460850e-01 -2.18679562e-01 7.51152217e-01
3.61794055e-01 1.05665371e-01 -3.45519006e-01 2.50064403e-01
1.50945401e+00 -5.45220733e-01 3.40407759e-01 -8.64023387e-01
7.40678072e-01 -1.12635225e-01 9.50335637e-02 8.52020502e-01
-4.64590639e-01 3.13516736e-01 7.57583022e-01 -6.90373778e-01
-1.04055619e+00 -1.56799662e+00 -4.55321193e-01 1.04875684e+00
-7.40909949e-02 -5.29616356e-01 -4.77216691e-01 -1.04465461e+00
6.17232263e-01 7.81335533e-01 -8.79331350e-01 -6.49019599e-01
-3.86487186e-01 -8.83495092e-01 9.88831162e-01 7.70401418e-01
2.49636471e-01 -7.24085450e-01 -1.48200005e-01 -7.96632394e-02
5.07728875e-01 -5.33074856e-01 -5.93068719e-01 5.53525567e-01
-7.52420902e-01 -1.16349769e+00 -4.19103354e-01 -5.91567576e-01
7.30700135e-01 -3.36832181e-02 7.33703852e-01 -2.39790529e-01
-7.58691847e-01 1.21548975e+00 -3.33746284e-01 -6.44292116e-01
-2.39584967e-01 -2.67816603e-01 7.50683546e-01 2.08079785e-01
1.07041597e+00 -7.80911207e-01 -4.80522662e-01 2.22031295e-01
-1.14227343e+00 -1.09423912e+00 4.50859487e-01 8.50808024e-01
4.53861028e-01 2.97458440e-01 5.73450863e-01 -5.59965074e-01
8.46091270e-01 -5.18576026e-01 -6.53034449e-01 1.77372210e-02
-1.04137897e+00 4.54859167e-01 6.23659670e-01 -5.27497172e-01
-2.35842839e-01 -1.34222493e-01 1.26916647e-01 -5.98327160e-01
-2.83851475e-01 -6.69812858e-02 2.31762659e-02 -4.04732049e-01
1.03533971e+00 6.42319918e-01 4.83958542e-01 -4.63458985e-01
4.77724880e-01 6.92649066e-01 4.29475129e-01 -6.76940441e-01
9.72752929e-01 5.66456318e-01 2.21795380e-01 -1.34810913e+00
-3.02929550e-01 -1.09004724e+00 -1.04936075e+00 4.20103222e-01
8.48860502e-01 -1.95899516e-01 -6.22337759e-01 6.50207102e-01
-7.16498792e-01 -1.54992491e-01 -7.96239018e-01 5.72366416e-01
-8.74697506e-01 6.63749635e-01 -4.10811931e-01 -9.22315478e-01
-2.70803452e-01 -4.95613426e-01 1.00141478e+00 -2.46397659e-01
-8.01817924e-02 -1.50040591e+00 5.70388615e-01 -8.02590191e-01
5.17177284e-01 1.51736706e-01 1.16591203e+00 -1.58614552e+00
1.99113429e-01 -8.95693958e-01 -3.11541796e-01 7.96432853e-01
4.11107183e-01 1.47588268e-01 -7.61685252e-01 -4.70688254e-01
7.68795386e-02 2.12158322e-01 5.30513346e-01 9.41402987e-02
1.24560034e+00 -1.32107690e-01 -3.34606111e-01 7.96773851e-01
1.43336177e+00 8.30972269e-02 4.81542796e-01 4.23858643e-01
8.10464799e-01 1.65690914e-01 2.39988849e-01 2.20872059e-01
-1.70958444e-01 4.06782597e-01 8.31727311e-02 1.93731617e-02
6.45314634e-01 -1.90194778e-03 6.68228209e-01 8.85640264e-01
1.27107427e-01 5.89114428e-01 -1.00589299e+00 5.38835287e-01
-2.01186061e+00 -1.07204449e+00 -1.89006403e-01 2.47712612e+00
3.37777674e-01 -9.00296941e-02 4.36584651e-01 4.34293300e-01
6.04543984e-01 -2.53762722e-01 -4.51981515e-01 -9.87482190e-01
2.99681891e-02 3.87811273e-01 7.12152302e-01 3.65934044e-01
-1.39200258e+00 7.02335238e-01 6.62662029e+00 8.95398080e-01
-9.14124966e-01 -1.15414612e-01 -1.07550509e-01 4.82266933e-01
1.13840818e-01 5.32989800e-02 -7.30462372e-01 6.13309443e-01
1.02376246e+00 -3.51662666e-01 -3.92679013e-02 1.40462232e+00
-4.11728591e-01 1.36223748e-01 -1.47985172e+00 1.41366243e+00
7.94645995e-02 -8.62405121e-01 6.87109903e-02 4.72615510e-01
1.56181470e-01 -1.81279182e-01 3.51225555e-01 6.20878994e-01
1.73274890e-01 -1.20202982e+00 -1.66376442e-01 5.58589876e-01
5.83094478e-01 -9.97600555e-01 8.11020970e-01 1.47334948e-01
-1.24323022e+00 9.90908518e-02 -7.62744069e-01 1.08336225e-01
-3.33794266e-01 6.06106460e-01 -5.70234954e-01 2.70450383e-01
8.04567277e-01 5.64127326e-01 -6.96782231e-01 9.00364816e-01
4.06960189e-01 6.71335101e-01 -5.37455261e-01 2.74751067e-01
3.16899657e-01 -4.72120106e-01 7.61979103e-01 1.53309107e+00
5.34435153e-01 -1.20404221e-01 2.60146081e-01 6.39614224e-01
6.94410682e-01 4.56154764e-01 -1.36120260e+00 -1.12867996e-01
2.50753701e-01 1.03357482e+00 -6.63928092e-01 -2.49354169e-01
-5.41770339e-01 1.42098832e+00 3.17412436e-01 1.68533981e-01
-7.17175305e-01 -1.08512354e+00 1.16817081e+00 7.09096640e-02
4.32426780e-01 -3.59677136e-01 1.29732281e-01 -1.09900403e+00
-5.13647981e-02 -3.49460900e-01 6.98309302e-01 -8.80638659e-02
-1.87858737e+00 1.43642351e-01 2.37080291e-01 -1.21188617e+00
-2.89150447e-01 -1.33613610e+00 -1.04621685e+00 8.72113824e-01
-9.77669835e-01 -7.79449046e-01 1.95782602e-01 1.14451003e+00
-8.87771919e-02 -4.99166697e-01 1.32464361e+00 -6.50903359e-02
-2.63762504e-01 1.06767607e+00 5.83992124e-01 5.23029208e-01
8.42126906e-01 -2.08614850e+00 4.11583662e-01 4.39198911e-01
9.52791721e-02 6.21928692e-01 4.65554059e-01 -3.03039491e-01
-1.55407894e+00 -1.08115482e+00 3.22277427e-01 -1.04465652e+00
9.27050292e-01 -2.38636479e-01 -1.35098076e+00 7.41525173e-01
-3.64335030e-01 6.62150443e-01 1.21582782e+00 3.66711646e-01
-7.60169566e-01 -1.39770135e-01 -1.36074567e+00 2.69847214e-01
5.37271500e-01 -7.38417327e-01 -1.03846931e+00 3.21267456e-01
6.42261505e-01 2.72164971e-01 -1.02882409e+00 4.43488598e-01
2.19761193e-01 -9.39538777e-01 1.04761887e+00 -9.46690023e-01
-6.00961387e-01 -3.37894052e-01 -4.81214374e-01 -1.33679819e+00
-5.94731688e-01 -5.20484090e-01 -5.17742753e-01 9.76483226e-01
1.69011489e-01 -1.34162545e+00 4.20785218e-01 4.67453390e-01
2.22634766e-02 -7.31115520e-01 -1.03175414e+00 -1.53386176e+00
4.51055646e-01 -6.70501590e-01 4.34594095e-01 1.12885642e+00
3.78544956e-01 2.73790932e-03 -3.76224518e-04 4.65417206e-01
7.29976714e-01 -2.09067389e-01 8.28322530e-01 -1.49268675e+00
-1.42871350e-01 -6.29851341e-01 -1.54118168e+00 -4.14025635e-01
2.56171644e-01 -9.60042119e-01 -3.05695266e-01 -7.05466390e-01
-3.11326414e-01 -4.14843440e-01 -7.66736686e-01 3.55963141e-01
5.68115376e-02 5.38702607e-02 -5.26869535e-01 2.02016652e-01
-5.11822641e-01 4.08627808e-01 2.13054776e-01 1.35971531e-01
-4.08737808e-01 -1.47811677e-02 -2.54893929e-01 1.21881986e+00
8.18006575e-01 -2.50752896e-01 -3.82366329e-01 3.49033445e-01
-2.58575436e-02 -7.20318496e-01 4.86837775e-01 -1.20827842e+00
1.60522714e-01 -1.76547579e-02 4.46973532e-01 -6.00726247e-01
1.22441337e-01 -7.70569384e-01 -6.60333872e-01 2.71522939e-01
2.96749175e-02 6.24487817e-01 2.97964483e-01 9.60659325e-01
-1.52620882e-01 -2.38036737e-01 8.00000727e-01 3.25062901e-01
-1.06165564e+00 5.01227319e-01 -4.58437771e-01 7.80956671e-02
1.49750769e+00 -3.31850350e-01 1.01812808e-02 -1.27328143e-01
-6.96571410e-01 -1.49834961e-01 4.23905909e-01 1.25461161e-01
8.18119347e-01 -1.72001863e+00 -4.22106385e-01 6.61530137e-01
6.23318017e-01 -2.60439545e-01 -1.02359563e-01 1.10177040e+00
-5.21859109e-01 3.01567763e-01 3.78560871e-02 -9.14308012e-01
-1.11922348e+00 1.00730109e+00 2.80364007e-01 9.59426537e-02
-1.08889282e+00 7.45271623e-01 2.38793105e-01 -7.51612306e-01
1.00289717e-01 -4.34888154e-01 1.67203039e-01 -1.24704510e-01
8.61584961e-01 5.34791887e-01 2.19941083e-02 -4.36568797e-01
-4.66578692e-01 7.33186305e-01 -2.39975378e-01 -1.86507087e-02
9.11392808e-01 3.05668056e-01 -3.58479619e-01 1.03542769e+00
1.62903404e+00 6.45835996e-02 -6.93981469e-01 -2.45938271e-01
4.05501693e-01 -4.98794407e-01 -2.62274086e-01 -1.24100760e-01
-2.34208375e-01 8.12211096e-01 9.62847054e-01 7.23563731e-01
9.63309407e-01 2.65522182e-01 5.54081738e-01 7.22623229e-01
-9.52751189e-03 -1.25915504e+00 1.26808852e-01 7.20141828e-01
6.91945195e-01 -1.23025072e+00 -3.34007174e-01 -2.21872870e-02
-2.12323427e-01 1.33409154e+00 5.92213392e-01 -8.40580285e-01
1.44914079e+00 1.15756743e-01 2.37570051e-02 -3.89940441e-01
-2.44752273e-01 -1.07898824e-01 5.39099634e-01 8.46492052e-01
1.23401120e-01 1.34139299e-01 -9.79738031e-03 4.45956409e-01
-2.43940741e-01 -7.21377730e-01 1.42223716e-01 1.00393522e+00
-7.97909617e-01 -1.04210258e+00 -4.87259239e-01 7.63582587e-01
-3.30008686e-01 -1.72253940e-02 -3.58181894e-01 9.71417844e-01
-5.11072278e-02 3.21269333e-01 3.03606987e-01 -4.77271885e-01
3.37753415e-01 7.75128067e-01 3.02658141e-01 -4.34865654e-01
2.43356228e-01 -5.55673182e-01 -9.70524371e-01 -7.52554655e-01
9.41084996e-02 -5.04163265e-01 -1.20245790e+00 -5.24564087e-01
-3.45368743e-01 5.85258901e-01 6.98921561e-01 9.27125812e-01
4.18805987e-01 9.46631748e-03 6.45557821e-01 -4.50033247e-01
-1.09172451e+00 -7.34184027e-01 -1.28646910e+00 4.95498031e-01
6.43582761e-01 -7.21227646e-01 -9.08351243e-01 -3.10424089e-01] | [7.6325459480285645, 2.5241799354553223] |
8545f6c9-f2b9-40f1-b95d-0d9f2e959adb | don-t-treat-the-symptom-find-the-cause | 2306.12850 | null | https://arxiv.org/abs/2306.12850v1 | https://arxiv.org/pdf/2306.12850v1.pdf | Don't Treat the Symptom, Find the Cause! Efficient Artificial-Intelligence Methods for (Interactive) Debugging | In the modern world, we are permanently using, leveraging, interacting with, and relying upon systems of ever higher sophistication, ranging from our cars, recommender systems in e-commerce, and networks when we go online, to integrated circuits when using our PCs and smartphones, the power grid to ensure our energy supply, security-critical software when accessing our bank accounts, and spreadsheets for financial planning and decision making. The complexity of these systems coupled with our high dependency on them implies both a non-negligible likelihood of system failures, and a high potential that such failures have significant negative effects on our everyday life. For that reason, it is a vital requirement to keep the harm of emerging failures to a minimum, which means minimizing the system downtime as well as the cost of system repair. This is where model-based diagnosis comes into play. Model-based diagnosis is a principled, domain-independent approach that can be generally applied to troubleshoot systems of a wide variety of types, including all the ones mentioned above, and many more. It exploits and orchestrates i.a. techniques for knowledge representation, automated reasoning, heuristic problem solving, intelligent search, optimization, stochastics, statistics, decision making under uncertainty, machine learning, as well as calculus, combinatorics and set theory to detect, localize, and fix faults in abnormally behaving systems. In this thesis, we will give an introduction to the topic of model-based diagnosis, point out the major challenges in the field, and discuss a selection of approaches from our research addressing these issues. | ['Patrick Rodler'] | 2023-06-22 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty', 'decision-making'] | ['medical', 'reasoning', 'reasoning'] | [-2.16605067e-01 -3.48674394e-02 -2.92959273e-01 2.88215458e-01
-2.31237207e-02 -6.98308647e-01 2.24816516e-01 3.74945611e-01
1.65322334e-01 5.98743439e-01 -4.79127795e-01 -7.37628639e-01
-8.20392966e-01 -8.96148205e-01 -1.41352594e-01 -5.50037265e-01
-1.91756874e-01 5.84138572e-01 3.80916625e-01 -1.48928180e-01
5.06929874e-01 7.14788735e-01 -1.55901992e+00 -3.73065650e-01
5.95434546e-01 1.28905773e+00 -6.26571551e-02 3.79617840e-01
2.14790136e-01 6.55976951e-01 -5.12320995e-01 -1.77385822e-01
-1.88455269e-01 -1.25055432e-01 -4.82043892e-01 9.12497863e-02
-9.25237894e-01 -1.35490879e-01 -1.16015919e-01 9.52009141e-01
1.17986903e-01 1.27774933e-02 4.58469599e-01 -1.44725990e+00
-2.26045430e-01 3.25289577e-01 -4.12236989e-01 2.59142786e-01
2.44261980e-01 2.32607380e-01 5.72200179e-01 -4.06390637e-01
7.27686659e-02 6.93535388e-01 4.51066643e-01 2.39890423e-02
-8.80349457e-01 -3.37225944e-01 2.35501811e-01 3.77233177e-01
-1.35019433e+00 -3.84194791e-01 3.76993448e-01 -6.12422407e-01
1.26356626e+00 4.19559449e-01 5.81937075e-01 5.74757278e-01
6.00858867e-01 1.95341691e-01 5.15619099e-01 -6.07579172e-01
5.89085102e-01 3.94160509e-01 8.74756500e-02 5.63476086e-01
4.79326099e-01 9.19548273e-02 -1.25853777e-01 -3.76573443e-01
3.68560016e-01 4.09335613e-01 -2.00128123e-01 -7.63937160e-02
-7.76286840e-01 4.43363965e-01 -2.57097185e-01 5.86660624e-01
-4.66471523e-01 8.28959048e-02 2.00842544e-01 1.72706217e-01
4.22803424e-02 5.49264371e-01 -6.17385924e-01 -2.83498496e-01
-6.03850245e-01 -1.44215360e-01 1.03727293e+00 7.06181049e-01
3.79289597e-01 8.71934146e-02 2.86717325e-01 3.34084332e-01
3.30800205e-01 4.68956530e-01 4.69484717e-01 -6.48953497e-01
5.15171997e-02 8.04193377e-01 2.00137481e-01 -8.02493513e-01
-4.53509778e-01 -5.63591242e-01 -7.93131173e-01 6.47065118e-02
-1.27599970e-01 -9.19633135e-02 -3.41876209e-01 1.32945156e+00
1.80584759e-01 2.72273600e-01 -3.29308957e-01 5.00687420e-01
-4.20589328e-01 4.25106078e-01 -3.65479797e-01 -6.21245623e-01
1.04253256e+00 -2.56317586e-01 -5.78857899e-01 3.77676934e-02
2.24413946e-01 -6.06137455e-01 5.86456597e-01 9.90532160e-01
-1.02039051e+00 1.80074051e-02 -1.07129610e+00 7.57976770e-01
-3.39891106e-01 -2.54594028e-01 5.83602250e-01 7.58260906e-01
-8.89178455e-01 7.60109544e-01 -9.55513000e-01 -3.69500846e-01
1.50872711e-02 3.85448545e-01 -4.76652384e-02 -2.11394522e-02
-9.54209864e-01 1.28840005e+00 -2.39181463e-02 1.63203537e-01
-6.15313828e-01 -4.62499410e-01 -4.48842674e-01 3.45342785e-01
7.59331703e-01 -4.75520164e-01 1.15772176e+00 -1.99315771e-01
-1.28139949e+00 8.29989612e-02 1.15404338e-01 -4.14396226e-01
1.84218645e-01 5.88050298e-02 -7.94006407e-01 -2.19226107e-01
-2.72601008e-01 -6.70443296e-01 4.83216286e-01 -7.22641885e-01
-7.42163479e-01 -4.20772821e-01 -6.89824298e-02 -1.94220424e-01
-2.30776057e-01 4.52343263e-02 8.46371800e-02 5.51465563e-02
1.22268558e-01 -7.40335643e-01 -5.14677882e-01 -5.10987103e-01
-4.59840387e-01 -9.35516730e-02 9.87992704e-01 -4.84935045e-01
1.47257388e+00 -1.87864661e+00 1.13985978e-01 7.38875389e-01
-9.59681273e-02 1.42043546e-01 6.18554354e-01 1.03077102e+00
-3.52270249e-03 2.47057751e-01 -1.12116605e-01 2.02177092e-01
2.50970423e-01 4.26524431e-02 -3.42892289e-01 4.41291124e-01
2.48203665e-01 4.36313272e-01 -6.96587682e-01 8.19683596e-02
5.72707593e-01 1.75255999e-01 -1.48216203e-01 9.88615304e-02
-2.79094070e-01 1.63126051e-01 -6.30982876e-01 9.32036281e-01
-6.35192394e-02 -4.20760870e-01 4.43292975e-01 2.31944829e-01
-1.30203545e-01 5.15069515e-02 -1.35230196e+00 7.79137790e-01
-8.04948568e-01 3.36971790e-01 8.57172906e-02 -1.19056666e+00
6.34290159e-01 5.37945509e-01 6.88443959e-01 -7.61995375e-01
3.04865211e-01 2.50043064e-01 -6.71311021e-02 -6.55264139e-01
2.53165215e-01 1.82018504e-01 -1.90419361e-01 8.38680267e-01
-2.89492309e-01 -2.57862378e-02 1.57780051e-01 1.45305887e-01
1.60141790e+00 -3.49499345e-01 6.56263709e-01 -1.76946923e-01
4.71806824e-01 -2.51518548e-01 4.97707784e-01 1.74737319e-01
5.96120358e-02 1.79040805e-02 6.40348554e-01 -1.85402200e-01
-6.87386751e-01 -8.33766639e-01 -1.03868946e-01 3.67670923e-01
1.53703645e-01 -9.30020437e-02 -5.14099896e-01 -3.21203500e-01
1.24987982e-01 9.10933554e-01 -1.14592284e-01 -5.72744906e-01
-1.52825519e-01 -7.17399836e-01 -9.44849197e-03 2.89682478e-01
1.19948454e-01 -6.04149342e-01 -6.16260886e-01 4.66354311e-01
4.08777893e-01 -9.20346260e-01 4.96114716e-02 3.91248137e-01
-8.07657182e-01 -1.29198182e+00 -4.06800099e-02 -1.38769999e-01
6.51640415e-01 -3.19484100e-02 9.98053968e-01 4.58001703e-01
-5.08004665e-01 5.77860475e-01 -2.58213282e-01 -3.45310181e-01
-3.68625849e-01 -4.40958917e-01 4.30169195e-01 -1.42984148e-02
8.30813944e-02 -7.69796908e-01 -3.22948158e-01 4.89162058e-01
-8.15874040e-01 -5.54639637e-01 4.81085211e-01 5.49813032e-01
5.20010650e-01 1.14612353e+00 9.12921846e-01 -7.47839928e-01
6.40064120e-01 -9.15204167e-01 -8.78956974e-01 5.35940707e-01
-1.11262202e+00 -1.73552111e-01 6.40008688e-01 -2.77384073e-01
-4.89071399e-01 -3.84697542e-02 2.31196299e-01 -3.02125871e-01
6.03490649e-03 5.40693820e-01 -4.98428822e-01 -1.93152763e-02
3.74362290e-01 4.53758538e-02 -8.03481787e-02 -1.63420483e-01
2.21719146e-02 7.73685634e-01 4.35474992e-01 -3.06050986e-01
6.86184585e-01 2.93282837e-01 3.43817264e-01 -6.11699939e-01
-3.24007452e-01 -4.19911921e-01 -1.57134384e-01 -3.86725992e-01
3.11382532e-01 -3.47958833e-01 -1.14517915e+00 3.31889123e-01
-7.35925972e-01 -8.42210203e-02 -2.04390287e-01 2.18163922e-01
-1.49699926e-01 4.86593023e-02 -3.70761067e-01 -1.34278488e+00
-1.14353731e-01 -9.80846226e-01 5.07149518e-01 4.70567286e-01
-3.21529895e-01 -1.07496977e+00 -1.13160051e-01 1.14547528e-01
6.76004350e-01 4.09067750e-01 1.14043641e+00 -5.87832153e-01
-8.07640910e-01 -7.23680913e-01 1.74254373e-01 3.91429037e-01
3.31542671e-01 3.11013877e-01 -6.04160488e-01 -1.97216928e-01
3.09352756e-01 2.04178065e-01 7.43452460e-02 1.69046566e-01
1.09806335e+00 -3.93678904e-01 -5.45525312e-01 -1.00003846e-01
1.37560678e+00 5.17404020e-01 5.71377754e-01 -5.73298782e-02
1.83798015e-01 7.17897594e-01 8.27572823e-01 8.50794673e-01
3.99748862e-01 5.61219096e-01 7.78673708e-01 2.77651399e-01
5.71348488e-01 1.12427406e-01 9.16159451e-02 5.82483828e-01
1.63360879e-01 -4.21592414e-01 -1.04238212e+00 4.99832064e-01
-1.86094511e+00 -7.84610808e-01 8.46192054e-03 2.70079684e+00
3.87741566e-01 3.88905585e-01 6.71002688e-03 6.10514581e-01
7.56049395e-01 -8.74186635e-01 -7.04536915e-01 -3.41674358e-01
5.50490737e-01 7.34390244e-02 6.27824426e-01 1.85778931e-01
-6.02524936e-01 1.70846716e-01 5.98146200e+00 5.89187264e-01
-1.10130548e+00 9.92056951e-02 7.80850291e-01 -6.42730072e-02
-1.91872835e-01 1.48487315e-01 -4.84379739e-01 8.16479921e-01
1.46264780e+00 -3.03305030e-01 7.30427027e-01 8.24797511e-01
4.33459342e-01 -6.33006930e-01 -1.06718731e+00 7.41825044e-01
-3.15548033e-01 -9.29866970e-01 -9.34441149e-01 3.25934470e-01
5.95375538e-01 -2.02823222e-01 -6.77280426e-02 -5.41637689e-02
2.16283605e-01 -9.31278884e-01 7.58362055e-01 7.84656584e-01
3.37396473e-01 -1.02631009e+00 8.63261163e-01 5.63120127e-01
-7.88386226e-01 -4.83017594e-01 2.25084126e-01 -3.02163810e-01
4.87337977e-01 1.28861117e+00 -5.54669440e-01 5.76956689e-01
5.64381540e-01 1.43419728e-01 -5.26597314e-02 1.16033971e+00
-6.48303851e-02 4.49913859e-01 -5.21873057e-01 -1.80600002e-01
-4.59642470e-01 4.53791656e-02 2.58932143e-01 3.94423395e-01
4.69393462e-01 7.57410005e-02 3.27606872e-02 5.86203158e-01
3.63275021e-01 -4.91700172e-01 -4.01444584e-01 -1.38123140e-01
8.71311963e-01 1.24004960e+00 -1.09800601e+00 -1.32669210e-01
-2.19844177e-01 3.88262868e-01 -3.47627312e-01 1.27755821e-01
-7.88499057e-01 -4.56087828e-01 5.15299737e-01 4.36611176e-01
-1.75934911e-01 -3.66290152e-01 -4.08059061e-01 -7.64016569e-01
1.76676050e-01 -8.37031662e-01 2.17605352e-01 -3.74373585e-01
-1.16743720e+00 2.40119025e-01 2.02995865e-03 -1.15022469e+00
-7.35779226e-01 -4.14821714e-01 -6.87956214e-01 6.53790534e-01
-1.21454298e+00 -3.46662641e-01 8.64499584e-02 4.24454361e-01
-5.59054408e-03 -1.69401407e-01 6.06945574e-01 5.11089802e-01
-8.98178995e-01 1.25255454e-02 1.43461630e-01 -5.98518014e-01
1.67764097e-01 -8.94751966e-01 2.78777957e-01 9.31399345e-01
3.77260931e-02 3.88883859e-01 6.11702383e-01 -4.95490044e-01
-1.84101164e+00 -5.81054330e-01 7.41876185e-01 -3.01115781e-01
9.57836449e-01 -2.54854381e-01 -6.68904245e-01 2.21849039e-01
-2.34358639e-01 -1.06189474e-01 3.80953610e-01 1.26973197e-01
4.40536924e-02 -3.16004366e-01 -1.42630816e+00 2.48803839e-01
4.32336688e-01 -4.75153714e-01 2.52155699e-02 6.66015089e-01
4.28508162e-01 -1.95485830e-01 -8.10557008e-01 2.26360112e-01
4.19921696e-01 -1.12691891e+00 6.63963556e-01 -7.44377747e-02
-1.54515609e-01 -3.42323899e-01 2.61648651e-02 -1.05066454e+00
-1.45878047e-01 -7.42524445e-01 -5.51844537e-01 1.36552572e+00
4.60987717e-01 -1.07792890e+00 4.92789149e-01 9.57808256e-01
-1.80791363e-01 -9.35026288e-01 -1.06178844e+00 -9.13989127e-01
-4.57953811e-01 -5.44028521e-01 8.54281723e-01 8.98125827e-01
4.40099746e-01 -8.27852935e-02 1.13355987e-01 4.41805452e-01
4.81542975e-01 2.29645148e-02 1.56935573e-01 -1.25023973e+00
-8.17149699e-01 -4.34228122e-01 -5.24267137e-01 -2.40399882e-01
-3.12701106e-01 -1.40458852e-01 -3.06464471e-02 -1.36590254e+00
-5.82269356e-02 -7.16145098e-01 -5.09639740e-01 4.01407927e-01
3.68594766e-01 -1.43892959e-01 -1.89418703e-01 2.52649963e-01
-5.45958698e-01 5.85720467e-04 4.84874666e-01 1.23282768e-01
-1.03565440e-01 4.82082695e-01 -6.84110999e-01 6.73727691e-01
8.34254086e-01 -3.64534855e-01 -4.83753741e-01 1.45940930e-02
6.82323992e-01 4.08861697e-01 3.31037968e-01 -1.07043433e+00
5.24316192e-01 -4.53042448e-01 -8.22673440e-02 -2.02364579e-01
1.63364261e-01 -1.23893046e+00 5.08493900e-01 7.46665597e-01
3.35836112e-01 1.30043373e-01 -1.61748856e-01 5.56630731e-01
-1.32988676e-01 -4.09778118e-01 4.73712891e-01 1.52231336e-01
-5.22518933e-01 1.44275486e-01 -6.13003314e-01 -4.15214986e-01
1.44319081e+00 2.91422717e-02 -4.69996929e-01 -5.15228093e-01
-5.51151037e-01 5.05746841e-01 3.63432378e-01 4.41666126e-01
2.70207137e-01 -6.12445235e-01 2.85706557e-02 1.55246630e-01
-2.15020582e-01 -1.48029402e-01 2.66703486e-01 1.00652647e+00
-7.15486035e-02 4.28003401e-01 2.83624202e-01 -3.04050535e-01
-8.40648949e-01 5.58690250e-01 2.45485306e-01 -3.93937111e-01
-4.26647067e-02 4.24546182e-01 -4.62084532e-01 7.67607465e-02
1.66412070e-01 -3.58990908e-01 -9.16452613e-04 -3.02323550e-01
5.40713310e-01 7.25972354e-01 4.51207042e-01 1.44865420e-02
-6.35757685e-01 2.55291402e-01 2.65947253e-01 -2.45467722e-02
1.18311977e+00 -2.71876186e-01 -3.56817931e-01 4.42875922e-01
3.45556647e-01 -9.22367945e-02 -7.13124812e-01 1.61045223e-01
4.02209997e-01 -1.20894380e-01 2.05870003e-01 -1.02320671e+00
-9.92961526e-01 5.53592980e-01 3.47730398e-01 1.10300660e+00
1.43754399e+00 -2.49794442e-02 5.74374378e-01 3.92928749e-01
1.05347705e+00 -1.05766463e+00 -7.89043307e-02 2.89801806e-01
3.91036808e-01 -7.86971390e-01 2.18436882e-01 -1.00601152e-01
-3.46788228e-01 9.93020177e-01 3.05705786e-01 -2.27433555e-02
1.07755709e+00 6.96488857e-01 -6.02230728e-01 -7.22588645e-03
-1.04027820e+00 -2.35815644e-02 -2.89183587e-01 5.51284075e-01
-7.46221244e-02 9.92430523e-02 -5.14917336e-02 9.37029064e-01
1.88534632e-01 -1.59789640e-02 7.07667470e-01 9.50231194e-01
-6.11438453e-01 -1.44418705e+00 -4.88451391e-01 6.45723879e-01
-5.20146966e-01 1.84579581e-01 -1.86892822e-01 5.67116857e-01
2.10533291e-01 1.30908084e+00 -2.86933687e-02 -6.93310857e-01
3.93172175e-01 -9.42359492e-02 1.00861527e-01 -5.88045597e-01
-3.73421371e-01 -2.11395591e-01 1.83126867e-01 -6.45610631e-01
6.89853802e-02 -8.22218359e-01 -1.20093286e+00 -5.65652370e-01
-8.49112749e-01 1.63844094e-01 1.19965148e+00 1.18218136e+00
5.07687271e-01 6.45529270e-01 1.03029859e+00 -4.72398251e-01
-6.52399540e-01 -7.27708459e-01 -9.58898783e-01 -5.26052237e-01
-1.97596513e-02 -6.85467720e-01 -4.22790051e-01 -3.83784324e-01] | [6.189904689788818, 2.624246597290039] |
c4038707-e35c-49f1-85f7-408a65d1208b | simple-baselines-for-human-pose-estimation | 1804.06208 | null | http://arxiv.org/abs/1804.06208v2 | http://arxiv.org/pdf/1804.06208v2.pdf | Simple Baselines for Human Pose Estimation and Tracking | There has been significant progress on pose estimation and increasing
interests on pose tracking in recent years. At the same time, the overall
algorithm and system complexity increases as well, making the algorithm
analysis and comparison more difficult. This work provides simple and effective
baseline methods. They are helpful for inspiring and evaluating new ideas for
the field. State-of-the-art results are achieved on challenging benchmarks. The
code will be available at https://github.com/leoxiaobin/pose.pytorch. | ['Haiping Wu', 'Yichen Wei', 'Bin Xiao'] | 2018-04-17 | simple-baselines-for-human-pose-estimation-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Bin_Xiao_Simple_Baselines_for_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Bin_Xiao_Simple_Baselines_for_ECCV_2018_paper.pdf | eccv-2018-9 | ['2d-human-pose-estimation'] | ['computer-vision'] | [-1.71224907e-01 -3.27871263e-01 -4.30186003e-01 -4.29400623e-01
-9.29381430e-01 -7.23309100e-01 2.99822837e-01 -1.92590252e-01
-3.86390865e-01 5.64167857e-01 1.91102475e-01 1.15926258e-01
3.51409227e-01 -3.85446608e-01 -5.45756757e-01 -6.08611226e-01
-3.45593780e-01 6.27978861e-01 3.97420526e-01 -2.82743335e-01
1.94641426e-01 4.87715393e-01 -1.29196823e+00 -5.18941917e-02
3.22426349e-01 8.14186931e-01 1.64235651e-01 8.21192145e-01
3.92095953e-01 2.53844947e-01 -5.06563723e-01 -6.88636303e-01
4.08319712e-01 -2.95606464e-01 -8.34250748e-01 -9.25162286e-02
1.14959729e+00 -2.10915491e-01 -4.63136733e-01 1.11930597e+00
1.00324118e+00 4.80116084e-02 9.01966020e-02 -1.09192705e+00
-7.00944290e-02 2.76434839e-01 -6.19034588e-01 3.25810105e-01
6.03528976e-01 8.74385238e-02 1.07118571e+00 -9.48561370e-01
7.40279317e-01 1.30746603e+00 9.14197445e-01 5.42301476e-01
-9.30608511e-01 -6.96901262e-01 2.63747007e-01 2.00066566e-01
-1.41396439e+00 -4.96817380e-01 5.46143115e-01 -1.48222789e-01
6.19308352e-01 5.10314643e-01 7.69721329e-01 1.13871574e+00
4.20401931e-01 1.08316112e+00 7.79253900e-01 -1.28242493e-01
-1.93111017e-01 -1.96120054e-01 -3.15549076e-02 7.35496402e-01
2.75490642e-01 1.30723342e-01 -6.26143992e-01 5.63164800e-02
8.00632894e-01 -1.15883596e-01 -2.16917425e-01 -7.51002967e-01
-1.32060647e+00 5.93656957e-01 5.94233930e-01 1.83395088e-01
-1.07568186e-02 5.96787632e-01 5.13665438e-01 2.50486314e-01
4.40183192e-01 4.11520064e-01 -4.41149265e-01 -5.30821860e-01
-9.56125796e-01 7.25774765e-01 8.49469364e-01 1.02751172e+00
2.05118924e-01 -2.54268140e-01 2.52839267e-01 6.04865134e-01
3.58201832e-01 7.00023890e-01 3.65060493e-02 -1.15911078e+00
4.31660324e-01 1.09909521e-02 1.26909360e-01 -1.12959683e+00
-5.73634863e-01 -7.40395725e-01 -3.86119634e-01 4.63204943e-02
6.16376817e-01 1.08620161e-02 -8.21306884e-01 1.51734757e+00
4.78530735e-01 3.20709683e-02 -6.31730914e-01 1.10731757e+00
9.36552405e-01 4.77539778e-01 -8.17639753e-02 1.66780472e-01
1.38714385e+00 -1.32799613e+00 -7.52048790e-01 -5.12904167e-01
4.94868815e-01 -1.39261460e+00 8.31673086e-01 5.40030599e-01
-1.30555296e+00 -4.60813582e-01 -9.11806762e-01 -3.51041108e-02
-8.74464363e-02 2.67615169e-01 1.00726044e+00 7.06826866e-01
-8.84737253e-01 6.84583187e-01 -1.45359254e+00 -6.96560621e-01
5.25093496e-01 5.68157434e-01 -1.03244729e-01 -1.80437118e-01
-8.52764487e-01 8.69438469e-01 2.04841062e-01 3.87028813e-01
-5.06521344e-01 -5.58433890e-01 -7.38536417e-01 -5.86267292e-01
5.84333718e-01 -6.43685222e-01 1.81518781e+00 -3.62372011e-01
-1.37124705e+00 7.61108935e-01 -1.06418006e-01 -3.29358935e-01
9.25448954e-01 -7.60538101e-01 -3.84136438e-01 -1.98228911e-01
-1.89638492e-02 9.06886041e-01 3.87229860e-01 -9.10137653e-01
-4.71995145e-01 -5.44711947e-01 8.31725225e-02 5.08323789e-01
3.61468308e-02 2.68723935e-01 -1.06153631e+00 -8.28719079e-01
3.88764560e-01 -1.41185760e+00 -2.95400977e-01 2.33846530e-01
-3.92995596e-01 -6.20983988e-02 7.63675272e-01 -5.61882436e-01
1.29167426e+00 -1.86776531e+00 1.73553199e-01 -1.43976912e-01
1.31352350e-01 2.64954895e-01 -6.61544353e-02 5.45064032e-01
9.06566307e-02 -2.41383284e-01 1.88456044e-01 -4.55453277e-01
-1.72418565e-01 -6.80641830e-02 -2.46315256e-01 8.43830943e-01
-2.61955827e-01 1.02533889e+00 -9.17496979e-01 -4.00967062e-01
4.12330091e-01 5.36502421e-01 -5.86566806e-01 -1.54517293e-01
-2.29435274e-03 8.20072472e-01 -3.64163935e-01 8.96140039e-01
6.93899095e-01 -3.16224605e-01 9.27614048e-02 -4.20788705e-01
-2.02054858e-01 3.73861670e-01 -1.43911970e+00 1.94363201e+00
-1.81649029e-01 7.53476024e-01 5.23100235e-02 -5.66035926e-01
4.98398066e-01 9.98840318e-04 5.14672816e-01 -4.13583636e-01
4.68429238e-01 2.73354650e-01 2.40923807e-01 -1.33535162e-01
5.47008753e-01 9.60090607e-02 -5.95795065e-02 1.06203258e-01
-7.97912106e-02 -2.40634605e-01 4.16707635e-01 1.29876614e-01
7.68791318e-01 5.25193512e-01 4.26544458e-01 -2.40867719e-01
3.94514591e-01 1.27975151e-01 4.76037562e-01 3.75157416e-01
-5.88758945e-01 7.43555725e-01 -3.98040339e-02 -6.39148712e-01
-8.96897614e-01 -1.32861042e+00 -3.89083147e-01 1.07275200e+00
4.55828309e-01 -7.44005680e-01 -5.20391524e-01 -3.13893646e-01
1.38687983e-01 -5.89136267e-03 -5.46083868e-01 3.71987790e-01
-9.81405377e-01 -5.55259943e-01 4.56581533e-01 9.48867738e-01
3.72103721e-01 -7.55402803e-01 -3.70291322e-01 1.19693108e-01
-3.90871376e-01 -1.10810864e+00 -6.66933358e-01 -1.04252823e-01
-1.21460438e+00 -1.01654041e+00 -9.21593845e-01 -7.89835751e-01
6.64208710e-01 3.16977352e-01 1.23826623e+00 -6.83083606e-04
-6.15395308e-01 3.36392790e-01 -1.69792101e-01 -4.57268685e-01
2.85246164e-01 3.56118619e-01 2.66347945e-01 -6.29685581e-01
1.42635658e-01 -2.96933323e-01 -8.43119442e-01 7.06772089e-01
-4.10769582e-01 -1.92043379e-01 4.92039531e-01 6.26871228e-01
8.04269016e-01 -5.77650785e-01 -2.72926167e-02 -8.55403364e-01
2.25617185e-01 1.70125067e-01 -8.77887964e-01 -1.12994671e-01
-5.60523868e-01 -8.81224200e-02 1.48595780e-01 -3.52301329e-01
-7.59277642e-01 2.72686332e-01 -4.02168423e-01 -1.10562943e-01
2.29290742e-02 1.37035817e-01 1.54326543e-01 -5.21642029e-01
6.38761878e-01 -1.20795727e-01 -5.92596196e-02 -6.54528737e-01
3.46455336e-01 3.30567420e-01 5.18191397e-01 -4.07401979e-01
9.16741490e-01 6.69624031e-01 1.57010704e-02 -8.89920294e-01
-9.91137624e-01 -8.25182438e-01 -5.07319927e-01 -2.78231025e-01
5.05496740e-01 -9.22490835e-01 -7.68076837e-01 5.53411722e-01
-9.41160977e-01 -2.12205917e-01 9.49369743e-02 6.38507843e-01
-5.20843208e-01 3.82109016e-01 -7.63289690e-01 -4.45445865e-01
-2.11048305e-01 -1.34468472e+00 1.24362051e+00 3.42626721e-01
-5.50761342e-01 -1.14055169e+00 4.68624085e-01 6.72685146e-01
3.61566186e-01 1.43665731e-01 -3.93651247e-01 -2.42828384e-01
-9.46932673e-01 -4.96473253e-01 2.28038933e-02 -3.39864343e-02
-1.46694794e-01 -1.29107630e-03 -6.99426293e-01 -7.60283053e-01
-4.25191641e-01 -4.20107454e-01 7.75504351e-01 6.31056607e-01
1.10827124e+00 4.81364504e-02 -7.33111620e-01 8.08164239e-01
1.24468017e+00 -8.98281187e-02 5.67729890e-01 2.73425609e-01
6.73081577e-01 2.80592084e-01 1.02726030e+00 7.50091523e-02
3.48334730e-01 1.20551610e+00 2.40703657e-01 -2.71198303e-02
-4.14913028e-01 -7.82118738e-02 1.49650604e-01 8.87656033e-01
-5.03749549e-01 2.42546550e-03 -1.25899065e+00 3.09312820e-01
-1.76650465e+00 -9.80009079e-01 -3.71085137e-01 2.09748983e+00
6.56791806e-01 1.72800660e-01 2.90148854e-01 -1.18231960e-01
6.03255689e-01 4.56252128e-01 -3.70252669e-01 -4.58238646e-02
1.29216328e-01 1.73585191e-01 7.61905789e-01 6.34042263e-01
-1.49426019e+00 1.23544037e+00 6.76260996e+00 8.16749632e-01
-1.23027480e+00 2.78543849e-02 5.15918255e-01 -4.18434083e-01
3.25963467e-01 -1.04144141e-01 -1.28080750e+00 2.38957629e-01
5.52262187e-01 -2.63644755e-01 3.24418962e-01 9.61315036e-01
6.08090013e-02 -3.03251326e-01 -1.01068854e+00 1.25962090e+00
1.26603350e-01 -1.21577609e+00 -5.36777020e-01 1.34305850e-01
6.34381175e-01 4.64314431e-01 2.93443292e-01 1.16581723e-01
-1.47109507e-02 -6.36759639e-01 6.40313625e-01 7.02225743e-03
7.01258183e-01 -6.73219204e-01 9.31979299e-01 6.70692474e-02
-1.63800001e+00 2.99859494e-01 -2.90454119e-01 -3.21289003e-01
3.50875437e-01 4.10567284e-01 -6.42268717e-01 4.56928253e-01
8.59203041e-01 9.03594613e-01 -7.20007360e-01 1.55214858e+00
-3.98714960e-01 3.82878572e-01 -5.21707058e-01 -1.54951036e-01
3.81211340e-02 7.26906024e-03 6.43734515e-01 1.21044445e+00
4.88999160e-03 -1.24600679e-01 5.27477503e-01 1.52807534e-01
-1.13346592e-01 7.20397159e-02 -4.22037125e-01 -8.86483714e-02
4.64541435e-01 1.38223290e+00 -1.17966127e+00 -2.96404194e-02
-4.19712842e-01 9.24922228e-01 1.31230101e-01 2.80250423e-02
-1.22327340e+00 -1.70294419e-01 8.79130483e-01 1.95114926e-01
2.25584030e-01 -5.76057971e-01 -3.85697752e-01 -1.10158551e+00
2.10116461e-01 -7.58293986e-01 3.47947001e-01 -2.32327655e-01
-1.09503114e+00 3.80222172e-01 -7.51217967e-03 -1.62876439e+00
-8.89548957e-02 -7.62915134e-01 -2.83841759e-01 3.38913590e-01
-1.05884838e+00 -9.28797126e-01 -4.29942310e-01 1.97173148e-01
6.76169813e-01 3.16020459e-01 8.22299957e-01 6.45954013e-01
-5.71991324e-01 8.34290802e-01 -4.83222343e-02 4.12807673e-01
1.10303557e+00 -9.20874715e-01 9.35056329e-01 7.66426086e-01
2.93677658e-01 7.61422098e-01 9.58278537e-01 -6.28615439e-01
-1.70441747e+00 -8.16018820e-01 7.63280630e-01 -1.02448988e+00
9.51758504e-01 -6.36989594e-01 -2.87388593e-01 8.59785199e-01
9.73763987e-02 4.79590409e-02 5.21602035e-01 5.42041302e-01
-1.07539035e-01 1.49084022e-02 -7.49227524e-01 7.54168928e-01
1.32622862e+00 -5.45190759e-02 -5.05840003e-01 5.80435693e-01
3.05058867e-01 -1.30236161e+00 -7.48341858e-01 5.54161549e-01
8.51545811e-01 -7.52278924e-01 1.26487756e+00 -2.93922842e-01
1.74322486e-01 -4.15974647e-01 -2.05241609e-02 -8.88937414e-01
-3.98386538e-01 -5.84321320e-01 -4.28454578e-01 6.06041789e-01
4.65312004e-01 -4.66140896e-01 1.35585570e+00 2.94407606e-01
1.41679135e-03 -1.19568384e+00 -7.87149489e-01 -1.28783488e+00
-2.17438117e-01 -4.11396593e-01 -1.03081070e-01 6.57228887e-01
-8.78090858e-02 5.05276203e-01 -4.25288826e-01 3.54338847e-02
7.73692250e-01 2.02504769e-01 1.03436232e+00 -1.06851518e+00
-3.08558702e-01 -4.92608130e-01 -8.07675242e-01 -1.49939489e+00
-3.70459229e-01 -7.27336645e-01 5.81254214e-02 -1.42832267e+00
1.09428562e-01 -2.28127345e-01 -4.40648608e-02 2.80479252e-01
-1.63429826e-01 8.95055354e-01 5.65460503e-01 3.27444524e-01
-9.78440046e-01 3.86371762e-01 1.49175179e+00 1.10091753e-01
-7.55651295e-02 3.76398087e-01 -4.90661502e-01 8.37602556e-01
1.36974788e+00 -4.30430740e-01 -1.16979353e-01 -4.35420990e-01
6.40922561e-02 -3.14736754e-01 3.18453223e-01 -1.43745434e+00
2.66525269e-01 9.11635533e-02 4.74745780e-01 -9.91610944e-01
7.74192452e-01 -6.82840407e-01 7.68958181e-02 9.20340478e-01
1.61422212e-02 4.45914030e-01 3.90880793e-01 4.01817083e-01
-2.42180616e-01 2.20176503e-01 8.56389761e-01 -8.26770365e-02
-9.07221317e-01 5.51443815e-01 3.22920233e-01 3.27298909e-01
9.49024320e-01 -3.19405884e-01 -2.58576542e-01 -5.42889416e-01
-7.89077342e-01 3.89998496e-01 6.22178555e-01 6.26070499e-01
3.58212858e-01 -1.49733710e+00 -6.27720356e-01 -1.26652226e-01
1.21407323e-01 -2.26505443e-01 1.03017576e-01 1.32645214e+00
-9.10214245e-01 7.15820253e-01 -2.22294435e-01 -7.95098782e-01
-1.72704494e+00 1.63290277e-01 2.52663016e-01 -1.99579120e-01
-5.71965873e-01 1.42942786e+00 -1.60537772e-02 -4.15904284e-01
5.61473429e-01 -1.24655455e-01 1.85779437e-01 -8.86346772e-02
6.68427765e-01 7.20196784e-01 -1.00756641e-02 -6.28395081e-01
-8.03869188e-01 8.73009741e-01 -2.06672281e-01 -5.60362674e-02
1.22644520e+00 1.04307160e-02 2.83912104e-02 3.62707525e-01
1.05832303e+00 2.95003563e-01 -1.09164119e+00 3.04047274e-03
-9.16875899e-02 -7.44400740e-01 -5.11253886e-02 -6.84493482e-01
-1.05762053e+00 7.03412354e-01 8.18774045e-01 -7.80128911e-02
9.22049046e-01 2.03975558e-01 8.82798672e-01 3.21841002e-01
8.73358965e-01 -1.11844778e+00 7.96405450e-02 7.60502398e-01
8.63231659e-01 -1.31854343e+00 6.24804020e-01 -9.46917415e-01
-3.08968991e-01 9.41984236e-01 6.50092423e-01 -4.36167300e-01
5.92433512e-01 5.53868651e-01 3.25650603e-01 -1.64999977e-01
-2.51267612e-01 -8.49202275e-02 5.38648903e-01 3.94837230e-01
1.09268653e+00 1.19746342e-01 -6.28846407e-01 1.62001505e-01
-5.79076529e-01 -4.84578591e-03 9.73170698e-02 9.52134788e-01
-3.03286016e-01 -1.55161810e+00 -5.43911517e-01 2.55504966e-01
-5.63780963e-01 -1.46111116e-01 -4.72356647e-01 9.86927450e-01
-2.19400406e-01 5.66941381e-01 -1.72725603e-01 -4.19816792e-01
3.12186807e-01 -1.75462157e-01 1.06711149e+00 -6.74271643e-01
-3.71608675e-01 3.26906472e-01 2.40611240e-01 -1.23514760e+00
-3.04384202e-01 -9.18357968e-01 -1.23297226e+00 -7.06419706e-01
-3.38655680e-01 -1.17879502e-01 5.46066225e-01 3.65397185e-01
3.69223803e-01 3.99539322e-01 1.38342321e-01 -1.16920555e+00
-5.76634645e-01 -8.57333064e-01 -1.89359546e-01 9.62957274e-03
2.71964595e-02 -8.55460286e-01 1.94698554e-02 -1.25638470e-01] | [7.17161226272583, -0.7985823154449463] |
eb330d1f-c33d-48d6-81d2-eefb7a7ead1d | a-novel-filter-approach-for-band-selection | 2210.15477 | null | https://arxiv.org/abs/2210.15477v1 | https://arxiv.org/pdf/2210.15477v1.pdf | A Novel Filter Approach for Band Selection and Classification of Hyperspectral Remotely Sensed Images Using Normalized Mutual Information and Support Vector Machines | Band selection is a great challenging task in the classification of hyperspectral remotely sensed images HSI. This is resulting from its high spectral resolution, the many class outputs and the limited number of training samples. For this purpose, this paper introduces a new filter approach for dimension reduction and classification of hyperspectral images using information theoretic (normalized mutual information) and support vector machines SVM. This method consists to select a minimal subset of the most informative and relevant bands from the input datasets for better classification efficiency. We applied our proposed algorithm on two well-known benchmark datasets gathered by the NASA's AVIRIS sensor over Indiana and Salinas valley in USA. The experimental results were assessed based on different evaluation metrics widely used in this area. The comparison with the state of the art methods proves that our method could produce good performance with reduced number of selected bands in a good timing. Keywords: Dimension reduction, Hyperspectral images, Band selection, Normalized mutual information, Classification, Support vector machines | ['Ahmed Hammouch', 'Elkebir Sarhrouni', 'Asma Elmaizi', 'Hasna Nhaila'] | 2022-10-27 | null | null | null | null | ['classification-of-hyperspectral-images'] | ['computer-vision'] | [ 8.65092933e-01 -7.09480345e-01 -1.71483180e-03 -3.18676233e-01
-3.02711964e-01 -6.18262291e-01 2.92691469e-01 1.91619009e-01
-3.00211072e-01 1.04731870e+00 -2.23718479e-01 -7.02806339e-02
-1.03690696e+00 -1.09021354e+00 1.58593193e-01 -9.36683476e-01
-4.37804401e-01 2.90680438e-01 -3.28008160e-02 -2.72095591e-01
5.06781042e-01 7.44700491e-01 -2.20629382e+00 2.22669646e-01
1.12713754e+00 1.05892146e+00 3.93292069e-01 6.41448736e-01
1.72859237e-01 2.11143628e-01 -9.20003876e-02 5.16735673e-01
7.14818239e-01 -6.23287022e-01 -7.67687201e-01 3.11849892e-01
1.94046259e-01 4.79882136e-02 1.06601045e-01 1.48947203e+00
4.42030042e-01 6.39630914e-01 1.04549372e+00 -9.79419112e-01
-3.71550173e-02 6.02209270e-01 -8.12820494e-01 4.32780445e-01
-1.91517845e-01 -2.20934242e-01 1.01655591e+00 -6.34559035e-01
2.79677153e-01 8.20150793e-01 4.00235206e-01 -7.71312490e-02
-1.30125344e+00 -8.53077590e-01 -3.85106504e-01 4.60438371e-01
-1.48222983e+00 7.16459900e-02 6.08047307e-01 -7.53443122e-01
7.63896644e-01 9.78711784e-01 6.92510545e-01 -1.03234932e-01
-9.35234055e-02 2.20832884e-01 1.46979558e+00 -6.26085877e-01
2.74187863e-01 3.76412153e-01 7.41904259e-01 4.86120403e-01
5.56561053e-01 2.66099870e-01 -5.61446175e-02 -3.58116060e-01
1.30641818e-01 4.30681631e-02 -5.61026156e-01 -2.83155382e-01
-8.28383744e-01 1.01863992e+00 4.02192980e-01 5.14753759e-01
-4.85672891e-01 -9.05011177e-01 3.93182337e-01 3.62971425e-01
4.91131723e-01 4.55731541e-01 -5.49505234e-01 6.09090745e-01
-1.00456595e+00 -7.03976750e-02 5.09954929e-01 3.86453956e-01
9.91450965e-01 -2.33597979e-02 7.33924285e-02 1.05223310e+00
2.39284441e-01 7.27718055e-01 5.83300352e-01 -2.76792258e-01
8.50056484e-02 7.91401327e-01 -2.35593803e-02 -1.21473110e+00
-6.56103015e-01 -6.02532566e-01 -9.67728913e-01 5.73071420e-01
-1.22544006e-01 -2.05308542e-01 -8.87527764e-01 1.07331645e+00
1.82719558e-01 -5.47788665e-02 4.13299769e-01 1.03948331e+00
5.59388399e-01 9.70692873e-01 8.18663687e-02 -5.81368744e-01
1.04599452e+00 -5.09631932e-01 -4.09615755e-01 -1.59595624e-01
2.51540244e-01 -8.61264229e-01 6.65718734e-01 6.50232315e-01
-1.62849352e-01 -4.04366970e-01 -1.32921827e+00 9.11595166e-01
-6.27852142e-01 6.38596416e-01 8.41827571e-01 8.46016645e-01
-3.37372422e-01 8.33835721e-01 -5.09795308e-01 -6.33740783e-01
3.85746837e-01 4.15569812e-01 -6.12846613e-01 8.60716999e-02
-1.05094707e+00 6.93784237e-01 1.17925489e+00 -4.64773476e-02
-8.26538429e-02 -3.67556930e-01 -5.57499468e-01 9.24094617e-02
2.10710820e-02 -3.19948886e-03 3.95116746e-01 -1.12252307e+00
-1.09782135e+00 7.05764771e-01 2.58024544e-01 -3.94671738e-01
1.49405131e-03 1.16438337e-01 -5.13466537e-01 1.87984094e-01
-1.55180320e-01 6.05432875e-02 5.10363102e-01 -1.00572419e+00
-1.05150628e+00 -9.03705239e-01 -5.83699346e-01 2.80236751e-01
-6.55356109e-01 -2.75090337e-01 7.98582077e-01 -4.34127033e-01
8.03847194e-01 -9.87041652e-01 -2.44616345e-01 -6.55399144e-01
-3.49915415e-01 1.40875652e-01 1.14140916e+00 -6.58969283e-01
9.51611161e-01 -2.14817953e+00 8.91233981e-02 7.62714207e-01
-4.47727233e-01 4.73649889e-01 7.46290386e-02 4.88212675e-01
-4.79601443e-01 7.15795532e-02 -6.17429376e-01 6.43645823e-01
-5.23176134e-01 -2.00083748e-01 -1.76616997e-01 7.63858914e-01
-1.03276476e-01 -2.65011936e-01 -6.39266610e-01 -2.20361769e-01
5.87521255e-01 4.03993577e-01 -5.46764545e-02 3.45884450e-02
5.21299131e-02 2.59537190e-01 -3.45755458e-01 6.40912414e-01
1.01476645e+00 -5.09263650e-02 7.25539774e-02 -8.06561887e-01
-3.07727367e-01 -4.84111041e-01 -1.53604472e+00 1.20212364e+00
5.73350564e-02 6.77872777e-01 -2.53679335e-01 -1.39148045e+00
1.03648615e+00 1.70920968e-01 6.22544885e-01 -1.75909966e-01
2.27373734e-01 2.57031262e-01 -8.67413729e-03 -6.04963064e-01
2.62068182e-01 -3.21713269e-01 5.80253124e-01 2.45244369e-01
-2.79474203e-02 -3.83161843e-01 4.50919330e-01 -3.59235555e-01
3.31989527e-01 -3.31998557e-01 9.16352570e-01 -6.63544357e-01
8.69860291e-01 4.62530404e-01 4.26951140e-01 4.05727983e-01
-2.47689009e-01 2.33646557e-01 -1.98334053e-01 -2.45771676e-01
-8.42172682e-01 -4.23082322e-01 -6.84601486e-01 8.43350768e-01
5.68914525e-02 3.13909799e-01 -4.57035333e-01 -4.03092295e-01
2.01015435e-02 7.98001945e-01 -4.27742332e-01 1.22251958e-02
1.45471096e-01 -1.41255081e+00 1.45047456e-01 -3.32152843e-01
7.86123037e-01 -5.71254075e-01 -6.34059727e-01 5.13046794e-02
1.97444052e-01 -4.55839843e-01 3.41194689e-01 3.48501980e-01
-1.20169675e+00 -1.17627680e+00 -3.64641577e-01 -6.22458220e-01
4.42695618e-01 7.42264688e-01 4.41173166e-01 -3.57667774e-01
-7.10078120e-01 -2.83450723e-01 -6.65747881e-01 -6.91978216e-01
-1.37271062e-01 2.51244325e-02 -2.19512433e-01 1.80386841e-01
6.39002204e-01 -5.13963759e-01 -4.96175557e-01 2.40369201e-01
-1.12811911e+00 -1.50422171e-01 6.80410802e-01 1.06172287e+00
5.33249140e-01 1.13359761e+00 3.03098083e-01 -8.71950269e-01
3.97945523e-01 -4.87861156e-01 -9.37008679e-01 2.25817889e-01
-8.50409448e-01 -9.04847458e-02 5.06298244e-01 1.12683289e-01
-9.73796666e-01 3.19322824e-01 4.50977325e-01 3.74597847e-01
-3.66525471e-01 8.38501513e-01 5.41268773e-02 -4.98085976e-01
1.27102923e+00 4.06902701e-01 3.31655204e-01 -4.42989558e-01
-1.50113344e-01 1.10949910e+00 1.12136818e-01 6.81965500e-02
8.69281590e-01 5.04324853e-01 4.06477183e-01 -1.26398921e+00
-7.41754174e-01 -9.62007403e-01 -6.51006579e-01 -1.49184689e-01
4.94981498e-01 -7.79951274e-01 -3.55035186e-01 3.68978590e-01
-3.79004836e-01 3.46373171e-01 8.05222020e-02 9.94933307e-01
-1.79359481e-01 3.65895838e-01 -4.75618616e-02 -1.08491707e+00
-6.55544221e-01 -9.15359676e-01 4.75951821e-01 5.20075202e-01
1.96538270e-01 -6.49552941e-01 -3.82918939e-02 2.68390775e-01
4.89893287e-01 4.81255293e-01 1.10473418e+00 -7.93948948e-01
-4.84642573e-02 -3.42091709e-01 -3.70702684e-01 7.42176652e-01
7.37101316e-01 1.96754038e-01 -9.36952949e-01 -4.26558256e-01
9.04257894e-02 -1.11231714e-01 1.09261239e+00 4.92801070e-01
9.50377524e-01 -3.68988588e-02 -3.72830957e-01 6.72133744e-01
2.23663402e+00 6.68115020e-01 3.86458963e-01 6.52957261e-01
2.39915401e-01 6.77396178e-01 1.11567688e+00 7.72910118e-01
-6.28597677e-01 -5.72518706e-02 5.72913349e-01 -8.40139985e-02
4.80385184e-01 5.13611019e-01 -3.56440485e-01 2.95768738e-01
-3.04239929e-01 -1.90719590e-01 -9.63092744e-01 1.69433847e-01
-1.62285423e+00 -1.33252108e+00 -3.51657867e-01 2.40287995e+00
5.54196000e-01 -3.93278152e-01 -7.51694515e-02 1.04112387e+00
8.63923907e-01 1.34712636e-01 -3.39539826e-01 1.15480855e-01
-4.77470726e-01 2.44751051e-01 9.99749482e-01 5.65851450e-01
-1.72557056e+00 5.67494333e-01 5.38452959e+00 7.44252622e-01
-1.47056842e+00 -2.06606597e-01 5.78651488e-01 7.42697641e-02
2.74738729e-01 -1.95040122e-01 -4.45584089e-01 1.14583887e-01
5.98069012e-01 -4.09569174e-01 4.84661490e-01 7.07844675e-01
2.45688215e-01 -4.73924339e-01 -3.23969573e-01 9.60539043e-01
1.06429927e-01 -1.07716620e+00 2.15015739e-01 -1.33278117e-01
9.94290471e-01 5.51825836e-02 -3.78072485e-02 -2.95000553e-01
-1.66455045e-01 -9.82299984e-01 9.51135606e-02 5.31394899e-01
4.02920842e-01 -1.11481988e+00 8.18300366e-01 3.89669389e-01
-1.04907584e+00 -5.28561890e-01 -6.94417000e-01 -1.27403215e-01
-5.37618399e-01 9.66483355e-01 -8.09670150e-01 7.75809526e-01
5.63880682e-01 7.82290518e-01 -4.20655400e-01 1.38865137e+00
3.97150129e-01 7.04704344e-01 -3.55553061e-01 -1.87622309e-01
2.90328264e-01 -6.85542166e-01 6.78332984e-01 1.07003129e+00
5.03082991e-01 5.60860097e-01 2.97255695e-01 2.70619571e-01
6.76973879e-01 8.23776901e-01 -7.74626613e-01 -1.37198880e-01
4.78004366e-01 1.19895554e+00 -8.26155066e-01 -2.56405622e-01
-1.74383298e-01 6.28140211e-01 -4.89361256e-01 9.66627970e-02
-4.13301528e-01 -7.98118055e-01 5.45012653e-01 -2.34828159e-01
-8.98043215e-02 -2.95958128e-02 -4.18989390e-01 -7.58830130e-01
-4.10569489e-01 -8.76293778e-01 8.49250376e-01 -5.01351476e-01
-1.22771192e+00 8.94100547e-01 9.13872346e-02 -1.61855626e+00
2.05916688e-01 -7.82515287e-01 -7.58890994e-03 1.14606369e+00
-1.61127436e+00 -6.76132977e-01 -6.39112771e-01 5.13242364e-01
4.91366148e-01 -6.47886872e-01 1.11397576e+00 1.98775277e-01
-5.32646894e-01 -1.91578716e-01 7.57563591e-01 -2.02731863e-01
4.47995573e-01 -1.05336046e+00 -6.59194648e-01 9.11218405e-01
-2.33075172e-01 1.50014535e-01 9.01768625e-01 -6.20768309e-01
-1.16603839e+00 -1.11523747e+00 2.45600745e-01 5.53219140e-01
3.84934008e-01 5.06112397e-01 -7.24050641e-01 2.55376875e-01
1.00879580e-01 -1.48873851e-01 1.26487923e+00 -3.07732731e-01
7.69260451e-02 -5.20248115e-01 -1.50187194e+00 1.39066413e-01
2.76103973e-01 -1.05048977e-01 -3.32918465e-01 9.21448708e-01
-4.49479967e-02 2.24268511e-01 -8.50867033e-01 7.04043508e-01
5.26345372e-01 -1.17137039e+00 9.23235893e-01 -4.68269348e-01
7.96484798e-02 -4.95634735e-01 -4.91451919e-01 -1.53721631e+00
-6.80428743e-01 1.06609287e-02 8.93216372e-01 8.57096851e-01
5.60513020e-01 -7.06171095e-01 7.90074229e-01 2.10343197e-01
3.80346984e-01 -6.00839220e-02 -4.05177295e-01 -6.85593128e-01
-4.37874943e-01 1.57523602e-01 5.60438097e-01 1.25989604e+00
3.54987830e-02 2.40847915e-01 -3.62972528e-01 6.22315347e-01
1.07846892e+00 5.20733535e-01 2.20083252e-01 -2.01825857e+00
-5.57356067e-02 -4.22961056e-01 -8.74645591e-01 2.53379971e-01
-1.99105442e-02 -9.29495871e-01 -7.47932643e-02 -1.37447393e+00
2.49016970e-01 -4.74264473e-01 -4.64594811e-01 4.27777290e-01
-1.81397706e-01 3.32297027e-01 -1.83473788e-02 2.19133243e-01
4.39032495e-01 1.22245975e-01 8.82780075e-01 -3.94725472e-01
-5.50966859e-01 1.65251136e-01 -4.84366864e-01 6.27219439e-01
1.16503584e+00 -3.88087779e-01 -7.41955757e-01 -3.67903374e-02
-1.67480886e-01 -2.62939185e-02 -1.38859630e-01 -1.46664786e+00
4.22019139e-02 -6.81606948e-01 5.93461037e-01 -6.33597493e-01
1.08779691e-01 -1.11051297e+00 5.63881457e-01 7.32077062e-01
-1.83035895e-01 -6.04850709e-01 2.12930992e-01 3.57606798e-01
-3.70680183e-01 -5.23726225e-01 1.33791876e+00 -2.93218315e-01
-1.19342780e+00 2.18137071e-01 -2.67143250e-01 -7.94797897e-01
1.30065417e+00 -4.08880532e-01 -2.48100534e-01 1.32137269e-01
-7.20214665e-01 -2.38866016e-01 1.36007130e-01 -5.53605594e-02
4.06753659e-01 -9.85370696e-01 -9.92622137e-01 4.47009951e-01
3.38098675e-01 -7.05758452e-01 1.93666160e-01 4.44818020e-01
-8.90488088e-01 5.25534391e-01 -7.91960597e-01 -5.96244156e-01
-1.74079609e+00 4.16015312e-02 3.90472174e-01 6.73419237e-02
-5.80954105e-02 8.22879732e-01 -4.63646650e-01 -2.61856407e-01
-1.67478815e-01 -8.39631259e-03 -9.60875452e-01 4.59949702e-01
6.84332311e-01 5.86640239e-01 4.03598756e-01 -7.91328430e-01
-3.59846145e-01 6.59059644e-01 3.86006534e-01 1.65148079e-02
1.54443014e+00 7.62879997e-02 -5.74835300e-01 3.26556057e-01
1.18609238e+00 -1.79228008e-01 -4.49128628e-01 -2.05220819e-01
-1.88059211e-02 -9.22443449e-01 6.27104521e-01 -8.68865788e-01
-8.97381127e-01 4.85733509e-01 1.43554437e+00 5.06753445e-01
1.57745838e+00 -8.12201738e-01 -9.56515968e-03 7.21254647e-01
3.67882252e-01 -1.22418225e+00 -8.48450005e-01 2.98072278e-01
6.44667745e-01 -1.61622632e+00 4.04961169e-01 -7.75835156e-01
-4.34473872e-01 1.46150243e+00 3.02896321e-01 4.50070165e-02
1.06334651e+00 -8.19491073e-02 6.16965555e-02 3.17154005e-02
-2.25376874e-01 -4.70651805e-01 2.84419626e-01 6.20934367e-01
6.21256709e-01 4.31810439e-01 -8.87247443e-01 5.43695912e-02
-1.63513064e-01 -1.02872506e-01 3.71305674e-01 1.07998765e+00
-1.24625814e+00 -7.91931450e-01 -7.58358896e-01 1.06549597e+00
-3.23912412e-01 -1.56117022e-01 -1.85066149e-01 5.89246571e-01
3.18302331e-03 1.08727503e+00 -3.31470847e-01 -5.86973310e-01
2.29667842e-01 1.42240539e-01 1.32645845e-01 -5.81571639e-01
-2.29240492e-01 -1.06403597e-01 -6.87740669e-02 -6.28316104e-02
-7.19280779e-01 -6.56158566e-01 -1.03891706e+00 5.81775941e-02
-7.90217578e-01 6.71470106e-01 1.12865472e+00 7.47962296e-01
-2.08498025e-03 1.59068584e-01 1.04804242e+00 -9.14731681e-01
-8.57656956e-01 -1.19473922e+00 -1.34524190e+00 3.27562928e-01
2.16792732e-01 -5.73577225e-01 -6.50545657e-01 1.92847162e-01] | [9.80390453338623, -1.8463932275772095] |
b184c74a-96ab-4a59-8c8b-2196fafbfa07 | matrix-completion-with-variational-graph | 1811.01662 | null | http://arxiv.org/abs/1811.01662v1 | http://arxiv.org/pdf/1811.01662v1.pdf | Matrix Completion With Variational Graph Autoencoders: Application in Hyperlocal Air Quality Inference | Inferring air quality from a limited number of observations is an essential
task for monitoring and controlling air pollution. Existing inference methods
typically use low spatial resolution data collected by fixed monitoring
stations and infer the concentration of air pollutants using additional types
of data, e.g., meteorological and traffic information. In this work, we focus
on street-level air quality inference by utilizing data collected by mobile
stations. We formulate air quality inference in this setting as a graph-based
matrix completion problem and propose a novel variational model based on graph
convolutional autoencoders. Our model captures effectively the spatio-temporal
correlation of the measurements and does not depend on the availability of
additional information apart from the street-network topology. Experiments on a
real air quality dataset, collected with mobile stations, shows that the
proposed model outperforms state-of-the-art approaches. | ['Evaggelia Tsiligianni', 'Tien Huu Do', 'Nikos Deligiannis', 'Frank Pasveer', 'Duc Minh Nguyen', 'Valerio Panzica La Manna', 'Wilfried Philips', 'Angel Lopez Aguirre'] | 2018-11-05 | null | null | null | null | ['air-quality-inference'] | ['miscellaneous'] | [-5.89645691e-02 -3.47220600e-01 -1.67795084e-02 -2.65694350e-01
-4.59217787e-01 -4.05634165e-01 4.88269895e-01 5.24229586e-01
-3.58894348e-01 7.43240654e-01 1.58420384e-01 -4.38411534e-01
-6.96128070e-01 -1.49005747e+00 -9.43333566e-01 -6.15375578e-01
5.32072484e-02 4.02431130e-01 -1.44600004e-01 1.01270694e-02
-6.46580815e-01 3.57193023e-01 -9.76579010e-01 -2.12741256e-01
8.09377491e-01 6.90025151e-01 3.04910779e-01 8.58205318e-01
-9.01022777e-02 3.01424652e-01 -2.24703103e-01 -1.28031254e-01
1.70684010e-01 -9.24451798e-02 -3.13778609e-01 2.06926823e-01
4.35349405e-01 -2.30239913e-01 -7.07918346e-01 1.05731642e+00
2.16052741e-01 5.84636211e-01 4.10121828e-01 -1.13124788e+00
-3.63873482e-01 2.22536445e-01 -9.94948074e-02 5.41536152e-01
-6.15473874e-02 1.54465452e-01 1.18132770e+00 -3.24855596e-01
-6.03614040e-02 1.01822340e+00 7.79572725e-01 6.03453219e-02
-1.48995376e+00 -4.92537230e-01 5.54323792e-01 1.23014987e-01
-1.63516557e+00 -1.26189888e-01 7.37796783e-01 -5.92205882e-01
6.72836065e-01 3.58554423e-01 6.86388314e-01 9.51410294e-01
9.82970297e-02 1.98224243e-02 9.26009774e-01 1.50444359e-01
4.29880410e-01 6.38626441e-02 2.41401508e-01 6.45421803e-01
6.66606486e-01 2.02507466e-01 -1.18533866e-02 -4.11913276e-01
5.39520800e-01 8.15064013e-01 -3.23092729e-01 2.36954376e-01
-7.58802056e-01 9.65203226e-01 8.22710216e-01 2.91878015e-01
-9.20961499e-01 3.42768222e-01 -2.82698274e-01 1.49155483e-01
9.33337986e-01 -4.37342487e-02 -4.44889694e-01 3.70678931e-01
-1.05729198e+00 3.35987061e-02 6.31194592e-01 8.41214120e-01
1.17207515e+00 4.12702113e-02 -6.81129247e-02 4.17446285e-01
8.98158312e-01 1.16848505e+00 -3.33831072e-01 -6.53980911e-01
6.88945413e-01 4.40612704e-01 2.51175225e-01 -1.33401000e+00
-6.67599797e-01 -5.23006558e-01 -1.20245123e+00 -2.98860371e-01
2.96023190e-01 -7.57731795e-01 -5.57626367e-01 1.63094366e+00
5.79425931e-01 8.62351894e-01 -4.24081475e-01 7.81484306e-01
8.75261426e-01 9.52232480e-01 -5.23362160e-02 -1.28492519e-01
1.08720076e+00 -4.88257736e-01 -9.28424358e-01 1.57446638e-01
9.12636239e-03 4.19494435e-02 5.30765235e-01 5.89197576e-02
-7.21857607e-01 -6.45397842e-01 -6.58835113e-01 4.22813863e-01
-1.02799428e+00 -6.87773004e-02 3.58714938e-01 7.71618247e-01
-8.75775814e-01 7.12596595e-01 -1.19839311e+00 -1.26767665e-01
2.58615702e-01 3.33530605e-01 -1.66586831e-01 -2.05251768e-01
-1.37775540e+00 2.41895631e-01 -6.33956417e-02 5.13798654e-01
-9.00960922e-01 -9.63425934e-01 -8.52711260e-01 1.16494566e-01
3.66017610e-01 -1.14203775e+00 8.24566722e-01 -4.41420600e-02
-1.06064546e+00 -1.54263407e-01 -4.57460493e-01 -3.76333952e-01
1.85780242e-01 -1.77119210e-01 -9.13567483e-01 1.49734020e-01
6.78545088e-02 -1.46576524e-01 9.73533750e-01 -1.11879575e+00
-5.47960341e-01 -4.75789696e-01 3.93295288e-01 -3.91871721e-01
-3.82493377e-01 -5.50126314e-01 -2.43393302e-01 1.20199071e-02
-4.49841052e-01 -1.14168346e+00 -5.10349751e-01 -6.49387166e-02
-5.64768374e-01 -2.86486596e-01 5.45690000e-01 -7.60807991e-01
1.41990221e+00 -1.75102162e+00 -7.41184130e-02 5.42158961e-01
5.89946270e-01 2.24807963e-01 -1.28342763e-01 6.77156448e-01
2.15766072e-01 6.52921051e-02 -4.79503989e-01 -6.67508364e-01
2.41206676e-01 5.23806572e-01 8.61791819e-02 8.72485042e-01
2.25825712e-01 7.87655592e-01 -1.22800195e+00 -2.58090466e-01
6.07237160e-01 8.39496911e-01 -5.16740680e-01 2.43214518e-01
-3.18634123e-01 9.35291946e-01 -6.71220720e-01 -4.25825045e-02
7.43665397e-01 -5.63206971e-01 8.27070698e-02 -9.58645567e-02
-1.48287371e-01 3.52575153e-01 -1.42526722e+00 1.59909928e+00
-9.86817837e-01 5.16820312e-01 4.34262782e-01 -9.00962114e-01
2.21160412e-01 7.14495003e-01 5.41700721e-01 -4.51580107e-01
-4.73512076e-02 -4.06774282e-01 -2.27434427e-01 -7.01072037e-01
8.80008638e-02 -2.86175404e-02 3.44721645e-01 4.52953517e-01
-2.21936658e-01 1.80557698e-01 1.46987215e-01 1.42491475e-01
9.79891658e-01 -4.36537564e-01 -1.31763101e-01 -1.19524345e-01
5.62383354e-01 -2.70556808e-01 3.96766394e-01 9.03949440e-01
1.52089670e-01 1.49034366e-01 -1.04214035e-01 -1.33685723e-01
-8.01651478e-01 -1.27324831e+00 -2.91262329e-01 7.28144944e-01
-2.88920581e-01 -5.04963994e-01 -5.72711468e-01 -5.57147443e-01
2.73992628e-01 6.51149452e-01 -8.41131091e-01 1.80758208e-01
-1.97717115e-01 -1.18610418e+00 2.17909366e-01 3.25108469e-01
4.52146530e-01 -2.47443467e-01 9.63226706e-02 3.48070413e-01
-3.51519674e-01 -1.29767120e+00 -1.32160693e-01 -4.43703383e-01
-8.93089592e-01 -1.22733843e+00 -2.36890748e-01 9.62224007e-02
6.15164578e-01 5.68143010e-01 1.30206740e+00 -5.04661277e-02
-7.16781691e-02 7.47818291e-01 1.03256926e-01 -5.77202857e-01
-1.93992093e-01 1.69128761e-01 1.38862357e-01 7.07864404e-01
3.55819523e-01 -1.02856302e+00 -7.20963538e-01 -4.51177433e-02
-1.14990497e+00 -4.30933893e-01 4.01535094e-01 4.18136686e-01
7.32123971e-01 8.05419028e-01 4.92479593e-01 -8.73878837e-01
6.39399111e-01 -1.06510913e+00 -9.60027397e-01 -2.02470124e-01
-5.61886668e-01 5.35710603e-02 8.13800275e-01 -1.11446656e-01
-9.34292376e-01 -1.48181260e-01 -8.50860104e-02 -4.72038776e-01
-6.30945683e-01 6.72922134e-01 -1.72937959e-01 3.10587227e-01
2.87758976e-01 4.39817049e-02 -4.26609844e-01 -8.97113979e-01
4.52591658e-01 6.75032675e-01 4.45945933e-02 -2.62810081e-01
1.14844251e+00 9.18963909e-01 4.72157806e-01 -1.38240123e+00
-1.00916147e+00 -8.47833335e-01 -8.27207386e-01 -1.87358871e-01
1.15920162e+00 -1.28672242e+00 -8.58698010e-01 -1.98293962e-02
-1.01646292e+00 -3.25755388e-01 -9.77763087e-02 9.36391175e-01
3.29347551e-01 8.20942130e-03 -4.59694505e-01 -1.04673743e+00
-1.05416812e-01 -6.91694081e-01 1.18665671e+00 -1.14042893e-01
2.01327577e-01 -1.93144584e+00 6.28927529e-01 3.69845033e-01
5.07229686e-01 3.54135573e-01 6.57992780e-01 -1.76539868e-01
-8.74668658e-01 -3.24656755e-01 -3.76712650e-01 3.53278816e-01
6.43675923e-01 -3.15568328e-01 -1.03036499e+00 -3.14978153e-01
-8.15187842e-02 4.30336237e-01 9.74697649e-01 5.43654144e-01
1.07329917e+00 -6.50515079e-01 -3.16894382e-01 5.11026084e-01
1.58359325e+00 -4.90789860e-01 9.50890034e-02 -4.04891521e-01
1.24200988e+00 3.99838775e-01 -2.15626493e-01 5.29911339e-01
5.83008528e-01 4.32371050e-01 8.00429583e-01 -2.42185190e-01
1.43444791e-01 -2.97472149e-01 2.48641404e-03 1.03499675e+00
-1.50395676e-01 -4.47975397e-01 -6.57179713e-01 7.41469681e-01
-1.88299918e+00 -1.02041399e+00 -7.62405396e-01 2.13045001e+00
3.56172979e-01 -3.46889257e-01 2.24706337e-01 7.42783770e-02
6.09116852e-01 5.06967306e-01 -3.55987251e-01 1.32136539e-01
3.21678698e-01 3.66138458e-01 8.06068242e-01 9.60170567e-01
-1.22275198e+00 2.76302934e-01 5.89822912e+00 2.39022121e-01
-7.67935812e-01 5.08236587e-01 -9.72312167e-02 -1.93038806e-01
-4.46958452e-01 -3.39502007e-01 -7.60557294e-01 5.91911018e-01
1.43941402e+00 3.98713976e-01 7.68708646e-01 2.87051290e-01
8.64378691e-01 1.84974626e-01 -9.27429676e-01 6.83011889e-01
-3.69271517e-01 -1.15677524e+00 -7.99972862e-02 4.50796545e-01
8.48936439e-01 6.20757639e-01 8.83690342e-02 -4.17172275e-02
5.79797804e-01 -1.00482643e+00 9.09794308e-03 1.11586022e+00
3.02440286e-01 -6.45879388e-01 6.19071066e-01 3.90623778e-01
-1.48427296e+00 7.35047236e-02 -4.13795292e-01 -3.53026718e-01
2.09361330e-01 1.19953609e+00 -6.70666039e-01 9.18935478e-01
6.42413020e-01 1.12494254e+00 -5.03008604e-01 8.67108285e-01
-2.54306525e-01 1.26420105e+00 -6.89760089e-01 -7.49320537e-02
4.17001635e-01 -6.62604034e-01 7.22094536e-01 1.31799614e+00
1.73374325e-01 1.16869040e-01 2.62152851e-01 1.25476289e+00
-1.30698264e-01 -2.57348478e-01 -9.49254632e-01 3.42998132e-02
2.30416432e-01 1.21145499e+00 8.22551474e-02 -3.77347618e-01
-9.20739114e-01 5.45366466e-01 1.72188818e-01 7.93402731e-01
-8.58188987e-01 -1.57976061e-01 1.04846931e+00 1.71585992e-01
4.92067009e-01 -7.31670856e-01 2.20760435e-01 -1.05095458e+00
-7.04427883e-02 -2.43742719e-01 2.74617463e-01 -4.22998697e-01
-1.39456511e+00 8.14898126e-03 1.30708948e-01 -9.40236032e-01
6.03651926e-02 -5.25144756e-01 -9.38981891e-01 1.00414753e+00
-2.00429249e+00 -7.91249335e-01 -4.28069979e-01 9.87406254e-01
6.95719123e-02 3.41805995e-01 6.76017463e-01 7.11220980e-01
-1.03549445e+00 -6.30339468e-03 5.80451608e-01 4.39121962e-01
-8.48076120e-02 -1.36818790e+00 3.86783600e-01 9.22088265e-01
5.37044823e-01 5.76939642e-01 5.84751785e-01 -5.40308297e-01
-1.63571906e+00 -2.00140643e+00 8.75446498e-01 -5.24791002e-01
9.37314153e-01 -3.59247535e-01 -8.01078975e-01 6.60648048e-01
1.49289608e-01 2.57259041e-01 1.02976429e+00 2.98035353e-01
-1.48293033e-01 -4.84995484e-01 -8.49923611e-01 2.29804739e-01
8.63655090e-01 -8.47124219e-01 -3.04209173e-01 7.95472980e-01
7.76766837e-01 1.56907350e-01 -1.17393088e+00 -5.00046797e-02
8.43649358e-02 -4.55615461e-01 1.05402863e+00 -7.70417511e-01
2.84296162e-02 -5.33227623e-01 -2.45256513e-01 -1.68911004e+00
-7.31553316e-01 -1.94121718e-01 -2.34406084e-01 9.52044904e-01
5.94512761e-01 -7.07677722e-01 4.49526429e-01 6.27106190e-01
1.83554873e-01 -5.15250377e-02 -7.92133510e-01 -7.77007222e-01
1.40306754e-02 -8.10566127e-01 9.70742702e-01 9.15183008e-01
-5.66004097e-01 5.22165000e-01 -5.00532448e-01 1.19867289e+00
9.82859194e-01 1.31319463e-01 6.89616501e-01 -1.64915633e+00
-5.15722752e-01 6.55498728e-02 -3.05965506e-02 -1.01097393e+00
4.42454249e-01 -9.15616274e-01 -1.40430376e-01 -1.77555943e+00
-1.81946546e-01 -2.65587807e-01 -5.52521586e-01 3.36171910e-02
-2.45374322e-01 1.15787961e-01 -1.47468120e-01 -4.46799815e-01
-6.06673539e-01 6.48934007e-01 1.00180447e+00 -6.53192818e-01
-3.31598103e-01 2.81462729e-01 -2.27113768e-01 5.72328866e-01
1.00144219e+00 -7.41588295e-01 -5.48590541e-01 -8.80752385e-01
5.53486049e-01 9.79564041e-02 7.45782554e-01 -1.05862880e+00
2.72663891e-01 -4.50425833e-01 4.46854591e-01 -5.45423210e-01
6.42875791e-01 -1.28850067e+00 3.44530642e-01 2.93902636e-01
-2.45017409e-01 1.08853318e-01 1.17305607e-01 1.48575974e+00
7.85372406e-02 1.43716782e-01 2.09996492e-01 -2.32424125e-01
-1.19883880e-01 9.66962695e-01 -6.88687027e-01 -3.45716923e-02
3.80300671e-01 5.06282508e-01 5.76970400e-03 -5.20032048e-01
-9.69723821e-01 4.10118997e-01 -9.15116668e-02 1.79901049e-01
4.48057622e-01 -1.34973395e+00 -8.96901608e-01 1.75155342e-01
8.38823766e-02 -8.08952842e-04 4.43314493e-01 8.92188489e-01
1.21052243e-01 8.47754180e-01 5.84730148e-01 -6.25659108e-01
-9.47288513e-01 7.41465986e-01 5.71675181e-01 -3.77031416e-01
-5.64147055e-01 3.52795333e-01 3.99465486e-02 -4.71890807e-01
-1.55509487e-01 -6.65055513e-01 -2.13321239e-01 -3.51829343e-02
5.08379161e-01 8.83676708e-01 2.29563355e-01 -6.75170064e-01
-3.34734350e-01 3.42336982e-01 6.89241111e-01 -7.32315406e-02
1.35659885e+00 -5.01711369e-01 -7.45638534e-02 9.59781110e-01
1.47279644e+00 1.99290320e-01 -1.05464160e+00 -5.83734155e-01
-4.92958963e-01 -3.75980973e-01 6.66169822e-01 -2.33216941e-01
-1.40496516e+00 1.21385324e+00 5.36550045e-01 5.86098552e-01
7.38967180e-01 -2.86153704e-01 1.03504658e+00 6.51853204e-01
1.03855534e-02 -9.26553845e-01 -4.07486111e-01 3.10316950e-01
3.84047687e-01 -1.35243785e+00 -5.30373491e-02 -1.93266705e-01
2.77173638e-01 6.18899703e-01 1.01743564e-02 -2.93491423e-01
1.51827729e+00 -3.88235807e-01 -1.36279225e-01 -4.27801669e-01
-7.38421381e-01 -6.91897511e-01 5.15934825e-01 7.30734587e-01
1.50287136e-01 5.43033898e-01 3.69440377e-01 2.97019452e-01
-5.91789894e-02 -2.20529824e-01 1.84846282e-01 3.99092138e-01
-2.42467329e-01 -8.58630300e-01 -4.71217752e-01 7.26216197e-01
-5.03610730e-01 -2.82815337e-01 -1.16522878e-01 4.44430947e-01
2.90588140e-01 1.80761862e+00 1.96466044e-01 -2.05214709e-01
3.98343682e-01 -2.03221887e-02 1.40660599e-01 -5.28108239e-01
-3.57932270e-01 -3.08594763e-01 2.86952592e-02 -5.24771094e-01
-7.59034812e-01 -5.89444280e-01 -6.50940120e-01 -6.77640021e-01
-2.02462122e-01 4.20614660e-01 6.87705338e-01 1.16628122e+00
4.45483387e-01 1.10139143e+00 8.27118814e-01 -7.12420762e-01
-1.71984151e-01 -8.51596475e-01 -9.78264749e-01 4.41633850e-01
1.22676110e+00 -5.09968817e-01 -4.61206019e-01 -2.31498122e-01] | [6.253604888916016, 2.524998426437378] |
24994a39-9a48-4e6b-a0b0-9e77034fc67b | mask-free-video-instance-segmentation | 2303.15904 | null | https://arxiv.org/abs/2303.15904v1 | https://arxiv.org/pdf/2303.15904v1.pdf | Mask-Free Video Instance Segmentation | The recent advancement in Video Instance Segmentation (VIS) has largely been driven by the use of deeper and increasingly data-hungry transformer-based models. However, video masks are tedious and expensive to annotate, limiting the scale and diversity of existing VIS datasets. In this work, we aim to remove the mask-annotation requirement. We propose MaskFreeVIS, achieving highly competitive VIS performance, while only using bounding box annotations for the object state. We leverage the rich temporal mask consistency constraints in videos by introducing the Temporal KNN-patch Loss (TK-Loss), providing strong mask supervision without any labels. Our TK-Loss finds one-to-many matches across frames, through an efficient patch-matching step followed by a K-nearest neighbor selection. A consistency loss is then enforced on the found matches. Our mask-free objective is simple to implement, has no trainable parameters, is computationally efficient, yet outperforms baselines employing, e.g., state-of-the-art optical flow to enforce temporal mask consistency. We validate MaskFreeVIS on the YouTube-VIS 2019/2021, OVIS and BDD100K MOTS benchmarks. The results clearly demonstrate the efficacy of our method by drastically narrowing the gap between fully and weakly-supervised VIS performance. Our code and trained models are available at https://github.com/SysCV/MaskFreeVis. | ['Fisher Yu', 'Chi-Keung Tang', 'Yu-Wing Tai', 'Henghui Ding', 'Martin Danelljan', 'Lei Ke'] | 2023-03-28 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Ke_Mask-Free_Video_Instance_Segmentation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Ke_Mask-Free_Video_Instance_Segmentation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-instance-segmentation', 'patch-matching'] | ['computer-vision', 'computer-vision'] | [ 4.04006764e-02 -3.08247030e-01 -4.70787585e-01 -2.51492649e-01
-7.75896370e-01 -6.65779173e-01 5.13040900e-01 -2.08904997e-01
-5.88021040e-01 5.36379755e-01 6.83860481e-02 -1.64419532e-01
1.81419432e-01 -2.24950179e-01 -8.28547716e-01 -5.62734842e-01
2.88882852e-02 1.57758132e-01 7.29276538e-01 2.26588026e-02
1.09383665e-01 2.19330370e-01 -1.55296099e+00 2.74637312e-01
7.14673758e-01 1.19386148e+00 2.60632902e-01 5.69803953e-01
2.08667621e-01 8.60624373e-01 -8.65778029e-02 -5.07798553e-01
6.37171865e-01 -4.39791083e-01 -9.38092232e-01 2.28647068e-01
9.49585259e-01 -5.30256927e-01 -6.85396492e-01 7.83160090e-01
3.38603973e-01 3.50022465e-01 2.23190889e-01 -1.36112714e+00
-3.40607971e-01 2.09234446e-01 -7.04563618e-01 5.23548841e-01
2.44370520e-01 5.30365527e-01 1.09058559e+00 -9.21461761e-01
8.50120366e-01 9.19706762e-01 7.36769199e-01 5.87672710e-01
-1.49273300e+00 -5.48930287e-01 2.82883584e-01 3.53066593e-01
-1.41252005e+00 -6.70072079e-01 4.75043923e-01 -6.83809936e-01
9.45052862e-01 1.93785667e-01 7.65423059e-01 1.06917548e+00
-1.41707480e-01 8.50459993e-01 1.02856708e+00 -1.11849993e-01
4.09985036e-02 -1.53518066e-01 -2.57878825e-02 1.01841688e+00
-8.19312856e-02 2.21139684e-01 -6.89445734e-01 1.56650677e-01
8.93902123e-01 5.98442554e-02 -5.14634669e-01 -5.32175064e-01
-1.18693626e+00 5.39174259e-01 4.25730854e-01 -1.88943148e-02
-2.14067400e-01 4.31054592e-01 5.48091173e-01 1.63058832e-01
4.58908349e-01 1.21691495e-01 -4.92334723e-01 -4.69138473e-01
-1.28068733e+00 2.43246764e-01 4.98229921e-01 8.93565834e-01
8.75997007e-01 -1.12416707e-01 -4.43338126e-01 7.28786528e-01
9.69923735e-02 3.23484361e-01 1.52443707e-01 -1.48418558e+00
3.63572717e-01 2.81267613e-01 5.37768821e-04 -8.26205909e-01
-9.55950022e-02 -2.96639770e-01 -3.60032678e-01 1.87101305e-01
4.63577628e-01 1.36150658e-01 -1.23489571e+00 1.68056929e+00
6.38483703e-01 8.23929071e-01 -4.46018338e-01 1.19289315e+00
7.09618151e-01 4.51578975e-01 7.25248754e-02 7.79922754e-02
1.13082647e+00 -1.42149782e+00 -5.12198687e-01 -1.07846469e-01
6.38888776e-01 -6.90589607e-01 1.24706244e+00 3.34354490e-01
-1.23136377e+00 -5.02114415e-01 -7.65318811e-01 -2.93132603e-01
9.92127955e-02 -7.70494714e-02 6.59281611e-01 3.89151931e-01
-1.01799440e+00 6.96637928e-01 -1.31510592e+00 -6.02154881e-02
8.48554075e-01 4.32029992e-01 -5.49082637e-01 -2.59692311e-01
-7.35322833e-01 5.04067302e-01 2.29430571e-01 3.46772484e-02
-1.04026353e+00 -1.11858428e+00 -1.04643905e+00 -1.50392175e-01
7.66038656e-01 -5.89138329e-01 1.23963654e+00 -1.00613999e+00
-1.44031739e+00 9.31546509e-01 -4.00409102e-01 -6.33009672e-01
9.09772038e-01 -2.96713799e-01 -1.00387938e-01 5.13912499e-01
2.46910736e-01 1.03059900e+00 8.41695070e-01 -1.15524137e+00
-6.37384713e-01 2.30974734e-01 1.51478514e-01 4.61367751e-03
-2.01474532e-01 6.08938336e-02 -1.14842629e+00 -6.56998694e-01
-1.81728989e-01 -1.00420916e+00 -1.50597498e-01 4.03214127e-01
-3.14207405e-01 -4.88084555e-02 9.30666089e-01 -7.36908019e-01
1.38988507e+00 -2.27664351e+00 3.84781510e-02 -4.50016633e-02
3.11295390e-01 4.82043862e-01 -2.68363923e-01 9.28021371e-02
2.40790784e-01 -8.53524059e-02 -4.56150740e-01 -7.43822217e-01
-9.54737142e-02 2.84956992e-01 -1.03090003e-01 6.29836321e-01
3.02324891e-01 1.02571654e+00 -9.86841440e-01 -6.86054647e-01
5.60631454e-01 5.72882950e-01 -9.96068299e-01 1.33725107e-01
-2.96921313e-01 5.45589685e-01 -8.63211676e-02 7.26531744e-01
5.35497248e-01 -5.32712162e-01 -1.26831606e-01 -4.55945969e-01
-7.47581869e-02 4.03041571e-01 -1.11299801e+00 2.10256910e+00
-2.56264389e-01 7.87618399e-01 1.89052254e-01 -9.66477096e-01
3.58589768e-01 1.58012450e-01 7.91127563e-01 -7.39833176e-01
1.89748645e-01 2.25943089e-01 -4.57813665e-02 -5.20330667e-01
3.29783499e-01 2.08709061e-01 2.85235405e-01 1.73593342e-01
1.79835424e-01 2.52700821e-02 6.04305685e-01 2.83282071e-01
1.20731843e+00 6.38982534e-01 -1.34413540e-01 -2.06183851e-01
3.96895409e-01 3.97940837e-02 8.60728383e-01 5.84893465e-01
-5.20424068e-01 9.52556908e-01 3.65079045e-01 -3.38573515e-01
-8.50822151e-01 -9.26067770e-01 -2.36456081e-01 8.31872702e-01
5.16927421e-01 -5.71060479e-01 -7.56222188e-01 -8.77581954e-01
1.95611835e-01 1.82904795e-01 -6.08016551e-01 4.27451394e-02
-8.12625825e-01 -2.25638285e-01 4.86971021e-01 5.45629919e-01
5.13553500e-01 -9.66711462e-01 -5.84092557e-01 3.33253533e-01
-2.99224913e-01 -1.64362001e+00 -9.48604584e-01 -7.11750984e-02
-7.67592549e-01 -1.18166387e+00 -6.07970536e-01 -6.34733498e-01
6.07728243e-01 3.13183993e-01 1.19625092e+00 2.74337888e-01
-5.02991199e-01 3.23887318e-01 -2.96167582e-01 6.11278079e-02
1.13604320e-02 -2.32828985e-04 -1.86333969e-01 7.64776096e-02
1.74462155e-01 -4.14997935e-01 -1.04123497e+00 6.04746103e-01
-8.44397426e-01 1.45056382e-01 1.76444128e-01 8.11052024e-01
8.21906984e-01 -4.24972206e-01 1.71151429e-01 -7.15160251e-01
-1.29394308e-01 -2.37686902e-01 -7.09618330e-01 -9.86771286e-02
-5.57528019e-01 -1.49527073e-01 3.44620436e-01 -4.97693598e-01
-7.48963058e-01 1.58572093e-01 -5.92643693e-02 -9.58995402e-01
2.91016009e-02 2.11970568e-01 1.97469398e-01 -3.21109653e-01
4.30929452e-01 1.86541751e-02 5.44127002e-02 -4.02661681e-01
2.80716896e-01 2.80261159e-01 8.50160003e-01 -6.44780397e-01
7.34036326e-01 8.06866765e-01 -1.38894394e-01 -5.78085184e-01
-9.47593451e-01 -7.30376244e-01 -5.10514975e-01 -3.19558412e-01
8.88876379e-01 -1.01210248e+00 -6.38895929e-01 4.22048151e-01
-8.70298147e-01 -8.20409000e-01 -3.11794341e-01 5.06930888e-01
-5.36008835e-01 5.26723802e-01 -8.61691356e-01 -3.54459465e-01
-2.69463211e-01 -1.29358029e+00 1.04537773e+00 3.85893770e-02
-2.39165857e-01 -7.79914558e-01 -5.27616926e-02 7.16070890e-01
2.54322648e-01 2.71960676e-01 2.62881160e-01 -1.77870929e-01
-9.13031697e-01 3.63057703e-02 -3.97078574e-01 5.60476243e-01
6.90958574e-02 1.48628399e-01 -9.13835526e-01 -3.62559646e-01
-5.03312945e-01 -4.10040289e-01 1.19078326e+00 5.45691788e-01
1.37413931e+00 -8.29254240e-02 -2.21359015e-01 1.03570795e+00
1.35570121e+00 -4.19627354e-02 5.74907005e-01 3.99198115e-01
1.08409441e+00 3.49418312e-01 7.14238524e-01 2.85288215e-01
4.81893808e-01 8.46186936e-01 4.61639613e-01 -2.67674863e-01
-5.87369263e-01 -1.96562022e-01 2.63046473e-01 5.45610845e-01
-1.21152349e-01 -3.55210036e-01 -8.20478022e-01 8.29479158e-01
-2.00380611e+00 -8.60012293e-01 -3.80478315e-02 2.12698746e+00
9.27647352e-01 2.47454166e-01 2.54322827e-01 2.39869859e-02
5.64656913e-01 3.18918109e-01 -5.36116600e-01 1.88953876e-02
-1.34147378e-03 3.16640168e-01 5.26831806e-01 7.32310534e-01
-1.33032203e+00 1.17461634e+00 5.25060797e+00 7.80234575e-01
-1.24683702e+00 2.59703904e-01 6.81959033e-01 -5.78070402e-01
-1.18922800e-01 1.28955454e-01 -5.95818639e-01 6.98746443e-01
5.55949509e-01 2.80050933e-01 5.21831512e-01 5.18820763e-01
4.46919560e-01 -1.57555282e-01 -1.24185145e+00 1.14602709e+00
-6.52055889e-02 -1.72648847e+00 -2.66627103e-01 -3.98557633e-02
9.01647270e-01 4.07750130e-01 -1.21116331e-02 2.16914807e-02
3.85247916e-02 -8.58391404e-01 1.04802704e+00 1.59665167e-01
9.32205319e-01 -4.31188256e-01 4.25081015e-01 -1.25849649e-01
-1.38493836e+00 1.03500262e-01 -2.90737022e-02 6.66639954e-02
3.90377194e-01 4.40903515e-01 -2.42288962e-01 3.73242319e-01
1.06589675e+00 1.12969708e+00 -3.07027072e-01 1.28704000e+00
-1.41059592e-01 7.29773939e-01 -4.90540296e-01 4.44672912e-01
4.82074380e-01 -1.78947300e-01 4.42758441e-01 1.37292635e+00
1.35782361e-02 -2.54159365e-02 4.78849083e-01 8.71700287e-01
-2.76708722e-01 -1.58237338e-01 -1.22695580e-01 9.88484323e-02
4.23406452e-01 1.22377586e+00 -7.73369133e-01 -3.73290092e-01
-6.10118032e-01 1.11694658e+00 2.27810696e-01 3.59355181e-01
-1.11014926e+00 -2.76015531e-02 9.01583791e-01 3.64225239e-01
7.35358655e-01 -2.07057774e-01 -2.37635195e-01 -1.25383234e+00
3.07665467e-01 -8.29021215e-01 3.62739235e-01 -4.89103943e-01
-1.15150404e+00 4.27472979e-01 1.75760556e-02 -1.21715057e+00
3.88661809e-02 -4.11073446e-01 -4.71273959e-01 5.17128468e-01
-1.80811739e+00 -1.01946270e+00 -4.49606270e-01 7.09434688e-01
6.21273339e-01 2.77611494e-01 2.27236062e-01 8.34844053e-01
-8.22872877e-01 7.28637397e-01 -2.25518063e-01 2.90657312e-01
7.97896683e-01 -1.01549017e+00 4.26845223e-01 1.08063126e+00
2.28448868e-01 3.37754399e-01 5.27103961e-01 -4.94639724e-01
-1.23010254e+00 -1.14352012e+00 6.29204512e-01 -4.20722812e-01
6.86281085e-01 -5.08135319e-01 -9.37722921e-01 6.30128622e-01
3.38246152e-02 7.97412395e-01 2.08374456e-01 -2.99840719e-01
-5.26538074e-01 -7.83227012e-02 -1.00794971e+00 5.45242071e-01
1.46502542e+00 -4.33018744e-01 -1.74457040e-02 2.80990452e-01
8.14635515e-01 -6.48162663e-01 -8.23284030e-01 4.79064822e-01
5.60901403e-01 -1.16383088e+00 1.05137551e+00 -4.53357756e-01
4.59303737e-01 -5.16339362e-01 3.62841412e-02 -7.66066134e-01
-2.12819919e-01 -9.18891609e-01 -4.39556599e-01 1.17508602e+00
3.98582578e-01 -5.05440891e-01 1.03971815e+00 7.49612033e-01
-3.77833635e-01 -1.04256105e+00 -1.03197777e+00 -9.15310979e-01
-2.60452837e-01 -5.33719480e-01 3.07997257e-01 9.47720945e-01
-3.63206238e-01 -1.35815620e-01 -3.99049431e-01 1.90756824e-02
6.81936026e-01 5.69405481e-02 8.43326271e-01 -8.39926660e-01
-5.92438102e-01 -5.75828791e-01 -4.47464883e-01 -1.34644282e+00
1.72344789e-01 -8.12174797e-01 1.04761784e-04 -1.39844930e+00
9.93223712e-02 -7.11513340e-01 -3.39486867e-01 8.35696459e-01
-3.15011978e-01 7.56795168e-01 3.26077104e-01 3.27242911e-01
-7.93204010e-01 5.33556283e-01 1.36087775e+00 -5.62580377e-02
-2.48282135e-01 -3.06837231e-01 -2.52722621e-01 6.88281298e-01
7.29261279e-01 -5.62230170e-01 -3.89902949e-01 -6.02409601e-01
-1.84194833e-01 -2.30679750e-01 7.01885164e-01 -9.49999154e-01
1.13366298e-01 -1.28911376e-01 5.55550605e-02 -3.89300257e-01
4.61019725e-01 -7.54631281e-01 -2.29356103e-02 4.49717253e-01
-1.04164757e-01 3.80324312e-02 2.55861759e-01 5.01169264e-01
-2.33997181e-01 1.74899828e-02 9.18664277e-01 1.14937358e-01
-9.42723811e-01 7.68421948e-01 1.99579298e-01 5.61080217e-01
1.04313517e+00 -4.41259474e-01 -1.83481023e-01 -9.83559415e-02
-4.17705595e-01 3.40041518e-01 6.25088990e-01 4.49049950e-01
4.55792308e-01 -1.11685777e+00 -6.25018418e-01 1.90973446e-01
1.29494742e-01 2.38136932e-01 2.99511284e-01 1.26367652e+00
-7.83679962e-01 1.23282872e-01 5.17852642e-02 -9.45239604e-01
-1.26583624e+00 3.79943132e-01 4.13044572e-01 -7.46885985e-02
-1.04355037e+00 1.10431445e+00 2.99168050e-01 -1.26115024e-01
3.70749146e-01 -3.45169961e-01 3.55399936e-01 -2.89110810e-01
3.48415911e-01 5.12039483e-01 -7.84152839e-03 -6.30391240e-01
-4.92122382e-01 6.28130138e-01 -1.12944536e-01 -1.17988139e-01
1.11614835e+00 -1.20876923e-01 1.73890799e-01 5.85459992e-02
1.32524467e+00 -1.25894710e-01 -2.00462461e+00 -3.67236525e-01
-2.02559650e-01 -8.38521004e-01 1.17879599e-01 -6.17538691e-01
-1.43940592e+00 6.93691015e-01 4.13113028e-01 -1.75088242e-01
1.00171149e+00 8.83405469e-03 1.29287314e+00 -1.31943196e-01
2.22265661e-01 -9.96357799e-01 9.95343700e-02 3.03002447e-01
5.04641533e-01 -1.32575715e+00 -7.16813654e-02 -6.82908416e-01
-5.74147940e-01 6.54023588e-01 7.48025894e-01 -2.89735436e-01
5.48480034e-01 4.47115481e-01 2.14441359e-01 -1.33689955e-01
-7.01903582e-01 -3.03048253e-01 5.19803464e-01 3.85623455e-01
2.65627861e-01 -3.22510272e-01 -2.15014324e-01 2.21041180e-02
3.60161737e-02 1.73386335e-01 1.85217664e-01 9.09356117e-01
-8.15033261e-03 -1.06817579e+00 1.53193460e-03 4.27180767e-01
-6.15849078e-01 -1.88895270e-01 8.46214890e-02 7.20722735e-01
1.17682084e-01 8.85503352e-01 1.19844571e-01 -2.45582506e-01
2.40775481e-01 -2.19195545e-01 6.14715695e-01 -3.53963643e-01
-5.74082136e-01 1.70629770e-01 4.36544083e-02 -1.13081360e+00
-6.03286326e-01 -7.44089901e-01 -1.39210141e+00 -4.32882935e-01
-2.45436832e-01 -1.49065435e-01 3.65064263e-01 8.77323806e-01
6.68689489e-01 4.04643327e-01 3.69584918e-01 -1.02505684e+00
-2.50637054e-01 -4.91266847e-01 -1.08876213e-01 7.62557745e-01
5.44308841e-01 -7.95388520e-01 -3.62683654e-01 3.04173917e-01] | [9.160717010498047, 0.0061780186370015144] |
9d27c44e-f735-42a7-8406-35598aeabc29 | towards-surgical-context-inference-and | 2302.14237 | null | https://arxiv.org/abs/2302.14237v2 | https://arxiv.org/pdf/2302.14237v2.pdf | Towards Surgical Context Inference and Translation to Gestures | Manual labeling of gestures in robot-assisted surgery is labor intensive, prone to errors, and requires expertise or training. We propose a method for automated and explainable generation of gesture transcripts that leverages the abundance of data for image segmentation. Surgical context is detected using segmentation masks by examining the distances and intersections between the tools and objects. Next, context labels are translated into gesture transcripts using knowledge-based Finite State Machine (FSM) and data-driven Long Short Term Memory (LSTM) models. We evaluate the performance of each stage of our method by comparing the results with the ground truth segmentation masks, the consensus context labels, and the gesture labels in the JIGSAWS dataset. Our results show that our segmentation models achieve state-of-the-art performance in recognizing needle and thread in Suturing and we can automatically detect important surgical states with high agreement with crowd-sourced labels (e.g., contact between graspers and objects in Suturing). We also find that the FSM models are more robust to poor segmentation and labeling performance than LSTMs. Our proposed method can significantly shorten the gesture labeling process (~2.8 times). | ['Homa Alemzadeh', 'Ian Reyes', 'Zongyu Li', 'Kay Hutchinson'] | 2023-02-28 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 6.49089873e-01 6.28380775e-01 -3.18533957e-01 -5.86274326e-01
-1.26391840e+00 -9.07732785e-01 2.32972294e-01 -1.41140511e-02
-3.93706828e-01 1.64239198e-01 3.92728060e-01 -3.66390586e-01
-2.57941168e-02 -5.00491858e-02 -6.55411661e-01 -5.20927310e-01
5.68012036e-02 7.36821115e-01 -1.17387727e-01 5.38499020e-02
3.87996316e-01 3.89291227e-01 -9.14870620e-01 6.00454807e-01
4.31579322e-01 8.48084688e-01 4.29028869e-01 9.64405358e-01
-4.75302994e-01 7.47228920e-01 -6.39506519e-01 -1.31987957e-02
3.88868421e-01 -6.67997539e-01 -1.42341495e+00 1.67081013e-01
3.34316492e-01 -5.38499951e-01 -7.18823075e-02 7.99109101e-01
3.30123812e-01 -8.72680768e-02 4.31727916e-01 -6.54272914e-01
-2.22863391e-01 7.57372856e-01 -4.78902102e-01 -2.62169868e-01
4.30446267e-01 3.11946511e-01 5.67205369e-01 -6.73557043e-01
1.14471173e+00 1.18079185e+00 5.30353189e-01 9.67222393e-01
-7.84866214e-01 -6.69087112e-01 1.43996105e-01 -4.20217127e-01
-9.37258661e-01 -3.91911209e-01 4.35745895e-01 -7.75590479e-01
7.51233280e-01 1.19764566e-01 7.45882332e-01 1.16043782e+00
5.26884198e-01 9.21517909e-01 9.60645676e-01 -5.76485217e-01
1.03826970e-01 -5.21218002e-01 1.44399345e-01 1.42552602e+00
-9.87449810e-02 9.83357579e-02 -6.33961380e-01 -1.95525680e-03
1.36513865e+00 1.90757111e-01 4.78032418e-02 -1.42713889e-01
-1.90333784e+00 4.08154339e-01 5.63683748e-01 4.25132632e-01
-4.28439379e-01 4.10565674e-01 1.94363892e-01 -1.11662827e-01
4.68399413e-02 7.90347815e-01 -4.09835130e-01 -3.94087851e-01
-1.01532722e+00 -4.61802065e-01 8.37779045e-01 1.38333070e+00
2.91261315e-01 -4.84145939e-01 -3.20052564e-01 4.06169057e-01
1.80095583e-01 2.37535998e-01 4.77387220e-01 -1.36094046e+00
2.20969498e-01 6.85412884e-01 1.63823560e-01 -6.81670129e-01
-8.31892371e-01 -1.12610068e-02 -4.10740942e-01 6.80755675e-02
4.53788400e-01 -2.68151015e-01 -1.78432155e+00 1.24465275e+00
1.75439060e-01 1.82096839e-01 -1.44450724e-01 1.01816976e+00
9.15053606e-01 -5.29964603e-02 3.09654564e-01 -1.55841962e-01
1.10314488e+00 -1.26652861e+00 -9.83671308e-01 -3.97579491e-01
1.06420612e+00 -8.90283883e-01 8.52943182e-01 1.41337693e-01
-9.61953580e-01 -2.68261582e-01 -7.14850307e-01 -1.87767953e-01
9.81903821e-02 4.58632320e-01 7.94833541e-01 1.76833823e-01
-1.06477368e+00 6.61162794e-01 -1.59950936e+00 -5.17957747e-01
5.94391763e-01 7.58200645e-01 -2.81139225e-01 9.46333930e-02
-1.69066116e-01 1.05100381e+00 1.62472412e-01 4.60101813e-01
-1.07210493e+00 -1.15563907e-01 -1.14418602e+00 -5.96987307e-01
4.06780094e-01 -4.28312540e-01 1.69530332e+00 -9.20496881e-01
-1.70652223e+00 1.32960188e+00 -4.79849547e-01 2.86940448e-02
4.38509852e-01 -2.71478325e-01 2.29565859e-01 3.55782270e-01
2.16184109e-02 1.20126545e+00 6.29899740e-01 -1.35215580e+00
-4.35552061e-01 -3.45466793e-01 -9.00997892e-02 1.52306542e-01
2.70286798e-01 4.54169586e-02 -4.23617661e-01 -4.07545656e-01
7.94568181e-01 -1.48371410e+00 -5.26332200e-01 1.54430136e-01
-8.28803897e-01 -1.61979869e-02 4.36817080e-01 -9.49543774e-01
7.23001182e-01 -2.11845040e+00 1.64299712e-01 1.77014723e-01
3.09145004e-01 4.87568416e-03 -3.00488770e-01 6.30625859e-02
3.77026141e-01 2.36934051e-01 -2.66917408e-01 -4.96663958e-01
-2.80220568e-01 7.60652959e-01 -1.32477432e-01 4.75654691e-01
-1.46990597e-01 1.37642550e+00 -1.22253740e+00 -9.55643833e-01
4.38791513e-01 9.51056629e-02 -2.25737080e-01 2.78600574e-01
-3.02354157e-01 1.11594844e+00 -5.00314534e-01 1.15103221e+00
-1.14697620e-01 -4.25967276e-01 5.11694014e-01 3.81275639e-02
1.11825205e-01 2.04750568e-01 -4.94268328e-01 2.68350506e+00
-5.11995018e-01 5.04104614e-01 2.23422766e-01 -2.90559411e-01
7.68943012e-01 4.86883134e-01 6.73828185e-01 2.54509412e-02
4.62790728e-01 4.57890987e-01 1.77980006e-01 -9.00673151e-01
1.77330598e-01 -2.54376799e-01 -2.43651003e-01 6.66665971e-01
2.59981036e-01 -5.47649562e-01 -3.13736558e-01 -6.47469833e-02
1.08745193e+00 4.65143710e-01 2.62228280e-01 -9.45584550e-02
-4.10462588e-01 5.58624327e-01 4.07486469e-01 8.89981925e-01
-4.62664992e-01 7.32801259e-01 1.85882822e-01 -6.60292387e-01
-6.03489995e-01 -7.12616742e-01 4.20269459e-01 9.78315592e-01
4.69628096e-01 -5.80632733e-03 -1.07438338e+00 -1.04042363e+00
-5.24625599e-01 5.94447136e-01 -8.57492447e-01 1.36834113e-02
-9.22389805e-01 -1.12110503e-01 4.90394264e-01 9.17673707e-01
-2.22550943e-01 -1.51326394e+00 -1.13262296e+00 1.27308503e-01
-3.71667683e-01 -1.23926342e+00 -5.67187250e-01 3.11635315e-01
-1.20799446e+00 -1.26565230e+00 -6.13742709e-01 -1.38340795e+00
1.25788748e+00 -1.86719652e-02 8.44685853e-01 3.58683169e-01
-5.85636914e-01 4.31241095e-01 -4.43998605e-01 -6.02968633e-01
-6.81015790e-01 6.77620694e-02 -1.78936288e-01 -6.36707425e-01
3.23906660e-01 8.08774829e-02 -6.45833373e-01 4.95488137e-01
-7.00445592e-01 5.06613910e-01 8.93886924e-01 7.71879792e-01
7.88376391e-01 -7.89410770e-01 -9.21684429e-02 -7.34598517e-01
2.87100792e-01 1.67289507e-02 -2.01149881e-01 3.22818041e-01
-2.73988158e-01 9.36018825e-02 -9.10381973e-02 -4.80534405e-01
-8.19287896e-01 7.30262280e-01 2.07083315e-01 -6.08450174e-01
-5.78878582e-01 3.13430965e-01 5.44493198e-01 -2.67378867e-01
5.43153346e-01 -2.43461967e-01 2.98455834e-01 -2.82204449e-01
3.33746761e-01 6.21803999e-01 7.38353431e-01 -5.40603757e-01
3.05263162e-01 6.30545616e-01 -2.07083032e-01 -2.69445300e-01
-1.14395630e+00 -4.79850471e-01 -1.17407334e+00 -3.51762027e-01
1.17483711e+00 -3.52280676e-01 -6.19978189e-01 3.41537356e-01
-1.52488744e+00 -9.45174634e-01 -2.50112027e-01 7.25963354e-01
-7.35655963e-01 -1.80639338e-03 -1.24033570e+00 -6.46394312e-01
-4.55128312e-01 -1.61882865e+00 1.68563378e+00 4.91476543e-02
-8.20287347e-01 -7.03999102e-01 -2.34195456e-01 5.37049055e-01
-3.71986032e-02 7.02740610e-01 8.83489490e-01 -7.01487958e-01
-5.96405745e-01 -4.86905962e-01 6.75283596e-02 -1.45714477e-01
5.13364255e-01 -9.38044339e-02 -6.33532763e-01 -1.27444835e-02
-9.62348580e-02 -4.20372665e-01 4.89636332e-01 7.85955846e-01
1.17940235e+00 -2.22031891e-01 -8.42439353e-01 4.74084944e-01
8.24263573e-01 3.50235939e-01 2.92430609e-01 6.06460907e-02
8.15876424e-01 9.25113559e-01 9.51868892e-01 -6.13470562e-02
1.14951573e-01 2.71651924e-01 3.52686048e-01 -2.99870402e-01
-1.89975768e-01 -2.75336504e-01 4.40541049e-03 1.02922094e+00
-1.65291086e-01 1.70706168e-01 -1.37249196e+00 6.39596164e-01
-1.89282238e+00 -2.78425574e-01 -1.06032491e-01 1.68394959e+00
9.01458859e-01 -8.57790858e-02 -3.95124167e-01 -4.36568975e-01
5.83076417e-01 -1.77095592e-01 -5.94191492e-01 -2.57812977e-01
6.32448256e-01 3.00335824e-01 6.59187078e-01 6.97629929e-01
-1.01881313e+00 1.36655021e+00 7.24179316e+00 3.62631351e-01
-1.09280944e+00 1.90476298e-01 5.61575532e-01 -4.08732384e-01
2.05351651e-01 -3.02338213e-01 -2.17130899e-01 -1.30499648e-02
5.76220989e-01 6.14320457e-01 3.45003814e-01 6.99636877e-01
2.83201247e-01 -1.56137004e-01 -1.50303614e+00 1.01564932e+00
2.06965014e-01 -1.44806361e+00 -2.19154716e-01 -2.17843372e-02
9.81078565e-01 6.03752360e-02 -2.10692108e-01 -4.24262844e-02
5.43415248e-01 -1.30826759e+00 7.37141848e-01 4.42487895e-01
1.22867286e+00 -4.30240929e-02 9.04901683e-01 3.11085612e-01
-7.78929889e-01 -8.05294290e-02 2.29483977e-01 5.77375665e-02
3.81785423e-01 -6.39939457e-02 -1.44132543e+00 1.04185589e-01
3.82711381e-01 3.68822545e-01 -3.32679749e-02 6.20573640e-01
-5.54944277e-01 3.72047186e-01 -2.11143196e-01 -1.42393373e-02
5.77645719e-01 1.45494297e-01 4.29927021e-01 1.14143586e+00
1.43297568e-01 4.84656900e-01 5.43457806e-01 6.83786869e-01
-9.27353054e-02 -2.15372264e-01 -4.71060932e-01 -3.15569073e-01
2.11831018e-01 1.20818174e+00 -1.41338527e+00 -3.46351147e-01
8.50461274e-02 1.12810302e+00 -1.09457381e-01 3.49105358e-01
-4.86049294e-01 2.42179651e-02 3.38660538e-01 -2.37347856e-01
-1.64821431e-01 -5.43475926e-01 -9.57267821e-01 -5.62655926e-01
4.99349758e-02 -7.79984057e-01 2.20429271e-01 -7.63620317e-01
-6.52471542e-01 6.74645066e-01 -2.94806302e-01 -1.25019574e+00
-4.83376950e-01 -7.96206772e-01 -2.13280961e-01 5.18138707e-01
-9.88798499e-01 -1.44199955e+00 -6.97894633e-01 2.27018431e-01
9.06026483e-01 4.38654870e-01 1.23685908e+00 -3.20630580e-01
-2.44076401e-01 2.51276970e-01 -4.57969010e-01 7.53523469e-01
6.20442033e-01 -1.04289937e+00 4.34077650e-01 5.59303641e-01
2.67501742e-01 9.19972003e-01 4.77359742e-01 -1.07050371e+00
-1.39668596e+00 -9.52172399e-01 5.11333466e-01 -9.42184448e-01
3.83415550e-01 -2.75907874e-01 -2.76814640e-01 1.30401814e+00
-1.42120108e-01 -6.11406229e-02 7.53566921e-01 -6.22594263e-03
1.04189962e-01 7.70797014e-01 -1.19072986e+00 5.45418322e-01
1.45142376e+00 -4.89545524e-01 -7.61890948e-01 6.87896430e-01
8.57202470e-01 -1.34155416e+00 -6.58574522e-01 5.11306047e-01
8.58471930e-01 -3.43690962e-01 4.58873808e-01 -7.10509360e-01
7.34151423e-01 -1.43121171e-03 1.64951324e-01 -1.11524308e+00
1.07558787e-01 -8.14456046e-01 1.59359008e-01 2.69761831e-01
6.23970747e-01 -4.23149914e-02 9.95002031e-01 1.15417862e+00
-6.10650480e-01 -1.00513160e+00 -8.70088995e-01 -4.93517429e-01
-2.08017722e-01 -3.42600971e-01 2.08055705e-01 1.08791208e+00
3.99850816e-01 -2.88058698e-01 4.93885390e-02 8.54242593e-02
2.55888224e-01 4.37239289e-01 4.57799196e-01 -9.29420292e-01
-1.30917076e-02 -4.20958817e-01 -2.28391603e-01 -9.47694421e-01
2.13273212e-01 -9.44006026e-01 1.00537229e+00 -2.18064141e+00
1.53108716e-01 -4.98582661e-01 -1.24771357e-01 1.12391567e+00
9.05010756e-03 1.77648276e-01 1.81303442e-01 4.91262823e-01
-6.69343770e-01 -2.07026109e-01 1.57165146e+00 -9.58701372e-02
-5.91472208e-01 -3.06116313e-01 -3.55366766e-01 1.15581071e+00
6.24880612e-01 -6.60998285e-01 2.05450997e-01 -9.11548674e-01
-3.00177515e-01 3.83249879e-01 2.82618612e-01 -6.27645910e-01
4.56638992e-01 -2.64591157e-01 2.13608876e-01 -3.70055377e-01
2.38608927e-01 -7.31570184e-01 -7.82473832e-02 8.46540689e-01
-7.92457640e-01 -2.86156595e-01 1.47367269e-01 4.57894057e-01
4.58159894e-02 -1.36297300e-01 3.38151038e-01 -7.01839685e-01
-6.50287867e-01 1.02966018e-01 -4.11868989e-01 -3.26828569e-01
8.91427457e-01 -3.98538738e-01 -4.56349999e-02 -2.46879101e-01
-1.05628669e+00 2.35522807e-01 4.98021990e-01 6.00636303e-01
7.76118219e-01 -7.67262399e-01 -4.20682728e-01 -3.13453563e-02
-3.54203209e-03 6.88072145e-01 -4.00796272e-02 1.19955087e+00
-8.20041537e-01 4.54931289e-01 6.26988038e-02 -8.78578782e-01
-1.23953938e+00 2.16986418e-01 3.44268471e-01 -9.77529734e-02
-8.92382085e-01 1.29026103e+00 5.48006669e-02 -6.37738824e-01
5.17542779e-01 -9.98705387e-01 2.82195330e-01 -3.96222919e-01
3.09239089e-01 6.58135489e-02 -8.83744508e-02 -3.45359236e-01
-4.04846191e-01 6.74604058e-01 8.47294554e-02 -2.37093866e-01
1.11150944e+00 1.47614852e-01 -5.04406132e-02 5.84494591e-01
8.45653474e-01 -2.77162790e-01 -1.29609001e+00 2.60738167e-03
2.47763813e-01 -1.77436799e-01 -2.77698860e-02 -1.33239365e+00
-8.73058259e-01 8.03877950e-01 5.15952706e-01 -4.14797097e-01
6.52911544e-01 4.04428631e-01 1.29541886e+00 4.10334975e-01
8.98887336e-01 -1.07934976e+00 2.36581311e-01 3.79295021e-01
7.63862431e-01 -1.30512726e+00 -4.29174095e-01 -6.00298226e-01
-8.51042807e-01 1.09662271e+00 5.72456360e-01 1.34764820e-01
1.47644430e-01 6.61125124e-01 8.91254485e-01 -6.85184062e-01
-7.32218772e-02 6.99366489e-03 3.59372944e-01 2.74391532e-01
4.78817075e-01 4.46319252e-01 -2.61710398e-02 3.60000730e-01
-2.66960949e-01 1.53357863e-01 4.02611941e-01 1.42118001e+00
-2.57557929e-01 -6.40307248e-01 -1.92553267e-01 7.35091507e-01
-3.89395893e-01 -1.34623125e-01 -8.32982898e-01 5.31255603e-01
-2.83134449e-02 1.06824100e+00 3.31306048e-02 -3.83248091e-01
1.29490688e-01 3.06609333e-01 6.91548049e-01 -1.11721778e+00
-7.13820636e-01 3.00348550e-01 4.84740222e-03 -1.13971603e+00
-6.40695333e-01 -4.18286145e-01 -2.00796175e+00 5.23240745e-01
-4.40742671e-01 -2.49651760e-01 9.94984090e-01 1.32426012e+00
4.13810670e-01 7.12693274e-01 -1.07828833e-01 -1.22028840e+00
-2.27968171e-01 -9.47691262e-01 -1.13198832e-01 4.11990434e-01
4.60894406e-01 -4.90150720e-01 3.95889990e-02 3.63755405e-01] | [14.087203979492188, -3.3244340419769287] |
49ff3543-03ca-47ea-8d90-ca21a5162043 | vartani-spellcheck-automatic-context | 2012.07652 | null | https://arxiv.org/abs/2012.07652v1 | https://arxiv.org/pdf/2012.07652v1.pdf | Vartani Spellcheck -- Automatic Context-Sensitive Spelling Correction of OCR-generated Hindi Text Using BERT and Levenshtein Distance | Traditional Optical Character Recognition (OCR) systems that generate text of highly inflectional Indic languages like Hindi tend to suffer from poor accuracy due to a wide alphabet set, compound characters and difficulty in segmenting characters in a word. Automatic spelling error detection and context-sensitive error correction can be used to improve accuracy by post-processing the text generated by these OCR systems. A majority of previously developed language models for error correction of Hindi spelling have been context-free. In this paper, we present Vartani Spellcheck - a context-sensitive approach for spelling correction of Hindi text using a state-of-the-art transformer - BERT in conjunction with the Levenshtein distance algorithm, popularly known as Edit Distance. We use a lookup dictionary and context-based named entity recognition (NER) for detection of possible spelling errors in the text. Our proposed technique has been tested on a large corpus of text generated by the widely used Tesseract OCR on the Hindi epic Ramayana. With an accuracy of 81%, the results show a significant improvement over some of the previously established context-sensitive error correction mechanisms for Hindi. We also explain how Vartani Spellcheck may be used for on-the-fly autocorrect suggestion during continuous typing in a text editor environment. | ['Abhijit Mustafi', 'Aditya Pal'] | 2020-12-14 | null | null | null | null | ['spelling-correction'] | ['natural-language-processing'] | [ 5.29537618e-01 -5.97425282e-01 3.54572833e-01 -4.00697291e-01
-7.88770676e-01 -8.52706075e-01 4.39094305e-01 7.09464133e-01
-8.85063171e-01 1.07730794e+00 -1.25389010e-01 -6.29511356e-01
-4.10017893e-02 -6.83284581e-01 -3.62138391e-01 -3.00720066e-01
2.43417129e-01 8.12212110e-01 4.45341289e-01 -6.69764042e-01
1.05032563e+00 5.93879759e-01 -1.33602035e+00 3.29458863e-01
1.34250283e+00 1.37843609e-01 4.88416642e-01 1.19168162e+00
-3.39391381e-01 6.20408833e-01 -1.28762376e+00 -4.42572951e-01
-5.86488284e-02 -5.25330842e-01 -8.74308169e-01 -4.41859245e-01
5.08329332e-01 -4.40718234e-02 -1.98763292e-02 1.04245484e+00
4.97853845e-01 -7.65354410e-02 6.67064667e-01 -6.82568252e-01
-8.35193932e-01 7.05859780e-01 -1.23189382e-01 4.82775360e-01
5.35068452e-01 -2.60920227e-01 6.51011765e-01 -9.25982177e-01
8.32174897e-01 8.95273507e-01 7.28204072e-01 4.61516410e-01
-9.53410149e-01 -5.22492886e-01 -6.16248369e-01 4.87090737e-01
-1.62519979e+00 3.02444659e-02 4.99618411e-01 -4.22286928e-01
1.36681294e+00 6.71115756e-01 3.74702215e-01 4.71632481e-01
4.69487160e-01 7.17026591e-01 1.48386514e+00 -1.17379034e+00
1.45442128e-01 3.30215067e-01 5.37088990e-01 4.58982497e-01
3.68267477e-01 -1.75332472e-01 -3.82127255e-01 5.05402684e-02
2.53190011e-01 -3.46507668e-01 4.39911596e-02 5.54969192e-01
-8.58426988e-01 8.65283847e-01 -3.62348706e-01 5.48230886e-01
-9.58883613e-02 -4.51526731e-01 7.00763464e-01 3.90796840e-01
2.82608926e-01 6.93059564e-01 -4.36778396e-01 -4.82206434e-01
-1.41681552e+00 3.67056370e-01 6.62724137e-01 1.16415989e+00
5.31539559e-01 7.07818791e-02 -2.38269135e-01 1.10720730e+00
2.01560110e-01 6.32808089e-01 8.64819705e-01 -1.10266306e-01
6.10112369e-01 5.84927678e-01 1.17218375e-01 -8.04941297e-01
-8.97424445e-02 1.89393297e-01 -3.09704304e-01 4.86500204e-01
6.16831183e-01 -6.16806522e-02 -1.09560692e+00 1.09021556e+00
8.42981339e-02 -6.65607333e-01 -2.06416519e-03 5.00899613e-01
9.27902386e-02 9.90095139e-01 -7.73172230e-02 -8.64393711e-02
1.16385233e+00 -6.06889307e-01 -8.70704353e-01 -1.01904638e-01
9.39593375e-01 -1.46793079e+00 9.63711858e-01 6.47377133e-01
-7.48102367e-01 -3.55301797e-01 -1.29044509e+00 -2.42095307e-01
-7.86642373e-01 3.21548730e-01 3.04592457e-02 1.17386222e+00
-6.39896393e-01 6.45814121e-01 -3.58351022e-01 -7.43985951e-01
-3.19525182e-01 1.53262362e-01 -2.76717484e-01 8.20477381e-02
-1.17279077e+00 1.05399323e+00 7.14846492e-01 5.98877259e-02
-2.49106675e-01 -4.87103730e-01 -6.73934281e-01 -2.82679498e-01
-8.74281023e-03 3.41660261e-01 1.08358908e+00 -7.67066956e-01
-1.39471781e+00 1.11854064e+00 -1.37184009e-01 -4.61086780e-01
5.96803904e-01 -4.02889103e-01 -9.42488134e-01 1.25847667e-01
6.78597167e-02 -1.17533050e-01 6.18887961e-01 -7.02282250e-01
-7.63937414e-01 -4.62180078e-01 -7.29554713e-01 -4.17388156e-02
-3.29957642e-02 7.92646587e-01 -1.47812113e-01 -1.13939321e+00
-2.24236935e-01 -1.05026412e+00 1.77787334e-01 -4.99008119e-01
-2.99467295e-01 -1.52657032e-01 9.74537432e-01 -1.46166325e+00
1.96333039e+00 -1.74929118e+00 -4.15588468e-01 4.18446839e-01
-6.12523615e-01 1.05326986e+00 7.39655495e-02 7.29849875e-01
-2.22645067e-02 9.78712365e-02 -5.38014829e-01 8.02142397e-02
-1.70871615e-01 2.35346988e-01 -1.55800313e-01 4.26578224e-01
1.60402507e-01 4.10023034e-01 -8.51892710e-01 -6.86568856e-01
2.99342036e-01 3.72539520e-01 -2.28438035e-01 1.78451389e-01
-8.72797915e-04 -1.66509151e-01 2.06432551e-01 5.58982849e-01
6.68759704e-01 7.72823811e-01 8.82540122e-02 3.59955668e-01
-6.60040677e-01 4.98913586e-01 -1.52318013e+00 1.46442521e+00
-2.84404159e-01 9.54617262e-01 -5.22472382e-01 -6.86175466e-01
1.38994443e+00 1.20884106e-02 -3.40624452e-01 -6.80978060e-01
-1.10151187e-01 7.21678615e-01 -6.67218119e-02 -4.77078885e-01
1.32875228e+00 2.32040465e-01 -9.15597603e-02 4.14087802e-01
-6.36968464e-02 -2.74674688e-02 7.01816142e-01 1.80165216e-01
6.39287472e-01 3.18453640e-01 5.57438433e-01 -2.95739174e-01
8.68274570e-01 3.37375343e-01 3.05094093e-01 1.13240492e+00
-1.71096772e-01 7.94161916e-01 -1.28621683e-01 -1.08921014e-01
-1.45830357e+00 -5.88621318e-01 -3.38400722e-01 9.69667077e-01
-4.63920295e-01 -5.40252924e-01 -9.19652164e-01 -4.86618102e-01
-3.17098379e-01 1.21048164e+00 -4.24549669e-01 1.45690754e-01
-1.07394123e+00 -7.58991659e-01 9.45617259e-01 3.68122011e-01
2.39828289e-01 -1.29861486e+00 -4.79564458e-01 6.95932746e-01
-2.88125835e-02 -6.11041903e-01 -4.93000776e-01 3.39199305e-01
-6.79485857e-01 -6.64370239e-01 -8.73302341e-01 -9.59390819e-01
4.57284838e-01 -1.01854883e-01 5.97949564e-01 2.53162205e-01
-6.59617364e-01 -1.54920980e-01 -7.75799334e-01 -6.13496840e-01
-1.09740305e+00 1.87563270e-01 -1.48277640e-01 -3.46171200e-01
7.54267752e-01 4.60287072e-02 -1.14244729e-01 2.54215777e-01
-1.01421607e+00 -4.49227333e-01 3.03584158e-01 6.10082209e-01
3.94362926e-01 -2.05728084e-01 1.43762499e-01 -1.28160715e+00
5.79952955e-01 -1.55268624e-01 -7.13287354e-01 6.44051909e-01
-9.33555603e-01 2.15269521e-01 8.42442572e-01 -2.57750541e-01
-1.21330774e+00 -2.92875171e-01 -4.72737342e-01 6.19399071e-01
-4.47490960e-01 4.26770359e-01 -2.23282706e-02 -2.14897886e-01
8.61121774e-01 5.62558711e-01 -3.07933927e-01 -7.21116602e-01
7.94976577e-02 1.45027781e+00 8.35746586e-01 -3.46992791e-01
5.75898290e-01 -1.79204255e-01 -3.27699333e-01 -1.29321420e+00
-1.65071130e-01 -6.92987978e-01 -7.75725842e-01 -1.20110303e-01
6.93300307e-01 -3.90324980e-01 -2.25142851e-01 1.09930217e+00
-1.17560303e+00 -2.25498959e-01 -1.61216445e-02 2.48678982e-01
-2.92916924e-01 9.32552397e-01 -8.10594916e-01 -1.12486315e+00
-6.82116807e-01 -6.80359244e-01 6.64289594e-01 4.33186531e-01
-4.44628626e-01 -8.18544030e-01 4.31486547e-01 2.36826658e-01
2.89539993e-01 -4.84088138e-02 1.05748427e+00 -9.75184262e-01
8.74551013e-02 -4.49485838e-01 -1.04583181e-01 4.43706334e-01
-8.48420486e-02 4.78590935e-01 -5.96500874e-01 -3.82221155e-02
-3.73092830e-01 1.36002719e-01 3.70886743e-01 -1.35964140e-01
6.57707155e-01 -2.82647699e-01 1.45620897e-01 3.77058864e-01
1.70414937e+00 7.36873567e-01 1.07096887e+00 9.42708254e-01
6.39390171e-01 2.99192011e-01 1.05949438e+00 6.94327831e-01
4.95813750e-02 6.44877732e-01 -2.89566934e-01 4.03834432e-01
-1.26410127e-01 -3.50178629e-01 3.99421066e-01 9.18718159e-01
-3.03844530e-02 -3.38514835e-01 -1.08129764e+00 7.28291273e-01
-1.42960835e+00 -1.11457539e+00 -8.31877649e-01 2.52713037e+00
1.26823699e+00 2.07858786e-01 2.27884844e-01 7.89513350e-01
1.03043079e+00 -3.02521080e-01 2.79683501e-01 -1.32722926e+00
-3.34862381e-01 5.75334013e-01 9.55851614e-01 9.38719392e-01
-8.23969960e-01 1.23880613e+00 5.86123800e+00 8.75921607e-01
-1.17343795e+00 -1.16571695e-01 -1.76132657e-02 3.57239038e-01
7.68706435e-03 5.96214533e-02 -1.32573569e+00 7.92182624e-01
1.23542833e+00 -1.21696508e-02 3.55188996e-01 7.45345116e-01
1.50385559e-01 -5.01568317e-01 -5.32311201e-01 7.96053648e-01
5.03707945e-01 -1.25577283e+00 -2.10857634e-02 -2.67483562e-01
6.28550351e-01 -1.44014150e-01 -4.80853856e-01 1.74422473e-01
3.07669304e-02 -5.66637933e-01 1.14568031e+00 4.07125801e-01
9.78336930e-01 -8.84267867e-01 6.54816747e-01 -4.69266474e-02
-6.89840913e-01 2.25513220e-01 -5.49723983e-01 -1.20915724e-02
8.42311084e-02 3.91539156e-01 -1.03441703e+00 2.71950394e-01
4.02230769e-01 3.47595252e-02 -1.00752985e+00 1.42099214e+00
-3.52750659e-01 1.00031090e+00 -4.12237018e-01 -6.15767121e-01
1.42443866e-01 -2.18447521e-01 7.14534044e-01 2.08286691e+00
4.96329635e-01 -1.03128456e-01 -4.55829978e-01 3.22938323e-01
5.38989663e-01 6.49206638e-01 -2.48596102e-01 -1.89917773e-01
5.53844750e-01 8.26688826e-01 -8.09450388e-01 -5.63382566e-01
-1.07150979e-01 1.68003845e+00 1.54613540e-01 4.52934317e-02
-6.03573143e-01 -1.48097289e+00 1.75559416e-01 -1.25504863e-02
4.52853292e-01 -4.85226035e-01 -5.79177678e-01 -9.06128168e-01
-2.03046985e-02 -1.01717091e+00 5.56507349e-01 -6.98790312e-01
-7.67130256e-01 6.26319051e-01 -2.85502952e-02 -8.11274886e-01
-1.67961508e-01 -1.04500997e+00 -5.64164817e-01 1.17749393e+00
-1.17683077e+00 -8.69066238e-01 3.82392295e-02 3.93237561e-01
8.16696227e-01 -2.14218795e-01 8.52732539e-01 4.58770961e-01
-5.29211521e-01 1.00701284e+00 6.70389891e-01 3.44359189e-01
1.14898705e+00 -1.60873890e+00 5.40804446e-01 1.38682652e+00
1.95201442e-01 7.01086104e-01 9.85034406e-01 -1.14762866e+00
-9.48318660e-01 -9.92168128e-01 1.94400144e+00 -3.46221209e-01
6.20961666e-01 -4.99633968e-01 -1.11361158e+00 5.14055789e-01
4.19287309e-02 -6.70776188e-01 8.35710764e-01 -2.91087538e-01
-2.32323602e-01 1.92925483e-02 -1.14426351e+00 5.48693538e-01
4.15836394e-01 -6.31097138e-01 -1.01339650e+00 2.83276051e-01
2.30723973e-02 -2.19012380e-01 -5.23812294e-01 -4.26221997e-01
6.25787854e-01 -7.48600245e-01 3.67883116e-01 -4.99397457e-01
2.81574488e-01 -6.16306186e-01 -2.30008755e-02 -1.20126438e+00
-3.65407467e-01 -9.72341061e-01 5.24494171e-01 1.74985564e+00
5.76160371e-01 -3.35666567e-01 3.28023523e-01 5.85522830e-01
-1.60562158e-01 2.14520425e-01 -6.87840223e-01 -9.50644135e-01
7.26570562e-02 -5.02461374e-01 3.68912578e-01 8.45128000e-01
2.65664011e-01 5.44630066e-02 -5.74898124e-01 3.77098955e-02
4.12347913e-01 -3.87878478e-01 4.08881873e-01 -9.45032656e-01
-1.87883556e-01 -1.59366623e-01 -5.32436371e-01 -2.78733581e-01
-3.62372965e-01 -8.80366802e-01 4.13504660e-01 -1.10540748e+00
-1.90871596e-01 -5.07134378e-01 -2.24493048e-03 3.81781846e-01
-3.74489278e-01 4.02379215e-01 4.68790650e-01 2.09726110e-01
-1.20734811e-01 -1.37328014e-01 3.46732616e-01 6.42802939e-02
-3.76500964e-01 -2.20853731e-01 -1.61884487e-01 3.54069918e-01
8.90826941e-01 -9.55170751e-01 1.78379431e-01 -3.86187285e-01
4.48539644e-01 -4.29292679e-01 -1.56410918e-01 -1.20159662e+00
5.18419333e-02 -2.26771712e-01 2.18634203e-01 -7.83618927e-01
-4.18152571e-01 -4.91627276e-01 -8.86858851e-02 4.64884073e-01
-4.14123356e-01 5.84409535e-01 2.11907968e-01 8.89253318e-02
9.06669945e-02 -1.07320821e+00 9.27015841e-01 -5.36120385e-02
-1.15812552e+00 -3.11963916e-01 -1.29237890e+00 8.85503441e-02
9.33023036e-01 -5.35402954e-01 -3.50276589e-01 2.52459973e-01
-1.22856468e-01 -4.92963880e-01 6.09127343e-01 3.07629347e-01
3.25546056e-01 -1.01773918e+00 -8.17791879e-01 4.39927489e-01
3.34091038e-01 -8.47059429e-01 2.46865749e-02 5.39911926e-01
-1.55552173e+00 5.53189039e-01 -6.64928675e-01 -1.00390837e-01
-1.64069307e+00 4.39416140e-01 -9.15120021e-02 -2.01687753e-01
-3.90713632e-01 7.29546189e-01 -8.08711410e-01 -2.72716433e-01
5.51415496e-02 -9.64587927e-02 -2.57963657e-01 5.98324798e-02
9.31404412e-01 7.11184740e-01 4.37049419e-01 -6.80679202e-01
-4.88013029e-01 3.43789846e-01 -5.21143854e-01 -4.61402744e-01
9.49557424e-01 -2.23352984e-01 -1.86575085e-01 5.52949429e-01
9.94749784e-01 5.16197443e-01 -3.19839180e-01 1.20592192e-01
3.71226162e-01 -5.64042211e-01 -1.87146232e-01 -1.26394260e+00
-5.14489710e-02 8.85887861e-01 8.48461211e-01 6.22376539e-02
8.18377316e-01 -7.30046690e-01 9.10790205e-01 4.62639481e-01
1.08459815e-01 -2.00741315e+00 -8.14581692e-01 9.80328918e-01
6.66397810e-01 -1.06273437e+00 -2.20865324e-01 -2.98373811e-02
-6.85783386e-01 1.58672369e+00 3.65221709e-01 -1.94455869e-02
1.34335548e-01 3.62665504e-01 3.26070696e-01 4.25501406e-01
-7.37941116e-02 -6.99519888e-02 6.76761344e-02 8.81886184e-01
8.33464503e-01 1.44267887e-01 -1.27549314e+00 2.29459792e-01
-3.85199875e-01 2.68075876e-02 9.95191216e-01 1.38602579e+00
-5.96029520e-01 -1.42653620e+00 -9.49719608e-01 2.41210550e-01
-6.60248578e-01 -4.22103882e-01 -5.23598313e-01 5.54700613e-01
2.00062767e-02 1.00170660e+00 -1.07052140e-01 -2.49102190e-01
1.09803982e-01 4.53987867e-01 4.68430161e-01 -4.93680328e-01
-1.10814440e+00 -1.68634295e-01 2.73782372e-01 9.30866133e-03
5.44625036e-02 -1.00612390e+00 -1.18427455e+00 -4.49293703e-01
-3.99717480e-01 2.83844590e-01 1.07060754e+00 1.01934266e+00
1.53122796e-02 -2.31640235e-01 1.72603786e-01 -1.98232904e-01
-4.93065953e-01 -1.09221399e+00 -7.31061935e-01 5.32888889e-01
-1.98325273e-02 5.10482676e-02 -4.04238142e-02 2.17438191e-01] | [10.824790954589844, 10.601958274841309] |
06bea677-4dfc-4450-8d23-311bb5e05169 | reasoning-with-latent-structure-refinement | 2005.06312 | null | https://arxiv.org/abs/2005.06312v3 | https://arxiv.org/pdf/2005.06312v3.pdf | Reasoning with Latent Structure Refinement for Document-Level Relation Extraction | Document-level relation extraction requires integrating information within and across multiple sentences of a document and capturing complex interactions between inter-sentence entities. However, effective aggregation of relevant information in the document remains a challenging research question. Existing approaches construct static document-level graphs based on syntactic trees, co-references or heuristics from the unstructured text to model the dependencies. Unlike previous methods that may not be able to capture rich non-local interactions for inference, we propose a novel model that empowers the relational reasoning across sentences by automatically inducing the latent document-level graph. We further develop a refinement strategy, which enables the model to incrementally aggregate relevant information for multi-hop reasoning. Specifically, our model achieves an F1 score of 59.05 on a large-scale document-level dataset (DocRED), significantly improving over the previous results, and also yields new state-of-the-art results on the CDR and GDA dataset. Furthermore, extensive analyses show that the model is able to discover more accurate inter-sentence relations. | ['Ivan Sekulić', 'Zhijiang Guo', 'Wei Lu', 'Guoshun Nan'] | 2020-05-13 | reasoning-with-latent-structure-refinement-1 | https://aclanthology.org/2020.acl-main.141 | https://aclanthology.org/2020.acl-main.141.pdf | acl-2020-6 | ['document-level-relation-extraction'] | ['natural-language-processing'] | [ 1.14527270e-01 5.95303774e-01 -4.90617752e-01 -5.86755216e-01
-1.05861425e+00 -6.40089035e-01 8.55268896e-01 8.19943666e-01
-2.67107617e-02 8.54939520e-01 6.80377543e-01 -4.50539440e-01
-4.79329169e-01 -1.06455123e+00 -5.52032709e-01 4.66563143e-02
-3.37928921e-01 6.88753247e-01 4.29471016e-01 -2.74886787e-01
3.08473743e-02 2.08567038e-01 -1.18351305e+00 7.56860793e-01
1.09907615e+00 7.24984169e-01 -1.33675724e-01 6.53108895e-01
-5.35953164e-01 1.53195250e+00 -6.97857797e-01 -9.55118299e-01
-3.59775662e-01 -4.30437595e-01 -1.37355399e+00 -8.07313770e-02
4.74654734e-01 2.95405388e-02 -2.70488597e-02 8.75808001e-01
-1.09885417e-01 -2.18402758e-01 5.51205575e-01 -9.15125310e-01
-5.63709021e-01 1.39993787e+00 -4.68416989e-01 2.54055440e-01
7.15027273e-01 -7.68833280e-01 1.79438376e+00 -7.31139541e-01
1.09652209e+00 1.30187488e+00 5.99629819e-01 8.73863325e-02
-1.13492799e+00 -3.33017409e-01 4.54359502e-01 2.49720097e-01
-1.27210629e+00 -1.97993308e-01 7.17649579e-01 -3.29850346e-01
1.61419809e+00 4.60189015e-01 4.16860431e-01 9.35166597e-01
3.58098179e-01 7.43477106e-01 9.05242920e-01 -7.36276865e-01
-1.72437698e-01 -1.28103629e-01 8.82465899e-01 7.69661844e-01
7.67620385e-01 -8.22063565e-01 -7.24824488e-01 -2.48050958e-01
3.49088162e-02 -3.38347226e-01 1.27615318e-01 9.15467553e-03
-9.46099520e-01 5.88945270e-01 1.63548142e-01 5.99970460e-01
-4.13584381e-01 -2.57246077e-01 2.99638540e-01 1.20130152e-01
7.13301539e-01 6.37370050e-01 -6.32572174e-01 -6.14098310e-02
-8.88496339e-01 3.78211677e-01 1.26725972e+00 1.19254363e+00
4.41603720e-01 -7.52733469e-01 -4.67481762e-01 6.80954099e-01
3.26813400e-01 1.02470875e-01 -1.43306762e-01 -7.42883384e-01
1.18358970e+00 1.18953872e+00 -1.79006577e-01 -1.30301297e+00
-5.25449574e-01 -5.04216909e-01 -7.54498124e-01 -7.26881921e-01
1.45350695e-01 -9.07889754e-02 -3.67541522e-01 1.42030478e+00
4.03093547e-01 -2.98229218e-01 4.08072025e-01 2.45626584e-01
1.11597359e+00 3.25565845e-01 8.41852725e-02 -5.92459321e-01
1.56580532e+00 -1.05487776e+00 -1.19171154e+00 -4.15779918e-01
1.09298921e+00 -5.56217670e-01 5.43474257e-01 2.91546851e-01
-1.04410195e+00 -2.83329219e-01 -1.05176461e+00 -4.93627101e-01
-5.14774561e-01 8.74923393e-02 8.41629386e-01 9.03488696e-02
-8.30644667e-01 3.28076154e-01 -6.06377959e-01 -2.36551687e-01
3.14158708e-01 2.55819947e-01 -3.83803815e-01 -1.63305312e-01
-1.60912585e+00 1.07765603e+00 5.62905312e-01 9.66838654e-03
-1.27869487e-01 -6.08669043e-01 -1.01463628e+00 8.90379474e-02
1.04060566e+00 -8.26015592e-01 9.89687324e-01 -1.31436717e-02
-9.85365510e-01 6.80508971e-01 -6.46117747e-01 -5.85707486e-01
1.01646550e-01 -5.02158701e-01 -5.59679568e-01 1.47898301e-01
2.82741368e-01 8.07626871e-04 6.97287321e-02 -1.17187083e+00
-6.96147799e-01 -5.44585586e-01 4.92501616e-01 9.01255608e-02
-4.03600693e-01 4.72897887e-01 -8.20281148e-01 -5.16758800e-01
9.40250456e-02 -6.62857473e-01 -5.34289889e-02 -7.88647652e-01
-9.79457915e-01 -6.43917203e-01 7.32883453e-01 -7.62594521e-01
1.94238710e+00 -1.51221704e+00 3.67514670e-01 2.15929165e-01
5.61779857e-01 -3.54201123e-02 1.02517724e-01 7.79045522e-01
-4.94655035e-03 4.89093781e-01 -1.68588057e-01 -4.65234637e-01
5.45651093e-02 3.97858679e-01 -3.17103386e-01 -2.72806644e-01
4.36818242e-01 1.14048874e+00 -9.85379279e-01 -1.02796364e+00
-2.81691581e-01 2.22528443e-01 -2.97768623e-01 1.00657038e-01
-3.73640209e-01 -5.88578247e-02 -6.71307981e-01 5.96307039e-01
3.82200450e-01 -5.34806609e-01 8.01579237e-01 -3.23470443e-01
2.52135068e-01 9.20778871e-01 -9.71385658e-01 1.45933592e+00
-3.76244038e-01 4.30921197e-01 -1.29334912e-01 -8.58687758e-01
1.00881886e+00 1.20962292e-01 3.79195422e-01 -5.74243248e-01
-2.42933169e-01 -3.47123444e-02 -9.51725170e-02 -6.87930763e-01
6.96835518e-01 2.45422915e-01 -4.41368222e-01 3.90268624e-01
9.26057473e-02 -1.81053132e-01 9.66892660e-01 9.03953493e-01
1.50293374e+00 -9.28810462e-02 5.01837611e-01 -1.76101416e-01
7.44240820e-01 9.21023339e-02 3.48909527e-01 7.27152109e-01
4.82917637e-01 1.32778570e-01 1.12946868e+00 -1.28170311e-01
-5.23792386e-01 -5.86799204e-01 -1.34625405e-01 8.17034364e-01
1.67757198e-02 -1.45560944e+00 -5.58652043e-01 -1.30539501e+00
-4.49826568e-02 9.42790449e-01 -6.70592070e-01 1.45271629e-01
-6.47248864e-01 -7.99613714e-01 5.38137496e-01 6.43335462e-01
3.38448346e-01 -6.27951324e-01 2.17312425e-01 2.44399071e-01
-6.89816117e-01 -1.92270303e+00 4.29165810e-02 1.47916183e-01
-6.24545634e-01 -1.23266399e+00 2.91127294e-01 -4.64065433e-01
6.98724091e-01 -1.94271833e-01 1.55693936e+00 1.40609950e-01
5.06240455e-03 3.44822630e-02 -6.14133358e-01 -1.27555564e-01
-5.41053534e-01 5.27249932e-01 -2.65855849e-01 -4.29297000e-01
5.72226644e-01 -3.26721787e-01 6.02346584e-02 -5.35917468e-02
-6.50840521e-01 2.44271606e-01 4.79995668e-01 6.33495212e-01
4.65601027e-01 5.02986252e-01 4.49174285e-01 -1.50987947e+00
1.00732720e+00 -4.35137898e-01 -3.02617997e-01 7.76910067e-01
-8.30534995e-01 4.01323438e-01 6.80312395e-01 1.20644383e-01
-1.35910916e+00 -3.17646623e-01 6.77898303e-02 5.01961768e-01
1.27405956e-01 1.08960319e+00 -2.07380056e-01 6.29465222e-01
3.28466266e-01 -3.39247644e-01 -7.29390204e-01 -2.98488200e-01
5.80090582e-01 7.05212533e-01 4.61909294e-01 -8.75521004e-01
8.27576756e-01 1.91299587e-01 3.36605370e-01 -5.40276468e-01
-1.67016602e+00 -3.97066474e-01 -1.16263187e+00 1.15941569e-01
7.22359002e-01 -8.39388132e-01 -5.10488272e-01 -1.21109076e-01
-1.49843085e+00 -5.95738813e-02 -3.62460688e-02 1.88117713e-01
5.08329049e-02 4.67233241e-01 -8.07335198e-01 -7.78649390e-01
-4.07633275e-01 -7.61660457e-01 1.21587598e+00 -1.90112829e-01
-8.28839719e-01 -1.23722887e+00 1.65642247e-01 9.04288530e-01
-1.99619472e-01 3.78765911e-01 1.15358520e+00 -8.20238352e-01
-5.89908063e-01 -3.56853455e-01 -3.79673898e-01 3.02050728e-02
4.37325537e-01 3.56086880e-01 -5.96934199e-01 1.97452873e-01
-4.86139268e-01 -2.43841708e-01 7.59760439e-01 -1.58085123e-01
9.24576879e-01 -5.02781570e-01 -7.03324914e-01 -3.79500389e-02
9.52868164e-01 -9.40807015e-02 3.73382717e-01 2.88353562e-01
8.99540067e-01 8.68209183e-01 7.14368701e-01 1.24542937e-01
1.06307614e+00 8.05577278e-01 6.81221038e-02 1.00443035e-01
-1.17143631e-01 -4.10527527e-01 1.92013100e-01 1.27431023e+00
-2.58123249e-01 -3.43310803e-01 -1.10437703e+00 5.54062843e-01
-2.10049343e+00 -1.04393959e+00 -7.59495199e-01 1.47108698e+00
1.33952761e+00 5.62298715e-01 -1.58881605e-01 1.75423875e-01
3.67336333e-01 1.54498011e-01 -5.48448507e-03 -3.78970951e-01
-3.29934269e-01 1.99676976e-01 1.62357867e-01 8.79039645e-01
-1.07354403e+00 1.04826951e+00 6.95173645e+00 6.74248695e-01
-2.30545387e-01 -1.37534440e-01 2.26941958e-01 1.27015546e-01
-5.65109968e-01 4.09368217e-01 -1.39146483e+00 1.04143269e-01
1.07926142e+00 -2.19810352e-01 -4.65695886e-03 5.51911950e-01
-8.72475933e-03 -1.92570582e-01 -1.20624483e+00 2.44356558e-01
1.93978816e-01 -1.49222708e+00 1.49559870e-01 1.02191262e-01
7.64325500e-01 -3.75103116e-01 -6.58987045e-01 3.67657959e-01
8.21041703e-01 -7.83784449e-01 4.49143678e-01 6.76531553e-01
3.21307927e-01 -7.63270438e-01 1.01963222e+00 4.38942164e-01
-1.39525425e+00 8.00819471e-02 7.70318434e-02 -1.95375219e-01
1.70514911e-01 1.01086545e+00 -9.63000834e-01 1.29780066e+00
5.80064476e-01 1.02915084e+00 -1.17050552e+00 7.87553936e-02
-7.32657313e-01 5.45845985e-01 -8.51948708e-02 -2.65026778e-01
1.89437792e-02 -6.84058145e-02 3.04604232e-01 1.56712019e+00
-5.18220849e-02 -6.35555293e-03 7.69842118e-02 6.56501770e-01
-3.98463309e-01 1.77292377e-01 -6.54889941e-01 -2.14418039e-01
5.41018724e-01 1.44069743e+00 -5.92365801e-01 -6.18497491e-01
-6.23383164e-01 5.86752474e-01 9.72743452e-01 2.36523598e-01
-5.78829050e-01 -4.67916846e-01 1.60827309e-01 -2.00577937e-02
3.37532252e-01 -4.10307229e-01 -2.60910898e-01 -1.31259894e+00
5.34209311e-01 -7.90028453e-01 7.61007190e-01 -7.41090834e-01
-1.27196050e+00 6.62442088e-01 4.14544255e-01 -6.90324247e-01
-5.72022498e-01 -3.73349726e-01 -2.40343466e-01 5.05767465e-01
-1.31019127e+00 -1.42484486e+00 -1.12018675e-01 2.24338219e-01
2.70634145e-01 1.69986278e-01 9.64780748e-01 1.52828097e-01
-7.50142157e-01 3.52403700e-01 -3.32949132e-01 4.50189143e-01
5.78634799e-01 -1.47999454e+00 4.73241508e-01 1.11335683e+00
5.65748334e-01 1.15014267e+00 5.44926822e-01 -9.73637223e-01
-1.44478440e+00 -9.84064519e-01 1.82598376e+00 -1.02738166e+00
1.18852019e+00 -6.06736243e-01 -9.59452689e-01 1.04653919e+00
4.16049749e-01 -2.02778116e-01 7.45160997e-01 8.58940542e-01
-6.52729988e-01 -2.05444559e-01 -6.52871013e-01 5.65227866e-01
1.33576286e+00 -7.15250611e-01 -1.06202555e+00 4.25593883e-01
9.72876847e-01 -5.59723198e-01 -1.34993279e+00 6.36539578e-01
1.80688769e-01 -7.18564987e-01 8.25059474e-01 -7.52189636e-01
8.37625206e-01 -2.18073875e-01 -1.17680721e-01 -9.10746157e-01
-2.18289480e-01 -6.07516527e-01 -9.55063164e-01 1.76189184e+00
1.02319884e+00 -3.81105155e-01 3.50689054e-01 8.54601979e-01
-4.71957847e-02 -9.03893888e-01 -5.31617582e-01 -6.85106337e-01
-1.97679922e-01 -7.16030717e-01 5.08932710e-01 9.96785104e-01
7.49409378e-01 9.76038933e-01 -6.88264221e-02 2.82091200e-01
6.91434741e-01 4.27152425e-01 6.68578506e-01 -1.43548441e+00
-2.21587211e-01 -2.46173635e-01 -7.13309348e-02 -8.29201758e-01
6.90307558e-01 -1.03748989e+00 -2.09946170e-01 -2.27160692e+00
5.65704405e-01 -8.17473009e-02 2.00509690e-02 6.57485366e-01
-4.04323995e-01 -1.80014029e-01 -1.98284850e-01 7.05051571e-02
-1.04553771e+00 1.57150537e-01 1.33234286e+00 -2.92816639e-01
-2.03734245e-02 -2.68006742e-01 -1.06106353e+00 6.96534753e-01
4.30143803e-01 -5.62073290e-01 -4.83384639e-01 -2.68618435e-01
7.04480827e-01 1.60058551e-02 3.52550298e-02 -4.11078870e-01
5.18392801e-01 -1.97651163e-01 8.63023698e-02 -8.63309443e-01
1.70217112e-01 -5.69403768e-01 -2.77115479e-02 8.42463505e-03
-6.46379888e-01 -8.55520144e-02 4.08971906e-02 4.04222757e-01
-5.43086946e-01 -2.31150925e-01 -4.19105813e-02 7.25880563e-02
-3.21906269e-01 -7.68627003e-02 -5.28183617e-02 3.69876176e-01
8.33869040e-01 2.80695140e-01 -8.74198973e-01 -1.96270317e-01
-6.40196443e-01 4.25608128e-01 -7.43493345e-03 5.12306809e-01
3.60127181e-01 -1.20589876e+00 -7.38410115e-01 -2.25040033e-01
1.97303221e-01 2.94034213e-01 -9.01772603e-02 8.23883057e-01
-1.79361090e-01 8.38089108e-01 6.17039919e-01 -2.78908759e-01
-1.71658409e+00 4.20181692e-01 -1.44245461e-01 -1.19057882e+00
-4.92632598e-01 4.88504410e-01 -4.74947810e-01 -2.44281769e-01
-1.41722783e-01 -7.27540076e-01 -4.99961793e-01 3.91338795e-01
4.44386631e-01 8.59265402e-02 3.36220473e-01 -6.54342532e-01
-6.09451771e-01 6.07522428e-01 -4.31516707e-01 7.90925696e-02
1.42000782e+00 -2.10651457e-01 -7.76522517e-01 4.18926179e-01
9.72495377e-01 5.08103967e-01 -4.53370124e-01 -4.46756214e-01
6.85712099e-01 -6.45922720e-02 -7.25288540e-02 -7.80559897e-01
-4.95432824e-01 4.49781626e-01 -8.45713913e-01 7.96272159e-01
7.38052487e-01 6.02101922e-01 5.29365420e-01 7.82488585e-01
2.63363332e-01 -9.18452144e-01 2.76934188e-02 7.85326242e-01
9.87577856e-01 -1.09214604e+00 6.25107706e-01 -1.04801321e+00
-6.42769873e-01 1.09221339e+00 6.16595984e-01 4.04155940e-01
5.43617249e-01 8.26367736e-01 -1.88038319e-01 -5.51072598e-01
-1.18212736e+00 -2.23391429e-01 6.63226962e-01 2.00089678e-01
8.29254389e-01 -6.11761771e-02 -3.68168622e-01 8.37965727e-01
-3.87360513e-01 -1.74742296e-01 3.50207180e-01 9.51569557e-01
-4.73724343e-02 -1.53647256e+00 -3.34562808e-02 4.47643429e-01
-4.79554176e-01 -2.62490362e-01 -9.97565687e-01 9.16930437e-01
-1.27683222e-01 1.30381024e+00 -1.59787267e-01 -3.59480739e-01
4.98301953e-01 2.35103235e-01 5.03092349e-01 -8.13395441e-01
-4.47684020e-01 -1.94339141e-01 1.01310718e+00 -4.07934636e-01
-8.64477396e-01 -6.25862002e-01 -1.44214988e+00 -1.49866760e-01
-4.30197507e-01 2.71878362e-01 1.22932248e-01 1.39218247e+00
5.22682846e-01 8.76008093e-01 3.49166930e-01 1.18743256e-01
8.22820142e-02 -1.05270743e+00 -3.41686040e-01 4.20087218e-01
1.19207934e-01 -4.62205261e-01 -1.59461156e-01 1.59889475e-01] | [9.262495994567871, 8.639062881469727] |
2c33cfc7-3e97-4c27-b31b-5c6c7311081b | beating-the-worlds-best-at-super-smash-bros | 1702.06230 | null | http://arxiv.org/abs/1702.06230v3 | http://arxiv.org/pdf/1702.06230v3.pdf | Beating the World's Best at Super Smash Bros. with Deep Reinforcement Learning | There has been a recent explosion in the capabilities of game-playing
artificial intelligence. Many classes of RL tasks, from Atari games to motor
control to board games, are now solvable by fairly generic algorithms, based on
deep learning, that learn to play from experience with minimal knowledge of the
specific domain of interest. In this work, we will investigate the performance
of these methods on Super Smash Bros. Melee (SSBM), a popular console fighting
game. The SSBM environment has complex dynamics and partial observability,
making it challenging for human and machine alike. The multi-player aspect
poses an additional challenge, as the vast majority of recent advances in RL
have focused on single-agent environments. Nonetheless, we will show that it is
possible to train agents that are competitive against and even surpass human
professionals, a new result for the multi-player video game setting. | ['Joshua B. Tenenbaum', 'William F. Whitney', 'Vlad Firoiu'] | 2017-02-21 | null | null | null | null | ['board-games'] | ['playing-games'] | [-2.47864515e-01 5.53701743e-02 4.64730114e-02 4.61817086e-01
-4.88869429e-01 -8.48729193e-01 5.11070192e-01 -2.48534903e-01
-8.72843266e-01 1.00835419e+00 -3.44690531e-01 -4.31837469e-01
-5.06535769e-01 -6.78924203e-01 -6.18763685e-01 -5.86719930e-01
-5.85230649e-01 8.69522691e-01 5.73882103e-01 -1.06170321e+00
-1.06224880e-01 2.70319194e-01 -1.59774423e+00 1.54173866e-01
3.54908139e-01 7.63906300e-01 3.42040032e-01 1.17464840e+00
5.78699589e-01 1.58826530e+00 -6.89337432e-01 -1.70050576e-01
6.21028185e-01 -4.51470524e-01 -8.35140646e-01 -2.31925547e-01
-3.08991075e-01 -2.29609936e-01 -5.36997437e-01 7.56660938e-01
5.72574615e-01 6.31938815e-01 1.93799600e-01 -1.53751957e+00
2.59611964e-01 5.33949971e-01 -2.77969033e-01 4.22279567e-01
4.12537545e-01 4.63734359e-01 1.12674546e+00 3.94472927e-02
7.77806461e-01 1.00910306e+00 5.76993763e-01 7.12704122e-01
-8.01751256e-01 -6.88468754e-01 1.56906083e-01 3.98387581e-01
-9.50591683e-01 7.80671239e-02 3.41708750e-01 -4.11325455e-01
1.07105172e+00 2.09917560e-01 9.60728586e-01 1.03830254e+00
1.54324993e-01 8.89017105e-01 1.17563725e+00 -1.75391525e-01
5.20210743e-01 -3.96570325e-01 -5.23177564e-01 5.84913492e-01
8.67672488e-02 5.50825596e-01 -4.33546543e-01 6.44359291e-02
9.38137650e-01 -4.28022891e-01 -2.21774857e-02 -6.94088995e-01
-9.73304987e-01 1.11266673e+00 1.77152008e-01 2.73268729e-01
-4.46024358e-01 4.61751014e-01 4.10506874e-01 7.74977863e-01
-6.21909052e-02 9.94833410e-01 -5.37258029e-01 -7.22740710e-01
-4.95987117e-01 9.60028708e-01 1.29629982e+00 4.56699848e-01
3.01178217e-01 2.09880412e-01 5.22791386e-01 3.01718831e-01
5.40670305e-02 2.53185511e-01 4.36931968e-01 -1.33978903e+00
5.22124410e-01 2.05352783e-01 3.90831888e-01 -9.27070737e-01
-8.85929585e-01 -4.65342373e-01 -5.39457381e-01 9.56467330e-01
6.56534374e-01 -6.86876178e-01 -4.31740642e-01 1.80643296e+00
3.36198628e-01 4.29444104e-01 2.54764885e-01 1.03819251e+00
4.37299013e-01 6.34446859e-01 -1.15685321e-01 7.37412833e-04
1.07168341e+00 -1.02956915e+00 -3.33426774e-01 -6.61295116e-01
5.31919062e-01 -1.35918498e-01 7.65594482e-01 8.43307078e-01
-1.30268645e+00 -2.99309283e-01 -1.24263442e+00 4.20315951e-01
-2.64952511e-01 -4.59909081e-01 1.09484184e+00 5.89512885e-01
-1.05022347e+00 7.11136103e-01 -1.17532182e+00 -3.61898124e-01
1.99861050e-01 8.74875903e-01 -4.93956268e-01 4.54845130e-02
-1.41759288e+00 1.30503571e+00 7.43603289e-01 -1.32604972e-01
-1.28267407e+00 -4.42215472e-01 -7.12766171e-01 8.57129414e-03
1.02587461e+00 -5.58588028e-01 1.77222371e+00 -1.14968014e+00
-2.06235099e+00 7.41130710e-01 7.84721255e-01 -7.32000113e-01
7.95406938e-01 -1.35935485e-01 6.97139427e-02 -9.56672356e-02
-7.23090535e-03 2.51753896e-01 4.94610518e-01 -9.22409892e-01
-1.13062632e+00 -9.85361040e-02 9.35506761e-01 7.30176032e-01
2.71258324e-01 7.77538866e-02 -5.13283126e-02 -1.45987973e-01
-3.31811756e-01 -1.14548063e+00 -6.97471976e-01 -4.46848333e-01
9.88186598e-02 -2.63459682e-01 4.56497669e-01 -1.64428562e-01
8.05119812e-01 -1.85241008e+00 6.89063966e-01 -9.33806673e-02
3.10589522e-01 3.84715378e-01 -1.63018167e-01 6.64615691e-01
4.96968441e-02 -3.06857914e-01 2.14545399e-01 8.52967575e-02
3.09224874e-01 5.15228152e-01 -3.62708271e-02 4.47564155e-01
-2.00914919e-01 9.45250690e-01 -1.22116482e+00 -2.28367168e-02
4.56365980e-02 -1.38546452e-01 -8.44844222e-01 1.98023513e-01
-4.49334949e-01 4.80242342e-01 -6.49491012e-01 -7.84775708e-03
3.45039903e-03 -1.53193241e-02 3.66587698e-01 6.57908142e-01
-2.34510481e-01 9.57403481e-02 -1.36730707e+00 1.79050171e+00
-5.00093460e-01 6.35773778e-01 5.30078948e-01 -1.20499885e+00
2.49294579e-01 2.95354903e-01 6.91693604e-01 -8.15732658e-01
3.75365466e-01 3.77088822e-02 6.31295800e-01 -4.27997231e-01
5.30443609e-01 -5.65270245e-01 -4.73568112e-01 6.91931427e-01
9.87734273e-02 -5.67661703e-01 6.10498607e-01 -4.61436361e-02
1.59775805e+00 2.56466448e-01 3.57025862e-01 -1.64088652e-01
2.18552664e-01 5.29594958e-01 5.08965790e-01 1.22012699e+00
-2.99101144e-01 2.17721298e-01 7.14301348e-01 -5.65672874e-01
-8.98031056e-01 -1.00560021e+00 6.71481907e-01 1.43686843e+00
2.35195264e-01 -4.10846829e-01 -7.32861578e-01 -3.05833161e-01
-2.33466893e-01 3.32692683e-01 -5.25511682e-01 -3.20721447e-01
-8.65357697e-01 -7.64192164e-01 5.80843329e-01 3.23527753e-01
4.88052189e-01 -1.43015456e+00 -1.35487700e+00 6.76059425e-01
-8.95751640e-02 -1.26440060e+00 1.15421124e-01 3.35420460e-01
-4.48390812e-01 -1.27062976e+00 -4.73510742e-01 -6.42279208e-01
-3.26169848e-01 -1.01945996e-02 1.34168470e+00 4.93558981e-02
-2.94982284e-01 6.71933174e-01 -3.45182627e-01 -6.41559064e-01
-5.00516772e-01 3.63028675e-01 2.88143933e-01 -5.05949080e-01
-6.23813160e-02 -6.22535408e-01 -2.97197551e-01 2.39772037e-01
-7.36605585e-01 1.98362991e-01 3.97152990e-01 8.13415527e-01
1.13283545e-01 5.88237047e-01 2.42283523e-01 -5.69339633e-01
8.70202184e-01 -5.37075579e-01 -8.14621329e-01 -1.70427233e-01
1.36580855e-01 -1.86116755e-01 8.19599271e-01 -5.38878322e-01
-7.15799809e-01 -3.19651607e-03 -3.63873579e-02 -3.80267203e-02
-9.05331001e-02 6.74398541e-01 1.75717231e-02 -1.73570544e-01
7.59753227e-01 1.01670556e-01 1.93687141e-01 -8.46109390e-02
1.41652539e-01 1.76007554e-01 4.89497989e-01 -6.70959115e-01
9.06970859e-01 3.67345750e-01 1.57158211e-01 -8.38569939e-01
-7.71727443e-01 -1.70398876e-01 -3.64920229e-01 -3.63857031e-01
9.16292667e-01 -9.08861756e-01 -1.56237638e+00 6.60140216e-01
-8.47816110e-01 -1.14612520e+00 -4.49026376e-01 5.56697726e-01
-1.22120726e+00 -5.11058010e-02 -7.18167782e-01 -7.75800586e-01
2.95599639e-01 -1.10426807e+00 5.15056968e-01 3.61061782e-01
-2.26925481e-02 -9.77391601e-01 6.87888503e-01 4.49754328e-01
4.22066569e-01 2.06959754e-01 5.26041269e-01 -6.08063221e-01
-5.54237306e-01 -6.20705895e-02 3.30012262e-01 -8.18331912e-02
-2.61393249e-01 -3.98517936e-01 -6.31979406e-01 -5.95804691e-01
4.33364064e-02 -7.04681039e-01 3.34335476e-01 3.45357090e-01
5.60226500e-01 2.29358934e-02 4.03923318e-02 2.94806957e-01
1.21779287e+00 4.71692324e-01 5.30832112e-01 1.02734733e+00
3.28425825e-01 3.64614159e-01 7.11885393e-01 5.56552470e-01
5.45268416e-01 8.15790474e-01 7.39219189e-01 -4.36538197e-02
3.20743978e-01 -8.59346986e-02 4.46240246e-01 5.26429713e-01
-7.34469950e-01 -2.90399998e-01 -9.38153148e-01 1.43378362e-01
-2.25672460e+00 -1.28672457e+00 3.37341309e-01 1.90486479e+00
5.46605289e-01 3.55072409e-01 5.87455511e-01 1.78315431e-01
4.00572509e-01 6.33207783e-02 -6.03470445e-01 -3.85017186e-01
-1.98865309e-02 5.99883258e-01 4.10475671e-01 3.86603802e-01
-1.20095706e+00 1.32749307e+00 6.47069359e+00 1.03949881e+00
-8.67701888e-01 1.40972942e-01 2.43208066e-01 -3.40021551e-01
4.88123596e-01 -2.39492431e-01 -3.57886642e-01 1.17157690e-01
8.49808514e-01 -3.53578001e-01 1.02904630e+00 9.02895093e-01
3.17303509e-01 -3.85067761e-01 -9.00019526e-01 9.85162973e-01
-1.10100701e-01 -1.17051888e+00 -6.87435925e-01 2.06736699e-01
7.23895192e-01 2.82221317e-01 2.12154493e-01 9.26437140e-01
1.17503607e+00 -1.40047634e+00 7.03807175e-01 -3.14113917e-03
2.07230017e-01 -1.08196390e+00 7.67696500e-01 7.89709747e-01
-1.05391431e+00 -4.78690982e-01 -3.33486646e-01 -8.86050165e-01
-2.12666411e-02 -5.24432123e-01 -5.25212348e-01 4.76578772e-01
6.52082562e-01 3.91686976e-01 -5.74693196e-02 1.23406541e+00
-2.46432021e-01 2.73502588e-01 -4.65169489e-01 -2.95305550e-01
7.74072409e-01 -2.64635086e-01 4.34673935e-01 5.24617851e-01
5.99028654e-02 6.61038458e-01 5.47259867e-01 3.90287369e-01
2.64580101e-01 -1.88698933e-01 -7.10987210e-01 -1.53483212e-01
-2.15903133e-01 1.14460015e+00 -8.76224220e-01 8.74031410e-02
-2.40621254e-01 8.63985360e-01 5.29074430e-01 1.22583948e-01
-9.26589966e-01 -6.01572283e-02 1.06128764e+00 -1.63587049e-01
3.66301805e-01 -4.69780445e-01 2.52933472e-01 -1.13878429e+00
-4.06654179e-01 -1.47784519e+00 5.28952062e-01 -5.77104092e-01
-9.14341629e-01 6.22455895e-01 -1.49701014e-01 -1.10160518e+00
-7.06954002e-01 -7.41467237e-01 -5.41106880e-01 1.52721092e-01
-1.06956232e+00 -7.86929488e-01 -1.04706837e-02 7.69897401e-01
5.59804499e-01 -4.33594346e-01 6.93972707e-01 1.71728879e-01
-5.30415356e-01 1.09441578e-01 2.43153691e-01 7.98362792e-02
-8.66630767e-03 -1.24350452e+00 2.74744391e-01 6.04267895e-01
3.49695653e-01 -9.65966061e-02 1.07718050e+00 -2.60525197e-01
-1.74859416e+00 -5.13641119e-01 -1.47197079e-02 -6.71349168e-01
1.09657872e+00 -5.02218068e-01 -2.84188360e-01 7.75339723e-01
9.03137177e-02 -2.99208254e-01 1.98616073e-01 1.82157487e-01
8.57739449e-02 1.71129733e-01 -7.59160161e-01 8.90392900e-01
1.08687472e+00 -3.46422642e-01 -5.82890630e-01 3.86299968e-01
2.83893853e-01 -8.78536165e-01 -3.72312754e-01 6.20304756e-02
5.28547347e-01 -1.11021841e+00 1.09994996e+00 -1.24457526e+00
4.12860930e-01 -2.21424296e-01 8.86047110e-02 -1.70421624e+00
-2.33183250e-01 -9.53162909e-01 6.18632622e-02 2.74100572e-01
2.23264158e-01 -2.90442079e-01 1.19303060e+00 4.83157337e-01
1.43144369e-01 -4.34331328e-01 -1.13233399e+00 -1.00985527e+00
4.13072109e-01 -6.50052845e-01 1.72089785e-01 7.23351598e-01
4.76579696e-01 3.66472244e-01 -6.45643353e-01 -1.16711613e-02
3.87713194e-01 -1.27265872e-02 9.07344997e-01 -1.25675178e+00
-1.04065502e+00 -6.06800258e-01 -6.91262603e-01 -9.38817799e-01
3.59791875e-01 -4.82936382e-01 3.23417693e-01 -1.26958120e+00
4.74402979e-02 -3.66973788e-01 -1.14950739e-01 2.89414287e-01
9.94722173e-02 3.07878703e-01 6.65402055e-01 -1.28486931e-01
-1.28683925e+00 2.78776884e-01 1.48017418e+00 -3.97123694e-02
-2.16728508e-01 5.03829718e-01 -3.51266921e-01 1.01011276e+00
9.73788619e-01 -3.24347436e-01 -5.99551857e-01 -3.80396456e-01
7.12619364e-01 3.42862278e-01 1.42439231e-01 -1.44142306e+00
5.85446000e-01 -4.03459668e-01 -2.64276981e-01 1.87771827e-01
6.08402729e-01 -7.30902135e-01 1.46880165e-01 9.36625302e-01
-8.49466175e-02 2.86432564e-01 3.88902009e-01 4.85864639e-01
-1.24864802e-01 -2.83660650e-01 6.45775437e-01 -5.38743496e-01
-1.06741261e+00 4.18926865e-01 -1.18110323e+00 6.92914069e-01
1.38887465e+00 -1.45184845e-01 -5.43622114e-02 -8.77598643e-01
-9.56357241e-01 3.70222807e-01 5.69095872e-02 2.97846466e-01
1.35365948e-01 -8.30345094e-01 -7.02917099e-01 -2.75108010e-01
-4.09200668e-01 -5.54825813e-02 3.88497949e-01 5.75382113e-01
-8.76245320e-01 2.94096738e-01 -7.15300143e-01 -2.70954706e-02
-1.20297742e+00 4.75145191e-01 8.16519439e-01 -8.31972659e-01
-6.89606190e-01 8.55642915e-01 3.20749164e-01 -6.32291853e-01
2.50365525e-01 -8.13494623e-02 -3.39473754e-01 -1.27237663e-01
5.25761485e-01 3.53898644e-01 -5.88961765e-02 -3.73504668e-01
-2.29606211e-01 1.48314506e-01 6.46980628e-02 -4.79126662e-01
1.64082956e+00 3.27897936e-01 3.10863823e-01 2.91705906e-01
3.93562555e-01 -4.08615410e-01 -1.45702016e+00 6.43414110e-02
1.24058761e-01 -4.72710580e-02 -2.61147171e-02 -6.86572909e-01
-1.10611475e+00 6.86273575e-01 1.96513474e-01 4.56670314e-01
8.09060633e-01 -1.32343888e-01 5.67965329e-01 7.92549431e-01
1.01659143e+00 -1.24093914e+00 3.87381971e-01 1.09132075e+00
5.77109396e-01 -1.08983278e+00 -2.23640233e-01 1.11736991e-01
-8.23397160e-01 9.19448435e-01 7.97587216e-01 -5.31050861e-01
3.20607215e-01 6.32734835e-01 2.20681783e-02 -3.41741025e-01
-1.05830753e+00 -5.91785014e-01 -3.22089851e-01 8.98648441e-01
-2.90430099e-01 6.70345314e-03 1.74383186e-02 8.43073964e-01
-6.28233850e-01 7.70491064e-02 8.59721839e-01 1.09651458e+00
-5.98980129e-01 -9.97639775e-01 -2.58750618e-01 9.74418670e-02
-4.53365594e-01 2.34926254e-01 -4.20418769e-01 1.32404137e+00
7.00783879e-02 9.62315023e-01 -1.74124129e-02 -3.48071992e-01
4.27624613e-01 -4.35053796e-01 8.67775679e-01 -6.90030515e-01
-9.38506246e-01 -4.05602396e-01 2.23972306e-01 -9.18758690e-01
-1.86354578e-01 -7.22613454e-01 -1.21190071e+00 -6.98225081e-01
-7.66367987e-02 3.78658831e-01 4.32195932e-01 1.26401091e+00
-1.68495849e-01 5.80606878e-01 3.55012685e-01 -1.08258903e+00
-6.12790167e-01 -4.83560562e-01 -8.98919702e-01 6.02975823e-02
2.13754639e-01 -8.24217558e-01 -2.90428102e-01 -5.92457473e-01] | [3.5889205932617188, 1.4699798822402954] |
3b84822b-72f1-4432-902c-b7bf7d63f620 | surgical-vqla-transformer-with-gated-vision | 2305.11692 | null | https://arxiv.org/abs/2305.11692v1 | https://arxiv.org/pdf/2305.11692v1.pdf | Surgical-VQLA: Transformer with Gated Vision-Language Embedding for Visual Question Localized-Answering in Robotic Surgery | Despite the availability of computer-aided simulators and recorded videos of surgical procedures, junior residents still heavily rely on experts to answer their queries. However, expert surgeons are often overloaded with clinical and academic workloads and limit their time in answering. For this purpose, we develop a surgical question-answering system to facilitate robot-assisted surgical scene and activity understanding from recorded videos. Most of the existing VQA methods require an object detector and regions based feature extractor to extract visual features and fuse them with the embedded text of the question for answer generation. However, (1) surgical object detection model is scarce due to smaller datasets and lack of bounding box annotation; (2) current fusion strategy of heterogeneous modalities like text and image is naive; (3) the localized answering is missing, which is crucial in complex surgical scenarios. In this paper, we propose Visual Question Localized-Answering in Robotic Surgery (Surgical-VQLA) to localize the specific surgical area during the answer prediction. To deal with the fusion of the heterogeneous modalities, we design gated vision-language embedding (GVLE) to build input patches for the Language Vision Transformer (LViT) to predict the answer. To get localization, we add the detection head in parallel with the prediction head of the LViT. We also integrate GIoU loss to boost localization performance by preserving the accuracy of the question-answering model. We annotate two datasets of VQLA by utilizing publicly available surgical videos from MICCAI challenges EndoVis-17 and 18. Our validation results suggest that Surgical-VQLA can better understand the surgical scene and localize the specific area related to the question-answering. GVLE presents an efficient language-vision embedding technique by showing superior performance over the existing benchmarks. | ['Hongliang Ren', 'Lalithkumar Seenivasan', 'Mobarakol Islam', 'Long Bai'] | 2023-05-19 | null | null | null | null | ['answer-generation'] | ['natural-language-processing'] | [-1.03484280e-01 3.15870017e-01 -2.67619878e-01 -1.84573438e-02
-1.08621991e+00 -6.38927996e-01 -4.38257232e-02 3.36834699e-01
-4.79622155e-01 2.31233120e-01 4.35758144e-01 -6.54842675e-01
-1.32884234e-02 -5.17831504e-01 -7.54586518e-01 -5.23698688e-01
2.74487227e-01 9.93732139e-02 2.60310411e-01 -1.20572433e-01
1.34904217e-02 2.58151889e-01 -1.21352780e+00 7.53960133e-01
7.39705503e-01 1.17002845e+00 7.83735573e-01 8.46771240e-01
-3.05448085e-01 1.26809740e+00 -2.57365614e-01 -2.47084439e-01
3.54896724e-01 -5.32807171e-01 -8.00463200e-01 -2.14406729e-01
3.92333061e-01 -3.73816133e-01 -8.15780103e-01 9.04882014e-01
5.33342898e-01 -3.40960026e-01 3.34736347e-01 -1.08254540e+00
-3.15579236e-01 2.48669982e-01 -3.94667238e-01 3.62527043e-01
3.87175739e-01 3.54499489e-01 6.84331596e-01 -1.22231042e+00
9.04472172e-01 8.85986865e-01 6.09608829e-01 4.62098688e-01
-8.47476602e-01 -5.24925530e-01 9.21505466e-02 3.97722781e-01
-1.24834037e+00 -1.44599855e-01 8.91988754e-01 -5.64632297e-01
6.36817992e-01 3.24635357e-01 7.77201355e-01 9.51834142e-01
7.84462750e-01 9.96194065e-01 8.73479664e-01 -1.64633557e-01
-1.20781340e-01 4.20550585e-01 -1.01734690e-01 1.31450391e+00
-1.21064015e-01 8.95192698e-02 -5.01542866e-01 -6.77342117e-02
9.45631921e-01 4.57590222e-01 -7.03849614e-01 -7.04746127e-01
-1.60693216e+00 8.55171025e-01 9.43289995e-01 1.20493948e-01
-4.97543216e-01 1.85387075e-01 6.28216803e-01 2.25407943e-01
-1.77059770e-01 3.19937050e-01 -7.61815086e-02 8.01985785e-02
-6.05989873e-01 -2.99724728e-01 8.80232155e-01 9.25341904e-01
6.01337910e-01 -2.73092598e-01 -6.06630743e-01 5.81251502e-01
3.70279551e-01 2.79052347e-01 6.65549219e-01 -7.50959516e-01
4.70090955e-01 1.05426252e+00 -1.81903392e-01 -1.01394439e+00
-6.57181442e-01 -4.83313233e-01 -8.65666568e-01 1.42185822e-01
3.02810818e-01 1.75603941e-01 -1.06860769e+00 1.22495127e+00
3.23815703e-01 1.18403010e-01 9.61105153e-02 1.20422959e+00
1.57038426e+00 3.53729427e-01 -3.18205878e-02 8.62242654e-02
1.73705137e+00 -1.30446339e+00 -7.07785010e-01 -4.99645948e-01
9.86874938e-01 -7.49324262e-01 1.26065004e+00 5.75468130e-02
-7.18119144e-01 -5.25433958e-01 -8.81825745e-01 -2.93947965e-01
-1.46745563e-01 7.68425465e-01 5.21385968e-01 1.81256130e-01
-9.74196911e-01 -2.70308346e-01 -7.73728073e-01 -3.07529867e-01
2.93450385e-01 4.75513250e-01 -8.07713091e-01 -3.71901125e-01
-8.85146260e-01 9.10942554e-01 7.63091519e-02 3.94421011e-01
-1.28100431e+00 -7.36353517e-01 -1.33584189e+00 -1.35636643e-01
5.01039147e-01 -9.02759969e-01 1.06971467e+00 -7.33254790e-01
-1.13071787e+00 1.04108012e+00 -1.13653643e-02 -2.43490234e-01
4.65498537e-01 8.02775249e-02 -2.84864195e-02 5.81605017e-01
7.50845149e-02 6.94534123e-01 7.44449437e-01 -1.25202179e+00
-4.12199974e-01 -4.79729325e-01 4.62223977e-01 2.97537982e-01
-1.72194406e-01 -5.48890650e-01 -7.88182437e-01 -5.46663702e-01
2.10887089e-01 -9.42739308e-01 -3.71642709e-01 5.89378059e-01
-2.78269351e-01 1.44650713e-01 7.97176957e-01 -9.55265939e-01
1.22347009e+00 -2.35867929e+00 7.02982023e-02 -1.54643789e-01
5.64676583e-01 -1.00664003e-02 -2.09010810e-01 1.01587303e-01
-5.91575466e-02 -4.29223388e-01 1.98032975e-01 -3.10450286e-01
-5.63357472e-01 5.10935128e-01 -2.66624808e-01 6.56314433e-01
-7.48887807e-02 1.18594682e+00 -8.80896688e-01 -1.13035059e+00
4.15637195e-01 2.00999111e-01 -8.81323338e-01 5.91019511e-01
7.94112384e-02 6.36534572e-01 -4.74340796e-01 1.04746044e+00
2.11071312e-01 -4.32269365e-01 -7.13370219e-02 -7.29530811e-01
1.51201412e-01 -1.31871656e-01 -7.06292152e-01 2.26253104e+00
-8.03712606e-01 4.57918882e-01 5.84173024e-01 -9.57199931e-01
7.11120725e-01 2.83677757e-01 5.99602163e-01 -8.14944685e-01
2.65963048e-01 2.81491846e-01 -1.41795641e-02 -9.81487751e-01
1.44625351e-01 -2.62732059e-02 -1.43432215e-01 -2.60834634e-01
2.92492598e-01 -2.48029664e-01 -2.45112672e-01 4.42570180e-01
1.25997818e+00 1.17624586e-03 3.33250314e-01 -4.38394994e-02
7.24341512e-01 4.10163999e-01 4.22309816e-01 7.73390114e-01
-5.47675848e-01 7.19177246e-01 5.23608983e-01 -5.16544461e-01
-4.82631832e-01 -1.13262939e+00 8.13929066e-02 7.17218816e-01
5.45110762e-01 -3.02035302e-01 -3.49989772e-01 -1.18380833e+00
-5.00951856e-02 2.76050627e-01 -8.69001150e-01 -3.94730926e-01
-6.43494546e-01 5.60221821e-02 1.73775285e-01 6.55860841e-01
1.47171030e-02 -1.00642645e+00 -8.03024590e-01 1.31064579e-01
-4.85809773e-01 -1.36681306e+00 -7.75176585e-01 3.02468568e-01
-7.50993371e-01 -1.28724098e+00 -7.96991825e-01 -1.19569385e+00
8.40403855e-01 3.33420247e-01 9.17099833e-01 -1.49269942e-02
-7.22304583e-01 8.10653031e-01 -2.82061011e-01 -4.13683653e-01
-3.37970883e-01 -8.94834995e-02 -3.37576091e-01 -3.68949473e-02
3.29338089e-02 7.57137910e-02 -1.20354712e+00 3.07828933e-01
-8.28132153e-01 3.15938264e-01 9.87016618e-01 1.07223856e+00
7.64090240e-01 -8.95480931e-01 1.89976141e-01 -4.95748848e-01
3.29480022e-01 -5.34578681e-01 -4.41893131e-01 4.82059658e-01
-1.54431179e-01 -3.03691141e-02 5.89264810e-01 -4.68868583e-01
-5.56128800e-01 4.09172565e-01 -1.40648991e-01 -8.97175252e-01
1.54532686e-01 5.82658529e-01 2.33340591e-01 -3.02045375e-01
6.83399320e-01 2.75492668e-01 3.57454896e-01 2.06080023e-02
1.16988949e-01 4.69920576e-01 6.82819486e-01 -4.33420725e-02
4.77071375e-01 5.04484057e-01 -1.41183808e-01 -5.28422654e-01
-8.03224325e-01 -7.34668672e-01 -1.15973972e-01 -4.72653031e-01
1.00603259e+00 -1.04394519e+00 -1.04616451e+00 -2.39291891e-01
-1.27165675e+00 -5.03589809e-02 -3.04263413e-01 7.97861218e-01
-6.01656854e-01 4.54825073e-01 -7.61865675e-01 -3.95793855e-01
-6.68609202e-01 -1.78211474e+00 1.39038968e+00 2.49574557e-01
1.27382919e-01 -8.11311364e-01 8.01822618e-02 6.44487441e-01
3.40216011e-01 2.53790230e-01 9.40404415e-01 -4.81806695e-01
-8.83467734e-01 -4.63913262e-01 -3.79113168e-01 1.80946931e-01
-3.47468331e-02 -6.98150396e-01 -7.19641685e-01 -2.91978538e-01
-6.82792664e-02 -2.05081388e-01 7.71490812e-01 4.79756355e-01
1.47602999e+00 -8.22110027e-02 -5.20130754e-01 8.52259457e-01
1.29467726e+00 -4.69344258e-02 3.18318903e-01 9.44235846e-02
8.08207870e-01 5.25331259e-01 7.99089372e-01 2.95459837e-01
7.22145379e-01 5.24916291e-01 9.66732562e-01 -3.95707130e-01
-3.23852748e-01 -3.10282469e-01 3.84759843e-01 9.51445222e-01
3.06868494e-01 6.82099089e-02 -1.19446671e+00 7.58240283e-01
-1.91834140e+00 -3.96084219e-01 -2.27346346e-02 1.87600398e+00
7.35258877e-01 -2.52281815e-01 -4.19557571e-01 -3.65940541e-01
2.07415700e-01 1.59690939e-02 -4.71855849e-01 -1.70628820e-02
3.87877464e-01 5.34035005e-02 6.35215938e-01 4.39835608e-01
-1.03569722e+00 6.44827187e-01 5.09315300e+00 7.00186372e-01
-1.22997057e+00 3.46971214e-01 2.73643672e-01 -1.58400267e-01
-2.82371163e-01 -2.21898183e-01 -5.65328240e-01 9.58767533e-02
4.12451923e-01 1.02315612e-01 5.96014736e-03 1.04465234e+00
6.31799623e-02 -2.13411465e-01 -1.20868397e+00 1.50264764e+00
4.54708248e-01 -1.56034410e+00 -9.33621824e-03 -4.66591939e-02
2.90360183e-01 -6.46555573e-02 -1.37773864e-02 5.91934741e-01
-2.22461119e-01 -1.14252818e+00 4.84378487e-01 8.03378642e-01
8.71430039e-01 -2.27454618e-01 1.08649004e+00 3.37515861e-01
-1.30468762e+00 -4.06444073e-01 -1.15462542e-01 4.61515546e-01
-9.90019813e-02 1.12722330e-01 -1.19504666e+00 4.80787754e-01
7.57126570e-01 4.86304700e-01 -5.70327640e-01 1.12200534e+00
-1.22561321e-01 2.33089522e-01 -1.66282266e-01 1.48650095e-01
2.92680800e-01 1.68681547e-01 3.66405904e-01 9.27559793e-01
3.71016532e-01 9.54766348e-02 3.96895617e-01 7.05744922e-01
-1.04454555e-01 2.92749316e-01 -6.79951370e-01 1.90046895e-02
-1.36502340e-01 1.42538369e+00 -4.65464890e-01 1.24097560e-02
-7.27559924e-01 1.15476382e+00 1.71978712e-01 3.79350722e-01
-9.07708585e-01 -2.43885010e-01 3.87722611e-01 2.56984323e-01
1.74489990e-02 -1.98689774e-01 -2.00646833e-01 -1.23463178e+00
1.49089128e-01 -8.32482219e-01 6.39885724e-01 -8.37010920e-01
-7.87426889e-01 6.76146209e-01 -3.66197109e-01 -1.62567604e+00
-8.53731558e-02 -9.32025969e-01 -3.19722682e-01 6.80142820e-01
-1.60895908e+00 -1.54599476e+00 -1.06465340e+00 9.71291065e-01
7.71319509e-01 5.57851139e-03 8.86907339e-01 2.21857488e-01
-3.14644396e-01 4.91454899e-01 -4.94764268e-01 3.86086047e-01
9.47318852e-01 -9.95949090e-01 -6.11502826e-01 4.57589477e-01
8.28364771e-03 4.52495694e-01 3.17819118e-01 -3.62346351e-01
-1.91412759e+00 -9.04822469e-01 4.34429973e-01 -6.46333516e-01
6.66908085e-01 -3.32419515e-01 -5.87922215e-01 6.50131762e-01
1.88552421e-02 5.20717502e-01 7.83465803e-01 -5.02858460e-01
-2.60997899e-02 -2.78741598e-01 -9.45226073e-01 6.13054037e-01
6.73420072e-01 -7.49256611e-01 -6.01426125e-01 4.34816182e-01
7.92376161e-01 -8.57396483e-01 -8.98118675e-01 7.83282220e-01
5.44447839e-01 -7.00316191e-01 9.66810763e-01 -3.30567479e-01
6.30647659e-01 -4.02551472e-01 -1.50523052e-01 -7.84908175e-01
1.99956700e-01 -8.29702336e-03 -1.42275944e-01 3.66891772e-01
3.56980443e-01 -2.92879045e-01 8.38505507e-01 3.44326854e-01
-4.17086184e-01 -1.18586612e+00 -1.07463574e+00 1.81956328e-02
-2.87470758e-01 -3.45138162e-01 -1.46131694e-01 6.76933348e-01
2.79048309e-02 1.87407322e-02 -9.34128389e-02 2.90570021e-01
3.16632777e-01 3.23732287e-01 7.63476670e-01 -7.18464553e-01
-3.51773381e-01 -1.21223584e-01 -5.91398001e-01 -1.01286173e+00
-7.55845010e-02 -1.11467361e+00 1.55664384e-01 -2.07696390e+00
2.86758780e-01 -4.11910526e-02 -3.85359496e-01 5.53726673e-01
-7.97255188e-02 2.69492090e-01 1.36483416e-01 1.57860100e-01
-5.57182491e-01 4.55851883e-01 1.56760120e+00 -3.01217020e-01
-5.14764376e-02 -1.79921255e-01 -3.54450941e-01 7.68197715e-01
3.00206691e-01 -3.39322984e-01 -4.02810633e-01 -3.86335194e-01
2.38043174e-01 6.56609416e-01 7.86487520e-01 -8.39000404e-01
7.12575972e-01 1.66697666e-01 3.17120999e-01 -7.48141646e-01
4.06601697e-01 -1.16483545e+00 -3.08041334e-01 7.97213197e-01
-7.85070360e-02 -4.59193923e-02 3.48546237e-01 5.77077985e-01
-6.55475080e-01 2.50376426e-02 6.38224542e-01 -2.67314762e-01
-9.06844258e-01 4.79173362e-01 -2.66992390e-01 2.68763769e-03
1.26754737e+00 -1.93794250e-01 -2.43240222e-01 -3.25358301e-01
-8.80822003e-01 4.82916802e-01 2.43431628e-01 6.59062028e-01
1.30067992e+00 -1.10127592e+00 -5.75028539e-01 2.18316466e-01
7.35636771e-01 1.08102605e-01 7.33648062e-01 1.66770184e+00
-1.10669398e+00 4.43745136e-01 1.47685498e-01 -8.51557910e-01
-1.17283177e+00 8.00881505e-01 6.54016018e-01 -5.27512789e-01
-7.96460986e-01 8.66543949e-01 9.32046652e-01 -5.93669951e-01
4.30918187e-01 -5.81762850e-01 -2.95710415e-01 -4.61134575e-02
3.01920056e-01 -4.12905693e-01 1.18831201e-02 -3.57348651e-01
-6.29004955e-01 5.45255244e-01 -2.50062421e-02 3.44469786e-01
8.73753190e-01 6.38007447e-02 8.42003226e-02 1.45830229e-01
1.36261523e+00 -3.83823812e-02 -8.81456256e-01 -3.16593260e-01
-4.25129771e-01 -2.80565321e-01 2.56601065e-01 -7.12125897e-01
-1.05072296e+00 1.01138711e+00 9.58402455e-01 -4.86276209e-01
1.08676779e+00 3.94378573e-01 7.66281426e-01 1.53959289e-01
4.22595024e-01 -7.04288483e-01 3.78610551e-01 3.46207827e-01
1.20748401e+00 -1.48829877e+00 -3.68867964e-01 -3.71963471e-01
-8.33398521e-01 1.09065008e+00 9.37872648e-01 1.67527556e-01
6.91976964e-01 3.67102891e-01 4.45463955e-01 -4.31390822e-01
-4.57061201e-01 -2.40871593e-01 5.71698546e-01 2.17528880e-01
2.14473799e-01 -4.54281308e-02 -1.54654995e-01 7.05170393e-01
5.06574772e-02 -2.71639489e-02 2.90538430e-01 9.71170127e-01
-2.09646091e-01 -5.24140120e-01 -2.68725395e-01 5.01219630e-01
-4.14463192e-01 -2.13357866e-01 -1.68209169e-02 7.07287550e-01
1.09610252e-01 9.11790013e-01 -2.94187486e-01 -3.69771153e-01
5.66999972e-01 -2.33007625e-01 1.65260538e-01 -6.80152893e-01
-7.77629614e-01 -1.04904979e-01 -1.85603231e-01 -1.11538076e+00
-6.60365261e-03 -7.97271729e-02 -1.32871938e+00 4.11237448e-01
-2.17632189e-01 3.09326798e-02 8.09033692e-01 5.89185119e-01
3.67081434e-01 8.53149772e-01 3.11391711e-01 -5.68526983e-01
-4.50056463e-01 -6.44253790e-01 -2.60300457e-01 1.66380003e-01
6.77912414e-01 -5.08148253e-01 -3.32900882e-01 -1.87643930e-01] | [14.156536102294922, -3.2407596111297607] |
f6627c01-37e8-44c6-b188-a7c2919bf702 | an-improved-model-for-voicing-silent-speech | 2106.01933 | null | https://arxiv.org/abs/2106.01933v2 | https://arxiv.org/pdf/2106.01933v2.pdf | An Improved Model for Voicing Silent Speech | In this paper, we present an improved model for voicing silent speech, where audio is synthesized from facial electromyography (EMG) signals. To give our model greater flexibility to learn its own input features, we directly use EMG signals as input in the place of hand-designed features used by prior work. Our model uses convolutional layers to extract features from the signals and Transformer layers to propagate information across longer distances. To provide better signal for learning, we also introduce an auxiliary task of predicting phoneme labels in addition to predicting speech audio features. On an open vocabulary intelligibility evaluation, our model improves the state of the art for this task by an absolute 25.8%. | ['Dan Klein', 'David Gaddy'] | 2021-06-03 | null | https://aclanthology.org/2021.acl-short.23 | https://aclanthology.org/2021.acl-short.23.pdf | acl-2021-5 | ['electromyography-emg'] | ['medical'] | [ 3.52426797e-01 4.52904642e-01 1.18027739e-01 -4.66046482e-01
-1.23008132e+00 -3.48211825e-01 7.93865323e-02 -6.24453247e-01
-2.95565635e-01 4.49947566e-01 6.53320730e-01 -7.51170516e-02
1.78322047e-01 -3.38485897e-01 -6.37569129e-01 -3.70733261e-01
-2.37494469e-01 -5.54814860e-02 -1.32341117e-01 -1.57308459e-01
-2.04362854e-01 3.60730290e-01 -1.52071047e+00 7.38589227e-01
5.76032400e-01 1.19729817e+00 1.85276344e-01 8.86165440e-01
2.59856790e-01 6.60002768e-01 -8.16024005e-01 -2.15642810e-01
6.99321330e-02 -4.36258495e-01 -9.22145784e-01 5.29605262e-02
4.85553980e-01 -4.15695578e-01 -4.14051354e-01 5.75244188e-01
8.13645840e-01 2.20984980e-01 6.07144356e-01 -1.02268600e+00
-5.95418155e-01 8.83252978e-01 1.35265887e-01 9.82787982e-02
5.23889959e-01 3.68914306e-01 1.26258552e+00 -9.45672333e-01
4.41177011e-01 1.14604235e+00 7.75648117e-01 9.38921988e-01
-1.08180809e+00 -8.12221169e-01 5.40435016e-02 3.96313518e-01
-1.20721209e+00 -9.48313475e-01 8.72237027e-01 -2.00633883e-01
1.45798087e+00 4.01591569e-01 6.09272838e-01 1.61292255e+00
-3.24026085e-02 8.65069091e-01 8.63251686e-01 -1.95723861e-01
-1.82848219e-02 -1.10348172e-01 -2.56995648e-01 4.46735710e-01
-7.44796753e-01 4.74497229e-01 -9.55080688e-01 2.21247256e-01
4.72867429e-01 -3.92499000e-01 -5.08925498e-01 3.30837727e-01
-9.55758870e-01 5.58572292e-01 4.99096155e-01 2.50597656e-01
-3.06663066e-01 2.96021461e-01 1.96890980e-01 6.91288173e-01
4.43310380e-01 5.01431346e-01 -5.86629927e-01 -7.51736939e-01
-8.53782594e-01 2.08709855e-02 7.72816479e-01 5.19887090e-01
2.58323669e-01 4.99471903e-01 -1.30929112e-01 1.08042789e+00
2.70722777e-01 4.13587689e-01 7.04201996e-01 -1.27404845e+00
5.07315934e-01 -1.92141812e-02 -3.49250078e-01 -4.64993149e-01
-4.74248171e-01 -6.84227169e-01 -3.58999997e-01 2.18495905e-01
1.76139191e-01 -2.45958701e-01 -9.83979404e-01 1.68088698e+00
-3.31904292e-01 3.37473571e-01 -2.68910415e-02 1.06380057e+00
8.78015816e-01 7.11770535e-01 -2.22790599e-01 -7.83670023e-02
9.46729422e-01 -1.08209813e+00 -8.38788569e-01 -1.10134974e-01
5.26495695e-01 -5.93689859e-01 1.26106012e+00 9.28254902e-01
-1.06703949e+00 -5.61947882e-01 -9.53589201e-01 -1.21085480e-01
-7.33758435e-02 1.38491392e-01 3.81361514e-01 5.36693215e-01
-1.30159521e+00 9.15603697e-01 -9.01667595e-01 -2.27210615e-02
3.85885119e-01 6.04248762e-01 -4.64424253e-01 3.29575866e-01
-1.27386487e+00 8.44048202e-01 -1.09598063e-01 -2.03610733e-01
-9.47742343e-01 -6.70958698e-01 -9.42943394e-01 9.34127495e-02
3.68822254e-02 -4.40360785e-01 1.40573716e+00 -9.31158960e-01
-2.17770815e+00 3.90547097e-01 -1.24589600e-01 -3.07264417e-01
1.74772471e-01 -3.45750630e-01 -6.75404072e-01 2.48375252e-01
-4.74698365e-01 8.95564198e-01 9.90237713e-01 -7.59387136e-01
-3.35229516e-01 -1.23493470e-01 -9.26815122e-02 2.17604145e-01
-5.59490144e-01 1.76529270e-02 -1.52512016e-02 -8.95600319e-01
-1.79676563e-01 -9.68609869e-01 2.89417475e-01 -1.34699002e-01
-1.22901402e-01 -4.13459331e-01 5.65459132e-01 -9.79502559e-01
1.09293365e+00 -2.25369549e+00 3.59151721e-01 9.87763256e-02
2.12920278e-01 2.55855769e-01 -4.85386372e-01 2.88591027e-01
-1.29879385e-01 -2.15125345e-02 -2.00688183e-01 -6.11245573e-01
7.32687861e-02 2.04787150e-01 -3.76626670e-01 2.25662038e-01
7.01363444e-01 8.53622258e-01 -6.88961685e-01 -6.58316389e-02
2.42861430e-03 7.70833671e-01 -9.20540452e-01 2.61240721e-01
-1.23531753e-02 4.99432683e-01 1.79137871e-01 5.60190797e-01
