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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
80a5353d-822d-4d2d-b717-cf0f3fac2ab3 | pinpointing-why-object-recognition | 2304.05391 | null | https://arxiv.org/abs/2304.05391v1 | https://arxiv.org/pdf/2304.05391v1.pdf | Pinpointing Why Object Recognition Performance Degrades Across Income Levels and Geographies | Despite impressive advances in object-recognition, deep learning systems' performance degrades significantly across geographies and lower income levels raising pressing concerns of inequity. Addressing such performance gaps remains a challenge, as little is understood about why performance degrades across incomes or geographies. We take a step in this direction by annotating images from Dollar Street, a popular benchmark of geographically and economically diverse images, labeling each image with factors such as color, shape, and background. These annotations unlock a new granular view into how objects differ across incomes and regions. We then use these object differences to pinpoint model vulnerabilities across incomes and regions. We study a range of modern vision models, finding that performance disparities are most associated with differences in texture, occlusion, and images with darker lighting. We illustrate how insights from our factor labels can surface mitigations to improve models' performance disparities. As an example, we show that mitigating a model's vulnerability to texture can improve performance on the lower income level. We release all the factor annotations along with an interactive dashboard to facilitate research into more equitable vision systems. | ['Mark Ibrahim', 'Diane Bouchacourt', 'Caner Hazirbas', 'Melissa Hall', 'Megan Richards', 'Laura Gustafson'] | 2023-04-11 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [ 1.70170709e-01 -2.12381989e-01 -3.19976419e-01 -6.94169879e-01
-3.64790618e-01 -5.50282478e-01 5.46318889e-01 3.87903869e-01
-5.60901463e-01 3.51212710e-01 6.91248715e-01 -6.14786863e-01
-1.35209322e-01 -7.68851399e-01 -9.24988449e-01 -2.46970236e-01
1.35477901e-01 -1.04509190e-01 -3.41337055e-01 -2.11275786e-01
4.86791641e-01 3.00218254e-01 -1.67492437e+00 2.33843669e-01
1.21616161e+00 6.97166562e-01 -7.49329701e-02 5.94293833e-01
9.41472277e-02 4.85622585e-01 -5.93951583e-01 -8.23120952e-01
5.52559912e-01 -1.95932299e-01 -3.43065500e-01 5.52588841e-03
1.51221800e+00 -6.79256022e-01 -3.01541924e-01 1.19284892e+00
3.65684122e-01 -2.24107906e-01 6.59295559e-01 -1.26678622e+00
-1.37487602e+00 6.01631224e-01 -1.08613086e+00 2.75399595e-01
-9.49457958e-02 6.05400026e-01 8.66587222e-01 -3.64951223e-01
3.09309691e-01 1.56618512e+00 1.04983509e+00 8.88254493e-02
-1.40513444e+00 -7.75747299e-01 3.95671695e-01 2.70226836e-01
-1.22360849e+00 -5.95600307e-01 3.37194473e-01 -9.19668436e-01
8.33649576e-01 3.66708875e-01 7.51751184e-01 6.87835395e-01
5.63495867e-02 3.31319600e-01 1.26003158e+00 -2.47614846e-01
-1.03581108e-01 5.75041398e-02 3.50146443e-01 7.88964391e-01
8.02765310e-01 6.89201578e-02 -7.64951825e-01 -2.09781602e-02
8.26505423e-01 4.49179336e-02 -3.42868790e-02 -2.01076612e-01
-7.85429060e-01 6.85657024e-01 6.37127697e-01 -3.55617464e-01
-1.25605632e-02 4.13766891e-01 3.02419811e-02 9.85708088e-02
6.62250340e-01 6.36797249e-01 -3.90804470e-01 -4.42491882e-02
-6.28622949e-01 4.33413446e-01 3.34010363e-01 3.36807549e-01
8.47803354e-01 1.26356736e-01 -1.27938688e-01 8.73706520e-01
9.91602149e-03 9.01812077e-01 -2.58304447e-01 -1.41741121e+00
6.11583352e-01 4.95397002e-01 3.57924163e-01 -1.63294685e+00
-5.54484904e-01 -4.73668844e-01 -3.56605709e-01 6.90458000e-01
9.43964779e-01 -2.14586228e-01 -8.61322105e-01 1.82809877e+00
-1.11515395e-01 -1.12904996e-01 -4.83500183e-01 8.24152112e-01
3.32114518e-01 1.81122452e-01 4.56649363e-01 4.48286116e-01
1.33475721e+00 -6.26736343e-01 -1.53456479e-01 -4.44482923e-01
3.80623341e-01 -7.66363144e-01 1.24356782e+00 2.53794670e-01
-9.37946022e-01 -6.46385252e-01 -8.35226417e-01 -2.68077821e-01
-4.71988827e-01 2.23268997e-02 8.77559125e-01 1.25009310e+00
-1.05520988e+00 5.67565739e-01 -7.38306999e-01 -3.90086859e-01
8.73874128e-01 2.21086577e-01 -5.07922843e-02 2.02322323e-02
-6.30002141e-01 9.72571790e-01 -3.45313787e-01 -2.18630984e-01
-5.89866221e-01 -1.25149524e+00 -9.97895241e-01 1.96467876e-01
4.83312868e-02 -5.99711299e-01 8.41241419e-01 -1.38440740e+00
-6.66212678e-01 9.23914254e-01 1.36429316e-03 -4.28366959e-01
4.62511897e-01 -3.67102802e-01 -2.85904080e-01 -2.51932293e-01
2.33906969e-01 8.09913516e-01 6.05975866e-01 -1.20143580e+00
-8.36716771e-01 -7.72865117e-01 2.38610685e-01 2.29816586e-01
-5.79873800e-01 4.13726538e-01 -1.08229615e-01 -5.42780221e-01
-4.01071578e-01 -8.02698731e-01 -1.98167473e-01 2.29042724e-01
-3.35943669e-01 2.23119825e-01 2.68808722e-01 -1.04097152e+00
1.02920890e+00 -1.95137310e+00 -2.99516767e-01 1.12249792e-01
4.81278837e-01 6.86465278e-02 -2.37744525e-01 -2.94328004e-01
5.43649755e-02 5.08586943e-01 9.08440277e-02 -6.96499199e-02
1.83078468e-01 -1.45838276e-01 -2.90708542e-02 5.74647665e-01
5.55140913e-01 8.48633766e-01 -6.17998481e-01 -1.03243858e-01
3.84396970e-01 5.29610515e-01 -8.92309129e-01 -5.06106913e-01
1.28647760e-01 3.22049141e-01 -1.27531052e-01 7.61168778e-01
9.79157746e-01 -1.10598959e-01 7.76331965e-03 -9.51229557e-02
-4.66524482e-01 -1.12394787e-01 -7.10842848e-01 1.10174382e+00
-2.68664747e-01 9.83032703e-01 5.27743697e-02 -7.98625708e-01
6.48977637e-01 -6.38060272e-01 1.96783468e-01 -1.22116518e+00
-1.60207935e-02 -4.18537892e-02 3.77419919e-01 -5.78676164e-01
7.79437721e-01 2.36949027e-01 2.81945569e-03 2.74272084e-01
-6.16294086e-01 6.27001673e-02 -2.08168104e-01 -6.44806921e-02
9.07263875e-01 -4.08750959e-02 -2.32719451e-01 -5.31876683e-01
-3.01175505e-01 1.57022789e-01 7.01788843e-01 9.36821699e-01
-5.27732313e-01 7.59921908e-01 6.63045704e-01 -6.82067394e-01
-1.55164862e+00 -1.34186375e+00 -3.04670364e-01 1.46864724e+00
8.88489708e-02 -5.58488332e-02 -8.01330030e-01 -2.85255313e-01
7.00402379e-01 3.83466244e-01 -9.21566069e-01 -1.66369140e-01
-3.13759476e-01 -1.03304648e+00 5.96926749e-01 7.59461641e-01
6.23754799e-01 -4.32037383e-01 -8.65779519e-01 -4.97703940e-01
2.00598180e-01 -7.42465317e-01 -2.86972314e-01 -4.01655525e-01
-4.44396228e-01 -1.12676084e+00 -8.57323349e-01 -3.09963524e-01
6.97375536e-01 4.83427882e-01 1.50842309e+00 3.41655612e-01
-5.95561922e-01 3.62164527e-01 -5.47551811e-02 -8.43220174e-01
5.88452965e-02 -2.45947894e-02 7.11826459e-02 -1.71195507e-01
3.99416536e-01 -2.16091141e-01 -1.04577661e+00 3.32845300e-01
-5.01270473e-01 2.37247273e-01 1.75750509e-01 4.78003293e-01
1.66169614e-01 -1.78946629e-01 4.99956280e-01 -9.32065427e-01
4.14581567e-01 -7.65554249e-01 -8.52870762e-01 3.24857980e-01
-7.64168918e-01 -1.61941975e-01 1.82895228e-01 -2.36472920e-01
-1.21781206e+00 -1.16605341e-01 3.28234553e-01 1.37643293e-01
-1.85899779e-01 1.68018535e-01 1.01967134e-01 -1.48011535e-01
9.19727623e-01 -6.18377268e-01 -1.55164450e-01 -5.25621235e-01
3.86134148e-01 4.53795910e-01 7.35675097e-01 -8.57427001e-01
6.61368728e-01 6.73973620e-01 -2.19098270e-01 -8.30395520e-01
-7.07904637e-01 -1.07828500e-02 -3.08028907e-01 -2.68953323e-01
9.21806157e-01 -1.02740872e+00 -6.78046405e-01 6.65826619e-01
-6.77060962e-01 -6.65245712e-01 1.24016115e-02 4.65005755e-01
-1.90959573e-01 -7.48481527e-02 -3.36626321e-01 -7.63935268e-01
1.46772608e-01 -1.07641625e+00 9.15348470e-01 7.58309007e-01
-2.45187938e-01 -8.86602581e-01 -2.13630661e-01 7.77566016e-01
6.77545667e-01 3.95620584e-01 1.21999097e+00 -1.31637886e-01
-4.57381785e-01 4.40480113e-01 -1.01606309e+00 3.06002617e-01
2.87384726e-02 2.49094889e-01 -1.02109957e+00 6.33991584e-02
-6.47998214e-01 -2.53320336e-01 1.07743406e+00 7.99365282e-01
1.53844583e+00 -1.28214329e-01 -6.80262595e-02 9.74329293e-01
1.45237339e+00 8.36016759e-02 5.32185555e-01 5.84861159e-01
1.00342882e+00 9.17153955e-01 1.79041326e-01 3.99059504e-01
7.27935910e-01 4.17691559e-01 3.53078961e-01 -6.13322854e-01
-3.89668584e-01 -7.05740452e-02 5.67641519e-02 -9.61043015e-02
-2.56190717e-01 1.98488310e-01 -1.40105021e+00 7.46643066e-01
-1.61500347e+00 -9.18513715e-01 -3.04795414e-01 2.21674633e+00
5.57214200e-01 6.90519363e-02 2.67565906e-01 -3.93364668e-01
6.99257493e-01 1.60292387e-01 -1.14707804e+00 -4.77330953e-01
-3.59349579e-01 1.56687379e-01 1.07907832e+00 6.32325053e-01
-9.07936811e-01 7.87850142e-01 7.63932610e+00 4.40246493e-01
-1.07606840e+00 -2.14556068e-01 1.62455833e+00 -4.32960957e-01
-6.46603703e-01 -2.11936399e-01 -7.09353864e-01 5.43052137e-01
7.63518810e-01 -7.41892904e-02 6.09674752e-01 6.67639911e-01
4.98119891e-01 -4.50907558e-01 -8.45684230e-01 8.98759305e-01
1.95298821e-01 -1.40380800e+00 -3.19495082e-01 2.02475190e-01
1.11778462e+00 1.39211565e-01 8.31593454e-01 1.03407718e-01
8.73024821e-01 -1.58777845e+00 8.52504909e-01 5.81428647e-01
1.04072654e+00 -8.72796893e-01 2.22621560e-01 -3.70987803e-01
-8.84501994e-01 -3.76928002e-01 -4.96289641e-01 -7.86076069e-01
-3.06431174e-01 3.61663967e-01 -3.48899126e-01 -1.65354252e-01
1.15365970e+00 3.69102448e-01 -1.16973853e+00 9.57502306e-01
2.75809169e-01 5.35109043e-01 -2.67196596e-01 5.28859854e-01
-5.57623729e-02 -1.45655230e-01 -1.40241295e-01 1.05245340e+00
2.34947398e-01 -3.38481888e-02 9.60235577e-03 1.00672495e+00
-1.29611492e-01 -1.73697710e-01 -6.67329550e-01 5.95258027e-02
5.91696918e-01 8.35599422e-01 -5.64821303e-01 -2.58323073e-01
-7.92350709e-01 5.72794020e-01 3.84744674e-01 5.71772635e-01
-9.24781144e-01 -2.26123258e-01 1.33766401e+00 2.80893087e-01
-2.52257645e-01 -3.03428352e-01 -1.01634741e+00 -1.03344381e+00
-1.66012064e-01 -8.72545838e-01 8.22590292e-02 -8.46319377e-01
-1.26707506e+00 -3.19411516e-01 3.26102525e-02 -3.00151378e-01
4.05487955e-01 -6.99335337e-01 -3.78788352e-01 1.08102894e+00
-1.61859310e+00 -1.11745918e+00 -4.93424863e-01 3.73963147e-01
2.91016251e-01 -4.85033467e-02 2.54270107e-01 3.99451345e-01
-8.27335775e-01 7.62779355e-01 2.46855423e-01 2.33751372e-01
9.05902326e-01 -1.25061309e+00 9.33893681e-01 7.96129584e-01
-7.63836950e-02 5.98348379e-01 6.74170792e-01 -6.89846575e-01
-1.10781932e+00 -8.79958987e-01 3.47040623e-01 -8.23135734e-01
6.51237547e-01 -9.93830189e-02 -6.51597798e-01 7.60997415e-01
1.28582790e-01 -2.64240474e-01 7.55647123e-01 6.22156143e-01
-8.29575419e-01 -2.74602175e-01 -1.22542357e+00 7.83911943e-01
1.35134304e+00 -5.33496082e-01 -7.28494078e-02 9.30876285e-02
6.43230855e-01 -9.46220607e-02 -7.89900720e-01 1.63824037e-01
9.94428873e-01 -1.37156093e+00 1.20372367e+00 -6.71674252e-01
8.00891817e-01 1.10647611e-01 -4.14953530e-01 -1.40743041e+00
-3.93411547e-01 -2.88771510e-01 5.16156137e-01 1.31317866e+00
4.30624247e-01 -7.24346101e-01 9.03124392e-01 1.27109492e+00
-5.88480532e-02 -4.79996592e-01 -3.93676370e-01 -5.09778380e-01
5.08976758e-01 -4.21507567e-01 1.00156248e+00 1.08167851e+00
-3.51110131e-01 -2.43604392e-01 -2.44217113e-01 2.34199554e-01
6.16480827e-01 -1.86477333e-01 8.86180699e-01 -1.11607671e+00
-8.37359428e-02 -7.98064113e-01 -4.25338477e-01 -4.68728572e-01
-1.47096803e-02 -4.72763062e-01 -3.78716171e-01 -1.23800993e+00
7.25416541e-01 -8.96716833e-01 -2.31007084e-01 5.83698750e-01
-4.16187406e-01 6.63801670e-01 5.53784013e-01 -2.24496368e-02
-3.30518156e-01 -4.96839546e-02 1.01864243e+00 -2.83705890e-01
7.15802088e-02 -5.91112077e-01 -1.44848192e+00 9.85389113e-01
8.80704522e-01 -8.51593018e-02 -2.74386942e-01 -1.32610631e+00
5.62751174e-01 -5.41604817e-01 5.95698357e-01 -1.09328330e+00
-1.79600388e-01 -7.54698634e-01 9.70507503e-01 -1.52563885e-01
1.64285585e-01 -5.25726199e-01 -1.40874475e-01 4.84821200e-01
-3.76214415e-01 3.38916719e-01 3.16505611e-01 2.28568867e-01
4.76224095e-01 3.56378824e-01 8.58905792e-01 1.82598494e-02
-7.16817021e-01 1.64394498e-01 2.77828220e-02 1.62821382e-01
8.11950803e-01 -2.45968685e-01 -1.15161574e+00 -3.21199298e-01
-2.34842137e-01 3.46227616e-01 1.19485354e+00 4.12497342e-01
1.51641518e-01 -1.19590676e+00 -9.04451907e-01 3.00520003e-01
6.00367412e-02 -4.21628863e-01 3.60539973e-01 5.31310678e-01
-7.20225155e-01 2.39434130e-02 -7.43516326e-01 -4.29743648e-01
-1.19468760e+00 2.48516217e-01 4.92182255e-01 3.42544496e-01
-2.15896070e-01 9.81465042e-01 5.12290835e-01 -1.88390657e-01
9.75958630e-02 -3.46638918e-01 -5.70759736e-03 6.04289547e-02
7.19855845e-01 8.44325781e-01 -3.00514996e-01 -6.68942988e-01
-1.51633665e-01 9.13427770e-01 9.84142944e-02 6.13723025e-02
1.28007340e+00 -3.38966280e-01 1.32982820e-01 1.05523489e-01
9.07062709e-01 2.06158251e-01 -1.86170542e+00 2.01302841e-01
-1.48659438e-01 -1.15215409e+00 -5.51040992e-02 -1.02521229e+00
-1.05480778e+00 7.36762762e-01 1.05977726e+00 2.32212722e-01
1.07746732e+00 -1.28492534e-01 6.13213599e-01 1.82429329e-03
1.83872297e-01 -1.16022813e+00 4.48657833e-02 2.88575321e-01
5.53214252e-01 -1.59494460e+00 2.98584640e-01 -1.07178599e-01
-5.34108400e-01 5.05864203e-01 7.05949426e-01 -1.07895494e-01
5.45890450e-01 2.27812037e-01 2.22693026e-01 -1.40509233e-01
-3.06167901e-01 -3.46724838e-01 1.93620175e-01 1.16298151e+00
4.73692268e-01 5.97344100e-01 -7.44595751e-02 2.55988300e-01
-6.56611145e-01 -3.46853703e-01 4.28641796e-01 3.72303367e-01
-7.39609361e-01 -7.21456766e-01 -9.42464769e-01 8.76237452e-01
-6.71264946e-01 -1.01089701e-01 -4.44550037e-01 6.68885112e-01
6.51192963e-01 9.64054525e-01 7.27232516e-01 -3.98459285e-01
1.25087991e-01 -2.23936841e-01 5.31585574e-01 -5.78560591e-01
-4.83951718e-01 -3.59729111e-01 3.35972130e-01 -6.20952368e-01
-1.17580211e-02 -7.37507820e-01 -7.70038366e-01 -1.03475916e+00
1.30138636e-01 -7.27856815e-01 9.06168461e-01 5.93033433e-01
3.89492840e-01 4.19475406e-01 4.12967563e-01 -8.82569611e-01
-2.15061322e-01 -4.44309860e-01 -4.66832042e-01 7.74914086e-01
5.21960497e-01 -6.01391912e-01 -1.51162967e-01 1.56432450e-01] | [12.897370338439941, 1.2204035520553589] |
8ca75246-951c-4ffa-bf20-593e7c6dd999 | learning-individual-speaking-styles-for | 2005.08209 | null | https://arxiv.org/abs/2005.08209v1 | https://arxiv.org/pdf/2005.08209v1.pdf | Learning Individual Speaking Styles for Accurate Lip to Speech Synthesis | Humans involuntarily tend to infer parts of the conversation from lip movements when the speech is absent or corrupted by external noise. In this work, we explore the task of lip to speech synthesis, i.e., learning to generate natural speech given only the lip movements of a speaker. Acknowledging the importance of contextual and speaker-specific cues for accurate lip-reading, we take a different path from existing works. We focus on learning accurate lip sequences to speech mappings for individual speakers in unconstrained, large vocabulary settings. To this end, we collect and release a large-scale benchmark dataset, the first of its kind, specifically to train and evaluate the single-speaker lip to speech task in natural settings. We propose a novel approach with key design choices to achieve accurate, natural lip to speech synthesis in such unconstrained scenarios for the first time. Extensive evaluation using quantitative, qualitative metrics and human evaluation shows that our method is four times more intelligible than previous works in this space. Please check out our demo video for a quick overview of the paper, method, and qualitative results. https://www.youtube.com/watch?v=HziA-jmlk_4&feature=youtu.be | ['C. V. Jawahar', 'Vinay Namboodiri', 'Rudrabha Mukhopadhyay', 'K R Prajwal'] | 2020-05-17 | learning-individual-speaking-styles-for-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Prajwal_Learning_Individual_Speaking_Styles_for_Accurate_Lip_to_Speech_Synthesis_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Prajwal_Learning_Individual_Speaking_Styles_for_Accurate_Lip_to_Speech_Synthesis_CVPR_2020_paper.pdf | cvpr-2020-6 | ['speaker-specific-lip-to-speech-synthesis', 'lip-to-speech-synthesis'] | ['computer-vision', 'computer-vision'] | [ 1.88674718e-01 3.85155380e-01 -4.36846524e-01 -2.10711986e-01
-1.09214270e+00 -5.01209438e-01 5.68858087e-01 -5.65475464e-01
7.49507546e-02 7.91238427e-01 6.84047103e-01 -3.34714085e-01
4.27520603e-01 -2.21782811e-02 -5.69858551e-01 -4.83793855e-01
4.70868409e-01 2.42681086e-01 -9.11151916e-02 -7.88352117e-02
1.86365917e-01 3.17984313e-01 -1.85071826e+00 2.31750742e-01
6.20467544e-01 8.22767437e-01 3.74323785e-01 9.09548700e-01
-1.51015565e-01 6.45995438e-01 -5.53393841e-01 -3.75932962e-01
4.21208106e-02 -6.36188030e-01 -8.61772656e-01 1.56893760e-01
4.61858064e-01 -4.11603689e-01 -2.82165855e-01 1.08862102e+00
1.11807942e+00 2.95154508e-02 5.08512676e-01 -1.53072584e+00
-5.27514815e-01 5.98773718e-01 -1.30234435e-01 -1.55990347e-01
7.99226522e-01 6.86903238e-01 9.26110625e-01 -1.04410923e+00
6.00369394e-01 1.37914944e+00 3.97493184e-01 1.24574769e+00
-1.01356912e+00 -7.95092285e-01 1.68845523e-02 -1.46208322e-02
-1.39693594e+00 -1.69273245e+00 7.96863198e-01 -3.48544478e-01
7.05958605e-01 4.62891489e-01 3.81436735e-01 1.63466763e+00
-3.81987274e-01 1.10688877e+00 8.64519894e-01 -4.89107788e-01
6.15724511e-02 2.23764002e-01 -3.97737443e-01 2.72493958e-01
-2.11317241e-01 1.40587524e-01 -9.14558411e-01 1.47247195e-01
3.34310383e-01 -5.06200433e-01 -7.21853673e-01 -1.23280622e-01
-1.46493363e+00 5.11668980e-01 -1.58335313e-01 1.14237368e-01
-2.00952843e-01 5.41406460e-02 3.80142659e-01 1.83441535e-01
3.72729421e-01 1.69095457e-01 -2.90363759e-01 -5.25411427e-01
-9.61004376e-01 2.96904236e-01 9.16662276e-01 1.22192287e+00
2.46610269e-01 2.60408968e-01 -2.68429637e-01 1.04042768e+00
4.42600369e-01 6.88737690e-01 7.03040361e-01 -1.05541432e+00
4.78626788e-01 -1.02172591e-01 1.72776118e-01 -3.66248101e-01
-1.49224028e-01 1.91923827e-01 -5.44268131e-01 1.95035651e-01
4.85939890e-01 -3.94296795e-01 -6.64247930e-01 2.00573587e+00
3.14775407e-01 5.13871573e-02 2.35714301e-01 7.32842743e-01
1.14304936e+00 5.43931127e-01 -2.49013435e-02 -6.00197136e-01
1.21551740e+00 -1.11704576e+00 -1.24374068e+00 -2.58679062e-01
2.91084051e-01 -1.09093475e+00 1.49690425e+00 2.01478928e-01
-1.42173350e+00 -5.04566491e-01 -6.95411325e-01 -1.41317427e-01
-1.22150153e-01 2.48377055e-01 2.95392334e-01 6.08812749e-01
-1.38523233e+00 1.73673823e-01 -4.95573848e-01 -5.05394459e-01
3.52354556e-01 7.91291893e-02 -3.08596939e-01 2.19746619e-01
-1.15142572e+00 5.83046734e-01 -1.02945626e-01 -9.08459350e-02
-7.88389504e-01 -5.65223038e-01 -1.08899188e+00 -2.99732447e-01
4.34337437e-01 -4.37889755e-01 1.92429876e+00 -1.00168359e+00
-2.18811440e+00 8.84589791e-01 -7.83303797e-01 -3.03924710e-01
7.64056325e-01 -1.02711678e-01 -3.30403686e-01 1.75485894e-01
-8.97773355e-02 1.05130970e+00 1.05564260e+00 -1.39689016e+00
-4.53893661e-01 -6.32912666e-02 -9.68165174e-02 2.69695520e-01
-1.08861834e-01 3.13840508e-01 -4.58830476e-01 -6.34461224e-01
-3.10489029e-01 -8.81445825e-01 4.43882525e-01 2.64722884e-01
-4.99987692e-01 -5.06241441e-01 7.43134737e-01 -6.83913708e-01
1.16558456e+00 -2.18185019e+00 -1.03004403e-01 -4.98940766e-01
6.92031309e-02 3.16472828e-01 -3.18716504e-02 4.60387558e-01
4.08336753e-03 3.69570315e-01 -1.43229038e-01 -9.28391933e-01
2.50390172e-01 -1.65927425e-01 -4.47882056e-01 3.82860184e-01
-7.19426200e-02 1.17405581e+00 -8.24682534e-01 -7.15204895e-01
2.60085911e-01 6.58014834e-01 -3.56012225e-01 4.43666071e-01
-1.90894827e-01 4.80330974e-01 -9.46480855e-02 7.83284068e-01
5.00379801e-01 1.19262405e-01 -9.65436101e-02 -4.82380614e-02
-2.01686859e-01 5.48491538e-01 -9.94023502e-01 1.67013991e+00
-7.33686686e-01 1.14677215e+00 5.94298482e-01 -5.08097351e-01
6.82737648e-01 9.47776020e-01 2.77543306e-01 -5.16647518e-01
1.50011867e-01 3.15852135e-01 -1.75162449e-01 -8.35523427e-01
2.10360155e-01 -2.42536545e-01 1.11509480e-01 3.26090127e-01
-2.70428061e-02 -4.66837883e-01 -7.16852583e-03 -1.29045054e-01
5.19012332e-01 6.48406520e-02 4.74203944e-01 -1.45837694e-01
6.41856134e-01 -6.68193817e-01 1.85504332e-01 3.23473632e-01
-6.60685718e-01 7.37972081e-01 3.73677760e-01 1.58915520e-01
-7.72605419e-01 -1.04426062e+00 -1.35034278e-01 1.10732996e+00
-2.02079508e-02 -4.32859719e-01 -1.18721151e+00 -3.41425866e-01
-2.61408538e-01 8.68944585e-01 -3.49759191e-01 1.91095442e-01
-4.68222916e-01 5.85401766e-02 6.39856458e-01 3.25590342e-01
2.19990820e-01 -1.52829671e+00 -3.33276808e-01 -1.15825586e-01
-6.38092518e-01 -1.34634352e+00 -9.72005844e-01 -4.84951913e-01
-2.36489609e-01 -7.15169489e-01 -9.72744584e-01 -8.29866827e-01
3.63534242e-01 9.90171283e-02 8.06084156e-01 -5.70077039e-02
-1.24874614e-01 4.27039534e-01 -9.92847979e-02 -6.27168477e-01
-8.18482220e-01 -6.37173578e-02 4.92736220e-01 1.61900505e-01
2.39625826e-01 -3.69557709e-01 -5.14871359e-01 3.45255882e-01
-6.04101181e-01 3.47263843e-01 3.03063482e-01 6.16942465e-01
3.02217782e-01 -4.17529076e-01 8.43823195e-01 -1.45217434e-01
9.10602689e-01 -2.78957874e-01 -3.23355973e-01 1.32142380e-02
-2.72711158e-01 -2.40995124e-01 4.74811435e-01 -6.06276453e-01
-8.12805295e-01 3.73444036e-02 -6.07492745e-01 -5.35786152e-01
-5.70091963e-01 -2.73462534e-01 -6.47441566e-01 2.49776304e-01
5.28984666e-01 4.32784349e-01 3.77225697e-01 -5.05720258e-01
3.62956941e-01 1.46304178e+00 5.46686590e-01 -3.74893010e-01
5.09465754e-01 4.18873698e-01 -4.58215147e-01 -1.25768995e+00
-4.04321045e-01 -3.00224721e-01 -5.48082471e-01 -4.77297813e-01
7.48540640e-01 -8.86314392e-01 -1.17784679e+00 6.82302058e-01
-1.29344225e+00 -8.78620923e-01 -4.17169541e-01 5.00104368e-01
-1.07325971e+00 3.27183634e-01 -1.40608802e-01 -1.06595862e+00
-4.43871528e-01 -1.44598556e+00 1.36946201e+00 1.55852094e-01
-4.74514931e-01 -7.81037092e-01 -5.79881407e-02 6.44635499e-01
5.34697115e-01 -7.39171430e-02 1.62010059e-01 -5.39603472e-01
-2.78930724e-01 1.34418234e-01 2.51724455e-03 4.74016815e-01
5.33111334e-01 8.20839629e-02 -1.30797780e+00 -1.80016160e-01
-1.42969802e-01 -4.68333155e-01 5.64567804e-01 4.37877029e-01
1.02898824e+00 -7.89757311e-01 -2.63602287e-01 5.05835295e-01
8.75226140e-01 6.32661581e-02 4.62284952e-01 -2.50045091e-01
4.80786353e-01 9.37057137e-01 4.68595147e-01 3.26851517e-01
4.71878231e-01 1.03597462e+00 2.41661981e-01 9.90539417e-02
-8.63649368e-01 -5.60943127e-01 7.31969893e-01 7.19798207e-01
2.24915162e-01 -6.34470403e-01 -8.50688815e-01 8.08329523e-01
-1.39155412e+00 -1.01758099e+00 2.94177353e-01 2.16361213e+00
1.19228780e+00 -3.95915806e-02 4.23805743e-01 2.70689040e-01
8.80534291e-01 3.23557347e-01 -4.38539505e-01 -4.10186768e-01
2.24495102e-02 -7.45391399e-02 4.82351631e-02 1.16661799e+00
-7.85974920e-01 1.16869128e+00 5.85597754e+00 7.71987021e-01
-1.55592370e+00 1.22895330e-01 4.85683590e-01 -4.55099523e-01
-3.17841202e-01 -3.88722628e-01 -8.74348283e-01 5.56672037e-01
1.16992414e+00 -3.66747975e-01 6.54995084e-01 5.51101565e-01
7.80417740e-01 1.46870703e-01 -1.21009445e+00 1.19518721e+00
2.65046865e-01 -1.12634993e+00 -6.21439191e-03 3.19000799e-03
3.64532650e-01 3.66408452e-02 2.26368248e-01 3.03656552e-02
-9.57276523e-02 -1.25078678e+00 1.11355793e+00 5.26116550e-01
1.39239299e+00 -3.45175833e-01 1.62454948e-01 4.47488904e-01
-1.15872753e+00 9.21952501e-02 1.47461608e-01 9.42184553e-02
4.69088793e-01 -5.20510075e-04 -9.08519924e-01 2.54473034e-02
4.40073520e-01 4.64242548e-01 -6.50279298e-02 7.85870850e-01
-3.74820143e-01 6.19141877e-01 -2.59186715e-01 -2.18751162e-01
-1.57991037e-01 4.50225085e-01 9.55659389e-01 1.24973524e+00
3.62396181e-01 -3.92872654e-02 -2.60437965e-01 9.37080145e-01
-2.87694871e-01 3.41301858e-01 -8.78557146e-01 -1.03133572e-02
5.69689691e-01 1.00871313e+00 -1.47669420e-01 -1.66922286e-01
-2.54390419e-01 8.05277228e-01 -1.07639141e-01 4.29709584e-01
-6.32781863e-01 -2.88355291e-01 1.01176512e+00 4.73956883e-01
9.45759267e-02 -1.20777100e-01 -9.14359167e-02 -9.95968103e-01
2.60069996e-01 -1.16765034e+00 -3.19748223e-01 -1.00782788e+00
-6.90096259e-01 8.51018250e-01 2.80309450e-02 -1.08738065e+00
-7.75729895e-01 -4.72758591e-01 -6.49954200e-01 9.04460311e-01
-1.72352850e+00 -1.19430840e+00 -2.86028504e-01 6.93313181e-01
1.27859104e+00 -2.29346946e-01 7.87391484e-01 2.70816498e-02
-3.54415715e-01 9.34378207e-01 -2.25108728e-01 1.04599990e-01
9.71004605e-01 -8.85340869e-01 5.75373769e-01 7.16255248e-01
1.92202404e-02 4.76687223e-01 8.76944482e-01 -4.24232185e-01
-1.29393363e+00 -8.23049009e-01 1.30583370e+00 -4.53863770e-01
5.04814863e-01 -7.92881727e-01 -5.61131060e-01 4.88288939e-01
5.02826810e-01 7.38500478e-03 5.85532546e-01 -4.70087230e-01
-8.83909836e-02 -2.78441142e-02 -1.27074432e+00 9.40963030e-01
1.21681702e+00 -6.50727451e-01 -4.24760312e-01 3.02159101e-01
1.05143273e+00 -4.08215255e-01 -5.53707838e-01 3.58890325e-01
8.02289307e-01 -8.43165338e-01 8.47340763e-01 -3.89826983e-01
2.28934258e-01 -1.08318776e-01 -1.81431562e-01 -1.15945292e+00
3.85967404e-01 -1.51482713e+00 -1.29659772e-01 1.56145835e+00
4.17915374e-01 -7.64019489e-01 5.24998903e-01 3.40443820e-01
-1.24636978e-01 -8.05153430e-01 -1.06149852e+00 -6.46362066e-01
2.06534848e-01 -6.19355500e-01 7.51678050e-01 4.14677590e-01
1.98889554e-01 2.30834216e-01 -4.28131193e-01 4.19311523e-02
6.12784207e-01 -1.51064470e-01 9.60190058e-01 -8.71332347e-01
1.01330355e-02 -6.97979569e-01 -2.61438061e-02 -1.31785250e+00
5.57227433e-01 -6.55422568e-01 3.61700863e-01 -1.43622029e+00
-2.16361538e-01 -4.55441475e-02 3.46986830e-01 4.38475013e-01
-3.70412506e-02 1.85350761e-01 3.49167734e-01 1.44615486e-01
-1.95635110e-01 5.90797901e-01 1.43134248e+00 -1.17256656e-01
-2.94728726e-01 4.47223395e-01 -8.18675160e-01 6.35795474e-01
1.03626847e+00 -6.23545721e-02 -6.06939912e-01 -2.72747487e-01
-1.68021291e-01 2.31120959e-01 3.36072356e-01 -8.18652272e-01
3.73282164e-01 -2.57722437e-01 -3.48222494e-01 -3.11975360e-01
7.44752526e-01 -5.11632740e-01 -2.94956297e-01 1.64458841e-01
-5.62079549e-01 -2.13007063e-01 4.20830637e-01 2.18534797e-01
-1.76052615e-01 -1.14639118e-01 9.12872136e-01 2.08349153e-02
-4.95545298e-01 2.69827157e-01 -3.56887311e-01 4.37972516e-01
8.20385814e-01 -3.13734978e-01 -1.46179646e-01 -1.01507854e+00
-6.69917583e-01 5.06215356e-02 4.58198309e-01 7.35017657e-01
6.78629458e-01 -1.30013335e+00 -9.98744428e-01 3.58152062e-01
1.51047915e-01 -9.65171382e-02 2.08907694e-01 8.20479512e-01
-9.60721299e-02 7.39158928e-01 -3.27655450e-02 -6.01428330e-01
-1.57934558e+00 4.75853592e-01 5.44693351e-01 3.82908583e-01
-4.36002374e-01 8.42478991e-01 1.54684350e-01 -2.38195166e-01
6.42483115e-01 -3.66757125e-01 -6.75923154e-02 8.45674518e-03
6.25577509e-01 1.72844708e-01 -1.72991008e-01 -1.18194449e+00
-3.66876304e-01 5.81262767e-01 2.81899333e-01 -4.72547889e-01
8.58592629e-01 -4.94914591e-01 2.49814212e-01 5.64590096e-01
1.29790890e+00 6.46436572e-01 -1.24863315e+00 -2.35485807e-02
-4.75984275e-01 -4.21380311e-01 -2.52573222e-01 -7.37361670e-01
-7.57098913e-01 1.11566305e+00 3.15658897e-01 1.86661690e-01
1.08388758e+00 2.62475252e-01 8.53175938e-01 1.25587313e-02
8.38908702e-02 -9.15315330e-01 9.59672630e-02 2.33944029e-01
1.34734118e+00 -1.37614357e+00 -4.50042427e-01 -3.51391822e-01
-8.97339225e-01 8.30422461e-01 3.40420395e-01 5.65912604e-01
6.04151845e-01 5.11970818e-01 3.42469215e-01 3.42533052e-01
-8.42424929e-01 -3.21159065e-01 2.70121425e-01 8.31486940e-01
4.96628046e-01 1.44315138e-01 4.79512922e-02 3.99910957e-01
-7.65038908e-01 1.12106167e-01 3.69022071e-01 4.38273460e-01
-3.24522376e-01 -1.03760350e+00 -2.66295046e-01 -5.11452742e-02
-5.18792629e-01 -2.15848327e-01 -5.88546515e-01 6.62517905e-01
-9.39941704e-02 1.30700648e+00 -7.62827173e-02 -1.12701170e-01
2.63875872e-01 4.28389221e-01 2.76693135e-01 -4.78386164e-01
-1.60295710e-01 8.51300210e-02 1.06628381e-01 -5.65588355e-01
-2.80359894e-01 -9.17040706e-01 -1.08627439e+00 -2.42134035e-01
-8.23683217e-02 -1.57368049e-01 9.71719444e-01 6.58028245e-01
4.17058647e-01 3.04740101e-01 6.35930598e-01 -1.07446039e+00
-6.22276664e-01 -1.16094375e+00 -1.82329342e-01 3.01031530e-01
1.00873828e+00 -5.38929701e-01 -8.80110979e-01 2.07596973e-01] | [14.312758445739746, 4.97475004196167] |
740fd5cc-5c6e-47a4-80ad-4ff130db65a7 | visual-relationship-detection-with-language-1 | 1904.07798 | null | http://arxiv.org/abs/1904.07798v1 | http://arxiv.org/pdf/1904.07798v1.pdf | Visual Relationship Detection with Language prior and Softmax | Visual relationship detection is an intermediate image understanding task
that detects two objects and classifies a predicate that explains the
relationship between two objects in an image. The three components are
linguistically and visually correlated (e.g. "wear" is related to "person" and
"shirt", while "laptop" is related to "table" and "on") thus, the solution
space is huge because there are many possible cases between them. Language and
visual modules are exploited and a sophisticated spatial vector is proposed.
The models in this work outperformed the state of arts without costly
linguistic knowledge distillation from a large text corpus and building complex
loss functions. All experiments were only evaluated on Visual Relationship
Detection and Visual Genome dataset. | ['Jaewon Jung', 'Jongyoul Park'] | 2019-04-16 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [ 1.29175439e-01 -1.34545460e-01 -1.89813182e-01 -4.43072706e-01
-2.74182200e-01 -5.70168495e-01 7.25571692e-01 3.95983011e-01
-3.41159731e-01 4.11103487e-01 -4.00010534e-02 -5.10465920e-01
-7.88137838e-02 -5.55012405e-01 -9.23461378e-01 -3.96532983e-01
2.16900691e-01 3.95456046e-01 2.80717254e-01 -1.45065278e-01
2.49471948e-01 2.27960512e-01 -1.67350769e+00 7.05996513e-01
2.21924037e-01 1.03680623e+00 6.58789217e-01 4.67879385e-01
-4.32371229e-01 9.39775884e-01 -6.79901838e-01 -6.95876658e-01
-8.13322812e-02 -3.38463604e-01 -8.03452075e-01 3.41579020e-01
8.42209756e-01 3.24808769e-02 -1.14402667e-01 1.13087022e+00
1.66606769e-01 1.37537509e-01 7.04140246e-01 -1.56779027e+00
-1.26813614e+00 5.87456942e-01 -9.62443948e-01 2.72143304e-01
6.65420711e-01 -1.34529799e-01 1.06939864e+00 -1.00621057e+00
5.74687779e-01 1.72030354e+00 1.62785068e-01 2.17913568e-01
-1.23820281e+00 -5.92262208e-01 4.31455225e-01 6.91903055e-01
-1.49107122e+00 -1.68312177e-01 7.48270035e-01 -7.62733757e-01
1.07801092e+00 4.01740611e-01 6.13821208e-01 9.40018117e-01
2.16307794e-03 9.84517634e-01 8.46632600e-01 -5.90174377e-01
-2.23819271e-01 6.50350749e-01 2.67781913e-01 9.79887486e-01
2.93296650e-02 -1.42633066e-01 -5.55554748e-01 2.47415319e-01
4.62541610e-01 3.60555737e-03 -6.60629645e-02 -5.39441347e-01
-1.15275073e+00 5.83226144e-01 4.49391276e-01 3.57232511e-01
-3.97776300e-03 2.00882986e-01 2.76420563e-01 1.87446430e-01
5.12371957e-02 2.20259011e-01 -1.40468881e-01 2.84324318e-01
-5.84146678e-01 -4.74970490e-02 5.35084128e-01 1.32486367e+00
8.80339563e-01 -3.20167631e-01 -1.77414849e-01 8.10472131e-01
4.17572051e-01 4.36738431e-01 3.13096732e-01 -6.71212018e-01
6.72951460e-01 9.09970701e-01 5.51011972e-02 -1.42669797e+00
-4.26911354e-01 -1.53705895e-01 -5.60256660e-01 1.67584687e-01
4.06772524e-01 1.96369424e-01 -9.19629216e-01 1.36795640e+00
1.87991500e-01 6.81126267e-02 7.41379112e-02 1.07390487e+00
1.33174193e+00 5.57863653e-01 3.04772735e-01 9.81874093e-02
2.22412372e+00 -1.11608541e+00 -8.30605567e-01 -7.66360283e-01
3.65195960e-01 -1.04534781e+00 1.43513954e+00 6.39790967e-02
-1.09407640e+00 -9.74803448e-01 -9.91920471e-01 -5.91810465e-01
-8.62910569e-01 5.80851316e-01 9.17035043e-01 2.41368413e-01
-8.28533232e-01 1.09250352e-01 -4.04322118e-01 -5.96207678e-01
2.34341785e-01 -9.40485112e-03 -4.67312932e-01 -7.39527401e-03
-1.00577557e+00 9.89957094e-01 4.88078713e-01 8.59193578e-02
-5.87332785e-01 -3.11531395e-01 -1.07954562e+00 9.86410584e-03
4.81721491e-01 -8.15811276e-01 6.55435801e-01 -9.64480579e-01
-8.36233318e-01 1.49860620e+00 -3.86972934e-01 -2.60200948e-01
4.16757137e-01 -2.21864372e-01 -6.77175879e-01 -1.34946429e-03
2.57640243e-01 8.35823953e-01 8.31383407e-01 -1.48316967e+00
-7.78264880e-01 -4.11671460e-01 2.86803335e-01 3.83984566e-01
-1.90737620e-01 4.79370117e-01 -1.00830781e+00 -3.91674668e-01
4.11092997e-01 -8.05149972e-01 3.80829036e-01 2.70447493e-01
-7.07952738e-01 -3.47200423e-01 1.09412003e+00 -7.18744099e-01
1.02453721e+00 -2.26877165e+00 7.59525523e-02 -3.70179079e-02
1.27393380e-02 7.54561499e-02 8.39387327e-02 1.36844099e-01
-3.94947559e-01 1.53176010e-01 3.03772390e-01 -1.90836206e-01
2.46771742e-02 2.06864148e-01 -2.97042042e-01 2.90971726e-01
-5.31208375e-03 8.35390031e-01 -8.14454317e-01 -8.11567187e-01
3.59692067e-01 5.81898153e-01 -1.90607801e-01 1.84322134e-01
-1.83361888e-01 2.58263946e-01 -4.88142408e-02 7.44207859e-01
3.97351563e-01 -4.73432273e-01 1.43845975e-01 -8.06429267e-01
-1.48293737e-03 6.69419616e-02 -1.17119038e+00 1.64465404e+00
-4.33669180e-01 9.59372342e-01 -3.53871882e-01 -8.29480767e-01
9.72630918e-01 6.46954253e-02 -1.06754713e-01 -7.80766666e-01
2.87502885e-01 -3.53780627e-01 -2.73320079e-02 -1.17174256e+00
2.00339794e-01 1.35743245e-01 -2.67384015e-02 1.79999359e-02
-3.83254945e-01 -3.79384346e-02 3.84565592e-01 1.11688361e-01
5.55763066e-01 4.79814708e-01 4.59857643e-01 -8.93328711e-02
7.84469426e-01 3.09395790e-02 3.99776012e-01 4.74860936e-01
-1.72822624e-01 3.45571816e-01 7.39041924e-01 -2.40595654e-01
-6.74893498e-01 -1.21863329e+00 1.72710553e-01 1.18398082e+00
8.09193969e-01 -5.93027055e-01 -3.40742201e-01 -5.70951760e-01
5.47070429e-02 9.92183924e-01 -6.84420049e-01 4.99361493e-02
-4.55758363e-01 -2.24001259e-01 2.40588605e-01 7.80259490e-01
5.78474522e-01 -8.45532596e-01 -7.02746689e-01 -3.33453059e-01
-3.76933485e-01 -1.46178246e+00 -5.26232004e-01 2.44841166e-02
-2.57411063e-01 -1.17400181e+00 -1.57990903e-01 -1.28572893e+00
8.68956685e-01 4.69188631e-01 1.40207851e+00 1.23981066e-01
-6.33201718e-01 2.51573354e-01 -7.28268549e-02 -5.90326667e-01
-1.51277021e-01 -6.38131380e-01 -8.31131041e-02 5.33711351e-02
5.96688330e-01 -1.79093048e-01 -5.27712822e-01 2.95682132e-01
-4.78864998e-01 4.03033435e-01 5.13325632e-01 6.99080706e-01
8.39809954e-01 3.83442074e-01 -2.56407768e-01 -7.39365518e-01
3.53413850e-01 -3.08138072e-01 -4.79240119e-01 7.71656215e-01
-1.56372026e-01 8.68517309e-02 4.37732041e-01 -6.58779681e-01
-9.35735881e-01 3.79107118e-01 5.35444796e-01 -6.10446095e-01
-4.28763121e-01 1.53100580e-01 -4.54946280e-01 3.79173696e-01
3.67751986e-01 7.42298737e-02 -2.88774401e-01 -3.25276285e-01
6.48993134e-01 3.74010414e-01 7.38215566e-01 -3.99256885e-01
6.13994420e-01 6.87858403e-01 1.10053383e-01 -7.74517596e-01
-1.01135337e+00 -4.29533452e-01 -7.73756087e-01 -1.53380394e-01
1.23421896e+00 -9.23842549e-01 -9.76102471e-01 -1.63604766e-02
-1.27901042e+00 1.94876447e-01 9.56893116e-02 5.45417607e-01
-4.53288376e-01 3.47481877e-01 -3.58161956e-01 -7.51602352e-01
2.01101732e-02 -1.03075790e+00 1.16334081e+00 2.61500031e-01
-3.64998877e-01 -9.64982688e-01 -6.16236866e-01 6.10427141e-01
-1.34631380e-01 9.22820643e-02 1.19228351e+00 -3.30569088e-01
-4.00587171e-01 -2.07960848e-02 -7.24679172e-01 7.56085739e-02
2.48411745e-01 3.44307303e-01 -7.64659882e-01 -1.03034876e-01
-1.32227585e-01 -1.74929962e-01 6.51511610e-01 1.37555346e-01
1.20007992e+00 -2.13532791e-01 -7.06594467e-01 4.63746041e-01
1.39194667e+00 5.03355622e-01 3.54086190e-01 3.30297321e-01
1.18711388e+00 9.98431504e-01 7.59467185e-01 -4.90402384e-03
6.13314092e-01 9.55823183e-01 3.48412305e-01 -5.11686027e-01
-4.37405646e-01 -3.02421927e-01 2.85324514e-01 2.74810940e-01
1.14597395e-01 -4.34873849e-01 -9.85447824e-01 5.24709344e-01
-2.07233214e+00 -1.03392589e+00 -3.84241372e-01 1.81682980e+00
5.64437270e-01 8.66444632e-02 -1.49347246e-01 -5.28271273e-02
9.59896266e-01 -1.00128122e-01 -5.35047889e-01 -2.38048762e-01
-3.73496711e-01 -3.34111333e-01 3.38753194e-01 3.94357294e-01
-1.33321488e+00 1.21540380e+00 6.24482250e+00 6.57509267e-01
-1.00580847e+00 -3.69788497e-03 6.21780694e-01 1.61316656e-02
-4.74312641e-02 2.09739119e-01 -9.23794746e-01 2.97596246e-01
1.87753543e-01 1.23262547e-01 2.21087754e-01 7.01722383e-01
1.75389089e-02 -3.36829543e-01 -1.41187453e+00 1.49144423e+00
4.79603261e-01 -1.07126796e+00 3.23523551e-01 -3.91892195e-01
1.81496009e-01 -5.08737564e-01 2.61347175e-01 4.81909513e-02
2.27397814e-01 -1.17730916e+00 1.00754619e+00 5.29210567e-01
6.14638746e-01 -4.94525343e-01 2.92934507e-01 1.77752659e-01
-1.48689735e+00 -6.59450367e-02 -4.49736327e-01 -8.72557536e-02
1.95288539e-01 1.37150899e-01 -8.94980073e-01 3.18745106e-01
1.01029861e+00 7.56735206e-01 -9.13279235e-01 6.03968203e-01
-7.58812487e-01 5.71060441e-02 -2.49760486e-02 1.14284670e-02
4.83072586e-02 -2.36544803e-01 4.91360992e-01 1.17089963e+00
1.39261171e-01 -2.14635029e-01 2.24088416e-01 1.18649614e+00
1.76277533e-01 1.22385636e-01 -6.76224470e-01 2.00607628e-01
3.26958954e-01 1.25858378e+00 -9.60447192e-01 -5.18856525e-01
-6.60144150e-01 1.22839248e+00 3.24942380e-01 5.13012350e-01
-1.26691687e+00 -3.45152050e-01 5.61453402e-01 3.24414611e-01
3.46538186e-01 -6.59404835e-03 -6.31274283e-02 -9.87988889e-01
1.24494575e-01 -4.51201588e-01 5.15492618e-01 -1.42310917e+00
-1.17099798e+00 5.20310462e-01 9.57044289e-02 -1.10369122e+00
-6.08705506e-02 -8.45102310e-01 -3.05644691e-01 8.50238621e-01
-1.35712314e+00 -1.28919041e+00 -4.03483212e-01 8.21715951e-01
5.70568502e-01 -1.82026625e-01 7.52302766e-01 3.28779131e-01
-6.05437100e-01 5.73425829e-01 -4.88163561e-01 2.54256397e-01
5.41471660e-01 -1.16937768e+00 2.32358903e-01 9.53309119e-01
5.33073008e-01 7.25856960e-01 7.99385071e-01 -6.23645842e-01
-1.25805163e+00 -8.53295028e-01 1.20395958e+00 -5.51568091e-01
8.01763356e-01 -5.02601624e-01 -8.32633913e-01 7.63074458e-01
4.47914213e-01 -6.73993081e-02 5.50635815e-01 -2.14153714e-02
-8.58258247e-01 3.18998396e-02 -8.53617132e-01 8.71080518e-01
1.18883181e+00 -7.91927874e-01 -7.02329993e-01 5.20304263e-01
5.23954034e-01 -4.09735054e-01 -5.77557564e-01 2.34191537e-01
3.46855938e-01 -8.69713426e-01 1.44761288e+00 -6.47259653e-01
2.69033462e-01 -8.32256973e-01 -1.50610223e-01 -8.69547546e-01
-3.08632106e-01 -1.78967848e-01 -2.17241868e-01 1.33966374e+00
4.60956067e-01 -1.27044365e-01 4.61466789e-01 5.92281520e-01
2.92679995e-01 -3.42518359e-01 -5.70598960e-01 -9.03871238e-01
-5.13337195e-01 -2.40942404e-01 4.75638121e-01 1.14591086e+00
5.91078997e-02 7.79104114e-01 -3.41909736e-01 2.80973941e-01
4.53952074e-01 3.64635527e-01 5.54921627e-01 -9.17922556e-01
-4.58515942e-01 -6.17341280e-01 -6.80064380e-01 -1.01838648e+00
3.08864355e-01 -9.34768677e-01 -1.71681628e-01 -1.75394738e+00
4.41139728e-01 -1.20790072e-01 -1.38717934e-01 9.02834177e-01
-3.25752437e-01 1.88219681e-01 1.82582602e-01 6.38380572e-02
-6.50224864e-01 1.13487855e-01 1.10149908e+00 -6.39008164e-01
2.46613957e-02 -3.42676640e-01 -7.63112307e-01 1.07029581e+00
4.87013638e-01 -1.38551578e-01 -7.28146374e-01 -7.28669524e-01
4.31198031e-01 7.96618536e-02 6.93898320e-01 -7.17556357e-01
3.44861805e-01 -1.47051334e-01 4.80361938e-01 -6.60461962e-01
5.74168026e-01 -1.10413349e+00 2.41214529e-01 2.66851395e-01
-3.53548825e-01 3.27980548e-01 2.39030555e-01 5.39859533e-01
-3.25526834e-01 -6.42459616e-02 5.39409816e-01 -1.67596936e-02
-1.29550338e+00 -1.20577112e-01 1.20469090e-02 -2.44995520e-01
1.39740288e+00 -3.95680636e-01 -4.34821278e-01 -3.31029356e-01
-8.95136535e-01 2.76078314e-01 2.55249470e-01 8.85524094e-01
7.38004446e-01 -1.23814559e+00 -5.12711585e-01 7.59784281e-02
4.94157702e-01 -2.42046028e-01 1.05607875e-01 5.15099287e-01
-3.15952718e-01 2.35463947e-01 -1.69894814e-01 -6.45677567e-01
-1.79123330e+00 1.01546144e+00 2.94540763e-01 2.63041139e-01
-7.00991869e-01 1.12156844e+00 5.28955579e-01 1.42869547e-01
6.31852627e-01 -3.38449478e-01 -6.55834615e-01 2.29068965e-01
4.86047059e-01 2.10241624e-03 -4.11361009e-01 -1.12135041e+00
-6.59696460e-01 8.99028420e-01 -7.73056503e-03 2.02377617e-01
7.69021094e-01 -4.62981761e-01 -2.37705126e-01 7.16645718e-01
1.22344100e+00 -1.56181797e-01 -9.40182090e-01 -2.80681342e-01
6.84568286e-02 -4.68783766e-01 -1.45817772e-01 -7.84309506e-01
-1.09970474e+00 9.88228977e-01 7.53347576e-01 1.47660434e-01
1.09500098e+00 4.69442397e-01 3.12104851e-01 3.62886548e-01
2.07080618e-01 -1.03316045e+00 2.17688113e-01 3.11625004e-01
1.00470281e+00 -1.46584404e+00 2.96956390e-01 -8.66082251e-01
-1.02354205e+00 8.95317376e-01 8.85887861e-01 2.77871430e-01
5.47580183e-01 9.04168859e-02 1.15114577e-01 -5.49123824e-01
-5.63676536e-01 -5.08602083e-01 7.67040730e-01 5.44046402e-01
6.38487399e-01 1.92883477e-01 -1.62941843e-01 3.93168181e-01
-2.78759927e-01 -6.13156259e-01 1.28888087e-02 5.85106492e-01
-8.29655975e-02 -7.56356001e-01 -3.77519995e-01 -1.51849017e-02
-1.37944445e-01 -2.07166970e-01 -4.92338628e-01 9.30765390e-01
5.79456985e-01 1.11664391e+00 4.53195453e-01 -2.67871648e-01
4.13898528e-01 -2.38165736e-01 5.00454128e-01 -5.19634783e-01
-3.99748862e-01 6.38713539e-02 3.85531671e-02 -5.66216469e-01
-7.14254975e-01 -5.48207641e-01 -1.52016461e+00 -9.28103272e-03
-1.42565385e-01 -3.27631474e-01 6.00779295e-01 8.60637307e-01
2.07110599e-01 7.36646056e-01 8.89507905e-02 -2.47891918e-01
3.11075896e-01 -4.97341514e-01 -3.37144881e-01 8.18305254e-01
-2.67196372e-02 -8.82654428e-01 -4.28420566e-02 4.34173375e-01] | [10.355036735534668, 1.604426383972168] |
6b904112-6f44-43ee-b71d-1f59abeb1be5 | visual-relationship-detection-based-on-guided | 1805.10802 | null | http://arxiv.org/abs/1805.10802v1 | http://arxiv.org/pdf/1805.10802v1.pdf | Visual Relationship Detection Based on Guided Proposals and Semantic Knowledge Distillation | A thorough comprehension of image content demands a complex grasp of the
interactions that may occur in the natural world. One of the key issues is to
describe the visual relationships between objects. When dealing with real world
data, capturing these very diverse interactions is a difficult problem. It can
be alleviated by incorporating common sense in a network. For this, we propose
a framework that makes use of semantic knowledge and estimates the relevance of
object pairs during both training and test phases. Extracted from precomputed
models and training annotations, this information is distilled into the neural
network dedicated to this task. Using this approach, we observe a significant
improvement on all classes of Visual Genome, a challenging visual relationship
dataset. A 68.5% relative gain on the recall at 100 is directly related to the
relevance estimate and a 32.7% gain to the knowledge distillation. | ['Françoise Prêteux', 'François Plesse', 'Bertrand Delezoide', 'Alexandru Ginsca'] | 2018-05-28 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [ 2.66060859e-01 3.69709805e-02 -4.39085588e-02 -5.52132249e-01
-9.29702595e-02 -6.29363477e-01 7.79869616e-01 4.99511361e-01
-5.89552343e-01 6.14964843e-01 2.22883239e-01 1.21263504e-01
-1.84682071e-01 -7.01134861e-01 -9.36514556e-01 -3.93727005e-01
6.02186769e-02 5.66286147e-01 3.87398154e-01 -2.75883287e-01
2.02690318e-01 4.92403328e-01 -1.78968060e+00 4.68984693e-01
6.23257756e-01 1.09383762e+00 6.38754487e-01 5.78917444e-01
-3.65538627e-01 1.02535439e+00 -6.16834939e-01 -6.42855287e-01
1.01781962e-02 -1.05045661e-01 -9.40839171e-01 -3.24613415e-03
5.79925179e-01 -1.84702918e-01 -9.82591063e-02 9.70188916e-01
2.13982269e-01 1.70270666e-01 7.15916991e-01 -1.22307003e+00
-3.50582004e-01 6.84063673e-01 -5.40535867e-01 2.15214357e-01
3.93105179e-01 7.57135674e-02 1.10669219e+00 -8.21651399e-01
8.20513368e-01 1.19798315e+00 3.29106227e-02 2.07143605e-01
-1.34524810e+00 -4.19488728e-01 2.60437697e-01 6.57165170e-01
-1.27413797e+00 -2.91159064e-01 6.24716878e-01 -6.47040546e-01
1.01142418e+00 2.34114498e-01 8.42771292e-01 9.08313394e-01
-1.80892676e-01 8.26839209e-01 8.49188447e-01 -5.15129268e-01
1.17469439e-02 5.07179856e-01 4.32666689e-01 3.07022810e-01
3.74954760e-01 -1.05830416e-01 -6.95466220e-01 4.64302957e-01
5.57853520e-01 -2.34831814e-02 -3.72387648e-01 -6.53728306e-01
-1.00997043e+00 4.16043967e-01 8.52620602e-01 3.55959475e-01
-2.96586692e-01 7.25935474e-02 2.55922437e-01 1.17449962e-01
1.24213859e-01 7.80612946e-01 -2.32486904e-01 -1.09565206e-01
-6.13440037e-01 1.86616406e-01 9.52322960e-01 8.54598582e-01
7.56969273e-01 -5.52615941e-01 -9.85228047e-02 8.81295860e-01
2.15135619e-01 3.21490854e-01 4.45572287e-02 -6.96091950e-01
5.33248842e-01 9.46509600e-01 1.27439737e-01 -1.25305951e+00
-3.83504897e-01 -6.57500505e-01 -6.57915533e-01 1.68343082e-01
6.10197127e-01 2.95996606e-01 -8.73006344e-01 1.80898905e+00
3.76701355e-01 1.75623484e-02 1.01005934e-01 8.29637408e-01
9.61987674e-01 4.39796835e-01 3.08255434e-01 4.18556221e-02
1.58840859e+00 -9.75836098e-01 -6.83037877e-01 -7.54189610e-01
3.86703283e-01 -7.02622712e-01 8.95184994e-01 3.85240376e-01
-1.04360521e+00 -6.18321180e-01 -1.16350663e+00 -3.10213625e-01
-6.40112936e-01 2.86425427e-02 6.37550533e-01 1.15077503e-01
-9.70275342e-01 3.84477228e-01 -4.29155886e-01 -4.02707249e-01
5.07051170e-01 4.55493093e-01 -6.07924640e-01 -4.64255303e-01
-9.69935596e-01 1.18481803e+00 7.47060180e-01 3.13354656e-02
-6.19416237e-01 -6.91331744e-01 -6.77727997e-01 3.47117543e-01
7.71992147e-01 -7.31786311e-01 1.04038787e+00 -1.02429187e+00
-7.11403191e-01 9.43344116e-01 -1.47210345e-01 -3.29703718e-01
4.55056846e-01 -4.44381237e-01 -4.48104218e-02 1.02675945e-01
-1.16695724e-01 9.64435637e-01 5.97992599e-01 -1.56706893e+00
-6.57230616e-01 -2.90789574e-01 4.31643367e-01 4.70657825e-01
-3.90600592e-01 -1.53473943e-01 -9.16118443e-01 -3.40693861e-01
-1.38343558e-01 -9.57471967e-01 -4.44098450e-02 2.18307361e-01
-3.14899236e-01 -1.55249566e-01 7.39263475e-01 -6.52479172e-01
8.30596030e-01 -2.05603600e+00 4.16111648e-01 2.51673430e-01
4.74861413e-01 4.32272583e-01 -2.86739498e-01 4.14791733e-01
-1.44604921e-01 1.80161446e-02 1.07419066e-01 -3.41894895e-01
-5.66978455e-02 1.77466020e-01 -1.54170096e-01 -2.24748462e-01
2.27250427e-01 1.10931516e+00 -9.34537828e-01 -3.38255167e-01
3.04877698e-01 5.64984262e-01 -4.04494882e-01 5.15195549e-01
-3.42808545e-01 2.69487947e-01 -1.58677384e-01 2.73117900e-01
6.22300684e-01 -5.52137792e-01 4.59535718e-01 -6.22718751e-01
4.52004552e-01 1.37556732e-01 -1.04576683e+00 1.74088621e+00
-4.48696524e-01 8.45167637e-01 -4.40302998e-01 -9.20162976e-01
7.56517112e-01 -7.20817372e-02 1.65334612e-01 -9.75826502e-01
1.47689149e-01 -1.79211661e-01 1.29257843e-01 -6.73019886e-01
5.28876841e-01 1.25325397e-01 2.41870522e-01 3.17616671e-01
1.97412625e-01 -3.72975245e-02 4.41879898e-01 5.25312185e-01
9.95918453e-01 8.96036923e-02 5.10686755e-01 -3.82712623e-03
3.92972857e-01 4.77964319e-02 1.57460988e-01 6.23509467e-01
1.80993840e-01 3.07885021e-01 7.58195400e-01 -6.27566874e-01
-7.62624502e-01 -1.00901365e+00 2.45388478e-01 8.72239292e-01
4.50687110e-01 -4.72995669e-01 -4.88924593e-01 -6.61218762e-01
-2.81492248e-02 7.22833991e-01 -7.95391619e-01 -4.46596116e-01
-3.93333524e-01 -4.88847524e-01 -7.42691085e-02 6.34575784e-01
6.05390668e-01 -1.10802257e+00 -8.35060120e-01 -1.86903089e-01
-3.49355161e-01 -1.59044206e+00 5.42841665e-03 1.38149321e-01
-5.56573749e-01 -1.24931324e+00 -2.59771138e-01 -7.35798180e-01
7.39561141e-01 4.32052255e-01 1.65892315e+00 1.79463536e-01
-5.07065594e-01 2.91991532e-01 -3.89025033e-01 -5.46850801e-01
-2.40377918e-01 1.80017315e-02 -4.32281673e-01 -7.91285187e-02
4.96160865e-01 -4.70563412e-01 -5.18419683e-01 2.76926011e-01
-8.01024497e-01 5.27836561e-01 6.72646523e-01 7.10063756e-01
4.46093440e-01 -2.82182526e-02 8.13880935e-02 -7.63985395e-01
3.91518891e-01 -3.61399621e-01 -4.01417166e-01 5.10878623e-01
-3.38070303e-01 1.98487267e-01 1.92212105e-01 -5.51801562e-01
-1.13768983e+00 2.52664894e-01 2.45401695e-01 -3.00549954e-01
-2.84300804e-01 5.28698027e-01 -3.11305970e-01 -4.02420014e-02
5.94474494e-01 -1.29328161e-01 -1.77444875e-01 -2.15775043e-01
4.85054463e-01 3.37234408e-01 5.16483545e-01 -5.43353140e-01
5.86933732e-01 3.59125048e-01 6.47971556e-02 -7.38080025e-01
-1.19514227e+00 -4.93174225e-01 -7.43156195e-01 -3.64816457e-01
9.23918426e-01 -8.84108722e-01 -8.27171087e-01 2.24348128e-01
-1.27120268e+00 -3.85400683e-01 -2.71625489e-01 4.07566041e-01
-2.95562357e-01 4.61811833e-02 6.39077723e-02 -5.45598626e-01
-3.50509696e-02 -1.11598158e+00 7.91345835e-01 3.35425794e-01
-3.91071022e-01 -9.50320423e-01 -1.28273323e-01 5.62857509e-01
3.02571595e-01 2.98943013e-01 1.00976372e+00 -6.52217448e-01
-8.64673674e-01 -1.02592103e-01 -7.51262069e-01 2.52539605e-01
1.22899011e-01 -1.91871315e-01 -1.08244693e+00 -6.77477047e-02
-1.65462598e-01 -4.65317667e-01 8.70341539e-01 1.79813966e-01
1.23599362e+00 -4.09066714e-02 -4.67871636e-01 2.76003361e-01
1.50214756e+00 8.61767754e-02 6.30690932e-01 1.02769129e-01
9.64136004e-01 1.01786721e+00 6.72617555e-01 2.35518962e-01
6.61591351e-01 9.61487532e-01 7.86592901e-01 -6.42917380e-02
-2.80852377e-01 -1.68782085e-01 -1.34669542e-01 3.48638326e-01
-1.35688245e-01 -4.32904273e-01 -1.25210619e+00 7.21337318e-01
-2.04824805e+00 -8.75947356e-01 -1.38232052e-01 2.11304164e+00
8.84815037e-01 1.74631238e-01 -1.49324149e-01 -1.42866690e-02
5.15317678e-01 -4.45850333e-03 -4.62386787e-01 -1.45946860e-01
1.11705504e-01 1.49633680e-02 1.50805101e-01 3.55818599e-01
-7.95341134e-01 9.81163561e-01 6.48713017e+00 6.89917684e-01
-1.04053819e+00 -1.49962738e-01 5.63742638e-01 -9.12475288e-02
-1.16887070e-01 -1.11696921e-01 -6.19249225e-01 1.63325042e-01
7.17219532e-01 -3.48187312e-02 5.03967345e-01 5.38574040e-01
-5.62485382e-02 -6.32480562e-01 -1.43232131e+00 1.14660525e+00
2.94423997e-01 -1.18499279e+00 1.93697646e-01 -5.71126603e-02
5.81857264e-01 -2.04646707e-01 5.40929288e-02 1.30935654e-01
4.20432955e-01 -1.24818861e+00 6.23938978e-01 6.95066512e-01
5.67595184e-01 -8.45296443e-01 7.34100461e-01 3.82888347e-01
-1.05372047e+00 9.50041227e-03 -3.49839389e-01 -1.49990067e-01
5.18700965e-02 7.25486696e-01 -1.22651148e+00 4.55575198e-01
7.74523735e-01 8.63423347e-01 -7.80091405e-01 1.14504075e+00
-6.05691493e-01 2.20250353e-01 -1.49724782e-01 -1.25260100e-01
-1.21687956e-01 1.45582333e-01 3.42705160e-01 1.08934927e+00
-7.73839578e-02 -9.58919078e-02 1.06528163e-01 8.10567021e-01
-2.41237059e-01 3.50255258e-02 -6.50675058e-01 -1.23275787e-01
4.08669859e-01 1.41251004e+00 -7.92302668e-01 -5.23995996e-01
-2.88360119e-01 9.69923139e-01 8.64088178e-01 3.44056666e-01
-8.19303095e-01 -1.20232694e-01 5.74868560e-01 2.92793922e-02
3.11622500e-01 -2.04055503e-01 -2.54426837e-01 -9.68204141e-01
2.69820958e-01 -7.82997847e-01 2.36765712e-01 -1.15190411e+00
-1.13084102e+00 5.17873526e-01 3.43598068e-01 -9.77523029e-01
-2.77650833e-01 -7.31583655e-01 -2.74316877e-01 8.04202318e-01
-1.57031977e+00 -1.24661398e+00 -9.64288831e-01 4.32219833e-01
4.16903824e-01 2.21288905e-01 6.73075736e-01 1.06386609e-01
-2.81961590e-01 3.93939883e-01 -4.74240690e-01 4.30256762e-02
6.66223764e-01 -1.29860353e+00 2.67077595e-01 6.94260538e-01
4.67250437e-01 6.21829033e-01 7.67155468e-01 -5.49408317e-01
-1.19055915e+00 -8.88701320e-01 9.86936510e-01 -5.89072049e-01
5.02435386e-01 -3.95911723e-01 -9.80500221e-01 6.21003687e-01
2.84143418e-01 6.10278249e-02 6.98891461e-01 1.79235831e-01
-7.40064859e-01 -1.17673740e-01 -9.32316244e-01 6.19681716e-01
1.28634632e+00 -5.34189343e-01 -7.02103615e-01 1.13484137e-01
5.34750104e-01 -4.11870956e-01 -7.97843516e-01 2.15591565e-01
7.46687233e-01 -1.00166214e+00 1.13829410e+00 -5.85093141e-01
6.97645783e-01 -2.22550333e-01 -1.35117248e-01 -1.38860929e+00
-1.68792725e-01 2.35937163e-02 -9.78574082e-02 1.08664131e+00
4.11753237e-01 -2.79159427e-01 5.77184737e-01 7.85736978e-01
3.91290188e-01 -4.65854108e-01 -4.33542073e-01 -6.20620668e-01
-3.44986975e-01 -3.73433352e-01 3.31552684e-01 9.54352379e-01
-6.26948625e-02 6.13736331e-01 -3.26543450e-01 1.40887409e-01
5.12035966e-01 8.59702285e-03 9.24207270e-01 -1.47428024e+00
-3.33632410e-01 -3.05237919e-01 -6.70875490e-01 -9.28587437e-01
4.64185216e-02 -7.47037411e-01 3.17119770e-02 -1.90883934e+00
6.27088904e-01 -3.21980715e-01 -2.75834739e-01 5.59153199e-01
-2.60347277e-01 2.91548938e-01 4.81892794e-01 1.71881579e-02
-7.49101698e-01 2.42454156e-01 1.26958072e+00 -2.93907911e-01
-3.52808870e-02 -4.44745630e-01 -7.70474255e-01 7.21004069e-01
7.92014122e-01 -4.06263143e-01 -8.11826825e-01 -7.34383523e-01
6.16798341e-01 -1.10567339e-01 6.00869596e-01 -1.13213181e+00
2.87816435e-01 -2.04268560e-01 5.05953252e-01 -5.07102907e-01
6.85416043e-01 -1.13787889e+00 3.50704610e-01 2.79128462e-01
-5.09910047e-01 1.88893303e-02 3.47951531e-01 6.17014170e-01
-3.21751565e-01 -1.82412863e-01 4.90248382e-01 -1.18835472e-01
-1.09880698e+00 -8.04731324e-02 2.65038341e-01 8.00045282e-02
1.26237023e+00 -2.48238653e-01 -4.79059190e-01 -4.39468473e-01
-9.27841008e-01 2.98361272e-01 4.00053948e-01 6.97889745e-01
6.38034225e-01 -1.19998169e+00 -4.85733211e-01 9.57661197e-02
4.51583773e-01 8.86869803e-02 3.61413926e-01 6.49953246e-01
-2.19784975e-01 1.92636803e-01 -5.57438076e-01 -7.01404810e-01
-1.65929294e+00 4.85974163e-01 2.32409149e-01 -3.99982810e-01
-2.75013566e-01 9.13441300e-01 3.84668887e-01 1.73923925e-01
3.68077397e-01 -2.20867574e-01 -7.03080475e-01 3.00277829e-01
7.09532142e-01 9.30126011e-02 3.31520662e-02 -6.86141908e-01
-3.91949952e-01 6.68528795e-01 -3.56396824e-01 5.52557781e-02
1.39658475e+00 -8.26999247e-02 -1.77999586e-01 5.13277352e-01
1.10012007e+00 -1.69244647e-01 -1.18874359e+00 -3.86878878e-01
7.49510527e-02 -5.81247270e-01 -1.12942450e-01 -1.21379864e+00
-1.09688926e+00 1.02323508e+00 5.45327306e-01 1.60784528e-01
1.14432085e+00 2.63482660e-01 1.56690702e-01 6.10251069e-01
2.70661771e-01 -7.09994137e-01 4.06142384e-01 4.50888067e-01
1.09852290e+00 -1.45012712e+00 7.41645247e-02 -7.68473804e-01
-8.41921806e-01 9.95042562e-01 9.10064340e-01 6.51513273e-03
4.55525160e-01 1.67772397e-01 -3.62461843e-02 -4.29825664e-01
-9.74217594e-01 -3.91217262e-01 6.22718751e-01 6.18349910e-01
2.64000863e-01 -4.19250838e-02 -9.88920480e-02 2.57810384e-01
4.72747022e-03 -1.03560679e-01 3.14459801e-01 7.12075591e-01
-4.11170870e-01 -1.02027941e+00 6.53920881e-03 3.24299067e-01
-6.97125271e-02 -4.49293219e-02 -6.10641241e-01 7.06513226e-01
2.69109845e-01 8.33731711e-01 2.72459418e-01 -3.75081927e-01
4.40600961e-01 -6.79899231e-02 5.31559050e-01 -6.82781160e-01
-4.58046496e-01 -4.37162101e-01 2.28981644e-01 -7.26950526e-01
-6.98054850e-01 -2.37882346e-01 -1.22370827e+00 5.86271286e-04
-2.26251110e-01 -1.59819096e-01 7.46384203e-01 1.05139482e+00
2.79446721e-01 8.34691584e-01 8.50941241e-02 -7.30551779e-01
-7.41828606e-02 -7.25372851e-01 -1.83212698e-01 8.91557455e-01
1.51054963e-01 -9.07992184e-01 -2.98668761e-02 1.79225966e-01] | [10.456360816955566, 1.6993961334228516] |
4a6d3a7c-dad6-40b5-ae10-f91cac508f0a | towards-explainable-ai-for-channel-estimation | 2307.00952 | null | https://arxiv.org/abs/2307.00952v1 | https://arxiv.org/pdf/2307.00952v1.pdf | Towards Explainable AI for Channel Estimation in Wireless Communications | Research into 6G networks has been initiated to support a variety of critical artificial intelligence (AI) assisted applications such as autonomous driving. In such applications, AI-based decisions should be performed in a real-time manner. These decisions include resource allocation, localization, channel estimation, etc. Considering the black-box nature of existing AI-based models, it is highly challenging to understand and trust the decision-making behavior of such models. Therefore, explaining the logic behind those models through explainable AI (XAI) techniques is essential for their employment in critical applications. This manuscript proposes a novel XAI-based channel estimation (XAI-CHEST) scheme that provides detailed reasonable interpretability of the deep learning (DL) models that are employed in doubly-selective channel estimation. The aim of the proposed XAI-CHEST scheme is to identify the relevant model inputs by inducing high noise on the irrelevant ones. As a result, the behavior of the studied DL-based channel estimators can be further analyzed and evaluated based on the generated interpretations. Simulation results show that the proposed XAI-CHEST scheme provides valid interpretations of the DL-based channel estimators for different scenarios. | ['Laurent Clavier', 'Ali J. Ghandour', 'Yahia Medjahdi', 'Abdul Karim Gizzini'] | 2023-07-03 | null | null | null | null | ['decision-making'] | ['reasoning'] | [ 3.64051700e-01 1.65310398e-01 -3.32309157e-01 -5.23021281e-01
-3.68970484e-01 -1.36673272e-01 4.59560841e-01 -3.86998989e-02
1.42912775e-01 1.12716079e+00 -1.11994475e-01 -8.95862460e-01
-5.13739705e-01 -7.20333278e-01 -6.83313668e-01 -9.37943876e-01
-4.03801292e-01 3.26963276e-01 -3.70247126e-01 -1.44442022e-01
1.57772914e-01 6.15041435e-01 -1.35632026e+00 1.14145786e-01
8.16065252e-01 1.44785357e+00 1.16718739e-01 9.09856915e-01
2.01868534e-01 8.80735636e-01 -6.02513850e-01 -1.65501818e-01
1.44191965e-01 -8.06093872e-01 -2.65870214e-01 -8.24221671e-02
-3.15047830e-01 -3.70372146e-01 -3.14830303e-01 1.03057826e+00
4.37635571e-01 -3.54874820e-01 7.75348485e-01 -1.60207021e+00
-7.20358491e-02 1.00126612e+00 -2.15732798e-01 2.19621167e-01
-1.45630062e-01 -5.93661517e-02 8.85566771e-01 -5.77848077e-01
2.72606686e-03 1.23918140e+00 3.77301335e-01 3.93569946e-01
-7.70944834e-01 -1.08470833e+00 1.81747913e-01 4.84495729e-01
-1.42990136e+00 -6.23197615e-01 8.94007564e-01 -3.46109234e-02
4.56877619e-01 2.09837109e-01 6.37608409e-01 1.09339643e+00
5.25585949e-01 8.21948469e-01 1.00463378e+00 -7.50516236e-01
4.52630967e-01 4.25390124e-01 6.74379393e-02 7.16693819e-01
6.45532846e-01 2.47746900e-01 -6.67059243e-01 -1.33263290e-01
7.18678117e-01 -4.77871180e-01 -1.14992239e-01 -4.97320481e-02
-1.02384198e+00 7.41268277e-01 3.06815773e-01 1.37032941e-01
-5.79393327e-01 6.91370845e-01 1.27432242e-01 4.39231992e-01
1.33386940e-01 2.49827623e-01 -1.39636502e-01 -7.03136548e-02
-6.91960514e-01 1.03770398e-01 7.04412460e-01 1.27722323e+00
6.21514022e-01 5.43163419e-01 -5.02761304e-02 4.50092033e-02
8.68388534e-01 6.79904163e-01 1.11691095e-02 -5.88235080e-01
4.14300889e-01 9.06109214e-02 1.48360148e-01 -1.10512638e+00
-5.76114297e-01 -1.28906202e+00 -1.17432857e+00 1.05268769e-02
3.04030448e-01 -5.36539137e-01 -7.63309181e-01 1.55574346e+00
-3.27782720e-01 4.19218093e-01 3.36564779e-01 8.74161541e-01
6.96528912e-01 5.99666834e-01 2.28191406e-01 -1.68666214e-01
1.05867791e+00 -5.67313254e-01 -8.67952347e-01 -1.38141811e-01
7.26928651e-01 -4.16293263e-01 4.88496035e-01 4.83969629e-01
-5.43180346e-01 -6.05713129e-01 -1.47196507e+00 5.81396401e-01
-1.61192432e-01 4.77022916e-01 8.66363049e-01 1.03503382e+00
-9.43516076e-01 6.82134181e-02 -4.40556824e-01 -3.71201277e-01
4.71816927e-01 6.84484601e-01 9.99103710e-02 2.34404176e-01
-1.63436806e+00 6.56807601e-01 3.57127905e-01 7.34026194e-01
-1.23501921e+00 -1.85102969e-01 -4.12932932e-01 3.67114723e-01
1.91015586e-01 -7.32723117e-01 1.09610319e+00 -1.15064597e+00
-1.37695706e+00 1.66053146e-01 -3.16152394e-01 -1.17055178e+00
5.12132645e-01 -1.76201444e-02 -7.33685732e-01 3.95016298e-02
-2.59837747e-01 3.64234865e-01 1.10938311e+00 -1.49898398e+00
-8.35264802e-01 -2.65293628e-01 -8.44856817e-03 1.03674896e-01
-1.37621164e-01 -4.27945763e-01 -2.01551512e-01 -3.33963513e-01
1.16772823e-01 -1.08078742e+00 -3.51842463e-01 -1.44198731e-01
-7.69886434e-01 2.98503399e-01 8.45976472e-01 -3.38857800e-01
1.40747988e+00 -1.93914759e+00 -3.42156380e-01 8.99295926e-01
4.41649944e-01 1.60952851e-01 1.34292454e-01 3.77354652e-01
2.56334245e-01 2.16211036e-01 1.17325023e-01 -1.37178525e-01
-1.20940477e-01 2.09602982e-01 -2.68899709e-01 4.01966542e-01
-8.81922171e-02 6.66206181e-01 -6.00988507e-01 -2.64603138e-01
4.77284998e-01 4.06226784e-01 -4.00710791e-01 1.11994669e-01
-1.65607452e-01 8.08216929e-01 -9.98082101e-01 7.26650596e-01
5.39123714e-01 -1.61085054e-01 1.84978962e-01 -5.01175284e-01
1.11136690e-01 -1.05634574e-02 -9.50406194e-01 9.93888140e-01
-7.61805713e-01 1.02820730e+00 -9.01513249e-02 -1.33714700e+00
1.14065802e+00 6.20882034e-01 3.20453197e-01 -7.90279329e-01
5.79686880e-01 5.43798149e-01 6.93192422e-01 -2.47618556e-01
1.12744235e-01 9.79409367e-02 -3.10000926e-02 2.60554761e-01
-3.68952066e-01 2.64145702e-01 -4.77043748e-01 6.44108728e-02
7.48959780e-01 -3.47422361e-01 5.24794459e-01 -4.59772885e-01
9.43257153e-01 -1.78685963e-01 5.18286884e-01 1.12173200e+00
-4.38262731e-01 1.22783042e-01 3.63520116e-01 -5.04171073e-01
-8.65701318e-01 -7.30943441e-01 -1.70907214e-01 2.68030286e-01
5.34713686e-01 3.03452928e-02 -5.16170084e-01 -3.72032404e-01
-1.59724221e-01 9.66485679e-01 -4.68600154e-01 -4.58847284e-01
-1.01010866e-01 -8.41969550e-01 8.54068696e-01 3.76871556e-01
1.03054655e+00 -7.16044545e-01 -7.67683446e-01 4.62035060e-01
1.44710809e-01 -1.21508873e+00 6.34634554e-01 5.73770881e-01
-6.44541621e-01 -5.73020458e-01 -4.66255516e-01 -2.83746511e-01
5.95310569e-01 6.29851967e-02 8.60689461e-01 1.66649386e-01
1.71792477e-01 -2.07641032e-02 -4.38741654e-01 -7.14272022e-01
-5.57609260e-01 2.62647092e-01 9.46282372e-02 4.39957589e-01
3.96403164e-01 -5.06355047e-01 -5.51055491e-01 3.62414479e-01
-3.02641898e-01 1.16925649e-01 9.34166670e-01 7.69234776e-01
3.13870370e-01 5.35523713e-01 1.06330299e+00 -9.42875266e-01
9.04261053e-01 -5.98111868e-01 -5.86709976e-01 3.94933283e-01
-8.01409602e-01 3.13144356e-01 8.09236288e-01 6.05330765e-02
-1.15936720e+00 -6.76612630e-02 -2.28702351e-01 -1.26261845e-01
-2.49416634e-01 6.07448220e-01 -6.85128093e-01 -3.12391013e-01
4.86451685e-01 2.26273611e-01 -3.08983952e-01 2.25013751e-03
9.70508233e-02 1.16209221e+00 2.44395360e-01 -4.92450446e-01
7.42000103e-01 2.77676433e-01 4.66273785e-01 -1.02906239e+00
-6.04721069e-01 -5.33943325e-02 -3.20623994e-01 -5.08185148e-01
5.50267696e-01 -9.13086474e-01 -6.82258904e-01 3.52608442e-01
-1.08850682e+00 -6.51823655e-02 3.54751915e-01 7.23174989e-01
-5.15709162e-01 -9.58956555e-02 -9.54007208e-02 -1.26507783e+00
-2.51758486e-01 -1.51425755e+00 9.29655850e-01 2.84129739e-01
-3.30664635e-01 -1.12857985e+00 -7.24898160e-01 2.59343445e-01
6.55827999e-01 -2.38005966e-02 1.30872750e+00 -6.91927016e-01
-9.46349502e-01 -2.08991453e-01 -3.34081143e-01 -3.83218750e-03
-1.43396761e-02 -1.55657411e-01 -1.26843154e+00 4.72482778e-02
-2.89966911e-01 3.10939878e-01 5.75204074e-01 8.96473289e-01
1.29361355e+00 -1.45673111e-01 -4.73096937e-01 6.49043500e-01
1.41932797e+00 7.70498395e-01 8.55369627e-01 2.99951643e-01
4.14419711e-01 1.16033316e-01 7.63717890e-01 6.69583678e-01
1.85069859e-01 5.91323793e-01 8.59508693e-01 -3.25695634e-01
1.54617161e-01 -2.03868568e-01 8.83921534e-02 2.63079226e-01
-8.19705129e-02 -9.30971861e-01 -6.44980669e-01 1.60479307e-01
-2.00837874e+00 -8.73612523e-01 -3.82851303e-01 2.15454912e+00
4.46316972e-02 4.76491034e-01 -2.30655864e-01 6.45196855e-01
5.79968095e-01 -6.59308210e-02 -7.35866070e-01 -5.54398060e-01
-1.58552706e-01 -1.23429872e-01 9.16650772e-01 5.42740107e-01
-1.05239356e+00 7.45285273e-01 5.96853304e+00 7.36416399e-01
-1.46350634e+00 -1.60169557e-01 1.04437792e+00 6.26887858e-01
-4.84840602e-01 1.40107170e-01 -7.94084191e-01 3.93044770e-01
1.06308925e+00 -2.17817381e-01 2.96566665e-01 7.01577187e-01
8.36984098e-01 -3.41194153e-01 -9.66952443e-01 1.17945409e+00
-2.51709908e-01 -1.32834995e+00 2.86237746e-01 2.27400064e-01
5.55489302e-01 -4.88732189e-01 3.70840609e-01 -3.30211595e-02
-1.32162943e-01 -1.22545826e+00 7.89713562e-01 6.57245040e-01
5.08454144e-01 -9.43618536e-01 1.14578640e+00 2.18211189e-01
-9.68868315e-01 -4.59498286e-01 -2.00853407e-01 -2.67929524e-01
8.20287690e-02 9.77554262e-01 -1.30616641e+00 6.88032508e-01
2.85765469e-01 6.16058648e-01 -3.62953663e-01 1.00389135e+00
-2.87434340e-01 1.13563633e+00 -2.50313014e-01 -3.33122164e-01
2.20655292e-01 -4.44739312e-03 6.24501050e-01 9.06797111e-01
5.18159211e-01 -1.62369721e-02 -2.77945042e-01 8.56838346e-01
1.26548409e-01 -1.51588321e-01 -6.85796499e-01 -8.51502195e-02
8.06529641e-01 8.62726092e-01 -7.08816707e-01 -3.67686749e-01
-3.13626647e-01 5.40049851e-01 -4.87645268e-01 6.26145601e-01
-1.01126122e+00 -2.28072926e-01 5.45129716e-01 3.10578436e-01
-9.24222320e-02 -2.92717934e-01 -7.99909949e-01 -7.46846914e-01
-3.04055572e-01 -8.43329549e-01 1.74229890e-02 -6.75247312e-01
-6.83312595e-01 6.76365197e-01 1.30322501e-02 -1.47408342e+00
-2.95559227e-01 -4.37107444e-01 -4.98857588e-01 7.75092363e-01
-1.71320224e+00 -1.06497407e+00 -4.85381961e-01 5.60394764e-01
3.48431319e-01 -5.41537702e-01 7.99792349e-01 5.17405808e-01
-6.38518393e-01 7.97392309e-01 -5.41049242e-02 -8.39674845e-02
1.92753032e-01 -6.87886178e-01 -7.75961298e-03 9.93788064e-01
5.26127554e-02 5.82269907e-01 1.10427999e+00 -3.37131530e-01
-1.36151564e+00 -1.07915032e+00 6.27688169e-01 2.73930550e-01
2.45664060e-01 -1.47374988e-01 -2.03384027e-01 5.74356258e-01
-1.15473149e-02 -1.29499137e-01 4.63213384e-01 -8.86021331e-02
3.96206200e-01 -5.85428476e-01 -8.55549753e-01 5.56936145e-01
7.69900620e-01 -3.37239653e-01 2.33994469e-01 7.19766840e-02
-1.02694422e-01 -1.35792857e-02 -3.13278794e-01 4.47469205e-01
7.04024494e-01 -9.91452277e-01 7.34084606e-01 -3.66947621e-01
-5.89715689e-02 -4.57994223e-01 -2.86912620e-01 -1.21069241e+00
5.84990717e-02 -6.11803949e-01 -2.46503562e-01 6.80062175e-01
8.73928487e-01 -7.96704412e-01 9.04579222e-01 5.23685753e-01
-6.04289770e-02 -6.80568576e-01 -8.53070498e-01 -5.44957161e-01
-4.81423914e-01 -7.39433229e-01 7.29134858e-01 5.16854942e-01
-3.16770136e-01 1.92693084e-01 -7.18352616e-01 6.70263946e-01
7.72834897e-01 -5.73097989e-02 8.66325080e-01 -1.20724058e+00
-2.39464194e-01 -2.51845092e-01 -8.94949555e-01 -1.17645836e+00
-4.50209267e-02 -5.11843085e-01 -1.11430183e-01 -1.33067012e+00
-3.29306841e-01 -1.08434129e+00 -4.10686016e-01 1.28470108e-01
1.36089072e-01 3.21251228e-02 -1.12362042e-01 1.42508343e-01
-2.27195323e-01 5.71813881e-01 8.88692915e-01 -1.26024008e-01
-5.48898280e-02 6.81534052e-01 -7.73135304e-01 5.63802421e-01
1.28323519e+00 -2.47201160e-01 -8.57659221e-01 -1.97247446e-01
2.68026531e-01 2.80412942e-01 4.84651953e-01 -1.31215549e+00
2.90904194e-01 8.80872533e-02 4.09756541e-01 -4.93246943e-01
3.22912872e-01 -1.22719836e+00 2.36722127e-01 4.94812608e-01
-3.59613210e-01 -2.49574259e-01 -1.35508683e-02 6.20169997e-01
-3.52740079e-01 9.29373968e-03 4.34241086e-01 1.09851182e-01
-1.12901950e+00 1.94469109e-01 -7.75009334e-01 -4.06925291e-01
1.16532993e+00 -4.63359088e-01 3.83846350e-02 -1.22479153e+00
-7.80186236e-01 -3.13437283e-02 -3.37894350e-01 9.35664549e-02
7.28965759e-01 -1.09599078e+00 -5.85123181e-01 3.74688238e-01
1.76482484e-01 -3.33776325e-01 -3.58276814e-02 9.26386237e-01
-6.71195805e-01 9.74509597e-01 -2.07588390e-01 -6.91154003e-01
-1.05581284e+00 -5.74458344e-03 7.81190395e-01 -3.93461026e-02
-2.36466557e-01 5.96769571e-01 1.19689487e-01 2.07215875e-01
4.56052035e-01 -3.58098537e-01 -2.70040572e-01 -3.98557007e-01
6.05360977e-02 2.18172416e-01 1.39897782e-02 -4.95582700e-01
-3.05695295e-01 3.47383976e-01 2.22577453e-01 -9.63383839e-02
8.36457193e-01 -5.47924936e-01 3.74962389e-01 1.23865880e-01
8.29846382e-01 -5.94162829e-02 -9.12867606e-01 -5.11626108e-03
-4.82004136e-03 -2.83529937e-01 4.74269986e-01 -9.05582786e-01
-1.17181730e+00 1.17251527e+00 7.89064646e-01 3.98233123e-02
1.32099199e+00 -3.44405293e-01 5.40796280e-01 5.02026618e-01
7.40590394e-01 -9.03514802e-01 -6.30030692e-01 -1.38948653e-02
5.25549471e-01 -1.14748204e+00 7.39603937e-02 -6.44384503e-01
-4.11473304e-01 1.33317327e+00 3.16333294e-01 3.63154471e-01
1.02045298e+00 2.80589134e-01 2.43430912e-01 -1.57132789e-01
-6.79350972e-01 -2.22341254e-01 1.47894636e-01 6.10827088e-01
3.26770186e-01 3.73034656e-01 -1.96468979e-01 8.11816692e-01
-4.20760252e-02 -3.99588570e-02 6.22251451e-01 5.33906281e-01
-4.43401188e-01 -9.57485139e-01 -3.73356819e-01 5.91925383e-01
-8.00383985e-02 -7.24385381e-02 -3.77935737e-01 8.40252578e-01
1.18304834e-01 1.28596640e+00 -2.86401719e-01 -6.33049190e-01
-1.48837313e-01 -3.70847464e-01 1.84177421e-02 -5.69509994e-03
5.33377519e-03 -3.11911814e-02 5.06512582e-01 -1.99378684e-01
-4.93278950e-01 -3.38922411e-01 -1.34404528e+00 -1.87803149e-01
-5.14757156e-01 3.25900406e-01 7.58830965e-01 1.32933033e+00
4.17885810e-01 8.22669446e-01 6.86904490e-01 -3.75341237e-01
-3.33229691e-01 -6.17083430e-01 -6.81646705e-01 -2.03900099e-01
4.34566110e-01 -8.20081234e-01 -3.23807508e-01 -2.50367105e-01] | [6.220977306365967, 1.4668123722076416] |
628ce173-8088-4eba-a0c7-41e727088db5 | layout-guided-indoor-panorama-inpainting-with | 2301.05624 | null | https://arxiv.org/abs/2301.05624v1 | https://arxiv.org/pdf/2301.05624v1.pdf | Layout-guided Indoor Panorama Inpainting with Plane-aware Normalization | We present an end-to-end deep learning framework for indoor panoramic image inpainting. Although previous inpainting methods have shown impressive performance on natural perspective images, most fail to handle panoramic images, particularly indoor scenes, which usually contain complex structure and texture content. To achieve better inpainting quality, we propose to exploit both the global and local context of indoor panorama during the inpainting process. Specifically, we take the low-level layout edges estimated from the input panorama as a prior to guide the inpainting model for recovering the global indoor structure. A plane-aware normalization module is employed to embed plane-wise style features derived from the layout into the generator, encouraging local texture restoration from adjacent room structures (i.e., ceiling, floor, and walls). Experimental results show that our work outperforms the current state-of-the-art methods on a public panoramic dataset in both qualitative and quantitative evaluations. Our code is available at https://ericsujw.github.io/LGPN-net/ | ['Hung-Kuo Chu', 'Jheng-Wei Su', 'Cheng-Hsiu Chen', 'Chao-Chen Gao'] | 2023-01-13 | null | null | null | null | ['image-inpainting'] | ['computer-vision'] | [ 5.73422015e-01 -1.48993835e-01 3.78293917e-02 -4.79053319e-01
-7.53699601e-01 -2.79604435e-01 3.05603713e-01 -2.96816945e-01
2.18291253e-01 6.49506211e-01 6.40001655e-01 5.46197034e-02
6.38866937e-03 -1.13301432e+00 -1.21153295e+00 -4.68684137e-01
2.51807958e-01 -6.71126246e-02 -1.43596798e-01 -2.46480271e-01
2.10562423e-02 3.95894349e-01 -1.36149716e+00 4.83283162e-01
9.84651029e-01 8.61736774e-01 4.51582640e-01 8.14712346e-01
3.80909562e-01 8.96888316e-01 -2.45998129e-01 -2.41017863e-01
5.63027620e-01 -5.60772002e-01 -4.67780083e-01 4.30154681e-01
1.23196077e+00 -6.64390206e-01 -7.69000530e-01 9.03137028e-01
3.42438459e-01 1.94357067e-01 2.04691857e-01 -8.17597568e-01
-5.68761230e-01 -1.62798092e-02 -6.37381732e-01 -2.52142906e-01
5.36209047e-01 1.04459025e-01 8.68622005e-01 -8.55623662e-01
8.07247877e-01 1.22326529e+00 6.77956939e-01 7.48601779e-02
-1.17552245e+00 -6.10078692e-01 2.24673659e-01 9.50849429e-02
-1.11596918e+00 -4.55055892e-01 1.37466526e+00 -5.24930470e-02
4.17933375e-01 4.81349766e-01 6.47480726e-01 1.16887152e+00
6.28960490e-01 7.30411530e-01 1.49028325e+00 -4.89003658e-01
3.73472944e-02 -3.90144169e-01 -5.25465310e-01 1.00496864e+00
-2.49698862e-01 2.03865588e-01 -5.78689277e-01 5.73119260e-02
1.13536525e+00 4.06215489e-01 -5.71551502e-01 -6.99466944e-01
-9.95189011e-01 3.97467852e-01 8.39655340e-01 -4.78738472e-02
-3.70739937e-01 1.91748321e-01 -5.93744144e-02 1.82105795e-01
6.50037169e-01 1.98610365e-01 -2.22199440e-01 7.40011632e-02
-1.06033349e+00 4.38204139e-01 3.79028469e-01 9.68852520e-01
9.97916818e-01 -2.94330239e-01 -4.10166010e-02 1.21132994e+00
1.05165809e-01 5.71716249e-01 -7.13467598e-02 -1.26122022e+00
8.48557413e-01 4.74267513e-01 9.51571688e-02 -1.09717989e+00
-1.15013018e-01 -1.86805367e-01 -8.52846980e-01 2.51759410e-01
2.16935873e-01 1.40535012e-01 -1.07562959e+00 1.34375119e+00
2.91744441e-01 1.89742222e-01 -3.33650708e-01 8.57448041e-01
6.30757451e-01 8.09455276e-01 -5.22727430e-01 1.85121521e-01
1.17407787e+00 -1.58710778e+00 -7.45368898e-01 -6.74239755e-01
-1.69564993e-03 -1.32089984e+00 1.40642953e+00 4.24911171e-01
-1.05320084e+00 -5.59368670e-01 -1.04665363e+00 -6.04288876e-01
1.33853937e-02 2.49890968e-01 4.34098721e-01 4.56116110e-01
-7.60578871e-01 5.10003328e-01 -8.50007296e-01 -2.45194420e-01
3.88771355e-01 -6.14683181e-02 -4.29049671e-01 -9.11808431e-01
-8.02527606e-01 5.32727599e-01 -1.66918069e-01 1.42904222e-01
-8.05569470e-01 -7.53330052e-01 -1.08974981e+00 -1.10180769e-02
3.70288283e-01 -8.14857900e-01 1.17989779e+00 -9.61286068e-01
-1.59564757e+00 5.46772063e-01 -1.88545570e-01 -3.34996171e-03
7.01353371e-01 -6.46787643e-01 -1.79735690e-01 1.25203028e-01
2.35507637e-01 6.38234615e-01 8.61363530e-01 -1.62777591e+00
-4.32556927e-01 -2.18829811e-01 1.91783175e-01 6.34098828e-01
-1.17214248e-02 -4.56187427e-01 -7.14192569e-01 -9.65261161e-01
3.00851196e-01 -8.27137172e-01 -2.79349148e-01 3.61365080e-01
-5.57197213e-01 7.60823309e-01 1.00975597e+00 -1.22056127e+00
1.05279601e+00 -2.18126321e+00 7.23152086e-02 2.04400480e-01
-6.47445321e-02 -2.76193500e-01 2.28861067e-02 4.17738289e-01
9.25481170e-02 -4.38614577e-01 -6.54216647e-01 -8.39009881e-01
-2.16931701e-01 1.66359797e-01 -5.10502815e-01 4.36910301e-01
-1.45880461e-01 6.25566959e-01 -8.96480143e-01 -2.50031769e-01
8.44752014e-01 8.12519193e-01 -6.72095060e-01 4.03655857e-01
-1.40725821e-01 5.39281487e-01 -2.31238469e-01 8.71472299e-01
9.86952782e-01 7.41192997e-02 -1.11168250e-02 -1.51397973e-01
-6.24048077e-02 4.43960786e-01 -9.88196909e-01 2.23053813e+00
-1.17420089e+00 7.19961464e-01 3.43313992e-01 -2.48647034e-01
8.58059347e-01 -1.12366721e-01 2.20027179e-01 -8.12952697e-01
-1.54599056e-01 1.27546951e-01 -4.59754556e-01 -1.28918678e-01
7.77862608e-01 8.35740343e-02 -6.92639127e-02 1.61486104e-01
-1.88217461e-01 -6.83735132e-01 -2.74709642e-01 -1.54411485e-02
9.21670198e-01 4.49947000e-01 1.57501832e-01 -2.90761646e-02
4.09176886e-01 -4.19640720e-01 4.11605418e-01 4.78164405e-01
1.69799775e-01 1.55249321e+00 1.19967878e-01 -6.65207744e-01
-1.30907714e+00 -1.25145698e+00 -1.89067367e-02 8.35906148e-01
3.66252691e-01 -6.38962805e-01 -9.64996874e-01 -4.78586704e-01
-4.82674152e-01 6.11129582e-01 -6.93882763e-01 2.17327718e-02
-8.04942846e-01 -5.55775642e-01 -1.30211441e-02 3.68009835e-01
1.10812604e+00 -8.74569356e-01 -5.24724960e-01 2.93942094e-01
-7.01772690e-01 -1.20268917e+00 -8.45888257e-01 -3.93252522e-02
-8.86893451e-01 -1.02849579e+00 -8.43408346e-01 -7.82279730e-01
6.80150807e-01 4.23809826e-01 1.14334679e+00 -1.25543728e-01
-4.04773802e-01 1.97086811e-01 -1.50516897e-01 9.52193737e-02
-1.72969237e-01 -3.71018797e-02 -5.35869360e-01 -4.86306213e-02
-3.38589817e-01 -9.94825959e-01 -1.01671648e+00 3.67527544e-01
-9.18963194e-01 7.48905122e-01 7.04950690e-01 8.49209249e-01
7.43166804e-01 2.12091729e-01 -4.66527045e-01 -7.87393272e-01
1.69131011e-01 -9.18090641e-02 -5.21469772e-01 7.70330280e-02
-1.86898798e-01 -2.67716438e-01 8.22448909e-01 -4.37954515e-02
-1.35853541e+00 2.07348943e-01 -3.76098037e-01 -3.22308838e-01
-3.57604384e-01 -7.13291243e-02 -6.07803643e-01 -1.26184046e-01
3.59898686e-01 4.03310746e-01 -5.45357168e-01 -5.31340599e-01
2.68095940e-01 2.75511801e-01 5.88953614e-01 -5.96694171e-01
9.45548654e-01 8.80710304e-01 -1.11282848e-01 -9.98846233e-01
-1.04197431e+00 -1.19530395e-01 -5.57519257e-01 -2.21273720e-01
7.37598181e-01 -1.13464808e+00 7.67562613e-02 7.36700118e-01
-1.07519031e+00 -7.85748243e-01 -2.62956887e-01 1.10642754e-01
-6.98243558e-01 3.43230277e-01 -7.33214915e-01 -3.51560324e-01
-2.92381197e-01 -1.04356027e+00 1.50653780e+00 2.73110807e-01
-1.13619342e-02 -8.48100066e-01 2.14724183e-01 7.31445193e-01
2.77481854e-01 5.48679650e-01 6.97274685e-01 7.20967829e-01
-1.09357452e+00 4.93602417e-02 -1.70777723e-01 4.42578614e-01
5.00256360e-01 -1.32681638e-01 -1.16065300e+00 -9.82662216e-02
6.95016757e-02 -1.22013360e-01 8.68439853e-01 4.56117541e-01
1.38813639e+00 -4.83044237e-01 -1.45860776e-01 1.11495590e+00
1.44293284e+00 -1.42755792e-01 1.07106662e+00 5.58760047e-01
1.14286160e+00 5.52257001e-01 7.68086493e-01 2.93793321e-01
3.99939299e-01 7.25970030e-01 4.42658067e-01 -3.53828669e-01
-5.52265465e-01 -8.30425262e-01 4.26239252e-01 5.87758660e-01
-5.59778847e-02 -3.18408281e-01 -6.83355629e-01 3.76050711e-01
-1.67046273e+00 -6.82230592e-01 1.24039918e-01 2.18314648e+00
8.14534366e-01 8.47375393e-02 -4.94666368e-01 7.07759485e-02
3.70094776e-01 8.05161953e-01 -1.42994091e-01 -2.72389501e-01
-1.19549178e-01 3.77623975e-01 5.45406342e-01 1.04026532e+00
-1.23531210e+00 9.68985617e-01 4.67057276e+00 9.72946763e-01
-1.09111035e+00 6.15875907e-02 1.08403206e+00 3.31521258e-02
-2.40528494e-01 2.36343760e-02 -3.80479097e-01 3.81682962e-01
4.16485012e-01 5.80550611e-01 6.49016798e-01 9.70172822e-01
2.52422661e-01 -3.39908302e-01 -7.58752704e-01 9.82343912e-01
1.09884620e-01 -1.22846127e+00 -1.88947231e-01 1.29439577e-01
1.20804012e+00 -1.11458510e-01 2.18565002e-01 -1.21263735e-01
1.59028009e-01 -1.02642608e+00 7.78093278e-01 4.99357432e-01
8.70727837e-01 -8.43806922e-01 4.44774657e-01 5.25377244e-02
-1.26928937e+00 -3.69221158e-02 -2.57628113e-01 -1.83167353e-01
2.30827749e-01 8.67157042e-01 -5.18645883e-01 7.38565803e-01
9.38394725e-01 7.63152778e-01 -6.09811723e-01 9.43050742e-01
-6.32542729e-01 3.79325628e-01 -3.47562313e-01 7.93586671e-01
2.08312839e-01 -3.51308525e-01 3.64888430e-01 1.08108640e+00
4.57574338e-01 -1.62074596e-01 1.01970412e-01 9.18654382e-01
-2.86034107e-01 -4.69545014e-02 -7.18808413e-01 5.54623425e-01
2.10919216e-01 1.28530514e+00 -7.63281822e-01 -2.02554986e-01
-3.21016967e-01 1.45455861e+00 1.40840814e-01 4.05018300e-01
-7.86019444e-01 -3.90504748e-01 6.54035091e-01 4.99804795e-01
3.09379369e-01 -3.69505048e-01 -5.47415137e-01 -1.15713406e+00
3.14293116e-01 -1.00679684e+00 -4.57730219e-02 -9.34201360e-01
-9.49833453e-01 4.23334002e-01 -1.69713706e-01 -1.29529679e+00
2.16937900e-01 -3.22145253e-01 -7.39042342e-01 6.72838271e-01
-1.43734586e+00 -1.56665802e+00 -7.75764704e-01 5.62859476e-01
8.08830082e-01 4.50196892e-01 7.80674398e-01 4.60718781e-01
-3.14053714e-01 2.98303932e-01 2.57186770e-01 1.30466655e-01
9.25593257e-01 -1.12443233e+00 5.60267091e-01 1.24423504e+00
-1.61778450e-03 3.87568802e-01 6.90358996e-01 -5.54259300e-01
-1.31036294e+00 -1.34467149e+00 6.75706029e-01 -3.42554659e-01
-3.28142894e-04 -8.14935029e-01 -5.26740491e-01 6.46826506e-01
2.89825499e-01 1.45134866e-01 2.13284373e-01 -2.39511028e-01
-2.88303465e-01 -3.09374392e-01 -1.07282913e+00 9.38897073e-01
1.17909050e+00 -5.22858739e-01 -2.49968275e-01 2.73686051e-01
5.65177619e-01 -9.45051432e-01 -5.71961820e-01 4.57083285e-01
5.98741233e-01 -1.37962925e+00 1.26224387e+00 5.01873493e-01
1.11466873e+00 -4.45588678e-01 -3.49293888e-01 -1.37617815e+00
-2.37650052e-01 -6.44864082e-01 -1.24715544e-01 1.07876706e+00
-6.25795498e-02 -2.21251011e-01 8.21130276e-01 3.84025782e-01
-3.80771607e-01 -7.24542379e-01 -7.19992518e-01 -4.04443979e-01
-2.75486261e-01 -3.55259210e-01 6.63472831e-01 7.54485130e-01
-6.34282887e-01 1.89962491e-01 -7.46013522e-01 2.23617673e-01
6.69636190e-01 5.34238636e-01 1.06547952e+00 -5.39690733e-01
-4.35691893e-01 -1.20464914e-01 -6.32772073e-02 -1.31703556e+00
-6.87309653e-02 -3.20298374e-01 2.99052745e-01 -1.67043078e+00
-1.39356935e-02 -3.82089615e-01 3.93127277e-02 3.12842607e-01
-2.28676528e-01 6.13649249e-01 1.17022417e-01 -1.33772530e-02
-2.46144101e-01 8.67145181e-01 1.68155217e+00 -6.48440197e-02
-2.94773370e-01 2.88550816e-02 -4.71110731e-01 8.87307525e-01
8.01927447e-01 -3.78807127e-01 -3.90443116e-01 -5.76605558e-01
2.25631855e-02 -1.08636715e-01 5.57009876e-01 -1.31839192e+00
-1.72951698e-01 -2.32056648e-01 8.20387244e-01 -4.92643446e-01
7.58388937e-01 -9.77582216e-01 1.60187557e-01 2.77524561e-01
-7.56329075e-02 -1.45763904e-02 1.77668080e-01 5.39759338e-01
-2.72249311e-01 3.39035243e-01 8.57908845e-01 -1.45183474e-01
-5.11732459e-01 3.63877267e-01 -4.16774824e-02 -5.48440479e-02
5.26380420e-01 -1.03762090e-01 -2.34213114e-01 -6.06734753e-01
-1.91068485e-01 -1.13961235e-01 1.12363470e+00 5.15356481e-01
8.64600897e-01 -1.26292694e+00 -3.57016593e-01 4.50042933e-01
-7.07306862e-02 5.74582517e-01 6.16721034e-01 6.05037391e-01
-1.15507936e+00 3.93173471e-02 -4.28331435e-01 -4.53468591e-01
-1.28147733e+00 1.49452612e-01 3.25912297e-01 -4.28255677e-01
-9.60956693e-01 8.53115559e-01 6.42812431e-01 -7.12214768e-01
1.27133057e-01 -7.15541840e-01 3.79031807e-01 -5.08469522e-01
3.56011987e-01 3.26381803e-01 8.07650760e-02 -5.53729475e-01
-1.59778818e-01 7.41261721e-01 -2.44761910e-02 -2.04592153e-01
1.51939535e+00 -3.30827296e-01 -1.29529476e-01 2.31249049e-01
1.13283110e+00 5.27596891e-01 -1.95925975e+00 -1.24546289e-01
-5.29646456e-01 -1.12336838e+00 5.97800687e-02 -7.21940935e-01
-1.06153440e+00 9.05685306e-01 7.05047607e-01 -5.84649265e-01
1.40653682e+00 -3.87857109e-01 1.15500069e+00 8.59843343e-02
4.03654844e-01 -1.04979455e+00 1.70822337e-01 4.19391096e-01
1.17925894e+00 -1.01823103e+00 3.33862036e-01 -7.98357129e-01
-3.30987662e-01 1.15690625e+00 6.99728847e-01 -4.76443976e-01
5.02180219e-01 3.11094761e-01 4.99415286e-02 1.09808885e-01
-2.16658428e-01 4.17290717e-01 2.99247235e-01 4.58187044e-01
3.08652222e-01 2.18034327e-01 2.64052898e-01 1.64545760e-01
-6.41760230e-01 -1.62280187e-01 5.64454973e-01 1.02491033e+00
-1.57036707e-01 -1.14343369e+00 -7.85761833e-01 1.40675738e-01
-1.60585091e-01 -3.07283401e-01 -3.24145108e-01 7.08508193e-01
3.76587600e-01 6.86345696e-01 -1.31193042e-01 -2.57439405e-01
4.00115073e-01 -3.63636404e-01 6.53133690e-01 -5.13213694e-01
-2.43238464e-01 2.14191839e-01 1.11358933e-01 -1.05217242e+00
-2.10612252e-01 -4.64970499e-01 -6.58886313e-01 -3.09388906e-01
1.70369297e-01 -4.22766358e-01 4.57767546e-01 5.41042149e-01
2.80151159e-01 6.65767729e-01 6.06514394e-01 -1.35650206e+00
3.13095748e-01 -8.47307563e-01 -4.42873865e-01 2.89076328e-01
5.75435221e-01 -4.63214606e-01 -1.83012962e-01 1.59059167e-01] | [10.196840286254883, -2.342907667160034] |
3037ca55-1647-473d-a578-6adfa9c89c33 | esta-an-esports-trajectory-and-action-dataset | 2209.09861 | null | https://arxiv.org/abs/2209.09861v1 | https://arxiv.org/pdf/2209.09861v1.pdf | ESTA: An Esports Trajectory and Action Dataset | Sports, due to their global reach and impact-rich prediction tasks, are an exciting domain to deploy machine learning models. However, data from conventional sports is often unsuitable for research use due to its size, veracity, and accessibility. To address these issues, we turn to esports, a growing domain that encompasses video games played in a capacity similar to conventional sports. Since esports data is acquired through server logs rather than peripheral sensors, esports provides a unique opportunity to obtain a massive collection of clean and detailed spatiotemporal data, similar to those collected in conventional sports. To parse esports data, we develop awpy, an open-source esports game log parsing library that can extract player trajectories and actions from game logs. Using awpy, we parse 8.6m actions, 7.9m game frames, and 417k trajectories from 1,558 game logs from professional Counter-Strike tournaments to create the Esports Trajectory and Actions (ESTA) dataset. ESTA is one of the largest and most granular publicly available sports data sets to date. We use ESTA to develop benchmarks for win prediction using player-specific information. The ESTA data is available at https://github.com/pnxenopoulos/esta and awpy is made public through PyPI. | ['Claudio Silva', 'Peter Xenopoulos'] | 2022-09-20 | null | null | null | null | ['log-parsing'] | ['computer-code'] | [-4.59674001e-01 -5.54481268e-01 -6.07514918e-01 5.52623868e-02
-1.14858532e+00 -8.22009683e-01 4.43361133e-01 4.85808961e-02
-5.59344649e-01 5.53600311e-01 6.21261120e-01 -1.33462831e-01
-2.27728814e-01 -1.05009234e+00 -6.91695094e-01 -1.56393990e-01
-1.17207669e-01 4.39447612e-01 6.57939196e-01 -5.35284162e-01
2.05176309e-01 2.40206555e-01 -1.65292561e+00 5.36694467e-01
2.59805679e-01 8.07525098e-01 -1.42823666e-01 9.01414752e-01
2.02083662e-01 1.03163838e+00 -3.07532549e-01 -5.64168930e-01
5.73052824e-01 -3.11495185e-01 -7.90982783e-01 -5.46696901e-01
1.12259090e-01 -2.99012303e-01 -8.78412545e-01 4.79865134e-01
4.90372896e-01 3.17651570e-01 2.11435594e-02 -1.33902919e+00
-1.14058882e-01 7.52528608e-01 -5.05005479e-01 8.21607530e-01
4.75739866e-01 5.87528110e-01 1.11278021e+00 -4.76593822e-01
8.65667462e-01 5.61451614e-01 1.04068911e+00 3.39999646e-01
-8.82179558e-01 -8.71445656e-01 -2.00999096e-01 4.06842649e-01
-1.22135794e+00 -4.61897403e-01 4.84555542e-01 -5.63044786e-01
8.59192431e-01 4.54028934e-01 1.20801008e+00 1.44789350e+00
1.56088620e-01 8.63260150e-01 9.60146189e-01 1.50425375e-01
1.89909041e-01 -7.09282756e-01 2.90619910e-01 1.96413249e-01
-9.45543200e-02 2.24025711e-01 -1.11901498e+00 -1.58686534e-01
5.66695750e-01 5.88527471e-02 3.11269194e-01 7.74725825e-02
-9.90010679e-01 6.95159197e-01 9.71843079e-02 -3.53322476e-01
-5.52493274e-01 4.40938950e-01 7.47239113e-01 1.25891849e-01
5.04068792e-01 1.31921962e-01 -2.32376829e-01 -1.53876460e+00
-7.34412909e-01 9.65974927e-01 7.04350650e-01 6.10480428e-01
3.74252319e-01 2.88374331e-02 8.42010826e-02 8.28134000e-01
-1.29492328e-01 4.22816157e-01 2.97941118e-01 -1.23662150e+00
7.32159555e-01 5.24997413e-01 -1.96568772e-01 -9.81422484e-01
-4.92985994e-01 -1.03879310e-01 -1.93114102e-01 -9.88984760e-03
7.32901514e-01 -7.93130249e-02 -5.12739241e-01 1.54538131e+00
4.05077964e-01 4.18714970e-01 -3.24465305e-01 1.15425467e+00
9.83917832e-01 4.13047075e-01 2.46838450e-01 4.14075881e-01
1.32981300e+00 -6.60157621e-01 -1.75948799e-01 -3.51780891e-01
5.67417026e-01 -4.39503938e-01 1.20407629e+00 5.38099527e-01
-1.11990476e+00 -2.28801861e-01 -5.02742290e-01 -7.35208439e-03
-2.64654011e-01 -3.01842958e-01 9.79345620e-01 2.62980163e-01
-4.57649142e-01 5.80104947e-01 -1.21983254e+00 -4.26840335e-01
6.04588032e-01 3.18316855e-02 -4.68194157e-01 1.03138760e-01
-1.18351638e+00 3.75740945e-01 3.41429591e-01 -3.65085661e-01
-8.20107281e-01 -9.97120142e-01 -7.50295401e-01 -1.87703326e-01
7.79692113e-01 -6.97643459e-02 1.35615504e+00 -1.52773529e-01
-1.23360622e+00 8.17782581e-01 4.36332881e-01 -5.76538980e-01
5.56466818e-01 -4.57522511e-01 -2.94499695e-01 -4.75508422e-02
5.68851233e-01 4.13673222e-02 -9.88502651e-02 -3.54298830e-01
-9.08006191e-01 -4.24404681e-01 6.85033798e-02 2.22093761e-01
1.68598052e-02 4.12540138e-01 -8.25534523e-01 -5.89799941e-01
-1.28332883e-01 -1.14717078e+00 -2.11712033e-01 -6.59179628e-01
-5.19685745e-01 -1.22508094e-01 2.74928778e-01 -7.84303665e-01
1.48564243e+00 -2.14534593e+00 -4.19685319e-02 -2.43199281e-02
2.86103100e-01 1.78499162e-01 8.53585601e-02 9.48695898e-01
1.28449872e-01 -7.86739364e-02 2.75195748e-01 -1.39620468e-01
1.05352595e-01 3.30011070e-01 -2.32481852e-01 4.91039753e-01
-4.05524313e-01 9.41215217e-01 -8.48044753e-01 -3.96498442e-01
1.84244037e-01 -1.50999054e-01 -8.88524413e-01 -1.85146138e-01
-8.25187117e-02 4.36072111e-01 -5.81746936e-01 8.53092372e-01
2.73338139e-01 1.85406402e-01 -4.72761206e-02 3.32510710e-01
-4.85929459e-01 6.69162333e-01 -1.12041807e+00 1.89573824e+00
7.78952241e-02 6.80137873e-01 -2.31107604e-02 -9.56090450e-01
5.78048944e-01 -7.13001043e-02 1.14854133e+00 -1.02212656e+00
1.22767806e-01 6.49681091e-02 -2.34375432e-01 -6.23806953e-01
9.05814111e-01 1.08662926e-01 -8.79864335e-01 6.68671787e-01
-4.37046103e-02 2.16993868e-01 7.08969533e-01 2.40981460e-01
1.77814043e+00 3.11069429e-01 7.14952592e-04 1.55984923e-01
-2.82327294e-01 7.62664974e-01 1.15297449e+00 7.88960993e-01
-3.81150275e-01 5.77953398e-01 5.01433074e-01 -7.16682732e-01
-1.07600975e+00 -1.33314490e+00 7.10421577e-02 1.50974607e+00
-7.02490509e-02 -1.03225851e+00 -4.47284222e-01 -1.26539320e-01
1.29280239e-01 1.34893417e-01 -3.33773106e-01 -5.45562692e-02
-9.42705035e-01 -6.95831299e-01 1.10988295e+00 7.03164697e-01
6.47592545e-01 -1.03794622e+00 -5.23379982e-01 4.23510373e-01
-6.36547267e-01 -1.17685378e+00 -3.44524771e-01 -2.16371357e-01
-5.09813428e-01 -1.48504114e+00 -1.02658018e-01 -9.59381908e-02
-6.94020331e-01 1.77237734e-01 1.45048296e+00 -2.55052358e-01
-5.80758035e-01 3.94406289e-01 -5.41315794e-01 -6.47558868e-01
5.46115218e-03 5.34667075e-01 2.40109757e-01 -3.41192126e-01
9.21578765e-01 -8.49183679e-01 -5.46254218e-01 2.09519759e-01
-5.66257834e-01 2.24314973e-01 1.34885848e-01 2.31288210e-01
9.36454892e-01 -1.82194665e-01 3.97467256e-01 -7.40329862e-01
5.44287801e-01 -1.15424573e+00 -5.14091313e-01 -4.75479037e-01
-2.08564788e-01 -7.03163445e-01 3.30748916e-01 -3.04253608e-01
-5.87805688e-01 -9.17417556e-02 -3.33806694e-01 -3.83610904e-01
-3.26109648e-01 7.55066574e-01 2.53740072e-01 5.94254196e-01
9.08606052e-01 2.79076219e-01 7.58517981e-02 -8.19802344e-01
2.37091511e-01 6.73624814e-01 1.00564837e+00 -7.89320350e-01
6.05180204e-01 4.67125118e-01 -2.24510416e-01 -8.45137298e-01
-7.62315631e-01 -7.70534456e-01 -3.85216683e-01 -6.79573715e-01
9.45827782e-01 -1.18091655e+00 -1.18237758e+00 6.65533900e-01
-2.98420310e-01 -6.36764824e-01 -5.47823668e-01 6.55730784e-01
-7.51599610e-01 -4.65790592e-02 -1.10942650e+00 -7.37822771e-01
-2.17275292e-01 -7.40633249e-01 8.36282611e-01 3.07472050e-01
-4.48316336e-01 -4.34561193e-01 6.26245379e-01 9.24036562e-01
3.32075357e-02 5.57897866e-01 2.07834721e-01 -7.00139523e-01
-5.31372488e-01 -4.42473024e-01 1.30861104e-01 -2.07220376e-01
-2.51106858e-01 4.45272494e-03 -5.56892335e-01 6.01429977e-02
-6.17076576e-01 -4.93904263e-01 6.86444819e-01 4.77748930e-01
1.13091969e+00 4.24458459e-02 -1.35015860e-01 8.04918706e-01
1.12529206e+00 1.89112907e-05 6.33160532e-01 9.30518866e-01
7.05476999e-01 2.58420914e-01 8.59420359e-01 8.39880526e-01
7.92774498e-01 9.04389441e-01 4.14106309e-01 2.65679240e-01
-9.51940119e-02 -6.83969736e-01 2.82847911e-01 8.13735127e-01
-7.35385120e-01 -3.53322066e-02 -1.18456340e+00 6.59955919e-01
-2.01911378e+00 -1.66264462e+00 -6.05126798e-01 1.92432868e+00
6.74996972e-01 2.77859271e-01 1.01429486e+00 1.69275746e-01
2.67349869e-01 2.70994037e-01 -4.67889607e-01 -3.24439824e-01
-3.87768261e-02 4.48916227e-01 8.49263430e-01 -1.20081313e-01
-1.24403191e+00 1.09462535e+00 6.14465904e+00 1.15069079e+00
-6.64783239e-01 2.70443857e-01 1.33985832e-01 -1.07311583e+00
1.54248804e-01 1.68326534e-02 -8.02304864e-01 8.45077455e-01
1.30799913e+00 -2.86852360e-01 4.71305370e-01 9.99741793e-01
4.86933827e-01 -2.63796449e-01 -5.56486726e-01 1.06739759e+00
-5.15175462e-01 -1.79709911e+00 -6.33217156e-01 6.68309867e-01
2.09570199e-01 7.51953363e-01 -7.73968175e-02 8.45419705e-01
8.23925436e-01 -9.44192648e-01 1.00088465e+00 4.96660888e-01
7.03791976e-01 -9.45192218e-01 4.41726983e-01 5.66431940e-01
-1.44980383e+00 -2.73177117e-01 -1.89128518e-01 -7.79727519e-01
4.30336714e-01 2.10751757e-01 -2.30881855e-01 4.75883216e-01
1.36497009e+00 1.06478059e+00 -4.40218270e-01 1.02641809e+00
3.13743472e-01 1.19494069e+00 -5.42060256e-01 2.14842558e-01
2.96060383e-01 -4.12037939e-01 6.60776496e-01 9.28386688e-01
2.17802450e-01 4.96935308e-01 3.60224843e-01 4.07838732e-01
-9.76517722e-02 -3.58763635e-02 -7.05811143e-01 -1.93814427e-01
7.92807758e-01 1.04878473e+00 -5.59288085e-01 -2.96727326e-02
-5.29980719e-01 3.94083440e-01 3.51235777e-01 -1.11933656e-01
-1.16508055e+00 -7.62534365e-02 1.53202689e+00 6.43150926e-01
-1.86698660e-01 -4.07063872e-01 -2.39756584e-01 -1.12590468e+00
-9.76814888e-03 -1.07926786e+00 8.12062860e-01 -4.95410383e-01
-1.25202894e+00 2.12963015e-01 2.13621095e-01 -1.40818250e+00
-4.23024029e-01 -2.73307174e-01 -8.16537797e-01 5.55986583e-01
-7.20335662e-01 -1.10147607e+00 -3.70119840e-01 6.84961438e-01
5.93609869e-01 -2.56261021e-01 3.46013695e-01 5.94213307e-01
-9.59503829e-01 3.33360821e-01 -1.07826618e-03 5.21656871e-01
4.57527339e-01 -9.17039037e-01 7.43320405e-01 7.13973343e-01
1.74710840e-01 6.24416582e-02 5.63618839e-01 -8.43739033e-01
-1.64054990e+00 -9.16646838e-01 3.16330612e-01 -1.00115001e+00
1.13492644e+00 -3.58862936e-01 -6.53773606e-01 9.91836607e-01
-2.71110952e-01 -4.30216268e-02 1.11481595e+00 4.63851333e-01
-4.84714136e-02 -6.55180216e-02 -4.66339976e-01 6.73785686e-01
1.49776256e+00 -5.26651621e-01 -5.02086878e-01 8.86051208e-02
1.45824432e-01 -7.58351624e-01 -1.22692084e+00 6.83168918e-02
7.37846076e-01 -1.14302754e+00 1.14305651e+00 -9.08973157e-01
6.89655542e-01 9.08375829e-02 -2.14000806e-01 -1.04571128e+00
-3.11518490e-01 -3.60143781e-01 -5.50230267e-03 1.31401980e+00
1.79837704e-01 -2.67864406e-01 1.27426207e+00 8.71187925e-01
-3.37760955e-01 -6.95804656e-01 -1.08393991e+00 -8.80442679e-01
2.75164813e-01 -1.24825740e+00 8.69892836e-01 7.61957884e-01
2.57366359e-01 -4.92464565e-02 -5.76620877e-01 -1.98432371e-01
7.23919451e-01 2.50898689e-01 1.36431730e+00 -1.27639830e+00
-5.65323889e-01 -3.32657427e-01 -6.97326660e-01 -7.15910375e-01
-7.48059601e-02 -8.80179822e-01 -1.61172181e-01 -1.30920994e+00
2.45799974e-01 -6.86810791e-01 4.81800325e-02 6.29259408e-01
2.27326781e-01 7.20072448e-01 3.68431002e-01 4.76592779e-01
-7.86260009e-01 2.66795784e-01 7.74442971e-01 2.71459192e-01
-2.92884916e-01 2.92484373e-01 -4.69698697e-01 8.22525620e-01
9.48216498e-01 -5.11888921e-01 -1.53890595e-01 -3.40079963e-01
4.18713778e-01 2.72669852e-01 4.17270362e-01 -1.24586010e+00
1.26979440e-01 -5.75041056e-01 8.51145834e-02 -7.03475118e-01
6.80984795e-01 -6.32574484e-02 5.40804029e-01 2.27820382e-01
-7.93737397e-02 9.78271142e-02 2.85752475e-01 6.23231888e-01
-2.85382092e-01 2.24391103e-01 8.51526633e-02 -2.31158093e-01
-1.01579785e+00 7.09519148e-01 -8.75534892e-01 7.03819215e-01
1.10185313e+00 -4.20725137e-01 -4.30928111e-01 -3.88334483e-01
-8.98508966e-01 3.12945276e-01 5.03361046e-01 6.15441561e-01
1.51487812e-01 -1.42454422e+00 -8.64053488e-01 -8.21430162e-02
2.36686558e-01 -2.24069878e-01 7.29571700e-01 8.46934676e-01
-7.55864680e-01 9.79468152e-02 -4.52114850e-01 -4.65588570e-01
-1.23713207e+00 2.36221001e-01 -4.01709974e-02 -2.45098636e-01
-1.22034526e+00 5.30464590e-01 -4.20494825e-01 -5.24947703e-01
-2.85356015e-01 1.63574051e-02 -2.39478812e-01 4.33537737e-02
6.85839236e-01 7.57247746e-01 -1.50850276e-03 -7.88997293e-01
-3.87772173e-01 1.44490421e-01 4.23206538e-01 -1.61870629e-01
1.72828460e+00 6.90260828e-02 3.90085757e-01 5.81622243e-01
6.11094832e-01 2.61356384e-01 -1.45356941e+00 -1.55309394e-01
8.81013423e-02 -7.21400917e-01 -1.49449691e-01 -2.28892952e-01
-1.18584824e+00 4.35498774e-01 2.14958102e-01 3.59331876e-01
8.95327210e-01 1.30476832e-01 1.21486485e+00 -2.16629848e-01
7.60298491e-01 -1.35539043e+00 -9.34357196e-02 7.37889349e-01
3.35821450e-01 -1.00550306e+00 -2.13469088e-01 -1.28157422e-01
-8.86040330e-01 5.34375548e-01 7.46220827e-01 -3.28600556e-01
6.21560991e-01 5.06510139e-01 -6.99910894e-02 -4.42048579e-01
-1.01361740e+00 -3.60186338e-01 -1.58128679e-01 5.07159710e-01
8.14583302e-02 2.31937349e-01 -3.38125736e-01 1.32870579e+00
-8.84521604e-01 3.15694541e-01 7.32229233e-01 1.05409193e+00
-2.45016724e-01 -1.01020575e+00 -3.35289955e-01 7.31593668e-01
-7.83162832e-01 9.70004946e-02 -1.76807240e-01 8.98084164e-01
7.29455799e-02 9.04876411e-01 3.32519978e-01 -9.01681960e-01
6.88520432e-01 -8.98950547e-02 7.72523731e-02 -5.94835162e-01
-7.88477898e-01 -2.75692552e-01 4.89621133e-01 -1.23604190e+00
6.23780079e-02 -9.14793730e-01 -1.18748522e+00 -1.24361825e+00
3.76467258e-01 2.05002695e-01 4.89032060e-01 7.49875546e-01
5.98902643e-01 3.60254288e-01 1.42532945e-01 -9.43572879e-01
-9.51463953e-02 -8.37695837e-01 -8.55423689e-01 5.28522372e-01
-3.90473157e-01 -8.13947439e-01 1.18665263e-01 -1.61214754e-01] | [6.61002779006958, 0.35257670283317566] |
c79b85a7-57ee-4296-b7d5-5d6f5e3b4fcc | modeling-transitions-of-focal-entities-for | null | null | https://aclanthology.org/2021.acl-long.255 | https://aclanthology.org/2021.acl-long.255.pdf | Modeling Transitions of Focal Entities for Conversational Knowledge Base Question Answering | Conversational KBQA is about answering a sequence of questions related to a KB. Follow-up questions in conversational KBQA often have missing information referring to entities from the conversation history. In this paper, we propose to model these implied entities, which we refer to as the focal entities of the conversation. We propose a novel graph-based model to capture the transitions of focal entities and apply a graph neural network to derive a probability distribution of focal entities for each question, which is then combined with a standard KBQA module to perform answer ranking. Our experiments on two datasets demonstrate the effectiveness of our proposed method. | ['Jing Jiang', 'Yunshi Lan'] | 2021-08-01 | null | null | null | acl-2021-5 | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-1.45762697e-01 6.34603798e-01 7.24869072e-02 -5.66405952e-01
-7.01532364e-01 -5.09379268e-01 5.73204637e-01 2.47461259e-01
-2.02567503e-01 1.03323603e+00 7.69274294e-01 -3.76003951e-01
-1.33515298e-01 -1.01911485e+00 -5.99782407e-01 -1.17037885e-01
1.97519273e-01 8.84505570e-01 6.93792880e-01 -6.99620903e-01
1.08377680e-01 -3.70303690e-02 -8.36588681e-01 5.39378405e-01
1.18898630e+00 7.29004741e-01 6.11166656e-02 8.90351355e-01
-7.87596345e-01 1.64092481e+00 -9.51298475e-01 -1.02621853e+00
-4.95126158e-01 -8.06172132e-01 -1.64017081e+00 -6.03322536e-02
7.82995373e-02 -5.80194712e-01 -6.63529992e-01 9.03121591e-01
3.10043305e-01 5.49814582e-01 6.50148332e-01 -1.26318038e+00
-8.65979671e-01 9.52428639e-01 5.38447388e-02 4.42114145e-01
9.38911557e-01 -7.09201321e-02 1.50793159e+00 -5.37581146e-01
6.58458233e-01 1.50373149e+00 3.75349492e-01 6.42586350e-01
-5.76237857e-01 -2.41438463e-01 4.25450951e-01 6.91052556e-01
-1.09322309e+00 -2.46895820e-01 1.18638813e+00 -1.10049307e-01
1.00280941e+00 1.52087674e-01 7.03205347e-01 7.61568785e-01
1.25989914e-01 8.88616562e-01 5.09156108e-01 -1.96246728e-01
-4.94054742e-02 -8.25346820e-03 7.18745589e-01 8.32092881e-01
-1.56531692e-01 -8.52500916e-01 -4.30832684e-01 -6.03109777e-01
2.23770082e-01 -2.50136256e-01 -5.35816550e-01 -3.01122397e-01
-6.70327008e-01 1.07656193e+00 6.38456285e-01 1.45562783e-01
-4.59776044e-01 1.41297981e-01 2.67058820e-01 5.04741609e-01
5.17638028e-01 3.31232816e-01 -3.93311948e-01 7.42265955e-02
1.27308434e-02 2.36622632e-01 1.48792648e+00 8.90786231e-01
1.01536679e+00 -6.38032734e-01 -4.13837373e-01 1.00337434e+00
5.38142025e-01 1.44396707e-01 1.77064106e-01 -9.52158391e-01
7.96069503e-01 1.18748343e+00 4.09029931e-01 -1.31043446e+00
-3.98336411e-01 -2.90045086e-02 -5.68272233e-01 -1.12923205e+00
7.44123831e-02 -3.56200665e-01 -5.02243757e-01 1.67959630e+00
6.77983582e-01 4.44387436e-01 4.62322980e-01 7.88772941e-01
1.41954768e+00 9.55098689e-01 8.68880674e-02 -2.52853513e-01
1.23314917e+00 -1.30915892e+00 -1.04416800e+00 -2.24030524e-01
7.18362629e-01 -2.27634266e-01 9.58315969e-01 -3.36217940e-01
-8.30032289e-01 -2.65677184e-01 -6.41341746e-01 -4.19943541e-01
-2.16900393e-01 -4.00786400e-01 5.25547504e-01 2.34484702e-01
-1.19049394e+00 1.32337078e-01 -4.24063265e-01 -2.83469737e-01
6.75766394e-02 3.85491043e-01 2.16268953e-02 2.36996096e-02
-1.94329727e+00 1.09162068e+00 4.16032851e-01 4.79886979e-01
-7.33022749e-01 -3.07816327e-01 -9.65791523e-01 2.56781429e-01
6.87931597e-01 -1.26806593e+00 1.69398296e+00 -5.12840986e-01
-1.72246885e+00 4.45871741e-01 -4.65141118e-01 -3.68481785e-01
-6.15807027e-02 -1.55762240e-01 -6.64133728e-01 3.77608031e-01
3.35701741e-02 3.52666885e-01 4.32908386e-01 -1.33624780e+00
-8.39353144e-01 -9.03256461e-02 8.95368814e-01 6.54587388e-01
-2.16641575e-01 -3.25164169e-01 -7.53605366e-01 4.02346104e-02
-1.22197106e-01 -6.54318035e-01 -1.19417205e-01 -8.06489885e-01
-5.47044814e-01 -9.29243982e-01 6.76643193e-01 -9.05088305e-01
1.55828905e+00 -1.38663447e+00 4.79803503e-01 2.23018173e-02
6.09390140e-01 -8.58239597e-04 -1.29031509e-01 8.26905608e-01
4.25101936e-01 -5.14262430e-02 -2.07515106e-01 -5.62027283e-02
-4.72809635e-02 3.55952471e-01 -5.49386203e-01 -5.78404553e-02
2.96249270e-01 1.30924547e+00 -1.19640648e+00 -6.59675479e-01
-3.33577871e-01 9.38847736e-02 -3.59855831e-01 5.78742087e-01
-7.62297273e-01 2.46117249e-01 -9.26307321e-01 2.90614903e-01
3.70016634e-01 -5.70081651e-01 3.07795107e-01 -1.83353335e-01
7.48783827e-01 7.86720335e-01 -6.64086521e-01 1.53446841e+00
-3.51670474e-01 3.53296965e-01 2.22155955e-02 -8.47889602e-01
8.51776302e-01 2.22713217e-01 -1.21371612e-01 -4.75076377e-01
6.61663115e-02 -1.11877173e-01 4.30094264e-02 -8.42472851e-01
7.97307253e-01 -1.36532053e-01 -2.19345227e-01 5.92114627e-01
2.58525133e-01 -4.42764640e-01 4.41090055e-02 8.92857969e-01
1.14370847e+00 -2.46747017e-01 2.44316161e-01 1.54200226e-01
1.04296529e+00 1.91300586e-01 2.19571814e-01 7.43211448e-01
-2.89633900e-01 -1.24417908e-01 8.56614172e-01 -1.90359458e-01
-3.29314500e-01 -8.57850313e-01 6.51486874e-01 1.01107419e+00
4.90595132e-01 -5.05089641e-01 -5.45829892e-01 -1.26435459e+00
-7.97982514e-02 9.16021883e-01 -6.07045233e-01 -5.41735530e-01
-7.32910514e-01 -2.09297389e-01 3.41542751e-01 2.51174957e-01
5.05732715e-01 -1.32214296e+00 2.00699031e-01 3.04423243e-01
-8.67130160e-01 -1.18358719e+00 -5.69973707e-01 -3.68898451e-01
-6.65865839e-01 -1.12933636e+00 -4.74343866e-01 -9.38193262e-01
3.21758777e-01 2.16644317e-01 1.47280896e+00 3.88945609e-01
5.01509190e-01 9.23995376e-01 -4.88520622e-01 -1.04783677e-01
-4.93525654e-01 3.68834585e-01 -3.62349361e-01 1.91330627e-01
3.94909710e-01 -4.49393153e-01 -8.78432214e-01 1.14723824e-01
-7.29478300e-01 -2.84864753e-01 3.06162983e-01 6.99653506e-01
1.38793498e-01 -2.73361176e-01 1.13638270e+00 -1.28299320e+00
1.49389994e+00 -8.59634519e-01 -1.62873849e-01 7.92606235e-01
-1.56941086e-01 3.73234637e-02 5.36401331e-01 -3.12745482e-01
-1.40701091e+00 -4.16784078e-01 -1.16888970e-01 3.07773966e-02
1.93985298e-01 1.09698462e+00 -2.83270776e-01 1.87277034e-01
3.05125266e-01 1.53199092e-01 -3.53122443e-01 -6.47553131e-02
7.77762353e-01 7.28524506e-01 4.15251195e-01 -5.56488872e-01
4.89992023e-01 1.95507526e-01 -4.87909079e-01 -5.40320098e-01
-1.29548419e+00 -7.78005123e-01 -3.50752056e-01 -4.16756362e-01
7.55001307e-01 -6.53622031e-01 -9.86812890e-01 3.83499146e-01
-1.40230703e+00 -2.95371890e-01 -9.72443372e-02 1.07274733e-01
-3.77970159e-01 6.44241393e-01 -8.90446484e-01 -9.27113831e-01
-5.40138006e-01 -6.55671418e-01 7.81443000e-01 5.39245307e-01
-1.79359227e-01 -1.41946995e+00 4.51076537e-01 8.45691025e-01
1.42738059e-01 -3.83737683e-01 1.13275683e+00 -9.00611579e-01
-5.72098494e-01 -8.91704708e-02 -2.06483200e-01 5.10109775e-02
3.50443691e-01 -3.02885354e-01 -5.80569148e-01 1.62643328e-01
-1.38312802e-02 -4.35102224e-01 8.58974814e-01 1.42230121e-02
7.76116431e-01 -5.34013748e-01 -4.16924804e-01 -1.16593763e-01
8.45190704e-01 1.70120373e-01 5.52448750e-01 -9.28240865e-02
7.80882537e-01 9.32836711e-01 5.56265414e-01 1.45872250e-01
1.33245540e+00 4.01270032e-01 2.98433959e-01 3.46277475e-01
1.09281570e-01 -4.57690388e-01 1.98469654e-01 1.64003742e+00
4.26575005e-01 -7.46479273e-01 -9.63176489e-01 8.36613536e-01
-2.10617709e+00 -7.74866283e-01 -3.05338711e-01 1.43112016e+00
1.00733197e+00 -2.42664088e-02 -5.02071381e-02 -5.88224351e-01
9.04734075e-01 2.07779303e-01 -6.57829523e-01 -2.65349269e-01
-2.28465535e-02 -6.96325768e-03 -3.46177071e-01 1.07610226e+00
-7.71171927e-01 1.13544953e+00 6.14462471e+00 5.10307014e-01
-4.18835074e-01 -8.60800222e-02 4.24239188e-01 5.70255160e-01
-7.45557845e-01 2.24642009e-01 -6.85787320e-01 1.31073222e-01
9.50230122e-01 -6.98386371e-01 2.69052327e-01 4.36459005e-01
-1.59370512e-01 -5.80498613e-02 -9.08465505e-01 4.89779413e-01
3.87436569e-01 -1.36542189e+00 6.07538164e-01 -5.00814378e-01
5.35594702e-01 -3.31462979e-01 -5.43749630e-01 8.56988311e-01
7.84720719e-01 -6.87615812e-01 6.80671483e-02 9.39373136e-01
-7.34012667e-03 -8.94563198e-01 9.92255390e-01 4.98924613e-01
-1.20410299e+00 -6.75814524e-02 -5.15083075e-01 -7.44770691e-02
4.04851496e-01 2.52189279e-01 -1.27780569e+00 1.14254618e+00
4.93900418e-01 4.59672511e-01 -4.08393741e-01 7.76194513e-01
-6.68407679e-01 7.95466781e-01 6.29594258e-04 -7.37720609e-01
2.26918399e-01 -3.47564012e-01 3.19265693e-01 9.21450257e-01
2.98751555e-02 7.47129500e-01 9.87202376e-02 5.82414448e-01
-5.81861496e-01 2.84256160e-01 -4.92017627e-01 -1.70822084e-01
5.85639656e-01 1.29981709e+00 -3.19748640e-01 -5.26236594e-01
-5.37850976e-01 1.26786709e+00 9.71896052e-01 5.34092069e-01
-6.24374390e-01 -4.14706528e-01 1.07373394e-01 -3.80617857e-01
1.93193123e-01 2.98483446e-02 6.10652804e-01 -1.29672098e+00
4.55775782e-02 -7.14224398e-01 9.92715180e-01 -1.15018082e+00
-1.77332723e+00 5.41966259e-01 7.18764514e-02 -6.90450370e-01
-4.47878301e-01 -2.10792109e-01 -9.35040891e-01 8.82553995e-01
-1.55956435e+00 -1.16327488e+00 -4.30570662e-01 8.91148031e-01
2.39771441e-01 2.06241101e-01 6.92343652e-01 -3.33481207e-02
-4.28093195e-01 2.05323428e-01 -4.74682599e-01 3.60644788e-01
5.45501471e-01 -1.24829221e+00 2.89907694e-01 3.52441519e-01
1.50252551e-01 7.96487272e-01 5.33031702e-01 -8.88426960e-01
-1.42421508e+00 -9.31281567e-01 1.21143675e+00 -7.01449633e-01
1.06998670e+00 -1.45346656e-01 -1.14777839e+00 1.06372929e+00
6.30744457e-01 -4.70031887e-01 9.31411743e-01 3.46451312e-01
-1.06644601e-01 -6.52132630e-02 -6.87240124e-01 7.23924696e-01
9.56377745e-01 -8.66721809e-01 -1.39010298e+00 5.15862763e-01
1.45267415e+00 -4.02271330e-01 -6.23484194e-01 4.07586992e-01
1.65849961e-02 -4.60138500e-01 7.07955301e-01 -1.03261101e+00
3.81242394e-01 -2.22841695e-01 1.10440664e-01 -1.59113753e+00
-7.88474679e-02 -5.13107359e-01 -6.88447177e-01 1.40057254e+00
7.03104496e-01 -6.90163910e-01 7.75808513e-01 5.78467667e-01
-2.33078003e-01 -5.99649489e-01 -9.30159986e-01 -1.40595227e-01
-1.91143185e-01 -7.23439902e-02 6.65311456e-01 1.12438321e+00
4.20050979e-01 1.23202860e+00 -2.90730029e-01 4.63005990e-01
6.60659447e-02 4.59697038e-01 8.04118752e-01 -1.17420685e+00
-2.40425512e-01 -7.07985535e-02 -8.06600153e-02 -1.65755641e+00
6.78909540e-01 -6.95154309e-01 2.34760553e-01 -2.29438305e+00
3.78913820e-01 -2.18474753e-02 -1.75642267e-01 2.72438210e-03
-9.08101380e-01 -3.32547456e-01 -2.89982051e-01 1.64951049e-02
-1.15786874e+00 1.10180736e+00 1.48546302e+00 -4.30531055e-01
-4.01463121e-01 1.07977837e-01 -7.30463862e-01 5.37069619e-01
6.03803575e-01 -2.87809134e-01 -8.76056790e-01 -3.74669969e-01
5.73884189e-01 7.04838157e-01 6.96761236e-02 -3.52427214e-01
6.99471593e-01 -1.59833729e-01 -2.71406353e-01 -7.12272942e-01
5.76032102e-01 -6.54578686e-01 -2.12532222e-01 1.27373964e-01
-5.79383373e-01 -4.97677177e-02 -5.25501929e-02 9.90022242e-01
-5.91072083e-01 -3.42319667e-01 -1.35649666e-01 -1.17733255e-02
-8.11865211e-01 3.23400497e-01 -2.56330162e-01 3.54746848e-01
8.31173420e-01 2.83961475e-01 -6.74399257e-01 -1.38960958e+00
-9.39356148e-01 9.68401611e-01 -2.06645191e-01 4.89745438e-01
7.69045711e-01 -1.22587335e+00 -7.53702700e-01 -7.10723281e-01
2.57523865e-01 1.20444298e-01 5.18542588e-01 5.85924566e-01
-3.99692565e-01 4.41764414e-01 3.26807141e-01 -2.85717130e-01
-1.11562240e+00 3.87854666e-01 6.79613531e-01 -6.27284527e-01
-3.25639993e-01 1.11357820e+00 1.26074597e-01 -8.56699824e-01
-1.52603351e-02 -2.03912824e-01 -9.42929804e-01 2.27855042e-01
2.42754072e-01 3.71181667e-02 -1.04375176e-01 -4.81619239e-01
-3.71803463e-01 1.34225622e-01 -1.97691485e-01 -1.89187333e-01
1.03089619e+00 -5.90688765e-01 -5.94710410e-01 6.53137684e-01
8.12523603e-01 1.08427338e-01 -7.27608323e-01 -5.35988629e-01
2.51710981e-01 7.92991892e-02 -4.44274545e-01 -6.82395816e-01
-7.26092100e-01 6.80404186e-01 -2.45134532e-01 6.50642276e-01
8.88244689e-01 5.83826840e-01 1.01626861e+00 9.32753026e-01
2.41159126e-01 -7.17161655e-01 2.95941532e-01 8.64049137e-01
1.12204862e+00 -8.87238324e-01 -2.64798880e-01 -6.96691096e-01
-7.81802893e-01 9.70635831e-01 9.85759437e-01 1.40406132e-01
5.70525110e-01 -4.19581622e-01 1.77109376e-01 -5.93109608e-01
-1.05881357e+00 -3.40950996e-01 3.49058837e-01 3.86552930e-01
2.21854877e-02 -1.72951370e-01 -3.41107011e-01 8.53965878e-01
-2.20570311e-01 -1.87223747e-01 5.83523989e-01 8.13201427e-01
-6.24962747e-01 -8.79739702e-01 1.61759049e-01 4.29828137e-01
-1.55085027e-01 -2.00362489e-01 -1.20508373e+00 5.39840579e-01
-5.49739420e-01 1.54805684e+00 -7.90860206e-02 -5.71450293e-01
5.30527294e-01 6.24984860e-01 2.03703418e-01 -7.26257741e-01
-6.39277339e-01 -6.51490510e-01 7.06991255e-01 -1.72856376e-01
-6.11103415e-01 -1.25943840e-01 -1.35862505e+00 -1.79229558e-01
-7.20564246e-01 8.32642317e-01 3.05787213e-02 1.17723870e+00
4.42472190e-01 6.51083171e-01 5.56862116e-01 2.25267231e-01
-2.84714848e-01 -1.20398927e+00 -2.01171607e-01 4.30753440e-01
3.42607409e-01 -3.70892733e-01 -4.88453686e-01 -1.72304794e-01] | [10.924246788024902, 7.8814849853515625] |
f416e60a-2d22-4f8e-b746-3ade273f06de | generation-of-radiology-findings-in-chest-x | 2306.10448 | null | https://arxiv.org/abs/2306.10448v1 | https://arxiv.org/pdf/2306.10448v1.pdf | Generation of Radiology Findings in Chest X-Ray by Leveraging Collaborative Knowledge | Among all the sub-sections in a typical radiology report, the Clinical Indications, Findings, and Impression often reflect important details about the health status of a patient. The information included in Impression is also often covered in Findings. While Findings and Impression can be deduced by inspecting the image, Clinical Indications often require additional context. The cognitive task of interpreting medical images remains the most critical and often time-consuming step in the radiology workflow. Instead of generating an end-to-end radiology report, in this paper, we focus on generating the Findings from automated interpretation of medical images, specifically chest X-rays (CXRs). Thus, this work focuses on reducing the workload of radiologists who spend most of their time either writing or narrating the Findings. Unlike past research, which addresses radiology report generation as a single-step image captioning task, we have further taken into consideration the complexity of interpreting CXR images and propose a two-step approach: (a) detecting the regions with abnormalities in the image, and (b) generating relevant text for regions with abnormalities by employing a generative large language model (LLM). This two-step approach introduces a layer of interpretability and aligns the framework with the systematic reasoning that radiologists use when reviewing a CXR. | ['Dorin Comaniciu', 'Oladimeji Farri', 'Sasa Grbic', 'Constantin Suciu', 'Lucian Mihai Itu', 'Florin Ghesu', 'Awais Mansoor', 'Bogdan Georgescu', 'Sanjeev Kumar Karn', 'George Marica', 'Manuela Daniela Danu'] | 2023-06-18 | null | null | null | null | ['image-captioning'] | ['computer-vision'] | [ 8.39222968e-01 6.34294212e-01 1.49741054e-01 -5.85534930e-01
-1.16461658e+00 -7.31137753e-01 6.24002457e-01 8.24045837e-01
-2.70530432e-01 5.61197996e-01 6.46807671e-01 -9.76568222e-01
-2.95018643e-01 -6.18099809e-01 -5.60190737e-01 -3.65685374e-01
2.86296397e-01 6.25239134e-01 -8.16996992e-02 3.77113849e-01
4.42554563e-01 4.53468531e-01 -1.13052046e+00 6.30213857e-01
5.76685786e-01 7.85605133e-01 5.79151571e-01 1.12497663e+00
-2.32943341e-01 1.39685392e+00 -7.65219569e-01 -2.80338913e-01
-8.74749720e-02 -9.58485186e-01 -1.17668021e+00 7.99461424e-01
9.44899842e-02 -4.83584911e-01 -4.67782281e-03 8.98674488e-01
4.32042748e-01 -1.06440492e-01 7.71987975e-01 -6.27652645e-01
-6.77391648e-01 6.93802655e-01 -4.11645889e-01 5.53971767e-01
6.57377362e-01 1.54577717e-01 5.02765834e-01 -7.11685479e-01
6.47672653e-01 8.80373716e-01 2.89054006e-01 5.01960039e-01
-7.80993581e-01 -9.61041749e-02 2.11178526e-01 -1.33185729e-01
-1.08634531e+00 -1.73112839e-01 3.87504727e-01 -6.89947605e-01
6.05111420e-01 6.64636254e-01 6.17183924e-01 8.32851887e-01
4.27143931e-01 5.21018684e-01 1.15821636e+00 -6.27750516e-01
3.48480105e-01 2.03125224e-01 1.57082438e-01 4.40785587e-01
3.24546456e-01 -4.25647676e-01 -8.60982016e-03 -2.26133689e-01
8.52746487e-01 3.36226016e-01 -3.53798121e-01 1.58925563e-01
-1.40765524e+00 7.61912286e-01 1.97616458e-01 4.66533482e-01
-8.32699895e-01 3.17476913e-02 2.73033828e-01 -1.06828846e-01
2.62319535e-01 3.29525024e-01 1.67471528e-01 -8.71409103e-02
-1.15906262e+00 3.97118889e-02 6.79653823e-01 7.53793895e-01
3.41392606e-02 -5.07339895e-01 -5.68734467e-01 5.28048873e-01
3.43433708e-01 4.27150220e-01 6.41654313e-01 -5.69851935e-01
5.23515165e-01 6.14877045e-01 4.76179272e-02 -8.07694912e-01
-3.14115465e-01 -4.40050066e-01 -6.06396854e-01 -5.76507114e-02
8.76450092e-02 1.80525891e-02 -1.24482858e+00 1.15343583e+00
1.90982893e-01 -5.26813984e-01 7.55464211e-02 9.60087538e-01
1.06068969e+00 4.85891342e-01 3.97189975e-01 -3.62381399e-01
2.02292347e+00 -9.20987904e-01 -8.93944979e-01 -4.18238729e-01
4.61576194e-01 -1.06405520e+00 9.77468014e-01 1.23720840e-01
-1.49897444e+00 -4.08036411e-01 -8.87106121e-01 -2.00199842e-01
-1.34194240e-01 3.67452383e-01 2.04809725e-01 1.85907528e-01
-1.10466063e+00 1.99158534e-01 -7.29755938e-01 -3.70992452e-01
2.21863970e-01 2.06361506e-02 -2.27698043e-01 -2.75716037e-01
-8.35365057e-01 1.11976922e+00 3.94015491e-01 2.46165469e-01
-6.25963569e-01 -7.10194647e-01 -8.33097517e-01 -5.00848927e-02
7.14565277e-01 -8.93720329e-01 1.90679193e+00 -5.03511012e-01
-9.16393697e-01 1.05223894e+00 -3.88486177e-01 -2.33669370e-01
6.78755701e-01 -1.28879538e-02 -3.41424137e-01 5.89116812e-01
2.37740159e-01 5.78141034e-01 6.32216334e-01 -1.40357947e+00
-5.35169661e-01 -2.30863988e-01 2.79650807e-01 2.65058666e-01
3.99284631e-01 2.19751418e-01 -6.39865816e-01 -8.62925529e-01
3.14261675e-01 -8.71795774e-01 -5.54628015e-01 -4.89452370e-02
-5.45890212e-01 9.31390598e-02 5.80948174e-01 -9.64387596e-01
1.52061737e+00 -2.00807452e+00 -2.53715962e-01 2.99828827e-01
5.18442929e-01 7.38966763e-02 4.31730747e-01 4.15332377e-01
-3.37918639e-01 4.17564958e-01 -5.65269947e-01 -1.68273211e-01
-2.47792765e-01 1.32162303e-01 -3.49162459e-01 1.23755373e-01
2.83263803e-01 1.04263020e+00 -9.72933948e-01 -8.93486023e-01
3.15278113e-01 5.41248620e-01 -1.33602157e-01 3.27711225e-01
-6.31614104e-02 8.11257899e-01 -5.98261058e-01 5.07088125e-01
4.20241982e-01 -1.02543056e+00 2.03722492e-01 -2.08660826e-01
-6.12757355e-02 5.09442627e-01 -8.02295148e-01 1.49784434e+00
-7.10303903e-01 2.73802072e-01 -2.41694599e-02 -4.45179075e-01
4.64978278e-01 7.19749153e-01 3.38022679e-01 -5.17558634e-01
1.86182756e-03 1.36975303e-01 1.74222529e-01 -9.37954664e-01
5.04257202e-01 -3.10769588e-01 -4.21037823e-02 1.15620339e+00
-6.19595170e-01 -3.50036711e-01 3.57150614e-01 5.21195471e-01
1.18300974e+00 -1.13198236e-01 9.01243806e-01 8.01102445e-02
7.04808116e-01 1.65480152e-01 -1.58715904e-01 8.95904124e-01
6.79624155e-02 1.15735769e+00 5.57865024e-01 -3.11711639e-01
-1.06496525e+00 -9.86224890e-01 -1.17001191e-01 3.50879818e-01
-2.11321607e-01 -3.51013660e-01 -6.87146425e-01 -8.13279033e-01
-4.40179318e-01 9.52089310e-01 -8.62049699e-01 2.02699259e-01
-6.25973821e-01 -6.74496353e-01 1.53870374e-01 6.35057986e-01
1.12447441e-01 -1.36173654e+00 -1.31236625e+00 3.32162589e-01
-4.49649811e-01 -1.16336322e+00 -6.23895288e-01 3.17854621e-02
-7.92707741e-01 -1.11238348e+00 -1.01523685e+00 -5.64510703e-01
1.34625828e+00 1.35997534e-01 1.31684041e+00 3.26842606e-01
-7.46522069e-01 8.14237952e-01 -5.53447247e-01 -6.01292968e-01
-9.21396315e-01 -2.56470680e-01 -7.41237581e-01 -3.92405212e-01
-9.24626831e-03 -8.40813890e-02 -8.50414693e-01 -5.87052368e-02
-1.47508526e+00 5.95102429e-01 1.18751454e+00 4.56310570e-01
9.41338301e-01 -1.58157289e-01 2.09012404e-01 -1.23526239e+00
1.03104150e+00 -6.37102604e-01 -8.50387570e-03 5.67963123e-01
-4.04811352e-01 1.27925485e-01 4.89519298e-01 -2.32818767e-01
-1.13442385e+00 -9.29029509e-02 -1.43841133e-01 -1.07500680e-01
-4.57954526e-01 6.13797069e-01 5.19950449e-01 3.84256750e-01
3.71992767e-01 4.16172624e-01 6.13357648e-02 -1.47279233e-01
6.23058304e-02 7.02824891e-01 5.95179439e-01 -2.24143565e-01
5.44169843e-01 4.79268789e-01 -2.78379302e-02 -3.52219313e-01
-1.09451568e+00 -5.29824436e-01 -6.34434521e-01 -4.93699670e-01
1.04298258e+00 -4.39018220e-01 -3.90219599e-01 -3.67644787e-01
-1.17001832e+00 2.41759494e-01 -4.95962471e-01 5.41413426e-01
-3.71549785e-01 5.60364246e-01 -5.58564305e-01 -8.60786259e-01
-4.97549802e-01 -1.63295376e+00 1.27941275e+00 1.13788946e-02
-6.12488627e-01 -9.82111156e-01 -2.66562819e-01 4.04382080e-01
4.33897227e-01 4.50097352e-01 1.22286999e+00 -5.48955441e-01
-4.92428988e-01 -3.01032603e-01 -3.77612352e-01 -5.85427769e-02
5.06542146e-01 -1.70522645e-01 -7.41242826e-01 1.11128867e-01
4.26559031e-01 -5.22069596e-02 3.99784118e-01 5.13414502e-01
1.40521061e+00 -2.95756608e-01 -1.28518611e-01 1.22748047e-01
1.23969841e+00 6.26116753e-01 4.39998180e-01 1.55868381e-01
4.13191199e-01 8.38211715e-01 6.73964202e-01 3.22653800e-01
5.59526742e-01 1.89414144e-01 2.99987078e-01 -4.80228692e-01
-2.41978496e-01 -2.78578877e-01 -1.27229512e-01 7.14405775e-01
3.04062665e-02 -3.39286208e-01 -1.16716564e+00 4.49483305e-01
-1.59956634e+00 -5.86775720e-01 -3.25798988e-01 1.87537587e+00
7.62508214e-01 -1.51964799e-01 -2.70175576e-01 1.07302018e-01
6.27267540e-01 5.91400117e-02 -3.22214991e-01 -3.18608999e-01
3.16448301e-01 1.55039147e-01 3.73302788e-01 4.77010041e-01
-7.78535008e-01 1.93931580e-01 6.72596216e+00 2.96603441e-01
-1.16663659e+00 3.01849861e-02 1.09677029e+00 -6.52107522e-02
-5.05156517e-01 -1.18538097e-01 -4.74344581e-01 3.77156317e-01
8.79149973e-01 1.06324889e-01 -6.75610155e-02 6.13665760e-01
5.59987843e-01 -6.62788093e-01 -1.38336706e+00 8.52761447e-01
3.41916174e-01 -1.26218450e+00 3.58461678e-01 1.69395745e-01
3.39021206e-01 -5.08338213e-01 3.27656604e-02 -1.15895167e-01
-5.84381856e-02 -1.11914492e+00 8.62869740e-01 5.24661124e-01
1.06720257e+00 -2.16615558e-01 8.89040709e-01 2.26242140e-01
-9.45961535e-01 2.30914965e-01 2.10640486e-02 3.18288445e-01
6.10917747e-01 5.94558418e-01 -1.51770282e+00 6.67931795e-01
4.34146464e-01 1.18146874e-01 -6.64259970e-01 9.10908401e-01
-4.66944367e-01 4.67919320e-01 1.64993331e-01 1.77768007e-01
4.21102107e-01 -1.09294653e-01 3.65873575e-01 1.42604065e+00
4.32564557e-01 6.17291272e-01 1.82034373e-01 8.07734728e-01
1.17699668e-01 2.95548022e-01 -5.60141981e-01 -2.79324561e-01
-1.50138922e-02 1.34251904e+00 -1.43528819e+00 -9.44782734e-01
-3.45230490e-01 8.11625183e-01 -2.20714957e-01 9.87129733e-02
-7.44582951e-01 5.90578467e-02 -3.54306936e-01 4.38330173e-01
1.62805229e-01 1.86565980e-01 -4.42986578e-01 -6.94584966e-01
4.13766593e-01 -9.53985810e-01 4.55189228e-01 -1.15706933e+00
-8.46104860e-01 8.57089281e-01 2.20850334e-01 -1.40788054e+00
-5.67138314e-01 -3.33248198e-01 -5.09134293e-01 8.91403258e-01
-1.53327775e+00 -9.32357609e-01 -6.39927208e-01 1.33109748e-01
8.62135828e-01 5.14985025e-01 8.35442483e-01 1.37201428e-01
-1.03817627e-01 -5.68234771e-02 -5.51470339e-01 1.28695797e-02
3.99920225e-01 -1.31321371e+00 2.44283438e-01 8.28800738e-01
-2.69360840e-01 9.59539652e-01 6.65339053e-01 -8.07511985e-01
-9.64058101e-01 -8.74812603e-01 1.16011941e+00 -2.93003887e-01
3.57322544e-01 -1.04998155e-02 -9.27595675e-01 5.89178920e-01
2.50442177e-01 -1.71106845e-01 9.08609927e-01 -6.44647241e-01
2.39632398e-01 4.27557766e-01 -1.19723105e+00 6.40168607e-01
7.44552672e-01 -4.94498879e-01 -9.86245215e-01 5.65644145e-01
6.19173527e-01 -6.92844391e-01 -8.62967074e-01 1.56710595e-01
2.72445351e-01 -5.26316702e-01 9.13907111e-01 -4.56793755e-01
8.56997192e-01 -3.95112574e-01 2.90480405e-01 -1.04550850e+00
-1.19250394e-01 -2.21853077e-01 1.97496876e-01 7.09019482e-01
6.40794694e-01 -2.68453956e-01 4.06017482e-01 9.11413491e-01
-2.32113257e-01 -7.30450451e-01 -3.76554847e-01 1.80030018e-02
-5.26794553e-01 -5.68704426e-01 4.19196039e-01 6.53183103e-01
-7.30471611e-02 1.78024359e-02 7.24565983e-02 2.13786855e-01
2.74265230e-01 2.27211699e-01 2.45693594e-01 -9.06933129e-01
-3.38204771e-01 -2.00420454e-01 -2.99102385e-02 -8.32670450e-01
-4.51492369e-01 -8.47829163e-01 1.79074854e-01 -2.34046841e+00
5.29686034e-01 -2.72518545e-01 -2.08024066e-02 3.60477269e-01
-3.59133780e-01 1.16937377e-01 1.55120194e-01 3.57550055e-01
-5.29665112e-01 -2.88926870e-01 1.57372129e+00 8.07164237e-02
9.15044174e-03 1.47298679e-01 -1.04939926e+00 7.89298475e-01
4.90786374e-01 -6.42811477e-01 -5.66784143e-01 -4.04851496e-01
3.00477624e-01 5.98608971e-01 4.76968288e-01 -7.30710328e-01
3.20627451e-01 -9.07383338e-02 4.50084239e-01 -9.92963016e-01
1.18548810e-01 -8.63312125e-01 1.21474139e-01 6.88717425e-01
-6.01288199e-01 6.41158640e-01 8.55037719e-02 3.85689288e-01
-5.01170874e-01 -6.46585763e-01 4.86262530e-01 -8.22314680e-01
-3.36525947e-01 -3.46890390e-02 -7.75352061e-01 -7.60477707e-02
1.17176342e+00 -4.20639098e-01 -5.17182834e-02 -5.58997393e-01
-1.06163228e+00 6.39164373e-02 2.33211949e-01 2.04667121e-01
9.02050912e-01 -6.95616484e-01 -7.94533134e-01 8.17002729e-02
1.23794138e-01 4.64460999e-01 5.09009004e-01 1.15610063e+00
-8.42960417e-01 7.32092381e-01 2.77868420e-01 -6.62862241e-01
-1.37893808e+00 4.88266885e-01 5.49458414e-02 -7.68659770e-01
-9.05446231e-01 4.73016292e-01 5.24469733e-01 1.19349442e-01
-7.63006683e-04 -7.63410270e-01 -4.41927761e-01 2.24655494e-02
8.23448479e-01 -1.08047321e-01 3.14837277e-01 -5.48879325e-01
-1.79056585e-01 4.82206583e-01 -4.93421674e-01 -3.88938516e-01
1.22700286e+00 -3.50144416e-01 -1.88325942e-01 4.55650985e-01
7.68173158e-01 9.49571654e-02 -7.46258914e-01 -1.28619343e-01
-5.56301232e-03 -2.57918060e-01 -4.41630790e-03 -1.18597710e+00
-6.59988642e-01 7.96912491e-01 1.83297724e-01 3.45892102e-01
1.20011342e+00 4.28309470e-01 4.76644427e-01 1.08489546e-03
1.00227334e-01 -8.57543826e-01 1.09426789e-01 2.36068759e-02
1.15374243e+00 -1.07880354e+00 2.75447756e-01 -5.21748006e-01
-1.02713740e+00 1.23232186e+00 2.27120459e-01 4.20472980e-01
3.58549178e-01 3.89232427e-01 3.69455725e-01 -5.96897602e-01
-6.90182745e-01 7.18239322e-02 4.58877802e-01 1.91296950e-01
8.07252109e-01 -1.64432060e-02 -5.92193902e-01 3.45355928e-01
-1.61469296e-01 6.39159307e-02 7.22743630e-01 1.50956845e+00
-2.82902390e-01 -8.84184599e-01 -8.26283395e-01 6.36052132e-01
-9.34465230e-01 -2.75255740e-01 -4.51991528e-01 7.18039036e-01
-7.37636909e-02 9.10703659e-01 2.47746073e-02 1.48869082e-01
3.69457483e-01 1.05257161e-01 3.30394745e-01 -1.01662147e+00
-8.36776316e-01 2.86530107e-02 7.44001940e-02 -2.33418539e-01
-4.68376726e-01 -6.00108266e-01 -1.48465776e+00 3.28874409e-01
1.31275281e-01 2.14297161e-01 9.94460881e-01 1.06110096e+00
1.47948027e-01 1.13730538e+00 1.40404060e-01 -3.25875849e-01
-3.98055911e-01 -1.05331779e+00 -4.83447239e-02 5.76810479e-01
4.36720878e-01 -2.00401068e-01 -2.01829314e-01 4.36877906e-01] | [15.045442581176758, -1.3982497453689575] |
4b36dc0d-8ecd-4b29-bf38-81136b8a9380 | memnet-a-persistent-memory-network-for-image | 1708.02209 | null | http://arxiv.org/abs/1708.02209v1 | http://arxiv.org/pdf/1708.02209v1.pdf | MemNet: A Persistent Memory Network for Image Restoration | Recently, very deep convolutional neural networks (CNNs) have been attracting
considerable attention in image restoration. However, as the depth grows, the
long-term dependency problem is rarely realized for these very deep models,
which results in the prior states/layers having little influence on the
subsequent ones. Motivated by the fact that human thoughts have persistency, we
propose a very deep persistent memory network (MemNet) that introduces a memory
block, consisting of a recursive unit and a gate unit, to explicitly mine
persistent memory through an adaptive learning process. The recursive unit
learns multi-level representations of the current state under different
receptive fields. The representations and the outputs from the previous memory
blocks are concatenated and sent to the gate unit, which adaptively controls
how much of the previous states should be reserved, and decides how much of the
current state should be stored. We apply MemNet to three image restoration
tasks, i.e., image denosing, super-resolution and JPEG deblocking.