1.92941651e-01 1.95050746e-01 -2.45805174e-01 4.20173258e-02
2.45630801e-01 1.09676909e+00 -9.87463713e-01 2.07822037e+00
-1.00161910e+00 9.05214071e-01 1.37068227e-01 -7.42833614e-01
7.79428065e-01 8.43118131e-01 3.69697839e-01 -3.22826803e-01
1.06394872e-01 1.53343499e-01 2.65373051e-01 -6.06536448e-01
1.40082479e-01 -3.13925713e-01 1.80396542e-01 3.78856748e-01
4.86564189e-01 -2.39464536e-01 -5.16593158e-01 -2.85509557e-01
1.27398932e+00 1.11931391e-01 -2.49820992e-01 2.41334230e-01
2.90755361e-01 -4.83419150e-01 4.16285127e-01 1.49880782e-01
3.87489796e-02 9.25895333e-01 1.40799776e-01 1.61653623e-01
-6.01876915e-01 -1.37179077e+00 -1.53321683e-01 1.14699030e+00
-5.47576189e-01 -7.58805573e-01 -7.90113449e-01 -6.84024334e-01
-1.22984936e-02 6.02443933e-01 -3.09477538e-01 -4.12995875e-01
-4.50366110e-01 4.59715463e-02 8.83492589e-01 9.57606971e-01
-3.64230238e-02 -1.32412899e+00 -2.58272886e-01 1.53774157e-01
-4.17883337e-01 -1.06075966e+00 -7.47376919e-01 3.99920464e-01
-9.03907955e-01 -4.71481144e-01 -5.37086308e-01 -8.96236658e-01
6.03430206e-03 -3.03895324e-01 7.10160196e-01 -2.04122528e-01
-2.41655841e-01 2.22646564e-01 -4.21676934e-01 -5.37109792e-01
-4.20437604e-01 3.37374657e-01 4.03078854e-01 -9.27568451e-02
2.57627070e-01 -8.55715573e-01 -2.94992775e-01 9.72398669e-02
-8.23764682e-01 -2.73429036e-01 6.36029780e-01 1.02515209e+00
2.47729555e-01 -3.45903248e-01 9.51015472e-01 -1.65533915e-01
8.10749710e-01 -2.88505495e-01 7.91859403e-02 -2.99542636e-01
-4.16220784e-01 1.06028959e-01 6.15701377e-01 -6.35074675e-01
-6.75747275e-01 6.25872388e-02 -8.74977827e-01 -6.65222228e-01
-2.95854419e-01 3.55523229e-01 -3.28505456e-01 1.29377708e-01
5.64216137e-01 6.31401539e-02 2.68778175e-01 -7.59686708e-01
4.40136492e-01 1.51101947e+00 6.12540305e-01 -1.22271128e-01
4.43346441e-01 3.43082063e-02 -4.47747350e-01 -9.96759593e-01
-5.36517859e-01 -1.99369565e-01 -4.84420985e-01 -1.71695471e-01
6.64782763e-01 -8.30881298e-01 -7.87542522e-01 4.15228367e-01
-1.09689820e+00 -5.10627449e-01 -3.58107239e-01 8.44441831e-01
-8.95066917e-01 7.29667470e-02 -8.70841503e-01 -7.99371421e-01
-3.80858928e-01 -1.08479226e+00 1.23230588e+00 -3.07689041e-01
-6.44315541e-01 -6.49776578e-01 -1.56522527e-01 4.49528486e-01
5.61573148e-01 -3.11079592e-01 5.15610278e-01 -7.87996471e-01
-1.76988617e-01 1.14130855e-01 3.23983520e-01 9.50631261e-01
5.16340017e-01 -3.78816992e-01 -1.55178452e+00 -2.17368379e-01
9.77326035e-02 -5.37534177e-01 1.05911171e+00 1.55176491e-01
1.33741415e+00 -3.46498132e-01 3.65511440e-02 9.03898239e-01
5.67443430e-01 1.13078311e-01 5.94662249e-01 -1.81622073e-01
5.05157709e-01 6.01179898e-01 3.22153926e-01 2.85727829e-01
2.14756295e-01 7.85462439e-01 2.61215985e-01 6.45281002e-03
-6.29542887e-01 -3.39105517e-01 8.22718918e-01 1.36675847e+00
-6.00141250e-02 -1.50216268e-02 -6.21192515e-01 5.66392183e-01
-1.30814445e+00 -9.80336010e-01 6.05414093e-01 1.75194705e+00
1.21199167e+00 1.73325166e-01 2.59660095e-01 5.71478546e-01
2.09955975e-01 6.12344518e-02 -4.37730432e-01 -7.09597886e-01
1.79526702e-01 7.28010952e-01 -8.37744921e-02 7.65128911e-01
-8.67564917e-01 9.03158069e-01 7.29033947e+00 7.11213827e-01
-1.54163599e+00 1.10480651e-01 1.10153377e-01 -7.03103781e-01
-3.82248700e-01 -5.15975058e-01 -4.72750366e-01 3.59968156e-01
1.41962516e+00 3.54193985e-01 8.62625062e-01 4.95746821e-01
2.92149544e-01 4.73622680e-01 -1.25476742e+00 1.10766757e+00
1.95683539e-01 -9.41433966e-01 -2.02794135e-01 -1.68815181e-02
3.14944655e-01 2.79247314e-01 3.73822898e-01 3.19457620e-01
2.72119474e-02 -1.41029167e+00 5.93177378e-01 6.37008786e-01
1.05850816e+00 -6.88404024e-01 4.98062402e-01 1.06614493e-01
-8.56210709e-01 -1.62594765e-01 1.40771922e-02 -4.05474573e-01
1.48587480e-01 2.55102634e-01 -1.21190727e+00 1.61084950e-01
6.96351707e-01 7.87035823e-01 -1.15455084e-01 7.13645339e-01
-3.79825115e-01 1.10907292e+00 -4.87163246e-01 -3.61944251e-02
5.95727898e-02 3.28721195e-01 6.29628420e-01 1.21636367e+00
2.57047027e-01 -4.99673188e-02 -2.22420871e-01 7.13161826e-01
-2.66815782e-01 3.43431570e-02 -6.51730597e-01 -1.00162372e-01
4.33786780e-01 8.44610393e-01 2.55872697e-01 5.23345582e-02
-2.82774150e-01 1.19021118e+00 3.10245216e-01 3.79767776e-01
-6.20090306e-01 -9.37871099e-01 1.09019876e+00 1.73584297e-01
4.08491582e-01 -3.73995334e-01 -1.45696402e-01 -9.53483641e-01
2.80746967e-01 -9.84838247e-01 -3.65468152e-02 -8.39698017e-01
-1.04264724e+00 8.43821943e-01 -4.21717525e-01 -1.21555078e+00
-8.61331105e-01 -7.57992148e-01 -5.67786992e-01 1.08607137e+00
-1.51909709e+00 -9.10029948e-01 -1.08325738e-03 5.21002114e-01
8.18699837e-01 -2.66772658e-01 1.06622732e+00 3.84159088e-01
-1.19770356e-01 9.52730954e-01 -1.85946301e-01 2.89779425e-01
9.05638874e-01 -1.14900744e+00 6.91701114e-01 4.93460447e-01
6.53208315e-01 4.78258133e-01 4.90350366e-01 -2.75644451e-01
-1.18195117e+00 -9.54937339e-01 9.45545197e-01 -5.60157478e-01
7.63330996e-01 -4.95368451e-01 -6.70201004e-01 7.30450332e-01
1.94214404e-01 2.77284142e-02 9.11769152e-01 1.61185473e-01
-3.41694713e-01 -3.40294957e-01 -8.38202834e-01 4.50976253e-01
1.54989946e+00 -9.19331789e-01 -9.27538991e-01 -1.14772141e-01
1.04802978e+00 -3.72516394e-01 -1.06252766e+00 3.10000092e-01
7.23789215e-01 -4.15464997e-01 7.76121318e-01 -7.37673998e-01
2.95222014e-01 8.26557726e-02 -2.88269550e-01 -2.02395391e+00
-1.24843761e-01 -8.77665937e-01 -2.50873387e-01 1.03131425e+00
7.13862836e-01 -4.85220581e-01 5.80279171e-01 5.74044771e-02
-6.21201456e-01 -9.73795652e-01 -1.17799616e+00 -8.83286893e-01
1.46737784e-01 -9.03693676e-01 6.08516634e-01 4.99973387e-01
6.16755486e-01 3.47146928e-01 -3.60287517e-01 -3.19799781e-02
8.87833163e-02 -2.12489441e-01 2.17096478e-01 -1.08688068e+00
-5.30308843e-01 -3.91246378e-01 -5.62810421e-01 -1.24043798e+00
3.88760746e-01 -1.30768418e+00 2.80456990e-01 -1.40831983e+00
-4.37082171e-01 -1.30877703e-01 -5.31679690e-01 7.41902113e-01
3.92457843e-02 5.39564431e-01 3.44026029e-01 -3.11326414e-01
-2.36649558e-01 7.75930285e-01 1.14780068e+00 -3.01687390e-01
-3.21694672e-01 1.99198455e-01 -6.50294960e-01 5.12003779e-01
9.54935908e-01 -3.90240192e-01 -3.64311725e-01 -6.06782675e-01
-2.80338198e-01 6.86434209e-02 2.55655438e-01 -1.28466976e+00
-4.81263958e-02 3.07504594e-01 3.69668901e-01 -1.91764966e-01
9.45437551e-01 -6.40723228e-01 -2.53141165e-01 3.11763138e-01
-7.85713375e-01 -1.21809021e-01 2.66851276e-01 2.40670815e-01
-2.65468687e-01 7.66106546e-02 3.32279950e-01 2.60718107e-01
-3.43585074e-01 1.04013555e-01 -6.09885037e-01 -1.05634136e-02
3.54063153e-01 -5.97244166e-02 -4.77576293e-02 -8.45037222e-01
-1.35279632e+00 -4.17725518e-02 7.64201209e-02 9.71987307e-01
9.71448660e-01 -1.43309462e+00 -7.68230081e-01 6.20432258e-01
7.16493428e-02 -4.39112514e-01 -1.81727961e-01 7.91945517e-01
7.02942088e-02 4.86582905e-01 -2.52356678e-01 -5.93139887e-01
-1.24983585e+00 1.28755972e-01 3.57720166e-01 3.87347192e-01
-7.03715980e-01 1.07475066e+00 -3.19267929e-01 -5.64870775e-01
7.85920858e-01 -6.95954204e-01 -1.36638850e-01 -9.36233178e-02
5.39365768e-01 3.35759163e-01 4.46442336e-01 -5.02829909e-01
-3.52954537e-01 3.30906987e-01 2.88525760e-01 -4.99275029e-01
1.46880293e+00 -7.09778965e-02 4.11868989e-01 7.61186242e-01
1.43780243e+00 2.78577507e-01 -1.43225873e+00 -1.56022370e-01
-1.25026450e-01 -2.00677603e-01 2.56523609e-01 -1.24706376e+00
-1.18584347e+00 1.22311366e+00 6.27098501e-01 -2.12988362e-01
1.22454548e+00 2.39143506e-01 1.10598838e+00 3.41874897e-01
1.37413844e-01 -1.09642148e+00 1.93344802e-01 4.72809732e-01
1.13210988e+00 -9.42954421e-01 -6.58812940e-01 -2.75389463e-01
-8.19587648e-01 1.22425854e+00 3.43291461e-01 -4.30480652e-02
7.87697256e-01 7.37278163e-01 3.56255412e-01 1.63424522e-01
-9.33874965e-01 -3.03055465e-01 6.42719150e-01 9.39001143e-01
6.63326204e-01 1.30173236e-01 -2.56242529e-02 1.07086504e+00
-8.07817578e-01 1.29226983e-01 2.35549197e-01 4.14828181e-01
-4.50731933e-01 -1.06078184e+00 -1.66518167e-01 6.03557706e-01
-5.92217684e-01 -3.00604105e-01 -5.07295310e-01 3.32118124e-01
-7.52487630e-02 1.35271454e+00 1.94426090e-01 -1.01682746e+00
4.75833148e-01 4.53297675e-01 5.64729631e-01 -7.10635900e-01
-8.06402624e-01 -1.08938040e-02 3.73704672e-01 -7.95338988e-01
-2.50781011e-02 -5.90331852e-01 -1.46987760e+00 1.68487698e-01
-2.70961732e-01 1.73839759e-02 8.07668328e-01 1.01017284e+00
4.31523383e-01 9.94619787e-01 7.46362269e-01 -8.83946121e-01
-9.21321929e-01 -1.41371012e+00 -5.12338936e-01 3.47339988e-01
7.67255306e-01 -5.20964384e-01 -5.82463145e-01 -1.69072915e-02] | [15.02472972869873, 5.934204578399658] |
ca47afd9-5827-40a4-92b2-89429e984e82 | improving-transition-based-dependency-parsing | null | null | https://aclanthology.org/D12-1029 | https://aclanthology.org/D12-1029.pdf | Improving Transition-Based Dependency Parsing with Buffer Transitions | null | ["Carlos G{\\'o}mez-Rodr{\\'\\i}guez", "Daniel Fern{\\'a}ndez-Gonz{\\'a}lez"] | 2012-07-01 | null | null | null | emnlp-2012-7 | ['transition-based-dependency-parsing'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.3037567138671875, 3.831505060195923] |
70ab8713-1344-4774-8230-af50157df4b7 | efficient-similarity-based-passive-filter | 2210.17416 | null | https://arxiv.org/abs/2210.17416v1 | https://arxiv.org/pdf/2210.17416v1.pdf | Efficient Similarity-based Passive Filter Pruning for Compressing CNNs | Convolution neural networks (CNNs) have shown great success in various applications. However, the computational complexity and memory storage of CNNs is a bottleneck for their deployment on resource-constrained devices. Recent efforts towards reducing the computation cost and the memory overhead of CNNs involve similarity-based passive filter pruning methods. Similarity-based passive filter pruning methods compute a pairwise similarity matrix for the filters and eliminate a few similar filters to obtain a small pruned CNN. However, the computational complexity of computing the pairwise similarity matrix is high, particularly when a convolutional layer has many filters. To reduce the computational complexity in obtaining the pairwise similarity matrix, we propose to use an efficient method where the complete pairwise similarity matrix is approximated from only a few of its columns by using a Nystr\"om approximation method. The proposed efficient similarity-based passive filter pruning method is 3 times faster and gives same accuracy at the same reduction in computations for CNNs compared to that of the similarity-based pruning method that computes a complete pairwise similarity matrix. Apart from this, the proposed efficient similarity-based pruning method performs similarly or better than the existing norm-based pruning methods. The efficacy of the proposed pruning method is evaluated on CNNs such as DCASE 2021 Task 1A baseline network and a VGGish network designed for acoustic scene classification. | ['Mark D. Plumbley', 'Arshdeep Singh'] | 2022-10-27 | null | null | null | null | ['scene-classification'] | ['computer-vision'] | [ 2.45793208e-01 -2.17954427e-01 7.54412830e-01 -4.78668839e-01
-3.05617124e-01 -2.32871369e-01 1.75903678e-01 2.90353298e-01
-1.13249350e+00 4.35001463e-01 -3.93192738e-01 -3.98634642e-01
-5.37048638e-01 -1.08827913e+00 -6.64637566e-01 -6.97560668e-01
4.97373343e-02 -1.31367773e-01 7.63372838e-01 -1.20870462e-02
2.53582835e-01 6.18817270e-01 -2.00524092e+00 2.45946839e-01
6.04446173e-01 1.55208266e+00 5.12111127e-01 6.72431529e-01
-1.09322399e-01 2.28117213e-01 -5.24632633e-01 -4.85165238e-01
7.12468565e-01 -1.50382847e-01 -4.69724894e-01 -5.63410699e-01
8.66730869e-01 -1.41957611e-01 -4.45690572e-01 1.40266621e+00
5.70955455e-01 6.97775424e-01 2.57455379e-01 -8.25961471e-01
5.84438220e-02 5.96036017e-01 -2.98243344e-01 5.02532423e-01
-3.15647870e-01 -4.30296689e-01 7.69562721e-01 -9.88584578e-01
4.28528577e-01 9.12104547e-01 1.24108076e+00 2.94717133e-01
-8.61013830e-01 -1.01491892e+00 6.30513281e-02 2.54185945e-01
-1.75384045e+00 -4.94030058e-01 5.28454781e-01 -3.22404765e-02
1.38076866e+00 2.92094111e-01 7.64392197e-01 2.29628280e-01
1.26386611e-02 1.93313852e-01 5.86984992e-01 -4.86229539e-01
2.89555371e-01 -1.42212123e-01 3.99780333e-01 1.08380127e+00
5.87992549e-01 -2.69962400e-01 -5.68991899e-01 -1.56415224e-01
3.62953216e-01 1.87078476e-01 -8.84471014e-02 1.38891518e-01
-6.14608467e-01 6.57930672e-01 4.75617588e-01 4.45708036e-01
-1.49590895e-01 4.18009460e-01 5.55353940e-01 4.16101187e-01
3.72931093e-01 1.25433326e-01 -4.36313748e-01 7.31683150e-02
-1.48544979e+00 4.48047966e-01 8.33547413e-01 9.87154305e-01
7.84921408e-01 2.98091054e-01 1.08244844e-01 1.07804191e+00
2.33586699e-01 2.54000485e-01 4.14882094e-01 -5.57838619e-01
5.60505867e-01 5.63993990e-01 -4.02025014e-01 -1.39113235e+00
-4.18099672e-01 -6.42709434e-01 -1.02377260e+00 3.15014064e-01
3.35910857e-01 -2.88345456e-01 -9.62942183e-01 1.60403419e+00
1.47638738e-01 4.44613576e-01 1.13841176e-01 5.42970479e-01
1.10502768e+00 7.25660741e-01 -1.38353616e-01 -1.74223199e-01
1.25374460e+00 -6.54093385e-01 -3.83439600e-01 -2.95882840e-02
6.10937595e-01 -1.18004501e+00 4.09772754e-01 5.17646551e-01
-1.12731338e+00 -7.06731558e-01 -1.43375731e+00 -3.39317508e-03
-4.48170424e-01 5.19801497e-01 3.34652781e-01 7.87611961e-01
-1.06525815e+00 1.21955061e+00 -8.42777610e-01 -2.90746480e-01
6.98782027e-01 7.80709088e-01 -2.64530808e-01 1.07170939e-02
-7.88869143e-01 5.49836814e-01 5.84169924e-01 1.81219041e-01
-2.45383412e-01 -7.93654561e-01 -5.57856500e-01 4.47783440e-01
5.69490418e-02 -3.54838401e-01 1.04157460e+00 -4.26042050e-01
-1.20647764e+00 2.60260344e-01 -1.24581337e-01 -8.60597789e-01
2.27422148e-01 -2.93874323e-01 -3.27211320e-01 1.11604914e-01
-3.28097999e-01 6.67211950e-01 7.35260010e-01 -4.00367111e-01
-9.53266621e-01 -1.42540574e-01 -2.13791117e-01 2.24566132e-01
-7.13892996e-01 -8.96332115e-02 -3.25800180e-01 -7.74021149e-01
7.33279705e-01 -7.24986732e-01 -3.52522880e-01 2.08890408e-01
-2.22266510e-01 -1.35332108e-01 1.02506793e+00 -3.75894099e-01
1.24077117e+00 -2.33146429e+00 -3.61812115e-01 3.76166672e-01
2.69617319e-01 8.86574566e-01 -1.05376810e-01 1.51485682e-01
5.08663952e-02 3.04653384e-02 -9.60626230e-02 -5.52664638e-01
-5.38926363e-01 2.39495710e-01 -1.03504688e-01 3.67785305e-01
1.42526656e-01 3.69383441e-03 -5.84598780e-01 -4.84774381e-01
4.90643121e-02 6.77794576e-01 -8.78684163e-01 -1.49345517e-01
2.54070461e-01 -4.34394747e-01 -5.19296825e-02 2.98838228e-01
1.08289039e+00 3.67688626e-01 -7.99226202e-03 -8.13237965e-01
-4.50029552e-01 2.83133000e-01 -1.59557939e+00 1.41660500e+00
-3.55955899e-01 1.01367259e+00 -4.51944321e-02 -1.24816155e+00
1.27964008e+00 3.51922274e-01 3.06680769e-01 -4.22227919e-01
4.76231188e-01 3.54412079e-01 3.97408068e-01 -1.64481699e-01
5.79938173e-01 -1.88712552e-02 5.34395933e-01 9.00339335e-02
4.16536152e-01 3.52687873e-02 3.83945942e-01 -5.47529683e-02
1.30621135e+00 -3.21091622e-01 1.40602976e-01 -5.87624192e-01
7.03824878e-01 -1.79230720e-01 8.89675260e-01 7.61704683e-01
-8.03347491e-03 5.10721207e-01 1.41791254e-01 -6.44318759e-01
-1.00684464e+00 -8.25241745e-01 -3.04325521e-01 6.63095057e-01
-3.55851874e-02 -7.14467108e-01 -8.38398397e-01 -2.08206326e-01
-1.46339536e-01 1.51799873e-01 -2.15004340e-01 -2.17341751e-01
-8.06999087e-01 -6.72925353e-01 9.64491963e-01 5.91553986e-01
1.05267453e+00 -6.19866133e-01 -9.48974252e-01 2.41065189e-01
2.38610595e-01 -1.14809620e+00 -1.13125496e-01 6.63266420e-01
-1.36477208e+00 -1.04252434e+00 -4.02697682e-01 -1.01045728e+00
6.41806006e-01 4.00211304e-01 5.49129426e-01 2.40336493e-01
-4.79972750e-01 -3.49362403e-01 -3.18016768e-01 -5.84854186e-01
1.56373739e-01 8.88282806e-03 9.31098983e-02 -1.68339685e-01
4.16975230e-01 -8.46786439e-01 -6.32135570e-01 1.33521408e-01
-5.79008043e-01 -2.20168367e-01 6.00214839e-01 8.43623221e-01
7.93162107e-01 5.52133381e-01 2.65176684e-01 -9.28358972e-01
6.29873335e-01 -2.15691607e-02 -8.40980530e-01 -9.18296576e-02
-4.53062862e-01 3.27327810e-02 9.59559083e-01 -5.81131637e-01
-1.06209576e+00 3.25670481e-01 -3.93961906e-01 -4.71538872e-01
2.17682179e-02 3.85416240e-01 2.46798128e-01 -5.86721539e-01
5.88321388e-01 1.08866692e-01 -2.35453859e-01 -8.05194795e-01
-1.40871793e-01 3.41675639e-01 5.46287775e-01 -2.22661808e-01
8.67403209e-01 3.82460058e-01 3.66410404e-01 -1.22622168e+00
-4.69598204e-01 -5.25622666e-01 -4.77011770e-01 3.40538211e-02
5.33815682e-01 -7.64007211e-01 -8.58865738e-01 3.71773392e-01
-1.34369063e+00 2.71217257e-01 -1.27441630e-01 7.52347887e-01
1.61572993e-02 1.73488259e-01 -4.83480960e-01 -7.66207814e-01
-7.10421920e-01 -1.03081536e+00 5.03942728e-01 3.20589572e-01
-9.78118181e-02 -3.85016143e-01 -2.37361595e-01 -2.45827153e-01
4.66159225e-01 4.15283628e-02 1.00401056e+00 -7.95428634e-01
-5.27315199e-01 -4.05024379e-01 -5.40849686e-01 6.65538430e-01
-2.58133501e-01 4.03475240e-02 -9.54873502e-01 -5.81158474e-02
1.07950062e-01 1.28364623e-01 1.08723426e+00 3.57723773e-01
1.15813196e+00 -8.26517045e-02 -2.63019681e-01 8.89964104e-01
1.76453078e+00 5.94923794e-01 3.86103779e-01 3.10448706e-02
4.29230720e-01 1.98380664e-01 4.51419920e-01 4.78085577e-01
-4.55360264e-01 2.06140608e-01 2.94225663e-01 6.07107356e-02
-2.22661585e-01 1.11550510e-01 -2.55884528e-01 1.18651080e+00
-2.56795585e-01 -5.90338698e-03 -6.90907717e-01 4.49646294e-01
-1.70775139e+00 -9.96764421e-01 -2.60703564e-01 2.14206624e+00
5.26795447e-01 4.04003322e-01 -4.39197063e-01 7.94218600e-01
4.24783438e-01 -1.38547663e-02 -1.86135724e-01 -8.34838033e-01
-4.64723706e-02 8.47351015e-01 6.16101861e-01 2.84477770e-01
-1.08422637e+00 7.56104410e-01 5.80348349e+00 9.29050326e-01
-8.75347316e-01 1.34365708e-01 2.05052465e-01 -5.01870811e-01
4.37723249e-01 -1.43792465e-01 -1.25786602e+00 7.87474141e-02
8.45380127e-01 1.52789727e-01 1.09401837e-01 1.08503044e+00
2.59598177e-02 -3.70238751e-01 -9.11995173e-01 1.22037971e+00
-1.15166791e-01 -1.65617716e+00 5.46802953e-02 -2.74717838e-01
5.37248015e-01 2.53415823e-01 -1.59477592e-01 3.06488797e-02
-2.62424499e-01 -6.51598215e-01 8.13215852e-01 4.81593966e-01
4.51316267e-01 -1.21836531e+00 9.94403660e-01 1.97753504e-01
-1.65459371e+00 -1.43029287e-01 -1.01573789e+00 -3.00920874e-01
-2.49914959e-01 9.00437295e-01 -6.07117534e-01 1.60812601e-01
1.24307859e+00 3.18952233e-01 -4.13100868e-01 1.46690238e+00
5.81168592e-01 4.46441531e-01 -6.98057711e-01 -2.77821720e-01
4.13430780e-01 -2.88216025e-01 4.90700424e-01 1.48261070e+00
7.04875708e-01 2.25431874e-01 -3.17706585e-01 5.27799070e-01
-1.99265286e-01 2.30623767e-01 -6.10187709e-01 4.74907942e-02
6.36487484e-01 1.09989619e+00 -1.09489167e+00 -2.89683461e-01
-5.17494500e-01 6.28027380e-01 3.64497632e-01 2.83927340e-02
-5.09338260e-01 -1.15766346e+00 1.05381370e+00 -3.81308571e-02
6.01751029e-01 -3.50070179e-01 -3.28296691e-01 -5.32139480e-01
1.37779891e-01 -6.27951205e-01 1.58705518e-01 -3.55703562e-01
-5.32402039e-01 1.16323662e+00 -7.73810782e-03 -1.23997140e+00
2.04423636e-01 -7.11251915e-01 -6.74460292e-01 9.37638521e-01
-1.15945852e+00 -5.79094768e-01 -4.00251567e-01 6.62389338e-01
5.07469773e-01 -4.05033797e-01 7.29671955e-01 7.11248279e-01
-5.32418549e-01 8.57357740e-01 5.68266436e-02 1.41233742e-01
3.23764265e-01 -6.18563533e-01 2.41051853e-01 1.06627738e+00
2.04140738e-01 9.26189721e-01 5.89879036e-01 -4.88367289e-01
-1.00283933e+00 -1.09116447e+00 1.01273739e+00 6.47864997e-01
2.04726294e-01 -3.36951703e-01 -9.00584340e-01 2.46765107e-01
4.27672639e-02 3.37393850e-01 6.45257056e-01 5.20167351e-02
-3.34782898e-01 -5.61078191e-01 -1.23757553e+00 6.47360682e-01
1.23709714e+00 -2.93911278e-01 -3.45031291e-01 1.80140942e-01
5.59385240e-01 -3.84914637e-01 -6.12662196e-01 4.23818976e-01
7.52849638e-01 -1.23390687e+00 9.66129184e-01 -1.52668074e-01
-5.73153142e-03 -4.42342073e-01 -4.89792049e-01 -8.07267427e-01
-1.32265106e-01 -5.01932025e-01 2.87327349e-01 9.82650578e-01
3.04806918e-01 -4.91686076e-01 1.04172134e+00 2.36197278e-01
-4.13851708e-01 -9.70456362e-01 -1.09107625e+00 -9.81900096e-01
-3.40892285e-01 -7.52468765e-01 4.13791478e-01 4.68174577e-01
-4.15265828e-01 -9.13372040e-02 6.19715713e-02 3.24287623e-01
5.79090059e-01 -2.15903088e-01 3.51055801e-01 -1.47978079e+00
-1.96747467e-01 -4.10158962e-01 -8.80281448e-01 -7.71998465e-01
-3.03567350e-01 -7.90866256e-01 9.39694494e-02 -1.10461891e+00
-2.14288309e-01 -6.29100084e-01 -1.80463105e-01 2.97900558e-01
4.07256395e-01 5.48905611e-01 2.57183075e-01 -1.64839715e-01
-8.13232437e-02 2.90104419e-01 7.14803696e-01 9.83532146e-02
-3.09432477e-01 1.89779878e-01 -2.31619254e-01 1.18715429e+00
7.21720040e-01 -8.82134199e-01 -6.25273824e-01 -6.48587048e-01
3.01901996e-01 -3.73188853e-01 1.84223413e-01 -1.81601810e+00
8.75854969e-01 3.95612836e-01 3.86886567e-01 -9.88850236e-01
6.63254023e-01 -8.93253207e-01 1.43601105e-01 7.85782158e-01
-1.14458770e-01 2.58421421e-01 5.72525203e-01 5.96714258e-01
-4.47595537e-01 -8.14857364e-01 1.06902444e+00 7.23134866e-03
-5.49259543e-01 1.74902290e-01 -5.91456354e-01 -2.51152843e-01
6.13242090e-01 -6.83549345e-01 6.65185302e-02 2.05219328e-01
-7.92532504e-01 -4.08256799e-01 -3.47910047e-01 -6.76919892e-02
9.78902578e-01 -1.18897986e+00 -2.78795272e-01 3.56539845e-01
-4.36660081e-01 2.04675511e-01 2.16542691e-01 6.62844419e-01
-8.93765450e-01 4.00372505e-01 -3.75244468e-01 -4.35287029e-01
-1.89118552e+00 -1.52759373e-01 3.66375029e-01 4.53621754e-03
-7.74492204e-01 1.42007875e+00 -2.54494786e-01 2.51431651e-02
5.46602428e-01 -7.87146449e-01 -1.84746772e-01 3.92287597e-02
6.02145493e-01 7.21261919e-01 6.58388317e-01 -3.34392548e-01
-4.31167454e-01 7.85874486e-01 -8.06248635e-02 7.03901798e-02
1.49658346e+00 2.96536714e-01 -1.85697556e-01 -7.72272646e-02
1.49492836e+00 -3.09934586e-01 -7.75548697e-01 -4.43042815e-01
-1.12033583e-01 -4.60965872e-01 3.37384403e-01 -6.28576279e-02
-1.45243704e+00 8.81237209e-01 8.68265986e-01 -1.49718761e-01
1.44403720e+00 -6.47905707e-01 9.24441934e-01 1.06553197e+00
1.46614406e-02 -1.22476757e+00 -1.69838876e-01 7.81275749e-01
5.94530761e-01 -8.59722078e-01 2.43292645e-01 -7.16206312e-01
1.55293211e-01 1.44622695e+00 6.31045520e-01 -6.49709225e-01
1.47812188e+00 5.89466572e-01 -3.54287624e-01 -1.31015688e-01
-6.44013584e-01 -8.81676599e-02 3.26881319e-01 5.01722097e-01
4.70201939e-01 -2.93945491e-01 -6.32445157e-01 5.34179449e-01
-5.11818647e-01 -2.63004422e-01 2.11839691e-01 9.50699925e-01
-6.43921733e-01 -9.15786028e-01 -5.77939749e-02 7.82410383e-01
-4.08230752e-01 -4.81317431e-01 8.36469680e-02 7.42332458e-01
7.09962904e-01 9.58286941e-01 4.61178184e-01 -8.15046072e-01
4.03137147e-01 1.24424674e-01 3.96284968e-01 -6.36636317e-01
-9.34607327e-01 3.26206572e-02 1.07377358e-01 -5.76471806e-01
-3.11973572e-01 -4.92573053e-01 -1.25125921e+00 -4.13718581e-01
-5.32622993e-01 -1.11711165e-03 1.05419683e+00 9.81416285e-01
2.44463131e-01 5.91231525e-01 1.36655897e-01 -6.50464892e-01
-4.97730136e-01 -7.66986430e-01 -4.57553864e-01 1.88876800e-02
2.88344398e-02 -5.45364559e-01 -7.25873634e-02 -1.39107302e-01] | [8.537041664123535, 3.024709939956665] |
46e06b5d-daef-4ef9-8fa9-e1db9aecedcf | data-fusion-for-radio-frequency-slam-with | 2206.09746 | null | https://arxiv.org/abs/2206.09746v1 | https://arxiv.org/pdf/2206.09746v1.pdf | Data Fusion for Radio Frequency SLAM with Robust Sampling | Precise indoor localization remains a challenging problem for a variety of essential applications. A promising approach to address this problem is to exchange radio signals between mobile agents and static physical anchors (PAs) that bounce off flat surfaces in the indoor environment. Radio frequency simultaneous localization and mapping (RF-SLAM) methods can be used to jointly estimates the time-varying location of agents as well as the static locations of the flat surfaces. Recent work on RF-SLAM methods has shown that each surface can be efficiently represented by a single master virtual anchor (MVA). The measurement model related to this MVA-based RF-SLAM method is highly nonlinear. Thus, Bayesian estimation relies on sampling-based techniques. The original MVA-based RF-SLAM method employs conventional "bootstrap" sampling. In challenging scenarios it was observed that the original method might converge to incorrect MVA positions corresponding to local maxima. In this paper, we introduce MVA-based RF-SLAM with an improved sampling technique that succeeds in the aforementioned challenging scenarios. Our simulation results demonstrate significant performance advantages. | ['Florian Meyer', 'Mingchao Liang', 'Wenyu Zhang', 'Bryan Teague', 'Erik Leitinger'] | 2022-06-20 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [ 1.08667873e-01 -1.14922293e-01 1.07503958e-01 -2.31638998e-01
-1.18269312e+00 -6.54941022e-01 6.74942315e-01 1.47807643e-01
-4.66853797e-01 1.40180254e+00 -4.16229397e-01 -1.41347408e-01
-3.16938519e-01 -8.57841611e-01 -9.09416378e-01 -8.11483324e-01
-5.50651550e-01 5.70096374e-01 3.46132129e-01 -1.00087270e-01
3.37225497e-01 5.48274577e-01 -1.14690113e+00 -5.45739055e-01
8.74662220e-01 7.46972144e-01 6.92400217e-01 8.71956468e-01
-6.47150874e-02 7.59043768e-02 -1.10120857e+00 3.25797826e-01
-4.49175872e-02 -6.93267677e-03 -1.92913845e-01 -7.73163557e-01
1.80933401e-01 7.86052942e-02 8.31532404e-02 7.77347863e-01
5.40492058e-01 1.17745213e-01 4.19391245e-01 -1.38819051e+00
2.24936470e-01 2.19935998e-01 -6.12074971e-01 -3.28803472e-02
9.84504342e-01 -8.04523349e-01 2.26081491e-01 -7.10527301e-01
2.94699997e-01 8.88367891e-01 1.31150365e+00 9.24236551e-02
-1.18596268e+00 -7.18300700e-01 -2.91086137e-02 -2.18446195e-01
-2.03363752e+00 -4.96820956e-01 6.21811450e-01 -1.89918391e-02
4.04502809e-01 4.17060643e-01 5.71722984e-01 1.08336270e+00
8.87743950e-01 1.89847693e-01 1.23077011e+00 -4.92162287e-01
7.56197870e-01 8.68267119e-02 -2.35403866e-01 5.72967231e-01
9.53729391e-01 -8.44429657e-02 -8.29145372e-01 -7.81255484e-01
7.53628969e-01 -3.56100239e-02 -5.28264046e-01 -6.95352376e-01
-1.37968564e+00 5.52899718e-01 6.13016486e-01 4.29179966e-01
-6.03429317e-01 6.92133904e-01 -4.81832296e-01 1.04134344e-01
1.52332723e-01 5.11255205e-01 -3.08812559e-01 -2.17297226e-01
-8.85908961e-01 2.37054855e-01 9.58622992e-01 1.21262121e+00
8.16582024e-01 -1.86005563e-01 4.19324696e-01 3.41792345e-01
8.16907823e-01 1.29748452e+00 2.66032778e-02 -8.35462689e-01
3.84480208e-01 -2.01130942e-01 1.13820159e+00 -1.21011150e+00
-7.00808406e-01 -7.03522325e-01 -6.29945457e-01 -3.35592851e-02
3.93643796e-01 -4.60768908e-01 -6.24462068e-01 1.74108529e+00
6.69844508e-01 6.15944326e-01 1.96555078e-01 4.74283546e-01
2.32893556e-01 5.70008159e-01 -3.34514648e-01 -3.05630326e-01
9.64619875e-01 -3.75308841e-01 -9.27532613e-01 -4.98476446e-01
4.25306916e-01 -8.95272255e-01 4.16417688e-01 2.70324677e-01
-6.39153957e-01 -2.93427408e-01 -1.45927715e+00 8.43448281e-01
-1.44576818e-01 -1.94714606e-01 6.58728659e-01 8.34396541e-01
-1.32580554e+00 1.15615554e-01 -1.15289271e+00 -4.47984964e-01
-2.36700952e-01 6.43210769e-01 -2.85768602e-02 -8.02289248e-02
-1.10613489e+00 8.62669945e-01 -4.13932294e-01 4.02522624e-01
-6.47611737e-01 -3.90664041e-01 -8.69797826e-01 -6.65245414e-01
4.51633334e-02 -7.65160084e-01 1.12733281e+00 -1.51676938e-01
-1.71649504e+00 -3.89793366e-02 -9.57714856e-01 -5.68029225e-01
4.79749501e-01 -2.42461100e-01 -4.15601015e-01 -1.53172761e-01
6.29872501e-01 -2.97634751e-02 4.91813689e-01 -1.80515158e+00
-5.76079786e-01 -3.11066359e-01 -2.28211924e-01 1.30894944e-01
5.82253993e-01 -6.91303551e-01 1.77083572e-03 -1.88145146e-01
1.02752590e+00 -1.22380567e+00 -4.75819528e-01 -4.12127942e-01
-2.97182590e-01 2.63468504e-01 4.82210785e-01 -1.09054323e-03
6.96683705e-01 -1.79806221e+00 -4.27384615e-01 6.71068847e-01
-2.14879945e-01 -6.34135067e-01 2.78085530e-01 7.23849475e-01
7.60823846e-01 -1.46282002e-01 9.63583961e-02 -6.73627675e-01
-1.58385634e-01 2.29721844e-01 -3.88919562e-01 1.19494569e+00
-7.82255590e-01 4.49366719e-01 -1.25264633e+00 -3.07140440e-01
2.87829876e-01 9.09129858e-01 -1.77834496e-01 -1.34862766e-01
4.93636936e-01 8.92003119e-01 -7.71280944e-01 7.27872968e-01
1.10388339e+00 -9.51969326e-02 3.10507238e-01 7.66491964e-02
-4.50169265e-01 3.18439722e-01 -1.48193264e+00 1.96382380e+00
-9.89749372e-01 6.18590713e-01 5.06246626e-01 -3.63368601e-01
1.15564442e+00 2.69360185e-01 5.51473856e-01 -5.28733671e-01
-2.60665983e-01 7.50282168e-01 -5.76064050e-01 8.28193650e-02
7.79207408e-01 6.81430101e-02 -3.74769866e-01 2.08293032e-02
-4.47226048e-01 -2.41310015e-01 -5.90533376e-01 -6.77055819e-03
1.39049995e+00 2.37195879e-01 5.02210736e-01 -5.74839413e-01
7.13123620e-01 -3.80290933e-02 5.70466757e-01 1.22027290e+00
-3.19231339e-02 1.83124989e-01 -5.77925682e-01 -1.00822687e-01
-2.29731649e-01 -1.28596818e+00 -3.83587807e-01 4.74141896e-01
1.18708897e+00 -3.55843455e-01 -3.58750492e-01 -1.52182028e-01
3.00859272e-01 4.83601421e-01 -4.36834246e-01 2.99683005e-01
-7.25907326e-01 -6.77112341e-01 3.11663389e-01 -1.22033022e-01
5.68336427e-01 -3.49825248e-02 -7.66134918e-01 5.99005640e-01
-5.23609817e-01 -9.11685705e-01 1.88805148e-01 2.37633675e-01
-5.28680563e-01 -8.16698790e-01 -6.45155370e-01 -4.55544204e-01
8.60745668e-01 1.03044200e+00 6.70411587e-01 -7.33741745e-02
2.88190424e-01 5.33820689e-01 -2.63481975e-01 -5.04875779e-01
-1.03452526e-01 3.08143515e-02 6.32834613e-01 1.01499029e-01
-3.87491938e-03 -7.15699613e-01 -7.17090726e-01 8.61106575e-01
1.72170341e-01 -3.20713669e-01 4.97767687e-01 4.22481567e-01
8.28029335e-01 -5.23393042e-02 6.22907877e-01 -3.13942313e-01
2.71468401e-01 -7.20224977e-01 -1.02052927e+00 1.31972834e-01
-3.71874064e-01 -7.57255405e-02 3.29247355e-01 -2.69422621e-01
-6.14797533e-01 2.91403770e-01 3.27691138e-02 3.29510868e-01
7.60433078e-02 1.95935294e-01 -1.23433292e-01 -9.88095641e-01
6.36073887e-01 2.83465385e-01 -3.66154164e-01 -2.89437443e-01
-7.99884200e-02 7.64990568e-01 4.46763247e-01 -5.40522575e-01
1.06702185e+00 9.16121840e-01 4.78054136e-01 -1.02693808e+00
-5.58506668e-01 -5.52156568e-01 -2.42261633e-01 -3.68992090e-01
2.65293837e-01 -1.05293775e+00 -8.84484649e-01 1.53228685e-01
-1.21010923e+00 -1.22924827e-01 1.81139365e-01 9.17501211e-01
-5.84110975e-01 3.34669381e-01 9.91308913e-02 -1.32784963e+00
1.21941715e-01 -1.02753603e+00 1.30593514e+00 3.18006575e-01
-3.23830485e-01 -1.00465274e+00 6.29401147e-01 -1.18860848e-01
8.06895435e-01 5.87506592e-01 -1.55489787e-01 -1.51232183e-01
-8.71584833e-01 -5.65486312e-01 3.80667388e-01 -9.77999449e-01
4.15146023e-01 -7.54772007e-01 -8.03840697e-01 -7.73652494e-01
2.78188616e-01 4.42327082e-01 2.17914373e-01 8.15937757e-01
1.48718640e-01 9.34275426e-03 -1.12516844e+00 7.16793180e-01
1.55251300e+00 2.64839679e-01 3.43263984e-01 7.05582440e-01
2.52520114e-01 -6.33746833e-02 1.05405438e+00 3.51618916e-01
6.66159332e-01 1.12669754e+00 6.32861674e-01 1.57457918e-01
2.85056531e-01 -3.55568469e-01 3.27089429e-01 4.40121680e-01
3.94375771e-02 -4.97387290e-01 -6.84463620e-01 3.82834107e-01
-1.85916519e+00 -5.73679745e-01 -4.69961435e-01 2.53639722e+00
1.87997594e-01 -7.02126026e-02 -6.51530087e-01 3.34094465e-02
8.73677135e-01 2.65533686e-01 -3.26653004e-01 3.92402202e-01
1.95227817e-01 -2.59931087e-01 1.33552182e+00 1.18720901e+00
-9.49544549e-01 5.03005028e-01 6.23713970e+00 3.10802162e-01
-8.79944563e-01 3.66665393e-01 -4.96304065e-01 4.53081846e-01
-2.73204774e-01 1.70502678e-01 -1.04081488e+00 5.55458665e-01
1.00182867e+00 4.89282049e-02 2.82382280e-01 7.90193737e-01
3.08057517e-01 -8.42995942e-01 -5.45559168e-01 1.00611627e+00
-1.82603240e-01 -1.23943162e+00 -5.39095581e-01 4.72425342e-01
7.72102892e-01 3.19533408e-01 -1.26593590e-01 -4.73139808e-02
4.02133077e-01 -5.30950069e-01 7.05416858e-01 5.31940103e-01
5.35456538e-01 -5.83082616e-01 1.02352977e+00 3.97874027e-01
-1.70667875e+00 1.81063667e-01 -3.89516473e-01 -3.86045873e-02
6.22350752e-01 8.18156898e-01 -1.27695966e+00 8.65512490e-01
4.95490193e-01 1.54608309e-01 -1.20967902e-01 1.37998974e+00
-3.08261037e-01 2.25484103e-01 -9.76942182e-01 -3.78573865e-01
6.06795922e-02 -6.81082020e-03 8.63093615e-01 7.42314637e-01
9.16726351e-01 -4.35107738e-01 2.51217753e-01 4.32247758e-01
4.78864372e-01 -2.37753913e-01 -8.25264335e-01 9.26266670e-01
1.21338689e+00 8.96059990e-01 -7.48760343e-01 6.72429204e-02
3.06986901e-03 7.33408272e-01 -2.61087686e-01 5.99541485e-01
-9.32972372e-01 -5.37973464e-01 4.86089885e-01 2.98314959e-01
-1.02394330e-03 -1.07076609e+00 9.63965729e-02 -9.95454788e-01
-6.29711300e-02 -2.19973586e-02 -3.05827886e-01 -3.94998074e-01
-5.58791280e-01 4.48479950e-01 -1.35618106e-01 -1.38507962e+00
-3.89882624e-01 1.66715294e-01 -2.34647527e-01 7.73928106e-01
-1.73667908e+00 -9.54658866e-01 -7.40517080e-01 6.55663073e-01
1.63570374e-01 2.17133135e-01 9.78137195e-01 1.18355729e-01
2.64136136e-01 1.79781765e-01 7.16680050e-01 -5.04009604e-01
7.23359823e-01 -1.21577895e+00 3.89397413e-01 8.31964016e-01
1.82323113e-01 9.86077785e-01 1.22619569e+00 -9.46291387e-01
-1.80275941e+00 -8.87772977e-01 8.09345186e-01 -4.60819989e-01
4.04316455e-01 -6.32242978e-01 -3.00874919e-01 6.44553244e-01
-3.09877425e-01 4.28316966e-02 4.15191680e-01 3.17046195e-02
4.06199425e-01 -2.60090858e-01 -1.40203798e+00 2.75524855e-01
6.80942297e-01 -2.46332049e-01 -1.55241817e-01 4.56115365e-01
3.40442240e-01 -5.76640308e-01 -5.10980248e-01 4.91106451e-01
6.38186634e-01 -8.16973984e-01 1.11430013e+00 8.77750576e-01
-1.36571383e+00 -9.62526083e-01 -5.99660695e-01 -1.31851554e+00
1.72925279e-01 -1.10181284e+00 -7.82722533e-02 9.46066141e-01
1.56457156e-01 -1.32421422e+00 7.35900164e-01 -4.19833474e-02
1.52230889e-01 -2.91758746e-01 -1.60453641e+00 -9.15528536e-01
-7.77604640e-01 -4.29811835e-01 8.16930294e-01 5.27373374e-01
-3.10934246e-01 -6.88806921e-02 -3.95587236e-01 1.13486981e+00
1.26241183e+00 -1.50718451e-01 1.26322520e+00 -1.40793061e+00
-1.63760912e-02 4.42968518e-01 -4.38821673e-01 -1.50554681e+00
-7.87690654e-02 -3.28116089e-01 6.66703939e-01 -1.82843971e+00
-5.87390065e-01 -1.21563733e+00 -1.41597003e-01 -3.04823518e-01
3.71069074e-01 8.79443213e-02 -4.41732019e-01 4.02851373e-01
-7.17654943e-01 3.68270248e-01 4.85935301e-01 1.80894867e-01
-5.44287205e-01 8.50400567e-01 -3.57448608e-02 7.31899679e-01
6.68940067e-01 -7.88213074e-01 -5.09677231e-01 -3.51428002e-01
5.20768523e-01 4.49687958e-01 4.60008770e-01 -1.58222640e+00
7.57312596e-01 -1.74720526e-01 3.17139030e-01 -9.31981087e-01
8.42776716e-01 -1.07372808e+00 7.72066593e-01 7.67250955e-01
2.68317819e-01 1.10454582e-01 -1.02724135e-01 1.20094383e+00
1.37175977e-01 -1.30271137e-01 4.36874241e-01 8.53742659e-02
-3.69034857e-01 -1.37617543e-01 -5.42365253e-01 -6.11865640e-01
1.07863796e+00 -2.17045769e-01 -3.32639217e-01 -8.56283665e-01
-3.60289842e-01 1.08267955e-01 7.31110990e-01 -4.61348891e-02
5.92083812e-01 -1.25478280e+00 -1.71417370e-01 1.70022652e-01
-1.97831929e-01 1.47684693e-01 2.19209865e-02 9.51739311e-01
-5.60675621e-01 5.93360424e-01 3.25418562e-01 -9.26671088e-01
-9.49708045e-01 -5.33816032e-02 4.47151810e-01 -8.32271762e-03
-2.81957656e-01 7.86513090e-01 -2.73127168e-01 -4.79819268e-01
2.08893284e-01 -3.11167061e-01 2.72049665e-01 -4.60693508e-01
4.49157536e-01 3.78262371e-01 -8.27513859e-02 -7.54939497e-01
-9.43971097e-01 1.14269900e+00 4.68473583e-01 -6.34919524e-01
1.06762767e+00 -9.36054826e-01 6.72788871e-03 5.58991730e-01
8.42576861e-01 1.02196789e+00 -8.92816842e-01 -3.53737511e-02
2.05943003e-01 -6.25766098e-01 -4.60585169e-02 -3.93653870e-01
-2.00083137e-01 1.03522025e-01 7.01807320e-01 1.87988833e-01
4.31635022e-01 -1.27912298e-01 4.17390317e-01 6.63691819e-01
1.88239443e+00 -6.84441626e-01 -5.10968745e-01 2.29154944e-01
5.27719736e-01 -9.73724544e-01 3.09960723e-01 -4.64148581e-01
2.62439221e-01 8.65286946e-01 8.84468257e-02 -1.87247664e-01
6.21039867e-01 4.67704087e-01 3.46566290e-02 8.97879750e-02
-8.62946138e-02 2.69714832e-01 -3.24407011e-01 8.52047086e-01
-6.69594407e-02 1.67451844e-01 -5.33197746e-02 2.69860685e-01
-4.65756893e-01 -1.17177032e-01 6.12894833e-01 1.42171049e+00
-8.01508009e-01 -9.60123003e-01 -9.89216447e-01 -2.29851097e-01
-9.36996713e-02 3.93419474e-01 2.90294260e-01 1.00215256e+00
-1.74691856e-01 1.21339202e+00 -2.02323601e-01 -4.10131603e-01
8.99239779e-02 -4.76169765e-01 5.14035344e-01 1.17384857e-02
2.03533620e-01 7.17892274e-02 4.64432575e-02 -7.96772242e-01
-6.72208071e-01 -7.92387664e-01 -1.49298120e+00 -3.99161428e-02
-5.88653028e-01 8.83716702e-01 1.53302896e+00 7.45132864e-01
4.76339310e-01 2.91676164e-01 9.36904430e-01 -1.20609581e+00
-2.89296091e-01 -7.16433346e-01 -6.33485496e-01 -6.95008397e-01
8.57584059e-01 -1.02294660e+00 -6.27013803e-01 -8.27081382e-01] | [6.233766555786133, 0.9882333874702454] |
d84829b7-73fb-404e-b5bb-8f07781ddfb0 | self-concordant-analysis-of-frank-wolfe | 2002.04320 | null | https://arxiv.org/abs/2002.04320v3 | https://arxiv.org/pdf/2002.04320v3.pdf | Self-Concordant Analysis of Frank-Wolfe Algorithms | Projection-free optimization via different variants of the Frank-Wolfe (FW), a.k.a. Conditional Gradient method has become one of the cornerstones in optimization for machine learning since in many cases the linear minimization oracle is much cheaper to implement than projections and some sparsity needs to be preserved. In a number of applications, e.g. Poisson inverse problems or quantum state tomography, the loss is given by a self-concordant (SC) function having unbounded curvature, implying absence of theoretical guarantees for the existing FW methods. We use the theory of SC functions to provide a new adaptive step size for FW methods and prove global convergence rate O(1/k) after k iterations. If the problem admits a stronger local linear minimization oracle, we construct a novel FW method with linear convergence rate for SC functions. | ['Pavel Dvurechensky', 'Petr Ostroukhov', 'Mathias Staudigl', 'Shimrit Shtern', 'Kamil Safin'] | 2020-02-11 | null | null | null | null | ['quantum-state-tomography'] | ['medical'] | [ 2.22651035e-01 2.40283936e-01 -8.36828202e-02 -3.82070810e-01
-8.93557608e-01 -5.70792139e-01 3.12948406e-01 -8.03722516e-02
-6.68238342e-01 1.07929277e+00 8.74279365e-02 -4.90443558e-01
-2.30543002e-01 -7.59725034e-01 -8.60811412e-01 -1.22586846e+00
6.05486296e-02 5.15501380e-01 3.26632150e-02 -2.83416003e-01
3.87333095e-01 2.93368965e-01 -5.43925226e-01 -3.65223289e-01
8.06340098e-01 9.40717518e-01 2.83141923e-03 6.34695172e-01
-2.51503866e-02 3.78428996e-01 2.19662189e-01 -4.91464615e-01
7.59857357e-01 -5.88225245e-01 -8.94784451e-01 -8.92373398e-02
4.17139113e-01 -2.66604424e-02 -3.61616939e-01 1.52933466e+00
3.56209040e-01 1.58956364e-01 6.01615489e-01 -7.98759758e-01
-3.29181999e-01 5.08717418e-01 -3.92905086e-01 -1.24685720e-01
9.08583775e-02 -1.37186393e-01 1.10750055e+00 -1.11698985e+00
6.67595863e-01 8.86996865e-01 8.20669472e-01 5.36556482e-01
-1.81358647e+00 -3.52643371e-01 -2.99268901e-01 1.70503661e-01
-1.25166452e+00 -4.51646298e-01 7.39757121e-01 -1.97166264e-01
4.79968339e-01 5.07634401e-01 4.95218545e-01 5.97374439e-01
1.89453214e-01 5.73154449e-01 1.54902542e+00 -6.34141862e-01
2.85714418e-01 5.65273404e-01 1.05329879e-01 9.09794509e-01
3.00653219e-01 -1.31829575e-01 -4.51951742e-01 -3.84900242e-01
6.97239935e-01 -9.41991881e-02 -6.26061141e-01 -6.70221865e-01
-1.16829658e+00 1.15226042e+00 4.03293759e-01 8.95321965e-02
-9.51244608e-02 1.44845858e-01 1.66570783e-01 5.16932964e-01
3.51349741e-01 3.05316299e-01 -2.25844890e-01 -2.81371903e-02
-9.83595729e-01 2.52907902e-01 1.12280488e+00 6.53488100e-01
1.01604784e+00 -9.82364342e-02 9.70095247e-02 3.79637510e-01
1.97118402e-01 6.66644037e-01 -1.64253090e-03 -1.14660287e+00
5.78072965e-01 -1.00726550e-02 3.88079315e-01 -7.44439900e-01
-2.80299425e-01 -4.26814854e-01 -1.03481710e+00 1.86642796e-01
7.48956382e-01 -3.20530832e-02 -3.68030906e-01 1.88199639e+00
5.80472827e-01 -3.05997469e-02 -5.93892969e-02 1.01672268e+00
-2.13853925e-01 6.81433737e-01 -6.76860809e-01 -5.50501227e-01
6.52908444e-01 -6.49083614e-01 -5.82440197e-01 -7.27645978e-02
7.59561360e-01 -6.30791605e-01 1.03310418e+00 4.03752923e-01
-1.18546116e+00 1.72863320e-01 -9.64878619e-01 -1.33757144e-01
1.17219314e-01 -1.01003371e-01 7.43613899e-01 7.76834130e-01
-1.08302522e+00 9.22111869e-01 -8.72634172e-01 -1.44984484e-01
1.84611574e-01 4.76096570e-01 -6.22259974e-01 -9.21454877e-02
-6.97463691e-01 7.94520676e-01 2.85062551e-01 1.82470441e-01
-8.14223528e-01 -6.13618553e-01 -6.20913446e-01 -1.11402169e-01
3.76890898e-01 -4.42941934e-01 7.65896380e-01 -4.95830119e-01
-1.79317653e+00 7.78941035e-01 -3.43467623e-01 -6.00136518e-01
8.60239506e-01 -2.07133472e-01 2.06307963e-01 1.50881425e-01
4.50046323e-02 -2.02124238e-01 8.99074733e-01 -7.25476384e-01
-1.07448988e-01 -5.60241878e-01 -9.64429509e-03 1.64788544e-01
-2.76354879e-01 -1.80708736e-01 1.60139084e-01 5.25471047e-02
5.00320315e-01 -1.07642710e+00 -5.24689317e-01 1.39601707e-01
-4.11430806e-01 1.07830375e-01 5.42983711e-01 -5.35566330e-01
9.05115962e-01 -1.87316275e+00 3.94229174e-01 4.01023537e-01
3.53121102e-01 -1.07328393e-01 1.78495094e-01 6.00776553e-01
1.10783190e-01 -1.29458234e-01 -8.11443865e-01 -3.71578574e-01
4.04837169e-02 5.50880134e-02 -3.41254592e-01 1.22223818e+00
-1.99909508e-01 6.81751609e-01 -7.83257067e-01 -1.97132617e-01
2.03437377e-02 2.47727275e-01 -9.25783813e-01 -1.13510676e-01
1.64096907e-01 6.75467253e-01 -4.44992065e-01 1.95909157e-01
9.09905195e-01 -3.92738611e-01 1.10828988e-01 -9.29785296e-02
-3.42045277e-01 3.59363794e-01 -1.31838548e+00 1.60112035e+00
-5.05621314e-01 5.06961405e-01 7.94378400e-01 -1.46388364e+00
5.20701587e-01 2.20081374e-01 3.56563032e-01 -1.39314860e-01
-2.15738162e-01 6.18983030e-01 -2.71530002e-01 -6.13408983e-02
2.21769676e-01 -7.25167394e-01 7.76202008e-02 2.77333468e-01
1.57941447e-03 -2.23864645e-01 -7.62977824e-02 4.29517835e-01
1.22665763e+00 -1.80125892e-01 3.29610378e-01 -7.15477526e-01
6.40216529e-01 -8.63384083e-02 7.39018559e-01 9.22152936e-01
-1.42337501e-01 6.49668992e-01 4.67085898e-01 -3.01097810e-01
-1.12260342e+00 -1.09867096e+00 -5.13071060e-01 6.41901374e-01
6.49585947e-02 -3.09060812e-01 -6.55683041e-01 -3.92130435e-01
-1.60683930e-01 4.05426294e-01 -4.65043843e-01 -1.54405236e-01
-5.91989636e-01 -1.04657924e+00 3.60566556e-01 -4.97060716e-02
8.67907107e-01 -4.07933384e-01 -1.84374526e-01 2.36145243e-01
7.61276782e-02 -1.05548036e+00 -6.49809361e-01 2.93753922e-01
-1.17273486e+00 -9.21484828e-01 -6.73630774e-01 -4.26982701e-01
7.41415143e-01 1.51628688e-01 6.90146744e-01 -3.90332371e-01
5.54067791e-02 3.20829928e-01 1.85979366e-01 7.02027455e-02
-1.45144641e-01 -1.61366478e-01 3.40571433e-01 1.99635729e-01
-3.30262817e-02 -7.11619735e-01 -6.52931809e-01 1.05784439e-01
-7.55682051e-01 -1.14123095e-02 5.35022080e-01 1.19680405e+00
7.58930087e-01 -1.42819881e-01 9.56065860e-03 -1.17188573e+00
4.89354193e-01 -2.92161614e-01 -1.16981435e+00 2.98662353e-02
-8.42879176e-01 3.58710736e-01 9.44603443e-01 -2.70557672e-01
-9.51751709e-01 3.20004821e-01 -1.44101650e-01 -1.70314714e-01
4.33180392e-01 5.55941999e-01 -4.20917198e-02 -7.53520727e-01
6.17234588e-01 5.04203677e-01 -2.25254539e-02 -5.95922649e-01
4.05330747e-01 2.56136447e-01 5.10266542e-01 -7.44119942e-01
1.11947763e+00 9.80945706e-01 5.54182887e-01 -1.12796640e+00
-1.07864141e+00 -6.65108800e-01 -4.37694728e-01 8.99795145e-02
7.65729249e-01 -4.79079157e-01 -7.43073702e-01 1.95485711e-01
-9.60890353e-01 -2.15685487e-01 -4.66199428e-01 8.19210768e-01
-7.59952247e-01 6.46844625e-01 -7.01721191e-01 -1.15681267e+00
-2.96750695e-01 -9.02793765e-01 6.39692783e-01 4.97591943e-02
3.57123554e-01 -1.06945109e+00 3.71934891e-01 2.68004149e-01
6.04390860e-01 1.09890506e-01 6.41437888e-01 -2.90555209e-01
-6.70563221e-01 -4.38701361e-01 -1.03824869e-01 6.30825758e-01
-3.16111624e-01 -5.96454322e-01 -6.50024354e-01 -4.45027858e-01
9.04393971e-01 -2.47682393e-01 9.11429703e-01 3.58643442e-01
9.87514675e-01 -7.50685036e-01 -6.05328530e-02 1.15929651e+00
1.74421096e+00 -5.67343771e-01 5.47312438e-01 -4.64174338e-02
7.05515385e-01 2.42754728e-01 1.60374567e-01 4.45975631e-01
-3.07170786e-02 4.40480173e-01 2.74217665e-01 3.16480994e-01
3.36933881e-01 -2.48008177e-01 5.42397261e-01 1.35527217e+00
-2.10471943e-01 6.44732773e-01 -4.47161257e-01 4.55995977e-01
-1.84375870e+00 -1.01961243e+00 -4.35688019e-01 2.82447767e+00
1.11834693e+00 6.77666813e-02 -1.09728560e-01 1.04519188e-01
4.36010808e-01 4.58019674e-02 -4.56770390e-01 -3.05596530e-01
-2.61607677e-01 6.32563353e-01 1.17596614e+00 1.01924932e+00
-7.40591347e-01 8.39523792e-01 6.62594843e+00 8.95934522e-01
-1.07334232e+00 6.48760378e-01 2.66726434e-01 5.89824701e-03
-3.20919752e-01 5.61043799e-01 -7.71865129e-01 4.20926154e-01
6.32790327e-01 -9.66520309e-02 6.59750938e-01 9.33581471e-01
1.66861057e-01 -1.95655480e-01 -8.43752205e-01 1.08106768e+00
-3.37789387e-01 -1.51667404e+00 -5.30389905e-01 5.07424295e-01
1.06289828e+00 3.70535940e-01 -1.78888645e-02 1.70665253e-02
1.70808267e-02 -8.82925332e-01 3.01954269e-01 2.21873105e-01
7.09944069e-01 -6.85681581e-01 4.95166421e-01 4.76365685e-01
-8.50567222e-01 1.48505807e-01 -8.38902235e-01 -1.97629198e-01
3.91332984e-01 1.23578203e+00 -4.59280461e-01 3.00963908e-01
2.96963543e-01 2.51124531e-01 1.59351632e-01 1.06726837e+00
-1.83342367e-01 9.13761854e-01 -9.81570661e-01 -5.90056926e-02
3.85916263e-01 -1.22475541e+00 1.07543778e+00 1.07964575e+00
2.97237933e-01 1.24961719e-01 -3.97636788e-03 9.76621747e-01
-2.17520639e-01 3.40697289e-01 -5.18323600e-01 -1.50299668e-01
9.12179127e-02 1.43002069e+00 -5.21626949e-01 -6.81602433e-02
-4.90767717e-01 1.15356576e+00 3.21433753e-01 3.72807413e-01
-4.00058925e-01 -4.68609393e-01 5.56035876e-01 2.05825806e-01
3.62963617e-01 -6.70558751e-01 -4.16822970e-01 -1.60962236e+00
3.22716773e-01 -4.79462355e-01 1.60725668e-01 4.72797938e-02
-1.18464029e+00 1.35914199e-02 -2.36321226e-01 -5.91192544e-01
6.33985847e-02 -8.07921708e-01 -8.54497731e-01 9.42694366e-01
-1.37627995e+00 -5.89145184e-01 1.84670746e-01 7.21623182e-01
-1.02393910e-01 1.90094873e-01 7.24063575e-01 2.13845029e-01
-3.33975852e-01 3.50386173e-01 8.20715249e-01 -9.96743515e-02
4.79022443e-01 -1.65683329e+00 -8.09339806e-02 1.00410593e+00
2.43527636e-01 7.34723151e-01 9.78036523e-01 -4.02619094e-01
-1.98391068e+00 -6.65332317e-01 7.85666883e-01 -3.15945894e-01
9.59531903e-01 -4.72357690e-01 -6.95334733e-01 6.39986515e-01
-1.51608258e-01 2.76145607e-01 4.49914306e-01 9.63354781e-02
-1.78461328e-01 -2.01178044e-01 -1.25728321e+00 4.22375530e-01
7.72508442e-01 -8.43854070e-01 -1.41798779e-01 8.45706582e-01
3.32485706e-01 -2.51208067e-01 -7.37957180e-01 2.01368883e-01
3.91096532e-01 -1.01787472e+00 8.48324478e-01 -4.50288504e-01
7.59092495e-02 -3.01771581e-01 -4.45990622e-01 -1.06931055e+00
-1.43686906e-01 -1.20678341e+00 -2.29964897e-01 5.10933280e-01
3.94112170e-01 -1.10121882e+00 9.16273832e-01 6.18119657e-01
-2.42756996e-02 -8.50503802e-01 -1.21533835e+00 -9.43878651e-01
3.03454190e-01 -4.77913707e-01 -2.61980772e-01 8.29255760e-01
1.44189239e-01 3.89557779e-01 -6.91245854e-01 1.21507555e-01
1.18887746e+00 2.29416341e-01 6.28272474e-01 -8.56982529e-01
-8.33148837e-01 -2.90296555e-01 -4.31380093e-01 -1.37749708e+00
-1.00296803e-01 -1.22278082e+00 -7.97776058e-02 -1.13719082e+00
3.47945631e-01 -4.85285610e-01 -1.25086695e-01 -5.82474656e-02
8.35056454e-02 2.04210460e-01 3.74561027e-02 3.16806912e-01
-4.73217189e-01 7.21162617e-01 1.08395767e+00 9.51077342e-02
-1.88526273e-01 3.74005467e-01 -3.81401777e-01 7.43622720e-01
6.54697776e-01 -6.93774700e-01 -3.78480926e-02 -1.38647869e-01
5.84051490e-01 2.93553472e-02 1.84651300e-01 -8.50937545e-01
3.87127548e-01 -1.84436187e-01 -2.80826204e-02 -1.93320140e-01
5.33483207e-01 -3.92310083e-01 2.53124200e-02 5.68711877e-01
-1.43630549e-01 -3.31001133e-01 -5.14784992e-01 6.39472485e-01
-2.36578248e-02 -6.90810263e-01 1.16162312e+00 -2.61196315e-01
1.19471394e-01 4.15504456e-01 1.82555206e-02 2.76077211e-01
4.90184933e-01 2.05273762e-01 1.32778794e-01 -6.20961905e-01
-6.10886097e-01 -7.25270137e-02 5.13802886e-01 -7.63625205e-01
6.76410854e-01 -1.40405762e+00 -7.75552630e-01 -5.76388836e-02
-3.16673279e-01 -5.49750694e-04 -8.15088227e-02 1.64699769e+00
-6.82672501e-01 4.81754899e-01 2.38669902e-01 -4.93062645e-01
-8.54044974e-01 1.89492315e-01 3.08529615e-01 -3.54875475e-01
-8.16472709e-01 1.13051283e+00 3.12470436e-01 -5.34739256e-01
-9.20306966e-02 2.53374465e-02 7.16135979e-01 -5.23036778e-01
6.38099015e-01 4.58090723e-01 2.63407920e-02 -5.20094454e-01
-1.06931582e-01 4.69462484e-01 -7.28686750e-02 -6.04681909e-01
1.43475747e+00 -9.01270658e-02 -5.72396576e-01 4.14570302e-01
1.59749246e+00 4.25514132e-01 -1.16232157e+00 -3.86691868e-01
8.31811056e-02 -5.06422520e-01 2.20102444e-01 -9.10536572e-02
-9.03156638e-01 9.57212329e-01 3.67024362e-01 1.42776892e-01
7.97652721e-01 -4.70437035e-02 8.97526860e-01 8.94160151e-01
5.49317300e-01 -1.30255401e+00 -4.23472553e-01 5.40387273e-01
5.88400960e-01 -1.20017684e+00 8.01186338e-02 -2.84198403e-01
-2.07009181e-01 1.15857720e+00 -1.35314792e-01 -5.32504320e-01
7.70415008e-01 2.95734070e-02 -4.39466745e-01 4.07095626e-02
-3.37530941e-01 -8.21462572e-02 2.46544927e-01 4.75549623e-02
1.58799484e-01 1.20449245e-01 -8.90345454e-01 5.92885204e-02
-2.15389982e-01 -2.86004871e-01 6.56010270e-01 7.25832880e-01
-5.04247129e-01 -1.26536417e+00 -2.04593971e-01 6.35003924e-01
-6.04037523e-01 -2.60953039e-01 -1.34667218e-01 3.65227640e-01
-4.48797137e-01 5.95011473e-01 -7.14572430e-01 -9.59143322e-03
-2.58956999e-01 2.14874834e-01 1.02408159e+00 -4.35528070e-01
-2.70426162e-02 -6.13949224e-02 -3.34551871e-01 -8.15147579e-01
-3.18707526e-01 -6.38083577e-01 -1.08407795e+00 -5.86737514e-01
-6.05028749e-01 4.44776446e-01 9.81978297e-01 1.15884674e+00
-2.55749617e-02 -4.30297345e-01 8.37124765e-01 -6.39077663e-01
-1.26617086e+00 -7.03339577e-01 -8.40363085e-01 1.93101063e-01
3.71122360e-01 -2.73785502e-01 -8.27553332e-01 -4.40015614e-01] | [6.491551876068115, 4.612423896789551] |
29f9fc87-7f14-45fd-aa96-5fba6a381cc6 | target-aware-spatio-temporal-reasoning-via | 2305.12397 | null | https://arxiv.org/abs/2305.12397v1 | https://arxiv.org/pdf/2305.12397v1.pdf | Target-Aware Spatio-Temporal Reasoning via Answering Questions in Dynamics Audio-Visual Scenarios | Audio-visual question answering (AVQA) is a challenging task that requires multistep spatio-temporal reasoning over multimodal contexts. To achieve scene understanding ability similar to humans, the AVQA task presents specific challenges, including effectively fusing audio and visual information and capturing question-relevant audio-visual features while maintaining temporal synchronization. This paper proposes a Target-aware Joint Spatio-Temporal Grounding Network for AVQA to address these challenges. The proposed approach has two main components: the Target-aware Spatial Grounding module, the Tri-modal consistency loss and corresponding Joint audio-visual temporal grounding module. The Target-aware module enables the model to focus on audio-visual cues relevant to the inquiry subject by exploiting the explicit semantics of text modality. The Tri-modal consistency loss facilitates the interaction between audio and video during question-aware temporal grounding and incorporates fusion within a simpler single-stream architecture. Experimental results on the MUSIC-AVQA dataset demonstrate the effectiveness and superiority of the proposed method over existing state-of-the-art methods. Our code will be availiable soon. | ['Jianqin Yin', 'Yuanyuan Jiang'] | 2023-05-21 | null | null | null | null | ['scene-understanding'] | ['computer-vision'] | [-3.28514688e-02 -2.91682512e-01 3.66509240e-03 -3.23432267e-01
-1.49398553e+00 -4.76413071e-01 4.38602507e-01 3.91231596e-01
-1.82675526e-01 1.17870346e-01 4.89942789e-01 4.55903299e-02
-4.12929624e-01 -4.87063974e-01 -5.99064648e-01 -5.24839997e-01
-1.17485598e-01 1.56895131e-01 6.83470845e-01 -1.54900715e-01
2.43347377e-01 1.29007414e-01 -2.03274798e+00 1.11239517e+00
6.43413246e-01 1.35631883e+00 3.11480463e-01 1.05528951e+00
-4.81043875e-01 1.21935892e+00 -4.32776153e-01 9.02143940e-02
-1.94647402e-01 -3.65233749e-01 -1.06814790e+00 1.03254132e-01
8.39619398e-01 -3.27554792e-01 -2.88331300e-01 6.70846581e-01
6.20744944e-01 6.37470245e-01 2.42847458e-01 -1.61425054e+00
-4.73771960e-01 2.69941360e-01 -4.90141302e-01 4.82693613e-01
8.73865485e-01 1.86824426e-01 1.39218855e+00 -1.04413199e+00
4.60168749e-01 1.45467985e+00 4.09514964e-01 1.33648589e-01
-7.89449036e-01 -5.53298831e-01 5.40417373e-01 1.15726423e+00
-1.63505566e+00 -5.02720058e-01 1.08522332e+00 -4.08438623e-01
8.18287313e-01 4.38506246e-01 7.22267449e-01 7.76995778e-01
-1.99316502e-01 1.12731910e+00 6.48285687e-01 -2.69798815e-01
2.58451551e-01 -1.81608364e-01 1.41225547e-01 6.55149758e-01
-8.50606263e-01 -3.16413075e-01 -1.42322445e+00 -2.71040589e-01
5.19490540e-01 -9.21772644e-02 -2.68875688e-01 -5.82000792e-01
-1.43642199e+00 4.95311290e-01 4.42198634e-01 4.55959648e-01
-6.57110751e-01 3.79428774e-01 5.38061619e-01 1.63701966e-01
6.33349940e-02 -7.68501237e-02 -1.09699331e-01 -2.53625721e-01
-1.20657921e+00 4.29407924e-01 2.05563083e-01 9.91334200e-01
4.87472266e-01 2.69272812e-02 -7.25799978e-01 6.58893466e-01
5.46392381e-01 3.90585065e-01 1.36586919e-01 -1.34734833e+00
8.59445214e-01 4.73138899e-01 1.30085751e-01 -1.15846884e+00
-3.66326869e-01 -7.46220350e-02 -5.14971912e-01 -2.04952300e-01
3.72610003e-01 3.94646287e-01 -5.72010934e-01 1.73348820e+00
8.34370375e-01 4.47823793e-01 -1.27161935e-01 1.16300941e+00
1.29272068e+00 1.09850061e+00 1.65109500e-01 -3.47572714e-01
1.81676984e+00 -9.83252883e-01 -1.22533357e+00 1.34554029e-01
1.75573364e-01 -8.43731999e-01 1.38886154e+00 3.35703045e-01
-1.36085832e+00 -7.10929990e-01 -9.20723677e-01 -4.96914864e-01
-1.93693399e-01 -4.06279694e-03 2.24879786e-01 1.02437407e-01
-1.07860732e+00 -1.42732292e-01 -6.34398341e-01 -5.08955657e-01
1.90383151e-01 8.15417990e-03 -2.29630306e-01 -1.35899723e-01
-1.25662231e+00 2.65777975e-01 3.29813600e-01 2.13866428e-01
-1.14343131e+00 -8.44361544e-01 -7.98387110e-01 4.18390669e-02
7.24815607e-01 -8.27643573e-01 1.58301675e+00 -7.60470986e-01
-1.39845383e+00 4.03023571e-01 -6.22968554e-01 -2.04422221e-01
8.62948075e-02 -2.80484855e-01 -5.57603121e-01 1.13658786e+00
1.73621014e-01 6.36111617e-01 1.05637765e+00 -1.23007965e+00
-8.21216226e-01 -4.27559912e-01 1.59474775e-01 6.08162045e-01
-3.62180024e-01 5.78725673e-02 -7.67665088e-01 -6.53174698e-01
3.41545492e-01 -3.92937034e-01 3.77176374e-01 3.90789092e-01
8.04222301e-02 -3.47945184e-01 1.09486532e+00 -7.32629180e-01
1.38119519e+00 -2.29124212e+00 1.98583663e-01 8.76721144e-02
6.20549619e-02 -1.45743877e-01 -3.46894205e-01 7.10619688e-01
4.03035842e-02 -5.04883528e-01 1.37357950e-01 -2.68414110e-01
4.20829765e-02 -9.19657573e-02 -5.83254457e-01 3.83468568e-01
1.28698617e-01 8.04746985e-01 -1.00261319e+00 -1.09452271e+00
2.33206898e-01 5.68363190e-01 -6.10128343e-01 4.09151316e-01
-3.80932420e-01 3.77900630e-01 -4.83795404e-01 9.25557196e-01
3.69007438e-01 -2.37522542e-01 -2.89971679e-01 -6.14080012e-01
-1.49675846e-01 2.59618938e-01 -1.44885468e+00 2.34759736e+00
-2.67820269e-01 6.29393756e-01 3.52143854e-01 -7.51978278e-01
6.69848740e-01 7.20986962e-01 5.76679230e-01 -1.02101982e+00
-1.03668876e-01 -1.52212009e-01 -5.02934039e-01 -9.83577490e-01
7.97571480e-01 -6.02970365e-03 -1.87017079e-02 3.93321753e-01
3.09946984e-01 -2.42959231e-01 -7.18880445e-03 6.33271933e-01
8.33284974e-01 2.44571924e-01 5.74714970e-03 1.17860742e-01
8.14667165e-01 9.77890864e-02 2.97465086e-01 5.06950796e-01
-6.09713495e-01 7.16814697e-01 1.27796412e-01 -2.26966128e-01
-5.50844789e-01 -1.28074467e+00 4.19239074e-01 1.59023058e+00
3.76148939e-01 -6.97733998e-01 -3.48051935e-01 -3.48647028e-01
-3.02170843e-01 6.56471848e-01 -5.94848514e-01 3.47706713e-02
-5.39609253e-01 2.54205942e-01 4.94122356e-01 5.51361203e-01
5.78666031e-01 -8.85998249e-01 -6.11122012e-01 9.87026468e-02
-1.11767709e+00 -1.16136193e+00 -6.49567723e-01 -3.51092726e-01
-6.16764903e-01 -9.83190238e-01 -7.15586543e-01 -7.16195643e-01
-9.34300348e-02 7.75062501e-01 9.32029545e-01 -3.52495052e-02
-2.16137797e-01 1.08642197e+00 -5.24539471e-01 -2.74832398e-01
2.21059233e-01 -2.09528506e-01 -3.93633008e-01 4.89259839e-01
1.16591148e-01 -6.25930130e-01 -7.11432636e-01 3.70972246e-01
-9.89736557e-01 9.20334889e-04 1.33268520e-01 5.15878320e-01
9.56947267e-01 -1.36291951e-01 6.20524943e-01 -3.15432921e-02
4.43407476e-01 -4.26133066e-01 -2.94210643e-01 5.01970589e-01
1.16370477e-01 -2.60123044e-01 2.12221116e-01 -5.46499968e-01
-1.30621517e+00 -1.28621794e-02 6.25784695e-02 -7.23773837e-01
-9.57067385e-02 4.85273540e-01 -2.75371611e-01 2.97221035e-01
4.26577926e-01 3.28846008e-01 -2.42379367e-01 -2.65912771e-01
7.77068377e-01 4.54064578e-01 9.91447806e-01 -7.02649534e-01
6.73165262e-01 7.95706093e-01 -1.34834386e-02 -9.40920353e-01
-8.47461998e-01 -1.08049214e+00 -4.37161505e-01 -8.73726070e-01
1.25368774e+00 -1.26338816e+00 -1.10644746e+00 7.30036572e-02
-1.27818263e+00 2.42739506e-02 -4.41950649e-01 5.05461097e-01
-8.00723970e-01 6.47616029e-01 -3.71737957e-01 -1.04436648e+00
-2.64314562e-01 -7.48559356e-01 1.38100207e+00 7.63196796e-02
-2.22226605e-01 -5.92455566e-01 -3.38283763e-03 9.07675326e-01
1.37636274e-01 8.16952512e-02 7.64340639e-01 -2.93700337e-01
-9.11964297e-01 1.92018554e-01 -3.33991408e-01 -2.68964171e-01
-1.54419795e-01 -9.43677202e-02 -1.27739298e+00 -5.56090400e-02
1.11749237e-02 -5.38730025e-01 7.69273877e-01 3.40923071e-01
1.13852727e+00 -1.50282100e-01 2.01958328e-01 2.74122059e-01
1.01583600e+00 1.98512554e-01 2.37303734e-01 2.43410721e-01
6.33960605e-01 8.13226998e-01 1.24681365e+00 6.31541312e-01
9.71233130e-01 8.35392773e-01 8.32197785e-01 5.81152812e-02
-2.12222472e-01 -2.46987328e-01 3.07594031e-01 1.09625149e+00
2.61815816e-01 -1.50575265e-01 -9.17004883e-01 1.02109838e+00
-2.21323180e+00 -1.33605552e+00 -3.15026075e-01 1.65326071e+00
6.63765073e-01 -4.10909027e-01 3.68228585e-01 5.37939847e-01
5.04404843e-01 4.93068665e-01 -2.68467188e-01 -1.71335310e-01
-4.44805026e-02 -2.47964606e-01 -4.93157804e-01 4.93981838e-01
-1.03963017e+00 5.43910563e-01 5.57429028e+00 9.46050942e-01
-7.15926766e-01 3.60359013e-01 -4.15808745e-02 -4.46212500e-01
-5.22677302e-01 -1.47246689e-01 -3.04549843e-01 -1.17875323e-01
6.92805588e-01 1.35833640e-02 2.17127800e-01 2.46270776e-01
4.48694199e-01 -2.92795599e-01 -1.19105840e+00 1.26857007e+00
3.77512068e-01 -1.30729687e+00 2.61800647e-01 -5.35510480e-01
3.04306835e-01 -4.70599115e-01 2.02201515e-01 1.55498430e-01
-2.73977757e-01 -6.52947962e-01 1.33931243e+00 8.92844975e-01
5.16726792e-01 -8.35034907e-01 2.86748379e-01 2.68776447e-01
-1.97167563e+00 -1.74467176e-01 1.55513465e-01 2.07588404e-01
4.43838388e-01 2.29147464e-01 -5.52499473e-01 1.02186561e+00
1.18564916e+00 5.97809374e-01 -4.68888015e-01 1.04194832e+00
1.10141546e-01 4.93928671e-01 -2.09055692e-01 2.80979633e-01
3.08368385e-01 1.89885452e-01 7.75437891e-01 1.12680018e+00
4.42004204e-01 4.06711876e-01 1.36652127e-01 7.01847970e-01
4.02063072e-01 3.58303338e-01 -2.47825205e-01 8.93221051e-03
5.43899655e-01 9.69941497e-01 -4.03305173e-01 -2.72872597e-01
-5.06017327e-01 7.63693035e-01 -3.01137520e-03 5.66198945e-01
-9.41654444e-01 -3.36579412e-01 3.85287642e-01 -3.92965116e-02
5.16479373e-01 -2.83971012e-01 -1.07014641e-01 -9.78456140e-01
2.67068386e-01 -9.51372862e-01 1.05700219e+00 -1.55304396e+00
-9.48899984e-01 4.86553162e-01 1.97263569e-01 -1.44666350e+00
-1.53966933e-01 -4.49521653e-02 -3.90987635e-01 6.60576761e-01
-1.64929581e+00 -1.53796387e+00 -6.00626230e-01 1.39219272e+00
7.39907444e-01 6.25822768e-02 6.33825183e-01 3.85161370e-01
-1.91337526e-01 3.13123882e-01 -5.45971155e-01 -2.98226625e-01
8.93615365e-01 -9.22686517e-01 -3.60002905e-01 8.91990840e-01
3.31980973e-01 2.59809881e-01 7.24788606e-01 -3.79559726e-01
-1.61551368e+00 -9.87397194e-01 9.72034693e-01 -2.10826680e-01
6.84252083e-01 -1.67000473e-01 -1.17200291e+00 4.31251973e-01
4.87020195e-01 -8.80541950e-02 7.03715682e-01 -2.29063686e-02
-7.31397748e-01 -4.48050380e-01 -8.34317029e-01 4.97006148e-01
7.59575129e-01 -1.18783975e+00 -8.80770802e-01 1.11017928e-01
1.10876250e+00 -2.69848496e-01 -7.47668564e-01 2.39085510e-01
4.68132615e-01 -9.64729071e-01 1.20642197e+00 -2.75886208e-01
7.66364485e-02 -8.18677902e-01 -5.28878510e-01 -5.72319984e-01
-2.04755664e-01 -7.03853786e-01 -3.50101531e-01 1.53676271e+00
-1.25836730e-01 1.61682412e-01 4.44381684e-01 1.19744204e-01
-1.23266466e-01 -2.91266620e-01 -1.33205318e+00 -5.50046682e-01
-6.22729838e-01 -1.00603724e+00 5.57475388e-01 8.95237803e-01
2.08167091e-01 3.64243120e-01 -4.29422140e-01 6.02867782e-01
4.91049856e-01 3.44609022e-01 6.31661594e-01 -9.35519755e-01
-1.94667473e-01 -1.96728364e-01 -2.45247722e-01 -1.17606878e+00
-1.27709761e-01 -5.57183743e-01 2.43410751e-01 -1.78969574e+00
3.57146077e-02 2.78660148e-01 -3.26053768e-01 4.27555829e-01
-6.60624704e-04 1.13517620e-01 4.83650804e-01 2.49310341e-02
-1.34094405e+00 9.11939859e-01 1.33559024e+00 -3.73631299e-01
-2.76020259e-01 -3.95116448e-01 -2.51716197e-01 5.38520873e-01
3.55239600e-01 -2.51628280e-01 -7.89321899e-01 -5.11754513e-01
3.36400479e-01 6.48127913e-01 7.34637201e-01 -1.03680229e+00
7.66371608e-01 -2.75595814e-01 3.65153104e-02 -1.36554456e+00
9.50231493e-01 -1.04745460e+00 -2.73226615e-04 3.11880466e-03
-4.95612770e-01 2.11235523e-01 4.53080595e-01 7.85215199e-01
-7.65561998e-01 2.81569004e-01 4.66566712e-01 1.89109311e-01
-8.89539182e-01 1.71647578e-01 -5.68412125e-01 1.93387017e-01
1.02128792e+00 -1.33141503e-01 -2.19729975e-01 -8.88790190e-01
-8.34544539e-01 5.53257585e-01 -2.57401556e-01 5.05064189e-01
9.98875856e-01 -1.71430755e+00 -5.26319802e-01 -1.15691490e-01
5.46202421e-01 -1.34862497e-01 9.65750515e-01 9.82768178e-01
-1.41847715e-01 5.33432364e-01 -1.92502718e-02 -1.10973406e+00
-1.57923126e+00 5.78910947e-01 1.75988719e-01 3.25902514e-02
-3.95872504e-01 9.83785689e-01 3.52916837e-01 -6.61024870e-03
7.74045467e-01 -3.94570887e-01 -4.75282907e-01 3.70451510e-01
7.60131478e-01 4.76269513e-01 -2.17610281e-02 -8.89906764e-01
-5.36095679e-01 7.52875924e-01 3.63490969e-01 -5.86885273e-01
1.04245067e+00 -7.30758965e-01 -2.67385086e-03 1.04451644e+00
9.25670624e-01 -3.16378206e-01 -1.00257802e+00 -6.54897630e-01
-2.63983101e-01 -4.08517092e-01 1.54935166e-01 -7.04579890e-01
-7.38098979e-01 1.29661834e+00 6.75254762e-01 1.86577499e-01
1.56740856e+00 1.02651551e-01 6.89645708e-01 3.30155104e-01
9.78147984e-02 -1.16099048e+00 8.11212420e-01 3.80510032e-01
1.35078275e+00 -9.74645615e-01 -8.27116221e-02 -4.48833853e-01
-7.01896727e-01 9.46589649e-01 6.57251835e-01 3.77556086e-01
6.28394067e-01 -2.51638025e-01 2.16279924e-01 -3.66135806e-01
-1.16692114e+00 -3.70964617e-01 8.00182700e-01 5.80671489e-01
2.67114546e-02 -3.65950346e-01 1.48120284e-01 7.56580532e-01
1.98815435e-01 -7.97227100e-02 5.62468767e-02 1.07013369e+00
-5.34563959e-01 -5.95116317e-01 -7.95696020e-01 -2.96912938e-01
-2.69638836e-01 1.62989944e-02 -2.90841222e-01 6.30354166e-01
6.03660196e-02 1.40768611e+00 1.07596010e-01 -2.58827269e-01
5.31035781e-01 2.97871858e-01 3.84440094e-01 -2.71313518e-01
-8.01745355e-01 4.76222724e-01 1.24255251e-02 -1.06822801e+00
-8.75949025e-01 -7.72361577e-01 -1.46935189e+00 8.04511309e-02
-6.81996793e-02 2.34759048e-01 4.70402330e-01 9.07836616e-01
4.44476277e-01 7.91524172e-01 3.48709464e-01 -6.92007422e-01
-9.01518762e-02 -6.77943110e-01 -4.36963588e-01 3.10717523e-01
7.51219749e-01 -4.81884480e-01 -7.26677030e-02 3.19304943e-01] | [10.46689510345459, 1.0706003904342651] |
28686502-e962-4b9b-b4dc-f060c23d6551 | disentangling-structure-and-style-political | 2304.02247 | null | https://arxiv.org/abs/2304.02247v1 | https://arxiv.org/pdf/2304.02247v1.pdf | Disentangling Structure and Style: Political Bias Detection in News by Inducing Document Hierarchy | We address an important gap in detection of political bias in news articles. Previous works that perform supervised document classification can be biased towards the writing style of each news outlet, leading to overfitting and limited generalizability. Our approach overcomes this limitation by considering both the sentence-level semantics and the document-level rhetorical structure, resulting in a more robust and style-agnostic approach to detecting political bias in news articles. We introduce a novel multi-head hierarchical attention model that effectively encodes the structure of long documents through a diverse ensemble of attention heads. While journalism follows a formalized rhetorical structure, the writing style may vary by news outlet. We demonstrate that our method overcomes this domain dependency and outperforms previous approaches for robustness and accuracy. Further analysis demonstrates the ability of our model to capture the discourse structures commonly used in the journalism domain. | ['James Thorne', 'Jiyoung Han', 'Jaemin Jung', 'Yejin Cho', 'Jiwoo Hong'] | 2023-04-05 | null | null | null | null | ['document-classification'] | ['natural-language-processing'] | [-1.67135838e-02 1.62387520e-01 -8.80951822e-01 -5.12148499e-01
-7.73452222e-01 -8.87752533e-01 1.17625093e+00 3.94238800e-01
-4.48575467e-01 6.98608994e-01 1.32330251e+00 -6.49191022e-01
4.29618284e-02 -6.78297460e-01 -6.34324849e-01 -3.10895681e-01
6.09384775e-01 3.52535307e-01 1.10785298e-01 -5.74054718e-01
7.94456422e-01 3.65283266e-02 -1.04067242e+00 5.28219998e-01
8.57097328e-01 4.49454159e-01 -1.08427741e-01 6.60289407e-01
-3.38002533e-01 1.64270318e+00 -9.18382764e-01 -5.18231094e-01
-2.23132908e-01 -5.50815403e-01 -1.12964630e+00 -3.18995342e-02
9.82113123e-01 -1.99932322e-01 -3.96154135e-01 9.32567120e-01
4.82643604e-01 -4.58864748e-01 8.42569709e-01 -6.31080151e-01
-9.02474642e-01 1.18376982e+00 -6.18464291e-01 6.66453779e-01
2.81597972e-01 -2.90603518e-01 1.51757085e+00 -4.78543252e-01
9.46015656e-01 1.62999690e+00 7.00922847e-01 3.45291615e-01
-1.29228520e+00 -4.58119750e-01 6.75964117e-01 -7.11175567e-03
-4.60486561e-01 -5.25777996e-01 1.06367862e+00 -9.55613852e-01
4.66211766e-01 2.67103165e-01 4.82566714e-01 1.77674627e+00
3.95893395e-01 7.08572388e-01 1.55783772e+00 -4.64304417e-01
-3.90648246e-02 5.44567667e-02 9.65560675e-01 4.91004050e-01
4.76083308e-01 -3.43325824e-01 -5.15025377e-01 -5.27841926e-01
2.74074256e-01 -3.07806015e-01 -1.89518064e-01 6.39683828e-02
-1.12509453e+00 1.28998923e+00 4.46153998e-01 4.82439935e-01
-1.96646661e-01 -4.81932471e-03 7.45151281e-01 5.68328500e-01
7.32477307e-01 9.47705626e-01 -2.73830622e-01 -9.50612649e-02
-1.13852143e+00 4.98866975e-01 1.19218218e+00 7.54995346e-01
2.01083198e-01 -7.78848305e-02 -8.02112937e-01 8.71611774e-01
1.00167468e-01 5.86750507e-01 6.81635261e-01 -8.32729042e-01
5.19537330e-01 5.53865850e-01 1.62381664e-01 -1.48700416e+00
-5.05053937e-01 -6.64093196e-01 -4.28306699e-01 -2.43519068e-01
6.34114742e-01 -2.26550549e-01 -5.91185808e-01 1.84428501e+00
1.03201792e-01 -9.81538475e-01 -1.46504134e-01 7.95860887e-01
1.03849566e+00 3.95262152e-01 1.31066710e-01 -1.18766919e-01
1.68150818e+00 -1.01139092e+00 -8.34241033e-01 -6.44852102e-01
5.32751560e-01 -8.16304147e-01 1.30603397e+00 -2.66466856e-01
-9.63316262e-01 -1.15499005e-01 -9.33682084e-01 -3.69473159e-01
-3.07636280e-02 1.03667881e-02 2.46598542e-01 3.94264698e-01
-4.95184302e-01 2.48278573e-01 -2.51437902e-01 -3.24439555e-01
4.49219912e-01 -2.87047446e-01 2.15814441e-01 2.60994554e-01
-1.21636677e+00 1.02662742e+00 6.93492293e-02 -4.23027515e-01
-3.96517396e-01 -6.14811003e-01 -6.43544316e-01 -6.08318225e-02
4.72773999e-01 -6.68253958e-01 1.56546450e+00 -1.35699511e+00
-1.34840310e+00 1.16730106e+00 -4.14792508e-01 -4.58175123e-01
7.25961566e-01 -2.44288132e-01 -1.37700319e-01 -1.76006466e-01
5.61639428e-01 -1.71205893e-01 8.99651945e-01 -1.12798214e+00
-3.59037966e-01 -3.60753357e-01 1.73754811e-01 2.20565438e-01
-5.34437120e-01 4.38184768e-01 -9.79785249e-02 -9.84058857e-01
-6.23356327e-02 -8.44850421e-01 2.70938873e-02 -5.96666574e-01
-6.25676453e-01 -2.76823133e-01 5.57958961e-01 -8.69350135e-01
1.65800738e+00 -1.88101304e+00 2.63405502e-01 3.13849524e-02
4.83997732e-01 -1.66661263e-01 3.26325923e-01 5.37927568e-01
4.63836849e-01 3.44036102e-01 5.83404191e-02 1.13595044e-02
1.22901537e-01 1.88819617e-01 -6.84622407e-01 5.24160385e-01
-3.02926060e-02 9.68275905e-01 -9.48675156e-01 -7.04519093e-01
-4.98588860e-01 5.71072213e-02 -6.79054320e-01 -1.09803386e-01
-3.68448108e-01 3.35382640e-01 -5.60809910e-01 5.72001755e-01
1.98635057e-01 -5.65734804e-01 5.62587917e-01 -7.93143809e-02
-3.47634584e-01 1.19674134e+00 -4.30879772e-01 1.09524989e+00
-3.66819471e-01 1.18838179e+00 3.51160824e-01 -9.26733077e-01
6.95412934e-01 -1.30421435e-02 -6.05738461e-02 -7.52112508e-01
2.62863547e-01 2.37409040e-01 3.16597402e-01 -4.85412568e-01
6.92415893e-01 -1.50064752e-01 -5.11186957e-01 6.39207780e-01
-2.52436131e-01 4.65343110e-02 3.59586477e-01 3.40527594e-01
7.87031949e-01 -3.89986724e-01 4.47133660e-01 -9.26880777e-01
2.48968095e-01 2.08713427e-01 6.17358446e-01 1.33063877e+00
-9.49546173e-02 4.06237066e-01 1.03101277e+00 -5.01432538e-01
-1.22634768e+00 -4.23388928e-01 -3.71086836e-01 1.72504771e+00
-2.73518443e-01 -4.83440369e-01 -5.94142377e-01 -9.97025132e-01
2.77158320e-01 8.06015909e-01 -9.58261311e-01 2.42690504e-01
-8.46782863e-01 -8.27621043e-01 5.97794831e-01 3.15585226e-01
3.41292948e-01 -9.03691173e-01 -5.32259345e-01 1.89138800e-01
-5.03870308e-01 -1.07119977e+00 -4.55813855e-01 -7.37463459e-02
-5.54443300e-01 -1.18346453e+00 -6.41396403e-01 -6.28945351e-01
3.67656857e-01 1.34556657e-02 1.63852644e+00 -5.49995452e-02
3.66019607e-01 2.63323367e-01 -2.15888560e-01 -9.00172174e-01
-8.65877092e-01 6.35107219e-01 -2.87570268e-01 -2.62715220e-01
4.24236268e-01 -6.64599836e-02 -4.33171064e-01 -8.40231478e-02
-4.94718134e-01 4.81011532e-02 3.87204766e-01 1.11663854e+00
-2.16581628e-01 -5.82331538e-01 6.42438412e-01 -1.50658500e+00
1.12935197e+00 -7.20131636e-01 -3.35893691e-01 -1.86746884e-02
-7.97104657e-01 1.27365932e-01 4.76390541e-01 -2.85007954e-01
-1.06581962e+00 -8.97376537e-01 -6.97548464e-02 5.46496272e-01
1.58910766e-01 7.64847577e-01 3.11888754e-01 3.11494827e-01
9.84458685e-01 -2.12729067e-01 1.11902624e-01 -4.21264470e-01
1.37847766e-01 8.69072378e-01 2.51966864e-01 -7.41119742e-01
5.30675769e-01 4.75728273e-01 -3.90329093e-01 -9.03998673e-01
-1.71573484e+00 -1.84605196e-01 -1.64859220e-01 -3.21313851e-02
5.33161104e-01 -1.04142845e+00 -2.55298883e-01 1.53601438e-01
-1.26528454e+00 -2.18070537e-01 -1.42837539e-01 3.30146253e-01
-2.11816177e-01 3.44244450e-01 -1.15607810e+00 -6.82415009e-01
-4.33210731e-01 -8.69454563e-01 9.08448517e-01 -3.02395612e-01
-9.69934642e-01 -1.15431511e+00 2.94486433e-01 5.88417649e-01
2.77718037e-01 4.42922145e-01 1.10540330e+00 -8.73305380e-01
1.33714378e-01 -1.79346965e-03 -3.14370006e-01 -1.05533144e-02
-8.11356902e-02 1.80153668e-01 -7.87212372e-01 -1.92771807e-01
8.10827985e-02 -4.26502734e-01 1.09594405e+00 5.47237277e-01
6.27737701e-01 -9.28452849e-01 -1.73358679e-01 2.13172153e-01
1.05962193e+00 -4.16699320e-01 3.97601612e-02 1.06797171e+00
7.64681876e-01 7.68685758e-01 1.71694040e-01 2.21620277e-01
5.85289121e-01 4.90265459e-01 -1.72340795e-02 -1.41168535e-01
-6.16037585e-02 -2.56523043e-01 4.13044184e-01 8.28848302e-01
7.49477819e-02 -1.77697495e-01 -1.13271630e+00 7.51800895e-01
-1.84154916e+00 -1.32798386e+00 -4.99702096e-01 1.61227024e+00
1.13433504e+00 4.22554672e-01 3.66973758e-01 4.54429202e-02
5.94561934e-01 6.10041797e-01 -1.75698102e-01 -7.30601132e-01
-5.00873387e-01 -3.55927944e-01 6.07320428e-01 6.83766842e-01
-1.03171194e+00 8.33075643e-01 7.38778305e+00 3.51759374e-01
-1.09035814e+00 4.34178412e-02 5.11831164e-01 -1.40613064e-01
-6.36945426e-01 -1.16283581e-01 -8.85763049e-01 7.96025872e-01
6.22517526e-01 -2.16047138e-01 -7.29854852e-02 8.89645875e-01
2.75784701e-01 1.91397503e-01 -8.81583929e-01 3.16332668e-01
2.35730380e-01 -1.59285641e+00 7.95584917e-02 1.68135524e-01
9.74766791e-01 2.40708739e-01 -2.73635481e-02 3.15569013e-01
6.16040051e-01 -7.97902405e-01 1.26894355e+00 5.58329634e-02
4.49760169e-01 -2.87358224e-01 7.62336135e-01 4.39604551e-01
-1.54712945e-01 -4.48374957e-01 -1.04847610e-01 -5.96214592e-01
-6.02590665e-02 7.16716290e-01 -7.77936637e-01 1.98586628e-01
6.86070383e-01 6.10384285e-01 -5.89709640e-01 1.41922072e-01
-3.51380825e-01 1.20465839e+00 -1.47801423e-02 -4.70072508e-01
6.04897499e-01 1.97650000e-01 8.51505578e-01 1.67382598e+00
-1.22156419e-01 -2.00435042e-01 1.83789343e-01 6.21650815e-01
-3.29478920e-01 3.84034902e-01 -6.65224075e-01 -2.07753703e-01
4.45019424e-01 8.51510048e-01 -5.01637459e-01 -5.74503243e-01
-5.02818167e-01 5.65641820e-01 7.88989127e-01 3.79925072e-01
-4.78205711e-01 8.04952234e-02 4.01579261e-01 3.03060383e-01
1.78207412e-01 -9.12206769e-02 -7.57894754e-01 -1.35491955e+00
4.32642661e-02 -1.37841177e+00 6.01703286e-01 -1.98266998e-01
-1.52540886e+00 4.93673235e-01 -1.34638518e-01 -4.82244045e-01
-1.33261949e-01 -5.09467304e-01 -5.49450517e-01 7.13730097e-01
-1.47790265e+00 -1.23544323e+00 -1.69551343e-01 -3.97306029e-03
6.89136267e-01 -1.74037561e-01 4.91239369e-01 -1.74621865e-01
-4.93527770e-01 4.36891347e-01 3.51145238e-01 3.86729360e-01
1.04255295e+00 -1.46565652e+00 3.19004804e-01 5.25071383e-01
-4.48549613e-02 8.06253791e-01 1.12159073e+00 -6.72647953e-01
-8.63148987e-01 -7.94063628e-01 1.50932801e+00 -9.06855345e-01
1.05928612e+00 -4.09414917e-01 -7.62487769e-01 7.68698335e-01
6.45827770e-01 -8.87384772e-01 9.75608051e-01 8.81291628e-01
-8.91784608e-01 3.04747015e-01 -6.44977868e-01 8.75838518e-01
9.33914423e-01 -6.37344301e-01 -1.27114010e+00 3.73929918e-01
4.34405684e-01 -4.23136830e-01 -6.72231555e-01 1.11665696e-01
6.79159224e-01 -8.69230688e-01 5.34894526e-01 -9.76467490e-01
1.10732734e+00 2.11638510e-01 5.89234419e-02 -1.31285608e+00
-9.14488852e-01 -4.54769373e-01 -1.94553390e-01 1.30111420e+00
4.99290556e-01 -6.01653159e-01 3.87692541e-01 3.40940833e-01
3.18589360e-02 -5.07419407e-01 -5.32689571e-01 -3.69300097e-01
5.91458678e-01 -3.86628620e-02 2.93313473e-01 1.42209017e+00
1.80544600e-01 1.07199109e+00 -4.58752185e-01 -2.74315715e-01
4.16762829e-01 5.99340796e-01 7.08917618e-01 -1.41752505e+00
-2.82746017e-01 -9.81983185e-01 7.67645612e-02 -1.09018767e+00
5.05927265e-01 -9.25079942e-01 -2.34024420e-01 -1.51692116e+00
4.24933016e-01 -1.18718676e-01 1.61022525e-02 2.20312569e-02
-3.70839626e-01 9.14979801e-02 1.62628621e-01 5.73162913e-01
-4.31425393e-01 1.29199162e-01 1.20200419e+00 -3.24250638e-01
-7.08628669e-02 -3.34735572e-01 -1.37128985e+00 9.19343293e-01
7.54201233e-01 -5.16240656e-01 2.52734516e-02 -6.29598379e-01
6.24279022e-01 -5.47187448e-01 3.11959445e-01 -3.69655728e-01
-3.71015840e-03 -3.23713124e-01 3.70458722e-01 -1.83746919e-01
-2.68223643e-01 -2.08014771e-01 -5.64363956e-01 4.22088087e-01
-9.00558054e-01 5.11352979e-02 -6.01335950e-02 7.17966557e-01
-1.88390762e-01 -1.62641048e-01 6.71977520e-01 -4.20194209e-01
-8.62299949e-02 -2.07066119e-01 -8.12966347e-01 6.89568579e-01
4.30780113e-01 1.90480679e-01 -7.39762425e-01 -5.15400052e-01
-2.61972398e-01 1.79434493e-01 6.81233346e-01 4.61614609e-01
-2.97702938e-01 -1.18176770e+00 -1.17098701e+00 -1.69589937e-01
1.59907222e-01 -4.54413176e-01 -1.73603922e-01 9.41836894e-01
-3.55674744e-01 4.86042768e-01 1.83856681e-01 -4.19139832e-01
-1.16469753e+00 3.61453116e-01 3.99023294e-02 -4.74345773e-01
-5.41286349e-01 6.60637677e-01 7.76963383e-02 -2.41929099e-01
-8.54538828e-02 -2.68009543e-01 -4.63543624e-01 6.22308016e-01
4.29078788e-01 4.03584212e-01 -1.74186707e-01 -7.63805807e-01
-1.71520293e-01 3.29616964e-01 -4.62322921e-01 -1.54569179e-01
1.12389016e+00 -2.96006680e-01 -2.08723277e-01 9.20998037e-01
8.99929821e-01 7.62613058e-01 -7.79761612e-01 -4.15981263e-01
4.27536994e-01 -2.04282165e-01 2.24390447e-01 -8.02881241e-01
-2.16821104e-01 3.46430957e-01 -3.28731030e-01 7.28090942e-01
3.39894354e-01 2.07821712e-01 5.06687820e-01 2.81149387e-01
-2.70624489e-01 -1.57119346e+00 -6.63337633e-02 9.26896453e-01
1.11952591e+00 -1.49714029e+00 1.79776311e-01 -2.06843600e-01
-7.58834243e-01 1.05957031e+00 4.01342034e-01 -1.08742930e-01
3.74703377e-01 2.17133969e-01 5.59590101e-01 -5.85384369e-01
-7.27197051e-01 1.53803334e-01 4.86580640e-01 -4.51144278e-02
1.20250499e+00 -1.31033957e-01 -1.02038586e+00 5.28225839e-01
-6.57463849e-01 -4.97709483e-01 5.20197392e-01 9.55470026e-01
-4.28116143e-01 -6.99250579e-01 -5.47700644e-01 6.37988985e-01
-1.02068651e+00 -1.83493927e-01 -9.92302895e-01 7.21207798e-01
-3.98648232e-01 8.61416280e-01 1.76961482e-01 9.88305956e-02
2.47519374e-01 2.94385463e-01 2.86805481e-01 -6.08125389e-01
-8.82801950e-01 1.16679139e-01 7.61146486e-01 -2.18130857e-01
-5.74948132e-01 -7.04638362e-01 -5.91660619e-01 -5.10522783e-01
-2.53569838e-02 3.75741124e-01 3.61753374e-01 9.16886091e-01
3.68785113e-01 5.47924995e-01 5.34941018e-01 -3.98655802e-01
-9.93655264e-01 -1.32152772e+00 -6.49490580e-02 7.76102722e-01
8.04519236e-01 -5.14984846e-01 -5.21429121e-01 -8.98539871e-02] | [8.846879005432129, 10.099246978759766] |
18574af3-7965-4dcc-ace2-042a7ad2a4e3 | warping-of-radar-data-into-camera-image-for | 2012.12809 | null | https://arxiv.org/abs/2012.12809v2 | https://arxiv.org/pdf/2012.12809v2.pdf | Warping of Radar Data into Camera Image for Cross-Modal Supervision in Automotive Applications | We present an approach to automatically generate semantic labels for real recordings of automotive range-Doppler (RD) radar spectra. Such labels are required when training a neural network for object recognition from radar data. The automatic labeling approach rests on the simultaneous recording of camera and lidar data in addition to the radar spectrum. By warping radar spectra into the camera image, state-of-the-art object recognition algorithms can be applied to label relevant objects, such as cars, in the camera image. The warping operation is designed to be fully differentiable, which allows backpropagating the gradient computed on the camera image through the warping operation to the neural network operating on the radar data. As the warping operation relies on accurate scene flow estimation, we further propose a novel scene flow estimation algorithm which exploits information from camera, lidar and radar sensors. The proposed scene flow estimation approach is compared against a state-of-the-art scene flow algorithm, and it outperforms it by approximately 30% w.r.t. mean average error. The feasibility of the overall framework for automatic label generation for RD spectra is verified by evaluating the performance of neural networks trained with the proposed framework for Direction-of-Arrival estimation. | ['Reinhold Haeb-Umbach', 'Tobias Breddermann', 'Ridha Farhoud', 'Ernst Warsitz', 'Tai Fei', 'Christopher Grimm'] | 2020-12-23 | null | null | null | null | ['direction-of-arrival-estimation', 'scene-flow-estimation'] | ['audio', 'computer-vision'] | [ 8.91952276e-01 -1.95180804e-01 9.68881100e-02 -5.46677887e-01
-6.39072299e-01 -6.18836999e-01 7.04245389e-01 -2.10598618e-01
-6.78076327e-01 4.99192536e-01 -3.26706022e-01 -2.39968121e-01
-2.06819743e-01 -8.73646617e-01 -4.96396303e-01 -7.60655582e-01
-4.35660556e-02 4.12290603e-01 1.68502867e-01 8.95302445e-02
3.38426024e-01 1.23594093e+00 -1.84170938e+00 -7.55814184e-03
4.11894560e-01 1.24672055e+00 1.22932024e-01 9.61286783e-01
-1.54378787e-01 9.63791788e-01 -6.17525220e-01 -6.06709272e-02
5.48301876e-01 -1.98622286e-01 -1.89615369e-01 2.69491702e-01
1.05284441e+00 -3.29553396e-01 -5.33244014e-01 1.18663681e+00
1.47534117e-01 4.42987978e-01 7.76732862e-01 -1.27229655e+00
-1.39865756e-01 -2.39800885e-02 -2.07383856e-01 2.27590010e-01
-1.18291028e-01 -3.75292413e-02 8.19923997e-01 -9.84051585e-01
8.29432189e-01 8.13861489e-01 3.15763414e-01 2.88640678e-01
-9.27665472e-01 -7.51381040e-01 -2.94127073e-02 4.68909293e-01
-1.38670158e+00 -4.22816217e-01 1.21647429e+00 -6.83138669e-01
5.25541604e-01 1.48219362e-01 5.33546448e-01 5.94993114e-01
1.40975371e-01 1.94572434e-01 1.01960075e+00 -3.80622804e-01
3.50954890e-01 2.60338545e-01 2.42987812e-01 9.03811038e-01
2.08987057e-01 6.36909366e-01 -5.87262988e-01 3.29252958e-01
4.51551080e-01 -2.87380844e-01 -2.44808450e-01 -5.58981240e-01
-1.05908549e+00 8.01742017e-01 3.77325922e-01 5.36512062e-02
-3.97091538e-01 2.08541617e-01 3.55590135e-01 -2.70460024e-02
5.06832659e-01 2.14808881e-01 1.13875486e-01 1.53740391e-01
-1.19689453e+00 2.33966652e-02 8.28785360e-01 7.67167628e-01
1.07080686e+00 5.11245906e-01 7.09457174e-02 3.62761229e-01
2.23524764e-01 9.71493483e-01 -2.34200247e-02 -1.01713753e+00
3.79723817e-01 4.09333348e-01 4.87922579e-02 -1.12743592e+00
-4.84085590e-01 -5.70397198e-01 -5.28913021e-01 6.78772807e-01
3.36904705e-01 -2.08294570e-01 -9.58810151e-01 1.43512726e+00
4.83247519e-01 6.01496637e-01 3.56987655e-01 1.16783547e+00
5.26746154e-01 8.07913244e-01 -1.10549442e-01 -4.65783238e-01
9.82880533e-01 -6.48752928e-01 -7.63272643e-01 -1.84912011e-01
4.89475489e-01 -6.80216610e-01 3.88869196e-01 4.71308738e-01
-5.64991713e-01 -9.58736837e-01 -1.34413350e+00 3.23254168e-01
-4.70251530e-01 5.13168931e-01 2.05508754e-01 8.01633656e-01
-5.96731901e-01 3.89610916e-01 -5.12066782e-01 -1.22792557e-01
1.11951903e-01 1.82860300e-01 -2.28648201e-01 -1.18059814e-01
-1.09726608e+00 1.07802594e+00 5.23083270e-01 5.29193163e-01
-1.17361283e+00 -7.63944805e-01 -8.82942438e-01 -1.97914198e-01
3.69674116e-01 -2.16141343e-01 8.79668057e-01 -8.80418241e-01
-1.51227200e+00 7.46187806e-01 3.25644389e-02 -5.55303872e-01
5.11680365e-01 -2.42139235e-01 -7.40433335e-01 5.46402395e-01
-1.86720759e-01 7.14653134e-01 1.19063878e+00 -1.43372154e+00
-1.00844586e+00 -1.71647772e-01 -1.28684998e-01 1.47471175e-01
-7.69324526e-02 -2.19764322e-01 9.38657075e-02 -2.68041342e-01
-2.06388295e-01 -9.65033770e-01 1.43517941e-01 -1.27585024e-01
-2.65979737e-01 2.50643641e-01 1.33072078e+00 -6.23473346e-01
8.13862443e-01 -2.00823355e+00 -1.76282823e-01 4.66028571e-01
-1.21305466e-01 2.98847705e-01 -9.69164148e-02 7.01117441e-02
-2.41593227e-01 -6.86536968e-01 -4.85209376e-01 -3.86243202e-02
-2.08163977e-01 3.37032489e-02 -6.83810174e-01 9.50529158e-01
2.76197225e-01 4.04581755e-01 -6.74258053e-01 -3.22319597e-01
7.48820305e-01 5.73385417e-01 -1.07342310e-01 2.34711707e-01
-1.02960631e-01 4.42481995e-01 -1.33059233e-01 4.44646001e-01
8.09406102e-01 4.11827266e-01 -8.98617432e-02 -6.12752438e-01
-5.06313205e-01 -9.86331254e-02 -1.21682918e+00 1.33919036e+00
-6.35661185e-01 1.15713525e+00 1.99892800e-02 -1.22583354e+00
1.52340388e+00 1.27188563e-01 7.09687531e-01 -5.84922552e-01
2.51923859e-01 1.40441775e-01 -7.47232065e-02 -4.52005237e-01
6.56444371e-01 -4.95626897e-01 -5.76675050e-02 3.53249729e-01
3.39694284e-02 -2.21860111e-01 2.58089095e-01 -1.40437663e-01
8.91545236e-01 1.68671921e-01 -1.58921499e-02 -6.26665205e-02
1.17423558e+00 3.78147185e-01 2.55230099e-01 4.84119028e-01
-1.22911029e-01 3.59378643e-02 7.93155283e-03 -6.58668399e-01
-9.47000206e-01 -9.79821742e-01 -1.13660738e-01 8.88789713e-01
2.73710966e-01 2.85524338e-01 -7.65812159e-01 -7.67897069e-01
-7.45965317e-02 1.00363564e+00 -4.12326604e-01 -2.21502244e-01
-7.80714154e-01 -3.47415537e-01 5.00098646e-01 2.74567395e-01
6.28093898e-01 -8.24944437e-01 -1.09075594e+00 3.09595466e-01
1.44220933e-01 -1.85143995e+00 -5.61721511e-02 1.77101746e-01
-7.72457957e-01 -1.06746614e+00 -3.41208905e-01 -5.71300805e-01
6.59922183e-01 3.25083703e-01 6.44633412e-01 -4.82465476e-01
-6.21837676e-01 8.14679146e-01 -1.36884958e-01 -2.85996944e-01
-6.18556678e-01 -1.17313586e-01 6.40821606e-02 6.62296355e-01
3.09720695e-01 -2.67724276e-01 -1.94528013e-01 3.12076122e-01
-7.83257544e-01 -7.40240067e-02 5.94878674e-01 4.23478931e-01
4.04936045e-01 2.25565180e-01 1.78047881e-01 -7.31524706e-01
2.31382683e-01 5.37383407e-02 -1.29965711e+00 2.12254357e-02
-4.46110457e-01 1.96065843e-01 6.18283331e-01 -2.26326421e-01
-1.27797234e+00 6.06648326e-01 1.00920253e-01 -7.58625269e-01
-6.46531940e-01 2.05528349e-01 6.33701384e-02 -4.34301645e-01
5.99197865e-01 1.68974832e-01 -8.19574371e-02 -1.51286542e-01
5.92844725e-01 6.21012330e-01 1.10341692e+00 -2.00022593e-01
1.21103573e+00 8.13694537e-01 5.78206599e-01 -1.31848454e+00
-8.91684115e-01 -7.78298557e-01 -9.72035646e-01 -1.07511079e+00
1.18836033e+00 -6.98567986e-01 -6.97543323e-01 4.24555182e-01
-1.34285247e+00 6.45711496e-02 -3.80428404e-01 8.89899850e-01
-6.22944593e-01 1.57273799e-01 -5.22911735e-02 -1.16833150e+00
-1.71313047e-01 -9.15783465e-01 1.01644170e+00 -2.04560123e-02
2.75053233e-01 -1.02722347e+00 4.82698195e-02 3.85693282e-01
8.99357274e-02 3.93832177e-01 7.79499233e-01 -6.61772788e-01
-8.93588185e-01 -4.83113945e-01 -3.68472278e-01 7.09867775e-01
-1.58686563e-02 3.43725813e-04 -1.20406556e+00 6.91532865e-02
2.28029206e-01 5.68964109e-02 9.26541090e-01 4.10707504e-01
5.48758030e-01 1.49612620e-01 -1.83317646e-01 6.16445422e-01
1.66408145e+00 4.93709058e-01 2.84149438e-01 1.60792276e-01
9.04733777e-01 1.01496923e+00 9.40793753e-01 2.75890499e-01
-1.56081125e-01 7.90332675e-01 6.70156837e-01 1.11184448e-01
-3.07494104e-01 2.74085812e-02 6.75771177e-01 6.69927001e-01
5.45217544e-02 -7.09057450e-02 -8.35733354e-01 3.01033705e-01
-1.43672132e+00 -9.46105361e-01 -2.91704386e-01 2.30914617e+00
1.15178578e-01 3.44627529e-01 -2.40008399e-01 3.45118225e-01
9.84626889e-01 3.12700063e-01 -4.02046382e-01 -4.27776992e-01
2.43022859e-01 2.44820014e-01 9.06483829e-01 9.52073097e-01
-1.30022645e+00 8.64592791e-01 5.59026718e+00 5.67646503e-01
-1.43982553e+00 -7.85568543e-03 5.78848273e-02 7.09924698e-02
3.54536250e-02 -8.73830169e-02 -1.19163489e+00 1.11539528e-01
1.08650708e+00 2.55498830e-02 3.89582574e-01 6.39131367e-01
2.71085680e-01 -7.44994581e-02 -9.79116619e-01 8.84223998e-01
5.81477404e-01 -1.32504976e+00 2.90243085e-02 4.71244529e-02
5.11632919e-01 -2.15340108e-01 -1.38507068e-01 -4.49108295e-02
3.82454097e-02 -6.08064115e-01 7.60455191e-01 6.97747588e-01
6.62142634e-01 -9.21401918e-01 6.92160308e-01 2.26246789e-01
-1.39293408e+00 -1.81705624e-01 -2.94421583e-01 8.66758302e-02
5.97235501e-01 7.91837990e-01 -1.16302288e+00 5.44588208e-01
5.50247245e-02 6.57036304e-01 -3.59694749e-01 7.49037325e-01
-3.83234054e-01 5.03422499e-01 -2.99223721e-01 1.52894288e-01
4.74315703e-01 -5.19636214e-01 8.31145287e-01 1.37705767e+00
6.24036491e-01 -2.03715816e-01 2.47384116e-01 8.16913307e-01
1.87436387e-01 -1.56130940e-01 -1.00373387e+00 -1.24647589e-02
3.31692487e-01 1.61127818e+00 -8.49860787e-01 -3.16904664e-01
-1.41547561e-01 4.05293912e-01 -2.58988440e-01 4.21867460e-01
-1.04283953e+00 -5.03470838e-01 4.84867156e-01 -3.63192074e-02
4.27078903e-01 -4.60122496e-01 -7.09636062e-02 -4.27813023e-01
-2.72258252e-01 -2.25930348e-01 2.46388793e-01 -1.06762099e+00
-8.79510045e-01 7.49035120e-01 1.81003705e-01 -1.62125957e+00
-3.23896617e-01 -1.03928447e+00 -3.38521421e-01 6.80817902e-01
-1.84371901e+00 -1.19267130e+00 -8.31716299e-01 6.41886115e-01
4.62747544e-01 -4.87291425e-01 4.32696313e-01 3.57592195e-01
-3.12653393e-01 -2.75829621e-02 -3.75384152e-01 1.32150799e-01
5.69859564e-01 -8.86001468e-01 7.20746443e-02 1.14460588e+00
4.04556572e-01 -2.59898543e-01 7.49599099e-01 -7.14529157e-01
-1.14504087e+00 -1.45426095e+00 6.63377821e-01 2.94655859e-02
7.86131263e-01 -4.24539953e-01 -5.01030326e-01 4.78381544e-01
1.09840907e-01 1.30178690e-01 3.39181751e-01 -6.58182502e-01
-1.93214193e-01 -5.63803136e-01 -7.34097064e-01 1.46295518e-01
7.53092647e-01 -7.17838466e-01 -5.43325603e-01 2.43105844e-01
4.29266453e-01 -2.69512147e-01 -5.75866342e-01 4.12845492e-01
3.27295274e-01 -8.15726817e-01 1.06332684e+00 -3.28288317e-01
2.42136434e-01 -5.72908819e-01 -2.82144725e-01 -1.10599244e+00
-3.21004204e-02 -4.26416308e-01 -3.14127281e-02 9.52167928e-01
1.87194183e-01 -4.41835374e-01 9.44338799e-01 8.95671397e-02
-2.98849642e-01 5.29631339e-02 -1.07378316e+00 -8.58914554e-01
-5.13288379e-01 -8.38504970e-01 -2.87980549e-02 7.31802166e-01
-6.54660583e-01 3.87284487e-01 -4.36488092e-01 6.18065357e-01
1.04975581e+00 2.09795773e-01 6.74464822e-01 -1.49472082e+00
3.17586176e-02 -4.26532328e-02 -6.89272046e-01 -7.99789429e-01
7.00012326e-01 -1.03597927e+00 3.90889019e-01 -1.20348763e+00
-4.81235445e-01 -1.79652527e-01 -3.48609239e-02 1.69073157e-02
3.97240877e-01 4.16257888e-01 5.55740952e-01 1.53644159e-01
-3.77682708e-02 3.70030552e-01 1.13872015e+00 -2.59972185e-01
-1.06616803e-01 -5.42302877e-02 7.69003853e-02 8.38623881e-01
6.49917006e-01 -5.67013979e-01 -3.37168843e-01 -8.56580436e-02
-5.92395328e-02 1.79543033e-01 6.43490851e-01 -1.54340792e+00
6.40676141e-01 9.06283259e-02 2.08001375e-01 -8.67448390e-01
6.83399737e-01 -1.30113339e+00 1.33521184e-01 5.21340549e-01
-2.93502569e-01 -8.60498473e-02 7.56635293e-02 5.70393026e-01
-3.67548019e-01 -5.28284907e-01 9.04570520e-01 1.39035970e-01
-1.09660745e+00 5.87736666e-02 -6.63216412e-01 -1.43040583e-01
1.24106562e+00 -4.74706560e-01 -4.08286184e-01 -3.96841526e-01
-4.82971072e-01 -4.18175042e-01 -1.94600195e-01 2.34108165e-01
7.26945698e-01 -1.36073458e+00 -5.99009633e-01 5.63895166e-01
8.96326229e-02 -3.16105127e-01 1.62388131e-01 7.64015019e-01
-7.67737091e-01 6.31318033e-01 -4.52795744e-01 -6.65371776e-01
-1.26276612e+00 5.49379408e-01 4.85024244e-01 -1.24714933e-02
-5.26592791e-01 4.36639935e-01 1.37684077e-01 -1.81679517e-01
9.57806781e-02 -2.85461426e-01 -3.98947626e-01 3.98183107e-01
3.43549639e-01 5.91786861e-01 2.32003316e-01 -1.00468683e+00
-3.91930431e-01 1.02555859e+00 2.63432235e-01 -3.93256456e-01
1.02002072e+00 1.32407516e-01 1.85989007e-01 4.36582804e-01
1.27874875e+00 -7.95745403e-02 -1.38414097e+00 -4.63182032e-02
-1.03569411e-01 -3.18798244e-01 6.78229511e-01 -5.89530468e-01
-1.17612731e+00 1.12344873e+00 8.72760713e-01 1.27812713e-01
1.25221455e+00 -5.00104129e-01 5.44892132e-01 6.05262756e-01
2.34944582e-01 -1.23860693e+00 -1.69674251e-02 5.13347208e-01
5.75361431e-01 -8.18581879e-01 7.87414536e-02 -5.45910597e-01
-6.26188338e-01 1.50684166e+00 1.87111109e-01 -4.23898607e-01
5.73970377e-01 5.08033276e-01 2.75099188e-01 -2.42648438e-01
-4.77300167e-01 -4.14554775e-01 4.14764166e-01 6.04301929e-01
-1.02760345e-01 -1.65348545e-01 1.86193764e-01 -3.54118645e-01
-3.02568059e-02 -3.78838070e-02 6.72401845e-01 6.94647014e-01
-8.39147091e-01 -6.63187146e-01 -6.88078463e-01 7.05689192e-02
1.45782322e-01 9.61781517e-02 -3.94088268e-01 9.89818156e-01
2.76786417e-01 8.09857845e-01 4.84852940e-01 -4.87714708e-01
5.27671039e-01 2.71432728e-01 3.80789667e-01 -2.50381202e-01
-1.71274439e-01 -4.71964926e-02 3.10036838e-01 -5.16060889e-01
-7.97622800e-01 -6.63156986e-01 -1.23717284e+00 1.13322102e-01
-2.02301905e-01 1.02631234e-01 1.09920871e+00 9.41022515e-01
-1.22249655e-01 7.14343488e-01 8.53283226e-01 -9.83620465e-01
-3.74062687e-01 -5.53835750e-01 -7.23590255e-01 1.20652087e-01
4.33847636e-01 -9.08017576e-01 -6.03616059e-01 5.83190560e-01] | [8.036446571350098, -1.3646435737609863] |
425e82d3-ec5e-4283-bfd7-a8d2a8f03a82 | learning-3d-representations-from-2d-pre | 2212.06785 | null | https://arxiv.org/abs/2212.06785v1 | https://arxiv.org/pdf/2212.06785v1.pdf | Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders | Pre-training by numerous image data has become de-facto for robust 2D representations. In contrast, due to the expensive data acquisition and annotation, a paucity of large-scale 3D datasets severely hinders the learning for high-quality 3D features. In this paper, we propose an alternative to obtain superior 3D representations from 2D pre-trained models via Image-to-Point Masked Autoencoders, named as I2P-MAE. By self-supervised pre-training, we leverage the well learned 2D knowledge to guide 3D masked autoencoding, which reconstructs the masked point tokens with an encoder-decoder architecture. Specifically, we first utilize off-the-shelf 2D models to extract the multi-view visual features of the input point cloud, and then conduct two types of image-to-point learning schemes on top. For one, we introduce a 2D-guided masking strategy that maintains semantically important point tokens to be visible for the encoder. Compared to random masking, the network can better concentrate on significant 3D structures and recover the masked tokens from key spatial cues. For another, we enforce these visible tokens to reconstruct the corresponding multi-view 2D features after the decoder. This enables the network to effectively inherit high-level 2D semantics learned from rich image data for discriminative 3D modeling. Aided by our image-to-point pre-training, the frozen I2P-MAE, without any fine-tuning, achieves 93.4% accuracy for linear SVM on ModelNet40, competitive to the fully trained results of existing methods. By further fine-tuning on on ScanObjectNN's hardest split, I2P-MAE attains the state-of-the-art 90.11% accuracy, +3.68% to the second-best, demonstrating superior transferable capacity. Code will be available at https://github.com/ZrrSkywalker/I2P-MAE. | ['Hongsheng Li', 'Peng Gao', 'Yu Qiao', 'Liuhui Wang', 'Renrui Zhang'] | 2022-12-13 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhang_Learning_3D_Representations_From_2D_Pre-Trained_Models_via_Image-to-Point_Masked_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhang_Learning_3D_Representations_From_2D_Pre-Trained_Models_via_Image-to-Point_Masked_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-point-cloud-classification', '3d-point-cloud-linear-classification', 'few-shot-3d-point-cloud-classification'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 7.94692412e-02 3.54180008e-01 -4.43707943e-01 -2.95850098e-01
-1.11221099e+00 -5.23945987e-01 6.10071480e-01 -3.03569436e-01
-2.56745070e-02 1.70386672e-01 4.06619497e-02 -2.02015102e-01
2.16956720e-01 -6.77069426e-01 -1.22953379e+00 -6.10651612e-01
1.04909360e-01 4.07268286e-01 3.22804570e-01 -1.04631111e-01
1.95004061e-01 6.98836625e-01 -1.66860712e+00 4.68452573e-01
5.17545283e-01 1.31328022e+00 6.82106674e-01 3.78879100e-01
-1.97803617e-01 5.51742792e-01 -6.26221523e-02 -1.30695015e-01
4.81753021e-01 6.33227825e-02 -5.22288382e-01 2.38334239e-01
6.80722117e-01 -6.96060896e-01 -6.10734284e-01 7.12941289e-01
3.99391800e-01 -2.82698959e-01 7.12613225e-01 -8.89224052e-01
-7.64831483e-01 4.73156907e-02 -8.03906739e-01 -1.34721175e-01
2.94288725e-01 2.72062689e-01 9.29544806e-01 -1.45443642e+00
7.44921923e-01 1.03466392e+00 6.06685519e-01 5.20060480e-01
-1.27632868e+00 -6.20698810e-01 1.85927033e-01 -1.74184088e-02
-1.51491094e+00 -4.27769661e-01 1.12456501e+00 -5.50978839e-01
1.38265049e+00 -9.48914094e-04 6.68098152e-01 1.10261142e+00
7.98880607e-02 9.14756596e-01 1.12742972e+00 -3.89523059e-01
-7.45615140e-02 5.89339584e-02 -2.25264683e-01 9.22301769e-01
-1.09840050e-01 5.24549246e-01 -7.33150244e-01 2.44768962e-01
1.11029232e+00 2.27772236e-01 -2.58991271e-01 -5.93939722e-01
-1.13090682e+00 5.86737514e-01 8.52212846e-01 -3.98621783e-02
-4.18239981e-01 1.02505766e-01 -2.31386032e-02 1.44957915e-01
6.48822367e-01 2.47719482e-01 -6.65380478e-01 1.64988607e-01
-1.08235753e+00 -3.96974012e-02 2.37031311e-01 1.04984534e+00
1.16178226e+00 8.84069502e-02 1.77089497e-01 8.15883279e-01
5.29724717e-01 8.24225307e-01 1.52100921e-01 -8.83211672e-01
7.20819652e-01 7.31964767e-01 -1.71687052e-01 -7.52748072e-01
-2.59055674e-01 -5.58609605e-01 -8.44329476e-01 5.07041991e-01
3.07394639e-02 4.83027905e-01 -1.32493651e+00 1.44815636e+00
2.20002353e-01 1.39710223e-02 -7.32260421e-02 9.93214488e-01
8.87358069e-01 6.05621278e-01 -3.21023345e-01 1.88980564e-01
1.14258385e+00 -9.75655079e-01 -1.04845397e-01 -4.79731679e-01
4.38122779e-01 -7.19488859e-01 1.05584466e+00 2.53622502e-01
-1.14513934e+00 -7.95641482e-01 -1.08338606e+00 -4.14231271e-01
-2.33571589e-01 1.29599959e-01 4.49122012e-01 1.08291067e-01
-9.92753386e-01 3.17690462e-01 -1.04332709e+00 3.41190249e-02
8.13111424e-01 3.84305835e-01 -8.09045076e-01 -5.16846359e-01
-8.65314007e-01 9.29374039e-01 2.94992059e-01 2.44094580e-02
-1.26553714e+00 -9.51510012e-01 -1.09545553e+00 -8.92674774e-02
1.79530099e-01 -8.05661917e-01 1.00668085e+00 -6.73945129e-01
-1.25734568e+00 1.36930704e+00 -2.42628694e-01 -1.12071313e-01
3.74206394e-01 -1.17517687e-01 -1.51877701e-01 5.12868106e-01
3.39859486e-01 1.16805208e+00 1.01266444e+00 -1.80917668e+00
-3.99645358e-01 -4.86429006e-01 9.74142924e-02 2.52356082e-01
-2.89905053e-02 -5.12798727e-01 -8.02955627e-01 -6.40112698e-01
5.73898137e-01 -7.58708358e-01 -7.86528438e-02 3.05797070e-01
-5.37948072e-01 1.84948131e-01 7.17568338e-01 -7.04970062e-01
6.08065963e-01 -2.40497231e+00 1.68952182e-01 2.49092337e-02
4.39840466e-01 4.81391028e-02 -2.76047319e-01 2.63984352e-01
-2.75319844e-01 -9.28807706e-02 -2.64527202e-01 -7.48050749e-01
-6.11367710e-02 2.48266846e-01 -4.15144444e-01 4.81855094e-01
6.22319341e-01 1.07860196e+00 -6.60193861e-01 -3.12646508e-01
6.68675721e-01 5.98259628e-01 -8.20797622e-01 4.60387707e-01
-2.22425207e-01 5.29864550e-01 -4.22875434e-01 8.52703393e-01
9.25777197e-01 -5.97512841e-01 -2.79781461e-01 -4.82345313e-01
-1.10038728e-01 3.87074262e-01 -6.72115266e-01 2.20550489e+00
-7.42214739e-01 3.90685439e-01 -2.11389866e-02 -1.06695771e+00
8.76604795e-01 1.42905384e-01 5.98509014e-01 -8.56960654e-01
-1.37642503e-01 4.49958175e-01 -4.59918916e-01 -4.11613405e-01
2.39448488e-01 -2.10967496e-01 1.37289437e-02 5.88998571e-02
4.20541555e-01 -4.53960180e-01 -5.21370828e-01 1.55898452e-01
9.23222423e-01 5.17225802e-01 1.41003281e-01 1.79427430e-01
3.01151782e-01 5.33514246e-02 1.48354337e-01 4.97072935e-01
1.48413897e-01 1.15695167e+00 1.50211275e-01 -3.43514204e-01
-1.19962728e+00 -1.27064693e+00 -1.81843042e-01 5.75561643e-01
2.57791013e-01 -2.59567589e-01 -3.67859334e-01 -7.27493703e-01
1.70287549e-01 6.83380306e-01 -5.34877002e-01 -2.96561658e-01
-5.38175762e-01 -2.44344965e-01 2.99995959e-01 5.96152186e-01
5.74804425e-01 -5.66027522e-01 -3.89869958e-01 5.96904010e-02
-7.91191682e-02 -1.33049572e+00 -3.47899407e-01 6.41910791e-01
-1.00553775e+00 -8.03286850e-01 -6.62170947e-01 -7.76219547e-01
8.17513406e-01 6.85558796e-01 1.10260499e+00 1.38671383e-01
-7.45447725e-02 3.13531429e-01 -3.56374204e-01 -8.47488791e-02
-2.31098577e-01 1.49648622e-01 -8.47773626e-02 -2.85698235e-01
2.31625691e-01 -9.46728230e-01 -7.40857244e-01 3.43955129e-01
-8.17910194e-01 4.65510130e-01 1.07162464e+00 8.90688121e-01
1.14561105e+00 -4.01130497e-01 2.72411317e-01 -4.22560215e-01
-4.42346662e-01 -4.21842963e-01 -4.38859493e-01 -7.59021491e-02
-4.91146475e-01 7.81325921e-02 5.67375660e-01 -2.44085878e-01
-7.37755954e-01 4.15080696e-01 -5.54748178e-01 -1.10978937e+00
-2.55964756e-01 2.98360050e-01 -4.26893175e-01 -2.75937825e-01
5.15025973e-01 6.73517108e-01 7.90742710e-02 -6.83753848e-01
5.06999791e-01 5.95481515e-01 3.43187302e-01 -4.20618981e-01
1.07177889e+00 7.08415210e-01 -1.04534484e-01 -7.16006994e-01
-1.17233860e+00 -4.22788650e-01 -8.53163540e-01 -1.26064066e-02
9.83735025e-01 -1.38596952e+00 -3.04402858e-01 3.89266104e-01
-1.11557162e+00 -3.94993782e-01 -3.30623031e-01 4.83971566e-01
-7.89797008e-01 2.38707930e-01 -4.00727898e-01 -3.72431338e-01
-9.22913700e-02 -1.27269852e+00 1.66082597e+00 -2.16559678e-01
1.79224104e-01 -6.70946479e-01 -1.91303059e-01 6.18297756e-01
1.09166794e-01 4.47734855e-02 9.18005586e-01 -3.96690488e-01
-1.12549365e+00 -2.08670259e-01 -3.73748779e-01 5.55110693e-01
6.84340820e-02 -4.21035439e-01 -1.29614925e+00 -2.00089306e-01
1.90360785e-01 -4.62665796e-01 9.44297910e-01 3.77529055e-01
1.40511680e+00 -7.52014443e-02 -4.69006687e-01 1.10627711e+00
1.39093781e+00 -3.16151440e-01 5.10188162e-01 3.04944992e-01
1.04762590e+00 2.95048445e-01 4.76455063e-01 2.63570577e-01
6.18861496e-01 7.45956302e-01 8.40600371e-01 -2.74480224e-01
-4.76871073e-01 -8.40516090e-01 2.94453889e-01 9.92969036e-01
-1.13517329e-01 2.12461269e-03 -8.77302051e-01 4.05534595e-01
-1.41611540e+00 -8.16146910e-01 1.04509667e-01 1.99190319e+00
7.11624801e-01 3.04122776e-01 -3.18884343e-01 -3.40892375e-02
3.90859336e-01 5.11640370e-01 -6.38159990e-01 1.61461473e-01
-2.08632559e-01 1.34961277e-01 6.42660975e-01 5.82001269e-01
-9.99133348e-01 1.04799807e+00 5.08152723e+00 1.03835714e+00
-1.36066270e+00 2.46469066e-01 6.85042918e-01 -2.17935681e-01
-6.00458622e-01 6.09414233e-03 -9.42593396e-01 3.59223843e-01
5.98357260e-01 4.61697042e-01 4.46699411e-01 9.40251827e-01
1.34861156e-01 2.25686625e-01 -1.24776042e+00 1.26045203e+00
1.52112722e-01 -1.60745704e+00 1.56249329e-01 3.21147621e-01
7.49965370e-01 5.16859353e-01 1.23950586e-01 2.31000349e-01
-5.01224324e-02 -1.06965017e+00 1.07355022e+00 3.62165630e-01
1.28508115e+00 -4.06524986e-01 2.90683657e-01 5.56820512e-01
-1.06014645e+00 1.19201474e-01 -4.00583953e-01 1.86966449e-01
2.67656684e-01 7.23830521e-01 -7.28943050e-01 7.98055351e-01
7.16737628e-01 1.00080776e+00 -4.25921530e-01 6.18016779e-01
-4.14079219e-01 4.44576353e-01 -4.71341342e-01 4.95780557e-01
3.39139163e-01 1.29529074e-01 4.33930337e-01 8.80647182e-01
4.43345129e-01 -1.46221220e-02 1.56933710e-01 9.42821980e-01
-1.62222385e-01 -3.32176387e-01 -8.17394018e-01 2.98271000e-01
4.44773495e-01 1.04459524e+00 -4.91490364e-01 -2.04403698e-01
-7.21345246e-01 1.15860176e+00 5.71190357e-01 3.33041430e-01
-8.03593636e-01 9.93632227e-02 5.97098112e-01 4.76861149e-01
7.75925517e-01 -5.20029843e-01 -3.95089656e-01 -1.34264302e+00
2.51511693e-01 -8.04302514e-01 5.17988056e-02 -9.57924962e-01
-1.29355586e+00 7.51561999e-01 -3.51731991e-03 -1.57694292e+00
-1.89473987e-01 -8.97826254e-01 -1.88581392e-01 9.95299220e-01
-1.90718210e+00 -1.66788399e+00 -4.15075630e-01 5.77019036e-01
6.53911889e-01 -8.52402002e-02 8.04220319e-01 3.21997792e-01
-9.91179347e-02 4.58615512e-01 -7.46782050e-02 1.27479341e-03
6.15926206e-01 -1.01270223e+00 5.15080154e-01 5.88265121e-01
2.47712150e-01 3.42852443e-01 1.43844545e-01 -5.18797398e-01
-1.69477201e+00 -1.08775592e+00 6.50644779e-01 -6.89424276e-01
4.20889497e-01 -5.76850593e-01 -8.82702947e-01 7.64762878e-01
-1.46638095e-01 4.15234476e-01 5.25041282e-01 -2.25057706e-01
-6.79115951e-01 -4.86425720e-02 -8.72744977e-01 3.28748852e-01
1.35885119e+00 -8.60287964e-01 -6.90110624e-01 2.76749223e-01
8.54386806e-01 -7.11582124e-01 -8.56284320e-01 5.12885332e-01
3.44587147e-01 -1.10371482e+00 1.42872381e+00 -3.01629812e-01
9.32796776e-01 -4.17025834e-01 -7.08923936e-01 -1.19290066e+00
-2.80138463e-01 -2.00753018e-01 -3.47469121e-01 8.51947010e-01
4.28699970e-01 -4.92212266e-01 8.59059274e-01 2.41281971e-01
-6.84398532e-01 -1.12946010e+00 -1.10090244e+00 -7.74931192e-01
1.41427219e-01 -6.51184261e-01 4.70704108e-01 8.69468570e-01
-6.03199303e-01 2.56504685e-01 -2.85737574e-01 5.55580914e-01
6.76945269e-01 4.02694851e-01 8.02473545e-01 -7.75952041e-01
-4.26731408e-01 -2.91753680e-01 -4.92110640e-01 -1.84581459e+00
2.17619017e-01 -1.27107322e+00 -3.09245158e-02 -1.45866120e+00
4.24095802e-03 -6.62005663e-01 -2.40024224e-01 5.94586909e-01
2.88406946e-02 5.28074205e-01 1.60755336e-01 4.63982105e-01
-3.16130280e-01 8.94512236e-01 1.56892765e+00 -2.04879567e-01
-5.46112619e-02 -1.90534711e-01 -6.86572075e-01 6.56926453e-01
5.51687658e-01 -4.36492741e-01 -2.98944294e-01 -7.97141969e-01
-5.76453954e-02 5.83003834e-02 7.54683137e-01 -8.04436445e-01
2.24098447e-03 -3.89889963e-02 7.08252192e-01 -1.00484419e+00
8.01368773e-01 -9.61949706e-01 -3.23261926e-03 1.67488635e-01
5.93119152e-02 -1.77965268e-01 1.69938937e-01 5.00709534e-01
-4.54318762e-01 -1.03449911e-01 6.18087947e-01 -3.04290533e-01
-7.39425838e-01 6.58851802e-01 1.35682642e-01 -9.59664956e-02
8.71654689e-01 -4.85101819e-01 -2.09545568e-01 -2.02894270e-01
-6.88927114e-01 7.28062168e-02 8.66475761e-01 3.91957700e-01
9.70273256e-01 -1.34385777e+00 -5.09720504e-01 6.33242309e-01
3.39957088e-01 6.41385913e-01 5.23788512e-01 7.61426330e-01
-5.14870524e-01 3.93941224e-01 -1.29450604e-01 -1.03989375e+00
-8.20046246e-01 7.48581886e-01 2.57959932e-01 -3.69651169e-02
-9.64558184e-01 8.41337085e-01 5.12812078e-01 -5.97130358e-01
4.94217221e-03 -3.18667471e-01 2.92361766e-01 -1.62187889e-01
2.22567707e-01 -2.71266460e-01 2.39343971e-01 -6.93982244e-01
-4.61699724e-01 9.43134606e-01 -1.09326474e-01 -1.45753454e-02
1.71898401e+00 -8.09450522e-02 1.17071211e-01 3.85956287e-01
1.66275167e+00 -3.78503185e-03 -1.76171172e+00 -1.77815333e-01
-4.55594301e-01 -7.39879370e-01 3.17727000e-01 -6.69461489e-01
-1.17020369e+00 1.22885752e+00 5.29071510e-01 -9.21718553e-02
1.05348420e+00 3.62807482e-01 6.97743237e-01 2.20226675e-01
4.27028060e-01 -5.27873337e-01 2.51578301e-01 4.92188632e-01
1.01039171e+00 -1.53234446e+00 -5.57597838e-02 -5.97979844e-01
-5.50006986e-01 9.54431593e-01 7.14402795e-01 -3.17316741e-01
8.39054763e-01 1.71118632e-01 1.11536935e-01 -3.82976145e-01
-6.61960423e-01 -1.39995292e-01 4.23167348e-01 7.37656474e-01
9.39894691e-02 -1.02275148e-01 6.38534963e-01 5.10448515e-01
-1.40127897e-01 -1.81593388e-01 -1.82369573e-03 6.54629648e-01
-2.55872965e-01 -9.10828233e-01 -1.55182004e-01 3.53996605e-01
1.09186219e-02 -2.95223862e-01 -1.96094979e-02 8.18310738e-01
1.69995412e-01 3.99193078e-01 1.00889452e-01 -6.40949249e-01
4.13732827e-01 -6.14233315e-02 6.82910144e-01 -7.30098486e-01
-1.88242644e-01 1.42243251e-01 -1.19560786e-01 -7.57650733e-01
-3.94809902e-01 -3.97400916e-01 -1.18609893e+00 -8.00219774e-02
-1.99053362e-01 -3.64326507e-01 6.64264381e-01 8.18969369e-01
7.03474700e-01 4.56654966e-01 9.29680407e-01 -1.49552798e+00
-5.15467286e-01 -6.87879503e-01 -2.51860887e-01 3.21122259e-01
5.40786088e-01 -8.98881733e-01 -3.85970175e-01 -2.45305616e-02] | [8.235313415527344, -3.3830065727233887] |
3570a465-eeae-4eaa-b9c1-11d92b4fd9de | fighting-malicious-media-data-a-survey-on | 2212.05667 | null | https://arxiv.org/abs/2212.05667v1 | https://arxiv.org/pdf/2212.05667v1.pdf | Fighting Malicious Media Data: A Survey on Tampering Detection and Deepfake Detection | Online media data, in the forms of images and videos, are becoming mainstream communication channels. However, recent advances in deep learning, particularly deep generative models, open the doors for producing perceptually convincing images and videos at a low cost, which not only poses a serious threat to the trustworthiness of digital information but also has severe societal implications. This motivates a growing interest of research in media tampering detection, i.e., using deep learning techniques to examine whether media data have been maliciously manipulated. Depending on the content of the targeted images, media forgery could be divided into image tampering and Deepfake techniques. The former typically moves or erases the visual elements in ordinary images, while the latter manipulates the expressions and even the identity of human faces. Accordingly, the means of defense include image tampering detection and Deepfake detection, which share a wide variety of properties. In this paper, we provide a comprehensive review of the current media tampering detection approaches, and discuss the challenges and trends in this field for future research. | ['Yu-Gang Jiang', 'Larry S. Davis', 'Zuxuan Wu', 'Jingjing Chen', 'Chao Zhang', 'Zhenxin Li', 'Junke Wang'] | 2022-12-12 | null | null | null | null | ['face-swapping'] | ['computer-vision'] | [ 2.32217655e-01 -2.66559631e-01 -1.10395383e-02 8.62314180e-02
-3.56103450e-01 -6.22581244e-01 6.52417064e-01 -1.12891532e-01
-9.89568606e-02 4.86716211e-01 -1.46239743e-01 -3.26667666e-01
6.14900887e-01 -8.57710481e-01 -7.54778445e-01 -9.80992973e-01
1.10305727e-01 -4.58015501e-01 -3.91917452e-02 -3.15239951e-02
3.39337200e-01 2.97488034e-01 -1.49927533e+00 4.91851747e-01
5.02127945e-01 1.28716564e+00 -1.34404868e-01 7.30114460e-01
9.01379902e-03 1.10990441e+00 -7.43557513e-01 -1.16474152e+00
1.13997096e-02 -4.23992604e-01 -5.24665833e-01 4.80995029e-01
4.05321896e-01 -9.68593657e-01 -8.91772449e-01 1.57777798e+00
2.00806633e-01 -4.34172928e-01 2.50749946e-01 -1.76596797e+00
-1.02368915e+00 3.64403725e-01 -6.88497305e-01 9.54383090e-02
1.66287363e-01 1.76870719e-01 3.71661186e-01 -7.84602642e-01
4.01556492e-01 1.39831316e+00 4.61415529e-01 7.52228081e-01
-7.53720641e-01 -1.02960873e+00 -1.46448597e-01 5.21946132e-01
-1.40000856e+00 -6.86093271e-01 9.28339899e-01 -5.37498116e-01
2.80675352e-01 3.34354788e-01 4.50443715e-01 1.50232828e+00
5.05432189e-01 9.50730026e-01 1.12696052e+00 -2.29198277e-01
1.33660719e-01 2.91390866e-01 -4.51497257e-01 7.44299471e-01
3.37181330e-01 3.39295685e-01 -4.56229925e-01 -3.23565632e-01
6.87449694e-01 3.48851345e-02 -1.73828810e-01 -1.09679803e-01
-7.95011818e-01 1.09559476e+00 2.69468516e-01 1.96038023e-01
-1.32035092e-01 3.01976204e-01 5.48814654e-01 2.52204597e-01
2.19038516e-01 1.08490668e-01 3.19705576e-01 -3.05480342e-02
-7.77229667e-01 8.62048045e-02 4.82042491e-01 6.69427276e-01
6.43034518e-01 4.06697452e-01 3.93700719e-01 3.68430495e-01
4.04034674e-01 5.75091124e-01 3.18195909e-01 -6.90820158e-01
4.25244719e-01 3.75081122e-01 1.59572676e-01 -2.05355811e+00
2.03890622e-01 2.30453357e-01 -9.38296616e-01 1.80296689e-01
3.00178170e-01 -2.29031503e-01 -6.49805903e-01 1.37011564e+00
4.15185243e-01 2.47965500e-01 -1.41842887e-01 1.00171626e+00
8.64628315e-01 7.20349371e-01 9.59085152e-02 -1.66058794e-01
1.31807101e+00 -7.44844913e-01 -1.03553736e+00 -2.27181092e-01
4.49747294e-01 -9.13132787e-01 6.24404669e-01 4.41896558e-01
-9.37387586e-01 -5.50077379e-01 -1.24629354e+00 7.67665356e-02
-4.07108337e-01 -8.54156166e-02 6.60340488e-01 1.19089675e+00
-8.26756895e-01 3.44915301e-01 -6.65974379e-01 1.29347533e-01
9.64677095e-01 1.95088908e-01 -3.83500993e-01 -2.85881907e-01
-1.42008770e+00 5.04115701e-01 2.39600018e-01 3.87958199e-01
-1.22521329e+00 -9.77424011e-02 -8.23003888e-01 -2.15907633e-01
4.18039501e-01 -2.60449320e-01 1.09790683e+00 -1.42992556e+00
-1.19863081e+00 1.09997714e+00 2.37431943e-01 -3.87094080e-01
8.50693107e-01 -1.02339633e-01 -9.04213250e-01 3.66355687e-01
-1.44239441e-01 5.90789497e-01 1.72128952e+00 -1.32125616e+00
-4.46359396e-01 -3.43343943e-01 4.06009480e-02 -1.99643344e-01
-6.54262245e-01 4.55459714e-01 -1.65515661e-01 -6.86665833e-01
-3.17052275e-01 -8.89898419e-01 4.05275524e-02 3.23054403e-01
-6.32170975e-01 1.63606763e-01 1.31519818e+00 -8.77347112e-01
1.19937038e+00 -2.39377642e+00 -1.21956326e-01 -1.54564351e-01
6.76560819e-01 6.60578012e-01 -3.27946022e-02 4.99196708e-01
7.69611523e-02 5.05841494e-01 9.61127281e-02 -1.92137912e-01
-2.61391163e-01 -2.84916442e-02 -5.11206031e-01 8.46313596e-01
2.40029335e-01 1.00963879e+00 -8.50411534e-01 -3.95371646e-01
2.13633195e-01 4.87205952e-01 -2.31402695e-01 1.39785483e-01
-4.47112955e-02 3.09659511e-01 -5.42115450e-01 7.28977144e-01
1.02753162e+00 -2.10767239e-01 1.62517756e-01 -1.21359266e-01
8.44486356e-02 -1.21889368e-01 -7.49662936e-01 8.85734737e-01
-2.44669281e-02 1.00960612e+00 3.20435345e-01 -8.87702227e-01
6.26293361e-01 2.73678929e-01 2.05831096e-01 -5.82463861e-01
7.14652956e-01 1.59387469e-01 -6.84176981e-02 -7.97339737e-01
7.49448895e-01 -7.92705417e-02 -5.08033335e-02 5.13957798e-01
-3.27188075e-01 3.31245542e-01 -2.27952167e-01 2.12329879e-01
7.61312068e-01 -3.22638631e-01 1.43822610e-01 3.99858356e-01
3.32265049e-01 -3.63518775e-01 2.72645295e-01 5.42777956e-01
-6.32029235e-01 2.59369224e-01 4.39092904e-01 -4.49260414e-01
-1.09643292e+00 -7.87273228e-01 2.85714298e-01 8.28857541e-01
5.74181080e-01 -2.48666108e-01 -1.02698815e+00 -9.14037406e-01
3.77557315e-02 2.08448991e-01 -5.96058309e-01 -6.24611139e-01
-4.05340701e-01 -7.12123632e-01 9.42382514e-01 1.47462383e-01
9.51238215e-01 -1.03577912e+00 -3.15724105e-01 -7.47315586e-02
-4.04090464e-01 -1.53788579e+00 -3.23652774e-01 -7.29208410e-01
-5.04488647e-01 -1.19187629e+00 -5.31275094e-01 -7.48470068e-01
6.47392511e-01 7.74659455e-01 5.70281327e-01 5.04111886e-01
-1.76484793e-01 1.73894182e-01 -5.11489689e-01 -3.44088405e-01
-1.06305313e+00 -5.10381520e-01 9.01304036e-02 6.16795480e-01
4.08597231e-01 -6.81000203e-02 -4.87926930e-01 3.13763261e-01
-1.28721499e+00 -4.02437970e-02 3.78791064e-01 5.60677171e-01
-1.93788502e-02 6.46668017e-01 4.36399430e-01 -6.76229477e-01
5.39714992e-01 -5.48735023e-01 -5.23248434e-01 6.56374693e-02
-1.91768929e-01 -5.70236325e-01 5.01532495e-01 -7.56131113e-01
-7.95754910e-01 -3.49618942e-01 -1.17139056e-01 -6.48474455e-01
-1.79208010e-01 3.30438018e-01 -4.68198329e-01 -6.30900979e-01
2.67550975e-01 4.45937693e-01 1.66915208e-01 -1.18646808e-01
1.34501547e-01 9.04855311e-01 3.62224489e-01 1.32417217e-01
9.63984251e-01 7.49932766e-01 -2.69646913e-01 -1.21310818e+00
-4.06993151e-01 -2.47448124e-02 -1.94849610e-01 -5.64642727e-01
7.60633230e-01 -7.85539806e-01 -7.81173050e-01 1.39539576e+00
-1.34257984e+00 2.03831986e-01 4.81678724e-01 1.03443682e-01
-2.20724463e-01 1.07441926e+00 -9.94569480e-01 -8.45790505e-01
-7.80374631e-02 -1.31912506e+00 8.92704248e-01 8.47907811e-02
-5.31930998e-02 -9.00266469e-01 -5.05219698e-01 8.50688994e-01
2.02520549e-01 3.82566959e-01 7.55481482e-01 -4.10921842e-01
-8.24737608e-01 -6.04463875e-01 -3.54565233e-01 7.18866587e-01
8.07017162e-02 6.56201690e-02 -1.08274055e+00 -2.89128333e-01
3.05628359e-01 -4.56955582e-01 5.74920237e-01 7.50213712e-02
1.26244724e+00 -6.89503014e-01 -2.29597330e-01 4.48723435e-01
1.16216290e+00 4.03247237e-01 1.17536712e+00 2.36964226e-01
9.47118104e-01 4.85873848e-01 4.23730612e-01 5.33113062e-01
1.19444549e-01 4.34153527e-01 8.74971986e-01 -7.24108890e-02
1.69007316e-01 -5.46743512e-01 5.83823621e-01 6.05354011e-01
3.24716359e-01 -6.75781369e-01 -5.77840686e-01 7.25567639e-02
-1.53272128e+00 -1.08356833e+00 -1.88387364e-01 1.94973183e+00
4.58278149e-01 -1.17083564e-01 7.06620067e-02 3.37454915e-01
1.28319418e+00 3.93001109e-01 -6.22519255e-01 -3.76415163e-01
-2.17534021e-01 -3.67064029e-01 4.72291857e-01 1.55759662e-01
-1.35702658e+00 1.10320425e+00 6.37744474e+00 8.93710732e-01
-1.36098337e+00 1.59626752e-01 8.65474224e-01 2.90875971e-01
1.88201014e-02 -1.91629589e-01 -6.17943168e-01 8.94597471e-01
6.87162220e-01 2.10346907e-01 4.61571127e-01 7.47433364e-01
3.07581276e-01 -2.08534263e-02 -7.88453698e-01 1.30035460e+00
3.92584264e-01 -1.44963384e+00 2.15492681e-01 4.91558850e-01
6.31149769e-01 -4.75530744e-01 4.93051112e-01 -1.42159820e-01
-1.29982069e-01 -9.22710121e-01 8.55306983e-01 -1.43600553e-02
6.80435658e-01 -8.10703337e-01 7.19247460e-01 2.63332218e-01
-6.27530575e-01 -1.38236448e-01 -3.71697754e-01 -1.13273367e-01
1.12239212e-01 3.45466554e-01 -5.53593695e-01 8.10132623e-02
5.96297264e-01 6.49311304e-01 -4.19897169e-01 6.89263463e-01
-1.79322615e-01 6.44921541e-01 2.31172636e-01 -6.92305639e-02
1.68300182e-01 6.06330484e-02 5.84990084e-01 8.81673336e-01
2.19976202e-01 -2.20035687e-01 5.87914027e-02 8.87717605e-01
-4.30171102e-01 -1.32940903e-01 -6.54561698e-01 -7.23933101e-01
4.36408877e-01 1.09386384e+00 -6.02401435e-01 -1.79077700e-01
-4.37011153e-01 1.21077669e+00 -1.18942402e-01 1.41567916e-01
-1.03182125e+00 -2.18350053e-01 7.71777689e-01 2.49202728e-01
2.49386102e-01 -3.20685387e-01 2.16384470e-01 -1.33224916e+00
3.30720246e-02 -1.30471647e+00 1.55145586e-01 -7.65439689e-01
-1.24425626e+00 1.76377594e-01 -4.60920542e-01 -1.23080111e+00
6.56068325e-02 -5.40161073e-01 -3.80985945e-01 2.22395837e-01
-1.39714098e+00 -1.13767600e+00 -3.57684970e-01 5.79445124e-01
3.41578722e-01 -2.22697571e-01 3.87577981e-01 3.88803303e-01
-6.59155607e-01 7.92604923e-01 3.92694883e-02 6.97687566e-01
2.74741054e-01 -2.41724551e-01 5.06357074e-01 9.93227661e-01
5.49023598e-02 2.81269193e-01 6.47418976e-01 -6.71507001e-01
-1.60506630e+00 -1.00376403e+00 6.18368626e-01 -9.12179276e-02
8.65880907e-01 -5.19745290e-01 -8.81342351e-01 4.22759980e-01
1.70823097e-01 -1.69741422e-01 7.16831684e-01 -5.88390768e-01
-5.28453112e-01 1.25979081e-01 -1.39053333e+00 6.79278910e-01
5.69613636e-01 -7.99571574e-01 -6.91764429e-02 2.06867382e-01
4.90432113e-01 -3.76669206e-02 -4.00979578e-01 1.43946037e-02
6.45296752e-01 -1.16368651e+00 8.65813971e-01 -7.04244256e-01
7.35951483e-01 -3.86126488e-02 -1.54128596e-01 -9.30791795e-01
-1.53130412e-01 -9.58279610e-01 -4.50133801e-01 1.28431332e+00
-2.12004498e-01 -4.46830869e-01 8.41485023e-01 4.58191305e-01
3.63299638e-01 -2.85767674e-01 -7.89588988e-01 -6.26609504e-01
3.58993337e-02 -5.00172377e-01 4.42006767e-01 1.12304354e+00
-7.22353235e-02 -6.08936250e-02 -1.09588444e+00 3.89555424e-01
7.55779088e-01 -1.91059694e-01 6.84896588e-01 -6.49074912e-01
6.50523081e-02 -3.43051672e-01 -9.08570170e-01 -1.02480042e+00
1.89778879e-01 -4.74662185e-01 -2.65509605e-01 -7.37516701e-01
2.14791149e-01 -7.64412880e-02 1.03363313e-01 1.13427743e-01
-3.62278409e-02 7.24645734e-01 2.06496701e-01 2.73376763e-01
-4.99534011e-01 5.09720802e-01 1.35844612e+00 -2.84186482e-01
3.66231948e-01 -4.17445190e-02 -8.31081033e-01 8.90108943e-01
9.24403667e-01 -3.90373230e-01 -1.77996725e-01 -4.57388401e-01
5.14099836e-01 1.82661582e-02 7.95417845e-01 -7.36896932e-01
-6.31950051e-02 -2.87001610e-01 3.94408762e-01 -2.18337834e-01
4.27732646e-01 -7.96331346e-01 -1.90145507e-01 6.49068594e-01
-2.24794567e-01 9.14279372e-03 9.47702453e-02 8.90851080e-01
-2.75755197e-01 -2.13904709e-01 1.10660136e+00 -2.20210969e-01
-8.60010922e-01 4.34525549e-01 -9.40817058e-01 -1.36960477e-01
1.32895136e+00 -4.89769161e-01 -3.76347661e-01 -8.15642416e-01
-5.13962209e-01 -2.63222933e-01 4.72958177e-01 8.72314155e-01
1.08055282e+00 -1.36965454e+00 -5.73510289e-01 3.28012764e-01
6.42702281e-02 -6.83786035e-01 5.67505658e-01 4.17363584e-01
-4.65576500e-01 1.91746233e-03 -2.28485271e-01 -3.23283583e-01
-1.56889701e+00 1.01348543e+00 2.15032518e-01 3.61964464e-01
-1.91779256e-01 7.97594905e-01 3.96667868e-01 3.81133348e-01
7.12838471e-02 2.91276693e-01 2.88979951e-02 -5.33003323e-02
1.06087399e+00 6.15360320e-01 -1.33906379e-01 -1.12674761e+00
-1.94399029e-01 6.24761730e-02 -3.77359688e-01 3.25560093e-01
7.32836664e-01 -4.30918723e-01 -1.71651095e-01 -1.13857668e-02
1.41551602e+00 6.16749004e-02 -1.11581504e+00 1.79815497e-02
-3.49337965e-01 -1.05048585e+00 2.87621677e-01 -4.26566541e-01
-1.40755296e+00 1.05393052e+00 4.77163136e-01 6.05803311e-01
9.49263752e-01 -4.82644252e-02 1.24997962e+00 -9.08395350e-02
4.39818799e-01 -1.00299644e+00 5.04050851e-01 3.11310384e-02
6.88714027e-01 -1.41895604e+00 -1.52290925e-01 -5.60909808e-01
-6.59379125e-01 9.54195142e-01 4.77377117e-01 -4.42198925e-02
6.24639869e-01 2.54937291e-01 5.61168604e-02 -1.20279342e-01
-2.83892393e-01 3.39410037e-01 -6.83024600e-02 7.92101979e-01
2.34313235e-02 8.33324939e-02 1.50908411e-01 3.00985932e-01
-3.49406782e-03 -1.69036344e-01 7.41924405e-01 8.78055513e-01
-3.51412594e-01 -8.74759674e-01 -5.92588186e-01 1.41892210e-01
-7.56585658e-01 1.96932182e-02 -8.08546782e-01 5.36926866e-01
3.42908889e-01 1.33029187e+00 -1.01427577e-01 -8.48600924e-01
-3.34235549e-01 -3.89427006e-01 4.66696411e-01 1.19991936e-02
-2.44475693e-01 -7.57251829e-02 -8.08218494e-02 -3.31007630e-01
-2.80344397e-01 -6.22114718e-01 -6.61489129e-01 -1.03252006e+00
-5.73392272e-01 -2.55465865e-01 6.83581889e-01 9.77520645e-01
4.32023257e-01 -1.63826004e-01 9.57395494e-01 -6.98301136e-01
-5.76137662e-01 -5.25255322e-01 -6.31927609e-01 4.38588768e-01
5.49200892e-01 -6.27016664e-01 -4.25687253e-01 1.72809482e-01] | [12.653860092163086, 1.1642169952392578] |
0a647ee2-0e7a-4eda-b478-1849448055f9 | bag-of-tricks-for-effective-language-model | 2302.09268 | null | https://arxiv.org/abs/2302.09268v1 | https://arxiv.org/pdf/2302.09268v1.pdf | Bag of Tricks for Effective Language Model Pretraining and Downstream Adaptation: A Case Study on GLUE | This technical report briefly describes our JDExplore d-team's submission Vega v1 on the General Language Understanding Evaluation (GLUE) leaderboard, where GLUE is a collection of nine natural language understanding tasks, including question answering, linguistic acceptability, sentiment analysis, text similarity, paraphrase detection, and natural language inference. [Method] We investigate several effective strategies and choose their best combination setting as the training recipes. As for model structure, we employ the vanilla Transformer with disentangled attention as the basic block encoder. For self-supervised training, we employ the representative denoising objective (i.e., replaced token detection) in phase 1 and combine the contrastive objective (i.e., sentence embedding contrastive learning) with it in phase 2. During fine-tuning, several advanced techniques such as transductive fine-tuning, self-calibrated fine-tuning, and adversarial fine-tuning are adopted. [Results] According to our submission record (Jan. 2022), with our optimized pretraining and fine-tuning strategies, our 1.3 billion model sets new state-of-the-art on 4/9 tasks, achieving the best average score of 91.3. Encouragingly, our Vega v1 is the first to exceed powerful human performance on the two challenging tasks, i.e., SST-2 and WNLI. We believe our empirically successful recipe with a bag of tricks could shed new light on developing efficient discriminative large language models. | ['DaCheng Tao', 'Yibing Zhan', 'Li Shen', 'Bo Du', 'Juhua Liu', 'Keqin Peng', 'Liang Ding', 'Qihuang Zhong'] | 2023-02-18 | null | null | null | null | ['linguistic-acceptability'] | ['natural-language-processing'] | [ 1.77643761e-01 1.71135310e-02 -2.42581397e-01 -5.13637483e-01
-1.30808377e+00 -7.11513281e-01 6.14893198e-01 -7.59978546e-03
-5.54141223e-01 5.72458148e-01 5.99495649e-01 -3.72053295e-01
1.63731232e-01 -5.39242148e-01 -8.56279612e-01 -4.55803305e-01
3.64442557e-01 5.47605336e-01 -2.49651730e-01 -5.89682400e-01
1.96032673e-01 -1.10645056e-01 -1.08294165e+00 5.92438340e-01
1.04933119e+00 9.98408318e-01 6.69431910e-02 7.48659492e-01
-2.90429026e-01 9.95179296e-01 -5.88509917e-01 -8.89080346e-01
-7.66980425e-02 -2.10790381e-01 -1.13501930e+00 -4.49495286e-01
3.89636487e-01 -4.85019505e-01 -1.72510356e-01 9.31766748e-01
6.83715105e-01 1.32237241e-01 6.11482441e-01 -1.05278039e+00
-1.02838826e+00 9.76990044e-01 -3.76203656e-01 2.43586063e-01
3.19633961e-01 4.56233233e-01 1.49277842e+00 -1.12398839e+00
3.17289263e-01 1.56812131e+00 6.80676699e-01 9.05146003e-01
-1.27866590e+00 -7.74550259e-01 2.08140492e-01 2.61940837e-01
-1.07133758e+00 -6.91669166e-01 6.74895585e-01 -2.29727775e-01
1.37551486e+00 1.65413573e-01 1.83734909e-01 1.57928312e+00
1.86332360e-01 1.28469634e+00 1.02992904e+00 -4.56481129e-01
1.46397129e-01 3.27784240e-01 4.55736190e-01 7.57960021e-01
-7.77394325e-02 3.39161456e-02 -5.93759596e-01 -2.50360757e-01
1.15097515e-01 -3.48756611e-01 -1.61051974e-01 8.39234516e-02
-1.16624010e+00 1.13312304e+00 3.13904285e-01 1.44208550e-01
-9.12453160e-02 2.53664911e-01 8.43308747e-01 5.32079875e-01
7.82074153e-01 7.14573562e-01 -9.45292830e-01 -3.33028287e-01
-7.92867482e-01 2.72855252e-01 9.01038766e-01 8.26453567e-01
6.72711909e-01 8.29899311e-02 -4.24334407e-01 1.11262178e+00
3.02857637e-01 4.46118444e-01 5.24320841e-01 -8.71826589e-01
7.42390215e-01 4.40828294e-01 -3.02013367e-01 -5.74624598e-01
-2.51596630e-01 -4.67040122e-01 -1.00723279e+00 -2.81995356e-01
2.02333301e-01 -3.22693914e-01 -8.90701115e-01 1.98695254e+00
-4.42685820e-02 8.71173441e-02 3.11385572e-01 6.27002299e-01
1.16461325e+00 6.67645633e-01 2.71065325e-01 2.22339988e-01
1.49464750e+00 -1.27876222e+00 -5.94574153e-01 -5.55853307e-01
9.27877247e-01 -8.21432531e-01 1.55688322e+00 4.16326970e-01
-1.09753931e+00 -6.30105674e-01 -1.08714342e+00 -5.76916516e-01
-4.11788493e-01 3.03455070e-02 7.16471195e-01 5.28301895e-01
-9.88600612e-01 3.97994161e-01 -6.79608345e-01 -2.32912540e-01
2.89759248e-01 1.91346243e-01 -3.04658920e-01 -1.90459251e-01
-1.63774014e+00 1.02572846e+00 2.96023279e-01 8.78490061e-02
-1.00426805e+00 -9.51594830e-01 -8.84378135e-01 1.27939954e-01
3.48845810e-01 -1.27042091e+00 1.32167852e+00 -7.24390626e-01
-1.61498046e+00 1.26431561e+00 -1.94716439e-01 -5.99869132e-01
2.19661951e-01 -6.44936502e-01 -2.37796828e-01 -1.29663154e-01
1.29744336e-01 5.88775158e-01 8.45753253e-01 -1.03961849e+00
-9.14290398e-02 -2.19124362e-01 1.90872371e-01 1.58025667e-01
-4.03022677e-01 1.56861946e-01 -3.75429392e-01 -7.25215256e-01
-4.31171685e-01 -6.17643058e-01 -1.29393741e-01 -5.35817027e-01
-6.93398774e-01 -4.26159352e-01 3.68462384e-01 -7.84437120e-01
1.32011890e+00 -2.12186480e+00 2.87519425e-01 -8.24302062e-02
3.81160825e-01 2.72580773e-01 -5.11265934e-01 4.80261773e-01
-1.11462735e-01 3.92863601e-01 -2.69721389e-01 -8.40217829e-01
3.67669761e-01 9.23350602e-02 -6.30859375e-01 7.78244585e-02
3.36790919e-01 1.32597828e+00 -9.45778430e-01 -2.21888244e-01
3.88578735e-02 1.79379344e-01 -8.41860592e-01 4.51688588e-01
-4.01950657e-01 1.34112045e-01 -5.88413298e-01 5.05562901e-01
4.98071820e-01 -4.15731460e-01 -1.20263390e-01 -4.09978479e-01
2.80097276e-01 7.85337985e-01 -6.69152975e-01 1.98720133e+00
-8.04038823e-01 4.47637588e-01 2.03001887e-01 -9.99458313e-01
7.83058822e-01 1.29507929e-01 -7.27916285e-02 -6.25747621e-01
8.12344998e-02 1.06316730e-01 -2.34456569e-01 -6.18726969e-01
4.44150448e-01 -2.58229762e-01 -3.75834078e-01 5.17303646e-01
4.26168442e-01 -5.51055849e-01 -5.01050912e-02 6.56455994e-01
1.21403444e+00 -6.35086074e-02 3.45967203e-01 -3.74019891e-01
5.23399353e-01 -1.80959478e-01 2.20499024e-01 1.01547515e+00
-1.10245198e-01 4.98896509e-01 6.21406496e-01 -1.39442116e-01
-9.59715903e-01 -1.15278578e+00 1.69430062e-01 1.43251717e+00
-1.85854405e-01 -6.24067545e-01 -9.36754584e-01 -7.23000407e-01
2.96361968e-02 1.01842737e+00 -7.70706177e-01 -5.81568539e-01
-4.32278723e-01 -8.18205893e-01 1.02258801e+00 5.79793453e-01
6.11067951e-01 -1.11908257e+00 1.98061794e-01 1.19721524e-01
-4.41161007e-01 -9.60169911e-01 -7.73220658e-01 3.74424011e-01
-6.08605802e-01 -7.41718829e-01 -3.57925236e-01 -7.88716555e-01
2.68070847e-01 -3.40404026e-02 1.69183922e+00 1.01070605e-01
-2.11847728e-04 3.60240579e-01 -3.57420027e-01 -2.40905911e-01
-5.02971947e-01 4.10250813e-01 3.85078490e-02 -1.95809782e-01
4.94788408e-01 -3.33410561e-01 -5.25624752e-01 -3.43082007e-03
-8.43581378e-01 3.11627220e-02 4.96808171e-01 1.38074243e+00
4.47581440e-01 -5.19820690e-01 7.19955623e-01 -1.10934603e+00
1.11548126e+00 -5.66398501e-01 -2.63418138e-01 5.60201406e-01
-5.44518709e-01 3.29357296e-01 7.10368216e-01 -2.29649097e-01
-1.10155845e+00 -5.08503735e-01 -5.14808834e-01 -2.02828631e-01
2.29319967e-02 5.49037337e-01 -4.16903555e-01 2.54081190e-01
7.84094989e-01 2.28165492e-01 -1.73294663e-01 -4.90800083e-01
8.09313476e-01 7.16252267e-01 5.01547813e-01 -8.03535998e-01
6.51035845e-01 1.22496270e-01 -7.10910797e-01 -4.81891245e-01
-1.37816942e+00 -3.99788678e-01 -1.78309679e-01 3.65393102e-01
9.49208796e-01 -1.21016765e+00 -8.90373826e-01 5.78181982e-01
-1.37915874e+00 -5.38267434e-01 -2.22167328e-01 2.18380064e-01
-4.75947082e-01 3.40186238e-01 -8.48402262e-01 -4.20470297e-01
-9.34721589e-01 -1.12727892e+00 1.24517274e+00 -1.61020942e-02
-4.65990782e-01 -1.19808686e+00 2.87103832e-01 8.84998858e-01
5.91307878e-01 -1.72589332e-01 1.11952341e+00 -8.68048728e-01
-3.45342010e-01 1.68335885e-01 -2.25507230e-01 9.21406686e-01
-5.21503724e-02 -1.79976672e-01 -1.25251496e+00 -3.36536020e-01
1.93838477e-01 -9.74999428e-01 1.27891898e+00 2.68774122e-01
1.33939826e+00 -2.96388626e-01 -5.59661761e-02 8.05141807e-01
1.14316392e+00 -3.10455948e-01 4.91804332e-01 2.67027467e-01
7.76095808e-01 4.11993235e-01 3.59154880e-01 2.41459250e-01
6.24728143e-01 4.75611746e-01 1.95241123e-01 -2.84797214e-02
-8.75850618e-02 -4.31942523e-01 6.58366561e-01 1.21971786e+00
3.63370180e-01 -5.11566699e-01 -7.83762097e-01 3.64528596e-01
-1.69905448e+00 -8.98543656e-01 2.37623185e-01 1.88107336e+00
1.06599689e+00 3.60599369e-01 -3.65337968e-01 -3.21395606e-01
3.48566562e-01 4.65362191e-01 -7.30776131e-01 -7.17039943e-01
-3.82294267e-01 4.48829681e-01 2.23950043e-01 8.02515805e-01
-9.92845058e-01 1.27441585e+00 6.10729599e+00 1.18598759e+00
-8.16277266e-01 3.25930357e-01 7.85953104e-01 -1.37477964e-01
-8.50947976e-01 -7.65967071e-02 -9.08223629e-01 3.49818289e-01
1.02587283e+00 -2.17297658e-01 7.04047084e-01 6.43294036e-01
9.50828567e-02 1.45205513e-01 -1.22511828e+00 7.96536624e-01
2.18958452e-01 -1.37358057e+00 8.11618716e-02 -4.88207877e-01
7.01840580e-01 4.17566568e-01 1.77951127e-01 9.99274313e-01
6.06189668e-01 -1.29071140e+00 4.80956644e-01 4.18912172e-01
7.90985703e-01 -6.77544713e-01 7.80906320e-01 3.75971347e-01
-8.08784723e-01 8.35862160e-02 -1.24144487e-01 6.29209429e-02
3.02186787e-01 6.95755482e-01 -6.33020461e-01 4.18065429e-01
5.58294713e-01 7.78414071e-01 -3.59377325e-01 3.57607007e-01
-4.26512867e-01 1.12238538e+00 -2.90321946e-01 -1.73922896e-01
2.29220957e-01 -2.94494499e-02 5.31774759e-01 1.38142133e+00
-2.35617369e-01 1.76102947e-02 -1.11116312e-01 1.05952370e+00
-5.18074036e-01 7.18672723e-02 -4.20601606e-01 -7.14401677e-02
4.62751746e-01 1.09342790e+00 -5.69766238e-02 -5.50873280e-01
-3.79025102e-01 1.06486928e+00 7.19526827e-01 4.66154605e-01
-7.20786333e-01 -4.61117625e-01 8.20311844e-01 -4.02221888e-01
2.17924237e-01 -2.28969995e-02 -5.76623201e-01 -1.56121886e+00
-1.91398710e-03 -1.24006641e+00 4.19866800e-01 -7.63354778e-01
-1.79156828e+00 7.98056960e-01 -6.88280240e-02 -6.81940854e-01
-4.47444350e-01 -6.47066236e-01 -6.76682532e-01 9.45329785e-01
-1.50626934e+00 -1.25524163e+00 2.16745446e-03 6.24999285e-01
6.54464364e-01 -2.24295869e-01 1.03094876e+00 2.10546270e-01
-6.92759216e-01 1.08313823e+00 3.12453687e-01 1.15871377e-01
9.12971139e-01 -1.35309577e+00 8.23198557e-01 6.19517803e-01
3.07433419e-02 8.67061853e-01 6.09881938e-01 -4.14766252e-01
-1.37176633e+00 -8.89818609e-01 1.01393390e+00 -7.37954378e-01
1.11084044e+00 -8.23373497e-01 -9.49653924e-01 7.32891142e-01
3.97309244e-01 -3.09202373e-01 7.14483380e-01 3.99817497e-01
-6.72385335e-01 -1.56837001e-01 -9.90743220e-01 5.87125182e-01
9.44281578e-01 -9.85079169e-01 -8.84664536e-01 4.62703437e-01
1.29790556e+00 -4.09682840e-01 -7.91686475e-01 3.72317642e-01
3.03275913e-01 -6.98595703e-01 1.04365849e+00 -9.66760516e-01
7.50807166e-01 2.35491082e-01 -2.52631873e-01 -1.55601573e+00
-2.40801632e-01 -7.23656893e-01 -5.88930734e-02 1.41081536e+00
5.95825791e-01 -7.05991983e-01 5.87582171e-01 4.82059449e-01
-2.81726003e-01 -1.04330695e+00 -8.01066816e-01 -4.93844509e-01
4.97238815e-01 -6.58874512e-01 3.95050496e-01 9.01341558e-01
-6.79715574e-02 9.02149737e-01 -3.44896495e-01 -2.21816227e-01
6.06061697e-01 2.08504163e-02 7.76656568e-01 -7.70672143e-01
-6.88409984e-01 -4.99038160e-01 2.50882983e-01 -1.44342482e+00
6.15770638e-01 -1.04492164e+00 1.67529862e-02 -1.30810440e+00
3.93299848e-01 -1.87921613e-01 -4.07685459e-01 5.37144005e-01
-5.90814769e-01 -1.05037928e-01 1.48841113e-01 -1.51865900e-01
-6.01402819e-01 9.19392228e-01 1.14668715e+00 -4.14374262e-01
2.43191682e-02 -2.34357603e-02 -1.20008159e+00 5.14709473e-01
8.38140309e-01 -3.64190131e-01 -3.49761695e-01 -7.50650883e-01
3.65366399e-01 -2.29008824e-01 4.45125371e-01 -2.56408840e-01
7.25122020e-02 1.22419029e-01 -1.37051836e-01 -3.37283254e-01
5.06881595e-01 -2.07668290e-01 -4.81226325e-01 2.78271049e-01
-7.41064191e-01 1.48472518e-01 5.00373363e-01 3.16669971e-01
-4.07693028e-01 -2.65235454e-01 5.75287521e-01 -2.13036180e-01
-6.86225414e-01 1.17259495e-01 -1.70033604e-01 7.81504571e-01
3.51337463e-01 2.01616719e-01 -6.03493929e-01 -4.45219696e-01
-5.84221244e-01 5.88899970e-01 6.07575662e-03 6.58856869e-01
5.89336693e-01 -1.03455174e+00 -1.16887975e+00 1.25472918e-01
1.67033702e-01 -3.06160804e-02 4.49872136e-01 6.05099261e-01
-1.00388654e-01 4.48448151e-01 3.17718267e-01 -3.11740696e-01
-1.02437997e+00 3.63091946e-01 3.48658323e-01 -8.95291924e-01
-1.96420580e-01 1.38123310e+00 4.77688074e-01 -8.66193831e-01
1.98612243e-01 -4.09390181e-01 -3.31028625e-02 -1.10175170e-01
4.91999030e-01 1.30226165e-01 2.17613563e-01 -8.71090293e-02
-5.23050547e-01 3.99297893e-01 -5.55903137e-01 -3.37074660e-02
1.27582896e+00 -9.12190080e-02 -2.69123077e-01 3.52927536e-01
1.61948800e+00 1.82279032e-02 -8.65444422e-01 -4.90665793e-01
-1.78840756e-01 6.24180324e-02 1.47564292e-01 -1.10575199e+00
-8.29275846e-01 9.98448670e-01 2.17233509e-01 2.32111067e-02
1.15117693e+00 1.87351316e-01 9.99946415e-01 5.54626167e-01
-1.02603401e-03 -9.77133095e-01 2.08927810e-01 1.01487446e+00
1.07305586e+00 -1.42491508e+00 -1.70339242e-01 -1.22698903e-01
-8.33788157e-01 7.04118550e-01 7.38348424e-01 -5.19752502e-02
5.71251571e-01 2.94548422e-01 1.43332556e-01 -2.36433119e-01
-1.28396833e+00 7.06180707e-02 3.57055157e-01 1.94885328e-01
5.00898123e-01 6.81136101e-02 -2.32524872e-02 9.24923182e-01
-4.78991181e-01 -1.67852029e-01 1.12379991e-01 3.27736288e-01
-1.66148826e-01 -9.96874094e-01 -5.16361669e-02 5.14658272e-01
-4.49979067e-01 -8.02700877e-01 -3.35522234e-01 5.26079118e-01
-1.60902247e-01 1.10759521e+00 -1.94494739e-01 -5.45495331e-01
4.97852445e-01 5.80725782e-02 3.91736627e-01 -6.69757366e-01
-1.08259606e+00 -5.13909280e-01 3.76789689e-01 -6.81137145e-01
-1.10129341e-01 -3.50879967e-01 -1.08490431e+00 -4.54885572e-01
-3.29280406e-01 1.51384711e-01 4.98259217e-01 1.23307288e+00
4.93640721e-01 4.31822598e-01 5.43419063e-01 -4.32568043e-01
-1.08872807e+00 -1.15825248e+00 -1.60070151e-01 4.49882537e-01
4.98481482e-01 -3.14044207e-01 -5.58465242e-01 -1.22074127e-01] | [10.966463088989258, 8.247320175170898] |
57bba24c-28d7-497c-9d8e-7938559d865b | finread-a-transfer-learning-based-tool-to | null | null | https://aclanthology.org/2021.icon-main.81 | https://aclanthology.org/2021.icon-main.81.pdf | FinRead: A Transfer Learning Based Tool to Assess Readability of Definitions of Financial Terms | Simplified definitions of complex terms help learners to understand any content better. Comprehending readability is critical for the simplification of these contents. In most cases, the standard formula based readability measures do not hold good for measuring the complexity of definitions of financial terms. Furthermore, some of them works only for corpora of longer length which have at least 30 sentences. In this paper, we present a tool for evaluating readability of definitions of financial terms. It consists of a Light GBM based classification layer over sentence embeddings (Reimers et al., 2019) of FinBERT (Araci, 2019). It is trained on glossaries of several financial textbooks and definitions of various financial terms which are available on the web. The extensive evaluation shows that it outperforms the standard benchmarks by achieving a AU-ROC score of 0.993 on the validation set. | ['Sunny Kumar Singh', 'Sudip Naskar', 'Shovon Sengupta', 'Sohom Ghosh'] | null | null | null | null | icon-2021-12 | ['sentence-embeddings', 'sentence-embeddings'] | ['methodology', 'natural-language-processing'] | [-2.49251142e-01 2.35889912e-01 -1.03033073e-01 -1.49142370e-01
-3.75108212e-01 -8.79720867e-01 7.88522542e-01 8.46773088e-01
-7.37016678e-01 5.90539396e-01 4.36439425e-01 -6.31208122e-01
-3.72149378e-01 -9.26425636e-01 -4.73354459e-01 -5.71317673e-02
2.64416426e-01 3.39213103e-01 1.88329592e-01 -4.98375118e-01
6.43176734e-01 2.64602304e-01 -1.53958023e+00 1.10679321e-01
1.01302755e+00 9.41760600e-01 2.20760718e-01 8.24776292e-01
-4.99125779e-01 8.51619065e-01 -8.48916054e-01 -9.36195612e-01
-8.04490373e-02 -9.80543345e-02 -1.16475749e+00 -3.88055384e-01
7.76825845e-01 -1.87289119e-01 -3.05134922e-01 1.02492344e+00
2.65452534e-01 -6.59243912e-02 9.20111537e-01 -6.96388781e-01
-1.52616131e+00 7.50791907e-01 1.68672144e-01 6.59143925e-01
5.94794095e-01 -2.39811689e-01 1.36546707e+00 -1.12319708e+00
4.08524007e-01 1.08379781e+00 7.23201454e-01 4.81751025e-01
-7.48178959e-01 -2.09493697e-01 -1.56089023e-01 4.67075139e-01
-1.02100730e+00 -1.69659592e-02 2.24176869e-01 -5.93009174e-01
1.45297408e+00 4.61390048e-01 7.38493204e-01 9.07561719e-01
3.86784613e-01 2.68419027e-01 9.13341999e-01 -9.17118907e-01
-1.19437449e-01 3.87688965e-01 9.45822954e-01 6.15273118e-01
8.88219893e-01 -2.08292469e-01 -3.93537968e-01 4.65689898e-01
3.97849172e-01 -1.88023925e-01 -3.08080256e-01 1.33810222e-01
-9.44040477e-01 1.09996879e+00 1.95739001e-01 6.75856888e-01
9.09422413e-02 -1.66841999e-01 6.44691110e-01 6.31366074e-01
5.07367551e-01 1.10786080e+00 -3.10767055e-01 -2.94596136e-01
-6.65515959e-01 4.49169636e-01 1.03393388e+00 9.86549318e-01
1.60199702e-01 -2.81898111e-01 -2.21193373e-01 6.88230217e-01
2.45309979e-01 4.07653421e-01 8.84471238e-01 -6.49937570e-01
6.92273378e-01 8.78197134e-01 2.21554935e-02 -8.77333879e-01
-4.71252978e-01 -5.85855424e-01 -4.34620827e-01 2.57080585e-01
5.96993208e-01 -2.55748369e-02 -5.17845333e-01 1.17021871e+00
-2.79286057e-01 -4.91412997e-01 1.40513197e-01 5.51631808e-01
1.41336346e+00 6.39648199e-01 -9.84761342e-02 8.79206583e-02
1.50921321e+00 -1.18786097e+00 -9.14585888e-01 -2.50476208e-02
6.48407757e-01 -9.38781738e-01 1.45591259e+00 2.83756465e-01
-1.22182202e+00 -8.04673672e-01 -1.39203227e+00 -5.57372153e-01
-1.12461591e+00 -6.16191886e-02 2.70322114e-01 8.67644370e-01
-9.25936460e-01 7.99415648e-01 -2.92500943e-01 -4.17646170e-01
1.35738244e-02 -8.98291618e-02 -1.63556412e-01 -5.21811247e-02
-1.34031427e+00 1.40844810e+00 5.43607891e-01 -3.12633097e-01
-3.96695882e-01 -8.63321066e-01 -9.21641052e-01 4.28324193e-01
-1.05092563e-01 -5.16625941e-01 1.42979360e+00 -6.63748384e-01
-1.26029181e+00 1.06363630e+00 4.22924161e-01 -4.48392421e-01
4.98569965e-01 -7.64556170e-01 -4.11849201e-01 6.37016594e-02
8.86459500e-02 1.54764831e-01 2.15890363e-01 -7.09989369e-01
-4.66653943e-01 1.17516682e-01 8.47524881e-01 3.05705190e-01
-7.09555507e-01 -6.33430779e-02 -2.26988763e-01 -6.83649540e-01
-3.90881687e-01 -3.83189052e-01 4.51058924e-01 -1.47355273e-01
-5.19877151e-02 -8.49620402e-01 3.29636037e-02 -9.45683956e-01
1.70576811e+00 -1.89355719e+00 -7.34996274e-02 -1.86993569e-01
5.36421120e-01 5.67324638e-01 -2.23176762e-01 7.47563422e-01
-1.97355419e-01 6.15522206e-01 1.79610610e-01 -1.47147804e-01
5.62849522e-01 5.24357930e-02 -7.89026096e-02 5.39772771e-02
2.77162966e-04 9.02176619e-01 -1.10649014e+00 -4.72737223e-01
4.69813555e-01 4.29080844e-01 -3.22725266e-01 4.93687600e-01
-6.05419166e-02 -4.89439309e-01 -2.02011496e-01 2.83012271e-01
4.86162513e-01 -1.54885516e-01 -1.70342043e-01 1.44653544e-01
-3.57852548e-01 7.62668312e-01 -8.87593925e-01 1.44309032e+00
-7.91290641e-01 8.92782271e-01 -9.27532315e-01 -7.12323666e-01
7.54499078e-01 2.95510083e-01 -9.48316827e-02 -9.42409039e-01
1.77262232e-01 2.90008903e-01 -3.46670710e-02 -8.16185236e-01
5.50408006e-01 5.55873290e-02 -3.95356901e-02 3.88654083e-01
1.68947309e-01 -5.01961350e-01 1.00585186e+00 1.12482816e-01
1.06992161e+00 -6.45927787e-02 5.05935073e-01 -4.59471792e-01
7.28280663e-01 -1.98255032e-01 -5.65773882e-02 8.08699429e-01
-9.36878920e-02 6.44876003e-01 3.86373669e-01 -5.32050431e-01
-1.18339634e+00 -1.45771670e+00 -4.77024555e-01 1.01874185e+00
-1.44814923e-01 -9.83423352e-01 -9.41868126e-01 -3.86917323e-01
-2.27403939e-02 1.06395924e+00 -6.41831994e-01 2.66487565e-04
-2.91494668e-01 -3.73997211e-01 2.46801466e-01 4.21359271e-01
5.37175059e-01 -9.86294627e-01 -1.02024639e+00 2.26774409e-01
4.21451554e-02 -8.85009527e-01 -3.95743966e-01 1.39697269e-01
-5.47309399e-01 -1.30379665e+00 -8.66773069e-01 -9.37244117e-01
2.59927720e-01 -8.55688304e-02 1.77371287e+00 4.15211469e-01
-2.01032460e-01 4.24630672e-01 -8.44934225e-01 -7.93135643e-01
-5.01580715e-01 2.47200117e-01 -1.40806839e-01 -8.96030843e-01
7.38685071e-01 -3.73858474e-02 -1.92200020e-01 -3.42044801e-01
-1.08443046e+00 -1.38239279e-01 5.65800905e-01 8.32851529e-01
9.13583636e-02 4.04741727e-02 3.66365105e-01 -9.33182001e-01
1.25119269e+00 -2.87149727e-01 -4.19259399e-01 7.00278461e-01
-1.06652486e+00 2.61037499e-01 7.78448522e-01 -2.90853530e-01
-6.37707531e-01 -9.22452569e-01 -2.54617631e-01 2.31162697e-01
1.90964080e-02 7.95076907e-01 2.20876753e-01 -3.39783989e-02
7.29888856e-01 3.47102992e-02 -3.82596850e-01 -6.33779109e-01
1.32713333e-01 6.68789148e-01 1.75442487e-01 -2.94161975e-01
8.58283997e-01 -6.49435461e-01 -2.25221306e-01 -7.46583581e-01
-1.34560907e+00 -4.20143425e-01 -7.13076949e-01 -2.55912244e-01
7.76143551e-01 -4.87979859e-01 -4.02553260e-01 3.38741302e-01
-1.13428450e+00 4.27256152e-02 -3.03751856e-01 5.47298908e-01
-1.60637885e-01 5.27438700e-01 -7.71206558e-01 -5.60923278e-01
-3.61684769e-01 -6.76694930e-01 4.50802624e-01 5.61649919e-01
-4.42759722e-01 -1.41979790e+00 3.90688658e-01 2.17996001e-01
6.93526268e-01 3.15439165e-01 1.24271238e+00 -7.89529443e-01
3.14269438e-02 -2.62060374e-01 -2.06130460e-01 7.79757798e-01
7.68112391e-02 1.57274142e-01 -8.51418316e-01 -5.76819889e-02
5.14407875e-03 -5.38095057e-01 8.93937826e-01 8.22890550e-02
9.36543405e-01 -5.75577915e-01 3.54394227e-01 3.21397930e-01
1.75398922e+00 9.66601297e-02 6.50914490e-01 9.68283474e-01
5.15173316e-01 5.73724568e-01 3.99136335e-01 4.20911074e-01
4.59799200e-01 1.96958438e-01 2.26355016e-01 1.19220681e-01
-1.48002282e-01 1.19606983e-02 4.58827585e-01 1.47147417e+00
-4.70219366e-02 -4.00623798e-01 -1.16548836e+00 5.25591373e-01
-1.08800101e+00 -1.02574170e+00 -1.91233486e-01 2.05036521e+00
1.25358391e+00 5.60483992e-01 -4.47437614e-02 4.78765905e-01
2.67431587e-01 3.00249934e-01 -4.47388627e-02 -1.17305732e+00
-2.98841540e-02 7.81288743e-01 3.56523842e-01 6.86844230e-01
-1.00318527e+00 5.12263656e-01 6.38732052e+00 5.09077430e-01
-5.88379681e-01 -1.04568765e-01 2.96997517e-01 2.09569097e-01
-5.23885727e-01 -3.63578945e-01 -7.67945766e-01 4.97920156e-01
1.28755975e+00 -6.51870489e-01 1.26278892e-01 5.54274082e-01
-1.65343568e-01 1.58203505e-02 -1.31700993e+00 9.51808333e-01
3.19498211e-01 -1.18463647e+00 3.54087532e-01 -4.28901553e-01
5.48948109e-01 -3.11463237e-01 2.75324643e-01 6.81788027e-01
3.02306265e-02 -1.59559751e+00 7.82844722e-01 6.78454340e-01
7.15785205e-01 -7.86914170e-01 1.30516720e+00 1.03674240e-01
-9.08527613e-01 1.83771953e-01 -6.74836338e-01 -5.53949356e-01
-2.56950110e-01 5.16789138e-01 -4.84585643e-01 4.95539337e-01
5.60128152e-01 5.98482847e-01 -1.13514304e+00 9.51426029e-01
-3.91968489e-01 3.57123971e-01 -8.26995540e-03 -7.38144636e-01
2.96424329e-01 -2.90317591e-02 -3.67659144e-02 1.53744090e+00
2.55357504e-01 2.82920152e-02 -3.87144595e-01 7.46565044e-01
-7.74584934e-02 5.04583538e-01 -3.58733505e-01 -3.25120091e-01
3.64374101e-01 9.54898238e-01 -2.24848121e-01 -3.35419118e-01
-8.09122920e-01 6.80449724e-01 5.47713757e-01 -4.82010469e-02
-8.02177489e-01 -1.03286231e+00 2.63583034e-01 -1.24776460e-01
5.21188006e-02 -1.91133738e-01 -3.76906991e-01 -1.25069034e+00
2.97699869e-01 -8.93777072e-01 3.55318159e-01 -6.77583873e-01
-1.30697453e+00 6.60759866e-01 9.26090553e-02 -9.92696464e-01
-8.19929466e-02 -1.29463601e+00 -7.05820441e-01 1.01058114e+00
-1.73218846e+00 -6.53299510e-01 -4.83549893e-01 -2.52777785e-02
5.89696586e-01 -2.12530896e-01 1.11491120e+00 2.57033378e-01
-5.30823052e-01 8.32421601e-01 4.33907419e-01 2.91786015e-01
5.16887248e-01 -1.93820143e+00 4.50343430e-01 7.81164587e-01
1.99625984e-01 7.92561352e-01 9.40932274e-01 -2.16747299e-01
-8.10216248e-01 -6.65257752e-01 1.67649353e+00 -6.01548314e-01
8.31534803e-01 -1.59117863e-01 -9.36669648e-01 3.32953781e-01
8.14448893e-01 -7.68529236e-01 9.46576893e-01 1.31272838e-01
-6.08260393e-01 -1.06628224e-01 -8.24294627e-01 5.14916182e-01
6.94646299e-01 -5.20114183e-01 -1.58729291e+00 5.94100177e-01
7.41158485e-01 -3.62284571e-01 -1.36330700e+00 5.51207960e-02
5.91778398e-01 -1.19291055e+00 7.81645060e-01 -5.38590252e-01
8.86245966e-01 2.06894532e-01 1.54956743e-01 -1.31316936e+00
-3.65108758e-01 -1.36059001e-01 -4.98934686e-02 1.24033260e+00
5.69956899e-01 -4.82264102e-01 2.22758681e-01 5.74203551e-01
-1.14878684e-01 -8.50921810e-01 -7.69762874e-01 -9.47889566e-01
8.85374904e-01 -1.43964738e-01 4.94679362e-01 8.75210106e-01
3.28096718e-01 4.69859779e-01 4.87381697e-01 -3.44551355e-01
3.25921357e-01 -1.32900789e-01 3.71579587e-01 -1.23381555e+00
-1.72855645e-01 -8.23619187e-01 -5.35895228e-01 -8.33585203e-01
1.42343014e-01 -8.15116346e-01 -5.09562671e-01 -1.99058652e+00
1.75716892e-01 1.53863192e-01 -5.33580542e-01 1.89855378e-02
-2.76318103e-01 -1.47469327e-01 2.56100118e-01 -6.65414110e-02
-4.46305841e-01 2.35427722e-01 1.20199239e+00 -3.11484814e-01
4.19980943e-01 -3.69032145e-01 -5.97736120e-01 6.18356109e-01
9.78936970e-01 -1.75688833e-01 -4.54642713e-01 -8.34457219e-01
4.65063781e-01 -5.04805624e-01 4.60564606e-02 -1.33751070e+00
-1.29182220e-01 -1.72811240e-01 2.98161924e-01 -3.31579506e-01
-6.59223944e-02 -5.88579595e-01 -4.50352013e-01 6.53341115e-01
-7.12621629e-01 8.74015033e-01 1.50499837e-02 -6.37882203e-02
-4.49271291e-01 -1.12575233e+00 7.50976384e-01 -3.40496659e-01
-5.48215210e-01 -1.21835098e-01 -1.99875444e-01 6.54423535e-01
7.11903512e-01 -5.13286814e-02 -5.24699092e-01 -1.88923627e-01
-2.98613578e-01 8.14880505e-02 3.98964554e-01 4.88219112e-01
5.49211621e-01 -1.14424467e+00 -8.47523749e-01 -1.84866235e-01
3.40048403e-01 -1.38695478e-01 -3.73006791e-01 4.18340474e-01
-1.47120547e+00 9.51373756e-01 -3.14561218e-01 3.93910483e-02
-1.20878482e+00 4.95930910e-01 4.90355998e-01 -4.82798338e-01
-5.42547762e-01 1.05210102e+00 -3.62542160e-02 -7.83260316e-02
5.03400505e-01 -1.11962819e+00 -9.42053318e-01 2.28798941e-01
1.01881337e+00 5.74051142e-01 8.68216455e-02 -3.75491351e-01
-8.95221829e-02 8.55493844e-01 -1.41114980e-01 4.10302766e-02
1.57495272e+00 -8.53910372e-02 -4.53107655e-02 8.46062064e-01
1.48912156e+00 8.23330134e-02 -4.47048545e-01 -6.25284463e-02
5.58690190e-01 -3.46482277e-01 -2.35789791e-01 -8.97220910e-01
-5.35040855e-01 1.00295591e+00 6.18469119e-01 7.44953752e-01
9.62835968e-01 -4.01737869e-01 7.09893048e-01 6.70882642e-01
3.10520479e-03 -1.39773571e+00 9.72290486e-02 1.07389629e+00
1.08239305e+00 -1.11605239e+00 3.63663211e-02 -1.38900936e-01
-2.60503888e-01 1.96396816e+00 4.47403044e-01 -2.77068257e-01
5.03970265e-01 -6.43419325e-02 2.08163649e-01 -2.56789297e-01
-5.85364342e-01 -6.83846399e-02 6.75839484e-01 4.77862358e-01
1.03684819e+00 -6.90011680e-02 -1.01838028e+00 6.41431928e-01
-6.49800897e-01 -2.36246794e-01 8.98052514e-01 5.77565968e-01
-6.70513451e-01 -9.71128702e-01 -8.97245109e-02 5.89354575e-01
-8.40120673e-01 -5.59948862e-01 -5.03493369e-01 1.16838288e+00
1.69806272e-01 1.06156743e+00 -5.09895235e-02 -1.12004027e-01
5.73582470e-01 3.63513917e-01 5.84922791e-01 -9.86833096e-01
-8.55162144e-01 -7.16893077e-01 2.66962767e-01 -1.48134515e-01
-4.00835901e-01 -4.83441412e-01 -9.36024129e-01 -6.66423917e-01
-1.90328002e-01 4.58572477e-01 2.65569925e-01 9.83391047e-01
-2.93555409e-01 6.31108403e-01 3.90828103e-01 -2.91628510e-01
-1.02113080e+00 -1.28529608e+00 -7.37712622e-01 8.51212621e-01
2.60736108e-01 -6.95824623e-01 -7.72290826e-01 1.91664100e-02] | [10.995467185974121, 10.08178424835205] |
35d9cb5a-4859-40c0-b9cb-21b8275f02b3 | gapx-generalized-autoregressive-paraphrase | 2210.01979 | null | https://arxiv.org/abs/2210.01979v1 | https://arxiv.org/pdf/2210.01979v1.pdf | GAPX: Generalized Autoregressive Paraphrase-Identification X | Paraphrase Identification is a fundamental task in Natural Language Processing. While much progress has been made in the field, the performance of many state-of-the-art models often suffer from distribution shift during inference time. We verify that a major source of this performance drop comes from biases introduced by negative examples. To overcome these biases, we propose in this paper to train two separate models, one that only utilizes the positive pairs and the other the negative pairs. This enables us the option of deciding how much to utilize the negative model, for which we introduce a perplexity based out-of-distribution metric that we show can effectively and automatically determine how much weight it should be given during inference. We support our findings with strong empirical results. | ['Ser-Nam Lim', 'Hayden Housen', 'Renyu Li', 'Yifei Zhou'] | 2022-10-05 | null | null | null | null | ['paraphrase-identification'] | ['natural-language-processing'] | [ 2.31646359e-01 4.11106944e-02 -5.31567633e-01 -5.29960573e-01
-7.35394597e-01 -7.63273180e-01 8.22452486e-01 3.85204345e-01
-7.26817548e-01 7.24376738e-01 2.28276089e-01 -6.07181191e-01
7.60710537e-02 -6.48803234e-01 -5.38524985e-01 -5.09206593e-01
3.41431737e-01 7.40315139e-01 2.50788510e-01 -1.60815254e-01
5.45050502e-01 3.55771035e-01 -1.32600284e+00 1.72155827e-01
9.12361324e-01 4.94177014e-01 -3.31218876e-02 5.31649232e-01
-2.40911692e-01 7.55457878e-01 -5.77740312e-01 -7.65821159e-01
2.44866580e-01 -4.89664644e-01 -8.31054509e-01 -1.62190869e-01
3.25104266e-01 -2.95801222e-01 -1.79916680e-01 1.21796679e+00
4.04746443e-01 5.99925928e-02 9.17151630e-01 -1.23557210e+00
-3.84025484e-01 6.65425777e-01 -6.64928794e-01 6.28121734e-01
3.91610652e-01 1.62040025e-01 1.24525368e+00 -8.72736454e-01
6.73401833e-01 1.30770636e+00 4.45415586e-01 3.54219496e-01
-1.30780125e+00 -5.13418138e-01 3.25453803e-02 2.27820083e-01
-1.20796943e+00 -4.76282686e-01 7.60744393e-01 -4.60650414e-01
9.09539640e-01 1.81733161e-01 5.89095950e-01 9.95112360e-01
1.74867466e-01 8.49795520e-01 1.22727048e+00 -7.74478316e-01
1.76035449e-01 2.27125987e-01 5.45304060e-01 5.55754781e-01
5.57412326e-01 -5.97621277e-02 -4.40185636e-01 -2.18020409e-01
4.58046019e-01 -2.13257134e-01 -1.63126945e-01 -2.66737014e-01
-9.49032009e-01 9.42482471e-01 -2.52744532e-03 4.08034295e-01
-2.48473719e-01 -1.14955902e-01 2.44368494e-01 3.88235748e-01
4.95165795e-01 5.96293211e-01 -3.45314085e-01 -6.33026958e-01
-1.17039263e+00 3.63012880e-01 9.72466588e-01 6.53098583e-01
6.11198485e-01 -5.31132758e-01 -2.57738769e-01 1.16704798e+00
8.05216655e-02 2.85469294e-01 6.51368141e-01 -7.72810698e-01
6.15142107e-01 5.52925050e-01 2.08565995e-01 -9.47852135e-01
-2.13074386e-01 -4.30040985e-01 -5.41524947e-01 -1.46478683e-01
8.00861120e-01 -1.12445757e-01 -7.89654195e-01 1.91485476e+00
-1.84717961e-02 -4.36341345e-01 -2.73460329e-01 6.67979181e-01
-2.08244305e-02 3.69879186e-01 7.44218230e-02 -3.06659460e-01
1.30198634e+00 -7.92251468e-01 -7.01118648e-01 -6.35132730e-01
6.91499531e-01 -9.61569667e-01 1.15284836e+00 4.50056970e-01
-1.21573532e+00 -2.11461872e-01 -1.04007912e+00 -3.42967734e-02
-4.84826416e-02 -9.89629235e-03 6.85698628e-01 5.93858659e-01
-7.61023104e-01 7.09681749e-01 -8.13146591e-01 -5.41036129e-01
2.29083940e-01 1.61116108e-01 -1.98857769e-01 -1.34799719e-01
-1.21636403e+00 1.36753082e+00 2.17325956e-01 -8.08379054e-02
-2.67847061e-01 -3.24786723e-01 -6.18071318e-01 2.36387193e-01
4.46274400e-01 -7.39680588e-01 1.42639744e+00 -1.15565717e+00
-1.23196125e+00 1.00069296e+00 -6.45144284e-01 -4.40721005e-01
8.04561079e-01 -2.61409521e-01 -8.81353095e-02 9.48607102e-02
1.62881523e-01 3.47658008e-01 8.33396554e-01 -1.07215583e+00
-5.08098602e-01 -4.43930417e-01 3.01472154e-02 9.11883190e-02
-4.28008080e-01 1.80950999e-01 -5.96124530e-01 -7.30652452e-01
2.27232516e-01 -1.13559544e+00 2.24779807e-02 -6.57163784e-02
-3.91771287e-01 -4.35771108e-01 4.12511565e-02 -6.05047643e-01
1.65453434e+00 -1.75057960e+00 7.05627128e-02 2.89648741e-01
7.12129921e-02 2.68141657e-01 2.64600967e-03 5.70208967e-01
-5.19851036e-02 1.80749685e-01 -2.59561718e-01 -2.32957780e-01
1.22757889e-01 2.11190715e-01 -4.41940069e-01 2.74134755e-01
3.92326802e-01 6.57276511e-01 -1.14929986e+00 -4.24399525e-01
-6.91106990e-02 1.01577796e-01 -6.47059739e-01 1.94205791e-01
2.70581506e-02 -2.04947308e-01 -2.70548940e-01 2.78920770e-01
7.60341585e-01 -3.91232252e-01 4.65415716e-01 -2.26569176e-02
1.45779610e-01 8.11280787e-01 -9.32213783e-01 1.30001199e+00
-3.24822247e-01 7.19179332e-01 -2.20486239e-01 -1.05983651e+00
5.78992248e-01 -5.40280007e-02 -1.74756814e-02 -5.93712270e-01
1.57673344e-01 1.81549981e-01 4.96105283e-01 -6.39888108e-01
6.09142959e-01 -5.03633976e-01 1.54262915e-01 7.11794972e-01
-6.81235939e-02 -9.54023972e-02 4.44191039e-01 3.65951478e-01
1.04183400e+00 -5.56380525e-02 4.37083125e-01 -2.70802051e-01
3.41936260e-01 -1.24774329e-01 4.51233923e-01 9.43266094e-01
-2.95999110e-01 3.57660621e-01 9.38312232e-01 4.58990969e-02
-1.19526517e+00 -1.09222674e+00 -5.59747703e-02 9.94523823e-01
-1.28216073e-01 -4.19375628e-01 -8.01497400e-01 -7.51930773e-01
1.51767895e-01 1.03946352e+00 -5.67596555e-01 -3.12728107e-01
-5.60895026e-01 -8.08032155e-01 6.10012412e-01 6.46299660e-01
1.17679290e-01 -6.74422801e-01 -3.77174854e-01 1.21124066e-01
-4.00733590e-01 -8.96106660e-01 -3.46952647e-01 6.56539872e-02
-9.11921322e-01 -1.09055364e+00 -5.61327398e-01 -4.26697791e-01
6.23542130e-01 4.16298985e-01 1.40136731e+00 2.15247869e-02
2.11992916e-02 1.88363910e-01 -2.31461987e-01 -4.62133944e-01
-6.45877242e-01 2.57758498e-01 1.91662814e-02 -3.73246193e-01
9.44252253e-01 -5.01974761e-01 -4.32716072e-01 1.48864761e-01
-7.69977450e-01 6.94596320e-02 7.43981361e-01 9.86773670e-01
1.85015887e-01 -2.74578512e-01 5.82591355e-01 -1.02290380e+00
9.56395626e-01 -5.17792165e-01 -3.42821658e-01 4.27088946e-01
-8.03837240e-01 3.79195362e-01 5.69378793e-01 -4.07414943e-01
-7.78270304e-01 -3.60542983e-01 -2.97424734e-01 -2.04409838e-01
-2.88259499e-02 6.40801847e-01 7.20659345e-02 3.14470857e-01
5.17864883e-01 8.53970796e-02 8.55102390e-02 -3.12386274e-01
2.48732075e-01 8.10989738e-01 2.65540838e-01 -7.30460286e-01
7.08316803e-01 2.33208656e-01 -2.46488988e-01 -1.00298321e+00
-1.23347843e+00 -5.71633518e-01 -4.78777647e-01 -3.51976864e-02
5.52880824e-01 -7.46926546e-01 -3.64163250e-01 2.33940050e-01
-1.09819758e+00 -1.01674199e-01 2.27751378e-02 4.70188409e-01
-3.37547928e-01 6.66171014e-01 -6.08732283e-01 -1.02238095e+00
-2.56884456e-01 -8.16999912e-01 6.37433231e-01 8.55057761e-02
-7.72100866e-01 -1.00952184e+00 2.55700707e-01 3.83738816e-01
3.50833923e-01 -2.74123877e-01 1.11771822e+00 -1.00209606e+00
-2.16435820e-01 -3.04242373e-01 -2.02067629e-01 4.72253650e-01
2.03737877e-02 1.56132847e-01 -9.66940999e-01 -2.98537761e-01
2.20219210e-01 -2.14913934e-01 1.08088732e+00 1.31292716e-01
1.04276299e+00 -1.56493768e-01 -3.17062080e-01 1.74139768e-01
1.30528367e+00 -1.91973358e-01 5.71618795e-01 2.38564700e-01
4.07152385e-01 6.45678341e-01 5.61343491e-01 2.11570695e-01
3.33168179e-01 6.66354477e-01 -9.33741103e-04 2.43432000e-01
2.17498779e-01 -4.68049616e-01 3.14425141e-01 9.66531992e-01
2.06294030e-01 -4.64197576e-01 -9.78225708e-01 7.34099984e-01
-1.88481736e+00 -1.15896213e+00 9.74390730e-02 2.45872808e+00
1.05646074e+00 6.24402761e-01 1.65423185e-01 3.72820258e-01
5.61070263e-01 4.76985760e-02 -3.51642251e-01 -3.87947381e-01
-2.83744689e-02 2.50619650e-01 3.56776267e-01 7.53745377e-01
-7.52843797e-01 7.95028567e-01 6.73969793e+00 8.64788651e-01
-9.93982673e-01 -3.10345083e-01 5.48859954e-01 -1.29357085e-01
-4.41454262e-01 1.45467877e-01 -5.82439065e-01 6.12924457e-01
8.19093823e-01 -3.37283999e-01 3.66043419e-01 7.10059404e-01
1.84913352e-01 -4.25565958e-01 -1.39464343e+00 8.17986667e-01
2.12847322e-01 -7.74123609e-01 1.72046766e-01 -3.06189079e-02
3.23826224e-01 -6.28069788e-02 -2.24510700e-01 3.48812282e-01
2.43777856e-01 -8.15925002e-01 5.77212393e-01 3.38920832e-01
4.82219577e-01 -6.04641438e-01 6.89106405e-01 5.81393123e-01
-6.83359385e-01 -6.48935288e-02 -5.17741144e-01 -3.43297482e-01
2.55978536e-02 8.75344276e-01 -9.52033222e-01 2.24060059e-01
1.68300018e-01 3.81322354e-01 -5.56587219e-01 9.85452890e-01
-6.30676389e-01 1.07120717e+00 -3.72392923e-01 -2.81706989e-01
2.07384061e-02 -3.37210327e-01 5.22445261e-01 1.43098438e+00
2.25995317e-01 -1.07770845e-01 -1.01706766e-01 8.51361215e-01
-1.62002236e-01 1.71242237e-01 -5.44344425e-01 -2.57671773e-01
6.62952065e-01 1.09527576e+00 -5.92515647e-01 -6.70624256e-01
-4.38702315e-01 9.39752758e-01 7.25862086e-01 2.03735128e-01
-6.52305961e-01 -4.62128252e-01 5.87145627e-01 8.48975554e-02
2.40947142e-01 -1.08391263e-01 -4.11778927e-01 -1.45975220e+00
1.78456143e-01 -7.57317543e-01 3.69784623e-01 -5.29003143e-01
-1.69936168e+00 1.35000661e-01 1.58233300e-01 -9.26114142e-01
-6.04121149e-01 -5.78031540e-01 -5.61566889e-01 1.04424441e+00
-1.49967706e+00 -4.82728153e-01 1.32886367e-02 5.92169724e-02
5.18703043e-01 2.09947571e-01 5.80989540e-01 3.36062759e-01
-6.37770772e-01 7.43824005e-01 1.00306906e-01 2.77630240e-01
9.72063482e-01 -1.28521347e+00 4.03927118e-01 8.66744161e-01
3.20562907e-02 1.09330916e+00 1.07652414e+00 -3.85576218e-01
-1.12710571e+00 -4.59163636e-01 1.48610663e+00 -5.80321550e-01
8.63566101e-01 -2.98909932e-01 -1.02457893e+00 6.16136968e-01
-8.23087711e-03 -4.64450657e-01 7.75262773e-01 3.42008084e-01
-5.89996338e-01 8.79772678e-02 -9.77569401e-01 7.29643524e-01
8.43819022e-01 -6.01066053e-01 -1.12433672e+00 1.62479490e-01
3.67835850e-01 -6.28832728e-02 -4.99669403e-01 1.23467147e-01
7.92156219e-01 -1.20244014e+00 6.95162714e-01 -7.51406550e-01
7.35885143e-01 1.10555468e-02 5.39554060e-02 -1.41648388e+00
-2.21071362e-01 -2.45421484e-01 -4.30127122e-02 1.22352636e+00
5.49052536e-01 -7.18599856e-01 7.34570086e-01 8.40261102e-01
2.73094624e-01 -7.63294935e-01 -7.43557632e-01 -8.15896332e-01
4.93441969e-01 -3.66688102e-01 2.18537167e-01 1.03990912e+00
3.71044636e-01 8.02092969e-01 -9.49962065e-02 -2.95918047e-01
4.96415913e-01 2.16491930e-02 6.97944462e-01 -1.25477099e+00
-4.83938247e-01 -6.32115126e-01 -1.16554625e-01 -1.38237536e+00
1.68459654e-01 -7.32314467e-01 -2.50381399e-02 -1.45171475e+00
6.49428904e-01 -1.71658397e-01 -2.51343369e-01 3.47754151e-01
-5.79653919e-01 3.03451624e-02 2.06483826e-01 1.88399613e-01
-4.04579610e-01 3.96791995e-01 8.35982978e-01 1.18385836e-01
4.44920473e-02 -5.79725020e-02 -7.76402652e-01 8.33103538e-01
8.12247097e-01 -5.38499594e-01 -3.89113158e-01 -4.29793298e-01
4.92485702e-01 -1.46655887e-01 2.06898898e-01 -7.12357223e-01
-2.75066234e-02 -1.12282731e-01 3.08963686e-01 -3.38592291e-01
1.73604071e-01 -5.92157304e-01 -3.98005396e-01 5.30305445e-01
-5.27299345e-01 3.13966185e-01 1.10649616e-01 4.64136392e-01
-1.05001800e-01 -4.81150746e-01 8.04757416e-01 -5.05360402e-02
-9.32543278e-02 -1.56269610e-01 -3.97286236e-01 4.80697721e-01
6.09122396e-01 3.04357354e-02 -3.91148865e-01 -5.32692432e-01
-3.38791490e-01 6.82940334e-02 6.97150707e-01 3.46586227e-01
2.80381501e-01 -1.12626362e+00 -6.11675560e-01 1.20491005e-01
2.00500593e-01 -4.99953806e-01 -9.48769674e-02 9.62046266e-01
-3.34887892e-01 4.87459689e-01 6.63577989e-02 -3.38619649e-01
-1.19517744e+00 4.66446728e-01 1.47539377e-01 -5.25229096e-01
-1.36599377e-01 6.56617880e-01 -4.03886512e-02 -1.55303344e-01
1.03913680e-01 -2.11882412e-01 8.47986201e-04 1.91823468e-01
5.03460824e-01 4.11600679e-01 1.54222995e-01 -4.26236391e-01
-2.65407592e-01 2.50062048e-01 -3.37627828e-01 -4.61082816e-01
1.06533682e+00 -2.17898246e-02 -2.91247845e-01 6.65002704e-01
9.87144411e-01 2.01449901e-01 -7.98881471e-01 -2.09297210e-01
1.83274522e-01 -6.49876297e-01 -2.20302910e-01 -7.83192694e-01
-4.63216871e-01 1.14439833e+00 1.57616034e-01 3.45193833e-01
7.23564327e-01 -3.29058409e-01 6.54527187e-01 5.19718349e-01
2.05988869e-01 -1.23319030e+00 -1.55893385e-01 6.39158607e-01
5.18995941e-01 -1.29402697e+00 1.57006741e-01 -3.12548012e-01
-6.16423190e-01 8.98560762e-01 3.50570381e-01 -1.93943173e-01
4.63274866e-01 1.95660368e-02 1.69934612e-02 1.31572820e-02
-9.49090838e-01 -1.57256443e-02 2.10809037e-01 2.62507886e-01
8.81037414e-01 -1.11104824e-01 -8.76664698e-01 4.15084183e-01
-3.85247678e-01 -2.44908500e-02 4.39718902e-01 9.01335239e-01
-4.53958988e-01 -1.37529230e+00 -1.59370303e-01 8.05660188e-01
-6.50197983e-01 -3.07046890e-01 -6.92995071e-01 7.99527109e-01
-4.02531534e-01 1.04969740e+00 1.41789392e-01 -2.40384430e-01
2.86072016e-01 4.06584680e-01 8.47842395e-01 -5.15216351e-01
-5.47094345e-01 -2.49078706e-01 2.75742680e-01 -2.01497674e-01
-2.53197283e-01 -5.79189122e-01 -8.73080015e-01 -4.74761754e-01
-5.01551569e-01 2.87238985e-01 4.83494073e-01 1.05109453e+00
2.52821743e-01 1.15938477e-01 3.35980088e-01 -5.93482316e-01
-1.28809524e+00 -1.23921192e+00 -3.21643591e-01 6.93880677e-01
2.53455579e-01 -6.18866444e-01 -6.99838221e-01 -2.69684702e-01] | [11.065670013427734, 9.06217098236084] |
2df2616a-659d-4922-a4fc-16f2b18949b1 | adaptdhm-adaptive-distribution-hierarchical | 2211.12105 | null | https://arxiv.org/abs/2211.12105v1 | https://arxiv.org/pdf/2211.12105v1.pdf | AdaptDHM: Adaptive Distribution Hierarchical Model for Multi-Domain CTR Prediction | Large-scale commercial platforms usually involve numerous business domains for diverse business strategies and expect their recommendation systems to provide click-through rate (CTR) predictions for multiple domains simultaneously. Existing promising and widely-used multi-domain models discover domain relationships by explicitly constructing domain-specific networks, but the computation and memory boost significantly with the increase of domains. To reduce computational complexity, manually grouping domains with particular business strategies is common in industrial applications. However, this pre-defined data partitioning way heavily relies on prior knowledge, and it may neglect the underlying data distribution of each domain, hence limiting the model's representation capability. Regarding the above issues, we propose an elegant and flexible multi-distribution modeling paradigm, named Adaptive Distribution Hierarchical Model (AdaptDHM), which is an end-to-end optimization hierarchical structure consisting of a clustering process and classification process. Specifically, we design a distribution adaptation module with a customized dynamic routing mechanism. Instead of introducing prior knowledge for pre-defined data allocation, this routing algorithm adaptively provides a distribution coefficient for each sample to determine which cluster it belongs to. Each cluster corresponds to a particular distribution so that the model can sufficiently capture the commonalities and distinctions between these distinct clusters. Extensive experiments on both public and large-scale Alibaba industrial datasets verify the effectiveness and efficiency of AdaptDHM: Our model achieves impressive prediction accuracy and its time cost during the training stage is more than 50% less than that of other models. | ['Haoyuan Hu', 'Lixia Wu', 'Minfang Lu', 'Yuanlin Liu', 'Huiwen Zheng', 'Jinyun Li'] | 2022-11-22 | null | null | null | null | ['click-through-rate-prediction'] | ['miscellaneous'] | [-2.73991436e-01 -4.36864525e-01 -8.82917821e-01 -4.33967084e-01
-2.88700193e-01 -5.60536027e-01 3.21786582e-01 3.14896666e-02
1.05295897e-01 5.38341761e-01 -1.99852780e-01 -4.24569786e-01
-5.92216969e-01 -1.04037857e+00 -4.54567492e-01 -6.51852727e-01
2.16213182e-01 9.56007540e-01 5.37328422e-01 1.46963531e-02
1.73267975e-01 2.88812965e-01 -1.38568509e+00 3.66399437e-01
1.15032256e+00 1.24805903e+00 4.66164410e-01 4.64870445e-02
-3.37702245e-01 5.02044559e-01 -5.91511190e-01 -2.62389779e-01
4.23540741e-01 -2.39380464e-01 -4.05512869e-01 3.24275494e-01
-1.70855254e-01 -2.85902619e-01 -2.02453375e-01 9.52323258e-01
1.70658946e-01 7.04888999e-02 6.59867585e-01 -1.47002208e+00
-5.07949293e-01 7.24777758e-01 -8.69628727e-01 8.28276128e-02
-1.49815291e-01 1.19407080e-01 1.19620228e+00 -4.27286476e-01
4.53147262e-01 8.76124561e-01 3.88430506e-01 2.11274251e-01
-1.37706721e+00 -9.79142725e-01 6.24098063e-01 1.75773531e-01
-1.44811070e+00 1.35995597e-01 1.01861906e+00 -5.23305416e-01
3.61728549e-01 5.45035526e-02 4.49379116e-01 9.56037045e-01
-3.49585600e-02 6.61306679e-01 8.21565390e-01 1.72439113e-01
4.88722652e-01 5.41645408e-01 2.70184726e-01 8.16687271e-02
3.90771538e-01 -2.00596347e-01 -2.88323224e-01 -4.22002286e-01
9.60995615e-01 4.34584826e-01 2.21455038e-01 -8.82014155e-01
-1.01200306e+00 9.52711642e-01 1.71431899e-01 1.84684351e-01
-5.07830679e-01 -4.13502842e-01 2.32343853e-01 3.69986802e-01
2.89199084e-01 1.86658934e-01 -7.56083786e-01 4.13459875e-02
-8.16754758e-01 3.51597458e-01 9.27181482e-01 1.39199471e+00
1.01809430e+00 -2.73652554e-01 1.28110468e-01 1.09586179e+00
2.98639446e-01 9.24122706e-02 4.74952757e-01 -8.13809335e-01
6.07503891e-01 1.03561449e+00 8.14323425e-02 -1.20203972e+00
-2.52577066e-01 -6.01700127e-01 -1.00718045e+00 -2.42096603e-01
5.88653028e-01 -1.65354669e-01 -6.21157169e-01 1.57974327e+00
5.67518115e-01 -1.08285453e-02 -2.57566243e-01 8.91617239e-01
2.53015041e-01 5.43980837e-01 2.22356588e-01 -2.20184147e-01
1.18678498e+00 -7.89068341e-01 -4.81426656e-01 -9.36418772e-02
2.66002357e-01 -5.52818835e-01 1.11863852e+00 6.59642816e-01
-7.27265894e-01 -5.61444879e-01 -1.07091486e+00 4.37025428e-01
-3.02452415e-01 1.38537407e-01 8.41918290e-01 5.97806931e-01
-2.16919631e-01 4.84542221e-01 -4.82005715e-01 -1.71940118e-01
4.04595613e-01 5.46590090e-01 1.34449959e-01 -1.87437654e-01
-1.17651021e+00 2.73836374e-01 4.64699775e-01 -4.55734044e-01
-7.49270201e-01 -8.38696718e-01 -2.77789712e-01 1.60014525e-01
6.11873269e-01 -6.14417255e-01 1.05019605e+00 -9.45824623e-01
-1.52677095e+00 2.66007811e-01 -4.17597182e-02 -3.85101736e-01
4.54147190e-01 1.01258442e-01 -7.57809520e-01 -9.24652964e-02
3.36424895e-02 1.06458530e-01 8.42844307e-01 -1.34512806e+00
-1.13174832e+00 -4.61854219e-01 8.96686241e-02 1.09702162e-01
-6.90099299e-01 -1.84622362e-01 -7.68645585e-01 -5.44593513e-01
2.08176807e-01 -7.55366147e-01 -5.49373984e-01 -3.62155974e-01
-2.88053691e-01 -3.47361982e-01 8.79929662e-01 -3.86124074e-01
1.75945795e+00 -2.19881606e+00 -1.03914462e-01 7.57749975e-01
4.44000781e-01 -9.03185084e-02 1.90705154e-02 2.78627723e-01
6.86093792e-02 9.81656313e-02 1.13691084e-01 2.22316429e-01
2.52820522e-01 1.69332266e-01 -1.86605588e-01 2.34185234e-01
-1.04030572e-01 2.77225465e-01 -5.65365493e-01 -4.34397459e-01
2.67693490e-01 6.11521304e-02 -8.67771327e-01 1.39984444e-01
-5.26514411e-01 2.42014483e-01 -8.17984402e-01 7.96158612e-01
1.00243151e+00 -8.71833146e-01 9.68495011e-01 -2.75265008e-01
1.88448980e-01 2.44366527e-01 -1.51504219e+00 1.38418996e+00
-3.52816761e-01 -2.09842920e-01 2.86948293e-01 -1.25454199e+00
1.11302543e+00 -3.68482694e-02 1.04653072e+00 -6.41161680e-01
1.50422184e-02 1.17875412e-01 1.20194398e-01 8.13596398e-02
3.99123102e-01 8.24938565e-02 -2.96068311e-01 2.51938820e-01
-1.96771458e-01 4.22339082e-01 3.28652889e-01 7.02795163e-02
1.02970421e+00 -2.50914574e-01 3.28760207e-01 -2.61195302e-01
4.01190519e-01 2.57737249e-01 9.64357555e-01 4.98044878e-01
-2.26205796e-01 3.03894281e-01 6.97140813e-01 -2.93253660e-01
-1.04602945e+00 -1.10739350e+00 -1.65372476e-01 1.36770904e+00
6.67206287e-01 -4.55226958e-01 -3.93566847e-01 -9.22941029e-01
4.46327001e-01 5.13977885e-01 -1.72895595e-01 4.95466925e-02
-3.49282265e-01 -9.48566616e-01 -3.42874750e-02 3.83461118e-01
5.09877026e-01 -4.55171973e-01 1.91944838e-01 4.10257369e-01
-5.71590327e-02 -1.04262030e+00 -3.46304327e-01 2.78638393e-01
-9.30955648e-01 -1.20726180e+00 -5.18158853e-01 -8.28165054e-01
5.13298035e-01 3.16451699e-01 1.09579158e+00 -3.71826798e-01
1.55739278e-01 -1.67782664e-01 -3.21819991e-01 -7.75056481e-02
-2.97695041e-01 5.35742223e-01 -3.57103422e-02 6.52906746e-02
8.41473639e-01 -1.00884867e+00 -7.61809170e-01 1.04453838e+00
-6.70002341e-01 -2.39674345e-01 8.79549921e-01 6.98456645e-01
6.28820121e-01 7.48820066e-01 9.00171876e-01 -1.17663527e+00
7.68602908e-01 -1.04508841e+00 -7.73478806e-01 2.04870909e-01
-9.94497716e-01 -3.31096709e-01 8.70051920e-01 -8.74005854e-01
-9.51390803e-01 2.20809355e-01 2.36517504e-01 -6.71340704e-01
-3.86685789e-01 5.33491671e-01 -5.66075683e-01 3.35691661e-01
4.16236609e-01 2.55315095e-01 1.50286213e-01 -6.84405029e-01
3.68008673e-01 9.46273685e-01 1.04976475e-01 -7.91536391e-01
8.76722336e-01 2.18450442e-01 -3.25699002e-01 -4.92985249e-01
-5.65590024e-01 -7.65058815e-01 -6.08889699e-01 -5.67075470e-03
4.57851619e-01 -1.03707385e+00 -8.73699546e-01 1.47309184e-01
-6.45995080e-01 -1.82277948e-01 -1.51898796e-02 3.91862541e-01
-3.20777148e-01 3.34134191e-01 -4.85930353e-01 -4.43578839e-01
1.52495116e-01 -1.00858295e+00 5.72969317e-01 1.08405448e-01
-1.43430799e-01 -9.10307586e-01 -3.51505876e-01 5.02374947e-01
4.49578583e-01 -7.21891448e-02 1.33508384e+00 -1.05115795e+00
-7.93212473e-01 -1.16753981e-01 -3.95894051e-01 4.08597618e-01
2.84671932e-01 -2.60847956e-01 -3.40260595e-01 -1.87995568e-01
-2.26728410e-01 8.50601345e-02 3.86087179e-01 2.59790391e-01
1.75449860e+00 -2.42127299e-01 -6.14516973e-01 3.24489892e-01
1.41866601e+00 5.12198031e-01 4.35781419e-01 3.76790792e-01
6.61442995e-01 5.44655621e-01 8.74599636e-01 8.72345209e-01
4.33645338e-01 6.97058201e-01 3.06591541e-01 -5.90768196e-02
2.70853043e-01 -3.06840777e-01 2.16179535e-01 7.53508031e-01
1.46168068e-01 -1.60127997e-01 -5.79299092e-01 5.03071010e-01
-1.91914809e+00 -8.62392664e-01 1.18308358e-01 2.24279284e+00
9.81983960e-01 3.80210459e-01 7.09540248e-01 -5.37128150e-02
8.25950861e-01 -5.51109910e-02 -1.04343629e+00 -4.57213670e-02
1.56744003e-01 -1.39520720e-01 6.56455338e-01 -4.90394495e-02
-1.13390493e+00 7.10144699e-01 5.84995031e+00 1.39174688e+00
-9.35615659e-01 -1.38887048e-01 6.46654189e-01 -9.47514400e-02
-2.30326638e-01 -1.65655226e-01 -1.11774468e+00 8.84795487e-01
9.66255903e-01 -4.16331083e-01 5.15378237e-01 1.46700585e+00
1.22997545e-01 2.32461303e-01 -9.54728544e-01 8.84011567e-01
-4.81451571e-01 -1.32627320e+00 1.83763981e-01 4.96206880e-01
5.63744009e-01 -7.37294331e-02 5.56132644e-02 4.97373909e-01
8.31359625e-01 -4.69264418e-01 3.25650394e-01 2.88287736e-02
5.26459634e-01 -1.04740834e+00 3.39637816e-01 4.60544735e-01
-1.30637825e+00 -4.95700985e-01 -5.00533402e-01 1.61716059e-01
-2.02928893e-02 9.14792538e-01 -7.42010951e-01 6.53611898e-01
6.69265747e-01 8.07940066e-01 -1.65472522e-01 8.91690910e-01
2.48907894e-01 6.87387884e-01 -2.91683346e-01 -6.21849555e-04
6.38185591e-02 -4.54891950e-01 2.37407580e-01 8.47951889e-01
3.29546511e-01 -1.21233694e-01 7.71599591e-01 9.26103592e-01
6.69044480e-02 1.42277598e-01 -3.05185705e-01 -1.37236014e-01
9.30553257e-01 1.20307624e+00 -5.43505669e-01 -3.14680815e-01
-6.48646951e-01 5.65742731e-01 2.60499329e-03 4.05813873e-01
-1.12458754e+00 -4.89624500e-01 1.06233943e+00 5.85584104e-01
6.11235023e-01 -1.92974553e-01 -6.20026112e-01 -9.05791104e-01
-7.85222352e-02 -1.08706379e+00 6.01041734e-01 -7.42399544e-02
-1.91017866e+00 2.11357787e-01 -4.30069417e-02 -1.69461012e+00
-1.30067453e-01 -6.49495661e-01 -1.55933648e-01 7.33622849e-01
-1.43649673e+00 -1.00581086e+00 -2.46355996e-01 7.78604150e-01
5.45316279e-01 -4.42720711e-01 4.66399044e-01 7.74243057e-01
-7.64266610e-01 6.92444921e-01 4.00961250e-01 9.06505994e-03
7.57969499e-01 -1.07384396e+00 -1.44366533e-01 3.45130086e-01
-4.28629458e-01 9.64053869e-01 4.35263723e-01 -5.91264009e-01
-1.23597336e+00 -1.28421283e+00 4.23258096e-01 -2.04056695e-01
9.32320178e-01 -4.34138685e-01 -8.49189222e-01 5.01910031e-01
3.41477781e-03 -3.20806503e-01 1.01254141e+00 5.91430902e-01
-3.41518193e-01 -6.43094182e-01 -1.12202871e+00 3.44427794e-01
9.67474520e-01 -7.06223547e-02 -2.91462898e-01 4.57138687e-01
7.10022211e-01 -1.11660615e-01 -1.38241351e+00 3.26909304e-01
4.20652509e-01 -9.24278140e-01 9.96069789e-01 -4.42481667e-01
5.43413460e-01 -4.28035319e-01 -2.51904726e-01 -1.04255259e+00
-6.78187966e-01 -5.08556783e-01 -2.93304801e-01 1.32050157e+00
5.11925995e-01 -7.66215086e-01 1.07494366e+00 6.73338771e-01
1.42141908e-01 -7.60418594e-01 -5.87151408e-01 -9.44572687e-01
1.46373529e-02 -1.80451334e-01 1.07583833e+00 1.19822788e+00
3.89855467e-02 4.27765101e-01 -3.27642262e-01 3.96047711e-01
5.95141232e-01 6.74109161e-01 1.08430910e+00 -1.62589943e+00
-5.72759390e-01 -5.41306794e-01 -1.18068978e-02 -1.58917022e+00
-8.55474174e-02 -6.53604150e-01 -4.20749307e-01 -1.23866880e+00
2.43824169e-01 -1.02593470e+00 -6.31764174e-01 2.93678075e-01
2.31399640e-01 -3.97897035e-01 -1.32602111e-01 5.19239247e-01
-8.93527091e-01 3.17747444e-01 1.27266908e+00 -5.99361584e-02
-4.98508602e-01 4.44834143e-01 -1.12496448e+00 6.14411354e-01
8.21168363e-01 -2.92890519e-01 -7.47467816e-01 -5.03614210e-02
-1.88851193e-01 1.66545119e-02 3.53864580e-02 -8.71997178e-01
3.14913571e-01 -6.19991779e-01 4.44258392e-01 -7.59743452e-01
9.71936807e-03 -1.11105728e+00 3.02917004e-01 7.53272250e-02
-1.90505460e-01 -1.60919249e-01 -2.91515086e-02 8.69427383e-01
-3.37937772e-01 2.66994148e-01 7.81413972e-01 -8.79782960e-02
-9.16142583e-01 5.62295735e-01 -2.73922443e-01 -1.92801148e-01
1.30158901e+00 -2.58445442e-01 -1.17387377e-01 -1.42937124e-01
-7.18826532e-01 4.20038640e-01 4.10989583e-01 4.79240626e-01
1.81270316e-01 -1.27410555e+00 -1.58250332e-01 3.21443349e-01
1.58102423e-01 1.31204367e-01 2.70923853e-01 5.87241590e-01
4.84571140e-03 3.16401064e-01 -3.46795395e-02 -6.23272240e-01
-1.02067435e+00 7.20226526e-01 1.25488415e-01 -6.24991119e-01
-3.24318886e-01 3.00877780e-01 4.18594778e-01 -5.83024025e-01
1.81070164e-01 -3.27457696e-01 -2.64854193e-01 9.33729261e-02
2.53027081e-01 4.66766208e-01 -1.30186155e-01 -6.60189241e-02
-2.97930777e-01 2.21132964e-01 -4.03409570e-01 3.95487547e-01
1.35574019e+00 -3.03519338e-01 1.84866264e-02 2.48399660e-01
8.52011800e-01 -1.30925760e-01 -1.28855062e+00 -3.78971964e-01
1.03549838e-01 -5.38914800e-01 -3.64470668e-02 -1.05099821e+00
-1.25282848e+00 5.38650393e-01 2.59268403e-01 4.35261846e-01
1.30910957e+00 -7.26311654e-03 9.27087665e-01 1.19993538e-01
5.39036870e-01 -1.39592922e+00 1.35517240e-01 3.13270062e-01
2.49225348e-01 -9.68081415e-01 -3.41331065e-02 -7.07979381e-01
-8.31646860e-01 8.00307155e-01 1.02686489e+00 -5.64182810e-02
9.74537671e-01 1.81845605e-01 -2.64488876e-01 -3.16094048e-02
-7.43306518e-01 6.97116330e-02 8.07281956e-02 6.42907679e-01
5.60099855e-02 2.40639523e-01 -3.70156080e-01 1.31666672e+00
2.19856873e-02 1.39997616e-01 1.37952995e-02 6.87042654e-01
-5.23662329e-01 -1.49075532e+00 3.79150198e-03 7.02621996e-01
-3.48922759e-01 2.17223734e-01 -1.44912899e-01 9.88159418e-01
1.89175308e-01 9.37327683e-01 7.74046853e-02 -8.02476645e-01
5.55044472e-01 -2.05148280e-01 -8.24490413e-02 -5.59850097e-01
-4.11931455e-01 4.78852987e-01 6.91487594e-03 -5.48810065e-01
-7.11063221e-02 -6.30098581e-01 -1.23985863e+00 -6.10929012e-01
-2.27088869e-01 2.09751472e-01 4.93977278e-01 5.51784754e-01
7.99470007e-01 6.33464336e-01 9.79304075e-01 -3.63719314e-01
-9.02910650e-01 -7.88176954e-01 -1.13395500e+00 5.50329328e-01
-1.43577874e-01 -1.01558399e+00 -2.53854424e-01 -5.58121316e-02] | [9.968993186950684, 5.325347900390625] |
01fec7bf-ceff-45f4-a713-fcb0134956fd | safe-ds-a-domain-specific-language-to-make | 2302.14548 | null | https://arxiv.org/abs/2302.14548v2 | https://arxiv.org/pdf/2302.14548v2.pdf | Safe-DS: A Domain Specific Language to Make Data Science Safe | Due to the long runtime of Data Science (DS) pipelines, even small programming mistakes can be very costly, if they are not detected statically. However, even basic static type checking of DS pipelines is difficult because most are written in Python. Static typing is available in Python only via external linters. These require static type annotations for parameters or results of functions, which many DS libraries do not provide. In this paper, we show how the wealth of Python DS libraries can be used in a statically safe way via Safe-DS, a domain specific language (DSL) for DS. Safe-DS catches conventional type errors plus errors related to range restrictions, data manipulation, and call order of functions, going well beyond the abilities of current Python linters. Python libraries are integrated into Safe-DS via a stub language for specifying the interface of its declarations, and an API-Editor that is able to extract type information from the code and documentation of Python libraries, and automatically generate suitable stubs. Moreover, Safe-DS complements textual DS pipelines with a graphical representation that eases safe development by preventing syntax errors. The seamless synchronization of textual and graphic view lets developers always choose the one best suited for their skills and current task. We think that Safe-DS can make DS development easier, faster, and more reliable, significantly reducing development costs. | ['Günter Kniesel-Wünsche', 'Lars Reimann'] | 2023-02-28 | null | null | null | null | ['type'] | ['speech'] | [-3.23775440e-01 9.63275433e-02 1.90135073e-02 -3.43599677e-01
-2.83678293e-01 -1.20102966e+00 4.01143849e-01 5.69999337e-01
-2.08276957e-01 2.91707307e-01 -2.76553929e-01 -7.30598390e-01
1.20080732e-01 -1.08891845e+00 -6.88265145e-01 -1.75307151e-02
-7.75381848e-02 1.18049167e-01 8.98162067e-01 -1.21237114e-01
2.39636406e-01 5.38938880e-01 -1.95779550e+00 3.48645896e-01
1.20061421e+00 6.18639529e-01 4.18719739e-01 7.10964382e-01
-7.63279974e-01 4.89081025e-01 -7.18671083e-01 -4.25919473e-01
-9.02160108e-02 1.06676906e-01 -7.05672741e-01 -8.18866849e-01
1.00818813e-01 -3.66099238e-01 4.59161460e-01 8.67017567e-01
3.67902517e-01 -6.07904732e-01 6.62147775e-02 -1.50909078e+00
1.17089376e-01 7.35274196e-01 -6.70562685e-02 -3.48581195e-01
5.54601312e-01 5.95906198e-01 5.40167272e-01 -6.00312233e-01
8.36590409e-01 1.00428891e+00 7.76231587e-01 6.39316559e-01
-1.40080714e+00 -4.89743739e-01 -4.01714951e-01 -1.63939312e-01
-1.14678192e+00 -3.78747642e-01 3.06134313e-01 -1.06252897e+00
1.21473718e+00 6.95896208e-01 5.96612155e-01 9.04130876e-01
-1.19790636e-01 2.49111339e-01 8.05724442e-01 -4.80709434e-01
4.59711045e-01 5.93977809e-01 6.19508088e-01 4.06683326e-01
5.24294853e-01 -2.45525971e-01 -2.16994032e-01 -4.15467799e-01
4.35045451e-01 -2.58830011e-01 -1.57529101e-01 -4.28498477e-01
-1.22448337e+00 3.09121221e-01 -2.17533022e-01 5.26362658e-01
2.75469095e-01 1.42338365e-01 9.47071552e-01 2.95842737e-01
-1.00012636e-02 4.82756346e-01 -9.70497489e-01 -6.95545673e-01
-4.62956935e-01 3.32837611e-01 1.19855356e+00 1.29727674e+00
7.80191898e-01 -3.73915657e-02 2.12221462e-02 6.37886405e-01
2.16118023e-01 3.92410934e-01 2.32166231e-01 -6.16799176e-01
1.59928471e-01 1.32544863e+00 9.70669538e-02 -7.10252523e-01
-3.99835199e-01 -2.88317740e-01 -1.86071381e-01 8.82429540e-01
7.08623588e-01 7.41932914e-02 -2.93950010e-02 1.27150226e+00
6.09083474e-01 -4.81099725e-01 3.60709280e-02 3.57898653e-01
8.23206067e-01 2.02572852e-01 6.21991344e-02 1.07981116e-01
1.18955743e+00 -2.21901208e-01 -4.05670792e-01 -1.08155042e-01
1.32231486e+00 -6.26864076e-01 1.39758801e+00 5.03734648e-01
-1.11633515e+00 -3.23824108e-01 -8.97935927e-01 -3.56377751e-01
-7.56937265e-01 1.45390093e-01 5.90695739e-01 8.14748049e-01
-1.12803602e+00 7.71438479e-01 -9.87080336e-01 -2.83633113e-01
2.01232150e-01 3.09769750e-01 -2.82352298e-01 6.31435096e-01
-6.72156096e-01 7.00633228e-01 4.12628382e-01 -1.78759560e-01
-5.11668563e-01 -1.35830331e+00 -1.04262877e+00 1.61428005e-02
3.53322029e-01 -5.02468586e-01 1.46096671e+00 -8.40727091e-01
-1.08001602e+00 1.12426341e+00 1.06936418e-01 -1.05115585e-01
7.85451829e-01 -1.48007378e-01 -2.89720923e-01 -4.78919804e-01
1.82840794e-01 -1.53969070e-02 3.19509655e-01 -9.96718347e-01
-5.01598358e-01 -3.58768374e-01 4.21785951e-01 -7.31273949e-01
-2.52940625e-01 5.16014636e-01 -3.99366766e-01 -3.36145401e-01
-5.31428814e-01 -4.49655175e-01 6.47633523e-03 3.59465599e-01
-2.60225207e-01 -1.39494151e-01 6.89547956e-01 -7.38780737e-01
1.50347388e+00 -2.40443397e+00 -9.30430442e-02 1.77015021e-01
2.28390127e-01 5.61087966e-01 3.33624542e-01 4.52533215e-01
-2.36219063e-01 3.39279026e-01 -2.81050712e-01 6.75206184e-02
5.78488767e-01 1.95719898e-01 -1.14105277e-01 2.89435118e-01
1.04765676e-01 3.14416230e-01 -9.52169955e-01 -9.57513079e-02
1.93219379e-01 2.87412733e-01 -6.89415395e-01 3.82918715e-01
-7.34109342e-01 2.52058893e-01 -3.60161632e-01 4.67810810e-01
9.94989097e-01 2.10104119e-02 4.01732624e-01 2.94692498e-02
-1.05136061e+00 8.46429646e-01 -1.68895304e+00 1.66268682e+00
-7.69541025e-01 3.04396927e-01 3.55864942e-01 -6.21254683e-01
1.06254089e+00 9.32294726e-02 -1.54018819e-01 -5.25284111e-01
-3.81950319e-01 5.53960800e-01 -3.02214563e-01 -5.59326410e-01
2.77349502e-01 7.00067401e-01 -2.66562611e-01 6.27208352e-01
-3.24364632e-01 -1.48155972e-01 6.37326419e-01 2.63703197e-01
1.12633967e+00 8.95862162e-01 3.29962313e-01 -3.65415454e-01
7.37929344e-01 1.31557241e-01 6.79411173e-01 6.37823224e-01
7.37092078e-01 2.45200947e-01 1.24255025e+00 -3.12206209e-01
-1.17641592e+00 -9.35298324e-01 -3.74644727e-01 1.06114364e+00
-5.78840554e-01 -1.25878465e+00 -9.28657591e-01 -6.81444347e-01
-1.09031856e-01 8.27534080e-01 -2.59641975e-01 9.40718800e-02
-6.42152727e-01 -2.55328566e-01 8.47185731e-01 5.81285775e-01
1.95902497e-01 -7.71823168e-01 -1.08978999e+00 1.70991495e-01
5.00458837e-01 -8.40486884e-01 -6.79463595e-02 -2.46788487e-02
-6.77673399e-01 -1.16416550e+00 -2.09364548e-01 -4.84657943e-01
4.45870191e-01 -2.19809845e-01 1.30780745e+00 4.55158412e-01
-5.65868258e-01 1.81722984e-01 -3.37738246e-01 -4.72143382e-01
-8.88761640e-01 -5.34747541e-02 -6.65871382e-01 -6.18088603e-01
1.38505325e-01 -6.85476065e-01 -1.38463214e-01 2.55610913e-01
-1.09052455e+00 2.09163710e-01 -2.04356655e-01 3.79709691e-01
-6.44174917e-03 -2.58566022e-01 -3.00156511e-02 -1.05642080e+00
1.30462825e-01 -2.88347661e-01 -1.48870409e+00 1.63449317e-01
-4.89734232e-01 2.17776299e-01 9.79405582e-01 9.40987095e-02
-8.98082316e-01 6.08373284e-02 -4.87895668e-01 3.22139800e-01
-3.26045245e-01 5.61341226e-01 -6.47491813e-01 1.34286404e-01
8.51280034e-01 1.08441174e-01 2.38641590e-01 -9.93980765e-01
-4.39043567e-02 6.62624717e-01 3.76284689e-01 -8.41042101e-01
7.81019926e-01 4.14102562e-02 -2.77022153e-01 -6.72144294e-01
1.11203134e-01 -9.02604684e-02 -5.14493883e-01 -2.38267872e-02
3.67700964e-01 -5.63733637e-01 -9.73073959e-01 7.43042231e-01
-1.40263402e+00 -6.61669075e-01 -2.78485477e-01 -2.03002140e-01
-1.22003749e-01 4.26741779e-01 -2.70889610e-01 -6.75041795e-01
-2.09322393e-01 -1.47385061e+00 9.46220875e-01 -1.10249773e-01
-5.12674332e-01 -9.37034786e-01 7.34615251e-02 -3.29191327e-01
6.72397733e-01 4.72995251e-01 1.05414891e+00 -4.56578463e-01
-6.58704460e-01 -1.59812912e-01 -2.00355157e-01 2.48233825e-01
-2.34477416e-01 9.83133316e-01 -1.06676066e+00 1.63748413e-01
-4.36638445e-01 4.42511052e-01 4.52231942e-03 -2.90923208e-01
1.13619089e+00 -3.38702172e-01 -4.55775917e-01 6.38272047e-01
1.53728509e+00 -9.55819339e-02 8.42982888e-01 8.67502332e-01
5.93160391e-01 6.76349103e-01 4.27444905e-01 7.01078236e-01
5.12676239e-01 8.63147736e-01 4.41262275e-01 -8.45828727e-02
6.39724061e-02 1.09609261e-01 5.64253211e-01 1.64208859e-01
2.13763371e-01 4.47243243e-01 -1.32492149e+00 2.91197747e-01
-1.74805164e+00 -8.03079307e-01 -1.31856263e+00 2.75168824e+00
1.23656607e+00 1.10827073e-01 5.75771511e-01 4.97174829e-01
3.95504683e-01 -4.38344002e-01 -3.10953502e-02 -7.20251203e-01
1.45509183e-01 3.10103506e-01 4.18161899e-01 5.34016311e-01
-6.64137244e-01 4.78289366e-01 5.57819271e+00 4.16892827e-01
-1.21262765e+00 8.75177011e-02 -3.04208249e-01 3.10725868e-01
-7.97726512e-01 4.15834159e-01 -1.10423028e+00 6.70704842e-01
1.01338542e+00 -5.36372900e-01 2.93988824e-01 1.30147088e+00
5.07253289e-01 -3.08313549e-01 -1.49169016e+00 3.65669847e-01
-5.76552808e-01 -1.53830504e+00 -4.47368383e-01 -1.66113302e-01
-9.63621680e-03 -1.00038201e-01 -5.64939499e-01 1.64922729e-01
2.90556282e-01 -6.74569666e-01 1.09087098e+00 5.76932371e-01
8.63026679e-01 -4.63516831e-01 3.83858263e-01 3.37817967e-01
-9.59027171e-01 3.23232651e-01 -2.61372775e-02 -1.92482099e-01
-3.06013256e-01 1.09430027e+00 -9.15384710e-01 6.85645163e-01
9.26657915e-01 6.47220314e-01 -9.71827030e-01 1.23191977e+00
-1.04687460e-01 2.97585636e-01 -5.26799500e-01 -8.27619061e-02
-5.67167878e-01 9.46701020e-02 4.94463682e-01 1.94759166e+00
1.68106094e-01 -6.06662035e-01 -2.28671342e-01 1.11345935e+00
7.20320046e-01 2.15869352e-01 -6.15758479e-01 -7.74835125e-02
3.60250950e-01 1.24120331e+00 -5.06524026e-01 -3.32889050e-01
-4.51738060e-01 3.79582614e-01 6.99116588e-02 -1.54560320e-02
-7.12576747e-01 -7.52229393e-01 6.94080353e-01 4.65802521e-01
2.04094648e-01 -4.40363228e-01 -6.81194723e-01 -1.15252185e+00
5.52154541e-01 -1.14420438e+00 3.34300399e-01 -7.45571077e-01
-8.39126170e-01 3.12084556e-01 3.35155725e-02 -9.40482140e-01
-2.89865285e-01 -5.98389268e-01 -6.20526195e-01 8.67804766e-01
-1.05400729e+00 -8.87051284e-01 -3.50093275e-01 3.21234912e-01
-2.61215955e-01 3.73850733e-01 1.16046000e+00 6.58750474e-01
-6.11013174e-01 6.57277942e-01 1.12860491e-02 1.19636603e-01
6.24526083e-01 -1.62164140e+00 7.74909496e-01 1.02741492e+00
-8.51738870e-01 1.20163536e+00 9.40974593e-01 -5.11683285e-01
-1.81215394e+00 -7.01550901e-01 9.64866579e-01 -3.54821563e-01
9.26537156e-01 -9.41235125e-01 -1.22407758e+00 4.95374054e-01
-1.80932522e-01 -1.26661718e-01 4.96181637e-01 1.98832955e-02
-8.41292918e-01 -4.23598796e-01 -1.07462680e+00 4.35165137e-01
6.95039928e-01 -4.71997261e-01 -2.50303149e-01 3.53318125e-01
5.37966490e-01 -6.64710522e-01 -1.21374655e+00 -1.23037435e-01
6.14626050e-01 -1.25891411e+00 8.20027769e-01 3.05973273e-02
2.15435028e-01 -8.56375039e-01 1.91238910e-01 -7.42720962e-01
4.45271552e-01 -8.18675995e-01 2.18492955e-01 1.94507277e+00
4.71002370e-01 -9.64520454e-01 1.15084365e-01 1.07828128e+00
-4.40035909e-01 9.35740471e-02 -2.60865688e-01 -8.41719329e-01
-1.01733275e-01 -8.78531933e-01 1.03144884e+00 1.10754275e+00
3.61760974e-01 -1.79602802e-01 4.24204290e-01 1.71110019e-01
1.72739610e-01 -7.70334750e-02 1.35769820e+00 -1.41387606e+00
-7.76467979e-01 -6.47198975e-01 -3.14591169e-01 -3.31092387e-01
-1.95057079e-01 -8.33697557e-01 -1.74848795e-01 -1.37506521e+00
-2.55471528e-01 -9.17463779e-01 8.78919423e-01 8.71450126e-01
2.18247309e-01 -1.28498018e-01 -1.59426764e-01 -1.20029889e-01
-5.61343171e-02 -3.56672108e-01 5.08602679e-01 2.27847785e-01
-1.60066515e-01 -1.10801600e-01 -3.99854809e-01 8.82151663e-01
5.24992645e-01 -7.35932946e-01 4.39133123e-02 -4.09785718e-01
8.25451970e-01 -1.16287105e-01 6.72525346e-01 -9.63791251e-01
1.73070475e-01 -1.72718242e-01 -1.20875545e-01 -4.78410780e-01
-2.69369334e-01 -8.90969157e-01 8.24276209e-01 5.08271158e-01
-9.71590132e-02 3.56565788e-03 5.97056448e-01 -2.66319841e-01
1.61868379e-01 -8.12846363e-01 4.28663462e-01 -3.26073527e-01
-6.57862246e-01 -1.26704901e-01 -4.67395484e-01 3.88081372e-03
1.15350318e+00 -2.21060306e-01 -6.16435528e-01 5.21149397e-01
-5.44659436e-01 2.78936550e-02 1.29315233e+00 2.43070066e-01
-2.25434318e-01 -5.56715488e-01 -2.09901571e-01 5.56678295e-01
3.18671674e-01 2.50436157e-01 -4.14013900e-02 7.35044062e-01
-9.75413144e-01 -1.31346390e-01 -1.34569421e-01 -7.07396269e-01
-1.35448718e+00 5.81284642e-01 2.75591046e-01 1.25632137e-01
-6.90284610e-01 3.26320469e-01 -1.33183897e-01 -6.53636336e-01
1.49251193e-01 -5.82847595e-01 1.19009279e-01 3.59554850e-02
1.16858864e+00 4.51677978e-01 6.65424168e-01 -7.99506530e-02
-7.82655895e-01 5.53009093e-01 2.60436773e-01 2.49882578e-03
1.49951959e+00 3.31295431e-01 -6.94353163e-01 5.98177910e-01
7.70891249e-01 6.56388819e-01 -1.02794576e+00 4.98962194e-01
2.84587204e-01 -5.14269173e-01 -5.20866215e-01 -1.00759125e+00
-5.42569399e-01 9.78636265e-01 1.62470639e-01 6.73170269e-01
8.55246484e-01 2.28392985e-03 1.19985647e-01 -8.48662704e-02
6.09310985e-01 -5.19818068e-01 -7.64403164e-01 3.48600477e-01
8.26489389e-01 -5.91046810e-01 -2.09311485e-01 -6.69850826e-01
-2.89958324e-02 1.67816889e+00 5.29639423e-01 3.36315989e-01
2.22737998e-01 1.13406992e+00 -1.60666123e-01 1.79199651e-02
-6.55133247e-01 -1.53534651e-01 -2.69757122e-01 1.04618025e+00
7.22427905e-01 -3.27058613e-01 -7.26867318e-01 8.14649642e-01
-1.71781972e-01 5.54927349e-01 1.00652587e+00 1.30524862e+00
-2.66818136e-01 -1.77203548e+00 -6.52565062e-01 2.37224088e-03
-5.29994249e-01 -1.43550411e-01 -2.80778348e-01 8.56566310e-01
2.97399879e-01 5.12415469e-01 1.26973331e-01 3.04373559e-02
4.81507123e-01 2.19438951e-02 4.12354738e-01 -7.96804905e-01
-1.19935274e+00 -1.97294191e-01 6.11062586e-01 -7.22395837e-01
4.23400179e-02 -7.84465015e-01 -1.46995068e+00 -5.13855875e-01
7.03083351e-02 9.58956704e-02 9.89174724e-01 8.79908264e-01
7.81040192e-01 2.61409730e-01 1.50657728e-01 -4.19250399e-01
-2.18416557e-01 -3.90296012e-01 -1.52563274e-01 -1.62315480e-02
1.96933359e-01 -3.89329076e-01 -3.34299058e-01 2.93548524e-01] | [7.812972068786621, 7.727957725524902] |
b11b7f80-b952-4e27-9a07-91ea76705414 | rigidity-preserving-image-transformations-and | 2201.13065 | null | https://arxiv.org/abs/2201.13065v2 | https://arxiv.org/pdf/2201.13065v2.pdf | Rigidity Preserving Image Transformations and Equivariance in Perspective | We characterize the class of image plane transformations which realize rigid camera motions and call these transformations `rigidity preserving'. In particular, 2D translations of pinhole images are not rigidity preserving. Hence, when using CNNs for 3D inference tasks, it can be beneficial to modify the inductive bias from equivariance towards translations to equivariance towards rigidity preserving transformations. We investigate how equivariance with respect to rigidity preserving transformations can be approximated in CNNs, and test our ideas on both 6D object pose estimation and visual localization. Experimentally, we improve on several competitive baselines. | ['Fredrik Kahl', 'Axel Flinth', 'Georg Bökman', 'Lucas Brynte'] | 2022-01-31 | null | null | null | null | ['6d-pose-estimation'] | ['computer-vision'] | [-1.15593500e-01 3.44774097e-01 -2.30972230e-01 -4.01913583e-01
-4.08535480e-01 -1.25358629e+00 7.79906332e-01 -6.24400020e-01
-5.30874610e-01 2.15335667e-01 5.75199246e-01 -2.86513746e-01
5.52484505e-02 -7.08005488e-01 -1.35884750e+00 -5.92343867e-01
1.42573223e-01 6.07134640e-01 7.69080222e-02 -1.04045101e-01
1.73663050e-01 1.20522249e+00 -8.39238346e-01 -3.68403532e-02
-4.83280867e-02 3.63449484e-01 -3.36589009e-01 7.91983962e-01
3.90146136e-01 3.41863871e-01 -4.94340109e-03 -4.97799873e-01
8.52378130e-01 1.96846426e-01 -1.22228086e+00 1.85719714e-01
9.16792810e-01 -7.74091780e-01 -7.33742356e-01 1.17436790e+00
3.39059055e-01 2.55193532e-01 7.10392058e-01 -8.51746082e-01
-9.84972179e-01 6.31919026e-01 -5.66096723e-01 1.96364298e-01
2.52523482e-01 1.42547460e-02 9.16043103e-01 -9.11788166e-01
1.09180343e+00 1.79097378e+00 5.94366610e-01 7.55572796e-01
-1.47683489e+00 -2.76781797e-01 2.69134969e-01 -2.12872967e-01
-1.18682766e+00 -6.49934173e-01 8.58188391e-01 -4.19761628e-01
1.10834861e+00 3.68765056e-01 3.62965167e-01 1.36192894e+00
2.71724463e-01 5.81849217e-01 7.03309238e-01 -9.74930525e-02
-1.58919513e-01 -3.41772437e-01 -3.05128664e-01 8.60121071e-01
3.36652130e-01 -3.69108020e-04 -1.60522088e-01 9.45640504e-02
1.55884647e+00 -3.81074138e-02 -2.31938690e-01 -9.09496486e-01
-1.54142416e+00 7.86617815e-01 7.64209986e-01 -6.52028620e-02
-3.91782857e-02 5.23616254e-01 1.39078185e-01 3.24511439e-01
2.51547575e-01 7.26308882e-01 -4.18234110e-01 1.82615519e-01
-1.42275870e-01 4.15442407e-01 5.77445984e-01 1.30747521e+00
6.07671618e-01 -8.14590752e-02 1.57617122e-01 4.77369159e-01
1.85149670e-01 5.64119458e-01 1.12398870e-01 -1.37930250e+00
4.84062582e-01 2.36694559e-01 1.32942379e-01 -9.98060226e-01
-6.58482373e-01 -3.26554000e-01 -7.96962678e-01 -5.64238764e-02
4.96939510e-01 4.45594192e-02 -9.52023923e-01 2.06927729e+00
1.99913025e-01 -2.08901033e-01 -9.18890387e-02 1.08307528e+00
5.16794682e-01 1.95430085e-01 -3.75900716e-01 1.25856131e-01
1.04211891e+00 -7.74443150e-01 -2.58061588e-01 -2.27576301e-01
6.12999260e-01 -9.55130637e-01 1.02065074e+00 -9.41667631e-02
-1.44873095e+00 -3.48370314e-01 -7.56248534e-01 -5.47164261e-01
2.82896403e-02 -1.74422294e-01 4.34315443e-01 2.68076301e-01
-1.27865279e+00 6.49236798e-01 -1.21161425e+00 -4.61665511e-01
2.68956393e-01 7.00986803e-01 -9.43860590e-01 1.72162384e-01
-7.86024272e-01 1.17606139e+00 2.51311243e-01 1.18158914e-01
-9.20035779e-01 -5.22307217e-01 -8.08721483e-01 -2.12952346e-01
1.77463964e-01 -1.01510835e+00 1.31695628e+00 -5.15265524e-01
-1.53243947e+00 1.36069345e+00 -1.74026966e-01 -5.32288730e-01
9.34069991e-01 -2.36213446e-01 3.53123695e-01 2.28090696e-02
-1.58297077e-01 1.12155616e+00 8.07683945e-01 -1.08804524e+00
-2.75625020e-01 -3.18530798e-01 7.06595421e-01 6.16667688e-01
-7.21175550e-03 -3.83757651e-02 -7.22470224e-01 -6.17074013e-01
5.83382607e-01 -1.55923593e+00 -3.55464756e-01 3.02173197e-01
-7.01958537e-01 -2.15433184e-02 7.75267243e-01 -4.31718051e-01
2.23563626e-01 -2.06327248e+00 6.91840231e-01 3.10115516e-01
3.20504695e-01 -3.55405629e-01 -2.93681294e-01 -1.16392434e-01
-1.91327199e-01 2.48248369e-01 2.03944921e-01 -1.98219046e-01
3.03840309e-01 4.99285609e-01 -3.83272648e-01 9.61906433e-01
2.83839285e-01 1.30225194e+00 -5.00024319e-01 -1.03628300e-01
4.34724450e-01 5.40729582e-01 -1.08965182e+00 -3.75414267e-02
-4.93691536e-03 7.91674495e-01 -3.87408614e-01 4.99789089e-01
7.93669164e-01 -5.87914810e-02 -3.54125388e-02 -5.69473386e-01
9.03026462e-02 3.71325105e-01 -9.20746326e-01 1.80014920e+00
-4.61889327e-01 6.77851498e-01 -8.82591233e-02 -9.84848022e-01
6.05653167e-01 4.96595912e-02 4.31601524e-01 -2.49710217e-01
4.28510994e-01 -7.95329362e-02 -6.30556196e-02 -1.05913863e-01
4.34095949e-01 -1.03037059e-01 8.20995867e-02 1.56651735e-01
3.46862040e-02 -3.07506025e-01 -4.30358887e-01 -1.91377118e-01
8.69115055e-01 2.37687692e-01 3.60329263e-02 -3.96566540e-01
3.05514753e-01 -3.83435488e-01 3.09913248e-01 5.85707724e-01
-9.22432765e-02 8.37631702e-01 5.24035275e-01 -8.52815747e-01
-1.53363240e+00 -1.51454484e+00 -1.64879173e-01 7.87303388e-01
2.04480663e-01 1.07228877e-02 -7.92806804e-01 -6.93208873e-01
-1.66539866e-02 1.59885198e-01 -7.06834674e-01 -3.54505152e-01
-1.04078066e+00 -4.68100369e-01 6.21781170e-01 8.38513792e-01
5.96302688e-01 -4.64507788e-01 -2.80333608e-01 -7.09791258e-02
-2.50365764e-01 -1.46376491e+00 -8.07767391e-01 2.26999640e-01
-1.08701563e+00 -8.80064011e-01 -6.51983559e-01 -7.95181155e-01
9.79830205e-01 5.51269293e-01 1.08885515e+00 -4.79055911e-01
2.09040478e-01 4.72447634e-01 2.29139486e-03 -9.66127366e-02
-2.40509182e-01 3.51857215e-01 3.42474341e-01 -2.69971043e-01
1.01074606e-01 -7.88218737e-01 -5.29902518e-01 5.01186073e-01
-6.66876554e-01 1.05585195e-01 5.21919727e-01 3.66371840e-01
8.86290848e-01 -3.19051117e-01 -5.10588825e-01 -6.64168775e-01
9.88206714e-02 1.79770201e-01 -7.23878086e-01 4.36526490e-03
8.38855579e-02 5.22446036e-01 3.82445723e-01 -6.06630623e-01
-7.24509597e-01 4.71128702e-01 -2.50935644e-01 -1.02846587e+00
-2.44582389e-02 -1.01787597e-01 -3.46472263e-01 -6.74341023e-01
7.49180973e-01 -1.98291495e-01 -2.28892028e-01 -4.37630355e-01
8.14638793e-01 -6.02179207e-02 8.88132036e-01 -8.60924542e-01
1.39614177e+00 9.92641687e-01 4.36616123e-01 -5.98239541e-01
-8.03344607e-01 -2.25921497e-02 -8.58198047e-01 1.91660970e-02
1.24028349e+00 -9.41379964e-01 -8.37218642e-01 1.59011811e-01
-1.42215192e+00 -7.72917196e-02 -2.25255296e-01 5.54721713e-01
-9.46165800e-01 3.75544637e-01 -5.89881957e-01 3.00750006e-02
-3.53432670e-02 -1.51339936e+00 1.27998543e+00 -1.89062446e-01
-4.15165871e-01 -1.00213742e+00 -7.25578666e-02 1.17873788e-01
2.18581930e-01 2.68198133e-01 8.92079890e-01 -4.79231089e-01
-8.26581299e-01 -1.58875495e-01 -1.07393667e-01 2.01097757e-01
8.07301626e-02 -4.07352000e-02 -8.53721976e-01 -5.01515150e-01
4.89809141e-02 -5.22455722e-02 6.59846187e-01 6.29051268e-01
1.26348197e+00 -7.35787332e-01 -3.35502103e-02 1.28916514e+00
1.04973900e+00 -2.00623587e-01 6.15885198e-01 5.60286343e-01
1.21050572e+00 4.06738013e-01 -3.50042135e-02 -9.02206004e-02
2.27442831e-01 8.34770381e-01 7.21007347e-01 -1.40592143e-01
-2.38989145e-02 -3.53601396e-01 3.34479421e-01 7.60272026e-01
-6.82592630e-01 9.52978507e-02 -6.67775631e-01 3.69182557e-01
-1.55310035e+00 -6.50355875e-01 1.67558014e-01 2.13062477e+00
6.80259705e-01 3.14391226e-01 4.04352508e-02 -3.08307618e-01
6.87135339e-01 2.52755016e-01 -5.18549979e-01 -4.81103361e-01
-3.21891606e-01 8.26331154e-02 1.13490999e+00 8.91028702e-01
-1.33278978e+00 1.16966975e+00 7.10091829e+00 2.08295390e-01
-1.14365315e+00 -5.73798344e-02 2.04635367e-01 -1.50180981e-01
-4.05065417e-01 1.02373976e-02 -7.30809450e-01 -8.74505416e-02
3.32077622e-01 7.19017312e-02 6.15717709e-01 1.00738835e+00
5.95743246e-02 7.77724326e-01 -1.64946592e+00 1.00429153e+00
-4.10416089e-02 -1.49081635e+00 6.37087405e-01 2.34374970e-01
9.16180849e-01 2.88720161e-01 2.90325940e-01 -3.17333251e-01
5.95851839e-01 -1.05760682e+00 9.25603509e-01 2.76937276e-01
6.38342083e-01 -8.39318812e-01 5.40524304e-01 3.27929668e-02
-9.09997523e-01 3.68319631e-01 -7.81115472e-01 2.06805959e-01
5.94615713e-02 7.08509162e-02 -8.31292868e-01 3.13223958e-01
6.82439387e-01 6.43083930e-01 -3.15284908e-01 5.35254955e-01
-3.25115055e-01 -2.41547953e-02 -4.55824703e-01 4.74100292e-01
3.82947594e-01 -1.84657261e-01 8.07997048e-01 1.15974092e+00
1.77327394e-01 8.22727010e-02 -3.52583826e-02 8.34756017e-01
-6.42606080e-01 -3.95476073e-01 -1.02891362e+00 2.70205230e-01
2.83393383e-01 9.88291502e-01 -7.98316061e-01 -3.35389078e-02
-2.67805904e-01 1.08912623e+00 2.36109078e-01 4.14179176e-01
-9.22492683e-01 1.32188514e-01 1.17352307e+00 1.23556219e-01
3.70580912e-01 -7.25678384e-01 -2.49099955e-01 -1.44884610e+00
4.30681147e-02 -6.54872239e-01 -1.32213477e-02 -9.11955893e-01
-1.00866246e+00 2.37678707e-01 1.54566437e-01 -1.08231175e+00
1.64575987e-02 -8.92790914e-01 -4.14090723e-01 5.79992592e-01
-1.07770491e+00 -1.23878181e+00 -1.44811571e-01 8.08265328e-01
4.42833960e-01 3.13860297e-01 4.92068321e-01 3.12328129e-03
1.13834431e-02 6.84199989e-01 -2.40959242e-01 4.22438323e-01
6.14286423e-01 -1.27367163e+00 9.36773121e-01 1.02691519e+00
3.40101838e-01 1.04715061e+00 8.32580447e-01 -1.82388514e-01
-1.78312576e+00 -1.30498111e+00 5.67456007e-01 -1.13034749e+00
5.38709342e-01 -4.14248079e-01 -7.52279341e-01 1.48924232e+00
-1.21670373e-01 6.88676015e-02 -3.59267890e-02 -2.74712294e-02
-8.72710824e-01 1.11693233e-01 -9.47580934e-01 1.06289828e+00
1.58633590e+00 -7.59760678e-01 -7.09876657e-01 4.64449644e-01
9.38207269e-01 -1.02879047e+00 -9.21642423e-01 5.39833844e-01
5.76576054e-01 -4.77156758e-01 1.51912475e+00 -1.03944421e+00
3.58210713e-01 -4.26446736e-01 -3.40141237e-01 -1.11434329e+00
-6.30700350e-01 -7.22921312e-01 1.45529285e-01 6.06718302e-01
2.64014691e-01 -5.19096553e-01 7.12808251e-01 4.96137023e-01
-1.90880328e-01 -2.23951280e-01 -1.11226201e+00 -9.78447676e-01
6.35850489e-01 -1.65526584e-01 5.71880579e-01 1.07970524e+00
-5.23281813e-01 2.81117946e-01 -3.91195893e-01 3.93463612e-01
4.83343065e-01 -2.02307239e-01 1.05137861e+00 -8.55791926e-01
-2.58347332e-01 -7.01876462e-01 -7.36969292e-01 -1.48868430e+00
4.00609314e-01 -1.07401276e+00 -1.01119794e-01 -1.06732714e+00
3.67536604e-01 1.80587590e-01 1.87765080e-02 5.76520920e-01
4.04826641e-01 5.11740446e-01 9.75363404e-02 4.47763354e-01
-3.48023415e-01 2.76694596e-01 1.60533047e+00 -1.70917124e-01
-1.19790584e-01 -1.97647795e-01 -6.26987517e-01 1.03505957e+00
6.86242878e-01 -2.78052896e-01 -3.91418755e-01 -1.04077375e+00
4.93680924e-01 -3.12963873e-01 5.92698932e-01 -4.81689304e-01
1.20237493e-03 -3.45019013e-01 4.12272006e-01 -4.44246411e-01
1.40103728e-01 -6.47123694e-01 2.89470673e-01 3.36477250e-01
-5.37815332e-01 4.99354422e-01 1.31316230e-01 2.25691646e-01
1.29900813e-01 2.51145244e-01 1.08443415e+00 -2.95956612e-01
-3.30786705e-01 6.33270919e-01 -2.01909602e-01 8.79642740e-02
4.86726791e-01 -2.65590817e-01 -2.25896642e-01 -3.62406671e-01
-8.57730806e-01 -2.96506375e-01 1.05560327e+00 5.93921304e-01
5.62457263e-01 -1.53287923e+00 -4.49387401e-01 4.63922806e-02
-9.16264802e-02 4.98208821e-01 -9.62785557e-02 8.16217542e-01
-9.67373967e-01 5.94921708e-01 -4.56589252e-01 -8.18261802e-01
-1.25841701e+00 5.33193767e-01 6.84355915e-01 1.37191743e-01
-8.06020856e-01 8.58253300e-01 8.17523837e-01 -7.58407235e-01
3.52280498e-01 -7.10344613e-01 1.18509598e-01 -6.24611199e-01
5.93709536e-02 1.40051290e-01 1.53272882e-01 -7.82786071e-01
-5.11765838e-01 1.28363621e+00 -3.19109038e-02 -2.06244081e-01
1.23569596e+00 -1.45576194e-01 -2.28045970e-01 -3.53277661e-02
1.60928893e+00 1.06779002e-01 -1.47415233e+00 -1.61012709e-01
-2.84635335e-01 -4.10197347e-01 -2.50373960e-01 -1.78434506e-01
-1.10430050e+00 8.40407372e-01 2.81921834e-01 -2.27537617e-01
8.13206434e-01 3.68775129e-01 3.04427832e-01 1.10885954e+00
1.66215315e-01 -6.90734029e-01 -6.24826439e-02 8.96429777e-01
1.16873503e+00 -8.87766361e-01 -9.19539481e-02 -4.42395777e-01
-2.12853700e-01 1.02740574e+00 4.95891005e-01 -6.83352292e-01
5.32081425e-01 6.54064491e-02 -9.84373316e-02 -1.40360072e-01
-4.72687930e-01 1.41925111e-01 5.27018428e-01 5.95971584e-01
2.14421973e-01 9.49655399e-02 1.85063388e-02 -1.98374391e-01
-8.02492917e-01 -4.94676024e-01 4.49266523e-01 5.23927629e-01
-2.18471646e-01 -8.58945131e-01 -3.97803754e-01 -1.81951270e-01
-4.40975934e-01 1.85616985e-01 -8.68652165e-01 1.04594743e+00
-6.17120489e-02 2.70433277e-01 2.48372510e-01 -4.36522752e-01
5.28850496e-01 -3.02238643e-01 1.05841506e+00 -3.94972652e-01
2.07424872e-02 2.01060399e-01 3.23939100e-02 -7.58760333e-01
-6.20007336e-01 -5.43625534e-01 -1.07287610e+00 -7.41846681e-01
-4.31994013e-02 -3.98066103e-01 4.77020234e-01 9.84247148e-01
7.82404542e-02 9.89724472e-02 6.21528566e-01 -1.05695629e+00
-7.44391739e-01 -5.27951717e-01 4.23061550e-02 5.33258498e-01
4.68756765e-01 -6.74773335e-01 -5.31186163e-01 1.58338428e-01] | [8.888689041137695, 2.3692519664764404] |
6d118bc7-13c4-4b7b-bf85-f5bcf3aac30c | i4u-system-description-for-nist-sre-20-cts | 2211.01091 | null | https://arxiv.org/abs/2211.01091v1 | https://arxiv.org/pdf/2211.01091v1.pdf | I4U System Description for NIST SRE'20 CTS Challenge | This manuscript describes the I4U submission to the 2020 NIST Speaker Recognition Evaluation (SRE'20) Conversational Telephone Speech (CTS) Challenge. The I4U's submission was resulted from active collaboration among researchers across eight research teams - I$^2$R (Singapore), UEF (Finland), VALPT (Italy, Spain), NEC (Japan), THUEE (China), LIA (France), NUS (Singapore), INRIA (France) and TJU (China). The submission was based on the fusion of top performing sub-systems and sub-fusion systems contributed by individual teams. Efforts have been spent on the use of common development and validation sets, submission schedule and milestone, minimizing inconsistency in trial list and score file format across sites. | ['Luis Buera', 'Longbiao Wang', 'Alfonso Ortega Giménez', 'Haizhou Li', 'Ruijie Tao', 'Hitoshi Yamamoto', 'Koji Okabe', 'Boning Zhang', 'Sandro Cumani', 'Md Sahidullah', 'Xuechen Liu', 'Héctor Deldago', 'Meng Liu', 'Ignacio Viñals Bailo', 'Rohan Kumar Das', 'Pierre-Michel Bousquet', 'Mickael Rouvier', 'Qiongqiong Wang', 'Tianyu Liang', 'Liang He', 'Hanwu Sun', 'Andreas Nautsch', 'Claudio Vair', 'Daniele Colibro', 'Tomi Kinnunen', 'Kong Aik Lee'] | 2022-11-02 | null | null | null | null | ['speaker-recognition'] | ['speech'] | [-1.61674097e-02 -8.77360441e-03 2.83057570e-01 -6.58122420e-01
-1.29933226e+00 -5.61557949e-01 6.25677347e-01 -2.92906046e-01
-2.38539413e-01 8.08260441e-01 6.21797442e-01 -6.76032722e-01
-1.10446084e-02 -9.67819914e-02 -1.33764714e-01 8.19787979e-02
-1.83279455e-01 3.43293101e-01 5.09634875e-02 -2.41221502e-01
4.11143810e-01 3.32445800e-01 -1.34829378e+00 4.00732517e-01
6.53245986e-01 7.60887980e-01 1.63366318e-01 8.51046562e-01
1.85877651e-01 4.62971509e-01 -4.67500538e-01 -2.11054221e-01
1.68719620e-01 -4.28967953e-01 -1.05195081e+00 -3.19199473e-01
2.55049169e-01 -7.48805180e-02 -4.12412882e-01 9.23958659e-01
7.41117001e-01 2.22895190e-01 3.76110196e-01 -1.02517593e+00
-1.15343481e-01 5.12398839e-01 -3.14193904e-01 3.15547377e-01
5.50503731e-01 5.65892570e-02 8.19482088e-01 -9.38301086e-01
5.36575198e-01 1.31704652e+00 8.11692953e-01 5.28654873e-01
-6.76819742e-01 -1.01340950e+00 2.14792341e-02 2.77865529e-02
-1.45911229e+00 -8.71447623e-01 3.19256067e-01 -1.54744968e-01
1.29593396e+00 7.07238257e-01 -7.41008744e-02 1.07362831e+00
-5.05737484e-01 8.19383025e-01 1.25388515e+00 -7.25487649e-01
3.45917135e-01 1.95727155e-01 4.88579810e-01 3.32122147e-01
-2.63549566e-01 -7.35246614e-02 -6.24409795e-01 -3.63855302e-01
4.58847702e-01 -7.45039940e-01 -3.66783708e-01 3.19991708e-01
-1.21308815e+00 5.55830598e-01 -1.85136825e-01 6.28636003e-01
-3.10656250e-01 -2.99944818e-01 5.15112877e-01 6.60347283e-01
2.92948812e-01 -1.04707927e-02 -7.57123768e-01 -9.51688170e-01
-8.77892375e-01 1.94705635e-01 9.93071198e-01 1.01298928e+00
2.28441313e-01 3.28975677e-01 1.13771014e-01 1.24195433e+00
6.10082388e-01 4.53568548e-01 3.33239287e-01 -1.10134304e+00
7.88378596e-01 1.16361585e-02 4.55618143e-01 -5.69651961e-01
-1.50034934e-01 2.67104823e-02 -4.45266962e-01 -2.07952708e-01
9.43273157e-02 -6.74224198e-01 -5.96625566e-01 1.64458513e+00
-1.34311855e-01 6.12292625e-02 8.69061202e-02 8.42344820e-01
7.40713179e-01 6.88946307e-01 -2.13883594e-02 -4.27919239e-01
9.24916804e-01 -6.64809406e-01 -5.94347060e-01 9.91681665e-02
4.14989024e-01 -1.15525484e+00 5.86180508e-01 8.12288582e-01
-1.30473626e+00 -7.00649977e-01 -1.07308698e+00 5.58796704e-01
-2.42027923e-01 2.47088403e-01 2.04456195e-01 1.24722469e+00
-1.59693408e+00 1.90693095e-01 -5.16391516e-01 -9.22956526e-01
-1.02863863e-01 7.33987033e-01 -2.02298090e-01 -1.80446491e-01
-1.22115183e+00 7.95208871e-01 -4.30531800e-02 2.60793835e-01
-6.78971827e-01 -3.14671367e-01 -5.21995246e-01 -4.16075826e-01
-2.09137008e-01 -2.20997762e-02 1.23039830e+00 -5.24907410e-01
-1.81103921e+00 5.12573421e-01 -5.16767383e-01 -1.56247467e-01
4.57972527e-01 1.75467819e-01 -1.01593101e+00 -3.18695784e-01
6.51217252e-02 6.35922253e-01 3.87701690e-02 -7.98553646e-01
-8.88350904e-01 -5.72333097e-01 -1.16781339e-01 2.63333470e-01
4.80678529e-02 8.50454032e-01 -2.99465775e-01 -4.91089791e-01
1.93778604e-01 -9.33576941e-01 1.19340874e-01 -9.45594132e-01
-2.22624168e-01 -2.72257000e-01 1.17678738e+00 -1.15172875e+00
1.17980218e+00 -2.20895958e+00 -5.09470180e-02 2.99635559e-01
-3.46136630e-01 4.50800240e-01 -1.36873081e-01 7.09604025e-01
-1.23858415e-01 1.22453831e-01 -3.89080867e-02 -4.00962472e-01
9.64136720e-02 -4.08589356e-02 -1.86040655e-01 3.82499754e-01
-4.25328046e-01 4.18068707e-01 -5.79914987e-01 1.48987826e-02
2.66011745e-01 2.40373313e-01 -9.76175740e-02 -4.89479788e-02
1.78767279e-01 4.52808619e-01 -4.00482446e-01 4.23276246e-01
6.86974287e-01 3.89575541e-01 3.15643489e-01 2.97301877e-02
-5.48266709e-01 8.00949097e-01 -1.67508674e+00 1.37236428e+00
-4.04137164e-01 7.10228145e-01 7.23925650e-01 -6.78436756e-01
1.09003413e+00 1.01387691e+00 2.62708634e-01 -2.97707111e-01
-4.15449440e-02 8.89951766e-01 -6.16510399e-02 -2.78955072e-01
6.13107979e-01 -1.70838945e-02 7.08148926e-02 8.08604419e-01
-9.70254019e-02 -9.05524865e-02 -4.79987524e-02 7.47156292e-02
7.77045906e-01 -3.37508023e-01 -6.78292736e-02 -3.81807715e-01
9.86602068e-01 -4.81858343e-01 5.61317742e-01 5.86494267e-01
-9.48893428e-01 5.38797677e-01 2.30511297e-02 -1.79204736e-02
-1.07606161e+00 -8.76646280e-01 -3.57174277e-01 8.80936980e-01
-4.81011808e-01 -3.09106827e-01 -8.07164669e-01 -5.00616789e-01
-4.75122213e-01 7.86176682e-01 -1.28598392e-01 3.96106631e-01
-8.01730692e-01 -1.31188467e-01 1.15466142e+00 3.18522573e-01
9.00585294e-01 -1.27265728e+00 1.37274906e-01 2.03607529e-01
-4.01598185e-01 -1.03393221e+00 -9.44015145e-01 1.88426331e-01
-6.49576187e-01 -4.78305817e-01 -9.55799162e-01 -9.61049378e-01
8.31893235e-02 2.96597093e-01 4.84225303e-01 -5.74005365e-01
3.50560509e-02 5.93389809e-01 -2.20531747e-01 -3.38107347e-01
-6.28637314e-01 6.04744703e-02 4.58023429e-01 -2.55475819e-01
5.97927988e-01 -1.72264516e-01 -4.63476539e-01 8.27874720e-01
-2.11239994e-01 -3.89737219e-01 9.43018124e-02 7.07212090e-01
-6.24451861e-02 -3.33915770e-01 1.19300437e+00 -3.39895815e-01
9.40271020e-01 -1.04780793e-01 -4.81287748e-01 2.85568476e-01
-5.35233736e-01 -5.83369613e-01 -3.05915847e-02 2.74189830e-01
-1.33419037e+00 -2.54316777e-01 -3.17939937e-01 -6.08151853e-02
-3.79457861e-01 2.74225444e-01 -4.17590171e-01 -1.41602546e-01
2.94339925e-01 3.33949268e-01 8.00430477e-02 -4.21453595e-01
-2.07401663e-02 1.88567936e+00 3.99328947e-01 -6.52306139e-01
2.23971590e-01 -1.19895868e-01 -7.88216650e-01 -1.19952095e+00
9.34199691e-02 -7.91773796e-01 -4.42125708e-01 -4.37910706e-01
6.23753905e-01 -9.89016891e-01 -5.38536072e-01 7.81807065e-01
-1.27850497e+00 -1.46610532e-02 2.14807734e-01 9.76728857e-01
-3.95112112e-02 6.41150773e-01 -7.37166286e-01 -1.04549468e+00
-3.31997633e-01 -1.32348669e+00 7.27776587e-01 4.22633171e-01
-8.35555673e-01 -7.82022655e-01 1.27523080e-01 1.05862737e+00
6.86774194e-01 -4.19214696e-01 3.12298030e-01 -6.26370549e-01
-3.95974487e-01 -4.51130778e-01 -3.22073966e-01 7.96160221e-01
1.21252678e-01 -6.24317266e-02 -1.25674736e+00 -4.83366966e-01
5.81241287e-02 -3.09198558e-01 3.29076201e-01 4.77563888e-01
6.59368515e-01 3.81618254e-02 -2.35670432e-01 -1.78860545e-01
6.67608976e-01 9.55144465e-01 6.05267823e-01 1.31011844e-01
1.01113908e-01 6.70830667e-01 3.85817647e-01 5.11207581e-02
6.07099116e-01 9.59339261e-01 -5.68930030e-01 4.77010190e-01
-9.25673321e-02 1.61954969e-01 8.55646551e-01 1.69111526e+00
-2.66948402e-01 -1.94314159e-02 -9.42737103e-01 7.08823979e-01
-1.62684357e+00 -8.36374581e-01 -4.15576100e-01 2.65819311e+00
6.87078178e-01 -1.66509319e-02 5.06424606e-01 8.94690156e-02
7.75706947e-01 7.94435963e-02 -6.05862848e-02 -8.56880546e-01
-1.49075732e-01 2.14620754e-01 1.82390004e-01 7.58150876e-01
-6.48897946e-01 6.07455790e-01 7.06016445e+00 8.67919922e-01
-1.11642313e+00 2.72750616e-01 7.00350404e-01 3.46345343e-02
-2.51787663e-01 -1.79553568e-01 -9.07592475e-01 3.20879012e-01
1.66875923e+00 -4.14112031e-01 5.49206495e-01 4.72109795e-01
2.84573436e-01 3.12387019e-01 -6.22337639e-01 8.01310718e-01
7.42585957e-03 -1.17899895e+00 -7.32444346e-01 2.99848169e-01
7.51957953e-01 8.31312239e-01 -1.55860260e-01 4.71483797e-01
4.51928943e-01 -9.18846965e-01 5.41926324e-01 6.88547119e-02
8.32326472e-01 -8.83264184e-01 7.07119286e-01 1.45973936e-01
-1.04898238e+00 1.26527492e-02 1.14148252e-01 1.17020369e-01
1.18996792e-01 1.74707472e-01 -9.74082053e-01 9.12324727e-01
6.43443465e-01 1.08236797e-01 3.20294872e-02 5.49868286e-01
3.33316296e-01 1.11247015e+00 -3.88989240e-01 -3.33392352e-01
2.73724556e-01 -2.14577407e-01 7.29181647e-01 1.62398446e+00
1.73555464e-01 2.00374395e-01 6.13824837e-02 2.45727360e-01
6.10559173e-02 -6.05988093e-02 -2.43083581e-01 1.09726481e-01
6.26163006e-01 9.20872867e-01 -4.08857197e-01 -8.77707526e-02
-5.03008306e-01 6.72134399e-01 -3.41663420e-01 4.97516483e-01
-3.74503762e-01 -4.69610602e-01 8.50041211e-01 -1.37823582e-01
2.80645669e-01 -3.05951029e-01 -4.33540791e-01 -6.97583735e-01
3.29600535e-02 -1.38101220e+00 1.94454193e-01 -6.16069317e-01
-9.60007727e-01 8.80993366e-01 -2.50414549e-03 -8.13903451e-01
-5.28259754e-01 -4.33518738e-01 -5.68334162e-01 1.45309031e+00
-7.14931190e-01 -6.82154894e-01 3.05736810e-01 4.86567199e-01
5.74825585e-01 -6.72135592e-01 1.01815534e+00 7.72724926e-01
-5.96977592e-01 9.80664194e-01 1.01702332e-01 -2.86163632e-02
4.64153439e-01 -8.68012309e-01 6.12753510e-01 4.58154142e-01
-3.12073827e-02 1.00133348e+00 5.41955888e-01 -5.00227213e-01
-1.27084935e+00 -9.38812256e-01 1.46470499e+00 -4.91830945e-01
5.93191326e-01 -3.27309042e-01 -6.17187858e-01 6.76351726e-01
5.00520289e-01 -6.79934561e-01 7.73699462e-01 3.88866931e-01
-2.37713367e-01 -1.07557103e-01 -1.27418780e+00 3.69188756e-01
5.94708264e-01 -1.10727000e+00 -4.27555412e-01 3.18641216e-01
3.63999933e-01 -3.58244389e-01 -8.69052768e-01 2.87224829e-01
7.01361775e-01 -7.05926001e-01 4.72750068e-01 -8.33550021e-02
-1.48912191e-01 -4.94226247e-01 -7.26889789e-01 -9.16630387e-01
1.62529290e-01 -8.97050142e-01 4.61801827e-01 1.37787783e+00
6.38781190e-01 -7.97463357e-01 6.58621728e-01 6.39867902e-01
-4.26273674e-01 -3.95615548e-01 -1.55970252e+00 -7.52840877e-01
-8.24944228e-02 -1.04833162e+00 4.42454964e-01 8.46221030e-01
1.89759657e-01 2.90374994e-01 -3.68066728e-01 -8.06989148e-02
4.91532475e-01 -5.58126509e-01 7.39549935e-01 -8.26131225e-01
-3.55308384e-01 -4.66867745e-01 -4.07190770e-01 -9.13389921e-01
-3.07588011e-01 -5.37347198e-01 -5.25452904e-02 -1.36073947e+00
-9.03312936e-02 -3.73150289e-01 -3.20880145e-01 2.86014915e-01
-5.23254685e-02 -6.81635886e-02 2.51266032e-01 2.62797266e-01
-4.46006685e-01 3.45816091e-03 7.68824816e-01 -1.02216959e-01
-1.35478884e-01 4.52889502e-01 -7.62653291e-01 9.98702422e-02
6.59789383e-01 -3.37408185e-02 -1.90033048e-01 -2.37745449e-01
-5.76172769e-01 2.68385381e-01 5.35374545e-02 -9.56805289e-01
3.11906248e-01 2.06356332e-01 -1.31239653e-01 -6.68623388e-01
4.52040583e-01 -3.39969188e-01 4.37252998e-01 3.64395350e-01
-1.90933019e-01 -2.11594589e-02 3.70368689e-01 1.32256851e-01
-4.02864069e-01 -6.47618398e-02 5.60430348e-01 -1.01756081e-01
-5.07877231e-01 -1.82264552e-01 -4.46041971e-01 -2.87687033e-01
8.31295907e-01 -4.22293603e-01 -2.68489897e-01 -6.50760353e-01
-5.51783383e-01 4.87986356e-01 1.79425944e-02 5.79311490e-01
3.07952523e-01 -9.15694475e-01 -9.73947942e-01 1.47982270e-01
-1.55018374e-01 -6.90243483e-01 4.08277571e-01 7.37893462e-01
-3.28816742e-01 9.17544067e-01 4.88855392e-02 -5.20714521e-01
-1.50596654e+00 -4.73507762e-01 1.12040922e-01 -1.78151011e-01
-8.22172984e-02 1.00015724e+00 -4.32716876e-01 -1.08735955e+00
5.33312857e-01 8.10211599e-02 4.44014743e-02 -1.42476827e-01
5.51589608e-01 6.05554402e-01 4.34091598e-01 -7.49887586e-01
-6.91548645e-01 2.01440558e-01 -4.34679687e-01 -8.95607173e-01
1.20074058e+00 -3.05075437e-01 5.95757142e-02 6.33879721e-01
1.29592800e+00 7.21967146e-02 -6.31319880e-01 2.02342704e-01
1.87627718e-01 1.11912563e-01 2.08018407e-01 -1.02439702e+00
-5.31605065e-01 5.58382034e-01 7.75219440e-01 2.16254182e-02
7.58390844e-01 -2.28704333e-01 8.18778157e-01 2.06412390e-01
8.17184985e-01 -1.39713538e+00 -5.52669406e-01 8.83417308e-01
1.01574528e+00 -8.38541031e-01 -3.99759412e-01 -1.12625986e-01
-7.11379468e-01 9.60017681e-01 7.68787116e-02 5.02156198e-01
8.99052620e-01 1.45387426e-01 4.07041848e-01 1.55128986e-01
-7.50611126e-01 2.78173029e-01 8.62554535e-02 7.73832917e-01
9.82766569e-01 3.42568010e-01 -6.35538936e-01 6.09441638e-01
-9.90421548e-02 5.16002774e-01 4.48499113e-01 9.75816667e-01
-3.13879967e-01 -1.33352566e+00 -5.17465115e-01 4.70761865e-01
-2.77134687e-01 -1.55493513e-01 -6.09447360e-01 6.84842467e-01
-4.55872193e-02 1.28937352e+00 -2.51074642e-01 -6.73926175e-01
4.85913008e-01 3.16756099e-01 1.35274470e-01 -4.04801965e-01
-1.16604114e+00 3.83503169e-01 9.05093074e-01 -8.37758705e-02
-2.40757957e-01 -1.24780273e+00 -9.06835139e-01 -5.56522846e-01
-1.97982654e-01 9.44782197e-01 9.86152649e-01 7.83654273e-01
5.43460190e-01 3.50165516e-01 9.06920850e-01 -5.21000326e-01
-7.26263463e-01 -1.45363903e+00 -7.64217675e-01 -3.01565915e-01
-2.01139972e-01 -2.10726306e-01 -4.07564729e-01 -2.73994625e-01] | [14.316417694091797, 6.231239318847656] |
72b800c7-e533-44d7-a380-91ac1d03849e | sense-imagine-act-multimodal-perception | 2305.04750 | null | https://arxiv.org/abs/2305.04750v1 | https://arxiv.org/pdf/2305.04750v1.pdf | Sense, Imagine, Act: Multimodal Perception Improves Model-Based Reinforcement Learning for Head-to-Head Autonomous Racing | Model-based reinforcement learning (MBRL) techniques have recently yielded promising results for real-world autonomous racing using high-dimensional observations. MBRL agents, such as Dreamer, solve long-horizon tasks by building a world model and planning actions by latent imagination. This approach involves explicitly learning a model of the system dynamics and using it to learn the optimal policy for continuous control over multiple timesteps. As a result, MBRL agents may converge to sub-optimal policies if the world model is inaccurate. To improve state estimation for autonomous racing, this paper proposes a self-supervised sensor fusion technique that combines egocentric LiDAR and RGB camera observations collected from the F1TENTH Gym. The zero-shot performance of MBRL agents is empirically evaluated on unseen tracks and against a dynamic obstacle. This paper illustrates that multimodal perception improves robustness of the world model without requiring additional training data. The resulting multimodal Dreamer agent safely avoided collisions and won the most races compared to other tested baselines in zero-shot head-to-head autonomous racing. | ['Ram Vasudevan', 'Yulun Zhuang', 'Hanxi Wan', 'Chetan Reddy', 'Elena Shrestha'] | 2023-05-08 | null | null | null | null | ['continuous-control', 'model-based-reinforcement-learning'] | ['playing-games', 'reasoning'] | [-1.39619142e-01 3.79688740e-01 -2.54989475e-01 -3.18101525e-01
-9.09693360e-01 -4.03250098e-01 6.19714200e-01 2.71944702e-01
-1.04573274e+00 1.10004389e+00 -8.57087038e-03 8.68057013e-02
-2.85025418e-01 -6.34625554e-01 -9.85526741e-01 -6.51277721e-01
-4.33454335e-01 1.00184369e+00 3.24195176e-01 -7.85186708e-01
2.71266967e-01 3.78453463e-01 -1.81588984e+00 5.71034923e-02
7.98110247e-01 5.65625906e-01 5.09692132e-01 1.14351654e+00
4.82594699e-01 1.12058842e+00 -3.71519893e-01 1.37020215e-01
3.86223525e-01 -1.95458487e-01 -5.22236407e-01 1.66929170e-01
3.36414546e-01 -2.52376109e-01 -4.25861239e-01 7.70125270e-01
4.88602608e-01 7.15302229e-01 1.86651856e-01 -1.69824076e+00
2.19151303e-01 4.08021659e-01 -3.20286632e-01 -6.98520243e-02
5.25651574e-01 7.30190814e-01 6.84594452e-01 -1.47220209e-01
6.11052752e-01 1.47740459e+00 5.10011852e-01 4.12292540e-01
-1.39020944e+00 -5.96847892e-01 2.22112879e-01 6.41571879e-01
-1.23781371e+00 -3.93788457e-01 2.94007808e-01 -1.59653738e-01
1.05181575e+00 -4.11399044e-02 9.77251530e-01 1.29404628e+00
6.01115584e-01 7.36678481e-01 1.39873946e+00 -1.25502065e-01
7.07740366e-01 -3.11627954e-01 -2.98647135e-01 1.21302402e+00
1.74145952e-01 7.01251209e-01 -9.29696083e-01 -4.51501906e-02
3.56978297e-01 -3.80351156e-01 1.21957950e-01 -8.88972640e-01
-1.33777225e+00 7.73338974e-01 5.69523692e-01 -4.48445559e-01
-5.83890617e-01 4.12595063e-01 1.94338530e-01 3.29212695e-01
-1.94154993e-01 7.78274298e-01 -3.43337417e-01 -5.46647310e-01
-4.54755336e-01 8.65873158e-01 6.44066870e-01 7.85661697e-01
8.84169698e-01 2.06246108e-01 1.61144555e-01 4.88640189e-01
1.74176231e-01 8.33565652e-01 3.11706841e-01 -1.62435591e+00
4.21256781e-01 5.07379532e-01 6.43551111e-01 -5.36817491e-01
-8.58703315e-01 -9.39638093e-02 -4.41015474e-02 1.04767764e+00
4.22482669e-01 -4.70352501e-01 -1.07410479e+00 1.69475818e+00
5.73419750e-01 3.59838665e-01 6.48606896e-01 1.14939010e+00
2.00521067e-01 5.49803257e-01 4.61169034e-02 3.14620920e-02
1.02847469e+00 -9.85072076e-01 -7.22548962e-01 -6.05912626e-01
5.77840984e-01 -2.25854710e-01 9.08108890e-01 7.48877406e-01
-9.26273942e-01 -7.17920423e-01 -1.30447781e+00 1.94026843e-01
-3.69563818e-01 -3.70019376e-02 4.00215030e-01 3.69892001e-01
-8.18661690e-01 5.98537862e-01 -1.35594487e+00 -5.34625113e-01
6.91009015e-02 6.82557404e-01 -6.03242934e-01 -8.67708772e-02
-9.61063981e-01 1.42761600e+00 7.09618390e-01 -5.01486547e-02
-1.71136308e+00 -2.58192569e-01 -1.08114564e+00 -4.74780202e-01
8.91089797e-01 -4.74999040e-01 1.51095867e+00 -4.36568916e-01
-1.96098638e+00 3.68752718e-01 1.95018858e-01 -9.79957104e-01
3.30890268e-01 -7.17696071e-01 -2.84896135e-01 3.74620035e-02
1.75442383e-01 9.34272587e-01 6.92758203e-01 -1.48188913e+00
-1.05714417e+00 -4.90168899e-01 2.84513086e-01 7.26958513e-01
4.64066714e-01 -9.67452109e-01 -2.19496880e-02 6.95649311e-02
9.15975496e-02 -1.49495649e+00 -5.65963089e-01 -2.10864633e-01
-1.08439147e-01 -1.12110823e-01 8.19898188e-01 -9.72536281e-02
3.71927798e-01 -1.64502680e+00 5.83108902e-01 9.11347941e-03
-2.10399434e-01 1.20999506e-02 -1.77536547e-01 6.78018272e-01
4.40086752e-01 -7.17842698e-01 -1.14203274e-01 -4.63508308e-01
1.24542452e-01 8.70178521e-01 -2.63075739e-01 7.15761423e-01
-2.67229766e-01 7.97765195e-01 -1.09878922e+00 -4.33043122e-01
5.35079896e-01 1.75227195e-01 -4.01853114e-01 3.74863178e-01
-5.73370516e-01 6.54860377e-01 -2.77850032e-01 3.69081676e-01
1.68435916e-01 4.27906930e-01 3.47091675e-01 6.90231025e-02
-5.20663857e-01 -1.43258125e-01 -1.25282705e+00 2.44779706e+00
-3.70647311e-01 5.15936732e-01 3.08386624e-01 -9.23844635e-01
7.26985931e-01 -1.79944038e-02 4.57364291e-01 -9.99721766e-01
1.67741701e-01 -6.06325567e-02 -7.93844834e-02 -7.44083703e-01
7.67511249e-01 -2.67705947e-01 -4.44702566e-01 1.00712091e-01
1.75322652e-01 -6.59350872e-01 -1.89736336e-02 1.36905953e-01
1.09632313e+00 7.96312332e-01 1.25176847e-01 -6.19802177e-02
2.43655562e-01 7.88113713e-01 6.47276819e-01 9.25382376e-01
-3.65458101e-01 1.27768233e-01 -2.37297639e-02 -5.08384764e-01
-7.84727395e-01 -1.25299358e+00 4.75012422e-01 1.19259489e+00
7.57321477e-01 -2.63277888e-01 -7.45486438e-01 -3.53057891e-01
3.03532928e-02 1.19533777e+00 -7.40472138e-01 -5.06788731e-01
-7.38672614e-01 -5.21063983e-01 4.29618210e-01 2.10478902e-01
5.20026982e-01 -1.02804375e+00 -1.58143795e+00 4.63318557e-01
-3.64216536e-01 -1.35953426e+00 1.76924482e-01 2.88905472e-01
-6.12245083e-01 -1.31906545e+00 -1.03258707e-01 -2.13910177e-01
1.96397856e-01 7.63892680e-02 7.80719578e-01 -3.07775944e-01
-3.32074910e-01 8.40250969e-01 -3.15183938e-01 -5.88992417e-01
-4.85819548e-01 -1.92447558e-01 6.25423133e-01 -2.44933620e-01
1.15837611e-01 -2.02245131e-01 -4.05175149e-01 2.81676888e-01
-3.28244060e-01 2.00409964e-01 6.18375301e-01 6.66393220e-01
5.53619683e-01 1.92239620e-02 2.09981114e-01 -1.23878941e-01
5.60144782e-01 -1.29374042e-01 -9.47487831e-01 1.38365120e-01
-4.89891857e-01 3.65587443e-01 3.97420853e-01 -4.02934790e-01
-1.16433012e+00 3.63551706e-01 1.12326458e-01 -3.11260074e-01
-3.99537683e-01 4.10954244e-02 1.96628958e-01 9.92298573e-02
8.81456316e-01 5.80970757e-02 2.45042235e-01 -1.39954656e-01
5.43682575e-01 1.94543898e-01 8.72916698e-01 -8.34419012e-01
7.51220942e-01 8.58101547e-01 3.42865765e-01 -8.54719341e-01
-7.93081164e-01 -3.18475068e-01 -6.57893419e-01 -5.48998594e-01
1.19135499e+00 -1.14280009e+00 -1.37634504e+00 3.48634183e-01
-5.96852422e-01 -9.73653197e-01 -5.56429327e-01 1.01665485e+00
-1.23881769e+00 -5.39640822e-02 -2.78733909e-01 -1.15355718e+00
3.40255201e-01 -1.18039453e+00 1.07889891e+00 3.66501749e-01
-9.78433192e-02 -6.28557801e-01 7.21355498e-01 4.33318585e-01
-4.63162549e-02 5.41698158e-01 3.65419954e-01 -6.84360266e-02
-5.14235675e-01 7.95759037e-02 4.80312109e-01 -3.21610749e-01
-1.54336348e-01 -5.90549469e-01 -8.26146483e-01 -6.83432758e-01
-3.34651351e-01 -8.66174042e-01 6.64629996e-01 3.29147041e-01
4.80220139e-01 -2.86528617e-02 -3.73912871e-01 3.01036745e-01
1.28719568e+00 1.87961787e-01 3.27294528e-01 7.50727594e-01
4.99378085e-01 8.04934025e-01 1.36018252e+00 6.05901718e-01
8.29536021e-01 8.04041862e-01 1.09453511e+00 2.99689062e-02
2.24674776e-01 -4.89051431e-01 6.17829621e-01 2.33955935e-01
-1.75740972e-01 -6.94361776e-02 -9.38244045e-01 3.98260772e-01
-2.13094974e+00 -9.91939366e-01 1.15017451e-01 2.23660374e+00
3.87254089e-01 2.82892019e-01 2.65956670e-01 -1.70414194e-01
1.79428682e-01 -2.75822240e-03 -1.07113671e+00 -3.21903318e-01
4.56840955e-02 -1.34570479e-01 8.51870656e-01 8.12744558e-01
-9.47765172e-01 1.26444519e+00 5.89836025e+00 5.12527168e-01
-7.42064416e-01 1.18992858e-01 -5.51825427e-02 -4.64787573e-01
4.44147021e-01 2.21574619e-01 -8.48163009e-01 1.52769554e-02
1.25948429e+00 1.65667906e-01 6.94109797e-01 7.84601390e-01
5.40039718e-01 -8.49208117e-01 -9.66276586e-01 9.36815858e-01
-4.80200127e-02 -1.22168779e+00 -6.27677143e-01 1.18375488e-01
5.58858216e-01 5.01177549e-01 4.28931136e-03 8.17176282e-01
1.15264022e+00 -1.01987326e+00 1.02025032e+00 6.64216518e-01
1.76436096e-01 -1.01780188e+00 4.29348886e-01 8.86948824e-01
-9.46000278e-01 -6.71465576e-01 -3.38594615e-01 -4.06588882e-01
4.50933576e-01 -3.19378048e-01 -9.16670620e-01 5.48552036e-01
7.00646102e-01 3.74912798e-01 -4.98548359e-01 8.82094622e-01
-7.48039633e-02 2.45473921e-01 -4.68588501e-01 2.82990597e-02
6.41779363e-01 -3.61716986e-01 8.25941443e-01 5.82225084e-01
9.17220041e-02 3.13522786e-01 8.22260916e-01 3.99065197e-01
5.77845454e-01 -3.89041573e-01 -7.04591334e-01 3.03482026e-01
2.18019523e-02 1.01669037e+00 -5.74780047e-01 -3.83007109e-01
1.35753229e-01 8.11627924e-01 5.39042056e-01 4.11369205e-01
-9.20330226e-01 4.53286543e-02 9.23463523e-01 -1.07163973e-01
3.17319781e-02 -7.61915326e-01 2.32626975e-01 -8.21717203e-01
-5.41671336e-01 -8.84622276e-01 2.93606430e-01 -9.21489120e-01
-5.74831426e-01 4.55324680e-01 3.83003294e-01 -1.20181358e+00
-4.78050411e-01 -4.11768824e-01 -1.37570426e-01 1.69533953e-01
-1.39078343e+00 -1.04273033e+00 -4.74973112e-01 5.97562492e-01
8.22980106e-01 -2.91650265e-01 8.45765173e-01 -3.88861001e-01
-3.80172670e-01 -7.94296637e-02 -1.16640516e-01 -3.91105592e-01
6.43040419e-01 -1.45072782e+00 3.19142751e-02 5.21891296e-01
1.14078134e-01 1.76031888e-02 1.26973128e+00 -7.72943854e-01
-1.73517239e+00 -7.63341725e-01 -2.19830081e-01 -3.87241632e-01
4.68319148e-01 -3.04185748e-01 -5.95259368e-01 6.29490614e-01
4.13520128e-01 -7.22496137e-02 2.37789556e-01 -3.36571261e-02
1.69757783e-01 -1.24121092e-01 -1.06483495e+00 8.36551130e-01
7.99323082e-01 -1.59868434e-01 -7.45402932e-01 3.10194969e-01
5.80309451e-01 -7.21108258e-01 -6.12376153e-01 3.26080382e-01
7.34282553e-01 -1.07875049e+00 8.70121419e-01 -8.21884811e-01
-2.22599238e-01 -4.96142089e-01 -2.70099789e-01 -1.73644507e+00
4.27980274e-02 -7.24479795e-01 -2.51855582e-01 2.20629930e-01
-3.29850875e-02 -3.43477249e-01 9.84625936e-01 3.59363258e-01
-2.38850683e-01 -3.77883017e-01 -1.19117522e+00 -8.45483541e-01
-1.44419521e-01 -4.47338372e-01 2.67851591e-01 4.84399050e-01
2.94869449e-02 2.66268522e-01 -6.44306839e-01 6.30723536e-01
1.06883478e+00 -4.85125333e-02 1.29487491e+00 -9.97091055e-01
-2.06707582e-01 1.53861687e-01 -4.35783923e-01 -8.91084015e-01
5.14306188e-01 -3.92602354e-01 6.57940984e-01 -1.60233808e+00
-4.14842337e-01 -3.83487433e-01 -2.91845147e-02 4.48204726e-01
1.80607781e-01 -1.34964881e-04 2.59498328e-01 -1.29680231e-01
-9.83716369e-01 9.27134931e-01 1.27666223e+00 -2.43421998e-02
-4.49716747e-01 -1.80563003e-01 9.32852477e-02 7.88093626e-01
1.12339640e+00 -4.67233121e-01 -7.05978155e-01 -3.14866662e-01
1.35838509e-01 4.83710289e-01 6.39348567e-01 -1.59540367e+00
5.50244749e-01 -4.84753698e-01 1.60680503e-01 -7.37974465e-01
1.22451353e+00 -8.23076427e-01 -3.08883134e-02 9.43904340e-01
-4.16161537e-01 3.06139886e-01 4.10053134e-01 1.08701742e+00
1.13643751e-01 6.54129013e-02 7.39187956e-01 -3.27834934e-01
-1.32124114e+00 -7.90048204e-03 -9.39698756e-01 4.09822427e-02
1.42136765e+00 -1.75623789e-01 -2.21210390e-01 -5.21632850e-01
-9.30502236e-01 7.74103582e-01 4.10010338e-01 6.15628421e-01
6.82661474e-01 -9.96688187e-01 -6.60885394e-01 1.67142153e-01
8.80912244e-02 -1.22839406e-01 2.98378557e-01 8.13703179e-01
-4.71465856e-01 2.22273931e-01 -6.94222927e-01 -7.75375664e-01
-1.18253458e+00 4.25014079e-01 4.91711050e-01 -1.05636612e-01
-9.25426841e-01 5.60506821e-01 -2.12369382e-01 -8.28039825e-01
1.88527063e-01 -2.39122361e-01 -2.13497818e-01 -2.28901029e-01
2.51963794e-01 5.97104073e-01 -2.62831002e-01 -8.25356424e-01
-1.26900077e-01 4.28756118e-01 1.16821945e-01 -8.92755151e-01
1.13062274e+00 -3.98784131e-01 6.20679379e-01 7.30615497e-01
5.85589945e-01 -4.13885832e-01 -1.98053229e+00 1.57396123e-02
1.21752560e-01 -3.03374141e-01 2.95186758e-01 -8.00224721e-01
-3.30184340e-01 6.11470044e-01 1.15158916e+00 -2.95209050e-01
5.58097780e-01 -2.62334168e-01 6.24953091e-01 1.03961778e+00
8.26570749e-01 -1.65706360e+00 4.67629284e-01 5.21235406e-01
7.89022028e-01 -1.65307319e+00 2.37464368e-01 2.95152932e-01
-1.14229906e+00 8.33577931e-01 9.72117245e-01 -3.91303688e-01
1.00208730e-01 2.09206745e-01 3.35417002e-01 -2.67137319e-01
-1.02830255e+00 -5.67705631e-01 -1.15533166e-01 8.64978254e-01
-6.57917023e-01 1.85989976e-01 4.04391378e-01 -2.21475899e-01
-4.93702441e-01 -1.83421120e-01 6.26315117e-01 1.25011098e+00
-8.96091461e-01 -9.43890929e-01 -7.13569403e-01 1.02773085e-02
2.11530790e-01 7.20528126e-01 4.97365110e-02 1.23973536e+00
1.93248168e-01 1.11689484e+00 5.55417575e-02 -3.52445990e-01
5.67612469e-01 -1.10873379e-01 8.31333756e-01 -4.53700960e-01
-3.66389930e-01 -8.90531018e-02 2.51187325e-01 -1.22188723e+00
-4.11164522e-01 -7.58442998e-01 -1.76352727e+00 -8.59409794e-02
4.70556319e-03 2.09974244e-01 8.32077086e-01 1.00431144e+00
1.30884901e-01 6.42964542e-01 1.94462270e-01 -1.21318257e+00
-5.56452036e-01 -6.67956471e-01 -3.43486339e-01 1.91316366e-01
6.70101881e-01 -1.22416711e+00 1.34462371e-01 -1.93474457e-01] | [4.738095760345459, 1.090179204940796] |
2789e7b0-763f-45ed-9a83-0519fd09c452 | filling-in-the-details-perceiving-from-low | 1604.04125 | null | http://arxiv.org/abs/1604.04125v1 | http://arxiv.org/pdf/1604.04125v1.pdf | Filling in the details: Perceiving from low fidelity images | Humans perceive their surroundings in great detail even though most of our
visual field is reduced to low-fidelity color-deprived (e.g. dichromatic) input
by the retina. In contrast, most deep learning architectures are
computationally wasteful in that they consider every part of the input when
performing an image processing task. Yet, the human visual system is able to
perform visual reasoning despite having only a small fovea of high visual
acuity. With this in mind, we wish to understand the extent to which
connectionist architectures are able to learn from and reason with low acuity,
distorted inputs. Specifically, we train autoencoders to generate full-detail
images from low-detail "foveations" of those images and then measure their
ability to reconstruct the full-detail images from the foveated versions. By
varying the type of foveation, we can study how well the architectures can cope
with various types of distortion. We find that the autoencoder compensates for
lower detail by learning increasingly global feature functions. In many cases,
the learnt features are suitable for reconstructing the original full-detail
image. For example, we find that the networks accurately perceive color in the
periphery, even when 75\% of the input is achromatic. | ['Michael L. Wick', 'Marc Pomplun', 'Farahnaz Ahmed Wick'] | 2016-04-14 | null | null | null | null | ['foveation'] | ['computer-vision'] | [ 9.74968597e-02 1.30160555e-01 5.45539618e-01 -2.41909459e-01
-3.40576395e-02 -7.07455754e-01 4.67076391e-01 -3.86174142e-01
-5.25322020e-01 5.73864281e-01 1.35462865e-01 -2.94047058e-01
-1.66983122e-03 -1.04679537e+00 -8.22254539e-01 -6.44226909e-01
4.73728836e-01 -8.26580264e-03 1.34413913e-01 -2.68339366e-01
3.22825074e-01 6.74963593e-01 -2.07244325e+00 5.78848720e-01
6.92001224e-01 7.85122991e-01 4.79741096e-01 1.02831507e+00
2.53295183e-01 6.52597785e-01 -6.47380173e-01 -2.76092589e-01
5.92905521e-01 -6.41340613e-01 -8.34752262e-01 1.15231387e-01
7.95943975e-01 -7.06105292e-01 -5.98940313e-01 1.13471115e+00
2.92052507e-01 2.64629405e-02 6.58629298e-01 -6.34379268e-01
-1.05811715e+00 -1.80459842e-01 -2.56381154e-01 4.59463447e-01
5.19972265e-01 5.95392823e-01 5.65548360e-01 -9.27063763e-01
4.39558119e-01 1.26015675e+00 3.08064997e-01 5.05197525e-01
-1.52541065e+00 4.56536636e-02 -8.24170932e-03 1.58074334e-01
-1.09568357e+00 -6.96920872e-01 4.21999902e-01 -3.83278131e-01
1.28013813e+00 9.74314883e-02 9.69833136e-01 5.74746192e-01
5.71704090e-01 1.41505748e-01 1.52472878e+00 -5.37354648e-01
1.25590593e-01 1.91464722e-02 -4.19844389e-01 6.22575521e-01
6.13106310e-01 4.38013762e-01 -3.92135471e-01 2.39140809e-01
1.14429283e+00 8.87939632e-02 -5.67871094e-01 -1.48521131e-02
-1.19940400e+00 3.05399776e-01 9.51747775e-01 4.06822234e-01
-5.93264699e-01 7.53975287e-03 -3.57349098e-01 5.54888189e-01
-2.24264618e-02 9.63323593e-01 -2.50794679e-01 2.72736102e-01
-5.69711447e-01 4.10808027e-02 4.77181226e-01 5.43098748e-01
9.57613111e-01 2.31582552e-01 1.40506342e-01 5.19071341e-01
-3.64354663e-02 2.38617912e-01 5.79195797e-01 -1.46133959e+00
-9.19695850e-03 5.10089695e-01 2.89455205e-01 -6.55294359e-01
-3.47793579e-01 -3.77794534e-01 -7.42100000e-01 1.16182339e+00
8.97091806e-01 -2.03045771e-01 -1.06307423e+00 1.57950532e+00
-1.58009842e-01 -4.17033404e-01 2.14635998e-01 1.26296580e+00
3.63021165e-01 5.94292998e-01 -2.11591259e-01 -2.91765463e-02
1.26325917e+00 -5.35059929e-01 -3.40615243e-01 -7.21412301e-01
4.29624803e-02 -6.61763191e-01 1.38258517e+00 5.56020975e-01
-1.42139030e+00 -9.72346663e-01 -1.16841674e+00 -4.67935413e-01
-4.25215453e-01 -4.65432256e-02 5.92854261e-01 3.03500682e-01
-1.56646895e+00 6.47772968e-01 -6.32516563e-01 -5.15954494e-01
3.51510465e-01 7.50922784e-02 -6.75535619e-01 -3.64821345e-01
-8.16420913e-01 1.17549264e+00 3.93633604e-01 1.20407388e-01
-8.32666457e-01 -5.96270263e-01 -6.71200097e-01 3.78063411e-01
-1.38502136e-01 -1.19058573e+00 1.26261449e+00 -1.43677223e+00
-1.08821654e+00 9.35726464e-01 -2.28243038e-01 -2.93273538e-01
2.03485638e-01 1.35200754e-01 -2.26456165e-01 4.73716587e-01
-3.19733828e-01 7.17833877e-01 9.86130953e-01 -1.44083643e+00
-4.80049312e-01 -5.62601984e-01 4.81868416e-01 2.79673040e-01
1.53653994e-01 -2.91398436e-01 2.40544826e-01 -2.78674692e-01
2.49637634e-01 -6.08737051e-01 8.61014128e-02 4.87826377e-01
7.14748800e-02 2.82877743e-01 3.24784130e-01 -6.49248779e-01
4.47344303e-01 -2.28455138e+00 -3.96404564e-02 -6.19710935e-03
5.28331697e-01 2.82873750e-01 -2.31599525e-01 3.19332153e-01
-2.78236121e-01 2.10874975e-02 -1.36274338e-01 7.95892403e-02
-3.46023023e-01 2.54855663e-01 -2.01229393e-01 3.84806514e-01
4.08818305e-01 9.05246973e-01 -8.16453099e-01 -2.28486098e-02
1.09440848e-01 7.15458810e-01 -7.13610470e-01 2.00731263e-01
-1.79484754e-03 3.56890023e-01 1.16319187e-01 2.69577622e-01
6.43124223e-01 -1.42102465e-01 4.63906489e-02 -2.00794220e-01
-2.81688601e-01 1.27587661e-01 -6.95919931e-01 1.63243484e+00
-4.58605826e-01 1.16459835e+00 -4.65132967e-02 -8.35375965e-01
6.22207522e-01 1.79679379e-01 -1.82297915e-01 -1.24203408e+00
1.63698643e-01 2.90669322e-01 7.47362733e-01 -4.55871701e-01
1.45077586e-01 -6.30982637e-01 5.90560019e-01 4.03269976e-01
2.73512467e-03 -3.87712836e-01 3.63133522e-03 -5.01810685e-02
9.66040134e-01 -1.33346885e-01 3.89612556e-01 5.59776165e-02
6.62786290e-02 -6.16494194e-02 2.92264782e-02 6.93293691e-01
-2.22048327e-01 9.60583270e-01 4.20488954e-01 -7.85970509e-01
-1.44936347e+00 -1.43676674e+00 -1.30306497e-01 6.63423657e-01
-3.56579274e-02 2.59563953e-01 -8.57338488e-01 6.18546829e-02
-1.97064191e-01 8.27474475e-01 -6.85803592e-01 -4.39815223e-01
-9.85991582e-02 -3.42036426e-01 2.60355592e-01 4.11822945e-01
7.59019792e-01 -1.39283442e+00 -1.21222758e+00 -5.66594154e-02
7.19273239e-02 -7.67707586e-01 2.84583587e-02 2.99943574e-02
-6.87715113e-01 -1.21340263e+00 -8.26676071e-01 -8.31215382e-01
9.47355092e-01 5.80357373e-01 1.13832760e+00 1.82958141e-01
-5.99365592e-01 4.00084406e-01 -2.39409730e-02 -2.43760854e-01
-3.39161664e-01 -6.66761100e-01 -4.86322865e-02 -9.47209448e-02
3.27473938e-01 -7.16891885e-01 -9.22912478e-01 4.39492203e-02
-1.06551456e+00 4.26276997e-02 9.79447901e-01 7.57086277e-01
4.34821874e-01 3.73505056e-01 1.47853002e-01 -4.81567681e-01
6.87107444e-01 -7.06687793e-02 -4.07962739e-01 3.45036238e-02
-3.23604673e-01 2.01904356e-01 8.66523325e-01 -3.50898385e-01
-1.27158463e+00 -2.19943151e-01 2.68456452e-02 -8.56852904e-02
-5.71545243e-01 1.96142241e-01 2.40636095e-02 -2.64385015e-01
1.35494184e+00 3.45600277e-01 2.88860537e-02 -1.69123039e-01
2.18596712e-01 2.78601915e-01 6.69265985e-01 -2.71640003e-01
8.13251734e-01 6.58716381e-01 7.47829629e-03 -8.10077131e-01
-5.58053970e-01 1.28164425e-01 -6.79014683e-01 -1.62882298e-01
9.55083370e-01 -7.77762949e-01 -1.02492058e+00 4.92752582e-01
-1.11470020e+00 -4.98437077e-01 -6.54826522e-01 7.00134158e-01
-7.84052372e-01 -1.48731191e-02 -4.11662191e-01 -6.84497058e-01
3.90646964e-01 -8.70381653e-01 5.47315419e-01 4.27068442e-01
1.38787448e-01 -5.74822247e-01 2.46694610e-02 -1.35827288e-01
5.40537596e-01 1.89950839e-01 1.40340745e+00 1.69372916e-01
-6.68655097e-01 -1.05401911e-01 -5.73769033e-01 6.54196799e-01
1.31147832e-01 -1.26165420e-01 -1.39006329e+00 -2.72624075e-01
1.64746076e-01 -3.55712533e-01 1.05404687e+00 5.22478163e-01
1.08240616e+00 -4.27435488e-01 2.41975263e-01 7.39895165e-01
1.75658309e+00 3.70694429e-01 1.07850909e+00 1.32646859e-01
1.69512719e-01 8.77585351e-01 -1.81767698e-02 3.69409807e-02
2.35548526e-01 1.42857939e-01 4.28955555e-01 -3.16726714e-01
-6.31397426e-01 -2.78445989e-01 -1.14858877e-02 2.34071538e-01
-4.97094661e-01 3.13141793e-02 -8.67221117e-01 5.67276180e-01
-1.12181604e+00 -1.15878892e+00 2.50675172e-01 2.23309875e+00
9.37264681e-01 3.21057677e-01 9.37741809e-03 3.75960469e-01
4.75621372e-01 -7.77963102e-02 -8.35447609e-01 -7.01083362e-01
-3.11880171e-01 2.63828933e-01 1.90159544e-01 5.50690711e-01
-3.85391206e-01 6.24690890e-01 7.35856009e+00 -3.21371928e-02
-1.06209087e+00 -3.63866240e-01 5.77221334e-01 -1.36139646e-01
-3.55126143e-01 1.09098926e-01 4.52665128e-02 2.07031369e-01
9.17396605e-01 -4.13236618e-02 1.18339717e+00 4.85478640e-01
-6.01004884e-02 -5.34730732e-01 -1.27155149e+00 1.00875509e+00
4.76966314e-02 -9.98775840e-01 2.08724335e-01 4.69679497e-02
6.20780647e-01 -2.21094817e-01 4.87345308e-01 3.29467580e-02
1.60930514e-01 -1.43630576e+00 5.14422119e-01 1.02464485e+00
9.78667974e-01 -5.03605068e-01 3.73066455e-01 5.81150651e-01
-5.30557752e-01 -3.71586114e-01 -8.94146085e-01 -6.10642672e-01
-3.21390837e-01 3.54909420e-01 -6.07136488e-01 -8.82521048e-02
8.02357316e-01 1.34974748e-01 -9.15994287e-01 1.29731178e+00
1.22944964e-02 -5.73055632e-03 -7.86708966e-02 2.51228601e-01
-7.83829838e-02 -5.36357760e-02 1.60421729e-01 5.50089002e-01
4.89269942e-01 5.36813200e-01 -5.74623644e-01 1.10533881e+00
2.23255157e-02 -3.91387016e-01 -9.25754786e-01 -1.58222336e-02
1.91174403e-01 8.36288452e-01 -2.76286781e-01 -1.81520239e-01
-6.08153164e-01 1.03080809e+00 3.72956544e-01 9.36194837e-01
-7.78896809e-02 -6.65869236e-01 6.91018462e-01 3.35532874e-01
9.20328796e-02 -3.62766951e-01 -2.19039053e-01 -1.12132931e+00
-9.88517608e-03 -8.21825862e-01 -2.01748341e-01 -1.78421628e+00
-8.66650522e-01 8.61942768e-01 -2.46429935e-01 -8.24162424e-01
-4.08585757e-01 -1.03349650e+00 -5.31455994e-01 1.23123431e+00
-1.52128422e+00 -7.48726010e-01 -5.76864719e-01 9.07188475e-01
3.03245068e-01 -9.20991600e-02 9.06718671e-01 -1.19141832e-01
1.71203643e-01 2.43007049e-01 -1.19969293e-01 7.03675076e-02
6.22741401e-01 -1.42091286e+00 4.08233523e-01 9.02011991e-01
2.78348684e-01 7.18699336e-01 7.47797310e-01 -1.94042504e-01
-1.22709882e+00 -5.48050463e-01 7.43185937e-01 -4.01762813e-01
1.54961675e-01 -2.31032506e-01 -9.99948025e-01 6.36156380e-01
4.46281523e-01 2.30214298e-01 3.08371812e-01 -1.73156500e-01
-7.00651526e-01 -2.29138032e-01 -1.37738431e+00 8.16623390e-01
9.17264521e-01 -9.15259778e-01 -8.87746572e-01 3.48327905e-02
5.42066872e-01 -7.01863319e-02 -4.90290225e-01 -8.63540322e-02
8.46261799e-01 -1.63023114e+00 8.28185201e-01 -6.72304213e-01
5.14456928e-01 -4.45249647e-01 -1.08435974e-01 -1.55700338e+00
-6.23833656e-01 -1.36507392e-01 2.50240564e-01 5.31352460e-01
3.06086123e-01 -9.17499125e-01 4.10696268e-01 5.28706849e-01
-5.61025850e-02 -2.84801036e-01 -6.20667100e-01 -4.20808673e-01
1.31097510e-01 4.98456508e-02 3.90654355e-01 5.28722882e-01
-2.97803611e-01 3.00317947e-02 1.02565594e-01 4.15634438e-02
5.71069479e-01 1.31439209e-01 3.83042246e-01 -1.20311749e+00
-4.18502390e-01 -3.97155643e-01 -5.21177173e-01 -7.28124857e-01
-7.61960447e-02 -5.50259471e-01 -4.85774316e-02 -1.69312179e+00
3.50371227e-02 1.55592069e-01 -1.07552215e-01 3.77918035e-01
1.15362793e-01 4.71902817e-01 3.25252265e-01 2.11421400e-01
1.75561503e-01 2.85417251e-02 1.82886624e+00 9.68988463e-02
-6.92251790e-03 -1.33799344e-01 -1.11096764e+00 8.97089601e-01
8.25079978e-01 5.28103765e-03 -4.70451713e-01 -8.64636838e-01
5.36240816e-01 -5.77549599e-02 8.21128070e-01 -1.26506197e+00
2.75575131e-01 -1.38446733e-01 1.23280835e+00 -1.59226343e-01
3.34963143e-01 -8.60346019e-01 2.77419500e-02 4.06351179e-01
-4.31486398e-01 5.36050349e-02 2.68996865e-01 4.66163784e-01
-1.97483256e-01 -9.50094014e-02 1.09207189e+00 -6.34124279e-01
-7.33674884e-01 -8.09246749e-02 -5.80938935e-01 2.46009822e-05
6.31404042e-01 -5.46723485e-01 -7.16351211e-01 -4.96669710e-01
-8.69178236e-01 -5.57322741e-01 8.88196170e-01 -7.47150108e-02
8.76947939e-01 -1.16272664e+00 -5.59080541e-01 6.75986826e-01
-1.92446180e-03 -7.54775405e-02 3.27557325e-01 3.71966034e-01
-9.02885675e-01 4.75578040e-01 -9.98932958e-01 -1.50267094e-01
-4.09945041e-01 9.63226914e-01 7.35197067e-01 7.05287158e-01
-4.41232204e-01 7.63826787e-01 7.86141038e-01 2.23318473e-01
-1.52440831e-01 -5.10459304e-01 -1.88327604e-03 -4.08863783e-01
6.67843163e-01 -4.92976122e-02 -7.08758235e-02 -4.14219946e-01
-1.92184765e-02 7.37991929e-01 1.63496643e-01 -2.11463720e-01
1.08876932e+00 -1.47954464e-01 -5.90223782e-02 3.93191338e-01
9.73007143e-01 -2.96678431e-02 -1.75118971e+00 -6.94127008e-02
-6.53420448e-01 -7.66081691e-01 2.92073749e-02 -1.05145538e+00
-8.95440102e-01 1.39990473e+00 4.94650424e-01 4.23020929e-01
1.64001513e+00 -1.15497708e-01 2.30750442e-01 7.16140926e-01
3.43734145e-01 -7.65951455e-01 2.63444930e-01 2.38817304e-01
1.07635629e+00 -1.08686280e+00 -2.20434755e-01 1.51194215e-01
-5.03125131e-01 1.11270773e+00 6.57705784e-01 -5.91587663e-01
3.23093027e-01 -1.08624890e-01 8.76376685e-03 -1.77121207e-01
-9.04989839e-01 -5.01799047e-01 3.79108697e-01 1.11533594e+00
3.18881512e-01 -2.52972066e-01 3.67314368e-01 3.28943171e-02
-5.15674531e-01 -3.41063350e-01 8.39022100e-01 4.97619957e-01
-8.88061643e-01 -3.60549390e-01 -4.34594333e-01 3.81753802e-01
-1.91516399e-01 -1.72052354e-01 -5.65553963e-01 7.87836790e-01
3.33609939e-01 7.73927748e-01 3.98397416e-01 -4.15992364e-02
4.09287095e-01 7.96271116e-02 1.05986106e+00 -7.33154476e-01
-1.07319385e-01 -2.02810951e-02 -3.53937596e-01 -6.22410893e-01
-3.42902184e-01 -3.93002212e-01 -1.01870048e+00 -3.65871221e-01
2.86874563e-01 -4.18236047e-01 5.35044730e-01 6.94069684e-01
1.12808391e-01 5.33684909e-01 3.30775768e-01 -9.21753347e-01
-2.61463821e-01 -7.94089496e-01 -8.40151608e-01 4.75793719e-01
1.05174220e+00 -3.90949577e-01 -6.46197438e-01 2.35080838e-01] | [10.087371826171875, 2.2902441024780273] |
5969f3bd-28b7-47d5-959f-3b2d94dff5d3 | angular-luminance-for-material-segmentation | 2009.10825 | null | https://arxiv.org/abs/2009.10825v1 | https://arxiv.org/pdf/2009.10825v1.pdf | Angular Luminance for Material Segmentation | Moving cameras provide multiple intensity measurements per pixel, yet often semantic segmentation, material recognition, and object recognition do not utilize this information. With basic alignment over several frames of a moving camera sequence, a distribution of intensities over multiple angles is obtained. It is well known from prior work that luminance histograms and the statistics of natural images provide a strong material recognition cue. We utilize per-pixel {\it angular luminance distributions} as a key feature in discriminating the material of the surface. The angle-space sampling in a multiview satellite image sequence is an unstructured sampling of the underlying reflectance function of the material. For real-world materials there is significant intra-class variation that can be managed by building a angular luminance network (AngLNet). This network combines angular reflectance cues from multiple images with spatial cues as input to fully convolutional networks for material segmentation. We demonstrate the increased performance of AngLNet over prior state-of-the-art in material segmentation from satellite imagery. | ['Kristin Dana', 'Jia Xue', 'Matthew Purri'] | 2020-09-22 | null | null | null | null | ['material-recognition'] | ['computer-vision'] | [ 9.15877521e-01 -6.54851317e-01 -2.02240497e-01 -4.97061104e-01
-8.71389925e-01 -6.92195475e-01 4.79298949e-01 -1.42981127e-01
-4.56717581e-01 4.59833443e-01 -1.65847704e-01 3.08026858e-02
1.35518655e-01 -1.06110299e+00 -9.71355855e-01 -8.91750455e-01
2.55059540e-01 1.65409654e-01 4.15374577e-01 7.73168132e-02
4.08756047e-01 7.96502769e-01 -1.70768034e+00 5.68947792e-01
5.44396162e-01 1.34904301e+00 5.83393097e-01 9.52341616e-01
-3.06346834e-01 4.76512849e-01 -3.34293038e-01 4.39515412e-01
5.34475565e-01 -1.66295156e-01 -8.51947844e-01 8.37054729e-01
1.10052204e+00 -6.13833547e-01 -1.26404464e-01 1.10820496e+00
-3.22167538e-02 2.35061571e-01 9.31636512e-01 -8.11922491e-01
-4.16986495e-01 7.16803819e-02 -6.19524658e-01 2.13450983e-01
5.11203229e-01 2.66211897e-01 1.12760365e+00 -8.14447105e-01
6.20586157e-01 9.64061737e-01 6.92842364e-01 -4.04339433e-02
-1.40158617e+00 -8.75593200e-02 1.80042654e-01 1.71938494e-01
-1.39465368e+00 -2.43921399e-01 9.95378077e-01 -5.24648905e-01
7.92322755e-01 3.98578167e-01 8.99808466e-01 5.94859779e-01
3.99295278e-02 8.77766430e-01 1.36280262e+00 -3.40051711e-01
2.13560775e-01 -1.34363666e-01 -1.33316502e-01 4.85699117e-01
-3.41047086e-02 -1.28079697e-01 -5.83095968e-01 1.26050204e-01
1.07687449e+00 2.73503393e-01 -5.56347013e-01 -3.81391853e-01
-1.24491000e+00 3.21815968e-01 5.43507099e-01 -5.22799464e-03
-4.49811101e-01 2.06108689e-01 -1.49353102e-01 1.06272221e-01
6.97815478e-01 1.60814598e-01 -3.79383683e-01 1.64771080e-01
-1.30402303e+00 1.78590696e-02 5.59382200e-01 6.86208010e-01
1.33254707e+00 1.67811230e-01 3.79889697e-01 8.14547598e-01
1.56899735e-01 1.03481877e+00 1.75759140e-02 -1.37313783e+00
1.74626186e-01 5.30104578e-01 2.24331930e-01 -8.89120638e-01
-3.03975463e-01 1.00199126e-01 -5.04561782e-01 4.38950211e-01
6.95524037e-01 2.04902172e-01 -1.06955338e+00 1.32231200e+00
2.72739530e-01 -1.72768950e-01 -1.83928326e-01 8.90095115e-01
5.39056778e-01 7.23351359e-01 -4.53763694e-01 -2.76135094e-02
1.20263112e+00 -4.75439638e-01 -2.57679135e-01 -5.56478798e-01
-4.68933769e-03 -8.15076172e-01 1.03266037e+00 5.78462422e-01
-1.03775442e+00 -4.61024404e-01 -1.05140054e+00 -4.30342667e-02
-3.48278284e-01 7.35702515e-02 6.34367764e-01 3.53734314e-01
-9.99990106e-01 5.24546325e-01 -9.63115394e-01 -3.01949739e-01
5.27045012e-01 1.24367572e-01 -3.49505037e-01 -5.08515716e-01
-5.73604345e-01 5.65754950e-01 1.62116215e-01 2.15693444e-01
-7.85464048e-01 -7.48916924e-01 -8.03427756e-01 -4.09275353e-01
3.86718988e-01 -3.51592869e-01 9.43720460e-01 -1.44648445e+00
-1.29224503e+00 1.04488361e+00 -2.41230547e-01 5.65417707e-02
4.69516665e-01 -1.29633904e-01 -6.71602562e-02 7.29955912e-01
6.15867339e-02 7.19606221e-01 1.04446387e+00 -1.45950937e+00
-5.82686365e-01 -3.15894037e-01 4.12214361e-03 4.09780234e-01
1.63491011e-01 -3.00245941e-01 -4.03555095e-01 -3.63721579e-01
5.48965514e-01 -7.87788212e-01 -2.15355568e-02 4.32356387e-01
-4.36309636e-01 2.11138040e-01 8.03487062e-01 -6.84587061e-01
5.08166254e-01 -2.03865623e+00 -2.58501507e-02 1.14990585e-01
-1.45496696e-01 -1.84135526e-01 -1.52866930e-01 6.33716956e-02
1.20003775e-01 -7.22284317e-02 -5.37916183e-01 2.38759890e-01
-4.10231918e-01 8.12267736e-02 -2.63213050e-02 8.83949101e-01
2.36958802e-01 5.57744265e-01 -6.39034390e-01 -3.52094471e-01
6.56069696e-01 6.23281181e-01 -3.79129797e-01 1.80935115e-01
-5.62546551e-01 4.38998878e-01 -1.93463340e-01 9.66260433e-01
9.36888576e-01 -2.62142539e-01 -6.10187314e-02 -6.57046854e-01
-2.48376235e-01 1.19670123e-01 -1.26642656e+00 1.66162527e+00
-3.93295914e-01 1.08299184e+00 4.01722163e-01 -8.51043403e-01
6.97884023e-01 -4.46803235e-02 8.07644010e-01 -5.71931064e-01
-3.10568586e-02 6.75866678e-02 -3.34497124e-01 -3.91594052e-01
7.51724541e-01 7.25057200e-02 2.98242897e-01 3.66164595e-01
-2.72292465e-01 -6.06409192e-01 7.30298534e-02 -1.32924244e-01
7.35589206e-01 2.59469658e-01 -2.00015362e-02 -3.21179956e-01
3.55317563e-01 1.36891663e-01 4.05929953e-01 7.36477017e-01
8.26443061e-02 1.04545355e+00 1.12396859e-01 -4.87671584e-01
-1.22005594e+00 -1.30357015e+00 -4.51044291e-01 9.43300962e-01
2.91852117e-01 1.91784546e-01 -6.87086523e-01 -1.08452179e-01
-2.50928197e-02 3.50687563e-01 -4.96337086e-01 4.13704544e-01
-5.09267211e-01 -8.43172312e-01 -4.49886434e-02 5.59790909e-01
8.60861063e-01 -9.17234004e-01 -6.78664446e-01 5.01398966e-02
-3.80790800e-01 -1.52899265e+00 -3.07134897e-01 3.74354333e-01
-8.36235404e-01 -1.17020273e+00 -5.48044801e-01 -5.28027713e-01
7.40150154e-01 7.79765308e-01 1.44750702e+00 -1.12770572e-01
-6.79609835e-01 8.97384405e-01 -1.32412270e-01 2.93703657e-02
-1.57252312e-01 -1.15748174e-01 -2.74900049e-01 1.60237402e-01
2.99326450e-01 -5.23926735e-01 -9.00465608e-01 3.65263224e-01
-1.11871326e+00 2.84429878e-01 4.23329532e-01 5.06030560e-01
8.72350574e-01 8.18497688e-02 -2.78426021e-01 -7.04800963e-01
-2.40808189e-01 -2.23334759e-01 -6.42455935e-01 1.27144381e-01
-4.47661951e-02 -2.28921190e-01 2.85052925e-01 -2.21878827e-01
-1.04860413e+00 2.51742870e-01 1.96632192e-01 -1.32903323e-01
-6.21273458e-01 2.47404203e-01 -2.50134200e-01 -1.17199853e-01
3.94461632e-01 3.76286745e-01 -7.79125839e-02 -2.24744320e-01
3.40664208e-01 6.42035782e-01 7.44935334e-01 -6.30377352e-01
4.61955488e-01 1.07894742e+00 1.43702894e-01 -1.63369346e+00
-1.00972617e+00 -7.52342403e-01 -8.41987252e-01 -4.78986561e-01
1.24222946e+00 -9.92030740e-01 -6.58456504e-01 7.53467202e-01
-8.76592278e-01 -6.90870047e-01 -1.83812752e-01 2.93011129e-01
-6.84396982e-01 2.92379171e-01 -5.29473782e-01 -9.00520802e-01
-5.98514304e-02 -1.19959784e+00 1.41715860e+00 2.89604902e-01
1.52381197e-01 -1.09509492e+00 -5.45752227e-01 5.67543089e-01
3.72031629e-01 2.56097585e-01 5.80447137e-01 1.87221959e-01
-1.13101387e+00 -1.43371508e-01 -3.90676200e-01 6.59881890e-01
4.49868023e-01 2.54809529e-01 -1.11728847e+00 1.10486113e-02
-1.05466820e-01 -2.50706732e-01 1.09043324e+00 8.36504459e-01
1.21362138e+00 5.98354675e-02 4.39572372e-02 8.60946178e-01
1.77244997e+00 -6.13029413e-02 6.43264830e-01 5.95535994e-01
1.09897459e+00 5.18197536e-01 4.89002973e-01 2.20936030e-01
2.22770944e-01 6.50795698e-01 5.43084621e-01 -2.96900839e-01
-1.40116751e-01 2.93019325e-01 4.77998227e-01 4.76206988e-01
-2.07445130e-01 -1.88619241e-01 -9.12841797e-01 5.10421813e-01
-1.14469159e+00 -1.08085179e+00 -2.20428184e-01 2.31150866e+00
8.68114054e-01 -3.39798667e-02 -7.84926340e-02 1.10607460e-01
6.29558444e-01 6.12163067e-01 -7.48712838e-01 -4.37737955e-03
-6.07616961e-01 -8.32755938e-02 1.20084453e+00 6.16666675e-01
-1.34105206e+00 7.80323267e-01 6.99024010e+00 7.54204273e-01
-1.26389122e+00 -5.05992591e-01 8.02159190e-01 9.46380273e-02
-4.18295234e-01 -1.69850484e-01 -7.55482912e-01 1.61542267e-01
4.38482732e-01 4.03480649e-01 7.24880219e-01 6.50590837e-01
2.45544195e-01 -7.40478516e-01 -9.78528380e-01 1.07810879e+00
2.72322804e-01 -1.15268159e+00 -1.20255321e-01 3.31004225e-02
9.55175161e-01 4.21811789e-01 2.65780389e-01 -6.78061604e-01
2.77082175e-01 -8.33743930e-01 7.06288278e-01 7.42242634e-01
9.54992056e-01 -3.84940803e-01 3.39197218e-01 1.00571103e-01
-1.42818272e+00 1.63188145e-01 -2.98278928e-01 2.64963955e-01
1.87633693e-01 7.17125237e-01 -6.28016233e-01 3.01218480e-01
7.91140616e-01 1.08718777e+00 -6.83175623e-01 7.26756930e-01
-1.12954520e-01 4.93133903e-01 -6.33348465e-01 4.18473363e-01
2.72905231e-01 -7.46455133e-01 4.17706400e-01 1.10402524e+00
1.42877415e-01 -9.62558091e-02 5.23506880e-01 7.85115480e-01
-2.20056660e-02 -2.44726971e-01 -5.85222900e-01 -1.02345102e-01
2.26936102e-01 1.31099272e+00 -1.17510927e+00 -3.99778962e-01
-6.97991669e-01 1.06740844e+00 -1.85612828e-01 5.97853661e-01
-2.81176209e-01 1.49558574e-01 7.04246461e-01 2.38160074e-01
5.38658619e-01 -6.68298066e-01 -3.61324102e-01 -1.02299416e+00
5.08942045e-02 -8.32263827e-01 -2.47715265e-02 -1.20589268e+00
-1.41141999e+00 1.13970026e-01 -1.67743228e-02 -1.35813916e+00
8.50107223e-02 -9.66111243e-01 -3.76186401e-01 8.11949611e-01
-1.50381982e+00 -1.17224443e+00 -6.12131059e-01 4.25287247e-01
8.68938625e-01 2.23612756e-01 4.63410974e-01 -1.20544687e-01
-2.06426352e-01 -3.86932760e-01 3.62316579e-01 2.00197682e-01
4.64652538e-01 -1.50431013e+00 7.59720653e-02 9.71693933e-01
1.57890201e-01 1.82879373e-01 6.74512267e-01 -5.20347238e-01
-1.66798437e+00 -9.71446574e-01 -3.52214426e-02 -4.75835860e-01
5.36903322e-01 -1.42571002e-01 -8.71756792e-01 5.96428633e-01
1.76467627e-01 1.97665885e-01 5.11887550e-01 -3.15636307e-01
-3.19306821e-01 -1.61547065e-01 -9.25621688e-01 3.71375024e-01
6.90916240e-01 -8.84984672e-01 -1.76502407e-01 3.78450274e-01
3.23623627e-01 -3.83628339e-01 -8.92131984e-01 2.13438913e-01
7.74368882e-01 -1.10777211e+00 1.31075907e+00 -1.73646715e-02
4.86007839e-01 -4.51423645e-01 -8.23218942e-01 -1.17162979e+00
1.33764774e-01 -1.04878411e-01 5.08746147e-01 8.94020021e-01
1.49332970e-01 -3.32687616e-01 1.01868284e+00 7.15702653e-01
2.93249786e-02 -2.04200014e-01 -6.62395656e-01 -6.02756560e-01
-5.99340834e-02 -7.85194576e-01 1.95763335e-01 7.28161871e-01
-7.98275471e-01 -2.36503072e-02 -1.46203175e-01 2.94715106e-01
1.02296352e+00 4.53343958e-01 6.09782696e-01 -1.21924078e+00
-1.11699872e-01 -4.36410189e-01 -4.56046551e-01 -1.33225155e+00
3.33797857e-02 -7.42533624e-01 3.62784058e-01 -1.65415835e+00
1.80367768e-01 -3.71308774e-01 3.09186950e-02 1.37712017e-01
-5.74603155e-02 6.27985954e-01 4.43116017e-02 1.69953376e-01
-6.21818423e-01 2.44213298e-01 1.18545043e+00 -4.56211925e-01
-7.30175152e-02 -1.96299404e-01 -1.10728443e-01 9.75020111e-01
7.86709249e-01 -5.86453564e-02 -1.39026552e-01 -7.09133685e-01
3.42209816e-01 -1.23558193e-03 5.08667052e-01 -9.34586704e-01
-4.82924059e-02 -5.23725033e-01 7.62935877e-01 -9.60243404e-01
5.66045165e-01 -1.05381966e+00 1.48810416e-01 -1.28791347e-01
-1.93750679e-01 -1.63969517e-01 -9.83798802e-02 6.91614628e-01
-7.41518736e-02 -5.46703339e-02 9.32996392e-01 -5.53463399e-01
-9.45858896e-01 3.36734235e-01 -5.65478861e-01 7.96246454e-02
3.73928785e-01 -5.58356166e-01 -3.76717687e-01 -2.32712373e-01
-4.62539971e-01 -2.60236710e-01 9.38659370e-01 -1.73325762e-02
6.26202703e-01 -1.11173332e+00 -4.51747358e-01 3.36557060e-01
-4.12260927e-02 3.27945620e-01 3.43674898e-01 9.10206258e-01
-1.07468677e+00 6.86261943e-03 -2.55610257e-01 -1.24268115e+00
-1.18593180e+00 -7.18753263e-02 5.85798204e-01 4.26246822e-01
-7.01966345e-01 6.76118076e-01 2.44016856e-01 -1.48379862e-01
-1.11068167e-01 -4.46489275e-01 7.25305080e-02 1.37698874e-01
4.14595038e-01 4.46628779e-01 6.79709539e-02 -8.09906185e-01
-2.70082325e-01 1.06556177e+00 1.96501389e-01 -3.27956766e-01
1.57151580e+00 -4.28110957e-01 -4.06141162e-01 9.64145482e-01
1.29566312e+00 -1.28681272e-01 -1.81236303e+00 -3.60516071e-01
-2.84832597e-01 -8.25927973e-01 5.51966488e-01 -4.21344340e-01
-1.25508368e+00 8.44806790e-01 6.47141576e-01 4.34454083e-01
1.22921443e+00 -1.43690571e-01 4.86619681e-01 4.31213260e-01
2.41131470e-01 -1.33924651e+00 6.86085895e-02 3.07094663e-01
6.30321860e-01 -1.57787836e+00 4.53251868e-01 -4.93380368e-01
-5.02672911e-01 1.42575502e+00 3.30910802e-01 -1.34394169e-01
6.60182834e-01 2.92366236e-01 2.39972636e-01 -2.45035544e-01
-3.64534557e-01 -4.39437002e-01 4.08803642e-01 6.08578861e-01
3.89192075e-01 1.36391252e-01 4.52772081e-01 -4.03298944e-01
-1.52863394e-02 -3.70010495e-01 4.96227443e-01 9.23449337e-01
-8.61622930e-01 -6.95824921e-01 -6.30175591e-01 5.72027326e-01
-3.40215147e-01 -1.79653212e-01 -1.30380616e-01 5.92234790e-01
-2.99767964e-02 9.43341494e-01 6.20502770e-01 2.27217935e-02
-1.16594933e-01 -1.18642122e-01 6.48858011e-01 -4.35225308e-01
-2.92282104e-02 3.38970989e-01 1.81780069e-03 -7.47127175e-01
-1.01250172e+00 -1.03613079e+00 -1.03424776e+00 -4.70186137e-02
-1.11212194e-01 -4.42533046e-01 9.37048376e-01 9.55583036e-01
-4.23008680e-01 4.51621592e-01 7.92991519e-01 -1.48141646e+00
1.72954232e-01 -7.09848881e-01 -9.11509871e-01 4.19996142e-01
6.23002350e-01 -5.21483600e-01 -8.42506707e-01 6.02080762e-01] | [9.663843154907227, -2.7515478134155273] |
627dfbab-76e5-4dc8-bd71-d852ff9032d2 | symmetric-optimistic-natural-policy-gradient | 2210.12812 | null | https://arxiv.org/abs/2210.12812v2 | https://arxiv.org/pdf/2210.12812v2.pdf | Symmetric (Optimistic) Natural Policy Gradient for Multi-agent Learning with Parameter Convergence | Multi-agent interactions are increasingly important in the context of reinforcement learning, and the theoretical foundations of policy gradient methods have attracted surging research interest. We investigate the global convergence of natural policy gradient (NPG) algorithms in multi-agent learning. We first show that vanilla NPG may not have parameter convergence, i.e., the convergence of the vector that parameterizes the policy, even when the costs are regularized (which enabled strong convergence guarantees in the policy space in the literature). This non-convergence of parameters leads to stability issues in learning, which becomes especially relevant in the function approximation setting, where we can only operate on low-dimensional parameters, instead of the high-dimensional policy. We then propose variants of the NPG algorithm, for several standard multi-agent learning scenarios: two-player zero-sum matrix and Markov games, and multi-player monotone games, with global last-iterate parameter convergence guarantees. We also generalize the results to certain function approximation settings. Note that in our algorithms, the agents take symmetric roles. Our results might also be of independent interest for solving nonconvex-nonconcave minimax optimization problems with certain structures. Simulations are also provided to corroborate our theoretical findings. | ['Asuman Ozdaglar', 'Kaiqing Zhang', 'Sarath Pattathil'] | 2022-10-23 | null | null | null | null | ['policy-gradient-methods'] | ['methodology'] | [-2.21561775e-01 1.43182620e-01 -5.07198572e-01 3.05954844e-01
-7.78234065e-01 -6.34778202e-01 1.87189817e-01 3.38234067e-01
-9.29574728e-01 1.41672361e+00 8.65549669e-02 -4.25108701e-01
-7.43854821e-01 -5.52036583e-01 -8.23217392e-01 -1.02228653e+00
-5.43719590e-01 5.03521025e-01 -2.70636708e-01 -4.60757732e-01
2.87468493e-01 1.59850463e-01 -1.05869460e+00 -3.84471864e-01
1.24991679e+00 7.60092139e-01 1.73003867e-01 8.12024117e-01
3.29308599e-01 6.08578265e-01 -4.38886225e-01 -1.65128231e-01
6.49412572e-01 -3.42623949e-01 -6.57288492e-01 2.69061208e-01
1.29022121e-01 -4.51851010e-01 2.21853685e-02 1.39579844e+00
6.69719994e-01 5.97595870e-01 4.82571751e-01 -1.43378246e+00
-3.10365170e-01 6.34664893e-01 -1.04053807e+00 1.33331433e-01
2.05760878e-02 1.51931211e-01 1.30046904e+00 -4.85246122e-01
2.93620080e-01 1.47577345e+00 3.78072500e-01 8.14039290e-01
-1.07613206e+00 -2.92050123e-01 5.83551049e-01 1.62487715e-01
-8.62550914e-01 5.20674884e-03 2.35873684e-01 -3.35606992e-01
5.39029896e-01 2.67478883e-01 7.25897729e-01 6.79380238e-01
2.00552512e-02 1.07895982e+00 1.13143051e+00 -3.57581109e-01
4.61913079e-01 -5.75663671e-02 -2.16810182e-01 8.81033480e-01
5.22039056e-01 2.35987306e-01 -2.07517326e-01 -4.68594968e-01
8.01571667e-01 -1.56118155e-01 -5.27978539e-01 -6.47064447e-01
-9.98142958e-01 1.41819406e+00 1.17936842e-01 -3.66845950e-02
-7.80520976e-01 3.33988309e-01 2.51763403e-01 6.13781750e-01
5.23103774e-01 6.32812560e-01 -4.11042124e-01 -3.52511853e-01
-4.60345000e-01 7.00937510e-01 1.15275884e+00 7.59862304e-01
5.22971928e-01 2.62813300e-01 -9.97897312e-02 7.79893875e-01
2.69878000e-01 5.58115959e-01 1.89038903e-01 -1.28519237e+00
8.32741678e-01 -1.99781600e-02 6.90442979e-01 -7.31943905e-01
-4.73402709e-01 -4.81353194e-01 -7.42482424e-01 4.89428341e-01
8.08272839e-01 -9.64162707e-01 -2.21968845e-01 2.02138948e+00
5.39595604e-01 1.08742863e-01 1.65264711e-01 1.24578786e+00
-2.30609193e-01 6.57693386e-01 -2.37341493e-01 -8.84842753e-01
9.08700049e-01 -9.55354393e-01 -6.20696783e-01 -1.94406688e-01
6.54123008e-01 -3.66235167e-01 1.15076685e+00 3.48918080e-01
-1.31694353e+00 1.79512933e-01 -8.56219828e-01 6.55907333e-01
1.45074189e-01 -3.20686698e-01 5.35997510e-01 4.50123191e-01
-1.14737630e+00 6.10340714e-01 -5.55916965e-01 -1.18809976e-01
2.58928835e-01 5.49361348e-01 2.18241081e-01 1.22591734e-01
-1.04259670e+00 8.65902126e-01 4.33047473e-01 1.17273971e-01
-9.51471567e-01 -8.33869517e-01 -5.27025998e-01 6.07440695e-02
1.16021919e+00 -6.27277136e-01 1.61383832e+00 -1.00885665e+00
-1.82511222e+00 1.16766796e-01 2.43765816e-01 -4.28145140e-01
9.95310783e-01 -1.33870393e-01 2.66261816e-01 1.68185532e-01
1.10030167e-01 4.12408635e-02 9.83616590e-01 -1.14593315e+00
-1.01611221e+00 -2.08156496e-01 6.80094004e-01 8.33310068e-01
-4.86500740e-01 -2.89721429e-01 2.17261538e-01 -3.49213302e-01
-6.60941064e-01 -1.04797339e+00 -8.39925230e-01 -4.50145416e-02
-1.74663499e-01 -2.22207665e-01 5.62382221e-01 -2.44238138e-01
1.03894675e+00 -1.74700868e+00 5.20947158e-01 3.55096310e-01
1.70799851e-01 1.45052284e-01 -2.69888133e-01 5.16867220e-01
2.28015229e-01 1.10018373e-01 -2.22463205e-01 -9.61477682e-02
2.83609360e-01 3.25926811e-01 -3.23026597e-01 8.76766086e-01
-2.98197448e-01 7.53966570e-01 -1.14272475e+00 -2.35865131e-01
-7.62677714e-02 -2.49215096e-01 -9.60289896e-01 -7.04358369e-02
-4.08487648e-01 2.93054938e-01 -9.35885310e-01 1.78202197e-01
4.22865957e-01 -3.67000312e-01 2.76240259e-01 4.05996829e-01
-2.94350535e-01 -3.86875868e-01 -1.42609358e+00 1.18745458e+00
-6.20642483e-01 2.10441440e-01 8.88459563e-01 -1.44734502e+00
1.21476680e-01 2.64306188e-01 9.76781726e-01 -4.11131889e-01
-8.33971575e-02 2.10428312e-01 2.21142665e-01 -5.01312613e-01
5.16327739e-01 -2.90616423e-01 5.85135184e-02 6.64896846e-01
-4.02953297e-01 -9.13347676e-02 6.62268698e-01 1.30867749e-01
6.84542120e-01 -3.82071525e-01 3.97993386e-01 -6.94398284e-01
5.08289576e-01 -1.39860198e-01 5.71372390e-01 1.10995018e+00
-2.41310209e-01 -1.42789766e-01 8.43604088e-01 7.80104324e-02
-1.08356690e+00 -6.39192224e-01 1.45384327e-01 1.33226681e+00
2.15971574e-01 -4.79274020e-02 -5.17216206e-01 -5.80523670e-01
5.08673251e-01 3.93138438e-01 -6.16747797e-01 8.04604888e-02
-6.88572049e-01 -1.25784719e+00 -5.87007264e-03 3.94359324e-03
4.66142118e-01 -7.81526089e-01 -6.05004251e-01 5.97784340e-01
9.54277366e-02 -8.08139980e-01 -8.26031983e-01 -3.19614932e-02
-7.26494014e-01 -1.31122661e+00 -1.11258566e+00 -5.19684851e-01
5.63313901e-01 -2.23548408e-03 7.21403897e-01 -5.28777540e-02
2.01291144e-01 1.05921292e+00 -5.63583523e-02 -4.87901241e-01
-3.47728521e-01 2.25792341e-02 4.32002634e-01 1.65071502e-01
-4.46918160e-01 -3.99039567e-01 -6.69859231e-01 2.88797528e-01
-9.21060562e-01 -1.23970233e-01 3.26748550e-01 9.91773188e-01
3.40399921e-01 -8.35004300e-02 9.02935863e-01 -7.24940777e-01
1.51079321e+00 -6.59771085e-01 -1.17375648e+00 3.31495076e-01
-6.67874038e-01 1.90887019e-01 9.51003790e-01 -7.75815308e-01
-7.41723120e-01 -3.25172484e-01 3.07892323e-01 -3.06589305e-01
5.50035357e-01 7.19489872e-01 2.27136135e-01 -3.17680597e-01
4.52706069e-01 1.37535825e-01 4.57897753e-01 -9.43895653e-02
4.49984312e-01 3.90848517e-01 -2.65531093e-02 -1.13208020e+00
6.26737297e-01 3.46765429e-01 2.21347943e-01 -8.70144129e-01
-6.51467919e-01 -2.49066621e-01 2.14550465e-01 -3.03779155e-01
4.46650088e-01 -6.59342647e-01 -1.58817148e+00 3.28864247e-01
-8.78309071e-01 -7.16154635e-01 -4.81462449e-01 7.01222420e-01
-1.21501708e+00 4.10093904e-01 -7.28733182e-01 -1.26980495e+00
-1.22404888e-01 -1.08418298e+00 3.39630187e-01 4.64276910e-01
5.33117294e-01 -1.34447694e+00 2.97924250e-01 -7.94231147e-02
3.96816015e-01 8.76622126e-02 7.69007564e-01 -3.69042695e-01
-2.89607346e-01 2.92962313e-01 1.36619702e-01 1.03781603e-01
1.88529324e-02 -3.76778066e-01 -2.08328187e-01 -8.51882219e-01
1.61025375e-01 -5.90208888e-01 5.51467299e-01 8.57455552e-01
9.05280769e-01 -9.44432259e-01 -1.09847791e-01 3.41298878e-01
1.57391727e+00 1.92299485e-01 -1.45774081e-01 4.87181813e-01
4.73816335e-01 4.96175826e-01 6.41535640e-01 1.10013521e+00
3.47903252e-01 5.37086070e-01 6.67215109e-01 2.15159561e-02
7.62247801e-01 2.40065195e-02 4.27849114e-01 5.34107029e-01
-2.61198789e-01 -2.41955936e-01 -6.15330994e-01 3.50557029e-01
-2.46119261e+00 -1.00376225e+00 2.98174053e-01 2.33935523e+00
1.07370639e+00 -3.25788438e-01 6.90798223e-01 -2.61954397e-01
9.81213212e-01 7.65270144e-02 -1.19604075e+00 -5.01837194e-01
-2.23248810e-01 -3.76305319e-02 9.32953835e-01 9.11064625e-01
-9.51959968e-01 6.46470904e-01 6.29710436e+00 9.42764938e-01
-9.11144018e-01 1.73769712e-01 6.28053248e-01 -4.55885559e-01
-6.30791858e-03 -3.59608948e-01 -5.29893458e-01 5.17524183e-01
5.97527206e-01 -6.75760508e-01 9.95952070e-01 8.45094085e-01
7.60686636e-01 -1.91935584e-01 -7.74638772e-01 1.02074194e+00
-4.24472362e-01 -1.20722508e+00 -5.16849875e-01 5.25651872e-01
1.13353741e+00 -8.03592801e-02 1.59116209e-01 3.90208721e-01
7.39305556e-01 -7.81464100e-01 5.03744066e-01 9.37809199e-02
4.77663636e-01 -1.15766895e+00 2.62056530e-01 5.80041289e-01
-9.49286878e-01 -7.28137076e-01 -5.23254454e-01 -3.34842682e-01
2.85031796e-01 2.61960357e-01 -5.23687840e-01 3.79446983e-01
1.84316739e-01 4.90978450e-01 2.08526909e-01 1.39435935e+00
5.20696528e-02 4.54585761e-01 -5.58602512e-01 -5.87419868e-01
7.44730353e-01 -7.35142589e-01 8.31580460e-01 6.83689058e-01
1.54486418e-01 1.68545529e-01 7.02466607e-01 4.41487253e-01
-5.40211052e-02 4.96235222e-01 -3.52845103e-01 -5.85708506e-02
1.92025244e-01 1.11028111e+00 -5.34118176e-01 -1.98579907e-01
-5.09637833e-01 7.12418616e-01 5.78466773e-01 7.16284156e-01
-6.81211293e-01 -1.56245306e-01 1.21027291e+00 -3.21350574e-01
3.47203434e-01 -2.95057625e-01 1.72929838e-01 -1.20532298e+00
2.11333364e-01 -1.01172888e+00 6.23635232e-01 2.23702967e-01
-1.43555570e+00 -8.62345919e-02 1.23684458e-01 -9.03375030e-01
-6.99727416e-01 -4.94860500e-01 -5.07868528e-01 4.26317990e-01
-1.46115255e+00 -3.89313608e-01 5.70678115e-01 7.67789781e-01
5.18178940e-01 -2.05666229e-01 3.88481468e-01 1.92759246e-01
-6.75610900e-01 4.73589838e-01 8.87855351e-01 -2.96085417e-01
2.30679303e-01 -1.58766496e+00 -4.58492309e-01 6.15654051e-01
-2.19025314e-01 2.00027406e-01 9.51015890e-01 -3.55924696e-01
-1.81930208e+00 -7.44097650e-01 -2.66774118e-01 3.66360217e-01
1.14777565e+00 -1.67156190e-01 -5.26550174e-01 3.79128903e-01
1.47960767e-01 -1.24626961e-02 2.10531250e-01 1.10806823e-01
4.06293780e-01 7.75915757e-02 -1.03985763e+00 8.56404006e-01
8.32438827e-01 -3.37624960e-02 9.05522034e-02 7.19476819e-01
4.47135925e-01 -5.52537560e-01 -6.30292952e-01 -3.32485102e-02
3.12820017e-01 -4.19264227e-01 7.85280108e-01 -1.13525224e+00
9.69653353e-02 -4.04641731e-03 -4.27536331e-02 -1.95645118e+00
-2.30973065e-01 -1.21875679e+00 -4.96221513e-01 5.93862891e-01
4.46146429e-01 -9.11205471e-01 9.35553372e-01 5.68051398e-01
8.39659572e-02 -1.10060918e+00 -1.13564849e+00 -1.01883793e+00
6.55295551e-01 -9.82340574e-02 1.72515035e-01 8.61345947e-01
4.14080083e-01 2.72415996e-01 -8.62154007e-01 1.25787050e-01
8.45969081e-01 8.12063441e-02 5.50726533e-01 -7.66157568e-01
-8.08088899e-01 -8.37291241e-01 1.14846878e-01 -1.33950484e+00
4.37105238e-01 -6.01406395e-01 3.83892246e-02 -1.28601146e+00
7.00512454e-02 -5.00452340e-01 -7.94011950e-02 1.38396949e-01
-3.63916755e-01 -4.41539347e-01 6.51530802e-01 6.58322871e-02
-8.43141496e-01 9.64861333e-01 1.69045246e+00 -2.69355357e-01
-4.76495653e-01 5.73322475e-01 -7.44051754e-01 5.58475912e-01
9.66644883e-01 -2.77734637e-01 -7.52877891e-01 -2.95202821e-01
2.45958850e-01 3.75768512e-01 4.12478335e-02 -5.70377290e-01
3.20091486e-01 -9.68862772e-01 -2.54230917e-01 1.76154539e-01
2.45321751e-01 -6.39338851e-01 -4.25349355e-01 8.33150327e-01
-5.95263958e-01 1.72453940e-01 -1.54552668e-01 8.27816725e-01
1.95036754e-01 -6.39399827e-01 9.20050085e-01 -2.01038271e-01
-2.67871857e-01 7.29310691e-01 -6.50554180e-01 7.11345136e-01
1.13976836e+00 1.98205709e-01 -1.46829233e-01 -9.48999166e-01
-6.69370770e-01 8.09827924e-01 3.39083783e-02 -4.90148515e-02
5.28258443e-01 -1.23467076e+00 -9.13308263e-01 -4.03275311e-01
-4.55776006e-01 -2.04276547e-01 -3.88097242e-02 9.83006597e-01
-1.00480966e-01 2.37059519e-01 5.96035458e-02 -1.99778527e-01
-8.46965671e-01 5.32678008e-01 6.75593734e-01 -4.01915073e-01
-3.74556273e-01 4.29426342e-01 1.61132753e-01 -1.13708094e-01
3.28704715e-01 -2.43942082e-01 -1.54074421e-02 2.10196286e-01
5.40065646e-01 6.10213041e-01 -3.39474410e-01 -4.25484031e-02
5.93707822e-02 2.90234774e-01 -2.33964205e-01 -7.33884990e-01
1.44804478e+00 -2.75613904e-01 3.86514850e-02 1.74051091e-01
1.00960648e+00 -1.02510341e-01 -1.57972848e+00 -3.43877494e-01
9.02649686e-02 -3.29200864e-01 -3.45734879e-02 -3.34526926e-01
-1.26846147e+00 3.05453598e-01 3.92592579e-01 5.22411942e-01
7.72704482e-01 -2.75353819e-01 3.65374058e-01 7.89756596e-01
5.83119512e-01 -1.55666304e+00 1.58432499e-01 6.93681598e-01
8.58881354e-01 -1.16293633e+00 3.78956832e-02 1.20645203e-01
-7.22537816e-01 1.10408223e+00 6.20103538e-01 -3.04774106e-01
6.16641462e-01 1.32939965e-01 -2.53081262e-01 9.17979255e-02
-8.22775245e-01 -4.63429332e-01 -1.88263446e-01 4.35314089e-01
-7.07799057e-03 2.19312087e-01 -9.27754998e-01 2.53541529e-01
-1.34842927e-02 -3.22797686e-01 9.11949754e-01 1.01682460e+00
-5.61298609e-01 -1.11940181e+00 -4.59860295e-01 6.83721662e-01
-5.54593921e-01 4.77225557e-02 1.00782834e-01 8.40240657e-01
-6.28295600e-01 1.01935542e+00 -2.26088658e-01 3.49493772e-01
1.84431419e-01 -5.10505557e-01 6.84836864e-01 -2.73033082e-01
-4.00571197e-01 1.64150313e-01 -1.14766397e-01 -4.95236933e-01
-4.40254152e-01 -7.53078997e-01 -1.04414630e+00 -5.14482677e-01
-2.87526309e-01 6.55782223e-01 5.39139509e-01 1.00170004e+00
5.47011122e-02 2.42077727e-02 9.18975353e-01 -7.18502760e-01
-1.42692137e+00 -7.82679856e-01 -7.54369020e-01 7.85553828e-02
7.03809679e-01 -6.95952296e-01 -4.62938994e-01 -6.36854887e-01] | [4.220066070556641, 2.620173692703247] |
067491f6-9331-40d7-9afc-8de3290dbfff | self-supervised-learning-for-neural | 2304.01023 | null | https://arxiv.org/abs/2304.01023v1 | https://arxiv.org/pdf/2304.01023v1.pdf | Self-Supervised learning for Neural Architecture Search (NAS) | The objective of this internship is to propose an innovative method that uses unlabelled data, i.e. data that will allow the AI to automatically learn to predict the correct outcome. To reach this stage, the steps to be followed can be defined as follows: (1) consult the state of the art and position ourself against it, (2) come up with ideas for development paths, (3) implement these ideas, (4) and finally test them to position ourself against the state of the art, and then start the sequence again. During my internship, this sequence was done several times and therefore gives the tracks explored during the internship. | ['Samuel Ducros'] | 2023-04-03 | null | null | null | null | ['architecture-search'] | ['methodology'] | [ 1.63369909e-01 3.66775006e-01 -2.71719158e-01 -3.79765511e-01
-3.47555339e-01 -6.60598159e-01 5.25770664e-01 1.49565861e-01
-1.26803041e-01 1.04761100e+00 1.14974245e-01 -5.23993790e-01
-5.54667175e-01 -8.04881752e-01 -4.46372092e-01 -4.61343735e-01
-9.60653648e-02 8.81811202e-01 4.81095582e-01 -3.70448828e-01
1.03402734e+00 5.66881418e-01 -1.75279105e+00 2.78819859e-01
9.77554560e-01 4.37930137e-01 4.99954633e-02 6.40967965e-01
-2.19476089e-01 9.66366351e-01 -4.28805172e-01 -5.23871899e-01
7.10371360e-02 -8.64064515e-01 -1.56718063e+00 1.77512646e-01
-1.97773069e-01 9.92155895e-02 2.17488155e-01 9.04488623e-01
6.37112781e-02 1.89421996e-01 7.49024868e-01 -1.13997650e+00
-8.76164436e-02 7.71612823e-01 -1.89747721e-01 1.27525121e-01
4.94378358e-01 -1.52373910e-01 6.01836741e-01 -7.36404419e-01
8.80316377e-01 1.12478054e+00 4.00229096e-01 4.64051187e-01
-9.19400811e-01 -7.89025486e-01 1.04094334e-01 3.52195024e-01
-1.25100243e+00 -4.28101480e-01 7.67341256e-01 -4.88631248e-01
5.92915177e-01 1.71175420e-01 1.13313317e+00 8.66560757e-01
1.03739671e-01 7.47262061e-01 1.14175510e+00 -9.41205859e-01
4.20748919e-01 4.44974184e-01 3.83785069e-01 8.93839359e-01
2.85952300e-01 1.28917918e-01 -3.56930941e-01 9.01507959e-02
4.84163672e-01 -1.97646454e-01 1.90223053e-01 -3.06950301e-01
-7.95767367e-01 7.83910513e-01 1.72733366e-02 8.78622293e-01
-3.41806114e-01 -2.51894265e-01 7.40832239e-02 5.88381350e-01
2.59262264e-01 1.07608712e+00 -4.23273981e-01 -4.19719964e-01
-1.02761948e+00 3.47766221e-01 1.08559036e+00 4.92346585e-01
1.01279378e+00 -4.99742448e-01 2.15563700e-01 6.27922952e-01
4.36703533e-01 -2.23188877e-01 5.14749408e-01 -1.10984111e+00
1.87701404e-01 1.07387674e+00 1.82912678e-01 -7.24680483e-01
-4.34078604e-01 -2.12429494e-01 -1.95234567e-01 4.18433368e-01
4.59708542e-01 -4.15549219e-01 -1.01703095e+00 1.45456338e+00
2.16830447e-02 1.22216828e-01 2.46329874e-01 4.07049358e-01
6.20737493e-01 6.32717967e-01 -1.81952640e-01 -2.08483860e-01
9.21754539e-01 -9.26739812e-01 -5.37092090e-01 -1.23058014e-01
9.59266186e-01 -7.71726191e-01 6.88278735e-01 9.15898263e-01
-1.16902363e+00 -7.77996600e-01 -1.12092686e+00 5.50106406e-01
-6.23129070e-01 1.35730609e-01 8.83126080e-01 8.23957682e-01
-1.24878573e+00 1.05627644e+00 -9.79561090e-01 -4.85342741e-01
-1.76093224e-02 4.36277628e-01 -1.14381015e-01 -6.90088561e-03
-1.27879512e+00 1.08826625e+00 7.67666399e-01 3.91878188e-02
-9.63024378e-01 -1.77924410e-01 -5.99222660e-01 -2.52182126e-01
5.72391808e-01 -4.72662538e-01 1.00537169e+00 -1.22661471e+00
-1.63820922e+00 8.21964979e-01 -6.26059175e-02 -7.23963603e-02
3.17709088e-01 2.96906792e-02 -3.37579519e-01 -2.75573492e-01
1.96021721e-01 6.77261472e-01 8.97615552e-02 -1.24091780e+00
-1.12839115e+00 -4.28365469e-01 3.32889557e-01 5.44330180e-02
-3.74065638e-01 8.21238831e-02 -7.91466594e-01 -2.64642030e-01
3.36390495e-01 -9.48122799e-01 -2.71223485e-01 -6.95047379e-01
-3.21395934e-01 -6.14794433e-01 3.03165644e-01 -3.48676443e-01
1.45678866e+00 -1.77918243e+00 4.40463454e-01 4.12644029e-01
-2.00301647e-01 1.97306827e-01 3.27697694e-02 7.89450586e-01
-3.04238498e-01 2.01256409e-01 2.66721509e-02 -2.65881903e-02
-2.22291827e-01 1.18964463e-01 1.22080438e-01 6.96937889e-02
5.02804033e-02 6.71976805e-01 -1.09557045e+00 -4.97971207e-01
3.50310802e-02 -2.34959066e-01 -3.58116627e-01 1.20324783e-01
-1.63544506e-01 3.80185366e-01 -8.19539666e-01 5.78772068e-01
1.49740607e-01 -2.36334652e-02 3.70139748e-01 2.44306535e-01
-6.43781304e-01 4.56629068e-01 -1.40756714e+00 1.51363993e+00
-9.72257778e-02 6.44696593e-01 -3.92917871e-01 -1.25025403e+00
1.08657372e+00 2.26193905e-01 3.17198247e-01 -2.72939056e-01
1.67636842e-01 2.92519331e-01 1.24163561e-01 -7.84051001e-01
3.37026417e-01 -2.20890671e-01 2.07472276e-02 6.65130258e-01
-4.72412333e-02 -3.71534750e-02 5.94036102e-01 -1.43613189e-01
1.13092673e+00 4.37622041e-01 2.89119899e-01 -1.22253940e-01
8.40527117e-01 4.48477089e-01 4.97550070e-01 7.52162158e-01
1.99356839e-01 6.29422441e-02 6.88088298e-01 -7.86427915e-01
-9.22576427e-01 -7.91532815e-01 4.91327746e-03 7.88256168e-01
-9.42203030e-02 -3.50387394e-01 -6.74006939e-01 -9.32957590e-01
-4.78842020e-01 7.87574649e-01 -9.92381573e-01 -2.87613235e-02
-4.70639735e-01 -1.92389965e-01 3.28892320e-01 3.87614787e-01
5.34538805e-01 -1.30437362e+00 -7.24559903e-01 2.49100298e-01
3.52020338e-02 -2.04723537e-01 4.10358667e-01 3.91286016e-01
-9.20424759e-01 -8.94400120e-01 -6.08546734e-01 -8.74868929e-01
7.64329433e-01 -2.05923960e-01 1.12448299e+00 3.32989872e-01
-2.12131832e-02 -9.04716551e-03 -6.69655383e-01 -6.48939252e-01
-5.84397435e-01 2.56829500e-01 -1.77910373e-01 -4.41728026e-01
3.06250364e-01 -5.27623892e-01 -1.72915399e-01 2.81384647e-01
-7.85169542e-01 5.30611686e-02 8.35161984e-01 7.46720374e-01
3.67074579e-01 4.88021523e-01 5.48386991e-01 -1.14308262e+00
5.03966033e-01 -2.78334707e-01 -6.45175457e-01 6.05927050e-01
-9.73520398e-01 2.35015959e-01 3.15534532e-01 -1.65263608e-01
-9.23455656e-01 9.24847275e-02 -2.69399583e-01 -7.67105967e-02
-5.54817200e-01 6.99910343e-01 -1.72801837e-01 1.14558108e-01
6.19362116e-01 3.63469049e-02 -2.95992941e-01 -4.44030225e-01
3.34527157e-02 7.82729030e-01 2.88940907e-01 -4.10607100e-01
6.84828341e-01 2.22389512e-02 7.78937191e-02 -3.65820318e-01
-1.02270555e+00 -3.25201303e-01 -9.75261033e-01 -3.94030869e-01
8.49753499e-01 -2.21619695e-01 -4.05044109e-01 2.05454707e-01
-9.20959234e-01 -4.28452939e-01 -1.86680347e-01 7.19681501e-01
-4.50159073e-01 3.00032534e-02 -2.83954173e-01 -1.03973424e+00
1.79158077e-01 -1.06069684e+00 3.18822473e-01 5.47645032e-01
-4.75304484e-01 -1.31255615e+00 4.08745617e-01 5.21054745e-01
-1.38837412e-01 7.30751157e-02 9.44221497e-01 -9.84000921e-01
-3.82565767e-01 -2.47407854e-01 2.32949227e-01 2.04434365e-01
1.60855159e-01 2.08463505e-01 -6.04611814e-01 -1.68257281e-01
-1.16274856e-01 -3.52427393e-01 4.87753808e-01 2.52941161e-01
1.16077220e+00 -7.81456828e-02 -8.56621265e-01 4.32633638e-01
1.08349490e+00 7.30680943e-01 8.35129678e-01 9.85856652e-01
1.38984546e-01 1.19307983e+00 9.45465565e-01 5.96990203e-03
4.18615073e-01 5.08047163e-01 1.94831118e-01 -1.30888402e-01
2.40270332e-01 -2.68256664e-01 3.91495615e-01 9.22552526e-01
-2.13788539e-01 -2.73805052e-01 -1.15890181e+00 6.49289966e-01
-1.91265035e+00 -8.31422448e-01 -1.57922119e-01 1.94988477e+00
5.94307840e-01 7.48621345e-01 4.32504982e-01 4.21327382e-01
3.80955875e-01 -3.10630053e-01 -2.22873554e-01 -8.41879547e-01
5.06560802e-01 3.56701612e-01 2.60949165e-01 2.02583060e-01
-8.38051617e-01 1.19595540e+00 7.10829496e+00 5.91069281e-01
-7.52115369e-01 -2.32151657e-01 6.34481788e-01 4.89574760e-01
-3.34298313e-01 5.48100650e-01 -8.12511683e-01 2.31650904e-01
1.08615816e+00 -3.71722609e-01 4.84310925e-01 5.87155282e-01
-7.99661502e-02 -2.30547145e-01 -1.25279868e+00 2.20624760e-01
2.45651081e-01 -1.10728765e+00 3.43943126e-02 1.03301227e-01
6.17884338e-01 -3.72668266e-01 -2.96715975e-01 4.53629345e-01
6.86332345e-01 -1.16812539e+00 3.29197943e-01 6.18339002e-01
1.78015411e-01 -1.14394760e+00 9.23209429e-01 8.23723614e-01
-7.93598115e-01 -2.03766286e-01 -3.35482746e-01 -3.89691710e-01
-3.03970963e-01 9.08464789e-02 -1.09097016e+00 1.25652599e+00
4.58317965e-01 5.97828865e-01 -4.95907277e-01 1.31917441e+00
-4.77943599e-01 5.87524772e-01 6.58501312e-02 -4.25991654e-01
4.79053944e-01 -2.67250210e-01 3.47762734e-01 9.73052502e-01
5.73591232e-01 1.69646010e-01 1.53521314e-01 7.35575378e-01
2.73862153e-01 2.14081798e-02 -6.43254101e-01 -1.09416686e-01
2.52715379e-01 1.04516411e+00 -1.08039033e+00 -1.74504876e-01
-1.23238772e-01 8.36962521e-01 1.44728690e-01 7.30992407e-02
-2.63146937e-01 -5.84663510e-01 9.65976119e-02 9.09234062e-02
2.47672975e-01 3.06402594e-01 -6.56302646e-02 -7.11686194e-01
-3.96377057e-01 -8.85273337e-01 3.93308580e-01 -8.58240247e-01
-8.31418991e-01 7.58053064e-01 2.46734008e-01 -1.17700326e+00
-5.08133769e-01 -4.71316844e-01 -7.64019489e-01 8.68720651e-01
-1.10308135e+00 -8.28907907e-01 -9.52700526e-02 2.95586914e-01
6.24377072e-01 -2.92549640e-01 9.01726246e-01 1.57897696e-01
-6.16407275e-01 3.85906458e-01 -1.77586168e-01 1.61672458e-01
3.98945361e-01 -1.13438261e+00 2.45662346e-01 7.32829273e-01
1.97032228e-01 6.85937047e-01 7.27393091e-01 -9.68548715e-01
-1.11339319e+00 -7.04309821e-01 1.18831742e+00 -5.93771875e-01
7.24002361e-01 1.85665488e-01 -7.09246576e-01 4.89020020e-01
-8.69759098e-02 -7.12257862e-01 7.16804266e-01 4.08103764e-01
4.13235664e-01 3.31812091e-02 -8.63467693e-01 5.42100132e-01
8.69719326e-01 -4.02925648e-02 -1.08451939e+00 2.52261162e-01
2.41911441e-01 -4.34789091e-01 -7.39004612e-01 4.27188218e-01
5.98411024e-01 -1.00537241e+00 5.42766035e-01 -8.20810318e-01
5.95093250e-01 -2.15647534e-01 3.02102029e-01 -1.39784312e+00
-4.32470977e-01 -5.90378821e-01 3.49289179e-01 1.34020352e+00
9.25556600e-01 -5.09279788e-01 9.67486560e-01 2.77946293e-01
-3.09806257e-01 -9.67248797e-01 -4.58062470e-01 -7.34477639e-01
1.11410916e-01 -3.51926923e-01 5.86302817e-01 1.13109159e+00
1.63359702e-01 1.50727689e-01 -2.45006680e-01 -2.26653963e-01
2.72760689e-01 3.79401028e-01 7.73524940e-01 -1.80489206e+00
-7.03608766e-02 -4.27783728e-01 -3.42775106e-01 -8.86706412e-01
4.97840829e-02 -9.13959861e-01 -7.91282021e-03 -1.85985994e+00
8.40673745e-02 -5.80323160e-01 -6.05043530e-01 6.49157107e-01
-1.25062749e-01 1.30565926e-01 -4.02515158e-02 3.80626947e-01
-5.58270931e-01 -2.12504670e-01 1.07124364e+00 3.14340979e-01
-4.99861985e-01 3.49153101e-01 -1.04483581e+00 1.08700657e+00
6.03655040e-01 -6.22439802e-01 -3.76741886e-01 1.54283270e-01
4.25337166e-01 4.60884213e-01 -9.32506546e-02 -1.26830184e+00
2.30432302e-01 -3.88440073e-01 8.25941324e-01 -7.91578531e-01
2.68689785e-02 -5.78228474e-01 3.50755870e-01 7.96182334e-01
-4.85317171e-01 1.35999486e-01 1.36494800e-01 6.46762997e-02
-2.45140746e-01 -9.88168597e-01 3.72289836e-01 -2.84714788e-01
-5.74743152e-01 -1.76375821e-01 -5.82826972e-01 -3.09817016e-01
1.29741943e+00 -4.94529486e-01 9.02288929e-02 -1.01074971e-01
-1.01613891e+00 5.62838852e-01 6.17268562e-01 1.98057681e-01
2.51986027e-01 -9.97931063e-01 -5.48710704e-01 1.68551132e-01
-3.25582325e-02 -4.22223508e-01 -3.09300303e-01 7.52986789e-01
-4.64305639e-01 5.84837139e-01 -4.58920896e-01 -3.65570962e-01
-1.29184210e+00 5.22325993e-01 1.53731450e-01 -6.62711263e-01
-3.86108845e-01 9.56925571e-01 -3.21004301e-01 -2.60327488e-01
2.96030849e-01 1.32949203e-01 -1.06570280e+00 2.98873842e-01
4.76766467e-01 2.49300122e-01 1.05804719e-01 -1.09550178e-01
-4.92674440e-01 6.38984442e-01 -1.01462223e-01 -4.82158899e-01
1.70124900e+00 -2.09534690e-02 1.30106946e-02 8.34228396e-01
7.52851009e-01 -8.64644721e-02 -7.46725440e-01 1.56876341e-01
4.98262018e-01 -5.23132980e-01 -2.90206745e-02 -1.07802331e+00
-7.50386536e-01 5.98926723e-01 5.91614544e-01 4.95435148e-01
1.25900650e+00 6.28952309e-02 2.89392292e-01 4.84486848e-01
5.31411469e-01 -1.15380204e+00 -1.65166017e-02 4.95182037e-01
4.31372762e-01 -7.97920108e-01 2.09488273e-01 -1.39208227e-01
-7.57096469e-01 1.55166626e+00 5.03962576e-01 1.39422834e-01
4.44639385e-01 1.34992406e-01 -1.20362900e-01 -4.38308477e-01
-1.06554055e+00 -2.84607381e-01 3.62590611e-01 4.25454587e-01
4.72718060e-01 -3.05465877e-01 -6.89572752e-01 2.64445990e-01
-2.91794240e-01 5.40194035e-01 5.11955202e-01 1.07484496e+00
-5.94606578e-01 -1.73017812e+00 -4.45238411e-01 3.55251074e-01
-5.02688110e-01 2.59760052e-01 -9.85805154e-01 9.34141099e-01
6.32354617e-01 1.19030404e+00 -2.07332090e-01 -7.15796947e-01
5.22428632e-01 2.23160520e-01 4.14790154e-01 -7.22257376e-01
-8.07484150e-01 7.08390325e-02 3.22813362e-01 -4.30326641e-01
-6.27643943e-01 -7.68436253e-01 -1.15377069e+00 -1.92068040e-01
-3.12860429e-01 7.14281797e-01 5.82495868e-01 1.15580273e+00
-1.50674313e-01 5.82479477e-01 8.67905676e-01 -6.87435389e-01
-7.19756410e-02 -9.63289380e-01 -1.28120407e-01 2.16409434e-02
-2.00402454e-01 -8.68458450e-01 -4.13852721e-01 -1.45535111e-01] | [8.545801162719727, 4.664829730987549] |
f42ce96b-7dcf-4757-a32f-f741d626b025 | supervising-neural-attention-models-for-video | 1707.06029 | null | http://arxiv.org/abs/1707.06029v1 | http://arxiv.org/pdf/1707.06029v1.pdf | Supervising Neural Attention Models for Video Captioning by Human Gaze Data | The attention mechanisms in deep neural networks are inspired by human's
attention that sequentially focuses on the most relevant parts of the
information over time to generate prediction output. The attention parameters
in those models are implicitly trained in an end-to-end manner, yet there have
been few trials to explicitly incorporate human gaze tracking to supervise the
attention models. In this paper, we investigate whether attention models can
benefit from explicit human gaze labels, especially for the task of video
captioning. We collect a new dataset called VAS, consisting of movie clips, and
corresponding multiple descriptive sentences along with human gaze tracking
data. We propose a video captioning model named Gaze Encoding Attention Network
(GEAN) that can leverage gaze tracking information to provide the spatial and
temporal attention for sentence generation. Through evaluation of language
similarity metrics and human assessment via Amazon mechanical Turk, we
demonstrate that spatial attentions guided by human gaze data indeed improve
the performance of multiple captioning methods. Moreover, we show that the
proposed approach achieves the state-of-the-art performance for both gaze
prediction and video captioning not only in our VAS dataset but also in
standard datasets (e.g. LSMDC and Hollywood2). | ['Sang-Hun Lee', 'Yeonhwa Kim', 'Jongwook Choi', 'Youngjae Yu', 'Kyung Yoo', 'Gunhee Kim'] | 2017-07-19 | supervising-neural-attention-models-for-video-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Yu_Supervising_Neural_Attention_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Yu_Supervising_Neural_Attention_CVPR_2017_paper.pdf | cvpr-2017-7 | ['eye-tracking'] | ['computer-vision'] | [ 1.88618466e-01 8.04832429e-02 -1.38779692e-02 -5.50252676e-01
-4.96847212e-01 -3.50208938e-01 5.66242337e-01 -3.15307647e-01
-3.99717003e-01 5.26196241e-01 3.00013542e-01 -1.41289413e-01
1.67064294e-01 -5.09184897e-02 -7.66377568e-01 -4.69338894e-01
2.38206193e-01 5.09608053e-02 7.38952532e-02 -1.30892709e-01
5.55514276e-01 -1.85043216e-01 -1.73672187e+00 3.23209882e-01
8.91381383e-01 1.21369076e+00 5.54423392e-01 8.78353715e-01
1.61543190e-01 1.21356368e+00 -5.92007279e-01 -5.65676749e-01
-2.65290022e-01 -7.92445898e-01 -7.99902320e-01 -3.11208963e-01
9.01404917e-01 -4.28787142e-01 -2.94520646e-01 8.02786171e-01
6.22522235e-01 4.54782039e-01 7.21060753e-01 -1.73398769e+00
-1.54817140e+00 3.14995736e-01 -7.60407209e-01 6.13488376e-01
4.96042460e-01 4.67868328e-01 1.10081720e+00 -1.17888057e+00
4.37059283e-01 1.12770259e+00 3.06171566e-01 1.17408013e+00
-7.44784594e-01 -7.13410437e-01 1.95383579e-01 6.58681989e-01
-1.19531202e+00 -7.05516517e-01 7.27589726e-01 -5.21526277e-01
8.32577884e-01 4.08463627e-01 1.29832789e-01 1.41422713e+00
5.39842285e-02 1.10608792e+00 4.89549279e-01 -4.35153246e-01
-1.27096415e-01 5.12196086e-02 2.36975908e-01 4.80802625e-01
-2.58381456e-01 2.15020459e-02 -1.07336950e+00 3.44777435e-01
3.77392888e-01 4.46776934e-02 -6.15181684e-01 1.52504984e-02
-1.35293019e+00 7.22636044e-01 8.99470866e-01 -5.33649884e-02
-4.99478459e-01 2.96745181e-01 2.26594120e-01 -1.60167530e-01
5.66337466e-01 3.40927869e-01 -2.56519634e-02 -2.25466758e-01
-9.39059258e-01 8.66338760e-02 1.68692082e-01 1.26002717e+00
4.79115427e-01 -2.68203914e-01 -1.12301338e+00 6.55284762e-01
6.54300988e-01 8.21966827e-01 5.31716704e-01 -1.08899915e+00
8.12454104e-01 3.40741158e-01 3.07099611e-01 -1.07346451e+00
-2.91498333e-01 1.12812147e-01 -6.11749947e-01 -2.89683521e-01
2.28060901e-01 -2.32454076e-01 -7.38447666e-01 2.10623240e+00
2.44530067e-02 2.26508290e-01 -1.82908684e-01 1.28245723e+00
1.08963132e+00 6.39479876e-01 4.16107625e-01 -3.45350564e-01
1.56359124e+00 -1.53694475e+00 -1.07823074e+00 -3.11617911e-01
5.05853772e-01 -4.75268722e-01 1.43941689e+00 7.93693066e-02
-1.30852604e+00 -7.27265775e-01 -6.58548594e-01 -5.11367559e-01
1.57200709e-01 2.12618500e-01 1.33582577e-01 3.76337290e-01
-1.33651590e+00 2.73537427e-01 -5.44158816e-01 -4.84746546e-01
6.47406459e-01 2.65606165e-01 -4.87490930e-02 1.24165297e-01
-1.20714736e+00 7.04970181e-01 -1.10772461e-01 3.06929737e-01
-8.12333524e-01 -4.98277873e-01 -7.92660236e-01 3.01299304e-01
1.62372768e-01 -8.42592061e-01 1.61110902e+00 -1.33324575e+00
-1.38922441e+00 7.34364212e-01 -8.31712663e-01 -3.94943744e-01
1.88214630e-01 -5.27892828e-01 -2.13365257e-01 4.15247232e-01
1.24608226e-01 1.37510908e+00 9.94284928e-01 -9.37781692e-01
-6.25868738e-01 -1.55121490e-01 2.33949214e-01 3.21024776e-01
-6.49077356e-01 5.87519467e-01 -4.72063094e-01 -5.48154294e-01
-6.90207362e-01 -1.02692747e+00 3.36512595e-01 1.10302791e-01
-4.90741402e-01 -6.59192860e-01 7.62681305e-01 -6.81996882e-01
1.53371561e+00 -2.07575846e+00 2.53568500e-01 -4.35941249e-01
4.01923507e-01 1.75254330e-01 -2.53835946e-01 1.28824249e-01
-6.26575574e-02 1.71715677e-01 -7.25236908e-02 -9.20966923e-01
5.82663678e-02 -3.35967571e-01 -4.98166591e-01 1.93875507e-01
3.06127727e-01 1.28083909e+00 -1.11553514e+00 -6.18358850e-01
-1.62829459e-01 5.56651175e-01 -5.48203707e-01 5.80930650e-01
-2.01212093e-01 5.96867025e-01 -1.30731612e-01 3.53424489e-01
3.38104993e-01 -8.24435294e-01 -3.72383028e-01 -3.42237242e-02
-3.67407538e-02 1.14891656e-01 -3.37001681e-02 1.65300739e+00
-2.48127714e-01 1.20842564e+00 -3.00349861e-01 -3.36441338e-01
6.40030921e-01 4.49999064e-01 3.46501246e-02 -8.83837342e-01
3.60282272e-01 -3.54802340e-01 9.06386878e-03 -9.42196131e-01
7.85478294e-01 3.29976410e-01 1.28273025e-01 9.16153252e-01
1.56904131e-01 8.55523765e-01 1.69899613e-01 3.26021314e-01
6.20023906e-01 1.43861726e-01 -5.85057475e-02 -1.32065699e-01
7.30682015e-01 -2.68998384e-01 7.28009418e-02 7.65312493e-01
-6.85293078e-01 9.80017602e-01 4.70317006e-01 -7.19540119e-02
-8.82834077e-01 -5.26203334e-01 2.92743951e-01 2.01246762e+00
3.39596540e-01 -3.07036012e-01 -1.33209431e+00 -7.84634054e-01
-4.84050751e-01 9.87731934e-01 -1.08227122e+00 -3.01125765e-01
-3.75789106e-01 -2.44213820e-01 3.57575178e-01 7.45218039e-01
3.22238386e-01 -1.69096076e+00 -7.89603531e-01 -4.07574713e-01
-5.78129351e-01 -1.13926661e+00 -1.19576931e+00 -7.29287326e-01
-3.07588488e-01 -1.10743201e+00 -1.24351251e+00 -7.47448981e-01
7.50198960e-01 6.02813065e-01 1.11252749e+00 2.20019281e-01
2.83603430e-01 4.63847041e-01 -4.61531967e-01 -5.11873424e-01
-5.05683683e-02 2.61125475e-01 2.11932108e-01 3.86723816e-01
7.99097061e-01 4.48630638e-02 -8.91549647e-01 5.26107848e-01
-5.76052487e-01 3.08718264e-01 2.55463183e-01 6.15824878e-01
4.34472933e-02 -1.17865562e+00 6.11383438e-01 -5.10113120e-01
8.91451657e-01 -8.12913716e-01 -2.50714958e-01 5.10826945e-01
-3.38092268e-01 -1.21705592e-01 3.14282268e-01 -3.70829731e-01
-1.00102878e+00 -2.99521625e-01 2.58670598e-01 -1.02042747e+00
-1.89617366e-01 3.19122255e-01 2.94506967e-01 1.89174265e-01
7.19869375e-01 1.72970250e-01 5.36679588e-02 -2.52165943e-01
2.38772660e-01 1.07124162e+00 5.00789464e-01 -2.71849744e-02
3.44042748e-01 1.81762725e-01 -2.38339588e-01 -5.14074206e-01
-1.28560174e+00 -3.40269893e-01 -5.08774102e-01 -6.42242312e-01
1.34257424e+00 -9.28048849e-01 -1.37922263e+00 4.39173609e-01
-1.57564080e+00 -2.81779706e-01 3.37151915e-01 3.13962013e-01
-6.53168619e-01 1.16034560e-01 -3.48862350e-01 -9.58566010e-01
-8.40812504e-01 -1.08492410e+00 1.22459686e+00 4.93091434e-01
-3.86611164e-01 -7.88479149e-01 1.88978706e-02 4.99920040e-01
4.60591018e-01 -2.12294981e-01 2.75939703e-01 -6.49869978e-01
-6.05613470e-01 1.51783660e-01 -5.29666603e-01 3.80586125e-02
-1.17887691e-01 3.70309278e-02 -1.20668423e+00 -1.88584775e-01
-1.64987475e-01 -5.32358706e-01 7.45324016e-01 5.43165743e-01
1.25941706e+00 -3.73595685e-01 -3.79118264e-01 4.08618271e-01
7.04476893e-01 -1.09006194e-02 6.93948507e-01 1.48154899e-01
9.98231947e-01 9.20184970e-01 7.18113720e-01 2.46176004e-01
7.21436977e-01 7.79659033e-01 6.21131778e-01 2.77220994e-01
-3.75521295e-02 -1.83117837e-01 4.64074790e-01 5.41544080e-01
-1.71368435e-01 -8.23815227e-01 -1.00653851e+00 6.04255259e-01
-2.04388571e+00 -1.17913604e+00 -3.69364291e-01 2.02243090e+00
7.05274761e-01 -2.85942167e-01 3.27118635e-01 -3.99792671e-01
1.17421043e+00 8.45497251e-02 -6.34394407e-01 -3.46229911e-01
1.62353843e-01 -4.05956805e-01 -5.63364401e-02 2.94231206e-01
-1.01350522e+00 8.01378191e-01 6.09241247e+00 2.92872161e-01
-1.24644947e+00 2.88800448e-01 6.28936350e-01 -6.45061791e-01
-5.41164726e-02 -5.08768260e-01 -7.73987710e-01 1.01010644e+00
1.36330628e+00 -1.60819292e-01 4.05666560e-01 5.82070649e-01
4.69365120e-01 5.51076755e-02 -1.25854540e+00 1.19723547e+00
7.01736927e-01 -1.14779806e+00 -1.85838580e-01 -1.13098517e-01
6.24179661e-01 1.70956656e-01 5.29164791e-01 2.56105572e-01
-2.69217759e-01 -1.14882004e+00 9.65698600e-01 9.33939517e-01
9.79383647e-01 -4.62269753e-01 5.66456974e-01 3.22068930e-01
-5.80561340e-01 -2.88774729e-01 -8.14676359e-02 -4.62896377e-02
4.97467607e-01 -3.63702476e-01 -8.24124336e-01 2.92581273e-03
8.70933115e-01 9.22022700e-01 -8.89884710e-01 1.10879242e+00
-4.58452672e-01 6.42623186e-01 2.36689225e-01 -4.71449465e-01
2.40184695e-01 2.68039674e-01 4.52983886e-01 8.43332410e-01
4.36329007e-01 1.43200099e-01 -6.30274832e-01 1.07589102e+00
-4.40325797e-01 -7.54150301e-02 -3.20895612e-01 -3.05742901e-02
5.16201794e-01 1.07991242e+00 -1.61423355e-01 -2.31863290e-01
-5.02161801e-01 8.44366133e-01 5.61060607e-01 6.81212664e-01
-1.17484403e+00 -4.48514789e-01 6.02376163e-01 1.14946924e-01
5.49806774e-01 1.31521642e-01 -4.78800833e-02 -9.66798902e-01
-8.64485744e-03 -4.76131350e-01 2.12924868e-01 -1.64353156e+00
-1.33502913e+00 9.07456934e-01 -1.19376101e-01 -1.26103103e+00
-5.86772561e-01 -4.79056448e-01 -7.42168784e-01 1.19522870e+00
-1.54854512e+00 -1.06284761e+00 -7.06431806e-01 7.71783531e-01
6.96898460e-01 -1.91249564e-01 4.87791479e-01 1.75365239e-01
-8.13278377e-01 1.03877091e+00 -1.25759035e-01 1.14044003e-01
1.14854944e+00 -1.01194799e+00 3.99393827e-01 7.97636092e-01
-8.65287893e-03 6.99859977e-01 7.93835163e-01 -3.60324442e-01
-8.66812587e-01 -1.03506804e+00 1.24563694e+00 -9.82049108e-01
4.12857741e-01 -3.42032552e-01 -9.35753703e-01 8.67915332e-01
1.01553309e+00 -1.37537671e-02 8.39630365e-01 6.23104814e-03
-9.50995013e-02 2.19437957e-01 -7.56332457e-01 6.66576684e-01
1.13590395e+00 -6.06433511e-01 -6.07586265e-01 2.23545805e-01
9.65586960e-01 -4.12270844e-01 -2.14553341e-01 2.52655335e-02
4.09668982e-01 -8.83924067e-01 6.38674915e-01 -9.71954107e-01
9.45796132e-01 -2.18339100e-01 2.51519293e-01 -9.34971511e-01
-5.01728773e-01 -7.83908784e-01 -6.15514696e-01 1.22018027e+00
3.67535830e-01 -8.34663808e-02 5.00280201e-01 9.20165896e-01
-2.37014696e-01 -4.50647384e-01 -6.14042342e-01 -2.53477305e-01
-2.43783101e-01 -1.79044515e-01 6.06800139e-01 6.76418781e-01
5.82116283e-02 8.43843162e-01 -7.18711317e-01 -5.36061041e-02
4.39079642e-01 -1.13345735e-01 7.32475936e-01 -1.12262952e+00
2.27891624e-01 -4.46176350e-01 -1.47701487e-01 -1.27216613e+00
4.61850554e-01 -6.39840007e-01 3.65975440e-01 -1.27672052e+00
3.73636633e-01 2.08725899e-01 -4.72693682e-01 4.25784737e-01
-6.63993180e-01 4.60177958e-01 5.17259359e-01 5.34397185e-01
-1.32534134e+00 7.51324058e-01 1.33495533e+00 4.90997545e-02
-4.87683192e-02 -8.15792382e-02 -9.76106226e-01 4.14859056e-01
3.66511941e-01 -3.49408269e-01 -4.71758008e-01 -9.73315716e-01
3.47850323e-01 -6.10666946e-02 6.05170786e-01 -8.07538152e-01
6.74074113e-01 -1.24275252e-01 1.65617421e-01 -5.55928767e-01
3.37691993e-01 -4.77830023e-01 -5.25317192e-01 -2.53980875e-01
-9.12525237e-01 2.30729133e-01 7.96446130e-02 6.68464065e-01
-8.71872753e-02 -3.74916881e-01 4.06020671e-01 3.66919309e-01
-6.10443592e-01 4.97274667e-01 -2.59425528e-02 3.22708696e-01
1.11642659e+00 -8.00290033e-02 -6.64626956e-01 -7.40821004e-01
-7.03911126e-01 4.80169564e-01 3.65864366e-01 9.06522334e-01
5.12581348e-01 -1.52124536e+00 -6.92767680e-01 5.81140369e-02
3.76479447e-01 -1.66052163e-01 5.24296224e-01 1.16187668e+00
-1.28401041e-01 7.82440841e-01 -2.39362672e-01 -9.48897660e-01
-1.32016242e+00 8.82346869e-01 1.64997458e-01 4.17867631e-01
-3.02620307e-02 1.22981822e+00 6.47355020e-01 2.18905956e-01
3.45517069e-01 -9.06526372e-02 -7.47589469e-01 1.87166855e-01
1.10259163e+00 1.51349857e-01 -3.36230218e-01 -1.20286858e+00
-3.75917763e-01 3.70340914e-01 -9.90463048e-02 4.57334593e-02
1.05792677e+00 -7.28192627e-01 8.92868713e-02 3.35019559e-01
1.15872574e+00 -4.15278673e-01 -1.58710372e+00 -1.65012240e-01
-2.28507444e-01 -5.45396626e-01 -3.55129875e-02 -8.84620428e-01
-1.03481591e+00 1.37103987e+00 4.14358318e-01 3.42163771e-01
1.08419251e+00 1.16386533e-01 8.60835612e-01 2.29753390e-01
-1.98338151e-01 -7.68982887e-01 3.26315492e-01 6.27522647e-01
1.12457788e+00 -1.62013960e+00 -6.29715383e-01 1.37764692e-01
-1.24313915e+00 7.28903055e-01 1.06384420e+00 1.56384990e-01
3.21475774e-01 -6.12394512e-01 4.14741822e-02 -1.62379503e-01
-1.18044496e+00 -1.49646744e-01 7.27050364e-01 5.74412405e-01
5.18919885e-01 -4.60628361e-01 9.86548513e-02 9.57450986e-01
-1.20187834e-01 3.12661320e-01 4.20431256e-01 5.11783302e-01
-3.92464489e-01 -2.97965884e-01 -2.97211945e-01 3.55918914e-01
-5.08013546e-01 -3.86418462e-01 -3.03712726e-01 3.29389393e-01
-3.12268317e-01 9.92903054e-01 6.45163655e-01 -3.10951740e-01
2.17350662e-01 1.49190381e-01 2.77975738e-01 -4.68255013e-01
-4.98844773e-01 -3.86612713e-01 -2.09921420e-01 -5.75610876e-01
-6.98229253e-01 -6.76692188e-01 -8.88318300e-01 -2.92992324e-01
-3.16016287e-01 3.83377708e-02 3.55693161e-01 9.87839341e-01
9.22351956e-01 5.84842741e-01 6.26027763e-01 -1.13033617e+00
-3.84807825e-01 -1.34926653e+00 -5.12742139e-02 5.75552881e-01
7.04516768e-01 -7.98836410e-01 -4.04561371e-01 4.44288164e-01] | [14.015166282653809, 0.053274061530828476] |
5f25624c-9546-4934-8aed-7525bdd9d814 | robust-pose-estimation-in-crowded-scenes-with | null | null | http://proceedings.neurips.cc/paper/2021/hash/31857b449c407203749ae32dd0e7d64a-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/31857b449c407203749ae32dd0e7d64a-Paper.pdf | Robust Pose Estimation in Crowded Scenes with Direct Pose-Level Inference | Multi-person pose estimation in crowded scenes is challenging because overlapping and occlusions make it difficult to detect person bounding boxes and infer pose cues from individual keypoints. To address those issues, this paper proposes a direct pose-level inference strategy that is free of bounding box detection and keypoint grouping. Instead of inferring individual keypoints, the Pose-level Inference Network (PINet) directly infers the complete pose cues for a person from his/her visible body parts. PINet first applies the Part-based Pose Generation (PPG) to infer multiple coarse poses for each person from his/her body parts. Those coarse poses are refined by the Pose Refinement module through incorporating pose priors, and finally are fused in the Pose Fusion module. PINet relies on discriminative body parts to differentiate overlapped persons, and applies visual body cues to infer the global pose cues. Experiments on several crowded scenes pose estimation benchmarks demonstrate the superiority of PINet. For instance, it achieves 59.8% AP on the OCHuman dataset, outperforming the recent works by a large margin. | ['Gang Hua', 'Shiliang Zhang', 'Dongkai Wang'] | 2021-12-01 | null | https://openreview.net/forum?id=AvHeCmK2fsE | https://openreview.net/pdf?id=AvHeCmK2fsE | neurips-2021-12 | ['multi-person-pose-estimation'] | ['computer-vision'] | [-4.74349000e-02 5.38085885e-02 6.56657293e-02 -2.30465427e-01
-7.23281622e-01 -4.12364691e-01 4.95365322e-01 7.18848407e-02
-3.59213352e-01 6.05865896e-01 4.05003637e-01 5.56292653e-01
1.49695531e-01 -5.45287371e-01 -6.17901623e-01 -5.03022969e-01
-1.33044243e-01 9.65402067e-01 4.97849762e-01 -1.89317837e-01
-1.96264699e-01 2.79335588e-01 -1.48361897e+00 -6.56289011e-02
6.96651220e-01 9.63611603e-01 -1.72902599e-01 6.94373548e-01
3.19355845e-01 3.40505004e-01 -8.36111188e-01 -6.44521117e-01
4.18555945e-01 -8.98986533e-02 -4.20136273e-01 3.28711301e-01
1.14974105e+00 -6.24907613e-01 -1.74635842e-01 7.59147942e-01
6.45536900e-01 2.47809902e-01 6.09805465e-01 -1.36032474e+00
1.10413164e-01 1.37321174e-01 -1.17090011e+00 -4.00289893e-02
1.01364553e+00 7.91855156e-02 8.21592033e-01 -9.60756660e-01
5.22432446e-01 1.65452504e+00 7.35847950e-01 4.24391270e-01
-1.10991836e+00 -7.63591826e-01 5.52777290e-01 -1.80447146e-01
-1.78104103e+00 -3.09036672e-01 6.59114420e-01 -6.30132556e-01
4.09692675e-01 2.56548494e-01 9.54493940e-01 9.14207816e-01
4.70192209e-02 9.68599677e-01 8.00579011e-01 -3.67470309e-02
-1.16796922e-02 -4.83964592e-01 1.18961046e-02 8.47200513e-01
6.32305086e-01 -2.13568375e-01 -8.80142152e-01 -4.55050617e-01
8.53276312e-01 2.47563884e-01 -8.65181759e-02 -5.74291527e-01
-1.28050375e+00 5.53538442e-01 6.23154223e-01 -4.51892287e-01
-4.67490613e-01 2.38691270e-01 2.24428609e-01 -5.74153125e-01
3.75361860e-01 1.11756071e-01 -1.98531538e-01 2.26242200e-01
-1.06668270e+00 9.24733400e-01 5.65001786e-01 1.24780118e+00
7.75866508e-01 -3.80317241e-01 -7.06409514e-01 5.54324508e-01
4.97544765e-01 7.03180552e-01 -1.47920728e-01 -9.56495762e-01
8.42611849e-01 7.13095129e-01 4.93447721e-01 -1.26206779e+00
-5.35676301e-01 -4.08517092e-01 -7.26959229e-01 -1.45551547e-01
7.32061148e-01 -3.33437324e-01 -1.00038314e+00 1.60912120e+00
9.88971889e-01 -6.01013824e-02 -4.85096663e-01 1.34971023e+00
1.02573311e+00 2.73102969e-01 1.54500827e-01 1.00862809e-01
1.90039992e+00 -1.13486111e+00 -5.38440168e-01 -8.18414092e-01
-7.89278671e-02 -4.62113202e-01 4.32223141e-01 3.86063427e-01
-1.02903497e+00 -7.35868335e-01 -8.76643121e-01 -1.75422847e-01
-2.01353943e-03 3.92846137e-01 5.88590741e-01 6.19436443e-01
-5.14681399e-01 7.61866868e-02 -8.68938386e-01 -2.90611297e-01
3.97869825e-01 3.99073720e-01 -4.01936203e-01 -6.30086660e-02
-9.34509397e-01 5.62367320e-01 3.30521613e-01 4.07722592e-01
-8.96963537e-01 -4.46571171e-01 -1.27247524e+00 -2.92637229e-01
8.32782805e-01 -1.16430426e+00 1.07958937e+00 -3.56257498e-01
-1.31739128e+00 6.61456466e-01 -4.85070437e-01 -2.41407409e-01
1.06767154e+00 -8.22779477e-01 4.60923202e-02 3.71360898e-01
4.64411378e-01 8.51074100e-01 1.06058693e+00 -1.21846664e+00