Comprehensive experiments demonstrate the necessity of the MemNet and its
unanimous superiority on all three tasks over the state of the arts. Code is
available at https://github.com/tyshiwo/MemNet. | ['Ying Tai', 'Jian Yang', 'Xiaoming Liu', 'Chunyan Xu'] | 2017-08-07 | memnet-a-persistent-memory-network-for-image-1 | http://openaccess.thecvf.com/content_iccv_2017/html/Tai_MemNet_A_Persistent_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Tai_MemNet_A_Persistent_ICCV_2017_paper.pdf | iccv-2017-10 | ['color-image-denoising', 'jpeg-artifact-correction'] | ['computer-vision', 'computer-vision'] | [-2.37435848e-02 -1.04498520e-01 -2.48537853e-01 -1.04766399e-01
4.20570858e-02 1.73990190e-01 4.17772233e-01 -3.34561914e-01
-3.46887887e-01 6.72180891e-01 5.52266777e-01 -1.86449394e-01
1.65739641e-01 -1.09937143e+00 -6.75807357e-01 -1.06697643e+00
2.47924119e-01 -2.45094612e-01 4.73599792e-01 -1.84550375e-01
3.32542866e-01 1.60820350e-01 -1.50165796e+00 4.23632920e-01
7.32025862e-01 9.43555951e-01 7.54542112e-01 3.71618360e-01
-1.85974151e-01 1.33138454e+00 -3.66511822e-01 1.08605005e-01
1.52866952e-02 -5.20499408e-01 -8.77002776e-01 1.66886076e-01
2.13588312e-01 -6.87902927e-01 -1.11145568e+00 1.33208156e+00
4.60570753e-01 2.97233313e-01 1.09607726e-01 -6.17730558e-01
-1.14848018e+00 5.14800489e-01 -2.98692763e-01 7.26300657e-01
-1.18790865e-01 3.58883828e-01 5.49593091e-01 -7.98694551e-01
4.77846771e-01 1.28163445e+00 3.94501448e-01 6.84975803e-01
-9.83130932e-01 -7.53508031e-01 3.40012819e-01 5.60575426e-01
-1.29493117e+00 -6.29252434e-01 6.91733420e-01 -1.94443747e-01
9.46983278e-01 -6.07602932e-02 6.84325814e-01 9.55220461e-01
6.59372211e-01 6.59125745e-01 1.13982809e+00 -1.86585933e-01
3.44067782e-01 -1.76725313e-01 2.50706315e-01 6.85443163e-01
-1.60406530e-02 5.77018932e-02 -7.34439969e-01 1.22459076e-01
1.19162726e+00 4.53637689e-01 -4.60339189e-01 1.38750330e-01
-9.81546998e-01 6.84097886e-01 7.73836315e-01 5.28650761e-01
-7.95353711e-01 3.22917461e-01 1.80381015e-01 5.54981411e-01
4.17273700e-01 -7.98761621e-02 -1.87070400e-01 2.75016934e-01
-8.35669935e-01 -2.36204669e-01 4.91970301e-01 5.72744787e-01
8.76984000e-01 1.97670847e-01 -1.23592727e-01 9.61550951e-01
1.75589189e-01 6.81593642e-02 8.43112111e-01 -9.60431576e-01
1.89300492e-01 6.33484781e-01 -1.04258090e-01 -1.22381961e+00
-1.84925422e-02 -5.31334162e-01 -1.60112762e+00 1.68092921e-01
-2.12976307e-01 1.24447547e-01 -1.30041003e+00 1.82569027e+00
-9.16822720e-03 5.80959558e-01 -9.34175774e-03 8.99004579e-01
8.68777752e-01 1.12781572e+00 1.73350662e-01 -3.97672355e-01
1.18771362e+00 -1.04621935e+00 -8.45783889e-01 -6.46937430e-01
7.24782199e-02 -5.62677443e-01 8.69831860e-01 2.85500765e-01
-1.24061465e+00 -8.31770778e-01 -1.22294855e+00 -1.97754577e-01
-1.32907942e-01 -8.82852599e-02 5.39542377e-01 1.83059573e-02
-1.40091443e+00 9.11590695e-01 -1.01726067e+00 -1.49067074e-01
6.77715361e-01 3.80007982e-01 -1.63944855e-01 -2.76748061e-01
-1.38088977e+00 7.49912620e-01 5.60394049e-01 4.56178576e-01
-1.17737436e+00 -1.64400607e-01 -6.04054570e-01 2.53404200e-01
5.74124306e-02 -7.15140045e-01 8.15834284e-01 -1.17889154e+00
-1.35997748e+00 7.80883193e-01 -2.77359307e-01 -4.46248293e-01
1.88304305e-01 -1.13057047e-01 -5.34938931e-01 1.24964364e-01
-8.33961368e-02 6.54497027e-01 1.22123587e+00 -9.27917004e-01
-5.26130736e-01 -2.05402792e-01 1.24060646e-01 1.40526906e-01
-7.87643611e-01 -2.32060298e-01 -6.43954754e-01 -7.88919151e-01
6.59271896e-01 -7.31551707e-01 -3.71238351e-01 5.51752746e-02
-1.98636860e-01 -1.26378193e-01 6.73361599e-01 -8.57109606e-01
1.47284365e+00 -2.43102455e+00 4.80007827e-01 -2.42255583e-01
2.92285800e-01 2.74759442e-01 -2.84478635e-01 1.06231853e-01
-2.25681201e-01 5.14922664e-02 -2.01044992e-01 -1.00057654e-01
-4.13481116e-01 4.79710817e-01 -5.55474639e-01 4.21526253e-01
-1.50349187e-02 8.12138677e-01 -7.98894703e-01 -3.22114587e-01
1.88426644e-01 6.13174260e-01 -4.23274338e-01 3.06086600e-01
-4.05880362e-02 5.32099307e-01 -4.61240381e-01 3.55283886e-01
8.46489668e-01 -6.57798946e-01 1.97407901e-01 -2.23396406e-01
-3.10017616e-01 4.79954422e-01 -1.14368641e+00 1.65295231e+00
-4.37444627e-01 4.98221278e-01 7.90174380e-02 -9.87211466e-01
8.49620223e-01 3.51006985e-01 1.10444322e-01 -1.08464766e+00
1.14534408e-01 -3.05946637e-03 -1.44851446e-01 -4.60591316e-01
5.53865850e-01 -1.98065534e-01 2.73715615e-01 5.06610215e-01
1.49366900e-01 3.63344997e-01 -3.70242000e-02 2.03522727e-01
1.02855778e+00 -2.13201717e-01 1.40625417e-01 -3.70621234e-01
5.75070024e-01 -2.04070613e-01 9.14938271e-01 6.56524777e-01
-2.38261804e-01 5.42094350e-01 3.55280638e-01 -8.91131997e-01
-9.64528799e-01 -9.32687879e-01 1.10466726e-01 1.10599160e+00
5.21661043e-01 -2.38152936e-01 -5.74777782e-01 -7.79901296e-02
-3.95392329e-01 3.95879328e-01 -7.86258340e-01 -6.85949683e-01
-8.48831534e-01 -8.68126035e-01 2.77214777e-02 3.69108230e-01
1.17234612e+00 -1.42298949e+00 -4.38864797e-01 3.82997632e-01
-3.44476044e-01 -7.26959109e-01 -3.85941058e-01 1.17177673e-01
-1.32936513e+00 -7.03790128e-01 -6.32149994e-01 -1.07818234e+00
7.82190979e-01 4.14036810e-01 1.07649505e+00 5.57629228e-01
1.71910331e-01 -1.04039595e-01 -1.29178971e-01 3.29196870e-01
-4.35719103e-01 1.40612960e-01 -3.85752544e-02 1.78859718e-02
1.07441641e-01 -1.05646896e+00 -1.04462230e+00 2.95372277e-01
-1.16475344e+00 4.85888392e-01 6.85887992e-01 9.59304571e-01
6.66461051e-01 2.72847414e-01 5.21544635e-01 -8.75719190e-01
4.95371044e-01 -5.68503439e-01 -2.20294386e-01 8.11475217e-02
-5.40525436e-01 1.25102684e-01 4.69007015e-01 -6.26383185e-01
-1.13032246e+00 -2.45343149e-01 -1.76730022e-01 -4.33591962e-01
1.42835170e-01 5.40414453e-01 -1.87519863e-01 2.24778548e-01
4.90379632e-01 6.19801998e-01 -8.16674158e-02 -5.39635003e-01
5.79193495e-02 5.54553986e-01 6.62627637e-01 -1.57314003e-01
5.26743770e-01 4.26362306e-01 -4.46904242e-01 -5.74349701e-01
-8.89687836e-01 9.16257575e-02 -6.06500685e-01 -8.53429213e-02
8.23539197e-01 -1.18882871e+00 -4.14241284e-01 9.64423895e-01
-1.14481962e+00 -5.16399503e-01 -2.01379821e-01 8.79841596e-02
-2.61751741e-01 2.13169694e-01 -1.18628073e+00 -2.90028065e-01
-3.99413198e-01 -1.05633247e+00 2.39098191e-01 5.98841667e-01
1.58451021e-01 -9.43673372e-01 -7.38514960e-02 8.54429156e-02
8.07479203e-01 -2.08918989e-01 1.05770695e+00 1.49137378e-01
-9.42540884e-01 1.89403370e-01 -2.67235935e-01 6.82293713e-01
1.58116832e-01 -5.10132611e-01 -9.34066832e-01 -5.51655591e-01
4.26882535e-01 -1.64079875e-01 1.63613677e+00 4.07862037e-01
1.35285640e+00 -5.83377957e-01 -2.05010876e-01 7.51931310e-01
1.43239260e+00 1.00878477e-01 1.31396341e+00 4.61094171e-01
6.90748513e-01 5.58948666e-02 9.17546600e-02 2.17843533e-01
4.28604633e-01 1.98506594e-01 6.88512266e-01 -7.30267400e-03
-5.74696898e-01 -1.90269038e-01 4.61142778e-01 1.34960687e+00
-1.85006306e-01 -1.35309771e-01 -7.60017455e-01 5.89972019e-01
-1.74354589e+00 -9.81438398e-01 3.50742042e-01 1.87678063e+00
1.10907900e+00 1.94271252e-01 -7.39349663e-01 -1.07917883e-01
9.77324307e-01 8.29543829e-01 -8.42561305e-01 -1.35177016e-01
-1.96076989e-01 2.06789523e-01 2.13469684e-01 3.57422203e-01
-9.67506826e-01 1.03230536e+00 5.53409815e+00 7.65384674e-01
-1.37794900e+00 3.91700089e-01 9.57993448e-01 9.22371596e-02
-1.96510419e-01 1.56396925e-01 -4.97837454e-01 5.65937579e-01
8.59541655e-01 1.57387052e-02 6.38110638e-01 5.41356027e-01
4.42133918e-02 -2.65323937e-01 -6.31579816e-01 9.21786785e-01
-6.29197881e-02 -1.52467740e+00 3.94008785e-01 -4.76256683e-02
9.23198521e-01 1.19311869e-01 3.36336672e-01 2.60900408e-01
2.19579622e-01 -1.03901207e+00 5.93061388e-01 7.62473762e-01
6.30842209e-01 -7.13105381e-01 5.00968397e-01 5.06205678e-01
-9.62529719e-01 -3.73607278e-01 -9.09563601e-01 -2.20513180e-01
4.13699672e-02 7.95168519e-01 5.81144653e-02 1.61558241e-01
8.58604789e-01 9.87776160e-01 -2.85082459e-01 6.74898982e-01
-5.06377518e-01 7.00115621e-01 1.67628333e-01 4.32548732e-01
1.41977072e-01 -7.05491826e-02 3.39038491e-01 9.00875449e-01
3.04584682e-01 4.95048612e-01 1.20868683e-02 8.95025134e-01
-4.51647669e-01 -3.62810314e-01 -8.71484578e-02 8.09620693e-02
4.59745556e-01 1.04880404e+00 -5.75404108e-01 -5.67794204e-01
-2.75494784e-01 1.32189727e+00 3.40018988e-01 5.11203289e-01
-4.85764116e-01 -1.90374181e-01 7.70709693e-01 1.20698012e-01
4.32187527e-01 -3.02886039e-01 -2.71937579e-01 -1.29369891e+00
-1.83977693e-01 -7.17279971e-01 4.21515763e-01 -8.06462049e-01
-1.02107775e+00 8.44636083e-01 -5.77136219e-01 -1.13355625e+00
5.26270829e-02 -2.69617379e-01 -7.39113927e-01 1.05753362e+00
-1.69642043e+00 -8.24781299e-01 -4.54282194e-01 7.70635545e-01
6.15055919e-01 -6.25404567e-02 7.97772229e-01 4.14866298e-01
-8.16977561e-01 1.06339477e-01 -3.49796936e-02 2.46133640e-01
4.92182076e-01 -6.80170476e-01 3.15357953e-01 1.02885628e+00
-3.01871479e-01 8.85306656e-01 5.71394205e-01 -7.14496851e-01
-1.38316047e+00 -1.18021107e+00 7.55989850e-01 5.53450994e-02
5.20889223e-01 -3.26972343e-02 -1.38725448e+00 6.80426478e-01
3.56577337e-01 2.82395631e-01 5.87277338e-02 -2.00654462e-01
-3.20015848e-01 -1.59759775e-01 -8.86163652e-01 5.29002964e-01
1.06347895e+00 -6.62934601e-01 -5.84212780e-01 1.10494763e-01
6.77201211e-01 -3.99262607e-01 -5.08729219e-01 2.47600004e-01
2.98237711e-01 -1.02521300e+00 9.47256446e-01 -3.42155486e-01
8.06336224e-01 -2.59712368e-01 -1.06599815e-01 -1.28756475e+00
-8.12481701e-01 -2.03811795e-01 -3.67850661e-01 9.92170990e-01
-5.95855154e-02 -7.53744900e-01 6.17373347e-01 4.23232794e-01
-2.01658845e-01 -7.21809745e-01 -1.08363557e+00 -3.62860799e-01
1.56939670e-01 1.13321014e-01 4.73020852e-01 9.41858828e-01
-4.94628906e-01 3.26800436e-01 -6.47090018e-01 3.87608051e-01
4.24505144e-01 2.80208081e-01 1.70078725e-01 -1.07974780e+00
-1.55341312e-01 -1.62835836e-01 -2.46014714e-01 -1.51022565e+00
7.55719841e-02 -8.29606533e-01 6.37076376e-03 -1.65512884e+00
5.12166917e-01 -3.35359901e-01 -8.21155906e-01 7.73071766e-01
-1.83654130e-01 3.70540440e-01 1.35454163e-02 7.54050732e-01
-5.76771259e-01 7.27719665e-01 1.43961477e+00 -2.83664644e-01
-1.72808558e-01 -1.83032095e-01 -8.72137129e-01 8.24188590e-01
9.88839567e-01 -4.82597858e-01 -2.80705690e-01 -9.48502123e-01
3.05641025e-01 3.99282575e-02 5.09965777e-01 -1.12879884e+00
7.09512949e-01 -1.62729084e-01 6.21123552e-01 -3.52421314e-01
2.22497448e-01 -4.31212902e-01 3.43148857e-01 6.57049894e-01
-2.85933942e-01 7.96234831e-02 1.47862032e-01 4.72214311e-01
-4.26714242e-01 -4.47609089e-02 1.05132270e+00 -6.09552920e-01
-1.11139488e+00 6.17219925e-01 -4.18035030e-01 -2.72206903e-01
5.11957526e-01 -2.02054739e-01 -3.94477636e-01 -2.67253757e-01
-8.69315922e-01 8.45606904e-03 4.15564507e-01 5.35109818e-01
8.68171811e-01 -1.35548770e+00 -6.58903003e-01 3.16146880e-01
-4.41128820e-01 3.13665494e-02 8.13242555e-01 6.50460541e-01
-2.55658746e-01 -2.40491219e-02 -4.94733125e-01 -3.46412808e-01
-9.15753841e-01 5.11203051e-01 4.95463729e-01 -3.10047299e-01
-8.48923385e-01 7.99126387e-01 3.83577496e-01 -1.55942524e-02
7.40136802e-02 -4.67149355e-02 -3.55933428e-01 -3.03921830e-02
9.02244925e-01 3.00965726e-01 4.72396100e-03 -4.94094491e-01
-9.71310660e-02 3.09047282e-01 -3.77990663e-01 1.82479680e-01
1.70447993e+00 -4.65865344e-01 -7.68008113e-01 3.19295615e-01
9.20230985e-01 -5.36096931e-01 -1.53195715e+00 -5.44660032e-01
-3.34634513e-01 -3.73680264e-01 5.28975010e-01 -5.63131094e-01
-1.66061318e+00 1.00163710e+00 7.91149318e-01 -4.77158092e-02
1.50946164e+00 -2.03573164e-02 9.94744539e-01 2.57156193e-01
3.35166007e-01 -1.06352544e+00 3.14470053e-01 7.95470059e-01
9.53578949e-01 -8.84150386e-01 9.25291516e-03 -2.21896656e-02
-2.23289624e-01 9.83560264e-01 5.98204613e-01 -4.01146710e-01
8.63826752e-01 2.22299635e-01 -6.96498230e-02 -1.31005943e-01
-9.56891835e-01 2.11059719e-01 -1.55514389e-01 2.03935876e-01
1.29835680e-01 -1.03030756e-01 -1.44049078e-01 4.87293780e-01
1.34648904e-01 2.92267859e-01 5.02226591e-01 9.71074283e-01
-7.53388107e-01 -9.33065832e-01 -1.86630547e-01 3.67321402e-01
-3.76663089e-01 -2.91722625e-01 2.88190782e-01 1.15547396e-01
1.99876770e-01 7.70171046e-01 2.35969231e-01 -5.11673570e-01
4.14179377e-02 -1.35223523e-01 2.33656362e-01 -6.16007149e-01
-4.21763211e-01 -2.09684640e-01 -3.98126274e-01 -6.82774365e-01
-4.20781225e-01 -3.52966338e-01 -1.22279835e+00 -6.82515740e-01
-8.08399096e-02 -8.01067576e-02 2.50199795e-01 7.62869358e-01
4.81478125e-01 7.99371719e-01 6.16200626e-01 -7.63317406e-01
-4.54211265e-01 -1.03259218e+00 -5.08214116e-01 2.14060917e-01
5.50806701e-01 -3.49389583e-01 -2.79299915e-01 2.31085718e-01] | [11.11847972869873, -2.0987372398376465] |
087c754a-0176-46e1-a445-56c59288bc51 | ppr10k-a-large-scale-portrait-photo | 2105.09180 | null | https://arxiv.org/abs/2105.09180v1 | https://arxiv.org/pdf/2105.09180v1.pdf | PPR10K: A Large-Scale Portrait Photo Retouching Dataset with Human-Region Mask and Group-Level Consistency | Different from general photo retouching tasks, portrait photo retouching (PPR), which aims to enhance the visual quality of a collection of flat-looking portrait photos, has its special and practical requirements such as human-region priority (HRP) and group-level consistency (GLC). HRP requires that more attention should be paid to human regions, while GLC requires that a group of portrait photos should be retouched to a consistent tone. Models trained on existing general photo retouching datasets, however, can hardly meet these requirements of PPR. To facilitate the research on this high-frequency task, we construct a large-scale PPR dataset, namely PPR10K, which is the first of its kind to our best knowledge. PPR10K contains $1, 681$ groups and $11, 161$ high-quality raw portrait photos in total. High-resolution segmentation masks of human regions are provided. Each raw photo is retouched by three experts, while they elaborately adjust each group of photos to have consistent tones. We define a set of objective measures to evaluate the performance of PPR and propose strategies to learn PPR models with good HRP and GLC performance. The constructed PPR10K dataset provides a good benchmark for studying automatic PPR methods, and experiments demonstrate that the proposed learning strategies are effective to improve the retouching performance. Datasets and codes are available: https://github.com/csjliang/PPR10K. | ['Lei Zhang', 'Xuansong Xie', 'Miaomiao Cui', 'Hui Zeng', 'Jie Liang'] | 2021-05-19 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Liang_PPR10K_A_Large-Scale_Portrait_Photo_Retouching_Dataset_With_Human-Region_Mask_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Liang_PPR10K_A_Large-Scale_Portrait_Photo_Retouching_Dataset_With_Human-Region_Mask_CVPR_2021_paper.pdf | cvpr-2021-1 | ['photo-retouching'] | ['computer-vision'] | [ 4.46828008e-01 -1.32096961e-01 -3.43345463e-01 -1.92662105e-01
-8.24451387e-01 -4.32458013e-01 2.64167726e-01 -3.77179623e-01
-1.85231432e-01 6.09412909e-01 -8.38811994e-02 -6.49553835e-02
6.27248958e-02 -7.06814110e-01 -7.75949240e-01 -6.92346931e-01
4.05230254e-01 7.95765314e-03 3.99854273e-01 -2.30963126e-01
3.43093902e-01 5.15420198e-01 -1.28862762e+00 1.61522672e-01
9.14609730e-01 8.78253222e-01 3.34038407e-01 7.75818527e-01
1.95297271e-01 2.50659525e-01 -6.44102395e-01 -6.91804945e-01
5.39686263e-01 -4.64813650e-01 -6.99366927e-01 6.56593263e-01
6.88845813e-01 -2.14164808e-01 -3.34539443e-01 1.16721237e+00
5.05235255e-01 2.66482472e-01 6.16532326e-01 -1.38071966e+00
-1.33466601e+00 5.59924364e-01 -1.22133648e+00 3.73648196e-01
2.74778992e-01 4.18978989e-01 1.07710135e+00 -8.04097235e-01
5.74661255e-01 1.00909960e+00 6.44526303e-01 7.33815372e-01
-1.24062431e+00 -9.55322623e-01 2.59733230e-01 2.20238134e-01
-1.59508264e+00 -2.99953073e-01 1.03102946e+00 -2.10329592e-01
3.58494133e-01 7.87286103e-01 6.85659885e-01 9.71589208e-01
1.14909731e-01 8.42864215e-01 1.19147730e+00 -2.19544217e-01
1.04052415e-02 1.92796230e-01 -3.09044402e-02 3.19071144e-01
5.25889844e-02 -3.17139328e-01 -2.19953671e-01 2.20392540e-01
1.02969205e+00 3.54949273e-02 -6.04891062e-01 1.31947130e-01
-1.05818665e+00 2.90660262e-01 6.59019947e-01 4.43539321e-01
-1.45286888e-01 -9.60631575e-03 -2.80707330e-02 3.72848026e-02
5.68674564e-01 4.50001389e-01 -1.99315563e-01 1.75353736e-01
-9.17685568e-01 -5.52000031e-02 1.87144130e-01 1.20519853e+00
9.31566000e-01 -1.20039731e-01 -4.30939585e-01 1.36638069e+00
2.50528064e-02 6.90989614e-01 2.73893267e-01 -9.61308002e-01
4.01325583e-01 6.35482192e-01 2.25132197e-01 -1.39511919e+00
-5.75392134e-02 1.57158053e-03 -1.22244227e+00 -5.83120137e-02
1.61442831e-01 -7.78563619e-02 -1.12381530e+00 1.22444880e+00
3.17235321e-01 6.52062893e-02 -2.44186163e-01 1.18986368e+00
9.44883525e-01 1.21522236e+00 9.96674001e-02 -4.14099127e-01
1.38889432e+00 -9.85443950e-01 -8.21699202e-01 -3.08327496e-01
-3.51488702e-02 -1.06777978e+00 1.44386530e+00 3.44816953e-01
-1.13239408e+00 -7.67325759e-01 -9.02271450e-01 -2.32214659e-01
-2.87129790e-01 5.81798375e-01 2.82100230e-01 5.41711509e-01
-1.09106350e+00 3.33457828e-01 -1.06862478e-01 -4.73553836e-01
6.92204118e-01 7.49735013e-02 -2.54138142e-01 -3.65579069e-01
-1.00519264e+00 5.37063658e-01 1.98268354e-01 5.77865720e-01
-7.04651117e-01 -7.91427195e-01 -6.05982125e-01 -1.76448196e-01
5.99913597e-01 -3.16853583e-01 1.10444772e+00 -1.37989366e+00
-1.44677556e+00 1.18716824e+00 7.81629756e-02 2.93647945e-02
6.32921457e-01 4.26621325e-02 -8.09432864e-01 2.27625161e-01
-3.64814955e-03 8.33717585e-01 1.19885135e+00 -1.67964983e+00
-7.14700520e-01 -1.79056358e-02 -1.23343810e-01 2.29087710e-01
-3.28652948e-01 -7.85880685e-02 -1.06608951e+00 -8.97691786e-01
-1.63649842e-01 -8.38926613e-01 -2.02028275e-01 -9.85581707e-03
-7.30130970e-01 -2.24644914e-01 7.31735945e-01 -8.68896842e-01
1.28283000e+00 -2.22710443e+00 -1.81874812e-01 8.86006951e-02
2.60727525e-01 4.70065594e-01 -4.63023275e-01 2.27676153e-01
-5.02055660e-02 5.58134675e-01 -4.41626221e-01 -1.69211581e-01
-9.07339007e-02 1.95622578e-01 -2.17964873e-01 3.77882719e-01
2.99383193e-01 9.37464058e-01 -7.40466356e-01 -6.45892799e-01
2.31090292e-01 2.35856220e-01 -3.56918313e-02 1.62729219e-01
-2.14882016e-01 8.16317648e-02 -2.95735002e-01 8.47727895e-01
1.05446327e+00 -2.50814348e-01 -1.11575879e-01 -4.76075649e-01
-4.93550599e-02 -4.09914494e-01 -1.00814533e+00 1.25282240e+00
-1.06983595e-01 6.59335136e-01 -2.37117618e-01 -3.92451704e-01
1.20508850e+00 -1.21279269e-01 5.06697714e-01 -9.65060532e-01
3.20681445e-02 -6.35059252e-02 -4.74954635e-01 -4.95938241e-01
1.03043151e+00 2.50751860e-02 -1.49297535e-01 2.15313166e-01
-4.50072587e-01 -1.68201238e-01 3.38599354e-01 2.43992627e-01
7.59008229e-01 -2.27297068e-01 2.13912994e-01 -1.63704872e-01
5.16005754e-01 1.21397421e-01 8.90738010e-01 4.24616694e-01
-3.86732608e-01 1.02718997e+00 2.92814732e-01 -4.13588077e-01
-1.31529343e+00 -1.08460557e+00 -9.89653468e-02 8.63550842e-01
7.76310503e-01 -1.08998604e-01 -7.43545651e-01 -6.29170001e-01
-2.29067340e-01 6.45025074e-01 -7.29731083e-01 1.79841891e-02
-5.62256694e-01 -8.79477382e-01 1.53806657e-01 2.30640158e-01
9.55245852e-01 -1.26859796e+00 2.70492565e-02 -2.81930298e-01
-3.00515980e-01 -1.06969690e+00 -1.27954471e+00 -6.20126367e-01
-3.93425524e-01 -1.26382613e+00 -1.22731340e+00 -9.77848172e-01
1.05591321e+00 8.04608643e-01 1.08559144e+00 1.07255615e-01
-3.98435682e-01 4.30968523e-01 -6.05877817e-01 1.60822645e-01
-2.83607930e-01 1.96463183e-01 -3.03843617e-01 3.88338864e-01
3.28984391e-03 -3.59947354e-01 -8.80247593e-01 9.04275060e-01
-1.21682358e+00 3.66597533e-01 9.49523330e-01 4.33652490e-01
1.03483558e+00 3.93088102e-01 3.53083372e-01 -9.33155239e-01
5.01440167e-01 -1.28495246e-01 -4.92198884e-01 5.11267006e-01
-4.45703775e-01 -6.38121963e-01 4.82171714e-01 -9.09410834e-01
-1.14895046e+00 -2.49609485e-01 -5.02874814e-02 -5.11814117e-01
-8.53357092e-02 -1.71734944e-01 -3.65697950e-01 -1.33114770e-01
5.38830936e-01 3.18495810e-01 -4.58261698e-01 -3.16801250e-01
4.90616709e-01 6.25073195e-01 6.08777046e-01 -2.47400641e-01
1.03988338e+00 2.91727662e-01 -4.54484046e-01 -8.89348030e-01
-6.53252125e-01 -3.75877410e-01 -4.32599068e-01 -6.60427749e-01
8.45483005e-01 -6.28424168e-01 -6.46441698e-01 4.95884150e-01
-7.59580612e-01 -6.64700985e-01 -3.82439673e-01 -7.28817135e-02
-2.18395844e-01 4.02482837e-01 -5.07646620e-01 -7.58245170e-01
-4.01636094e-01 -8.29564214e-01 1.04289949e+00 7.92654514e-01
9.36614200e-02 -5.38585246e-01 -2.37222649e-02 5.56688666e-01
9.48728621e-02 1.53707966e-01 5.58131516e-01 9.44840163e-02
-6.75612271e-01 -1.52071919e-02 -6.11777842e-01 4.29621577e-01
2.93676823e-01 4.19881165e-01 -7.19034016e-01 -2.16252863e-01
-4.16216701e-01 1.96659639e-02 8.09277773e-01 4.72979307e-01
1.33816409e+00 -6.27213240e-01 -2.93837219e-01 4.50256497e-01
1.47123528e+00 3.60612184e-01 1.26909888e+00 2.89073259e-01
8.89358163e-01 3.82816941e-01 8.55817139e-01 3.33155096e-01
4.49236780e-01 6.19284689e-01 2.44212955e-01 -5.25862575e-01
-4.73914742e-01 -5.26081085e-01 3.28293085e-01 7.42994726e-01
-1.41236648e-01 -5.41305602e-01 -5.96508563e-01 6.67824745e-01
-1.64149749e+00 -9.05195951e-01 -2.70885944e-01 2.02013636e+00
9.17128682e-01 -2.09765390e-01 1.54592797e-01 1.24752320e-01
1.17311001e+00 5.07291377e-01 -6.12655103e-01 -2.21992463e-01
-3.23298931e-01 -3.91714543e-01 6.63138092e-01 3.64940196e-01
-9.69948590e-01 1.09841406e+00 5.90316868e+00 1.30633283e+00
-9.01136220e-01 -9.27308053e-02 1.28513944e+00 1.45309851e-01
-5.89115322e-01 -1.55596331e-01 -8.73220384e-01 7.19705224e-01
1.88828826e-01 -1.40817799e-02 6.29411876e-01 6.50346935e-01
3.39450806e-01 -2.78628916e-01 -5.73253453e-01 1.43971479e+00
3.54169071e-01 -1.15940046e+00 1.90389737e-01 -1.88777551e-01
1.08703136e+00 -5.24782538e-01 4.03361827e-01 9.28482935e-02
2.27211356e-01 -9.55879688e-01 4.58966821e-01 7.71043718e-01
9.76124167e-01 -6.87477708e-01 4.71701294e-01 -6.23435192e-02
-1.28524792e+00 1.45385534e-01 -6.23970985e-01 4.34329152e-01
1.19819053e-01 6.49689794e-01 -5.83589435e-01 5.94927192e-01
9.59376395e-01 9.23521876e-01 -9.81779456e-01 1.07711089e+00
-3.23978066e-01 3.30839783e-01 7.48514831e-02 3.42940204e-02
-4.99546342e-02 -3.69818568e-01 4.60493714e-01 1.03546238e+00
3.28421354e-01 4.21379387e-01 -3.85655239e-02 9.08756435e-01
-4.33849841e-01 2.75836706e-01 -1.69903964e-01 -7.46243298e-02
4.02753115e-01 1.76632321e+00 -8.36541176e-01 -3.51132482e-01
-9.60943475e-03 1.27057099e+00 -9.08546001e-02 6.06312037e-01
-1.02578378e+00 -2.10381821e-01 5.49979448e-01 3.80879909e-01
3.19270164e-01 -6.27029836e-02 -2.23454922e-01 -1.04095101e+00
1.51865348e-01 -8.21693838e-01 4.45829570e-01 -1.42652512e+00
-1.65880525e+00 6.77869141e-01 -2.31840312e-01 -1.35911894e+00
5.15382409e-01 -1.58112898e-01 -5.58556557e-01 7.07350731e-01
-1.67601848e+00 -1.22325897e+00 -7.12312877e-01 6.26625717e-01
8.89370024e-01 3.40270400e-01 1.09388798e-01 3.16572368e-01
-1.09288287e+00 7.44172931e-01 -1.90251753e-01 1.54023707e-01
8.85212660e-01 -9.45618451e-01 3.85412157e-01 1.10038257e+00
9.55408067e-02 1.17020972e-01 6.37358546e-01 -7.01311886e-01
-1.25063908e+00 -1.40818679e+00 6.08395159e-01 -3.39670867e-01
3.36791515e-01 -1.42345652e-01 -1.01938140e+00 4.81941640e-01
2.41309002e-01 -8.47912766e-03 2.87744373e-01 -3.02337229e-01
-1.74440995e-01 -6.12532973e-01 -1.19509912e+00 9.82530236e-01
1.00595105e+00 -1.79523796e-01 -3.02501947e-01 2.92017668e-01
7.58728802e-01 -4.47139353e-01 -9.41892505e-01 2.93709666e-01
4.32142735e-01 -8.17434251e-01 9.16407943e-01 1.72512412e-01
3.92778158e-01 -5.78296363e-01 1.72093749e-01 -1.18367720e+00
-5.37893355e-01 -8.61333072e-01 4.13883179e-01 1.82611787e+00
3.17069829e-01 -5.48513949e-01 6.43619001e-01 9.19651151e-01
4.47072014e-02 -5.79784513e-01 -5.36060750e-01 -5.82755923e-01
-3.87957722e-01 -1.53355569e-01 6.16014540e-01 8.83784950e-01
-5.07443190e-01 1.84418753e-01 -7.76917577e-01 3.46223652e-01
4.68608469e-01 1.52559146e-01 8.50550354e-01 -6.51775956e-01
-6.11320417e-03 -5.77611923e-01 -1.08299434e-01 -1.00031555e+00
-3.08602810e-01 -3.98739666e-01 1.63178995e-01 -1.69409931e+00
3.77726197e-01 -6.15599155e-01 -6.95833415e-02 5.36609590e-01
-4.61470544e-01 9.09016669e-01 2.98554659e-01 5.27945638e-01
-5.83740413e-01 4.89892840e-01 1.88270426e+00 -4.20049727e-01
-4.31440771e-01 -3.43601517e-02 -8.47056627e-01 4.20084685e-01
9.48705196e-01 -1.48969695e-01 -2.92317659e-01 -4.00450468e-01
1.58708334e-01 -2.81727552e-01 4.25675482e-01 -7.38854587e-01
1.25500798e-01 -5.84681928e-01 5.87148726e-01 -5.60865581e-01
2.73946017e-01 -8.56175601e-01 5.50346851e-01 6.92664087e-02
-2.86929190e-01 3.06557082e-02 2.00715423e-01 5.49704790e-01
9.07237902e-02 9.67515707e-02 9.91104186e-01 1.17426561e-02
-1.11398065e+00 6.29442394e-01 -1.34051278e-01 -2.50593759e-02
1.30336821e+00 -5.23744702e-01 -5.26397586e-01 -2.56095439e-01
-2.99249560e-01 2.68830508e-01 7.35712111e-01 4.90909308e-01
8.30107391e-01 -1.43255913e+00 -6.70102596e-01 1.02769984e-02
1.56795919e-01 -8.86444654e-03 9.45503473e-01 6.25725567e-01
-5.00818670e-01 -1.33838177e-01 -3.23328912e-01 -3.58559519e-01
-1.42359054e+00 8.49832475e-01 1.35921717e-01 4.53743413e-02
-8.92364144e-01 6.10796034e-01 3.19715619e-01 3.03108320e-02
7.18510672e-02 -4.26946551e-01 -2.99708784e-01 -4.23289463e-02
6.69412971e-01 2.98666030e-01 -3.35582733e-01 -7.62675464e-01
-5.14892004e-02 1.00629222e+00 -6.40257448e-02 1.30593732e-01
1.08147085e+00 -5.81110775e-01 5.01562692e-02 3.11066538e-01
9.47059095e-01 9.57743227e-02 -1.55853105e+00 -1.36019602e-01
-4.96931106e-01 -9.32317376e-01 -1.90239355e-01 -8.48506629e-01
-1.37815619e+00 4.11500841e-01 4.62668687e-01 3.22058469e-01
1.70319676e+00 4.36266698e-02 1.13291848e+00 -1.61125511e-01
1.98086426e-01 -1.15104270e+00 4.05089706e-01 -2.30046764e-01
1.17513525e+00 -1.15824747e+00 1.68330699e-01 -6.55243099e-01
-1.09996676e+00 7.61476994e-01 7.33151793e-01 -4.39242125e-01
1.41361326e-01 -1.79661378e-01 9.15260687e-02 8.34282041e-02
-3.82561415e-01 -2.50665277e-01 6.34718657e-01 7.32351542e-01
-1.44491553e-01 1.23889118e-01 -2.63002425e-01 4.84623581e-01
-7.76847228e-02 -2.00920135e-01 4.90045011e-01 2.29643777e-01
-3.31032276e-01 -8.30612540e-01 -6.34023845e-01 2.57193238e-01
-1.37272760e-01 2.67130397e-02 -5.13411641e-01 9.76461887e-01
2.21517339e-01 1.00492394e+00 -2.08832659e-02 -4.61715519e-01
3.19029331e-01 -5.92384040e-01 3.14622492e-01 -2.00915754e-01
-3.42925340e-01 1.10527694e-01 -1.53511748e-01 -4.22895283e-01
-5.52866697e-01 -2.94321209e-01 -8.35201919e-01 -6.04362369e-01
-1.54811040e-01 -3.08316685e-02 1.71180934e-01 5.17231166e-01
2.78916150e-01 6.20867610e-01 9.14902270e-01 -7.92496085e-01
2.47111976e-01 -8.37674618e-01 -8.40861082e-01 4.64638799e-01
5.90375327e-02 -2.01202720e-01 -3.05660367e-01 4.49746639e-01] | [11.243208885192871, -1.2784942388534546] |
6de4d837-4f82-492b-b1ca-ccffa05dddaf | ceil-generalized-contextual-imitation | 2306.14534 | null | https://arxiv.org/abs/2306.14534v1 | https://arxiv.org/pdf/2306.14534v1.pdf | CEIL: Generalized Contextual Imitation Learning | In this paper, we present \textbf{C}ont\textbf{E}xtual \textbf{I}mitation \textbf{L}earning~(CEIL), a general and broadly applicable algorithm for imitation learning (IL). Inspired by the formulation of hindsight information matching, we derive CEIL by explicitly learning a hindsight embedding function together with a contextual policy using the hindsight embeddings. To achieve the expert matching objective for IL, we advocate for optimizing a contextual variable such that it biases the contextual policy towards mimicking expert behaviors. Beyond the typical learning from demonstrations (LfD) setting, CEIL is a generalist that can be effectively applied to multiple settings including: 1)~learning from observations (LfO), 2)~offline IL, 3)~cross-domain IL (mismatched experts), and 4) one-shot IL settings. Empirically, we evaluate CEIL on the popular MuJoCo tasks (online) and the D4RL dataset (offline). Compared to prior state-of-the-art baselines, we show that CEIL is more sample-efficient in most online IL tasks and achieves better or competitive performances in offline tasks. | ['Huazhe Xu', 'Donglin Wang', 'Zifeng Zhuang', 'Yachen Kang', 'Li He', 'Jinxin Liu'] | 2023-06-26 | null | null | null | null | ['imitation-learning', 'd4rl'] | ['methodology', 'robots'] | [-3.85781899e-02 2.02550933e-01 -4.11483884e-01 -3.65679264e-01
-9.74017382e-01 -7.57306695e-01 6.20955527e-01 -2.44721219e-01
-9.52779531e-01 8.11100006e-01 1.35441601e-01 -6.40254915e-01
-2.57038146e-01 -1.14419714e-01 -1.28205931e+00 -4.49928999e-01
-9.69604701e-02 6.83904171e-01 -8.63307491e-02 -2.89869085e-02
2.63212770e-01 4.77993131e-01 -1.06980085e+00 -9.72667709e-02
1.01640224e+00 9.52033937e-01 4.28406119e-01 9.68138337e-01
1.08909264e-01 1.16952264e+00 -4.96629804e-01 -3.63383383e-01
4.68928963e-01 -2.77228981e-01 -8.10139298e-01 -2.14169651e-01
4.12030935e-01 -9.13716495e-01 -6.31534636e-01 1.04337823e+00
6.22564793e-01 4.53396320e-01 9.44528997e-01 -1.48806095e+00
-7.78919935e-01 5.74319541e-01 -4.16115701e-01 1.94310948e-01
5.10611951e-01 6.32718742e-01 1.01091671e+00 -1.02805054e+00
7.56882966e-01 1.57932818e+00 4.80162323e-01 6.19910955e-01
-1.27008951e+00 -6.91909373e-01 5.83363891e-01 8.40138197e-02
-7.17321515e-01 -2.23212421e-01 5.49059927e-01 -5.14507532e-01
9.75229979e-01 -3.49621296e-01 3.41145158e-01 1.88293195e+00
8.53606388e-02 1.58609092e+00 1.20458233e+00 -3.99174303e-01
2.23473579e-01 8.57623387e-03 -1.06852315e-01 8.80778015e-01
-1.77108809e-01 5.53344190e-01 -6.85504377e-01 -1.45183697e-01
8.95896375e-01 -3.37231420e-02 -2.29290262e-01 -6.14512026e-01
-1.48260295e+00 9.36059475e-01 3.56346518e-01 -2.39092335e-01
-2.30612308e-01 7.35166490e-01 3.64756972e-01 9.15152133e-01
2.37786368e-01 5.92706323e-01 -7.27608442e-01 -3.81308377e-01
-3.53704005e-01 7.31031179e-01 9.35990632e-01 1.38385487e+00
7.44010508e-01 1.50514677e-01 -2.59953946e-01 5.55319369e-01
3.55472803e-01 6.83485270e-01 4.36202615e-01 -1.50761890e+00
9.62989926e-01 1.10493854e-01 7.76652753e-01 -1.97404832e-01
-2.59759754e-01 -1.77870050e-01 -2.29460090e-01 5.31510651e-01
5.24649680e-01 -5.56263506e-01 -5.83045602e-01 2.08901215e+00
3.46429825e-01 1.29795164e-01 2.27363467e-01 9.11392570e-01
3.38461995e-01 4.64283288e-01 -1.12337276e-01 1.70287743e-01
6.35634720e-01 -1.44112980e+00 -5.28025568e-01 -3.70983630e-01
5.71944714e-01 -3.19816232e-01 1.45288253e+00 4.09219056e-01
-8.86269867e-01 -5.18726349e-01 -9.78368759e-01 -2.30556846e-01
-1.75803199e-01 2.91799188e-01 4.12990481e-01 -1.10858142e-01
-1.04528546e+00 9.36896682e-01 -7.49224126e-01 -2.77429938e-01
1.52578562e-01 3.07429969e-01 -2.25842029e-01 -3.74483503e-02
-1.27933896e+00 9.64900613e-01 2.62315154e-01 -1.31871700e-01
-1.79332721e+00 -8.08867216e-01 -9.89042044e-01 -2.90817380e-01
5.55954099e-01 -6.33156121e-01 1.86137772e+00 -1.03706849e+00
-2.11957049e+00 5.91642380e-01 9.57021043e-02 -6.86024964e-01
1.15489459e+00 -8.18594575e-01 -1.17623411e-01 3.04936379e-01
3.08635533e-01 9.27974164e-01 1.14629698e+00 -1.02115774e+00
-5.91547310e-01 -8.47164765e-02 3.11372548e-01 4.75855321e-01
-5.47744753e-03 4.03732620e-03 -2.75212020e-01 -6.74873650e-01
-6.34141922e-01 -1.00440812e+00 -6.15682602e-02 3.68250012e-01
-3.52319032e-01 -6.11764550e-01 7.24589169e-01 -5.80132902e-01
9.87147331e-01 -2.14241409e+00 4.92100835e-01 -1.35329425e-01
1.79419279e-01 2.19168171e-01 -3.37297559e-01 6.91535592e-01
2.49663606e-01 -1.97269574e-01 -1.14038944e-01 -5.44547856e-01
7.49892294e-01 2.72626281e-01 -5.04605234e-01 6.10343754e-01
-4.67201173e-02 1.12817585e+00 -1.22680700e+00 -4.20728952e-01
2.39183560e-01 -1.05578929e-01 -7.05662072e-01 7.46493638e-01
-7.05075204e-01 6.89443409e-01 -7.44512558e-01 5.45164168e-01
9.84649509e-02 -2.73580641e-01 -5.63432761e-02 3.46316636e-01
2.71438956e-02 1.57748573e-02 -9.59158123e-01 2.02323627e+00
-5.64615071e-01 7.25879967e-01 2.87147790e-01 -9.50678349e-01
6.05353713e-01 2.65171379e-01 1.88860029e-01 -3.51373434e-01
1.88622311e-01 2.08152696e-01 -2.78836310e-01 -7.04224348e-01
1.77182078e-01 1.35275975e-01 -2.52767235e-01 6.22719347e-01
3.26680839e-01 -1.27321035e-01 -2.59378046e-01 2.92072147e-01
1.05045760e+00 8.10536981e-01 2.28083655e-01 -2.57731080e-01
1.62329257e-01 -1.43772230e-01 4.19450849e-01 1.13778961e+00
-7.64054239e-01 1.15365580e-01 5.24485826e-01 -1.59296572e-01
-9.80639815e-01 -1.30080569e+00 2.38574579e-01 1.56159425e+00
3.03529799e-01 1.53044224e-01 -6.97092235e-01 -1.07541990e+00
5.10112286e-01 9.04311597e-01 -7.06321657e-01 -3.00977658e-02
-6.70382559e-01 2.47486189e-01 7.80876458e-01 7.08152354e-01
7.10115254e-01 -1.26317847e+00 -7.52225101e-01 2.83050686e-01
3.20730731e-02 -9.77903605e-01 -9.69686627e-01 2.83446431e-01
-5.38300276e-01 -9.56986368e-01 -8.40998173e-01 -7.50318527e-01
5.04611492e-01 -1.12654760e-01 1.01380622e+00 -5.63608110e-01
-4.04839553e-02 9.65462744e-01 -3.00255477e-01 -6.52374506e-01
-2.21988335e-01 -8.43695775e-02 5.87811530e-01 -2.75976509e-01
9.53152701e-02 -3.21090192e-01 -8.64577055e-01 2.44368121e-01
-5.62203586e-01 -1.12347685e-01 7.62010276e-01 1.02098417e+00
3.75612855e-01 -7.51606405e-01 6.86087072e-01 -7.66990721e-01
8.70917737e-01 -4.57904607e-01 -6.88479006e-01 2.46047929e-01
-6.94850385e-01 4.32010859e-01 1.06662750e+00 -9.24702346e-01
-1.26975417e+00 -3.03549707e-01 3.30436304e-02 -1.18156779e+00
-7.32549727e-02 9.06728283e-02 7.79172555e-02 8.30024108e-03
7.39572346e-01 1.75173879e-01 1.61829054e-01 -5.20637035e-01
6.32603526e-01 6.85140908e-01 5.28974056e-01 -1.23769975e+00
7.62611270e-01 3.37317824e-01 -4.50164378e-01 -3.13364774e-01
-9.63975668e-01 -3.49194765e-01 -4.52816695e-01 -1.87146336e-01
9.34050381e-01 -9.67907846e-01 -1.02949214e+00 2.71636069e-01
-9.10464704e-01 -1.20634580e+00 -2.88666904e-01 6.91112041e-01
-1.19499409e+00 3.39144588e-01 -7.52098680e-01 -7.68118501e-01
-7.28224814e-02 -1.33059514e+00 1.09103525e+00 1.14982054e-01
-3.31550315e-02 -1.14725792e+00 2.18444213e-01 1.72895238e-01
1.27091795e-01 1.60517469e-01 6.86618507e-01 -7.49307394e-01
-5.69233418e-01 2.37317249e-01 3.59323435e-02 4.45039898e-01
-3.98539513e-01 -3.53577107e-01 -7.23976016e-01 -7.10611820e-01
-2.32323214e-01 -1.15194011e+00 5.77523708e-01 1.28238335e-01
1.16048789e+00 -5.56850672e-01 -2.36718625e-01 7.04586327e-01
1.14807522e+00 1.93967938e-01 -2.38052625e-02 4.61280763e-01
5.00295043e-01 4.44964677e-01 1.09541368e+00 5.97181737e-01
3.98033798e-01 5.08464932e-01 4.45464700e-01 1.83699012e-01
2.23182157e-01 -9.07184839e-01 9.40036952e-01 4.01212841e-01
3.32559109e-01 -2.11465508e-01 -6.00184500e-01 6.73083842e-01
-2.29300690e+00 -7.80550301e-01 5.94283044e-01 1.85122681e+00
9.33187187e-01 -6.66130409e-02 2.04102084e-01 -7.25427449e-01
3.78083944e-01 2.69744813e-01 -1.32666385e+00 -4.78058219e-01
2.81558394e-01 7.21657649e-02 4.99555260e-01 5.73402107e-01
-1.14989185e+00 1.05284405e+00 6.51064110e+00 6.91540360e-01
-7.23861754e-01 1.69365495e-01 1.93578020e-01 -1.94081083e-01
-2.20230758e-01 -4.29964401e-02 -8.78917038e-01 4.15529817e-01
7.14413226e-01 -5.78687973e-02 9.34809744e-01 1.19148695e+00
-9.19378828e-03 9.48346555e-02 -1.75059903e+00 8.58285487e-01
-6.95582330e-02 -9.37422216e-01 -2.50498652e-01 -1.50005415e-01
7.86528826e-01 3.98811668e-01 3.60824287e-01 9.87098157e-01
1.11579859e+00 -9.39352453e-01 1.03141904e+00 4.12083596e-01
1.10754204e+00 -5.17408669e-01 1.89237401e-01 7.94431031e-01
-8.86891365e-01 -4.11572754e-01 -4.01100904e-01 4.06395793e-02
-2.04279069e-02 -2.13915989e-01 -6.86744452e-01 4.23592716e-01
8.49120438e-01 7.45398283e-01 4.05294336e-02 3.13833296e-01
-7.22877026e-01 4.04708654e-01 -1.57117561e-01 -3.89754206e-01
7.67134964e-01 -2.69952416e-01 6.49612486e-01 1.14059544e+00
-1.03259526e-01 -1.52051710e-02 4.78512526e-01 1.04941678e+00
-2.70912826e-01 -3.23045492e-01 -8.45585942e-01 -1.11507520e-01
5.95768452e-01 7.36525238e-01 3.21024239e-01 -2.63274223e-01
-4.14200783e-01 1.30648696e+00 9.46260870e-01 8.60544324e-01
-1.07677829e+00 -5.95003545e-01 9.25891757e-01 -4.64394510e-01
5.55749059e-01 -2.57186294e-01 5.24643719e-01 -1.35063398e+00
-7.07040727e-02 -1.15168452e+00 3.30367446e-01 -6.75316930e-01
-1.39989948e+00 8.05553421e-02 2.16271564e-01 -1.26125288e+00
-5.76595247e-01 -8.74858856e-01 -5.08454680e-01 6.88558638e-01
-1.62872577e+00 -9.18254018e-01 1.52041376e-01 8.99970651e-01
7.49456644e-01 -3.30934048e-01 3.74457866e-01 -4.71702665e-02
-5.37974596e-01 7.77187169e-01 4.65434551e-01 1.17250510e-01
9.87334132e-01 -1.72979975e+00 2.82645941e-01 6.17530465e-01
4.46078852e-02 6.46686077e-01 7.15943456e-01 -3.72731417e-01
-1.62658453e+00 -1.05090237e+00 3.41536492e-01 -6.56776845e-01
1.12599444e+00 -4.15687591e-01 -4.38712746e-01 1.44631433e+00
1.92651555e-01 -5.24134561e-02 2.42970645e-01 -1.86868057e-01
-4.27768588e-01 -1.28651052e-04 -1.20239937e+00 9.03321266e-01
1.38168907e+00 -8.17957878e-01 -8.88254225e-01 4.54556227e-01
1.16888750e+00 -6.02182746e-01 -1.01122677e+00 1.53366804e-01
5.96731842e-01 -7.30571210e-01 1.01456559e+00 -9.79958832e-01
4.33766693e-01 -2.63902843e-02 -1.13815337e-01 -1.44301713e+00
-8.46021101e-02 -1.17475677e+00 -5.51044583e-01 7.48651862e-01
2.60094106e-01 -9.05466855e-01 3.68555814e-01 4.92974818e-01
-2.38744304e-01 -9.65560257e-01 -9.66654778e-01 -9.55254674e-01
4.87580925e-01 -2.25891888e-01 2.75166512e-01 5.88173211e-01
3.58393714e-02 1.23769301e-03 -6.31567836e-01 1.40913278e-01
9.19151247e-01 3.06716919e-01 1.16113448e+00 -6.36957169e-01
-1.00664186e+00 -3.17265183e-01 4.96141016e-01 -1.96828592e+00
8.13530624e-01 -8.22308540e-01 2.95444369e-01 -1.12664521e+00
4.43394706e-02 -3.90803754e-01 -3.70085746e-01 2.50959933e-01
-1.60417989e-01 -5.78426301e-01 3.29614103e-01 2.82198608e-01
-9.31021035e-01 8.06793332e-01 1.43275845e+00 -1.45248871e-03
3.99435274e-02 -7.15766326e-02 -4.02817369e-01 7.59514809e-01
4.50779498e-01 -2.29511887e-01 -6.53705001e-01 -7.71236956e-01
3.02518737e-02 3.53761971e-01 3.37360501e-01 -4.23025757e-01
3.03736866e-01 -2.62179047e-01 2.38187715e-01 -4.43617970e-01
3.60091865e-01 -6.07566714e-01 -4.32289630e-01 5.76544285e-01
-6.81021154e-01 1.04043193e-01 -1.46878123e-01 1.08375609e+00
1.18969508e-01 -2.91373670e-01 7.30525911e-01 -3.24942470e-01
-9.36757743e-01 5.61522841e-01 -2.80745745e-01 4.82183754e-01
1.03360796e+00 1.23647600e-01 -4.79545206e-01 -5.55862904e-01
-5.84068060e-01 9.49959934e-01 4.95845675e-01 3.12574327e-01
6.70755148e-01 -1.11690760e+00 -4.27094907e-01 -5.91806509e-02
2.00904116e-01 1.16152868e-01 -1.21346906e-01 8.24141979e-01
-3.55251908e-01 4.24878120e-01 1.24543920e-01 -5.56825936e-01
-5.65177143e-01 6.85478747e-01 4.26203310e-01 -5.22847116e-01
-7.86824524e-01 9.89786327e-01 4.86344725e-01 -1.08244419e+00
7.69078732e-01 -3.24595720e-01 2.31506571e-01 -4.48758423e-01
1.80291891e-01 5.97073555e-01 -7.14480639e-01 7.99962655e-02
-1.66792393e-01 3.57998133e-01 -1.62129864e-01 -5.44169128e-01
1.11853123e+00 -4.17158678e-02 5.84469259e-01 6.79171383e-01
1.24469137e+00 -3.91861469e-01 -2.21757984e+00 -4.04499859e-01
8.13517421e-02 -3.62988442e-01 -4.65439528e-01 -8.70660782e-01
-3.18038940e-01 8.93216431e-01 2.91962773e-01 -5.22183001e-01
5.52631617e-01 4.80344370e-02 6.10133171e-01 8.55266213e-01
7.53408253e-01 -1.40302920e+00 8.06986034e-01 5.13548553e-01
1.01195955e+00 -1.37119210e+00 -1.96471125e-01 6.18508637e-01
-1.02200067e+00 8.10493767e-01 7.43830204e-01 -5.06951332e-01
4.70180273e-01 9.79788154e-02 -4.57422286e-02 1.56709645e-02
-9.82782722e-01 -9.76260081e-02 -3.04908603e-02 5.65138638e-01
-1.34119809e-01 4.80059199e-02 6.31725565e-02 3.81647229e-01
1.21633142e-01 -1.00954426e-02 6.80112168e-02 1.06612682e+00
-4.56581950e-01 -9.28155124e-01 -7.07525909e-02 2.67683268e-01
-3.43519241e-01 2.72894621e-01 -4.24343169e-01 1.13500774e+00
-2.65462548e-01 8.54776084e-01 -2.26533353e-01 -1.27546534e-01
3.43965024e-01 -3.04901004e-02 6.67972267e-01 -3.34988147e-01
-5.99962473e-01 -1.45791396e-01 6.14572726e-02 -1.01077902e+00
-1.15795635e-01 -6.88058078e-01 -1.22329235e+00 -1.95428699e-01
-3.91316265e-02 -1.11251473e-01 3.63001466e-01 8.16813529e-01
3.37383509e-01 2.29290634e-01 8.24859500e-01 -9.65917528e-01
-1.56435037e+00 -1.03480220e+00 -5.14340460e-01 5.66831052e-01
6.71209216e-01 -8.70615482e-01 -8.12531710e-01 -4.27930772e-01] | [4.1255645751953125, 1.9578382968902588] |
8af4338e-39d4-48bb-ac38-d482d82e045e | meta-generative-attack-on-person | 2301.06286 | null | https://arxiv.org/abs/2301.06286v1 | https://arxiv.org/pdf/2301.06286v1.pdf | Meta Generative Attack on Person Reidentification | Adversarial attacks have been recently investigated in person re-identification. These attacks perform well under cross dataset or cross model setting. However, the challenges present in cross-dataset cross-model scenario does not allow these models to achieve similar accuracy. To this end, we propose our method with the goal of achieving better transferability against different models and across datasets. We generate a mask to obtain better performance across models and use meta learning to boost the generalizability in the challenging cross-dataset cross-model setting. Experiments on Market-1501, DukeMTMC-reID and MSMT-17 demonstrate favorable results compared to other attacks. | ['A V Subramanyam'] | 2023-01-16 | null | null | null | null | ['person-re-identification'] | ['computer-vision'] | [-7.16413110e-02 -6.11377120e-01 -2.82655954e-01 -5.57512999e-01
-7.08926678e-01 -8.51547837e-01 9.02789593e-01 -1.35864586e-01
-5.20861626e-01 7.48984635e-01 -9.82639566e-03 -7.31157064e-02
-7.19787404e-02 -5.16825557e-01 -6.19774461e-01 -2.01329663e-01
-1.50112122e-01 3.80151093e-01 5.45618646e-02 -2.93089956e-01
1.50953243e-02 3.51504326e-01 -7.40610361e-01 4.32107002e-01
7.68871903e-01 1.76494807e-01 -6.63902640e-01 6.97668910e-01
4.45517570e-01 3.01767766e-01 -6.42635942e-01 -1.33109534e+00
8.32356334e-01 -1.72622114e-01 -7.94582844e-01 -4.03977662e-01
9.40559804e-01 -2.89408892e-01 -6.75018907e-01 1.15415168e+00
1.11734319e+00 -1.56263053e-01 7.56128967e-01 -1.67686796e+00
-7.35388994e-01 6.26107574e-01 -9.27392662e-01 4.63340849e-01
5.76625943e-01 2.35381499e-01 4.43714976e-01 -5.11834621e-01
3.02822024e-01 1.69037187e+00 1.07897615e+00 1.19727361e+00
-1.24856782e+00 -1.39334035e+00 2.15216517e-01 1.05737615e-02
-1.48200238e+00 -3.94169211e-01 4.28790450e-01 -1.44282430e-01
5.74764252e-01 3.01476777e-01 -4.90480363e-02 1.99359870e+00
5.56907132e-02 7.13069916e-01 1.58113539e+00 2.40695830e-02
-3.44960541e-01 3.95447195e-01 4.10648763e-01 2.64408112e-01
3.94205987e-01 4.22850430e-01 -3.14744651e-01 -5.15830457e-01
3.09631169e-01 3.88815254e-02 -1.33882672e-01 4.60273065e-02
-7.85586298e-01 7.25286841e-01 2.39298478e-01 1.38861001e-01
1.12448886e-01 -1.24846399e-01 8.46112669e-01 5.64138293e-01
3.19011629e-01 3.51451874e-01 -4.93081808e-01 6.82255179e-02
-8.57055783e-01 6.54733181e-01 7.06476212e-01 8.84288967e-01
1.93842426e-01 6.53014481e-02 -2.14584768e-01 9.17151988e-01
-1.25662023e-02 6.25880837e-01 5.94934106e-01 -3.71415138e-01
1.00987124e+00 2.72479177e-01 1.02231398e-01 -1.00713801e+00
-3.95017654e-01 -7.06631422e-01 -1.23671710e+00 5.11630103e-02
4.78777349e-01 -3.69922638e-01 -9.03916955e-01 1.98119473e+00
2.43651211e-01 8.96840870e-01 2.83764541e-01 4.58597064e-01
6.92496300e-01 2.76134193e-01 3.75003815e-01 3.27356964e-01
1.38619471e+00 -7.95506477e-01 -4.68802661e-01 -2.82172501e-01
7.26446271e-01 -7.23818064e-01 6.67696834e-01 1.35744274e-01
-8.77315521e-01 -8.12492728e-01 -1.14071143e+00 3.13453108e-01
-7.14149833e-01 -4.62057814e-02 3.23647529e-01 1.28304100e+00
-6.15185499e-01 4.84281480e-01 -3.87493193e-01 -5.81400335e-01
6.13283396e-01 5.85270286e-01 -7.39856243e-01 4.68834266e-02
-1.53688002e+00 1.11996424e+00 4.11662579e-01 -9.72674638e-02
-8.54808211e-01 -1.18298101e+00 -5.78571439e-01 -2.53247857e-01
-4.24280688e-02 -7.38332629e-01 9.37479913e-01 -6.82979941e-01
-1.04650450e+00 1.14223516e+00 1.61710113e-01 -1.07769239e+00
1.04130197e+00 -3.86566669e-01 -9.69423056e-01 -3.73858571e-01
8.14522728e-02 7.34850347e-01 6.44980311e-01 -1.38990569e+00
-5.23088932e-01 -4.88776386e-01 2.20105588e-01 -9.35370624e-02
-3.25802237e-01 1.05463386e-01 -2.96938151e-01 -8.41140509e-01
-5.60310006e-01 -1.29011643e+00 1.09952632e-02 -6.64055228e-01
-6.88030601e-01 1.30410358e-01 9.46311295e-01 -7.81595111e-01
1.07620513e+00 -1.93872356e+00 -1.23943076e-01 2.89565563e-01
-1.00529119e-02 8.80852818e-01 -2.45389342e-01 4.56833482e-01
-4.40952301e-01 4.64810342e-01 -4.14120629e-02 -7.40795851e-01
-1.21224560e-01 1.02943748e-01 -3.42519462e-01 4.75032955e-01
1.42124757e-01 8.98258328e-01 -4.36602473e-01 -4.84638691e-01
1.20875880e-01 2.75199592e-01 -3.79137605e-01 1.44337595e-01
4.79051679e-01 6.53307557e-01 -4.82394844e-01 6.29717588e-01
1.33806205e+00 2.00663269e-01 -6.47476315e-02 -2.02099293e-01
4.30418342e-01 -1.77461237e-01 -1.26943183e+00 1.26061940e+00
-3.05412173e-01 1.90936983e-01 -9.13531780e-02 -8.04445624e-01
6.53111517e-01 9.96504575e-02 1.51171207e-01 -6.39514804e-01
1.42497629e-01 -7.43573159e-02 2.88152486e-01 -3.17261010e-01
4.86747444e-01 -5.58370873e-02 -3.17216218e-01 3.15043718e-01
4.30258550e-02 5.76699853e-01 5.13543487e-02 3.05485308e-01
6.40048862e-01 -1.90592021e-01 1.17481180e-01 -2.76586533e-01
7.37369597e-01 -3.71022075e-01 5.39935291e-01 1.14938939e+00
-6.75754547e-01 7.28811324e-01 3.12994570e-02 -3.97710025e-01
-9.42777872e-01 -1.04190230e+00 -2.69057035e-01 9.16520238e-01
2.24200130e-01 -3.05973917e-01 -8.78446877e-01 -1.20207930e+00
4.53827500e-01 4.15293336e-01 -9.70394969e-01 -3.81102562e-01
-7.25022078e-01 -1.21042037e+00 1.55839789e+00 6.01906180e-01
9.94429350e-01 -3.01878691e-01 3.38194966e-01 -8.65664855e-02
-9.51767862e-02 -1.47248936e+00 -8.50947976e-01 -4.16177750e-01
-5.06351352e-01 -1.25392413e+00 -8.63410115e-01 -7.37043262e-01
3.00922751e-01 2.20727891e-01 9.37847972e-01 3.99741530e-02
-2.27413654e-01 3.81566167e-01 -1.71202272e-01 -4.60270941e-01
-5.68612278e-01 3.78325760e-01 5.72757363e-01 1.99459255e-01
4.60653722e-01 -4.16556567e-01 -4.91674691e-01 6.33221507e-01
-7.32427597e-01 -2.29321569e-01 3.02793264e-01 7.72240818e-01
-1.11510538e-01 -1.70203652e-02 8.15078616e-01 -1.24962234e+00
5.62749386e-01 -6.71284199e-01 -2.53121287e-01 5.84546983e-01
-6.03288293e-01 -1.93226069e-01 5.55838108e-01 -9.40904140e-01
-1.09821415e+00 -4.48003232e-01 -1.98190868e-01 -2.31620878e-01
-2.10497193e-02 -1.65296629e-01 -3.83892059e-01 -2.49286056e-01
6.91559434e-01 1.04585424e-01 -1.49924546e-01 -5.96500337e-01
8.14616382e-02 7.26371288e-01 6.98614120e-01 -6.60915971e-01
1.38477993e+00 3.51555973e-01 -1.23171933e-01 -1.60201430e-01
-6.47243679e-01 -3.22668850e-01 -7.56586552e-01 1.82604492e-01
8.13724637e-01 -1.29940903e+00 -7.72710562e-01 9.10178542e-01
-8.70975494e-01 -1.25269309e-01 5.45940399e-01 3.06983829e-01
-2.35254597e-02 5.76172829e-01 -7.37388432e-01 -5.90165496e-01
-6.84423029e-01 -1.11170173e+00 6.99619710e-01 4.71863925e-01
-1.87226266e-01 -1.21702456e+00 1.11746103e-01 6.04492903e-01
4.61102962e-01 4.35325712e-01 4.38798785e-01 -1.25321603e+00
-2.66116485e-03 -5.25297701e-01 -3.34281266e-01 3.31029564e-01
2.34922349e-01 -1.52191699e-01 -1.11194611e+00 -8.33395958e-01
-4.98560041e-01 -3.19720030e-01 7.79203534e-01 -1.25428945e-01
1.09480882e+00 -2.55873501e-01 -5.51820159e-01 6.86152220e-01
1.22825015e+00 -1.58204615e-01 5.91745377e-01 5.56880355e-01
8.75758171e-01 4.27004337e-01 2.72516191e-01 2.34551311e-01
4.79225218e-01 9.55968559e-01 3.21049020e-02 -8.49134922e-02
-1.43969983e-01 -5.01916409e-01 2.29607403e-01 7.43229389e-02
1.75421283e-01 -3.18711728e-01 -8.96314204e-01 3.72987837e-01
-1.74252439e+00 -1.32917976e+00 -4.27813306e-02 2.19959497e+00
7.63603687e-01 2.77951956e-01 6.38869882e-01 -1.95888817e-01
1.05603051e+00 -6.10671937e-03 -3.99636507e-01 -4.27464008e-01
-3.83203149e-01 4.22301330e-02 8.22801709e-01 4.90665048e-01
-1.36500764e+00 1.01090527e+00 6.88954544e+00 8.94160867e-01
-8.63791406e-01 3.03814501e-01 8.88487458e-01 -8.67698193e-02
2.60673046e-01 -2.24538341e-01 -1.35274386e+00 7.42270827e-01
1.02703130e+00 -2.90921330e-01 1.69030890e-01 4.57070023e-01
-2.28345037e-01 4.58941579e-01 -1.16580391e+00 1.16854203e+00
2.60210037e-01 -9.14068282e-01 2.63854474e-01 1.62473470e-01
1.02408433e+00 -7.00462535e-02 4.82880116e-01 8.38445425e-01
5.30807853e-01 -1.25225604e+00 2.67582953e-01 2.38606542e-01
5.26287556e-01 -9.15906549e-01 8.87569904e-01 2.92965770e-01
-8.81162167e-01 -8.17353725e-02 -4.02694464e-01 3.79726112e-01
6.36721179e-02 -5.93705848e-02 -8.40164125e-01 9.52875614e-01
6.86443031e-01 5.83020568e-01 -1.10167551e+00 8.95693898e-01
4.84957188e-01 4.83193129e-01 -1.63397565e-01 4.88230824e-01
2.72102896e-02 2.54972011e-01 4.15896893e-01 1.52124715e+00
-8.49528760e-02 -1.78306937e-01 2.74875700e-01 4.70094889e-01
-2.55766541e-01 -7.91879594e-02 -6.27931833e-01 3.09215844e-01
4.97839749e-01 9.76471543e-01 1.32161621e-02 -3.51106912e-01
-3.76451850e-01 1.16426778e+00 2.72799253e-01 3.70370209e-01
-1.21142542e+00 2.61344612e-02 1.02543402e+00 -3.26895081e-02
-1.67419210e-01 3.69938165e-02 -1.37107730e-01 -1.39452672e+00
-1.30263656e-01 -1.34872794e+00 8.01336527e-01 -1.79602325e-01
-1.95983386e+00 6.23731792e-01 4.32018667e-01 -1.30862296e+00
-3.48678499e-01 -5.85225940e-01 -7.17126369e-01 1.10225093e+00
-1.41112018e+00 -1.82594657e+00 -7.90334418e-02 8.76677871e-01
2.76470721e-01 -6.28574908e-01 6.79981709e-01 6.71265244e-01
-9.24348831e-01 1.71905327e+00 3.36975157e-02 6.66903436e-01
1.24548388e+00 -1.18547511e+00 8.87726188e-01 1.08016384e+00
-1.01793848e-01 9.44435894e-01 6.98392332e-01 -8.86771560e-01
-1.16263342e+00 -1.27615690e+00 2.80242503e-01 -9.59898472e-01
4.14682329e-01 -5.36260605e-01 -9.82235312e-01 8.89662385e-01
2.60823995e-01 3.17495130e-02 9.07823324e-01 2.92873323e-01
-1.08311832e+00 -2.04332978e-01 -1.67348766e+00 6.09466791e-01
1.03386629e+00 -6.99492157e-01 -3.88628721e-01 1.63936391e-01
3.27628285e-01 -5.80262005e-01 -1.04958761e+00 6.81382418e-01
7.41268039e-01 -7.09858119e-01 1.51942921e+00 -1.26520181e+00
-1.35268673e-01 -7.95639306e-02 -2.01252297e-01 -1.22548699e+00
-1.61825567e-01 -7.46365666e-01 5.60932085e-02 2.00338888e+00
4.93427545e-01 -8.02231789e-01 7.55275011e-01 8.44100237e-01
5.22644758e-01 -2.37984091e-01 -8.84382844e-01 -1.11847210e+00
6.63436174e-01 -3.29735249e-01 8.53077650e-01 1.24593544e+00
-3.43919933e-01 2.65109986e-02 -1.15286160e+00 6.21778607e-01
1.07578075e+00 -2.52272904e-01 1.31884432e+00 -1.05143738e+00
-3.57045949e-01 -2.20368579e-01 -4.44409877e-01 -4.41849679e-01
5.26342392e-01 -9.23241735e-01 -8.04404676e-01 -6.53147519e-01
6.59956038e-01 -5.31374454e-01 -3.14827979e-01 3.56559962e-01
-4.82783794e-01 6.34830415e-01 4.50535953e-01 1.86670080e-01
-4.39190656e-01 3.40398580e-01 8.23685229e-01 -4.11804020e-01
-1.25447437e-02 2.49548793e-01 -8.60331893e-01 3.35908830e-01
1.08191025e+00 -4.35210168e-01 -2.08772227e-01 -3.24688166e-01
-3.96549761e-01 -4.23804343e-01 5.39972544e-01 -1.12433767e+00
1.89263821e-01 1.24542415e-01 8.22672248e-01 -3.57506216e-01
3.18952531e-01 -6.67424142e-01 3.30366194e-01 5.39804757e-01
-3.87713850e-01 4.98408437e-01 5.35738111e-01 5.84670544e-01
1.11256279e-01 -7.63144791e-02 1.04221749e+00 -2.09432561e-02
-3.91535431e-01 7.61822104e-01 3.72911781e-01 3.15377176e-01
9.72968996e-01 -1.93854615e-01 -5.98956048e-01 -2.02768341e-01
-7.25998998e-01 5.82076132e-01 4.18655425e-01 1.01522732e+00
1.70852333e-01 -1.52169728e+00 -1.24452353e+00 2.61176586e-01
3.48471791e-01 -9.39524770e-01 5.73044181e-01 4.54421997e-01
-9.61793512e-02 3.30787122e-01 -5.10190845e-01 -4.83790696e-01
-1.69003773e+00 7.06240177e-01 7.94613779e-01 -4.87862319e-01
-2.61385739e-01 5.29591799e-01 1.97122872e-01 -8.66119146e-01
6.62914515e-02 3.75187844e-01 -1.01704195e-01 -2.16163337e-01
8.84128094e-01 5.32062650e-01 -1.69414699e-01 -1.03419757e+00
-5.51533163e-01 4.68941987e-01 -7.73987532e-01 6.83433488e-02
9.40786481e-01 -4.99327481e-02 4.18224126e-01 -1.79778546e-01
1.50951695e+00 9.63339210e-02 -9.58352029e-01 -2.46738046e-01
-6.23461865e-02 -6.68350756e-01 -5.98153591e-01 -9.03656006e-01
-9.01946306e-01 4.91655409e-01 1.15756571e+00 1.09727867e-03
6.83358192e-01 -4.24301565e-01 8.88786614e-01 1.17566720e-01
3.46248567e-01 -8.10805559e-01 -5.11546806e-02 2.99427390e-01
6.70814037e-01 -1.73549044e+00 6.21845201e-02 -4.46686834e-01
-7.17800498e-01 6.09533548e-01 1.05738771e+00 -9.93424281e-02
6.23177171e-01 1.53948441e-01 3.20933089e-02 2.57445693e-01
-3.63423347e-01 7.31226653e-02 3.51538956e-01 9.04620349e-01
1.66132480e-01 7.74998814e-02 -1.18916303e-01 4.48195726e-01
-2.29350686e-01 -2.00881198e-01 3.51759940e-01 5.75238764e-01
4.35245991e-01 -1.70925403e+00 -4.61349517e-01 1.37701422e-01
-1.08278036e+00 1.11828521e-01 -5.11553109e-01 1.16946268e+00
9.08250883e-02 1.04863036e+00 -3.57931554e-01 -7.18699634e-01
4.68232602e-01 6.86003342e-02 5.55381477e-01 -1.34340346e-01
-1.18809199e+00 -5.79300582e-01 2.03824013e-01 -1.73940569e-01
-3.94530773e-01 -6.04440272e-01 -5.18955171e-01 -8.26379895e-01
-1.21653803e-01 1.21598884e-01 2.74926782e-01 8.61218095e-01
4.75649446e-01 -1.44880060e-02 7.55736113e-01 -3.52900922e-01
-8.74398232e-01 -1.28175902e+00 -3.00569594e-01 1.02044225e+00
1.82711691e-01 -6.02520645e-01 -2.16872245e-01 -2.60701776e-01] | [14.66025447845459, 1.0205045938491821] |
9c7aeba3-0e9b-4d1f-8628-fbda913650d1 | unifying-flow-stereo-and-depth-estimation | 2211.05783 | null | https://arxiv.org/abs/2211.05783v1 | https://arxiv.org/pdf/2211.05783v1.pdf | Unifying Flow, Stereo and Depth Estimation | We present a unified formulation and model for three motion and 3D perception tasks: optical flow, rectified stereo matching and unrectified stereo depth estimation from posed images. Unlike previous specialized architectures for each specific task, we formulate all three tasks as a unified dense correspondence matching problem, which can be solved with a single model by directly comparing feature similarities. Such a formulation calls for discriminative feature representations, which we achieve using a Transformer, in particular the cross-attention mechanism. We demonstrate that cross-attention enables integration of knowledge from another image via cross-view interactions, which greatly improves the quality of the extracted features. Our unified model naturally enables cross-task transfer since the model architecture and parameters are shared across tasks. We outperform RAFT with our unified model on the challenging Sintel dataset, and our final model that uses a few additional task-specific refinement steps outperforms or compares favorably to recent state-of-the-art methods on 10 popular flow, stereo and depth datasets, while being simpler and more efficient in terms of model design and inference speed. | ['Andreas Geiger', 'DaCheng Tao', 'Fisher Yu', 'Hamid Rezatofighi', 'Jianfei Cai', 'Jing Zhang', 'Haofei Xu'] | 2022-11-10 | null | null | null | null | ['stereo-depth-estimation', 'stereo-matching-1'] | ['computer-vision', 'computer-vision'] | [ 9.36643332e-02 -5.76758757e-02 -1.32247671e-01 -4.26643848e-01
-6.43949151e-01 -5.46188831e-01 8.95329416e-01 -4.54405457e-01
-5.00380218e-01 5.74297130e-01 7.18934417e-01 1.36054412e-01
-4.28271368e-02 -4.99981284e-01 -6.86367691e-01 -3.54567349e-01
2.73862094e-01 5.22248507e-01 2.84013540e-01 -1.51778370e-01
4.68142927e-01 3.46486032e-01 -1.68919528e+00 2.35735998e-01
7.74476111e-01 8.87889028e-01 4.74711865e-01 8.90114844e-01
2.06388667e-01 1.06162310e+00 -1.50900623e-02 -2.73904681e-01
5.08856595e-01 -1.08132243e-01 -1.30312872e+00 1.05687149e-01
1.41715252e+00 -8.49168003e-01 -7.88184822e-01 6.60118341e-01
6.05106473e-01 4.71755743e-01 5.70785999e-01 -1.11315584e+00
-6.34141386e-01 -2.24158570e-01 -5.18503129e-01 4.27815735e-01
7.24400938e-01 3.13167781e-01 1.26568007e+00 -1.01233947e+00
1.00941551e+00 1.68997216e+00 5.59508204e-01 7.22955585e-01
-1.61882675e+00 -3.23214173e-01 4.08254474e-01 4.55352306e-01
-8.91833305e-01 -8.57485056e-01 4.89931613e-01 -8.40323567e-01
1.41894901e+00 -1.18962437e-01 7.43222833e-01 1.16108084e+00
1.21126316e-01 9.62259650e-01 8.00035834e-01 1.04237208e-02
-1.63345993e-01 -2.08173692e-01 -6.01993725e-02 7.90685773e-01
3.89866680e-02 4.34456050e-01 -8.53365302e-01 2.10335016e-01
1.31720114e+00 -9.63733867e-02 -6.41805768e-01 -8.21664810e-01
-1.49392545e+00 7.49904931e-01 6.27829134e-01 2.66468972e-02
-7.08627850e-02 3.46592426e-01 3.71505082e-01 2.69829959e-01
5.05854011e-01 3.60501170e-01 -5.17952681e-01 -2.31658578e-01
-7.91576087e-01 5.27593315e-01 7.42620230e-01 1.04497612e+00
1.10216951e+00 2.92134099e-02 -1.73783019e-01 5.87722540e-01
2.85278887e-01 3.16778839e-01 2.63550490e-01 -1.60591173e+00
6.13077343e-01 2.51990646e-01 2.35081717e-01 -9.09926236e-01
-3.93037528e-01 -2.57944703e-01 -6.81098580e-01 3.83112967e-01
5.44130862e-01 1.08301222e-01 -8.52397323e-01 1.94734073e+00
2.55345881e-01 6.21867836e-01 -4.03327756e-02 1.12810230e+00
9.38518226e-01 4.38680261e-01 -2.07623333e-01 3.65945399e-01
1.08076394e+00 -1.40454733e+00 -4.43264842e-01 -6.43659234e-01
5.89376569e-01 -6.96835577e-01 8.53727162e-01 1.40339971e-01
-1.29172099e+00 -9.25484300e-01 -8.30614626e-01 -9.94145334e-01
-9.77755412e-02 -3.82990837e-01 8.76069546e-01 1.36458054e-01
-1.46006358e+00 8.08046162e-01 -8.46980512e-01 -4.05037642e-01
5.61507881e-01 4.40689147e-01 -8.01221430e-01 -3.39061230e-01
-7.44267106e-01 1.04050922e+00 -7.11292867e-03 3.30574997e-02
-9.43431556e-01 -1.22306025e+00 -1.47498071e+00 -9.70048979e-02
-2.98708491e-02 -1.73826265e+00 1.21849048e+00 -6.08914256e-01
-1.54841244e+00 1.23867559e+00 -6.97168171e-01 -1.82672828e-01
6.55325174e-01 -6.53980136e-01 1.96306527e-01 1.77496761e-01
3.50364625e-01 1.36313462e+00 8.29129517e-01 -1.02603054e+00
-7.17615664e-01 -4.45542365e-01 4.01466906e-01 4.22704428e-01
2.52502978e-01 -3.69043440e-01 -5.35445154e-01 -3.94045889e-01
2.00494509e-02 -8.14866662e-01 -2.54562855e-01 4.18662727e-01
-9.43012759e-02 2.64677238e-02 6.56716168e-01 -3.81107897e-01
7.08462298e-01 -1.95004344e+00 8.16422224e-01 -3.06174397e-01
5.30814290e-01 -7.70100057e-02 -3.00955296e-01 -2.34423745e-02
2.00990718e-02 -2.17651442e-01 -1.24675393e-01 -9.13122296e-01
6.69935020e-04 3.44534963e-01 -1.81258738e-01 5.50874174e-01
3.56179118e-01 1.11191237e+00 -1.13981390e+00 -3.80128384e-01
6.84076548e-01 6.22164428e-01 -1.21897364e+00 5.48237503e-01
1.10124148e-01 9.13254201e-01 -2.84554362e-01 3.08281630e-01
6.69427812e-01 -4.53483969e-01 -3.35315287e-01 -5.22513509e-01
-8.39880481e-02 6.02688074e-01 -1.15755200e+00 2.49676943e+00
-7.57840574e-01 8.98117781e-01 9.25408825e-02 -8.64872575e-01
5.90168893e-01 1.93803057e-01 4.00537074e-01 -6.89538181e-01
-6.51270449e-02 1.20083496e-01 -2.24610016e-01 -6.15141988e-01
5.61971664e-01 7.09520653e-02 3.37111771e-01 3.38244319e-01
5.37180364e-01 -4.74999726e-01 2.61928350e-01 1.93576068e-01
8.56609166e-01 6.66296184e-01 2.31816769e-01 -3.83746624e-01
6.80025458e-01 -1.86133161e-01 5.86245835e-01 7.38881350e-01
-3.66786182e-01 9.48397458e-01 1.39734074e-01 -7.65670478e-01
-9.81066048e-01 -1.17888892e+00 -1.58098593e-01 8.91327024e-01
3.83244812e-01 -2.69251257e-01 -1.82529539e-01 -5.84974051e-01
3.70214671e-01 1.18735254e-01 -8.01230311e-01 1.03495605e-01
-6.78730130e-01 -7.34381527e-02 7.57219717e-02 6.62050247e-01
6.48686409e-01 -9.71541345e-01 -8.03666592e-01 1.77989930e-01
-3.82625818e-01 -1.55576169e+00 -7.61412501e-01 -2.17178792e-01
-1.04559457e+00 -1.16443086e+00 -7.81486154e-01 -6.34421706e-01
3.77804637e-01 6.13605380e-01 1.55321491e+00 -7.74769112e-02
-2.88228333e-01 6.78440392e-01 1.04395553e-01 2.37956509e-01
1.90076023e-01 1.58737570e-01 -1.59876153e-01 -7.08609968e-02
1.40444204e-01 -8.85837317e-01 -9.89839375e-01 2.33421892e-01
-5.15072286e-01 2.62003958e-01 2.65765041e-01 1.02311945e+00
2.21119568e-01 -1.14169824e+00 5.39926551e-02 -6.17542446e-01
6.69585615e-02 -1.98595762e-01 -5.66964865e-01 -1.66009232e-01
-1.65518463e-01 3.76944005e-01 2.22147465e-01 -5.37296571e-02
-1.29469430e+00 1.45752236e-01 -2.53252517e-02 -7.88993061e-01
-2.06366956e-01 -1.01089433e-01 3.05596609e-02 -4.30719256e-01
4.71147776e-01 4.03940640e-02 1.28530622e-01 -4.73545223e-01
6.27496898e-01 7.05162669e-03 8.89725864e-01 -5.66274166e-01
7.63040781e-01 9.68596697e-01 2.10552201e-01 -6.29913807e-01
-1.08893967e+00 -6.56774104e-01 -1.10716820e+00 -3.83995324e-02
1.11797357e+00 -1.32972348e+00 -9.26465631e-01 5.52918971e-01
-1.50159204e+00 -2.87416309e-01 -2.42540330e-01 6.74721122e-01
-1.05015957e+00 3.47709656e-01 -7.19466329e-01 -3.33281517e-01
-6.48986250e-02 -1.42939460e+00 1.53172684e+00 9.58348438e-03
-4.14178342e-01 -1.61562800e+00 3.66165817e-01 6.41914725e-01
4.02363122e-01 1.70339242e-01 4.85519975e-01 -3.39794680e-02
-9.65742946e-01 3.84684741e-01 -6.04004502e-01 1.98549792e-01
1.22556224e-01 -2.31228396e-01 -1.34621704e+00 -4.07734245e-01
-2.13193446e-01 -6.22177243e-01 1.31069922e+00 6.70573294e-01
8.97285104e-01 2.31825992e-01 -2.58055031e-01 1.35990858e+00
1.25748742e+00 -3.00326556e-01 5.86765170e-01 3.55868042e-01
1.10434234e+00 7.92148709e-01 2.66253263e-01 3.39160919e-01
8.13519716e-01 7.08032250e-01 5.38516700e-01 -2.73335993e-01
-2.69598931e-01 -1.91949740e-01 2.38170832e-01 4.94867265e-01
-3.36916268e-01 1.61355421e-01 -6.20640814e-01 7.43971825e-01
-2.07044673e+00 -1.21484137e+00 1.30040318e-01 1.88467753e+00
4.65326190e-01 -3.30467932e-02 -1.23500519e-01 -1.92417115e-01
2.41397068e-01 5.80955684e-01 -5.70506811e-01 -2.72368610e-01
-9.66964737e-02 2.97808051e-01 2.14145407e-01 1.12997353e+00
-1.17620790e+00 1.01326478e+00 7.21391487e+00 2.05366373e-01
-8.55355024e-01 5.79911284e-03 4.54114884e-01 -1.97459355e-01
-5.07663488e-01 9.37229469e-02 -7.40379989e-01 -3.07008568e-02
2.01293051e-01 -1.51802778e-01 3.58072579e-01 4.47627276e-01
1.95567161e-01 -1.13609992e-01 -1.86158848e+00 1.29729879e+00
1.29081845e-01 -1.72867739e+00 2.34355465e-01 1.75757915e-01
9.35534179e-01 3.33763301e-01 -3.32615413e-02 2.92247701e-02
4.40492064e-01 -9.50748563e-01 5.76285720e-01 5.88185430e-01
6.64118111e-01 -2.46410802e-01 4.55871433e-01 -8.69141892e-02
-1.23888445e+00 -1.66617390e-02 -1.97872832e-01 -4.86141175e-01
3.87395412e-01 3.12492967e-01 9.40554664e-02 6.98043704e-01
8.15318227e-01 1.58291066e+00 -3.11114192e-01 1.03026414e+00
4.19095568e-02 -2.96150982e-01 -1.58099070e-01 7.17959702e-01
3.95314902e-01 -3.46145630e-01 5.41824520e-01 1.15155661e+00
1.38986573e-01 -2.70716786e-01 8.36827233e-02 1.02039492e+00
1.23889454e-01 -3.42163444e-01 -7.76537597e-01 7.29244053e-01
2.01902837e-01 1.02454424e+00 9.23692286e-02 -4.17345822e-01
-7.44788826e-01 1.19135332e+00 7.52470613e-01 5.50809145e-01
-4.52974409e-01 -3.32107730e-02 1.46954465e+00 -7.67990351e-02
4.79735404e-01 -2.16437325e-01 -2.21032768e-01 -1.73074353e+00
-3.38827819e-02 -3.95876676e-01 2.97418863e-01 -8.86178732e-01
-1.32638013e+00 5.19315362e-01 -2.46103406e-02 -1.18997025e+00
-6.99330926e-01 -7.69249618e-01 -5.61071992e-01 1.04977679e+00
-2.02527332e+00 -1.00129843e+00 -7.85659850e-01 9.97184157e-01
6.58703506e-01 1.33510783e-01 6.39320850e-01 4.48428869e-01
-3.57603759e-01 2.45943889e-01 -4.44835126e-01 -8.25057998e-02
9.31036055e-01 -1.16381824e+00 7.82320976e-01 7.52657712e-01
4.85842340e-02 3.75911832e-01 4.22592998e-01 -1.16472743e-01
-1.29062319e+00 -8.60273004e-01 9.91145909e-01 -8.20342600e-01
5.86830258e-01 -4.47434008e-01 -7.14631617e-01 1.01253641e+00
2.37544268e-01 5.02301276e-01 2.55677283e-01 1.34173766e-01
-6.46494091e-01 1.84150562e-02 -9.33977902e-01 4.76056755e-01
1.81917191e+00 -8.11659336e-01 -6.92796588e-01 2.59752162e-02
6.45108759e-01 -6.19152725e-01 -7.19735265e-01 2.91674048e-01
7.61614740e-01 -1.52134943e+00 1.30449438e+00 -8.11321080e-01
6.69978440e-01 -1.73351765e-01 -2.55647272e-01 -1.23535275e+00
-7.52935946e-01 -7.95560241e-01 -1.95168748e-01 7.91062057e-01
2.09198058e-01 -5.58505714e-01 6.79463387e-01 5.93032002e-01
-2.16673657e-01 -4.26600873e-01 -9.86522973e-01 -5.63210547e-01
4.90303189e-02 -4.00903910e-01 2.97894806e-01 8.70979548e-01
-1.03277192e-01 4.36524928e-01 -5.17017365e-01 -1.17995679e-01
8.56785715e-01 2.53587246e-01 1.14291191e+00 -1.21908081e+00
-4.12824601e-01 -5.10357380e-01 -7.23638654e-01 -1.95931232e+00
4.97262388e-01 -8.07801247e-01 -1.69958308e-01 -1.50739253e+00
1.47873089e-01 -4.11188900e-02 8.01159814e-02 1.68088287e-01
-2.76464820e-01 2.66395301e-01 4.39870030e-01 2.40625903e-01
-5.14451206e-01 7.59299755e-01 1.78495800e+00 -7.20744487e-03
-1.42407075e-01 -2.70625234e-01 -5.21996856e-01 6.61216378e-01
1.96628422e-01 5.41396998e-02 -4.87524092e-01 -1.10429597e+00
-4.73789498e-02 1.30306870e-01 6.73726737e-01 -9.25240874e-01
2.77596831e-01 -8.28100368e-02 3.76389951e-01 -3.21041763e-01
7.12682545e-01 -6.40294433e-01 -1.52063861e-01 3.28486264e-01
-1.79543719e-01 1.64361507e-01 2.33316660e-01 4.75720853e-01
-3.44638079e-01 3.54038924e-01 7.37951517e-01 -2.48421103e-01
-1.13552916e+00 7.42403924e-01 -1.34031698e-02 5.85653245e-01
6.68139040e-01 -5.82351923e-01 -4.38503265e-01 -5.61878145e-01
-8.73350203e-01 4.42981124e-01 4.74128425e-01 6.30968034e-01
6.63445592e-01 -1.32573569e+00 -7.84315884e-01 4.82468605e-01
1.83585957e-01 3.24778378e-01 4.22468811e-01 8.13646734e-01
-4.41472203e-01 5.96947372e-01 -6.65227473e-01 -1.08134234e+00
-8.45001698e-01 3.66290927e-01 6.67714775e-01 -1.52584255e-01
-8.03597987e-01 8.45120490e-01 8.58610094e-01 -4.37583208e-01
1.42950803e-01 -4.07012582e-01 -1.00575276e-01 -1.19354457e-01
5.48562825e-01 4.76874888e-01 -2.08209887e-01 -7.70463526e-01
-3.57940823e-01 1.28016877e+00 -1.42822778e-02 -8.17283764e-02
1.21153283e+00 -4.55284327e-01 7.03817159e-02 2.54001200e-01
1.46105444e+00 -4.59167123e-01 -2.06064606e+00 -3.83458614e-01
-3.90795231e-01 -9.98524725e-01 1.38849169e-01 -1.86161891e-01
-1.26650548e+00 1.14887154e+00 1.67262033e-01 -4.57005024e-01
8.86229753e-01 4.41202596e-02 5.83200693e-01 2.94306368e-01
2.22993940e-01 -6.09266341e-01 5.78365251e-02 8.05441141e-01
9.46430266e-01 -1.65885484e+00 -2.05773398e-01 -5.84286213e-01
-4.59813595e-01 9.63918328e-01 9.13742125e-01 -5.01523793e-01
7.60549963e-01 1.50633529e-01 -4.00924720e-02 -8.84521306e-02
-1.09244692e+00 -5.51360071e-01 5.59914708e-01 8.84014189e-01
4.70902205e-01 -5.45018077e-01 3.51407230e-01 -2.85796255e-01
-2.48235047e-01 7.41765425e-02 2.36621141e-01 7.94416368e-01
-1.39237612e-01 -9.74866927e-01 1.40568316e-01 9.11728665e-02
-7.79293105e-02 -8.82663354e-02 -4.86367233e-02 8.38414609e-01
-4.16823514e-02 7.04793930e-01 6.09669209e-01 -2.54243553e-01
4.71838653e-01 -2.28804052e-01 1.01315272e+00 -5.21677315e-01
-4.96602356e-01 -1.92156106e-01 1.51909515e-01 -1.42658484e+00
-9.28619325e-01 -5.49937546e-01 -7.57892013e-01 -6.08954549e-01
9.03568268e-02 -3.76951993e-01 5.48323914e-02 1.00451195e+00
7.05949306e-01 3.23786765e-01 5.20229697e-01 -1.63086474e+00
-1.52648181e-01 -6.73069179e-01 -2.38533095e-01 8.31502557e-01
9.27903235e-01 -1.01778877e+00 -4.51618850e-01 1.84943095e-01] | [8.644646644592285, -1.946570873260498] |
5bfa3355-0f45-47c6-9da4-62797c30533d | model-unit-exploration-for-sequence-to | 1902.01955 | null | https://arxiv.org/abs/1902.01955v2 | https://arxiv.org/pdf/1902.01955v2.pdf | On the Choice of Modeling Unit for Sequence-to-Sequence Speech Recognition | In conventional speech recognition, phoneme-based models outperform grapheme-based models for non-phonetic languages such as English. The performance gap between the two typically reduces as the amount of training data is increased. In this work, we examine the impact of the choice of modeling unit for attention-based encoder-decoder models. We conduct experiments on the LibriSpeech 100hr, 460hr, and 960hr tasks, using various target units (phoneme, grapheme, and word-piece); across all tasks, we find that grapheme or word-piece models consistently outperform phoneme-based models, even though they are evaluated without a lexicon or an external language model. We also investigate model complementarity: we find that we can improve WERs by up to 9% relative by rescoring N-best lists generated from a strong word-piece based baseline with either the phoneme or the grapheme model. Rescoring an N-best list generated by the phonemic system, however, provides limited improvements. Further analysis shows that the word-piece-based models produce more diverse N-best hypotheses, and thus lower oracle WERs, than phonemic models. | ['Antoine Bruguier', 'Anjuli Kannan', 'Rohit Prabhavalkar', 'Kazuki Irie', 'David Rybach', 'Patrick Nguyen'] | 2019-02-05 | null | null | null | null | ['sequence-to-sequence-speech-recognition'] | ['speech'] | [ 4.40897405e-01 2.02517107e-01 -1.32873967e-01 -3.16831201e-01
-1.50711989e+00 -6.47830844e-01 5.82097471e-01 1.12592448e-02
-7.44466364e-01 5.83310127e-01 6.11570477e-01 -8.11785340e-01
1.64182350e-01 -3.64668339e-01 -6.62386179e-01 -4.03900027e-01
5.26707232e-01 6.13769591e-01 2.00424954e-01 -2.82403916e-01
2.75115013e-01 1.28257051e-02 -1.36063480e+00 3.20345581e-01
8.49384487e-01 3.59774888e-01 5.79878449e-01 9.37148750e-01
-2.45842356e-02 4.16388959e-01 -7.52613783e-01 -6.11620545e-01
1.45488232e-01 -7.62266934e-01 -6.45479441e-01 -2.74967462e-01
6.65628910e-01 -8.29125270e-02 -5.00368893e-01 8.63692045e-01
6.77101791e-01 1.53576329e-01 5.93828857e-01 -5.20593762e-01
-7.81384945e-01 9.76243079e-01 -1.07227907e-01 3.81885916e-01
2.91565806e-01 1.45489007e-01 1.23709655e+00 -1.13641536e+00
4.57607180e-01 1.34940135e+00 4.41826731e-01 6.57404125e-01
-1.16414380e+00 -4.92285967e-01 1.00919388e-01 3.18906307e-01
-1.60366380e+00 -1.19940054e+00 2.39445627e-01 -1.15391545e-01
1.92135918e+00 5.35920680e-01 3.54247540e-01 9.67943192e-01
2.43791252e-01 5.23928106e-01 8.90629470e-01 -8.53217781e-01
-9.15025175e-02 3.03624123e-02 1.61340401e-01 5.26548505e-01
9.65436548e-02 2.57921010e-01 -7.66935647e-01 2.22556874e-01
5.06359220e-01 -7.52329051e-01 -5.28066456e-01 3.67614031e-01
-1.06001019e+00 6.66902065e-01 -4.73343916e-02 2.59941876e-01
-2.96662778e-01 1.52815625e-01 2.87113458e-01 3.86137903e-01
2.54339784e-01 6.91825449e-01 -5.49652398e-01 -5.86584747e-01
-1.19414425e+00 -1.58123091e-01 8.10251296e-01 1.09169674e+00
4.58136201e-01 5.01336694e-01 -4.08814698e-01 1.18664539e+00
2.68403441e-01 5.41410804e-01 8.85998130e-01 -6.94985151e-01
8.36081088e-01 -1.45408466e-01 -1.64020777e-01 -2.20199287e-01
-7.53331631e-02 -4.51817662e-01 -2.39522472e-01 -2.86803275e-01
4.72317219e-01 -1.56216189e-01 -1.42423618e+00 1.83063304e+00
-2.28586048e-01 -2.11582586e-01 2.27635249e-01 5.99938214e-01
7.68893480e-01 1.11902976e+00 1.41467988e-01 -2.76836127e-01
1.32654393e+00 -1.27804220e+00 -9.81101215e-01 -6.76979721e-01
7.83255219e-01 -9.46095407e-01 1.39980423e+00 2.92742103e-01
-1.51214433e+00 -6.58083618e-01 -1.07995582e+00 -3.07358295e-01
-3.44587952e-01 3.56949829e-02 4.57120277e-02 9.12003160e-01
-1.39044654e+00 3.14739197e-01 -6.87554359e-01 -3.15147549e-01
-3.33881259e-01 2.42614940e-01 -1.67669132e-01 -9.23049375e-02
-1.26341999e+00 1.39271629e+00 7.52099156e-01 -9.90507379e-02
-8.00627530e-01 -3.97784263e-01 -1.07539070e+00 3.40924323e-01
3.52047496e-02 -2.50821322e-01 1.59427142e+00 -6.70712233e-01
-1.65909147e+00 7.58889437e-01 -6.22846782e-01 -4.85573560e-01
3.57825472e-03 8.81157890e-02 -6.30174637e-01 -3.16959858e-01
-3.89415085e-01 8.02444756e-01 5.10692060e-01 -9.55475509e-01
-7.74368107e-01 7.83866048e-02 -2.60882914e-01 5.94123900e-01
-1.38589382e-01 2.02995017e-01 -5.93334436e-01 -6.13577366e-01
3.23242731e-02 -9.18543398e-01 2.22248137e-01 -9.40391481e-01
-4.54387695e-01 -1.84230179e-01 3.20360512e-01 -1.21098709e+00
1.64725208e+00 -2.05433702e+00 1.01462625e-01 8.93300250e-02
-4.07995582e-01 4.30342615e-01 -3.90858680e-01 4.93335992e-01
-1.36325315e-01 3.94380569e-01 -2.60840267e-01 -4.78616059e-01
1.13745905e-01 3.08706254e-01 -7.00279847e-02 1.08791448e-01
1.74146175e-01 1.10299349e+00 -5.66920698e-01 -1.77800074e-01
3.61944377e-01 4.35394943e-01 -5.63426971e-01 -5.34118265e-02
1.25537142e-02 1.96622014e-02 4.86513525e-01 4.36097592e-01
1.92971751e-01 1.41420722e-01 3.28311801e-01 1.60899296e-01
-5.33414148e-02 1.27738559e+00 -6.75183952e-01 1.48804963e+00
-8.87690604e-01 8.07394505e-01 -7.62344897e-02 -5.02385199e-01
6.79759800e-01 6.73901558e-01 -3.65652531e-01 -1.12386334e+00
-1.62542433e-01 5.76427519e-01 7.98753560e-01 2.01439355e-02
8.68733883e-01 -3.78205746e-01 2.08440721e-02 1.88191205e-01
2.14887992e-01 -3.30974758e-01 2.01496363e-01 -1.04270779e-01
1.17774868e+00 -3.46990049e-01 3.07184130e-01 -1.99528724e-01
2.60681450e-01 -1.69464916e-01 1.96803063e-01 9.41854417e-01
5.85347563e-02 7.69707084e-01 5.49684651e-02 3.34981382e-01
-1.07994938e+00 -1.14555752e+00 -3.28061610e-01 1.41124880e+00
-1.11943409e-01 -5.72286963e-01 -8.77054691e-01 -3.28201264e-01
-2.12408304e-01 1.51694918e+00 -5.65214222e-03 -3.56428504e-01
-1.00523329e+00 -6.88073218e-01 9.18647289e-01 5.36685467e-01
-1.59971103e-01 -1.25841951e+00 -6.18339032e-02 5.14898300e-01
-2.52415419e-01 -1.14373660e+00 -8.62255037e-01 4.73300785e-01
-6.08078718e-01 -3.92658770e-01 -6.49897099e-01 -1.01339293e+00
2.44610339e-01 -6.18033484e-02 1.24066329e+00 -2.22200211e-02
2.92576998e-01 1.55995339e-01 -4.57894355e-01 -2.66412675e-01
-7.73510098e-01 4.60888654e-01 -1.18064739e-01 -4.97347921e-01
4.11638707e-01 -1.06673829e-01 -8.08816999e-02 8.09971690e-02
-5.80157399e-01 1.15785405e-01 7.08162427e-01 9.18583572e-01
6.11086369e-01 -2.70232230e-01 8.91232908e-01 -7.70013511e-01
6.83819771e-01 -3.00007373e-01 -4.50294554e-01 2.72533119e-01
-7.88809299e-01 1.98393166e-01 6.98599577e-01 -3.50331455e-01
-1.04803801e+00 -1.96005002e-01 -7.64319003e-01 -1.35796918e-02
-3.38245481e-02 7.30565548e-01 -4.70658243e-01 2.69438177e-01
5.29405117e-01 4.92312521e-01 -3.80127341e-01 -6.34472847e-01
5.31079590e-01 9.22443867e-01 6.85340226e-01 -2.54139900e-01
5.33782661e-01 -4.17043835e-01 -8.03236961e-01 -9.96748805e-01
-5.26410580e-01 -4.83046025e-01 -8.34026784e-02 1.05104588e-01
8.92039597e-01 -8.83661091e-01 -1.20589115e-01 3.44343096e-01
-1.44245338e+00 -5.62766314e-01 -3.70923162e-01 6.68851614e-01
-3.81925851e-01 1.61533102e-01 -9.61448848e-01 -7.02659965e-01
-2.53264725e-01 -1.36366284e+00 1.02383840e+00 -2.64003687e-02
-5.13615370e-01 -9.85295832e-01 -1.64609980e-02 2.87209541e-01
4.71667290e-01 -7.33088136e-01 1.24583852e+00 -1.07800114e+00
-4.55248952e-01 -1.24231078e-01 3.26698273e-02 4.50742841e-01
2.78385639e-01 -2.82942563e-01 -1.00867963e+00 -3.28472912e-01
-5.25282845e-02 -1.66134253e-01 8.72655511e-01 4.08187836e-01
9.28508639e-01 -4.33097541e-01 -4.69591953e-02 6.00838244e-01
1.12478900e+00 6.20528042e-01 7.96424389e-01 8.27543810e-02
7.64140487e-01 4.45274442e-01 2.05851063e-01 -1.72097012e-01
5.18837929e-01 8.69242847e-01 -2.98715848e-02 1.33124534e-02
-6.22187495e-01 -4.57845449e-01 7.93408930e-01 1.78304434e+00
3.05833280e-01 -1.09243464e+00 -1.06464958e+00 6.93014085e-01
-1.25200188e+00 -6.61346078e-01 -1.48615301e-01 2.44704056e+00
1.15960407e+00 3.60675037e-01 -1.91123247e-01 9.30670798e-02
8.81848574e-01 2.57900447e-01 -2.35513985e-01 -9.22625184e-01
-2.63419509e-01 6.64714873e-01 6.92748249e-01 1.13993728e+00
-5.21690071e-01 1.47537315e+00 7.18915224e+00 1.17288637e+00
-1.01279080e+00 3.30371499e-01 3.46971393e-01 -1.92759633e-01
-5.86096883e-01 -2.41938327e-02 -1.15999568e+00 4.70802665e-01
1.86169410e+00 -2.73972124e-01 8.30341041e-01 5.17352104e-01
9.97969657e-02 5.04229702e-02 -1.45230365e+00 9.81114268e-01
2.58677214e-01 -1.05354869e+00 2.15709835e-01 7.05929697e-02
6.29229367e-01 2.99867868e-01 -3.95218208e-02 5.53352475e-01
3.05088073e-01 -1.32158005e+00 1.03360927e+00 2.38714427e-01
1.08531749e+00 -7.74210155e-01 5.70377231e-01 2.74414361e-01
-1.10576522e+00 3.81394029e-01 -3.08621734e-01 6.79452494e-02
5.19039512e-01 1.82819605e-01 -9.55093324e-01 3.84571731e-01
1.16528980e-01 1.56982616e-01 -4.73773897e-01 1.12019813e+00
-3.54397088e-01 1.23710740e+00 -2.39076123e-01 4.86334227e-02
2.59011537e-01 7.41834417e-02 6.30470335e-01 1.70094764e+00
4.55146700e-01 -6.23266734e-02 -1.76262096e-01 4.27101880e-01
-2.23025590e-01 1.67387366e-01 -3.70058328e-01 -2.51605392e-01
8.89894366e-01 7.54505098e-01 -4.71447051e-01 -4.36215222e-01
-5.01801431e-01 1.12509704e+00 4.35141891e-01 4.17816430e-01
-6.00030363e-01 -7.11147785e-01 4.62293446e-01 4.61878022e-03
3.25827509e-01 -3.85754764e-01 -2.52595693e-01 -9.01592255e-01
-2.31355038e-02 -1.26945961e+00 5.68708517e-02 -7.46903360e-01
-1.02622604e+00 8.70897174e-01 -1.16382137e-01 -5.17281950e-01
-9.12371159e-01 -6.31058335e-01 -4.34628606e-01 1.36103928e+00
-1.45021391e+00 -5.25228322e-01 2.95555443e-01 5.05221672e-02
9.53706026e-01 -1.41584545e-01 9.22896922e-01 3.67987096e-01
-4.62333620e-01 1.04628575e+00 2.40146413e-01 2.10783616e-01
6.05725408e-01 -1.32040346e+00 1.05730987e+00 1.09374046e+00
6.38464332e-01 6.50400698e-01 6.03560090e-01 -6.42810106e-01
-1.14499557e+00 -9.79175746e-01 1.40703475e+00 -7.19914496e-01
3.88321251e-01 -5.92831790e-01 -1.16277492e+00 9.02548611e-01
6.18738294e-01 -5.45457721e-01 5.78000605e-01 2.53738701e-01
-1.87915519e-01 1.75401077e-01 -6.59415007e-01 7.75106907e-01
1.12400079e+00 -9.64186907e-01 -9.22389627e-01 1.46163389e-01
1.09903014e+00 -5.24507523e-01 -6.21494710e-01 2.15912357e-01
4.70511347e-01 -6.49502277e-01 5.97663939e-01 -5.13015091e-01
7.97814727e-02 -1.49730518e-02 -3.98201138e-01 -1.86904132e+00
-3.44875634e-01 -4.09019440e-01 -1.52098775e-01 1.18705404e+00
1.03270996e+00 -6.51292980e-01 4.17641699e-01 2.61588842e-01
-8.16404700e-01 -5.05932808e-01 -1.16433501e+00 -1.14722157e+00
2.97020435e-01 -5.93825459e-01 5.14399886e-01 4.65099573e-01
3.18276472e-02 6.47153974e-01 -1.11386403e-01 2.80078333e-02
2.15069260e-02 -6.41433537e-01 2.26650834e-01 -5.93649328e-01
-5.71479201e-01 -6.55603766e-01 -1.40462682e-01 -1.36875176e+00
2.86187410e-01 -1.12228549e+00 6.63201153e-01 -1.65573883e+00
-1.53044567e-01 -3.18609953e-01 -3.40991288e-01 6.11433864e-01
-2.97758728e-01 1.12419583e-01 3.36769909e-01 -7.70174414e-02
-2.83036023e-01 5.71522295e-01 8.04629683e-01 9.13823098e-02
-3.25322747e-01 -3.31430346e-01 -6.01433516e-01 5.65320432e-01
7.15871751e-01 -4.25706804e-01 -4.35612112e-01 -8.80770743e-01
-1.25404745e-01 2.24126875e-01 -9.59131718e-02 -9.65110719e-01
2.60119468e-01 1.96736634e-01 1.15219034e-01 -4.35728520e-01
6.15907490e-01 -2.90777802e-01 -4.53672856e-02 2.97486275e-01
-5.03604233e-01 4.90024090e-01 5.55116057e-01 1.35830954e-01
-2.90003270e-01 -5.86566091e-01 7.96536982e-01 7.18045142e-03
-5.50934732e-01 -1.72872573e-01 -8.31831872e-01 4.24581736e-01
3.08397591e-01 -4.82496560e-01 -3.77094656e-01 -6.23826504e-01
-4.71134514e-01 -1.71769157e-01 3.28399599e-01 5.16363442e-01
4.53559667e-01 -1.24512696e+00 -9.08048987e-01 3.22656661e-01
9.14008841e-02 -3.91956210e-01 -2.02967554e-01 5.12731731e-01
-3.90012532e-01 8.39617074e-01 2.59697258e-01 -2.57151365e-01
-1.06306601e+00 1.27278537e-01 5.05069375e-01 -1.26848429e-01
-1.48533791e-01 1.07105577e+00 1.30243674e-01 -5.67180276e-01
4.61493164e-01 -4.30673122e-01 2.00232387e-01 -4.82332781e-02
3.39274317e-01 4.92126018e-01 6.10936224e-01 -9.00124669e-01
-4.93653566e-01 1.84493884e-01 -2.44903475e-01 -8.04955244e-01
7.52072334e-01 -2.45303288e-01 3.88039500e-01 6.35085821e-01
1.09977341e+00 3.75173360e-01 -8.29251826e-01 4.92047518e-02
1.68622315e-01 -1.19379543e-01 2.20338956e-01 -1.04277682e+00
-6.74849510e-01 8.60530019e-01 3.13945830e-01 1.65004041e-02
8.88488591e-01 -4.79321517e-02 9.38164234e-01 3.04516196e-01
2.61604667e-01 -1.15658677e+00 -3.87942791e-01 1.08835101e+00
8.45912635e-01 -8.84302795e-01 -5.84377527e-01 -2.29516461e-01
-6.16783738e-01 6.38972700e-01 6.68653071e-01 2.53865927e-01
2.83107102e-01 1.94498986e-01 9.03623402e-02 1.58257037e-01
-9.47161317e-01 -4.40122157e-01 5.74650466e-01 4.64480489e-01
8.92291188e-01 1.28459990e-01 -3.65008116e-01 5.84719598e-01
-9.20864820e-01 -6.25558496e-01 4.46348906e-01 6.11227751e-01
-6.28023148e-01 -1.28710783e+00 -2.91239947e-01 6.39294982e-01
-4.52927262e-01 -1.05178189e+00 -3.68434429e-01 8.33475053e-01
-1.25328600e-01 1.21061981e+00 3.76912266e-01 -4.40472782e-01
4.16894197e-01 5.71695745e-01 5.16342700e-01 -1.13367116e+00
-6.49317503e-01 2.61077613e-01 5.72876692e-01 -2.52012998e-01
2.36631483e-01 -6.16361439e-01 -1.25380528e+00 -1.75592080e-01
-7.71959901e-01 2.76364297e-01 6.49139643e-01 9.80337739e-01
8.41364041e-02 5.62428415e-01 1.56176224e-01 -5.63528776e-01
-4.84489590e-01 -1.28273523e+00 -3.20958823e-01 -1.36053875e-01
2.12668777e-01 -1.54156655e-01 -5.05150259e-01 -5.37460074e-02] | [14.335619926452637, 6.832771301269531] |
e9e5c0d7-3707-48ac-a330-ab8c6d062293 | learnable-online-graph-representations-for-3d | 2104.11747 | null | https://arxiv.org/abs/2104.11747v1 | https://arxiv.org/pdf/2104.11747v1.pdf | Learnable Online Graph Representations for 3D Multi-Object Tracking | Tracking of objects in 3D is a fundamental task in computer vision that finds use in a wide range of applications such as autonomous driving, robotics or augmented reality. Most recent approaches for 3D multi object tracking (MOT) from LIDAR use object dynamics together with a set of handcrafted features to match detections of objects. However, manually designing such features and heuristics is cumbersome and often leads to suboptimal performance. In this work, we instead strive towards a unified and learning based approach to the 3D MOT problem. We design a graph structure to jointly process detection and track states in an online manner. To this end, we employ a Neural Message Passing network for data association that is fully trainable. Our approach provides a natural way for track initialization and handling of false positive detections, while significantly improving track stability. We show the merit of the proposed approach on the publicly available nuScenes dataset by achieving state-of-the-art performance of 65.6% AMOTA and 58% fewer ID-switches. | ['Luc van Gool', 'Martin Danelljan', 'Alexander Liniger', 'Dengxin Dai', 'Jan-Nico Zaech'] | 2021-04-23 | null | null | null | null | ['3d-multi-object-tracking'] | ['computer-vision'] | [ 1.08431496e-01 -4.20766085e-01 -1.52862072e-01 -1.94657862e-01
-5.05667984e-01 -7.01350868e-01 6.92488670e-01 3.66925657e-01
-5.42486966e-01 4.41964447e-01 -5.81673682e-01 -4.64417845e-01
-2.13780239e-01 -7.96112061e-01 -8.53394508e-01 -6.31814718e-01
-1.53458327e-01 7.97890782e-01 8.52738678e-01 -3.82956751e-02
1.33936837e-01 9.34470475e-01 -1.88650179e+00 -3.65531594e-01
5.37829936e-01 9.70341265e-01 3.95460159e-01 7.57339239e-01
-1.39630750e-01 4.83114332e-01 -9.97780487e-02 -2.21632764e-01
4.61867929e-01 2.01424584e-02 -3.29530567e-01 2.32919246e-01
7.81730294e-01 3.18535394e-03 -3.09330583e-01 1.10309243e+00
5.35027266e-01 1.22412287e-01 5.54804027e-01 -1.47288537e+00
-1.17355824e-01 2.30312496e-01 -7.75498688e-01 3.85284126e-01
4.69812714e-02 3.49017531e-01 9.11801517e-01 -6.32010579e-01
4.69467968e-01 1.17244136e+00 7.68163860e-01 4.66689765e-01
-1.27838647e+00 -8.00544858e-01 3.35000336e-01 2.32755467e-01
-1.44405878e+00 -4.18777764e-01 5.22039413e-01 -6.23923898e-01
8.41115892e-01 1.40069902e-01 6.37926579e-01 5.95863760e-01
1.80252090e-01 5.74413836e-01 9.98616993e-01 -2.66371429e-01
1.48181304e-01 1.36283964e-01 3.30398858e-01 9.19133067e-01
7.23341584e-01 2.40231559e-01 -5.08657396e-01 -2.63833888e-02
5.70369482e-01 1.82728365e-01 2.33122975e-01 -9.66223359e-01
-1.07687306e+00 7.44707406e-01 4.87433761e-01 -4.85422015e-02
-3.33388686e-01 3.93843830e-01 1.97753355e-01 1.24650732e-01
2.09087983e-01 -1.01013809e-01 -1.55961588e-01 1.45322591e-01
-6.42915905e-01 4.61994350e-01 5.54622352e-01 1.25765765e+00
7.83359826e-01 -6.71763793e-02 -2.27159947e-01 4.74181950e-01
4.12168205e-01 8.49613070e-01 -2.42053926e-01 -6.32058859e-01
2.06091687e-01 6.98030412e-01 3.82908136e-01 -9.26237226e-01
-6.00690305e-01 -3.92918229e-01 -6.62610710e-01 5.41663766e-01
3.36950839e-01 3.85476500e-02 -1.10018981e+00 1.70855093e+00
8.96354139e-01 4.25143600e-01 -1.51503995e-01 8.58924925e-01
4.72396731e-01 3.93352330e-01 -1.06965512e-01 -6.27626926e-02
1.40032661e+00 -7.22798944e-01 -4.96484429e-01 -3.24548006e-01
4.76848423e-01 -7.13797629e-01 4.94766921e-01 2.25618154e-01
-9.81711149e-01 -6.33249938e-01 -9.82726038e-01 1.71632543e-01
-2.99588174e-01 5.56848161e-02 6.43146873e-01 7.49433875e-01
-9.47405338e-01 3.23884308e-01 -1.25788391e+00 -6.98433399e-01
4.57836211e-01 7.61846304e-01 -1.35537311e-01 -7.92812742e-03
-4.86881196e-01 1.06763935e+00 5.93428433e-01 2.00245216e-01
-8.21256697e-01 -3.66225004e-01 -7.61821389e-01 -2.19653800e-01
7.51237214e-01 -7.84609497e-01 1.16378510e+00 -3.26158702e-02
-1.30449998e+00 7.59364784e-01 -2.46381357e-01 -7.00630367e-01
4.58396435e-01 -3.80176604e-01 -3.23183000e-01 -2.30718017e-01
8.46419204e-03 6.71580017e-01 8.76378596e-01 -1.12328815e+00
-1.10981238e+00 -4.14324731e-01 -7.61409551e-02 1.82470270e-02
-2.29818672e-01 -7.90876895e-02 -6.97256207e-01 -1.60137787e-01
2.24567875e-01 -1.34517324e+00 -3.95557284e-01 3.83666486e-01
-2.66840160e-01 -3.83262604e-01 9.97246802e-01 1.42407075e-01
8.45050335e-01 -1.73179579e+00 7.06516355e-02 1.19765408e-01
2.73515642e-01 4.38178599e-01 1.14616856e-01 2.62630492e-01
5.90819120e-01 -4.19718832e-01 5.75416721e-02 -5.15512943e-01
6.25948012e-02 2.71023631e-01 -4.73795980e-02 6.35325909e-01
4.28979635e-01 7.01783419e-01 -9.72381473e-01 -5.15491188e-01
5.91902196e-01 3.97409260e-01 -6.14709854e-01 7.08105639e-02
-4.39873010e-01 5.08265018e-01 -5.77868044e-01 5.99919319e-01
7.52964020e-01 -3.68706256e-01 9.88877416e-02 -1.14015386e-01
-5.19779801e-01 2.91882068e-01 -1.68454361e+00 1.57223153e+00
-2.33961478e-01 3.20067763e-01 2.12149918e-02 -8.06950212e-01
9.17809725e-01 -6.01790510e-02 6.81607604e-01 -2.61319935e-01
2.85859525e-01 7.57766664e-02 5.70941996e-03 -2.59087950e-01
7.24290788e-01 1.09753527e-01 8.16832203e-03 3.03704798e-01
-4.51540239e-02 -5.71045801e-02 2.93909550e-01 -2.54575592e-02
1.31019425e+00 1.34337649e-01 2.73005337e-01 -9.47597176e-02
5.92453897e-01 4.08088475e-01 6.11168623e-01 9.91143286e-01
-2.08359465e-01 8.52343738e-02 -1.99475795e-01 -3.82923514e-01
-8.47283125e-01 -1.06440496e+00 -3.50587629e-02 6.19407594e-01
4.87045139e-01 -2.40864635e-01 -5.13385460e-02 -5.60151219e-01
3.39893699e-01 2.84529448e-01 -2.82332808e-01 -1.62322223e-01
-6.56471312e-01 -5.11124671e-01 3.31889540e-01 5.45123518e-01
2.02865213e-01 -7.69717574e-01 -9.57848489e-01 4.77545232e-01
2.13996708e-01 -1.38687336e+00 -3.35393488e-01 4.18746322e-01
-9.56076682e-01 -9.66144145e-01 -3.30704868e-01 -6.80865884e-01
5.81352174e-01 7.78242230e-01 8.73615563e-01 6.77130073e-02
-3.96603137e-01 2.05496714e-01 -2.00419307e-01 -6.20185435e-01
-3.54107767e-01 3.29895079e-01 4.17144775e-01 -3.28401178e-02
4.53102589e-01 -4.35528219e-01 -3.94121230e-01 4.86775428e-01
-5.71920335e-01 -6.17112741e-02 7.51196086e-01 6.00960672e-01
7.36856878e-01 -2.44776830e-02 2.33951852e-01 -6.69716775e-01
9.94395651e-03 -3.30105364e-01 -1.38660431e+00 9.75741260e-03
-6.03465080e-01 2.28830203e-01 2.21243784e-01 -5.51011324e-01
-5.54085970e-01 7.73224175e-01 -1.89058743e-02 -7.58368313e-01
-2.08239600e-01 -1.07033681e-02 1.15405172e-01 -5.50894797e-01
5.04238725e-01 1.19785510e-01 -5.14505096e-02 -5.18814802e-01
4.97914970e-01 4.22656357e-01 6.66034043e-01 -5.40953100e-01
1.23794985e+00 6.75186515e-01 4.57379460e-01 -6.34607315e-01
-8.37724090e-01 -8.94141436e-01 -6.69951022e-01 -3.83321911e-01
7.50074446e-01 -9.68316257e-01 -1.20833600e+00 2.49292597e-01
-1.11109316e+00 5.08078095e-03 -1.51088089e-02 5.18012583e-01
-5.08014321e-01 8.92507136e-02 -1.80155590e-01 -1.10321248e+00
-2.12255105e-01 -1.03400171e+00 1.18619323e+00 3.63172591e-01
1.49491310e-01 -5.78481376e-01 2.06961706e-01 1.38114337e-02
2.25513443e-01 2.92273432e-01 4.51238930e-01 -4.09262240e-01
-1.22402489e+00 -2.60203570e-01 -3.43253404e-01 -2.38870084e-01
2.15207636e-01 -1.29387513e-01 -8.17035973e-01 -5.21512032e-01
-3.69442701e-01 -1.85496122e-01 8.19634497e-01 3.69557470e-01
6.55321419e-01 1.44484952e-01 -8.63038301e-01 4.36582476e-01
1.48612261e+00 6.84304461e-02 -5.33614820e-03 3.55843008e-01
7.24716246e-01 2.78659850e-01 8.93661857e-01 4.96861249e-01
5.33697248e-01 9.42049205e-01 6.77485764e-01 1.50079593e-01
-1.07296892e-01 -1.23011731e-01 1.13866687e-01 5.02381384e-01
6.71188459e-02 -3.00108999e-01 -9.66749609e-01 6.67892098e-01
-2.15563416e+00 -8.93577218e-01 -3.64309311e-01 2.25291491e+00
4.86609071e-01 6.07640862e-01 2.87603647e-01 -2.34416332e-02
8.91594768e-01 -6.02041483e-02 -6.05400264e-01 2.20236614e-01
2.47558191e-01 7.82516375e-02 9.20371652e-01 4.29174602e-01
-1.27056837e+00 1.09849536e+00 5.26509905e+00 3.48085940e-01
-1.06185174e+00 -6.31175772e-06 -4.32927340e-01 -8.51688534e-02
3.21605861e-01 2.43857037e-02 -1.59141123e+00 2.20359623e-01
7.90612459e-01 1.11593604e-01 1.91398144e-01 6.79493904e-01
9.21518430e-02 -6.75870106e-02 -1.07124996e+00 9.76993859e-01
-1.91753730e-01 -1.43139958e+00 -1.44428700e-01 1.16808869e-01
5.77941835e-01 2.83220917e-01 -1.21156834e-01 1.48739234e-01
6.25098765e-01 -4.30074066e-01 8.85795653e-01 3.10933292e-01
4.20780152e-01 -6.18571281e-01 4.05161947e-01 4.74168837e-01
-1.63889945e+00 -1.69849750e-02 -3.35455596e-01 -6.92516863e-02
4.34037387e-01 4.23910230e-01 -1.12437081e+00 6.95309401e-01
7.82849312e-01 6.41979039e-01 -4.21944886e-01 1.59545159e+00
1.32396996e-01 3.45670700e-01 -8.10335755e-01 -2.83584803e-01
2.09004313e-01 1.43654138e-01 8.87434542e-01 1.03686118e+00
2.84087002e-01 -8.79966095e-02 6.72584355e-01 5.50190151e-01
9.80181396e-02 -3.14060181e-01 -8.84929419e-01 3.38960625e-02
7.13449836e-01 1.36539030e+00 -9.76482451e-01 -2.21677244e-01
-4.83405888e-01 5.33814430e-01 4.25756067e-01 -2.47271627e-01
-9.35628712e-01 -7.02931210e-02 8.08494449e-01 2.80085802e-02
8.93718779e-01 -6.56846225e-01 -6.62963390e-02 -9.08604324e-01
3.55039276e-02 -4.53459144e-01 3.92943203e-01 -3.21041435e-01
-1.10269916e+00 4.25040126e-01 -8.80272165e-02 -1.46764004e+00
-1.54822305e-01 -5.38615525e-01 -2.67080039e-01 5.85609019e-01
-1.81928062e+00 -1.26275647e+00 -5.63506782e-01 5.59359670e-01
3.65471721e-01 1.48743048e-01 4.40136164e-01 7.09355772e-01
-5.08799732e-01 3.16483796e-01 -1.90690026e-01 -1.10009640e-01
6.83404446e-01 -1.11024570e+00 7.56523073e-01 1.01409578e+00
4.63525146e-01 3.78407121e-01 9.99417424e-01 -7.64082730e-01
-1.81825209e+00 -1.40758717e+00 6.80563390e-01 -6.00210547e-01
7.36877561e-01 -5.32555103e-01 -6.24345839e-01 7.67804265e-01
-1.30539015e-01 2.31539533e-01 1.04281805e-01 -1.43255806e-02
1.97854880e-02 -2.20751777e-01 -9.09969926e-01 5.31638682e-01
1.39043629e+00 -9.70070586e-02 -4.06861007e-01 3.23408842e-01
5.78495920e-01 -7.36714959e-01 -6.15317404e-01 4.29039359e-01
4.69713300e-01 -5.50491273e-01 1.08847082e+00 -5.03663301e-01
-5.82475364e-01 -1.01462376e+00 -4.30701301e-02 -8.09768915e-01
-3.85268956e-01 -7.81852305e-01 -4.74159449e-01 1.13927352e+00
9.87713560e-02 -4.99421060e-01 1.02729142e+00 4.58996780e-02
-3.59456718e-01 -3.03110182e-01 -9.23846781e-01 -1.07513607e+00
-5.25972068e-01 -4.84827608e-01 3.25544566e-01 5.73157430e-01
-7.83750653e-01 3.84729147e-01 -4.45741743e-01 9.01707888e-01
1.05391228e+00 3.57282341e-01 1.27527893e+00 -1.70304477e+00
-1.07657284e-01 -1.95594564e-01 -6.49277389e-01 -1.22887266e+00
-6.23575337e-02 -9.65709329e-01 4.08701003e-01 -1.30285525e+00
3.08890175e-02 -9.14623320e-01 -3.12435597e-01 5.01867056e-01
-8.15129504e-02 2.70398706e-01 3.25846583e-01 1.53385654e-01
-8.77688050e-01 3.92553777e-01 8.68102193e-01 -1.98914573e-01
-3.17977130e-01 2.64925480e-01 -3.35758567e-01 5.47572255e-01
6.23270750e-01 -9.27105010e-01 -1.14367068e-01 -5.92448294e-01
-1.10117430e-02 -7.75870457e-02 6.47898197e-01 -1.30234241e+00
7.16463387e-01 -1.60143122e-01 1.45946711e-01 -1.29166460e+00
6.68368816e-01 -1.17644429e+00 2.53805339e-01 6.85105026e-01
1.80729002e-01 3.41716379e-01 4.43637192e-01 8.56127739e-01
1.70930579e-01 -1.91836972e-02 8.01665306e-01 6.81083873e-02
-9.23487961e-01 5.19381046e-01 -7.92576894e-02 -2.31080249e-01
1.21941805e+00 -1.86572656e-01 -1.89434662e-01 1.43933922e-01
-3.41677010e-01 6.67378604e-01 5.14855087e-01 5.66447258e-01
3.53649110e-01 -1.27055216e+00 -4.54983145e-01 1.65033534e-01
2.93311030e-01 1.25223979e-01 -4.55991626e-02 9.94311929e-01
-1.77615106e-01 5.20659447e-01 -2.60815531e-01 -1.37018132e+00
-1.32727969e+00 6.58850908e-01 7.79052228e-02 -1.68627709e-01
-9.36640799e-01 5.81217229e-01 -3.88249487e-01 -3.51886570e-01
3.71189147e-01 -4.53023016e-01 2.64232513e-02 -9.12170112e-02
3.21423620e-01 3.17977250e-01 1.28112629e-01 -5.34135997e-01
-6.25766456e-01 8.64458323e-01 -2.68988967e-01 1.25048934e-02
1.21640992e+00 -1.18374340e-01 3.52622032e-01 3.99946451e-01
4.86931294e-01 -1.45660684e-01 -1.38380563e+00 -4.57451224e-01
3.97769809e-01 -4.03873771e-01 1.59791201e-01 -3.36863935e-01
-7.40350544e-01 6.09605074e-01 9.23200250e-01 2.46369436e-01
6.72215581e-01 1.38937131e-01 7.21392632e-01 8.53286207e-01
7.39731193e-01 -6.00471795e-01 -1.95863232e-01 5.76272249e-01
1.12789869e-01 -1.38977814e+00 -5.08555397e-02 -3.62632871e-01
-1.28035352e-01 8.54194283e-01 6.43199980e-01 -1.95820346e-01
4.70787823e-01 4.70938891e-01 -3.30691691e-03 -2.92794317e-01
-7.53090680e-01 -8.50569427e-01 2.20597312e-01 5.17191470e-01
-8.55378807e-02 -2.75633365e-01 -1.29762748e-02 -9.35748816e-02
1.85035571e-01 8.10157880e-02 1.58283547e-01 1.40177262e+00
-7.49001622e-01 -1.16509235e+00 -5.03044963e-01 3.83159995e-01
-1.49045601e-01 2.80583441e-01 -5.45582213e-02 8.83165598e-01
1.06559597e-01 8.70140612e-01 7.98828006e-02 -2.99649119e-01
5.59413970e-01 5.54768462e-03 7.14404166e-01 -7.05330789e-01
-4.26540583e-01 -2.64367973e-03 1.10070467e-01 -4.26348478e-01
-7.31743872e-01 -8.77877593e-01 -1.24493515e+00 -2.35889718e-01
-7.08005667e-01 -2.12743238e-01 7.58988798e-01 9.02881622e-01
4.92653877e-01 5.54468155e-01 4.38803554e-01 -9.79225636e-01
-6.07693017e-01 -5.18265009e-01 -1.40267730e-01 7.85600021e-02
6.53297186e-01 -1.27630937e+00 1.27810955e-01 -5.51847741e-02] | [6.614516735076904, -2.2010135650634766] |
a9ca7751-c0c1-4639-9c88-e881cdce8081 | geometric-feature-learning-for-3d-meshes | 2112.01801 | null | https://arxiv.org/abs/2112.01801v3 | https://arxiv.org/pdf/2112.01801v3.pdf | Mesh Convolution with Continuous Filters for 3D Surface Parsing | Geometric feature learning for 3D surfaces is critical for many applications in computer graphics and 3D vision. However, deep learning currently lags in hierarchical modeling of 3D surfaces due to the lack of required operations and/or their efficient implementations. In this paper, we propose a series of modular operations for effective geometric feature learning from 3D triangle meshes. These operations include novel mesh convolutions, efficient mesh decimation and associated mesh (un)poolings. Our mesh convolutions exploit spherical harmonics as orthonormal bases to create continuous convolutional filters. The mesh decimation module is GPU-accelerated and able to process batched meshes on-the-fly, while the (un)pooling operations compute features for up/down-sampled meshes. We provide open-source implementation of these operations, collectively termed Picasso. Picasso supports heterogeneous mesh batching and processing. Leveraging its modular operations, we further contribute a novel hierarchical neural network for perceptual parsing of 3D surfaces, named PicassoNet++. It achieves highly competitive performance for shape analysis and scene segmentation on prominent 3D benchmarks. The code, data and trained models are available at https://github.com/EnyaHermite/Picasso. | ['Ajmal Mian', 'Mubarak Shah', 'Naveed Akhtar', 'Huan Lei'] | 2021-12-03 | null | null | null | null | ['scene-parsing', 'scene-segmentation'] | ['computer-vision', 'computer-vision'] | [ 1.04392886e-01 -1.04643568e-01 2.77141392e-01 -3.55489850e-01
-6.23242378e-01 -4.41153079e-01 4.59622771e-01 4.74382311e-01
-1.78789109e-01 -1.12730920e-01 -2.54501492e-01 -4.11251247e-01
2.35939816e-01 -1.27930975e+00 -1.12561452e+00 -3.37031245e-01
-3.61024827e-01 5.39700031e-01 3.63506854e-01 -5.65764382e-02
4.56474364e-01 1.22490132e+00 -1.88496494e+00 4.82677937e-01
5.53786278e-01 1.40112996e+00 1.46530345e-01 1.03012466e+00
-6.18142724e-01 1.59053989e-02 -1.34474295e-03 -2.13140368e-01
3.94163966e-01 3.52649659e-01 -7.43260443e-01 1.19196065e-01
8.76720130e-01 -3.82744044e-01 -7.07217902e-02 8.34821343e-01
6.27604604e-01 -3.55555192e-02 7.76396751e-01 -9.70876038e-01
-4.72161412e-01 1.13077506e-01 -4.77884293e-01 -1.48370862e-01
8.40037540e-02 -7.55835846e-02 9.51203465e-01 -1.48401678e+00
6.80046737e-01 1.55182695e+00 1.12891006e+00 2.92480946e-01
-1.12848485e+00 -3.93486112e-01 -2.09978390e-02 -3.31006795e-01
-1.28258395e+00 -2.03481287e-01 8.19175005e-01 -5.24764359e-01
1.28180039e+00 3.57279718e-01 8.36196959e-01 6.13050938e-01
2.88129419e-01 6.77434802e-01 6.51540756e-01 -1.96684137e-01
2.84956932e-01 -4.70987260e-01 3.31780277e-02 1.20041358e+00
1.81263357e-01 -1.86507419e-01 -1.66586831e-01 -2.99666941e-01
1.55036485e+00 1.19704075e-01 1.24789350e-01 -5.02095461e-01
-8.98127854e-01 7.74130821e-01 5.87576032e-01 -2.16791362e-01
-4.84534860e-01 5.03907084e-01 4.35316056e-01 2.79539496e-01
9.27901626e-01 2.33337671e-01 -6.99782550e-01 6.89712465e-02
-6.61795378e-01 3.89776289e-01 8.55412424e-01 1.14213169e+00
1.08612025e+00 -1.46697372e-01 1.59020081e-01 9.07715321e-01
1.57705322e-01 4.31831658e-01 -3.94765258e-01 -1.18885577e+00
7.31228292e-02 8.69186938e-01 -1.78233385e-01 -1.00983155e+00
-6.90828621e-01 -1.59523472e-01 -7.32122004e-01 5.72139263e-01
1.95051298e-01 1.56080216e-01 -1.20325983e+00 1.01255155e+00
7.98013210e-01 2.97774792e-01 -5.56648970e-01 6.76908731e-01
1.24874854e+00 6.05329394e-01 -1.03284419e-01 5.68741083e-01
1.47342312e+00 -8.13023508e-01 1.19132884e-01 1.66169971e-01
5.03106594e-01 -1.06092727e+00 1.00575042e+00 1.40057728e-01
-1.41344488e+00 -5.93515277e-01 -8.70737314e-01 -6.95073068e-01
-4.63780969e-01 1.00240875e-02 8.69768977e-01 1.55662417e-01
-1.24448037e+00 9.60553288e-01 -1.21314847e+00 -2.78211534e-01
1.08896184e+00 6.06121659e-01 -1.93154529e-01 1.10947005e-01
-5.03970206e-01 3.30252767e-01 -2.66997870e-02 -4.35082056e-02
-6.31685376e-01 -9.70138848e-01 -1.24050701e+00 7.92475883e-03
8.95856991e-02 -1.08581161e+00 1.25974441e+00 -4.59792197e-01
-1.53940046e+00 1.31883204e+00 -1.32958755e-01 -1.33751750e-01
2.60161042e-01 -2.79597253e-01 3.04356873e-01 1.95804194e-01
-5.20478134e-05 8.48868966e-01 9.29470897e-01 -1.22031856e+00
-5.26218534e-01 -2.96848089e-01 -3.09164133e-02 3.48297507e-03
2.89145727e-02 -2.83652276e-01 -6.41486585e-01 -5.48634529e-01
4.82249916e-01 -6.03223860e-01 -3.69023383e-01 4.85749006e-01
-5.33280313e-01 -4.33964461e-01 7.15610325e-01 -3.52731586e-01
7.11803913e-01 -2.19632721e+00 -4.02431041e-02 3.58696252e-01
5.62688172e-01 9.17138625e-03 -1.78436637e-01 3.89861345e-01
1.73272952e-01 1.54353887e-01 -3.78795087e-01 -8.28868747e-01
1.56039149e-01 9.21566784e-02 -7.27775544e-02 3.53588790e-01
7.41399705e-01 1.01425493e+00 -7.00046241e-01 -4.05245930e-01
6.70128167e-01 8.91672015e-01 -9.35960472e-01 5.78708313e-02
-5.26835144e-01 3.67501825e-01 -4.72232878e-01 9.58606184e-01
1.09247220e+00 -4.86931413e-01 -3.54582101e-01 -3.82291019e-01
-4.58287716e-01 3.54277492e-01 -1.11788845e+00 1.85187244e+00
-6.38230026e-01 4.05153126e-01 3.28618884e-01 -7.04397440e-01
9.62947309e-01 -3.80119495e-02 5.32670736e-01 -3.71496230e-01
4.29155976e-01 5.46186507e-01 -7.23530889e-01 -2.92371422e-01
4.76413637e-01 2.84174979e-01 7.00332895e-02 8.58905464e-02
1.92727819e-02 -8.57168436e-01 -7.86561146e-02 -1.67001680e-01
1.16425693e+00 3.23992819e-01 8.73708129e-02 -4.82561469e-01
3.43503296e-01 2.24230587e-01 2.87506700e-01 5.49655974e-01
2.82390356e-01 6.81353450e-01 4.58352745e-01 -8.00142765e-01
-1.34718931e+00 -1.22064209e+00 -4.41658497e-01 9.88403320e-01
1.62747175e-01 -4.17980939e-01 -7.57943153e-01 -2.76887745e-01
4.38979268e-01 9.00780112e-02 -6.55593336e-01 4.52911675e-01
-7.97028184e-01 -1.58296272e-01 2.20088258e-01 5.50728559e-01
4.06897366e-01 -1.05966318e+00 -7.07194269e-01 1.98906735e-01
8.09100389e-01 -9.49695230e-01 -4.62671995e-01 1.71723038e-01
-1.09244716e+00 -1.16313469e+00 -5.33524811e-01 -1.12229955e+00
7.97046125e-01 3.39258671e-01 1.44558799e+00 3.01729530e-01
-6.58632457e-01 5.27743518e-01 -8.25176090e-02 -5.36817729e-01
-1.67524040e-01 2.37404123e-01 -4.91862446e-01 -2.36096546e-01
1.46712750e-01 -8.66891742e-01 -8.11420679e-01 4.95883413e-02
-7.30980396e-01 4.38635558e-01 5.32379985e-01 5.84486425e-01
1.31486320e+00 -2.79953212e-01 3.81457619e-02 -9.83422816e-01
2.32166484e-01 -4.94057924e-01 -7.66290188e-01 -2.67963767e-01
7.41550177e-02 -1.17454663e-01 3.94689113e-01 -7.91173652e-02
-6.78335965e-01 1.81151718e-01 -7.43225276e-01 -5.43644488e-01
-4.87112820e-01 2.07449123e-01 -4.97055240e-02 -3.80426645e-01
3.20311487e-01 -1.73740506e-01 -4.61974479e-02 -7.63103306e-01
3.42270911e-01 4.37470078e-01 2.83626616e-01 -8.22224796e-01
6.34071708e-01 8.76053154e-01 3.01636338e-01 -1.31703806e+00
-7.74865568e-01 -3.63184601e-01 -7.34281182e-01 -1.48207560e-01
1.06973255e+00 -7.78839350e-01 -7.68898368e-01 7.90889144e-01
-1.45921218e+00 -6.61720872e-01 -2.92097330e-01 -6.32120250e-03
-7.23322451e-01 1.46923661e-02 -9.12477732e-01 -5.26872337e-01
-6.72766268e-01 -1.29566729e+00 1.68077075e+00 2.69512385e-01
-7.26348311e-02 -9.91530478e-01 -2.02848956e-01 -7.94404820e-02
4.13557708e-01 7.13199675e-01 1.18203139e+00 -2.20615551e-01
-9.20768023e-01 -2.56038867e-02 -5.07317960e-01 2.47241899e-01
-4.42644255e-03 3.30096096e-01 -9.77530420e-01 -1.39067993e-01
-3.24203402e-01 -1.77171096e-01 7.92409658e-01 6.15695655e-01
1.54011548e+00 1.84715569e-01 -1.83379680e-01 1.16445911e+00
1.41217744e+00 -2.73934245e-01 2.43456095e-01 1.32296294e-01
1.06554377e+00 4.58732277e-01 9.94372293e-02 4.39706564e-01
6.06142402e-01 3.54365051e-01 6.39830470e-01 -3.20473254e-01
-2.32273757e-01 -5.97363971e-02 -1.40603930e-01 9.71199393e-01
-2.13083133e-01 1.37531206e-01 -9.86308515e-01 3.89561564e-01
-1.50133502e+00 -3.16876441e-01 -6.66533947e-01 1.90076160e+00
5.09934843e-01 2.83487141e-01 -1.71631142e-01 -1.51276931e-01
5.82026660e-01 1.44325346e-01 -6.45554721e-01 -8.49270105e-01
6.33202028e-04 1.10020483e+00 4.42545861e-01 6.97394907e-01
-1.42527103e+00 1.20572877e+00 4.75371313e+00 8.79452050e-01
-1.08325410e+00 6.67311475e-02 5.49547553e-01 1.83992192e-01
-3.32884967e-01 -3.11679393e-01 -7.34347522e-01 2.62768893e-03
4.06177640e-01 4.35367942e-01 1.93371087e-01 9.39151704e-01
1.24624027e-02 1.30593792e-01 -9.87690747e-01 1.03854907e+00
-2.86586285e-01 -1.72261524e+00 2.81764269e-01 6.81441724e-02
7.63544261e-01 5.09558797e-01 -1.12572968e-01 -8.80765542e-02
1.18673079e-01 -9.71409857e-01 8.03297877e-01 4.39400375e-01
9.68297422e-01 -8.85499299e-01 3.37288350e-01 -6.59806356e-02
-1.66965652e+00 4.89817679e-01 -4.45704937e-01 -1.77056089e-01
5.24542890e-02 8.31419528e-01 -4.80390847e-01 4.11992550e-01
6.33062661e-01 6.77637815e-01 -3.26581270e-01 1.01744854e+00
3.18048261e-02 1.98326692e-01 -6.03315890e-01 9.07086655e-02
3.17142993e-01 -1.05707116e-01 4.58083659e-01 1.23989987e+00
3.60231906e-01 6.58254251e-02 2.78636873e-01 9.71162796e-01
-3.88455957e-01 2.11093426e-01 -4.85360175e-01 1.14109352e-01
7.13893533e-01 1.24993551e+00 -1.11234617e+00 -2.81813443e-01
-5.21262407e-01 1.02979076e+00 4.99022603e-01 7.24624172e-02
-5.38352013e-01 -5.92114449e-01 1.10315061e+00 2.39528045e-01
4.74096000e-01 -7.68947005e-01 -7.82295704e-01 -7.57806361e-01
-5.70239425e-02 -2.06314594e-01 -1.15447836e-02 -3.23487848e-01
-1.41906595e+00 4.26000237e-01 -2.76117682e-01 -7.95506716e-01
6.66759133e-01 -1.00063074e+00 -8.12762856e-01 7.34698534e-01
-1.52355754e+00 -1.32410264e+00 -5.70834339e-01 4.13218558e-01
6.50509059e-01 2.34682381e-01 8.22343767e-01 2.76689678e-01
-2.86362529e-01 2.59081990e-01 -2.69765377e-01 1.08743705e-01
1.07280791e-01 -1.36441183e+00 1.31235707e+00 3.49197268e-01
-7.18224347e-02 3.88342798e-01 2.89620280e-01 -7.84750581e-01
-1.65513515e+00 -1.31290245e+00 6.23558700e-01 -2.58284807e-01
4.31184202e-01 -6.63322031e-01 -9.63452280e-01 4.97281343e-01
-3.14501435e-01 1.55232340e-01 4.79888618e-01 -1.64932430e-01
-2.83168882e-01 3.82556289e-01 -1.05696023e+00 6.08830750e-01
1.50941002e+00 -3.74181569e-01 -2.11347714e-01 3.90902013e-01
9.13458943e-01 -8.48399937e-01 -1.19187379e+00 5.54721296e-01
5.23463905e-01 -1.08256543e+00 1.25489235e+00 -3.39047253e-01
5.57365179e-01 -1.28507316e-01 -2.67544389e-01 -7.24323809e-01
-3.12785000e-01 -4.12091702e-01 -1.09656088e-01 7.74916232e-01
2.08698273e-01 -6.27800882e-01 9.53633547e-01 2.28006035e-01
-8.11341405e-01 -1.26852334e+00 -9.26278472e-01 -3.80585134e-01
2.17826352e-01 -8.64216566e-01 7.94560492e-01 7.61453450e-01
-8.54993522e-01 -1.29502311e-01 3.12748879e-01 3.87978584e-01
8.35040390e-01 3.95725727e-01 9.15150166e-01 -1.52232873e+00
-2.29714508e-03 -7.58021891e-01 -4.08349454e-01 -1.34378874e+00
-6.66793883e-02 -1.20537555e+00 -1.53696239e-01 -1.82752836e+00
-3.57120186e-01 -6.58755362e-01 2.52781779e-01 3.46450567e-01
6.53544664e-02 5.26751876e-01 -5.65262213e-02 -6.28305376e-02
-4.03249979e-01 4.55345780e-01 1.45980394e+00 9.64883193e-02
-4.36712503e-01 -3.54782492e-02 -2.41240695e-01 1.08452940e+00
9.40271258e-01 -1.04765624e-01 1.17898464e-01 -8.50281775e-01
1.94378927e-01 -3.22238743e-01 6.30755305e-01 -9.06769037e-01
1.05248436e-01 1.25096723e-01 4.96916860e-01 -8.64150345e-01
5.44887185e-01 -5.84628284e-01 1.27499946e-03 3.54601443e-01
1.83397621e-01 1.28133714e-01 4.57711577e-01 1.98766649e-01
1.83504492e-01 5.88039774e-03 8.52060735e-01 -3.37579489e-01
-6.76959813e-01 8.36103916e-01 -1.00805983e-01 5.37327258e-03
7.28484750e-01 -2.92742074e-01 -8.19113920e-05 2.68201828e-01
-8.30635905e-01 1.63516656e-01 8.53128552e-01 2.15818167e-01
8.99900496e-01 -1.38040447e+00 -5.66380799e-01 4.71939325e-01
-1.93707764e-01 8.54604602e-01 5.07409036e-01 6.29224598e-01
-1.31954956e+00 4.06818092e-02 -1.23385437e-01 -7.79769957e-01
-1.04479575e+00 1.98670831e-02 2.74629682e-01 8.93371403e-02
-9.30252850e-01 1.22521102e+00 2.74350375e-01 -6.89969122e-01
1.39590874e-01 -8.20280969e-01 3.77644569e-01 -1.60620242e-01
2.54138142e-01 5.93681693e-01 3.07761759e-01 -3.55571806e-01
-2.51838923e-01 1.00235224e+00 1.43858537e-01 4.58122045e-01
1.66195810e+00 1.53497115e-01 -4.62571740e-01 3.86151582e-01
1.36660075e+00 -9.04814980e-04 -1.42321277e+00 -1.14645183e-01
2.08422448e-02 -3.63739371e-01 7.84598812e-02 -1.32461488e-01
-9.75446165e-01 1.07986319e+00 3.63082379e-01 2.53038734e-01
8.71796012e-01 2.88665950e-01 1.18280411e+00 2.47940093e-01
5.09707868e-01 -8.09272707e-01 -2.30742231e-01 9.44856346e-01
9.71680284e-01 -1.02952874e+00 -8.03366378e-02 -9.50260699e-01
3.30056727e-01 1.39699650e+00 3.51305336e-01 -7.57006943e-01
1.15170228e+00 5.36698759e-01 -2.09154531e-01 -7.85104513e-01
-3.63516718e-01 -2.66671091e-01 3.95233959e-01 3.63611460e-01
3.77880216e-01 2.64724433e-01 9.78873670e-02 4.09139454e-01
-3.89448881e-01 -3.88978273e-01 8.61433595e-02 1.09204531e+00
-6.45949304e-01 -9.76394117e-01 -2.61503935e-01 7.27150679e-01
-1.86698347e-01 -1.23910867e-01 -1.50725514e-01 6.15777731e-01
3.39451969e-01 3.44151020e-01 5.52267432e-01 -3.26833487e-01
6.32861495e-01 -1.39264435e-01 6.22758031e-01 -8.91921699e-01
-5.31017780e-01 1.45594895e-01 -5.98687492e-02 -8.78595054e-01
-9.05902088e-02 -6.08636856e-01 -1.62768888e+00 -4.25181746e-01
2.08516330e-01 -3.12542886e-01 9.04772222e-01 5.68742096e-01
7.83192515e-01 5.90257943e-01 5.36284447e-01 -1.88415790e+00
3.57589647e-02 -5.73653162e-01 -4.10559803e-01 2.64379025e-01
1.45262524e-01 -7.49621272e-01 -1.42967671e-01 -1.11060038e-01] | [8.116226196289062, -3.6572461128234863] |
a56c4a8b-31e5-4749-b532-2e33076d81ac | convolutions-through-the-lens-of-tensor | 2307.02275 | null | https://arxiv.org/abs/2307.02275v1 | https://arxiv.org/pdf/2307.02275v1.pdf | Convolutions Through the Lens of Tensor Networks | Despite their simple intuition, convolutions are more tedious to analyze than dense layers, which complicates the generalization of theoretical and algorithmic ideas. We provide a new perspective onto convolutions through tensor networks (TNs) which allow reasoning about the underlying tensor multiplications by drawing diagrams, and manipulating them to perform function transformations, sub-tensor access, and fusion. We demonstrate this expressive power by deriving the diagrams of various autodiff operations and popular approximations of second-order information with full hyper-parameter support, batching, channel groups, and generalization to arbitrary convolution dimensions. Further, we provide convolution-specific transformations based on the connectivity pattern which allow to re-wire and simplify diagrams before evaluation. Finally, we probe computational performance, relying on established machinery for efficient TN contraction. Our TN implementation speeds up a recently-proposed KFAC variant up to 4.5x and enables new hardware-efficient tensor dropout for approximate backpropagation. | ['Felix Dangel'] | 2023-07-05 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [-2.04073831e-01 5.61489128e-02 2.83770472e-01 -4.03636724e-01
6.43059090e-02 -8.29065859e-01 4.58203673e-01 1.38364151e-01
-6.09746397e-01 3.00172031e-01 3.42681557e-01 -8.34533155e-01
-2.48311728e-01 -8.75324309e-01 -8.47790420e-01 -4.69997436e-01
-6.59821212e-01 8.25976804e-02 6.57840222e-02 -4.61835414e-01
1.26722872e-01 8.16764235e-01 -1.30795658e+00 6.44377351e-01
4.40939277e-01 1.31674063e+00 -2.59239227e-01 9.53665912e-01
-2.82150835e-01 7.81033814e-01 -2.83748895e-01 -1.19482195e+00
5.60786664e-01 1.61475047e-01 -1.15315235e+00 -2.63717920e-01
6.28333807e-01 -7.76612937e-01 -5.12958169e-01 1.03358817e+00
1.29560798e-01 -4.34143879e-02 3.14930379e-01 -1.26678371e+00
-6.45294368e-01 1.30174112e+00 -1.78803392e-02 2.77955204e-01
-8.80426392e-02 1.12379082e-01 1.22451150e+00 -8.85539234e-01
3.71616006e-01 1.28380311e+00 1.17685056e+00 2.98868299e-01
-1.64027297e+00 -2.76119858e-01 2.35865023e-02 6.39137253e-02
-1.16252542e+00 -3.74749035e-01 7.06083402e-02 -3.82819712e-01
1.32656753e+00 5.29098213e-01 9.60619867e-01 3.36369663e-01
-5.94316311e-02 7.39625573e-01 6.10098839e-01 -1.83575481e-01
1.79945081e-01 3.45698521e-02 4.67207193e-01 1.05580485e+00
3.23307335e-01 -2.22355351e-01 -3.94041151e-01 -2.23380148e-01
9.39002931e-01 1.39939422e-02 -4.71237332e-01 -2.00556397e-01
-1.68159139e+00 6.32463396e-01 8.87468517e-01 2.42759123e-01
-2.02577680e-01 6.92511678e-01 6.54734850e-01 8.11418295e-01
1.58865690e-01 5.22250772e-01 -6.22367263e-01 -1.26487285e-01
-7.95049250e-01 3.27799469e-01 1.21897030e+00 1.17999685e+00
9.89238262e-01 -1.28634442e-02 -1.07522316e-01 3.62045199e-01
-1.03834145e-01 1.04016401e-01 1.46768525e-01 -1.02876163e+00
4.42866087e-01 4.93890226e-01 -3.61057073e-01 -6.45226657e-01
-5.59158564e-01 -5.83083987e-01 -1.06223202e+00 3.43693532e-02
7.05771625e-01 -5.35681881e-02 -6.59603357e-01 1.45336080e+00
-1.94170605e-02 -2.31097028e-01 -5.10326624e-02 7.52499759e-01
1.34258419e-01 2.47544125e-01 -1.65324166e-01 2.75688857e-01
1.53758156e+00 -8.71643901e-01 -3.85843426e-01 3.89936000e-01
1.14081657e+00 -5.26411533e-01 1.01982832e+00 3.81214082e-01
-1.45417404e+00 -1.25396028e-01 -1.18282270e+00 -6.95639491e-01
-6.80113137e-01 1.85896739e-01 1.61873341e+00 6.66660666e-01
-1.67341924e+00 1.26106369e+00 -9.13862467e-01 -5.99695034e-02
6.80937111e-01 5.73171079e-01 -4.98390168e-01 1.53297439e-01
-8.95376921e-01 7.75977910e-01 4.44145411e-01 5.95449507e-01
-3.76410544e-01 -1.45502651e+00 -6.26430869e-01 4.41730469e-01
-1.75495654e-01 -1.07860315e+00 1.40990794e+00 -6.59115195e-01
-1.23245025e+00 4.90138859e-01 3.63454819e-02 -1.10459781e+00
4.53357786e-01 -1.05625689e-02 -1.36829332e-01 1.14232965e-01
-3.57688069e-01 6.34690702e-01 6.56135499e-01 -3.37562352e-01
-3.89105916e-01 -3.39168966e-01 7.78415084e-01 3.45324017e-02
-5.69435775e-01 -2.65378594e-01 -2.32283905e-01 -6.09107316e-01
1.84378088e-01 -5.97486198e-01 -3.28226656e-01 3.92031550e-01
-6.13787770e-01 1.15756281e-01 1.91517860e-01 -3.27512234e-01
1.18861806e+00 -2.37730122e+00 2.85121560e-01 4.56080407e-01
1.03249359e+00 1.18393719e-01 5.50825009e-03 5.16735256e-01
-1.61280125e-01 1.77083045e-01 -1.47185206e-01 -3.24870616e-01
4.53211635e-01 3.10608149e-01 -5.24825871e-01 3.63430053e-01
4.74259466e-01 1.07525897e+00 -7.35420644e-01 -5.59549145e-02
-8.70356523e-03 5.70308685e-01 -1.13707209e+00 -1.83289215e-01
-6.35653511e-02 -3.86900514e-01 -9.56461281e-02 4.17793989e-01
8.27594578e-01 -3.12840283e-01 4.75318059e-02 -1.00324535e+00
-3.72303426e-01 6.66373432e-01 -1.18869114e+00 1.71425533e+00
-5.72807491e-01 8.11641455e-01 1.80734187e-01 -9.13731635e-01
1.78448617e-01 7.77512882e-03 3.13753039e-02 -1.86173856e-01
2.28490442e-01 4.09215361e-01 8.21787491e-02 -4.44867276e-02
7.49908268e-01 3.19883004e-02 3.51490170e-01 6.57742202e-01
5.38124979e-01 -1.25248030e-01 4.37916875e-01 5.72135270e-01
1.22514927e+00 -2.94001192e-01 -3.48675162e-01 -5.52875519e-01
2.00469732e-01 -1.51018113e-01 -3.09280992e-01 6.29269063e-01
3.46112162e-01 1.74595162e-01 9.32636857e-01 -8.38716686e-01
-1.28049707e+00 -1.21641111e+00 -2.90889055e-01 1.12236965e+00
-4.91644293e-01 -7.99174130e-01 -8.88594151e-01 -1.97558090e-01
1.99164584e-01 3.50655615e-01 -7.28111446e-01 -5.20278327e-02
-4.41782624e-01 -7.11876750e-01 1.09211290e+00 7.13619173e-01
5.76194823e-01 -1.63838670e-01 -3.54534566e-01 1.52898386e-01
3.95254701e-01 -1.09678066e+00 -5.44642568e-01 6.20560229e-01
-1.20381749e+00 -8.05604815e-01 -6.74514115e-01 -5.33491313e-01
7.52017498e-01 2.31705755e-02 9.78076458e-01 1.38045982e-01
-1.83292612e-01 2.26174235e-01 2.18570381e-01 6.49346262e-02
-2.12697953e-01 2.37493277e-01 4.36027572e-02 7.74570368e-03
1.94472939e-01 -1.06667376e+00 -6.40174031e-01 -6.41106144e-02
-1.05379117e+00 2.80413061e-01 7.57147968e-01 7.88624585e-01
2.89409488e-01 -1.53374687e-01 -4.00658011e-01 -7.78638065e-01
8.30781817e-01 -1.02331243e-01 -6.89762831e-01 2.35194176e-01
-3.59785497e-01 9.57293808e-01 7.01850951e-01 -4.54002142e-01
-4.93179679e-01 -1.80548042e-01 -4.23791036e-02 -3.71672183e-01
5.89085042e-01 4.89620626e-01 3.91625553e-01 -6.26426518e-01
7.50304878e-01 6.29786402e-02 8.05244148e-02 -4.67450023e-01
9.04172778e-01 1.68041810e-01 7.57534802e-01 -8.14518750e-01
8.09109569e-01 6.19211733e-01 2.04159409e-01 -6.36726916e-01
-6.03381097e-01 -1.50958281e-02 -6.26384139e-01 2.52817899e-01
5.26638627e-01 -7.84672916e-01 -1.47745812e+00 3.61426324e-01
-1.43127930e+00 -3.40755969e-01 -4.85544235e-01 5.26683927e-01
-2.66716570e-01 1.20324172e-01 -1.12017679e+00 -5.34202993e-01
-4.53148425e-01 -1.13097262e+00 7.25559473e-01 -3.53834391e-01
1.19214147e-01 -1.03954887e+00 -4.96446967e-01 -2.55757511e-01
8.16851735e-01 -3.17439049e-01 1.15183711e+00 -3.98802251e-01
-1.07935882e+00 7.71024078e-02 -8.66521299e-01 6.48847222e-01
-6.73650801e-01 7.41560478e-03 -9.79558170e-01 1.53407119e-02
-2.33567327e-01 6.59381300e-02 1.03783286e+00 -5.52940033e-02
1.29942763e+00 -6.64954841e-01 -6.15418442e-02 1.34754848e+00
1.37194371e+00 -6.35182321e-01 5.60916066e-01 1.90557372e-02
7.80027688e-01 2.16134429e-01 -5.03433883e-01 4.63865250e-01
4.85395491e-01 2.12757036e-01 2.93082058e-01 -2.31341366e-02
-6.83583170e-02 -1.45205840e-01 8.27046931e-02 1.07775545e+00
-5.77843308e-01 5.89025199e-01 -4.48923111e-01 2.15341017e-01
-1.43474162e+00 -7.52137601e-01 -2.85979509e-01 2.03711510e+00
9.69172060e-01 2.56098449e-01 -2.01434325e-02 4.00477260e-01
2.01463729e-01 -1.79713234e-01 -3.52403581e-01 -6.42005205e-01
-1.01860650e-02 8.47367883e-01 1.32043409e+00 5.60170829e-01
-8.94804299e-01 6.26579642e-01 7.20130873e+00 4.96182442e-01
-9.18472946e-01 8.61535147e-02 2.87723631e-01 -6.73005134e-02
-7.54962444e-01 2.55563855e-02 -6.36058629e-01 -1.01780340e-01
9.32402194e-01 -1.40587315e-01 1.14400935e+00 7.37448215e-01
-4.85395998e-01 3.66103411e-01 -1.48034859e+00 1.06732702e+00
-5.82088590e-01 -1.81956780e+00 2.55780876e-01 2.87224408e-02
3.05919677e-01 3.00113231e-01 -5.75833917e-02 1.68807030e-01
5.21915555e-01 -9.57385004e-01 8.85705829e-01 4.39805120e-01
8.69793296e-01 -5.53763390e-01 3.49873334e-01 -3.63789469e-01
-1.25302565e+00 -2.39111125e-01 -5.39948344e-01 -4.59039927e-01
-1.09145187e-01 7.65216351e-01 -6.64185464e-01 3.73930484e-01
5.40360808e-01 3.94878000e-01 -4.64240670e-01 8.79432559e-01
2.21479461e-01 1.84521556e-01 -7.56307602e-01 -3.65203619e-01
5.07065356e-01 -1.74881101e-01 1.22871563e-01 1.39475155e+00
2.71898508e-01 2.33172160e-02 -5.65940142e-01 1.14240587e+00
-2.76907086e-01 -8.72659460e-02 -3.58989060e-01 -3.23116660e-01
3.10385048e-01 1.25843096e+00 -4.72652227e-01 -3.91873717e-01
-4.08807993e-01 1.07714355e+00 6.42938554e-01 5.74365199e-01
-6.23283565e-01 -8.14466298e-01 1.11334038e+00 2.94685867e-02
4.76658016e-01 -6.22651339e-01 -4.77967918e-01 -1.45216846e+00
3.51881444e-01 -4.94891524e-01 5.88170160e-03 -6.30333960e-01
-9.93437111e-01 7.05114305e-01 -6.53760806e-02 -8.54061544e-01
2.88386829e-02 -1.19359565e+00 -2.07631111e-01 8.85975003e-01
-1.35893261e+00 -9.35602248e-01 -4.17872369e-02 7.35831439e-01
-3.40224534e-01 1.94095939e-01 9.47386026e-01 6.56666875e-01
-2.42669001e-01 9.83747303e-01 -4.92286496e-02 3.21791887e-01
-1.35960609e-01 -1.36034966e+00 9.03728366e-01 7.43104815e-01
1.17255218e-01 1.15721190e+00 3.86787206e-01 7.68334791e-02
-2.14810300e+00 -6.96461976e-01 8.40441525e-01 -6.62706196e-02
1.31327343e+00 -6.22663319e-01 -6.96801424e-01 1.02500498e+00
-1.03277691e-01 2.92716801e-01 3.33031058e-01 3.68139595e-01
-1.03222752e+00 -4.39981014e-01 -8.67561638e-01 9.17404115e-01
1.46577120e+00 -8.44810128e-01 -2.89064180e-03 3.58374387e-01
8.74105394e-01 -5.85896909e-01 -1.16069508e+00 -9.49315280e-02
8.92546952e-01 -1.07055402e+00 9.90028262e-01 -7.54330277e-01
2.02213958e-01 -3.02552372e-01 -3.26867163e-01 -1.09218585e+00
-3.46538723e-01 -1.01875997e+00 -1.53870612e-01 6.38673842e-01
6.25696182e-01 -9.60444033e-01 3.46758187e-01 6.64255083e-01
-4.51139390e-01 -7.32260942e-01 -5.87390125e-01 -5.46267331e-01
-7.78181255e-02 -9.31209981e-01 1.01280355e+00 8.91806006e-01
4.84309763e-01 3.58753018e-02 2.42856160e-01 1.31422520e-01
4.55834478e-01 -2.72513121e-01 4.35621709e-01 -9.65609133e-01
-6.72626555e-01 -9.76035774e-01 -9.46514726e-01 -1.46755898e+00
-2.50260562e-01 -1.24070752e+00 -5.10349631e-01 -1.10385072e+00
-2.38162011e-01 -6.03006005e-01 3.87849510e-02 6.56876326e-01
5.39552152e-01 4.52389598e-01 -5.99888153e-03 3.24498527e-02
-1.46953791e-01 1.29740119e-01 1.24498200e+00 -1.37723863e-01
1.54826641e-01 -1.54588163e-01 -7.02275455e-01 5.02901971e-01
4.63350922e-01 2.54372180e-01 -3.85733634e-01 -1.00489664e+00
8.73098075e-01 -2.89499193e-01 6.36292756e-01 -1.05684817e+00
4.58149940e-01 5.51904857e-01 2.17939019e-01 -3.24161053e-01
2.98914403e-01 -5.36749303e-01 3.30038890e-02 4.38929737e-01
-5.39233565e-01 4.49256212e-01 4.14329886e-01 7.75346085e-02
3.06889974e-02 -4.29729894e-02 2.95976222e-01 -5.35181798e-02
-4.35653687e-01 5.45216620e-01 -1.94357689e-02 -6.14391416e-02
3.92338902e-01 4.54030558e-02 -4.56059724e-01 -1.70229226e-01
-1.01865602e+00 -4.42999788e-02 2.23471582e-01 -2.97644347e-01
4.72054452e-01 -1.54643393e+00 -4.16675985e-01 4.45641845e-01
-1.66292593e-01 -1.42509267e-02 1.49970934e-01 1.29426134e+00
-1.02561843e+00 4.74918664e-01 -2.07824707e-01 -2.96465963e-01
-8.06713581e-01 5.10901392e-01 5.43402076e-01 -4.48306417e-03
-6.54046118e-01 1.02401173e+00 7.57029466e-03 -1.97318301e-01
4.87017810e-01 -1.39125144e+00 5.98408997e-01 -1.24699503e-01
8.80368531e-01 3.37339699e-01 5.75024128e-01 1.78943560e-01
-2.45235369e-01 1.76490143e-01 -6.61406964e-02 -1.56092420e-01
1.16935492e+00 1.95684388e-01 -7.61454403e-01 2.88148522e-01
1.60363054e+00 -2.90973395e-01 -9.64530170e-01 -2.85007268e-01
-2.33943433e-01 -1.72334835e-02 9.22330096e-03 -4.33169663e-01
-1.22421038e+00 1.13832462e+00 7.73149803e-02 4.97224420e-01
1.11038399e+00 -1.05332933e-01 6.68451905e-01 1.13210618e+00
1.87078863e-01 -5.60555995e-01 -5.31531751e-01 6.24773502e-01
6.05069280e-01 -3.18415731e-01 -1.44799516e-01 -4.70554322e-01
-1.66098066e-02 1.62944913e+00 -6.94045424e-02 -2.01414153e-01
1.03007078e+00 8.23175728e-01 -3.61093670e-01 -2.71760553e-01
-8.77298355e-01 -2.54283875e-01 1.94624528e-01 4.54906493e-01
5.08022666e-01 2.06741095e-01 -5.88816851e-02 1.68889940e-01
-8.12152207e-01 -7.90042430e-02 3.85924041e-01 4.78758752e-01
-8.89822170e-02 -8.56913924e-01 3.50779369e-02 6.58970237e-01
-3.29897642e-01 -8.62878323e-01 8.31973329e-02 7.46554971e-01
-1.23020187e-02 1.44095540e-01 4.67164934e-01 -6.26442909e-01
1.76895469e-01 1.30384177e-01 1.00304544e+00 -1.45716682e-01
-6.70745730e-01 -5.50833941e-01 3.05622071e-01 -7.63075650e-01
-1.17069833e-01 -3.45872253e-01 -1.27014875e+00 -1.12845790e+00
-1.86964065e-01 8.49123765e-03 1.00701714e+00 8.25666249e-01
6.06681764e-01 4.81746614e-01 2.49167040e-01 -8.49219024e-01
-9.48330343e-01 -8.48848283e-01 -5.38253844e-01 1.51538566e-01
5.03681660e-01 -2.41335511e-01 -2.44983375e-01 1.68175399e-02] | [6.120996475219727, 4.999866485595703] |
16ffce5f-26ef-49af-b920-06273c58a896 | on-device-model-fine-tuning-with-label | 2211.01163 | null | https://arxiv.org/abs/2211.01163v1 | https://arxiv.org/pdf/2211.01163v1.pdf | On-Device Model Fine-Tuning with Label Correction in Recommender Systems | To meet the practical requirements of low latency, low cost, and good privacy in online intelligent services, more and more deep learning models are offloaded from the cloud to mobile devices. To further deal with cross-device data heterogeneity, the offloaded models normally need to be fine-tuned with each individual user's local samples before being put into real-time inference. In this work, we focus on the fundamental click-through rate (CTR) prediction task in recommender systems and study how to effectively and efficiently perform on-device fine-tuning. We first identify the bottleneck issue that each individual user's local CTR (i.e., the ratio of positive samples in the local dataset for fine-tuning) tends to deviate from the global CTR (i.e., the ratio of positive samples in all the users' mixed datasets on the cloud for training out the initial model). We further demonstrate that such a CTR drift problem makes on-device fine-tuning even harmful to item ranking. We thus propose a novel label correction method, which requires each user only to change the labels of the local samples ahead of on-device fine-tuning and can well align the locally prior CTR with the global CTR. The offline evaluation results over three datasets and five CTR prediction models as well as the online A/B testing results in Mobile Taobao demonstrate the necessity of label correction in on-device fine-tuning and also reveal the improvement over cloud-based learning without fine-tuning. | ['Guihai Chen', 'Chengfei Lyu', 'Shaojie Tang', 'Fan Wu', 'Chaoyue Niu', 'Yucheng Ding'] | 2022-10-21 | null | null | null | null | ['click-through-rate-prediction'] | ['miscellaneous'] | [-7.75269866e-02 -5.93521953e-01 -5.41774392e-01 -5.27047992e-01
-7.77715087e-01 -5.89125991e-01 6.57798201e-02 -6.45190701e-02
-2.96231449e-01 5.71704805e-01 -3.78085613e-01 -4.81461376e-01
-2.47821689e-01 -6.37706220e-01 -8.01171482e-01 -6.48174167e-01
1.96022585e-01 7.41645634e-01 2.57883728e-01 3.17207128e-02
-1.28264176e-02 5.83542585e-02 -1.48565900e+00 5.56903243e-01
1.07906413e+00 1.27054942e+00 2.44625971e-01 4.54680651e-01
-1.18124802e-02 3.10936421e-01 -5.90393662e-01 -4.45061147e-01
3.81344318e-01 1.55270755e-01 -1.74235761e-01 -3.39626580e-01
3.82071227e-01 -6.26326621e-01 -8.45559314e-02 1.11902392e+00
7.11153269e-01 1.08374730e-01 3.07280988e-01 -1.20660174e+00
-5.67045093e-01 5.66245198e-01 -6.33713663e-01 2.39032134e-01
-8.29787701e-02 -5.74585609e-02 8.67002189e-01 -7.23817050e-01
2.22926110e-01 7.58421659e-01 6.86066091e-01 6.27663434e-01
-9.58385110e-01 -1.01863432e+00 5.51199198e-01 1.46752015e-01
-1.43188334e+00 -3.54260117e-01 3.71718913e-01 -4.67697233e-01
4.23762202e-01 5.45227468e-01 3.81942749e-01 1.01019049e+00
5.79262301e-02 8.73622358e-01 7.35342503e-01 -3.89445871e-02
5.15918076e-01 6.80692434e-01 2.33460888e-01 2.63862699e-01
2.84371823e-01 -1.87384620e-01 -5.69557428e-01 -3.35842013e-01
4.80473340e-01 5.23331404e-01 1.56513035e-01 -2.93651819e-01
-8.09175789e-01 5.70314527e-01 1.54193044e-01 4.11945432e-02
-3.37422997e-01 -1.69470236e-02 3.75220090e-01 1.74902245e-01
6.09092355e-01 3.13281268e-02 -1.27565444e+00 -2.50093520e-01
-9.02389705e-01 -8.02956372e-02 7.23945677e-01 1.12845838e+00
6.29058719e-01 -4.67982203e-01 -4.51160908e-01 8.90156150e-01
1.28512815e-01 6.90065205e-01 6.58670366e-01 -5.99702060e-01
5.74884355e-01 4.39937055e-01 4.79964197e-01 -8.36910367e-01
-2.86104754e-02 -8.38348210e-01 -7.74408221e-01 -6.39300168e-01
3.01862419e-01 -3.70415747e-01 -5.23721337e-01 1.66879523e+00
5.98821819e-01 3.02823424e-01 -6.07402146e-01 8.60533834e-01
1.88526154e-01 3.62427503e-01 2.52171643e-02 -4.79782313e-01
1.16748071e+00 -1.04150987e+00 -6.30981207e-01 3.30028459e-02
8.38283002e-01 -5.76126397e-01 1.54162967e+00 5.87636113e-01
-9.06229079e-01 -7.26448834e-01 -9.83682096e-01 2.05370948e-01
-3.97755712e-01 5.53153872e-01 7.90219426e-01 8.93334568e-01
-6.37120962e-01 5.64771712e-01 -7.11530626e-01 -1.66346412e-02
4.42159027e-01 7.70483136e-01 1.74043640e-01 -5.91061711e-02
-1.08643138e+00 1.02797538e-01 -1.20807394e-01 1.44741088e-01
-5.42065978e-01 -9.95249391e-01 3.05335879e-01 1.27682090e-01
5.61324000e-01 -6.95546865e-01 1.42991507e+00 -8.79138052e-01
-1.53748858e+00 3.73284906e-01 -2.30601355e-01 -1.04441240e-01
6.74679458e-01 -3.05177033e-01 -6.65408731e-01 -6.05645418e-01
6.71921298e-02 1.65947899e-01 8.99732351e-01 -9.93760347e-01
-1.04552603e+00 -6.03346944e-01 -3.22289616e-02 7.80107379e-02
-7.03352153e-01 -3.45778376e-01 -9.55421090e-01 -4.94245201e-01
-1.99335307e-01 -1.22989392e+00 3.50801125e-02 -3.51238251e-01
-3.51265460e-01 -2.27958679e-01 8.32740664e-01 -4.12878305e-01
1.65719926e+00 -2.24848700e+00 -5.93190432e-01 4.01614308e-01
2.54291147e-01 4.94954705e-01 -6.89938990e-03 -1.20268919e-01
1.39971957e-01 1.23836078e-01 6.64783061e-01 -3.55757415e-01
7.56457150e-02 1.49828076e-01 -2.82879442e-01 2.15051115e-01
-4.83411968e-01 7.00533390e-01 -8.93909276e-01 -7.81402811e-02
-1.04457684e-01 2.61573941e-01 -8.58171046e-01 9.77399275e-02
-2.70666957e-01 4.51796621e-01 -5.54950655e-01 5.85402548e-01
9.57230926e-01 -6.87807798e-01 4.29627270e-01 -3.37180406e-01
1.64737895e-01 3.97484243e-01 -1.23709643e+00 1.16295278e+00
-7.93139994e-01 -3.40334550e-02 1.22970650e-02 -3.79468888e-01
4.91311848e-01 -7.12270513e-02 6.61208272e-01 -1.11734235e+00
1.13915754e-02 3.56406093e-01 -1.94569662e-01 -2.89099187e-01
5.26371777e-01 3.89860183e-01 -1.34455478e-02 7.83054054e-01
-4.91849184e-01 9.81270671e-01 -1.27230883e-01 3.53193358e-02
8.65869403e-01 -2.63200760e-01 -2.05666780e-01 -8.04942399e-02
3.11525673e-01 -4.55812454e-01 6.60692453e-01 9.67253804e-01
-1.00882322e-01 1.51155025e-01 4.25128751e-02 -3.12702209e-01
-7.80634880e-01 -8.74182701e-01 -5.35247624e-02 1.87842441e+00
1.72069922e-01 -5.88167608e-01 -8.54300320e-01 -1.02125013e+00
2.43393645e-01 4.31673229e-01 -5.05501926e-01 -1.64083585e-01
-2.94418514e-01 -9.51629937e-01 6.81064203e-02 4.34912980e-01
3.85831416e-01 -4.71454531e-01 8.00551940e-03 2.42052495e-01
-2.04001755e-01 -7.66630054e-01 -1.06857586e+00 1.75691858e-01
-8.42772901e-01 -7.21824706e-01 -2.17049450e-01 -1.87179565e-01
7.59268582e-01 6.81184769e-01 8.41829717e-01 1.20428778e-01
1.27685010e-01 4.14626189e-02 -2.46217459e-01 -3.77065331e-01
1.38797343e-01 5.78097820e-01 3.75172734e-01 3.62861395e-01
6.19517267e-01 -4.38205808e-01 -1.04021668e+00 9.87066388e-01
-6.72895253e-01 -2.25278229e-01 3.11511844e-01 7.52997994e-01
9.22069550e-01 3.09645236e-01 6.68977022e-01 -1.29894435e+00
7.47560918e-01 -7.15476632e-01 -5.93977213e-01 3.67038608e-01
-1.11805558e+00 -3.05193126e-01 5.98768532e-01 -9.82603967e-01
-6.88077688e-01 -1.63272172e-01 1.57515686e-02 -4.14483517e-01
3.36739510e-01 2.26897150e-01 -3.59860510e-01 1.97116315e-01
5.04607320e-01 -1.74174726e-01 -4.84401733e-01 -8.45510483e-01
3.53899986e-01 1.23357308e+00 1.60369709e-01 -4.36527491e-01
4.35939550e-01 2.59235919e-01 -4.90106851e-01 -2.59244442e-01
-1.08082545e+00 -6.42886043e-01 -3.44668597e-01 -1.06994957e-01
2.93940425e-01 -1.17763710e+00 -1.17441022e+00 2.73392022e-01
-6.01843238e-01 -5.58314383e-01 -2.76554793e-01 2.96220988e-01
-8.41246992e-02 -4.55687707e-03 -4.70510423e-01 -8.21008563e-01
-6.22126460e-01 -1.08370245e+00 1.12829626e+00 3.24235499e-01
-4.64070663e-02 -7.78480709e-01 -1.72313079e-01 5.35876036e-01
4.80231076e-01 -7.23838985e-01 7.32382596e-01 -7.52883077e-01
-6.40994012e-01 -5.15091419e-01 -3.45864087e-01 2.70810217e-01
2.48460382e-01 -2.43109792e-01 -1.12535250e+00 -6.48138940e-01
-8.31016377e-02 2.53705144e-01 3.20016533e-01 3.31834763e-01
1.69116879e+00 -4.67124790e-01 -6.09830499e-01 5.02127469e-01
1.09772158e+00 2.63403028e-01 2.81152844e-01 5.86775728e-02
8.54488969e-01 5.88735603e-02 1.07786953e+00 7.18414187e-01
4.90123630e-01 1.06844592e+00 1.99141592e-01 2.24442214e-01
1.39395028e-01 -5.96108437e-01 4.16830152e-01 1.16602004e+00
2.03665823e-01 -3.63673627e-01 -3.28902394e-01 1.70506462e-01
-2.14023137e+00 -5.55855393e-01 -8.32770243e-02 2.92826676e+00
8.90675247e-01 1.56789407e-01 3.10721248e-01 -2.03868017e-01
8.08876872e-01 -4.38410789e-01 -1.03299022e+00 -9.87757593e-02
3.02522153e-01 -1.02180354e-01 9.02739584e-01 3.82686973e-01
-8.26272845e-01 6.99100196e-01 5.83762121e+00 1.15549147e+00
-1.40523076e+00 7.50176013e-01 8.28985989e-01 -5.81562877e-01
-3.85961980e-01 -3.07224512e-01 -1.28117061e+00 1.23624551e+00
1.22956944e+00 1.93845332e-01 8.60438466e-01 1.08581924e+00
3.42364937e-01 1.97552815e-02 -1.14474487e+00 1.32438409e+00
-3.74135911e-01 -1.15923142e+00 -2.30436370e-01 3.52669626e-01
9.54269528e-01 2.99396366e-01 4.05641139e-01 5.86812139e-01
1.46740660e-01 -3.87331575e-01 5.82056701e-01 6.25411093e-01
1.18199694e+00 -6.57517910e-01 6.83162332e-01 5.82104623e-01
-9.56667483e-01 -3.93727243e-01 -2.58061022e-01 1.29288524e-01
-1.47629470e-01 1.11210418e+00 -9.23964560e-01 1.57183811e-01
9.07066703e-01 3.82731408e-01 -7.00606167e-01 7.62136102e-01
2.70791322e-01 5.84761500e-01 -4.14386034e-01 -1.21686377e-01
-4.09521848e-01 -2.05997080e-01 -6.96078166e-02 8.86511266e-01
4.59174454e-01 -2.55429715e-01 9.81100947e-02 2.84142137e-01
-4.02888060e-01 2.07996443e-01 1.49803147e-01 5.41440621e-02
7.50412464e-01 1.34293211e+00 -2.53928721e-01 -4.53718126e-01
-2.53410310e-01 9.17912662e-01 1.61627889e-01 2.34568492e-01
-1.12679100e+00 4.91993316e-03 8.84888113e-01 5.50774693e-01
3.74499887e-01 9.82078444e-03 -3.89983982e-01 -1.26443577e+00
2.56835848e-01 -7.68491447e-01 2.32436433e-01 -6.87752843e-01
-1.59897745e+00 2.71494210e-01 -5.35277605e-01 -1.27534556e+00
-3.68944965e-02 -2.53068417e-01 -3.25107694e-01 8.11926842e-01
-1.21569061e+00 -8.22955251e-01 -3.49411458e-01 8.22715938e-01
3.24019581e-01 -4.71527167e-02 5.62108934e-01 1.17140877e+00
-8.14866722e-01 1.49314225e+00 6.22165382e-01 -4.85525966e-01
1.07332766e+00 -8.41683686e-01 2.65017509e-01 3.91766518e-01
-1.25241160e-01 9.98924792e-01 3.67580563e-01 -8.67447078e-01
-1.43887532e+00 -1.35313570e+00 7.49590516e-01 -4.89525288e-01
5.85359693e-01 -6.25615358e-01 -8.27768743e-01 5.65191686e-01
-5.30605793e-01 1.18971184e-01 8.57432008e-01 5.68391383e-01
-2.88107425e-01 -8.38149548e-01 -1.13439441e+00 5.80284476e-01
1.16444004e+00 -6.14653647e-01 5.80119967e-01 7.08853483e-01
5.84602773e-01 -3.63934249e-01 -8.45001578e-01 2.25166947e-01
7.91971147e-01 -6.67899549e-01 7.19241440e-01 -6.43572748e-01
-3.48191887e-01 -3.42310220e-01 -2.77340353e-01 -9.40938592e-01
-4.00438666e-01 -9.10816133e-01 -4.65905964e-01 1.38504922e+00
4.63626534e-01 -6.07618809e-01 1.11532485e+00 8.30947459e-01
2.40322143e-01 -7.57411599e-01 -6.14944458e-01 -4.74499941e-01
-4.11721200e-01 -4.79337573e-01 8.81572187e-01 9.28964376e-01
-4.15812522e-01 4.18094963e-01 -6.65451884e-01 1.59688830e-01
2.99083948e-01 3.00107673e-02 9.84451175e-01 -1.36488843e+00
-6.16486430e-01 1.10237651e-01 4.03473973e-01 -1.40490341e+00
-3.69170219e-01 -6.60517037e-01 -7.86246881e-02 -8.27907205e-01
3.49720538e-01 -1.26716220e+00 -6.27268910e-01 4.15591568e-01
-1.40944287e-01 2.39971891e-01 -7.32818339e-03 5.94381571e-01
-1.16063952e+00 1.59578949e-01 1.11262238e+00 -3.98844667e-02
-5.38248301e-01 6.78049326e-01 -9.95944619e-01 3.71763408e-01
7.32366025e-01 -6.71917439e-01 -6.23368800e-01 -3.42867970e-01
7.21111298e-01 -2.07349494e-01 -1.48970157e-01 -6.74339831e-01
4.25830215e-01 -1.58896863e-01 3.89078766e-01 -6.19768381e-01
4.78462018e-02 -9.56178606e-01 3.52686733e-01 4.48992215e-02
-2.85022169e-01 8.18753690e-02 -1.18597537e-01 8.38280380e-01
3.83219630e-01 8.86458755e-02 5.98433614e-01 1.95674986e-01
-2.17978626e-01 7.12594748e-01 -8.29080585e-03 -1.69748455e-01
8.16642761e-01 1.08962562e-02 -4.14351344e-01 -2.06947282e-01
-7.78554261e-01 2.37839803e-01 3.51583749e-01 6.35198116e-01
-1.28955673e-02 -1.42510903e+00 3.15947123e-02 3.70880783e-01
1.05038628e-01 -3.03887516e-01 7.66444147e-01 9.88328338e-01
2.17769414e-01 4.06968474e-01 2.65227139e-01 -6.27632558e-01
-1.14972091e+00 7.04187751e-01 2.95338511e-01 -4.86543715e-01
9.93561372e-03 7.52925515e-01 1.92005262e-01 -5.27834594e-01
4.74100769e-01 -1.88329577e-01 2.59519130e-01 9.74811390e-02
6.78921938e-01 5.42672157e-01 5.51277101e-01 1.08481327e-03
-2.82251686e-01 2.83036947e-01 -4.95116234e-01 2.43102625e-01
9.76169109e-01 -6.24336004e-01 7.40323886e-02 5.18911600e-01
1.24271393e+00 5.12518883e-01 -1.28480494e+00 -4.19200510e-01
-3.80066216e-01 -7.70435750e-01 4.60628942e-02 -1.14364684e+00
-1.33090174e+00 6.12643123e-01 1.08649731e+00 2.65401840e-01
1.05226123e+00 -2.19357416e-01 1.14117455e+00 2.87550122e-01
5.51279008e-01 -1.60220456e+00 -4.24950710e-03 2.64208317e-01
8.75348598e-02 -1.29159164e+00 -9.99442935e-02 -2.24161208e-01
-5.53198636e-01 4.68846798e-01 6.81627691e-01 3.48961234e-01
9.47349072e-01 1.30852982e-01 -5.38771600e-02 1.80123940e-01
-8.87511730e-01 2.51148254e-01 1.72117114e-01 4.80819702e-01
1.93575934e-01 5.27109027e-01 -4.58332971e-02 1.21688437e+00
1.19059920e-01 3.38562459e-01 8.20124671e-02 4.85732436e-01
-1.25032842e-01 -1.37869751e+00 -3.84417921e-02 9.25438285e-01
-5.21351516e-01 -7.27727264e-02 -6.24900572e-02 1.67336062e-01
5.79241335e-01 9.36403573e-01 5.13614304e-02 -9.26937759e-01
4.36668992e-01 -1.91266820e-01 1.47271380e-01 -4.17024523e-01
-7.55148232e-01 3.40997696e-01 -1.19037353e-01 -6.92367077e-01
8.96297768e-02 -6.89025700e-01 -8.91196072e-01 -7.33634114e-01
-6.03056014e-01 2.42588669e-01 9.34435904e-01 8.55849385e-01
9.86418247e-01 5.58059275e-01 1.01860213e+00 -5.87714970e-01
-7.80135214e-01 -8.61987412e-01 -7.35558689e-01 1.75256610e-01
1.59980923e-01 -5.56244314e-01 -3.10183674e-01 -2.93898612e-01] | [5.925634860992432, 6.259609222412109] |
0ac0fe1a-d019-46d9-ae8e-e61d07268338 | learning-target-oriented-dual-attention-for | 1908.04441 | null | https://arxiv.org/abs/1908.04441v1 | https://arxiv.org/pdf/1908.04441v1.pdf | Learning Target-oriented Dual Attention for Robust RGB-T Tracking | RGB-Thermal object tracking attempt to locate target object using complementary visual and thermal infrared data. Existing RGB-T trackers fuse different modalities by robust feature representation learning or adaptive modal weighting. However, how to integrate dual attention mechanism for visual tracking is still a subject that has not been studied yet. In this paper, we propose two visual attention mechanisms for robust RGB-T object tracking. Specifically, the local attention is implemented by exploiting the common visual attention of RGB and thermal data to train deep classifiers. We also introduce the global attention, which is a multi-modal target-driven attention estimation network. It can provide global proposals for the classifier together with local proposals extracted from previous tracking result. Extensive experiments on two RGB-T benchmark datasets validated the effectiveness of our proposed algorithm. | ['Xiao Wang', 'Rui Yang', 'Jin Tang', 'Yabin Zhu', 'Chenglong Li'] | 2019-08-12 | null | null | null | null | ['rgb-t-tracking'] | ['computer-vision'] | [-8.34158659e-02 -3.45698327e-01 -4.10232246e-01 -1.74546093e-01
-9.24209118e-01 -5.40737212e-01 4.75042075e-01 -6.49997711e-01
-4.60494131e-01 3.38363439e-01 9.98660251e-02 -1.75660159e-02
5.31821810e-02 -1.22225940e-01 -5.65391779e-01 -1.02372992e+00
6.15846992e-01 1.19735606e-01 5.24956286e-01 1.11853845e-01
1.69562370e-01 4.72609311e-01 -1.48110366e+00 1.89086720e-01
4.77159232e-01 1.45270300e+00 1.75869495e-01 8.96438479e-01
-8.97527710e-02 8.31937551e-01 -3.49106550e-01 -1.97823644e-01
3.90458554e-01 -5.49850576e-02 -4.17704552e-01 -1.15619585e-01
7.34584033e-01 -3.76051217e-01 -3.18395883e-01 8.91669214e-01
6.00119054e-01 1.40973046e-01 4.89708573e-01 -1.69278169e+00
-1.00407064e+00 2.14976221e-01 -8.84325087e-01 5.66786289e-01
2.39591271e-01 7.32225776e-01 6.98470652e-01 -8.98537695e-01
3.22129011e-01 1.33492625e+00 8.56168687e-01 8.27665508e-01
-7.62620926e-01 -7.93087840e-01 3.44362676e-01 3.29760224e-01
-1.06767070e+00 -1.54960141e-01 1.01950943e+00 -4.39632624e-01
8.03142130e-01 2.80285299e-01 7.94884086e-01 1.18746078e+00
2.46999145e-01 1.10227454e+00 1.19497812e+00 -1.74292743e-01
-3.25750083e-01 2.54323453e-01 1.97931141e-01 6.76792562e-01
3.22713017e-01 6.02698624e-01 -4.38402802e-01 1.46599533e-02
5.74088037e-01 2.78863341e-01 -2.72977650e-01 -3.31182301e-01
-1.12948430e+00 4.77142423e-01 1.08842611e+00 2.83551395e-01
-2.95231432e-01 8.19515049e-01 2.05006868e-01 -2.34342441e-01
4.29738611e-01 -2.58004367e-01 -4.00769264e-01 3.00375044e-01
-5.92284858e-01 -8.61521810e-02 -6.97553009e-02 1.17593908e+00
5.80919385e-01 2.15186447e-01 -6.62797689e-01 1.22945108e-01
1.13502157e+00 1.05016851e+00 3.75557572e-01 -6.24630392e-01
3.08564752e-01 8.23752999e-01 2.24462792e-01 -6.06578350e-01
-5.07569790e-01 -2.88461387e-01 -5.80241382e-01 4.79453504e-01
4.59377497e-01 -1.01823233e-01 -1.18244016e+00 1.64414871e+00
5.77200651e-01 3.81646454e-01 -8.26728269e-02 1.52018559e+00
1.29869688e+00 3.82846326e-01 5.18423796e-01 7.97778368e-02
1.48700726e+00 -1.02392924e+00 -8.02264988e-01 -3.98820758e-01
1.71839476e-01 -9.35814023e-01 5.29509544e-01 -1.56891227e-01
-7.91011274e-01 -1.05291092e+00 -7.42158890e-01 -2.67813027e-01
-6.19250655e-01 5.26608586e-01 6.36442721e-01 1.00580072e+00
-8.49493444e-01 -1.90355815e-02 -8.87769938e-01 -6.11066878e-01
6.07596517e-01 4.80634272e-01 -3.04621398e-01 6.24957606e-02
-9.87018168e-01 9.65764821e-01 4.73896295e-01 6.08027399e-01
-1.01546013e+00 -5.64084172e-01 -6.09167874e-01 -2.69839376e-01
3.81197095e-01 -9.95391428e-01 1.17097151e+00 -9.74041581e-01
-1.52978063e+00 7.03557312e-01 -6.74338713e-02 -6.49334937e-02
2.12685317e-01 -3.94462168e-01 -1.85165077e-01 -5.55465892e-02
2.09691972e-02 7.17887819e-01 1.16210735e+00 -1.36728632e+00
-7.46811092e-01 -6.04001105e-01 -2.57903844e-01 1.41410217e-01
-2.12035745e-01 2.08815768e-01 -8.23802888e-01 -5.07066727e-01
1.08100198e-01 -9.57383275e-01 -2.71265686e-01 3.80664289e-01
-4.22148287e-01 -5.49470246e-01 1.28031123e+00 -4.44622010e-01
7.38338232e-01 -1.87504935e+00 2.13165551e-01 -7.40555115e-03
1.91389993e-01 3.00948054e-01 2.58529782e-02 -1.11069791e-01
7.50697926e-02 -5.93174621e-02 4.35125470e-01 -3.01741421e-01
1.38989717e-01 -1.42532513e-01 -1.78686813e-01 7.43115842e-01
1.91141546e-01 1.56676090e+00 -8.54015350e-01 -9.34590340e-01
6.39375746e-01 5.79022646e-01 -1.02685966e-01 3.51491958e-01
-2.89729416e-01 6.06074393e-01 -9.31707501e-01 1.21347857e+00
7.23021805e-01 -2.59744376e-01 -2.24071473e-01 -8.50434899e-01
-2.53752947e-01 -4.58598882e-01 -7.14665353e-01 1.66645265e+00
-4.93602175e-03 6.91480935e-01 -9.62205604e-02 -4.15569335e-01
7.16941595e-01 1.15623496e-01 5.87917328e-01 -6.46592081e-01
7.40514874e-01 -8.68029669e-02 -2.56929189e-01 -7.01202810e-01
5.27183115e-01 -1.51810572e-01 -1.55633330e-01 5.81072688e-01
7.91523680e-02 3.92854303e-01 -4.85976964e-01 -4.55966257e-02
7.36818552e-01 8.14489782e-01 -1.61973655e-01 2.86271304e-01
4.70662117e-01 3.08513433e-01 6.19262278e-01 8.31380844e-01
-6.79364741e-01 5.59445381e-01 -3.85355920e-01 -4.49452013e-01
-8.11357319e-01 -9.37184513e-01 1.85487002e-01 1.36759663e+00
6.33074641e-01 -5.31566627e-02 -3.12220186e-01 -9.51209605e-01
2.08937183e-01 2.29958892e-01 -1.03530848e+00 -3.10335547e-01
-4.72263128e-01 -4.50938314e-01 3.90045553e-01 9.86855090e-01
4.48720515e-01 -1.14856768e+00 -9.75804448e-01 -1.73130617e-01
-1.34532958e-01 -9.81213629e-01 -4.70506221e-01 2.68185943e-01
-8.62390399e-01 -1.13001132e+00 -8.03111434e-01 -3.35864276e-01
3.98866147e-01 6.84990644e-01 6.35473251e-01 5.28172478e-02
-3.78318340e-01 1.10009563e+00 -3.48416716e-01 -4.67002541e-01
1.85142204e-01 -1.37821943e-01 3.99816036e-02 2.41202489e-01
5.56535125e-01 3.24777693e-01 -7.59864092e-01 4.27627176e-01
-5.02531588e-01 -2.15638623e-01 9.58542168e-01 5.26502967e-01
4.73954022e-01 -4.39386457e-01 -9.88344550e-02 -4.87286523e-02
3.35920379e-02 -2.89184630e-01 -6.88285589e-01 6.22730434e-01
1.23169683e-01 5.41725270e-02 -7.97433034e-02 -5.86228788e-01
-1.32514513e+00 5.09109437e-01 2.18219921e-01 -1.22651982e+00
-1.47369951e-01 -2.83745509e-02 -9.25690308e-02 -7.22062588e-01
2.49507830e-01 2.85055131e-01 -2.21362799e-01 -5.33343852e-01
7.48315215e-01 5.65344870e-01 4.65589494e-01 -4.70855623e-01
1.24573135e+00 6.01790309e-01 -1.28503025e-01 -4.54104096e-01
-1.00503564e+00 -7.03629732e-01 -7.37607360e-01 -9.62619185e-01
1.39163375e+00 -8.47929060e-01 -1.28703225e+00 3.29761922e-01
-1.16542923e+00 -1.46632418e-01 -5.91908842e-02 6.54065967e-01
-5.03187537e-01 1.99166894e-01 -2.48629138e-01 -1.24843287e+00
-6.01159632e-01 -1.10433400e+00 1.76789105e+00 6.33020878e-01
5.13370037e-01 -8.36701393e-01 1.55550554e-01 4.69353259e-01
6.57488108e-01 1.45558462e-01 9.97657701e-03 -1.50421634e-01
-1.19337547e+00 -3.75502408e-01 -6.76742256e-01 -1.05685532e-01
-9.83211994e-02 6.74153119e-02 -1.35291958e+00 -3.48443717e-01
-2.11074635e-01 -4.43269759e-01 1.15871084e+00 5.88607848e-01
1.05469525e+00 2.24104717e-01 -9.02520061e-01 7.52483547e-01
1.56333935e+00 3.34882848e-02 4.08136070e-01 3.62380743e-01
1.16997361e+00 6.97969645e-02 1.07528341e+00 2.13991299e-01
8.93765464e-02 9.39990222e-01 7.18053639e-01 -1.24289408e-01
-3.11959326e-01 6.94785267e-02 4.88779545e-01 2.70230681e-01
-4.12255049e-01 -1.22854812e-02 -8.47566247e-01 2.52265334e-01
-2.05316782e+00 -1.06026256e+00 -2.41640463e-01 1.81396973e+00
2.48440638e-01 8.41301978e-02 2.61040449e-01 -3.28362137e-01
7.35051394e-01 -1.11044645e-01 -6.54469132e-01 2.85619885e-01
-2.99026556e-02 -1.21414691e-01 8.75946462e-01 -2.01151296e-02
-1.35545647e+00 9.37676668e-01 6.04443121e+00 7.69467115e-01
-1.10260081e+00 5.08305907e-01 1.95170995e-02 -1.52702495e-01
6.38260767e-02 2.57583186e-02 -1.04190278e+00 4.06138659e-01
5.51807165e-01 1.21765234e-01 -2.57416219e-01 9.17476654e-01
-1.67265564e-01 3.05374693e-02 -7.93568432e-01 1.00678158e+00
1.95534587e-01 -8.49592566e-01 -2.55417079e-01 -2.00362671e-02
3.72732431e-01 1.81195125e-01 3.44450891e-01 4.63310182e-01
2.65773118e-01 -6.19587243e-01 9.31719661e-01 1.15959668e+00
4.13022578e-01 -3.22037816e-01 7.33549595e-01 -3.02520581e-02
-1.85653877e+00 -3.24283153e-01 -3.24397981e-01 5.22611201e-01
7.85309598e-02 -1.95086077e-01 -4.51105654e-01 9.29336071e-01
1.15321922e+00 9.64306295e-01 -9.37162161e-01 1.45438802e+00
-1.21929556e-01 2.98967838e-01 -3.17438394e-01 -1.91195190e-01
3.82389933e-01 2.23407671e-01 6.01739645e-01 9.88482654e-01
3.13339621e-01 7.00109899e-02 3.79859328e-01 9.63159621e-01
3.32978159e-01 -2.49826342e-01 -5.78159034e-01 2.41404101e-01
9.55217183e-02 1.62912989e+00 -6.47760510e-01 -1.53376371e-01
-5.69314241e-01 7.52164423e-01 3.39567065e-01 4.88391608e-01
-1.36823380e+00 2.26496533e-02 4.16762680e-01 -3.06103617e-01
6.48419380e-01 -1.10141605e-01 -1.94294173e-02 -8.89078498e-01
-3.52635801e-01 -1.98651448e-01 4.78124201e-01 -1.59505773e+00
-1.36318135e+00 5.53507686e-01 -9.24399048e-02 -1.68967080e+00
4.05038774e-01 -9.91605937e-01 -5.21123528e-01 7.56473601e-01
-1.48690498e+00 -1.87764299e+00 -7.81315565e-01 7.56439626e-01
2.30836973e-01 -6.09186366e-02 1.02312014e-01 3.02911550e-01
-9.19311404e-01 5.58551073e-01 -2.94325322e-01 2.57495672e-01
7.92363942e-01 -1.08050597e+00 -1.05697185e-01 7.70656645e-01
-6.04114532e-02 6.69425189e-01 3.99620771e-01 -7.12082088e-01
-1.98612988e+00 -1.39096868e+00 1.53609887e-01 -1.00034094e+00
5.87469637e-01 3.24765965e-02 -6.59706891e-01 9.58123863e-01
5.72051823e-01 4.68647718e-01 4.14630711e-01 -1.60667464e-01
-3.60603333e-01 -2.86690235e-01 -9.02437031e-01 1.93019748e-01
7.16449738e-01 -3.41638714e-01 -6.67101741e-01 2.84622967e-01
7.16561854e-01 -4.86350417e-01 -8.89293969e-01 4.35701996e-01
7.68570483e-01 -6.93612456e-01 1.28825033e+00 -5.39848089e-01
-3.86635423e-01 -8.33956778e-01 -1.22051544e-01 -7.29645312e-01
-5.04709303e-01 -2.62654722e-01 -2.81450510e-01 1.33328080e+00
-1.61471933e-01 -2.59433359e-01 9.07914877e-01 5.34945667e-01
-3.21162254e-01 -2.81299919e-01 -9.93368328e-01 -6.54473782e-01
-4.22915965e-01 -5.66786051e-01 2.67559022e-01 5.60784042e-01
-5.01598477e-01 1.53727472e-01 -6.59578443e-01 3.06028366e-01
1.19867933e+00 5.01762331e-01 8.29949439e-01 -1.14725292e+00
1.60503626e-01 -6.89505339e-01 -4.76627320e-01 -8.44425082e-01
5.70222512e-02 -8.52925122e-01 3.64992529e-01 -1.40142000e+00
5.40078998e-01 -4.36031014e-01 -6.72513127e-01 7.08922684e-01
-5.12939870e-01 7.01217651e-01 3.59094501e-01 1.46825433e-01
-1.29332936e+00 8.89491081e-01 1.21593690e+00 -3.74284625e-01
8.82843435e-02 -1.36040255e-01 -4.34809595e-01 3.90847772e-01
7.83142209e-01 -6.29013538e-01 9.99124348e-02 -4.16167587e-01
-1.52044430e-01 -1.13907471e-01 8.97058547e-01 -1.27995646e+00
7.96772420e-01 -2.59564489e-01 8.88669074e-01 -1.26870024e+00
5.91746509e-01 -1.39670300e+00 1.81471363e-01 5.08154273e-01
1.36010628e-02 -8.30898806e-02 5.41075110e-01 1.06433284e+00
1.95688263e-01 2.37488523e-01 7.06582308e-01 -4.22421731e-02
-1.14800429e+00 5.65101087e-01 -1.64463729e-01 -3.16384256e-01
1.34294057e+00 -4.67012525e-01 -4.77848798e-01 2.69407153e-01
-6.84299827e-01 4.57003444e-01 2.12931201e-01 7.91205943e-01
7.60914922e-01 -1.93742764e+00 -3.41341048e-01 -3.10263671e-02
4.87860918e-01 -4.81890649e-01 3.74014616e-01 1.26952767e+00
4.50132787e-02 7.07439363e-01 -4.08366889e-01 -1.13088667e+00
-1.58894825e+00 9.89927351e-01 6.19891286e-01 1.76772043e-01
-4.65723068e-01 9.25717592e-01 -1.87514946e-02 -2.54880816e-01
4.28706229e-01 -4.76856679e-01 -2.95869052e-01 -1.27147064e-01
3.25497657e-01 1.88200518e-01 -3.50715250e-01 -1.12365723e+00
-8.12759876e-01 1.37345636e+00 2.17073351e-01 1.64592579e-01
8.96487594e-01 -3.90496373e-01 1.52914464e-01 3.19281548e-01
8.84663403e-01 -5.48775077e-01 -1.31492352e+00 -4.33748871e-01
-6.62962049e-02 -6.07812643e-01 4.14725631e-01 -6.88494146e-01
-1.57557237e+00 8.31670105e-01 1.20426798e+00 -7.72588281e-03
1.20940828e+00 1.23474926e-01 4.91279840e-01 9.81854051e-02
2.34098807e-01 -6.40738010e-01 5.19370377e-01 1.13415249e-01
7.21838236e-01 -1.66216648e+00 5.12124673e-02 -1.46961510e-01
-5.43519497e-01 8.90486717e-01 1.20791030e+00 8.64670351e-02
5.27298391e-01 1.21107183e-01 2.54237533e-01 -3.64935845e-01
-7.49975502e-01 -1.06450403e+00 7.84729123e-01 7.60345161e-01
2.04472348e-01 -4.24554229e-01 3.40875357e-01 5.23020208e-01
7.06408739e-01 4.23348276e-03 -3.15997630e-01 1.07913291e+00
-6.73341036e-01 -7.85781384e-01 -9.53632534e-01 1.81920707e-01
-2.11318851e-01 2.28308678e-01 -6.41354978e-01 8.73460293e-01
3.29201579e-01 8.89973044e-01 -2.87208319e-01 -5.85188568e-01
1.73471034e-01 1.64829697e-02 7.63776064e-01 -1.88190401e-01
-8.81290972e-01 4.06305492e-01 -4.63730633e-01 -7.95553684e-01
-1.02796030e+00 -5.09532332e-01 -7.93657362e-01 1.14946505e-02
-7.55694389e-01 -8.63359869e-02 7.18367219e-01 8.12399387e-01
1.28818035e-01 8.27163815e-01 3.16976994e-01 -1.47392070e+00
-7.04089701e-02 -9.71296191e-01 -2.32923985e-01 2.15788439e-01
5.68967760e-01 -1.22679305e+00 -1.35723263e-01 -2.07360044e-01] | [6.347821235656738, -2.194688320159912] |
21c6f58b-9fc2-4f40-a3db-e7a2a9e55e0f | skeletal-movement-to-color-map-a-novel | 1807.07033 | null | http://arxiv.org/abs/1807.07033v1 | http://arxiv.org/pdf/1807.07033v1.pdf | Skeletal Movement to Color Map: A Novel Representation for 3D Action Recognition with Inception Residual Networks | We propose a novel skeleton-based representation for 3D action recognition in
videos using Deep Convolutional Neural Networks (D-CNNs). Two key issues have
been addressed: First, how to construct a robust representation that easily
captures the spatial-temporal evolutions of motions from skeleton sequences.
Second, how to design D-CNNs capable of learning discriminative features from
the new representation in a effective manner. To address these tasks, a
skeletonbased representation, namely, SPMF (Skeleton Pose-Motion Feature) is
proposed. The SPMFs are built from two of the most important properties of a
human action: postures and their motions. Therefore, they are able to
effectively represent complex actions. For learning and recognition tasks, we
design and optimize new D-CNNs based on the idea of Inception Residual networks
to predict actions from SPMFs. Our method is evaluated on two challenging
datasets including MSR Action3D and NTU-RGB+D. Experimental results indicated
that the proposed method surpasses state-of-the-art methods whilst requiring
less computation. | ['Alain Crouzil', 'Sergio A. Velastin', 'Pablo Zegers', 'Louahdi Khoudour', 'Huy Hieu Pham'] | 2018-07-18 | null | null | null | null | ['3d-human-action-recognition'] | ['computer-vision'] | [ 3.85394484e-01 -2.48457208e-01 -3.95019174e-01 -2.38349885e-01
-2.30133235e-01 1.13155821e-03 5.12808442e-01 -5.47981977e-01
-2.91044623e-01 3.11772794e-01 4.40349311e-01 2.00107738e-01
-1.49051607e-01 -5.58583915e-01 -7.38798976e-01 -7.18387604e-01
-2.24891946e-01 2.56799962e-02 3.92347991e-01 -2.30652705e-01
3.55181694e-01 9.50146914e-01 -1.78940964e+00 1.99315399e-01
4.68922973e-01 1.27999222e+00 1.92073602e-02 5.28874874e-01
8.83722380e-02 1.04030168e+00 -3.09963405e-01 1.43315166e-01
4.08154994e-01 -5.27710497e-01 -7.29930878e-01 4.90055501e-01
3.58646035e-01 -5.69983006e-01 -8.01182568e-01 5.82315743e-01
4.65766102e-01 3.47082406e-01 7.28532612e-01 -9.74600852e-01
-4.51281846e-01 -5.20757250e-02 -5.35245776e-01 2.46615112e-01
4.89547998e-01 3.34795207e-01 7.15434551e-01 -6.39209092e-01
7.29593575e-01 1.48945975e+00 5.61509192e-01 6.86000228e-01
-7.85839736e-01 -4.01158392e-01 1.47395536e-01 4.65296358e-01
-9.97441709e-01 -3.74184638e-01 8.61914873e-01 -5.32155335e-01
1.04690170e+00 5.44557348e-02 9.28479373e-01 1.35618210e+00
3.26888472e-01 1.37580216e+00 6.89296603e-01 -1.26194939e-01
-3.93848196e-02 -7.63458967e-01 -1.99911088e-01 8.77512097e-01
-1.21824503e-01 -2.65029520e-02 -6.84481025e-01 2.46805355e-01
1.26105297e+00 2.99069911e-01 -3.48050863e-01 -6.44033194e-01
-1.34488630e+00 4.91270572e-01 4.01510954e-01 3.50044072e-01
-5.88646293e-01 4.02536511e-01 4.92359489e-01 5.96378036e-02
4.17424798e-01 6.72696754e-02 -6.55339181e-01 -5.84989607e-01
-4.68746424e-01 1.83592066e-01 2.91745067e-01 6.20390594e-01
5.34415960e-01 2.02092037e-01 -2.44791195e-01 6.76796496e-01
2.11416289e-01 3.88451666e-01 8.31953287e-01 -9.68562663e-01
5.24128139e-01 9.83166814e-01 -8.69265050e-02 -1.27452314e+00
-5.82310259e-01 -9.39055160e-02 -9.73658741e-01 9.97335389e-02
3.08269411e-01 2.46171638e-01 -1.04476643e+00 1.50632012e+00
3.33068788e-01 3.14485908e-01 -6.22965917e-02 9.44118023e-01
7.00835407e-01 5.07920802e-01 -9.60348248e-02 9.74944383e-02
8.98597538e-01 -9.91729856e-01 -6.67155325e-01 -1.33258224e-01
6.32053554e-01 -2.80395746e-01 8.10248435e-01 2.97612429e-01
-1.07974148e+00 -1.06985331e+00 -9.85959053e-01 -4.30087261e-02
-1.27250895e-01 5.13849616e-01 6.03924990e-01 1.75914362e-01
-8.05840850e-01 1.01662159e+00 -1.24125183e+00 -3.39045197e-01
5.40792763e-01 3.84177983e-01 -6.16061449e-01 9.41721946e-02
-9.94616091e-01 7.74144530e-01 4.52746272e-01 3.02141041e-01
-1.14092875e+00 -1.64335951e-01 -1.04739833e+00 -1.52075350e-01
3.80637527e-01 -5.71489513e-01 1.09909737e+00 -9.67166543e-01
-1.76459885e+00 8.34583759e-01 1.00842267e-01 -2.55128473e-01
3.67231846e-01 -5.25115550e-01 -2.15483844e-01 4.29165691e-01
-4.87351753e-02 4.17196035e-01 1.09980404e+00 -7.62744486e-01
-4.60015297e-01 -6.94770873e-01 9.59554836e-02 2.39083081e-01
-4.38677877e-01 -2.05917791e-01 -6.51008129e-01 -8.07483137e-01
3.21611106e-01 -9.80620921e-01 -2.07269147e-01 4.61836636e-01
-1.64358065e-01 -3.22625607e-01 9.68742013e-01 -7.82005906e-01
9.99521196e-01 -2.07162976e+00 7.11635590e-01 -1.89140096e-01
1.36383787e-01 7.82050014e-01 -2.62212127e-01 2.57849306e-01
-9.13541317e-02 -2.35627994e-01 -3.79099064e-02 -1.50136247e-01
-1.01939313e-01 4.83654976e-01 1.47940099e-01 5.54010749e-01
4.34500664e-01 1.07212579e+00 -8.14309955e-01 -4.17093456e-01
4.58170503e-01 5.42852998e-01 -4.02667671e-01 3.68455380e-01
-2.93129236e-01 7.81631529e-01 -8.30161572e-01 8.05435717e-01
2.98970938e-01 -1.06913388e-01 5.98160438e-02 -2.98427135e-01
1.40813082e-01 -1.70282729e-03 -1.07153773e+00 2.09763813e+00
-2.83879757e-01 2.85766423e-01 -2.49092132e-01 -1.42880511e+00
1.12415755e+00 1.89094335e-01 9.16794240e-01 -7.10742116e-01
3.18125963e-01 1.83064967e-01 -3.42146963e-01 -9.47145224e-01
1.02669902e-01 1.34139761e-01 9.45247710e-02 3.33187968e-01
2.90920764e-01 1.13134369e-01 6.74073100e-02 -4.11930025e-01
1.26471424e+00 7.75834978e-01 4.28777903e-01 2.16255307e-01
8.74674141e-01 -3.47657830e-01 8.33451211e-01 1.71792731e-01
-3.70163709e-01 5.47398746e-01 4.31665778e-01 -8.73433053e-01
-5.87119222e-01 -7.81997204e-01 3.82189065e-01 7.65849113e-01
1.23129606e-01 -2.40905583e-01 -5.57734668e-01 -9.25260484e-01
-3.60721052e-02 6.59199581e-02 -8.91527176e-01 -3.98896903e-01
-8.22281182e-01 -2.37096369e-01 4.68465358e-01 1.07159710e+00
8.33978415e-01 -1.28107786e+00 -1.04598784e+00 2.40716159e-01
-9.18098465e-02 -1.17794645e+00 -4.65677440e-01 1.39419725e-02
-1.28746045e+00 -1.39283824e+00 -1.05105531e+00 -6.56269550e-01
4.48295325e-01 4.74681050e-01 6.94583297e-01 2.80055515e-02
-3.42343777e-01 6.09755993e-01 -6.74250662e-01 1.66349616e-02
-1.35746032e-01 -2.31901348e-01 1.35410488e-01 3.52920890e-01
3.64835620e-01 -8.82624388e-01 -7.22958565e-01 4.86092240e-01
-1.01175249e+00 -2.43100505e-02 1.02271116e+00 6.90837801e-01
6.86302006e-01 -2.27355957e-02 2.05381185e-01 -2.24642321e-01
1.66200802e-01 -2.08420932e-01 -1.29200488e-01 2.70910293e-01
5.69898337e-02 1.68693617e-01 6.05105400e-01 -5.63279390e-01
-9.03988063e-01 5.65970242e-01 -1.58319414e-01 -7.90430069e-01
-3.84659439e-01 3.01246315e-01 -3.03930461e-01 5.21907723e-03
3.50492269e-01 5.20677269e-01 1.52807564e-01 -7.42371857e-01
3.10809642e-01 4.67476845e-01 6.47771835e-01 -5.74029326e-01
5.05456626e-01 5.67840517e-01 2.53884524e-01 -1.04220569e+00
-8.24839950e-01 -5.59126139e-01 -1.25632215e+00 -5.15128314e-01
1.20898330e+00 -8.00108016e-01 -6.63300991e-01 1.06722295e+00
-1.16934621e+00 -2.04062432e-01 -2.57196248e-01 6.59984827e-01
-1.08130217e+00 7.01741159e-01 -5.61057031e-01 -6.05107129e-01
-1.70135170e-01 -1.09153056e+00 1.33355653e+00 1.58871129e-01
-7.86354672e-03 -8.28219533e-01 1.95357189e-01 5.54808557e-01
5.18750101e-02 9.25091565e-01 7.53183722e-01 -3.01376343e-01
-4.91378725e-01 -2.49707729e-01 1.64133348e-02 6.96632445e-01
4.02400732e-01 1.43753234e-02 -5.24888813e-01 -1.22879468e-01
5.59128169e-03 -5.61156631e-01 9.45418954e-01 4.37314749e-01
1.49798441e+00 -1.28671393e-01 -3.00882787e-01 7.70957351e-01
1.01303315e+00 3.51715714e-01 8.78643572e-01 3.12197626e-01
8.59499037e-01 4.02447462e-01 6.18209183e-01 6.83920383e-01
1.78488210e-01 9.07760262e-01 6.53244078e-01 9.79048535e-02
-1.46535724e-01 -3.61789644e-01 6.53695881e-01 9.91815150e-01
-7.66954422e-01 1.75947204e-01 -7.05882013e-01 3.94463003e-01
-2.16623878e+00 -8.51221204e-01 -4.13762815e-02 1.76889277e+00
5.32100439e-01 1.55955300e-01 2.66887754e-01 3.36066246e-01
4.90050435e-01 4.78755385e-01 -6.89368188e-01 -9.25511271e-02
1.12469599e-01 3.99027944e-01 8.17606375e-02 -2.50977159e-01
-1.38536727e+00 8.76142383e-01 5.46302080e+00 5.83844960e-01
-1.04150772e+00 -2.37303674e-01 1.26681894e-01 1.25693619e-01
4.73089576e-01 -4.19584543e-01 -5.42636812e-01 2.46348560e-01
5.23000717e-01 1.59725159e-01 1.23280203e-02 9.64807808e-01
3.19514811e-01 9.40673053e-02 -1.09434474e+00 1.03611279e+00
2.05272213e-01 -1.07583988e+00 1.24599852e-01 -7.56412372e-02
6.76049292e-01 -2.24014193e-01 -8.11562538e-02 1.86033458e-01
-2.48368718e-02 -9.40095663e-01 5.69448709e-01 9.70135093e-01
6.63471401e-01 -6.91610157e-01 6.05619013e-01 3.61311734e-01
-1.43856168e+00 -2.62074620e-01 -2.83642679e-01 -1.72769025e-01
1.02568887e-01 1.87713325e-01 -2.94408113e-01 7.31464088e-01
7.36910641e-01 1.68442345e+00 -5.22890031e-01 8.18500102e-01
-4.60121095e-01 2.12082729e-01 1.04800843e-01 2.02464107e-02
4.57096666e-01 -9.22057852e-02 3.85663599e-01 9.06722784e-01
4.24112201e-01 2.01239735e-01 1.57225654e-01 5.98988056e-01
2.05095038e-01 -9.75308493e-02 -6.68067694e-01 -4.51986641e-01
-2.29943663e-01 9.41315055e-01 -6.32991850e-01 -6.87170774e-02
-3.37120444e-01 1.04857337e+00 1.69904679e-01 1.20633997e-01
-7.43230700e-01 -7.95801654e-02 9.50237989e-01 5.64158075e-02
6.76462770e-01 -5.95964253e-01 3.27755421e-01 -1.30481100e+00
1.23742677e-01 -1.01679659e+00 2.11279243e-01 -7.05631852e-01
-9.99367714e-01 2.56642014e-01 -3.71741503e-02 -1.65307522e+00
-2.76430607e-01 -9.18942630e-01 -4.73060608e-01 2.63658404e-01
-1.30324829e+00 -1.23502111e+00 -5.18546760e-01 9.54261303e-01
8.10467720e-01 -2.29622200e-01 7.16687799e-01 1.99590459e-01
-6.10454082e-01 1.25137806e-01 -6.06060736e-02 3.85198563e-01
1.89407468e-01 -9.31777298e-01 4.60992068e-01 6.91289544e-01
2.32087374e-01 3.37883443e-01 5.88091053e-02 -5.31877160e-01
-1.66580188e+00 -1.00851536e+00 4.22110856e-01 -2.98274368e-01
4.34990466e-01 6.78906143e-02 -5.86538374e-01 6.34143889e-01
-3.70511562e-01 1.39636323e-01 4.76314634e-01 -3.36772293e-01
-1.28660411e-01 -2.60788221e-02 -6.71958625e-01 4.38340187e-01
1.53796637e+00 -3.94640923e-01 -7.83816159e-01 2.65617043e-01
3.54659438e-01 -4.26455945e-01 -8.72260332e-01 7.12282479e-01
8.50401223e-01 -1.16158438e+00 1.15375435e+00 -8.05659831e-01
7.20398724e-01 -2.19243407e-01 -1.50671318e-01 -1.19152081e+00
-1.96727678e-01 -4.82165754e-01 -7.57790983e-01 5.70393562e-01
-4.33846056e-01 -1.46085083e-01 8.25775445e-01 5.83385751e-02
-1.49970353e-01 -1.10710275e+00 -8.80864739e-01 -9.26928163e-01
-2.97182083e-01 -4.04122382e-01 4.64039266e-01 6.91124022e-01
-4.04970735e-01 8.66091698e-02 -6.68682933e-01 -1.50091425e-01
3.78709316e-01 3.46072428e-02 1.06545615e+00 -1.21298742e+00
-2.10220367e-01 -3.58397245e-01 -1.11386037e+00 -1.60896075e+00
3.44902843e-01 -6.14667714e-01 -7.40137650e-03 -1.64281726e+00
2.94657834e-02 1.44633353e-01 -3.77150774e-01 6.51158094e-01
1.69773698e-02 -5.12883998e-02 2.14422196e-01 2.00955868e-01
-7.12255955e-01 1.10403585e+00 1.78218877e+00 -3.11677724e-01
-5.33222593e-02 3.12203109e-01 -6.53890297e-02 9.03461576e-01
6.53886199e-01 -2.13275313e-01 -4.40476239e-01 -4.24277216e-01
-3.31596553e-01 2.69379586e-01 5.70243359e-01 -1.43599713e+00
-6.22060932e-02 -2.94510961e-01 5.49690723e-01 -9.30656612e-01
6.13568485e-01 -7.85746634e-01 -4.97593638e-03 6.40418291e-01
-2.78664798e-01 -7.20125586e-02 -4.99485899e-03 7.67296135e-01
-3.53240222e-01 5.31653129e-02 6.31941020e-01 -4.02279824e-01
-1.17331338e+00 5.72773933e-01 -3.51402879e-01 -7.59457797e-02
1.15798473e+00 -5.12381971e-01 1.68780416e-01 -3.45618337e-01
-7.73659885e-01 -4.11000587e-02 2.58533001e-01 7.66500950e-01
1.00154960e+00 -1.60414064e+00 -3.58751923e-01 3.43159556e-01
8.76169130e-02 5.27528822e-02 3.11500579e-01 8.06790948e-01
-7.07863212e-01 5.56604505e-01 -8.36963892e-01 -6.93963945e-01
-1.23999333e+00 4.23973024e-01 3.62020433e-01 -3.22357953e-01
-7.80341506e-01 6.98706150e-01 1.17411390e-02 -4.56997484e-01
4.81485605e-01 -5.49937606e-01 -4.23565924e-01 -1.29233226e-01
3.82990390e-01 5.18645585e-01 -1.75767183e-01 -9.67864871e-01
-4.34795350e-01 1.02949107e+00 3.39098990e-01 4.09760326e-01
1.79237783e+00 1.87401444e-01 9.97485146e-02 2.41277218e-01
1.41005778e+00 -8.02424192e-01 -1.81569266e+00 -2.24642828e-01
5.55714825e-03 -7.54888296e-01 -8.98620114e-02 -2.93639362e-01
-1.55036044e+00 1.15198088e+00 5.89439631e-01 -2.19203785e-01
1.27453792e+00 -2.35770926e-01 1.20358264e+00 5.23588061e-01
3.73352855e-01 -1.23864174e+00 8.76806855e-01 5.76713085e-01
1.05234492e+00 -9.38844979e-01 8.56605545e-02 2.14645583e-02
-5.52729011e-01 1.63807654e+00 7.59457946e-01 -1.16700917e-01
6.61619008e-01 -2.72178531e-01 -1.33108839e-01 -4.87862647e-01
-3.15741748e-01 -4.24220353e-01 6.58686697e-01 6.99841678e-01
1.71277449e-01 -2.61787683e-01 -1.71719030e-01 5.08547187e-01
3.45322102e-01 3.51244330e-01 8.41604024e-02 1.26157498e+00
-3.66777331e-01 -1.01718485e+00 -7.42428079e-02 2.00132027e-01
-1.60772964e-01 6.94074810e-01 -6.45274878e-01 8.95386159e-01
3.03937614e-01 5.74909210e-01 -2.62450457e-01 -8.68897021e-01
5.32701552e-01 -1.91166869e-03 6.68956816e-01 -5.55076897e-01
-1.13075607e-01 -1.64094523e-01 -1.49832442e-01 -1.18597186e+00
-1.19892514e+00 -6.97790384e-01 -1.02982950e+00 1.59308106e-01
-1.10635065e-01 -5.40178835e-01 5.14234126e-01 1.06547868e+00
2.90840030e-01 5.66355348e-01 7.36398458e-01 -1.37685549e+00
-7.22431779e-01 -8.22958529e-01 -7.13465214e-01 5.72755635e-01
2.08557412e-01 -1.14813852e+00 -2.06930876e-01 2.19700068e-01] | [7.839609146118164, 0.377939373254776] |
1da21ea4-5a54-4dba-8d0e-75e53caf23e9 | video-relation-detection-via-tracklet-based | 2108.08669 | null | https://arxiv.org/abs/2108.08669v1 | https://arxiv.org/pdf/2108.08669v1.pdf | Video Relation Detection via Tracklet based Visual Transformer | Video Visual Relation Detection (VidVRD), has received significant attention of our community over recent years. In this paper, we apply the state-of-the-art video object tracklet detection pipeline MEGA and deepSORT to generate tracklet proposals. Then we perform VidVRD in a tracklet-based manner without any pre-cutting operations. Specifically, we design a tracklet-based visual Transformer. It contains a temporal-aware decoder which performs feature interactions between the tracklets and learnable predicate query embeddings, and finally predicts the relations. Experimental results strongly demonstrate the superiority of our method, which outperforms other methods by a large margin on the Video Relation Understanding (VRU) Grand Challenge in ACM Multimedia 2021. Codes are released at https://github.com/Dawn-LX/VidVRD-tracklets. | ['Jun Xiao', 'Yifeng Huang', 'Long Chen', 'Kaifeng Gao'] | 2021-08-19 | null | null | null | null | ['video-visual-relation-detection'] | ['computer-vision'] | [-2.29647189e-01 -1.23407096e-02 -5.71907103e-01 -1.64996028e-01
-8.98444295e-01 -4.61283922e-01 7.35297501e-01 1.01166531e-01
-1.44473597e-01 2.88680941e-01 4.22389358e-01 -3.60560954e-01
1.25059724e-01 -5.04711151e-01 -1.10162890e+00 -1.29467845e-01
-7.99442232e-02 5.74835181e-01 8.82602453e-01 -1.30643705e-02
-1.46834895e-01 1.01122536e-01 -1.38956285e+00 6.03954613e-01
3.94898027e-01 1.11698282e+00 1.71981946e-01 7.25933671e-01
8.64249319e-02 1.36617374e+00 -3.91720124e-02 -5.67552745e-01
1.52238831e-01 -2.20803291e-01 -6.85084581e-01 -1.14917971e-01
6.93757474e-01 -3.68969828e-01 -1.24792743e+00 7.93540299e-01
3.16932887e-01 3.47890481e-02 3.23987991e-01 -1.58851969e+00
-8.92656922e-01 7.76113808e-01 -7.86070466e-01 6.78465605e-01
6.63106322e-01 1.64726079e-01 1.26563680e+00 -1.14583182e+00
9.69655395e-01 1.47229183e+00 4.30072993e-01 2.68124878e-01
-8.65375638e-01 -8.34093392e-01 3.88353437e-01 9.09230769e-01
-1.65351629e+00 -5.65642715e-01 5.04033923e-01 -5.91564775e-01
1.00013828e+00 2.55543321e-01 7.96222985e-01 1.04927564e+00
6.60796696e-03 1.30089223e+00 4.66038555e-01 -1.09233007e-01
-1.75029039e-01 -1.11444242e-01 1.75294861e-01 9.44153905e-01
1.47214055e-01 -3.75975366e-03 -9.59303916e-01 9.73187685e-02
7.83649921e-01 -1.72660306e-01 -5.27407050e-01 -6.16051197e-01
-1.34263813e+00 3.71820331e-01 3.52704912e-01 4.86483537e-02
-2.57150769e-01 5.10870159e-01 5.53820074e-01 2.35135019e-01
5.81506371e-01 -2.48954684e-01 -2.79794693e-01 -2.36497283e-01
-6.33675456e-01 1.93990678e-01 6.18534029e-01 1.53378403e+00
3.46421689e-01 -3.77636373e-01 -7.61661053e-01 4.79950607e-01
6.13437355e-01 3.10355604e-01 -9.85821560e-02 -8.79168332e-01
6.82230294e-01 5.44757426e-01 -7.56832361e-02 -1.03898573e+00
7.13450760e-02 -1.82309553e-01 -3.88590753e-01 -3.05029213e-01
4.57290858e-02 4.15649593e-01 -8.98852706e-01 1.30987918e+00
6.62490308e-01 7.39224613e-01 -1.28276482e-01 1.20061016e+00
1.30746686e+00 8.72141898e-01 2.89730877e-01 -1.11681260e-01
1.64707804e+00 -1.34759927e+00 -8.46870840e-01 -3.92573923e-02
5.88897645e-01 -7.74164021e-01 9.61347997e-01 1.26157910e-01
-1.13689971e+00 -6.75770164e-01 -9.14417744e-01 -4.40803558e-01
-3.17688938e-03 8.82536173e-02 5.14688551e-01 8.33394974e-02
-1.02715802e+00 4.71526384e-02 -1.18511808e+00 -5.91525972e-01
8.98935139e-01 7.75664076e-02 -3.29937130e-01 -1.89612821e-01
-9.70558405e-01 5.52689612e-01 6.12034500e-01 1.68653637e-01
-1.21950793e+00 -1.00088954e+00 -9.86661553e-01 -5.66020980e-02
8.32858205e-01 -8.76060605e-01 1.40556872e+00 -4.92852777e-01
-1.01079988e+00 8.71320426e-01 -3.75210673e-01 -4.69902962e-01
6.78816795e-01 -6.22880161e-01 -4.85824943e-01 3.49472314e-01
9.34355706e-02 5.96560240e-01 5.55251300e-01 -1.19176579e+00
-8.73643994e-01 9.48593393e-02 1.23931229e-01 3.13938469e-01
-1.67250693e-01 3.57571632e-01 -1.54639387e+00 -5.50016999e-01
-6.73782304e-02 -7.84727633e-01 2.09711432e-01 3.95464599e-01
-6.13106549e-01 -7.20970988e-01 1.14726353e+00 -6.19006336e-01
1.32940888e+00 -2.30752492e+00 1.99742150e-02 -1.30936623e-01
6.11586809e-01 3.99749368e-01 -1.51797459e-01 4.97186333e-01
-9.88794789e-02 -1.37185350e-01 2.48458996e-01 -4.47932988e-01
1.49982423e-01 1.54393330e-01 -3.47703338e-01 5.75145185e-01
2.75036931e-01 1.24790418e+00 -1.06620443e+00 -8.24716687e-01
1.64160848e-01 5.81956804e-01 -4.15369362e-01 3.78087431e-01
-6.93962395e-01 1.88241884e-01 -7.70161152e-01 7.75983810e-01
4.55280066e-01 -5.87709963e-01 2.80680686e-01 -7.16498137e-01
-2.34167740e-01 4.15448964e-01 -9.15549934e-01 1.88243616e+00
6.75644428e-02 9.78767276e-01 -5.04452229e-01 -7.81757653e-01
5.84508240e-01 3.28391194e-01 5.66411138e-01 -8.72723222e-01
1.28314927e-01 -2.38813043e-01 -3.09208483e-01 -6.55928791e-01
5.26752710e-01 5.49923897e-01 2.59372860e-01 1.47662446e-01
1.05539352e-01 6.78087592e-01 3.19010556e-01 7.28146076e-01
1.27876163e+00 7.21478641e-01 1.38507947e-01 7.79749826e-02
4.61074919e-01 1.82535470e-01 7.13315189e-01 5.29264271e-01
-3.46615285e-01 5.92407167e-01 6.86044097e-01 -3.91525000e-01
-9.55698848e-01 -1.18932211e+00 2.20012963e-01 9.63390231e-01
6.83965445e-01 -1.05325508e+00 -2.42752984e-01 -9.00304139e-01
5.74242771e-02 5.17574608e-01 -7.22412646e-01 -1.20894969e-01
-8.25637221e-01 -1.38543636e-01 4.51894879e-01 7.21270204e-01
4.68303740e-01 -7.94199049e-01 -3.07993829e-01 -3.82494815e-02
-3.50345403e-01 -1.62282693e+00 -7.92810082e-01 -3.05618465e-01
-3.21620256e-01 -1.17397559e+00 -3.23486000e-01 -9.29177046e-01
4.00514752e-01 6.03333533e-01 1.22387993e+00 1.14767477e-01
-1.80281341e-01 5.47249734e-01 -5.63710749e-01 -2.31667727e-01
-2.25025564e-01 -8.35533664e-02 -2.38049120e-01 1.31493241e-01
4.59939808e-01 -2.88291901e-01 -7.31408477e-01 4.60555732e-01
-5.43437302e-01 3.89816165e-01 3.79983783e-01 3.28477710e-01
1.15783620e+00 -1.30969167e-01 8.80660713e-02 -8.69988680e-01
3.62775326e-02 -5.87845623e-01 -8.28861594e-01 5.84220648e-01
-2.68816352e-01 -2.14257166e-01 1.23140872e-01 -3.29502046e-01
-9.50791180e-01 -4.40586656e-02 1.79266095e-01 -1.15264583e+00
1.22506104e-01 4.27317560e-01 -3.03282231e-01 8.88893530e-02
3.19621801e-01 1.82110906e-01 -4.16849136e-01 -3.16753894e-01
6.47985339e-01 2.43579537e-01 6.96894944e-01 -4.06864971e-01
1.19660103e+00 5.46084046e-01 -3.04778099e-01 -4.33169901e-01
-1.03398073e+00 -7.49002695e-01 -4.30423766e-01 -5.40656686e-01
1.18161511e+00 -1.39978981e+00 -9.10947502e-01 5.22310100e-02
-1.24337804e+00 -5.97589791e-01 -2.22024322e-01 4.65417266e-01
-4.92440104e-01 2.95046568e-01 -6.16914868e-01 -4.66572165e-01
-3.15249801e-01 -8.83207738e-01 1.14412212e+00 3.10930043e-01
8.92551094e-02 -7.03982830e-01 1.70424562e-02 6.23511255e-01
-1.46504000e-01 1.45689592e-01 3.97334456e-01 -5.49057484e-01
-1.58620083e+00 1.58064783e-01 -6.44799054e-01 -1.09662957e-01
-1.10432394e-01 1.07974276e-01 -8.62402976e-01 -1.89140499e-01
-7.58085668e-01 -2.82323211e-01 9.86156344e-01 3.01342905e-01
1.12057376e+00 -7.38612115e-02 -8.25160980e-01 8.54417026e-01
1.32417011e+00 3.54846418e-01 7.16176808e-01 3.72273475e-01
1.23761964e+00 1.13772646e-01 1.12400973e+00 3.26454580e-01
1.10718155e+00 9.17655528e-01 4.53902751e-01 -2.73498264e-03
-7.06142485e-01 -5.26939154e-01 3.20592463e-01 8.66403461e-01
-5.80968037e-02 -6.48432553e-01 -9.77747798e-01 6.51364088e-01
-2.49341941e+00 -1.08750367e+00 -3.92634243e-01 1.78124881e+00
6.17523789e-01 2.33014435e-01 1.97249621e-01 -2.66448021e-01
7.78415501e-01 2.90233880e-01 -4.27411318e-01 1.97922975e-01
-7.23323300e-02 -1.56927407e-01 4.34542358e-01 3.83667678e-01
-1.30614591e+00 1.39970160e+00 4.82411385e+00 6.46763682e-01
-6.69365764e-01 4.00644720e-01 4.03494865e-01 8.71933848e-02
-2.77944475e-01 2.35115439e-01 -8.50966632e-01 1.23322614e-01
6.78898156e-01 -3.53729457e-01 1.00560151e-01 6.04981661e-01
1.82369411e-01 7.45938905e-03 -1.23014235e+00 1.18457413e+00
1.80882812e-01 -1.69440603e+00 1.32462091e-03 -1.65459961e-01
3.85057032e-01 1.22258388e-01 -8.42787921e-02 3.95782948e-01
1.96423233e-01 -5.99188864e-01 1.09815931e+00 5.19172907e-01
8.09594274e-01 -4.50006723e-01 3.07132959e-01 -2.31224403e-01
-1.91364253e+00 2.23518595e-01 -2.42582619e-01 2.24835157e-01
4.07464236e-01 3.71223569e-01 -9.30064023e-01 7.82961845e-01
9.14089262e-01 1.31947947e+00 -4.82921749e-01 1.48867393e+00
-3.87065560e-01 8.89181972e-01 -3.44843268e-01 1.45590290e-01
-5.10872491e-02 1.60450608e-01 6.88872159e-01 1.25264609e+00
7.59438425e-02 3.46069634e-01 4.31873083e-01 7.03175545e-01
-2.92364150e-01 -9.55003574e-02 -3.47395808e-01 -2.47122906e-02
6.94202363e-01 1.22610879e+00 -6.98093712e-01 -4.39758480e-01
-7.56399274e-01 1.01100910e+00 2.82077670e-01 3.78108680e-01
-1.39589798e+00 2.05647960e-01 7.49334633e-01 2.40612447e-01
7.78308988e-01 -1.72326133e-01 3.32020670e-01 -1.32409668e+00
2.15609625e-01 -5.47468543e-01 7.49609351e-01 -1.05905628e+00
-1.05366004e+00 6.01924181e-01 1.33487806e-01 -1.41761041e+00
2.68613726e-01 -3.00868809e-01 -3.47706020e-01 3.96214068e-01
-1.67231309e+00 -1.46087265e+00 -5.24372041e-01 7.88551927e-01
5.50971746e-01 1.08319290e-01 3.14657032e-01 8.37041438e-01
-8.34570289e-01 6.68534696e-01 -4.85968530e-01 4.33149457e-01
7.51510859e-01 -1.00545847e+00 7.96418786e-01 9.67801332e-01
4.75049227e-01 2.59679556e-01 6.62341893e-01 -8.91144872e-01
-1.85989153e+00 -1.59587574e+00 9.24405873e-01 -6.70658171e-01
9.72369313e-01 -5.59258342e-01 -9.56246495e-01 1.13485742e+00
4.05750573e-01 6.52102292e-01 4.43902105e-01 -9.29835737e-02
-6.43048108e-01 -2.04023868e-01 -3.55656564e-01 5.97624183e-01
1.57963312e+00 -6.64795101e-01 -4.23693419e-01 5.21834970e-01
1.10343277e+00 -8.09572518e-01 -9.81177449e-01 4.52520400e-01
4.79061276e-01 -6.28531635e-01 1.08449495e+00 -5.11491060e-01
2.89170325e-01 -8.91742647e-01 -1.51187077e-01 -5.27729154e-01
-5.51635504e-01 -6.88677251e-01 -7.20441878e-01 1.46686804e+00
3.09040487e-01 -1.92720100e-01 8.09420347e-01 1.03317499e-01
-4.36891884e-01 -7.60332227e-01 -6.71457231e-01 -7.94864476e-01
-7.51625955e-01 -5.72597921e-01 2.01607317e-01 9.54065442e-01
-1.92691326e-01 4.65392411e-01 -4.04662400e-01 6.30813003e-01
6.49165988e-01 3.04640621e-01 9.49912608e-01 -7.23134518e-01
-3.94049466e-01 -1.84820518e-01 -8.31278086e-01 -1.43929577e+00
-5.81142642e-02 -8.62190485e-01 2.78871898e-02 -1.90925384e+00
3.42353612e-01 -3.84102881e-01 -4.31654185e-01 6.92118466e-01
-3.02216798e-01 3.62050295e-01 4.32380795e-01 2.95292586e-01
-1.40504324e+00 7.64086068e-01 1.21614897e+00 -2.88669288e-01
3.11974473e-02 -3.25210303e-01 -4.62187499e-01 4.36070293e-01
4.95684862e-01 -5.37661672e-01 -6.84086919e-01 -6.58477068e-01
2.12154686e-01 1.61051333e-01 5.60027659e-01 -9.35328722e-01
4.23421770e-01 2.33858358e-02 2.61349976e-01 -9.53656375e-01
4.92924541e-01 -5.98043203e-01 3.20944041e-01 2.14835122e-01
-2.95295238e-01 1.92722365e-01 2.60452449e-01 7.92607963e-01
-2.61738688e-01 6.11091495e-01 2.79825389e-01 3.28577131e-01
-1.24959886e+00 9.21190858e-01 -1.91885494e-02 2.76591808e-01
1.28531635e+00 -8.47092345e-02 -8.49981070e-01 -3.08627427e-01
-4.35631901e-01 6.57133579e-01 2.36782357e-01 7.88244307e-01
7.33524024e-01 -1.45691717e+00 -8.67578745e-01 -3.11965227e-01
5.94040275e-01 3.87679003e-02 3.11828762e-01 1.09252083e+00
-5.27353406e-01 2.02431634e-01 2.04631224e-01 -7.83374190e-01
-1.45531392e+00 7.90401280e-01 1.21941082e-01 3.57367471e-02
-1.16634166e+00 1.21424627e+00 5.63573718e-01 3.13085139e-01
3.93002659e-01 -2.69722432e-01 -1.40233666e-01 -1.17668040e-01
6.10660732e-01 -3.05406358e-02 -1.43287495e-01 -7.05450714e-01
-5.90955973e-01 4.75564927e-01 -2.85093725e-01 2.36968383e-01
1.15307748e+00 -1.26010835e-01 1.69565156e-01 2.63949722e-01
1.17071056e+00 -2.56966680e-01 -1.37823534e+00 -4.87753332e-01
-6.15776815e-02 -5.91375411e-01 3.25192548e-02 -5.29943585e-01
-1.28566647e+00 4.92297441e-01 5.08667886e-01 8.72630998e-02
1.11987138e+00 5.81771374e-01 9.19998586e-01 4.98447055e-03
2.01470464e-01 -5.74772358e-01 1.35991171e-01 2.56313533e-01
7.42342174e-01 -1.07816005e+00 8.32012519e-02 -9.61923838e-01
-5.68135440e-01 8.19266677e-01 8.57623816e-01 -2.38256883e-02
6.28047943e-01 2.83381790e-01 1.04247347e-01 -3.59538257e-01
-1.34441566e+00 -4.48847473e-01 5.42550147e-01 5.10466933e-01
3.61404479e-01 -1.36521459e-01 -1.86569825e-01 2.34035864e-01
3.54496360e-01 1.86903507e-01 1.21951401e-01 9.63080168e-01
-1.76163942e-01 -1.10355341e+00 -4.57909517e-02 2.05335170e-01
-1.29057437e-01 -2.22379982e-01 -1.16849519e-01 8.30813050e-01
1.17428437e-01 6.67891085e-01 1.21753648e-01 -6.23067915e-01
5.60301185e-01 -3.52538615e-01 4.83157396e-01 -3.68380815e-01
-2.76951522e-01 2.45017350e-01 5.36335170e-01 -1.01960874e+00
-5.44716358e-01 -9.11232233e-01 -1.48437881e+00 -1.67242959e-01
-2.26055697e-01 -2.05127448e-02 -6.06931299e-02 6.13178253e-01
6.57737970e-01 7.72606730e-01 2.54920185e-01 -4.05560404e-01
2.69708067e-01 -5.39006710e-01 2.22647414e-02 2.75182545e-01
9.25817192e-02 -8.55238676e-01 3.35906059e-01 4.59079415e-01] | [9.383708953857422, 0.7387280464172363] |
acf287cd-207f-4b68-9f29-9d5f8a2d5383 | breaking-the-representation-bottleneck-of | 2211.12781 | null | https://arxiv.org/abs/2211.12781v1 | https://arxiv.org/pdf/2211.12781v1.pdf | Breaking the Representation Bottleneck of Chinese Characters: Neural Machine Translation with Stroke Sequence Modeling | Existing research generally treats Chinese character as a minimum unit for representation. However, such Chinese character representation will suffer two bottlenecks: 1) Learning bottleneck, the learning cannot benefit from its rich internal features (e.g., radicals and strokes); and 2) Parameter bottleneck, each individual character has to be represented by a unique vector. In this paper, we introduce a novel representation method for Chinese characters to break the bottlenecks, namely StrokeNet, which represents a Chinese character by a Latinized stroke sequence (e.g., "ao1 (concave)" to "ajaie" and "tu1 (convex)" to "aeaqe"). Specifically, StrokeNet maps each stroke to a specific Latin character, thus allowing similar Chinese characters to have similar Latin representations. With the introduction of StrokeNet to neural machine translation (NMT), many powerful but not applicable techniques to non-Latin languages (e.g., shared subword vocabulary learning and ciphertext-based data augmentation) can now be perfectly implemented. Experiments on the widely-used NIST Chinese-English, WMT17 Chinese-English and IWSLT17 Japanese-English NMT tasks show that StrokeNet can provide a significant performance boost over the strong baselines with fewer model parameters, achieving 26.5 BLEU on the WMT17 Chinese-English task which is better than any previously reported results without using monolingual data. Code and scripts are freely available at https://github.com/zjwang21/StrokeNet. | ['Min Zhang', 'Xuebo Liu', 'Zhijun Wang'] | 2022-11-23 | null | null | null | null | ['nmt'] | ['computer-code'] | [ 2.36853436e-01 -3.51968586e-01 -4.82967824e-01 -1.72719121e-01
-8.21234643e-01 -6.96982145e-01 6.85330153e-01 -2.79029518e-01
-6.56032383e-01 7.09524632e-01 5.41652799e-01 -8.05858374e-01
6.54376209e-01 -6.60233855e-01 -7.34964311e-01 -7.15397656e-01
5.53261220e-01 4.12724495e-01 -2.72675931e-01 -2.98959792e-01
2.33579934e-01 3.27188492e-01 -7.17163742e-01 3.78279060e-01
1.07114732e+00 3.83909315e-01 4.51706231e-01 5.47826469e-01
-6.33216381e-01 3.84679079e-01 -7.27178276e-01 -3.57401878e-01
8.02922770e-02 -5.55876613e-01 -8.15271199e-01 -4.82194781e-01
4.28825915e-01 -4.47744757e-01 -4.29129779e-01 9.21873450e-01
5.57291865e-01 7.90187623e-03 7.63391376e-01 -6.56990469e-01
-1.26016176e+00 1.10992813e+00 -5.67911506e-01 -1.87226027e-01
2.27405000e-02 -7.39316642e-02 8.52369547e-01 -1.25544453e+00
7.08755970e-01 1.25891352e+00 4.58442450e-01 1.09130120e+00
-8.94421816e-01 -9.48441625e-01 2.74736941e-01 1.52777776e-01
-1.34361708e+00 -2.25966170e-01 4.67534006e-01 -1.39556170e-01
9.64440107e-01 4.36312377e-01 3.86661559e-01 1.36924958e+00
-1.12476796e-01 1.19715273e+00 1.03154469e+00 -5.99211812e-01
-1.88654855e-01 8.66737068e-02 2.65735358e-01 4.41545665e-01
1.29129216e-01 -1.53381512e-01 -1.87083319e-01 2.22374648e-01
7.92742968e-01 -1.40351027e-01 -1.00701883e-01 1.04189903e-01
-1.47285283e+00 8.48614931e-01 5.02021074e-01 4.54070032e-01
-8.36948752e-02 3.48589063e-01 4.69998151e-01 2.69229263e-01
2.33045593e-01 3.76116216e-01 -5.09223521e-01 -3.02201867e-01
-5.29813290e-01 -4.34115455e-02 5.31430364e-01 1.33280921e+00
6.98254704e-01 5.00655890e-01 -3.08045208e-01 1.07011271e+00
2.46835090e-02 9.38902259e-01 8.27320278e-01 -4.77120310e-01
7.93043494e-01 3.05608898e-01 -1.61353484e-01 -3.91586661e-01
-1.36184588e-01 -2.97983110e-01 -1.19851530e+00 -1.54973626e-01
3.25484246e-01 -4.59737211e-01 -1.08352196e+00 1.87514496e+00
-3.12531143e-01 -2.03232691e-02 3.16232711e-01 7.00806320e-01
1.01215923e+00 1.12299287e+00 8.20699483e-02 8.79340172e-02
1.37502503e+00 -1.25692534e+00 -7.19148993e-01 -3.59421998e-01
7.46619880e-01 -1.16936517e+00 1.55160534e+00 1.30168080e-01
-8.83049667e-01 -7.73358285e-01 -1.04254329e+00 -2.59540230e-01
-6.02903605e-01 6.20375037e-01 5.40998399e-01 6.95378423e-01
-7.57832050e-01 3.69746119e-01 -5.91414750e-01 -4.03809845e-01
2.11185917e-01 1.86782911e-01 -2.98747122e-01 -1.83758542e-01
-1.30597007e+00 1.02813995e+00 4.63894874e-01 3.37330252e-02
-7.30253160e-01 -5.90095103e-01 -7.75462270e-01 -5.86695224e-02
1.52108505e-01 -2.95226216e-01 1.17154622e+00 -9.72378492e-01
-1.71214616e+00 5.47573805e-01 -4.07179028e-01 -1.65497631e-01
4.57172751e-01 -6.94722295e-01 -4.65147555e-01 -1.11740440e-01
-1.37366131e-01 8.70873690e-01 4.57741797e-01 -1.15214157e+00
-3.24087322e-01 6.83943704e-02 -4.07580197e-01 3.21186125e-01
-7.54737079e-01 4.37543392e-01 -8.06866646e-01 -1.20057476e+00
-5.91549724e-02 -1.12708759e+00 -2.03112751e-01 -2.92074978e-01
-5.75542331e-01 -3.17874849e-01 7.37428188e-01 -9.80559528e-01
1.42993140e+00 -1.96995199e+00 1.75957158e-01 3.30735482e-02
-2.32798055e-01 7.36322761e-01 -5.04086733e-01 4.75296855e-01
3.83969247e-02 5.44709623e-01 -3.85933995e-01 -2.71406084e-01
-1.38625130e-01 3.01600963e-01 -3.61813545e-01 2.05416977e-01
2.55260885e-01 1.19116008e+00 -7.79329062e-01 -1.99516073e-01
1.59210235e-01 4.90568906e-01 -4.60255861e-01 1.41698286e-01
1.17755989e-02 3.79522443e-01 -4.33307379e-01 4.99211103e-01
7.89991736e-01 1.94825903e-01 1.82125986e-01 -5.58965281e-03
-4.72496659e-01 4.42170709e-01 -9.05557215e-01 1.74361420e+00
-5.18935263e-01 6.93067670e-01 -3.73982757e-01 -6.09277964e-01
1.13692558e+00 3.49635124e-01 -1.74652860e-01 -6.17085636e-01
1.54181495e-01 4.35806483e-01 8.43442827e-02 -1.22209534e-01
6.34009778e-01 2.03339607e-01 -4.21171933e-01 4.62673604e-01
-1.38111740e-01 -1.71987042e-01 1.72517180e-01 1.62542641e-01
5.46930373e-01 3.19274515e-01 1.41961679e-01 -4.06539947e-01
5.70951283e-01 -3.22287194e-02 7.27249026e-01 6.73416138e-01
1.74829513e-01 8.26883376e-01 2.78492153e-01 -2.81461239e-01
-1.42505431e+00 -1.08646238e+00 9.81823914e-03 1.01062417e+00
-5.55475205e-02 -4.73313242e-01 -9.24064517e-01 -3.74536365e-01
-2.43362010e-01 8.62691879e-01 -4.65020448e-01 -1.50562003e-01
-1.27398598e+00 -6.30928874e-01 9.83970284e-01 9.10784960e-01
6.30293727e-01 -1.41075420e+00 -1.08485632e-01 2.78891504e-01
-2.24639669e-01 -1.06370568e+00 -1.13037086e+00 6.92607462e-02
-7.40962207e-01 -5.25378883e-01 -1.22008979e+00 -1.22485292e+00
6.31177783e-01 3.37846935e-01 8.74830306e-01 1.83503345e-01
-1.79787904e-01 -1.44678563e-01 -5.64677417e-01 -4.07901347e-01
-5.50156355e-01 3.62347513e-01 8.39420706e-02 -3.02839428e-01
4.87167895e-01 1.09712714e-02 -4.78241742e-01 1.02841698e-01
-6.28169537e-01 4.56414789e-01 9.74233568e-01 9.05219972e-01
4.72002119e-01 -8.64674926e-01 4.02173221e-01 -8.66218328e-01
7.38596320e-01 -2.37371191e-01 -2.39448741e-01 5.68873882e-01
-4.53184336e-01 5.67696244e-02 9.75767851e-01 -9.11372066e-01
-1.05982196e+00 -1.20348096e-01 -2.09852904e-01 -1.18128784e-01
-3.85963470e-02 5.27378619e-01 -3.93675953e-01 2.70378858e-01
4.02097523e-01 6.61007524e-01 -1.04080297e-01 -7.62045741e-01
7.52708375e-01 9.13832247e-01 6.88035607e-01 -8.57944548e-01
8.11500490e-01 -8.85977119e-04 -4.90295202e-01 -9.95217681e-01
-2.22712159e-01 -1.79687858e-01 -7.03037143e-01 2.09119886e-01
1.09780395e+00 -9.88655329e-01 -3.66712630e-01 8.72409880e-01
-1.35906959e+00 -4.75424588e-01 6.35149255e-02 6.46324873e-01
-1.51776478e-01 6.05464399e-01 -1.16134119e+00 -3.91451061e-01
-7.44631648e-01 -1.13457072e+00 4.37928766e-01 4.57833052e-01
-2.52456814e-01 -6.33111954e-01 -6.58161864e-02 1.32094443e-01
6.50757253e-01 -1.30734026e-01 1.22751904e+00 -5.74614584e-01
-2.40531951e-01 -1.57898009e-01 -4.13491905e-01 6.92030907e-01
1.71605006e-01 1.43350109e-01 -6.04739547e-01 -4.49146301e-01
-4.64777470e-01 -3.31584513e-01 8.09812546e-01 1.54052749e-02
1.14033222e+00 -2.33207554e-01 -7.47915953e-02 8.73507857e-01
1.22082090e+00 4.49802458e-01 8.10798943e-01 3.64384890e-01
1.06303060e+00 2.03396231e-01 1.13369204e-01 7.74354786e-02
2.97017962e-01 6.09680057e-01 -1.15025416e-01 -4.67327446e-01
-6.49228215e-01 -4.72408086e-01 6.79109693e-01 1.52773392e+00
-2.10984007e-01 -3.01508605e-01 -9.11505938e-01 3.71811241e-01
-1.53204656e+00 -6.02983117e-01 -2.73260146e-01 2.18311405e+00
1.03889012e+00 7.02623054e-02 -2.79089719e-01 -2.67403483e-01
7.71443725e-01 1.78192705e-02 -5.14575362e-01 -8.89574468e-01
-6.59451246e-01 4.41492498e-01 5.07674396e-01 3.81465197e-01
-9.60072815e-01 1.67118335e+00 5.47976255e+00 9.99163508e-01
-1.15541828e+00 6.42667264e-02 4.69952762e-01 2.40116552e-01
-5.24107158e-01 1.18597373e-01 -9.86191690e-01 4.49385881e-01
7.57367730e-01 -1.44030184e-01 5.43984890e-01 5.43990791e-01
-4.31314148e-02 3.89205873e-01 -7.69215941e-01 8.71856570e-01
1.88071132e-01 -1.42047358e+00 5.99901915e-01 2.48724725e-02
9.13645267e-01 1.47180587e-01 1.05522908e-01 6.70092225e-01
4.53529418e-01 -1.14584374e+00 7.89166093e-01 2.68915981e-01
1.23996580e+00 -7.99818754e-01 4.55541998e-01 1.63359210e-01
-1.35625386e+00 2.52998978e-01 -6.70287132e-01 2.13432416e-01
1.09825712e-02 -1.13712236e-01 -4.46398497e-01 5.42208076e-01
3.57582152e-01 6.19500518e-01 -5.18132091e-01 7.91198552e-01
-6.95209980e-01 8.68563414e-01 -4.38003950e-02 -4.30403888e-01
5.99153638e-01 -5.82794011e-01 2.74974704e-01 1.74151206e+00
5.74339330e-01 -5.70317752e-05 6.97413608e-02 7.33405590e-01
-3.03726345e-01 5.57816803e-01 -4.23333973e-01 -1.30173221e-01
6.72759712e-01 9.55575526e-01 -4.73131716e-01 -6.06683135e-01
-6.64126873e-01 1.29457700e+00 4.56582636e-01 6.53102994e-01
-7.46388376e-01 -7.07266271e-01 8.39147508e-01 -4.42278087e-01
2.20623165e-01 -4.50564146e-01 -4.41277772e-01 -1.28329647e+00
-2.01434270e-01 -1.23135257e+00 7.16850013e-02 -6.44506454e-01
-1.04770875e+00 9.53463018e-01 -2.12472320e-01 -1.28652310e+00
-6.04316555e-02 -8.58972609e-01 -9.53473687e-01 1.24639559e+00
-1.19884145e+00 -1.30582833e+00 6.86449581e-04 6.13033414e-01
8.06714058e-01 -5.12874544e-01 1.06179929e+00 2.61738867e-01
-8.47346425e-01 1.13819897e+00 5.54782569e-01 6.62031710e-01
8.86923611e-01 -1.20861506e+00 9.83771443e-01 1.02518189e+00
2.98071116e-01 9.47107315e-01 2.57944286e-01 -7.30981767e-01
-1.46451461e+00 -9.53233480e-01 1.15461898e+00 -3.37434173e-01
5.66188753e-01 -6.12398565e-01 -1.04211080e+00 6.18930042e-01
4.48308021e-01 -4.40605611e-01 6.65114105e-01 3.07940063e-03
-5.60537696e-01 1.81927219e-01 -4.73567098e-01 1.18928182e+00
8.84979069e-01 -6.11323774e-01 -5.31198800e-01 7.86063597e-02
8.23952615e-01 -3.08466524e-01 -4.68626380e-01 3.23076338e-01
5.88231683e-01 -2.76664674e-01 8.87574971e-01 -8.15240383e-01
3.64118159e-01 -2.33635262e-01 -2.32224017e-01 -1.30442035e+00
-4.02663976e-01 -6.21746302e-01 2.73568705e-02 1.28788280e+00
6.84629679e-01 -4.53339934e-01 5.75250685e-01 2.11331367e-01
-4.51999545e-01 -6.42119706e-01 -8.25980604e-01 -9.24561501e-01
9.46473718e-01 -2.65824378e-01 6.93998396e-01 1.01235652e+00
-2.26218939e-01 3.04388940e-01 -8.23792398e-01 -3.05161089e-01
2.45025635e-01 -1.59587543e-02 7.18617916e-01 -7.88142204e-01
-2.17888243e-02 -9.32227075e-01 2.75233358e-01 -1.44472218e+00
6.84498250e-02 -1.26340985e+00 1.83731885e-04 -1.51803052e+00
4.53928381e-01 -4.03113246e-01 -3.40228140e-01 5.52200079e-01
-5.02888978e-01 2.17761502e-01 5.65292716e-01 2.54263431e-01
-5.66711910e-02 5.88681102e-01 1.36412561e+00 -1.72463700e-01
-1.05236806e-01 -2.08832368e-01 -6.56554401e-01 4.98389870e-01
1.14975691e+00 -3.29912812e-01 2.13435874e-03 -9.85041916e-01
-6.60503805e-02 -3.18125963e-01 -2.51105845e-01 -7.17253983e-01
1.69713259e-01 -1.20942838e-01 5.45297027e-01 -5.33809602e-01
2.07096264e-01 -2.14169323e-01 -2.04868808e-01 8.14140975e-01
-2.88325667e-01 4.45056200e-01 3.59597832e-01 4.27736714e-02
-5.40390760e-02 -4.17497635e-01 5.53743780e-01 -1.42021120e-01
-8.11694145e-01 3.14414501e-01 -7.09578872e-01 8.47478881e-02
6.10150158e-01 -1.90008670e-01 -5.26511848e-01 -2.09493741e-01
-1.87657952e-01 2.68716455e-01 3.56843621e-01 8.30472410e-01
6.51078045e-01 -1.55160701e+00 -1.22238934e+00 2.25782260e-01
7.40501285e-02 -3.08078706e-01 1.09513171e-01 4.30647969e-01
-8.29815149e-01 4.93909985e-01 -2.91538268e-01 -1.23263344e-01
-1.16634703e+00 4.39727485e-01 -8.70024599e-03 -1.91510588e-01
-6.13287270e-01 8.10850441e-01 2.39578441e-01 -6.59874439e-01
3.64165366e-01 -2.26874992e-01 -2.56694585e-01 -7.48065561e-02
6.44028664e-01 3.40700746e-01 -2.09172145e-01 -7.07135022e-01
-2.72405624e-01 8.97026777e-01 -4.29753482e-01 -7.24204555e-02
9.67004299e-01 2.03318954e-01 -1.32906064e-01 2.36304998e-01
1.27851796e+00 2.59473830e-01 -9.72335935e-01 -4.38039780e-01
1.37519568e-01 -1.26844242e-01 -3.38593960e-01 -8.12292457e-01
-8.49913001e-01 1.20873332e+00 5.74486852e-01 -6.23610318e-01
7.84186900e-01 -1.57313287e-01 1.04788256e+00 5.43712437e-01
1.20488137e-01 -1.21495283e+00 1.92354852e-03 1.02702677e+00
1.06619632e+00 -1.09875560e+00 -2.95895666e-01 -1.93129927e-01
-9.69361722e-01 1.34492075e+00 7.17046857e-01 -9.27504972e-02
2.77257174e-01 1.70840293e-01 5.71817219e-01 3.97542030e-01
-4.62642878e-01 -1.06762901e-01 4.96062249e-01 3.37927788e-01
8.02748561e-01 5.44297576e-01 -6.60132945e-01 6.45653486e-01
-2.91523248e-01 -5.25943756e-01 5.22865951e-01 6.60656810e-01
-2.78308541e-01 -1.61390460e+00 -2.73676783e-01 2.43333712e-01
-3.51759881e-01 -7.10582614e-01 -4.31923062e-01 9.42323506e-01
-1.23879528e-02 8.32204223e-01 1.28888220e-01 -4.11053896e-01
2.86549360e-01 1.23583607e-01 2.17523023e-01 -6.29675329e-01
-7.52082467e-01 3.91824603e-01 -9.35182497e-02 -7.29973540e-02
1.30186841e-01 -6.04231894e-01 -1.45315671e+00 -3.98063213e-01
-1.91917941e-02 4.80755568e-02 6.67693853e-01 5.72740376e-01
1.16019234e-01 4.39033300e-01 5.23111582e-01 -4.67746019e-01
-6.17923141e-01 -1.30573750e+00 -1.13927297e-01 4.23213005e-01
-1.38107926e-01 3.49193737e-02 -1.06644914e-01 -2.85786390e-01] | [10.45827579498291, 10.251154899597168] |
726c6a73-2b01-4be7-8385-9ac47f2eb3de | variational-inference-for-bayesian-neural | 2305.00934 | null | https://arxiv.org/abs/2305.00934v1 | https://arxiv.org/pdf/2305.00934v1.pdf | Variational Inference for Bayesian Neural Networks under Model and Parameter Uncertainty | Bayesian neural networks (BNNs) have recently regained a significant amount of attention in the deep learning community due to the development of scalable approximate Bayesian inference techniques. There are several advantages of using a Bayesian approach: Parameter and prediction uncertainties become easily available, facilitating rigorous statistical analysis. Furthermore, prior knowledge can be incorporated. However, so far, there have been no scalable techniques capable of combining both structural and parameter uncertainty. In this paper, we apply the concept of model uncertainty as a framework for structural learning in BNNs and hence make inference in the joint space of structures/models and parameters. Moreover, we suggest an adaptation of a scalable variational inference approach with reparametrization of marginal inclusion probabilities to incorporate the model space constraints. Experimental results on a range of benchmark datasets show that we obtain comparable accuracy results with the competing models, but based on methods that are much more sparse than ordinary BNNs. | ['Geir Storvik', 'Aliaksandr Hubin'] | 2023-05-01 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [-1.10913841e-02 3.62525731e-02 -1.46009356e-01 -6.98774099e-01
-7.63394177e-01 -1.92505240e-01 7.74962008e-01 -1.92699850e-01
-4.26928103e-01 1.15835416e+00 4.45714258e-02 -2.38716915e-01
-5.33823431e-01 -6.96551025e-01 -7.82067597e-01 -8.22025776e-01
1.01028524e-01 6.79020047e-01 3.84698540e-01 3.82891893e-01
2.10809663e-01 6.58390701e-01 -1.48564267e+00 -1.82008550e-01
7.23934293e-01 1.15825450e+00 6.80366531e-02 1.25353843e-01
-9.72787663e-02 6.12114549e-01 -2.57731259e-01 -5.40605128e-01
-2.05495078e-02 2.20687706e-02 -5.17064095e-01 -4.00208652e-01
5.55832803e-01 -7.56215036e-01 -5.12690306e-01 1.14705479e+00
3.91740471e-01 4.17306542e-01 1.02007961e+00 -1.12243164e+00
-2.52032697e-01 9.93037403e-01 -5.31981409e-01 1.70551956e-01
-1.68321714e-01 -2.50524253e-01 1.17876494e+00 -8.29541385e-01
2.99421191e-01 1.53193545e+00 8.74695659e-01 3.05897444e-01
-1.68579566e+00 -6.90869272e-01 1.37686223e-01 4.46649820e-01
-1.65886283e+00 -5.32331884e-01 7.78530240e-01 -2.44907230e-01
1.02628112e+00 -1.25833124e-01 4.34213966e-01 1.23921227e+00
2.38966540e-01 8.75522077e-01 8.95892441e-01 -3.50484043e-01
7.28071630e-01 1.35913193e-01 1.16254762e-01 5.26681483e-01
4.65012193e-01 2.34193787e-01 -5.27900279e-01 -3.46033216e-01
7.33771384e-01 4.96776700e-02 8.03678930e-02 -5.21091700e-01
-6.17967367e-01 1.11772811e+00 2.22193241e-01 -6.91523477e-02
-1.58267409e-01 7.12211192e-01 2.96790123e-01 -3.58819574e-01
5.21356761e-01 -2.64875628e-02 -4.47449416e-01 -1.10991485e-01
-1.32530630e+00 5.34182310e-01 1.00177205e+00 9.45182204e-01
7.01893508e-01 1.85650274e-01 9.33708996e-02 9.00212109e-01
9.48053479e-01 4.49582458e-01 -1.02512263e-01 -1.41179836e+00
2.18115419e-01 -9.00511071e-02 1.86970025e-01 -9.01042640e-01
-9.50826928e-02 -3.45622212e-01 -1.03275061e+00 1.14618324e-01
3.64800602e-01 -1.79435797e-02 -9.17646646e-01 1.87672675e+00
2.74547875e-01 5.31704664e-01 -1.96876526e-01 4.77814734e-01
5.29659212e-01 6.59782887e-01 -7.29001164e-02 -1.00883923e-01
1.03811264e+00 -5.00510037e-01 -8.04302990e-01 -1.14425197e-01
1.54523402e-01 -4.18802023e-01 4.05269384e-01 7.83581913e-01
-9.60259378e-01 -3.33727688e-01 -1.13549280e+00 -2.39897985e-02
-1.49866402e-01 -2.44788364e-01 8.35896909e-01 7.93462634e-01
-9.39765334e-01 8.50672603e-01 -1.45057666e+00 1.07016666e-02
5.62278688e-01 3.76174629e-01 -9.73944366e-02 -1.58267856e-01
-1.16517818e+00 9.99225259e-01 9.37090874e-01 3.98976833e-01
-1.02892041e+00 -7.07890749e-01 -9.08999443e-01 1.52797952e-01
4.51258272e-01 -5.81605673e-01 1.32224488e+00 -3.48098487e-01
-1.56910622e+00 1.40657678e-01 -2.59809822e-01 -6.74000084e-01
4.54602331e-01 -2.32186928e-01 -5.41017801e-02 5.51934913e-02
-3.53761226e-01 8.58577251e-01 8.51034880e-01 -1.09774494e+00
-5.37121892e-01 -1.92551583e-01 1.03801474e-01 -1.87059566e-01
-1.82689771e-01 -2.99315881e-02 -4.59624201e-01 -3.09421420e-01
3.18198353e-01 -8.24508607e-01 -2.52523869e-01 1.20739296e-01
-4.71633673e-01 -3.90910417e-01 5.89109719e-01 -6.80220425e-01
1.20362413e+00 -1.92553544e+00 2.18786791e-01 4.25841689e-01
2.12167844e-01 4.47598360e-02 2.25536242e-01 3.64994556e-01
2.14499503e-01 1.13496296e-01 -5.74680865e-01 -5.70784628e-01
4.76447970e-01 7.59593487e-01 -5.42873383e-01 3.00822467e-01
1.15038894e-01 7.00739741e-01 -5.63925028e-01 -6.54482186e-01
2.23046839e-01 7.61400342e-01 -7.72511244e-01 -8.57713744e-02
-3.06165218e-01 2.88526881e-02 -4.74440813e-01 4.13853019e-01
8.90399575e-01 -1.85604140e-01 3.69827718e-01 -2.74088293e-01
8.84253830e-02 3.25326473e-01 -1.53086722e+00 1.45802593e+00
-2.09065571e-01 5.58528960e-01 3.83181982e-02 -1.11994445e+00
7.55024850e-01 1.58722997e-01 2.20475256e-01 6.53296039e-02
1.34265870e-01 1.66682258e-01 5.06886542e-02 7.24840090e-02
2.75885999e-01 -2.08717406e-01 1.75125584e-01 3.89808744e-01
3.81519526e-01 -3.18149954e-01 3.38648021e-01 2.23174185e-01
8.26385438e-01 5.87758482e-01 2.23650455e-01 -3.26089621e-01
3.07689086e-02 -4.41922188e-01 8.27113152e-01 1.09830534e+00
-4.94542606e-02 4.43659604e-01 4.96449977e-01 -2.66948998e-01
-1.05774736e+00 -1.40403461e+00 -8.09045970e-01 7.53831029e-01
-3.83527428e-01 -3.15735370e-01 -6.92072093e-01 -1.99572638e-01
2.64454246e-01 8.42731237e-01 -4.91676211e-01 -1.71420425e-02
-2.94298500e-01 -1.04268920e+00 6.08992517e-01 9.40987408e-01
4.15368885e-01 -7.30952024e-01 -4.47727174e-01 3.90920967e-01
6.96144179e-02 -1.02793896e+00 1.14580959e-01 3.31652761e-01
-1.24091446e+00 -5.79026461e-01 -3.56476277e-01 -1.66865364e-01
2.59618074e-01 -3.86010796e-01 8.83530557e-01 -3.05155158e-01
-2.76545435e-01 2.39834040e-01 1.00745767e-01 -4.77167815e-01
-3.23737979e-01 -9.20474604e-02 3.76160085e-01 -2.54194647e-01
3.71617824e-01 -8.90253842e-01 -3.02917361e-01 1.13111891e-01
-9.16243017e-01 3.96921597e-02 5.45992494e-01 7.79425323e-01
5.79307497e-01 2.89631814e-01 5.70523858e-01 -6.76876187e-01
1.69149339e-01 -3.69843543e-01 -1.03050089e+00 1.92666888e-01
-8.91509771e-01 3.89536500e-01 5.24888597e-02 -3.29375386e-01
-1.59725833e+00 -1.87941000e-01 -1.61694035e-01 -3.90753686e-01
-2.59939253e-01 1.02429056e+00 -2.73050070e-01 2.35521868e-01
3.38832170e-01 -7.42580369e-02 -6.54678643e-02 -7.93327689e-01
3.08711171e-01 4.75177556e-01 3.52539092e-01 -1.00885129e+00
5.98736167e-01 4.23991263e-01 3.71248692e-01 -7.27900684e-01
-1.03839850e+00 -4.16353904e-02 -7.51538694e-01 3.64307351e-02
5.62942684e-01 -8.28420937e-01 -6.05961502e-01 5.76797426e-01
-1.13289046e+00 -1.46069452e-01 5.40857529e-03 8.36116016e-01
-4.41645205e-01 6.44109488e-01 -6.98564410e-01 -1.11143780e+00
1.01599619e-02 -1.24971652e+00 6.95604026e-01 2.12087974e-01
-1.43003866e-01 -1.19107890e+00 -1.12410538e-01 2.83369839e-01
3.68624300e-01 4.92690690e-02 9.90742385e-01 -7.01217353e-01
-8.70176196e-01 -1.92216933e-01 -4.68217462e-01 5.04508436e-01
-8.07010978e-02 3.53306621e-01 -1.12647688e+00 -1.74335986e-01
-4.15406190e-02 -2.44898781e-01 1.20703900e+00 8.76873732e-01
1.25220561e+00 -1.54880568e-01 -4.16335940e-01 4.91627872e-01
1.35815084e+00 -1.10896371e-01 5.26973188e-01 -7.09418803e-02
5.79210281e-01 5.17007411e-01 2.37390161e-01 6.04879260e-01
2.76530594e-01 6.07501268e-01 4.95919317e-01 6.36963546e-01
2.20662311e-01 -1.91878691e-01 8.81422684e-02 7.29153574e-01
-2.00411007e-01 -7.12283477e-02 -9.86439228e-01 3.84729207e-01
-1.99757993e+00 -9.71764207e-01 -2.19020452e-02 2.16065335e+00
1.23252106e+00 3.42732787e-01 -2.64236361e-01 -6.69051707e-02
5.78962386e-01 8.86258017e-03 -8.78358424e-01 -9.25089121e-02
8.36613327e-02 2.31446445e-01 4.10965711e-01 5.77460468e-01
-1.00305784e+00 6.03406668e-01 7.15994883e+00 1.12047613e+00
-4.62252259e-01 1.15986150e-02 5.38693130e-01 -2.44475603e-01
-3.71162921e-01 1.16946846e-01 -1.35771513e+00 3.57924670e-01
1.22106016e+00 1.21937796e-01 4.51930791e-01 8.66467476e-01
1.98841959e-01 -3.28649580e-01 -1.37869155e+00 7.79769897e-01
-2.68054545e-01 -1.37342465e+00 -3.75100337e-02 3.36081147e-01
7.91157663e-01 2.13897139e-01 -4.91060922e-03 3.67456287e-01
6.86959982e-01 -1.06563926e+00 7.58538485e-01 9.71831083e-01
4.87839490e-01 -1.01686728e+00 8.29818130e-01 4.79441494e-01
-7.63506234e-01 4.48509380e-02 -8.84070992e-01 -1.93905495e-02
2.27808446e-01 9.33931768e-01 -6.04271948e-01 3.95184815e-01
8.97559702e-01 7.29770303e-01 -1.87214911e-01 1.06043637e+00
-2.87257135e-01 8.96047056e-01 -8.85650396e-01 3.07590775e-02
9.48691666e-02 -3.99056226e-01 3.57201457e-01 9.87403631e-01
3.01977545e-01 -2.28433505e-01 -5.97770028e-02 1.42622221e+00
1.31667564e-02 -3.93403411e-01 -2.81406850e-01 -3.95584442e-02
8.58123302e-01 9.87383187e-01 -5.97969949e-01 -1.71548650e-01
-3.88575971e-01 1.49188831e-01 4.03288871e-01 5.00997782e-01
-8.05372298e-01 -1.28028184e-01 4.57521886e-01 -3.72320324e-01
7.18648612e-01 -3.26924056e-01 -1.73949540e-01 -1.02188766e+00
-1.50384694e-01 -5.82316816e-01 2.21505925e-01 -7.21197426e-01
-1.43677998e+00 4.97761704e-02 8.39495122e-01 -4.62592244e-01
-6.30308688e-01 -1.03320241e+00 -3.65360081e-01 9.66308117e-01
-1.35250437e+00 -9.79122221e-01 2.83756584e-01 1.40692428e-01
9.29771438e-02 -4.47505787e-02 7.84854412e-01 2.74169631e-02
-6.16475642e-01 4.86372143e-01 7.71033168e-01 -7.04430118e-02
4.60851640e-01 -1.35297394e+00 -2.39670137e-03 7.40774155e-01
9.30267572e-02 9.62610245e-01 9.70887303e-01 -6.33854866e-01
-1.21475816e+00 -7.81618834e-01 4.62097168e-01 -5.32368243e-01
8.40458632e-01 -4.17736262e-01 -1.03797281e+00 8.63014817e-01
-7.43125379e-02 -4.30123173e-02 7.29047000e-01 6.39744401e-01
-5.84408164e-01 -2.10236073e-01 -1.14561594e+00 4.17608500e-01
7.83929169e-01 -5.17779171e-01 -7.94175565e-01 -8.58968794e-02
5.08982062e-01 -1.98544204e-01 -1.04588485e+00 6.61413729e-01
8.66349518e-01 -9.50205028e-01 1.07835793e+00 -3.86400342e-01
3.11176062e-01 -1.65760159e-01 -6.21833444e-01 -1.01491225e+00
-1.46823645e-01 -3.26360852e-01 -6.42972410e-01 1.37282693e+00
2.98556864e-01 -6.30152702e-01 7.78604686e-01 1.10238719e+00
-1.52784437e-01 -6.93308115e-01 -1.30480599e+00 -7.94334650e-01
2.88329452e-01 -9.80342805e-01 5.67022204e-01 4.52216864e-01
-2.96162039e-01 3.94072048e-02 -3.60671163e-01 1.19921677e-01
1.17375064e+00 -1.19560085e-01 3.39310497e-01 -1.72793579e+00
-6.25478029e-01 -5.16700506e-01 -3.97818595e-01 -1.09235203e+00
5.04122794e-01 -7.50500619e-01 1.89714313e-01 -1.45306122e+00
5.56327462e-01 -2.17559755e-01 -5.45097172e-01 4.75778311e-01
8.50948393e-02 9.81717929e-02 -3.22945654e-01 1.79187790e-01
-4.29011703e-01 7.37956524e-01 5.16757190e-01 3.89217623e-02
2.84137517e-01 -4.84239645e-02 -4.88850474e-01 1.08860457e+00
6.33363008e-01 -6.00814044e-01 -5.55546165e-01 -3.45223427e-01
4.26936507e-01 -1.99509040e-01 5.92281818e-01 -8.72979999e-01
2.16585100e-01 -1.80804864e-01 4.73201871e-01 -1.04557335e+00
6.51082516e-01 -6.95845366e-01 3.47272813e-01 1.72953516e-01
-3.05614740e-01 -5.78068137e-01 3.36141169e-01 1.07073522e+00
-7.94910863e-02 -6.96871340e-01 7.97605991e-01 1.56044617e-01
-4.72958654e-01 3.46565068e-01 -4.35398012e-01 -1.95781872e-01
4.17530090e-01 -1.56593949e-01 9.64243561e-02 -2.13769346e-01
-7.13932395e-01 3.19817252e-02 8.35573524e-02 -7.22206160e-02
5.88689744e-01 -1.28638589e+00 -5.87126791e-01 -7.60340691e-02
-3.05273801e-01 3.10415417e-01 3.14018220e-01 7.61225164e-01
-2.89855927e-01 6.18051887e-01 6.67735413e-02 -7.53695369e-01
-7.65563726e-01 1.93083078e-01 2.60314494e-01 -1.90012589e-01
-4.71111268e-01 8.69310141e-01 6.61614910e-02 -6.35315359e-01
5.08351088e-01 -4.68115032e-01 7.30262175e-02 8.73849988e-02
5.54057419e-01 6.19941354e-01 -8.46640617e-02 -2.60418981e-01
-2.49824315e-01 6.59159273e-02 -2.83095777e-01 -3.79635692e-01
1.55492759e+00 -2.08976403e-01 -2.41794780e-01 6.55873954e-01
7.64153779e-01 -4.70249057e-01 -1.70319438e+00 -6.28535390e-01
2.50913143e-01 -2.43681967e-01 6.31297469e-01 -7.76865721e-01
-7.58910477e-01 1.13729942e+00 4.80479240e-01 5.41236736e-02
6.89075112e-01 4.58792336e-02 3.26969385e-01 8.77960205e-01
4.05882359e-01 -1.17046034e+00 -3.25385451e-01 6.92289829e-01
6.51348650e-01 -1.06747794e+00 2.59232283e-01 1.40488110e-02
-4.48915474e-02 1.07745910e+00 3.29057544e-01 8.52058642e-03
1.11219728e+00 3.59762341e-01 -4.94437426e-01 4.97698635e-02
-8.23720932e-01 2.67252564e-01 4.24621671e-01 5.59876144e-01
3.23817790e-01 -9.01966318e-02 1.31134972e-01 7.70862699e-01
-2.35390346e-02 1.73739046e-02 2.60837853e-01 6.63674593e-01
-6.32292628e-01 -1.05296779e+00 -2.46586114e-01 6.21518075e-01
-5.81433237e-01 -3.31554741e-01 3.22251260e-01 7.21872211e-01
-3.57312292e-01 8.36336076e-01 -9.99385142e-04 1.63747996e-01
-1.74252003e-01 2.66718358e-01 7.04292893e-01 -5.08387983e-01
4.59583074e-01 2.34300494e-01 7.01241642e-02 -4.51340884e-01
-4.63016927e-01 -1.07286227e+00 -1.09541893e+00 -4.10975367e-01
-5.57034731e-01 1.06451884e-01 9.22226548e-01 1.43966854e+00
1.92754626e-01 3.61567914e-01 4.33549359e-02 -1.11512506e+00
-1.20330489e+00 -1.15709245e+00 -9.47030604e-01 -3.34407419e-01
2.98711602e-02 -1.02121782e+00 -5.21383286e-01 -1.32125705e-01] | [7.1942949295043945, 3.8685572147369385] |
d0305c51-3ce9-4d06-bd10-489a37e39903 | 3d-medical-point-transformer-introducing | 2112.04863 | null | https://arxiv.org/abs/2112.04863v2 | https://arxiv.org/pdf/2112.04863v2.pdf | 3D Medical Point Transformer: Introducing Convolution to Attention Networks for Medical Point Cloud Analysis | General point clouds have been increasingly investigated for different tasks, and recently Transformer-based networks are proposed for point cloud analysis. However, there are barely related works for medical point clouds, which are important for disease detection and treatment. In this work, we propose an attention-based model specifically for medical point clouds, namely 3D medical point Transformer (3DMedPT), to examine the complex biological structures. By augmenting contextual information and summarizing local responses at query, our attention module can capture both local context and global content feature interactions. However, the insufficient training samples of medical data may lead to poor feature learning, so we apply position embeddings to learn accurate local geometry and Multi-Graph Reasoning (MGR) to examine global knowledge propagation over channel graphs to enrich feature representations. Experiments conducted on IntrA dataset proves the superiority of 3DMedPT, where we achieve the best classification and segmentation results. Furthermore, the promising generalization ability of our method is validated on general 3D point cloud benchmarks: ModelNet40 and ShapeNetPart. Code is released. | ['Weidong Cai', 'Dongnan Liu', 'Tiange Xiang', 'Yang song', 'Dingxin Zhang', 'Heng Wang', 'Chaoyi Zhang', 'Jianhui Yu'] | 2021-12-09 | null | null | null | null | ['3d-part-segmentation'] | ['computer-vision'] | [-1.65114179e-01 5.95280454e-02 -1.18471779e-01 -2.84259290e-01