-8.61932695e-01 -8.59734058e-01 -1.14405014e-01 4.11786526e-01
1.64723784e-01 -5.57549484e-02 -1.28795373e+00 -6.13536775e-01
4.63995069e-01 -1.04505980e+00 -3.07441503e-01 1.08846270e-01
-8.25914919e-01 -2.82098979e-01 4.54394579e-01 -1.01526058e+00
1.00802588e+00 -1.61569846e+00 3.94970745e-01 3.38315070e-01
7.00547099e-01 -1.72170565e-01 2.58904129e-01 1.34432763e-01
3.46896112e-01 -4.63679224e-01 3.76453213e-02 -8.14346731e-01
2.19499514e-01 6.42259493e-02 4.32681851e-02 7.49657691e-01
1.09585240e-01 9.58763003e-01 -7.51041114e-01 -8.71532977e-01
2.35547811e-01 5.34866810e-01 -5.55199027e-01 2.48340592e-01
-9.05116498e-02 6.25257373e-01 -5.18605113e-01 1.03476858e+00
8.15932751e-01 -1.25584111e-01 1.20632112e-01 -3.60811085e-01
1.19269334e-01 -1.51784942e-01 -1.64142382e+00 1.76874375e+00
2.60739833e-01 6.80778772e-02 2.36854777e-01 -1.63623303e-01
6.88217044e-01 1.06051236e-01 5.16960859e-01 -1.48675069e-01
2.15745389e-01 -2.23727047e-01 -3.85629356e-01 -1.90523416e-01
5.52858055e-01 9.38450620e-02 -6.66234612e-01 -1.28688797e-01
1.41991321e-02 1.75622031e-01 3.44728291e-01 1.72629833e-01
7.40717888e-01 5.68313956e-01 4.91523117e-01 2.55426187e-02
5.77491224e-01 -1.61006212e-01 8.99201930e-01 6.79250240e-01
-4.22051191e-01 7.57607758e-01 1.67393520e-01 -5.63456059e-01
-6.71563327e-01 -1.34984696e+00 1.66633993e-01 1.38718486e+00
6.14938438e-01 -7.22042322e-01 -8.56770694e-01 -6.68539584e-01
2.42040366e-01 -7.09668100e-02 -7.64201343e-01 2.91885138e-01
-7.44033694e-01 -5.71343958e-01 5.06368637e-01 7.31325209e-01
5.77362537e-01 -3.40451211e-01 -7.19132781e-01 9.15847272e-02
-7.09409654e-01 -1.31886625e+00 -8.71518493e-01 -4.83960927e-01
-3.68707240e-01 -1.28150201e+00 -1.23348737e+00 -2.95819968e-01
7.87068605e-01 7.48151988e-02 9.89303827e-01 -5.70625179e-02
-2.86378235e-01 3.20914209e-01 -1.60273463e-01 -5.19188583e-01
3.89034897e-01 -5.18496670e-02 3.96944076e-01 1.36662960e-01
1.82641834e-01 -2.11596251e-01 -9.25839007e-01 6.02677882e-01
-8.32826272e-02 2.58033015e-02 5.79456568e-01 3.78497452e-01
7.22903192e-01 -4.90189232e-02 -2.76861340e-02 -5.07014751e-01
1.53443202e-01 -6.57384396e-02 -5.31883359e-01 1.22791961e-01
4.12583239e-02 -2.56095171e-01 6.18701167e-02 -3.22934389e-01
-1.03138900e+00 5.67564249e-01 1.01753049e-01 -3.37987632e-01
-3.58846605e-01 -2.32240915e-01 -6.26204014e-01 9.70512480e-02
4.94520336e-01 9.89698246e-02 -3.54465514e-01 -5.76758444e-01
2.52226800e-01 2.36483663e-01 1.04109454e+00 -1.00612187e+00
1.13395464e+00 7.08183348e-01 -4.51562330e-02 -7.08449602e-01
-1.05463040e+00 -6.37176871e-01 -1.07853901e+00 -4.87915248e-01
1.31585610e+00 -1.36117494e+00 -1.10125840e+00 4.33336526e-01
-1.23821938e+00 1.33882836e-01 9.94629562e-02 4.09444213e-01
-2.69550562e-01 5.42990029e-01 -5.54621279e-01 -9.60229695e-01
-3.70489717e-01 -1.04367328e+00 1.90361643e+00 2.73757994e-01
-5.35966039e-01 -3.91415060e-01 -5.33618294e-02 7.98597336e-01
-4.45180923e-01 8.21836710e-01 -1.88909605e-01 -2.89623886e-01
-5.04058719e-01 -5.09921849e-01 -1.67853180e-02 -2.86768645e-01
-2.58872867e-01 -2.82550365e-01 -9.35333967e-01 -5.11641383e-01
-5.27430475e-01 -7.70779774e-02 7.23494470e-01 4.86174971e-01
7.98373878e-01 -3.41079712e-01 -7.14726627e-01 7.23382473e-01
8.71924341e-01 -4.55209851e-01 2.26997793e-01 1.41451612e-01
1.13455677e+00 7.69096255e-01 8.20759058e-01 6.49557650e-01
9.10915911e-01 9.31925058e-01 3.10400546e-01 -2.78299823e-02
-4.07119729e-02 -6.02861404e-01 3.66038978e-01 1.27408817e-01
-5.58077395e-01 7.97053203e-02 -9.41927075e-01 2.26470590e-01
-1.94924486e+00 -8.23573649e-01 -1.21259153e-01 2.15170217e+00
5.31874597e-01 1.51840672e-01 8.04987192e-01 -9.73148048e-02
1.02296007e+00 1.25223935e-01 -5.62287986e-01 5.92920244e-01
1.29127666e-01 -3.64655882e-01 5.03761411e-01 4.45826054e-01
-1.48957539e+00 9.13181663e-01 5.87187910e+00 6.12199187e-01
-3.29304636e-01 2.49442402e-02 3.17605406e-01 -2.32117757e-01
4.49225843e-01 -1.88183427e-01 -1.36400402e+00 4.14213121e-01
-3.38613316e-02 3.01872790e-01 -5.66481724e-02 9.09697354e-01
2.66631301e-02 -4.87661064e-01 -1.15856397e+00 1.27138138e+00
2.40269884e-01 -6.86340392e-01 -1.08194657e-01 3.17439586e-01
6.53418422e-01 -4.77058947e-01 -2.65144676e-01 3.25795710e-01
2.66019970e-01 -8.76803219e-01 1.16667986e+00 7.18011379e-01
5.31419516e-01 -9.93341684e-01 6.37966990e-01 6.30483150e-01
-1.83702993e+00 1.80245861e-02 -3.54024857e-01 -3.58623594e-01
2.94626474e-01 4.35180902e-01 -7.91084528e-01 6.55782938e-01
8.17107320e-01 3.77252191e-01 -7.39775360e-01 1.05781984e+00
-4.72909898e-01 1.77825436e-01 -5.10122776e-01 2.51964718e-01
-2.58598983e-01 4.81012352e-02 6.87410116e-01 1.19338298e+00
2.96518784e-02 8.93342718e-02 1.03166962e+00 7.44194925e-01
1.72255874e-01 -3.23859453e-02 3.57789434e-02 6.82185769e-01
5.18083155e-01 1.43937552e+00 -7.68643439e-01 -4.53438759e-01
-6.05727546e-02 1.22797310e+00 2.84445256e-01 2.48567045e-01
-1.05921102e+00 -6.75261542e-02 8.12824488e-01 4.00634527e-01
2.62254685e-01 -3.52493554e-01 -2.20325566e-03 -1.23921371e+00
3.68332118e-01 -7.01422751e-01 6.39172673e-01 -6.35458410e-01
-1.13500953e+00 2.82272726e-01 3.89181942e-01 -1.11378586e+00
-3.08656693e-01 -3.38648796e-01 -3.89267594e-01 8.72288942e-01
-8.70936394e-01 -1.58140886e+00 -6.41654789e-01 6.75569654e-01
4.40531641e-01 2.83688635e-01 3.62445295e-01 5.73951192e-02
-8.60018134e-01 7.45258093e-01 -8.52274477e-01 6.65317535e-01
7.63440371e-01 -1.29451513e+00 3.38712066e-01 9.67452884e-01
-1.67952746e-01 8.14772367e-01 8.76700401e-01 -1.20372081e+00
-1.34105635e+00 -1.14371872e+00 6.69853508e-01 -9.19177830e-01
7.16516450e-02 -7.16180623e-01 -3.56602758e-01 6.82149589e-01
-3.00600171e-01 9.63695198e-02 5.02389491e-01 1.65503651e-01
-3.98996562e-01 -6.35905787e-02 -9.77847338e-01 5.88928163e-01
1.32951343e+00 -9.98933315e-02 -7.96192110e-01 2.22588435e-01
5.04780233e-01 -9.33943212e-01 -8.85422945e-01 4.33125079e-01
6.53491557e-01 -8.86617839e-01 1.45491505e+00 -1.84499756e-01
1.05542250e-01 -7.98478663e-01 -2.71639917e-02 -7.76678860e-01
-3.93211901e-01 -5.56026816e-01 -4.91050750e-01 1.17585921e+00
-1.39760166e-01 -3.02824557e-01 8.85427117e-01 7.80271828e-01
1.54013306e-01 -4.56145316e-01 -9.13195133e-01 -6.91370606e-01
-4.79916453e-01 -3.86041611e-01 7.75860488e-01 3.70854735e-01
-2.10780665e-01 4.75806594e-01 -6.86389267e-01 6.38003290e-01
1.13339710e+00 1.63569435e-01 1.51018333e+00 -1.31682622e+00
-4.23551500e-01 -1.28325149e-01 -5.55591226e-01 -1.44371521e+00
-1.68625370e-01 -2.49001861e-01 3.36351097e-01 -1.61273527e+00
4.53618407e-01 6.50069341e-02 3.88856530e-01 5.17550647e-01
-6.71133041e-01 5.21617770e-01 4.91681367e-01 2.77916282e-01
-1.06001174e+00 4.74164724e-01 1.40444875e+00 -1.79848507e-01
-2.13375734e-03 9.03544948e-02 -6.45270586e-01 1.23907185e+00
2.52178520e-01 -2.37572730e-01 -1.10076465e-01 -2.12911859e-01
1.44140095e-01 1.07430838e-01 9.49466765e-01 -1.48147476e+00
3.83404583e-01 -1.78184547e-03 1.24774039e+00 -1.16116011e+00
8.16242576e-01 -5.54915547e-01 7.31301606e-02 5.54288566e-01
1.69991106e-01 -1.41349539e-01 4.92058769e-02 8.11829269e-01
9.49161053e-02 4.34301943e-01 5.25723279e-01 -1.79661199e-01
-7.18513608e-01 6.67098343e-01 1.05123796e-01 1.95141897e-01
1.08450973e+00 -4.63055611e-01 2.99554635e-02 -2.99285889e-01
-9.26281512e-01 5.69915116e-01 4.47762907e-01 4.75783646e-01
5.18548608e-01 -1.32975912e+00 -8.22971940e-01 8.33972469e-02
3.05775017e-01 5.53390980e-01 3.19394559e-01 1.00182223e+00
-2.77402043e-01 1.77271366e-01 1.16540238e-01 -9.66630220e-01
-1.58313525e+00 3.62805635e-01 3.75954986e-01 -1.31235406e-01
-5.89918971e-01 1.05720937e+00 4.99300987e-01 -4.01488483e-01
2.53293097e-01 -2.05450207e-01 -1.85797766e-01 2.33626604e-01
7.99168766e-01 7.27216721e-01 -3.67046654e-01 -1.24779308e+00
-7.65734255e-01 1.01103282e+00 1.72746047e-01 -1.74865887e-01
1.00270998e+00 -2.40166128e-01 5.28710186e-02 1.47433402e-02
8.03805590e-01 3.26123506e-01 -1.70866752e+00 -3.07013512e-01
-3.96231443e-01 -6.58192813e-01 -5.57070553e-01 -5.18099010e-01
-8.20046723e-01 6.25612378e-01 2.63080716e-01 -4.40660536e-01
8.26558948e-01 2.34008968e-01 8.74566972e-01 2.90297151e-01
7.01948345e-01 -1.26578987e+00 2.11793080e-01 4.30154532e-01
9.83421922e-01 -1.12378299e+00 4.56653297e-01 -8.12603652e-01
-5.39112866e-01 8.08075666e-01 9.61762011e-01 -1.45541862e-01
1.90071389e-01 1.79710880e-01 -1.54262811e-01 -3.38742316e-01
-1.71792470e-02 -4.65680718e-01 8.15134227e-01 6.79170668e-01
1.86591789e-01 3.33843291e-01 1.57778163e-03 8.82752478e-01
-5.09512126e-01 -3.49604756e-01 -2.66957134e-01 7.89351523e-01
-5.72611213e-01 -7.31464803e-01 -1.31653547e+00 2.41729081e-01
-2.66213387e-01 3.61753464e-01 -5.90294898e-01 7.34431028e-01
6.64323628e-01 9.32634652e-01 -6.11081980e-02 -4.50908989e-01
5.17209649e-01 -9.32798013e-02 5.69736600e-01 -6.28283143e-01
-4.43737537e-01 4.86296624e-01 1.13898546e-01 -8.54558170e-01
-2.94885457e-01 -9.52913761e-01 -1.25058246e+00 -2.19492197e-01
-1.26889408e-01 -6.12135343e-02 1.01294354e-01 9.75083470e-01
8.97982344e-02 4.39782530e-01 -3.12647820e-02 -1.58882082e+00
-2.77993888e-01 -8.96076262e-01 -3.63293409e-01 5.45181990e-01
3.11196685e-01 -1.09266686e+00 1.03740923e-01 3.22247595e-02] | [7.113887310028076, -0.7899867296218872] |
472cb65b-b0ee-41a6-8f29-308e697f0b2f | learning-a-dilated-residual-network-for-sar | 1709.02898 | null | http://arxiv.org/abs/1709.02898v3 | http://arxiv.org/pdf/1709.02898v3.pdf | Learning a Dilated Residual Network for SAR Image Despeckling | In this paper, to break the limit of the traditional linear models for
synthetic aperture radar (SAR) image despeckling, we propose a novel deep
learning approach by learning a non-linear end-to-end mapping between the noisy
and clean SAR images with a dilated residual network (SAR-DRN). SAR-DRN is
based on dilated convolutions, which can both enlarge the receptive field and
maintain the filter size and layer depth with a lightweight structure. In
addition, skip connections and residual learning strategy are added to the
despeckling model to maintain the image details and reduce the vanishing
gradient problem. Compared with the traditional despeckling methods, the
proposed method shows superior performance over the state-of-the-art methods on
both quantitative and visual assessments, especially for strong speckle noise. | ['Qiang Zhang', 'Xiaoshuang Ma', 'Jie Li', 'Zhen Yang', 'Qiangqiang Yuan'] | 2017-09-09 | null | null | null | null | ['sar-image-despeckling'] | ['computer-vision'] | [ 3.51578832e-01 -2.05342412e-01 6.81633711e-01 -5.46418071e-01
-6.06711924e-01 -3.09306741e-01 4.30173308e-01 -6.48944199e-01
-5.18495142e-01 5.31678617e-01 3.42688829e-01 -2.40832582e-01
-3.86133879e-01 -5.99777579e-01 -4.98587072e-01 -9.05676842e-01
6.62699193e-02 -3.45351815e-01 1.22938141e-01 -1.77579463e-01
1.45288885e-01 7.32898474e-01 -9.66530979e-01 3.37774545e-01
8.81450951e-01 1.16831481e+00 4.62344557e-01 4.63488311e-01
2.83269286e-01 1.07712114e+00 -3.09887677e-01 -3.56190801e-02
6.49528384e-01 -2.60700792e-01 -2.33329892e-01 -2.11969148e-02
7.76651561e-01 -5.90421677e-01 -9.29934800e-01 1.26609230e+00
9.33033288e-01 1.95835516e-01 1.89247042e-01 -3.05370748e-01
-1.14645064e+00 5.42157352e-01 -7.37806857e-01 3.65543544e-01
-4.26904589e-01 1.96036920e-01 3.12157273e-01 -1.09109437e+00
2.93233514e-01 1.21501422e+00 9.75522041e-01 4.61694032e-01
-1.17017472e+00 -6.94390774e-01 -2.43887696e-02 1.67455133e-02
-1.25990820e+00 -6.30691171e-01 6.85317159e-01 -3.14764917e-01
6.47684813e-01 2.60204494e-01 3.58317107e-01 6.98406041e-01
4.44075882e-01 3.57703596e-01 1.36518753e+00 -2.57906616e-01
-1.28274634e-02 -4.46802378e-01 8.32209438e-02 5.14918208e-01
2.02808097e-01 7.04467773e-01 7.06217065e-03 3.03173006e-01
1.13239670e+00 3.53909492e-01 -5.40805817e-01 -2.50547349e-01
-1.34928238e+00 6.70196831e-01 9.85472500e-01 3.41423541e-01
-4.54536170e-01 5.72118945e-02 7.81616494e-02 4.92531478e-01
5.34593403e-01 5.89915693e-01 -2.22497568e-01 6.50882542e-01
-1.35865450e+00 1.34973437e-01 2.25946605e-01 3.52148294e-01
5.97989082e-01 6.01677060e-01 -5.01807272e-01 7.01345623e-01
2.54251599e-01 7.45680809e-01 3.59615773e-01 -8.20233583e-01
2.77702749e-01 3.33995908e-01 1.87210903e-01 -9.65463698e-01
-4.35131550e-01 -1.14581406e+00 -1.55060315e+00 8.02101851e-01
6.24468252e-02 -2.44361550e-01 -1.41743743e+00 1.39710128e+00
-1.49121791e-01 4.01395231e-01 6.48210227e-01 1.20261991e+00
1.12350869e+00 6.32978201e-01 -2.40137056e-01 -1.75260156e-01
8.72620642e-01 -1.12678456e+00 -7.49407053e-01 -8.05490732e-01
-2.18442013e-03 -9.70534921e-01 7.36918867e-01 1.74632579e-01
-1.05366254e+00 -8.77129734e-01 -1.20834672e+00 -1.12705067e-01
7.86811300e-03 5.91262639e-01 4.63738710e-01 1.95761919e-01
-9.72444773e-01 6.61222339e-01 -7.17091680e-01 1.44962639e-01
5.35761595e-01 7.71470368e-02 -2.41627723e-01 -4.72560197e-01
-1.09887445e+00 8.92221689e-01 2.14942232e-01 9.20829117e-01
-9.93009686e-01 -6.55927896e-01 -8.82882953e-01 2.75105596e-01
8.74662679e-03 -6.15417600e-01 8.13722312e-01 -9.56076205e-01
-1.47275734e+00 6.52192175e-01 2.92555302e-01 -6.12837255e-01
4.98940438e-01 -5.52158833e-01 -5.83718836e-01 2.02294160e-02
-2.56752610e-01 4.84237671e-01 1.19656312e+00 -1.18476474e+00
-6.92713708e-02 -2.27107212e-01 -1.14436127e-01 1.98332787e-01
2.01635450e-01 -7.32125863e-02 1.01455830e-01 -1.11803770e+00
5.28543234e-01 -3.48819882e-01 -6.46397650e-01 2.39595529e-02
-1.27010912e-01 5.96657813e-01 1.08973563e+00 -1.02430427e+00
1.02238798e+00 -2.43603826e+00 4.50947806e-02 -2.95500960e-02
4.09623295e-01 8.72100294e-01 -5.02725303e-01 9.06089437e-04
-3.34843755e-01 -4.16148484e-01 -5.60444057e-01 5.41656315e-02
-4.18451250e-01 -2.28657439e-01 -4.46889192e-01 6.18347585e-01
1.85180288e-02 1.17817461e+00 -6.43914461e-01 -2.73403898e-02
4.29158241e-01 6.14277363e-01 -6.16155937e-02 3.17295551e-01
1.45258486e-01 6.67663276e-01 -2.70996243e-01 4.74979252e-01
1.42863107e+00 -8.15618038e-02 -2.29391873e-01 -6.03834867e-01
-2.77157724e-01 -2.92176515e-01 -1.12101853e+00 1.44264936e+00
-5.14660537e-01 6.58489227e-01 5.11811793e-01 -1.03180158e+00
1.26970279e+00 -4.57738042e-02 5.44398427e-02 -8.93318713e-01
9.98497233e-02 2.75144249e-01 -4.92341071e-03 -4.27923769e-01
1.93360403e-01 -3.21926117e-01 2.77058125e-01 -4.02162187e-02
-3.09841722e-01 -3.83935310e-02 -4.15915221e-01 -1.88743263e-01
9.90224302e-01 -3.31846885e-02 2.12932721e-01 -2.61506438e-01
7.40950525e-01 -2.57290989e-01 6.27583683e-01 7.72436917e-01
1.51243880e-01 8.74470055e-01 -1.95598841e-01 -7.97814250e-01
-9.89374876e-01 -1.29436362e+00 -1.57139122e-01 3.78168911e-01
2.29078427e-01 3.64066929e-01 -5.63080132e-01 -4.43211079e-01
-3.73680174e-01 4.28632140e-01 -2.85302401e-01 -1.39759928e-01
-7.71930397e-01 -7.50594556e-01 5.44388175e-01 3.69823277e-01
1.38007581e+00 -1.05661833e+00 -5.09463787e-01 2.59794742e-01
-1.00058578e-01 -1.23502600e+00 -5.47717512e-01 1.16270915e-01
-9.69373584e-01 -8.55251491e-01 -8.73808146e-01 -1.16329765e+00
8.16138268e-01 5.88962853e-01 8.45025778e-01 -1.61593571e-01
-4.23681736e-01 -3.91757786e-01 -3.76585484e-01 1.79816745e-02
-1.49049148e-01 -4.05674487e-01 -2.72460490e-01 3.77529487e-02
2.79087611e-02 -6.52627766e-01 -9.62183595e-01 1.44858718e-01
-1.18573833e+00 1.13257200e-01 1.30833197e+00 1.30650580e+00
4.46385324e-01 3.16632628e-01 4.39470619e-01 -5.49558401e-01
4.20764655e-01 -3.74469720e-02 -8.83757412e-01 2.97646374e-01
-4.72266048e-01 9.30984765e-02 8.07600737e-01 -3.72131586e-01
-1.34784174e+00 1.41597107e-01 -8.27914327e-02 -5.24356186e-01
6.15543276e-02 5.96675158e-01 2.63936874e-02 -6.18015289e-01
7.72909522e-01 4.12123233e-01 1.28219575e-01 -7.42437959e-01
2.51708716e-01 6.97418272e-01 1.03434348e+00 3.84293079e-01
1.13496292e+00 5.44663846e-01 -7.40712658e-02 -7.33613908e-01
-1.13382578e+00 -2.24093840e-01 -5.81165433e-01 3.96783762e-02
8.31349254e-01 -1.23742330e+00 -3.43864083e-01 7.76292264e-01
-1.00241911e+00 -3.96794945e-01 -3.29304755e-01 8.82870555e-01
-3.15907568e-01 6.06699288e-01 -7.07217872e-01 -5.34172535e-01
-7.01544225e-01 -8.62109959e-01 8.40374947e-01 4.06804383e-01
6.04943156e-01 -6.35978341e-01 -1.31725490e-01 1.93964705e-01
1.01486504e+00 2.30128467e-01 6.33689046e-01 -6.46342430e-03
-7.24348366e-01 -1.21615663e-01 -7.06704080e-01 1.00587988e+00
-6.61550015e-02 -8.99011731e-01 -8.19397092e-01 -8.41562808e-01
4.10318732e-01 -1.83855802e-01 1.41844368e+00 7.92516172e-01
1.04781532e+00 -3.89873058e-01 1.32733017e-01 1.27238691e+00
1.83623481e+00 1.89800262e-02 1.27229333e+00 4.38952476e-01
4.48498458e-01 3.72926891e-01 4.53440517e-01 1.49264276e-01
-3.37703675e-01 2.34414920e-01 4.75967884e-01 -8.80436420e-01
-5.85526049e-01 -6.37806803e-02 3.86183381e-01 4.43366170e-01
1.50182977e-01 -1.20606132e-01 -4.98442948e-01 3.71006489e-01
-1.74283409e+00 -1.05165756e+00 -1.92162357e-02 2.06314468e+00
4.75993186e-01 -5.22965565e-02 -9.20675159e-01 -1.74376994e-01
6.97611392e-01 6.88367903e-01 -6.33417666e-01 -2.11860463e-02
-5.82356036e-01 3.87918830e-01 9.86273050e-01 6.40162408e-01
-1.30999315e+00 1.05829895e+00 6.81143093e+00 8.93729627e-01
-1.51579916e+00 -5.46533987e-02 5.43695152e-01 1.89860582e-01
-1.31200943e-02 -5.73481880e-02 -4.52819556e-01 1.87261581e-01
4.50165212e-01 3.48740816e-01 6.41809404e-01 5.99213421e-01
4.10615772e-01 1.46296084e-01 -4.73109931e-01 1.13170695e+00
1.38933107e-01 -1.56776726e+00 -8.03253353e-02 -2.59160608e-01
8.68246675e-01 3.07034433e-01 1.15976110e-01 1.82295963e-01
2.14116499e-01 -1.31345427e+00 3.08428913e-01 9.37624514e-01
1.02689350e+00 -5.85219383e-01 1.18345273e+00 2.30971575e-01
-9.23015296e-01 -2.94268876e-01 -6.93825066e-01 -6.04775697e-02
2.15178281e-01 9.18807030e-01 -1.14954963e-01 6.84109688e-01
7.44194865e-01 6.94183707e-01 -6.37334704e-01 1.10703003e+00
-1.89851359e-01 4.07257825e-01 8.61090571e-02 6.34438694e-01
4.50021058e-01 -2.81130373e-01 6.84688985e-01 1.19284654e+00
5.76453626e-01 2.01423600e-01 2.03669101e-01 1.00576651e+00
1.19668178e-01 -4.02027607e-01 -5.72119057e-01 5.31191118e-02
1.43343806e-01 1.40345204e+00 -3.90949339e-01 -1.29595414e-01
-2.70001262e-01 1.00258911e+00 -1.07025869e-01 6.02832913e-01
-5.60721636e-01 -6.62550271e-01 3.67901772e-01 1.59203872e-01
6.40889227e-01 -3.51846159e-01 -2.22967327e-01 -1.12773788e+00
9.58912298e-02 -9.46544945e-01 -1.02964723e-02 -1.01245332e+00
-1.43259108e+00 1.06160855e+00 -2.72081017e-01 -1.56907165e+00
1.89722374e-01 -5.80423594e-01 -6.84366524e-01 1.17719114e+00
-2.06594133e+00 -1.41190374e+00 -6.74963534e-01 5.13587534e-01
3.55808318e-01 -4.92974341e-01 5.19258976e-01 2.67410129e-01
-9.75198373e-02 2.02136502e-01 4.04567152e-01 2.01553643e-01
6.43862963e-01 -6.75690889e-01 6.43383324e-01 1.43046427e+00
-4.58168775e-01 4.54249859e-01 6.41602516e-01 -6.33310914e-01
-9.69226122e-01 -1.38467431e+00 6.49269998e-01 4.25909698e-01
3.36950392e-01 -9.24300924e-02 -9.74617004e-01 5.76188922e-01
3.50248128e-01 4.53299254e-01 7.67446607e-02 -5.92199445e-01
-3.45184654e-01 -4.16775405e-01 -1.20524859e+00 3.32793832e-01
7.78271914e-01 -2.66951621e-01 -7.17288136e-01 1.43453106e-01
6.72066748e-01 -4.72582668e-01 -3.78316939e-01 9.24870670e-01
3.71555090e-01 -1.00674963e+00 1.23568738e+00 -1.30780295e-01
3.62678379e-01 -7.12733746e-01 -2.56087422e-01 -1.33361959e+00
-9.17572498e-01 -6.74211919e-01 1.24887988e-01 8.03454340e-01
1.44856036e-01 -5.28743327e-01 6.26643896e-01 -2.14210972e-01
-4.52209294e-01 -5.69315434e-01 -7.76599526e-01 -8.25913191e-01
-1.01052053e-01 1.90394610e-01 3.35340649e-01 7.76404500e-01
-8.11217070e-01 3.70868325e-01 -8.80892754e-01 5.47519803e-01
9.12168920e-01 2.55207419e-01 2.66309857e-01 -1.10247779e+00
-3.03098500e-01 -6.05480038e-02 -3.94864559e-01 -1.34625649e+00
-4.08944897e-02 -8.45245898e-01 2.72412509e-01 -1.71589589e+00
-2.26892866e-02 -2.50439197e-01 -3.06408525e-01 3.02614272e-01
8.02042242e-03 5.84095299e-01 8.67982879e-02 3.76807749e-01
-2.87048936e-01 5.77365160e-01 1.53478765e+00 -2.61213392e-01
-2.06286892e-01 6.45338520e-02 -6.74628794e-01 7.95556247e-01
7.72879064e-01 -2.54799783e-01 -1.25845090e-01 -9.52284157e-01
-3.99720892e-02 2.02207893e-01 7.68141329e-01 -1.24603832e+00
4.53013718e-01 2.54877545e-02 8.38163972e-01 -7.72780180e-01
1.76472694e-01 -7.80570626e-01 8.58528465e-02 6.11076593e-01
-3.37746918e-01 -7.92366937e-02 1.24816984e-01 6.45789623e-01
-5.62707782e-01 -2.69035567e-02 1.40401351e+00 -2.28886560e-01
-8.07751179e-01 2.93361276e-01 -1.86015487e-01 -2.20651582e-01
7.42564797e-01 -3.00070614e-01 -4.47737545e-01 -3.01077157e-01
-7.61382103e-01 9.24955010e-02 5.16343452e-02 1.47727579e-01
1.15444231e+00 -1.14804900e+00 -1.07959020e+00 7.30347872e-01
-2.91033059e-01 -3.06826867e-02 9.10585523e-01 8.83374393e-01
-9.03123438e-01 1.70466781e-01 -5.20863116e-01 -1.98679507e-01
-1.14691126e+00 4.36439872e-01 7.22147942e-01 -4.84253287e-01
-1.08564270e+00 6.92028999e-01 3.71832132e-01 -4.61488366e-01
2.84668714e-01 -5.58129586e-02 -2.57385820e-01 -4.90120500e-01
7.97178030e-01 2.24164352e-01 7.08095497e-03 -4.88893449e-01
-1.51761353e-01 7.29775310e-01 -1.27274379e-01 8.97979736e-02
1.70848274e+00 -3.42996240e-01 -3.64135712e-01 -1.88622072e-01
9.07376528e-01 1.49779290e-01 -1.49558115e+00 -6.04754210e-01
-3.95450771e-01 -6.68125212e-01 7.05199480e-01 -1.16899371e+00
-1.31045568e+00 8.20145190e-01 1.24515557e+00 -2.89259523e-01
1.53511775e+00 -4.18170750e-01 1.01940167e+00 5.26740670e-01
-2.55950183e-01 -7.20286846e-01 -5.65133663e-03 6.71243787e-01
1.10542190e+00 -1.18202603e+00 2.16256887e-01 -2.64632195e-01
-5.45808673e-01 1.33305466e+00 4.57871795e-01 -5.50634623e-01
7.54194558e-01 4.59608406e-01 4.67911780e-01 -3.13968927e-01
-2.52179295e-01 -2.38758504e-01 2.48623550e-01 6.28061533e-01
1.85777754e-01 -1.70146525e-01 -3.75298381e-01 4.22360301e-01
2.98026390e-02 8.93773809e-02 4.12040532e-01 5.09078681e-01
-7.36121655e-01 -7.05134094e-01 -4.09824371e-01 3.45783085e-01
-4.69528615e-01 -5.72768211e-01 6.38825670e-02 4.91348326e-01
1.94522381e-01 8.96331251e-01 4.84738909e-02 -3.62511575e-01
4.87312496e-01 -6.12698197e-01 2.62467355e-01 -3.16222519e-01
-5.42941034e-01 3.15131247e-01 -2.77408510e-01 -3.85459960e-01
-4.74983484e-01 -6.03913628e-02 -8.80549550e-01 -1.57392681e-01
-3.08581471e-01 -1.94902322e-03 3.84014159e-01 8.22975159e-01
3.54528010e-01 7.50553846e-01 7.71063149e-01 -8.34511161e-01
-9.26644862e-01 -1.16242337e+00 -8.12474012e-01 5.74470498e-02
7.75327802e-01 -1.54387191e-01 -5.81566095e-01 -6.63058907e-02] | [10.492218971252441, -2.241219997406006] |
a75ff156-ffaf-41f7-ab78-9ec94d3ac3d0 | transcending-traditional-boundaries | 2306.14374 | null | https://arxiv.org/abs/2306.14374v1 | https://arxiv.org/pdf/2306.14374v1.pdf | Transcending Traditional Boundaries: Leveraging Inter-Annotator Agreement (IAA) for Enhancing Data Management Operations (DMOps) | This paper presents a novel approach of leveraging Inter-Annotator Agreement (IAA), traditionally used for assessing labeling consistency, to optimize Data Management Operations (DMOps). We advocate for the use of IAA in predicting the labeling quality of individual annotators, leading to cost and time efficiency in data production. Additionally, our work highlights the potential of IAA in forecasting document difficulty, thereby boosting the data construction process's overall efficiency. This research underscores IAA's broader application potential in data-driven research optimization and holds significant implications for large-scale data projects prioritizing efficiency, cost reduction, and high-quality data. | ['Harksoo Kim', 'Chanjun Park', 'NamHyeok Kim', 'Damrin Kim'] | 2023-06-26 | null | null | null | null | ['management'] | ['miscellaneous'] | [-2.84505755e-01 3.33542049e-01 -5.85666120e-01 -3.80061835e-01
-5.54264605e-01 -8.29662561e-01 8.15297514e-02 8.87175918e-01
-4.23486322e-01 3.92657220e-01 4.72315609e-01 -4.05437201e-01
-6.70398295e-01 -4.39340562e-01 -1.01592667e-01 -9.19692293e-02
3.98874044e-01 5.01480937e-01 -5.44689357e-01 2.67498583e-01
8.17529976e-01 7.15624213e-01 -1.50604987e+00 5.28560102e-01
1.02790105e+00 6.41901135e-01 1.94044694e-01 2.80267805e-01
-4.41361457e-01 9.60315108e-01 -9.39664245e-01 -7.81459272e-01
3.33246171e-01 2.30737180e-01 -1.33428264e+00 8.71176720e-02
3.77532393e-01 2.47458264e-01 3.29483539e-01 8.78278315e-01
3.76918197e-01 -2.65256673e-01 3.63105536e-01 -1.47480416e+00
-1.13420057e+00 6.82208776e-01 -3.38945508e-01 3.06970984e-01
3.67720991e-01 2.15102360e-02 1.45187128e+00 -8.89741838e-01
8.20363462e-01 8.69277060e-01 8.66013885e-01 4.76147681e-02
-1.10613167e+00 -6.00675642e-01 -1.82465255e-01 2.76715130e-01
-1.51189768e+00 -5.13028800e-01 2.26876810e-01 -1.05705798e+00
9.78323221e-01 5.10269761e-01 5.06878853e-01 1.92974895e-01
-3.26831010e-03 4.80853260e-01 1.37397337e+00 -9.66924131e-01
1.86718702e-01 5.12641609e-01 5.56029260e-01 6.27803087e-01
8.65865171e-01 -6.18205369e-01 -7.92635381e-01 -2.14252695e-01
2.60293275e-01 -3.70384336e-01 4.97508533e-02 4.53027412e-02
-8.58188808e-01 4.90636557e-01 -8.55687484e-02 5.50256908e-01
-4.91417795e-01 -3.31886679e-01 3.84943306e-01 3.75522166e-01
6.88614607e-01 1.51274157e+00 -4.81729388e-01 -5.80135822e-01
-8.65219712e-01 1.83297068e-01 9.78118002e-01 1.29837608e+00
5.67457259e-01 -3.50846916e-01 -6.50363266e-01 9.85962391e-01
4.80607748e-01 1.26215279e-01 4.42698896e-02 -1.43318009e+00
4.35682744e-01 1.14236403e+00 5.20840526e-01 -9.40307140e-01
-6.46233320e-01 -3.49322081e-01 1.51383167e-03 -5.48274033e-02
1.98859647e-01 3.74331713e-01 -5.40119231e-01 1.23710275e+00
-5.25468327e-02 -8.08018088e-01 -1.69884600e-02 6.53418601e-01
5.42588413e-01 1.99534029e-01 3.72298509e-01 -4.23090667e-01
1.57912171e+00 -7.67766416e-01 -1.24257922e+00 9.17911530e-06
1.18406248e+00 -9.90039885e-01 1.05731320e+00 6.30373597e-01
-1.30730581e+00 -6.80041909e-01 -6.66314900e-01 -4.45119172e-01
-5.00160277e-01 2.71512210e-01 6.78267062e-01 6.49348557e-01
-8.49313378e-01 3.99941385e-01 -5.29188991e-01 -6.08349979e-01
6.61455512e-01 -3.79423611e-02 -1.79667056e-01 -1.81571737e-01
-6.16372526e-01 1.55332255e+00 5.16032040e-01 2.71749273e-02
-2.39614189e-01 -9.57277775e-01 -4.37488377e-01 2.11258698e-02
2.42405429e-01 -2.93937474e-01 1.09898484e+00 -2.57537156e-01
-6.88663483e-01 9.81130064e-01 -1.47370964e-01 3.69013585e-02
-5.58147170e-02 -2.92905420e-01 -7.21503556e-01 -4.26396549e-01
3.93724561e-01 2.97493368e-01 -3.94963026e-01 -1.39377081e+00
-8.12767625e-01 -5.87853253e-01 -1.57698587e-01 3.56588423e-01
-8.94241512e-01 5.07941127e-01 -2.88249612e-01 -3.99399638e-01
3.95448785e-03 -8.35309148e-01 1.29774854e-01 -2.16515541e-01
-8.70347694e-02 -7.98320174e-01 3.42560619e-01 -1.02390397e+00
1.97112298e+00 -1.81802964e+00 1.19755998e-01 3.83160174e-01
3.82924080e-01 4.54826832e-01 -9.51226056e-02 7.79241681e-01
2.82600552e-01 7.51248956e-01 4.44267720e-01 -2.10036635e-01
7.21150711e-02 2.07497671e-01 6.22468814e-02 2.16469377e-01
1.65211290e-01 8.31510603e-01 -9.56715286e-01 -6.26833975e-01
-1.41029894e-01 -2.69442797e-01 -2.60900676e-01 5.52479088e-01
-8.66934564e-03 2.15298474e-01 -8.90559852e-02 8.89706671e-01
3.82485241e-01 -4.12722707e-01 4.86138523e-01 -3.71495008e-01
-9.00496840e-01 3.01638842e-01 -8.98342133e-01 1.43705118e+00
-3.80254358e-01 8.60871911e-01 4.53585684e-02 -5.95506430e-01
1.34810352e+00 2.81294107e-01 4.55126494e-01 -7.74421811e-01
-1.61325261e-01 2.68413544e-01 1.20285921e-01 -1.02502024e+00
8.90817344e-01 4.40406621e-01 -2.67314948e-02 7.65326381e-01
-5.91421761e-02 -1.73589718e-02 6.11186624e-01 9.84829664e-02
1.02713609e+00 -1.32050991e-01 3.54484797e-01 -7.15678096e-01
1.61852643e-01 4.88812506e-01 5.40268898e-01 5.21271944e-01
-5.54914549e-02 -5.35611548e-02 4.64364767e-01 -4.79800105e-01
-1.59562624e+00 -1.94759995e-01 -4.22984123e-01 1.14280558e+00
-2.73047030e-01 -6.19955182e-01 -5.98466218e-01 -6.99509382e-01
1.86135456e-01 1.07636154e+00 -4.87393439e-01 1.79353297e-01
-1.41627058e-01 -4.31142390e-01 6.31944954e-01 5.51825941e-01
-4.45066914e-02 -7.99417377e-01 -4.48723823e-01 1.46433473e-01
-3.41284961e-01 -9.55182850e-01 -1.31119922e-01 3.33505822e-03
-5.88894069e-01 -1.08658600e+00 -4.05114144e-02 -4.50483561e-01
7.76228726e-01 5.83429754e-01 1.31842899e+00 4.98105913e-01
-2.44653031e-01 3.62605959e-01 -6.58741057e-01 -8.65485370e-01
-5.76626301e-01 2.91519254e-01 -7.19970884e-03 -8.40869427e-01
9.78542924e-01 5.51567450e-02 -1.28926530e-01 6.07648253e-01
-7.30082929e-01 5.76270521e-02 3.61335814e-01 5.25053203e-01
3.80977571e-01 3.12878698e-01 6.86417341e-01 -1.10109282e+00
1.17528975e+00 -5.34474552e-01 -5.98297000e-01 1.10294974e+00
-1.84801626e+00 -9.70848501e-02 3.64778936e-01 2.95800984e-01
-1.10073113e+00 -6.34384155e-01 4.47715670e-01 6.35850430e-02
-1.97688825e-02 9.83923912e-01 9.25087333e-02 -1.85088031e-02
6.41009808e-01 -8.32401693e-01 9.45379212e-02 -6.31133378e-01
2.33496562e-01 1.00859809e+00 4.23832208e-01 -6.37812436e-01
3.67436796e-01 9.76260155e-02 -2.45127454e-01 -1.79177552e-01
-1.14215112e+00 -9.35166657e-01 -8.46321762e-01 -4.78575468e-01
8.32059562e-01 -8.41771066e-01 -6.56983912e-01 -7.68036097e-02
-1.17688978e+00 1.85487643e-01 -3.49649519e-01 4.08419818e-01
-1.47298491e-02 8.15528631e-02 -2.05579311e-01 -8.00326645e-01
-3.44313025e-01 -1.11924231e+00 6.36202335e-01 1.68735087e-01
-8.72595549e-01 -9.37549412e-01 2.74433613e-01 1.30635560e+00
2.44422734e-01 -1.74484290e-02 1.11544132e+00 -7.83910453e-01
-3.22336942e-01 -2.18497276e-01 -4.65824544e-01 3.28761160e-01
9.21551958e-02 4.88929570e-01 -7.10625470e-01 -2.02696800e-01
-3.66009682e-01 -3.59527022e-01 -1.23407051e-01 -5.91151714e-02
1.43036687e+00 -4.02444959e-01 -1.06167935e-01 -2.08905954e-02
1.38125658e+00 2.35272914e-01 2.84437180e-01 7.41756618e-01
8.80903482e-01 1.17958832e+00 1.22341728e+00 7.53030658e-01
7.15433121e-01 7.46582270e-01 8.81321281e-02 -1.26320094e-01
-1.56080082e-01 8.84982571e-02 -2.83075422e-01 1.38820469e+00
-4.15205598e-01 -4.31568295e-01 -1.51871073e+00 5.12413085e-01
-1.96211827e+00 -7.14131057e-01 -6.42439425e-01 1.93195105e+00
7.58773208e-01 -1.56607106e-01 8.72392021e-03 1.76503330e-01
6.00708663e-01 -2.96650797e-01 -7.40581155e-02 -9.73596931e-01
3.68824713e-02 -1.66025874e-03 8.60609591e-01 1.11395150e-01
-3.81225765e-01 5.09360492e-01 7.22296476e+00 3.38264644e-01
-2.52114475e-01 2.31727824e-01 5.02007961e-01 1.41631648e-01
-7.29260087e-01 2.52165377e-01 -8.51668239e-01 4.24904168e-01
1.13887393e+00 -5.61428189e-01 5.47200382e-01 9.43667352e-01
4.88655150e-01 -1.88206032e-01 -1.18595743e+00 4.62989479e-01
-1.18734399e-02 -1.75529385e+00 -2.33475760e-01 3.32146734e-01
9.49011445e-01 -3.57195854e-01 -2.61293203e-01 -7.99285714e-04
7.57688642e-01 -8.94283295e-01 9.02824223e-01 8.17239285e-01
4.23101693e-01 -7.56889880e-01 1.13402796e+00 2.49252379e-01
-7.15382278e-01 -2.90921956e-01 -3.54517788e-01 -1.46442160e-01
8.15059841e-02 7.62590110e-01 -1.30604708e+00 6.82681382e-01
7.85536110e-01 4.07592684e-01 -7.44308829e-01 7.75729179e-01
1.23786569e-01 3.54815185e-01 3.64660382e-01 2.36625105e-01
-9.98644009e-02 -2.39649326e-01 -3.59791219e-02 1.69058430e+00
2.36555919e-01 4.72542532e-02 2.28352457e-01 7.36902297e-01
-2.13875622e-01 2.63252825e-01 -5.29140294e-01 -4.73565340e-01
1.53602147e+00 1.31905520e+00 -5.80614030e-01 -2.11968467e-01
-4.29124117e-01 1.98374644e-01 6.87992275e-01 -4.00834419e-02
-2.16254413e-01 -2.31650963e-01 3.11944216e-01 -5.31836152e-02
-2.23595887e-01 -3.69089216e-01 -1.27412701e+00 -4.29971576e-01
-4.03284542e-02 -1.01240242e+00 5.09727776e-01 -8.81587446e-01
-1.18276620e+00 2.38361508e-01 5.33343991e-03 -1.06258821e+00
3.25424016e-01 -4.83105540e-01 -3.90395671e-02 1.06334305e+00
-1.08262694e+00 -1.29373300e+00 -6.64490521e-01 -2.89152741e-01
1.44150078e-01 -1.64893717e-01 9.29645956e-01 4.35414582e-01
-8.46818626e-01 4.65201825e-01 4.49166298e-02 -1.90821826e-01
8.68781090e-01 -1.41461504e+00 7.72698075e-02 8.50663126e-01
-2.78323349e-02 9.96155500e-01 3.82586360e-01 -8.70853364e-01
-1.31964540e+00 -9.42894816e-01 1.27043903e+00 -8.61113667e-01
7.64397740e-01 -1.22043751e-01 -1.06890571e+00 4.01068389e-01
3.71334046e-01 -4.51124668e-01 1.44506037e+00 7.53651679e-01
-4.07249331e-01 -1.90801784e-01 -1.26017547e+00 -7.46791542e-04
8.27141583e-01 -7.53746450e-01 -4.66410846e-01 5.41869640e-01
3.99512231e-01 -2.88399428e-01 -2.00772715e+00 2.26824984e-01
6.54972196e-01 -4.40272152e-01 3.30067009e-01 -7.51404226e-01
4.58543062e-01 -3.76294464e-01 5.49798422e-02 -1.17029536e+00
-1.09887433e+00 -2.08425000e-01 -1.21771973e-02 1.62180376e+00
4.82118189e-01 -9.79411155e-02 3.83940399e-01 1.60551620e+00
-4.84950542e-01 -5.10758817e-01 -5.29300690e-01 -8.25929582e-01
-2.58249938e-01 -3.91816765e-01 8.85154545e-01 1.51134396e+00
2.42183089e-01 -1.30523801e-01 -2.91065603e-01 3.25788230e-01
3.18379998e-01 -1.62360802e-01 8.61744821e-01 -1.58305264e+00
3.43395561e-01 -3.71351540e-01 -9.64953378e-02 -1.83231816e-01
-8.49096384e-03 -1.03973019e+00 -2.38989424e-02 -1.80984354e+00
3.42834622e-01 -1.01992548e+00 -3.23270023e-01 6.25722349e-01
-2.50155777e-01 7.57281110e-02 2.36763462e-01 8.21785331e-01
-8.26858640e-01 -4.80630010e-01 1.05506611e+00 3.96863818e-01
-1.38794035e-01 -5.49735844e-01 -1.18018353e+00 4.39095616e-01
4.94582325e-01 -7.57456183e-01 -2.25619346e-01 -8.32206786e-01
6.64537907e-01 -4.79140699e-01 -1.80477127e-01 -6.14318252e-01
5.62965453e-01 -5.62798858e-01 4.75379862e-02 -3.97600114e-01
-4.58330303e-01 -9.52106893e-01 7.11945355e-01 1.11163363e-01
-8.91143143e-01 5.70356011e-01 -6.69844523e-02 1.17028341e-01
-6.61150832e-03 -9.72135365e-01 4.16112214e-01 -7.48745278e-02
-7.40973234e-01 -6.87374994e-02 -2.30737820e-01 3.29960026e-02
1.02699649e+00 -2.45479807e-01 -9.25081313e-01 5.11447668e-01
-3.88405919e-01 3.25433671e-01 8.67492318e-01 3.74431819e-01
-1.08903967e-01 -1.28903472e+00 -6.82947457e-01 -2.22918317e-01
6.95974410e-01 -5.28881662e-02 -3.95464636e-02 7.07057595e-01
-6.88450277e-01 5.88550270e-01 -4.26812381e-01 -3.58884245e-01
-1.18492568e+00 3.79802614e-01 -3.85961026e-01 -3.44732910e-01
-8.89335126e-02 7.78788745e-01 -6.75007820e-01 -7.96615332e-02
1.60292283e-01 6.17311075e-02 -2.82003254e-01 3.41057450e-01
4.11371678e-01 1.06703317e+00 4.67015535e-01 -3.63637805e-01
-1.91628829e-01 -3.00397519e-02 -2.86214560e-01 8.82923827e-02
1.36400223e+00 -3.59170675e-01 -7.62835860e-01 6.83939397e-01
7.04863846e-01 2.07215816e-01 -6.07992291e-01 -2.80489117e-01
9.23595011e-01 -8.77253532e-01 3.46335888e-01 -1.08714378e+00
-8.05429757e-01 2.76101112e-01 4.68180686e-01 7.34499156e-01
8.99843574e-01 3.37460339e-02 -6.75490946e-02 4.00778085e-01
3.53483737e-01 -1.88056362e+00 -1.22824358e-02 4.91997832e-03
9.53305781e-01 -1.29974413e+00 5.72093070e-01 -2.29425162e-01
-8.52458060e-01 1.37492859e+00 8.37039351e-01 7.37851024e-01
1.66611314e-01 3.15448284e-01 6.21520728e-02 -7.21210837e-01
-7.39141047e-01 2.32933849e-01 5.26548684e-01 5.78311801e-01
9.01820242e-01 4.54906374e-01 -7.56853163e-01 2.95807928e-01
4.25072340e-03 1.43606171e-01 6.32654428e-01 1.17027068e+00
-4.61894840e-01 -1.42430818e+00 -3.82071942e-01 9.21274900e-01
-5.01829803e-01 -1.16763651e-01 -4.58230704e-01 5.53763270e-01
5.56448460e-01 1.15372050e+00 1.78170055e-01 -3.02597106e-01
4.43488926e-01 2.69777685e-01 1.11457959e-01 -1.01602566e+00
-7.67596722e-01 -3.10167938e-01 8.16613853e-01 -4.66011047e-01
-9.73677099e-01 -8.86796176e-01 -6.89004838e-01 -5.43893337e-01
-6.72429264e-01 5.59902191e-01 1.13601339e+00 8.58681858e-01
8.72060359e-01 4.17507172e-01 6.32177472e-01 2.98426062e-01
-1.65405735e-01 -1.18336654e+00 -1.83696717e-01 4.58278716e-01
-3.62336606e-01 -6.54881418e-01 2.26518348e-01 2.70209283e-01] | [9.525323867797852, 8.56096363067627] |
a860541c-8970-4ee9-b47b-5f7fd7e39be3 | learning-to-discover-and-detect-objects | 2210.10774 | null | https://arxiv.org/abs/2210.10774v2 | https://arxiv.org/pdf/2210.10774v2.pdf | Learning to Discover and Detect Objects | We tackle the problem of novel class discovery and localization (NCDL). In this setting, we assume a source dataset with supervision for only some object classes. Instances of other classes need to be discovered, classified, and localized automatically based on visual similarity without any human supervision. To tackle NCDL, we propose a two-stage object detection network Region-based NCDL (RNCDL) that uses a region proposal network to localize regions of interest (RoIs). We then train our network to learn to classify each RoI, either as one of the known classes, seen in the source dataset, or one of the novel classes, with a long-tail distribution constraint on the class assignments, reflecting the natural frequency of classes in the real world. By training our detection network with this objective in an end-to-end manner, it learns to classify all region proposals for a large variety of classes, including those not part of the labeled object class vocabulary. Our experiments conducted using COCO and LVIS datasets reveal that our method is significantly more effective than multi-stage pipelines that rely on traditional clustering algorithms. Furthermore, we demonstrate the generality of our approach by applying our method to a large-scale Visual Genome dataset, where our network successfully learns to detect various semantic classes without direct supervision. | ['Aljoša Ošep', 'Laura Leal-Taixé', 'Deva Ramanan', 'Ismail Elezi', 'Vladimir Fomenko'] | 2022-10-19 | null | null | null | null | ['novel-class-discovery', 'novel-class-discovery'] | ['computer-vision', 'methodology'] | [ 4.20646280e-01 1.46469235e-01 -3.09841990e-01 -4.88663435e-01
-7.52927780e-01 -9.71753836e-01 4.94360328e-01 3.66722256e-01
-5.62018633e-01 4.23664778e-01 -2.60883868e-01 4.99112578e-03
8.58921632e-02 -7.32833743e-01 -1.04779291e+00 -7.22531259e-01
7.55662024e-02 8.62670302e-01 5.35396755e-01 4.22926575e-01
2.12833181e-01 6.15089417e-01 -1.49967265e+00 6.01692080e-01
3.09720039e-01 9.48670387e-01 3.97195697e-01 4.61981356e-01
-3.36495275e-03 6.34959042e-01 -5.90383649e-01 1.35107577e-01
2.41776675e-01 -3.14098030e-01 -8.90660524e-01 3.30672622e-01
6.61097050e-01 -7.29190782e-02 -3.50994579e-02 1.00181282e+00
2.40460068e-01 1.40940383e-01 7.62963891e-01 -1.09702039e+00
-4.71693724e-01 3.90554547e-01 -8.52947235e-01 3.75516802e-01
-6.96507618e-02 2.28796065e-01 1.17442465e+00 -1.11512733e+00
1.06730068e+00 1.16186202e+00 5.26110888e-01 3.79212469e-01
-1.50595689e+00 -5.67526996e-01 4.55775648e-01 6.47583157e-02
-1.62355554e+00 -3.83156449e-01 5.18567920e-01 -5.29167295e-01
7.59580672e-01 -1.15076102e-01 4.16891307e-01 7.63385296e-01
-2.63162941e-01 8.42634141e-01 7.51734674e-01 -2.87528425e-01
5.64644814e-01 3.16765785e-01 1.13315210e-01 7.23074913e-01
8.36588144e-02 -1.54546350e-01 -4.04923946e-01 -1.00622714e-01
5.32191277e-01 2.24851176e-01 -1.25774339e-01 -9.72674608e-01
-1.32731903e+00 6.48746431e-01 7.34625876e-01 3.40646654e-01
-2.71411389e-01 1.63794547e-01 3.37877572e-01 -1.72077805e-01
6.04879200e-01 4.06215280e-01 -7.54537284e-01 3.87428015e-01
-9.31314230e-01 3.19729820e-02 3.56140316e-01 9.44602787e-01
9.58010733e-01 -6.24168515e-01 -2.12553501e-01 7.49811411e-01
2.76045054e-01 1.93046540e-01 3.21703911e-01 -8.03611338e-01
1.49586245e-01 8.88529718e-01 -1.01002567e-01 -7.15307057e-01
-5.32433867e-01 -7.83413947e-01 -5.54859638e-01 1.33010596e-01
5.38755596e-01 2.67705441e-01 -1.08246863e+00 1.89781475e+00
7.71933854e-01 1.80934310e-01 3.87082770e-02 8.28785360e-01
6.94124162e-01 5.31834245e-01 -6.79337457e-02 9.99788269e-02
1.38894880e+00 -8.15264642e-01 -1.68655545e-01 -2.97902644e-01
6.39746845e-01 -3.28403920e-01 9.93409157e-01 2.90391833e-01
-5.24871230e-01 -5.93210995e-01 -7.12830663e-01 -4.77536097e-02
-6.01340115e-01 5.15625656e-01 5.04612446e-01 1.28426656e-01
-9.20602798e-01 2.47589141e-01 -7.66084731e-01 -6.43699586e-01
1.05016863e+00 8.97381827e-02 -4.34781879e-01 -2.56277084e-01
-4.45917934e-01 4.12034363e-01 8.06012332e-01 -1.89087465e-01
-1.35522783e+00 -8.09617937e-01 -8.76315832e-01 1.51955411e-01
6.82491362e-01 -3.96391332e-01 9.40572441e-01 -1.11016250e+00
-6.62641585e-01 1.17853343e+00 -2.78813660e-01 -2.19821885e-01
1.53005108e-01 2.72807091e-01 -2.52625514e-02 3.55888218e-01
4.70835894e-01 1.10985792e+00 6.66538358e-01 -1.40543127e+00
-9.80136454e-01 -4.46445674e-01 -1.30140027e-02 -9.77227688e-02
-4.83768359e-02 -2.36222632e-02 -5.54018378e-01 -4.41359192e-01
1.89912587e-01 -8.63533020e-01 -1.37707382e-01 5.88609040e-01
-5.13021052e-01 -5.21410644e-01 9.06404316e-01 -3.64951670e-01
5.67665756e-01 -2.45812273e+00 1.24353223e-01 3.13067228e-01
4.35416043e-01 6.44179657e-02 -2.39993513e-01 2.38871388e-02
-1.26234800e-01 -1.29969135e-01 -5.73257267e-01 -1.07817814e-01
-1.70639709e-01 1.01867206e-01 -2.41927028e-01 7.03757226e-01
5.03780544e-01 7.88743615e-01 -1.11187327e+00 -3.56142640e-01
-1.00270145e-01 3.05912226e-01 -5.43809891e-01 2.90805101e-01
-5.42725086e-01 3.77030641e-01 -1.88345924e-01 8.90577555e-01
5.25815904e-01 -6.65120184e-01 3.02941978e-01 -3.00456703e-01
1.79539695e-02 -3.09527609e-02 -9.85129178e-01 1.78917384e+00
-1.32745251e-01 6.53566003e-01 5.11499420e-02 -1.19117379e+00
6.20349884e-01 -3.90929505e-02 3.13592732e-01 -2.65513986e-01
-1.12447515e-01 6.19000345e-02 6.14805408e-02 -3.17208022e-01
-3.12793930e-03 -1.38080837e-02 -2.78831422e-02 5.32689333e-01
4.69258368e-01 2.61084259e-01 3.18358928e-01 3.81700009e-01
1.30632007e+00 2.57419318e-01 3.54625762e-01 -1.26131073e-01
1.33383408e-01 2.96361923e-01 7.97121882e-01 9.39673245e-01
-3.52263331e-01 6.98811948e-01 7.49747574e-01 -3.93998146e-01
-9.91659224e-01 -1.19427609e+00 -2.32439712e-01 1.39704883e+00
6.48717508e-02 -1.08756907e-01 -5.12201548e-01 -1.26849413e+00
1.33458525e-01 4.51332897e-01 -8.80696535e-01 -1.61897480e-01
-2.20805317e-01 -6.57087862e-01 3.96285385e-01 5.20729244e-01
1.79754093e-01 -1.14738321e+00 -6.57481670e-01 -4.83244322e-02
3.59876752e-02 -1.09545803e+00 -4.55829501e-01 4.65677649e-01
-4.86799806e-01 -1.48690617e+00 -6.00006640e-01 -9.98365402e-01
1.16125178e+00 3.72223973e-01 1.12973702e+00 8.95733163e-02
-7.50931561e-01 3.28614771e-01 -2.10080996e-01 -2.80331671e-01
-1.91283211e-01 5.59939444e-02 -1.93559572e-01 1.77527219e-01
4.04384404e-01 -1.92907706e-01 -6.76245213e-01 3.12139064e-01
-9.32679057e-01 -9.21714455e-02 6.79406881e-01 7.02088296e-01
9.08337593e-01 1.84432283e-01 5.82058787e-01 -1.11425316e+00
-2.23546267e-01 -8.00621212e-01 -7.78876424e-01 5.53591728e-01
-2.19343930e-01 -1.97679609e-01 4.63805676e-01 -5.38095891e-01
-7.94361234e-01 5.87609172e-01 2.88895190e-01 -5.45601964e-01
-5.47575593e-01 3.66800934e-01 -3.28074723e-01 1.63622662e-01
6.64966106e-01 3.31241608e-01 -7.21968710e-02 -5.45761764e-01
6.38061941e-01 6.14819407e-01 7.06577599e-01 -3.05308998e-01
7.66578853e-01 7.82109976e-01 -2.75430799e-01 -5.81749022e-01
-1.27728462e+00 -8.72793257e-01 -1.13784289e+00 -8.86830017e-02
9.33024585e-01 -1.18998933e+00 -5.93104839e-01 1.23575784e-01
-9.97841179e-01 -4.27505434e-01 -4.07160610e-01 2.99587607e-01
-2.83999383e-01 6.54173270e-02 -3.55435073e-01 -5.15990317e-01
2.17372775e-02 -1.00955796e+00 1.35025489e+00 1.51339471e-01
-6.76901191e-02 -7.43943512e-01 -9.72079560e-02 3.18028569e-01
7.71086365e-02 1.87250271e-01 1.03566861e+00 -9.63018835e-01
-6.81802750e-01 -2.18198106e-01 -5.45102537e-01 9.68707725e-02
3.02369297e-01 -4.97807823e-02 -9.90382195e-01 -5.30133486e-01
-4.96133268e-01 -5.35945594e-01 1.18397212e+00 1.61411867e-01
1.48757279e+00 -2.67670932e-03 -7.75380433e-01 5.98187566e-01
1.38268459e+00 1.59339961e-02 4.47605141e-02 1.65177844e-02
7.68559635e-01 8.05798054e-01 6.83934033e-01 2.75647998e-01
3.76165777e-01 6.46171570e-01 5.71693420e-01 -4.22003478e-01
-1.19307697e-01 -1.75426006e-01 1.90636680e-01 -1.10555507e-01
6.30955577e-01 -3.24587762e-01 -8.65891516e-01 7.60173321e-01
-1.82787681e+00 -7.86495388e-01 3.54778588e-01 2.10873818e+00
8.60965848e-01 -4.20496128e-02 3.49034876e-01 -2.67330348e-01
9.52594876e-01 -1.55561984e-01 -1.05876100e+00 2.61104017e-01
-1.26327559e-01 4.73623686e-02 4.04624075e-01 2.39473525e-02
-1.55052304e+00 1.09911048e+00 5.88709307e+00 7.50319779e-01
-9.73208964e-01 8.66408646e-02 8.76502812e-01 -2.18234524e-01
-6.18011281e-02 1.29134879e-01 -8.08648109e-01 1.82219848e-01
4.49222833e-01 2.72268325e-01 3.71977985e-01 8.92854393e-01
-3.36630866e-02 -8.66624266e-02 -1.29931355e+00 6.27238452e-01
2.36169055e-01 -1.37275350e+00 -6.80607855e-02 -1.06722640e-03
5.85442305e-01 2.49005601e-01 -8.11595917e-02 2.61031747e-01
4.41195786e-01 -8.38659346e-01 8.24290752e-01 3.37939799e-01
8.60675037e-01 -6.39039695e-01 5.33994913e-01 5.07046402e-01
-1.12802804e+00 -2.38568217e-01 -6.56071961e-01 2.90799290e-01
-2.85324752e-01 6.23741865e-01 -1.20161545e+00 1.40532538e-01
7.79977918e-01 1.01520753e+00 -7.83028841e-01 1.13747430e+00
-2.88982332e-01 6.98115528e-01 -3.80795300e-01 2.58163720e-01
1.85504809e-01 8.05902109e-02 2.92240143e-01 1.10552394e+00
2.30078623e-02 -2.92114969e-02 5.62324524e-01 1.25845587e+00
-3.09083223e-01 -1.14105068e-01 -5.21017313e-01 -6.84266239e-02
5.36503911e-01 1.58448458e+00 -1.33242011e+00 -5.10921657e-01
-2.74933785e-01 9.26697254e-01 7.92346001e-01 3.33746374e-01
-7.10711360e-01 -4.40444559e-01 4.93683577e-01 5.50104538e-03
6.83002174e-01 1.24978021e-01 1.64943293e-01 -1.01009345e+00
4.23132144e-02 -6.36889219e-01 7.61867046e-01 -7.33663797e-01
-1.43141735e+00 2.28157550e-01 -1.03821270e-01 -1.01651597e+00
1.14028275e-01 -7.27833927e-01 -4.52505976e-01 4.66420084e-01
-1.44624460e+00 -1.36699092e+00 -1.87291265e-01 3.28758538e-01
5.28026581e-01 -5.15883081e-02 4.71626550e-01 2.27567688e-01
-6.63652360e-01 4.47414637e-01 1.65028185e-01 3.50697339e-01
7.56251037e-01 -1.32285535e+00 1.65261567e-01 8.49581361e-01
3.98411870e-01 5.39730787e-01 2.86458373e-01 -5.50165057e-01
-9.72874880e-01 -1.75793517e+00 5.57672739e-01 -4.85376239e-01
3.66998136e-01 -8.67824376e-01 -1.01396000e+00 8.85084033e-01
-2.23866105e-01 6.92660987e-01 7.92380154e-01 -3.28191221e-02
-7.16821074e-01 -4.54529151e-02 -1.34174240e+00 4.97530922e-02
1.16967583e+00 -4.75516289e-01 -3.62341762e-01 7.48642921e-01
9.44740713e-01 -1.85507298e-01 -6.24381542e-01 2.71544248e-01
3.32289994e-01 -5.98518729e-01 9.64430213e-01 -7.93093741e-01
3.16406190e-01 -8.31369102e-01 -2.20222741e-01 -1.17195165e+00
-4.19411331e-01 1.88163400e-01 1.20180473e-01 1.29508531e+00
6.59559309e-01 -4.74642754e-01 7.38115251e-01 -2.97408737e-02
-1.20584376e-01 -5.86395025e-01 -1.02245164e+00 -5.93336105e-01
-4.24000882e-02 -1.72814235e-01 4.61583883e-01 1.02451551e+00
-2.95079291e-01 3.49465728e-01 4.33147512e-02 4.93664205e-01
6.97564721e-01 5.28959394e-01 7.57848620e-01 -1.27661908e+00
-3.85513455e-01 -2.50039577e-01 -4.74113792e-01 -9.41568315e-01
3.44169080e-01 -1.26243949e+00 4.88212973e-01 -1.33623195e+00
8.35946023e-01 -6.87346160e-01 -5.28304398e-01 9.77201402e-01
-1.66637614e-01 5.49729526e-01 -9.37980339e-02 3.57810587e-01
-1.25630212e+00 3.08953255e-01 7.28355348e-01 -3.84434551e-01
-5.04526906e-02 -6.61900640e-02 -6.58122003e-01 7.43427515e-01
4.05355543e-01 -6.84632897e-01 -2.58044302e-01 -2.56139040e-01
2.52218433e-02 -3.75117928e-01 5.49729645e-01 -8.69318128e-01
1.32248953e-01 -5.83335161e-02 8.70034456e-01 -8.73003662e-01
3.66883986e-02 -9.09986973e-01 -1.51493743e-01 3.51483524e-01
-7.41276622e-01 -4.20001775e-01 1.42011885e-02 8.45118284e-01
8.16139877e-02 3.20305787e-02 1.02221358e+00 -9.73772779e-02
-8.25302899e-01 3.46439928e-01 -3.98941606e-01 7.07782730e-02
1.24195993e+00 -1.73701271e-02 -4.48770702e-01 1.56509712e-01
-7.53527880e-01 3.48460168e-01 6.74575925e-01 3.80687833e-01
5.45472145e-01 -1.00944054e+00 -5.66925287e-01 2.32492968e-01
8.35174799e-01 4.86576766e-01 3.70482430e-02 5.05915821e-01
-2.34762490e-01 2.06407547e-01 2.13192731e-01 -1.13478231e+00
-1.16979671e+00 8.67970288e-01 3.82307798e-01 -7.53840655e-02
-5.90349197e-01 9.46060717e-01 7.43851185e-01 -9.30703163e-01
2.72237957e-01 -1.79090098e-01 -2.06216857e-01 7.53038600e-02
4.01805192e-01 -4.91129272e-02 -1.43670276e-01 -7.09042609e-01
-7.14762866e-01 4.24679577e-01 -2.44983509e-01 3.74933660e-01
1.41429222e+00 -5.28350770e-02 -3.13491374e-01 5.05631983e-01
1.21410406e+00 -2.59155869e-01 -1.40373325e+00 -6.47571206e-01
2.52201527e-01 -3.62294972e-01 -5.56874312e-02 -1.02412045e+00
-1.25911570e+00 6.13518059e-01 8.67256105e-01 -3.08700979e-01
1.01831234e+00 6.86359465e-01 3.27684969e-01 4.47516561e-01
2.68484920e-01 -8.70435655e-01 1.76662669e-01 2.90569067e-01
4.96453643e-01 -1.42168701e+00 -6.31309748e-02 -4.34897959e-01
-3.37784648e-01 9.65433896e-01 7.83084989e-01 9.11359712e-02
5.44592798e-01 -4.77524921e-02 4.95632663e-02 -4.30813611e-01
-8.98615181e-01 -4.31190878e-01 3.15110922e-01 6.98276401e-01
2.55428731e-01 -5.24076112e-02 3.12919766e-01 2.79261768e-01
3.42667103e-01 -2.31901586e-01 3.05106789e-01 9.81838167e-01
-7.14546502e-01 -6.18271828e-01 -2.03153670e-01 5.60270250e-01
-2.97052622e-01 5.69941252e-02 -4.67258453e-01 6.07894421e-01
4.45421249e-01 7.77707040e-01 5.23403943e-01 -1.99800767e-02
1.30059049e-01 -4.45031263e-02 2.53361076e-01 -1.02710354e+00
-2.85417050e-01 6.26443177e-02 -3.75264704e-01 -5.97873092e-01
-3.41634601e-01 -7.01944113e-01 -1.49605381e+00 5.01374781e-01
-4.01524186e-01 4.55686636e-02 5.15891016e-01 1.02225888e+00
5.59636116e-01 5.51549017e-01 6.64352357e-01 -6.88330770e-01
-1.67730361e-01 -7.68840313e-01 -5.55293202e-01 4.89274919e-01
3.46333116e-01 -6.58547461e-01 -2.44557396e-01 1.76145241e-01] | [9.48861312866211, 1.2621123790740967] |
ec96872f-3958-4aaf-b5c3-4161ddbf4933 | few-shot-learning-based-on-multi-stage | 2109.11806 | null | https://arxiv.org/abs/2109.11806v2 | https://arxiv.org/pdf/2109.11806v2.pdf | A Multi-stage Transfer Learning Framework for Diabetic Retinopathy Grading on Small Data | Diabetic retinopathy (DR) is one of the major blindness-causing diseases currently known. Automatic grading of DR using deep learning methods not only speeds up the diagnosis of the disease but also reduces the rate of misdiagnosis. However,problems such as insufficient samples and imbalanced class distribution in small DR datasets have constrained the improvement of grading performance. In this paper, we apply the idea of multi-stage transfer learning into the grading task of DR. The new transfer learning technique utilizes multiple datasets with different scales to enable the model to learn more feature representation information. Meanwhile, to cope with the imbalanced problem of small DR datasets, we present a class-balanced loss function in our work and adopt a simple and easy-to-implement training method for it. The experimental results on IDRiD dataset show that our method can effectively improve the grading performance on small data, obtaining scores of 0.7961 and 0.8763 in terms of accuracy and quadratic weighted kappa, respectively. Our method also outperforms several state-of-the-art methods. | ['Junxing Zhang', 'Bin Wang', 'Lei Shi'] | 2021-09-24 | null | null | null | null | ['diabetic-retinopathy-grading'] | ['medical'] | [-3.69952142e-01 -1.18413374e-01 -1.84527546e-01 -8.15947473e-01
-8.89721453e-01 -6.70445338e-02 -7.62491524e-02 -3.07721168e-01
-3.40667278e-01 8.82821679e-01 3.24280053e-01 -7.70857483e-02
-2.51963139e-01 -1.00286520e+00 -1.74643993e-01 -8.50719929e-01
4.75218236e-01 4.44618672e-01 1.44427955e-01 4.76745367e-02
2.33158559e-01 3.33766967e-01 -1.65874398e+00 4.40957069e-01
1.70783925e+00 9.36239779e-01 5.27760619e-03 2.67191023e-01
-1.42153785e-01 9.20163393e-01 -5.89787126e-01 -4.32781905e-01
2.29958594e-01 -2.49656096e-01 -5.07850230e-01 -6.32683262e-02
6.07243478e-01 -7.46566176e-01 -3.50775212e-01 9.88678634e-01
1.14268553e+00 -2.13722706e-01 6.68483615e-01 -6.71587527e-01
-1.03526950e+00 9.07428935e-02 -1.01828945e+00 1.33533657e-01
-1.25992820e-01 1.51821181e-01 5.81311166e-01 -5.59534907e-01
2.84404069e-01 1.06139326e+00 7.11836874e-01 6.66940987e-01
-8.34381104e-01 -8.81274998e-01 -2.60533243e-01 6.12302005e-01
-1.13525975e+00 -2.52532750e-01 4.70942557e-01 -7.10612833e-01
3.46160889e-01 4.13639694e-02 5.91937482e-01 4.78846312e-01
3.38216498e-02 5.91465771e-01 1.49902260e+00 -5.59420250e-02
1.28346011e-01 -8.82100537e-02 1.88904837e-01 7.10624695e-01
5.00748098e-01 -1.15883090e-02 2.27534920e-01 6.87624961e-02
8.52810383e-01 1.52542725e-01 -2.03185230e-01 -2.08455905e-01
-8.17863405e-01 9.36594665e-01 7.72559583e-01 -9.11550596e-02
-2.49507546e-01 -2.46968836e-01 3.39833796e-01 3.03802550e-01
5.34427106e-01 1.35507002e-01 -4.95969862e-01 1.26289487e-01
-6.15033090e-01 1.02108382e-01 2.11068198e-01 3.67409974e-01
5.11094689e-01 -2.76665896e-01 -5.19655049e-01 1.20932508e+00
2.12994277e-01 6.19118512e-01 5.71288347e-01 -1.03786159e+00
3.93032193e-01 1.12126935e+00 -1.57953482e-02 -6.17533207e-01
-5.07655680e-01 -6.31618261e-01 -1.12252522e+00 5.43126166e-01
5.27630925e-01 -2.98990786e-01 -1.32171810e+00 1.34369016e+00
3.20252925e-01 -7.04567209e-02 -1.22632030e-02 1.14820981e+00
1.01423466e+00 3.12967837e-01 -1.63640063e-02 -1.46386743e-01
1.14502370e+00 -9.29029047e-01 -6.13020837e-01 1.34338394e-01
6.98283613e-01 -5.15935779e-01 1.18072319e+00 5.47921717e-01
-9.08831477e-01 -4.10987705e-01 -7.00063050e-01 -3.52164388e-01
-8.24500471e-02 7.77602375e-01 8.29684913e-01 4.05784249e-01
-8.48794281e-01 2.09481597e-01 -4.55529630e-01 -3.54813576e-01
1.03587198e+00 3.15305740e-01 -3.13392758e-01 -6.37226820e-01
-7.99104452e-01 9.56789315e-01 -1.35317266e-01 1.49155736e-01
-4.29513693e-01 -7.59083033e-01 -3.10894042e-01 -1.00873947e-01
-3.01954728e-02 -9.71622765e-01 9.83142614e-01 -7.07349241e-01
-1.26709580e+00 1.08341277e+00 -1.12699317e-02 1.96673162e-02
6.81873798e-01 -3.24462831e-01 -4.23827261e-01 2.27502346e-01
1.98657308e-02 5.38734198e-01 3.53558481e-01 -8.46990526e-01
-7.73999393e-01 -8.60032439e-01 -6.22720793e-02 4.72591408e-02
-3.73100340e-01 6.14910536e-02 4.79596443e-02 -2.60013312e-01
-1.41621560e-01 -4.98156875e-01 -2.09236905e-01 4.15432453e-01
-2.36812923e-02 -4.85943109e-01 3.55951369e-01 -9.24938262e-01
1.08355129e+00 -2.03988338e+00 -3.13344551e-03 -5.24797104e-02
6.67168677e-01 6.51739836e-01 -1.33308750e-02 -2.02741548e-01
1.13329761e-01 4.28996235e-02 2.34367922e-02 1.32047221e-01
-3.83099228e-01 1.86266378e-01 1.75280541e-01 3.56496185e-01
2.26444244e-01 7.48685539e-01 -7.21053481e-01 -4.86460954e-01
1.08750507e-01 6.17570579e-01 -5.14734566e-01 4.53128546e-01
1.61540538e-01 4.55989957e-01 -5.79278111e-01 7.77626395e-01
8.19958925e-01 -4.82407212e-01 -4.05809999e-01 -3.97893548e-01
-1.60876587e-01 -4.57970351e-02 -7.99356103e-01 1.23517382e+00
-2.10851684e-01 3.76080334e-01 -4.57388878e-01 -1.05343080e+00
1.09888315e+00 5.34424447e-02 2.62777179e-01 -8.70751202e-01
1.67453945e-01 3.44938368e-01 2.15217754e-01 -1.05252099e+00
-3.32497239e-01 -1.77081879e-02 4.88104552e-01 1.33755103e-01
-2.62557924e-01 4.20815736e-01 1.49444059e-01 -3.36447001e-01
9.51510787e-01 -1.40556157e-01 2.84464091e-01 4.07088324e-02
4.51872081e-01 -2.51136959e-01 8.57441902e-01 2.57663101e-01
-3.83544713e-01 7.56760955e-01 5.16763389e-01 -6.26438439e-01
-9.05676007e-01 -8.33632171e-01 -4.76364195e-01 5.18149137e-01
-9.91567522e-02 3.14587682e-01 -4.91444647e-01 -6.83494866e-01
2.97991097e-01 2.62168109e-01 -7.57676482e-01 -1.95723027e-01
-1.96007118e-01 -1.31595671e+00 3.00989747e-01 7.75924623e-01
8.06813836e-01 -6.06695473e-01 -6.52771071e-02 -2.58328706e-01
-7.44474158e-02 -5.55469573e-01 -1.73848942e-01 -5.53465486e-01
-1.21216023e+00 -1.27767098e+00 -1.21557891e+00 -8.57271671e-01
8.82203400e-01 3.82756650e-01 8.03366840e-01 -8.97755474e-03
-5.00927210e-01 -2.17654571e-01 -1.99802652e-01 -3.93661946e-01
-4.92240908e-03 -1.39042512e-01 -2.23725751e-01 1.30700558e-01
8.28071237e-01 -4.78933364e-01 -1.06950474e+00 2.54913449e-01
-5.37439346e-01 -1.68669090e-01 1.18379962e+00 7.53897369e-01
5.98936200e-01 -2.34950837e-02 9.97347713e-01 -7.72301078e-01
4.59961325e-01 -1.94023460e-01 -5.67140281e-01 3.30496728e-01
-1.15211082e+00 -9.54710618e-02 2.45992318e-01 -3.76423180e-01
-1.01977503e+00 -2.87097037e-01 1.17963888e-01 -1.29413947e-01
-1.16600022e-01 4.56712782e-01 -1.16590045e-01 -3.31203490e-01
6.27021134e-01 -7.14893043e-02 2.07392797e-01 -7.80642748e-01
9.84619334e-02 1.14413595e+00 3.26689363e-01 -1.05469078e-01
5.72404385e-01 2.60709047e-01 -3.63809504e-02 -2.30212659e-01
-1.34670281e+00 -3.97144169e-01 -2.62185186e-01 -1.68843508e-01
9.46218610e-01 -1.29758441e+00 -6.01046979e-01 9.02850330e-01
-8.14289927e-01 -2.26897478e-01 -4.53881323e-02 6.71411812e-01
-1.94062173e-01 1.53517365e-01 -5.90377212e-01 -4.49734837e-01
-6.33652508e-01 -8.99428666e-01 8.07889163e-01 6.81627691e-01
3.73608410e-01 -6.56729102e-01 3.01752448e-01 9.39493239e-01
6.04939044e-01 2.35856697e-01 1.23314035e+00 -2.71765798e-01
-3.90742570e-01 -3.28377306e-01 -9.92191494e-01 7.49220371e-01
3.71691108e-01 9.39700827e-02 -1.02335882e+00 -2.58072942e-01
-4.12519485e-01 -4.99065518e-01 1.05320919e+00 7.01901555e-01
1.42452192e+00 -1.16599493e-01 -1.24436043e-01 6.48860693e-01
1.61354578e+00 2.23572105e-01 1.11730599e+00 4.08250719e-01
6.92457378e-01 5.89935005e-01 5.91187716e-01 2.49297410e-01
8.58065069e-01 3.48504215e-01 2.44122937e-01 -5.91344118e-01
-4.47175562e-01 9.57448930e-02 -6.49709329e-02 7.18542755e-01
-5.81237018e-01 1.63020641e-01 -8.80214393e-01 6.05449736e-01
-1.76220393e+00 -7.23178685e-01 -3.99513602e-01 2.30895948e+00
1.10336423e+00 -2.19653044e-02 8.00773501e-02 -6.34231046e-02
8.11744809e-01 -4.09404576e-01 -6.02703691e-01 -5.51974252e-02
-9.25711095e-02 1.18009023e-01 3.56959969e-01 1.98377624e-01
-8.78882349e-01 5.45868695e-01 5.84144831e+00 3.42469394e-01
-1.26451385e+00 3.03875878e-02 7.80655026e-01 -1.80501446e-01
1.10620558e-02 -4.34210867e-01 -7.72560775e-01 6.86522424e-01
5.38788259e-01 -7.29318336e-02 2.96679050e-01 6.27374172e-01
5.28305829e-01 -6.69915974e-02 -8.33541214e-01 9.19925988e-01
5.61498292e-02 -9.29216385e-01 1.87693819e-01 1.31295398e-01
8.43414426e-01 1.28980935e-01 7.93686658e-02 3.77655983e-01
1.11943543e-01 -1.16908860e+00 -3.53555739e-01 9.85315382e-01
1.05674088e+00 -9.09956276e-01 1.41314411e+00 -1.18716791e-01
-5.31161726e-01 -2.05622733e-01 -7.22969830e-01 -1.25856906e-01
-2.06310675e-01 1.22818232e+00 -4.98572171e-01 3.69167149e-01
8.96944880e-01 8.50310922e-01 -8.58737051e-01 1.72669709e+00
-2.93102413e-01 6.28490746e-01 1.95924833e-01 1.87147573e-01
-2.72270650e-01 -4.82318580e-01 5.36713004e-02 7.63476908e-01
4.16875243e-01 3.16836923e-01 -3.04675214e-02 6.55213118e-01
-6.89443201e-02 2.87323058e-01 -3.05660635e-01 1.83031514e-01
1.47408679e-01 1.21871734e+00 2.09200352e-01 -1.65234089e-01
-5.39966404e-01 4.29802448e-01 3.18819970e-01 3.86769116e-01
-4.96005863e-01 -7.12723792e-01 4.45972919e-01 1.39232993e-01
2.72751123e-01 2.78449982e-01 -5.87185323e-01 -1.28343022e+00
1.75009176e-01 -7.38645852e-01 4.04584855e-01 -1.07334197e+00
-1.71574998e+00 3.26225579e-01 -6.42243981e-01 -1.33533359e+00
2.87088245e-01 -4.99871463e-01 -6.30079448e-01 9.83337939e-01
-2.11358333e+00 -1.21851492e+00 -8.65015686e-01 4.40884829e-01
-2.64433287e-02 -4.36527520e-01 6.87560260e-01 7.91004419e-01
-8.80428731e-01 6.11393511e-01 2.49343514e-01 2.60685503e-01
1.44369698e+00 -1.38110936e+00 -4.29259092e-01 5.50124049e-01
-9.37332511e-01 2.79332638e-01 1.96132347e-01 -4.53249186e-01
-8.11392963e-01 -1.25368655e+00 8.92645776e-01 -1.76594034e-01
4.66399163e-01 3.13645035e-01 -1.03620589e+00 3.63548994e-01
-2.72954434e-01 1.88486487e-01 9.12158847e-01 2.07010612e-01
-4.50287074e-01 -8.65120590e-01 -1.52057564e+00 3.14618856e-01
8.39839399e-01 -1.65263236e-01 -6.08832896e-01 3.33666623e-01
5.27158499e-01 -2.96552896e-01 -1.21867144e+00 6.94767714e-01
7.57437408e-01 -9.44158852e-01 7.13405371e-01 -7.92414069e-01
8.54137659e-01 -3.84040534e-01 1.32643720e-02 -1.21801603e+00
-4.54599798e-01 4.07121703e-02 -1.57473132e-01 1.31177974e+00
2.46702254e-01 -9.63683069e-01 5.21087408e-01 4.90444511e-01
-1.16288975e-01 -9.09336329e-01 -6.48540735e-01 -3.83482724e-01
3.19670618e-01 4.50386196e-01 4.15330410e-01 9.39361274e-01
-4.59504634e-01 2.76273251e-01 -9.24657732e-02 2.54640788e-01
8.72801065e-01 3.32558274e-01 5.75865090e-01 -1.70348108e+00
-5.07414807e-03 -2.75730282e-01 -5.94041228e-01 -5.88292241e-01
-3.82849693e-01 -7.80268729e-01 -3.14161927e-01 -2.09055209e+00
6.61577702e-01 -5.29743373e-01 -5.87410092e-01 7.39301264e-01
-4.10246432e-01 3.50227296e-01 -4.97379303e-01 1.97083399e-01
-4.32670146e-01 5.30831158e-01 1.75655842e+00 -1.99890077e-01
-1.75840110e-01 1.89874142e-01 -1.20703328e+00 5.03945291e-01
1.00102091e+00 -2.18361348e-01 -3.67347717e-01 -7.25106239e-01
2.01550633e-01 -2.85206884e-01 3.14239651e-01 -9.57259417e-01
-7.58630931e-02 -2.39929676e-01 7.50413418e-01 -4.16256100e-01
-1.70038253e-01 -4.35750395e-01 -3.60653698e-01 3.99717480e-01
-2.40644112e-01 -4.61949736e-01 -6.69388920e-02 3.42639595e-01
-1.51250511e-01 2.46636510e-01 1.10238457e+00 9.57377031e-02
-4.31977630e-01 4.76148129e-01 6.29886240e-02 5.25111035e-02
8.84864926e-01 -5.75074255e-02 -9.51788783e-01 -9.94552821e-02
-4.46498930e-01 6.35540783e-01 3.86816442e-01 2.64612675e-01
5.71737647e-01 -1.31804252e+00 -1.19027889e+00 -3.55758257e-02
4.79296923e-01 3.13196629e-01 3.62749845e-01 1.17715096e+00
-4.88671362e-01 1.60556614e-01 -4.94847983e-01 -5.33587277e-01
-1.33695364e+00 2.57237017e-01 5.84844470e-01 -1.18878141e-01
-6.76614165e-01 6.32353187e-01 1.28442824e-01 -3.52883935e-01
2.62381643e-01 -3.81920516e-01 -6.89084232e-01 1.20971642e-01
9.35931981e-01 7.63245702e-01 1.68830901e-01 3.60432491e-02
-8.93471316e-02 1.01787651e+00 -4.53155816e-01 6.47534251e-01
1.64109325e+00 -1.07751995e-01 -4.43407983e-01 2.31168360e-01
9.32343423e-01 -8.36790875e-02 -1.08098900e+00 -2.01215774e-01
-2.48023808e-01 -5.96937537e-01 4.66598153e-01 -1.51435232e+00
-1.46665502e+00 1.01643574e+00 1.31988204e+00 -1.50628746e-01
1.27661848e+00 -8.87894854e-02 9.72434878e-01 3.11332107e-01
2.37063795e-01 -8.53865564e-01 -1.94852993e-01 -3.05120200e-02
6.96015716e-01 -1.58819795e+00 1.12452924e-01 -2.68972248e-01
-5.68261504e-01 1.09754407e+00 8.45702767e-01 -1.46173924e-01
2.31744751e-01 -1.52086407e-01 6.36495709e-01 -4.66193520e-02
-6.26027942e-01 -2.07524076e-01 4.60774899e-01 8.16992819e-01
3.74153316e-01 1.50199622e-01 -8.20397973e-01 5.94236970e-01
1.39449432e-01 7.55505383e-01 7.19938576e-01 4.16760534e-01
-6.94955468e-01 -1.02327418e+00 2.48343553e-02 1.03863978e+00
-4.53033239e-01 1.43810540e-01 -3.52832466e-01 4.29947495e-01
2.89793164e-01 7.68451214e-01 -3.70422774e-03 -3.34433824e-01
3.61257643e-01 -2.27940291e-01 4.69669521e-01 -4.80526358e-01
-4.85643223e-02 -9.93201882e-02 8.00530761e-02 -4.64379311e-01
-6.24373078e-01 -3.65819842e-01 -1.15867662e+00 -1.80579871e-01
-2.83042103e-01 -2.85784245e-01 3.51170212e-01 9.46400464e-01
6.00792229e-01 5.94748914e-01 8.38511586e-01 -9.16637778e-02
-8.74686003e-01 -1.06567335e+00 -5.71956038e-01 1.14374585e-01
4.66892511e-01 -7.21282661e-01 -2.82816142e-01 8.41106940e-03] | [15.808928489685059, -3.974264621734619] |
981a9e4d-f43e-40fc-9081-a2a97872af5b | tganet-text-guided-attention-for-improved | 2205.04280 | null | https://arxiv.org/abs/2205.04280v1 | https://arxiv.org/pdf/2205.04280v1.pdf | TGANet: Text-guided attention for improved polyp segmentation | Colonoscopy is a gold standard procedure but is highly operator-dependent. Automated polyp segmentation, a precancerous precursor, can minimize missed rates and timely treatment of colon cancer at an early stage. Even though there are deep learning methods developed for this task, variability in polyp size can impact model training, thereby limiting it to the size attribute of the majority of samples in the training dataset that may provide sub-optimal results to differently sized polyps. In this work, we exploit size-related and polyp number-related features in the form of text attention during training. We introduce an auxiliary classification task to weight the text-based embedding that allows network to learn additional feature representations that can distinctly adapt to differently sized polyps and can adapt to cases with multiple polyps. Our experimental results demonstrate that these added text embeddings improve the overall performance of the model compared to state-of-the-art segmentation methods. We explore four different datasets and provide insights for size-specific improvements. Our proposed text-guided attention network (TGANet) can generalize well to variable-sized polyps in different datasets. | ['Sharib Ali', 'Ulas Bagci', 'Debesh Jha', 'Nikhil Kumar Tomar'] | 2022-05-09 | null | null | null | null | ['polyp-segmentation'] | ['computer-vision'] | [ 2.68845409e-01 3.87786895e-01 -4.99359131e-01 -3.95211071e-01
-6.74717844e-01 -4.89093095e-01 5.21650165e-02 7.41538644e-01
-7.06581950e-01 2.28939250e-01 4.74687994e-01 -5.43360591e-01
-5.59699237e-02 -9.09744978e-01 -7.54473269e-01 -5.39247870e-01
-3.78388613e-01 3.51340145e-01 3.63817483e-01 -6.06146678e-02
6.66171908e-02 2.29959026e-01 -8.84043336e-01 4.83588248e-01
1.22540033e+00 7.45068610e-01 4.03951824e-01 8.56343150e-01
1.01689234e-01 2.56547213e-01 -3.04123789e-01 -2.64300793e-01
4.62731957e-01 -2.33324811e-01 -6.39867842e-01 -1.36013895e-01
3.96744549e-01 -1.28102049e-01 -2.81739235e-01 1.00413978e+00
4.48473811e-01 1.38239171e-02 7.65752137e-01 -2.87169874e-01
-9.52015281e-01 7.06996560e-01 -3.99765640e-01 4.69010949e-01
-5.75857982e-02 1.46072865e-01 8.68752778e-01 -3.42592716e-01
4.28337276e-01 8.97756279e-01 1.03488970e+00 6.65798187e-01
-9.49443877e-01 -1.78361475e-01 4.01253909e-01 -3.51059675e-01
-8.41372788e-01 1.37746885e-01 2.42828265e-01 -4.47967112e-01
7.49003232e-01 3.32166314e-01 1.06489074e+00 8.74341547e-01
6.17786765e-01 8.14780056e-01 6.10574067e-01 -3.10491920e-01
-1.13944091e-01 7.56707788e-02 3.31968106e-02 8.08535516e-01
8.71815920e-01 1.35702282e-01 9.46349502e-02 -3.43364365e-02
9.49075699e-01 3.83449227e-01 -5.13944328e-01 -4.93087828e-01
-1.42795098e+00 1.12396801e+00 1.18653333e+00 3.50737929e-01
-9.63901207e-02 1.48235545e-01 5.68444431e-01 2.22601518e-02
2.94428468e-01 1.02114809e+00 -5.65700173e-01 2.10137695e-01
-8.36799443e-01 -7.44599896e-03 9.15427446e-01 6.22964561e-01
3.28111917e-01 -2.86905557e-01 -5.54757297e-01 7.33252585e-01
3.15817028e-01 1.58473060e-01 1.17966056e+00 -2.86606431e-01
3.62007409e-01 1.12556636e+00 -1.03895389e-01 -7.84288406e-01
-9.07309115e-01 -8.85253787e-01 -6.58988297e-01 -1.11196861e-01
4.92033571e-01 -2.85332620e-01 -1.48964310e+00 1.43649185e+00
1.81042969e-01 -9.80441645e-02 -1.32908389e-01 8.96914244e-01
8.84010673e-01 1.19845174e-01 2.43655890e-01 4.52595115e-01
1.58852220e+00 -1.42413735e+00 -2.39350528e-01 -6.95115030e-01
9.68936324e-01 -6.23556972e-01 1.03136277e+00 -4.42267116e-03
-8.15760553e-01 -2.72852957e-01 -1.10627210e+00 -4.23926264e-01
-4.01516765e-01 3.23808700e-01 8.44354868e-01 9.42459047e-01
-8.80195320e-01 5.29448271e-01 -1.31357825e+00 -5.35914958e-01
5.44436455e-01 5.24451911e-01 -1.98643152e-02 -1.31300911e-01
-9.92220283e-01 8.26663733e-01 4.08391476e-01 -1.11130163e-01
-6.95507109e-01 -9.54902291e-01 -1.16022038e+00 3.88252527e-01
2.33660400e-01 -1.07309806e+00 1.17675388e+00 -9.43205655e-01
-1.31500924e+00 9.13349509e-01 5.58363125e-02 -8.78463805e-01
5.11750937e-01 -9.84440818e-02 -3.57102901e-02 1.86522812e-01
-5.72253391e-03 8.41087699e-01 6.93713486e-01 -7.07868993e-01
-6.58481598e-01 -1.48035005e-01 1.58317938e-01 2.62320340e-01
-4.29551095e-01 -5.75544000e-01 -5.24555981e-01 -8.24369371e-01
-2.19883639e-02 -1.12512183e+00 -9.43699300e-01 4.35226500e-01
-2.77202278e-01 2.33658046e-01 1.77868634e-01 -6.08922303e-01
1.19872844e+00 -2.18355799e+00 5.25785573e-02 -1.94902420e-02
2.73817331e-01 4.35878873e-01 -5.88415973e-02 7.28216469e-02
2.50680387e-01 4.50400770e-01 -4.66824532e-01 -1.39987037e-01
-2.50508994e-01 2.09757522e-01 3.43064338e-01 6.11959696e-01
4.61422831e-01 1.14388549e+00 -1.19298208e+00 -5.36808610e-01
3.76499921e-01 3.84371698e-01 -9.97488260e-01 -1.37137314e-02
-2.86503851e-01 4.13608402e-01 -5.66291511e-01 6.57625377e-01
4.59888160e-01 -6.68779671e-01 7.37009197e-02 -1.58817396e-01
3.24601233e-02 1.65635750e-01 -4.54575628e-01 1.87036228e+00
-7.46379495e-01 6.00563169e-01 -4.38016206e-02 -9.61702406e-01
4.89378870e-01 1.33982837e-01 3.18823725e-01 -2.85836399e-01
3.42014492e-01 5.01641154e-01 7.44270027e-01 -6.69285536e-01
5.25950015e-01 1.09233037e-01 -2.13328991e-02 1.40083045e-01
-1.25594869e-01 -4.76660654e-02 3.20448220e-01 -1.61837667e-01
9.37715173e-01 -1.95187077e-01 7.13097930e-01 -6.55689657e-01
2.83517987e-01 1.83340669e-01 4.63184327e-01 9.67588067e-01
-4.05924648e-01 7.42041647e-01 3.73717636e-01 -6.33493721e-01
-8.58096898e-01 -7.04898477e-01 -5.73223174e-01 1.01418149e+00
2.92055607e-01 6.69655874e-02 -3.83179545e-01 -9.66075301e-01
1.86504826e-01 3.23376000e-01 -1.13346696e+00 -2.16432795e-01
-6.35405779e-01 -1.18255532e+00 2.40458429e-01 7.63981879e-01
2.47217447e-01 -7.06236362e-01 -9.85454082e-01 3.21193010e-01
1.62152603e-01 -7.53936768e-01 -7.51198113e-01 3.74698400e-01
-1.12788427e+00 -1.39706111e+00 -1.24371195e+00 -1.18558764e+00
1.26968288e+00 1.68833375e-01 1.07153213e+00 3.42243373e-01
-5.91625214e-01 3.15818578e-01 -5.51197588e-01 -6.68063045e-01
-5.08855939e-01 6.51750088e-01 -5.07093310e-01 -2.27201223e-01
3.54895025e-01 7.40601495e-03 -1.32374322e+00 8.09220597e-02
-9.27160025e-01 3.16069857e-03 9.27732348e-01 1.21663070e+00
5.92468739e-01 -4.72249806e-01 1.89605191e-01 -1.17812085e+00
6.15518272e-01 -5.32352984e-01 -3.64718974e-01 1.08413436e-01
-4.31109816e-01 1.85049102e-01 6.47673309e-01 -7.12462544e-01
-6.63720965e-01 4.10532653e-02 -2.20226362e-01 1.03661912e-02
2.16667667e-01 7.88321495e-01 5.81297338e-01 -3.96055937e-01
9.77511823e-01 -1.74227403e-04 7.32977502e-03 -2.40390763e-01
4.32800390e-02 4.23031867e-01 1.95586458e-01 -1.76839069e-01
3.26017797e-01 5.28454542e-01 -4.75977100e-02 -4.79883403e-01
-1.04079258e+00 -7.50468671e-01 -3.09138209e-01 5.12558997e-01
9.12588954e-01 -9.33621764e-01 -1.74506932e-01 4.89051379e-02
-6.05029404e-01 -5.91294169e-01 -5.01721919e-01 6.42679214e-01
-2.20449015e-01 3.20810050e-01 -7.98397660e-01 -2.45823726e-01
-5.31728268e-01 -1.42468631e+00 9.90231872e-01 4.95715410e-01
-2.31524244e-01 -1.47993374e+00 1.21325970e-01 -1.44848228e-01
6.12843752e-01 5.71568966e-01 8.82579625e-01 -1.04139197e+00
-4.50901717e-01 -4.25305694e-01 -3.81123990e-01 8.64045098e-02
3.14005136e-01 -1.43642023e-01 -6.08842611e-01 -5.20162880e-01
-4.54154387e-02 -1.76100917e-02 1.51493728e+00 9.38890576e-01
1.29960752e+00 -1.46290392e-01 -8.63167644e-01 1.14887571e+00
1.45167959e+00 -5.09218313e-02 2.75342911e-01 5.18363714e-01
5.30316532e-01 2.16827780e-01 3.08459878e-01 8.82508382e-02
3.19809228e-01 2.06954688e-01 6.13100111e-01 -4.20745254e-01
-2.75969952e-01 -1.47171244e-01 -1.89431325e-01 6.02923810e-01
1.76108822e-01 -3.08054805e-01 -9.39787149e-01 1.07847524e+00
-1.56900823e+00 -4.20796484e-01 -2.20325757e-02 2.16513872e+00
8.00058126e-01 1.43430335e-02 -2.75971323e-01 -1.24280162e-01
4.40754056e-01 8.88697132e-02 -7.44070172e-01 -4.47356969e-01
2.40343645e-01 1.38274193e-01 9.19444680e-01 3.26550841e-01
-1.34931207e+00 4.42976296e-01 6.38698530e+00 3.19066346e-01
-1.45822215e+00 2.85659172e-02 7.28302717e-01 -2.23613158e-02
-2.55346388e-01 -3.56239855e-01 -6.21112883e-01 3.60051721e-01
6.12108767e-01 -1.11530945e-01 -1.18322019e-02 9.25492585e-01
-1.51283234e-01 1.01953400e-02 -1.37173963e+00 4.98667657e-01
6.51215762e-02 -1.33034503e+00 7.81050464e-03 -4.89857420e-02
1.09154236e+00 4.22009051e-01 3.78617495e-01 3.73686492e-01
1.04096383e-01 -1.16475523e+00 1.91443339e-01 1.65486068e-01
9.25555825e-01 -2.56461680e-01 9.80605185e-01 -1.92492045e-02
-1.13450766e+00 -3.77837300e-01 -4.98887688e-01 2.68224448e-01
1.82672814e-02 4.72832382e-01 -1.18408179e+00 3.88117939e-01
4.51753855e-01 8.38954926e-01 -8.07463348e-01 1.66458738e+00
-5.18629812e-02 4.92620617e-01 -3.78375769e-01 -2.26918444e-01
5.01690507e-01 1.96698979e-01 4.83013600e-01 1.43218052e+00
6.33710742e-01 -2.50005245e-01 2.23417953e-01 6.82317555e-01
-1.13172881e-01 2.72579670e-01 -9.75420475e-02 -1.21762760e-01
1.71605900e-01 1.07147455e+00 -9.43318844e-01 -2.36788347e-01
-5.30064046e-01 8.50940883e-01 2.85492092e-01 1.95938796e-01
-6.28204405e-01 -3.84131193e-01 3.42308789e-01 1.38735220e-01
5.62321424e-01 1.66714191e-01 -4.39539552e-01 -1.17516458e+00
-1.36067182e-01 -6.65639400e-01 6.48408651e-01 6.68231472e-02
-1.21662796e+00 5.47085345e-01 -4.43809301e-01 -1.36603713e+00
7.75767043e-02 -9.77876484e-01 -7.74320066e-01 6.19611919e-01
-1.97005570e+00 -1.16092217e+00 -4.24064875e-01 2.85540372e-02
6.55790687e-01 1.29622772e-01 9.34357226e-01 1.55819446e-01
-3.63235593e-01 9.56198037e-01 2.31668726e-01 3.65209103e-01
8.17185760e-01 -1.59723997e+00 4.10905182e-01 6.99903786e-01
-1.07291117e-01 8.22297394e-01 4.07310724e-01 -5.58968782e-01
-1.04201663e+00 -1.46470141e+00 5.31393349e-01 -2.78900325e-01
4.79585767e-01 2.26687007e-02 -8.54971766e-01 8.18418860e-01
-1.13470770e-01 3.12631458e-01 9.13412333e-01 7.68066645e-02
-5.21801151e-02 1.64870739e-01 -1.16034985e+00 8.31741810e-01
8.16483974e-01 6.05734736e-02 -4.25713569e-01 5.07940292e-01
9.24878597e-01 -1.07991385e+00 -1.05115843e+00 4.88660604e-01
4.94892210e-01 -8.80191386e-01 1.00103784e+00 -3.04631799e-01
5.72955370e-01 9.75500047e-02 3.92335534e-01 -1.47165680e+00
-4.71746713e-01 -3.31232369e-01 9.88258868e-02 1.65706143e-01
8.49563241e-01 -9.30144131e-01 8.63167703e-01 5.63308001e-01
-5.64392209e-01 -1.20149088e+00 -7.56462276e-01 -4.01713699e-01
4.82806653e-01 1.57966912e-01 5.89585245e-01 7.63362169e-01
1.25466868e-01 -4.74780232e-01 1.57005727e-01 3.43250245e-01
1.60180390e-01 2.02571377e-01 2.76117444e-01 -1.15357161e+00
-4.64708567e-01 -6.96316481e-01 -4.22567129e-01 -1.23962712e+00
-4.87127006e-01 -1.20342851e+00 -2.67499313e-02 -1.84333622e+00
1.61365762e-01 -4.76700395e-01 -4.34595615e-01 3.27787906e-01
-7.07860410e-01 7.66019374e-02 7.18739182e-02 -5.45318350e-02
-2.26430729e-01 1.24530680e-01 1.78703046e+00 -2.78877318e-01
-5.23124635e-01 3.44301730e-01 -9.45844054e-01 6.68318093e-01
8.19575787e-01 -5.92652798e-01 -2.57721156e-01 -7.26412535e-01
2.40473002e-01 -2.26618290e-01 1.51520789e-01 -1.04999793e+00
1.24340795e-01 6.24181032e-02 5.87281883e-01 -9.64586511e-02
-1.26716942e-01 -7.56939292e-01 -4.29692507e-01 1.06479371e+00
-5.39336979e-01 -3.66329759e-01 5.37226140e-01 8.40144694e-01
-1.17727123e-01 -5.57243645e-01 7.85073280e-01 -5.03737390e-01
-1.74088687e-01 4.74106878e-01 -4.11129355e-01 2.56998241e-01
9.07589972e-01 -3.37899476e-01 -2.10332587e-01 4.89603579e-02
-6.03683054e-01 4.64330643e-01 5.41938066e-01 2.92387366e-01
2.90446043e-01 -8.86130393e-01 -8.13051343e-01 1.44913018e-01
3.48487735e-01 6.14131927e-01 2.99411416e-01 1.12850380e+00
-9.52540934e-01 5.61402798e-01 8.78003910e-02 -6.74759448e-01
-8.47448945e-01 5.32755256e-01 6.49272859e-01 -7.72075176e-01
-7.86811590e-01 1.13572240e+00 3.64217252e-01 -3.98282021e-01
1.52449474e-01 -1.26486278e+00 -2.82138884e-01 -2.80024290e-01
4.60592777e-01 -8.23721662e-02 -6.66692257e-02 6.29900908e-03
-9.75356810e-03 5.90183735e-01 -5.36557734e-01 4.85890925e-01
1.13794291e+00 2.73245797e-02 2.78136224e-01 1.26894265e-01
9.91727889e-01 9.72057730e-02 -1.23917282e+00 -6.62335232e-02
-1.86254412e-01 -4.02967840e-01 1.45627514e-01 -8.99704218e-01
-1.16357553e+00 7.76903331e-01 8.05317283e-01 5.53564914e-02
9.55430329e-01 -2.76936173e-01 9.83315170e-01 2.06999227e-01
-7.70588368e-02 -6.80136979e-01 -1.06173180e-01 2.32791111e-01
7.36680984e-01 -1.64243889e+00 1.23580366e-01 -3.79949063e-01
-4.69446540e-01 1.25411260e+00 4.78373021e-01 -4.34160680e-01
4.92494315e-01 1.75246000e-01 2.74324954e-01 -4.56870086e-02
-3.24956238e-01 -2.81503461e-02 6.22466505e-01 6.54300392e-01
6.92023575e-01 1.67907298e-01 -4.73723710e-01 6.91651285e-01
-6.68001622e-02 -1.04847915e-01 5.74253678e-01 8.74870896e-01
-5.68839312e-01 -1.03096020e+00 -1.36229426e-01 1.14434147e+00
-8.13652754e-01 -2.88555324e-01 1.04808055e-01 9.71993566e-01
1.32437527e-01 5.33294380e-01 1.24538556e-01 2.52595544e-01
1.51701540e-01 -2.67893136e-01 4.02550697e-01 -1.01381648e+00
-1.04205823e+00 -3.65535766e-02 -5.58956824e-02 -2.57673979e-01
-2.18141571e-01 -4.95933712e-01 -1.23538518e+00 3.75123620e-01
-4.52464342e-01 -1.22951359e-01 4.26543981e-01 5.19061327e-01
3.21961582e-01 9.05958891e-01 2.99865697e-02 -8.79210711e-01
-6.96356654e-01 -9.94841099e-01 -1.71703562e-01 5.22637427e-01
7.29507506e-01 -3.45302820e-01 -5.15003741e-01 -4.26984169e-02] | [14.56131649017334, -2.89798641204834] |
0cbddec5-3159-47b3-b0fd-09b0c1f59819 | the-effect-of-noise-level-on-causal | 2108.11320 | null | https://arxiv.org/abs/2108.11320v1 | https://arxiv.org/pdf/2108.11320v1.pdf | The Effect of Noise Level on Causal Identification with Additive Noise Models | In recent years a lot of research has been conducted within the area of causal inference and causal learning. Many methods have been developed to identify the cause-effect pairs in models and have been successfully applied to observational real-world data in order to determine the direction of causal relationships. Many of these methods require simplifying assumptions, such as absence of confounding, cycles, and selection bias. Yet in bivariate situations causal discovery problems remain challenging. One class of such methods, that also allows tackling the bivariate case, is based on Additive Noise Models (ANMs). Unfortunately, one aspect of these methods has not received much attention until now: what is the impact of different noise levels on the ability of these methods to identify the direction of the causal relationship. This work aims to bridge this gap with the help of an empirical study. For this work, we considered bivariate cases, which is the most elementary form of a causal discovery problem where one needs to decide whether X causes Y or Y causes X, given joint distributions of two variables X, Y. Furthermore, two specific methods have been selected, \textit{Regression with Subsequent Independence Test} and \textit{Identification using Conditional Variances}, which have been tested with an exhaustive range of ANMs where the additive noises' levels gradually change from 1% to 10000% of the causes' noise level (the latter remains fixed). Additionally, the experiments in this work consider several different types of distributions as well as linear and non-linear ANMs. The results of the experiments show that these methods can fail to capture the true causal direction for some levels of noise. | ['Benjamin Kap'] | 2021-08-24 | null | null | null | null | ['causal-identification'] | ['reasoning'] | [ 3.78624737e-01 -1.61167130e-01 -4.02927667e-01 -2.10829556e-01
-1.92919254e-01 -2.96495587e-01 8.44758511e-01 1.86965808e-01
-3.22988868e-01 1.03828800e+00 2.30068833e-01 -6.90659821e-01
-8.21004629e-01 -9.35110807e-01 -7.54685462e-01 -6.69745743e-01
-3.96243066e-01 1.98148072e-01 2.43766829e-01 -4.56261598e-02
2.62870133e-01 1.25986651e-01 -1.50918198e+00 -2.12294117e-01
8.44235301e-01 3.55582595e-01 2.29852609e-02 4.76250380e-01
-4.69357744e-02 6.63257003e-01 -5.67745566e-01 -2.06676960e-01
1.32055342e-01 -6.65306509e-01 -5.66025078e-01 -2.43172839e-01
-1.43826216e-01 -1.10214844e-01 -1.90187380e-01 8.45120013e-01
4.02272493e-01 -6.27518967e-02 8.72623742e-01 -1.41665375e+00
-3.87560457e-01 8.45593333e-01 -5.90629339e-01 2.70994902e-01
3.84483516e-01 2.64866110e-02 7.57246256e-01 -3.08361053e-01
3.13080788e-01 1.67182255e+00 3.97148162e-01 2.08790824e-02
-1.52366495e+00 -9.10527349e-01 1.02330968e-01 2.45761529e-01
-1.26398873e+00 -9.49573368e-02 6.49624705e-01 -5.82180917e-01
3.52767557e-01 2.73306936e-01 2.45157138e-01 1.24531424e+00
3.82888675e-01 1.80362478e-01 1.48923278e+00 -7.54638851e-01
3.80715370e-01 1.98193148e-01 4.01716083e-01 4.18181568e-02
6.49303913e-01 6.20974362e-01 -3.55889648e-01 -3.74804109e-01
8.48186016e-01 -1.02796100e-01 -1.11694783e-01 -7.45637119e-02
-9.40834045e-01 1.03359497e+00 8.95231888e-02 6.84099734e-01
-3.65128517e-01 1.61589056e-01 1.28208995e-01 3.84576052e-01
3.49935502e-01 2.49927253e-01 -3.21471423e-01 1.46588059e-02
-4.14399177e-01 3.57611656e-01 7.30094433e-01 4.17592764e-01
4.85064805e-01 -9.30551589e-02 -9.77209657e-02 4.94588614e-01
3.32832843e-01 5.85932970e-01 1.73446909e-01 -4.47997063e-01
4.20924306e-01 5.83550811e-01 3.80346268e-01 -1.18904710e+00
-3.73610407e-01 -2.45653287e-01 -9.12827134e-01 4.45804931e-02
7.46107519e-01 -4.92424190e-01 -8.27633917e-01 1.93123972e+00
2.77118713e-01 5.99684298e-01 -1.18807301e-01 6.98655605e-01
5.41745305e-01 4.71649319e-01 3.98389846e-01 -6.52922630e-01
1.02878022e+00 6.42738715e-02 -9.28682208e-01 6.34560212e-02
2.57735610e-01 -6.80572152e-01 9.85989034e-01 3.07557166e-01
-6.98527932e-01 -4.69042361e-01 -5.72894990e-01 6.08812332e-01
-2.81781137e-01 -2.18835726e-01 7.28285789e-01 8.09883296e-01
-5.60747147e-01 4.67770725e-01 -6.35848522e-01 -3.20996612e-01
-1.56295568e-01 2.54637927e-01 -8.41493681e-02 -1.36758983e-01
-1.71603596e+00 7.56132722e-01 1.71691030e-01 1.12502910e-01
-8.38804245e-01 -7.73026466e-01 -4.00779098e-01 1.51877865e-01
6.57118320e-01 -5.94703555e-01 9.10151005e-01 -9.88006294e-01
-9.74876344e-01 2.90825248e-01 -2.59847492e-01 -2.38721102e-01
5.23178637e-01 -2.01087266e-01 -4.94195938e-01 -4.22558725e-01
2.89873153e-01 -1.27520561e-01 4.61423635e-01 -1.34679186e+00
-5.24747670e-01 -5.97638249e-01 -3.17958393e-03 -2.50571996e-01
5.34658786e-03 3.64589363e-01 -2.40546931e-02 -5.93696952e-01
-1.43083818e-02 -8.36930752e-01 -3.00928712e-01 -6.69995189e-01
-5.49070776e-01 -3.18512619e-01 4.04088885e-01 -2.61596024e-01
1.46989489e+00 -2.04430056e+00 2.27575228e-02 4.32850242e-01
-2.01569676e-01 -6.93576485e-02 1.37320533e-01 7.92691529e-01
-4.28109378e-01 4.25881922e-01 -2.97724694e-01 9.78643969e-02
-1.65883079e-01 3.76654983e-01 -2.78762221e-01 3.64979714e-01
1.29187793e-01 3.50895613e-01 -7.30147958e-01 -2.01848611e-01
2.36904532e-01 3.74095917e-01 -2.64741987e-01 1.78653046e-01
2.14427803e-03 5.18598914e-01 -4.72825080e-01 1.49118528e-01
4.65889156e-01 -3.30592096e-02 4.81529683e-01 2.86296576e-01
-5.50590575e-01 1.18424706e-01 -1.64859879e+00 6.49342358e-01
-2.39690647e-01 2.77207464e-01 -1.14016347e-01 -1.39456952e+00
8.93072963e-01 5.34860790e-01 4.74909574e-01 -4.83626336e-01
2.43131742e-01 1.71394452e-01 5.63526332e-01 -7.28239357e-01
-1.37011230e-01 -5.26194692e-01 -3.02721169e-02 1.79509401e-01
-4.40016240e-01 2.01152459e-01 4.06760395e-01 -1.70535535e-01
1.18057823e+00 -2.48093635e-01 5.24431407e-01 -2.05318317e-01
3.19442809e-01 -2.48334147e-02 8.50148559e-01 1.06961608e+00
2.38346562e-01 2.87197560e-01 1.00838041e+00 -1.53610274e-01
-7.52910435e-01 -1.03805208e+00 -3.11607987e-01 5.68603218e-01
1.35694206e-01 -3.91052738e-02 -2.54228324e-01 -2.15465799e-01
1.14356175e-01 8.95097196e-01 -7.99436808e-01 -6.92457706e-02
-4.18744653e-01 -1.42687261e+00 3.24566454e-01 2.48253718e-01
3.32954794e-01 -8.29730034e-01 -5.60708225e-01 1.19680882e-01
-6.66339919e-02 -6.34044111e-01 3.67003143e-01 1.94546804e-01
-8.73449922e-01 -1.51155233e+00 -3.21691096e-01 -1.97460607e-01
4.25684810e-01 3.38390917e-02 9.11326408e-01 -1.76795766e-01
-1.42346218e-01 1.88315183e-01 -3.45569104e-01 -6.18344188e-01
-4.57673609e-01 -2.44987234e-01 5.17245829e-02 6.86806887e-02
3.97968650e-01 -7.67523468e-01 -2.71381289e-01 4.93864477e-01
-1.02550387e+00 -4.01409268e-01 7.45837986e-01 7.64067173e-01
2.35157683e-01 5.28023243e-01 8.57077420e-01 -1.00880742e+00
7.15856731e-01 -9.47736323e-01 -7.63199627e-01 2.46193007e-01
-5.96207857e-01 1.00216698e-02 4.57398295e-01 -6.31994665e-01
-1.11754727e+00 -2.68785298e-01 1.99347869e-01 -1.79559261e-01
-5.12507796e-01 8.58352065e-01 -3.48906994e-01 5.58605373e-01
5.38990200e-01 -2.06857115e-01 -1.53858706e-01 -4.26200241e-01
-8.10347684e-03 4.10812438e-01 1.79134503e-01 -5.29191136e-01
7.03121543e-01 2.23174065e-01 3.87639970e-01 -5.67683816e-01
-4.03135061e-01 -1.90731511e-01 -4.17472661e-01 -4.67651077e-02
7.42393672e-01 -5.62816858e-01 -7.27513850e-01 2.25234464e-01
-9.13352251e-01 -3.05719078e-01 4.87689115e-02 9.64545429e-01
-2.04064563e-01 5.53677119e-02 -1.64550021e-01 -1.29020071e+00
3.27967644e-01 -1.07658732e+00 3.85393679e-01 1.26028672e-01
-2.65620887e-01 -1.07131267e+00 1.96524546e-01 -6.36572838e-02
2.21178547e-01 3.82604450e-01 1.34029090e+00 -4.78097379e-01
-3.01271081e-01 -1.79308310e-01 -6.07552454e-02 -1.52223974e-01
4.06680256e-01 2.56404191e-01 -5.26578724e-01 7.82630220e-02
-2.06060987e-03 2.02343658e-01 5.73429167e-01 8.65822852e-01
8.37286770e-01 -2.44018927e-01 -5.79553425e-01 -1.11647762e-01
1.38233030e+00 6.59035146e-01 7.05835342e-01 2.64396876e-01
3.36393982e-01 7.69160867e-01 5.88609159e-01 4.86857444e-01
1.99928164e-01 7.38216460e-01 5.84409535e-01 -1.69696450e-01
3.44532609e-01 -1.44870758e-01 1.26482293e-01 1.38058469e-01
-2.16734096e-01 -4.48855251e-01 -9.28297281e-01 5.00168383e-01
-1.82776105e+00 -1.17700613e+00 -8.07664394e-01 2.54134345e+00
7.24543750e-01 2.12140977e-01 2.83109546e-01 4.90346432e-01
8.27324033e-01 -2.24798635e-01 -2.59857595e-01 -2.77024090e-01
-1.42000094e-02 3.14042270e-02 4.85040933e-01 5.20061195e-01
-9.62192297e-01 2.60921448e-01 6.98625755e+00 4.47838694e-01
-9.16482687e-01 -1.34359255e-01 7.10371017e-01 3.10718089e-01
-3.51448655e-01 3.66755813e-01 -6.32569373e-01 6.66696131e-01
1.02411270e+00 -4.51017451e-03 -1.14400852e-02 3.35972995e-01
9.99793768e-01 -6.26512051e-01 -8.87192965e-01 4.53609794e-01
-5.32832682e-01 -6.35319531e-01 -2.68113643e-01 1.66202724e-01
7.40677118e-01 -5.96988916e-01 -1.00798830e-01 2.14802146e-01
6.81252062e-01 -1.16183126e+00 2.11470664e-01 7.89200366e-01
4.41090703e-01 -6.91643894e-01 1.07817006e+00 4.79259878e-01
-6.55574322e-01 -3.64094526e-01 -1.64222345e-01 -6.78442001e-01
2.67358840e-01 1.12171137e+00 -6.58819735e-01 7.61374831e-01
6.78268075e-01 2.86376953e-01 -1.57510489e-01 1.10679293e+00
-3.47211927e-01 1.17780232e+00 -4.68238801e-01 -9.65940207e-02
4.74914946e-02 -1.94022402e-01 4.93329614e-01 9.19209898e-01
4.51528877e-01 2.23073423e-01 -3.41077819e-02 9.59327877e-01
4.05289352e-01 9.28792730e-02 -6.88428998e-01 2.47319013e-01
6.14422143e-01 6.37408078e-01 -7.01160491e-01 -6.95227087e-02
-5.13595343e-01 2.73092892e-02 -2.95370489e-01 5.56993306e-01
-8.23943615e-01 7.54293948e-02 3.82613748e-01 4.31763917e-01
-1.37361690e-01 -5.52141294e-02 -4.29763287e-01 -6.98549390e-01
-3.47217560e-01 -8.37841153e-01 5.54091156e-01 -3.84148061e-01
-1.33669317e+00 -1.47752210e-01 7.22287118e-01 -7.30616808e-01
-3.45935732e-01 -2.80692011e-01 -8.43152583e-01 1.03933501e+00
-1.07113290e+00 -7.05883026e-01 -1.47014633e-01 4.83400255e-01
2.82302350e-01 7.50721097e-02 5.23438931e-01 3.63501757e-01
-7.20806897e-01 1.50576964e-01 1.67697877e-01 -8.42549577e-02
6.43460810e-01 -1.19503009e+00 -2.53126234e-01 7.70875931e-01
-2.77623564e-01 7.36597896e-01 1.14659047e+00 -1.05710530e+00
-1.14089394e+00 -6.44008815e-01 1.01368010e+00 -2.85966903e-01
7.88476646e-01 -1.79986909e-01 -7.98365593e-01 5.03845751e-01
-1.04145054e-03 -3.44529510e-01 5.30973554e-01 5.31590462e-01
-1.19492896e-02 -1.64830804e-01 -8.78721058e-01 6.11092210e-01
8.11865032e-01 1.87240854e-01 -5.63360870e-01 1.96867008e-02
3.91825080e-01 2.17925832e-01 -7.34563053e-01 7.74863183e-01
3.79442722e-01 -1.13978684e+00 8.34625363e-01 -5.85137844e-01
4.79917169e-01 -3.48330945e-01 -4.81120981e-02 -1.22174561e+00
-1.47722632e-01 -1.99143916e-01 4.54785734e-01 1.44791973e+00
4.45200384e-01 -7.15357125e-01 2.42805451e-01 7.00103462e-01
4.18334007e-01 -3.76965523e-01 -9.94904399e-01 -7.00612605e-01
1.50673434e-01 -5.49092472e-01 4.39814121e-01 1.04012656e+00
-2.26595625e-01 5.37584245e-01 -5.09222627e-01 2.09321707e-01
5.31690359e-01 5.52839711e-02 8.71229708e-01 -1.47053254e+00
-3.59694213e-01 -3.45801800e-01 -2.97413647e-01 -3.70464295e-01
-7.06260130e-02 -2.33402148e-01 -1.59512743e-01 -1.34665060e+00
2.93182224e-01 -6.28303945e-01 -7.26160109e-02 3.65312368e-01
-4.88197684e-01 -3.78255576e-01 -1.06607184e-01 1.05826929e-01
2.44724020e-01 3.27283531e-01 8.17349732e-01 7.39857033e-02
-4.62465018e-01 4.94640172e-01 -5.22039294e-01 6.55671418e-01
7.74012148e-01 -7.07569778e-01 -6.98754728e-01 1.27284899e-01
3.59848619e-01 4.64444816e-01 7.13368833e-01 -4.81640905e-01
-1.26700491e-01 -6.67198598e-01 1.23633154e-01 -4.41771716e-01
-1.33034736e-01 -8.77177179e-01 7.95361221e-01 4.48932201e-01
-3.73631299e-01 2.20067158e-01 1.09658390e-01 6.53909922e-01
-2.78717667e-01 -3.86568457e-01 3.17941546e-01 -4.46339957e-02
-4.05690640e-01 -3.20874065e-01 -4.49001968e-01 -2.39513695e-01
1.02339709e+00 2.23136060e-02 -1.85761854e-01 -5.15975475e-01
-8.30297828e-01 9.42597464e-02 -1.26409769e-01 3.39495212e-01
1.76566318e-01 -9.86226737e-01 -8.37425828e-01 -2.62249261e-01
-2.74756432e-01 -3.50056827e-01 2.20500231e-01 1.20248938e+00
1.82295561e-01 5.24291098e-01 1.06078371e-01 -4.05645549e-01
-1.25948393e+00 7.51841843e-01 2.09613860e-01 -3.55106443e-01
-1.68678090e-01 2.47381449e-01 4.42425489e-01 -1.06324039e-01
9.42841843e-02 -2.03149229e-01 -4.34592813e-01 2.05508024e-01
3.40283215e-01 6.82698667e-01 -4.14504051e-01 -2.69910425e-01
-3.16057235e-01 4.11083847e-01 4.34355855e-01 -1.79379717e-01
1.22993600e+00 -1.94834962e-01 -2.59323686e-01 9.01330650e-01
6.86895132e-01 8.20673779e-02 -7.97019720e-01 4.51241620e-02
2.79601991e-01 -4.56910014e-01 -1.56520844e-01 -8.99559796e-01
-7.10066199e-01 6.29476905e-01 6.46019340e-01 7.35253036e-01
1.11555004e+00 -7.11412132e-02 -3.16464037e-01 -2.34211072e-01
1.72459066e-01 -6.54163003e-01 -2.25104928e-01 1.63369134e-01
7.57921636e-01 -1.28627634e+00 -9.84539241e-02 -6.33445978e-01
-2.22585633e-01 8.30403030e-01 4.28034306e-01 -1.78338271e-02
7.38894820e-01 3.00073326e-01 -1.99880794e-01 -8.09594989e-02
-6.92214191e-01 -3.93549204e-01 1.48066506e-01 4.68204618e-01
6.92722976e-01 2.39276335e-01 -1.04943740e+00 5.41027725e-01
1.12421419e-02 2.21118301e-01 4.50866878e-01 5.56364298e-01
-1.69859454e-01 -1.14089143e+00 -9.03038919e-01 4.14179116e-01
-6.50452197e-01 -7.37767294e-03 -4.03777599e-01 1.37596774e+00
3.32058281e-01 1.45895088e+00 1.20033398e-02 -8.04604962e-02
4.68162507e-01 -7.23751038e-02 1.69063453e-02 -3.67107868e-01
-4.06260371e-01 2.99605250e-01 2.06197396e-01 -1.32242531e-01
-8.05423141e-01 -8.44661117e-01 -9.31102693e-01 -4.23892200e-01
-4.78905439e-01 3.11080784e-01 3.26887965e-01 1.08329368e+00
-1.12040021e-01 6.03524327e-01 5.36576509e-01 -2.62541115e-01
-4.56015617e-01 -1.17809927e+00 -6.08041644e-01 3.13793391e-01
3.95824127e-02 -1.07708037e+00 -6.29604638e-01 -1.65633410e-01] | [7.799682140350342, 5.236228942871094] |
fd5f7b9f-40d4-411e-91cf-a31e45b88e4c | merlot-reserve-neural-script-knowledge | 2201.02639 | null | https://arxiv.org/abs/2201.02639v4 | https://arxiv.org/pdf/2201.02639v4.pdf | MERLOT Reserve: Neural Script Knowledge through Vision and Language and Sound | As humans, we navigate a multimodal world, building a holistic understanding from all our senses. We introduce MERLOT Reserve, a model that represents videos jointly over time -- through a new training objective that learns from audio, subtitles, and video frames. Given a video, we replace snippets of text and audio with a MASK token; the model learns by choosing the correct masked-out snippet. Our objective learns faster than alternatives, and performs well at scale: we pretrain on 20 million YouTube videos. Empirical results show that MERLOT Reserve learns strong multimodal representations. When finetuned, it sets state-of-the-art on Visual Commonsense Reasoning (VCR), TVQA, and Kinetics-600; outperforming prior work by 5%, 7%, and 1.5% respectively. Ablations show that these tasks benefit from audio pretraining -- even VCR, a QA task centered around images (without sound). Moreover, our objective enables out-of-the-box prediction, revealing strong multimodal commonsense understanding. In a fully zero-shot setting, our model obtains competitive results on four video tasks, even outperforming supervised approaches on the recently proposed Situated Reasoning (STAR) benchmark. We analyze why audio enables better vision-language representations, suggesting significant opportunities for future research. We conclude by discussing ethical and societal implications of multimodal pretraining. | ['Yejin Choi', 'Ali Farhadi', 'Jack Hessel', 'Aditya Kusupati', 'Mohammadreza Salehi', 'Yanpeng Zhao', 'Youngjae Yu', 'Ximing Lu', 'Jiasen Lu', 'Rowan Zellers'] | 2022-01-07 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zellers_MERLOT_Reserve_Neural_Script_Knowledge_Through_Vision_and_Language_and_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zellers_MERLOT_Reserve_Neural_Script_Knowledge_Through_Vision_and_Language_and_CVPR_2022_paper.pdf | cvpr-2022-1 | ['visual-commonsense-reasoning'] | ['reasoning'] | [ 4.35010135e-01 3.26681763e-01 -3.04171652e-01 -2.32226476e-01
-1.04489446e+00 -5.61726689e-01 6.84238613e-01 -1.08473115e-01
-5.39602995e-01 4.80566770e-01 7.86911786e-01 -1.35083914e-01
8.15567374e-02 -3.48260939e-01 -1.09580576e+00 -3.37262034e-01
-1.72059029e-01 1.91321343e-01 -8.19445774e-02 -3.30680460e-01
1.76336303e-01 -1.67282611e-01 -1.72943985e+00 9.63935912e-01
6.72380984e-01 1.07331300e+00 4.81051486e-03 9.31797802e-01
2.57472366e-01 1.52023995e+00 -4.71195906e-01 -8.77034724e-01
5.37414402e-02 -3.40180725e-01 -9.80519176e-01 8.53820965e-02
9.09879744e-01 -7.14791894e-01 -7.24692583e-01 7.96115458e-01
3.56991559e-01 3.57007593e-01 5.27389765e-01 -1.43758333e+00
-1.31826341e+00 7.43259549e-01 -4.98562723e-01 2.70615309e-01
9.06973004e-01 7.39041328e-01 1.29607511e+00 -6.92613900e-01
7.86355019e-01 1.60695875e+00 5.07959545e-01 7.63379037e-01
-1.11050892e+00 -6.02301240e-01 1.72492191e-01 6.65539980e-01
-1.11167097e+00 -7.13256955e-01 3.79690617e-01 -4.59758759e-01
1.25517368e+00 2.57402331e-01 6.24080002e-01 1.82949221e+00
-4.65369150e-02 1.15299439e+00 8.08270037e-01 -1.13118380e-01
6.10463321e-02 -2.11611852e-01 -1.68041587e-01 8.97785306e-01
-2.58033901e-01 -9.21831205e-02 -1.08099282e+00 5.76207489e-02
5.31568050e-01 -1.01738647e-01 -4.25866634e-01 -1.92059144e-01
-1.57386267e+00 6.43112302e-01 4.50920343e-01 -1.30124083e-02
-2.37703159e-01 7.93125808e-01 6.25147104e-01 5.10774493e-01
7.66590834e-02 5.03864408e-01 -2.63262708e-02 -6.20783508e-01
-8.51595879e-01 9.32262987e-02 4.92816269e-01 7.75842369e-01
4.07004118e-01 1.28814071e-01 -2.93066829e-01 6.53814197e-01
4.74275239e-02 7.52636194e-01 3.41619372e-01 -1.64640701e+00
5.02981126e-01 1.38293520e-01 8.92823562e-03 -9.60247040e-01
-2.12822467e-01 -1.37535170e-01 -4.65071619e-01 3.17781270e-02
3.47985178e-01 -8.05429146e-02 -9.76336479e-01 1.80290961e+00
-2.61095256e-01 5.45845032e-01 1.75071716e-01 1.04884291e+00
1.05255723e+00 7.22709537e-01 2.17569262e-01 6.79727271e-02
1.33315814e+00 -9.99288380e-01 -6.32228792e-01 -3.10957432e-01
5.28061390e-01 -3.52008671e-01 1.53085864e+00 9.02681470e-01
-1.18880701e+00 -4.85861152e-01 -8.82727027e-01 -5.43276608e-01
-1.77965954e-01 1.54228851e-01 5.22818446e-01 4.26531434e-01
-1.16256607e+00 4.42609042e-01 -6.50899470e-01 -5.17128825e-01
6.63994849e-01 -8.98651406e-03 -5.64223647e-01 -2.59160519e-01
-1.09489357e+00 8.47595990e-01 6.62531331e-02 -9.46902782e-02
-1.60623860e+00 -7.03618705e-01 -1.05309367e+00 5.20930625e-02
6.31996155e-01 -8.71181667e-01 1.32596946e+00 -1.28001797e+00
-1.36611831e+00 1.16791606e+00 -1.93516329e-01 -9.14067209e-01
5.25359690e-01 -6.95227265e-01 -4.98369515e-01 8.90970588e-01
2.31241092e-01 1.26346529e+00 1.21867514e+00 -1.23989832e+00
-3.56631786e-01 -1.96513403e-02 7.00931430e-01 2.45724007e-01
-2.82432318e-01 -1.12529583e-01 -4.54433441e-01 -4.96305972e-01
-5.39522827e-01 -7.29120076e-01 2.16868728e-01 1.60990536e-01
-3.67514104e-01 -8.41689110e-02 6.32703006e-01 -8.39301705e-01
9.64197397e-01 -2.51107669e+00 4.20697689e-01 -2.41678506e-01
4.33706284e-01 -2.02934265e-01 -6.56384885e-01 4.51340318e-01
-1.48758292e-01 2.51150399e-01 6.39491156e-02 -5.34197450e-01
2.91297466e-01 1.80026621e-01 -6.79886043e-01 4.30810332e-01
3.11404645e-01 1.02536082e+00 -1.07881892e+00 -4.98575568e-01
2.68842995e-01 5.78284740e-01 -9.54900146e-01 -3.87632027e-02
-4.32856679e-01 1.82541639e-01 2.18297035e-01 8.12270761e-01
1.61628053e-01 -4.21776205e-01 -8.67817700e-02 -3.92737985e-01
2.71599710e-01 7.53102750e-02 -5.37897229e-01 2.31213093e+00
-3.47941935e-01 1.19726551e+00 -1.65349454e-01 -9.21455681e-01
2.63269573e-01 3.28268170e-01 2.87281960e-01 -9.59765851e-01
2.06169672e-02 -1.28034070e-01 -6.52828142e-02 -1.03295374e+00
4.56620753e-01 -1.20078616e-01 -1.37043148e-01 1.56606540e-01
4.18415129e-01 -7.83505738e-02 1.19406208e-01 8.01034093e-01
1.22850621e+00 2.18582362e-01 8.16906020e-02 3.61697018e-01
1.07759319e-01 -2.59281863e-02 8.22208226e-02 9.62669849e-01
-3.95956844e-01 6.51626348e-01 9.56051350e-01 -2.26363093e-01
-7.25236833e-01 -1.30739844e+00 3.01376730e-01 1.52803147e+00
2.40257740e-01 -6.98149502e-01 -6.85690224e-01 -6.06422961e-01
4.41451743e-02 1.10501194e+00 -1.04678702e+00 -2.39298984e-01
-2.32063830e-01 5.67807704e-02 9.32109177e-01 5.10768831e-01
4.64487672e-01 -1.07758296e+00 -8.27221632e-01 -2.86507308e-01
-6.46419406e-01 -1.43432355e+00 -2.81417727e-01 -2.65189469e-01
-3.72120589e-01 -1.11334395e+00 -5.95267653e-01 -2.63339311e-01
8.77990723e-02 4.21980888e-01 1.19505715e+00 9.97944325e-02
-4.47259009e-01 1.22795045e+00 -5.45809329e-01 -1.08250082e-01
-1.83861628e-01 -4.58675712e-01 -4.75791879e-02 3.95143479e-02
1.60599291e-01 -4.66663361e-01 -6.36838555e-01 6.25692755e-02
-7.75925517e-01 6.42477348e-02 4.07784343e-01 8.32789779e-01
2.73797005e-01 -2.92964786e-01 3.68797451e-01 -3.47432256e-01
3.25491309e-01 -6.27281308e-01 4.76119518e-02 1.66477337e-01
1.01846352e-01 -1.29157439e-01 3.33544105e-01 -4.37889457e-01
-9.86774981e-01 -3.24390024e-01 1.45025596e-01 -9.68675017e-01
-2.72740930e-01 3.25960129e-01 1.20814152e-01 2.47324139e-01
8.44543576e-01 1.52657405e-01 9.34929028e-02 1.52011132e-02
8.75247836e-01 3.06577444e-01 1.05933368e+00 -7.27675498e-01
5.41526616e-01 8.45781088e-01 -2.89896995e-01 -1.02381074e+00
-8.53293180e-01 -3.40922356e-01 -2.56417632e-01 -4.03199375e-01
1.32731485e+00 -1.34236395e+00 -1.15123665e+00 -4.57657054e-02
-1.27252972e+00 -4.98423994e-01 -3.92633289e-01 3.99150878e-01
-8.99919271e-01 5.02705812e-01 -5.65250874e-01 -8.21370780e-01
8.01424682e-02 -1.04218090e+00 1.23236406e+00 -4.55884635e-02
-4.72070515e-01 -7.33405530e-01 -1.46877065e-01 9.76563334e-01
7.43205398e-02 1.93346933e-01 6.96445465e-01 -5.50671160e-01
-6.92691565e-01 2.70365417e-01 -4.95260447e-01 3.07785600e-01
-4.13704962e-01 8.65943879e-02 -1.26026535e+00 -7.95027465e-02
-5.48630476e-01 -1.09822202e+00 1.41355371e+00 2.15415925e-01
1.33334744e+00 -4.05988574e-01 -3.60178687e-02 6.52891755e-01
1.14516425e+00 -1.05707266e-03 7.87705660e-01 3.79662246e-01
5.53219497e-01 4.72779512e-01 5.19687414e-01 7.05761492e-01
6.08425856e-01 2.87805885e-01 1.06413400e+00 2.06316009e-01
-3.12292635e-01 -3.17779660e-01 8.35904837e-01 1.31036893e-01
-4.23371553e-01 -3.93415719e-01 -8.50890577e-01 6.88874364e-01
-1.83471835e+00 -1.64280617e+00 2.82881767e-01 1.77300549e+00
6.35855079e-01 1.15702832e-02 3.22422385e-01 -1.25716599e-02
2.83388704e-01 4.46262866e-01 -6.00009620e-01 -3.59795094e-01
-2.52216011e-01 1.26884043e-01 3.25129852e-02 5.01425147e-01
-1.06496048e+00 9.67992842e-01 6.52609491e+00 7.27842867e-01
-1.00241017e+00 1.66237816e-01 3.79093498e-01 -6.74838603e-01
-5.11151552e-01 -9.84263048e-02 -1.70608863e-01 2.59136647e-01
8.80559385e-01 1.68486074e-01 9.69247699e-01 6.55231237e-01
1.76238403e-01 -2.23058909e-01 -1.42266142e+00 1.47474384e+00
6.92593634e-01 -1.49002147e+00 3.06781232e-01 -1.65643558e-01
5.13367772e-01 2.40064621e-01 5.25501549e-01 6.26109004e-01
1.49915054e-01 -1.46227252e+00 1.08631837e+00 5.80291390e-01
9.48002338e-01 -5.09464741e-01 4.25958991e-01 2.12998148e-02
-7.62608886e-01 -4.01198119e-01 -1.46796510e-01 -1.01740733e-01
2.11835325e-01 1.21217571e-01 -6.03716373e-01 3.01475406e-01
8.70492280e-01 9.65439975e-01 -7.26220250e-01 7.62473643e-01
-2.36214414e-01 5.65944433e-01 -4.10529077e-02 1.54337570e-01
3.37110192e-01 4.69799310e-01 5.82783818e-01 1.50369918e+00
2.78230876e-01 1.77780718e-01 -6.82375720e-03 8.39147747e-01
-3.98585498e-01 -2.31233507e-01 -7.22672284e-01 -4.86834228e-01
3.68792087e-01 8.45147789e-01 -2.47642726e-01 -3.88585418e-01
-4.60687429e-01 1.35721600e+00 1.98431239e-01 5.92572868e-01
-1.33896494e+00 -2.55234033e-01 6.23318553e-01 -1.19840920e-01
3.80523980e-01 -6.03070445e-02 -1.67115591e-02 -1.41308892e+00
-1.71361133e-01 -1.31717384e+00 5.41426003e-01 -1.46401691e+00
-1.23468435e+00 3.31173807e-01 6.42013997e-02 -1.12430620e+00
-3.21172118e-01 -8.38987768e-01 -4.06706005e-01 3.82718481e-02
-1.37879336e+00 -1.19502342e+00 -5.22319436e-01 9.35719073e-01
7.58932531e-01 -1.82335287e-01 6.59812510e-01 -5.02708182e-02
-2.67755181e-01 5.38003802e-01 -3.72941494e-01 2.32172191e-01
9.60370183e-01 -1.14294541e+00 5.04467897e-02 4.52238709e-01
4.38694000e-01 5.83454132e-01 7.79377937e-01 -3.63317668e-01
-1.48089635e+00 -5.71834445e-01 4.19010371e-01 -7.48912871e-01
1.06817782e+00 -1.23449467e-01 -5.66841424e-01 9.72511292e-01
8.56240928e-01 -2.73132063e-02 9.10691321e-01 2.26615176e-01
-1.04797387e+00 4.21583690e-02 -9.80971396e-01 7.27623403e-01
1.22450995e+00 -1.12019801e+00 -9.29980278e-01 3.06808412e-01
1.04360795e+00 -3.12284976e-01 -6.83823764e-01 2.90845241e-02
9.60633099e-01 -1.28037751e+00 1.14104259e+00 -8.60511959e-01
1.14395833e+00 4.79703583e-02 -6.02821112e-01 -1.21573997e+00
-5.72515130e-02 -9.28271055e-01 -3.64952505e-01 8.51451457e-01
1.29489213e-01 -1.70622602e-01 4.72583711e-01 3.20064574e-01
-9.97298136e-02 -4.60492045e-01 -9.28139031e-01 -6.35796785e-01
-1.91193044e-01 -8.17109048e-01 1.60763845e-01 9.76541102e-01
4.08075273e-01 4.70685929e-01 -6.04672134e-01 6.99851066e-02
6.63741589e-01 -3.19264740e-01 7.85632432e-01 -7.66177714e-01
-6.29306853e-01 -4.64631826e-01 -5.54217696e-01 -1.10386968e+00
4.21455383e-01 -6.35170519e-01 -2.74341613e-01 -1.44224381e+00
3.67954314e-01 5.36395788e-01 -2.59733647e-01 6.83901846e-01
1.12405501e-01 4.67605591e-01 6.00386143e-01 -9.06553715e-02
-1.19056177e+00 4.93507147e-01 1.27610564e+00 -5.48417628e-01
1.90439120e-01 -7.19708979e-01 -7.69087493e-01 8.45542073e-01
4.89415526e-01 4.51904014e-02 -6.83723986e-01 -8.03282857e-01
5.61335206e-01 2.18861148e-01 1.13791907e+00 -9.50282454e-01
-3.91647853e-02 -1.70986593e-01 2.92435318e-01 -4.69668478e-01
9.15498435e-01 -6.02108955e-01 -3.04122865e-01 1.31660029e-01
-7.61549294e-01 -1.06441595e-01 3.19107711e-01 7.93385744e-01
-1.51760757e-01 3.00392620e-02 4.73958015e-01 -1.97591230e-01
-1.11397004e+00 -1.75561011e-01 -3.67719889e-01 3.96482408e-01
8.37244928e-01 -2.23170325e-01 -8.94553304e-01 -9.98862267e-01
-1.05089605e+00 3.26802671e-01 3.52327198e-01 4.52265263e-01
8.85231078e-01 -1.20684052e+00 -6.11119092e-01 -2.36564264e-01
2.88729161e-01 -4.69733149e-01 6.75577879e-01 1.02035189e+00
-3.42580587e-01 3.31043392e-01 -2.59439021e-01 -6.56735182e-01
-1.14160180e+00 7.37546206e-01 1.34616584e-01 3.44079435e-01
-5.60331106e-01 1.11987519e+00 4.50113833e-01 2.30859984e-02
5.63735187e-01 -4.70026284e-01 -4.01233733e-02 1.39154911e-01
6.14497781e-01 4.47463810e-01 -4.75007921e-01 -5.37925363e-01
-3.88581872e-01 3.89760882e-01 2.14482740e-01 -5.09751201e-01
1.22261727e+00 -2.38909379e-01 9.17910933e-02 6.16113305e-01
1.19991481e+00 9.51543450e-02 -1.37722385e+00 8.11155066e-02
-5.10276198e-01 -4.87705499e-01 -1.70427129e-01 -1.09154570e+00
-6.24491274e-01 1.23311269e+00 3.81854624e-01 2.37798057e-02
1.05935192e+00 2.32364267e-01 7.22230732e-01 7.01121330e-01
1.62926644e-01 -9.74250674e-01 8.83415043e-01 3.46130848e-01
1.22269154e+00 -1.32650125e+00 -1.12730116e-01 -1.72959901e-02
-1.21051812e+00 1.10837340e+00 5.06634533e-01 -1.71731547e-01
1.15596868e-01 -5.95973358e-02 -1.54109657e-01 -1.91490844e-01
-1.19896591e+00 -3.60770732e-01 4.37183887e-01 8.98373485e-01
2.58435160e-01 -6.21406697e-02 4.64932352e-01 6.46525145e-01
-1.87676996e-01 1.37157291e-01 5.38975418e-01 5.06903648e-01
-4.55839127e-01 -2.61880130e-01 -3.31383526e-01 7.50867054e-02
-3.52162570e-01 -2.29546204e-01 -4.84789759e-01 9.10322666e-01
2.65524447e-01 1.08214092e+00 1.82757080e-01 -6.03240967e-01
1.45177424e-01 1.00593269e-01 7.80346036e-01 -3.07791948e-01
-3.22530955e-01 -1.95899978e-03 5.30257225e-01 -1.13289297e+00
-5.89852929e-01 -6.51326656e-01 -1.32444894e+00 -3.63129914e-01
1.40720323e-01 -2.83444405e-01 1.73787802e-01 9.05093253e-01
4.01838005e-01 5.15071392e-01 7.49705285e-02 -9.89915490e-01
-5.08249044e-01 -5.99165916e-01 -2.01365337e-01 5.04260063e-01
7.20383286e-01 -7.51549661e-01 -5.18319488e-01 3.80707920e-01] | [10.455110549926758, 1.145527958869934] |
71e38a22-b28f-429c-9ac1-f180cb5213ce | efficient-entity-candidate-generation-for-low | 2206.15163 | null | https://arxiv.org/abs/2206.15163v1 | https://arxiv.org/pdf/2206.15163v1.pdf | Efficient Entity Candidate Generation for Low-Resource Languages | Candidate generation is a crucial module in entity linking. It also plays a key role in multiple NLP tasks that have been proven to beneficially leverage knowledge bases. Nevertheless, it has often been overlooked in the monolingual English entity linking literature, as naive approaches obtain very good performance. Unfortunately, the existing approaches for English cannot be successfully transferred to poorly resourced languages. This paper constitutes an in-depth analysis of the candidate generation problem in the context of cross-lingual entity linking with a focus on low-resource languages. Among other contributions, we point out limitations in the evaluation conducted in previous works. We introduce a characterization of queries into types based on their difficulty, which improves the interpretability of the performance of different methods. We also propose a light-weight and simple solution based on the construction of indexes whose design is motivated by more complex transfer learning based neural approaches. A thorough empirical analysis on 9 real-world datasets under 2 evaluation settings shows that our simple solution outperforms the state-of-the-art approach in terms of both quality and efficiency for almost all datasets and query types. | ['Robert West', 'Akhil Arora', 'Alberto García-Durán'] | 2022-06-30 | null | https://aclanthology.org/2022.lrec-1.690 | https://aclanthology.org/2022.lrec-1.690.pdf | lrec-2022-6 | ['cross-lingual-entity-linking'] | ['natural-language-processing'] | [-1.87032565e-01 1.59960508e-01 -6.72516704e-01 -1.11696310e-01
-1.18291628e+00 -7.24282026e-01 6.72005892e-01 3.80162776e-01
-8.24792862e-01 1.14238334e+00 2.46153012e-01 -3.81407857e-01
-3.38327974e-01 -9.31020737e-01 -9.70446348e-01 -1.48075372e-01
3.42344157e-02 9.03433859e-01 4.76899832e-01 -5.95684707e-01
-6.51434660e-02 3.07078034e-01 -1.43106651e+00 1.69661477e-01
1.07036412e+00 4.88481492e-01 9.90896765e-03 3.41534466e-02
-4.45704490e-01 4.79595691e-01 -5.57756662e-01 -1.10284340e+00
1.98603328e-03 -4.68780324e-02 -9.87615108e-01 -6.29407406e-01
6.34055495e-01 1.89010724e-01 -1.89871147e-01 1.01412773e+00
7.98127294e-01 -2.28882834e-01 4.94792074e-01 -1.14575732e+00
-8.34910631e-01 1.08664513e+00 -2.68230051e-01 5.78449294e-02
4.16660368e-01 -4.62038070e-01 1.40846682e+00 -9.14828062e-01
1.05760980e+00 1.00728202e+00 8.10447514e-01 2.73736298e-01
-1.11665821e+00 -6.96862936e-01 9.71905887e-03 3.59298706e-01
-1.38404906e+00 -3.97778273e-01 5.23125708e-01 -3.15576971e-01
1.08780229e+00 1.36042178e-01 2.25750342e-01 8.26707423e-01
-3.64035398e-01 8.13481629e-01 9.71405029e-01 -7.54776895e-01
-3.07450444e-01 4.32710379e-01 2.31639538e-02 6.11248434e-01
6.90010905e-01 -6.02185875e-02 -5.33840358e-01 -1.91450194e-01
3.28717589e-01 -7.46075034e-01 -2.20216051e-01 -6.34116292e-01
-1.48737276e+00 9.35333729e-01 5.70296347e-01 6.08254015e-01
-2.12711096e-01 -2.16113850e-02 4.91585433e-01 3.66278261e-01
4.82020915e-01 7.66539216e-01 -7.30115175e-01 1.08999513e-01
-8.82837236e-01 3.51051509e-01 1.09372342e+00 1.29982603e+00
7.08261192e-01 -3.84116530e-01 -2.16315433e-01 8.17800701e-01
7.59501234e-02 4.23757762e-01 1.68550953e-01 -7.30437219e-01
9.24998164e-01 7.20389962e-01 3.51595968e-01 -9.24498975e-01
-4.45443243e-01 -6.43683374e-01 -4.46076274e-01 -3.55797648e-01
5.94341218e-01 -2.17572808e-01 -1.56127974e-01 1.89696658e+00
3.15652072e-01 -2.67901748e-01 2.42115438e-01 5.89928091e-01
1.08046150e+00 2.51684576e-01 2.66951889e-01 -4.48931307e-02
1.61978412e+00 -1.10179234e+00 -7.95888722e-01 -8.09195638e-02
7.31344640e-01 -9.29028630e-01 8.95997167e-01 -9.70540568e-02
-1.00848973e+00 -5.46572506e-01 -9.52097833e-01 -3.36148232e-01
-9.40627635e-01 4.16351765e-01 9.10562575e-01 6.96773350e-01
-8.96442592e-01 4.40944374e-01 -4.30623442e-01 -8.25942576e-01
9.10337120e-02 3.81037712e-01 -4.84508365e-01 6.43770769e-02
-1.69998920e+00 1.38341105e+00 8.95319283e-01 -1.23722933e-01
1.06179357e-01 -7.32207775e-01 -6.84473455e-01 -2.92089302e-02
8.68766308e-01 -9.32856798e-01 1.12630236e+00 -4.35782492e-01
-1.05962050e+00 1.01690638e+00 -1.82170853e-01 -5.64843416e-01
5.04029572e-01 -4.51499432e-01 -5.06226540e-01 -3.81702296e-02
3.16104501e-01 8.22860420e-01 1.25945687e-01 -1.09370887e+00
-6.53103828e-01 6.41542375e-02 2.71551669e-01 3.22266042e-01
-4.88879889e-01 2.34438077e-01 -8.35142553e-01 -7.88836718e-01
-2.53668427e-01 -8.38209748e-01 1.35967180e-01 -2.09647655e-01
-3.35741639e-01 -5.07310688e-01 1.48455098e-01 -6.75151527e-01
1.61131656e+00 -1.67278898e+00 2.23837569e-01 -4.67239618e-02
-9.71422419e-02 3.32018465e-01 -1.41380141e-02 8.84632230e-01
9.65536162e-02 2.42264926e-01 -1.14444382e-01 -6.61676973e-02
3.66573095e-01 -1.93101000e-02 -3.64430964e-01 -4.47710305e-02
3.24114919e-01 1.25017738e+00 -1.01789582e+00 -8.82386029e-01
-3.58318418e-01 3.74469340e-01 -4.90146130e-01 6.44787252e-02
-2.41445735e-01 4.91212308e-02 -5.06222904e-01 7.17803001e-01
1.61290035e-01 -2.62802929e-01 4.02114362e-01 -5.52108467e-01
-2.03547567e-01 5.78745306e-01 -1.02235448e+00 1.82587671e+00
-3.08122993e-01 3.92450154e-01 -3.56568336e-01 -7.54004657e-01
7.53113210e-01 5.73964238e-01 4.74723160e-01 -7.09267020e-01
-2.67170429e-01 6.82645142e-01 -1.10195123e-01 -6.05842113e-01
1.05667675e+00 1.49094522e-01 -3.61741364e-01 3.36643726e-01
9.93621945e-02 2.02210233e-01 7.08870292e-01 1.87672868e-01
7.28538156e-01 6.08637810e-01 4.27194625e-01 -1.51044443e-01
4.26025212e-01 3.02373141e-01 3.95497084e-01 7.51992226e-01
1.85852833e-02 1.88343912e-01 2.49868140e-01 -8.82908776e-02
-9.92705226e-01 -7.86111891e-01 -1.82350248e-01 1.24602807e+00
5.76858409e-02 -2.21771941e-01 -6.61727905e-01 -7.68182695e-01
1.29526109e-01 6.30645156e-01 -4.60323423e-01 5.51818758e-02
-8.69286180e-01 -1.02511823e+00 9.85779345e-01 5.36070406e-01
3.65278870e-01 -1.24747252e+00 -4.58873093e-01 3.92328322e-01
-6.32057011e-01 -1.42194843e+00 -8.70581940e-02 1.15419254e-01
-6.67603791e-01 -9.87005413e-01 -7.09562242e-01 -9.17156577e-01
2.12368086e-01 8.15195777e-03 1.52646303e+00 2.59070359e-02
1.38865709e-01 2.78712541e-01 -3.17912340e-01 -3.65562916e-01
-2.86733449e-01 9.49649930e-01 -7.39262924e-02 -4.47023422e-01
8.51416647e-01 -1.41041920e-01 -1.96007311e-01 2.01079190e-01
-7.54051089e-01 -2.28848428e-01 7.99618125e-01 8.27868223e-01
4.18951690e-01 -1.67673096e-01 1.00266743e+00 -1.17362893e+00
6.89850450e-01 -6.69588268e-01 -5.41509032e-01 8.62364113e-01
-7.94385314e-01 3.39205742e-01 1.35980308e-01 -2.49792412e-01
-1.10239971e+00 -2.26540297e-01 -6.89517846e-03 2.32180282e-01
8.09479803e-02 9.57791150e-01 -1.43513545e-01 -4.45326716e-02
5.44808805e-01 -1.29909083e-01 -5.30141771e-01 -6.67017221e-01
7.77013719e-01 4.23152953e-01 5.31687558e-01 -1.06045520e+00
9.15563524e-01 2.31703922e-01 -2.24969164e-01 -3.93283784e-01
-8.43128264e-01 -3.98806840e-01 -8.77816558e-01 1.48254439e-01
6.63835824e-01 -1.07695103e+00 -4.79917794e-01 4.35844623e-03
-1.16881764e+00 7.53258588e-03 -1.55262411e-01 5.03184319e-01
-3.59772831e-01 2.98089027e-01 -6.83701634e-01 -2.21724197e-01
-3.76020998e-01 -9.41037953e-01 8.88621628e-01 1.38282239e-01
-2.61647075e-01 -1.14015543e+00 2.80314445e-01 4.05402988e-01
5.58349550e-01 6.61829188e-02 1.22392213e+00 -1.00242138e+00
-6.16146326e-01 -2.28008419e-01 -4.21106189e-01 -3.59491706e-01
-4.48278487e-02 4.50634194e-04 -8.99531901e-01 -1.04091190e-01
-7.17299700e-01 -4.35244322e-01 8.25338662e-01 -4.05440759e-03
4.69622821e-01 -4.64055426e-02 -5.68433464e-01 2.34814212e-01
1.60241616e+00 -9.29583460e-02 4.32185829e-01 9.70680594e-01
5.05784988e-01 9.39282238e-01 7.51400650e-01 -1.20700620e-01
8.15095246e-01 1.02267241e+00 1.58074543e-01 -2.64558107e-01
-3.51080954e-01 -4.16167259e-01 9.94536355e-02 1.00466216e+00
-1.55461833e-01 -3.99566919e-01 -7.89141119e-01 7.93937743e-01
-1.91269910e+00 -8.77165496e-01 -3.52765201e-03 2.12585306e+00
1.25674546e+00 1.38024494e-01 1.84633508e-01 -1.43308640e-01
7.05319941e-01 -5.54977655e-02 -2.09561899e-01 -5.93487509e-02
-4.98161495e-01 2.28048757e-01 5.89386046e-01 3.89496982e-01
-1.32125843e+00 1.21516871e+00 5.91114044e+00 8.23284507e-01
-7.93346822e-01 1.57179624e-01 2.37552579e-02 2.08619416e-01
-3.88902009e-01 5.96278347e-02 -1.27389610e+00 1.86844230e-01
7.99908459e-01 -4.07673568e-01 1.25674576e-01 5.77956080e-01
-4.43234205e-01 1.59402058e-01 -1.26929998e+00 4.05474901e-01
3.15982550e-02 -1.25056934e+00 -1.81382708e-02 -1.20796412e-01
6.20407581e-01 1.98879242e-01 3.16473506e-02 7.16516435e-01
3.74698192e-01 -7.67177820e-01 6.98926747e-01 4.76499259e-01
7.28534460e-01 -5.52821517e-01 9.35320318e-01 -6.42762519e-04
-1.15696859e+00 2.16214284e-01 -3.52162689e-01 3.93313378e-01
2.97832847e-01 5.39949298e-01 -6.74356461e-01 1.19810915e+00
6.21892571e-01 5.38632751e-01 -6.54796958e-01 1.12050235e+00
-3.50195676e-01 4.11520183e-01 -2.91961432e-01 -8.86254534e-02
2.05146521e-01 6.36407956e-02 3.97475839e-01 1.55820441e+00
3.69817555e-01 -3.64623874e-01 9.95653868e-02 7.79587567e-01
-4.90542233e-01 5.36486506e-01 -6.91068649e-01 -1.34823456e-01
9.29155529e-01 1.27974033e+00 -4.74237651e-01 -3.36167991e-01
-6.95794523e-01 4.37127292e-01 7.16263354e-01 3.70600998e-01
-7.92552590e-01 -6.22545838e-01 1.23013921e-01 -9.70486850e-02
4.20984566e-01 -1.15769759e-01 1.61003266e-02 -1.22023785e+00
3.16795141e-01 -8.43550622e-01 7.69842625e-01 -3.86538118e-01
-1.44003463e+00 7.99090087e-01 3.52448493e-01 -1.25129139e+00
-4.80754673e-01 -4.57967043e-01 4.87471856e-02 8.82561207e-01
-2.15619183e+00 -1.33999002e+00 3.27353850e-02 2.90165216e-01
1.65480554e-01 -2.53911585e-01 1.03443587e+00 9.31771934e-01
-6.00421548e-01 8.15438211e-01 1.17215887e-01 3.54698360e-01
1.23460102e+00 -1.20829213e+00 4.37915146e-01 8.25126588e-01
3.34930360e-01 8.60867083e-01 4.90803629e-01 -6.24475896e-01
-1.24982977e+00 -9.25117135e-01 1.75334811e+00 -4.78881091e-01
9.53828573e-01 -6.97896257e-02 -9.44790483e-01 6.93594158e-01
6.09236598e-01 -2.19797373e-01 7.26588666e-01 4.45524067e-01
-5.27436197e-01 4.12565880e-02 -8.17343652e-01 5.61260641e-01
9.17937815e-01 -5.07189155e-01 -8.60409617e-01 2.76753426e-01
7.97456324e-01 -3.68580103e-01 -1.28469217e+00 5.78914046e-01
4.99052167e-01 -5.79940438e-01 1.19011950e+00 -8.15990090e-01
2.59121150e-01 -3.41908842e-01 -1.68748304e-01 -9.76870537e-01
-2.33329058e-01 -4.06210423e-01 -2.76018858e-01 1.56242311e+00
8.11407149e-01 -6.72943234e-01 5.19310951e-01 3.91610831e-01
1.42921105e-01 -7.60070205e-01 -8.85961711e-01 -8.07485580e-01
3.67983609e-01 -3.21258120e-02 7.72950470e-01 1.20433724e+00
4.98890551e-03 5.23333728e-01 -1.44394264e-01 3.66553329e-02
4.25158352e-01 3.66004974e-01 6.17203355e-01 -1.43008518e+00
-1.36027694e-01 -5.28675199e-01 8.08266327e-02 -7.53431737e-01
3.51790488e-01 -1.14917266e+00 -2.27467403e-01 -1.58701801e+00
2.44442746e-01 -9.17342842e-01 -3.66360635e-01 5.44606149e-01
-4.36674654e-01 3.03691357e-01 1.03357263e-01 3.41917455e-01
-6.48654699e-01 3.53446007e-01 7.74080873e-01 -7.90701360e-02
3.13210599e-02 -1.87240809e-01 -7.46908903e-01 5.05926192e-01
7.18573332e-01 -6.24994457e-01 -1.78549185e-01 -8.76077533e-01
8.72825146e-01 -1.19872019e-01 5.42937368e-02 -7.34133124e-01
2.97835439e-01 2.04704762e-01 -2.91340370e-02 -5.75218916e-01
1.46148488e-01 -8.22472930e-01 6.51151538e-02 3.94850641e-01
-5.91427982e-01 4.27443504e-01 8.83203596e-02 3.33954543e-01
-3.67029816e-01 -3.91596854e-01 1.27254516e-01 -1.08821884e-01
-8.48464131e-01 5.71979247e-02 2.40969762e-01 5.57344913e-01
6.83558285e-01 7.07210451e-02 -5.04936099e-01 -8.50809366e-02
-3.70214880e-01 2.38580689e-01 3.51061672e-01 6.72888041e-01
-2.45317165e-02 -1.42855382e+00 -9.59804654e-01 -3.03568512e-01
4.49654788e-01 -3.93989563e-01 -3.72417420e-01 9.03074324e-01
-4.29120451e-01 8.96475852e-01 -1.20458171e-01 -1.96865067e-01
-9.63198125e-01 6.99709713e-01 3.97742838e-02 -8.23669314e-01
-3.53260964e-01 5.00137627e-01 -3.81595343e-01 -7.45040059e-01
4.06572044e-01 7.85532594e-02 -6.74595535e-01 5.36737382e-01
2.96111293e-02 3.45497340e-01 2.89462060e-01 -6.73246384e-01
-2.71517724e-01 5.89016199e-01 -1.24649338e-01 -1.82518378e-01
1.20604992e+00 -1.20452419e-01 -2.05198526e-01 3.76794428e-01
9.15255010e-01 3.04328531e-01 -4.21537876e-01 -6.83646142e-01
6.87206089e-01 3.91444303e-02 -4.22568589e-01 -9.37500536e-01
-9.13540304e-01 7.54626274e-01 3.13420475e-01 1.15814388e-01
8.54891539e-01 4.01154310e-02 8.77024889e-01 7.72925019e-01
5.82907796e-01 -1.09175050e+00 -3.53347063e-01 4.96859670e-01
6.33445144e-01 -1.32243359e+00 1.08742639e-01 -5.52038789e-01
-4.41558093e-01 9.98399913e-01 4.75670278e-01 3.54251713e-01
3.19121063e-01 4.98994552e-02 3.86446454e-02 -1.80550203e-01
-5.13806105e-01 -5.67710578e-01 5.34288943e-01 4.16082114e-01
8.40541661e-01 -9.05392617e-02 -7.78521419e-01 4.22133088e-01
-2.23696843e-01 2.33829115e-02 1.00054078e-01 9.39862609e-01
-2.26894379e-01 -1.53036392e+00 -8.75176415e-02 7.45370314e-02
-7.56114006e-01 -4.14134324e-01 -2.87865251e-01 1.26019073e+00
1.26616061e-01 6.76046669e-01 -1.87307313e-01 -4.79092337e-02
3.87128353e-01 2.91060150e-01 5.20896196e-01 -4.48417246e-01
-8.54923666e-01 -1.68791071e-01 5.73577583e-01 -3.00221354e-01
-1.02087390e+00 -5.77653646e-01 -9.10426497e-01 -8.25787932e-02
-4.64519799e-01 4.67123091e-01 6.27134562e-01 7.09596872e-01
4.92781669e-01 3.47073197e-01 8.54621083e-03 -2.77038515e-01
-3.79994422e-01 -9.35369909e-01 -4.43775058e-02 3.57659072e-01
-1.73394278e-01 -8.33473027e-01 1.98877409e-01 2.38268692e-02] | [9.503652572631836, 8.850025177001953] |
5b5f4288-da87-4830-a314-e2f4974e2e1e | learning-audio-visual-dynamics-using-scene | 2210.16472 | null | https://arxiv.org/abs/2210.16472v1 | https://arxiv.org/pdf/2210.16472v1.pdf | Learning Audio-Visual Dynamics Using Scene Graphs for Audio Source Separation | There exists an unequivocal distinction between the sound produced by a static source and that produced by a moving one, especially when the source moves towards or away from the microphone. In this paper, we propose to use this connection between audio and visual dynamics for solving two challenging tasks simultaneously, namely: (i) separating audio sources from a mixture using visual cues, and (ii) predicting the 3D visual motion of a sounding source using its separated audio. Towards this end, we present Audio Separator and Motion Predictor (ASMP) -- a deep learning framework that leverages the 3D structure of the scene and the motion of sound sources for better audio source separation. At the heart of ASMP is a 2.5D scene graph capturing various objects in the video and their pseudo-3D spatial proximities. This graph is constructed by registering together 2.5D monocular depth predictions from the 2D video frames and associating the 2.5D scene regions with the outputs of an object detector applied on those frames. The ASMP task is then mathematically modeled as the joint problem of: (i) recursively segmenting the 2.5D scene graph into several sub-graphs, each associated with a constituent sound in the input audio mixture (which is then separated) and (ii) predicting the 3D motions of the corresponding sound sources from the separated audio. To empirically evaluate ASMP, we present experiments on two challenging audio-visual datasets, viz. Audio Separation in the Wild (ASIW) and Audio Visual Event (AVE). Our results demonstrate that ASMP achieves a clear improvement in source separation quality, outperforming prior works on both datasets, while also estimating the direction of motion of the sound sources better than other methods. | ['Anoop Cherian', 'Narendra Ahuja', 'Moitreya Chatterjee'] | 2022-10-29 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [ 2.06348971e-01 -3.25275719e-01 2.31428340e-01 2.46855319e-01
-1.01315689e+00 -8.55113506e-01 5.75979352e-01 -4.18518372e-02
6.71640635e-02 9.90126934e-03 5.10802805e-01 3.75491492e-02
-3.96035761e-02 -1.53909579e-01 -6.28862202e-01 -9.47899997e-01
-1.92596003e-01 1.85507610e-01 4.40179020e-01 3.94862652e-01
1.10223837e-01 5.11738360e-01 -1.90201950e+00 4.04044151e-01
1.89973429e-01 1.26045024e+00 3.78030926e-01 1.28135753e+00
1.37789333e-02 9.77228403e-01 -5.06595731e-01 3.90489548e-01
2.84426361e-01 -6.04285121e-01 -5.27704835e-01 4.21776921e-01
7.67023683e-01 -2.76598096e-01 -4.35790926e-01 7.62306988e-01
4.28703129e-01 1.33216172e-01 7.60136366e-01 -1.52513039e+00
-5.59011735e-02 1.92285940e-01 -7.60664940e-01 4.10658836e-01
6.12322748e-01 5.59382476e-02 9.97507691e-01 -1.00497842e+00
4.09136564e-01 1.34842145e+00 2.76242286e-01 2.98571080e-01
-1.25276315e+00 -5.16232550e-01 2.11023569e-01 8.41654018e-02
-1.34242451e+00 -8.32800448e-01 1.20212138e+00 -1.04764938e+00
7.43653357e-01 2.07728952e-01 5.31800389e-01 1.07645392e+00
-5.79288229e-02 8.65167677e-01 5.47811866e-01 -2.76228905e-01
3.45303982e-01 -6.95920289e-02 -4.61093672e-02 3.12720150e-01
-3.44681561e-01 9.48975384e-02 -8.84780526e-01 -3.02679539e-01
7.86150813e-01 -2.45969415e-01 -6.73545897e-01 -5.32904446e-01
-1.12727642e+00 3.72696400e-01 2.55592227e-01 1.62114322e-01
-3.08569491e-01 2.85831958e-01 -3.81583571e-02 -3.52366492e-02
5.39037645e-01 -1.00097395e-01 -1.39347866e-01 -6.65555596e-02
-1.08464646e+00 3.35013628e-01 5.31486690e-01 5.57792902e-01
3.53457302e-01 2.49361724e-01 7.79891908e-02 6.67914987e-01
8.21932435e-01 5.39114594e-01 3.16863656e-01 -1.21741426e+00
4.17230040e-01 1.90654740e-01 1.95950881e-01 -9.96500611e-01
-2.03198031e-01 -2.93488532e-01 -5.81634700e-01 3.48690987e-01
5.77741504e-01 -9.90203470e-02 -9.39520478e-01 1.89015269e+00
6.93172157e-01 7.84418285e-01 -1.74164593e-01 1.20362926e+00
8.91830444e-01 1.06841159e+00 -2.37986311e-01 -2.09546179e-01
1.24546218e+00 -8.48530948e-01 -4.05939162e-01 -2.85449117e-01
2.02209186e-02 -8.00429106e-01 5.36507666e-01 3.81789654e-01
-1.29851127e+00 -9.32896674e-01 -7.57126451e-01 -6.99470704e-03
1.43798396e-01 6.59375638e-02 -1.26191273e-01 9.78957638e-02
-1.09661853e+00 2.20668942e-01 -9.89247382e-01 3.98742147e-02
1.60180241e-01 -5.57502583e-02 -2.34947786e-01 9.60100368e-02
-6.78598404e-01 1.72693312e-01 -1.14585698e-01 1.32352756e-02
-1.53553271e+00 -9.17625964e-01 -9.75054801e-01 9.25712585e-02
3.03316623e-01 -7.09798992e-01 1.31815088e+00 -1.14685011e+00
-1.21483147e+00 7.58631527e-01 -5.72634161e-01 -2.30716065e-01
4.45857435e-01 -3.13573956e-01 -2.87353635e-01 4.17450994e-01
2.67217308e-01 5.71696758e-01 1.42972577e+00 -1.61858821e+00
-9.80716169e-01 -4.10195172e-01 -2.06598222e-01 4.11723733e-01
1.22878641e-01 -7.14296848e-02 -7.31189191e-01 -6.28860414e-01
2.02713907e-01 -8.22370112e-01 2.41019949e-01 1.28869593e-01
-4.74714130e-01 5.73784634e-02 9.19327796e-01 -7.58420646e-01
9.28135931e-01 -2.61277747e+00 6.92217171e-01 -1.38598964e-01
5.23481011e-01 -3.99783216e-02 -3.08947057e-01 9.46542770e-02
-2.24220529e-01 -3.96631896e-01 -7.62888417e-02 -7.39971876e-01
-1.94769084e-01 -2.67205179e-01 -7.65990257e-01 5.43743610e-01
5.20721860e-02 4.05340016e-01 -1.04838276e+00 -2.30885789e-01
1.49548411e-01 7.10569799e-01 -5.14935255e-01 5.08667171e-01
-9.88011956e-02 5.90816379e-01 -7.59451985e-02 5.51224768e-01
5.79022884e-01 -6.25668615e-02 -1.65354013e-01 -1.84178092e-02
-1.78734884e-02 3.43393892e-01 -1.62020612e+00 1.57269561e+00
-2.28797019e-01 1.03601420e+00 7.53097177e-01 -5.54808974e-01
6.73563480e-01 5.79703689e-01 7.02627361e-01 -3.14332515e-01
-2.46649031e-02 2.11839005e-01 -2.50653386e-01 -5.33742428e-01
2.93989033e-01 -1.82128772e-01 1.51801169e-01 3.52888763e-01
2.86216646e-01 -2.10837841e-01 -7.91568458e-02 3.05575013e-01
1.01143360e+00 1.06037417e-02 -1.55228868e-01 1.50552496e-01
4.27568167e-01 -4.09180045e-01 4.12395746e-01 5.45536101e-01
-2.99388081e-01 9.08648729e-01 4.06071514e-01 6.66669244e-03
-7.55037129e-01 -1.54110205e+00 1.89291596e-01 8.31481576e-01
2.70647705e-01 -4.54597652e-01 -6.01005197e-01 -4.97515291e-01
1.12933554e-01 3.29345226e-01 -3.44619244e-01 -8.59218612e-02
-5.41336000e-01 -3.09529215e-01 3.82945627e-01 3.76436293e-01
6.02389760e-02 -6.66055083e-01 -8.00221622e-01 9.73049775e-02
-5.60523391e-01 -1.29600382e+00 -6.35306716e-01 2.09741443e-01
-5.03301084e-01 -9.48180377e-01 -6.63109779e-01 -7.22030282e-01
1.37637541e-01 7.71467328e-01 9.15917635e-01 -4.11801785e-01
-1.97312772e-01 7.81005323e-01 -4.91589531e-02 -4.08642769e-01
-3.91627192e-01 -5.03575027e-01 2.55672485e-01 7.12858379e-01
-2.53114820e-01 -8.86003017e-01 -5.37175000e-01 2.10108817e-01
-8.56865168e-01 1.79322168e-01 1.04130737e-01 2.62421697e-01
6.18719816e-01 3.11569721e-01 2.55418748e-01 9.94051844e-02
1.13341115e-01 -7.05431104e-01 -4.89199877e-01 -2.67908722e-01
3.39263111e-01 -2.64862180e-01 5.03973901e-01 -6.44603491e-01
-9.13510382e-01 4.28087801e-01 1.63484991e-01 -1.00092888e+00
-5.28716326e-01 -1.12315994e-02 -3.08489352e-01 4.12061036e-01
3.79670739e-01 2.59900302e-01 -2.12591380e-01 -7.68133283e-01
2.13294059e-01 6.52028739e-01 9.04709637e-01 -1.49796233e-01
7.40983844e-01 8.44687283e-01 1.06919140e-01 -1.27689517e+00
-8.43125224e-01 -7.92914152e-01 -5.37839830e-01 -5.24393737e-01
1.04503524e+00 -1.23200071e+00 -4.74464238e-01 6.62394285e-01
-1.25825548e+00 -4.41433519e-01 -2.38147601e-01 6.32270336e-01
-4.82522726e-01 1.88439772e-01 -4.63184237e-01 -1.12947226e+00
2.85716712e-01 -1.16469920e+00 1.58745897e+00 6.90207034e-02
-1.92704409e-01 -8.28597188e-01 5.34399033e-01 4.76705372e-01
-2.06487209e-01 2.35593900e-01 7.43206084e-01 -3.59155536e-01
-7.07849741e-01 -7.67215416e-02 7.95664489e-02 3.38752538e-01
1.44053534e-01 2.23859981e-01 -1.62414205e+00 -9.92428213e-02
3.47886145e-01 4.75753620e-02 9.83040631e-01 8.85118484e-01
5.11691213e-01 -3.06608230e-02 -2.52967417e-01 7.01394141e-01
1.15102673e+00 3.52465242e-01 1.58790033e-02 -1.74150988e-01
9.72222149e-01 8.48488271e-01 3.57241303e-01 3.41693372e-01
2.09298760e-01 1.05753636e+00 8.00059795e-01 -2.71000117e-02
-6.65406048e-01 -4.39714402e-01 7.46277869e-01 6.53333366e-01
2.50188172e-01 -4.06719178e-01 -8.35461736e-01 7.25861371e-01
-1.80551434e+00 -9.00028527e-01 -2.15237364e-01 2.12501001e+00
4.86991793e-01 -6.99598938e-02 4.48821843e-01 4.12033439e-01
7.21872032e-01 4.55889255e-01 -4.77141827e-01 2.11764425e-01
-1.36135697e-01 -1.00645602e-01 -2.42935389e-01 8.60060632e-01
-1.28403175e+00 4.56034631e-01 5.15938091e+00 6.46693826e-01
-1.30672312e+00 -2.09288783e-02 3.45871389e-01 -6.82334840e-01
-2.27864541e-04 1.07395034e-02 -7.52610803e-01 5.47298312e-01
8.82080078e-01 1.56749204e-01 4.68489647e-01 4.32312429e-01
4.02723640e-01 -2.01481253e-01 -1.35332537e+00 1.23848951e+00
1.44146934e-01 -1.19361281e+00 8.74752700e-02 8.57212208e-03
4.61156994e-01 3.76386903e-02 1.57171249e-01 -1.88254133e-01
2.78070103e-02 -7.78496146e-01 1.45846403e+00 3.92639130e-01
4.08456445e-01 -4.80119944e-01 7.87496641e-02 5.53957283e-01
-1.52494276e+00 -1.01097062e-01 1.19907767e-01 -6.85112849e-02
4.22037870e-01 5.48077762e-01 -6.05821967e-01 5.32303989e-01
9.63696063e-01 9.18599486e-01 -2.22461358e-01 1.11690855e+00
-1.82390407e-01 5.99243522e-01 -3.71124953e-01 6.39841557e-01
7.98343346e-02 -6.25961600e-03 1.32297528e+00 1.28690863e+00
3.81616026e-01 -1.85547516e-01 8.57245028e-02 9.95771587e-01
1.43918559e-01 -3.11320007e-01 -6.07155621e-01 1.60020396e-01
3.58517200e-01 1.15027165e+00 -7.30689585e-01 -1.02933571e-01
-1.83555305e-01 8.23277056e-01 -1.17095284e-01 6.01409972e-01
-8.53847980e-01 -1.73779607e-01 1.09971857e+00 2.50340223e-01
5.62164485e-01 -2.33530104e-01 1.73249125e-01 -1.08013606e+00
6.83185607e-02 -5.72696328e-01 3.40631336e-01 -1.21774328e+00
-8.50433350e-01 5.61806738e-01 -3.75855006e-02 -1.34275496e+00
-2.42187753e-01 -4.54288602e-01 -6.44340515e-01 8.75547767e-01
-1.36265874e+00 -7.71576881e-01 -3.72726947e-01 7.72073627e-01
6.80874586e-01 2.27004617e-01 2.59014934e-01 2.59367704e-01
-5.35031199e-01 7.70505867e-04 -2.23267674e-02 6.53763711e-02
5.42250276e-01 -1.05131519e+00 1.15564473e-01 9.98889387e-01
6.68890893e-01 -1.53961986e-01 6.63551450e-01 -4.53845859e-01
-1.35783386e+00 -1.16210544e+00 7.99233019e-01 -6.47779644e-01
6.97071552e-01 -5.38668573e-01 -7.96381414e-01 5.35088480e-01
-1.60042211e-01 5.64864688e-02 4.27278697e-01 -6.20118558e-01
-3.39172244e-01 -7.51672313e-02 -6.37690842e-01 4.20081377e-01
9.46578979e-01 -8.22083116e-01 -4.91503209e-01 5.26682511e-02
7.38596559e-01 -4.04225886e-01 -2.20795199e-01 1.45179495e-01
4.25556988e-01 -1.14625430e+00 1.39538336e+00 -3.92613798e-01
5.70306480e-01 -5.91522634e-01 -4.16000456e-01 -1.19783604e+00
-2.04576433e-01 -7.48205960e-01 -5.30019760e-01 1.35326755e+00
-6.77223280e-02 -6.00285083e-02 3.87523264e-01 1.84744615e-02
-4.28135656e-02 -3.95343572e-01 -1.10898530e+00 -7.18611538e-01
-4.13067549e-01 -9.22614813e-01 9.33506265e-02 7.33181536e-01
-3.10493290e-01 6.94326222e-01 -2.86529869e-01 7.17583716e-01
7.39391685e-01 2.99906284e-01 8.39625299e-01 -1.21586227e+00
-6.83446586e-01 -5.23763001e-01 -4.66516554e-01 -1.34180975e+00
1.63059428e-01 -7.77474284e-01 2.78001428e-01 -1.33829212e+00
-1.31710516e-02 4.84800227e-02 -1.26903132e-01 4.08170791e-03
-7.64044225e-02 1.20877482e-01 3.77443254e-01 4.74028438e-01
-3.35272163e-01 5.37975550e-01 9.13336098e-01 -1.80152804e-01
-5.64490676e-01 1.83800787e-01 -3.39253664e-01 1.04349124e+00
2.23839238e-01 -4.83294368e-01 -5.82269847e-01 -6.68977797e-01
-1.43996468e-02 6.20980620e-01 8.40328753e-01 -1.01504529e+00
2.61636078e-01 3.84118743e-02 2.62423664e-01 -6.28343821e-01
6.93536282e-01 -8.18171620e-01 2.29229525e-01 1.01139583e-01
-3.09109420e-01 -3.26447487e-01 3.70991915e-01 9.21867132e-01
-3.75336856e-01 8.84916708e-02 7.69513905e-01 1.60745278e-01
-4.31284219e-01 1.73073068e-01 -5.93221962e-01 2.21672535e-01
9.40291166e-01 -1.59291834e-01 -3.76323424e-02 -6.53564870e-01
-9.78383362e-01 -5.16675599e-02 1.22379549e-01 4.85081702e-01
7.38514066e-01 -1.35931814e+00 -6.98819160e-01 4.37054873e-01
-2.12415397e-01 2.74777889e-01 4.70650911e-01 9.21943903e-01
-1.20716095e-01 2.96365559e-01 1.56429186e-01 -1.11808455e+00
-1.53561354e+00 6.29648268e-01 5.83637476e-01 9.17213932e-02
-6.44344270e-01 1.19134784e+00 1.00515091e+00 1.47221595e-01
6.94323957e-01 -5.29132903e-01 -1.96548209e-01 2.74069846e-01
6.99943364e-01 6.01185501e-01 -2.59711355e-01 -9.67138052e-01
-4.27220553e-01 8.33941162e-01 5.44654608e-01 -5.87166190e-01
1.29066551e+00 -3.55953157e-01 3.85920727e-03 9.79968369e-01
1.43791509e+00 4.43567812e-01 -1.82116425e+00 -2.82497823e-01
-3.33988488e-01 -5.12799919e-01 2.71696210e-01 -4.03231770e-01
-1.16686928e+00 1.21365905e+00 4.65196759e-01 4.31376427e-01
1.33806241e+00 4.76021171e-01 6.10778153e-01 -4.39864725e-01
-5.04553542e-02 -4.26738858e-01 4.76989150e-01 2.99397469e-01
8.32492352e-01 -7.44246602e-01 -4.65519845e-01 -3.97878885e-01
-4.79623914e-01 1.07296276e+00 2.38111690e-01 -1.05198182e-01
7.64641643e-01 3.94116014e-01 1.46350995e-01 -2.76510477e-01
-8.87195349e-01 -1.81977734e-01 7.30144143e-01 5.64442456e-01
5.45866322e-03 -2.39948317e-01 1.02467382e+00 6.34892583e-01
-6.02129400e-02 -4.33725029e-01 3.01442116e-01 7.45205641e-01
-3.82667929e-01 -4.26270634e-01 -7.62619555e-01 -2.74701566e-01
-3.04099619e-01 1.17471933e-01 -7.91736901e-01 3.23238164e-01
2.28729770e-01 1.19880557e+00 3.76656175e-01 -4.29528922e-01
3.22603196e-01 1.46356329e-01 4.39332426e-01 -5.13877034e-01
-2.08329722e-01 9.04319346e-01 -3.05588156e-01 -5.90161502e-01
-4.04204369e-01 -7.72241831e-01 -1.07473958e+00 1.79304898e-01
-2.36922614e-02 5.04618026e-02 3.90454680e-01 6.88812613e-01
3.66627127e-01 6.90104246e-01 7.47454703e-01 -1.51392889e+00
-1.52049676e-01 -5.19124210e-01 -8.16914082e-01 3.61432791e-01
1.16183233e+00 -5.73604524e-01 -9.80355918e-01 4.56798345e-01] | [14.85724925994873, 4.979794979095459] |
d4c57173-f9c7-46fd-8ca3-19984984f782 | recent-advances-in-transient-imaging-a | 1611.00939 | null | http://arxiv.org/abs/1611.00939v1 | http://arxiv.org/pdf/1611.00939v1.pdf | Recent Advances in Transient Imaging: A Computer Graphics and Vision Perspective | Transient imaging has recently made a huge impact in the computer graphics
and computer vision fields. By capturing, reconstructing, or simulating light
transport at extreme temporal resolutions, researchers have proposed novel
techniques to show movies of light in motion, see around corners, detect
objects in highly-scattering media, or infer material properties from a
distance, to name a few. The key idea is to leverage the wealth of information
in the temporal domain at the pico or nanosecond resolution, information
usually lost during the capture-time temporal integration. This paper presents
recent advances in this field of transient imaging from a graphics and vision
perspective, including capture techniques, analysis, applications and
simulation. | ['Adrian Jarabo', 'Julio Marco', 'Diego Gutierrez', 'Belen Masia'] | 2016-11-03 | null | null | null | null | ['pico'] | ['natural-language-processing'] | [ 5.89695871e-01 -9.75166917e-01 6.43396437e-01 5.34851141e-02
-3.47054958e-01 -7.24834383e-01 5.35360992e-01 -9.85914320e-02
-4.55727994e-01 6.37265563e-01 -1.46709725e-01 -2.16276065e-01
1.00951791e-01 -5.09596050e-01 -1.76143900e-01 -1.08539450e+00
2.02353567e-01 1.65589288e-01 5.89170814e-01 3.32378566e-01
4.39560264e-01 7.83739328e-01 -1.70159984e+00 1.20402150e-01
4.52895820e-01 9.51526463e-01 1.73764542e-01 1.00611579e+00
-1.92855007e-03 8.37564647e-01 -8.44552442e-02 -7.93767348e-03
-1.10924847e-01 -5.12788236e-01 -1.68379713e-02 -1.43722426e-02
4.93430406e-01 -5.87607086e-01 -5.10182858e-01 9.41778481e-01
3.63563925e-01 2.50980526e-01 3.26171130e-01 -7.32528269e-01
-1.61071777e-01 -2.93465167e-01 -7.72181869e-01 5.91855168e-01
8.25022995e-01 5.14874160e-01 4.19930257e-02 -7.79134929e-01
8.80032063e-01 8.99680495e-01 8.19550991e-01 4.90884751e-01
-1.05880415e+00 -3.61784220e-01 -3.91271442e-01 9.41009726e-03
-1.01679230e+00 -4.87466097e-01 9.44384873e-01 -5.48197150e-01
7.65667140e-01 2.94065684e-01 8.79348516e-01 8.79912734e-01
5.94586372e-01 2.65816897e-01 1.55561650e+00 -3.31086069e-01
2.69390166e-01 -1.01739511e-01 2.74922073e-01 8.23813379e-01
1.34354487e-01 5.69375515e-01 -4.82664317e-01 -3.20360154e-01
1.07766593e+00 2.07355246e-01 -5.60862124e-01 5.54200262e-03
-1.05848897e+00 1.54570431e-01 1.26611426e-01 1.07594617e-01
-4.11611021e-01 3.93665940e-01 1.41039878e-01 1.05389878e-01
5.41936517e-01 3.00897479e-01 4.38271612e-02 -5.57201684e-01
-9.27128315e-01 1.59638748e-01 6.21680737e-01 3.85262668e-01
4.72552210e-01 1.70380790e-02 1.10877462e-01 8.99702404e-03
2.29707345e-01 9.56103981e-01 8.77496898e-02 -1.35714281e+00
-1.92563891e-01 1.09269120e-01 6.42965376e-01 -8.10798407e-01
-1.39363125e-01 5.04954569e-02 -8.05505991e-01 8.02593887e-01
5.06839812e-01 -1.58259958e-01 -8.71388376e-01 1.05572307e+00
6.89454377e-01 7.00186670e-01 -1.52659893e-01 1.09506822e+00
6.85878754e-01 8.88688266e-01 -3.97513896e-01 -7.44222522e-01
1.02387369e+00 -4.39042032e-01 -5.87205112e-01 -6.24946468e-02
-1.41637117e-01 -9.91701961e-01 6.68350816e-01 6.24323845e-01
-1.41401052e+00 -2.42443144e-01 -6.66101694e-01 -3.86906080e-02
1.98884785e-01 -4.37393636e-01 5.45801759e-01 2.60273278e-01
-8.95389378e-01 6.84232116e-01 -1.08700967e+00 -1.44223511e-01
2.04500794e-01 -5.52780293e-02 9.56043974e-02 -5.09883165e-01
-4.28640693e-01 4.36565071e-01 -5.15102744e-01 -1.50755197e-01
-5.61299860e-01 -1.02558446e+00 -2.73510456e-01 -1.61067978e-01
7.17474297e-02 -1.10793173e+00 1.11019492e+00 -5.91530323e-01
-1.60161293e+00 7.03652322e-01 -5.50765216e-01 -2.45901197e-01
6.44169509e-01 -4.72930968e-02 -3.14087778e-01 5.85933805e-01
-2.26910621e-01 -1.74392760e-01 9.27855313e-01 -1.12621105e+00
-2.86440462e-01 -4.04682070e-01 -1.88221976e-01 -1.42767662e-02
3.60103816e-01 2.49150917e-01 -3.61233115e-01 -1.13417283e-01
2.10897662e-02 -8.40255797e-01 -3.11797857e-01 4.31028694e-01
-1.38344139e-01 4.27322417e-01 1.12013721e+00 -1.67737827e-01
7.80962765e-01 -1.98739302e+00 -3.15653905e-02 -3.14506918e-01
4.63760823e-01 4.90723461e-01 2.66753584e-01 7.14235067e-01
3.86530727e-01 -5.35594881e-01 -1.81072816e-01 -4.60189998e-01
-7.27226913e-01 -1.34237304e-01 -6.44003808e-01 5.82337856e-01
-4.26154166e-01 6.09386444e-01 -9.29246008e-01 -2.14722633e-01
6.75246596e-01 8.64934564e-01 -1.77359655e-01 -6.42132573e-03
-2.44570658e-01 1.11781907e+00 -6.84552550e-01 6.63685739e-01
9.67306554e-01 -5.26365221e-01 -1.22207753e-01 -2.67538399e-01
-1.07059944e+00 -1.40623942e-01 -9.79241788e-01 1.11877644e+00
-5.64907908e-01 1.10694838e+00 3.83528978e-01 -3.73384953e-01
2.32795149e-01 2.36367986e-01 7.33066738e-01 -9.30168271e-01
1.84601337e-01 4.29615602e-02 -4.95630413e-01 -7.88727999e-01
4.75504756e-01 -5.80858707e-01 3.69565427e-01 6.30218625e-01
-7.88084149e-01 -2.28405163e-01 -2.65640337e-02 8.29562396e-02
1.38896179e+00 -1.23516820e-01 3.01118046e-02 2.76412338e-01
2.14616656e-01 2.24477232e-01 1.22694299e-01 7.93618381e-01
-1.70416653e-01 7.84950376e-01 -2.08406925e-01 -5.97523391e-01
-1.05132592e+00 -1.15848148e+00 -1.47731021e-01 3.73966187e-01
4.69240427e-01 -1.02061786e-01 -3.47443193e-01 1.19004279e-01
-5.47993481e-02 2.76619792e-01 -3.63006383e-01 1.47830471e-01
-8.04498255e-01 -4.56375927e-01 6.60111308e-02 3.26313943e-01
5.29753864e-01 -9.22347367e-01 -1.39787686e+00 2.48657733e-01
1.22772276e-01 -1.44062626e+00 -1.89265400e-01 -3.84010941e-01
-8.48835647e-01 -1.10133708e+00 -7.12057173e-01 -2.98763007e-01
3.64498138e-01 9.48146760e-01 1.06312466e+00 -2.85184179e-02
-9.72917259e-01 9.07447815e-01 -8.83824229e-02 -3.16420376e-01
-1.61934674e-01 -9.99924481e-01 6.94888160e-02 1.89766049e-01
-1.85166836e-01 -8.70458066e-01 -8.69215965e-01 1.33261815e-01
-7.22197652e-01 2.87915975e-01 2.41569176e-01 1.99732572e-01
5.77591956e-01 -4.24090140e-02 -2.99827754e-01 -7.59948075e-01
1.95184395e-01 -2.52508074e-01 -9.73845780e-01 -5.39259659e-03
-2.17944488e-01 -3.20478141e-01 8.16700876e-01 -5.96721411e-01
-1.29923689e+00 -1.95093378e-01 1.91629454e-01 -6.36792541e-01
-2.29465067e-01 4.13012594e-01 4.14985329e-01 -4.83578235e-01
4.95709896e-01 3.27357590e-01 -1.35536939e-01 -3.16570729e-01
3.97428349e-02 1.38669580e-01 6.95421934e-01 -3.93632799e-01
6.51014090e-01 1.52348554e+00 4.24791276e-01 -1.43576026e+00
-7.66916633e-01 -6.74261510e-01 -2.70305306e-01 -6.43388212e-01
7.11925447e-01 -5.57295680e-01 -1.02876580e+00 8.64406645e-01
-1.21935308e+00 -4.55374807e-01 -4.18642491e-01 8.44053328e-01
-3.21579069e-01 3.30300629e-01 -7.08393633e-01 -9.67390716e-01
-1.16788737e-01 -6.97080493e-01 1.12122452e+00 6.77772522e-01
1.14786223e-01 -1.08470750e+00 5.50552011e-01 6.08800054e-02
7.22692847e-01 5.68723679e-01 6.10945642e-01 8.48826170e-01
-1.24909484e+00 -2.26313606e-01 -4.99133110e-01 -3.36205512e-01
-6.26614094e-02 4.79499817e-01 -9.21898067e-01 -1.76374182e-01
5.70011318e-01 1.81118771e-02 8.99241269e-01 1.05105436e+00
7.45430171e-01 1.11886814e-01 -6.54503226e-01 1.04090655e+00
1.85034811e+00 3.01736504e-01 4.32511747e-01 -2.64405042e-01
6.54803216e-01 3.04999739e-01 2.74682552e-01 6.78763092e-01
-1.35892496e-01 7.62290120e-01 2.34687716e-01 4.20310162e-02
-3.64398211e-01 3.36772174e-01 1.82386756e-01 7.30239689e-01
-3.80444437e-01 -4.09758717e-01 -8.16426873e-01 1.50462002e-01
-1.54883075e+00 -1.38345361e+00 -5.28831482e-01 2.13205194e+00
3.54219794e-01 -1.61046222e-01 -2.68619239e-01 -9.46360454e-02
4.14221227e-01 1.51294664e-01 -7.82970428e-01 -1.59097672e-01
5.62888905e-02 1.31854638e-01 3.90967011e-01 6.97178543e-01
-4.65459704e-01 5.73894799e-01 6.87836933e+00 1.90624624e-01
-1.69263196e+00 5.21720573e-02 1.22611828e-01 -2.10455418e-01
-4.61200565e-01 1.46978661e-01 -6.00998402e-01 4.80658054e-01
7.81578839e-01 -2.72923827e-01 5.07209182e-01 1.64428085e-01
7.56398439e-01 -8.52520108e-01 -7.80949831e-01 1.23860312e+00
-8.44098628e-02 -1.47616982e+00 -2.07869578e-02 9.75725874e-02
7.49259949e-01 2.38909543e-01 5.44335127e-01 -6.61440611e-01
4.13675159e-02 -7.81227767e-01 3.67113441e-01 1.11122763e+00
9.54188883e-01 -4.19468611e-01 1.23483092e-01 5.75982034e-01
-1.17546380e+00 2.73182303e-01 -2.14097321e-01 -3.27633321e-01
8.64004076e-01 1.09868252e+00 -3.53024870e-01 1.28991038e-01
6.20371521e-01 7.00920820e-01 -9.26811472e-02 1.37992227e+00
2.23653898e-01 6.21146023e-01 -4.75017428e-01 -3.72271240e-02
2.54886448e-01 -7.57954776e-01 8.71303737e-01 8.92519295e-01
5.92520714e-01 8.07425678e-01 -2.03503855e-02 7.99384177e-01
3.40385109e-01 -8.20443094e-01 -6.54410064e-01 7.11589530e-02
1.30171612e-01 1.32178879e+00 -9.80185211e-01 -1.38384119e-01
-5.45112848e-01 9.33075607e-01 -2.84080446e-01 5.59861004e-01
-8.85429502e-01 -1.79506049e-01 6.29493237e-01 6.11718774e-01
1.56798959e-01 -8.65739584e-01 -1.66117260e-03 -1.15501475e+00
2.07512472e-02 5.15813269e-02 -2.40455627e-01 -1.30071783e+00
-1.08845294e+00 3.44864845e-01 -1.13544889e-01 -1.14153111e+00
9.84755009e-02 -5.92044771e-01 -9.25930619e-01 7.51840591e-01
-1.48653519e+00 -7.48948574e-01 -5.77657461e-01 8.11777413e-01
3.22886348e-01 5.64557791e-01 4.50837106e-01 1.61327049e-01
-1.27704889e-01 -4.34544116e-01 8.48892093e-01 -3.34566772e-01
3.89812648e-01 -7.57875919e-01 2.86192656e-01 8.46194506e-01
-1.78236272e-02 6.06372774e-01 1.04119146e+00 -4.24662918e-01
-1.86534965e+00 -6.02066696e-01 3.46397966e-01 -4.70348835e-01
5.54714680e-01 -1.38521314e-01 -8.48925829e-01 1.77412465e-01
-7.87726603e-03 6.07248068e-01 2.66409338e-01 -5.65926075e-01
-1.14304513e-01 1.55838668e-01 -1.27318501e+00 4.66315389e-01
9.15663660e-01 -7.34523475e-01 -4.07698527e-02 1.16750322e-01
2.65757948e-01 -5.09877205e-01 -1.48631260e-01 1.34291485e-01
9.84104693e-01 -1.46763706e+00 1.00589716e+00 -1.27646506e-01
2.98732281e-01 -2.86151826e-01 2.03585327e-01 -1.02018869e+00
-4.30043414e-02 -8.86096120e-01 -1.22306153e-01 8.33765805e-01
-3.17556411e-01 -6.28330767e-01 9.51367915e-01 6.70206666e-01
-1.74206242e-01 -3.38044465e-01 -1.05439723e+00 -5.49802542e-01
-5.04266560e-01 -4.71292913e-01 -1.41335607e-01 7.62169659e-01
-5.21015465e-01 1.49868861e-01 -2.35557377e-01 2.88870573e-01
1.08659184e+00 7.35274374e-01 4.60898399e-01 -1.43086219e+00
-3.65419239e-01 3.63975652e-02 -2.84904599e-01 -1.29096711e+00
-2.20284209e-01 -3.37070733e-01 3.00742723e-02 -1.40700221e+00
2.93374002e-01 -2.70980835e-01 1.41649023e-01 -6.78421855e-01
1.63458601e-01 4.06837970e-01 1.01191908e-01 4.57249582e-01
-6.21302843e-01 4.24785204e-02 1.40381658e+00 2.78393358e-01
-1.06523484e-01 1.39337212e-01 3.85446325e-02 9.72392321e-01
2.86552310e-01 -2.71351993e-01 -2.39477351e-01 -6.87126935e-01
3.99091452e-01 5.03904223e-01 8.72223496e-01 -1.36269188e+00
6.89072669e-01 -2.88288385e-01 4.24064279e-01 -5.15647233e-01
7.68236399e-01 -7.79583931e-01 4.35893148e-01 3.78097206e-01
4.91476953e-02 1.43740714e-01 1.83049724e-01 7.98908114e-01
-2.97547281e-02 -2.39172596e-02 1.10905230e+00 -4.89769667e-01
-6.04573369e-01 4.30531681e-01 -7.03672588e-01 1.74078837e-01
1.01911604e+00 -3.63818347e-01 -5.25354385e-01 -3.25426072e-01
-5.32249868e-01 -3.34720165e-01 9.18791890e-01 -3.65288585e-01
9.86867964e-01 -9.22310650e-01 -3.49201888e-01 2.00858116e-01
-2.25823507e-01 -3.28455180e-01 6.31494999e-01 1.00288820e+00
-8.63040507e-01 2.78734922e-01 -3.91300172e-02 -7.91782796e-01
-1.58030832e+00 3.94356728e-01 2.77455539e-01 6.07352331e-02
-1.04769278e+00 8.73181343e-01 1.65016294e-01 4.82686162e-01
-4.11756545e-01 -2.76579320e-01 2.82070160e-01 -2.97015965e-01
8.78871381e-01 7.32575238e-01 -2.19854251e-01 -4.47180986e-01
-1.75345868e-01 1.41271579e+00 1.77445441e-01 -4.76227254e-01
1.24558020e+00 -4.10802424e-01 -4.03134674e-02 8.34312558e-01
1.13006055e+00 4.48369831e-01 -1.60130966e+00 -2.69242465e-01
-6.07842267e-01 -7.71443367e-01 1.97609067e-01 -5.04896045e-01
-8.12348247e-01 1.18256247e+00 4.00606632e-01 4.13933694e-01
1.08984685e+00 1.16480500e-01 1.23672295e+00 1.32252090e-02
6.28022313e-01 -5.67999363e-01 5.63204251e-02 5.34688950e-01
1.50551006e-01 -9.38458443e-01 9.93605256e-02 -4.70180333e-01
-1.41423836e-01 1.35558593e+00 2.88201068e-02 -2.77594328e-01
8.57630968e-01 9.66722190e-01 -2.41762195e-02 -3.47328752e-01
-8.16701949e-01 1.10055111e-01 -1.36855394e-01 5.33558309e-01
2.92313069e-01 -2.95882165e-01 1.02766998e-01 -4.39239532e-01
4.13584709e-01 2.04619542e-01 7.70817220e-01 1.00914967e+00
-6.20464861e-01 -6.73114121e-01 -3.69916558e-01 2.85642505e-01
-3.20802957e-01 -5.50055616e-02 -1.75632909e-01 2.88252831e-01
-2.70651668e-01 7.77295113e-01 2.22575054e-01 1.82605654e-01
1.02361761e-01 -3.07491601e-01 9.67624485e-01 -3.96534115e-01
-1.43311739e-01 -4.60260548e-02 -2.66786098e-01 -7.88675308e-01
-8.51963997e-01 -8.33765030e-01 -1.34688401e+00 -4.75341916e-01
-1.74986631e-01 1.95769500e-02 7.55647063e-01 8.49233150e-01
2.98108101e-01 3.36230278e-01 6.43880486e-01 -1.22454858e+00
2.07365200e-01 -3.58159423e-01 -5.88516295e-01 1.55031636e-01
9.24918115e-01 -3.91615719e-01 -7.40605950e-01 6.80266172e-02] | [9.747343063354492, -2.708648920059204] |
7db5758e-746c-4705-9dc1-8457eff3188f | multi-level-cnn-for-lung-nodule | 1901.00276 | null | http://arxiv.org/abs/1901.00276v1 | http://arxiv.org/pdf/1901.00276v1.pdf | Multi-level CNN for lung nodule classification with Gaussian Process assisted hyperparameter optimization | This paper investigates lung nodule classification by using deep neural
networks (DNNs). Hyperparameter optimization in DNNs is a computationally
expensive problem, where evaluating a hyperparameter configuration may take
several hours or even days. Bayesian optimization has been recently introduced
for the automatically searching of optimal hyperparameter configurations of
DNNs. It applies probabilistic surrogate models to approximate the validation
error function of hyperparameter configurations, such as Gaussian processes,
and reduce the computational complexity to a large extent. However, most
existing surrogate models adopt stationary covariance functions to measure the
difference between hyperparameter points based on spatial distance without
considering its spatial locations. This distance-based assumption together with
the condition of constant smoothness throughout the whole hyperparameter search
space clearly violates the property that the points far away from optimal
points usually get similarly poor performance even though each two of them have
huge spatial distance between them. In this paper, a non-stationary kernel is
proposed which allows the surrogate model to adapt to functions whose
smoothness varies with the spatial location of inputs, and a multi-level
convolutional neural network (ML-CNN) is built for lung nodule classification
whose hyperparameter configuration is optimized by using the proposed
non-stationary kernel based Gaussian surrogate model. Our algorithm searches
the surrogate for optimal setting via hyperparameter importance based
evolutionary strategy, and the experiments demonstrate our algorithm
outperforms manual tuning and well-established hyperparameter optimization
methods such as Random search, Gaussian processes with stationary kernels, and
recently proposed Hyperparameter Optimization via RBF and Dynamic coordinate
search. | ['Steven Su', 'Juan Lyu', 'Sai Ho Ling', 'Miao Zhang', 'Huiqi Li'] | 2019-01-02 | null | null | null | null | ['lung-nodule-classification'] | ['medical'] | [-3.16296518e-01 -9.14613754e-02 -1.45947903e-01 -1.25632361e-01
-7.36699760e-01 -3.29138488e-01 1.49587646e-01 -3.35222296e-03
-6.95140779e-01 7.22165644e-01 -8.06081221e-02 -2.62290210e-01
-8.75341654e-01 -8.38660479e-01 -5.50744116e-01 -1.46025085e+00
2.57902294e-01 7.97251165e-01 5.32041132e-01 1.83549717e-01
1.58344328e-01 8.11472416e-01 -1.12604523e+00 -2.91728020e-01
7.72595227e-01 1.03087509e+00 2.68980354e-01 6.29304588e-01
-2.28371441e-01 -1.27444670e-01 -6.63782239e-01 -3.08446530e-02
1.33051604e-01 8.58011469e-02 -3.72374505e-01 -3.28770965e-01
-1.04345202e-01 -2.47403011e-02 -2.10667342e-01 1.17137396e+00
6.69445693e-01 3.25926244e-01 1.12215889e+00 -9.58181441e-01
-6.65521920e-01 4.70205039e-01 -3.12872261e-01 1.77806064e-01
-7.08051801e-01 3.48642707e-01 4.39648151e-01 -4.82467562e-01
-1.30753532e-01 1.04859054e+00 9.60206211e-01 3.45572740e-01
-9.05105650e-01 -4.31176603e-01 -8.15633982e-02 1.11383639e-01
-1.91432345e+00 1.15076572e-01 4.77979720e-01 -5.80546379e-01
6.51021719e-01 7.96977207e-02 6.03513777e-01 9.87662971e-01
5.56897700e-01 3.34657967e-01 4.76517081e-01 -3.21878225e-01
3.43455046e-01 2.77715713e-01 1.21856317e-01 8.12736392e-01
5.62890410e-01 7.20369965e-02 2.17615604e-01 -6.79763854e-01
1.20678496e+00 1.10572658e-01 -4.21856970e-01 -3.74040812e-01
-1.39091897e+00 9.01422441e-01 6.04159355e-01 3.69095266e-01
-4.79987651e-01 3.46035928e-01 1.84471056e-01 -5.23091137e-01
-5.40799275e-02 5.63707769e-01 -5.58895826e-01 -4.77879643e-02
-4.93628621e-01 1.18469909e-01 8.74116242e-01 8.41534674e-01
4.02083427e-01 -7.63002783e-02 -7.16962397e-01 9.19695914e-01
5.08638918e-01 5.59632301e-01 7.94339478e-01 -5.82254410e-01
2.76255876e-01 5.10559261e-01 8.71286541e-02 -9.70731258e-01
-5.95725298e-01 -7.15726078e-01 -1.21241307e+00 3.75618152e-02