-5.65962017e-01 -1.90832898e-01 2.73320884e-01 5.95884204e-01
6.33272007e-02 4.44859684e-01 1.32148638e-01 -1.58058345e-01
-4.12145406e-01 -9.71576333e-01 -8.60366464e-01 -8.07385027e-01
-3.26593965e-01 5.54753721e-01 2.96575457e-01 -3.39863211e-01
1.00361764e-01 9.36093986e-01 -1.08079875e+00 3.24570447e-01
8.47129285e-01 9.56455529e-01 2.25061908e-01 4.28880721e-01
-4.73477334e-01 2.37625733e-01 -3.68940115e-01 -1.95417017e-01
-3.37059312e-02 2.14601114e-01 -7.83667326e-01 -9.55247283e-02
2.08618134e-01 -4.64124642e-02 -3.37579489e-01 1.07774913e+00
8.35493207e-01 -2.15961173e-01 5.00542939e-01 -1.27662396e+00
-9.46204841e-01 2.43400022e-01 -5.34699321e-01 4.29882795e-01
1.89104244e-01 2.13983372e-01 1.09255791e+00 -9.01728749e-01
6.51920080e-01 1.40763152e+00 6.84406579e-01 5.98732233e-01
-8.49872887e-01 -6.52545333e-01 3.34260494e-01 1.40546843e-01
-1.36932826e+00 6.00646846e-02 8.32128942e-01 -1.09787643e-01
1.06813300e+00 4.13764596e-01 1.00044775e+00 1.19530332e+00
4.78515744e-01 1.04003441e+00 8.02597046e-01 1.64491490e-01
-3.43055576e-02 -1.84888199e-01 2.55465239e-01 7.83764064e-01
2.36098245e-01 2.37026084e-02 -3.36228192e-01 -4.15778786e-01
1.21116889e+00 5.88862717e-01 -5.34251630e-01 -2.66658843e-01
-1.24508679e+00 6.56815529e-01 1.25876784e+00 5.56643844e-01
-5.34943104e-01 3.65535885e-01 3.96333009e-01 9.07327011e-02
7.21528411e-01 4.35640216e-01 -6.10564590e-01 3.40307355e-01
-3.66484523e-01 3.88664424e-01 1.96918681e-01 1.10422170e+00
5.28502464e-01 -3.33586633e-01 -3.74266982e-01 7.74794221e-01
5.10557532e-01 4.54640478e-01 2.58195609e-01 -5.47676802e-01
2.84639776e-01 1.19828105e+00 -4.04886872e-01 -1.36044264e+00
-6.17862225e-01 -8.23591948e-01 -1.07354224e+00 -1.73920751e-01
5.33834435e-02 3.90270740e-01 -9.65165854e-01 1.30441642e+00
6.18796349e-01 6.35946572e-01 -2.23858714e-01 1.13053036e+00
1.46543980e+00 4.43610370e-01 2.04333737e-01 1.53720438e-01
1.55315030e+00 -5.98951936e-01 -5.42012811e-01 3.78753841e-02
8.21583807e-01 -3.12735975e-01 8.69161129e-01 1.24974303e-01
-1.00413024e+00 -4.62307394e-01 -6.21737599e-01 -2.92402238e-01
-5.74198782e-01 -3.14603299e-01 8.76988530e-01 8.60341191e-02
-1.03882337e+00 5.86569488e-01 -9.84122634e-01 -3.23830903e-01
1.20358145e+00 6.37555718e-01 -2.24229589e-01 -4.04952347e-01
-1.14490569e+00 4.45442796e-01 -1.97782949e-01 1.02512091e-01
-6.54009163e-01 -1.11640918e+00 -6.20390773e-01 5.90557307e-02
-7.88591579e-02 -1.19530845e+00 7.04550505e-01 -3.12337130e-01
-1.08458507e+00 9.77238894e-01 -1.43020675e-01 -1.55024603e-01
1.63190320e-01 3.90716232e-02 -1.38040885e-01 4.27454174e-01
1.36683166e-01 9.66672361e-01 2.91404426e-01 -1.29155362e+00
-2.79731899e-01 -9.52658474e-01 6.98304027e-02 2.95869708e-01
-2.00257778e-01 -2.74433881e-01 -7.41933763e-01 -5.48544109e-01
5.33714473e-01 -7.21170425e-01 -6.82937264e-01 4.00809705e-01
-5.07269144e-01 -6.33782208e-01 7.63873458e-01 -2.83547789e-01
7.81948447e-01 -2.14411783e+00 2.05106676e-01 3.22164774e-01
7.04176068e-01 -3.39878686e-02 -2.21247971e-01 1.56751946e-01
-6.18864708e-02 4.27550882e-01 -1.22907557e-01 -1.61031336e-01
-3.43943328e-01 3.92790258e-01 -1.43998682e-01 4.60179776e-01
5.71553528e-01 1.66487134e+00 -9.68275785e-01 -7.68760443e-01
2.52923369e-01 7.72569537e-01 -6.49255872e-01 -1.79473504e-01
-3.68295521e-01 6.16912246e-01 -1.28580356e+00 1.25845492e+00
7.74310410e-01 -8.40731919e-01 -3.80191475e-01 -3.48116338e-01
3.59884560e-01 -1.86806560e-01 -2.60959506e-01 2.08420348e+00
-2.74720877e-01 2.50518978e-01 -3.38315330e-02 -1.03309631e+00
8.79677474e-01 2.66972244e-01 9.64211643e-01 -6.54896200e-01
3.32793832e-01 -7.82191940e-03 -9.87349674e-02 -6.39164031e-01
2.37357672e-02 9.49857831e-02 1.41478434e-01 -2.20309258e-01
-7.10900053e-02 -2.11556211e-01 -6.60438955e-01 2.50032067e-01
1.24079502e+00 -2.19754502e-01 3.63235213e-02 -3.12752247e-01
3.73599976e-01 1.21964484e-01 2.41610244e-01 4.08220172e-01
-2.86723942e-01 6.27654374e-01 4.72578138e-01 -5.61756492e-01
-5.91904700e-01 -9.41963136e-01 -4.59267557e-01 5.82145274e-01
6.10609174e-01 -3.87560636e-01 -3.69240850e-01 -7.09890366e-01
3.81549746e-01 2.66162753e-01 -8.15691292e-01 -3.13987195e-01
-4.00158018e-01 -8.72580945e-01 4.95589375e-01 7.57450819e-01
2.64202148e-01 -8.51733983e-01 -2.36793414e-01 1.98099881e-01
-1.66693497e-02 -9.94409323e-01 -1.48042992e-01 -2.41933301e-01
-1.36441195e+00 -1.14702082e+00 -7.88009286e-01 -6.77243471e-01
8.49962056e-01 3.07524264e-01 1.02847064e+00 6.34569466e-01
-4.84327435e-01 5.39872885e-01 -3.65513831e-01 -6.94296181e-01
2.37151206e-01 6.59614876e-02 -3.46089959e-01 -1.77572176e-01
4.70823199e-01 -6.63935721e-01 -8.12726378e-01 3.15127939e-01
-6.93411112e-01 -7.01761320e-02 7.99897730e-01 9.08883274e-01
1.10814500e+00 -3.45862031e-01 4.41983223e-01 -8.69945228e-01
6.19743764e-01 -7.60291040e-01 -1.37737945e-01 8.55964497e-02
-2.55429804e-01 -2.95568913e-01 2.85931319e-01 -7.86263347e-02
-4.76262838e-01 -1.76142603e-01 -3.92312050e-01 -1.12947631e+00
-2.67242819e-01 4.62290734e-01 -1.59924641e-01 -4.39839602e-01
5.00574946e-01 2.23644674e-01 1.19278185e-01 -2.37637550e-01
8.93022493e-02 3.34520578e-01 1.02366455e-01 -5.38121700e-01
4.46099818e-01 8.18561852e-01 4.27874506e-01 -9.78124321e-01
-5.51924169e-01 -5.28717220e-01 -4.37105715e-01 -8.90760049e-02
1.04950333e+00 -7.73373604e-01 -1.16402602e+00 2.36641750e-01
-1.34551942e+00 4.15829103e-03 -1.85646385e-01 3.08817983e-01
-4.11311328e-01 1.76916167e-01 -6.14801228e-01 -4.14419115e-01
-4.42154080e-01 -1.31189096e+00 1.81268430e+00 -4.55534384e-02
1.26886651e-01 -1.22032571e+00 -9.69474316e-02 1.93269476e-01
2.80680954e-01 6.33287549e-01 1.13961804e+00 -6.90809429e-01
-1.17829919e+00 -1.28465056e-01 -5.28704643e-01 -3.26367587e-01
1.53320655e-01 -1.80583537e-01 -1.00232458e+00 -1.91496044e-01
-1.90700337e-01 -1.08951718e-01 8.80775869e-01 8.06249619e-01
1.88916230e+00 1.89768653e-02 -1.20509958e+00 8.80799770e-01
1.28611636e+00 -1.93991110e-01 5.56320667e-01 -9.81412753e-02
9.49246347e-01 3.64297718e-01 4.54474002e-01 1.36387408e-01
4.96018589e-01 3.89360189e-01 9.99548674e-01 -3.26677591e-01
3.58271226e-03 -1.10381708e-01 -3.66162926e-01 9.11038995e-01
-2.67060012e-01 -3.43069017e-01 -1.16644084e+00 6.80229485e-01
-1.68333852e+00 -5.31796157e-01 -3.32836002e-01 1.46998131e+00
3.55359316e-01 -2.93563437e-02 -3.42835844e-01 -1.37200952e-01
5.06881416e-01 5.03964126e-02 -7.62491882e-01 1.33109674e-01
-6.04032055e-02 4.88919675e-01 3.33382905e-01 1.51716813e-01
-9.23415899e-01 9.08680499e-01 5.61586523e+00 8.32026422e-01
-1.14366281e+00 1.93303734e-01 6.96057618e-01 -1.09088086e-01
-6.82452261e-01 -4.07866925e-01 -4.71528172e-01 2.20812187e-01
4.18395132e-01 1.76019982e-01 -6.40421659e-02 5.78995109e-01
6.71785772e-02 5.73308229e-01 -1.01310766e+00 1.23332131e+00
-1.59155175e-01 -1.69310236e+00 4.41895813e-01 3.64562273e-01
3.25599849e-01 4.76852566e-01 9.20484290e-02 2.03486577e-01
8.64808783e-02 -1.22349930e+00 -2.09555663e-02 7.04233229e-01
6.46329343e-01 -5.63451469e-01 7.05908954e-01 2.09880382e-01
-1.18322086e+00 6.27664402e-02 -7.15561569e-01 4.47016507e-01
-3.13032512e-03 6.09004915e-01 -7.59347677e-01 9.58683729e-01
8.94613206e-01 1.29080367e+00 -4.99176800e-01 1.16465914e+00
4.85704467e-02 3.55220020e-01 -4.00113553e-01 -3.50619346e-01
3.37039113e-01 1.04610816e-01 7.77536154e-01 8.41419280e-01
2.98960865e-01 4.76783931e-01 1.34062752e-01 1.17979014e+00
-1.19954094e-01 2.27455407e-01 -7.81159580e-01 1.14599417e-03
2.76497722e-01 1.17779768e+00 -7.73452878e-01 -7.84168616e-02
-2.97302097e-01 7.49526024e-01 3.39705646e-01 4.37166959e-01
-7.77211368e-01 -5.18661067e-02 9.75677371e-01 1.24669790e-01
1.25294566e-01 -7.22952113e-02 -3.85052174e-01 -9.65166926e-01
-1.32727951e-01 -5.05307317e-01 3.30231130e-01 -9.20594096e-01
-1.66954207e+00 5.40254474e-01 -8.34962726e-02 -1.28351808e+00
6.88771367e-01 -7.46665180e-01 -7.22441614e-01 7.10172892e-01
-1.75853562e+00 -1.43488026e+00 -5.99788845e-01 7.00544417e-01
3.77526700e-01 5.51348031e-02 6.61270857e-01 3.52727622e-01
-2.76618630e-01 4.26396340e-01 -2.62969196e-01 2.43565775e-02
1.65029064e-01 -1.12133634e+00 4.88591164e-01 -1.36724472e-01
7.80380517e-02 7.01026261e-01 1.04733318e-01 -8.06320846e-01
-1.69717062e+00 -1.43873990e+00 5.15607357e-01 -7.64484286e-01
2.65722513e-01 -3.62632632e-01 -1.19753504e+00 5.63089669e-01
-1.87953129e-01 5.34764588e-01 6.23123229e-01 1.89921215e-01
4.65625264e-02 9.19994488e-02 -1.14954841e+00 4.94652689e-01
1.58007503e+00 -2.84918725e-01 -3.74820441e-01 6.90549910e-01
1.29798365e+00 -7.06022739e-01 -1.19854963e+00 6.09800398e-01
1.21824384e-01 -5.20099938e-01 1.25019157e+00 -8.62085283e-01
4.20617521e-01 -2.67861694e-01 -2.61712551e-01 -1.28907549e+00
-6.48219645e-01 9.19359084e-03 4.11263993e-03 7.04156637e-01
2.46413723e-01 -8.04861963e-01 9.28530991e-01 8.81511718e-02
-7.13502824e-01 -1.49132097e+00 -1.14556682e+00 -3.94473642e-01
2.62030363e-01 -5.00244319e-01 1.11816120e+00 1.16162360e+00
-1.99133962e-01 1.58153191e-01 3.13808173e-01 6.29085660e-01
5.05916238e-01 1.75133497e-01 5.82479835e-01 -1.43393135e+00
8.05207342e-02 -4.61104035e-01 -8.57313037e-01 -1.04857075e+00
1.84660286e-01 -1.44661283e+00 -5.64525127e-01 -1.85614848e+00
-4.27720249e-02 -7.60982871e-01 -6.00013316e-01 3.25235307e-01
-1.81194037e-01 1.55658682e-03 -1.88452557e-01 2.01590449e-01
-5.23473442e-01 7.82436073e-01 1.95642650e+00 -5.29285371e-01
-7.06662983e-02 -3.47943604e-03 -7.21385777e-01 3.77896875e-01
6.24702573e-01 -3.09817970e-01 -3.88354957e-01 -6.64572418e-01
2.25941613e-01 1.04236387e-01 9.68516171e-01 -8.70021164e-01
3.66334736e-01 -2.81946640e-02 6.91371620e-01 -8.52988362e-01
5.56264818e-01 -9.59719539e-01 -1.61299258e-02 3.61165732e-01
-3.63806426e-03 -1.37330323e-01 3.52920353e-01 8.93258750e-01
-2.48932526e-01 4.37229216e-01 4.51169014e-01 -4.19507176e-01
-2.98986912e-01 1.26763344e+00 2.55870938e-01 -9.90051925e-02
1.07260036e+00 -3.56996119e-01 -2.82367945e-01 1.09715275e-01
-8.84960115e-01 6.67618990e-01 3.39798659e-01 4.67425853e-01
1.06057477e+00 -1.59833765e+00 -6.16866112e-01 3.09729069e-01
4.58844692e-01 6.26892209e-01 6.91312671e-01 9.70083475e-01
-5.42983234e-01 4.71382737e-01 -1.04857616e-01 -1.35778177e+00
-9.78015184e-01 6.77722037e-01 5.92745900e-01 -5.63586950e-02
-8.57538342e-01 1.11525631e+00 6.17668211e-01 -5.57461381e-01
-6.41285777e-02 -1.01867437e+00 -2.51574695e-01 -2.61507601e-01
7.69883990e-02 1.06308877e-01 1.46512702e-01 -4.58226651e-01
-6.72708988e-01 9.65324521e-01 -2.12203003e-02 6.31434619e-01
1.59331334e+00 1.57411590e-01 -2.50919402e-01 1.68258905e-01
1.26131296e+00 -3.85901928e-01 -7.32075453e-01 -2.55011052e-01
-3.30447286e-01 -4.47758496e-01 2.71263808e-01 -4.92787868e-01
-1.35633528e+00 1.33466065e+00 6.63727105e-01 5.41599467e-02
9.14721429e-01 5.04283726e-01 9.15328741e-01 2.57628024e-01
5.07124186e-01 -3.75849158e-01 5.03039025e-02 1.80053100e-01
1.11372316e+00 -1.16989017e+00 -8.91161487e-02 -7.40204930e-01
-2.85292208e-01 8.42161715e-01 7.17926323e-01 -2.71311343e-01
1.06703186e+00 -7.80343786e-02 -1.28287271e-01 -1.05131805e+00
-7.88088560e-01 -3.82851809e-02 3.19820791e-01 7.10672438e-01
2.43447170e-01 7.24492595e-02 2.68275350e-01 7.38389075e-01
-1.78268086e-02 1.40284086e-02 -1.62458897e-01 6.47784829e-01
-2.46175170e-01 -7.55147219e-01 -1.31234109e-01 7.44656086e-01
-4.21789527e-01 -3.83223966e-02 -6.12604856e-01 9.28893209e-01
1.52987346e-01 3.51299882e-01 2.08354965e-01 -2.90856779e-01
5.87901890e-01 -3.12885195e-01 5.85599959e-01 -6.05698586e-01
-6.34745717e-01 1.62249744e-01 -4.77788001e-01 -7.53525734e-01
-2.73780704e-01 -4.91064519e-01 -1.58583319e+00 -5.64155504e-02
-3.98913920e-01 -2.59430315e-02 4.44581509e-01 6.17499769e-01
7.78291285e-01 9.46038902e-01 3.82256806e-01 -7.19403625e-01
-1.11387469e-01 -6.41973913e-01 -4.58622932e-01 3.51941079e-01
4.38959479e-01 -8.35508168e-01 7.42278770e-02 -5.25011539e-01] | [7.960484504699707, -3.6002018451690674] |
5129c75a-f27b-4778-9f11-fc916b294417 | deep-multi-task-learning-for-anomalous | 1907.00749 | null | https://arxiv.org/abs/1907.00749v1 | https://arxiv.org/pdf/1907.00749v1.pdf | Deep Multi-Task Learning for Anomalous Driving Detection Using CAN Bus Scalar Sensor Data | Corner cases are the main bottlenecks when applying Artificial Intelligence (AI) systems to safety-critical applications. An AI system should be intelligent enough to detect such situations so that system developers can prepare for subsequent planning. In this paper, we propose semi-supervised anomaly detection considering the imbalance of normal situations. In particular, driving data consists of multiple positive/normal situations (e.g., right turn, going straight), some of which (e.g., U-turn) could be as rare as anomalous situations. Existing machine learning based anomaly detection approaches do not fare sufficiently well when applied to such imbalanced data. In this paper, we present a novel multi-task learning based approach that leverages domain-knowledge (maneuver labels) for anomaly detection in driving data. We evaluate the proposed approach both quantitatively and qualitatively on 150 hours of real-world driving data and show improved performance over baseline approaches. | ['Teruhisa Misu', 'Dario Pompili', 'Vidyasagar Sadhu'] | 2019-06-28 | null | null | null | null | ['supervised-anomaly-detection', 'semi-supervised-anomaly-detection'] | ['computer-vision', 'computer-vision'] | [ 1.76941037e-01 -6.59235269e-02 -1.61511943e-01 -5.80267191e-01
-6.11609995e-01 -4.79466677e-01 6.42619967e-01 7.01892793e-01
-8.81554782e-02 4.92766649e-01 -7.87072331e-02 -9.09649611e-01
-1.18251085e-01 -5.86565733e-01 -7.67041743e-01 -4.36166734e-01
-1.64040849e-01 5.83512187e-01 4.62392181e-01 -6.97212875e-01
5.58598220e-01 5.99069476e-01 -1.92557490e+00 3.98870319e-01
1.08420873e+00 1.04232776e+00 -6.27175868e-01 6.89206958e-01
-1.77970558e-01 8.76198888e-01 -7.49566197e-01 -4.01658863e-02
4.77119863e-01 2.14313660e-02 -5.95478237e-01 1.20464806e-02
5.69504499e-01 -3.59680831e-01 -1.46510988e-01 8.27971697e-01
-2.03407630e-02 2.19054595e-01 8.04800868e-01 -2.10381055e+00
2.30479151e-01 7.40293339e-02 -7.53362000e-01 8.15089643e-01
2.64069527e-01 5.71411252e-01 9.17416632e-01 -6.01737440e-01
3.77532952e-02 8.72418880e-01 3.88389796e-01 2.89424479e-01
-9.09350634e-01 -6.27359271e-01 4.03332442e-01 7.45572567e-01
-9.27532971e-01 -3.96189243e-01 1.11678934e+00 -6.67745590e-01
1.13359559e+00 2.37541184e-01 3.13945234e-01 1.03252447e+00
7.09008813e-01 9.72178161e-01 7.84017265e-01 -1.46294773e-01
5.42812526e-01 -2.03439400e-01 4.67895001e-01 1.68777809e-01
4.53289986e-01 2.69584358e-01 -3.48470539e-01 -2.61380851e-01
-2.57332683e-01 3.04242820e-01 6.22722387e-01 -1.48911670e-01
-1.10998392e+00 6.61877513e-01 -2.40162797e-02 4.17229533e-02
-5.75357974e-01 -2.34836951e-01 8.47553611e-01 5.45846105e-01
3.67701948e-01 4.84216869e-01 -4.19962853e-01 -3.47536266e-01
-8.85182381e-01 6.91351593e-01 6.03241146e-01 8.94776046e-01
7.93783724e-01 3.38430822e-01 -2.80833870e-01 5.23361802e-01
-3.53766419e-02 4.33130503e-01 3.63927126e-01 -5.98885357e-01
7.04963624e-01 8.34196687e-01 1.91178024e-01 -1.02285814e+00
-6.72134876e-01 1.96888093e-02 -6.16737187e-01 4.33518678e-01
5.56054711e-01 -2.36862347e-01 -9.98895049e-01 1.24406874e+00
2.57329971e-01 4.01603699e-01 2.86261201e-01 6.83330178e-01
3.42028707e-01 5.60360312e-01 1.43904105e-01 1.24012925e-01
1.07093573e+00 -7.03147471e-01 -8.36652040e-01 -6.11748874e-01
1.02783799e+00 -7.30115473e-01 1.00912154e+00 5.97253799e-01
-5.57621539e-01 -4.60465640e-01 -1.05372679e+00 5.14090717e-01
-6.65873766e-01 -2.73939818e-01 4.32594299e-01 4.30520713e-01
-3.81390840e-01 2.60217607e-01 -7.57107615e-01 -2.22220719e-01
3.39776397e-01 1.17494293e-01 -2.96768099e-01 1.87976025e-02
-1.13857126e+00 9.02893722e-01 5.15543818e-01 6.95239753e-02
-1.02965868e+00 -4.72909331e-01 -1.04051507e+00 -5.19043744e-01
8.30620408e-01 1.53625950e-01 1.51221013e+00 -6.24011576e-01
-7.09821820e-01 5.88737249e-01 -3.05423349e-01 -7.92248130e-01
5.25022805e-01 -3.66029680e-01 -1.13990355e+00 -3.42338860e-01
1.68973565e-01 2.06057429e-01 8.58270764e-01 -1.11350501e+00
-1.16647077e+00 -3.86972696e-01 1.39506441e-02 -1.23864837e-01
-1.75001938e-02 3.89673300e-02 4.29875821e-01 -3.30094397e-01
-1.33774534e-01 -1.02810419e+00 -3.74435008e-01 -5.10861814e-01
-7.59958327e-01 -4.75539923e-01 1.31354415e+00 -4.59322929e-01
1.48151517e+00 -2.20466471e+00 -6.03483021e-01 5.09512722e-01
1.55245140e-01 4.82120723e-01 1.02531679e-01 4.55666423e-01
-3.07316810e-01 -1.86114222e-01 -2.29108170e-01 1.06678218e-01
-8.50544050e-02 5.57346165e-01 -5.95068157e-01 4.98444915e-01
5.96580327e-01 6.66146636e-01 -8.14789474e-01 -2.75574267e-01
4.89067823e-01 -5.21903157e-01 -4.33586985e-01 2.44581729e-01
-2.60158300e-01 4.19571221e-01 -6.30334496e-01 1.00566697e+00
5.44926286e-01 3.51673633e-01 -4.54713225e-01 7.79940188e-02
-2.23424032e-01 6.79963231e-02 -9.99291182e-01 1.03291476e+00
-2.78946042e-01 7.22736537e-01 -5.38842082e-01 -1.36178279e+00
9.75841582e-01 1.59389436e-01 5.05361080e-01 -9.63624358e-01
1.36671960e-02 1.81083411e-01 5.47460794e-01 -7.76583433e-01
5.62848091e-01 9.04523507e-02 -5.75869739e-01 4.64315832e-01
-5.84917963e-01 -1.49343073e-01 4.07108337e-01 9.99364778e-02
1.50392377e+00 -2.02423051e-01 5.00247777e-01 -9.97370854e-02
6.09216869e-01 5.61021566e-01 8.15083504e-01 6.45829320e-01
-7.73612797e-01 2.63314784e-01 8.92085075e-01 -9.87372935e-01
-1.06166744e+00 -9.43972051e-01 1.83099076e-01 8.88060212e-01
1.10229336e-01 -2.62802720e-01 -4.50123012e-01 -8.56839478e-01
4.09869641e-01 1.32454014e+00 -4.92977172e-01 -7.20310092e-01
-7.26544797e-01 -5.24963617e-01 5.42936563e-01 5.35167396e-01
1.88230634e-01 -1.13258219e+00 -9.09438014e-01 2.45011032e-01
-6.11043535e-02 -1.32476377e+00 -8.63324329e-02 2.82671034e-01
-6.37046158e-01 -1.16072786e+00 3.35241556e-01 -1.91136599e-01
6.05585396e-01 2.52431571e-01 1.10411811e+00 -6.21467829e-02
-3.39747310e-01 4.78821099e-02 -3.51365358e-01 -9.94102657e-01
-6.50287509e-01 -1.26518428e-01 3.92811179e-01 2.40596399e-01
1.00019348e+00 -3.70323092e-01 -4.43776935e-01 6.46022081e-01
-8.82661223e-01 -2.95913577e-01 7.45922446e-01 4.85137910e-01
5.12215734e-01 2.18908906e-01 8.37191164e-01 -8.72209966e-01
7.06271827e-01 -9.14628565e-01 -4.21483606e-01 6.42068684e-02
-8.07026684e-01 -3.59228700e-02 9.65148866e-01 -3.29754233e-01
-8.16812158e-01 9.56029370e-02 -3.28880586e-02 -4.61346775e-01
-1.12860107e+00 3.08104277e-01 -1.08329451e-03 2.93740600e-01
7.95920908e-01 7.03636184e-03 3.46979275e-02 -9.39280391e-02
-4.62498926e-02 8.42887819e-01 4.88796890e-01 -5.06843746e-01
8.95992458e-01 3.56909215e-01 5.28477393e-02 -6.40599966e-01
-6.34594798e-01 -7.48320401e-01 -6.02178156e-01 -5.01973152e-01
5.38273871e-01 -7.06063569e-01 -4.07415241e-01 5.33630013e-01
-6.21203601e-01 -3.00432056e-01 -2.13293523e-01 1.47724837e-01
-5.78580618e-01 2.18236700e-01 -2.06305999e-02 -1.00504076e+00
1.09356225e-01 -1.12525737e+00 1.13280952e+00 1.31709138e-02
-6.45359635e-01 -7.04697967e-01 5.72458580e-02 3.41560215e-01
3.87597799e-01 6.11720443e-01 8.11636627e-01 -1.43359470e+00
-1.29525363e-01 -6.15648389e-01 1.38075620e-01 2.35606372e-01
2.40446091e-01 2.86155730e-01 -9.37667251e-01 -1.83866788e-02
-4.36381549e-01 -1.03022911e-01 5.57607830e-01 9.80580002e-02
1.32745159e+00 -2.80753285e-01 -3.88451040e-01 5.59696108e-02
8.32164645e-01 5.85420609e-01 6.44037783e-01 5.55884004e-01
7.24792302e-01 6.81165576e-01 1.37992370e+00 5.67330420e-01
4.86560941e-01 6.52431428e-01 5.35241842e-01 -7.90672898e-02
3.18319738e-01 1.35191292e-01 4.04698044e-01 1.37342066e-01
2.36428663e-01 -1.13584578e-01 -1.56579161e+00 8.32224190e-01
-2.04342508e+00 -1.03148687e+00 -3.74788016e-01 2.09782600e+00
2.65153646e-01 7.81637549e-01 4.53483373e-01 6.53554916e-01
5.06710410e-01 9.34055969e-02 -9.53334868e-01 -9.59577799e-01
1.03363506e-01 -1.86767370e-01 4.47683483e-01 7.84931481e-02
-1.50977027e+00 7.56783247e-01 5.96256447e+00 5.62483251e-01
-9.97966886e-01 -1.87527180e-01 7.01030016e-01 -2.42890432e-01
-1.06161408e-01 -1.22728325e-01 -6.92654014e-01 5.67895889e-01
1.33962667e+00 -2.28280976e-01 -5.04459925e-02 9.99622405e-01
3.86273324e-01 -3.11987430e-01 -1.18665087e+00 7.08012462e-01
5.22732064e-02 -7.86320448e-01 -1.69278204e-01 -6.00480996e-02
6.00804090e-01 6.02423139e-02 2.97985412e-02 5.93651772e-01
1.96076676e-01 -9.10912693e-01 5.94004869e-01 2.82707721e-01
3.04100543e-01 -9.54760313e-01 9.07869875e-01 5.27844548e-01
-9.21198547e-01 -4.18285906e-01 1.20420367e-01 -3.33453119e-01
1.51124910e-01 8.16691220e-01 -9.08232272e-01 4.96715039e-01
7.23855138e-01 6.99617684e-01 -7.34096229e-01 8.42259169e-01
9.18488428e-02 7.57669687e-01 -4.17982370e-01 2.10277930e-01
4.04606640e-01 -2.59090848e-02 8.74633133e-01 9.92973566e-01
1.84548676e-01 -2.67473787e-01 4.40208137e-01 3.66926134e-01
6.04611516e-01 -1.62090749e-01 -1.34674633e+00 5.66552319e-02
2.33608946e-01 1.17329109e+00 -5.96980453e-01 -2.50276595e-01
-7.28460848e-01 5.47815561e-01 -4.76909354e-02 3.88749152e-01
-9.05506730e-01 -4.12253916e-01 1.13722134e+00 3.18478674e-01
-7.33668134e-02 -1.57508597e-01 -5.16605854e-01 -8.21722269e-01
3.77759963e-01 -9.47881043e-01 5.25768578e-01 -2.34134004e-01
-1.35988295e+00 4.81394976e-01 2.26586163e-01 -1.74796128e+00
-7.55958498e-01 -5.60544729e-01 -1.21065187e+00 3.67979556e-01
-1.70593345e+00 -8.37343454e-01 -4.07813847e-01 5.42533457e-01
6.58246994e-01 -5.44967592e-01 4.01300371e-01 2.64616370e-01
-8.44650209e-01 5.93734860e-01 -3.72017562e-01 1.04413209e-02
8.10100377e-01 -1.29559898e+00 8.88450205e-01 1.25733840e+00
-1.61088496e-01 1.92015931e-01 1.03306699e+00 -7.24547982e-01
-1.27044034e+00 -1.46687698e+00 6.52221203e-01 -6.31950736e-01
8.52621257e-01 -1.49616286e-01 -1.10407591e+00 8.38732958e-01
-1.50158525e-01 3.15412134e-01 7.32360840e-01 1.44671738e-01
-2.88167000e-01 -2.30610266e-01 -1.23264563e+00 5.37839115e-01
7.57181227e-01 -2.22933084e-01 -8.35560143e-01 3.41124982e-01
2.20472932e-01 -6.52134538e-01 -3.89384538e-01 7.23362267e-01
2.08339736e-01 -1.08945239e+00 7.08209455e-01 -1.07317233e+00
4.42491233e-01 -5.07019997e-01 2.75809225e-02 -1.43922544e+00
1.29098311e-01 -5.04205763e-01 -3.76349568e-01 7.24860430e-01
5.97298682e-01 -9.33096290e-01 5.73239863e-01 7.35549569e-01
-6.63921833e-01 -7.86848426e-01 -1.09629750e+00 -8.99096429e-01
-3.87918949e-03 -1.06895185e+00 7.94684470e-01 7.27742374e-01
9.30275023e-02 -1.30996972e-01 -3.56396407e-01 2.79889762e-01
4.40066665e-01 1.19077101e-01 1.09290063e+00 -1.40783918e+00
3.00593972e-01 -4.22032475e-01 -9.72605407e-01 -4.33775425e-01
3.69356513e-01 -4.68259513e-01 2.15017483e-01 -1.10268760e+00
-2.44572833e-01 -4.64857459e-01 -5.29695809e-01 7.78987586e-01
-2.26151094e-01 -4.25470918e-02 -3.37549776e-01 9.69910845e-02
-7.42705464e-01 1.34622321e-01 6.26418471e-01 -1.32621437e-01
-2.08596379e-01 3.70837957e-01 -5.66609144e-01 8.17162216e-01
1.23051858e+00 -3.81684750e-01 -3.43932688e-01 -2.56001968e-02
2.19188258e-01 -2.03957319e-01 3.06433648e-01 -1.29765284e+00
2.98713177e-01 -5.63183844e-01 6.87043518e-02 -8.42956722e-01
-4.40924056e-02 -9.87692058e-01 -2.65561432e-01 4.17538226e-01
-1.88303262e-01 5.61362207e-01 5.90977192e-01 7.72788525e-01
-5.53680539e-01 1.61455691e-01 3.90516043e-01 4.75011975e-01
-1.32537115e+00 3.44486415e-01 -7.89431155e-01 1.75535470e-01
1.87655771e+00 -3.29783976e-01 -4.41912532e-01 -2.88846523e-01
-4.12692070e-01 7.22621143e-01 2.27802649e-01 1.02219772e+00
6.64007366e-01 -1.19628179e+00 -6.40558898e-01 5.78743637e-01
9.87308800e-01 9.68137756e-02 3.05394322e-01 9.65147436e-01
-2.35478058e-01 3.58988374e-01 -3.91366541e-01 -7.41592765e-01
-1.13565290e+00 5.44262171e-01 2.01489270e-01 -5.28403856e-02
-6.87025905e-01 9.77120847e-02 -1.01976342e-01 -4.74430948e-01
1.21244021e-01 -5.24982154e-01 -3.44717652e-01 3.91061381e-02
6.30238891e-01 3.34535927e-01 5.02386212e-01 -8.03149343e-01
-6.32015765e-01 2.88619816e-01 -3.88351977e-01 3.52136254e-01
8.90298247e-01 2.18774587e-01 3.45965147e-01 8.62482548e-01
6.12235665e-01 -3.36911857e-01 -9.50120389e-01 -1.11297794e-01
4.18675691e-01 -5.59870660e-01 -8.52534175e-02 -6.27542973e-01
-9.05304611e-01 8.07298541e-01 5.42515755e-01 4.09254104e-01
1.07576632e+00 -5.81668727e-02 1.01223719e+00 6.17845833e-01
2.90502399e-01 -1.42606652e+00 -9.42872837e-03 5.58907032e-01
5.21580994e-01 -1.79687905e+00 -2.67594695e-01 -5.82014360e-02
-1.14497447e+00 9.08717752e-01 1.10763884e+00 -1.17782056e-01
6.75761521e-01 3.93222928e-01 3.02170813e-01 -4.06548530e-01
-1.02574337e+00 -2.28485093e-01 3.94042134e-01 4.67363596e-01
5.64574376e-02 3.86483669e-01 1.16229542e-01 2.03665450e-01
-9.38763171e-02 -3.18820149e-01 7.07093537e-01 1.11765301e+00
-4.41011041e-01 -7.36526906e-01 -3.08005542e-01 1.08558536e+00
-3.99445474e-01 3.89154673e-01 -3.53968948e-01 6.38810873e-01
1.72143996e-01 1.22839129e+00 6.03986740e-01 -6.98183119e-01
9.18031096e-01 3.13904762e-01 -3.02236736e-01 -4.90836740e-01
-3.41918528e-01 -5.06646574e-01 2.08338648e-01 -9.06398952e-01
-6.53608814e-02 -9.39224601e-01 -1.36969221e+00 -3.46152842e-01
1.83890000e-01 -1.18359573e-01 5.42017400e-01 1.32631350e+00
4.23879147e-01 7.65630066e-01 8.02551448e-01 -4.36169893e-01
-4.07843977e-01 -7.37093389e-01 -3.49693656e-01 7.46653497e-01
7.71477818e-01 -9.97899711e-01 -6.41862750e-01 -2.18267173e-01] | [7.49802827835083, 2.4420418739318848] |
ea093bf1-124e-465b-9045-6808fa20a529 | sparkly-a-simple-yet-surprisingly-strong-tf | null | null | https://dl.acm.org/doi/abs/10.14778/3583140.3583163 | https://dl.acm.org/doi/pdf/10.14778/3583140.3583163 | Sparkly: A Simple yet Surprisingly Strong TF/IDF Blocker for Entity Matching | Blocking is a major task in entity matching. Numerous blocking solutions have been developed, but as far as we can tell, blocking using the well-known tf/idf measure has received virtually no attention. Yet, when we experimented with tf/idf blocking using Lucene, we found it did quite well. So in this paper we examine tf/idf blocking in depth. We develop Sparkly, which uses Lucene to perform top-k tf/idf blocking in a distributed share-nothing fashion on a Spark cluster. We develop techniques to identify good attributes and tokenizers that can be used to block on, making Sparkly completely automatic. We perform extensive experiments showing that Sparkly outperforms 8 state-of-the-art blockers. Finally, we provide an in-depth analysis of Sparkly's performance, regarding both recall/output size and runtime. Our findings suggest that (a) tf/idf blocking needs more attention, (b) Sparkly forms a strong baseline that future blocking work should compare against, and (c) future blocking work should seriously consider top-k blocking, which helps improve recall, and a distributed share-nothing architecture, which helps improve scalability, predictability, and extensibility. | ['AnHai Doan', 'Yash Govind', 'Derek Paulsen'] | 2023-04-20 | null | null | null | proceedings-of-the-vldb-endowment-2023-4 | ['blocking'] | ['natural-language-processing'] | [-5.93155265e-01 -4.46546137e-01 -4.18732613e-01 -7.23925829e-01
-8.53441834e-01 -6.26167715e-01 5.22189736e-01 5.31334043e-01
-7.27785051e-01 3.66274834e-01 6.18788183e-01 -7.30993390e-01
-1.22942194e-01 -9.11597967e-01 -7.59312391e-01 -3.92549396e-01
-2.01041207e-01 6.13694847e-01 5.05397379e-01 -2.81783879e-01
8.17918330e-02 5.28624654e-01 -1.71613932e+00 7.98211992e-01
6.86362684e-01 4.55807269e-01 -2.30406389e-01 5.50076425e-01
-3.10401350e-01 9.47864056e-01 -8.08926284e-01 -5.97629368e-01
4.62182939e-01 2.92422533e-01 -1.21610653e+00 -8.72884691e-01
6.45646036e-01 -5.07751584e-01 -2.78158635e-01 5.07603884e-01
5.16237497e-01 -1.51543856e-01 1.90155163e-01 -1.46524775e+00
-2.18212381e-01 8.80013168e-01 -5.81737995e-01 5.87175429e-01
2.83265978e-01 2.06822947e-01 1.03300560e+00 -6.40050173e-01
6.44315779e-01 1.08979368e+00 8.53527069e-01 2.69384325e-01
-1.34608781e+00 -1.03667188e+00 1.10735081e-01 5.60357682e-02
-1.61419702e+00 -7.11739182e-01 3.22258435e-02 -3.07675511e-01
1.54891467e+00 6.69183850e-01 3.85808021e-01 4.15701538e-01
-2.00167075e-02 8.95734668e-01 1.05164063e+00 -6.74824357e-01
4.57513556e-02 3.33220959e-01 4.80859011e-01 1.99187815e-01
2.99371362e-01 2.09914133e-01 -6.77720785e-01 -7.11120725e-01
2.60597348e-01 4.19220654e-03 1.52359053e-01 -2.20869243e-01
-1.15877736e+00 6.80135429e-01 3.93705100e-01 3.46520096e-01
-3.60425532e-01 3.81581634e-01 7.42729783e-01 5.09113371e-01
4.19016570e-01 4.08062458e-01 -5.46492517e-01 -2.74364889e-01
-1.13601327e+00 6.10781014e-01 1.05782390e+00 1.03863084e+00
9.84650671e-01 -6.20301247e-01 -3.43016356e-01 8.56788516e-01
2.41083857e-02 5.32446325e-01 -1.17559031e-01 -8.54471147e-01
5.39100826e-01 6.28115952e-01 2.42411047e-01 -8.37425709e-01
-5.11353970e-01 -2.40949560e-02 -3.60545963e-01 3.59693915e-02
4.10247177e-01 8.11854154e-02 -5.62422693e-01 1.62052214e+00
4.41646367e-01 2.67135183e-04 -3.99571449e-01 7.96457410e-01
7.38389730e-01 4.71210361e-01 4.22028124e-01 1.30164012e-01
1.65885997e+00 -9.70043004e-01 -4.74276066e-01 -1.43947348e-01
1.26110411e+00 -1.06661713e+00 8.68033946e-01 1.79673880e-01
-1.11178362e+00 -3.39895159e-01 -8.07027757e-01 -1.71202570e-01
-4.61285561e-01 -3.53002578e-01 1.20703638e+00 1.02503419e+00
-1.03877115e+00 1.78673193e-01 -1.10122550e+00 -4.65107113e-01
4.61713746e-02 5.09556711e-01 -3.73062402e-01 -1.54014528e-01
-1.22556686e+00 8.69291723e-01 8.66765380e-02 -3.38512629e-01
-6.34306133e-01 -9.30526793e-01 -2.49253660e-01 1.24075584e-01
3.58769268e-01 -8.68809819e-01 1.41994941e+00 -4.71033305e-01
-6.42707109e-01 1.11192214e+00 -5.00482976e-01 -4.46732432e-01
2.87498146e-01 -3.09762031e-01 -5.40553808e-01 -1.31180108e-01
3.08510333e-01 6.88449681e-01 1.13042079e-01 -1.05637383e+00
-9.32343543e-01 -4.60449219e-01 2.54187196e-01 2.13110596e-02
-5.58408797e-01 4.62273717e-01 -5.92217982e-01 -4.84872192e-01
-2.59629518e-01 -9.14882779e-01 -8.47049579e-02 -4.52527672e-01
-1.16947100e-01 -5.40414274e-01 1.93297341e-01 -3.52105349e-01
1.89251792e+00 -2.20746374e+00 -7.44568110e-01 6.03423119e-01
4.67304021e-01 3.58271509e-01 3.88813019e-02 1.00032854e+00
-2.01921072e-02 3.14493150e-01 3.02493572e-01 1.27957225e-01
-3.37754078e-02 -7.87927508e-02 -3.13252062e-01 5.08258641e-01
-3.33888441e-01 8.57916355e-01 -7.10249603e-01 -5.88650644e-01
-1.99297033e-02 3.41887563e-01 -6.74215257e-01 9.80456844e-02
-4.40244265e-02 -4.45567280e-01 -4.58218396e-01 5.90032578e-01
7.63647079e-01 -2.75045604e-01 8.90099779e-02 2.44684896e-04
-4.99686927e-01 6.62762821e-01 -1.21846569e+00 1.36426663e+00
-4.58765060e-01 3.30493897e-01 3.04356396e-01 -8.34292829e-01
9.73753810e-01 1.99457526e-01 5.94939172e-01 -9.24634814e-01
-4.05987263e-01 3.89223188e-01 -5.36145866e-02 -3.13300461e-01
8.04432631e-01 2.37146944e-01 -5.85232191e-02 8.60769033e-01
-6.40344381e-01 2.65801847e-01 3.03164542e-01 7.12669015e-01
1.69905460e+00 -4.14014906e-01 -7.64709860e-02 -6.53606057e-01
2.52057016e-01 6.05696201e-01 4.77096528e-01 1.17341912e+00
-1.50580838e-01 1.95131868e-01 2.16041580e-01 -7.43215382e-01
-1.10832798e+00 -7.90537059e-01 -3.50569487e-01 1.66417658e+00
1.68361589e-01 -1.10125148e+00 -6.92855239e-01 -4.89339948e-01
5.69084644e-01 6.11261666e-01 -3.48167777e-01 8.50324780e-02
-8.30235600e-01 -6.76818073e-01 1.00986862e+00 6.36789024e-01
5.06944895e-01 -9.19589818e-01 -6.02635741e-01 3.76282990e-01
-1.63042992e-01 -7.22599506e-01 -4.60490406e-01 4.51650262e-01
-8.92548203e-01 -8.53689492e-01 -3.96691322e-01 -6.58595979e-01
2.01570243e-01 8.72788370e-01 1.57683206e+00 5.40994108e-01
-2.90859848e-01 2.77701885e-01 -2.50128865e-01 -5.92893362e-01
-2.85006940e-01 3.59005213e-01 -3.89837146e-01 -7.67300248e-01
1.30655742e+00 -3.85051310e-01 -9.48053002e-01 9.16064262e-01
-7.53067553e-01 -5.57278916e-02 4.34811920e-01 6.21937633e-01
1.84355140e-01 -9.68147963e-02 5.06051362e-01 -1.39376771e+00
6.51827991e-01 -3.90919834e-01 -3.68044943e-01 2.69388139e-01
-1.07595980e+00 -8.35740864e-02 3.79199684e-01 -5.35289245e-03
-9.89162147e-01 -1.48848012e-01 -2.29517907e-01 -5.75620346e-02
-1.56005755e-01 2.30813324e-01 2.15415120e-01 3.13994765e-01
1.28033495e+00 -2.74203837e-01 -1.67882647e-02 -5.28776646e-01
4.52932417e-01 1.13696051e+00 4.80044127e-01 -1.05510974e+00
3.93592566e-01 7.70144880e-01 -6.71982408e-01 -4.26934123e-01
-3.26043665e-01 -1.15752792e+00 -8.76917988e-02 2.25937515e-01
3.28509241e-01 -1.33413756e+00 -1.13277948e+00 2.09008485e-01
-9.35423851e-01 -4.28088546e-01 -8.29241201e-02 2.80902684e-01
-2.41402641e-01 4.35168818e-02 -9.81705785e-01 -5.78601778e-01
-5.28858721e-01 -9.48658705e-01 1.03377962e+00 -7.15729743e-02
-5.42312205e-01 -4.60163027e-01 3.32438916e-01 6.11704648e-01
7.79454112e-01 -4.89425629e-01 7.33660579e-01 -7.80604899e-01
-7.01123595e-01 -1.29610807e-01 -4.31933522e-01 -9.31503996e-02
-3.91153842e-01 -1.07805409e-01 -8.79937530e-01 -3.58311087e-01
-4.35562372e-01 -7.36886263e-02 9.17366922e-01 3.01551789e-01
9.18908656e-01 -1.22573063e-01 -1.04237294e+00 5.08401155e-01
1.28454018e+00 2.36256778e-01 7.83465743e-01 1.04741824e+00
4.47353721e-01 6.43193364e-01 7.91858017e-01 4.25358862e-01
7.08209574e-01 8.63163888e-01 -9.90336984e-02 -4.48397875e-01
-1.51792854e-01 -1.24230824e-01 2.52092499e-02 5.57821929e-01
1.79296553e-01 2.09646616e-02 -1.31067359e+00 8.62792373e-01
-1.97963536e+00 -1.03908467e+00 -3.67811143e-01 2.25410271e+00
9.08464193e-01 2.06142470e-01 6.14922881e-01 -2.14075483e-02
7.94373274e-01 -8.81347656e-02 -2.79277474e-01 -8.03093255e-01
4.26593348e-02 3.71257335e-01 9.64065373e-01 1.29432246e-01
-9.08012092e-01 1.00138867e+00 6.50352812e+00 8.05511594e-01
-9.88728225e-01 2.07606524e-01 7.11095750e-01 -2.88616657e-01
-5.23768365e-01 3.29323649e-01 -1.13602912e+00 4.41337228e-01
9.76544380e-01 -3.64817649e-01 4.14941669e-01 1.02846551e+00
4.83928360e-02 -2.85502404e-01 -1.24046421e+00 8.38987947e-01
-5.09413421e-01 -1.29481685e+00 -4.85520393e-01 1.15275010e-01
6.19019508e-01 3.55263978e-01 -3.80537659e-01 5.49054265e-01
7.24479377e-01 -9.59712267e-01 4.72982705e-01 2.86686942e-02
5.73983610e-01 -7.71452427e-01 7.10532248e-01 3.42919320e-01
-1.12630677e+00 2.29645707e-02 -4.46788460e-01 -1.55017436e-01
-4.56368225e-03 1.04564345e+00 -8.17918837e-01 3.07486296e-01
1.25228131e+00 1.72853932e-01 -5.71272433e-01 1.10391295e+00
2.22308576e-01 8.13874483e-01 -6.95604682e-01 -1.35185242e-01
9.69483703e-02 1.92330629e-01 1.76504776e-02 1.83984971e+00
-2.24142876e-02 1.50118396e-01 1.20477393e-01 3.87945294e-01
-1.07436344e-01 4.20354486e-01 -4.63530272e-01 3.51532787e-01
1.07837796e+00 1.05619836e+00 -7.45621204e-01 -5.67063153e-01
-6.87529445e-01 2.73691535e-01 3.79778832e-01 2.62227565e-01
-7.58097231e-01 -4.49195594e-01 9.65365946e-01 4.07401770e-01
2.07704827e-01 1.96194530e-01 -5.58889568e-01 -8.45200956e-01
1.13200448e-01 -1.13778305e+00 7.40667224e-01 -3.79584193e-01
-1.21354246e+00 4.35408831e-01 2.24867105e-01 -6.04231119e-01
7.71388644e-03 -4.16373201e-02 -3.41712266e-01 1.00166893e+00
-1.07038462e+00 -7.65028477e-01 -2.86602408e-01 6.89031482e-01
4.41831797e-02 1.10018656e-01 7.67496169e-01 7.98428178e-01
-2.47872308e-01 8.86441588e-01 2.23129064e-01 2.48213075e-02
1.37096953e+00 -9.43418801e-01 8.14806163e-01 5.21381557e-01
1.63842976e-01 1.35986137e+00 8.04701388e-01 -6.43498302e-01
-1.48380077e+00 -5.98309040e-01 1.15304923e+00 -5.29133916e-01
4.14090067e-01 -4.22944814e-01 -1.04614162e+00 7.76125133e-01
-2.41068918e-02 2.56320108e-02 5.08172035e-01 1.17191386e+00
-8.37287962e-01 -6.04405642e-01 -9.27192211e-01 3.88153255e-01
1.21544898e+00 -5.78616619e-01 -4.21950877e-01 4.87474531e-01
5.05282938e-01 -3.35544705e-01 -9.81860936e-01 3.46087039e-01
8.47366273e-01 -1.38358545e+00 9.76999819e-01 -4.40741062e-01
9.38966572e-02 -3.02028537e-01 1.33498922e-01 -9.36314821e-01
-5.61548650e-01 -4.63512480e-01 3.75075251e-01 1.49757314e+00
4.76954699e-01 -7.24627554e-01 1.14496469e+00 1.09453440e+00
2.13479307e-02 -6.41113162e-01 -5.70682466e-01 -9.40872848e-01
1.96837664e-01 -5.23137808e-01 1.11454105e+00 9.98555958e-01
2.55724221e-01 1.91445097e-01 -1.61550686e-01 -1.49656668e-01
8.38226750e-02 9.47204381e-02 1.41282141e+00 -8.68564069e-01
-1.62748396e-01 -3.24011415e-01 -1.31206274e-01 -1.24644518e+00
-1.24430753e-01 -9.51443255e-01 -1.84696391e-01 -1.44643545e+00
5.25411546e-01 -1.26855636e+00 -3.05849701e-01 6.86854601e-01
-2.30836183e-01 -3.55602503e-02 6.01880103e-02 7.19482899e-01
-5.78827798e-01 -2.46451050e-01 5.06672442e-01 -2.65895218e-01
-2.94394344e-01 -7.88890719e-02 -9.41924810e-01 9.47155580e-02
5.69026649e-01 -8.09652030e-01 3.39923333e-03 -6.13227904e-01
4.75614578e-01 4.88714576e-02 8.37261006e-02 -7.26728857e-01
6.54944479e-01 4.64231567e-03 9.27674770e-02 -5.66536903e-01
-9.84840468e-02 -7.33743131e-01 4.51253653e-01 2.97268033e-01
-3.74600410e-01 5.53376198e-01 3.14832568e-01 1.37270644e-01
-3.40437800e-01 -8.70446861e-03 2.38141954e-01 -1.51412293e-01
-6.35456800e-01 1.51958922e-02 -2.29513526e-01 1.77930668e-01
9.26070511e-01 -1.32653475e-01 -9.13862109e-01 -1.28209412e-01
-9.45215076e-02 5.62458396e-01 8.33078265e-01 2.81395555e-01
-3.86442579e-02 -1.06161129e+00 -8.43305588e-01 1.00608900e-01
5.06385148e-01 -2.80439943e-01 1.86295345e-01 7.61386216e-01
-7.05405533e-01 5.50463080e-01 1.09854631e-01 -6.98583424e-01
-1.50480771e+00 9.68072891e-01 -1.87189385e-01 -3.62476796e-01
-8.81286860e-01 8.90712619e-01 2.64242262e-01 -4.11437452e-01
3.45816553e-01 2.90242970e-01 4.36038256e-01 -1.46875724e-01
6.64549708e-01 6.17443204e-01 6.30671442e-01 -1.94949239e-01
-7.57427871e-01 7.71078765e-02 -5.47089517e-01 -1.62421539e-01
1.19221473e+00 1.29687116e-02 -3.71697932e-01 4.29267064e-02
1.13241959e+00 4.14851010e-01 -4.43911880e-01 -1.51081949e-01
2.88966775e-01 -7.73110032e-01 -2.08811630e-02 -9.33410287e-01
-7.92652071e-01 5.06062448e-01 4.13644612e-01 4.85157758e-01
1.06662881e+00 -8.56956642e-04 9.85918164e-01 3.13506365e-01
4.95257437e-01 -1.12794065e+00 -6.84857428e-01 2.17545465e-01
2.52789110e-01 -9.99678612e-01 2.51031935e-01 -4.56980675e-01
-3.95782560e-01 7.21132100e-01 6.95851207e-01 3.43425758e-02
6.75935984e-01 7.86479652e-01 3.67886692e-01 -3.82344604e-01
-1.18218505e+00 -7.44765699e-02 -2.24570259e-01 2.93138504e-01
9.27447319e-01 2.55642712e-01 -6.69013381e-01 1.18694350e-01
-3.97173077e-01 1.22243211e-01 -4.77644019e-02 1.08267176e+00
-3.31816703e-01 -1.52409720e+00 -7.80655205e-01 7.70311415e-01
-7.12320507e-01 -2.17506006e-01 -1.51508436e-01 9.57563818e-01
-6.43696915e-03 1.17336464e+00 8.44803452e-02 -3.02512497e-01
4.93104309e-01 9.12409276e-02 1.79917857e-01 -4.83208835e-01
-1.48728311e+00 -5.34193292e-02 4.29100454e-01 -7.92914033e-01
-1.29720598e-01 -4.76583034e-01 -9.69757915e-01 -1.07502985e+00
-5.17761707e-01 5.33056021e-01 8.62201095e-01 3.95369947e-01
7.07031012e-01 4.26784717e-02 4.21908796e-01 9.12560895e-03
-5.41931152e-01 -7.29353130e-01 -3.86082709e-01 6.82179689e-01
-1.42905474e-01 -4.53811377e-01 -3.42070580e-01 -3.28894436e-01] | [9.413714408874512, 8.423264503479004] |
2b05f893-e400-4c1f-9e71-f2cfca1ca7f7 | on-computing-universal-plans-for-partially | 2305.16203 | null | https://arxiv.org/abs/2305.16203v2 | https://arxiv.org/pdf/2305.16203v2.pdf | On Computing Universal Plans for Partially Observable Multi-Agent Path Finding | Multi-agent routing problems have drawn significant attention nowadays due to their broad industrial applications in, e.g., warehouse robots, logistics automation, and traffic control. Conventionally, they are modelled as classical planning problems. In this paper, we argue that it is beneficial to formulate them as universal planning problems. We therefore propose universal plans, also known as policies, as the solution concepts, and implement a system called ASP-MAUPF (Answer Set Programming for Multi-Agent Universal Plan Finding) for computing them. Given an arbitrary two-dimensional map and a profile of goals for the agents, the system finds a feasible universal plan for each agent that ensures no collision with others. We use the system to conduct some experiments, and make some observations on the types of goal profiles and environments that will have feasible policies, and how they may depend on agents' sensors. We also demonstrate how users can customize action preferences to compute more efficient policies, even (near-)optimal ones. | ['Fangzhen Lin', 'Fengming Zhu'] | 2023-05-25 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [-3.38842794e-02 5.04659355e-01 -2.05347076e-01 -5.15928388e-01
-3.02647799e-01 -7.99632788e-01 3.51051599e-01 5.04803240e-01
-4.52443480e-01 9.30799961e-01 -4.40780111e-02 -3.73243988e-01
-6.42157912e-01 -1.23666096e+00 -4.42932546e-01 -6.41757309e-01
-4.13807482e-01 1.09680402e+00 5.61522007e-01 -5.02758026e-01
4.39040095e-01 6.09520733e-01 -1.48725486e+00 -4.61852290e-02
7.63191700e-01 6.77820265e-01 7.34973073e-01 5.62203288e-01
-2.37553775e-01 5.03828704e-01 -6.18182600e-01 2.69840751e-02
3.46553952e-01 -1.31182298e-01 -1.24735582e+00 5.18322289e-01
-6.67934835e-01 -2.80372828e-01 2.92204022e-01 1.09118915e+00
1.89294759e-02 3.73188108e-01 3.37933153e-01 -1.82824337e+00
-3.05240631e-01 7.48390615e-01 -3.77424031e-01 -2.63080835e-01
6.69472158e-01 1.21138848e-01 7.48140872e-01 5.92463128e-02
7.30141401e-01 1.37864769e+00 4.43199351e-02 5.23157001e-01
-9.84784484e-01 -2.60903221e-02 5.36874115e-01 3.25004548e-01
-1.18158591e+00 -1.82349950e-01 5.31779468e-01 4.41024490e-02
7.64882743e-01 5.21440566e-01 2.67906338e-01 4.58445132e-01
4.79050606e-01 5.43610096e-01 1.08224559e+00 -5.46778977e-01
6.75300539e-01 1.73691332e-01 -1.08965561e-02 5.06569684e-01
2.66021460e-01 -2.36700282e-01 1.45943239e-01 -2.83029139e-01
7.32654989e-01 -1.22853741e-01 -8.17848593e-02 -7.25224912e-01
-1.17256796e+00 9.20069158e-01 5.12271151e-02 2.06206083e-01
-8.05432320e-01 8.04666877e-02 1.44058466e-01 3.82050365e-01
-1.53505147e-01 7.32475400e-01 -6.58381343e-01 -6.19335379e-03
2.38994211e-01 6.85990989e-01 1.14980638e+00 1.39475298e+00
1.01297319e+00 -4.02244210e-01 1.87692851e-01 5.39140582e-01
3.07143420e-01 3.13483357e-01 1.69488434e-02 -1.27639508e+00
4.48222548e-01 6.33420110e-01 9.34376478e-01 -9.11048770e-01
-8.90818834e-01 3.55267406e-01 -1.83623418e-01 2.43706509e-01
4.25440222e-01 -4.68610913e-01 -5.42373598e-01 1.53296757e+00
5.68703711e-01 4.37544286e-02 4.06228453e-01 9.54245090e-01
9.78058204e-02 8.02310526e-01 9.68561023e-02 -7.03727245e-01
1.43687308e+00 -9.27718878e-01 -6.88782752e-01 -3.85807812e-01
7.82360613e-01 -4.74484742e-01 8.39924276e-01 3.14355016e-01
-1.12282753e+00 8.94965678e-02 -8.33573461e-01 5.16668022e-01
-5.14081717e-01 -5.68662167e-01 7.42006898e-01 5.31431019e-01
-1.05422115e+00 5.39680481e-01 -6.71774387e-01 -7.31787503e-01
-3.23783517e-01 5.41766167e-01 -1.12757809e-01 -2.39933759e-01
-8.28879952e-01 1.12736583e+00 5.95187902e-01 -4.09410559e-02
-8.90848577e-01 2.05801487e-01 -8.09909284e-01 -8.62221867e-02
1.19983613e+00 -4.59238023e-01 1.55716658e+00 -4.62968677e-01
-1.64191353e+00 4.13223356e-01 9.45978016e-02 -1.84613869e-01
6.88655227e-02 3.52089882e-01 -4.17951047e-01 3.17228660e-02
3.31236869e-01 3.81596178e-01 1.40580252e-01 -1.47440636e+00
-1.20276880e+00 -4.69652116e-01 9.14552152e-01 4.09465730e-01
1.21829413e-01 2.63015896e-01 -2.38045692e-01 1.87542409e-01
3.03660899e-01 -1.06202173e+00 -1.26136053e+00 -6.50377452e-01
-5.52864730e-01 -5.13106763e-01 3.08150351e-01 1.79248065e-01
7.46453047e-01 -1.68073165e+00 2.23786145e-01 5.38172483e-01
-2.77547657e-01 -3.22408170e-01 -5.09785235e-01 6.48469985e-01
5.31724572e-01 2.38197464e-02 -6.97757900e-02 -3.11103854e-02
4.83367831e-01 8.01791191e-01 -1.71899665e-02 3.57154936e-01
-2.03272343e-01 4.82136130e-01 -1.04994082e+00 -2.46614024e-01
3.84728283e-01 -3.49141061e-01 -3.46431524e-01 1.72057033e-01
-9.24714744e-01 2.30431184e-01 -1.11337483e+00 4.57512945e-01
6.73972130e-01 1.35599896e-01 8.96234095e-01 3.54718149e-01
-5.39054573e-01 5.23989387e-02 -1.57983935e+00 1.38894928e+00
-2.95485109e-01 -2.17503682e-01 4.89810228e-01 -9.13832366e-01
8.65080893e-01 2.51157433e-01 7.87017226e-01 -6.86593652e-01
2.84199446e-01 2.20839381e-01 -7.21033141e-02 -6.01441026e-01
9.00212824e-01 -1.03966288e-01 -5.03172934e-01 6.70507014e-01
-5.22224545e-01 -1.76334441e-01 5.78223467e-01 -3.00363660e-01
1.11711991e+00 5.19284867e-02 4.55985457e-01 -5.16278803e-01
6.51947796e-01 4.71282333e-01 6.73046470e-01 7.74968207e-01
-1.86023995e-01 -1.26731679e-01 5.60136437e-01 -6.12219632e-01
-4.83272761e-01 -6.05005145e-01 3.05872172e-01 1.05405641e+00
1.03961158e+00 -3.23316991e-01 -5.60793996e-01 -5.00715077e-01
-2.42481217e-01 8.75320375e-01 -1.62800446e-01 2.98953056e-01
-7.64627516e-01 -6.86854839e-01 -1.29060239e-01 1.94061905e-01
2.12975308e-01 -1.38411069e+00 -1.35933483e+00 6.71753943e-01
-2.03739300e-01 -1.22342741e+00 -1.67793393e-01 2.04954922e-01
-5.33358216e-01 -1.39392245e+00 -1.45137817e-01 -6.94986641e-01
7.81154752e-01 5.66594124e-01 8.42335165e-01 1.16109572e-01
2.58399397e-01 6.34505391e-01 -5.80249250e-01 -4.71697539e-01
-4.41317260e-01 7.29854703e-02 2.54625171e-01 -2.03606740e-01
2.18710497e-01 -2.91561425e-01 -2.92512625e-01 7.88622618e-01
-8.95718694e-01 -1.46857724e-01 4.32489812e-01 3.30636762e-02
7.92137206e-01 7.55128324e-01 6.48884237e-01 -8.03788126e-01
1.11843944e+00 -4.93526697e-01 -8.43753040e-01 5.51686168e-01
-5.47938347e-01 4.40395363e-02 6.28266394e-01 1.06600657e-01
-9.92544174e-01 -1.20074384e-01 8.92901570e-02 1.96355388e-01
-5.81082344e-01 6.68473184e-01 -5.40573835e-01 4.61039227e-03
1.96341544e-01 -2.14232326e-01 -1.26525760e-01 -2.98616230e-01
3.99614483e-01 3.49149704e-01 2.55777180e-01 -1.02559578e+00
4.45534021e-01 2.49809653e-01 1.99070916e-01 -5.90280592e-01
-2.87166566e-01 -3.61031473e-01 -1.54824093e-01 -2.49215573e-01
8.45036924e-01 -1.97934121e-01 -1.24638009e+00 -5.05816191e-02
-1.15127420e+00 -5.94265282e-01 -1.34873286e-01 2.27645069e-01
-1.14142799e+00 5.35191782e-02 -3.10653508e-01 -1.07971907e+00
3.24124873e-01 -1.51817930e+00 6.11683965e-01 4.50737119e-01
-7.56626055e-02 -8.31382871e-01 6.76690266e-02 -1.13501720e-01
2.94399709e-01 3.22008342e-01 1.02135229e+00 -6.38574898e-01
-8.53215277e-01 1.84438393e-01 1.44724980e-01 -4.62237984e-01
2.31699809e-01 -3.35235745e-01 -3.01786631e-01 -1.77275851e-01
-2.29620427e-01 2.24269088e-02 -7.06827044e-02 5.60371220e-01
8.44773710e-01 -6.49936020e-01 -7.83533096e-01 -3.70187573e-02
1.63701153e+00 9.82417762e-01 6.49296284e-01 8.74134839e-01
-2.34098174e-02 1.05440462e+00 1.30262589e+00 7.01219738e-01
8.29995573e-01 9.03134048e-01 9.15906191e-01 3.99181217e-01
8.35991323e-01 2.35421315e-01 2.45221570e-01 2.59101450e-01
-3.53039265e-01 -5.26685417e-01 -7.52720952e-01 4.30866957e-01
-2.31303573e+00 -7.43832707e-01 -1.39391767e-02 2.05603456e+00
2.85055906e-01 -8.69914144e-03 4.09731925e-01 -1.91912517e-01
8.04878294e-01 -2.30852123e-02 -4.43622142e-01 -8.10183406e-01
1.32399723e-01 -2.37646624e-01 7.66601920e-01 7.24453747e-01
-1.03369033e+00 8.23941767e-01 5.91666222e+00 1.71164647e-01
-5.07722616e-01 2.48147566e-02 3.54546517e-01 3.78930032e-01
-4.08018321e-01 5.55359125e-02 -5.38039565e-01 9.48013067e-02
7.74643958e-01 -3.74878377e-01 7.64064074e-01 8.30376625e-01
3.23565930e-01 -4.79451090e-01 -1.02579689e+00 4.56897527e-01
-4.97279167e-01 -1.23423433e+00 -4.05292094e-01 2.10043505e-01
5.20920813e-01 -3.17662448e-01 -3.98095429e-01 2.10231915e-01
9.21620429e-01 -6.53096974e-01 8.35559726e-01 3.52914967e-02
5.99095747e-02 -1.19481635e+00 8.91800463e-01 7.40238190e-01
-1.18528569e+00 -3.87419194e-01 -5.93786299e-01 -3.17853868e-01
6.49738431e-01 9.66256484e-02 -9.05260682e-01 1.04940069e+00
4.44446653e-01 -4.34607677e-02 1.93887129e-01 9.72733080e-01
-1.69539064e-01 -4.26987410e-01 -5.58186412e-01 -4.11474645e-01
4.60909367e-01 -5.54122210e-01 4.72082853e-01 5.00779748e-01
4.18353021e-01 5.06394029e-01 9.39987421e-01 5.14599144e-01
3.95294100e-01 8.30970779e-02 -6.73965096e-01 1.76956311e-01
6.23468280e-01 1.12754858e+00 -1.25404549e+00 -1.09389154e-02
-3.53814512e-01 5.44416010e-01 6.24100342e-02 3.64891559e-01
-8.67982626e-01 -2.38843292e-01 9.86408174e-01 -1.98136464e-01
-8.69540423e-02 -2.53977358e-01 -1.31912142e-01 -5.43440759e-01
-1.01397820e-01 -6.86427772e-01 5.67732871e-01 -6.67791307e-01
-7.94398308e-01 6.47579074e-01 3.77184689e-01 -9.77107584e-01
-4.73978847e-01 -5.45615256e-01 -4.72960502e-01 4.28573698e-01
-1.64999855e+00 -7.32409596e-01 -5.41618764e-02 6.05251908e-01
5.20367801e-01 3.70122902e-02 9.84413326e-01 -1.03148185e-01
-4.61418301e-01 -2.16749102e-01 -4.02987659e-01 -5.50317466e-01
1.64775681e-02 -1.07053268e+00 -1.20595589e-01 6.88097179e-01
-3.18031043e-01 3.92561018e-01 1.18158734e+00 -5.20963192e-01
-1.77289069e+00 -9.79372859e-01 5.40311337e-01 -2.75691189e-02
5.63994467e-01 2.66868412e-01 -3.94006938e-01 1.18688297e+00
5.07343590e-01 -5.95927775e-01 2.76896745e-01 8.86288658e-02
4.97814924e-01 -9.47812125e-02 -1.47202599e+00 1.04661489e+00
1.08190691e+00 4.50893253e-01 -2.68303752e-01 6.01744950e-01
8.67065787e-01 -3.57781410e-01 -6.90364659e-01 3.58439773e-01
1.94915816e-01 -9.82101738e-01 6.31776452e-01 -6.37312591e-01
-4.93387468e-02 -5.58799148e-01 -3.97602588e-01 -1.61483264e+00
-4.69637394e-01 -7.29787707e-01 3.84303719e-01 9.79015350e-01
4.72858638e-01 -1.08439517e+00 8.16655576e-01 1.14195490e+00
-4.22829866e-01 -7.00292528e-01 -9.30896103e-01 -8.43820214e-01
-2.76948184e-01 -3.94494951e-01 1.37174892e+00 8.05547655e-01
4.77867961e-01 1.30359337e-01 -2.01800406e-01 7.53533065e-01
5.12463272e-01 5.50681949e-01 7.87657559e-01 -1.14051247e+00
-1.23639479e-01 -2.80308068e-01 -2.63843313e-02 -1.06181347e+00
2.23768145e-01 -4.70852077e-01 4.44586903e-01 -1.92129624e+00
-3.32285315e-01 -1.26560092e+00 -1.06843285e-01 6.32578254e-01
5.50825059e-01 -5.73490560e-01 4.24679637e-01 -4.59875166e-02
-1.03352427e+00 1.64716467e-01 1.43248343e+00 4.94691469e-02
-5.43728232e-01 3.19358200e-01 -9.06094968e-01 7.08903491e-01
1.30328798e+00 -2.48414308e-01 -6.65091217e-01 -5.51980019e-01
3.36521924e-01 7.25664318e-01 -2.34516278e-01 -6.78258359e-01
4.19905961e-01 -1.27572274e+00 -5.94839871e-01 -4.13672060e-01
3.75068039e-01 -1.26996851e+00 4.05028075e-01 5.32746375e-01
-1.52088627e-01 4.91647929e-01 -8.76406580e-03 3.93555462e-01
3.52824554e-02 -7.24008858e-01 3.42696428e-01 -4.99091148e-01
-1.11764622e+00 2.13402867e-01 -7.16905773e-01 -5.49545586e-01
1.85455203e+00 -1.23373836e-01 -5.88621795e-01 -3.67562801e-01
-7.22805142e-01 9.11818624e-01 5.31411648e-01 2.23097593e-01
6.51043594e-01 -9.75034654e-01 -3.65156233e-01 -1.27310231e-01
1.54122133e-02 2.06389189e-01 1.16037913e-01 5.99113166e-01
-4.60631788e-01 6.31219625e-01 -3.37336332e-01 -2.07003653e-01
-8.91511858e-01 9.19841230e-01 1.68453738e-01 -4.07755315e-01
-4.28221881e-01 2.95451880e-01 4.39874418e-02 -5.71934521e-01
1.28240168e-01 -4.21590000e-01 -3.59071970e-01 -2.25847095e-01
5.92394173e-01 4.94916052e-01 -2.46572286e-01 -2.77346432e-01
-5.69185317e-01 4.00450677e-01 1.54472068e-01 -2.16157883e-01
1.38128972e+00 -4.18499768e-01 -3.96639258e-01 -8.34075455e-03
2.58325785e-01 -9.30276960e-02 -9.01073039e-01 -4.49918471e-02
4.44136351e-01 -4.60368693e-01 -5.53516090e-01 -5.57354391e-01
-7.95155525e-01 5.12925722e-02 7.52940848e-02 1.08808589e+00
1.13961995e+00 2.75697112e-01 5.02701163e-01 6.07539177e-01
1.49372268e+00 -1.16541874e+00 -2.34022796e-01 6.19426847e-01
7.74457633e-01 -7.16348708e-01 -1.68057099e-01 -7.93169320e-01
-7.09377408e-01 1.07040071e+00 7.15867996e-01 -5.64559512e-02
1.98681861e-01 4.64623213e-01 9.95118096e-02 -2.98895538e-01
-8.72510433e-01 -4.91429090e-01 -8.79330754e-01 9.43110943e-01
-4.63436812e-01 6.39880598e-01 -5.27220786e-01 4.79290992e-01
3.05718593e-02 -6.11416139e-02 1.10428929e+00 1.26852775e+00
-9.72579598e-01 -1.48181355e+00 -6.84436440e-01 2.39195690e-01
1.95503514e-02 7.27635682e-01 -3.81563939e-02 8.46567869e-01
9.60327014e-02 1.52514338e+00 6.45919517e-02 -3.43582064e-01
7.95458913e-01 -3.39745939e-01 5.68205535e-01 -7.52603173e-01
-2.65956759e-01 -1.16015173e-01 6.84364676e-01 -6.77977264e-01
-6.86774492e-01 -6.80244505e-01 -1.75167000e+00 -4.36274499e-01
-4.08169962e-02 4.41973656e-01 4.49680030e-01 1.02045214e+00
1.36032358e-01 5.09433329e-01 6.00773454e-01 -7.62146831e-01
-5.69169044e-01 -5.18420637e-01 -7.59726286e-01 -1.05226874e-01
-1.43790334e-01 -8.23815882e-01 1.55866608e-01 -3.92660707e-01] | [4.908859729766846, 1.6509590148925781] |
0ea369e4-ed3f-4700-9282-3d27187a2b2d | super-resolution-and-image-re-projection-for | 2210.11129 | null | https://arxiv.org/abs/2210.11129v1 | https://arxiv.org/pdf/2210.11129v1.pdf | Super-Resolution and Image Re-projection for Iris Recognition | Several recent works have addressed the ability of deep learning to disclose rich, hierarchical and discriminative models for the most diverse purposes. Specifically in the super-resolution field, Convolutional Neural Networks (CNNs) using different deep learning approaches attempt to recover realistic texture and fine grained details from low resolution images. In this work we explore the viability of these approaches for iris Super-Resolution (SR) in an iris recognition environment. For this, we test different architectures with and without a so called image re-projection to reduce artifacts applying it to different iris databases to verify the viability of the different CNNs for iris super-resolution. Results show that CNNs and image re-projection can improve the results specially for the accuracy of recognition systems using a complete different training database performing the transfer learning successfully. | ['Fernando Alonso-Fernandez', 'Andreas Uhl', 'Eduardo Ribeiro'] | 2022-10-20 | null | null | null | null | ['iris-recognition'] | ['computer-vision'] | [ 2.63498336e-01 1.02030829e-01 1.43729514e-02 -4.90945727e-01
-6.58285677e-01 -9.14816260e-02 5.78900099e-01 -4.38551664e-01
-1.67650670e-01 7.85114288e-01 4.98311937e-01 2.46509522e-01
-4.20032322e-01 -7.39198625e-01 -5.54117620e-01 -6.77400410e-01
1.87272370e-01 4.25767899e-01 2.93424409e-02 -3.62165868e-01
3.03691983e-01 1.04010952e+00 -2.08016229e+00 8.66282701e-01
7.02409446e-01 5.19781768e-01 -3.22050780e-01 8.00632000e-01
2.24649921e-01 7.45440185e-01 -6.22644603e-01 -2.49418288e-01
5.76619685e-01 -3.10826898e-01 -9.03542578e-01 -3.55120674e-02
1.39892364e+00 -6.11793756e-01 -1.14663601e-01 1.01634300e+00
7.60624051e-01 -1.15153879e-01 2.32120425e-01 -1.20466918e-01
-8.33274782e-01 5.10371327e-01 -6.60236657e-01 5.48340023e-01
4.41510409e-01 1.61145523e-01 3.24545860e-01 -5.61096728e-01
7.31983781e-01 1.27043045e+00 8.10160697e-01 6.58912420e-01
-1.49514484e+00 -6.17435753e-01 -7.86976933e-01 -1.57758277e-02
-1.39638937e+00 -6.03596628e-01 3.62875015e-01 -3.62326622e-01
7.06345260e-01 3.21985900e-01 3.05063099e-01 1.12251019e+00
6.78721294e-02 -9.05524269e-02 1.95186079e+00 -4.29292202e-01
-2.00371370e-01 3.70581895e-01 5.65337650e-02 5.59574366e-01
3.85163218e-01 8.43322754e-01 -5.30111372e-01 2.50625491e-01
1.38191235e+00 -3.50201488e-01 -4.98774610e-02 1.37030259e-01
-9.71938908e-01 5.07956922e-01 5.30019343e-01 7.80991375e-01
-2.45338872e-01 -3.81152868e-01 9.04406905e-02 3.44184816e-01
5.07006347e-01 8.79957557e-01 -1.92670092e-01 1.91194698e-01
-1.22261119e+00 5.89611717e-02 4.45462197e-01 3.78866762e-01
5.67893386e-01 2.62048781e-01 -6.40154839e-01 7.86576271e-01
-3.69795531e-01 -1.62725702e-01 5.64756751e-01 -6.87060833e-01
3.71251628e-02 5.02951264e-01 4.89653908e-02 -8.62134159e-01
-5.38862586e-01 -1.00003278e+00 -1.12000525e+00 6.97791696e-01
5.93332469e-01 7.95639902e-02 -1.27453983e+00 1.17543864e+00
2.48262927e-01 6.18822277e-01 2.14615270e-01 1.16993988e+00
1.21830475e+00 2.23605148e-02 3.12187560e-02 2.26150975e-01
1.18929577e+00 -6.68510377e-01 -4.15724546e-01 3.05662811e-01
4.55439538e-02 -1.09541178e+00 9.31695700e-01 5.32305062e-01
-1.22049892e+00 -1.22026384e+00 -9.89028811e-01 -3.90965074e-01
-1.97908401e-01 5.56120455e-01 3.56002629e-01 6.72391713e-01
-1.36379564e+00 1.10260415e+00 -5.72847068e-01 -4.22393084e-01
6.00575030e-01 7.01147616e-01 -6.60837650e-01 -7.78788999e-02
-8.63843501e-01 9.19967532e-01 3.82722974e-01 5.63361943e-02
-6.37077391e-01 -6.68030858e-01 -5.39200366e-01 1.79180488e-01
-3.56707215e-01 -7.32506037e-01 5.25057554e-01 -1.18771875e+00
-1.76996315e+00 1.40254533e+00 -7.05556870e-02 -6.76871479e-01
3.40361446e-01 -1.78894103e-01 -4.03054953e-01 6.96584135e-02
-3.19071919e-01 3.43497694e-01 1.07592845e+00 -1.18913710e+00
-5.19196332e-01 -5.65267265e-01 6.46418780e-02 -5.54867573e-02
-3.86835262e-02 3.70632738e-01 5.76785244e-02 -5.92129230e-01
4.67647500e-02 -7.63668060e-01 -9.32727531e-02 -5.53614259e-01
-3.61319929e-01 1.37727767e-01 3.54281098e-01 -9.22259808e-01
7.28576303e-01 -2.02499032e+00 2.98455209e-01 -7.42361546e-02
2.79843420e-01 7.54380345e-01 -3.11561286e-01 -1.03428222e-01
-4.68769372e-01 1.68455213e-01 8.82318541e-02 -4.45949346e-01
-7.49004066e-01 1.49883792e-01 1.21812234e-02 3.85100693e-01
2.81156600e-01 7.24308670e-01 -2.74688184e-01 -3.73148799e-01
5.10145068e-01 1.16753078e+00 -2.23624319e-01 1.26413688e-01
2.30025336e-01 1.17540848e+00 -1.18309140e-01 5.49287260e-01
1.02810550e+00 -3.10140163e-01 -6.07934520e-02 -4.49383706e-01
-2.16644570e-01 -6.36031926e-02 -1.19700265e+00 1.55972064e+00
-4.02644068e-01 5.81696689e-01 -5.43463789e-02 -7.70411968e-01
9.66841221e-01 3.64879012e-01 2.17410535e-01 -1.04850852e+00
-6.62479624e-02 9.91764739e-02 -2.78659314e-02 -3.51452202e-01
6.06102467e-01 -3.29959780e-01 7.22183049e-01 4.67677042e-02
2.21251056e-01 3.23076814e-01 -2.90996343e-01 -5.26514709e-01
5.40544152e-01 1.63553566e-01 1.96122140e-01 -3.04426700e-01
7.75293767e-01 -9.35717672e-02 2.86287427e-01 7.10360944e-01
6.94293007e-02 1.12762225e+00 1.32141158e-01 -9.82854784e-01
-1.45265818e+00 -6.63567781e-01 -6.60889924e-01 6.61215663e-01
-2.46544167e-01 1.99114740e-01 -8.33704889e-01 -2.07426652e-01
-2.01991841e-01 1.33910745e-01 -8.32445085e-01 2.34623164e-01
-6.64970994e-01 -9.81344938e-01 5.68556666e-01 2.02760994e-01
7.36623108e-01 -9.86394048e-01 -4.78345096e-01 -1.41145572e-01
1.65089399e-01 -1.30573869e+00 3.39700460e-01 -3.74608994e-01
-1.06461918e+00 -1.10512638e+00 -7.28466094e-01 -6.70370162e-01
5.54394782e-01 -2.43883893e-01 1.19157171e+00 2.31194139e-01
-6.94858134e-01 -3.68171558e-02 -2.25445658e-01 -3.93172614e-02
-5.38896501e-01 1.64278314e-01 -1.86598785e-02 2.89912194e-01
3.44339818e-01 -6.57850564e-01 -5.97511411e-01 1.24354236e-01
-7.87197351e-01 5.34267529e-05 8.97285879e-01 8.28296125e-01
7.03275681e-01 2.22766533e-01 3.26506607e-02 -1.02757883e+00
5.23419261e-01 7.16890022e-02 -8.77676129e-01 9.24850777e-02
-6.37716711e-01 2.15293452e-01 4.77147132e-01 -2.46458158e-01
-1.08790290e+00 -2.22059265e-02 -8.40023905e-02 -4.62816387e-01
-7.66029775e-01 5.36307432e-02 3.79778266e-01 -7.39999712e-01
1.14726961e+00 2.41360351e-01 -2.69210916e-02 -9.13891315e-01
1.52500585e-01 5.00965595e-01 6.19970500e-01 -2.86997288e-01
9.17550325e-01 7.01640606e-01 5.24234295e-01 -7.19024062e-01
-9.35114086e-01 -1.57122537e-01 -1.06052649e+00 2.96633542e-01
1.01823056e+00 -9.41721618e-01 -6.52052879e-01 4.81845289e-01
-7.50667274e-01 -7.74132982e-02 -4.47068900e-01 6.05884194e-01
-3.63912135e-01 2.20200717e-01 -7.89033234e-01 -4.17118758e-01
-2.94672996e-01 -1.29367757e+00 1.07220685e+00 8.75316143e-01
3.38842332e-01 -8.29016507e-01 2.85084933e-01 8.12494993e-01
8.74884427e-01 6.35276139e-01 5.93729019e-01 -2.90998191e-01
-8.03038299e-01 5.57444729e-02 -6.40109837e-01 4.35158879e-01
2.04008833e-01 -5.83481826e-02 -1.47197163e+00 -5.13376653e-01
-7.92964771e-02 -1.52145743e-01 9.87449586e-01 5.34194827e-01
1.12755537e+00 -3.61176729e-01 1.02734238e-01 1.26847744e+00
1.83359993e+00 -1.50947973e-01 1.14915240e+00 4.61559862e-01
5.05109668e-01 7.01238632e-01 2.22516567e-01 3.40664349e-02
-2.45049164e-01 9.31977093e-01 3.05605620e-01 -6.51752234e-01
-8.01498592e-01 -2.25799484e-03 -2.04004154e-01 1.09691836e-01
-8.63352060e-01 3.91701072e-01 -6.50720417e-01 4.73304570e-01
-1.24294412e+00 -1.10025334e+00 -3.90920863e-02 2.33373690e+00
9.06790972e-01 -2.76823282e-01 -5.40161319e-02 -1.83932617e-01
6.93974018e-01 -3.80218178e-02 -3.19958091e-01 -4.85917091e-01
-3.40441793e-01 1.11516619e+00 4.87840712e-01 7.02453852e-01
-1.03602231e+00 1.09789693e+00 6.59615135e+00 7.10383117e-01
-1.48460853e+00 5.39637990e-02 7.38458872e-01 -1.33574292e-01
3.06448072e-01 -3.03443253e-01 -9.55617070e-01 4.59278710e-02
1.15811920e+00 4.43202913e-01 5.61411977e-01 4.80981350e-01
1.24787435e-01 -7.98463728e-03 -7.44458735e-01 1.00605309e+00
6.96782619e-02 -1.61525202e+00 2.19776079e-01 1.95198655e-01
9.84812617e-01 6.81854114e-02 6.48034751e-01 -2.07414344e-01
1.65168241e-01 -1.90031922e+00 -3.57458532e-01 9.68934059e-01
1.29186594e+00 -7.29431927e-01 1.06790280e+00 -1.77895620e-01
-8.00043285e-01 2.80778427e-02 -6.37964666e-01 4.85721976e-02
-5.60863078e-01 4.11880821e-01 -1.05298448e+00 7.07067490e-01
9.21943069e-01 5.48061252e-01 -1.14998627e+00 9.94743407e-01
1.25580728e-01 3.97842407e-01 -3.08514088e-02 7.33528376e-01
-6.24198839e-02 -1.96404055e-01 6.80883110e-01 1.05391085e+00
1.30488142e-01 3.95968370e-02 -4.04117852e-01 1.05208945e+00
2.23938257e-01 5.79071157e-02 -3.64011556e-01 1.90617874e-01
-9.50136259e-02 1.28201663e+00 -4.36983705e-01 -1.27588466e-01
-2.02220604e-01 8.39439154e-01 2.75521964e-01 3.98041964e-01
-2.53183663e-01 4.42562103e-02 7.01972604e-01 5.45764685e-01
3.06978583e-01 2.99062040e-02 -5.48213840e-01 -1.16234577e+00
-2.64925927e-01 -1.08659542e+00 2.16104999e-01 -8.38732719e-01
-1.07040632e+00 1.03813636e+00 -2.79116780e-01 -1.11430192e+00
-1.76651701e-01 -4.76985663e-01 -2.72994936e-01 1.48803496e+00
-1.91110265e+00 -1.57486236e+00 -6.61011219e-01 8.81608903e-01
3.08375478e-01 -6.18255913e-01 8.13979447e-01 1.97494745e-01
-4.90585476e-01 6.34562373e-01 -2.79406514e-02 2.33716324e-01
8.21107209e-01 -1.26295829e+00 1.00865588e-01 9.36340690e-01
2.93179244e-01 6.92839444e-01 6.73704565e-01 -4.41438586e-01
-8.24977458e-01 -9.67451870e-01 5.10945857e-01 -6.14219785e-01
-1.00475825e-01 1.33065775e-01 -1.12811780e+00 5.44536591e-01
3.15603971e-01 2.50631839e-01 5.67420840e-01 4.11411971e-01
-3.88452172e-01 -1.68639317e-01 -1.56080651e+00 1.71077758e-01
5.65032125e-01 -6.42427206e-01 -5.99311709e-01 2.50186652e-01
2.74185985e-01 -7.92861700e-01 -1.35458803e+00 7.33100057e-01
4.79761362e-01 -1.70120990e+00 1.24499929e+00 -8.81451666e-01
7.55036533e-01 -3.36580276e-01 9.55711901e-02 -8.72825205e-01
-5.22749484e-01 -3.06466371e-01 1.79122731e-01 8.28661621e-01
1.70285314e-01 -5.53610563e-01 8.90929520e-01 4.37119395e-01
2.54128665e-01 -4.81357843e-01 -9.67926979e-01 -3.46134871e-01
1.53613642e-01 5.22365570e-01 7.70445883e-01 1.05998695e+00
-8.93715143e-01 7.85044432e-02 -5.88290334e-01 6.96620524e-01
7.68438578e-01 2.68838733e-01 8.22946250e-01 -1.53552961e+00
-3.47343534e-01 -2.71416634e-01 -5.95769048e-01 -4.83658276e-02
-1.58100531e-01 -6.14906907e-01 -6.94741666e-01 -1.20201492e+00
2.64269620e-01 -2.55089104e-01 -2.38618642e-01 3.85358512e-01
8.48689675e-02 7.14876115e-01 -1.65477410e-01 5.17214656e-01
-4.86370288e-02 -1.99313134e-01 1.47523332e+00 1.65545985e-01
-3.20895940e-01 9.84166712e-02 -5.91365874e-01 4.09176797e-01
8.31222653e-01 -1.16534926e-01 -5.03965653e-02 -2.96863347e-01
4.01738919e-02 2.82698452e-01 6.88645303e-01 -1.27256346e+00
1.12392686e-01 2.84319043e-01 8.50264072e-01 -1.45544231e-01
2.73981214e-01 -6.26844466e-01 3.63339543e-01 2.58809894e-01
-4.38876539e-01 -4.05017018e-01 5.69142878e-01 1.17064036e-01
-5.79059124e-01 1.04875349e-01 1.40352190e+00 -4.27913934e-01
-4.33857143e-01 2.57660866e-01 4.40164715e-01 -3.63202512e-01
6.36176646e-01 -5.08106947e-01 -4.51470613e-01 4.04362604e-02
-1.33189964e+00 -5.71828663e-01 7.33868718e-01 3.81606668e-01
6.97522402e-01 -8.10230553e-01 -1.12919104e+00 6.55203402e-01
-1.32889748e-01 -6.17802404e-02 5.46718180e-01 8.31282139e-01
-7.37636924e-01 6.02988660e-01 -9.36135828e-01 -5.94966888e-01
-1.78357089e+00 6.70592368e-01 1.05120635e+00 -4.70375001e-01
-7.83087432e-01 7.28418648e-01 1.38861593e-02 -2.02855065e-01
-1.01352558e-01 -1.80860862e-01 -8.37927282e-01 -3.33075434e-01
7.83889055e-01 2.12477148e-01 1.96991056e-01 -6.00951076e-01
2.25950330e-01 1.10562050e+00 -3.14987093e-01 2.72851229e-01
1.40625727e+00 -3.17907706e-02 -4.36654925e-01 -1.56200767e-01
8.44628751e-01 9.95558351e-02 -1.00602889e+00 -3.76381338e-01
-1.92897081e-01 -8.09724927e-01 3.77374887e-01 -1.07934308e+00
-1.28802049e+00 8.54278445e-01 1.35362911e+00 -1.93957403e-01
1.41196978e+00 -2.72594810e-01 1.84271231e-01 -1.81819871e-01
3.68446290e-01 -6.64657950e-01 -2.90007442e-01 1.05951838e-01
8.71452868e-01 -1.41055620e+00 1.45443320e-01 -2.73085266e-01
-3.95227283e-01 1.45840514e+00 5.62902033e-01 -3.84853423e-01
4.68853056e-01 1.54433340e-01 1.10869057e-01 -3.26905161e-01
-2.31976688e-01 -4.89396542e-01 7.63879180e-01 8.06609631e-01
6.38983488e-01 1.21414792e-02 -1.25460446e-01 1.49187865e-02
-3.66809398e-01 3.86802375e-01 7.97453344e-01 6.46146536e-02
-8.05088133e-02 -1.10241544e+00 -4.48755085e-01 3.04399908e-01
-7.45481908e-01 -1.93575546e-01 -2.62119979e-01 7.62865543e-01
3.85761857e-01 6.30080819e-01 -6.40985509e-03 -3.86513919e-01
1.68410242e-01 -2.35942066e-01 6.79651618e-01 -7.80616641e-01
-9.98219550e-01 -1.13290966e-01 -1.24522470e-01 -8.75769258e-01
-8.34156513e-01 -3.86454076e-01 -3.98454756e-01 -2.89805293e-01
9.97208953e-02 -2.59817094e-01 5.32849431e-01 6.92426562e-01
5.44179499e-01 6.10660553e-01 3.76958489e-01 -6.73337340e-01
-3.70238692e-01 -1.18701053e+00 -6.82139993e-01 4.81814414e-01
8.13810706e-01 -4.15532053e-01 -2.66925544e-01 -1.53598739e-02] | [3.767946243286133, -3.624748706817627] |
1c9adb6e-6a9c-4576-991e-97b3db92a4b0 | improving-action-quality-assessment-using | 2102.10555 | null | https://arxiv.org/abs/2102.10555v2 | https://arxiv.org/pdf/2102.10555v2.pdf | Improving Action Quality Assessment using Weighted Aggregation | Action quality assessment (AQA) aims at automatically judging human action based on a video of the said action and assigning a performance score to it. The majority of works in the existing literature on AQA divide RGB videos into short clips, transform these clips to higher-level representations using Convolutional 3D (C3D) networks, and aggregate them through averaging. These higher-level representations are used to perform AQA. We find that the current clip level feature aggregation technique of averaging is insufficient to capture the relative importance of clip level features. In this work, we propose a learning-based weighted-averaging technique. Using this technique, better performance can be obtained without sacrificing too much computational resources. We call this technique Weight-Decider(WD). We also experiment with ResNets for learning better representations for action quality assessment. We assess the effects of the depth and input clip size of the convolutional neural network on the quality of action score predictions. We achieve a new state-of-the-art Spearman's rank correlation of 0.9315 (an increase of 0.45%) on the MTL-AQA dataset using a 34 layer (2+1)D ResNet with the capability of processing 32 frame clips, with WD aggregation. | ['Md. Bakhtiar Hasan', 'Hasibul Himel', 'Moshiur Farazi', 'Md. Hasanul Kabir', 'Fakhruddin Gazzali', 'Shafkat Farabi'] | 2021-02-21 | null | null | null | null | ['action-quality-assessment'] | ['computer-vision'] | [ 2.38925576e-01 -1.45598650e-01 -9.15088579e-02 -4.38401163e-01
-7.15001225e-01 -2.28035539e-01 3.97238374e-01 1.33487806e-01
-5.87992251e-01 3.98845315e-01 4.52494711e-01 1.30524442e-01
-2.36698642e-01 -8.15274477e-01 -4.23765153e-01 -5.03398955e-01
-3.24464679e-01 -2.38335222e-01 5.43208599e-01 -8.63026232e-02
3.55384976e-01 6.20643318e-01 -1.85833549e+00 8.38220596e-01
6.11408353e-01 1.64267194e+00 -3.37305307e-01 1.04192996e+00
6.53726282e-03 1.24726665e+00 -9.01824832e-01 -3.09866101e-01
6.42129362e-01 -5.96990585e-01 -8.71986151e-01 1.21163137e-01
7.41765797e-01 -6.66227460e-01 -4.50476110e-01 6.39335990e-01
5.22696316e-01 2.66815931e-01 4.53632623e-01 -1.32650936e+00
-2.41448373e-01 4.30000573e-01 -4.06807095e-01 5.34326971e-01
5.41394055e-01 3.86464030e-01 1.15377367e+00 -6.80725217e-01
4.07499909e-01 1.24633527e+00 6.55814886e-01 3.49480003e-01
-7.89915800e-01 -5.11339009e-01 -1.10924497e-01 8.35485458e-01
-1.08025932e+00 -3.05975944e-01 6.02652788e-01 -2.80441701e-01
1.28699470e+00 6.37660027e-02 1.00738406e+00 7.16048956e-01
4.91478503e-01 8.15503120e-01 1.09459496e+00 -2.41286784e-01
4.88795698e-01 -6.60818219e-01 -9.24565345e-02 6.92143798e-01
8.19557384e-02 -1.88383520e-01 -9.14297640e-01 2.10488722e-01
7.41696417e-01 -3.53853814e-02 1.54866591e-01 -1.42019898e-01
-1.15858662e+00 4.86244529e-01 5.50487578e-01 4.06365752e-01
-7.27657616e-01 5.16216576e-01 5.75903893e-01 5.16673446e-01
4.59859908e-01 6.30604863e-01 -4.49674010e-01 -5.11493146e-01
-1.07062519e+00 4.00274009e-01 4.37939882e-01 3.14482868e-01
5.22770703e-01 7.42578730e-02 -5.84052265e-01 6.19365335e-01
-6.69012368e-02 2.82411933e-01 6.83480561e-01 -1.66096675e+00
2.24597439e-01 8.92738700e-01 1.86953872e-01 -1.05408847e+00
-6.49641752e-01 -1.19927965e-01 -5.48063159e-01 1.01290882e+00
6.60879254e-01 -7.06963241e-02 -8.88612032e-01 1.47328115e+00
6.12314418e-02 2.71541893e-01 1.46208890e-02 1.04910886e+00
5.88884115e-01 4.38892305e-01 1.43180519e-01 -1.64998516e-01
1.17790711e+00 -8.52416098e-01 -6.37471199e-01 1.97862878e-01
6.24596119e-01 -5.71972370e-01 1.21226120e+00 8.41496944e-01
-1.30282605e+00 -9.17544246e-01 -1.39701855e+00 -9.09917504e-02
-3.68105829e-01 2.19550729e-01 5.73460937e-01 5.50729573e-01
-1.14090538e+00 1.26873565e+00 -8.56121600e-01 -2.26349995e-01
6.53266668e-01 3.93578619e-01 -4.80446965e-01 -1.00227157e-02
-1.24252284e+00 1.05200756e+00 2.05339432e-01 -2.16290638e-01
-8.13379407e-01 -7.88806140e-01 -5.71149290e-01 2.67315879e-02
2.21380100e-01 -5.37849307e-01 1.19902909e+00 -1.14359415e+00
-1.58140361e+00 6.58753991e-01 1.26454636e-01 -8.56410027e-01
6.07636452e-01 -4.86533374e-01 -4.85034823e-01 4.81502950e-01
-2.73380950e-02 8.55834007e-01 7.98203349e-01 -5.76808631e-01
-1.00428236e+00 -4.92999673e-01 6.87786698e-01 1.49918020e-01
-3.88587326e-01 -5.76040149e-02 -1.89003110e-01 -5.78962266e-01
1.03533035e-02 -7.98513651e-01 -9.50816199e-02 5.04931271e-01
3.28215778e-01 -4.52267289e-01 5.06059766e-01 -6.32851005e-01
1.45270133e+00 -2.07543802e+00 8.36092159e-02 1.66196600e-02
2.56369084e-01 5.16962051e-01 -3.73377830e-01 1.20119266e-01
3.69824581e-02 -1.33087263e-01 5.96028566e-02 -2.24244446e-01
-7.32728988e-02 4.20173332e-02 2.15740025e-01 3.83113921e-01
4.79468316e-01 5.89636981e-01 -8.59421968e-01 -5.30813396e-01
4.49103564e-01 5.56915760e-01 -6.41860664e-01 1.76064402e-01
1.39267626e-03 -4.90575284e-03 -2.66601205e-01 6.06816530e-01
3.57546210e-01 7.01686516e-02 -1.56173080e-01 -5.56160152e-01
-1.05254315e-01 1.54813737e-01 -1.28451645e+00 1.97019684e+00
-5.10082841e-01 9.13247645e-01 -5.23232758e-01 -6.30154848e-01
1.00341547e+00 1.77874252e-01 9.66881275e-01 -1.01356006e+00
2.42967099e-01 1.18474863e-01 1.66396677e-01 -6.70931339e-01
4.96947587e-01 1.58179179e-01 -5.65591641e-02 4.01840955e-01
2.72224635e-01 6.89219832e-02 4.91408288e-01 1.38948057e-02
1.65626895e+00 3.09355646e-01 3.78369749e-01 -7.35488161e-02
4.77272123e-01 -2.12527528e-01 3.65389705e-01 5.47849596e-01
-6.75650656e-01 6.22877240e-01 6.42983139e-01 -8.00125420e-01
-1.19089639e+00 -8.17291379e-01 2.83827126e-01 1.20054018e+00
-1.24897078e-01 -5.08928359e-01 -8.73670399e-01 -6.52059495e-01
-1.34269586e-02 4.81669873e-01 -9.29508150e-01 -4.61556584e-01
-5.98088980e-01 -3.07577461e-01 6.06795251e-01 8.86516511e-01
7.90500104e-01 -1.28254652e+00 -1.25257158e+00 3.19074214e-01
8.07743240e-03 -1.01729572e+00 -1.64410323e-02 -1.21825844e-01
-8.99185658e-01 -1.23285174e+00 -6.64333105e-01 -8.07247758e-02
2.44295597e-01 2.03242287e-01 1.11581242e+00 -5.41761692e-04
-1.89159557e-01 3.33754420e-01 -7.77492046e-01 -2.14220017e-01
-1.51990473e-01 -1.57403484e-01 1.84268668e-01 8.70417431e-02
4.86487925e-01 -6.10370040e-01 -8.75651896e-01 2.28182212e-01
-1.01307273e+00 -1.47826403e-01 7.51146257e-01 2.91634232e-01
5.24298072e-01 9.74268019e-02 3.43005776e-01 -2.46591136e-01
7.48335838e-01 -1.37402639e-02 -2.41529211e-01 -6.06732769e-03
-4.34672087e-01 8.48382711e-02 4.79787111e-01 -4.28500116e-01
-6.50560856e-01 9.75825116e-02 -1.57141089e-01 -7.32972741e-01
-1.49736807e-01 2.28937149e-01 1.63510427e-01 -1.95063464e-02
8.21896434e-01 -2.78471589e-01 -7.07228780e-02 -1.94561049e-01
3.16316187e-01 3.77054602e-01 5.72597265e-01 -1.82953089e-01
3.07670087e-01 3.69500637e-01 3.38674694e-01 -3.26881677e-01
-1.03344178e+00 -3.63639116e-01 -8.76719296e-01 -8.11669886e-01
1.12787962e+00 -7.67782152e-01 -9.51723039e-01 5.82773805e-01
-7.80449629e-01 -3.94834220e-01 -5.70776343e-01 6.41168475e-01
-7.90993452e-01 6.55714348e-02 -4.99603033e-01 -6.60884798e-01
-2.86329150e-01 -9.76857603e-01 9.04893279e-01 2.25980729e-01
-4.33322608e-01 -3.40200037e-01 1.57765999e-01 4.28319901e-01
5.39964378e-01 5.82682788e-01 5.06287992e-01 -3.13209563e-01
-1.67137861e-01 -3.63906085e-01 -2.50683904e-01 6.55407012e-01
1.57731771e-02 5.10893129e-02 -1.01286483e+00 1.11041576e-01
-2.44363666e-01 -4.27863210e-01 1.02670145e+00 5.17535746e-01
1.27183676e+00 -6.80502132e-02 5.47887325e-01 1.47818327e-01
1.31523108e+00 1.40946984e-01 1.00528991e+00 4.14760530e-01
5.98980486e-01 3.15663397e-01 9.53278661e-01 8.54928493e-01
2.23420355e-02 7.89366722e-01 6.70097470e-01 7.62184560e-02
-4.13038701e-01 9.63756219e-02 6.28962994e-01 4.71409768e-01
-7.97672808e-01 -2.84299888e-02 -6.88767552e-01 3.63435566e-01
-1.72512734e+00 -1.09449482e+00 3.30086611e-02 2.05603600e+00
6.34507358e-01 4.51401711e-01 4.76829708e-01 7.80737400e-01
3.51161331e-01 2.68807322e-01 -5.76816320e-01 -6.68860853e-01
2.05250248e-01 4.72706348e-01 4.69609559e-01 2.39640132e-01
-1.14788961e+00 6.96099520e-01 5.95996904e+00 8.21667016e-01
-9.15847003e-01 -5.19284345e-02 5.14487326e-01 -4.79425400e-01
3.60875160e-01 -4.02963549e-01 -3.22544545e-01 3.86501312e-01
1.21634138e+00 4.75735404e-02 2.91748285e-01 8.35913241e-01
5.04905283e-01 -4.24677521e-01 -1.08342147e+00 1.04651642e+00
1.98416978e-01 -1.14599240e+00 9.07698125e-02 -1.06714673e-01
6.66987777e-01 -2.04982415e-01 -1.40349597e-01 2.80892134e-01
8.86156335e-02 -9.58547175e-01 7.51874745e-01 9.32247341e-01
5.91671884e-01 -9.40333784e-01 1.06804395e+00 -2.92223513e-01
-1.19053125e+00 -3.78939271e-01 -4.94624108e-01 -4.32879597e-01
-3.70381512e-02 6.30208433e-01 -3.78129184e-01 2.90207654e-01
9.27613378e-01 7.64187217e-01 -7.85875380e-01 1.07958424e+00
-9.86987129e-02 4.62529808e-01 1.06900902e-02 -4.56370674e-02
3.27664554e-01 2.69719273e-01 2.37941638e-01 1.05017042e+00
5.33754528e-01 3.43127728e-01 -1.00979440e-01 3.49372834e-01
-1.05424479e-01 1.18669093e-01 -1.63783520e-01 9.61228758e-02
3.81552391e-02 1.09433711e+00 -6.71616018e-01 -4.89939123e-01
-3.80840003e-01 9.83660996e-01 4.48505208e-02 -2.08773360e-01
-9.02771652e-01 -4.09204811e-01 9.22241271e-01 1.51731029e-01
3.20302129e-01 -2.59208292e-01 -2.45401159e-01 -7.73409486e-01
1.21563058e-02 -8.03905308e-01 4.91509140e-01 -1.03859055e+00
-1.05419981e+00 4.53136981e-01 -1.76567316e-01 -1.66576004e+00
-2.32003540e-01 -6.67685688e-01 -5.60513258e-01 3.07093352e-01
-1.16551113e+00 -8.17379355e-01 -6.51261628e-01 6.82278216e-01
7.45046496e-01 -7.99125060e-02 7.01981843e-01 3.26726198e-01
-1.91877097e-01 4.71097648e-01 -3.99594665e-01 7.13368878e-02
6.95123434e-01 -1.28594708e+00 2.18015268e-01 6.57731533e-01
2.29070261e-02 -1.43890455e-01 7.34973013e-01 -2.70890594e-01
-1.11168063e+00 -9.57677126e-01 8.15625012e-01 -4.07194942e-01
6.38222635e-01 1.12599932e-01 -7.47174323e-01 -9.10374243e-03
1.08796842e-01 1.83046475e-01 7.18568921e-01 -1.46276370e-01
-3.71370792e-01 -4.82058823e-01 -1.33954251e+00 3.32884818e-01
1.05513489e+00 -3.76527399e-01 -5.53182781e-01 -8.01900402e-02
6.05943501e-01 -8.43943954e-02 -1.26756382e+00 3.98106605e-01
1.03980410e+00 -1.51116586e+00 8.99118721e-01 -5.61152101e-01
8.64872456e-01 -3.85028630e-01 -2.15297058e-01 -1.17841756e+00
-4.62302536e-01 -1.94225609e-02 -2.25165844e-01 7.62539148e-01
1.70136884e-01 6.69693053e-02 7.86872387e-01 5.93784332e-01
-1.36018842e-01 -9.51050997e-01 -9.49225366e-01 -7.85392702e-01
-1.49471194e-01 -7.60037422e-01 5.51603615e-01 6.46039307e-01
3.99632007e-02 -1.12143606e-01 -4.88564551e-01 -2.13406131e-01
2.49494821e-01 -2.58522838e-01 5.87145567e-01 -1.20009780e+00
-1.59556285e-01 -5.71518064e-01 -1.11055708e+00 -3.32359314e-01
-1.46881491e-01 -3.85854572e-01 -1.14197314e-01 -1.62764728e+00
-5.45227118e-02 1.48503214e-01 -8.68522108e-01 5.98492086e-01
4.20396626e-02 7.12469697e-01 4.64735210e-01 -2.27586497e-02
-1.03358412e+00 3.10448587e-01 1.25922644e+00 -1.17281467e-01
-1.97157666e-01 -8.46810564e-02 -1.94057658e-01 7.59198487e-01
1.03914225e+00 -2.78066069e-01 -3.72107565e-01 -4.09593940e-01
1.93944350e-01 -4.31128033e-02 3.09562415e-01 -1.92964625e+00
-6.73696771e-02 -8.20856094e-02 7.54974604e-01 -3.92939329e-01
3.91724139e-01 -8.39680612e-01 -2.79353950e-02 5.74488997e-01
-7.85347402e-01 1.22022025e-01 4.73685153e-02 2.25444660e-01
-2.38511205e-01 -1.48428768e-01 8.55804384e-01 -1.84551582e-01
-1.12273192e+00 2.33268842e-01 -3.26178342e-01 -2.36719951e-01
9.51027691e-01 -3.91365618e-01 -5.31656519e-02 -3.83526593e-01
-8.83807778e-01 -2.25151390e-01 3.06637347e-01 4.74532604e-01
5.75625718e-01 -1.58100867e+00 -7.08900094e-01 -8.15582722e-02
1.13397732e-01 -5.68904400e-01 3.86241466e-01 6.66730464e-01
-6.09973252e-01 1.80346921e-01 -1.02909887e+00 -3.23423028e-01
-1.39219272e+00 2.12471515e-01 2.64275312e-01 -4.54603940e-01
-4.23023462e-01 6.38353884e-01 -4.75589633e-01 2.97439367e-01
2.66520768e-01 -5.62040508e-01 -5.30700445e-01 3.94635499e-01
9.00685608e-01 8.83517742e-01 1.98407903e-01 -6.15175664e-01
-3.61037254e-01 6.02348685e-01 2.95160025e-01 -9.47766155e-02
1.39278817e+00 2.20302507e-01 2.59011835e-01 5.11561036e-01
1.23386598e+00 -3.72220069e-01 -1.41718364e+00 8.31126496e-02
-1.60898373e-01 -7.09465683e-01 1.52859628e-01 -7.43501186e-01
-1.34621537e+00 8.99621487e-01 1.20180905e+00 1.71216741e-01
1.47975981e+00 -1.07368335e-01 7.94038951e-01 3.76377255e-01
3.91383141e-01 -1.39502454e+00 5.90770185e-01 2.23109022e-01
9.38675821e-01 -1.06609130e+00 2.10192844e-01 1.45000562e-01
-7.00329602e-01 1.24923098e+00 6.58576250e-01 -5.25873184e-01
3.42585593e-01 2.85477061e-02 1.15199052e-01 -1.22414514e-01
-6.71452880e-01 -4.17591184e-01 3.22810501e-01 4.51790959e-01
3.18660408e-01 1.47187412e-01 -4.05416101e-01 3.48970860e-01
-8.79638493e-02 3.82810414e-01 5.65633714e-01 9.71605361e-01
-7.90347338e-01 -7.85104692e-01 -1.30450549e-02 6.81993306e-01
-6.93973541e-01 2.43349299e-01 -3.78122181e-01 4.83749479e-01
3.74550611e-01 9.37726438e-01 1.89535439e-01 -8.92312467e-01
6.95470810e-01 2.23467611e-02 5.06830215e-01 -1.31250277e-01
-6.80368185e-01 -2.60618508e-01 2.04050124e-01 -1.37541831e+00
-1.04262888e+00 -6.36833727e-01 -1.18461215e+00 -4.51805949e-01
1.00747250e-01 -3.21898103e-01 5.53577840e-01 8.98779213e-01
2.64665097e-01 7.92383015e-01 5.64549923e-01 -1.03489625e+00
-2.09671855e-01 -1.07189226e+00 -5.36259055e-01 6.56909049e-01
1.52313292e-01 -8.34996998e-01 -2.58975506e-01 1.24479048e-01] | [8.017416954040527, 0.6206842064857483] |
18ff5f17-4c45-4d41-99c0-c2f44694d5a5 | single-shot-freestyle-dance-reenactment | 2012.01158 | null | https://arxiv.org/abs/2012.01158v2 | https://arxiv.org/pdf/2012.01158v2.pdf | Single-Shot Freestyle Dance Reenactment | The task of motion transfer between a source dancer and a target person is a special case of the pose transfer problem, in which the target person changes their pose in accordance with the motions of the dancer. In this work, we propose a novel method that can reanimate a single image by arbitrary video sequences, unseen during training. The method combines three networks: (i) a segmentation-mapping network, (ii) a realistic frame-rendering network, and (iii) a face refinement network. By separating this task into three stages, we are able to attain a novel sequence of realistic frames, capturing natural motion and appearance. Our method obtains significantly better visual quality than previous methods and is able to animate diverse body types and appearances, which are captured in challenging poses, as shown in the experiments and supplementary video. | ['Lior Wolf', 'Oron Ashual', 'Oran Gafni'] | 2020-12-02 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Gafni_Single-Shot_Freestyle_Dance_Reenactment_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Gafni_Single-Shot_Freestyle_Dance_Reenactment_CVPR_2021_paper.pdf | cvpr-2021-1 | ['pose-transfer'] | ['computer-vision'] | [ 5.81898928e-01 1.85554594e-01 3.93346816e-01 -2.00386405e-01
-4.12103087e-01 -5.07390320e-01 5.76016486e-01 -6.73493385e-01
-4.87367779e-01 8.54391634e-01 -2.36054540e-01 1.99865192e-01
4.69460815e-01 -6.15776122e-01 -8.71831417e-01 -6.31876826e-01
2.31057018e-01 7.41737485e-01 4.95859593e-01 -2.62476414e-01
-1.25833496e-01 6.01465225e-01 -1.61551237e+00 1.61361709e-01
5.44657648e-01 8.18125963e-01 1.40697941e-01 8.74663413e-01
3.46882641e-01 6.60279393e-01 -5.88982046e-01 -5.85321009e-01
2.71873951e-01 -8.98421526e-01 -8.46845865e-01 5.16518950e-01
8.08725893e-01 -4.91388768e-01 -2.85084546e-01 9.15717602e-01
4.08712298e-01 3.85718495e-01 5.88643014e-01 -1.15570021e+00
-1.99220896e-01 2.19415322e-01 -6.63600087e-01 -1.28292799e-01
7.92977512e-01 2.24078059e-01 4.28408295e-01 -6.53467417e-01
9.53867257e-01 1.39624608e+00 6.11254334e-01 1.30046213e+00
-1.26945806e+00 -6.08832240e-01 1.02829032e-01 1.78021684e-01
-1.23924267e+00 -5.23161888e-01 7.85530329e-01 -4.13045466e-01
2.22676367e-01 3.11422884e-01 1.18741703e+00 1.42941034e+00
-4.93873283e-02 7.68638194e-01 8.72163296e-01 -3.57985377e-01
5.93651794e-02 -2.53627807e-01 -6.24319375e-01 5.91729164e-01
-3.44889686e-02 2.66952127e-01 -4.60659236e-01 9.59869176e-02
1.20434535e+00 -1.19861580e-01 -5.23228168e-01 -3.92227560e-01
-1.06373560e+00 4.67438668e-01 3.04414570e-01 1.98614731e-01
-3.47967148e-01 3.81055653e-01 -1.15544483e-01 -3.50921266e-02
3.25788170e-01 2.08935231e-01 -1.44266725e-01 1.54127762e-01
-1.12400615e+00 3.75215769e-01 8.46275926e-01 7.93684602e-01
4.89836574e-01 3.08229506e-01 -1.16814330e-01 4.29739177e-01
3.85607123e-01 6.61648512e-01 3.08428884e-01 -1.25767541e+00
7.30631426e-02 1.33087426e-01 3.13974977e-01 -9.00504827e-01
-2.99283773e-01 -4.68543768e-02 -7.58186996e-01 4.49915439e-01
7.31404424e-01 -3.67248118e-01 -1.00227165e+00 1.99758363e+00
7.97219694e-01 4.84858453e-01 -8.78199562e-02 1.25251782e+00
8.18063140e-01 5.49075365e-01 2.98981667e-02 -3.26656818e-01
1.32203746e+00 -1.10366833e+00 -8.17114055e-01 -2.77905732e-01
-1.37347832e-01 -7.59465992e-01 7.61490762e-01 3.97750497e-01
-1.67862844e+00 -7.26824462e-01 -9.03571963e-01 -6.92112371e-02
1.63018972e-01 -4.85543348e-03 4.97988880e-01 4.42575276e-01
-1.28823733e+00 8.18067968e-01 -8.63515556e-01 -4.19091284e-01
3.89007002e-01 4.97960806e-01 -5.31585634e-01 9.67319459e-02
-9.84888077e-01 6.61565363e-01 2.23219886e-01 3.79579604e-01
-1.14771092e+00 -4.36582416e-01 -1.11199284e+00 -1.74153194e-01
3.89314741e-01 -1.23256528e+00 1.16548300e+00 -1.74910378e+00
-1.98839760e+00 1.13329494e+00 -5.14885299e-02 -4.17481735e-02
1.07361639e+00 -3.73669922e-01 -1.62417740e-01 3.19477290e-01
-2.09582418e-01 9.10116792e-01 1.22455668e+00 -1.46401966e+00
-5.59999704e-01 -3.75712961e-01 5.13860583e-02 3.23704869e-01
8.78612250e-02 -5.82111664e-02 -9.08976078e-01 -7.70431697e-01
-2.00569436e-01 -1.13212538e+00 -1.50757745e-01 3.65530461e-01
-3.27352285e-01 1.00002848e-01 8.28952014e-01 -7.35762596e-01
7.64855206e-01 -2.06783223e+00 8.53616834e-01 1.56921986e-02
1.51203603e-01 4.96813804e-01 -1.93851292e-01 3.96257862e-02
-1.07558534e-01 -2.24776015e-01 -3.12603056e-01 -6.20594859e-01
-3.14370841e-01 2.41295934e-01 7.04764575e-02 5.17047167e-01
6.15874827e-02 9.60148335e-01 -1.00522900e+00 -6.24857783e-01
8.86077210e-02 6.68342471e-01 -5.43277383e-01 5.46475887e-01
-3.87378752e-01 1.04577422e+00 -1.80046782e-01 3.78976882e-01
5.68492711e-01 -1.54971495e-01 3.82085115e-01 -2.70761639e-01
2.07289040e-01 -3.46094072e-01 -1.23373365e+00 1.90190005e+00
-6.37949482e-02 4.85346824e-01 4.24409211e-01 -5.90816021e-01
6.69937670e-01 4.03436840e-01 5.01567364e-01 -3.13031763e-01
4.02605474e-01 5.30448146e-02 -5.54124974e-02 -5.91314197e-01
3.35881203e-01 -4.28402275e-01 2.07606539e-01 4.20739770e-01
1.43949568e-01 -1.90276653e-01 2.81907052e-01 -1.92244589e-01
7.60619342e-01 6.28901184e-01 -4.84316796e-02 1.62981600e-02
5.66145360e-01 -2.56528854e-01 5.27082503e-01 3.46964329e-01
-1.22493245e-02 1.00579429e+00 1.04129612e-01 -6.37552559e-01
-1.01865804e+00 -1.20325601e+00 2.72540003e-01 7.31486022e-01
4.13712174e-01 1.17017543e-02 -1.14501214e+00 -6.12080991e-01
-2.40733013e-01 3.07915121e-01 -7.20604777e-01 -1.18238203e-01
-1.00423551e+00 -5.18211663e-01 4.09662426e-01 5.15403211e-01
5.70020437e-01 -1.34766066e+00 -7.69844115e-01 2.97295023e-02
-5.09840488e-01 -1.22811139e+00 -9.84268129e-01 -3.60979885e-01
-7.06530035e-01 -1.27449727e+00 -1.13508236e+00 -9.49547410e-01
7.50248671e-01 -5.48878610e-02 1.19289398e+00 3.72846097e-01
-2.12553233e-01 4.00877833e-01 -2.19092295e-01 -8.80330428e-02
-6.34705067e-01 -1.66868776e-01 9.17931125e-02 4.22977328e-01
-2.81604439e-01 -6.91729963e-01 -8.29869926e-01 4.16318476e-01
-1.06345773e+00 3.94567430e-01 4.21946287e-01 6.31553948e-01
4.61487621e-01 -2.51967609e-01 -4.40116525e-02 -8.34605455e-01
4.83132973e-02 1.49480561e-02 -4.59047914e-01 2.28096426e-01
5.32493144e-02 -2.87570626e-01 5.29582441e-01 -7.79765725e-01
-1.10695648e+00 6.10147774e-01 -3.13521504e-01 -5.81807971e-01
-2.43535131e-01 -3.52578551e-01 -3.62086535e-01 -2.22290993e-01
5.58496237e-01 1.72742635e-01 2.24115834e-01 -3.14137399e-01
3.57911527e-01 2.01219529e-01 9.75605011e-01 -5.30660212e-01
1.13003838e+00 6.60094142e-01 8.81388858e-02 -8.68922710e-01
-6.06767058e-01 3.55012529e-02 -9.97553885e-01 -6.68309510e-01
1.17131031e+00 -8.17221463e-01 -8.48183990e-01 7.06775963e-01
-1.30553579e+00 -7.43383288e-01 -3.97820860e-01 4.06698942e-01
-8.33188891e-01 3.00637841e-01 -7.66982079e-01 -4.49432462e-01
-2.16465950e-01 -1.01658177e+00 1.27483475e+00 5.30590773e-01
-2.85436809e-01 -9.82133210e-01 1.31226238e-02 4.45114851e-01
3.57680283e-02 8.60548079e-01 4.87338245e-01 -7.59526640e-02
-6.37948573e-01 1.84433714e-01 1.94176212e-01 2.34305799e-01
1.97040662e-01 1.23770550e-01 -7.86783934e-01 -5.55164933e-01
-1.76761597e-02 -2.23394021e-01 6.07842505e-01 4.34783220e-01
7.73743927e-01 -2.40516797e-01 -2.73832440e-01 9.06301200e-01
1.08265543e+00 3.61374915e-01 9.09043133e-01 1.43891275e-02
8.94654989e-01 6.40655518e-01 4.14568037e-01 6.60931244e-02
1.36690646e-01 1.00920379e+00 4.12929296e-01 -4.22393084e-01
-4.70076591e-01 -2.38748744e-01 3.80661011e-01 4.36493099e-01
-6.97883546e-01 -1.88619152e-01 -4.99235362e-01 2.26590604e-01
-1.84461057e+00 -9.68679309e-01 -1.56538889e-01 2.13179374e+00
7.86532223e-01 -2.48885393e-01 4.30097133e-01 -7.68365934e-02
8.13728392e-01 -3.86480577e-02 -5.80228686e-01 -2.27309037e-02
1.37792483e-01 1.37519464e-01 -8.70376378e-02 4.99094725e-01
-1.08423328e+00 1.06150222e+00 6.69889164e+00 4.56163019e-01
-1.15776038e+00 6.74469694e-02 5.63716710e-01 -8.90241712e-02
-1.80006400e-01 -3.38892967e-01 -5.11858821e-01 3.53242874e-01
3.56440932e-01 1.46962836e-01 5.27063489e-01 4.35871691e-01
9.81234536e-02 1.06041789e-01 -1.29409337e+00 9.71688628e-01
5.42074382e-01 -1.11897933e+00 2.24800378e-01 -8.27799588e-02
6.72063649e-01 -6.59109354e-01 5.84442616e-02 -1.49741024e-02
8.41080099e-02 -1.16894960e+00 1.01639128e+00 7.14488268e-01
9.23232257e-01 -6.13077700e-01 4.44094479e-01 1.83136493e-01
-1.04373789e+00 2.48087883e-01 -3.44794728e-02 9.29426998e-02
4.47533190e-01 3.71216331e-03 -1.53072625e-01 6.26925170e-01
6.81767523e-01 6.19916081e-01 -3.48369181e-01 9.70112920e-01
-4.53583598e-01 1.28316104e-01 -1.33085459e-01 3.13417017e-01
-1.19858168e-01 -3.43445450e-01 6.51312768e-01 8.70208561e-01
2.85789222e-01 1.91273823e-01 4.63113114e-02 8.82973850e-01
-7.92893320e-02 -8.02498609e-02 -3.67380321e-01 3.90285790e-01
-1.58459380e-01 1.37078440e+00 -8.59504402e-01 -2.35909730e-01
-1.96391940e-02 1.50977886e+00 1.62101939e-01 4.57563162e-01
-9.46020901e-01 9.31015536e-02 4.74805921e-01 2.34516561e-01
4.19701099e-01 -7.13536218e-02 4.64243293e-01 -1.25442207e+00
-3.28118540e-02 -9.59120631e-01 1.86434299e-01 -1.01188815e+00
-1.02192807e+00 1.01312280e+00 -3.57151069e-02 -1.20007694e+00
-5.57290435e-01 -4.33891654e-01 -4.70773429e-01 5.74703038e-01
-1.03471363e+00 -1.26563811e+00 -5.53769886e-01 8.38309288e-01
5.55757403e-01 1.60285756e-01 6.34874105e-01 5.16321182e-01
-4.66861665e-01 3.91588181e-01 -3.75272214e-01 2.58629054e-01
6.83341980e-01 -1.19305611e+00 3.13826591e-01 6.86146617e-01
3.09666783e-01 5.33341579e-02 7.95826674e-01 -5.72576165e-01
-1.10021198e+00 -1.18955553e+00 4.49315578e-01 -4.78943855e-01
4.23914082e-02 -3.88795733e-01 -8.56100321e-01 7.95040548e-01
1.90749973e-01 6.92242086e-02 3.35668802e-01 -4.80718434e-01
2.20816344e-01 -1.32164851e-01 -1.09056401e+00 7.37992644e-01
1.24651849e+00 -8.64790604e-02 -5.10376334e-01 2.09781706e-01
3.10663044e-01 -9.42004442e-01 -6.06003344e-01 3.18516344e-01
8.48190069e-01 -9.90154982e-01 1.08962989e+00 -6.38837874e-01
4.73405749e-01 -4.31876421e-01 2.48133764e-01 -1.16040480e+00
-1.08784080e-01 -8.87053013e-01 -1.61066160e-01 1.10618031e+00
-1.38432076e-02 -1.02846436e-01 8.48773718e-01 5.38998187e-01
2.80126482e-01 -4.28987831e-01 -8.16244543e-01 -6.53614461e-01
-6.22293167e-02 -7.02214800e-03 4.42074597e-01 7.58176208e-01
-5.81407845e-01 4.30646122e-01 -7.81758428e-01 5.47008552e-02
7.68529058e-01 1.72390208e-01 9.87192869e-01 -1.23010480e+00
-5.76646388e-01 -1.24575093e-01 -3.30770969e-01 -1.21475482e+00
2.68717527e-01 -6.10948145e-01 2.39646152e-01 -1.25161266e+00
2.08904862e-01 4.81769033e-02 2.80768871e-01 3.65392327e-01
-2.75903940e-01 8.33928943e-01 4.74821568e-01 2.18281120e-01
-5.70611179e-01 5.44956505e-01 1.88818026e+00 9.01506022e-02
-2.44120300e-01 2.60028750e-01 -2.59011149e-01 1.12210023e+00
3.51783633e-01 -5.34601927e-01 -3.32958788e-01 -3.95528257e-01
-1.20239250e-01 5.25024414e-01 6.60790503e-01 -1.10427976e+00
9.24901962e-02 -6.71515241e-02 8.75017226e-01 -3.41727346e-01
6.29337788e-01 -8.56090426e-01 7.72626936e-01 7.34591722e-01
-9.36124027e-02 8.85397345e-02 2.12739483e-01 5.88059247e-01
2.71456819e-02 -8.54106843e-02 9.64992583e-01 -2.80788004e-01
-6.06543183e-01 4.67330039e-01 -2.63969749e-01 1.23371080e-01
1.10950947e+00 -2.48193398e-01 6.47365451e-02 -6.20597720e-01
-1.22920442e+00 1.00743212e-01 6.81014657e-01 4.12191302e-01
6.70005322e-01 -1.50946879e+00 -8.13292921e-01 2.90865928e-01
-4.30679947e-01 1.19016267e-01 3.11477870e-01 8.10381591e-01
-7.63395727e-01 -2.93928415e-01 -4.77312803e-01 -8.00304055e-01
-1.51921070e+00 4.15883839e-01 7.19630122e-01 -6.00323714e-02
-7.64874578e-01 6.55262291e-01 4.68750119e-01 -1.18843526e-01
1.01839423e-01 7.32148960e-02 -2.39788517e-01 -1.17209800e-01
6.82075381e-01 1.60284624e-01 -4.01195914e-01 -1.13569582e+00
-2.04733446e-01 9.66696739e-01 2.42368236e-01 -2.42700413e-01
1.18247104e+00 -2.00008497e-01 -2.97267362e-02 3.98909360e-01
8.26812029e-01 -1.24472186e-01 -1.84790576e+00 5.43523766e-02
-4.63212073e-01 -6.13387346e-01 -5.02627075e-01 -5.69859505e-01
-1.41905093e+00 6.70331180e-01 6.19309187e-01 -1.47708908e-01
1.36664331e+00 -3.24356765e-03 9.94866371e-01 -5.86708970e-02
5.40372372e-01 -8.36845100e-01 4.71647292e-01 2.30790570e-01
9.67438281e-01 -9.37606573e-01 -1.63686380e-01 -5.04786909e-01
-4.14302319e-01 1.18384290e+00 7.41462052e-01 -1.82066217e-01
3.14664692e-01 1.81070074e-01 2.45430410e-01 2.08102502e-02
-4.58488584e-01 -3.66407841e-01 5.95975757e-01 8.32674325e-01
1.09451361e-01 -2.61546344e-01 -1.98403656e-01 4.18959498e-01
-2.03054994e-01 9.15665850e-02 4.84533578e-01 5.11740625e-01
-4.20213342e-02 -1.10490000e+00 -5.67870855e-01 -2.94019163e-01
-4.18589264e-01 3.17310214e-01 -5.23107290e-01 1.10371089e+00
3.35704595e-01 8.12625468e-01 1.67472214e-01 -2.61491835e-01
4.74133819e-01 -1.02519967e-01 1.02126336e+00 -5.87106347e-01
-6.00051284e-01 3.47714663e-01 -9.67995524e-02 -7.68554509e-01
-7.86274314e-01 -6.77079022e-01 -1.07203150e+00 -3.09109658e-01
1.20744817e-02 -6.92306459e-02 2.06900164e-01 1.02533197e+00
-1.10945269e-01 7.68230975e-01 4.19233590e-01 -1.57516491e+00
5.24527133e-02 -6.95655406e-01 -6.99347973e-01 7.87249625e-01
3.41250479e-01 -5.94869018e-01 -2.51865596e-01 5.52496314e-01] | [11.016790390014648, -0.8139723539352417] |
e99e9975-a0e1-4346-8efb-e94dd5f6638e | on-the-stability-and-generalization-of-1 | 2302.09815 | null | https://arxiv.org/abs/2302.09815v1 | https://arxiv.org/pdf/2302.09815v1.pdf | On the Stability and Generalization of Triplet Learning | Triplet learning, i.e. learning from triplet data, has attracted much attention in computer vision tasks with an extremely large number of categories, e.g., face recognition and person re-identification. Albeit with rapid progress in designing and applying triplet learning algorithms, there is a lacking study on the theoretical understanding of their generalization performance. To fill this gap, this paper investigates the generalization guarantees of triplet learning by leveraging the stability analysis. Specifically, we establish the first general high-probability generalization bound for the triplet learning algorithm satisfying the uniform stability, and then obtain the excess risk bounds of the order $O(n^{-\frac{1}{2}} \mathrm{log}n)$ for both stochastic gradient descent (SGD) and regularized risk minimization (RRM), where $2n$ is approximately equal to the number of training samples. Moreover, an optimistic generalization bound in expectation as fast as $O(n^{-1})$ is derived for RRM in a low noise case via the on-average stability analysis. Finally, our results are applied to triplet metric learning to characterize its theoretical underpinning. | ['Feng Zheng', 'Tieliang Gong', 'Weifu Li', 'Bin Gu', 'Xue Jiang', 'Hong Chen', 'Jun Chen'] | 2023-02-20 | null | null | null | null | ['person-re-identification', 'metric-learning', 'metric-learning'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 2.19288006e-01 4.18433435e-02 1.03215210e-01 -6.43476784e-01
-1.18183351e+00 -3.82163018e-01 9.58430246e-02 8.34591463e-02
-5.22060037e-01 8.91034365e-01 -4.85467881e-01 -4.93354410e-01
-5.63930392e-01 -3.87786597e-01 -6.93489134e-01 -1.00796664e+00
-2.01739132e-01 2.61927783e-01 -2.26685911e-01 1.41834691e-01
1.32353336e-01 2.91044861e-01 -1.47224307e+00 -4.59447026e-01
1.02412748e+00 1.50369513e+00 -4.35142994e-01 4.78809059e-01
2.14052945e-01 2.04829678e-01 -2.21977681e-01 -1.01830900e+00
5.23058832e-01 -6.38357997e-01 -5.57771921e-01 3.75095569e-02
6.92844093e-01 3.99013311e-02 -1.63237691e-01 1.41364253e+00
6.74678504e-01 3.21127862e-01 8.53890240e-01 -1.37790239e+00
-5.85654140e-01 2.73393393e-01 -8.10559452e-01 2.27971196e-01
1.39487147e-01 -1.93997681e-01 1.13138521e+00 -1.10835528e+00
-3.32707427e-02 1.07669139e+00 7.37004042e-01 5.58392525e-01
-1.16053569e+00 -9.04781759e-01 4.86162268e-02 1.29721209e-01
-1.49135411e+00 -3.73881817e-01 5.90903640e-01 -4.67231274e-01
3.77784252e-01 4.28416103e-01 1.96236163e-01 9.66771543e-01
2.82765424e-04 7.63415277e-01 1.24977922e+00 -6.61569893e-01
1.55081615e-01 4.52215016e-01 2.27426916e-01 9.09632087e-01
5.94174922e-01 1.98632583e-01 -6.99382186e-01 -1.55522496e-01
5.68263948e-01 -2.47118250e-01 -2.67298948e-02 -4.12103683e-01
-8.60077500e-01 8.22316706e-01 7.64956549e-02 2.07949593e-03
-9.73190889e-02 1.89439178e-01 3.14497203e-01 3.42216820e-01
5.77973843e-01 -6.06431887e-02 -2.63078958e-01 -1.40519723e-01
-7.83203483e-01 1.16224373e-02 7.31729865e-01 1.04258001e+00
6.33489370e-01 1.17322661e-01 2.20192056e-02 8.96898329e-01
4.27237600e-01 9.03680086e-01 1.09151483e-01 -8.32045674e-01
8.08107018e-01 2.50256985e-01 2.09798172e-01 -9.68953013e-01
-3.43111634e-01 -5.07473171e-01 -1.05080843e+00 7.69341290e-02
7.23815918e-01 -2.40664780e-01 -2.47206330e-01 2.06322432e+00
3.20752561e-01 1.58979818e-01 -2.18492448e-01 6.25018537e-01
2.11373940e-01 7.91048259e-02 -2.37209946e-01 -5.38701594e-01
1.14144313e+00 -3.05906922e-01 -3.76108229e-01 2.64056977e-02
6.64400756e-01 -6.22280717e-01 1.02807581e+00 3.52113336e-01
-1.02285063e+00 -2.25395694e-01 -1.00052154e+00 1.56807378e-01
-4.20828257e-03 4.46983278e-02 5.96294761e-01 1.39529967e+00
-9.00974929e-01 6.00945592e-01 -6.89291060e-01 -4.01730716e-01
5.22973120e-01 6.65809751e-01 -3.39478552e-01 -1.89489903e-04
-9.00433004e-01 5.44495761e-01 8.15345068e-03 4.96923059e-01
-5.44927537e-01 -6.40363514e-01 -7.13325560e-01 -2.68030554e-01
2.93671876e-01 -6.37741506e-01 9.02510762e-01 -4.29368973e-01
-1.30544209e+00 9.39280510e-01 -2.52304494e-01 -4.77799177e-01
8.11528683e-01 -9.43457782e-02 -3.78761768e-01 -4.03733104e-02
1.65684998e-01 1.37943476e-01 8.45860898e-01 -9.64506269e-01
-6.83219135e-01 -9.45301831e-01 -3.93125713e-02 5.40922694e-02
-7.09423125e-01 1.38273478e-01 -8.42933282e-02 -6.08963013e-01
2.94205308e-01 -1.10963392e+00 -8.37357044e-02 5.91747873e-02
-3.09753418e-01 -3.20809454e-01 3.87267232e-01 -3.90911669e-01
1.23884201e+00 -2.17040658e+00 4.84535582e-02 5.83511412e-01
1.39335081e-01 1.03858434e-01 2.01410174e-01 1.65233836e-01
-1.98798589e-02 1.58220842e-01 -3.84171605e-01 -6.10910594e-01
3.11570704e-01 -1.53548136e-01 -2.62581855e-01 9.96811032e-01
-2.53817141e-01 6.23241365e-01 -7.77173042e-01 -4.64064479e-01
2.20425036e-02 3.22396189e-01 -3.25044066e-01 -2.24725664e-01
3.46465707e-01 2.29418933e-01 -5.16149163e-01 6.73433900e-01
6.86744094e-01 -4.85992044e-01 1.48464486e-01 -2.08027318e-01
1.19969465e-01 -2.55771428e-01 -1.45149124e+00 1.29853153e+00
-3.63376051e-01 5.72516680e-01 4.94372435e-02 -1.32444966e+00
7.81920075e-01 2.98676819e-01 6.56113088e-01 -4.43848640e-01
2.66441554e-01 4.19589132e-01 -2.74517685e-01 -2.40873486e-01
1.32657394e-01 -4.83916014e-01 -1.16891898e-01 6.95416033e-01
-9.31081921e-02 4.47201133e-01 9.73639786e-02 -3.65214422e-02
6.55948162e-01 -3.80561016e-02 -1.14189051e-02 -3.85484248e-01
5.13355494e-01 -6.82398498e-01 7.04200387e-01 8.52652431e-01
-4.54239905e-01 4.05097783e-01 4.14240241e-01 -6.79195970e-02
-8.23516726e-01 -1.20470679e+00 -5.38635612e-01 1.03789842e+00
1.86959743e-01 -1.49313405e-01 -7.72415102e-01 -7.10404396e-01
1.30362064e-01 4.76883650e-01 -8.21240008e-01 -4.34196919e-01
-3.55679125e-01 -1.29681063e+00 7.13514626e-01 5.68375945e-01
5.45424998e-01 -5.25869310e-01 -4.18634772e-01 -2.82137722e-01
-3.55528183e-02 -1.03420031e+00 -6.64490998e-01 5.96879683e-02
-8.71162713e-01 -1.02436674e+00 -5.36177874e-01 -6.89177275e-01
6.86456800e-01 2.54177600e-01 6.88033462e-01 -2.89816618e-01
-3.20442736e-01 7.22688913e-01 -4.43500690e-02 -3.89360309e-01
9.26924311e-03 -2.02772990e-01 8.35231960e-01 6.16738796e-01
6.54723227e-01 -5.19727647e-01 -5.33279657e-01 4.51927334e-01
-5.61217964e-01 -4.81260508e-01 4.84502107e-01 8.52820218e-01
5.90477347e-01 -1.23262338e-01 5.63260972e-01 -5.59164584e-01
4.93425459e-01 -2.80372530e-01 -8.89727831e-01 6.03205562e-01
-9.79325533e-01 7.20940977e-02 4.33474183e-01 -3.11470836e-01
-8.60261261e-01 -1.34551823e-01 2.77738959e-01 -3.62255663e-01
2.45199054e-01 3.27122927e-01 -1.71015844e-01 -3.15131277e-01
6.73584580e-01 3.13125104e-01 5.07676341e-02 -4.90292281e-01
3.27281564e-01 6.54649615e-01 4.67621773e-01 -9.03465331e-01
1.04528558e+00 5.80660999e-01 4.67094511e-01 -8.34285319e-01
-1.19659758e+00 -2.72504181e-01 -6.67946637e-01 -2.66485959e-01
5.59432030e-01 -6.12892389e-01 -1.27385986e+00 2.26059154e-01
-6.19862854e-01 1.48223966e-01 -1.44760713e-01 8.11894774e-01
-8.66636217e-01 7.00631380e-01 -3.27032506e-01 -1.38845932e+00
-2.71491945e-01 -9.34801042e-01 6.79140091e-01 1.86907589e-01
1.45140201e-01 -1.08262575e+00 -1.55186221e-01 5.66472471e-01
1.02691382e-01 3.03849220e-01 8.12208533e-01 -6.06991172e-01
-3.41001689e-01 -4.01150167e-01 -3.28152239e-01 5.91078520e-01
2.12944418e-01 -3.52894515e-01 -9.81693685e-01 -6.02580011e-01
1.87460005e-01 -3.59925658e-01 7.07016110e-01 3.28459293e-01
1.12340069e+00 -2.65164256e-01 -2.42806822e-01 6.84415162e-01
1.17518413e+00 -1.73424762e-02 3.21467429e-01 7.61938691e-02
5.72919667e-01 6.31855726e-01 5.17099619e-01 7.71480799e-01
1.56186506e-01 5.11464655e-01 2.42497236e-01 2.44635820e-01
2.07965031e-01 -1.03088632e-01 3.48924071e-01 5.38376033e-01
-8.55510705e-04 6.69632480e-02 -7.06854165e-01 2.78860807e-01
-1.71920967e+00 -9.94411409e-01 -3.99190895e-02 2.95012021e+00
7.61094034e-01 9.20983031e-02 3.64310741e-01 2.92719185e-01
1.00383997e+00 -2.79572159e-01 -8.00393701e-01 -1.70372903e-01
-3.96675766e-01 1.40076101e-01 5.50841033e-01 5.48002183e-01
-1.07332265e+00 5.52134752e-01 5.82422400e+00 9.92133260e-01
-7.54671097e-01 -6.34519532e-02 7.62292564e-01 -2.39500687e-01
-6.58151284e-02 -1.07385600e-02 -1.01147020e+00 5.34939885e-01
9.36772048e-01 -3.54524106e-01 4.39107418e-01 5.87677836e-01
3.25791650e-02 5.24910577e-02 -1.26439357e+00 1.51718688e+00
2.90983588e-01 -8.95225704e-01 -2.31715923e-04 4.36111510e-01
6.55870259e-01 -2.45546445e-01 6.06448889e-01 4.40129451e-02
1.10568121e-01 -9.28081512e-01 5.19570172e-01 3.39597315e-01
9.54919636e-01 -9.31776345e-01 4.39377636e-01 2.45081812e-01
-1.22226453e+00 -3.45076680e-01 -4.29445654e-01 2.79493809e-01
-3.02773621e-02 8.36795032e-01 -3.48084837e-01 7.45672524e-01
8.01022708e-01 5.59028983e-01 -4.81140047e-01 1.06102204e+00
2.26319760e-01 5.67980945e-01 -6.79571807e-01 2.35406887e-02
-1.85170844e-02 -7.09751666e-01 5.89280367e-01 9.45714951e-01
4.15580332e-01 1.39333934e-01 -7.34996572e-02 4.83974189e-01
-2.42967591e-01 2.95323014e-01 -3.46846312e-01 6.70190379e-02
4.84220445e-01 9.78454828e-01 -5.18619895e-01 5.85725792e-02
-3.53729546e-01 8.68048668e-01 3.89393896e-01 3.47152710e-01
-8.25931072e-01 -6.41847849e-01 6.25512660e-01 -1.81038037e-01
3.24421793e-01 -2.26492554e-01 -5.10859132e-01 -1.11553812e+00
6.67621911e-01 -4.24382657e-01 6.27392888e-01 3.48554854e-03
-1.67392468e+00 4.29535210e-01 8.32997113e-02 -1.36809921e+00
-2.94569761e-01 -7.99950778e-01 -1.89533025e-01 8.36401701e-01
-1.08442438e+00 -8.05584371e-01 5.56963719e-02 7.62361944e-01
-2.12246850e-01 -3.60099584e-01 6.75033748e-01 5.16518176e-01
-9.96061206e-01 1.51209950e+00 5.84140778e-01 2.43490502e-01
5.95321894e-01 -1.22133303e+00 -9.51635316e-02 8.66954207e-01
2.22046629e-01 6.09284520e-01 6.89293087e-01 -2.65531898e-01
-1.45843077e+00 -8.90915632e-01 9.14186954e-01 -6.97996974e-01
6.86218202e-01 -4.62510705e-01 -5.94088316e-01 5.11420012e-01
-4.37449843e-01 2.68072128e-01 1.23799431e+00 2.35222280e-01
-7.90567517e-01 -5.07782698e-01 -1.40371132e+00 6.53508008e-01
1.25769448e+00 -7.05889821e-01 -1.67292163e-01 4.62249935e-01
2.21240610e-01 -1.12562209e-01 -1.02140653e+00 4.47499007e-01
7.44405389e-01 -1.10322666e+00 9.46611106e-01 -8.68581057e-01
-2.12496743e-01 -1.41812503e-01 -7.85310566e-01 -7.70769656e-01
2.50975993e-02 -1.02827168e+00 -2.66131669e-01 1.15568841e+00
5.38507223e-01 -8.08493912e-01 9.77573276e-01 8.27820599e-01
1.43116817e-01 -8.85898232e-01 -1.50524879e+00 -1.20120323e+00
3.35589647e-01 -5.32214940e-01 1.55152857e-01 7.03261971e-01
5.74510582e-02 9.69604477e-02 -6.31960392e-01 8.63213390e-02
1.30250299e+00 4.90552522e-02 3.62795651e-01 -1.38437092e+00
-3.89276564e-01 -4.40126926e-01 -5.43581009e-01 -1.04497969e+00
2.74207085e-01 -9.82106209e-01 -2.76098520e-01 -9.04829621e-01
3.30221534e-01 -6.60683870e-01 -7.23706901e-01 1.33279294e-01
-2.15974122e-01 3.51179421e-01 4.33625244e-02 2.91093867e-02
-6.78715587e-01 8.04237068e-01 6.38570607e-01 4.85267416e-02
2.00051293e-02 4.28693801e-01 -8.08998644e-01 6.02910817e-01
6.72947645e-01 -2.73802400e-01 -4.33471799e-01 -3.35210383e-01
2.95735836e-01 -1.41537003e-02 3.30830216e-01 -8.57534587e-01
1.80153713e-01 -4.58036289e-02 2.79517323e-01 -3.24881673e-01
4.36069429e-01 -6.55651331e-01 -2.46550709e-01 2.82199681e-01
-3.96209270e-01 1.60321459e-01 -6.99894354e-02 9.16606247e-01
1.79239213e-01 -2.41973177e-01 9.81061280e-01 2.54377067e-01
6.17338046e-02 6.51466191e-01 2.11461838e-02 2.94408262e-01
1.01726842e+00 -3.33937109e-01 6.76795915e-02 -3.86865646e-01
-5.28206110e-01 1.83527574e-01 -9.91027430e-03 1.74011201e-01
5.58485091e-01 -1.53871894e+00 -8.28722954e-01 1.16053112e-01
1.32807255e-01 -3.40350300e-01 3.52503866e-01 1.11034000e+00
8.20313469e-02 4.22120988e-01 1.71177775e-01 -6.63731694e-01
-1.05122280e+00 4.52407688e-01 4.91782397e-01 7.11171106e-02
-2.78302640e-01 1.06159675e+00 -5.51093630e-02 -1.59657866e-01
5.94379008e-01 1.83885619e-01 4.34567362e-01 3.50666940e-02
5.40275216e-01 7.94982731e-01 1.42746076e-01 -6.92214847e-01
-5.51384687e-01 7.21721470e-01 -2.03880176e-01 -3.47299039e-01
9.04255390e-01 -2.61737466e-01 -2.23261058e-01 4.77557123e-01
1.59014356e+00 -2.49168262e-01 -1.29972553e+00 -4.59903181e-01
3.16819429e-01 -5.47665656e-01 -2.50557542e-01 -3.44523758e-01
-8.67737651e-01 9.57143843e-01 9.61261868e-01 1.13391139e-01
9.43150878e-01 -6.61087707e-02 5.56171894e-01 6.70564890e-01
5.59331000e-01 -1.23180389e+00 6.11529313e-02 2.58812308e-01
5.70779443e-01 -1.39598429e+00 -3.78644769e-03 -2.14593142e-01
-3.22259188e-01 9.25059497e-01 3.86529058e-01 9.66770351e-02
1.04997206e+00 -2.64727175e-01 -2.07424179e-01 1.70686319e-01
-3.12039226e-01 2.81639770e-02 3.61645430e-01 4.78611916e-01
4.24058616e-01 1.08478971e-01 -1.82787135e-01 4.94537234e-01
-3.05542707e-01 -2.01284081e-01 2.33040974e-01 6.08737886e-01
-3.75621408e-01 -1.03460765e+00 -2.08765358e-01 6.45925760e-01
-5.78836083e-01 1.37161929e-02 -1.42572001e-01 3.97599548e-01
-2.57028937e-01 1.22507060e+00 -1.90428212e-01 -3.97643983e-01
2.27724761e-01 1.40142143e-01 8.32646847e-01 -1.97377294e-01
-1.00987911e-01 -2.53950626e-01 -3.83497089e-01 -3.30945820e-01
-3.72774839e-01 -1.07066202e+00 -9.09901500e-01 -3.94856215e-01
-3.51546943e-01 3.55908155e-01 4.59910631e-01 1.13874829e+00
1.74295053e-01 -3.10716480e-01 9.79506314e-01 -4.32024777e-01
-1.19644868e+00 -7.12530017e-01 -8.15111339e-01 2.67031699e-01
3.22059542e-01 -6.91564202e-01 -6.66958809e-01 -2.76083469e-01] | [7.639455795288086, 4.147721290588379] |
0ff99deb-4459-45d3-b968-5ea74fcc0ba2 | an-arabic-tweets-sentiment-analysis-dataset | null | null | https://aclanthology.org/2020.osact-1.1 | https://aclanthology.org/2020.osact-1.1.pdf | An Arabic Tweets Sentiment Analysis Dataset (ATSAD) using Distant Supervision and Self Training | As the number of social media users increases, they express their thoughts, needs, socialise and publish their opinions reviews. For good social media sentiment analysis, good quality resources are needed, and the lack of these resources is particularly evident for languages other than English, in particular Arabic. The available Arabic resources lack of from either the size of the corpus or the quality of the annotation. In this paper, we present an Arabic Sentiment Analysis Corpus collected from Twitter, which contains 36K tweets labelled into positive and negative. We employed distant supervision and self-training approaches into the corpus to annotate it. Besides, we release an 8K tweets manually annotated as a gold standard. We evaluated the corpus intrinsically by comparing it to human classification and pre-trained sentiment analysis models, Moreover, we apply extrinsic evaluation methods exploiting sentiment analysis task and achieve an accuracy of 86{\%}. | ['Richard Johansson', 'Stergios Chatzikyriakidis', 'Simon Dobnik', 'Kathrein Abu Kwaik', 'Motaz Saad'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['arabic-sentiment-analysis'] | ['natural-language-processing'] | [-1.13652393e-01 3.16670537e-01 -3.18033338e-01 -5.72237074e-01
-5.54966450e-01 -7.66049325e-01 5.99020422e-01 6.33356631e-01
-7.84516096e-01 7.37154007e-01 2.70336270e-01 -1.11188479e-01
5.97986519e-01 -6.30436540e-01 -2.25827798e-01 -4.54245180e-01
3.56724739e-01 3.33235562e-01 1.00483805e-01 -6.91257000e-01
3.25966686e-01 -3.88608612e-02 -1.37187636e+00 4.39943999e-01
1.02269351e+00 1.05814910e+00 -2.06245095e-01 3.76078457e-01
-1.83884397e-01 9.74867165e-01 -8.03165078e-01 -1.21309507e+00
-1.51417717e-01 -2.65735179e-01 -9.34681892e-01 2.60193467e-01
-3.28086376e-01 6.89476058e-02 4.28173095e-01 1.07025468e+00
2.37527668e-01 -9.24311951e-02 6.63406014e-01 -1.07228649e+00
-5.47851741e-01 6.72261894e-01 -6.04203999e-01 -6.35686666e-02
2.56009638e-01 -2.91898012e-01 1.18648934e+00 -1.14017928e+00
6.79595888e-01 6.99469507e-01 5.16393483e-01 3.21246207e-01
-4.93320048e-01 -6.06310844e-01 5.35925031e-02 -2.27822557e-01
-1.10395515e+00 -4.84196305e-01 7.09096313e-01 -4.86763895e-01
7.47394621e-01 2.33486132e-03 6.27023399e-01 1.07649052e+00
-5.69380522e-02 7.31437206e-01 1.28827453e+00 -7.92594969e-01
1.97762266e-01 8.83297324e-01 1.90684795e-01 3.34013462e-01
4.66322340e-02 -8.58215511e-01 -5.67920208e-01 5.97204864e-02
-1.57972634e-01 -1.97168082e-01 1.22852810e-01 1.14343956e-01
-9.66546535e-01 8.46046031e-01 2.78956056e-01 4.91459608e-01
-2.53448308e-01 -5.36960125e-01 7.25833952e-01 2.76751131e-01
1.02746952e+00 4.71396118e-01 -7.21009016e-01 -2.04655886e-01
-7.44665563e-01 -2.37187058e-01 8.89584482e-01 8.90129805e-01
9.16211486e-01 -3.67312163e-01 3.64970118e-01 9.24647450e-01
5.27718246e-01 8.02787900e-01 7.74026453e-01 -4.44531202e-01
4.71595109e-01 1.08598423e+00 2.98268259e-01 -1.41218197e+00
-4.86442089e-01 -3.08319151e-01 -6.70100927e-01 -3.26695770e-01
2.70118833e-01 -4.91065294e-01 -3.83454263e-01 1.24471867e+00
3.40434223e-01 -6.01782143e-01 4.74212140e-01 7.96524465e-01
8.76655877e-01 5.32468140e-01 1.66510850e-01 -3.42560977e-01
1.45372760e+00 -1.11862099e+00 -9.70016539e-01 -3.50093573e-01
1.04923809e+00 -1.08309901e+00 1.36663318e+00 4.34110433e-01
-8.57006550e-01 -2.61623114e-01 -1.00136864e+00 2.40947619e-01
-7.79632926e-01 4.31916326e-01 4.83422607e-01 8.28968346e-01
-7.34741747e-01 2.84670979e-01 -7.82421708e-01 -3.39073777e-01
3.96409631e-01 1.73564404e-01 -5.10625005e-01 3.61269921e-01
-1.48939300e+00 9.74825740e-01 2.10298911e-01 1.91949099e-01
-1.55568272e-01 5.72661720e-02 -8.97883892e-01 -4.56981599e-01
1.46139950e-01 1.61553770e-01 1.21009374e+00 -1.61919165e+00
-1.45182562e+00 1.25083351e+00 -1.93430409e-01 -8.70090201e-02
3.48035306e-01 -1.59618378e-01 -9.13399160e-01 1.69913098e-01
4.93159145e-01 3.15352976e-01 5.50752580e-01 -1.13100541e+00
-8.55240285e-01 -3.90122831e-01 2.49198347e-01 1.21046230e-01
-1.01739323e+00 4.69665140e-01 -4.32699591e-01 -5.52905560e-01
-1.88128740e-01 -1.12672162e+00 -2.79334307e-01 -7.32989550e-01
-4.47537869e-01 -2.43064970e-01 6.29361629e-01 -6.18268788e-01
1.38583374e+00 -2.27269959e+00 -2.30694532e-01 3.83549482e-01
9.31677967e-02 1.46914557e-01 2.30992034e-01 5.45212746e-01
1.02766588e-01 3.45745146e-01 -1.85435966e-01 -5.58369339e-01
1.92811508e-02 1.58900842e-01 -3.29351217e-01 3.77100378e-01
4.38187718e-01 5.87630808e-01 -1.09224379e+00 -7.53896534e-01
-3.23614299e-01 3.71679574e-01 -3.52417499e-01 6.88805282e-02
-5.04756868e-02 5.26070356e-01 -6.64842188e-01 6.15070820e-01
3.85462970e-01 -5.63401043e-01 1.84717238e-01 -1.32040665e-01
-1.48801520e-01 3.57939094e-01 -8.17989469e-01 1.16425419e+00
-6.62342846e-01 5.11071146e-01 -9.60432589e-02 -7.06857383e-01
1.05785632e+00 3.85789245e-01 3.38633835e-01 -6.95816278e-01
5.77510059e-01 4.45951909e-01 -2.73900062e-01 -4.93898869e-01
9.38407421e-01 -1.12746269e-01 -4.92688268e-01 7.95510590e-01
-1.45302549e-01 -1.96045950e-01 6.38766170e-01 3.89140069e-01
4.93584186e-01 -1.24533959e-01 2.93367773e-01 -2.48879701e-01
9.30581927e-01 3.27225447e-01 1.26153588e-01 -2.90530380e-02
-2.28469372e-01 4.62304354e-01 7.11277723e-01 -1.88103288e-01
-9.02616680e-01 -6.53310567e-02 -3.36668253e-01 1.37962043e+00
-2.32945735e-04 -7.92281091e-01 -9.78405774e-01 -1.00876296e+00
-4.82982308e-01 3.40549499e-01 -6.08337998e-01 2.05150083e-01
-1.50948569e-01 -1.02783823e+00 3.55381817e-01 2.55892336e-01
4.71872956e-01 -1.29481745e+00 -3.00863147e-01 1.09569654e-01
-4.75030243e-01 -1.41254425e+00 -1.32884100e-01 1.77912414e-01
-2.11027801e-01 -1.08848858e+00 -4.64236408e-01 -7.69546449e-01
9.05055821e-01 -4.63763475e-02 1.15117145e+00 2.24930450e-01
6.63770318e-01 -2.15207390e-03 -9.69960451e-01 -8.82051528e-01
-6.50631130e-01 5.62649965e-01 1.50921553e-01 2.39134714e-01
7.88783967e-01 -1.24150766e-02 -3.47068429e-01 5.57158113e-01
-9.30214942e-01 -1.55889571e-01 2.39761829e-01 6.00377142e-01
1.51736811e-01 3.77013572e-02 9.26506162e-01 -1.33060014e+00
7.35222757e-01 -4.78763610e-01 -1.48988232e-01 -1.43901736e-01
-5.81200898e-01 -4.17844623e-01 7.13673651e-01 -5.74122481e-02
-9.49906528e-01 -4.39663306e-02 -3.39019388e-01 4.56360638e-01
-3.50102372e-02 9.44647312e-01 -1.38115706e-02 2.51774013e-01
7.50052571e-01 -2.37421244e-01 1.14129126e-01 -4.69838008e-02
1.88940704e-01 1.42408502e+00 -5.00804605e-03 -1.84114039e-01
4.80042487e-01 4.78497535e-01 -6.72898531e-01 -8.00433993e-01
-1.65450895e+00 -5.58850288e-01 -6.96481407e-01 -3.37239593e-01
7.86784887e-01 -1.06152356e+00 -6.11916900e-01 5.02642035e-01
-8.38394403e-01 -1.58260733e-01 -2.26451010e-02 3.89931709e-01
-1.00750141e-01 2.96050876e-01 -6.30557120e-01 -8.30066860e-01
-4.47437733e-01 -1.11081588e+00 9.38630521e-01 8.31846371e-02
-6.77029908e-01 -1.18081069e+00 3.65974344e-02 7.18714893e-01
3.88210624e-01 2.40464851e-01 3.49695355e-01 -9.57297444e-01
4.83282328e-01 -7.55114198e-01 -1.26480341e-01 5.80823243e-01
2.80414492e-01 2.74231076e-01 -1.10542238e+00 -9.29725692e-02
-7.59971887e-02 -8.41969013e-01 3.49983871e-01 -2.84242004e-01
7.61806488e-01 -4.93168890e-01 -5.38455285e-02 -3.04248393e-01
9.53162611e-01 -2.05428258e-01 3.99033010e-01 7.19985187e-01
5.60542047e-01 1.15709531e+00 9.95354235e-01 6.36000574e-01
8.39268565e-01 3.57223123e-01 3.44608724e-01 -2.35979289e-01
6.44664943e-01 1.53689170e-02 5.70314169e-01 1.52159607e+00
-1.02131583e-01 -3.94385040e-01 -1.07189858e+00 5.79128027e-01
-1.63185489e+00 -5.81496119e-01 -4.20219243e-01 1.86773682e+00
1.24555290e+00 3.82099122e-01 1.38581514e-01 4.23696697e-01
6.50978506e-01 5.46919405e-02 -5.41244112e-02 -4.37845767e-01
-3.57984245e-01 2.72516962e-02 3.60302448e-01 4.08266664e-01
-1.20200849e+00 1.10431969e+00 5.71991301e+00 6.66890383e-01
-1.22210133e+00 2.71229923e-01 9.50640857e-01 1.82457268e-01
-2.41799980e-01 -2.48674989e-01 -8.61032426e-01 5.17220080e-01
1.07780218e+00 9.82828662e-02 -1.56264424e-01 8.78453732e-01
2.41007984e-01 -6.98224157e-02 -5.69584429e-01 5.99026799e-01
3.08864176e-01 -8.88749778e-01 -3.52761358e-01 -3.67380194e-02
8.83624911e-01 1.49340317e-01 -2.13111788e-02 3.72236609e-01
1.59758806e-01 -7.93541193e-01 8.44945729e-01 5.90670258e-02
6.28455698e-01 -8.40831876e-01 1.32266879e+00 4.17748541e-01
-7.95922220e-01 3.16034883e-01 -5.21272570e-02 -3.00385773e-01
2.32098147e-01 7.10628808e-01 -6.38792455e-01 2.67071754e-01
8.24576795e-01 1.13611877e+00 -8.67554128e-01 2.06026882e-02
-6.43371999e-01 7.34086573e-01 -1.85749620e-01 -6.10379517e-01
3.47382456e-01 -3.19047004e-01 -2.48916224e-02 1.42364657e+00
-1.12017885e-01 -5.27040884e-02 2.22132519e-01 -6.12421855e-02
-2.77563959e-01 8.25419426e-01 -3.45234126e-01 -4.33794111e-01
9.71731842e-02 1.78268039e+00 -1.08236122e+00 -4.83742177e-01
-5.75125158e-01 7.15573132e-01 2.63091654e-01 1.21109828e-01
-5.96328676e-01 -4.61271763e-01 1.31710649e-01 7.17898235e-02
-8.31337720e-02 9.38604772e-02 -1.36645183e-01 -1.37823009e+00
2.09858432e-01 -1.02740061e+00 2.42027342e-01 -6.60258293e-01
-1.39913356e+00 1.07258868e+00 -5.34068108e-01 -1.32912731e+00
-2.20254019e-01 -6.83203876e-01 -1.74310952e-01 5.88496983e-01
-1.71412289e+00 -1.18275654e+00 -2.73532093e-01 3.23490590e-01
1.41320139e-01 -2.64996290e-01 9.05085266e-01 5.02241075e-01
-5.54866791e-01 5.81262767e-01 -1.27180398e-01 4.30957437e-01
1.11333919e+00 -1.14723253e+00 5.34509569e-02 5.41634619e-01
-1.46791205e-01 6.40703857e-01 7.18148410e-01 -4.50544536e-01
-1.02691448e+00 -9.52355623e-01 1.47567558e+00 -8.31617177e-01
1.25569975e+00 -3.00664663e-01 -7.24973202e-01 5.39475858e-01
4.28991467e-01 -1.05921395e-01 1.34763849e+00 2.74067014e-01
-1.45446882e-01 8.93950537e-02 -9.84493196e-01 4.36523080e-01
3.94114703e-01 -5.00749767e-01 -3.55226547e-01 4.85760331e-01
4.64405209e-01 -3.51564348e-01 -9.16602194e-01 2.32337236e-01
4.56524432e-01 -6.99592829e-01 3.29836667e-01 -5.40528059e-01
9.84375238e-01 -2.73143888e-01 -1.29180267e-01 -1.28016984e+00
2.49064878e-01 -2.26009130e-01 2.80205965e-01 1.54045284e+00
9.61379588e-01 -6.01730704e-01 8.26591074e-01 5.83577752e-01
4.64905687e-02 -6.51976526e-01 -1.73945904e-01 -2.32473668e-02
-7.96981230e-02 -6.71057045e-01 4.71643656e-01 1.44570267e+00
6.91578507e-01 7.05337763e-01 -1.29128456e-01 -3.64175402e-02
8.60856473e-03 4.55450565e-02 8.65034282e-01 -1.10751247e+00
2.44859949e-01 -3.72071594e-01 -1.48512810e-01 -6.55065656e-01
3.91166717e-01 -7.09691346e-01 -6.84959069e-02 -1.29956794e+00
-5.97013608e-02 -5.99561810e-01 -9.13376212e-02 6.03545845e-01
-1.68583021e-01 8.31168056e-01 -1.72959000e-01 3.24405044e-01
-1.02726173e+00 4.29338068e-01 1.11434650e+00 -7.42698908e-02
-1.76122308e-01 -8.15125555e-02 -1.04805756e+00 1.13385594e+00
7.82376826e-01 -3.78721416e-01 -1.41359726e-02 -1.02703266e-01
1.25019503e+00 -4.99204010e-01 -3.11371744e-01 -4.27799642e-01
-7.23992884e-02 -3.01873758e-02 1.42604709e-01 -4.53211069e-01
9.24624428e-02 -8.89284313e-01 -5.28812706e-01 1.52878106e-01
-3.01188022e-01 1.25678852e-01 -5.03217950e-02 1.48018345e-01
-6.65961683e-01 -4.16030109e-01 6.44932449e-01 1.86350942e-02
-3.97786677e-01 -8.60823784e-03 -5.26476502e-01 2.02071965e-01
8.52007270e-01 -6.43894896e-02 -3.05454582e-01 -4.92492884e-01
-7.28969276e-01 4.08978611e-01 6.99689627e-01 5.06442547e-01
1.57286704e-01 -1.11359620e+00 -7.71807551e-01 -3.12446021e-02
7.30795145e-01 3.26544456e-02 -1.36235982e-01 9.27078605e-01
-5.75569272e-01 9.29602087e-02 3.08140870e-02 -3.91468704e-01
-1.00680017e+00 2.38410518e-01 -1.99450642e-01 -2.50571758e-01
5.62813953e-02 4.56364512e-01 -3.09078276e-01 -7.05803335e-01
-8.76125470e-02 4.00803387e-02 -1.13846326e+00 8.57953548e-01
7.12118030e-01 9.70185697e-02 3.28878015e-01 -1.31202495e+00
-3.92125010e-01 3.53110522e-01 -5.93266562e-02 -1.94496453e-01
1.36297798e+00 -2.94074923e-01 -5.03297985e-01 6.02464557e-01
1.14972734e+00 5.56999266e-01 -5.26198447e-01 -1.31483480e-01
2.49135554e-01 -3.12994391e-01 -1.81586191e-01 -5.35434783e-01
-9.49794590e-01 5.84620774e-01 3.19534950e-02 7.97545016e-01
9.89467800e-01 -3.77761610e-02 7.04206884e-01 4.55771327e-01
2.11861297e-01 -1.47725046e+00 2.04939067e-01 9.64578748e-01
6.35457039e-01 -1.75092566e+00 -6.52942508e-02 -4.74160314e-01
-1.21705770e+00 9.85005438e-01 4.85030651e-01 1.24051599e-02
9.35595155e-01 1.44714847e-01 7.11441398e-01 -2.91351080e-01
-4.16576892e-01 -2.31889978e-01 1.44961357e-01 2.33348861e-01
9.53833997e-01 -2.09494662e-02 -6.61067724e-01 1.08851802e+00
-7.56499231e-01 -2.27407157e-01 7.62202740e-01 9.11911011e-01
-3.42383385e-01 -1.05492520e+00 -1.18744865e-01 2.77479798e-01
-1.06558621e+00 -1.22678682e-01 -5.47371149e-01 4.74558085e-01
-4.95622829e-02 1.38628626e+00 -5.33579476e-02 -2.43104577e-01
3.38230073e-01 -2.09389552e-01 -3.07279676e-01 -7.95742571e-01
-8.91903400e-01 1.03680618e-01 4.56413835e-01 -6.90464256e-03
-1.12573254e+00 -6.63107276e-01 -1.27427590e+00 -1.61863744e-01
-6.26927793e-01 6.85181677e-01 9.30965245e-01 1.18614030e+00
1.88924417e-01 1.91651419e-01 1.09564555e+00 -5.08855641e-01
-1.09563254e-01 -1.20226908e+00 -3.20565253e-01 5.83169222e-01
1.54386386e-01 -1.81495488e-01 -5.96383393e-01 4.42991912e-01] | [11.129498481750488, 6.933278560638428] |
ab2fee0d-2021-4b0c-a230-ad46f779546c | modeling-heterogeneous-hierarchies-with | 2110.14923 | null | https://arxiv.org/abs/2110.14923v2 | https://arxiv.org/pdf/2110.14923v2.pdf | Modeling Heterogeneous Hierarchies with Relation-specific Hyperbolic Cones | Hierarchical relations are prevalent and indispensable for organizing human knowledge captured by a knowledge graph (KG). The key property of hierarchical relations is that they induce a partial ordering over the entities, which needs to be modeled in order to allow for hierarchical reasoning. However, current KG embeddings can model only a single global hierarchy (single global partial ordering) and fail to model multiple heterogeneous hierarchies that exist in a single KG. Here we present ConE (Cone Embedding), a KG embedding model that is able to simultaneously model multiple hierarchical as well as non-hierarchical relations in a knowledge graph. ConE embeds entities into hyperbolic cones and models relations as transformations between the cones. In particular, ConE uses cone containment constraints in different subspaces of the hyperbolic embedding space to capture multiple heterogeneous hierarchies. Experiments on standard knowledge graph benchmarks show that ConE obtains state-of-the-art performance on hierarchical reasoning tasks as well as knowledge graph completion task on hierarchical graphs. In particular, our approach yields new state-of-the-art Hits@1 of 45.3% on WN18RR and 16.1% on DDB14 (0.231 MRR). As for hierarchical reasoning task, our approach outperforms previous best results by an average of 20% across the three datasets. | ['Jure Leskovec', 'Hongyu Ren', 'Rex Ying', 'Yushi Bai'] | 2021-10-28 | null | http://proceedings.neurips.cc/paper/2021/hash/662a2e96162905620397b19c9d249781-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/662a2e96162905620397b19c9d249781-Paper.pdf | neurips-2021-12 | ['ancestor-descendant-prediction'] | ['graphs'] | [-4.16377753e-01 8.21265399e-01 -3.96533638e-01 -6.85499683e-02
-2.53807575e-01 -7.65658557e-01 4.27391440e-01 5.49239039e-01
-9.03781876e-02 4.67141330e-01 6.99520946e-01 -2.44195163e-01
-7.00635493e-01 -1.29003561e+00 -7.54801154e-01 -3.39286476e-01
-3.30716103e-01 1.04435420e+00 6.24253094e-01 -4.21736807e-01
-4.57786381e-01 3.81981820e-01 -1.29715884e+00 5.13995647e-01
7.13213265e-01 7.60526299e-01 -4.20154393e-01 5.43135762e-01
8.86887908e-02 1.39229524e+00 -2.23404437e-01 -9.07261074e-01
1.70677736e-01 2.56223649e-01 -1.42113686e+00 -9.56587791e-02
6.91083193e-01 -1.64041325e-01 -1.01540303e+00 8.02072167e-01
2.05912054e-01 9.20782164e-02 7.30697334e-01 -1.31105494e+00
-1.10235536e+00 1.20571911e+00 -2.61015892e-01 2.84476038e-02
5.05132556e-01 -4.37402904e-01 1.93654847e+00 -9.71373022e-01
8.67481351e-01 1.28144228e+00 6.25605702e-01 1.78304061e-01
-1.36545372e+00 -3.52141500e-01 1.01705514e-01 6.22966826e-01
-1.82283998e+00 -9.35681164e-02 3.11901867e-01 -4.18466151e-01
1.49611449e+00 3.43033105e-01 8.31699431e-01 5.50878406e-01
1.19779490e-01 6.34416044e-01 7.67828286e-01 -3.20619017e-01
-6.18670620e-02 -1.75772056e-01 7.69166410e-01 1.20179009e+00
8.25001478e-01 -4.28509653e-01 -7.48499632e-01 -2.61118025e-01
7.23000169e-01 -3.27778459e-01 -2.42099062e-01 -7.96297669e-01
-1.19378555e+00 9.12226737e-01 1.06691098e+00 9.11858082e-02
-2.09959343e-01 3.09194624e-01 3.61464858e-01 2.39919901e-01
-2.10116059e-01 7.87390411e-01 -5.12413621e-01 4.11682457e-01
-2.99427986e-01 2.06482500e-01 1.15766680e+00 1.32528055e+00
8.95593107e-01 -3.40319991e-01 -2.75690645e-01 5.76155245e-01
1.89606607e-01 1.15590589e-02 -4.22603488e-02 -9.03270900e-01
6.10848904e-01 1.41519868e+00 -3.00000399e-01 -1.30083597e+00
-6.64118528e-01 -5.27834296e-01 -8.73138309e-01 -5.77536941e-01
1.02613017e-01 4.94826794e-01 -7.90216148e-01 1.48448956e+00
3.75613421e-01 5.05982339e-02 3.41544986e-01 5.04490495e-01
1.43691885e+00 3.23920935e-01 -1.59533560e-01 1.81719884e-01
1.76328826e+00 -1.12488270e+00 -6.30294502e-01 -4.10534665e-02
8.70707929e-01 -8.09698701e-02 9.12506580e-01 1.56442270e-01
-1.01997995e+00 -1.16640337e-01 -1.03916335e+00 -6.30978644e-01
-9.51868951e-01 -1.68590501e-01 1.08978796e+00 5.09417415e-01
-1.25784862e+00 6.52309954e-02 -6.23722494e-01 -4.35477614e-01
3.95119488e-02 5.11653066e-01 -7.23118484e-01 -3.30470860e-01
-1.66283929e+00 9.26904261e-01 9.79112327e-01 -1.79677329e-03
-6.57749236e-01 -1.03565323e+00 -1.06238496e+00 3.79269153e-01
8.99143815e-01 -1.10271788e+00 6.00244701e-01 3.44691396e-01
-7.03822076e-01 8.36642325e-01 1.10202476e-01 -5.89861095e-01
1.83179706e-01 -3.28512132e-01 -5.43983281e-01 3.67162347e-01
1.39652863e-01 6.38662457e-01 1.38332605e-01 -1.31272590e+00
-5.47616720e-01 -4.75431651e-01 7.39712775e-01 3.60649586e-01
-5.72248042e-01 -3.86211336e-01 -1.16210687e+00 -2.89868116e-01
2.46693164e-01 -1.09987986e+00 1.74740046e-01 -6.82960212e-01
-8.59865785e-01 -6.18756652e-01 6.39415503e-01 -6.30421579e-01
1.71060932e+00 -1.73501945e+00 6.24410033e-01 4.92799848e-01
9.70516443e-01 -2.59125412e-01 3.72052282e-01 6.82789564e-01
7.64672980e-02 3.51273686e-01 4.00098786e-02 -4.70458008e-02
4.63522136e-01 6.01000905e-01 -4.01097924e-01 1.23251632e-01
-1.61668569e-01 1.41347551e+00 -8.33943725e-01 -6.24808550e-01
-1.01760291e-01 4.13570940e-01 -7.74652302e-01 -6.39005378e-02
-1.74639612e-01 -4.35017437e-01 -1.88678637e-01 8.29314053e-01
3.84358674e-01 -7.41643608e-01 8.50891531e-01 -6.75994277e-01
4.48114336e-01 2.15894982e-01 -1.20448470e+00 1.36221731e+00
-6.03925474e-02 3.15476179e-01 -4.57185417e-01 -5.41147888e-01
4.93733346e-01 1.74384922e-01 3.76382619e-01 -3.32665682e-01
-1.86603799e-01 -3.82116577e-03 8.72261301e-02 -3.93451899e-02
7.28343189e-01 4.27761257e-01 -4.07793075e-01 -3.46978642e-02
2.31326804e-01 -3.17352824e-02 5.48038960e-01 1.02586055e+00
1.72289479e+00 -5.12809992e-01 4.34187323e-01 -4.46627229e-01
3.05272043e-01 -4.19089422e-02 5.93641460e-01 7.32231557e-01
3.28550786e-01 1.96345404e-01 1.01913679e+00 -5.68032026e-01
-8.08235288e-01 -1.28036821e+00 -3.23786922e-02 1.15294743e+00
3.52467954e-01 -1.23990941e+00 -2.39663884e-01 -9.50939894e-01
4.80329603e-01 5.01040876e-01 -9.90487099e-01 -2.49056280e-01
-4.17438298e-01 -5.93926013e-01 8.73888969e-01 8.09283376e-01
4.83673871e-01 -7.99626052e-01 -1.08026721e-01 1.11454204e-02
-1.38994932e-01 -1.83643508e+00 -1.79574683e-01 9.01877210e-02
-4.24575031e-01 -1.54891145e+00 2.08357155e-01 -8.06399882e-01
5.96232057e-01 9.62260365e-02 1.52973652e+00 2.98980772e-01
-3.52195889e-01 7.19981551e-01 -4.69684809e-01 2.88696736e-01
1.40310124e-01 4.66798753e-01 1.98333651e-01 -2.43989930e-01
2.39627942e-01 -5.73089063e-01 -3.17349702e-01 2.66191900e-01
-1.01762080e+00 2.53837388e-02 4.52795625e-01 7.95789659e-01
5.88885665e-01 7.30074644e-01 2.29092062e-01 -1.23524952e+00
5.98193109e-01 -3.16625357e-01 -4.61916983e-01 7.42755175e-01
-7.39212990e-01 2.24161848e-01 5.65728843e-01 3.99695076e-02
-7.12392569e-01 -1.95377931e-01 4.45105284e-01 -4.59066331e-01
3.26511800e-01 8.82514238e-01 -2.69673884e-01 5.79913147e-03
6.87655628e-01 -1.11969031e-01 -8.15723419e-01 -1.11024007e-01
9.21652436e-01 1.07998803e-01 6.43922389e-01 -8.69054019e-01
9.58245933e-01 4.97893095e-01 3.11458349e-01 -7.72773445e-01
-1.09956205e+00 -7.71509945e-01 -9.02018428e-01 1.88796505e-01
1.04332840e+00 -1.15524542e+00 -1.13958383e+00 -2.17134461e-01
-9.82055068e-01 -2.14407191e-01 -1.39470577e-01 2.93733720e-02
-2.80164808e-01 3.25856268e-01 -9.44181442e-01 -2.49333709e-01
-3.14018697e-01 -8.14261496e-01 1.00131571e+00 -1.23036534e-01
-3.26079309e-01 -1.00374997e+00 -8.38417783e-02 7.93276429e-01
-4.51840647e-02 1.83301657e-01 1.57810223e+00 -5.93576789e-01
-1.00302362e+00 -2.89870612e-02 -5.04889131e-01 -9.00110677e-02
-2.97594666e-02 -1.06194593e-01 -3.70765060e-01 -2.57250428e-01
-9.42972302e-01 -6.63510144e-01 1.15916085e+00 -2.94352353e-01
7.78485954e-01 -4.64937001e-01 -5.59137046e-01 5.58852851e-01
1.50817871e+00 -2.12749630e-01 8.01441908e-01 2.68506289e-01
1.36519265e+00 3.59437704e-01 1.94111988e-01 1.00338995e-01
1.09048808e+00 6.86846256e-01 4.25512493e-01 1.85439676e-01
-2.08078131e-01 -4.68380034e-01 -7.35345390e-03 9.62382138e-01
-2.96184957e-01 -4.48650345e-02 -1.38614988e+00 7.66341925e-01
-2.15084791e+00 -7.98324764e-01 -5.08649766e-01 1.79448068e+00
1.04431581e+00 1.55502826e-01 1.01045720e-01 2.35908502e-03
4.23706710e-01 6.54224604e-02 -2.76105702e-01 -1.28273740e-01
-3.81183118e-01 1.01649955e-01 7.07928181e-01 8.20581138e-01
-1.14201641e+00 1.42440403e+00 6.12382317e+00 5.66689014e-01
-1.44261092e-01 4.36608866e-03 -1.62736788e-01 1.58542126e-01
-5.13075411e-01 1.26701236e-01 -1.09638226e+00 -2.56402224e-01
5.04076481e-01 -2.40809053e-01 5.19210339e-01 6.73094034e-01
-8.86424720e-01 1.82971925e-01 -1.15683293e+00 9.24414754e-01
-4.68111131e-03 -1.50675678e+00 5.27452409e-01 2.64312208e-01
1.04200637e+00 -1.15289882e-01 -2.32519999e-01 7.44988263e-01
1.00048399e+00 -1.21819293e+00 3.52356613e-01 3.74978870e-01
6.10825837e-01 -8.27856719e-01 7.99742103e-01 -2.22597198e-04
-1.77700841e+00 -9.89607722e-02 -5.06103992e-01 2.10039496e-01
-1.51572093e-01 5.29826641e-01 -9.13877547e-01 1.11551452e+00
7.86362350e-01 5.15672386e-01 -9.97950494e-01 5.77164650e-01
-6.32200480e-01 3.46025825e-01 -2.66828030e-01 2.84770817e-01
1.02915771e-01 1.06807895e-01 2.36615568e-01 1.29353249e+00
-2.31596470e-01 2.81731009e-01 3.12546283e-01 6.36780620e-01
-5.60003519e-01 -2.18423568e-02 -6.00970030e-01 -2.98453122e-01
7.49139607e-01 1.25078344e+00 -6.95353150e-01 -3.93870324e-01
-3.63372386e-01 8.73322606e-01 1.00214660e+00 4.73820210e-01
-9.90978420e-01 -4.31835324e-01 4.45080906e-01 6.51830435e-02
3.39636713e-01 -3.71190786e-01 -8.89503136e-02 -1.39235413e+00
4.43783551e-02 -7.48767972e-01 1.09468818e+00 -7.25709796e-01
-1.21486962e+00 5.05296350e-01 1.36781752e-01 -3.12695354e-01
1.80350482e-01 -5.85325778e-01 1.55052796e-01 3.83803427e-01
-1.18216395e+00 -1.53929925e+00 -6.13984585e-01 9.01300490e-01
-9.65518132e-02 -1.15922637e-01 1.09806037e+00 2.93126494e-01
-5.07492721e-01 6.85176551e-01 -2.69754946e-01 3.38217825e-01
4.11356598e-01 -1.76790428e+00 2.62804359e-01 5.02273858e-01
4.34293896e-01 9.66889620e-01 4.34384376e-01 -8.51518929e-01
-1.63294017e+00 -1.09371567e+00 1.05654085e+00 -8.69215131e-01
1.12936628e+00 -4.82791781e-01 -1.05241108e+00 1.33625615e+00
2.18113244e-01 1.92687228e-01 6.78560495e-01 8.99951816e-01
-1.17196441e+00 -1.24732718e-01 -6.33670449e-01 8.23480487e-01
1.50929034e+00 -8.40372562e-01 -9.17343915e-01 3.69689286e-01
1.23261988e+00 -5.21121919e-01 -1.53657198e+00 8.16849768e-01
5.71341991e-01 -8.09183955e-01 1.19813371e+00 -9.57330287e-01
4.17075783e-01 -5.18836558e-01 -3.98694903e-01 -9.68085468e-01
-8.08121741e-01 -4.83632028e-01 -1.00234771e+00 1.01774013e+00
5.95675886e-01 -5.69464326e-01 9.77240264e-01 5.96504450e-01
4.36338075e-02 -9.63084996e-01 -7.05864012e-01 -1.00697124e+00
-2.35304177e-01 -1.58631176e-01 5.68410635e-01 1.33633137e+00
3.45471919e-01 8.47594738e-01 -6.45256042e-02 7.86362469e-01
7.49554634e-01 2.19856068e-01 8.28060508e-01 -1.34547496e+00
-1.82529464e-01 -2.70458907e-01 -8.37766707e-01 -7.81986356e-01
5.06706536e-01 -1.28343308e+00 -6.15054905e-01 -2.32406783e+00
5.30324817e-01 -2.58722305e-01 -3.80369753e-01 8.98066163e-01
-1.15096666e-01 4.63779978e-02 1.84342816e-01 1.22054555e-01
-1.07254326e+00 5.55238008e-01 8.35875154e-01 -4.14209157e-01
-1.58298746e-01 -7.09301710e-01 -9.22911882e-01 6.14854634e-01
4.57507282e-01 -1.88180611e-01 -7.32800603e-01 -4.38361436e-01
8.43617082e-01 -1.55897394e-01 4.06446606e-01 -7.41006970e-01
8.14365804e-01 2.22240165e-01 5.17724454e-02 -5.84831893e-01
4.63335544e-01 -8.53336751e-01 3.35340232e-01 1.94208726e-01
-7.48917833e-02 7.30252638e-02 1.59373328e-01 7.54636347e-01
-3.41241837e-01 2.94481307e-01 1.70300215e-01 5.85720539e-02
-9.99421537e-01 2.57472813e-01 2.59798914e-01 3.81819546e-01
9.41368639e-01 1.13981385e-02 -9.13254321e-01 -1.38427287e-01
-1.14153135e+00 7.18054771e-01 2.48449147e-01 4.67590928e-01
6.36581481e-01 -1.63198555e+00 -4.98609662e-01 -3.64230931e-01
5.34721613e-01 2.22811326e-01 2.57832229e-01 8.18799257e-01
-6.41330779e-01 7.69852817e-01 1.44442692e-01 -2.68949687e-01
-1.33923900e+00 8.10308397e-01 2.71523982e-01 -8.50613058e-01
-7.45917320e-01 8.90086651e-01 2.53654480e-01 -6.00568593e-01
2.95817256e-01 -3.99694562e-01 -1.19978942e-01 3.13601017e-01
6.37449920e-02 5.14453411e-01 1.98842600e-01 -6.37690425e-01
-6.44842088e-01 4.97890890e-01 -3.14146191e-01 2.47243643e-01
1.12553334e+00 3.15256983e-01 -6.77837670e-01 2.85023302e-01
1.08672237e+00 2.53008813e-01 -3.92305344e-01 -4.66945380e-01
3.40600282e-01 -1.45493671e-01 -1.19775973e-01 -6.92684233e-01
-7.29030609e-01 2.37176955e-01 -3.38445306e-01 7.02070177e-01
7.47688711e-01 5.81750214e-01 4.51360643e-01 9.00970817e-01
6.03990674e-01 -9.31470752e-01 1.39757067e-01 9.20118988e-01
1.10860229e+00 -7.15365827e-01 4.25765663e-01 -1.07253230e+00
-7.04047143e-01 7.87305534e-01 7.90542245e-01 1.15446478e-01
6.25763774e-01 -3.33741345e-02 -6.31549001e-01 -8.55597913e-01
-9.86130178e-01 -5.71451306e-01 7.14292347e-01 2.15779126e-01
1.90965071e-01 5.19249201e-01 4.42750938e-02 6.49335921e-01
-5.33475459e-01 -4.67164487e-01 3.35752428e-01 6.29283726e-01
-3.79475653e-01 -7.98563778e-01 -8.26998204e-02 4.54657763e-01
-1.19921878e-01 -3.13371271e-01 -9.37104642e-01 1.20053136e+00
5.75720631e-02 8.38976204e-01 -2.26339445e-01 -7.41332293e-01
6.97337925e-01 3.33966725e-02 5.81078231e-01 -9.04963195e-01
-3.42925072e-01 -4.04810041e-01 5.12956858e-01 -6.82733715e-01
-4.08526370e-03 -1.44961432e-01 -1.63096929e+00 -5.50460696e-01
-2.33601868e-01 2.41980731e-01 -2.71003813e-01 5.85132480e-01
3.93568486e-01 8.46061885e-01 -1.42413378e-01 7.33227134e-02
-8.79747570e-02 -6.34900510e-01 -1.03418720e+00 4.48381156e-01
-1.52109325e-01 -7.98670113e-01 -2.43855014e-01 -1.77652538e-01] | [8.77641773223877, 7.830065727233887] |
026867bf-a75d-4f6d-81a6-22711aac75fb | localization-of-ice-rink-for-broadcast-hockey | 2104.10847 | null | https://arxiv.org/abs/2104.10847v1 | https://arxiv.org/pdf/2104.10847v1.pdf | Localization of Ice-Rink for Broadcast Hockey Videos | In this work, an automatic and simple framework for hockey ice-rink localization from broadcast videos is introduced. First, video is broken into video-shots by a hierarchical partitioning of the video frames, and thresholding based on their histograms. To localize the frames on the ice-rink model, a ResNet18-based regressor is implemented and trained, which regresses to four control points on the model in a frame-by-frame fashion. This leads to the projection jittering problem in the video. To overcome this, in the inference phase, the trajectory of the control points on the ice-rink model are smoothed, for all the consecutive frames of a given video-shot, by convolving a Hann window with the achieved coordinates. Finally, the smoothed homography matrix is computed by using the direct linear transform on the four pairs of corresponding points. A hockey dataset for training and testing the regressor is gathered. The results show success of this simple and comprehensive procedure for localizing the hockey ice-rink and addressing the problem of jittering without affecting the accuracy of homography estimation. | ['Alexander Wong', 'John Zelek', 'David A. Clausi', 'Pascale Berunelle Walters', 'Mehrnaz Fani'] | 2021-04-22 | null | null | null | null | ['homography-estimation'] | ['computer-vision'] | [-1.62721902e-01 -3.14170659e-01 -6.66512921e-03 -1.73973545e-01
-5.87432623e-01 -3.85703206e-01 3.62800986e-01 1.95309650e-02
-5.31155586e-01 3.58976364e-01 1.33415470e-02 4.13185179e-01
-1.38436958e-01 -4.26061064e-01 -9.35317814e-01 -8.54455352e-01
-4.37842727e-01 1.08611673e-01 5.21627128e-01 -9.67385396e-02
2.21485868e-01 4.00855929e-01 -1.67100072e+00 2.45386615e-01
1.74620256e-01 1.06331658e+00 -7.16279000e-02 8.03454399e-01
4.26388383e-01 8.09370995e-01 -8.28745544e-01 -1.10641696e-01
4.65493113e-01 -4.33334231e-01 -4.58966583e-01 1.15671501e-01
8.36617470e-01 -4.95036602e-01 -8.78148139e-01 9.51967597e-01
-2.18118988e-02 3.39427322e-01 3.61903667e-01 -1.15397990e+00
1.57885298e-01 2.04175919e-01 -4.06004399e-01 1.51462555e-01
5.72808802e-01 -2.06925482e-01 7.08693743e-01 -7.40874529e-01
1.09188914e+00 1.03200042e+00 1.10389566e+00 -3.12211066e-02
-1.10849476e+00 -6.28547132e-01 -3.09136182e-01 7.70882726e-01
-1.80689704e+00 -3.53874683e-01 7.68300951e-01 -3.47361505e-01
5.85048079e-01 4.04082626e-01 1.18822801e+00 5.98994076e-01
4.03285891e-01 4.61951762e-01 8.71822357e-01 -4.00714815e-01
2.37956002e-01 -1.55193523e-01 1.10878401e-01 5.78146696e-01
-7.08478391e-02 -7.91139603e-02 -7.44627416e-01 9.92732495e-02
6.56313658e-01 -6.41666278e-02 -1.98087141e-01 -4.82738107e-01
-9.00106728e-01 5.48773825e-01 3.49442869e-01 2.49680102e-01
-5.17133296e-01 2.44261220e-01 5.17014802e-01 2.17303291e-01
4.92263019e-01 1.45380422e-01 -1.59614384e-01 -1.64094210e-01
-1.63761568e+00 2.45630488e-01 7.63280630e-01 8.21519017e-01
8.82183492e-01 -2.09903549e-02 1.03242733e-01 3.16647917e-01
1.53987914e-01 4.20702606e-01 2.30627477e-01 -9.86478448e-01
2.42309585e-01 3.20749402e-01 -1.66750327e-01 -1.58959854e+00
-5.01140714e-01 2.34888449e-01 -6.21045709e-01 1.61783189e-01
4.96644676e-01 -1.37057185e-01 -8.89888644e-01 1.25551295e+00
5.83029807e-01 6.19476080e-01 -1.49006486e-01 1.12281108e+00
6.24113202e-01 9.01924729e-01 -2.05087677e-01 -4.18297827e-01
1.27319157e+00 -5.13176143e-01 -8.35776269e-01 1.67952448e-01
4.18433756e-01 -6.76084399e-01 4.66563523e-01 4.05423939e-01
-8.82374346e-01 -6.34732425e-01 -1.18397224e+00 6.72048032e-02
-1.61992162e-01 3.31440359e-01 4.23006751e-02 6.59480393e-02
-9.34516370e-01 7.00779080e-01 -8.86741519e-01 -5.04629195e-01
-2.10860908e-01 2.57228523e-01 -6.32999480e-01 2.08026186e-01
-1.09022915e+00 9.27132845e-01 4.36480552e-01 5.06494224e-01
-9.66241419e-01 -3.74608815e-01 -8.79336953e-01 3.13421562e-02
4.73082393e-01 -8.85165483e-02 8.61602247e-01 -1.09733605e+00
-1.51042044e+00 6.54634297e-01 -3.33226956e-02 -8.48034918e-01
5.69043219e-01 -3.61378521e-01 -3.14903468e-01 5.61280310e-01
1.33968785e-01 4.82785851e-01 1.49939609e+00 -1.14779961e+00
-7.20633924e-01 -2.68140256e-01 -1.29065394e-01 2.87003011e-01
-9.32216868e-02 1.09737895e-01 -1.05558240e+00 -5.12081742e-01
5.22867084e-01 -1.16875815e+00 1.04750097e-01 -4.17030543e-01
-2.55464226e-01 9.28479508e-02 1.07956970e+00 -1.28423452e+00
1.50418520e+00 -2.49188924e+00 3.08298290e-01 4.94026363e-01
7.03835785e-02 1.14630714e-01 1.58579439e-01 3.89452815e-01
-4.27542031e-02 -5.45990884e-01 -3.55696976e-02 -1.09541290e-01
-3.48272145e-01 8.64065625e-03 -9.44996029e-02 1.06031215e+00
-3.68486643e-01 2.66558230e-01 -4.36146080e-01 -5.40077209e-01
3.79507095e-01 5.57451129e-01 -4.83930886e-01 1.70570195e-01
-9.68287215e-02 1.93582594e-01 1.00367528e-03 2.27312133e-01
7.08184242e-01 5.28923392e-01 2.74369180e-01 -4.82871145e-01
-5.19367278e-01 -1.01891220e-01 -1.34661186e+00 1.64376616e+00
-3.76629829e-02 1.10339916e+00 1.13491811e-01 -8.09660792e-01
9.38688457e-01 2.61008948e-01 9.01516557e-01 -4.08803046e-01
2.73976505e-01 -1.59154743e-01 -2.90553421e-01 -7.10339367e-01
8.32381785e-01 3.41503024e-01 -1.44789800e-01 -1.35304168e-01
3.78304332e-01 -4.55697477e-02 5.07518291e-01 1.05054602e-01
8.83511305e-01 3.23702753e-01 1.45077407e-01 -2.49924064e-01
5.29100895e-01 2.57127821e-01 5.32756448e-01 4.67113256e-01
-2.45721471e-02 7.22957969e-01 6.37845099e-01 -7.86669254e-01
-1.15310264e+00 -4.37097371e-01 -3.75730102e-03 8.45619380e-01
4.07847971e-01 -8.84423435e-01 -1.42180324e+00 -3.79046470e-01
-6.62491545e-02 3.58573526e-01 -7.44828761e-01 -1.20224535e-01
-8.22452903e-01 -1.83097824e-01 6.30709052e-01 1.21053576e-01
7.20563173e-01 -8.07625771e-01 -8.15040231e-01 2.47421801e-01
-3.65166575e-01 -1.24176288e+00 -4.52069402e-01 1.51923001e-01
-6.63017809e-01 -1.25452816e+00 -4.76848602e-01 -7.57487655e-01
6.80687964e-01 4.23439503e-01 5.01720488e-01 2.05115303e-02
-2.87523896e-01 3.22410971e-01 -3.42916578e-01 1.36556715e-01
-3.30209970e-01 -2.86900043e-01 2.54369825e-01 2.35520616e-01
4.23360258e-01 -1.71651810e-01 -3.09782624e-01 6.38527989e-01
-6.71431959e-01 -1.76542655e-01 2.14666069e-01 5.55624902e-01
7.42616057e-01 3.53317052e-01 -3.05273414e-01 -7.04926550e-02
1.45135418e-01 -1.05144978e-01 -1.10040545e+00 1.20283931e-01
-3.85110117e-02 -1.96448609e-01 6.52191162e-01 -3.42826784e-01
-6.16463482e-01 3.95416796e-01 2.02091157e-01 -8.95280421e-01
-9.31081250e-02 3.76826435e-01 5.16833253e-02 -3.59337717e-01
5.04912317e-01 2.21034408e-01 6.04637414e-02 -2.03766003e-01
2.77046770e-01 3.92199188e-01 8.88368368e-01 2.44107749e-02
1.04401255e+00 5.52619398e-01 7.36801047e-03 -1.49816573e+00
-5.21951795e-01 -7.03331411e-01 -9.63542998e-01 -7.96687484e-01
1.34417748e+00 -8.86506855e-01 -8.02421331e-01 5.28387666e-01
-1.14549601e+00 -4.27486263e-02 -1.28915116e-01 8.66169989e-01
-4.76732224e-01 5.97497702e-01 -6.05758786e-01 -4.85583514e-01
-4.47974913e-02 -1.10175431e+00 7.18089163e-01 2.61597037e-01
-2.98536658e-01 -5.19483626e-01 3.38066190e-01 1.01577118e-01
-2.06173941e-01 -3.90488334e-04 2.89903909e-01 -3.65210712e-01
-6.37374759e-01 -6.50993824e-01 1.33631825e-01 5.47049284e-01
-3.67307454e-01 4.55555677e-01 -6.59758031e-01 -3.15341502e-01
2.27925405e-01 1.18323676e-01 6.29345417e-01 8.51204932e-01
7.46465623e-01 -2.19263569e-01 -7.41964132e-02 1.02772582e+00
1.26483357e+00 2.70767480e-01 6.80847287e-01 6.57524288e-01
6.61202431e-01 3.99393618e-01 9.59876001e-01 4.88154441e-01
2.58945376e-01 9.90572155e-01 2.80418426e-01 3.67575064e-02
-2.90249940e-03 -2.95231611e-01 5.32718062e-01 7.99688637e-01
-5.26022375e-01 7.95463026e-02 -6.74720883e-01 2.52161652e-01
-1.72992122e+00 -1.20855463e+00 -2.50797778e-01 2.41391873e+00
1.86319992e-01 6.37334660e-02 1.66247249e-01 2.21459657e-01
9.30187404e-01 3.08229774e-01 -9.34075639e-02 -2.02727869e-01
-9.27193929e-03 -1.89629510e-01 8.66949975e-01 4.73251969e-01
-1.43003058e+00 1.17347741e+00 6.52255774e+00 7.45661557e-01
-1.26360977e+00 -2.02696517e-01 -4.69889045e-02 5.04680686e-02
4.91989404e-01 3.92724484e-01 -8.97747040e-01 5.21937311e-01
7.82103360e-01 6.72506317e-02 5.67210138e-01 8.49573374e-01
4.83394742e-01 -5.66304922e-01 -7.33589113e-01 1.11309123e+00
4.93153542e-01 -1.26854348e+00 -3.23975533e-01 -7.88100064e-02
4.35959071e-01 -2.42281452e-01 -2.63247699e-01 3.54231521e-02
-3.72549593e-01 -4.60189879e-01 9.82984722e-01 6.89528525e-01
5.62858582e-01 -8.86823893e-01 7.30067313e-01 8.81018564e-02
-1.30844760e+00 7.59434304e-04 -5.42587757e-01 1.01065235e-02
1.07620232e-01 1.31541178e-01 -8.58063519e-01 3.48610520e-01
9.93199646e-01 6.26656890e-01 -6.14683986e-01 1.20756721e+00
-2.19744712e-01 6.07929170e-01 -5.72037816e-01 2.10876986e-01
2.53892899e-01 -6.74125671e-01 8.05410981e-01 1.31208348e+00
4.26023513e-01 1.29311025e-01 8.82081091e-02 2.16082588e-01
1.64975047e-01 5.62426671e-02 -5.31144023e-01 3.20478618e-01
3.29725564e-01 1.35784900e+00 -9.48738515e-01 -5.13854444e-01
-1.73551187e-01 9.39210713e-01 -2.97065973e-01 3.70688081e-01
-1.17432690e+00 -5.08568585e-01 4.45014417e-01 3.54814798e-01
5.43870807e-01 -3.80017847e-01 1.56062916e-01 -9.84791577e-01
1.14578679e-01 -9.84049201e-01 3.97767395e-01 -7.40967155e-01
-3.25258076e-01 5.99532306e-01 2.65409410e-01 -1.59579706e+00
-4.39082831e-01 -6.83493838e-02 -4.69473630e-01 2.39914849e-01
-6.63128257e-01 -8.77515435e-01 -5.12404740e-01 8.08225691e-01
5.05325973e-01 -1.30520910e-01 4.95688409e-01 2.57031620e-01
-4.69385028e-01 2.30213627e-01 3.18507016e-01 1.81802914e-01
7.28861451e-01 -8.01675498e-01 1.25526562e-01 1.14604688e+00
3.07449520e-01 2.69430310e-01 9.63211596e-01 -7.91151762e-01
-1.42291284e+00 -6.43553257e-01 6.53285027e-01 -2.03720555e-01
6.92344487e-01 -5.04559517e-01 -8.78688574e-01 7.43171632e-01
1.17182046e-01 -2.18658403e-01 1.86768711e-01 -3.19225520e-01
1.41213927e-02 -4.62676883e-01 -8.51095378e-01 3.63163382e-01
4.08338577e-01 -4.94044811e-01 -6.98419452e-01 1.42329544e-01
2.72060603e-01 -6.19957745e-01 -7.68214583e-01 1.53861910e-01
6.03194773e-01 -1.18110633e+00 8.67209375e-01 -1.19332001e-01
2.66515881e-01 -5.94745815e-01 2.24642716e-02 -1.09841967e+00
-2.01896369e-01 -7.94814587e-01 2.07133204e-01 9.62190211e-01
-2.28745371e-01 1.76440254e-01 8.54925036e-01 1.85286820e-01
-1.99487314e-01 -1.08238876e-01 -1.17473650e+00 -3.99618596e-01
-7.15019107e-01 -3.60842377e-01 5.42167714e-03 6.53888583e-01
8.75612572e-02 1.39996810e-02 -8.86918366e-01 4.84307528e-01
7.83084393e-01 -2.45819598e-01 1.08065248e+00 -9.48691666e-01
2.07876638e-02 1.99973941e-01 -1.08269107e+00 -1.00781119e+00
1.10293619e-01 -3.87430906e-01 4.49253142e-01 -1.04883885e+00
-2.47946993e-01 1.45406306e-01 2.10052654e-01 -1.05440721e-03
3.42945427e-01 4.35848415e-01 4.21113193e-01 4.05201763e-01
-6.80203915e-01 2.07368344e-01 7.51056612e-01 6.46202192e-02
-3.45074534e-01 -6.99933171e-02 4.54808027e-01 1.27126420e+00
4.72802848e-01 -6.90217674e-01 7.49683529e-02 -9.95897055e-02
-2.31656909e-01 5.26183903e-01 4.62952375e-01 -1.66980588e+00
7.29907215e-01 1.99257061e-01 4.79539514e-01 -1.01501763e+00
5.79024792e-01 -1.04242015e+00 4.84898686e-01 3.73258501e-01
-1.53726161e-01 1.35038584e-01 2.41497569e-02 4.60792631e-01
-5.91424346e-01 -2.18312889e-01 8.38783205e-01 1.79768071e-01
-9.94719148e-01 1.14007063e-01 -7.07494259e-01 -3.52465034e-01
1.32446182e+00 -5.51201284e-01 3.03064436e-01 -4.34515834e-01
-9.16002214e-01 -1.22835234e-01 4.07354325e-01 1.82257503e-01
6.17054045e-01 -1.10057747e+00 -3.49518687e-01 5.55380821e-01
-2.87667453e-01 -1.60013333e-01 6.73529923e-01 1.01529479e+00
-1.14084840e+00 7.34530389e-02 -3.99060696e-01 -8.70848000e-01
-1.77007902e+00 4.86216873e-01 2.38850743e-01 1.38457716e-01
-8.44357967e-01 4.35928553e-01 -2.43045494e-01 2.77603209e-01
4.95818704e-01 -4.48453814e-01 -3.23037833e-01 5.32380044e-01
3.90979886e-01 6.54090762e-01 2.33259816e-02 -1.23188376e+00
-3.57108414e-01 8.95885348e-01 1.61157206e-01 -1.27669454e-01
1.26380467e+00 -2.39875302e-01 -3.50715041e-01 3.17284435e-01
1.34290552e+00 1.26196770e-02 -1.53004050e+00 2.09185511e-01
-4.75278236e-02 -4.95747149e-01 -6.89970255e-02 2.47446704e-03
-8.55136096e-01 3.86357605e-01 5.87530196e-01 2.20444113e-01
1.01079476e+00 -3.14615965e-01 5.55198848e-01 4.50233191e-01
1.76907405e-01 -1.39099634e+00 -2.81811774e-01 5.31829059e-01
5.92218518e-01 -6.80676401e-01 1.10395461e-01 -8.36873353e-02
-5.61019421e-01 1.48064780e+00 3.15086275e-01 -6.04756594e-01
5.49721718e-01 1.99216038e-01 2.38783821e-01 -1.91205114e-01
-2.31665209e-01 1.21173069e-01 2.38612473e-01 3.24592203e-01
2.47977804e-02 -2.55271912e-01 -3.88480812e-01 2.70794258e-02
-4.22002077e-01 -1.44411745e-02 6.38429821e-01 6.86025560e-01
-5.01380861e-01 -4.45293993e-01 -9.09243345e-01 -7.19947964e-02
-2.42551476e-01 1.19158126e-01 -3.40162486e-01 1.04581165e+00
2.59435415e-01 7.02457368e-01 3.07084858e-01 -7.10541785e-01
4.79095787e-01 -8.95671770e-02 2.19980046e-01 2.11457536e-02
-5.93578279e-01 4.95175749e-01 -4.16115634e-02 -8.19514453e-01
-4.17904437e-01 -6.35056496e-01 -1.07275105e+00 -4.56264555e-01
-2.29164436e-01 4.14955348e-01 7.70927191e-01 7.47479498e-01
-1.84214432e-02 2.40843549e-01 6.67892456e-01 -1.24195576e+00
-2.71449357e-01 -7.11347878e-01 -8.08828712e-01 4.87885356e-01
3.22880596e-01 -3.88234437e-01 -5.48750103e-01 3.74132246e-01] | [8.310474395751953, 0.01546634454280138] |
735d189a-71d8-4828-bb12-37402f0363b5 | autonomous-driving-on-curvy-roads-without | null | null | https://ieeexplore.ieee.org/document/9703250 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9703250 | Autonomous Driving on Curvy Roads Without Reliance on Frenet Frame: A Cartesian-Based Trajectory Planning Method | Curvy roads are a particular type of urban road scenario, wherein the curvature of the road centerline changes drastically. This paper is focused on the trajectory planning task for autonomous driving on a curvy road. The prevalent on-road trajectory planners in the Frenet frame cannot impose accurate restrictions on the trajectory curvature, thus easily making the resultant trajectories beyond the ego vehicle's kinematic capability. Regarding planning in the Cartesian frame, selection-based methods suffer from the curse of dimensionality. By contrast, optimization-based methods in the Cartesian frame are more flexible to find optima in the continuous solution space, but the new challenges are how to tackle the intractable collision-avoidance constraints and nonconvex kinematic constraints. An iterative computation framework is proposed to accumulatively handle the complex constraints. Concretely, an intermediate problem is solved in each iteration, which contains linear and tractably scaled collision-avoidance constraints and softened kinematic constraints. Compared with the existing optimization-based planners, our proposal is less sensitive to the initial guess especially when it is not kinematically feasible. The efficiency of the proposed planner is validated by both simulations and real-world experiments. Source codes of this work are available at https://github.com/libai1943/CartesianPlanner. | ['Youmin Zhang', 'Li Li', 'Yakun Ouyang', 'Bai Li'] | 2022-02-03 | null | null | null | ieee-transactions-on-intelligent-8 | ['trajectory-planning'] | ['robots'] | [-2.09029123e-01 3.02074160e-02 -3.54252219e-01 -1.14799216e-01
-3.97184938e-01 -5.09589136e-01 4.27959681e-01 -1.92827642e-01
-5.53690195e-01 9.06462908e-01 -2.20622912e-01 -8.65235686e-01
-4.60176438e-01 -9.88597631e-01 -5.85432589e-01 -7.75145054e-01
1.39013501e-02 3.17364603e-01 2.31467828e-01 -6.27165139e-01
4.61849660e-01 6.99602187e-01 -1.49433672e+00 -7.52954602e-01
1.53835106e+00 5.16668260e-01 3.18952441e-01 3.77520055e-01
1.19691372e-01 -1.18787996e-01 2.10398346e-01 -1.99860498e-01
2.95126319e-01 7.07808286e-02 -5.67965865e-01 -8.13210197e-03
-9.56482440e-02 -1.73672557e-01 -4.13105965e-01 1.15622711e+00
2.74692029e-01 5.57530820e-01 5.08813441e-01 -1.57861841e+00
8.49947333e-02 2.99866851e-02 -4.26995486e-01 6.79156482e-02
1.22314274e-01 2.99742818e-01 6.35094404e-01 -1.02973127e+00
6.14610970e-01 1.13186514e+00 5.33502877e-01 1.44257307e-01
-5.40220141e-01 -4.19180810e-01 3.63023490e-01 5.73286951e-01
-1.75166225e+00 -3.66315335e-01 7.63586462e-01 -3.11458319e-01
5.48555315e-01 4.56377566e-01 5.80835462e-01 5.57580173e-01
3.95068169e-01 4.40814942e-01 6.27745450e-01 8.70032683e-02
1.43391609e-01 -8.25586468e-02 5.84525689e-02 5.79778194e-01
4.78259444e-01 4.15698409e-01 9.74645168e-02 9.69157442e-02
5.01180530e-01 -9.18750316e-02 -2.82601595e-01 -4.96392041e-01
-1.35944676e+00 8.60879421e-01 4.96202856e-01 -5.08546643e-03
-3.74350756e-01 2.67156865e-02 2.74536729e-01 -5.64804375e-02
1.41637787e-01 2.84128666e-01 -1.26915485e-01 -4.96244460e-01
-5.91474593e-01 7.99408197e-01 5.04176676e-01 1.52757847e+00
8.55617702e-01 7.52433240e-02 2.17745185e-01 4.41515088e-01
4.13932353e-01 6.96585894e-01 -1.41298383e-01 -9.51519489e-01
8.45119059e-01 5.06833494e-01 5.01319826e-01 -1.19806731e+00
-7.46627271e-01 -3.69297773e-01 -7.51998305e-01 2.86314517e-01
5.64791381e-01 -4.76513594e-01 -5.69605052e-01 1.20214641e+00
8.06476235e-01 1.45856768e-01 9.37640816e-02 1.16972387e+00
4.56375599e-01 7.33115435e-01 -2.56797045e-01 -2.88666964e-01
1.07751203e+00 -1.01103461e+00 -9.36921895e-01 9.37710553e-02
7.92856693e-01 -8.47445607e-01 8.57699990e-01 2.67644860e-02
-1.03380609e+00 -3.45535964e-01 -1.02709591e+00 -1.00685731e-02
-4.98627722e-01 1.78776994e-01 2.96134859e-01 6.06764317e-01
-8.33637357e-01 2.41249353e-01 -7.75795579e-01 -1.76311359e-01
7.19102845e-02 2.24876896e-01 -1.12970784e-01 -4.91310395e-02
-1.29568088e+00 1.28956854e+00 4.32083309e-01 6.71537101e-01
-5.17149031e-01 -6.39775157e-01 -7.56083310e-01 -2.84791052e-01
7.20987558e-01 -3.36745799e-01 1.10047460e+00 -1.39327407e-01
-1.81323779e+00 1.43543273e-01 -4.76693004e-01 -1.25759766e-01
1.09392476e+00 -6.41473904e-02 -4.01445597e-01 -1.80339456e-01
8.89028013e-02 2.76659638e-01 4.81731296e-01 -1.31420732e+00
-8.97284389e-01 5.00998236e-02 2.15584144e-01 6.07308269e-01
2.21276119e-01 -4.33926910e-01 -5.76042473e-01 -2.06853569e-01
2.87565738e-01 -1.40245378e+00 -7.65408933e-01 -9.69794020e-02
-5.56927025e-01 -2.66176194e-01 1.09767044e+00 -1.69784904e-01
1.34838986e+00 -1.86886275e+00 -1.29778869e-04 4.80444014e-01
-7.84055516e-02 3.25892150e-01 2.63971806e-01 7.11814344e-01
3.24461222e-01 3.54694240e-02 -3.41522127e-01 8.26383159e-02
-7.60486955e-03 2.61188090e-01 -2.07538471e-01 8.08397830e-01
3.01364111e-03 6.41485214e-01 -1.14286137e+00 -4.85875577e-01
4.74661499e-01 3.62762004e-01 -5.53089738e-01 -1.50739744e-01
9.46028978e-02 6.24145985e-01 -1.01492417e+00 4.43307459e-01
1.12895024e+00 4.35305685e-01 -1.41284421e-01 -4.40649316e-02
-1.08196294e+00 1.12466261e-01 -1.43633699e+00 1.45890665e+00
-5.28082788e-01 4.03283715e-01 2.18209788e-01 -8.82262230e-01
9.86545980e-01 1.31340444e-01 5.83909512e-01 -5.30203044e-01
2.28153452e-01 4.69150126e-01 5.78313172e-02 -7.55580842e-01
9.72639024e-01 2.36403391e-01 -1.80352151e-01 2.28948854e-02
-8.31526101e-01 -4.81598169e-01 3.80308777e-01 -1.06314480e-01
5.78872025e-01 2.07574844e-01 2.11613849e-01 -5.92836261e-01
8.72428954e-01 5.83567381e-01 7.81652868e-01 1.73200086e-01
-2.44908765e-01 2.80670077e-01 2.58337945e-01 -4.22892690e-01
-1.06893337e+00 -8.87530088e-01 -5.87435007e-01 2.73642272e-01
9.94281054e-01 -2.71922141e-01 -6.65469825e-01 -3.15726876e-01
1.14015698e-01 8.37572277e-01 -8.30081925e-02 6.64040521e-02
-9.85853732e-01 -7.75651634e-01 2.15007856e-01 9.01177600e-02
5.72599351e-01 -3.76773417e-01 -8.19757760e-01 3.09922427e-01
-2.82372445e-01 -1.16082191e+00 -5.30894279e-01 -5.17873287e-01
-6.14210725e-01 -1.18631208e+00 -4.64170486e-01 -5.99808156e-01
7.92055130e-01 6.84076607e-01 2.67360538e-01 1.98285937e-01
1.06113613e-01 2.59750634e-02 -2.53122419e-01 -3.14060628e-01
-1.69125944e-01 2.72761941e-01 1.22906119e-01 3.24477628e-02
-3.01689692e-02 -2.52888262e-01 -7.23176956e-01 9.29653406e-01
-4.55791354e-01 3.70351762e-01 2.85444647e-01 5.42993724e-01
7.17576325e-01 5.43527842e-01 6.95684135e-01 -4.25753027e-01
5.10924339e-01 -7.87767112e-01 -8.53584409e-01 -2.73958713e-01
-5.54882824e-01 -1.61309540e-01 7.68525302e-01 -1.56947091e-01
-1.20837986e+00 2.85529703e-01 -3.48303944e-01 -6.69960454e-02
3.83668318e-02 5.62790394e-01 -2.85208255e-01 -2.87492096e-01
2.57380635e-01 5.92371449e-02 -9.33329612e-02 -1.67820722e-01
5.19357204e-01 4.34871703e-01 4.25849497e-01 -5.99798501e-01
1.16727197e+00 5.97140014e-01 4.18730110e-01 -9.72298384e-01
-6.68997839e-02 -5.25251150e-01 -8.96090031e-01 -6.94342732e-01
7.29119241e-01 -6.11397028e-01 -9.64761019e-01 3.00580055e-01
-1.10133386e+00 -1.31303102e-01 2.25254484e-02 7.19110429e-01
-7.60328233e-01 4.33043480e-01 6.54768273e-02 -8.29002678e-01
8.96270275e-02 -1.57044721e+00 5.42197406e-01 3.98927063e-01
1.25299692e-01 -9.72506464e-01 -1.83167532e-01 6.76530600e-02
5.03918529e-01 6.26723826e-01 5.71117043e-01 2.17002228e-01
-8.81017745e-01 -1.92531332e-01 -1.56841829e-01 -4.26053882e-01
-1.01398490e-01 2.30680376e-01 -2.63747811e-01 -2.13228479e-01
-1.61416933e-01 3.68799388e-01 3.42467248e-01 2.45109424e-01
8.41608107e-01 -3.87965679e-01 -6.49316192e-01 6.41282022e-01
1.39495850e+00 5.04722953e-01 5.83593428e-01 6.24763787e-01
7.46086717e-01 6.28589272e-01 1.30909574e+00 3.94336551e-01
9.46191430e-01 8.69135857e-01 7.96615899e-01 -4.49422263e-02
3.07614923e-01 -3.54544759e-01 1.45024389e-01 7.78396964e-01
-4.47856277e-01 -2.74280399e-01 -1.13516009e+00 7.79774070e-01
-2.15141892e+00 -7.94662058e-01 -9.57617819e-01 2.12236786e+00
4.54410136e-01 -1.04117217e-02 -9.19119455e-03 1.86815098e-01
8.46002281e-01 1.98877808e-02 -6.05539083e-01 -5.86559534e-01
2.45460480e-01 -5.96201062e-01 1.16481817e+00 9.20250237e-01
-9.97560859e-01 8.95606339e-01 5.58514738e+00 9.24171448e-01
-1.15807402e+00 2.90047238e-03 5.70732541e-02 4.15834486e-02
-4.38305289e-01 2.10886270e-01 -1.00152063e+00 5.86632550e-01
8.16291511e-01 -5.39447784e-01 3.07823151e-01 7.62088239e-01
9.78380382e-01 -3.53616178e-01 -4.40263540e-01 6.22159302e-01
-5.53911865e-01 -1.33547401e+00 -3.89955819e-01 1.13396339e-01
7.79898882e-01 4.72975634e-02 5.99967353e-02 1.48021290e-02
-1.86596345e-02 -8.15896928e-01 8.07032168e-01 6.24178290e-01
7.02868700e-01 -1.12740839e+00 5.47940910e-01 6.50492430e-01
-1.47644305e+00 3.65768075e-02 -2.81801969e-01 -1.45047218e-01
8.46559167e-01 4.78482455e-01 -8.83065939e-01 9.19599116e-01
3.68070096e-01 7.70097673e-01 -9.90661010e-02 1.39875746e+00
-1.58572271e-01 2.22394228e-01 -4.46671814e-01 -2.50685573e-01
7.08845019e-01 -7.93776810e-01 1.05294406e+00 9.79285181e-01
5.37173390e-01 1.96290240e-01 3.48976701e-01 7.46080995e-01
5.47580540e-01 2.29045376e-01 -7.99351752e-01 5.21687746e-01
5.78667581e-01 1.16330600e+00 -5.66740990e-01 -9.48117822e-02
-3.64072770e-01 1.83956251e-01 -2.84557045e-02 7.02062190e-01
-1.26664293e+00 -5.98423541e-01 8.22038233e-01 3.82977992e-01
-5.31770922e-02 -7.57102728e-01 -4.83302563e-01 -9.02337074e-01
1.90281838e-01 -3.13910574e-01 -2.27575734e-01 -3.01929384e-01
-4.69657153e-01 4.72948581e-01 4.20761555e-01 -1.61330640e+00
-4.00906280e-02 -2.94212550e-01 -8.26017618e-01 9.08256769e-01
-1.97239470e+00 -8.86252165e-01 -3.72428298e-01 5.07957518e-01
5.72456837e-01 1.43215254e-01 2.77187914e-01 4.96129692e-01
-7.67217934e-01 3.23711395e-01 4.73418117e-01 -4.58330065e-01
1.52816236e-01 -8.42776358e-01 3.10786903e-01 1.02689171e+00
-9.63366926e-01 5.09858370e-01 9.53496397e-01 -6.45711958e-01
-1.82044542e+00 -1.27355039e+00 9.37228262e-01 -1.58219971e-02
7.17061520e-01 -8.07537958e-02 -7.74786711e-01 4.64901239e-01
-1.76944271e-01 -9.12864693e-03 -1.53291807e-01 -3.23471934e-01
4.79634792e-01 -1.59551552e-03 -1.11117721e+00 8.97718132e-01
1.20405042e+00 1.17741220e-01 -9.01222676e-02 3.75580370e-01
5.15307128e-01 -8.89122009e-01 -8.63043249e-01 6.66570127e-01
5.57412505e-01 -5.17553627e-01 8.55818748e-01 -1.33883938e-01
-4.91644777e-02 -8.46808732e-01 2.06417814e-01 -1.22310841e+00
-9.07564238e-02 -1.08457255e+00 2.89016694e-01 8.88740122e-01
4.53837335e-01 -1.05892813e+00 7.06504226e-01 5.28612375e-01
-7.37492561e-01 -1.18863463e+00 -1.35369289e+00 -8.74447763e-01
2.09634155e-01 -6.57925546e-01 8.41619194e-01 6.91408515e-01
1.83620393e-01 -1.95131063e-01 -2.55601406e-01 5.46118021e-01
5.01489580e-01 3.61796990e-02 1.00851738e+00 -8.14578116e-01
5.37132800e-01 -4.85511422e-01 -2.13534877e-01 -1.20467901e+00
1.37841836e-01 -8.79128098e-01 4.08677340e-01 -1.49432039e+00
-5.55258155e-01 -1.34273219e+00 4.87410843e-01 5.67572564e-02
4.44506854e-03 -3.40465367e-01 9.29298922e-02 1.89222723e-01
-1.43112674e-01 8.62414062e-01 1.80621171e+00 3.19919214e-02
-4.78939503e-01 4.44870532e-01 -2.01993525e-01 8.25018823e-01
1.13947845e+00 -2.30633840e-01 -8.23581517e-01 -3.88611019e-01
2.25009754e-01 3.03236574e-01 2.46523902e-01 -8.13039064e-01
5.21672606e-01 -7.64069378e-01 -6.35220885e-01 -1.10746527e+00
3.61160219e-01 -9.25684392e-01 3.26517224e-01 5.06118715e-01
1.45453274e-01 3.57017815e-01 1.52981028e-01 6.09411716e-01
-1.90088928e-01 -2.61719644e-01 7.72770107e-01 2.11464688e-01
-8.29896569e-01 5.75519621e-01 -6.83114469e-01 -1.40652228e-02
1.55990005e+00 -4.38780189e-01 -3.30556750e-01 -1.35560617e-01
-4.36026573e-01 7.73025692e-01 5.81577599e-01 4.48598087e-01
6.32718265e-01 -1.35117984e+00 -6.83595896e-01 1.59919873e-01
-4.39641774e-02 4.24425811e-01 3.98764133e-01 1.29750109e+00
-8.23393703e-01 7.57572114e-01 -1.11240394e-01 -5.46161652e-01
-7.11797655e-01 5.02424121e-01 4.17906791e-01 5.41764311e-02
-8.32717538e-01 3.05112928e-01 4.38019000e-02 -4.84477967e-01
-2.19584584e-01 -5.36407471e-01 -3.31380665e-01 -1.36353597e-01
1.06243864e-01 9.36353445e-01 -4.74341176e-02 -1.22664571e+00
-3.86553079e-01 8.96054447e-01 3.39647591e-01 -7.81749859e-02
9.58188534e-01 -6.59276545e-01 1.55854121e-01 3.95796038e-02
1.05846488e+00 -1.83552597e-02 -1.36515641e+00 1.22546323e-01
-7.48765394e-02 -6.31937921e-01 9.43336412e-02 -1.07859217e-01
-8.98523629e-01 6.11707509e-01 1.81141600e-01 9.71915945e-02
6.89574242e-01 -6.26887262e-01 1.07277012e+00 2.77620822e-01
5.50836980e-01 -1.31142449e+00 -6.47571206e-01 8.58631849e-01
9.87467945e-01 -9.79307413e-01 1.38361529e-01 -9.12639558e-01
-5.87645531e-01 1.22459340e+00 5.47576427e-01 -2.37625286e-01
8.72602522e-01 -1.22006414e-02 -1.99692383e-01 4.08305675e-02
-3.72743100e-01 -2.68163979e-01 7.45266601e-02 4.30332780e-01
3.82861169e-03 3.21151227e-01 -1.04315817e+00 9.38786715e-02
-5.17508984e-01 -2.38800719e-01 8.42248380e-01 8.52659822e-01
-5.10984540e-01 -9.45221484e-01 -5.12594581e-01 7.48026371e-02
8.94129425e-02 3.01687121e-01 5.24749577e-01 1.07161045e+00
3.82996857e-01 1.20120955e+00 -6.24356493e-02 -1.83628872e-01
6.50289774e-01 -4.28252548e-01 -1.42576680e-01 -1.92663446e-01
-2.74830498e-02 -1.99425384e-01 5.02115369e-01 -6.24735653e-01
-2.58730888e-01 -1.08278680e+00 -1.86277592e+00 -5.45936048e-01
-4.06550437e-01 2.80790597e-01 9.73717511e-01 9.32906449e-01
2.88825423e-01 3.30881327e-01 8.97202373e-01 -1.00284767e+00
-2.98504382e-01 -3.86384100e-01 -1.87969789e-01 -1.19654797e-01
4.37169850e-01 -1.16070044e+00 -2.69714057e-01 -3.47924888e-01] | [5.246136665344238, 1.643401026725769] |
ae180025-ea7b-4302-a916-539217e1e030 | smart-parking-space-detection-under-hazy | 2201.05858 | null | https://arxiv.org/abs/2201.05858v1 | https://arxiv.org/pdf/2201.05858v1.pdf | Smart Parking Space Detection under Hazy conditions using Convolutional Neural Networks: A Novel Approach | Limited urban parking space combined with urbanization has necessitated the development of smart parking systems that can communicate the availability of parking slots to the end users. Towards this, various deep learning based solutions using convolutional neural networks have been proposed for parking space occupation detection. Though these approaches are robust to partial obstructions and lighting conditions, their performance is found to degrade in the presence of haze conditions. Looking in this direction, this paper investigates the use of dehazing networks that improves the performance of parking space occupancy classifier under hazy conditions. Additionally, training procedures are proposed for dehazing networks to maximize the performance of the system on both hazy and non-hazy conditions. The proposed system is deployable as part of existing smart parking systems where limited number of cameras are used to monitor hundreds of parking spaces. To validate our approach, we have developed a custom hazy parking system dataset from real-world task-driven test set of RESIDE-\b{eta} dataset. The proposed approach is tested against existing state-of-the-art parking space detectors on CNRPark-EXT and hazy parking system datasets. Experimental results indicate that there is a significant accuracy improvement of the proposed approach on the hazy parking system dataset. | ['Rajendra Kumar Roul', 'Jajati Keshari Sahoo', 'Gaurav Satyanath'] | 2022-01-15 | null | null | null | null | ['parking-space-occupancy'] | ['computer-vision'] | [-2.62966305e-01 -7.97773227e-02 3.49067718e-01 -2.86695004e-01
-2.90276140e-01 1.64456457e-01 8.18432689e-01 -3.40390682e-01
-6.73453987e-01 8.42536569e-01 -1.15066119e-01 -6.17658556e-01
1.34792268e-01 -1.11829197e+00 -4.78472352e-01 -7.96815574e-01
1.79602280e-01 4.02872086e-01 5.20104468e-01 -6.02162302e-01
1.81082755e-01 4.51725662e-01 -2.14846778e+00 -1.44589245e-01
8.19072247e-01 7.67856956e-01 6.94986522e-01 7.47495353e-01
9.32552516e-02 4.26488996e-01 -3.78401101e-01 3.74616265e-01
7.27841198e-01 5.02752006e-01 -1.82545155e-01 3.07631046e-01
6.41116321e-01 -3.96016747e-01 -4.84989226e-01 8.49562764e-01
7.04570770e-01 1.56136289e-01 5.53357124e-01 -1.32244527e+00
-5.52257061e-01 1.51680000e-02 -1.35621145e-01 7.84212410e-01
-5.92107296e-01 4.95245814e-01 5.43060541e-01 -1.43864202e+00
-2.42428645e-01 1.05015385e+00 4.48729306e-01 -2.64311898e-02
-6.98302388e-01 -7.64208198e-01 -5.45847714e-01 5.68809628e-01
-1.70871556e+00 -4.80439603e-01 6.10330403e-01 -2.92278260e-01
1.39793432e+00 1.25161245e-01 4.32285011e-01 4.09767419e-01
1.89147532e-01 6.51891708e-01 1.31071234e+00 -4.97878522e-01
4.29004371e-01 6.08520031e-01 -2.53982954e-02 4.29562449e-01
2.33684555e-01 4.49464619e-01 -9.79380235e-02 4.47363406e-01
2.38899410e-01 1.72906339e-01 5.03078222e-01 -2.44744152e-01
-6.18911266e-01 1.16485465e+00 6.95863545e-01 2.72193015e-01
-4.61173683e-01 -2.19592489e-02 2.56433599e-02 -1.28074318e-01
2.98960924e-01 5.53273223e-03 -9.58336815e-02 1.01677597e-01
-1.27543199e+00 3.22378546e-01 2.59669989e-01 8.53134274e-01
1.09291112e+00 5.61327219e-01 1.70361385e-01 9.30624425e-01
6.01293266e-01 8.43655050e-01 4.75717485e-01 -6.10034108e-01
4.59947169e-01 5.86205602e-01 1.99410424e-01 -7.74518728e-01
-2.81177282e-01 -2.18159094e-01 -4.52326536e-01 6.10746682e-01
-1.86712313e-02 3.43597054e-01 -1.54853606e+00 9.79344606e-01
1.74196020e-01 1.38363540e-01 3.38045396e-02 8.93603742e-01
7.54749715e-01 6.55413628e-01 2.23197371e-01 5.97921073e-01
1.29079318e+00 -9.86208916e-01 -4.14694041e-01 -4.80268329e-01
4.16545928e-01 -7.73315012e-01 1.01557350e+00 1.73198730e-02
-7.98495591e-01 -8.16599667e-01 -1.38237917e+00 6.04143552e-02
-1.29742336e+00 1.68727264e-01 2.15960503e-01 1.04408586e+00
-1.41031122e+00 3.64972912e-02 -7.11982906e-01 -7.29165852e-01
5.36921978e-01 5.18190324e-01 6.15314767e-02 -1.90382123e-01
-1.47157419e+00 1.35822988e+00 4.31464583e-01 2.61829704e-01
-1.14648652e+00 -1.73616171e-01 -8.86885703e-01 1.03970423e-01
-1.47335991e-01 -1.81241810e-01 8.78178060e-01 -3.02468985e-01
-8.18545878e-01 8.29350948e-01 1.58012599e-01 -8.01030159e-01
1.47131443e-01 1.99540462e-02 -6.90250278e-01 -3.27696711e-01
5.01511037e-01 1.10960019e+00 5.97241998e-01 -1.01400912e+00
-9.53028262e-01 -4.32386994e-01 -1.20661885e-01 4.03028071e-01
-2.64484048e-01 -3.42868298e-01 -1.57965228e-01 8.43368992e-02
-2.95166999e-01 -1.05313003e+00 -1.72786117e-01 -7.71089345e-02
-3.65703791e-01 -1.88912436e-01 1.59760213e+00 -4.08171445e-01
8.61542344e-01 -2.16154099e+00 -1.01722252e+00 9.49064568e-02
-3.28068137e-02 8.38113844e-01 2.76025563e-01 2.46374786e-01
1.07613571e-01 -2.14465499e-01 -1.67814910e-01 -2.98343122e-01
3.09004605e-01 7.27681458e-01 4.34546452e-03 6.13556027e-01
2.91981399e-01 8.32435787e-01 -2.03069046e-01 -7.04346538e-01
1.31342769e+00 6.79268360e-01 -2.18796328e-01 -8.34467337e-02
2.20213294e-01 -8.39012787e-02 -1.53900266e-01 1.07142794e+00
9.93008792e-01 4.25491109e-03 -7.33774126e-01 3.05294931e-01
-6.36697710e-01 1.50250450e-01 -9.63770986e-01 8.66721869e-01
-6.61624730e-01 1.01263380e+00 2.13996276e-01 -1.27803457e+00
1.15883267e+00 -6.01007268e-02 -1.56817902e-02 -1.34075332e+00
2.02277258e-01 2.81458408e-01 -9.38504264e-02 -3.53158057e-01
1.14661205e+00 -7.21737817e-02 1.15624033e-01 -2.61485934e-01
-2.98527807e-01 2.02788860e-01 -1.27447367e-01 -1.77487075e-01
1.06653905e+00 -4.70593810e-01 1.54455543e-01 -3.88595909e-01
7.21584618e-01 2.44220436e-01 3.72070968e-02 7.40549266e-01
-8.75140369e-01 3.76464784e-01 -7.38769412e-01 -6.49483979e-01
-1.66490662e+00 -1.34444380e+00 -5.05487442e-01 8.33309114e-01
1.57533556e-01 2.67672151e-01 -4.70707893e-01 5.56817241e-02
1.84528857e-01 8.44209969e-01 -5.31669319e-01 6.14093542e-02
-3.53655607e-01 -8.25153053e-01 5.05872846e-01 4.73680973e-01
1.22594261e+00 -1.11920071e+00 -9.21041608e-01 2.07714394e-01
2.62384772e-01 -1.31746900e+00 1.69398636e-01 3.41785938e-01
-4.53453213e-01 -8.65630627e-01 -5.32698810e-01 -7.48307049e-01
3.28258574e-01 8.28361690e-01 8.99454415e-01 1.32676482e-01
-6.29235506e-01 2.21711978e-01 -8.99343267e-02 -7.33277977e-01
-1.82156742e-01 -7.35781416e-02 -5.88071048e-02 -4.08706069e-01
1.15256774e+00 -5.76718926e-01 -9.45387542e-01 4.94145393e-01
-8.08335960e-01 -2.29104176e-01 7.38563895e-01 7.39525557e-01
3.89571548e-01 7.76559293e-01 3.37334216e-01 -1.56233132e-01
3.15123975e-01 -8.58592391e-01 -8.50197375e-01 -4.52036530e-01
-1.11397517e+00 -1.77977040e-01 3.85825694e-01 1.47249490e-01
-6.52650535e-01 2.00370818e-01 -1.03969283e-01 -4.57985997e-01
-6.34285510e-01 -6.56861588e-02 -7.64574930e-02 -9.33550671e-02
5.70275366e-01 4.63539213e-01 -9.82033908e-02 -1.15968093e-01
3.31791133e-01 1.21993065e+00 7.90746391e-01 6.92043751e-02
1.08707190e+00 7.17474520e-01 -1.10444135e-03 -1.36154306e+00
2.06944458e-02 -1.01439905e+00 -4.84281659e-01 -3.44655156e-01
1.10357916e+00 -1.32014406e+00 -2.87033379e-01 2.63285786e-01
-6.00563943e-01 -4.10135323e-03 1.59086004e-01 1.38425201e-01
-2.43081152e-01 2.34682634e-01 1.87598169e-01 -1.27410626e+00
-5.30489028e-01 -1.20472324e+00 1.11838758e+00 4.53698784e-01
3.26922357e-01 -9.24068332e-01 -1.22414805e-01 7.59571552e-01
9.51706946e-01 1.70239341e-02 3.97750437e-01 -3.68741840e-01
-9.77405787e-01 -4.26887721e-01 -3.16441357e-01 4.75591987e-01
2.44254041e-02 -4.78750199e-01 -1.33231044e+00 -9.03280824e-03
-4.15810823e-01 3.35892774e-02 9.26886678e-01 4.85904902e-01
6.86217368e-01 -8.48817229e-02 -2.20459566e-01 1.55812263e-01
1.53927076e+00 -2.35699746e-03 1.12482512e+00 1.17796075e+00
3.90497148e-01 1.50530785e-01 8.35576892e-01 4.30147886e-01
6.22349083e-01 6.59683764e-01 1.17083478e+00 -3.54942888e-01
-1.69086993e-01 -4.95126359e-02 3.02051663e-01 1.31936342e-01
4.00341332e-01 4.40948829e-02 -1.07741797e+00 1.34817767e+00
-1.63658500e+00 -8.62556875e-01 -2.28434324e-01 1.97415102e+00
4.63597896e-03 3.03290576e-01 2.64889985e-01 3.26793760e-01
7.74855912e-01 3.12835090e-02 -3.53217423e-01 -7.06707656e-01
-1.85135216e-01 4.79181170e-01 1.28214526e+00 4.43805099e-01
-1.36905336e+00 1.09407163e+00 5.77103806e+00 5.65919578e-01
-1.15069902e+00 1.97618648e-01 2.05057636e-01 -3.29225101e-02
9.64189172e-02 -1.84640408e-01 -9.27219748e-01 9.64336753e-01
1.44152069e+00 1.45637155e-01 2.82492757e-01 1.27259839e+00
8.18696916e-01 -5.62529743e-01 -1.97792843e-01 8.73291612e-01
-1.67602628e-01 -1.30138087e+00 -5.10707915e-01 5.50956368e-01
4.59308267e-01 7.87774324e-01 4.22755867e-01 6.05539978e-01
5.24221539e-01 -1.24657810e+00 3.01576942e-01 1.67599786e-02
4.74048227e-01 -6.92772031e-01 1.00265586e+00 4.58213866e-01
-1.21801353e+00 -2.30380788e-01 -7.22814679e-01 -3.88480812e-01
1.00243345e-01 3.27625602e-01 -1.85783136e+00 -5.99185079e-02
1.01775849e+00 2.60908544e-01 -1.03397715e+00 1.10327399e+00
2.26897925e-01 5.46963096e-01 -5.64809203e-01 -1.16594248e-01
7.57391870e-01 1.93014786e-01 1.44254461e-01 1.14276993e+00
5.78959763e-01 -2.79301912e-01 6.80700690e-02 1.11430025e+00
5.61745524e-01 -2.26192787e-01 -1.15117085e+00 2.88276821e-01
6.31821692e-01 1.33961904e+00 -5.14767468e-01 -2.46790484e-01
-5.65051913e-01 4.26490396e-01 -1.16462357e-01 -4.50408012e-02
-8.10239553e-01 -7.52587169e-02 7.18953371e-01 6.47578895e-01
8.62379014e-01 -4.36153024e-01 -3.46496820e-01 -3.47403854e-01
-2.46727422e-01 -2.31962651e-01 -1.02627464e-02 -1.14916837e+00
-8.99430454e-01 2.14732438e-01 -3.51102613e-02 -1.10735393e+00
1.49072871e-01 -8.43857527e-01 -9.37227786e-01 1.00076044e+00
-2.15274024e+00 -1.29498208e+00 -6.23651028e-01 7.64887393e-01
8.45034182e-01 -5.94962060e-01 6.10894203e-01 7.19241083e-01
-3.53030086e-01 2.77551979e-01 1.87472552e-01 -9.46773067e-02
4.05369252e-01 -1.15423930e+00 3.03816885e-01 1.15466261e+00
-3.34848583e-01 2.06243664e-01 1.08320034e+00 -3.82327676e-01
-1.00984621e+00 -1.56401467e+00 7.13268399e-01 -4.81256813e-01
4.78050917e-01 -2.15583771e-01 -6.26144290e-01 2.76425749e-01
3.14357370e-01 4.24404070e-02 4.26127940e-01 -3.76029134e-01
5.64075261e-02 -1.34475559e-01 -1.28294957e+00 3.15512657e-01
8.59002843e-02 -5.41296721e-01 -6.10486507e-01 2.53159732e-01
1.29695058e-01 -7.22984374e-02 -3.63391399e-01 3.77318621e-01
1.25504017e-01 -1.07532609e+00 1.00119627e+00 1.36540443e-01
1.35248289e-01 -7.67382979e-01 -7.92177141e-01 -9.33192909e-01
-2.99128264e-01 2.42039680e-01 2.01747358e-01 8.42651486e-01
3.73802304e-01 -3.60534728e-01 1.36372352e+00 6.44941330e-01
-5.50861359e-01 -3.16977322e-01 -1.29344630e+00 -7.14283705e-01
3.65917273e-02 -3.36799532e-01 4.96267140e-01 6.71216607e-01
-5.77713549e-01 2.14853495e-01 -5.24542689e-01 6.84391022e-01
5.89432180e-01 -5.76205671e-01 1.04131997e+00 -1.06017828e+00
4.16981906e-01 5.56792356e-02 -1.07461190e+00 -2.59887934e-01
-5.81949018e-02 -7.30900466e-01 1.43375695e-01 -1.70545745e+00
1.87876914e-02 -4.37339753e-01 -4.21359241e-01 3.97677094e-01
1.69592369e-02 3.56477380e-01 -4.55355868e-02 2.53648013e-01
-5.68762124e-01 6.50768161e-01 6.16145670e-01 -4.33157384e-01
7.84689933e-02 1.26645444e-02 -3.72911483e-01 1.97193071e-01
1.10561275e+00 -3.75402533e-02 -4.15806919e-01 -2.17715815e-01
-2.90983319e-01 -6.55908346e-01 7.37056375e-01 -1.67224348e+00
2.03981921e-01 1.52199835e-01 4.55804527e-01 -1.42569077e+00
7.05478013e-01 -1.06215036e+00 9.94265899e-02 3.37396890e-01
4.28941995e-01 2.27414742e-01 3.82676125e-01 4.89989311e-01
-1.49161994e-01 2.34806865e-01 1.06520605e+00 -3.02483022e-01
-1.52245140e+00 1.29029423e-01 -6.24139428e-01 -6.95840120e-01
1.16389620e+00 -1.04423141e+00 -6.37367070e-01 -2.64052987e-01
-3.23530108e-01 4.88384128e-01 4.67215061e-01 5.68879783e-01
7.86593199e-01 -1.37928891e+00 -3.24048519e-01 3.43647033e-01
3.55196476e-01 6.31872341e-02 1.71510130e-01 7.59351969e-01
-6.58835053e-01 1.00177765e+00 -5.31666160e-01 -8.88870120e-01
-1.47113931e+00 5.22568762e-01 4.39751863e-01 2.56841481e-01
-5.95914066e-01 2.26763725e-01 -9.94142741e-02 -5.81575930e-01
-9.31338742e-02 -4.14807908e-02 -2.52476156e-01 -6.16204739e-01
3.18815082e-01 3.81067336e-01 2.28840202e-01 -1.00841057e+00
-4.67616051e-01 4.24632877e-02 -3.40716466e-02 4.61379625e-02
1.27163291e+00 -6.57551646e-01 3.79870385e-01 8.81200209e-02
9.45576191e-01 -5.61836898e-01 -9.82410312e-01 -2.07905307e-01
-8.25813115e-02 -4.96949524e-01 5.14842093e-01 -5.81677735e-01
-8.27719033e-01 1.07084548e+00 1.59579170e+00 2.18497261e-01
9.12078857e-01 -2.43382808e-03 7.59117544e-01 3.92119080e-01
1.59208864e-01 -1.59823763e+00 -2.08636433e-01 6.23900652e-01
1.37256131e-01 -1.77025855e+00 -5.98845631e-02 1.83075666e-01
-5.31930029e-01 6.71078503e-01 8.09843361e-01 -1.30365029e-01
9.02294934e-01 2.79730614e-02 2.57844329e-01 -5.43591142e-01
-4.55751240e-01 -6.64488673e-01 -4.06443357e-01 9.20824349e-01
-8.58991519e-02 6.00859940e-01 4.86319214e-02 -2.08806008e-01
-5.68464756e-01 -5.97382002e-02 6.52376950e-01 1.18808317e+00
-1.35347283e+00 -6.62820578e-01 -7.87574053e-01 3.89763772e-01
-7.67411739e-02 -2.24078432e-01 -1.19589597e-01 1.18627250e+00
4.60000604e-01 1.20796990e+00 1.63816258e-01 -4.27385330e-01
1.61880970e-01 1.15542941e-01 -2.42860079e-01 -3.99264395e-01
2.15507582e-01 -1.28736973e-01 1.31246507e-01 -3.08111191e-01
-3.25918257e-01 -5.01482069e-01 -1.04844725e+00 -5.70540309e-01
-4.06620979e-01 -5.00249974e-02 1.20058358e+00 9.92700398e-01
2.86674708e-01 4.97191161e-01 7.44100392e-01 -8.59684587e-01
-2.39234582e-01 -1.03470409e+00 -9.55164850e-01 1.05503157e-01
3.55716944e-01 -1.16921890e+00 -3.54252517e-01 -3.65150124e-01] | [8.082958221435547, -1.1061218976974487] |
13928fa1-9939-4621-87da-b82478e21379 | multi-label-ranking-mining-multi-label-and | 2101.00583 | null | https://arxiv.org/abs/2101.00583v1 | https://arxiv.org/pdf/2101.00583v1.pdf | Multi-label Ranking: Mining Multi-label and Label Ranking Data | We survey multi-label ranking tasks, specifically multi-label classification and label ranking classification. We highlight the unique challenges, and re-categorize the methods, as they no longer fit into the traditional categories of transformation and adaptation. We survey developments in the last demi-decade, with a special focus on state-of-the-art methods in deep learning multi-label mining, extreme multi-label classification and label ranking. We conclude by offering a few future research directions. | ['Lihi Dery'] | 2021-01-03 | null | null | null | null | ['extreme-multi-label-classification'] | ['methodology'] | [ 6.19414151e-01 -2.11854056e-01 -5.79481363e-01 -8.63290966e-01
-1.17646468e+00 -5.56935310e-01 3.99497598e-01 5.05296350e-01
-4.66766834e-01 7.98178315e-01 5.38478419e-02 2.83338457e-01
-3.69013250e-01 -3.36752683e-01 -2.26836696e-01 -5.51639736e-01
1.87981918e-01 8.59402418e-01 -4.56931531e-01 -4.50240448e-02
1.89721808e-01 2.33933493e-01 -1.97165382e+00 9.69938934e-01
3.41038167e-01 1.08882654e+00 -4.45320755e-01 7.12516248e-01
-7.40592554e-02 1.01683831e+00 -5.27910888e-01 -5.10320902e-01
-3.21988851e-01 3.28870155e-02 -1.41338229e+00 -1.37771323e-01
1.12336898e+00 -1.87602192e-01 4.77776676e-02 9.75084484e-01
8.62585843e-01 1.91652998e-01 1.10080624e+00 -1.32638073e+00
-8.95261884e-01 5.39687514e-01 -7.60871172e-01 -1.17201224e-01
2.07353413e-01 -9.47876811e-01 1.60314751e+00 -9.73082960e-01
6.63435936e-01 1.26416552e+00 1.12797868e+00 8.97188902e-01
-9.93060350e-01 -8.31049860e-01 2.76230007e-01 3.06848109e-01
-1.19316709e+00 -2.51085967e-01 5.95458448e-01 -6.37076378e-01
1.03497827e+00 3.74086171e-01 -8.06329697e-02 1.36293137e+00
2.34943792e-01 9.34848726e-01 1.47950149e+00 -7.77487695e-01
-4.73969042e-01 6.87886924e-02 5.80078423e-01 9.56647635e-01
-1.41788870e-01 4.58139405e-02 -5.61801314e-01 -5.92148960e-01
-7.19676241e-02 7.83189293e-03 5.24499655e-01 -3.63384664e-01
-1.16896427e+00 1.09232855e+00 -1.26084298e-01 5.01429856e-01
1.36975542e-01 3.89711171e-01 1.02691948e+00 7.19154954e-01
1.05760002e+00 4.72368956e-01 -8.62658739e-01 4.47859436e-01
-1.07383692e+00 2.54238367e-01 3.08967561e-01 7.92002678e-01
7.57608533e-01 -5.95430620e-02 -3.65956247e-01 1.52941215e+00
2.99251914e-01 2.10880592e-01 6.23337626e-01 -8.03100228e-01
-8.35797116e-02 2.36645535e-01 -2.24839568e-01 -3.23286921e-01
-1.20824456e+00 -5.97883642e-01 -9.48390245e-01 8.40929747e-02
-1.12947211e-01 -5.67062497e-02 -6.21771514e-01 1.55663967e+00
1.06181115e-01 -2.79517062e-02 -1.53665379e-01 -1.04250445e-03
1.36524713e+00 8.56135786e-02 5.18334925e-01 -4.90111679e-01
1.57701480e+00 -1.42082226e+00 -7.35457778e-01 -9.17421579e-02
1.06476951e+00 -8.08551967e-01 9.26868141e-01 2.87483841e-01
-7.69629419e-01 -6.48081005e-01 -8.16325486e-01 -1.57968938e-01
-7.98584521e-01 1.55297264e-01 9.36401367e-01 5.80705822e-01
-1.04273164e+00 7.44357765e-01 7.41375163e-02 -4.36347425e-01
3.53613645e-01 6.16783857e-01 -4.43904042e-01 -4.51080538e-02
-1.58155859e+00 1.01981318e+00 3.88846129e-01 -5.37875891e-01
-9.05561507e-01 -6.79682434e-01 -6.26351595e-01 -3.96744311e-01
1.29394352e-01 -7.60947943e-01 1.72146297e+00 -7.60505855e-01
-1.22853792e+00 1.76020026e+00 4.77988534e-02 1.63631886e-01
4.75962348e-02 -9.43815932e-02 -9.61447835e-01 -3.53055239e-01
3.04688722e-01 6.17877126e-01 4.95758832e-01 -1.53799963e+00
-1.41476870e+00 -3.51292759e-01 3.60680334e-02 2.66535908e-01
-6.58432245e-01 7.59947658e-01 4.51630384e-01 -7.54729509e-01
-3.42191011e-01 -1.06550360e+00 -1.26807496e-01 -8.08004916e-01
-3.98610979e-01 -7.49705672e-01 5.02619088e-01 1.23204783e-01
1.38298452e+00 -1.69869137e+00 1.08253462e-02 -2.63734579e-01
5.62563121e-01 -1.36243239e-01 -4.02581304e-01 5.42804420e-01
-6.05165958e-01 3.65662932e-01 4.73198116e-01 -8.62836897e-01
1.63880780e-01 -1.52469650e-01 -3.15326810e-01 5.99451363e-01
-3.72203171e-01 1.26100433e+00 -1.18456066e+00 -7.33233154e-01
-2.56853388e-03 2.71801595e-02 6.10577203e-02 -1.90086216e-01
-1.02663442e-01 3.54835033e-01 -2.87771910e-01 1.13838637e+00
3.38697612e-01 -6.13284111e-01 2.89169222e-01 -3.38338554e-01
-5.97254150e-02 4.21573482e-02 -9.45179164e-01 1.59121490e+00
-7.27209687e-01 3.57591927e-01 -2.89061397e-01 -9.28894877e-01
5.84949970e-01 4.95986164e-01 1.17815912e+00 -4.73165035e-01
9.23257321e-02 5.36155879e-01 -6.22662723e-01 -5.98074161e-02
7.00004339e-01 -3.55886757e-01 -6.55648768e-01 8.25873375e-01
4.53233808e-01 3.96127254e-01 1.76125079e-01 -1.65932626e-01
7.78753817e-01 -2.08862841e-01 8.30039382e-01 -3.24361593e-01
5.00534892e-01 -3.69512111e-01 2.87128329e-01 1.00191975e+00
-1.54631063e-01 4.10983145e-01 1.99725658e-01 -9.90134418e-01
-8.21672380e-01 -6.16737187e-01 -4.63387370e-01 2.64771533e+00
-2.44073689e-01 -4.71829265e-01 -1.33604333e-01 -1.38579428e+00
1.54226825e-01 2.37372324e-01 -1.13117254e+00 5.18171862e-02
-4.05678451e-01 -1.24374056e+00 8.87131512e-01 3.54430914e-01
1.01531222e-01 -1.28769219e+00 -3.76299359e-02 2.14204118e-01
-2.35721171e-01 -7.21117139e-01 -2.75971502e-01 8.97457957e-01
-7.10668385e-01 -9.14424539e-01 -5.73680162e-01 -1.38982010e+00
3.70765537e-01 2.56628722e-01 1.79111803e+00 6.64082840e-02
-1.00690186e-01 3.52714419e-01 -6.51576877e-01 -2.64792114e-01
-6.42575264e-01 6.17195725e-01 3.32915157e-01 -1.83961615e-01
5.77237010e-01 -2.99027711e-01 -6.66243061e-02 3.40778530e-01
-7.21950471e-01 -8.76448676e-02 4.97617245e-01 1.06967235e+00
9.41882789e-01 -1.10397510e-01 1.11791790e+00 -1.88436615e+00
8.31659436e-01 -3.98843408e-01 7.09616989e-02 8.49764764e-01
-1.34138608e+00 -1.37634605e-01 4.06704307e-01 -5.83250403e-01
-6.65606260e-01 2.07099080e-01 -2.29893997e-01 -1.54780259e-03
-4.28252161e-01 4.82118845e-01 3.11441898e-01 -3.69709194e-01
9.03158307e-01 -2.17859015e-01 -5.80695331e-01 -6.72112525e-01
7.47684181e-01 1.06725645e+00 1.11934565e-01 -4.41931635e-01
2.29105845e-01 4.27954972e-01 2.07129076e-01 -8.94546881e-02
-1.69091272e+00 -7.94848263e-01 -1.06183076e+00 -4.89014745e-01
7.59690881e-01 -8.50443363e-01 -7.49600410e-01 6.94126964e-01
-9.93148625e-01 -1.63702801e-01 -5.84609248e-02 -1.64020151e-01
-6.00171745e-01 3.55909705e-01 -1.09199703e+00 -2.43725374e-01
-7.22543418e-01 -1.00724757e+00 1.44831181e+00 -1.62745360e-04
-3.78546119e-01 -1.26737821e+00 5.22489250e-01 6.50738180e-01
1.86065167e-01 3.80839258e-01 1.29244053e+00 -9.89678562e-01
4.35606092e-01 -1.82409465e-01 -1.49873555e-01 1.37351707e-01
1.86896864e-02 -3.25760365e-01 -1.42966068e+00 -6.09492064e-01
-5.74675858e-01 -1.10786963e+00 1.21993840e+00 3.25771153e-01
1.32116115e+00 1.82416007e-01 -5.98526716e-01 5.03493011e-01
1.36264002e+00 -7.32233888e-03 -1.18213981e-01 5.25599599e-01
9.21533167e-01 5.56696177e-01 9.12204385e-01 3.06972951e-01
5.92979610e-01 9.08230245e-01 4.17191505e-01 -2.74035871e-01
-2.15175033e-01 2.66916454e-01 7.06243590e-02 1.08499551e+00
9.44068946e-04 -3.19077581e-01 -6.90562069e-01 1.48379058e-01
-1.93508911e+00 -8.03996086e-01 -2.24049330e-01 1.61146915e+00
1.07452512e+00 -2.83604026e-01 1.95709229e-01 -7.89109319e-02
7.89429486e-01 4.37297165e-01 -6.68237627e-01 -5.14086723e-01
-2.16477543e-01 2.00002149e-01 6.54090345e-01 4.02344018e-01
-1.99593806e+00 1.15015316e+00 8.23131466e+00 1.13020527e+00
-6.18071020e-01 7.52543747e-01 6.49702907e-01 -6.00538589e-03
-7.23363012e-02 -5.32660902e-01 -1.17013955e+00 -2.49344055e-02
9.27844286e-01 3.15915495e-01 4.23389852e-01 1.01124990e+00
-6.64260387e-01 5.26634812e-01 -1.27451634e+00 9.96907711e-01
3.90583783e-01 -1.04875910e+00 6.59900084e-02 1.43974259e-01
1.06574857e+00 3.72887045e-01 4.25605744e-01 7.17319429e-01
5.96766174e-01 -1.09662306e+00 4.57946748e-01 3.52477878e-01
1.55197334e+00 -7.72678316e-01 7.99610794e-01 -1.96392104e-01
-1.24148452e+00 -4.87718284e-01 -3.73741448e-01 1.33296773e-01
-3.75689007e-02 7.22519100e-01 -2.19665006e-01 6.03550076e-01
5.37275851e-01 1.05246413e+00 -8.98420632e-01 5.51546037e-01
9.99726355e-02 2.09709167e-01 3.66553426e-01 1.76387131e-01
4.39854264e-02 3.12443405e-01 -9.31129456e-02 1.53635943e+00
-5.15608005e-02 -4.59229827e-01 6.54238760e-01 -1.56176716e-01
-4.85368222e-01 4.32632327e-01 -4.52095300e-01 2.70568699e-01
2.89675772e-01 1.75440776e+00 -6.82953894e-01 -4.63867038e-01
-8.34581017e-01 9.62679088e-01 6.69817984e-01 -3.67119126e-02
-9.77841139e-01 -2.11035326e-01 4.73404467e-01 -5.20230830e-01
-3.19262445e-01 3.67623061e-01 -4.18674737e-01 -9.93643105e-01
-6.53394163e-01 -1.08451807e+00 1.04328823e+00 -3.67878526e-01
-1.77879632e+00 5.74526489e-01 -8.84072185e-02 -1.07254243e+00
-2.47117758e-01 -7.48323321e-01 3.53326827e-01 2.69932032e-01
-1.70852470e+00 -1.80756402e+00 -1.40443146e-01 2.26308912e-01
5.96908212e-01 -6.06806159e-01 1.59828329e+00 1.00025249e+00
-3.91909271e-01 1.20540679e+00 8.48811090e-01 -2.45566741e-01
1.47541714e+00 -1.58284032e+00 3.59192014e-01 -2.89436013e-01
6.05439723e-01 1.74220234e-01 1.52206898e-01 -3.87960345e-01
-2.65030026e-01 -1.40003610e+00 1.24517822e+00 -8.35494995e-01
5.57484925e-01 -3.81395519e-01 -2.93578744e-01 7.94966936e-01
5.69288358e-02 -1.39135038e-02 1.52300906e+00 8.19083631e-01
-6.74016535e-01 -9.50025469e-02 -9.94399428e-01 2.03602642e-01
1.01692247e+00 -8.87635648e-01 -1.63886230e-02 1.00665081e+00
5.65552354e-01 -1.94498971e-01 -1.21094680e+00 8.31590652e-01
1.01584387e+00 -5.69312036e-01 1.26270664e+00 -1.00221288e+00
2.81125247e-01 2.28170648e-01 -1.84441000e-01 -1.36554694e+00
-1.11542630e+00 -6.90655112e-02 -3.18113208e-01 1.24333656e+00
5.16528487e-01 -4.06162351e-01 6.53166890e-01 -1.09808624e-01
-2.56849706e-01 -1.00932121e+00 -7.84709632e-01 -5.28232098e-01
4.83809114e-01 -8.74587148e-02 4.60219800e-01 1.58654475e+00
-5.54956011e-02 9.08636451e-01 -9.53140795e-01 -5.92225730e-01
5.59342742e-01 3.56988758e-01 1.57008648e-01 -1.92943037e+00
-2.65998453e-01 -7.87089050e-01 8.37403685e-02 -6.54812574e-01
9.28845823e-01 -1.36588717e+00 -2.22955883e-01 -1.46696055e+00
8.94728839e-01 -4.16259527e-01 -1.29138815e+00 8.76221657e-01
-6.87334538e-02 8.98171604e-01 -4.01636027e-03 4.34296817e-01
-1.59292543e+00 -2.29842309e-02 9.36576307e-01 -5.22803187e-01
2.99821705e-01 3.16175640e-01 -9.99443412e-01 6.76533043e-01
8.60521853e-01 -9.52138007e-01 -2.67058164e-01 -4.36480731e-01
7.84047604e-01 -2.97353119e-01 -3.35741550e-01 -7.57271409e-01
-2.29647204e-01 -2.29726404e-01 2.24678308e-01 -5.06368160e-01
6.18804283e-02 -5.70192993e-01 1.90243088e-02 1.79575086e-01
-1.06681943e+00 3.29803079e-01 -1.18695915e-01 5.13460219e-01
2.53553726e-02 -6.45785451e-01 1.04350233e+00 -4.05945748e-01
-9.79863107e-01 3.93698633e-01 -5.59670031e-01 1.21052772e-01
9.22018945e-01 2.11012632e-01 -4.87605751e-01 3.43153365e-02
-1.01740324e+00 2.64863729e-01 1.70519382e-01 8.77956033e-01
1.22263283e-01 -1.85083008e+00 -8.03284943e-01 -4.08911735e-01
7.54828870e-01 -7.34256923e-01 2.68185258e-01 4.76716548e-01
2.56070122e-02 6.62820518e-01 -3.52586657e-02 -4.03622016e-02
-1.84792054e+00 6.24602735e-01 4.90944028e-01 -1.07911563e+00
-6.98574334e-02 7.68719316e-01 1.21538922e-01 -8.90515387e-01
3.34725201e-01 1.84597105e-01 -9.12096441e-01 6.44250870e-01
5.70171177e-01 5.47942340e-01 2.58254975e-01 -7.55902767e-01
-5.43770552e-01 8.30379426e-01 -4.69636679e-01 3.31464946e-01
1.18119073e+00 -2.13510096e-01 -5.43939888e-01 9.72662926e-01
1.56473994e+00 -3.73479158e-01 -7.56543800e-02 -3.14279497e-01
4.23979491e-01 8.04860592e-02 8.79315436e-02 -1.37722993e+00
-7.34531879e-01 5.63162625e-01 9.28942919e-01 2.56305695e-01
8.96215320e-01 3.19522440e-01 6.84027374e-01 6.56831384e-01
4.07040924e-01 -1.53664231e+00 3.05555731e-01 9.65393841e-01
4.52297956e-01 -1.35297358e+00 2.66999245e-01 -1.58056822e-02
-5.37049949e-01 9.74181414e-01 6.65277660e-01 3.04097533e-01
9.50294256e-01 -6.36319891e-02 3.58257949e-01 -5.80348015e-01
-7.05038130e-01 -2.71127909e-01 4.04316097e-01 4.57489580e-01
1.07861078e+00 1.97175205e-01 -2.75180280e-01 3.58244002e-01
1.75938502e-01 -2.41649508e-01 1.25973284e-01 6.87384844e-01
-5.71618140e-01 -1.68408394e+00 3.23496796e-02 9.86278832e-01
-8.82570982e-01 -2.30319843e-01 -5.03399730e-01 3.00420940e-01
6.14013135e-01 1.10695016e+00 -2.73180306e-01 -9.27363873e-01
2.83860475e-01 5.97151041e-01 5.05401552e-01 -1.04471457e+00
-9.06513333e-01 -2.90101796e-01 3.54730248e-01 -2.79808581e-01
-7.97188938e-01 -7.21173704e-01 -7.61046827e-01 -2.24568248e-01
-7.56804645e-01 -7.38325864e-02 5.57413816e-01 8.21847916e-01
1.76445663e-01 7.59815097e-01 6.20905340e-01 -6.87550724e-01
-6.10772550e-01 -1.34178162e+00 -8.68943095e-01 5.09714007e-01
3.56003165e-01 -8.94682229e-01 -9.96122733e-02 -1.67005733e-01] | [9.599181175231934, 4.311259746551514] |
ef6304e3-0714-44bd-9eea-9d2c897cfb91 | sentence-level-adaptation-for-low-resource | null | null | https://aclanthology.org/W19-6807 | https://aclanthology.org/W19-6807.pdf | Sentence-Level Adaptation for Low-Resource Neural Machine Translation | null | ['Yash Kumar Lal', 'Aaron Mueller'] | 2019-08-01 | null | null | null | ws-2019-8 | ['low-resource-neural-machine-translation'] | ['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] | [-7.509810924530029, 3.680128335952759] |
3e8432f1-fc35-4a38-a18c-b3aeea86ec14 | fine-grained-visual-classification-with-high-1 | 2303.06442 | null | https://arxiv.org/abs/2303.06442v2 | https://arxiv.org/pdf/2303.06442v2.pdf | Fine-grained Visual Classification with High-temperature Refinement and Background Suppression | Fine-grained visual classification is a challenging task due to the high similarity between categories and distinct differences among data within one single category. To address the challenges, previous strategies have focused on localizing subtle discrepancies between categories and enhencing the discriminative features in them. However, the background also provides important information that can tell the model which features are unnecessary or even harmful for classification, and models that rely too heavily on subtle features may overlook global features and contextual information. In this paper, we propose a novel network called ``High-temperaturE Refinement and Background Suppression'' (HERBS), which consists of two modules, namely, the high-temperature refinement module and the background suppression module, for extracting discriminative features and suppressing background noise, respectively. The high-temperature refinement module allows the model to learn the appropriate feature scales by refining the features map at different scales and improving the learning of diverse features. And, the background suppression module first splits the features map into foreground and background using classification confidence scores and suppresses feature values in low-confidence areas while enhancing discriminative features. The experimental results show that the proposed HERBS effectively fuses features of varying scales, suppresses background noise, discriminative features at appropriate scales for fine-grained visual classification.The proposed method achieves state-of-the-art performance on the CUB-200-2011 and NABirds benchmarks, surpassing 93% accuracy on both datasets. Thus, HERBS presents a promising solution for improving the performance of fine-grained visual classification tasks. code: https://github.com/chou141253/FGVC-HERBS | ['Cheng-Hung Lin', 'Yu-Yung Kao', 'Po-Yung Chou'] | 2023-03-11 | fine-grained-visual-classification-with-high | https://arxiv.org/abs/2303.06442 | https://arxiv.org/pdf/2303.06442.pdf | null | ['fine-grained-image-classification'] | ['computer-vision'] | [ 0.15352783 -0.7279219 -0.14267574 -0.43234104 -0.49737298 -0.446844
0.55630195 0.2203819 -0.29622158 0.7013547 0.12639724 -0.0292681
-0.14779297 -0.7537077 -0.41238534 -1.1145238 0.18517976 -0.08505777
0.78984684 -0.02668878 0.23263443 0.6656298 -1.8918871 0.6461259
0.9376809 1.4451182 0.21348837 0.43984637 -0.3464322 0.6979295
-0.5014537 0.14152238 0.2253704 -0.19252092 -0.4220717 0.08813398
0.69488674 -0.1025572 -0.12575403 1.2976584 0.2990115 0.0540823
0.721532 -1.3671124 -0.6859291 0.08025507 -1.0219936 0.7483774
-0.1980922 0.14054152 0.64564294 -1.0526717 0.17507118 1.2294962
0.6568274 0.49070862 -1.2648969 -0.986661 0.6440473 0.40097868
-1.6708534 -0.22615814 0.686513 -0.67857146 0.55289257 0.53279436
0.53938925 0.8709793 0.22986981 0.48379162 1.5526139 -0.27148008
0.16077358 0.13754639 0.58238804 0.6480186 0.32660815 0.12047432
-0.47174704 -0.05946045 0.64134187 0.36187384 -0.34858108 -0.46550193
-1.099345 0.7517463 0.7840654 0.41886008 -0.28695428 -0.2277312
0.26789874 0.11189979 0.53672695 -0.06966012 -0.5218329 0.2739896
-0.7742551 0.04415324 0.2408706 0.63659346 1.0180031 -0.10316996
-0.68976057 0.9545368 0.2958751 0.42206505 0.45130938 -0.53490216
0.35594162 0.87401366 -0.0211352 -1.1286945 -0.34751204 -0.6161446
-1.3159438 0.40342835 0.37720865 0.1487626 -1.3175079 1.6059526
0.550661 0.3818061 -0.36652768 1.0012877 1.0432097 0.6969735
0.3276024 -0.21859114 1.4994586 -1.1364745 -0.42222652 -0.27292624
0.05066659 -0.7375013 1.109158 0.30188045 -0.54503596 -1.0770367
-1.040513 0.05191622 -0.59979767 0.23864323 0.49434158 0.46731386
-0.8974379 0.52474403 -0.6897492 -0.239484 0.69826096 0.22427729
-0.3931962 -0.15703887 -0.93698055 0.65484285 0.47518447 0.3926173
-0.84598804 -0.7089908 -0.69298446 0.14116523 0.288589 -0.5403283
0.67383075 -0.9302058 -0.9282169 0.72776103 -0.16450927 -0.08919265
0.39467007 -0.00542465 -0.47891086 -0.0975601 0.25416493 0.65196675
1.0952383 -1.217577 -1.1770077 -0.5139505 -0.22133306 0.11388084
-0.336888 0.05842627 -0.5799181 -0.83942866 0.14476155 -0.6570709
-0.20788322 0.03828415 -0.22812901 -0.32360715 1.0709985 -0.49388626
1.2764744 -2.3410418 -0.02445638 0.19742468 0.4856307 0.42096058
-0.09259082 -0.14895436 0.03851225 0.18357883 0.06287487 0.0680344
-0.08750623 -0.01025718 -0.17021559 0.45679086 0.45197305 0.7391599
-0.57811683 -0.5318017 0.58925843 0.6599824 -0.258511 0.1257535
0.13428994 0.45867565 -0.6157316 0.83006084 1.0131097 -0.35674685
-0.2183027 -0.57731587 -0.25770894 -0.24229436 -1.5181375 1.124996
-0.07439379 0.41735315 0.14577542 -0.78639174 0.9599105 -0.32161006
0.00775128 -0.8267381 0.20044886 0.0067718 -0.10843472 -0.21937566
0.15715437 -0.1043594 -0.07416624 -0.15948156 -0.10209813 0.16500391
0.19240393 0.01974135 0.83544755 0.03579598 0.41238618 -0.372029
0.80268365 -0.16483586 0.79582286 0.8231432 -0.5626086 0.4640423
0.11288989 -0.50193 -0.63682055 -1.1131296 -0.24964793 1.3430706
0.62483186 -0.37806088 -0.55234027 -0.88059217 0.2311838 0.33033326
-0.9996767 -0.2996476 -0.34768832 -0.9252764 0.23560794 0.71139574
0.7754511 -0.96717554 -0.42828545 -0.12364669 -0.2087348 -0.8738329
-0.46011686 0.55912733 -0.7131611 -0.9579267 -0.51953405 -0.78737324
0.6062138 0.7889579 1.0232723 0.27483 -0.45973495 -0.26010323
-0.33828285 -0.22419187 -0.03684794 -0.10847867 -0.17155308 0.29298595
0.4840899 -0.32216468 -0.68239725 0.56404185 -0.6222909 0.12950106
0.7564494 1.0468309 0.9147999 0.37547228 0.44010815 -0.78758806
0.22610041 -0.27249682 -0.5478394 0.25747898 -0.47179195 -0.10142338
0.77096546 -0.4180608 -1.1032441 -0.06920405 0.0668499 -0.48386508
-0.5569312 0.03073177 -0.31083864 -0.28067616 0.58595735 0.20829321
-0.55845124 -0.65227795 0.16975771 0.6133725 0.4433762 -0.49052423
0.8862223 0.49094126 -0.05693504 -0.57475185 -1.0245444 -0.87524927
-0.78893566 -0.0862451 0.78629506 -1.024199 -0.2810291 0.61912805
-0.53827405 -0.14598341 -0.22064254 0.17014596 -0.02196123 0.3211819
-0.4547173 -0.58025634 -0.3169433 -1.1153057 1.1819626 0.79308563
0.11316401 -0.67096275 -0.3187715 0.28130355 0.51375294 0.2673834
0.8834919 -0.34661865 -0.5614163 0.04008342 -0.8497243 0.38214943
0.41944885 0.0355447 -1.2056688 -0.44162402 -0.21528405 -0.3327361
1.3407618 0.45683178 1.4159278 -0.05478599 -0.5181284 0.8317143
1.5668079 0.09962584 0.43708012 0.5222154 0.8260385 0.34102145
0.8441179 0.36561784 0.12925476 0.66963744 0.44297594 -0.40281132
-0.4402354 0.01455578 0.1188792 0.44888762 -0.02231113 0.20120057
-0.63135064 0.537176 -1.8312737 -0.84915 -0.23602681 1.9857459
0.84512746 0.29260376 0.0863838 0.15844868 0.9992183 0.23463708
-0.5715159 -0.08203092 -0.31734025 0.21981587 0.38290232 0.3229401
-1.6163092 1.0060495 4.7708173 1.246943 -1.1939907 0.13295464
0.82965404 0.09197863 0.18464881 -0.2794155 -1.1900315 0.6013262
0.5109407 0.16103111 0.24519858 0.7698841 0.1806756 -0.1724299
-0.6360016 0.9515101 0.01027671 -1.3503785 0.25700372 -0.29820493
0.7079095 -0.10055188 -0.04096628 0.43514833 0.1510082 -1.0390661
0.8509639 0.44333193 0.73057425 -0.7726755 0.7644181 0.32148075
-1.6054379 -0.27069446 -0.5428359 0.01266669 -0.43079665 0.5887827
-0.45631617 0.48281768 1.2046685 0.64588046 -1.2352607 1.1747481
-0.00498908 0.47904846 -0.06900157 0.04529598 0.2747401 -0.02286862
0.1475957 1.5228889 0.0847073 -0.03926544 0.49388686 0.64728713
0.24082802 0.02925319 -0.17413063 0.39275658 0.3410029 1.6453156
-1.1152425 -0.3716515 -0.45327917 0.74196804 0.4649349 0.28917286
-1.0262593 -0.36548907 0.6973104 0.06377556 0.58084685 0.11317676
-0.41805863 -1.2630583 -0.02923059 -0.9609735 0.7122141 -0.72184485
-1.3834239 0.7022935 -0.17916152 -1.1589587 0.28809723 -0.786498
-0.38887158 0.9811353 -1.7772605 -1.353819 -0.98605514 0.9554571
0.6165779 0.03632508 0.47398385 0.33287266 -0.6502588 0.4949684
0.13982867 0.08149143 0.82402456 -1.2709316 -0.01439073 1.0372385
-0.03360297 0.4730755 0.51189935 -0.5571007 -0.90741646 -1.4752939
0.5128665 -0.16435112 0.5256986 -0.44302425 -1.1187817 0.19919848
-0.23074718 0.5135279 0.42295438 0.01780141 -0.593762 -0.5340858
-1.1132351 0.41556364 1.0029299 -0.35822752 -0.4868137 0.04951472
0.4431475 -0.1354694 -0.6126814 0.67763466 0.4956117 -1.1064519
1.0606748 -0.4867961 0.11363428 -0.821695 -0.3517941 -1.2059209
-0.9733198 0.04154909 0.02140383 1.4566203 0.07521684 -0.4358061
0.66554505 0.15319909 -0.03294563 -0.63609874 -0.83864856 -0.6081306
-0.12629315 -0.14134857 0.5359255 0.86567837 -0.7515658 0.26958615
-0.18647265 0.38513708 0.6843706 0.5715124 0.681429 -1.359536
-0.03663194 -0.5827186 -0.5069929 -0.6821626 -0.08221342 -0.77499783
0.14644226 -1.3976402 0.64418423 -0.50140584 -0.79694927 0.6949433
-0.5901511 0.8111174 0.3109453 0.20144323 -0.872483 0.44842282
1.285898 -0.3877722 -0.01709534 -0.0775409 -0.989012 0.8599437
0.6633339 -0.290808 -0.1141017 -0.06523712 -0.4067435 -0.51484096
0.6705444 -1.1134421 0.1024166 -0.47574946 1.1007339 -0.6938893
0.08814809 -0.8302296 0.11574817 0.45429575 0.01856002 -0.22591498
0.5483294 0.5796874 -0.37656197 0.07157554 1.3144125 0.07377997
-1.3462912 0.32593054 -0.15388232 0.1707449 1.0889369 -0.34595847
-0.7211275 0.1898152 -0.74551105 0.11793653 0.46920994 0.56134504
0.5789464 -1.443629 -0.7214465 0.59424603 0.3677192 -0.04186796
0.6504698 0.74734044 -0.3253164 0.32570288 -0.48545334 -0.96711147
-1.6526858 0.62958837 0.34613207 -0.3899695 -0.47057635 0.98821276
0.84442866 -0.08809122 0.2387385 -0.47663704 -0.51697844 0.13925979
0.7746602 0.3141379 0.26050627 -0.9325379 -0.697933 0.94674367
-0.24326661 0.628083 1.1619756 -0.33678642 -0.07996369 0.37759078
0.90852636 0.13615756 -1.5020065 -0.27825466 -0.14630924 -0.7344666
0.20642644 -1.0665348 -1.1029737 0.91427 1.0303017 0.23818286
1.5920702 0.03001358 0.28757948 -0.14037485 0.18645999 -0.8629326
0.06198075 0.5338727 0.82753736 -1.2907214 -0.0097189 -0.5894183
-0.47127348 0.92995846 0.96676093 -0.15553081 0.70091087 0.39811158
0.14816557 0.06225389 -0.70546794 -0.3021457 0.7030227 0.67681473
0.1714544 0.14436589 0.02772546 0.8546287 0.18841341 -0.06742673
-0.01054724 0.6815977 -0.69713414 -0.77237874 -0.56580573 0.6207913
-0.47822106 -0.17250831 -0.43813583 0.7725678 0.825713 1.0203804
0.11191926 -0.43204963 0.3663552 -0.13983528 0.32068858 -0.47382662
-0.65074503 0.30589876 -0.05891787 -0.72121733 -0.38060755 -0.5677978
-0.9728865 -0.2696418 -0.47097182 0.11235517 0.22332406 0.7974018
0.3403823 0.7466597 0.5950774 -0.9028408 -0.39375654 -0.86013126
-0.6881553 0.55342925 0.4881295 -0.9877135 -0.4830984 -0.09779433] | [9.606197357177734, 2.0029964447021484] |
c7512df7-c086-46d5-9fd9-b2cf520b6753 | window-transformer-for-dialogue-document-a | null | null | https://link.springer.com/article/10.1007/s13042-023-01792-y | https://link.springer.com/article/10.1007/s13042-023-01792-y | Window transformer for dialogue document: a joint framework for causal emotion entailment | The Causal Emotion Entailment (CEE) task aims to extract all potential pairs of emotions and corresponding causes from the unannotated emotion document in the conversational context. Most existing methods to solve CEE task follow a two-stage pipeline framework, in which the first stage is to identify emotional clauses and cause clauses and extract clause representation,separately. And in the second stage is to construct the final emotion and cause pairs. However, they ignore the effect of the distance between clauses on emotion-cause pair matching. Here, we construct a joint framework with Window Transformer to handle this problem. The pre-trained BERT and RoBERTa are used as the text encoder to generate a local representation of clauses in a given document. Meanwhile, we feed it into 2D Window Transformer to make the clause representation sensitive to the context within the Window and to obtain the dependencies between clauses. At the same time, the document ranks the candidate clauses to extract causal emotion entailments, which enhances the representation of clause pairs (emotion pairs and cause pairs) by kernel-based relative position embedding. Experimental results indicate that the framework acquires state-of-the-art results on the benchmark dataset. | ['Geng Tu & Runguo Wei', 'Hao liu', 'Dazhi Jiang'] | 2023-02-24 | null | null | null | international-journal-of-machine-learning-and-2 | ['causal-emotion-entailment'] | ['natural-language-processing'] | [ 2.80949622e-01 1.19620115e-01 -1.77724913e-01 -9.34354961e-01
-7.42423534e-01 -5.33971190e-01 5.92913210e-01 1.63694009e-01
-2.52896130e-01 3.18602949e-01 6.96992517e-01 1.23673163e-01
6.45167977e-02 -7.54387975e-01 -4.79219228e-01 -6.61682904e-01
-1.25837162e-01 1.65254593e-01 -2.36202374e-01 -2.32648775e-01
2.19419390e-01 -1.00564361e-01 -1.43831754e+00 9.64024842e-01
7.24247575e-01 1.25943565e+00 -4.16341051e-02 3.75150144e-01
-6.06509566e-01 1.25095963e+00 -5.90630651e-01 -5.59813082e-01
-3.48480046e-01 -6.79366291e-01 -7.98291385e-01 -4.13095094e-02
-2.91387856e-01 -1.70337588e-01 -4.65502366e-02 8.03872645e-01
3.05828899e-01 1.48893699e-01 5.79043031e-01 -1.23866999e+00
-4.94816512e-01 8.98562431e-01 -4.82516050e-01 8.12882259e-02
6.00228846e-01 -3.22644413e-01 1.48895752e+00 -1.17737949e+00
4.86194909e-01 1.39544547e+00 8.08137879e-02 6.34421349e-01
-6.90837443e-01 -7.87807167e-01 6.08457983e-01 4.35116202e-01
-1.02910149e+00 -2.66678542e-01 1.14306819e+00 -5.43501638e-02
1.19665515e+00 3.96962732e-01 8.02955508e-01 1.05726361e+00
1.30260706e-01 1.12057340e+00 7.59298801e-01 -3.08813661e-01
1.11769058e-01 1.80957615e-01 3.58047038e-01 5.71389377e-01
-6.63177490e-01 -1.91115141e-01 -7.29543388e-01 -6.49456233e-02
1.34931549e-01 -6.52861521e-02 -4.39846605e-01 5.66443503e-02
-1.15205789e+00 1.07919240e+00 5.15205622e-01 2.64089227e-01
-4.93949831e-01 -1.09168299e-01 6.76337123e-01 2.67336249e-01
6.96018696e-01 2.15877056e-01 -5.40479541e-01 -1.30066827e-01
-4.44084406e-01 3.47081274e-01 7.14140952e-01 8.01463902e-01