5.12717426e-01 -2.93460667e-01 -1.01232421e+00 1.39957225e+00
5.36561549e-01 3.78618598e-01 -4.53802533e-02 8.36821258e-01
6.71140134e-01 4.56560940e-01 1.90130219e-01 -2.51781732e-01
1.44249749e+00 -5.53028941e-01 -5.60798466e-01 4.66752797e-01
4.16136384e-01 -6.04889631e-01 1.07621694e+00 2.07634434e-01
-7.87716389e-01 -3.54341090e-01 -9.75379109e-01 4.87571597e-01
-2.09306851e-01 3.09486568e-01 2.50191420e-01 7.53232896e-01
-8.63612950e-01 6.55789375e-01 -1.04682744e+00 -2.07544095e-03
5.05947948e-01 5.55415988e-01 2.43387312e-01 3.07463408e-01
-1.25117159e+00 6.82393730e-01 7.87412584e-01 5.03893793e-01
-6.47747755e-01 -1.00493240e+00 -3.27689171e-01 1.95409521e-01
4.02978390e-01 -1.06464183e+00 1.21755087e+00 -4.82566237e-01
-1.82124901e+00 1.48136750e-01 1.44559771e-01 -2.66005605e-01
5.57059050e-01 3.72226313e-02 -3.17653775e-01 -1.48730297e-02
-4.69945103e-01 4.30866331e-01 1.03760326e+00 -7.67533660e-01
-3.31287265e-01 -1.86443448e-01 -4.91620362e-01 1.95711419e-01
-4.14962500e-01 -3.38102490e-01 -6.91167772e-01 -4.88530695e-01
3.02167922e-01 -1.10752499e+00 -5.78117609e-01 6.75344234e-03
-7.15046287e-01 -3.68549794e-01 6.58492148e-01 -1.64765179e-01
1.50659370e+00 -1.90052891e+00 -5.64203374e-02 7.28336453e-01
2.58643270e-01 3.58123422e-01 2.56424874e-01 -2.43646339e-01
7.47329816e-02 3.88522595e-01 -3.55964929e-01 4.88126963e-01
-6.34918585e-02 1.91997811e-01 2.58217454e-01 5.37536860e-01
7.01051354e-02 8.78481567e-01 -7.10926533e-01 -7.45611429e-01
1.79196775e-01 8.41344714e-01 -3.32921743e-01 -5.59681840e-02
-2.00139880e-01 7.37760663e-02 -8.02601337e-01 4.91447896e-01
6.39061153e-01 -5.42721808e-01 -3.38136554e-01 -6.37662768e-01
7.21685588e-02 -5.28522968e-01 -1.39482236e+00 9.31450546e-01
-3.98337424e-01 3.18859547e-01 -3.24612111e-01 -7.27407455e-01
1.00870979e+00 3.93194050e-01 4.88552988e-01 2.31849059e-01
2.84223944e-01 1.55267343e-01 2.25454316e-01 -7.43985653e-01
4.59234901e-02 -4.46747728e-02 3.43491822e-01 2.31766533e-02
-3.34098339e-01 -6.41627535e-02 -2.38337591e-01 -4.87841934e-01
1.17480326e+00 -1.67332023e-01 5.19857407e-01 -3.13523889e-01
6.83859885e-01 -8.76211282e-03 4.83642071e-01 8.97847831e-01
-1.80242136e-01 7.71211445e-01 5.45682669e-01 -4.50308651e-01
-9.52532709e-01 -1.09874129e+00 -5.87430894e-01 4.00442421e-01
9.92612243e-02 2.71410704e-01 -6.98157907e-01 -7.25759268e-01
1.15633905e-01 5.98399639e-01 -7.51861155e-01 -3.50198358e-01
-6.81046247e-01 -1.14621651e+00 5.93974233e-01 5.29844522e-01
7.24439919e-01 -9.57767665e-01 -5.45857370e-01 2.93089330e-01
2.32355595e-01 -5.18009186e-01 -4.44350541e-01 3.21959257e-01
-9.72085774e-01 -1.10654330e+00 -1.06901002e+00 -5.76903284e-01
8.40036631e-01 -1.91634923e-01 7.74415076e-01 -4.25838381e-02
-5.58358729e-01 1.42991975e-01 1.62351225e-02 -3.77534479e-01
-3.39758098e-01 2.89050490e-01 -2.52044778e-02 -1.70294151e-01
3.57451886e-01 -3.66139829e-01 -8.20081234e-01 6.75341964e-01
-9.62520897e-01 -4.25600767e-01 1.08655584e+00 9.08639908e-01
1.10383832e+00 3.49251628e-01 2.73363948e-01 -7.02646792e-01
9.21893060e-01 -6.33811653e-01 -7.29612529e-01 5.14825046e-01
-6.11826062e-01 4.87625778e-01 5.18103719e-01 -9.52419162e-01
-9.87713933e-01 4.59711216e-02 2.44009301e-01 -9.13193405e-01
-2.65048027e-01 4.13517952e-01 -1.65233836e-01 -1.97199494e-01
1.03300071e+00 4.37759385e-02 8.67051035e-02 -1.45646811e-01
1.07650079e-01 4.53072488e-01 4.09753293e-01 -5.54625928e-01
8.19344580e-01 3.86836022e-01 4.06254411e-01 -6.63336158e-01
-6.70156479e-01 -4.08509552e-01 -5.83054662e-01 -5.13230339e-02
9.84059036e-01 -3.25561136e-01 -1.08005488e+00 2.75229454e-01
-1.04409802e+00 -9.88323092e-02 -2.30534568e-01 9.62889910e-01
-5.86650610e-01 2.27166176e-01 -2.15985224e-01 -5.97974598e-01
-4.10328448e-01 -1.40738773e+00 7.81913280e-01 4.98247594e-01
-1.46432435e-02 -1.16020644e+00 -3.57755087e-02 -2.02651232e-01
5.27680218e-01 4.09439988e-02 1.26710331e+00 -8.41064155e-01
-6.33594573e-01 -4.33069825e-01 -2.48590469e-01 2.91592211e-01
1.95612654e-01 2.79811740e-01 -5.40996671e-01 3.88743319e-02
2.13565126e-01 3.23993355e-01 4.92659390e-01 1.25064719e+00
1.83134401e+00 -2.68645316e-01 -6.61156714e-01 1.11111069e+00
1.48514462e+00 2.68139064e-01 5.37270308e-01 4.80859786e-01
6.36183500e-01 1.98945016e-01 4.13793564e-01 4.74468291e-01
-2.33669445e-01 3.42573673e-01 4.89553213e-01 1.00978851e-01
3.27628642e-01 2.43351310e-01 -3.73885959e-01 4.25400496e-01
-2.60253549e-02 -3.29031497e-01 -1.11748552e+00 2.52889961e-01
-1.96651840e+00 -5.39112151e-01 -9.35307667e-02 2.35299945e+00
7.83135951e-01 2.08677500e-01 -2.54445940e-01 -3.42566013e-01
1.08670962e+00 -3.67638648e-01 -8.89173090e-01 5.39530739e-02
1.32032856e-01 -1.23179927e-01 9.56157744e-01 1.59933120e-01
-1.13031256e+00 4.12825525e-01 5.60427809e+00 1.40083611e+00
-9.89388227e-01 -1.60809427e-01 6.28288805e-01 -1.78656846e-01
-9.73122045e-02 -2.87153363e-01 -1.23559999e+00 5.82263291e-01
5.98022342e-01 -9.96814445e-02 6.50794134e-02 1.08334351e+00
3.89208347e-01 -4.60562408e-02 -7.13105917e-01 1.18266976e+00
-2.69064128e-01 -1.42718422e+00 4.97647189e-02 1.16740786e-01
7.07698345e-01 -1.90277491e-02 3.14896345e-01 1.15023784e-01
3.06005269e-01 -1.28457189e+00 1.50862247e-01 1.02150881e+00
3.49072546e-01 -9.04816806e-01 9.96685386e-01 3.02637696e-01
-9.54280555e-01 -9.97496322e-02 -6.13137901e-01 1.07330954e+00
-5.09536313e-03 8.08524728e-01 -1.35348201e+00 1.83647409e-01
7.55693555e-01 -2.18938459e-02 -4.98359561e-01 1.76161909e+00
1.99258211e-03 6.69596612e-01 -7.73830533e-01 -6.90346658e-01
3.07075262e-01 -3.03781390e-01 8.84059489e-01 1.00565314e+00
7.50755906e-01 -1.90743376e-02 -7.85515308e-02 1.05018139e+00
4.24151957e-01 2.22477570e-01 -1.78097188e-01 1.40197620e-01
7.68680215e-01 1.27270794e+00 -8.45629930e-01 -4.89303134e-02
2.16955468e-01 3.20541948e-01 -3.59973460e-02 6.17220640e-01
-1.12875807e+00 -6.24401152e-01 4.20201838e-01 1.79011282e-02
4.06521827e-01 -1.14515424e-01 -4.21613693e-01 -3.97521615e-01
6.05203584e-03 -2.71126628e-01 4.72545058e-01 -5.88689685e-01
-1.35024333e+00 7.22950816e-01 3.60431820e-01 -1.43405759e+00
-1.37978315e-01 -7.21831918e-01 -7.45462656e-01 1.03627801e+00
-1.28908145e+00 -7.09115624e-01 -4.68416005e-01 5.79472303e-01
2.30829999e-01 -3.63171488e-01 5.57074845e-01 -3.81023735e-01
-6.59549832e-01 7.06888378e-01 4.11431164e-01 -1.71933725e-01
5.62421858e-01 -1.07052660e+00 -4.18359697e-01 1.99777514e-01
-5.37278175e-01 6.25428975e-01 6.35322750e-01 -7.34630704e-01
-9.72320437e-01 -1.22540426e+00 1.71982810e-01 -1.84078991e-01
6.17373705e-01 3.34890783e-01 -1.25665629e+00 5.19442856e-02
-3.50286782e-01 2.31980473e-01 5.51469386e-01 -9.88006443e-02
1.98608100e-01 -2.53350921e-02 -1.18649256e+00 8.42811346e-01
5.99298775e-01 5.95355704e-02 -1.54958859e-01 6.32779717e-01
8.73871148e-01 -3.57046098e-01 -1.17603850e+00 8.31207454e-01
3.18830192e-01 -5.39695144e-01 1.18755472e+00 -3.99661481e-01
-1.12390757e-01 -4.50735062e-01 1.78606883e-01 -1.20799589e+00
-5.25242329e-01 -7.55930245e-01 -8.60114470e-02 8.32980275e-01
5.68230212e-01 -7.80953467e-01 1.06035447e+00 6.33605778e-01
-2.13198364e-01 -1.39642739e+00 -1.02902305e+00 -7.91164279e-01
-3.69398147e-02 -3.45728904e-01 7.27273643e-01 3.91730934e-01
-7.43699074e-01 -2.69513220e-01 5.11795580e-01 6.52746260e-01
6.21990979e-01 -3.60907316e-01 2.98140913e-01 -1.23691678e+00
-4.52509671e-01 -9.64770973e-01 -3.17573339e-01 -7.02307105e-01
1.25450185e-02 -6.56488955e-01 3.45859706e-01 -1.40095890e+00
-4.54529114e-02 -8.05195630e-01 -3.77812147e-01 5.28369173e-02
-2.96560615e-01 -5.42730272e-01 -5.79344690e-01 4.47750270e-01
-2.44883373e-01 5.74917257e-01 1.37656498e+00 6.91479370e-02
-6.25859082e-01 7.37303674e-01 -3.80520850e-01 1.06024504e+00
8.48146379e-01 -6.51498079e-01 -4.81114149e-01 -8.53244513e-02
1.59105554e-01 1.29836109e-02 3.83202821e-01 -1.06859398e+00
5.38483977e-01 -4.03207839e-01 7.52177954e-01 -7.52246797e-01
2.11063460e-01 -1.02389681e+00 3.44928801e-01 3.10501724e-01
-3.17426383e-01 -1.39033943e-01 9.75040123e-02 9.52119648e-01
1.04141630e-01 -8.57914865e-01 1.08769953e+00 -1.35630295e-01
-2.15779930e-01 7.17184961e-01 -3.36038083e-01 -2.20948264e-01
1.20927572e+00 -5.50515532e-01 -7.57694691e-02 1.49944555e-02
-8.36941063e-01 1.26203001e-01 -7.72488536e-03 -3.71668078e-02
6.92754984e-01 -1.39957476e+00 -3.97454619e-01 8.89358521e-02
-9.29788351e-02 5.50157070e-01 3.83965135e-01 1.06158638e+00
-6.85329854e-01 3.68882984e-01 4.01824385e-01 -1.03913713e+00
-1.05123949e+00 3.61861110e-01 8.98764551e-01 -2.87177503e-01
-5.33121884e-01 1.01783109e+00 5.52939288e-02 -3.19494605e-01
5.00050783e-01 -5.94229758e-01 -3.52213681e-01 -3.17473710e-01
1.15744285e-01 4.52444673e-01 1.20014384e-01 2.78423969e-02
-1.03678577e-01 7.91185260e-01 1.35559335e-01 -1.41922794e-02
1.10757411e+00 1.69019088e-01 -3.72873284e-02 2.65624166e-01
1.30564010e+00 -4.04011846e-01 -1.18081033e+00 -3.63903344e-01
-1.49786589e-03 -1.83800593e-01 3.73056352e-01 -5.00200450e-01
-9.30121541e-01 4.68140155e-01 8.69254291e-01 -1.87702086e-02
8.99655402e-01 2.34541763e-02 5.83402395e-01 8.10152590e-01
-3.14481169e-01 -1.09224939e+00 1.15827322e-01 5.26190281e-01
8.37277114e-01 -1.09376383e+00 2.20105425e-02 -1.54415876e-01
-6.12051487e-01 1.60072196e+00 8.57562721e-01 -1.43583357e-01
1.27133787e+00 3.04689616e-01 -7.59189576e-02 -2.36889720e-01
-6.01891994e-01 4.05029207e-02 5.43129265e-01 5.99836171e-01
-1.08548634e-01 8.78882706e-02 -5.03749177e-02 5.25581360e-01
-9.42471772e-02 -2.86750197e-01 1.60205252e-02 5.93435943e-01
-6.25371277e-01 -5.83213091e-01 -7.80911446e-01 7.31114447e-01
-2.50329167e-01 -1.09739237e-01 5.05205572e-01 1.01886499e+00
3.40764709e-02 3.58876169e-01 2.79165208e-01 -1.61546748e-02
2.91454166e-01 -5.47814555e-02 2.41913915e-01 -4.24678624e-01
-4.14278418e-01 2.61629134e-01 -4.75951076e-01 -2.42811441e-01
1.25692010e-01 -5.47034204e-01 -1.47479737e+00 2.06244260e-01
-7.20349371e-01 7.62630850e-02 9.93395865e-01 8.23442161e-01
1.18709013e-01 6.43646240e-01 4.64181930e-01 -4.61056024e-01
-1.13490891e+00 -8.46860826e-01 -4.43193436e-01 -2.02209800e-01
1.29607663e-01 -6.95735812e-01 -6.32773042e-01 -3.52061272e-01] | [6.856664180755615, 3.913001775741577] |
efa9ba85-517c-4f23-9950-ded744142a50 | the-computation-of-cyclic-peptide-with-prolin | 1511.01388 | null | http://arxiv.org/abs/1511.01388v1 | http://arxiv.org/pdf/1511.01388v1.pdf | The Computation of Cyclic Peptide with Prolin-Prolin Bond as Fusion Inhibitor of DENV Envelope Protein through Molecular Docking and Molecular Dynamics Simulation | A disease that caused by dengue virus (DENV) has become the major health
problem of the world. Nowadays, no effective treatment is available to overcome
the disease due to the level of dengue virus pathogeneses. A novel treatment
method such as antiviral drug is highly necessary for coping with the dengue
disease. Envelope protein is one of the non-structural proteins of DENV, which
engaged in the viral fusion process. The fusion process is mediated by the
conformational change in the protein structure from dimer to trimer state. The
previous research showed the existing cavity on the dimer structure of the
envelope protein. The existing ligand could get into cavity of the envelope
protein, stabilize the dimer structure or hamper the transition of dimer
protein into trimer. In this fashion, the fusion process can be prevented. The
aim of this research is designing the cyclic peptide with prolin-prolin bond as
fusion inhibitor of DENV envelope protein through molecular docking and
molecular dynamics simulation. The screening of 3,883 cyclic peptides, each of
them connected by prolin-prolin bond, through molecular docking resulted in
five best ligands. The pharmacological and toxicity character of these five
ligands were analised in silico. The result showed that PYRRP was the best
ligand. PAWRP was also chosen as the best ligand because it showed good
affinity with protein cavity. Stability of ligand-protein complex was analyzed
by molecular dynamics simulation. The result showed that PYRRP ligand was able
to support the stability of DENV envelope protein dimer structure at 310 K and
312 K. While PAWRP ligand actively formed complex with the DENV envelope
protein at 310 K compared to 312 K. Thus the PYRRP ligand has a potential to be
developed as DENV fusion inhibitor. | [] | 2015-10-27 | null | null | null | null | ['molecular-docking'] | ['medical'] | [-1.73925459e-01 -4.79106754e-01 -9.88542661e-02 -8.00170153e-02
1.81594983e-01 -7.24587381e-01 2.27505237e-01 1.20932736e-01
-4.27087873e-01 1.24886906e+00 1.49915934e-01 -3.87745231e-01
1.87750310e-01 -8.35442841e-01 -4.83232528e-01 -1.03850627e+00
-1.78212002e-01 6.08834386e-01 6.53065816e-02 -3.86278808e-01
2.62345433e-01 9.57762420e-01 -1.02909184e+00 1.20241486e-01
1.08921790e+00 1.86231315e-01 5.85535407e-01 2.23485470e-01
-3.71001542e-01 -5.07082418e-02 -2.65309513e-01 1.03090607e-01
1.65536374e-01 -3.49289626e-01 -3.16058338e-01 -3.13571513e-01
-5.45298636e-01 -4.03699189e-01 3.18721980e-01 5.93370020e-01
5.10380208e-01 -8.88737589e-02 1.02650177e+00 -9.80437875e-01
-3.57320786e-01 -1.10819004e-01 -6.22492552e-01 3.97907794e-02
4.68269974e-01 -4.05181944e-02 2.68931448e-01 -1.12161756e+00
8.26517522e-01 1.38773942e+00 2.83693701e-01 5.60822606e-01
-1.20148063e+00 -9.22766864e-01 -3.75811845e-01 1.74205080e-01
-1.45441437e+00 -5.81744909e-02 5.45515537e-01 -6.06314480e-01
1.38056040e+00 -1.78845879e-02 7.67597735e-01 5.60068905e-01
9.38060582e-01 -6.26494735e-02 1.27122366e+00 -2.45514475e-02
2.26509273e-02 8.13646149e-03 2.60708153e-01 4.48177963e-01
5.01442552e-01 2.63381332e-01 1.92011409e-02 -6.95896387e-01
4.39718038e-01 4.36610758e-01 -4.08608139e-01 -2.93369383e-01
-3.69027644e-01 1.10275412e+00 2.11333081e-01 3.39109600e-01
-5.70247829e-01 -5.93389332e-01 2.48989657e-01 -4.16756868e-02
-4.44659323e-01 -3.67048837e-04 -6.07411027e-01 1.33209765e-01
-3.60217541e-01 5.17072618e-01 6.50568724e-01 4.40085381e-02
6.82081699e-01 -1.36984065e-01 2.76609749e-01 3.08204651e-01
8.17155957e-01 8.58869433e-01 3.09405625e-02 -2.22850099e-01
-1.06237315e-01 9.14778233e-01 3.83322388e-01 -6.87579036e-01
-5.73475122e-01 3.53176355e-01 -6.58902466e-01 3.18561673e-01
3.88104737e-01 -4.75800097e-01 -9.29480731e-01 1.85231578e+00
4.80131745e-01 8.04239586e-02 4.56086755e-01 9.47257221e-01
8.88995767e-01 1.36605453e+00 5.41593671e-01 -1.09275186e+00
1.80332530e+00 -9.50028300e-02 -6.63447142e-01 4.21523422e-01
1.86363921e-01 -1.07794917e+00 2.28904590e-01 2.03573331e-01
-6.18297875e-01 -1.40016586e-01 -1.03816819e+00 6.54376566e-01
-1.85503826e-01 -2.01233745e-01 5.31146705e-01 5.63453197e-01
-5.96907020e-01 1.23568632e-01 -9.04039323e-01 -8.80100310e-01
-2.88869917e-01 7.91714370e-01 -6.66966140e-01 -7.79360831e-02
-1.24746418e+00 1.07725573e+00 6.45393908e-01 -4.60535064e-02
-5.69685936e-01 -2.07305729e-01 -2.21139088e-01 -2.94344991e-01
-2.25045174e-01 -6.64194345e-01 5.03338695e-01 -5.08873761e-01
-1.37481213e+00 6.83956683e-01 -4.24330860e-01 -3.06275219e-01
-1.57245651e-01 1.70667954e-02 -4.20192391e-01 1.47880930e-02
-4.64693308e-02 5.23729205e-01 3.45369130e-02 -1.13199139e+00
-3.04071277e-01 -7.48680174e-01 -5.09296000e-01 2.98026323e-01
7.02768207e-01 4.46288794e-01 2.49797285e-01 -3.14965576e-01
8.44069347e-02 -1.09958005e+00 -2.65490532e-01 -7.45513499e-01
-6.80252612e-02 -6.30481184e-01 1.15945482e+00 -2.05941498e-01
8.40196013e-01 -1.71201074e+00 5.93217276e-02 5.18068850e-01
-5.14921658e-02 7.30398953e-01 2.65734822e-01 1.35273695e+00
-3.25744748e-01 8.38408694e-02 3.34794186e-02 1.00395894e+00
-5.94040573e-01 1.98211566e-01 -3.80805939e-01 5.86577415e-01
-1.08217426e-01 5.13483822e-01 -4.86726165e-01 -1.57292232e-01
7.01080039e-02 8.66945565e-01 -4.14116174e-01 3.70677441e-01
-3.31265002e-01 6.12618446e-01 -1.00891674e+00 6.32172525e-01
1.44069290e+00 2.28995040e-01 6.91677570e-01 -3.63934040e-01
-3.52904081e-01 -2.19743714e-01 -8.94300520e-01 8.67321193e-01
6.54235899e-01 -8.18308070e-02 2.32904077e-01 -6.08570874e-01
1.22315538e+00 6.51096582e-01 5.29788196e-01 -5.30597210e-01
2.05617979e-01 2.29144678e-01 2.16885120e-01 -4.46290463e-01
-1.96286410e-01 -5.44255614e-01 4.75361645e-01 9.77022275e-02
-5.46324670e-01 4.37330663e-01 1.80257559e-01 7.78164491e-02
6.05214715e-01 3.04516070e-02 4.36119199e-01 -5.30485988e-01
8.90914917e-01 4.85649332e-02 7.99396515e-01 2.42725499e-02
3.82475345e-03 -3.08554024e-01 3.36560696e-01 -5.96670985e-01
-8.69725049e-01 -1.15992010e+00 -6.60624921e-01 5.48507631e-01
2.84246981e-01 -9.66210216e-02 -6.73187435e-01 8.33867192e-02
-6.96878135e-02 2.29165912e-01 -6.16267882e-02 5.49421273e-02
-5.72094440e-01 -1.07589912e+00 3.65288526e-01 -4.27733362e-01
4.96413201e-01 -1.16070545e+00 -5.53796530e-01 5.12714922e-01
1.36574611e-01 -4.48676944e-01 4.62496802e-02 2.85672545e-01
-1.14542425e+00 -1.22965407e+00 -2.85903931e-01 -8.78748894e-01
2.91061789e-01 8.84511024e-02 1.85723826e-01 2.17672154e-01
2.96402890e-02 -5.32493174e-01 -2.38014132e-01 -4.23874885e-01
-3.13458681e-01 -4.77226168e-01 5.09973764e-01 -4.68291432e-01
1.08469737e+00 -7.76539385e-01 -8.80591273e-01 2.47193575e-01
-8.05259764e-01 -1.14675798e-01 5.35771608e-01 5.43448091e-01
4.64936048e-01 -1.23433150e-01 5.59246659e-01 -4.76732820e-01
7.80889452e-01 -5.93237817e-01 -7.59622216e-01 3.81247103e-02
-6.17385805e-01 2.94512808e-01 4.87989575e-01 -1.42208919e-01
-9.33673859e-01 1.00440927e-01 -2.69571811e-01 3.30533832e-01
-1.06319144e-01 5.30741811e-01 -7.39799440e-01 1.44509580e-02
3.41598451e-01 3.44097108e-01 4.47480530e-01 -5.60232878e-01
-3.19706589e-01 5.56452572e-01 1.57142639e-01 -5.51305115e-01
5.22324502e-01 1.85844198e-01 3.56714070e-01 -1.03880358e+00
2.49382854e-01 -1.70449063e-01 1.22304045e-01 1.31175831e-01
1.36529326e+00 -7.85939097e-01 -1.25086010e+00 3.94881397e-01
-1.48826683e+00 5.21475852e-01 1.09756231e+00 8.82347047e-01
6.26804009e-02 5.56004524e-01 -6.67531550e-01 -6.52435124e-01
-7.00634837e-01 -1.44147646e+00 1.21982433e-01 3.67247134e-01
-5.99953756e-02 -4.99822110e-01 8.07249904e-01 1.75098911e-01
2.06670508e-01 5.96308351e-01 1.40480888e+00 -6.83364034e-01
-7.83947170e-01 1.54106289e-01 -1.31363228e-01 -2.76028305e-01
1.69855207e-01 4.08543408e-01 -4.02492642e-01 -4.14542139e-01
-3.65972370e-02 -6.15412332e-02 5.42852044e-01 6.62913263e-01
-1.86875045e-01 -2.90271312e-01 -7.76448131e-01 6.67117834e-01
1.79131627e+00 1.37923872e+00 1.02624810e+00 3.17047507e-01
2.45374292e-01 3.94344211e-01 8.40378940e-01 3.91308039e-01
-2.15902515e-02 8.32094908e-01 3.56774211e-01 1.28296101e-02
7.20300555e-01 -1.87834173e-01 4.84457374e-01 2.30543762e-01
-5.70374310e-01 -3.29610556e-01 -8.29699755e-01 -1.36769712e-01
-1.47673905e+00 -1.02685475e+00 -5.70139170e-01 2.17503500e+00
8.53151202e-01 -1.98865920e-01 1.47024944e-01 -3.95166487e-01
4.13463712e-01 -4.64662939e-01 -3.59261751e-01 -8.68342042e-01
-2.04911828e-01 3.85526329e-01 5.18876970e-01 8.06538939e-01
-5.50951481e-01 8.97455037e-01 5.76216507e+00 4.90404069e-01
-1.29993808e+00 -4.32755053e-01 -1.08062029e-01 2.69601524e-01
-2.36953422e-01 6.41638935e-01 -8.95090461e-01 7.05677867e-01
7.56016314e-01 -1.24348767e-01 1.72162846e-01 3.90606403e-01
6.93205476e-01 -4.57204252e-01 -6.32472873e-01 9.88354623e-01
-6.31444037e-01 -1.22549391e+00 3.47991049e-01 5.83314002e-01
3.59790504e-01 -2.01190889e-01 -5.02451241e-01 -1.78130418e-01
-6.76477104e-02 -1.22315955e+00 -3.64308029e-01 5.28036416e-01
3.25102776e-01 -1.25773668e+00 9.90089357e-01 3.61087084e-01
-1.30118465e+00 2.79741019e-01 -5.17056525e-01 -1.06086038e-01
4.60706413e-01 2.49488682e-01 -9.67432439e-01 1.26644418e-01
4.25763071e-01 -3.04760244e-02 8.27896073e-02 5.93343318e-01
1.20854616e-01 2.03168720e-01 -4.96867180e-01 1.90620851e-02
2.21090853e-01 -1.13077760e+00 5.58921993e-01 9.16205764e-01
4.05564681e-02 8.96879792e-01 4.47023571e-01 5.08256197e-01
3.89841676e-01 6.17165267e-01 -9.42217529e-01 -1.46727040e-02
6.34830117e-01 7.60648489e-01 -5.82473636e-01 -1.05812922e-01
-2.96108484e-01 5.87674558e-01 -9.74669307e-03 4.86890525e-01
-6.77587271e-01 -2.31576473e-01 1.05175936e+00 3.58632475e-01
2.14710936e-01 5.00491587e-04 5.21216094e-01 -4.78830278e-01
-4.11022484e-01 -9.00920510e-01 1.97755471e-01 -2.80438930e-01
-7.21439898e-01 4.36817944e-01 1.53952956e-01 -5.10591865e-01
6.11775778e-02 -4.44061160e-01 -6.47768915e-01 1.30459285e+00
-9.25791144e-01 -1.05765843e+00 8.47820863e-02 5.16562223e-01
8.92024413e-02 -2.66593128e-01 1.19857168e+00 4.82104309e-02
-6.59159958e-01 3.91185284e-04 4.27232563e-01 -4.37669605e-01
8.45864117e-01 -5.09790480e-01 -5.32398939e-01 3.89845490e-01
-9.41900373e-01 1.28649497e+00 1.10979486e+00 -1.30831027e+00
-1.34705901e+00 -4.67529923e-01 9.90173280e-01 -1.13871351e-01
1.57911986e-01 -6.54578209e-02 -8.45241129e-01 4.61678833e-01
5.40348709e-01 -8.91199172e-01 1.15252900e+00 -6.17216587e-01
3.36866304e-02 1.95763424e-01 -1.38694715e+00 4.64907110e-01
1.54513374e-01 -8.51956829e-02 -7.41823435e-01 1.70248985e-01
3.93761873e-01 -2.06481852e-02 -8.88184667e-01 6.66583896e-01
9.01875198e-01 -1.05839479e+00 9.08113182e-01 -8.50206673e-01
-1.86854407e-01 -9.10681605e-01 -2.06380129e-01 -6.63147032e-01
-3.66163582e-01 -5.00929832e-01 2.19132066e-01 1.02919936e+00
2.20025778e-01 -5.86432338e-01 6.53439403e-01 5.67590415e-01
3.58220875e-01 -8.13927889e-01 -8.02632749e-01 -3.34194064e-01
3.90577577e-02 4.71860945e-01 3.67384493e-01 6.12402856e-01
-9.50294826e-03 7.96311915e-01 -5.01823306e-01 1.80927347e-02
5.13957381e-01 2.21082196e-02 7.10742831e-01 -1.32139361e+00
2.41271362e-01 8.35199058e-02 -3.33481550e-01 -5.88990927e-01
-1.29714519e-01 -6.91929400e-01 -9.03912604e-01 -1.73655748e+00
4.64065582e-01 -1.34303018e-01 1.81043178e-01 9.78868529e-02
2.67951429e-01 -3.03217024e-01 -2.65860315e-02 3.38392913e-01
2.72824019e-01 2.59643823e-01 8.68678451e-01 2.49233902e-01
-9.28375423e-01 -1.03283435e-01 -3.57661963e-01 4.26995724e-01
9.31491137e-01 -5.61662078e-01 -6.65529132e-01 4.32484627e-01
9.20520723e-02 3.76808703e-01 -2.95335770e-01 -2.25921020e-01
-1.45460248e-01 -6.50879800e-01 4.75652695e-01 -1.13076675e+00
3.92551631e-01 -1.24447131e+00 1.09409451e+00 1.04199064e+00
8.53178859e-01 4.36136425e-01 2.44547322e-01 1.72916815e-01
9.81962755e-02 -9.81977135e-02 9.61546957e-01 -5.65779917e-02
-4.85581666e-01 1.13930687e-01 -1.05436051e+00 -3.02643210e-01
1.41464436e+00 -8.02181721e-01 -2.12574735e-01 1.76880062e-01
-6.76429808e-01 2.84102112e-01 8.05439532e-01 9.04319435e-02
6.85617805e-01 -1.04981458e+00 -5.82620323e-01 2.98132420e-01
-4.74943444e-02 -5.43281972e-01 4.02697444e-01 8.42790723e-01
-1.25822079e+00 8.05893362e-01 -8.11277688e-01 -5.54747939e-01
-1.67722487e+00 9.85815763e-01 4.39282507e-01 -1.63269535e-01
-1.89608991e-01 3.25966030e-01 4.65838909e-01 -9.93850827e-02
-3.42888422e-02 1.36268049e-01 -8.25973451e-01 -1.35039464e-01
5.64429283e-01 2.82084167e-01 -2.56710857e-01 -9.24641013e-01
-9.95405793e-01 1.07995629e+00 -3.41641843e-01 3.29399645e-01
1.25394130e+00 1.64374318e-02 -7.31703222e-01 -3.88586581e-01
1.02048731e+00 2.90278584e-01 -6.46376729e-01 2.51826823e-01
-9.78117250e-03 -2.49670222e-01 -4.79329288e-01 -8.08649302e-01
-3.81289661e-01 1.94505855e-01 1.11197221e+00 -5.84105372e-01
8.43141913e-01 -1.01417810e-01 8.35664630e-01 3.70043933e-01
4.97752726e-01 -5.67882538e-01 -5.18748701e-01 4.77618456e-01
7.21917033e-01 -1.01327860e+00 -5.78637496e-02 -2.41507649e-01
-5.81773818e-01 9.54769790e-01 6.00230038e-01 -1.54350802e-01
7.40550995e-01 1.87667057e-01 1.53007075e-01 -4.29776967e-01
-8.99643481e-01 3.30277771e-01 -1.62882656e-01 8.00269127e-01
8.57087433e-01 1.11066520e-01 -1.49711013e+00 4.34974879e-01
1.11726947e-01 2.30308294e-01 3.00023824e-01 9.50613499e-01
-8.79810989e-01 -1.81046903e+00 -7.88067937e-01 6.45925924e-02
-5.65515757e-01 1.04413228e-02 -6.67715251e-01 8.63469124e-01
4.08800870e-01 1.00305974e+00 -2.18718335e-01 -2.97941640e-02
1.75018072e-01 2.67045170e-01 4.33017969e-01 -6.30796552e-02
-3.62875342e-01 5.65292120e-01 -1.54540420e-01 -1.20114461e-01
-4.24431562e-01 -1.96824044e-01 -1.96679354e+00 -7.92484939e-01
-7.71414489e-02 9.87816930e-01 4.42193449e-01 7.96113729e-01
2.74950415e-01 -4.97032464e-01 7.20356524e-01 -2.68858433e-01
3.39310290e-03 -7.85712719e-01 -8.06513786e-01 2.15751588e-01
2.29325071e-01 -8.24265480e-01 -1.88150126e-02 -2.46074229e-01] | [4.673398971557617, 5.103201866149902] |
b3df66d4-e573-4096-964b-adc339a394cc | reference-twice-a-simple-and-unified-baseline | 2301.01156 | null | https://arxiv.org/abs/2301.01156v2 | https://arxiv.org/pdf/2301.01156v2.pdf | Reference Twice: A Simple and Unified Baseline for Few-Shot Instance Segmentation | Few Shot Instance Segmentation (FSIS) requires models to detect and segment novel classes with limited several support examples. In this work, we explore a simple yet unified solution for FSIS as well as its incremental variants, and introduce a new framework named Reference Twice (RefT) to fully explore the relationship between support/query features based on a Transformer-like framework. Our key insights are two folds: Firstly, with the aid of support masks, we can generate dynamic class centers more appropriately to re-weight query features. Secondly, we find that support object queries have already encoded key factors after base training. In this way, the query features can be enhanced twice from two aspects, i.e., feature-level and instance-level. In particular, we firstly design a mask-based dynamic weighting module to enhance support features and then propose to link object queries for better calibration via cross-attention. After the above steps, the novel classes can be improved significantly over our strong baseline. Additionally, our new framework can be easily extended to incremental FSIS with minor modification. When benchmarking results on the COCO dataset for FSIS, gFSIS, and iFSIS settings, our method achieves a competitive performance compared to existing approaches across different shots, e.g., we boost nAP by noticeable +8.2/+9.4 over the current state-of-the-art FSIS method for 10/30-shot. We further demonstrate the superiority of our approach on Few Shot Object Detection. Code and model will be available. | ['Xiangtai Li', 'Yong liu', 'Chengjie Wang', 'Yabiao Wang', 'Xintian Shen', 'Chao Xu', 'Zhucun Xue', 'Jiangning Zhang', 'Yue Han'] | 2023-01-03 | null | null | null | null | ['few-shot-object-detection'] | ['computer-vision'] | [ 4.66044486e-01 8.34557600e-03 -4.27597642e-01 -4.37320381e-01
-9.80392039e-01 -4.39804465e-01 5.54029942e-01 8.28505158e-02
-4.14221972e-01 3.72886807e-01 4.37742025e-02 3.18694144e-01
-2.41313249e-01 -7.65779912e-01 -7.13139296e-01 -4.59437728e-01
2.54373074e-01 4.52905804e-01 1.10468447e+00 -2.94587642e-01
3.37886512e-01 2.03450814e-01 -1.83936834e+00 2.69231588e-01
9.52359140e-01 1.11633909e+00 1.33425325e-01 4.14793849e-01
-2.87532479e-01 5.49360335e-01 -6.01600289e-01 -5.63106358e-01
3.67320776e-01 -2.98810750e-01 -6.40852630e-01 2.30276898e-01
5.36047637e-01 -4.25915927e-01 -4.26869869e-01 8.78280044e-01
6.25578165e-01 3.54554445e-01 5.53516507e-01 -1.22345912e+00
-4.55169618e-01 7.32334912e-01 -7.77537525e-01 3.43379676e-01
1.16226608e-02 3.95055532e-01 1.21237969e+00 -1.20084405e+00
6.33259535e-01 1.21516454e+00 4.83835459e-01 4.67821658e-01
-1.04894352e+00 -5.94799161e-01 4.26369786e-01 6.34999752e-01
-1.30823421e+00 -5.55923820e-01 5.09809732e-01 -2.74353445e-01
1.02383482e+00 3.37045729e-01 6.17829382e-01 8.34613681e-01
-4.39739227e-01 1.36444044e+00 6.08167231e-01 -3.92100960e-01
2.35847309e-01 2.55401134e-01 5.44084430e-01 6.47185028e-01
6.50081262e-02 -1.06315784e-01 -5.44558585e-01 7.18150213e-02
3.12679738e-01 3.93730372e-01 -1.90141007e-01 -3.84143353e-01
-1.09915614e+00 9.39430237e-01 5.50848424e-01 2.34093830e-01
-7.22997338e-02 1.53508827e-01 4.85402435e-01 -9.79155488e-03
4.80777502e-01 4.43392724e-01 -3.87913883e-01 -1.64955169e-01
-1.18286800e+00 3.29308391e-01 5.33997774e-01 1.19338548e+00
8.57213914e-01 -1.60637438e-01 -9.67988610e-01 1.15855014e+00
3.65213081e-02 2.72837430e-01 3.39098811e-01 -9.38519120e-01
4.26087111e-01 6.26167715e-01 -6.84783608e-02 -5.30390739e-01
-1.28874674e-01 -7.17486382e-01 -3.11529934e-01 -9.74960029e-02
9.00772065e-02 9.22445655e-02 -1.26653647e+00 1.60334754e+00
5.20066082e-01 6.57489598e-01 -1.08816147e-01 6.67800844e-01
1.01958895e+00 5.24618566e-01 -8.49838853e-02 -1.18120551e-01
1.49776554e+00 -1.34366572e+00 -5.20883441e-01 -3.35570872e-01
3.87468427e-01 -5.27000129e-01 1.31206417e+00 1.20417729e-01
-1.10051858e+00 -5.57590365e-01 -1.16478944e+00 5.32737635e-02
-3.32749426e-01 7.75644779e-02 6.20841980e-01 5.73742330e-01
-5.87413907e-01 7.10747361e-01 -8.44507933e-01 -3.93746138e-01
7.62512565e-01 2.30476111e-01 2.02397659e-01 -1.43796042e-01
-1.18769348e+00 5.40568709e-01 5.26272774e-01 -3.37233812e-01
-9.15093541e-01 -1.02868402e+00 -9.28146660e-01 2.75001734e-01
1.07857382e+00 -5.99593222e-01 1.43676674e+00 -5.12108564e-01
-1.25299442e+00 5.93896985e-01 -3.12833190e-01 -5.58791459e-01
4.48246479e-01 -3.05181175e-01 -1.81166336e-01 3.30935955e-01
3.97022247e-01 8.61512601e-01 7.68607378e-01 -1.07297218e+00
-9.57667053e-01 -2.29492068e-01 3.66375595e-01 1.05057351e-01
-5.42887688e-01 9.89535376e-02 -1.10904455e+00 -6.58436120e-01
-9.44814906e-02 -7.41257608e-01 -1.32552907e-01 9.46622863e-02
-3.58957499e-01 -3.29528064e-01 5.94264030e-01 -2.09967718e-01
1.53674376e+00 -2.27687502e+00 -1.29107758e-02 -6.61877990e-02
2.69892037e-01 5.62135518e-01 -2.79816926e-01 1.92778170e-01
1.51621774e-01 2.98124403e-02 -4.85366225e-01 -4.77438807e-01
1.93628356e-01 1.77999407e-01 -2.16084644e-01 8.61613676e-02
5.00944853e-01 1.18405139e+00 -8.73715520e-01 -5.83622217e-01
1.82984382e-01 1.68399930e-01 -7.36061633e-01 -4.92943972e-02
-2.59108394e-01 -1.17610440e-01 -2.95094073e-01 8.29507709e-01
6.54485822e-01 -3.14563751e-01 -3.65005523e-01 -2.22155944e-01
6.90850094e-02 1.25998393e-01 -1.34325469e+00 1.69332528e+00
-2.82323211e-01 3.37511450e-01 -2.83882439e-01 -1.03582263e+00
6.87215865e-01 -4.23656069e-02 3.65403920e-01 -5.95514655e-01
2.08646827e-03 8.63364711e-02 -3.15117128e-02 -4.57848758e-01
4.83210683e-01 -1.54679373e-01 1.78972632e-02 2.70306110e-01
2.65318602e-01 9.47980061e-02 5.79543889e-01 5.20828068e-01
1.11078501e+00 6.21642508e-02 1.95312083e-01 1.65069342e-01
2.12055132e-01 -1.29358903e-01 7.19871402e-01 9.98830140e-01
-3.56506437e-01 8.09530735e-01 4.10781413e-01 6.35009408e-02
-5.90073764e-01 -9.81535316e-01 -3.01484436e-01 1.50221682e+00
3.72211456e-01 -5.22007108e-01 -6.91482663e-01 -9.81973112e-01
1.24842815e-01 8.61436427e-01 -7.13711202e-01 -3.22093904e-01
-4.03514862e-01 -7.06856191e-01 3.93593341e-01 9.22605157e-01
5.05007505e-01 -1.05723441e+00 -7.67488122e-01 3.34478945e-01
5.37723079e-02 -1.03574502e+00 -6.70817196e-01 2.29806527e-01
-6.77564859e-01 -9.79825914e-01 -8.73167157e-01 -4.77753311e-01
1.73025146e-01 7.50475466e-01 1.01545644e+00 2.21052781e-01
-5.16128540e-01 2.53596932e-01 -6.95957184e-01 -5.59444726e-01
1.39781997e-01 3.11868876e-01 -2.62120783e-01 2.87334882e-02
3.68888021e-01 -4.84095782e-01 -8.02581310e-01 4.01170224e-01
-1.03773832e+00 7.33751729e-02 4.91728902e-01 8.56999397e-01
5.69198906e-01 -2.75903136e-01 6.98150516e-01 -1.10942841e+00
1.80007443e-01 -5.42129934e-01 -2.58315802e-01 5.50646424e-01
-5.96890450e-01 -2.20635347e-02 1.38916507e-01 -4.68169153e-01
-1.19385958e+00 -6.72731996e-02 -1.65268257e-01 -5.27187228e-01
-5.83630204e-02 3.59246135e-01 -2.41030544e-01 1.27302215e-01
7.51239955e-01 6.97201490e-02 -3.61067593e-01 -5.18515527e-01
5.41687429e-01 6.19522512e-01 3.04595411e-01 -4.50490952e-01
8.50888491e-01 4.61785644e-01 -4.89204884e-01 -5.97071111e-01
-1.23125970e+00 -9.25193787e-01 -4.68146116e-01 3.25786769e-02
5.43916285e-01 -8.28098178e-01 -2.83043295e-01 3.91406685e-01
-8.52374554e-01 -2.32091516e-01 -7.49109447e-01 2.84769088e-01
-3.65958244e-01 4.29728776e-01 -5.86559772e-01 -7.08036482e-01
-4.07007426e-01 -1.07247627e+00 1.31908226e+00 4.77208704e-01
1.67606305e-02 -5.36068082e-01 -2.39959010e-03 3.15197647e-01
3.82419497e-01 2.95631513e-02 6.35681450e-01 -9.40296829e-01
-6.14668250e-01 -1.29539251e-01 -4.90731448e-01 1.32466897e-01
5.22771850e-02 9.11503732e-02 -1.04324317e+00 -2.63866812e-01
-1.80654064e-01 -4.14749146e-01 1.49331570e+00 2.62890846e-01
1.06605256e+00 -5.68699930e-03 -4.73165423e-01 8.12911034e-01
1.32777119e+00 1.59789473e-01 5.86524665e-01 2.95611858e-01
6.05608940e-01 4.92811948e-01 1.12568271e+00 6.74066484e-01
2.75478959e-01 8.23850513e-01 3.12395930e-01 -2.75139753e-02
-3.06643367e-01 -1.16996132e-01 1.31867409e-01 4.23616648e-01
5.10850139e-02 -2.03735474e-03 -6.28664136e-01 6.63733780e-01
-1.98514438e+00 -1.05087614e+00 1.70724377e-01 2.05698490e+00
7.78206825e-01 4.28490669e-01 3.69117260e-01 2.24974781e-01
8.21791053e-01 1.89659134e-01 -9.60746944e-01 5.78790084e-02
1.96686521e-01 4.99072433e-01 2.72542417e-01 1.40910640e-01
-1.12465608e+00 1.19069004e+00 5.17080021e+00 1.26691568e+00
-7.82414556e-01 2.69375086e-01 4.55797344e-01 -5.65396726e-01
-3.10396194e-01 7.16578290e-02 -1.19593835e+00 3.57165366e-01
5.62261701e-01 -4.06638116e-01 3.39156479e-01 1.02657700e+00
-2.42122993e-01 2.18180269e-02 -1.00213552e+00 8.45512986e-01
2.51062781e-01 -1.42198753e+00 -1.50339350e-01 -2.42580920e-01
6.99322283e-01 1.24758489e-01 -2.76497323e-02 9.30382550e-01
1.80222373e-02 -4.41816002e-01 8.68742526e-01 3.19258064e-01
8.46982300e-01 -7.35215068e-01 5.68840146e-01 3.66686136e-01
-1.36802268e+00 -3.17508221e-01 -5.19621015e-01 2.71128088e-01
1.69720158e-01 4.99325871e-01 -5.98958254e-01 6.70343280e-01
7.25759447e-01 7.56992459e-01 -8.55100393e-01 1.48057902e+00
-1.81473553e-01 6.89009249e-01 -3.71731281e-01 2.17531286e-02
1.97099879e-01 1.92743465e-01 4.12884146e-01 1.27666402e+00
3.08568269e-01 1.38842270e-01 2.12107345e-01 8.56478393e-01
-2.36803815e-02 9.20682400e-02 -1.12184376e-01 7.86318723e-03
5.74195743e-01 1.39901292e+00 -9.17654991e-01 -7.70430028e-01
-4.59531307e-01 9.07418072e-01 3.52554679e-01 2.39825442e-01
-1.26755631e+00 -7.80790210e-01 4.91501153e-01 1.98431835e-01
7.87519395e-01 3.93201828e-01 -1.71153359e-02 -1.39436674e+00
-4.80579399e-03 -7.94103086e-01 5.59258521e-01 -3.74855638e-01
-1.27541757e+00 4.61214840e-01 3.63043845e-01 -1.14129615e+00
-7.17210919e-02 -3.76692027e-01 -8.86347055e-01 3.76101106e-01
-1.53784585e+00 -1.15987325e+00 -4.75746363e-01 4.98942643e-01
1.00287545e+00 4.08426970e-02 4.65370089e-01 3.55302751e-01
-7.62857318e-01 9.39248383e-01 -3.94156799e-02 -1.07685603e-01
7.52351105e-01 -1.17897022e+00 6.62833571e-01 8.82863939e-01
4.94291782e-01 4.68694299e-01 4.84856784e-01 -5.44488549e-01
-9.31087196e-01 -1.27032650e+00 3.32795203e-01 -3.74758750e-01
6.02497756e-01 -3.49654406e-01 -1.14270616e+00 5.06435692e-01
-5.12509085e-02 1.68722093e-01 6.33020163e-01 1.61016881e-01
-4.23571557e-01 -2.18648985e-01 -1.05395555e+00 5.02309263e-01
1.51988840e+00 -2.19509840e-01 -6.29266202e-01 3.24417591e-01
1.12003624e+00 -3.51413935e-01 -5.47102451e-01 7.35737026e-01
4.46232826e-01 -1.07890832e+00 9.99139667e-01 -5.34820080e-01
2.41720200e-01 -2.92928964e-01 -1.93934947e-01 -9.91455495e-01
-4.72980350e-01 -4.92288381e-01 -4.49398339e-01 1.61160243e+00
4.63340372e-01 -4.83829796e-01 7.36275375e-01 3.70620310e-01
-3.43484491e-01 -1.18341804e+00 -7.99140453e-01 -8.88284445e-01
-2.61467308e-01 -6.54674768e-01 7.79511392e-01 5.30236185e-01
-1.27494574e-01 5.02630651e-01 -2.42594406e-01 -9.78923813e-02
4.01239485e-01 2.90856272e-01 7.79320359e-01 -1.06244361e+00
-7.29117393e-01 -4.56154644e-01 -2.60614008e-01 -1.14618862e+00
-1.09766237e-01 -8.58863413e-01 1.50416613e-01 -1.62665117e+00
6.38824880e-01 -4.00991499e-01 -5.76153457e-01 6.97156310e-01
-7.12041616e-01 3.21024686e-01 5.76854289e-01 1.75070673e-01
-1.03617883e+00 7.57863581e-01 1.02444172e+00 -2.28338316e-01
-3.33383054e-01 2.61801720e-01 -8.22634161e-01 6.48297131e-01
6.67998314e-01 -5.02499938e-01 -5.17181098e-01 -1.09091297e-01
-8.10551494e-02 -3.19525599e-01 2.71808177e-01 -1.16269433e+00
1.04045525e-01 -6.44948632e-02 2.10924253e-01 -6.87625468e-01
3.13215762e-01 -3.48955631e-01 -2.78418809e-01 2.35152707e-01
-2.04090774e-01 -5.40346086e-01 1.61457226e-01 6.93209112e-01
-1.35740444e-01 -6.73419774e-01 8.03138316e-01 -8.28747526e-02
-8.95946980e-01 5.93912840e-01 1.84835330e-01 3.51721048e-01
1.19766760e+00 -3.93309891e-01 -4.04569536e-01 -1.33556873e-01
-6.56125605e-01 4.74402308e-01 3.88370126e-01 5.15106022e-01
6.56075597e-01 -1.24892735e+00 -5.58792949e-01 8.67865905e-02
6.46314979e-01 5.07479310e-02 4.62760627e-01 9.18156326e-01
-2.52265781e-02 1.68146640e-01 2.03102276e-01 -7.05575466e-01
-1.25422966e+00 7.17722178e-01 -2.27129497e-02 -3.45418870e-01
-6.11923218e-01 1.11320102e+00 4.04241115e-01 -1.32573724e-01
4.70156848e-01 -2.20780611e-01 -2.05187261e-01 3.44149768e-01
7.75763512e-01 4.39575493e-01 -1.30106993e-02 -3.05096149e-01
-3.84434342e-01 5.66313207e-01 -4.35369730e-01 -3.54697965e-02
1.44547677e+00 7.70731345e-02 3.81467730e-01 3.47080112e-01
1.00431585e+00 -1.27613381e-01 -1.51795495e+00 -5.36111653e-01
-2.20356286e-02 -6.06775165e-01 -2.02270746e-01 -6.88798666e-01
-1.10100102e+00 9.79248643e-01 6.53491378e-01 8.66572782e-02
1.20231116e+00 3.82939637e-01 8.39021564e-01 2.93620825e-01
1.75982863e-01 -1.11379373e+00 3.61199886e-01 4.42547828e-01
6.75901830e-01 -1.38061166e+00 -1.84769511e-01 -5.57662189e-01
-8.12509239e-01 7.75901377e-01 7.05393493e-01 4.42643873e-02
3.97080868e-01 1.32777244e-01 -3.79011393e-01 -3.58671963e-01
-7.64436066e-01 -7.45335281e-01 4.98069316e-01 3.85190427e-01
1.87973678e-01 -1.98929712e-01 -1.29233941e-01 7.55622268e-01
1.33956686e-01 9.02599022e-02 1.34415284e-01 9.47398841e-01
-8.36045265e-01 -9.10983384e-01 -5.23030348e-02 7.37121344e-01
-2.70693362e-01 -2.26630166e-01 -8.34061578e-02 7.54685462e-01
3.55169207e-01 8.55032384e-01 -2.41730735e-02 -3.75034720e-01
6.31623387e-01 1.14285454e-01 4.51614708e-01 -1.15359640e+00
-4.72892672e-01 1.93665847e-01 -8.18558261e-02 -6.22263253e-01
-3.64615202e-01 -5.58401048e-01 -1.30454254e+00 9.69245583e-02
-8.33565712e-01 -3.17629352e-02 2.65910119e-01 8.92899513e-01
4.70593393e-01 7.21559107e-01 5.21219790e-01 -9.06217337e-01
-7.44129956e-01 -1.01200926e+00 -4.15936828e-01 3.71024460e-01
-4.41998057e-02 -9.40367281e-01 -4.13810343e-01 -2.77376235e-01] | [9.606552124023438, 1.8139228820800781] |
e908ae5a-a454-4764-999e-2e1d62a0a9e0 | simalign-high-quality-word-alignments-without | 2004.08728 | null | https://arxiv.org/abs/2004.08728v4 | https://arxiv.org/pdf/2004.08728v4.pdf | SimAlign: High Quality Word Alignments without Parallel Training Data using Static and Contextualized Embeddings | Word alignments are useful for tasks like statistical and neural machine translation (NMT) and cross-lingual annotation projection. Statistical word aligners perform well, as do methods that extract alignments jointly with translations in NMT. However, most approaches require parallel training data, and quality decreases as less training data is available. We propose word alignment methods that require no parallel data. The key idea is to leverage multilingual word embeddings, both static and contextualized, for word alignment. Our multilingual embeddings are created from monolingual data only without relying on any parallel data or dictionaries. We find that alignments created from embeddings are superior for four and comparable for two language pairs compared to those produced by traditional statistical aligners, even with abundant parallel data; e.g., contextualized embeddings achieve a word alignment F1 for English-German that is 5 percentage points higher than eflomal, a high-quality statistical aligner, trained on 100k parallel sentences. | ['François Yvon', 'Hinrich Schütze', 'Masoud Jalili Sabet', 'Philipp Dufter'] | 2020-04-18 | null | https://aclanthology.org/2020.findings-emnlp.147 | https://aclanthology.org/2020.findings-emnlp.147.pdf | findings-of-the-association-for-computational | ['multilingual-word-embeddings'] | ['methodology'] | [ 8.88048634e-02 9.75788981e-02 -6.32598460e-01 -2.90023446e-01
-1.38770413e+00 -8.94234359e-01 6.62362456e-01 2.73687065e-01
-1.06655908e+00 7.66331494e-01 7.30849683e-01 -5.65745652e-01
5.36840439e-01 -4.26347196e-01 -7.08245039e-01 -3.22047025e-01
4.49483186e-01 8.26843262e-01 -4.85823691e-01 -4.58641648e-01
-4.50263694e-02 -9.04086903e-02 -6.86656356e-01 -2.03043446e-01
1.11020052e+00 9.81478989e-02 2.83397317e-01 4.41369712e-01
-4.15700346e-01 -3.40873927e-01 -2.29565620e-01 -7.27321863e-01
4.72053558e-01 -5.21146357e-01 -7.83551157e-01 -7.30853975e-01
9.01371121e-01 1.19479872e-01 -1.34974495e-01 1.09077382e+00
7.08890915e-01 -3.05945963e-01 5.22136450e-01 -7.84898639e-01
-1.20919013e+00 1.01750314e+00 -3.94425601e-01 3.01018149e-01
3.24091643e-01 1.47285461e-01 1.71760499e+00 -1.37208438e+00
7.41335452e-01 9.04214263e-01 1.01949012e+00 5.37680387e-01
-1.64133453e+00 -6.70389354e-01 -1.00838937e-01 -7.01445714e-02
-1.32398212e+00 -6.94944084e-01 2.51224458e-01 -3.49348813e-01
1.55293131e+00 1.42532051e-01 6.19355440e-01 1.31562293e+00
5.70975780e-01 4.73484725e-01 9.58593369e-01 -9.29579258e-01
-3.07164073e-01 2.37342417e-02 1.16131358e-01 4.45881546e-01
4.21928287e-01 -5.20221703e-02 -7.25839019e-01 -1.47011623e-01
3.41319114e-01 -3.92408013e-01 3.74471694e-02 2.03581192e-02
-1.85248554e+00 9.58635449e-01 -7.49600753e-02 5.62861621e-01
-3.99245679e-01 8.93928558e-02 4.45832908e-01 5.87231219e-01
7.55217075e-01 1.09521616e+00 -6.68909788e-01 -3.76788825e-01
-1.00492096e+00 3.46158706e-02 5.11674285e-01 1.11886728e+00
8.80241811e-01 1.85364008e-01 -3.72658558e-02 1.01269031e+00
5.20819984e-03 1.09562790e+00 1.03563237e+00 -5.38573265e-01
7.79334068e-01 1.34215206e-01 -1.71553627e-01 -7.04064608e-01
-1.81153387e-01 -1.99989244e-01 -3.88742328e-01 -3.66622418e-01
3.72333378e-01 -2.96562701e-01 -7.85493135e-01 2.04510069e+00
7.60616213e-02 -1.08468547e-01 3.71378332e-01 5.84752142e-01
4.08180088e-01 7.06006467e-01 3.84695679e-02 -5.89971282e-02
1.51315069e+00 -1.22455812e+00 -8.73629689e-01 -7.97728717e-01
1.14790952e+00 -1.50608134e+00 1.55311656e+00 -1.92608044e-01
-1.09027302e+00 -5.41085005e-01 -1.13683736e+00 -3.76983494e-01
-3.41951668e-01 7.07970886e-03 4.33329433e-01 6.07109845e-01
-1.13105428e+00 4.54123944e-01 -9.11439896e-01 -6.18417799e-01
-2.20615808e-02 3.93181682e-01 -9.47214305e-01 -8.92361701e-02
-1.31120992e+00 1.58279824e+00 1.60456434e-01 -1.88752517e-01
-2.85825402e-01 -6.15632057e-01 -1.09143353e+00 -3.84653211e-01
-3.62471074e-01 -5.40057242e-01 1.17354429e+00 -9.39305067e-01
-1.31268787e+00 1.19365847e+00 -5.00366926e-01 -3.94976139e-01
7.47656003e-02 -6.49796426e-01 -4.84052718e-01 -5.13557136e-01
5.44290006e-01 7.38756239e-01 2.60305345e-01 -6.08281910e-01
-4.53355104e-01 -3.18984568e-01 -4.59039927e-01 3.90899807e-01
-7.55267799e-01 2.68658996e-01 -3.60217899e-01 -7.12748826e-01
-6.70585409e-02 -1.20311880e+00 -5.27684748e-01 -5.18548608e-01
-3.68838072e-01 -1.55834034e-01 1.81100368e-01 -1.19810128e+00
1.24532318e+00 -1.78081644e+00 5.28144479e-01 3.30314226e-02
-1.45522013e-01 3.12341779e-01 -8.34728062e-01 5.83678126e-01
-1.15515195e-01 3.69591981e-01 -2.05107123e-01 -5.08648813e-01
1.03904918e-01 4.61439252e-01 -1.51815563e-01 4.37689066e-01
5.56117415e-01 1.30766332e+00 -9.68009591e-01 -5.12235045e-01
-1.44458070e-01 1.68995008e-01 -7.09259331e-01 1.13294534e-01
9.13348794e-02 3.15121353e-01 -2.66340002e-02 4.36129451e-01
3.05638343e-01 3.14394504e-01 6.70206726e-01 -1.44791275e-01
-2.59115726e-01 8.93693328e-01 -5.38303018e-01 2.20358348e+00
-1.00694335e+00 7.92516530e-01 -3.36297065e-01 -7.47628808e-01
1.00598562e+00 4.85848576e-01 2.95477539e-01 -7.87846863e-01
-8.62361044e-02 8.50972354e-01 2.86697149e-01 -2.76043504e-01
9.01366532e-01 -2.40103215e-01 -4.47715402e-01 8.00592959e-01
3.90385091e-01 -3.15014660e-01 1.33313611e-01 -8.70027393e-02
1.20798898e+00 2.98401862e-01 5.06750107e-01 -5.12534857e-01
-3.47615639e-03 4.12699491e-01 8.06600094e-01 2.11108550e-01
1.18480548e-01 4.10377473e-01 1.21449538e-01 -4.78844970e-01
-1.85716009e+00 -1.17395663e+00 -2.03958303e-01 1.06532311e+00
-2.21054465e-01 -7.84470916e-01 -7.97977209e-01 -5.93912125e-01
2.00218353e-02 8.05841327e-01 -2.99272567e-01 5.01497127e-02
-1.19011164e+00 -8.51270318e-01 8.25611472e-01 6.42987072e-01
-4.22733277e-01 -9.47501957e-01 1.71994060e-01 4.71977353e-01
-4.27877486e-01 -1.30788267e+00 -1.09873176e+00 3.28119487e-01
-8.75782490e-01 -5.93157947e-01 -3.97509724e-01 -1.08195198e+00
6.75471544e-01 -1.38783336e-01 1.39134610e+00 -2.24023908e-01
9.08627436e-02 4.81804870e-02 -2.17510462e-01 -1.92011714e-01
-7.69996583e-01 8.38270426e-01 7.71075904e-01 -5.50354242e-01
9.90995109e-01 -7.50452936e-01 -7.79202953e-02 2.49704063e-01
-4.76710379e-01 5.87754212e-02 9.20666397e-01 1.22957873e+00
7.14821696e-01 -1.21537197e+00 5.90444446e-01 -8.47809792e-01
7.76309907e-01 -3.25787485e-01 -3.42966378e-01 1.83511585e-01
-8.45587611e-01 3.81768763e-01 7.49911249e-01 -5.46605885e-01
-3.99855852e-01 -1.69176966e-01 -2.75172383e-01 2.06638109e-02
1.96533710e-01 5.31464934e-01 -6.77169040e-02 3.24781567e-01
9.44115281e-01 9.38236937e-02 1.36948213e-01 -5.65682054e-01
7.73049057e-01 8.80347311e-01 6.40662491e-01 -7.88941681e-01
1.09121835e+00 -1.93177059e-01 -4.61647570e-01 -5.30894816e-01
-6.16587818e-01 -3.38861585e-01 -1.15458763e+00 4.57295269e-01
1.15691078e+00 -9.42383766e-01 2.40320534e-01 -1.11764312e-01
-1.53004074e+00 -2.27758244e-01 -3.26887488e-01 9.03538167e-01
-3.84609073e-01 1.45403698e-01 -6.63584232e-01 5.05137518e-02
-6.26340449e-01 -1.20222902e+00 1.01686049e+00 -2.57458419e-01
-1.03164375e+00 -1.23899031e+00 7.47476995e-01 2.91635811e-01
5.08778036e-01 -7.86980018e-02 1.02486038e+00 -1.11333537e+00
-2.00521439e-01 -3.30315173e-01 1.20164707e-01 5.31457305e-01
5.60204148e-01 -1.97886303e-01 -6.31862342e-01 -3.75509173e-01
-4.20647562e-01 -2.78041214e-01 3.16414833e-01 8.48011374e-02
2.87956774e-01 -3.86544496e-01 -3.50302488e-01 5.74400604e-01
1.19466281e+00 -2.74190158e-01 3.53085637e-01 3.83050263e-01
1.02050626e+00 4.21059787e-01 4.61997479e-01 -3.70862216e-01
3.63540113e-01 9.92945611e-01 -2.36239702e-01 -3.43280733e-01
-2.38775149e-01 -4.70476329e-01 8.76335561e-01 2.17884731e+00
5.61112538e-02 1.30210012e-01 -1.20996439e+00 9.64296222e-01
-1.58364332e+00 -5.08329928e-01 -2.06357241e-01 2.22168350e+00
1.42181993e+00 -7.73052871e-02 -3.28058600e-01 -3.87365729e-01
5.02314389e-01 1.45886138e-01 -1.74127236e-01 -9.19551134e-01
-4.13696527e-01 9.26671803e-01 8.97038877e-01 8.49730968e-01
-7.81390846e-01 1.54421866e+00 6.85308886e+00 8.28066409e-01
-8.95931721e-01 6.59265757e-01 1.50699943e-01 -1.26892567e-01
-9.24320936e-01 2.53837556e-01 -9.39442873e-01 4.05803233e-01
1.37946570e+00 -3.37504238e-01 4.04721230e-01 6.47203863e-01
4.35616560e-02 3.63273740e-01 -1.51483715e+00 7.72123039e-01
1.66470721e-01 -1.38845110e+00 -5.55334054e-02 1.89759657e-01
1.08307016e+00 6.29082620e-01 3.51809785e-02 1.91932544e-01
8.48147810e-01 -1.32187843e+00 4.82325405e-01 1.00835465e-01
1.22731066e+00 -6.11137807e-01 1.01513124e+00 -1.18239429e-02
-9.20182049e-01 5.34270942e-01 -5.20829260e-01 3.31188440e-02
4.54829276e-01 4.86119896e-01 -8.97064626e-01 5.84710658e-01
2.06621379e-01 7.84540594e-01 -4.30199265e-01 3.17495495e-01
-4.58642513e-01 8.01937401e-01 -3.22434604e-01 3.02234069e-02
3.95427346e-01 -5.34463525e-01 4.96485353e-01 1.42696655e+00
6.44982696e-01 -5.99385142e-01 2.41032600e-01 3.13237429e-01
-2.52625018e-01 8.01883340e-01 -9.26115692e-01 -4.63697433e-01
7.34514117e-01 1.11046720e+00 -6.29164055e-02 -1.86263308e-01
-6.95984364e-01 1.23462975e+00 6.69337749e-01 1.32422447e-01
-6.93859696e-01 -2.66425818e-01 1.18275142e+00 -6.92237988e-02
-2.93730199e-02 -7.58603454e-01 -4.80255544e-01 -1.31093216e+00
1.89070851e-01 -1.28129852e+00 -4.76810634e-02 -3.67325544e-01
-1.44995403e+00 9.75983918e-01 -3.48834664e-01 -1.07902348e+00
-5.34375787e-01 -7.71129131e-01 -4.35023814e-01 1.38355553e+00
-1.21708512e+00 -1.46317315e+00 3.97812933e-01 -3.13715748e-02
5.31834602e-01 -6.42181158e-01 1.52813661e+00 5.07960439e-01
-5.02840936e-01 1.05696487e+00 3.14468533e-01 4.18936789e-01
1.24367368e+00 -1.28774965e+00 1.31044197e+00 1.06894469e+00
9.32855904e-01 9.10495698e-01 5.82990885e-01 -6.02209210e-01
-1.38874543e+00 -9.92301285e-01 1.80177033e+00 -1.09303868e+00
1.07398903e+00 -5.10092914e-01 -6.94413304e-01 1.07043612e+00
7.69370317e-01 -8.40875506e-02 1.18987548e+00 6.30179822e-01
-6.20549500e-01 1.01270154e-01 -5.05069613e-01 1.02621305e+00
1.17329431e+00 -9.01310384e-01 -8.83917928e-01 4.03323174e-01
1.07581019e+00 -3.37637633e-01 -1.05985141e+00 1.28523603e-01
6.65042996e-01 -6.69786483e-02 7.08126724e-01 -1.09192181e+00
4.65900451e-01 -1.95518628e-01 -6.06910348e-01 -1.89027441e+00
-1.91256553e-01 -6.72723234e-01 4.90519822e-01 1.17003059e+00
1.07659304e+00 -8.35006297e-01 5.07862687e-01 5.85143603e-02
-4.29752916e-01 -5.91489255e-01 -1.17079926e+00 -1.11342740e+00
7.78609395e-01 -3.72701794e-01 7.33132958e-01 1.48265290e+00
1.70524478e-01 7.60963798e-01 -3.62816066e-01 -3.20268311e-02
2.57184237e-01 -1.31714776e-01 9.75541472e-01 -8.68922770e-01
-2.16867909e-01 -4.29828495e-01 -4.37834114e-01 -1.03123784e+00
6.05036795e-01 -1.58532560e+00 1.00865729e-01 -1.46376705e+00
8.71820971e-02 -6.88464344e-01 -2.19179735e-01 7.18762517e-01
-4.51163977e-01 5.01636147e-01 -9.24590304e-02 2.74885148e-01
6.49291426e-02 5.74796498e-01 7.92464018e-01 -2.02561557e-01
-6.69012740e-02 -6.36817992e-01 -5.72119296e-01 4.81252670e-01
8.69740963e-01 -8.23585868e-01 8.21376145e-02 -1.08652174e+00
3.56838316e-01 -5.17426789e-01 -4.22933817e-01 -6.42244101e-01
3.60168666e-02 -2.76124507e-01 3.74968499e-02 -2.47764830e-02
2.09185988e-01 -3.99342149e-01 8.47052634e-02 1.46016538e-01
-2.78655440e-01 1.01803374e+00 2.93387383e-01 4.47157770e-02
-2.97163099e-01 -2.87754536e-01 4.31235224e-01 1.61954492e-01
-3.38901311e-01 2.15507701e-01 -2.53832877e-01 4.05705690e-01
5.83675027e-01 -8.88707265e-02 -2.38606736e-01 1.49851469e-02
-3.65169078e-01 7.56974071e-02 7.30137587e-01 6.03433669e-01
9.28034037e-02 -1.95467854e+00 -1.07670212e+00 2.44040295e-01
4.03199106e-01 -2.31549934e-01 -5.38645148e-01 9.15169954e-01
-4.87637490e-01 3.63623351e-01 -3.69316131e-01 -4.36304182e-01
-1.16572726e+00 1.71609968e-01 -5.22550233e-02 -3.98987144e-01
-3.80638868e-01 6.85664117e-01 -2.95593619e-01 -1.12561798e+00
-4.15477902e-01 -2.46728137e-01 3.81576687e-01 6.92445934e-02
8.32382962e-02 -1.33701488e-01 2.52211779e-01 -8.61678302e-01
-3.57834369e-01 8.00901294e-01 -2.92008430e-01 -5.31978965e-01
1.35337973e+00 1.11068644e-01 -3.19846034e-01 5.88253677e-01
1.17234278e+00 8.31271768e-01 -5.18469691e-01 -6.19723976e-01
4.13001150e-01 -3.87789220e-01 -3.20593178e-01 -3.36655200e-01
-5.04118025e-01 8.64038527e-01 2.63795733e-01 -3.48958015e-01
4.23800886e-01 -4.70263809e-02 1.19954717e+00 3.67791265e-01
2.50115424e-01 -1.25280023e+00 -9.03997943e-02 7.43810475e-01
5.62293410e-01 -1.12632620e+00 -1.87475160e-01 8.90427902e-02
-5.47663391e-01 9.10831213e-01 5.89967489e-01 -4.81832959e-02
2.20482260e-01 5.20125270e-01 6.23132110e-01 1.31800845e-01
-5.75655162e-01 1.91040449e-02 3.81872594e-01 5.97666681e-01
8.81437302e-01 3.17446798e-01 -6.77770853e-01 4.32029963e-01
-7.95931756e-01 -7.52226114e-01 3.50681394e-01 5.10320663e-01
-1.62858754e-01 -2.03030491e+00 -1.76627100e-01 3.21354568e-01
-5.51950097e-01 -9.39475477e-01 -2.39480391e-01 6.93080306e-01
-1.32924803e-02 5.01982212e-01 3.84760857e-01 -4.72896993e-01
1.64242417e-01 5.45004010e-01 4.70694661e-01 -1.05700254e+00
-4.09954846e-01 -1.91719681e-01 5.55017233e-01 -4.65655714e-01
-1.99527100e-01 -8.24486971e-01 -8.98054957e-01 -3.61623436e-01
-3.96539360e-01 4.38989520e-01 8.24008703e-01 1.01938891e+00
2.88712889e-01 1.57211363e-01 4.59377587e-01 -6.78750932e-01
-5.30992746e-01 -1.24437594e+00 8.04373026e-02 2.97508031e-01
3.21864486e-02 -1.68601111e-01 -1.33706912e-01 5.63833863e-02] | [11.268411636352539, 10.215616226196289] |
2a4981a2-adde-4c73-a1c3-7210e2e41771 | mmd-fuse-learning-and-combining-kernels-for | 2306.08777 | null | https://arxiv.org/abs/2306.08777v1 | https://arxiv.org/pdf/2306.08777v1.pdf | MMD-FUSE: Learning and Combining Kernels for Two-Sample Testing Without Data Splitting | We propose novel statistics which maximise the power of a two-sample test based on the Maximum Mean Discrepancy (MMD), by adapting over the set of kernels used in defining it. For finite sets, this reduces to combining (normalised) MMD values under each of these kernels via a weighted soft maximum. Exponential concentration bounds are proved for our proposed statistics under the null and alternative. We further show how these kernels can be chosen in a data-dependent but permutation-independent way, in a well-calibrated test, avoiding data splitting. This technique applies more broadly to general permutation-based MMD testing, and includes the use of deep kernels with features learnt using unsupervised models such as auto-encoders. We highlight the applicability of our MMD-FUSE test on both synthetic low-dimensional and real-world high-dimensional data, and compare its performance in terms of power against current state-of-the-art kernel tests. | ['Arthur Gretton', 'Antonin Schrab', 'Felix Biggs'] | 2023-06-14 | null | null | null | null | ['hypothesis-testing', 'hypothesis-testing'] | ['methodology', 'miscellaneous'] | [ 2.83923090e-01 -1.43466219e-01 -4.46376801e-02 -4.79143053e-01
-7.97063529e-01 -7.41243660e-01 7.36352682e-01 3.18545192e-01
-8.64873171e-01 9.71377134e-01 -1.23187326e-01 -3.60174030e-01
-9.13168132e-01 -9.61544991e-01 -6.91417515e-01 -9.33301747e-01
-5.54597735e-01 5.98613799e-01 5.56053638e-01 2.19590977e-01
3.61370236e-01 6.09853208e-01 -1.65935516e+00 -1.67257950e-01
1.16410553e+00 8.44376266e-01 -1.62571937e-01 8.67141426e-01
3.43422264e-01 4.45798784e-02 -6.16137564e-01 -4.20227528e-01
2.56892294e-01 -3.51496577e-01 -6.55387104e-01 -3.01636998e-02
4.85129595e-01 9.21837911e-02 -6.08088672e-02 1.12814534e+00
6.89264953e-01 1.35162577e-01 1.17048264e+00 -1.14183998e+00
-5.08426487e-01 3.48623961e-01 -2.29364872e-01 2.20630884e-01
9.30724759e-03 -1.85177270e-02 7.68890083e-01 -7.50759363e-01
2.03466877e-01 7.17764854e-01 7.10901558e-01 4.04568799e-02
-1.85182035e+00 -4.43550736e-01 -6.21207654e-01 7.14716092e-02
-1.34150839e+00 -1.64263427e-01 2.28690267e-01 -5.40729702e-01
6.87310755e-01 2.56899685e-01 5.28684676e-01 9.93943155e-01
4.02805567e-01 3.79685789e-01 1.52486169e+00 -5.61792552e-01
6.43555045e-01 1.41082972e-01 8.53584185e-02 4.86464977e-01
7.28022814e-01 1.35779098e-01 -1.96809694e-01 -6.59546614e-01
8.97556722e-01 -4.50623065e-01 -2.43219718e-01 -9.02337134e-01
-1.41132700e+00 9.53551531e-01 -3.48032832e-01 2.74014860e-01
-1.89774439e-01 -2.11628452e-01 4.13593143e-01 2.47195229e-01
7.13557422e-01 5.35551190e-01 -7.90427387e-01 -1.58758864e-01
-1.06890070e+00 2.44093820e-01 1.10064566e+00 5.27935088e-01
7.78548956e-01 -1.08119227e-01 -4.67725188e-01 8.32722604e-01
-6.40997887e-02 7.35192001e-01 3.68457645e-01 -5.13229549e-01
-6.81248307e-02 2.29413569e-01 7.08705410e-02 -4.20037001e-01
-4.38653231e-01 -5.68591654e-01 -9.64286029e-01 3.53996992e-01
6.61770761e-01 -1.43910185e-01 -7.93735802e-01 1.98822594e+00
3.69939536e-01 6.60599828e-01 6.20509572e-02 6.09422147e-01
4.57749814e-02 1.71661779e-01 -9.32923332e-02 -2.40246177e-01
1.08480990e+00 -1.83388427e-01 -2.14280531e-01 2.81328827e-01
8.53742123e-01 -6.00670815e-01 9.30874050e-01 5.79376638e-01
-8.40882599e-01 -3.88541728e-01 -1.11679184e+00 4.69079763e-01
-4.89757508e-01 2.47820780e-01 4.60078388e-01 1.03944957e+00
-1.06126797e+00 7.84571946e-01 -6.16345704e-01 -3.30437720e-01
3.32518965e-01 5.63376606e-01 -6.22884274e-01 1.19796753e-01
-1.29265296e+00 7.62370884e-01 4.96771187e-01 -2.31026858e-03
-5.82132757e-01 -9.08712685e-01 -7.39909887e-01 -5.54136038e-02
2.34801397e-01 -4.97601032e-01 6.82760894e-01 -4.31244701e-01
-1.47702312e+00 7.99903214e-01 4.17035431e-01 -6.50193214e-01
4.25427675e-01 -1.94854036e-01 -5.79697549e-01 7.34610716e-03
9.91749987e-02 1.93048075e-01 8.36630821e-01 -6.60438418e-01
-3.64876032e-01 -1.81794390e-01 -3.03604931e-01 -2.89861709e-01
-3.57810140e-01 -2.33345345e-01 -1.00953005e-01 -5.78097999e-01
-3.04771721e-01 -8.74023378e-01 -2.05017880e-01 -3.84584069e-01
-4.99932379e-01 -1.66875049e-01 5.61931551e-01 -3.51581722e-01
1.18917704e+00 -2.10520387e+00 7.65966177e-02 7.31188774e-01
-3.53568755e-02 3.52966934e-01 -2.72959024e-01 4.71906990e-01
-2.50949711e-01 -9.20046344e-02 -6.94001377e-01 1.96764395e-01
2.02373669e-01 1.54256210e-01 -1.56159759e-01 9.36065733e-01
5.78868151e-01 6.56005681e-01 -6.98318183e-01 -2.17350721e-01
2.58702725e-01 5.14731884e-01 -4.18540359e-01 8.67155492e-02
1.46122351e-01 9.15612355e-02 -1.36898518e-01 8.84443820e-02
9.46303427e-01 -1.20587558e-01 4.57948484e-02 -7.96610788e-02
-1.22226797e-01 3.93550359e-02 -1.45194733e+00 1.29864895e+00
-2.13721380e-01 4.74211544e-01 -4.03330714e-01 -1.32754552e+00
1.01607132e+00 -1.61929324e-01 2.15118095e-01 -4.48087424e-01
-1.11838005e-01 4.11575794e-01 2.29069844e-01 -2.99023956e-01
8.30671471e-03 -8.83649290e-02 -9.91276130e-02 4.15609747e-01
4.24744695e-01 -6.38348982e-02 4.03725177e-01 -8.71809348e-02
1.35863054e+00 -7.04321042e-02 5.74899435e-01 -1.02792263e+00
5.72228014e-01 -5.91915131e-01 1.74227595e-01 9.83004510e-01
1.62794724e-01 4.03513283e-01 9.16228473e-01 6.36686608e-02
-1.13302100e+00 -1.58828533e+00 -9.74557221e-01 5.69233835e-01
-3.16779166e-01 -6.14479696e-03 -7.34803498e-01 -6.95784032e-01
3.70867014e-01 7.31833756e-01 -1.03108108e+00 -3.92453671e-01
-8.36541429e-02 -1.24587488e+00 8.79942834e-01 1.91381022e-01
4.46975946e-01 -6.07245207e-01 -7.10246623e-01 -4.56853621e-02
6.07953548e-01 -8.29890311e-01 -2.15784296e-01 7.21974194e-01
-7.32846737e-01 -1.31692111e+00 -7.01998949e-01 -4.60692257e-01
5.57117045e-01 -3.38014096e-01 9.40185606e-01 -7.16703594e-01
-3.19052726e-01 5.49116135e-01 -2.39963129e-01 -7.30507001e-02
-2.45492220e-01 -4.78229709e-02 2.85683155e-01 1.49100170e-01
4.20427650e-01 -5.58090150e-01 -4.48630363e-01 4.57424015e-01
-1.34677541e+00 -5.28159499e-01 9.73569572e-01 1.02891552e+00
4.91898894e-01 3.00134420e-01 6.38658047e-01 -8.50159347e-01
8.46521795e-01 -5.77172816e-01 -8.77956808e-01 3.48392010e-01
-6.30484521e-01 5.79127848e-01 6.55380547e-01 -7.99839199e-01
-4.80361581e-01 -3.51700157e-01 3.00467461e-01 -1.98572204e-01
-3.18295747e-01 4.27511811e-01 -1.55048415e-01 -3.20851892e-01
6.74602687e-01 3.55840594e-01 1.19421840e-01 -1.33190140e-01
2.96484947e-01 4.48074013e-01 5.25688112e-01 -7.58134305e-01
8.14026773e-01 2.32845470e-01 6.11249447e-01 -8.94743145e-01
-4.89483356e-01 -4.75749165e-01 -8.60831499e-01 9.55131054e-02
6.24329269e-01 -3.73517275e-01 -8.64839077e-01 5.97071350e-01
-5.93688548e-01 -5.00084400e-01 -5.64033329e-01 9.92403209e-01
-8.28953743e-01 4.01964188e-01 -1.25643179e-01 -7.78044224e-01
1.22003436e-01 -9.94549215e-01 9.74900663e-01 -5.04023470e-02
-1.00394888e-02 -1.32565510e+00 6.20480835e-01 -4.69407290e-01
2.26610348e-01 2.05315307e-01 1.14669418e+00 -1.31290805e+00
-1.76466312e-02 -3.98999840e-01 -2.91003197e-01 8.44145477e-01
2.37219289e-01 1.29744895e-02 -7.52507985e-01 -3.21831822e-01
-1.79703757e-01 -2.86970198e-01 1.01494491e+00 5.91484427e-01
1.09557009e+00 -1.51628599e-01 -9.47908387e-02 5.05477786e-01
1.51181638e+00 -3.45589727e-01 6.84129894e-01 1.98771656e-01
1.57536581e-01 5.66698253e-01 4.66879517e-01 4.48907226e-01
-2.60422796e-01 5.84035933e-01 5.55942655e-02 4.06227782e-02
3.45908552e-01 1.90594718e-01 2.81420022e-01 4.90515083e-01
1.26959696e-01 -1.33153304e-01 -6.97324336e-01 5.28467536e-01
-1.75508785e+00 -9.72661734e-01 -1.78084686e-01 3.02638602e+00
9.87137616e-01 1.62349090e-01 3.10868770e-01 2.39466906e-01
7.17333138e-01 -1.25720724e-01 -4.61041361e-01 -4.17864233e-01
-4.66768742e-01 9.97036278e-01 8.53354931e-01 2.71783799e-01
-1.26503837e+00 3.24774712e-01 6.82248068e+00 1.22681606e+00
-8.77624512e-01 -8.33795071e-02 4.53559875e-01 1.67550430e-01
-2.26007223e-01 5.46742231e-02 -5.00155270e-01 5.41403770e-01
1.29309607e+00 -2.33270451e-01 1.20357476e-01 6.85273111e-01
1.57329421e-02 -5.64161718e-01 -1.02923405e+00 6.51160300e-01
6.84102997e-02 -1.06516623e+00 -1.43126130e-01 4.10091728e-01
8.43964994e-01 -2.34239891e-01 3.19772005e-01 1.65432617e-01
1.54829636e-01 -1.03662336e+00 1.79554746e-01 6.92717731e-01
8.95277321e-01 -1.09237182e+00 9.99672592e-01 1.38277426e-01
-6.74581528e-01 2.82854468e-01 -4.96717781e-01 2.15208635e-01
-3.88521761e-01 1.49765301e+00 -8.57997596e-01 5.87532222e-01
3.52449745e-01 3.29363912e-01 -6.34976685e-01 1.35201073e+00
8.80751014e-03 8.38989973e-01 -5.28156996e-01 -1.83346570e-01
7.13273287e-02 -4.48058754e-01 4.15944070e-01 1.41949189e+00
5.89971364e-01 -4.12853152e-01 -1.20202236e-01 7.54659176e-01
2.63557464e-01 1.91478491e-01 -5.23254335e-01 6.53505251e-02
3.96202952e-01 1.11604798e+00 -6.94885552e-01 -1.97880581e-01
-3.67571861e-01 1.03901708e+00 3.12330425e-01 2.90969968e-01
-8.06476593e-01 -7.38122284e-01 7.60551929e-01 4.64613922e-02
5.73314846e-01 -3.45076710e-01 1.17281258e-01 -8.90182197e-01
3.51033658e-02 -5.35513699e-01 4.00835842e-01 -3.29439580e-01
-1.78172743e+00 6.38334751e-02 4.69632149e-01 -1.26820052e+00
-3.26250225e-01 -1.06510699e+00 -5.37080109e-01 9.98992145e-01
-1.26736820e+00 -5.78332365e-01 3.10674071e-01 4.04903889e-01
-1.93541467e-01 -5.83477505e-02 1.06146061e+00 3.13210934e-02
-1.93211362e-01 6.56704187e-01 5.28067708e-01 7.32834339e-02
1.02923369e+00 -1.41597545e+00 1.90258861e-01 6.85760438e-01
1.20790839e-01 5.11700571e-01 6.92075193e-01 -6.14085734e-01
-1.11601973e+00 -1.03148079e+00 4.10359502e-01 -3.32492560e-01
1.01718986e+00 -4.78476614e-01 -1.02752638e+00 1.41016915e-01
-1.68524697e-01 2.00645819e-01 1.05307865e+00 2.89268613e-01
-5.24193525e-01 -2.32433099e-02 -1.25966072e+00 3.62495661e-01
7.76309013e-01 -4.01507467e-01 -4.65565056e-01 1.90046147e-01
2.08827302e-01 9.21231881e-02 -1.18879473e+00 8.12515438e-01
5.71530223e-01 -1.27386367e+00 8.74341667e-01 -7.50220954e-01
6.04045801e-02 -3.08734208e-01 -1.12889171e-01 -1.44794142e+00
-4.45379585e-01 -4.76316303e-01 9.08260345e-02 8.39426458e-01
3.34885627e-01 -1.00308359e+00 4.96000141e-01 6.36875033e-02
2.09688231e-01 -9.11732435e-01 -1.33441484e+00 -1.13104737e+00
8.97481963e-02 -6.46482706e-01 3.93925816e-01 9.28168595e-01
-2.65226271e-02 -1.15829654e-01 -2.10837677e-01 4.16301101e-01
8.86517048e-01 -2.49547243e-01 8.00388038e-01 -1.40277874e+00
-6.72668397e-01 -4.80980366e-01 -1.10801399e+00 -5.26884794e-01
2.39311934e-01 -7.88891971e-01 -1.41836286e-01 -9.44816291e-01
2.31962636e-01 -3.16150695e-01 -4.93979931e-01 1.43511787e-01
-1.24142051e-01 2.15657994e-01 -5.40052712e-01 -2.18320727e-01
-3.56518865e-01 4.43398923e-01 9.18304265e-01 1.90247789e-01
-4.88581248e-02 6.10570014e-02 -3.51075023e-01 3.95835370e-01
7.72844136e-01 -3.58900815e-01 -4.83887643e-01 4.31203216e-01
2.70865470e-01 -3.11995238e-01 6.01877511e-01 -1.38134205e+00
1.14947043e-01 -1.66311696e-01 6.85040653e-01 -3.26749355e-01
-2.75452286e-02 -4.76320267e-01 -8.10383260e-02 2.79811919e-01
-4.00416762e-01 -1.83911607e-01 4.79455322e-01 5.99972427e-01
1.09280832e-02 -2.31269136e-01 8.36086452e-01 4.89431143e-01
-3.81405115e-01 2.15116218e-01 -2.00518057e-01 2.32416168e-01
1.18620455e+00 -2.57859588e-01 -3.85910600e-01 -1.55762926e-01
-5.20195901e-01 -9.81785432e-02 3.38938922e-01 -8.63720551e-02
5.21587193e-01 -1.29362917e+00 -8.85255992e-01 6.06064975e-01
3.86394501e-01 -3.54359299e-01 2.84353226e-01 1.16427696e+00
-1.24821343e-01 3.84399146e-01 -3.23020667e-01 -6.85920715e-01
-8.59894216e-01 4.62856382e-01 9.67289880e-02 -3.32757890e-01
-1.74453899e-01 4.78965282e-01 2.19925389e-01 -4.85982269e-01
-2.75077492e-01 -3.01369280e-01 1.39392436e-01 -1.01560533e-01
4.48873162e-01 4.00574356e-01 2.09225073e-01 -1.98301151e-01
-2.88835198e-01 6.03731453e-01 1.88758686e-01 -2.60890007e-01
1.12896144e+00 3.85050625e-01 -2.24970102e-01 6.85122967e-01
1.31757736e+00 3.24506432e-01 -1.04658067e+00 -1.26282116e-02
6.86678803e-03 -4.13383573e-01 -2.35537693e-01 -6.50816500e-01
-3.18849057e-01 5.50945520e-01 5.47820747e-01 4.36743140e-01
1.15051734e+00 3.82037088e-02 3.77211235e-02 3.69069725e-01
2.39901021e-01 -1.08402359e+00 -1.16323031e-01 4.93432045e-01
3.89103174e-01 -7.90800095e-01 5.64009659e-02 -3.35582465e-01
-2.45350584e-01 1.32229424e+00 1.69479296e-01 -4.44863170e-01
8.26080978e-01 4.60383534e-01 -4.62063164e-01 1.99051142e-01
-5.38703382e-01 -5.05197763e-01 6.47275329e-01 9.05021369e-01
2.69461364e-01 1.47155717e-01 -4.83857363e-01 2.79887170e-01
-3.88486944e-02 -7.50306919e-02 3.56382966e-01 5.52332580e-01
-3.88935953e-01 -1.24480939e+00 -1.95741355e-01 9.50544536e-01
-1.19273573e-01 -1.99386001e-01 -1.14760183e-01 1.00194955e+00
-1.21023012e-02 4.12774533e-01 2.86027670e-01 -3.41964364e-01
1.06301278e-01 3.50227207e-01 7.63275325e-01 -4.08652931e-01
7.48204216e-02 -1.59423128e-01 -4.06849757e-03 -3.31117302e-01
-4.95509028e-01 -6.49293005e-01 -7.41439164e-01 -2.04323411e-01
-4.24742073e-01 3.57692480e-01 4.46518809e-01 9.19115067e-01
3.27864975e-01 2.31969237e-01 5.37018001e-01 -5.10770082e-01
-8.30283165e-01 -1.00777054e+00 -9.46541131e-01 3.16694856e-01
1.04889266e-01 -9.03694630e-01 -7.37603605e-01 -3.59165668e-01] | [7.467051982879639, 4.17130708694458] |
3930d8c4-9e93-4596-a095-6685272c52cd | bvi-vfi-a-video-quality-database-for-video | 2210.00823 | null | https://arxiv.org/abs/2210.00823v2 | https://arxiv.org/pdf/2210.00823v2.pdf | BVI-VFI: A Video Quality Database for Video Frame Interpolation | Video frame interpolation (VFI) is a fundamental research topic in video processing, which is currently attracting increased attention across the research community. While the development of more advanced VFI algorithms has been extensively researched, there remains little understanding of how humans perceive the quality of interpolated content and how well existing objective quality assessment methods perform when measuring the perceived quality. In order to narrow this research gap, we have developed a new video quality database named BVI-VFI, which contains 540 distorted sequences generated by applying five commonly used VFI algorithms to 36 diverse source videos with various spatial resolutions and frame rates. We collected more than 10,800 quality ratings for these videos through a large scale subjective study involving 189 human subjects. Based on the collected subjective scores, we further analysed the influence of VFI algorithms and frame rates on the perceptual quality of interpolated videos. Moreover, we benchmarked the performance of 28 classic and state-of-the-art objective image/video quality metrics on the new database, and demonstrated the urgent requirement for more accurate bespoke quality assessment methods for VFI. To facilitate further research in this area, we have made BVI-VFI publicly available at https://github.com/danielism97/BVI-VFI-database. | ['David Bull', 'Fan Zhang', 'Duolikun Danier'] | 2022-10-03 | null | null | null | null | ['video-frame-interpolation'] | ['computer-vision'] | [ 3.44590582e-02 -7.99410760e-01 -9.44539830e-02 -4.12609398e-01
-6.56754315e-01 -2.66411394e-01 1.59961447e-01 4.85060811e-02
-3.03099751e-01 8.33207130e-01 3.67988974e-01 -1.06946968e-01
-1.34928999e-02 -6.36627257e-01 -7.17953980e-01 -3.94217819e-01
-4.46551561e-01 -3.79223615e-01 3.52414072e-01 -8.92043579e-04
5.48367143e-01 1.86984271e-01 -1.79873157e+00 4.02198672e-01
1.08692229e+00 1.25190926e+00 2.23345250e-01 8.17901552e-01
4.61829692e-01 6.54891312e-01 -8.60480726e-01 -4.44222718e-01
1.81195781e-01 -5.37275910e-01 -4.98102546e-01 1.85465515e-02
4.45745647e-01 -6.48583174e-01 -4.29566294e-01 1.02998328e+00
6.28368855e-01 2.02009827e-01 3.05740416e-01 -1.35985494e+00
-8.09289992e-01 2.76150089e-02 -3.26218963e-01 8.25856805e-01
8.55755806e-01 4.26773727e-01 8.12773883e-01 -1.02279651e+00
5.59650064e-01 1.11830354e+00 4.74027872e-01 2.14744195e-01
-1.00241733e+00 -7.85575867e-01 -4.25380319e-01 9.01657820e-01
-1.61544466e+00 -6.45643890e-01 4.86517549e-01 -5.07523477e-01
6.05441630e-01 2.59835064e-01 6.94485128e-01 8.01072419e-01
4.01490003e-01 5.70959210e-01 9.79360759e-01 -2.62207448e-01
2.39855766e-01 -1.31253347e-01 -1.11609280e-01 3.56918544e-01
2.91859031e-01 3.49434048e-01 -7.32677162e-01 3.21147144e-02
1.10305202e+00 -4.38480228e-01 -7.25152254e-01 -7.45476037e-02
-1.47999656e+00 5.55629134e-01 1.24849945e-01 3.42146188e-01
-2.55608886e-01 -1.37827575e-01 5.64631164e-01 5.46830714e-01
4.75523621e-01 1.13983780e-01 -3.48390415e-02 -5.44028640e-01
-1.02804983e+00 3.61645073e-01 4.53688800e-01 9.20831203e-01
5.53609908e-01 1.37799054e-01 -2.47985810e-01 9.26466644e-01
3.36767495e-01 6.28626347e-01 2.32939184e-01 -1.55951476e+00
5.21075308e-01 4.52607721e-02 4.37456638e-01 -1.43343329e+00
5.46176322e-02 2.50433628e-02 -8.24600518e-01 5.26705921e-01
4.81475472e-01 1.75329655e-01 -2.85860360e-01 1.33890295e+00
-2.26841327e-02 2.11757451e-01 -2.95940578e-01 1.27992427e+00
8.55375648e-01 8.67917895e-01 -3.53799835e-02 -4.15095925e-01
9.83810127e-01 -6.63351595e-01 -7.81564713e-01 4.22366172e-01
4.71761711e-02 -1.08044600e+00 1.16804338e+00 7.34440982e-01
-1.58209884e+00 -1.09006393e+00 -1.19110155e+00 7.33154565e-02
2.22875014e-01 -2.83244606e-02 1.73387006e-01 7.53819108e-01
-1.39888191e+00 5.10816276e-01 -4.14777011e-01 -7.79743865e-02
3.56908143e-01 1.09891556e-01 -2.89349884e-01 -3.15432280e-01
-1.36811376e+00 7.10551500e-01 1.61945224e-01 1.25495847e-02
-1.04194975e+00 -7.90051341e-01 -6.95765793e-01 -1.01007141e-01
3.35962415e-01 -5.83570480e-01 1.10938966e+00 -1.04553568e+00
-1.35544538e+00 5.36366880e-01 -5.43142200e-01 -1.46558315e-01
4.32888299e-01 -2.16483548e-01 -9.24882233e-01 4.45276976e-01
7.88397565e-02 4.90612656e-01 9.95273650e-01 -1.24709320e+00
-7.89251029e-01 -9.38745737e-02 1.45119280e-01 2.53728449e-01
-4.59506065e-02 3.44428331e-01 -4.88285094e-01 -8.81865382e-01
-2.16057092e-01 -5.85909307e-01 3.92429650e-01 3.47686708e-01
1.29104689e-01 -8.18558410e-02 5.53915143e-01 -9.28795338e-01
1.65685105e+00 -2.38942051e+00 -1.96172576e-02 8.13932791e-02
2.55607665e-01 5.46775162e-01 -2.21356541e-01 1.64271384e-01
1.27915844e-01 1.88697249e-01 -3.98484897e-03 1.55348510e-01
-3.82825464e-01 2.97137216e-04 1.33433059e-01 5.10755956e-01
-9.29256156e-02 4.28665698e-01 -1.09693325e+00 -5.41696429e-01
5.10148764e-01 6.18534446e-01 -7.66202629e-01 4.27505016e-01
2.37213761e-01 6.79606795e-01 4.14767824e-02 8.70214581e-01
8.56011689e-01 -1.12134285e-01 -3.26885223e-01 -4.90685880e-01
-2.23696634e-01 -8.57166871e-02 -1.30165732e+00 1.56075835e+00
-2.19142362e-01 1.04446483e+00 1.06251620e-01 -7.17676699e-01
6.97026432e-01 6.80315435e-01 5.82593620e-01 -8.33832383e-01
2.97013652e-02 3.10334325e-01 -1.70004070e-02 -7.06951916e-01
6.04183972e-01 1.10622950e-01 4.66525763e-01 -5.23030907e-02
-3.56209427e-02 7.74822012e-02 7.68350720e-01 2.05579568e-02
7.01044202e-01 8.76430646e-02 1.75979197e-01 -2.51075298e-01
8.13319921e-01 -3.09766680e-01 5.46695769e-01 5.20892262e-01
-7.68368900e-01 8.61903369e-01 1.38168722e-01 -3.75132978e-01
-1.22219384e+00 -1.39629364e+00 -2.61890680e-01 7.77697265e-01
5.54273784e-01 -5.27420580e-01 -6.74324453e-01 7.23669007e-02
-2.17283472e-01 3.38460088e-01 -2.70241678e-01 2.33024180e-01
-3.99810135e-01 -4.20720637e-01 3.96269709e-01 2.25232601e-01
8.79287362e-01 -1.08981502e+00 -5.02088487e-01 2.74582505e-01
-8.34146976e-01 -1.06724095e+00 -6.46607339e-01 -9.36009049e-01
-8.51298928e-01 -1.15191436e+00 -1.08199215e+00 -5.99852264e-01
2.35603690e-01 7.06777215e-01 1.42988825e+00 4.34300601e-01
-1.63135186e-01 2.80696213e-01 -5.80421448e-01 -3.22830379e-02
-4.92273510e-01 -3.73318523e-01 9.27068666e-02 1.11677451e-02
2.99220141e-02 -4.74827677e-01 -9.98123467e-01 9.64628816e-01
-1.01192546e+00 2.75898948e-02 1.03677489e-01 4.04064864e-01
4.89709735e-01 3.26245040e-01 4.70197350e-01 -1.19157068e-01
6.72841370e-01 -4.71787632e-01 -6.20934546e-01 -1.76649429e-02
-3.99475843e-01 -4.53041762e-01 4.28430468e-01 -2.07296848e-01
-1.08878052e+00 -9.42653716e-01 -3.73023152e-01 -2.29517996e-01
-5.91765717e-02 4.08449084e-01 -1.89559028e-01 -1.54977426e-01
6.24709904e-01 1.27832349e-02 -2.58742254e-02 -6.03447258e-02
-7.43811131e-02 7.66577065e-01 7.36546934e-01 -4.72100317e-01
5.72041154e-01 3.11819583e-01 -2.79308826e-01 -1.09660339e+00
-2.83169657e-01 -4.15616542e-01 -2.31487229e-01 -8.47702801e-01
8.18056345e-01 -1.11030364e+00 -7.52537787e-01 7.67852902e-01
-1.02190804e+00 -3.00782621e-01 1.54605076e-01 7.63412654e-01
-6.32885456e-01 7.64270246e-01 -6.78524137e-01 -5.52416027e-01
-1.28440976e-01 -1.52293527e+00 6.99451268e-01 2.14832604e-01
-2.41618425e-01 -8.01470935e-01 1.43211556e-03 5.13711214e-01
7.35268831e-01 9.40857530e-02 5.70367038e-01 5.10975242e-01
-7.06665218e-01 2.78761327e-01 -4.68676537e-01 5.87916672e-01
2.42732897e-01 3.86260033e-01 -6.57054484e-01 -4.65959638e-01
1.05898576e-02 -1.01282895e-01 2.81340867e-01 7.09513545e-01
1.13399827e+00 -5.44001050e-02 3.73908728e-01 6.63840711e-01
1.60981131e+00 4.85138148e-01 1.20141363e+00 4.74454194e-01
3.93001199e-01 1.80334270e-01 8.38072836e-01 7.34850883e-01
2.38660142e-01 8.58067453e-01 3.32571596e-01 1.32407531e-01
-2.71459043e-01 6.96831644e-02 4.16812360e-01 8.77806306e-01
-8.50097656e-01 -4.11239982e-01 -5.99655211e-01 4.30466294e-01
-1.25917602e+00 -1.28513670e+00 -1.05064832e-01 2.33044147e+00
7.38282681e-01 3.16924788e-02 2.87490129e-01 6.22829854e-01
8.75226319e-01 1.42500564e-01 -2.30714679e-01 -2.22911149e-01
-1.40694082e-01 -9.36864242e-02 2.20542327e-01 4.38431948e-01
-8.65053415e-01 3.02267015e-01 6.67034721e+00 7.87580550e-01
-1.11049557e+00 -5.66409603e-02 7.98028648e-01 -3.74388658e-02
-9.00634676e-02 -2.83907056e-01 -3.84549469e-01 8.90501976e-01
1.15415156e+00 -4.40485895e-01 6.15472853e-01 4.58015263e-01
1.04291153e+00 -5.26045561e-01 -9.00442541e-01 1.33258879e+00
-4.39307978e-03 -1.26796556e+00 2.82296222e-02 6.02553189e-02
7.09282935e-01 -2.61860549e-01 1.30349323e-01 -4.95669842e-02
-4.57942486e-01 -9.71216977e-01 8.50016952e-01 5.34082830e-01
1.15130854e+00 -5.44275522e-01 7.28971839e-01 -1.65570527e-01
-1.39623594e+00 6.59367219e-02 -4.73140866e-01 -1.48336396e-01
4.53803569e-01 7.07115591e-01 -8.74215737e-02 6.10569000e-01
1.09147358e+00 7.57732749e-01 -4.91224349e-01 1.56447744e+00
-3.98474336e-02 8.06281626e-01 3.44522633e-02 4.39084142e-01
-2.01483443e-01 -2.72312492e-01 4.90493804e-01 1.05688512e+00
5.76671720e-01 2.82494634e-01 -2.46346697e-01 5.26717246e-01
1.16292901e-01 1.62200451e-01 -3.33744466e-01 1.61223248e-01
5.14562309e-01 7.31672525e-01 -3.42861831e-01 -4.92816359e-01
-7.68861175e-01 8.07609320e-01 -4.16935414e-01 5.33198118e-01
-1.15135527e+00 -1.94162354e-01 9.32947814e-01 3.48215252e-01
1.84901908e-01 -5.02050817e-01 -1.73199385e-01 -1.34877777e+00
1.88478634e-01 -1.17330348e+00 1.18420616e-01 -1.21137285e+00
-1.04538870e+00 5.46510696e-01 2.96415985e-01 -1.63400781e+00
-1.34250954e-01 -3.33626330e-01 -3.80324602e-01 9.29828048e-01
-1.55308485e+00 -2.13431716e-01 -7.77718782e-01 8.81139576e-01
8.40969265e-01 -1.81741547e-02 4.60435241e-01 8.43920112e-01
-3.39328498e-01 5.52540600e-01 1.45151466e-01 -6.88874125e-02
9.73245382e-01 -5.64107656e-01 2.57220954e-01 8.30379665e-01
-2.08562210e-01 3.98720950e-01 8.27340961e-01 -4.91735280e-01
-1.31192005e+00 -7.33522058e-01 7.34353423e-01 -2.52902638e-02
3.13622057e-01 3.11829954e-01 -1.11094034e+00 1.54480189e-01
3.68471652e-01 4.66529541e-02 6.00183427e-01 -5.90780616e-01
4.91255485e-02 -2.06058517e-01 -1.29751444e+00 4.91486132e-01
1.08375347e+00 -3.74854237e-01 -1.64013952e-01 -2.48559013e-01
3.72183263e-01 -2.50292748e-01 -1.32177401e+00 4.73084748e-01
6.90338790e-01 -1.63249159e+00 1.17481172e+00 3.22913915e-01
6.70800507e-01 -5.56910694e-01 -3.32608134e-01 -1.22994137e+00
-4.60313648e-01 -3.87830108e-01 -3.38983443e-03 1.03236318e+00
9.68272686e-02 -3.88884991e-01 4.90419447e-01 5.50109982e-01
-3.13119031e-02 -5.57780027e-01 -8.45310152e-01 -8.25311124e-01
-3.62768412e-01 -5.09232581e-01 4.90969956e-01 5.09663224e-01
-1.25603646e-01 -1.91764951e-01 -6.23952508e-01 3.15581895e-02
7.14739978e-01 -3.46832931e-01 7.44038105e-01 -8.21955144e-01
-1.80485487e-01 -3.53870451e-01 -7.62565076e-01 -1.02212739e+00
-3.48972648e-01 -3.99099261e-01 -2.29858890e-01 -1.43464875e+00
1.06743023e-01 -1.22601978e-01 -1.54726461e-01 -1.97248369e-01
-3.80747616e-01 6.43059194e-01 3.86636466e-01 5.00733912e-01
-4.85175818e-01 5.18433213e-01 1.48215556e+00 6.25782982e-02
-5.16757593e-02 -6.23514839e-02 -3.53783160e-01 5.59840798e-01
8.66048455e-01 8.49376619e-02 -5.87763429e-01 -6.31340861e-01
-1.18424753e-02 4.66724187e-01 2.96331197e-01 -1.50696504e+00
-8.34281296e-02 -2.37388670e-01 4.58198577e-01 -4.70598012e-01
2.28446171e-01 -7.04879582e-01 4.15541947e-01 1.79365486e-01
-1.46223113e-01 4.05253470e-01 4.79837544e-02 2.98769563e-01
-6.34458184e-01 -1.44257993e-01 8.35145175e-01 -8.72075930e-03
-1.21981657e+00 3.64463449e-01 -5.34451962e-01 2.27721199e-01
9.53250349e-01 -6.02864683e-01 -6.36645779e-02 -6.73541069e-01
-4.54203784e-01 -1.10532470e-01 8.00262809e-01 3.43210965e-01
1.04527688e+00 -1.40177608e+00 -1.01389790e+00 3.29192042e-01
6.94419593e-02 -4.42801744e-01 4.83459622e-01 6.77389443e-01
-9.36672747e-01 2.28218868e-01 -7.04109192e-01 -7.46961176e-01
-1.36251605e+00 5.98668635e-01 4.86090453e-03 1.35851607e-01
-4.99594778e-01 4.05123174e-01 4.23971042e-02 4.64120060e-01
2.09802389e-01 -3.08632940e-01 -1.53287172e-01 -2.87254602e-01
8.73221993e-01 9.11951125e-01 4.98721413e-02 -9.49291587e-01
-1.75585598e-01 4.70681041e-01 4.68689740e-01 -2.10422635e-01
9.15175736e-01 -6.80321634e-01 1.32491514e-01 3.19714755e-01
1.22095788e+00 -8.77162591e-02 -1.30298722e+00 6.48801476e-02
-3.60829413e-01 -1.28479874e+00 -1.25457317e-01 -5.22442579e-01
-1.15439606e+00 6.12333179e-01 8.25706482e-01 1.81976050e-01
1.55547881e+00 -5.37778378e-01 9.08801913e-01 -3.58505666e-01
7.33893216e-01 -8.28631163e-01 1.03933610e-01 3.14705789e-01
1.00399232e+00 -1.27867961e+00 2.59416886e-02 -6.10329986e-01
-5.02561808e-01 9.83445168e-01 3.23756099e-01 -2.88761146e-02
5.47103047e-01 -3.40885036e-02 1.50541171e-01 2.64606088e-01
-4.65014189e-01 2.40469545e-01 2.90956527e-01 7.28447139e-01
8.75996590e-01 -9.33134705e-02 -6.85836375e-01 -2.92678885e-02
-4.21273001e-02 7.01940656e-01 6.79029286e-01 6.64031386e-01
-4.40532744e-01 -9.50611711e-01 -6.87667727e-01 3.47866148e-01
-6.32338226e-01 -1.27194412e-02 6.04342997e-01 5.53986073e-01
1.33671954e-01 1.39863026e+00 -2.32457906e-01 -3.93343031e-01
3.78721148e-01 -4.65418339e-01 4.35178012e-01 -6.51159063e-02
-1.66054502e-01 -1.71776727e-01 1.57791674e-02 -7.95199335e-01
-9.19654191e-01 -5.84901929e-01 -8.56399715e-01 -7.87331343e-01
1.64455548e-02 1.47011459e-01 5.68407178e-01 5.61830223e-01
1.75585106e-01 3.98798764e-01 6.30064309e-01 -1.20266974e+00
-9.44461524e-02 -7.43706405e-01 -5.67315757e-01 7.03303695e-01
2.39872381e-01 -8.07866275e-01 -4.16183054e-01 4.53510702e-01] | [11.6927490234375, -1.8765909671783447] |
fbd069c9-a9a3-46da-bea2-de6932fe012a | an-attempt-towards-interpretable-audio-visual | 1812.02872 | null | http://arxiv.org/abs/1812.02872v1 | http://arxiv.org/pdf/1812.02872v1.pdf | An Attempt towards Interpretable Audio-Visual Video Captioning | Automatically generating a natural language sentence to describe the content
of an input video is a very challenging problem. It is an essential multimodal
task in which auditory and visual contents are equally important. Although
audio information has been exploited to improve video captioning in previous
works, it is usually regarded as an additional feature fed into a black box
fusion machine. How are the words in the generated sentences associated with
the auditory and visual modalities? The problem is still not investigated. In
this paper, we make the first attempt to design an interpretable audio-visual
video captioning network to discover the association between words in sentences
and audio-visual sequences. To achieve this, we propose a multimodal
convolutional neural network-based audio-visual video captioning framework and
introduce a modality-aware module for exploring modality selection during
sentence generation. Besides, we collect new audio captioning and visual
captioning datasets for further exploring the interactions between auditory and
visual modalities for high-level video understanding. Extensive experiments
demonstrate that the modality-aware module makes our model interpretable on
modality selection during sentence generation. Even with the added
interpretability, our video captioning network can still achieve comparable
performance with recent state-of-the-art methods. | ['Justin Goodman', 'Marc Moore', 'Yapeng Tian', 'Chenliang Xu', 'Chenxiao Guan'] | 2018-12-07 | null | null | null | null | ['audio-captioning', 'audio-visual-video-captioning'] | ['audio', 'computer-vision'] | [ 5.83184123e-01 1.24164127e-01 -1.10854916e-01 -3.59396726e-01
-9.04077947e-01 -3.24330568e-01 6.10128939e-01 6.99704839e-03
-1.72773868e-01 6.36908352e-01 7.27270067e-01 -1.48049325e-01
3.39383036e-01 -3.24315488e-01 -1.05831718e+00 -4.46436703e-01
3.71261269e-01 5.90360537e-02 -7.39208516e-03 -2.17558220e-01
1.08051449e-01 -8.71232376e-02 -1.97144878e+00 1.06629515e+00
5.48113763e-01 1.11278582e+00 5.30527711e-01 9.09128308e-01
-3.88854384e-01 1.05542922e+00 -5.09838343e-01 -3.35522860e-01
-3.11423182e-01 -8.50139081e-01 -9.23206925e-01 3.46316129e-01
4.07970786e-01 -3.06452632e-01 -3.12644750e-01 6.87211990e-01
5.22588074e-01 7.49279782e-02 5.45167565e-01 -1.71725869e+00
-8.53495657e-01 9.13825154e-01 -3.17860872e-01 3.33180055e-02
9.30105329e-01 2.93203503e-01 1.05983329e+00 -9.36364233e-01
4.99053717e-01 1.30291402e+00 7.30086267e-02 7.29284346e-01
-8.68186831e-01 -4.39702034e-01 3.88753921e-01 7.82349765e-01
-1.31017828e+00 -5.79620957e-01 1.02833498e+00 -4.18509424e-01
6.94065094e-01 5.21796644e-01 7.56509304e-01 1.49377191e+00
-1.88635945e-01 1.18080914e+00 4.45181310e-01 -4.23482418e-01
-1.63616184e-02 1.49692729e-01 -2.84708023e-01 4.75760460e-01
-3.41475695e-01 -3.27555686e-01 -9.10779119e-01 1.13613203e-01
5.40997982e-01 -1.21154428e-01 -6.62970126e-01 -7.80924559e-02
-1.63191056e+00 7.47467518e-01 3.10975343e-01 2.81893402e-01
-3.94256085e-01 4.02216375e-01 4.87880886e-01 9.69791263e-02
1.69410467e-01 4.05318946e-01 -5.87209202e-02 -2.24498376e-01
-7.83262014e-01 -2.52698679e-02 3.80060852e-01 8.34368825e-01
5.29550910e-01 1.15106575e-01 -7.03227699e-01 7.54459500e-01
5.69630027e-01 5.24127424e-01 4.06151712e-01 -9.45445359e-01
6.34654284e-01 4.65528786e-01 2.15304699e-02 -1.04399478e+00
-2.30284750e-01 -8.67572874e-02 -8.13312709e-01 -3.28721136e-01
4.43832241e-02 -1.45223394e-01 -7.53534257e-01 1.78090060e+00
-7.98925478e-03 5.37329674e-01 1.72850549e-01 1.25080466e+00
1.43389976e+00 1.00190175e+00 3.74418408e-01 -3.65866661e-01
1.73042023e+00 -1.01421809e+00 -1.01452565e+00 -1.80587739e-01
2.03823373e-01 -8.04323256e-01 1.02991927e+00 7.36325085e-02
-1.06271100e+00 -8.15739214e-01 -8.80703807e-01 -1.40144601e-01
1.21945376e-02 1.75412476e-01 4.75514084e-01 1.43374220e-01
-1.03807735e+00 -1.35229260e-01 -3.70403469e-01 -3.16366494e-01
1.99553400e-01 7.41482750e-02 -5.31658173e-01 4.01881896e-02
-1.53435731e+00 6.13282621e-01 5.40876091e-01 6.10109158e-02
-1.01459146e+00 -4.47328746e-01 -1.17960691e+00 1.86464131e-01
3.95434558e-01 -1.10134804e+00 1.39478540e+00 -1.41161120e+00
-1.36479032e+00 4.96899635e-01 -5.80985904e-01 -4.39036101e-01
7.13855550e-02 1.23199239e-01 -4.05964434e-01 7.78068721e-01
-9.33074802e-02 1.26913774e+00 1.11555445e+00 -1.56093597e+00
-5.62731564e-01 3.93681675e-02 2.12700233e-01 5.21044672e-01
-5.21047473e-01 -2.22200751e-02 -6.39207602e-01 -7.24735379e-01
-1.25095427e-01 -6.90802276e-01 2.35713452e-01 -8.80068243e-02
-4.40104574e-01 -2.59917259e-01 7.56716967e-01 -7.41581261e-01
1.18687284e+00 -2.19477582e+00 3.17421049e-01 -3.30200315e-01
1.15815863e-01 1.12034418e-01 -4.35177714e-01 5.01512349e-01
-8.87795910e-02 2.46141538e-01 -9.87669304e-02 -4.95295465e-01
-3.32824327e-02 -8.49128291e-02 -5.35708904e-01 -5.72260283e-02
5.69593310e-01 1.04508257e+00 -9.91564870e-01 -7.18001842e-01
3.25105250e-01 8.79080355e-01 -4.69403058e-01 4.47776705e-01
-4.71174419e-01 7.24506915e-01 -3.62234503e-01 6.07573807e-01
3.25433224e-01 -3.12108517e-01 -1.21308088e-01 -5.58997214e-01
2.31418744e-01 -7.72574395e-02 -8.22237134e-01 1.85687125e+00
-5.34803152e-01 1.02830994e+00 -9.71194580e-02 -9.54209030e-01
7.80790627e-01 8.67307544e-01 4.27372336e-01 -6.58624589e-01
1.04902752e-01 -1.00719534e-01 -2.78256804e-01 -1.08821809e+00
6.91970408e-01 -1.86103597e-01 -2.36833423e-01 2.50682324e-01
1.96972519e-01 7.84789037e-04 1.05453841e-01 3.60422045e-01
7.63567328e-01 1.26639515e-01 5.86606599e-02 5.65877676e-01
8.08571458e-01 -1.25186369e-01 1.46819383e-01 5.21357000e-01
-1.00994267e-01 1.20925498e+00 4.28354949e-01 -1.95028067e-01
-8.86028349e-01 -9.68364775e-01 2.36315534e-01 1.35053349e+00
3.23213398e-01 -4.87994522e-01 -6.49651945e-01 -3.88148993e-01
-4.67555016e-01 7.13812768e-01 -5.33906221e-01 -2.37122089e-01
-2.15658337e-01 -2.40674227e-01 5.18732607e-01 5.82509458e-01
4.13596720e-01 -1.31251562e+00 -2.16752604e-01 1.27135560e-01
-1.05192733e+00 -1.54924631e+00 -5.71590900e-01 -3.69521230e-01
-3.45398933e-01 -9.08173740e-01 -8.39543939e-01 -9.64875400e-01
5.53441942e-01 3.87806088e-01 1.09502363e+00 2.84488034e-02
4.62421626e-02 7.63066709e-01 -7.57233858e-01 -4.74211037e-01
-5.44015288e-01 -5.06985513e-03 -1.69487715e-01 5.27476072e-01
2.99877942e-01 -2.94934839e-01 -6.21154189e-01 1.89992726e-01
-1.26961720e+00 6.97547436e-01 5.10629535e-01 7.17378676e-01
3.76357436e-01 -4.35086071e-01 6.85906708e-01 -1.28060296e-01
7.07430720e-01 -6.13987207e-01 1.19998544e-01 3.29280227e-01
3.35191309e-01 8.20477754e-02 6.43590868e-01 -6.23275161e-01
-1.05260873e+00 4.63843703e-01 -2.53378004e-01 -5.45035064e-01
-4.31126714e-01 6.18710756e-01 -3.83855700e-01 3.56275976e-01
2.84399778e-01 5.61136723e-01 1.46573976e-01 -2.32961953e-01
4.65951562e-01 1.04057014e+00 7.05889225e-01 -2.82919258e-01
5.67115307e-01 5.31123042e-01 -1.70102939e-01 -9.72418189e-01
-7.93720543e-01 -4.11448956e-01 -3.62940818e-01 -6.48896396e-01
1.36053765e+00 -1.32835579e+00 -8.39643419e-01 1.00219518e-01
-1.63001108e+00 1.83973476e-01 4.19522747e-02 4.90402132e-01
-8.03388238e-01 5.89070559e-01 -2.18234986e-01 -8.24654043e-01
-2.77063459e-01 -1.47296262e+00 1.50999415e+00 2.00717553e-01
-2.96824098e-01 -6.62247002e-01 -1.55600607e-01 8.41158628e-01
1.60029396e-01 1.27883434e-01 7.03318119e-01 -5.32175124e-01
-6.00303411e-01 1.21831317e-02 -3.90297294e-01 1.12380378e-01
-5.73616959e-02 -4.89127189e-02 -1.12717378e+00 5.41339368e-02
-2.19923660e-01 -3.60739291e-01 1.02910769e+00 2.33487874e-01
1.40379071e+00 -3.07193398e-01 -2.33087894e-02 3.92059594e-01
1.00007558e+00 1.97597250e-01 7.72905231e-01 1.77262455e-01
8.52362454e-01 6.60782218e-01 5.93889296e-01 4.08891261e-01
6.17242455e-01 8.73133659e-01 8.87205064e-01 -1.70144275e-01
-1.62002608e-01 -4.65314895e-01 6.08741879e-01 8.02044749e-01
1.46614909e-01 -6.78855836e-01 -6.52163088e-01 6.34652078e-01
-2.10538960e+00 -1.26968682e+00 -1.51447296e-01 1.78873241e+00
7.38085270e-01 -2.96252787e-01 1.06554843e-01 1.43843919e-01
9.26359355e-01 1.92450553e-01 -2.30955601e-01 -3.43803406e-01
-2.12516010e-01 -3.47553283e-01 -2.62679130e-01 4.09182340e-01
-1.20201623e+00 6.14163280e-01 5.37044334e+00 7.14170456e-01
-1.17947376e+00 3.85465920e-02 6.17511511e-01 -9.99291018e-02
-6.11860633e-01 -2.23649114e-01 -4.78589058e-01 6.00079119e-01
1.03628421e+00 4.42844182e-02 2.42700711e-01 4.43483502e-01
5.68886161e-01 1.98494028e-02 -1.23363388e+00 1.38416815e+00
5.90475976e-01 -1.39578199e+00 6.09943032e-01 -4.68254238e-01
3.79139125e-01 -4.20112640e-01 6.86659440e-02 1.76596075e-01
-5.85890174e-01 -1.26271796e+00 8.50639164e-01 7.47363508e-01
8.05341601e-01 -5.39067328e-01 8.11995149e-01 8.33830237e-02
-1.35431552e+00 -1.80451609e-02 -5.81702031e-02 -5.29872738e-02
5.78639150e-01 2.88014472e-01 -9.77653146e-01 5.13971865e-01
5.39348781e-01 6.38653100e-01 -6.47300363e-01 1.07974243e+00
-2.56053478e-01 5.13688445e-01 5.75865656e-02 -7.36762881e-02
1.61163226e-01 3.85449201e-01 7.22464144e-01 1.20803750e+00
5.58865309e-01 -6.04782477e-02 6.53463453e-02 7.62534082e-01
-1.38869718e-01 1.88499913e-01 -7.00209856e-01 -3.51538658e-01
2.30394036e-01 1.13834000e+00 -5.05097985e-01 -3.67977113e-01
-5.34479856e-01 1.01051188e+00 -2.11805582e-01 4.51226950e-01
-1.06294262e+00 -2.42810458e-01 6.00190938e-01 1.02534324e-01
2.42501140e-01 -3.33135463e-02 6.72140419e-02 -1.37172496e+00
7.90923610e-02 -8.32562149e-01 1.60695240e-01 -1.38151455e+00
-1.05190730e+00 7.81372070e-01 3.66474204e-02 -1.56805360e+00
-5.03791988e-01 -4.21327174e-01 -5.54751396e-01 6.40297830e-01
-1.31377590e+00 -1.29307425e+00 -5.55520296e-01 6.39949262e-01
9.85369146e-01 -2.36450747e-01 6.52435958e-01 1.75700054e-01
-2.76371568e-01 4.06053454e-01 -3.38506967e-01 1.69009924e-01
7.92185068e-01 -8.68148744e-01 3.72113585e-02 7.37554550e-01
3.51436049e-01 2.34855428e-01 9.70369637e-01 -4.24356699e-01
-1.54509556e+00 -1.22298551e+00 8.26580942e-01 -3.21390867e-01
4.83198404e-01 -3.16530645e-01 -8.66042793e-01 4.21543360e-01
7.79921532e-01 -3.78088444e-01 9.83884394e-01 -4.36362773e-01
-1.73808858e-01 7.37682357e-02 -5.73728144e-01 7.12376475e-01
8.80318999e-01 -7.79805183e-01 -7.04063356e-01 2.76273727e-01
1.25961745e+00 -7.42269531e-02 -5.21003842e-01 4.21371639e-01
3.77577335e-01 -7.38162160e-01 1.15303016e+00 -5.68635046e-01
9.49710786e-01 -6.22083068e-01 -9.70448554e-02 -1.28293252e+00
6.15981920e-03 -6.16992176e-01 -1.70631006e-01 1.48418891e+00
4.59630132e-01 8.23907703e-02 4.55101162e-01 2.02570900e-01
-1.33066237e-01 -3.09697628e-01 -8.68099630e-01 -2.66751707e-01
-5.44855237e-01 -7.40066469e-01 6.06547117e-01 7.21331537e-01
2.44287476e-01 8.78614962e-01 -8.13867152e-01 7.51965493e-02
2.96341568e-01 1.31131783e-01 6.79332197e-01 -9.83547866e-01
-1.50150552e-01 -3.09966743e-01 -4.50617105e-01 -1.07655048e+00
3.81017596e-01 -9.63476360e-01 2.55499125e-01 -1.87904418e+00
3.94104868e-01 4.93074983e-01 -1.93279251e-01 5.14043510e-01
-3.74921799e-01 5.24791181e-01 5.38847983e-01 3.85367870e-02
-1.01948106e+00 8.75576556e-01 1.35952199e+00 -4.37100112e-01
-1.74376726e-01 -2.34767869e-01 -9.40398037e-01 5.18266559e-01
5.55232286e-01 -6.82907701e-02 -5.67436934e-01 -6.58462763e-01
3.99298698e-01 5.63229978e-01 6.72989011e-01 -8.61068726e-01
1.87091514e-01 -2.42594913e-01 3.30750257e-01 -6.02127612e-01
6.95755184e-01 -8.54655266e-01 1.36202425e-01 6.54719099e-02
-7.14555383e-01 -7.00859027e-03 2.22369522e-01 6.48884118e-01
-7.25620806e-01 -1.05586961e-01 4.26826388e-01 -1.20600305e-01
-9.57706571e-01 1.31799936e-01 -7.71917999e-01 -1.84553698e-01
1.03837848e+00 -1.87833473e-01 -3.28242332e-01 -1.10729599e+00
-8.84600282e-01 3.27401787e-01 1.66661337e-01 8.65051150e-01
1.08518660e+00 -1.62053561e+00 -8.84386480e-01 -1.75986346e-02
4.66772228e-01 -2.73330808e-01 5.01399398e-01 7.01207995e-01
-2.70954281e-01 5.05477071e-01 -1.79227650e-01 -9.10493612e-01
-1.42276251e+00 5.51698208e-01 -5.24807274e-02 2.24750116e-01
-1.63740218e-01 7.14069545e-01 3.47511679e-01 1.18410848e-01
3.66456926e-01 -9.74407941e-02 -6.43308103e-01 3.37510198e-01
8.38542640e-01 -1.26290694e-01 -3.43636930e-01 -1.03844690e+00
-2.16870353e-01 3.69634718e-01 3.03233802e-01 -1.41346201e-01
1.12343431e+00 -6.08484924e-01 -2.53846589e-02 5.09082496e-01
1.37244928e+00 -4.98414069e-01 -9.71888959e-01 3.08589246e-02
-4.09259439e-01 -2.11161748e-01 -8.67380872e-02 -5.42196572e-01
-1.04028666e+00 1.30507255e+00 4.42399591e-01 3.02714378e-01
1.36756432e+00 1.75215378e-01 9.06876683e-01 1.62740022e-01
-5.25998399e-02 -7.68361866e-01 4.96734500e-01 4.39236164e-01
1.23421907e+00 -1.53777134e+00 -4.28362578e-01 -4.23275858e-01
-1.11039019e+00 1.24528873e+00 8.21127474e-01 5.41854322e-01
5.51346801e-02 -7.96588659e-02 5.13123684e-02 9.21790749e-02
-8.57421994e-01 -3.91155601e-01 5.84512174e-01 6.28622711e-01
4.56095338e-01 -2.25716963e-01 -5.76115288e-02 7.65399337e-01
-6.31816313e-02 5.22621721e-03 6.44260705e-01 6.06110156e-01
-5.29488087e-01 -9.23545957e-01 -6.21554196e-01 1.51634350e-01
-3.86647373e-01 -2.05963910e-01 -5.62037528e-01 2.20780551e-01
-8.85943174e-02 1.24714518e+00 2.11299449e-01 -6.16017103e-01
3.57059948e-02 2.74794489e-01 1.85463697e-01 -5.66562176e-01
-4.30959344e-01 1.09064914e-01 -1.82359051e-02 -3.84492278e-01
-7.44230449e-01 -3.22913855e-01 -1.26126683e+00 1.72524542e-01
-3.44277918e-02 2.62582272e-01 6.22494519e-01 1.11257470e+00
5.18449783e-01 7.51708448e-01 5.14774442e-01 -1.04986560e+00
7.96528980e-02 -8.42944384e-01 -2.83726528e-02 6.20934129e-01
5.67996502e-01 -4.01000232e-01 -3.58733952e-01 5.01474023e-01] | [10.658875465393066, 0.85406494140625] |
bf83ab4b-d473-4f6c-aa4d-7d25a35d29d0 | the-search-for-agreement-on-logical-fallacy | null | null | https://aclanthology.org/2022.lrec-1.471 | https://aclanthology.org/2022.lrec-1.471.pdf | The Search for Agreement on Logical Fallacy Annotation of an Infodemic | We evaluate an annotation schema for labeling logical fallacy types, originally developed for a crowd-sourcing annotation paradigm, now using an annotation paradigm of two trained linguist annotators. We apply the schema to a variety of different genres of text relating to the COVID-19 pandemic. Our linguist (as opposed to crowd-sourced) annotation of logical fallacies allows us to evaluate whether the annotation schema category labels are sufficiently clear and non-overlapping for both manual and, later, system assignment. We report inter-annotator agreement results over two annotation phases as well as a preliminary assessment of the corpus for training and testing a machine learning algorithm (Pattern-Exploiting Training) for fallacy detection and recognition. The agreement results and system performance underscore the challenging nature of this annotation task and suggest that the annotation schema and paradigm must be iteratively evaluated and refined in order to arrive at a set of annotation labels that can be reproduced by human annotators and, in turn, provide reliable training data for automatic detection and recognition systems. | ['Clare Voss', 'Peter Sutor', 'Douglas Summers-Stay', 'Jeffrey Micher', 'Stephanie M. Lukin', 'Taylor Hudson', 'Austin Blodgett', 'Claire Bonial'] | null | null | null | null | lrec-2022-6 | ['logical-fallacies'] | ['miscellaneous'] | [ 4.50526804e-01 4.37265486e-01 -4.33219178e-03 -6.68636620e-01
-1.01337898e+00 -1.07166231e+00 7.62092769e-01 7.56362259e-01
-4.07975703e-01 7.45934308e-01 4.68920261e-01 -4.32444006e-01
-7.23331794e-02 -2.89337426e-01 -2.16185689e-01 -3.37866902e-01
3.29341769e-01 1.21209788e+00 3.77667695e-01 -1.69328436e-01
1.85855597e-01 1.92791149e-01 -1.51441622e+00 6.86675608e-01
8.51769388e-01 8.07360351e-01 5.74464444e-04 4.34857965e-01
-2.06463888e-01 1.14569223e+00 -7.53821611e-01 -6.35525346e-01
2.02466697e-01 -5.19169688e-01 -1.27514184e+00 1.27230361e-01
2.93443829e-01 1.79328427e-01 6.59460545e-01 9.53874588e-01
4.66552019e-01 -1.58093423e-01 7.80959129e-01 -1.00815153e+00
-3.66780013e-01 5.66757917e-01 -6.88803345e-02 1.56974792e-01
9.92526531e-01 9.47974995e-02 1.01198971e+00 -6.60900295e-01
1.13330424e+00 1.10910213e+00 1.14278567e+00 4.28710341e-01
-1.10643804e+00 -4.17634815e-01 -2.02972814e-01 1.01055056e-01
-1.48804855e+00 -4.24932778e-01 1.72546610e-01 -1.18586969e+00
1.00331366e+00 1.70716539e-01 3.54032159e-01 8.78220320e-01
-2.75623292e-01 2.52478540e-01 1.22334802e+00 -9.31960464e-01
4.64045286e-01 6.52499676e-01 2.45769009e-01 5.90445876e-01
5.96600533e-01 -3.50728512e-01 -4.70393956e-01 -7.40697384e-01
1.46511585e-01 -6.51502371e-01 -1.63771152e-01 -8.89743268e-02
-9.39612269e-01 7.71508455e-01 -8.11558738e-02 8.07316601e-01
-5.16380072e-01 -3.26717466e-01 7.26217628e-01 2.95174159e-02
4.56703067e-01 6.49382830e-01 -4.39724207e-01 -1.62761286e-01
-9.34481740e-01 2.29840785e-01 1.14274025e+00 6.80374026e-01
5.56385517e-01 -4.25719857e-01 -4.69664276e-01 8.70497704e-01
3.36666137e-01 3.77808303e-01 2.73360610e-01 -1.05040359e+00
4.42367196e-01 1.01170790e+00 6.64367676e-01 -1.04634476e+00
-5.36067307e-01 6.13052696e-02 -1.71043620e-01 1.58535153e-01
8.37083220e-01 -2.37669423e-02 -2.97560126e-01 1.41446638e+00
4.76098239e-01 -3.70131165e-01 4.53535840e-02 6.89680934e-01
5.59955060e-01 1.20016243e-02 4.89269704e-01 -3.62408221e-01
1.95010591e+00 -3.76240224e-01 -8.68074656e-01 1.02586284e-01
1.27684939e+00 -8.05348754e-01 9.23064291e-01 5.46068311e-01
-8.29114676e-01 -2.67758876e-01 -8.42525363e-01 5.09290844e-02
-2.40002483e-01 3.64222601e-02 1.07481763e-01 1.04834831e+00
-7.67286241e-01 1.65755272e-01 -6.89161360e-01 -8.66231322e-01
3.03258032e-01 2.10225955e-02 -4.25869554e-01 4.60322015e-02
-1.23538136e+00 1.30204475e+00 5.46762347e-01 -9.71447527e-02
-6.29721880e-01 -1.87900797e-01 -6.64580703e-01 -2.07712963e-01
3.94494742e-01 -4.22330678e-01 1.41678071e+00 -1.10804200e+00
-8.95562112e-01 1.51506746e+00 -2.56134808e-01 -3.99745673e-01
5.28596222e-01 1.80844903e-01 -5.36962569e-01 2.52044648e-01
5.54485500e-01 1.26156196e-01 2.01853737e-01 -1.36014855e+00
-9.97414470e-01 -2.88469821e-01 -1.48153573e-01 8.92098993e-02
-1.98240712e-01 8.49176943e-01 6.88428581e-02 -2.58662134e-01
-1.37552366e-01 -7.34999061e-01 1.91529654e-02 -3.16112906e-01
-2.40460262e-01 -6.51714861e-01 4.29208040e-01 -8.89922678e-01
1.32373381e+00 -1.49780607e+00 -3.29280645e-01 2.64380157e-01
2.76934087e-01 4.05839831e-01 1.89702347e-01 5.04924119e-01
-3.65221165e-02 3.78759950e-01 -2.78301328e-01 -2.39219487e-01
3.81392948e-02 2.40159571e-01 -8.27286392e-03 5.37571132e-01
7.02389181e-02 4.28940833e-01 -1.28648198e+00 -9.15250659e-01
-2.37151995e-01 1.63425177e-01 -5.07685304e-01 2.60493338e-01
-3.08860898e-01 6.82215393e-01 -4.82458383e-01 5.29235959e-01
2.49998197e-01 -2.51070231e-01 7.97636986e-01 -1.04235142e-01
-5.23317009e-02 4.52455014e-01 -1.10611951e+00 1.29349482e+00
-8.60064849e-02 4.01558906e-01 -2.18147449e-02 -7.54758835e-01
1.15112936e+00 8.10189009e-01 4.42880809e-01 -5.32865882e-01
2.43964240e-01 7.12630272e-01 -4.42461446e-02 -1.11012018e+00
4.33367729e-01 -2.53351748e-01 -3.57168019e-01 1.07886147e+00
8.28444436e-02 -9.25686285e-02 4.84265804e-01 -2.61164531e-02
9.46574986e-01 1.91089615e-01 7.43875742e-01 -5.99422395e-01
7.94179976e-01 5.78252137e-01 7.38960385e-01 7.09896028e-01
-4.87864822e-01 3.29565048e-01 4.80879217e-01 -8.60469282e-01
-1.17262578e+00 -2.03644201e-01 -3.62552106e-01 1.18973207e+00
-3.75713795e-01 -3.60814631e-01 -1.18386686e+00 -9.42083061e-01
-3.89214844e-01 7.22784042e-01 -6.90889180e-01 2.73508012e-01
-5.06162047e-01 -7.33092070e-01 8.85474980e-01 2.95110375e-01
2.89009549e-02 -1.08631527e+00 -1.00366187e+00 4.59418714e-01
-4.96689856e-01 -1.19420242e+00 -1.84035778e-01 1.86430007e-01
-6.52088374e-02 -1.68599164e+00 -2.51747668e-01 -8.01931262e-01
5.76139152e-01 -3.90202492e-01 1.19290686e+00 3.85653883e-01
-3.11804153e-02 3.56905043e-01 -5.91090143e-01 -4.90349591e-01
-1.04364479e+00 -7.06684962e-02 3.35797995e-01 -8.22957978e-02
1.04323542e+00 1.38022183e-02 -1.43520862e-01 5.28480470e-01
-8.84426773e-01 -2.37653598e-01 -1.75398573e-01 5.75807512e-01
2.38105446e-01 -2.32967392e-01 4.67305869e-01 -1.12805426e+00
8.02372098e-01 -4.02201444e-01 -5.91398716e-01 6.66366458e-01
-6.04780614e-01 -1.01602681e-01 2.60067523e-01 -2.08251894e-01
-1.20420885e+00 3.48020755e-02 -9.96751636e-02 2.21462920e-01
-5.74185431e-01 5.24138808e-01 3.46172489e-02 1.57494426e-01
1.33264625e+00 -3.27410221e-01 -8.50591511e-02 -5.28910637e-01
1.86737657e-01 1.22958291e+00 5.33129394e-01 -7.63202846e-01
4.52087253e-01 2.94697762e-01 -4.88360286e-01 -2.36491367e-01
-9.86633003e-01 -6.10354185e-01 -9.17977214e-01 -1.28249988e-01
9.99573767e-01 -8.31825674e-01 -5.69563806e-01 1.11213304e-01
-1.56861019e+00 -2.54508793e-01 -3.09154689e-01 2.78813511e-01
-4.42055672e-01 3.16336304e-01 -2.39329070e-01 -1.08358109e+00
-3.26560318e-01 -9.32573080e-01 9.74476039e-01 -1.96466267e-01
-1.02442586e+00 -1.17498004e+00 6.54014945e-01 8.07800174e-01
1.61485225e-01 3.78146738e-01 8.08457375e-01 -1.44392800e+00
3.08904469e-01 -4.72403616e-01 -1.40402749e-01 6.32878318e-02
9.59973857e-02 -1.84054017e-01 -1.10100937e+00 5.71707599e-02
1.54727340e-01 -6.86349988e-01 1.57982498e-01 -2.35871568e-01
2.48828322e-01 -4.89364117e-01 -3.84935141e-01 -2.36530811e-01
1.18439949e+00 3.25497717e-01 4.36495036e-01 5.72931230e-01
4.11982089e-01 1.12554431e+00 5.37000597e-01 3.17142844e-01
5.96330345e-01 9.55537558e-01 -2.76587605e-01 1.86565384e-01
3.69175449e-02 -1.51004821e-01 -9.09651164e-03 3.49469572e-01
-1.16497479e-01 -3.33737373e-01 -1.69998181e+00 4.32405323e-01
-1.98117399e+00 -9.25719619e-01 -3.54079753e-01 2.12825322e+00
1.02749455e+00 -1.06997311e-01 4.40646380e-01 3.08397114e-01
9.75365102e-01 -4.85250056e-01 1.34704918e-01 -5.92754006e-01
-4.18478660e-02 -2.03575358e-01 9.25577730e-02 6.45211697e-01
-9.15630460e-01 8.29558432e-01 7.38897419e+00 5.84265828e-01
-5.82234800e-01 6.12757027e-01 5.38843930e-01 4.58611876e-01
-2.57631779e-01 -1.40093891e-02 -7.73916125e-01 3.58865291e-01
1.01347792e+00 3.27072702e-02 1.28831357e-01 5.12551546e-01
3.33717734e-01 -2.18441725e-01 -1.14471400e+00 4.87009227e-01
2.82645971e-01 -1.29032648e+00 -2.85043657e-01 -1.05089016e-01
7.67550826e-01 -4.31764238e-02 -9.05239940e-01 9.94754583e-02
6.56840026e-01 -9.10048246e-01 1.07359397e+00 4.47051704e-01
8.86873960e-01 -2.11004376e-01 1.21742332e+00 3.46232831e-01
-9.78424370e-01 -5.30072153e-02 1.09408788e-01 -1.42258286e-01
4.51648146e-01 3.66707027e-01 -1.12267637e+00 4.92929012e-01
6.08916640e-01 1.98008537e-01 -5.84353924e-01 8.12876999e-01
-2.60636717e-01 5.50813675e-01 -3.11145842e-01 -2.42853463e-01
9.09508690e-02 2.35700801e-01 4.40048277e-01 1.49308789e+00
-7.85561949e-02 2.54758775e-01 5.38215101e-01 7.98374653e-01
2.46473506e-01 3.21946442e-01 -3.95768523e-01 2.95422953e-02
7.19684482e-01 1.14755690e+00 -8.99812400e-01 -5.05134165e-01
-2.20000520e-01 3.80965352e-01 4.77951646e-01 2.40655541e-01
-4.10648167e-01 -1.38115853e-01 -2.32230481e-02 3.41211557e-01
-5.86363785e-02 2.12268934e-01 -6.24400914e-01 -7.58051872e-01
-8.26824009e-02 -1.12113345e+00 7.44149804e-01 -7.21790969e-01
-1.19165289e+00 9.68928277e-01 2.16437593e-01 -1.19254553e+00
-6.52779043e-01 -5.34797192e-01 -4.68152642e-01 8.56203735e-01
-9.07579243e-01 -1.25078261e+00 -3.83115560e-01 1.88601390e-01
-2.81485226e-02 -2.68155843e-01 1.21531332e+00 1.98261067e-01
-4.27348584e-01 4.06893641e-01 -3.19717407e-01 1.33460134e-01
6.78912878e-01 -9.70373154e-01 -3.36883664e-02 6.81859195e-01
-4.01596054e-02 4.32915717e-01 9.54502523e-01 -8.48593950e-01
-7.02791631e-01 -9.77787733e-01 1.62439466e+00 -1.05757415e+00
7.46841967e-01 -3.38286221e-01 -1.05038369e+00 5.23094594e-01
8.90473928e-03 -2.09388465e-01 1.05459797e+00 2.34548718e-01
-6.42288148e-01 3.53333503e-01 -1.49101675e+00 8.95298049e-02
8.50705683e-01 -6.42149806e-01 -9.50427115e-01 6.82415366e-01
2.50808477e-01 -1.32301331e-01 -8.46479952e-01 1.01398557e-01
4.04777646e-01 -7.97489345e-01 3.73749644e-01 -7.00956941e-01
7.12988079e-02 -5.50603628e-01 -2.37234294e-01 -7.03090191e-01
-1.87477037e-01 -5.66842258e-01 4.65380400e-01 1.41474736e+00
5.81450343e-01 -4.98215824e-01 3.14198881e-01 9.28347945e-01
4.32360508e-02 -2.96055913e-01 -9.69381213e-01 -6.17770374e-01
-6.77250251e-02 -5.91822326e-01 5.24386823e-01 1.31490946e+00
6.73689723e-01 3.58816862e-01 -2.87433993e-02 1.03366924e-02
1.83305442e-01 -3.32125753e-01 6.71390355e-01 -1.49445546e+00
-1.15600796e-02 -4.73055035e-01 -4.63105857e-01 1.55172963e-02
8.54429826e-02 -9.69755530e-01 2.60676026e-01 -1.54073846e+00
3.32471788e-01 -7.88052857e-01 1.77214220e-01 8.84002090e-01
-1.53997838e-01 4.63851571e-01 -6.42180890e-02 7.68047273e-01
-9.02557313e-01 -1.96154088e-01 4.89149898e-01 1.21045239e-01
-1.96732849e-01 -4.16339993e-01 -8.56506288e-01 8.44402254e-01
5.60018778e-01 -9.82023776e-01 -2.56716111e-03 -3.05266619e-01
6.77174985e-01 -3.48336190e-01 2.13034421e-01 -7.84365356e-01
3.09936881e-01 -1.91824779e-01 -7.53457323e-02 -6.10192418e-02
-4.11370099e-01 -6.91939533e-01 4.09464121e-01 6.91848755e-01
-4.64813054e-01 -2.54838943e-01 6.44374713e-02 2.40879700e-01
5.90403005e-02 -7.47207284e-01 7.00896144e-01 -1.68143228e-01
-2.87794024e-01 -3.01394969e-01 -6.82178617e-01 3.26743752e-01
1.04263055e+00 -2.82080770e-01 -4.69505817e-01 -1.58883959e-01
-7.52058923e-01 2.24638849e-01 6.93984210e-01 9.09385011e-02
-2.72608936e-01 -1.19757938e+00 -1.04807937e+00 -1.33990645e-01
5.05009830e-01 -2.24582613e-01 -1.20848678e-01 9.92418051e-01
-7.62245536e-01 5.05378008e-01 -7.87908807e-02 -6.91785991e-01
-1.02421927e+00 6.69199824e-01 4.02103543e-01 -5.61424375e-01
-3.23115110e-01 5.32750309e-01 -1.65523678e-01 -5.51824093e-01
1.82899728e-01 -6.58832192e-02 -6.27173662e-01 3.63235205e-01
8.38236451e-01 3.27690661e-01 2.01592460e-01 -1.00950801e+00
-6.30484104e-01 3.51618707e-01 2.27691025e-01 -2.74661571e-01
1.01445222e+00 -1.24897800e-01 -1.96929350e-01 4.61302698e-01
5.89215040e-01 3.14146578e-02 -5.74992180e-01 -3.82536829e-01
6.19320929e-01 -9.93146226e-02 -4.33708221e-01 -1.00736964e+00
-1.80249885e-01 4.99042213e-01 5.60710549e-01 5.89113593e-01
5.76110899e-01 1.05797812e-01 1.83416829e-01 3.44930470e-01
5.47825336e-01 -1.25487947e+00 -1.97235301e-01 4.82384503e-01
9.08719420e-01 -1.14953148e+00 -8.13124329e-02 -4.55213875e-01
-9.12777483e-01 9.24585819e-01 3.58246148e-01 3.20517689e-01
3.59635860e-01 4.14374679e-01 4.62535590e-01 -4.95445222e-01
-5.93604624e-01 -1.00495122e-01 4.01473403e-01 9.05161738e-01
7.96880960e-01 4.11863774e-02 -7.86376178e-01 8.54813695e-01
-6.14770055e-02 2.72535920e-01 3.95227313e-01 9.24250782e-01
-5.97388685e-01 -1.16250110e+00 -7.28294969e-01 2.57382631e-01
-4.90928024e-01 5.57074435e-02 -7.04795778e-01 4.87751216e-01
7.52575278e-01 1.17713833e+00 1.03629164e-01 -1.19855031e-01
4.10226047e-01 4.96697962e-01 2.74244070e-01 -9.64356899e-01
-1.15847862e+00 -3.78980972e-02 8.88330936e-01 -2.95425385e-01
-9.12627518e-01 -9.17949677e-01 -1.03929138e+00 9.37383398e-02
-3.54253441e-01 6.07649684e-01 3.96976680e-01 1.46122813e+00
9.55014378e-02 -2.55996380e-02 2.33880520e-01 -5.87216437e-01
-3.13157439e-01 -1.30695450e+00 -2.03486964e-01 8.55228603e-01
-1.99977800e-01 -7.00418890e-01 -2.01116711e-01 5.58382928e-01] | [9.71573257446289, 4.965119361877441] |
e7d4f7d5-560b-4f68-9b3b-ceb8c9d08bf0 | omni-detr-omni-supervised-object-detection | 2203.16089 | null | https://arxiv.org/abs/2203.16089v1 | https://arxiv.org/pdf/2203.16089v1.pdf | Omni-DETR: Omni-Supervised Object Detection with Transformers | We consider the problem of omni-supervised object detection, which can use unlabeled, fully labeled and weakly labeled annotations, such as image tags, counts, points, etc., for object detection. This is enabled by a unified architecture, Omni-DETR, based on the recent progress on student-teacher framework and end-to-end transformer based object detection. Under this unified architecture, different types of weak labels can be leveraged to generate accurate pseudo labels, by a bipartite matching based filtering mechanism, for the model to learn. In the experiments, Omni-DETR has achieved state-of-the-art results on multiple datasets and settings. And we have found that weak annotations can help to improve detection performance and a mixture of them can achieve a better trade-off between annotation cost and accuracy than the standard complete annotation. These findings could encourage larger object detection datasets with mixture annotations. The code is available at https://github.com/amazon-research/omni-detr. | ['Stefano Soatto', 'Bernt Schiele', 'Nuno Vasconcelos', 'Gurumurthy Swaminathan', 'Hao Yang', 'Zhaowei Cai', 'Pei Wang'] | 2022-03-30 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Wang_Omni-DETR_Omni-Supervised_Object_Detection_With_Transformers_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wang_Omni-DETR_Omni-Supervised_Object_Detection_With_Transformers_CVPR_2022_paper.pdf | cvpr-2022-1 | ['semi-supervised-object-detection'] | ['computer-vision'] | [-7.48497620e-02 1.46565273e-01 -4.65159148e-01 -5.99556088e-01
-1.09040427e+00 -6.43096805e-01 6.78261936e-01 1.28268480e-01
-5.58781683e-01 4.50577974e-01 7.24838376e-02 -9.34791565e-02
1.44836336e-01 -4.86514330e-01 -8.07683110e-01 -4.69701290e-01
1.00991661e-02 7.04564154e-01 4.82803762e-01 2.52465487e-01
-4.36712652e-02 -2.87705427e-03 -1.69028604e+00 3.49541515e-01
5.94707906e-01 8.09108198e-01 2.79222220e-01 6.55659080e-01
-4.63165455e-02 6.86122417e-01 -1.79578274e-01 -5.29266655e-01
4.62804914e-01 1.13907158e-01 -6.94937229e-01 1.91919040e-02
8.12187731e-01 -2.90059477e-01 -2.90070087e-01 1.13146257e+00
5.71366787e-01 -1.68254480e-01 7.63746858e-01 -1.28515160e+00
-7.01558948e-01 9.42934513e-01 -1.08751762e+00 1.51856989e-01
1.48028880e-01 4.50298756e-01 1.21854663e+00 -1.25234425e+00
3.62756610e-01 1.38183594e+00 6.69674993e-01 6.68124795e-01
-1.38929462e+00 -9.89756465e-01 1.13211907e-01 4.45512906e-02
-1.50142217e+00 -2.79469371e-01 3.40141743e-01 -5.28516889e-01
3.79752338e-01 2.70540893e-01 4.27668333e-01 1.01075065e+00
-4.33762789e-01 1.28110838e+00 1.12893283e+00 -5.88283777e-01
-1.34907022e-01 5.80144405e-01 4.33654457e-01 8.74173939e-01
3.85820776e-01 1.12202339e-01 -5.64390242e-01 -1.72236696e-01
3.72111857e-01 1.14394113e-01 -2.66523939e-02 -5.00838518e-01
-1.09967422e+00 7.17178047e-01 6.88909113e-01 2.06520855e-01
9.68207270e-02 2.52780944e-01 2.73100019e-01 5.75602278e-02
6.20865405e-01 4.79080319e-01 -5.43229938e-01 1.15957454e-01
-9.23111558e-01 7.30213076e-02 5.86592972e-01 1.15326428e+00
8.66242886e-01 -3.24214160e-01 -6.05573714e-01 9.77409899e-01
5.83973169e-01 4.15480912e-01 3.00489783e-01 -7.96406269e-01
1.81748495e-01 7.14198709e-01 2.66204178e-01 -2.93276399e-01
-3.79898727e-01 -7.52744794e-01 -1.99615300e-01 2.13313147e-01
6.96505189e-01 -2.07531452e-01 -1.22206473e+00 2.01877022e+00
6.62334681e-01 5.65909326e-01 -3.75476509e-01 1.04767263e+00
8.51606965e-01 1.87939659e-01 3.56087178e-01 4.11146283e-01
1.78657401e+00 -1.28669453e+00 -2.54510850e-01 -2.83623099e-01
9.44253147e-01 -7.22741902e-01 1.22015655e+00 1.83171228e-01
-9.55697000e-01 -6.31253064e-01 -9.70807910e-01 5.96383493e-03
-3.97990286e-01 6.13124430e-01 7.40526497e-01 6.58085108e-01
-8.92164886e-01 1.83537513e-01 -9.67993855e-01 -4.15481359e-01
8.87788713e-01 2.06680119e-01 -2.27397904e-01 -2.60989130e-01
-7.74262011e-01 6.78823650e-01 4.71199006e-01 -2.23305672e-01
-1.31349349e+00 -8.95122111e-01 -5.00255823e-01 7.91161060e-02
5.85901678e-01 -8.16514373e-01 1.46939778e+00 -8.16164613e-01
-8.10955882e-01 1.28887928e+00 6.38376698e-02 -5.40501773e-01
6.92331910e-01 -2.63202012e-01 1.14594430e-01 4.46274765e-02
4.00767058e-01 1.17895019e+00 5.92446625e-01 -1.21527624e+00
-1.19428086e+00 -6.13791168e-01 1.91521585e-01 7.07301646e-02
-5.73837757e-01 3.73516195e-02 -6.76422715e-01 -3.56473982e-01
-1.10629685e-01 -1.22137368e+00 -2.12022647e-01 2.03003243e-01
-4.40771133e-01 -7.96023488e-01 9.05586243e-01 -1.00720979e-01
7.57369339e-01 -2.11674809e+00 -3.36972773e-01 -1.69205651e-01
4.62428570e-01 2.97249168e-01 -3.10939670e-01 -4.71932068e-02
6.25732467e-02 9.19489283e-03 3.72651778e-02 -7.48652756e-01
3.82207446e-02 -9.29942057e-02 -2.38857679e-02 5.77958763e-01
2.98910409e-01 9.36981618e-01 -1.12186503e+00 -6.10137045e-01
1.70998216e-01 3.16692024e-01 -4.41368699e-01 2.87972599e-01
-3.29947382e-01 3.83523583e-01 -4.31954086e-01 8.74507844e-01
6.90748215e-01 -5.63866496e-01 -3.73080820e-02 -1.12237237e-01
1.01089478e-01 2.13199407e-01 -1.22032309e+00 1.65219915e+00
-3.50413799e-01 5.90908587e-01 2.01561570e-01 -1.01019096e+00
4.34105873e-01 2.41338789e-01 4.09483939e-01 -3.07754636e-01
1.20908514e-01 1.73932552e-01 -9.17753726e-02 -2.39654452e-01
2.59477019e-01 3.19730967e-01 6.69787526e-02 4.15775567e-01
5.03703356e-01 2.14548722e-01 4.59013343e-01 5.58991969e-01
1.07401621e+00 3.03966969e-01 -1.06757125e-02 -4.31375295e-01
2.11715728e-01 2.02916265e-01 4.93045688e-01 1.13950503e+00
-3.00265551e-01 5.65461218e-01 -9.43024457e-03 -1.35012791e-01
-9.45007563e-01 -1.00518167e+00 -4.29243654e-01 1.66738176e+00
1.44114614e-01 -2.84812182e-01 -6.11827195e-01 -9.12655890e-01
1.62845716e-01 5.07513285e-01 -5.31211138e-01 -1.84746623e-01
-1.96868673e-01 -9.45302725e-01 6.63705766e-01 6.78671539e-01
5.52772917e-02 -8.16968620e-01 -4.58895087e-01 4.25701477e-02
-6.08441532e-02 -1.14529371e+00 -5.49356759e-01 4.53114122e-01
-6.43833816e-01 -9.89396155e-01 -6.52332187e-01 -6.59923136e-01
7.39644408e-01 7.72250652e-01 1.40031862e+00 2.39721537e-01
-6.35327339e-01 5.62794030e-01 -2.36910701e-01 -8.06896150e-01
-1.38250887e-01 2.06458747e-01 -2.43213214e-03 -1.74074583e-02
5.68160772e-01 -4.73400891e-01 -8.62976611e-01 5.59329450e-01
-8.64700735e-01 7.19961673e-02 6.09025478e-01 7.72998989e-01
4.17455137e-01 -5.56681693e-01 4.27623183e-01 -9.58591223e-01
4.35314961e-02 -7.03669250e-01 -6.95457995e-01 1.51221365e-01
-7.25707352e-01 5.25160991e-02 1.94551185e-01 -7.64600575e-01
-1.03194392e+00 3.23353469e-01 6.98981434e-02 -5.06353676e-01
-2.47786626e-01 1.25705585e-01 9.03409868e-02 -2.23363966e-01
9.39190149e-01 -1.45776004e-01 -3.12064856e-01 -6.85354292e-01
7.30461121e-01 7.53127933e-01 4.74240571e-01 -6.38553679e-01
7.79179275e-01 6.56700134e-01 -4.18081552e-01 -3.14145148e-01
-1.46373248e+00 -1.00656593e+00 -5.49747050e-01 -1.92549273e-01
6.85448945e-01 -1.40668344e+00 -5.14302433e-01 1.50948659e-01
-8.10049415e-01 -3.84013563e-01 -3.44689131e-01 4.66324419e-01
-3.01733673e-01 7.74486884e-02 -6.86682105e-01 -1.00386298e+00
-2.34894842e-01 -1.00843954e+00 1.31210792e+00 5.72456658e-01
-2.61398870e-02 -6.86057150e-01 6.18452253e-03 7.14718044e-01
2.45392069e-01 -1.45941108e-01 3.56983960e-01 -8.86736035e-01
-7.58933663e-01 -3.58625293e-01 -5.30024707e-01 1.27734363e-01
-2.31967479e-01 -8.09524134e-02 -1.24931538e+00 -4.77583051e-01
-5.18274128e-01 -8.63474011e-01 1.22057450e+00 2.22269282e-01
1.15772295e+00 -1.07682697e-01 -8.17341924e-01 3.97675842e-01
1.33085859e+00 -4.54337865e-01 6.22364953e-02 4.00325619e-02
9.70220149e-01 4.83320653e-01 7.22779512e-01 4.03110534e-01
6.17967486e-01 8.00039470e-01 6.42247319e-01 -3.06282520e-01
-5.62118828e-01 -3.32121462e-01 1.87365487e-01 1.92953318e-01
8.66644531e-02 -7.77159184e-02 -9.63504553e-01 8.80169034e-01
-2.06372237e+00 -8.53280067e-01 -2.82009602e-01 2.16583753e+00
1.04312980e+00 3.30231041e-01 5.46708405e-01 -2.16985315e-01
8.78792048e-01 -2.48913392e-01 -5.12753248e-01 3.06115478e-01
1.97264180e-01 -1.59886926e-01 8.15024972e-01 2.72436559e-01
-1.22135007e+00 9.93805170e-01 6.04897022e+00 1.16677392e+00
-9.35356259e-01 6.36390150e-01 7.44839132e-01 -3.37093502e-01
-8.61692280e-02 3.78152914e-02 -1.20730150e+00 4.74875212e-01
9.32792723e-01 2.32096508e-01 1.06220022e-01 1.17410171e+00
-1.57291800e-01 -8.11513290e-02 -1.15036750e+00 7.06310987e-01
-2.24358544e-01 -1.10776806e+00 -3.09989184e-01 -2.64409389e-02
8.45882356e-01 5.87228715e-01 5.06704487e-02 6.23821080e-01
9.09568131e-01 -5.49814165e-01 9.22943771e-01 1.88860983e-01
8.01839352e-01 -2.70437211e-01 5.65548003e-01 5.83571374e-01
-1.24081123e+00 -2.62293726e-01 -3.94865423e-01 2.03084081e-01
-4.13004570e-02 6.71081543e-01 -9.46579278e-01 2.48118460e-01
1.00774205e+00 4.62671846e-01 -9.29518461e-01 1.32863021e+00
-1.67014599e-01 1.05322862e+00 -7.15601385e-01 -1.67122688e-02
3.00682038e-01 1.10696480e-01 4.87137556e-01 1.51791310e+00
-8.66441149e-03 -1.83176056e-01 6.52783871e-01 9.62271810e-01
-3.10549289e-01 -1.14212602e-01 -4.35322225e-01 3.11842829e-01
5.94129860e-01 1.84578514e+00 -9.11433995e-01 -6.37135804e-01
-5.92611790e-01 4.40981239e-01 6.61698580e-01 1.03027798e-01
-9.59776282e-01 1.25405192e-01 4.15612131e-01 3.39248627e-01
2.83771902e-01 1.38033271e-01 -3.19634885e-01 -1.18719053e+00
-2.43629470e-01 -6.74177468e-01 7.68758476e-01 -6.79228544e-01
-1.46830952e+00 5.17163724e-02 1.77471906e-01 -1.13693643e+00
1.56060413e-01 -5.37758172e-01 -5.31408787e-01 4.57717896e-01
-1.47725999e+00 -1.42378354e+00 -4.46059287e-01 3.89999837e-01
4.83802319e-01 1.51142940e-01 4.58163798e-01 6.14320397e-01
-6.30470037e-01 7.83331692e-01 -1.21589288e-01 3.17818731e-01
8.96508992e-01 -1.49697709e+00 8.83599818e-02 7.64541984e-01
8.04612637e-01 3.25929075e-01 6.83888912e-01 -3.02166045e-01
-1.21808732e+00 -1.24326479e+00 3.47049713e-01 -1.00389230e+00
6.39588118e-01 -4.85756963e-01 -7.31441617e-01 7.76033521e-01
3.31884669e-03 6.04649067e-01 6.80538237e-01 3.49583834e-01
-7.17520833e-01 -8.52248296e-02 -1.20094562e+00 4.90449667e-01
1.33921647e+00 -3.77188265e-01 -2.34116450e-01 7.05195427e-01
7.36784458e-01 -3.55540812e-01 -6.29152715e-01 5.07234514e-01
5.41196704e-01 -7.55016148e-01 9.77231681e-01 -6.20946050e-01
1.42569110e-01 -4.98427987e-01 1.56435668e-02 -1.04779840e+00
-5.20313323e-01 -2.78588176e-01 -2.08935410e-01 1.42839587e+00
7.87124038e-01 -5.30881703e-01 9.29083288e-01 5.43239236e-01
-4.36414964e-02 -7.89902151e-01 -7.22041011e-01 -7.68479347e-01
-1.35784417e-01 -4.35828358e-01 3.33340734e-01 7.63114095e-01
-1.74428403e-01 6.18172824e-01 -2.79659063e-01 2.27678552e-01
8.94630492e-01 1.10432141e-01 9.13009048e-01 -1.40209901e+00
-3.40996653e-01 -4.23411578e-01 -4.03979927e-01 -1.09647751e+00
4.99845706e-02 -1.17600572e+00 1.50310874e-01 -1.27102220e+00
8.07294905e-01 -7.59184957e-01 -5.95450044e-01 8.03245604e-01
-5.17788470e-01 9.56256092e-01 3.30373228e-01 4.07670885e-01
-1.17969775e+00 3.12963456e-01 1.04109359e+00 -3.42125535e-01
-2.94400528e-02 1.27493888e-01 -8.92463267e-01 7.50062227e-01
6.99418128e-01 -9.79534388e-01 -9.79013145e-02 -4.10344332e-01
-1.09378554e-01 -1.24629140e-01 3.43168020e-01 -1.01191044e+00
1.84316486e-01 2.29310319e-01 3.43709916e-01 -5.38784623e-01
2.70883858e-01 -7.64604390e-01 -2.80281961e-01 4.79036093e-01
-4.92703199e-01 -3.67794216e-01 3.92300822e-02 7.49059260e-01
9.06766504e-02 -3.45870763e-01 9.78112280e-01 -1.65080339e-01
-5.10322452e-01 4.61431414e-01 9.63381976e-02 2.38001645e-01
1.07422173e+00 -1.00645967e-01 -3.49233061e-01 -1.35824606e-01
-6.56212389e-01 5.07851958e-01 2.35633090e-01 4.26747739e-01
3.50076007e-03 -1.36749089e+00 -9.43557322e-01 -1.77027255e-01
5.68705857e-01 -6.13040738e-02 -1.07790932e-01 1.03006732e+00
1.82335675e-01 2.47614458e-01 9.22713801e-02 -1.09456432e+00
-1.34960091e+00 5.11375844e-01 2.12540805e-01 -3.36024970e-01
-4.13724959e-01 1.05455506e+00 3.62627417e-01 -6.36142850e-01
4.68248099e-01 -3.29884514e-02 -6.27749637e-02 2.58113001e-03
5.77517927e-01 3.18321228e-01 -1.22337714e-01 -2.94204682e-01
-3.31451416e-01 3.32896680e-01 -1.68104410e-01 -2.28640400e-02
1.16210496e+00 -6.79277852e-02 2.69466519e-01 2.06494778e-01
8.90839577e-01 -1.06669776e-02 -1.40462899e+00 -5.49439609e-01
2.15982571e-01 -4.26782817e-01 6.92954808e-02 -7.58842707e-01
-1.13646281e+00 6.40288413e-01 8.89700413e-01 2.48514012e-01
6.14002764e-01 4.42082524e-01 5.11962056e-01 2.30888069e-01
3.94408256e-01 -8.06159019e-01 2.29773417e-01 1.31995708e-01
3.02312285e-01 -1.73967266e+00 -1.60376832e-01 -3.76047522e-01
-3.74360234e-01 5.22787988e-01 1.04385376e+00 -7.25056976e-02
5.97263575e-01 4.56942499e-01 1.14799723e-01 -2.03528002e-01
-8.60719621e-01 -7.35445201e-01 5.38487136e-01 3.90874028e-01
6.63964808e-01 1.92458510e-01 -2.62621492e-01 3.59155148e-01
2.40031675e-01 -7.18425810e-02 2.08940104e-01 7.32213438e-01
-5.52357078e-01 -9.38095689e-01 -5.47759652e-01 6.94983006e-01
-6.65201426e-01 4.31174971e-02 -2.48891771e-01 6.20225191e-01
2.75549114e-01 1.03215170e+00 7.11320573e-03 -7.98245519e-02
2.07218558e-01 1.01508372e-01 4.77476448e-01 -1.06212437e+00
-4.57418263e-01 -7.96506032e-02 5.23109455e-03 -5.62276959e-01
-4.25346017e-01 -4.43882793e-01 -1.12407815e+00 9.11473036e-02
-8.70489836e-01 1.15770273e-01 5.90873599e-01 6.61961079e-01
4.48012352e-01 3.38619858e-01 3.27956080e-01 -9.82161283e-01
-8.55401278e-01 -1.31515181e+00 -2.82817513e-01 6.10873938e-01
1.31361902e-01 -8.70385766e-01 -3.34008515e-01 -5.13956584e-02] | [9.285476684570312, 1.2763880491256714] |
beee0d21-ce80-4754-9e58-f25c69de2ad2 | deep-learning-based-spatio-temporal-facial | 2305.00552 | null | https://arxiv.org/abs/2305.00552v1 | https://arxiv.org/pdf/2305.00552v1.pdf | Deep Learning-based Spatio Temporal Facial Feature Visual Speech Recognition | In low-resource computing contexts, such as smartphones and other tiny devices, Both deep learning and machine learning are being used in a lot of identification systems. as authentication techniques. The transparent, contactless, and non-invasive nature of these face recognition technologies driven by AI has led to their meteoric rise in popularity in recent years. While they are mostly successful, there are still methods to get inside without permission by utilising things like pictures, masks, glasses, etc. In this research, we present an alternate authentication process that makes use of both facial recognition and the individual's distinctive temporal facial feature motions while they speak a password. Because the suggested methodology allows for a password to be specified in any language, it is not limited by language. The suggested model attained an accuracy of 96.1% when tested on the industry-standard MIRACL-VC1 dataset, demonstrating its efficacy as a reliable and powerful solution. In addition to being data-efficient, the suggested technique shows promising outcomes with as little as 10 positive video examples for training the model. The effectiveness of the network's training is further proved via comparisons with other combined facial recognition and lip reading models. | ['Garika Akshay', 'Pangoth Santhosh Kumar'] | 2023-04-30 | null | null | null | null | ['face-recognition', 'visual-speech-recognition'] | ['computer-vision', 'speech'] | [-7.82596916e-02 -1.36827677e-01 -4.58659589e-01 -2.22075194e-01
-2.32487723e-01 -2.49348044e-01 6.98849738e-01 -4.06961292e-01
-6.16784632e-01 6.06478512e-01 -2.92206705e-01 -3.88987571e-01
1.69129968e-01 -3.73492837e-01 -1.45672202e-01 -8.57792974e-01
1.95641845e-01 -1.71538610e-02 -6.10002838e-02 -1.55183328e-02
1.24818057e-01 6.15439177e-01 -1.95224965e+00 3.32546234e-02
5.14031470e-01 1.42123306e+00 -1.86051622e-01 3.51380289e-01
-1.06184870e-01 3.33175898e-01 -4.66122240e-01 -8.37595165e-01
3.82945180e-01 -2.35423483e-02 -3.94904494e-01 4.67424048e-03
3.25160950e-01 -5.71850777e-01 -1.29179239e-01 8.32575023e-01
8.55951428e-01 -2.31994763e-01 3.47469509e-01 -1.31940281e+00
-5.05414486e-01 8.70047957e-02 -7.86079228e-01 -1.27780542e-01
6.89854860e-01 3.16286951e-01 4.47251946e-01 -1.02367663e+00
1.08752109e-01 9.13258553e-01 8.06395113e-01 1.02857363e+00
-7.66724288e-01 -1.17813241e+00 -3.55974644e-01 2.03046039e-01
-1.65061402e+00 -9.28435385e-01 7.11886227e-01 -2.05864012e-01
9.17584062e-01 1.37750760e-01 6.91786051e-01 1.30141878e+00
7.13224337e-03 6.86433315e-01 1.15863013e+00 -6.37092710e-01
1.03531845e-01 7.34088540e-01 -1.11542806e-01 7.91281760e-01
3.66648793e-01 -1.05459124e-01 -5.93892217e-01 -1.87696785e-01
4.57196444e-01 1.70746863e-01 -1.32401943e-01 -1.23451494e-01
-7.90745080e-01 5.85052788e-01 -1.72724172e-01 5.22483170e-01
-4.01772767e-01 -1.46951556e-01 3.96526247e-01 1.11946523e-01
4.51731771e-01 -4.06310940e-03 -4.13850576e-01 -4.38107729e-01
-1.10671711e+00 -2.86041826e-01 8.67003500e-01 5.14750898e-01
4.21934038e-01 2.66982019e-01 4.04362112e-01 8.79845619e-01
5.81580877e-01 5.77160418e-01 9.10521686e-01 -4.74863917e-01
1.71482697e-01 6.17031813e-01 8.51983279e-02 -9.86266315e-01
-2.85668075e-01 3.48166972e-02 -1.07597315e+00 3.91991615e-01
6.49506927e-01 -1.07606038e-01 -8.39704394e-01 1.63550580e+00
3.58896583e-01 4.43249971e-01 -6.04532696e-02 5.33831656e-01
8.27852845e-01 3.92386258e-01 1.70319095e-01 -3.34407508e-01
1.57990420e+00 -3.93744290e-01 -9.08667922e-01 1.08091213e-01
4.20808882e-01 -7.10646749e-01 1.09799814e+00 7.31787562e-01
-8.86925459e-01 -6.66209638e-01 -9.15145874e-01 2.62895077e-01
-5.99526882e-01 4.23651546e-01 8.88506353e-01 1.64445019e+00
-1.28472054e+00 4.23829526e-01 -5.68976998e-01 -7.15983808e-01
6.48286939e-01 9.54980612e-01 -7.43705928e-01 1.23317868e-01
-1.05392611e+00 6.73720777e-01 -5.03097996e-02 2.06841439e-01
-3.46283764e-01 -1.78450003e-01 -6.42029941e-01 -7.08371475e-02
4.41841781e-02 -2.98144728e-01 8.54561508e-01 -1.01161849e+00
-2.05591941e+00 1.17042172e+00 -3.98348898e-01 -3.96450013e-01
4.28457230e-01 -3.01094323e-01 -7.69231915e-01 1.58322737e-01
-4.60549414e-01 6.05581641e-01 1.32618654e+00 -7.10178077e-01
-4.36900795e-01 -4.25861865e-01 -2.17176139e-01 -1.76961809e-01
-8.68775845e-01 3.40396464e-01 -3.63567770e-01 -2.60612965e-01
-1.83685631e-01 -8.16949248e-01 2.42007151e-01 1.23509154e-01
-3.21875930e-01 -4.10028249e-01 1.25206268e+00 -7.52040982e-01
1.17813909e+00 -2.20141888e+00 -5.06108820e-01 3.10718566e-01
2.77442783e-02 9.52155888e-01 1.39764071e-01 1.87778905e-01
-4.28890362e-02 4.62662458e-01 3.81951965e-02 -6.76077962e-01
-1.60510644e-01 -1.50129512e-01 -1.92906167e-02 5.89215875e-01
1.91907093e-01 7.61691332e-01 -3.40860993e-01 -4.20162141e-01
4.22646284e-01 1.06155157e+00 -3.05837721e-01 1.25140920e-01
4.57752794e-01 3.17317069e-01 -2.22322732e-01 1.05282450e+00
8.75069201e-01 3.55974515e-03 3.44451293e-02 -1.43990710e-01
4.04398851e-02 -3.78515832e-02 -1.25697458e+00 1.24057388e+00
-3.46961796e-01 5.80731750e-01 2.14931697e-01 -8.42365861e-01
8.96590471e-01 7.52522051e-01 4.29034501e-01 -5.17469823e-01
3.59904677e-01 3.19705635e-01 -1.05095215e-01 -6.95336461e-01
1.59994155e-01 -4.25974429e-02 4.77503866e-01 5.11348784e-01
6.46167761e-03 2.32098818e-01 -2.03401089e-01 -3.18547994e-01
6.48171306e-01 1.76682740e-01 4.19450611e-01 -5.93588538e-02
1.01233351e+00 -8.13436806e-01 2.29775891e-01 3.96458477e-01
-5.23658931e-01 2.86972225e-01 -1.78008214e-01 -4.99802470e-01
-5.99728525e-01 -4.47460055e-01 -2.34998748e-01 6.66521072e-01
-5.59370443e-02 -3.80283237e-01 -1.04465342e+00 -5.28348386e-01
-1.26361415e-01 9.58730876e-02 -4.44045663e-01 -1.07949331e-01
-3.73902172e-01 -5.60770869e-01 9.08054054e-01 2.64523596e-01
9.57645833e-01 -1.09233487e+00 -5.90712726e-01 5.40459789e-02
1.20432399e-01 -1.25076008e+00 -2.19932318e-01 -3.25930893e-01
-8.96854699e-01 -8.33654940e-01 -9.53225255e-01 -6.98918283e-01
4.37765747e-01 2.08756879e-01 5.93559623e-01 4.23165828e-01
-3.35375309e-01 5.62733769e-01 -7.93851614e-02 -4.82253790e-01
-1.08030833e-01 1.11278770e-02 7.79030561e-01 7.05714285e-01
9.34846520e-01 -5.88554859e-01 -6.53970242e-01 3.29859376e-01
-6.21919513e-01 -2.65774101e-01 4.44131494e-01 6.80215299e-01
-2.61896639e-03 -1.42650157e-01 6.26978576e-01 -4.69523728e-01
6.63349926e-01 -2.66833901e-01 -3.37560207e-01 1.68011725e-01
-7.74879277e-01 -2.80386031e-01 4.66104448e-01 -7.73035228e-01
-9.45583284e-01 1.62230968e-01 -4.80007946e-01 -2.85773516e-01
-5.52408576e-01 1.23025194e-01 -3.69766504e-01 -4.47200537e-01
3.12864214e-01 3.01051110e-01 4.00109708e-01 -5.30431926e-01
-6.81659132e-02 1.25704670e+00 2.97294497e-01 -1.99622244e-01
8.46971929e-01 4.14978325e-01 -2.50673085e-01 -1.36491764e+00
-1.08995579e-01 -4.07963127e-01 -5.94485641e-01 -3.19303960e-01
7.88133264e-01 -7.64342248e-01 -1.47285450e+00 1.08302414e+00
-9.24686849e-01 8.61570239e-02 1.23039164e-01 5.50508559e-01
-1.02883965e-01 5.96744835e-01 -4.55656856e-01 -1.52758479e+00
-5.51009893e-01 -1.09907699e+00 8.41035783e-01 5.69413245e-01
-2.02968463e-01 -9.37512159e-01 -3.17098349e-01 5.21320820e-01
7.66272783e-01 -4.91574779e-02 3.30505937e-01 -6.47167027e-01
-2.38804474e-01 -7.80227721e-01 -1.15199804e-01 5.76002240e-01
5.18975079e-01 2.25292847e-01 -1.56559420e+00 -3.30307364e-01
2.53964067e-01 -3.64902198e-01 3.62150341e-01 1.94998816e-01
1.13986564e+00 -4.44995403e-01 -3.70243490e-01 6.35814250e-01
1.10505354e+00 4.15876627e-01 8.50411296e-01 1.75680012e-01
5.16783953e-01 6.01142406e-01 1.36853725e-01 4.29241419e-01
1.69207886e-01 7.71570385e-01 2.78016835e-01 -1.36047423e-01
3.22074890e-02 -1.15163833e-01 4.72943664e-01 6.64791346e-01
-4.93138283e-01 -4.85726222e-02 -8.36782873e-01 1.29198551e-01
-1.23227215e+00 -1.23707366e+00 2.50481993e-01 2.53261638e+00
7.56334066e-01 -9.66031626e-02 2.75976866e-01 6.12420321e-01
6.58132792e-01 -1.66440845e-01 -3.82199943e-01 -5.09206176e-01
-1.48560837e-01 5.36114275e-01 1.87152833e-01 2.31424034e-01
-1.09731197e+00 8.77476513e-01 6.14885950e+00 7.37235308e-01
-1.67533338e+00 5.72017469e-02 8.73394907e-01 1.24694079e-01
2.48671860e-01 -4.65561181e-01 -9.86962020e-01 5.88650763e-01
1.23517907e+00 2.50864744e-01 2.84769207e-01 7.92174816e-01
2.91206747e-01 -2.38782227e-01 -8.37348998e-01 1.46620882e+00
1.47692516e-01 -1.14517057e+00 -3.82395461e-02 3.88368726e-01
1.01266153e-01 -4.64064837e-01 3.95686358e-01 1.31808713e-01
-4.18895692e-01 -1.32422924e+00 3.39501888e-01 3.56650949e-01
1.19224739e+00 -6.92847729e-01 9.06950116e-01 3.31296653e-01
-1.12347233e+00 -2.16102362e-01 -8.82477015e-02 -1.63086429e-01
-8.25571641e-02 2.06564933e-01 -8.56371939e-01 3.07876468e-01
7.25763798e-01 5.37729442e-01 -4.94809896e-01 9.43513393e-01
4.18905877e-02 8.71287107e-01 -6.35159135e-01 -2.40761891e-01
-1.02279216e-01 -1.51068985e-01 3.10792457e-02 1.08223617e+00
3.84777367e-01 8.49756077e-02 -2.90398538e-01 2.61329353e-01
-8.63698125e-02 3.54626387e-01 -8.30481887e-01 -1.94846630e-01
3.57325584e-01 1.44606519e+00 -6.05758369e-01 -1.25421643e-01
-5.60115874e-01 1.01225150e+00 -2.72216618e-01 2.22014382e-01
-6.80468917e-01 -3.41481179e-01 7.50602484e-01 2.59459555e-01
-8.31005648e-02 -1.92671508e-01 -1.65321916e-01 -1.04060578e+00
2.86595821e-02 -1.07969058e+00 -4.41423543e-02 -5.71924686e-01
-9.76156652e-01 7.33942330e-01 -3.44374597e-01 -1.29664087e+00
-3.36273342e-01 -7.42841721e-01 -4.48822528e-01 9.90763128e-01
-1.45053923e+00 -1.30842829e+00 -3.44225556e-01 1.00487113e+00
2.64035076e-01 -6.46627486e-01 1.15745580e+00 5.44349134e-01
-6.39854133e-01 1.11238706e+00 -1.68424487e-01 3.67454857e-01
6.47956371e-01 -6.76356971e-01 2.24519268e-01 5.96987426e-01
2.79467553e-01 9.65608478e-01 3.96880090e-01 -2.63481379e-01
-1.58657134e+00 -2.53823757e-01 8.59850943e-01 -2.30212167e-01
3.34386110e-01 -5.06212950e-01 -8.27622771e-01 1.67391941e-01
2.88716495e-01 3.79493497e-02 9.50656652e-01 -2.44546950e-01
-3.51019531e-01 -3.88733208e-01 -1.60516083e+00 5.26498735e-01
7.18143582e-01 -6.39562070e-01 -1.62017763e-01 1.48368895e-01
-1.72539353e-02 -1.43118168e-03 -5.72237492e-01 3.07243794e-01
1.00766170e+00 -1.10201490e+00 8.46715271e-01 -4.11297530e-01
-2.39387706e-01 -1.13012359e-01 3.17488573e-02 -4.95221138e-01
2.27586180e-01 -1.12281680e+00 -2.88960367e-01 1.52186513e+00
3.08364749e-01 -1.03547478e+00 1.10202694e+00 9.25341189e-01
6.56693578e-01 -7.50093222e-01 -1.19214439e+00 -6.01643562e-01
-4.15033937e-01 -4.88578230e-01 5.50686121e-01 8.88042092e-01
-2.80014309e-03 6.02670088e-02 -7.17543662e-01 -7.87428841e-02
5.56147575e-01 -5.53330779e-01 8.45726073e-01 -1.47876120e+00
2.43717469e-02 -3.12979549e-01 -7.04184353e-01 -7.57366955e-01
1.50678515e-01 -4.24066186e-01 -4.53815937e-01 -8.95717204e-01
-6.37851581e-02 -5.09711444e-01 -2.61391282e-01 6.99056745e-01
1.42782092e-01 7.68849730e-01 1.05037764e-01 2.02541575e-01
2.29673926e-02 3.09136480e-01 6.90954149e-01 -2.80081183e-02
-2.75164366e-01 4.67675775e-01 -5.71051478e-01 7.63016641e-01
9.57456708e-01 -8.01380724e-02 -3.19944322e-01 1.49127930e-01
-6.98693320e-02 -2.19280362e-01 1.00007020e-01 -1.20992541e+00
2.83640355e-01 2.41286635e-01 5.17143369e-01 -6.01116754e-02
7.46197641e-01 -1.04523122e+00 1.27337009e-01 4.02940214e-01
1.63818166e-01 9.12771448e-02 4.13963139e-01 2.37476915e-01
-1.53092980e-01 -4.40307893e-02 8.23040903e-01 1.05534963e-01
-6.66557789e-01 3.27349901e-01 -4.45021629e-01 -5.07226110e-01
1.11092448e+00 -9.70544219e-01 7.34762624e-02 -5.92681944e-01
-5.12514114e-01 -4.20347005e-01 3.59858006e-01 5.32319784e-01
7.52401352e-01 -1.09548843e+00 -3.68099689e-01 7.74547398e-01
-2.82040387e-01 -6.62518919e-01 6.02238029e-02 9.85086620e-01
-2.12595329e-01 5.33535600e-01 -2.75236100e-01 -6.48844421e-01
-1.77691281e+00 3.68911088e-01 3.54894757e-01 1.57469556e-01
-5.15465379e-01 7.48001635e-01 -2.88314432e-01 -6.57112822e-02
5.35252631e-01 5.14876097e-02 -4.49563682e-01 4.95501682e-02
1.02874136e+00 6.39434531e-02 4.12827395e-02 -7.68271685e-01
-6.34328127e-01 8.24202061e-01 -3.73580283e-03 -1.05913162e-01
1.20395792e+00 -6.41975105e-02 1.37039730e-02 2.95922786e-01
1.00157678e+00 3.66386235e-01 -1.01007617e+00 6.61171004e-02
9.11307037e-02 -4.30323988e-01 -2.47488618e-01 -7.03589678e-01
-1.22704887e+00 1.11453199e+00 1.20758402e+00 3.11552107e-01
1.12524056e+00 -3.96190941e-01 7.51149416e-01 2.95697659e-01
5.75403810e-01 -8.87732923e-01 -6.97953850e-02 1.89534098e-01
4.24735993e-01 -1.49543691e+00 -3.73299234e-02 -1.91070065e-01
-5.17506242e-01 1.20766997e+00 4.00631547e-01 4.01950151e-01
8.92358899e-01 4.17825222e-01 1.81816772e-01 1.75034389e-01
-3.95664304e-01 1.68780573e-02 2.55719393e-01 9.59684908e-01
5.92532098e-01 -2.17276752e-01 -1.86011627e-01 5.35666645e-01
-3.23231478e-04 3.50604028e-01 3.45677644e-01 8.09033930e-01
-2.40229473e-01 -1.27212679e+00 -3.16742599e-01 3.19183260e-01
-9.43488300e-01 -2.29891166e-02 -3.02275360e-01 8.77971947e-01
1.04689777e-01 1.18570721e+00 -2.65072972e-01 -6.35413527e-01
-8.61306265e-02 4.65961069e-01 2.13699669e-01 -1.72201961e-01
-6.69090152e-01 -2.04811037e-01 -1.61286980e-01 -5.53746223e-01
-7.19348609e-01 -5.04100263e-01 -8.44893575e-01 -5.99360466e-01
-4.53850895e-01 2.54330076e-02 1.03319657e+00 1.01238644e+00
4.39649552e-01 -3.61327022e-01 6.72291279e-01 -9.45060790e-01
-4.43947792e-01 -1.02802658e+00 -4.91396606e-01 1.65701970e-01
3.06545585e-01 -6.89878285e-01 -4.40881222e-01 -1.61811169e-02] | [13.261495590209961, 1.1654514074325562] |
fd15b52f-9519-4c6d-b3cf-6e707a980e03 | a-new-paradigm-for-generative-adversarial | 2306.13641 | null | https://arxiv.org/abs/2306.13641v1 | https://arxiv.org/pdf/2306.13641v1.pdf | A New Paradigm for Generative Adversarial Networks based on Randomized Decision Rules | The Generative Adversarial Network (GAN) was recently introduced in the literature as a novel machine learning method for training generative models. It has many applications in statistics such as nonparametric clustering and nonparametric conditional independence tests. However, training the GAN is notoriously difficult due to the issue of mode collapse, which refers to the lack of diversity among generated data. In this paper, we identify the reasons why the GAN suffers from this issue, and to address it, we propose a new formulation for the GAN based on randomized decision rules. In the new formulation, the discriminator converges to a fixed point while the generator converges to a distribution at the Nash equilibrium. We propose to train the GAN by an empirical Bayes-like method by treating the discriminator as a hyper-parameter of the posterior distribution of the generator. Specifically, we simulate generators from its posterior distribution conditioned on the discriminator using a stochastic gradient Markov chain Monte Carlo (MCMC) algorithm, and update the discriminator using stochastic gradient descent along with simulations of the generators. We establish convergence of the proposed method to the Nash equilibrium. Apart from image generation, we apply the proposed method to nonparametric clustering and nonparametric conditional independence tests. A portion of the numerical results is presented in the supplementary material. | ['Faming Liang', 'Qifan Song', 'Sehwan Kim'] | 2023-06-23 | null | null | null | null | ['clustering'] | ['methodology'] | [ 1.94048807e-01 6.36379719e-02 -2.76443595e-03 3.67348753e-02
-7.03862369e-01 -5.91284156e-01 6.04412794e-01 -4.05099303e-01
-1.99498355e-01 1.12716687e+00 2.18066797e-02 -2.56266057e-01
-5.88121451e-02 -9.53489065e-01 -6.80158496e-01 -1.24755669e+00
3.05408686e-01 6.99006319e-01 -1.62311137e-01 2.45549574e-01
8.58419016e-03 3.98225933e-01 -1.16155636e+00 -2.95213193e-01
1.22974026e+00 6.38097346e-01 -7.85387829e-02 7.46581137e-01
1.55126229e-01 6.02941990e-01 -9.33678687e-01 -4.61994439e-01
1.82228163e-01 -1.22921419e+00 -3.19218069e-01 2.67287847e-02
-4.03446883e-01 -3.08837712e-01 1.34201631e-01 1.23751950e+00
4.50408369e-01 2.51978695e-01 1.29060602e+00 -1.45514011e+00
-5.07710874e-01 3.31051499e-01 -5.80519617e-01 -2.33080015e-01
-1.25567690e-01 -1.96624473e-02 6.15268409e-01 -6.48974419e-01
3.08617383e-01 1.01183259e+00 3.78483027e-01 6.64151609e-01
-1.30392408e+00 -6.76528394e-01 -3.99989903e-01 -1.13294579e-01
-1.55631232e+00 -2.89413959e-01 8.32207978e-01 -5.57868898e-01
1.36049151e-01 1.56802297e-01 6.52904212e-01 1.20895815e+00
3.92151237e-01 5.74591994e-01 1.15982711e+00 -5.65137684e-01
8.68316770e-01 1.34377658e-01 -5.70820451e-01 5.83932936e-01
3.45841676e-01 2.07815871e-01 -1.88347712e-01 -2.94507504e-01
1.06709754e+00 -1.40714824e-01 -8.50551799e-02 -3.37649584e-01
-6.72507465e-01 1.23393798e+00 2.07958385e-01 1.06283940e-01
-4.57719505e-01 3.52166146e-01 -1.77692890e-01 -1.55068696e-01
3.20563823e-01 9.42640938e-03 2.16064885e-01 -6.10634238e-02
-1.08431137e+00 1.45477474e-01 8.98424149e-01 6.18040681e-01
5.96229613e-01 4.79454994e-01 -3.00133824e-01 5.78196466e-01
5.37983954e-01 7.25672722e-01 3.36408049e-01 -9.25970495e-01
1.34725630e-01 9.07187164e-02 1.78997114e-01 -6.94800615e-01
1.90230593e-01 -6.15155756e-01 -1.35235417e+00 3.25439513e-01
4.78238940e-01 -6.06162071e-01 -9.76306498e-01 1.96267033e+00
3.83770049e-01 3.87389898e-01 -1.51397251e-02 6.29568279e-01
2.00604439e-01 8.32432151e-01 1.54329333e-02 -4.65644270e-01
6.63129330e-01 -5.61533868e-01 -6.97608709e-01 3.13482642e-01
1.61808580e-01 -6.29849017e-01 7.73207128e-01 2.28158429e-01
-1.00251722e+00 -3.92030239e-01 -8.90448809e-01 6.46289945e-01
1.61727175e-01 1.86292976e-01 3.64422023e-01 1.00558209e+00
-9.75263059e-01 4.99288350e-01 -1.08684814e+00 -2.16305107e-01
3.68940413e-01 9.49778885e-04 2.01059446e-01 1.48398980e-01
-9.50866163e-01 5.13028979e-01 4.03127551e-01 1.71687841e-01
-1.09261036e+00 -1.12930417e-01 -6.18198097e-01 1.19676732e-01
1.83162183e-01 -9.81614590e-01 1.03472900e+00 -1.10473895e+00
-2.09578633e+00 3.05248439e-01 -2.48060614e-01 -3.06839108e-01
7.11547792e-01 1.42649338e-01 -1.80409655e-01 3.03800795e-02
1.53809875e-01 3.07996154e-01 1.13029671e+00 -1.38184714e+00
-1.81505308e-01 3.99234481e-02 -3.21806461e-01 3.90705690e-02
4.89773005e-02 -3.08922499e-01 -2.93855697e-01 -9.18358326e-01
2.08190847e-02 -9.68809009e-01 -2.53872603e-01 -5.70091605e-01
-7.40627170e-01 -2.47444995e-02 4.01582450e-01 -4.88016307e-01
1.03977561e+00 -2.13186169e+00 1.94889665e-01 5.66741467e-01
-5.25633655e-02 -7.07878396e-02 2.63551682e-01 5.70023000e-01
1.11295745e-01 1.33902550e-01 -6.43173039e-01 -3.76883209e-01
6.63043633e-02 2.48132840e-01 -3.78101498e-01 6.26772344e-01
1.59221087e-02 7.29250848e-01 -8.34473073e-01 -4.10707265e-01
1.06797539e-01 4.10575241e-01 -6.67649209e-01 3.63086045e-01
-1.38764948e-01 8.70595992e-01 -4.84231472e-01 4.03321356e-01
7.59262323e-01 -2.17947930e-01 2.91913539e-01 1.81633577e-01
6.74760118e-02 -2.25065678e-01 -1.20964801e+00 1.01641941e+00
-1.78282231e-01 1.07283585e-01 -1.66556522e-01 -1.19720435e+00
8.91844749e-01 3.05801332e-01 2.74979025e-01 -2.96652652e-02
2.61848152e-01 2.27147356e-01 1.49669543e-01 -1.36633620e-01
1.85017258e-01 -6.77714407e-01 -1.65605441e-01 7.32151687e-01
1.28515214e-01 -3.45275998e-01 1.61715493e-01 2.18797937e-01
8.47979903e-01 7.03658909e-02 4.41767424e-01 -1.35397330e-01
4.18759584e-01 -2.44547293e-01 5.25479019e-01 1.01601923e+00
2.74385810e-01 7.10628510e-01 5.94845831e-01 1.74539715e-01
-1.13207841e+00 -1.52727103e+00 3.43465358e-02 2.77917564e-01
-1.09331071e-01 -7.47356191e-02 -9.56779361e-01 -5.27332902e-01
-3.67025375e-01 9.91702020e-01 -6.03954136e-01 -3.72186184e-01
-2.20669866e-01 -9.59272623e-01 3.89345348e-01 3.16569090e-01
8.12682033e-01 -1.00099361e+00 -2.45067790e-01 1.43541381e-01
-2.11972386e-01 -6.33632600e-01 -2.19651729e-01 7.02678263e-02
-9.09534693e-01 -7.76887000e-01 -9.95226502e-01 -4.67340559e-01
9.81160402e-01 -3.85103643e-01 8.26312900e-01 -1.26353323e-01
1.65865123e-01 1.34939864e-01 -2.71887541e-01 -2.05981344e-01
-9.80221093e-01 1.26480356e-01 -3.99439484e-02 1.40566841e-01
-2.04456493e-01 -6.02878273e-01 -5.04252434e-01 4.25624907e-01
-1.07402778e+00 8.24103951e-02 5.35692811e-01 8.99539530e-01
7.29478121e-01 5.32256246e-01 8.40140641e-01 -8.16039562e-01
6.84620321e-01 -6.30532265e-01 -8.32017899e-01 1.16431087e-01
-3.52605075e-01 2.65472263e-01 8.82610619e-01 -3.93605143e-01
-1.12510824e+00 8.59735683e-02 -1.68072596e-01 -4.36738998e-01
-7.33061135e-02 5.50304294e-01 -4.12177980e-01 2.16853023e-01
3.09766859e-01 3.10684353e-01 9.20510013e-03 -3.53898942e-01
3.49198490e-01 4.02846336e-01 6.40083849e-01 -7.63025165e-01
1.22177720e+00 4.50975716e-01 2.60338366e-01 -5.07417381e-01
-6.60094082e-01 1.88358992e-01 -3.10714304e-01 -2.47860953e-01
9.49202240e-01 -6.13532066e-01 -6.43627286e-01 7.43285835e-01
-8.65512550e-01 -4.08346295e-01 -5.78852952e-01 7.03256011e-01
-8.60169291e-01 1.00416541e-01 -3.51771802e-01 -1.15842974e+00
-5.07202074e-02 -9.18801129e-01 5.86453974e-01 4.31540549e-01
-4.90205130e-03 -1.06773210e+00 3.99596363e-01 -6.37900596e-03
2.84271151e-01 6.17060006e-01 1.08088815e+00 -4.36135828e-01
-4.63259459e-01 -2.91593879e-01 1.61416829e-01 5.05647957e-01
3.02296370e-01 1.77613407e-01 -6.82290137e-01 -3.58842731e-01
4.52077508e-01 2.00474765e-02 6.65742218e-01 6.83785021e-01
1.15028870e+00 -4.17544365e-01 -2.31303185e-01 5.69159389e-01
1.51601899e+00 4.83905345e-01 7.55235910e-01 -4.03263941e-02
4.52539444e-01 2.15997428e-01 2.03602374e-01 6.48333430e-01
5.36534376e-02 4.15959030e-01 3.14239860e-01 9.64831188e-02
1.76471382e-01 -5.93286991e-01 2.91512549e-01 8.42080176e-01
-2.32103050e-01 -5.86184323e-01 -8.15944850e-01 3.57737035e-01
-1.75144804e+00 -1.07745755e+00 2.12824121e-01 2.59377813e+00
8.51842403e-01 1.17912732e-01 2.38560066e-01 8.81660804e-02
1.10654616e+00 -2.45497778e-01 -5.67995727e-01 -2.14707389e-01
-6.19665124e-02 4.38484311e-01 4.32417274e-01 4.81126875e-01
-7.03371942e-01 6.22719347e-01 6.39535141e+00 9.98225629e-01
-9.71406639e-01 1.45416215e-01 7.89890587e-01 1.19911544e-01
-4.53232527e-01 1.67682543e-01 -4.83594835e-01 9.56054866e-01
8.17042887e-01 -3.12525421e-01 4.84843612e-01 5.87215960e-01
2.46356070e-01 -4.36919242e-01 -8.46090138e-01 7.16923833e-01
2.42503220e-03 -1.11399674e+00 8.92624408e-02 3.14630628e-01
1.18908596e+00 -4.90877062e-01 2.11059317e-01 1.07498810e-01
6.38654470e-01 -1.12804198e+00 7.31468737e-01 6.79815114e-01
8.04064453e-01 -1.19977891e+00 8.01622808e-01 6.47704065e-01
-7.99385667e-01 7.70222545e-02 -2.84188777e-01 9.56249610e-02
3.32216442e-01 9.00279641e-01 -8.79082799e-01 5.03670931e-01
2.71268964e-01 1.72325253e-01 -2.96315402e-01 1.11403656e+00
-5.38148403e-01 1.01553690e+00 -3.63712013e-01 -9.36394855e-02
-7.76645765e-02 -7.77782559e-01 5.01948893e-01 5.90736270e-01
8.33275557e-01 -1.59394041e-01 -1.74815968e-01 1.20606041e+00
-1.77213550e-01 -9.64787230e-02 -4.46268380e-01 -2.72608250e-01
5.68509042e-01 9.19010103e-01 -1.06315124e+00 -2.63456762e-01
3.79752517e-02 9.98163581e-01 1.33281767e-01 6.45181239e-01
-1.08493471e+00 -2.97209293e-01 1.01931226e-02 -3.05490606e-02
4.65906888e-01 -2.79248983e-01 -2.76342273e-01 -1.04208946e+00
-1.55187383e-01 -7.10212469e-01 2.69253671e-01 -5.84531367e-01
-1.30873430e+00 4.00324672e-01 2.60529280e-01 -1.20287716e+00
-8.66377115e-01 -1.08612612e-01 -1.01194942e+00 9.49899912e-01
-9.80710328e-01 -5.78277171e-01 -1.07201040e-01 6.34841383e-01
-9.65254265e-04 -2.60688812e-01 7.38253772e-01 -2.75336597e-02
-6.64662719e-01 6.07303619e-01 7.22886086e-01 8.99428353e-02
3.82437289e-01 -1.15767026e+00 7.20313489e-02 9.36922133e-01
-2.38000080e-02 4.13537800e-01 8.29103827e-01 -8.16716611e-01
-8.69403005e-01 -9.72410202e-01 4.97249961e-01 -7.56622031e-02
4.68609244e-01 -3.13300550e-01 -5.07631958e-01 5.41983187e-01
1.63603142e-01 -2.56679267e-01 6.27456844e-01 -3.99990976e-01
2.94242740e-01 1.27877071e-01 -1.40965950e+00 5.96293688e-01
4.27141309e-01 -2.28324279e-01 -2.11424213e-02 2.14185923e-01
2.80436665e-01 -2.06991091e-01 -5.40433943e-01 3.04144323e-01
2.53745586e-01 -1.09538364e+00 5.77041149e-01 -1.70709580e-01
4.59213823e-01 -4.56534833e-01 -1.01511404e-01 -1.58678901e+00
-1.54929876e-01 -7.95393407e-01 -2.48723969e-01 1.50615025e+00
2.05050021e-01 -8.16208720e-01 8.54342520e-01 1.40871361e-01
1.92421511e-01 -4.82475728e-01 -1.31066930e+00 -9.22078192e-01
3.34374905e-01 -2.20639318e-01 5.60135543e-01 4.48019981e-01
-5.76779187e-01 1.39844701e-01 -3.79705250e-01 1.48166209e-01
8.96197677e-01 1.59224644e-01 7.79289842e-01 -9.27716851e-01
-7.31889129e-01 -2.42502525e-01 -1.00133479e-01 -1.02225959e+00
1.62724674e-01 -7.48807132e-01 2.89106935e-01 -1.36401808e+00
3.30169559e-01 -4.79634106e-01 -2.31430251e-02 -1.16093501e-01
-1.35948241e-01 1.95478380e-01 1.62058651e-01 3.09143871e-01
-1.42161429e-01 8.39871228e-01 1.20061851e+00 1.00987330e-01
-1.10496201e-01 5.58339298e-01 -5.01822650e-01 6.01505220e-01
9.99484241e-01 -6.53012693e-01 -5.48118830e-01 6.42923638e-02
2.41399154e-01 2.72313803e-01 4.48036253e-01 -1.08081102e+00
5.40898589e-04 -8.33133385e-02 5.45452178e-01 -4.44426537e-01
3.01291227e-01 -6.18840158e-01 6.11028790e-01 5.14955938e-01
8.87156092e-03 -1.40354633e-01 -1.94553554e-01 6.87653303e-01
-1.20477423e-01 -6.57329738e-01 9.52603519e-01 -9.37683973e-03
2.06332222e-01 7.63191953e-02 -8.02343607e-01 1.75293028e-01
9.01927590e-01 -2.00852696e-02 -6.71434123e-03 -8.85991395e-01
-7.29740083e-01 -1.37009442e-01 6.24991298e-01 -3.11665505e-01
4.51625139e-01 -1.57448828e+00 -6.11959279e-01 3.58197153e-01
-5.26758909e-01 1.29030079e-01 1.29452929e-01 7.44689703e-01
-2.80818313e-01 -4.37512435e-02 -7.68253133e-02 -5.39480805e-01
-5.35870016e-01 2.87006259e-01 4.28169996e-01 -2.62517065e-01
-3.47819254e-02 5.10259569e-01 2.84408540e-01 -2.73186296e-01
-8.91711935e-02 6.06727898e-02 6.00892119e-02 -2.88472682e-01
5.64855151e-02 4.58362877e-01 -2.13901088e-01 -4.05430645e-01
-1.82113722e-01 3.94937277e-01 4.18895870e-01 -6.30850375e-01
1.07469046e+00 -4.69690673e-02 -1.34854689e-01 4.62643266e-01
9.33202207e-01 1.87549859e-01 -1.28025734e+00 1.25835165e-01
-4.58824188e-01 -3.47653896e-01 -2.27393612e-01 -6.03992105e-01
-1.12658083e+00 6.08491421e-01 4.42039698e-01 4.20653254e-01
1.08792055e+00 9.82915610e-03 4.62842047e-01 -8.66888538e-02
1.71359673e-01 -9.89799857e-01 4.46078479e-02 2.43824810e-01
5.47481239e-01 -9.28456962e-01 -1.05442576e-01 -1.58112332e-01
-5.81487536e-01 9.27070200e-01 3.84523332e-01 -3.32608491e-01
7.30588377e-01 2.83196718e-01 -1.79055810e-01 2.49121189e-01
-2.88960129e-01 3.57530899e-02 6.65753558e-02 6.55183971e-01
1.71358168e-01 2.30345190e-01 -3.21944028e-01 4.73754764e-01
-3.69048208e-01 1.11678861e-01 5.29177368e-01 7.40793645e-01
-1.28615662e-01 -1.18050921e+00 -4.45790142e-01 3.58432353e-01
-3.44527394e-01 -8.78933892e-02 -4.27782498e-02 5.67920625e-01
1.34927481e-01 9.69062507e-01 2.50463337e-01 -2.22586524e-02
-3.00547749e-01 1.26375407e-01 4.74711925e-01 -4.45574820e-01
1.87349599e-02 3.64321083e-01 -4.06157970e-01 -3.53940688e-02
-3.21798921e-01 -8.71210933e-01 -1.11882102e+00 -3.36729169e-01
-3.38749170e-01 4.61618841e-01 6.21606112e-01 9.92214978e-01
4.72714491e-02 4.11683828e-01 9.79147851e-01 -6.65341437e-01
-7.24208176e-01 -8.87268305e-01 -7.00060368e-01 1.66234374e-01
-8.30290243e-02 -5.74166656e-01 -7.70657599e-01 -3.76577154e-02] | [7.023630142211914, 3.8706438541412354] |
b37e296f-c911-4cc9-89ce-869fc8eaee00 | design-in-the-dark-learning-deep-generative | null | null | https://openreview.net/forum?id=WQVouCWioh | https://openreview.net/pdf?id=WQVouCWioh | Design in the Dark: Learning Deep Generative Models for De Novo Protein Design | The design of novel protein sequences is providing paths towards the development of novel therapeutics and materials.
Generative modelling approaches to design are emerging and to date have required conditioning on 3D protein structure-derived information, and unconditional models of protein sequences have so far performed poorly.
Thus, it is unknown if unconditional generative models can learn a distribution of sequences that captures structure information without it being explicitly provided, and so be of use in important tasks like de novo protein sequence design, where it is not possible to condition on structure.
Here, we demonstrate that it is possible to use unconditioned generative models to produce realistic samples of protein sequences.
We progressively grow a dataset of over half a million synthetic sequences for training autoregressive language models, using an iterative framework we call DARK.
It begins by training an autoregressive model on an initial sample of synthetic sequences, sampling from it, and refining the samples thus generated, which are then used for subsequent rounds of training.
Using the confidence measures provided by AlphaFold and other measures of sample quality, we show that our approach matches or exceeds the performance of prior methods that use weak conditioning on explicit structural information, and improves after each iteration of DARK.
Crucially, the DARK framework and the trained models are entirely unsupervised; strong structural signal is an objective, but no model is ever conditioned on any specific structural state.
The trained model indirectly learns to incorporate a structural signal into its learned sequence distribution, as this signal is strongly represented in the makeup of the training set at each step.