7.04978228e-01 -6.03878617e-01 -3.80372047e-01 7.91483581e-01
3.38409960e-01 -7.37400204e-02 4.65196252e-01 -2.41898417e-01
7.00964391e-01 1.08513176e+00 -2.53666639e-01 -1.13230097e+00
-1.16422750e-01 -2.48600945e-01 -6.51534855e-01 -4.81565237e-01
-2.74173737e-01 -3.06183636e-01 -4.19870079e-01 1.66119337e+00
6.20185554e-01 3.61016333e-01 4.38063145e-01 9.91655171e-01
1.05334461e+00 1.23376155e+00 -1.54309748e-02 -3.25737059e-01
1.62918890e+00 -1.08314180e+00 -1.02951241e+00 -5.66895187e-01
7.44370341e-01 -8.51262093e-01 1.13944769e+00 1.55141905e-01
-7.54994094e-01 -1.63273573e-01 -9.35331464e-01 -3.24372560e-01
-2.47317821e-01 2.65449435e-01 7.23017693e-01 -1.86040610e-01
-3.10809731e-01 1.67551920e-01 -3.59123349e-01 2.03427464e-01
7.98191577e-02 4.32519428e-02 -2.93399960e-01 -1.96083099e-01
-1.92701268e+00 7.98755109e-01 4.53892082e-01 5.05902410e-01
-4.34598267e-01 -7.64037549e-01 -1.24712908e+00 3.71108115e-01
2.77828187e-01 -2.37247750e-01 1.02411294e+00 -1.25411546e+00
-1.46718705e+00 8.52258503e-01 -5.09417713e-01 -1.01534560e-01
-6.44817054e-02 -4.03813601e-01 -5.38043857e-01 -3.13184224e-02
-1.30090984e-02 5.17536104e-01 6.39918387e-01 -8.13203335e-01
-7.45276928e-01 -2.32617766e-01 1.08734742e-01 5.93894958e-01
-2.49221876e-01 4.19150561e-01 -6.65634394e-01 -4.34468180e-01
-8.26990139e-03 -6.14616454e-01 9.02599767e-02 -3.65390688e-01
-3.18665326e-01 -8.88557792e-01 8.99350107e-01 -6.00383222e-01
1.48092699e+00 -2.44384313e+00 4.39768940e-01 1.23300247e-01
1.24736659e-01 -3.46508205e-01 -2.40949795e-01 3.23479265e-01
-5.99769711e-01 -1.50579497e-01 -1.85879037e-01 -2.40473703e-01
1.93905532e-01 2.46761128e-01 -6.82993650e-01 1.15663879e-01
8.57099712e-01 8.01225185e-01 -7.89575696e-01 -6.48794115e-01
6.89539090e-02 5.33318639e-01 -5.76991975e-01 9.01935041e-01
-3.14366579e-01 -1.36061519e-01 -5.17888844e-01 2.08214149e-01
5.44232547e-01 1.39228776e-01 3.14932138e-01 -4.88013357e-01
7.81470835e-02 7.33053386e-01 -1.14797795e+00 1.32110775e+00
-6.46798015e-01 5.40993035e-01 5.02246479e-03 -9.93112028e-01
1.13315284e+00 4.68415320e-01 2.36394614e-01 -5.53193271e-01
2.59398073e-01 -4.32890430e-02 -1.85310524e-02 -8.88765991e-01
1.83444843e-01 -4.25638914e-01 -4.97697771e-01 4.29793537e-01
2.86473241e-03 -1.19635172e-01 5.87211885e-02 2.88031816e-01
9.05632317e-01 -4.87352423e-02 2.28747055e-01 7.96086639e-02
6.99657798e-01 -2.29890421e-01 8.27242017e-01 -3.65038067e-01
-1.41681405e-02 2.11719558e-01 1.09834993e+00 -5.18236816e-01
-4.01426047e-01 -7.68698692e-01 1.89614788e-01 1.11899650e+00
1.97373748e-01 -6.93666101e-01 -5.95486879e-01 -8.39657724e-01
-3.72759014e-01 8.80750000e-01 -9.63135004e-01 -4.35949504e-01
-6.06958449e-01 -7.04838395e-01 2.52291799e-01 3.83449018e-01
5.17311096e-01 -1.21489787e+00 -4.98823732e-01 2.45216280e-01
-7.55574465e-01 -8.44327807e-01 -5.69405973e-01 3.81387889e-01
-2.02557534e-01 -1.18662477e+00 -1.35761037e-01 -9.31093514e-01
5.40921509e-01 -1.92089871e-01 1.03324986e+00 -4.71150838e-02
-1.22988530e-01 -1.58919081e-01 -5.18289864e-01 -3.52774411e-01
5.13989963e-02 -1.62884817e-01 -4.70130771e-01 3.73648137e-01
8.41690063e-01 -2.42979139e-01 -4.90880430e-01 -3.25636007e-02
-8.22750032e-01 1.79016501e-01 5.78797042e-01 9.31292772e-01
5.19612312e-01 1.82146043e-01 4.15994287e-01 -8.96369338e-01
9.85618651e-01 -7.37601280e-01 -2.17494741e-01 2.95907795e-01
-4.00254846e-01 2.00128704e-01 5.26070595e-01 -5.59407175e-01
-1.37133920e+00 -5.99243538e-03 -2.30425119e-01 -2.86513567e-01
3.28485705e-02 9.17521179e-01 -6.02160752e-01 8.47460330e-01
5.36585003e-02 1.25411183e-01 -4.10533786e-01 -8.15243050e-02
4.38178509e-01 7.12947428e-01 3.72717947e-01 -5.02153397e-01
3.80029559e-01 -8.36115889e-03 -4.50904787e-01 -2.99643189e-01
-1.14085805e+00 -4.56077069e-01 -2.81281531e-01 -3.04342508e-01
1.09303212e+00 -9.82219636e-01 -5.92136800e-01 1.12472288e-01
-1.61182034e+00 -8.10218044e-03 3.97361405e-02 5.89279711e-01
-1.12479910e-01 -1.00940101e-01 -8.90556633e-01 -8.03738713e-01
-4.78107452e-01 -9.16625440e-01 1.22760510e+00 3.44954997e-01
-4.30142879e-01 -6.82518482e-01 5.98124713e-02 2.41600633e-01
4.44856696e-02 3.63826036e-01 1.33898342e+00 -5.72495282e-01
-1.89110219e-01 -1.98376209e-01 -2.15834603e-01 1.40812531e-01
9.90801156e-02 3.24012935e-01 -9.13381755e-01 1.92994088e-01
2.22989753e-01 -5.11375666e-01 7.95296371e-01 2.02827863e-02
9.55402672e-01 -5.60457230e-01 -1.65236503e-01 2.23764881e-01
1.06716919e+00 1.94225550e-01 6.19849801e-01 -8.69047046e-02
4.56910074e-01 1.14459503e+00 1.04659355e+00 3.93624067e-01
5.20890415e-01 3.79856914e-01 3.26240093e-01 -2.80763179e-01
4.04226154e-01 -2.60078192e-01 6.05335057e-01 1.04667914e+00
3.28897417e-01 -1.38072193e-01 -5.62302649e-01 5.50059736e-01
-1.96169746e+00 -1.07059598e+00 -2.26106882e-01 1.53247249e+00
1.23767447e+00 1.90801024e-01 -4.32710469e-01 5.33370972e-02
7.66033113e-01 3.59182656e-01 -3.03533375e-01 -6.24673665e-01
6.58868754e-04 1.88443527e-01 -4.62866724e-01 5.99889040e-01
-9.89843428e-01 9.89623368e-01 4.81225777e+00 8.52544963e-01
-1.19375384e+00 -4.83632833e-02 7.08087444e-01 -3.89575690e-01
-5.58308721e-01 1.03093395e-02 -4.26726967e-01 3.20474803e-01
5.41509807e-01 -1.67145640e-01 3.62337828e-01 9.37883973e-01
1.01892166e-01 5.17102480e-02 -1.26400495e+00 9.09197867e-01
1.53888315e-01 -1.05102706e+00 -2.02963501e-01 -4.39835966e-01
3.14557225e-01 -3.17435652e-01 -2.37942964e-01 8.27368557e-01
-1.94597498e-01 -9.56967652e-01 6.64811730e-01 2.37119332e-01
5.79893470e-01 -1.30391562e+00 1.08936346e+00 2.32444495e-01
-1.55052376e+00 1.07854813e-01 -3.39934707e-01 -2.97478318e-01
2.78573725e-02 9.26023722e-01 -6.09164715e-01 4.00440872e-01
6.28390908e-01 7.34106243e-01 -1.72462940e-01 2.60760576e-01
-7.85062253e-01 5.57229459e-01 -1.11554794e-01 -4.25652713e-01
2.94688910e-01 -1.70597687e-01 1.87115520e-01 1.51230657e+00
2.81325784e-02 3.87773693e-01 -7.52068982e-02 1.04022920e+00
-2.09989801e-01 2.98197359e-01 -2.91331828e-01 3.96208167e-02
5.44232547e-01 1.64963150e+00 -3.78210306e-01 -2.37794116e-01
-3.13606828e-01 1.37957621e+00 7.78994620e-01 2.96829045e-01
-1.18739796e+00 -8.54418933e-01 8.16958368e-01 -5.19050717e-01
1.83993071e-01 3.21291894e-01 1.87319443e-01 -1.23037696e+00
3.56118232e-01 -8.41947019e-01 5.06882191e-01 -7.39677727e-01
-1.33935106e+00 7.15112746e-01 -7.57838786e-02 -7.69621015e-01
-3.53215307e-01 -4.39525694e-01 -1.25262320e+00 8.78091455e-01
-1.56737888e+00 -1.13220060e+00 -2.40567252e-01 6.42621100e-01
6.65544271e-01 4.03907508e-01 1.05709743e+00 2.34040454e-01
-8.41205299e-01 3.48040223e-01 -5.84534228e-01 3.38728696e-01
8.34543705e-01 -1.21953130e+00 -1.04900070e-01 7.62000084e-01
-8.04166943e-02 8.09772015e-01 5.35071015e-01 -5.27964830e-01
-1.17840111e+00 -1.09345400e+00 1.54137003e+00 -2.12632507e-01
5.16898572e-01 -8.44005942e-01 -9.79475141e-01 5.62389493e-01
5.47470152e-01 -2.61176582e-02 1.03337932e+00 5.03372788e-01
-5.50281405e-01 -2.11753204e-01 -7.46335328e-01 5.15068114e-01
5.91948152e-01 -7.99638212e-01 -1.03849936e+00 7.23035559e-02
1.07594562e+00 -4.40542877e-01 -5.92732787e-01 3.40576410e-01
4.66539890e-01 -7.82025278e-01 4.81995791e-01 -6.27850831e-01
1.23523974e+00 -2.34814212e-01 -7.22286329e-02 -1.36739779e+00
-1.95194528e-01 -3.01667273e-01 5.02819419e-02 1.82698274e+00
6.76569104e-01 -1.30820259e-01 2.85223365e-01 7.79625952e-01
-5.84873073e-02 -1.33160734e+00 -7.40432262e-01 1.21581713e-02
-2.61676520e-01 -4.49373037e-01 7.95682371e-01 1.22728527e+00
6.46778703e-01 1.04110050e+00 -1.27346188e-01 1.16629511e-01
7.75846466e-02 7.00669527e-01 3.75352263e-01 -7.36183047e-01
-1.49474934e-01 -3.17616105e-01 1.58106983e-01 -8.22504699e-01
7.00493872e-01 -8.06225836e-01 3.25077802e-01 -1.41717350e+00
3.87823761e-01 -1.95994884e-01 -3.41626197e-01 4.60845619e-01
-7.81006038e-01 -4.00560766e-01 -3.93181331e-02 -1.48657724e-01
-5.61093032e-01 9.60360289e-01 9.17787313e-01 -1.61619946e-01
-1.16095372e-01 -3.57825935e-01 -6.77337766e-01 6.91549599e-01
6.87677741e-01 -4.60686117e-01 -5.64919353e-01 -2.49662384e-01
4.67717141e-01 -7.15354010e-02 2.09072888e-01 -3.26331407e-01
5.97034097e-02 -3.96959960e-01 4.45374250e-01 -5.71964681e-01
1.93202689e-01 -8.66890192e-01 -2.99257636e-01 1.60863921e-01
-7.48200476e-01 1.11016765e-01 4.89377379e-02 3.61105412e-01
-6.81376874e-01 -1.13252431e-01 4.23601657e-01 1.75398871e-01
-6.75469875e-01 7.26717860e-02 -3.34622622e-01 2.31046185e-01
1.07511592e+00 4.10768658e-01 -1.53747424e-01 -3.15465301e-01
-2.70180553e-01 5.24622381e-01 -2.68860400e-01 6.90202713e-01
9.20383513e-01 -1.56150460e+00 -7.91241586e-01 1.64645255e-01
3.41438681e-01 2.60853142e-01 3.06932002e-01 5.75085700e-01
1.57599822e-01 7.30015263e-02 1.54894918e-01 -1.58924431e-01
-1.60909903e+00 5.20256937e-01 3.41789544e-01 -6.31294787e-01
-2.18599945e-01 1.34780908e+00 3.63812596e-01 -5.56302249e-01
2.58323550e-01 -5.08387744e-01 -3.77779990e-01 3.90776932e-01
6.46394849e-01 -4.34855193e-01 -1.50023833e-01 -3.41475159e-01
-6.87975228e-01 3.02642703e-01 -8.52744132e-02 -1.08883210e-01
1.37776375e+00 6.60940120e-03 -6.56914055e-01 5.32474279e-01
1.65014184e+00 1.22192197e-01 -8.72464240e-01 -8.82646292e-02
-1.84539836e-02 -8.35949257e-02 1.07252538e-01 -6.25877619e-01
-9.70467329e-01 1.03049433e+00 1.14088066e-01 -1.32397324e-01
1.34143770e+00 2.82683671e-01 8.09046268e-01 1.35074183e-01
-3.45776111e-01 -1.08199525e+00 1.76259860e-01 6.42732739e-01
1.19438481e+00 -9.45256650e-01 -3.62244517e-01 -7.02656865e-01
-8.94374669e-01 1.25563657e+00 8.76359582e-01 -7.16828704e-02
4.63635743e-01 4.75562572e-01 1.10517457e-01 -4.74988669e-01
-1.29134309e+00 -8.39797780e-02 4.05441999e-01 1.77039132e-01
9.25027907e-01 9.78795364e-02 -5.45385301e-01 1.29760540e+00
-2.59230912e-01 -3.23608905e-01 -5.09833135e-02 6.96402371e-01
-1.79386765e-01 -9.83022630e-01 -9.71225835e-03 1.78673819e-01
-2.92349160e-01 -3.45502198e-01 -9.28067267e-01 5.28490901e-01
3.60762596e-01 1.04236531e+00 2.53239632e-01 -6.42189503e-01
4.60599869e-01 3.45502794e-01 1.50951833e-01 -5.90221107e-01
-8.25368822e-01 2.31891274e-01 4.39359158e-01 -7.48670936e-01
-3.50406408e-01 -3.76323342e-01 -1.78459740e+00 -4.55038249e-02
-4.39309686e-01 5.47450721e-01 3.94004554e-01 1.12500799e+00
2.67970294e-01 7.99402714e-01 9.68886495e-01 -3.75493616e-01
-1.92803696e-01 -9.46490765e-01 -3.83486331e-01 6.62074506e-01
2.64936417e-01 -2.97552079e-01 -3.60472620e-01 1.05942192e-03] | [12.61863899230957, 6.216686248779297] |
ee91e3eb-68a4-483e-b23f-8ded137b01f4 | brain-tumor-classification-by-cascaded | 2112.14320 | null | https://arxiv.org/abs/2112.14320v1 | https://arxiv.org/pdf/2112.14320v1.pdf | Brain Tumor Classification by Cascaded Multiscale Multitask Learning Framework Based on Feature Aggregation | Brain tumor analysis in MRI images is a significant and challenging issue because misdiagnosis can lead to death. Diagnosis and evaluation of brain tumors in the early stages increase the probability of successful treatment. However, the complexity and variety of tumors, shapes, and locations make their segmentation and classification complex. In this regard, numerous researchers have proposed brain tumor segmentation and classification methods. This paper presents an approach that simultaneously segments and classifies brain tumors in MRI images using a framework that contains MRI image enhancement and tumor region detection. Eventually, a network based on a multitask learning approach is proposed. Subjective and objective results indicate that the segmentation and classification results based on evaluation metrics are better or comparable to the state-of-the-art. | ['Shadrokh Samavi', 'Pejman Khadivi', 'Nader Karimi', 'Zahra Sobhaninia'] | 2021-12-28 | null | null | null | null | ['brain-tumor-segmentation'] | ['medical'] | [ 1.91042915e-01 -1.61431313e-01 -3.03380370e-01 -3.21511775e-01
-6.11360431e-01 -6.76127672e-02 3.50447476e-01 4.06582981e-01
-7.36475408e-01 6.10059142e-01 -1.63092345e-01 -2.16107801e-01
-1.75278366e-01 -5.66908777e-01 1.69333011e-01 -9.91888225e-01
7.25622894e-03 6.39207006e-01 4.71362472e-01 2.12646246e-01
3.33749533e-01 5.81640661e-01 -1.09516919e+00 2.46909738e-01
1.08594429e+00 9.75825489e-01 6.43617570e-01 3.38891417e-01
-4.26166534e-01 5.36808968e-01 -5.60571253e-01 9.07444283e-02
2.99945492e-02 -1.18460312e-01 -7.83184707e-01 5.26756406e-01
-6.62374943e-02 -2.02394232e-01 1.09565437e-01 1.36323690e+00
3.51804525e-01 -2.95031350e-02 9.62518275e-01 -9.46269453e-01
-2.98415750e-01 4.32940632e-01 -7.62834013e-01 4.29253727e-01
-1.56355560e-01 -2.59848863e-01 1.55353054e-01 -7.98216820e-01
1.53139621e-01 4.71138895e-01 3.87401551e-01 3.98340911e-01
-7.45551169e-01 -5.97271800e-01 5.81102036e-02 3.61217886e-01
-1.40368116e+00 8.18253830e-02 5.27232468e-01 -8.01248133e-01
4.32174295e-01 1.14973702e-01 6.45914614e-01 6.69943452e-01
7.21641064e-01 7.27375031e-01 1.23231387e+00 -2.23480850e-01
2.41896123e-01 5.33963880e-03 3.69965047e-01 8.88813257e-01
3.40391666e-01 -8.62977654e-02 1.76123813e-01 1.32988051e-01
6.86184347e-01 4.67467368e-01 -1.49807751e-01 -4.19047385e-01
-1.19317162e+00 7.34930515e-01 3.98519784e-01 8.58931899e-01
-5.70092976e-01 -2.97000349e-01 4.62284327e-01 -1.82398170e-01
4.65019703e-01 -1.22228093e-01 3.79326940e-02 3.35489631e-01
-1.24924850e+00 -3.38466883e-01 2.91828126e-01 3.03221971e-01
1.75193056e-01 -5.81386276e-02 -1.27464697e-01 8.30551684e-01
2.39961326e-01 3.56171161e-01 1.09521794e+00 -3.06655616e-01
-6.95470395e-03 5.19031346e-01 -3.37004155e-01 -6.96904182e-01
-8.78353715e-01 -9.11437809e-01 -1.09112203e+00 5.19854724e-01
4.29968148e-01 3.24350037e-02 -1.23515880e+00 1.05082679e+00
2.33401805e-01 -1.56517122e-02 2.79436763e-02 8.23800147e-01
1.01368594e+00 2.74425328e-01 3.72357070e-01 -2.13302955e-01
1.44354033e+00 -1.26586187e+00 -8.67063165e-01 -1.52362540e-01
4.92270857e-01 -5.32010913e-01 6.12375796e-01 5.64109623e-01
-8.27893019e-01 -3.34450483e-01 -7.27717578e-01 4.60476667e-01
-3.51135671e-01 4.49406475e-01 7.15342224e-01 7.11793244e-01
-8.26932609e-01 5.87998144e-02 -9.97298181e-01 -2.79150158e-01
5.22846162e-01 4.55450058e-01 -2.64915347e-01 1.24861203e-01
-5.40238440e-01 1.23452008e+00 6.21995628e-01 -3.03763524e-02
-9.33928728e-01 -4.33482587e-01 -3.92859846e-01 -1.08996317e-01
4.38281715e-01 -4.04463768e-01 1.15169585e+00 -9.18347597e-01
-1.23604894e+00 8.90732348e-01 -8.82646739e-02 -4.49752808e-01
5.27877986e-01 4.34268981e-01 -4.66634840e-01 4.21277225e-01
-8.28531850e-03 5.82981706e-01 7.18871355e-01 -1.03862429e+00
-9.93368745e-01 -5.04282176e-01 -3.92668366e-01 2.88694799e-01
-1.70525879e-01 1.68349296e-01 -7.49701411e-02 -5.58915854e-01
2.78036982e-01 -6.09965444e-01 -4.73567516e-01 -1.28200382e-01
-3.24733227e-01 -8.17443505e-02 9.41563427e-01 -8.72605085e-01
8.74828815e-01 -1.92228067e+00 5.33242226e-02 4.73211318e-01
7.14198112e-01 9.94932279e-02 2.18661204e-01 -4.94213432e-01
-1.04224131e-01 7.61683211e-02 -2.15509385e-01 1.83222734e-03
-4.99471456e-01 3.46916206e-02 5.23412824e-01 6.53955042e-01
-2.48082668e-01 6.31943882e-01 -6.11620247e-01 -8.01833451e-01
4.22130376e-01 3.94673198e-01 3.53767835e-02 -8.30248743e-02
2.98461288e-01 8.01993072e-01 -5.47014177e-01 1.02410364e+00
3.84851307e-01 -3.49837661e-01 -2.20917478e-01 -2.61017233e-01
-1.75769776e-01 -5.00962198e-01 -7.82806098e-01 1.33587110e+00
-3.53860438e-01 5.37898779e-01 1.66085139e-01 -1.37896585e+00
7.11486995e-01 6.37122095e-01 9.75905955e-01 -4.90612060e-01
6.53736591e-01 4.20124948e-01 4.16461051e-01 -8.20993841e-01
-7.75903836e-03 -6.53835908e-02 3.26914877e-01 2.16432512e-01
-1.89532086e-01 -1.69471249e-01 2.42029354e-01 -2.30263069e-01
8.95871103e-01 -4.23977584e-01 4.98021394e-01 -3.62399817e-01
7.75526524e-01 2.85337921e-02 5.07085025e-01 2.77745366e-01
-5.79475164e-01 3.13013136e-01 -5.81836812e-02 -4.39624667e-01
-6.67417169e-01 -1.12059760e+00 -4.31906015e-01 6.28519773e-01
8.07728693e-02 4.42956328e-01 -9.67004776e-01 -5.44057727e-01
-3.44546556e-01 4.54746991e-01 -5.68706989e-01 1.46030068e-01
-2.46115863e-01 -1.01689053e+00 3.52291204e-02 3.81057888e-01
7.41268575e-01 -7.74838924e-01 -8.25545013e-01 2.21074343e-01
-3.04067850e-01 -1.06724584e+00 -3.00560087e-01 1.61650151e-01
-1.10089779e+00 -1.15105546e+00 -1.28021789e+00 -1.10047305e+00
1.17271602e+00 3.04550588e-01 5.08078873e-01 1.97427869e-01
-5.76721370e-01 1.45176426e-01 -3.86654437e-01 -4.93389517e-01
-4.74779874e-01 1.80604234e-01 -2.78455943e-01 4.93798405e-02
2.29617521e-01 -3.33829731e-01 -3.35622132e-01 2.82866180e-01
-8.84848058e-01 2.54852891e-01 9.96077776e-01 7.47071743e-01
7.05534458e-01 4.71201062e-01 3.67523909e-01 -6.08644783e-01
6.98330820e-01 -3.47647905e-01 -5.96957922e-01 5.27355611e-01
-6.72920704e-01 -3.21842402e-01 2.94422686e-01 -4.88632888e-01
-8.64612997e-01 2.45284855e-01 4.41379920e-02 -2.42436409e-01
-5.06468058e-01 6.45645380e-01 2.35122263e-01 -5.78865826e-01
4.98957366e-01 2.79072523e-01 2.96546429e-01 -5.91786839e-02
-2.77596653e-01 6.27170384e-01 5.57732463e-01 -1.63621455e-02
3.63450706e-01 3.97036135e-01 1.72565207e-01 -7.21483648e-01
-5.27590990e-01 -5.20708561e-01 -7.36744761e-01 -7.42808223e-01
1.16253150e+00 -5.34416676e-01 -3.63684624e-01 6.87998235e-01
-7.32790351e-01 -8.02287534e-02 1.47724107e-01 8.94227326e-01
-2.17770070e-01 3.75204742e-01 -5.96155882e-01 -5.76380432e-01
-4.43958521e-01 -1.81906104e+00 5.91189086e-01 3.97621632e-01
5.07387966e-02 -1.16334236e+00 -3.95137191e-01 5.78493595e-01
6.72776520e-01 2.18441248e-01 9.42264676e-01 -7.08299816e-01
-4.21285868e-01 -3.91969532e-01 -2.60612428e-01 2.30801348e-02
3.91837358e-01 1.35410160e-01 -6.53084338e-01 -1.87798724e-01
1.98546112e-01 -6.69348799e-03 7.59410858e-01 1.02660382e+00
1.43887007e+00 2.66531467e-01 -7.03895390e-01 3.76899451e-01
1.42876780e+00 7.17457473e-01 2.00117901e-01 3.39866281e-01
4.37159747e-01 6.62869692e-01 3.52791250e-01 1.65402830e-01
2.93027490e-01 2.85024166e-01 6.01356626e-01 -4.46325094e-01
-1.03294984e-01 5.82656980e-01 -1.52078256e-01 8.22949767e-01
-2.11136922e-01 -1.22802556e-01 -1.36061704e+00 4.57731694e-01
-1.40630186e+00 -8.65056574e-01 -1.81136876e-01 1.77528369e+00
4.09067571e-01 9.26467404e-02 5.81574552e-02 3.06355178e-01
1.07421613e+00 -3.51023823e-01 -4.43369329e-01 -6.02843687e-02
-1.07124239e-01 2.63025961e-03 7.24492729e-01 3.48027080e-01
-1.24899471e+00 5.55466354e-01 7.02498055e+00 9.57480311e-01
-1.54857302e+00 5.46632171e-01 9.21886086e-01 1.63552418e-01
7.46905953e-02 -5.69703937e-01 -5.28621852e-01 4.67153847e-01
5.08766830e-01 -9.25504714e-02 8.98827538e-02 6.39885306e-01
2.24867791e-01 -6.51035428e-01 -5.87890625e-01 9.92412210e-01
2.57364482e-01 -1.14090800e+00 -1.68627322e-01 2.90876508e-01
6.82342947e-01 1.87155217e-01 1.16333008e-01 1.24956109e-03
-6.95347264e-02 -9.65511799e-01 6.33233011e-01 6.44023240e-01
4.94434059e-01 -6.54651046e-01 9.67404127e-01 5.35200119e-01
-1.11784124e+00 -2.12451473e-01 5.64017566e-03 2.42583871e-01
2.38930941e-01 6.85143113e-01 -9.12913024e-01 3.60113859e-01
4.59667116e-01 3.44286352e-01 -5.59956849e-01 1.66420949e+00
-1.77397639e-01 3.88780057e-01 -9.82842594e-02 -1.67331263e-01
1.31961957e-01 -3.23923230e-01 3.62268776e-01 9.00363326e-01
5.25336087e-01 1.04640141e-01 6.77257121e-01 6.55718446e-01
3.76549900e-01 5.19911647e-01 -3.33300531e-01 9.89734530e-02
3.94074582e-02 1.51286888e+00 -1.59227896e+00 -5.12906671e-01
-5.70288956e-01 6.35990560e-01 -6.57860637e-02 1.54010966e-01
-8.73263121e-01 -3.42209674e-02 -1.05837956e-01 -1.79202259e-01
-7.97398537e-02 -1.99305832e-01 -7.15727270e-01 -9.31744814e-01
-2.39897326e-01 -4.16498810e-01 2.56057620e-01 -4.09249812e-01
-8.76676619e-01 9.16395426e-01 1.27022967e-01 -1.03910303e+00
8.19867384e-03 -6.33643925e-01 -8.75600159e-01 4.23073441e-01
-1.43997025e+00 -9.12183702e-01 -5.33120334e-01 5.02761543e-01
6.72799945e-01 -4.18995351e-01 8.01501751e-01 2.90565789e-01
-8.16037357e-01 3.12433064e-01 -2.44051889e-02 1.01552539e-01
2.55574286e-01 -9.96143699e-01 -6.23643816e-01 8.13814342e-01
-4.18175489e-01 -1.75627962e-01 3.41631204e-01 -6.01304233e-01
-5.09756744e-01 -9.34502840e-01 5.47777653e-01 4.95175362e-01
6.02586746e-01 2.47886240e-01 -7.94240177e-01 2.84921646e-01
3.85916591e-01 4.91163731e-02 7.37738132e-01 -3.44130963e-01
3.90852064e-01 -1.41894430e-01 -1.52091467e+00 4.91192460e-01
4.20654863e-01 -1.61167294e-01 -4.45964366e-01 5.86462438e-01
-6.58037513e-02 -3.52301866e-01 -8.53804111e-01 6.81291759e-01
4.21742499e-01 -6.63161039e-01 7.03000844e-01 8.68854970e-02
6.35997877e-02 -1.39398992e-01 2.32631713e-01 -1.40681207e+00
-4.22303587e-01 3.49448681e-01 4.58014458e-01 5.47945440e-01
6.79044008e-01 -5.49143195e-01 8.90514374e-01 6.36982143e-01
-2.21725434e-01 -8.41304660e-01 -8.25947821e-01 -6.78001165e-01
1.06479369e-01 -9.81291980e-02 2.02248156e-01 8.25140953e-01
-9.61307213e-02 -1.18234232e-01 2.33112678e-01 1.72479779e-01
8.18539858e-01 1.93073079e-02 -8.86224285e-02 -1.51454628e+00
2.54400462e-01 -1.03979301e+00 -5.77910542e-01 -2.31104061e-01
2.24599227e-01 -1.04572845e+00 -1.03266090e-01 -1.81806910e+00
3.77961874e-01 -3.49918395e-01 -4.43946630e-01 5.91735959e-01
7.13847503e-02 7.43519887e-02 -1.00307323e-01 3.44376236e-01
-2.18002558e-01 2.11264715e-01 1.46179557e+00 -4.99189943e-01
2.06227005e-02 2.55780220e-01 -4.24093381e-02 9.45644140e-01
1.21105540e+00 -4.28298265e-01 -4.17728931e-01 -4.21363741e-01
-6.02260709e-01 2.93452054e-01 1.42448708e-01 -1.35051334e+00
6.10144138e-01 -3.05796891e-01 4.79913831e-01 -7.46480286e-01
2.91507483e-01 -1.01823854e+00 5.95336184e-02 8.93726170e-01
-2.46507585e-01 2.01854631e-01 -3.03770415e-04 1.62058756e-01
-4.56145257e-01 -6.33478403e-01 1.02265239e+00 -3.42297286e-01
-8.07711780e-01 4.03349489e-01 -1.15872252e+00 -4.91467625e-01
1.62809551e+00 -5.25415242e-01 1.66350801e-03 -1.51088461e-01
-1.14487219e+00 6.86026290e-02 -5.27721755e-02 9.81358886e-02
8.33557367e-01 -1.06266522e+00 -5.88262439e-01 8.31629783e-02
3.14085670e-02 -7.35262558e-02 2.00535610e-01 1.42699230e+00
-6.34458244e-01 2.51904339e-01 -6.13270402e-01 -7.12069333e-01
-1.62980974e+00 2.97026604e-01 6.33456409e-01 -2.97680676e-01
-4.72043514e-01 6.99020684e-01 1.56414837e-01 -3.07167470e-02
5.38329244e-01 -6.75325871e-01 -6.88103139e-01 -8.14910606e-03
4.64059234e-01 3.02060485e-01 9.30808634e-02 -6.51487410e-01
-1.37829110e-01 4.29155350e-01 -1.78445205e-01 -2.04717387e-02
1.05010056e+00 -3.67266200e-02 -2.91664213e-01 9.69524905e-02
8.38079929e-01 -2.93790191e-01 -5.75781524e-01 -2.18352914e-01
8.22748393e-02 -1.50469601e-01 6.77801609e-01 -1.02871025e+00
-1.58721483e+00 8.79802465e-01 1.10692835e+00 9.13695246e-02
1.23348963e+00 -2.05859959e-01 5.70566714e-01 2.30197310e-01
4.58309680e-01 -1.10853112e+00 -3.95310968e-02 3.36574376e-01
5.68556905e-01 -1.51069295e+00 -2.53988355e-01 -6.34511888e-01
-4.88340884e-01 1.38807237e+00 6.64977014e-01 6.41203001e-02
1.05216467e+00 3.82520914e-01 2.33637080e-01 -1.44291863e-01
-1.93329975e-01 -1.82464272e-01 2.50965178e-01 5.56172609e-01
3.19219738e-01 2.92064011e-01 -6.18554533e-01 5.02095342e-01
8.55819955e-02 9.52451527e-02 2.90020108e-01 8.71977389e-01
-9.97957706e-01 -1.03347170e+00 -6.36137724e-01 7.06176460e-01
-6.84141874e-01 1.27184957e-01 -3.92650925e-02 5.69272280e-01
3.35969329e-01 8.89615655e-01 -5.96080571e-02 -1.93385020e-01
-1.66255191e-01 -1.07792430e-01 5.24039030e-01 -4.59800869e-01
-4.67712402e-01 3.52922320e-01 -2.08691657e-01 -7.32820202e-03
-6.11449599e-01 -5.64467072e-01 -1.67292905e+00 2.19128951e-01
-5.34069359e-01 2.74267197e-01 1.13953567e+00 1.22031951e+00
-1.80536643e-01 8.92961860e-01 4.80237782e-01 -6.56648695e-01
-3.04871500e-01 -8.88889968e-01 -6.80291593e-01 -5.11203930e-02
1.07860178e-01 -8.37709665e-01 -1.04968073e-02 -5.79961091e-02] | [14.698792457580566, -2.5222814083099365] |
7bd2ba27-e22e-4e9b-a35e-e68696065b92 | collective-wisdom-improving-low-resource | 2010.05445 | null | https://arxiv.org/abs/2010.05445v1 | https://arxiv.org/pdf/2010.05445v1.pdf | Collective Wisdom: Improving Low-resource Neural Machine Translation using Adaptive Knowledge Distillation | Scarcity of parallel sentence-pairs poses a significant hurdle for training high-quality Neural Machine Translation (NMT) models in bilingually low-resource scenarios. A standard approach is transfer learning, which involves taking a model trained on a high-resource language-pair and fine-tuning it on the data of the low-resource MT condition of interest. However, it is not clear generally which high-resource language-pair offers the best transfer learning for the target MT setting. Furthermore, different transferred models may have complementary semantic and/or syntactic strengths, hence using only one model may be sub-optimal. In this paper, we tackle this problem using knowledge distillation, where we propose to distill the knowledge of ensemble of teacher models to a single student model. As the quality of these teacher models varies, we propose an effective adaptive knowledge distillation approach to dynamically adjust the contribution of the teacher models during the distillation process. Experiments on transferring from a collection of six language pairs from IWSLT to five low-resource language-pairs from TED Talks demonstrate the effectiveness of our approach, achieving up to +0.9 BLEU score improvement compared to strong baselines. | ['Gholamreza Haffari', 'Wray Buntine', 'Fahimeh Saleh'] | 2020-10-12 | null | https://aclanthology.org/2020.coling-main.302 | https://aclanthology.org/2020.coling-main.302.pdf | coling-2020-8 | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [ 2.30020851e-01 5.67058921e-02 -2.14118227e-01 -3.82503510e-01
-1.40409732e+00 -7.45625198e-01 5.01670063e-01 -1.56128123e-01
-6.93423510e-01 1.12517130e+00 2.51235098e-01 -6.88670278e-01
1.78523600e-01 -4.31129456e-01 -1.04806149e+00 -6.24287307e-01
3.96220684e-01 1.00584948e+00 6.24304228e-02 -4.80558723e-01
-2.28317261e-01 9.91884992e-02 -8.22006881e-01 3.30980986e-01
1.41809464e+00 4.31123793e-01 6.50043190e-01 4.21384871e-01
-2.57925093e-01 3.74728799e-01 -6.78169906e-01 -6.69170678e-01
4.03350711e-01 -7.48997688e-01 -1.06075275e+00 -3.07290852e-01
6.15398288e-01 -2.15491056e-01 -1.32682160e-01 8.84874940e-01
6.04650080e-01 8.95256773e-02 6.55906975e-01 -7.04813957e-01
-7.22930491e-01 1.07996309e+00 -4.90871280e-01 5.08526385e-01
-2.84955138e-03 4.16017324e-02 9.64169741e-01 -9.97842312e-01
5.93130589e-01 1.22660768e+00 2.07490280e-01 6.93735540e-01
-1.37280142e+00 -9.06827033e-01 4.83106598e-02 2.29420975e-01
-1.10548544e+00 -5.52787602e-01 4.62074101e-01 -1.05076402e-01
1.04962146e+00 -1.09567948e-01 3.27392250e-01 1.07461059e+00
-1.48388281e-01 7.67739475e-01 1.35066843e+00 -6.62453115e-01
-2.15992600e-01 3.67468894e-01 -2.68215030e-01 4.94815677e-01
-3.79786342e-02 -1.99179560e-01 -7.84177423e-01 -1.55372359e-03
5.44980764e-01 -5.31049669e-01 -3.99713695e-01 -2.05800548e-01
-1.45243466e+00 6.25235140e-01 4.11327183e-01 5.86041987e-01
-2.76252985e-01 -1.29255593e-01 3.00299853e-01 8.74738395e-01
6.46034479e-01 6.40172184e-01 -8.41300786e-01 -2.49450028e-01
-9.33584273e-01 7.64124617e-02 6.23417914e-01 1.02664816e+00
9.14298177e-01 -4.73126508e-02 -3.26504976e-01 1.13159764e+00
5.37051354e-03 6.72599196e-01 6.46903813e-01 -5.16767561e-01
1.07889426e+00 2.91567594e-01 5.59693342e-03 -3.66769165e-01
2.29170978e-01 -4.24512029e-01 -6.47618473e-01 -3.60619158e-01
5.00789046e-01 -3.57558757e-01 -8.25549603e-01 1.99289334e+00
3.68227482e-01 9.00042504e-02 3.74553561e-01 7.78814375e-01
5.29491484e-01 8.74752223e-01 -6.63269265e-03 -2.91427255e-01
1.11465454e+00 -1.24484408e+00 -4.32200730e-01 -4.67347175e-01
7.79169679e-01 -1.07288635e+00 1.24311972e+00 2.41207257e-02
-1.30811155e+00 -5.49677968e-01 -9.71644878e-01 -1.30770534e-01
-4.73410860e-02 8.56453627e-02 4.28972654e-02 1.54506356e-01
-1.13721073e+00 7.05060005e-01 -6.11933947e-01 -3.51067543e-01
1.80737689e-01 4.82011199e-01 -4.28463697e-01 -4.39737737e-01
-1.33296287e+00 1.31945562e+00 5.06875455e-01 -3.35430214e-03
-9.00460839e-01 -7.87104249e-01 -4.68952328e-01 -2.91927084e-02
3.01397234e-01 -8.42581213e-01 1.48959100e+00 -1.26453424e+00
-1.80517232e+00 7.42656171e-01 -2.04345614e-01 -3.29626471e-01
6.55711889e-01 -3.12320381e-01 1.78219546e-02 -7.60754347e-02
1.15351565e-01 7.37410188e-01 6.88755810e-01 -1.04943109e+00
-6.20088518e-01 -6.30472749e-02 1.20836832e-01 6.99954689e-01
-4.65762496e-01 3.06287467e-01 -4.09710556e-01 -5.93354225e-01
-1.84467267e-02 -9.39854205e-01 -1.70931399e-01 -5.34905493e-01
-2.79551260e-02 -2.59334773e-01 3.75413716e-01 -1.01124454e+00
9.56897378e-01 -1.87639678e+00 5.90951324e-01 -8.48813951e-02
-2.20966309e-01 5.24849296e-01 -5.41933775e-01 4.52487618e-01
2.09791645e-01 4.75572795e-02 -2.06383124e-01 -2.23292366e-01
-2.30035067e-01 3.89267802e-01 -1.67801961e-01 5.13410568e-02
3.87068748e-01 1.01974618e+00 -1.16645694e+00 -5.48805058e-01
-1.73181385e-01 3.84106159e-01 -4.56079155e-01 5.57530940e-01
-2.58360803e-01 9.10905302e-01 -3.88982356e-01 2.12070182e-01
3.64278644e-01 -3.68287414e-02 4.63084817e-01 1.49719380e-02
1.54524401e-01 8.97273779e-01 -6.62899315e-01 1.89045346e+00
-9.07158494e-01 5.29960811e-01 -2.85185110e-02 -9.72196102e-01
9.34150398e-01 4.86545324e-01 6.26922399e-02 -8.02827597e-01
-6.24460354e-02 6.16130888e-01 5.38219094e-01 -3.55831891e-01
4.95593190e-01 -4.46414649e-01 9.23350453e-02 7.14045703e-01
3.42892528e-01 -2.25778267e-01 3.00562065e-02 3.20351645e-02
8.73187840e-01 3.00943971e-01 8.16444904e-02 -4.37478513e-01
3.70603085e-01 5.53828944e-03 6.50406122e-01 4.61753786e-01
-9.45313275e-02 3.63406509e-01 5.78895137e-02 -7.59477839e-02
-1.06665242e+00 -1.00788236e+00 2.20514789e-01 1.35994375e+00
-1.42651722e-01 -1.64698318e-01 -9.96981025e-01 -8.73604417e-01
-3.22326899e-01 6.82386339e-01 -1.84370816e-01 -1.52244836e-01
-1.24009025e+00 -5.59773743e-01 4.90750670e-01 3.57864618e-01
4.72116590e-01 -9.21135843e-01 -1.72646508e-01 3.44052225e-01
-6.88661456e-01 -1.24246728e+00 -7.05648124e-01 2.28218466e-01
-9.61585820e-01 -5.81153631e-01 -8.96960378e-01 -1.05208230e+00
7.34323978e-01 4.58239734e-01 1.50402653e+00 -2.12310143e-02
4.75044698e-01 -1.99380860e-01 -3.31658185e-01 -1.41602352e-01
-9.14171934e-01 6.53299332e-01 2.76331365e-01 -1.77589700e-01
3.29010546e-01 -5.42270899e-01 -3.12398791e-01 2.40637705e-01
-4.84965652e-01 4.42097992e-01 8.53811443e-01 1.06709552e+00
5.78870654e-01 -3.89465630e-01 7.88920522e-01 -6.52125061e-01
6.28010929e-01 -4.30578202e-01 -4.08535361e-01 6.73483849e-01
-4.16764498e-01 3.08995157e-01 8.42595994e-01 -6.89746320e-01
-1.14942539e+00 -2.68749952e-01 1.89105541e-01 -4.19213027e-01
1.69953719e-01 4.91293371e-01 -3.96295220e-01 1.39508441e-01
6.35149896e-01 2.65523702e-01 -3.15122426e-01 -6.43856943e-01
6.29480243e-01 7.38608003e-01 4.50864911e-01 -1.10892677e+00
8.83099377e-01 -2.65392721e-01 -5.21942377e-01 -4.90309656e-01
-7.82518804e-01 -2.63547242e-01 -9.59928989e-01 1.42486379e-01
5.77904165e-01 -1.07496023e+00 8.43323842e-02 3.35241646e-01
-1.29219854e+00 -7.40950823e-01 -1.11795641e-01 6.09289825e-01
-4.44043964e-01 1.56367663e-02 -6.77312076e-01 -1.59293145e-01
-5.21504998e-01 -1.41776454e+00 7.08088934e-01 1.38719112e-01
-9.67559218e-02 -1.01115286e+00 2.18489587e-01 6.44193172e-01
6.17212117e-01 -4.56046015e-01 1.20045221e+00 -7.50874758e-01
-4.54907924e-01 3.00647914e-01 -1.12019911e-01 5.41060746e-01
2.58703560e-01 -4.40891176e-01 -8.68877232e-01 -5.35553396e-01
-1.69739023e-01 -6.26439273e-01 5.28316319e-01 -1.80753320e-02
6.18993938e-01 -4.33647454e-01 -1.86875582e-01 3.91845942e-01
1.17816496e+00 -9.27611738e-02 2.56534189e-01 2.46829242e-01
7.34921932e-01 6.41285002e-01 6.55611992e-01 -3.66591632e-01
5.20098627e-01 9.97091115e-01 -1.95158303e-01 -4.73121135e-03
-3.15965861e-01 -4.30419177e-01 7.39078760e-01 1.64091575e+00
-1.11205928e-01 -6.20626435e-02 -1.03805172e+00 7.00116575e-01
-1.64434445e+00 -5.24469912e-01 3.03125620e-01 2.31560063e+00
1.54789686e+00 -1.72417611e-02 -2.87406724e-02 -4.57170397e-01
6.84520245e-01 -1.80213168e-01 -3.31955433e-01 -5.08830786e-01
-1.26737520e-01 3.58858526e-01 2.98860759e-01 5.99749982e-01
-5.90363562e-01 1.50007391e+00 5.71025229e+00 9.45352614e-01
-1.20334983e+00 6.08731627e-01 4.84580040e-01 -1.79061413e-01
-3.80562961e-01 7.20271170e-02 -8.10901761e-01 3.24591160e-01
1.34924316e+00 -4.43604529e-01 6.51867628e-01 3.04291487e-01
2.22307459e-01 1.40758663e-01 -1.33689141e+00 5.42118669e-01
-2.01321230e-03 -8.88339460e-01 1.96804523e-01 2.89691146e-02
1.06038928e+00 5.14336586e-01 -6.87246099e-02 5.40177584e-01
5.86970747e-01 -8.56462717e-01 5.44708908e-01 -6.00453168e-02
7.65036464e-01 -6.81465089e-01 6.27082109e-01 6.24238312e-01
-7.31077731e-01 4.08505708e-01 -3.70813429e-01 7.82037452e-02
7.81075060e-02 2.76255459e-01 -1.30569625e+00 7.29263544e-01
5.84934056e-01 3.09074849e-01 -3.20133358e-01 6.05273962e-01
-5.23785353e-01 9.40285802e-01 -2.36272469e-01 7.92864412e-02
3.31456125e-01 -2.80911535e-01 3.78500253e-01 1.34279752e+00
4.53733861e-01 -5.52743906e-03 1.44578487e-01 6.36837661e-01
-5.14906466e-01 4.21682090e-01 -5.86804867e-01 -9.77390353e-03
6.57614529e-01 1.02128327e+00 -2.37968937e-01 -5.01936615e-01
-5.05791724e-01 1.16511977e+00 8.88262153e-01 3.16490024e-01
-4.99571592e-01 2.81297937e-02 6.09356701e-01 -7.91443139e-02
2.43730202e-01 -2.95531154e-01 7.25829676e-02 -1.36477137e+00
2.38582030e-01 -1.27919352e+00 2.18704998e-01 -4.71080869e-01
-1.33242416e+00 8.41524720e-01 7.58927688e-02 -1.17796469e+00
-6.01095796e-01 -3.31812084e-01 -4.86847311e-01 1.37967753e+00
-1.68106937e+00 -1.35021329e+00 2.77092755e-01 4.31788981e-01
8.39062750e-01 -2.58979380e-01 8.35130036e-01 3.04208875e-01
-3.89165580e-01 9.08719659e-01 3.52680743e-01 6.10053390e-02
1.04666448e+00 -1.21554279e+00 6.29747272e-01 9.58925605e-01
3.83911639e-01 5.67288697e-01 6.30580604e-01 -5.10386169e-01
-1.38345754e+00 -1.06655300e+00 1.41389203e+00 -5.09696066e-01
7.13192821e-01 -3.48503798e-01 -1.21934152e+00 6.87668741e-01
6.62472367e-01 -2.72058278e-01 7.11122096e-01 2.13328496e-01
-4.16415870e-01 -6.55189343e-03 -9.68112648e-01 7.56987631e-01
9.19806123e-01 -6.48182511e-01 -8.81832242e-01 3.26822847e-01
7.76498616e-01 -4.63704020e-01 -9.32213187e-01 2.91745245e-01
2.65621275e-01 -3.83457720e-01 7.18962610e-01 -7.97356844e-01
4.76867795e-01 -1.03503756e-01 -1.48768276e-01 -2.07815552e+00
-1.31890565e-01 -6.19575262e-01 9.05967802e-02 1.39654565e+00
7.40069807e-01 -4.61322665e-01 3.70926887e-01 2.51199484e-01
-1.04299262e-01 -7.42836297e-01 -1.07105505e+00 -9.38936830e-01
8.62405419e-01 -6.49755523e-02 6.21440053e-01 1.16461384e+00
1.88492030e-01 8.92923236e-01 -3.44078064e-01 -5.29258326e-02
2.93414205e-01 9.12912488e-02 8.68829548e-01 -7.84001112e-01
-6.40593588e-01 -3.28839004e-01 2.69525677e-01 -1.13511121e+00
4.83513325e-01 -1.27694452e+00 2.83149391e-01 -1.36441612e+00
3.56070220e-01 -6.76968634e-01 -3.64180505e-01 6.62407756e-01
-5.99098027e-01 6.00572713e-02 3.42867881e-01 2.99660981e-01
-2.95870811e-01 5.42339742e-01 1.41094601e+00 -1.81034029e-01
-2.71894395e-01 -1.27633646e-01 -5.77397048e-01 3.96395952e-01
8.69274437e-01 -6.32953763e-01 -4.48550463e-01 -1.15974081e+00
4.39806096e-02 1.05350651e-01 -1.32873207e-01 -5.26999831e-01
9.33047384e-02 -3.16014200e-01 -2.29242612e-02 -9.80694368e-02
1.41358867e-01 -6.63855195e-01 -1.66205764e-01 2.61573762e-01
-3.56649786e-01 2.31751874e-01 2.75895208e-01 1.37450010e-01
-1.82818219e-01 -2.56219625e-01 7.49554992e-01 -3.07504088e-01
-3.44334573e-01 1.81253076e-01 -6.51373118e-02 3.69610190e-01
4.94420141e-01 2.57119626e-01 -2.34993011e-01 -1.94556206e-01
-3.03169966e-01 2.99004257e-01 4.61897582e-01 6.82226360e-01
2.59907067e-01 -1.34014964e+00 -1.28924477e+00 2.74668336e-02
1.13041461e-01 7.85049647e-02 -7.36040249e-02 8.20782065e-01
-1.39932379e-01 3.93609732e-01 -2.91860521e-01 -4.91071224e-01
-1.17722583e+00 2.43341759e-01 4.26396012e-01 -5.51148236e-01
-4.02985692e-01 8.23012233e-01 2.13359386e-01 -7.46678114e-01
-1.22864984e-01 -1.07380569e-01 1.02947503e-01 -2.08456874e-01
3.27504098e-01 1.63412511e-01 3.25513422e-01 -8.87019813e-01
-1.63402870e-01 4.40043390e-01 -5.33708334e-01 -5.29539645e-01
1.10780132e+00 -2.41677150e-01 -6.79734424e-02 4.81356531e-01
1.09470940e+00 -8.64488259e-02 -1.11342204e+00 -8.03568125e-01
5.98544814e-02 -2.57226825e-01 -1.37172595e-01 -9.92528498e-01
-7.94928849e-01 1.13056386e+00 2.21051082e-01 -4.78595287e-01
9.95684266e-01 8.33358467e-02 9.17025983e-01 5.95361531e-01
5.43225348e-01 -1.09561801e+00 -1.99549254e-02 8.12106490e-01
7.83140421e-01 -1.33269787e+00 -3.92550766e-01 -1.87589273e-01
-5.02033174e-01 7.70231068e-01 7.57433176e-01 2.00944737e-01
1.40885472e-01 1.86693799e-02 4.09062535e-01 2.91330814e-01
-9.65228677e-01 -6.21414371e-02 4.00225312e-01 3.94008934e-01
6.94442391e-01 3.12069714e-01 -3.46976191e-01 3.14285725e-01
-5.44422507e-01 -1.24283962e-01 2.76976913e-01 7.88798273e-01
-3.69001836e-01 -1.52764940e+00 -1.46253601e-01 2.04858422e-01
-5.34832597e-01 -4.26789075e-01 -3.93256575e-01 5.75994670e-01
-1.07160341e-02 9.34040189e-01 -2.03230411e-01 -2.84053326e-01
3.22184205e-01 3.32841098e-01 7.88251519e-01 -8.57598186e-01
-9.31005657e-01 6.77566081e-02 1.29556179e-01 -1.41334906e-01
-3.63688320e-01 -6.28118634e-01 -1.05299389e+00 -1.89215943e-01
-3.66289616e-01 4.41576242e-01 5.57500720e-01 1.28125429e+00
2.38246411e-01 2.20206454e-01 6.06060624e-01 -6.53267086e-01
-9.25628662e-01 -1.31924260e+00 3.96082923e-02 1.96545929e-01
1.61019042e-01 -3.17704201e-01 -9.64748412e-02 1.02525301e-01] | [11.639171600341797, 10.224075317382812] |
bf95a9bc-dee8-457b-8990-80de5fb6e2c1 | hr-crime-human-related-anomaly-detection-in | 2108.00246 | null | https://arxiv.org/abs/2108.00246v1 | https://arxiv.org/pdf/2108.00246v1.pdf | HR-Crime: Human-Related Anomaly Detection in Surveillance Videos | The automatic detection of anomalies captured by surveillance settings is essential for speeding the otherwise laborious approach. To date, UCF-Crime is the largest available dataset for automatic visual analysis of anomalies and consists of real-world crime scenes of various categories. In this paper, we introduce HR-Crime, a subset of the UCF-Crime dataset suitable for human-related anomaly detection tasks. We rely on state-of-the-art techniques to build the feature extraction pipeline for human-related anomaly detection. Furthermore, we present the baseline anomaly detection analysis on the HR-Crime. HR-Crime as well as the developed feature extraction pipeline and the extracted features will be publicly available for further research in the field. | ['Estefanía Talavera', 'Maya Aghaei', 'Alina Matei', 'Kayleigh Boekhoudt'] | 2021-07-31 | null | null | null | null | ['anomaly-detection-in-surveillance-videos', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'methodology'] | [ 1.76002190e-01 -2.76108354e-01 5.91838777e-01 -2.27201968e-01
-4.74291682e-01 -5.79579771e-01 8.83249819e-01 8.03351760e-01
-3.45772773e-01 3.44001576e-02 1.86300635e-01 -3.75840098e-01
-1.69117935e-02 -7.82500923e-01 -2.42676169e-01 -3.00028473e-01
-5.44888437e-01 1.12023756e-01 5.07091641e-01 -1.04796536e-01
6.91604197e-01 8.78526926e-01 -1.71366704e+00 5.75222492e-01
4.27037776e-01 8.61764491e-01 -5.92028201e-01 9.21813726e-01
-4.52440372e-03 6.54040396e-01 -6.24094129e-01 -5.28940022e-01
3.71877253e-01 -1.43514410e-01 -7.64086485e-01 1.30872250e-01
7.89676189e-01 -5.22988677e-01 -4.98737037e-01 8.30735028e-01
1.98620453e-01 2.02698827e-01 8.20136309e-01 -1.34561658e+00
-4.58635330e-01 -1.59565493e-01 -8.63248706e-01 1.25347984e+00
6.22394800e-01 3.60362500e-01 6.92655265e-01 -1.03018081e+00
7.76879430e-01 1.12642002e+00 4.28786337e-01 4.28451180e-01
-9.83436763e-01 -3.82925421e-01 -7.76918679e-02 4.52031940e-01
-1.13471758e+00 -6.37819946e-01 8.95051181e-01 -9.44526613e-01
1.39378488e+00 2.23984346e-01 6.84132338e-01 1.11859345e+00
1.77248567e-02 6.81953490e-01 7.87759721e-01 -5.89140177e-01
2.48951167e-01 -3.47224832e-01 6.03494704e-01 9.20623481e-01
4.88033623e-01 1.50617927e-01 -4.57598150e-01 -4.66627598e-01
5.91915607e-01 2.45063096e-01 4.76527363e-02 -1.24955356e-01
-6.65778339e-01 9.54853356e-01 2.02670068e-01 1.79900512e-01
-4.11910981e-01 -2.62853086e-01 9.14381266e-01 4.41382051e-01
1.00067091e+00 4.62507993e-01 -2.82434002e-02 -4.03509736e-01
-8.89531434e-01 4.88476038e-01 5.15577905e-02 5.22901773e-01
4.62705404e-01 -2.70730513e-03 -4.22856659e-01 7.21204698e-01
2.14391649e-01 2.60798752e-01 -1.46224558e-01 -7.95926213e-01
5.65789580e-01 1.03082407e+00 -1.34316325e-01 -1.36787641e+00
-5.67307591e-01 3.91858429e-01 -5.09010136e-01 4.00952846e-01
5.86029947e-01 -2.79385480e-04 -1.08322835e+00 1.02025485e+00
4.40540016e-01 4.33288574e-01 -1.46305144e-01 6.61465347e-01
8.10753822e-01 4.70349550e-01 1.62291542e-01 1.63078681e-01
1.34277594e+00 -5.10570347e-01 -6.17869854e-01 -2.75903165e-01
8.05059612e-01 -6.96072221e-01 8.04810286e-01 3.96610528e-01
-5.73382676e-01 -3.14503862e-03 -7.99603820e-01 1.47428751e-01
-8.96713555e-01 -1.33151844e-01 6.08213425e-01 7.25463927e-01
-7.98197329e-01 3.06499094e-01 -1.04022884e+00 -9.42138672e-01
1.11006916e+00 -2.85595477e-01 -9.16709542e-01 -2.13491261e-01
-6.53158545e-01 8.28358829e-01 2.55931526e-01 -2.80647844e-01
-8.42036486e-01 -5.64231277e-01 -1.32735062e+00 -3.38625669e-01
2.55462706e-01 1.10268220e-01 9.11850512e-01 -3.49058092e-01
-4.53015357e-01 1.47978640e+00 -2.67886966e-01 -5.42660832e-01
4.66484070e-01 -4.28647280e-01 -8.06336045e-01 5.72195113e-01
3.08759332e-01 6.04799353e-02 8.37598383e-01 -8.94008040e-01
-8.74059975e-01 -6.90944493e-01 -7.60239735e-02 -4.83327717e-01
-3.50024790e-01 7.61977494e-01 -3.45558584e-01 -9.85060275e-01
-2.24575877e-01 -5.99263489e-01 -1.83232859e-01 -1.09763313e-02
-4.92788374e-01 -1.34291843e-01 1.61867476e+00 -7.47112513e-01
1.53657234e+00 -2.40036702e+00 -5.19395113e-01 5.82786500e-01
2.57231414e-01 6.03653431e-01 -2.15174809e-01 5.06699443e-01
-4.06744242e-01 5.00508621e-02 -5.79304099e-01 -4.97801721e-01
-2.92680442e-01 3.74292023e-02 -4.05165702e-01 9.52075422e-01
6.67600989e-01 5.38672924e-01 -9.84771073e-01 -5.15289843e-01
6.76168084e-01 1.23641782e-01 -4.20507222e-01 2.43296340e-01
3.11100036e-01 3.70624542e-01 -3.17917228e-01 9.85819101e-01
6.63355529e-01 4.78978276e-01 -7.59661138e-01 4.86533433e-01
-1.64050490e-01 -1.68349817e-01 -8.98409605e-01 1.35445559e+00
1.58343241e-01 1.04975295e+00 -7.10139610e-03 -8.62164736e-01
7.78608561e-01 2.03120917e-01 4.28621978e-01 -7.05615103e-01
-9.20443833e-02 -1.00864068e-01 -3.12921286e-01 -8.43892813e-01
5.72164416e-01 4.95169818e-01 -1.85364977e-01 6.03419125e-01
2.48201601e-02 2.67965794e-01 7.60872781e-01 3.79872292e-01
1.66701293e+00 -1.87118709e-01 4.87347096e-01 -2.07033344e-02
6.00012600e-01 3.28286678e-01 3.81889582e-01 8.40432644e-01
-5.84630430e-01 9.32361066e-01 5.86041391e-01 -1.04844642e+00
-1.16030943e+00 -1.01640379e+00 -4.49005187e-01 8.94316792e-01
-5.33153176e-01 -6.91376150e-01 -8.74600172e-01 -9.99377549e-01
-7.11393431e-02 7.15943933e-01 -9.70335424e-01 1.32519245e-01
-5.33228278e-01 -9.53528047e-01 8.35802913e-01 4.62906629e-01
3.25100243e-01 -1.50479126e+00 -1.09346294e+00 -5.44489436e-02
1.79722041e-01 -1.17981815e+00 1.10966219e-02 -4.38365251e-01
-3.59249145e-01 -1.79045069e+00 -2.47470140e-01 -4.49397527e-02
8.06603611e-01 1.84481189e-01 1.06155884e+00 4.51662213e-01
-1.12521029e+00 6.86633945e-01 -7.15526581e-01 -8.87772262e-01
-2.18272552e-01 -4.50471431e-01 -1.92439377e-01 1.93586200e-01
9.75952566e-01 -2.91576564e-01 -4.14450884e-01 -2.41822600e-01
-1.13425350e+00 -4.68868852e-01 -5.33609539e-02 3.31327796e-01
2.23117203e-01 1.11824296e-01 4.44990657e-02 -9.44755733e-01
7.25578368e-01 -6.93623483e-01 -6.17690802e-01 -1.74876586e-01
8.52259174e-02 -6.22522295e-01 3.78339410e-01 2.60377884e-01
-1.00188589e+00 6.69470280e-02 -5.99758998e-02 -3.68262559e-01
-1.08504045e+00 2.42640451e-01 3.62004161e-01 2.33119488e-01
7.45131493e-01 -3.68005149e-02 -5.72761357e-01 -4.44178939e-01
-3.24262418e-02 6.57474875e-01 8.54538500e-01 -2.35639900e-01
7.75359690e-01 6.83982611e-01 1.55798852e-01 -1.20753169e+00
-7.44001210e-01 -9.75619733e-01 -8.77820373e-01 -2.68611729e-01
1.17900348e+00 -4.68093544e-01 -1.01673298e-01 9.97826099e-01
-1.41760302e+00 -8.26625973e-02 -7.69889206e-02 -1.44700050e-01
-1.83899581e-01 6.04187727e-01 -5.43312609e-01 -1.21989036e+00
-4.14482355e-01 -6.41125500e-01 1.29409647e+00 -1.26769757e-02
-3.78330201e-01 -9.04010296e-01 6.54682994e-01 2.48367622e-01
2.67161578e-01 1.04518855e+00 7.09809482e-01 -9.86811578e-01
9.44489986e-02 -6.31328046e-01 -3.48520100e-01 -4.89721037e-02
5.30692637e-02 5.69484353e-01 -1.20407629e+00 -2.19780952e-01
-6.16727829e-01 3.30259614e-02 1.11856866e+00 1.43271372e-01
1.35524845e+00 -1.51703224e-01 -3.49512100e-01 3.89509320e-01
1.19594574e+00 2.49548763e-01 9.24002409e-01 9.79471564e-01
8.05007875e-01 7.91424334e-01 8.27941418e-01 8.05399954e-01
2.02581823e-01 5.84874094e-01 4.92788911e-01 -2.33274177e-01
3.25792313e-01 -6.27475232e-02 4.11776230e-02 -3.48644078e-01
-3.21497828e-01 -9.61222425e-02 -1.71388113e+00 1.03596878e+00
-2.06566787e+00 -1.57038128e+00 -4.84694391e-01 1.94369507e+00
-3.61023657e-02 7.76392445e-02 4.27164793e-01 6.73936188e-01
6.15540683e-01 2.19860494e-01 1.42614365e-01 -7.82267153e-01
4.56876047e-02 3.00155282e-01 -8.21333937e-03 1.30540803e-01
-1.97581577e+00 8.89295697e-01 6.78195715e+00 4.96860176e-01
-7.10369885e-01 -1.86113846e-02 5.81044257e-01 -1.88455179e-01
4.89522904e-01 -3.16138595e-01 -3.11881125e-01 5.17663896e-01
9.38988626e-01 2.52871931e-01 6.04582243e-02 8.89895082e-01
3.96103323e-01 -3.70678753e-01 -8.96369040e-01 8.13756883e-01
1.10438325e-01 -1.13720179e+00 8.94090219e-04 -1.72949154e-02
3.38616133e-01 -7.10423663e-02 -1.60512775e-01 5.16973771e-02
1.39781848e-01 -9.95500326e-01 3.99698764e-01 3.96657735e-01
7.28466809e-01 -1.29184544e+00 9.59747732e-01 -6.80210516e-02
-1.12365043e+00 -2.88838267e-01 -2.52051860e-01 -1.89085782e-01
2.27839664e-01 7.40195572e-01 -6.22969449e-01 3.02239507e-01
1.40122187e+00 1.07658947e+00 -1.39445317e+00 1.05778635e+00
-2.60787606e-02 5.66070974e-01 -1.08495474e-01 8.49569023e-01
3.39728028e-01 -1.12736262e-01 8.86867523e-01 1.87688875e+00
3.48078102e-01 2.41705641e-01 1.46421477e-01 4.84748989e-01
5.26080430e-01 -2.38324627e-02 -1.37590051e+00 -6.68080226e-02
1.53497100e-01 1.47922850e+00 -8.97076428e-01 -1.96717992e-01
-6.72778726e-01 8.55410039e-01 5.79501569e-01 1.85420081e-01
-5.24352252e-01 -5.90955079e-01 8.27453196e-01 3.85063738e-01
1.42200083e-01 -9.20461491e-02 -8.92962143e-02 -1.19663370e+00
7.17637390e-02 -7.77794361e-01 9.39470291e-01 -4.57023680e-01
-1.59926677e+00 7.07634568e-01 4.72228736e-01 -1.24373245e+00
-3.12758178e-01 -7.85600126e-01 -1.46164262e+00 3.42950702e-01
-1.12445760e+00 -1.40640688e+00 -4.99257743e-01 9.53592539e-01
5.57664514e-01 -5.81590772e-01 1.00022352e+00 1.70702815e-01
-1.09099376e+00 3.30382079e-01 -4.42355871e-01 8.02163541e-01
4.64521855e-01 -1.24887252e+00 9.34077084e-01 1.72418785e+00
1.28541037e-01 3.49911600e-01 6.73226058e-01 -8.01996827e-01
-6.39484286e-01 -1.33092713e+00 5.62301874e-01 -9.65865552e-01
7.97181964e-01 -5.02166986e-01 -1.15869129e+00 8.42778504e-01
1.94954306e-01 4.02813286e-01 7.99854875e-01 -2.18196660e-01
-3.07817400e-01 4.93144423e-01 -1.56265414e+00 4.93826568e-01
9.56219673e-01 -5.41744769e-01 -7.96945035e-01 4.69853841e-02
-1.03354134e-01 -3.19093466e-01 -3.70999604e-01 2.92295575e-01
-2.64282003e-02 -1.29046988e+00 9.02600944e-01 -6.70521617e-01
6.98063493e-01 -3.23676944e-01 7.43570700e-02 -1.16591084e+00
-2.70111591e-01 -1.80588275e-01 -3.64054143e-01 1.47679901e+00
1.22535758e-01 -6.96565807e-01 4.45986927e-01 5.72964787e-01
-1.37151316e-01 -4.49330240e-01 -1.12573910e+00 -3.98266584e-01
-2.89299935e-01 -8.96394074e-01 4.41998452e-01 9.21521008e-01
2.05661997e-01 -2.77628452e-01 -2.14493975e-01 3.42373550e-01
7.21799433e-01 -3.50516379e-01 9.66274798e-01 -1.42608356e+00
1.92742601e-01 -2.26707324e-01 -1.16014886e+00 2.12495357e-01
1.00232080e-01 -3.78041267e-01 -2.60426491e-01 -1.21063805e+00
2.84440994e-01 -2.82342844e-02 -1.32135615e-01 7.04934359e-01
-2.65653193e-01 6.28391743e-01 6.20032428e-03 -4.71208282e-02
-6.39518082e-01 -8.51742327e-02 5.25241315e-01 -5.51485680e-02
-1.59985218e-02 -1.95646718e-01 -1.65981740e-01 1.12193525e+00
7.95149922e-01 -5.48036337e-01 -7.64641985e-02 -2.01743767e-01
-1.94126636e-01 -5.55941999e-01 6.73556805e-01 -1.02841401e+00
-8.80783647e-02 -8.32655579e-02 7.32944667e-01 -7.99910009e-01
1.25990778e-01 -8.97228360e-01 -4.47889566e-01 2.24472418e-01
-6.30313382e-02 7.32566178e-01 5.35303354e-01 5.70068300e-01
-3.13396573e-01 -1.04337908e-01 6.88244343e-01 -2.20913485e-01
-1.05515695e+00 4.14772958e-01 -8.84674311e-01 7.40556493e-02
1.53209329e+00 -2.71858424e-01 -7.64639616e-01 -3.00896972e-01
-5.50382853e-01 1.86720982e-01 7.02267647e-01 6.79958642e-01
9.75534499e-01 -1.15992606e+00 -9.08830762e-01 5.12082219e-01
6.15277290e-01 -1.51435897e-01 2.38594368e-01 5.97749114e-01
-8.98553252e-01 2.96143591e-01 -7.50733554e-01 -4.69857574e-01
-1.68722844e+00 8.80243421e-01 -3.14246789e-02 -3.42615128e-01
-9.94207442e-01 3.51377666e-01 -6.66679516e-02 -1.95623413e-01
-7.10151047e-02 3.71373773e-01 -8.24342370e-01 5.22532389e-02
1.24029076e+00 8.55184734e-01 1.35846734e-01 -1.16418898e+00
-7.94724762e-01 3.86666805e-01 -7.70908371e-02 -5.89395650e-02
1.67613971e+00 5.98796643e-03 -1.81969792e-01 3.43109965e-02
8.20846558e-01 -1.26235515e-01 -7.42147744e-01 1.13427497e-01
2.86743402e-01 -1.05027139e+00 -4.07357141e-02 -2.64728785e-01
-1.19845998e+00 9.30471897e-01 8.53646457e-01 5.15958190e-01
1.23627293e+00 -4.79234792e-02 4.44804221e-01 2.02009365e-01
2.11852446e-01 -1.05083346e+00 2.25835126e-02 4.90242541e-01
9.86737490e-01 -1.39605987e+00 2.57328842e-02 -4.52717751e-01
-5.27797699e-01 1.16175795e+00 9.25991178e-01 -3.11047286e-01
5.33762753e-01 2.30427504e-01 2.64690369e-01 -9.03882265e-01
-4.22214687e-01 -2.21659169e-01 5.87087870e-01 9.64832366e-01
4.28951859e-01 -2.09580809e-01 9.71419960e-02 2.35734552e-01
-5.57604693e-02 -5.51657975e-01 6.79340541e-01 1.35014820e+00
-2.71102816e-01 -5.76636970e-01 -7.72142708e-01 6.24850690e-01
-9.56739426e-01 1.30336091e-01 -8.08684230e-01 7.58971691e-01
1.84350505e-01 1.23581517e+00 5.88328242e-01 -2.52618082e-02
7.03431547e-01 -4.30988632e-02 9.43120196e-02 -5.98024905e-01
-5.03670514e-01 -3.12296838e-01 2.23271027e-01 -9.56708908e-01
-2.75139689e-01 -9.05035079e-01 -9.64180052e-01 -4.91341412e-01
1.62720323e-01 -3.60472351e-01 7.09442124e-02 8.93287003e-01
3.06388795e-01 1.76514506e-01 2.42902458e-01 -1.15439785e+00
5.10128140e-01 -9.25368130e-01 -4.68689084e-01 1.07657468e+00
6.93949580e-01 -7.71272480e-01 -4.24529910e-01 -7.02468976e-02] | [7.876350402832031, 1.4772660732269287] |
83ed47d6-ef02-404d-933e-08d0a9ab4703 | unsupervised-high-fidelity-facial-texture | 2110.04760 | null | https://arxiv.org/abs/2110.04760v1 | https://arxiv.org/pdf/2110.04760v1.pdf | Unsupervised High-Fidelity Facial Texture Generation and Reconstruction | Many methods have been proposed over the years to tackle the task of facial 3D geometry and texture recovery from a single image. Such methods often fail to provide high-fidelity texture without relying on 3D facial scans during training. In contrast, the complementary task of 3D facial generation has not received as much attention. As opposed to the 2D texture domain, where GANs have proven to produce highly realistic facial images, the more challenging 3D geometry domain has not yet caught up to the same levels of realism and diversity. In this paper, we propose a novel unified pipeline for both tasks, generation of both geometry and texture, and recovery of high-fidelity texture. Our texture model is learned, in an unsupervised fashion, from natural images as opposed to scanned texture maps. To the best of our knowledge, this is the first such unified framework independent of scanned textures. Our novel training pipeline incorporates a pre-trained 2D facial generator coupled with a deep feature manipulation methodology. By applying precise 3DMM fitting, we can seamlessly integrate our modeled textures into synthetically generated background images forming a realistic composition of our textured model with background, hair, teeth, and body. This enables us to apply transfer learning from the domain of 2D image generation, thus, benefiting greatly from the impressive results obtained in this domain. We provide a comprehensive study on several recent methods comparing our model in generation and reconstruction tasks. As the extensive qualitative, as well as quantitative analysis, demonstrate, we achieve state-of-the-art results for both tasks. | ['Ron Kimmel', 'Ibrahim Jubran', 'Ron Slossberg'] | 2021-10-10 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 5.82979620e-01 5.42979181e-01 3.21960390e-01 -2.93947875e-01
-9.61545706e-01 -2.67325908e-01 9.06952202e-01 -3.70532334e-01
1.07688554e-01 5.25621712e-01 -1.01181015e-01 2.91663408e-02
2.22465098e-01 -9.92921889e-01 -8.94597054e-01 -7.06766903e-01
3.16742420e-01 7.68015563e-01 4.54240069e-02 -3.88069749e-01
-1.06520779e-01 7.24195182e-01 -2.06131411e+00 2.98960567e-01
5.52662909e-01 1.24196184e+00 -1.10065214e-01 3.53961915e-01
-1.39042273e-01 2.31024548e-01 -2.63395846e-01 -7.03865886e-01
4.27633613e-01 -6.25121117e-01 -5.26626706e-01 4.88434613e-01
8.45400870e-01 -5.17139435e-01 5.51228523e-02 5.99500954e-01
5.58049500e-01 -2.80524343e-01 7.05348074e-01 -8.88099372e-01
-5.03721952e-01 -9.15111378e-02 -7.74463415e-01 -7.45127499e-01
5.19836426e-01 1.73332214e-01 3.98554564e-01 -9.46306825e-01
1.00375724e+00 1.27289331e+00 6.25962675e-01 9.30441976e-01
-1.47867274e+00 -5.06028175e-01 -2.69164830e-01 -4.83732462e-01
-1.14406657e+00 -6.77837789e-01 9.63864446e-01 -4.81294960e-01
3.51481915e-01 4.57010679e-02 9.01460767e-01 1.39559674e+00
-2.87381429e-02 6.09948218e-01 1.47856832e+00 -6.07773900e-01
1.27555147e-01 1.27537102e-01 -8.54046404e-01 9.20843840e-01
-2.08340175e-02 2.20527187e-01 -5.72611272e-01 6.07732460e-02
1.16848171e+00 -1.49472877e-01 -3.28014269e-02 -4.67393368e-01
-8.52388740e-01 4.97360319e-01 4.37818579e-02 8.21267813e-02
-3.09909672e-01 3.67774636e-01 1.80130918e-02 9.05763432e-02
9.28528786e-01 3.07761431e-02 -1.96261451e-01 -2.76824906e-02
-1.14199686e+00 5.02145171e-01 5.31553268e-01 9.19902384e-01
1.00621915e+00 3.04117918e-01 -1.68760158e-02 8.44392121e-01
2.32298851e-01 9.36577082e-01 -6.22956827e-03 -9.60395396e-01
-5.15654124e-02 4.53064114e-01 3.95395309e-02 -9.70408976e-01
3.80774327e-02 -2.13127702e-01 -7.52461612e-01 8.29874337e-01
5.28108060e-01 1.02532774e-01 -1.22920060e+00 1.73892009e+00
7.33097792e-01 5.59066907e-02 -2.45796084e-01 8.07271898e-01
7.39306927e-01 2.76843369e-01 -1.98867157e-01 3.86967994e-02
1.24011600e+00 -5.17636597e-01 -4.29456413e-01 1.18092716e-01
6.15609586e-02 -1.00033975e+00 1.03801286e+00 5.07192731e-01
-1.43660152e+00 -3.66777956e-01 -7.28094518e-01 -1.88029677e-01
-8.76370743e-02 -3.98064218e-02 6.66341960e-01 7.92943120e-01
-1.23624921e+00 6.54931009e-01 -8.57164681e-01 -5.56083441e-01
7.92693138e-01 1.81553468e-01 -8.66483271e-01 -2.38462403e-01
-7.68340647e-01 8.69798779e-01 -3.77118379e-01 1.59558639e-01
-9.49564874e-01 -8.46125305e-01 -8.94656539e-01 -3.56860101e-01
2.28386655e-01 -1.08916509e+00 1.27003026e+00 -1.02939999e+00
-2.01641679e+00 1.45373702e+00 -2.46219575e-01 1.21174548e-02
8.45660031e-01 -3.04449275e-02 1.01665922e-01 1.19749382e-01
-5.39180413e-02 8.01057756e-01 1.09648550e+00 -1.57722533e+00
-1.53563663e-01 -4.53885704e-01 -2.16139287e-01 -9.20394883e-02
2.22680978e-02 -1.81486577e-01 -5.04049540e-01 -6.23274326e-01
3.52011994e-02 -7.37816751e-01 -1.28164560e-01 6.00356519e-01
-1.50370777e-01 2.62174636e-01 7.66328156e-01 -6.16947234e-01
5.34197688e-01 -1.95461309e+00 2.55634815e-01 2.79877782e-01
1.15842402e-01 2.16283137e-03 -1.38818488e-01 3.90911996e-01
3.89557704e-02 1.53615132e-01 -4.09360021e-01 -8.98185194e-01
1.96710184e-01 1.46038249e-01 -2.99452484e-01 3.55055630e-01
6.18847907e-01 9.69254494e-01 -7.42269993e-01 -3.37929517e-01
2.39808768e-01 9.20860231e-01 -6.93252444e-01 2.03086093e-01
-4.15200710e-01 8.76050353e-01 -3.85263026e-01 8.64328206e-01
8.17462742e-01 -4.51200940e-02 1.51573256e-01 -1.63598001e-01
1.12582915e-01 -9.78742391e-02 -8.47256720e-01 2.00257516e+00
-6.38821959e-01 2.29245290e-01 2.87568539e-01 -7.49541044e-01
1.12722886e+00 4.69933212e-01 5.88260353e-01 -1.06312990e+00
1.63518116e-01 5.46434820e-01 -4.60711628e-01 -3.16881478e-01
2.34188437e-01 -7.71504939e-01 3.61421816e-02 5.37792385e-01
2.31119290e-01 -9.10898805e-01 -1.51848167e-01 -1.33739933e-01
6.86980963e-01 8.90354156e-01 -7.31601790e-02 3.28780897e-02
2.32573539e-01 -1.44269675e-01 2.34180778e-01 3.02997112e-01
4.64231849e-01 1.18279040e+00 5.45036316e-01 -4.96101558e-01
-1.47191930e+00 -1.08195710e+00 -2.23636374e-01 4.69134390e-01
-2.09016308e-01 -2.75180578e-01 -1.17252815e+00 -3.57340127e-01
6.34728447e-02 2.38822058e-01 -1.01574135e+00 1.26251951e-01
-5.01316547e-01 -5.95962226e-01 5.68023801e-01 3.79394814e-02
3.88946682e-01 -9.51011896e-01 -4.43293959e-01 5.36683165e-02
1.67014882e-01 -1.17110157e+00 -1.01450518e-01 -3.18616807e-01
-8.12744439e-01 -9.55625176e-01 -9.05332327e-01 -6.00969970e-01
7.29461253e-01 -1.38085261e-01 1.25361156e+00 2.64595807e-01
-5.85862696e-01 3.80774170e-01 -1.96989492e-01 -4.18783516e-01
-5.26563644e-01 -2.65192240e-01 -5.97816333e-02 4.17786717e-01
-3.26091886e-01 -1.02587605e+00 -6.70193136e-01 2.43761644e-01
-1.18514085e+00 5.18621743e-01 5.89209199e-01 8.01144361e-01
7.95159400e-01 -3.57647151e-01 3.86829555e-01 -9.52647030e-01
1.78613678e-01 -2.23690003e-01 -4.95826691e-01 -2.19530263e-03
-3.30143899e-01 3.47260348e-02 4.46456283e-01 -2.31537014e-01
-1.19332480e+00 1.10171370e-01 -4.80755359e-01 -4.83378500e-01
-5.10567546e-01 6.33706897e-02 -2.69382745e-01 -1.90703318e-01
5.92776656e-01 2.59634852e-01 5.24013996e-01 -6.70171261e-01
4.49412584e-01 3.60036194e-01 3.65617603e-01 -9.61382091e-01
9.44518864e-01 9.50095057e-01 3.24139863e-01 -7.62380958e-01
-6.83173418e-01 2.99229741e-01 -7.01190829e-01 -2.68059283e-01
6.99565351e-01 -7.55534589e-01 -5.09953678e-01 7.94725597e-01
-9.79744136e-01 -7.77533472e-01 -5.30786574e-01 -2.35638879e-02
-9.90755141e-01 2.39751816e-01 -4.68249232e-01 -8.08813512e-01
-2.02083796e-01 -1.17126489e+00 1.73455048e+00 -9.26126242e-02
-9.39873308e-02 -8.72241497e-01 2.37409361e-02 5.06559789e-01
6.50964797e-01 1.01897490e+00 8.01439881e-01 3.96903038e-01
-7.30801821e-01 -1.20338991e-01 -3.75380404e-02 2.05005363e-01
2.65350401e-01 1.21948391e-01 -1.35506630e+00 6.53896332e-02
-2.02613458e-01 -5.22347808e-01 6.65447950e-01 1.10112220e-01
9.08476889e-01 -1.99566949e-02 -1.04326122e-01 7.19322562e-01
1.28507292e+00 -2.14674145e-01 8.79847229e-01 -4.18225676e-02
6.67864382e-01 9.42743003e-01 3.77106547e-01 3.20221424e-01
2.12194011e-01 1.01804793e+00 4.17686254e-01 -4.31191206e-01
-7.22367525e-01 -5.20307064e-01 1.75262868e-01 3.48324120e-01
-4.40242499e-01 8.03448446e-03 -7.83805668e-01 3.57216477e-01
-1.48672378e+00 -8.97996783e-01 2.53346525e-02 2.21533442e+00
9.70199645e-01 -1.34736955e-01 -9.91781335e-03 1.18967511e-01
3.29475373e-01 -1.13331474e-01 -3.50990325e-01 -2.75228500e-01
-2.20441878e-01 9.17627096e-01 -8.24058428e-02 5.39725721e-01
-8.19349170e-01 1.10736978e+00 6.17453146e+00 9.52772915e-01
-1.44658327e+00 9.53164771e-02 9.43269551e-01 -2.11347193e-01
-7.31380522e-01 7.06439884e-03 -5.08213341e-01 3.22033495e-01
6.42356873e-01 3.78633261e-01 3.89606357e-01 5.28484106e-01
2.93476731e-01 -1.47433534e-01 -9.79689062e-01 1.09403884e+00
1.20683521e-01 -1.37833524e+00 3.07575792e-01 3.24406534e-01
9.20753360e-01 -4.30882573e-01 3.01007509e-01 -1.78154424e-01
1.67749807e-01 -1.40487814e+00 1.08631229e+00 8.53489518e-01
1.45834911e+00 -5.16568720e-01 2.48330608e-01 2.37464786e-01
-7.44808674e-01 4.86071646e-01 -3.17610577e-02 1.85448885e-01
4.34685081e-01 1.01553059e+00 -6.00515962e-01 7.58228064e-01
6.55633748e-01 5.98452389e-01 -3.08374971e-01 5.57687998e-01
-2.42337361e-01 3.25842768e-01 -4.33853984e-01 4.81427073e-01
-1.27528965e-01 -3.61403823e-01 2.13736445e-01 8.49871457e-01
5.94988942e-01 -4.97373939e-02 -1.96424082e-01 1.15347433e+00
-1.20586473e-02 1.48501486e-01 -8.31159711e-01 3.88841219e-02
-7.72801265e-02 1.39823699e+00 -6.47722661e-01 -1.43570974e-01
-2.86130875e-01 9.86348033e-01 3.30400169e-01 1.83082744e-01
-6.42201602e-01 7.54587725e-02 6.74163878e-01 7.12660789e-01
2.08503887e-01 -3.08265418e-01 -3.07557315e-01 -1.10606349e+00
1.10663719e-01 -8.30505192e-01 -3.69136244e-01 -9.38402832e-01
-1.26278400e+00 7.13049650e-01 -1.03936829e-01 -1.02300346e+00
-3.69765550e-01 -5.94297290e-01 -3.24987441e-01 1.07447815e+00
-1.57032192e+00 -1.72533667e+00 -3.91141057e-01 5.09841144e-01
8.65737125e-02 -2.02241121e-03 1.07230163e+00 3.58487904e-01
-3.11122537e-01 5.33055007e-01 -2.48335287e-01 -1.86635926e-01
7.86152482e-01 -8.98665965e-01 4.98277754e-01 4.26519513e-01
-1.38858005e-01 3.75340998e-01 5.65362692e-01 -5.21933496e-01
-1.55320728e+00 -8.93629551e-01 5.69048822e-01 -6.69787228e-01
1.94254786e-01 -6.68057382e-01 -5.96306443e-01 4.08764780e-01
-1.29667819e-02 1.17337503e-01 4.29995894e-01 -2.39143401e-01
-4.22189534e-01 -1.69863954e-01 -1.29201174e+00 6.09152138e-01
1.33902431e+00 -3.87653530e-01 -2.69201789e-02 1.84861258e-01
1.88673854e-01 -6.25029564e-01 -1.00516915e+00 5.70912063e-01
8.98148715e-01 -1.31835783e+00 7.96790481e-01 -2.33971417e-01
7.84077108e-01 -3.08736295e-01 -1.43686086e-01 -1.09578049e+00
1.59334272e-01 -8.10248733e-01 7.10595474e-02 1.21202457e+00
2.38904476e-01 -4.27405745e-01 1.03127396e+00 4.10301238e-01
-1.95440158e-01 -9.69708204e-01 -7.27503061e-01 -4.70509380e-01
1.81039751e-01 -3.58931184e-01 7.83210814e-01 7.37827957e-01
-5.72816074e-01 -8.36704224e-02 -5.13922036e-01 -2.98163205e-01
7.14852691e-01 4.38733757e-01 1.22422040e+00 -1.23914647e+00
-3.23525250e-01 -4.97406214e-01 -4.39678580e-01 -8.21052611e-01
1.76663250e-01 -8.43465328e-01 4.59828861e-02 -1.40166342e+00
1.36544272e-01 -6.80810511e-01 5.21133959e-01 5.28762281e-01
2.14540884e-01 9.29077744e-01 -1.39988795e-01 9.97185358e-04
3.78774628e-02 7.17431664e-01 1.78684247e+00 1.96444407e-01
1.95077732e-01 -3.98529589e-01 -8.48816216e-01 5.91439068e-01
5.63644648e-01 -1.88813761e-01 -3.32231164e-01 -6.12169266e-01
2.36629814e-01 7.39535242e-02 6.33454919e-01 -7.96687543e-01
-2.87168324e-01 -1.35324180e-01 5.68534374e-01 -2.34596301e-02
7.43258715e-01 -6.32678449e-01 6.39455438e-01 -1.30897835e-02
1.72112167e-01 -3.64660114e-01 3.15146744e-01 3.27140599e-01
-1.70850784e-01 3.64743978e-01 9.25418735e-01 -3.04609418e-01
-3.31244558e-01 5.82348824e-01 -1.78636730e-01 -2.27155477e-01
8.48258972e-01 -6.38794422e-01 1.34248352e-02 -5.07885993e-01
-8.59121084e-01 -4.08008844e-01 1.08473825e+00 2.94233710e-01
5.87834954e-01 -1.36724985e+00 -8.29286814e-01 6.07567132e-01
-2.69641876e-02 3.93008053e-01 4.61720705e-01 7.66552150e-01
-6.25380397e-01 -7.29005858e-02 -4.12767649e-01 -7.23827600e-01
-9.40672457e-01 1.92105293e-01 4.97283846e-01 -7.55154639e-02
-5.38726091e-01 5.06982982e-01 4.40573364e-01 -7.35926747e-01
-2.35714793e-01 -3.61688882e-02 2.26298854e-01 -1.23961180e-01
3.56336981e-01 -1.46072119e-01 4.67892498e-01 -7.20303833e-01
7.27539882e-02 1.16109037e+00 3.88816953e-01 -3.45934004e-01
1.53279734e+00 1.11213028e-01 -2.15274796e-01 1.57074437e-01
9.45628405e-01 3.18789780e-01 -1.57432234e+00 1.38028637e-01
-4.42280233e-01 -6.29187763e-01 -2.19963491e-01 -6.75917745e-01
-1.27109969e+00 1.04952657e+00 3.02650392e-01 -2.55461156e-01
1.14199388e+00 -7.65279215e-03 8.78447592e-01 -3.11415762e-01
7.33993948e-01 -6.31235540e-01 2.39762813e-01 3.27363819e-01
1.05317211e+00 -9.87524271e-01 -2.00886145e-01 -6.85460269e-01
-2.47583479e-01 1.07334185e+00 3.37708116e-01 -2.12021708e-01
6.30079150e-01 4.75122631e-01 1.70970753e-01 -2.86367238e-01
-6.38659716e-01 -1.47127777e-01 1.97898090e-01 7.90848792e-01
5.08627474e-01 -1.93037301e-01 9.31736529e-02 1.50674641e-01
-4.11718994e-01 1.23977005e-01 2.92661905e-01 8.41150641e-01
-1.01201283e-02 -1.55224907e+00 -4.01593834e-01 1.04966015e-01
-4.38452244e-01 5.91700673e-02 -3.97646934e-01 8.04109514e-01
2.32802436e-01 5.11009336e-01 1.12835906e-01 -2.39083186e-01
4.12052572e-01 1.76459283e-01 1.16143560e+00 -6.81069613e-01
-1.98920980e-01 2.47477770e-01 -4.19210419e-02 -7.20549881e-01
-4.89266485e-01 -5.99930525e-01 -7.67662585e-01 -5.23864865e-01
8.11164752e-02 -3.07943553e-01 7.94658303e-01 6.95159733e-01
5.15766144e-01 1.63576946e-01 4.27960813e-01 -1.44776869e+00
1.92286484e-02 -7.81284571e-01 -7.82681108e-01 7.42766321e-01
2.06598669e-01 -9.19208050e-01 -7.90687501e-02 2.41111055e-01] | [12.709869384765625, -0.3193938136100769] |
f51671e9-0254-48cd-a236-1b8d03135fb8 | carigan-caricature-generation-through-weakly | 1811.00445 | null | http://arxiv.org/abs/1811.00445v2 | http://arxiv.org/pdf/1811.00445v2.pdf | CariGAN: Caricature Generation through Weakly Paired Adversarial Learning | Caricature generation is an interesting yet challenging task. The primary
goal is to generate plausible caricatures with reasonable exaggerations given
face images. Conventional caricature generation approaches mainly use low-level
geometric transformations such as image warping to generate exaggerated images,
which lack richness and diversity in terms of content and style. The recent
progress in generative adversarial networks (GANs) makes it possible to learn
an image-to-image transformation from data, so that richer contents and styles
can be generated. However, directly applying the GAN-based models to this task
leads to unsatisfactory results because there is a large variance in the
caricature distribution. Moreover, some models require strictly paired training
data which largely limits their usage scenarios. In this paper, we propose
CariGAN overcome these problems. Instead of training on paired data, CariGAN
learns transformations only from weakly paired images. Specifically, to enforce
reasonable exaggeration and facial deformation, facial landmarks are adopted as
an additional condition to constrain the generated image. Furthermore, an
attention mechanism is introduced to encourage our model to focus on the key
facial parts so that more vivid details in these regions can be generated.
Finally, a Diversity Loss is proposed to encourage the model to produce diverse
results to help alleviate the `mode collapse' problem of the conventional
GAN-based models. Extensive experiments on a new large-scale `WebCaricature'
dataset show that the proposed CariGAN can generate more plausible caricatures
with larger diversity compared with the state-of-the-art models. | ['Yang Gao', 'Wenbin Li', 'Jiebo Luo', 'Jing Huo', 'Wei Xiong', 'Haofu Liao'] | 2018-11-01 | null | null | null | null | ['caricature'] | ['computer-vision'] | [ 2.11351305e-01 3.87328982e-01 2.41955873e-02 -3.56877387e-01
-5.13719559e-01 -3.85596246e-01 6.25836015e-01 -6.76658511e-01
7.46101961e-02 8.04852188e-01 1.07574709e-01 2.45411605e-01
3.12464178e-01 -9.62875247e-01 -8.50435913e-01 -7.27541447e-01
4.88841981e-01 2.00670704e-01 -1.53033450e-01 -5.13161838e-01
-2.13301759e-02 6.38788521e-01 -1.45837724e+00 1.14819288e-01
1.21858692e+00 8.03076506e-01 1.90496236e-01 2.21783102e-01
-2.09940493e-01 4.53624398e-01 -8.02828670e-01 -8.12000096e-01
2.93644577e-01 -7.67961204e-01 -2.78170049e-01 2.84525752e-01
5.96017182e-01 -5.21317780e-01 -2.68939793e-01 1.11107063e+00
5.04417002e-01 -1.27795085e-01 6.80932462e-01 -1.37093019e+00
-1.07359672e+00 2.77731031e-01 -9.54163313e-01 -5.43124318e-01
2.49537602e-01 3.09111625e-01 6.85538173e-01 -9.09011364e-01
6.90502644e-01 1.60312366e+00 4.62626398e-01 1.07302487e+00
-1.45609725e+00 -9.90167022e-01 8.37979317e-02 -1.70931175e-01
-1.39438033e+00 -4.56414461e-01 1.25524354e+00 1.04351901e-02
1.16861358e-01 4.84580159e-01 8.00341606e-01 1.43971634e+00
9.40971971e-02 7.49531686e-01 1.07264233e+00 -2.93014079e-01
1.70827314e-01 1.77079618e-01 -8.15736830e-01 6.36536717e-01
1.48188338e-01 1.14210941e-01 -1.37849480e-01 3.80201451e-02
1.39806306e+00 1.33078322e-01 -2.96741903e-01 -2.00576633e-01
-9.76922512e-01 9.38605547e-01 7.68849790e-01 1.99063584e-01
-4.38781530e-01 1.87929705e-01 5.73158124e-03 1.69408061e-02
4.52258140e-01 7.09986150e-01 4.17162143e-02 2.36724168e-01
-8.15708101e-01 6.74512386e-01 2.87216574e-01 1.05243254e+00
8.42179358e-01 5.35426795e-01 -2.92526037e-01 1.07216620e+00
1.71128333e-01 5.84234774e-01 3.18641990e-01 -9.87614810e-01
5.09602189e-01 6.10324383e-01 -3.66689935e-02 -1.25430691e+00
1.92182511e-01 -2.94532359e-01 -1.20571089e+00 6.07705176e-01
2.88163722e-01 -2.27204055e-01 -1.16662765e+00 1.90780616e+00
2.68489450e-01 8.58528763e-02 -5.49621694e-02 1.01031578e+00
7.70147145e-01 7.59332597e-01 3.54888961e-02 -3.48746702e-02
1.19904554e+00 -8.93027604e-01 -8.95699561e-01 -2.78549552e-01
1.43459767e-01 -9.31845486e-01 1.42062783e+00 1.80697575e-01
-1.32298160e+00 -6.62506044e-01 -8.47999811e-01 -2.80646533e-02
-8.24346244e-02 6.68365434e-02 5.88046849e-01 5.03347278e-01
-9.09092665e-01 3.97785783e-01 -4.62133974e-01 1.50390804e-01
7.56416738e-01 9.93185490e-02 -3.87675196e-01 -1.50668502e-01
-1.13960409e+00 7.61362135e-01 1.35033444e-01 2.57927150e-01
-7.67894924e-01 -7.24358141e-01 -9.60945785e-01 -5.05674742e-02
1.29297227e-01 -6.75978363e-01 9.22681630e-01 -1.42411876e+00
-1.76644647e+00 6.82598889e-01 6.23747930e-02 4.30728234e-02
9.16959882e-01 -3.51634622e-02 -3.16613048e-01 -2.94435229e-02
-1.07527256e-01 1.28160512e+00 1.10686767e+00 -1.68520188e+00
-2.39186093e-01 -9.42244828e-02 1.23338856e-01 1.30741075e-01
-3.46698970e-01 -2.57314861e-01 -6.58601463e-01 -1.29025614e+00
-1.03603095e-01 -1.04100513e+00 -3.06879938e-01 2.60793328e-01
-5.61585724e-01 5.54703251e-02 9.87706900e-01 -6.11903250e-01
9.42476928e-01 -2.01605701e+00 3.78412068e-01 9.30171907e-02
1.73412561e-01 2.77336955e-01 -4.80599374e-01 2.00642794e-01
-1.50512844e-01 4.58077222e-01 -3.40350449e-01 -5.60519040e-01
-1.99467525e-01 3.70492101e-01 -3.57559919e-01 7.29203969e-02
5.49525976e-01 1.08460748e+00 -7.04296768e-01 -5.52880645e-01
1.31995454e-01 7.97409534e-01 -8.37280869e-01 4.38168526e-01
-4.29617643e-01 8.15270483e-01 -4.76024300e-01 6.62650287e-01
9.80958283e-01 2.30694339e-01 -1.62656084e-01 -1.64222896e-01
2.12896824e-01 -3.38623792e-01 -7.61586607e-01 1.70525098e+00
-6.29365742e-01 3.95887643e-01 -1.06978387e-01 -5.76185524e-01
1.14967108e+00 2.53870070e-01 1.57964841e-01 -5.92286944e-01
2.35000983e-01 9.24315900e-02 -8.64120945e-03 -3.43365788e-01
1.71726972e-01 -4.17723030e-01 -7.44157657e-02 1.61721259e-01
-2.84626186e-01 -6.88347697e-01 -8.35580677e-02 -8.24748129e-02
4.03353542e-01 3.41127157e-01 -1.76764041e-01 -5.44540547e-02
5.47691047e-01 -2.50570089e-01 7.73544192e-01 2.36479007e-02
2.61839539e-01 1.16391075e+00 5.09435296e-01 -5.02481878e-01
-1.33624184e+00 -1.03327143e+00 -1.58420894e-02 4.08607543e-01
2.05019712e-01 -1.59223914e-01 -1.19011211e+00 -7.18187392e-01
-2.23589838e-01 7.05559015e-01 -8.39846194e-01 -3.60196561e-01
-7.43177891e-01 -5.13627291e-01 4.27851140e-01 5.34599662e-01
7.76845396e-01 -1.36087787e+00 -2.08766267e-01 3.44818756e-02
-2.63937324e-01 -9.05721188e-01 -7.92784452e-01 -6.93126798e-01
-7.28617728e-01 -6.93316519e-01 -1.12981939e+00 -7.70561278e-01
1.19198668e+00 1.68623310e-02 8.95559609e-01 1.29597172e-01
-3.21473688e-01 -3.88864130e-01 -2.35245988e-01 -5.08270979e-01
-6.36006892e-01 -1.52738601e-01 -7.48026967e-02 1.77456290e-01
-1.62546575e-01 -6.96922183e-01 -7.92685151e-01 3.95159245e-01
-1.24633729e+00 4.54864591e-01 7.16077030e-01 9.83492255e-01
5.17878354e-01 1.18541278e-01 6.73845053e-01 -8.69528174e-01
6.51013136e-01 -2.14881897e-01 -4.38611031e-01 1.17948264e-01
-3.52720588e-01 1.04072683e-01 1.05158293e+00 -8.44555378e-01
-1.36820459e+00 -1.60742387e-01 -2.98868477e-01 -8.24359894e-01
-1.57714978e-01 3.28985043e-02 -6.16948187e-01 -6.04974441e-02
5.98840117e-01 1.72402665e-01 3.35749865e-01 -3.75765651e-01
4.89749342e-01 3.01558256e-01 5.28928101e-01 -6.67088985e-01
1.15818107e+00 4.08377022e-01 -7.04222172e-02 -5.82593679e-01
-4.45393682e-01 6.20000064e-01 -3.85625988e-01 -1.41157955e-01
8.02104652e-01 -7.64449179e-01 -4.74897861e-01 5.01605988e-01
-1.17757463e+00 -2.85956800e-01 -3.11901271e-01 1.56382442e-01
-5.66657007e-01 1.66526765e-01 -4.92575824e-01 -6.62229478e-01
-3.14279854e-01 -1.26740861e+00 1.09230006e+00 5.03381371e-01
-3.07953954e-01 -6.96531355e-01 -1.65908426e-01 3.32803726e-01
4.55254644e-01 7.31746733e-01 9.17951226e-01 1.87755436e-01
-5.72705448e-01 -1.68136433e-01 -1.21448144e-01 4.34695721e-01
5.10310411e-01 3.12022060e-01 -7.68001020e-01 -2.46441185e-01
-2.72325337e-01 -3.47193956e-01 4.88124043e-01 1.07943676e-01
1.49439144e+00 -6.49281859e-01 -1.27729967e-01 7.65306652e-01
1.20816338e+00 2.78584450e-01 1.11047792e+00 -4.70035002e-02
8.93364012e-01 7.92991400e-01 5.68799496e-01 2.48196214e-01
6.30516410e-02 7.36878037e-01 5.28285086e-01 -5.13780594e-01
-4.06933397e-01 -8.09988916e-01 2.01642320e-01 3.45894217e-01
-3.54030699e-01 -2.05500498e-01 -4.09887493e-01 3.78070712e-01
-1.53187239e+00 -9.41327691e-01 2.41597995e-01 2.01856565e+00
1.03462303e+00 -3.86655442e-02 7.58287013e-02 -3.60322073e-02
8.23076904e-01 1.56805068e-01 -5.12474656e-01 -3.51102680e-01
-3.92351449e-02 2.07776174e-01 1.81754734e-02 3.02534789e-01
-6.32675469e-01 1.12098551e+00 5.71893501e+00 1.06940567e+00
-1.39122772e+00 -1.64821327e-01 1.04124379e+00 -1.82590649e-01
-8.48972261e-01 -2.19684422e-01 -5.47944784e-01 8.11437309e-01
2.09593713e-01 1.21115344e-02 4.80111092e-01 9.40415502e-01
3.11684728e-01 2.77340978e-01 -7.37814188e-01 1.16133451e+00
7.43190646e-02 -1.35702407e+00 4.65253085e-01 1.43079057e-01
1.03909421e+00 -8.48401964e-01 4.48887587e-01 1.76680073e-01
3.05862218e-01 -1.30299020e+00 7.72673845e-01 5.64277768e-01
1.30462980e+00 -1.19239771e+00 4.88402337e-01 2.20742851e-01
-8.95129085e-01 2.01538399e-01 -5.09706795e-01 2.79341787e-01
2.13084146e-01 4.06848848e-01 -5.62714279e-01 3.29976618e-01
5.36299646e-01 2.53941894e-01 -4.66990799e-01 5.17084539e-01
-4.97372836e-01 1.69798076e-01 -1.45486519e-01 1.52158707e-01
1.71699628e-01 -5.13105631e-01 4.11446452e-01 7.10278213e-01
5.61710656e-01 9.83006358e-02 -1.16375834e-01 1.31564367e+00
-3.27578157e-01 1.11118920e-01 -8.60164404e-01 1.41524926e-01
4.64282036e-01 1.29527235e+00 -2.96189845e-01 -2.18610898e-01
-1.14502668e-01 1.18395197e+00 2.99705952e-01 4.06228006e-01
-1.09689522e+00 -3.07737917e-01 7.36824572e-01 3.56078386e-01
9.21807662e-02 4.67924140e-02 -3.07588726e-01 -1.15856588e+00
3.22518195e-03 -1.15529156e+00 -1.19330399e-01 -8.44704688e-01
-1.24451506e+00 1.04908645e+00 -4.93319742e-02 -1.31218910e+00
-4.07550752e-01 -9.29389969e-02 -9.76709962e-01 1.04625702e+00
-1.33288288e+00 -1.53834677e+00 -5.52435279e-01 5.88442028e-01
6.79620087e-01 -1.86603203e-01 6.63244605e-01 1.22851320e-01
-6.18397593e-01 1.02581036e+00 -3.18459779e-01 1.16125263e-01
8.00127268e-01 -9.64949012e-01 4.62271512e-01 7.53527939e-01
-2.85131428e-02 6.33391619e-01 7.76572466e-01 -6.74687445e-01
-1.16006374e+00 -1.32485366e+00 4.20755059e-01 -2.77943105e-01
7.49194324e-02 -4.39993232e-01 -1.00679410e+00 4.79502827e-01
2.49707058e-01 1.40506640e-01 3.44992757e-01 -3.84565383e-01
-2.79792875e-01 -2.89497435e-01 -1.42646694e+00 1.11970174e+00
1.14868307e+00 -1.03195779e-01 -3.31119299e-01 6.34408602e-03
5.67603290e-01 -3.07983786e-01 -7.70942450e-01 5.89374125e-01
5.80526531e-01 -9.65285003e-01 9.28192139e-01 -4.13912654e-01
8.80888641e-01 -3.06878239e-01 3.24074775e-01 -1.68278909e+00
-2.30909631e-01 -8.69364798e-01 8.96054879e-02 1.54674244e+00
3.00611585e-01 -5.30756891e-01 9.58929956e-01 6.78020775e-01
3.85286883e-02 -8.44255209e-01 -5.55635333e-01 -6.90819800e-01
2.96478122e-01 1.03918828e-01 1.12862647e+00 9.18868899e-01
-3.82510453e-01 2.27287605e-01 -7.35572815e-01 -1.48395419e-01
5.26235759e-01 1.63959593e-01 1.14226508e+00 -8.88964772e-01
-6.78208396e-02 -4.58559990e-01 -1.76047891e-01 -8.78543675e-01
2.27331787e-01 -5.02312720e-01 4.96435650e-02 -1.11924028e+00
-5.26206419e-02 -7.92079628e-01 2.61106759e-01 4.77081686e-01
-4.17282403e-01 6.23074412e-01 3.44422519e-01 2.06124514e-01
3.47644240e-01 9.65314209e-01 2.11929440e+00 -7.06581473e-02
-1.83700711e-01 -1.16901942e-01 -9.60563123e-01 7.31732905e-01
7.97190666e-01 -1.14378676e-01 -7.10067630e-01 -4.84788567e-01
-1.14471592e-01 1.04239851e-01 3.71309876e-01 -7.86297381e-01
-2.62922645e-01 -4.95433092e-01 7.33525455e-01 -2.52096862e-01
5.45799196e-01 -7.05257893e-01 5.67511976e-01 3.32666993e-01
-1.46309853e-01 4.13802415e-02 1.84437141e-01 3.28464270e-01
-3.66444409e-01 1.14968352e-01 1.18461299e+00 -1.87185809e-01
-2.95180947e-01 6.80013537e-01 1.31978884e-01 -1.94619015e-01
1.15420103e+00 -2.30250448e-01 -5.38412146e-02 -6.50246024e-01
-4.06594187e-01 2.95401942e-02 9.32547271e-01 7.09120989e-01
7.96681881e-01 -1.85591626e+00 -9.52391028e-01 5.34139335e-01
4.17091139e-02 4.19799894e-01 4.31805283e-01 3.13712955e-01
-6.44735038e-01 3.85972932e-02 -5.31540930e-01 -3.75137627e-01
-1.07884812e+00 6.69634938e-01 1.67017683e-01 -1.49447303e-02
-6.09168053e-01 6.38791680e-01 8.16964388e-01 -2.88280219e-01
-1.55817837e-01 -8.45282823e-02 -1.02866143e-01 -2.72729874e-01
4.98059928e-01 -5.22351675e-02 -3.28567445e-01 -6.73675537e-01
-5.45054786e-02 7.45854318e-01 -9.02928039e-02 -6.75579533e-02
1.25307786e+00 9.03022960e-02 3.35494988e-02 -2.48703301e-01
1.10823894e+00 2.37385616e-01 -1.65671718e+00 8.67933556e-02
-7.54104018e-01 -8.80339384e-01 -1.41969517e-01 -6.12300456e-01
-1.58278656e+00 9.39574718e-01 3.03330749e-01 -9.41818580e-02
1.27088368e+00 -2.36637875e-01 9.75926757e-01 -2.54643649e-01
3.11794996e-01 -7.66580820e-01 3.74019831e-01 -1.19235083e-01
1.57729995e+00 -1.13580203e+00 -2.65331239e-01 -6.43601775e-01
-9.32355404e-01 9.36962664e-01 1.07926941e+00 -2.78551579e-01
1.77505568e-01 1.82608128e-01 1.47747412e-01 5.03589734e-02
-4.11727309e-01 1.79316267e-01 3.64403218e-01 6.97215915e-01
3.22899371e-01 -9.36842710e-02 -2.90375948e-01 4.10609990e-01
-5.15564203e-01 -1.98058262e-01 3.63476992e-01 4.84522521e-01
5.75348781e-03 -1.28789437e+00 -5.81283569e-01 1.05480492e-01
-3.19772571e-01 6.44243658e-02 -3.33941638e-01 9.83518362e-01
3.51590246e-01 6.72567725e-01 5.40985465e-02 -3.42004836e-01
3.90206724e-01 -3.50994319e-01 6.16982341e-01 -5.32116115e-01
-2.30407462e-01 2.50712544e-01 -1.88802689e-01 -4.03967917e-01
-3.31071652e-02 -2.61226267e-01 -1.04860210e+00 -6.53380632e-01
-2.76041538e-01 1.16590988e-02 4.50687230e-01 5.79571307e-01
3.90661806e-01 5.43492794e-01 9.55223620e-01 -1.02766764e+00
-3.72929960e-01 -9.28550601e-01 -4.95856017e-01 7.94855237e-01
7.62798637e-02 -7.31046021e-01 -2.77210951e-01 1.48728013e-01] | [12.161757469177246, -0.35785606503486633] |
1a8df427-9afa-4e9c-9982-6ebe4fe5bebd | robust-face-recognition-with-structural | 1506.00481 | null | http://arxiv.org/abs/1506.00481v1 | http://arxiv.org/pdf/1506.00481v1.pdf | Robust Face Recognition with Structural Binary Gradient Patterns | This paper presents a computationally efficient yet powerful binary framework
for robust facial representation based on image gradients. It is termed as
structural binary gradient patterns (SBGP). To discover underlying local
structures in the gradient domain, we compute image gradients from multiple
directions and simplify them into a set of binary strings. The SBGP is derived
from certain types of these binary strings that have meaningful local
structures and are capable of resembling fundamental textural information. They
detect micro orientational edges and possess strong orientation and locality
capabilities, thus enabling great discrimination. The SBGP also benefits from
the advantages of the gradient domain and exhibits profound robustness against
illumination variations. The binary strategy realized by pixel correlations in
a small neighborhood substantially simplifies the computational complexity and
achieves extremely efficient processing with only 0.0032s in Matlab for a
typical face image. Furthermore, the discrimination power of the SBGP can be
enhanced on a set of defined orientational image gradient magnitudes, further
enforcing locality and orientation. Results of extensive experiments on various
benchmark databases illustrate significant improvements of the SBGP based
representations over the existing state-of-the-art local descriptors in the
terms of discrimination, robustness and complexity. Codes for the SBGP methods
will be available at
http://www.eee.manchester.ac.uk/research/groups/sisp/software/. | ['Weilin Huang', 'Hujun Yin'] | 2015-06-01 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [ 3.22583884e-01 -5.52063346e-01 -4.41402912e-01 -5.44217587e-01
-3.24824065e-01 -2.68544644e-01 5.34657896e-01 -8.45023170e-02
-2.37028226e-01 5.08228898e-01 -1.62604507e-02 -1.27430514e-01
-4.53194678e-01 -9.10718143e-01 -2.55891740e-01 -1.05190778e+00
-4.73372221e-01 -3.18652064e-01 3.51806939e-01 -3.04472804e-01
5.70376694e-01 1.05338454e+00 -1.87357700e+00 3.46631855e-01
5.14982045e-01 1.60398495e+00 -4.52391757e-03 5.80215573e-01
1.05901383e-01 4.42639142e-01 -3.82630438e-01 -2.01137215e-01
4.04853821e-01 -2.22387567e-01 -3.35280985e-01 1.79991156e-01
7.22410023e-01 -5.97633459e-02 -2.15899318e-01 1.33623302e+00
8.54749501e-01 4.05512145e-03 6.69707000e-01 -1.01083434e+00
-7.94407070e-01 -3.12189609e-01 -7.88283408e-01 4.67929125e-01
6.84264660e-01 -1.66387215e-01 8.12236667e-01 -1.23997927e+00
5.49959362e-01 1.22454035e+00 7.34558463e-01 2.52018660e-01
-1.04842579e+00 -6.14245951e-01 -2.69867450e-01 4.37394857e-01
-1.64716291e+00 -5.93906343e-01 1.14533782e+00 -1.22513250e-01
7.75421858e-01 7.04304874e-01 3.96947652e-01 6.36290789e-01
3.22104126e-01 4.28828567e-01 1.43139970e+00 -6.08940125e-01
1.58765420e-01 -2.59283604e-03 5.15283011e-02 1.39051759e+00
1.32847816e-01 2.00354964e-01 -6.86766982e-01 -2.11242184e-01
8.86901915e-01 7.26111680e-02 -4.39802587e-01 -4.07065630e-01
-6.77033067e-01 6.48124218e-01 5.41943133e-01 5.18552542e-01
-2.08887771e-01 -6.29867539e-02 2.31610909e-01 5.21719933e-01
3.53625953e-01 -5.24176061e-02 -6.62184134e-02 1.81413203e-01
-1.01309049e+00 -7.56951869e-02 5.96093714e-01 5.08222461e-01
9.16269004e-01 6.23914450e-02 4.47241095e-04 1.21758246e+00
2.65815705e-01 6.98417127e-01 6.96559608e-01 -7.45974600e-01
1.85925826e-01 5.89508832e-01 -2.43731603e-01 -1.93459773e+00
-3.59359741e-01 -8.10529590e-02 -8.89073491e-01 5.85960925e-01
2.06177905e-01 2.86878675e-01 -1.03033900e+00 1.34480596e+00
3.15598845e-01 7.15664728e-03 -2.44615644e-01 8.31900358e-01
9.97759879e-01 6.49863243e-01 -2.43053883e-01 -1.06366560e-01
1.34838593e+00 -5.60528994e-01 -4.71426129e-01 -9.15770531e-02
2.21619248e-01 -8.27137351e-01 8.19592416e-01 2.75179803e-01
-1.02211714e+00 -7.44059443e-01 -1.21582258e+00 1.86469376e-01
-6.59955680e-01 3.92853081e-01 7.86094308e-01 1.01830626e+00
-1.51097858e+00 6.62648141e-01 -5.48687816e-01 -3.04327875e-01
3.63438129e-01 6.27432823e-01 -6.89846635e-01 -2.03709379e-01
-9.46753502e-01 4.42590088e-01 5.63611686e-02 3.15961003e-01
-1.86806142e-01 -2.93094009e-01 -1.11622512e+00 -4.61266041e-02
-1.68865159e-01 1.80205598e-01 4.74918902e-01 -9.91968036e-01
-1.55407548e+00 1.12010467e+00 -4.11619902e-01 1.41842619e-01
2.70406425e-01 4.22574908e-01 -8.11739743e-01 7.23833740e-01
-5.89637086e-02 5.77352345e-01 1.26235521e+00 -8.18447590e-01
-4.83769208e-01 -5.21122932e-01 -2.83020586e-01 7.85559043e-02
-4.81750488e-01 3.33736002e-01 -4.24879998e-01 -7.65626729e-01
5.37350833e-01 -6.68445110e-01 1.25245731e-02 3.61274868e-01
6.66698813e-02 -7.12188408e-02 1.09488404e+00 -5.93854845e-01
1.20142269e+00 -2.24896932e+00 -2.52867699e-01 7.93066740e-01
5.29928319e-02 3.18796962e-01 -2.17685521e-01 2.04848647e-01
-2.31983393e-01 -9.98626091e-03 -1.37771875e-01 1.66975841e-01
-1.33742630e-01 3.99203785e-02 -1.14241391e-02 8.39984357e-01
1.65375471e-01 7.74993539e-01 -6.84542358e-01 -6.14685953e-01
2.01716527e-01 6.31519794e-01 -2.65220612e-01 -6.22640401e-02
5.83441019e-01 -2.68600136e-02 -4.10898030e-01 1.11413181e+00
1.08365476e+00 6.37197495e-02 1.55841902e-01 -3.33307505e-01
-3.04964036e-02 -1.75131127e-01 -1.21195149e+00 1.00999010e+00
-2.07690716e-01 7.00874984e-01 2.59313166e-01 -1.30134153e+00
1.34695160e+00 8.67584869e-02 4.02396798e-01 -9.40256357e-01
1.58916175e-01 3.51616263e-01 -2.47217819e-01 -3.24512810e-01
2.83668935e-01 -5.21327816e-02 1.81432351e-01 1.60293579e-01
1.32403031e-01 -2.26410944e-03 4.41482931e-01 -1.79713666e-01
6.66005909e-01 -1.68148220e-01 3.88365686e-01 -7.26742089e-01
9.97648180e-01 -6.27615690e-01 5.75105429e-01 5.58196008e-01
-3.68632048e-01 5.09303272e-01 2.36686110e-01 -5.24625421e-01
-4.84688848e-01 -1.06212080e+00 -6.98881805e-01 1.03003383e+00
2.53970325e-01 -2.99492031e-01 -4.70800519e-01 -4.41599429e-01
2.21417218e-01 -3.30129206e-01 -6.71774328e-01 -5.06360680e-02
-4.73372728e-01 -1.04342675e+00 3.65306616e-01 4.95912373e-01
1.01014555e+00 -1.02368617e+00 -6.57248199e-01 -7.06994012e-02
2.42199108e-01 -8.84470522e-01 -4.72797126e-01 -8.40752050e-02
-8.19115579e-01 -9.36427295e-01 -8.98034513e-01 -1.20747948e+00
9.64888096e-01 3.80900621e-01 7.78407037e-01 2.63732642e-01
-8.34600091e-01 5.14543235e-01 -2.68237799e-01 2.14467999e-02
9.57573056e-02 -5.73563933e-01 1.30228251e-01 4.26386893e-01
4.01727378e-01 -5.43096542e-01 -7.79362142e-01 5.55587888e-01
-7.02757299e-01 -4.68061745e-01 5.73501468e-01 1.09090376e+00
6.82689786e-01 3.97555023e-01 2.13213831e-01 -5.44595540e-01
6.75359726e-01 -1.86911717e-01 -5.50184309e-01 3.58402371e-01
-4.39158708e-01 -6.57130480e-02 3.33808869e-01 -4.01353627e-01
-9.98720109e-01 -9.63741690e-02 -8.39720443e-02 4.64923605e-02
-1.81768373e-01 3.57047498e-01 -4.15472984e-02 -1.15131032e+00
6.13961101e-01 4.85921234e-01 1.14943586e-01 -3.02920967e-01
4.25486416e-02 7.64185190e-01 5.21200836e-01 -6.44753337e-01
4.54830676e-01 6.41395092e-01 3.70868683e-01 -1.38401794e+00
-1.70341253e-01 -7.20961928e-01 -4.96605694e-01 -3.16637486e-01
3.46532762e-01 -4.61721748e-01 -6.59653962e-01 6.80839479e-01
-4.89679277e-01 -7.80810323e-03 2.09839493e-02 1.28962934e-01
-3.95174950e-01 8.47635925e-01 -7.81576097e-01 -7.24973500e-01
-2.84448624e-01 -1.02723289e+00 9.74153697e-01 4.80898410e-01
-2.13991553e-02 -1.25726461e+00 -2.94005066e-01 -1.18086308e-01
6.61359668e-01 4.41411614e-01 6.97175562e-01 1.00092269e-01
-1.25426695e-01 -4.87834632e-01 -2.83061177e-01 4.29773062e-01
3.38687867e-01 1.65961504e-01 -8.27279747e-01 -4.88703817e-01
1.51675358e-01 -9.89273265e-02 9.07380462e-01 2.56957322e-01
1.17820299e+00 -3.35250139e-01 -2.97804236e-01 8.67598474e-01
1.53478527e+00 2.22104236e-01 6.94695950e-01 2.30779544e-01
2.19712585e-01 6.18013144e-01 4.16755706e-01 4.29055542e-01
-2.18076244e-01 6.71637118e-01 7.78942928e-02 -6.06087804e-01
-3.09511572e-01 -2.60488242e-02 3.99643332e-01 5.12865067e-01
-2.90899754e-01 2.62284607e-01 -5.43479323e-01 3.32864195e-01
-1.35356665e+00 -1.06851041e+00 2.04679459e-01 2.06712818e+00
8.25961769e-01 -1.41970932e-01 -1.65443406e-01 5.60372889e-01
7.41117120e-01 5.32186449e-01 -1.30593181e-01 -7.72327065e-01
-5.12714744e-01 6.51542604e-01 6.04619265e-01 4.79371250e-01
-1.30223191e+00 8.02987874e-01 6.69579697e+00 1.12510741e+00
-1.48323750e+00 -2.10880831e-01 9.02486026e-01 3.12183380e-01
1.44835174e-01 -4.13481832e-01 -9.46331859e-01 4.63087797e-01
5.87278783e-01 1.36250984e-02 1.84531838e-01 8.50951672e-01
-8.05750936e-02 -3.63074511e-01 -5.01200497e-01 1.47246623e+00
2.88370192e-01 -1.15844440e+00 7.62307048e-02 9.70180631e-02
7.87186623e-01 -2.56506175e-01 5.82424879e-01 -3.20476919e-01
-2.54004180e-01 -1.14498615e+00 4.94200736e-01 3.28518748e-01
1.16965103e+00 -6.32250011e-01 6.46497905e-01 -3.29327047e-01
-1.40868485e+00 -2.45454878e-01 -7.87912488e-01 1.93914752e-02
-3.95638287e-01 7.47076809e-01 -3.75777125e-01 3.54965061e-01
9.67234612e-01 7.92107046e-01 -7.93386698e-01 9.46301222e-01
-2.73639977e-01 4.56550121e-01 -5.44549406e-01 -9.90522876e-02
2.70726919e-01 -6.30169928e-01 3.18859816e-01 1.58000159e+00
4.08354670e-01 9.11607221e-02 -1.44438762e-02 3.59254271e-01
1.67553842e-01 3.78330171e-01 -6.28534079e-01 2.04085290e-01
2.27498800e-01 1.39966643e+00 -8.45984221e-01 -2.22715378e-01
-3.93186152e-01 1.12684059e+00 6.14584237e-02 4.95882183e-01
-3.35378885e-01 -9.24950838e-01 6.83769882e-01 -3.38793360e-02
4.37348306e-01 -2.14763835e-01 -1.45838335e-01 -7.51051009e-01
2.45865554e-01 -8.97003293e-01 5.42783797e-01 -3.80811870e-01
-1.06407404e+00 7.66885698e-01 -2.33881861e-01 -9.88185942e-01
-5.90309389e-02 -1.26007140e+00 -6.37125790e-01 7.19501615e-01
-1.65789270e+00 -9.11924601e-01 -4.37497884e-01 9.50408876e-01
4.01606113e-02 -1.28957689e-01 6.44372702e-01 1.67632625e-01
-3.85828078e-01 9.89630282e-01 2.20546827e-01 1.46648765e-01
5.44561148e-01 -1.01044381e+00 -1.67753249e-02 7.96659112e-01
1.90559641e-01 5.83736479e-01 3.41432482e-01 -2.78354287e-01
-1.34666622e+00 -6.89332485e-01 8.64984512e-01 6.83028102e-02
4.48658168e-01 -2.27420524e-01 -8.18775892e-01 2.02328619e-02
-1.86680168e-01 4.24420983e-01 6.28305078e-01 -2.59038806e-01
-5.02310038e-01 -4.92414147e-01 -1.52629030e+00 5.74337542e-01
1.13976347e+00 -6.65808618e-01 -3.35783541e-01 3.65637392e-01
-4.14594442e-01 -1.76970154e-01 -8.43829572e-01 5.10519326e-01
7.82757103e-01 -1.38495576e+00 1.23800039e+00 -6.86876522e-03
-1.20296367e-01 -2.92384505e-01 -3.14472854e-01 -8.84027004e-01
-5.25145710e-01 -6.46456003e-01 1.87484965e-01 8.97758007e-01
1.65724769e-01 -1.10647416e+00 7.83059120e-01 3.14576924e-01
3.13732713e-01 -9.00239348e-01 -1.12389219e+00 -9.56582725e-01
-3.69892478e-01 -9.94836465e-02 4.34633225e-01 6.78031445e-01
3.64333570e-01 -3.76571894e-01 -1.00607842e-01 1.48695642e-02
7.49900818e-01 2.81798989e-01 1.80221781e-01 -8.96812081e-01
-4.66283001e-02 -7.27000833e-01 -1.24450254e+00 -1.00235426e+00
1.00328937e-01 -8.42302084e-01 -1.87855974e-01 -1.02673805e+00
-1.07526660e-01 -5.11711597e-01 -5.96266747e-01 6.50054276e-01
-7.20329583e-02 1.05154562e+00 -1.42400727e-01 5.52827157e-02
-1.94337204e-01 3.14948440e-01 1.03953016e+00 -1.69302434e-01
-7.45773166e-02 -1.76578611e-01 -4.25658733e-01 6.87054753e-01
1.02798975e+00 -5.54541945e-02 -2.11256370e-01 8.45142920e-03
-3.24375570e-01 -3.61375898e-01 1.67935669e-01 -1.16264343e+00
2.36832544e-01 1.37268379e-03 7.72331893e-01 -5.72158359e-02
5.41942477e-01 -4.61886466e-01 -1.23860218e-01 5.26900411e-01
1.06935896e-01 9.31270644e-02 1.90491155e-01 2.82269925e-01
-6.72541797e-01 -1.54471725e-01 1.20565689e+00 -2.39595603e-02
-1.24781847e+00 3.08651119e-01 -1.74328387e-01 -5.52582145e-01
9.29862738e-01 -8.91202748e-01 -1.16455153e-01 -3.11328560e-01
-4.49431509e-01 -3.41646433e-01 4.73510027e-01 1.09789990e-01
1.03585052e+00 -1.40318310e+00 -5.00008345e-01 9.44298804e-01
-5.24331480e-02 -9.45654452e-01 7.37861842e-02 7.65143692e-01
-8.95849466e-01 5.02633631e-01 -6.86354697e-01 -5.43929338e-01
-1.68701100e+00 3.98343533e-01 4.49769825e-01 7.55077749e-02
-5.74875534e-01 1.06205869e+00 1.40766099e-01 8.65754089e-04
2.87098046e-02 -1.17166698e-01 -3.18128556e-01 -4.79791611e-02
8.77311230e-01 4.82248724e-01 3.06076467e-01 -1.03006494e+00
-6.07204258e-01 1.30857849e+00 1.82390988e-01 3.79554108e-02
1.11807740e+00 5.56496792e-02 -5.31567335e-01 -2.97577530e-01
1.65183425e+00 2.49744549e-01 -1.12413728e+00 -5.80180511e-02
3.21303681e-02 -8.64865780e-01 1.96369797e-01 -5.47901869e-01
-1.37445915e+00 7.79280126e-01 9.34027970e-01 6.38089254e-02
1.70987058e+00 -2.28782117e-01 3.46481115e-01 1.34972438e-01
4.50327247e-01 -1.01012099e+00 3.85977291e-02 1.87521830e-01
1.19306469e+00 -9.85703707e-01 1.12807028e-01 -7.95106411e-01
-3.04811239e-01 1.42495990e+00 7.84696788e-02 -2.72311658e-01
1.04268110e+00 2.99806535e-01 2.08808005e-01 -2.55611151e-01
-5.81051409e-02 -1.99560866e-01 6.79528654e-01 7.64750779e-01
3.72737437e-01 1.03447773e-01 -6.18571639e-01 -1.23206891e-01
-5.10333329e-02 -2.23896742e-01 4.80374768e-02 1.06903768e+00
-6.20427966e-01 -1.05084896e+00 -5.11231005e-01 2.62951523e-01
-6.01571381e-01 -9.65232253e-02 -2.96183795e-01 5.42859018e-01
-1.72711723e-02 9.34937477e-01 -4.88493219e-02 -3.55634511e-01
1.74339116e-01 -8.09227452e-02 6.71978414e-01 -8.14712197e-02
-1.53487056e-01 -3.27646062e-02 -1.14479654e-01 -9.14188206e-01
-4.83136415e-01 -3.79821837e-01 -8.91909957e-01 -3.04473221e-01
-1.67941395e-02 2.70903677e-01 6.48670912e-01 2.69303173e-01
4.53921080e-01 -1.83988139e-01 8.83412361e-01 -1.06619775e+00
-4.74453211e-01 -6.98948741e-01 -9.35997784e-01 4.79792446e-01
2.19975337e-01 -7.93494999e-01 -5.21912098e-01 -4.16158251e-02] | [10.55929183959961, -0.35410141944885254] |
d99a84d7-9202-47d6-9ffc-4c47611fdcb1 | interpretability-then-what-editing-machine | 2206.15465 | null | https://arxiv.org/abs/2206.15465v1 | https://arxiv.org/pdf/2206.15465v1.pdf | Interpretability, Then What? Editing Machine Learning Models to Reflect Human Knowledge and Values | Machine learning (ML) interpretability techniques can reveal undesirable patterns in data that models exploit to make predictions--potentially causing harms once deployed. However, how to take action to address these patterns is not always clear. In a collaboration between ML and human-computer interaction researchers, physicians, and data scientists, we develop GAM Changer, the first interactive system to help domain experts and data scientists easily and responsibly edit Generalized Additive Models (GAMs) and fix problematic patterns. With novel interaction techniques, our tool puts interpretability into action--empowering users to analyze, validate, and align model behaviors with their knowledge and values. Physicians have started to use our tool to investigate and fix pneumonia and sepsis risk prediction models, and an evaluation with 7 data scientists working in diverse domains highlights that our tool is easy to use, meets their model editing needs, and fits into their current workflows. Built with modern web technologies, our tool runs locally in users' web browsers or computational notebooks, lowering the barrier to use. GAM Changer is available at the following public demo link: https://interpret.ml/gam-changer. | ['Rich Caruana', 'Jennifer Wortman Vaughan', 'Mihaela Vorvoreanu', 'Duen Horng Chau', 'Mark E. Nunnally', 'Peter Stella', 'Harsha Nori', 'Alex Kale', 'Zijie J. Wang'] | 2022-06-30 | null | null | null | null | ['additive-models', 'model-editing'] | ['methodology', 'natural-language-processing'] | [ 1.21871345e-01 4.80227888e-01 -2.09908739e-01 -4.84673172e-01
-3.82724226e-01 -5.12956202e-01 -9.92394686e-02 6.34687841e-01
-1.42044425e-01 3.01957250e-01 4.67812061e-01 -1.02967870e+00
-2.50721157e-01 -4.93398517e-01 -3.43417078e-01 7.54738674e-02
2.20403925e-01 7.39793837e-01 -2.83163100e-01 5.69275804e-02
4.62537736e-01 8.51289332e-02 -9.54469740e-01 6.91462636e-01
1.12014151e+00 8.62706676e-02 -7.87738562e-02 7.42948353e-01
-1.82848111e-01 1.38256240e+00 -1.97341502e-01 -3.50571543e-01
2.72170097e-01 -3.21777344e-01 -6.21022224e-01 -6.29542470e-01
2.10691243e-01 -6.04128063e-01 2.97622412e-01 5.43193996e-01
5.33902705e-01 -2.60197669e-01 2.72977799e-01 -1.37141240e+00
-8.62115085e-01 5.22688448e-01 -2.34969392e-01 1.83929339e-01
4.60088462e-01 1.05079663e+00 5.51133394e-01 -4.15556371e-01
5.37751675e-01 1.22436726e+00 1.24801052e+00 5.95076680e-01
-1.48194242e+00 -7.65086353e-01 6.04930334e-02 -5.88072417e-03
-1.09416127e+00 -4.13452208e-01 1.46514580e-01 -1.16232681e+00
1.18930936e+00 7.23900020e-01 8.58528674e-01 1.06603432e+00
2.55869269e-01 2.64743060e-01 1.01060534e+00 -3.00565213e-01
3.20372224e-01 4.29031372e-01 5.38700581e-01 8.37301195e-01
3.44710976e-01 -3.13667767e-02 -5.21130264e-01 -6.40641451e-01
7.06338167e-01 4.07641917e-01 -1.13902107e-01 1.15294717e-01
-1.16389573e+00 6.20589375e-01 2.81259287e-02 -1.08706668e-01
-5.67883372e-01 1.41550601e-01 2.16307994e-02 1.44179285e-01
6.88474953e-01 6.65461361e-01 -5.32478988e-01 -7.98466206e-01
-5.88477850e-01 2.89402515e-01 1.00542533e+00 6.84137225e-01
2.43290678e-01 -2.75622159e-01 -6.29582256e-02 7.49959171e-01
5.83947539e-01 2.40417063e-01 -4.82964329e-02 -9.87785697e-01
-3.16322446e-02 1.13622737e+00 3.52443755e-01 -1.04043388e+00
-7.60974884e-01 -2.17387542e-01 -6.13389790e-01 5.37246048e-01
5.25552332e-01 -2.68160075e-01 -5.13722360e-01 1.32062197e+00
3.39628607e-01 7.68822059e-02 -4.88365203e-01 6.31014049e-01
7.32637107e-01 -7.67622963e-02 8.37750256e-01 1.26873598e-01
1.23840618e+00 -4.30015832e-01 -6.06272161e-01 -2.02221990e-01
1.16868031e+00 -6.44160748e-01 1.46562064e+00 6.04161561e-01
-1.13568735e+00 -2.05997407e-01 -7.89415359e-01 -1.77313149e-01
-3.66308808e-01 -4.26718324e-01 4.71082181e-01 7.56412804e-01
-1.00282478e+00 8.62518430e-01 -1.18553185e+00 -5.80947578e-01
7.14841127e-01 1.03452869e-01 -4.44473438e-02 7.52736554e-02
-8.04949045e-01 1.25458813e+00 1.26767904e-01 -2.74933577e-01
-2.31266558e-01 -1.66410220e+00 -3.94935161e-01 -2.73076240e-02
1.25867918e-01 -1.44079459e+00 1.34945488e+00 -7.22163796e-01
-6.98088586e-01 6.56829953e-01 -9.73415226e-02 -1.65142193e-01
8.61389101e-01 -4.56468284e-01 -2.54674375e-01 -5.21318793e-01
-2.36268148e-01 3.69360417e-01 6.95644245e-02 -7.78000772e-01
-5.65644801e-01 -3.89751375e-01 -8.50630030e-02 -5.86294420e-02
-5.27590550e-02 3.28786522e-01 2.02664807e-01 -3.98616135e-01
-2.90951699e-01 -5.75388312e-01 -5.62548757e-01 3.61037105e-01
-1.73135653e-01 8.54096860e-02 5.11030495e-01 -1.08928382e+00
1.78058732e+00 -1.86826015e+00 -3.69402170e-01 2.08871603e-01
6.46841407e-01 4.37088311e-01 3.32787693e-01 5.19258678e-01
-2.43743569e-01 7.03506231e-01 -2.26862013e-01 -1.73536256e-01
1.67568788e-01 -2.96861231e-02 6.31995052e-02 -9.57005471e-03
9.50778276e-02 9.87326264e-01 -1.25036371e+00 -2.67015666e-01
5.80100417e-01 5.64337432e-01 -8.01691592e-01 1.66302279e-01
-3.38489503e-01 6.22509956e-01 -2.35832870e-01 4.17193085e-01
6.49844110e-01 -5.21607041e-01 5.64650536e-01 2.58945405e-01
-2.14747041e-01 4.55602705e-01 -1.20297849e+00 1.08441401e+00
-3.33216250e-01 3.30030292e-01 -1.82161570e-01 -3.08708757e-01
7.49015927e-01 7.12324679e-02 3.27825963e-01 -2.22609535e-01
-3.33333135e-01 7.08086863e-02 -1.07459659e-02 -9.20984745e-01
-5.02469316e-02 5.76860681e-02 4.83422071e-01 9.38995779e-01
-7.30654359e-01 -7.42611196e-03 -1.81645006e-01 -1.76907517e-02
1.43602741e+00 1.39046684e-01 8.13692391e-01 -1.53233647e-01
-2.70802468e-01 2.41889924e-01 5.05180955e-01 9.96433198e-01
8.24832693e-02 3.66696417e-01 7.02628255e-01 -1.04948032e+00
-1.04680026e+00 -1.01073861e+00 -4.98539470e-02 1.10253704e+00
-6.99582994e-01 -7.28760958e-01 -6.54711127e-01 -4.85024512e-01
2.98227131e-01 1.44506431e+00 -5.87303996e-01 -2.36643121e-01
-1.51676521e-01 -6.87354863e-01 5.39171159e-01 5.41970253e-01
3.49858440e-02 -8.83388221e-01 -1.10690570e+00 3.61948907e-01
-5.24229445e-02 -3.21005583e-01 -4.40714538e-01 -3.13746035e-01
-8.81977081e-01 -1.46642005e+00 8.45673904e-02 2.35919957e-03
5.71368933e-01 -7.46305147e-03 1.25861013e+00 4.65207130e-01
-6.44520998e-01 6.98484898e-01 -2.45035470e-01 -1.10039580e+00
-9.59603548e-01 -4.03150707e-01 -1.78224668e-01 -7.43887663e-01
9.46747124e-01 -7.89846957e-01 -8.21760058e-01 3.51996511e-01
-7.98605800e-01 8.54533434e-01 1.02618523e-01 3.75561774e-01
1.44024432e-01 -6.38985932e-01 4.68070418e-01 -1.33946657e+00
9.44110990e-01 -7.52727509e-01 -2.56773502e-01 3.74082714e-01
-1.38617373e+00 -2.55417019e-01 3.44447285e-01 -9.73761380e-02
-6.73514247e-01 -7.93591440e-02 2.28634663e-02 -5.90509847e-02
-3.25758457e-01 6.96745336e-01 4.83978502e-02 3.25159341e-01
1.08414316e+00 -3.00555110e-01 2.90519297e-01 -7.87550867e-01
3.03279936e-01 8.74777734e-01 1.65881768e-01 -2.13568404e-01
2.27183923e-01 3.34658436e-02 -4.22319561e-01 -3.44224125e-01
-3.15448195e-01 -1.54452428e-01 -5.62534094e-01 -2.45202437e-01
6.22978568e-01 -7.31076658e-01 -1.02674115e+00 2.75850713e-01
-1.00792587e+00 -8.59896064e-01 -2.64571011e-01 3.28463554e-01
-2.69171834e-01 1.10749625e-01 -8.18017647e-02 -8.44714761e-01
-5.69342315e-01 -7.49460161e-01 4.20520872e-01 1.21813633e-01
-1.38232160e+00 -1.15998507e+00 1.88080803e-01 5.00968754e-01
7.74175882e-01 5.74580193e-01 1.23968792e+00 -9.20174420e-01
-3.95777613e-01 -8.95379558e-02 -2.69700676e-01 9.01156962e-02
3.96907032e-01 4.00701582e-01 -8.54243159e-01 2.39893913e-01
-3.30695480e-01 2.77363628e-01 8.94148499e-02 3.89720827e-01
1.34671581e+00 -8.22806120e-01 -4.03818071e-01 4.41477299e-01
7.99168468e-01 5.33585012e-01 5.37411213e-01 3.75046581e-01
7.60350406e-01 4.70130831e-01 2.83639461e-01 8.35606217e-01
6.59564137e-01 4.24486518e-01 1.25290707e-01 -2.95732200e-01
2.06564888e-01 -4.36337352e-01 2.15612128e-02 2.13017538e-01
-2.05360919e-01 1.47952706e-01 -1.62359357e+00 1.16899274e-01
-2.31376505e+00 -7.86841512e-01 -4.47975129e-01 2.22114491e+00
8.64560843e-01 1.49863169e-01 2.52002031e-01 -4.82240319e-01
2.00130537e-01 -3.43960971e-01 -8.13630760e-01 -8.77073705e-01
5.14815748e-01 1.07899539e-01 3.04783195e-01 6.56649709e-01
-7.46610999e-01 4.15543646e-01 6.58332348e+00 1.03585042e-01
-9.34974074e-01 3.37739497e-01 9.06618416e-01 -5.24893701e-01
-7.00212955e-01 1.97885595e-02 -6.72714561e-02 5.86432457e-01
1.11032438e+00 -6.44062817e-01 6.90585375e-01 1.08963287e+00
1.17552567e+00 1.08631402e-02 -1.51293898e+00 8.55653405e-01
-4.26548630e-01 -1.57147217e+00 -3.08189154e-01 -6.31401986e-02
2.18810052e-01 6.83338894e-03 1.35430858e-01 1.90424249e-02
9.56996977e-01 -1.47540462e+00 4.97296602e-01 1.24935281e+00
6.77863181e-01 -5.31274796e-01 3.27304721e-01 3.59734595e-01
-3.56244415e-01 2.02385876e-02 2.31339466e-02 -5.41635096e-01