Our work demonstrates a way of unconditionally sampling sequences and structures jointly, and in an unsupervised way. | ['David T. Jones', 'Shaun M. Kandathil', 'Lewis Moffat'] | 2021-09-29 | null | null | null | null | ['protein-design'] | ['medical'] | [ 7.50021160e-01 1.81321427e-01 -5.43690361e-02 -2.50173002e-01
-7.60356009e-01 -7.84660697e-01 5.72966456e-01 2.40060836e-01
-3.94252270e-01 1.10022962e+00 2.09032670e-01 -5.83393812e-01
2.63773620e-01 -6.95493400e-01 -1.10896695e+00 -1.04339600e+00
9.00399014e-02 9.90581632e-01 1.68354914e-01 -1.11921698e-01
1.75749198e-01 5.12896538e-01 -1.26350796e+00 3.65082979e-01
7.86053777e-01 2.96022296e-01 4.80558962e-01 8.20755780e-01
-7.56956488e-02 5.37959039e-01 -3.47618967e-01 -5.33617809e-02
-3.49768661e-02 -8.14081967e-01 -6.40741169e-01 9.42469090e-02
-2.50691473e-01 -1.82768404e-01 1.85104623e-01 6.52551353e-01
4.82157081e-01 4.99201287e-03 9.98320937e-01 -4.64425683e-01
-5.90943396e-01 4.38077837e-01 -2.85562903e-01 -1.95431009e-01
3.03270012e-01 6.24321401e-01 9.11193550e-01 -9.10702705e-01
7.38085091e-01 8.90418172e-01 4.93472755e-01 6.94740832e-01
-1.86883903e+00 -3.68604660e-01 -2.40805140e-03 -4.40605670e-01
-1.07151246e+00 -4.49206144e-01 6.14253044e-01 -6.57876551e-01
1.08701241e+00 1.43265573e-03 6.34361863e-01 1.34220111e+00
2.63899773e-01 5.78461170e-01 1.02619648e+00 -4.27501231e-01
6.82539105e-01 5.34089468e-02 -9.67832096e-03 6.31305933e-01
1.72416940e-02 2.50038981e-01 -4.05192882e-01 -5.99216938e-01
5.86626530e-01 3.39244828e-02 -1.60861820e-01 -6.65809870e-01
-8.63332808e-01 9.83457386e-01 -4.77059484e-02 2.20077470e-01
-5.89946151e-01 4.43012342e-02 8.76495615e-02 6.39663860e-02
2.49829993e-01 5.77747822e-01 -6.17235661e-01 -2.05064386e-01
-1.13564992e+00 3.72199833e-01 7.57768452e-01 7.54234850e-01
8.15050483e-01 -9.04667154e-02 1.56068131e-01 6.13851428e-01
3.20209235e-01 3.11597139e-01 3.49719822e-01 -6.87858999e-01
-2.50925332e-01 2.73172766e-01 2.89768726e-01 -3.31165403e-01
-1.12875909e-01 -4.52180743e-01 -4.91588533e-01 1.81638628e-01
4.31667358e-01 -1.94166973e-01 -1.22881126e+00 1.99884903e+00
3.74506086e-01 -3.08635961e-02 4.32323888e-02 4.78452981e-01
2.93772578e-01 6.81026101e-01 3.10438633e-01 -5.36454141e-01
8.96229148e-01 -3.19909424e-01 -2.52703369e-01 -9.89195630e-02
5.40140450e-01 -7.49108434e-01 8.45615745e-01 7.17199564e-01
-1.12819588e+00 -4.55265641e-01 -1.07361042e+00 1.08876698e-01
-1.95853412e-01 -7.36199915e-02 6.32142127e-01 7.59209812e-01
-8.76718640e-01 8.04253161e-01 -1.06887186e+00 -8.81474316e-02
4.38865513e-01 4.65038806e-01 -3.17310750e-01 -1.17530800e-01
-1.05052769e+00 7.86076427e-01 4.90935802e-01 -1.38857901e-01
-1.35457075e+00 -7.87713766e-01 -6.74616277e-01 -4.29074131e-02
1.68536410e-01 -8.54643464e-01 1.07805824e+00 -8.76497567e-01
-1.60410285e+00 6.22492075e-01 -3.98029506e-01 -4.93217885e-01
3.05124134e-01 3.31485420e-02 1.07295044e-01 -1.31867863e-02
-1.87158436e-01 6.48792505e-01 8.41396630e-01 -1.25620210e+00
-7.30235577e-02 -2.53002226e-01 -1.36249438e-01 -1.59815699e-01
2.43781567e-01 -1.14806108e-01 -5.32803033e-03 -4.29221958e-01
-9.19107795e-02 -9.36178863e-01 -7.02119410e-01 -3.04406673e-01
-2.63675243e-01 3.30202878e-02 2.42386475e-01 -5.37003398e-01
8.87458742e-01 -1.81101274e+00 3.88097286e-01 6.26218557e-01
1.88061476e-01 2.95047611e-01 -1.03538357e-01 6.87439978e-01
-3.37554753e-01 2.15225935e-01 -6.26805067e-01 1.53734758e-01
-8.67321193e-02 2.40596369e-01 -3.52028012e-01 2.73887485e-01
5.57371140e-01 8.10518920e-01 -9.61579144e-01 -5.47550730e-02
7.46627226e-02 7.08457351e-01 -9.69043434e-01 2.83337265e-01
-9.13730621e-01 6.62633538e-01 -4.63284791e-01 2.42463753e-01
4.37347114e-01 -6.67265534e-01 7.15879858e-01 -3.88289392e-02
1.13742888e-01 3.15538734e-01 -8.23145151e-01 1.56430125e+00
-1.47626862e-01 -7.08258152e-02 -4.41514283e-01 -9.92946744e-01
9.61478949e-01 4.44246471e-01 5.99527419e-01 -1.22865312e-01
5.21174781e-02 1.58313677e-01 2.65014976e-01 -1.60366952e-01
-2.61768997e-02 -7.87967265e-01 2.38662362e-01 5.87267756e-01
2.01467782e-01 -1.22982711e-01 2.55063087e-01 2.52395004e-01
1.18014276e+00 6.40340030e-01 2.76928842e-01 -6.66206703e-02
3.28539819e-01 9.35963765e-02 5.13429940e-01 6.56911910e-01
2.40757957e-01 5.21784186e-01 7.36732423e-01 -1.61151841e-01
-1.67408407e+00 -1.17639470e+00 -1.17315337e-01 9.34456766e-01
-5.07360160e-01 -5.12517393e-01 -7.83846319e-01 -5.69942117e-01
-1.90493733e-01 7.33999014e-01 -7.15285420e-01 -2.57824838e-01
-4.72633511e-01 -1.11829126e+00 2.30219871e-01 3.74041528e-01
-4.46718812e-01 -9.96267259e-01 -3.39421004e-01 6.45910203e-01
3.13752741e-01 -5.78031719e-01 -4.10479337e-01 7.06126153e-01
-7.28924215e-01 -9.44066405e-01 -7.64838755e-01 -5.55433631e-01
8.50393653e-01 -3.49456310e-01 1.06992567e+00 -1.93654880e-01
-4.67253894e-01 2.73846220e-02 -6.43744171e-02 -3.93224895e-01
-9.00867939e-01 -1.36078089e-01 -1.68641657e-02 -9.43872705e-02
4.18293506e-01 -9.84462678e-01 -5.40072620e-01 3.47801074e-02
-1.08211458e+00 1.59053206e-01 7.92363822e-01 1.22874677e+00
7.93731213e-01 -3.49168450e-01 6.80027902e-01 -1.18034554e+00
4.51035261e-01 -4.99898583e-01 -4.98663306e-01 1.81356654e-01
-4.50051248e-01 6.58621907e-01 5.89757264e-01 -5.05145550e-01
-1.03852320e+00 3.96177709e-01 -5.65513968e-01 -2.36055270e-01
-2.11348847e-01 7.04129517e-01 -3.02230775e-01 2.45354578e-01
9.69039619e-01 5.00007033e-01 4.75394219e-01 -4.43004578e-01
3.93833786e-01 3.07898283e-01 1.72607362e-01 -1.01096165e+00
4.93904293e-01 2.97394782e-01 -5.28365709e-02 -7.98692107e-01
-4.85707194e-01 -3.02713037e-01 -5.44382870e-01 2.74105698e-01
7.28389382e-01 -5.99805713e-01 -7.05907941e-01 1.66029796e-01
-8.74420524e-01 -4.29156393e-01 -2.88114995e-01 4.47416037e-01
-9.73505735e-01 4.29975718e-01 -5.62857330e-01 -8.58860970e-01
-1.32192761e-01 -1.39089108e+00 9.62818027e-01 -2.38755122e-01
-5.23312509e-01 -8.63744318e-01 4.00555909e-01 1.09555833e-01
2.90997624e-01 2.67490208e-01 1.27866507e+00 -7.94887960e-01
-6.96365297e-01 -3.86373624e-02 3.66396427e-01 3.94964904e-01
3.02275658e-01 1.47841886e-01 -7.71641731e-01 -2.37362310e-01
6.20587915e-02 -5.76593339e-01 8.03008556e-01 3.73439819e-01
9.38012540e-01 -1.64759576e-01 -3.79712373e-01 1.57532498e-01
1.19746161e+00 3.28393489e-01 7.29046822e-01 -1.98156849e-01
3.68183225e-01 5.17242432e-01 2.32636169e-01 4.01357800e-01
-3.23924899e-01 3.33253264e-01 -9.98976827e-03 -1.04443118e-01
3.95720005e-01 -4.89448369e-01 4.29530263e-01 4.97680366e-01
-9.79109332e-02 -1.84537962e-01 -7.20943093e-01 3.58984590e-01
-1.62827241e+00 -1.16264141e+00 3.61867934e-01 2.26477814e+00
1.56148708e+00 3.81978244e-01 2.98409611e-01 -1.32783502e-01
4.05742347e-01 -3.28974813e-01 -8.77404928e-01 -2.54837304e-01
-3.13653722e-02 8.79070222e-01 3.56156498e-01 6.50709152e-01
-6.70917392e-01 8.42074275e-01 7.30350828e+00 7.36769319e-01
-9.87141609e-01 -3.12156618e-01 7.64243603e-01 -2.25887056e-02
-7.94702649e-01 4.53877807e-01 -7.44847298e-01 4.86325532e-01
1.23322642e+00 -2.74561234e-02 2.52380192e-01 7.37677336e-01
3.85843396e-01 1.11298468e-02 -1.26373816e+00 4.21940774e-01
-3.06267112e-01 -1.54531324e+00 1.25345051e-01 2.94567317e-01
5.93281984e-01 -7.80942142e-02 3.81807126e-02 3.83785069e-02
9.60985661e-01 -1.29739618e+00 2.29894057e-01 6.50814176e-01
4.91217852e-01 -8.86728525e-01 3.60061824e-01 6.84672832e-01
-4.70033437e-01 4.40177172e-01 -2.72719651e-01 2.59166032e-01
2.05756769e-01 7.70815372e-01 -1.24116313e+00 2.43290305e-01
-2.33130623e-02 3.95217031e-01 -9.54268873e-03 5.81880152e-01
-1.53394967e-01 9.27319944e-01 -5.23253560e-01 -1.08990379e-01
1.65217414e-01 -3.80216956e-01 3.25729877e-01 1.10181749e+00
1.35558769e-01 1.01819612e-01 4.01966512e-01 1.06997514e+00
1.57197416e-01 8.52772146e-02 -4.63985503e-01 -4.63776141e-01
8.71966854e-02 8.01728070e-01 -5.16158283e-01 -3.63625199e-01
-2.72789657e-01 6.80287421e-01 2.09453672e-01 5.82728446e-01
-6.43672466e-01 7.16936812e-02 5.21171510e-01 2.64682591e-01
6.27842963e-01 -3.00427645e-01 -1.01511488e-02 -9.34527636e-01
-3.17354590e-01 -1.31667340e+00 1.73128679e-01 -6.59748733e-01
-1.23670053e+00 1.89190179e-01 -3.23468745e-01 -6.28116548e-01
-6.47318244e-01 -6.16595507e-01 -2.34314054e-01 1.27984631e+00
-9.95058537e-01 -8.56736839e-01 5.42730808e-01 1.18957984e-03
2.53100723e-01 5.83015755e-03 1.02130377e+00 -9.26119909e-02
-1.75815493e-01 4.38003242e-01 3.38009357e-01 -2.32082680e-01
5.92605472e-01 -1.11658084e+00 3.42721462e-01 5.14233232e-01
9.53117236e-02 1.28843057e+00 1.17610502e+00 -9.40646350e-01
-1.28645468e+00 -8.52908134e-01 5.83059251e-01 -6.78607881e-01
4.78152245e-01 -4.72602546e-01 -1.17314172e+00 6.30161762e-01
-1.03646532e-01 -3.77634168e-01 1.11937463e+00 1.26569882e-01
-2.71699816e-01 4.36275125e-01 -1.14660800e+00 4.29084927e-01
8.11702013e-01 -4.66133624e-01 -4.83927190e-01 4.31036055e-01
7.51074016e-01 -1.95086375e-01 -8.90232444e-01 4.09474969e-01
5.08794308e-01 -8.31234932e-01 9.83468771e-01 -1.08975852e+00
3.36725175e-01 -5.02464175e-01 -1.72380626e-01 -1.10816467e+00
-3.84597808e-01 -8.85339558e-01 -8.81161988e-02 8.66577506e-01
9.29529488e-01 -5.01154482e-01 1.07243371e+00 5.86293221e-01
-8.44184533e-02 -9.54249024e-01 -4.13541913e-01 -5.70299327e-01
3.97545010e-01 -2.46136427e-01 5.62528551e-01 5.41388750e-01
-5.46880029e-02 5.61232567e-01 -3.18101913e-01 -1.90767899e-01
5.59680700e-01 9.05702263e-02 8.04811954e-01 -8.77969325e-01
-1.00250089e+00 -2.17725143e-01 -2.34426484e-01 -1.19861829e+00
-1.74646005e-01 -8.29052389e-01 1.48700625e-01 -1.21849644e+00
3.69774580e-01 -3.42319846e-01 -1.91817328e-01 3.01390707e-01
-2.59595662e-01 -1.61807194e-01 -3.27238500e-01 5.62473685e-02
-8.39343816e-02 6.40558541e-01 1.15155101e+00 2.85438839e-02
-2.04114079e-01 -2.35337522e-02 -8.46730471e-01 5.05324721e-01
6.67457938e-01 -3.16651911e-01 -6.82303190e-01 4.40657616e-01
2.92017967e-01 -8.28339458e-02 1.76503003e-01 -5.21969080e-01
-1.44955099e-01 -3.23601484e-01 6.95050001e-01 -5.88539481e-01
2.50085205e-01 -4.67695802e-01 6.55486882e-01 5.41022599e-01
-4.96734858e-01 -4.01433468e-01 -3.21938097e-02 6.95494235e-01
2.42519215e-01 -4.01616305e-01 8.43954325e-01 -3.82086009e-01
-4.49505746e-02 4.00718361e-01 -6.47183597e-01 4.35516657e-03
7.86083281e-01 -2.11395174e-01 2.70413220e-01 -1.90009400e-01
-1.19392824e+00 -5.20509928e-02 8.03298473e-01 -1.38947651e-01
4.84204441e-01 -8.34058583e-01 -6.41529560e-01 3.46489996e-01
2.64244620e-02 -4.68603298e-02 1.92198455e-02 5.19424438e-01
-3.41892689e-01 3.39785427e-01 8.14300776e-02 -7.95272529e-01
-1.01638067e+00 9.01250124e-01 1.90691739e-01 -3.73782903e-01
-3.00105304e-01 6.09620869e-01 5.57957888e-01 -3.27714294e-01
-1.69253141e-01 8.35994631e-03 1.54198751e-01 -2.95399994e-01
4.52392906e-01 -3.41748267e-01 6.25852123e-02 -2.09801167e-01
-8.87333304e-02 2.25334242e-01 -5.15858173e-01 -1.70523494e-01
1.62494802e+00 4.09309477e-01 9.32610780e-02 3.48350108e-01
9.90185916e-01 -2.27387641e-02 -1.50313067e+00 -1.95861369e-01
3.14379409e-02 -5.22206873e-02 -4.94280607e-01 -7.71808684e-01
-3.25016320e-01 7.80475497e-01 3.84261429e-01 -2.55857587e-01
7.23049104e-01 9.36665758e-02 6.44638896e-01 3.65060568e-01
4.77291375e-01 -8.54669809e-01 2.04833865e-01 3.71312618e-01
5.37946999e-01 -8.29499900e-01 -5.05293272e-02 -2.21971571e-01
-3.60864013e-01 8.88656437e-01 8.32831562e-02 -7.56387636e-02
5.10248244e-01 5.31939924e-01 -3.01778227e-01 -2.10368857e-01
-1.04920101e+00 -1.57824859e-01 1.16831809e-01 7.52953172e-01
7.91894019e-01 -8.46470892e-02 -2.40567893e-01 3.90780002e-01
-1.17007330e-01 2.50277489e-01 1.50780842e-01 1.18968892e+00
-7.18495965e-01 -1.80505180e+00 -1.37932912e-01 4.71399248e-01
-5.47384143e-01 -3.23328912e-01 -5.28267980e-01 3.09110880e-01
-9.24971178e-02 5.41868865e-01 -3.65669340e-01 -8.55776370e-02
-2.45830100e-02 5.69754422e-01 8.44174027e-01 -7.83624470e-01
-2.75433093e-01 3.78936857e-01 9.65828672e-02 -1.38672397e-01
-2.90936470e-01 -7.66623557e-01 -1.35965133e+00 -2.29957208e-01
-3.77020627e-01 5.77249289e-01 5.19670129e-01 9.27474678e-01
5.18601298e-01 3.93001735e-01 5.57935059e-01 -6.62067354e-01
-6.32606387e-01 -6.99421287e-01 -1.75876334e-01 2.55877852e-01
2.23148182e-01 -5.04791915e-01 -1.12860933e-01 5.50405204e-01] | [4.719488143920898, 5.587850570678711] |
f402c77c-7a51-4040-be0e-394f00bf7940 | symfm6d-symmetry-aware-multi-directional | 2307.00306 | null | https://arxiv.org/abs/2307.00306v1 | https://arxiv.org/pdf/2307.00306v1.pdf | SyMFM6D: Symmetry-aware Multi-directional Fusion for Multi-View 6D Object Pose Estimation | Detecting objects and estimating their 6D poses is essential for automated systems to interact safely with the environment. Most 6D pose estimators, however, rely on a single camera frame and suffer from occlusions and ambiguities due to object symmetries. We overcome this issue by presenting a novel symmetry-aware multi-view 6D pose estimator called SyMFM6D. Our approach efficiently fuses the RGB-D frames from multiple perspectives in a deep multi-directional fusion network and predicts predefined keypoints for all objects in the scene simultaneously. Based on the keypoints and an instance semantic segmentation, we efficiently compute the 6D poses by least-squares fitting. To address the ambiguity issues for symmetric objects, we propose a novel training procedure for symmetry-aware keypoint detection including a new objective function. Our SyMFM6D network significantly outperforms the state-of-the-art in both single-view and multi-view 6D pose estimation. We furthermore show the effectiveness of our symmetry-aware training procedure and demonstrate that our approach is robust towards inaccurate camera calibration and dynamic camera setups. | ['Gerhard Neumann', 'Ngo Anh Vien', 'Hanna Ziesche', 'Sebastian Koch', 'Fabian Duffhauss'] | 2023-07-01 | null | null | null | null | ['camera-calibration', 'pose-estimation', 'keypoint-detection', '6d-pose-estimation-1', '6d-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 1.23996949e-02 -2.48169437e-01 2.46418100e-02 -4.92243439e-01
-9.75678146e-01 -8.68817925e-01 4.31724608e-01 -5.63769676e-02
-3.54482234e-01 -5.02582416e-02 -1.53961658e-01 2.30543718e-01
-1.11766905e-01 -3.95364821e-01 -1.17823052e+00 -4.41772997e-01
4.55211967e-01 9.30657327e-01 4.86553222e-01 6.13786988e-02
3.80815238e-01 9.23848867e-01 -1.52206016e+00 -8.80538672e-02
5.52969396e-01 1.01882863e+00 4.24136132e-01 6.97943449e-01
2.55048931e-01 1.21341147e-01 -3.22693497e-01 -2.38713890e-01
6.24519587e-01 1.77082479e-01 -5.35439849e-01 5.84653139e-01
9.89807427e-01 -7.23483741e-01 -1.58118963e-01 9.38043535e-01
3.81154597e-01 2.33801872e-01 3.68347853e-01 -1.36056566e+00
1.67300791e-01 -1.35927081e-01 -7.20691204e-01 -1.81189269e-01
5.69511354e-01 -1.20570101e-01 8.44978571e-01 -1.12493503e+00
7.15960026e-01 1.17435360e+00 7.10470140e-01 2.31258050e-01
-1.15669239e+00 -4.90967333e-01 4.30890709e-01 9.14938003e-02
-1.30494475e+00 -3.18550885e-01 9.96928096e-01 -4.02949631e-01
9.82579231e-01 5.00673167e-02 6.73356414e-01 8.87210548e-01
-2.61998884e-02 7.86211610e-01 8.63137484e-01 -2.81065673e-01
1.20156139e-01 -1.88811049e-01 -2.17897892e-01 7.76092649e-01
3.41129273e-01 -1.55070737e-01 -6.65673256e-01 4.21674810e-02
1.05912733e+00 3.47413719e-01 -5.06418198e-02 -1.39343882e+00
-1.48934662e+00 5.40236115e-01 2.99462408e-01 -2.85289109e-01
-1.99926183e-01 2.03823432e-01 6.53437451e-02 -2.17738822e-01
4.27820832e-01 4.77911562e-01 -7.94328511e-01 -7.80062601e-02
-6.59069955e-01 4.51649487e-01 4.07486737e-01 1.17865491e+00
7.38267243e-01 -3.82918030e-01 5.22325456e-01 5.32119215e-01
3.61385614e-01 7.60920525e-01 -1.15969039e-01 -1.45421970e+00
5.88900506e-01 7.89601743e-01 3.79675478e-01 -1.07641995e+00
-6.99995279e-01 -4.32386875e-01 -2.81979769e-01 1.37276590e-01
3.45284462e-01 2.31984258e-01 -8.75798643e-01 1.40399563e+00
8.68036807e-01 8.29122588e-02 -2.37958208e-01 1.13578498e+00
6.34209991e-01 2.82944381e-01 -5.68190336e-01 7.00211897e-02
1.21958995e+00 -9.79789078e-01 -2.47767627e-01 -4.84997392e-01
3.52678239e-01 -9.45377171e-01 6.51632547e-01 4.74316448e-01
-1.06520760e+00 -3.89867634e-01 -1.06089437e+00 -3.45972806e-01
-1.09717079e-01 3.40471536e-01 5.45831025e-01 3.02634627e-01
-6.50986612e-01 2.63324916e-01 -1.09204435e+00 -3.18605721e-01
1.58562317e-01 5.90493023e-01 -7.69349515e-01 -1.42726868e-01
-4.03337181e-01 1.02407181e+00 3.20127010e-01 8.96360949e-02
-8.21821868e-01 -6.64283097e-01 -1.00054920e+00 -2.03773811e-01
9.56896782e-01 -1.02729988e+00 1.40265059e+00 -3.08610201e-01
-1.38323665e+00 1.06610298e+00 -3.03489298e-01 -7.73379207e-02
6.92856491e-01 -7.43224621e-01 2.46290341e-01 3.68356228e-01
2.35435590e-01 6.11910343e-01 7.64275253e-01 -1.59113514e+00
-6.81193769e-01 -8.38054419e-01 2.14559123e-01 7.24720716e-01
1.51380643e-01 -2.99830705e-01 -1.00354862e+00 -2.51373440e-01
9.11896586e-01 -1.29906476e+00 -3.62456799e-01 3.14564496e-01
-4.27364498e-01 1.47898152e-01 8.87873530e-01 -3.74763846e-01
3.98718715e-01 -1.93015575e+00 4.44276184e-01 1.76734582e-01
2.17145592e-01 4.25412841e-02 1.25731990e-01 5.77348471e-02
1.11964673e-01 -4.71004367e-01 6.31379634e-02 -8.19996119e-01
-3.00771520e-02 3.53022695e-01 -2.59460174e-02 7.41924345e-01
-5.27549796e-02 5.29181302e-01 -8.09056878e-01 -3.10455590e-01
7.30441570e-01 6.49559915e-01 -6.13563895e-01 3.09768051e-01
-3.06581318e-01 4.70784515e-01 -4.51005101e-01 7.13522255e-01
9.48258519e-01 -2.40339503e-01 -4.11948673e-02 -6.29341364e-01
-1.23438291e-01 1.47606269e-01 -1.47961485e+00 2.27969027e+00
-5.07225037e-01 2.56641954e-01 2.61353612e-01 -6.69576049e-01
6.66409492e-01 8.19036663e-02 7.35295177e-01 -2.16389343e-01
1.97313890e-01 3.68333727e-01 -5.62222421e-01 -3.05325568e-01
5.09514213e-01 9.51601788e-02 -4.45693657e-02 2.71179557e-01
1.90908551e-01 -5.34119189e-01 -7.55213797e-02 7.58643448e-02
6.93773270e-01 5.58282793e-01 2.90782422e-01 2.19815984e-01
3.90993804e-01 -1.46626681e-01 4.57108527e-01 4.06402647e-01
2.67115030e-02 9.48550045e-01 2.43100852e-01 -5.57315528e-01
-1.10521555e+00 -1.13486671e+00 7.08320960e-02 5.42778552e-01
5.58622003e-01 -4.59621400e-01 -5.87766886e-01 -7.02591896e-01
3.09585541e-01 4.19681162e-01 -2.56308943e-01 6.02804236e-02
-6.19563043e-01 -4.51896071e-01 -1.91663057e-01 7.13862836e-01
3.47032249e-01 -1.22866683e-01 -1.03614855e+00 3.79663371e-02
-3.27986628e-01 -1.81804037e+00 -4.95800197e-01 2.92943567e-01
-9.41889048e-01 -1.27240944e+00 -5.93679368e-01 -4.67048436e-01
7.86081135e-01 8.41008544e-01 9.16888535e-01 -4.15039629e-01
-1.19286910e-01 7.38368869e-01 -1.52412489e-01 -2.57309407e-01
-6.74523711e-02 -5.30059375e-02 2.26212591e-01 -4.15784270e-02
1.90586999e-01 -5.48795879e-01 -8.00447106e-01 6.56097114e-01
-4.45846885e-01 2.10178494e-01 5.40077865e-01 3.36551785e-01
1.03337407e+00 -9.05762315e-02 -2.75123388e-01 -4.03823614e-01
-2.93304145e-01 1.39807329e-01 -1.11613333e+00 4.46046218e-02
-1.69804946e-01 -2.31871959e-02 1.93792984e-01 -2.02919111e-01
-8.46118212e-01 8.01889122e-01 6.68815970e-02 -9.58374560e-01
-2.93550014e-01 -5.61947562e-03 -4.56119865e-01 -4.52903390e-01
2.19079837e-01 -1.02148823e-01 -6.17034212e-02 -6.01336539e-01
3.93792540e-01 2.54568428e-01 6.39872849e-01 -5.03307939e-01
9.93621409e-01 8.44661534e-01 3.78213257e-01 -6.82711124e-01
-1.21667171e+00 -9.31081772e-01 -1.19762683e+00 -3.94048721e-01
1.01776898e+00 -1.27851772e+00 -9.63799596e-01 6.34949446e-01
-1.41205764e+00 1.75302565e-01 4.85688960e-03 6.67322636e-01
-7.73951948e-01 4.08598453e-01 -1.57976359e-01 -7.13002086e-01
-4.82522361e-02 -1.41977239e+00 1.92292356e+00 -8.60686526e-02
-4.93602939e-02 -5.79296708e-01 -1.94322377e-01 7.34813511e-01
-3.07501882e-01 4.88433361e-01 4.45992857e-01 -3.08617800e-01
-1.02639520e+00 -2.97070712e-01 -1.19359307e-01 9.81757045e-02
-8.59259069e-03 -6.86999485e-02 -8.17333698e-01 -1.76835746e-01
2.15930596e-01 -1.63221344e-01 5.89370489e-01 4.65582043e-01
1.00252914e+00 9.32087228e-02 -3.54005307e-01 9.86823142e-01
1.47212064e+00 -6.48524659e-03 -6.84759095e-02 6.23315036e-01
1.12102640e+00 5.20938814e-01 8.15471053e-01 3.81272137e-01
6.56795621e-01 1.04583466e+00 1.01515257e+00 7.99301639e-02
3.15384269e-01 -2.97116369e-01 1.02715254e-01 6.43794060e-01
-1.20085880e-01 -1.08243555e-01 -9.51087058e-01 3.37011099e-01
-1.82013690e+00 -4.49130327e-01 -2.88448393e-01 2.30949211e+00
2.08622798e-01 2.14880556e-01 1.24645866e-01 1.99695118e-02
5.10719359e-01 9.49160978e-02 -8.37399244e-01 1.63924649e-01
8.03652629e-02 -1.40023366e-01 7.99744070e-01 5.00497937e-01
-1.30296671e+00 8.51135492e-01 5.69178247e+00 2.73045570e-01
-8.79759789e-01 5.85453212e-03 6.53262511e-02 -2.67744064e-01
-1.56126305e-01 7.50223501e-03 -1.06487811e+00 -4.96952161e-02
2.16685444e-01 5.12148082e-01 1.55830517e-01 1.04600275e+00
-5.87819815e-02 -3.61054659e-01 -1.22331786e+00 1.43017280e+00
4.64823514e-01 -1.07735085e+00 -1.22474931e-01 8.70902836e-02
8.25078785e-01 2.76661903e-01 3.89858987e-03 -4.32639062e-01
-4.18533087e-02 -3.78781199e-01 9.60380614e-01 2.54874915e-01
4.28945988e-01 -7.85900235e-01 4.71261322e-01 4.96232271e-01
-1.07528174e+00 -1.02959797e-02 -1.89219117e-01 1.34758875e-01
4.61009026e-01 5.50178885e-01 -7.86467135e-01 6.65386379e-01
7.00774670e-01 7.73177028e-01 -6.08903170e-01 9.08041596e-01
-2.17248410e-01 -2.69966692e-01 -6.91843092e-01 3.94890964e-01
1.43094078e-01 -1.72272116e-01 8.01852822e-01 5.43618977e-01
5.25005698e-01 1.14556350e-01 3.01557899e-01 6.57813370e-01
1.54410705e-01 -4.04859215e-01 -3.64064693e-01 2.99029678e-01
2.99413353e-01 1.24234986e+00 -1.05515921e+00 -2.01520827e-02
-4.37631458e-01 1.15638578e+00 8.10337216e-02 1.48736626e-01
-7.62061179e-01 1.33937910e-01 7.44283676e-01 1.97786063e-01
5.65337479e-01 -7.86261439e-01 -1.18278593e-01 -1.56432045e+00
4.46187645e-01 -6.59260035e-01 2.62205093e-03 -1.14056253e+00
-8.67032647e-01 3.21163207e-01 3.87847126e-01 -1.43402362e+00
-1.48261249e-01 -9.22660470e-01 2.73353467e-03 4.38212872e-01
-1.36174226e+00 -1.38964927e+00 -4.47894186e-01 4.99729961e-01
7.00570643e-01 2.09520355e-01 5.99992990e-01 3.77072357e-02
-1.83907866e-01 2.11308405e-01 -6.96879029e-02 -1.92568824e-01
7.30059326e-01 -1.28417075e+00 5.39269745e-01 8.71072352e-01
2.36701474e-01 5.20095706e-01 6.21640742e-01 -4.33810949e-01
-1.87849402e+00 -8.03546727e-01 6.65728509e-01 -8.15386236e-01
3.40596378e-01 -7.38882661e-01 -4.62678164e-01 8.95288885e-01
-4.68162268e-01 1.99479133e-01 2.84615040e-01 1.70739498e-02
-3.29859823e-01 -2.47386515e-01 -9.31840837e-01 3.83708626e-01
1.34025586e+00 -4.37559873e-01 -4.59801912e-01 4.64224577e-01
8.01199377e-01 -1.24133527e+00 -7.67794132e-01 6.57299638e-01
6.26660883e-01 -1.10647976e+00 1.36036253e+00 -1.08958542e-01
1.03934549e-01 -5.63392401e-01 -3.49161208e-01 -9.94238079e-01
2.34583393e-02 -4.89565372e-01 -2.51435578e-01 8.19225967e-01
-9.01572779e-02 -4.28715348e-01 1.01416397e+00 5.06829679e-01
-2.04984739e-01 -5.31487465e-01 -1.00533926e+00 -6.48679972e-01
-4.96905804e-01 -6.65662646e-01 4.30771023e-01 5.80677450e-01
-6.33475900e-01 3.30652088e-01 -3.04475874e-01 7.92682886e-01
9.31992650e-01 5.24747908e-01 1.29594660e+00 -1.32114577e+00
-2.71050274e-01 -7.48060122e-02 -7.36533403e-01 -1.47172856e+00
9.27598029e-02 -4.37601268e-01 7.52953887e-02 -1.41835904e+00
2.25824237e-01 -1.20922454e-01 2.25391328e-01 1.42538846e-01
7.26005435e-02 2.59522915e-01 2.71533340e-01 1.12714864e-01
-8.38741183e-01 3.53970557e-01 1.34161186e+00 2.23247737e-01
-1.03803083e-01 1.34383246e-01 -2.64408082e-01 1.21062303e+00
3.45745444e-01 -4.18671340e-01 -3.89487535e-01 -7.11927772e-01
5.19189298e-01 1.29819706e-01 7.88844228e-01 -1.03803229e+00
2.60724872e-01 -8.08908418e-02 6.51560366e-01 -1.24207056e+00
8.93913686e-01 -1.28850174e+00 1.69632778e-01 1.25070557e-01
5.86508140e-02 1.98948190e-01 3.90892997e-02 6.55743003e-01
2.22053826e-02 3.08654197e-02 5.40334582e-01 -4.21416581e-01
-7.20203876e-01 4.65235740e-01 -7.24605564e-03 -1.63227856e-01
1.12753141e+00 -4.35518384e-01 9.21814963e-02 -2.94284672e-01
-7.83322394e-01 3.64973038e-01 9.26108539e-01 4.95513439e-01
7.25168347e-01 -1.17080235e+00 -2.02512518e-01 5.01870036e-01
2.65200913e-01 8.93241942e-01 2.08869338e-01 8.78144979e-01
-7.71552801e-01 4.02232438e-01 5.70325330e-02 -1.29632068e+00
-1.37047780e+00 4.09633398e-01 4.88692164e-01 1.37433231e-01
-5.70609152e-01 8.77878606e-01 3.29065025e-01 -8.30594480e-01
2.53632873e-01 -6.16120160e-01 1.46115676e-01 -1.66296527e-01
1.48411602e-01 4.82073009e-01 3.17930728e-01 -1.02675915e+00
-5.13446927e-01 1.29140592e+00 -6.18200488e-02 -1.06128395e-01
1.55157280e+00 -3.15093845e-01 1.49228107e-02 4.39338714e-01
1.38925231e+00 -3.00085172e-02 -1.71538413e+00 -1.31557748e-01
-3.65433395e-01 -7.67439663e-01 7.58429524e-03 -4.66347486e-01
-9.65601206e-01 7.72318006e-01 4.55689251e-01 -4.91650701e-01
9.63983595e-01 7.58818984e-02 8.00306976e-01 5.82891107e-01
5.64656794e-01 -1.11189353e+00 2.39858225e-01 5.03978252e-01
7.46781230e-01 -1.34028077e+00 3.42939705e-01 -7.61289954e-01
-3.69797915e-01 1.21667290e+00 6.56537414e-01 -2.15920404e-01
4.48069870e-01 7.63737485e-02 5.74974269e-02 -2.48437136e-01
-2.85413861e-01 8.47993568e-02 5.19157410e-01 4.86042887e-01
-1.13982946e-01 -2.64831424e-01 3.27176481e-01 2.33849958e-02
-1.75139681e-01 -3.94711554e-01 3.51761132e-01 1.02533054e+00
-2.64375210e-01 -9.84514594e-01 -6.09106719e-01 -6.61495700e-02
-2.69328594e-01 4.50198919e-01 -4.46778834e-01 9.59686041e-01
2.41287097e-01 6.36921644e-01 1.18524916e-01 -2.85278201e-01
6.43620193e-01 -7.46663287e-02 1.03085518e+00 -5.96957088e-01
-1.70967788e-01 5.49026787e-01 -1.94150940e-01 -1.08589554e+00
-8.74437690e-01 -9.85849977e-01 -1.09092057e+00 2.20205672e-02
-6.03694022e-01 -4.22622621e-01 9.51847196e-01 1.05873382e+00
4.04739320e-01 2.57179618e-01 4.63723898e-01 -1.45596790e+00
-5.35141110e-01 -3.54352474e-01 -4.16331470e-01 2.50202507e-01
6.28782213e-01 -9.92766798e-01 -3.60160679e-01 -7.67762735e-02] | [7.530863285064697, -2.5684778690338135] |
ccf88a33-ecb8-4bac-a3e2-9018a18ae1fc | deep-point-wise-prediction-for-action | 1909.07725 | null | https://arxiv.org/abs/1909.07725v1 | https://arxiv.org/pdf/1909.07725v1.pdf | Deep Point-wise Prediction for Action Temporal Proposal | Detecting actions in videos is an important yet challenging task. Previous works usually utilize (a) sliding window paradigms, or (b) per-frame action scoring and grouping to enumerate the possible temporal locations. Their performances are also limited to the designs of sliding windows or grouping strategies. In this paper, we present a simple and effective method for temporal action proposal generation, named Deep Point-wise Prediction (DPP). DPP simultaneously predicts the action existing possibility and the corresponding temporal locations, without the utilization of any handcrafted sliding window or grouping. The whole system is end-to-end trained with joint loss of temporal action proposal classification and location prediction. We conduct extensive experiments to verify its effectiveness, generality and robustness on standard THUMOS14 dataset. DPP runs more than 1000 frames per second, which largely satisfies the real-time requirement. The code is available at https://github.com/liluxuan1997/DPP. | ['Huaping Liu', 'Luxuan Li', 'Fuchun Sun', 'Tao Kong'] | 2019-09-17 | null | null | null | null | ['temporal-action-proposal-generation'] | ['computer-vision'] | [ 1.78796723e-01 -4.56109673e-01 -6.59098268e-01 -1.54350579e-01
-7.77014375e-01 -4.54520941e-01 6.53019309e-01 -3.54438871e-01
-3.24451715e-01 5.77494204e-01 2.85054892e-01 -1.41976476e-01
4.01528999e-02 -4.55829501e-01 -4.21974152e-01 -7.22893476e-01
-3.85973811e-01 -8.16520825e-02 8.75071347e-01 1.30886823e-01
3.12660575e-01 2.97274232e-01 -1.51100445e+00 4.49137181e-01
5.22415936e-01 1.21219623e+00 2.57207930e-01 8.26029360e-01
3.88778657e-01 9.70660627e-01 -3.89715374e-01 -1.22596741e-01
5.53572237e-01 -6.27741039e-01 -4.78731424e-01 2.60162234e-01
2.37777174e-01 -7.64472663e-01 -6.42153263e-01 6.82024062e-01
5.69972575e-01 3.01455140e-01 2.77646273e-01 -1.31247282e+00
-1.44488111e-01 3.27718198e-01 -5.50148964e-01 4.73823041e-01
4.72164541e-01 5.34886599e-01 9.60378408e-01 -8.49018276e-01
5.13098001e-01 9.47653174e-01 5.38954198e-01 4.70926255e-01
-6.67615116e-01 -6.38774753e-01 2.61079520e-01 6.17547035e-01
-1.36759472e+00 -6.25138402e-01 7.33557761e-01 -3.31030190e-01
8.82436693e-01 1.65088609e-01 7.23897934e-01 1.03624833e+00
9.32032764e-02 1.17970574e+00 6.12608314e-01 -1.40750587e-01
2.14877725e-01 -4.77235585e-01 -3.65742236e-01 6.20217264e-01
-2.35078514e-01 -4.18486912e-03 -5.76399267e-01 -1.07602231e-01
9.57441092e-01 3.64006758e-01 -2.56231278e-01 -1.62378311e-01
-1.52561653e+00 4.60395038e-01 9.43494067e-02 1.46265477e-01
-4.19803590e-01 1.76211521e-01 5.81931412e-01 7.48207867e-02
2.78162301e-01 3.55757400e-02 -4.81064081e-01 -5.60928404e-01
-1.05159318e+00 3.60572189e-01 1.37592897e-01 9.15897667e-01
3.54513109e-01 -8.08961838e-02 -5.16283870e-01 6.23646438e-01
6.54371679e-02 1.63641170e-01 7.03608036e-01 -1.26448739e+00
6.14218235e-01 4.35323536e-01 3.52323741e-01 -8.68663490e-01
-2.04289511e-01 -9.63805895e-03 -5.43000162e-01 6.99401051e-02
4.35717076e-01 -3.49169940e-01 -6.58098817e-01 1.48421705e+00
5.29793739e-01 6.49393439e-01 -2.65813410e-01 9.50613916e-01
5.22522449e-01 7.46849477e-01 1.29319906e-01 -3.08581412e-01
1.13150311e+00 -1.23302746e+00 -5.69777668e-01 1.89825390e-02
6.73216224e-01 -8.14623177e-01 9.14303124e-01 4.72311437e-01
-1.22379422e+00 -7.17653573e-01 -6.98865294e-01 -2.14802843e-04
1.57184243e-01 6.77271962e-01 5.79893112e-01 2.16544509e-01
-7.39872932e-01 7.51792014e-01 -1.15220773e+00 -2.01335296e-01
5.48089027e-01 2.36581311e-01 -2.37729788e-01 2.46242404e-01
-1.05988193e+00 3.87264282e-01 4.36753690e-01 -1.13603603e-02
-9.10870135e-01 -3.12613457e-01 -6.86460853e-01 -2.69302651e-02
6.22082353e-01 -3.09221983e-01 1.47310829e+00 -1.02098536e+00
-1.67716229e+00 4.33592647e-01 -3.46968114e-01 -5.81384897e-01
8.55768263e-01 -4.47383910e-01 -3.01687300e-01 3.40422839e-01
7.55278319e-02 7.71736562e-01 7.79550791e-01 -4.60403115e-01
-1.04182744e+00 -6.92966133e-02 1.56393856e-01 1.89521357e-01
-3.08320850e-01 2.81307995e-01 -7.47177303e-01 -1.02731192e+00
7.38835102e-03 -9.94491279e-01 -2.62199879e-01 2.54568100e-01
-3.15516770e-01 -5.76236427e-01 6.46240294e-01 -6.33097172e-01
1.56790447e+00 -2.25100350e+00 -1.40355304e-01 -3.24617535e-01
-2.56245673e-01 4.20857996e-01 -1.93673417e-01 4.51726854e-01
-7.02516809e-02 -1.90335855e-01 5.01215234e-02 -2.06438422e-01
1.80917848e-02 -1.36039868e-01 -3.49187285e-01 5.26339293e-01
1.68440476e-01 7.31837928e-01 -8.61382723e-01 -7.07983434e-01
4.96607661e-01 2.71413445e-01 -5.09271264e-01 1.83417886e-01
-2.00296566e-01 3.66974264e-01 -5.81764638e-01 8.71857941e-01
3.58312011e-01 -1.38731688e-01 2.47135758e-01 -2.48913601e-01
-2.90579289e-01 5.01343727e-01 -1.18662047e+00 1.56208599e+00
-5.18479533e-02 5.91202915e-01 -5.13061404e-01 -8.03589761e-01
6.47387385e-01 4.44115072e-01 8.97243142e-01 -6.64402604e-01
2.36131437e-02 1.40621243e-02 -1.84229523e-01 -6.78518951e-01
5.34879625e-01 6.57630190e-02 7.22366869e-02 3.35059375e-01
-9.38183591e-02 5.49281657e-01 6.02310002e-01 -3.45720574e-02
1.44571185e+00 6.88281476e-01 7.65953064e-01 3.41263920e-01
4.04422522e-01 -4.21042182e-02 1.02010155e+00 4.40290183e-01
-6.64470077e-01 7.88266122e-01 5.69486499e-01 -5.24233639e-01
-9.80908573e-01 -9.38156307e-01 1.53192893e-01 1.25607538e+00
3.46987426e-01 -6.05401158e-01 -3.81556392e-01 -8.85943592e-01
-2.36264378e-01 4.10185993e-01 -3.00273061e-01 4.41297218e-02
-8.21330428e-01 -4.10473466e-01 3.97897452e-01 7.56328404e-01
5.71877003e-01 -1.34843993e+00 -8.63410413e-01 2.67587394e-01
-4.55341965e-01 -1.04863358e+00 -7.15734422e-01 -3.73207122e-01
-8.74056339e-01 -1.20652556e+00 -9.46032047e-01 -5.39223790e-01
4.82574612e-01 5.38143277e-01 6.49020970e-01 -8.24732482e-02
-2.25255415e-01 7.66363889e-02 -7.00157702e-01 3.21856886e-02
6.59665391e-02 -1.77667603e-01 -4.40114103e-02 5.71318865e-02
2.67823309e-01 -5.07901788e-01 -1.09911525e+00 7.89731383e-01
-6.25415266e-01 1.90693378e-01 6.81503832e-01 4.65329468e-01
5.83472729e-01 1.21518292e-01 3.82991046e-01 -2.38451257e-01
1.62792683e-01 -4.80872571e-01 -5.80376625e-01 2.48985618e-01
4.74479534e-02 -3.34756285e-01 5.16427517e-01 -6.26030803e-01
-1.05066514e+00 4.19978648e-01 -1.40880108e-01 -4.69025880e-01
-3.55227292e-01 1.74143553e-01 -6.16768003e-02 1.74106151e-01
4.58339810e-01 5.14385283e-01 -1.66542113e-01 -4.11349356e-01
1.14628173e-01 4.72719282e-01 3.76668632e-01 -1.92863211e-01
3.62367779e-01 5.37994325e-01 -2.85841078e-01 -4.72486764e-01
-5.25446177e-01 -6.01840913e-01 -5.71753800e-01 -5.23853302e-01
6.78194642e-01 -1.01805842e+00 -7.40319490e-01 4.82393980e-01
-9.52481210e-01 -6.31152034e-01 1.25138881e-02 7.21347451e-01
-7.02329636e-01 6.92646682e-01 -6.93544030e-01 -7.39347696e-01
-1.96196914e-01 -9.43716645e-01 1.17609549e+00 1.24082014e-01
-3.58899176e-01 -4.41136181e-01 -8.72655492e-03 4.14819539e-01
-8.64428375e-03 2.00613126e-01 1.25871062e-01 -2.39261702e-01
-7.07913876e-01 -3.97254795e-01 -8.00868273e-02 2.10080072e-01
1.67061567e-01 3.23816359e-01 -4.38829511e-01 -9.17298794e-02
-2.77375609e-01 -2.59790272e-01 8.74271989e-01 6.07810557e-01
1.41607070e+00 -3.82444352e-01 -2.53533512e-01 3.07835311e-01
1.12918842e+00 5.44796705e-01 8.54158282e-01 3.10714543e-01
3.93889725e-01 3.30966592e-01 1.23798287e+00 9.81641471e-01
3.01821291e-01 1.10723388e+00 3.46990258e-01 2.56533206e-01
-8.26837793e-02 -7.90779665e-02 8.28860641e-01 2.99447030e-01
-4.60268080e-01 -3.91603380e-01 -6.68511212e-01 5.67840755e-01
-2.20912743e+00 -1.49414504e+00 6.96167424e-02 2.22485328e+00
6.78188860e-01 2.96590179e-01 5.59750915e-01 1.06150955e-01
7.93240130e-01 4.15553510e-01 -5.33790350e-01 1.20838031e-01
1.35994449e-01 -1.20553575e-01 4.38903898e-01 6.64507747e-02
-1.54440057e+00 8.53905916e-01 5.59047842e+00 1.23961771e+00
-1.06389880e+00 2.58982629e-02 6.83453143e-01 -5.32022238e-01
3.84092271e-01 -9.76203158e-02 -8.16973209e-01 8.40048313e-01
5.71736336e-01 1.85111240e-02 6.85637370e-02 8.95672679e-01
6.23564899e-01 -3.70312870e-01 -8.06459248e-01 9.04168844e-01
-1.60423934e-01 -1.27361369e+00 -8.97493437e-02 -2.28132546e-01
5.42310536e-01 -2.71618385e-02 -1.22466788e-01 3.13218534e-01
6.42371774e-02 -4.67029572e-01 8.49585295e-01 4.21865880e-01
6.04912519e-01 -5.44599950e-01 4.85671550e-01 3.20011854e-01
-1.37661397e+00 -3.30576628e-01 -1.51348010e-01 -2.80871868e-01
4.93370801e-01 3.07656646e-01 -5.89050233e-01 3.37721080e-01
6.36100292e-01 1.15678549e+00 -4.28317070e-01 1.34852457e+00
-3.76787633e-01 7.50352740e-01 -2.84955651e-01 -5.42125758e-03
2.88420498e-01 -5.87179745e-03 4.77614403e-01 1.03323233e+00
7.16097355e-01 3.16283345e-01 2.66506195e-01 1.12135388e-01
2.01893076e-01 -6.32290095e-02 -2.08593190e-01 3.15247625e-02
4.30334687e-01 1.13248920e+00 -8.30489814e-01 -4.03549671e-01
-5.92610359e-01 9.29861009e-01 -6.21066913e-02 1.33549675e-01
-1.35242677e+00 -2.03820378e-01 6.12637639e-01 2.29631096e-01
6.87646925e-01 -3.29487115e-01 2.10403483e-02 -1.15158951e+00
3.26047361e-01 -7.38402069e-01 5.50662816e-01 -7.79105902e-01
-9.56929386e-01 2.76500076e-01 -1.70735235e-03 -2.04499602e+00
-1.04085647e-01 -3.23109776e-01 -8.83099973e-01 1.23732418e-01
-1.07506919e+00 -8.10894608e-01 -2.73215175e-01 5.61167955e-01
1.14612687e+00 -7.26700351e-02 3.10355037e-01 5.07776916e-01
-8.82563472e-01 4.31872308e-01 -1.32534534e-01 9.92181748e-02
8.36726367e-01 -8.27666461e-01 3.17724347e-01 9.79263842e-01
-6.74454868e-02 2.15848982e-01 6.77280068e-01 -5.31926990e-01
-1.03104639e+00 -1.28324211e+00 9.74158227e-01 -2.13265881e-01
7.03660131e-01 -5.48098311e-02 -6.22313380e-01 5.89586675e-01
4.39480413e-03 1.70515105e-01 5.90376258e-01 -2.35619888e-01
1.53433338e-01 -2.34014690e-01 -7.82969058e-01 9.02845740e-01
1.28751898e+00 -8.64487067e-02 -3.03810537e-01 5.10625124e-01
4.67910677e-01 -3.67646784e-01 -5.68749905e-01 5.58443308e-01
7.84608841e-01 -1.30584455e+00 9.67456281e-01 -4.52289611e-01
7.98886299e-01 -5.29567897e-01 6.23211302e-02 -6.07791066e-01
-3.94495904e-01 -8.47774208e-01 -3.06628168e-01 9.62139010e-01
2.11470708e-01 -2.43484363e-01 9.24095571e-01 4.97947246e-01
-1.55817971e-01 -1.16783714e+00 -9.71779764e-01 -8.73970270e-01
-6.42005503e-01 -5.10726511e-01 4.16354865e-01 6.81280732e-01
8.69956836e-02 -2.33574703e-01 -7.58780956e-01 3.92315499e-02
2.52835691e-01 2.64335394e-01 8.62130404e-01 -4.22472209e-01
-4.35275048e-01 -5.38202286e-01 -6.50974572e-01 -1.45625734e+00
-1.73073173e-01 -2.13226557e-01 6.17873110e-02 -1.30684555e+00
1.29649788e-01 -2.16382295e-01 -4.53931510e-01 6.90876365e-01
-1.56619668e-01 4.07679617e-01 1.95622280e-01 3.35228205e-01
-1.34231186e+00 7.76034951e-01 1.06578398e+00 2.21927971e-01
-2.00799331e-01 3.82879525e-01 -1.53464869e-01 8.91984940e-01
1.08314264e+00 -3.24161142e-01 -4.80260074e-01 -1.89492017e-01
-9.35458764e-02 3.27981114e-01 5.84641933e-01 -1.25225270e+00
1.67580247e-01 -4.85593200e-01 2.70592213e-01 -7.61948407e-01
4.68311518e-01 -5.42810321e-01 1.14969410e-01 4.53456610e-01
-4.45390999e-01 -7.45939538e-02 -4.67604995e-02 5.33431232e-01
-2.69184381e-01 3.93101312e-02 5.97747505e-01 3.17462459e-02
-1.18518782e+00 4.95165795e-01 -3.30645651e-01 -2.57831186e-01
1.54153037e+00 -4.61124927e-01 -1.66636184e-01 -2.81393588e-01
-7.72632718e-01 2.67588437e-01 3.46000761e-01 3.99802297e-01
6.02801621e-01 -1.57998550e+00 -5.01522243e-01 -1.44791558e-01
9.34187397e-02 -3.80183905e-01 6.35410488e-01 1.13809860e+00
-5.03856003e-01 3.38296235e-01 -1.91610888e-01 -3.99692684e-01
-1.66978371e+00 4.95182395e-01 1.61402121e-01 -1.84521884e-01
-7.31430829e-01 7.80812085e-01 1.82728797e-01 1.92008793e-01
3.00098896e-01 -2.35849753e-01 -2.12857902e-01 -6.18919432e-02
7.19300091e-01 6.22801006e-01 -4.44255799e-01 -5.11376858e-01
-3.75038296e-01 4.24575537e-01 5.57088815e-02 8.59755874e-02
1.16514826e+00 -9.39459428e-02 2.40307510e-01 1.98669806e-01
8.58466804e-01 -2.98990965e-01 -1.88117456e+00 -2.21136585e-01
-1.28667235e-01 -8.29239070e-01 -2.89316207e-01 -3.98228139e-01
-1.03742695e+00 6.53664827e-01 4.39898133e-01 1.19309433e-01
1.30386961e+00 -1.13840692e-01 1.14232445e+00 1.26220450e-01
4.07159150e-01 -1.10597229e+00 2.82324880e-01 2.58651257e-01
8.45493734e-01 -1.29118609e+00 1.92009211e-01 -2.85742700e-01
-8.62195015e-01 1.07353222e+00 6.92556024e-01 -1.73511818e-01
4.75058556e-01 -1.16260469e-01 -1.84112862e-01 1.65787607e-01
-9.96964514e-01 -2.71302283e-01 2.32279390e-01 2.03976735e-01
4.66978490e-01 6.84362799e-02 -5.87974727e-01 3.72732192e-01
2.22294465e-01 2.42775008e-01 3.53808761e-01 1.01472628e+00
-5.40833771e-01 -9.72262442e-01 -1.39771536e-01 3.43925953e-01
-4.87402350e-01 1.44043207e-01 -1.44412726e-01 5.71322560e-01
3.62542540e-01 9.25057173e-01 3.29782218e-02 -4.13619220e-01
1.76677525e-01 -2.36115873e-01 4.22716945e-01 -4.21784550e-01
-1.90543413e-01 3.19918215e-01 1.08379155e-01 -9.66491163e-01
-5.68607032e-01 -1.08945751e+00 -1.24959755e+00 -2.10287943e-01
-1.95021674e-01 -2.23074511e-01 1.19539879e-01 7.75853872e-01
6.41020417e-01 2.81061202e-01 7.66526461e-01 -1.07898092e+00
-4.20607418e-01 -9.51650143e-01 -3.47616524e-01 4.18039620e-01
1.45467401e-01 -7.01816022e-01 -2.79724714e-03 3.36196363e-01] | [8.371603965759277, 0.4133458435535431] |
a5d41da8-a844-4a60-8a94-bef43eebe509 | learning-feature-matching-via-matchable | 2307.01447 | null | https://arxiv.org/abs/2307.01447v1 | https://arxiv.org/pdf/2307.01447v1.pdf | Learning Feature Matching via Matchable Keypoint-Assisted Graph Neural Network | Accurately matching local features between a pair of images is a challenging computer vision task. Previous studies typically use attention based graph neural networks (GNNs) with fully-connected graphs over keypoints within/across images for visual and geometric information reasoning. However, in the context of feature matching, considerable keypoints are non-repeatable due to occlusion and failure of the detector, and thus irrelevant for message passing. The connectivity with non-repeatable keypoints not only introduces redundancy, resulting in limited efficiency, but also interferes with the representation aggregation process, leading to limited accuracy. Targeting towards high accuracy and efficiency, we propose MaKeGNN, a sparse attention-based GNN architecture which bypasses non-repeatable keypoints and leverages matchable ones to guide compact and meaningful message passing. More specifically, our Bilateral Context-Aware Sampling Module first dynamically samples two small sets of well-distributed keypoints with high matchability scores from the image pair. Then, our Matchable Keypoint-Assisted Context Aggregation Module regards sampled informative keypoints as message bottlenecks and thus constrains each keypoint only to retrieve favorable contextual information from intra- and inter- matchable keypoints, evading the interference of irrelevant and redundant connectivity with non-repeatable ones. Furthermore, considering the potential noise in initial keypoints and sampled matchable ones, the MKACA module adopts a matchability-guided attentional aggregation operation for purer data-dependent context propagation. By these means, we achieve the state-of-the-art performance on relative camera estimation, fundamental matrix estimation, and visual localization, while significantly reducing computational and memory complexity compared to typical attentional GNNs. | ['Jiayi Ma', 'Zizhuo Li'] | 2023-07-04 | null | null | null | null | ['visual-localization'] | ['computer-vision'] | [ 3.86675596e-02 -6.51284158e-02 -1.67359114e-01 5.28293326e-02
-6.62685931e-01 -4.10163969e-01 5.52859843e-01 3.52328956e-01
-3.84369045e-01 3.22304010e-01 6.83394447e-02 4.20260280e-02
-3.80435258e-01 -9.02878642e-01 -9.83024120e-01 -5.72914958e-01
-1.12888396e-01 2.77213633e-01 4.80387062e-01 -3.67617747e-03
2.97760457e-01 8.91204715e-01 -1.70520639e+00 7.92351216e-02
6.92598760e-01 1.29774785e+00 4.59983885e-01 5.89603424e-01
-1.22872636e-01 5.70565224e-01 -3.90685439e-01 -2.15746298e-01
5.07476509e-01 -2.11017981e-01 -4.95801359e-01 -1.95816413e-01
1.19980264e+00 -4.41456288e-01 -7.78165460e-01 1.20269096e+00
3.78520340e-01 3.27523887e-01 2.92103440e-01 -1.28134012e+00
-8.01995575e-01 3.11537743e-01 -6.57283723e-01 2.79789954e-01
4.15642560e-01 4.09357071e-01 1.12323797e+00 -1.05008733e+00
7.97620773e-01 1.17820764e+00 4.59205270e-01 1.14329197e-01
-1.26961112e+00 -6.94920361e-01 6.61287308e-01 3.33711624e-01
-1.71464634e+00 -5.12884974e-01 1.09527922e+00 -3.30751687e-02
9.94427443e-01 4.19719249e-01 1.22812057e+00 8.40201735e-01
1.30937591e-01 5.27168691e-01 4.81854796e-01 -7.70216435e-02
2.84936279e-01 -2.43373707e-01 3.71008143e-02 8.28378618e-01
3.38894546e-01 1.46497369e-01 -9.07724500e-01 -1.69215128e-01
1.16342759e+00 4.83249873e-01 -5.78059316e-01 -7.91498601e-01
-1.46755338e+00 6.92236364e-01 1.12199199e+00 4.16357033e-02
-7.31502235e-01 4.66330439e-01 -8.62260014e-02 3.49702567e-01
4.84126853e-03 5.16556203e-01 -2.19813496e-01 1.93485916e-01
-1.00152493e+00 4.48434912e-02 4.88555849e-01 1.35886073e+00
1.30227220e+00 -7.07552359e-02 -3.62451196e-01 3.65204275e-01
1.05303876e-01 6.70904756e-01 6.27623945e-02 -8.30188334e-01
5.45818508e-01 1.10519516e+00 -1.51697502e-01 -1.69643247e+00
-4.53611732e-01 -5.08177340e-01 -1.05299401e+00 1.30772144e-01
1.80555701e-01 2.98679322e-01 -9.66290236e-01 1.68145502e+00
5.26902258e-01 2.13867188e-01 -3.40505630e-01 1.18308520e+00
8.36988270e-01 4.77612346e-01 -5.15200235e-02 -6.54248819e-02
1.31244206e+00 -8.00027370e-01 -2.03135446e-01 -3.78740460e-01
1.80349290e-01 -7.41032422e-01 1.05091059e+00 1.44993827e-01
-9.53441024e-01 -4.57991958e-01 -1.06379247e+00 -4.36576307e-01
-3.21642131e-01 -2.25824155e-02 7.64734149e-01 2.82033067e-02
-1.20123863e+00 4.23708349e-01 -4.95721608e-01 -2.42669061e-01
5.93642473e-01 3.65816295e-01 -5.28171003e-01 -4.78040189e-01
-8.03899050e-01 4.13450152e-01 9.95841473e-02 2.13275746e-01
-7.67474711e-01 -8.77806723e-01 -8.99498284e-01 2.71866292e-01
7.12904632e-01 -8.96289229e-01 4.98817295e-01 -8.94347966e-01
-1.06508791e+00 3.41375560e-01 -1.25345469e-01 -3.73534232e-01
4.25621897e-01 -1.30770579e-01 -1.09922864e-01 5.61231852e-01
1.12431146e-01 9.68528807e-01 1.35009980e+00 -9.92659152e-01
-5.26424289e-01 -3.05978119e-01 2.84225315e-01 4.26566750e-01
-1.63260564e-01 -4.05562311e-01 -9.00386214e-01 -6.65117800e-01
6.15913033e-01 -7.86476552e-01 -2.13929117e-01 5.26779473e-01
-4.16790754e-01 -1.11864395e-01 7.60102749e-01 -2.88101524e-01
1.06617308e+00 -2.23343873e+00 2.49635026e-01 5.29828072e-01
6.63665473e-01 -4.45104875e-02 -4.14008677e-01 2.51103282e-01
1.56861357e-02 -2.71224827e-01 7.13785514e-02 -2.33803988e-01
-3.17482591e-01 9.97932330e-02 -1.97501227e-01 5.38142145e-01
4.14905012e-01 1.12358832e+00 -1.10452020e+00 -2.82347441e-01
5.64072013e-01 5.85816085e-01 -7.22841084e-01 -2.17084959e-03
-1.75717652e-01 -1.96343735e-02 -5.97424388e-01 8.97987843e-01
7.83827722e-01 -4.30406570e-01 -5.75271109e-03 -8.12614560e-01
3.00237928e-02 8.50775689e-02 -1.46838820e+00 1.93981493e+00
-3.63997042e-01 6.13915622e-01 1.60611436e-01 -7.09678173e-01
7.28151977e-01 -4.20114487e-01 2.74169475e-01 -8.43854606e-01
-9.46964473e-02 -8.21513012e-02 -3.55704248e-01 -4.86385152e-02
8.16964626e-01 7.03336537e-01 1.37357935e-01 2.02682704e-01
-4.14451286e-02 -4.48700525e-02 -3.78700234e-02 6.49816930e-01
1.22016871e+00 -7.15312064e-02 3.07965547e-01 -2.91223168e-01
5.06333947e-01 -1.87104404e-01 4.03670907e-01 8.64990354e-01
-6.32925853e-02 7.36333132e-01 4.38804954e-01 -5.82894444e-01
-8.13410938e-01 -1.03162682e+00 2.56987214e-01 9.34483051e-01
7.19602942e-01 -6.57206714e-01 -3.57363999e-01 -5.65914869e-01
2.03055009e-01 1.21055879e-01 -5.77897668e-01 -1.98509634e-01
-6.48323715e-01 -1.81773826e-01 7.28458166e-02 3.63193214e-01
5.04859567e-01 -8.50725889e-01 -7.88887978e-01 1.54628918e-01
-4.20949347e-02 -8.82398427e-01 -8.21075082e-01 2.20840156e-01
-6.16724968e-01 -1.47497141e+00 -5.52803516e-01 -4.67037499e-01
1.04206502e+00 8.62648427e-01 1.16842866e+00 4.76659268e-01
-3.87051940e-01 5.80460727e-01 -2.31011212e-01 1.39499202e-01
1.46109000e-01 8.06825384e-02 -2.48713970e-01 1.53410807e-01
1.48941740e-01 -5.13105750e-01 -1.05134225e+00 2.62626708e-01
-7.47837365e-01 1.46500226e-02 7.73750186e-01 8.61429274e-01
9.57144916e-01 -3.32545072e-01 2.31953651e-01 -2.17312425e-01
1.12718321e-01 -2.86836028e-01 -9.03151810e-01 1.74198583e-01
-4.18265969e-01 1.76670983e-01 6.22704625e-01 -6.51572585e-01
-3.05333018e-01 1.45133153e-01 4.22050834e-01 -8.71390224e-01
2.31223688e-01 2.84859072e-02 -1.76482797e-01 -6.82740688e-01
6.03267014e-01 2.65289098e-01 2.46480620e-03 -1.21901847e-01
5.51873803e-01 1.72985606e-02 3.20308357e-01 -3.63464445e-01
8.02736461e-01 6.38900876e-01 2.93391883e-01 -7.99139678e-01
-4.16904479e-01 -4.80847567e-01 -2.99498290e-01 -2.47180775e-01
5.28415799e-01 -9.86444175e-01 -8.12718570e-01 3.85902852e-01
-1.08223093e+00 -1.01344801e-01 -5.40827155e-01 3.37180674e-01
-3.11371177e-01 4.26124394e-01 -3.37083369e-01 -4.67374325e-01
-4.29469407e-01 -1.37427187e+00 1.26954460e+00 1.62348539e-01
-2.06962209e-02 -4.79047269e-01 -5.36262572e-01 -8.30591470e-02
4.14667010e-01 -3.68823111e-02 6.80993974e-01 -3.04734886e-01
-1.63670838e+00 -3.05105299e-01 -7.02980518e-01 -7.79606262e-03
-2.08036542e-01 -1.80414572e-01 -8.63555193e-01 -4.95979220e-01
-3.62879515e-01 -1.09543853e-01 9.79273081e-01 3.91563267e-01
1.04443121e+00 -2.37031773e-01 -4.74994600e-01 9.65047300e-01
1.37252319e+00 -2.74119049e-01 4.04087692e-01 1.85401753e-01
1.00597453e+00 4.13991868e-01 4.52186257e-01 2.82086253e-01
3.99053514e-01 6.02163911e-01 7.54182875e-01 -1.24350257e-01
-4.37725306e-01 -3.50704491e-01 -3.23336534e-02 3.76340121e-01
2.40520209e-01 -1.07941791e-01 -6.08913779e-01 6.05212271e-01
-1.84944093e+00 -7.10719764e-01 5.24297431e-02 2.37658000e+00
4.37233418e-01 -3.08968145e-02 -2.15204209e-01 -2.99548894e-01
7.38708079e-01 5.39340854e-01 -7.38198817e-01 2.54135758e-01
-2.51784384e-01 1.32886153e-02 6.89201653e-01 4.27314937e-01
-8.78074586e-01 8.76864791e-01 4.96382189e+00 9.09022689e-01
-1.07143784e+00 -1.48372427e-01 5.57629585e-01 -5.63347280e-01
-4.73627478e-01 1.63628951e-01 -7.65468657e-01 3.40638995e-01
2.08765358e-01 1.19155996e-01 8.18648636e-01 9.23429310e-01
-2.09919140e-01 -4.16432351e-01 -1.23270667e+00 1.39598215e+00
1.48165166e-01 -1.73176730e+00 4.76361305e-01 2.02270538e-01
6.00846589e-01 2.69103974e-01 4.72676232e-02 -4.52957451e-02
-1.67067405e-02 -6.55610383e-01 8.48326564e-01 7.17534244e-01
7.82211423e-01 -9.62925613e-01 3.66868764e-01 2.39253435e-02
-1.43521273e+00 -2.24981621e-01 -6.39305234e-01 1.97470337e-01
-6.86087012e-02 8.71046782e-01 -2.98451364e-01 6.27527595e-01
7.75469005e-01 6.43993378e-01 -6.94542885e-01 1.04356182e+00
-9.92936343e-02 -5.57696857e-02 -7.29549944e-01 -1.13003487e-02
3.79640967e-01 -1.53512388e-01 7.88710296e-01 8.06977868e-01
3.33683670e-01 -2.44600140e-02 2.63370603e-01 1.06675112e+00
-2.68235534e-01 -1.20613165e-02 -6.67595506e-01 4.35873896e-01
9.25411940e-01 1.28441358e+00 -9.37662661e-01 -4.72773343e-01
-4.64220881e-01 1.08471704e+00 5.90727746e-01 5.58819354e-01
-5.70327401e-01 -3.60674948e-01 9.64373291e-01 2.62101352e-01
5.58267355e-01 -2.15897411e-01 -4.57127951e-02 -1.13593328e+00
3.65891933e-01 -6.59832299e-01 2.84142971e-01 -7.30626106e-01
-1.00721097e+00 5.60125291e-01 4.16328311e-02 -1.26484430e+00
3.60576138e-02 -1.77265137e-01 -4.70866144e-01 8.20082426e-01
-1.63990653e+00 -1.31609392e+00 -6.98952377e-01 8.64236772e-01
3.95236343e-01 7.18097836e-02 3.39213312e-01 1.75194785e-01
-2.78256088e-01 7.14586556e-01 -2.80395091e-01 9.19855759e-03
5.68377733e-01 -8.88709128e-01 7.13377416e-01 1.03630280e+00
3.42987418e-01 8.10082138e-01 1.38544321e-01 -7.55196750e-01
-1.88897026e+00 -1.17374051e+00 4.55381662e-01 -2.85491586e-01
5.02704442e-01 -5.64153254e-01 -9.54735219e-01 2.71224141e-01
-2.26477847e-01 5.25193989e-01 -4.23919857e-02 -1.18569337e-01
-5.87662935e-01 -3.69545400e-01 -8.53355110e-01 8.59660089e-01
1.48451650e+00 -7.23611355e-01 -1.41915143e-01 2.16674894e-01
8.84388149e-01 -3.58981162e-01 -5.15568256e-01 1.45772859e-01
3.68583918e-01 -1.11320007e+00 1.31526005e+00 -4.45054984e-03
-6.47774562e-02 -5.62793493e-01 -1.52027473e-01 -9.36018705e-01
-5.72997987e-01 -7.97544718e-01 -3.30744892e-01 9.78318095e-01
-4.36628051e-02 -6.78620040e-01 8.10149789e-01 5.04771054e-01
2.89545637e-02 -6.66164458e-01 -1.17952240e+00 -5.91499627e-01
-6.06536210e-01 -4.77311999e-01 8.25276077e-01 7.56071150e-01
-4.65971529e-01 1.45103291e-01 -2.13947043e-01 4.41577464e-01
7.79155970e-01 3.85416359e-01 7.98258901e-01 -1.09233320e+00
-1.57081947e-01 -5.12737334e-01 -8.42449248e-01 -1.49836028e+00
-1.28927991e-01 -7.66120970e-01 -1.47770017e-01 -1.16802323e+00
-5.48294187e-03 -5.98317027e-01 -6.83512539e-02 3.52158070e-01
-1.43804863e-01 3.68906170e-01 4.67256844e-01 2.38815829e-01
-7.56223500e-01 4.13398743e-01 1.26568389e+00 -2.20910460e-01
-2.69455642e-01 -4.08586323e-01 -5.74765384e-01 5.36714494e-01
2.67648667e-01 -2.86366940e-01 -5.23353875e-01 -3.82930070e-01
5.29137731e-01 -1.35986075e-01 9.43076849e-01 -1.09194195e+00
6.55151486e-01 -4.80337664e-02 6.90392137e-01 -7.61380255e-01
3.84295434e-01 -1.03572702e+00 1.82236582e-01 1.06465548e-01
3.69073488e-02 2.47280046e-01 7.27587268e-02 1.00254726e+00
-3.78876105e-02 1.69434234e-01 5.00277638e-01 -1.52985215e-01
-9.52750742e-01 7.10483670e-01 -1.23115778e-01 -3.36518176e-02
1.06559086e+00 -5.75438559e-01 -2.93679535e-01 -2.61750460e-01
-3.80677015e-01 2.00873479e-01 7.75990129e-01 3.01165521e-01
8.93768489e-01 -1.32960594e+00 -2.64245540e-01 8.94983172e-01
2.71062464e-01 5.72067320e-01 5.68404675e-01 7.78343260e-01
-3.65166217e-01 1.95335910e-01 4.09418568e-02 -9.40002620e-01
-9.94581580e-01 7.78879523e-01 2.31149524e-01 8.08125138e-02
-9.04921472e-01 9.45811570e-01 4.28456068e-01 6.86728656e-02
3.39146495e-01 -7.22836971e-01 1.75448522e-01 2.12813511e-01
5.78806937e-01 3.20527196e-01 1.44936025e-01 -5.23227811e-01
-3.80128950e-01 9.21487689e-01 -2.80138999e-01 4.30122226e-01
9.23171818e-01 -3.85848373e-01 -1.17377162e-01 -6.97551295e-02
1.22067022e+00 -1.01931363e-01 -1.58951998e+00 -3.65320861e-01
-3.99261564e-01 -7.39143252e-01 4.07876253e-01 -3.48273188e-01
-1.20333993e+00 5.43787599e-01 4.33475703e-01 6.06503636e-02
1.20908177e+00 9.96694714e-03 6.71640158e-01 5.39214611e-01
6.18058920e-01 -9.26840425e-01 2.14397132e-01 3.42538446e-01
9.13943529e-01 -1.20827925e+00 1.59367859e-01 -4.19315010e-01
-1.65973246e-01 1.03749108e+00 7.01638579e-01 -2.73369968e-01
4.80666071e-01 4.78205457e-02 -2.33444080e-01 -4.80808467e-01
-6.56837702e-01 -1.51136324e-01 4.51454937e-01 6.81062996e-01
-2.85667062e-01 -1.61519855e-01 3.88554931e-01 -1.35866208e-02
1.35442674e-01 -3.65188748e-01 1.64026320e-01 7.10936546e-01
-3.00118059e-01 -5.79736531e-01 -3.65705788e-01 5.44762433e-01
2.16015890e-01 -5.04739523e-01 -1.96623996e-01 8.70277524e-01
-1.13360822e-01 6.27869725e-01 4.27395493e-01 -2.35471696e-01
3.73206377e-01 -3.79239142e-01 3.18582684e-01 -3.58187884e-01
-5.87328434e-01 9.15852636e-02 -2.86508113e-01 -1.23502922e+00
-2.54825771e-01 -3.43721181e-01 -1.05522501e+00 -4.53600705e-01
-4.56734061e-01 -2.19631910e-01 3.63927513e-01 5.58341324e-01
9.91248310e-01 4.01202142e-01 4.36885715e-01 -1.16901183e+00
-3.72276664e-01 -5.57636201e-01 -2.84639567e-01 3.27939093e-01
7.45880842e-01 -6.76399410e-01 -4.07319814e-01 -4.95410949e-01] | [7.893206596374512, -1.9614654779434204] |
a764ea34-27ec-47c6-8514-af3514489e32 | hindsight-learning-for-mdps-with-exogenous | 2207.06272 | null | https://arxiv.org/abs/2207.06272v2 | https://arxiv.org/pdf/2207.06272v2.pdf | Hindsight Learning for MDPs with Exogenous Inputs | Many resource management problems require sequential decision-making under uncertainty, where the only uncertainty affecting the decision outcomes are exogenous variables outside the control of the decision-maker. We model these problems as Exo-MDPs (Markov Decision Processes with Exogenous Inputs) and design a class of data-efficient algorithms for them termed Hindsight Learning (HL). Our HL algorithms achieve data efficiency by leveraging a key insight: having samples of the exogenous variables, past decisions can be revisited in hindsight to infer counterfactual consequences that can accelerate policy improvements. We compare HL against classic baselines in the multi-secretary and airline revenue management problems. We also scale our algorithms to a business-critical cloud resource management problem -- allocating Virtual Machines (VMs) to physical machines, and simulate their performance with real datasets from a large public cloud provider. We find that HL algorithms outperform domain-specific heuristics, as well as state-of-the-art reinforcement learning methods. | ['Adith Swaminathan', 'Ishai Menache', 'Jennifer Neville', 'Jingling Li', 'Hugo Barbalho', 'Luke Marshall', 'Ching-An Cheng', 'Felipe Frujeri', 'Sean R. Sinclair'] | 2022-07-13 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 6.77565709e-02 5.11333406e-01 -7.19475567e-01 -1.98942542e-01
-6.96769893e-01 -6.26095772e-01 4.05494690e-01 -4.52010669e-02
-5.68522096e-01 1.23119414e+00 4.25753385e-01 -1.00540817e+00
-3.02282095e-01 -6.05817676e-01 -7.96085596e-01 -5.82996786e-01
-3.61071616e-01 9.35699940e-01 -5.17132342e-01 3.37095857e-01
2.21797287e-01 1.84521675e-01 -1.03746247e+00 1.36342525e-01
5.78471720e-01 1.18535352e+00 1.66919887e-01 7.54505396e-01
1.68365479e-01 1.21646416e+00 -4.06396657e-01 -2.98832297e-01
7.71433592e-01 3.22663873e-01 -7.57230937e-01 3.17880511e-01
-5.03125012e-01 -7.62981653e-01 -2.24853456e-01 8.84142339e-01
5.24280429e-01 2.19127297e-01 6.76802158e-01 -1.53539586e+00
-5.01650453e-01 9.76265192e-01 -1.01952350e+00 3.43345255e-01
1.39789134e-01 7.88739026e-01 1.22929728e+00 1.12085477e-01
5.10687172e-01 1.61228752e+00 1.19707964e-01 4.17476267e-01
-1.47677803e+00 -4.12871689e-01 8.13012660e-01 1.20124444e-01
-4.76585478e-01 -3.16908121e-01 3.80106628e-01 -4.15369630e-01
1.08510542e+00 -3.23008858e-02 3.28472048e-01 1.18580472e+00
4.57979709e-01 1.03582716e+00 1.47477019e+00 -4.90877360e-01
7.60330796e-01 1.39098629e-01 -4.25895929e-01 1.40158281e-01
4.18888360e-01 6.00210607e-01 -1.19874932e-01 -6.61096096e-01
4.74432975e-01 2.19305173e-01 1.21004224e-01 -4.93406504e-01
-8.96507561e-01 8.39325726e-01 -2.85316110e-01 -4.41444546e-01
-1.22765899e+00 4.83907700e-01 2.67288178e-01 4.65574563e-01
2.12578207e-01 9.00756419e-01 -9.64900136e-01 -1.91084757e-01
-5.80331922e-01 5.76488256e-01 9.63702381e-01 8.46882582e-01
3.29892606e-01 1.07746921e-01 -7.50565350e-01 2.28356048e-01
9.37156677e-02 5.83586335e-01 2.33566687e-01 -1.55067408e+00
7.33838081e-01 -4.89284754e-01 1.26770818e+00 -3.21236402e-01
-2.95064062e-01 -3.14468145e-01 -2.24604040e-01 2.38663718e-01
1.69768542e-01 -9.44969714e-01 -7.69684017e-01 1.75427186e+00
2.30836600e-01 -2.55536474e-02 3.04258205e-02 8.30667138e-01
-7.04759657e-01 7.92806983e-01 1.51353016e-01 -8.41342866e-01
1.04167414e+00 -7.98488081e-01 -6.20525658e-01 -2.60623187e-01
4.00510997e-01 -3.80029261e-01 7.69731224e-01 8.21550667e-01
-1.09607184e+00 -4.30904441e-02 -6.80063903e-01 8.74957860e-01
1.32057428e-01 -4.61647451e-01 9.09956336e-01 2.21253768e-01
-6.13941789e-01 8.14491332e-01 -7.22180009e-01 1.86122969e-01
5.45049250e-01 1.41674682e-01 4.19198513e-01 -4.50070528e-03
-1.18635547e+00 9.51561034e-01 1.23964719e-01 -2.91262716e-01
-1.65790474e+00 -8.57366621e-01 -1.39023408e-01 2.81216532e-01
1.39372623e+00 -8.62868071e-01 2.15088534e+00 -8.47609639e-01
-1.58561265e+00 3.91240954e-01 3.15350562e-01 -1.02536225e+00
8.73562574e-01 -2.75216341e-01 -4.32741761e-01 -7.71906450e-02
2.45282933e-01 1.59636170e-01 1.10767555e+00 -1.23916113e+00
-1.10699654e+00 -2.68227458e-01 2.42775172e-01 2.82882392e-01
1.52164951e-01 -9.35061872e-02 3.77230406e-01 -2.41653934e-01
-7.30832219e-01 -1.07273018e+00 -1.09787488e+00 -8.81384373e-01
-6.43959999e-01 -9.04894397e-02 2.17707381e-01 -3.53636205e-01
1.21457934e+00 -1.70225978e+00 -2.24958688e-01 4.41587344e-03
-3.58843505e-02 -3.18374038e-01 9.97809619e-02 4.10316467e-01
-9.61003229e-02 4.81465936e-01 2.50597358e-01 -1.07779428e-01
5.88740885e-01 3.72271985e-01 -7.90748000e-01 3.83323520e-01
-1.86785266e-01 6.00445688e-01 -7.87049770e-01 -2.20589623e-01
1.93736076e-01 -8.72377455e-01 -4.91289109e-01 3.52792531e-01
-9.79066730e-01 5.67010194e-02 -8.97330046e-01 5.13612866e-01
5.78627765e-01 -3.23249757e-01 5.69430709e-01 6.79815590e-01
9.05582868e-03 3.15068588e-02 -1.24302804e+00 9.67428803e-01
-8.53364825e-01 2.03902945e-02 2.62784809e-01 -1.01774740e+00
-9.55036189e-03 2.50293940e-01 6.61654413e-01 -6.28436446e-01
-5.79239801e-02 -2.08506182e-01 -1.17491901e-01 -6.21906281e-01
3.28072488e-01 -5.35931706e-01 -2.79991031e-01 1.02670503e+00
-3.56289178e-01 -8.80072266e-02 -1.09429896e-01 1.62890002e-01
1.28299057e+00 -2.10556239e-02 7.73337245e-01 -2.72699326e-01
-3.43944401e-01 1.72312587e-01 1.09328139e+00 1.24961758e+00
-2.94181406e-01 -2.29426712e-01 1.14378631e+00 -4.97323006e-01
-1.11034966e+00 -9.18087840e-01 1.98871970e-01 1.23193073e+00
-3.16752791e-01 1.98325455e-01 -4.11231190e-01 -7.76681066e-01
9.63041365e-01 1.38542426e+00 -7.37508476e-01 3.10761243e-01
4.25740890e-02 -8.81900609e-01 -4.01513696e-01 4.99609768e-01
1.47588542e-02 -9.13446307e-01 -7.68807530e-01 5.92141032e-01
2.97536463e-01 -1.16051066e+00 -3.20740849e-01 2.46472284e-01
-6.25336647e-01 -9.75336552e-01 -3.79298151e-01 5.97758651e-01
4.38561529e-01 1.56618252e-01 1.22015274e+00 -7.66895115e-01
-2.54324228e-01 4.91579980e-01 2.82750845e-01 -1.08718729e+00
-3.88311177e-01 -1.84707567e-01 6.47372723e-01 -5.54729588e-02
2.35872701e-01 -4.50216800e-01 -7.95509517e-01 1.68215134e-03
-5.11742830e-01 -1.47772744e-01 7.71513104e-01 7.89506435e-01
7.66437709e-01 3.78265768e-01 7.51981318e-01 -1.40044451e+00
8.91827345e-01 -7.72520542e-01 -1.15814555e+00 4.97133523e-01
-1.07877266e+00 4.15908098e-01 9.86474812e-01 -4.24530834e-01
-1.45524073e+00 -1.60241961e-01 8.98387909e-01 -6.37644529e-01
-2.06295758e-01 1.97846338e-01 -1.51038408e-01 8.10395718e-01
4.37219352e-01 -1.13557778e-01 -1.88379481e-01 -3.66598606e-01
4.55605954e-01 6.36700511e-01 2.91147083e-01 -1.24627924e+00
5.25204003e-01 4.11459327e-01 2.43163511e-01 -4.25959229e-02
-1.10356307e+00 -1.48437291e-01 7.34171942e-02 7.99457505e-02
5.20086288e-01 -1.08318472e+00 -1.19534063e+00 -2.08098233e-01
-5.38993478e-01 -5.43893754e-01 -5.22638261e-01 4.52828884e-01
-1.41013598e+00 -1.62837014e-01 -4.02178794e-01 -1.33793330e+00
3.66189778e-02 -8.96067083e-01 5.32184184e-01 2.99981594e-01
1.46436825e-01 -8.68337333e-01 1.98478460e-01 4.47876781e-01
2.28736505e-01 3.03081155e-01 1.22058928e+00 -5.90807378e-01
-8.65635514e-01 1.44594058e-01 1.88410804e-01 1.74162418e-01
-1.18126735e-01 1.02988668e-01 -7.29142010e-01 -5.82520604e-01
-5.37628168e-03 -4.39752787e-01 5.50891697e-01 7.90642023e-01
1.82279325e+00 -1.12906539e+00 -5.56883454e-01 3.04106921e-01
1.47217476e+00 5.21858752e-01 2.76375543e-02 6.91915393e-01
-1.11904629e-01 5.47549486e-01 1.25943708e+00 1.41000962e+00
2.56156832e-01 3.51001978e-01 6.44198954e-01 2.83463389e-01
1.03939199e+00 -5.75905144e-01 3.24450403e-01 -3.09441984e-01
-1.98681522e-02 -2.40324274e-01 -9.10692036e-01 6.51409090e-01
-2.13567662e+00 -1.16881371e+00 6.31150961e-01 2.34185100e+00
6.15070105e-01 5.66702783e-01 4.86974120e-01 -4.81163412e-01
5.54018319e-01 1.70178279e-01 -1.35586536e+00 -9.60662663e-01
2.08591700e-01 -1.26237452e-01 1.19822419e+00 3.67621899e-01
-9.31733131e-01 6.64347053e-01 6.75504303e+00 5.58816850e-01
-6.89176261e-01 7.92660285e-03 1.17437625e+00 -8.87497127e-01
-3.75262082e-01 1.46201015e-01 -7.71046877e-01 5.13066709e-01
1.47718716e+00 -7.95661449e-01 8.79959881e-01 1.32950878e+00
6.71887636e-01 -3.11970502e-01 -1.41094124e+00 7.87315071e-01
-8.46200228e-01 -1.46722507e+00 -2.81732351e-01 3.28374654e-01
1.27760530e+00 6.67185113e-02 3.23967069e-01 6.17339373e-01
1.70882273e+00 -1.11424685e+00 4.99466568e-01 5.94292939e-01
5.31855941e-01 -1.25356579e+00 5.63084722e-01 5.29208362e-01
-2.31614247e-01 -1.16895342e+00 -4.55786884e-01 -3.01908404e-01
2.40391433e-01 6.28757894e-01 -1.20062220e+00 2.82424301e-01
4.40941572e-01 1.35193899e-01 2.19119474e-01 4.44635600e-01
-3.26184362e-01 7.68683136e-01 -1.00531153e-01 8.17312673e-02
6.10522389e-01 -1.13892421e-01 2.42855847e-01 7.50665605e-01
-6.77344278e-02 2.61742383e-01 5.81848562e-01 8.41029704e-01
-8.27896819e-02 -2.62828439e-01 -5.97835362e-01 -4.62773860e-01
6.74102306e-01 1.08110726e+00 -2.35272273e-01 -3.31243873e-01
-2.66210169e-01 4.45877790e-01 2.61299242e-03 6.07885838e-01
-6.04152620e-01 -3.52828950e-02 1.24611747e+00 -5.75302392e-02
2.67369628e-01 7.89333582e-02 -3.06818038e-01 -9.44556594e-01
-6.80219904e-02 -1.17671847e+00 6.26465619e-01 -5.63967586e-01
-1.38590634e+00 -2.76323318e-01 -4.55699116e-02 -6.94140196e-01
-8.65186393e-01 -6.36888921e-01 -4.56917673e-01 7.49637306e-01
-1.60955167e+00 -3.58313859e-01 6.96150899e-01 1.92585140e-01
4.38193411e-01 -4.44068871e-02 4.41533536e-01 -6.59480333e-01
-7.83902466e-01 -5.99057153e-02 6.92323029e-01 -3.94054204e-01
4.42680627e-01 -1.63602304e+00 4.90121424e-01 7.15700507e-01
-2.67536908e-01 2.23605365e-01 9.51215565e-01 -3.66163373e-01
-1.65137243e+00 -9.32189703e-01 8.74888003e-02 -4.62325394e-01
9.46061909e-01 -1.47518054e-01 -3.21905553e-01 9.34996784e-01
3.31472278e-01 -1.00122988e-01 4.94875759e-01 3.51864904e-01
-2.22121477e-02 -2.57638305e-01 -1.17347455e+00 5.64527273e-01
9.72231627e-01 -2.73857415e-01 -5.91713905e-01 6.57416582e-01
1.01690972e+00 -1.26463562e-01 -6.96188033e-01 2.39608306e-02
2.17481405e-01 -5.04435956e-01 6.27118230e-01 -1.73939812e+00
4.88076895e-01 2.15112507e-01 -3.15937102e-01 -1.74386406e+00
-5.69011927e-01 -1.55666399e+00 -5.07550895e-01 8.65248382e-01
5.01808703e-01 -6.20082676e-01 6.10814571e-01 1.02513480e+00
5.63837111e-01 -7.84513652e-01 -8.08266461e-01 -9.26460981e-01
3.31029236e-01 -2.27348477e-01 1.15586841e+00 8.36757839e-01
-1.16382673e-01 7.37171015e-03 -5.08735359e-01 3.20360750e-01
1.09478462e+00 7.88121462e-01 7.42000043e-01 -8.58648956e-01
-8.94792974e-01 -3.12291831e-01 4.92883950e-01 -7.42743850e-01
5.35996974e-01 -2.02592000e-01 -3.92036140e-02 -1.35459948e+00
3.86645347e-01 -6.79548144e-01 -7.88732946e-01 4.27503049e-01
-7.36527815e-02 -1.15167356e+00 4.16687012e-01 -6.27276823e-02
-1.08617640e+00 5.69786012e-01 1.16881871e+00 1.13651603e-01
-2.35447362e-01 6.06710494e-01 -1.24418211e+00 6.56065702e-01
8.25857639e-01 -5.15649617e-01 -5.13270617e-01 -2.37728894e-01
2.62658119e-01 9.88268971e-01 -1.15824893e-01 -3.34445029e-01
8.58931511e-04 -1.38740659e+00 3.59162241e-01 -4.34982300e-01
-2.24204570e-01 -5.81228137e-01 1.91566825e-01 6.43389463e-01
-7.34950781e-01 3.34346473e-01 -1.28585368e-01 1.04339242e+00
3.14197659e-01 1.15869366e-01 7.41833746e-01 -6.40163422e-01
-7.30052531e-01 6.62405372e-01 -6.31283402e-01 4.62487370e-01
1.37155461e+00 5.22914588e-01 -3.85955095e-01 -7.71499038e-01
-6.86526299e-01 9.32813346e-01 2.59989172e-01 3.10949445e-01
1.69478636e-02 -7.90587604e-01 -5.69199920e-01 -5.12419045e-01
-1.28631994e-01 -5.24275564e-02 3.25331628e-01 3.91686976e-01
2.03234464e-01 9.01502252e-01 -2.01769620e-01 -1.24588996e-01
-5.28308988e-01 1.27926719e+00 3.87664676e-01 -8.59038711e-01
3.97240464e-03 4.75295544e-01 2.37190098e-01 -1.83768481e-01
1.20045975e-01 -1.08047776e-01 4.39551175e-01 1.11548699e-01
5.47287107e-01 5.39995372e-01 -2.96118587e-01 5.87069750e-01
-1.21488169e-01 -4.88132626e-01 -3.58961225e-01 -5.44378519e-01
1.53489435e+00 -3.46084565e-01 3.82230461e-01 4.20519084e-01
4.09021676e-01 -2.35082224e-01 -1.79296374e+00 -6.92859173e-01
3.99592251e-01 -7.36910760e-01 2.07033932e-01 -1.19786417e+00
-7.82921135e-01 7.22123146e-01 1.16695777e-01 3.30624908e-01
1.02900934e+00 -3.38625252e-01 1.49919391e-01 4.35275316e-01
8.99486721e-01 -1.77419364e+00 -1.40491694e-01 -7.26236701e-02
5.32197237e-01 -1.26854694e+00 2.17806801e-01 4.23454136e-01
-1.21838117e+00 5.06613076e-01 7.27973759e-01 -5.42904325e-02
5.76319098e-01 4.01351213e-01 -2.29503274e-01 3.09872270e-01
-2.00953364e+00 -1.20548576e-01 -5.86781025e-01 5.25679231e-01
-3.03026408e-01 9.93773878e-01 3.62108424e-02 7.96207190e-01
-1.86931267e-02 2.44719297e-01 8.78646076e-01 1.07884288e+00
-2.11947963e-01 -9.02206182e-01 -4.67784256e-01 8.39264214e-01
-9.59341228e-01 1.21487975e-01 -6.59948140e-02 7.36127377e-01
-3.65097821e-01 8.53678465e-01 3.37084420e-02 1.60119295e-01
2.79593110e-01 -1.00880470e-02 2.99463570e-01 -6.44649565e-01
-6.91487864e-02 -8.39575753e-02 1.30024090e-01 -9.02249575e-01
1.29547879e-01 -1.03391075e+00 -9.09120202e-01 -5.44921100e-01
1.63153604e-01 2.01788276e-01 6.26070976e-01 8.97756934e-01
7.21932173e-01 5.48432827e-01 1.54293883e+00 -3.48459095e-01
-2.08003163e+00 -7.27682412e-01 -1.01160920e+00 1.25508577e-01
4.72223520e-01 -7.30414569e-01 -4.73073542e-01 -5.05173683e-01] | [4.285417079925537, 2.787506580352783] |
4978c436-4d56-43ce-8697-9ec494f279a5 | a-survey-on-text-to-sql-parsing-concepts | 2208.13629 | null | https://arxiv.org/abs/2208.13629v1 | https://arxiv.org/pdf/2208.13629v1.pdf | A Survey on Text-to-SQL Parsing: Concepts, Methods, and Future Directions | Text-to-SQL parsing is an essential and challenging task. The goal of text-to-SQL parsing is to convert a natural language (NL) question to its corresponding structured query language (SQL) based on the evidences provided by relational databases. Early text-to-SQL parsing systems from the database community achieved a noticeable progress with the cost of heavy human engineering and user interactions with the systems. In recent years, deep neural networks have significantly advanced this task by neural generation models, which automatically learn a mapping function from an input NL question to an output SQL query. Subsequently, the large pre-trained language models have taken the state-of-the-art of the text-to-SQL parsing task to a new level. In this survey, we present a comprehensive review on deep learning approaches for text-to-SQL parsing. First, we introduce the text-to-SQL parsing corpora which can be categorized as single-turn and multi-turn. Second, we provide a systematical overview of pre-trained language models and existing methods for text-to-SQL parsing. Third, we present readers with the challenges faced by text-to-SQL parsing and explore some potential future directions in this field. | ['Yongbin Li', 'Fei Huang', 'Luo Si', 'Jian Sun', 'Rongyu Cao', 'Ruiying Geng', 'Binhua Li', 'Jinyang Li', 'Min Yang', 'Lihan Wang', 'Binyuan Hui', 'Bowen Qin'] | 2022-08-29 | null | null | null | null | ['text-to-sql'] | ['computer-code'] | [ 2.22359255e-01 5.44089735e-01 -2.35238299e-01 -1.17988718e+00
-1.42008460e+00 -6.02562606e-01 4.62332904e-01 4.14157659e-01
-1.02902234e-01 4.12074327e-01 1.38046846e-01 -7.41057038e-01
3.76088202e-01 -1.42419040e+00 -1.24637115e+00 2.62681484e-01
3.26910436e-01 9.72367704e-01 2.33908936e-01 -2.82614440e-01
1.72611922e-01 2.34711766e-01 -1.18353271e+00 1.02153802e+00
4.98295367e-01 1.13240540e+00 1.16904497e-01 8.76623333e-01
-1.12579107e+00 1.33926404e+00 -6.83133304e-01 -9.26635861e-01
-5.63507713e-02 -1.66671664e-01 -1.22467542e+00 -3.46861184e-01
5.25038421e-01 -5.78697801e-01 -5.80600835e-02 8.90247107e-01
1.97306007e-01 -4.63381618e-01 -8.81497860e-02 -9.99147177e-01
-6.57726407e-01 9.22714114e-01 6.53959140e-02 -2.51061112e-01
6.47522330e-01 -1.65257335e-01 1.40760541e+00 -1.17223239e+00
7.60756850e-01 1.39695919e+00 5.66247940e-01 2.51680404e-01
-8.85033965e-01 -1.68331549e-01 9.23748885e-04 -1.29725382e-01
-7.40156770e-01 -3.83289427e-01 6.41236365e-01 -3.19924921e-01
1.72928691e+00 1.78282410e-01 2.14471653e-01 6.93949401e-01
3.36073577e-01 1.00900578e+00 6.30755067e-01 -7.14896023e-01
5.35439216e-02 7.01933503e-02 4.39661354e-01 8.84699762e-01
1.60960525e-01 -2.33354911e-01 -7.22756088e-01 -1.63564548e-01
5.90812266e-01 -2.10204437e-01 5.11258602e-01 -2.53583401e-01
-9.44454670e-01 1.17778194e+00 2.52035737e-01 1.17719062e-01
-3.49686325e-01 1.52874991e-01 8.10340464e-01 2.78931528e-01
1.95171967e-01 5.67512929e-01 -7.92384565e-01 -2.17789993e-01
-7.29851425e-01 4.41664845e-01 1.27738845e+00 1.38021147e+00
7.41851807e-01 8.10070038e-02 9.36801285e-02 7.20056057e-01
4.27251726e-01 2.75069565e-01 1.99674994e-01 -1.13660610e+00
1.02752745e+00 9.63478267e-01 -1.78693190e-01 -6.46766126e-01
-1.75817013e-01 2.10552618e-01 -4.27860349e-01 -1.57857165e-01
1.93029672e-01 -7.54290149e-02 -8.58239412e-01 1.26552618e+00
2.18788296e-01 -1.04947770e+00 3.73737127e-01 2.81068504e-01
1.37747216e+00 8.34022701e-01 3.09877872e-01 1.06693916e-01
1.63295865e+00 -7.19433844e-01 -7.20044136e-01 -6.19156182e-01
6.94775283e-01 -7.10043669e-01 1.29153562e+00 2.32763499e-01
-1.51057875e+00 -8.69350910e-01 -9.22334909e-01 -7.12073088e-01
-7.38592863e-01 1.13096267e-01 7.62237310e-01 6.26196623e-01
-1.03598702e+00 2.20803544e-01 -1.04715908e+00 -5.74711859e-01
3.27365607e-01 3.05246562e-01 -3.43505681e-01 -1.40393367e-02
-1.29239881e+00 6.77105308e-01 6.45903170e-01 1.12888161e-02
-4.24395174e-01 -5.44552922e-01 -1.12092590e+00 5.15134558e-02
8.01152647e-01 -7.76002705e-01 1.84634972e+00 -4.16155130e-01
-1.47736502e+00 1.27607965e+00 -5.78394949e-01 -5.10073602e-01
1.88535139e-01 -4.80279803e-01 -1.19478621e-01 1.63898408e-01
4.76582944e-01 6.83240712e-01 2.41202071e-01 -1.04700148e+00
-6.04901433e-01 -4.44936991e-01 3.19504023e-01 -1.54350564e-01
3.18355232e-01 6.89981580e-01 -8.86564195e-01 -5.05693376e-01
2.84424633e-01 -2.55287886e-01 -4.97590005e-02 -1.61681980e-01
-7.86228895e-01 -6.06647551e-01 3.16891015e-01 -7.39079654e-01
1.11751759e+00 -1.65198433e+00 -4.29173112e-01 -2.41597861e-01
2.19406396e-01 7.21097132e-03 -2.03510420e-03 9.76052761e-01
-1.21312767e-01 4.51336265e-01 -1.66516677e-01 -2.99670964e-01
1.44927979e-01 3.02829921e-01 -8.87632608e-01 -5.05615473e-01
9.39495444e-01 1.30756605e+00 -5.21165013e-01 -7.77839482e-01
-3.77826183e-03 5.14220297e-02 -6.28670454e-01 9.87127125e-01
-9.24823344e-01 -1.20383643e-01 -6.15671933e-01 9.04125631e-01
4.15662795e-01 -2.16055959e-01 4.87267405e-01 -2.27566004e-01
6.57478273e-02 1.12138748e+00 -6.41237438e-01 1.38350129e+00
-4.85387325e-01 5.17146885e-01 9.23595428e-02 -1.31018114e+00
1.28773928e+00 1.57673135e-01 2.01515555e-01 -9.37794626e-01
-1.56840280e-01 4.61780429e-01 -5.05670428e-01 -7.72104144e-01
8.34139884e-01 -2.04819411e-01 -6.95512176e-01 3.23258460e-01
2.16938704e-01 -5.36490202e-01 2.33564988e-01 3.48851830e-01
9.30802822e-01 4.01778311e-01 5.92563391e-01 2.13960007e-01
4.48460340e-01 4.36925262e-01 3.11926603e-01 9.71472204e-01
3.45625073e-01 5.04753172e-01 8.74007165e-01 -8.69113326e-01
-9.60042417e-01 -1.11609721e+00 8.20916668e-02 1.35443735e+00
-5.67510605e-01 -5.47384620e-01 -9.16847765e-01 -8.39236379e-01
-7.74835274e-02 9.47875082e-01 -3.61526012e-01 2.21570209e-01
-1.11738980e+00 -5.22371233e-01 8.70175302e-01 9.63456571e-01
6.23334587e-01 -1.59354007e+00 -6.64263725e-01 4.62905943e-01
-2.44456977e-01 -1.50614643e+00 1.58105075e-01 6.75810874e-01
-1.12247181e+00 -9.56278384e-01 6.18287846e-02 -1.15111792e+00
3.57303202e-01 -4.42362785e-01 1.95923424e+00 1.12995535e-01
-1.87940612e-01 2.28718922e-01 -3.39045763e-01 -7.52042711e-01
-9.60207701e-01 2.55443066e-01 -8.23666751e-01 -5.56276917e-01
9.86800671e-01 -1.24541083e-02 7.84619749e-02 -6.16875989e-03
-1.15222597e+00 9.65496674e-02 4.80026782e-01 5.88205457e-01
9.44811821e-01 -1.69260874e-01 5.70252955e-01 -1.27354860e+00
6.91162527e-01 -2.91453421e-01 -7.64042914e-01 5.76126397e-01
-4.67260420e-01 2.79535532e-01 7.93754876e-01 2.91131675e-01
-1.09617293e+00 1.74989626e-01 -6.01123333e-01 2.30632097e-01
-2.92961299e-01 1.04513919e+00 -6.01180732e-01 6.60181701e-01
4.19206142e-01 2.53890634e-01 -2.48865381e-01 -4.86493826e-01
7.43206680e-01 6.62149429e-01 7.29647636e-01 -9.15424526e-01
4.16991621e-01 8.34910199e-02 -2.44442046e-01 -5.90409994e-01
-9.65887487e-01 4.58264090e-02 -7.74011374e-01 2.91919798e-01
1.15369320e+00 -7.01671541e-01 -5.77638328e-01 4.59018677e-01
-1.64291418e+00 -5.07074237e-01 -4.65666741e-01 -2.72790074e-01
-5.93863606e-01 2.77494788e-01 -8.75817299e-01 -5.73598444e-01
-5.51482141e-01 -1.03336608e+00 1.31699407e+00 1.12626545e-01
-2.02365875e-01 -8.91140223e-01 -1.33309200e-01 6.61796749e-01
3.56473029e-01 1.28095493e-01 1.59283721e+00 -1.11112940e+00
-8.45715940e-01 -5.46980619e-01 -3.61952633e-01 2.98842639e-01
-8.57695639e-02 7.74645507e-02 -7.12038815e-01 2.61799365e-01
3.67838234e-01 -9.24100876e-01 3.73571634e-01 1.04156934e-01
1.26626790e+00 -2.67065704e-01 -1.68472677e-01 5.23433268e-01
1.32048070e+00 4.40828592e-01 4.81103778e-01 4.09811884e-01
4.89094108e-01 8.93049538e-01 7.39324391e-01 6.51638061e-02
9.40723717e-01 3.92873555e-01 5.01068115e-01 -7.17445016e-02
7.38427043e-02 -7.24975824e-01 1.78695813e-01 8.53451729e-01
6.20216310e-01 -3.63118052e-01 -1.23687625e+00 3.07757258e-01
-1.53648770e+00 -4.20641690e-01 -1.36008069e-01 1.80996358e+00
1.03791833e+00 4.84889418e-01 -2.50714272e-01 -2.59536386e-01
4.66306657e-01 2.20767900e-01 -6.16425157e-01 -8.66646945e-01
-3.27908136e-02 5.64074278e-01 -2.87969727e-02 5.56767762e-01
-1.09058928e+00 1.33802855e+00 6.98342514e+00 1.42122284e-01
-1.09169090e+00 -1.91191807e-01 7.72976398e-01 3.35784316e-01
-2.94832200e-01 9.89431441e-02 -1.25630653e+00 -2.25131527e-01
1.34129441e+00 4.16732989e-02 1.85114145e-01 1.19433880e+00
-1.03741951e-01 -8.99709240e-02 -1.67762649e+00 6.39028490e-01
-4.64906963e-03 -1.53231716e+00 5.30085027e-01 -3.35658997e-01
-8.05725530e-02 2.14868397e-01 -4.21207845e-01 8.00826132e-01
4.78854358e-01 -1.22302473e+00 6.72501147e-01 4.52081561e-01
9.25077140e-01 -3.93570244e-01 8.06914389e-01 2.38303453e-01
-1.34341335e+00 1.76733181e-01 -4.26215649e-01 1.81659207e-01
2.67536014e-01 4.85982329e-01 -8.76843512e-01 6.82951391e-01
8.42368007e-01 4.18906093e-01 -7.77490497e-01 4.79438975e-02
-1.12119272e-01 4.47010219e-01 -1.97203338e-01 -1.33577392e-01
1.31007627e-01 -1.17102684e-02 -1.37640581e-01 1.34270072e+00
1.44795761e-01 4.61384468e-02 2.12308213e-01 1.35960662e+00
-5.08686185e-01 9.14292857e-02 -6.29493177e-01 -5.76723814e-01
4.11239386e-01 8.20584297e-01 -4.48951155e-01 -7.40884483e-01
-8.15121412e-01 5.81550539e-01 4.59563166e-01 1.21411562e-01
-5.05187750e-01 -8.61926794e-01 5.00448570e-02 3.40313882e-01
3.16673785e-01 -1.73999965e-01 -5.68252981e-01 -9.93494809e-01
5.48543513e-01 -1.17859781e+00 5.28124154e-01 -1.10407317e+00
-1.32017350e+00 8.72829378e-01 1.07197881e-01 -5.39626539e-01
-7.44742811e-01 -7.67384887e-01 -4.77064848e-01 9.49849427e-01
-1.32676709e+00 -1.08273518e+00 -2.49720529e-01 2.62407869e-01
8.13041866e-01 -3.53790224e-01 1.10241055e+00 1.81218505e-01
-3.58686954e-01 3.80309075e-01 -4.66223866e-01 8.48945022e-01
5.84054828e-01 -1.29007387e+00 1.25403714e+00 7.27332652e-01
3.50438096e-02 1.05595112e+00 1.81914583e-01 -8.67940247e-01
-1.79659688e+00 -1.15390360e+00 1.59254861e+00 -7.84033775e-01
6.10429466e-01 -6.66460335e-01 -1.07642865e+00 1.14719152e+00
3.67003947e-01 -4.77322303e-02 5.42064369e-01 -8.69523883e-02
-4.31089789e-01 -2.21390054e-01 -1.04052818e+00 3.59285057e-01
4.65034425e-01 -1.04216576e+00 -1.02634883e+00 2.56040871e-01
1.28593612e+00 -6.33211613e-01 -1.03179383e+00 4.65588093e-01
3.75322282e-01 -1.16407490e+00 8.81656408e-01 -9.59129989e-01
9.21046793e-01 2.31170431e-02 -5.46283484e-01 -2.56422311e-01
3.64352971e-01 -6.15121722e-01 -1.40080377e-02 1.21313941e+00
4.53774780e-01 -4.44641471e-01 1.11777329e+00 9.95711923e-01
-3.70687842e-01 -9.75252330e-01 -6.85717344e-01 -2.68557400e-01
4.76724058e-01 -8.85280788e-01 7.22893417e-01 5.36787689e-01
-8.18918347e-02 6.27690613e-01 2.43204728e-01 -4.19766940e-02
2.14924708e-01 3.32948625e-01 1.03440213e+00 -1.06207287e+00
-3.16689074e-01 -1.47068396e-01 1.63969383e-01 -1.62074220e+00
-5.25103658e-02 -8.78894329e-01 3.10313702e-01 -1.86745405e+00
6.43752739e-02 -2.18663178e-02 3.68357033e-01 4.98108089e-01
-1.45835593e-01 -5.52069664e-01 2.03974053e-01 4.24228683e-02
-4.77977782e-01 9.08071250e-02 6.19907677e-01 -1.28825620e-01
-2.40710318e-01 -1.62619613e-02 -9.59040463e-01 5.22841930e-01
7.24491179e-01 -6.82465315e-01 -2.14085683e-01 -9.13952708e-01
7.83105493e-01 6.07439339e-01 7.95210972e-02 -5.46141922e-01
1.15921952e-01 -5.54480478e-02 1.30583823e-01 -1.11958873e+00
2.71719784e-01 -4.65871483e-01 -3.46082866e-01 3.67512494e-01
-6.56407535e-01 6.39009178e-01 2.31839836e-01 1.96437061e-01
-8.42038214e-01 -4.21861678e-01 5.22890270e-01 -5.93354285e-01
-4.45045561e-01 -1.19574755e-01 -4.87641096e-01 5.26200771e-01
4.02303547e-01 2.64235400e-02 -4.22705531e-01 -5.44173896e-01
-4.96849954e-01 2.63648510e-01 -7.64016360e-02 5.00675023e-01
7.51816392e-01 -9.21088815e-01 -4.89648163e-01 3.92812133e-01
1.24059349e-01 5.75681925e-01 -4.47357833e-01 1.67918816e-01
-9.24171150e-01 1.06035388e+00 1.64600030e-01 -6.10831499e-01
-9.33407366e-01 4.21819746e-01 3.96507442e-01 -7.09851265e-01
-3.17465961e-01 7.23153234e-01 7.30411485e-02 -1.17104578e+00
4.13808942e-01 -9.03793812e-01 -2.52236542e-03 -2.87693560e-01
1.47518888e-01 -1.74974188e-01 4.29357857e-01 -1.74552038e-01
-4.56566274e-01 5.32180071e-01 -1.29245371e-01 -6.29021376e-02
1.41360331e+00 3.13729346e-01 -4.65494424e-01 4.45659578e-01
1.23325849e+00 -1.76607549e-01 -5.48524916e-01 -4.15012658e-01
5.80825210e-01 7.20875561e-02 -4.24573809e-01 -9.25581336e-01
-8.93608034e-01 1.27987623e+00 -3.55404094e-02 5.23848653e-01
8.97697508e-01 3.48942816e-01 1.09676957e+00 1.10816729e+00
2.72179395e-01 -1.13117850e+00 2.59916067e-01 9.18600738e-01
1.05219483e+00 -1.49506152e+00 -1.61147416e-01 -5.04837036e-01
-5.95387757e-01 1.74706280e+00 7.55556047e-01 2.73381025e-02
6.72941029e-01 5.46378553e-01 2.51710385e-01 -5.09622335e-01
-7.02708900e-01 2.28828356e-01 -9.81288403e-02 5.35394728e-01
9.47251201e-01 -2.67537057e-01 2.18143374e-01 9.67953324e-01
-5.34940600e-01 2.45511159e-01 4.51389104e-01 1.34164429e+00
-3.91823083e-01 -1.64281762e+00 -1.89997256e-01 5.72552562e-01
-7.16748118e-01 -3.59281033e-01 -7.72028506e-01 9.83215213e-01
-3.93966913e-01 1.14786065e+00 2.02680573e-01 -2.55221784e-01
7.31989026e-01 5.94814241e-01 3.16179954e-02 -1.11120188e+00
-1.03214705e+00 1.22615900e-02 5.10728776e-01 -7.34359443e-01
-1.11789100e-01 -3.41208369e-01 -1.72193098e+00 -1.17350683e-01
2.84214944e-01 5.31863002e-03 7.95644939e-01 7.35350668e-01
5.15043974e-01 3.84647697e-01 1.40969649e-01 -2.72238940e-01
-7.37494111e-01 -7.12088346e-01 9.83613655e-02 4.89053652e-02
7.69542754e-02 3.11763346e-01 2.09973902e-01 2.27738544e-01] | [9.98831558227539, 7.846551418304443] |
a2e0c286-a336-49a1-8137-2a47fdda1f79 | enlighten-anything-when-segment-anything | 2306.10286 | null | https://arxiv.org/abs/2306.10286v3 | https://arxiv.org/pdf/2306.10286v3.pdf | Enlighten Anything: When Segment Anything Model Meets Low-Light Image Enhancement | Image restoration is a low-level visual task, and most CNN methods are designed as black boxes, lacking transparency and intrinsic aesthetics. Many unsupervised approaches ignore the degradation of visible information in low-light scenes, which will seriously affect the aggregation of complementary information and also make the fusion algorithm unable to produce satisfactory fusion results under extreme conditions. In this paper, we propose Enlighten-anything, which is able to enhance and fuse the semantic intent of SAM segmentation with low-light images to obtain fused images with good visual perception. The generalization ability of unsupervised learning is greatly improved, and experiments on LOL dataset are conducted to show that our method improves 3db in PSNR over baseline and 8 in SSIM. Zero-shot learning of SAM introduces a powerful aid for unsupervised low-light enhancement. The source code of Enlighten Anything can be obtained from https://github.com/zhangbaijin/enlighten-anything | ['Shanying Zhu', 'Chaochen Gu', 'Hao Tang', 'Xiaofeng Zhang', 'Qihan Zhao'] | 2023-06-17 | null | null | null | null | ['image-enhancement', 'low-light-image-enhancement', 'image-restoration'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 2.24143907e-01 -3.72214347e-01 3.80304419e-02 -3.17553282e-01
-4.78151798e-01 -3.71642947e-01 5.48427254e-02 -9.63131636e-02
-2.10983261e-01 8.55970800e-01 2.47837484e-01 -1.33960485e-01
9.81861651e-02 -8.94772172e-01 -6.54926717e-01 -1.05534542e+00
6.15605950e-01 -6.27747059e-01 2.87255079e-01 -3.43333811e-01
1.30872861e-01 2.40431115e-01 -1.50144100e+00 5.52289248e-01
1.10643101e+00 7.95763791e-01 5.18505573e-01 4.95615184e-01
-1.50258139e-01 8.11296225e-01 -4.82293874e-01 -3.53576779e-01
4.22256589e-01 -4.21997875e-01 -4.13708687e-01 2.12294489e-01
5.73845506e-01 -6.00160122e-01 -4.52457845e-01 1.46453118e+00
6.47039533e-01 7.39799365e-02 3.75369370e-01 -1.08887899e+00
-8.98834527e-01 3.74348938e-01 -6.55671299e-01 2.28758782e-01
1.06001817e-01 4.26792771e-01 8.28788936e-01 -9.73810196e-01
3.45824391e-01 1.09225500e+00 4.37772453e-01 3.89753520e-01
-1.14714932e+00 -8.07204425e-01 -7.01738149e-02 2.71034569e-01
-9.74170983e-01 -5.94889581e-01 9.39675510e-01 2.06317324e-02
4.70337510e-01 3.36543173e-01 6.83614135e-01 8.61275136e-01
4.19673949e-01 7.92217851e-01 1.39553654e+00 -4.11545873e-01
8.59649032e-02 -3.08067389e-02 -9.96005237e-02 6.72439814e-01
5.25369883e-01 3.40315968e-01 -6.57902122e-01 4.59941447e-01
6.46210968e-01 5.08169949e-01 -4.18840945e-01 1.87262237e-01
-1.00473809e+00 3.46752971e-01 9.06415403e-01 1.87786445e-01
-3.34035039e-01 5.60956355e-03 -1.09256566e-01 2.26051211e-01
6.31307423e-01 2.57480115e-01 -8.01039860e-02 3.10889244e-01
-8.27539146e-01 -1.50861055e-01 1.83021545e-01 5.71676135e-01
9.66208398e-01 3.90367806e-01 -1.21247947e-01 9.21182573e-01
2.20063791e-01 7.31287181e-01 -4.46149781e-02 -1.45516825e+00
3.25399265e-02 4.79910791e-01 -2.98768375e-02 -9.92195427e-01
-2.13640288e-01 -6.38950706e-01 -1.00521564e+00 7.48848557e-01
1.62391961e-01 -1.78828850e-01 -9.95151103e-01 1.39319384e+00
4.86668646e-02 -3.87225747e-02 1.50423855e-01 1.08272552e+00
1.28856313e+00 8.71377945e-01 -4.35603447e-02 -3.28922063e-01
1.16831934e+00 -1.23204732e+00 -1.09479916e+00 -4.27105039e-01
-4.27509956e-02 -1.19320345e+00 1.14701176e+00 4.91981536e-01
-1.39967573e+00 -6.94130599e-01 -1.18709946e+00 -3.25226814e-01
-1.48375198e-01 1.65055230e-01 7.66755581e-01 5.08652151e-01
-1.04156041e+00 4.93112922e-01 -5.53349614e-01 -2.36414760e-01
8.06670487e-01 -2.22996753e-02 -2.71852642e-01 -4.82155323e-01
-8.59978557e-01 6.75054193e-01 3.43008637e-01 3.65504205e-01
-9.16760683e-01 -4.82855290e-01 -5.57982981e-01 -1.45137608e-01
4.96292382e-01 -5.79248548e-01 8.88627827e-01 -8.34372461e-01
-1.28131068e+00 6.28172398e-01 -3.36452514e-01 -5.77310212e-02
2.61445910e-01 -1.52222231e-01 -4.18188691e-01 3.96768928e-01
1.14529766e-01 7.58543491e-01 9.25016701e-01 -1.87382102e+00
-5.19821644e-01 -1.82101712e-01 2.23028153e-01 3.03037226e-01
-5.49009562e-01 -1.21000655e-01 -3.88839155e-01 -7.03517437e-01
3.99285048e-01 -5.17440140e-01 -1.61048263e-01 1.52197137e-01
-3.65919828e-01 2.70536125e-01 9.79609489e-01 -8.12713861e-01
8.24506640e-01 -2.26286125e+00 -5.60064375e-01 -3.53356242e-01
2.81484962e-01 3.97524595e-01 -1.58720180e-01 2.88496733e-01
-3.22354324e-02 8.47249329e-02 -4.72073793e-01 -1.64410949e-01
-3.93539041e-01 2.26012707e-01 -1.18775390e-01 5.03655732e-01
6.12522587e-02 9.72366035e-01 -8.89918804e-01 -5.32669961e-01
5.09353697e-01 6.38696849e-01 -3.37267965e-01 1.77897036e-01
3.15018483e-02 6.82923079e-01 -2.11182654e-01 1.02985144e+00
1.01194859e+00 -8.69731531e-02 -1.86028361e-01 -7.97548890e-01
-1.63294733e-01 -2.18757674e-01 -9.30556476e-01 1.69971001e+00
-4.19657588e-01 8.06088626e-01 9.82736871e-02 -6.40573323e-01
8.51100147e-01 5.80668785e-02 3.83339942e-01 -1.08433497e+00
3.30461800e-01 1.75003670e-02 -2.02357382e-01 -5.73260307e-01
4.15638119e-01 -1.84552327e-01 4.12682950e-01 2.81785846e-01
-1.60873324e-01 -3.42494905e-01 2.21323758e-01 3.14961433e-01
6.62649572e-01 4.87639382e-02 -4.40682285e-02 -8.24230984e-02
2.92304963e-01 -2.80489653e-01 7.79309392e-01 6.77303016e-01
-2.33043611e-01 7.27098525e-01 -5.57682849e-02 -2.40452051e-01
-9.72731173e-01 -1.40889323e+00 -5.91971837e-02 9.46901083e-01
7.68499970e-01 -1.61076605e-01 -5.24245739e-01 -2.12050363e-01
-4.51093137e-01 8.52405131e-01 -1.68373033e-01 -3.26768577e-01
-1.64452076e-01 -8.39143217e-01 1.48674369e-01 2.73661524e-01
1.09940720e+00 -1.03472531e+00 -4.26134944e-01 -6.05826490e-02
-6.66372359e-01 -1.22247112e+00 -2.68455833e-01 2.42314577e-01
-8.69201481e-01 -8.88938487e-01 -5.47674358e-01 -8.64381015e-01
8.79740953e-01 9.00549054e-01 8.97323072e-01 3.32772762e-01
-4.89388108e-01 8.76274407e-02 -4.53803688e-01 -5.62100351e-01
-1.84968427e-01 -5.63765943e-01 -2.33082548e-01 -1.51619865e-02
8.23311061e-02 -7.39491105e-01 -9.71255302e-01 2.09661528e-01
-1.12268543e+00 4.03730333e-01 7.29388535e-01 8.48539591e-01
5.11391699e-01 7.03491211e-01 3.39400858e-01 -5.85398316e-01
4.42601502e-01 1.24490157e-01 -2.99991727e-01 2.27194186e-02
-6.73697650e-01 -3.07546854e-01 5.07254779e-01 -1.77698523e-01
-1.64195824e+00 -5.87505586e-02 -1.22826040e-01 -2.02344120e-01
-3.22748125e-01 3.00150365e-02 -3.70576054e-01 -2.99776316e-01
6.35271847e-01 3.11513275e-01 -1.87597722e-01 -5.10718644e-01
4.67807680e-01 6.77893937e-01 7.86930203e-01 -2.71112233e-01
8.49854648e-01 9.61698532e-01 7.51558915e-02 -7.93773413e-01
-1.07171357e+00 -3.18806052e-01 -4.69701350e-01 -5.41533887e-01
9.18359935e-01 -1.14201772e+00 -7.08951712e-01 6.20421231e-01
-8.78889799e-01 -4.00844902e-01 -3.96573275e-01 3.71028990e-01
-2.74087548e-01 5.60444832e-01 -8.19519579e-01 -7.84176946e-01
-3.83489728e-01 -1.14601707e+00 6.77495003e-01 7.92304814e-01
4.17345762e-01 -7.57692814e-01 -4.88089800e-01 8.54721606e-01
4.30467606e-01 6.77212998e-02 7.40872562e-01 1.73051327e-01
-7.62372732e-01 1.66357905e-02 -6.03228927e-01 7.87266731e-01
2.19823882e-01 8.81395265e-02 -1.26441121e+00 -3.11052799e-01
1.17137395e-01 -2.13440359e-01 1.27052665e+00 5.74430764e-01
1.17572904e+00 -1.35066390e-01 6.02127537e-02 7.66731918e-01
1.64694953e+00 3.53287607e-01 1.03595066e+00 3.36252719e-01
7.33191311e-01 5.41686177e-01 5.87016761e-01 2.80584365e-01
2.59082943e-01 2.35610053e-01 8.97748351e-01 -1.03741956e+00
-7.82844901e-01 -3.03634852e-02 3.99230480e-01 7.12868154e-01
-3.19939405e-01 -2.50722408e-01 -5.33361614e-01 5.20219505e-01
-1.60626376e+00 -9.49650586e-01 -5.12714267e-01 1.88879728e+00
8.96741569e-01 -1.17476517e-02 -2.95245260e-01 1.50314584e-01
7.92176008e-01 2.66878814e-01 -3.79372388e-01 -1.70476250e-02
-6.73572659e-01 4.73889299e-02 5.96970618e-01 4.97211277e-01
-9.55199718e-01 1.00120139e+00 5.85100842e+00 8.86177480e-01
-1.10005486e+00 3.74603599e-01 7.58007765e-01 -1.60262778e-01
-2.50264168e-01 1.27511173e-01 -3.51944208e-01 5.18202841e-01
2.84426033e-01 4.85972650e-02 4.67150748e-01 2.38641784e-01
5.53341568e-01 -4.82997090e-01 -4.11466599e-01 1.06852055e+00
1.53544754e-01 -1.16056812e+00 -8.78094509e-02 -4.26678285e-02
1.17360663e+00 -1.60943314e-01 3.40317130e-01 -3.08371276e-01
3.46387357e-01 -8.19245636e-01 4.67840552e-01 7.73444951e-01
6.49239600e-01 -7.47889876e-01 7.14073062e-01 1.68238625e-01
-9.27602649e-01 -2.20527962e-01 -6.21278226e-01 1.03481078e-04
3.10861170e-01 8.93410504e-01 -3.87152642e-01 7.24941254e-01
1.02923238e+00 9.38645542e-01 -7.10758507e-01 1.23655581e+00
-6.75754845e-01 6.37623131e-01 1.21349826e-01 4.46684837e-01
-1.69050379e-03 -3.15935165e-01 6.88028157e-01 1.06660783e+00
1.99792489e-01 1.43500820e-01 2.22934663e-01 8.98515522e-01
-6.32166415e-02 -8.41933340e-02 -4.00787324e-01 -3.69629823e-02
5.03904410e-02 1.49577069e+00 -9.21579003e-01 -2.94076592e-01
-4.03682321e-01 1.12401247e+00 -4.66443360e-01 6.99084580e-01
-6.07222319e-01 -3.00131559e-01 2.67810613e-01 1.14051975e-01
-5.84679432e-02 -1.54150814e-01 -6.97155893e-01 -1.18892956e+00
-3.81447189e-03 -6.99007630e-01 1.48055553e-01 -1.42126918e+00
-1.22266221e+00 4.24869746e-01 -2.81671852e-01 -1.46285045e+00
6.18168890e-01 -4.61343288e-01 -6.49116099e-01 6.30750597e-01
-1.71829522e+00 -1.40920568e+00 -8.41567695e-01 7.68104017e-01
7.08491087e-01 -5.90318330e-02 3.39065224e-01 3.85492384e-01
-4.23849165e-01 1.58858940e-01 3.31175566e-01 -8.52101743e-02
1.00337422e+00 -1.03743231e+00 -2.00347885e-01 1.42899680e+00
4.72800583e-02 4.22267526e-01 7.61666179e-01 -6.80881560e-01
-1.12070239e+00 -1.10827100e+00 4.61943477e-01 -9.13223922e-02
2.99838632e-01 7.60594681e-02 -8.33955526e-01 2.82923162e-01
6.56867445e-01 -4.42788564e-02 6.35364830e-01 -2.99600214e-01
-2.46074378e-01 -4.56558377e-01 -1.11325967e+00 8.21338058e-01
9.42080617e-01 -4.74174768e-01 -3.52392882e-01 4.82178360e-01
7.29986966e-01 -8.02310035e-02 -6.42957747e-01 5.76920807e-01
4.71497536e-01 -1.43003643e+00 1.28331923e+00 -2.20582969e-02
7.82783568e-01 -6.27740324e-01 -3.67978930e-01 -1.17195117e+00
-2.85770833e-01 -2.39856184e-01 2.64382869e-01 1.25340116e+00
2.54544318e-01 -6.37968123e-01 4.66136009e-01 9.23967138e-02
-2.66150713e-01 -5.12570143e-01 -4.24155235e-01 -5.72055101e-01
-2.49838591e-01 -4.02977258e-01 3.50076318e-01 8.38854909e-01
-4.25081551e-01 2.42328227e-01 -5.66373587e-01 2.83357114e-01
1.12458825e+00 2.04225257e-01 5.31868815e-01 -9.56768274e-01
-9.13982168e-02 -2.72564054e-01 -1.86338022e-01 -6.42723441e-01
-1.35399684e-01 -8.51072967e-01 1.81802735e-01 -2.14405298e+00
5.09211004e-01 7.44395051e-03 -4.67939138e-01 7.84402788e-01
-2.08330631e-01 1.01997650e+00 4.33303595e-01 3.92756164e-01
-6.97145760e-01 5.96225798e-01 1.72131050e+00 -4.50865567e-01
-1.04360692e-01 -1.99809864e-01 -9.72568810e-01 8.64451528e-01
8.13992500e-01 -3.09603930e-01 -4.35021400e-01 -4.85470563e-01
1.06843956e-01 -1.76212937e-01 6.55880451e-01 -1.07928336e+00
3.54633301e-01 -2.45062321e-01 7.47063041e-01 -6.11276150e-01
5.08723795e-01 -7.70815194e-01 9.75857079e-02 3.11589479e-01
-8.14803541e-02 -5.43816328e-01 1.46999180e-01 3.63343090e-01
-2.53396004e-01 -2.52113074e-01 1.03180051e+00 -2.89441466e-01
-8.48552883e-01 1.58889338e-01 -2.38279939e-01 -3.41806322e-01
7.42257893e-01 -4.33105856e-01 -7.76210666e-01 -5.83429813e-01
-6.50696099e-01 -2.22595409e-02 7.70947576e-01 1.60029814e-01
8.70863438e-01 -1.24063933e+00 -7.35670507e-01 1.90079361e-01
-1.40100345e-02 -1.72494292e-01 8.16473603e-01 9.06814337e-01
-6.25467837e-01 -1.19404145e-01 -5.86356342e-01 -4.40733850e-01
-1.38665676e+00 3.33405018e-01 9.36204419e-02 2.77894765e-01
-6.98231220e-01 8.46439481e-01 5.41752577e-01 -2.66627744e-02
7.82219972e-03 7.64472112e-02 -5.13230376e-02 -1.59002513e-01
6.08819902e-01 4.76906836e-01 4.94425930e-02 -6.12365425e-01
-1.48757309e-01 5.30462265e-01 4.18011174e-02 3.44267525e-02
1.49975693e+00 -5.50711811e-01 -4.17678565e-01 2.50664383e-01
9.55681205e-01 1.91550702e-01 -1.37022388e+00 -1.52519733e-01
-8.01421702e-01 -9.46505308e-01 5.48153520e-01 -1.05027103e+00
-1.43036294e+00 1.15856731e+00 9.19555664e-01 9.44277719e-02
1.76246095e+00 -1.12655662e-01 8.96675229e-01 2.71481544e-01
1.00219235e-01 -9.89113688e-01 5.07473946e-01 8.14323500e-02
7.61824369e-01 -1.37605810e+00 2.13882461e-01 -6.29064083e-01
-6.53350353e-01 1.04833508e+00 6.19174898e-01 3.51937860e-02
3.60161841e-01 3.51746857e-01 3.42241049e-01 -6.92422464e-02
-4.92917418e-01 -5.38458169e-01 1.17193192e-01 8.14035654e-01
3.10902297e-01 -1.74219161e-01 -2.39294216e-01 3.18476021e-01
3.62003744e-02 -1.29227459e-01 7.43203342e-01 7.95262456e-01
-7.08873630e-01 -8.73936296e-01 -5.07672310e-01 2.58846819e-01
-5.23193777e-01 -2.85335392e-01 -1.48740888e-01 2.68015891e-01
3.89273673e-01 1.48439634e+00 -2.34173432e-01 -4.97457623e-01
9.64086428e-02 -4.24140692e-01 3.42530996e-01 -2.89041042e-01
-1.31315008e-01 5.59699178e-01 -1.77953258e-01 -5.65645158e-01
-7.42688358e-01 -1.26013175e-01 -1.22769392e+00 -3.50247562e-01
-2.54112452e-01 -3.18973273e-01 6.51348591e-01 7.90342510e-01
1.48183322e-02 7.87070930e-01 6.89327121e-01 -8.63690615e-01
2.41662756e-01 -6.27651632e-01 -8.08585525e-01 3.59445691e-01
2.87522078e-01 -5.29104888e-01 -5.13563693e-01 4.43492591e-01] | [10.823568344116211, -2.4409420490264893] |
7d43666f-399e-43b7-9000-ba1c17c22b2c | detection-of-chinese-stock-market-bubbles | 1905.09640 | null | http://arxiv.org/abs/1905.09640v2 | http://arxiv.org/pdf/1905.09640v2.pdf | Detection of Chinese Stock Market Bubbles with LPPLS Confidence Indicator | We present an advance bubble detection methodology based on the Log Periodic
Power Law Singularity (LPPLS) confidence indicator for the early causal
identification of positive and negative bubbles in the Chinese stock market
using the daily data on the Shanghai Shenzhen CSI 300 stock market index from
January 2002 through April 2018. We account for the damping condition of LPPLS
model in the search space and implement the stricter filter conditions for the
qualification of the valid LPPLS fits by taking account of the maximum relative
error, performing the Lomb log-periodic test of the detrended residual, and
unit-root tests of the logarithmic residual based on both the Phillips-Perron
test and Dickey-Fuller test to improve the performance of LPPLS confidence
indicator. Our analysis shows that the LPPLS detection strategy diagnoses the
positive bubbles and negative bubbles corresponding to well-known historical
events, implying the detection strategy based on the LPPLS confidence indicator
has an outstanding performance to identify the bubbles in advance. We find that
the probability density distribution of the estimated beginning time of bubbles
appears to be skewed and the mass of the distribution is concentrated on the
area where the price starts to have an obvious super-exponentially growth. This
study is the first work in the literature that identifies the existence of
bubbles in the Chinese stock market using the daily data of CSI 300 index with
the advance bubble detection methodology of LPPLS confidence indicator. We have
shown that it is possible to detect the potential positive and negative bubbles
and crashes ahead of time, which in turn limits the bubble sizes and eventually
minimizes the damages from the bubble crash. | [] | 2019-06-13 | null | null | null | null | ['causal-identification'] | ['reasoning'] | [-8.05538595e-01 -1.42447233e-01 -7.01035634e-02 6.06437027e-01
-4.14908767e-01 -7.87444651e-01 4.84592915e-01 1.96002260e-01
3.85166355e-03 8.46813560e-01 1.04302138e-01 -9.98209476e-01
-1.60632640e-01 -9.90909874e-01 -5.31794548e-01 -8.58253181e-01
-5.99910676e-01 2.26750240e-01 4.65837777e-01 -1.42012676e-02
5.31449974e-01 1.91849887e-01 -1.22607481e+00 -4.53888923e-01
1.01338601e+00 1.11843956e+00 1.21320508e-01 4.90479738e-01
2.27816164e-01 6.09871268e-01 -5.04092991e-01 -7.66302943e-02
5.66222668e-01 -3.90979379e-01 -1.04461744e-01 -4.84729111e-01
-4.73497778e-01 -4.38358396e-01 1.28337562e-01 1.15099514e+00
2.93741643e-01 -1.39030114e-01 4.34414655e-01 -1.22786760e+00
-6.43492997e-01 9.14555490e-01 -7.50151396e-01 1.03430033e+00
2.72119105e-01 -4.43584938e-03 1.13927305e+00 -1.17963529e+00
2.13214636e-01 6.53567493e-01 6.55537546e-01 -3.98140222e-01
-6.92508459e-01 -6.55117035e-01 -2.96763659e-01 1.00234404e-01
-1.19096053e+00 2.16306299e-01 6.35015786e-01 -8.78934801e-01
7.50699043e-01 3.45857948e-01 9.95789349e-01 2.89440006e-01
4.63711172e-01 2.27250218e-01 9.07098472e-01 -3.40091705e-01
3.30808192e-01 9.00709853e-02 3.03851277e-01 1.49717465e-01
8.31524849e-01 6.05526209e-01 -3.18495244e-01 -3.75044644e-01
8.79707515e-01 -1.88023984e-01 -1.31723419e-01 4.87247258e-01
-8.93260419e-01 1.13510907e+00 8.40994269e-02 7.56902039e-01
-4.14254487e-01 -3.55503887e-01 2.56154597e-01 1.58089340e-01
8.55279207e-01 3.95652562e-01 -6.27868831e-01 -2.14832932e-01
-1.01709783e+00 -1.57099981e-02 9.25235569e-01 5.60821652e-01
3.36200416e-01 2.59787619e-01 3.35316882e-02 -2.20342740e-01
2.50861168e-01 9.69443321e-01 7.26178169e-01 -6.04086578e-01
4.32125360e-01 3.20601672e-01 4.06575054e-01 -9.25029933e-01
-1.41390592e-01 -7.71587431e-01 -5.29244304e-01 3.43419403e-01
5.14551640e-01 -4.45051879e-01 -8.78880844e-02 1.36313689e+00
3.76555435e-02 4.75969344e-01 1.29226789e-01 4.46792662e-01
2.38775015e-01 1.11534071e+00 -3.95301312e-01 -9.22258496e-01
1.27622044e+00 -4.54002023e-01 -6.25122190e-01 1.98676184e-01
4.35602367e-01 -3.39905620e-01 8.62284720e-01 3.64633590e-01
-9.23307061e-01 -2.56904781e-01 -9.95066285e-01 6.61098123e-01
-3.64289954e-02 -4.12492529e-02 3.29778939e-01 3.83742154e-01
-7.13287413e-01 4.96018142e-01 -9.96797681e-01 2.74080962e-01
-4.60565239e-01 -1.06180809e-01 5.37585430e-02 9.72953260e-01
-1.34539771e+00 7.55830467e-01 3.90957981e-01 4.22102064e-01
-4.41452771e-01 -8.46383691e-01 -5.78957379e-01 4.96785522e-01
2.20512804e-02 9.53543112e-02 7.36879468e-01 -7.86165357e-01
-9.97656286e-01 2.05854028e-01 9.93630514e-02 -8.49102855e-01
5.62702358e-01 1.44714922e-01 -2.59669602e-01 2.49637097e-01
3.09981138e-01 -5.99495769e-01 4.28041607e-01 -6.89606607e-01
-9.66806829e-01 -5.00894263e-02 -4.93975937e-01 -2.60055333e-01
-2.58691013e-01 1.46776006e-01 5.49037158e-01 -8.06906104e-01
4.01409753e-02 -5.51397741e-01 1.87382132e-01 -9.71616209e-01
-7.27667734e-02 -6.32007420e-01 4.95127857e-01 -1.31196463e+00
1.74029636e+00 -2.41181564e+00 -6.43036485e-01 3.98223102e-01
-6.21634945e-02 -1.47322223e-01 5.77910781e-01 4.70339447e-01
-3.74721557e-01 4.23014402e-01 -1.29234344e-01 1.79550543e-01
5.28579205e-02 -1.81331620e-01 -8.78137350e-01 7.76243031e-01
3.81865650e-02 4.67624277e-01 -7.25447536e-01 -1.29098639e-01
-2.52975207e-02 -2.78655410e-01 -2.39883229e-01 1.34521320e-01
3.20170633e-02 1.60778612e-01 -1.46402925e-01 4.34687227e-01
8.11385989e-01 -1.92843467e-01 -4.78808433e-01 3.50151658e-01
-1.05085707e+00 2.19165027e-01 -1.01380765e+00 2.87351429e-01
1.56326607e-01 6.94947839e-01 -2.10531458e-01 -8.40656877e-01
9.55242515e-01 6.50374055e-01 5.61654389e-01 -6.34799361e-01
-1.51753262e-01 7.10424185e-01 9.00036655e-03 -6.44434512e-01
1.81385010e-01 -3.85817885e-01 6.78085387e-02 3.55820626e-01
-3.96449327e-01 1.19618520e-01 6.68971121e-01 5.69203822e-03
6.86381757e-01 -3.18828106e-01 4.17318135e-01 -6.85467303e-01
4.56773013e-01 -1.77290961e-01 7.19740391e-01 3.43756378e-01
-6.75409138e-02 3.54305178e-01 1.09441090e+00 3.64965647e-02
-9.77006733e-01 -9.55571294e-01 -3.41636986e-01 5.41598439e-01
-5.52654639e-02 2.31057599e-01 -3.07732642e-01 -1.84359521e-01
1.83338419e-01 1.07126737e+00 -7.00608492e-01 2.42888018e-01
-4.64339882e-01 -1.10826826e+00 -3.97462398e-02 3.70520741e-01
3.80766124e-01 -1.11761534e+00 -7.87931383e-01 1.17894523e-01
-1.83440745e-02 -5.41037917e-01 -1.71028450e-01 1.79230571e-01
-6.74248815e-01 -1.11482513e+00 -6.61014259e-01 -8.22408736e-01
5.51808894e-01 -2.09115833e-01 7.60339975e-01 9.43582207e-02
4.63188529e-01 -1.48460150e-01 -3.87341499e-01 -4.87937957e-01
-3.94578546e-01 -3.52076083e-01 -1.11477353e-01 -4.61194292e-02
1.00864865e-01 -5.02839327e-01 -8.43765020e-01 2.38309920e-01
-5.26574194e-01 -3.72940511e-01 4.58351299e-02 6.11915946e-01
1.18147828e-01 6.92748129e-01 1.09290183e+00 7.49428570e-02
5.73215544e-01 -9.71185565e-01 -1.39836526e+00 2.05004826e-01
-1.05683947e+00 -2.03547761e-01 2.64361233e-01 -4.73581433e-01
-8.94813001e-01 -5.62245727e-01 8.19108635e-02 -3.22851818e-03
5.06808043e-01 1.19392836e+00 3.97772700e-01 3.01451653e-01
2.23237276e-01 1.54648557e-01 -8.17310810e-02 -4.31062162e-01
-2.91660488e-01 4.24763262e-01 6.38038695e-01 2.98639182e-02
9.21485186e-01 4.20347273e-01 -1.23722978e-01 -5.36491394e-01
-2.96513379e-01 -6.84151530e-01 -2.69765347e-01 -1.59743354e-01
7.32312500e-01 -9.56659973e-01 -8.48882794e-01 7.54179418e-01
-8.46516192e-01 -1.22238331e-01 -3.68541449e-01 7.97277868e-01
-7.57796988e-02 4.10776794e-01 -1.00702739e+00 -1.40079665e+00
-3.57487679e-01 -7.03748465e-01 1.43126458e-01 4.26938832e-01
9.46865752e-02 -1.30899394e+00 4.35436547e-01 -3.38461131e-01
2.27709740e-01 6.13799632e-01 7.71366358e-01 -1.08674276e+00
-3.84456426e-01 -4.91224378e-01 1.22329041e-01 3.92837197e-01
8.42690766e-02 7.02341735e-01 -3.05908710e-01 -3.91423881e-01
8.20172608e-01 6.50399208e-01 5.86281061e-01 8.03451955e-01
-1.63505618e-02 -6.23461843e-01 1.27571980e-02 1.78477526e-01
1.65238202e+00 6.31351411e-01 3.21308851e-01 6.61876559e-01
-1.16487324e-01 3.42399120e-01 6.22020245e-01 8.55700970e-01
1.49066389e-01 -5.25305979e-02 4.98430550e-01 -1.53150298e-02
5.61793923e-01 -2.59327769e-01 7.44260252e-01 9.11317289e-01
4.73910943e-03 -1.27828896e-01 -9.77749765e-01 9.27162409e-01
-1.52487326e+00 -1.26107419e+00 -6.70477033e-01 2.52054286e+00
6.87533081e-01 5.01769602e-01 4.27429050e-01 3.48500460e-01
7.99832046e-01 -1.61498070e-01 -1.04540668e-01 -1.20094836e-01
-3.15830410e-01 -3.93365324e-02 8.24698269e-01 7.25932837e-01
-7.79040933e-01 9.55155119e-02 6.31313705e+00 5.94518960e-01
-1.21149671e+00 1.37747958e-01 6.30927622e-01 3.55905682e-01
-4.60988700e-01 3.87142986e-01 -7.68462181e-01 1.03437257e+00
9.11896229e-01 -9.46333826e-01 -5.70433810e-02 7.01116860e-01
7.60701120e-01 -6.97952151e-01 -4.63924229e-01 4.49017018e-01
-2.74047196e-01 -1.09912348e+00 -5.74005544e-01 4.35984373e-01
7.47935116e-01 -8.28808546e-02 1.23606548e-01 -6.39454322e-03
-1.35678530e-01 -5.37173808e-01 1.16606617e+00 6.34844184e-01
3.23193401e-01 -6.86456144e-01 1.16568935e+00 6.34444654e-01
-1.19510400e+00 -5.11776865e-01 -1.90427586e-01 -5.41545331e-01
4.25780773e-01 1.04946280e+00 -8.07684481e-01 3.97781760e-01
5.53288221e-01 5.51979601e-01 -4.34016913e-01 1.27566767e+00
-2.57512987e-01 1.23913658e+00 -7.16385782e-01 5.45357540e-02
-3.23114619e-02 -7.30984449e-01 6.32379651e-01 5.04285157e-01
1.05962157e+00 1.78876426e-02 -2.98143715e-01 9.62437451e-01
6.25617981e-01 -6.47452995e-02 -3.73655856e-01 -1.69794448e-02
5.80926836e-01 5.92015803e-01 -9.33423221e-01 -2.77008593e-01
-7.08249986e-01 3.19620967e-02 -4.75585192e-01 2.92880207e-01
-7.47242033e-01 -1.65817812e-01 -1.64377883e-01 4.00830925e-01
5.93234599e-01 -3.47662866e-01 -9.72611904e-01 -1.01086128e+00
1.89138010e-01 -3.15230101e-01 5.12287498e-01 -5.65514684e-01
-1.12586796e+00 3.01950097e-01 7.03853890e-02 -1.18770266e+00
-7.11818933e-01 -3.20544481e-01 -1.30328381e+00 1.02433491e+00
-1.32936823e+00 -3.21575761e-01 4.09641922e-01 2.46707246e-01
1.34699300e-01 4.85959882e-03 1.70953423e-01 -3.43244001e-02
-6.42574251e-01 -4.20100354e-02 5.47923446e-01 2.96304435e-01
3.04394290e-02 -1.25587845e+00 2.71712035e-01 1.45670545e+00
-2.24486768e-01 4.45461929e-01 1.20589888e+00 -1.23289955e+00
-6.30012393e-01 -5.36660142e-02 1.20758235e+00 -1.57622591e-01
1.37691605e+00 -9.87171605e-02 -9.16264832e-01 7.03300238e-01
3.20829004e-01 -3.09597552e-01 1.59211695e-01 -4.38541532e-01
3.79846603e-01 -2.33354762e-01 -8.68332565e-01 5.34268692e-02
-1.49914160e-01 -1.83613256e-01 -7.87185609e-01 4.58333462e-01
6.57387316e-01 1.19377755e-01 -9.16506827e-01 5.83464861e-01
3.93895656e-01 -1.16053414e+00 4.86725777e-01 1.96667656e-01
-4.43237610e-02 -1.98814884e-01 3.92208159e-01 -8.99188936e-01
-3.73168170e-01 -8.44543934e-01 6.67479038e-02 1.41089046e+00
6.80877388e-01 -1.08920443e+00 2.54760593e-01 6.16699040e-01
1.75291598e-01 -4.20545042e-01 -1.48143256e+00 -7.16636956e-01
2.75814325e-01 -1.46197602e-02 4.91472036e-01 6.05018079e-01
4.87380654e-01 -2.87847549e-01 -1.40064955e-01 5.26813269e-01
4.74711448e-01 3.62010837e-01 2.73163527e-01 -1.38312995e+00
-1.41298801e-01 -6.05178595e-01 1.06635511e-01 -5.17681003e-01
-8.98375586e-02 -2.03389972e-01 -1.05787464e-03 -9.84997332e-01
6.52625188e-02 -2.16339231e-01 -3.08592588e-01 -8.51001889e-02
-1.27095520e-01 -2.54025638e-01 7.88512360e-03 3.56091321e-01
3.50124061e-01 4.11058873e-01 6.79655790e-01 3.62920403e-01
-2.74458557e-01 5.64319134e-01 -2.30195165e-01 5.51831543e-01
6.80564344e-01 -3.75623316e-01 -2.57328063e-01 4.09693718e-01
9.84552085e-01 5.84916353e-01 4.10299540e-01 -7.89845467e-01
2.25971192e-01 -2.10748374e-01 -1.73602864e-01 -1.15959454e+00
-5.90202928e-01 -5.90175331e-01 3.30886841e-01 9.33640540e-01
2.91645229e-01 5.19933641e-01 -2.67110169e-02 4.04074758e-01
-4.03961092e-01 -5.97733498e-01 2.93815047e-01 -6.37135953e-02
-1.77881598e-01 4.16943245e-02 -6.63754702e-01 -2.02024579e-02
1.26104736e+00 -3.30734923e-02 -3.37801158e-01 -4.57379818e-01
-5.98428905e-01 3.90597761e-01 1.94399118e-01 1.23453410e-02
2.82027692e-01 -1.01898754e+00 -9.67977583e-01 3.11191559e-01
-7.42428660e-01 -4.30360168e-01 -9.73625183e-02 1.24981940e+00
-7.95857728e-01 4.17646915e-01 1.87909812e-01 -2.19605610e-01
-7.52121389e-01 9.84283328e-01 5.02989173e-01 -3.90452087e-01
-5.04060626e-01 5.98763287e-01 1.36107355e-01 6.38377130e-01
-2.14200038e-02 -7.19250739e-01 -4.58974242e-01 7.00926483e-01
4.93815064e-01 8.24732184e-01 -1.38268322e-01 -4.33170319e-01
-3.29087287e-01 3.18019092e-01 8.24840307e-01 -3.85133862e-01
1.42714155e+00 -3.98308814e-01 -6.35477543e-01 9.43025053e-01
7.62472808e-01 6.21074617e-01 -1.41648769e+00 3.15366626e-01
4.79517281e-01 -5.72202988e-02 -3.69941533e-01 -4.53463614e-01
-9.39291835e-01 4.22940969e-01 2.24771127e-01 1.22105944e+00
9.03238773e-01 -2.31471658e-02 6.46445155e-01 -5.83599508e-01
1.66208461e-01 -9.44232404e-01 -3.72201204e-01 3.94640416e-01
9.80345726e-01 -9.21834230e-01 -6.58917204e-02 -1.05347045e-01
-3.34649861e-01 1.09006524e+00 -1.24008441e-02 -5.72239637e-01
1.19715881e+00 4.28562015e-01 -2.63052821e-01 1.71414278e-02
-5.41204572e-01 8.53750482e-02 3.80712263e-02 -2.53083613e-02
-5.53778000e-02 1.59829527e-01 -1.06409061e+00 8.53447676e-01
-5.38939834e-01 -2.51571178e-01 6.50315464e-01 4.45840567e-01
-6.06346130e-01 -3.37950706e-01 -6.81650877e-01 1.32011876e-01
-7.83145189e-01 -3.57838929e-01 2.59795904e-01 8.35416853e-01
-4.17429628e-03 1.05251539e+00 4.81621593e-01 1.20154262e-01
-8.03407505e-02 9.12171900e-02 -3.82306904e-01 -7.19049796e-02
-5.57983994e-01 3.90428156e-01 -2.03788251e-01 1.99784458e-01
-2.66262323e-01 -1.10137081e+00 -1.40409803e+00 -3.53288293e-01
-8.73572588e-01 8.75912607e-01 3.10177386e-01 1.17869353e+00
-1.02801725e-01 -9.81339142e-02 1.07723916e+00 -5.65198183e-01
-8.98783624e-01 -1.31348264e+00 -1.12155879e+00 3.28257382e-02
8.42329144e-01 -4.19584423e-01 -1.46133566e+00 -1.55380413e-01] | [4.740936279296875, 4.183809280395508] |
e5cae96e-4223-499d-96df-b67380e07455 | eventea-benchmarking-entity-alignment-for | 2211.02817 | null | https://arxiv.org/abs/2211.02817v1 | https://arxiv.org/pdf/2211.02817v1.pdf | EventEA: Benchmarking Entity Alignment for Event-centric Knowledge Graphs | Entity alignment is to find identical entities in different knowledge graphs (KGs) that refer to the same real-world object. Embedding-based entity alignment techniques have been drawing a lot of attention recently because they can help solve the issue of symbolic heterogeneity in different KGs. However, in this paper, we show that the progress made in the past was due to biased and unchallenging evaluation. We highlight two major flaws in existing datasets that favor embedding-based entity alignment techniques, i.e., the isomorphic graph structures in relation triples and the weak heterogeneity in attribute triples. Towards a critical evaluation of embedding-based entity alignment methods, we construct a new dataset with heterogeneous relations and attributes based on event-centric KGs. We conduct extensive experiments to evaluate existing popular methods, and find that they fail to achieve promising performance. As a new approach to this difficult problem, we propose a time-aware literal encoder for entity alignment. The dataset and source code are publicly available to foster future research. Our work calls for more effective and practical embedding-based solutions to entity alignment. | ['Wei Hu', 'Guangyao Li', 'Zequn Sun', 'Xiaobin Tian'] | 2022-11-05 | null | null | null | null | ['entity-alignment', 'entity-alignment'] | ['knowledge-base', 'natural-language-processing'] | [-1.04319818e-01 4.13075238e-01 -6.85774922e-01 -4.88764435e-01
-5.31978369e-01 -5.08831441e-01 3.60778123e-01 8.41797411e-01
-1.44235417e-01 7.32070804e-01 4.78385776e-01 -2.16746196e-01
-3.66443008e-01 -1.04562938e+00 -6.05467379e-01 -2.29875714e-01
-3.80996197e-01 7.64550328e-01 3.62192988e-01 -3.82240534e-01
7.14143068e-02 2.06426144e-01 -1.31110168e+00 2.42135480e-01
8.91367376e-01 4.47923243e-01 -3.64112467e-01 1.12301201e-01
-1.66056678e-01 9.96309042e-01 -3.86603236e-01 -1.13758826e+00
2.47219920e-01 -3.68892163e-01 -1.20665348e+00 -5.40738940e-01
5.93717396e-01 1.21094007e-02 -4.71899182e-01 1.10807860e+00
7.68346608e-01 -2.51318008e-01 3.51913035e-01 -1.85250568e+00
-8.03169549e-01 1.10858917e+00 -5.47470272e-01 1.99801162e-01
4.64754075e-01 -4.03872669e-01 1.58007348e+00 -6.67712390e-01
1.04440784e+00 9.13137913e-01 1.02571690e+00 1.06798932e-01
-1.06099570e+00 -4.63453919e-01 -9.05213580e-02 8.95165920e-01
-1.78249943e+00 -5.13924360e-01 5.58755159e-01 -1.45078212e-01
1.31948829e+00 4.72510576e-01 7.56027997e-01 7.85553932e-01
3.90468575e-02 6.54259980e-01 6.77208543e-01 -6.76785409e-01
2.20485236e-02 2.12451145e-01 1.97955862e-01 7.29988456e-01
9.39956665e-01 -2.31111929e-01 -5.90443492e-01 -4.00862277e-01
2.63109624e-01 -4.53609854e-01 -2.33698577e-01 -7.81498969e-01
-1.38911712e+00 7.92959929e-01 5.80702364e-01 6.16182446e-01
-2.72743881e-01 1.38132676e-01 5.42080402e-01 3.83816034e-01
3.42781663e-01 8.43249679e-01 -5.18480182e-01 -7.18906671e-02
-6.02773666e-01 4.78374571e-01 1.08483183e+00 1.24085307e+00
7.87232697e-01 -4.09682393e-01 2.06549279e-02 5.22722363e-01
1.05015352e-01 5.83621934e-02 2.22754925e-01 -9.54553962e-01
7.10264921e-01 9.86181974e-01 -1.04974896e-01 -1.52517378e+00
-3.97365034e-01 -2.58773357e-01 -6.28562689e-01 -3.39961648e-01
1.88790306e-01 2.62135744e-01 -3.63918841e-01 1.90287244e+00
5.39751410e-01 2.96426266e-01 3.26682448e-01 5.64721942e-01
7.54009247e-01 1.98496789e-01 -1.84577086e-03 -4.50213812e-02
1.42564237e+00 -9.88961041e-01 -8.97143602e-01 -1.98618528e-02
1.27624834e+00 -6.49403870e-01 6.35007143e-01 -3.52776915e-01
-1.05668867e+00 -1.89823471e-02 -1.15067112e+00 -1.84795186e-01
-6.98484123e-01 -3.39842737e-01 9.69358563e-01 6.46313131e-01
-9.80063319e-01 5.70779324e-01 -6.32533431e-01 -6.69201672e-01
1.51369259e-01 4.54383790e-01 -7.93247163e-01 2.78897975e-02
-1.59409893e+00 1.29070520e+00 7.29016066e-01 -8.32112953e-02
1.58824354e-01 -7.05156446e-01 -9.64199662e-01 1.31845504e-01
7.46392906e-01 -9.99246180e-01 8.49050164e-01 -4.80548620e-01
-7.05609500e-01 9.93711174e-01 -1.19349189e-01 -6.03035212e-01
1.77230179e-01 -1.07530899e-01 -7.62247205e-01 1.47792818e-02
3.96571577e-01 3.92273605e-01 -2.27254070e-02 -1.15756559e+00
-7.49038339e-01 -5.39815962e-01 3.33027601e-01 2.60471046e-01
-6.05156302e-01 7.16455281e-02 -4.54779297e-01 -5.35113513e-01
2.64827698e-01 -9.80610847e-01 -1.12886289e-02 -4.50059384e-01
-6.69258356e-01 -1.47378907e-01 3.54147643e-01 -5.58109760e-01
1.59157920e+00 -1.84441984e+00 3.95713300e-01 2.96617419e-01
4.05887991e-01 1.37880921e-01 -6.37074932e-04 9.25514460e-01
-3.24615479e-01 4.41282064e-01 -1.16619386e-01 -5.66253290e-02
1.99706554e-01 2.72154421e-01 -1.64503366e-01 2.75116026e-01
1.47948027e-01 1.29903233e+00 -1.08106589e+00 -1.00629437e+00
-2.64297843e-01 2.06312746e-01 -5.32009661e-01 -9.52837169e-02
6.29455596e-02 -7.96163976e-02 -2.96580195e-01 9.35407281e-01
4.80413765e-01 -3.83749694e-01 7.74618208e-01 -7.71749854e-01
9.10308138e-02 4.43242520e-01 -1.35195553e+00 1.68061101e+00
-7.79408670e-04 4.55772579e-01 -3.70161176e-01 -9.70587671e-01
4.61156279e-01 3.18203866e-01 8.28097820e-01 -6.99033439e-01
-2.19705224e-01 5.11991978e-01 1.82028681e-01 -5.27529478e-01
1.14231908e+00 1.20862730e-01 -1.61933526e-01 3.21929216e-01
1.03051245e-01 1.74750447e-01 3.83273840e-01 5.74172854e-01
1.35306931e+00 -7.21282437e-02 6.39743745e-01 -9.90430713e-02
3.17666799e-01 3.08822155e-01 6.95071697e-01 4.66252625e-01
-1.42203897e-01 1.77364096e-01 5.43098629e-01 -4.84508187e-01
-1.14669156e+00 -8.87703180e-01 -1.97483674e-01 6.47909880e-01
3.86496097e-01 -1.11541378e+00 -3.79195929e-01 -8.21179152e-01
2.80575275e-01 5.27099192e-01 -5.79290450e-01 -3.13209504e-01
-8.12965631e-01 -8.74925852e-01 8.15492034e-01 7.79895782e-01
2.85264969e-01 -5.63326776e-01 -2.62612522e-01 3.56369913e-01
-5.80353618e-01 -1.34275520e+00 -6.20683730e-02 1.06472775e-01
-6.48860216e-01 -1.20201206e+00 -2.64005959e-01 -8.42960835e-01
6.52659893e-01 6.09655008e-02 1.49907744e+00 2.44597182e-01
-1.73365906e-01 3.21705461e-01 -5.03922820e-01 -3.23794395e-01
-2.54695922e-01 5.78372717e-01 1.79988191e-01 -4.25508469e-01
6.24902368e-01 -6.20150447e-01 -2.91843325e-01 4.27195758e-01
-7.50867546e-01 -4.29625139e-02 4.47094738e-01 7.93106556e-01
4.79814559e-01 6.87349439e-02 5.82612872e-01 -1.17642140e+00
5.55150568e-01 -6.72115624e-01 -4.02680695e-01 6.85184836e-01
-1.14059210e+00 2.30005860e-01 2.34940916e-01 -2.05503311e-02
-5.59784532e-01 -2.56804049e-01 1.77368484e-02 -5.12740202e-02
4.01164830e-01 8.94267857e-01 -1.58286586e-01 -1.33953720e-01
4.46615309e-01 -1.68953508e-01 -4.22684401e-01 -1.95656821e-01
5.46352625e-01 4.99200314e-01 6.12266600e-01 -7.67548323e-01
8.38052809e-01 3.24432224e-01 9.70552713e-02 -1.62168339e-01
-6.75422549e-01 -4.47780013e-01 -6.49341285e-01 2.13383988e-01
5.70117593e-01 -1.02958965e+00 -5.38304329e-01 5.12384139e-02
-1.11522710e+00 8.35878551e-02 -5.59908152e-01 3.92908603e-01
-4.64337230e-01 5.30994534e-01 -6.22457922e-01 -3.12601537e-01
-2.53981084e-01 -8.64904344e-01 9.32871401e-01 -6.82139676e-03
-6.14269435e-01 -1.02336216e+00 4.87152159e-01 2.54133552e-01
3.61909479e-01 2.66327173e-01 1.22500372e+00 -9.89744782e-01
-7.35506177e-01 -2.80850351e-01 -1.83954582e-01 -2.66335577e-01
2.56676257e-01 4.12282050e-02 -4.75404501e-01 -2.51298875e-01
-4.87370461e-01 -1.38644397e-01 3.77941012e-01 -2.11515397e-01
5.57638645e-01 -3.78667414e-01 -7.83883095e-01 6.54808760e-01
1.62199950e+00 1.92051791e-02 7.83189595e-01 7.48094738e-01
9.06236172e-01 7.03611612e-01 8.13612759e-01 2.34488502e-01
1.00468373e+00 1.22651649e+00 3.55001122e-01 -1.04389586e-01
-1.28740296e-01 -3.63845140e-01 -4.05707024e-02 1.25081611e+00
-3.30504686e-01 -3.59804034e-01 -9.81524587e-01 9.23444510e-01
-2.11363888e+00 -1.21898460e+00 -3.78304034e-01 2.05230951e+00
1.01701939e+00 -1.91966653e-01 3.87688205e-02 5.65777756e-02
6.57086134e-01 4.21971753e-02 -1.44058853e-01 -1.66346997e-01
-4.59065080e-01 9.33617800e-02 5.94161093e-01 3.32204759e-01
-8.97814155e-01 9.54331338e-01 6.38605213e+00 4.41674203e-01
-6.96559429e-01 5.36777116e-02 1.52093112e-01 2.05617607e-01
-7.14517593e-01 3.96898985e-01 -9.23529744e-01 4.71438855e-01
8.30024540e-01 -7.35342026e-01 1.21505447e-01 6.11354232e-01
-6.86037064e-01 7.82805532e-02 -1.32580733e+00 8.60694230e-01
1.34960450e-02 -1.54909348e+00 -1.66998819e-01 1.40925020e-01
6.70829415e-01 -2.38970183e-02 -1.80080295e-01 3.07372004e-01
4.56642210e-01 -8.78582060e-01 4.86447215e-01 2.54107475e-01
5.34585178e-01 -6.09021902e-01 1.08018064e+00 -2.52915800e-01
-1.50230813e+00 1.50947452e-01 -3.94821852e-01 3.06660861e-01
2.63996542e-01 6.05268896e-01 -8.35649729e-01 1.35186517e+00
6.86288714e-01 7.05313385e-01 -7.30367422e-01 9.02458370e-01
-1.04883254e-01 2.18525454e-01 -2.73844630e-01 2.61944056e-01
-1.50146767e-01 -9.78157967e-02 4.67326403e-01 1.07631207e+00
4.16840523e-01 -2.28968393e-02 -1.09996505e-01 6.34155691e-01
-3.32725227e-01 3.59555632e-01 -9.41640973e-01 -2.41542071e-01
8.97495031e-01 1.13019824e+00 -5.44995666e-01 -3.91651481e-01
-6.62759364e-01 1.01083374e+00 6.97405696e-01 4.48842011e-02
-1.02244329e+00 -4.51936394e-01 6.87794685e-01 2.69354507e-02
2.25207612e-01 -8.60494748e-02 -5.99261187e-02 -1.45253003e+00
3.52572829e-01 -1.00053108e+00 7.16496527e-01 -4.89298105e-01
-1.24978793e+00 4.50423390e-01 1.92847341e-01 -9.84662175e-01
-3.47744226e-01 -2.30627552e-01 -1.22208610e-01 5.77051580e-01
-1.52373397e+00 -1.14265537e+00 -1.46005794e-01 4.81746525e-01
-2.18038052e-01 -1.93632185e-01 1.11818874e+00 8.76137018e-01
-4.50000674e-01 8.82771909e-01 -4.07555439e-02 2.02752993e-01
1.07806182e+00 -1.33404100e+00 6.59360588e-01 8.22730958e-01
5.60175955e-01 8.89662683e-01 8.18422973e-01 -8.49705875e-01
-1.60720074e+00 -8.82549703e-01 1.58417809e+00 -6.84569776e-01
8.72648478e-01 -1.06375270e-01 -9.27162111e-01 1.28266013e+00
3.51379603e-01 1.09008268e-01 9.19851422e-01 5.78124881e-01
-5.63794136e-01 -1.59628987e-01 -9.50063169e-01 7.42392242e-01
1.47191453e+00 -5.29886663e-01 -6.71522319e-01 4.01498109e-01
7.47084260e-01 -4.49195147e-01 -1.46329081e+00 8.15721631e-01
5.39931774e-01 -8.26520383e-01 1.08530831e+00 -9.33992684e-01
3.44095558e-01 -4.42479670e-01 -3.82646799e-01 -1.27937007e+00
-5.02689719e-01 -3.13514769e-01 -4.60372686e-01 1.52906680e+00
5.04358232e-01 -8.55931818e-01 8.07681143e-01 5.49963236e-01
-3.19106467e-02 -8.40339482e-01 -8.37285578e-01 -1.00137091e+00
-2.02801973e-01 7.92970434e-02 1.05160105e+00 1.70686805e+00
4.56507832e-01 5.45701742e-01 -2.92910904e-01 3.92689228e-01
4.34588194e-01 5.55796325e-01 9.12068248e-01 -1.16542041e+00
-1.61737740e-01 -1.84982076e-01 -1.02367747e+00 -3.56774032e-01
2.35033616e-01 -1.15127277e+00 -5.01949131e-01 -1.67474949e+00
5.04692376e-01 -7.23356009e-01 -2.92478800e-01 8.06942999e-01
-3.78537416e-01 1.33863717e-01 4.61967364e-02 2.87011862e-01
-6.39320254e-01 3.83043110e-01 5.77950001e-01 -2.71520048e-01
1.78574666e-01 -6.19267464e-01 -1.02869380e+00 3.41204762e-01
6.45034611e-01 -7.83635318e-01 -4.48873430e-01 -4.94745284e-01
7.81735241e-01 -1.48258239e-01 1.62917688e-01 -7.26677954e-01
4.88448262e-01 -4.27432358e-02 -1.23969600e-01 -3.62082720e-01
1.77564830e-01 -9.89501536e-01 7.43904352e-01 3.85995954e-01
-2.05510855e-01 5.85229933e-01 -7.43073784e-03 4.02860612e-01
-5.73102772e-01 -1.88028842e-01 2.14817479e-01 5.97829260e-02
-9.24125910e-01 2.47818545e-01 4.01210546e-01 2.99209327e-01
1.21973109e+00 -1.70697302e-01 -7.38073766e-01 -3.26475948e-01
-4.77232724e-01 2.56405503e-01 8.84302318e-01 3.56091380e-01
3.01885545e-01 -1.60739625e+00 -8.29450250e-01 -3.11012298e-01
5.23382843e-01 -2.93334037e-01 -1.30816862e-01 1.10830510e+00
-3.97644788e-01 3.45178843e-01 -3.32703233e-01 -1.95516646e-01
-1.45789123e+00 6.11539423e-01 1.42173961e-01 -5.43715179e-01
-5.80146849e-01 6.32157981e-01 -3.66774142e-01 -4.65656340e-01
-2.41989382e-02 -1.27729680e-02 -3.09361401e-03 1.77850023e-01
8.77095833e-02 4.21392828e-01 3.12474191e-01 -8.06409180e-01
-8.00077677e-01 4.25676525e-01 -2.45986149e-01 4.23548333e-02
1.40070903e+00 -2.91853789e-02 -4.84671026e-01 3.80673170e-01
1.18269408e+00 5.51828384e-01 -8.37228969e-02 -2.74069071e-01
5.03392100e-01 -6.55010939e-01 -4.68446583e-01 -5.55712104e-01
-1.05735886e+00 2.59150743e-01 2.35806093e-01 4.65666384e-01
9.33308125e-01 1.42428771e-01 8.27143788e-01 4.87943292e-01
6.02917552e-01 -9.64948714e-01 -3.75064492e-01 3.57076138e-01
3.72971058e-01 -1.17766726e+00 4.48242188e-01 -8.30456495e-01
-5.36108553e-01 8.01103830e-01 6.48360193e-01 3.19626987e-01
2.18473643e-01 3.06214631e-01 -1.35447457e-01 -5.80489814e-01
-6.89204454e-01 -3.47407490e-01 4.19486091e-02 4.86389726e-01
5.43624461e-01 8.46889743e-04 -5.82500756e-01 3.85646135e-01
-2.45666057e-01 -1.02162234e-01 4.10043716e-01 1.12759662e+00
1.12870194e-01 -1.63477719e+00 7.58754760e-02 3.79176855e-01
-4.70261246e-01 -3.48527044e-01 -6.32140994e-01 1.13714802e+00
1.22895412e-01 5.11132896e-01 -1.65635243e-01 -5.74675143e-01
4.14955199e-01 2.02983662e-01 7.11731017e-01 -4.60662097e-01
-5.46974063e-01 -7.24591315e-01 5.82368195e-01 -5.78401148e-01
-6.26077116e-01 -6.51889145e-01 -1.15449488e+00 -6.38208747e-01
-6.51966035e-01 3.24858695e-01 3.37109178e-01 5.85491538e-01
7.87950456e-01 3.60085338e-01 3.46394122e-01 -3.99721228e-02
-3.53738904e-01 -5.49176872e-01 -5.56926727e-01 6.44152105e-01
-2.40336537e-01 -6.67342782e-01 -6.80255964e-02 -9.92012396e-02] | [8.803617477416992, 7.9632568359375] |
c500ec56-5409-4b29-b19f-73efd0108b77 | a-richly-annotated-dataset-for-pedestrian | 1603.07054 | null | http://arxiv.org/abs/1603.07054v3 | http://arxiv.org/pdf/1603.07054v3.pdf | A Richly Annotated Dataset for Pedestrian Attribute Recognition | In this paper, we aim to improve the dataset foundation for pedestrian
attribute recognition in real surveillance scenarios. Recognition of human
attributes, such as gender, and clothes types, has great prospects in real
applications. However, the development of suitable benchmark datasets for
attribute recognition remains lagged behind. Existing human attribute datasets
are collected from various sources or an integration of pedestrian
re-identification datasets. Such heterogeneous collection poses a big challenge
on developing high quality fine-grained attribute recognition algorithms.
Furthermore, human attribute recognition are generally severely affected by
environmental or contextual factors, such as viewpoints, occlusions and body
parts, while existing attribute datasets barely care about them. To tackle
these problems, we build a Richly Annotated Pedestrian (RAP) dataset from real
multi-camera surveillance scenarios with long term collection, where data
samples are annotated with not only fine-grained human attributes but also
environmental and contextual factors. RAP has in total 41,585 pedestrian
samples, each of which is annotated with 72 attributes as well as viewpoints,
occlusions, body parts information. To our knowledge, the RAP dataset is the
largest pedestrian attribute dataset, which is expected to greatly promote the
study of large-scale attribute recognition systems. Furthermore, we empirically
analyze the effects of different environmental and contextual factors on
pedestrian attribute recognition. Experimental results demonstrate that
viewpoints, occlusions and body parts information could assist attribute
recognition a lot in real applications. | ['Haibin Ling', 'Xiaotang Chen', 'Zhang Zhang', 'Kaiqi Huang', 'Dangwei Li'] | 2016-03-23 | null | null | null | null | ['pedestrian-attribute-recognition'] | ['computer-vision'] | [-9.41889510e-02 -5.80303490e-01 -1.78241640e-01 -9.58488882e-01
-1.23109221e-01 -4.37462419e-01 7.84319520e-01 3.95548999e-01
-3.07277083e-01 8.38205218e-01 3.98774087e-01 2.73298800e-01
2.33910039e-01 -1.11146021e+00 -5.94259918e-01 -9.58465993e-01
4.61426303e-02 5.97824454e-01 1.34256288e-01 -1.43537074e-01
-1.64318249e-01 4.49350834e-01 -2.14253068e+00 1.91444114e-01
7.43654907e-01 1.30688918e+00 -1.87880665e-01 4.49469000e-01
-2.01591551e-02 2.47907475e-01 -7.63706744e-01 -1.00177729e+00
3.52148980e-01 -3.98441032e-03 -4.42555338e-01 5.05409300e-01
8.53520393e-01 -4.98410046e-01 -2.57788092e-01 1.01077294e+00
5.80969274e-01 -5.17633148e-02 7.30498970e-01 -1.69086206e+00
-7.72151053e-01 1.04312077e-01 -4.18189853e-01 -1.19742051e-01
5.88417053e-01 5.66288471e-01 6.96789086e-01 -7.54521787e-01
3.26891214e-01 1.42054951e+00 8.05292070e-01 4.43875551e-01
-9.24106896e-01 -7.16844916e-01 4.99551624e-01 5.61326683e-01
-1.57014894e+00 -4.15389746e-01 7.44408011e-01 -4.64214116e-01
2.81789243e-01 5.82623005e-01 1.16748869e+00 1.57939339e+00
-1.92217469e-01 9.17096376e-01 1.29587376e+00 -1.83017254e-02
-5.66082671e-02 4.04596001e-01 2.44512230e-01 5.03963768e-01
5.53702593e-01 1.27392992e-01 -2.59408146e-01 -1.97805941e-01
4.26479876e-01 2.70277292e-01 -6.35042880e-03 -3.75618488e-01
-1.44967902e+00 4.85011727e-01 2.29321837e-01 -3.84164125e-01
-3.61812681e-01 -4.85659748e-01 6.74699485e-01 1.73862487e-01
3.25565189e-01 -9.60153341e-02 -5.28945506e-01 8.07579234e-02
-1.99374259e-01 2.83021271e-01 6.47807181e-01 1.64415824e+00
8.35330606e-01 -2.30048206e-02 -3.58839214e-01 1.06215501e+00
2.57518828e-01 1.06556523e+00 1.87385362e-02 -7.37665176e-01
3.50674480e-01 9.11002457e-01 1.28117755e-01 -1.18713558e+00
-4.11074013e-01 -5.88480122e-02 -1.24671519e+00 -1.42998472e-01
6.76539063e-01 3.99180464e-02 -7.22741127e-01 1.52048683e+00
6.37342691e-01 -1.34687006e-01 -1.67183638e-01 1.10189712e+00
1.35118675e+00 3.20642352e-01 7.07825124e-01 -6.98710606e-02
1.97460938e+00 -6.17388248e-01 -6.43288314e-01 -2.40396723e-01
1.23815484e-01 -9.08443511e-01 9.05906975e-01 1.82155013e-01
-5.33248007e-01 -9.81648922e-01 -7.14973509e-01 2.34984726e-01
-7.36975431e-01 4.72730905e-01 8.52728844e-01 1.04393446e+00
-5.35289466e-01 -8.76406357e-02 -2.21948281e-01 -7.96799004e-01
6.36712193e-01 2.68643737e-01 -6.94514096e-01 -2.81605572e-01
-1.20764720e+00 6.93050921e-01 1.42952099e-01 1.16909720e-01
-8.87782753e-01 -4.70547259e-01 -9.65486825e-01 -4.62547183e-01
4.88440365e-01 -1.01407015e+00 7.40156412e-01 -6.50739610e-01
-9.92810369e-01 1.16088116e+00 -1.93597525e-01 -1.43936157e-01
5.55089593e-01 -1.36512488e-01 -8.95440996e-01 -3.07969600e-01
3.60118717e-01 6.11687124e-01 7.02513695e-01 -1.26820350e+00
-9.37719941e-01 -8.94265294e-01 8.37977082e-02 -1.51093146e-02
-4.31352675e-01 1.04376413e-01 -2.64611214e-01 -5.98259926e-01
-2.44799376e-01 -1.00204551e+00 -1.76024169e-01 2.43166685e-01
-4.11898047e-01 -3.37574333e-01 7.42921531e-01 -5.29312789e-01
7.67157614e-01 -1.98525393e+00 -2.14135498e-01 -9.41064581e-02
2.17109248e-01 2.69114673e-01 1.06796198e-01 9.39401910e-02
1.70555383e-01 -9.01014209e-02 3.13588865e-02 -2.13568822e-01
7.73049071e-02 4.65320647e-01 6.75911307e-02 6.19676292e-01
1.74727961e-01 6.22663200e-01 -7.82889426e-01 -9.09595191e-01
5.37762225e-01 5.62886536e-01 -1.08843513e-01 3.54220301e-01
3.23696919e-02 5.32286644e-01 -8.50851417e-01 1.46247113e+00
8.73235464e-01 4.92911696e-01 -3.45897645e-01 -7.13743269e-01
-9.38940421e-02 -4.85723436e-01 -1.28940511e+00 1.20561743e+00
-8.98547471e-03 5.01541257e-01 7.52024278e-02 -8.93325746e-01
1.37584996e+00 2.63605773e-01 7.10123897e-01 -4.93938446e-01
2.87085682e-01 -3.49050947e-02 -3.78865480e-01 -6.51295662e-01
5.38217783e-01 1.74115077e-01 -4.47826892e-01 -2.35887364e-01
-1.69567570e-01 2.47110426e-01 3.57306510e-01 -9.50949416e-02
4.55809027e-01 -6.53802454e-02 5.46217918e-01 -2.20299184e-01
8.81946325e-01 1.88305184e-01 8.79181862e-01 6.62735641e-01
-8.70611668e-01 5.58076978e-01 1.69593588e-01 -1.28297746e+00
-1.36896157e+00 -1.02816737e+00 -4.75095153e-01 1.19576752e+00
3.69369566e-01 -3.11195105e-01 -6.65242970e-01 -6.36160493e-01
8.51236656e-02 2.34661371e-01 -6.13763988e-01 -1.95859298e-01
-4.73174334e-01 -1.23789692e+00 7.40087450e-01 7.37154186e-01
1.13213301e+00 -9.59400177e-01 -2.26325244e-01 7.02341050e-02
-3.97590041e-01 -1.54503906e+00 -3.91243070e-01 -4.63413119e-01
-3.80081952e-01 -1.43277061e+00 -5.89419782e-01 -4.61300015e-01
6.09036028e-01 4.08681005e-01 1.48455071e+00 1.65104136e-01
-3.61077160e-01 5.12485147e-01 -4.65153396e-01 -6.80314541e-01
-1.58307597e-01 -1.54613927e-01 5.33943057e-01 5.37139297e-01
9.12313759e-01 -2.54989177e-01 -7.14678645e-01 1.11294198e+00
-3.10160786e-01 -3.36621813e-02 6.71354890e-01 7.89227724e-01
5.50168931e-01 -1.23714162e-02 4.19936329e-01 -6.83399081e-01
1.49800360e-01 -3.82169247e-01 -4.93515462e-01 3.25612843e-01
-2.87051439e-01 -4.64808494e-01 6.59142911e-01 -4.88496840e-01
-1.21633792e+00 2.34995916e-01 -1.74017474e-01 1.06962204e-01
-1.25982952e+00 -3.25398803e-01 -1.12790740e+00 1.08732365e-01
4.52487648e-01 2.69947678e-01 -1.29341885e-01 -3.96697849e-01
1.84527755e-01 7.85289943e-01 8.49393904e-01 -1.21151447e+00
6.10763133e-01 4.27963972e-01 2.11735159e-01 -9.22221243e-01
-8.66094530e-01 -6.25727952e-01 -1.11756313e+00 -5.22050500e-01
1.06773555e+00 -1.12101901e+00 -9.09714699e-01 9.54638541e-01
-9.93806660e-01 3.60863537e-01 -1.98374853e-01 4.06955332e-01
-3.86350781e-01 3.66265625e-01 -3.74434382e-01 -7.96359181e-01
-1.52815908e-01 -1.23472178e+00 1.20914328e+00 4.11458611e-01
4.40711342e-02 -5.09385347e-01 -5.65990925e-01 7.48023987e-01
2.58991182e-01 5.54470956e-01 4.99886334e-01 -6.10413373e-01
-5.80106556e-01 -3.92438889e-01 -3.97872806e-01 1.05740510e-01
2.02584624e-01 1.32951304e-01 -9.80104804e-01 -1.48321226e-01
-4.95167106e-01 -2.15568483e-01 6.05215251e-01 2.35320121e-01
1.12387490e+00 -2.69133270e-01 -4.26153719e-01 7.29226947e-01
8.15495253e-01 2.08211467e-01 6.10582352e-01 5.29767811e-01
1.06220007e+00 7.87673831e-01 1.04730821e+00 8.34922612e-01
8.02595317e-01 9.23282087e-01 4.87646431e-01 -6.99148029e-02
-2.47739553e-01 -8.02308619e-02 1.49733618e-01 3.57832998e-01
-6.92066967e-01 -7.29557648e-02 -8.45956147e-01 3.64601910e-01
-1.65933990e+00 -1.08813155e+00 -6.82021379e-01 2.23926711e+00
5.96747935e-01 -1.27878577e-01 5.68427086e-01 2.28728324e-01
9.36525822e-01 2.13057250e-02 -4.66413230e-01 1.12498082e-01
-6.06791556e-01 -6.44866765e-01 3.99565786e-01 -1.26049921e-01
-1.75118256e+00 7.21026182e-01 5.47796965e+00 6.81174576e-01
-3.93619835e-01 -2.54938990e-01 8.72922421e-01 3.74532193e-01
3.27910841e-01 -4.09530252e-01 -1.28042483e+00 6.65726721e-01
4.94002759e-01 3.03591967e-01 -6.38670474e-02 1.10857260e+00
-1.60222650e-01 6.42863438e-02 -1.00038886e+00 1.03530037e+00
1.28512934e-01 -6.16604626e-01 1.44144043e-01 3.45088355e-02
4.52513456e-01 -6.75153255e-01 -1.73010454e-01 3.93843174e-01
3.39684308e-01 -7.56346762e-01 5.39189279e-01 5.65691948e-01
8.36471260e-01 -9.11140203e-01 1.08599186e+00 2.07665607e-01
-1.68162298e+00 -8.10328051e-02 -7.58139789e-01 -6.82962313e-02
1.18051805e-01 5.01948416e-01 -5.03546476e-01 5.62543333e-01
1.06214237e+00 8.32541406e-01 -9.81912255e-01 1.09240663e+00
-3.60486321e-02 3.68559033e-01 -1.86304778e-01 5.26843742e-02
-2.25430936e-01 -2.90771693e-01 4.64335591e-01 1.24641061e+00
1.84872955e-01 4.86073524e-01 7.14124978e-01 3.48593205e-01
2.35171184e-01 1.45498097e-01 -6.92966759e-01 3.81774306e-01
6.09180093e-01 1.42444468e+00 -3.81579310e-01 -5.23281753e-01
-7.55572319e-01 5.37400186e-01 -1.28847763e-01 2.65426368e-01
-8.76870096e-01 2.94985861e-01 1.29172671e+00 2.89625883e-01
-2.81449854e-02 -1.77281141e-01 -2.19164521e-01 -1.18370056e+00
1.12510391e-01 -1.04539847e+00 7.51275301e-01 -4.89296734e-01
-1.81940281e+00 6.01645470e-01 2.03846365e-01 -1.47517085e+00
9.62542370e-02 -7.47907043e-01 -2.87665516e-01 6.72689736e-01
-1.11176884e+00 -1.95489681e+00 -1.15310478e+00 8.79692852e-01
6.56211913e-01 -6.56545818e-01 8.92992198e-01 6.44973516e-01
-8.08195770e-01 7.08207309e-01 -4.58121061e-01 5.95595181e-01
9.43233192e-01 -1.00846362e+00 4.24965501e-01 5.00454962e-01
-2.25005552e-01 3.68097752e-01 8.41967165e-01 -7.24828362e-01
-1.46869218e+00 -1.29750907e+00 4.93878394e-01 -8.67721498e-01
3.43779445e-01 -3.55044186e-01 -7.93625534e-01 6.11764610e-01
-8.15398768e-02 4.41016942e-01 8.51928413e-01 2.28145197e-01
-3.83760065e-01 -5.39237916e-01 -1.26982796e+00 2.05127373e-01
1.29745269e+00 -1.77478403e-01 -3.35006565e-01 2.52743274e-01
3.47034991e-01 -2.57593960e-01 -1.25728524e+00 7.03291237e-01
6.99780285e-01 -1.04099739e+00 1.67781675e+00 -7.27412939e-01
1.75172642e-01 -4.57077712e-01 -4.95737791e-01 -1.21183908e+00
-3.79005790e-01 1.44341514e-01 1.14739746e-01 1.83363867e+00
-1.66683719e-01 -5.18687904e-01 7.40866542e-01 1.04594946e+00
-2.80908436e-01 -5.53701997e-01 -6.90261066e-01 -7.74160862e-01
-2.99040228e-01 -2.26940498e-01 1.28064418e+00 7.87306964e-01
-7.23740280e-01 1.69396341e-01 -7.77800977e-01 3.07673633e-01
1.11417234e+00 2.05911040e-01 1.40010452e+00 -1.56423330e+00
3.35123241e-01 -3.81881863e-01 -1.03662765e+00 -7.20383584e-01
8.04445744e-02 -1.87003702e-01 -2.49361470e-01 -1.22992969e+00
5.04718602e-01 -7.67215967e-01 1.03680730e-01 2.92662382e-01
-4.89975244e-01 2.67432392e-01 4.19246666e-02 1.77787140e-01
-6.68066144e-01 7.43885338e-01 1.26137233e+00 -6.46044552e-01
4.22064155e-01 4.07449841e-01 -4.29544389e-01 8.22606981e-01
8.81861269e-01 -9.84979868e-02 -8.20926726e-02 -2.58273274e-01
-3.78248155e-01 -6.97731003e-02 7.96324432e-01 -1.14145815e+00
2.80151278e-01 -5.28829694e-01 1.08080578e+00 -7.69207239e-01
6.13261640e-01 -1.16053867e+00 2.94478416e-01 5.14890663e-02
1.63770348e-01 2.96600401e-01 1.07517559e-02 5.52388191e-01
-2.79900730e-01 2.25253433e-01 6.73238933e-01 -3.78649086e-01
-1.26079524e+00 9.13469374e-01 -1.57957256e-01 1.33891076e-01
1.26555681e+00 -5.61673164e-01 -4.38676625e-01 -1.56858578e-01
-5.04341662e-01 2.15413958e-01 6.77939415e-01 5.17871857e-01
5.62360764e-01 -1.74929154e+00 -1.01528227e+00 4.65779871e-01
5.76440930e-01 -4.19983901e-02 3.28654319e-01 6.32884681e-01
-1.61373261e-02 1.62418559e-01 -8.40690494e-01 -8.25232863e-01
-1.61742377e+00 9.28514540e-01 5.41967377e-02 4.04458702e-01
-4.50542480e-01 5.02937615e-01 2.18511164e-01 -5.16750336e-01
1.35564238e-01 2.70594358e-01 -5.67192316e-01 3.49817991e-01
5.66128194e-01 5.75050354e-01 -2.68158346e-01 -1.30550230e+00
-4.86034125e-01 8.68635416e-01 1.09973967e-01 6.47025943e-01
1.00989211e+00 -3.65767002e-01 -2.30019391e-02 6.50404859e-03
7.93397069e-01 -1.75274134e-01 -1.25907159e+00 -3.28613430e-01
-2.44930819e-01 -9.57672000e-01 -6.78815305e-01 -5.17994225e-01
-1.25720155e+00 7.59303093e-01 6.50704920e-01 1.41539589e-01
1.08992183e+00 1.66522601e-04 8.47400665e-01 3.63339067e-01
8.39425743e-01 -1.16135359e+00 -1.97116330e-01 2.56257802e-01
6.13701344e-01 -1.84445870e+00 3.36431026e-01 -7.81033218e-01
-8.72005045e-01 9.12252247e-01 1.14942884e+00 4.01751369e-01
3.66825134e-01 2.08857402e-01 2.00457692e-01 -1.64543707e-02
-2.73751825e-01 -3.88018757e-01 4.49786812e-01 1.20594871e+00
3.55094403e-01 6.45781755e-01 2.26979945e-02 8.28366935e-01
-3.77348930e-01 -4.86423165e-01 2.45736808e-01 4.38733757e-01
-4.02492672e-01 -1.24902296e+00 -8.43521357e-01 4.61655319e-01
-3.91301811e-02 2.72306919e-01 -4.26778674e-01 8.08305562e-01
6.32660806e-01 1.00867403e+00 -7.52664134e-02 -5.42848051e-01
7.20317125e-01 -2.04907134e-01 2.78077662e-01 8.00433978e-02
-4.69833821e-01 -4.75391835e-01 5.54176569e-01 -4.61093307e-01
-5.26726544e-01 -9.18975949e-01 -5.55261493e-01 -7.60920584e-01
1.20722808e-01 -2.04795539e-01 3.79483521e-01 7.42995918e-01
2.31238641e-02 2.18349323e-01 4.91273493e-01 -8.12533438e-01
-2.52744526e-01 -8.18158805e-01 -3.63638192e-01 1.04409671e+00
8.75239745e-02 -1.22456932e+00 5.37401177e-02 3.33616734e-01] | [14.49011516571045, 0.9859794974327087] |
3a1bb1b5-1f25-4dab-ae83-3ec977bb3696 | do-you-follow-me-a-survey-of-recent-1 | null | null | https://aclanthology.org/2022.sigdial-1.33 | https://aclanthology.org/2022.sigdial-1.33.pdf | “Do you follow me?”: A Survey of Recent Approaches in Dialogue State Tracking | While communicating with a user, a task-oriented dialogue system has to track the user’s needs at each turn according to the conversation history. This process called dialogue state tracking (DST) is crucial because it directly informs the downstream dialogue policy. DST has received a lot of interest in recent years with the text-to-text paradigm emerging as the favored approach. In this review paper, we first present the task and its associated datasets. Then, considering a large number of recent publications, we identify highlights and advances of research in 2021-2022. Although neural approaches have enabled significant progress, we argue that some critical aspects of dialogue systems such as generalizability are still underexplored. To motivate future studies, we propose several research avenues. | ['Benoit Favre', 'Lina M. Rojas Barahona', 'Léo Jacqmin'] | null | null | null | null | sigdial-acl-2022-9 | ['dialogue-state-tracking'] | ['natural-language-processing'] | [ 2.73541182e-01 5.61033607e-01 -3.51130456e-01 -6.81601524e-01
-3.08282793e-01 -7.25399911e-01 9.54662919e-01 6.17773123e-02
-4.50604022e-01 8.72779965e-01 7.95172989e-01 -3.82525265e-01
4.20918204e-02 -3.33081901e-01 2.46950909e-01 -2.04890236e-01
1.30467042e-01 5.00313878e-01 9.80608072e-03 -7.92369425e-01
5.99221647e-01 2.43878558e-01 -1.23796713e+00 4.79157388e-01
7.91348577e-01 7.35817552e-01 1.93688929e-01 7.38647640e-01
-3.35142374e-01 8.97846818e-01 -7.89237976e-01 -3.74810815e-01
-1.01394840e-01 -9.08886611e-01 -1.46405649e+00 -6.24091066e-02
-2.44700871e-02 -6.51163280e-01 -4.15608883e-01 7.25753546e-01
7.43583500e-01 4.75970596e-01 4.29531991e-01 -1.06529009e+00
-1.31361172e-01 6.32452130e-01 1.44400984e-01 2.14304820e-01
8.08624864e-01 3.07461917e-01 1.11591029e+00 -5.34195483e-01
6.07707858e-01 1.38324010e+00 3.82354647e-01 1.14021122e+00
-1.00318575e+00 -2.33431160e-01 4.75572824e-01 1.82987288e-01
-5.54173708e-01 -9.21081960e-01 6.80749655e-01 -2.82065958e-01
1.13631749e+00 3.93984050e-01 8.13161850e-01 1.33847201e+00
-6.90483227e-02 1.22396982e+00 1.30099857e+00 -3.87341768e-01
1.28543213e-01 3.07414055e-01 3.77594531e-01 3.96276921e-01
-5.33672512e-01 -1.87633246e-01 -8.42635512e-01 -2.74302959e-01
5.34206688e-01 -7.61132956e-01 -4.25211012e-01 5.75895198e-02
-1.02161562e+00 9.42893088e-01 -5.65809719e-02 5.21430552e-01
-3.26419234e-01 -4.75782067e-01 7.62733340e-01 5.27340591e-01
4.71915126e-01 7.56478250e-01 -4.62127060e-01 -9.41752315e-01
-6.34655416e-01 4.05548811e-01 1.43745208e+00 6.42232060e-01
2.10943565e-01 -7.00100288e-02 -6.03142619e-01 1.25508523e+00
2.68514782e-01 -1.04004309e-01 3.98059845e-01 -1.06108725e+00
4.89205569e-01 6.09214127e-01 2.31538385e-01 -8.62472951e-01
-5.96913218e-01 1.77408367e-01 -7.10687220e-01 -3.18829894e-01
8.01001072e-01 -6.07373655e-01 -1.30940482e-01 1.81976509e+00
2.45015174e-01 -4.49274063e-01 2.27966011e-01 8.49608481e-01
7.03784347e-01 6.58019662e-01 4.88119014e-02 -4.14957941e-01
1.37348270e+00 -8.62366021e-01 -1.12905395e+00 -4.54643846e-01
4.23106611e-01 -5.55294514e-01 9.76620615e-01 4.14375693e-01
-1.13660562e+00 -2.64649928e-01 -7.34935105e-01 -1.50724724e-01
-1.94501311e-01 4.58786730e-03 6.82364821e-01 7.40256190e-01
-1.15642226e+00 4.17752892e-01 -6.12125874e-01 -8.82987201e-01
-1.48918808e-01 2.31196031e-01 -6.12817071e-02 4.45101678e-01
-1.76292610e+00 1.46273136e+00 2.07487345e-01 4.32971656e-01
-5.75702846e-01 -1.37475654e-01 -6.81078613e-01 -5.34737706e-02
3.72065961e-01 -3.84903520e-01 1.91619933e+00 -5.55954635e-01
-2.45932245e+00 7.44506180e-01 -3.49084914e-01 -4.27425146e-01
5.93479156e-01 -2.28342101e-01 -1.96979448e-01 1.49447277e-01
-3.95424664e-01 4.36500490e-01 4.19557244e-01 -8.93606722e-01
-1.03312325e+00 -3.03690910e-01 4.26464558e-01 7.41412163e-01
-3.45025003e-01 3.75562370e-01 -4.43727821e-01 -1.48996040e-01
2.79236566e-02 -9.90745366e-01 -1.38516262e-01 -2.72709042e-01
-5.45352280e-01 -7.13040888e-01 6.10692203e-01 -7.11978257e-01
1.52931094e+00 -1.67798865e+00 1.53767273e-01 -2.84122020e-01
2.14648440e-01 4.05130535e-01 1.17275968e-01 1.06669009e+00
2.17433125e-01 -4.66203094e-02 -1.79465249e-01 -3.56198043e-01
1.77093416e-01 -1.75012991e-01 -3.00471872e-01 2.87209183e-01
6.41210824e-02 8.88097227e-01 -9.07247603e-01 -2.44612619e-01
3.86027902e-01 9.81447250e-02 -3.99459124e-01 6.68077230e-01
-3.36002290e-01 7.70035684e-01 -5.79185784e-01 1.47324339e-01
1.68387547e-01 -1.73867688e-01 5.70438981e-01 -1.84862819e-02
-4.26560283e-01 9.84845281e-01 -6.25899673e-01 1.57146645e+00
-2.89942145e-01 9.58005548e-01 5.67125142e-01 -1.06393838e+00
9.15275574e-01 5.80378592e-01 4.29161400e-01 -5.13943613e-01
1.22897014e-01 -1.22848220e-01 3.71610224e-01 -7.61108339e-01
8.14501941e-01 -9.73615050e-02 -2.61943370e-01 7.47559249e-01
-1.55902818e-01 -1.08718641e-01 9.55982357e-02 3.57009977e-01
9.50962067e-01 2.90928781e-02 6.03284121e-01 -1.99859992e-01
6.41689181e-01 1.57445937e-01 3.31894428e-01 6.31791532e-01
-7.20421791e-01 1.37367696e-01 9.43179429e-01 -2.20050395e-01
-7.35189021e-01 -4.20087725e-01 -7.26787820e-02 1.29529631e+00
-2.24496424e-01 -4.26204473e-01 -9.60781276e-01 -6.18206441e-01
-1.97376966e-01 8.70364845e-01 -5.24377763e-01 -2.51198336e-02
-6.82628572e-01 -4.83128577e-01 6.64359093e-01 1.58226743e-01
8.03199291e-01 -1.31923425e+00 -6.06608212e-01 3.90141517e-01
-6.14851952e-01 -1.03645396e+00 -3.15061390e-01 -5.34277447e-02
-7.00072706e-01 -8.99988413e-01 -7.85407901e-01 -6.28699720e-01
2.15872154e-01 2.13681877e-01 9.07075346e-01 -4.56590243e-02
3.69166583e-01 4.46812809e-01 -3.43966901e-01 -9.16230679e-02
-8.27012002e-01 3.95999283e-01 1.49003267e-01 -9.40489918e-02
3.80261391e-01 -2.18750760e-01 -5.76656222e-01 4.72619385e-01
-2.28473365e-01 3.74964833e-01 1.85766697e-01 8.18251371e-01
-4.84670401e-01 -4.62841630e-01 9.60651696e-01 -1.09771597e+00
1.52467251e+00 -4.62228566e-01 -2.74779141e-01 2.12504700e-01
-6.83203459e-01 -2.12518245e-01 5.35536110e-01 -3.43221836e-02
-1.53294718e+00 -1.93475962e-01 -3.55479479e-01 4.94880348e-01
-3.49455476e-01 6.62491262e-01 -6.86774775e-02 2.54334420e-01
4.74254817e-01 2.53687918e-01 4.51126695e-01 -4.32821125e-01
3.19617242e-01 1.19569898e+00 2.50529051e-01 -6.55394256e-01
1.28615722e-01 1.68999713e-02 -6.83261693e-01 -1.17282391e+00
-6.84990644e-01 -4.17282194e-01 -6.96499705e-01 -6.99483991e-01
7.45591938e-01 -5.77553093e-01 -1.18773460e+00 5.67697465e-01
-1.31937099e+00 -6.77262485e-01 8.19427446e-02 2.71021515e-01
-5.79958618e-01 4.98585761e-01 -8.70621324e-01 -1.36932182e+00
-3.66253227e-01 -1.21369851e+00 4.58681494e-01 4.34280604e-01
-9.54672217e-01 -1.07105100e+00 -5.38015254e-02 6.43875301e-01
6.30077779e-01 -2.78558314e-01 7.64180005e-01 -1.20567083e+00
-7.27085918e-02 -1.27840281e-01 6.68617710e-02 1.61386520e-01
3.02962631e-01 -1.68931797e-01 -1.10592413e+00 -1.91602185e-01
3.34370494e-01 -5.36066115e-01 3.22224200e-01 2.41317257e-01
6.82056010e-01 -3.44311804e-01 -2.41552740e-01 -1.54574081e-01
4.30931389e-01 4.61376965e-01 2.99133539e-01 1.79273680e-01
1.77173242e-01 1.45258915e+00 7.94776320e-01 5.49939394e-01
8.30925763e-01 7.14273512e-01 2.53751110e-02 2.78803498e-01
7.77873993e-02 -3.47901165e-01 5.61696351e-01 1.00832450e+00
2.31379136e-01 -4.31631684e-01 -9.12378371e-01 3.25084567e-01
-1.90447783e+00 -9.36769724e-01 -3.45270149e-02 1.99273527e+00
1.12513280e+00 1.66605383e-01 4.14310157e-01 -4.19845283e-02
9.14808393e-01 4.72569793e-01 -7.11463273e-01 -6.29924595e-01
1.67885982e-02 -3.74010831e-01 -1.59340858e-01 7.58173048e-01
-7.30431616e-01 1.10947335e+00 7.42804909e+00 3.33044410e-01
-1.09081078e+00 -2.68777460e-01 7.25937188e-01 1.32738486e-01
1.28702568e-02 -3.19028161e-02 -7.68831193e-01 4.16950554e-01
1.05837226e+00 -3.77486974e-01 6.63691282e-01 5.47727883e-01
7.22198606e-01 -6.40569568e-01 -1.17924416e+00 5.20580113e-01
-3.78849916e-02 -9.28813279e-01 -4.17415529e-01 -8.07170384e-03
2.48463437e-01 -1.70679048e-01 -1.79221079e-01 4.86075759e-01
5.08481145e-01 -7.24784851e-01 2.29887694e-01 3.86672229e-01
5.83136916e-01 -3.97803277e-01 4.50888574e-01 7.06552148e-01
-8.31573665e-01 -8.92985146e-03 -3.74533032e-04 -4.66774970e-01
4.17211235e-01 1.68996841e-01 -1.19293571e+00 2.80978262e-01
2.14882404e-01 5.79964340e-01 -8.41707066e-02 5.90359747e-01
-3.68611008e-01 8.57013047e-01 -1.33225203e-01 -4.89016593e-01
2.74868101e-01 -4.16442722e-01 5.64817905e-01 1.13334608e+00
-2.17768982e-01 4.76745933e-01 1.55845448e-01 7.11653113e-01
1.63513631e-01 1.61054716e-01 -5.82109332e-01 -3.73733848e-01
7.13494480e-01 1.25741792e+00 -5.49354851e-01 -2.63447702e-01
-4.22461689e-01 1.01080632e+00 3.04305702e-01 2.80482411e-01
-2.83644348e-01 -1.24619260e-01 8.91864181e-01 -4.17964339e-01
-4.58118737e-01 -3.43550593e-01 -2.45124668e-01 -9.73223090e-01
-2.45628059e-01 -1.17040336e+00 3.29248130e-01 -3.62034649e-01
-1.08484697e+00 4.73829031e-01 -4.19835448e-02 -8.09143722e-01
-6.08628809e-01 -4.49160904e-01 -5.47768950e-01 1.02057111e+00
-1.19615638e+00 -5.76020062e-01 -1.81649309e-02 1.63943410e-01
1.05903983e+00 -1.50899127e-01 1.05392504e+00 -1.77475631e-01
-9.18095052e-01 4.86805052e-01 -4.29341719e-02 2.27517590e-01
8.77711952e-01 -1.16503263e+00 6.58970177e-01 1.92043483e-01
-4.02252585e-01 8.74191105e-01 8.76616836e-01 -6.72031581e-01
-1.52526402e+00 -2.98705071e-01 1.12196481e+00 -4.11741793e-01
6.19465053e-01 -3.19927007e-01 -8.50533724e-01 5.27405500e-01
7.15084791e-01 -8.20129514e-01 7.51854002e-01 3.48830640e-01
2.87904710e-01 3.59626919e-01 -1.06404567e+00 8.85151684e-01
8.28497291e-01 -7.84465909e-01 -6.63144648e-01 1.95645660e-01
3.95564616e-01 -6.86356723e-01 -8.89394343e-01 -2.05097087e-02
6.35228693e-01 -1.14863491e+00 4.45032656e-01 -6.89236879e-01
2.14611739e-01 3.32997769e-01 1.31033599e-01 -1.48593032e+00
-9.45463628e-02 -1.27556980e+00 -7.89992884e-03 1.22252595e+00
3.53100985e-01 -8.16042006e-01 8.59259188e-01 1.38530302e+00
-1.37495279e-01 -8.04574251e-01 -7.79797852e-01 -2.41517857e-01
1.76652744e-01 -2.55349666e-01 2.35180125e-01 1.04041195e+00
1.02245367e+00 8.71469855e-01 -6.72392249e-01 -4.47442532e-01
-1.50862813e-01 -7.89616108e-02 8.05525899e-01 -1.25861394e+00
9.59174931e-02 -6.90798819e-01 4.44096088e-01 -1.73054743e+00
2.50052184e-01 -5.99484921e-01 2.47002229e-01 -1.60682487e+00
-9.54646170e-02 -2.06476841e-02 1.58056691e-01 2.73687273e-01
-2.79232979e-01 -5.00856876e-01 2.35692874e-01 5.50656691e-02
-4.62671429e-01 7.92215288e-01 1.26809621e+00 6.46556541e-02
-3.02970052e-01 4.99289453e-01 -8.59138489e-01 5.90048254e-01
1.06962466e+00 5.91681600e-02 -4.22288835e-01 -4.77593318e-02
-8.22917968e-02 6.09037638e-01 -3.10204983e-01 -3.90307665e-01
4.70110714e-01 -3.33914608e-01 -1.03655510e-01 -4.71115589e-01
7.37339795e-01 -4.31594610e-01 -5.75274169e-01 3.96416843e-01
-1.00840604e+00 1.12248898e-01 1.46631345e-01 3.67317587e-01
-8.40596631e-02 -3.35276663e-01 6.07051194e-01 -1.90129951e-01
-6.98807657e-01 -1.39834300e-01 -1.13945425e+00 1.21675409e-01
7.73591876e-01 -7.50895888e-02 -2.45672345e-01 -9.64290202e-01
-5.99932611e-01 6.09141648e-01 9.45632383e-02 6.08560920e-01
3.94644678e-01 -8.27783048e-01 -6.24187350e-01 -1.72659114e-01
-7.70934001e-02 -3.69923770e-01 2.77050555e-01 7.83308327e-01
-2.09148228e-02 9.37565923e-01 -1.40846699e-01 -2.39728436e-01
-1.25536454e+00 -1.12353548e-01 4.30324465e-01 -1.76874593e-01
-6.39255404e-01 7.38771677e-01 -1.73268050e-01 -6.84064150e-01
4.94636238e-01 6.51078075e-02 -7.90459096e-01 5.03618836e-01
5.20524204e-01 5.00045538e-01 -1.30464584e-01 -5.41643858e-01
-8.30848515e-02 -2.25698009e-01 -2.26325005e-01 -5.43502390e-01
1.00814402e+00 -4.80238408e-01 -1.04937896e-01 8.38791668e-01
8.02968562e-01 -2.66092360e-01 -1.22239161e+00 -3.73342901e-01
3.38380188e-01 -2.98322648e-01 -3.11056525e-02 -1.06797969e+00
-6.79681063e-01 8.75073671e-01 8.44161659e-02 7.42819011e-01
7.32061923e-01 -3.05764467e-01 6.88983917e-01 5.40205657e-01
2.62735665e-01 -1.50143278e+00 3.70746925e-02 1.12015462e+00
1.05652487e+00 -1.29965341e+00 -1.83339760e-01 -2.52449036e-01
-1.07731450e+00 1.14782143e+00 8.99185777e-01 4.95141536e-01
3.56120795e-01 6.75746277e-02 2.74608135e-01 -1.62494510e-01
-1.12410760e+00 -1.69297874e-01 1.35427564e-02 4.92582858e-01
1.03437078e+00 -2.73740023e-01 -5.93883097e-01 4.81310904e-01
-4.01481181e-01 -1.69714227e-01 4.63143289e-01 8.81949425e-01
-4.40166861e-01 -1.37996769e+00 -7.95961767e-02 6.78581655e-01
-2.11385235e-01 2.64984071e-02 -9.62478220e-01 4.96920943e-01
-8.93942118e-01 1.50083029e+00 -2.18191490e-01 -4.63599861e-01
5.03820956e-01 4.64693397e-01 1.49419591e-01 -7.72495091e-01
-1.05395389e+00 -1.56453475e-01 8.33742797e-01 -4.67642963e-01
-2.36614063e-01 -9.39687192e-01 -1.06825233e+00 -3.78220737e-01
-2.94823557e-01 3.95804405e-01 6.19804263e-01 1.08435631e+00
3.08650374e-01 3.32323372e-01 6.95750237e-01 -6.72066212e-01
-8.31596971e-01 -1.38922358e+00 -2.36783609e-01 4.71122563e-02
2.32348129e-01 -4.10240114e-01 -1.19518980e-01 -3.44929039e-01] | [12.850491523742676, 7.941054821014404] |
b53ae2b9-9ce9-4be0-af75-0a1275a25d8a | program-transfer-for-answering-complex | null | null | https://openreview.net/forum?id=rn8YIulHv03 | https://openreview.net/pdf?id=rn8YIulHv03 | Program Transfer for Answering Complex Questions over Knowledge Bases | Program induction for answering complex questions over knowledge bases (KBs) aims to decompose a question into a multi-step program, whose execution against the KB produces the final answer. Learning to induce programs relies on a large number of parallel question-program pairs for the given KB. However, for most KBs, the gold program annotations are usually lacking, making learning difficult. In this paper, we propose the approach of program transfer, which aims to leverage the valuable program annotations on the rich-resourced KBs as external supervision signals to aid program induction for the low-resourced KBs that lack program annotations. For program transfer, we design a novel two-stage parsing framework with an efficient ontology-guided pruning strategy.
First, a sketch parser translates the question into a high-level program sketch, which is the composition of functions. Second, given the question and sketch, an argument parser searches the detailed arguments from the KB for functions. During the searching, we incorporate the KB ontology to prune the search space. The experiments on ComplexWebQuestions and WebQuestionSP show that our method outperforms SOTA methods significantly, demonstrating the effectiveness of program transfer and our framework.
| ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['program-induction'] | ['computer-code'] | [ 1.19416714e-01 3.34912151e-01 -5.58693588e-01 -4.47253793e-01
-1.02671123e+00 -7.51413286e-01 -6.55197203e-02 4.38235819e-01
-1.56865045e-01 4.12790984e-01 -9.96628478e-02 -7.48307467e-01
1.33016869e-01 -1.28899956e+00 -1.22485960e+00 -1.13788128e-01
2.59285629e-01 2.88349032e-01 9.00536418e-01 -2.42187595e-03
2.07960635e-01 -2.22532138e-01 -1.47970283e+00 6.79295599e-01
1.23921573e+00 7.73022771e-01 4.71132547e-01 7.90193558e-01
-6.63210392e-01 1.25331330e+00 -4.74557757e-01 -7.90709436e-01
-3.57153773e-01 -2.92061716e-01 -1.28480613e+00 -6.01631820e-01
2.84740478e-01 -4.14032310e-01 -8.97241086e-02 1.09711635e+00
7.05444291e-02 -8.11206326e-02 1.27120242e-01 -1.16461647e+00
-5.15721023e-01 1.00895655e+00 -2.42739782e-01 1.89057048e-02
6.97254002e-01 1.60999700e-01 1.83482277e+00 -6.68513715e-01
6.94551229e-01 1.13088548e+00 6.25836849e-01 5.31597912e-01
-1.25840545e+00 -4.33318108e-01 -1.18487753e-01 2.79599875e-01
-9.66823518e-01 -1.84523255e-01 8.34580362e-01 -4.59006071e-01
1.19530511e+00 2.78077513e-01 5.46411872e-01 4.07251835e-01
-3.14933330e-01 1.02351570e+00 5.56405365e-01 -7.74363697e-01
5.85519709e-03 6.58182323e-01 9.31527853e-01 1.43080449e+00
1.08606935e-01 -2.22284332e-01 -4.84598577e-01 -7.57793963e-01
2.18647644e-01 -2.29821309e-01 -3.42359602e-01 -4.15522933e-01
-7.70793796e-01 1.04210770e+00 1.26670212e-01 -3.24464329e-02
8.09450746e-02 9.99813005e-02 6.73518181e-01 5.10367334e-01
-2.35625412e-02 4.47362095e-01 -8.32252800e-01 -1.21692657e-01
-4.98059809e-01 3.81486148e-01 1.47617018e+00 1.14500713e+00
1.39529991e+00 -6.83122337e-01 -7.82790408e-02 7.41848230e-01
3.68239790e-01 4.20682847e-01 1.46358550e-01 -7.73257852e-01
1.00180995e+00 1.36144173e+00 -2.06545681e-01 -9.17381644e-01
6.35112971e-02 2.91114330e-01 -3.35042216e-02 -3.78750652e-01
3.78347665e-01 3.48752551e-02 -2.33947843e-01 1.58701575e+00
7.40642250e-01 -2.75794029e-01 1.75128058e-01 5.50006628e-01
1.20474744e+00 7.72790551e-01 1.66081905e-01 1.79433092e-01
1.75160551e+00 -1.27208364e+00 -2.58119047e-01 -2.88619608e-01
1.25508916e+00 -1.93869099e-01 1.66624010e+00 2.61036623e-02
-8.67292702e-01 -4.00806695e-01 -8.83423507e-01 -3.22564095e-01
-1.42748788e-01 2.80896455e-01 7.44233191e-01 4.88033980e-01
-6.51952088e-01 3.61672878e-01 -5.14956892e-01 -2.11781591e-01
3.28191578e-01 2.64449745e-01 -1.75682232e-01 -1.69582829e-01
-1.21101391e+00 4.60798979e-01 8.22154820e-01 -3.77596945e-01
-7.69702971e-01 -9.06831384e-01 -1.18688321e+00 3.11495721e-01
6.80248976e-01 -7.40422189e-01 1.49278069e+00 -7.99708903e-01
-1.58441544e+00 9.34853733e-01 -3.89014989e-01 -2.71933436e-01
-6.57506213e-02 -1.49566367e-01 7.23932013e-02 2.81572729e-01
2.19516486e-01 2.87033081e-01 4.62835342e-01 -9.72056508e-01
-9.00658906e-01 -3.71919692e-01 9.60364044e-01 2.27823313e-02
-4.61436301e-01 3.62297148e-01 -7.98658788e-01 -1.56210378e-01
-1.42755479e-01 -7.91397929e-01 9.71828625e-02 -2.52069652e-01
-3.54536176e-01 -6.95837557e-01 5.41691244e-01 -9.91373599e-01
1.61441123e+00 -2.25025892e+00 2.64661431e-01 1.88968390e-01
3.06619823e-01 -6.12400845e-02 1.19945575e-02 1.85238436e-01
2.11288080e-01 2.71226794e-01 -2.85965621e-01 2.84382135e-01
2.48649120e-01 4.63821203e-01 -6.00679040e-01 -1.11777723e-01
2.71214932e-01 1.06884789e+00 -8.99864674e-01 -9.22022700e-01
-3.94709110e-01 -5.25670707e-01 -1.15275276e+00 6.95607901e-01
-1.06708694e+00 -1.08643938e-02 -8.38885725e-01 7.21535861e-01
2.69579679e-01 -4.25174296e-01 3.04207832e-01 -1.04599699e-01
2.21543446e-01 7.37848639e-01 -8.99692655e-01 1.69751692e+00
-6.54924154e-01 3.18903953e-01 1.45883963e-01 -1.05621171e+00
7.96811461e-01 7.91978091e-02 -1.90380588e-01 -3.45421493e-01
-4.27533388e-01 3.26986581e-01 -1.65328339e-01 -1.11297369e+00
2.58637220e-01 6.62807524e-02 -4.91790771e-01 5.67949772e-01
2.48500422e-01 -3.01851958e-01 2.56502002e-01 4.55306143e-01
1.50690591e+00 3.73980522e-01 4.04554129e-01 -1.97938718e-02
1.01500046e+00 5.06673336e-01 7.09676802e-01 7.72844493e-01
2.44158849e-01 -2.94298232e-01 1.11418843e+00 -5.00145078e-01
-7.26785958e-01 -7.38990426e-01 2.81942576e-01 1.67066133e+00
-2.23749280e-02 -8.67017925e-01 -8.92957985e-01 -1.28680027e+00
-6.33624494e-02 5.68675518e-01 -3.07699233e-01 -1.27802476e-01
-1.01447892e+00 -3.85134459e-01 8.51714313e-01 5.69285631e-01
3.61359864e-01 -1.26741958e+00 -5.96371293e-01 1.50302485e-01
-5.13379395e-01 -1.00847983e+00 -3.86590004e-01 1.51982620e-01
-8.72621953e-01 -1.53043091e+00 -2.05200049e-03 -1.26588666e+00
9.14525747e-01 -1.72615528e-01 1.42665935e+00 5.16626537e-01
-5.62475808e-03 2.79512018e-01 -3.46868098e-01 -1.03142902e-01
-6.53228879e-01 2.89629370e-01 -6.56874180e-01 -5.06669283e-01
5.80716431e-01 -5.14032066e-01 -1.01011492e-01 3.49147655e-02
-7.49503911e-01 2.34012842e-01 6.39995635e-01 8.14760625e-01
5.19673586e-01 2.00836852e-01 2.10234851e-01 -1.46066844e+00
4.93742108e-01 -4.70516503e-01 -1.01702738e+00 5.76175332e-01
-2.28843838e-01 5.38727701e-01 8.82088482e-01 -4.27922100e-01
-1.19825757e+00 1.31823003e-01 -7.30717257e-02 2.85979849e-03
-9.62258577e-02 9.44607854e-01 -5.87564886e-01 -1.17793644e-03
8.21556747e-01 4.87714827e-01 -3.74586731e-01 -4.69958574e-01
4.91584122e-01 4.78342324e-01 5.99007428e-01 -1.49184597e+00
9.89419222e-01 5.81765324e-02 -5.32835662e-01 -4.78017956e-01
-7.60749221e-01 -4.99485433e-01 -4.45846498e-01 3.14495921e-01
7.31245220e-01 -6.78596139e-01 -7.86968172e-01 1.55468890e-02
-1.55885863e+00 -6.51466429e-01 -2.23201483e-01 -1.09900817e-01
-5.89044750e-01 3.79836261e-01 -7.91246593e-01 -5.62662542e-01
-5.99320889e-01 -1.22247386e+00 9.10546303e-01 2.07903221e-01
-2.18547583e-01 -7.81947672e-01 3.85602683e-01 6.33863568e-01
-1.71815217e-01 -3.02363127e-01 2.05878663e+00 -8.56363595e-01
-9.49137211e-01 -2.08236769e-01 -3.18242788e-01 3.38932067e-01
-2.54447281e-01 -1.31430507e-01 -6.77539766e-01 2.53640652e-01
-2.22775802e-01 -6.47741139e-01 4.59772587e-01 -3.74721885e-01
1.36215782e+00 -6.57979310e-01 -4.44926620e-01 6.66896105e-01
1.35792518e+00 4.13084626e-02 4.31967527e-01 4.72009510e-01
6.98832095e-01 6.45234942e-01 4.81474459e-01 -1.35315685e-02
7.42460430e-01 2.82958120e-01 -3.31952143e-03 4.47288483e-01
2.16139153e-01 -6.54204547e-01 5.28642416e-01 8.92045915e-01
3.82494241e-01 3.91437501e-01 -1.16865158e+00 6.78454280e-01
-1.83245194e+00 -5.19926548e-01 -3.55031565e-02 1.96561289e+00
1.49024737e+00 2.36542132e-02 -8.53972789e-03 -9.51771140e-02
5.18696845e-01 5.63144777e-03 -5.63285947e-01 -2.87824303e-01
4.82931107e-01 3.58912706e-01 2.62713293e-03 4.94543552e-01
-8.46189260e-01 1.12918687e+00 4.87207556e+00 6.57141805e-01
-7.12918639e-01 1.35778487e-01 2.85744648e-02 4.14173752e-01
-5.85570037e-01 5.20374298e-01 -1.19735575e+00 3.16538036e-01
1.04647672e+00 -3.80088896e-01 4.67596143e-01 1.36850464e+00
-3.79573494e-01 -1.24145873e-01 -1.42315805e+00 4.57478672e-01
-1.57321304e-01 -1.34332120e+00 -3.81067097e-02 -4.05113995e-01
3.16598088e-01 -2.12143555e-01 -4.04335171e-01 1.06204462e+00
4.83947843e-01 -3.94178033e-01 6.27077162e-01 1.40450209e-01
5.90312302e-01 -6.44802868e-01 4.62623447e-01 6.93991005e-01
-1.39680350e+00 -2.88894534e-01 -5.03268600e-01 2.35826597e-01
-4.11966860e-01 5.24621248e-01 -8.35219502e-01 4.21818614e-01
7.55667090e-01 3.17799330e-01 -5.16864419e-01 7.04054296e-01
-9.34379041e-01 7.61793315e-01 -2.01326609e-01 -3.83444518e-01
-1.02144629e-01 -1.35936022e-01 3.36128384e-01 1.36501527e+00
3.23987529e-02 4.43693221e-01 3.68502378e-01 1.30984151e+00
-4.42731589e-01 4.32933480e-01 -5.07286489e-01 -1.63269728e-01
3.98188055e-01 1.27224207e+00 -1.99880719e-01 -6.95416331e-01
-8.00754130e-01 7.37722874e-01 9.12878036e-01 2.73957312e-01
-9.49992180e-01 -7.61822641e-01 2.30303541e-01 -1.47126541e-01
4.07351375e-01 3.25032592e-01 -1.21847704e-01 -1.29516888e+00
6.63967252e-01 -1.16374135e+00 9.10507798e-01 -6.09675527e-01
-8.33640039e-01 2.55062461e-01 -2.65440363e-02 -4.57336813e-01
-1.86301962e-01 -3.89829874e-01 -8.02023113e-01 8.52746904e-01
-1.57907152e+00 -1.16112137e+00 -2.92849094e-01 5.54013610e-01
3.21927160e-01 1.15274467e-01 9.45194900e-01 2.67649174e-01
-6.47486746e-01 6.70193493e-01 -5.19566119e-01 5.12232542e-01
4.13231850e-01 -1.37148559e+00 2.13232443e-01 7.27432966e-01
-1.14413850e-01 1.09709728e+00 2.82028168e-01 -6.74241185e-01
-1.83679712e+00 -1.10580993e+00 8.50064456e-01 -5.82579136e-01
9.76915061e-01 -3.85029674e-01 -1.36316991e+00 7.75637329e-01
-1.86259910e-01 6.65050820e-02 8.95175338e-01 3.61271501e-01
-9.04996216e-01 -2.00474530e-01 -7.51879632e-01 4.63309586e-01
7.33758092e-01 -1.07467723e+00 -1.21940756e+00 2.75504529e-01
1.17326140e+00 -5.30563235e-01 -8.69673967e-01 2.03586936e-01
5.27179658e-01 -4.43659097e-01 7.36998737e-01 -8.63758683e-01
8.64033341e-01 -2.91635931e-01 -1.63111821e-01 -8.10445845e-01
1.60289541e-01 -4.53369647e-01 -7.31952250e-01 1.41834974e+00
6.92510247e-01 -5.45336068e-01 9.16585088e-01 7.37847507e-01
2.81474777e-02 -7.24146605e-01 -4.38509703e-01 -4.04486626e-01
-4.76121977e-02 -5.52310705e-01 5.88900924e-01 6.56347275e-01
4.33469623e-01 7.13367522e-01 2.68002212e-01 5.23188174e-01
4.99469995e-01 8.19109917e-01 1.24065828e+00 -1.17609441e+00
-9.49307978e-01 -1.00200638e-01 2.13249564e-01 -1.28786504e+00
6.82300687e-01 -1.13249588e+00 3.20359081e-01 -1.29565966e+00
5.85831285e-01 -7.21288681e-01 1.19525582e-01 7.49580979e-01
-5.35553634e-01 -6.55046284e-01 -2.03082994e-01 3.37858737e-01
-7.38268375e-01 1.69812322e-01 8.29476357e-01 -3.42749268e-01
-3.15574735e-01 6.22630417e-02 -7.13241041e-01 9.65107143e-01
4.14251983e-01 -8.01019788e-01 -4.43489045e-01 -3.68506312e-01
5.82690775e-01 4.85202849e-01 3.39542985e-01 -8.07319820e-01
5.91354907e-01 -1.85051635e-01 -4.09932345e-01 -4.59109902e-01
-2.29758948e-01 -6.75623298e-01 -3.06032985e-01 4.19889867e-01
-4.58097249e-01 -2.42599770e-01 2.11538106e-01 6.74755454e-01
-2.80381560e-01 -9.51706529e-01 5.00825107e-01 -3.26685250e-01
-8.38907897e-01 2.70800292e-01 5.24897501e-02 4.94721204e-01
7.26743221e-01 2.22703651e-01 -4.28730935e-01 1.57188103e-01
-5.02590001e-01 3.48410308e-01 4.95976031e-01 1.63533330e-01
3.24892819e-01 -7.75960386e-01 -2.87138224e-01 2.17128888e-01
5.16326189e-01 4.92802918e-01 -8.12128335e-02 7.05196917e-01
-7.19581068e-01 3.43418956e-01 3.58030498e-01 -4.89531010e-01
-1.51495230e+00 6.03294611e-01 1.10660300e-01 -7.20057547e-01
-5.86385190e-01 9.19063032e-01 4.67447191e-01 -1.07411385e+00
2.03031078e-01 -6.07070148e-01 -1.21562496e-01 -2.82375842e-01
6.73759699e-01 -1.27973527e-01 -9.19805020e-02 7.45232627e-02
-3.28828961e-01 3.42245758e-01 -1.52140200e-01 1.13724731e-01
1.45490396e+00 3.96840423e-01 -9.53592122e-01 1.07564248e-01
1.13680708e+00 1.92585140e-01 -7.87557304e-01 -5.78005493e-01
5.93514502e-01 -1.92310303e-01 -3.05802047e-01 -5.98171592e-01
-7.04552412e-01 8.82691443e-01 -2.85653859e-01 2.70663984e-02
9.86604929e-01 5.22120237e-01 8.51905584e-01 1.08121145e+00
5.87142348e-01 -7.06702173e-01 1.57899797e-01 7.45864093e-01
4.62839544e-01 -9.22435701e-01 -2.70389915e-01 -7.92515218e-01
-1.76716626e-01 1.16090858e+00 1.00340950e+00 2.60167778e-01
1.26605496e-01 3.39122206e-01 -4.03753757e-01 -2.44658396e-01
-8.66888762e-01 3.87207232e-02 8.89006108e-02 5.01113534e-01
1.08827114e-01 -2.57686824e-01 1.10251887e-03 1.24758196e+00
-1.61242142e-01 1.94370478e-01 2.12919518e-01 1.14021850e+00
-7.14673698e-01 -1.18279743e+00 -3.24564070e-01 4.58697438e-01
-4.12530929e-01 -4.65561926e-01 -2.70583272e-01 7.06430078e-01
4.88702208e-02 4.90183741e-01 -3.94415498e-01 -3.32991242e-01
5.77155471e-01 4.80128586e-01 4.67179805e-01 -1.15204465e+00
-6.47697210e-01 -5.68362713e-01 4.01642412e-01 -5.79258502e-01
9.31451172e-02 -2.63228834e-01 -1.55467570e+00 -1.19832203e-01
-6.24719083e-01 5.16148448e-01 3.47515881e-01 9.14416194e-01
2.07631797e-01 3.30165714e-01 3.09204280e-01 1.20666206e-01
-7.42519915e-01 -5.23463547e-01 -1.47472113e-01 3.09703439e-01
2.06759244e-01 -1.57019794e-01 -3.30286264e-01 4.45639461e-01] | [9.666708946228027, 7.597813606262207] |
2cdae8ec-9526-4f54-8bac-c5c72d257be0 | pose-invariant-object-recognition-for-event | 1903.07873 | null | http://arxiv.org/abs/1903.07873v1 | http://arxiv.org/pdf/1903.07873v1.pdf | Pose-Invariant Object Recognition for Event-Based Vision with Slow-ELM | Neuromorphic image sensors produce activity-driven spiking output at every
pixel. These low-power consuming imagers which encode visual change information
in the form of spikes help reduce computational overhead and realize complex
real-time systems; object recognition and pose-estimation to name a few.
However, there exists a lack of algorithms in event-based vision aimed towards
capturing invariance to transformations. In this work, we propose a methodology
for recognizing objects invariant to their pose with the Dynamic Vision Sensor
(DVS). A novel slow-ELM architecture is proposed which combines the
effectiveness of Extreme Learning Machines and Slow Feature Analysis. The
system, tested on an Intel Core i5-4590 CPU, can perform 10,000 classifications
per second and achieves 1% classification error for 8 objects with views
accumulated over 90 degrees of 2D pose. | ['Sunil Kukreja', 'Rohan Ghosh', 'Siyi Tang', 'Nitish Thakor', 'Mahdi Rasouli'] | 2019-03-19 | null | null | null | null | ['event-based-vision'] | ['computer-vision'] | [ 4.21747327e-01 -6.88689113e-01 5.17486274e-01 -2.88334429e-01
-9.25327465e-02 -6.36486351e-01 3.71528387e-01 -9.00750309e-02
-7.50429988e-01 6.79631650e-01 -6.51220381e-01 2.02270091e-01
-4.17049378e-02 -6.17689550e-01 -9.61924791e-01 -9.21689987e-01
1.34839505e-01 2.45360091e-01 4.60709900e-01 1.36619702e-01
6.85728431e-01 8.23041797e-01 -1.98048937e+00 2.90844560e-01
4.52609360e-01 1.43916821e+00 4.88809943e-01 7.84470141e-01
1.74053181e-02 4.79259491e-01 -5.79418123e-01 -8.00043419e-02
3.41319859e-01 -2.64694512e-01 1.39377683e-01 -1.33631304e-01
4.03778166e-01 6.11525886e-02 -2.43747070e-01 1.26235592e+00
7.38527656e-01 -3.83485496e-01 6.71283007e-01 -1.35835123e+00
-6.25441015e-01 1.97511882e-01 -3.95103782e-01 7.04697013e-01
2.93598026e-01 5.98158121e-01 1.27280772e-01 -7.45253384e-01
4.88915771e-01 8.19308758e-01 6.36060417e-01 6.79636955e-01
-1.31022179e+00 -4.92583662e-01 -2.30198249e-01 4.62546200e-01
-1.17212093e+00 -2.13165671e-01 7.79782951e-01 -2.76130021e-01
1.69817376e+00 1.18187882e-01 1.16509652e+00 1.21835697e+00
1.10558331e+00 2.91461319e-01 1.51596093e+00 -4.50174734e-02
1.02443790e+00 -2.04417422e-01 1.97903365e-01 5.80115378e-01
4.88572925e-01 -8.11697170e-02 -9.48542774e-01 -2.44326470e-03
6.55777991e-01 3.69783640e-01 -1.34018838e-01 -2.45713159e-01
-7.56879628e-01 1.03300691e-01 3.55644882e-01 3.00979704e-01
-4.57876176e-01 5.72457016e-01 1.68020770e-01 4.65628922e-01
-2.43090913e-01 4.49875623e-01 -3.49105477e-01 -3.86115789e-01
-5.28727710e-01 -5.20655848e-02 6.47020698e-01 7.38327980e-01
4.38806564e-01 4.62934732e-01 1.26401231e-01 2.83990204e-01
2.94087648e-01 8.33893001e-01 9.68038917e-01 -7.90183902e-01
-1.88833445e-01 1.07751322e+00 -2.29417801e-01 -7.79352784e-01
-4.35988545e-01 -1.91593304e-01 -6.97558582e-01 7.77092814e-01
1.01946704e-01 2.15706393e-01 -1.16117895e+00 1.51347744e+00
-9.00370330e-02 1.19553246e-01 1.53195500e-01 7.01225162e-01
3.61715108e-01 5.94823658e-01 -1.38168465e-02 -3.12860459e-01
1.24413693e+00 -2.87459791e-01 -5.49336970e-01 -3.55874270e-01
-2.89218515e-01 -1.71926901e-01 7.94078767e-01 5.75809836e-01
-1.05250812e+00 -5.81675589e-01 -1.44703591e+00 2.29212493e-01
-7.23358452e-01 -2.95742214e-01 7.06057489e-01 6.54967904e-01
-1.13486028e+00 3.87249410e-01 -1.07504678e+00 -4.01858360e-01
5.98449409e-01 6.53148651e-01 -9.21600312e-03 4.45352763e-01
-4.30268168e-01 8.51849079e-01 2.54271597e-01 -1.68309212e-01
-9.97103691e-01 -4.16322351e-01 -2.69862622e-01 1.05369121e-01
-4.56400543e-01 -5.10552943e-01 7.81667292e-01 -1.22869396e+00
-1.58295250e+00 1.14489484e+00 -5.71844950e-02 -7.13162303e-01
3.92259240e-01 3.20689499e-01 -4.41307575e-01 1.45956039e-01
-3.80479962e-01 6.30115449e-01 1.09811532e+00 -6.62101865e-01
-2.47839496e-01 -1.26811123e+00 -6.40012145e-01 -7.35969990e-02
-6.32295609e-01 -1.18355282e-01 1.40606403e-01 -9.06093121e-02
4.37325716e-01 -8.14454794e-01 -1.24154761e-02 4.39687580e-01
3.50616693e-01 -1.85287103e-01 1.12557399e+00 9.04965550e-02
4.37628597e-01 -2.18163681e+00 -1.41860293e-02 -1.26701996e-01
-5.48056252e-02 3.13294977e-01 1.73423916e-01 2.46025976e-02
6.55613095e-02 -5.20678461e-01 -2.75740981e-01 5.14862463e-02
-1.65996924e-01 2.36940563e-01 -6.03471518e-01 4.91826355e-01
1.87558949e-01 1.14507544e+00 -2.94388294e-01 -1.47107556e-01
3.45465153e-01 4.20140564e-01 -1.89787418e-01 -2.98937913e-02
-2.28332803e-01 2.54861474e-01 -2.03456283e-01 9.70843196e-01
7.12720752e-01 -5.67733236e-02 -4.04562987e-02 -1.73619658e-01
-3.99720699e-01 -4.69891548e-01 -9.53504980e-01 1.76071024e+00
-2.28487402e-02 8.40878308e-01 -3.75153154e-01 -9.50786889e-01
1.34866416e+00 -1.46022007e-01 5.07608593e-01 -1.20218813e+00
6.01671278e-01 3.51282746e-01 -8.96212086e-02 -4.97019947e-01
-2.24607587e-02 1.30778521e-01 -5.36122080e-03 2.44896159e-01
3.55992585e-01 -1.14528257e-02 -1.97152361e-01 -3.40889722e-01
1.49041820e+00 1.60547987e-01 5.67039177e-02 -2.55816907e-01
6.56225160e-02 -1.11680754e-01 4.48098421e-01 8.46685410e-01
-5.89557171e-01 2.34728366e-01 8.88974741e-02 -9.45381522e-01
-9.48728919e-01 -1.60850728e+00 -4.14508879e-01 5.38577855e-01
4.08803403e-01 3.20841908e-01 -8.03731322e-01 2.05719128e-01
3.16570997e-02 5.40394127e-01 -3.44438940e-01 -5.27840853e-01
-2.82469034e-01 -1.03004611e+00 6.00015700e-01 5.84507883e-01
7.64369726e-01 -1.22740197e+00 -1.98076725e+00 3.85279536e-01
7.12801576e-01 -8.87587011e-01 3.76880407e-01 8.58162463e-01
-1.08302534e+00 -7.70574570e-01 -3.89121175e-01 -8.60194564e-01
6.68433309e-01 -4.15363073e-01 6.69072509e-01 -6.94121599e-01
-1.23458242e+00 6.00975037e-01 -1.57080311e-02 -6.47784710e-01
3.53797525e-01 -4.94571626e-01 4.24037606e-01 1.08097240e-01
9.12005305e-01 -1.13477027e+00 -7.14429557e-01 -1.70479596e-01
-6.33409679e-01 -3.35603148e-01 6.37438595e-01 5.34332097e-01
9.36944425e-01 -3.48676801e-01 5.11226714e-01 -1.84994370e-01
3.85712504e-01 -6.94261491e-02 -9.35773551e-01 2.92042196e-01
-7.93393612e-01 2.55113363e-01 8.37028623e-01 -8.71773958e-01
-7.48078883e-01 5.40008783e-01 1.89103410e-01 -3.72258067e-01
-1.82321846e-01 -2.81693518e-01 -1.04451776e-02 -6.42787218e-01
8.04694831e-01 7.48702347e-01 -1.55738756e-01 -1.11611433e-01
-2.82739758e-01 7.55550802e-01 7.55042553e-01 3.50991972e-02
1.84249207e-01 6.25733078e-01 3.14867407e-01 -7.37147152e-01
2.05799583e-02 1.33225068e-01 -2.84183353e-01 -5.85924625e-01
7.26815760e-01 -7.18099117e-01 -1.24945617e+00 1.18869579e+00
-9.33344364e-01 -1.75481495e-02 -2.34595954e-01 2.90357411e-01
-9.55465198e-01 -2.26108551e-01 -4.80896264e-01 -9.64175701e-01
-7.51687944e-01 -8.30313981e-01 6.29509270e-01 8.27183723e-01
1.34528786e-01 -2.80336112e-01 2.04405725e-01 -1.58983171e-01
5.64034164e-01 2.91201651e-01 7.28069603e-01 -1.50265142e-01
-6.95351124e-01 -2.41414726e-01 1.97396383e-01 -5.83518334e-02
-3.05027872e-01 3.00796237e-02 -1.23039591e+00 -3.70614976e-01
6.15610540e-01 -5.19330561e-01 9.25909400e-01 4.86881256e-01
1.25831628e+00 4.72206287e-02 -3.94918591e-01 7.96832323e-01
2.03670454e+00 7.47643173e-01 8.69965792e-01 2.37936452e-01
2.00086623e-01 -1.19616017e-01 -9.47361067e-02 5.32274842e-01
-1.77667245e-01 3.16031158e-01 6.37631714e-01 4.53214258e-01
1.31182354e-02 2.83049196e-02 4.03494954e-01 6.06672883e-01
9.64839086e-02 -8.80516842e-02 -9.22291517e-01 3.19745302e-01
-1.52891529e+00 -8.95224214e-01 7.21496418e-02 2.06929207e+00
5.60765445e-01 3.22297961e-01 -3.52551192e-01 1.70125678e-01
6.43419027e-01 -3.22710931e-01 -1.43092966e+00 -6.65667772e-01
-4.47645634e-01 4.59538102e-01 5.22711039e-01 -2.33481526e-01
-3.96397799e-01 6.41007006e-01 6.07448149e+00 1.67049780e-01
-1.67735910e+00 -1.06208123e-01 2.89810389e-01 -5.51140845e-01
-1.75805446e-02 -1.58378169e-01 -7.09922075e-01 9.32718694e-01
1.16332150e+00 -8.61829892e-02 6.61754012e-01 8.97007048e-01
-3.28511864e-01 -3.57211620e-01 -1.03896618e+00 1.64729202e+00
3.91409457e-01 -1.10891986e+00 1.21986806e-01 -1.79844484e-01
5.00616670e-01 2.92503059e-01 4.81033266e-01 -4.63063791e-02
-2.41551474e-01 -8.07825387e-01 5.97039759e-01 9.76896882e-01
6.80466712e-01 -7.87980616e-01 3.05818021e-01 3.98073494e-01
-7.31927872e-01 -6.26106262e-01 -7.16810405e-01 -3.65972549e-01
-1.62936643e-01 4.05667305e-01 -3.84207010e-01 -3.62808555e-01
1.13358867e+00 2.90246785e-01 -6.16265833e-01 1.01828301e+00
3.71356100e-01 1.50176287e-01 -5.73931456e-01 -7.84702897e-01
-5.49812205e-02 -4.10843156e-02 5.90615630e-01 7.52175748e-01
6.31004512e-01 1.61233872e-01 -4.35502619e-01 1.14909470e+00
-1.59105901e-02 -4.43415254e-01 -8.20286214e-01 -3.95047106e-02
4.48315501e-01 1.18825591e+00 -1.13110328e+00 -1.05320141e-01
-2.14897916e-02 1.41596603e+00 2.09171161e-01 2.91218869e-02
-6.28316998e-01 -6.17938399e-01 4.65081036e-01 -1.59011543e-01
3.03368896e-01 -3.71140093e-01 -6.14041150e-01 -9.62517977e-01
2.13065684e-01 -4.44465458e-01 -1.71960685e-02 -1.06353700e+00
-8.74591410e-01 5.40152490e-01 -5.57155132e-01 -9.15418446e-01
-2.26657182e-01 -1.27749217e+00 -5.00582159e-01 3.33880097e-01
-9.96907055e-01 -7.21279979e-01 -5.35946548e-01 7.86343694e-01
4.20926303e-01 -3.07706118e-01 9.66796994e-01 -7.55359381e-02
-3.83434683e-01 4.77523446e-01 2.02428043e-01 -2.57471830e-01
2.84635395e-01 -1.05320883e+00 2.47513786e-01 6.71027362e-01
3.20319682e-01 2.00696975e-01 6.67604864e-01 -4.15968060e-01
-2.26426101e+00 -7.09479809e-01 4.36234802e-01 -3.13055336e-01
2.37386078e-01 -4.65457052e-01 -7.45856762e-01 4.05928820e-01
-2.08043829e-02 2.64805585e-01 3.66433114e-01 -6.45298302e-01
-4.14935976e-01 -7.30800688e-01 -1.60571933e+00 6.75112426e-01
1.12212348e+00 -5.73962867e-01 -6.59697354e-01 -1.20564498e-01
-5.28340265e-02 -1.08766027e-01 -5.10576427e-01 2.99327493e-01
6.78574264e-01 -8.87912691e-01 5.68068326e-01 -2.72917539e-01
-2.27321699e-01 -3.61860126e-01 -2.24409446e-01 -8.22328687e-01
-1.87689871e-01 -3.81447494e-01 -2.85562783e-01 8.20919275e-01
1.12702414e-01 -8.85556400e-01 7.48129249e-01 5.00911653e-01
8.75947252e-02 -5.91313481e-01 -1.44117200e+00 -8.96962345e-01
-3.80852431e-01 -2.04544753e-01 -4.48828004e-02 2.74652869e-01
-2.05899607e-02 1.36705548e-01 3.25564921e-01 1.57346845e-01
9.54527676e-01 3.70589465e-01 5.74882179e-02 -1.30542874e+00
-1.75585866e-01 -3.74422908e-01 -1.47896039e+00 -2.75595337e-01
1.60612855e-02 -8.59303951e-01 -4.80673872e-02 -1.06598914e+00
4.65174019e-01 2.07971349e-01 -5.28175235e-01 5.09546399e-01
4.88817155e-01 5.53721666e-01 -1.18725821e-01 1.18214250e-01
-6.73296750e-01 4.46595579e-01 3.56060624e-01 -1.72096461e-01
4.26050238e-02 -4.11410779e-01 -1.34238869e-01 7.28879273e-01
9.22331870e-01 -6.88950717e-01 -2.93353498e-01 -5.23327231e-01
4.07433748e-01 -3.34501326e-01 6.42828345e-01 -1.90027428e+00
7.90126920e-01 1.43121287e-01 1.14784360e+00 -4.90923792e-01
4.79866892e-01 -8.83797407e-01 6.07335865e-01 1.15358102e+00
-1.17098004e-01 3.51929843e-01 3.50151896e-01 6.55058444e-01
1.73110783e-01 -3.00366402e-01 1.00317931e+00 -3.83858293e-01
-9.84797299e-01 1.36665761e-01 -6.21500611e-01 -1.84848934e-01
1.52496815e+00 -8.24062824e-01 -3.59112471e-01 3.14707458e-01
-4.16481942e-01 -2.53616542e-01 6.49981141e-01 3.04731339e-01
1.05383670e+00 -1.12649620e+00 -1.38224617e-01 6.25300229e-01
1.07705228e-01 -6.40573621e-01 2.30141521e-01 3.48073840e-01
-5.33129156e-01 1.88688129e-01 -1.28866065e+00 -9.24621761e-01
-1.18717098e+00 7.10307956e-01 5.79264522e-01 5.87924361e-01
-5.78822136e-01 8.60190690e-01 -5.63007116e-01 1.80168033e-01
2.24184513e-01 -5.24738757e-03 4.16755192e-02 -1.12547517e-01
6.28701150e-01 3.10981005e-01 1.35711044e-01 -6.94749504e-02
-7.22364485e-01 8.24539185e-01 -2.17306595e-02 -2.87941843e-01
1.56532049e+00 1.15838364e-01 -2.33003825e-01 8.53316545e-01
1.05741811e+00 -8.66262853e-01 -1.65920711e+00 2.03475609e-01
8.47619623e-02 -1.12857610e-01 3.37788742e-03 -9.30625796e-01
-9.45034444e-01 9.92195249e-01 1.75402248e+00 -7.90210143e-02
1.48833835e+00 -1.10850044e-01 4.59533244e-01 1.00265491e+00
8.65513980e-01 -1.46735609e+00 3.82235199e-01 3.79124403e-01
7.58789897e-01 -8.07684124e-01 -4.06720728e-01 4.24177468e-01
-2.25088000e-01 1.18499172e+00 9.61428523e-01 -6.82967544e-01
3.13662529e-01 1.01731181e+00 -1.68316737e-01 -2.19300747e-01
-9.60545897e-01 1.86946407e-01 -3.84698480e-01 7.93991804e-01
-2.03023627e-01 1.17905855e-01 -2.35608801e-01 3.06004971e-01
5.94526902e-02 2.75094241e-01 3.27627629e-01 1.13982511e+00
-8.64153743e-01 -3.23129267e-01 -8.48723352e-02 5.02847254e-01
-3.63962710e-01 1.58376977e-01 -5.47023356e-01 -8.03042054e-02
1.08932473e-01 4.85254943e-01 5.54491222e-01 -3.10563892e-01
1.23107426e-01 2.59392798e-01 1.11933565e+00 1.16157740e-01
-4.85356987e-01 -5.50913274e-01 -9.15353417e-01 -6.63986862e-01
-5.32436930e-02 -4.85840708e-01 -1.46692276e+00 3.30016464e-02
2.53929645e-01 -5.22612631e-01 1.31612206e+00 5.23660421e-01
7.39400566e-01 4.76825982e-01 6.06427729e-01 -6.07773423e-01
-7.78374553e-01 -6.77671492e-01 -5.91042936e-01 2.65842706e-01
6.58965558e-02 -5.72186947e-01 -3.56869280e-01 1.42049387e-01] | [8.238561630249023, 2.3715646266937256] |
fe5a97ac-9c2c-45f5-a7f7-335a6e654fa1 | benchmark-dataset-for-automatic-damaged | 1812.05581 | null | http://arxiv.org/abs/1812.05581v1 | http://arxiv.org/pdf/1812.05581v1.pdf | Benchmark Dataset for Automatic Damaged Building Detection from Post-Hurricane Remotely Sensed Imagery | Rapid damage assessment is of crucial importance to emergency responders
during hurricane events, however, the evaluation process is often slow,
labor-intensive, costly, and error-prone. New advances in computer vision and
remote sensing open possibilities to observe the Earth at a different scale.