9.02522951e-02 6.09691978e-01 -1.70377517e+00 1.08915210e-01
6.43963158e-01 7.46229470e-01 -6.57588243e-01 9.70353007e-01
1.60067722e-01 7.84300566e-01 -2.66513944e-01 8.67937207e-02
-4.64719206e-01 -5.43375053e-02 4.91943687e-01 1.64926589e+00
1.29269898e-01 3.24781537e-01 -8.04531127e-02 1.27392995e+00
4.57642913e-01 3.92842554e-02 -5.19237638e-01 -3.24695379e-01
3.98784697e-01 9.56672430e-01 -2.15555653e-01 -2.31367499e-01
-2.08923042e-01 3.14263165e-01 1.00128263e-01 3.64136025e-02
-6.26671314e-01 -9.92283318e-03 1.00189149e+00 7.46370614e-01
-7.31995344e-01 -1.52759075e-01 -1.02224338e+00 -5.99496901e-01
-4.88881543e-02 -1.51165247e+00 5.98843217e-01 -1.03140259e+00
-1.04098141e+00 1.02574319e-01 1.25262365e-01 -1.03020859e+00
-3.66486430e-01 -4.56208974e-01 -5.95981836e-01 1.10225546e+00
-5.65314651e-01 -1.15585649e+00 -6.18137836e-01 9.67238918e-02
2.16860712e-01 2.46238306e-01 1.07642496e+00 1.00746274e-01
-5.42403221e-01 3.74951899e-01 6.55189231e-02 -2.53939599e-01
7.56756425e-01 -1.16185105e+00 9.37236965e-01 2.90618151e-01
-4.06590253e-01 1.26250410e+00 8.07660222e-01 -1.07210946e+00
-9.14133251e-01 -1.04670143e+00 9.10032749e-01 -1.17562199e+00
5.99290371e-01 -2.07510769e-01 -1.13746071e+00 1.12094796e+00
-1.43627776e-02 -6.54944897e-01 1.36312413e+00 3.13275188e-01
-2.55772442e-01 3.01907450e-01 -1.21248651e+00 8.83687377e-01
1.08991754e+00 -3.12819779e-01 -5.08707881e-01 7.13772416e-01
4.88168210e-01 -5.29310107e-01 -8.73501241e-01 1.85568795e-01
8.14495206e-01 -9.39306915e-01 8.02669942e-01 -1.35743308e+00
4.76114541e-01 -2.07959488e-01 3.20231050e-01 -1.24491942e+00
-3.57109904e-01 -9.23812807e-01 6.93799229e-03 7.94750512e-01
6.40609086e-01 -1.06695926e+00 3.41849595e-01 1.91699040e+00
-5.23388833e-02 -6.42227888e-01 -4.38244462e-01 -9.04382672e-03
-1.60447210e-01 -9.06116605e-01 8.42165053e-01 1.31743193e+00
6.01759613e-01 -3.01329464e-01 -2.27190539e-01 9.94361565e-02
3.71755600e-01 -4.58283395e-01 9.14258659e-01 -1.26431525e+00
-6.00720108e-01 -4.87394005e-01 -1.44642085e-01 -2.79643118e-01
-6.93874121e-01 -9.61123288e-01 -7.04332590e-01 -1.81270075e+00
4.99513835e-01 -6.52270377e-01 -2.87686557e-01 1.26926613e+00
-4.55933779e-01 -3.92581999e-01 3.52498829e-01 2.33889937e-01
-3.64037424e-01 -4.16384071e-01 7.24623442e-01 6.14766665e-02
-5.76236010e-01 4.50940616e-02 -1.43531811e+00 9.35869336e-01
9.09848928e-01 -6.77909672e-01 -3.92158836e-01 -4.10071969e-01
5.93774557e-01 -3.19510400e-01 8.29522669e-01 -7.31920362e-01
1.16856344e-01 -7.23144948e-01 4.24598843e-01 -2.17087835e-01
-1.55965209e-01 -6.57032132e-01 1.14208961e+00 7.90998101e-01
-4.46905792e-01 1.32352084e-01 6.87879145e-01 4.09424305e-02
7.97746241e-01 -1.88035548e-01 4.64984030e-01 -3.67139429e-01
-2.60068238e-01 -9.76346284e-02 -5.07503092e-01 6.35305718e-02
1.00901830e+00 -3.38532180e-02 -7.92189300e-01 -7.11983502e-01
-9.33245480e-01 4.80265558e-01 7.10447609e-01 4.65645403e-01
4.97624397e-01 -6.41900778e-01 -7.63925254e-01 1.51947156e-01
2.10356973e-02 -3.49483192e-02 5.71119905e-01 1.02387047e+00
-9.38161075e-01 2.41409957e-01 -1.10827491e-01 -2.85527527e-01
-1.41807079e+00 3.99078310e-01 5.40712833e-01 -6.39854819e-02
-7.39549637e-01 5.87404013e-01 9.19861905e-03 -5.59219658e-01
1.88916270e-02 -4.62499231e-01 2.16810018e-01 -2.38740802e-01
9.64862823e-01 7.75990069e-01 -1.01811960e-01 3.00108314e-01
-4.31016564e-01 -3.32763530e-02 -1.73198029e-01 1.85805917e-01
1.64849305e+00 -1.04386829e-01 -3.73206325e-02 5.58526635e-01
5.37126601e-01 -1.98476851e-01 -1.11580110e+00 2.57060051e-01
5.65007478e-02 -6.56446874e-01 -3.43907058e-01 -1.48817694e+00
-2.23055080e-01 6.59585595e-01 6.87253416e-01 1.68415189e-01
7.88232744e-01 -2.69829690e-01 2.01263040e-01 1.28813997e-01
-6.26558438e-02 -7.96887577e-01 -3.43022287e-01 -1.03973098e-01
1.05915225e+00 -9.90663588e-01 1.80085853e-01 -1.99457631e-01
-7.90254414e-01 9.77305055e-01 5.58344305e-01 5.61552227e-01
6.57169461e-01 3.75149548e-01 4.69088614e-01 -1.77946433e-01
-1.09134972e+00 4.98357981e-01 6.68611452e-02 9.92566168e-01
5.70025265e-01 4.43464816e-01 -2.34851316e-01 9.93851483e-01
-2.44889632e-01 9.06211257e-01 5.47385991e-01 7.74771869e-01
-1.80404320e-01 -1.14670098e+00 -4.99431014e-01 9.37994003e-01
-2.94708550e-01 -3.23384553e-01 -4.97090638e-01 6.19961977e-01
1.34785980e-01 1.08887088e+00 1.34149520e-02 -4.32926923e-01
5.77288508e-01 4.65593249e-01 -7.42839351e-02 -5.76787293e-01
-8.77557456e-01 -3.83301556e-01 4.56972629e-01 -7.48349011e-01
3.58447403e-01 -8.80783379e-01 -1.02038860e+00 -8.62426639e-01
3.57882500e-01 -3.21005255e-01 5.43818414e-01 5.97228467e-01
1.00937712e+00 5.69484949e-01 1.18699841e-01 -7.40791261e-02
-4.26478535e-01 -8.03931296e-01 -6.09961338e-02 3.15469742e-01
1.47356495e-01 -2.28154510e-01 -2.24562749e-01 2.52811313e-01] | [8.642718315124512, 6.030359745025635] |
ad8b4abf-fd6b-4fcd-bbce-89e50dfcec45 | functional2structural-cross-modality-brain | 2205.07854 | null | https://arxiv.org/abs/2205.07854v1 | https://arxiv.org/pdf/2205.07854v1.pdf | Functional2Structural: Cross-Modality Brain Networks Representation Learning | MRI-based modeling of brain networks has been widely used to understand functional and structural interactions and connections among brain regions, and factors that affect them, such as brain development and disease. Graph mining on brain networks may facilitate the discovery of novel biomarkers for clinical phenotypes and neurodegenerative diseases. Since brain networks derived from functional and structural MRI describe the brain topology from different perspectives, exploring a representation that combines these cross-modality brain networks is non-trivial. Most current studies aim to extract a fused representation of the two types of brain network by projecting the structural network to the functional counterpart. Since the functional network is dynamic and the structural network is static, mapping a static object to a dynamic object is suboptimal. However, mapping in the opposite direction is not feasible due to the non-negativity requirement of current graph learning techniques. Here, we propose a novel graph learning framework, known as Deep Signed Brain Networks (DSBN), with a signed graph encoder that, from an opposite perspective, learns the cross-modality representations by projecting the functional network to the structural counterpart. We validate our framework on clinical phenotype and neurodegenerative disease prediction tasks using two independent, publicly available datasets (HCP and OASIS). The experimental results clearly demonstrate the advantages of our model compared to several state-of-the-art methods. | ['Liang Zhan', 'Heng Huang', 'Paul Thompson', 'Alex Leow', 'Olusola Ajilore', 'Scott Mackin', 'Yalin Wang', 'Lei Guo', 'Xiyao Fu', 'Haoteng Tang'] | 2022-05-06 | null | null | null | null | ['disease-prediction'] | ['medical'] | [ 2.70518601e-01 1.81930482e-01 -1.50695518e-01 -4.80098486e-01
1.83536157e-01 -4.13855702e-01 4.51728076e-01 -1.23447031e-01
-1.24041677e-01 6.51920795e-01 2.14379713e-01 -1.06000558e-01
-5.09254873e-01 -7.94276714e-01 -5.85760057e-01 -7.50810087e-01
-5.10398388e-01 4.10352826e-01 2.83354968e-01 5.25299460e-03
-1.54413715e-01 4.85240459e-01 -9.41039801e-01 1.63937673e-01
8.47303569e-01 9.65435624e-01 2.28968114e-01 3.73638391e-01
-8.92720371e-02 5.83453417e-01 -2.05626488e-01 -2.00359032e-01
1.84886649e-01 -4.69660938e-01 -7.07240820e-01 1.59046918e-01
2.94673115e-01 1.59385398e-01 -6.25347137e-01 1.34130275e+00
3.71859998e-01 -2.79186368e-01 6.54521704e-01 -1.36630225e+00
-8.46263647e-01 5.39764762e-01 -5.97933710e-01 4.89929318e-01
-1.71098039e-02 1.71712533e-01 1.07191741e+00 -3.93399507e-01
9.59540129e-01 1.10497487e+00 4.60187167e-01 5.46668768e-01
-1.43881774e+00 -5.94158709e-01 4.81331229e-01 3.83349538e-01
-1.00351417e+00 -1.20677195e-01 9.22794700e-01 -7.91117132e-01
6.50136888e-01 2.65368912e-03 1.17529070e+00 1.29993939e+00
4.88371313e-01 3.69268179e-01 1.40042162e+00 1.94069862e-01
1.94150135e-02 -5.47862709e-01 2.54220098e-01 9.52588141e-01
3.78378093e-01 -2.56312825e-02 -4.04173940e-01 -1.17964402e-01
6.90366447e-01 1.22675173e-01 -7.30044663e-01 -8.40705276e-01
-1.52264011e+00 7.27606475e-01 8.11464846e-01 5.27159870e-01
-4.99081135e-01 1.03690803e-01 4.14777607e-01 3.54942888e-01
5.66434443e-01 1.18838005e-01 -4.43847895e-01 2.84201682e-01
-8.02762687e-01 -1.69674993e-01 6.56246185e-01 4.78044808e-01
4.61895853e-01 3.70744653e-02 8.52540731e-02 7.23065376e-01
5.69365919e-01 2.58303314e-01 6.90462768e-01 -4.26937819e-01
3.88240755e-01 9.60923970e-01 -7.42056549e-01 -1.13779867e+00
-9.31066394e-01 -5.90567589e-01 -1.09371972e+00 8.84435326e-02
4.12193924e-01 -2.65763402e-02 -8.48397553e-01 2.01453519e+00
2.42575735e-01 4.54504788e-01 -3.77382517e-01 9.08022702e-01
8.84040356e-01 1.15673602e-01 -1.13478556e-01 -1.54430509e-01
1.29506576e+00 -6.26380205e-01 -4.89186496e-01 -3.10672253e-01
5.40502608e-01 7.00515509e-02 7.72900283e-01 4.35924605e-02
-7.14726746e-01 -2.30332743e-02 -9.89086032e-01 1.00983121e-01
-3.53432745e-01 -1.85252473e-01 7.27790892e-01 2.49540091e-01
-1.23114586e+00 5.79604149e-01 -1.17243409e+00 -4.03855234e-01
7.47424722e-01 3.76589626e-01 -7.59156227e-01 -1.19150974e-01
-1.13861179e+00 7.86633253e-01 1.86984241e-01 3.28970641e-01
-7.49982953e-01 -8.89156044e-01 -8.09191883e-01 5.21084070e-02
2.51437098e-01 -7.76029766e-01 3.49902242e-01 -7.73062050e-01
-1.11070085e+00 8.74936104e-01 2.58818474e-02 -3.21417898e-01
6.61594510e-01 4.04011697e-01 -5.61384261e-01 3.67033362e-01
-7.89676141e-03 5.28096259e-01 6.81555092e-01 -8.60516727e-01
1.72009006e-01 -9.49323773e-01 -1.15924142e-01 -5.76071776e-02
-4.49394077e-01 -2.30454147e-01 -9.11854059e-02 -4.97752339e-01
3.71798515e-01 -9.66289580e-01 -1.43548682e-01 4.10811216e-01
-6.45301461e-01 3.62089500e-02 8.05273056e-01 -8.40796471e-01
8.74973178e-01 -1.75876987e+00 6.41254067e-01 4.84479100e-01
9.52194154e-01 -5.54973260e-03 -2.58878529e-01 6.14842586e-02
-5.16053617e-01 2.46234253e-01 -3.52460921e-01 5.26530296e-02
-2.67412126e-01 1.47200987e-01 3.02471668e-01 8.54494512e-01
1.06476061e-01 1.16833007e+00 -1.08083153e+00 -4.17280376e-01
-1.16643995e-01 7.35320151e-01 -3.34249437e-01 -8.09071586e-02
1.51724279e-01 7.61424541e-01 -5.87716639e-01 5.01187325e-01
3.42239767e-01 -5.63870728e-01 7.22960174e-01 -5.41040480e-01
3.00538629e-01 1.32876456e-01 -6.33582115e-01 1.62440622e+00
-1.66166406e-02 5.68658650e-01 1.35129496e-01 -1.60290563e+00
9.02825892e-01 3.14186722e-01 8.66371810e-01 -7.37976432e-01
1.97851181e-01 -6.86755627e-02 6.78987384e-01 -7.45249987e-01
-5.35232306e-01 -7.87068009e-02 4.27478105e-01 4.53348368e-01
2.55263418e-01 4.42893505e-01 3.56179953e-01 1.43803090e-01
1.53410411e+00 5.03657535e-02 2.58603305e-01 -4.48091209e-01
6.38762951e-01 -4.59968001e-01 7.86899686e-01 1.87976941e-01
-5.14027297e-01 3.03180665e-01 1.07526290e+00 -2.76918411e-01
-6.76174998e-01 -1.26079118e+00 -1.03071384e-01 6.46421969e-01
-1.82728261e-01 -1.85672998e-01 -5.80997586e-01 -9.99560177e-01
1.52261421e-01 -1.07091330e-01 -8.30613852e-01 -4.24926996e-01
-6.28512084e-01 -9.07509089e-01 4.86985236e-01 5.36090732e-01
2.69513786e-01 -7.98241138e-01 -2.15486497e-01 9.79948323e-03
-8.64237919e-02 -1.24121535e+00 -5.46759605e-01 -9.67064202e-02
-1.21398544e+00 -1.60656118e+00 -6.02729201e-01 -8.61165047e-01
1.04411411e+00 1.30724981e-01 8.94701958e-01 2.27828249e-01
-4.00846362e-01 3.27759564e-01 -9.82244238e-02 3.29045914e-02
-2.60950208e-01 -5.56342341e-02 8.58901739e-02 4.23548579e-01
-1.32170960e-01 -1.11914361e+00 -8.09209824e-01 2.05213398e-01
-9.16395307e-01 1.41618147e-01 7.47008920e-01 8.40737700e-01
6.58238828e-01 -8.65316391e-02 8.40313196e-01 -9.44531322e-01
6.10673845e-01 -7.45486259e-01 -4.02599961e-01 4.55768943e-01
-7.32133746e-01 9.10614803e-02 5.29444933e-01 -4.57699925e-01
-7.31776416e-01 3.11042648e-02 2.43094206e-01 -4.17662024e-01
8.38449299e-02 8.93747032e-01 -4.19193536e-01 -2.31625706e-01
2.97753364e-01 2.35035315e-01 4.56080765e-01 -4.68777657e-01
3.72462600e-01 1.54601574e-01 4.96829182e-01 -3.68101716e-01
5.07625997e-01 6.49204850e-01 4.07833397e-01 -6.64610922e-01
-5.54443002e-01 -2.83045441e-01 -1.02620471e+00 -4.42930102e-01
8.79395425e-01 -4.94169414e-01 -6.11063302e-01 4.30841058e-01
-8.27815413e-01 -2.40095347e-01 1.75128710e-02 4.35938269e-01
-3.82287890e-01 5.81397235e-01 -4.61483926e-01 -1.23651184e-01
-4.21772420e-01 -1.16938746e+00 6.85612023e-01 -1.74963012e-01
-1.72863808e-02 -1.45568871e+00 1.89038619e-01 4.23981071e-01
3.02476317e-01 6.25940979e-01 1.42856061e+00 -5.96270323e-01
-3.05097818e-01 -7.01893121e-02 -3.61468077e-01 3.13490540e-01
3.15750301e-01 -2.65024692e-01 -3.28562409e-01 -3.62412035e-01
-1.41275853e-01 -7.93912634e-02 8.46911073e-01 4.12861705e-01
1.09928930e+00 -1.80567190e-01 -3.83672804e-01 5.67158759e-01
1.16782236e+00 -8.11827257e-02 3.93773615e-01 -2.26609074e-02
9.84902918e-01 8.02573681e-01 -3.19327056e-01 2.95587210e-03
6.45028889e-01 4.94320661e-01 5.20734370e-01 -1.51381351e-03
-3.28349203e-01 2.76794489e-02 5.22188306e-01 1.03597486e+00
-2.63099521e-01 -7.69654363e-02 -1.01147509e+00 5.48712194e-01
-1.80976129e+00 -8.27180684e-01 -2.45084509e-01 2.03602648e+00
7.94931650e-01 1.67948753e-01 8.00312962e-03 -1.91672415e-01
8.65126312e-01 2.47454956e-01 -9.22105849e-01 2.19958782e-01
-1.76304415e-01 1.04038544e-01 2.07097635e-01 7.20220094e-04
-6.64912939e-01 4.39694345e-01 5.91171980e+00 -4.20193672e-02
-1.48427379e+00 2.78390020e-01 4.19385642e-01 -1.80533782e-01
-2.08783612e-01 -2.97151767e-02 -1.03194170e-01 5.36377847e-01
8.65378320e-01 -8.78138691e-02 6.21103227e-01 3.28460664e-01
2.69234985e-01 3.56601149e-01 -1.33534801e+00 8.48519087e-01
-1.48629828e-03 -1.15574765e+00 2.31808312e-02 3.58488828e-01
4.68752056e-01 3.83830637e-01 -2.26588830e-01 -7.64241815e-03
-1.53884655e-02 -1.02784848e+00 4.46532965e-01 1.00842607e+00
7.58282602e-01 -2.47500762e-01 5.02878666e-01 2.57972240e-01
-1.27342248e+00 9.16955322e-02 -6.20995164e-02 2.55845577e-01
1.49931744e-01 7.89807498e-01 -6.91521049e-01 6.15621746e-01
6.53408527e-01 1.18742001e+00 -7.79340565e-01 1.03592432e+00
-2.49194026e-01 6.35194182e-01 6.87423944e-02 3.94658238e-01
-1.03670456e-01 -6.09419048e-01 6.63065672e-01 8.95098150e-01
2.52196789e-01 -2.11468145e-01 2.72086352e-01 1.34219849e+00
-2.11275473e-01 2.87524760e-02 -7.38692999e-01 -4.87572819e-01
8.48162249e-02 1.53168988e+00 -9.50193226e-01 -4.53686565e-02
-6.42187059e-01 6.58769131e-01 6.47407174e-01 3.90697718e-01
-6.11964583e-01 -6.91242218e-02 7.68369019e-01 1.73136324e-01
5.22722974e-02 -4.22660500e-01 -1.21227942e-01 -1.45317471e+00
1.39123484e-01 -5.64369202e-01 3.13792437e-01 -5.29740274e-01
-1.60356236e+00 5.36970139e-01 -3.27892303e-02 -9.66586113e-01
2.92910561e-02 -6.84455216e-01 -7.67860651e-01 6.55500591e-01
-1.61123538e+00 -1.25910997e+00 -3.93681616e-01 8.23367178e-01
-1.02091461e-01 -3.71395908e-02 5.68452775e-01 3.13105524e-01
-8.00745010e-01 2.35122472e-01 -2.04694510e-01 3.34355831e-01
4.03580964e-01 -1.15756404e+00 5.61918272e-03 6.51009023e-01
8.64820257e-02 5.79353094e-01 1.37801170e-01 -8.83907974e-01
-1.52524567e+00 -1.18742096e+00 6.05361938e-01 -6.20498024e-02
1.16426313e+00 -4.58038986e-01 -1.08746636e+00 6.75333202e-01
-1.34063020e-01 5.70878386e-01 6.06962144e-01 2.83913016e-02
-4.41706181e-01 -1.32605657e-01 -9.72383678e-01 4.46436882e-01
1.52710521e+00 -5.87070167e-01 -4.97104585e-01 4.49422449e-01
4.41337198e-01 1.45311207e-02 -1.24560916e+00 5.24576962e-01
5.82135499e-01 -7.58908272e-01 9.66158926e-01 -8.71019900e-01
2.84213036e-01 -1.20962739e-01 1.40080184e-01 -1.75360060e+00
-3.44062924e-01 -1.25504524e-01 -3.93620908e-01 9.43282068e-01
2.91916192e-01 -1.01439273e+00 7.09798217e-01 3.47013712e-01
-1.91339120e-01 -1.05979824e+00 -1.02029228e+00 -8.71209383e-01
1.01019904e-01 -2.20774531e-01 5.29136181e-01 1.14084649e+00
-5.08949980e-02 4.04775441e-01 -1.35135219e-01 4.84519936e-02
6.27187908e-01 8.45395252e-02 2.45565400e-02 -1.73929799e+00
-1.35477483e-01 -7.60029852e-01 -9.53074336e-01 -3.80100965e-01
6.63874686e-01 -1.67479908e+00 -4.13279176e-01 -1.70242012e+00
3.12515587e-01 -1.92818567e-01 -7.16153443e-01 6.20033622e-01
2.15322986e-01 1.06296673e-01 4.16235551e-02 7.06985667e-02
-2.70224065e-01 6.42585874e-01 1.61328101e+00 -4.30180073e-01
2.07818160e-03 -3.25915605e-01 -6.99240923e-01 8.14533234e-01
7.05951333e-01 -2.85399228e-01 -6.89991236e-01 -3.64245504e-01
1.80080920e-01 1.05150975e-01 6.37978613e-01 -8.09061587e-01
2.50757873e-01 1.10664852e-02 4.16131765e-01 -1.20364800e-02
2.12393999e-01 -7.74033725e-01 1.54407874e-01 5.56551933e-01
-2.87870705e-01 1.67961836e-01 -4.12806273e-01 9.04194593e-01
-8.02507699e-02 2.40762770e-01 7.19728470e-01 -1.49442762e-01
-4.11666900e-01 8.25558007e-01 -1.22942664e-01 2.14171261e-01
9.61956143e-01 -1.20836094e-01 -5.34981489e-01 -1.76358044e-01
-1.02300072e+00 2.65713930e-01 1.13575369e-01 5.79418898e-01
7.99215913e-01 -1.31245029e+00 -6.39579237e-01 1.58233151e-01
3.47580686e-02 -3.59003574e-01 3.34178567e-01 1.58299792e+00
-1.41864687e-01 3.31258833e-01 -6.07303321e-01 -6.04727924e-01
-1.20805550e+00 2.45213747e-01 5.37990332e-01 -5.08329630e-01
-9.17169690e-01 4.91598636e-01 4.06407624e-01 -5.12117684e-01
-2.31021605e-02 -3.97346497e-01 -4.96423900e-01 1.42026708e-01
1.96301714e-01 3.59608352e-01 2.12354422e-01 -9.31274414e-01
-4.52539742e-01 3.98248851e-01 -3.23837250e-02 2.42870614e-01
1.68082500e+00 3.08398586e-02 -6.91752672e-01 3.62743884e-01
1.33466959e+00 -5.10537148e-01 -1.03711641e+00 -4.46395636e-01
1.32388249e-01 -1.78122550e-01 3.18448752e-01 -8.23586106e-01
-1.94502032e+00 8.03849816e-01 6.26975238e-01 9.60331336e-02
1.04306078e+00 1.96815893e-01 7.19107926e-01 3.51503760e-01
4.13961500e-01 -6.80860221e-01 8.73573050e-02 2.52098083e-01
1.12854612e+00 -1.02437568e+00 3.73593830e-02 -4.83849645e-01
-3.54422599e-01 1.27728975e+00 6.92904830e-01 -8.55411738e-02
1.12988985e+00 -4.94959876e-02 -3.79423887e-01 -8.46593559e-01
-7.05581903e-01 8.38094857e-03 6.74169362e-01 6.34363234e-01
3.62457067e-01 3.16366374e-01 -4.02368963e-01 5.96010089e-01
-5.90476058e-02 -5.97423129e-02 3.29294950e-01 7.99073935e-01
-1.48057519e-02 -1.19577467e+00 4.61039245e-02 9.88912761e-01
-2.62913495e-01 1.21022305e-02 -6.70871496e-01 6.98261678e-01
8.60872120e-02 5.84101975e-01 -3.80286761e-02 -3.23105663e-01
2.61585593e-01 3.08491915e-01 6.37803674e-01 -6.35765612e-01
-1.19143322e-01 -1.86897352e-01 -5.87296933e-02 -7.24975467e-01
-5.91676295e-01 -7.48951852e-01 -1.44960451e+00 3.86021547e-02
-9.92705896e-02 -4.51165736e-01 6.13746524e-01 1.06053352e+00
6.68351412e-01 7.98426509e-01 4.21022713e-01 -7.99815893e-01
-2.41624996e-01 -8.83538902e-01 -8.27016592e-01 5.07373214e-01
1.29657924e-01 -1.06937969e+00 -1.18692063e-01 -5.21716066e-02] | [12.398632049560547, 3.3807077407836914] |
be021dab-928e-4cff-b912-c8b90b63b5ec | numerical-methods-for-convex-multistage | 2303.15672 | null | https://arxiv.org/abs/2303.15672v1 | https://arxiv.org/pdf/2303.15672v1.pdf | Numerical Methods for Convex Multistage Stochastic Optimization | Optimization problems involving sequential decisions in a stochastic environment were studied in Stochastic Programming (SP), Stochastic Optimal Control (SOC) and Markov Decision Processes (MDP). In this paper we mainly concentrate on SP and SOC modelling approaches. In these frameworks there are natural situations when the considered problems are convex. Classical approach to sequential optimization is based on dynamic programming. It has the problem of the so-called ``Curse of Dimensionality", in that its computational complexity increases exponentially with increase of dimension of state variables. Recent progress in solving convex multistage stochastic problems is based on cutting planes approximations of the cost-to-go (value) functions of dynamic programming equations. Cutting planes type algorithms in dynamical settings is one of the main topics of this paper. We also discuss Stochastic Approximation type methods applied to multistage stochastic optimization problems. From the computational complexity point of view, these two types of methods seem to be complimentary to each other. Cutting plane type methods can handle multistage problems with a large number of stages, but a relatively smaller number of state (decision) variables. On the other hand, stochastic approximation type methods can only deal with a small number of stages, but a large number of decision variables. | ['Alexander Shapiro', 'Guanghui Lan'] | 2023-03-28 | null | null | null | null | ['stochastic-optimization', 'type'] | ['methodology', 'speech'] | [ 2.55996212e-02 1.45285614e-02 -1.04927614e-01 1.13657461e-02
-5.35651505e-01 -6.24988019e-01 5.68275213e-01 2.75374264e-01
-5.90835273e-01 9.56293106e-01 -3.37077916e-01 -3.70163649e-01
-5.85695326e-01 -6.18561566e-01 -3.73842686e-01 -1.11020577e+00
-7.48334303e-02 8.31988394e-01 3.20827633e-01 -2.89669245e-01
1.84698358e-01 5.05190194e-01 -1.63489819e+00 -2.19860181e-01
8.76204491e-01 1.04814732e+00 2.67933965e-01 8.52895439e-01
-3.02788198e-01 3.40051264e-01 -2.99875528e-01 -2.93763012e-01
2.30351746e-01 -2.76740342e-01 -5.61230481e-01 1.19175203e-01
-4.92455393e-01 2.41417497e-01 2.61621147e-01 1.04169846e+00
5.20098329e-01 5.63766360e-01 8.25733423e-01 -1.18981290e+00
8.72140527e-02 5.56437135e-01 -5.28014362e-01 2.56062716e-01
1.24091111e-01 3.92194480e-01 7.39276230e-01 -5.89635134e-01
5.49041629e-01 1.40010238e+00 5.08835435e-01 5.96495926e-01
-1.63564599e+00 7.83839375e-02 5.72294772e-01 5.89490905e-02
-1.08653843e+00 -1.10036001e-01 3.96672875e-01 -7.58178234e-01
7.51891851e-01 3.99878383e-01 8.16646636e-01 7.65584826e-01
4.48560327e-01 6.69552803e-01 1.33546460e+00 -2.56956846e-01
7.63472438e-01 2.41061926e-01 3.68291557e-01 3.79821151e-01
3.54599983e-01 2.16830283e-01 2.61938497e-02 -1.04449004e-01
4.82919037e-01 3.39723416e-02 -2.36262977e-02 -4.12838250e-01
-9.32674646e-01 9.12552834e-01 -3.35383505e-01 3.27438533e-01
-5.11053860e-01 3.01796377e-01 3.90903085e-01 3.75912786e-01
4.94669586e-01 3.44272435e-01 -4.78903353e-01 -2.86850601e-01
-7.65997827e-01 8.06713641e-01 1.25679755e+00 7.81175733e-01
3.49253386e-01 7.99525678e-02 -3.26125681e-01 2.52067745e-01
2.35029787e-01 7.70415187e-01 8.78917724e-02 -8.09211433e-01
7.98248410e-01 6.61434531e-02 6.31856680e-01 -7.39692152e-01
-6.15527570e-01 -4.00614232e-01 -8.92778635e-01 4.44459766e-01
8.91256630e-01 -6.32157922e-01 -6.12119138e-01 1.58351004e+00
4.09301251e-01 -3.50740016e-01 1.25669837e-01 6.74399853e-01
-3.50242645e-01 9.97035027e-01 -1.90712824e-01 -8.32374930e-01
1.13824666e+00 -5.60805798e-01 -9.48118150e-01 -1.12375412e-02
3.10716659e-01 -5.56626320e-01 8.05371404e-01 5.79337120e-01
-1.27905273e+00 -8.44995603e-02 -6.33807063e-01 5.23888707e-01
-3.58821809e-01 -8.49915519e-02 4.15278792e-01 8.12384725e-01
-9.99558449e-01 1.01103652e+00 -1.01187301e+00 -1.92156896e-01
-1.12534665e-01 5.03707290e-01 4.72295016e-01 4.22592670e-01
-9.35592890e-01 1.07601798e+00 2.15332121e-01 5.03328145e-01
-9.72301662e-01 -5.10030150e-01 -4.75188285e-01 1.62184954e-01
9.28135455e-01 -5.83923221e-01 1.28557575e+00 -6.21183932e-01
-1.84151244e+00 4.46612418e-01 -2.35820144e-01 -3.97649437e-01
1.16353989e+00 -6.17265552e-02 1.19816490e-01 -1.49261132e-01
-1.79391444e-01 -1.55358523e-01 9.13922727e-01 -8.78797770e-01
-4.74682689e-01 -4.13587153e-01 5.69758750e-02 1.66847572e-01
2.08722398e-01 2.69418597e-01 1.59802184e-01 -4.54295665e-01
-5.87565601e-02 -1.20394659e+00 -9.58066583e-01 -1.95727900e-01
-3.99723709e-01 -2.19332963e-01 5.21853387e-01 -2.16210872e-01
1.27855682e+00 -1.53325582e+00 9.90388513e-01 5.49078397e-02
-1.22476049e-01 1.59247182e-02 4.84033316e-01 7.15340674e-01
6.21578842e-02 2.78554440e-01 -5.90783834e-01 -4.63807791e-01
2.99663574e-01 2.49186173e-01 -2.69349515e-01 5.25787115e-01
1.41871274e-01 5.36149204e-01 -8.36303055e-01 -3.45924199e-01
2.10051686e-01 -2.02366859e-01 -6.08055413e-01 -2.67897815e-01
-5.62930226e-01 4.11677957e-01 -8.12986255e-01 2.67147899e-01
4.60407495e-01 2.45126158e-01 2.20606513e-02 3.11890602e-01
-7.45365620e-01 -2.77548790e-01 -1.78650057e+00 1.14251792e+00
-7.09708095e-01 3.28236282e-01 4.46913034e-01 -1.20651722e+00
5.47982812e-01 3.36655796e-01 6.19633257e-01 2.96968848e-01
2.39406973e-01 2.64565021e-01 -7.73577020e-02 -4.87368524e-01
4.46714759e-01 -6.98718071e-01 -3.81320119e-01 4.93478507e-01
-2.49302328e-01 -5.30543089e-01 5.65990865e-01 -2.91191936e-01
8.08010101e-01 7.77899474e-02 3.77292693e-01 -7.57430017e-01
7.34639883e-01 1.20111771e-01 6.49661362e-01 7.09003806e-01
-1.97565719e-01 8.72692838e-02 1.24498224e+00 -3.25729281e-01
-1.06282949e+00 -7.50309169e-01 -1.28016755e-01 6.95345283e-01
2.05286983e-02 -1.37485981e-01 -7.53059745e-01 -1.48615286e-01
-1.75828785e-02 7.53829300e-01 -6.49642169e-01 1.61941499e-01
-6.16445184e-01 -1.16074622e+00 -1.75129905e-01 5.30506158e-03
5.38013503e-02 -7.80235231e-01 -7.89411128e-01 5.63940942e-01
1.97405532e-01 -7.91988552e-01 -9.17855576e-02 3.60457033e-01
-1.06983066e+00 -8.24034691e-01 -1.03748703e+00 -1.52614459e-01
4.73350108e-01 -2.57006198e-01 7.74007857e-01 -5.96956849e-01
-3.83362956e-02 2.56141096e-01 2.60470454e-02 -7.97876477e-01
-4.72785920e-01 7.32864663e-02 4.66887057e-01 2.33711809e-01
-2.00398579e-01 -4.21102434e-01 -1.46548524e-01 4.75996494e-01
-8.56963754e-01 -1.87520236e-01 4.64564823e-02 6.70279264e-01
7.54475236e-01 3.68319958e-01 6.56240806e-02 -7.16152906e-01
7.32778907e-01 -3.10127050e-01 -1.26250994e+00 1.53927609e-01
-4.43026185e-01 5.07099211e-01 8.60055983e-01 -5.97803056e-01
-1.06370556e+00 9.43440720e-02 1.44970477e-01 -3.91345233e-01
1.99812874e-01 5.62035561e-01 -3.80334742e-02 2.76380390e-01
2.43351996e-01 2.43366599e-01 -1.04545586e-01 -6.39354885e-01
3.74375992e-02 3.17879528e-01 -1.50686309e-01 -9.11651909e-01
6.68838799e-01 4.79362577e-01 7.42713571e-01 -7.60588348e-01
-7.21771836e-01 -1.48444816e-01 -6.20760024e-01 -5.03322363e-01
8.43630672e-01 -2.34538287e-01 -1.02682877e+00 6.89841628e-01
-9.77873147e-01 -5.41318893e-01 -7.75333166e-01 3.78855020e-01
-1.02008951e+00 6.43372610e-02 -3.96023363e-01 -1.62626576e+00
1.61253586e-01 -1.37316120e+00 7.87022531e-01 2.23439887e-01
1.06828853e-01 -1.20883811e+00 2.40713850e-01 -9.84055623e-02
3.25006276e-01 6.99923515e-01 7.23974884e-01 -1.72295168e-01
-3.41866374e-01 -3.99961382e-01 4.66145664e-01 1.13957137e-01
-3.79181594e-01 2.90757090e-01 -3.16378355e-01 -2.13522613e-01
7.19382703e-01 3.90271008e-01 5.36274910e-01 8.15841079e-01
8.34012866e-01 -3.71468753e-01 -1.94022566e-01 3.21740985e-01
1.75815690e+00 3.47819686e-01 8.76614898e-02 2.85660207e-01
4.05454159e-01 1.17750955e+00 7.78539300e-01 6.76119089e-01
1.11445300e-01 8.86529922e-01 3.85440379e-01 5.06667912e-01
7.35850453e-01 1.18454605e-01 6.23714089e-01 4.58182633e-01
-3.34217340e-01 -3.49204808e-01 -1.17500830e+00 3.80464405e-01
-2.08998585e+00 -1.21491528e+00 -5.69916427e-01 2.39833856e+00
4.77920711e-01 3.05441529e-01 4.89371747e-01 2.68879414e-01
9.81754303e-01 -3.88431028e-02 -5.36188781e-01 -1.02676308e+00
-3.73439677e-02 -2.42626607e-01 8.14249575e-01 6.34212613e-01
-8.35575402e-01 5.95276654e-01 5.74260092e+00 8.91217172e-01
-8.41882825e-01 2.28206187e-01 6.72951639e-01 -4.31164235e-01
-2.16364115e-02 9.75686759e-02 -1.18159795e+00 9.24187958e-01
1.27358627e+00 -4.60273862e-01 4.34799701e-01 6.81780756e-01
8.40485871e-01 -5.83448708e-01 -9.01112080e-01 6.97909057e-01
-6.67930901e-01 -9.54498529e-01 -4.26002830e-01 4.33770686e-01
1.00713599e+00 -2.78758585e-01 1.69613585e-01 1.98744968e-01
4.33133304e-01 -7.31071651e-01 9.26726520e-01 8.29134047e-01
3.38653088e-01 -1.01575232e+00 7.92322516e-01 6.70189977e-01
-1.04850399e+00 -3.45309764e-01 -2.03875691e-01 -2.77182907e-01
8.11138570e-01 8.07787299e-01 2.59826332e-01 3.72267753e-01
3.74431878e-01 3.79310757e-01 1.17923856e-01 1.00417137e+00
-6.83168024e-02 3.67191106e-01 -8.63433659e-01 -6.34066284e-01
6.49051070e-01 -7.99278796e-01 1.09486842e+00 8.80048513e-01
2.28360817e-01 7.53065050e-02 2.15186045e-01 9.14392114e-01
8.68538439e-01 5.06077521e-02 -4.53054577e-01 -9.86225680e-02
-1.79194167e-01 7.94028759e-01 -1.04071212e+00 -2.31795609e-01
-1.78563043e-01 5.83200634e-01 -1.31442413e-01 1.64529532e-01
-7.03013122e-01 -2.96526849e-01 6.44315779e-01 7.63236135e-02
2.73536682e-01 -6.49116099e-01 -4.65674073e-01 -1.22190392e+00
9.08904895e-02 -3.78257483e-01 1.63476959e-01 -9.38459560e-02
-8.45063925e-01 1.86990932e-01 5.65487385e-01 -1.00926495e+00
-4.50812966e-01 -7.03816056e-01 -6.41031563e-01 9.14318979e-01
-9.60425377e-01 -4.46811736e-01 3.97254527e-01 4.05441672e-01
7.98133314e-01 1.87035754e-01 3.17402244e-01 -1.49560824e-01
-1.11621428e+00 -1.97721586e-01 9.81197655e-01 -7.01560915e-01
-2.19404802e-01 -1.58556426e+00 4.34512878e-03 9.29643929e-01
-5.51424742e-01 1.87159523e-01 1.47944307e+00 -4.35517550e-01
-1.57992268e+00 -5.89481771e-01 8.44282985e-01 -3.94470274e-01
1.08807039e+00 -4.79209423e-01 -6.31247282e-01 2.69872993e-01
-2.73119926e-01 -3.68090957e-01 1.10819958e-01 -4.14544046e-02
8.66141558e-01 -4.58467752e-02 -1.07519257e+00 6.64770961e-01
6.82813883e-01 1.50312604e-02 -1.94378912e-01 5.33797681e-01
4.12535965e-01 -2.10123911e-01 -8.80242169e-01 -9.74964425e-02
3.58296633e-01 -6.62242293e-01 6.43880188e-01 -4.83141959e-01
2.56678551e-01 -2.09972426e-01 1.68567702e-01 -1.27105737e+00
9.50180367e-02 -1.36004162e+00 -1.92824572e-01 1.12397814e+00
3.21870059e-01 -8.22023213e-01 6.47114515e-01 7.46162295e-01
1.23049848e-01 -1.13799739e+00 -1.21686924e+00 -1.19773543e+00
3.78040373e-01 -4.21927392e-01 2.04927847e-01 2.87603319e-01
3.30926938e-04 6.24174029e-02 -2.74969101e-01 -1.38371959e-01
7.87538767e-01 2.34819874e-01 3.78899455e-01 -1.19599450e+00
-6.75071776e-01 -7.07742929e-01 -1.27630174e-01 -6.60649300e-01
-8.83600265e-02 -4.65504259e-01 1.98771939e-01 -1.13820100e+00
-1.40912846e-01 -4.48683590e-01 -7.96301067e-02 -2.54946291e-01
-1.04381606e-01 -5.56340039e-01 4.43382055e-01 1.40610665e-01
-3.51097822e-01 6.98354542e-01 1.17587590e+00 1.84212193e-01
-6.18727267e-01 8.82008970e-01 -1.84886158e-01 8.19097400e-01
9.99027729e-01 -7.32440472e-01 -3.46350044e-01 -2.22886741e-01
4.87845719e-01 5.36315918e-01 4.63824049e-02 -8.98881674e-01
2.25461692e-01 -7.12886989e-01 -4.10719693e-01 -6.24623656e-01
3.11396211e-01 -6.03045762e-01 3.89280558e-01 8.58511925e-01
-2.89927453e-01 3.98285389e-01 1.08403221e-01 9.00150657e-01
-5.66638820e-02 -8.41990054e-01 1.07880068e+00 -4.13715780e-01
-1.95979923e-01 2.67434418e-01 -1.13426983e+00 5.16660176e-02
1.40613067e+00 -2.18250737e-01 5.49412727e-01 -3.17841083e-01
-1.36682880e+00 5.65155268e-01 4.09401029e-01 -3.40899155e-02
-1.96904149e-02 -7.47148275e-01 -4.59327936e-01 -6.00949526e-01
-6.55170441e-01 1.79916754e-01 4.10205662e-01 1.26602507e+00
-5.50294101e-01 6.40241325e-01 8.84576440e-02 -5.51432610e-01
-9.83096063e-01 8.33807230e-01 4.71727222e-01 -8.42685878e-01
-3.15111548e-01 6.94161296e-01 -7.44060129e-02 4.98401932e-02
-1.60082467e-02 -5.55881977e-01 -1.80856869e-01 7.35728800e-01
2.73505986e-01 9.82624650e-01 -3.32814902e-01 -4.73531693e-01
-2.00395301e-01 7.68507123e-01 1.99158967e-01 -7.12041378e-01
1.49333847e+00 -3.37060988e-01 -9.33924094e-02 9.65941787e-01
8.19038630e-01 -3.35639536e-01 -1.46812117e+00 -4.12531868e-02
3.21209759e-01 2.70479880e-02 2.02880293e-01 -3.00237268e-01
-8.54028583e-01 9.54191446e-01 4.02976364e-01 5.78826368e-01
9.12178874e-01 -3.15016091e-01 2.76294559e-01 3.23475420e-01
4.45239931e-01 -1.42374754e+00 -4.52505231e-01 5.78216434e-01
9.06187475e-01 -8.73496771e-01 1.61112566e-02 -4.32795346e-01
-6.44828975e-01 1.36108851e+00 1.97799981e-01 -3.48102540e-01
8.21851671e-01 4.65357631e-01 -5.85933268e-01 4.41132300e-02
-9.69885767e-01 -5.13961375e-01 -2.08714887e-01 3.07240695e-01
-1.27063110e-01 1.81100965e-01 -9.50641513e-01 4.36853528e-01
-1.66550849e-03 -1.76414058e-01 7.44758844e-01 1.03427076e+00
-3.01327318e-01 -1.09310830e+00 -5.63727677e-01 2.60929316e-01
-7.05568016e-01 8.75843316e-02 -8.68998542e-02 6.80908978e-01
-5.26055545e-02 8.49256754e-01 -8.75073746e-02 1.34008214e-01
5.16912580e-01 6.97735772e-02 4.23091918e-01 -6.34392440e-01
-7.20238864e-01 9.70880326e-04 2.84713544e-02 -5.45751214e-01
-1.96765915e-01 -1.31928074e+00 -6.92993343e-01 -2.65168667e-01
-4.21350420e-01 2.56598026e-01 1.06097507e+00 1.00101364e+00
-1.67970598e-01 3.47491324e-01 6.46297395e-01 -1.22632682e+00
-1.16746151e+00 -6.53183401e-01 -8.64093304e-01 -2.87682295e-01
2.52940774e-01 -6.95198178e-01 -4.34620470e-01 -3.44958544e-01] | [4.636254787445068, 2.725917339324951] |
0c5f4870-a606-4476-a4cc-4f6cd2e13718 | neural-sign-language-translation | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Camgoz_Neural_Sign_Language_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Camgoz_Neural_Sign_Language_CVPR_2018_paper.pdf | Neural Sign Language Translation | Sign Language Recognition (SLR) has been an active research field for the last two decades. However, most research to date has considered SLR as a naive gesture recognition problem. SLR seeks to recognize a sequence of continuous signs but neglects the underlying rich grammatical and linguistic structures of sign language that differ from spoken language. In contrast, we introduce the Sign Language Translation (SLT) problem. Here, the objective is to generate spoken language translations from sign language videos, taking into account the different word orders and grammar. We formalize SLT in the framework of Neural Machine Translation (NMT) for both end-to-end and pretrained settings (using expert knowledge). This allows us to jointly learn the spatial representations, the underlying language model, and the mapping between sign and spoken language. To evaluate the performance of Neural SLT, we collected the first publicly available Continuous SLT dataset, RWTH-PHOENIX-Weather 2014T. It provides spoken language translations and gloss level annotations for German Sign Language videos of weather broadcasts. Our dataset contains over .95M frames with >67K signs from a sign vocabulary of >1K and >99K words from a German vocabulary of >2.8K. We report quantitative and qualitative results for various SLT setups to underpin future research in this newly established field. The upper bound for translation performance is calculated at 19.26 BLEU-4, while our end-to-end frame-level and gloss-level tokenization networks were able to achieve 9.58 and 18.13 respectively. | ['Richard Bowden', 'Oscar Koller', 'Hermann Ney', 'Simon Hadfield', 'Necati Cihan Camgoz'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['sign-language-translation'] | ['computer-vision'] | [ 3.87535781e-01 -1.24795541e-01 -2.38539934e-01 -5.65320373e-01
-1.14938474e+00 -8.23782563e-01 9.45093751e-01 -8.27321231e-01
-8.80568981e-01 4.09584463e-01 7.42524385e-01 -2.62503922e-01
1.43914029e-01 -2.77051270e-01 -7.39743412e-01 -6.69309795e-01
-6.71783164e-02 3.38129371e-01 2.22933609e-02 -1.68154806e-01
-3.87502015e-02 4.42399889e-01 -1.39635396e+00 4.83695447e-01
5.46064436e-01 7.74669528e-01 7.91064054e-02 9.67821121e-01
-7.08045736e-02 7.58145332e-01 -6.12848878e-01 -3.23534131e-01
4.18243021e-01 -8.08029532e-01 -7.17484355e-01 -2.97429021e-02
1.00710881e+00 -5.52968979e-01 -6.44787908e-01 6.50844276e-01
8.27452362e-01 1.08349703e-01 5.75214386e-01 -1.04762936e+00
-7.47247994e-01 5.52307963e-01 -8.33371878e-02 -6.29663020e-02
3.39469582e-01 6.73598647e-01 1.12492025e+00 -8.62509906e-01
1.12396204e+00 1.23771691e+00 4.11018282e-01 1.02033448e+00
-5.88970959e-01 -6.83115065e-01 1.67718619e-01 1.09937891e-01
-1.28168845e+00 -4.87461746e-01 1.69160321e-01 -2.97559679e-01
1.22890460e+00 2.38774449e-01 8.03879559e-01 1.39711404e+00
-5.51338978e-02 1.05302823e+00 1.31452751e+00 -5.33921123e-01
1.16939366e-01 -6.05590045e-01 -4.19741943e-02 6.40423954e-01
-1.33234207e-02 3.76486361e-01 -9.77325082e-01 3.61038625e-01
7.97702610e-01 -3.68115962e-01 -1.92668140e-01 3.98003459e-02
-1.60223985e+00 5.48726916e-01 2.99359828e-01 2.18502328e-01
-2.53874928e-01 5.55184245e-01 3.82000625e-01 5.34914553e-01
8.32874924e-02 6.01311997e-02 -3.93735558e-01 -4.42554772e-01
-9.70774651e-01 2.24767491e-01 6.56208634e-01 1.07286382e+00
-5.50519396e-03 2.54402310e-01 -4.28762764e-01 6.49400532e-01
5.53537011e-01 1.26660347e+00 6.35334849e-01 -8.71061921e-01
6.59631610e-01 1.30652085e-01 -6.20852709e-02 -2.58041561e-01
-1.67475969e-01 -3.98303289e-03 -4.96670246e-01 2.92984784e-01
7.84257293e-01 -3.31891179e-01 -1.74613667e+00 1.80204892e+00
-1.11170888e-01 1.22152410e-01 2.99634844e-01 1.31961584e+00
1.01197839e+00 5.48875690e-01 1.53368279e-01 1.63423657e-01
1.37221622e+00 -9.63589072e-01 -6.26904547e-01 -2.13722140e-01
6.36307120e-01 -7.53687263e-01 1.28257108e+00 2.76879251e-01
-9.54509556e-01 -9.60857198e-02 -6.67886198e-01 -1.55822605e-01
-2.39953518e-01 3.79825890e-01 3.60415220e-01 4.99137402e-01
-1.24805260e+00 9.41378437e-03 -9.92906928e-01 -7.98388362e-01
2.43039846e-01 2.06958383e-01 -3.35500300e-01 -3.20112765e-01
-9.93227303e-01 1.06358564e+00 3.46915536e-02 3.55423510e-01
-8.00867260e-01 -4.04202551e-01 -6.89598262e-01 -4.45455551e-01
2.42014691e-01 -4.86331075e-01 1.49385285e+00 -1.03279662e+00
-1.97448885e+00 1.04068553e+00 -3.05801302e-01 -4.77399737e-01
8.89937222e-01 -3.50612491e-01 -4.24904406e-01 1.16566494e-01
-2.60755002e-01 9.00828958e-01 6.92144096e-01 -8.09544384e-01
-7.47880816e-01 -6.82584047e-02 -1.54678017e-01 4.41449344e-01
2.95574129e-01 4.92495209e-01 -5.29464602e-01 -8.17624271e-01
2.10559100e-01 -1.08708978e+00 1.58102050e-01 1.97781533e-01
2.93078786e-03 2.15449072e-02 4.85153258e-01 -1.14674938e+00
8.38186860e-01 -2.01377773e+00 3.04678261e-01 1.66780308e-01
-1.03981540e-01 3.43102276e-01 -6.57498360e-01 3.63280505e-01
3.50850284e-01 -4.43455167e-02 -4.29280758e-01 -3.18015426e-01
3.36551189e-01 5.63553333e-01 -6.01180732e-01 4.92575556e-01
1.15277216e-01 1.38884664e+00 -8.30338299e-01 -2.46552706e-01
2.24774316e-01 6.84485137e-01 -4.92424816e-01 5.12782335e-02
-3.45163643e-01 5.95582724e-01 -1.54775783e-01 8.66670012e-01
2.44290501e-01 1.86946496e-01 2.67333686e-01 -7.31986165e-02
-2.94613987e-01 4.26951081e-01 -8.62617493e-01 1.81621194e+00
-4.02602971e-01 9.97756779e-01 -8.12851191e-02 -5.64199269e-01
6.03207171e-01 4.16542679e-01 3.74593437e-01 -9.87932503e-01
1.47368744e-01 6.31176293e-01 1.35459229e-01 -7.43922055e-01
3.59026670e-01 -1.84634432e-01 -2.87909210e-01 5.11979580e-01
1.67253762e-01 -6.48156404e-02 2.46907473e-01 -1.86186612e-01
1.08359921e+00 3.86250049e-01 8.92941356e-02 2.11279914e-01
-1.51057839e-02 2.03264520e-01 1.89585358e-01 8.16287696e-01
-2.61050850e-01 7.25201964e-01 -1.63425859e-02 -4.02632236e-01
-9.49041724e-01 -1.28372443e+00 1.73219502e-01 1.18433285e+00
-1.85761139e-01 -9.79895592e-02 -5.88384688e-01 -5.51751256e-01
-2.72017777e-01 6.85950160e-01 -3.13795745e-01 1.89496428e-01
-1.17000842e+00 -4.60719436e-01 1.26603067e+00 6.67565703e-01
7.29492486e-01 -1.43387175e+00 -9.65282619e-01 -5.34800515e-02
-3.95343870e-01 -1.36833763e+00 -7.95956969e-01 -3.00786614e-01
-5.72718143e-01 -8.40459228e-01 -1.18765533e+00 -9.20097709e-01
5.12644291e-01 -1.68324828e-01 8.89856637e-01 -1.80282235e-01
-2.83690393e-01 6.10240400e-01 -7.60889232e-01 -2.25080684e-01
-4.52898026e-01 -6.02289550e-02 3.91956344e-02 -1.02699995e-01
5.42423904e-01 -1.82893366e-01 -4.23149496e-01 3.67532969e-01
-1.01328146e+00 7.95251355e-02 8.60928178e-01 7.94336677e-01
4.63194281e-01 -7.79280305e-01 -6.74630925e-02 1.00865290e-01
4.02591407e-01 2.20531344e-01 -6.00377858e-01 5.20599067e-01
-1.42015174e-01 2.66423047e-01 1.07518002e-01 -7.13578999e-01
-8.52960408e-01 1.66442230e-01 -1.40184626e-01 -1.05184197e-01
-2.92949736e-01 4.90054280e-01 -8.61297324e-02 8.37600790e-03
4.01060760e-01 7.11813450e-01 -6.60269111e-02 -3.22719097e-01
8.20732415e-01 8.36574554e-01 7.78950572e-01 -5.13494372e-01
6.98364198e-01 5.16522229e-01 -2.69340694e-01 -1.04660726e+00
-3.73346239e-01 -2.28834659e-01 -7.55822659e-01 -3.96098942e-01
9.55880940e-01 -8.04452062e-01 -5.93917549e-01 9.17510986e-01
-1.10374057e+00 -8.47457111e-01 -4.35966969e-01 8.18786025e-01
-6.78759754e-01 3.05834532e-01 -4.08707142e-01 -7.29552150e-01
-3.49859476e-01 -1.02073348e+00 1.22611642e+00 -1.16982438e-01
-4.01562363e-01 -5.97463489e-01 9.66126993e-02 2.87526786e-01
5.20002902e-01 2.12286144e-01 4.56307888e-01 -3.93270135e-01
-7.25548625e-01 -1.17501236e-01 -2.63351828e-01 4.45623219e-01
1.24558546e-01 -1.58579245e-01 -7.54566729e-01 -3.54906619e-01
-5.83394170e-01 -4.05254364e-01 1.01781690e+00 4.30575222e-01
2.08715677e-01 -4.29249614e-01 1.37348130e-01 6.69297099e-01
1.21150386e+00 1.77965507e-01 6.38509274e-01 1.80173650e-01
6.62160754e-01 4.59837735e-01 4.47874963e-01 1.77150339e-01
3.65804017e-01 8.51933002e-01 -4.47088256e-02 1.43966628e-02
-7.56514549e-01 -4.73784626e-01 8.96547139e-01 9.19666767e-01
-4.33226645e-01 -4.03991044e-01 -1.20210969e+00 6.01710081e-01
-1.66595423e+00 -9.08588111e-01 1.68884784e-01 2.08838201e+00
8.50092530e-01 -2.66704082e-01 1.54756159e-01 -1.66852087e-01
2.92702764e-01 2.26769730e-01 -5.58679461e-01 -2.73802221e-01
-5.15607357e-01 4.16900456e-01 8.45227420e-01 7.02861667e-01
-9.43801522e-01 1.45854092e+00 6.42020702e+00 4.77625757e-01
-1.58981097e+00 -7.94681460e-02 -1.64670259e-01 -4.75415140e-01
-7.57417232e-02 -2.11637288e-01 -8.04795086e-01 2.25075811e-01
9.62658107e-01 1.66281208e-01 8.25181007e-01 2.59279490e-01
4.84821886e-01 1.01402283e-01 -1.06601548e+00 1.05621505e+00
3.68781149e-01 -1.10047233e+00 3.58459085e-01 1.38966545e-01
6.66610539e-01 7.15279460e-01 1.19274832e-01 3.40764761e-01
5.86719811e-01 -1.29096437e+00 1.16369414e+00 6.69809759e-01
1.42390621e+00 -9.93740112e-02 6.22260451e-01 1.17192663e-01
-1.29501021e+00 1.16255105e-01 2.42897600e-01 -9.91828218e-02
5.99401057e-01 -1.93902329e-01 -7.21943021e-01 2.57789344e-01
5.74618638e-01 6.58554256e-01 -2.16372535e-01 9.35866714e-01
-6.20170951e-01 9.09165680e-01 -7.62147546e-01 -2.05212355e-01
6.36278510e-01 -1.69483609e-02 5.70261419e-01 1.53313649e+00
5.53376734e-01 1.68346390e-01 1.27397804e-02 4.05276120e-01
-1.36380494e-01 1.03843071e-01 -4.68105853e-01 -3.12253803e-01
1.73028812e-01 4.34481621e-01 -6.00377202e-01 -4.18195665e-01
-3.41268003e-01 1.30611229e+00 -2.45988414e-01 7.03504562e-01
-8.65171611e-01 -1.29633352e-01 8.68839920e-01 -1.14436910e-01
4.20003116e-01 -7.00724423e-01 -1.39394313e-01 -1.38811040e+00
4.25282568e-01 -1.02896488e+00 1.27532437e-01 -6.75857544e-01
-1.11847854e+00 4.90406811e-01 1.62463099e-01 -1.26706290e+00
-5.62254608e-01 -9.94703650e-01 -1.25782996e-01 8.72233510e-01
-1.62724113e+00 -1.55600929e+00 -3.00097734e-01 4.13183123e-01
6.96087539e-01 -1.39468580e-01 8.68715286e-01 3.10540676e-01
-5.04587479e-02 7.60317564e-01 2.11354017e-01 7.15955019e-01
6.83332562e-01 -9.47761893e-01 7.82028973e-01 9.94644463e-01
5.42098880e-01 4.98465717e-01 4.37333584e-01 -5.78995705e-01
-1.33265007e+00 -1.01407087e+00 1.38394105e+00 -6.00938022e-01
6.19915545e-01 -3.27326000e-01 -3.61452907e-01 7.73678660e-01
1.29060939e-01 3.93480584e-02 2.72094548e-01 -5.22830009e-01
-7.05264032e-01 2.58756816e-01 -8.72495830e-01 8.76346767e-01
1.61841738e+00 -7.69100010e-01 -6.81007624e-01 1.89311668e-01
4.38013017e-01 -7.92573929e-01 -5.65278172e-01 3.32392544e-01
1.26077294e+00 -3.83362532e-01 7.57471859e-01 -7.28154540e-01
1.95749074e-01 -3.49642247e-01 -6.10511661e-01 -1.08950293e+00
3.26611996e-01 -6.23128414e-01 8.75176042e-02 7.62282133e-01
4.45341587e-01 -5.68558335e-01 5.92798412e-01 5.52076280e-01
-1.34469345e-01 -4.42175865e-01 -1.31170809e+00 -1.08722949e+00
3.18541557e-01 -7.05710292e-01 4.31554794e-01 6.31122887e-01
-3.11756313e-01 -6.49855286e-02 -5.86103797e-01 6.86362609e-02
4.88472402e-01 5.44799790e-02 7.64722347e-01 -7.37763405e-01
-2.31487840e-01 -6.99358463e-01 -4.32103723e-01 -1.45590401e+00
1.50806382e-01 -1.05393100e+00 4.09110457e-01 -1.79144156e+00
-2.67372787e-01 1.34715527e-01 -4.51115053e-03 8.30326676e-01
4.98975098e-01 3.25949281e-01 4.55544859e-01 2.41680533e-01
-3.26222271e-01 4.15136635e-01 1.20088446e+00 -2.48962685e-01
-2.05592304e-01 -3.24992836e-01 -8.02532583e-02 5.71236968e-01
7.79812813e-01 -1.90857992e-01 -8.38090032e-02 -1.02538168e+00
-2.16385931e-01 -2.71232069e-01 6.01705074e-01 -7.19564080e-01
1.14071965e-01 -2.67108619e-01 -1.96033627e-01 -4.09645945e-01
3.15573514e-01 -6.28782630e-01 -1.82876468e-01 5.01339912e-01
-5.76444566e-01 -6.10155836e-02 3.84130925e-02 2.45806813e-01
-2.76985407e-01 3.60232890e-01 5.95708013e-01 -1.29310876e-01
-1.15389764e+00 1.22213483e-01 -5.23807466e-01 2.86970496e-01
5.75266898e-01 -3.76732051e-01 -2.15740234e-01 -5.37229419e-01
-7.03142464e-01 3.90953645e-02 1.81119278e-01 7.24132717e-01
8.20650101e-01 -1.32964909e+00 -1.10203063e+00 2.15000987e-01
3.01970720e-01 -1.66259840e-01 -1.00878216e-01 7.73870170e-01
-8.00848126e-01 6.81329250e-01 -2.13992640e-01 -6.09535515e-01
-1.44196069e+00 -3.53745013e-01 3.70159954e-01 5.10890000e-02
-9.94653881e-01 7.88869858e-01 -3.91001701e-01 -6.43155873e-01
5.00650585e-01 -1.03203237e+00 2.36806586e-01 -1.76831469e-01
4.82677549e-01 1.33221313e-01 -1.23176411e-01 -1.09424233e+00
-4.26356375e-01 8.80180001e-01 3.26545984e-01 -8.42787445e-01
1.18706262e+00 7.85561949e-02 2.77013868e-01 3.05415392e-01
1.05828607e+00 -1.60871476e-01 -1.23516881e+00 -3.78884941e-01
1.71395972e-01 -2.39380985e-01 -1.32547110e-01 -1.29003620e+00
-8.13401401e-01 7.77923465e-01 7.67786801e-01 -8.61864090e-01
9.72737312e-01 9.61204320e-02 1.00739872e+00 6.54898763e-01
4.26771313e-01 -1.02151358e+00 -1.49714947e-01 9.76439118e-01
1.09079647e+00 -1.12399316e+00 -4.71267790e-01 4.80811521e-02
-6.85870886e-01 9.28154111e-01 5.52564375e-02 -8.40775520e-02
2.35703662e-01 2.57707179e-01 7.10210443e-01 -2.51747873e-02
-3.37853491e-01 -4.46081877e-01 5.54028928e-01 6.42172813e-01
4.81077015e-01 2.52806962e-01 -4.05740649e-01 2.63850778e-01
-5.05980372e-01 3.38105619e-01 2.16155499e-01 1.08732152e+00
-2.69140393e-01 -1.02296197e+00 -2.86520481e-01 1.14182830e-01
-5.00323884e-02 -2.82519490e-01 -6.09207332e-01 8.81037116e-01
4.69259843e-02 6.90128624e-01 -1.39312148e-02 -3.30710948e-01
4.92438227e-01 4.02863324e-01 7.69454002e-01 -2.65815884e-01
-3.52408230e-01 -2.78457992e-05 2.00578138e-01 -6.19533002e-01
-7.88033664e-01 -8.96861434e-01 -1.44959664e+00 9.33007710e-03
1.74270198e-01 -3.68743509e-01 7.69296229e-01 1.21924889e+00
6.70110732e-02 2.68841892e-01 -1.54449865e-01 -7.17252195e-01
-6.79451168e-01 -1.02356768e+00 -2.92730868e-01 4.23090637e-01
3.86773169e-01 -3.35834205e-01 -2.44743973e-01 3.99833173e-01] | [9.194832801818848, -6.520578384399414] |
630f9dda-0caa-4f98-8503-ec8e5ece1f3d | identifying-consistent-statements-about | 1701.07696 | null | http://arxiv.org/abs/1701.07696v2 | http://arxiv.org/pdf/1701.07696v2.pdf | Identifying Consistent Statements about Numerical Data with Dispersion-Corrected Subgroup Discovery | Existing algorithms for subgroup discovery with numerical targets do not
optimize the error or target variable dispersion of the groups they find. This
often leads to unreliable or inconsistent statements about the data, rendering
practical applications, especially in scientific domains, futile. Therefore, we
here extend the optimistic estimator framework for optimal subgroup discovery
to a new class of objective functions: we show how tight estimators can be
computed efficiently for all functions that are determined by subgroup size
(non-decreasing dependence), the subgroup median value, and a dispersion
measure around the median (non-increasing dependence). In the important special
case when dispersion is measured using the average absolute deviation from the
median, this novel approach yields a linear time algorithm. Empirical
evaluation on a wide range of datasets shows that, when used within
branch-and-bound search, this approach is highly efficient and indeed discovers
subgroups with much smaller errors. | ['Luca M. Ghiringhelli', 'Mario Boley', 'Bryan R. Goldsmith', 'Jilles Vreeken'] | 2017-01-26 | null | null | null | null | ['subgroup-discovery'] | ['methodology'] | [ 2.69740224e-01 3.76992039e-02 -6.52936757e-01 -4.37954843e-01
-8.90012562e-01 -7.20069349e-01 5.11515319e-01 6.06740773e-01
-4.38508451e-01 1.21998191e+00 5.24198636e-02 -4.30403978e-01
-5.87052345e-01 -7.31087923e-01 -5.58020175e-01 -9.97282743e-01
-3.01820844e-01 7.55057096e-01 2.60890305e-01 1.79693967e-01
6.71254456e-01 4.18667972e-01 -1.54615700e+00 -1.67743221e-01
9.55167115e-01 1.04691184e+00 -3.29517752e-01 2.38892898e-01
2.51945227e-01 -7.15879127e-02 -5.97777665e-01 -2.30921343e-01
2.72427201e-01 -5.08025050e-01 -6.37837112e-01 2.19905339e-02
2.60050222e-02 -2.95832008e-02 5.97970724e-01 1.11438847e+00
3.04420501e-01 1.51476845e-01 7.81617582e-01 -1.10371923e+00
1.72059581e-01 8.54533851e-01 -8.76223087e-01 -1.51166124e-02
2.03285649e-01 -1.38777182e-01 1.06540334e+00 -4.06812429e-01
5.07768631e-01 1.22086823e+00 7.29046464e-01 -8.10742155e-02
-1.62630522e+00 -7.60648131e-01 1.03546478e-01 -1.53662413e-01
-1.61850011e+00 -4.36865002e-01 1.68441460e-01 -3.22782904e-01
4.36248094e-01 6.94404781e-01 3.22248995e-01 5.44531763e-01
2.36590967e-01 8.75901207e-02 1.15226686e+00 -2.85183996e-01
6.28924847e-01 2.49690905e-01 1.88009098e-01 3.58718365e-01
9.63185430e-01 1.30690262e-01 -3.88517737e-01 -8.43448937e-01
2.72587806e-01 -2.40274847e-01 -3.37677509e-01 -5.18285275e-01
-1.30721927e+00 1.08813810e+00 -1.71189278e-01 2.50247508e-01
-1.90653026e-01 1.36369690e-01 3.35240155e-01 3.00895333e-01
7.11171269e-01 5.19058704e-01 -4.89000618e-01 -1.60346925e-01
-1.16288435e+00 5.36490083e-01 1.03657210e+00 4.09313470e-01
4.65825915e-01 -4.45424557e-01 -1.89534336e-01 5.48775554e-01
1.68946311e-01 3.08350325e-01 1.53856918e-01 -9.92297828e-01
1.98400095e-01 3.72337818e-01 5.37139058e-01 -8.52525771e-01
-6.41603172e-01 -7.67751992e-01 -7.43983209e-01 1.21141553e-01
9.59472299e-01 -1.84712425e-01 -4.82539117e-01 1.86180866e+00
6.96985722e-01 -2.74979603e-02 -2.13644743e-01 5.64685047e-01
1.52250424e-01 3.76117676e-01 6.65151179e-02 -1.01529741e+00
1.22499192e+00 -1.43313020e-01 -5.10972083e-01 6.60838336e-02
7.43016839e-01 -6.53232515e-01 7.01726854e-01 5.96269488e-01
-1.02157724e+00 3.22983451e-02 -1.05815244e+00 3.97419900e-01
-2.77783703e-02 -2.00701743e-01 5.82824707e-01 7.71635592e-01
-7.13660359e-01 8.12125325e-01 -5.99305212e-01 -1.53871074e-01
3.75968277e-01 7.35792458e-01 -1.34709090e-01 1.69391379e-01
-7.32906640e-01 4.54635978e-01 3.91740441e-01 -7.35680014e-02
-3.15544367e-01 -9.78172004e-01 -3.41581106e-01 2.72664428e-01
6.41995430e-01 -7.56557465e-01 9.32126999e-01 -6.64439917e-01
-1.23407829e+00 6.68937743e-01 -4.90105748e-02 -4.69191134e-01
6.68741286e-01 2.95356989e-01 1.10106431e-01 -3.80601168e-01
8.51309821e-02 -1.41998532e-03 7.35886633e-01 -1.06754816e+00
-7.05154121e-01 -7.21711218e-01 -3.64416182e-01 5.48390746e-02
-1.71718270e-01 1.44151270e-01 1.73845887e-01 -3.52678627e-01
4.74087805e-01 -8.97662342e-01 -5.56953311e-01 -2.22106412e-01
-4.78791326e-01 -3.34359437e-01 3.50778788e-01 -2.60618299e-01
1.41806638e+00 -2.25721240e+00 -1.68975629e-02 7.57840931e-01
2.35261902e-01 -4.50887978e-01 4.33667272e-01 3.33537757e-01
-4.37336341e-02 2.59640217e-01 -4.01153564e-01 7.09557254e-03
5.58003038e-02 1.65801831e-02 -1.57000482e-01 1.11098409e+00
-4.06233758e-01 2.57238179e-01 -7.93527365e-01 -4.63513911e-01
-2.79910117e-01 -8.95701349e-02 -7.14444876e-01 -9.42504779e-02
-5.77320643e-02 2.48668268e-01 -5.98844290e-01 3.76578033e-01
7.24364281e-01 -3.96613806e-01 3.74733180e-01 1.44276336e-01
-3.31707656e-01 6.02964796e-02 -1.61757612e+00 8.10579658e-01
-5.64963222e-02 3.43990535e-01 2.49837503e-01 -1.19300556e+00
9.00047004e-01 1.28414929e-01 6.89614356e-01 2.53412761e-02
2.36193016e-01 6.12026453e-01 1.35534421e-01 4.34976555e-02
2.67670542e-01 -4.88520563e-01 -1.36343256e-01 5.93491018e-01
-4.77552682e-01 -1.87577885e-02 3.84931028e-01 -2.84926414e-01
1.11188185e+00 -5.26626348e-01 7.16775715e-01 -7.64035165e-01
3.51545602e-01 -1.94111928e-01 6.35521352e-01 9.67298448e-01
1.17476240e-01 4.78200376e-01 1.11788392e+00 -2.38275319e-01
-7.65212595e-01 -9.10601258e-01 -6.86429620e-01 9.76954579e-01
4.98544797e-02 -3.16424102e-01 -4.33655202e-01 -6.60391331e-01
3.90835136e-01 7.47414291e-01 -7.44716942e-01 -1.97159667e-02
-3.21187556e-01 -1.34777761e+00 1.10984266e-01 1.36664078e-01
6.69367388e-02 -4.28666115e-01 -5.72951078e-01 1.31533548e-01
1.56564891e-01 -6.84591472e-01 -4.29922968e-01 2.36770719e-01
-1.12174928e+00 -1.06330264e+00 -7.18198597e-01 -3.19211870e-01
6.55921400e-01 -9.86580476e-02 9.23889875e-01 -8.86787027e-02
7.32341707e-02 -1.62547722e-01 -6.74746558e-03 -4.83920127e-01
-2.04272449e-01 4.38716635e-02 3.27682830e-02 9.36649591e-02
1.26373187e-01 -6.34638071e-01 -4.73508507e-01 4.85444069e-01
-5.57156086e-01 -4.88555014e-01 3.08993310e-01 9.10094082e-01
6.26307607e-01 2.79070795e-01 7.26525247e-01 -9.56677794e-01
6.77737594e-01 -7.10090697e-01 -1.15114558e+00 1.68611631e-01
-1.02760613e+00 2.07839683e-01 2.73282379e-01 -3.89669389e-01
-6.73950016e-01 -6.26756772e-02 2.89939672e-01 1.36356935e-01
1.48204178e-01 5.89841187e-01 -2.20496953e-03 -1.80383921e-01
7.56931365e-01 -1.11802228e-01 1.45874202e-01 -4.44802195e-01
-8.03408101e-02 6.53061032e-01 2.17382610e-01 -4.59170580e-01
4.44950521e-01 4.89861697e-01 6.59857273e-01 -6.62287951e-01
-6.59370363e-01 -4.31245446e-01 -1.65810660e-01 1.07736126e-01
1.87129170e-01 -4.66305047e-01 -1.02185249e+00 -4.51129638e-02
-7.08152831e-01 -9.13047567e-02 -3.52284998e-01 7.68049598e-01
-6.90781415e-01 3.55619580e-01 9.44052041e-02 -9.83350575e-01
-2.52289385e-01 -1.21307290e+00 8.98186147e-01 -3.41453701e-02
-6.01945519e-01 -8.00076067e-01 -6.58368925e-03 4.62435000e-02
1.56418830e-01 6.07587755e-01 1.17280972e+00 -1.03816617e+00
-3.53665233e-01 -3.87778670e-01 -1.69240430e-01 -1.31147444e-01
2.31345147e-02 1.95958376e-01 -4.17526901e-01 -3.40511918e-01
7.63001153e-03 1.39663994e-01 6.26260340e-01 8.96235585e-01
1.44984794e+00 -4.99479532e-01 -6.95432305e-01 6.30420864e-01
1.20270324e+00 8.97696465e-02 1.40210971e-01 3.48260492e-01
4.68866490e-02 6.24806404e-01 7.62162089e-01 9.94570374e-01
3.38039286e-02 8.32459986e-01 2.23130271e-01 2.87334472e-01
5.18678904e-01 3.40149879e-01 3.70106138e-02 1.07185684e-01
-9.62876454e-02 -1.90356731e-01 -7.82615423e-01 5.48953712e-01
-1.89556766e+00 -7.60466754e-01 -2.09893495e-01 2.87709928e+00
1.15190232e+00 2.64691502e-01 4.71620083e-01 1.82462662e-01
6.81380510e-01 -1.16222380e-02 -4.35120314e-01 -5.03884971e-01
1.27724469e-01 5.76647110e-02 8.94515395e-01 5.50745726e-01
-7.06969261e-01 2.09827080e-01 6.96404934e+00 9.73208487e-01
-7.91834235e-01 -8.95314068e-02 7.83841729e-01 -1.49272934e-01
-3.31805646e-01 1.39190659e-01 -8.88787627e-01 6.64411068e-01
1.09255064e+00 -7.58516192e-01 1.89149886e-01 7.74444818e-01
3.94936562e-01 -4.35597897e-01 -1.28811908e+00 7.01593757e-01
-5.51530361e-01 -1.26242459e+00 -5.28715134e-01 4.84057397e-01
7.32881546e-01 -3.11578035e-01 1.29675707e-02 -1.60252839e-01
2.97809958e-01 -9.44625437e-01 4.48618233e-01 2.45393634e-01
6.71997905e-01 -1.15703154e+00 9.08115029e-01 3.61683726e-01
-5.91854274e-01 -2.07395777e-01 -3.01190645e-01 7.52182007e-02
-2.25249995e-02 1.20631051e+00 -9.50130641e-01 2.56153464e-01
5.74172735e-01 3.16269428e-01 -2.75615901e-01 1.41126883e+00
2.72850633e-01 6.50149941e-01 -8.55290294e-01 -2.04783991e-01
7.69120976e-02 -3.87074620e-01 8.47385705e-01 9.37950790e-01
5.16210496e-01 8.88281986e-02 -7.19050318e-02 5.49596608e-01
5.53367957e-02 3.71230125e-01 -3.62445295e-01 7.49565810e-02
6.37505412e-01 8.12528431e-01 -1.04902220e+00 -1.76895499e-01
1.44004757e-02 1.04517385e-01 -2.76089683e-02 1.96344897e-01
-5.80277562e-01 -4.78988737e-01 7.09118485e-01 2.93323576e-01
3.54106039e-01 -6.52311444e-02 -8.00699174e-01 -6.58814669e-01
1.61812957e-02 -1.03634191e+00 8.35604846e-01 6.20634072e-02
-1.04173028e+00 1.20868661e-01 5.44338346e-01 -9.03494120e-01
-6.22548759e-01 -3.34097594e-01 -3.15444946e-01 6.27728105e-01
-7.87306249e-01 -3.10188532e-01 1.60331592e-01 5.50881922e-02
1.08013995e-01 1.03872202e-01 5.83139122e-01 -3.66472192e-02
-3.92807126e-01 8.75615358e-01 6.01807952e-01 -6.62323415e-01
5.12252867e-01 -1.14674640e+00 -3.11775237e-01 4.08144921e-01
-2.76623785e-01 5.44407487e-01 1.40304828e+00 -6.56965554e-01
-1.13585579e+00 -6.38287544e-01 8.93304527e-01 -1.64431110e-01
7.55749881e-01 -1.97425663e-01 -7.35343277e-01 4.81026262e-01
-2.08242819e-01 4.46697697e-02 7.07254827e-01 6.46959960e-01
-2.58332328e-03 -3.74204487e-01 -1.41877699e+00 6.21609509e-01
8.04965258e-01 2.23392621e-01 -6.06171452e-02 6.53440177e-01
2.90847719e-01 -2.82123417e-01 -1.15360928e+00 6.82136834e-01
7.14585185e-01 -1.06984901e+00 7.79263794e-01 -4.45370376e-01
-6.40956759e-02 -2.25180179e-01 -1.22747004e-01 -1.14941764e+00
-8.55858400e-02 -9.30308282e-01 -6.34220392e-02 9.83820617e-01
5.35281360e-01 -9.98150468e-01 7.10994899e-01 5.23152947e-01
2.53139377e-01 -1.12990773e+00 -1.30037332e+00 -8.13354850e-01
7.26560354e-02 -1.31539196e-01 7.90421307e-01 6.88054681e-01
3.32739949e-02 -1.91262532e-02 7.66939949e-03 1.12895645e-01
1.05244434e+00 5.36742032e-01 6.08512461e-01 -1.49933553e+00
-3.57759029e-01 -7.78704464e-01 -4.56150025e-01 -5.80960512e-01
-1.91096067e-02 -5.11446178e-01 6.57036752e-02 -8.33931625e-01
3.02812546e-01 -6.69064999e-01 -1.13427825e-01 2.06094950e-01
-1.21313989e-01 6.31494867e-03 -4.74791706e-01 -3.30364518e-02
-3.53100389e-01 2.77445376e-01 8.32659185e-01 2.37812728e-01
-3.60512286e-01 5.64239681e-01 -8.65840673e-01 8.17844808e-01
7.71201074e-01 -9.14884090e-01 -2.50677466e-01 5.27410924e-01
3.61151755e-01 9.51714590e-02 4.83204238e-02 -6.88633442e-01
1.75251085e-02 -7.07945108e-01 2.93606222e-01 -6.67973459e-01
-6.15579896e-02 -5.50493658e-01 4.90486711e-01 5.45980394e-01
-4.96911556e-01 -9.30075347e-02 -8.31125155e-02 5.19085407e-01
7.54082799e-02 -4.84173238e-01 8.13085377e-01 2.21552551e-01
1.68464005e-01 3.36299874e-02 -1.89838082e-01 5.60605414e-02
1.17425787e+00 -1.43708333e-01 -1.65178001e-01 -5.02549350e-01
-5.83717227e-01 2.47449622e-01 5.04300237e-01 -1.84187502e-01
8.19863379e-02 -1.05031967e+00 -8.67780268e-01 -2.08922289e-02
1.85235702e-02 -1.86007097e-01 -7.62324706e-02 1.31195426e+00
-1.80827230e-01 3.53337228e-01 3.87812972e-01 -6.68894231e-01
-1.44333708e+00 3.76594335e-01 2.38313928e-01 -3.89878690e-01
-2.15298966e-01 8.28102231e-01 2.52007037e-01 5.11521548e-02
2.53945380e-01 -3.48844260e-01 6.33160025e-02 4.10210729e-01
5.40895939e-01 8.34809661e-01 4.95587811e-02 -2.55355984e-01
-4.51185942e-01 3.51243466e-01 3.08441259e-02 -1.01517111e-01
1.44015586e+00 -1.64386362e-01 -4.08348382e-01 4.74283159e-01
1.08855033e+00 2.26471052e-01 -1.00072944e+00 1.90208778e-02
2.47378632e-01 -5.77712595e-01 1.48594677e-01 -4.22692686e-01
-7.50940204e-01 2.24835888e-01 3.54221135e-01 6.88302159e-01
1.01567590e+00 1.11518905e-01 6.63333163e-02 2.08675951e-01
4.45397407e-01 -9.75024879e-01 -6.45820320e-01 1.01093844e-01
8.08800220e-01 -1.10470593e+00 7.17235565e-01 -5.04501283e-01
-1.93357870e-01 1.00226259e+00 1.08906485e-01 9.38223377e-02
8.03119719e-01 1.78306669e-01 -4.25046921e-01 9.06271022e-03
-8.11892927e-01 1.81872517e-01 2.96535373e-01 2.33288154e-01
3.31848711e-01 1.56940401e-01 -1.10685039e+00 5.58356583e-01
-3.54730904e-01 -2.05971152e-01 5.74441969e-01 5.96803665e-01
-5.46243608e-01 -1.08919311e+00 -5.42577922e-01 8.82075965e-01
-7.90509164e-01 1.96669534e-01 -1.20104901e-01 9.60308790e-01
-1.07375972e-01 9.20558929e-01 1.23537466e-01 1.38196766e-01
1.31277189e-01 -1.62322983e-01 3.70343775e-01 -3.97872657e-01
-4.62144971e-01 1.24579675e-01 3.55814219e-01 -5.98612964e-01
-1.79608405e-01 -1.10928190e+00 -9.89512444e-01 -5.72527945e-01
-5.50630629e-01 5.92992663e-01 5.96596301e-01 1.00189805e+00
2.36589208e-01 -6.15803190e-02 8.90206456e-01 -4.55993533e-01
-1.19300449e+00 -7.43757963e-01 -9.35923278e-01 -1.45357847e-01
4.49297577e-01 -8.60929012e-01 -8.17474425e-01 -4.31675851e-01] | [7.598612308502197, 4.6164984703063965] |
e204fb6b-8e68-4b31-9dbb-ac13851c2680 | eeg-based-emotion-recognition-using-genetic | null | null | https://ieeexplore.ieee.org/document/9588702/ | https://ieeexplore.ieee.org/document/9588702/ | EEG-Based Emotion Recognition Using Genetic Algorithm Optimized Multi-Layer Perceptron | Emotion Recognition is an important problem within Affective Computing and Human Computer Interaction. In recent years, various machine learning models have provided significant progress in the field of emotion recognition. This paper proposes a framework for EEG-based emotion recognition using Multi Layer Perceptron (MLP). Power Spectral Density features were used for quantifying the emotions in terms of valence-arousal scale and MLP is used for classification. Genetic algorithm is used to optimize the architecture of MLP. The proposed model identifies a. two classes of emotions viz. Low/High Valence with an average accuracy of 91.10% and Low/High Arousal with an average accuracy of 91.02%, b. four classes of emotions viz. High Valence-Low Arousal (HVLA), High Valence-High Arousal (HVHA), Low Valence-Low Arousal (LVLA) and Low Valence-High Arousal (HVHA) with 83.52% accuracy. The reported results are better compared to existing results in the literature. | ['Shyam Marjit'] | 2021-11-04 | null | null | null | 2021-international-symposium-of-asian-control | ['eeg', 'eeg-emotion-recognition', 'eeg'] | ['methodology', 'miscellaneous', 'time-series'] | [-6.47110939e-02 -7.62410164e-02 1.62184596e-01 -4.30963963e-01
-1.82369381e-01 -4.40565169e-01 2.77029932e-01 6.61747336e-01
-4.03433621e-01 1.05617356e+00 2.69973911e-02 2.89338857e-01
-2.97865212e-01 -5.69904566e-01 -2.00257838e-01 -8.33717644e-01
-1.87544554e-01 2.51929387e-02 -3.97614807e-01 1.12145543e-01
4.68435764e-01 5.06679952e-01 -1.84618998e+00 4.38824445e-01
4.45608467e-01 1.40291262e+00 -1.96284592e-01 7.57508278e-01
5.82282394e-02 6.53784633e-01 -7.19009638e-01 -5.90742007e-03
-1.07340194e-01 -4.41405952e-01 -6.65136278e-01 -3.20477635e-01
-5.16042888e-01 3.46235394e-01 4.62044656e-01 9.53332841e-01
5.71611285e-01 3.36650312e-01 8.66840720e-01 -1.47488081e+00
-3.24164987e-01 -5.87454848e-02 -6.97997451e-01 4.83101308e-01
4.55394536e-01 -3.83128583e-01 6.50450706e-01 -9.01567638e-01
1.28571123e-01 7.73735523e-01 5.71218789e-01 2.57168233e-01
-9.02904212e-01 -7.02827573e-01 -2.76924521e-01 5.40775418e-01
-1.66495991e+00 -6.76780269e-02 1.03993106e+00 -5.53754687e-01
1.42604184e+00 4.27932978e-01 1.05601597e+00 5.81148028e-01
7.89101839e-01 3.05659324e-01 1.79214525e+00 -4.03347939e-01
4.06803608e-01 8.40872645e-01 4.10076320e-01 3.20769846e-01
-2.89008045e-03 -1.69613034e-01 -4.96131152e-01 -7.62169808e-02
2.60782927e-01 -4.06715661e-01 9.82661694e-02 2.91412175e-01
-4.43321526e-01 7.90709496e-01 1.76214263e-01 6.69734716e-01
-1.14075339e+00 -1.98598295e-01 7.15403736e-01 1.33016661e-01
2.50207961e-01 4.64459687e-01 -4.66865718e-01 -4.96731132e-01
-5.99473000e-01 -6.32306114e-02 6.97327912e-01 4.34458941e-01
7.50908375e-01 2.06166953e-01 1.66078687e-01 1.09474790e+00
3.61143291e-01 1.66651711e-01 5.98255575e-01 -5.07475853e-01
-2.81228811e-01 7.65360236e-01 9.35241580e-02 -1.20033252e+00
-7.93654919e-01 -1.62431151e-01 -8.22691917e-01 3.33101153e-01
-4.04783994e-01 -5.32676995e-01 -4.84210789e-01 1.55985022e+00
7.31523558e-02 -8.44080225e-02 4.25897598e-01 8.17662477e-01
7.74541378e-01 1.22152519e+00 3.41595322e-01 -7.04392612e-01
1.56338274e+00 -4.39031392e-01 -9.57271695e-01 -1.14601664e-02
-2.49828324e-02 -6.90691650e-01 9.18593645e-01 6.72070682e-01
-1.05356526e+00 -4.89026338e-01 -1.14024377e+00 6.67366743e-01
-7.93655455e-01 1.95926577e-01 8.89554083e-01 1.04384029e+00
-1.04363906e+00 -3.95449065e-02 -4.51932967e-01 -4.96403188e-01
1.28828451e-01 6.61061347e-01 -3.63016427e-01 7.68908918e-01
-1.27937901e+00 9.46267426e-01 5.59912682e-01 1.52894214e-01
-3.58961999e-01 -2.80554861e-01 -4.80558097e-01 1.35786012e-01
-4.95946497e-01 -2.44580340e-02 7.68467784e-01 -1.57897019e+00
-1.75789380e+00 8.14947009e-01 -1.38144076e-01 -3.29137743e-02
-4.35278893e-01 1.25202835e-01 -8.76100659e-01 2.68732071e-01
-4.02024508e-01 5.51945269e-01 2.68470347e-01 -1.19379389e+00
-5.87451458e-01 -5.56922197e-01 -4.32989180e-01 5.05125523e-01
-5.09995997e-01 3.02252561e-01 2.69130796e-01 -1.62887663e-01
6.45599440e-02 -6.97912216e-01 1.78077117e-01 -7.12943137e-01
-6.06921427e-02 -4.88106579e-01 6.84411705e-01 -5.69928646e-01
1.21902215e+00 -1.89917994e+00 -1.63137615e-01 5.49177706e-01
-3.11156839e-01 2.32079968e-01 4.78083700e-01 4.62491959e-01
-3.97036880e-01 -1.92414030e-01 -5.29004708e-02 3.67364168e-01
-9.90625247e-02 1.81773081e-01 1.41882673e-01 2.79197425e-01
3.69700640e-01 3.42837542e-01 -5.42106748e-01 -4.01902199e-01
4.81295228e-01 7.90615201e-01 -3.43118489e-01 3.92929047e-01
3.86187673e-01 2.65830427e-01 -3.55165690e-01 8.45804155e-01
4.34023172e-01 3.52770180e-01 7.83410221e-02 -3.77574176e-01
-5.11125803e-01 -3.19076508e-01 -1.21254492e+00 7.66999364e-01
-5.89298546e-01 8.28738987e-01 3.34331281e-02 -1.05713952e+00
1.43861783e+00 7.35445738e-01 6.17917895e-01 -6.75787270e-01
6.74962997e-01 8.98495689e-02 1.01704262e-01 -9.13491786e-01
5.56315839e-01 -5.38572729e-01 -1.11378290e-01 3.26507986e-01
2.31418207e-01 9.28467214e-02 4.74836007e-02 -4.16943550e-01
7.52720416e-01 -1.54243097e-01 5.96117854e-01 -3.16137433e-01
7.71028638e-01 -3.62707525e-01 4.48358744e-01 1.29331946e-01
-5.72078168e-01 1.34624252e-02 6.12257421e-01 -2.71288186e-01
-7.98589885e-01 -8.40372026e-01 -3.61047804e-01 9.84374046e-01
3.47636230e-02 5.72935082e-02 -6.78884923e-01 2.48568594e-01
-5.74346900e-01 8.57144177e-01 -5.46366811e-01 -3.19948852e-01
6.85847402e-02 -1.19658887e+00 5.32239974e-01 4.99555081e-01
5.67055762e-01 -1.52815056e+00 -1.10327923e+00 7.15426952e-02
2.02065498e-01 -6.65116072e-01 5.76429665e-01 5.81865311e-01
-6.58255696e-01 -4.83003944e-01 -2.25075006e-01 -7.48468161e-01
5.57787836e-01 -6.65677905e-01 9.69711661e-01 -5.00612259e-01
-6.77928329e-01 3.58245164e-01 -4.98046875e-01 -9.77804065e-01
-1.58733372e-02 -5.09661913e-01 1.60697624e-01 2.47605294e-01
7.00970173e-01 -7.04279721e-01 -5.45285285e-01 -1.25640109e-01
-5.58489561e-01 -4.12153065e-01 8.66180420e-01 4.85015243e-01
7.38919795e-01 3.66063029e-01 1.05623424e+00 -4.97967213e-01
1.07365274e+00 -7.81185269e-01 -3.03763479e-01 9.51285809e-02
-7.28267789e-01 -3.91960800e-01 7.12405205e-01 -3.68735462e-01
-1.03441286e+00 1.03025220e-01 -4.77621794e-01 -3.85050550e-02
-6.03281677e-01 5.33427060e-01 -1.55813806e-02 -2.08063662e-01
4.33634967e-01 3.13041210e-01 -4.79582191e-01 2.75473371e-02
-1.48007721e-01 1.07100284e+00 2.75778413e-01 -2.21398249e-01
-3.22436929e-01 -1.97679788e-01 -5.48599176e-02 -1.09648454e+00
-2.79001355e-01 -3.96237522e-01 -3.19007576e-01 -7.18115628e-01
1.13906693e+00 -6.88169837e-01 -9.13235843e-01 3.77679259e-01
-7.88579822e-01 3.67069215e-01 1.95183843e-01 7.43654191e-01
-5.37719786e-01 -2.52129734e-01 -5.04607797e-01 -1.54485357e+00
-9.57572520e-01 -9.19495761e-01 3.36618721e-01 7.75556624e-01
-7.13012397e-01 -8.80045176e-01 8.16352740e-02 -2.18646992e-02
3.82542014e-01 6.70891643e-01 9.03827488e-01 -6.88000083e-01
4.72723097e-01 -5.34696639e-01 3.42188030e-03 4.27151591e-01
-2.78434288e-02 3.67365405e-02 -1.09598339e+00 2.21610263e-01
1.04109235e-01 -5.16696572e-01 1.60775688e-02 4.20222551e-01
1.04900432e+00 -3.21892887e-01 6.70220777e-02 2.15272740e-01
1.69300818e+00 1.21146917e+00 7.28208184e-01 2.15264529e-01
4.15681936e-02 2.74597913e-01 5.79842567e-01 6.70817792e-01
1.70183167e-01 2.00063214e-01 4.42184627e-01 1.64821044e-01
5.93939900e-01 2.40654066e-01 2.20430166e-01 9.46360171e-01
-2.26257518e-02 -1.99915290e-01 -9.01495934e-01 4.85763013e-01
-1.43576682e+00 -9.74419534e-01 -2.32494920e-01 1.89032364e+00
7.50354290e-01 2.42257677e-02 1.92217335e-01 7.22515762e-01
7.00330853e-01 -1.41978070e-01 -2.58676261e-01 -1.68325508e+00
1.20824073e-02 5.44494152e-01 2.63583791e-02 2.49538869e-01
-1.09370053e+00 5.13681114e-01 5.71247435e+00 2.03044057e-01
-1.33620858e+00 -1.60677820e-01 6.38296008e-01 -1.13446951e-01
1.97777957e-01 -3.92677665e-01 -2.93230265e-01 6.02901876e-01
1.31919837e+00 -1.29645020e-01 4.45509315e-01 8.47654164e-01
1.89921230e-01 -6.14656985e-01 -5.29029787e-01 1.29118097e+00
2.29518458e-01 -6.26932502e-01 -2.42331639e-01 -3.37763041e-01
6.08536541e-01 -3.22961926e-01 -1.12330765e-01 3.46847683e-01
-5.24460495e-01 -9.64586079e-01 3.46531153e-01 8.79250050e-01
4.50059444e-01 -1.41827226e+00 1.08220255e+00 1.28157511e-01
-1.08304870e+00 -1.55646592e-01 -3.73886108e-01 -3.51546824e-01
-6.41238019e-02 4.92290914e-01 -6.28791094e-01 2.83143789e-01
8.51738036e-01 3.32606435e-01 -1.29799515e-01 9.00063217e-01
-1.42820245e-02 6.43794239e-01 -3.75390559e-01 -6.63892388e-01
2.37222597e-01 -4.00988162e-01 2.64938891e-01 1.38529789e+00
3.34876448e-01 6.16982222e-01 -2.67185062e-01 4.56084162e-01
1.58585206e-01 5.86911619e-01 -4.10021931e-01 -2.69575953e-01
3.68000954e-01 1.72991073e+00 -9.06042576e-01 -3.13842118e-01
-1.43449986e-02 7.85411239e-01 -8.76426101e-02 -3.53753157e-02
-8.39776695e-01 -9.42412734e-01 3.22471172e-01 -5.16541779e-01
-3.42056125e-01 3.93350601e-01 -5.22791147e-01 -4.27682102e-01
-3.16179454e-01 -4.32072818e-01 4.28371906e-01 -9.90955412e-01
-9.52649474e-01 9.30226922e-01 -1.32725462e-01 -7.66190708e-01
-1.17550068e-01 -5.91048121e-01 -8.26026797e-01 1.19173086e+00
-9.49533999e-01 -6.44230127e-01 -3.80276829e-01 4.43005741e-01
4.28121418e-01 -2.60676295e-01 1.32792699e+00 3.56835663e-01
-5.02477705e-01 3.32944751e-01 -1.44059554e-01 -2.52326816e-01
3.53701860e-01 -1.18015265e+00 -9.92498040e-01 3.80500525e-01
-4.09114003e-01 3.61205816e-01 7.79322565e-01 -1.32444784e-01
-1.35658610e+00 -7.37208724e-01 9.63861585e-01 1.23019651e-01
2.76217341e-01 -5.88603094e-02 -7.99570084e-01 2.12911889e-01
5.93059897e-01 -3.24297816e-01 1.29224980e+00 -9.09559522e-03
3.04038376e-01 -2.17204213e-01 -1.57405996e+00 2.37839013e-01
-2.80483395e-01 -4.60121304e-01 -4.16729987e-01 2.59605423e-02
-2.53926069e-01 2.44504306e-03 -1.23270273e+00 2.24529460e-01
8.79953325e-01 -1.25819385e+00 4.69191551e-01 -1.34083167e-01
4.29159611e-01 -6.48850501e-02 -1.99212283e-01 -1.14485419e+00
-2.92675495e-01 -2.52792872e-02 4.99332175e-02 1.23087585e+00
4.20877755e-01 -5.60982823e-01 4.47915971e-01 5.40254593e-01
-1.78001940e-01 -1.13655806e+00 -8.35090280e-01 -1.72941685e-01
-2.90860921e-01 -3.17187428e-01 7.89816529e-02 1.00959563e+00
5.53108335e-01 5.84056795e-01 -2.26516649e-01 1.47270281e-02
1.49175718e-01 8.04482624e-02 1.73546700e-03 -1.27502024e+00
7.38207176e-02 -4.43484902e-01 -8.47066045e-01 3.42301726e-01
7.94392973e-02 -6.58717990e-01 -2.83001419e-02 -1.60398805e+00
3.51365775e-01 -6.92916065e-02 -8.32381368e-01 5.18138766e-01
1.90627337e-01 6.04514897e-01 -2.20938027e-01 -3.14308196e-01
-2.68580049e-01 2.43275210e-01 3.47128004e-01 3.71162653e-01
-4.44698274e-01 -6.90158829e-02 -5.69330573e-01 8.41843486e-01
1.38157260e+00 -1.80017576e-01 -4.64487851e-01 1.81421325e-01
3.69239390e-01 1.86494038e-01 -2.00771600e-01 -1.18885612e+00
-6.23586997e-02 9.96323419e-04 8.90423179e-01 -5.93994498e-01
6.75794601e-01 -8.26234400e-01 3.07671934e-01 4.33304757e-01
-3.45710099e-01 2.78603494e-01 4.23759937e-01 1.01755559e-01
-3.14454645e-01 -2.59496182e-01 1.05555999e+00 3.58757600e-02
-8.74598026e-01 -3.63956213e-01 -8.89586806e-01 -5.36004961e-01
1.52011144e+00 -4.49071676e-01 2.68630274e-02 -3.24802071e-01
-9.16006386e-01 -7.40176663e-02 -1.44888654e-01 4.14313912e-01
7.72609055e-01 -1.26518476e+00 -2.87462741e-01 3.07706892e-01
1.42342672e-01 -1.07335460e+00 3.18111986e-01 1.02990377e+00
-3.74226987e-01 4.46182370e-01 -1.01986432e+00 -2.55155146e-01
-1.46401525e+00 1.16883419e-01 2.83397526e-01 6.04355708e-02
2.75299083e-02 8.54946613e-01 -3.32353085e-01 1.45994425e-01
1.48244739e-01 3.88663888e-01 -1.05904460e+00 4.43757892e-01
4.38956827e-01 6.74231529e-01 3.19103375e-02 -1.10440314e+00
-5.96965909e-01 5.16121805e-01 3.99180412e-01 -9.29277018e-02
1.32951367e+00 4.01011519e-02 -4.60359752e-01 1.03190601e+00
1.36921942e+00 -3.74305159e-01 -4.24122065e-01 6.88892007e-01
8.39866027e-02 2.11288799e-02 2.87633866e-01 -1.28929830e+00
-5.56200206e-01 1.09955537e+00 1.18872786e+00 3.02468896e-01
1.78864145e+00 -4.29143488e-01 3.16790938e-01 1.02254957e-01
1.11573033e-01 -1.71246958e+00 -8.37322921e-02 4.54796195e-01
7.24357426e-01 -8.68970394e-01 -1.00674957e-01 4.44880873e-02
-1.11638784e+00 1.23270154e+00 5.34129977e-01 -2.40810469e-01
7.40766585e-01 3.37261438e-01 2.29362734e-02 -2.75637358e-01
-9.33089375e-01 3.55173484e-04 3.71801704e-01 3.42402697e-01
9.11569536e-01 3.08562785e-01 -9.05536771e-01 1.06162250e+00
-2.14389980e-01 -1.16630318e-02 4.69771802e-01 9.09830034e-01
-6.64869606e-01 -5.60108662e-01 -4.78921324e-01 7.58290529e-01
-8.21879148e-01 2.11664826e-01 -5.93599617e-01 4.80450153e-01
4.09748256e-01 1.23027956e+00 3.78905714e-01 -5.77111483e-01
1.92841083e-01 5.90810180e-01 3.71438742e-01 -2.17598423e-01
-9.46206987e-01 2.03938574e-01 3.79088684e-03 -2.27991462e-01
-4.64661986e-01 -2.86061019e-01 -1.61641371e+00 1.15690805e-01
3.17845978e-02 6.72848940e-01 1.29024494e+00 6.64751410e-01
2.49386087e-01 5.84543765e-01 7.43536711e-01 -6.67889833e-01
1.83305845e-01 -1.05284083e+00 -8.36358249e-01 1.40633017e-01
1.09891277e-02 -6.46013737e-01 -3.64892989e-01 6.99676797e-02] | [13.300179481506348, 3.275223970413208] |
8936be60-f2e0-4959-9c56-ba6531fff391 | comprehensive-dataset-of-face-manipulations | 2208.11776 | null | https://arxiv.org/abs/2208.11776v2 | https://arxiv.org/pdf/2208.11776v2.pdf | Comprehensive Dataset of Face Manipulations for Development and Evaluation of Forensic Tools | Digital media (e.g., photographs, video) can be easily created, edited, and shared. Tools for editing digital media are capable of doing so while also maintaining a high degree of photo-realism. While many types of edits to digital media are generally benign, others can also be applied for malicious purposes. State-of-the-art face editing tools and software can, for example, artificially make a person appear to be smiling at an inopportune time, or depict authority figures as frail and tired in order to discredit individuals. Given the increasing ease of editing digital media and the potential risks from misuse, a substantial amount of effort has gone into media forensics. To this end, we created a challenge dataset of edited facial images to assist the research community in developing novel approaches to address and classify the authenticity of digital media. Our dataset includes edits applied to controlled, portrait-style frontal face images and full-scene in-the-wild images that may include multiple (i.e., more than one) face per image. The goals of our dataset is to address the following challenge questions: (1) Can we determine the authenticity of a given image (edit detection)? (2) If an image has been edited, can we \textit{localize} the edit region? (3) If an image has been edited, can we deduce (classify) what edit type was performed? The majority of research in image forensics generally attempts to answer item (1), detection. To the best of our knowledge, there are no formal datasets specifically curated to evaluate items (2) and (3), localization and classification, respectively. Our hope is that our prepared evaluation protocol will assist researchers in improving the state-of-the-art in image forensics as they pertain to these challenges. | ['Kirill Trapeznikov', 'Brian DeCann'] | 2022-08-24 | null | null | null | null | ['image-forensics'] | ['computer-vision'] | [ 4.00922000e-01 2.91349608e-02 1.58437952e-01 -1.69540927e-01
-4.85476106e-01 -8.98718476e-01 5.07772982e-01 -4.12227921e-02
-1.76014483e-01 5.64840376e-01 -2.72613823e-01 -3.30528319e-01
1.70606166e-01 -7.03329980e-01 -6.97756469e-01 -3.77561241e-01
1.22027509e-01 5.90836592e-02 1.79343313e-01 1.10491529e-01
5.05601585e-01 8.51098120e-01 -1.58641076e+00 2.02909693e-01
5.14125586e-01 8.94655645e-01 -1.69003397e-01 6.19290948e-01
-3.43912952e-02 7.72644699e-01 -9.68926609e-01 -1.23972440e+00
3.94294828e-01 -3.95885438e-01 -8.34862173e-01 4.22397554e-01
7.73184419e-01 -8.18299472e-01 -3.87809366e-01 1.25964963e+00
3.46872449e-01 -2.39189461e-01 3.54272097e-01 -1.48698640e+00
-8.28580379e-01 2.06835777e-01 -7.31386125e-01 4.09489006e-01
6.03063643e-01 4.32504416e-01 2.18812540e-01 -7.11727679e-01
9.10241783e-01 1.17040968e+00 5.79639792e-01 4.16284919e-01
-9.79580522e-01 -1.08628678e+00 -3.79016459e-01 8.77355784e-02
-1.50828803e+00 -7.37032533e-01 7.86419928e-01 -5.88424444e-01
2.36655444e-01 3.43623608e-01 4.10189599e-01 1.29618490e+00
2.63251215e-02 2.50330746e-01 1.26276660e+00 -3.81923527e-01
-5.12517951e-02 4.52434570e-01 -3.40020955e-01 6.46443307e-01
3.49599868e-01 -2.28368267e-02 -7.01188803e-01 -3.43858957e-01
5.87235272e-01 -1.18582703e-01 -1.47958770e-01 2.68923104e-01
-7.17020810e-01 6.84760749e-01 -2.82161981e-01 1.78542718e-01
-3.11699718e-01 1.63918771e-02 4.81126398e-01 2.60972351e-01
4.13712680e-01 3.78751516e-01 1.54773727e-01 -4.35687989e-01
-1.27014685e+00 1.73141912e-01 6.35635555e-01 7.15748310e-01
5.77278316e-01 4.59450409e-02 2.63534278e-01 5.39546728e-01
-4.59504612e-02 4.54651088e-01 8.48582610e-02 -1.03578508e+00
3.18217158e-01 3.27663630e-01 5.85243478e-02 -1.45886040e+00
3.00455302e-01 1.88224956e-01 -3.67650777e-01 5.77188194e-01
5.45035481e-01 -6.29145978e-03 -5.11873662e-01 1.45722175e+00
5.36978364e-01 2.36408189e-01 -4.05899853e-01 7.65356660e-01
6.85251117e-01 2.78623790e-01 1.57413241e-02 -2.06921980e-01
1.58640850e+00 -1.27414748e-01 -8.17790270e-01 2.58450042e-02
2.86892980e-01 -1.34950900e+00 1.01022601e+00 5.39488614e-01
-1.20395923e+00 -3.81015956e-01 -9.75417852e-01 8.72169249e-03
-5.48452675e-01 -8.48243982e-02 3.02196026e-01 1.27175784e+00
-8.95840645e-01 5.72795868e-01 -4.52805609e-01 -5.03463387e-01
7.07750916e-01 1.79951549e-01 -7.53600061e-01 -1.48286089e-01
-9.59979653e-01 7.78589964e-01 -5.58540374e-02 -1.76312238e-01
-8.51053298e-01 -7.38552332e-01 -5.05264103e-01 -3.94009978e-01
4.84902442e-01 -1.42349347e-01 8.30854595e-01 -1.26623011e+00
-1.01261020e+00 1.52040815e+00 -1.64426584e-02 -2.86004424e-01
8.14823985e-01 2.34341342e-02 -8.91855776e-01 7.33208060e-01
3.37959290e-01 5.20330191e-01 1.24266362e+00 -1.28097260e+00
-3.80701035e-01 -5.58270812e-01 2.83562273e-01 -3.25480878e-01
-6.07510328e-01 7.45527565e-01 -3.00036013e-01 -8.78410459e-01
-3.43574554e-01 -8.76418650e-01 6.20439529e-01 6.28317297e-01
-5.71073055e-01 4.41192746e-01 1.45569694e+00 -1.16903961e+00
1.16562927e+00 -2.42296958e+00 -5.46642065e-01 -2.79531572e-02
1.96029902e-01 6.45674467e-01 1.60234138e-01 4.23620015e-01
-8.08572862e-03 7.68516839e-01 -8.00115690e-02 -3.10764700e-01
-2.74970651e-01 -2.19163243e-02 -3.98884237e-01 8.77939939e-01
1.48960263e-01 3.24021071e-01 -8.00393224e-01 -7.87989259e-01
1.56090185e-01 6.28124774e-01 -2.34681353e-01 9.08110291e-02
2.58105814e-01 3.25684667e-01 -2.45775193e-01 1.07374823e+00
1.04534721e+00 2.26162016e-01 -2.45439466e-02 -3.31222624e-01
-1.65235952e-01 -6.26434088e-02 -1.20712948e+00 8.91898096e-01
-1.06813721e-01 1.15311611e+00 3.76930445e-01 -3.57752413e-01
6.14288509e-01 2.73971200e-01 2.69466698e-01 -4.55611557e-01
4.49860655e-02 6.05899766e-02 -2.61592060e-01 -7.33906031e-01
7.88846135e-01 -5.67946024e-02 2.30609208e-01 8.10208797e-01
-2.46296301e-01 1.66687034e-02 1.23736270e-01 2.11307019e-01
1.15937781e+00 6.31288141e-02 6.21886514e-02 1.89429522e-01
2.76087970e-01 -2.10865036e-01 1.62750378e-01 5.14091194e-01
-6.23857915e-01 8.00656796e-01 5.87741733e-01 -3.15304756e-01
-1.09291458e+00 -1.00851822e+00 -2.05086380e-01 8.39877248e-01
1.47309542e-01 -3.54185402e-01 -1.10117996e+00 -6.33920252e-01
6.47340417e-02 6.08100057e-01 -5.06930649e-01 -2.02438328e-02
-3.84969145e-01 -2.74203569e-01 1.48780191e+00 7.32894018e-02
6.82079673e-01 -9.46950793e-01 -8.26247752e-01 -1.61280721e-01
-2.93599367e-01 -1.40143573e+00 -6.77092731e-01 -6.68832242e-01
-4.05093282e-01 -1.52287865e+00 -3.60479236e-01 -3.02359223e-01
8.67516577e-01 4.34845895e-01 7.06397891e-01 5.51407099e-01
-4.56306428e-01 6.94592357e-01 -3.99449587e-01 -3.59177023e-01
-6.38369441e-01 -5.86033463e-01 8.32825899e-02 4.67033774e-01
3.13084781e-01 -4.90930527e-01 -4.69582438e-01 5.29259622e-01
-1.43687963e+00 -4.39871788e-01 1.33657202e-01 2.96219230e-01
3.01776588e-01 2.86154062e-01 2.48872861e-01 -9.80013072e-01
7.56829202e-01 -5.08331597e-01 -2.99387813e-01 2.56919175e-01
-4.31934655e-01 -6.01342142e-01 4.40038264e-01 -6.51505232e-01
-9.09839749e-01 -3.92140508e-01 7.28066564e-02 -7.60880411e-01
-3.51470798e-01 1.80288926e-02 -1.29880905e-01 -5.43883085e-01
5.30595720e-01 4.06782590e-02 2.11602673e-01 -2.72648364e-01
1.07223988e-01 9.38814998e-01 7.18237519e-01 -4.91013616e-01
1.05166614e+00 8.75606358e-01 -3.48258577e-02 -9.73844588e-01
-3.82865161e-01 -1.73341125e-01 -5.48982024e-01 -7.36481965e-01
6.90476000e-01 -6.07193768e-01 -7.83683479e-01 7.57842779e-01
-1.08020139e+00 9.36265290e-02 6.02548718e-02 8.83341115e-03
-2.19518587e-01 9.47818220e-01 -4.76017892e-01 -1.08691549e+00
-3.25983874e-02 -1.03366983e+00 8.34280908e-01 1.77888319e-01
-4.16397065e-01 -7.19570220e-01 -5.16046882e-01 8.12962711e-01
3.72604132e-01 7.73184478e-01 5.53318739e-01 -4.45647031e-01
-4.25821871e-01 -4.64346588e-01 -2.39750266e-01 3.77386212e-01
1.82643786e-01 6.68333113e-01 -1.16947317e+00 -1.65037736e-01
1.20135710e-01 -3.05865973e-01 1.34435683e-01 -6.01058677e-02
1.21461642e+00 -6.03187859e-01 -1.61288828e-02 4.58346695e-01
1.10503685e+00 2.64099985e-01 1.16001880e+00 4.61647242e-01
4.32839692e-01 7.01130092e-01 8.45888495e-01 6.00551188e-01
9.96114910e-02 7.30115533e-01 6.21248782e-01 1.85085550e-01
-3.63695621e-01 -4.03011531e-01 5.42529106e-01 1.43017098e-01
-6.30744174e-02 -2.17813641e-01 -6.57955408e-01 3.37367505e-01
-1.19945526e+00 -1.39666617e+00 -2.86067784e-01 2.35167837e+00
6.13847256e-01 4.03077938e-02 1.46914259e-01 2.42439866e-01
1.20246172e+00 2.50860393e-01 -3.53078723e-01 -5.24042726e-01
-1.33841410e-01 1.08414009e-01 4.16622609e-01 1.36092873e-02
-1.02554548e+00 7.68022478e-01 5.95645809e+00 9.37508941e-01
-1.36731696e+00 3.18576187e-01 7.89699793e-01 -1.73533410e-01
-8.05604234e-02 2.60786474e-01 -6.84913218e-01 9.20421839e-01
7.67229736e-01 1.05619483e-01 4.90317971e-01 6.61547661e-01
3.30294043e-01 -2.87420273e-01 -7.74385631e-01 1.13389134e+00
5.44018507e-01 -1.25897825e+00 -1.01922780e-01 4.19610113e-01
3.47987473e-01 -7.03681350e-01 3.80745530e-01 -1.09131500e-01
-1.50370836e-01 -1.07455540e+00 1.03611958e+00 4.08130974e-01
1.14390314e+00 -7.50454426e-01 5.07219613e-01 1.05779409e-01
-8.93223643e-01 9.54761803e-02 -2.79673994e-01 9.39055756e-02
2.72532731e-01 4.83779639e-01 -6.95737123e-01 3.06626618e-01
7.58414090e-01 2.74118215e-01 -8.08283806e-01 9.24759805e-01
-1.75992429e-01 5.88859320e-01 -1.98954031e-01 3.96455646e-01
-2.82334816e-02 -2.13782310e-01 7.58406639e-01 1.01231134e+00
4.38410968e-01 3.79208773e-02 -3.50836962e-01 1.03998911e+00
-4.20167923e-01 -4.03755046e-02 -7.27285683e-01 -5.11696517e-01
8.14231575e-01 1.23122549e+00 -8.74370635e-01 5.56474738e-02
-2.38100454e-01 1.16071224e+00 -6.69789314e-02 1.48084566e-01
-9.55208361e-01 -3.41289490e-01 6.37234271e-01 7.80778408e-01
1.33664355e-01 -8.29527974e-02 -1.71729535e-01 -8.52880359e-01
3.11825186e-01 -1.01990068e+00 3.19661379e-01 -1.06548846e+00
-1.16296208e+00 4.42097574e-01 -4.47115786e-02 -1.18570364e+00
2.05649454e-02 -3.51631284e-01 -6.29180729e-01 5.04176497e-01
-1.12987447e+00 -1.31901097e+00 -4.22177553e-01 5.66528916e-01
2.68984646e-01 -2.00812295e-01 5.69925070e-01 4.89804000e-01
-3.99903208e-01 7.75839448e-01 -2.39452228e-01 4.13513899e-01
1.09539545e+00 -5.35985172e-01 -3.38005796e-02 1.13344681e+00
-6.87920302e-02 7.58774459e-01 7.93484330e-01 -8.56597722e-01
-1.44659138e+00 -9.93856847e-01 5.12285113e-01 -6.08700514e-01
6.54075086e-01 -1.62753552e-01 -7.01071084e-01 5.69425225e-01
3.76050249e-02 -1.93051435e-02 7.97429025e-01 -4.04758006e-01
-5.64303875e-01 1.27482206e-01 -1.80443335e+00 4.59176898e-01
7.83563733e-01 -9.06828880e-01 -3.51356298e-01 3.15851510e-01
1.56647980e-01 -2.65501022e-01 -9.43316698e-01 -4.22261702e-03
7.18314230e-01 -1.23319876e+00 9.83547688e-01 -4.13219601e-01
6.69523537e-01 -3.00791293e-01 -9.40008312e-02 -6.56797945e-01
4.20938224e-01 -8.94573450e-01 -1.92024305e-01 1.93411303e+00
-1.47684142e-01 -5.40603638e-01 5.77614129e-01 9.53957438e-01
1.80160150e-01 -3.95876646e-01 -1.02289486e+00 -8.07748377e-01
-2.63921201e-01 -4.02174860e-01 6.79002464e-01 1.30541050e+00
-3.32189262e-01 -5.13647914e-01 -5.83003402e-01 2.93529361e-01
6.89020395e-01 -4.02752697e-01 1.03690958e+00 -9.82645273e-01
-2.94024702e-02 -3.00362378e-01 -8.08761895e-01 -2.46673897e-01
2.78540011e-02 -4.14133847e-01 -6.75055206e-01 -8.33713531e-01
3.07480901e-01 -2.12284625e-01 5.18168569e-01 3.88731509e-01
9.75363255e-02 7.47217953e-01 4.98769403e-01 5.59991300e-01
-1.93078771e-01 -1.76692039e-01 9.72530603e-01 1.02331460e-01
3.36842567e-01 -1.72085464e-01 -8.49113286e-01 6.21829212e-01
5.81932545e-01 -6.78403437e-01 -1.50189921e-01 -1.08767547e-01
3.19371045e-01 -9.79928523e-02 7.62552202e-01 -9.04106915e-01
1.60605595e-01 -2.74591327e-01 3.16156596e-01 -3.65932167e-01
5.03128111e-01 -7.67098784e-01 5.54157853e-01 1.52778625e-01
9.96233523e-02 1.29310295e-01 1.24666244e-01 4.03349966e-01
-2.50208348e-01 -5.60242116e-01 1.02780032e+00 -2.83387870e-01
-4.47539359e-01 1.69268608e-01 -3.76012266e-01 1.58136457e-01
1.38952363e+00 -8.38043392e-01 -6.55035198e-01 -5.96071482e-01
-3.85760814e-01 -3.55576366e-01 1.11056554e+00 2.43770316e-01
6.02461159e-01 -1.08678341e+00 -4.00034636e-01 3.96962576e-02
-6.11636788e-02 -5.62912703e-01 2.42866114e-01 7.31180072e-01
-8.91934156e-01 -2.01682001e-01 -4.23650026e-01 -1.71666563e-01
-1.66915107e+00 4.81357843e-01 1.43602878e-01 2.75716931e-01
-4.00343895e-01 5.59465826e-01 -1.83140203e-01 -4.23918627e-02
-3.39759439e-02 5.27371883e-01 1.75931245e-01 3.46945584e-01
9.40622389e-01 6.86791599e-01 1.76451786e-03 -1.07995093e+00
-3.27082098e-01 2.73227066e-01 -5.93335330e-02 -5.46625466e-04
1.19427252e+00 -1.75507158e-01 -1.74104497e-01 -4.41255867e-02
1.15544963e+00 4.68239993e-01 -1.05350578e+00 3.85491103e-01
-3.72074932e-01 -1.18762171e+00 -1.49431974e-01 -5.92677474e-01
-1.22832727e+00 8.01223576e-01 6.42478645e-01 4.57898319e-01
1.04698193e+00 -2.26278573e-01 9.05141592e-01 -3.09579760e-01
4.64918315e-01 -1.28250468e+00 2.35141173e-01 -5.97888790e-02
6.66644812e-01 -1.07282388e+00 2.51215637e-01 -4.89806533e-01
-6.59968257e-01 1.28471148e+00 3.68163019e-01 1.99360356e-01
5.09586990e-01 4.05041933e-01 -9.18053612e-02 -3.28990430e-01
-5.23648672e-02 2.16085061e-01 -1.80499092e-01 8.71872902e-01
1.87468663e-01 -1.01870753e-01 -1.88704908e-01 2.87840664e-01
-3.49658281e-01 7.58514628e-02 9.72779214e-01 1.07943308e+00
-4.31739204e-02 -9.87415433e-01 -1.00433433e+00 4.23874408e-01
-1.02862644e+00 1.33218259e-01 -8.77244890e-01 9.50117409e-01
3.59757900e-01 1.24432051e+00 -4.84166555e-02 -3.09231490e-01
-3.80582623e-02 -3.28388698e-02 3.09123486e-01 -1.85821787e-01
-7.14650989e-01 -4.32731748e-01 2.53361136e-01 -3.97536427e-01
-4.69710112e-01 -9.04305816e-01 -8.52391958e-01 -9.61395264e-01
-2.45473936e-01 -3.38483125e-01 9.34114635e-01 8.22258115e-01
4.09166813e-01 -6.58375174e-02 4.28794861e-01 -7.25774825e-01
-2.79048651e-01 -5.96118629e-01 -6.27924562e-01 7.48686492e-01
9.12956521e-03 -8.24208975e-01 -6.19624972e-01 4.85546529e-01] | [12.524069786071777, 1.0587319135665894] |
c2ca5e44-82b5-4974-a216-11683042675f | nonconvex-matrix-factorization-from-rank-one | 1802.06286 | null | http://arxiv.org/abs/1802.06286v2 | http://arxiv.org/pdf/1802.06286v2.pdf | Nonconvex Matrix Factorization from Rank-One Measurements | We consider the problem of recovering low-rank matrices from random rank-one
measurements, which spans numerous applications including covariance sketching,
phase retrieval, quantum state tomography, and learning shallow polynomial
neural networks, among others. Our approach is to directly estimate the
low-rank factor by minimizing a nonconvex quadratic loss function via vanilla
gradient descent, following a tailored spectral initialization. When the true
rank is small, this algorithm is guaranteed to converge to the ground truth (up
to global ambiguity) with near-optimal sample complexity and computational
complexity. To the best of our knowledge, this is the first guarantee that
achieves near-optimality in both metrics. In particular, the key enabler of
near-optimal computational guarantees is an implicit regularization phenomenon:
without explicit regularization, both spectral initialization and the gradient
descent iterates automatically stay within a region incoherent with the
measurement vectors. This feature allows one to employ much more aggressive
step sizes compared with the ones suggested in prior literature, without the
need of sample splitting. | ['Yuxin Chen', 'Yuanxin Li', 'Yuejie Chi', 'Cong Ma'] | 2018-02-17 | null | null | null | null | ['quantum-state-tomography'] | ['medical'] | [ 3.69846314e-01 2.49594584e-01 -1.41709492e-01 -7.82285444e-03
-1.15142035e+00 -6.74244165e-01 4.40085709e-01 2.03928594e-02
-4.92352307e-01 9.11057830e-01 3.59364264e-02 -3.07841212e-01
-4.75575715e-01 -4.31931376e-01 -6.73697174e-01 -9.69639778e-01
-1.60399780e-01 6.34368002e-01 -2.70037234e-01 -9.19982269e-02
4.07591432e-01 4.97440249e-01 -8.78458798e-01 -5.44794321e-01
8.61881852e-01 8.96670580e-01 -7.17060491e-02 4.59940970e-01
2.31916174e-01 4.35698301e-01 -7.23151937e-02 -3.42120409e-01
4.61101234e-01 -2.15802625e-01 -6.43028498e-01 -4.03202027e-02
4.07534540e-01 -1.33623093e-01 -5.51330864e-01 1.42555213e+00
4.19389933e-01 2.04604521e-01 4.98339653e-01 -6.94930792e-01
-3.24005932e-01 7.43371785e-01 -4.35377538e-01 -1.56393826e-01
1.13493107e-01 -6.04638420e-02 1.27450693e+00 -9.44443047e-01
5.80214798e-01 6.73783660e-01 6.37191832e-01 2.82644540e-01
-1.76532984e+00 -5.87448597e-01 -1.98407724e-01 7.66941309e-02
-1.60760164e+00 -6.50896847e-01 9.90549922e-01 -5.00139356e-01
3.58912051e-01 1.12602159e-01 3.78498971e-01 8.39813411e-01
-3.07507306e-01 3.49591166e-01 1.09561396e+00 -5.63163280e-01
3.09601277e-01 1.93545729e-01 2.13579208e-01 9.74726140e-01
4.70511675e-01 1.32667795e-01 -6.66802406e-01 -4.96280462e-01
5.58601975e-01 -3.19001228e-01 -7.19661415e-01 -1.00570130e+00
-1.42573929e+00 7.64156818e-01 3.08013529e-01 3.61490279e-01
-2.13484660e-01 3.67577851e-01 9.74322632e-02 3.37176204e-01
2.40957767e-01 6.34000659e-01 -3.85634959e-01 5.95622994e-02
-9.78400528e-01 1.26842216e-01 9.94894505e-01 5.60673356e-01
1.07167923e+00 2.56559312e-01 2.10328043e-01 3.51970077e-01
7.64115825e-02 9.85185921e-01 -2.98930574e-02 -8.64492595e-01
6.26816809e-01 -5.35337254e-02 4.73413497e-01 -8.39566052e-01
-3.88667852e-01 -9.45650458e-01 -1.05766010e+00 2.11247448e-02
6.67572379e-01 -3.33544105e-01 -5.30868828e-01 2.00294375e+00
2.34560877e-01 2.97480524e-01 7.71243945e-02 1.09843469e+00
1.09878413e-01 4.57319051e-01 -5.57232559e-01 -4.68987375e-01
9.11113858e-01 -4.20251966e-01 -5.31221986e-01 -3.96165967e-01
5.99118888e-01 -6.50980413e-01 8.73517096e-01 5.28396368e-01
-9.32572722e-01 1.47711020e-02 -1.35281253e+00 1.15298048e-01
2.14833423e-01 3.13552856e-01 7.90120661e-01 7.72807837e-01
-7.88760066e-01 9.29024756e-01 -6.30480766e-01 2.53493279e-01
3.23514529e-02 5.13358891e-01 -5.74615300e-01 6.14227653e-02
-8.85422349e-01 7.29540408e-01 3.58314365e-01 4.55003351e-01
-6.93234622e-01 -7.88317859e-01 -5.87924182e-01 1.03981294e-01
7.02294588e-01 -5.60274899e-01 1.04077196e+00 -6.99234247e-01
-1.79807925e+00 6.64154530e-01 -3.82597506e-01 -5.28608739e-01
3.29936326e-01 -2.49507204e-01 5.17869145e-02 1.26529753e-01
2.22039502e-02 -2.33076826e-01 1.05552495e+00 -1.09267724e+00
8.01592618e-02 -4.71521795e-01 3.40129882e-02 4.01301980e-02
-1.66216224e-01 -5.95869064e-01 -2.84849286e-01 -1.94715023e-01
6.76755130e-01 -1.16948557e+00 -4.86389160e-01 -2.64101565e-01
-4.56100166e-01 1.99054360e-01 1.39873996e-01 -5.32161534e-01
9.28927302e-01 -2.17208290e+00 6.36652112e-01 6.81721389e-01
4.93564785e-01 -3.56212072e-03 -2.45846715e-02 4.81439054e-01
-4.17558514e-02 -1.87812924e-01 -5.01593351e-01 -2.98985004e-01
1.61481068e-01 -1.98324159e-01 -5.24986207e-01 1.03840482e+00
1.91446654e-02 7.60045230e-01 -9.94166136e-01 9.68682989e-02
1.95618853e-01 2.44936720e-01 -7.29861438e-01 -1.43422201e-01
-9.13876668e-03 6.69977367e-01 -5.07754683e-01 2.02338502e-01
5.66969633e-01 -5.47452271e-01 5.37836194e-01 -5.53254008e-01
-1.88002095e-01 3.27522725e-01 -1.63381398e+00 1.89216399e+00
-4.90426749e-01 6.65687799e-01 6.65406466e-01 -1.27526343e+00
5.31172693e-01 2.49927521e-01 5.70900142e-01 -4.62720424e-01
1.01262011e-01 7.14048624e-01 6.57432154e-02 -5.12635633e-02
5.16863942e-01 -4.01163310e-01 -1.39300212e-01 3.78870368e-01
3.40987407e-02 -1.12328544e-01 -6.02375809e-03 2.61187762e-01
7.71381080e-01 -2.18006909e-01 3.27616483e-01 -5.04973292e-01
6.19173586e-01 -3.49618167e-01 5.65184474e-01 9.17268217e-01
2.24837467e-01 4.14754033e-01 5.65388143e-01 -5.57428934e-02
-1.10110211e+00 -1.11474395e+00 -1.33463383e-01 5.80511808e-01
2.09088326e-01 -3.58649552e-01 -4.24004108e-01 -1.13506414e-01
-6.45391941e-02 3.76664370e-01 -3.36917341e-01 -3.29676345e-02
-6.54848039e-01 -6.55487955e-01 2.26358056e-01 -1.78358510e-01
2.78634876e-01 -3.27434897e-01 5.94120882e-02 1.98380232e-01
-2.47005261e-02 -1.35742748e+00 -4.67386514e-01 2.90486872e-01
-9.31222022e-01 -1.06734538e+00 -6.16361916e-01 -3.79666328e-01
8.22860241e-01 1.80788815e-01 7.14586794e-01 -3.12637866e-01
-7.50460774e-02 2.58242369e-01 1.35162890e-01 1.99939564e-01
-1.28936067e-01 1.37365118e-01 2.69513756e-01 2.71677375e-01
-2.42684543e-01 -7.73775458e-01 -4.72926378e-01 -1.22488782e-01
-6.85824513e-01 -7.32324421e-02 5.88928938e-01 1.01483893e+00
4.93027657e-01 -1.27756566e-01 9.89482105e-02 -8.63414705e-01
4.80521053e-01 -1.59497157e-01 -1.23659706e+00 6.17240034e-02
-7.08282769e-01 7.36426353e-01 8.81540179e-01 -2.53937751e-01
-4.98513520e-01 1.53863177e-01 1.86704725e-01 -2.45737925e-01
4.45227414e-01 6.30827904e-01 5.98391630e-02 -6.43699765e-01
7.19894648e-01 3.75420213e-01 -1.63757861e-01 -5.06349862e-01
6.95049763e-01 2.16765240e-01 6.71409369e-01 -8.09749484e-01
1.35964072e+00 6.05865479e-01 7.23712444e-01 -8.97219479e-01
-1.03222585e+00 -5.14163971e-01 -4.60863560e-01 1.08138941e-01
2.95501411e-01 -8.10118139e-01 -1.06753945e+00 2.08416685e-01
-8.50288033e-01 -1.89823091e-01 -2.55784303e-01 8.69351685e-01
-5.44369161e-01 8.35693836e-01 -4.46081936e-01 -9.67248857e-01
-3.63343030e-01 -9.98416126e-01 8.90333772e-01 -5.05920537e-02
2.60717124e-01 -8.63401353e-01 6.97210431e-02 2.88320422e-01
3.00208956e-01 1.34735584e-01 6.42515123e-01 -2.54641771e-01
-9.23386455e-01 -2.84373909e-01 -2.35317707e-01 2.07294837e-01
-1.96443707e-01 -5.56149960e-01 -8.54904532e-01 -5.77832580e-01
1.71463579e-01 -2.20688328e-01 9.90722060e-01 1.95329994e-01
9.08345699e-01 -3.19505364e-01 -2.75050923e-02 1.03586817e+00
1.50369465e+00 -3.61225486e-01 2.34331772e-01 -7.32995942e-02
7.63191342e-01 3.76029551e-01 1.97727889e-01 5.40988922e-01
3.07129398e-02 6.97976172e-01 3.41833234e-01 2.53182679e-01
2.45825872e-01 -2.22261205e-01 2.20122829e-01 1.02700341e+00
1.27300224e-03 2.26868704e-01 -7.20818877e-01 3.78486454e-01
-1.79795754e+00 -9.24295783e-01 -1.55328169e-01 2.87641072e+00
8.25471342e-01 7.09053278e-02 -3.38578492e-01 8.44611451e-02
3.54321688e-01 4.80377167e-01 -5.93343735e-01 9.43469757e-04
-1.27745062e-01 5.05797148e-01 9.58107173e-01 1.00797355e+00
-1.07170177e+00 7.88130939e-01 5.87131786e+00 8.11505377e-01
-1.27392006e+00 1.94447696e-01 1.37164176e-01 -1.68445259e-01
-4.83713180e-01 5.08563817e-01 -5.55322289e-01 2.22783104e-01
7.81073749e-01 -1.83042571e-01 9.84787643e-01 5.14237225e-01
1.31040022e-01 9.78109241e-02 -9.50459421e-01 1.28398919e+00
-1.90974563e-01 -1.47755897e+00 -3.67404610e-01 2.06313923e-01
8.24868560e-01 1.40508905e-01 2.21392903e-02 1.00475758e-01
9.57510341e-03 -8.03996801e-01 6.25115812e-01 5.64051449e-01
7.97323644e-01 -7.48047352e-01 3.71759564e-01 2.63714731e-01
-9.52599049e-01 -3.30728143e-02 -2.97734976e-01 -3.51117626e-02
4.09014493e-01 1.16521811e+00 -6.25852048e-01 5.97854972e-01
-1.29940227e-01 4.05165195e-01 1.60563469e-01 9.82213140e-01
-2.46273950e-01 7.09247947e-01 -6.83049679e-01 -1.03103772e-01
2.63198912e-01 -9.56923783e-01 1.01714027e+00 8.49261999e-01
4.45286244e-01 6.16160817e-02 2.59025991e-01 8.27721417e-01
-2.87307650e-01 1.02961116e-01 -4.27515626e-01 -3.77464533e-01
3.94753635e-01 1.23056269e+00 -3.55556667e-01 -2.79727089e-03
-2.02658698e-01 9.03230369e-01 3.60111833e-01 6.61880314e-01
-4.30387318e-01 -4.36584204e-01 5.16724646e-01 7.04664830e-03
3.91315579e-01 -7.11199999e-01 -3.44970614e-01 -1.55411935e+00
4.29210186e-01 -7.94431746e-01 -7.42557272e-02 -1.68569401e-01
-1.02561033e+00 1.28561005e-01 -3.11798126e-01 -9.61763740e-01
-4.13729995e-01 -6.80908978e-01 -2.33175129e-01 8.93981159e-01
-1.49512208e+00 -6.85107708e-01 1.34829968e-01 2.16423988e-01
-2.52686143e-01 -1.23699583e-01 7.52684057e-01 3.93380553e-01
-4.15209830e-01 5.81821322e-01 6.02452993e-01 6.80024102e-02
4.34008092e-01 -1.37085807e+00 -5.09990640e-02 1.06548476e+00
3.30569685e-01 9.04730499e-01 9.84561265e-01 -3.68470311e-01
-2.01102591e+00 -5.39267540e-01 6.94191992e-01 1.82797927e-02
1.31265140e+00 -4.97507125e-01 -6.26839697e-01 3.74962986e-01
-3.15912545e-01 1.02798074e-01 4.14150298e-01 4.60626483e-01
-4.88344491e-01 -3.41859013e-01 -8.53091538e-01 5.50373495e-01
7.93156326e-01 -9.72134113e-01 -7.98655078e-02 6.25071228e-01
4.77045208e-01 -5.45160949e-01 -8.04085195e-01 3.99518371e-01
4.95450139e-01 -6.15879357e-01 1.06452096e+00 -4.02405232e-01
-1.76810935e-01 -5.20848930e-01 -3.44577909e-01 -1.04280269e+00
-2.15476528e-01 -1.44066000e+00 -4.71505910e-01 6.04647815e-01
4.90506411e-01 -8.67419064e-01 9.91362751e-01 3.51069659e-01
-6.64865691e-03 -5.57095647e-01 -1.14517641e+00 -7.31792212e-01
-4.51639481e-02 -3.75353754e-01 1.58663720e-01 8.83618236e-01
-1.73638582e-01 5.23559093e-01 -7.66595483e-01 5.68028033e-01
1.30615842e+00 3.86460990e-01 7.64840484e-01 -1.23126245e+00
-8.88351083e-01 -4.30254161e-01 -3.17202866e-01 -1.43312192e+00
2.46909380e-01 -1.11167479e+00 3.85443643e-02 -1.01185250e+00
2.02280760e-01 -6.14840209e-01 -2.22482517e-01 9.19847563e-03
-6.63782842e-03 1.69447958e-01 1.83932334e-01 1.52335197e-01
-4.02765840e-01 6.48517966e-01 1.04908502e+00 -8.83276612e-02
-2.05378100e-01 3.12506258e-01 -6.52266145e-01 5.65942109e-01
7.81707585e-01 -5.73160410e-01 -3.18433404e-01 -2.46853992e-01
7.07264543e-01 3.98827016e-01 4.12882864e-01 -9.23191905e-01
3.89178991e-01 -4.22033072e-02 -2.43743151e-01 -2.73629606e-01
5.37691712e-01 -5.04991055e-01 2.77231276e-01 4.21869010e-01
-2.87065119e-01 -5.45295477e-01 -1.47089899e-01 7.54880965e-01
-2.58082766e-02 -6.78867042e-01 8.25584292e-01 1.63018987e-01
-2.94670016e-01 4.43797946e-01 -1.51197359e-01 1.38868451e-01
3.30820024e-01 2.08316669e-01 -7.04149455e-02 -5.17801940e-01
-7.56780088e-01 -1.01472445e-01 3.65026861e-01 -2.31148392e-01
2.73921609e-01 -1.24607944e+00 -5.89991152e-01 7.73247257e-02
-5.49251325e-02 -2.38411516e-01 9.38526168e-02 1.00758398e+00
-3.44884306e-01 6.89001262e-01 5.84699810e-01 -5.13633549e-01
-8.61989141e-01 2.50474095e-01 5.20992041e-01 -4.38552350e-01
-5.74314833e-01 6.76430166e-01 1.00895893e-02 -5.37514210e-01
1.18059404e-01 -1.68982133e-01 3.55135232e-01 -2.00017616e-01
3.71043801e-01 3.41260612e-01 9.99799669e-02 -3.95577878e-01
-1.17865622e-01 6.96569324e-01 9.65010226e-02 -5.19976795e-01
1.16227758e+00 7.97599107e-02 -3.57640892e-01 3.87967318e-01
1.42299712e+00 6.56834185e-01 -1.20467913e+00 -6.81790948e-01
2.68663675e-01 -2.27770299e-01 2.71139234e-01 -2.24706918e-01
-8.53038371e-01 7.12939680e-01 4.58321959e-01 7.92201012e-02
8.09399128e-01 -1.80774227e-01 7.20245063e-01 1.12480783e+00
6.84535563e-01 -9.48688567e-01 -1.01782389e-01 5.97115338e-01
5.37490487e-01 -1.01847517e+00 2.15603843e-01 -2.76920974e-01
8.80861282e-03 1.08329308e+00 -1.48697227e-01 -3.90741825e-01
4.31799084e-01 -3.04042529e-02 -3.41811031e-01 -6.16915151e-02
-3.24890167e-01 -2.66343236e-01 5.33559561e-01 2.39265725e-01
1.87161580e-01 2.71112204e-01 -3.61142159e-01 1.02735892e-01
-2.09457725e-01 -3.19275737e-01 5.41425347e-01 4.85876203e-01
-5.66081524e-01 -1.35654426e+00 -2.14523047e-01 4.16098177e-01
-4.34159845e-01 -3.30243915e-01 -7.28983134e-02 5.35101712e-01
-4.49763745e-01 7.17820227e-01 -5.09040952e-01 2.79635657e-02
6.92386180e-02 -5.22654913e-02 8.69998455e-01 -4.37348992e-01
-8.22261572e-02 4.12573963e-02 9.27252769e-02 -6.64795697e-01
-2.91455060e-01 -5.98687828e-01 -1.13657665e+00 -3.35750908e-01
-4.79510903e-01 5.05669415e-01 8.27968776e-01 9.24719691e-01
1.76771849e-01 -5.24416491e-02 7.31418610e-01 -7.60685563e-01
-1.12880802e+00 -6.62857294e-01 -5.94626844e-01 1.60897344e-01
6.89974785e-01 -3.53142858e-01 -7.05239534e-01 -5.16310036e-01] | [6.545301914215088, 4.62223482131958] |
afebb6f5-3341-429d-95fc-721a64dd1299 | optimizing-human-interpretable-dialog | 1605.03915 | null | http://arxiv.org/abs/1605.03915v2 | http://arxiv.org/pdf/1605.03915v2.pdf | Optimizing human-interpretable dialog management policy using Genetic Algorithm | Automatic optimization of spoken dialog management policies that are robust
to environmental noise has long been the goal for both academia and industry.
Approaches based on reinforcement learning have been proved to be effective.