However, substantial pre-processing work is still required in order to apply
state-of-the-art methodology for emergency response. To enable the comparison
of methods for automatic detection of damaged buildings from post-hurricane
remote sensing imagery taken from both airborne and satellite sensors, this
paper presents the development of benchmark datasets from publicly available
data. The major contributions of this work include (1) a scalable framework for
creating benchmark datasets of hurricane-damaged buildings and (2) public
sharing of the resulting benchmark datasets for Greater Houston area after
Hurricane Harvey in 2017. The proposed approach can be used to build other
hurricane-damaged building datasets on which researchers can train and test
object detection models to automatically identify damaged buildings. | ['Valentina Staneva', 'Youngjun Choe', 'Tessa Schneider', 'Sean Andrew Chen', 'Christopher Haberland', 'Andrew Escay'] | 2018-12-13 | null | null | null | null | ['damaged-building-detection'] | ['computer-vision'] | [ 3.46760482e-01 -5.34716368e-01 8.04665804e-01 -1.73471823e-01
-1.10735238e+00 -4.73623097e-01 4.04423892e-01 4.80812341e-01
-5.50566912e-01 6.60190225e-01 4.64503378e-01 -2.14562356e-01
-2.76444107e-01 -1.22896731e+00 -9.67544094e-02 -1.04900527e+00
-5.33621609e-01 6.15341842e-01 3.07421356e-01 -7.74411678e-01
1.78161787e-03 1.12057161e+00 -1.84726822e+00 1.93448350e-01
3.28457326e-01 6.60998940e-01 3.04866761e-01 9.73384261e-01
8.70952070e-01 4.36928183e-01 -5.14757395e-01 4.85602200e-01
5.13849676e-01 4.59569208e-02 -7.99491704e-01 8.60853493e-03
3.09324175e-01 -8.21806669e-01 -2.73778170e-01 6.01579726e-01
1.21394885e+00 2.45222688e-01 7.69469380e-01 -7.80057073e-01
-8.30407962e-02 4.04178858e-01 -4.25137401e-01 8.35965455e-01
4.32058811e-01 2.23702863e-01 6.94723666e-01 -8.98415685e-01
2.66546577e-01 1.02977264e+00 8.25957477e-01 -1.10395275e-01
-1.05462253e+00 -4.76278454e-01 -2.88486898e-01 3.20970654e-01
-1.80383718e+00 -2.76981473e-01 6.07486248e-01 -8.09751034e-01
1.09384859e+00 5.49763799e-01 6.66029453e-01 5.45468926e-01
-2.16159657e-01 5.95863879e-01 1.09034216e+00 -2.17229888e-01
7.85353035e-02 -5.54043353e-01 6.53449167e-03 4.28469926e-01
4.30650532e-01 5.09439230e-01 -1.43063748e-02 -4.00704324e-01
3.15353632e-01 4.95446235e-01 -4.45836693e-01 1.12994269e-01
-1.11809444e+00 9.41503406e-01 8.39216053e-01 5.41207910e-01
-7.75250375e-01 -1.48910016e-01 4.25743550e-01 1.77734748e-01
5.27129054e-01 4.41019356e-01 -1.52467459e-01 2.15364531e-01
-1.42268705e+00 3.53757232e-01 3.06534141e-01 1.70218408e-01
7.31960356e-01 1.17233828e-01 -3.06855701e-02 5.49591362e-01
1.39375612e-01 1.02519858e+00 -9.88880694e-02 -1.98970050e-01
3.51600915e-01 5.47809958e-01 3.40898186e-01 -1.46440005e+00
-9.71812546e-01 -2.89191872e-01 -9.29333925e-01 2.14641362e-01
-1.86313406e-01 -3.31350058e-01 -8.25540602e-01 8.25235844e-01
4.69100356e-01 -3.13531645e-02 -3.74849401e-02 1.14347374e+00
8.44779789e-01 7.83443987e-01 1.69708490e-01 5.81898019e-02
1.18753242e+00 -3.10111314e-01 -2.50319660e-01 -2.53184289e-01
6.15399718e-01 -6.39177144e-01 7.80293941e-01 8.19151625e-02
-3.54377657e-01 -3.50204170e-01 -9.17170942e-01 3.35482836e-01
-5.85638404e-01 1.86040133e-01 1.54449224e-01 1.77240491e-01
-9.73189056e-01 1.11949377e-01 -8.26134622e-01 -8.22794735e-01
3.45924139e-01 5.84385991e-02 -2.27636471e-01 -2.78585911e-01
-1.10279667e+00 1.25578403e+00 4.17962044e-01 3.72209817e-01
-1.69596088e+00 -7.05041647e-01 -8.98864567e-01 -1.41188148e-02
5.69765344e-02 -1.48055732e-01 7.88197696e-01 -1.54976711e-01
-2.63752997e-01 9.11650717e-01 7.10719526e-01 -6.03496730e-01
4.19938624e-01 -3.69656652e-01 -4.17132854e-01 3.67897958e-01
2.25279972e-01 4.33756858e-01 5.20523608e-01 -1.41438198e+00
-9.73768055e-01 -4.91153628e-01 4.15579602e-03 3.02190840e-01
-6.44010827e-02 4.80317473e-01 6.51527226e-01 -7.36135662e-01
3.36205475e-02 -7.57377803e-01 -1.80374905e-01 -3.44556391e-01
-1.16947025e-01 1.91685691e-01 1.11432576e+00 -1.08198678e+00
1.33753240e+00 -2.02755427e+00 -2.77367949e-01 4.22910899e-02
-1.73866838e-01 4.79951113e-01 -5.30475639e-02 1.01890957e+00
3.44458483e-02 4.50281538e-02 -7.70420969e-01 2.93046057e-01
-3.32478315e-01 1.33689970e-01 -8.02818298e-01 6.94478273e-01
-9.72110257e-02 4.30101514e-01 -8.81563127e-01 -3.11281353e-01
4.46959108e-01 5.72340310e-01 1.08190104e-01 3.53894919e-01
3.76039505e-01 6.24440908e-01 -4.08347785e-01 9.92885590e-01
6.73551500e-01 6.27643645e-01 -3.76397669e-01 -4.13545966e-02
-5.76292276e-01 5.18323435e-03 -1.09313393e+00 9.19615328e-01
-9.79937688e-02 5.91054022e-01 2.94581890e-01 -9.90745366e-01
1.13527560e+00 5.19614935e-01 5.78721941e-01 -2.13590831e-01
1.16102435e-01 -9.71117765e-02 -5.93418062e-01 -8.41391265e-01
4.88739759e-01 -3.55413884e-01 -3.52565378e-01 5.88776469e-01
-6.45936847e-01 -5.84504545e-01 1.13678664e-01 -1.73653290e-01
1.40630281e+00 -4.35110360e-01 3.22053939e-01 -4.18651968e-01
4.91949975e-01 6.37643993e-01 1.74683660e-01 6.80637956e-01
-2.84959078e-01 7.13540554e-01 -5.06881237e-01 -1.40956807e+00
-9.36776578e-01 -8.75218391e-01 -3.35793674e-01 1.31657362e+00
-2.80801505e-01 3.46930385e-01 -6.00138545e-01 -2.33833507e-01
-2.04584077e-02 7.16575682e-01 -4.44607168e-01 1.57267690e-01
-4.84788716e-01 -1.31461632e+00 1.06464696e+00 4.97429281e-01
7.41004050e-01 -1.23885047e+00 -1.28558743e+00 3.84615064e-01
-5.85058630e-01 -6.82494879e-01 1.82152435e-01 -9.76363197e-02
-7.34913707e-01 -1.27721715e+00 -6.24350071e-01 -6.75064981e-01
7.63509631e-01 9.37336266e-01 1.06494689e+00 5.90115726e-01
-9.55689311e-01 7.04709530e-01 -5.20891964e-01 -6.35695934e-01
4.67754528e-02 -1.65229067e-01 1.43143058e-01 -1.36504874e-01
3.34574610e-01 -5.64995050e-01 -7.69218743e-01 5.54868937e-01
-1.29232728e+00 -3.57383102e-01 4.70681518e-01 4.05276865e-01
5.13413846e-01 4.76179302e-01 4.61587995e-01 -4.05690111e-02
4.27062243e-01 -5.11356592e-01 -7.08631635e-01 1.22331537e-01
-1.64146185e-01 -4.32745874e-01 9.00936723e-02 7.70687759e-02
-1.01386988e+00 5.62005877e-01 -1.10438459e-01 4.25254583e-01
-8.36584508e-01 9.13774014e-01 3.99343759e-01 1.07739471e-01
1.17428243e+00 1.67246535e-01 -6.44192815e-01 -6.15572035e-01
-1.75122879e-02 9.54612911e-01 7.47178495e-01 -3.20348769e-01
1.15121710e+00 1.06127131e+00 -1.44745767e-01 -1.28176880e+00
-9.83170450e-01 -1.15259981e+00 -9.70387936e-01 -3.27208012e-01
9.65808511e-01 -1.15114558e+00 1.97808653e-01 5.78301847e-01
-9.09944415e-01 -3.67137313e-01 -9.41705704e-02 3.60106289e-01
-1.29117861e-01 2.04987407e-01 -1.09826624e-01 -1.04968429e+00
-8.35335732e-01 -6.29057646e-01 1.19243908e+00 -1.01817273e-01
1.12109669e-01 -6.24125659e-01 8.90679598e-01 6.43189549e-01
5.70554852e-01 9.27419543e-01 3.76051962e-01 -1.89211980e-01
-3.22543204e-01 -6.30528986e-01 -2.46966973e-01 1.48772076e-01
1.64488852e-01 1.53913032e-02 -8.69200587e-01 -8.30701768e-01
-1.37670174e-01 -3.44568700e-01 1.08480775e+00 3.49752426e-01
3.39088738e-01 -3.56176496e-01 -3.51652712e-01 2.02992484e-01
1.61539531e+00 -1.03523523e-01 8.16521645e-01 6.48687184e-01
3.54822725e-01 7.24547744e-01 9.29822385e-01 8.06270242e-01
4.51030582e-01 3.41666102e-01 7.36492097e-01 -3.50427181e-01
4.57978770e-02 3.33097458e-01 3.66293669e-01 1.73400387e-01
-4.72023278e-01 1.45074609e-03 -1.82733369e+00 1.30751920e+00
-1.56102467e+00 -1.47386527e+00 -3.90754551e-01 2.17922735e+00
3.36648941e-01 -5.61367214e-01 -2.93966774e-02 5.29523671e-01
7.84271538e-01 2.26538271e-01 6.96433932e-02 1.39538556e-01
-6.30613938e-02 -1.08708963e-01 6.26117468e-01 4.39476013e-01
-1.74027240e+00 6.09961808e-01 6.68799496e+00 3.93702269e-01
-1.05757606e+00 2.24211931e-01 1.26523688e-01 -9.59786400e-02
2.89637476e-01 -8.27149376e-02 -8.13817322e-01 -1.52524158e-01
7.62036502e-01 1.38973087e-01 2.93198407e-01 8.23880494e-01
6.21523142e-01 -4.22461569e-01 -2.99807340e-01 6.17749870e-01
1.32856160e-01 -9.83409345e-01 -2.37048298e-01 -1.14642292e-01
9.36055958e-01 6.98484957e-01 -3.25724065e-01 3.54198515e-02
4.11208659e-01 -6.64920449e-01 3.39903057e-01 5.04352093e-01
5.53437591e-01 -8.51406932e-01 1.15101361e+00 5.23771107e-01
-1.45350397e+00 -3.49229902e-01 -6.49151802e-01 -4.40429598e-01
2.33002350e-01 6.55704856e-01 -1.30477870e+00 3.69717389e-01
1.28484452e+00 2.54326314e-01 -7.25113988e-01 1.35163748e+00
-6.40155852e-01 6.91243947e-01 -5.53065240e-01 4.09140825e-01
4.10510033e-01 2.90897816e-01 7.58238673e-01 1.28628647e+00
5.63673496e-01 5.32534301e-01 6.71748579e-01 9.66879427e-02
4.97571677e-01 -9.47588757e-02 -1.12430143e+00 3.96618724e-01
2.55964309e-01 1.40231919e+00 -5.94089866e-01 -2.48375475e-01
-7.06437007e-02 4.02535230e-01 -9.98745337e-02 7.90262446e-02
-6.60737276e-01 -3.53498042e-01 4.56043333e-01 5.51035702e-01
2.48065948e-01 -5.31483114e-01 5.20475507e-02 -5.20732522e-01
-4.41965938e-01 -5.58488607e-01 7.73370564e-01 -9.50120211e-01
-1.01050460e+00 7.45893896e-01 4.46435452e-01 -1.26897550e+00
9.73782241e-02 -7.91868791e-02 -7.91351378e-01 4.99761760e-01
-1.73641491e+00 -1.32501292e+00 -9.38970208e-01 6.13802135e-01
3.99497896e-01 1.32148609e-01 9.03809488e-01 3.24251413e-01
-2.65861481e-01 -3.03555995e-01 1.29457593e-01 3.38123709e-01
5.11292875e-01 -8.66241097e-01 3.18575621e-01 1.08239651e+00
-1.44781977e-01 -2.21497968e-01 7.98153758e-01 -8.98010552e-01
-8.63990009e-01 -1.61613655e+00 8.10416996e-01 -3.53960305e-01
6.18556142e-01 1.85841143e-01 -9.21656430e-01 6.19359016e-01
-5.39516937e-03 -1.64503768e-01 7.68396735e-01 -2.72094756e-01
-1.47117386e-02 -2.10975528e-01 -1.44976199e+00 1.58390239e-01
5.54101288e-01 -3.88784230e-01 -9.97577965e-01 7.30336428e-01
5.53651638e-02 7.31144398e-02 -7.53679752e-01 5.37249863e-01
1.09545678e-01 -8.80583048e-01 1.09757340e+00 -1.85469165e-01
3.59722897e-02 -7.21321881e-01 -4.58815783e-01 -1.21152627e+00
-5.37760973e-01 -1.87209144e-01 5.42989314e-01 8.98679078e-01
1.49426043e-01 -2.92882502e-01 1.62250206e-01 2.41434246e-01
-3.55242103e-01 -2.16667563e-01 -1.10854793e+00 -7.39153743e-01
-2.82399714e-01 -3.41661006e-01 5.78964412e-01 5.59273303e-01
-3.99869353e-01 3.95033471e-02 -3.91070336e-01 1.04961514e+00
5.71555197e-01 8.85739699e-02 8.05998385e-01 -1.69069505e+00
4.48901862e-01 -1.39708607e-03 -5.46388149e-01 -1.83256149e-01
-5.36676288e-01 -4.95055169e-01 4.96407270e-01 -1.82589626e+00
1.10787086e-01 -3.80280286e-01 -8.76769274e-02 9.96822119e-01
7.01793432e-02 7.12361515e-01 -1.51061267e-01 5.28951287e-01
-2.79179633e-01 5.42029858e-01 5.96481383e-01 -4.18089837e-01
-1.33137912e-01 -1.50839806e-01 -2.17008010e-01 8.65370333e-01
9.48228002e-01 -9.26266193e-01 5.48913702e-02 -6.65074944e-01
2.68072248e-01 -2.50119567e-01 7.81583548e-01 -1.62044346e+00
1.38911009e-01 -2.32956365e-01 5.02259396e-02 -1.25095582e+00
4.55576293e-02 -1.01399624e+00 3.29921484e-01 8.75922978e-01
2.11925700e-01 2.70309269e-01 2.32547715e-01 4.48056191e-01
-2.74398237e-01 -2.12082699e-01 1.03652477e+00 -2.10489661e-01
-7.20020652e-01 3.17028433e-01 -9.87391174e-01 -2.38238022e-01
1.26340854e+00 7.13820532e-02 -2.81987965e-01 -2.07645625e-01
-4.99966353e-01 1.49532855e-01 4.12532508e-01 1.35365054e-01
7.89275229e-01 -1.10283756e+00 -1.48637331e+00 3.51598822e-02
3.22933763e-01 -1.69195458e-02 4.69436318e-01 7.76319563e-01
-1.18955660e+00 3.59686732e-01 -4.13067847e-01 -4.90176737e-01
-1.50931442e+00 6.17839336e-01 6.40149228e-04 -3.44354749e-01
-6.47308171e-01 5.13471305e-01 -1.66806862e-01 -5.04246354e-01
-1.00318968e-01 -1.84813559e-01 -4.88133520e-01 4.52057511e-01
1.12749720e+00 7.43162870e-01 3.23209643e-01 -1.14833927e+00
-5.18257499e-01 5.93119204e-01 4.98978108e-01 -1.47737801e-01
1.86219072e+00 -2.73688376e-01 -1.16619080e-01 -1.62160061e-02
5.83898902e-01 -3.70247871e-01 -9.93579805e-01 -5.39170578e-02
-1.04781322e-01 -4.06338632e-01 7.07074344e-01 -8.02813947e-01
-9.44678962e-01 9.24325168e-01 1.06128526e+00 3.96185070e-01
1.37011898e+00 -6.01383932e-02 8.39626014e-01 9.42067027e-01
6.03731632e-01 -1.14068997e+00 -2.38733307e-01 4.47755694e-01
1.67363036e+00 -1.29075694e+00 4.99502450e-01 1.15661450e-01
-5.22481263e-01 9.42212403e-01 -5.88062331e-02 -2.43201852e-01
6.56122088e-01 4.33875583e-02 2.75699705e-01 -6.07449651e-01
-1.10029295e-01 -7.14258134e-01 1.80523291e-01 9.35552478e-01
-4.51551192e-03 4.48074669e-01 -2.64055222e-01 -2.07153298e-02
-6.15053773e-02 -3.44085962e-01 5.27830958e-01 1.41684258e+00
-1.28648043e+00 -4.17335808e-01 -1.29215229e+00 3.20746154e-01
-2.92185545e-01 -3.44498843e-01 -3.67836446e-01 7.24054754e-01
2.67553907e-02 1.28847849e+00 -1.38761356e-01 -2.95380712e-01
4.38392937e-01 -1.95957556e-01 -1.50680229e-01 -6.48041725e-01
-7.44583070e-01 -1.70018122e-01 1.02021031e-01 -1.54928371e-01
-5.77348351e-01 -9.08124685e-01 -1.17189920e+00 -3.46804798e-01
-1.97772697e-01 4.46073487e-02 6.76763058e-01 8.67244422e-01
3.38755876e-01 4.35469374e-02 7.18029201e-01 -1.40515006e+00
-4.77210253e-01 -1.22409129e+00 -7.01275468e-01 1.89247444e-01
3.88269693e-01 -8.62159252e-01 -4.53226626e-01 7.93639943e-02] | [9.482156753540039, -1.2939746379852295] |
b0fe534f-19c1-4073-9b25-411330d063f6 | sketchaa-abstract-representation-for-abstract | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Yang_SketchAA_Abstract_Representation_for_Abstract_Sketches_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Yang_SketchAA_Abstract_Representation_for_Abstract_Sketches_ICCV_2021_paper.pdf | SketchAA: Abstract Representation for Abstract Sketches | What makes free-hand sketches appealing for humans lies with its capability as a universal tool to depict the visual world. Such flexibility at human ease, however, introduces abstract renderings that pose unique challenges to computer vision models. In this paper, we propose a purpose-made sketch representation for human sketches. The key intuition is that such representation should be abstract at design, so to accommodate the abstract nature of sketches. This is achieved by interpreting sketch abstraction on two levels: appearance and structure. We abstract sketch structure as a pre-defined coarse-to-fine visual block hierarchy, and average visual features within each block to model appearance abstraction. We then discuss three general strategies on how to exploit feature synergy across different levels of this abstraction hierarchy. The superiority of explicitly abstracting sketch representation is empirically validated on a number of sketch analysis tasks, including sketch recognition, fine-grained sketch-based image retrieval, and generative sketch healing. Our simple design not only yields strong results on all said tasks, but also offers intuitive feature granularity control to tailor for various downstream tasks. Code will be made publicly available. | ['Yi-Zhe Song', 'Honggang Zhang', 'Kaiyue Pang', 'Lan Yang'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['sketch-based-image-retrieval', 'sketch-recognition'] | ['computer-vision', 'computer-vision'] | [ 5.66813387e-02 -1.30964145e-01 -2.82048315e-01 -2.67724127e-01
-2.52698928e-01 -8.31538737e-01 1.19944608e+00 -1.60587877e-01
2.66647160e-01 2.38441259e-01 3.82070988e-01 -2.60912210e-01
-4.83567547e-03 -5.64788282e-01 -2.24397421e-01 -3.18945199e-01
3.40570584e-02 2.40821481e-01 5.40842898e-02 -3.45734596e-01
4.34825361e-01 1.17453873e+00 -1.70030034e+00 6.11193359e-01
3.70922267e-01 8.79889309e-01 8.06773901e-02 7.19977260e-01
-5.11978209e-01 3.18401784e-01 -5.92034340e-01 -5.15065730e-01
2.22838953e-01 -1.55343845e-01 -3.89370650e-01 3.56070340e-01
8.23571146e-01 -4.58233237e-01 -1.39682859e-01 5.77832282e-01
7.56908432e-02 -7.74761364e-02 9.06625688e-01 -1.35106111e+00
-9.86488640e-01 2.81736016e-01 -5.34244359e-01 -1.95617840e-01
4.41073805e-01 3.37937802e-01 1.45213509e+00 -1.24558067e+00
6.83173418e-01 1.59718585e+00 5.31490803e-01 6.37717128e-01
-1.65495241e+00 -5.86320460e-01 5.45805871e-01 -6.36048317e-01
-1.41382158e+00 -4.89067167e-01 9.15206909e-01 -6.69102967e-01
6.45116925e-01 6.52192056e-01 8.98692548e-01 8.36545467e-01
4.24569920e-02 8.77630711e-01 1.08319354e+00 -3.94028783e-01
2.28532687e-01 9.18295979e-02 2.20083911e-02 9.07602608e-01
3.51562113e-01 -8.48185793e-02 -7.18741834e-01 -4.38467771e-01
1.48909938e+00 2.81911850e-01 -7.87939429e-02 -6.50830805e-01
-1.14170456e+00 7.58607864e-01 4.16052639e-01 1.27682522e-01
-3.80584925e-01 5.82203209e-01 3.28088313e-01 2.31808409e-01
2.47268572e-01 5.96911848e-01 -5.46125360e-02 9.30225253e-02
-1.04428995e+00 4.99723911e-01 6.57603979e-01 1.03303587e+00
5.56637347e-01 2.39191324e-01 -3.54276925e-01 9.05731261e-01
2.56266952e-01 2.56831259e-01 -1.83144882e-02 -9.65455592e-01
1.05994068e-01 5.45223951e-01 1.74860999e-01 -9.72057998e-01
3.84572931e-02 -1.64612114e-01 -8.62120509e-01 9.05473232e-01
3.10783207e-01 4.97998595e-01 -9.88836825e-01 1.75907421e+00
-1.14657236e-02 -3.83830220e-01 -4.88672912e-01 9.41067517e-01
7.97977030e-01 5.54130077e-01 5.17479539e-01 4.70019400e-01
1.81228209e+00 -7.09065735e-01 -4.74789292e-01 1.53040979e-02
5.67976758e-02 -6.65377617e-01 1.67334759e+00 3.76116306e-01
-1.33935511e+00 -6.91974580e-01 -1.24665475e+00 -2.88099378e-01
-4.46972311e-01 2.99284607e-01 9.67309713e-01 6.86614573e-01
-1.03216684e+00 5.63031435e-01 -5.22791207e-01 -4.24064875e-01
5.16683817e-01 7.58156329e-02 -4.11348939e-01 1.61316141e-01
-7.11044252e-01 6.86269283e-01 -1.82732105e-01 -1.01033293e-01
-3.96216929e-01 -7.12605655e-01 -7.37998426e-01 4.59438145e-01
5.11836587e-03 -1.15264463e+00 1.17318738e+00 -9.06722128e-01
-1.38419139e+00 9.59778607e-01 -1.20240577e-01 4.09116633e-02
4.44036812e-01 1.33506889e-02 -5.76614551e-02 6.05479516e-02
-1.71606615e-01 8.75847042e-01 1.37244248e+00 -1.72925913e+00
-2.94926822e-01 -1.95112184e-01 2.41012156e-01 1.61699817e-01
-1.32884860e-01 -1.13411151e-01 -5.41271150e-01 -1.17312849e+00
1.04016863e-01 -9.16892469e-01 -1.65925533e-01 8.72008443e-01
-1.16155192e-01 -1.99261919e-01 8.22629690e-01 -2.83759385e-01
1.28661299e+00 -2.16307282e+00 2.05536515e-01 2.77048439e-01
6.03275120e-01 2.49593869e-01 -2.55233616e-01 9.44499552e-01
7.07314312e-02 3.45787704e-01 -1.01792961e-01 -5.54386377e-01
5.10027289e-01 2.18069613e-01 -8.23725343e-01 1.44060375e-02
3.64916414e-01 1.19592774e+00 -7.02358186e-01 -3.94248784e-01
3.20665985e-01 5.38406551e-01 -5.39470196e-01 2.06853732e-01
-2.82630473e-01 -2.29091439e-02 -5.54920375e-01 7.79202998e-01
6.41995966e-01 -3.70936543e-01 2.42698655e-01 -3.04674149e-01
-4.49452177e-02 3.84666510e-02 -1.15740800e+00 1.68803871e+00
-5.71816683e-01 6.35683954e-01 3.00514698e-01 -2.82704026e-01
9.59726691e-01 1.03076570e-01 -1.13301456e-01 -5.05248308e-01
-4.19669449e-01 -5.05806096e-02 -2.42017969e-01 7.23453090e-02
7.55084753e-01 -4.74916190e-01 -2.76480317e-01 7.40536511e-01
-2.62018174e-01 -6.81832969e-01 -4.33148481e-02 4.51708645e-01
6.27638102e-01 2.72412479e-01 4.92615074e-01 -4.24745172e-01
2.77973592e-01 -3.05013508e-01 -6.78106993e-02 9.36456680e-01
1.93727255e-01 7.64121234e-01 4.85112786e-01 -8.04401636e-01
-1.07790637e+00 -1.44656765e+00 1.48818372e-02 1.30199027e+00
3.84257659e-02 -8.94978166e-01 -4.20694709e-01 -3.46464187e-01
4.61675793e-01 4.53782707e-01 -7.66536236e-01 1.19768560e-01
-4.69160467e-01 7.26465741e-03 5.25596201e-01 7.27712154e-01
1.20093869e-02 -1.13743317e+00 -9.33540463e-01 -1.29428208e-01
5.25475085e-01 -7.60496080e-01 -6.57911420e-01 -1.94650576e-01
-8.33775580e-01 -4.94926900e-01 -9.96871889e-01 -4.82040197e-01
6.91589057e-01 4.96284544e-01 1.46403265e+00 7.80213058e-01
-5.35085917e-01 6.56752229e-01 -1.12491511e-01 -3.13932091e-01
-3.16368908e-01 -1.86714008e-01 -9.34853852e-02 -4.96060215e-02
-1.14076786e-01 -7.19237924e-01 -6.89286053e-01 1.58417866e-01
-1.15321231e+00 4.94728655e-01 7.24072993e-01 8.45118046e-01
3.86851251e-01 -5.00726581e-01 3.38635147e-01 -6.95129514e-01
1.15360510e+00 7.13080019e-02 -3.20683360e-01 3.29580337e-01
-2.87410259e-01 3.10643911e-01 5.58952153e-01 -4.51078355e-01
-9.04374480e-01 1.38214966e-02 -4.85345302e-03 -4.66413498e-01
-1.43341780e-01 2.15405747e-01 -3.76197360e-02 -6.69261962e-02
3.84233207e-01 8.89144465e-02 3.45943645e-02 -5.75834274e-01
9.08390164e-01 3.62098247e-01 3.58248115e-01 -1.22833538e+00
8.53180587e-01 5.33275366e-01 2.11474836e-01 -1.19945407e+00
-2.05767080e-01 -1.58447251e-02 -7.13869691e-01 -2.08473969e-02
5.51645100e-01 -6.32292628e-01 -7.97146797e-01 8.77485573e-02
-1.19806910e+00 -3.33564520e-01 -3.80033225e-01 -2.97945768e-01
-6.36929214e-01 3.12190205e-01 -6.42589629e-01 -9.26500261e-01
-3.20338756e-01 -1.15906417e+00 1.55191457e+00 3.22617739e-02
-7.66226053e-01 -8.29497099e-01 -1.70799166e-01 -2.26628542e-01
4.94232774e-01 2.93359250e-01 1.21997142e+00 9.68919098e-02
-9.58333075e-01 -7.46575892e-02 -6.15126967e-01 -1.05346046e-01
1.28354818e-01 3.19974154e-01 -9.85391498e-01 -2.53619581e-01
-6.22082174e-01 -2.62996703e-01 8.29301000e-01 7.88433701e-02
1.33802545e+00 -3.51344049e-01 -2.87061930e-01 6.22711599e-01
1.33672249e+00 -1.69484630e-01 4.99574721e-01 -1.30119249e-01
7.69342422e-01 5.16347289e-01 1.20793886e-01 5.32592833e-01
1.36359811e-01 1.03021789e+00 7.57797584e-02 -4.08666641e-01
-4.42877710e-01 -5.35320282e-01 -1.91271871e-01 4.64034498e-01
-2.00523645e-01 -3.14922654e-03 -7.10227609e-01 2.20695794e-01
-1.69281387e+00 -1.03678429e+00 2.79089987e-01 2.02840829e+00
5.82774699e-01 1.93943486e-01 3.32930028e-01 -1.11557461e-01
2.67976701e-01 3.94421518e-01 -2.10184917e-01 -8.99418056e-01
1.19382724e-01 3.59313279e-01 -1.03530541e-01 5.17872036e-01
-6.52557552e-01 9.10550952e-01 6.84957695e+00 7.00903416e-01
-1.19047582e+00 -3.67644250e-01 4.06640828e-01 8.83203447e-02
-8.02016437e-01 8.91897976e-02 -3.72162342e-01 2.42132634e-01
1.66823164e-01 3.84550318e-02 5.29406607e-01 6.31004453e-01
-2.29109839e-01 2.38512814e-01 -1.57301080e+00 1.24399519e+00
-2.56506562e-01 -1.82096827e+00 9.63507891e-01 4.53855135e-02
2.14049414e-01 -7.37160623e-01 4.30486828e-01 1.48861125e-01
2.70247757e-01 -1.36856008e+00 1.05848706e+00 6.43152356e-01
1.27928638e+00 -5.51248252e-01 -2.95861959e-01 1.47640482e-01
-1.64339304e+00 6.90971985e-02 -1.54487148e-01 -3.17947537e-01
9.45157185e-02 8.31095409e-03 -5.11442602e-01 2.68206149e-01
3.27792704e-01 3.31980228e-01 -5.87650955e-01 5.45697927e-01
-6.39484599e-02 -4.20353860e-02 -9.44035277e-02 -1.13548286e-01
3.85217816e-01 -1.68680921e-01 3.93519431e-01 1.79828358e+00
7.19706863e-02 3.87210667e-01 1.51830852e-01 1.31083226e+00
1.08589597e-01 -2.60533690e-01 -9.33174074e-01 -2.37496704e-01
7.34273493e-01 1.25889301e+00 -9.13209379e-01 -4.22907203e-01
-3.16507638e-01 1.15753877e+00 4.30021137e-01 5.26350081e-01
-4.50758457e-01 -5.22286296e-01 1.09220517e+00 3.25317532e-01
3.69876683e-01 -5.72900236e-01 -6.75474942e-01 -1.12938631e+00
-3.51842493e-02 -8.21557879e-01 2.27269083e-02 -1.06483448e+00
-1.36129701e+00 5.64648330e-01 1.00479126e-01 -1.09294784e+00
-1.93897888e-01 -8.11918557e-01 -8.26238632e-01 1.23349214e+00
-1.01697922e+00 -1.70240653e+00 -3.63174707e-01 2.56698281e-01
7.48430371e-01 6.73801750e-02 1.05189323e+00 -7.31685609e-02
-8.35944116e-02 7.85140336e-01 -3.33954334e-01 2.22396664e-02
4.16687220e-01 -1.23273587e+00 1.16369057e+00 3.89656246e-01
5.47513425e-01 1.30501115e+00 7.49172091e-01 -4.56528097e-01
-1.73235428e+00 -4.09925818e-01 5.23117483e-01 -8.13962102e-01
4.67553318e-01 -8.30851972e-01 -8.95861566e-01 5.52756846e-01
8.08827877e-02 1.08792754e-02 4.41292733e-01 2.73708314e-01
-9.70015049e-01 1.44671872e-01 -1.06796420e+00 1.16220081e+00
1.14129329e+00 -9.00311530e-01 -6.06929779e-01 -2.48092785e-01
4.38668430e-01 -3.69549654e-02 -8.70549202e-01 -1.30786048e-03
1.31120515e+00 -9.47646618e-01 1.49373007e+00 -8.26855004e-01
4.25909549e-01 -3.28607410e-01 -2.71422058e-01 -1.24386072e+00
-7.94941545e-01 -7.13954926e-01 -1.85545802e-01 1.00793159e+00
4.84860092e-02 -4.61906254e-01 7.17657268e-01 7.77935565e-01
1.44097716e-01 -9.06230867e-01 -4.15666848e-01 -6.89201772e-01
7.85479993e-02 -5.29170446e-02 9.37234998e-01 7.60774612e-01
1.41904324e-01 4.11426455e-01 -2.01797783e-01 -2.03037962e-01
6.91187501e-01 5.50489902e-01 1.08748209e+00 -1.55362797e+00
-2.73282260e-01 -1.13958275e+00 -3.14617813e-01 -1.26782775e+00
2.50286539e-04 -6.74825251e-01 -3.91906232e-01 -1.54828799e+00
1.05872750e-01 -8.40687931e-01 -4.86090183e-02 3.01893830e-01
-7.44013190e-02 4.50092524e-01 7.76366830e-01 4.38574612e-01
-2.62162983e-01 4.31830108e-01 1.47634816e+00 -1.29869178e-01
-1.44299185e-02 -2.62697875e-01 -9.02359903e-01 7.10385621e-01
5.09671926e-01 8.89302231e-03 -5.36774337e-01 -4.68223363e-01
-3.53173651e-02 2.66664803e-01 6.62924588e-01 -6.60037637e-01
-1.17387697e-01 -4.25689906e-01 6.79327786e-01 -4.71743166e-01
8.73612761e-01 -6.99330091e-01 3.65945041e-01 3.24052274e-01
-5.24331033e-01 2.60044038e-01 3.00921470e-01 4.16254044e-01
3.10903769e-02 8.29178691e-02 7.83393800e-01 -3.63691121e-01
-6.76011026e-01 3.35401952e-01 -2.55094856e-01 -2.82594651e-01
5.31869113e-01 -5.51220000e-01 -1.08175717e-01 -5.08641362e-01
-6.11594379e-01 -2.52957553e-01 9.82761085e-01 5.98803818e-01
7.23525167e-01 -1.69943154e+00 -5.14562249e-01 5.03579021e-01
3.65360886e-01 -4.60523725e-01 4.43095267e-02 9.35767516e-02
-3.98348898e-01 4.50702488e-01 -6.81529701e-01 -5.80356181e-01
-1.29141462e+00 6.93065524e-01 -3.77063975e-02 6.47430941e-02
-1.01656449e+00 6.94037318e-01 9.03371990e-01 -1.18035279e-01
2.38884583e-01 -6.39266968e-01 3.33288789e-01 -1.36719614e-01
7.33025908e-01 2.35816371e-03 -2.24696040e-01 -4.38278675e-01
-2.22377166e-01 8.27932954e-01 -6.76395446e-02 -3.32322359e-01
1.01099110e+00 -5.50260134e-02 6.62629455e-02 4.51959401e-01
8.92671347e-01 1.32867530e-01 -1.47844911e+00 4.00340185e-02
-8.21645781e-02 -7.27823436e-01 -3.54632020e-01 -8.07492256e-01
-5.90826809e-01 1.12283349e+00 3.66849042e-02 4.95188326e-01
7.07153261e-01 5.46395704e-02 5.31553745e-01 2.84503460e-01
4.56041336e-01 -5.24927795e-01 2.34646931e-01 1.89630345e-01
1.47195697e+00 -7.83880711e-01 3.19735587e-01 -2.19614938e-01
-7.57857203e-01 1.21431637e+00 4.00456280e-01 -3.64617914e-01
3.86209726e-01 3.55298191e-01 -1.12559557e-01 -4.80218858e-01
-7.49136567e-01 7.28337616e-02 6.65410340e-01 5.22113562e-01
6.13168836e-01 3.67826372e-01 -1.54639287e-02 4.59811151e-01
-1.08775264e-02 -8.57034847e-02 2.31049061e-01 9.86702025e-01
-4.93992150e-01 -1.21476781e+00 -4.11017865e-01 3.71697038e-01
-4.29382212e-02 -4.73785475e-02 -7.81820059e-01 9.83364522e-01
-2.02783674e-01 2.93901652e-01 7.19229952e-02 -1.20305993e-01
3.50884974e-01 1.42819911e-01 8.64099383e-01 -5.74272573e-01
-5.80623209e-01 -2.07252949e-01 -1.56931877e-02 -4.99436110e-01
1.13350369e-01 -3.75060141e-01 -9.36194301e-01 -3.74211222e-01
2.53910422e-01 -2.20238999e-03 6.59277737e-01 4.46483016e-01
4.57723528e-01 2.62087822e-01 1.51476264e-01 -1.46443689e+00
-6.02396250e-01 -4.47275668e-01 -6.61523283e-01 5.87679446e-01
6.52724206e-01 -9.21421707e-01 -5.71756996e-02 7.96221383e-03] | [11.723766326904297, 0.3266708552837372] |
c54b9746-bdce-4e2a-8a76-d03f99d793d3 | redwoodnlp-at-semeval-2021-task-7-ensembled | null | null | https://aclanthology.org/2021.semeval-1.171 | https://aclanthology.org/2021.semeval-1.171.pdf | RedwoodNLP at SemEval-2021 Task 7: Ensembled Pretrained and Lightweight Models for Humor Detection | An understanding of humor is an essential component of human-facing NLP systems. In this paper, we investigate several methods for detecting humor in short statements as part of Semeval-2021 Shared Task 7. For Task 1a, we apply an ensemble of fine-tuned pre-trained language models; for Tasks 1b, 1c, and 2a, we investigate various tree-based and linear machine learning models. Our final system achieves an F1-score of 0.9571 (ranked 24 / 58) on Task 1a, an RMSE of 0.5580 (ranked 18 / 50) on Task 1b, an F1-score of 0.5024 (ranked 26 / 36) on Task 1c, and an RMSE of 0.7229 (ranked 45 / 48) on Task 2a. | ['Ryan Chi', 'Nathan Chi'] | 2021-08-01 | null | null | null | semeval-2021 | ['humor-detection'] | ['natural-language-processing'] | [-3.09598058e-01 2.87964404e-01 -2.09894031e-01 7.44111240e-02
-6.97228611e-01 -6.42753184e-01 7.67068863e-01 3.93498063e-01
-3.47103089e-01 8.39837193e-01 5.81165016e-01 -6.37059391e-01
1.24774940e-01 -4.98684645e-01 -3.45316052e-01 -4.41471264e-02
1.31702110e-01 2.35166490e-01 1.67627111e-01 -6.03884995e-01
7.59225547e-01 1.23492911e-01 -9.01207030e-01 6.77195966e-01
8.63306463e-01 7.34880507e-01 -1.41172796e-01 9.62178111e-01
-1.46557406e-01 1.59958029e+00 -7.92458773e-01 -7.84561872e-01
-2.55312562e-01 -4.07995999e-01 -1.32023120e+00 -3.80824149e-01
2.81601369e-01 -2.57922728e-02 -6.15588762e-02 1.21957028e+00
2.19072536e-01 5.42042665e-02 7.82069981e-01 -1.12413836e+00
-4.48256135e-01 6.85123920e-01 -5.82222819e-01 1.42767563e-01
7.95967758e-01 9.26894695e-02 1.27407110e+00 -1.30248022e+00
5.38675845e-01 1.07276535e+00 7.45718360e-01 3.29282641e-01
-9.74064469e-01 -6.04481936e-01 -4.10201579e-01 2.94802248e-01
-1.07069767e+00 -3.86996835e-01 5.97388744e-01 -8.45247388e-01
9.24524903e-01 2.87716985e-01 1.56188130e-01 1.07294691e+00
3.73510092e-01 8.84464145e-01 1.43374920e+00 -4.34562385e-01
-4.01524454e-02 2.94237554e-01 4.11308765e-01 6.90138817e-01
5.60054556e-02 -4.59519506e-01 -8.41856301e-01 -3.11798662e-01
4.29000229e-01 -5.25308728e-01 -1.36391237e-01 7.52180159e-01
-1.08604777e+00 1.22424376e+00 2.81429768e-01 4.42988098e-01
-3.58527541e-01 -3.09354186e-01 5.57018757e-01 2.62382716e-01
4.42241341e-01 8.21701467e-01 -2.90361971e-01 -4.50935930e-01
-1.05415857e+00 5.10058045e-01 1.36745787e+00 9.13564265e-01
4.56397891e-01 -1.58144593e-01 -4.32826102e-01 1.05805922e+00
-1.32715086e-05 4.71268058e-01 6.03366435e-01 -9.07732546e-01
4.96717870e-01 5.23586154e-01 5.47790051e-01 -1.35436559e+00
-8.17442775e-01 -8.71638298e-01 -7.60228693e-01 -3.92296553e-01
4.78172332e-01 -1.37430737e-02 -4.46377665e-01 1.42847466e+00
-3.68465304e-01 -6.19216204e-01 -1.89738676e-01 8.24453712e-01
9.61009383e-01 7.83033311e-01 8.61421898e-02 -5.95748305e-01
1.31678855e+00 -1.50448680e+00 -5.39834499e-01 -5.87336302e-01
8.29512060e-01 -1.25641072e+00 1.50789535e+00 5.45023501e-01
-1.21523380e+00 -4.99828100e-01 -8.92777979e-01 -3.46430093e-01
-1.02155961e-01 2.67668664e-01 -8.58765934e-03 2.59185284e-01
-8.18979502e-01 3.82761329e-01 -7.66403973e-02 -2.70881832e-01
-1.44460842e-01 -2.24444479e-01 -1.44758239e-01 -8.31024647e-02
-1.34114671e+00 1.28151798e+00 5.21578379e-02 -4.06675011e-01
-6.63065314e-01 -5.41451037e-01 -5.36166966e-01 3.35598886e-02
2.07532808e-01 -4.06383723e-01 1.40296805e+00 -6.69799030e-01
-1.27840161e+00 1.10596704e+00 -4.95162517e-01 -5.39955795e-01
5.90212584e-01 -5.33003271e-01 -3.39068294e-01 -6.72286153e-02
5.24777830e-01 4.83910888e-02 4.78711694e-01 -9.90973473e-01
-5.03302872e-01 -4.68179770e-02 6.30166456e-02 -2.67482549e-02
-3.12930942e-01 4.62074578e-01 4.85732891e-02 -4.18718666e-01
-1.64018154e-01 -8.36377561e-01 3.96332331e-02 -5.93397737e-01
-6.25397325e-01 -2.39908934e-01 4.48194325e-01 -1.18573666e+00
1.84368002e+00 -1.73022616e+00 4.19767248e-03 2.06738207e-02
4.50570971e-01 2.97606945e-01 -1.54651301e-02 7.34385848e-01
1.59413889e-01 3.37757826e-01 -1.60190716e-01 -2.57055700e-01
3.71053852e-02 -5.44676423e-01 -3.15864742e-01 1.19497091e-01
-7.60012269e-02 8.94701302e-01 -1.01968348e+00 -2.79159188e-01
-1.59125313e-01 -2.14148406e-02 -2.48753101e-01 3.06327879e-01
-7.45851547e-02 7.15662818e-03 -2.38433897e-01 5.80157042e-01
3.71422619e-01 -5.44395685e-01 8.16275626e-02 1.30607828e-01
-3.19748819e-01 9.96478319e-01 -5.11391699e-01 1.01365829e+00
-5.70149839e-01 9.51843798e-01 -6.54546395e-02 -5.23621082e-01
1.16412568e+00 7.25727603e-02 1.73973680e-01 -7.59640217e-01
-1.61656889e-03 2.47758985e-01 6.32513314e-02 -5.43827891e-01
5.80726385e-01 -5.10639846e-01 -2.85536230e-01 6.51508689e-01
-2.53270090e-01 -7.46976882e-02 3.54267716e-01 5.57921827e-01
1.48257959e+00 -4.75444019e-01 5.80704212e-01 -5.69672227e-01
9.33750033e-01 -2.58849114e-02 1.91870525e-01 8.58718455e-01
-4.43937898e-01 4.89218533e-01 9.54447150e-01 -4.11118120e-01
-1.21031392e+00 -4.45821762e-01 -3.84141654e-02 1.17286003e+00
-4.02244598e-01 -8.74883652e-01 -5.52577257e-01 -6.88405812e-01
2.32602600e-02 1.34218264e+00 -3.81571949e-01 3.92543115e-02
-3.73896509e-01 -4.92189169e-01 6.14930272e-01 1.97179303e-01
5.48164606e-01 -1.01256120e+00 -5.22771299e-01 2.38260701e-01
-8.07784438e-01 -1.24611425e+00 -3.98342609e-01 2.30222079e-03
-5.09454787e-01 -1.12222075e+00 -6.03857100e-01 -3.87196541e-01
-9.75633878e-03 3.46323967e-01 1.49747598e+00 4.53722119e-01
5.21833062e-01 -1.18613420e-02 -5.14173865e-01 -3.07136953e-01
-5.48402131e-01 2.48375207e-01 1.12692058e-01 -3.90138894e-01
3.97373229e-01 -1.49811119e-01 -1.40514433e-01 2.73568779e-01
-4.39195216e-01 2.64202654e-01 3.82182389e-01 9.77657616e-01
-8.93889219e-02 -1.75479323e-01 7.41194904e-01 -9.43371475e-01
1.17588902e+00 -5.86120427e-01 -2.45361045e-01 2.18255848e-01
-9.63158667e-01 -4.29988921e-01 8.41644049e-01 -2.57010698e-01
-7.58277118e-01 -3.51913869e-01 -5.00645256e-03 4.25460115e-02
2.70014733e-01 1.04043913e+00 2.49055550e-01 2.92323947e-01
1.06042230e+00 5.56708984e-02 -3.07326078e-01 -4.04430151e-01
3.20260853e-01 1.12836981e+00 6.24142706e-01 -4.51804996e-01
9.04369652e-01 -1.72795802e-01 -3.20542127e-01 -9.56891894e-01
-1.63218379e+00 -7.66377866e-01 -6.53154016e-01 -1.83989912e-01
6.42756641e-01 -1.00554192e+00 -6.75787807e-01 5.20055175e-01
-1.42503202e+00 -4.69140559e-01 3.40761006e-01 2.11882398e-01
-2.83916622e-01 3.88283432e-01 -8.82244468e-01 -1.03089881e+00
-7.20251739e-01 -6.30921364e-01 5.93312681e-01 2.60078698e-01
-1.00841284e+00 -9.85815525e-01 8.42738077e-02 1.03503740e+00
3.59125316e-01 9.60239321e-02 1.10657835e+00 -1.01114380e+00
1.45509228e-01 -3.12925428e-01 -5.50804615e-01 4.47942823e-01
-9.76062864e-02 -5.72408848e-02 -7.50866830e-01 -2.06051439e-01
6.73632100e-02 -6.78909123e-01 6.93696082e-01 -4.91224341e-02
7.09720850e-01 -8.32613230e-01 1.51273414e-01 -2.42193893e-01
9.32591081e-01 -4.07886133e-02 7.27959752e-01 6.68960750e-01
4.47657287e-01 6.82922840e-01 5.32220840e-01 5.43506861e-01
5.64761758e-01 3.74462247e-01 1.44448997e-02 5.73766947e-01
8.23257565e-02 -5.87214887e-01 6.03670776e-01 8.82663786e-01
1.01315297e-01 -1.18482038e-01 -1.33651173e+00 7.16992795e-01
-1.82398641e+00 -9.91198242e-01 -9.42685783e-01 2.04517579e+00
9.03399348e-01 4.62951452e-01 5.03479064e-01 7.79128000e-02
4.67688441e-01 3.56294364e-01 -6.91004395e-02 -6.80858970e-01
-1.93767682e-01 -2.22256295e-02 -1.69167966e-02 7.93045640e-01
-8.80668223e-01 1.20294178e+00 6.03392649e+00 6.27063632e-01
-8.50607753e-01 2.61583901e-03 4.55376238e-01 2.70114746e-02
-3.19760978e-01 1.44151196e-01 -8.04278076e-01 6.76118016e-01
1.12608540e+00 -6.27282381e-01 4.15897876e-01 8.57502580e-01
3.08387756e-01 -3.43059808e-01 -8.25399220e-01 7.98802376e-01
4.32353497e-01 -1.04486227e+00 -7.74914399e-02 -8.76080897e-03
9.86984670e-01 -5.44769205e-02 -1.39714420e-01 7.04319715e-01
1.99204862e-01 -1.45232522e+00 7.14021027e-01 2.88158655e-01
4.86044109e-01 -5.65175056e-01 1.12114978e+00 8.87701869e-01
-4.43632811e-01 -1.94581792e-01 -4.43219364e-01 -8.37161899e-01
1.15571082e-01 1.06803370e+00 -9.87107813e-01 1.58194706e-01
5.66444457e-01 5.57883978e-01 -4.74215478e-01 7.35503793e-01
-1.00531614e+00 9.05814111e-01 -9.72308160e-04 -4.28776592e-01
2.36810371e-01 1.64667875e-01 6.26940131e-01 1.68711245e+00
-6.51249215e-02 1.73302606e-01 -2.22753044e-02 8.19571614e-01
-4.82546777e-01 3.58190924e-01 -3.06834877e-01 -2.31972709e-01
5.05193055e-01 1.22666276e+00 -2.39790857e-01 -4.02010620e-01
-1.98416337e-01 8.35981965e-01 5.55964649e-01 2.58413225e-01
-7.10996687e-01 -4.24376726e-01 1.56514928e-01 2.35633478e-01
-1.93007499e-01 -2.90201575e-01 -7.26273835e-01 -1.17084444e+00
-1.49748057e-01 -1.18604243e+00 4.52678561e-01 -9.34032440e-01
-1.50429845e+00 5.15590131e-01 -2.93084860e-01 -5.86213410e-01
-2.36295164e-01 -7.07582533e-01 -8.65557909e-01 1.15516233e+00
-1.03289437e+00 -1.00790024e+00 -1.67660639e-01 1.26712784e-01
4.90458906e-01 -2.30889931e-01 6.99291229e-01 5.49843535e-03
-3.72687757e-01 4.29568708e-01 -9.73528624e-02 2.57942468e-01
1.02852249e+00 -1.03980494e+00 1.44478664e-01 7.26658702e-01
3.07757538e-02 7.06574559e-01 9.91661906e-01 -7.88812995e-01
-7.84477651e-01 -9.00955975e-01 2.07492924e+00 -7.33400106e-01
1.04130542e+00 4.59712557e-02 -1.17746186e+00 5.93133628e-01
8.06329399e-02 -7.27836370e-01 6.88131988e-01 6.04698598e-01
-7.06116319e-01 4.02809441e-01 -9.44686115e-01 2.23729700e-01
6.28746688e-01 -1.04444933e+00 -7.64250100e-01 6.25130534e-01
4.84312207e-01 -5.04929185e-01 -1.01205218e+00 2.86288708e-01
6.86496735e-01 -1.03852522e+00 4.61084008e-01 -9.40040529e-01
1.37851608e+00 -1.12682693e-01 -1.79231584e-01 -1.03564775e+00
-7.50848770e-01 -5.14853239e-01 -3.04007940e-02 8.94938886e-01
8.29470158e-01 -2.47630656e-01 5.21176577e-01 6.12392604e-01
7.58320838e-02 -7.57150888e-01 -6.91239178e-01 -6.63267791e-01
6.64160430e-01 -1.88700378e-01 -2.75918394e-01 1.13700581e+00
4.64086533e-01 9.39047158e-01 -5.50482512e-01 -2.47146741e-01
4.29045439e-01 2.18326017e-01 8.95700157e-01 -1.04544759e+00
-8.49301964e-02 -6.92805409e-01 1.63915962e-01 -1.04931736e+00
3.02285314e-01 -7.11844027e-01 -3.27859670e-01 -1.61052227e+00
8.61335516e-01 2.96275496e-01 -1.76525101e-01 5.93210816e-01
-2.33703434e-01 1.04514480e-01 4.11415726e-01 5.81798792e-01
-4.12639737e-01 4.05246794e-01 7.75824666e-01 -1.10766061e-01
-1.80114210e-02 3.53672430e-02 -9.29318309e-01 7.86258519e-01
9.96723235e-01 -5.34710228e-01 4.72933007e-03 -9.50985253e-02
4.53323007e-01 6.44185096e-02 4.00335938e-01 -5.38956642e-01
3.75327051e-01 -3.71008992e-01 1.93214700e-01 -5.75684130e-01
-1.74237676e-02 8.92290995e-02 -3.45252395e-01 5.44161081e-01
-4.10962135e-01 -5.33255972e-02 1.73970088e-01 8.95685703e-02
-2.24567398e-01 -6.13478541e-01 8.28898728e-01 -1.61336094e-01
-3.15094516e-02 -5.65781891e-01 -6.40595257e-01 4.24693853e-01
4.21665877e-01 3.45954150e-01 -7.20159531e-01 -8.89511526e-01
-3.84350121e-01 2.70541817e-01 4.54366475e-01 4.74300593e-01
4.49791342e-01 -1.00863647e+00 -1.03989089e+00 -4.52681601e-01
2.23060488e-03 -3.73061240e-01 -8.29754472e-02 1.06029105e+00
-3.45323503e-01 7.21926808e-01 -6.64175227e-02 6.56327559e-03
-1.27153909e+00 1.89777821e-01 1.17578320e-01 -7.44910121e-01
-2.87883040e-02 7.97871768e-01 -1.05910525e-02 -2.92565465e-01
-1.35169685e-01 3.23673546e-01 -1.76782250e-01 1.40672445e-01
5.66422284e-01 7.78970897e-01 -1.54817984e-01 -8.23330045e-01
-4.34000015e-01 1.09746926e-01 -2.17722148e-01 -3.34551394e-01
1.15021515e+00 2.92014256e-02 -4.60758686e-01 6.40534401e-01
1.03760862e+00 3.44655424e-01 -3.04626703e-01 -2.41494432e-01
5.88388085e-01 -1.97944865e-01 -1.09902378e-02 -1.30894482e+00
-2.25708455e-01 8.71909618e-01 -3.59211832e-01 2.68010497e-01
6.70703053e-01 -1.96422562e-01 8.49107683e-01 5.53300440e-01
2.88637936e-01 -1.24558771e+00 4.11080331e-01 1.41011035e+00
1.31792343e+00 -1.22084451e+00 2.59218961e-01 -2.03508019e-01
-9.75748599e-01 1.42157197e+00 6.56795502e-01 2.82097101e-01
1.89324304e-01 -1.22568816e-01 2.14566384e-02 -3.37545484e-01
-8.56774151e-01 2.73558527e-01 5.03327668e-01 -2.47601673e-01
9.52883422e-01 1.59831941e-01 -9.15471494e-01 1.22401726e+00
-6.62522316e-01 -6.31792843e-02 7.33461261e-01 4.87681746e-01
-8.52886736e-01 -5.54919660e-01 -2.10187346e-01 2.53368407e-01
-5.89763284e-01 -2.80269563e-01 -1.15616477e+00 6.15288734e-01
-3.92179728e-01 1.31558824e+00 -4.11473125e-01 -6.98031008e-01
2.14633256e-01 4.73923624e-01 5.43567687e-02 -4.92507756e-01
-7.75896728e-01 -7.10915774e-02 6.74658537e-01 -3.73439580e-01
-8.01680386e-02 -4.19582993e-01 -1.03141499e+00 -8.10698390e-01
-2.46502459e-01 3.75130534e-01 4.99295980e-01 9.96510506e-01
1.37496948e-01 2.72725113e-02 5.53703129e-01 -6.12887777e-02
-9.37183917e-01 -1.35984743e+00 -4.01565194e-01 1.94278762e-01
3.32136415e-02 -3.84024292e-01 -7.11552978e-01 -4.28610034e-02] | [8.869324684143066, 11.076233863830566] |
b9851bb4-d1a6-414d-b310-fa5c526608c4 | vit-net-interpretable-vision-transformers | null | null | https://proceedings.mlr.press/v162/kim22g/kim22g.pdf | https://proceedings.mlr.press/v162/kim22g/kim22g.pdf | ViT-NeT: Interpretable Vision Transformers with Neural Tree Decoder | Vision transformers (ViTs), which have demonstrated a state-of-the-art performance in image classification, can also visualize global interpretations through attention-based contributions. How- ever, the complexity of the model makes it difficult to interpret the decision-making process, and the ambiguity of the attention maps can cause incorrect correlations between image patches. In this study, we propose a new ViT neural tree decoder (ViT-NeT). A ViT acts as a backbone, and to solve its limitations, the output contex- tual image patches are applied to the proposed NeT. The NeT aims to accurately classify fine-grained objects with similar inter-class correlations and different intra-class correlations. In addition, it describes the decision-making process through a tree structure and prototype and en- ables a visual interpretation of the results. The proposed ViT-NeT is designed to not only improve the classification performance but also provide a human-friendly interpretation, which is effective in resolving the trade-off between performance and interpretability. We compared the performance of ViT-NeT with other state-of-art methods using widely used fine-grained visual categorization benchmark datasets and experimentally proved that the proposed method is superior in terms of the classification performance and interpretability. The code and models are publicly available at https://github.com/jumpsnack/ViT-NeT. | ['Sangwon Kim; Jaeyeal Nam; Byoung Chul Ko'] | 2022-07-17 | null | null | null | icml-2022-7 | ['fine-grained-image-classification', 'fine-grained-visual-categorization'] | ['computer-vision', 'computer-vision'] | [ 1.49850219e-01 1.26887292e-01 7.83032030e-02 -4.57566559e-01
-1.88455999e-01 -4.79310185e-01 4.36024815e-01 1.13950267e-01
-4.64300811e-02 4.52319950e-01 9.55582131e-03 -5.31536102e-01
-1.47709176e-01 -6.41170025e-01 -6.23736560e-01 -7.28112519e-01
2.56135464e-01 3.30791026e-01 1.35835245e-01 8.67766961e-02
3.27495784e-01 3.04054737e-01 -1.71930695e+00 7.88797140e-01
1.07513428e+00 1.38574696e+00 3.93067837e-01 5.23347259e-01
-3.15979570e-01 8.55978310e-01 -8.33532512e-01 -5.15977085e-01
-1.73452869e-01 -4.22830284e-01 -8.74962628e-01 1.54882550e-01
4.89234954e-01 -7.23809004e-02 -5.91107868e-02 1.23073006e+00
1.26052767e-01 -2.24200025e-01 8.98831010e-01 -1.39634132e+00
-1.31175447e+00 3.92711520e-01 -7.88780212e-01 3.80011231e-01
-1.79212958e-01 2.04247952e-01 1.05948281e+00 -1.00836742e+00
2.23905385e-01 1.54714632e+00 4.76802915e-01 3.27763438e-01
-1.21752977e+00 -5.02231300e-01 5.19633532e-01 7.26345599e-01
-1.27805531e+00 -5.92776686e-02 5.61454833e-01 -6.22053385e-01
7.14928091e-01 4.66768920e-01 6.66269183e-01 1.04668343e+00
5.85518539e-01 6.83647811e-01 1.21184337e+00 -3.12832028e-01
2.75165558e-01 -6.65917993e-02 4.78693366e-01 6.80763364e-01
3.29218656e-01 2.37591728e-03 -3.57131720e-01 3.30963850e-01
8.49801540e-01 7.96158165e-02 -2.72897959e-01 -1.30871072e-01
-1.15735519e+00 6.30452156e-01 1.11618745e+00 3.03373218e-01
-3.16349477e-01 2.48919025e-01 2.98409194e-01 8.84664357e-02
3.78512412e-01 2.52319932e-01 -1.26475200e-01 2.83140391e-01
-5.40894747e-01 -2.02972144e-01 2.53730208e-01 5.13773084e-01
6.53979480e-01 1.02203749e-02 -5.85114837e-01 8.52664769e-01
3.98954451e-01 2.82152742e-01 4.01568681e-01 -7.38670349e-01
3.96736443e-01 8.19503605e-01 -1.94038555e-01 -1.24337435e+00
-3.21782649e-01 -8.06811929e-01 -1.32290244e+00 6.82256103e-01
4.32387859e-01 2.98908532e-01 -1.24134958e+00 1.51702118e+00
-1.09518161e-02 -6.99015781e-02 -9.41299498e-02 1.17914152e+00
1.06557429e+00 5.23378849e-01 2.96007186e-01 1.50352642e-01
1.81623971e+00 -1.28272855e+00 -8.70784283e-01 -4.57204252e-01
1.55728713e-01 -7.86564231e-01 1.28047574e+00 2.88095653e-01
-7.85137236e-01 -1.03097856e+00 -1.15875816e+00 -2.71652937e-01
-4.77204680e-01 5.17858922e-01 4.07818317e-01 3.06673825e-01
-1.09581161e+00 2.89469987e-01 -7.12383866e-01 -3.22574914e-01
7.74810791e-01 1.31134704e-01 -1.42984226e-01 1.91895932e-01
-8.64443421e-01 9.13056612e-01 4.67407882e-01 5.21495581e-01
-8.33819449e-01 -3.66026908e-01 -6.77938819e-01 4.89075482e-01
2.15624541e-01 -8.42280865e-01 1.28642464e+00 -1.36277425e+00
-1.03325224e+00 7.54675567e-01 -4.25117135e-01 -3.01696628e-01
4.96695787e-01 5.20962402e-02 -2.51233011e-01 2.32730266e-02
2.08917618e-01 8.74592841e-01 8.17980349e-01 -1.35697365e+00
-6.03097260e-01 -4.81593758e-01 1.40315264e-01 2.26836666e-01
-1.72833547e-01 -3.78204554e-01 -4.79715645e-01 -8.47631872e-01
1.28945649e-01 -5.88894308e-01 2.45437492e-02 2.88931102e-01
-4.47469503e-01 -2.31812268e-01 1.04883206e+00 -5.47048986e-01
1.14454401e+00 -2.33823538e+00 7.86909163e-02 -9.39486250e-02
5.36450446e-01 4.37636882e-01 -1.80560395e-01 1.18948773e-01
-1.98456883e-01 4.55268532e-01 -3.08326334e-01 -3.86103302e-01
-1.04458474e-01 3.63172323e-01 -2.99283862e-01 5.30988872e-02
3.40473235e-01 1.12694979e+00 -6.67326152e-01 -4.04397130e-01
4.34619635e-01 3.83902997e-01 -1.37912080e-01 1.45693660e-01
-2.92464226e-01 3.90509486e-01 -3.32601309e-01 5.87481320e-01
8.74658704e-01 -6.52246475e-01 1.69984370e-01 -4.53143924e-01
-5.89887910e-02 -7.45145231e-02 -8.68483186e-01 1.27564335e+00
-3.45422655e-01 1.01392186e+00 -3.51160347e-01 -1.03706217e+00
9.31962311e-01 -3.18599977e-02 -3.40081930e-01 -9.40327287e-01
1.89230651e-01 1.51459919e-02 2.04596668e-01 -4.75639492e-01
3.57033283e-01 2.76550353e-01 1.63382500e-01 2.41382062e-01
-1.24119915e-01 2.81857073e-01 -1.00328866e-02 -3.35745290e-02
5.74831247e-01 -9.55585483e-03 5.09990394e-01 -3.67323935e-01
6.07562423e-01 -3.32444794e-02 5.04295111e-01 7.97734022e-01
-6.88469112e-02 7.28521705e-01 8.04271936e-01 -7.28666842e-01
-8.67843211e-01 -1.12814760e+00 -1.22512117e-01 1.02613497e+00
4.87684399e-01 -3.75263244e-01 -8.98100197e-01 -7.66945958e-01
-2.32363865e-02 8.42763603e-01 -1.05980027e+00 -5.34608327e-02
-1.01953201e-01 -6.71961904e-01 4.58819926e-01 8.19947422e-01
1.03447950e+00 -1.14160860e+00 -6.36943221e-01 -1.70712247e-01
-4.61546421e-01 -9.21481788e-01 -3.10824156e-01 2.70269036e-01
-7.39114821e-01 -1.12170613e+00 -5.08950353e-01 -8.38580310e-01
1.00766540e+00 3.75470102e-01 1.09688413e+00 3.54214787e-01
-2.28637025e-01 -2.55269915e-01 -2.69591957e-01 -5.51437020e-01
-1.88693359e-01 2.14332342e-02 -3.99713933e-01 2.85905898e-01
3.43622118e-01 -2.79073119e-01 -7.55402684e-01 5.50736427e-01
-8.30520809e-01 5.64585805e-01 7.42979825e-01 1.00161767e+00
7.11948991e-01 5.82577195e-03 4.03893501e-01 -7.39206731e-01
6.24687433e-01 -1.83620736e-01 -4.70910519e-01 5.63505769e-01
-6.44260764e-01 2.68304408e-01 7.74780631e-01 -1.62947208e-01
-1.17225635e+00 -2.34142706e-01 -2.56506857e-02 -4.11895633e-01
-2.79037088e-01 3.98479104e-01 -4.69556265e-02 2.34388709e-02
5.72049797e-01 1.14392638e-01 -2.09841400e-01 -5.44379532e-01
3.24604899e-01 8.37663889e-01 5.32915652e-01 -3.51615757e-01
3.44969451e-01 3.68213624e-01 -1.71732143e-01 -5.67758977e-01
-1.03797638e+00 -4.04136404e-02 -7.68986940e-01 -1.52820766e-01
1.02685678e+00 -6.15895450e-01 -7.98871577e-01 6.58412457e-01
-1.40339768e+00 -2.43901566e-01 -3.11992336e-02 -4.05778689e-03
-2.71453470e-01 3.63753408e-01 -4.59138811e-01 -6.61058009e-01
-4.24444377e-01 -1.33473647e+00 1.05638552e+00 4.28300828e-01
-1.43615887e-01 -9.77461398e-01 -4.66167063e-01 4.32050198e-01
3.66658568e-01 1.52770400e-01 1.18599880e+00 -1.46638334e-01
-6.28641307e-01 4.62842286e-01 -9.12131310e-01 3.09046358e-01
1.24594249e-01 1.61956877e-01 -1.28919816e+00 -3.78725417e-02
-2.18677759e-01 -1.43052982e-02 1.09654510e+00 4.59349006e-01
1.82129347e+00 -3.74249846e-01 -3.17700267e-01 7.06949174e-01
1.30006099e+00 2.34387279e-01 7.62247682e-01 3.88192266e-01
7.45739579e-01 4.85757142e-01 5.58588922e-01 1.13589644e-01
5.00916123e-01 7.77927876e-01 7.41734743e-01 -5.31948805e-01
-4.56100404e-01 -3.21484268e-01 3.89919733e-03 5.60557961e-01
-1.10371239e-01 -5.07136405e-01 -9.16986823e-01 3.34302604e-01
-2.25001454e+00 -7.09100544e-01 -4.35502291e-01 1.95854974e+00
3.57607454e-01 1.46998584e-01 -2.82805204e-01 1.21040098e-01
8.46985698e-01 1.44093692e-01 -6.85224116e-01 -7.72838771e-01
-1.38119876e-01 1.65603384e-01 2.82643348e-01 4.80133712e-01
-1.01329768e+00 9.00966883e-01 6.01265574e+00 8.23163152e-01
-1.33480608e+00 1.25950962e-01 1.14302731e+00 3.63118887e-01
-1.09263398e-01 -2.36752525e-01 -4.61235881e-01 4.72867519e-01
4.84993219e-01 -3.90568040e-02 1.34134457e-01 6.78112984e-01
2.96923846e-01 -1.67912006e-01 -1.08314419e+00 1.08211040e+00
3.38805132e-02 -1.39790297e+00 4.65864897e-01 -3.89886871e-02
4.25214857e-01 -2.03043297e-01 2.51352727e-01 2.37863399e-02
1.63500421e-02 -1.36348510e+00 1.01298797e+00 4.15098578e-01
9.22988415e-01 -5.35208821e-01 1.03665233e+00 2.56040961e-01
-1.40635085e+00 -2.47925028e-01 -5.17298758e-01 -3.62138599e-01
-1.71341628e-01 5.24876535e-01 -6.13436520e-01 5.75041890e-01
1.05215228e+00 8.40515792e-01 -9.48381186e-01 9.83380079e-01
-7.14697480e-01 4.86730665e-01 1.12724729e-01 2.49296222e-02
3.30507725e-01 -8.00965950e-02 1.88063920e-01 1.21661663e+00
1.65633455e-01 -2.80332402e-04 -3.08082402e-02 1.25751364e+00
4.28106487e-02 -1.56168833e-01 -2.77374804e-01 9.52624008e-02
2.72178560e-01 1.25461280e+00 -9.27372515e-01 -5.63607275e-01
-4.95551191e-02 1.06226158e+00 5.66235065e-01 5.09604812e-01
-9.24078703e-01 -2.82576174e-01 7.77149260e-01 -9.46606770e-02
4.31381047e-01 1.90051291e-02 -9.60107744e-01 -1.11125219e+00
9.99894813e-02 -8.22434604e-01 4.50182557e-01 -1.23990190e+00
-1.27986479e+00 8.67925525e-01 -1.79107919e-01 -1.13595200e+00
1.91581622e-01 -7.81305373e-01 -6.70552671e-01 8.92532825e-01
-1.42480123e+00 -1.20340586e+00 -8.56026888e-01 3.52210701e-01
7.63048410e-01 1.58410430e-01 7.55091906e-01 1.05671555e-01
-6.63528025e-01 6.96746290e-01 -2.79151760e-02 1.31595522e-01
3.62861603e-01 -1.29006648e+00 5.58208823e-01 8.11864734e-01
2.68732548e-01 4.48989093e-01 5.04293382e-01 -3.09504896e-01
-8.04371297e-01 -1.14581430e+00 6.55142307e-01 -2.52906233e-01
3.46380472e-01 -4.59486425e-01 -1.09537125e+00 5.69775224e-01
3.77547354e-01 -9.36285481e-02 3.89020890e-01 9.79246795e-02
-4.93186474e-01 -3.00935924e-01 -9.38763559e-01 5.99514723e-01
9.62538183e-01 -3.92621845e-01 -7.87499428e-01 4.47708890e-02
5.43200910e-01 -4.12416339e-01 -3.27435106e-01 2.20810980e-01
7.04865694e-01 -1.23784268e+00 9.06917810e-01 -3.01069260e-01
6.83699191e-01 -7.64055669e-01 1.18104666e-01 -1.42632711e+00
-8.19647551e-01 1.26738340e-01 2.09622487e-01 1.11572528e+00
3.89381588e-01 -8.48755240e-01 2.96558440e-01 1.13668256e-01
-1.49850339e-01 -7.87356138e-01 -7.57063568e-01 -4.64109302e-01
-1.81444302e-01 -3.17300677e-01 5.91979563e-01 6.32445753e-01
-3.27109158e-01 5.18291175e-01 -1.72810778e-01 3.34744811e-01
6.63296402e-01 2.42857963e-01 3.98327321e-01 -1.17403841e+00
-1.96015641e-01 -7.37260401e-01 -6.05694652e-01 -1.03903043e+00
-1.02920711e-01 -7.55864382e-01 -4.42657396e-02 -1.92922986e+00
3.08169991e-01 -3.15512478e-01 -4.76523191e-01 8.00746858e-01
-3.21502149e-01 5.23300946e-01 4.44964826e-01 3.87851357e-01
-5.17060220e-01 4.58412260e-01 1.44610453e+00 -4.27175909e-01
1.11609891e-01 -2.26600975e-01 -9.15606618e-01 7.48971164e-01
8.78033042e-01 -3.02880138e-01 -5.14823735e-01 -8.45634520e-01
-1.43717527e-01 -2.30208144e-01 6.53654695e-01 -1.04086900e+00
2.13293865e-01 4.20418680e-02 7.81831563e-01 -6.68245912e-01
1.92335293e-01 -6.73752666e-01 2.31268659e-01 5.62563300e-01
-3.90156269e-01 1.82095379e-01 2.89380491e-01 5.55464327e-01
-5.69773018e-01 -7.53277987e-02 8.76425385e-01 -5.79639561e-02
-9.66667235e-01 4.96598333e-02 -3.48520100e-01 -2.19022527e-01
9.92901087e-01 -3.46720815e-01 -7.99944758e-01 -2.20218912e-01
-6.54131413e-01 2.56211907e-01 4.28400725e-01 5.88757396e-01
6.69205666e-01 -1.22671711e+00 -5.65855742e-01 3.19695681e-01
3.75823736e-01 -1.59366158e-04 3.85330468e-01 6.81148052e-01
-7.08004534e-01 4.57095087e-01 -5.28133571e-01 -9.98210609e-01
-1.47004580e+00 4.33518171e-01 5.83467245e-01 -7.43283406e-02
-5.45517623e-01 7.58523464e-01 1.07499564e+00 -9.51291472e-02
2.34442383e-01 -6.53272867e-01 -3.67248386e-01 -6.17810860e-02
6.15530670e-01 3.46608162e-02 -4.14703914e-04 -4.67861205e-01
-4.54524517e-01 8.46241355e-01 -3.90018001e-02 3.09011370e-01
8.69051099e-01 -2.46837735e-01 -1.88484579e-01 4.53958899e-01
7.90762842e-01 -5.68085909e-01 -1.23704576e+00 -1.01893388e-01
-2.30789602e-01 -5.47734499e-01 -1.02535365e-02 -1.19322240e+00
-1.18866694e+00 1.37703872e+00 8.16461504e-01 2.69366473e-01
1.32812929e+00 -2.28535328e-02 3.46975088e-01 1.05455853e-01
7.62956589e-02 -6.76912248e-01 7.44464546e-02 5.21374762e-01
1.13168705e+00 -1.33602738e+00 -2.12131321e-01 -6.06364429e-01
-8.65826488e-01 1.24951959e+00 8.99965763e-01 1.81847051e-01
3.38021398e-01 1.08252279e-01 2.99536943e-01 -2.94991821e-01
-8.29706311e-01 -1.98000178e-01 5.51300108e-01 6.35190010e-01
4.48962629e-01 3.54473382e-01 -1.70731202e-01 6.00318611e-01
-1.78271160e-01 -1.39648980e-02 2.31490374e-01 2.96538472e-01
-2.91557163e-01 -7.18621373e-01 -4.79910523e-01 3.47955674e-01
-2.40018427e-01 -1.29443988e-01 -6.11691475e-01 5.67820668e-01
4.43502128e-01 1.02184987e+00 2.98920304e-01 -4.73158956e-01
2.49456584e-01 -1.53953582e-01 1.82506651e-01 -3.20302457e-01
-6.21703088e-01 -1.93093076e-01 -7.53356367e-02 -4.61879939e-01
-2.14115158e-01 -1.27488598e-01 -1.13381481e+00 -3.52430016e-01
-9.04595107e-02 2.04558801e-02 5.34310520e-01 8.40625703e-01
6.51243508e-01 1.07846141e+00 3.23549062e-01 -6.09669447e-01
-2.43277684e-01 -9.72720981e-01 -2.87927002e-01 3.92965019e-01
3.37474048e-01 -7.85467446e-01 -2.86009163e-01 1.88226551e-01] | [10.137771606445312, 2.0016448497772217] |
c87007b0-127a-4d24-91e5-f326f7b5bd07 | fast-and-high-quality-blind-multi-spectral | 2103.09943 | null | https://arxiv.org/abs/2103.09943v4 | https://arxiv.org/pdf/2103.09943v4.pdf | Fast and High-Quality Blind Multi-Spectral Image Pansharpening | Blind pansharpening addresses the problem of generating a high spatial-resolution multi-spectral (HRMS) image given a low spatial-resolution multi-spectral (LRMS) image with the guidance of its associated spatially misaligned high spatial-resolution panchromatic (PAN) image without parametric side information. In this paper, we propose a fast approach to blind pansharpening and achieve state-of-the-art image reconstruction quality. Typical blind pansharpening algorithms are often computationally intensive since the blur kernel and the target HRMS image are often computed using iterative solvers and in an alternating fashion. To achieve fast blind pansharpening, we decouple the solution of the blur kernel and of the HRMS image. First, we estimate the blur kernel by computing the kernel coefficients with minimum total generalized variation that blur a downsampled version of the PAN image to approximate a linear combination of the LRMS image channels. Then, we estimate each channel of the HRMS image using local Laplacian prior to regularize the relationship between each HRMS channel and the PAN image. Solving the HRMS image is accelerated by both parallelizing across the channels and by fast numerical algorithms for each channel. Due to the fast scheme and the powerful priors we used on the blur kernel coefficients (total generalized variation) and on the cross-channel relationship (local Laplacian prior), numerical experiments demonstrate that our algorithm outperforms state-of-the-art model-based counterparts in terms of both computational time and reconstruction quality of the HRMS images. | ['Petros T. Boufounos', 'Hassan Mansour', 'Dehong Liu', 'Lantao Yu'] | 2021-03-17 | null | null | null | null | ['pansharpening'] | ['computer-vision'] | [ 5.93931913e-01 -5.23255765e-01 4.98584598e-01 2.31179059e-01
-9.07590270e-01 -7.51873791e-01 4.56004977e-01 -5.53454638e-01
-3.04408401e-01 5.52242994e-01 1.99678093e-01 -1.14278324e-01
-3.81642759e-01 -4.33045238e-01 -6.29732430e-01 -1.18671417e+00
2.77919620e-01 9.61951166e-02 4.95525971e-02 3.04819178e-02
1.49482593e-01 6.48781240e-01 -1.39500821e+00 1.33784607e-01
1.36180639e+00 8.58497322e-01 6.56176329e-01 1.06838310e+00
3.54183853e-01 4.05004293e-01 -2.14979738e-01 -9.80026796e-02
6.32522941e-01 -4.86612856e-01 -5.52657187e-01 5.85895419e-01
6.83880270e-01 -4.12325621e-01 -3.81870240e-01 1.53665209e+00
2.47014135e-01 1.94927752e-01 5.49466193e-01 -7.04173684e-01
-8.97099733e-01 4.35456298e-02 -1.26306319e+00 1.70782804e-02
8.81523937e-02 2.08320975e-01 5.57446122e-01 -1.03527069e+00
3.32230479e-01 8.76878798e-01 8.27750385e-01 -7.31475353e-02
-1.72778380e+00 -3.13544065e-01 -4.78776902e-01 2.47481018e-01
-1.54219306e+00 -4.51209247e-01 5.67059219e-01 -4.56873536e-01
3.89467031e-01 5.07856905e-01 4.19791907e-01 4.08314675e-01
1.50302602e-02 9.82797742e-02 1.64401412e+00 -5.58150709e-01
7.53626451e-02 -1.43386900e-01 3.53832543e-02 3.69040221e-01
3.36219400e-01 4.01511133e-01 -2.50893146e-01 -3.92387152e-01
1.09564114e+00 -1.27264604e-01 -9.17046249e-01 -4.01689142e-01
-1.25254238e+00 5.65514803e-01 5.09040713e-01 2.84555286e-01
-6.24328136e-01 -6.47760183e-02 -2.33451113e-01 4.28482220e-02
3.19188625e-01 5.11923850e-01 -4.16755944e-01 2.49152303e-01
-1.38184726e+00 2.50542909e-01 4.87750053e-01 5.46387732e-01
1.17763591e+00 -1.29220426e-01 -1.55989215e-01 9.65497434e-01
1.68420300e-01 9.74394441e-01 2.08518282e-01 -1.15608132e+00
2.15303078e-01 -2.38259826e-02 6.49928212e-01 -7.49649525e-01
-1.86362073e-01 -2.94462860e-01 -1.18653166e+00 2.03051731e-01
5.94216526e-01 -1.83025673e-01 -1.01598299e+00 1.32062352e+00
2.79610693e-01 3.77020299e-01 1.42527610e-01 1.32835543e+00
1.44344464e-01 1.00479221e+00 -4.45943713e-01 -6.67407632e-01
1.34131098e+00 -9.57265496e-01 -6.86116159e-01 -4.48772281e-01
-8.20275769e-02 -1.29609489e+00 7.23013699e-01 4.01905358e-01
-1.09097373e+00 -5.52961588e-01 -7.50201643e-01 -5.16763031e-02
1.05848629e-02 5.38174987e-01 1.51281565e-01 4.87408131e-01
-1.15756297e+00 5.07502317e-01 -5.06675899e-01 -4.82636429e-02
-4.08126302e-02 -1.30929247e-01 -1.75841436e-01 -3.44079614e-01
-8.70188117e-01 1.03562927e+00 1.79503039e-01 2.90110201e-01
-6.09811604e-01 -1.05021524e+00 -7.26500869e-01 -3.25697847e-02
1.81533813e-01 -6.25224829e-01 1.00058508e+00 -1.18658328e+00
-1.45996642e+00 6.69467568e-01 -2.70767361e-01 -3.40111613e-01
5.26262105e-01 -2.78930128e-01 -4.25182998e-01 5.20787597e-01
9.99599323e-02 1.43661469e-01 1.67900538e+00 -1.55015016e+00
-3.11072320e-01 -1.35815039e-01 -5.86772740e-01 4.67524469e-01
2.74765164e-01 1.45556182e-02 -6.24598026e-01 -8.81349862e-01
2.58890510e-01 -9.15802181e-01 -3.79497856e-01 -1.10087104e-01
-2.61077434e-01 1.01470339e+00 8.35273981e-01 -1.31105363e+00
8.18990171e-01 -2.50602174e+00 2.56047964e-01 1.03268467e-01
-4.33166418e-03 3.07537675e-01 -2.02246517e-01 2.12698027e-01
-4.04731721e-01 -6.99320138e-01 -8.36342931e-01 -8.20763335e-02
-4.43352729e-01 -9.02028531e-02 -4.53415126e-01 9.22012866e-01
-2.25266457e-01 6.59543812e-01 -6.91549480e-01 -7.96370804e-02
4.57791328e-01 8.39184642e-01 -1.80492952e-01 4.52565610e-01
-1.18549177e-02 5.72947204e-01 1.02263883e-01 2.00062171e-01
1.55221975e+00 -1.79708421e-01 7.00166374e-02 -7.61388779e-01
-5.74095845e-01 -5.78371286e-01 -1.36535943e+00 1.51044178e+00
-6.61478043e-01 3.97856951e-01 8.53284180e-01 -5.13713300e-01
5.98028779e-01 2.29167596e-01 3.82666379e-01 -4.19788212e-01
-4.12894219e-01 2.45246455e-01 -3.51548612e-01 -2.18578875e-01
6.24288857e-01 -3.63165081e-01 5.14571667e-01 3.76490384e-01
-1.92060634e-01 -7.14942038e-01 -9.04620960e-02 1.08040534e-01
4.57136840e-01 -4.88602370e-02 2.56111830e-01 -3.89165550e-01
7.31385052e-01 -3.61935832e-02 -2.03435514e-02 7.74309993e-01
5.89754954e-02 1.09589827e+00 5.74222989e-02 4.56157848e-02
-1.22666287e+00 -1.18342471e+00 -2.82540262e-01 6.09657109e-01
2.55585283e-01 3.14221717e-02 -1.12590706e+00 4.24566530e-02
-1.11365721e-01 6.43423021e-01 -3.51830363e-01 2.73340970e-01
-3.01645726e-01 -1.19790030e+00 4.18987386e-02 -9.49416533e-02
8.81228685e-01 -4.40210223e-01 -5.42819262e-01 5.63693456e-02
-2.33303308e-01 -1.31603324e+00 -8.10405493e-01 -9.82759967e-02
-6.64212346e-01 -1.20661032e+00 -1.12439322e+00 -5.02386153e-01
7.91794062e-01 8.72771621e-01 5.42291105e-01 -4.39926505e-01
-3.75894576e-01 3.72746587e-01 -7.96202943e-03 4.21374232e-01
-4.37408686e-01 -7.15441346e-01 -1.00561917e-01 6.86707020e-01
-4.94590640e-01 -5.67610562e-01 -5.33351004e-01 3.42745245e-01
-1.22616184e+00 5.94742060e-01 8.58149290e-01 8.88881981e-01
4.87568408e-01 5.52829325e-01 -1.96074873e-01 -3.49510580e-01
3.91957313e-01 -4.00746092e-02 -1.25053358e+00 1.78964719e-01
-6.19484007e-01 -9.75282118e-02 4.93949324e-01 -4.09495175e-01
-1.61859882e+00 3.97372603e-01 4.02178109e-01 -4.86469924e-01
-1.29294217e-01 2.81741768e-01 5.94768450e-02 -4.94376391e-01
7.67357171e-01 5.20441294e-01 1.24217398e-01 -8.13328326e-01
9.29233670e-01 8.15460086e-01 1.23079562e+00 -2.88248390e-01
1.32523501e+00 6.71322286e-01 1.22294284e-01 -1.35493839e+00
-7.45881379e-01 -8.37000191e-01 -5.61741471e-01 -9.15344730e-02
1.01036382e+00 -1.18440306e+00 -2.73128301e-01 1.02331936e+00
-1.12421381e+00 -4.23172802e-01 -5.14130555e-02 4.02210504e-01
-5.10776043e-01 9.54726100e-01 -7.98890650e-01 -7.60386527e-01
-3.15419883e-01 -1.02566314e+00 1.00767279e+00 3.28692734e-01
2.97689259e-01 -6.53220057e-01 3.03676035e-02 4.42254841e-01
7.50316858e-01 -2.16212407e-01 5.33579350e-01 5.90545475e-01
-7.34572828e-01 7.71634877e-02 -9.45772588e-01 4.39810872e-01
1.57211140e-01 -2.01503232e-01 -1.09992588e+00 -3.77021492e-01
5.09986103e-01 3.28231812e-01 9.20255065e-01 9.31130469e-01
8.85378420e-01 -4.85661864e-01 -2.15191394e-02 9.86776114e-01
1.87098038e+00 -3.70114356e-01 7.37354755e-01 2.82045126e-01
8.57083559e-01 4.09136087e-01 5.57821333e-01 3.36062551e-01
-4.86596711e-02 9.61157143e-01 2.39373028e-01 -3.74783963e-01
-3.50422591e-01 2.07684875e-01 3.53887618e-01 2.47677788e-01
-1.37566715e-01 4.23991382e-01 -6.20788693e-01 6.44362628e-01
-1.76228416e+00 -9.76106644e-01 -7.05628455e-01 2.50285435e+00
1.08745110e+00 -7.29656994e-01 -2.06922024e-01 -4.05977443e-02
9.37051892e-01 3.09187800e-01 -1.84813678e-01 7.02578202e-02
-4.70788330e-01 1.74982980e-01 1.04531777e+00 1.15349364e+00
-1.12292159e+00 7.06186831e-01 5.99741459e+00 7.80532420e-01
-1.05130422e+00 2.47829348e-01 3.69146019e-01 2.82654285e-01
-4.92371526e-03 1.63216040e-01 -2.76451081e-01 4.46115822e-01
7.96389341e-01 -1.70209318e-01 1.21511877e+00 3.71158808e-01
3.78397644e-01 -6.82045639e-01 -4.01497871e-01 1.37300503e+00
2.00875886e-02 -1.14034224e+00 -1.93375543e-01 1.51920822e-02
1.12818646e+00 7.75737092e-02 1.21471368e-01 -7.31373310e-01
1.21023916e-01 -8.38191986e-01 6.60357475e-01 6.41186237e-01
9.82614160e-01 -4.48418558e-01 4.61489826e-01 1.93097413e-01
-1.04865086e+00 -7.19040539e-03 -2.88483918e-01 9.58164930e-02
3.24575722e-01 1.15171564e+00 -2.70813406e-01 8.47524703e-01
7.47962654e-01 4.37092453e-01 -5.02393842e-01 1.07735252e+00
-2.75132567e-01 1.49286672e-01 -1.87498420e-01 9.31702495e-01
7.29347467e-02 -1.12261844e+00 6.25675499e-01 1.29857099e+00
5.71236491e-01 3.70045722e-01 -1.45338207e-01 1.09292519e+00
3.48958880e-01 -3.19427401e-01 -1.89240009e-01 1.82309940e-01
1.04308590e-01 1.49258649e+00 -3.57917100e-01 -3.30936491e-01
-4.95258689e-01 1.47233319e+00 -3.46001744e-01 9.90669370e-01
-7.06112981e-01 -6.61313608e-02 7.76431978e-01 -1.23358838e-01
5.99565387e-01 -3.29148144e-01 -5.57684898e-01 -1.25091493e+00
-1.79395169e-01 -1.01746535e+00 1.86905712e-01 -1.37581384e+00
-1.07005835e+00 3.19961250e-01 -1.07140608e-01 -1.13358843e+00
1.85275406e-01 -4.18602705e-01 -3.85976523e-01 1.59214580e+00
-1.64496148e+00 -1.21440482e+00 -5.27635753e-01 7.35239148e-01
1.86526701e-01 3.95411313e-01 6.26676500e-01 -1.98618043e-02
-2.24012122e-01 -3.37299705e-01 5.60605943e-01 -3.77788872e-01
8.19014013e-01 -1.08396733e+00 2.77411491e-01 1.47684062e+00
-3.36749882e-01 4.47873771e-01 1.06151438e+00 -5.78695416e-01
-1.39240253e+00 -1.06623137e+00 5.10520041e-01 -8.46167430e-02
7.76360035e-01 1.84616134e-01 -1.01401186e+00 5.19382179e-01
2.69806325e-01 1.04383528e-01 4.39128242e-02 -6.79722726e-01
-2.84943789e-01 -1.64418936e-01 -1.03984201e+00 3.43229651e-01
2.46284306e-01 -8.37592423e-01 -4.40528691e-01 2.51452059e-01
3.88046920e-01 -6.01147115e-01 -5.20115852e-01 1.90524310e-01
3.12038869e-01 -1.19543874e+00 1.21526706e+00 3.63411129e-01
2.75672108e-01 -9.29163337e-01 -1.09839335e-01 -1.60036361e+00
-5.84162235e-01 -9.49218452e-01 -1.19312391e-01 9.00424242e-01
2.02141985e-01 -5.96389234e-01 2.98401088e-01 4.20940489e-01
7.51476968e-03 9.56461430e-02 -6.36815965e-01 -5.66711366e-01
-3.36968601e-01 -2.06707060e-01 2.23571554e-01 1.03837180e+00
-3.48632038e-01 1.18393749e-01 -6.53948605e-01 1.05844414e+00
1.25233567e+00 4.87393230e-01 5.23889601e-01 -7.87348688e-01
-5.34435093e-01 -2.39866450e-01 2.04174057e-01 -9.57913756e-01
-2.55193084e-01 -3.30839038e-01 3.06725025e-01 -1.36105585e+00
3.47230136e-01 -5.65524586e-02 2.19594121e-01 -4.31664921e-02
-3.04394633e-01 3.25109094e-01 1.31398216e-01 3.16869438e-01
1.15619518e-01 1.23044834e-01 1.20339608e+00 -4.54263948e-02
-2.15465099e-01 -1.87790856e-01 -4.76059854e-01 5.36531448e-01
3.57562870e-01 -1.00612640e-01 -1.20904092e-02 -5.17393231e-01
-3.49353882e-03 2.76882559e-01 7.87861824e-01 -9.01195347e-01
2.93018192e-01 -2.40077958e-01 4.82009321e-01 -3.81476462e-01
4.49435890e-01 -7.77454257e-01 7.52753913e-01 3.31566572e-01
1.86395153e-01 -5.64849198e-01 1.97051719e-01 4.61264223e-01
-9.07851383e-02 -1.34456351e-01 1.53080082e+00 -5.55051230e-02
-5.26560187e-01 9.55323353e-02 -3.76812130e-01 -4.33579594e-01
7.34194815e-01 -1.44172451e-02 -4.16048497e-01 -5.30058086e-01
-4.32143152e-01 -3.08628410e-01 9.90799427e-01 -2.13254720e-01
4.17287290e-01 -9.94132340e-01 -6.31978214e-01 4.52412635e-01
-2.50780702e-01 -1.46005109e-01 7.42807150e-01 1.22055590e+00
-8.48027885e-01 2.34408945e-01 -1.52960375e-01 -4.25912172e-01
-1.08867788e+00 7.34544337e-01 7.14983881e-01 7.25347623e-02
-5.86609006e-01 7.82725394e-01 5.67736030e-01 -6.64942637e-02
-2.25444347e-01 -1.76589727e-01 2.03481928e-01 -2.02743635e-01
9.19401944e-01 6.81536973e-01 4.19098698e-02 -9.27108109e-01
-1.10423453e-01 9.80804920e-01 3.24796706e-01 -4.44591105e-01
1.21353960e+00 -5.15288770e-01 -6.94598198e-01 -1.83343306e-01
1.09491718e+00 3.92674029e-01 -1.69338572e+00 -1.97331339e-01
-4.38279808e-01 -8.94063711e-01 6.41835749e-01 -8.08883131e-01
-1.07205069e+00 5.65413952e-01 7.07700849e-01 1.48483500e-01
1.54766321e+00 -1.66550517e-01 5.43792844e-01 -2.70235658e-01
-1.16774268e-01 -9.60611403e-01 -2.77168363e-01 2.82791644e-01
9.96938765e-01 -8.42668295e-01 2.71196663e-01 -7.46266901e-01
-5.38003623e-01 1.11909544e+00 -2.36889292e-02 8.67115483e-02
3.82175922e-01 1.36079416e-01 3.96153092e-01 -3.37922834e-02
-9.28349886e-03 -2.61534423e-01 4.16445494e-01 7.39772975e-01
1.31854191e-01 1.12332307e-01 -5.73358312e-02 -6.27258196e-02
1.35550171e-01 -5.47196157e-02 8.26519728e-01 1.86461627e-01
-4.51509893e-01 -7.27847695e-01 -1.29765761e+00 8.36808383e-02
-1.88217461e-01 -5.33501029e-01 -5.33804111e-03 7.45506808e-02
-2.92643718e-02 1.03761649e+00 -4.88810539e-02 3.23468715e-01
1.50201604e-01 -2.45742649e-01 5.95963001e-01 -1.61797732e-01
6.21105125e-03 4.79380488e-01 -1.25820503e-01 -5.09114861e-01
-3.77375841e-01 -7.53575087e-01 -7.98324943e-01 -2.69085884e-01
-3.72937709e-01 1.83183104e-01 5.49349725e-01 7.40911305e-01
1.57743618e-01 -1.49563566e-01 6.13082051e-01 -1.39288819e+00
-4.58363742e-01 -9.52442765e-01 -1.15990329e+00 2.72051066e-01
8.77794147e-01 -1.74859673e-01 -6.73110127e-01 3.53162020e-01] | [11.590957641601562, -2.7301619052886963] |
0c684c2d-8a78-4c06-977b-7fb3c743f0c3 | polyformer-referring-image-segmentation-as | 2302.07387 | null | https://arxiv.org/abs/2302.07387v2 | https://arxiv.org/pdf/2302.07387v2.pdf | PolyFormer: Referring Image Segmentation as Sequential Polygon Generation | In this work, instead of directly predicting the pixel-level segmentation masks, the problem of referring image segmentation is formulated as sequential polygon generation, and the predicted polygons can be later converted into segmentation masks. This is enabled by a new sequence-to-sequence framework, Polygon Transformer (PolyFormer), which takes a sequence of image patches and text query tokens as input, and outputs a sequence of polygon vertices autoregressively. For more accurate geometric localization, we propose a regression-based decoder, which predicts the precise floating-point coordinates directly, without any coordinate quantization error. In the experiments, PolyFormer outperforms the prior art by a clear margin, e.g., 5.40% and 4.52% absolute improvements on the challenging RefCOCO+ and RefCOCOg datasets. It also shows strong generalization ability when evaluated on the referring video segmentation task without fine-tuning, e.g., achieving competitive 61.5% J&F on the Ref-DAVIS17 dataset. | ['R. Manmatha', 'Vijay Mahadevan', 'Ravi Kumar Satzoda', 'Yuting Zhang', 'Zhaowei Cai', 'Hui Ding', 'Jiang Liu'] | 2023-02-14 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Liu_PolyFormer_Referring_Image_Segmentation_As_Sequential_Polygon_Generation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Liu_PolyFormer_Referring_Image_Segmentation_As_Sequential_Polygon_Generation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['referring-expression-segmentation', 'video-semantic-segmentation'] | ['computer-vision', 'computer-vision'] | [ 6.30758703e-01 1.17563955e-01 -1.10212877e-01 -2.31061429e-01
-1.19231606e+00 -7.99254298e-01 3.82784843e-01 -8.88999328e-02
-4.09812331e-01 3.89869690e-01 -2.76991844e-01 -3.19353700e-01
5.90075493e-01 -8.22998524e-01 -1.30279529e+00 -4.37560797e-01
2.92221397e-01 4.31659371e-01 4.43946928e-01 1.24557987e-01
4.58582371e-01 4.25896078e-01 -1.30038607e+00 3.94119292e-01
1.04143548e+00 1.47496080e+00 3.83069336e-01 9.06216323e-01
-5.36325388e-02 6.08967066e-01 -5.29666543e-01 -7.05025315e-01
3.22446734e-01 -1.72966927e-01 -6.71206057e-01 3.90159488e-01
7.05038786e-01 -3.76856327e-01 -3.42345327e-01 9.65682685e-01
2.49547958e-01 -1.60794884e-01 6.13355041e-01 -8.45211685e-01
-6.85295641e-01 5.90512633e-01 -9.32792425e-01 -9.41640064e-02
4.23335314e-01 2.62064010e-01 9.44719851e-01 -1.12257576e+00
6.38058662e-01 1.24980235e+00 5.32672107e-01 1.37809128e-01
-1.14512873e+00 -6.18043482e-01 3.10391933e-01 1.53470129e-01
-1.78995633e+00 -2.64408648e-01 4.60112989e-01 -4.64190781e-01
9.98383045e-01 2.86128879e-01 4.97923136e-01 7.61799812e-01
1.49317235e-01 9.76378918e-01 6.96880758e-01 -1.61107525e-01
2.34552294e-01 -3.19798946e-01 -1.76606715e-01 8.06098998e-01
-2.30973467e-01 -1.48953587e-01 -4.61535752e-01 2.11093858e-01
1.04581439e+00 -2.26407245e-01 -4.77105737e-01 -2.08866335e-02
-1.30013764e+00 4.96063977e-01 6.39111936e-01 -1.10843882e-01
-4.71580774e-01 5.17349422e-01 3.65754992e-01 -1.34244576e-01
4.62691545e-01 4.10256416e-01 -4.06621963e-01 -2.74684012e-01
-1.24788249e+00 2.02480301e-01 4.70954329e-01 1.58922780e+00
7.17866659e-01 5.73141687e-02 -3.68968487e-01 6.29408896e-01
3.15173358e-01 8.86466742e-01 1.77398473e-01 -1.17943203e+00
9.70025420e-01 3.90451819e-01 7.26687685e-02 -1.01333356e+00
-6.81376383e-02 -4.86472398e-01 -7.43506193e-01 -2.33486950e-01
3.76687318e-01 -8.36042091e-02 -1.24274683e+00 1.19814432e+00
1.80093572e-01 2.70089805e-01 -2.06791118e-01 1.02915740e+00
6.01823688e-01 9.94575977e-01 -6.49364516e-02 5.26898168e-02
1.34816563e+00 -1.28586078e+00 -4.86696154e-01 -3.79679799e-01
4.23150241e-01 -8.50406170e-01 1.11509895e+00 7.01966405e-01
-1.27281117e+00 -6.03634953e-01 -1.08801162e+00 -4.71001506e-01
-1.88975647e-01 4.56587017e-01 4.43157643e-01 3.78842562e-01
-1.16422427e+00 4.65895802e-01 -8.80129516e-01 2.24980906e-01
6.84135675e-01 4.01151597e-01 -2.47278363e-02 -8.67434517e-02
-7.09339380e-01 2.25260109e-01 3.36996049e-01 2.33852059e-01
-4.92439955e-01 -8.46580446e-01 -7.82540858e-01 2.02387497e-01
5.44043183e-01 -5.89172542e-01 1.20431018e+00 -9.59632099e-01
-1.59585905e+00 7.72735655e-01 -4.53046173e-01 -6.24889255e-01
7.35198200e-01 -4.21854377e-01 -7.83826131e-03 5.30242324e-01
1.42026469e-01 1.31060827e+00 1.10112166e+00 -1.09319508e+00
-9.57934976e-01 -2.77768046e-01 -1.69547752e-01 1.99884459e-01
1.63841128e-01 -2.07491532e-01 -1.51125360e+00 -8.87113035e-01
3.06575626e-01 -9.63828981e-01 -3.41779888e-01 1.88698649e-01
-6.56574845e-01 -1.55099362e-01 6.67750597e-01 -8.99940550e-01
1.17108595e+00 -2.23516297e+00 1.54604003e-01 3.36316705e-01
2.41064399e-01 1.38542205e-01 -2.53692809e-02 -1.06189415e-01
5.02742417e-02 3.83145928e-01 -3.16308200e-01 -4.40902621e-01
-1.43343955e-02 -9.17078927e-02 -4.67778116e-01 3.16116005e-01
3.67688239e-01 1.17508614e+00 -6.40368700e-01 -5.78401864e-01
3.20415914e-01 4.13556159e-01 -8.07549357e-01 -8.58326182e-02
-5.42740047e-01 2.07528338e-01 -4.45429355e-01 7.72195578e-01
8.15631688e-01 -4.47401345e-01 -1.31804710e-02 -3.19333136e-01
-1.00284725e-01 -7.51305297e-02 -7.55106509e-01 1.81009424e+00
-4.44953531e-01 7.69254684e-01 -1.36449382e-01 -6.79216266e-01
8.99437785e-01 2.40388121e-02 3.72067302e-01 -9.79966104e-01
1.08490519e-01 3.11201841e-01 -4.52913582e-01 -7.31169507e-02
9.07709539e-01 6.03208125e-01 -3.52178127e-01 -1.35383159e-01
-2.92137057e-01 -4.31414008e-01 2.22748324e-01 1.01966329e-01
9.40447509e-01 3.94347161e-01 -2.88158637e-02 4.10889275e-02
5.27305007e-01 2.44011760e-01 4.34505194e-01 5.82258105e-01
1.77510142e-01 1.19537258e+00 7.13009357e-01 6.12819055e-03
-1.13150358e+00 -1.12264693e+00 -3.11829057e-02 7.13168442e-01
4.14243072e-01 -5.30297816e-01 -1.01804364e+00 -3.92931730e-01
-1.60159200e-01 7.01071918e-01 -3.77870858e-01 2.44492486e-01
-9.00831997e-01 -3.19248915e-01 3.61884952e-01 8.03985775e-01
7.43552089e-01 -7.38446653e-01 -5.05387306e-01 2.89776862e-01
-9.77520049e-02 -1.55237269e+00 -9.96078491e-01 2.36752369e-02
-8.59311104e-01 -8.58135879e-01 -9.90258098e-01 -9.53278184e-01
7.48732448e-01 1.25075027e-01 1.03940082e+00 -3.19791697e-02
-1.28216013e-01 3.38322185e-02 -3.37748677e-01 1.10800810e-01
1.77224795e-03 2.26581529e-01 -6.82026088e-01 -6.11312948e-02
-8.16523954e-02 -2.37136424e-01 -7.82574594e-01 4.03363109e-01
-6.27488494e-01 4.54626262e-01 7.33667374e-01 7.37696111e-01
1.34639204e+00 -3.93250614e-01 1.39702395e-01 -7.65307724e-01
1.45452753e-01 -1.21853478e-01 -9.71499741e-01 1.55869380e-01
-3.29085320e-01 -1.35927603e-01 9.70763862e-01 -1.73503011e-01
-6.55489922e-01 4.73278522e-01 -3.03953052e-01 -8.31333280e-01
3.36152762e-02 1.73644602e-01 -1.75264269e-01 -2.30325442e-02
3.33248258e-01 3.96681190e-01 -4.11261678e-01 -3.39175284e-01
6.31964326e-01 6.81939900e-01 1.02240562e+00 -4.80290741e-01
6.49666190e-01 2.38728151e-01 -2.30209544e-01 -7.86370814e-01
-4.92827535e-01 -3.16040307e-01 -4.77519721e-01 -1.00569859e-01
1.15218771e+00 -1.15109134e+00 -6.56045139e-01 5.65115750e-01
-1.44632709e+00 -6.25709474e-01 -1.14515588e-01 7.22307991e-03
-7.72385418e-01 4.16832536e-01 -6.71754420e-01 -3.24736416e-01
-4.96269614e-01 -1.48665655e+00 1.58534122e+00 2.21743107e-01
-6.52022809e-02 -4.57776368e-01 -7.96925306e-01 3.95645887e-01
-7.30286865e-03 1.77425653e-01 7.31520414e-01 -1.41315967e-01
-1.19023693e+00 -1.38736650e-01 -7.62616456e-01 2.62466401e-01
-2.91240275e-01 1.18511207e-01 -6.78983152e-01 -4.13342752e-03
-4.26179379e-01 -1.16882650e-02 6.84097886e-01 3.28957558e-01
1.83591819e+00 -2.01882586e-01 -4.62404370e-01 1.03636348e+00
1.43433464e+00 4.84752268e-01 8.83862495e-01 1.45763485e-02
9.54247653e-01 6.47837594e-02 8.83403838e-01 4.31159794e-01
6.27749801e-01 8.45649958e-01 3.72460365e-01 -5.50526157e-02
-4.09900486e-01 -3.92783225e-01 8.52335840e-02 6.38782740e-01
9.41815525e-02 -6.00915909e-01 -1.05595791e+00 3.64008844e-01
-1.70490921e+00 -4.80204582e-01 -2.91328847e-01 1.90161133e+00
7.46676743e-01 2.32598707e-01 -1.89105302e-01 -1.05430640e-01
8.11676800e-01 1.13694511e-01 -7.49022841e-01 -2.78623670e-01
2.06005555e-02 3.38441640e-01 1.02118397e+00 5.86717367e-01
-1.17421818e+00 1.39885807e+00 5.85460949e+00 1.26095259e+00
-1.28036153e+00 -5.91960438e-02 1.15830374e+00 7.46722966e-02
-8.06562304e-02 -1.83766842e-01 -7.65699625e-01 6.45048976e-01
7.20763385e-01 8.45579877e-02 6.46228135e-01 7.47607708e-01
2.37186164e-01 -2.70022333e-01 -9.93138254e-01 1.43774652e+00
7.39636496e-02 -1.48092139e+00 -9.68290400e-03 -6.97403252e-02
7.85046518e-01 3.60219851e-02 2.32874289e-01 1.72017857e-01
-6.59587309e-02 -9.91744041e-01 1.24595237e+00 5.60091317e-01
1.30923104e+00 -7.10163713e-01 4.46210802e-01 2.32795596e-01
-1.18627656e+00 1.72337160e-01 -2.87314475e-01 3.09800446e-01
3.15477341e-01 4.37507629e-01 -7.47752905e-01 4.55688179e-01
7.15382397e-01 7.69358099e-01 -5.79197288e-01 1.01280284e+00
-3.46212775e-01 7.32690871e-01 -5.62051475e-01 6.06707782e-02
3.03130358e-01 -3.02539855e-01 2.62158886e-02 1.20552015e+00
4.48069274e-01 1.45636544e-01 1.10558689e-01 1.02265501e+00
-4.24604625e-01 3.18338424e-02 -2.87717115e-02 -3.17173116e-02
6.25267446e-01 1.06875432e+00 -9.58539367e-01 -5.00764608e-01
-2.96848416e-01 1.54678226e+00 1.45273164e-01 4.98777628e-01
-1.32325482e+00 -4.99493450e-01 1.86632425e-01 3.21120098e-02
9.59202230e-01 -3.34305346e-01 -7.35269964e-01 -9.82587159e-01
2.86401331e-01 -8.00855100e-01 -2.04999715e-01 -9.94954944e-01
-7.69292414e-01 5.98251104e-01 -3.08799893e-01 -1.25544572e+00
-2.07406223e-01 -5.06136835e-01 -2.58097827e-01 9.24003601e-01
-1.49170220e+00 -1.02267003e+00 -3.52127790e-01 2.74279028e-01
7.12544560e-01 1.68213800e-01 3.38455886e-01 3.11165392e-01
-5.18602133e-01 8.79565120e-01 9.92056206e-02 2.26543725e-01
4.30614740e-01 -1.20428312e+00 9.79907930e-01 8.75034451e-01
7.12035000e-02 1.98629603e-01 4.65023994e-01 -6.42480969e-01
-1.62325919e+00 -1.55536854e+00 7.35194564e-01 -1.38081849e-01
5.15874445e-01 -5.78644872e-01 -6.91674411e-01 5.38682401e-01
-4.60306369e-03 1.98107421e-01 -5.08891381e-02 -6.65806532e-01
-1.56404763e-01 7.02884942e-02 -9.91988003e-01 7.89934874e-01
1.30545342e+00 -3.30752224e-01 -3.30240838e-02 4.59272861e-01
9.64730084e-01 -1.12396312e+00 -9.42237079e-01 1.35652646e-01
2.62143254e-01 -7.63217747e-01 1.01636207e+00 1.44906163e-01
7.40044951e-01 -3.99831891e-01 -3.02896678e-01 -8.40739310e-01
-3.44797932e-02 -6.70559943e-01 -3.30340117e-02 1.09555471e+00
6.20531917e-01 -4.16795403e-01 1.01080418e+00 4.49097544e-01
-5.35935342e-01 -1.15360916e+00 -8.44462335e-01 -5.77868402e-01
-6.05297461e-02 -7.58275390e-01 5.77566445e-01 3.37586284e-01
-4.60234076e-01 6.03579171e-02 -1.46126881e-01 2.49175087e-01
2.62996286e-01 3.23609114e-01 7.55002320e-01 -4.10523295e-01
-2.99161553e-01 -5.10240614e-01 -4.02280331e-01 -2.06309533e+00
8.50719027e-03 -7.97416151e-01 2.98925877e-01 -1.40013087e+00
-3.75948250e-01 -4.50563341e-01 4.00381684e-01 3.54042441e-01
-2.12180153e-01 4.68246430e-01 4.76245344e-01 9.96331796e-02
-8.35597336e-01 5.54036498e-01 1.45017195e+00 -2.58417428e-01
-1.78872541e-01 -1.19300351e-01 -5.47010303e-01 6.54631138e-01
6.87483549e-01 -1.63028076e-01 -1.82371572e-01 -8.28870714e-01
3.85272154e-03 4.24453586e-01 4.47104663e-01 -1.16106522e+00
2.11393803e-01 1.00508898e-01 4.51604187e-01 -1.14805424e+00
5.12878239e-01 -6.09841287e-01 -5.57390265e-02 1.15578800e-01
-2.07902476e-01 2.60277778e-01 1.13363795e-01 5.91464937e-01
-2.26697519e-01 -4.07545157e-02 4.79208887e-01 1.44512907e-01
-8.57373357e-01 4.75989252e-01 -1.75777107e-01 2.47291341e-01
1.06349528e+00 -4.93095785e-01 -1.37243882e-01 -1.82612151e-01
-6.50595307e-01 4.61353034e-01 6.45561278e-01 2.93569475e-01
8.06905329e-01 -1.16438293e+00 -3.99564505e-01 1.19385026e-01
-1.86533973e-01 5.28445482e-01 2.06764102e-01 7.54958570e-01
-9.82596397e-01 7.02458799e-01 2.83068955e-01 -9.22399938e-01
-9.72704649e-01 4.02820617e-01 2.67388344e-01 -8.50373432e-02
-7.95955837e-01 1.01330900e+00 2.64544189e-01 1.75362527e-02
2.92449355e-01 -7.72297144e-01 2.47429013e-01 -2.29366004e-01
3.11638564e-01 1.48595840e-01 7.93609917e-02 -6.33548796e-01
-1.79166734e-01 8.37210000e-01 -9.80304852e-02 -1.17438398e-01
9.69666362e-01 -3.97781990e-02 1.75563321e-01 3.96881104e-02
1.26655173e+00 -8.58298168e-02 -1.63809955e+00 -1.25041634e-01
-5.19291423e-02 -5.63256145e-01 4.20667119e-02 -7.69954264e-01
-1.39597952e+00 9.34600830e-01 3.60610306e-01 -1.42691448e-01
1.14979649e+00 -6.19258918e-02 1.13384628e+00 2.48172611e-01
5.45642495e-01 -9.50963557e-01 -4.69876267e-02 6.23761296e-01
1.03588343e+00 -9.27022934e-01 -1.56747103e-01 -1.07124519e+00
-6.17861331e-01 1.25948524e+00 6.31388843e-01 -3.12094837e-01
2.61535883e-01 2.90899605e-01 -2.31882602e-01 5.10028983e-03
-5.58465540e-01 1.85594577e-02 2.45239198e-01 4.11227316e-01
3.53259385e-01 1.48153648e-01 -1.62656829e-01 5.98625064e-01
-5.25097966e-01 4.33593802e-02 3.43061358e-01 4.92856354e-01
-1.79333374e-01 -8.43621135e-01 -4.73511964e-01 6.27543330e-01
-2.30685622e-01 -3.40507269e-01 -1.36678830e-01 6.86998963e-01
1.10879004e-01 7.30017245e-01 4.19255763e-01 -3.59865397e-01
4.19237643e-01 -3.13934207e-01 3.06562752e-01 -5.30613840e-01
-4.59505945e-01 3.46390188e-01 -1.35775451e-02 -8.62453103e-01
5.51603809e-02 -7.47052491e-01 -1.55782175e+00 -1.16552770e-01
-3.32295150e-01 -1.17516056e-01 5.38293719e-01 6.75485611e-01
5.49451947e-01 7.51904905e-01 4.79543239e-01 -1.06429827e+00
-3.25004131e-01 -5.81249118e-01 -1.81714565e-01 1.48235530e-01
-4.15807776e-02 -2.31669039e-01 -3.30983661e-02 2.52906471e-01] | [9.32702350616455, -0.22387461364269257] |
1df23239-a0d0-4a85-a2ac-b01625374ea4 | composing-distributed-data-intensive-web | 1901.09894 | null | http://arxiv.org/abs/1901.09894v1 | http://arxiv.org/pdf/1901.09894v1.pdf | Composing Distributed Data-intensive Web Services Using a Flexible Memetic Algorithm | Web Service Composition (WSC) is a particularly promising application of Web
services, where multiple individual services with specific functionalities are
composed to accomplish a more complex task, which must fulfil functional
requirements and optimise Quality of Service (QoS) attributes, simultaneously.
Additionally, large quantities of data, produced by technological advances,
need to be exchanged between services. Data-intensive Web services, which
manipulate and deal with those data, are of great interest to implement
data-intensive processes, such as distributed Data-intensive Web Service
Composition (DWSC). Researchers have proposed Evolutionary Computing (EC)
fully-automated WSC techniques that meet all the above factors. Some of these
works employed Memetic Algorithms (MAs) to enhance the performance of EC
through increasing its exploitation ability of in searching neighbourhood area
of a solution. However, those works are not efficient or effective. This paper
proposes an MA-based approach to solving the problem of distributed DWSC in an
effective and efficient manner. In particular, we develop an MA that hybridises
EC with a flexible local search technique incorporating distance of services.
An evaluation using benchmark datasets is carried out, comparing existing
state-of-the-art methods. Results show that our proposed method has the highest
quality and an acceptable execution time overall. | ['Soheila Sadeghiram', 'Hui Ma', 'Gang Chen'] | 2019-01-26 | null | null | null | null | ['service-composition'] | ['miscellaneous'] | [ 2.33995512e-01 -7.13445127e-01 2.47248486e-01 -4.18247789e-01
-3.55170041e-01 -3.69235516e-01 5.38680434e-01 4.79330532e-02
-4.16975498e-01 6.06483877e-01 -2.11964902e-02 -8.42630677e-03
-7.20624447e-01 -9.58471596e-01 4.99014976e-03 -1.05700493e+00
-1.36182263e-01 7.59160399e-01 5.43438554e-01 -3.90989155e-01
8.01311851e-01 6.80759072e-01 -2.35722971e+00 2.71834224e-01
1.17664015e+00 1.12934315e+00 5.93422234e-01 1.75611958e-01
-7.93606222e-01 -1.49746984e-01 -5.39098859e-01 -2.11997718e-01
2.49561042e-01 -7.28699803e-01 -5.19598305e-01 3.09857577e-01
-9.16129649e-01 4.79961812e-01 5.49669623e-01 1.18769491e+00
5.41765451e-01 3.69253337e-01 3.59595358e-01 -1.46641624e+00
-1.97902352e-01 4.64016557e-01 -5.45917988e-01 1.23386383e-01
3.68841976e-01 3.28466035e-02 6.46112621e-01 -5.85457623e-01
6.30855441e-01 1.00383770e+00 1.06841967e-01 5.40011406e-01
-7.56763101e-01 -3.25829059e-01 1.60728395e-01 7.20487773e-01
-1.37647641e+00 -3.73524725e-01 8.05840015e-01 2.35886052e-01
1.08357799e+00 8.00710797e-01 8.66402686e-01 4.02562827e-01
2.74389647e-02 3.91331047e-01 1.04633605e+00 -5.59000432e-01
8.58350575e-01 2.93577760e-01 -3.50617141e-01 1.34067714e-01
3.81943107e-01 -2.18656927e-01 -3.07770252e-01 -3.85971397e-01
-2.58298546e-01 1.37009233e-01 -1.84250727e-01 -3.87217373e-01
-6.65617347e-01 6.93071246e-01 -4.09097271e-03 8.31794679e-01
-7.75176883e-01 -4.33950722e-01 3.85264367e-01 2.16941535e-01
8.38776976e-02 3.43715340e-01 -5.36518931e-01 -6.10584974e-01
-4.23875511e-01 1.98734775e-01 1.03702116e+00 9.28200960e-01
2.40134761e-01 -4.95298058e-02 5.02112210e-01 7.51294971e-01
3.38213533e-01 1.31536499e-01 8.48148048e-01 -5.79682827e-01
1.07571796e-01 1.23443317e+00 -4.39110138e-02 -1.00023401e+00
-2.50786722e-01 -3.48238677e-01 -7.28225052e-01 2.07663208e-01
-2.71254808e-01 2.41812930e-01 -4.02525067e-01 1.14922535e+00
8.40947926e-01 5.04987035e-03 3.75761986e-01 1.10405564e+00
3.20652366e-01 8.13736796e-01 1.81329265e-01 -6.61022782e-01
1.39843309e+00 -7.10317552e-01 -7.92312562e-01 3.65592897e-01
2.14634567e-01 -8.93185019e-01 7.19515026e-01 5.24342656e-01
-9.74357247e-01 6.36809468e-02 -9.15443480e-01 6.01086020e-01
-7.08209097e-01 -5.19829273e-01 7.98697770e-01 8.04569185e-01
-8.72260630e-01 4.86419499e-01 -5.51434994e-01 -6.58825159e-01
1.13741376e-01 4.19860870e-01 -2.74189580e-02 -7.44450241e-02
-8.75631511e-01 7.83559203e-01 4.14000124e-01 8.42274278e-02
-4.42061245e-01 -2.91820794e-01 -2.16046825e-01 2.64335811e-01
7.40181625e-01 -2.71687955e-01 6.71186149e-01 -1.14604652e+00
-1.44348645e+00 3.34203482e-01 3.23350467e-02 -2.50739098e-01
3.44767600e-01 7.03157425e-01 -1.02465999e+00 -5.78562506e-02
-3.82708997e-01 -1.54885456e-01 3.50890607e-01 -1.38786852e+00
-1.29169476e+00 -7.03354955e-01 -1.77105397e-01 3.43553483e-01
-5.47205985e-01 5.57491899e-01 -3.51612657e-01 -1.42948836e-01
3.83256495e-01 -8.01254213e-01 -3.71548444e-01 -3.35111290e-01
2.02902451e-01 -3.60066414e-01 9.59453046e-01 -3.92462820e-01
1.39087641e+00 -1.84951723e+00 3.81295860e-01 8.00943792e-01
-2.28219897e-01 4.29471076e-01 1.16468938e-02 7.64682651e-01
4.32259500e-01 9.81834158e-02 -6.77043676e-01 1.32308736e-01
1.32546842e-01 4.79779452e-01 5.54401457e-01 2.65547991e-01
-9.12453532e-02 8.55664089e-02 -4.16645437e-01 -6.51438057e-01
-7.34083951e-02 4.28409874e-01 -5.40364802e-01 7.74530768e-02
-3.64861369e-01 2.14929089e-01 -6.87273502e-01 9.91777062e-01
7.20218837e-01 5.28813004e-02 6.09909415e-01 1.04572386e-01
-4.95345205e-01 -2.73327500e-01 -1.65375829e+00 1.44266272e+00
-5.50051630e-01 -2.10423753e-01 3.51146430e-01 -1.36782742e+00
1.08425903e+00 6.19323432e-01 1.01502287e+00 -8.90224218e-01
3.60725254e-01 7.77350307e-01 4.12391782e-01 -7.22268701e-01
1.56416982e-01 1.60864577e-01 2.22980991e-01 6.57621741e-01
-4.90921706e-01 7.62218237e-02 4.96787220e-01 -1.79913566e-01
1.06070864e+00 1.64683759e-01 3.24691951e-01 -3.83500814e-01
1.19007647e+00 9.83113497e-02 8.13229680e-01 6.12011086e-03
-1.69028237e-01 6.95652468e-03 1.03036083e-01 -4.06698346e-01
-1.10635746e+00 -3.52385223e-01 7.93335959e-02 7.21247971e-01
3.39043677e-01 -9.64085460e-02 -6.07591450e-01 -4.64561164e-01
-3.95262577e-02 7.76916504e-01 1.01225311e-02 -1.06805079e-01
-4.53701526e-01 -8.39869797e-01 2.18573119e-02 -1.60054028e-01
4.66046780e-01 -1.27763450e+00 -1.08731747e+00 6.64275348e-01
7.98706710e-02 -7.22104847e-01 4.80086217e-03 -1.66216865e-01
-8.23396206e-01 -8.34036171e-01 -3.04765344e-01 -6.95215702e-01
6.54235005e-01 4.00178194e-01 7.91011274e-01 5.85118830e-01
-4.91943568e-01 -1.06512562e-04 -8.91722441e-01 -3.86157364e-01
-4.45607364e-01 1.32055923e-01 -5.16786349e-05 4.74226832e-01
7.15875208e-01 -9.66507316e-01 -4.54436183e-01 5.91450155e-01
-1.22868550e+00 -1.28208861e-01 5.86678147e-01 4.46735054e-01
4.42442715e-01 7.39798665e-01 9.24026132e-01 -5.13293743e-01
8.12377334e-01 -7.90348530e-01 -8.96277368e-01 4.62636918e-01
-1.13801980e+00 1.36405766e-01 8.15862119e-01 -2.95743406e-01
-1.28464711e+00 -7.08998367e-02 -2.42839336e-01 2.30168447e-01
-8.01078305e-02 4.88146514e-01 -6.18793905e-01 -1.01596683e-01
2.95461938e-02 4.42660779e-01 3.12863857e-01 -7.92829394e-01
-2.00077802e-01 1.17789614e+00 -9.73261148e-02 -7.04585016e-01
5.19753516e-01 3.79568219e-01 1.38775021e-01 -2.55193174e-01
1.86973661e-01 -7.12426901e-01 -2.14231517e-02 -3.95419478e-01
5.32934964e-01 4.40316796e-02 -9.97570395e-01 2.11692512e-01
-7.64030993e-01 6.26636684e-01 2.15231374e-01 3.20975184e-01
-2.15982184e-01 5.60887642e-02 -8.24045390e-03 -1.22908247e+00
-4.73874569e-01 -1.44982696e+00 5.17833054e-01 4.47614580e-01
1.39699221e-01 -5.27298152e-01 -1.49309322e-01 2.53345281e-01
1.07199097e+00 2.96556413e-01 9.98718202e-01 -9.42121208e-01
-6.14557683e-01 -2.51155674e-01 9.71667245e-02 -6.96193725e-02
2.61968940e-01 1.45059749e-01 -1.72435284e-01 -1.53508127e-01
4.35940437e-02 3.10726315e-01 -1.43825054e-01 -2.37159222e-01
1.16289544e+00 -2.84762979e-01 -2.44157150e-01 3.56399119e-01
1.92532539e+00 1.08983946e+00 5.78568220e-01 7.66378522e-01
5.29030673e-02 9.02387619e-01 8.72984529e-01 1.06057286e+00
2.09513769e-01 8.93258095e-01 6.52783811e-01 3.65578353e-01
2.98977524e-01 4.51991469e-01 -2.17425134e-02 9.83712614e-01
-4.93284613e-01 -3.74563754e-01 -6.98554158e-01 4.60052669e-01
-2.00754428e+00 -9.41430151e-01 -2.20571175e-01 1.89023602e+00
7.44058311e-01 -1.21741742e-01 4.08597998e-02 6.51248991e-01
8.28981161e-01 -3.47732812e-01 -6.64187133e-01 -8.13616872e-01
-6.66657686e-02 1.54758364e-01 2.33471528e-01 -1.28918871e-01
-3.53507578e-01 4.24964935e-01 3.68341112e+00 7.53093302e-01
-9.13258731e-01 2.18368679e-01 8.59370679e-02 -1.60892755e-01
-7.69701004e-01 -2.46860255e-02 -4.34637219e-01 9.27204013e-01
1.13047433e+00 -4.41415429e-01 9.31004703e-01 8.84694636e-01
3.51990581e-01 -2.07741499e-01 -3.96419108e-01 7.35536635e-01
-1.02060428e-03 -1.14189517e+00 -1.28098533e-01 2.26447284e-01
1.02101862e+00 -4.14407879e-01 -3.69334906e-01 -1.53458670e-01
1.82125662e-02 -3.75471950e-01 4.31404859e-01 6.14118755e-01
1.35015830e-01 -1.17306447e+00 1.04902828e+00 3.61204028e-01
-1.16247964e+00 -4.56966937e-01 -2.77379662e-01 3.24022144e-01
3.74522746e-01 3.91057819e-01 -2.08710924e-01 9.37285185e-01
1.03147888e+00 -1.25235558e-01 -2.62504339e-01 1.39191306e+00
4.56901163e-01 -1.20693706e-02 -3.48820090e-01 -8.04374397e-01
3.61165345e-01 -7.18059361e-01 6.14386022e-01 6.88603163e-01
7.21382380e-01 3.31566572e-01 -6.24152198e-02 4.51566160e-01
3.42856526e-01 9.26211655e-01 -2.38349989e-01 -2.20841274e-01
6.72803700e-01 1.22647345e+00 -9.44471776e-01 -2.61055648e-01
-4.21500742e-01 6.98925734e-01 -1.87215909e-01 -8.34474936e-02
-7.92721510e-01 -5.38369715e-01 8.34224343e-01 -7.63881719e-03
1.51563615e-01 -4.77861501e-02 -2.38088503e-01 -6.40879095e-01
1.70515046e-01 -1.25500846e+00 5.31443894e-01 -3.25682640e-01
-1.09469223e+00 8.00651252e-01 -2.51893818e-01 -1.23172629e+00
-1.92149341e-01 -1.63652003e-01 -2.71055877e-01 7.31180131e-01
-1.39260387e+00 -7.22145796e-01 -4.23093587e-01 6.48443162e-01
7.74312317e-01 -6.16335869e-01 8.31214488e-01 6.18896723e-01
-3.99313569e-01 4.05607931e-02 4.67004359e-01 -9.31813240e-01
1.80822313e-01 -7.22609282e-01 -1.25392094e-01 9.74091172e-01
-2.60331601e-01 4.98866022e-01 7.77070343e-01 -5.51997602e-01
-1.98616457e+00 -4.50080514e-01 9.60361540e-01 4.82387871e-01
4.54942226e-01 -6.03403114e-02 -7.24270344e-01 -1.78080767e-01
1.18710838e-01 -2.08972529e-01 6.52276635e-01 -3.66449744e-01
3.02626103e-01 -5.49464405e-01 -1.83264983e+00 5.41587055e-01
1.21434772e+00 3.51665944e-01 -2.30601296e-01 1.39443755e-01
4.58235800e-01 1.90813735e-01 -7.84725428e-01 2.86978364e-01
5.37090123e-01 -1.01843417e+00 6.01432741e-01 -5.37560821e-01
1.77206099e-02 -5.07354915e-01 -3.63400102e-01 -1.10775459e+00
2.99701672e-02 -4.35197711e-01 4.10174876e-02 1.46911454e+00
3.73054266e-01 -9.79929388e-01 5.95568776e-01 8.65129769e-01
-1.14803046e-01 -7.64431536e-01 -9.48198438e-01 -7.11001992e-01
-7.71833479e-01 -1.16178170e-01 1.60116434e+00 8.82985055e-01
-3.59334871e-02 -4.53451902e-01 4.13470566e-02 4.35877964e-02
5.66907644e-01 2.30150521e-01 2.69371092e-01 -1.54130733e+00
-1.79261461e-01 -7.16543794e-01 -4.43467051e-01 2.49030545e-01
-1.60805076e-01 -6.90121889e-01 -7.46354386e-02 -1.72638035e+00
-7.29384348e-02 -8.20688963e-01 -4.30488020e-01 1.29520163e-01
2.37331703e-01 -1.01537518e-01 1.94064319e-01 2.40179837e-01
-4.93811548e-01 4.40128207e-01 9.16303992e-01 2.22877249e-01
-3.13397259e-01 2.46587828e-01 -5.64551651e-01 2.87414789e-01
9.27643716e-01 -7.46088207e-01 -5.25916994e-01 -6.21450916e-02
2.36100331e-01 -3.14003080e-02 -3.39438736e-01 -8.01897645e-01
4.61775362e-01 -6.46477997e-01 -1.34710059e-01 -3.00651908e-01
7.54354801e-03 -1.56864226e+00 9.22923207e-01 7.03945756e-01
-2.07810313e-01 5.02759933e-01 -2.79341787e-01 3.05692255e-01
-2.56841719e-01 -6.68230534e-01 6.89390779e-01 -3.08734506e-01
-9.47032630e-01 1.13422014e-01 -4.74514127e-01 -5.43297827e-01
1.52920365e+00 -3.20287347e-01 9.55441520e-02 -2.13527642e-02
-3.56195688e-01 1.67304963e-01 4.51968282e-01 6.05734050e-01
5.46052814e-01 -1.08719540e+00 -5.24661899e-01 8.82770345e-02
1.30949244e-01 -4.57817793e-01 2.07621157e-01 7.12242723e-01
-7.33947992e-01 2.74535149e-01 -4.80145752e-01 -1.98254466e-01
-1.31675994e+00 6.42144978e-01 -2.88725980e-02 6.34799153e-02
-5.04521370e-01 3.28150541e-01 -9.51891363e-01 -4.65218663e-01
1.87736094e-01 3.76130700e-01 -4.13189918e-01 1.06878228e-01
5.32407820e-01 6.69033170e-01 4.20674860e-01 -4.51039940e-01
-7.24660933e-01 3.71282846e-01 5.01867533e-01 -2.55135596e-01
1.65615940e+00 -2.29888320e-01 -5.93522549e-01 -1.70269147e-01
9.93927419e-01 -1.89386025e-01 -5.97276688e-01 -6.23676330e-02
5.54588079e-01 -7.51252055e-01 -2.62963790e-02 -1.14839172e+00
-1.26869798e+00 2.58452117e-01 7.62686849e-01 5.06547034e-01
1.66263688e+00 -3.54061395e-01 7.46795058e-01 -7.84788467e-03
7.30209053e-01 -1.51146305e+00 -5.67726851e-01 -2.35753179e-01
7.04611123e-01 -8.74114275e-01 -9.63480100e-02 -1.69209585e-01
-6.75780296e-01 1.09866571e+00 5.13864338e-01 2.93131977e-01
4.65210170e-01 4.72259253e-01 -2.40961894e-01 -5.10811321e-02
-9.67096925e-01 -3.65202963e-01 -6.96519390e-02 5.32405019e-01
-5.06892800e-02 -6.70485646e-02 -1.44031847e+00 5.74079633e-01
2.66623288e-01 -2.06993550e-01 2.07028925e-01 1.38642251e+00
-5.70689261e-01 -1.48487067e+00 -4.64127719e-01 2.86387801e-01
-4.65256155e-01 1.98275119e-01 -9.75981057e-02 5.59606493e-01
2.67614722e-01 1.25265932e+00 -3.93763781e-02 -1.20017573e-01
2.26740360e-01 3.65880914e-02 -2.19001640e-02 3.42129916e-02
-7.49428213e-01 1.19075641e-01 2.23064363e-01 -5.27834475e-01
-7.21642554e-01 -9.10479486e-01 -1.52780461e+00 -4.06356692e-01
-2.48061910e-01 7.86158383e-01 1.78602111e+00 8.98708045e-01
6.19767368e-01 6.18043184e-01 1.07539046e+00 -3.04116279e-01
-5.46558022e-01 -5.99495530e-01 -5.60279667e-01 4.26276445e-01
-5.87211668e-01 -6.51764870e-01 -4.03623432e-01 -2.13948891e-01] | [8.585481643676758, 6.933276653289795] |
0d06e045-331f-43c3-a8b9-e99244694ad3 | tata-a-multilingual-table-to-text-dataset-for | 2211.00142 | null | https://arxiv.org/abs/2211.00142v1 | https://arxiv.org/pdf/2211.00142v1.pdf | TaTa: A Multilingual Table-to-Text Dataset for African Languages | Existing data-to-text generation datasets are mostly limited to English. To address this lack of data, we create Table-to-Text in African languages (TaTa), the first large multilingual table-to-text dataset with a focus on African languages. We created TaTa by transcribing figures and accompanying text in bilingual reports by the Demographic and Health Surveys Program, followed by professional translation to make the dataset fully parallel. TaTa includes 8,700 examples in nine languages including four African languages (Hausa, Igbo, Swahili, and Yor\`ub\'a) and a zero-shot test language (Russian). We additionally release screenshots of the original figures for future research on multilingual multi-modal approaches. Through an in-depth human evaluation, we show that TaTa is challenging for current models and that less than half the outputs from an mT5-XXL-based model are understandable and attributable to the source data. We further demonstrate that existing metrics perform poorly for TaTa and introduce learned metrics that achieve a high correlation with human judgments. We release all data and annotations at https://github.com/google-research/url-nlp. | ['Clara Rivera', 'Ankur Parikh', 'Michael Chavinda', 'Jan A. Botha', 'Vitaly Nikolaev', 'Sebastian Ruder', 'Sebastian Gehrmann'] | 2022-10-31 | null | null | null | null | ['data-to-text-generation'] | ['natural-language-processing'] | [-6.15320802e-02 4.07469898e-01 -3.78166765e-01 -4.47985828e-01
-1.47370291e+00 -8.86678278e-01 8.43106925e-01 3.10620993e-01
-1.89058214e-01 1.17443264e+00 8.12892497e-01 -7.92964578e-01
4.25755411e-01 -7.02257037e-01 -7.89079547e-01 7.13600516e-02
3.32570970e-01 9.06419873e-01 -4.26259309e-01 -4.92450655e-01
-1.27209917e-01 -1.52897209e-01 -7.82250285e-01 7.35034823e-01
1.17476070e+00 1.04168996e-01 -9.68231112e-02 5.82975328e-01
-1.38150737e-01 6.29450738e-01 -6.52988970e-01 -9.50295568e-01
2.35770121e-01 -8.51074278e-01 -9.51949179e-01 -3.02761614e-01
3.97057265e-01 -2.22889751e-01 -2.23378882e-01 6.67352855e-01
6.95438921e-01 -3.90099049e-01 8.47514629e-01 -1.15250647e+00
-1.23698080e+00 1.22315037e+00 -5.77093005e-01 1.31266773e-01
7.83740222e-01 1.99798435e-01 1.01790357e+00 -1.23912930e+00
1.16430628e+00 1.49026942e+00 7.84531415e-01 4.89344001e-01
-1.15675652e+00 -8.43677402e-01 -9.50321034e-02 -2.37395704e-01
-1.35569561e+00 -8.34967732e-01 2.19925046e-01 -5.42624414e-01
1.07511461e+00 2.60591507e-01 4.35481399e-01 1.52862704e+00
-1.42001724e-02 6.28691316e-01 1.43012881e+00 -5.23373783e-01
-3.24833542e-01 1.13694102e-01 -3.52766305e-01 6.85492456e-01
3.61321181e-01 -3.35057914e-01 -7.12078094e-01 -2.32762247e-01
3.71628284e-01 -5.51555336e-01 -7.72448108e-02 2.56458938e-01
-1.79598939e+00 8.72885764e-01 3.49794738e-02 1.19332880e-01
-4.48205322e-02 -2.51884818e-01 3.03677708e-01 3.36370289e-01
6.80594325e-01 3.24349523e-01 -4.70366776e-01 -2.96307266e-01
-8.70727658e-01 4.53566402e-01 9.13681924e-01 1.34216547e+00
5.57401776e-01 1.75085217e-02 -1.10169135e-01 1.04038608e+00
2.43818089e-01 9.52620327e-01 2.64042616e-01 -7.87809372e-01
1.06267941e+00 4.94496495e-01 2.67825007e-01 -8.84688199e-01
-4.17232424e-01 9.09739174e-03 -7.12375820e-01 -3.77869308e-01
7.32744694e-01 -4.96234834e-01 -9.02842104e-01 1.71969426e+00
2.19355464e-01 -8.40351224e-01 2.96011090e-01 6.07410610e-01
1.02714586e+00 7.71125376e-01 8.81998241e-02 -6.59013689e-02
1.49716616e+00 -8.11085939e-01 -7.45108545e-01 -3.43013048e-01
8.81039083e-01 -1.14072800e+00 1.24171627e+00 9.33130644e-03
-1.07097912e+00 -2.56979465e-01 -7.51500547e-01 -4.79690373e-01
-8.07973921e-01 4.04510736e-01 4.15054798e-01 6.03354037e-01
-1.09486294e+00 6.69053644e-02 -7.16321111e-01 -9.34318006e-01
3.76456827e-01 -1.75457567e-01 -5.79414248e-01 -1.86762303e-01
-1.52548814e+00 1.04899812e+00 3.89344186e-01 -1.29371092e-01
-6.28062010e-01 -6.21077836e-01 -1.11088526e+00 -6.45252943e-01
1.41102865e-01 -6.25070989e-01 1.20573783e+00 -5.60713053e-01
-1.01901865e+00 1.15497291e+00 -3.71756941e-01 -2.38728642e-01
7.23031580e-01 -2.24017650e-01 -6.63168728e-01 -2.57842038e-02
7.83014953e-01 8.39258194e-01 2.29529455e-01 -1.13975155e+00
-5.79779029e-01 -2.64704287e-01 -1.29807845e-01 2.10339248e-01
1.89402625e-02 4.06439990e-01 -3.62281531e-01 -1.12165737e+00
-2.39637151e-01 -8.39605510e-01 -7.87805617e-02 -4.06931728e-01
-6.72514915e-01 7.06564868e-03 2.40751743e-01 -1.36695707e+00
1.44192207e+00 -1.58652675e+00 -1.16665632e-01 -9.53909457e-02
3.80601399e-02 -1.92113519e-01 -1.97983876e-01 1.00306702e+00
1.61383878e-02 5.98491549e-01 -4.83698517e-01 -2.10719168e-01
5.27244322e-02 7.75253326e-02 -3.27065378e-01 3.03087175e-01
3.09022248e-01 1.14055920e+00 -1.13399696e+00 -6.52009010e-01
-1.23571426e-01 3.25761616e-01 -4.22671139e-01 -1.88349754e-01
-8.55929330e-02 5.21649241e-01 -1.11518934e-01 1.04662716e+00
3.69566470e-01 -9.50494781e-03 4.39759582e-01 -9.86959785e-02
-2.84561694e-01 6.36764109e-01 -7.04785705e-01 1.71831369e+00
-4.75527287e-01 7.97650695e-01 -1.01098470e-01 -2.53931582e-01
5.98810732e-01 3.21629196e-01 1.60582080e-01 -8.01694214e-01
-1.11985974e-01 6.86554909e-01 8.16024840e-02 -3.59095067e-01
7.75106490e-01 -2.93948818e-02 -6.51974976e-01 6.83273554e-01
-1.59208961e-02 -3.00738037e-01 6.67252898e-01 5.82050800e-01
7.17127264e-01 3.31020385e-01 4.24475729e-01 -3.25608909e-01
2.70905364e-02 5.32327831e-01 5.99732280e-01 6.93521082e-01
4.47986796e-02 7.84813702e-01 4.18488383e-01 -2.65672296e-01
-1.48074627e+00 -1.01710641e+00 -2.63143539e-01 1.11193597e+00
-5.31013906e-01 -8.07591558e-01 -7.83845425e-01 -7.02381015e-01
-9.51079726e-02 1.03380299e+00 -7.17348039e-01 4.29408044e-01
-5.10756969e-01 -9.53633726e-01 1.09510410e+00 3.06309551e-01
1.50288463e-01 -8.63168597e-01 -3.47856551e-01 1.56839207e-01
-8.85564327e-01 -1.19285142e+00 -6.34007156e-01 -1.95293769e-01
-4.41574186e-01 -9.73226905e-01 -9.56288099e-01 -6.30581081e-01
6.31627083e-01 -2.33267546e-01 1.29418373e+00 -4.14498389e-01
-5.25305681e-02 1.13424763e-01 -4.64881063e-01 -4.84633684e-01
-9.49961245e-01 3.52205694e-01 9.32747275e-02 -5.66533327e-01
4.22498226e-01 -1.98142290e-01 -2.33278766e-01 7.96997845e-02
-7.19353199e-01 7.43283093e-01 4.84870970e-01 6.59491301e-01
4.35520977e-01 -5.94089031e-01 7.31723011e-01 -1.19015706e+00
7.71776438e-01 -8.30259800e-01 -1.99002326e-01 4.00822371e-01
-3.78958613e-01 -1.63858756e-01 5.69897592e-01 -7.91269541e-02
-9.42750096e-01 -2.02488780e-01 -5.66068403e-02 2.18637407e-01
-2.93956674e-03 8.55473578e-01 -8.57490152e-02 7.27483928e-01
8.13361764e-01 1.12123108e-02 -2.77335733e-01 -4.10080940e-01
7.25773931e-01 9.91708219e-01 5.84252238e-01 -6.79619670e-01
8.33267510e-01 2.84239620e-01 -5.70017338e-01 -7.32269645e-01
-7.55492568e-01 -2.62915716e-02 -7.22022235e-01 -1.49744675e-01
7.03105927e-01 -1.38763583e+00 -8.33906308e-02 4.74718690e-01
-1.04070568e+00 -7.28789210e-01 -5.67911156e-02 4.46153879e-01
-3.22305202e-01 -1.26407882e-02 -8.06391954e-01 -5.36323607e-01
-4.85635042e-01 -8.98557723e-01 1.07895792e+00 -7.30579570e-02
-7.55133390e-01 -1.03985012e+00 2.25751907e-01 5.88560164e-01
2.60786206e-01 5.41340232e-01 1.05881369e+00 -7.07807064e-01
3.86704574e-03 2.05363691e-01 -2.87906915e-01 -2.32986003e-01
4.76611018e-01 2.29954690e-01 -6.32911563e-01 -2.34121889e-01
-5.05099714e-01 -7.58964717e-01 6.17093980e-01 6.03960194e-02
3.50335568e-01 -6.44635320e-01 -2.18102843e-01 3.86446565e-01
1.10861695e+00 5.92456199e-02 3.02112371e-01 2.81563073e-01
8.45321298e-01 7.50079513e-01 5.54916322e-01 3.36838305e-01
1.21138370e+00 5.15277028e-01 -3.83970767e-01 -1.37225285e-01
-3.37867707e-01 -8.26539040e-01 6.12016559e-01 1.50897288e+00
1.73588485e-01 -4.05326605e-01 -1.43583882e+00 1.08702862e+00
-1.61473417e+00 -7.63619125e-01 -3.55141073e-01 1.91606212e+00
1.39445674e+00 -1.08877324e-01 1.78827554e-01 -4.06012982e-01
4.47617173e-01 1.77441195e-01 -2.31066957e-01 -5.03401995e-01
-5.46917677e-01 1.45287439e-01 4.95470643e-01 7.78534114e-01
-9.30173635e-01 1.20352447e+00 6.82509136e+00 7.75595248e-01
-6.80707335e-01 2.60180831e-01 8.32760930e-01 -5.84365651e-02
-8.31891537e-01 2.02190652e-01 -6.74094439e-01 4.41645890e-01
1.34031749e+00 -4.35293615e-01 4.39647198e-01 3.74482810e-01
4.69955802e-01 -8.10746662e-03 -9.68600392e-01 6.80005193e-01
2.41124183e-01 -1.20792270e+00 4.71656322e-01 3.83962952e-02
1.00212276e+00 2.51272142e-01 -8.88884738e-02 3.83280218e-01
8.15124869e-01 -1.21794236e+00 1.06693113e+00 3.33463192e-01
1.46792579e+00 -6.04687989e-01 4.14415240e-01 9.04543027e-02
-1.06704235e+00 3.85348439e-01 -2.27078557e-01 6.15943670e-02
4.70826060e-01 4.41128194e-01 -9.56261396e-01 7.97337174e-01
6.14199698e-01 7.66380548e-01 -8.90574276e-01 2.41676271e-01
-3.53269964e-01 6.95052803e-01 -3.24677974e-01 1.86421290e-01
2.20053360e-01 -1.35588974e-01 5.02550423e-01 1.56629574e+00
5.07860839e-01 1.14506125e-01 1.68132484e-01 8.62275064e-01
-4.88090545e-01 4.31449145e-01 -9.65201557e-01 -4.48444575e-01
5.80373347e-01 1.14975643e+00 -7.35409677e-01 -4.36131030e-01
-5.45693457e-01 9.07791793e-01 4.00137305e-01 4.23493356e-01
-6.85726404e-01 -4.98278975e-01 3.56664479e-01 3.29123437e-01
-1.47373617e-01 -1.87428236e-01 -2.82021016e-01 -1.35497224e+00
-3.84742655e-02 -1.45917284e+00 5.04669249e-01 -1.05258727e+00
-1.25493419e+00 6.18566692e-01 2.69825488e-01 -1.02607274e+00
-7.14400649e-01 -4.59888220e-01 -1.25419367e-02 1.17668271e+00
-1.02416432e+00 -1.62717819e+00 8.31639990e-02 4.54913080e-01
7.07882226e-01 -1.88069329e-01 9.31522131e-01 4.13235337e-01
-4.84958261e-01 6.93940639e-01 9.50817298e-03 5.76300621e-01
1.15876937e+00 -1.33824182e+00 1.15639925e+00 1.09504020e+00
1.12814993e-01 8.48424375e-01 6.55062616e-01 -1.14979041e+00
-1.08160257e+00 -1.16723609e+00 1.58966780e+00 -1.02221465e+00
9.55147326e-01 -7.11689711e-01 -5.73200762e-01 1.10010660e+00
7.35819042e-01 -6.13823354e-01 7.87849486e-01 6.03842176e-02
-3.54608715e-01 3.03531438e-01 -9.05920267e-01 9.13661897e-01
1.03159797e+00 -8.22970331e-01 -6.24639392e-01 5.00857294e-01
6.55846536e-01 -6.72868550e-01 -9.75493968e-01 1.37788177e-01
6.78068578e-01 -6.78978801e-01 4.02168989e-01 -6.73649848e-01
8.26067626e-01 -3.33772063e-01 -2.27445126e-01 -1.50982714e+00
3.53104272e-03 -8.02653909e-01 2.60580897e-01 1.43231368e+00
1.05186689e+00 -5.88516533e-01 1.51614130e-01 2.43621096e-01
-1.16688579e-01 -5.15209079e-01 -7.87613630e-01 -4.38400984e-01
5.48113942e-01 -5.34490347e-01 5.54413319e-01 1.43644178e+00
9.76795778e-02 4.74279732e-01 -4.72713590e-01 -1.85845822e-01
4.22536403e-01 -1.25760809e-01 8.33441317e-01 -5.82316041e-01
-9.54362564e-03 -2.11200058e-01 3.81120592e-01 -4.98480737e-01
-6.28312305e-02 -1.23726332e+00 -7.08772242e-02 -1.93172693e+00
4.41265881e-01 -3.77054781e-01 3.42839032e-01 8.81629229e-01
-2.90520549e-01 5.00176191e-01 2.08737373e-01 3.72152656e-01
-2.65061498e-01 2.11014137e-01 1.01117909e+00 -1.08629979e-01
-1.18391573e-01 -5.00656843e-01 -9.99718130e-01 4.85424012e-01
8.44798744e-01 -5.79918444e-01 -1.79476514e-01 -9.25458431e-01
5.33376753e-01 6.70203334e-03 5.12247533e-02 -6.08430266e-01
-1.53764829e-01 -3.33764017e-01 5.64898729e-01 -6.84944510e-01
3.92714404e-02 -1.48823157e-01 3.74250770e-01 3.95401686e-01
-4.85778838e-01 6.42707229e-01 3.13231766e-01 -5.03308587e-02
-4.09848504e-02 2.42758811e-01 3.89569461e-01 -2.96095043e-01
-1.96194649e-01 -5.81108481e-02 -5.81570446e-01 6.40857339e-01
5.46661615e-01 5.17972931e-02 -8.25002372e-01 -6.57251358e-01
-3.43605369e-01 5.06635606e-01 7.65612721e-01 7.04710066e-01
3.19109678e-01 -1.44853365e+00 -1.41318285e+00 -6.61896244e-02
4.13739413e-01 -2.55109608e-01 9.83453635e-03 7.68976748e-01
-7.98716366e-01 6.86617076e-01 -2.61542857e-01 -4.31971438e-02
-6.56621218e-01 1.80706501e-01 7.85814822e-02 -2.94096082e-01
-3.40346426e-01 2.82675236e-01 -2.85258088e-02 -1.08398390e+00
-4.02813107e-01 -3.41479599e-01 5.16644605e-02 3.06082368e-01
4.30151939e-01 3.72482985e-01 -1.86803743e-01 -1.14053702e+00
-4.20513690e-01 1.00615941e-01 3.94545868e-02 -8.20900559e-01
1.12973082e+00 -3.61848950e-01 -1.02216095e-01 7.85368443e-01
8.87200356e-01 8.03651333e-01 -5.74932694e-01 6.62597120e-02
-2.67719664e-02 -9.25255641e-02 -6.37471735e-01 -1.29077363e+00
-5.41811526e-01 5.94195127e-01 1.18048368e-02 -1.00312904e-01
7.49964476e-01 1.42050058e-01 7.08542645e-01 1.35253996e-01
2.85370111e-01 -1.21078408e+00 -3.70675176e-01 6.57265127e-01
1.09577608e+00 -1.25690091e+00 9.71763432e-02 -1.23221084e-01
-9.63644683e-01 8.14754128e-01 4.64822412e-01 5.86396337e-01
2.95991093e-01 1.72180936e-01 5.48728824e-01 -1.02779858e-01
-9.10586357e-01 -8.95637423e-02 2.46955633e-01 7.68255413e-01
9.76328671e-01 3.33440781e-01 -5.41346192e-01 6.69876277e-01
-9.13446248e-01 -2.67098665e-01 6.95862412e-01 8.15503061e-01
1.59593359e-01 -1.18727374e+00 -4.25667763e-01 5.79456329e-01
-7.80945599e-01 -5.70469737e-01 -7.34553218e-01 1.12059927e+00
-2.29222178e-02 1.07151985e+00 -6.04804270e-02 -2.73789048e-01
1.74249962e-01 2.56043434e-01 2.07302243e-01 -7.28568256e-01
-5.99503458e-01 9.43513066e-02 7.09526300e-01 -2.10557252e-01
-4.02134210e-01 -8.58077824e-01 -1.04570436e+00 -5.30324876e-01
2.90918201e-01 4.23513763e-02 4.49794143e-01 6.70712173e-01
3.86561990e-01 1.38228133e-01 2.50319481e-01 -4.30090249e-01
9.95576382e-02 -1.39904702e+00 -1.18738838e-01 3.55028003e-01
-3.33823683e-03 -1.09828487e-01 -8.45646486e-02 3.09891731e-01] | [11.509474754333496, 9.6818265914917] |
9ab77d61-d045-4179-8382-b0a5d8797a7c | meta-learning-for-airflow-simulations-with | 2306.10624 | null | https://arxiv.org/abs/2306.10624v1 | https://arxiv.org/pdf/2306.10624v1.pdf | Meta-Learning for Airflow Simulations with Graph Neural Networks | The field of numerical simulation is of significant importance for the design and management of real-world systems, with partial differential equations (PDEs) being a commonly used mathematical modeling tool. However, solving PDEs remains still a challenge, as commonly used traditional numerical solvers often require high computational costs. As a result, data-driven methods leveraging machine learning (more particularly Deep Learning) algorithms have been increasingly proposed to learn models that can predict solutions to complex PDEs, such as those arising in computational fluid dynamics (CFD). However, these methods are known to suffer from poor generalization performance on out-of-distribution (OoD) samples, highlighting the need for more efficient approaches. To this end, we present a meta-learning approach to enhance the performance of learned models on OoD samples. Specifically, we set the airflow simulation in CFD over various airfoils as a meta-learning problem, where each set of examples defined on a single airfoil shape is treated as a separate task. Through the use of model-agnostic meta-learning (MAML), we learn a meta-learner capable of adapting to new tasks, i.e., previously unseen airfoil shapes, using only a small amount of task-specific data. We experimentally demonstrate the efficiency of the proposed approach for improving the OoD generalization performance of learned models while maintaining efficiency. | ['Marc Schoenauer', 'Mouadh Yagoubi', 'Wenzhuo LIU'] | 2023-06-18 | null | null | null | null | ['meta-learning', 'management'] | ['methodology', 'miscellaneous'] | [ 3.80652472e-02 -5.21059453e-01 6.00404218e-02 -1.43951967e-01
-4.64750022e-01 -5.30199647e-01 4.56830204e-01 4.92890716e-01
-2.65746504e-01 6.33829415e-01 -5.36483169e-01 -3.10018003e-01
-4.18705434e-01 -9.15097892e-01 -7.96471477e-01 -5.99997342e-01
-1.36079386e-01 5.16097665e-01 -1.38856769e-01 -2.63892710e-01
2.65038997e-01 8.40655446e-01 -1.79498625e+00 2.95408636e-01
1.29131448e+00 1.06513703e+00 1.87142566e-01 8.08677971e-01
-2.69063890e-01 5.61409414e-01 -6.56294823e-01 7.58947730e-02
1.49831757e-01 -2.54514694e-01 -6.87048018e-01 5.55542856e-02
4.58869368e-01 2.11858824e-01 -1.34339094e-01 7.53692269e-01
3.17228317e-01 7.76585877e-01 8.30134988e-01 -1.18825650e+00
-2.31165648e-01 -1.39398903e-01 -1.25386640e-01 3.27026874e-01
-1.71405479e-01 4.15922254e-01 6.59314454e-01 -9.21774030e-01
1.19646035e-01 1.10726213e+00 6.38653934e-01 6.62635267e-01
-1.28097761e+00 -5.86305141e-01 1.05910242e-01 -1.33273512e-01
-9.88431334e-01 -1.06446845e-02 8.30505371e-01 -6.41837418e-01
7.14183629e-01 3.01416367e-01 6.93536222e-01 8.72821629e-01
4.46667850e-01 6.21168613e-01 1.08820009e+00 -8.49429891e-02
7.04249203e-01 3.28026980e-01 -2.14148477e-01 4.37552363e-01
2.97006696e-01 2.68351257e-01 -1.78777829e-01 -1.84949875e-01
4.90141898e-01 1.01664074e-01 -7.14844316e-02 -5.72860599e-01
-6.77618027e-01 7.78136432e-01 4.76491630e-01 3.10743213e-01
-2.69408405e-01 1.84209570e-02 4.41158831e-01 3.09305489e-01
7.56306291e-01 1.11043513e+00 -6.75043821e-01 -1.94588810e-01
-7.75646031e-01 8.61165941e-01 8.53613555e-01 4.82408524e-01
6.48589075e-01 3.42616916e-01 1.75723925e-01 8.77440870e-01
-3.02692596e-02 2.69946635e-01 6.20452821e-01 -6.71280742e-01
4.17951792e-01 8.04065764e-01 2.61609584e-01 -1.08994484e+00
-4.41084981e-01 -8.51558685e-01 -1.04627168e+00 4.03143108e-01
2.05088109e-01 -3.59979063e-01 -7.87062407e-01 1.37090075e+00
5.89468658e-01 5.01067340e-01 4.42648232e-02 8.32577407e-01
4.34738666e-01 9.91883099e-01 1.58955067e-01 -1.22696478e-02
7.34160066e-01 -8.98757577e-01 -2.50762969e-01 -3.11828792e-01
7.79072285e-01 -5.52953124e-01 9.46449399e-01 6.29739165e-01
-1.00265253e+00 -1.04966795e+00 -1.01314700e+00 3.18784148e-01
-6.70964718e-01 -2.41739094e-01 4.22977686e-01 5.34308851e-01
-4.80236411e-01 1.08192146e+00 -7.77789354e-01 1.53680101e-01
4.71510977e-01 3.18839878e-01 3.88661586e-03 -3.73474360e-02
-9.05411184e-01 7.25276709e-01 4.87774760e-01 1.02680333e-01
-1.18558717e+00 -1.48090303e+00 -6.83637202e-01 1.76923200e-01
2.40536153e-01 -6.33788049e-01 1.22598898e+00 -8.53502512e-01
-1.51553476e+00 5.52894950e-01 1.58492088e-01 -3.88632715e-01
5.74769497e-01 -3.79420012e-01 -5.34207344e-01 -1.64071351e-01
-4.05534685e-01 2.84862518e-01 1.18207431e+00 -1.57992697e+00
-8.43874633e-01 -4.27692682e-02 5.39448299e-02 7.97088072e-02
-5.40713847e-01 -6.23978615e-01 3.53324145e-01 -7.33925045e-01
-1.02458149e-01 -9.36112404e-01 -3.81426126e-01 -3.04247066e-03
-2.05079978e-03 -2.53667414e-01 1.05306947e+00 -2.41525844e-01
1.29809046e+00 -2.01387191e+00 4.29144681e-01 1.18053481e-01
2.73038149e-01 8.35656285e-01 -2.29302095e-03 3.29433024e-01
-1.46055683e-01 6.26631230e-02 -5.30152440e-01 -2.63530880e-01
-2.18226537e-01 2.50042289e-01 -3.99701148e-01 2.13584021e-01
4.67244297e-01 5.99331260e-01 -1.02528501e+00 -4.95583080e-02
5.49588919e-01 3.29365760e-01 -6.38361752e-01 6.92854226e-01
-7.76031911e-01 9.49355483e-01 -5.77314913e-01 3.75888795e-01
6.61775529e-01 -3.75565559e-01 -4.92666289e-03 -8.88186600e-03
-2.33144313e-01 -2.38751486e-01 -1.22385430e+00 1.55042827e+00
-1.21814156e+00 1.88024804e-01 6.66408986e-02 -1.44298494e+00
1.06716692e+00 7.84346163e-02 4.91016388e-01 -4.85447884e-01
2.19269753e-01 4.85565722e-01 2.58874893e-01 -7.06683934e-01
3.93105119e-01 -4.07168806e-01 1.43380448e-01 2.83723921e-01
-6.47152588e-02 -6.04421675e-01 -2.23613381e-02 -3.96627337e-01
7.43449867e-01 -1.71847865e-01 9.30329859e-02 -3.79989117e-01
8.11504543e-01 2.23497480e-01 2.75735229e-01 5.05056083e-01
6.66945130e-02 2.03629419e-01 -6.63892273e-03 -9.80607092e-01
-1.03082025e+00 -8.56012404e-01 -2.66783893e-01 9.55624402e-01
3.05981934e-02 -1.34327605e-01 -8.10557365e-01 -4.15277064e-01
4.98512238e-01 7.46910632e-01 -6.12438142e-01 -5.80960751e-01
-9.49238539e-01 -8.39576125e-01 1.60950109e-01 4.83235449e-01
4.03405458e-01 -1.01621807e+00 -8.84585023e-01 3.92198622e-01
5.87339520e-01 -9.36238170e-01 7.27247968e-02 2.45737329e-01
-1.33064914e+00 -1.15070438e+00 -7.84013152e-01 -7.88577855e-01
6.15559638e-01 -3.76132899e-03 9.27845061e-01 2.11820066e-01
-6.55246973e-01 3.92835319e-01 -1.60917509e-02 -2.87732482e-01
-7.27295041e-01 3.33337843e-01 4.71766740e-02 1.55390531e-01
-1.92043126e-01 -5.71942091e-01 -4.90883291e-01 1.33156195e-01
-1.03237998e+00 -3.54899019e-02 4.13096935e-01 1.08528137e+00
5.57813525e-01 5.24934530e-01 7.92174876e-01 -1.01069057e+00
8.87123823e-01 -8.00853968e-01 -8.10460746e-01 4.66509461e-02
-5.30500293e-01 6.00925721e-02 1.53772581e+00 -6.92747712e-01
-1.08916473e+00 -2.07635120e-01 1.08724669e-01 -9.18356597e-01
-4.32279915e-01 6.38217807e-01 6.14887960e-02 -3.36859763e-01
7.24124074e-01 2.41356418e-01 8.30118731e-02 -7.09363937e-01
8.73275101e-02 3.85881871e-01 3.63340467e-01 -1.07538843e+00
9.52444673e-01 2.16791898e-01 5.34387946e-01 -1.02312303e+00
-9.50042844e-01 -2.76163727e-01 -4.39789653e-01 -1.73492402e-01
4.06067550e-01 -5.03599346e-01 -8.14616799e-01 7.62684286e-01
-7.47330189e-01 -5.81601024e-01 -3.01083148e-01 4.26461518e-01
-4.61305410e-01 -1.84327289e-01 -3.06939960e-01 -8.93132329e-01
-1.66294008e-01 -1.07511556e+00 7.26526678e-01 4.16935265e-01
-2.14283079e-01 -1.53524792e+00 3.19040120e-01 1.06546953e-01
6.39638841e-01 6.84546769e-01 1.35460484e+00 -5.50024450e-01
-2.35309362e-01 -2.67853528e-01 1.78564638e-01 5.70649326e-01
3.80657494e-01 -1.60095301e-02 -1.13197339e+00 -6.25448883e-01
3.12551409e-01 -4.26904261e-01 6.23754561e-01 1.58977598e-01
1.82160199e+00 -2.10026577e-01 -3.83476913e-01 6.40318871e-01
1.28765082e+00 2.37085760e-01 -1.18474349e-01 1.25998810e-01
6.73333585e-01 5.35504401e-01 6.89285040e-01 6.33119881e-01
-1.31016091e-01 3.34773779e-01 5.12056828e-01 -2.46068649e-03
1.83187217e-01 -1.38172492e-01 -1.95407301e-01 7.31906474e-01
1.37202054e-01 -3.25628877e-01 -1.14130902e+00 5.51407397e-01
-1.55853701e+00 -4.39813614e-01 1.84879288e-01 2.16878343e+00
6.34866714e-01 1.51622057e-01 -4.97485586e-02 3.64101589e-01
5.11655748e-01 1.52541846e-01 -9.91678894e-01 -6.09458447e-01
3.27927798e-01 6.52739942e-01 -2.71573458e-02 2.21446574e-01
-1.32964027e+00 6.17062926e-01 5.19519806e+00 7.43876278e-01
-1.63044596e+00 -1.38662100e-01 8.01308930e-01 2.96815988e-02
-1.54724466e-02 -4.82857823e-01 -5.95300257e-01 5.73765755e-01
1.04204333e+00 -2.01453403e-01 5.33955872e-01 1.11924672e+00
2.30249882e-01 8.83737206e-02 -1.25498569e+00 8.71372581e-01
-2.29449168e-01 -1.40941155e+00 1.57682136e-01 -9.82675627e-02
1.21009457e+00 -4.08964545e-01 5.77321708e-01 4.82072622e-01
-9.71479267e-02 -1.38533378e+00 3.22917193e-01 3.90293986e-01
5.21911442e-01 -9.19358552e-01 5.22005677e-01 6.72927260e-01
-1.19769096e+00 -4.07327503e-01 -3.56351107e-01 -2.23268658e-01
-1.13373876e-01 5.30544817e-01 -6.55088782e-01 5.65540671e-01
5.60225427e-01 7.89036274e-01 -2.94287026e-01 1.19294262e+00
4.02699143e-01 5.11381507e-01 6.28813030e-03 -1.09218858e-01
4.19958889e-01 -5.70314340e-02 4.53095645e-01 8.26402962e-01
5.54213405e-01 1.55509979e-01 3.48912656e-01 9.62828279e-01
-2.36901611e-01 5.62141165e-02 -5.67569852e-01 -1.32188007e-01
2.36355796e-01 1.17519271e+00 -3.74109983e-01 -2.34054536e-01
-1.71232492e-01 4.42224354e-01 3.50079060e-01 1.98234499e-01
-7.28974581e-01 -1.43517554e-01 9.52868044e-01 2.48200387e-01
2.50886738e-01 -2.02738941e-01 -2.03562707e-01 -8.69628489e-01
-1.77675053e-01 -1.01053011e+00 4.22851384e-01 -2.56153405e-01
-1.53892243e+00 3.99594933e-01 3.22988885e-03 -1.52461874e+00
-2.00825617e-01 -1.00058734e+00 -7.27949560e-01 9.04376805e-01
-1.71093726e+00 -5.93697250e-01 -6.14212692e-01 2.27995202e-01
7.34703720e-01 -2.85474926e-01 8.23472083e-01 2.63795435e-01
-4.95678157e-01 3.41970295e-01 4.27200824e-01 -1.17016673e-01
4.29613262e-01 -1.23170984e+00 3.30213994e-01 3.15380335e-01
-2.14051709e-01 5.06924868e-01 6.35152757e-01 -4.77564096e-01
-1.66775048e+00 -1.53240991e+00 2.75498360e-01 -1.56466708e-01
4.97232318e-01 -1.33018032e-01 -1.33985615e+00 1.17908724e-01
-3.93982291e-01 5.63659012e-01 4.29673404e-01 -1.40745342e-01
5.47947474e-02 -3.43681782e-01 -1.38644230e+00 4.47577298e-01
7.38649487e-01 -3.15830559e-01 -4.13020581e-01 2.90958375e-01
4.27500039e-01 -8.28809857e-01 -9.85111356e-01 7.20429659e-01
3.36119026e-01 -7.28587747e-01 1.06445014e+00 -1.13204074e+00
5.92330456e-01 -1.44457698e-01 1.41783923e-01 -1.81673121e+00
7.24897832e-02 -3.92965615e-01 -6.55199826e-01 7.78368890e-01
4.81338892e-03 -6.61267042e-01 8.52313936e-01 4.73544210e-01
-3.56148899e-01 -1.22210252e+00 -8.77554297e-01 -9.13156807e-01
6.17776752e-01 -3.65584314e-01 6.49756014e-01 1.16033792e+00
-3.43057096e-01 -1.10337459e-01 -2.27482408e-01 3.35981250e-01
4.69957530e-01 4.27003771e-01 7.67998338e-01 -1.46706653e+00
-4.12637502e-01 -4.63481277e-01 -1.35029837e-01 -8.23852897e-01
4.47786778e-01 -8.72195363e-01 -6.64676353e-03 -9.02428985e-01
-4.86096233e-01 -9.02396023e-01 -4.16889846e-01 -7.05307648e-02
-2.08296657e-01 -2.52365351e-01 6.98083192e-02 -5.83760180e-02
-1.46893278e-01 7.08240509e-01 1.64503336e+00 -3.35642934e-01
-3.14407974e-01 4.36933845e-01 -1.45644039e-01 7.59449840e-01
9.24148381e-01 -5.30376494e-01 -5.79480052e-01 -3.29326302e-01
-1.63782649e-02 1.66818112e-01 4.94974107e-01 -1.39046049e+00
1.77753404e-01 -4.07010138e-01 5.16893685e-01 -2.01342016e-01
3.40153962e-01 -9.17693555e-01 -2.19712555e-02 8.79380107e-01
-3.73710126e-01 1.58063576e-01 7.47421443e-01 6.41468525e-01
-3.09530973e-01 -2.97678143e-01 9.83774364e-01 -2.84033239e-01
-6.05915546e-01 4.50630873e-01 -2.54559666e-01 4.67716128e-01
1.08267677e+00 5.98693229e-02 -1.71354618e-02 1.50165692e-01
-5.43806732e-01 1.57020897e-01 2.95322597e-01 5.49004495e-01
6.43230140e-01 -9.21487808e-01 -4.73666012e-01 4.28563982e-01
3.80937685e-03 5.62732637e-01 5.36279142e-01 4.13012266e-01
-6.47256196e-01 2.97007501e-01 -2.26534724e-01 -6.72871113e-01
-6.93948627e-01 6.14329636e-01 7.34592855e-01 -3.68098825e-01
-4.91898954e-01 8.22120488e-01 2.04844743e-01 -7.01080620e-01
1.02018550e-01 -6.37302399e-01 -2.52212472e-02 -9.96296480e-02
3.05126607e-01 7.27710605e-01 2.91331738e-01 -1.89298794e-01
-2.34302543e-02 6.13199174e-01 8.81428719e-02 3.99917305e-01
1.38561964e+00 5.23052692e-01 3.24170351e-01 5.48948824e-01
1.14996314e+00 -2.87612766e-01 -1.35663402e+00 -4.34875004e-02
7.55668432e-02 -5.85215688e-01 4.59183194e-02 -6.77580476e-01
-1.10465562e+00 1.17963934e+00 5.80154121e-01 1.32411063e-01
8.08812976e-01 -5.50394893e-01 1.01213729e+00 5.53915143e-01
1.95829034e-01 -8.71509731e-01 3.93585026e-01 4.33104217e-01
7.06709802e-01 -1.28829324e+00 -1.87411532e-01 -8.31773952e-02
-2.13144779e-01 1.41806364e+00 8.47292423e-01 -4.23640043e-01
8.78359675e-01 2.84339339e-01 -1.09154619e-01 6.49815351e-02
-7.86528111e-01 4.48442310e-01 4.96840715e-01 2.88641870e-01
2.15550765e-01 -6.73290938e-02 2.53484637e-01 5.79317451e-01
9.06734075e-03 1.60541341e-01 9.15810093e-02 1.09887803e+00
-2.94097364e-01 -1.09193170e+00 -2.79375315e-01 5.85250676e-01
-1.88444704e-02 2.18796849e-01 1.30421072e-01 8.11459243e-01
2.23672464e-01 6.29130363e-01 1.42338112e-01 -2.67014712e-01
4.61534828e-01 1.35511398e-01 4.07122046e-01 -8.39979947e-01
-9.72845912e-01 -5.64764559e-01 -4.05802578e-01 -1.88850418e-01
-9.83836651e-02 -3.76670420e-01 -1.12910748e+00 -3.65079314e-01
9.82595831e-02 3.73613685e-01 4.26266849e-01 9.84916031e-01
3.98494571e-01 8.29141498e-01 8.68678868e-01 -9.12462533e-01
-9.47598457e-01 -6.20921612e-01 -4.80566144e-01 5.03630042e-01
5.81878066e-01 -9.28353131e-01 -5.17539144e-01 -1.08775944e-01] | [6.45076322555542, 3.407475233078003] |
0ffd9d8e-0788-4ae1-b24d-b06710c76f67 | least-square-estimation-network-for-depth | 2203.03317 | null | https://arxiv.org/abs/2203.03317v2 | https://arxiv.org/pdf/2203.03317v2.pdf | Depth-Independent Depth Completion via Least Square Estimation | The depth completion task aims to complete a per-pixel dense depth map from a sparse depth map. In this paper, we propose an efficient least square based depth-independent method to complete the sparse depth map utilizing the RGB image and the sparse depth map in two independent stages. In this way can we decouple the neural network and the sparse depth input, so that when some features of the sparse depth map change, such as the sparsity, our method can still produce a promising result. Moreover, due to the positional encoding and linear procession in our pipeline, we can easily produce a super-resolution dense depth map of high quality. We also test the generalization of our method on different datasets compared to some state-of-the-art algorithms. Experiments on the benchmark show that our method produces competitive performance. | ['Yunkai Wang', 'Rong Xiong', 'Yue Wang', 'Zexi Chen', 'Xianze Fang'] | 2022-03-07 | null | null | null | null | ['depth-completion'] | ['computer-vision'] | [ 3.90672356e-01 2.25788400e-01 2.25277860e-02 -5.41548312e-01
-7.26959944e-01 -1.59995660e-01 1.68238550e-01 -2.19506145e-01
-3.07920963e-01 5.95929801e-01 3.34865808e-01 3.43538493e-01
1.57920301e-01 -1.11333561e+00 -8.10911775e-01 -7.00689077e-01
1.97937012e-01 3.75031322e-01 6.05269194e-01 3.18306014e-02
4.36251998e-01 4.41657931e-01 -1.67222965e+00 5.67281961e-01
5.60892522e-01 1.24700201e+00 5.51231682e-01 5.91911793e-01
-2.19083503e-01 9.37547982e-01 -1.06585316e-01 1.18609287e-01
5.50933957e-01 -2.11977094e-01 -7.75966465e-01 1.98175251e-01
7.86262333e-01 -9.76060808e-01 -6.50230765e-01 1.02665830e+00
3.41475397e-01 -9.94868875e-02 9.39308479e-02 -6.05344713e-01
-4.82746959e-02 2.59797275e-01 -8.76955152e-01 -2.44619533e-01
6.09113157e-01 -4.14601415e-02 7.45870769e-01 -1.03767633e+00
8.76214564e-01 1.18922901e+00 4.21980977e-01 4.38441545e-01
-1.14210534e+00 -5.33550560e-01 3.48577112e-01 7.96410963e-02
-1.39486766e+00 -5.31016588e-01 8.66852760e-01 -4.87748832e-02
1.10072887e+00 -1.48763314e-01 7.97261417e-01 7.03345537e-01
6.34129113e-03 6.42623782e-01 9.79888916e-01 -1.15471527e-01
4.90051597e-01 -2.92148113e-01 -1.91629872e-01 8.10275733e-01
1.86272219e-01 1.80099621e-01 -9.16996479e-01 2.28549123e-01
1.12811720e+00 2.51406550e-01 -3.22283834e-01 -4.03272778e-01
-1.18647325e+00 5.75498044e-01 7.85852194e-01 1.63193464e-01
-4.88253891e-01 3.95178199e-01 -1.00776672e-01 -3.17155151e-03
4.58948553e-01 2.58851480e-02 -3.69780838e-01 -1.27161056e-01
-1.21192932e+00 1.47963613e-01 6.89424634e-01 8.89621019e-01
1.47191787e+00 -2.92517334e-01 1.95775732e-01 3.92160743e-01
2.64688760e-01 5.26329935e-01 2.64697552e-01 -1.32936001e+00
6.15034282e-01 6.08985186e-01 -8.76471624e-02 -8.64304721e-01
-5.50438702e-01 -7.83281699e-02 -1.11044908e+00 4.70800132e-01
3.43858987e-01 1.20348148e-02 -1.02091324e+00 1.31054223e+00
2.74272233e-01 3.76264423e-01 7.71152005e-02 9.69806671e-01
9.22507286e-01 7.35285997e-01 -6.51467264e-01 -8.26328248e-02
8.35554421e-01 -1.10730863e+00 -6.57322764e-01 -8.10041130e-01
3.68124098e-01 -4.16937441e-01 8.00285876e-01 7.75849998e-01
-1.23300862e+00 -4.84496653e-01 -1.14960515e+00 -6.37229264e-01
1.31084099e-01 -1.47805378e-01 9.19470370e-01 1.27965853e-01
-1.19955254e+00 7.30078578e-01 -1.31090117e+00 -1.42568886e-01
5.18517256e-01 2.93875545e-01 -6.36131406e-01 -7.52825737e-01
-6.03245795e-01 3.93750638e-01 4.39040333e-01 1.18532613e-01
-1.02072477e+00 -5.98355711e-01 -1.09161866e+00 -1.41100418e-02
2.53159583e-01 -8.36295962e-01 1.09615254e+00 -5.90996444e-01
-1.60054457e+00 7.19747663e-01 -7.73936927e-01 -1.09043799e-01
2.61222899e-01 -2.17349514e-01 4.78361815e-01 5.27506173e-01
8.59773010e-02 1.00822842e+00 7.46405363e-01 -1.20546222e+00
-1.00850058e+00 -5.73789001e-01 1.06794156e-01 2.10992903e-01
-1.84083939e-01 -6.81988358e-01 -8.13437283e-01 -6.49373606e-02
1.01690733e+00 -6.56017184e-01 -5.25044441e-01 4.03441042e-01
-2.13132247e-01 3.52074027e-01 4.74755645e-01 -4.46260870e-01
9.72368479e-01 -2.26807976e+00 6.22220397e-01 2.12674648e-01
5.25418282e-01 -4.88174438e-01 4.14954359e-03 2.27250472e-01
1.96215466e-01 -1.99009895e-01 -5.62059104e-01 -1.04617238e+00
-4.60204005e-01 5.26316941e-01 -1.42004699e-01 5.15997052e-01
3.92568000e-02 8.16335738e-01 -1.01514542e+00 -2.90687293e-01
2.81501293e-01 5.91695070e-01 -9.15400803e-01 3.44674706e-01
-9.31602269e-02 6.49311304e-01 -3.57039392e-01 7.45502949e-01
9.33247805e-01 -3.60312819e-01 1.60035267e-02 -3.44854146e-01
-1.97165042e-01 3.48808348e-01 -1.28728843e+00 2.71446991e+00
-5.45730889e-01 4.99198437e-01 2.00004026e-01 -6.50522470e-01
8.22482765e-01 -4.58794981e-02 3.71435761e-01 -9.36743557e-01
-9.03908834e-02 2.58679628e-01 -4.24516916e-01 -7.25297332e-02
5.90256870e-01 -2.12032944e-01 2.73844659e-01 3.59852731e-01
7.86199942e-02 -6.06885791e-01 -6.97578937e-02 2.50972599e-01
1.28098869e+00 3.02890450e-01 3.75238620e-02 1.62665769e-01
2.73568958e-01 -1.63611978e-01 6.35390043e-01 5.34644723e-01
7.95433670e-02 1.22274828e+00 5.22735775e-01 -4.88027006e-01
-1.01169658e+00 -9.81339037e-01 -9.30892527e-02 4.28703666e-01
3.33402097e-01 -4.34274912e-01 -3.82497430e-01 -3.08293164e-01
-2.53592223e-01 3.06749139e-02 -6.63402379e-01 1.47395432e-01
-6.08609855e-01 -4.64021266e-01 -2.98015606e-02 5.03847420e-01
7.74655581e-01 -7.34885395e-01 -6.55134320e-01 1.74556449e-01
-1.22659802e-01 -1.50595427e+00 2.18080506e-02 4.78331119e-01
-1.35497451e+00 -7.98341870e-01 -6.40512764e-01 -7.55807340e-01
8.66076291e-01 6.44876897e-01 1.12217414e+00 4.99020889e-02
6.23927973e-02 -1.02427296e-01 -2.69108057e-01 2.32569322e-01
3.00152063e-01 1.02856942e-01 -3.40771526e-01 -1.47909656e-01
3.55736203e-02 -1.00046480e+00 -8.68829489e-01 -8.59932676e-02
-1.05821717e+00 5.01346111e-01 5.57249606e-01 5.44137716e-01
1.12116075e+00 3.22120398e-01 -1.58836413e-02 -8.45991373e-01
-2.65306830e-01 -4.66717035e-01 -7.86306441e-01 -3.99480730e-01
-4.81470346e-01 2.72846073e-01 3.59314442e-01 5.46363145e-02
-9.96389091e-01 7.95357704e-01 -4.52427834e-01 -5.35101235e-01
-4.45198789e-02 4.56244022e-01 -3.79237056e-01 -3.07099134e-01
5.30197263e-01 2.54806310e-01 -1.76178604e-01 -7.04980791e-01
3.94420952e-01 2.13355809e-01 6.10005081e-01 -2.62234360e-01
8.69620740e-01 1.18510294e+00 3.16504687e-01 -5.09835660e-01
-9.53366339e-01 -6.26398027e-01 -7.93201149e-01 1.98867723e-01
8.12861204e-01 -1.43957615e+00 -5.05157590e-01 6.64210081e-01
-1.27283776e+00 -5.85555375e-01 -3.10274035e-01 2.99303949e-01
-5.73845923e-01 4.32653189e-01 -7.91674197e-01 -4.67617780e-01
-2.43918076e-02 -1.06538749e+00 1.44954085e+00 1.91771895e-01
1.33872822e-01 -8.08566749e-01 1.86401218e-01 9.68259498e-02
2.49499619e-01 3.82428914e-01 3.72902095e-01 2.23015159e-01
-1.32201111e+00 7.95812979e-02 -4.63111371e-01 2.13103220e-01
1.13502234e-01 -2.55900413e-01 -1.20599127e+00 -3.31471771e-01
2.15491548e-01 -4.05497015e-01 1.16387177e+00 3.59614223e-01
1.13571405e+00 -2.22241864e-01 -2.18483388e-01 1.46309769e+00
1.91528428e+00 -2.34547004e-01 1.05708635e+00 4.72628534e-01
1.12528348e+00 6.88370287e-01 5.56047976e-01 5.86635292e-01
6.18362010e-01 4.25643355e-01 8.85296345e-01 -1.84909493e-01
-1.81456685e-01 -5.08390546e-01 1.71004295e-01 6.18194103e-01
2.84515880e-02 3.31947617e-02 -7.59390831e-01 5.32661617e-01
-1.82307935e+00 -7.31172800e-01 -1.51821092e-01 2.11537385e+00
8.16081345e-01 2.30441123e-01 -3.20138842e-01 2.75142878e-01
5.68802804e-02 4.95811969e-01 -6.55755699e-01 -5.06443344e-02
-3.08031917e-01 5.19680500e-01 7.23267674e-01 9.22957420e-01
-6.88412726e-01 9.93156314e-01 6.63848448e+00 5.12528896e-01
-1.03465915e+00 -6.88646175e-03 6.17299318e-01 -6.43015623e-01
-6.65698528e-01 -3.95142613e-03 -8.92867923e-01 1.65093586e-01
5.51743925e-01 1.67018026e-01 5.29292762e-01 8.52383852e-01
1.42886028e-01 -7.34198987e-01 -1.35956383e+00 1.51318705e+00
4.85082120e-02 -1.48104274e+00 -1.34710491e-01 1.93497628e-01
1.16218042e+00 2.84648418e-01 -1.85372353e-01 -1.87998369e-01
7.91923851e-02 -1.07285821e+00 5.95457494e-01 4.00413871e-01
9.43395078e-01 -6.57822192e-01 6.27235115e-01 4.73754615e-01
-1.15046060e+00 9.06681120e-02 -6.89508200e-01 -4.90837991e-01
4.28489029e-01 1.19526720e+00 -3.00005287e-01 5.25502086e-01
9.62855279e-01 1.25684726e+00 -4.05529767e-01 8.74494612e-01
-6.76316321e-01 6.84568193e-03 -5.53938091e-01 5.53080976e-01
2.29967043e-01 -1.01479717e-01 1.19584061e-01 7.84399986e-01
4.51393932e-01 3.77149552e-01 -1.39072910e-01 9.43838716e-01
4.27712016e-02 -3.28115851e-01 -7.43754625e-01 3.37971032e-01
2.47833759e-01 1.24379766e+00 -7.44290709e-01 -2.62265354e-01
-5.19383311e-01 1.38948894e+00 5.16628444e-01 3.96885931e-01
-2.77844995e-01 3.59526207e-03 7.83889771e-01 2.31958061e-01
5.25200427e-01 -4.86063272e-01 -8.04494143e-01 -1.34906375e+00
2.36524537e-01 -6.25249743e-01 -2.38107052e-02 -8.59952629e-01
-8.67638350e-01 7.06458628e-01 -2.92267084e-01 -1.17445111e+00
-3.30244869e-01 -3.50316107e-01 -2.81654119e-01 9.91308749e-01
-2.04369020e+00 -8.94170046e-01 -8.73823047e-01 7.36481309e-01
3.76287729e-01 3.34829807e-01 4.96794105e-01 3.83862048e-01
-2.65209943e-01 1.63251102e-01 -1.27096981e-01 -1.47822499e-01
4.76266265e-01 -1.08898604e+00 5.83195627e-01 9.81651187e-01
-4.89500165e-02 2.04229534e-01 3.08001131e-01 -6.91631675e-01
-1.57062006e+00 -8.36238384e-01 7.55847752e-01 -2.19286099e-01
3.85429859e-01 -4.15279150e-01 -9.97531712e-01 6.18825614e-01
-2.20043227e-01 2.52882153e-01 2.07042024e-01 -9.95505974e-02
-3.34381759e-01 -2.60616630e-01 -1.15470767e+00 2.25432679e-01
1.38409019e+00 -6.67232096e-01 -2.15271994e-01 1.43736452e-01
8.77002180e-01 -9.13510501e-01 -5.62132359e-01 3.96065235e-01
6.12512887e-01 -1.64088273e+00 1.07736850e+00 1.69775426e-01
1.12496090e+00 -4.60479081e-01 -4.26502466e-01 -1.03562534e+00
-4.24179494e-01 -3.42145979e-01 -3.54214281e-01 7.80504525e-01
1.05992593e-01 -4.38394308e-01 1.28374422e+00 6.13597989e-01
-1.44309968e-01 -7.35824525e-01 -1.02586257e+00 -3.01597238e-01
-2.10158557e-01 -5.29381692e-01 5.50506234e-01 5.16945839e-01
-2.29349747e-01 1.85193241e-01 -3.98908854e-01 4.67362016e-01
8.60852122e-01 1.95936620e-01 8.08120728e-01 -9.82732475e-01
-4.86102462e-01 -1.40313163e-01 -5.31242490e-01 -1.77476466e+00
3.68735418e-02 -5.79422951e-01 2.76735932e-01 -1.98163629e+00
2.44503811e-01 -4.16727990e-01 2.58800462e-02 6.41239405e-01
-1.67032272e-01 5.00169456e-01 -3.53861228e-02 2.11171493e-01
-5.02521813e-01 6.19696915e-01 1.33933759e+00 1.64187793e-03
-3.27874243e-01 -5.00613689e-01 -6.38208270e-01 7.26945639e-01
3.91412675e-01 -6.49544179e-01 -4.55033600e-01 -8.89086545e-01
5.91455519e-01 1.78408861e-01 1.56707764e-01 -1.14278650e+00
3.74282300e-01 -1.34492323e-01 5.18344104e-01 -6.50271237e-01
6.42588019e-01 -9.13816035e-01 4.80937511e-02 4.49957490e-01
1.29567981e-01 -3.07653636e-01 6.90937415e-02 4.18154180e-01
-4.78945524e-01 -1.53689250e-01 7.12662935e-01 -2.78869212e-01
-9.45923388e-01 6.99338496e-01 2.11487025e-01 -2.16876611e-01
6.90486908e-01 -4.86104608e-01 -2.84661382e-01 -4.78493243e-01
-4.75753605e-01 2.36526027e-01 9.24486756e-01 1.34431511e-01
9.61556613e-01 -1.25992441e+00 -6.38831854e-01 5.53552866e-01
-8.98151891e-04 1.03547096e+00 3.77194256e-01 5.93927622e-01
-8.81697416e-01 2.25627482e-01 -2.47781456e-01 -8.74550164e-01
-7.64227808e-01 3.89490306e-01 2.60377824e-01 -2.13225111e-01
-8.98058832e-01 1.04065359e+00 6.84544623e-01 -7.76030272e-02
3.93531531e-01 -5.69855571e-01 1.03546783e-01 -2.01327980e-01
9.80481029e-01 1.80563256e-01 1.33672625e-01 -3.04331869e-01
-3.92329603e-01 9.89604950e-01 2.97788262e-01 -2.75126696e-01
1.57245195e+00 -4.20327127e-01 -3.09334368e-01 4.58929747e-01
1.52579057e+00 7.12586120e-02 -1.82650828e+00 -3.03644657e-01
-4.97594774e-01 -8.40472341e-01 5.32962441e-01 -2.59162515e-01
-1.52776206e+00 9.34981346e-01 4.85055685e-01 -3.02118540e-01
1.41327596e+00 -2.45839041e-02 9.78812218e-01 3.53231758e-01
6.62730038e-01 -6.27823472e-01 2.41601691e-02 6.61656082e-01
7.75000334e-01 -1.53532910e+00 3.29376221e-01 -5.91746569e-01
-2.05995739e-01 1.04789007e+00 5.75996578e-01 -2.76873499e-01
6.73864901e-01 5.24521887e-01 -5.51759042e-02 -1.61479130e-01
-7.36473501e-01 -3.32240909e-01 6.34983089e-03 5.89423776e-01
1.21643342e-01 -3.96456659e-01 7.19785467e-02 2.28056699e-01
-2.49481499e-01 2.71036953e-01 5.53000927e-01 9.27626848e-01
-6.11536741e-01 -1.11404085e+00 -1.57003820e-01 8.32395777e-02
-6.60277382e-02 -3.28067571e-01 -1.69194564e-01 2.58240104e-01
1.86725874e-02 8.14319313e-01 2.62112558e-01 -4.43072677e-01
3.11201662e-01 -3.96822035e-01 6.40025914e-01 -9.43705857e-01
-9.72176418e-02 9.22357291e-02 -1.47522047e-01 -1.46766365e+00
-4.19042259e-01 -2.73758739e-01 -1.54154909e+00 -4.31577355e-01
2.67571360e-01 -3.61930221e-01 7.69937813e-01 8.70598555e-01
2.45974511e-01 4.06636596e-01 6.88296497e-01 -1.35713899e+00
1.95757940e-01 -7.06638634e-01 -6.83870018e-01 1.71512738e-01
5.49122393e-01 -3.17415923e-01 -6.29373610e-01 -6.77413344e-02] | [8.844282150268555, -2.653597116470337] |
69fe7c73-2e21-4765-bb40-80df76a03e89 | ftfdnet-learning-to-detect-talking-face-video | 2307.03990 | null | https://arxiv.org/abs/2307.03990v1 | https://arxiv.org/pdf/2307.03990v1.pdf | FTFDNet: Learning to Detect Talking Face Video Manipulation with Tri-Modality Interaction | DeepFake based digital facial forgery is threatening public media security, especially when lip manipulation has been used in talking face generation, and the difficulty of fake video detection is further improved. By only changing lip shape to match the given speech, the facial features of identity are hard to be discriminated in such fake talking face videos. Together with the lack of attention on audio stream as the prior knowledge, the detection failure of fake talking face videos also becomes inevitable. It's found that the optical flow of the fake talking face video is disordered especially in the lip region while the optical flow of the real video changes regularly, which means the motion feature from optical flow is useful to capture manipulation cues. In this study, a fake talking face detection network (FTFDNet) is proposed by incorporating visual, audio and motion features using an efficient cross-modal fusion (CMF) module. Furthermore, a novel audio-visual attention mechanism (AVAM) is proposed to discover more informative features, which can be seamlessly integrated into any audio-visual CNN architecture by modularization. With the additional AVAM, the proposed FTFDNet is able to achieve a better detection performance than other state-of-the-art DeepFake video detection methods not only on the established fake talking face detection dataset (FTFDD) but also on the DeepFake video detection datasets (DFDC and DF-TIMIT). | ['Yufei zha', 'Wei Huang', 'Feihan Yang', 'Junwen Xiong', 'Peng Zhang', 'Ganglai Wang'] | 2023-07-08 | null | null | null | null | ['face-detection', 'face-swapping', 'optical-flow-estimation', 'talking-face-generation', 'face-generation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-6.70062825e-02 -1.85578957e-01 -1.16349302e-01 5.09073101e-02
-3.34902555e-01 -2.13336915e-01 4.11983460e-01 -7.55910456e-01
-3.59494388e-02 4.26933855e-01 1.34313241e-01 3.02433930e-02
2.92521656e-01 -4.33433414e-01 -5.14163792e-01 -8.56153727e-01
2.94957161e-01 -3.42662424e-01 8.08956474e-02 -2.78899610e-01
3.06691587e-01 5.87309897e-01 -1.89716160e+00 7.02500582e-01
6.18218303e-01 1.24866319e+00 -1.66792832e-02 3.04904401e-01
-7.11827278e-02 6.58117533e-01 -7.53500521e-01 -6.07852578e-01
1.17113061e-01 -5.49879313e-01 -3.97290975e-01 1.48947701e-01
8.29527259e-01 -9.30802763e-01 -9.71503913e-01 1.28166831e+00
7.63734996e-01 -2.74709493e-01 4.55156505e-01 -1.85896432e+00
-6.00039244e-01 1.36511531e-02 -6.49223745e-01 3.27808768e-01
3.84308130e-01 5.87921977e-01 1.23321563e-01 -1.20883977e+00
5.55200458e-01 1.71739447e+00 5.81439376e-01 9.22020495e-01
-6.07675552e-01 -1.30084944e+00 -1.26099467e-01 8.47479105e-01
-1.55547345e+00 -1.04142857e+00 1.04599762e+00 -3.34934771e-01
5.04621685e-01 5.67914620e-02 8.69937778e-01 1.48550749e+00
2.18022078e-01 8.56731355e-01 5.51961720e-01 -3.01761590e-02
-4.65724528e-01 2.28380822e-02 -4.88054395e-01 8.93242836e-01
2.67410636e-01 4.23187613e-01 -7.71600723e-01 -4.15095612e-02
6.10882163e-01 9.97575372e-02 -7.32181311e-01 2.63409942e-01
-9.59766984e-01 6.56461000e-01 1.94522724e-01 4.51829880e-01
-1.97389498e-01 4.04752135e-01 6.89787865e-01 4.74741369e-01
3.45997483e-01 -2.81708241e-02 5.02113327e-02 2.18502879e-02
-1.00416613e+00 -3.70937330e-03 3.42078328e-01 5.53931534e-01
5.10827243e-01 4.36455488e-01 -3.42287779e-01 5.86460471e-01
4.80843633e-01 7.13069320e-01 7.47371614e-01 -9.82797265e-01
4.89173740e-01 2.46111184e-01 9.53372866e-02 -1.77761090e+00
-8.07039440e-02 1.28661871e-01 -7.61966944e-01 1.32833347e-01
3.87914956e-01 8.03018436e-02 -6.61404371e-01 1.67412829e+00
4.05504733e-01 7.80413389e-01 -1.09075718e-01 1.14262557e+00
1.20874023e+00 5.57212293e-01 -2.34081134e-01 -7.04856575e-01
1.29105306e+00 -7.31874585e-01 -1.44278109e+00 7.99922943e-02
5.06112218e-01 -9.86747861e-01 6.61559224e-01 4.75290805e-01
-6.98361039e-01 -6.33633137e-01 -1.04427719e+00 -6.02747723e-02
-2.79026866e-01 2.83933312e-01 5.42048633e-01 1.01732993e+00
-8.97915125e-01 2.69314438e-01 -3.78414929e-01 3.25415446e-03
9.15851414e-01 3.13815653e-01 -7.11240888e-01 -3.40684086e-01
-1.40746355e+00 6.70211136e-01 2.53769249e-01 6.20282173e-01
-1.34422553e+00 -3.77797216e-01 -8.30268323e-01 -2.93218270e-02
3.15063685e-01 -4.67433095e-01 8.07276905e-01 -1.40689993e+00
-1.61608946e+00 8.78941655e-01 -2.78486133e-01 5.00349812e-02
7.89913952e-01 1.36090681e-01 -8.99277031e-01 6.39524877e-01
-8.09972435e-02 7.44069755e-01 1.75901818e+00 -1.15258014e+00
-2.57700711e-01 -4.57854897e-01 -3.68559211e-01 -1.92817152e-02
-5.75814128e-01 2.28542060e-01 -1.80682316e-01 -7.29619980e-01
-9.21403244e-02 -5.42732358e-01 6.84212029e-01 5.40829122e-01
-3.15723032e-01 -3.66752744e-01 1.89214540e+00 -8.43767285e-01
1.31757665e+00 -2.43496823e+00 -1.73979282e-01 -2.70346075e-01
4.44335401e-01 7.94407666e-01 -3.28811496e-01 1.11699343e-01
-2.30172038e-01 1.12547427e-01 2.02913254e-01 -3.40141058e-01
-3.87196779e-01 -1.24888197e-01 -2.86619604e-01 9.89511549e-01
1.69127405e-01 8.64682972e-01 -1.00187993e+00 -6.21601284e-01
2.94049084e-01 6.98109150e-01 -4.89243537e-01 1.68594290e-02
1.45504743e-01 4.22034264e-01 -2.13779688e-01 9.35226679e-01
1.23687518e+00 3.66535276e-01 -2.60848045e-01 -4.28646803e-01
8.92678723e-02 -7.03772679e-02 -9.14985955e-01 1.44890380e+00
-2.29011014e-01 9.78430569e-01 4.17239100e-01 -8.69158924e-01
7.57827520e-01 6.37163997e-01 5.32103121e-01 -5.45189321e-01
5.87597370e-01 2.81757146e-01 1.08367782e-02 -1.22843826e+00
9.49823409e-02 -2.81048473e-02 3.16526502e-01 3.12486917e-01
2.92942435e-01 1.66116342e-01 -4.31419581e-01 -5.96766593e-03
7.85372913e-01 -2.29101360e-01 -3.02625984e-01 1.05800755e-01
8.32646072e-01 -4.87828195e-01 6.26653671e-01 2.31678978e-01
-7.17921138e-01 2.93157935e-01 4.33717281e-01 -3.60499293e-01
-6.30283415e-01 -6.08715653e-01 -1.57008767e-01 5.86514354e-01
3.88885707e-01 -3.53824645e-01 -8.16454768e-01 -8.11616063e-01
1.50158823e-01 1.18350692e-01 -4.60201412e-01 -7.67011285e-01
-5.28805673e-01 -3.59926969e-01 1.11825323e+00 -9.83771458e-02
9.74134684e-01 -1.03192008e+00 1.38652697e-02 1.09926157e-03
-6.14379644e-01 -1.31715667e+00 -7.36330092e-01 -8.25518131e-01
-4.07790422e-01 -1.22258544e+00 -5.37077725e-01 -8.63169432e-01
3.54873091e-01 7.32255101e-01 2.69900173e-01 4.91593659e-01
-4.77683455e-01 7.11785182e-02 -1.05268158e-01 -2.82236278e-01
-6.27548575e-01 -7.38630176e-01 4.37746048e-01 6.50188684e-01
4.55330998e-01 -2.15891525e-01 -7.21208155e-01 4.98359829e-01
-8.53720963e-01 -2.08231881e-01 9.19948742e-02 1.13144052e+00
-8.40165243e-02 2.33898535e-01 7.39246488e-01 -2.73960736e-02
5.02593517e-01 -4.06459361e-01 -2.86945671e-01 -4.35551740e-02
-2.20224515e-01 -6.42292440e-01 5.23647666e-01 -7.51432180e-01
-1.00364327e+00 -2.31596336e-01 -1.59337878e-01 -1.29490185e+00
-1.21526860e-01 -1.04667984e-01 -4.58346993e-01 -5.05701244e-01
2.56309927e-01 3.85146171e-01 5.46607792e-01 -2.03111038e-01
4.05567363e-02 1.16971922e+00 5.00648618e-01 1.60424843e-01
6.27109766e-01 7.54810333e-01 -4.07991745e-02 -1.08511114e+00
-1.83613002e-01 -3.24722320e-01 -2.55954057e-01 -7.23430991e-01
7.33165920e-01 -9.19702709e-01 -1.43558240e+00 1.32655835e+00
-1.60538292e+00 2.84313053e-01 5.65417051e-01 5.51390946e-01
-4.33012545e-01 1.02022457e+00 -7.78420568e-01 -1.04727733e+00
-1.32401556e-01 -1.39352894e+00 1.34689522e+00 -7.01629221e-02
3.46724927e-01 -5.97718775e-01 -5.30401886e-01 7.46940076e-01
2.84162879e-01 1.88468754e-01 5.01585126e-01 -1.69948921e-01
-6.74571931e-01 -2.84336776e-01 -3.83770347e-01 6.71409190e-01
3.98022652e-01 2.10929319e-01 -1.36925209e+00 -3.86760652e-01
3.76099974e-01 -2.72754818e-01 1.10426545e+00 2.78784305e-01
1.31979764e+00 -6.17905736e-01 -4.22565490e-01 6.67266548e-01
9.70589399e-01 1.45347893e-01 8.80591094e-01 -2.10522950e-01
9.34992731e-01 7.40194380e-01 5.46594858e-01 3.06754112e-01
1.43691108e-01 9.40697670e-01 9.11177516e-01 7.16337934e-02
-4.07198638e-01 -2.10872784e-01 9.13427711e-01 5.31243086e-01
1.13758534e-01 -3.76290768e-01 -3.52415323e-01 3.14662308e-01
-1.66405582e+00 -1.41906297e+00 -2.35761464e-01 2.10582995e+00
4.98255581e-01 -4.59717363e-01 6.47504057e-04 4.63084817e-01
1.08833683e+00 1.76550686e-01 -4.39158082e-01 -1.94321811e-01
-2.24842444e-01 -2.37888277e-01 1.80755556e-01 3.44999433e-01
-1.14469111e+00 1.14492249e+00 5.46893835e+00 1.30694115e+00
-1.62971199e+00 4.94942456e-01 3.59766334e-01 -1.30006686e-01
5.56937195e-02 -5.05693436e-01 -6.40903592e-01 9.87627208e-01
7.24623263e-01 2.56389499e-01 5.45751393e-01 4.58698481e-01
7.81647980e-01 1.51875407e-01 -6.92463160e-01 1.57312083e+00
4.96720403e-01 -1.33847988e+00 6.11806437e-02 2.14313835e-01
1.14300951e-01 -4.65337664e-01 2.74193585e-01 -5.37423939e-02
-5.57943285e-01 -1.16896617e+00 5.62629282e-01 4.46491182e-01
1.21800148e+00 -8.51581931e-01 8.47270966e-01 2.76884049e-01
-1.04423797e+00 -4.03577119e-01 -3.71997684e-01 2.66113490e-01
1.12712443e-01 3.63669902e-01 -8.04403722e-01 3.46609563e-01
8.34161937e-01 1.01365101e+00 -1.66482434e-01 7.56603420e-01
1.78247616e-02 3.37976873e-01 -1.67983219e-01 2.76586980e-01
2.58410890e-02 1.53632805e-01 8.24561834e-01 1.01797926e+00
4.58543479e-01 -2.43027106e-01 -1.94036856e-01 8.01513493e-01
-2.70712435e-01 -7.96282664e-03 -1.01324379e+00 -1.56248122e-01
4.75877732e-01 1.14058149e+00 -1.72204167e-01 -2.49966606e-01
-2.71483034e-01 1.16342890e+00 -2.97976047e-01 1.65207610e-01
-9.54995334e-01 -2.98556328e-01 1.16673946e+00 1.13686770e-01
1.78225294e-01 1.25361919e-01 4.62279916e-01 -1.38790810e+00
9.38556716e-02 -8.55978429e-01 1.19679831e-01 -9.11048591e-01
-1.17252123e+00 3.63814682e-01 -4.33041453e-01 -1.41389072e+00
-1.23424428e-02 -6.96063340e-01 -5.21831155e-01 6.02134049e-01
-1.67208338e+00 -1.09891367e+00 -6.73752069e-01 1.13033450e+00
5.05427420e-01 -3.94627571e-01 5.02247095e-01 6.40932500e-01
-7.78355122e-01 1.01764059e+00 -1.76946461e-01 3.49411339e-01
9.19639289e-01 -1.53291956e-01 -2.39202425e-01 6.83988094e-01
-4.62861419e-01 2.89131314e-01 4.11603898e-01 -6.12095475e-01
-1.56097877e+00 -1.13134325e+00 6.44994259e-01 -1.88723087e-01
4.89806592e-01 -3.27009022e-01 -8.92516851e-01 1.73054665e-01
1.25889331e-01 4.19618189e-01 3.93409401e-01 -9.86922085e-01
-4.96630013e-01 -2.66491264e-01 -1.61572862e+00 2.34911144e-01
1.15439177e+00 -8.66878271e-01 -1.99146613e-01 4.35376763e-01
6.08168960e-01 -6.17906228e-02 -5.29366374e-01 4.26302493e-01
7.73672104e-01 -1.20678329e+00 9.72692847e-01 -5.05260408e-01
2.12032154e-01 -5.25830090e-01 6.42048270e-02 -9.97757614e-01
1.20578386e-01 -9.61928248e-01 -3.76511842e-01 1.22197795e+00
-3.72211993e-01 -7.21968532e-01 6.27939939e-01 -3.67406130e-01
-9.30757001e-02 -3.81970465e-01 -1.49204874e+00 -7.82866359e-01
-3.57957661e-01 -4.01100516e-01 4.67919141e-01 1.21539509e+00
-1.60954725e-02 -7.24291056e-02 -9.40202057e-01 1.64675549e-01
5.49144566e-01 -3.93552542e-01 5.69492996e-01 -1.12174797e+00
2.25046396e-01 -5.20929873e-01 -9.86703336e-01 -1.02111518e+00
4.80262429e-01 -6.77551746e-01 -2.39472479e-01 -6.66257679e-01
-7.19202757e-02 8.86201188e-02 5.29359803e-02 3.41098249e-01
1.28217429e-01 6.22730672e-01 1.40849456e-01 3.20694506e-01
-5.92663959e-02 9.26588714e-01 1.77155733e+00 -3.91512603e-01
1.80801451e-01 -1.32512495e-01 -2.85175949e-01 5.54160655e-01
3.53145689e-01 -4.49567616e-01 -1.07503012e-01 -1.30609870e-01
-2.94235200e-01 5.09913802e-01 8.01053286e-01 -8.15304875e-01
2.24287421e-01 8.80683735e-02 5.29927731e-01 -3.59624565e-01
7.55741298e-01 -8.63326311e-01 -2.45427072e-01 6.80643916e-01
2.37071961e-01 -7.73722753e-02 2.46176258e-01 7.44674683e-01
-4.44451362e-01 -1.98630895e-02 1.01143336e+00 2.46047620e-02
-4.96150702e-01 4.45423752e-01 -6.94963038e-01 -4.43638265e-01
1.10297012e+00 -7.32341886e-01 -6.20131314e-01 -5.39573431e-01
-5.25756419e-01 -7.83753470e-02 2.30216920e-01 7.60878086e-01
1.13215411e+00 -1.56288254e+00 -6.36098325e-01 5.55561960e-01
8.08343813e-02 -5.22269905e-01 5.94175756e-01 1.11605704e+00
-3.69362354e-01 2.27718771e-01 -3.59724760e-01 -6.15508318e-01
-1.49435484e+00 7.52053678e-01 7.12481797e-01 5.94698250e-01
-3.78665090e-01 7.68221676e-01 1.78839147e-01 1.11389734e-01
2.43616596e-01 5.51863313e-02 -2.87527740e-01 2.98815221e-01
7.55783021e-01 5.31896114e-01 8.96786228e-02 -1.21924949e+00
-4.01261628e-01 5.60250819e-01 1.62404343e-01 3.78334552e-01
7.94328809e-01 -3.40586960e-01 -2.68031776e-01 -1.24254867e-01
1.48755884e+00 -4.41465378e-02 -1.04943991e+00 1.00390986e-01
-6.05907679e-01 -9.52308416e-01 3.44549060e-01 -3.22757185e-01
-1.52839112e+00 1.08672094e+00 9.06148016e-01 1.66028533e-02
1.11433303e+00 -1.42167315e-01 1.10585403e+00 1.16786852e-01
3.33380342e-01 -8.71156693e-01 7.01635480e-01 1.38839141e-01
9.96202052e-01 -1.34952843e+00 -4.78209883e-01 -7.04307735e-01
-3.03131312e-01 1.45650601e+00 7.03790307e-01 1.49928421e-01
7.89781630e-01 2.61042383e-03 3.87372151e-02 -1.96913660e-01
-4.86826867e-01 4.97418828e-02 9.35322046e-02 6.60228074e-01
-8.50986987e-02 -3.67352486e-01 -1.10650234e-01 3.65845352e-01
-1.13111278e-02 7.98486322e-02 5.65494299e-01 4.03015256e-01
-3.61485511e-01 -6.16422951e-01 -5.33896565e-01 1.82057142e-01
-5.57396591e-01 1.43980491e-03 -4.91405427e-01 6.13917112e-01
5.82612991e-01 1.25627601e+00 3.23841423e-02 -7.01558352e-01
-1.33923367e-01 -1.21343352e-01 5.42079210e-01 -2.28128910e-01
-2.66442180e-01 1.17861472e-01 -1.82923064e-01 -9.33210373e-01
-4.52364624e-01 -4.37435776e-01 -8.07456970e-01 -7.44277716e-01
-7.35787094e-01 -2.19748452e-01 5.63853204e-01 9.80268300e-01
4.37596262e-01 1.68042570e-01 9.04628932e-01 -1.03953075e+00
-1.85081333e-01 -1.09112072e+00 -5.46776295e-01 3.73937368e-01
8.28417659e-01 -1.12261796e+00 -8.43957007e-01 -8.32264125e-02] | [13.02138614654541, 1.2203302383422852] |
86883909-e7be-4c83-bd18-3da24095c9f3 | graph-bert-only-attention-is-needed-for | 2001.05140 | null | https://arxiv.org/abs/2001.05140v2 | https://arxiv.org/pdf/2001.05140v2.pdf | Graph-Bert: Only Attention is Needed for Learning Graph Representations | The dominant graph neural networks (GNNs) over-rely on the graph links, several serious performance problems with which have been witnessed already, e.g., suspended animation problem and over-smoothing problem. What's more, the inherently inter-connected nature precludes parallelization within the graph, which becomes critical for large-sized graph, as memory constraints limit batching across the nodes. In this paper, we will introduce a new graph neural network, namely GRAPH-BERT (Graph based BERT), solely based on the attention mechanism without any graph convolution or aggregation operators. Instead of feeding GRAPH-BERT with the complete large input graph, we propose to train GRAPH-BERT with sampled linkless subgraphs within their local contexts. GRAPH-BERT can be learned effectively in a standalone mode. Meanwhile, a pre-trained GRAPH-BERT can also be transferred to other application tasks directly or with necessary fine-tuning if any supervised label information or certain application oriented objective is available. We have tested the effectiveness of GRAPH-BERT on several graph benchmark datasets. Based the pre-trained GRAPH-BERT with the node attribute reconstruction and structure recovery tasks, we further fine-tune GRAPH-BERT on node classification and graph clustering tasks specifically. The experimental results have demonstrated that GRAPH-BERT can out-perform the existing GNNs in both the learning effectiveness and efficiency. | ['Congying Xia', 'Haopeng Zhang', 'Li Sun', 'Jiawei Zhang'] | 2020-01-15 | null | null | null | null | ['graph-structure-learning'] | ['graphs'] | [-6.79161474e-02 4.85797286e-01 5.66698313e-02 -2.93228269e-01
-6.09410554e-02 -2.13161588e-01 1.91540003e-01 2.24320680e-01
-3.85288984e-01 6.13225579e-01 -3.58287275e-01 -3.44615579e-01
-1.05074823e-01 -1.18132269e+00 -7.76041090e-01 -8.33620429e-01
-5.63294888e-01 6.70612037e-01 4.03278112e-01 -1.55795351e-01
-2.80075252e-01 6.57005012e-01 -1.12235343e+00 -2.34820619e-01
1.07399249e+00 6.95377290e-01 3.68680149e-01 6.03254914e-01
-2.58249879e-01 8.70588958e-01 -3.68225336e-01 -5.11062980e-01
2.43295595e-01 -8.46450999e-02 -5.41990578e-01 2.23517060e-01
2.89166361e-01 -2.58276999e-01 -8.79866660e-01 1.22731745e+00
4.13450092e-01 2.54507065e-01 2.30127215e-01 -1.41857541e+00
-4.90090430e-01 8.88068378e-01 -5.60188591e-01 9.39048380e-02
-1.95379362e-01 3.66621345e-01 9.59046066e-01 -5.38973927e-01
5.47737598e-01 1.28574526e+00 7.96856821e-01 4.43789750e-01
-1.19538653e+00 -7.62197912e-01 6.28798425e-01 1.38313159e-01
-1.47736645e+00 -9.24798194e-03 1.07872248e+00 -2.14203015e-01
8.89256358e-01 1.20249629e-01 8.34263027e-01 8.61106873e-01
-8.22207332e-02 5.41004598e-01 5.28338790e-01 -4.42982204e-02
2.47391406e-02 -1.68523327e-01 3.63512784e-01 1.19292760e+00
4.72007096e-01 -1.82672068e-02 7.77438655e-02 1.76745892e-01
9.48601723e-01 5.61207756e-02 -4.46512043e-01 -4.19287801e-01
-9.20177281e-01 8.19947958e-01 1.13249731e+00 1.19796865e-01
-3.01873595e-01 4.55069959e-01 6.53402686e-01 3.46284270e-01
5.43447435e-01 2.51638293e-01 -2.55903810e-01 3.62187564e-01
-6.96427524e-01 -2.18509063e-01 8.85172606e-01 1.17594194e+00
1.11996043e+00 3.55890691e-01 -1.03037462e-01 5.85918546e-01
1.11262113e-01 9.05662030e-02 2.33508840e-01 -2.39725217e-01
5.23454070e-01 1.03416991e+00 -5.90897560e-01 -1.32217515e+00
-6.50967360e-01 -6.44979000e-01 -1.49914658e+00 -1.23878144e-01
2.47204751e-01 -3.24344963e-01 -1.06906712e+00 1.69545877e+00
4.29443926e-01 7.14006543e-01 -1.68637305e-01 8.75195026e-01
1.26175010e+00 6.33656263e-01 8.33824053e-02 -1.27352506e-01
1.01353180e+00 -1.31264699e+00 -5.10308087e-01 -2.17012957e-01
8.14307749e-01 -1.04038835e-01 1.13529444e+00 1.43351734e-01
-7.66696215e-01 -5.41948855e-01 -1.03594327e+00 3.38689983e-02
-4.10913169e-01 2.85021029e-02 1.06460285e+00 3.08332026e-01
-1.26865792e+00 9.57626224e-01 -1.07186341e+00 -3.17534447e-01
5.74388385e-01 5.63863635e-01 -5.72081804e-01 -9.68680531e-02
-9.20694232e-01 3.86686921e-01 7.95224369e-01 4.79689837e-01
-9.61870611e-01 -3.13232929e-01 -1.00729167e+00 4.72440779e-01
7.34535456e-01 -5.33198893e-01 7.24652708e-01 -1.02964497e+00
-1.34910166e+00 5.03021538e-01 4.02792484e-01 -3.13729793e-01
2.92110085e-01 2.30201811e-01 -5.14124155e-01 2.12944210e-01
-2.20021024e-01 4.98155415e-01 6.62795484e-01 -1.00864697e+00
-1.48679614e-01 -2.60060638e-01 3.73912007e-01 2.10268229e-01
-6.31663859e-01 -3.15773129e-01 -9.42058563e-01 -6.07817471e-01
1.55489566e-02 -8.99113119e-01 -4.11908448e-01 -1.07723288e-01
-6.49734497e-01 -3.80809426e-01 9.13649857e-01 -4.73948002e-01
1.35874331e+00 -2.07320714e+00 1.62272304e-01 2.65515387e-01
7.57819176e-01 5.05711436e-01 -3.66806000e-01 4.38802958e-01
-1.60939589e-01 3.47258300e-02 -2.96832830e-01 -1.24770172e-01
-1.86365664e-01 4.56462622e-01 1.57643840e-01 5.51407933e-01
3.07280421e-01 1.12398291e+00 -8.97107303e-01 -7.58317471e-01
1.86120674e-01 3.34329277e-01 -6.55582607e-01 3.41770977e-01
-4.05161172e-01 2.15335339e-01 -5.18170655e-01 4.75105405e-01
9.33038950e-01 -8.78924847e-01 3.84100318e-01 -3.02670568e-01
4.04356509e-01 2.48816162e-02 -1.13233232e+00 1.55149472e+00
-3.48167717e-01 4.66214210e-01 4.37163919e-01 -1.48611975e+00
9.21654463e-01 2.75507453e-04 3.26005131e-01 -2.99925029e-01
6.97524771e-02 -1.06331073e-01 2.98878223e-01 -5.11158824e-01
1.34250835e-01 1.73255607e-01 1.90834656e-01 2.42772013e-01
2.76296198e-01 3.72407913e-01 4.55511749e-01 5.97894788e-01
1.50213361e+00 6.65647686e-02 6.95447177e-02 -3.93998116e-01
5.39131880e-01 -4.67071161e-02 5.31629205e-01 4.96070236e-01
6.67617321e-02 4.29110438e-01 6.70929015e-01 -4.35572743e-01
-5.91799021e-01 -7.39193380e-01 3.17920089e-01 1.12077963e+00
3.04969132e-01 -6.80894732e-01 -7.15729654e-01 -1.13781178e+00
-6.96104905e-03 2.26709321e-01 -3.55384439e-01 -2.82161504e-01
-7.49686897e-01 -8.05410564e-01 4.12862122e-01 5.98261476e-01
6.32330477e-01 -1.29015982e+00 -6.20911121e-02 4.66674834e-01
2.68601954e-01 -1.13410938e+00 -5.45669138e-01 2.86204755e-01
-8.75636220e-01 -1.26307082e+00 -3.54889929e-01 -1.07009888e+00
1.08626211e+00 4.09169674e-01 1.20251000e+00 8.41953695e-01
-1.75427333e-01 4.77974629e-03 -3.07172418e-01 3.21219079e-02
-1.97754949e-01 4.45681632e-01 -3.35578531e-01 1.09982267e-01
2.18909010e-02 -1.02258921e+00 -4.53503907e-01 2.55984008e-01
-7.29921579e-01 4.30511057e-01 6.97075725e-01 1.04485202e+00
3.60892296e-01 3.92571390e-01 6.09688699e-01 -1.43008244e+00
5.15447676e-01 -3.47839147e-01 -8.06103289e-01 1.89812973e-01
-5.68160117e-01 1.58687860e-01 9.64804530e-01 -5.27340829e-01
-6.65676177e-01 -2.09568515e-02 -2.09862784e-01 -7.62428701e-01
1.33325249e-01 9.54617441e-01 -5.70703149e-01 -1.91072851e-01
4.39398885e-01 -2.45163497e-02 4.31133322e-02 -4.40963119e-01
2.53963858e-01 3.62188369e-01 4.94739175e-01 -4.64283705e-01
9.95701015e-01 3.09879780e-01 1.76614165e-01 -7.53116608e-01
-5.81719935e-01 -2.99412429e-01 -4.94962543e-01 -1.41501352e-01
6.79294586e-01 -8.56192887e-01 -8.08366776e-01 5.35245240e-01
-9.99180913e-01 -7.69787312e-01 -1.49555011e-02 2.66398311e-01
-1.20143301e-03 5.67014694e-01 -9.07503188e-01 -4.98326331e-01
-5.16556323e-01 -1.12032032e+00 7.95158207e-01 2.58320183e-01
3.76153737e-01 -1.25639260e+00 -3.91091764e-01 -1.73353106e-01
3.93833369e-01 4.56524968e-01 1.02466381e+00 -7.75650501e-01
-8.16662908e-01 -2.09103152e-01 -6.64304256e-01 1.80166200e-01
8.94859880e-02 -7.40651786e-02 -7.12672949e-01 -7.42721319e-01
-4.90381956e-01 -3.21993351e-01 8.52997720e-01 1.13505125e-01
1.51870680e+00 -5.09399533e-01 -5.83184242e-01 1.03669763e+00
1.44491017e+00 -2.06299737e-01 3.61850142e-01 -7.38535589e-03
1.44480968e+00 4.66552287e-01 2.38118365e-01 6.39703348e-02
3.30526084e-01 3.48867297e-01 7.64453232e-01 -4.28928405e-01
-2.35164613e-01 -3.84870499e-01 1.12409770e-01 1.12712491e+00
-1.57913163e-01 -5.83923578e-01 -7.48726368e-01 1.68201998e-01
-2.01239824e+00 -5.51803589e-01 -2.59368122e-01 1.97120571e+00
4.70837444e-01 2.19654828e-01 4.90650497e-02 -1.46070078e-01
1.06092596e+00 3.40151876e-01 -7.55150437e-01 -2.48598494e-02
1.10102996e-01 1.41734749e-01 5.88814974e-01 3.62450331e-01
-1.12909436e+00 1.10426879e+00 4.77373981e+00 9.63190675e-01
-1.28265154e+00 -4.10737619e-02 5.50032556e-01 3.00490052e-01
-2.28175029e-01 1.84017271e-01 -5.25897086e-01 4.01617289e-01
7.91934371e-01 -2.05226094e-01 7.73330092e-01 9.45204675e-01
-1.02657072e-01 4.06008989e-01 -1.12842178e+00 1.00500953e+00
-2.36031443e-01 -1.22846246e+00 -9.21307132e-03 1.99737936e-01
3.48281026e-01 1.02747165e-01 -5.49245596e-01 8.45631421e-01
5.77347755e-01 -1.18760383e+00 1.37541950e-01 7.98306167e-02
9.67690110e-01 -6.43663824e-01 7.31130540e-01 5.59424281e-01
-1.62542570e+00 1.46114185e-01 -6.81968868e-01 -2.76383832e-02
8.06997251e-03 6.66011155e-01 -8.24402571e-01 9.56729531e-01
5.07735372e-01 9.29363906e-01 -6.95930421e-01 9.83004272e-01
-3.45930487e-01 7.74466276e-01 -5.15984952e-01 -1.23393066e-01
3.68916988e-01 -4.99368161e-01 4.42611903e-01 1.16648233e+00
2.42908090e-01 2.85225268e-02 6.46216512e-01 8.63173306e-01
-3.42209965e-01 6.10801131e-02 -5.17285109e-01 -3.90151441e-01
4.40934658e-01 1.75699401e+00 -1.00938845e+00 -4.77576256e-01
-5.41532457e-01 8.85414004e-01 9.30799246e-01 5.29367566e-01
-9.48076129e-01 -5.61197340e-01 2.63787657e-01 2.20527291e-01
3.07872176e-01 -1.76268786e-01 2.25940123e-01 -1.15838563e+00
-5.94885759e-02 -7.19337940e-01 5.76827288e-01 -6.51687920e-01
-1.41742313e+00 9.61569011e-01 -1.12730958e-01 -8.26979160e-01
1.76821753e-01 -6.54740572e-01 -8.94589007e-01 6.46190226e-01
-1.33154607e+00 -1.33803475e+00 -8.16790819e-01 7.30973542e-01
1.47735044e-01 -9.71436426e-02 6.26344323e-01 4.53570694e-01
-1.01373720e+00 7.15144813e-01 -2.68935263e-01 5.67688167e-01
3.55520934e-01 -1.40090680e+00 5.74443281e-01 8.84106159e-01
8.61044675e-02 5.55887938e-01 2.78928757e-01 -7.95912385e-01
-1.61129510e+00 -1.56043482e+00 3.05797428e-01 2.70534694e-01
8.25600624e-01 -7.04645872e-01 -1.25770330e+00 8.99676383e-01
9.40389261e-02 6.55092359e-01 5.89515716e-02 2.45369509e-01
-1.43233329e-01 -3.38908136e-01 -9.62127924e-01 6.28711462e-01
1.51038086e+00 -2.16568962e-01 5.14729097e-02 6.40532315e-01
1.09754264e+00 -5.49078226e-01 -8.14510405e-01 4.58278745e-01
-2.72964500e-02 -7.90217519e-01 8.37856233e-01 -7.13098228e-01
5.59988320e-02 -2.51154274e-01 2.57854223e-01 -1.38740575e+00
-5.13827801e-01 -7.76506960e-01 -1.05590492e-01 1.36785638e+00
1.60192862e-01 -8.61225188e-01 1.11076927e+00 4.13521916e-01
-3.40635478e-01 -8.46421182e-01 -6.41903043e-01 -7.25279748e-01
-3.66473973e-01 -1.08027749e-01 8.68829668e-01 9.95497346e-01
-2.80201644e-01 6.97672844e-01 -2.24365428e-01 3.98052573e-01
4.80006129e-01 2.19986737e-01 1.03117442e+00 -1.40177238e+00
-6.66432738e-01 -4.29730773e-01 -5.41534543e-01 -1.09943604e+00
3.92967314e-01 -1.43179929e+00 -1.41990811e-01 -1.58644021e+00
3.83472703e-02 -7.01480389e-01 -2.53758967e-01 6.86212718e-01
-4.54441845e-01 -1.01650640e-01 1.03297129e-01 -1.34625835e-02
-6.76197827e-01 6.51389897e-01 1.59728730e+00 -3.06986868e-01
-2.73379087e-01 -2.50758678e-02 -5.06321847e-01 7.01117575e-01
6.58381879e-01 -3.65867883e-01 -7.49769747e-01 -3.19366634e-01
1.49820289e-02 2.00877354e-01 4.02220011e-01 -9.95504797e-01
2.80970067e-01 -4.24003452e-02 -1.74148474e-02 -4.50710982e-01
9.90435258e-02 -9.28969681e-01 3.66718918e-01 3.39554161e-01
-6.43589254e-03 2.22762719e-01 1.05950408e-01 8.02076340e-01
-9.95818824e-02 -1.05040483e-01 6.60612702e-01 -2.07973421e-01
-8.06127250e-01 8.59823763e-01 1.41088799e-01 1.40783042e-01
9.76851881e-01 -1.31176144e-01 -4.45117712e-01 -2.44356200e-01
-7.86727548e-01 4.80454594e-01 3.03469718e-01 5.04375882e-02
4.87585366e-01 -1.29941034e+00 -5.20940185e-01 3.66199255e-01
2.55460739e-02 6.84883058e-01 2.58400351e-01 8.65041494e-01
-7.50442922e-01 -9.98236537e-02 1.14456527e-01 -5.94227910e-01
-1.15703428e+00 1.02004039e+00 3.76917213e-01 -6.08199298e-01
-1.14315200e+00 9.61213946e-01 6.45197749e-01 -6.25242114e-01
4.39431459e-01 -3.29987884e-01 -9.88232121e-02 -3.33256304e-01
2.69414261e-02 2.00817212e-01 6.22964837e-02 -3.65960658e-01
-2.04079852e-01 1.84225321e-01 -1.19872965e-01 5.17113984e-01
1.43512380e+00 1.41380966e-01 -3.57898861e-01 5.74045330e-02
1.22186494e+00 -3.62391829e-01 -1.40142488e+00 -2.35845640e-01
-4.93560284e-02 -2.53390640e-01 1.17079131e-01 -3.66358519e-01
-1.75778317e+00 8.88578296e-01 9.56387818e-02 3.49560022e-01
1.16782546e+00 -8.32099840e-02 8.09954822e-01 5.21430671e-01
3.71800393e-01 -8.10397387e-01 2.40509994e-02 2.40644798e-01
8.05850148e-01 -1.17323172e+00 5.70186004e-02 -8.06487262e-01
-3.15447032e-01 1.00217927e+00 1.03746057e+00 -5.03819406e-01
8.39030385e-01 2.93854833e-01 -3.33218753e-01 -4.77073848e-01
-7.04265714e-01 -2.59695292e-01 1.57134265e-01 5.73890626e-01
1.81792483e-01 1.63974002e-01 1.38262417e-02 5.35770535e-01
-7.36281723e-02 -2.78351039e-01 3.19866955e-01 5.23364007e-01
-2.79122859e-01 -9.19461370e-01 1.75919458e-01 8.16474319e-01
-4.34506834e-02 -1.96104810e-01 -2.77423054e-01 1.21729982e+00
-5.85787296e-02 6.04081452e-01 -3.49800177e-02 -5.22510052e-01
1.73580974e-01 -3.63779008e-01 2.23327518e-01 -7.25251138e-01
-6.70173764e-01 1.04478180e-01 3.46535623e-01 -6.10768437e-01
-4.85443622e-02 -8.01785067e-02 -1.39251697e+00 -5.25819421e-01
-6.77617073e-01 3.03403258e-01 2.45555565e-01 5.22278607e-01
4.43021953e-01 8.93355370e-01 3.19202602e-01 -8.63513410e-01
-4.42365646e-01 -1.15131640e+00 -6.55677378e-01 4.57665503e-01
1.20205343e-01 -4.96977210e-01 -4.11497414e-01 -3.78167003e-01] | [7.220090389251709, 6.1733245849609375] |
f0090045-9c31-47e6-8bb3-3fd3e2d4eb9d | free-supervision-from-video-games | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Krahenbuhl_Free_Supervision_From_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Krahenbuhl_Free_Supervision_From_CVPR_2018_paper.pdf | Free Supervision From Video Games | Deep networks are extremely hungry for data. They devour hundreds of thousands of labeled images to learn robust and semantically meaningful feature representations. Current networks are so data hungry that collecting labeled data has become as important as designing the networks themselves. Unfortunately, manual data collection is both expensive and time consuming. We present an alternative, and show how ground truth labels for many vision tasks are easily extracted from video games in real time as we play them. We interface the popular Microsoft DirectX rendering API, and inject specialized rendering code into the game as it is running. This code produces ground truth labels for instance segmentation, semantic labeling, depth estimation, optical flow, intrinsic image decomposition, and instance tracking. Instead of labeling images, a researcher now simply plays video games all day long. Our method is general and works on a wide range of video games. We collected a dataset of 220k training images, and 60k test images across 3 video games, and evaluate state of the art optical flow, depth estimation and intrinsic image decomposition algorithms. Our video game data is visually closer to real world images, than other synthetic dataset. | ['Philipp Krähenbühl'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['intrinsic-image-decomposition'] | ['computer-vision'] | [-2.64999773e-02 -1.43584251e-01 -1.07984170e-01 -3.96319747e-01
-5.66969931e-01 -9.98695254e-01 3.41169894e-01 -4.40558940e-01
-6.89456284e-01 6.04629874e-01 -7.02667385e-02 -3.52283001e-01
2.97170907e-01 -8.34687889e-01 -6.64464653e-01 -3.62885207e-01
-1.25040829e-01 7.15753734e-01 5.63381255e-01 -1.71819955e-01
2.61268079e-01 3.40377122e-01 -1.73690045e+00 9.95644480e-02
4.70332265e-01 9.44619536e-01 6.80499151e-02 1.20651150e+00
-1.29583269e-01 1.39667296e+00 -6.88534856e-01 -6.49291396e-01
7.28788793e-01 -3.84648085e-01 -1.09382176e+00 3.32929552e-01
8.23512077e-01 -8.44246030e-01 -5.28124690e-01 1.11376452e+00
3.76344502e-01 4.41180974e-01 1.96216866e-01 -1.84818363e+00
-3.48707080e-01 2.19779968e-01 -5.86819768e-01 4.76285696e-01
3.36764723e-01 4.02534872e-01 6.50484920e-01 -6.22717679e-01
7.79093802e-01 1.05675054e+00 3.76783043e-01 9.65260565e-01
-1.03105545e+00 -8.94680202e-01 -8.66917223e-02 9.43295434e-02
-9.99642074e-01 -5.12590766e-01 5.18767357e-01 -7.30987668e-01
7.05051124e-01 8.04824010e-02 1.11820149e+00 1.09747791e+00
-1.44427314e-01 6.99856758e-01 8.65433633e-01 -6.50463179e-02
4.43809450e-01 -6.78540543e-02 1.98614411e-02 1.07889807e+00
-2.44106129e-02 3.36110920e-01 -7.44240165e-01 1.05768025e-01
1.30722034e+00 -1.68129176e-01 -3.55971128e-01 -4.57257479e-01
-1.16476202e+00 7.59123802e-01 1.64132923e-01 -5.23351014e-01
-4.23453003e-02 5.80941975e-01 5.44336259e-01 2.64726728e-01
3.73351514e-01 3.09964329e-01 -5.16261816e-01 -8.99356306e-01
-1.03960896e+00 5.55753052e-01 8.08930278e-01 9.73019063e-01
1.01612628e+00 1.50657907e-01 2.92934239e-01 6.09958708e-01
9.91183892e-02 2.56375462e-01 5.32334983e-01 -2.05609083e+00
2.08388939e-01 4.65457290e-01 4.71799560e-02 -1.17422664e+00
-2.89001942e-01 9.65723172e-02 -5.27430832e-01 7.75578737e-01
9.91673470e-01 -3.94159943e-01 -9.55987155e-01 1.68546188e+00
2.67864764e-01 5.97602725e-01 -1.81827351e-01 1.33152235e+00
9.44932997e-01 4.40402716e-01 7.72067681e-02 3.15529704e-01
1.12196183e+00 -1.13559043e+00 -2.45063648e-01 -3.93830299e-01
5.77709377e-01 -7.10787892e-01 1.46628630e+00 6.96209431e-01
-1.23695612e+00 -4.93435115e-01 -8.12722445e-01 -3.43385965e-01
-7.55407512e-02 -1.80845797e-01 1.17944443e+00 9.51997638e-01
-1.31505251e+00 6.01442337e-01 -9.49009717e-01 -1.33485392e-01
6.23327792e-01 3.91483277e-01 -5.25242448e-01 -2.36215636e-01
-8.27543378e-01 4.15980190e-01 9.86059979e-02 -2.01657176e-01
-1.38208330e+00 -8.55672598e-01 -9.55828905e-01 -3.32757920e-01
3.92550528e-01 -5.99541903e-01 1.40531981e+00 -1.19859970e+00
-1.44388008e+00 1.30331516e+00 1.68447509e-01 -2.39149898e-01
6.05636537e-01 -1.00911170e-01 5.67733981e-02 4.42438155e-01
2.33800054e-01 1.15094101e+00 7.65526712e-01 -1.06001663e+00
-7.09173739e-01 -2.74309069e-01 6.74910367e-01 1.57082468e-01
-1.24353729e-02 1.50940955e-01 -6.45871580e-01 -4.26643938e-01
-8.53668526e-02 -8.36212516e-01 -2.78942943e-01 3.55008304e-01
-2.68198997e-01 1.28871799e-01 7.99430430e-01 -4.60904270e-01
5.39837360e-01 -2.01011705e+00 -1.11169368e-01 6.02101721e-02
8.33630919e-01 1.33883432e-01 -1.20055787e-01 -3.16101968e-01
-4.54619750e-02 5.08569255e-02 7.09403381e-02 -2.72082746e-01
-1.67393625e-01 4.52023506e-01 -4.50422652e-02 3.63398671e-01
-1.93391889e-01 9.56741810e-01 -1.13858414e+00 -5.78644097e-01
4.59413350e-01 3.41557473e-01 -9.60373759e-01 1.24456465e-01
-2.26172626e-01 5.30952632e-01 -5.90398014e-02 6.13808930e-01
5.47608793e-01 -3.63488704e-01 -1.25010997e-01 5.47024496e-02
2.84653246e-01 1.33720756e-01 -1.15429854e+00 2.05044937e+00
-3.50653529e-01 1.15993857e+00 5.09886593e-02 -9.53726649e-01
7.17354953e-01 3.44902538e-02 7.02734232e-01 -6.62719727e-01
3.91488701e-01 -1.91282108e-01 -1.70966357e-01 -5.62734008e-01
6.39223337e-01 1.37809977e-01 3.99489328e-02 7.64554024e-01
1.94766030e-01 -3.62518102e-01 4.43065524e-01 6.18999541e-01
1.28660178e+00 3.38444889e-01 -2.93955088e-01 -1.57858834e-01
4.94732410e-02 4.73910570e-01 6.61888480e-01 4.39695030e-01
-5.67649722e-01 8.09189200e-01 7.88895190e-01 -9.94016767e-01
-1.30556178e+00 -1.10009980e+00 4.00176406e-01 1.38857281e+00
3.69789094e-01 -5.62261641e-01 -1.08920956e+00 -4.54364151e-01
-4.38134253e-01 9.18220207e-02 -6.02845252e-01 -4.89256717e-02
-4.53926235e-01 -3.51073027e-01 6.49433196e-01 6.15523219e-01
7.41791964e-01 -1.23141086e+00 -8.51096988e-01 1.05193533e-01
-2.65846848e-01 -1.46750653e+00 -6.14358068e-01 -2.23493919e-01
-6.82564855e-01 -1.47584653e+00 -4.53697294e-01 -7.51989365e-01
5.11535764e-01 5.18498957e-01 1.64360404e+00 2.08003432e-01
-5.89406788e-01 4.99752343e-01 -1.83349609e-01 -9.47712287e-02
-2.20382333e-01 -1.38522685e-01 -7.21845925e-02 -2.59260207e-01
5.24794579e-01 -6.40661240e-01 -6.75931513e-01 4.11500663e-01
-8.46205533e-01 2.15262488e-01 -1.09103426e-01 3.97434205e-01
5.12732685e-01 -1.03064828e-01 -9.43935513e-02 -9.01603460e-01
6.32864177e-01 -1.07538141e-01 -7.97666788e-01 -1.63770631e-01
-5.88136427e-02 -2.88467020e-01 4.59487319e-01 -3.37815553e-01
-7.46635973e-01 2.15712622e-01 -5.98268490e-03 -6.85878456e-01
-3.06969732e-01 -1.18497923e-01 1.03839803e-02 -2.53163069e-01
1.01104939e+00 -5.24568781e-02 2.20714539e-01 -9.34241861e-02
5.22377253e-01 3.52668464e-01 9.67541218e-01 -8.69024098e-01
6.06276989e-01 6.69072509e-01 -2.85172880e-01 -7.36923277e-01
-7.01247573e-01 -3.13539326e-01 -5.32744110e-01 -6.22149885e-01
9.11366642e-01 -9.69735324e-01 -1.09625363e+00 6.51321888e-01
-1.01079261e+00 -9.26966012e-01 -6.31099284e-01 4.63689893e-01
-8.06853056e-01 1.24258339e-01 -7.34848559e-01 -1.92091182e-01
-6.40301928e-02 -1.52144933e+00 9.46905136e-01 3.74198347e-01
-1.42054334e-01 -1.05904877e+00 8.12288076e-02 5.81118584e-01
1.34219632e-01 2.64277846e-01 2.81547248e-01 7.32357129e-02
-8.33436012e-01 -1.81676835e-01 -6.34859324e-01 4.10875171e-01
-1.93698108e-02 3.27847838e-01 -1.07534242e+00 -4.12210599e-02
-2.81409085e-01 -8.26511085e-01 5.34833670e-01 5.85931242e-01
1.47836530e+00 -6.69791698e-02 6.35385737e-02 1.07377934e+00
1.30392075e+00 3.73429023e-02 8.15659761e-01 2.60463536e-01
1.03364158e+00 5.50648987e-01 3.99411142e-01 3.15140724e-01
5.70969224e-01 4.39363956e-01 6.45960689e-01 -1.86638415e-01
-7.58446977e-02 -1.26821712e-01 2.10981712e-01 4.21557277e-01
-4.58259135e-01 -7.94845540e-03 -1.07067716e+00 3.85000318e-01
-1.62411320e+00 -9.42668617e-01 -5.21205515e-02 2.01874518e+00
6.60870492e-01 2.83530504e-01 5.05780816e-01 1.00252710e-01
5.62058926e-01 3.13631794e-03 -4.73382175e-01 -3.07508677e-01
1.84869036e-01 3.73267472e-01 5.94598770e-01 6.52038753e-01
-1.05840373e+00 1.31777239e+00 6.73311758e+00 6.59394205e-01
-1.14739633e+00 1.31654277e-01 9.59559083e-01 -4.11541671e-01
-1.54284328e-01 -5.28559461e-02 -4.37484831e-01 3.03407401e-01
8.12501967e-01 9.45332868e-04 7.56581783e-01 1.00215089e+00
3.37624431e-01 -2.92174697e-01 -9.82174814e-01 1.51335967e+00
-9.76362377e-02 -1.67795086e+00 -1.74537867e-01 2.80647814e-01
6.69746637e-01 2.29889899e-01 -1.38831735e-01 2.02094782e-02
9.32606518e-01 -1.14566374e+00 7.44224727e-01 1.88161388e-01
9.58950520e-01 -6.23012125e-01 2.50459045e-01 1.61194861e-01
-9.77149546e-01 2.16542527e-01 -6.21836066e-01 -4.35432643e-01
9.13020000e-02 4.02574331e-01 -3.35848719e-01 -2.72772610e-01
8.14829588e-01 8.78994167e-01 -6.84690237e-01 1.06213593e+00
-2.69160569e-01 4.45579201e-01 -3.33093107e-01 2.38580689e-01
4.27072525e-01 -3.18375468e-01 8.99508297e-02 7.54363537e-01
-1.49372458e-01 -1.09650269e-02 2.99790889e-01 7.13958323e-01
-1.84458658e-01 -1.07294016e-01 -7.25102782e-01 5.57939224e-02
2.69255221e-01 1.46908569e+00 -1.29427278e+00 -5.00041306e-01
-3.26716751e-01 9.82934952e-01 1.51124388e-01 3.69644225e-01
-9.68275964e-01 -2.91290849e-01 1.19788110e+00 8.12699571e-02
-2.79753327e-01 -4.54734564e-01 -8.22783411e-02 -1.38585067e+00
-1.87761396e-01 -8.82265031e-01 7.86336586e-02 -1.10383224e+00
-8.81844819e-01 6.17710710e-01 -2.37349778e-01 -1.19928432e+00
-4.37156409e-01 -7.93722451e-01 -5.09285152e-01 3.84670675e-01
-1.09114361e+00 -5.77031255e-01 -9.36407328e-01 1.01647866e+00
6.24646246e-01 -1.55426010e-01 6.02905154e-01 6.23849154e-01
-4.82117712e-01 4.00661767e-01 -2.86650956e-01 7.73032248e-01
4.65941012e-01 -1.18219757e+00 7.65967071e-01 7.21820831e-01
5.24690688e-01 2.20782340e-01 4.36361432e-01 -1.74713165e-01
-1.04445732e+00 -8.37756753e-01 4.54985760e-02 -5.74095011e-01
5.26482761e-01 -5.62979341e-01 -5.62114179e-01 7.41587043e-01
-4.65293555e-03 6.74056113e-01 5.68142951e-01 -2.80422658e-01
-2.97441304e-01 -1.53465932e-02 -1.03350019e+00 5.18783629e-01
1.31714332e+00 -5.01768768e-01 1.24893948e-01 4.36945796e-01
6.56641304e-01 -7.63728440e-01 -4.34692234e-01 -2.33848780e-01
5.68345547e-01 -1.40658450e+00 1.20653701e+00 -8.44324768e-01
8.31008077e-01 -3.29548001e-01 -1.27426237e-01 -1.10466564e+00
1.84474900e-01 -6.44291699e-01 1.58770055e-01 8.91752541e-01
1.92577988e-01 -2.50969380e-01 1.75387907e+00 1.03097451e+00
6.16966262e-02 -4.78574008e-01 -5.18071592e-01 -4.99836981e-01
-7.82513097e-02 -9.75629091e-01 5.27433157e-01 1.16785717e+00
-2.42643863e-01 1.22691631e-01 -2.20471323e-01 -2.39504144e-01
9.15918767e-01 -1.73451155e-01 1.25282145e+00 -1.11343992e+00
-3.91858608e-01 -3.60840261e-01 -9.74513888e-01 -1.15566587e+00
1.95896417e-01 -6.67265058e-01 -1.31614491e-01 -1.26942599e+00
-5.94746247e-02 -6.47490919e-01 3.46422195e-01 5.44032872e-01
1.86913297e-01 9.75259066e-01 1.64354056e-01 7.15641305e-02
-6.42463803e-01 4.16441076e-02 1.58408010e+00 -1.06628120e-01
-1.10575758e-01 -2.02911481e-01 -6.65224731e-01 1.05317152e+00
9.10279274e-01 -3.12547803e-01 -8.92673254e-01 -7.56142080e-01
4.31182951e-01 1.61159977e-01 7.38766253e-01 -1.15288007e+00
2.35531047e-01 -2.87334055e-01 4.00601864e-01 -3.75280343e-02
4.60259318e-01 -5.36307395e-01 -9.90643725e-02 2.01638997e-01
-1.70609906e-01 7.67806321e-02 9.62514505e-02 1.63193747e-01
-1.77306175e-01 -1.93986923e-01 7.67578661e-01 -6.22620821e-01
-1.28255641e+00 7.00566649e-01 -2.87311673e-01 8.05726588e-01
1.01649284e+00 -6.10463202e-01 -4.70684528e-01 -7.19711244e-01
-8.94800067e-01 5.79355881e-02 7.79480577e-01 3.32111955e-01
6.95772588e-01 -1.05863190e+00 -3.49845588e-01 3.34752589e-01
-1.72748148e-01 3.14194202e-01 1.27241418e-01 3.44225019e-01
-1.38236523e+00 -1.66042939e-01 -6.11782432e-01 -7.18186319e-01
-1.08274531e+00 3.04924697e-01 6.23888373e-01 2.84580350e-01
-8.77101481e-01 1.27541053e+00 3.50475252e-01 -3.13479602e-01
1.94697261e-01 -3.47745240e-01 6.09555729e-02 -1.45778701e-01
7.64426112e-01 4.00551260e-01 -4.16759551e-02 -5.45281112e-01
-1.70879170e-01 6.83469534e-01 1.33958504e-01 -2.58071721e-01
1.10080481e+00 -1.12729080e-01 7.34084547e-02 -1.11746090e-02
1.26516783e+00 -4.32654351e-01 -1.87064099e+00 1.07961290e-01
-5.06714523e-01 -1.00074530e+00 1.75211161e-01 -3.25683802e-01
-1.86459255e+00 1.20912421e+00 5.14745533e-01 1.89951614e-01
1.12587512e+00 -7.15114996e-02 7.65237331e-01 2.38177463e-01
6.73297524e-01 -9.83920872e-01 4.32476521e-01 4.19483840e-01
2.07899362e-01 -1.34332573e+00 -2.87851363e-01 -4.00884300e-01
-6.21886194e-01 8.78702283e-01 8.88864040e-01 -5.23704827e-01
5.06050944e-01 5.56818426e-01 4.85504210e-01 -4.35790598e-01
-5.65606952e-01 -3.72560918e-02 -3.22254896e-01 1.03785491e+00
2.16529936e-01 -7.11911917e-02 3.29558641e-01 9.66667756e-02
-7.95487225e-01 1.90597504e-01 1.16196072e+00 7.49225557e-01
-3.82294506e-01 -9.39115942e-01 -2.09129542e-01 4.23248112e-01
-5.45524120e-01 6.76941220e-03 -1.52412906e-01 8.42964351e-01
1.11873932e-01 7.60405958e-01 3.99166048e-01 -4.44415659e-01
6.15361035e-02 -2.89401531e-01 6.17775083e-01 -5.67903280e-01
-1.94274023e-01 -2.90152371e-01 -1.14155106e-01 -1.09766614e+00
-6.87906265e-01 -3.63047987e-01 -1.27062678e+00 -8.62133265e-01
1.38527423e-01 2.25209109e-02 7.01958299e-01 7.33463228e-01
5.28025404e-02 3.79527360e-01 3.53997082e-01 -1.25824475e+00
3.62959594e-01 -4.41085875e-01 -6.61884606e-01 5.41694045e-01
2.86216348e-01 -7.44293094e-01 -2.64138550e-01 4.80779141e-01] | [8.724268913269043, -1.9732155799865723] |
9462838c-e33d-4a2e-be27-c847c6afd697 | large-language-models-are-human-level-prompt | 2211.01910 | null | https://arxiv.org/abs/2211.01910v2 | https://arxiv.org/pdf/2211.01910v2.pdf | Large Language Models Are Human-Level Prompt Engineers | By conditioning on natural language instructions, large language models (LLMs) have displayed impressive capabilities as general-purpose computers. However, task performance depends significantly on the quality of the prompt used to steer the model, and most effective prompts have been handcrafted by humans. Inspired by classical program synthesis and the human approach to prompt engineering, we propose Automatic Prompt Engineer (APE) for automatic instruction generation and selection. In our method, we treat the instruction as the "program," optimized by searching over a pool of instruction candidates proposed by an LLM in order to maximize a chosen score function. To evaluate the quality of the selected instruction, we evaluate the zero-shot performance of another LLM following the selected instruction. Experiments on 24 NLP tasks show that our automatically generated instructions outperform the prior LLM baseline by a large margin and achieve better or comparable performance to the instructions generated by human annotators on 19/24 tasks. We conduct extensive qualitative and quantitative analyses to explore the performance of APE. We show that APE-engineered prompts can be applied to steer models toward truthfulness and/or informativeness, as well as to improve few-shot learning performance by simply prepending them to standard in-context learning prompts. Please check out our webpage at https://sites.google.com/view/automatic-prompt-engineer. | ['Jimmy Ba', 'Harris Chan', 'Silviu Pitis', 'Keiran Paster', 'Ziwen Han', 'Andrei Ioan Muresanu', 'Yongchao Zhou'] | 2022-11-03 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 4.62122351e-01 2.22110674e-01 -5.12145102e-01 -4.62167263e-01
-1.27960920e+00 -6.21743262e-01 8.17316234e-01 3.84026557e-01
-4.81850773e-01 2.99705446e-01 5.36662936e-01 -6.18432045e-01
1.82548463e-01 -4.85030025e-01 -9.09274042e-01 -1.99993789e-01
3.72835696e-01 3.21293056e-01 2.26734430e-01 -2.62133360e-01
6.44946933e-01 -1.82934906e-02 -1.54661429e+00 4.60328579e-01
1.04494846e+00 3.57152104e-01 5.44981956e-01 7.54331887e-01
-2.59243399e-01 9.95396614e-01 -6.77399576e-01 -2.90667355e-01
-7.80581981e-02 -5.20731390e-01 -8.14527214e-01 -1.45501167e-01
3.92526895e-01 -3.48331898e-01 -1.37885973e-01 9.99258220e-01
3.59399915e-01 5.34961462e-01 4.10018533e-01 -1.09493995e+00
-5.04346848e-01 9.86577928e-01 -1.25141740e-01 1.27905905e-01
8.90186250e-01 6.80770218e-01 1.05821216e+00 -1.06686544e+00
5.11313438e-01 1.18682957e+00 2.83373147e-01 7.41221845e-01
-1.41714895e+00 -5.29445231e-01 8.72305632e-02 2.74398923e-01
-1.13224256e+00 -6.84880078e-01 6.93678617e-01 -5.07769883e-01
1.10068488e+00 2.38791436e-01 1.89859435e-01 1.23268807e+00
2.73357779e-01 1.09797227e+00 8.18210721e-01 -8.82739127e-01
3.14395756e-01 8.89070630e-02 4.46932375e-01 1.03395307e+00
-1.08405873e-01 2.93579549e-01 -8.52415025e-01 -3.55934381e-01
2.17264563e-01 -2.34085739e-01 -4.54205424e-01 -4.89379615e-02
-1.29590547e+00 1.04540622e+00 1.65478680e-02 9.14377496e-02
-4.84416842e-01 1.03019290e-01 4.27592635e-01 7.40239471e-02
-3.60962339e-02 1.16570544e+00 -3.92487615e-01 -5.54704547e-01
-9.14351642e-01 3.49841684e-01 9.66343820e-01 1.04236412e+00
7.95115352e-01 5.97180650e-02 -8.24490607e-01 7.18176663e-01
8.21498930e-02 1.52174383e-01 7.11881578e-01 -1.22910547e+00
3.86662245e-01 5.77659905e-01 2.26038992e-01 -6.61831319e-01
-1.36362806e-01 6.17680792e-03 -1.94551960e-01 8.20835456e-02
3.03352833e-01 -1.23558998e-01 -6.85591936e-01 1.73576844e+00
1.00664981e-01 6.15595616e-02 -2.54973263e-01 6.12811327e-01
5.89727342e-01 8.50625038e-01 3.69363844e-01 -1.55463621e-01
1.36418176e+00 -1.15414512e+00 -6.22198820e-01 -5.18597186e-01
8.44823897e-01 -8.50655019e-01 1.96815014e+00 5.79449117e-01
-8.70634735e-01 -7.36452401e-01 -1.00279820e+00 3.58935744e-02
5.71775511e-02 1.82594851e-01 4.58710700e-01 3.00906211e-01
-8.19387138e-01 5.84792793e-01 -8.08181822e-01 -8.36430266e-02
2.98720002e-01 7.28435367e-02 1.03972472e-01 -1.85324743e-01
-9.74181235e-01 7.67445207e-01 4.12192434e-01 -5.64538658e-01
-1.21071601e+00 -9.31400001e-01 -1.10978758e+00 1.94720611e-01
7.34292865e-01 -5.84176064e-01 2.10296273e+00 -7.53498971e-01
-1.62410963e+00 6.21818781e-01 -4.41999316e-01 -4.67974842e-01
1.53175220e-02 -5.12997746e-01 1.59653940e-03 6.88444683e-03
2.05124512e-01 9.02142525e-01 6.34452701e-01 -1.08100152e+00
-6.62428737e-01 1.85227364e-01 2.93911457e-01 -5.97422309e-02
-2.46208578e-01 2.53028840e-01 -3.17356169e-01 -5.32491744e-01
-4.99870151e-01 -9.48762059e-01 -3.57135385e-01 -5.70646226e-01
-3.65959615e-01 -6.43305123e-01 4.38193500e-01 -5.81406236e-01
1.63860261e+00 -2.10275745e+00 -2.09853113e-01 -2.10808486e-01
7.42395371e-02 4.79508102e-01 -5.81267297e-01 3.22888792e-01
1.06854647e-01 1.69369444e-01 -1.37405455e-01 -4.73148264e-02
3.60464573e-01 3.36222425e-02 -5.94984531e-01 -3.26438695e-02
2.21690148e-01 1.10273480e+00 -1.36344910e+00 -6.18196189e-01
2.01890185e-01 1.45653442e-01 -8.95696044e-01 8.31938624e-01
-8.19724798e-01 1.33583829e-01 -4.47030306e-01 5.36260664e-01
-2.06285954e-01 -3.76186103e-01 -1.42536715e-01 -7.25081712e-02
5.12310714e-02 8.34145665e-01 -7.61157215e-01 1.76825380e+00
-7.82700956e-01 6.15400910e-01 -2.81509936e-01 -5.55583954e-01
9.37201798e-01 3.61604303e-01 6.92717955e-02 -6.53039634e-01
-6.12606518e-02 2.82865893e-02 7.66459927e-02 -7.21985221e-01
7.07416832e-01 2.20293831e-03 -2.63335794e-01 7.42140532e-01
1.58833474e-01 -4.01112318e-01 4.39337760e-01 4.52247292e-01
1.30809784e+00 4.27676052e-01 7.79620945e-01 -1.49197981e-01
4.05184418e-01 3.26589078e-01 4.03734356e-01 1.10393500e+00
-2.81633854e-01 4.15822208e-01 3.96955281e-01 -3.51908207e-01
-7.11660028e-01 -5.87286353e-01 3.67465913e-01 1.73773611e+00
-3.25352818e-01 -7.59422600e-01 -9.42709804e-01 -8.86501908e-01
-2.18441144e-01 1.59603870e+00 -2.98975229e-01 -4.36647952e-01
-7.14335144e-01 -2.97377825e-01 4.51285660e-01 4.21166986e-01
9.65979416e-03 -1.38922751e+00 -9.80413139e-01 3.60279202e-01
-4.51353997e-01 -1.03431141e+00 -1.14772725e+00 3.70316207e-01
-5.41775405e-01 -9.53749001e-01 -8.49305913e-02 -5.52151740e-01
5.92397988e-01 2.32478529e-01 1.61883628e+00 1.23338670e-01
-3.98788601e-03 5.90399384e-01 -5.20379245e-01 -4.35996413e-01
-8.79051447e-01 5.03259487e-02 -9.46715660e-03 -4.36532438e-01
7.40360022e-01 -2.27502659e-01 -3.29306245e-01 1.19764574e-01
-8.51844013e-01 1.63574412e-01 5.92164218e-01 8.69662941e-01
4.60926503e-01 -3.10785025e-01 5.63014984e-01 -1.00711036e+00
9.92805898e-01 -3.24980497e-01 -6.52094245e-01 2.80857533e-01
-9.38648999e-01 5.57222545e-01 8.54919076e-01 -7.02304900e-01
-1.05436778e+00 2.71227479e-01 -4.56605516e-02 -3.95127892e-01
-4.87578243e-01 4.71246451e-01 -1.03955947e-01 6.40559494e-02
1.12720764e+00 1.99273959e-01 -4.39925730e-01 -2.47764856e-01
6.44309044e-01 5.14867842e-01 6.23738706e-01 -1.19522214e+00
6.05382681e-01 -3.69723886e-01 -4.42832232e-01 -4.30411249e-01
-1.00706398e+00 -3.53945374e-01 -3.33512455e-01 -4.83950339e-02
6.52714193e-01 -6.10940158e-01 -6.98994875e-01 -1.65373191e-01
-1.45827818e+00 -8.01364899e-01 -3.39265496e-01 2.80253768e-01
-6.32736027e-01 2.11983055e-01 -5.54790437e-01 -5.91525376e-01
-4.75287199e-01 -1.48163235e+00 1.00708532e+00 3.64487618e-01
-9.70722318e-01 -7.48984814e-01 9.23590362e-02 2.97604263e-01
3.37517887e-01 -1.35389090e-01 1.27582359e+00 -1.10854936e+00
-5.57413876e-01 -1.98439464e-01 2.46346578e-01 1.33853734e-01
5.75437546e-02 5.34990281e-02 -9.01460350e-01 1.13630287e-01
3.76379155e-02 -5.01978338e-01 4.11901563e-01 2.55563766e-01
1.20411396e+00 -7.93983579e-01 -2.25960493e-01 3.66265208e-01
1.19795728e+00 3.45262051e-01 1.88465446e-01 3.40445518e-01
6.47979319e-01 5.88382483e-01 9.00311470e-01 5.21819830e-01
2.07936227e-01 6.04526699e-01 1.33387402e-01 5.40316761e-01
-8.02044719e-02 -7.69419551e-01 7.09915102e-01 8.59236717e-01
4.67808545e-01 -1.29776046e-01 -1.12738800e+00 5.68831384e-01
-1.65619600e+00 -7.11152077e-01 1.67979404e-01 2.29756689e+00
1.38525867e+00 4.31191474e-01 -6.61441907e-02 -3.37524474e-01
3.41228694e-01 2.14763492e-01 -3.95426512e-01 -6.30788624e-01
3.87678623e-01 5.09902656e-01 1.24092840e-01 7.99018145e-01
-7.53887057e-01 1.05263126e+00 5.90167141e+00 1.01722527e+00
-8.07087660e-01 4.39334735e-02 4.08243239e-01 -1.26006231e-01
-4.87636358e-01 3.00119162e-01 -1.07547307e+00 3.22273970e-01
1.41787255e+00 -7.12005734e-01 7.07241356e-01 1.15506089e+00
4.89120603e-01 -5.33368811e-02 -1.57204950e+00 6.51793718e-01
8.36055502e-02 -1.32399750e+00 1.63809180e-01 -3.03653717e-01
8.44288170e-01 -2.19090536e-01 -1.96816936e-01 9.48269665e-01
7.15237558e-01 -9.61696744e-01 6.38343692e-01 4.35917050e-01
4.73888189e-01 -5.84794104e-01 2.88961053e-01 7.34105527e-01
-9.98469770e-01 -2.33201832e-02 -5.55604361e-02 -4.01365086e-02
4.42394521e-03 3.21428150e-01 -1.31193912e+00 -5.63628320e-03
7.40108192e-02 1.39088988e-01 -7.08269060e-01 7.86246896e-01
-8.39673102e-01 9.75466788e-01 6.56160638e-02 -3.74646038e-01
2.44551852e-01 1.49974659e-01 5.67624748e-01 1.59159231e+00
8.79454613e-02 3.03741604e-01 5.80887139e-01 1.25035799e+00
-2.45216206e-01 1.26279639e-02 -6.09281898e-01 -3.27051759e-01
8.39070380e-01 1.31196284e+00 -2.93963075e-01 -6.76685750e-01
-4.12981302e-01 6.38199627e-01 2.26548120e-01 4.24342126e-01
-8.17696452e-01 -5.81484079e-01 3.23834449e-01 1.61162272e-01
-1.46145284e-01 -7.08279759e-02 -2.49631450e-01 -8.78430486e-01
-1.63488656e-01 -1.54953158e+00 2.69978493e-01 -8.27799439e-01
-8.50604236e-01 5.38571894e-01 1.23673968e-01 -1.07178819e+00
-7.49261260e-01 -4.41792786e-01 -9.43303049e-01 7.08918750e-01
-1.14813936e+00 -5.12386322e-01 -2.08662853e-01 1.30437493e-01
1.17989266e+00 4.60576499e-03 8.19046319e-01 -9.19174179e-02
-5.61717689e-01 6.66775644e-01 -5.98551750e-01 -1.66120172e-01
8.83227825e-01 -1.32009077e+00 5.44164479e-01 1.13889599e+00
3.37561399e-01 9.29948747e-01 1.06895840e+00 -6.42768025e-01
-1.55339229e+00 -1.09357190e+00 9.94579613e-01 -6.62797391e-01
6.68230712e-01 -9.06752795e-02 -1.21585143e+00 7.24792242e-01
4.02318984e-01 -3.70535076e-01 6.51467085e-01 1.68589521e-02
-3.52390200e-01 2.68986404e-01 -7.41834283e-01 7.45448649e-01
6.06987536e-01 -6.76784515e-01 -9.78046119e-01 4.56756592e-01
1.19738579e+00 -4.52849656e-01 -5.59970021e-01 2.96821475e-01
2.13217273e-01 -6.39573991e-01 7.18811929e-01 -7.35678494e-01
7.35242009e-01 -1.55211419e-01 -1.88703120e-01 -1.36416101e+00
-3.40553820e-01 -9.30986583e-01 -4.02807325e-01 1.21118116e+00
3.49169582e-01 -1.17964171e-01 5.61353624e-01 1.05412567e+00
-5.06646633e-01 -9.51430380e-01 -2.30480060e-01 -6.73889160e-01
-1.70322254e-01 -5.42844653e-01 5.78651369e-01 5.94690025e-01
3.94872099e-01 7.01044679e-01 -1.87410805e-02 -9.01169181e-02
5.14092684e-01 1.30600214e-01 9.79349613e-01 -7.70042539e-01
-6.98234856e-01 -8.02110314e-01 5.02818763e-01 -1.32897162e+00
4.52304035e-01 -9.39675927e-01 7.89174199e-01 -1.19754076e+00
2.81220257e-01 -2.88635530e-02 -2.26894870e-01 6.85540795e-01
-7.72035658e-01 -5.01475275e-01 2.99823105e-01 2.98546385e-02
-8.48551512e-01 4.92133498e-01 9.82320070e-01 -9.23424140e-02
-5.36138892e-01 -1.73108149e-02 -8.08741570e-01 8.63551140e-01
7.47856855e-01 -5.40760577e-01 -3.13437819e-01 -2.66124666e-01
8.38168040e-02 2.86857158e-01 2.17816547e-01 -7.48040557e-01
4.91817862e-01 -5.45459926e-01 -2.02630043e-01 -1.60548925e-01
-1.15328923e-01 -1.94146261e-01 -3.27081770e-01 4.08459961e-01
-8.78556848e-01 1.88968539e-01 2.96673477e-01 2.79841900e-01
-1.40818432e-01 -7.52793491e-01 6.83427811e-01 -2.82580644e-01
-9.88410532e-01 3.29167000e-03 -4.92795289e-01 2.80010015e-01
6.27858877e-01 2.18221590e-01 -4.03599828e-01 -3.45070601e-01
-1.34492800e-01 2.03611046e-01 4.97999042e-01 4.44774806e-01
7.45905101e-01 -1.13165224e+00 -5.97249389e-01 -8.05864297e-03
4.13971484e-01 -9.90663916e-02 -2.39749417e-01 5.86951137e-01
-2.35112812e-02 7.09364414e-01 1.88214198e-01 -4.13623184e-01
-1.04992044e+00 7.30446577e-01 4.99936417e-02 -3.57344151e-01
-4.25054163e-01 8.01297009e-01 1.73377052e-01 -4.45631176e-01
5.10792553e-01 -4.69596326e-01 3.47292167e-03 -2.61081487e-01
7.53470123e-01 1.63961351e-01 -4.10486758e-02 1.41313681e-02
-2.08065495e-01 3.89104486e-02 -1.06512308e-01 -1.78161055e-01
1.10471940e+00 2.79243857e-01 1.41504690e-01 4.88772631e-01
9.52767968e-01 7.73001239e-02 -1.34420431e+00 -4.52956766e-01
4.72776949e-01 -3.78937304e-01 -1.52021632e-01 -9.16561663e-01
-3.54136527e-01 7.76745200e-01 8.86133090e-02 -2.29118749e-01
1.04387248e+00 -7.74641410e-02 8.04844499e-01 6.25542343e-01
3.55083317e-01 -8.43770623e-01 5.70821464e-01 6.86680973e-01
7.58832216e-01 -1.40355396e+00 -2.97854900e-01 5.47020882e-02
-7.63155878e-01 1.09196031e+00 9.93624687e-01 -6.97237765e-03
-3.59584275e-03 4.50070530e-01 -8.97610262e-02 -3.80050093e-02
-1.23109305e+00 1.02693364e-01 4.09537017e-01 3.61780822e-01
8.47765148e-01 2.01804310e-01 -2.20968857e-01 9.55021679e-01
-3.19542170e-01 2.33171493e-01 6.81058586e-01 1.02719200e+00
-8.63620937e-01 -1.10769880e+00 -4.49755996e-01 5.36825240e-01
-2.69955218e-01 -4.35765266e-01 -1.97683468e-01 3.62182617e-01
-2.10066289e-01 1.12883472e+00 -4.21140134e-01 -5.72547078e-01
4.18234795e-01 5.88715851e-01 3.38758945e-01 -1.09107399e+00
-7.25183487e-01 -2.27035820e-01 1.60467282e-01 -7.65725970e-01
1.37934327e-01 -5.01969635e-01 -1.35773146e+00 -9.77444723e-02
-2.40951598e-01 2.74817228e-01 3.82741928e-01 1.02045536e+00
3.48583251e-01 6.68764830e-01 4.34196085e-01 -7.43304789e-01
-1.11016405e+00 -1.03212953e+00 3.24389935e-01 4.43350315e-01
2.46391892e-01 -4.19322699e-01 -3.08594406e-01 4.40524131e-01] | [10.667472839355469, 8.177682876586914] |
02818012-af39-40ef-a386-c5fa36d6020c | opam-online-purchasing-behavior-analysis | 2102.01625 | null | https://arxiv.org/abs/2102.01625v1 | https://arxiv.org/pdf/2102.01625v1.pdf | OPAM: Online Purchasing-behavior Analysis using Machine learning | Customer purchasing behavior analysis plays a key role in developing insightful communication strategies between online vendors and their customers. To support the recent increase in online shopping trends, in this work, we present a customer purchasing behavior analysis system using supervised, unsupervised and semi-supervised learning methods. The proposed system analyzes session and user-journey level purchasing behaviors to identify customer categories/clusters that can be useful for targeted consumer insights at scale. We observe higher sensitivity to the design of online shopping portals for session-level purchasing prediction with accuracy/recall in range 91-98%/73-99%, respectively. The user-journey level analysis demonstrates five unique user clusters, wherein 'New Shoppers' are most predictable and 'Impulsive Shoppers' are most unique with low viewing and high carting behaviors for purchases. Further, cluster transformation metrics and partial label learning demonstrates the robustness of each user cluster to new/unlabelled events. Thus, customer clusters can aid strategic targeted nudge models. | ['Wenxi Li', 'Ebrahim Alareqi', 'Sohini Roychowdhury'] | 2021-02-02 | null | null | null | null | ['partial-label-learning'] | ['methodology'] | [-2.65435070e-01 -1.81128994e-01 -8.64070594e-01 -1.16274166e+00
-5.40235221e-01 -6.47727191e-01 1.89718634e-01 9.30233479e-01
-1.82017371e-01 8.83066654e-02 1.41735235e-02 -4.11024272e-01
-4.42236662e-01 -7.62388885e-01 -1.77453980e-01 -6.74292386e-01
-5.71654916e-01 8.53781521e-01 -1.85321152e-01 -3.09621960e-01
5.44430196e-01 2.33031526e-01 -1.34695768e+00 8.16817224e-01
6.28315270e-01 1.26653945e+00 2.53347963e-01 4.41927284e-01
-2.42702469e-01 7.38579333e-01 -9.25141051e-02 -4.31757212e-01
4.17640686e-01 -3.82424742e-01 -9.49787721e-02 3.13305914e-01
-3.53369623e-01 -1.54089689e-01 5.41446209e-01 7.94245541e-01
3.57193896e-03 4.84138774e-03 6.85955644e-01 -1.58792305e+00
-5.01067042e-01 1.27449441e+00 -5.86379230e-01 2.58539230e-01
3.66617769e-01 -3.43494654e-01 1.30564404e+00 -5.01315415e-01
5.33019841e-01 9.01236475e-01 6.55596256e-01 -1.02666654e-01
-1.50502455e+00 -6.49165511e-01 2.96876013e-01 4.21147138e-01
-1.23709333e+00 9.03362110e-02 9.20203626e-01 -5.40697992e-01
7.33078003e-01 3.07166696e-01 5.45143843e-01 9.03132737e-01
7.91798383e-02 1.12362432e+00 1.12316120e+00 -3.71356010e-01
5.62031984e-01 1.17934477e+00 8.38708282e-01 -7.10008442e-02
1.11082733e-01 -1.51854083e-01 -1.67754248e-01 1.56704076e-02
3.13646823e-01 5.84847450e-01 5.28045893e-01 -4.97706801e-01
-8.23254466e-01 1.37660265e+00 2.33400807e-01 3.85654867e-01
-7.38290429e-01 -6.65322483e-01 5.32801509e-01 6.39009178e-01
2.29660228e-01 3.20047528e-01 -7.36513376e-01 -3.16069037e-01
-7.62556612e-01 5.34680337e-02 1.34470093e+00 1.56674254e+00
6.05456412e-01 -2.07680017e-01 3.14318866e-01 8.04332674e-01
5.47419548e-01 -8.14376473e-02 5.33476830e-01 -5.92397988e-01
9.51428935e-02 6.91802382e-01 1.97839677e-01 -1.05938864e+00
-8.62581015e-01 -6.35159791e-01 -4.83687848e-01 -5.15037715e-01
1.76407605e-01 3.54868561e-01 -3.65291238e-01 9.62241292e-01
-1.39241796e-02 -6.87368095e-01 4.77491356e-02 5.84003627e-01
2.38459006e-01 5.06473184e-01 5.40682733e-01 -6.72479570e-01
1.32748199e+00 -7.16529846e-01 -6.30274057e-01 8.70921984e-02
8.58709335e-01 -8.26725125e-01 8.55152607e-01 9.28124487e-01
-7.41522253e-01 -6.32289588e-01 -7.44183064e-01 6.87503219e-01
-5.56709468e-01 -1.54994607e-01 1.14881885e+00 9.29856837e-01
-4.15871352e-01 4.49422091e-01 -5.87247908e-01 -7.71394968e-01
2.95118839e-01 5.14233053e-01 2.73244619e-01 -1.79956838e-01
-7.40769923e-01 5.32498777e-01 5.54944992e-01 -2.36752942e-01
-2.52353251e-01 -6.46346807e-01 -5.77124536e-01 1.62757244e-02
-5.60053671e-03 1.21645398e-01 1.24856877e+00 -9.38983798e-01
-7.64725447e-01 6.19230151e-01 1.90985247e-01 -6.84584856e-01
2.68866986e-01 3.42328459e-01 -1.13037324e+00 -1.23613052e-01
3.88269454e-01 4.25498664e-01 2.67832905e-01 -1.72148061e+00
-1.62572312e+00 -7.56691992e-01 -4.57006425e-01 -9.10025537e-02
-2.50995845e-01 -1.36760678e-02 3.83690260e-02 -2.84485430e-01
6.94501340e-01 -7.12374449e-01 -4.00060028e-01 -1.08711815e+00
-2.45251477e-01 -4.23934937e-01 7.91924238e-01 -7.23456979e-01
1.37055397e+00 -2.40152383e+00 -5.88753760e-01 8.28423798e-01
9.47789475e-02 -5.28648794e-01 3.98183733e-01 8.42184067e-01
-8.39632675e-02 3.57431546e-02 5.40152609e-01 -1.44048840e-01
5.92116475e-01 3.35030943e-01 -2.72730384e-02 2.28015453e-01
-2.47662947e-01 7.28407145e-01 -8.46969783e-01 -2.86490738e-01
4.98339027e-01 -1.78112552e-01 -5.91121197e-01 1.68312058e-01
2.58986768e-03 4.28312212e-01 -4.77722019e-01 1.29012573e+00
6.98426247e-01 -5.54028228e-02 4.72383827e-01 -3.39005530e-01
-2.68998653e-01 8.58160257e-02 -1.06931317e+00 8.92558575e-01
-3.74667019e-01 -4.24682572e-02 1.13202050e-01 -1.27709150e+00
1.15594208e+00 -2.86300421e-01 9.42483783e-01 -1.30881727e+00
2.56260425e-01 4.10855711e-02 1.93500131e-01 -5.54419518e-01
6.35482192e-01 -1.71029866e-01 -6.79579973e-01 5.11591375e-01
1.04409538e-01 7.85126805e-01 3.84162337e-01 1.81633919e-01
5.55390000e-01 -1.85373977e-01 3.53590459e-01 -4.45532143e-01
-5.47733903e-02 4.90852475e-01 5.37403584e-01 6.57548070e-01
-5.45338333e-01 5.65668158e-02 5.63247979e-01 -3.36375296e-01
-1.09442472e+00 -9.61684585e-01 -4.17532563e-01 1.53909039e+00
1.14678681e-01 -4.84913960e-02 -3.39214981e-01 -3.36124033e-01
2.47832656e-01 1.44587445e+00 -4.25249189e-01 -1.58508569e-01
-1.99262336e-01 -6.88416004e-01 -3.14865470e-01 7.29491889e-01
1.70543715e-01 -9.58349228e-01 -2.63532311e-01 6.83474243e-01
8.54973271e-02 -1.03108382e+00 -1.49236932e-01 6.88357592e-01
-9.11764741e-01 -7.87908912e-01 8.48295987e-02 -1.26846552e+00
5.59647322e-01 3.15628052e-01 1.10826111e+00 -5.91538727e-01
-2.08702788e-01 3.80936593e-01 -1.04933643e+00 -3.39642256e-01
-6.27241552e-01 2.73804873e-01 2.24906072e-01 1.15377545e-01
1.37940407e+00 -5.25920093e-01 -8.10088575e-01 9.75396216e-01
-4.58137274e-01 -4.25314933e-01 4.94854212e-01 5.10532737e-01
3.67892712e-01 6.34942353e-01 1.10440600e+00 -1.20568514e+00
6.41220272e-01 -1.20326662e+00 -2.35899955e-01 -1.36156809e-02
-1.32178700e+00 -6.56550944e-01 7.26144373e-01 -3.71156454e-01
-8.42567265e-01 1.10221226e-02 -1.29122257e-01 2.18217060e-01
-7.07126319e-01 5.19039452e-01 1.41166486e-02 4.76137191e-01
4.16465014e-01 1.71721995e-01 -6.13317378e-02 -5.72311521e-01
2.27986127e-01 1.00967634e+00 1.49757266e-01 -1.21258050e-01
2.45806515e-01 6.59271851e-02 -7.55793273e-01 -6.42236710e-01
-3.39664429e-01 -1.36430311e+00 -1.01397347e+00 -2.66844511e-01
7.26724327e-01 -7.93005586e-01 -1.32032108e+00 -1.62532806e-01
-1.13717437e-01 -1.15265757e-01 -3.51235121e-01 5.93889177e-01
-4.80258793e-01 9.22756940e-02 -7.71137655e-01 -1.16402113e+00
-7.47164264e-02 -8.31840575e-01 4.37891811e-01 1.56615540e-01
-8.33853066e-01 -1.29075468e+00 -4.67643052e-01 7.82324612e-01
2.47297078e-01 1.81091771e-01 1.15957224e+00 -1.66077125e+00
-2.83531725e-01 -7.21597075e-01 -1.08204335e-02 2.90698111e-01
1.54784605e-01 -4.39504832e-01 -3.92522782e-01 -5.68978131e-01
1.03301346e-01 1.20058618e-01 1.56500131e-01 5.31495631e-01
4.94354606e-01 -2.20913246e-01 -5.20220876e-01 8.43031704e-02
1.54142904e+00 1.30585361e+00 3.63784045e-01 5.55187523e-01
3.30132991e-01 1.04859555e+00 1.23952973e+00 9.26207483e-01
7.97356963e-01 2.43581772e-01 1.74336404e-01 2.25087628e-02
7.27093637e-01 -4.02831495e-01 3.05607080e-01 1.01139367e+00
3.12240183e-01 1.19954064e-01 -5.39517999e-01 4.93288845e-01
-2.05668068e+00 -9.13490355e-01 -3.63159269e-01 2.05273223e+00
5.33351041e-02 6.05766594e-01 8.48552406e-01 1.90964594e-01
5.96542656e-01 -5.86183310e-01 -5.98288476e-01 -7.88770139e-01
1.94711626e-01 -1.79874793e-01 8.20948064e-01 -6.41693994e-02
-8.50225031e-01 6.21350110e-01 6.10685205e+00 3.94167304e-01
-4.25232410e-01 -6.55844212e-02 1.00971031e+00 5.45895398e-02
-2.77338594e-01 -1.10599935e-01 -1.10444641e+00 7.81637430e-01
1.16970789e+00 6.39235750e-02 2.92204589e-01 1.40541351e+00
7.09103346e-01 -1.64273039e-01 -1.36309624e+00 8.98446620e-01
9.99764130e-02 -9.45586801e-01 -4.68936116e-01 4.07011539e-01
4.43451107e-01 -3.67891729e-01 -2.96495967e-02 8.17779243e-01
4.66443837e-01 -5.55403769e-01 5.11833012e-01 1.91388875e-01
-3.62319648e-02 -1.16395843e+00 9.46761012e-01 5.23909748e-01
-1.18837941e+00 -7.41904140e-01 -2.26151064e-01 5.30569777e-02
5.07817626e-01 3.20654005e-01 -1.21888769e+00 1.49687678e-01
1.02332795e+00 7.60768414e-01 -3.04309756e-01 6.34676397e-01
8.43793333e-01 6.87310278e-01 -2.37152621e-01 -2.27910027e-01
3.44392478e-01 -8.63536954e-01 -2.45390803e-01 1.45225298e+00
2.63726830e-01 1.07868992e-01 6.20340645e-01 6.85299575e-01
5.92710376e-01 4.79237497e-01 -2.64626712e-01 -3.04656804e-01
4.05399352e-01 1.47142565e+00 -1.20075107e+00 3.20671796e-04
-5.45943439e-01 8.68361294e-01 -3.61040026e-01 4.40991633e-02
-5.93270659e-01 -2.12460816e-01 3.51627767e-01 4.78008449e-01
5.90838134e-01 -7.12945610e-02 -6.47495449e-01 -7.16369331e-01
-3.54020268e-01 -6.89799905e-01 5.57227314e-01 -3.19632217e-02
-1.58473623e+00 3.91030870e-02 -1.43936396e-01 -1.32041359e+00
-2.43346557e-01 -5.49629509e-01 -4.20063913e-01 2.23953784e-01
-1.12385571e+00 -1.29537523e+00 -5.58926128e-02 4.96501654e-01
7.55978882e-01 -4.94600117e-01 7.37076879e-01 6.36935413e-01
-3.73521566e-01 7.45984733e-01 3.26658994e-01 -4.96479347e-02
3.46760720e-01 -1.38071263e+00 -7.17510805e-02 3.40603925e-02
-3.52700621e-01 6.96810961e-01 6.93230808e-01 -6.65041864e-01
-1.50445306e+00 -8.58450234e-01 9.78122771e-01 -1.84137896e-01
8.06278706e-01 -6.08429670e-01 -6.57094181e-01 7.15951920e-01
1.41757756e-01 -1.05033946e+00 1.65600288e+00 6.23205245e-01
-9.44626927e-02 -4.76081818e-01 -1.51889598e+00 2.49676287e-01
4.83243465e-01 -7.04979822e-02 -2.06909761e-01 4.28278983e-01
3.95226687e-01 4.36070710e-01 -1.67878091e+00 -6.26803637e-02
6.40811682e-01 -1.16892171e+00 7.75399566e-01 -4.32567626e-01
6.90362751e-02 4.25819635e-01 -4.99774128e-01 -8.88233662e-01
-9.14909899e-01 -3.24321717e-01 3.33638698e-01 1.56491256e+00
6.91115499e-01 -6.15344822e-01 1.10165548e+00 8.83806944e-01
-1.51928037e-01 -5.77581108e-01 -2.71621287e-01 -3.87429386e-01
-3.19963694e-01 -5.98466635e-01 4.29470837e-01 1.39166379e+00
8.01298678e-01 2.61748075e-01 -9.61098149e-02 7.93143287e-02
9.05796230e-01 4.67390418e-01 5.57460308e-01 -1.23715985e+00
-2.89178699e-01 -5.25268078e-01 -4.71309006e-01 -7.87712812e-01
-5.26266217e-01 -8.65026772e-01 -2.55651295e-01 -1.22632241e+00
1.75343677e-01 -8.47146690e-01 -5.64132512e-01 8.87090862e-02
6.45976305e-01 1.18536554e-01 2.62823002e-03 4.51118350e-01
-7.00573087e-01 -4.19586867e-01 6.33816063e-01 2.45781288e-01
-7.74194300e-01 6.01132452e-01 -1.09879446e+00 4.82245296e-01
1.01535404e+00 -1.11579739e-01 -3.72689515e-01 5.49433708e-01
2.95727283e-01 1.48198038e-01 -2.27664068e-01 -4.97145146e-01
1.44436523e-01 -1.32965535e-01 4.14679855e-01 -1.03991747e+00
-1.73938289e-01 -1.20709729e+00 3.55913848e-01 4.64133054e-01
-4.84436125e-01 1.47983417e-01 -2.60394365e-01 5.80931604e-01
-1.28077045e-01 -1.66608110e-01 4.90109831e-01 -3.26005340e-01
-1.08571649e+00 9.38844010e-02 -6.11084700e-01 -7.10980475e-01
1.54367399e+00 -6.98461652e-01 3.01191121e-01 -4.98520404e-01
-1.27764368e+00 5.08376300e-01 1.10392936e-01 6.02690279e-01
2.73763806e-01 -1.21822822e+00 -3.48805040e-01 4.35700655e-01
5.04416645e-01 -5.78428447e-01 3.17517638e-01 8.44402730e-01
-3.09750766e-01 6.03682160e-01 -3.89871687e-01 -8.01941693e-01
-9.29510236e-01 8.40051055e-01 -3.87533337e-01 -8.97445306e-02
-2.25063935e-01 4.30892825e-01 -3.37769568e-01 -4.96205956e-01
3.85489583e-01 -2.83669859e-01 -5.95335543e-01 8.33058894e-01
8.66978168e-02 6.85473680e-01 -1.35370325e-02 -7.13177025e-01
-7.79425129e-02 -5.98210143e-03 -8.30072284e-01 4.42621529e-01
1.50003242e+00 -6.40770972e-01 2.59946764e-01 8.33499074e-01
1.46701980e+00 -5.32858253e-01 -1.01232648e+00 -1.80851266e-01
6.55371070e-01 -4.34612125e-01 -1.03758164e-01 -9.48954284e-01
-8.58773887e-01 4.26645607e-01 9.59836662e-01 1.29955864e+00
9.99218881e-01 3.94279182e-01 9.32985425e-01 -3.57102528e-02
6.59045756e-01 -1.55150211e+00 -2.06831351e-01 -2.77390957e-01
7.26369694e-02 -1.75543654e+00 -2.89148301e-01 -4.02922004e-01
-1.17146969e+00 8.86086524e-01 3.67480397e-01 -5.98551184e-02
9.98855352e-01 7.00639486e-02 1.75673902e-01 -1.75816447e-01
-3.96579325e-01 -2.81685703e-02 -1.91580296e-01 6.65281534e-01
4.67927426e-01 6.77070081e-01 -3.89441758e-01 1.25274646e+00
-4.02351767e-01 -3.02570850e-01 2.33954057e-01 1.04370201e+00
-5.89308560e-01 -1.23043656e+00 1.31516205e-02 9.63605762e-01
-1.80800468e-01 9.37043577e-02 9.87084880e-02 1.01868880e+00
1.44651398e-01 1.12787509e+00 3.75145257e-01 -7.87414014e-01
6.61452472e-01 1.99648187e-01 6.24362193e-02 -5.32052338e-01
-8.12542379e-01 8.40246975e-01 8.90811756e-02 -3.81954342e-01
-2.87617564e-01 -1.23660541e+00 -1.30856633e+00 -5.25602579e-01
-3.91014099e-01 2.60548949e-01 1.01562750e+00 6.64520621e-01
2.34704792e-01 3.96991223e-01 1.32235384e+00 -4.17948216e-01
-6.39319777e-01 -9.24446344e-01 -1.26589429e+00 7.36072659e-01
-2.44125918e-01 -3.27756435e-01 -3.30075115e-01 3.97933066e-01] | [9.492427825927734, 5.907097339630127] |
c9c865c4-5edf-4c4b-8a46-b6cb5ec1b653 | 3d-human-pose-and-shape-regression-with | 2103.16507 | null | https://arxiv.org/abs/2103.16507v4 | https://arxiv.org/pdf/2103.16507v4.pdf | PyMAF: 3D Human Pose and Shape Regression with Pyramidal Mesh Alignment Feedback Loop | Regression-based methods have recently shown promising results in reconstructing human meshes from monocular images. By directly mapping raw pixels to model parameters, these methods can produce parametric models in a feed-forward manner via neural networks. However, minor deviation in parameters may lead to noticeable misalignment between the estimated meshes and image evidences. To address this issue, we propose a Pyramidal Mesh Alignment Feedback (PyMAF) loop to leverage a feature pyramid and rectify the predicted parameters explicitly based on the mesh-image alignment status in our deep regressor. In PyMAF, given the currently predicted parameters, mesh-aligned evidences will be extracted from finer-resolution features accordingly and fed back for parameter rectification. To reduce noise and enhance the reliability of these evidences, an auxiliary pixel-wise supervision is imposed on the feature encoder, which provides mesh-image correspondence guidance for our network to preserve the most related information in spatial features. The efficacy of our approach is validated on several benchmarks, including Human3.6M, 3DPW, LSP, and COCO, where experimental results show that our approach consistently improves the mesh-image alignment of the reconstruction. The project page with code and video results can be found at https://hongwenzhang.github.io/pymaf. | ['Zhenan Sun', 'LiMin Wang', 'Yebin Liu', 'Wanli Ouyang', 'Xinchi Zhou', 'Yating Tian', 'Hongwen Zhang'] | 2021-03-30 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Zhang_PyMAF_3D_Human_Pose_and_Shape_Regression_With_Pyramidal_Mesh_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Zhang_PyMAF_3D_Human_Pose_and_Shape_Regression_With_Pyramidal_Mesh_ICCV_2021_paper.pdf | iccv-2021-1 | ['3d-human-pose-and-shape-estimation', '3d-human-reconstruction', 'human-mesh-recovery'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.46128803e-01 1.25228390e-01 -2.47361049e-01 -5.49744129e-01
-4.86014366e-01 4.63194102e-02 3.09523433e-01 -4.09585238e-01
-3.34465504e-02 6.51150107e-01 2.10164756e-01 2.86205351e-01
-1.94459874e-02 -7.65960395e-01 -1.21941209e+00 -3.58712852e-01
2.71197617e-01 3.48519087e-01 1.37115642e-01 -7.84392804e-02
2.16828302e-01 4.33055013e-01 -1.42818499e+00 2.81129569e-01
7.05734551e-01 1.11799490e+00 4.23906147e-01 2.69635677e-01
1.21625131e-02 6.08185589e-01 1.86515637e-02 -2.41578370e-01
5.00111759e-01 -1.25389010e-01 -7.25193858e-01 3.34716707e-01
7.85831809e-01 -6.91678345e-01 -6.59028590e-01 1.01190901e+00
3.42448533e-01 -2.31574774e-01 4.55370754e-01 -1.02983558e+00
-5.74817955e-01 3.52179676e-01 -8.62289846e-01 -2.22381502e-01
3.13083112e-01 3.24683875e-01 7.57661283e-01 -1.30678988e+00
9.04453397e-01 1.33495164e+00 7.71759272e-01 3.12504023e-01
-1.25442028e+00 -9.33924019e-01 1.78559050e-01 1.84704930e-01
-1.49659956e+00 -5.45888126e-01 1.08360076e+00 -4.88105327e-01
5.99814653e-01 1.45125136e-01 8.45697999e-01 9.13242579e-01
3.15416396e-01 5.78706563e-01 8.38937163e-01 -1.28893912e-01
-3.37556034e-01 -4.76853512e-02 -4.74640816e-01 1.15615487e+00
-5.04012220e-02 3.07964534e-01 -7.98846126e-01 1.15405165e-01
1.52537298e+00 4.08031940e-02 -4.24986213e-01 -7.05191374e-01
-1.33924496e+00 5.52492082e-01 8.01467478e-01 -9.49281156e-02
-6.20198667e-01 4.52780664e-01 5.33332229e-02 1.16762854e-01
5.78384101e-01 2.73350924e-01 -5.36252201e-01 1.72672808e-01
-8.66037369e-01 2.69532353e-01 1.82456478e-01 9.55577672e-01
1.04395151e+00 3.86937670e-02 6.48126230e-02 1.00134385e+00
4.18172032e-01 5.98242640e-01 6.94659948e-02 -1.40397000e+00
6.10353291e-01 7.27711976e-01 1.24575190e-01 -1.32806087e+00
-3.77082735e-01 -4.14169252e-01 -9.10441101e-01 1.96382150e-01
1.43719360e-01 1.10003874e-01 -8.84535849e-01 1.38069272e+00
6.16060555e-01 4.12668079e-01 -6.19415164e-01 1.26814294e+00
9.09924865e-01 5.52300870e-01 -2.24525720e-01 2.02401072e-01
1.09088945e+00 -1.03049278e+00 -4.16166544e-01 -2.20639944e-01
2.52475947e-01 -7.79481769e-01 1.10210598e+00 1.96727365e-01
-1.24712181e+00 -7.03841627e-01 -9.32487905e-01 -2.48974696e-01
1.72250509e-01 3.80461782e-01 2.04256564e-01 -1.41101882e-01
-8.30784678e-01 8.04409027e-01 -9.38584268e-01 -5.48910573e-02
6.62845135e-01 3.81417900e-01 -4.66683745e-01 -1.25086501e-01
-1.04371428e+00 6.30535662e-01 7.83109739e-02 3.69195193e-01
-6.23177350e-01 -1.09771287e+00 -1.11243761e+00 -6.97075278e-02
3.14260840e-01 -9.71463740e-01 9.16853845e-01 -8.52277994e-01
-1.53082323e+00 7.52759516e-01 -1.34102747e-01 -1.68815807e-01
6.48016512e-01 -3.06582361e-01 8.74451473e-02 2.62596041e-01
2.19466522e-01 1.31281519e+00 9.91667628e-01 -1.51293528e+00
-6.14464343e-01 -1.58239767e-01 -7.48687908e-02 1.61447108e-01
-9.87690538e-02 -3.46774340e-01 -7.24816382e-01 -7.70395339e-01
3.76922488e-01 -8.54360640e-01 -2.52516747e-01 6.66831017e-01
-4.54879224e-01 9.98206064e-02 5.70372939e-01 -9.68335211e-01
9.41873431e-01 -1.99488282e+00 3.32808465e-01 3.94608647e-01
1.36337057e-01 -1.92001686e-01 -1.83895156e-01 -7.79433921e-02
8.47075805e-02 -2.88563192e-01 -3.89262289e-01 -5.48766792e-01
-2.06884012e-01 3.06513637e-01 -2.44620755e-01 6.42468333e-01
5.01196504e-01 1.17115831e+00 -7.54002094e-01 -6.73607051e-01
5.52623570e-01 8.21842194e-01 -7.06110120e-01 1.79475650e-01
-2.01103941e-01 8.69643092e-01 -3.48837346e-01 8.74926269e-01
7.23708689e-01 -5.00560462e-01 -2.82495022e-02 -7.19415963e-01
-1.78200623e-03 1.16645932e-01 -1.22638059e+00 1.91573668e+00
-6.80865943e-01 5.61001182e-01 1.18480638e-01 -8.61326814e-01
7.68122017e-01 9.29721221e-02 7.25578964e-01 -7.26291239e-01
1.95394039e-01 1.52915895e-01 -3.49120468e-01 -2.79180497e-01
3.77526522e-01 3.55181158e-01 2.45113418e-01 -5.81146516e-02
-1.02082022e-01 -2.44101346e-01 -2.41716608e-01 -1.63655013e-01
6.07605696e-01 6.34103775e-01 -3.40264887e-02 -2.62506474e-02
6.19989455e-01 -6.93777129e-02 7.88321197e-01 2.21989408e-01
2.19809130e-01 1.02605665e+00 1.30604446e-01 -6.31470859e-01
-1.38590181e+00 -9.47278559e-01 -3.00845027e-01 5.03024399e-01
3.62334341e-01 -4.23821747e-01 -5.12364089e-01 -4.74270284e-01
1.28676772e-01 1.63343295e-01 -6.90353692e-01 -2.30504498e-02
-8.93562555e-01 -2.36154199e-01 1.34120673e-01 6.64015651e-01
6.59987390e-01 -1.05004704e+00 -5.18240869e-01 2.78088003e-01
-2.94335961e-01 -1.19177616e+00 -5.92984200e-01 -2.51849234e-01
-9.30280149e-01 -9.55896914e-01 -7.43931472e-01 -7.65076220e-01
1.02959728e+00 7.95063600e-02 9.81504023e-01 4.81454819e-01
-2.86987990e-01 1.23267956e-01 1.52689159e-01 1.61900103e-01
-2.61045337e-01 2.03649756e-02 -1.64707303e-01 -1.67557560e-02
-4.13176209e-01 -7.95688331e-01 -8.90142143e-01 6.14104629e-01
-5.70160031e-01 6.22920752e-01 6.66163862e-01 9.90412951e-01
9.58024085e-01 -2.54110992e-01 2.63550848e-01 -8.18034589e-01
-1.16135791e-01 -3.54432970e-01 -7.59479463e-01 -7.35063255e-02
-5.96294582e-01 2.15212498e-02 5.52312374e-01 -3.61681551e-01
-8.95871341e-01 5.08188009e-01 -1.67347401e-01 -1.04545653e+00
1.02953158e-01 4.01907176e-01 -1.84695870e-01 -2.17430174e-01
2.93885559e-01 1.84942901e-01 6.73605651e-02 -5.99037409e-01
1.65557891e-01 3.42630416e-01 7.35853732e-01 -5.69893122e-01
9.64465797e-01 5.51826417e-01 1.05179824e-01 -5.49914122e-01
-8.11433375e-01 -2.00999007e-01 -6.47152841e-01 -3.92073721e-01
6.30090892e-01 -1.22382164e+00 -3.88397396e-01 5.98603547e-01
-1.14363635e+00 -4.37898666e-01 -1.41083747e-01 4.12056416e-01
-6.64971888e-01 3.48953545e-01 -7.42193162e-01 -2.03760698e-01
-3.58687073e-01 -1.32864559e+00 1.32113290e+00 6.85385317e-02
-1.92901209e-01 -7.24845648e-01 -3.46331239e-01 4.91777301e-01
2.68352896e-01 2.68409818e-01 8.59547377e-01 2.84798980e-01
-9.04927373e-01 -4.67448641e-04 -3.45253766e-01 2.86901742e-01
2.89381951e-01 8.66368935e-02 -8.08005154e-01 -2.10627183e-01
-3.26423258e-01 -2.11544290e-01 7.16819406e-01 5.38031518e-01
1.52843785e+00 -2.58717656e-01 -3.50405753e-01 9.45455253e-01
1.25002348e+00 -5.36868155e-01 5.72334290e-01 3.66579711e-01
1.20871317e+00 5.17766774e-01 7.61344790e-01 4.54971880e-01
6.09763205e-01 9.82086062e-01 5.33358753e-01 -2.25948080e-01
-6.14210844e-01 -5.89944005e-01 2.51202732e-01 7.97800183e-01
-7.59098083e-02 3.55042905e-01 -6.93092704e-01 3.68108660e-01
-1.81009126e+00 -6.06898248e-01 -1.69670284e-01 2.06010437e+00
1.07772398e+00 1.42419308e-01 -9.85605344e-02 -2.11205751e-01
6.45842969e-01 1.18532725e-01 -6.97619975e-01 1.07462257e-01
-6.21421374e-02 1.32047936e-01 5.06927311e-01 7.00523794e-01
-1.04774809e+00 9.99709427e-01 5.09826136e+00 6.93622410e-01
-1.31604278e+00 -5.82461059e-02 6.82901680e-01 -2.12428078e-01
-2.98263431e-01 -4.85025384e-02 -5.06008029e-01 5.96630812e-01
4.32303309e-01 1.64129049e-01 4.79373097e-01 7.13453770e-01
3.54855776e-01 -1.57683808e-02 -1.05016768e+00 1.17738700e+00
-1.46236897e-01 -1.68175793e+00 1.53472036e-01 1.84826285e-01
9.27651048e-01 2.44581923e-02 6.19426034e-02 -1.98009275e-02
-5.40298820e-02 -1.07523274e+00 9.00332868e-01 8.54804456e-01
9.23685670e-01 -7.42320001e-01 4.95947778e-01 1.42232135e-01
-1.31606066e+00 1.35284454e-01 -4.87901986e-01 1.80398654e-02
2.63975114e-01 7.01993406e-01 -7.08372176e-01 6.19299710e-01
8.66564274e-01 1.14550042e+00 -5.10644495e-01 8.42053890e-01
-2.50667185e-01 3.28881383e-01 -2.76611596e-01 6.28725708e-01
-1.36830479e-01 -1.83788240e-01 4.03807253e-01 7.33266294e-01
1.52435079e-01 -9.34546143e-02 1.28111854e-01 1.24750841e+00
-1.94239691e-01 1.13918543e-01 -2.25550294e-01 4.86414403e-01
5.55682003e-01 1.24205315e+00 -5.82531512e-01 -2.53747404e-01
-3.52368861e-01 9.42760885e-01 3.89725834e-01 1.21113710e-01
-9.26766694e-01 1.70567036e-01 6.81875467e-01 5.78628600e-01
4.62670624e-01 -2.11938486e-01 -5.70740163e-01 -9.92221057e-01
3.73975664e-01 -7.81350613e-01 -1.26786381e-01 -8.45106363e-01
-1.09625912e+00 4.04715478e-01 -6.08182363e-02 -1.46369922e+00
-2.26687603e-02 -2.93867677e-01 -2.52614528e-01 6.92474186e-01
-1.49208403e+00 -1.30445945e+00 -5.37048697e-01 4.54027206e-01
6.34756327e-01 1.11957654e-01 3.89817804e-01 4.62628692e-01
-4.61785138e-01 6.47444367e-01 -3.35484415e-01 1.62807360e-01
7.22811639e-01 -8.69826853e-01 4.71661538e-01 5.29874623e-01
6.70453608e-02 3.35615784e-01 5.56175828e-01 -9.16620851e-01
-1.43155706e+00 -1.39932871e+00 5.02173960e-01 -2.49087542e-01
2.96974570e-01 -1.66688979e-01 -1.01738763e+00 5.86379707e-01
-1.29518434e-01 4.22845930e-01 -4.17812355e-02 -3.68967295e-01
-1.67315483e-01 -4.32551764e-02 -9.83655572e-01 5.32110274e-01
1.30718184e+00 -3.39515030e-01 -2.15036869e-01 2.53783375e-01
5.17223775e-01 -9.38387930e-01 -1.16551018e+00 7.30458677e-01
7.63368487e-01 -7.63898253e-01 1.18405282e+00 -2.19855219e-01
9.68066216e-01 -5.27126491e-01 -8.56118724e-02 -1.19981599e+00
-3.46872240e-01 -2.07425669e-01 -2.14800045e-01 1.04873025e+00
1.78145781e-01 -3.14069718e-01 8.90142381e-01 4.78137851e-01
-1.36927962e-01 -1.35545087e+00 -9.84827518e-01 -4.22635406e-01
-4.62388098e-02 -2.01932400e-01 7.63732433e-01 8.81904602e-01
-5.14032125e-01 5.30897565e-02 -6.13848329e-01 4.00318444e-01
6.20829403e-01 1.96284801e-01 9.83519673e-01 -1.03657532e+00
-1.69854254e-01 -2.48960257e-01 -3.97814959e-01 -1.42497289e+00
3.62659246e-01 -8.42943192e-01 1.16438352e-01 -1.39878511e+00
2.26513539e-02 -5.78376770e-01 -8.38253275e-03 5.66085696e-01
-1.90953180e-01 6.78665340e-01 8.64855275e-02 3.95796537e-01
-2.48684064e-01 7.55020022e-01 1.60967088e+00 -1.74593881e-01
-2.54514869e-02 -1.41351879e-01 -2.51252264e-01 8.48583043e-01
7.35757649e-01 -3.92957330e-01 -2.37262875e-01 -6.15722477e-01
1.37650833e-01 5.52055277e-02 7.35848486e-01 -9.15784001e-01
2.30625719e-01 -2.25246042e-01 8.33555341e-01 -6.39711082e-01
5.28110981e-01 -9.03432429e-01 5.03380477e-01 4.60298419e-01
-2.09714219e-01 7.04988167e-02 -9.37291980e-02 3.60243946e-01
-5.57336248e-02 1.55867741e-01 7.42905855e-01 1.90562755e-02
-6.47707641e-01 6.22551858e-01 2.09077805e-01 -1.71895221e-01
6.50723040e-01 -4.32494760e-01 7.26943687e-02 -3.03798258e-01
-5.53494513e-01 4.08307672e-01 9.68047261e-01 5.59595466e-01
1.02761841e+00 -1.54245079e+00 -7.80390501e-01 4.28504556e-01
2.14802157e-02 5.61588347e-01 3.47996831e-01 1.02141643e+00
-5.61759472e-01 8.77951384e-02 -3.44550997e-01 -9.81080472e-01
-1.06038344e+00 2.49233752e-01 5.69122434e-01 -1.00070015e-02
-1.04950559e+00 7.43247509e-01 3.03594738e-01 -5.73764920e-01
2.73061723e-01 -5.06706178e-01 1.31198063e-01 -4.31802481e-01
1.70345575e-01 2.42177933e-01 4.03358787e-02 -8.74173701e-01
-3.12652349e-01 1.00412071e+00 -1.91160124e-02 2.25327656e-01
1.47977996e+00 -2.90336967e-01 -5.92586808e-02 1.68323740e-01
1.22849262e+00 1.78793296e-01 -1.89447248e+00 -4.02204275e-01
-3.03987086e-01 -7.47881532e-01 1.22945458e-01 -4.97266054e-01
-1.69420731e+00 5.16015947e-01 4.85454202e-01 -6.71685040e-01
8.97559106e-01 1.00873038e-01 1.04331636e+00 -3.52654629e-03
4.07846987e-01 -7.66075432e-01 7.11426064e-02 2.03562200e-01
1.22284007e+00 -1.21018374e+00 3.61233145e-01 -8.67859602e-01
-3.70315850e-01 1.07412434e+00 9.11113799e-01 -4.35089201e-01
7.58886755e-01 1.76639751e-01 -2.45852396e-02 -3.41930956e-01
-5.76478601e-01 2.97774345e-01 4.72315639e-01 3.62093866e-01
2.00256333e-01 -4.48019877e-02 5.63568659e-02 3.92920822e-01
-4.76442903e-01 5.52378818e-02 3.05465788e-01 6.83622301e-01
-1.94193214e-01 -9.54395592e-01 -3.93440604e-01 5.10623693e-01
-8.57012421e-02 -4.74872738e-02 -2.95165367e-02 5.65743864e-01
1.99551016e-01 5.09934366e-01 2.18054369e-01 -4.21456695e-01
4.87835199e-01 -4.36053395e-01 5.70276737e-01 -4.78658408e-01
-3.09331149e-01 7.32864290e-02 3.86745930e-02 -9.97707129e-01
-3.95152241e-01 -5.35921812e-01 -1.39944792e+00 -2.93680280e-01
-2.04269245e-01 -3.04501295e-01 5.01080990e-01 8.10573697e-01
5.67798078e-01 5.63193083e-01 6.89587891e-01 -1.26835966e+00
-3.36986452e-01 -7.82410920e-01 -1.30281433e-01 4.14706737e-01
2.40675405e-01 -1.08071744e+00 -1.05716191e-01 2.02244341e-01] | [7.14317512512207, -1.2999149560928345] |
60eec990-8ca1-4cf1-9399-679af992e5c6 | b-variational-autoencoders-and-transformers | 2304.03571 | null | https://arxiv.org/abs/2304.03571v1 | https://arxiv.org/pdf/2304.03571v1.pdf | $β$-Variational autoencoders and transformers for reduced-order modelling of fluid flows | Variational autoencoder (VAE) architectures have the potential to develop reduced-order models (ROMs) for chaotic fluid flows. We propose a method for learning compact and near-orthogonal ROMs using a combination of a $\beta$-VAE and a transformer, tested on numerical data from a two-dimensional viscous flow in both periodic and chaotic regimes. The $\beta$-VAE is trained to learn a compact latent representation of the flow velocity, and the transformer is trained to predict the temporal dynamics in latent space. Using the $\beta$-VAE to learn disentangled representations in latent-space, we obtain a more interpretable flow model with features that resemble those observed in the proper orthogonal decomposition, but with a more efficient representation. Using Poincar\'e maps, the results show that our method can capture the underlying dynamics of the flow outperforming other prediction models. The proposed method has potential applications in other fields such as weather forecasting, structural dynamics or biomedical engineering. | ['Ricardo Vinuesa', 'Scott T. M. Dawson', 'Abdulrahman Almashjary', 'Yuning Wang', 'M. A. Gómez', 'Carlos Sanmiguel Vila', 'Alberto Solera-Rico'] | 2023-04-07 | null | null | null | null | ['weather-forecasting'] | ['miscellaneous'] | [-3.20396781e-01 -1.19872138e-01 1.19195737e-01 1.30728677e-01
2.34993458e-01 -3.97679538e-01 6.13959610e-01 -4.45947707e-01
8.34821686e-02 6.08246505e-01 4.92339969e-01 -2.40791723e-01
-4.33677942e-01 -8.51006985e-01 -4.67224061e-01 -1.10751390e+00
-5.75897753e-01 4.15842682e-01 -3.36335868e-01 -3.52478832e-01
1.18156731e-01 5.66444933e-01 -1.28343856e+00 6.13698214e-02
7.78672457e-01 8.61163139e-01 -1.03628196e-01 8.51581454e-01
1.94555968e-01 9.83487308e-01 -2.88308442e-01 5.59867918e-02
3.18340123e-01 -5.68241715e-01 -5.42260051e-01 -2.46446073e-01
-7.02055618e-02 -1.66965336e-01 -7.84008861e-01 6.19015455e-01
1.30177727e-02 5.13567448e-01 1.18639755e+00 -1.01521492e+00
-5.89206159e-01 2.13798314e-01 -1.14434443e-01 3.17976654e-01
-6.24763444e-02 -1.19086709e-02 1.15823126e+00 -7.98061550e-01
7.01496899e-01 1.10544598e+00 7.14942575e-01 6.36399031e-01
-1.44721782e+00 -5.77928841e-01 -4.00824010e-01 -2.82672071e-03
-1.10291040e+00 -2.49642491e-01 1.20457160e+00 -1.03510833e+00
8.47560525e-01 3.25551331e-01 8.53750587e-01 1.20214164e+00
8.60831082e-01 4.09341127e-01 1.00372148e+00 -1.42291591e-01
1.45740867e-01 1.61809132e-01 6.94947019e-02 1.01197541e+00
1.37910053e-01 6.37138724e-01 -5.33735633e-01 -3.97686362e-01
9.77094948e-01 6.49770647e-02 -5.62827528e-01 -6.34393513e-01
-1.10784972e+00 1.28949869e+00 3.38997900e-01 4.60657835e-01
-3.75123799e-01 2.30074977e-03 2.50508130e-01 1.95119560e-01
7.20091283e-01 9.34197366e-01 -3.74687850e-01 -1.37270570e-01
-1.05188608e+00 4.03725922e-01 1.09086239e+00 1.67421579e-01
5.48007548e-01 5.95948875e-01 2.51159877e-01 3.41443747e-01
5.40468395e-01 5.00946045e-01 7.19435453e-01 -1.22764599e+00
2.57408440e-01 3.44564378e-01 1.42337024e-01 -1.32492089e+00
-3.64294529e-01 -3.22478473e-01 -1.26948845e+00 3.17525834e-01
2.43342534e-01 -2.27849931e-01 -4.99222308e-01 1.63486636e+00
-1.80401549e-01 5.65403819e-01 4.95850891e-01 1.01976418e+00
6.10129952e-01 1.11495316e+00 -3.13776582e-01 -4.21391636e-01
1.13222551e+00 -6.60481513e-01 -8.29553425e-01 1.31216198e-01
4.36716020e-01 -3.57690454e-01 5.21104693e-01 1.36327565e-01
-1.02259672e+00 -5.73827505e-01 -1.23174202e+00 3.23948599e-02
-2.07146004e-01 1.14840217e-01 6.08610988e-01 2.89669245e-01
-8.76475632e-01 1.28630972e+00 -1.39454734e+00 4.41489182e-02
-3.09275277e-02 3.84016298e-02 -5.34881175e-01 5.88733673e-01
-1.39317894e+00 8.08430195e-01 5.11827730e-02 3.57075959e-01
-9.13420498e-01 -7.28350461e-01 -1.18871450e+00 3.34577620e-01
-5.07966816e-01 -8.35285842e-01 5.74969232e-01 -3.09797287e-01
-1.78956139e+00 2.17961803e-01 -4.51125979e-01 -3.62295687e-01
3.69282842e-01 -8.49576443e-02 -3.44803751e-01 3.62543643e-01
7.72051886e-02 1.36430711e-01 1.05237341e+00 -1.01674020e+00
2.40337119e-01 -1.23112105e-01 -3.38765562e-01 3.91114615e-02
-4.80714589e-01 -7.13258743e-01 6.23963594e-01 -8.86246741e-01
2.51432866e-01 -1.14691305e+00 -3.06891203e-01 -1.99359640e-01
-1.78417981e-01 -2.92936694e-02 1.20714486e+00 -7.84781754e-01
1.34335101e+00 -1.88509595e+00 9.20295656e-01 1.06492996e-01
6.57291353e-01 3.05364639e-01 2.75814027e-01 5.31985104e-01
-3.66927505e-01 1.94026917e-01 -3.74053806e-01 -4.60614741e-01
-2.42002577e-01 3.78665268e-01 -8.68437409e-01 5.56947470e-01
1.96665823e-01 8.20333838e-01 -6.56687915e-01 -3.31635959e-02
4.45808887e-01 7.73527741e-01 -7.36620605e-01 3.53662968e-01
4.33127955e-02 8.15288305e-01 -4.71959203e-01 -2.01220855e-01
3.49366724e-01 -4.80941147e-01 8.95876437e-02 -1.95618663e-02
-1.62449658e-01 1.65183842e-02 -1.01362693e+00 1.20999491e+00
-5.05873144e-01 1.27343869e+00 3.58169153e-02 -1.39613748e+00
1.13817525e+00 6.36110246e-01 7.68458664e-01 -3.85714859e-01
2.02618942e-01 8.52854997e-02 4.93690521e-02 -8.32627237e-01
4.13708955e-01 -6.30248904e-01 1.79005209e-02 5.80540895e-01
2.53455520e-01 1.42938774e-02 5.50756976e-02 8.91505182e-02
8.18001091e-01 1.96603924e-01 7.18340874e-02 -7.52964437e-01
7.74177194e-01 -3.68647426e-01 6.29244447e-01 3.94130588e-01
-1.01765566e-01 2.58164644e-01 7.03744709e-01 -1.07906175e+00
-1.32341349e+00 -1.00126386e+00 -3.75478655e-01 3.68430287e-01
-2.40769405e-02 -4.87744033e-01 -4.95848954e-01 3.12731904e-03
4.34444658e-02 4.81802106e-01 -9.77679431e-01 -4.03901637e-01
-9.87154782e-01 -7.38697767e-01 4.61895704e-01 4.62861985e-01
3.91644567e-01 -1.12234390e+00 -8.22278857e-01 2.82644004e-01
-2.87895232e-01 -9.60642815e-01 -1.85730577e-01 -5.59586883e-02
-1.01646936e+00 -9.27818716e-01 -7.60454059e-01 -5.18838763e-01
3.94074589e-01 -2.76533127e-01 8.37476850e-01 -2.88254440e-01
-3.61015171e-01 1.67645946e-01 -5.94256222e-02 -3.46411057e-02
-5.84733546e-01 -2.42733255e-01 4.71633494e-01 1.99616402e-01
7.08721802e-02 -8.54338408e-01 -6.32714868e-01 2.23189980e-01
-6.97371185e-01 -4.49211709e-02 5.89605086e-02 1.16745436e+00
2.34456658e-01 5.64661063e-02 4.43898022e-01 -3.79982352e-01
7.77160704e-01 -5.87105274e-01 -5.32302558e-01 -1.60319284e-01
-5.85941494e-01 8.17388117e-01 1.11144412e+00 -4.31104541e-01
-1.03244042e+00 -4.11221117e-01 -6.01251572e-02 -9.93711054e-01
7.70872086e-02 5.06687284e-01 3.81141961e-01 1.77122548e-01
7.04597771e-01 4.75087315e-01 2.91182578e-01 -4.88958895e-01
3.59849147e-02 2.58275360e-01 2.35578328e-01 -6.14284515e-01
6.03734791e-01 5.81953466e-01 4.63836730e-01 -1.16519272e+00
-3.87947977e-01 -1.78107426e-01 -6.09457910e-01 -1.04213603e-01
1.11367261e+00 -7.07835078e-01 -1.12575114e+00 2.64661074e-01
-1.15275562e+00 -1.41631812e-01 -4.57118541e-01 9.85558152e-01
-7.55435467e-01 2.93280303e-01 -1.09700000e+00 -7.57280529e-01
-3.22218418e-01 -1.24299824e+00 8.21515262e-01 1.89835265e-01
-3.16062033e-01 -1.52327561e+00 7.85103381e-01 2.07916107e-02
3.53924423e-01 6.47600055e-01 1.02502823e+00 -1.42693073e-01
-3.03546041e-01 -9.17486567e-03 3.34495395e-01 3.63621265e-01
-3.45827453e-02 1.89897820e-01 -8.18679810e-01 -4.14093971e-01
5.43409407e-01 1.05025932e-01 1.23568666e+00 7.52881825e-01
7.20139265e-01 -7.20098794e-01 -3.26902598e-01 1.03933501e+00
1.14632773e+00 2.37976819e-01 3.20813239e-01 -2.03335419e-01
6.40267670e-01 7.06989467e-01 -3.58086586e-01 4.02886927e-01
8.98901448e-02 2.89835036e-01 1.62251964e-01 2.15042636e-01
3.15794170e-01 -2.40023717e-01 4.87589508e-01 1.44155312e+00
-6.07386172e-01 4.12907936e-02 -8.45399201e-01 3.51647377e-01
-1.74029350e+00 -1.31773841e+00 1.91379953e-02 1.65786445e+00
2.73808867e-01 -1.70101702e-01 -1.78534836e-01 1.68462768e-01
3.27340841e-01 5.97527981e-01 -5.31379282e-01 -8.18767786e-01
-7.24545270e-02 4.37314361e-01 3.26533094e-02 8.41409087e-01
-1.02930713e+00 5.46492457e-01 6.64278364e+00 1.94758981e-01
-1.48571754e+00 -1.22990355e-01 3.14604282e-01 1.76365599e-01
-3.45864117e-01 3.69011052e-02 -3.99503380e-01 4.51981694e-01
1.13309860e+00 -2.66811341e-01 3.72055560e-01 6.25574648e-01
2.63147622e-01 4.80439335e-01 -8.72367978e-01 8.00484836e-01
-9.97193251e-03 -1.71602368e+00 2.57591344e-02 3.30367535e-01
8.32492650e-01 -2.14502469e-01 2.77178526e-01 1.06280394e-01
6.90273882e-04 -1.22368717e+00 4.30633366e-01 1.01675868e+00
7.30244040e-01 -5.35955906e-01 6.57230854e-01 6.64801538e-01
-1.40343678e+00 -1.10003591e-01 -3.21134478e-01 -4.64624614e-01
2.00584173e-01 2.86736876e-01 -4.24893796e-01 4.46161747e-01
6.79779470e-01 1.26415944e+00 -6.54762536e-02 3.48349363e-01
1.30029097e-01 5.73253095e-01 -1.02752738e-01 -4.68368381e-02
1.36657789e-01 -7.40596652e-01 1.05119574e+00 9.27499294e-01
5.74077070e-01 3.64752650e-01 -1.87259048e-01 1.19525206e+00
1.92949384e-01 -2.02838421e-01 -8.29342723e-01 -3.55358779e-01
-1.87205940e-01 1.01584494e+00 -6.26563191e-01 -2.24148914e-01
1.07116774e-01 5.64147890e-01 1.11166582e-01 6.75326467e-01
-6.59020007e-01 -3.56521904e-01 1.02490282e+00 -6.30986914e-02
4.53164071e-01 -5.42377472e-01 -1.12445816e-01 -1.85623682e+00
-9.81897116e-02 -3.44156682e-01 1.23763837e-01 -6.65455043e-01
-1.21631694e+00 1.04814839e+00 1.98296025e-01 -1.48533857e+00
-7.93990672e-01 -1.05151165e+00 -8.01468909e-01 1.07141793e+00
-1.12825513e+00 -6.83793724e-01 -3.70564498e-02 5.55417776e-01
4.37162071e-02 -4.62655604e-01 1.25313473e+00 -1.11935087e-01
-5.44749498e-01 8.15557912e-02 5.52553773e-01 3.52914602e-01
-7.63432980e-02 -1.32671618e+00 6.00014366e-02 6.34968460e-01
2.78938025e-01 8.44329894e-01 1.15031898e+00 -4.32268471e-01
-1.04246008e+00 -7.20920026e-01 6.74517930e-01 -5.22533059e-01
6.21111870e-01 -3.37671965e-01 -1.03522706e+00 7.73237050e-01
2.03656256e-01 4.10304368e-01 8.48322511e-01 -1.47240385e-01
-1.28622562e-01 2.56703440e-02 -8.29605520e-01 3.33497792e-01
4.99468237e-01 -7.25824654e-01 -8.05145264e-01 -1.56054674e-02
5.16657889e-01 -3.14748526e-01 -1.05685866e+00 3.54874104e-01
1.00851011e+00 -1.02800405e+00 9.27681804e-01 -9.51208770e-01
8.48287463e-01 8.19078684e-02 1.11386195e-01 -1.59213972e+00
-3.94135177e-01 -6.94691181e-01 -5.87760150e-01 3.22928280e-01
1.48042753e-01 -8.34787011e-01 9.04681921e-01 4.24562395e-01
-1.96493231e-02 -1.07944357e+00 -1.13074636e+00 -4.41472739e-01
5.69693565e-01 -1.77259430e-01 2.06736684e-01 1.16184509e+00
2.73394048e-01 2.05255568e-01 -6.99880958e-01 2.17714995e-01
6.48266196e-01 3.98195058e-01 2.62585878e-01 -1.51249206e+00
-3.51884812e-01 -4.47462648e-01 -4.35886204e-01 -1.02970743e+00
6.37557268e-01 -9.28749323e-01 -2.61115462e-01 -1.07433891e+00
-2.28584304e-01 -9.86104459e-02 -1.86058983e-01 -4.47989665e-02
2.32233286e-01 4.32154089e-02 1.56613588e-01 4.99432683e-01
2.15763420e-01 1.12561023e+00 1.34605753e+00 2.21289787e-02
-4.23089266e-01 -6.42760172e-02 -1.27016306e-01 8.78662944e-01
6.05556548e-01 -4.03944582e-01 -3.15139025e-01 -1.13738477e-01
2.20405802e-01 7.61494696e-01 4.52748716e-01 -9.79970038e-01
1.46404818e-01 6.33157492e-02 5.63945413e-01 -2.35138834e-01
5.70424259e-01 -5.49962878e-01 6.77867383e-02 7.29222834e-01
-4.14665490e-01 2.52589822e-01 2.16221005e-01 6.44607306e-01
-4.88343477e-01 9.69811752e-02 6.27365470e-01 -2.11037531e-01
-1.76621988e-01 3.95341456e-01 -6.01669073e-01 -4.41614278e-02
9.23502386e-01 -5.23403510e-02 -2.41860285e-01 -4.64005172e-01
-1.25906014e+00 3.53436694e-02 -6.10232353e-04 4.97549713e-01
6.99755132e-01 -1.41244590e+00 -6.03958666e-01 9.75365043e-01
-2.81710058e-01 -2.63895929e-01 5.16910136e-01 6.19585693e-01
-9.33654189e-01 8.51638496e-01 -5.81062675e-01 -7.11936235e-01
-6.52677834e-01 3.28102767e-01 6.88461959e-01 -5.90287864e-01
-8.35430980e-01 6.16440654e-01 3.93756598e-01 -5.02928197e-01
-4.73541766e-01 -3.23953271e-01 -5.64930439e-01 8.36460441e-02
3.04346204e-01 5.94883144e-01 -4.88802433e-01 -1.26028550e+00
-8.73301774e-02 7.61776686e-01 4.35963929e-01 -3.24639022e-01
1.47164845e+00 1.61019608e-01 -2.40729615e-01 6.00740969e-01
1.54133928e+00 -9.01883934e-03 -1.48731303e+00 1.37724001e-02
-3.88240039e-01 -1.74214929e-01 7.81952962e-02 8.90777931e-02
-1.16327405e+00 1.36802864e+00 3.35593581e-01 5.15137911e-01
6.39559448e-01 -1.80741265e-01 6.99498832e-01 2.07119122e-01
-8.95030722e-02 -5.02992868e-01 5.49609587e-02 6.25039279e-01
1.02657676e+00 -9.10003901e-01 -2.01555237e-01 1.51439637e-01
-5.54141700e-01 1.61936784e+00 3.06383640e-01 -6.42664373e-01
1.31207955e+00 1.10924870e-01 -6.69633001e-02 -2.09246263e-01
-8.81486475e-01 4.38209653e-01 6.31500185e-01 1.25108317e-01
2.38537803e-01 1.28684595e-01 -1.49251716e-02 5.10812342e-01
-5.78502834e-01 -3.43764067e-01 7.17265725e-01 3.64149660e-01
3.48493713e-03 -7.73818076e-01 -1.79241270e-01 3.69676948e-01
-2.88018286e-01 2.81887710e-01 2.40628809e-01 7.32005119e-01
2.12540403e-02 5.35703778e-01 4.20299113e-01 -3.83445382e-01
2.35712714e-02 3.51671934e-01 1.28049463e-01 -1.88995764e-01
-8.56833905e-02 4.16640826e-02 -5.24783850e-01 -5.88641286e-01
-3.77217412e-01 -6.72528923e-01 -8.88458312e-01 -4.30433214e-01
-4.60288003e-02 7.30031967e-01 2.81235188e-01 8.96097302e-01
3.11456323e-01 5.28731108e-01 7.39073098e-01 -1.13848948e+00
-3.29612732e-01 -1.07474542e+00 -8.08694184e-01 4.95061606e-01
1.08148742e+00 -9.14185762e-01 -9.30818677e-01 1.46785840e-01] | [6.571417808532715, 3.4728281497955322] |
7089890c-ae2e-406d-b7d5-0d39721d2696 | efficient-online-decision-tree-learning-with | 2305.02093 | null | https://arxiv.org/abs/2305.02093v1 | https://arxiv.org/pdf/2305.02093v1.pdf | Efficient Online Decision Tree Learning with Active Feature Acquisition | Constructing decision trees online is a classical machine learning problem. Existing works often assume that features are readily available for each incoming data point. However, in many real world applications, both feature values and the labels are unknown a priori and can only be obtained at a cost. For example, in medical diagnosis, doctors have to choose which tests to perform (i.e., making costly feature queries) on a patient in order to make a diagnosis decision (i.e., predicting labels). We provide a fresh perspective to tackle this practical challenge. Our framework consists of an active planning oracle embedded in an online learning scheme for which we investigate several information acquisition functions. Specifically, we employ a surrogate information acquisition function based on adaptive submodularity to actively query feature values with a minimal cost, while using a posterior sampling scheme to maintain a low regret for online prediction. We demonstrate the efficiency and effectiveness of our framework via extensive experiments on various real-world datasets. Our framework also naturally adapts to the challenging setting of online learning with concept drift and is shown to be competitive with baseline models while being more flexible. | ['Morteza Haghir Chehreghani', 'Yuxin Chen', 'Ziyu Ye', 'Arman Rahbar'] | 2023-05-03 | null | null | null | null | ['medical-diagnosis'] | ['medical'] | [ 5.37394047e-01 6.06871724e-01 -6.03211820e-01 -6.83503032e-01
-1.20575380e+00 -5.67028046e-01 -4.09270227e-02 6.91023231e-01
-4.22220111e-01 7.03818560e-01 -2.55250186e-01 -3.12476128e-01
-5.04869401e-01 -7.72856772e-01 -9.28471148e-01 -9.12386358e-01
-2.21355259e-02 9.21555638e-01 -8.07303041e-02 2.99213111e-01
-6.38489053e-02 1.57812834e-01 -1.16544569e+00 -7.74879903e-02
9.20605659e-01 1.31653059e+00 2.15028495e-01 5.46883523e-01
2.46115804e-01 6.33402348e-01 -2.26683721e-01 -3.84425789e-01
5.89383125e-01 -2.54252970e-01 -8.89440119e-01 7.95776427e-01
-4.97974381e-02 -5.79121649e-01 5.49734682e-02 1.07337797e+00
3.51378798e-01 1.14808299e-01 3.53896350e-01 -1.17134464e+00
-1.66561544e-01 6.09942079e-01 -3.60124826e-01 -8.92850086e-02
2.21033186e-01 5.22928983e-02 1.32778049e+00 -4.04765368e-01
5.40206790e-01 8.53036404e-01 3.36931020e-01 5.97095072e-01
-1.30187690e+00 -1.61207438e-01 5.55400074e-01 1.68857902e-01
-1.02352059e+00 -4.53975886e-01 6.27894700e-01 -3.24403405e-01
1.97811380e-01 3.23371977e-01 6.20045960e-01 7.07685471e-01
1.60492063e-01 1.01164722e+00 8.54518473e-01 -3.25710654e-01
8.08204353e-01 3.70056480e-01 6.96272701e-02 7.15174317e-01
1.66648865e-01 1.33911353e-02 -6.22671366e-01 -4.44107234e-01
5.75238526e-01 5.54366231e-01 -3.04879934e-01 -8.04204345e-01
-9.75569248e-01 1.00072515e+00 4.27506626e-01 -2.66716540e-01
-5.92925489e-01 -5.29331900e-02 1.13612317e-01 5.21307588e-01
4.75296557e-01 2.55604219e-02 -6.66065931e-01 -2.77793743e-02
-5.27283669e-01 3.45580578e-01 9.85811174e-01 9.71952736e-01
7.85450995e-01 -6.60325885e-01 -3.25277776e-01 6.79220915e-01
2.34681323e-01 2.60964870e-01 1.98563620e-01 -1.09023583e+00
6.23803377e-01 5.46500027e-01 6.16499424e-01 -5.66717684e-01
-4.00031656e-01 -4.78065223e-01 -5.55164278e-01 -3.79266664e-02
5.33870101e-01 -3.73504311e-01 -7.32819378e-01 1.75526476e+00
9.86205757e-01 4.57641333e-02 -1.02437072e-01 8.99613857e-01
-8.37811306e-02 3.67778003e-01 -2.82725334e-01 -5.61539829e-01
1.06670880e+00 -8.69887710e-01 -7.13526607e-01 -3.10161442e-01
8.85032117e-01 -1.05654351e-01 7.24827349e-01 7.57405579e-01
-9.87647474e-01 1.09027289e-01 -9.08613980e-01 1.40033796e-01
1.70731872e-01 -1.14137337e-01 8.66217852e-01 6.91432536e-01
-6.63440228e-01 7.49571800e-01 -1.33114064e+00 -1.05247445e-01
7.00322926e-01 5.15606999e-01 -1.63135871e-01 -4.47118878e-01
-7.42498517e-01 3.67025793e-01 4.03010607e-01 2.71552563e-01
-1.19058001e+00 -5.54916441e-01 -7.23378658e-01 2.44798651e-03
1.24743187e+00 -9.03655052e-01 1.77354407e+00 -1.15487850e+00
-1.53107285e+00 4.38219726e-01 -6.82172999e-02 -7.35273421e-01
8.87123764e-01 -2.15729296e-01 1.59980372e-01 1.34594142e-01
5.60702682e-02 2.60179132e-01 7.74407268e-01 -7.43137062e-01
-1.08746731e+00 -6.71890855e-01 6.18393540e-01 4.44952041e-01
-4.51082110e-01 -3.62396777e-01 -2.26108432e-01 -1.28972307e-01
3.77081692e-01 -1.01828671e+00 -7.24117219e-01 3.99445713e-01
-4.23613906e-01 -1.73398182e-01 3.79020274e-01 -4.96574879e-01
1.03058982e+00 -1.98022032e+00 1.67534515e-01 2.57965744e-01
2.41259351e-01 -4.06908840e-01 3.45368862e-01 2.77900934e-01
3.88504237e-01 -1.39669582e-01 -6.63263977e-01 -3.56215775e-01
5.88334911e-02 4.21747983e-01 -1.35772839e-01 5.85007250e-01
-2.09825523e-02 8.45605195e-01 -1.07131612e+00 -3.56591314e-01
-2.59436131e-01 -1.61527544e-02 -7.53647923e-01 2.01600511e-02
-5.17882466e-01 4.25170124e-01 -9.14035082e-01 8.67009819e-01
3.70691806e-01 -6.79355323e-01 4.46324974e-01 6.75749660e-01
2.97853321e-01 2.08832428e-01 -1.39095163e+00 1.67551219e+00
-6.07441247e-01 5.54547459e-02 2.67422915e-01 -1.25109744e+00
3.57131273e-01 3.53411824e-01 6.02839768e-01 -2.14212716e-01
-3.60294767e-02 3.14200699e-01 1.69982724e-02 -3.50829631e-01
-5.86741976e-02 -2.15217695e-01 7.48217572e-03 3.86587024e-01
-2.50465184e-01 3.30893070e-01 -1.45848021e-01 6.20632246e-02
1.39658558e+00 -1.45391330e-01 4.79998469e-01 -5.15255332e-02
3.08675528e-01 2.70924577e-03 1.04629195e+00 9.53295350e-01
-3.14475000e-02 3.46723735e-01 8.10548246e-01 -4.30196017e-01
-5.97061217e-01 -8.30004752e-01 -1.85935214e-01 9.25725639e-01
-7.72782713e-02 1.02702063e-02 -6.55672610e-01 -1.16211557e+00
2.36104444e-01 4.92751032e-01 -7.35332072e-01 1.19829834e-01
-1.59948483e-01 -7.26687968e-01 -4.40353245e-01 4.00943041e-01
7.54161254e-02 -7.30612636e-01 -9.46868360e-01 5.33079028e-01
-2.66121458e-02 -7.57290184e-01 -6.26372099e-01 4.35725152e-01
-1.15082109e+00 -1.11896253e+00 -5.10251105e-01 -5.03664017e-01
9.07366276e-01 1.71329081e-01 8.56436253e-01 -4.50286008e-02
-2.79549956e-01 5.47760129e-01 -3.58178169e-01 -5.90166271e-01
-6.00298261e-03 2.00029790e-01 -7.53859356e-02 4.71487045e-01
2.38660490e-03 -2.39772841e-01 -8.60627234e-01 1.76495209e-01
-9.98839974e-01 -4.52890284e-02 6.52949810e-01 1.16284049e+00
9.88464355e-01 1.77909374e-01 5.85379779e-01 -1.56650579e+00
2.31810659e-01 -6.50270581e-01 -9.91728127e-01 3.39257032e-01
-8.69709015e-01 1.00893416e-01 5.41666090e-01 -3.59690696e-01
-9.26091373e-01 5.63853979e-01 1.65552840e-01 -3.77572298e-01
1.41358942e-01 6.41458154e-01 -3.15690994e-01 4.98223081e-02
3.28457355e-01 6.75243512e-02 -8.63310844e-02 -3.49474967e-01
5.17946109e-02 5.89627564e-01 6.87698200e-02 -5.10039508e-01
5.96513033e-01 6.62193656e-01 1.15658060e-01 -3.46531987e-01
-1.36412203e+00 -6.25430882e-01 -5.71838498e-01 1.04490407e-02
3.68023068e-01 -8.90284359e-01 -8.46615553e-01 2.09226906e-01
-7.36692011e-01 -3.76841903e-01 -6.54800773e-01 3.24804753e-01
-7.19540834e-01 1.03299413e-02 -3.73262137e-01 -1.20466828e+00
-1.64386541e-01 -1.03138602e+00 1.13937044e+00 1.39141068e-01
3.32429469e-01 -1.12725055e+00 -6.50798529e-02 5.82494974e-01
8.20458531e-02 3.71566504e-01 7.54664302e-01 -6.64965570e-01
-9.88366127e-01 -2.95914441e-01 2.73633629e-01 1.05478957e-01
1.89859331e-01 -5.31687796e-01 -8.17099273e-01 -4.39305156e-01
3.34915221e-01 -5.42445123e-01 7.31759489e-01 4.05919135e-01
1.39257026e+00 -8.80785584e-01 -3.88887286e-01 4.49517131e-01
1.32149220e+00 3.52655977e-01 1.17103914e-02 1.42970175e-01
4.16137248e-01 7.21753061e-01 1.08583522e+00 1.03939259e+00
5.80915391e-01 7.15770602e-01 4.78993326e-01 4.41152900e-01
6.82079792e-01 -2.86981940e-01 1.40246615e-01 4.84941959e-01
3.93670708e-01 -2.03143984e-01 -8.86149645e-01 6.59184992e-01
-2.18998504e+00 -3.34912866e-01 4.42300797e-01 2.66639924e+00
1.04005623e+00 2.08877310e-01 2.63243169e-01 1.10297665e-01
5.33782542e-01 -2.64172107e-01 -1.37995589e+00 2.42919195e-02
3.19054484e-01 5.33309393e-03 7.10263133e-01 4.34070915e-01
-1.08773077e+00 3.84445965e-01 4.98263884e+00 3.96466702e-01
-9.38544035e-01 3.44780177e-01 1.03701496e+00 -4.69192654e-01
-3.39993149e-01 2.14959741e-01 -8.31051588e-01 3.08311760e-01
8.58101666e-01 -1.82681531e-01 4.87470031e-01 1.13119054e+00
5.55829369e-02 -1.96624786e-01 -1.58886909e+00 9.34083104e-01
-4.17361081e-01 -9.88079011e-01 -4.84674007e-01 2.75535405e-01
7.63268888e-01 -1.74674720e-01 -2.24318225e-02 1.19973630e-01
4.29733276e-01 -6.44563735e-01 5.90439618e-01 4.60671842e-01
5.26618123e-01 -8.90213430e-01 5.54266274e-01 7.65418053e-01
-7.40040243e-01 -6.17267251e-01 -3.71235341e-01 3.89747927e-03
1.85835034e-01 8.79582942e-01 -9.73391175e-01 6.34458959e-01
4.02646124e-01 5.92592895e-01 -1.50623560e-01 1.33899117e+00
-2.01240048e-01 6.27138019e-01 -4.95697051e-01 -3.80219482e-02
1.88058883e-01 -1.70343220e-01 2.58214593e-01 5.05204082e-01
2.06847280e-01 1.87684551e-01 5.41496158e-01 4.95257795e-01
-1.11320376e-01 2.14711234e-01 -3.44542503e-01 -8.26879144e-02
5.62193274e-01 1.17668235e+00 -6.93498373e-01 -6.22472204e-02
-3.50226283e-01 1.02171266e+00 5.19837379e-01 5.93460836e-02
-5.80888927e-01 -9.55290571e-02 4.53673273e-01 9.15210545e-02
3.41257602e-01 1.75755337e-01 -2.93239862e-01 -1.05859649e+00
5.31708360e-01 -6.73093677e-01 6.87318504e-01 2.19803471e-02
-1.25264347e+00 3.45508717e-02 -2.51598358e-01 -1.09311855e+00
-4.93291974e-01 -2.86891013e-01 -1.70172572e-01 4.99098033e-01
-1.60007906e+00 -6.35642648e-01 -1.08084396e-01 5.28873384e-01
4.96461272e-01 4.09698755e-01 5.57779551e-01 7.88101330e-02
-6.55916631e-01 5.35808265e-01 2.88863420e-01 -2.96439946e-01
4.57082987e-01 -1.52593756e+00 8.72251913e-02 5.22367954e-01
9.46331620e-02 2.40761176e-01 4.16799664e-01 -5.02322078e-01
-1.79058409e+00 -1.20377660e+00 6.99214041e-01 -2.12365881e-01
5.33694267e-01 -5.72366595e-01 -8.49413097e-01 8.69858682e-01
-5.29977441e-01 3.04433316e-01 5.73408604e-01 1.74186334e-01
1.95410743e-01 -4.84269500e-01 -1.34885585e+00 2.24524334e-01
1.12796092e+00 -2.91617185e-01 -3.20463441e-02 8.84002149e-01
8.44899774e-01 -7.31805563e-01 -8.80949080e-01 2.81755090e-01
4.82957691e-01 -6.15510464e-01 5.00344813e-01 -6.27587795e-01
1.35652229e-01 -3.64654809e-02 -5.00171632e-02 -1.02218759e+00
-1.03256732e-01 -1.01373255e+00 -4.02147412e-01 9.05333698e-01
6.72345161e-01 -1.07403111e+00 1.04169893e+00 1.17533898e+00
2.49224737e-01 -1.26162112e+00 -1.16150498e+00 -7.20682383e-01
-3.16027015e-01 -3.26235861e-01 4.43869889e-01 7.30718255e-01
-6.08412586e-02 -1.86345745e-02 -3.73377860e-01 5.04105747e-01
7.01029062e-01 3.41823161e-01 5.34609556e-01 -1.32510364e+00
-9.12076771e-01 1.31151408e-01 -1.35662869e-01 -1.22713149e+00
-8.20516795e-02 -7.19625175e-01 3.08182240e-01 -1.31370807e+00
4.34989095e-01 -8.01353216e-01 -4.46137756e-01 6.48771644e-01
-2.92582929e-01 -5.01031518e-01 5.83160035e-02 1.21070251e-01
-9.02947426e-01 4.96665508e-01 1.08271945e+00 4.07901704e-02
-2.63339937e-01 7.47037351e-01 -9.17083561e-01 5.69278836e-01
6.31026804e-01 -7.98283160e-01 -6.58950984e-01 -3.06382269e-01
3.44809294e-01 8.37363124e-01 1.29557312e-01 -5.86010456e-01
5.45191348e-01 -4.46646214e-01 9.94147435e-02 -3.42313200e-01
2.37313747e-01 -9.76964951e-01 2.46392228e-02 6.51206434e-01
-5.12307346e-01 -2.45654568e-01 -4.61072356e-01 1.12092078e+00
2.23548226e-02 -4.26296651e-01 6.92494035e-01 -7.68178627e-02
-2.18099773e-01 8.21781456e-01 1.43894657e-01 1.18756428e-01
1.25776303e+00 1.62601382e-01 1.89161256e-01 -4.20602709e-01
-8.49718034e-01 7.13391006e-01 4.24577981e-01 3.40399854e-02
5.36285996e-01 -9.85034883e-01 -5.28064668e-01 8.66259709e-02
1.55231252e-01 6.38686299e-01 3.47099572e-01 9.14877951e-01
-9.01992470e-02 4.15396512e-01 4.33555216e-01 -5.72965503e-01
-9.24543798e-01 7.26935685e-01 1.81454599e-01 -5.96242666e-01
-5.94655693e-01 7.49693930e-01 2.70039111e-01 -3.57129931e-01
5.90236485e-01 -3.36923271e-01 1.51209429e-01 8.34775120e-02
4.70226496e-01 4.04105008e-01 2.62743831e-01 2.69369274e-01
-1.40226424e-01 2.71309949e-02 -4.49326485e-01 -2.06487983e-01
1.39413977e+00 -2.60807455e-01 1.91284299e-01 7.15955257e-01
7.35624731e-01 -3.53206158e-01 -1.72041094e+00 -7.90672600e-01
3.42457116e-01 -4.55637127e-01 1.66702289e-02 -8.83401513e-01
-1.14011633e+00 4.78959352e-01 5.45530021e-01 3.25630903e-01
1.34508789e+00 -3.81545015e-02 7.41429329e-01 8.42024565e-01
7.75065958e-01 -1.32605577e+00 -2.74357706e-01 -4.66527045e-02
5.09254634e-01 -1.50444746e+00 -1.55529410e-01 -4.06873047e-01
-6.94500625e-01 9.94813442e-01 2.77088702e-01 1.23738356e-01
7.16034770e-01 5.26058450e-02 -8.27208534e-02 3.66003923e-02
-1.28447461e+00 -1.71021577e-02 -3.33170928e-02 9.41463411e-02
-1.59651004e-02 3.27954262e-01 -1.33841559e-01 6.80851758e-01
6.35968000e-02 3.65318507e-02 4.18703467e-01 1.21241248e+00
-4.21031356e-01 -1.26255667e+00 -2.14582026e-01 8.06425095e-01
-7.22140849e-01 2.02168852e-01 -2.78213799e-01 3.43671799e-01
-8.91685337e-02 1.02597225e+00 -1.10517375e-01 7.11720288e-02
1.71207160e-01 6.73106909e-02 3.80529165e-01 -9.09812391e-01
-3.09516460e-01 1.38322860e-01 -6.49731159e-02 -7.08816350e-01
-2.60325849e-01 -1.11323774e+00 -1.00331688e+00 2.58296818e-01
-3.79048735e-01 -4.28811274e-03 6.51865959e-01 1.07794333e+00
2.32678637e-01 3.45706582e-01 1.21945643e+00 -3.14225554e-01
-1.30386639e+00 -6.19419158e-01 -9.17158186e-01 7.41478652e-02
5.46756506e-01 -5.17393172e-01 -1.79815024e-01 -7.27616474e-02] | [8.229109764099121, 4.118100166320801] |
da625c08-81c1-46ed-86dd-5998d07f73a2 | you-only-crash-once-improved-object-detection | 2303.04891 | null | https://arxiv.org/abs/2303.04891v1 | https://arxiv.org/pdf/2303.04891v1.pdf | You Only Crash Once: Improved Object Detection for Real-Time, Sim-to-Real Hazardous Terrain Detection and Classification for Autonomous Planetary Landings | The detection of hazardous terrain during the planetary landing of spacecraft plays a critical role in assuring vehicle safety and mission success. A cheap and effective way of detecting hazardous terrain is through the use of visual cameras, which ensure operational ability from atmospheric entry through touchdown. Plagued by resource constraints and limited computational power, traditional techniques for visual hazardous terrain detection focus on template matching and registration to pre-built hazard maps. Although successful on previous missions, this approach is restricted to the specificity of the templates and limited by the fidelity of the underlying hazard map, which both require extensive pre-flight cost and effort to obtain and develop. Terrestrial systems that perform a similar task in applications such as autonomous driving utilize state-of-the-art deep learning techniques to successfully localize and classify navigation hazards. Advancements in spacecraft co-processors aimed at accelerating deep learning inference enable the application of these methods in space for the first time. In this work, we introduce You Only Crash Once (YOCO), a deep learning-based visual hazardous terrain detection and classification technique for autonomous spacecraft planetary landings. Through the use of unsupervised domain adaptation we tailor YOCO for training by simulation, removing the need for real-world annotated data and expensive mission surveying phases. We further improve the transfer of representative terrain knowledge between simulation and the real world through visual similarity clustering. We demonstrate the utility of YOCO through a series of terrestrial and extraterrestrial simulation-to-real experiments and show substantial improvements toward the ability to both detect and accurately classify instances of planetary terrain. | ['Karthik Dantu', 'John Crassidis', 'Chris Gnam', 'Timothy Chase Jr'] | 2023-03-08 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [ 5.57865128e-02 -2.33870819e-01 1.34608507e-01 -2.49675646e-01
-8.37624192e-01 -8.53731811e-01 8.10840786e-01 1.42271638e-01
-5.70234299e-01 5.34949064e-01 -2.36318782e-01 -6.00278020e-01
-2.19193771e-01 -7.89569378e-01 -6.25427186e-01 -4.41514075e-01
-7.25119174e-01 8.58991861e-01 4.51698005e-01 -7.08561599e-01
1.52741268e-01 7.71643937e-01 -1.63138020e+00 -7.28462413e-02
6.61810517e-01 1.03407013e+00 2.25124925e-01 4.47199762e-01
3.25220823e-01 2.48936236e-01 -4.73183155e-01 3.99583995e-01
7.77437449e-01 -2.36344337e-03 -4.08174574e-01 -4.52302694e-02
3.70470345e-01 -1.95502385e-01 -3.21158707e-01 5.89499950e-01
4.18714195e-01 3.18179548e-01 9.42085445e-01 -1.24514687e+00
3.59545141e-01 -4.21580195e-01 -4.49409872e-01 1.25469357e-01
1.53622702e-01 4.14072871e-01 5.19574702e-01 -7.68993556e-01
3.42931837e-01 5.29287994e-01 1.02721334e+00 -1.86349414e-02
-1.04540336e+00 -7.33342886e-01 -7.92538300e-02 1.69660404e-01
-1.45969355e+00 -3.82148296e-01 3.39981765e-01 -9.46588933e-01
1.29487610e+00 4.96501513e-02 6.36816323e-01 8.05793405e-01
3.98151398e-01 -9.44689102e-03 8.67170453e-01 -1.73945725e-01
4.01588231e-01 -2.18305096e-01 -4.04405653e-01 9.55080092e-01
4.61727202e-01 6.78641617e-01 -5.08744776e-01 -6.23570606e-02
5.52041054e-01 -3.38528752e-01 -3.96771580e-01 -7.16619074e-01
-1.12996769e+00 7.54603922e-01 5.60732484e-01 -1.78810045e-01
-1.99232936e-01 8.64345487e-03 3.82235616e-01 3.38728547e-01
1.06205449e-01 6.72834754e-01 -3.94931957e-02 7.21954480e-02
-1.14401710e+00 2.50155658e-01 6.49845183e-01 7.60222316e-01
7.52362370e-01 5.49049199e-01 5.41969121e-01 2.97418833e-01
2.52927601e-01 6.82577074e-01 2.27847368e-01 -7.51720965e-01
3.27837467e-01 4.23051983e-01 3.44531208e-01 -7.35933542e-01
-6.88658655e-01 -4.67768788e-01 -4.93718207e-01 1.35746121e+00
2.81315237e-01 -5.01263291e-02 -1.36193609e+00 1.24303615e+00
3.50550473e-01 -2.12942004e-01 2.49343023e-01 1.12711310e+00
2.41969943e-01 5.18187344e-01 1.22750223e-01 5.32926500e-01
1.11326730e+00 -4.74442959e-01 1.35551140e-01 -9.43694651e-01
6.33469701e-01 -4.76222992e-01 9.23828244e-01 5.49833655e-01
-2.70748943e-01 -3.63739699e-01 -1.74093139e+00 1.16684183e-01
-5.76570690e-01 -1.56921774e-01 5.87895274e-01 4.43951666e-01
-8.44466984e-01 5.24995685e-01 -8.27782273e-01 -4.76133794e-01
3.78600180e-01 5.43192983e-01 -6.77280307e-01 4.44334485e-02
-1.15792286e+00 1.33214843e+00 2.24646196e-01 -5.47254384e-02
-1.69810057e+00 -4.93705451e-01 -1.09039044e+00 -4.56032008e-02
1.97486728e-01 -6.47651136e-01 1.14086831e+00 -9.77103412e-01
-7.52921879e-01 9.05397832e-01 2.11144984e-01 -8.23311031e-01
5.20734251e-01 -1.95034072e-01 -2.37118259e-01 3.54987420e-02
1.87887296e-01 6.12782657e-01 8.42966914e-01 -1.32269752e+00
-8.22363019e-01 -1.87057331e-01 -1.65656477e-01 5.59640229e-01
-7.74574373e-03 -2.49890491e-01 -2.29276687e-01 -2.50589818e-01
-3.54004209e-03 -9.01003122e-01 -2.17968866e-01 3.84583950e-01
2.13853195e-01 2.15621501e-01 9.06677663e-01 -8.68395686e-01
4.87728894e-01 -2.11548710e+00 7.25500584e-02 5.46094179e-01
1.86457321e-01 -4.33123037e-02 3.08292449e-01 5.88431537e-01
1.82001162e-02 -1.94482014e-01 -6.22477293e-01 -1.13894328e-01
2.26528402e-02 5.09180762e-02 -4.15289521e-01 8.08354914e-01
-2.55158525e-02 3.36565644e-01 -6.29476607e-01 -2.86673963e-01
3.25149745e-01 1.82907090e-01 -1.04058377e-01 1.52276114e-01
-2.65141577e-01 3.26585144e-01 -1.70757711e-01 7.94303358e-01
4.93585974e-01 2.95876712e-01 1.75243467e-01 1.02065489e-01
-5.80764890e-01 4.07189488e-01 -7.85016716e-01 1.39716423e+00
-3.94742817e-01 9.82471466e-01 3.91803831e-01 -8.07959259e-01
1.06629097e+00 -3.08754575e-02 1.63113937e-01 -1.01416111e+00
5.78381270e-02 2.90328979e-01 -1.64029196e-01 -3.57629001e-01
6.70507252e-01 -5.98390341e-01 -4.76679981e-01 1.18944727e-01
-2.25477561e-01 -5.63368976e-01 -2.01918319e-01 4.30178344e-02
9.47979450e-01 3.37261468e-01 2.14355156e-01 -4.18363243e-01
-2.19229922e-01 1.00822425e+00 4.77938622e-01 5.33298314e-01
-2.10520282e-01 6.54947400e-01 1.60592109e-01 -5.37032664e-01
-1.31124926e+00 -1.15596604e+00 -2.89235301e-02 8.26668739e-01
3.60159725e-01 8.34582001e-02 -2.84493953e-01 -5.72004139e-01
3.79812837e-01 5.97312450e-01 -4.53837365e-01 -3.30599368e-01
-2.89756060e-01 -6.31028712e-01 7.29561031e-01 5.95177710e-01
3.14695954e-01 -7.10678518e-01 -1.14284837e+00 9.18856859e-02
-6.19223677e-02 -8.13623011e-01 2.20618591e-01 6.57335520e-01
-6.09463453e-01 -9.86415207e-01 -3.24587613e-01 -6.91863179e-01
4.07632679e-01 4.97347116e-01 8.53443325e-01 -1.03085123e-01
-5.32758057e-01 2.07445458e-01 -6.64002970e-02 -5.14612317e-01
-9.66220126e-02 -2.66177416e-01 4.47241366e-01 -4.49128509e-01
5.17359041e-02 -5.08848786e-01 -4.67777342e-01 5.42119741e-01
-4.05808717e-01 3.11682783e-02 5.97396195e-01 6.00629389e-01
5.02498031e-01 2.00991258e-01 3.79355662e-02 -1.30894154e-01
2.47098133e-01 -6.72167838e-01 -1.00327659e+00 -7.05373138e-02
-5.46201348e-01 -3.11275814e-02 5.22412181e-01 -9.40321386e-02
-7.32455432e-01 3.22244734e-01 1.10647485e-01 -5.09005249e-01
-1.48956448e-01 7.99471021e-01 1.58507764e-01 -7.45308518e-01
1.09382391e+00 4.96696420e-02 7.05250129e-02 -2.38854453e-01
-1.18702009e-01 6.11888289e-01 8.69542062e-01 -4.37797248e-01
1.14612794e+00 8.01816702e-01 1.99243695e-01 -9.96626616e-01
-3.93719316e-01 -4.57199723e-01 -5.61151922e-01 -5.09937167e-01
7.58758307e-01 -1.19524133e+00 -3.41898412e-01 1.59580782e-01
-4.19323713e-01 -8.21890891e-01 1.11256070e-01 4.46228713e-01
-5.10687530e-01 5.81957698e-01 1.77830577e-01 -7.39196599e-01
-2.37916142e-01 -7.89104581e-01 1.11762357e+00 1.16253756e-01
-4.00552958e-01 -7.64356732e-01 2.44271502e-01 1.14914320e-01
3.72223765e-01 5.74422777e-01 7.41098762e-01 -1.82919666e-01
-6.90759718e-01 -3.84133846e-01 -9.74139199e-02 -1.96031943e-01
-1.37214020e-01 -4.49057102e-01 -9.04542983e-01 -5.71085274e-01
-1.38275996e-01 -3.84136945e-01 1.03989828e+00 -7.94186536e-03
4.26849604e-01 4.27118748e-01 -5.75265527e-01 1.04035854e+00
1.39726770e+00 3.61442715e-01 6.09474182e-01 1.19832361e+00
4.36048031e-01 9.43533063e-01 1.14306962e+00 4.20124501e-01
2.80644208e-01 7.34487474e-01 8.33685160e-01 -4.50349659e-01
1.35724396e-02 -8.17396790e-02 4.34307396e-01 -1.31986126e-01
-4.04967666e-02 8.18655267e-03 -1.56272697e+00 7.92816103e-01
-1.56681526e+00 -8.11053276e-01 1.23432307e-02 2.67738819e+00
2.34682932e-01 5.84210396e-01 -2.96412613e-02 -1.17700160e-01
2.57460028e-01 2.53430568e-02 -5.31233251e-01 -1.57960594e-01
-1.71908773e-02 -4.26303782e-02 1.05610251e+00 5.69138169e-01
-1.40552080e+00 9.50097084e-01 5.81884146e+00 4.73191053e-01
-1.22562933e+00 -1.05737515e-01 -5.39915496e-03 -1.42583162e-01
-1.18031897e-01 1.64348751e-01 -5.92503607e-01 6.55470490e-02
9.62819099e-01 -8.12568143e-02 3.83519351e-01 1.02146888e+00
1.67538032e-01 -4.91874307e-01 -8.83929729e-01 7.62767792e-01
8.03253725e-02 -1.11789870e+00 -3.75493765e-01 9.45991557e-03
1.65070668e-01 5.35580993e-01 -4.39507514e-03 3.66571337e-01
2.40681723e-01 -1.06165707e+00 1.01127648e+00 4.45378035e-01
1.17034161e+00 -7.48884737e-01 4.77902204e-01 1.77583009e-01
-1.27586782e+00 -2.13912815e-01 -3.95424902e-01 -1.03564814e-01
8.11093450e-02 8.25842991e-02 -1.32604849e+00 6.04556084e-01
1.02063084e+00 2.46472239e-01 -3.17072928e-01 1.34961331e+00
-8.04962888e-02 3.14712137e-01 -7.30748832e-01 2.41208032e-01
4.95674580e-01 4.61655222e-02 6.63505852e-01 1.08403754e+00
5.31231999e-01 -2.41562158e-01 3.58219117e-01 5.39743483e-01
4.08820271e-01 -5.83683670e-01 -1.16370487e+00 1.20266356e-01
2.81084687e-01 1.17297661e+00 -6.38691366e-01 -1.03068300e-01
-2.39564389e-01 6.38769090e-01 1.18830010e-01 1.56525448e-01
-7.80322254e-01 -6.00733757e-01 9.07334149e-01 5.75084507e-01
2.56222755e-01 -1.00378931e+00 -6.40807152e-01 -5.22189260e-01
-5.06137088e-02 -8.89994919e-01 2.51684815e-01 -9.58949924e-01
-7.61988819e-01 6.60549462e-01 6.17327690e-02 -1.82873774e+00
-1.21507473e-01 -6.78749919e-01 -6.13369107e-01 1.14420974e+00
-1.58959329e+00 -1.23375821e+00 -8.83742869e-01 4.34293985e-01
3.68782371e-01 -4.07582939e-01 6.97248340e-01 2.17920728e-03
9.57112908e-02 1.41726986e-01 2.05196530e-01 -3.35316241e-01
1.01266932e+00 -1.11328995e+00 8.84499788e-01 9.60824251e-01
-1.56792164e-01 2.07510561e-01 1.00879562e+00 -1.22155416e+00
-1.41025496e+00 -1.07483208e+00 4.62076515e-01 -4.79951024e-01
6.55874193e-01 -6.50304377e-01 -8.84512067e-01 7.54956245e-01
-4.55377512e-02 -5.87972879e-01 2.98698604e-01 7.56009817e-02
-2.37004787e-01 -1.22714497e-01 -9.75279808e-01 7.63577282e-01
7.37198353e-01 -7.22099960e-01 -7.54038274e-01 3.69807333e-01
2.52663553e-01 -4.64142472e-01 -4.95410830e-01 7.75022507e-01
6.18227422e-01 -9.13113058e-01 8.90954912e-01 -2.58130461e-01
-1.41190559e-01 -8.32348764e-01 -2.42740624e-02 -1.05369735e+00
-4.80916083e-01 -4.35292661e-01 4.97354656e-01 4.87465978e-01
4.51305091e-01 -3.77028286e-01 9.37574327e-01 3.48766059e-01
-8.15276682e-01 -6.42122403e-02 -1.08492124e+00 -9.24641609e-01
-1.83742225e-01 -3.81322801e-01 8.22854862e-02 8.57537508e-01
-6.16331771e-02 1.11578247e-02 -3.04709464e-01 8.10084403e-01
8.22016418e-01 1.38285443e-01 8.05885673e-01 -1.46368706e+00
-2.49414250e-01 -1.38317928e-01 -5.07622421e-01 -4.56451029e-01
-1.33547023e-01 -7.27805197e-01 5.96482635e-01 -1.43761194e+00
-3.10362577e-01 -5.83352923e-01 -5.99688999e-02 6.96997702e-01
3.09852540e-01 2.66926438e-01 -2.33835861e-01 3.84336233e-01
-3.12270999e-01 5.78571975e-01 3.79946649e-01 -1.63949698e-01
-1.24185542e-02 -2.95966059e-01 -1.78062245e-01 5.78478932e-01
8.15487504e-01 -4.64286417e-01 -1.91029161e-01 -6.07325912e-01
2.07331464e-01 6.14320263e-02 6.81062281e-01 -1.75863528e+00
3.97400737e-01 6.18181825e-02 5.70404887e-01 -4.60423976e-01
6.25571012e-01 -8.17113459e-01 1.26876041e-01 4.29677039e-01
4.02221322e-01 1.92932025e-01 7.66396403e-01 6.95599496e-01
-1.86901256e-01 -1.04534574e-01 8.56970370e-01 1.19916545e-02
-1.55278063e+00 2.60519795e-02 -9.49315071e-01 -2.81683117e-01
1.18582976e+00 -4.86006021e-01 -3.76834512e-01 -5.73274076e-01
-5.41029394e-01 4.53457862e-01 9.77344811e-01 3.25441390e-01
7.36151397e-01 -7.67508268e-01 -3.10506076e-01 5.31121492e-01
4.08049673e-01 -1.97258666e-02 1.26094535e-01 6.80674195e-01
-9.97659564e-01 2.87070751e-01 -6.48696899e-01 -4.83784705e-01
-1.22340703e+00 4.06668693e-01 7.23772824e-01 2.04236314e-01
-8.02130461e-01 6.38051927e-01 3.85674536e-01 -2.81877220e-01
1.29526526e-01 2.27091417e-01 -3.19994567e-03 -2.56785870e-01
2.79533625e-01 -1.07169040e-01 3.72401834e-01 -5.01710176e-01
-4.54060227e-01 5.76170623e-01 1.45939723e-01 -3.63434225e-01
1.12089133e+00 1.03217028e-01 4.67968524e-01 1.09891914e-01
5.34456670e-01 -1.47396058e-01 -1.71917450e+00 2.93151230e-01
1.60264507e-01 -2.71262586e-01 4.21648353e-01 -9.25233305e-01
-3.26282442e-01 1.02219033e+00 9.30537760e-01 -2.55950600e-01
8.24527860e-01 -2.60770559e-01 5.98734677e-01 6.99983478e-01
7.57083416e-01 -1.03995264e+00 -1.89575776e-01 7.05570638e-01
8.79198790e-01 -1.29303873e+00 5.80479801e-02 -1.22491606e-02
-8.78326237e-01 1.05490482e+00 6.17482781e-01 -1.15769230e-01
1.31432071e-01 5.52819252e-01 2.94612110e-01 -4.25549328e-01
-4.60938424e-01 -4.81617600e-01 2.53364176e-01 8.77693176e-01
-1.95419312e-01 -7.41766319e-02 3.39778155e-01 3.22099239e-01
-2.00243637e-01 -3.62876207e-01 2.81905115e-01 1.35811102e+00
-1.14876127e+00 -6.76827610e-01 -6.83012128e-01 1.64990231e-01
5.49495593e-02 -1.60413355e-01 -5.75899601e-01 1.11784434e+00
-3.62061523e-02 5.81972539e-01 1.09925292e-01 -6.28833652e-01
4.35602129e-01 4.41474915e-02 3.36678624e-01 -4.75190580e-01
-5.27608216e-01 -7.70852268e-02 7.13351667e-01 -4.43213135e-01
3.06761622e-01 -5.94042122e-01 -1.25267923e+00 -6.93390593e-02
2.69090414e-01 1.94554880e-01 9.13249314e-01 8.56432021e-01
3.36373597e-01 4.18785334e-01 2.98703998e-01 -1.49116874e+00
-5.08487880e-01 -5.40750325e-01 -5.61511576e-01 2.97931340e-02
4.34460282e-01 -1.21190178e+00 -3.58608007e-01 -1.52733013e-01] | [7.3794941902160645, -1.9046601057052612] |
20cdfdbf-8a30-4511-a8e5-ea699e79e52e | copula-based-deep-survival-models-for | 2306.11912 | null | https://arxiv.org/abs/2306.11912v1 | https://arxiv.org/pdf/2306.11912v1.pdf | Copula-Based Deep Survival Models for Dependent Censoring | A survival dataset describes a set of instances (e.g. patients) and provides, for each, either the time until an event (e.g. death), or the censoring time (e.g. when lost to follow-up - which is a lower bound on the time until the event). We consider the challenge of survival prediction: learning, from such data, a predictive model that can produce an individual survival distribution for a novel instance. Many contemporary methods of survival prediction implicitly assume that the event and censoring distributions are independent conditional on the instance's covariates - a strong assumption that is difficult to verify (as we observe only one outcome for each instance) and which can induce significant bias when it does not hold. This paper presents a parametric model of survival that extends modern non-linear survival analysis by relaxing the assumption of conditional independence. On synthetic and semi-synthetic data, our approach significantly improves estimates of survival distributions compared to the standard that assumes conditional independence in the data. | ['Rahul G. Krishnan', 'Russell Greiner', 'Michael Cooper', 'Ali Hossein Gharari Foomani'] | 2023-06-20 | null | null | null | null | ['survival-analysis'] | ['miscellaneous'] | [ 1.32233933e-01 1.38770603e-02 -4.47772503e-01 -6.92133307e-01
-9.13680077e-01 -6.13551378e-01 4.00779605e-01 7.63059676e-01
-3.11667919e-01 1.18378949e+00 2.20127717e-01 -6.30759776e-01
-3.56925100e-01 -9.30093169e-01 -6.09514296e-01 -9.37873662e-01
-5.52521586e-01 9.75783765e-01 -1.49361253e-01 2.89752722e-01
7.14357048e-02 5.13788581e-01 -1.19352460e+00 -9.39880162e-02
5.27375400e-01 6.63516164e-01 -2.80306160e-01 9.28560436e-01
9.51254144e-02 6.98830664e-01 -5.71384430e-01 -1.39630809e-01
-2.74387598e-01 -5.22153974e-01 -7.26744413e-01 -3.24843705e-01
-1.24044724e-01 -3.94543767e-01 -9.14983600e-02 4.26447809e-01
8.78683701e-02 -2.14994177e-01 1.28157604e+00 -1.76812851e+00
-3.24184984e-01 4.87378091e-01 -2.12889582e-01 2.71049216e-02
1.82800770e-01 -1.77101657e-01 8.12225223e-01 -1.82777077e-01
3.25827032e-01 6.64197028e-01 9.83378768e-01 4.98258412e-01
-1.60501587e+00 -2.94966877e-01 -3.94399732e-01 -9.96789802e-03
-1.27921295e+00 -3.50968719e-01 2.78704047e-01 -7.06078231e-01
3.36292922e-01 3.87844622e-01 4.96266127e-01 1.21253026e+00
5.61936677e-01 3.04293454e-01 1.10666680e+00 -3.55277419e-01
5.97859859e-01 2.13723630e-01 4.26170349e-01 2.28226408e-01
3.92876983e-01 7.60067284e-01 -1.13685794e-01 -6.15899444e-01
5.76334596e-01 3.35474521e-01 -3.57436612e-02 -3.65063846e-01
-1.21820629e+00 8.55113089e-01 1.03483982e-02 4.01588716e-02
-4.11617339e-01 1.69900924e-01 3.98334712e-01 5.14345884e-01
1.73646390e-01 -2.46536762e-01 -1.02529359e+00 -7.73730502e-03
-1.22187638e+00 1.35031208e-01 1.16394651e+00 8.16049039e-01
5.80936015e-01 -3.97357911e-01 -2.44089440e-01 7.66953528e-02
-5.46434484e-02 2.05957323e-01 2.33963609e-01 -5.64983964e-01
-5.12132384e-02 3.02327484e-01 6.37947917e-01 -2.31384754e-01
-7.29120791e-01 -5.68570077e-01 -1.04123163e+00 3.68272871e-01
1.12029886e+00 -2.23648146e-01 -7.97253728e-01 2.05019641e+00
1.82624191e-01 2.67090321e-01 2.47103050e-01 3.44129711e-01
4.73386437e-01 4.85100955e-01 3.94479096e-01 -6.80694580e-01
1.09133112e+00 -2.80054867e-01 -4.63586450e-01 -4.25964082e-03
1.02800369e+00 -1.98253393e-01 7.62944639e-01 2.64756203e-01
-7.38401294e-01 5.31530455e-02 -6.44001901e-01 3.61185759e-01
-2.08702862e-01 -1.21131085e-01 6.66113257e-01 8.39798629e-01
-7.69374728e-01 8.47896516e-01 -1.00712752e+00 -5.19763410e-01
2.86825836e-01 3.85958701e-01 -6.22428298e-01 -4.00558747e-02
-9.77839947e-01 8.26009631e-01 3.22543263e-01 -4.27129984e-01
-1.08241558e+00 -8.34897876e-01 -4.92199868e-01 2.23644763e-01
5.60232662e-02 -1.06533384e+00 9.44840968e-01 -1.14423442e+00
-8.75963986e-01 9.84893203e-01 -2.25477785e-01 -5.44281065e-01
6.72611833e-01 2.71358222e-01 -3.99898946e-01 -3.03044736e-01
2.14116499e-02 9.27192196e-02 5.65882266e-01 -9.89139140e-01
-6.71663404e-01 -6.81227744e-01 -2.32894674e-01 -2.48528227e-01
-8.96922201e-02 1.64009765e-01 1.06354922e-01 -5.52604795e-01
2.55289301e-03 -8.11015189e-01 -3.92639369e-01 1.00665674e-01
-6.04829133e-01 -1.32546425e-01 2.72585124e-01 -7.40136206e-01
1.28040528e+00 -2.09009099e+00 1.29474416e-01 2.32434869e-02
4.12883535e-02 -5.22727847e-01 2.88076609e-01 7.76373148e-01
-4.88465160e-01 1.96115062e-01 -4.86463904e-01 -5.75519860e-01
-3.23880523e-01 3.87562275e-01 -4.50266600e-02 8.50643098e-01
-1.70513734e-01 5.37589669e-01 -6.57025576e-01 -5.66146791e-01
-1.29201775e-02 4.22862232e-01 -3.11337054e-01 2.70519763e-01
1.33511186e-01 8.60773146e-01 -2.18436807e-01 4.83351022e-01
3.42802286e-01 -4.01877195e-01 5.96630834e-02 4.71235961e-01
-4.10817303e-02 -6.82896376e-02 -9.75147128e-01 1.17982924e+00
-3.17698210e-01 2.75880456e-01 -4.45542932e-01 -1.09701240e+00
8.42195213e-01 6.12907171e-01 5.84123015e-01 7.57433325e-02
2.40237519e-01 2.60719448e-01 -1.30379483e-01 -1.85784280e-01
-1.79271072e-01 -8.79892647e-01 -3.79307061e-01 4.11486804e-01
5.58073893e-02 4.22136337e-01 -1.76665530e-01 -2.61625201e-02
1.37186074e+00 -1.85500130e-01 8.18424582e-01 -3.79681885e-01
3.17635924e-01 2.03998104e-01 8.68943095e-01 8.53185058e-01
-6.36602193e-02 6.18344128e-01 9.32033777e-01 -4.24922109e-01
-1.17430270e+00 -1.35466933e+00 -5.96068680e-01 6.64523602e-01
-3.11315298e-01 -7.93394297e-02 -2.57519573e-01 -8.06653738e-01
2.15160280e-01 1.01069260e+00 -1.11633706e+00 -3.23888570e-01
-1.01593405e-01 -1.07247460e+00 2.95593709e-01 5.89079082e-01
-5.21170914e-01 -7.51748145e-01 -3.45717043e-01 3.15007806e-01
-3.28777693e-02 -1.84931576e-01 -2.07444578e-01 6.07117891e-01
-1.33964372e+00 -1.13841796e+00 -6.44721210e-01 -6.03427052e-01
7.38014042e-01 -5.01767218e-01 1.39880967e+00 2.37784058e-01
-7.88608119e-02 1.26245826e-01 -1.47857025e-01 -2.14882940e-01
-5.52434802e-01 -2.14087814e-01 6.35264395e-03 -9.10496786e-02
2.04785958e-01 -6.62791789e-01 -7.73360550e-01 2.76485384e-01
-8.22625220e-01 -1.77676842e-01 3.01517844e-01 1.04473484e+00
6.87658846e-01 3.54444176e-01 9.20518517e-01 -1.02352893e+00
-5.67370802e-02 -1.04613674e+00 -5.11126101e-01 3.59466195e-01
-7.13126123e-01 2.14511193e-02 7.60383427e-01 -4.61321563e-01
-6.32447183e-01 3.12403381e-01 3.48339021e-01 5.88351116e-02
-5.97247899e-01 6.13432586e-01 -2.12250605e-01 3.26424301e-01
4.35223043e-01 1.73220977e-01 -2.09995592e-03 -6.49201035e-01
-2.41037663e-02 4.93201941e-01 5.97460032e-01 -3.61431807e-01
4.87935573e-01 4.83235270e-01 6.18024647e-01 -3.25684369e-01
-7.08738506e-01 -3.98011923e-01 -9.10153508e-01 1.76746417e-02
5.77505231e-01 -7.90537417e-01 -9.46508944e-01 3.39824259e-01
-6.41949892e-01 -3.66792917e-01 -4.09109563e-01 7.67718196e-01
-9.11183417e-01 -8.37332010e-02 -3.18262100e-01 -9.60269213e-01
6.88031092e-02 -5.39001822e-01 6.29339814e-01 9.60501879e-02
-3.87686431e-01 -1.66093493e+00 2.41461590e-01 -6.84628189e-02
2.50880063e-01 8.43929589e-01 1.31428015e+00 -9.51649606e-01
-1.20950557e-01 -8.00162137e-01 1.53035939e-01 8.63699913e-02
8.84816274e-02 5.51538318e-02 -4.93839830e-01 -5.68023026e-01
-1.20672137e-01 -2.35768501e-02 7.40032494e-01 8.54254901e-01
1.13428175e+00 -3.46107900e-01 -7.66734302e-01 4.97806311e-01
1.50362754e+00 1.36042401e-01 5.41046321e-01 3.19895923e-01
1.07421577e-01 8.70129108e-01 4.98635441e-01 6.81426346e-01
4.53446478e-01 6.87261641e-01 4.88757461e-01 -1.56930953e-01
3.52258980e-01 -1.80731773e-01 1.18786320e-01 -3.93079184e-02
1.51550561e-01 -3.99074346e-01 -9.59372342e-01 8.08830142e-01
-1.66533303e+00 -7.66739428e-01 -7.12053120e-01 3.15308905e+00
8.89662445e-01 8.90592784e-02 5.06114721e-01 4.01841551e-01
6.83007419e-01 -4.97787923e-01 -5.40453315e-01 -1.44270808e-01
-1.24022871e-01 -8.76235031e-03 7.32910573e-01 6.91973150e-01
-9.08715248e-01 1.65193737e-01 6.62778282e+00 2.85291672e-01
-7.75055766e-01 3.97176668e-02 8.16187441e-01 1.68667603e-02
-1.54756710e-01 5.44938564e-01 -6.11200154e-01 6.92149937e-01
1.38575065e+00 -5.94758034e-01 7.25505352e-02 6.14275455e-01
1.72786966e-01 -1.65253311e-01 -1.66351295e+00 4.06729341e-01
-2.47020245e-01 -6.50886953e-01 -4.05434161e-01 2.54708737e-01
4.41483438e-01 -6.18842840e-01 -4.09604330e-03 1.94761828e-01
6.98015094e-01 -1.25279605e+00 5.99984467e-01 9.80872154e-01
1.17229569e+00 -9.18097019e-01 9.60853100e-01 7.80660808e-01
-5.79572439e-01 -1.65313646e-01 3.64832208e-02 -1.15064256e-01
4.40562099e-01 9.27744806e-01 -7.62083352e-01 5.14701545e-01
4.38412994e-01 5.23138940e-01 -5.15629590e-01 1.33178413e+00
-4.95208725e-02 1.13297963e+00 -3.49121034e-01 4.81486499e-01
-2.99086452e-01 3.86339203e-02 4.87460464e-01 7.88311183e-01
7.56317437e-01 8.78788158e-02 -4.00056839e-02 4.80880439e-01
3.44536453e-01 -3.92501540e-02 -3.98255795e-01 2.93336093e-01
3.60645384e-01 9.24111247e-01 -6.06351256e-01 -2.92874485e-01
-5.22783995e-01 8.03048730e-01 1.51375264e-01 2.74835229e-01
-6.01296544e-01 -7.19834790e-02 3.34114224e-01 5.05100429e-01
6.02494553e-02 -1.91742729e-03 -6.37362599e-01 -9.10775483e-01
-3.84357065e-01 -2.17274487e-01 1.01096702e+00 -5.35381436e-01
-1.59473264e+00 2.85774797e-01 6.14805939e-03 -1.17031860e+00
-5.08952796e-01 -3.13134760e-01 -8.11649144e-01 1.15584791e+00
-1.14676952e+00 -1.05013394e+00 -4.32631932e-02 6.67225897e-01
-2.89843529e-01 1.72647879e-01 1.16475224e+00 1.05112009e-01
-5.41283965e-01 6.27932787e-01 4.40973729e-01 -1.79312706e-01
8.72524202e-01 -1.51098633e+00 -6.80939555e-02 1.94291055e-01
-2.81516135e-01 3.82249713e-01 1.09783876e+00 -7.48533905e-01
-7.97186494e-01 -1.15259624e+00 1.39330018e+00 -9.51740563e-01
5.64450502e-01 -9.94277149e-02 -1.06884336e+00 9.82531011e-01
-6.09400809e-01 1.28394753e-01 1.08279645e+00 5.25787592e-01
-4.80452180e-02 -2.94824373e-02 -1.43126535e+00 1.89394087e-01
5.49246550e-01 -8.59174654e-02 -4.68309850e-01 3.03229928e-01
3.11781764e-01 3.81530821e-02 -9.61624384e-01 3.33999574e-01
7.20369816e-01 -1.15970802e+00 9.42637682e-01 -1.17809570e+00
4.77783263e-01 -2.49654263e-01 -1.26040816e-01 -1.17232728e+00
-3.88873279e-01 -2.73565352e-01 -1.44175902e-01 1.32334793e+00
5.01928926e-01 -6.08640611e-01 8.55186939e-01 8.67621124e-01
2.71313787e-01 -6.89017713e-01 -1.35270226e+00 -1.06255913e+00
5.14062464e-01 -2.31824473e-01 7.86551178e-01 9.85380471e-01
-1.42634258e-01 -2.86809206e-02 -6.05682552e-01 3.87450248e-01
9.59995866e-01 1.08567230e-01 5.63412845e-01 -1.75360966e+00
-2.88055092e-01 3.86447012e-02 -5.92812538e-01 -3.58236253e-01
1.38466060e-01 -6.22325540e-01 -2.47077689e-01 -1.33108532e+00
6.35361075e-01 -7.68764317e-01 -5.51325142e-01 5.07653177e-01
-2.29795992e-01 -8.67911354e-02 -5.34692109e-01 2.70770907e-01
-1.31167606e-01 2.18205169e-01 6.62720561e-01 4.66798961e-01
-1.21822484e-01 6.17385864e-01 -5.46044171e-01 5.04742205e-01
8.74548852e-01 -1.00424600e+00 -6.61435276e-02 2.72914678e-01
5.30773439e-02 1.30520940e+00 7.62418509e-01 -7.83770144e-01
1.87341213e-01 -4.19512779e-01 4.98127550e-01 -5.68964005e-01
1.79725830e-02 -1.16394019e+00 9.22081649e-01 7.28294849e-01
-3.96686018e-01 -1.15926996e-01 -2.11671591e-01 9.39085603e-01
-1.47767942e-02 -2.41296604e-01 6.53224587e-01 2.07824782e-01
6.00434490e-04 3.69777620e-01 -1.93169683e-01 -2.09006786e-01
1.28536844e+00 -1.71349555e-01 -8.54322091e-02 -4.98185456e-01
-1.32169616e+00 2.39384606e-01 9.17321324e-01 -1.70326792e-02
2.88528532e-01 -1.29377425e+00 -1.01439965e+00 1.80630758e-01
2.62669146e-01 -3.05281669e-01 2.47941762e-01 1.09988487e+00
4.43528853e-02 2.51642466e-01 -1.10537298e-01 -2.69692570e-01
-1.40853775e+00 8.63780856e-01 3.58756363e-01 -3.90306205e-01
-5.82424402e-01 5.18767655e-01 3.52343887e-01 -1.83558866e-01
6.51745424e-02 7.90791884e-02 -2.65466839e-01 -1.72261056e-02
3.67775828e-01 4.50396359e-01 -1.12998649e-01 -6.63677514e-01
-3.45652282e-01 9.45935100e-02 -9.61367600e-03 -2.06304535e-01
1.27433074e+00 -1.83067650e-01 -1.30911708e-01 1.07035065e+00
1.03827274e+00 -3.33047152e-01 -1.28617060e+00 -6.65076524e-02
1.30115911e-01 -5.95623851e-01 -4.50311601e-02 -1.13742590e+00
-7.43697524e-01 7.21551120e-01 4.90888804e-01 1.92827553e-01
1.11900091e+00 2.14991838e-01 2.75975525e-01 -2.45148301e-01
4.99300450e-01 -3.55156094e-01 -5.40530622e-01 4.70785201e-02
6.88709140e-01 -1.28223789e+00 -1.17295094e-01 -1.29863963e-01
-2.41497204e-01 9.91223454e-01 -5.68327270e-02 9.54165235e-02
8.41096222e-01 2.48232543e-01 -1.53855488e-01 6.41092509e-02
-9.93005395e-01 3.00341457e-01 9.88176093e-02 6.62632942e-01
5.29421031e-01 5.60033858e-01 -5.29491365e-01 1.03872085e+00
-2.19503716e-01 2.94758171e-01 6.18996322e-01 8.28488410e-01
-1.42302692e-01 -1.18786514e+00 -5.81231117e-01 8.85425746e-01
-7.41942465e-01 -1.37071381e-03 -1.07032128e-01 6.98427379e-01
-6.82452470e-02 9.05309141e-01 1.65441379e-01 6.18201606e-02
2.11727306e-01 1.35280415e-01 2.77702063e-01 -5.73381245e-01
-2.03467995e-01 -4.00072306e-01 3.40431035e-02 -8.43284801e-02
-7.24899918e-02 -1.09040225e+00 -1.19049656e+00 -7.41905570e-01
-4.13198113e-01 3.63730967e-01 5.48549294e-01 7.46938586e-01
-1.20130576e-01 3.96654606e-01 7.81761229e-01 -4.27248240e-01
-6.26772702e-01 -6.74895048e-01 -1.14166093e+00 3.92399490e-01
7.27337360e-01 -6.48049712e-01 -1.02691162e+00 4.57026929e-01] | [7.7765889167785645, 5.546986103057861] |
2a10ee6e-b6c2-402b-9121-21d9c230257f | nerf-pose-a-first-reconstruct-then-regress | 2203.04802 | null | https://arxiv.org/abs/2203.04802v1 | https://arxiv.org/pdf/2203.04802v1.pdf | NeRF-Pose: A First-Reconstruct-Then-Regress Approach for Weakly-supervised 6D Object Pose Estimation | Pose estimation of 3D objects in monocular images is a fundamental and long-standing problem in computer vision. Existing deep learning approaches for 6D pose estimation typically rely on the assumption of availability of 3D object models and 6D pose annotations. However, precise annotation of 6D poses in real data is intricate, time-consuming and not scalable, while synthetic data scales well but lacks realism. To avoid these problems, we present a weakly-supervised reconstruction-based pipeline, named NeRF-Pose, which needs only 2D object segmentation and known relative camera poses during training. Following the first-reconstruct-then-regress idea, we first reconstruct the objects from multiple views in the form of an implicit neural representation. Then, we train a pose regression network to predict pixel-wise 2D-3D correspondences between images and the reconstructed model. At inference, the approach only needs a single image as input. A NeRF-enabled PnP+RANSAC algorithm is used to estimate stable and accurate pose from the predicted correspondences. Experiments on LineMod and LineMod-Occlusion show that the proposed method has state-of-the-art accuracy in comparison to the best 6D pose estimation methods in spite of being trained only with weak labels. Besides, we extend the Homebrewed DB dataset with more real training images to support the weakly supervised task and achieve compelling results on this dataset. The extended dataset and code will be released soon. | ['Slobodan Ilic', 'Shaowu Yang', 'Benjamin Busam', 'Ivan Shugurov', 'Hao Yu', 'Fu Li'] | 2022-03-09 | null | null | null | null | ['6d-pose-estimation-1', '6d-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-1.95636265e-02 6.48757592e-02 -2.78485864e-01 -6.44763947e-01
-9.53756154e-01 -7.16637552e-01 4.72293377e-01 -4.13401514e-01
-5.01890779e-01 2.76722103e-01 -2.34822094e-01 5.92341758e-02
1.40420496e-01 -3.89996260e-01 -1.18602312e+00 -4.19083983e-01
2.92262226e-01 1.18779981e+00 5.01841068e-01 1.29891098e-01
1.36574671e-01 8.02250147e-01 -1.37841344e+00 7.21830409e-03
4.46147263e-01 1.11997032e+00 4.11319107e-01 5.67431867e-01
8.54160488e-02 3.86409968e-01 -2.35459521e-01 -4.68509912e-01
5.94196737e-01 -1.09919369e-01 -7.00794637e-01 6.11274362e-01
9.61704493e-01 -8.03728878e-01 -3.02687407e-01 7.72917449e-01
4.59585100e-01 -1.33280709e-01 5.96479237e-01 -1.07295287e+00
-6.92530796e-02 3.11181918e-02 -6.94873929e-01 -3.27975810e-01
6.02959514e-01 1.11598328e-01 8.06514800e-01 -1.16848266e+00
8.88723135e-01 1.17843771e+00 7.80808389e-01 4.86183494e-01
-1.32072866e+00 -5.58945954e-01 1.87970877e-01 -8.14873353e-02
-1.39983189e+00 -3.37946177e-01 8.47399294e-01 -4.57855135e-01
9.94954467e-01 8.16139132e-02 7.77811229e-01 1.18152630e+00
-2.11712360e-01 1.02076137e+00 1.04685557e+00 -1.94878504e-01
-4.48440760e-03 -8.02104771e-02 -2.34808967e-01 7.71622181e-01
4.88758944e-02 8.95059332e-02 -4.84993517e-01 1.68646574e-01
9.78907645e-01 6.06214888e-02 -1.12670198e-01 -1.14792395e+00
-1.30940270e+00 4.79037136e-01 5.84160686e-01 -1.86039984e-01
-3.25087994e-01 3.04147810e-01 8.03355724e-02 -1.08827740e-01
5.69944024e-01 2.48989642e-01 -7.20603406e-01 1.86022017e-02
-1.00697219e+00 4.50743943e-01 5.82432687e-01 1.09130073e+00
7.83670008e-01 -2.34579772e-01 2.53115654e-01 6.65776670e-01
4.17520374e-01 7.00001001e-01 4.50463034e-02 -1.22853458e+00
4.92906541e-01 5.68788171e-01 4.24608052e-01 -8.44596267e-01
-6.59062088e-01 -6.30992472e-01 -3.11743110e-01 2.38714904e-01
5.52374303e-01 3.04128230e-01 -1.04106629e+00 1.38689244e+00
6.92601264e-01 2.04195797e-01 -3.17343026e-01 1.32862902e+00
7.82794237e-01 3.93687040e-01 -4.00659829e-01 -6.46390207e-03
9.95960951e-01 -1.01845777e+00 -3.11066508e-01 -4.98115420e-01
2.78539509e-01 -8.72645497e-01 7.83024490e-01 3.82951677e-01
-1.12645984e+00 -6.02529287e-01 -9.15235519e-01 -3.39915246e-01
-1.75549109e-02 3.58308822e-01 4.52769399e-01 3.43119800e-01
-6.53290153e-01 3.62114191e-01 -1.01866972e+00 -3.43467385e-01
5.62560856e-01 4.22290266e-01 -7.04783261e-01 -2.04479158e-01
-5.03841519e-01 1.09506679e+00 4.26144689e-01 3.07997435e-01
-1.08313513e+00 -5.55184305e-01 -1.10587394e+00 -4.07751858e-01
6.71981573e-01 -8.42299104e-01 1.32245755e+00 -5.72062433e-01
-1.47054398e+00 1.35719609e+00 1.52407913e-02 -4.24351364e-01
9.71197367e-01 -6.83150470e-01 2.80010581e-01 1.61728144e-01
4.04433832e-02 1.02630138e+00 8.11548173e-01 -1.60876191e+00
-4.18682098e-01 -7.00598419e-01 1.35405511e-01 4.62754399e-01
2.35457674e-01 -2.83604294e-01 -1.16112053e+00 -2.65834570e-01
7.28168607e-01 -1.20681858e+00 -3.06192875e-01 4.63853538e-01
-4.66065854e-01 1.41575998e-02 8.12400818e-01 -6.45185113e-01
3.04403931e-01 -1.92139375e+00 4.47310299e-01 -9.41613168e-02
2.31825863e-04 1.52504876e-01 -1.69249400e-01 6.77921474e-02
1.03235371e-01 -4.20195401e-01 -2.06654742e-01 -8.27319801e-01
1.85350270e-03 2.81228364e-01 -1.74762890e-01 8.62137437e-01
1.61846861e-01 1.01308501e+00 -6.89018905e-01 -4.15282428e-01
7.06465065e-01 5.59097111e-01 -6.41774535e-01 4.01641935e-01
-5.97559631e-01 7.31414378e-01 -2.38093972e-01 6.83064938e-01
9.18543458e-01 -2.10781723e-01 3.17026787e-02 -6.04210436e-01
-4.18208204e-02 1.67964503e-01 -1.35426807e+00 2.31171465e+00
-4.42948014e-01 3.46211970e-01 2.90882457e-02 -8.49853873e-01
8.54746938e-01 9.68794078e-02 5.73230326e-01 -4.16803241e-01
2.76103735e-01 4.02808696e-01 -3.67918938e-01 -5.02223492e-01
2.30966061e-01 -4.02262956e-02 -6.16074130e-02 2.07042500e-01
3.06692451e-01 -8.09516668e-01 -6.33758977e-02 -1.60227835e-01
5.82997024e-01 1.07439983e+00 1.65488601e-01 2.84545422e-01
4.04666305e-01 1.32122263e-01 4.37173069e-01 2.78528601e-01
-1.08565830e-01 1.09830081e+00 2.03454033e-01 -5.50597847e-01
-1.16536427e+00 -1.08302391e+00 -2.51205593e-01 4.97632951e-01
5.48284411e-01 -1.87242240e-01 -6.18098974e-01 -9.01095569e-01
1.45811796e-01 5.88524699e-01 -3.12360704e-01 3.10562223e-01
-6.40742302e-01 -4.72054183e-01 1.84291273e-01 6.72519505e-01
3.81007522e-01 -6.99155033e-01 -7.28498876e-01 7.05544464e-03
-2.40483716e-01 -1.63554204e+00 -2.68154025e-01 2.35179737e-01
-9.74518538e-01 -1.14699793e+00 -7.03412235e-01 -6.57865703e-01
7.27204323e-01 2.59009570e-01 1.07838857e+00 -2.59182900e-01
-1.69689640e-01 2.59163558e-01 -1.32158892e-02 -1.97600812e-01
1.57879796e-02 -1.79670509e-02 3.17284651e-02 -1.27786294e-01
1.67573005e-01 -5.73122323e-01 -6.72651768e-01 6.31127357e-01
-4.42712247e-01 2.46793360e-01 7.76533008e-01 6.65566385e-01
1.02375484e+00 -3.27448964e-01 -1.11133158e-02 -7.40607858e-01
-3.57371062e-01 6.84769079e-02 -9.46974933e-01 -9.89612564e-02
-1.56085804e-01 -5.87359928e-02 2.12096825e-01 -3.38039398e-01
-1.02067482e+00 9.71630096e-01 -2.96486974e-01 -9.24486995e-01
-4.60479885e-01 1.00823969e-01 -3.06930572e-01 -5.87076843e-02
6.38335943e-01 1.49871528e-01 -4.45228480e-02 -6.87684476e-01
4.46172565e-01 4.03239906e-01 8.14098239e-01 -4.05241787e-01
1.10636866e+00 6.51599944e-01 1.67110693e-02 -5.27936161e-01
-1.27519965e+00 -4.93106425e-01 -1.25533783e+00 -2.96593130e-01
1.01133132e+00 -1.24841475e+00 -6.97746396e-01 5.42847931e-01
-1.31894708e+00 -3.49346429e-01 -1.99252646e-02 7.12851882e-01
-9.28924382e-01 2.50509828e-01 -5.45273185e-01 -6.95905745e-01
-1.39481574e-02 -1.26311040e+00 1.74300838e+00 -1.39213413e-01
-1.69302195e-01 -5.73930442e-01 -1.58371150e-01 1.05004072e+00
-2.10198358e-01 4.67835724e-01 4.02021587e-01 -3.37412536e-01
-1.04685199e+00 -5.02138734e-01 -1.91374063e-01 3.26251090e-01
-2.12318942e-01 -2.66541421e-01 -1.08791828e+00 -1.69760332e-01
3.57774422e-02 -6.26217544e-01 5.89133322e-01 3.65215570e-01
1.22600579e+00 9.46359336e-02 -2.44193643e-01 8.85206223e-01
1.42067564e+00 -2.37475947e-01 3.61771524e-01 2.01646432e-01
1.01974857e+00 7.09580541e-01 9.06573296e-01 1.95726454e-01
4.85997647e-01 1.01930726e+00 9.10419583e-01 1.73215650e-03
-1.48157433e-01 -5.54291368e-01 1.42993733e-01 6.05563223e-01
-7.60677308e-02 1.47524495e-02 -9.07692552e-01 2.22170994e-01
-1.69301605e+00 -5.80080807e-01 -2.91318923e-01 2.17088461e+00
6.43154025e-01 4.30913925e-01 9.92294252e-02 3.81742306e-02
3.69620681e-01 6.23449460e-02 -7.11019754e-01 2.27325022e-01
1.12081558e-01 -1.44046634e-01 5.77603817e-01 6.40199363e-01
-1.22824085e+00 1.24663305e+00 5.68915749e+00 5.63597918e-01
-1.14478350e+00 7.59289861e-02 3.89581025e-01 -1.87797874e-01
-1.65339168e-02 6.55472502e-02 -9.25697744e-01 7.72645846e-02
3.69266570e-01 6.18822455e-01 2.54578620e-01 1.05528009e+00
4.34840843e-02 -2.74873972e-01 -1.42724037e+00 1.35079753e+00
2.80741513e-01 -1.10863030e+00 -1.79949895e-01 1.31020416e-02
6.77513242e-01 2.28569895e-01 -9.33120921e-02 3.65682766e-02
8.60453118e-03 -9.28124309e-01 1.08582878e+00 3.53382558e-01
7.53719330e-01 -7.01033175e-01 6.76503539e-01 8.27700615e-01
-9.56535041e-01 1.67614758e-01 -3.07300717e-01 -1.31808808e-02
3.88160855e-01 5.15404284e-01 -9.41549957e-01 6.27465129e-01
8.41687918e-01 8.86381924e-01 -6.07281983e-01 9.55709279e-01
-4.73313779e-01 4.50142920e-02 -4.46317703e-01 3.73319566e-01
5.41254990e-02 -1.94032013e-01 4.81340319e-01 7.64265656e-01
1.31265163e-01 -8.53751749e-02 3.48224580e-01 8.23810339e-01
6.90568462e-02 -2.70629972e-01 -4.02015299e-01 2.50238985e-01
2.28962511e-01 1.32198691e+00 -9.14211631e-01 -2.91050375e-02
-2.77545840e-01 1.27331126e+00 3.90407830e-01 8.96186978e-02
-9.19246733e-01 3.00968379e-01 3.55861425e-01 2.00561598e-01
4.70821857e-01 -6.27947271e-01 -2.64475524e-01 -1.18706572e+00
2.71402776e-01 -6.45524085e-01 2.55462304e-02 -1.14589834e+00
-1.18235290e+00 4.94361550e-01 2.19480515e-01 -1.34026897e+00
-3.42159480e-01 -7.96138346e-01 4.24649343e-02 6.43396139e-01
-1.37059867e+00 -1.52301669e+00 -5.62933862e-01 3.15004349e-01
7.52058804e-01 1.98238283e-01 6.58089280e-01 1.77004322e-01
-2.97903091e-01 3.04735303e-01 -4.20195609e-01 7.25945979e-02
7.17790842e-01 -1.13625300e+00 3.61633360e-01 5.19466758e-01
4.98849869e-01 2.16807276e-01 6.63144290e-01 -5.26484966e-01
-1.62624598e+00 -1.01224661e+00 7.55955398e-01 -1.01680887e+00
2.48700470e-01 -7.20845222e-01 -5.43139279e-01 8.54282916e-01
-2.04571441e-01 4.01385307e-01 1.81683898e-01 -1.41567007e-01
-2.89827198e-01 -1.43645376e-01 -1.00233018e+00 4.18040186e-01
1.38937891e+00 -4.88218546e-01 -5.79874218e-01 4.41544980e-01
7.48220742e-01 -1.15748847e+00 -7.20823944e-01 6.10281885e-01
5.59954703e-01 -1.01502967e+00 1.25272799e+00 -2.18022883e-01
3.04565072e-01 -6.31800830e-01 -3.47059280e-01 -8.65741491e-01
1.78003207e-01 -2.38030076e-01 -1.06516294e-01 9.38403010e-01
2.15544075e-01 -1.23413153e-01 1.17446637e+00 4.00521874e-01
-1.29205093e-01 -8.64286423e-01 -8.73882592e-01 -7.16350555e-01
-2.56500751e-01 -7.19855547e-01 2.73117959e-01 6.87545121e-01
-8.29768538e-01 5.10885894e-01 -4.55028027e-01 4.26502436e-01
8.38630140e-01 4.58357871e-01 1.35272908e+00 -1.16560876e+00
-3.52315545e-01 -2.24764794e-02 -5.67916691e-01 -1.56346488e+00
3.81399482e-01 -8.12840462e-01 1.59418076e-01 -1.49464500e+00
1.92308053e-01 -2.95193970e-01 3.90786648e-01 3.08796793e-01
1.37638286e-01 7.25893974e-01 1.75637454e-01 3.01678747e-01
-6.87749267e-01 6.21178389e-01 1.46140635e+00 4.55556959e-02
-4.48426455e-02 2.50304401e-01 -4.19141278e-02 9.98187304e-01
3.11383873e-01 -4.37252015e-01 -3.92200142e-01 -4.94609416e-01
2.24216715e-01 1.82975426e-01 7.02784300e-01 -8.98329616e-01
9.88490656e-02 -2.86562536e-02 8.22397828e-01 -1.33798969e+00
9.42508340e-01 -1.31249261e+00 3.10577661e-01 3.89500320e-01
-1.21260852e-01 -1.25533625e-01 -1.57322939e-02 5.70802271e-01
-2.71802768e-02 -9.19988304e-02 6.87790871e-01 -2.72991806e-01
-7.79591143e-01 7.02164292e-01 2.84079432e-01 -1.10261545e-01
1.06484389e+00 -3.41890186e-01 1.88561246e-01 -2.66456217e-01
-7.10187972e-01 2.20065176e-01 7.90396988e-01 5.23058414e-01
7.02570677e-01 -1.21199524e+00 -6.37304425e-01 1.67055741e-01
2.42404878e-01 8.52550387e-01 1.11520968e-01 8.44292462e-01
-8.50908220e-01 3.12495768e-01 -5.41126840e-02 -1.35203350e+00
-1.10491431e+00 5.97640693e-01 5.57356596e-01 8.71855989e-02
-6.53999507e-01 7.87346065e-01 2.24306434e-01 -9.78517473e-01
3.83528382e-01 -2.45746389e-01 1.17501840e-01 -3.46854813e-02
1.30898967e-01 1.30733505e-01 1.54922843e-01 -8.68054390e-01
-4.83056605e-01 1.06919241e+00 1.02256134e-01 -2.18476102e-01
1.55015230e+00 -1.34098351e-01 4.55449671e-02 4.25400287e-01
1.35195947e+00 -1.86848998e-01 -1.85667574e+00 -2.27713734e-01
-1.28954083e-01 -7.79259324e-01 -1.44968107e-01 -6.47871614e-01
-1.08074868e+00 9.76060450e-01 5.85285366e-01 -6.14499450e-01
8.12471449e-01 3.67230684e-01 6.39459789e-01 3.35969537e-01
5.82597017e-01 -9.18758035e-01 3.91522437e-01 4.00997669e-01
1.10184431e+00 -1.43631852e+00 1.62657574e-01 -7.79776514e-01
-5.12260973e-01 1.08106434e+00 9.47537601e-01 -2.19174743e-01
3.71895075e-01 2.00280458e-01 1.71974003e-01 -1.25830457e-01
-3.38909268e-01 -7.81034231e-02 6.00597084e-01 5.59191287e-01
2.26814196e-01 -1.86410457e-01 5.44960834e-02 1.65929317e-01
-3.31060469e-01 -2.51954645e-01 8.27872530e-02 7.38165021e-01
-1.31629273e-01 -1.05831206e+00 -4.71266359e-01 -3.48886102e-02
-9.91844013e-02 3.93588722e-01 -5.12811959e-01 8.57655048e-01
2.81272441e-01 4.89324778e-01 1.02402285e-01 -3.18619758e-01
5.16290307e-01 -2.97467317e-02 9.63834047e-01 -6.60548151e-01
-2.07548454e-01 4.57773060e-01 1.40077798e-02 -9.13664341e-01
-6.85357273e-01 -7.18565404e-01 -1.24982798e+00 1.73810273e-02
-4.14310068e-01 -4.54521298e-01 8.89799893e-01 1.03754365e+00
6.22892864e-02 1.57355696e-01 4.08062935e-01 -1.55588222e+00
-5.65651774e-01 -7.29403913e-01 -2.78550297e-01 4.53978598e-01
2.59622216e-01 -7.71720350e-01 -1.57948717e-01 5.65596707e-02] | [7.5574846267700195, -2.5855729579925537] |
4dd68713-a73f-499c-a3e8-c03902328c8f | hybrid-deep-neural-networks-for-all-cause | 1810.08503 | null | http://arxiv.org/abs/1810.08503v1 | http://arxiv.org/pdf/1810.08503v1.pdf | Hybrid deep neural networks for all-cause Mortality Prediction from LDCT Images | Known for its high morbidity and mortality rates, lung cancer poses a
significant threat to human health and well-being. However, the same population
is also at high risk for other deadly diseases, such as cardiovascular disease.
Since Low-Dose CT (LDCT) has been shown to significantly improve the lung
cancer diagnosis accuracy, it will be very useful for clinical practice to
predict the all-cause mortality for lung cancer patients to take corresponding
actions. In this paper, we propose a deep learning based method, which takes
both chest LDCT image patches and coronary artery calcification risk scores as
input, for direct prediction of mortality risk of lung cancer subjects. The
proposed method is called Hybrid Risk Network (HyRiskNet) for mortality risk
prediction, which is an end-to-end framework utilizing hybrid imaging features,
instead of completely relying on automatic feature extraction. Our work
demonstrates the feasibility of using deep learning techniques for all-cause
lung cancer mortality prediction from chest LDCT images. The experimental
results show that the proposed HyRiskNet can achieve superior performance
compared with the neural networks with only image input and with other
traditional semi-automatic scoring methods. The study also indicates that
radiologist defined features can well complement convolutional neural networks
for more comprehensive feature extraction. | ['Mannudeep K. Kalra', 'Hengtao Guo', 'Ge Wang', 'Ruben De Man', 'Pingkun Yan'] | 2018-10-19 | null | null | null | null | ['lung-cancer-diagnosis'] | ['medical'] | [-7.24865049e-02 5.39998896e-03 -4.52160180e-01 -2.70477593e-01
-1.17336643e+00 1.49591174e-02 3.40183824e-01 2.66255647e-01
-4.81234670e-01 6.86861694e-01 3.31716210e-01 -6.33859038e-01
-1.81433350e-01 -1.14434063e+00 -3.25618847e-03 -7.97165692e-01
-6.77108616e-02 7.57920325e-01 3.63311410e-01 4.16510910e-01
-2.74382144e-01 6.69827819e-01 -1.01142240e+00 2.52388328e-01
5.51636100e-01 1.10524547e+00 2.21026167e-01 8.10565293e-01
-7.63006285e-02 1.20363593e+00 -1.02537543e-01 -1.63359568e-01
2.38014013e-02 -5.39974749e-01 -7.05268025e-01 -5.60089350e-01
3.58613610e-01 -5.94221354e-01 -4.17650938e-01 5.88835835e-01
7.64350355e-01 -3.31728101e-01 9.41470385e-01 -5.30234516e-01
-1.30448118e-01 2.12781563e-01 -3.08265388e-01 4.43045974e-01
-1.35643929e-01 2.92277932e-02 7.11944163e-01 -8.32385302e-01
-1.05678325e-03 8.24825764e-01 1.00843132e+00 7.33048022e-01
-5.50556064e-01 -7.42307842e-01 -5.54770529e-01 1.53501630e-01
-1.07295048e+00 1.88871488e-01 3.07315350e-01 -4.73330975e-01
5.48056364e-01 5.34513235e-01 6.32201612e-01 8.79766822e-01
4.93730307e-01 3.87764305e-01 8.50013912e-01 -2.39083916e-01
-1.31977359e-02 -9.91598740e-02 1.11988135e-01 1.28252614e+00
5.51356733e-01 4.86990541e-01 1.72049195e-01 -5.00586689e-01
9.46901023e-01 5.49562335e-01 -2.82333940e-01 -1.64773345e-01
-1.26221168e+00 8.99002135e-01 8.19517195e-01 5.54592073e-01
-4.33302194e-01 5.30792415e-01 5.16939461e-01 -2.28762254e-01
3.11148614e-01 -8.27873349e-02 -4.50657636e-01 2.70108342e-01
-1.01352549e+00 -5.78002222e-02 6.10989749e-01 -1.13302348e-02
1.01403892e-01 -2.44369939e-01 -6.35198653e-01 8.06869566e-01
5.22511661e-01 4.44973111e-01 1.04826045e+00 -6.58932269e-01
2.40700707e-01 6.92107320e-01 -3.06471914e-01 -5.97052693e-01
-7.03446627e-01 -6.35967672e-01 -1.44370389e+00 2.63651162e-01
3.26897413e-01 -6.89446554e-03 -1.07218099e+00 1.48919129e+00
1.99695572e-01 2.38502041e-01 5.58045181e-03 7.25201368e-01
1.18224537e+00 4.54674691e-01 6.59941316e-01 -3.01068544e-01
1.55381131e+00 -7.84369528e-01 -4.14364457e-01 1.90349683e-01
9.45482731e-01 -2.38510311e-01 9.51670468e-01 4.70857471e-02
-8.32136810e-01 -3.11059684e-01 -6.40517175e-01 -1.14330882e-02
-3.70526947e-02 6.16381288e-01 5.57842016e-01 8.51146877e-01
-7.75157273e-01 6.92679405e-01 -1.22346699e+00 -5.69240808e-01
7.25064337e-01 5.60077846e-01 -3.11744481e-01 -5.46256229e-02
-1.23536849e+00 9.99481499e-01 2.73171335e-01 -1.86335281e-01
-1.09997725e+00 -8.56737077e-01 -5.81049919e-01 3.03876102e-01
1.77377447e-01 -1.10924399e+00 1.23086786e+00 -4.60660964e-01
-1.09950304e+00 8.48850906e-01 4.25887555e-02 -5.13165593e-01
5.55207372e-01 -2.55169749e-01 -7.69867152e-02 4.17026937e-01
9.21778902e-02 2.16016456e-01 3.00387084e-01 -7.31312037e-01
-6.28730416e-01 -3.78211349e-01 -5.22002578e-01 1.93991348e-01
-6.70169413e-01 8.35372657e-02 -3.65344495e-01 -6.68691039e-01
2.63077915e-02 -8.23367000e-01 -7.15043962e-01 3.71354550e-01
-1.98430017e-01 -3.43526542e-01 8.58210683e-01 -7.84225464e-01
1.24696660e+00 -1.76595259e+00 -3.16351295e-01 1.26101688e-01
5.87022245e-01 4.82206136e-01 4.11453307e-01 -9.49199125e-02
-1.38398498e-01 4.55064803e-01 -4.82582659e-01 8.77386332e-02
-7.19900727e-01 -2.91284006e-02 4.31573153e-01 3.73826891e-01
1.37387976e-01 1.01035750e+00 -6.37525380e-01 -9.64502335e-01
4.11856115e-01 5.88990688e-01 -1.72086045e-01 4.60359991e-01
1.70921206e-01 4.31223482e-01 -7.07972825e-01 4.32576209e-01
3.82432550e-01 -5.55322111e-01 5.36662759e-03 -9.93170366e-02
2.50691324e-01 4.08025272e-02 -3.76686960e-01 1.13559723e+00
-6.09020233e-01 2.25791797e-01 -4.24602211e-01 -7.94741452e-01
7.17824996e-01 7.90638566e-01 9.14431512e-01 -3.37699383e-01
3.23295563e-01 2.95446754e-01 6.51692320e-03 -6.50912881e-01
-5.09830594e-01 -6.65424943e-01 1.56722423e-02 4.60925817e-01
-3.12524140e-01 6.60330355e-02 -3.34378302e-01 1.94823921e-01
1.50880885e+00 -4.01872963e-01 6.92168772e-01 -1.44686103e-01
8.94831777e-01 5.96844107e-02 4.40629154e-01 4.52873349e-01
-3.90986919e-01 8.77623320e-01 2.61326820e-01 -6.66971385e-01
-7.36228347e-01 -1.24683797e+00 -3.33528847e-01 5.62127590e-01
-2.90633440e-01 -1.66308992e-02 -5.87280512e-01 -1.13126552e+00
-2.01201051e-01 3.52021903e-01 -6.55582130e-01 -3.53141874e-01
-8.80591512e-01 -1.27282715e+00 8.02452326e-01 8.46702158e-01
8.18024158e-01 -9.00598764e-01 -7.17416883e-01 1.29691377e-01
-2.74745911e-01 -5.73588431e-01 -1.79474130e-01 3.79195362e-01
-1.37289035e+00 -1.48849523e+00 -1.28018558e+00 -7.61514664e-01
4.96591508e-01 1.62899852e-01 1.18881500e+00 6.49456561e-01
-6.56120539e-01 2.87194222e-01 -7.90165588e-02 -2.16591775e-01
-7.05768347e-01 1.75724804e-01 -2.24583521e-01 -4.20417368e-01
2.41232440e-01 -4.01715964e-01 -1.02577698e+00 1.23235404e-01
-7.03384221e-01 4.62393090e-03 1.17995310e+00 7.51907706e-01
8.35750639e-01 2.67595649e-01 5.58876574e-01 -1.14676666e+00
3.69764298e-01 -7.40662038e-01 2.49080602e-02 2.48701245e-01
-7.95400798e-01 -6.73163161e-02 7.47921288e-01 -1.15712807e-02
-1.09035969e+00 3.14395547e-01 -3.55744094e-01 -2.61541545e-01
-2.82370895e-01 5.02006233e-01 9.83178169e-02 7.43258223e-02
7.66303480e-01 5.86602129e-02 -1.44406497e-01 -5.27605176e-01
-2.47635201e-01 5.60403407e-01 5.18821955e-01 -1.71738669e-01
8.28156114e-01 3.52748305e-01 6.23618901e-01 -5.39861619e-01
-1.00912058e+00 -6.57997549e-01 -6.70839608e-01 -1.12263881e-01
1.34833574e+00 -1.04976070e+00 -4.27610487e-01 3.04530323e-01
-8.06771696e-01 -1.11618243e-01 -1.42295659e-01 8.23942780e-01
-3.56293440e-01 5.81294715e-01 -8.04660618e-01 -4.25728887e-01
-8.30039144e-01 -7.68550158e-01 9.10109341e-01 1.49232239e-01
7.98461959e-02 -1.25414741e+00 3.23694855e-01 3.13012064e-01
4.98509586e-01 3.82447720e-01 1.36489010e+00 -7.46627271e-01
-4.65986162e-01 -5.88410556e-01 -5.86632729e-01 4.22397882e-01
1.95154101e-01 -3.43901366e-01 -7.82772601e-01 -1.89197466e-01
1.41567945e-01 -4.81293768e-01 1.22015727e+00 6.77007437e-01
1.51570594e+00 -1.22732289e-01 -5.81976950e-01 7.09971309e-01
1.70211756e+00 1.31611735e-01 6.03699923e-01 1.06551774e-01
8.99511993e-01 2.24566892e-01 3.16021293e-01 2.15561002e-01
2.94668198e-01 3.16171587e-01 6.74572945e-01 -5.59523880e-01
-4.40241218e-01 9.46906507e-02 2.00904608e-02 6.94958568e-01
-3.57714891e-01 -1.79813817e-01 -1.23951244e+00 2.87173092e-01
-1.47065544e+00 -9.28098738e-01 -7.49220669e-01 2.21495390e+00
6.72978997e-01 -8.16808417e-02 2.14039311e-02 1.98980659e-01
6.28894091e-01 -2.05207154e-01 -4.56891745e-01 1.19020909e-01
4.05560136e-01 5.01898289e-01 6.32874489e-01 2.06036102e-02
-1.29557550e+00 2.23914042e-01 6.44557905e+00 7.50286281e-01
-1.08166623e+00 4.80130732e-01 1.14319515e+00 1.09479927e-01
1.00344382e-01 -2.26749584e-01 -5.99232733e-01 2.37052232e-01
1.09129298e+00 -6.64457381e-02 -4.25734490e-01 8.17093432e-01
3.48877788e-01 -1.29235595e-01 -9.56971645e-01 6.58035696e-01
-1.57777891e-01 -1.02259314e+00 1.05240986e-01 7.09687099e-02
2.50205636e-01 3.79732698e-02 -9.50078741e-02 3.29990715e-01
2.86951929e-01 -1.24198294e+00 -6.14277609e-02 5.32706082e-01
1.09485042e+00 -5.67828536e-01 1.14975131e+00 5.53378701e-01
-1.34223723e+00 -9.50654596e-02 -3.40936154e-01 1.86647803e-01
-2.88907010e-02 7.58882225e-01 -1.10759652e+00 5.16933382e-01
7.09280610e-01 5.22925377e-01 -9.08418953e-01 1.09285080e+00
5.22187836e-02 1.00555420e+00 -1.89210713e-01 3.04299197e-03
-3.36294286e-02 4.51349556e-01 -3.50063220e-02 1.15420616e+00
7.87630737e-01 3.92508566e-01 7.88153037e-02 5.95703363e-01
-6.44712597e-02 3.99162024e-01 -6.61904335e-01 3.60721171e-01
-1.02859877e-01 1.40875995e+00 -8.97364318e-01 -5.11347651e-01
-4.57774669e-01 7.92903185e-01 8.51796195e-02 -2.57876039e-01
-1.01119494e+00 6.49647340e-02 -7.57176057e-02 4.87171292e-01
-1.39131263e-01 1.74256653e-01 -4.47089761e-01 -9.82605815e-01
-5.41232288e-01 -3.02208453e-01 8.23527038e-01 -4.05179083e-01
-1.27140009e+00 5.50032914e-01 -1.29952148e-01 -1.30052829e+00
-7.43944719e-02 -6.00689888e-01 -1.03459573e+00 8.01223934e-01
-1.42430913e+00 -1.13606191e+00 -6.71654820e-01 6.18133366e-01
5.06154776e-01 -2.29819894e-01 1.00229347e+00 3.54869574e-01
-6.65185034e-01 4.17587578e-01 7.10556656e-02 2.29174182e-01
5.90800405e-01 -1.41983712e+00 -1.73362061e-01 3.64283174e-01
-3.35768253e-01 7.67569840e-02 -7.71913826e-02 -7.98210084e-01
-8.39516580e-01 -1.75645280e+00 8.48718762e-01 -5.18736482e-01
6.31232262e-02 3.32402349e-01 -8.71848941e-01 2.58829802e-01
-2.05889285e-01 4.15467232e-01 9.47058201e-01 -3.63747030e-01
1.02700450e-01 -1.83446519e-02 -1.22647786e+00 2.27964818e-01
5.94498396e-01 -2.20170245e-01 -3.99666727e-01 4.25230891e-01
5.04954517e-01 5.03253117e-02 -1.01501942e+00 8.95088911e-01
5.61908126e-01 -1.08295166e+00 1.39560997e+00 -3.97559583e-01
8.03192616e-01 3.74603570e-02 7.52519220e-02 -6.90123320e-01
-8.62253726e-01 5.44764936e-01 -8.38423148e-02 8.91858280e-01
4.03003305e-01 -2.18896300e-01 1.22654903e+00 6.26066327e-01
6.80982545e-02 -1.06702137e+00 -9.96539950e-01 -5.81466019e-01
5.12865901e-01 -2.38736972e-01 5.42736575e-02 6.58906937e-01
-6.18762016e-01 2.13532865e-01 -1.50157794e-01 1.48688942e-01
5.89557171e-01 -2.99088657e-01 3.14112276e-01 -1.50834262e+00
-3.90706986e-01 -4.80213612e-01 -3.50223094e-01 -2.97138363e-01
-2.54482210e-01 -1.12891698e+00 -9.25123841e-02 -2.02179027e+00
8.92435849e-01 -4.96646166e-01 -8.80618393e-01 3.49648595e-01
-5.79242706e-01 5.52040219e-01 -5.65677807e-02 6.11180842e-01
-3.09978098e-01 3.57590705e-01 1.35314369e+00 -1.04075439e-01
5.99913392e-03 4.55963552e-01 -4.33369607e-01 9.93538022e-01
9.75352764e-01 -8.10581803e-01 -2.35232934e-01 2.14164741e-02
-1.60307229e-01 5.75742245e-01 6.12401187e-01 -1.53895521e+00
-1.59081027e-01 -2.39434242e-01 8.01288486e-01 -6.35000765e-01
-4.78134900e-02 -1.12173033e+00 1.52653992e-01 1.32927036e+00
-3.66571367e-01 -2.75678545e-01 -9.25505310e-02 6.96507573e-01
-1.43878832e-01 -3.81519765e-01 1.08650267e+00 -5.33063650e-01
-2.13970825e-01 6.77486002e-01 -6.73249304e-01 -2.08274022e-01
1.32191014e+00 -4.76035438e-02 1.81546099e-02 -1.99201450e-01
-6.98281944e-01 -1.71814803e-02 -5.11975922e-02 -1.95253864e-01
6.22943878e-01 -1.48512566e+00 -7.03917861e-01 -9.48499739e-02
9.69357323e-03 1.92055047e-01 2.07607761e-01 1.08653533e+00
-9.08399165e-01 4.20450568e-01 8.80939066e-02 -4.95684773e-01
-1.46038449e+00 4.39997196e-01 7.20145047e-01 -1.13920414e+00
-7.74466693e-01 8.17292690e-01 6.83014572e-01 -2.11907983e-01
1.82857975e-01 -3.36006552e-01 -5.03554940e-01 -3.02521408e-01
2.52886295e-01 4.30380434e-01 6.40578195e-02 -3.59041095e-01
-2.98985958e-01 6.31512403e-01 -6.92876354e-02 3.15748274e-01
1.23032629e+00 2.02920705e-01 -3.97611372e-02 2.27895513e-01
1.19532788e+00 -2.13148326e-01 -6.42192543e-01 -7.38761127e-02
1.39281347e-01 -1.19612291e-01 4.24395621e-01 -1.16205823e+00
-1.33013964e+00 1.25667155e+00 1.29400682e+00 -4.92499992e-02
1.35869431e+00 1.94377363e-01 1.10880899e+00 5.35566866e-01
-1.15891779e-02 -4.32228297e-01 4.02055889e-01 6.88837543e-02
5.47735810e-01 -1.51668882e+00 1.45371109e-01 -4.44516093e-01
-4.60462600e-01 1.42005658e+00 7.34887302e-01 -2.26384103e-01
9.77527916e-01 1.72649547e-01 3.11418265e-01 -2.71772355e-01
-7.26671875e-01 -1.59762740e-01 2.07148194e-01 3.41784626e-01
8.55399787e-01 1.55645758e-01 -4.62913722e-01 8.81519437e-01
2.65158296e-01 3.84921193e-01 6.58469945e-02 7.25869119e-01
-8.74288261e-01 -1.17345858e+00 -5.52676320e-01 1.22714949e+00
-1.10236549e+00 -1.43736884e-01 -4.06418383e-01 7.19793320e-01
3.04974973e-01 3.67774576e-01 -2.80635208e-01 -2.81784415e-01
3.65029536e-02 2.18587533e-01 8.37371200e-02 -8.13972890e-01
-7.27713406e-01 -1.42231151e-01 -1.83235750e-01 -3.08528066e-01
-4.55287993e-01 -3.82256567e-01 -1.53517187e+00 2.79842243e-02
-2.77352661e-01 -1.43704712e-01 3.58331710e-01 7.47984886e-01
-1.44580677e-01 8.50630820e-01 5.97935557e-01 -4.08116013e-01
-5.38852155e-01 -9.68669593e-01 -3.98450643e-01 1.75455958e-01
2.07824096e-01 -5.12391210e-01 -3.38746995e-01 6.97085932e-02] | [15.326347351074219, -2.231928586959839] |
841f8bb0-358a-423e-a78a-6ed93b838921 | semantic-parsing-in-task-oriented-dialog-with | 2109.04500 | null | https://arxiv.org/abs/2109.04500v2 | https://arxiv.org/pdf/2109.04500v2.pdf | Semantic Parsing in Task-Oriented Dialog with Recursive Insertion-based Encoder | We introduce a Recursive INsertion-based Encoder (RINE), a novel approach for semantic parsing in task-oriented dialog. Our model consists of an encoder network that incrementally builds the semantic parse tree by predicting the non-terminal label and its positions in the linearized tree. At the generation time, the model constructs the semantic parse tree by recursively inserting the predicted non-terminal labels at the predicted positions until termination. RINE achieves state-of-the-art exact match accuracy on low- and high-resource versions of the conversational semantic parsing benchmark TOP (Gupta et al., 2018; Chen et al., 2020), outperforming strong sequence-to-sequence models and transition-based parsers. We also show that our model design is applicable to nested named entity recognition task, where it performs on par with state-of-the-art approach designed for that task. Finally, we demonstrate that our approach is 2-3.5 times faster than the sequence-to-sequence model at inference time. | ['Yi Zhang', 'Elman Mansimov'] | 2021-09-09 | null | null | null | null | ['nested-named-entity-recognition'] | ['natural-language-processing'] | [ 4.57388312e-01 9.03087556e-01 -8.88054296e-02 -7.52620399e-01
-1.16791463e+00 -9.70916867e-01 2.65782416e-01 -2.04017852e-02
-3.33965182e-01 6.39924467e-01 6.21056497e-01 -6.83221579e-01
4.16597426e-01 -7.77959526e-01 -7.81637192e-01 1.37054339e-01
-1.43282026e-01 1.04092836e+00 2.69133568e-01 -3.28499317e-01
-5.69817703e-03 -1.88412637e-01 -1.25036311e+00 7.91062653e-01
6.08365417e-01 6.46564841e-01 2.83135146e-01 1.02882588e+00
-8.68786335e-01 1.09624219e+00 -4.18871820e-01 -8.21282148e-01
-5.83246127e-02 -5.43085039e-01 -1.55566144e+00 -2.68178523e-01
1.70283377e-01 -2.90631741e-01 -9.44647193e-02 7.74093390e-01
1.74365327e-01 -1.28340542e-01 1.62073046e-01 -9.66119349e-01
-5.45730531e-01 1.03230786e+00 8.33944753e-02 -1.54368356e-01
5.65290153e-01 -1.98048294e-01 1.68119645e+00 -6.28264368e-01
1.01781690e+00 1.45002532e+00 8.03869486e-01 1.02298832e+00
-1.06752348e+00 -3.45915228e-01 3.72732282e-01 -5.99682555e-02
-6.23798549e-01 -3.35926026e-01 3.62179101e-01 -2.17658415e-01
1.41407692e+00 7.05519840e-02 1.69015631e-01 1.06708467e+00
2.49813944e-02 9.76151884e-01 9.17681396e-01 -4.19522226e-01
2.85327703e-01 -2.80999988e-01 5.05581856e-01 1.04579329e+00
-3.89541626e-01 -2.22493142e-01 -5.98796368e-01 -4.43985760e-01
2.10391402e-01 -5.57964623e-01 1.36047408e-01 -6.66910410e-02
-8.92211497e-01 9.13199961e-01 2.78744608e-01 2.08241105e-01
-2.59484619e-01 1.72960967e-01 6.66687369e-01 1.54299974e-01
4.98704255e-01 2.98952550e-01 -1.00759733e+00 -6.04063272e-01
-4.93159175e-01 2.50932246e-01 1.58470297e+00 1.19004440e+00
5.31542838e-01 -4.78803724e-01 -9.79039669e-02 9.47779000e-01
3.01399827e-01 -2.21551135e-02 3.85153383e-01 -1.39735317e+00
7.37788141e-01 4.70213562e-01 1.28622204e-01 -4.42448407e-02
-4.47208226e-01 4.25319485e-02 -1.65626675e-01 -2.89792538e-01
6.57014370e-01 -3.72319996e-01 -7.11491168e-01 2.10074615e+00
4.50889409e-01 2.03399360e-01 5.03213704e-01 5.49778104e-01
9.48896766e-01 7.73312211e-01 6.17599070e-01 8.83499533e-03
1.87223446e+00 -1.29748654e+00 -5.72250485e-01 -6.93583071e-01
9.09313321e-01 -4.36399639e-01 9.29404497e-01 -2.60094553e-02
-1.13212287e+00 -3.73984605e-01 -6.60506248e-01 -3.86063248e-01
-2.51645684e-01 -6.64008260e-02 7.14245677e-01 5.76450348e-01
-1.24011493e+00 6.51219428e-01 -8.62815917e-01 -5.17060041e-01
9.65915024e-02 -1.84016172e-02 -2.97620356e-01 8.44743922e-02
-1.25177240e+00 9.41942155e-01 4.71527159e-01 -1.02603495e-01
-7.11475194e-01 -7.11823404e-01 -1.08431554e+00 1.40894145e-01
3.21025342e-01 -8.42918873e-01 2.21851563e+00 -7.66080141e-01
-1.87404418e+00 1.05124426e+00 -8.38528991e-01 -7.22761869e-01
2.54478902e-01 -4.11178708e-01 -7.70537555e-02 9.33788642e-02
3.33694011e-01 1.00372708e+00 9.29087102e-02 -1.09524548e+00
-7.81892478e-01 -2.80708343e-01 3.51230025e-01 3.49342734e-01
3.02537441e-01 4.41154957e-01 -3.69222403e-01 -1.48007110e-01
2.20908314e-01 -9.83284712e-01 -3.62733573e-01 -3.43994945e-01
-5.35890698e-01 -5.59000969e-01 5.06129801e-01 -1.23915803e+00
9.75811243e-01 -1.88398373e+00 1.01763830e-01 -4.62607145e-01
1.56200901e-01 1.17986947e-01 -2.68520325e-01 7.03604162e-01
-1.21591337e-01 2.71905452e-01 -6.27995670e-01 -7.36390352e-01
1.92598045e-01 3.71358812e-01 -3.01309198e-01 -1.51686043e-01
3.98716331e-01 1.23294222e+00 -1.11080885e+00 -4.67268318e-01
-3.66631225e-02 7.83441141e-02 -7.78960109e-01 7.38115251e-01
-8.08534682e-01 4.15424556e-01 -4.75883752e-01 4.22796816e-01
2.45233417e-01 -3.81054461e-01 8.34872901e-01 1.43695042e-01
2.76392493e-02 1.38324940e+00 -4.89873797e-01 1.86936677e+00
-8.53197336e-01 4.94454771e-01 3.07598650e-01 -8.83266032e-01
8.13355386e-01 5.53664923e-01 1.89580381e-01 -4.48456347e-01
-1.88743696e-01 1.71813965e-01 -1.27826348e-01 -3.04246694e-01
5.52644849e-01 -2.61760890e-01 -5.46942294e-01 5.31776965e-01
1.89518511e-01 5.62577210e-02 2.31384244e-02 3.73566926e-01
1.43408668e+00 5.58871746e-01 3.23626608e-01 -4.44258787e-02
4.32966292e-01 1.67730466e-01 6.21010005e-01 7.08386600e-01
-1.04804270e-01 1.17570139e-01 9.58247304e-01 -4.02858615e-01
-1.16476631e+00 -1.06540692e+00 9.59695429e-02 1.67139399e+00
-2.43605405e-01 -7.39992082e-01 -1.26513755e+00 -1.06290829e+00
-1.69714242e-01 1.14113271e+00 -4.85667676e-01 2.95435160e-01
-1.05782044e+00 -2.62655348e-01 9.42481101e-01 5.83808720e-01
5.20747185e-01 -1.37918830e+00 -4.56053615e-01 7.18982577e-01
-6.24766529e-01 -1.49471200e+00 -1.97053924e-01 3.90377134e-01
-8.82100105e-01 -1.06761611e+00 -9.47830528e-02 -1.04979193e+00
2.57257521e-01 -2.42331699e-01 1.56227338e+00 -6.97998926e-02
1.38084725e-01 1.91352814e-01 -5.28512180e-01 -1.59786060e-01
-1.14684081e+00 3.96607608e-01 -6.80385232e-01 -4.93739367e-01
4.08762068e-01 -4.28198367e-01 -2.10746348e-01 2.05529615e-01
-4.06232327e-01 3.96292299e-01 1.81881756e-01 8.00636053e-01
2.82519668e-01 -4.60288405e-01 7.25527108e-01 -1.47990263e+00
5.13556421e-01 -4.92856771e-01 -4.31904793e-01 3.87527853e-01
-2.06828594e-01 3.94495040e-01 8.98639977e-01 2.40310445e-01
-1.46489739e+00 1.65068150e-01 -6.92000806e-01 3.18960994e-01
-3.77620429e-01 2.95079648e-01 -2.64341146e-01 5.54202199e-01
3.01831454e-01 2.19096020e-01 -1.12554669e-01 -7.68001139e-01
8.13536584e-01 8.91824245e-01 6.53166175e-01 -9.01914775e-01
2.63452530e-01 6.64122850e-02 -3.19073290e-01 -3.86402190e-01
-1.21695495e+00 -3.64760548e-01 -6.75983846e-01 2.41705924e-01
1.24256265e+00 -9.54089284e-01 -9.12280798e-01 2.65334934e-01
-1.71605480e+00 -6.83996618e-01 -1.12124905e-01 -1.19566649e-01
-7.15984583e-01 4.09968972e-01 -1.29317498e+00 -7.85788953e-01
-6.36032224e-01 -9.48244989e-01 1.30204749e+00 2.87227854e-02
-4.07901019e-01 -1.00123572e+00 6.81325570e-02 8.15470695e-01
1.84130892e-01 -2.23679960e-01 1.24376714e+00 -1.27170014e+00
-5.03059506e-01 2.73244530e-01 -2.10621223e-01 1.91187337e-01
-1.37335151e-01 -4.52245533e-01 -9.35040176e-01 1.31134570e-01
-1.82767153e-01 -4.16703433e-01 6.81392610e-01 4.49038334e-02
9.56559896e-01 -4.07683194e-01 -2.61415869e-01 2.48344094e-01
8.99066091e-01 3.40968490e-01 4.51471865e-01 3.36606413e-01
3.93058598e-01 8.96321714e-01 5.04138112e-01 8.50562304e-02
1.03092718e+00 5.95122278e-01 2.57122427e-01 2.94747204e-01
-3.76920588e-02 -7.45401680e-01 4.60080862e-01 8.10699046e-01
4.33303684e-01 -4.29435015e-01 -8.71573865e-01 5.63309908e-01
-1.89260340e+00 -7.93934047e-01 -1.65644228e-01 1.88549101e+00
1.04602599e+00 1.43389463e-01 1.47632090e-02 -5.90119720e-01
9.13749278e-01 2.88572609e-01 -4.75922823e-01 -9.46506202e-01
2.62305260e-01 5.23435950e-01 4.87800658e-01 8.59874308e-01
-1.22258043e+00 1.67286968e+00 7.02448893e+00 3.11488152e-01
-3.37171704e-01 4.27519649e-01 6.54501796e-01 4.42831457e-01
-3.38439941e-01 4.60736811e-01 -1.02982974e+00 4.57540631e-01
1.63032174e+00 1.67857930e-01 6.41889513e-01 1.05773783e+00
-1.25837654e-01 1.48575082e-01 -1.21990466e+00 2.87392884e-01
-1.65854380e-01 -1.30571103e+00 -3.36298227e-01 -2.43302062e-01
3.13741833e-01 2.48787731e-01 -7.21135914e-01 6.51252866e-01
1.09335351e+00 -9.40712988e-01 6.88870788e-01 -2.13738258e-05
5.02941787e-01 -4.06067401e-01 6.87042236e-01 3.91158521e-01
-1.28353405e+00 -1.65218432e-02 -1.58545941e-01 -8.00164342e-02
8.30300510e-01 2.99254358e-01 -1.14234471e+00 4.91776377e-01
4.35035884e-01 4.99603480e-01 -2.27037698e-01 2.10725889e-01
-7.98753500e-01 9.75060582e-01 -1.58131391e-01 -1.86853766e-01
4.21178490e-01 -1.83179423e-01 3.92305583e-01 1.52459955e+00
-1.04709372e-01 2.96901971e-01 2.36796945e-01 7.14278579e-01
-4.15824831e-01 -3.90653461e-02 -3.42413455e-01 -1.45251244e-01
8.00553083e-01 1.05908024e+00 -4.77124780e-01 -7.31165826e-01
-4.98125285e-01 1.25043619e+00 8.54899406e-01 3.93477529e-02
-7.26189435e-01 -1.77530631e-01 7.39480078e-01 -4.47874546e-01
5.65438688e-01 -2.25827530e-01 -2.87398100e-01 -9.40203726e-01
-1.32014137e-02 -9.60572481e-01 6.46794975e-01 -8.19821358e-01
-1.23881304e+00 6.72772884e-01 -1.17095821e-01 -4.31547672e-01
-7.37104654e-01 -7.36504614e-01 -6.05988920e-01 8.79666507e-01
-1.34103501e+00 -1.16624010e+00 1.28189906e-01 2.70908978e-03
8.19862485e-01 1.31595179e-01 1.25785947e+00 1.84825738e-03
-3.37902308e-01 3.70563954e-01 -1.71562016e-01 6.17047250e-01
4.15311664e-01 -1.41817057e+00 1.52657628e+00 6.40528500e-01
9.63943526e-02 5.75181723e-01 5.15075922e-01 -7.25439072e-01
-1.25355089e+00 -1.00219023e+00 1.59160936e+00 -7.04727829e-01
8.37016582e-01 -7.48523593e-01 -9.09115136e-01 1.30534935e+00
3.30746263e-01 -4.50512111e-01 6.86844885e-01 5.26213109e-01
-6.50820494e-01 5.20084739e-01 -1.05601752e+00 4.61159080e-01
1.48411381e+00 -7.24423707e-01 -1.07114828e+00 4.41872925e-01
1.31403458e+00 -7.47549832e-01 -8.31204891e-01 1.58897832e-01
5.72763026e-01 -9.70781684e-01 8.42680871e-01 -1.01276386e+00
8.03602457e-01 2.29023516e-01 -2.31608853e-01 -1.17946684e+00
-1.38431072e-01 -6.04308188e-01 3.57005335e-02 1.31446362e+00
7.91131854e-01 -6.15357399e-01 1.03623760e+00 7.66774595e-01
-5.28782606e-01 -6.57501817e-01 -1.14446712e+00 -5.19538760e-01
2.11205930e-01 -5.18676162e-01 5.68697631e-01 6.01406097e-01
2.73813635e-01 8.06636274e-01 -1.56578824e-01 9.95055214e-02
4.20015037e-01 2.64208496e-01 5.74433804e-01 -1.02761126e+00
-4.88631189e-01 -1.41647637e-01 5.99309094e-02 -1.54109573e+00
9.00161326e-01 -9.83350098e-01 4.83071953e-01 -1.79604137e+00
1.34477153e-01 -4.95372504e-01 4.25877050e-02 7.33333886e-01
-3.57568085e-01 -3.83784324e-01 3.02477419e-01 -1.23495787e-01
-7.22413838e-01 4.71601635e-01 8.24847937e-01 1.59917548e-01
-8.43868703e-02 -1.53490275e-01 -7.80445874e-01 6.43244267e-01
8.93285275e-01 -5.82564890e-01 -1.77781388e-01 -4.56110001e-01
2.26901527e-02 5.13327658e-01 1.03630103e-01 -5.06333470e-01
1.03506699e-01 -1.95187423e-02 -3.64353001e-01 -4.98009861e-01
4.36920345e-01 -3.65436345e-01 2.50847917e-02 3.89919311e-01
-8.42938900e-01 1.31555080e-01 1.20800093e-01 4.37732518e-01
-6.64783493e-02 -6.56180620e-01 4.55975354e-01 -5.10366440e-01
-7.29400754e-01 -1.34365708e-01 -5.06740749e-01 5.93602479e-01
6.20451570e-01 1.68812871e-01 -5.82057655e-01 -5.58962762e-01
-6.62249088e-01 2.90656686e-01 1.46152884e-01 5.06832421e-01
1.09270476e-01 -7.51859069e-01 -6.45396113e-01 -1.38767064e-01
1.69913191e-03 5.95747381e-02 3.02545913e-02 2.18091279e-01
-5.30663967e-01 6.12766743e-01 2.50749648e-01 -2.63374001e-01
-1.24328697e+00 2.56433994e-01 2.56456316e-01 -7.72337437e-01
-7.20790148e-01 9.19976711e-01 2.38102674e-01 -1.23036981e+00
2.60501113e-02 -2.69471586e-01 9.23271105e-02 -3.00205439e-01
3.84475112e-01 9.82342139e-02 4.65210006e-02 -4.48611319e-01
-5.33841610e-01 1.42024517e-01 -1.16450056e-01 -4.04910773e-01
1.23637831e+00 -5.17581180e-02 -3.72202545e-01 3.52182716e-01
1.28135395e+00 -1.44586846e-01 -1.15336633e+00 -3.78304005e-01
5.05140483e-01 6.75356388e-02 -4.88139838e-01 -1.06638217e+00
-4.46794063e-01 7.64254868e-01 -1.00232616e-01 7.09638894e-02
5.74432611e-01 4.14786607e-01 1.28929090e+00 6.65988028e-01
5.27310133e-01 -9.03229475e-01 -1.94910407e-01 1.08575392e+00
3.09166729e-01 -1.01329601e+00 -5.42136550e-01 -8.60074580e-01
-8.50255132e-01 9.23613191e-01 4.74991858e-01 2.21565217e-01
1.86348304e-01 5.11422694e-01 1.19580664e-01 -6.20859861e-02
-1.13016069e+00 -1.85939938e-01 -2.67228991e-01 4.99616474e-01
8.02125394e-01 3.42203528e-01 -3.35742861e-01 7.62829483e-01
-5.57465911e-01 -1.98152974e-01 3.04774880e-01 1.05804408e+00
-6.00793481e-01 -1.54501736e+00 7.81421289e-02 1.78612083e-01
-5.85678577e-01 -5.59387743e-01 -5.64053893e-01 5.88662088e-01
-3.30091774e-01 1.27361977e+00 2.95444071e-01 -2.39329904e-01
2.81959504e-01 9.42717314e-01 1.98774219e-01 -8.70995641e-01
-8.84247839e-01 -3.11567545e-01 1.21940041e+00 -7.93562233e-01
-1.42156616e-01 -6.48512125e-01 -1.69815123e+00 -1.94154799e-01
-1.29590556e-02 5.04300773e-01 7.61803150e-01 9.52341020e-01
7.23675072e-01 3.81386787e-01 5.36919057e-01 -4.38953966e-01
-6.58608317e-01 -9.90317166e-01 -1.15410574e-01 4.34237331e-01
-1.36805817e-01 -2.38275617e-01 -1.26693264e-01 1.70572311e-01] | [10.480875968933105, 9.303455352783203] |
6f786014-ba15-4279-9cc5-0a74d090b59e | smpconv-self-moving-point-representations-for | 2304.02330 | null | https://arxiv.org/abs/2304.02330v1 | https://arxiv.org/pdf/2304.02330v1.pdf | SMPConv: Self-moving Point Representations for Continuous Convolution | Continuous convolution has recently gained prominence due to its ability to handle irregularly sampled data and model long-term dependency. Also, the promising experimental results of using large convolutional kernels have catalyzed the development of continuous convolution since they can construct large kernels very efficiently. Leveraging neural networks, more specifically multilayer perceptrons (MLPs), is by far the most prevalent approach to implementing continuous convolution. However, there are a few drawbacks, such as high computational costs, complex hyperparameter tuning, and limited descriptive power of filters. This paper suggests an alternative approach to building a continuous convolution without neural networks, resulting in more computationally efficient and improved performance. We present self-moving point representations where weight parameters freely move, and interpolation schemes are used to implement continuous functions. When applied to construct convolutional kernels, the experimental results have shown improved performance with drop-in replacement in the existing frameworks. Due to its lightweight structure, we are first to demonstrate the effectiveness of continuous convolution in a large-scale setting, e.g., ImageNet, presenting the improvements over the prior arts. Our code is available on https://github.com/sangnekim/SMPConv | ['Eunbyung Park', 'Sanghyeon Kim'] | 2023-04-05 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Kim_SMPConv_Self-Moving_Point_Representations_for_Continuous_Convolution_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Kim_SMPConv_Self-Moving_Point_Representations_for_Continuous_Convolution_CVPR_2023_paper.pdf | cvpr-2023-1 | ['sequential-image-classification'] | ['computer-vision'] | [-1.75111637e-01 -2.89154440e-01 1.97433028e-03 -2.10408330e-01
-2.24710733e-01 -2.28309497e-01 3.95118266e-01 -3.26426089e-01
-7.31553197e-01 6.13204002e-01 1.31364703e-01 -2.82951772e-01
-1.05993599e-01 -7.62161851e-01 -8.67864370e-01 -5.84063053e-01
-1.70246169e-01 -4.34594154e-01 2.02647448e-01 -4.41479310e-02
4.04922850e-02 7.99845695e-01 -1.53489292e+00 2.65567809e-01
8.52059901e-01 1.05017233e+00 2.95629472e-01 6.83127522e-01
-5.65536844e-04 8.84261906e-01 -2.27531821e-01 -5.15931010e-01
3.08267325e-01 1.45567536e-01 -6.39873207e-01 -1.49132073e-01
5.43419600e-01 -4.51446921e-01 -6.58166111e-01 8.88489485e-01
5.33976138e-01 2.79649466e-01 2.14983910e-01 -1.21196997e+00
-9.54597533e-01 5.87986529e-01 -2.58884817e-01 8.37329403e-02
-2.76438177e-01 2.95293599e-01 8.85440230e-01 -1.11216366e+00
1.08904883e-01 1.15629435e+00 1.17266774e+00 4.31500316e-01
-1.21865487e+00 -6.77636504e-01 1.56220064e-01 1.58698782e-01
-1.45246232e+00 -4.70280617e-01 5.46367705e-01 -1.85685456e-01
1.10269403e+00 3.05834144e-01 6.04789317e-01 1.10868001e+00
-1.11117855e-01 9.61687684e-01 9.30596530e-01 -4.76406276e-01
1.00270964e-01 7.05082417e-02 1.16002344e-01 7.11761355e-01
5.88668026e-02 1.95953086e-01 -1.22777484e-01 -1.72027335e-01
1.26183486e+00 3.81255478e-01 -4.40198481e-01 -2.32518949e-02
-1.21505439e+00 7.54717708e-01 8.95193458e-01 1.53728724e-01
-4.94728684e-01 7.47355580e-01 4.65055406e-01 2.10889325e-01
4.89185929e-01 3.01871270e-01 -5.58817863e-01 -2.97583312e-01
-8.25895309e-01 2.08370388e-01 8.57014120e-01 8.80225062e-01
7.79908359e-01 1.39002904e-01 -1.35429695e-01 1.03594816e+00
-3.16681936e-02 1.96232781e-01 4.83262330e-01 -8.75378609e-01
2.67610997e-01 4.86485124e-01 -1.02879934e-01 -8.10712159e-01
-2.95289189e-01 -5.62863708e-01 -1.20843637e+00 4.51687545e-01
3.77322823e-01 -2.16714576e-01 -1.06304777e+00 1.51969957e+00
3.74145322e-02 6.67807758e-01 9.73126590e-02 6.60961211e-01
7.68319726e-01 6.26640499e-01 1.52459756e-01 4.35112089e-01
1.40125906e+00 -1.23191786e+00 -4.30417567e-01 6.81625754e-02
4.84829813e-01 -8.48794341e-01 1.27691436e+00 2.20099583e-01
-1.03532970e+00 -6.69608533e-01 -9.39043701e-01 -1.88787296e-01
-4.86558080e-01 5.04084110e-01 1.09312570e+00 4.70375657e-01
-1.27295387e+00 9.15375590e-01 -1.10022891e+00 -2.11694464e-01
9.46125388e-01 3.74322236e-01 -1.57349259e-01 6.22179210e-02
-1.08699429e+00 7.60101855e-01 5.86973310e-01 1.78544626e-01
-2.70733058e-01 -1.17453837e+00 -6.92191541e-01 3.72016430e-01
1.56808287e-01 -5.17223835e-01 1.27712941e+00 -9.25295532e-01
-1.60404360e+00 2.61301637e-01 2.52095852e-02 -6.43040001e-01
5.32378435e-01 -4.29820031e-01 -4.14477021e-01 4.19254974e-02
-4.25688893e-01 8.41720581e-01 7.64280796e-01 -7.01309621e-01
-5.78872442e-01 3.16856205e-01 2.47139618e-01 -1.04637936e-01
-8.22752416e-01 5.47481775e-02 -5.76821566e-01 -8.15800130e-01
-1.77472487e-01 -9.39106405e-01 -3.58745933e-01 4.05390471e-01
-1.48219794e-01 -3.06721866e-01 6.66753352e-01 -4.45914775e-01
1.35537601e+00 -2.43333864e+00 -5.27330458e-01 2.54010060e-03
1.99208409e-01 6.29598856e-01 -2.47063134e-02 3.63272250e-01
-1.42200634e-01 8.99067819e-02 -2.64431089e-01 -2.61682212e-01
9.77222528e-03 2.72206515e-01 -3.95939022e-01 2.77775645e-01
5.35392582e-01 1.19932675e+00 -7.04551220e-01 -1.23109549e-01
2.97110081e-01 1.04531229e+00 -5.71563184e-01 -1.16487920e-01
1.65799987e-02 2.56016362e-03 -2.73164362e-01 5.11507034e-01
7.43402064e-01 -5.96424043e-01 -4.54692960e-01 -4.96064067e-01
-4.03170824e-01 5.90698943e-02 -1.30555463e+00 1.57020223e+00
-5.66064477e-01 7.44807184e-01 -1.68775469e-02 -8.25091362e-01
6.52603567e-01 4.11470801e-01 1.70620039e-01 -4.35665965e-01
-5.95540814e-02 3.69086117e-01 -1.27689436e-01 -3.09169710e-01
4.48409885e-01 2.02300489e-01 3.55655849e-01 6.26832172e-02
8.46255645e-02 1.72360957e-01 2.16243982e-01 2.93350909e-02
1.05075037e+00 2.93645620e-01 7.59824440e-02 -3.01809043e-01
3.05389792e-01 -1.19466044e-01 2.53838599e-01 5.79956055e-01
8.29044357e-02 6.77325130e-01 2.72120804e-01 -6.90070391e-01
-1.09831405e+00 -8.91584158e-01 -4.06813353e-01 9.62526083e-01
-2.31730506e-01 -1.89299554e-01 -5.84730446e-01 -1.05150685e-01
1.74745679e-01 3.71356636e-01 -5.90480328e-01 6.41783103e-02
-7.29542613e-01 -7.60372877e-01 9.73717213e-01 9.72144306e-01
9.60662484e-01 -1.07921243e+00 -6.25521362e-01 3.21964473e-01
2.41443962e-01 -1.13731372e+00 -5.73639929e-01 2.08710939e-01
-1.03352058e+00 -7.51487672e-01 -9.81182575e-01 -9.45940733e-01
7.37586498e-01 2.13803276e-01 8.86805236e-01 2.52641946e-01
-5.61777949e-01 4.88242544e-02 -1.38598040e-01 -3.95725965e-01
-3.21336500e-02 1.25543505e-01 -1.34953139e-02 -8.88702199e-02
1.16222687e-01 -8.47685695e-01 -8.88816655e-01 1.93991467e-01
-1.14957082e+00 2.93515474e-01 9.27616656e-01 8.72641563e-01
3.34430397e-01 6.98578581e-02 4.69883800e-01 -8.81808281e-01
7.48371720e-01 -2.95842052e-01 -5.58632493e-01 4.31159511e-02
-5.19665480e-01 1.26676872e-01 8.08983088e-01 -8.33827972e-01
-1.00554991e+00 1.64489669e-03 -4.19210553e-01 -6.42388105e-01
-2.90299326e-01 4.66088146e-01 3.99613738e-01 -2.90214628e-01
7.83852041e-01 8.00585747e-02 1.28610775e-01 -5.83539367e-01
4.46696848e-01 4.49079126e-01 5.95126033e-01 -6.62762105e-01
7.61642396e-01 5.37668169e-01 -1.68077826e-01 -8.81254077e-01
-5.85117757e-01 -3.81406277e-01 -4.08202797e-01 1.96257010e-01
4.51700926e-01 -1.14812648e+00 -8.95950615e-01 7.37591922e-01
-1.03153801e+00 -5.33819854e-01 -4.11170423e-01 6.78378522e-01
-2.66507894e-01 2.10044399e-01 -9.98432755e-01 -5.16906559e-01
-5.71147442e-01 -1.05299616e+00 4.58531559e-01 4.32236671e-01
6.00715503e-02 -1.16599905e+00 -3.77334982e-01 -1.65054873e-01
9.70052183e-01 1.85554981e-01 6.15438759e-01 -2.99064934e-01
-5.68672359e-01 -2.57959515e-01 -6.71209931e-01 8.15963030e-01
1.18718773e-01 1.77246764e-01 -1.17783785e+00 -3.03775162e-01
-4.30518568e-01 -2.87919551e-01 9.65933204e-01 3.73969674e-01
1.42158318e+00 -3.82623464e-01 -1.44784495e-01 1.06961739e+00
1.59341478e+00 -2.55271375e-01 8.24772716e-01 3.31978589e-01
6.52130485e-01 -5.79312071e-02 -2.54298728e-02 4.17413622e-01
8.91331434e-02 2.46833906e-01 2.75809854e-01 -4.78675067e-01
-3.47448558e-01 -7.45882168e-02 1.66165978e-01 7.85003424e-01
-5.17848969e-01 7.42982030e-02 -6.44649923e-01 4.42174375e-01
-1.81913292e+00 -1.01200104e+00 -1.82101354e-01 1.99730504e+00
9.39795792e-01 2.02523261e-01 -5.99989630e-02 7.01474398e-03
4.77280408e-01 -8.87102410e-02 -3.82811278e-01 -2.63835758e-01
-8.61136764e-02 6.75575852e-01 8.39882553e-01 1.90051109e-01
-1.25973713e+00 8.93496633e-01 5.98473263e+00 1.00447989e+00
-1.34057045e+00 -3.87895815e-02 5.46130836e-01 -1.42744362e-01
1.05219752e-01 -1.37517825e-01 -8.19194317e-01 4.38737482e-01
7.92376280e-01 1.23216890e-01 5.22509336e-01 9.48095679e-01
1.45118430e-01 7.97696337e-02 -8.57335806e-01 9.56577718e-01
-4.25647050e-01 -1.72940564e+00 -1.69280007e-01 -3.31868697e-03
7.82474041e-01 3.36012572e-01 2.03893572e-01 5.01564682e-01
3.11111957e-01 -1.12477863e+00 5.85562348e-01 5.23168325e-01
9.40779746e-01 -8.43545973e-01 5.93797863e-01 2.38579020e-01
-1.33160377e+00 1.11688618e-02 -7.65750110e-01 -3.08331490e-01
-2.13394761e-01 7.09962726e-01 -6.47649586e-01 2.89603800e-01
8.58712316e-01 6.19992614e-01 -4.04004663e-01 1.38625884e+00
-1.18896045e-01 8.07476759e-01 -5.10561585e-01 -2.28360832e-01
5.16431630e-01 1.21851927e-02 8.79876222e-03 1.41822696e+00
3.72816622e-01 -4.94962595e-02 9.65666696e-02 1.01012611e+00
-3.36593360e-01 -8.39070138e-03 -2.58739740e-01 -1.31070865e-02
3.54049891e-01 1.46470428e+00 -5.31203508e-01 -2.89220572e-01
-8.10343862e-01 1.06997418e+00 5.52137733e-01 6.20677590e-01
-9.70352352e-01 -7.10325003e-01 9.76821601e-01 8.05224478e-02
5.39482236e-01 -4.02149796e-01 -3.38297814e-01 -1.02477968e+00
2.14301631e-01 -5.18968582e-01 1.14705056e-01 -5.75913072e-01
-1.32181013e+00 6.24610484e-01 -1.12474002e-01 -1.11421955e+00
2.75907516e-01 -8.29533994e-01 -8.43690276e-01 1.19566846e+00
-1.83895159e+00 -1.26016855e+00 -4.55848992e-01 6.97620332e-01
4.77418035e-01 1.10222176e-01 8.56233776e-01 5.91395140e-01
-4.06611860e-01 8.13646436e-01 2.88359702e-01 4.19327706e-01
5.13574243e-01 -1.06300664e+00 7.98777819e-01 6.79346800e-01
-1.27006173e-01 8.07414830e-01 2.97471046e-01 -2.72489816e-01
-1.34729970e+00 -1.19761348e+00 5.18516898e-01 -3.61921713e-02
8.40841949e-01 -2.53482103e-01 -1.20772517e+00 6.21719003e-01
6.43395931e-02 5.93477905e-01 4.59982097e-01 7.13444054e-02
-3.46515596e-01 -9.95538980e-02 -9.93696272e-01 8.24683666e-01
9.99138713e-01 -3.11548650e-01 -1.25230625e-01 7.26874396e-02
8.52474391e-01 -5.07278144e-01 -1.01784611e+00 3.60881776e-01
6.45556331e-01 -8.00482988e-01 1.23387194e+00 -3.86148185e-01
4.80467647e-01 -1.77260682e-01 9.59504247e-02 -1.24297488e+00
-6.83751047e-01 -6.61258519e-01 -3.30007821e-01 1.02399695e+00
4.24155027e-01 -1.04988933e+00 7.37131298e-01 7.44244874e-01
-3.02032083e-01 -1.15679669e+00 -6.90079331e-01 -7.81168997e-01
1.10763393e-01 -5.32876134e-01 6.12845898e-01 7.79220521e-01
-3.52121085e-01 -2.10999340e-01 -3.50877136e-01 2.99375713e-01
4.22251105e-01 -1.47929743e-01 5.93615174e-01 -1.06160736e+00
-4.13717538e-01 -5.95430613e-01 -4.82480764e-01 -1.26271665e+00
-1.15327686e-01 -8.29693556e-01 -1.86674774e-01 -1.32495177e+00
-1.51281670e-01 -7.12722540e-01 -5.66031933e-01 7.14128613e-01
-2.51482487e-01 5.04345536e-01 1.90386742e-01 2.78648794e-01
-2.05267757e-01 3.84235859e-01 1.38161170e+00 1.99429885e-01
-2.50520498e-01 1.96114480e-01 -6.55154705e-01 8.51520419e-01
1.08871841e+00 -7.73586780e-02 -3.20030093e-01 -4.76390183e-01
6.14454113e-02 -4.25892472e-01 6.65157795e-01 -1.40763283e+00
4.46115077e-01 -2.12932788e-02 6.71134591e-01 -3.50489557e-01
4.10843015e-01 -6.91443980e-01 2.24357992e-01 5.00954390e-01
-2.83925325e-01 2.07060371e-02 5.57406485e-01 3.65435988e-01
-1.41699627e-01 -1.81058779e-01 7.88956404e-01 -1.80486172e-01
-9.02062654e-01 6.11331165e-01 -9.41971466e-02 -1.49212986e-01
7.93378532e-01 -4.01156574e-01 -2.12472156e-01 3.72074805e-02
-5.85838258e-01 5.08995391e-02 2.77598172e-01 2.98064858e-01
7.07907081e-01 -1.46342385e+00 -7.61117458e-01 2.94659972e-01
-9.79452580e-02 9.05111581e-02 2.58599132e-01 9.26763415e-01
-7.41429806e-01 2.07714289e-01 -6.39847815e-02 -3.89345378e-01
-1.07234561e+00 3.47017556e-01 1.60726666e-01 -1.22207917e-01
-1.18529475e+00 1.05421293e+00 1.08167291e-01 -2.97891796e-01
4.53539699e-01 -7.96340585e-01 1.74240500e-01 -3.31726640e-01
7.06789076e-01 4.25180078e-01 1.07423358e-01 2.56761126e-02
-7.17628300e-02 2.12955073e-01 -1.57441020e-01 2.66250312e-01
1.61052215e+00 3.10423821e-01 -2.46080589e-02 1.42741442e-01
1.39661765e+00 -1.43272370e-01 -1.68182504e+00 -5.29086113e-01
-2.47802123e-01 -4.88288045e-01 1.62674174e-01 -6.25768244e-01
-1.15597749e+00 7.26523042e-01 5.12960672e-01 2.04582825e-01
1.15770006e+00 -2.36350805e-01 9.76410687e-01 5.81192613e-01
1.29510775e-01 -9.60708797e-01 -6.50967583e-02 7.18644917e-01
8.08382750e-01 -1.05780232e+00 1.23362159e-04 -4.27220613e-01
-4.26756799e-01 1.49633408e+00 4.76844221e-01 -4.65941936e-01
8.20570707e-01 5.55478930e-01 8.24338049e-02 3.68615389e-02
-6.71462595e-01 -1.48339882e-01 1.81491897e-01 3.39876682e-01
6.75852895e-01 1.24482416e-01 1.52439093e-02 1.85722217e-01
-1.30971298e-01 2.94027328e-01 3.06293428e-01 9.55485404e-01
-3.13847303e-01 -9.45357323e-01 -2.82346040e-01 7.11745381e-01
-5.97645104e-01 -5.13632894e-01 5.24603248e-01 8.32269669e-01
2.45296266e-02 4.13566440e-01 7.24240392e-02 -9.61836949e-02
4.19821888e-01 2.78478209e-02 3.46713781e-01 -2.50916243e-01
-7.10729241e-01 -1.75800368e-01 -1.25643238e-01 -4.59859937e-01
-3.02015811e-01 -2.09281921e-01 -1.35098219e+00 -6.23483956e-01
-3.97696525e-01 -1.75485298e-01 7.62587190e-01 5.60543418e-01
4.98821944e-01 7.40908980e-01 3.48332614e-01 -9.60506380e-01
-9.00879443e-01 -9.30930555e-01 -4.19916958e-01 4.64622468e-01
3.53208631e-01 -5.01678228e-01 -2.29243636e-01 6.66432455e-02] | [8.95788860321045, 2.344003677368164] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.