However, the numerical representation of dialog policy is
human-incomprehensible and difficult for dialog system designers to verify or
modify, which limits its practical application. In this paper we propose a
novel framework for optimizing dialog policies specified in domain language
using genetic algorithm. The human-interpretable representation of policy makes
the method suitable for practical employment. We present learning algorithms
using user simulation and real human-machine dialogs respectively.Empirical
experimental results are given to show the effectiveness of the proposed
approach. | ['Yonghong Yan', 'Weiqun Xu', 'Hang Ren'] | 2016-05-12 | null | null | null | null | ['user-simulation'] | ['natural-language-processing'] | [-1.66398883e-01 2.56429583e-01 8.05392787e-02 -6.19595826e-01
-2.24032700e-01 -7.23998189e-01 6.37210190e-01 9.53388661e-02
-5.99538863e-01 1.24439907e+00 2.47889966e-01 -5.03861129e-01
-2.81919539e-01 -6.01731777e-01 2.58618891e-02 -5.46876669e-01
1.64502487e-01 6.14895642e-01 1.89733699e-01 -6.19003177e-01
6.51777744e-01 3.95453870e-01 -1.36291265e+00 -8.76938775e-02
9.89884317e-01 4.48463589e-01 5.78139067e-01 8.06045294e-01
-5.11069775e-01 7.68022716e-01 -8.93248618e-01 2.16085121e-01
2.07287088e-01 -5.79859912e-01 -8.65469575e-01 5.09776592e-01
-5.20265341e-01 -3.95052403e-01 -2.33547345e-01 1.06241298e+00
5.22634089e-01 4.98437613e-01 6.29528522e-01 -1.21494699e+00
-3.56014192e-01 6.16912603e-01 9.88155827e-02 7.71733758e-04
4.59871143e-01 2.72613585e-01 6.93864048e-01 -1.54118538e-01
2.78163791e-01 1.68931007e+00 6.78943917e-02 7.13366449e-01
-9.25054848e-01 -2.77053386e-01 1.38166457e-01 1.49309874e-01
-1.20937824e+00 -2.88656801e-01 6.59766138e-01 -2.84557700e-01
1.10416508e+00 2.69411772e-01 3.86354059e-01 5.84532440e-01
2.38162830e-01 4.35752988e-01 1.56724799e+00 -7.88496017e-01
4.77160066e-01 8.24364483e-01 9.50350314e-02 6.39395237e-01
-2.30851382e-01 1.78503856e-01 -1.20331831e-01 -3.36312920e-01
7.05027461e-01 -3.04409653e-01 -2.28369772e-01 -1.30397961e-01
-5.62818706e-01 1.21306050e+00 -2.28561368e-02 3.01097840e-01
-3.49341065e-01 -3.63969266e-01 2.56303251e-01 5.47727942e-01
6.08741194e-02 5.06545484e-01 -4.55661952e-01 -4.85875517e-01
-1.87469348e-01 4.64887649e-01 1.28458250e+00 8.42010736e-01
1.76084384e-01 3.23561192e-01 1.45285726e-01 1.14681482e+00
6.69106543e-01 3.67520064e-01 6.45278156e-01 -8.99718225e-01
6.60741031e-02 6.72259808e-01 7.40190506e-01 -7.24006653e-01
-4.75016475e-01 3.63826096e-01 -4.28058475e-01 5.97052097e-01
6.46142840e-01 -8.22613657e-01 -4.20861810e-01 1.45081341e+00
3.76660526e-01 -2.05180079e-01 6.09616756e-01 9.47947025e-01
5.73883951e-01 1.03081310e+00 1.90473348e-01 -5.71778119e-01
1.26569235e+00 -7.44345069e-01 -1.14810169e+00 7.14408234e-02
1.82895020e-01 -9.11493003e-01 1.23279476e+00 4.78851706e-01
-8.46504807e-01 -3.57527614e-01 -9.83186722e-01 5.29118776e-01
-2.90404141e-01 -1.17734827e-01 3.32794249e-01 1.12102985e+00
-9.34027195e-01 3.97058457e-01 -4.78803754e-01 -6.63022757e-01
-4.25071657e-01 1.02459812e+00 8.07288140e-02 6.08339489e-01
-1.28868735e+00 1.19237471e+00 7.42672503e-01 8.62424299e-02
-3.50931525e-01 1.98290870e-01 -5.84589362e-01 1.21189453e-01
4.87675935e-01 -1.75555184e-01 1.85408616e+00 -7.56149113e-01
-2.43656826e+00 4.28132474e-01 1.99840870e-02 -5.81904054e-01
3.66563559e-01 -3.59072722e-02 -5.59098244e-01 8.92777275e-03
-6.14915967e-01 3.10000747e-01 6.31396413e-01 -1.28404224e+00
-7.71564305e-01 -8.00632387e-02 4.31462139e-01 6.33593142e-01
-4.49552089e-01 2.86483586e-01 9.46512148e-02 -3.23357344e-01
-1.26732513e-01 -1.01503015e+00 -5.24575770e-01 -4.52976227e-01
-3.03525150e-01 -3.48214716e-01 9.71409500e-01 -4.45833504e-01
1.29695415e+00 -1.71018243e+00 -1.75860196e-01 3.30920994e-01
-4.56009239e-01 3.74492586e-01 3.17537487e-01 9.09338713e-01
3.84794593e-01 -2.05217674e-02 1.43290684e-02 1.05911084e-01
2.97185212e-01 5.94657362e-01 -2.99185038e-01 1.75497308e-01
-2.98520803e-01 1.45399868e-01 -5.91314852e-01 -4.87317264e-01
6.09776080e-01 6.57379851e-02 -5.38591087e-01 8.41811001e-01
-5.36344886e-01 5.84523261e-01 -8.62187386e-01 2.80686855e-01
3.39013964e-01 2.45419115e-01 4.66722161e-01 1.91106200e-01
-3.00595969e-01 1.94807917e-01 -1.48998785e+00 9.77275610e-01
-5.91938496e-01 1.46032453e-01 4.00916427e-01 -8.15836549e-01
1.32319784e+00 6.02952659e-01 2.23445252e-01 -2.65961915e-01
4.89315480e-01 -7.03666881e-02 3.60580206e-01 -8.29398453e-01
5.22029161e-01 -4.33629811e-01 -1.18230723e-01 5.59078693e-01
-5.54718748e-02 -5.04125595e-01 9.06557590e-02 -2.01223418e-01
5.50958157e-01 -1.01920955e-01 8.34783554e-01 -3.55862886e-01
1.11214423e+00 1.08332649e-01 2.77224392e-01 7.26078272e-01
-4.47413385e-01 -1.01370767e-01 3.41261655e-01 -1.42792985e-01
-8.66737068e-01 -7.43064523e-01 1.17974930e-01 1.06197107e+00
1.66865632e-01 -2.76639927e-02 -9.85909402e-01 -5.03920317e-01
-1.60014689e-01 1.02523386e+00 -3.39596495e-02 9.19178203e-02
-5.50578594e-01 -4.02627468e-01 3.54982197e-01 5.69735616e-02
7.44044006e-01 -1.37840176e+00 -6.94765568e-01 5.77648342e-01
1.92624897e-01 -1.10127747e+00 -1.57707632e-01 1.12965681e-01
-7.44987488e-01 -7.83180594e-01 -3.28380048e-01 -8.27535093e-01
4.22459781e-01 1.03394501e-01 6.08740926e-01 3.42990249e-01
-4.46989276e-02 6.93918467e-01 -4.54440325e-01 -6.91892385e-01
-1.13508940e+00 -1.03973389e-01 2.99197584e-01 -2.82467395e-01
4.08778280e-01 -3.10649008e-01 -3.62232625e-01 5.95178604e-01
-8.21036398e-01 -3.24607670e-01 2.12081298e-01 1.03301775e+00
-1.66758284e-01 4.34960723e-01 1.04003882e+00 -8.16569626e-01
1.59536076e+00 -1.20465897e-01 -1.12437987e+00 4.97914046e-01
-6.38467669e-01 4.07396883e-01 7.29189396e-01 -1.93855911e-01
-1.64476860e+00 2.25922465e-01 -2.33196065e-01 3.20953131e-01
-6.39860332e-01 2.72711009e-01 -3.60606879e-01 -1.51196152e-01
4.93798763e-01 9.94606242e-02 3.58450502e-01 -3.48895550e-01
2.52146274e-01 1.28364336e+00 4.06796545e-01 -7.93311298e-01
5.60940266e-01 -2.49033421e-01 -3.37820947e-01 -1.25007427e+00
-3.80874164e-02 -4.69420731e-01 -3.35742235e-01 -4.22574162e-01
1.09276175e+00 -3.01467717e-01 -1.26854622e+00 3.62701803e-01
-1.02459431e+00 -3.06426734e-01 3.29118490e-01 4.28999633e-01
-6.93186283e-01 4.97840196e-01 -4.00414884e-01 -1.50128174e+00
-3.37728262e-01 -1.35932434e+00 4.17496383e-01 6.96523190e-01
-4.32689905e-01 -1.19846499e+00 -4.98933010e-02 2.55070686e-01
3.30132782e-01 -1.51758000e-01 1.03363788e+00 -1.01863408e+00
-2.98470825e-01 1.83300767e-03 2.91277111e-01 3.01484019e-01
5.45803487e-01 -5.06598391e-02 -9.14132416e-01 -1.01896100e-01
3.57046813e-01 -5.23869932e-01 -1.00107849e-01 1.74036816e-01
6.89751983e-01 -4.70454454e-01 -9.06176716e-02 -3.78078401e-01
1.29633868e+00 1.08964741e+00 3.93764853e-01 3.84693205e-01
-1.03037991e-01 7.94609368e-01 1.06665063e+00 9.47689056e-01
1.01845905e-01 8.26289952e-01 1.97904751e-01 2.75969833e-01
6.27999783e-01 -7.18032792e-02 3.77204835e-01 6.34313107e-01
1.55814812e-01 -5.73845506e-01 -8.30971897e-01 1.60474762e-01
-1.98976898e+00 -8.45313549e-01 2.70301878e-01 2.07475972e+00
9.09254849e-01 2.55976081e-01 3.02779853e-01 1.67453349e-01
7.66901672e-01 -2.00133920e-01 -2.11963326e-01 -9.98000443e-01
1.28851712e-01 8.58712569e-02 2.78186858e-01 9.85729754e-01
-9.75881934e-01 1.12695229e+00 5.66540146e+00 4.13358450e-01
-8.92790318e-01 -2.05883667e-01 2.95436293e-01 4.82523739e-01
6.73898309e-02 -7.58410292e-03 -7.61625826e-01 2.44768813e-01
9.56611514e-01 -5.51188946e-01 6.67882383e-01 7.54905701e-01
9.64513302e-01 -4.96466458e-01 -5.92533648e-01 6.94906473e-01
-5.84662259e-01 -1.00234103e+00 -2.29694590e-01 -2.01520160e-01
4.90235269e-01 -7.64613152e-01 -1.61464661e-01 2.10270926e-01
6.81150675e-01 -1.12784767e+00 1.29105687e-01 2.09819004e-01
-3.81356142e-02 -8.75614882e-01 8.12446415e-01 7.39084721e-01
-7.30508327e-01 -4.32454050e-01 -6.17811859e-01 -3.10529262e-01
2.80731469e-01 -3.75226676e-01 -1.61463451e+00 3.36131483e-01
3.12406152e-01 -3.07517946e-01 -6.05493151e-02 9.38689232e-01
-7.87638947e-02 6.50430143e-01 -3.64080817e-01 -9.74017441e-01
5.50311387e-01 -6.00618362e-01 5.08043468e-01 9.86904919e-01
2.13886797e-01 6.47463202e-01 6.12095356e-01 3.65859419e-01
5.34777939e-01 6.44442677e-01 -7.21036017e-01 -3.31300050e-01
6.74635470e-01 9.10149276e-01 -7.29547739e-01 -1.31812096e-01
-2.62061000e-01 6.08483791e-01 2.45707408e-02 3.12402844e-01
-6.26042366e-01 -2.45550603e-01 4.87139821e-01 -2.13766441e-01
-8.09436012e-03 -3.96663189e-01 -2.77301967e-01 -4.59855080e-01
-4.80051816e-01 -1.20957386e+00 3.25909197e-01 -3.60513121e-01
-1.03191507e+00 7.27709532e-01 3.64692926e-01 -1.15412748e+00
-7.93746889e-01 -7.32636034e-01 -6.10751390e-01 9.72677767e-01
-7.69900620e-01 -6.41376257e-01 -1.72530457e-01 5.73841155e-01
9.16145861e-01 -8.05649817e-01 1.19939244e+00 -2.36910209e-01
-4.21322733e-01 3.86860788e-01 1.88950673e-01 -2.71723509e-01
6.41084671e-01 -1.38110495e+00 -2.65250683e-01 2.84566641e-01
-2.56035089e-01 7.50349522e-01 1.39967501e+00 -5.81747353e-01
-1.29102838e+00 -4.86340910e-01 4.56286192e-01 2.13356927e-01
7.15828478e-01 -1.43008634e-01 -8.37649643e-01 1.96868062e-01
1.00756252e+00 -9.82915759e-01 5.49401343e-01 -1.66732952e-01
4.18566436e-01 7.24599063e-02 -1.39803720e+00 9.24233496e-01
2.57603079e-01 -2.73792893e-01 -8.00949693e-01 5.31499743e-01
4.06223357e-01 -5.21780729e-01 -6.31849229e-01 4.57879156e-02
1.60482317e-01 -1.01627171e+00 5.48739016e-01 -6.44886374e-01
-2.76002884e-01 -2.36540511e-01 -5.93815632e-02 -1.39012241e+00
2.80028969e-01 -1.13111746e+00 4.26986635e-01 1.26274025e+00
3.26932520e-01 -7.48785675e-01 7.24753141e-01 1.02306974e+00
1.36287212e-01 -4.99559164e-01 -7.43248641e-01 -7.08245754e-01
5.39372861e-02 -6.40808046e-02 4.51193511e-01 5.35594463e-01
4.75555420e-01 5.14683545e-01 -6.36353076e-01 3.08469683e-01
3.87247242e-02 -5.60708065e-03 7.74038434e-01 -8.25052321e-01
-6.24700904e-01 -2.55308479e-01 -8.27951804e-02 -1.08574867e+00
3.98038745e-01 -1.40292004e-01 4.70479518e-01 -1.35332894e+00
-4.90680695e-01 -4.64235365e-01 -9.60827246e-03 2.11371049e-01
-9.96375754e-02 -7.30123043e-01 2.31681347e-01 -2.77350485e-01
-9.40991417e-02 7.19713330e-01 9.94653344e-01 -1.12040266e-01
-6.29440665e-01 5.51792026e-01 -3.29693168e-01 8.19497108e-01
1.50068140e+00 -2.91601807e-01 -9.44618881e-01 2.65772551e-01
-5.71382523e-01 5.18682361e-01 -2.24622592e-01 -7.45926261e-01
2.40571409e-01 -6.26152396e-01 -1.38750151e-01 -2.77927637e-01
4.90445197e-01 -1.09831870e+00 -1.34430185e-01 6.55828774e-01
-5.53803921e-01 3.27391118e-01 2.63083130e-01 5.46220005e-01
-1.35635674e-01 -9.17468309e-01 8.39293897e-01 -1.74452901e-01
-9.14189517e-01 -3.88397604e-01 -8.11028898e-01 -3.11435968e-01
1.12703323e+00 -1.24833666e-01 1.84471253e-02 -1.18833494e+00
-7.36062229e-01 2.98272759e-01 1.84575871e-01 4.74120766e-01
5.58812141e-01 -6.73562407e-01 -2.43827432e-01 -1.85888827e-01
-3.09341460e-01 -7.38608837e-01 8.09236392e-02 9.35644284e-02
-6.93639398e-01 6.38668358e-01 -3.31158847e-01 -2.66762465e-01
-1.67398429e+00 4.76258159e-01 3.72050345e-01 -1.46512568e-01
-2.74258226e-01 4.17273313e-01 -1.75870493e-01 -6.39667034e-01
6.68833911e-01 -2.41391823e-01 -6.56018436e-01 -4.52634096e-01
3.29639167e-01 1.42618984e-01 -3.50659877e-01 -3.56039077e-01
-7.57883042e-02 7.29911476e-02 1.35160476e-01 -7.23927021e-01
9.60304499e-01 -3.91175032e-01 7.29050487e-02 3.91560644e-01
5.08118927e-01 -3.14912200e-02 -9.73182201e-01 -1.14380442e-01
4.31040049e-01 -3.88076544e-01 -3.63143355e-01 -9.62540030e-01
-2.89909035e-01 6.12647951e-01 8.83952677e-01 6.10215902e-01
9.54766989e-01 -6.83013141e-01 5.44163108e-01 1.07006586e+00
5.19468665e-01 -1.50244045e+00 1.67800952e-02 6.68976486e-01
1.06061804e+00 -1.42027664e+00 -1.45213008e-01 -5.46702683e-01
-1.04614162e+00 1.30627823e+00 9.27028298e-01 2.17647240e-01
7.30896533e-01 4.13247794e-01 6.53693438e-01 2.09161311e-01
-7.40952551e-01 -1.25413924e-01 -7.67645761e-02 6.78548932e-01
4.56556231e-01 -9.93927494e-02 -7.80332804e-01 3.99277806e-01
-2.21761867e-01 -9.32336301e-02 7.30804443e-01 1.20624828e+00
-9.47464466e-01 -1.66062272e+00 -6.88420475e-01 1.02101371e-01
-3.60604674e-01 3.38323206e-01 -5.95762253e-01 9.44619358e-01
-3.32919359e-01 1.67999184e+00 -5.21911621e-01 -1.72371805e-01
3.70641649e-01 3.09114397e-01 3.15263003e-01 -5.13893664e-01
-7.96921074e-01 2.94354349e-01 5.73246717e-01 1.35130957e-01
-3.66328180e-01 -4.05379832e-01 -1.81935942e+00 -1.29192173e-01
-2.15416208e-01 7.14278340e-01 5.34886479e-01 1.02157402e+00
-6.39961585e-02 1.76877901e-01 7.97579944e-01 -3.02817136e-01
-1.33463621e+00 -1.14180350e+00 -5.72323978e-01 1.46481484e-01
-2.22083062e-01 -8.39024603e-01 1.26333460e-02 -1.57706290e-02] | [13.043426513671875, 7.974075794219971] |
0bdfa056-185e-4bf6-8862-07764977c770 | category-level-global-camera-pose-estimation | 2209.14419 | null | https://arxiv.org/abs/2209.14419v1 | https://arxiv.org/pdf/2209.14419v1.pdf | Category-Level Global Camera Pose Estimation with Multi-Hypothesis Point Cloud Correspondences | Correspondence search is an essential step in rigid point cloud registration algorithms. Most methods maintain a single correspondence at each step and gradually remove wrong correspondances. However, building one-to-one correspondence with hard assignments is extremely difficult, especially when matching two point clouds with many locally similar features. This paper proposes an optimization method that retains all possible correspondences for each keypoint when matching a partial point cloud to a complete point cloud. These uncertain correspondences are then gradually updated with the estimated rigid transformation by considering the matching cost. Moreover, we propose a new point feature descriptor that measures the similarity between local point cloud regions. Extensive experiments show that our method outperforms the state-of-the-art (SoTA) methods even when matching different objects within the same category. Notably, our method outperforms the SoTA methods when registering real-world noisy depth images to a template shape by up to 20% performance. | ['Volkan Isler', 'Nicolai Häni', 'Selim Engin', 'Jun-Jee Chao'] | 2022-09-28 | null | null | null | null | ['point-cloud-registration'] | ['computer-vision'] | [ 4.78733405e-02 -2.07443327e-01 1.46644250e-01 -3.88260067e-01
-1.01455545e+00 -5.65120220e-01 6.16108179e-01 3.30909342e-01
-3.59666348e-01 2.27911398e-01 -2.61964798e-01 4.23263133e-01
-2.86810666e-01 -6.86902285e-01 -8.37852716e-01 -6.30753040e-01
2.74139792e-01 1.24294376e+00 5.98077893e-01 -1.52992129e-01
6.75744534e-01 1.09154236e+00 -1.62166548e+00 -2.38992214e-01
8.30697894e-01 8.50766361e-01 4.37209159e-01 3.65343578e-02
-2.24107176e-01 -2.87838072e-01 -1.95403889e-01 -3.56705874e-01
6.61423922e-01 2.25749508e-01 -6.70959711e-01 7.83885494e-02
9.84586000e-01 -5.89608289e-02 -1.00473769e-01 1.32589221e+00
4.39205050e-01 1.87446475e-01 2.64655471e-01 -1.19107640e+00
-5.43259643e-02 -7.52579197e-02 -7.35080481e-01 -1.60884351e-01
6.26015484e-01 -2.62445837e-01 6.32503867e-01 -1.23899877e+00
8.15400004e-01 1.16417825e+00 8.59752417e-01 1.76568046e-01
-1.15512705e+00 -7.96004176e-01 6.34932220e-02 2.05093235e-01
-1.84591019e+00 -4.14632857e-01 7.85869062e-01 -4.14931864e-01
9.06817496e-01 3.74280959e-01 6.15942419e-01 2.46089384e-01
2.57697463e-01 -9.69416928e-03 9.15266335e-01 -1.09466352e-01
1.20468676e-01 -2.13545457e-01 -3.54024693e-02 5.26652932e-01
2.69406617e-01 9.07238349e-02 -5.02381444e-01 -6.48139298e-01
8.91537666e-01 3.04434001e-01 -3.23255002e-01 -8.07803750e-01
-1.54056537e+00 4.75457162e-01 5.99381268e-01 3.01431328e-01
-4.13407892e-01 1.00855574e-01 -1.04004994e-01 3.23633552e-01
3.67870957e-01 3.09102893e-01 -2.88855702e-01 -1.87318698e-01
-8.41616809e-01 3.46093863e-01 4.71967220e-01 1.39513969e+00
1.24908543e+00 -6.40674651e-01 4.45724279e-01 6.52211726e-01
3.47629398e-01 5.82238019e-01 1.89292565e-01 -9.49301064e-01
4.37305331e-01 7.01995194e-01 1.24810666e-01 -1.31264925e+00
-3.79319489e-01 -2.43542269e-01 -7.34763026e-01 4.18397158e-01
1.16008453e-01 6.82536602e-01 -6.84713125e-01 1.10543978e+00
6.92147136e-01 5.72513998e-01 -2.57360578e-01 8.17228556e-01
6.83975220e-01 3.33844334e-01 -4.82744396e-01 -1.98100746e-01
1.10174739e+00 -6.85519218e-01 -4.46852118e-01 -2.18339369e-01
1.52068734e-01 -1.23845553e+00 4.77285266e-01 1.96130693e-01
-1.16600978e+00 -6.10397816e-01 -7.78016925e-01 -1.00151494e-01
6.34498373e-02 -2.63110936e-01 2.94170499e-01 1.26454636e-01
-1.08343351e+00 9.20480728e-01 -1.00591969e+00 -4.08348501e-01
1.39990762e-01 8.01159561e-01 -7.53374279e-01 -1.41423523e-01
-4.17493373e-01 8.47500801e-01 4.29100590e-03 4.64661270e-02
-1.87077612e-01 -8.55993927e-01 -6.94552183e-01 -2.55577385e-01
1.01046458e-01 -8.12904179e-01 8.93508315e-01 -3.16825479e-01
-1.39356172e+00 1.16259754e+00 -7.97039986e-01 4.08849195e-02
7.36921370e-01 -1.44893810e-01 -1.16236664e-01 -1.51413092e-02
3.04642320e-01 6.40724361e-01 7.69653678e-01 -1.64536238e+00
-6.61986291e-01 -6.41259968e-01 -3.90236616e-01 3.05219889e-01
3.44080001e-01 8.80250633e-02 -7.84345090e-01 -3.30617636e-01
1.34114814e+00 -1.13015044e+00 -2.97149241e-01 4.25423086e-01
-1.91945165e-01 -3.46281171e-01 8.37772012e-01 -1.32779717e-01
6.30807936e-01 -2.16221356e+00 3.03625971e-01 6.39493763e-01
2.39996642e-01 -2.58970588e-01 -5.22078248e-03 2.32493460e-01
-1.79204464e-01 -2.56062925e-01 -2.58679330e-01 -8.03534627e-01
-7.79752806e-02 2.68125981e-01 -2.15514839e-01 1.04568374e+00
-1.79743007e-01 5.68609953e-01 -8.66375446e-01 -5.41334271e-01
4.21268642e-01 4.28986818e-01 -4.06031847e-01 -2.94546578e-02
2.91163713e-01 6.12321436e-01 -3.83661032e-01 7.64111102e-01
1.23223245e+00 -8.73098895e-02 -3.58742923e-01 -4.56172764e-01
-2.65960991e-01 -7.23072216e-02 -1.56768882e+00 1.97610772e+00
-1.46779791e-01 2.87910491e-01 -1.01913407e-01 -5.62641680e-01
1.26706171e+00 1.62422642e-01 1.08311868e+00 -3.14518929e-01
1.33300334e-01 6.21027827e-01 -1.58629641e-01 -2.93956813e-03
6.72921896e-01 -1.09724529e-01 1.63567469e-01 1.42695501e-01
-2.32706845e-01 -8.39958251e-01 -3.91955107e-01 -3.03609788e-01
8.90990853e-01 4.61417697e-02 3.61321718e-01 -1.76450744e-01
5.93701363e-01 5.49730957e-02 6.23736262e-01 4.47962254e-01
-1.23634189e-01 1.23060441e+00 -2.92934090e-01 -5.78129709e-01
-1.00658822e+00 -1.11338055e+00 -5.98455727e-01 1.89815894e-01
8.03185403e-01 -3.50451529e-01 -4.34810966e-01 -3.20563525e-01
3.25540125e-01 6.91792816e-02 -3.37536812e-01 1.94386058e-02
-8.18360567e-01 -1.29979298e-01 -1.38594121e-01 3.20125341e-01
2.84619093e-01 -5.43067038e-01 -3.67264748e-01 2.21493945e-01
-1.28790244e-01 -1.21005821e+00 -6.24894857e-01 -2.36153409e-01
-1.27797031e+00 -1.00824714e+00 -5.26367009e-01 -8.21104944e-01
1.20996988e+00 6.85451269e-01 1.00583136e+00 3.97102267e-01
1.83789894e-01 3.10553879e-01 -1.65029272e-01 -2.10113645e-01
-2.64535517e-01 -6.70133233e-02 3.72314453e-01 -1.35392457e-01
5.17793119e-01 -6.83281600e-01 -3.55483502e-01 9.09513712e-01
-5.92845619e-01 -2.40784943e-01 2.40549400e-01 4.71000254e-01
1.41870725e+00 -4.33456786e-02 -2.03619525e-01 -3.20816845e-01
9.43891630e-02 -1.46819115e-01 -9.40823317e-01 1.38983622e-01
-4.38070655e-01 -4.34886478e-03 2.27342378e-02 -4.03046399e-01
-4.62519974e-01 6.40626013e-01 -4.63782810e-02 -9.73339856e-01
-8.13615546e-02 1.27766594e-01 -1.42648518e-01 -9.27010238e-01
4.58429277e-01 1.21960253e-01 1.01329826e-01 -6.00142121e-01
1.20100379e-01 4.10476685e-01 8.48892272e-01 -6.20346129e-01
1.43626249e+00 1.04746294e+00 3.07254285e-01 -5.05838513e-01
-3.35162520e-01 -1.06582105e+00 -1.37484610e+00 -7.09673986e-02
4.22708899e-01 -8.50212574e-01 -7.91383207e-01 4.75394964e-01
-1.49796104e+00 3.56693715e-01 -3.52181911e-01 5.61054707e-01
-7.34849036e-01 6.17471814e-01 -6.57768846e-02 -2.39030063e-01
-2.32866809e-01 -1.48311341e+00 1.49252248e+00 9.63174403e-02
-2.04784125e-01 -7.39753246e-01 3.87421459e-01 1.37407675e-01
1.33556277e-01 2.64997870e-01 2.12591648e-01 -3.10137749e-01
-9.34945405e-01 -5.07360637e-01 -3.33842449e-02 -4.01187688e-01
4.98605758e-01 1.18120223e-01 -8.56764734e-01 -5.02511144e-01
2.82939166e-01 4.03640658e-01 3.47476929e-01 2.37401381e-01
1.00641942e+00 2.09409550e-01 -7.08829939e-01 9.35702741e-01
1.50293708e+00 2.13223491e-02 7.71829128e-01 7.25078881e-01
8.26178610e-01 4.45566863e-01 9.88459170e-01 3.02436590e-01
4.35450196e-01 1.11573100e+00 7.12265074e-01 6.77219406e-02
1.23073936e-01 -1.92866568e-02 -1.57408878e-01 1.14993358e+00
-3.99200529e-01 3.65990788e-01 -1.00134456e+00 4.83334720e-01
-1.89515531e+00 -8.15243542e-01 -4.71756488e-01 2.65584898e+00
6.98608339e-01 1.28381960e-02 -3.09408307e-01 -5.76580390e-02
9.00874376e-01 -9.64420438e-02 -3.95379543e-01 1.17730059e-01
-1.05320424e-01 3.66692394e-01 6.68717921e-01 6.71819746e-01
-1.00312626e+00 9.27603245e-01 5.87023115e+00 5.45551538e-01
-1.03173518e+00 5.99135496e-02 -2.42129609e-01 2.11548194e-01
-2.18569160e-01 2.43178651e-01 -8.52125466e-01 2.93120503e-01
2.82936275e-01 -3.55785400e-01 2.36267090e-01 8.94634366e-01
-1.16359316e-01 -9.69529748e-02 -1.22454584e+00 1.47789919e+00
1.65745363e-01 -1.29998004e+00 -4.68521230e-02 2.85201102e-01
8.14194083e-01 4.13278103e-01 -2.78153270e-01 -4.84910578e-01
1.61169982e-03 -7.34967649e-01 7.64717042e-01 8.28165770e-01
6.53906345e-01 -6.31843328e-01 8.23614538e-01 2.94044375e-01
-1.43686771e+00 5.31595349e-01 -8.54956210e-01 2.69719929e-01
2.82043159e-01 5.40433109e-01 -6.23470426e-01 7.80295074e-01
8.82414937e-01 7.43471026e-01 -5.21442890e-01 1.62379301e+00
1.37917101e-01 -2.84045160e-01 -7.52910078e-01 5.60607731e-01
-1.57011658e-01 -6.19511247e-01 8.29910696e-01 5.68827868e-01
8.19075167e-01 2.28136107e-01 1.30630091e-01 6.69423521e-01
1.17071465e-01 4.82237898e-02 -5.97519875e-01 8.07125449e-01
8.28318894e-01 1.14021432e+00 -8.28227758e-01 -1.73688203e-01
-2.87230194e-01 1.13389957e+00 1.38151169e-01 -1.35571703e-01
-4.69644994e-01 -1.39058948e-01 9.78992403e-01 2.07762837e-01
7.33244792e-02 -4.73972827e-01 -3.27527881e-01 -1.10691297e+00
3.78332287e-01 -5.44924736e-01 7.45364055e-02 -8.48635018e-01
-1.34007275e+00 7.78901219e-01 6.88138902e-02 -2.00473952e+00
-6.67780340e-02 -1.77288637e-01 -5.18228054e-01 9.73451912e-01
-1.56050766e+00 -8.91964078e-01 -9.11381304e-01 7.03753769e-01
3.41096908e-01 1.05127782e-01 7.39758193e-01 3.07785660e-01
7.64459595e-02 3.72892022e-01 2.46847212e-01 -1.79711640e-01
8.71878147e-01 -9.87594545e-01 7.40432560e-01 6.28418148e-01
1.14103839e-01 8.47640812e-01 6.76777780e-01 -8.11431050e-01
-1.43464279e+00 -8.01877141e-01 9.50135767e-01 -7.25248873e-01
4.24941689e-01 -1.27967849e-01 -1.30643034e+00 6.88763022e-01
-2.35140070e-01 4.54525203e-01 1.41668633e-01 -2.51085222e-01
-8.96608010e-02 -1.51244730e-01 -1.42036617e+00 3.56350034e-01
1.32401872e+00 -4.73034203e-01 -7.55890906e-01 4.55013782e-01
4.65308040e-01 -1.20974362e+00 -1.22766435e+00 5.13719678e-01
3.53075892e-01 -8.22000086e-01 1.30169499e+00 7.40770251e-02
-2.40147680e-01 -8.46077740e-01 -3.20313007e-01 -1.30743086e+00
-4.31470752e-01 -7.06263304e-01 3.54291648e-01 1.06058455e+00
-2.34302312e-01 -8.65604639e-01 8.82804513e-01 6.13941371e-01
-3.31417322e-01 -2.97281265e-01 -1.48379505e+00 -1.02723706e+00
-1.09241255e-01 -2.65442193e-01 1.13422298e+00 1.23299646e+00
-3.57056707e-01 -2.96344638e-01 1.53627351e-01 8.21345389e-01
1.09075892e+00 3.94292891e-01 1.13888979e+00 -1.78074729e+00
3.36962283e-01 -3.77920628e-01 -1.14497554e+00 -9.28129911e-01
2.35624164e-01 -7.87583590e-01 1.97729334e-01 -1.33715975e+00
3.87429893e-02 -1.00640500e+00 1.38955861e-01 2.74431348e-01
-1.33109644e-01 2.71503717e-01 8.36952031e-02 8.31546247e-01
-1.64881900e-01 4.05728132e-01 1.14361620e+00 -3.69023271e-02
-1.49591610e-01 2.95324296e-01 -1.13868400e-01 7.83953667e-01
6.12690747e-01 -8.38263571e-01 1.16409637e-01 -6.05487287e-01
7.08178952e-02 -2.22706631e-01 4.10831273e-01 -1.25491059e+00
7.16006875e-01 -4.10435438e-01 2.56824605e-02 -1.05235612e+00
5.76498568e-01 -1.48264301e+00 9.30963278e-01 3.25577706e-01
3.83578628e-01 4.16197985e-01 1.35342151e-01 4.47099626e-01
-3.71120334e-01 -2.93591559e-01 8.38477671e-01 1.50604462e-02
-5.69227040e-01 6.64809585e-01 5.77719152e-01 -4.70547110e-01
1.06025720e+00 -7.85404444e-01 -1.17559768e-01 -4.58734743e-02
-6.60075366e-01 1.32082239e-01 1.30928111e+00 4.67342407e-01
9.02304530e-01 -1.55042779e+00 -7.08343446e-01 3.03224891e-01
4.59666491e-01 7.74444759e-01 -8.92589167e-02 7.88997293e-01
-6.70053661e-01 2.08316699e-01 -1.13669343e-01 -1.31775820e+00
-1.57438040e+00 3.72119278e-01 5.09605229e-01 3.39397997e-01
-8.25076699e-01 7.82923520e-01 9.02833641e-02 -7.33909369e-01
-2.17115209e-02 -5.98232388e-01 1.15755267e-01 -2.26143032e-01
2.88172305e-01 4.11782682e-01 6.31182730e-01 -1.25487602e+00
-6.42745078e-01 1.75665414e+00 1.47832707e-01 1.15893185e-01
1.40787482e+00 -1.62795365e-01 -5.30690253e-01 2.46129885e-01
1.27998960e+00 3.24319363e-01 -1.10826325e+00 -6.72809243e-01
-1.94140654e-02 -1.17081892e+00 -2.40785293e-02 6.79627061e-02
-1.07930326e+00 4.56502259e-01 7.77005315e-01 -2.71509826e-01
7.50948370e-01 3.50216150e-01 6.44777536e-01 4.43503529e-01
1.07289553e+00 -6.80727422e-01 -3.87248188e-01 3.63330811e-01
1.21270680e+00 -1.29989696e+00 5.32498062e-01 -8.45245004e-01
-1.74670443e-01 1.22536027e+00 4.16478842e-01 -3.30386609e-01
7.44310915e-01 9.88765731e-02 1.42568210e-02 -4.16925877e-01
-1.79264888e-01 3.34552340e-02 4.33324009e-01 6.90411866e-01
-1.76624820e-01 -1.35909155e-01 -9.64189023e-02 -2.38731548e-01
-5.83334982e-01 -1.84333548e-01 2.25701913e-01 1.01917243e+00
-4.59231645e-01 -1.33852053e+00 -8.88304651e-01 1.43081412e-01
5.39555922e-02 1.68618888e-01 -2.63238162e-01 7.17890203e-01
-1.16635017e-01 6.45241737e-01 6.12805605e-01 -3.50178510e-01
6.52712703e-01 -3.20203036e-01 5.82582772e-01 -5.99368274e-01
-6.04736984e-01 1.74482569e-01 -5.61996043e-01 -8.43112409e-01
-6.75945342e-01 -1.14482868e+00 -1.48402047e+00 -4.86453474e-01
-5.81875920e-01 7.07193613e-02 9.00520384e-01 7.39881933e-01
4.56310064e-01 -1.41655579e-01 8.83464456e-01 -1.33017707e+00
-2.78374314e-01 -4.84191507e-01 -3.59469205e-01 6.06511950e-01
3.27817976e-01 -9.62620854e-01 -4.35838878e-01 -7.18974993e-02] | [7.7400803565979, -2.877607583999634] |
9cbc9f11-49f4-4fef-9635-9c5c7dcd3efe | lipschitzness-effect-of-a-loss-function-on | 2303.16464 | null | https://arxiv.org/abs/2303.16464v1 | https://arxiv.org/pdf/2303.16464v1.pdf | Lipschitzness Effect of a Loss Function on Generalization Performance of Deep Neural Networks Trained by Adam and AdamW Optimizers | The generalization performance of deep neural networks with regard to the optimization algorithm is one of the major concerns in machine learning. This performance can be affected by various factors. In this paper, we theoretically prove that the Lipschitz constant of a loss function is an important factor to diminish the generalization error of the output model obtained by Adam or AdamW. The results can be used as a guideline for choosing the loss function when the optimization algorithm is Adam or AdamW. In addition, to evaluate the theoretical bound in a practical setting, we choose the human age estimation problem in computer vision. For assessing the generalization better, the training and test datasets are drawn from different distributions. Our experimental evaluation shows that the loss function with lower Lipschitz constant and maximum value improves the generalization of the model trained by Adam or AdamW. | ['Amin Gheibi', 'Mohammad Lashkari'] | 2023-03-29 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [-2.85850286e-01 1.70726016e-01 -3.24207127e-01 -5.63174069e-01
-1.49483129e-01 -6.50215968e-02 1.44357830e-01 1.11894928e-01
-8.35509121e-01 8.78716350e-01 -2.74605274e-01 -2.31673151e-01
-3.65836799e-01 -7.70411015e-01 -7.41672456e-01 -1.02382946e+00
2.54109725e-02 2.45428190e-01 4.93832398e-03 2.14281175e-02
2.04415411e-01 5.39630771e-01 -1.15276134e+00 -2.76053727e-01
1.07362318e+00 1.29408836e+00 -6.21880256e-02 5.10543644e-01
1.68170482e-01 3.26890975e-01 -8.08316231e-01 -8.28688145e-01
4.56203580e-01 -4.81724054e-01 -5.54052532e-01 -2.25103199e-01
4.19478029e-01 -3.57209593e-01 -1.30739778e-01 1.37240922e+00
5.84270597e-01 5.20567223e-02 8.22113216e-01 -1.43050528e+00
-6.29263818e-01 8.32263827e-01 -3.51527005e-01 2.55495667e-01
-2.18208402e-01 -1.04537204e-01 6.67137027e-01 -6.29059911e-01
7.20728412e-02 1.36157095e+00 7.86208868e-01 9.04234111e-01
-9.04031873e-01 -7.58476794e-01 1.56853676e-01 3.95501196e-01
-1.37168097e+00 -7.43988901e-02 7.77925432e-01 -3.30838263e-01
-4.33548354e-02 7.23366439e-02 3.51835251e-01 1.03759611e+00
2.28181258e-01 5.76888621e-01 7.65083253e-01 -4.08517122e-01
3.34314823e-01 5.81544280e-01 2.64175624e-01 7.44485319e-01
5.93858302e-01 -9.60317850e-02 -6.24106154e-02 -5.89930974e-02
8.24649811e-01 -3.43131721e-02 -4.26619142e-01 -9.19641256e-02
-4.98295575e-01 1.01346588e+00 4.41025913e-01 1.82251349e-01
-2.93641150e-01 2.76393324e-01 2.91929930e-01 5.12378097e-01
4.93177891e-01 6.06702507e-01 -4.28601235e-01 -1.06473841e-01
-5.45114875e-01 1.76198646e-01 8.00774634e-01 4.99152243e-01
2.97617763e-01 3.54463197e-02 -1.69253588e-01 1.00758195e+00
2.44000122e-01 3.80860150e-01 6.35875463e-01 -8.72098625e-01
4.97622758e-01 6.01849198e-01 1.95978209e-01 -8.24744582e-01
-4.27294284e-01 -5.70988834e-01 -7.14046001e-01 4.63884234e-01
1.15608883e+00 -5.13194919e-01 -5.42016447e-01 1.79684222e+00
3.17271203e-01 -1.01683505e-01 -5.58803743e-03 9.23504591e-01
5.59315920e-01 5.38457334e-01 7.73500577e-02 -1.42776221e-01
1.12502098e+00 -8.16609621e-01 -6.11683905e-01 1.36614442e-02
6.63072169e-01 -4.18810010e-01 1.14795303e+00 6.66577280e-01
-1.04523754e+00 -4.35243070e-01 -1.23869014e+00 1.75699845e-01
-2.03989923e-01 5.44402957e-01 3.16860646e-01 8.05047631e-01
-5.99729836e-01 8.59132946e-01 -8.69343519e-01 -3.43414217e-01
4.36950684e-01 4.07549798e-01 -1.37867123e-01 2.06592247e-01
-1.19563866e+00 9.21066105e-01 7.35552132e-01 1.15259819e-01
-5.83923221e-01 -6.79151893e-01 -3.81210327e-01 1.41651675e-01
-9.90897790e-02 -5.67747116e-01 1.20095146e+00 -1.30054092e+00
-1.44189155e+00 6.77911341e-01 3.43041807e-01 -8.79205942e-01
1.01665998e+00 -3.34992319e-01 -6.31141514e-02 -1.55167608e-02
-4.96673584e-01 3.03929895e-01 1.01292074e+00 -9.95892048e-01
-6.64133251e-01 -5.45590997e-01 2.13613644e-01 8.85712951e-02
-7.64432073e-01 -1.20991178e-01 -7.48309121e-02 -5.91815948e-01
-4.09913436e-02 -7.27949083e-01 -2.87295394e-02 2.80747265e-01
-2.03408986e-01 -4.87291574e-01 5.65697432e-01 -6.73843682e-01
1.33301032e+00 -2.18119645e+00 4.80347462e-02 2.95915514e-01
5.79774939e-02 2.51238406e-01 1.43694818e-01 6.80892915e-02
3.75141464e-02 1.95975155e-01 5.62341809e-02 -2.13521644e-01
-2.40540653e-02 1.15924455e-01 -1.22338787e-01 6.06260478e-01
9.44310427e-02 3.15407515e-01 -6.03451490e-01 -4.20253307e-01
-4.10814166e-01 5.61796248e-01 -3.79720151e-01 3.35272878e-01
1.81413114e-01 2.59963334e-01 -6.30084455e-01 3.36370081e-01
5.55431902e-01 3.78060900e-02 -2.33528703e-01 -1.04969308e-01
2.81261057e-01 -1.29987821e-01 -1.09081006e+00 9.42036510e-01
-4.55411166e-01 5.52786648e-01 -2.26580083e-01 -1.25033379e+00
1.40062761e+00 1.43160045e-01 1.45239428e-01 -4.78781760e-01
5.40700793e-01 3.07811916e-01 2.53887802e-01 -5.99753737e-01
9.25707743e-02 -2.24820197e-01 4.45345998e-01 1.55687854e-01
-1.05021968e-01 3.81591290e-01 1.19300656e-01 -4.06238288e-01
6.98311865e-01 -3.90338033e-01 1.11871272e-01 -4.86584783e-01
7.24536896e-01 -5.68635046e-01 6.22466326e-01 6.50509596e-01
-2.83637017e-01 4.72626597e-01 8.74748945e-01 -6.50594831e-01
-1.37704408e+00 -1.00013661e+00 -4.81812418e-01 1.02836192e+00
-1.53047862e-02 1.88046366e-01 -1.10996699e+00 -8.44073951e-01
8.35728645e-02 7.00968862e-01 -8.54330301e-01 -6.31166279e-01
-6.44158900e-01 -8.48848343e-01 6.70357764e-01 7.34392941e-01
6.41254306e-01 -6.35986030e-01 -5.59269428e-01 -1.64315730e-01
1.48737580e-01 -8.23049009e-01 -3.59438986e-01 8.15034360e-02
-1.09174120e+00 -1.04016721e+00 -1.04899073e+00 -7.25182116e-01
1.10982382e+00 -4.23965156e-01 7.86353290e-01 2.97433853e-01
2.16586098e-01 6.63717389e-02 -4.07089800e-01 -8.17216575e-01
-5.02972901e-01 4.24054980e-01 2.10460499e-01 -2.12798752e-02
2.30369523e-01 -3.91385138e-01 -8.77181172e-01 4.61282760e-01
-8.37426424e-01 -3.53138030e-01 4.09361750e-01 7.94260025e-01
2.13768780e-01 3.02906632e-01 7.07059503e-01 -7.28557348e-01
9.94634986e-01 -2.98698425e-01 -7.84190059e-01 3.58069837e-01
-8.57723057e-01 3.89537215e-01 1.01217473e+00 -9.11453128e-01
-8.06565344e-01 -1.60946742e-01 -2.61833251e-01 -3.67300957e-01
2.24124417e-01 2.94159889e-01 -2.17097223e-01 -1.06699392e-01
6.78143322e-01 -7.49413371e-02 6.47699684e-02 -5.57767034e-01
-3.94104198e-02 5.27828157e-01 3.29055905e-01 -8.06834936e-01
5.98089099e-01 2.25951999e-01 1.08129062e-01 -7.21332550e-01
-9.44497406e-01 1.95287094e-01 -3.84682089e-01 -2.83008248e-01
6.62399411e-01 -3.07279229e-01 -8.56355011e-01 5.61647475e-01
-1.06030941e+00 -3.37651670e-01 -2.07198143e-01 6.89592659e-01
-3.58884364e-01 -1.55339567e-02 -4.44295853e-01 -1.06460190e+00
-5.23904622e-01 -1.14597130e+00 4.05969650e-01 5.31681895e-01
-1.00381924e-02 -1.28671336e+00 -3.14499080e-01 2.14122795e-02
1.94138765e-01 3.33673835e-01 1.06001830e+00 -9.66781795e-01
1.64295807e-01 -4.34539229e-01 -8.16673338e-02 9.38558459e-01
-1.75422832e-01 1.68529749e-01 -8.06428492e-01 -3.44880044e-01
3.34320068e-01 -1.14555143e-01 7.17548847e-01 5.66615224e-01
1.60974097e+00 -6.27446830e-01 2.40540281e-02 7.57124782e-01
1.35124815e+00 3.31794083e-01 5.68402231e-01 6.35734320e-01
3.86449903e-01 5.54256201e-01 5.42218387e-01 5.24638951e-01
-8.99826586e-02 2.64669538e-01 5.02862811e-01 1.75739601e-01
3.76360536e-01 -1.65191203e-01 3.96344066e-01 3.52566034e-01
-2.71081656e-01 -1.39194587e-02 -9.55479503e-01 2.21422985e-01
-1.67670476e+00 -6.66197300e-01 1.90346435e-01 2.56515718e+00
8.68614912e-01 3.74734968e-01 3.57581943e-01 4.87867326e-01
8.43253553e-01 -2.56071866e-01 -6.66174233e-01 -7.35370517e-01
3.40798795e-02 -9.23069417e-02 6.53334081e-01 4.00222212e-01
-8.67630661e-01 3.28789175e-01 6.31213474e+00 9.27873313e-01
-1.43616533e+00 -1.34288430e-01 1.10727060e+00 -7.59708658e-02
8.88689309e-02 -3.66886675e-01 -1.04166198e+00 6.81889117e-01
8.32724333e-01 -2.16767728e-01 2.10546643e-01 1.16147196e+00
2.47204468e-01 5.81680425e-02 -1.35412729e+00 1.02665079e+00
-2.59611338e-01 -7.17697144e-01 -4.03650366e-02 9.88701582e-02
6.45944715e-01 -4.86620426e-01 3.86541545e-01 2.06877470e-01
-1.74600631e-01 -1.09432471e+00 7.08152294e-01 5.06831706e-01
2.74586350e-01 -1.11596942e+00 1.06098723e+00 5.43020189e-01
-5.26748836e-01 -4.07159597e-01 -6.60819113e-01 -7.72638060e-03
-2.97572643e-01 7.20091403e-01 -8.00463200e-01 -8.93780589e-02
5.70507348e-01 3.71795781e-02 -5.32915413e-01 1.20669281e+00
-3.14066291e-01 7.84613252e-01 -3.56139034e-01 -5.30813158e-01
1.41111061e-01 -4.32425022e-01 4.37846839e-01 8.94386411e-01
3.05597484e-01 -1.25211209e-01 -2.00794667e-01 8.99923086e-01
-2.76273966e-01 2.71458715e-01 -3.35423023e-01 -1.00119255e-01
5.57879984e-01 9.41808581e-01 -4.74240035e-01 -2.70090103e-02
-1.36822388e-01 6.67250752e-01 3.89992982e-01 4.74691212e-01
-1.04426014e+00 -5.82549691e-01 5.50960779e-01 3.67142439e-01
1.59390658e-01 -1.01280995e-01 -5.12759030e-01 -6.94698572e-01
2.73896366e-01 -6.49448633e-01 3.19014937e-01 -2.63096362e-01
-1.31683242e+00 5.79671621e-01 1.39924884e-01 -9.55034912e-01
5.82860261e-02 -9.53569233e-01 -8.17461073e-01 7.60854244e-01
-1.09567821e+00 -5.84154785e-01 -2.35475481e-01 2.80133605e-01
2.21973464e-01 -2.79108316e-01 2.58486331e-01 4.28239018e-01
-9.53141510e-01 1.24153018e+00 2.36478820e-01 3.91992986e-01
3.29517305e-01 -1.15080225e+00 -1.91995606e-01 8.17544222e-01
-4.13704544e-01 4.63757336e-01 1.10328972e+00 -1.05971582e-01
-8.09330523e-01 -9.17517245e-01 5.18682420e-01 -1.39710352e-01
5.83359957e-01 -2.17617936e-02 -1.03175080e+00 4.32012141e-01
-2.01748818e-01 -2.11906061e-01 3.96053076e-01 1.10230893e-01
-3.07799838e-02 -4.98747170e-01 -1.53155541e+00 6.80492282e-01
6.52006567e-01 -6.03139177e-02 -2.87445664e-01 3.47567312e-02
5.78359544e-01 -3.99442554e-01 -1.24329686e+00 5.04720390e-01
8.17933857e-01 -9.16936636e-01 8.37107778e-01 -8.16075325e-01
5.84478438e-01 2.78733760e-01 7.98420236e-02 -1.37290454e+00
3.39325778e-02 -2.29921281e-01 -2.50803947e-01 1.24237871e+00
5.71197152e-01 -9.00119007e-01 9.97565687e-01 8.96310747e-01
4.89515454e-01 -1.59433401e+00 -8.96265447e-01 -1.06540680e+00
5.96721768e-01 -2.10443437e-01 6.30686045e-01 5.08868575e-01
-2.77151972e-01 4.74735275e-02 -1.99139938e-01 -1.48014314e-02
4.45468962e-01 -4.56537992e-01 6.00066125e-01 -1.36460185e+00
-1.42257556e-01 -7.01207817e-01 -6.17280185e-01 -1.04598188e+00
1.79837808e-01 -4.06528056e-01 -2.00513124e-01 -1.17582107e+00
-6.48532063e-02 -4.09814268e-01 -4.38143432e-01 1.78861469e-01
-2.31888220e-01 -3.03583801e-01 1.09612294e-01 2.47276966e-02
3.37448642e-02 7.43128777e-01 1.41627431e+00 -2.49260087e-02
-2.44032979e-01 4.85114872e-01 -6.22430623e-01 1.00750232e+00
1.25250363e+00 -6.58455491e-01 -4.24079180e-01 -4.48528409e-01
3.62840503e-01 -3.74405801e-01 1.64626420e-01 -8.69489133e-01
-5.38577931e-03 -1.07816793e-01 4.29430664e-01 3.59299183e-02
1.58803105e-01 -1.01400089e+00 -4.14801270e-01 7.51156807e-01
-6.04653299e-01 1.55317634e-01 -1.27854347e-01 1.42968521e-01
-2.17075162e-02 -8.17994595e-01 1.30118489e+00 2.16798678e-01
-1.19973674e-01 4.16456521e-01 1.97023943e-01 1.63147584e-01
8.79569471e-01 -3.03160250e-01 -1.10287622e-01 -4.27642643e-01
-4.97946978e-01 2.81066000e-01 2.29095370e-01 2.70116240e-01
6.03064418e-01 -1.44409931e+00 -8.16086471e-01 4.55963649e-02
-1.25486597e-01 -8.35226402e-02 -3.46129686e-01 1.04214311e+00
-7.40936995e-01 1.22862764e-01 -2.12455541e-01 -3.00362170e-01
-1.25478792e+00 3.83736461e-01 7.78274119e-01 -9.85226333e-02
-2.38810807e-01 9.04085934e-01 -8.70078057e-03 -6.56564012e-02
8.31425369e-01 -5.45472383e-01 -2.60186523e-01 -2.01802656e-01
6.28261328e-01 8.13100696e-01 -8.88215378e-02 -2.02424616e-01
-3.43920216e-02 3.78670543e-01 -5.48075996e-02 1.78776998e-02
1.23127317e+00 4.14049178e-02 5.56380302e-02 5.42768180e-01
1.37760270e+00 -3.62937510e-01 -1.24719048e+00 -1.28970087e-01
-2.20594220e-02 -2.52825111e-01 -3.52820009e-02 -3.96826297e-01
-1.29674482e+00 8.10657680e-01 9.69837725e-01 4.24701333e-01
1.14377534e+00 -8.50977674e-02 6.80438697e-01 6.00783706e-01
6.23855181e-02 -1.25413096e+00 3.84014212e-02 2.72475958e-01
1.06503558e+00 -1.30670655e+00 -5.66840693e-02 -3.87292616e-02
-4.89886135e-01 1.48630846e+00 8.58598590e-01 -4.15504396e-01
8.97406936e-01 6.39021844e-02 4.43893149e-02 2.21151412e-01
-4.08048868e-01 3.52436811e-01 4.09150124e-01 4.15650696e-01
4.20960128e-01 -6.88000545e-02 -9.50034857e-01 8.89959872e-01
-4.46300715e-01 -1.06282748e-01 1.90954700e-01 2.14084074e-01
-5.72652578e-01 -1.01503015e+00 -3.82605076e-01 3.20785671e-01
-6.70236409e-01 2.27063909e-01 -3.16682756e-01 7.96669245e-01
2.39673167e-01 5.75844407e-01 -1.01198435e-01 -2.84359276e-01
3.88429552e-01 3.16221826e-02 6.56988502e-01 -7.37983510e-02
-3.34271878e-01 -6.15100861e-01 -9.54762250e-02 -2.40691289e-01
-3.11552603e-02 -3.68645936e-01 -1.41896999e+00 -4.45844084e-01
-3.93309265e-01 3.07463259e-01 8.07246029e-01 9.73703146e-01
-2.56425798e-01 1.87478945e-01 7.94100761e-01 -1.24215581e-01
-1.28810406e+00 -1.00045812e+00 -6.97247267e-01 4.83539224e-01
2.44697824e-01 -8.17621946e-01 -6.29981041e-01 -1.82610229e-01] | [7.918297290802002, 3.7010858058929443] |
4b941bd3-694d-4377-9a4a-9b4aeb6c61eb | pre-train-self-train-distill-a-simple-recipe | 2204.03642 | null | https://arxiv.org/abs/2204.03642v1 | https://arxiv.org/pdf/2204.03642v1.pdf | Pre-train, Self-train, Distill: A simple recipe for Supersizing 3D Reconstruction | Our work learns a unified model for single-view 3D reconstruction of objects from hundreds of semantic categories. As a scalable alternative to direct 3D supervision, our work relies on segmented image collections for learning 3D of generic categories. Unlike prior works that use similar supervision but learn independent category-specific models from scratch, our approach of learning a unified model simplifies the training process while also allowing the model to benefit from the common structure across categories. Using image collections from standard recognition datasets, we show that our approach allows learning 3D inference for over 150 object categories. We evaluate using two datasets and qualitatively and quantitatively show that our unified reconstruction approach improves over prior category-specific reconstruction baselines. Our final 3D reconstruction model is also capable of zero-shot inference on images from unseen object categories and we empirically show that increasing the number of training categories improves the reconstruction quality. | ['Shubham Tulsiani', 'Abhinav Gupta', 'Kalyan Vasudev Alwala'] | 2022-04-07 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Alwala_Pre-Train_Self-Train_Distill_A_Simple_Recipe_for_Supersizing_3D_Reconstruction_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Alwala_Pre-Train_Self-Train_Distill_A_Simple_Recipe_for_Supersizing_3D_Reconstruction_CVPR_2022_paper.pdf | cvpr-2022-1 | ['single-view-3d-reconstruction'] | ['computer-vision'] | [ 7.75352716e-02 2.33066782e-01 -3.21215272e-01 -7.88984478e-01
-9.54480410e-01 -8.14554453e-01 8.66150796e-01 -3.04700345e-01
3.93692078e-03 4.80828732e-02 4.07704622e-01 -5.95783331e-02
2.13055655e-01 -6.36041284e-01 -1.09747398e+00 -3.06000322e-01
3.09795290e-01 9.59124982e-01 4.50322270e-01 2.31486067e-01
3.32280457e-01 5.35970211e-01 -1.74410772e+00 2.68061608e-01
2.00812384e-01 9.27887142e-01 4.14655328e-01 4.61616695e-01
-1.37150243e-01 6.39934957e-01 -9.08480808e-02 -2.40474179e-01
6.21430814e-01 -1.06813625e-01 -9.73804414e-01 8.00742924e-01
1.23687756e+00 -7.01274514e-01 -2.46109337e-01 8.48036349e-01
1.16534494e-01 5.93150519e-02 1.10850334e+00 -9.55007613e-01
-6.50336146e-01 1.51142359e-01 -4.69336003e-01 -7.21303448e-02
4.15856659e-01 1.16824694e-01 1.15621805e+00 -1.14668930e+00
9.35763419e-01 1.50892055e+00 7.50609338e-01 7.73192167e-01
-1.52984583e+00 -3.70114684e-01 4.52157885e-01 -5.41276187e-02
-1.13119626e+00 -4.58776116e-01 5.96792638e-01 -7.47799814e-01
1.12898493e+00 -2.71708220e-01 6.06042862e-01 1.07780230e+00
-3.94409001e-01 9.44257975e-01 1.08918715e+00 -3.34688634e-01
1.89056829e-01 1.57943308e-01 4.25069332e-01 6.43840730e-01
2.93563247e-01 6.38115183e-02 -3.37957054e-01 1.04361318e-01
1.10695374e+00 2.68395573e-01 5.65999970e-02 -1.13947964e+00
-1.30015922e+00 7.45615900e-01 5.85657835e-01 -2.07088247e-01
-6.24160916e-02 3.74361634e-01 2.62282312e-01 2.61773895e-02
4.30432498e-01 1.25085607e-01 -5.42529106e-01 3.29426467e-01
-7.44863868e-01 1.96820349e-01 7.91494608e-01 1.35351217e+00
1.04162014e+00 -5.77713884e-02 2.21255079e-01 8.41300726e-01
4.49846715e-01 8.06015015e-01 -4.94817644e-02 -1.48369062e+00
7.59355351e-02 5.02919614e-01 1.94238201e-01 -5.04927099e-01
2.14089584e-02 -2.49076262e-01 -3.90442550e-01 2.54019678e-01
4.42819297e-01 3.45370471e-01 -1.45492208e+00 1.59849131e+00
2.98107862e-01 1.43673658e-01 1.70590989e-02 9.71309483e-01
1.00771046e+00 2.39856973e-01 2.66503394e-02 4.57716972e-01
1.12972689e+00 -1.04292023e+00 -3.20890695e-02 -4.91085142e-01
2.39549860e-01 -5.58221102e-01 1.01450717e+00 2.11891308e-01
-9.71752346e-01 -6.81346416e-01 -7.98838019e-01 -4.01522815e-01
-2.18740508e-01 -1.54113516e-01 9.37483728e-01 3.74234110e-01
-1.11130893e+00 2.89264381e-01 -8.30990970e-01 -7.01249123e-01
8.34128141e-01 4.62441780e-02 -5.80163181e-01 -5.72550952e-01
-4.24308479e-01 1.03712964e+00 2.24173114e-01 -5.96025527e-01
-1.85299027e+00 -8.05928290e-01 -1.04400146e+00 -3.74244988e-01
2.71183968e-01 -1.08199763e+00 1.38876808e+00 -6.92718327e-01
-1.16661561e+00 1.47460032e+00 -2.42971599e-01 -3.13344121e-01
2.52709419e-01 -1.59349307e-01 3.74714404e-01 3.74547601e-01
3.57536823e-01 1.22329569e+00 8.13173831e-01 -1.96112788e+00
-5.04490793e-01 -4.65273410e-01 3.89615148e-01 4.11493093e-01
1.55072555e-01 -3.09522033e-01 -6.09554052e-01 -2.64707506e-01
6.18732035e-01 -8.69116426e-01 -1.90196469e-01 3.21514010e-01
-3.06781739e-01 -7.80723691e-02 7.04091787e-01 -4.19197351e-01
1.10374980e-01 -1.90940475e+00 5.13061702e-01 -7.12018088e-02
1.98524430e-01 -5.33475459e-01 -1.71845958e-01 -5.17165326e-02
2.28988037e-01 -1.73797980e-02 -2.19319329e-01 -8.31433535e-01
1.10590629e-01 6.82824373e-01 -4.36682135e-01 5.20597935e-01
2.43733693e-02 9.03094053e-01 -8.57121229e-01 -4.51778412e-01
5.26272178e-01 3.96438420e-01 -9.28009808e-01 3.41419876e-01
-4.85093385e-01 3.67488861e-01 -2.40263551e-01 9.10271049e-01
5.97602546e-01 -6.35968745e-01 4.77124229e-02 -2.15519950e-01
4.15317625e-01 3.99973005e-01 -9.46232736e-01 2.28477383e+00
-6.66177213e-01 3.27155441e-01 -2.00986844e-02 -1.23567855e+00
7.99124658e-01 1.54631868e-01 2.96880692e-01 -1.53832048e-01
-6.08265288e-02 3.63725424e-02 -5.58315754e-01 -1.49196684e-01
1.14052914e-01 -4.92911935e-01 -2.04711810e-01 7.15044558e-01
6.92358196e-01 -8.23943377e-01 -3.25269192e-01 4.68275994e-01
7.87594199e-01 5.48116863e-01 1.07342759e-02 -2.39466771e-01
-8.96320865e-02 1.99333653e-01 2.20231488e-01 1.01145804e+00
-9.78803188e-02 9.21136439e-01 1.01563882e-03 -4.44811791e-01
-1.34596026e+00 -1.74918926e+00 -4.62043077e-01 8.03357720e-01
4.94708747e-01 -9.97945741e-02 -3.70349050e-01 -1.02944231e+00
2.29662195e-01 7.04983115e-01 -5.57391286e-01 1.88008785e-01
-6.54109046e-02 -2.68266767e-01 1.84611410e-01 6.31152809e-01
4.51516986e-01 -6.58635497e-01 -3.00611228e-01 -1.99939683e-01
-3.54821950e-01 -1.34737766e+00 -1.68261006e-01 2.31857151e-01
-1.26559973e+00 -1.15469658e+00 -4.32366371e-01 -8.99350584e-01
9.30835009e-01 9.60688174e-01 1.51547897e+00 -4.92704995e-02
-2.40121186e-01 1.14603579e+00 -1.75224632e-01 -3.66257876e-01
-3.64168435e-01 -2.01447293e-01 9.37901214e-02 -4.89602357e-01
5.68429887e-01 -8.16335320e-01 -3.14030647e-01 4.47494686e-01
-4.72569823e-01 3.29423100e-01 4.26019937e-01 6.35276675e-01
9.84784544e-01 -4.65494514e-01 4.20816600e-01 -8.27505112e-01
-3.99212152e-01 -5.63937604e-01 -5.59866607e-01 8.89541581e-02
-3.67334723e-01 9.65774134e-02 2.74419188e-01 -2.59689838e-01
-1.28236246e+00 2.69211978e-01 1.40148282e-01 -9.46498334e-01
-7.30361164e-01 -1.83880672e-01 -2.02988327e-01 2.58662719e-02
6.91879869e-01 2.81528741e-01 1.32272616e-01 -7.91421175e-01
7.54012644e-01 4.17290598e-01 5.54214537e-01 -7.32664824e-01
8.44484687e-01 9.25910175e-01 -2.04214796e-01 -5.82202017e-01
-1.46917069e+00 -7.01704800e-01 -1.03386855e+00 -1.53693184e-01
9.72146928e-01 -1.55417395e+00 -1.60063982e-01 3.49147618e-01
-1.07949197e+00 -5.17339766e-01 -3.03097546e-01 3.98811936e-01
-1.09861565e+00 3.37649316e-01 -5.93194962e-01 -3.62638950e-01
6.17606007e-02 -9.41917181e-01 1.57416117e+00 -2.30847135e-01
-6.75010905e-02 -1.11082554e+00 -1.69509664e-01 7.37585604e-01
4.01242077e-03 1.29833207e-01 8.19952786e-01 -3.87746930e-01
-1.22444546e+00 -4.73672710e-02 -3.62553626e-01 4.84596699e-01
1.30944446e-01 -5.09931266e-01 -1.17120063e+00 -1.76432997e-01
5.28883934e-02 -6.95398152e-01 1.08505404e+00 3.76861572e-01
1.05310750e+00 3.38421278e-02 -4.33998585e-01 7.01001108e-01
1.63676417e+00 -4.59297001e-01 4.61632639e-01 5.53487092e-02
9.89740551e-01 5.73897243e-01 5.46589077e-01 1.01832539e-01
6.33591413e-01 4.99080479e-01 5.81440032e-01 1.25928536e-01
-5.92540443e-01 -4.89874244e-01 2.85535336e-01 6.27429366e-01
-5.50771467e-02 1.14687704e-01 -7.95491278e-01 8.96983206e-01
-1.52044451e+00 -1.08945477e+00 2.99570799e-01 2.08713675e+00
6.78555012e-01 1.59913242e-01 6.76827133e-02 -3.59924793e-01
5.13154387e-01 -3.72099603e-04 -7.79223382e-01 3.49648595e-02
2.51595061e-02 9.84798595e-02 4.61136520e-01 6.34499550e-01
-1.10302126e+00 1.04648125e+00 7.67636442e+00 4.71050054e-01
-6.15747035e-01 2.48672232e-01 4.02067095e-01 -2.73954213e-01
-7.20649779e-01 3.71592879e-01 -1.00210130e+00 -1.50272802e-01
3.54679942e-01 2.13950440e-01 3.71426016e-01 1.17253304e+00
-3.04946214e-01 -9.72567722e-02 -1.51661718e+00 1.02029026e+00
3.71908695e-01 -1.40505147e+00 3.92057329e-01 3.62628669e-01
1.10987937e+00 5.32840550e-01 -1.28605768e-01 1.55800804e-01
9.27898765e-01 -9.19905663e-01 1.04931498e+00 5.82206547e-01
8.10097277e-01 -2.90557772e-01 2.90343791e-01 4.08435255e-01
-8.78206253e-01 7.98393115e-02 -5.19557774e-01 -1.19067542e-01
2.37927869e-01 5.38115501e-01 -8.20869625e-01 3.25732917e-01
8.08327436e-01 1.33386302e+00 -6.43121779e-01 7.00268745e-01
-3.52035403e-01 4.79022115e-01 -5.38411319e-01 3.98100585e-01
2.06238609e-02 -8.81821290e-02 5.34195900e-01 9.52694654e-01
9.25471336e-02 1.09124586e-01 4.46030617e-01 9.83985960e-01
-4.32736427e-02 -5.06689131e-01 -1.02888346e+00 3.45757186e-01
6.04810774e-01 9.24355447e-01 -7.35979617e-01 -5.71890831e-01
-6.16719425e-01 9.30570066e-01 5.64686418e-01 2.04684392e-01
-5.76537192e-01 2.73669213e-01 7.72456110e-01 2.67137170e-01
7.04452157e-01 -4.85809594e-01 -5.33056319e-01 -1.50370681e+00
-2.32007757e-01 -4.27533984e-01 2.61932015e-01 -1.04666352e+00
-1.61198914e+00 6.40705749e-02 4.41380799e-01 -1.13861120e+00
-2.87391484e-01 -6.88558340e-01 -1.23395637e-01 6.03443444e-01
-1.42522192e+00 -1.54111207e+00 -4.56770033e-01 5.90354919e-01
7.53158629e-01 -1.33599147e-01 9.10711646e-01 -3.63106169e-02
7.91825205e-02 1.68210164e-01 -1.24164470e-01 4.44019400e-02
5.60656667e-01 -1.35660958e+00 4.57704633e-01 7.14739442e-01
5.04060328e-01 6.05038583e-01 4.97848898e-01 -5.08002639e-01
-1.35870755e+00 -9.88966525e-01 4.17878240e-01 -1.36873412e+00
4.05486941e-01 -6.18546307e-01 -5.92321992e-01 1.23371553e+00
-1.11410029e-01 1.43095359e-01 6.87187672e-01 4.26258266e-01
-1.00500357e+00 2.08856836e-01 -1.18027675e+00 2.60573477e-01
1.73246694e+00 -8.04830551e-01 -9.95199800e-01 3.05078834e-01
7.07505405e-01 -2.85726070e-01 -1.04123652e+00 3.25983614e-01
5.68528712e-01 -1.05329788e+00 1.47302365e+00 -4.08864200e-01
6.09598577e-01 -3.45682412e-01 -8.44962656e-01 -1.06545699e+00
-3.03672761e-01 3.59793216e-01 -1.56520620e-01 9.00729954e-01
-4.77620139e-04 -2.91094601e-01 8.08362067e-01 3.82362068e-01
-3.47206473e-01 -3.51419181e-01 -5.85934997e-01 -7.28351235e-01
1.78264081e-01 -5.05327344e-01 2.82227308e-01 6.56851351e-01
-5.96607804e-01 5.71347058e-01 -1.45231053e-01 2.13890448e-01
1.34968996e+00 6.54232919e-01 1.20584738e+00 -1.38042867e+00
-4.95966136e-01 -1.58576146e-01 -5.81201494e-01 -1.57633579e+00
3.76976401e-01 -1.25219774e+00 1.59985766e-01 -1.83838844e+00
7.22759962e-01 -6.63387060e-01 7.27770552e-02 6.31098449e-01
1.88738599e-01 5.77661872e-01 2.16007456e-01 4.89093304e-01
-9.00630713e-01 4.48689193e-01 1.36230123e+00 -2.11123884e-01
2.71116376e-01 -2.66444892e-01 -8.30814362e-01 8.84350181e-01
3.37136477e-01 -3.26644868e-01 -6.69424057e-01 -8.92230451e-01
-2.94744134e-01 -3.00504804e-01 8.40860546e-01 -7.73817778e-01
-2.18041226e-01 -1.28968105e-01 5.79470277e-01 -8.32676470e-01
4.91488457e-01 -7.64569342e-01 1.03289180e-01 5.27924784e-02
-1.29141867e-01 -8.26900840e-01 1.49854213e-01 8.96608710e-01
5.61793186e-02 -1.90094471e-01 9.67101991e-01 -7.79354453e-01
-1.16595209e+00 3.90524864e-01 -6.73492029e-02 1.08171225e-01
7.25277483e-01 -4.88356143e-01 -8.84488896e-02 -3.77084553e-01
-9.46748793e-01 1.82140052e-01 1.14632320e+00 5.31808853e-01
5.98022163e-01 -1.35296810e+00 -4.91300762e-01 9.72645283e-02
4.29384708e-01 5.47674596e-01 1.38626114e-01 1.06432527e-01
-2.87855119e-01 3.36033732e-01 -2.26981789e-01 -1.33829796e+00
-8.87004852e-01 6.45828784e-01 4.48395818e-01 2.62309253e-01
-7.67066419e-01 8.43854189e-01 6.80975795e-01 -1.01542091e+00
3.30826849e-01 -1.61279619e-01 2.06674427e-01 -2.59336680e-01
2.75424123e-01 -7.58710643e-03 -1.66974291e-01 -7.03715384e-01
-3.34418088e-01 9.86290395e-01 -4.21297774e-02 -1.83697134e-01
1.51700819e+00 -3.83859724e-01 9.58172977e-02 6.84917688e-01
1.24906230e+00 -4.17124748e-01 -1.84040451e+00 -3.99495870e-01
-2.46708140e-01 -6.45468771e-01 5.42181097e-02 -7.61706769e-01
-8.51400018e-01 9.24650371e-01 3.82434934e-01 -4.06533331e-01
8.31735253e-01 8.13883185e-01 4.48001593e-01 6.12235725e-01
1.01012540e+00 -5.72355807e-01 5.13408124e-01 5.85870981e-01
6.61333084e-01 -1.67560399e+00 1.06540769e-01 -5.21498859e-01
-4.79066342e-01 6.87605858e-01 6.44350827e-01 -4.04204011e-01
7.83817887e-01 4.39093485e-02 4.28677090e-02 -4.88338590e-01
-6.79806352e-01 -3.41635913e-01 2.97896832e-01 9.90884423e-01
-1.65275782e-02 1.33705949e-02 4.92760390e-01 3.80816847e-01
-1.35607615e-01 -1.94603577e-01 4.59089369e-01 7.26598382e-01
-6.13051593e-01 -9.15389836e-01 -1.68446720e-01 4.93486136e-01
-2.89907912e-03 1.33566082e-01 -2.49237105e-01 7.78086841e-01
6.20151162e-02 6.30318522e-01 3.93792421e-01 -2.20756635e-01
1.84611708e-01 1.15607262e-01 1.17924380e+00 -1.17182112e+00
1.12777352e-01 4.76761200e-02 -2.02021506e-02 -5.94677687e-01
-8.68425369e-01 -8.05592954e-01 -1.22028661e+00 -9.59838927e-02
-2.79142112e-01 -2.91660488e-01 6.48871601e-01 9.01639044e-01
3.26472461e-01 -5.26678525e-02 5.90803564e-01 -1.30526543e+00
-4.81627971e-01 -8.17159832e-01 -5.44495404e-01 8.67677867e-01
3.19678873e-01 -1.18805301e+00 -5.80628037e-01 4.40826505e-01] | [8.372900009155273, -3.1174840927124023] |
5cdfc668-9e35-4895-b1fc-87704db3a546 | mix-n-match-ensemble-and-compositional | 2003.07329 | null | https://arxiv.org/abs/2003.07329v2 | https://arxiv.org/pdf/2003.07329v2.pdf | Mix-n-Match: Ensemble and Compositional Methods for Uncertainty Calibration in Deep Learning | This paper studies the problem of post-hoc calibration of machine learning classifiers. We introduce the following desiderata for uncertainty calibration: (a) accuracy-preserving, (b) data-efficient, and (c) high expressive power. We show that none of the existing methods satisfy all three requirements, and demonstrate how Mix-n-Match calibration strategies (i.e., ensemble and composition) can help achieve remarkably better data-efficiency and expressive power while provably maintaining the classification accuracy of the original classifier. Mix-n-Match strategies are generic in the sense that they can be used to improve the performance of any off-the-shelf calibrator. We also reveal potential issues in standard evaluation practices. Popular approaches (e.g., histogram-based expected calibration error (ECE)) may provide misleading results especially in small-data regime. Therefore, we propose an alternative data-efficient kernel density-based estimator for a reliable evaluation of the calibration performance and prove its asymptotically unbiasedness and consistency. Our approaches outperform state-of-the-art solutions on both the calibration as well as the evaluation tasks in most of the experimental settings. Our codes are available at https://github.com/zhang64-llnl/Mix-n-Match-Calibration. | ['T. Yong-Jin Han', 'Jize Zhang', 'Bhavya Kailkhura'] | 2020-03-16 | null | null | null | null | ['small-data'] | ['computer-vision'] | [-1.43716797e-01 -2.69987851e-01 -3.42083752e-01 -7.83589423e-01
-1.46721697e+00 -8.22046936e-01 4.85373288e-01 2.69542336e-01
-4.50948775e-01 9.82387900e-01 -3.42937112e-01 -6.23785436e-01
-3.67294133e-01 -5.67383230e-01 -8.62906933e-01 -1.01718760e+00
2.00180590e-01 5.83567619e-01 2.23020673e-01 1.42831862e-01
2.22254544e-01 4.15602207e-01 -1.66065812e+00 -2.41061375e-01
1.09716403e+00 1.23505819e+00 -2.88996220e-01 5.87514579e-01
3.68759692e-01 2.99943268e-01 -3.21159899e-01 -8.57253075e-01
3.92576635e-01 -2.49662519e-01 -5.41110992e-01 -3.85740221e-01
6.03203535e-01 -1.46629542e-01 -3.82443741e-02 1.22490239e+00
5.55733442e-01 -1.27984479e-01 8.73527944e-01 -1.62458932e+00
-4.94648278e-01 6.76107407e-01 -7.36046314e-01 1.46627516e-01
-9.39719900e-02 -9.09503996e-02 9.14795756e-01 -7.66141117e-01
1.68879375e-01 9.63489771e-01 9.06530082e-01 4.53589290e-01
-1.28672695e+00 -9.13491130e-01 -1.20095424e-01 -4.09391709e-02
-1.63107371e+00 -5.36625326e-01 3.47537547e-01 -4.37424481e-01
2.60389805e-01 3.82466704e-01 8.39239061e-02 1.14419603e+00
2.88502514e-01 4.57241535e-01 1.27606273e+00 -5.73382854e-01
5.92891932e-01 5.79628468e-01 5.65569222e-01 4.53042924e-01
7.02422202e-01 3.91587168e-01 -5.46534061e-01 -5.97908437e-01
3.82567555e-01 -4.40708816e-01 -2.23994374e-01 -6.75553083e-01
-9.29410994e-01 9.17490959e-01 1.36002526e-01 2.35594120e-02
1.03489995e-01 3.53534520e-01 3.87225598e-01 1.80046394e-01
5.71321428e-01 2.95882881e-01 -7.04840600e-01 -2.66987588e-02
-8.78986716e-01 4.96460080e-01 6.37637913e-01 1.31360161e+00
4.99608815e-01 -3.72586787e-01 -1.17190763e-01 6.31736457e-01
2.11366773e-01 5.81093550e-01 1.97288692e-01 -9.11044598e-01
3.91113758e-01 -9.12424549e-02 2.61705130e-01 -5.70879221e-01
-3.24715167e-01 -6.28930449e-01 -9.54241514e-01 1.76092312e-02
7.15876698e-01 -1.85751796e-01 -5.28406322e-01 1.92605519e+00
5.81731021e-01 1.48155555e-01 3.56468670e-02 6.62385643e-01
3.15439224e-01 1.66199878e-01 2.81097293e-01 -2.35039771e-01
1.37830842e+00 -4.19659466e-01 -5.65413475e-01 6.70079365e-02
7.75701702e-01 -6.52582586e-01 1.02280533e+00 3.53588760e-01
-8.45953047e-01 -5.10454535e-01 -1.25832605e+00 1.29199043e-01
-3.43169928e-01 2.19543800e-01 7.43798494e-01 1.18263364e+00
-6.74111366e-01 6.54910982e-01 -6.95816815e-01 -9.20148566e-02
4.12849903e-01 2.86845386e-01 -4.29344028e-01 -7.91591406e-03
-1.01249468e+00 8.77949357e-01 5.09099782e-01 -2.66250074e-01
-6.01670086e-01 -1.09679246e+00 -6.64087653e-01 -1.32431746e-01
2.76530683e-01 -6.38687253e-01 1.58685625e+00 -8.11648488e-01
-1.23459375e+00 8.39222968e-01 6.56775981e-02 -4.20755804e-01
9.00042534e-01 -2.76664257e-01 -3.05246145e-01 -2.29125038e-01
-8.59596208e-02 5.16821504e-01 3.67859155e-01 -1.20764315e+00
-8.47084582e-01 -4.75446701e-01 -2.64300793e-01 -5.26904687e-02
-3.02117765e-01 -8.30829144e-02 -2.94502079e-01 -4.91296947e-01
5.34629747e-02 -9.32707012e-01 3.28338929e-02 -2.91624703e-02
-3.97976279e-01 -9.98065993e-02 4.98082817e-01 -5.22407889e-01
1.27393985e+00 -2.14669085e+00 -2.59091884e-01 3.79500777e-01
-2.31057703e-02 1.64678499e-01 1.92332953e-01 1.92434937e-01
-2.02743992e-01 1.31367072e-01 -5.11023879e-01 -4.06033069e-01
1.71783417e-01 8.30553919e-02 -2.18099684e-01 9.10100996e-01
9.82259959e-02 7.37832129e-01 -6.90024495e-01 -6.04911327e-01
9.97915417e-02 3.51820499e-01 -3.39641720e-01 2.31809452e-01
-1.35578169e-02 3.38937998e-01 -2.25631624e-01 6.70582533e-01
9.77997959e-01 -1.14496402e-01 1.53877288e-01 -2.59483188e-01
1.55982718e-01 1.50434840e-02 -1.29037702e+00 1.27705586e+00
-3.25465530e-01 3.35280567e-01 -1.19952485e-01 -8.33054423e-01
8.72592092e-01 1.45845041e-01 2.32884973e-01 -3.24057549e-01
2.54912406e-01 5.79362333e-01 -9.71632674e-02 -1.27426581e-02
2.55860865e-01 -2.47598767e-01 -2.93236285e-01 2.56076336e-01
1.89731538e-01 9.69308615e-03 4.17436026e-02 7.16991276e-02
7.66422987e-01 1.17660463e-01 7.11451352e-01 -8.50537002e-01
3.94032985e-01 -3.09129655e-01 6.28628790e-01 8.72538090e-01
-3.35223794e-01 6.48275197e-01 5.61508060e-01 7.83710480e-02
-1.27288008e+00 -1.06539357e+00 -8.55180621e-01 1.02661693e+00
-3.07711046e-02 -1.14995293e-01 -9.69807923e-01 -6.94251537e-01
4.31780398e-01 1.08740091e+00 -7.53068388e-01 -1.62765115e-01
-8.61267224e-02 -1.11567438e+00 8.06609690e-01 8.23339462e-01
3.85191590e-01 -7.70600960e-02 -5.43561518e-01 -1.26775131e-01
-8.83420482e-02 -1.02470410e+00 -5.18844903e-01 4.87831354e-01
-7.78434694e-01 -1.17842817e+00 -5.75786769e-01 -2.38679156e-01
6.10639393e-01 -1.30489990e-01 1.08542955e+00 -9.79569107e-02
1.64429918e-01 4.83594388e-01 -2.30777308e-01 -7.75640190e-01
-3.81410450e-01 9.05208196e-03 1.76739097e-01 -1.28123343e-01
6.31999493e-01 -4.65880364e-01 -2.96234190e-01 5.36562383e-01
-7.90687323e-01 -2.64670014e-01 4.43291932e-01 9.46957290e-01
9.05439615e-01 -1.81257427e-01 6.76814079e-01 -1.31376159e+00
6.32529140e-01 -4.86001790e-01 -1.14384854e+00 7.06609309e-01
-1.18047750e+00 2.72183090e-01 3.55217665e-01 -5.19539654e-01
-8.83390069e-01 2.04907432e-01 -5.04858345e-02 -4.65602636e-01
4.17760499e-02 2.80031770e-01 -3.61382961e-01 -2.43008032e-01
1.03009403e+00 -2.17357755e-01 -2.52047986e-01 -3.88758272e-01
2.54884720e-01 9.29246485e-01 8.75668526e-01 -1.00797248e+00
7.72554338e-01 1.31013155e-01 2.91246504e-01 -3.36926579e-01
-9.87086654e-01 -5.70801497e-01 -7.36916006e-01 -6.48093317e-03
3.70289296e-01 -9.96619463e-01 -6.33728325e-01 4.61323082e-01
-7.53320634e-01 -1.12430036e-01 -2.23013163e-01 5.32078326e-01
-6.23841465e-01 4.30507898e-01 -3.05989176e-01 -1.09930682e+00
-4.24956620e-01 -1.05639255e+00 1.11686707e+00 2.62716174e-01
2.69566737e-02 -9.48728263e-01 5.96230961e-02 6.43320233e-02
3.17414701e-01 4.45163667e-01 7.91389108e-01 -9.42534029e-01
-1.10117614e-01 -3.99425000e-01 -2.95573026e-01 4.83609647e-01
-4.11932498e-01 1.53231874e-01 -1.42074239e+00 -3.05252701e-01
-1.50965959e-01 -3.73754382e-01 7.15512037e-01 4.72336441e-01
1.45740628e+00 -1.45113021e-01 -2.59064972e-01 7.32375979e-01
1.50244856e+00 -1.33244917e-01 6.17704332e-01 1.89460263e-01
2.61530966e-01 5.61413944e-01 8.31450701e-01 4.76514757e-01
2.73434848e-01 7.17877448e-01 1.53433979e-01 3.07068765e-01
2.86382496e-01 -2.67878264e-01 7.29765594e-02 5.69701016e-01
1.11021057e-01 -3.04413766e-01 -1.06198215e+00 2.89768398e-01
-1.99471951e+00 -4.77058411e-01 -3.01911861e-01 2.61575031e+00
9.71121132e-01 9.59729180e-02 1.93952829e-01 2.01431155e-01
7.86557376e-01 -5.00949740e-01 -6.15570784e-01 -1.75903752e-01
-2.10649431e-01 7.88181946e-02 9.38399255e-01 4.55289871e-01
-1.17878437e+00 3.60097259e-01 6.21898699e+00 1.11632073e+00
-6.01417303e-01 4.32627201e-01 8.96924496e-01 2.02968165e-01
-3.38979661e-01 1.56511024e-01 -1.12662530e+00 5.92151701e-01
1.34940577e+00 -2.64859796e-01 8.38788748e-02 1.14046228e+00
-4.42873895e-01 -3.75627786e-01 -1.38191688e+00 1.15108550e+00
-1.14583662e-02 -1.11404371e+00 -4.68199193e-01 1.18604906e-01
6.46452606e-01 -1.45311117e-01 7.62747303e-02 3.22408378e-01
2.56844491e-01 -8.65301788e-01 9.02161062e-01 5.40161610e-01
1.18543196e+00 -9.64609683e-01 1.20684075e+00 3.21719885e-01
-8.18792284e-01 3.28024141e-02 -6.17190838e-01 5.52808046e-01
-1.83624789e-01 1.15151775e+00 -2.30173633e-01 9.03302610e-01
7.04051316e-01 3.05478603e-01 -8.37554693e-01 1.06115210e+00
7.01365322e-02 6.61517143e-01 -5.98675370e-01 2.70094387e-02
-1.62619963e-01 -1.15701318e-01 2.03974903e-01 1.24100912e+00
4.73521441e-01 8.44429061e-02 -3.13782841e-01 8.08798492e-01
-1.80980742e-01 2.59785373e-02 -3.00004452e-01 2.54407465e-01
8.70576203e-01 1.03947914e+00 -5.06403685e-01 -2.05521882e-01
-8.30842480e-02 4.02526796e-01 4.33955461e-01 1.05540156e-01
-1.06211829e+00 -3.05355936e-01 4.12553847e-01 -5.32974601e-02
1.84496239e-01 2.61443276e-02 -4.24173415e-01 -1.02906358e+00
1.86388299e-01 -9.42011237e-01 7.41250515e-01 -5.57102084e-01
-1.63467181e+00 3.49516064e-01 5.00537038e-01 -1.13515413e+00
-2.36002222e-01 -8.29751015e-01 -2.97117591e-01 7.92154253e-01
-1.27962554e+00 -9.35668588e-01 -3.93693238e-01 4.88926291e-01
-1.65373921e-01 -1.81784973e-01 8.55162680e-01 3.02292615e-01
-5.38214207e-01 1.17428935e+00 5.48760951e-01 -1.34753913e-01
1.01999545e+00 -1.44209564e+00 7.97658265e-02 7.43314385e-01
2.96722744e-02 4.70720530e-01 8.91224146e-01 -3.76564860e-01
-1.22444594e+00 -9.15292442e-01 6.62235737e-01 -8.76508176e-01
5.65601707e-01 -4.51364070e-01 -8.46289814e-01 5.70488334e-01
-3.73034209e-01 2.39222124e-01 8.59980583e-01 3.82293761e-01
-7.34231293e-01 -3.90916079e-01 -1.38905311e+00 2.17719018e-01
7.20323861e-01 -3.57170045e-01 -2.80893564e-01 3.46564084e-01
4.20071065e-01 -5.73248863e-01 -1.10371256e+00 6.78955495e-01
8.07400703e-01 -1.22265518e+00 7.28542328e-01 -4.89056200e-01
1.12878844e-01 -1.19009636e-01 -5.76422274e-01 -1.01962578e+00
-1.52757764e-02 -4.15352613e-01 -2.09630191e-01 1.55303383e+00
4.47738290e-01 -1.00946558e+00 4.35787767e-01 8.39339256e-01
1.11849949e-01 -8.14786971e-01 -1.09138834e+00 -1.03386021e+00
5.70519328e-01 -5.83324432e-01 7.32693195e-01 8.69048953e-01
-1.81636348e-01 7.76175559e-02 -3.15482140e-01 2.72774965e-01
9.02380705e-01 -1.52514637e-01 7.57158935e-01 -1.30657423e+00
-3.66173357e-01 -3.18615437e-01 -4.12401974e-01 -6.54889107e-01
2.11312935e-01 -5.95138431e-01 5.66055924e-02 -7.60270298e-01
3.61157089e-01 -8.33505869e-01 -3.42051744e-01 3.17598432e-01
-2.31147677e-01 -1.29662603e-01 -1.56431794e-01 2.07333922e-01
-4.08622354e-01 2.88068533e-01 6.78717613e-01 2.36726522e-01
2.22366050e-01 1.96508661e-01 -9.36924756e-01 3.04865122e-01
1.05209947e+00 -7.51030385e-01 -3.79935741e-01 2.92651039e-02
2.36417785e-01 -7.07901120e-02 4.13451672e-01 -1.05685222e+00
2.47774988e-01 -1.11931384e-01 3.76511812e-01 -3.64439458e-01
-3.35922390e-02 -9.31373656e-01 3.93854737e-01 2.09507853e-01
-4.64656651e-01 4.05077003e-02 3.13881665e-01 6.91174865e-01
-1.87213160e-02 -4.99982208e-01 1.17652094e+00 3.74096096e-01
-1.86571360e-01 1.81845143e-01 3.36633503e-01 1.97961584e-01
1.09232712e+00 2.07914099e-01 -6.91698134e-01 -2.32513428e-01
6.72585657e-03 1.86579779e-01 4.89066154e-01 1.02311254e-01
5.80321439e-02 -1.28730714e+00 -8.92074466e-01 1.07337795e-01
5.14499128e-01 -1.24438688e-01 6.98910514e-03 9.91475642e-01
-3.38745505e-01 4.52616483e-01 4.09017764e-02 -7.43818700e-01
-1.14893806e+00 4.83187526e-01 5.60339570e-01 -1.21745050e-01
-2.44786173e-01 8.46559703e-01 -2.22267525e-04 -5.47348499e-01
5.58765233e-01 -2.15628624e-01 3.27104449e-01 -2.00038701e-01
5.37764132e-01 3.71218711e-01 3.05706561e-01 -3.16130191e-01
-4.58054304e-01 4.35506791e-01 1.35840818e-01 -1.59111768e-01
9.35788631e-01 1.75317805e-02 1.92628860e-01 5.15362859e-01
1.03143489e+00 -5.56001365e-02 -1.09883487e+00 -2.33316362e-01
1.89711511e-01 -5.49444318e-01 1.14954077e-01 -9.49069381e-01
-7.44668424e-01 9.27198112e-01 8.45201850e-01 -5.22876345e-02
1.17961192e+00 -6.38057366e-02 1.54346988e-01 3.48730564e-01
5.80885828e-01 -1.08945656e+00 -4.81580168e-01 -4.42694426e-02
7.04540014e-01 -1.29071867e+00 3.40210706e-01 -4.19943035e-01
-7.64472246e-01 1.04725540e+00 5.99542558e-01 1.45006761e-01
9.07862663e-01 4.41824436e-01 -1.10755146e-01 1.61043912e-01
-8.07360828e-01 1.94746330e-02 5.52774131e-01 7.76844740e-01
4.72577780e-01 2.10786656e-01 -3.56868804e-01 9.65867341e-01
-2.15314031e-01 -2.42586136e-01 3.39255184e-01 6.42066836e-01
-1.84250742e-01 -1.02755940e+00 -4.41616118e-01 6.14097178e-01
-4.08404857e-01 1.53928816e-01 -2.04117879e-01 1.07468832e+00
4.63463329e-02 8.92239451e-01 -9.64609757e-02 -3.19448918e-01
2.87744343e-01 2.47686967e-01 5.58641970e-01 -9.06337798e-02
-3.31857979e-01 -3.55493911e-02 1.41184032e-01 -4.65913773e-01
-4.77684855e-01 -9.29055989e-01 -7.37921357e-01 -6.90175951e-01
-7.97019780e-01 3.08576614e-01 7.93230951e-01 6.99414253e-01
2.15697110e-01 9.61137563e-02 4.44631338e-01 -2.63545811e-01
-1.37367344e+00 -1.00606108e+00 -7.64593482e-01 2.26263210e-01
1.45278186e-01 -8.82713020e-01 -6.41085446e-01 -1.23181216e-01] | [8.366475105285645, 4.191654682159424] |
3c831c9a-f10a-4db6-a05c-f303d8ccf673 | from-audio-to-symbolic-encoding | 2302.13401 | null | https://arxiv.org/abs/2302.13401v1 | https://arxiv.org/pdf/2302.13401v1.pdf | From Audio to Symbolic Encoding | Automatic music transcription (AMT) aims to convert raw audio to symbolic music representation. As a fundamental problem of music information retrieval (MIR), AMT is considered a difficult task even for trained human experts due to overlap of multiple harmonics in the acoustic signal. On the other hand, speech recognition, as one of the most popular tasks in natural language processing, aims to translate human spoken language to texts. Based on the similar nature of AMT and speech recognition (as they both deal with tasks of translating audio signal to symbolic encoding), this paper investigated whether a generic neural network architecture could possibly work on both tasks. In this paper, we introduced our new neural network architecture built on top of the current state-of-the-art Onsets and Frames, and compared the performances of its multiple variations on AMT task. We also tested our architecture with the task of speech recognition. For AMT, our models were able to produce better results compared to the model trained using the state-of-art architecture; however, although similar architecture was able to be trained on the speech recognition task, it did not generate very ideal result compared to other task-specific models. | ['Jiushuang Guo', 'Lingjie Kong', 'Shenli Yuan'] | 2023-02-26 | null | null | null | null | ['music-transcription', 'music-information-retrieval'] | ['music', 'music'] | [ 5.09686589e-01 2.07563341e-01 2.55852222e-01 -5.75514212e-02
-8.39233339e-01 -3.40615302e-01 7.08626151e-01 -7.49534816e-02
-4.69832152e-01 4.17161614e-01 2.55712748e-01 -1.35617331e-01
-2.91927636e-01 -5.87837338e-01 -6.30967796e-01 -4.28815633e-01
5.85759804e-02 6.63108885e-01 1.77908793e-01 -3.03608447e-01
1.62909076e-01 3.64835948e-01 -1.88287508e+00 7.01444566e-01
4.03214276e-01 9.29377735e-01 2.82141030e-01 5.85053086e-01
-4.84767050e-01 6.51710629e-01 -9.59165633e-01 -2.51306176e-01
5.17578050e-02 -7.64776766e-01 -8.17489922e-01 -2.72698730e-01
4.13420558e-01 1.48277700e-01 -3.44069451e-02 8.68009031e-01
6.24330819e-01 1.43061295e-01 7.34065354e-01 -9.65293705e-01
-3.99342269e-01 1.14605284e+00 1.12318374e-01 -1.05215162e-01
5.54744303e-01 -4.67049778e-01 8.88081849e-01 -9.90398526e-01
5.59325159e-01 1.03775048e+00 5.71571052e-01 5.71177363e-01
-1.25586116e+00 -4.89731491e-01 -2.53653467e-01 5.21268487e-01
-1.36794770e+00 -7.30507016e-01 7.83863544e-01 -2.88589418e-01
1.05918443e+00 4.18374747e-01 7.36969531e-01 1.25636005e+00
-1.37213826e-01 8.09187472e-01 9.26847279e-01 -7.63342738e-01
2.20997617e-01 1.35648042e-01 -2.27381289e-01 -9.34873149e-02
-3.62749040e-01 4.15329710e-02 -7.94230640e-01 8.90075713e-02
5.29284418e-01 -5.36565423e-01 -4.05358523e-01 -4.28304113e-02
-1.59357345e+00 5.09898782e-01 2.42820159e-01 1.10199392e+00
-4.70354646e-01 1.84938878e-01 5.74882567e-01 5.43239057e-01
1.13007091e-01 6.31163001e-01 -1.11697525e-01 -3.58343005e-01
-1.41614735e+00 1.78636894e-01 9.91453528e-01 3.58059108e-01
1.45900488e-01 4.80424494e-01 -3.15804183e-01 9.81080532e-01
-3.67000178e-02 2.28259280e-01 8.86102259e-01 -7.59493828e-01
2.31242210e-01 2.62742847e-01 -1.29024640e-01 -7.72609830e-01
-4.28277135e-01 -7.64739990e-01 -9.66424048e-01 2.09171921e-01
5.88338852e-01 2.28506595e-01 -6.84691072e-01 1.66172493e+00
-2.25892052e-01 3.19224566e-01 2.15056777e-01 1.03379858e+00
9.26320791e-01 8.75225961e-01 -4.49930876e-01 -4.07939762e-01
1.13973856e+00 -8.60795021e-01 -9.70404387e-01 8.64511356e-02
1.52389392e-01 -1.24596477e+00 1.02286565e+00 7.98167825e-01
-1.19200301e+00 -9.35653806e-01 -1.06354785e+00 1.48916617e-01
-4.06230688e-01 3.75595182e-01 1.69455320e-01 5.67162871e-01
-1.16162038e+00 8.06123197e-01 -4.77627069e-01 -4.82235730e-01
-2.64570743e-01 2.32595876e-01 -2.25053295e-01 4.91071016e-01
-1.36751056e+00 9.00852501e-01 4.71220136e-01 7.33644515e-02
-7.75893331e-01 -2.63410449e-01 -3.41863066e-01 2.51561761e-01
3.53008747e-01 -5.38755655e-01 1.50586390e+00 -1.24884140e+00
-2.09578395e+00 7.18435407e-01 -1.13333508e-01 -8.29709351e-01
2.83740014e-01 -1.85989633e-01 -5.94742417e-01 3.70893851e-02
-2.62862742e-01 6.74670756e-01 1.14091468e+00 -9.17330086e-01
-1.69982433e-01 -4.06948291e-02 -2.94041634e-01 2.67561823e-02
-2.64828771e-01 5.44142164e-02 -9.13377628e-02 -9.35664058e-01
5.17568439e-02 -9.05775011e-01 2.51750082e-01 -4.98868406e-01
-1.23458363e-01 -1.96176767e-01 5.85496247e-01 -6.64085090e-01
1.30519629e+00 -2.34647346e+00 5.63237190e-01 7.19993263e-02
-5.00648558e-01 5.92750371e-01 -3.18326563e-01 8.25882792e-01
-3.78925502e-01 -9.35335755e-02 -2.65952826e-01 -3.58844221e-01
2.98521489e-01 1.50193587e-01 -7.03408837e-01 -4.81046103e-02
8.65377784e-02 6.84917986e-01 -6.24868155e-01 -8.68867636e-02
1.50741562e-01 6.61998630e-01 -2.81023443e-01 1.58228979e-01
-3.79911780e-01 6.90598905e-01 1.40103817e-01 1.66261554e-01
1.26579925e-01 3.72296751e-01 2.19433531e-02 -2.37012818e-03
-2.53805518e-01 6.04216099e-01 -1.34506238e+00 2.03360105e+00
-5.87503493e-01 8.82011116e-01 2.37453785e-02 -1.25395501e+00
1.10954189e+00 9.63105857e-01 4.39266235e-01 -7.60572255e-01
1.17866158e-01 5.63186824e-01 3.12082291e-01 -3.74119371e-01
5.29569805e-01 -2.11258933e-01 3.60066146e-02 4.80172843e-01
1.95068881e-01 -3.94787699e-01 1.78732499e-01 -4.25841570e-01
8.42478454e-01 2.38385707e-01 2.25967318e-01 9.29936673e-03
8.28041315e-01 -6.93612248e-02 1.88663155e-02 5.87397039e-01
2.66811848e-01 9.42068756e-01 2.35880151e-01 -4.14278477e-01
-9.23641860e-01 -9.23494518e-01 -7.98250362e-02 1.14770174e+00
-3.79746586e-01 -5.61252236e-01 -1.05082083e+00 -7.89364651e-02
-4.74430233e-01 8.97962511e-01 -3.01463425e-01 -3.20584164e-03
-6.05683684e-01 -4.52036679e-01 8.86850238e-01 3.37661542e-02
3.18989843e-01 -1.57524133e+00 -5.31334281e-01 4.88132209e-01
-2.38414228e-01 -1.05543411e+00 -1.99120626e-01 2.19604522e-01
-6.02744758e-01 -7.52276778e-01 -1.07916296e+00 -8.88973176e-01
-2.24250734e-01 -6.52857497e-02 1.14258659e+00 -2.04620644e-01
-4.81610373e-02 1.63562119e-01 -7.08702147e-01 -7.79406011e-01
-9.14431036e-01 3.96436036e-01 1.27194598e-01 3.41970921e-01
1.31401673e-01 -8.74403775e-01 -1.76187992e-01 2.56113768e-01
-1.32450104e+00 4.92754467e-02 6.52297139e-01 6.82272911e-01
3.78560126e-01 1.16483018e-01 7.75253117e-01 -4.37469572e-01
9.16941941e-01 1.36261508e-01 -2.45835751e-01 4.68046172e-03
-1.87510148e-01 7.62452185e-02 7.24291503e-01 -6.21522546e-01
-5.59350610e-01 6.68694451e-02 -7.27530956e-01 -4.67819393e-01
-3.69276524e-01 6.80966556e-01 9.77705121e-02 1.43169269e-01
8.05744469e-01 3.97634357e-01 2.14788988e-02 -7.56972432e-01
2.48564839e-01 1.01398230e+00 6.11446977e-01 -2.73069203e-01
5.22949755e-01 8.50541815e-02 8.33666772e-02 -1.13970900e+00
-5.39175272e-01 -4.16382611e-01 -7.05556452e-01 -2.39726976e-01
9.24668431e-01 -5.19050121e-01 -4.82181996e-01 3.97893727e-01
-1.33246040e+00 -2.00430289e-01 -4.97657001e-01 5.88155568e-01
-8.71332705e-01 2.00964466e-01 -3.00629079e-01 -9.53136384e-01
-4.17563617e-01 -9.78935301e-01 1.07554793e+00 -2.82671481e-01
-5.22535563e-01 -6.43652380e-01 2.81375706e-01 1.37637347e-01
8.32316697e-01 6.27702475e-02 8.88315737e-01 -8.35060716e-01
-2.43351832e-01 -1.82240456e-01 3.28188896e-01 5.64757764e-01
1.41014708e-02 -1.96443945e-01 -1.31432104e+00 -1.68695182e-01
2.20802024e-01 -2.61373580e-01 1.01239824e+00 1.49335220e-01
9.31361079e-01 -9.28451568e-02 2.28095368e-01 1.57784566e-01
9.26814497e-01 3.24213207e-01 7.69228280e-01 3.10650378e-01
1.53650150e-01 7.02066302e-01 3.41049314e-01 1.23365708e-02
-2.26332530e-01 1.44617045e+00 3.80605489e-01 8.82375315e-02
-5.21916747e-01 -1.91797480e-01 7.92066097e-01 1.23735571e+00
-1.33724064e-01 -6.57099038e-02 -9.02712524e-01 4.02093619e-01
-1.87895238e+00 -1.16591489e+00 -9.23944339e-02 2.42105842e+00
6.45808458e-01 1.76626414e-01 3.94327134e-01 8.88772786e-01
5.41752577e-01 1.20298732e-02 -8.45391825e-02 -7.59995520e-01
-7.77396932e-02 6.60835087e-01 -2.62667984e-01 2.12405473e-01
-8.91030610e-01 6.47945464e-01 6.12900114e+00 1.06723785e+00
-1.50635409e+00 2.62971431e-01 -2.32288092e-01 -1.46251217e-01
1.55291751e-01 -3.43520284e-01 -2.77779996e-01 2.71493644e-01
1.45046163e+00 -3.03241261e-03 7.68903852e-01 4.46950287e-01
2.53291219e-01 2.83517689e-01 -1.32620382e+00 1.11423147e+00
1.99910671e-01 -1.04145503e+00 4.28852260e-01 -1.75668761e-01
4.48759496e-01 -1.96876094e-01 -2.30786316e-02 4.99482930e-01
-6.07362390e-01 -1.28709555e+00 1.08059025e+00 6.88531816e-01
6.67833030e-01 -6.90415561e-01 9.89519775e-01 6.65367782e-01
-1.14741695e+00 2.68719122e-02 -1.46221980e-01 -3.68561864e-01
2.23454863e-01 4.54805017e-01 -1.03860784e+00 7.47179091e-01
4.97650802e-01 5.09027898e-01 -3.27875495e-01 1.09552979e+00
-2.02474922e-01 6.30388558e-01 -2.62256742e-01 -1.35373762e-02
1.16704218e-01 -2.56614573e-03 8.95695686e-01 1.33628654e+00
6.85657859e-01 -4.99125212e-01 -6.25465065e-02 8.24177980e-01
1.62307546e-01 5.08686602e-01 -6.91218436e-01 -3.53932172e-01
5.51507771e-02 9.73344386e-01 -5.50598919e-01 -3.21075112e-01
-1.02398790e-01 9.04491305e-01 -5.25802858e-02 2.47771263e-01
-7.04360008e-01 -4.04349238e-01 3.50906044e-01 4.03819889e-01
2.74510264e-01 -1.48522481e-01 3.42196035e-05 -9.52516317e-01
-3.67166772e-02 -9.13185358e-01 -3.97262396e-03 -1.02823699e+00
-1.06584990e+00 1.01861596e+00 -9.15530398e-02 -1.58009970e+00
-8.57728064e-01 -6.48930907e-01 -6.19703829e-01 8.30627620e-01
-1.16294074e+00 -1.07374299e+00 6.18830025e-02 5.19629180e-01
7.44741678e-01 -3.83139700e-01 1.40784669e+00 5.16652524e-01
4.02671956e-02 4.53752220e-01 -3.80318798e-02 6.99928403e-02
8.27130616e-01 -1.10890758e+00 3.72408926e-01 6.28071547e-01
9.86167967e-01 5.44321418e-01 6.67103291e-01 -4.20470200e-02
-1.06992412e+00 -9.12215292e-01 1.05424607e+00 -1.02911890e-01
5.40584981e-01 -2.96755373e-01 -1.01462448e+00 3.14620852e-01
5.99746346e-01 -5.96525967e-01 6.17418170e-01 -1.06348544e-01
-1.75549209e-01 -1.16911240e-01 -5.43220460e-01 4.24399763e-01
8.10084462e-01 -7.10017145e-01 -1.06852627e+00 8.26252699e-02
6.27455294e-01 -4.10364926e-01 -7.41890371e-01 4.63712215e-01
5.62303603e-01 -1.12267900e+00 9.36029732e-01 -2.44558692e-01
2.85920233e-01 -5.63001454e-01 -2.01359153e-01 -1.40151012e+00
-9.51260030e-02 -6.58311427e-01 7.14020655e-02 1.24223506e+00
5.19589782e-01 -4.66472507e-01 3.27805758e-01 -3.00212771e-01
-3.12427312e-01 -2.66396344e-01 -1.32534885e+00 -9.82173920e-01
-3.83670256e-02 -7.28212535e-01 5.16509354e-01 7.15014875e-01
5.43159782e-04 6.62778854e-01 -4.76241112e-01 -1.55482411e-01
2.12916642e-01 2.29085013e-01 9.11741197e-01 -1.42089009e+00
-6.71379745e-01 -8.11113298e-01 -4.82466042e-01 -6.05381846e-01
2.89482743e-01 -1.31017935e+00 6.30413592e-02 -1.49092329e+00
-3.13037992e-01 1.41580358e-01 -2.74173141e-01 5.04502714e-01
4.03470337e-01 5.41894913e-01 6.24309301e-01 1.31443799e-01
-2.51158327e-01 4.01291668e-01 9.72148716e-01 -3.16674203e-01
-3.69736880e-01 3.06077659e-01 -1.72127515e-01 5.15209198e-01
8.21509123e-01 -5.00372171e-01 -2.38526508e-01 -2.83275753e-01
4.69254047e-01 1.94475994e-01 3.53320628e-01 -1.53405213e+00
3.09738457e-01 4.83262390e-01 7.40790814e-02 -6.25635564e-01
7.07143784e-01 -9.05296624e-01 7.46352553e-01 4.71265703e-01
-3.57641488e-01 -1.72563598e-01 3.32959086e-01 3.20583098e-02
-8.01867604e-01 -5.85099757e-01 4.81978506e-01 -1.43748030e-01
-4.04121280e-01 -4.87551510e-01 -6.78701580e-01 -4.21521157e-01
5.08277833e-01 -2.17103466e-01 2.99405809e-02 -6.28304362e-01
-1.07355368e+00 -5.21929622e-01 2.12362513e-01 6.66186810e-01
6.09046042e-01 -1.29437637e+00 -8.47910285e-01 2.98952878e-01
-8.58008340e-02 -2.49086395e-01 -2.72200331e-02 9.27985489e-01
-2.80233920e-01 7.65029609e-01 -3.52729321e-01 -7.89028227e-01
-1.35824800e+00 6.37553096e-01 3.24457914e-01 -1.41523093e-01
-3.76394153e-01 2.89148688e-01 -2.08098561e-01 -4.64310527e-01
6.35271370e-01 -4.55842018e-01 -5.97165525e-01 2.90526807e-01
6.31342888e-01 2.09248796e-01 5.21586001e-01 -6.84341490e-01
-8.05315003e-02 8.19568634e-01 4.95315999e-01 -5.95939338e-01
1.40265691e+00 2.17520535e-01 -2.72693932e-01 1.02091753e+00
8.09051871e-01 8.88907835e-02 -2.58480728e-01 -5.38801067e-02
1.88149989e-01 1.99902654e-02 -1.17076198e-02 -8.12816978e-01
-6.69185579e-01 1.17418253e+00 5.43946743e-01 6.98655665e-01
1.26912284e+00 -2.43227988e-01 6.87528610e-01 5.45433700e-01
5.58254957e-01 -9.35549200e-01 4.92939986e-02 7.69693255e-01
1.29801798e+00 -7.84100413e-01 -4.38282222e-01 -1.38782352e-01
-3.37967694e-01 1.47410941e+00 2.59536393e-02 -1.02674868e-02
3.21798205e-01 1.52268428e-02 1.11007042e-01 1.34262219e-01
-6.10351562e-01 -4.31421220e-01 6.20401919e-01 4.26718086e-01
6.11468852e-01 -5.17096408e-02 -4.59448516e-01 6.27783775e-01
-7.14081168e-01 2.67394632e-01 5.28210521e-01 5.40639579e-01
-3.46490175e-01 -1.43599796e+00 -7.83203721e-01 1.74481645e-02
-6.44647121e-01 -1.52769750e-02 -8.27153087e-01 6.46761298e-01
1.78223982e-01 1.01681805e+00 -1.08092718e-01 -4.65797186e-01
5.77790916e-01 5.78366280e-01 4.96366739e-01 -6.67344749e-01
-9.62212145e-01 3.41023117e-01 9.50918272e-02 -3.46477419e-01
-7.88878262e-01 -5.03409207e-01 -1.08037913e+00 1.22424580e-01
5.98726347e-02 4.24743325e-01 1.04360592e+00 8.55903864e-01
6.32763281e-02 9.97991264e-01 3.26171160e-01 -1.32405543e+00
-3.85210365e-01 -1.32009423e+00 -5.55676162e-01 4.87407625e-01
2.45870948e-01 -2.47686431e-01 -2.25850552e-01 1.65835679e-01] | [15.790192604064941, 5.361970901489258] |
5932c7db-d1bd-4cc7-b092-c1dc1b4fd1b0 | improving-neural-abstractive-document-1 | null | null | https://aclanthology.org/D18-1441 | https://aclanthology.org/D18-1441.pdf | Improving Neural Abstractive Document Summarization with Structural Regularization | Recent neural sequence-to-sequence models have shown significant progress on short text summarization. However, for document summarization, they fail to capture the long-term structure of both documents and multi-sentence summaries, resulting in information loss and repetitions. In this paper, we propose to leverage the structural information of both documents and multi-sentence summaries to improve the document summarization performance. Specifically, we import both structural-compression and structural-coverage regularization into the summarization process in order to capture the information compression and information coverage properties, which are the two most important structural properties of document summarization. Experimental results demonstrate that the structural regularization improves the document summarization performance significantly, which enables our model to generate more informative and concise summaries, and thus significantly outperforms state-of-the-art neural abstractive methods. | ['Xinyan Xiao', 'Yajuan Lyu', 'Wei Li', 'Yuanzhuo Wang'] | 2018-10-01 | null | null | null | emnlp-2018-10 | ['abstractive-sentence-summarization'] | ['natural-language-processing'] | [ 4.83782947e-01 1.06439739e-01 -5.09552956e-01 -2.12293029e-01
-1.01401401e+00 -3.89365911e-01 4.64853406e-01 6.98396504e-01
-2.03280702e-01 9.54503000e-01 1.13901234e+00 -5.60455844e-02
2.95724981e-02 -5.43730319e-01 -6.10992908e-01 -3.47229332e-01
9.13313404e-02 1.94356844e-01 -6.13392331e-02 2.68164556e-04
6.31751418e-01 3.75304036e-02 -1.34282386e+00 5.72983861e-01
1.46337342e+00 4.61578786e-01 3.41985792e-01 9.47254777e-01
-2.57941335e-01 7.45248199e-01 -8.54108989e-01 -1.16576478e-01
-3.05692554e-01 -8.93174410e-01 -8.27156007e-01 7.21988454e-02
6.74609959e-01 -5.97877920e-01 -4.30936515e-01 9.84067261e-01
6.46967053e-01 4.61457789e-01 8.87034357e-01 -4.18085545e-01
-6.79228365e-01 1.13403022e+00 -8.71498704e-01 3.28806847e-01
5.26729167e-01 -7.51611143e-02 1.40820181e+00 -4.73710120e-01
4.25514549e-01 1.21793878e+00 4.67998981e-01 5.32247066e-01
-1.05274820e+00 -2.29224518e-01 3.53988409e-01 -3.27375866e-02
-6.22401118e-01 -7.03935027e-01 8.35932434e-01 8.55849609e-02
1.33687496e+00 6.30595744e-01 7.95071244e-01 1.00142360e+00
5.05581677e-01 1.40677559e+00 1.27247036e-01 -3.11477393e-01
1.79020271e-01 -3.09526026e-01 6.19249105e-01 4.78809386e-01
8.13370764e-01 -4.70778406e-01 -5.10559499e-01 7.77949695e-04
3.30277652e-01 -3.66689675e-02 -4.06792551e-01 2.25442782e-01
-8.76911759e-01 8.36618662e-01 2.03245178e-01 4.89485949e-01
-6.37040019e-01 1.84315339e-01 8.18814814e-01 2.09563579e-02
7.36930549e-01 8.44098091e-01 2.71985941e-02 -2.88883567e-01
-1.56460476e+00 4.69502211e-01 8.57678413e-01 8.80449593e-01
4.35045749e-01 4.21978384e-01 -8.58925641e-01 9.74508226e-01
-1.37234896e-01 4.35017705e-01 8.31117392e-01 -9.06863034e-01
8.49370241e-01 6.57847822e-01 -1.28464967e-01 -1.08729541e+00
-3.75129431e-01 -6.29287243e-01 -1.44018865e+00 -8.94840121e-01
-3.68987709e-01 -2.53282160e-01 -8.88992608e-01 1.49662125e+00
-3.41595352e-01 -2.41639242e-01 2.03890875e-01 4.16670501e-01
1.14903545e+00 1.06607163e+00 -1.21075198e-01 -7.04332948e-01
1.01122534e+00 -1.27449429e+00 -1.04301465e+00 -4.89384234e-01
7.29759514e-01 -3.92738163e-01 9.87796485e-01 -1.61731225e-02
-1.79591024e+00 -4.95958686e-01 -1.18723869e+00 -1.95139065e-01
2.69529641e-01 3.01454097e-01 4.20987904e-01 2.11201131e-01
-9.14995968e-01 7.32557833e-01 -9.25930619e-01 -2.34578222e-01
6.30934119e-01 2.11833138e-02 5.42878583e-02 -4.19293977e-02
-1.00989497e+00 6.07273340e-01 7.88850427e-01 -1.90058485e-01
-5.43093026e-01 -6.64000630e-01 -1.04179728e+00 7.13319600e-01
2.04648152e-01 -1.16305399e+00 1.45666063e+00 -4.59527045e-01
-1.31362927e+00 2.44766146e-01 -6.09350145e-01 -7.01313615e-01
2.36264944e-01 -4.89757597e-01 1.13263324e-01 3.80292147e-01
4.02914509e-02 5.90816319e-01 4.61374134e-01 -1.15919185e+00
-6.45485044e-01 -1.99128166e-01 -2.55701840e-01 5.70842803e-01
-7.18590975e-01 -2.91591525e-01 -4.09980416e-01 -8.02350342e-01
-2.01119944e-01 -3.41446996e-01 -3.57976437e-01 -9.85234559e-01
-1.10188711e+00 -1.91204295e-01 4.42582607e-01 -9.55617666e-01
1.86637032e+00 -1.74090314e+00 3.22837740e-01 -2.74106681e-01
2.24495307e-01 5.22020400e-01 -4.75040883e-01 8.47938240e-01
3.08680445e-01 2.31259704e-01 -4.05612409e-01 -8.87129545e-01
-1.22684851e-01 -1.98607575e-02 -5.83109379e-01 4.37903963e-03
1.36146516e-01 1.35259378e+00 -9.08674598e-01 -5.27694702e-01
-1.45554081e-01 4.74198759e-02 -6.85079396e-01 2.86423594e-01
-4.45348561e-01 2.52534360e-01 -6.23718798e-01 1.72099143e-01
4.37278330e-01 -1.29150838e-01 8.33360255e-02 7.76809976e-02
1.64253749e-02 7.18892157e-01 -4.29970980e-01 1.81414235e+00
-3.56113136e-01 6.10713422e-01 -3.47748280e-01 -1.13527918e+00
7.56870866e-01 5.96450455e-02 3.69946420e-01 -7.30508149e-01
8.79117250e-02 8.32000375e-02 -2.64645070e-01 -4.10237551e-01
1.31687903e+00 3.55876684e-02 -6.02817312e-02 8.11001837e-01
-2.02278391e-01 -2.24761173e-01 9.16321993e-01 6.19974673e-01
9.10508931e-01 -3.21251601e-01 4.96218920e-01 -1.10113166e-01
4.07973975e-01 -2.33330697e-01 3.35589290e-01 1.05872190e+00
3.82049024e-01 8.35815787e-01 7.54697740e-01 1.85459092e-01
-1.26306808e+00 -6.95047915e-01 4.31512922e-01 8.43685329e-01
-1.01653062e-01 -6.62263691e-01 -1.06397414e+00 -4.64090258e-01
-1.61744073e-01 1.24365377e+00 -3.58009994e-01 -6.67002797e-01
-8.73628736e-01 -6.98541880e-01 6.79890871e-01 6.97399378e-01
4.43013012e-01 -1.11420178e+00 -5.95863879e-01 3.68171602e-01
-6.06744111e-01 -7.50925541e-01 -9.53952491e-01 -1.14081986e-01
-1.27590144e+00 -6.80767477e-01 -1.00189352e+00 -7.23222017e-01
4.23055470e-01 3.79125029e-01 1.01781797e+00 8.64722580e-02
3.45823057e-02 2.09075287e-01 -3.89352977e-01 -4.38399255e-01
-6.37036383e-01 8.17007661e-01 -1.00023255e-01 -3.45556080e-01
-6.95149004e-02 -6.84512496e-01 -4.89279687e-01 -3.62977475e-01
-1.16671610e+00 1.88082337e-01 8.37720215e-01 8.64243746e-01
3.68975520e-01 -9.78669450e-02 1.05946434e+00 -8.81788671e-01
1.46064782e+00 -1.52628228e-01 4.68445979e-02 4.17612702e-01
-5.82946777e-01 2.17579842e-01 9.31005239e-01 -3.06445926e-01
-1.11870992e+00 -5.84161043e-01 -2.98266530e-01 8.89895856e-03
1.01713955e-01 1.07279432e+00 3.34427468e-02 6.62524223e-01
4.65833455e-01 8.65049958e-01 -3.02406885e-02 -4.84028399e-01
3.47571373e-01 6.54883206e-01 6.03994429e-01 -2.81214446e-01
3.49115789e-01 9.25786123e-02 -1.21941119e-01 -1.15696192e+00
-1.23424911e+00 -6.25157654e-01 -4.36452687e-01 1.74625009e-01
4.75038826e-01 -7.26544678e-01 -3.44484746e-01 3.65372628e-01
-1.44958031e+00 1.93026680e-02 -6.35198891e-01 2.67024130e-01
-6.33537233e-01 1.16213346e+00 -8.26685667e-01 -8.31925273e-01
-1.21566164e+00 -7.75524616e-01 9.99117374e-01 5.13413131e-01
-4.95862186e-01 -9.66617644e-01 2.32477412e-01 1.03397325e-01
4.68132675e-01 1.49505377e-01 1.10294235e+00 -8.42778981e-01
-2.70595491e-01 -3.91078323e-01 -1.80155337e-01 5.44745088e-01
1.63520470e-01 -1.84177667e-01 -5.11947155e-01 -4.16698247e-01
8.70244205e-03 -3.14803332e-01 1.75837529e+00 9.10620689e-01
9.98498976e-01 -8.85420382e-01 -3.02562714e-01 4.42869186e-01
9.12187994e-01 -4.82964367e-02 7.78491378e-01 -6.88848039e-03
7.20801175e-01 5.05390763e-01 3.96110445e-01 6.30784154e-01
2.92442173e-01 1.48245052e-01 9.04330313e-02 1.22739315e-01
-2.29348481e-01 -5.71275532e-01 3.88897479e-01 1.47540414e+00
1.52287453e-01 -7.16555238e-01 -4.86541390e-01 6.42917275e-01
-2.12951732e+00 -1.17488408e+00 -6.88444898e-02 1.92280531e+00
1.06004810e+00 7.71478266e-02 4.31941524e-02 -8.04185942e-02
5.76734900e-01 7.58142710e-01 -7.29586601e-01 -5.92233002e-01
-3.21334600e-01 -1.98881969e-01 1.17129676e-01 4.74932730e-01
-8.41436028e-01 8.07789207e-01 6.91992331e+00 9.55398083e-01
-5.57055295e-01 -4.99087155e-01 6.19417906e-01 -4.35869813e-01
-6.61187232e-01 -2.84079641e-01 -8.63545477e-01 4.10475254e-01
9.76039350e-01 -7.42520154e-01 1.84397116e-01 5.43328464e-01
3.57327729e-01 -2.33007580e-01 -1.14221263e+00 5.62916517e-01
4.50352192e-01 -1.67539084e+00 7.86559761e-01 -1.00438908e-01
1.12701976e+00 -3.21390182e-01 -1.12777017e-01 3.71638596e-01
7.68600851e-02 -9.33928490e-01 4.09269750e-01 6.16163194e-01
7.68409908e-01 -9.45858419e-01 8.65172029e-01 7.38554835e-01
-9.58340168e-01 -1.08380266e-01 -6.33969367e-01 7.32482746e-02
5.66073060e-01 7.59901285e-01 -6.20043457e-01 8.86230290e-01
4.58815368e-03 1.18510020e+00 -5.46102524e-01 1.06653702e+00
-1.29314765e-01 7.39895105e-01 1.76984966e-02 -3.91150147e-01
4.65377450e-01 -2.85335332e-01 9.65059876e-01 1.49674475e+00
3.39772254e-01 -8.01059380e-02 1.77821472e-01 7.13169992e-01
-4.04333949e-01 6.09009936e-02 -4.51693714e-01 -4.79140371e-01
2.68272966e-01 6.75987959e-01 -2.45366976e-01 -5.30407846e-01
7.18964413e-02 9.15300429e-01 3.47620100e-01 4.70610410e-01
-4.37866718e-01 -7.05633938e-01 1.86534345e-01 -1.18653312e-01
2.87374258e-01 -1.17104732e-01 -5.38092792e-01 -1.32891667e+00
1.54413626e-01 -7.68537462e-01 2.78863639e-01 -3.52511078e-01
-9.86585021e-01 4.00903314e-01 9.76695567e-02 -8.08199584e-01
-5.04312158e-01 2.11085677e-01 -1.06465209e+00 6.32028699e-01
-1.54751265e+00 -1.07912838e+00 8.79485253e-03 -1.85309410e-01
1.09668505e+00 -2.01977193e-01 6.34138823e-01 -6.46676719e-02
-7.66696095e-01 6.30383849e-01 6.37860894e-01 -2.35673621e-01
4.39995140e-01 -1.09193635e+00 6.63222313e-01 9.80351686e-01
-1.77966118e-01 7.46620059e-01 8.40755880e-01 -9.32805657e-01
-1.11741245e+00 -1.18250561e+00 1.00465000e+00 -6.56153485e-02
2.55109489e-01 3.68485712e-02 -1.00603640e+00 5.74595332e-01
5.62967122e-01 -1.16993153e+00 6.75252318e-01 1.14217609e-01
-3.78523171e-02 1.28753200e-01 -5.97232521e-01 7.68709004e-01
8.75149786e-01 -1.79122537e-01 -9.64924932e-01 2.50323176e-01
1.00543404e+00 -2.42316619e-01 -3.72464299e-01 3.77348721e-01
4.96321976e-01 -8.46181512e-01 8.12304497e-01 -7.47306645e-01
1.16223371e+00 2.12730497e-01 1.82317466e-01 -1.68963718e+00
-4.20003712e-01 -5.78520119e-01 -6.79343045e-01 1.32809651e+00
4.05619711e-01 -4.08369213e-01 7.07916379e-01 1.57329172e-01
-5.26369274e-01 -8.73517871e-01 -6.58228397e-01 -6.11051440e-01
2.42009893e-01 1.38478577e-01 4.84478772e-01 3.45810443e-01
3.38930786e-01 8.72117639e-01 -5.90274572e-01 -5.44714332e-01
4.76604104e-01 3.60985160e-01 6.58391714e-01 -1.06191945e+00
-1.22571468e-01 -1.02793598e+00 2.52659410e-01 -1.67103195e+00
3.93623769e-01 -8.42892170e-01 2.07398206e-01 -2.31361628e+00
7.31446266e-01 4.00623202e-01 3.33978832e-02 9.00914222e-02
-7.08647132e-01 -2.95666635e-01 1.43989608e-01 2.57230639e-01
-9.63081658e-01 1.26876318e+00 1.40696990e+00 -5.09453833e-01
-5.47827363e-01 2.44411007e-01 -1.23470652e+00 4.41777915e-01
7.41842031e-01 -2.95827031e-01 -4.35474187e-01 -5.19794464e-01
1.31565854e-01 1.85834318e-01 -2.67246485e-01 -8.55054796e-01
3.56987834e-01 -5.22893108e-02 2.67184287e-01 -1.22336328e+00
1.60263449e-01 -4.98940274e-02 -4.43020225e-01 6.23118758e-01
-1.02222717e+00 -1.48855954e-01 2.64623553e-01 8.53093565e-01
-3.57274085e-01 -5.52744806e-01 6.37302577e-01 -2.01263547e-01
6.68965578e-02 1.85782656e-01 -4.98233527e-01 1.59523934e-01
4.90571469e-01 -2.04052404e-01 -3.96991342e-01 -7.03234017e-01
-8.98638442e-02 7.15750098e-01 3.14072520e-01 2.38569081e-01
7.88121998e-01 -1.03532660e+00 -1.25293326e+00 -6.76008761e-02
-1.85360581e-01 2.99792886e-01 5.48900902e-01 6.29589736e-01
-3.68285954e-01 9.02546763e-01 -7.05532134e-02 -3.01738650e-01
-1.17103517e+00 2.99328625e-01 -1.79718837e-01 -9.27174509e-01
-8.63211513e-01 5.79867840e-01 1.04788311e-01 -2.08682027e-02
4.51194167e-01 -4.48369145e-01 -6.48094714e-01 3.18628222e-01
8.69768560e-01 6.46064997e-01 -8.40930790e-02 -1.49893135e-01
1.39731631e-01 4.21468377e-01 -6.18377566e-01 1.01369455e-01
1.40123177e+00 -2.85018981e-01 -1.98365510e-01 3.38471860e-01
1.18654406e+00 -1.28811579e-02 -8.62227201e-01 -2.34797522e-01
2.33366843e-02 -1.36203066e-01 -7.64744505e-02 -5.45907199e-01
-6.88046515e-01 9.29232061e-01 -5.22640169e-01 3.36599499e-01
1.04571021e+00 -2.12710023e-01 1.47004914e+00 8.14863443e-01
-2.59981185e-01 -1.15250838e+00 2.73007244e-01 9.79333222e-01
1.04723549e+00 -7.45957792e-01 5.16954064e-01 -2.19770428e-02
-7.62313902e-01 1.08910704e+00 3.73559326e-01 -1.20781936e-01
-1.16334602e-01 -1.08979113e-01 -5.92625678e-01 -1.01055987e-02
-9.33715105e-01 9.48745832e-02 5.88605404e-01 3.76166314e-01
5.34370124e-01 -9.83117744e-02 -5.41622818e-01 6.91870213e-01
-3.23278010e-01 -2.48330921e-01 7.91940093e-01 6.67694867e-01
-9.33604658e-01 -7.07858443e-01 4.85873409e-02 1.04414785e+00
-4.93420511e-01 -2.75497675e-01 -6.21265054e-01 1.59864500e-01
-9.96556580e-01 1.07984579e+00 7.18758777e-02 -7.95618892e-02
4.69900340e-01 -5.95235899e-02 3.11214924e-01 -8.20293069e-01
-6.46775663e-01 1.69587478e-01 3.24378401e-01 -1.56210244e-01
-1.27864674e-01 -5.99881589e-01 -1.17671061e+00 -4.40229148e-01
-5.34255505e-01 4.25167561e-01 4.67495918e-01 8.50730598e-01
5.50486267e-01 1.01829660e+00 6.33541524e-01 -1.02249873e+00
-1.03379965e+00 -1.44602203e+00 -4.92193639e-01 1.76592782e-01
6.73052728e-01 1.09723225e-01 -2.92469501e-01 -1.28347069e-01] | [12.552239418029785, 9.518001556396484] |
9baf64e8-8472-4e1c-9203-a850caf72cfd | one-comment-from-one-perspective-an-effective | null | null | https://aclanthology.org/2020.coling-main.259 | https://aclanthology.org/2020.coling-main.259.pdf | One Comment from One Perspective: An Effective Strategy for Enhancing Automatic Music Comment | The automatic generation of music comments is of great significance for increasing the popularity of music and the music platform{'}s activity. In human music comments, there exists high distinction and diverse perspectives for the same song. In other words, for a song, different comments stem from different musical perspectives. However, to date, this characteristic has not been considered well in research on automatic comment generation. The existing methods tend to generate common and meaningless comments. In this paper, we propose an effective multi-perspective strategy to enhance the diversity of the generated comments. The experiment results on two music comment datasets show that our proposed model can effectively generate a series of diverse music comments based on different perspectives, which outperforms state-of-the-art baselines by a substantial margin. | ['Jie zhou', 'Jinchao Zhang', 'Zhiqiang Liu', 'Tengfei Huo'] | 2020-12-01 | null | null | null | coling-2020-8 | ['comment-generation'] | ['natural-language-processing'] | [ 1.74340140e-02 -2.16417477e-01 -1.24702774e-01 -1.11257516e-01
-8.40014338e-01 -8.53133261e-01 5.93188822e-01 -5.21223247e-02
2.08839715e-01 6.28252983e-01 9.62248802e-01 2.15610266e-01
1.37136266e-01 -5.34845173e-01 -4.97532897e-02 -7.27732658e-01
6.32296860e-01 2.67100275e-01 -9.24777053e-03 -5.60145855e-01
6.83448851e-01 -2.19153747e-01 -1.52736211e+00 6.27507567e-01
7.79700041e-01 5.06488621e-01 1.63824931e-01 5.19629538e-01
-1.96749642e-01 4.56245720e-01 -1.00072730e+00 -6.81056857e-01
2.72359312e-01 -1.01263642e+00 -3.42847705e-01 -2.01435462e-02
3.71565342e-01 -2.37246320e-01 1.19069800e-01 9.57444787e-01
9.17381704e-01 -5.53393066e-02 6.66593909e-01 -1.20669985e+00
-6.92205191e-01 1.40649509e+00 -4.39347893e-01 -1.96653903e-01
5.19782603e-01 -1.76295176e-01 1.54343092e+00 -9.16642308e-01
6.84990048e-01 9.71227765e-01 4.42538232e-01 6.40040100e-01
-8.70631993e-01 -9.10605431e-01 1.51042566e-01 -1.52466714e-01
-1.11269987e+00 -2.74296135e-01 1.04980755e+00 -4.97093409e-01
1.11948602e-01 5.58619499e-01 1.00013244e+00 1.17354739e+00
4.18656543e-02 1.05398381e+00 8.43909264e-01 -1.23206802e-01
7.08657131e-03 2.71273106e-01 -2.45147511e-01 -2.29001582e-01
2.03911051e-01 -3.27734858e-01 -9.51489151e-01 -3.13749373e-01
6.25236630e-01 -2.43306801e-01 -2.44506612e-01 1.54691532e-01
-1.62262475e+00 6.94561124e-01 2.28080843e-02 3.35859627e-01
-2.69827306e-01 -9.34448391e-02 6.59904599e-01 1.99751556e-02
5.73236704e-01 6.59421504e-01 -1.11920372e-01 -4.77678835e-01
-1.08903134e+00 8.24992359e-01 8.17280591e-01 1.03989470e+00
1.59838781e-01 9.93841365e-02 -5.76038718e-01 9.15301085e-01
2.29641408e-01 5.10889173e-01 6.17521524e-01 -8.09274733e-01
4.50636655e-01 7.27502763e-01 3.51963371e-01 -1.16243815e+00
1.95936523e-02 -7.44012594e-01 -9.75839972e-01 -1.23742543e-01
2.96985924e-01 -3.50093693e-01 -5.62616661e-02 1.57054257e+00
9.31825489e-02 1.21730380e-01 1.04889767e-02 7.81382263e-01
9.19571102e-01 7.77252793e-01 -3.61892998e-01 -3.47679913e-01
1.03373170e+00 -1.15318966e+00 -7.10578144e-01 6.03346936e-02
1.45374760e-01 -1.43381858e+00 1.18324399e+00 6.98670506e-01
-1.25262368e+00 -7.49189198e-01 -8.27315867e-01 1.03530422e-01
3.93931389e-01 3.65652323e-01 5.27227759e-01 3.70516211e-01
-7.12976038e-01 3.51275921e-01 6.36595022e-03 -7.32501745e-02
1.58940554e-01 -1.90621585e-01 3.12299486e-02 3.39572191e-01
-1.00398743e+00 1.88139342e-02 2.96379715e-01 -2.45738000e-01
-6.31672323e-01 -9.65606272e-01 -1.80107981e-01 -1.53561803e-02
7.75702968e-02 -7.89304078e-01 1.52031767e+00 -1.20228386e+00
-1.55541813e+00 6.22721374e-01 -2.19383001e-01 -2.76605934e-01
6.91302478e-01 -4.11796778e-01 -4.26793665e-01 -1.32404640e-01
2.92545944e-01 6.54828250e-01 9.60837722e-01 -1.43367553e+00
-7.05926836e-01 5.40210269e-02 1.50150314e-01 4.65746224e-01
-5.73917866e-01 2.06463620e-01 -4.29348379e-01 -1.24823713e+00
9.97574162e-03 -1.26527917e+00 -4.28118370e-02 -3.32047194e-01
-7.10105419e-01 -2.84287512e-01 3.83099228e-01 -3.54399890e-01
1.77927983e+00 -2.46164441e+00 9.19218063e-02 -1.47292376e-01
-7.33567029e-02 9.66902450e-02 4.08903658e-02 9.66718793e-01
8.70567858e-02 2.33808637e-01 -9.87381935e-02 -9.98139307e-02
1.65066347e-01 -2.18430206e-01 -7.84528792e-01 -1.93228275e-01
-4.65666801e-01 8.06014597e-01 -1.07512295e+00 -3.61892790e-01
-3.41353357e-01 2.23475903e-01 -7.96922028e-01 2.55621850e-01
-4.20639724e-01 6.24510229e-01 -3.92743111e-01 2.24349096e-01
3.74478400e-01 -3.05367887e-01 1.34847969e-01 -2.26568356e-02
-2.21893191e-01 5.60882986e-01 -1.23958719e+00 1.88857853e+00
-2.37132534e-01 4.70146567e-01 -3.52504909e-01 -1.62870988e-01
1.02556741e+00 6.04215622e-01 4.14837658e-01 -1.33656815e-01
1.17791750e-01 3.65692496e-01 2.56183833e-01 -2.20204502e-01
1.12929189e+00 -4.17976677e-01 -2.99830914e-01 1.03220773e+00
-4.98855144e-01 -2.24922761e-01 4.59275365e-01 4.34270531e-01
4.47325498e-01 1.30192062e-03 4.55172598e-01 -1.49525777e-01
4.90409315e-01 -3.68646905e-02 4.33365852e-01 5.94667196e-01
1.52037263e-01 1.01096106e+00 3.76371920e-01 2.43663434e-02
-1.03723574e+00 -9.03848708e-01 2.06073612e-01 1.15740156e+00
1.57320783e-01 -8.35943878e-01 -5.62510669e-01 -4.57414180e-01
-2.04602793e-01 6.41424179e-01 -3.61996025e-01 4.35316078e-02
-5.61088741e-01 -4.33818668e-01 6.35629892e-01 3.66950780e-01
3.07860255e-01 -1.35750163e+00 -2.21351385e-01 3.69836658e-01
-6.56187534e-01 -6.51450455e-01 -1.03898108e+00 -6.55300915e-01
-7.58988738e-01 -7.59144127e-01 -1.06934524e+00 -7.44932294e-01
4.55042720e-01 7.18565345e-01 1.37416780e+00 -3.91362682e-02
2.04737902e-01 5.16237970e-03 -7.85791516e-01 -7.32860863e-01
-6.43386722e-01 2.81545788e-01 -4.74298606e-03 8.36982504e-02
1.35181293e-01 -6.98334038e-01 -7.29137897e-01 2.74627239e-01
-1.10759509e+00 3.97079945e-01 4.33478832e-01 6.17915213e-01
3.90360266e-01 -7.12826103e-02 1.00511873e+00 -1.04123533e+00
1.21253145e+00 -5.03416896e-01 8.47163722e-02 -1.62256241e-01
-4.55066681e-01 -3.30977112e-01 7.81721413e-01 -5.19534945e-01
-1.11759758e+00 -4.18575913e-01 -1.46059796e-01 8.99127573e-02
-1.33030647e-02 6.50774360e-01 -6.51663169e-02 6.14711285e-01
6.36731863e-01 3.72312337e-01 -3.31455559e-01 -3.53871465e-01
4.62643594e-01 8.91583145e-01 3.26607525e-01 -4.85677719e-01
8.43155861e-01 5.04419744e-01 -3.92934948e-01 -5.30063272e-01
-1.11569381e+00 -6.85154557e-01 -2.85358667e-01 -4.23438102e-01
5.22051096e-01 -1.03339875e+00 -3.44847798e-01 5.55359483e-01
-1.18850148e+00 9.92033631e-02 -4.88174349e-01 3.52348387e-01
-4.43823993e-01 3.65082532e-01 -5.54998279e-01 -7.55208313e-01
-6.82759762e-01 -7.91942358e-01 1.09474051e+00 4.17523414e-01
-7.71314502e-01 -7.19304979e-01 2.87097007e-01 4.65665579e-01
2.75657028e-01 -4.46957387e-02 6.26379550e-01 -5.85793495e-01
-5.57295501e-01 -2.49850620e-02 7.32923076e-02 3.61960053e-01
3.90941858e-01 1.64841115e-01 -9.08233345e-01 -5.30805998e-02
-1.32717296e-01 -1.59307510e-01 6.62722468e-01 1.83511019e-01
7.55618393e-01 -4.61457729e-01 7.47124329e-02 2.26720318e-01
1.03274536e+00 1.22559913e-01 6.29426777e-01 1.32742509e-01
5.54343104e-01 4.95192051e-01 8.51628244e-01 9.29768026e-01
3.01685661e-01 6.49685979e-01 2.35231981e-01 2.75442958e-01
-9.92426947e-02 -6.80041313e-01 6.35414243e-01 1.62068129e+00
1.85249988e-02 -5.73520958e-01 -3.57359856e-01 8.11104059e-01
-1.95756888e+00 -1.33213019e+00 -3.51134419e-01 2.22837400e+00
9.00788546e-01 5.63642122e-02 3.36859345e-01 3.41212749e-01
8.11630845e-01 4.65465456e-01 -4.01743114e-01 -2.07354859e-01
-2.35858962e-01 -1.11939281e-01 -2.39506766e-01 7.61755183e-02
-7.23459423e-01 7.63150692e-01 6.33898067e+00 9.04413462e-01
-1.03361225e+00 -1.11665137e-01 2.66398758e-01 -2.94445574e-01
-8.72181416e-01 -4.38434584e-03 -8.70452464e-01 5.68970442e-01
3.09854746e-01 -8.16527843e-01 1.64625704e-01 8.26189458e-01
5.04598737e-01 3.32022309e-01 -8.23030531e-01 1.24326384e+00
4.86136913e-01 -1.34046447e+00 5.30978024e-01 1.58601608e-02
1.37337661e+00 -4.15975899e-01 1.67989343e-01 5.58453277e-02
9.43012908e-02 -7.11875439e-01 9.78126526e-01 3.57483983e-01
4.33539361e-01 -7.79833198e-01 4.83793408e-01 4.82658982e-01
-1.23269904e+00 8.64041001e-02 -2.37152547e-01 -2.00404972e-01
6.48621857e-01 7.85690188e-01 -7.12123275e-01 7.01984286e-01
2.56225348e-01 1.08398998e+00 -4.09418046e-01 1.17593408e+00
-4.68026489e-01 8.98220539e-01 1.49089396e-01 -3.43471199e-01
2.75023710e-02 -4.58950341e-01 8.59360397e-01 9.91024911e-01
7.79416323e-01 -5.52196093e-02 8.31038803e-02 7.78979063e-01
-2.78953761e-01 6.59730077e-01 -4.76161569e-01 -2.94539988e-01
5.02970874e-01 1.27394712e+00 -4.73672718e-01 -3.76748413e-01
-1.21536233e-01 9.63419795e-01 -1.85041159e-01 1.68179095e-01
-7.75855541e-01 2.46045403e-02 6.33171976e-01 2.62772739e-01
-3.34230624e-02 -9.42115039e-02 -4.59392488e-01 -1.23284173e+00
1.21655099e-01 -1.17007864e+00 2.56162167e-01 -9.38735485e-01
-1.64744723e+00 5.43168962e-01 -2.99608767e-01 -1.94757938e+00
-3.66177589e-01 1.17591880e-01 -8.08072031e-01 7.47287750e-01
-1.06187034e+00 -1.10166168e+00 -2.38259584e-01 3.56425196e-01
9.67994690e-01 -2.87535161e-01 7.95940518e-01 3.41441184e-01
9.09032393e-03 5.19328177e-01 4.52838503e-02 -4.99887057e-02
1.29205298e+00 -1.15611386e+00 4.75749761e-01 6.97470546e-01
5.79140961e-01 8.31119657e-01 7.48327315e-01 -6.59474134e-01
-1.01291108e+00 -9.15085912e-01 1.29484856e+00 -3.79694343e-01
6.54529989e-01 -5.42081073e-02 -5.34710050e-01 3.43930572e-01
6.17088437e-01 -9.27411199e-01 1.32917607e+00 1.84209719e-01
-3.19025964e-01 -2.19327603e-02 -4.53217179e-01 1.07210898e+00
1.03552210e+00 -3.80885214e-01 -5.49331665e-01 2.42369533e-01
6.71618044e-01 -2.48793140e-01 -5.80931067e-01 1.62683859e-01
8.70219529e-01 -1.37681520e+00 7.28277206e-01 -3.88127983e-01
9.23341036e-01 -5.45023859e-01 1.07394811e-02 -1.51547480e+00
-2.67300040e-01 -9.32417452e-01 1.85373068e-01 1.70530152e+00
5.22464991e-01 -1.88340634e-01 6.83704257e-01 -2.94978563e-02
-2.22929850e-01 -4.87649977e-01 -3.61596525e-01 -4.08152103e-01
-5.48853390e-02 -4.66011792e-01 7.23532319e-01 9.34895337e-01
1.38293535e-01 6.57675445e-01 -7.81771362e-01 -4.02914822e-01
1.04616545e-01 6.75883174e-01 1.26286399e+00 -1.35025525e+00
-5.77814639e-01 -7.37146139e-01 -3.77457999e-02 -1.36365974e+00
-4.48504724e-02 -9.86528039e-01 2.79875174e-02 -1.68653321e+00
5.36953926e-01 -5.03903389e-01 -1.73999369e-01 -1.29063383e-01
-5.53943515e-01 4.74926472e-01 5.09399295e-01 5.41590273e-01
-4.88728195e-01 5.48611939e-01 1.67466640e+00 -7.63160922e-03
-3.38287443e-01 3.47473949e-01 -1.35131013e+00 7.97398388e-01
1.00748062e+00 -3.82822007e-01 -7.21799016e-01 -2.60991991e-01
1.03899515e+00 -1.14915445e-01 -1.00968085e-01 -8.44575346e-01
5.95902652e-02 -1.55037507e-01 -2.27102518e-01 -7.27038383e-01
2.08684921e-01 -3.47672403e-01 3.78015220e-01 8.82174894e-02
-5.71163714e-01 1.72256306e-01 -2.49024078e-01 5.52000761e-01
-5.45526147e-01 -2.01865524e-01 4.44722921e-01 -2.66930252e-01
-1.00419506e-01 2.07495317e-01 -2.35680014e-01 4.75506812e-01
6.07882082e-01 -3.24040726e-02 -2.12050170e-01 -8.29427838e-01
-1.82480127e-01 -2.58043349e-01 5.99682629e-01 9.79005933e-01
4.36088860e-01 -1.61889982e+00 -1.25309563e+00 -4.24912535e-02
4.23423290e-01 -1.07337497e-01 2.90774256e-01 5.28398275e-01
-2.79629618e-01 9.32621732e-02 1.37213534e-02 -4.14645761e-01
-1.38936865e+00 1.43674895e-01 -3.49814206e-01 -3.17114085e-01
-4.69539046e-01 8.16750586e-01 3.30842942e-01 -3.28351766e-01
4.08221073e-02 -1.37432262e-01 -4.88075316e-01 4.41020221e-01
8.50618482e-01 3.33382964e-01 -3.62102211e-01 -8.33326995e-01
2.68645585e-01 4.80576545e-01 1.51712865e-01 -4.20355380e-01
9.20042157e-01 -1.13543548e-01 -2.38691643e-01 1.00936556e+00
5.22940397e-01 9.25561547e-01 -7.89040983e-01 4.93390001e-02
-2.83008248e-01 -6.43714070e-01 -6.38581097e-01 -6.92190111e-01
-8.20272624e-01 6.90651596e-01 -1.17388703e-01 5.45019984e-01
9.04291391e-01 -1.83737889e-01 1.19858468e+00 -6.78607374e-02
4.54414845e-01 -1.00544035e+00 3.13791066e-01 5.09755075e-01
1.14188969e+00 -9.07608211e-01 -2.18804497e-02 -4.88517463e-01
-1.11126137e+00 9.42576110e-01 4.47674632e-01 -1.56017944e-01
4.05686319e-01 5.53965271e-02 5.05242884e-01 1.80060357e-01
-8.43620598e-01 7.91987777e-02 4.12301570e-01 2.84111589e-01
9.42988455e-01 2.18268350e-01 -9.29054379e-01 8.18966568e-01
-7.61656940e-01 -6.23670369e-02 8.29484999e-01 4.20762867e-01
-5.17558396e-01 -1.54584658e+00 -2.56879568e-01 3.61808687e-01
-6.39734566e-01 -4.39222664e-01 -7.53838181e-01 2.74008423e-01
1.67995438e-01 1.22593904e+00 -3.56742024e-01 -3.93502027e-01
5.79552464e-02 1.23308048e-01 1.38959885e-01 -7.82662272e-01
-1.10056603e+00 5.09612918e-01 1.45756349e-01 4.70049158e-02
-7.40332365e-01 -7.99086988e-01 -1.08643699e+00 -2.37119630e-01
-3.07860851e-01 3.76133293e-01 3.58130664e-01 6.55007064e-01
4.54533517e-01 5.96470535e-01 8.29356849e-01 -6.72671795e-01
-4.33736861e-01 -1.16675496e+00 -7.98045337e-01 7.38879085e-01
-2.81428933e-01 -1.25038415e-01 -4.16936815e-01 3.14835578e-01] | [15.726689338684082, 5.7285075187683105] |
e2929f9f-0a36-45f9-87aa-9ae621699ef8 | bilinear-factor-matrix-norm-minimization-for | 1810.05186 | null | http://arxiv.org/abs/1810.05186v1 | http://arxiv.org/pdf/1810.05186v1.pdf | Bilinear Factor Matrix Norm Minimization for Robust PCA: Algorithms and Applications | The heavy-tailed distributions of corrupted outliers and singular values of
all channels in low-level vision have proven effective priors for many
applications such as background modeling, photometric stereo and image
alignment. And they can be well modeled by a hyper-Laplacian. However, the use
of such distributions generally leads to challenging non-convex, non-smooth and
non-Lipschitz problems, and makes existing algorithms very slow for large-scale
applications. Together with the analytic solutions to lp-norm minimization with
two specific values of p, i.e., p=1/2 and p=2/3, we propose two novel bilinear
factor matrix norm minimization models for robust principal component analysis.
We first define the double nuclear norm and Frobenius/nuclear hybrid norm
penalties, and then prove that they are in essence the Schatten-1/2 and 2/3
quasi-norms, respectively, which lead to much more tractable and scalable
Lipschitz optimization problems. Our experimental analysis shows that both our
methods yield more accurate solutions than original Schatten quasi-norm
minimization, even when the number of observations is very limited. Finally, we
apply our penalties to various low-level vision problems, e.g., text removal,
moving object detection, image alignment and inpainting, and show that our
methods usually outperform the state-of-the-art methods. | ['Zhi-Quan Luo', 'Fanhua Shang', 'Zhouchen Lin', 'Yuanyuan Liu', 'James Cheng'] | 2018-10-11 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [ 2.67465472e-01 -3.79328579e-01 1.14226751e-02 -9.86343920e-02
-9.79775250e-01 -4.00128424e-01 4.04282957e-01 -2.76742786e-01
-3.67630720e-01 6.89292908e-01 5.69280311e-02 4.29822840e-02
-3.96197885e-02 -1.43039584e-01 -9.07704294e-01 -1.12091398e+00
2.26435751e-01 1.09817691e-01 1.73940673e-01 7.03981444e-02
1.83384567e-01 4.97666240e-01 -1.53283596e+00 -1.76461682e-01
1.00979090e+00 9.94112194e-01 1.53057547e-02 5.74562430e-01
8.14544484e-02 5.92443585e-01 -2.51719773e-01 -3.52212638e-01
5.88243783e-01 -3.64069700e-01 -2.43829980e-01 6.82581365e-01
8.38298678e-01 -2.21788913e-01 -4.06946152e-01 1.68029118e+00
4.72269952e-01 3.04386884e-01 5.89250743e-01 -1.25470638e+00
-5.78490794e-01 -6.02009222e-02 -1.30394077e+00 6.64302260e-02
1.37763575e-01 4.93622981e-02 5.87969303e-01 -1.17842841e+00
4.59569067e-01 1.32402289e+00 9.48714256e-01 2.43182316e-01
-1.35287309e+00 -4.22147185e-01 1.46724209e-01 1.68852776e-01
-1.52896118e+00 -6.15171969e-01 8.45043361e-01 -5.92524946e-01
3.69842738e-01 3.80926430e-01 2.67088503e-01 8.11154306e-01
5.61950058e-02 8.32631350e-01 1.34216225e+00 -3.58461082e-01
1.03284121e-01 -8.15940052e-02 3.90688106e-02 6.70693755e-01
4.25210267e-01 -3.47458094e-01 -4.55862850e-01 -4.53350902e-01
1.05008233e+00 1.89989284e-01 -7.21101522e-01 -3.95053804e-01
-1.43729579e+00 7.65307665e-01 -1.68285474e-01 3.31599936e-02
-4.55929309e-01 8.60008821e-02 1.79603845e-01 -8.81859809e-02
5.42018414e-01 -1.59987092e-01 -2.27024674e-01 9.80951786e-02
-1.08356333e+00 2.14351848e-01 6.36916161e-01 1.10829473e+00
8.84969473e-01 1.24288857e-01 -1.82959601e-01 1.05018866e+00
3.53517532e-01 8.98042798e-01 1.73211142e-01 -1.25434947e+00
2.66243339e-01 9.77870300e-02 4.23062265e-01 -1.25748575e+00
-2.15752274e-01 -2.31828406e-01 -1.27770793e+00 1.42434582e-01
8.02679837e-01 1.24689065e-01 -7.34546900e-01 1.75573754e+00
4.39626098e-01 5.91836929e-01 -1.18559569e-01 1.05190098e+00
3.78999442e-01 6.77170157e-01 -5.55774510e-01 -8.45777094e-01
1.18737733e+00 -8.18441927e-01 -9.76516545e-01 -2.42174298e-01
1.33092567e-01 -1.02248287e+00 1.12755954e+00 5.71629882e-01
-1.15881288e+00 -3.36432099e-01 -7.97574401e-01 -1.32190034e-01
2.71331757e-01 7.54913613e-02 3.31968188e-01 4.05587971e-01
-7.98483431e-01 4.41984445e-01 -8.44441891e-01 -7.96427503e-02
1.36384115e-01 6.65555820e-02 -4.90861982e-01 -3.40133697e-01
-5.29627442e-01 5.93898535e-01 -2.12708354e-01 3.29818457e-01
-6.42964959e-01 -4.93455201e-01 -7.11535096e-01 -2.69179583e-01
4.64497447e-01 -4.53371525e-01 6.53694928e-01 -8.73516619e-01
-1.44873309e+00 1.10984552e+00 -5.13603151e-01 -1.47251427e-01
7.13482201e-01 -3.09656441e-01 -1.06609814e-01 7.20449314e-02
1.70421094e-01 1.42664820e-01 1.31201816e+00 -1.23748338e+00
-3.31985474e-01 -5.78059971e-01 -2.69891649e-01 2.89128095e-01
-3.75533849e-01 2.47924328e-01 -9.67427433e-01 -8.31536770e-01
4.49746639e-01 -9.32553291e-01 -3.66606832e-01 2.73171484e-01
-3.66599798e-01 1.49580464e-01 6.09584987e-01 -1.04570913e+00
9.02202845e-01 -2.37362123e+00 3.05898905e-01 2.01408893e-01
1.62727386e-01 5.18072508e-02 -7.26921186e-02 -4.33957241e-02
-2.90154051e-02 -2.25649163e-01 -5.28316975e-01 -3.67231011e-01
-2.94770114e-02 2.95992255e-01 -2.29012281e-01 1.15088701e+00
-8.39500502e-02 4.76297945e-01 -9.36013281e-01 -4.06898528e-01
2.36025110e-01 7.08898902e-01 -3.07764560e-01 -2.38992237e-02
2.05484152e-01 5.80013096e-01 -2.33572721e-01 4.99464124e-01
1.04138923e+00 -5.91644086e-02 -2.89063543e-01 -4.45102602e-01
-1.55347511e-01 -3.90563369e-01 -1.64763939e+00 1.59960556e+00
-1.94490716e-01 7.92650759e-01 8.47112298e-01 -9.76443291e-01
6.60873890e-01 6.38040751e-02 5.52990139e-01 -3.64852071e-01
-2.63765994e-02 3.41383785e-01 -4.19795841e-01 -4.21736896e-01
2.21418440e-01 -1.55288398e-01 4.46594954e-01 -7.17725381e-02
-3.92244548e-01 -2.70237148e-01 4.21692133e-01 2.17255503e-02
9.71411467e-01 -1.72410272e-02 1.60774708e-01 -4.96009588e-01
7.51809359e-01 -5.02969563e-01 1.09812498e+00 6.23521030e-01
-2.83185244e-01 1.06623399e+00 5.53049862e-01 -1.91414177e-01
-9.66217935e-01 -8.35085869e-01 -3.53326917e-01 8.04785728e-01
2.57250130e-01 -2.41995797e-01 -9.08353686e-01 -2.86821276e-01
-2.99857128e-02 3.98004711e-01 -2.83808649e-01 1.48884177e-01
-5.68569005e-01 -1.15669835e+00 3.84010792e-01 3.88256729e-01
6.03232443e-01 -4.60187346e-01 -2.40086615e-01 7.42778415e-03
-3.50885272e-01 -1.33555067e+00 -9.57922459e-01 -9.38606262e-02
-9.00310695e-01 -1.10145044e+00 -1.08188868e+00 -6.37738645e-01
9.92185175e-01 5.32652557e-01 9.51128602e-01 -3.20823491e-03
-1.80998251e-01 6.64520681e-01 -5.11770248e-02 -1.30081683e-01
2.81905261e-04 -8.13194156e-01 5.18646181e-01 5.27033925e-01
1.00093327e-01 -5.57862639e-01 -6.33919537e-01 6.57541394e-01
-1.00386238e+00 -9.82733369e-02 4.41933423e-01 1.07351041e+00
1.14107656e+00 1.97512403e-01 -1.16324440e-01 -7.96642661e-01
5.65439701e-01 8.29174444e-02 -9.65296745e-01 2.33715519e-01
-4.18810546e-01 -1.51198253e-01 5.48491120e-01 -7.36368656e-01
-9.16752517e-01 2.52497017e-01 1.11252703e-01 -9.58344698e-01
6.72177151e-02 2.94741839e-01 -3.69790822e-01 -4.20207471e-01
5.79001606e-01 3.58248949e-01 -3.59436795e-02 -3.78323764e-01
4.86599386e-01 4.25003260e-01 9.64614391e-01 -5.95816314e-01
1.04464376e+00 8.75255764e-01 1.90591604e-01 -1.35602212e+00
-7.68255293e-01 -7.56782711e-01 -4.51093912e-01 -5.82348779e-02
7.02215195e-01 -8.04146647e-01 -6.07781649e-01 9.09733534e-01
-1.11432064e+00 -1.84823081e-01 -3.30943704e-01 4.38538641e-01
-6.99208319e-01 1.03038907e+00 -5.65015674e-01 -1.02901816e+00
-1.23752728e-02 -1.27058160e+00 1.01536000e+00 2.57828422e-02
3.79514337e-01 -9.95768130e-01 -9.37754586e-02 3.77753079e-01
2.70912260e-01 4.57946509e-01 4.61702824e-01 3.25406715e-02
-4.26196188e-01 -2.40482062e-01 -3.40581983e-01 7.76603580e-01
-8.42525810e-02 -1.23973275e-02 -8.45900714e-01 -5.21791935e-01
6.20301306e-01 -4.95831370e-02 8.16001296e-01 7.40535140e-01
1.18161416e+00 -5.32131314e-01 -7.00714514e-02 1.14099848e+00
1.26921654e+00 -1.69747666e-01 7.52015889e-01 7.22320303e-02
9.20798123e-01 5.96285641e-01 4.99268681e-01 5.40628612e-01
5.57767972e-02 6.96074665e-01 4.28093165e-01 -1.11141592e-01
-2.42838096e-02 3.02320004e-01 5.53610027e-01 1.06283164e+00
-2.37023845e-01 1.81170285e-01 -7.65693009e-01 4.39331949e-01
-2.15651917e+00 -8.76571953e-01 -6.22235894e-01 2.71318579e+00
8.86499107e-01 -1.42386898e-01 -1.48971707e-01 2.23384276e-02
8.57547581e-01 1.95278719e-01 -6.80776060e-01 2.30961069e-01
-6.51561439e-01 -1.04914308e-01 1.02001619e+00 6.98155880e-01
-1.22399414e+00 6.92111850e-01 6.17286348e+00 1.11742485e+00
-8.88988972e-01 2.86460131e-01 5.50217092e-01 -1.26477731e-02
-8.24661553e-02 -1.30107207e-02 -6.90208197e-01 5.82060099e-01
4.11293596e-01 -7.73745552e-02 7.32754171e-01 7.28703916e-01
3.02863896e-01 -1.16581894e-01 -8.58674347e-01 1.66434538e+00
4.38885778e-01 -9.80490029e-01 -2.92464614e-01 1.53072730e-01
1.01862454e+00 -1.66468490e-02 2.68375576e-02 -2.61949599e-01
-5.46482913e-02 -8.31063867e-01 6.46653295e-01 4.80346739e-01
6.62995696e-01 -5.56263685e-01 5.56890011e-01 4.36020106e-01
-1.04455125e+00 1.42762780e-01 -8.77370059e-01 1.75935343e-01
2.35015586e-01 1.16730845e+00 1.52955547e-01 2.64400393e-01
7.91281760e-01 7.89369583e-01 -1.31603807e-01 1.23643219e+00
-2.59251535e-01 4.60901439e-01 -5.98290563e-01 4.95592654e-01
-6.95239827e-02 -9.27792609e-01 9.13306057e-01 1.11188769e+00
3.44015986e-01 2.20419973e-01 3.02900225e-01 6.28604949e-01
-9.20963064e-02 1.84258655e-01 -4.20151502e-01 2.66839653e-01
1.00623071e-01 1.34305418e+00 -7.14738727e-01 -2.46479824e-01
-7.66179323e-01 1.20408428e+00 -2.39383597e-02 8.51037621e-01
-8.23676825e-01 -2.41479546e-01 8.71272206e-01 1.39104232e-01
2.58472353e-01 -4.78037506e-01 -2.39588186e-01 -1.57820618e+00
3.83849680e-01 -1.13979983e+00 1.33284822e-01 -6.56849802e-01
-1.53696668e+00 2.56697327e-01 -1.34018913e-01 -1.08563566e+00
1.57007426e-01 -8.53431582e-01 -6.79719329e-01 8.28681886e-01
-1.47175491e+00 -9.89171386e-01 -4.13609684e-01 9.42149341e-01
3.79916549e-01 6.02918118e-02 4.44015712e-01 4.34069425e-01
-7.80227065e-01 3.70177269e-01 6.90229595e-01 2.35211495e-02
9.40709770e-01 -1.19421387e+00 2.53409129e-02 1.36937714e+00
-4.38380847e-03 6.42946005e-01 9.14220512e-01 -5.05914629e-01
-1.78411365e+00 -9.03870761e-01 3.15664411e-01 -3.16241413e-01
6.77454114e-01 -2.34962657e-01 -1.09331989e+00 6.65818512e-01
-1.01615824e-01 3.78717542e-01 3.09246272e-01 -3.32782388e-01
-2.51133949e-01 -1.30064711e-01 -1.02763557e+00 5.33364594e-01
9.85761583e-01 -3.70719612e-01 -2.03259021e-01 7.87078321e-01
2.28855729e-01 -5.34691334e-01 -6.69226825e-01 3.51828545e-01
3.08099508e-01 -1.05088079e+00 1.12885606e+00 -3.26160938e-01
-7.56145567e-02 -5.96826017e-01 -4.41415906e-01 -9.86937165e-01
-4.12499338e-01 -1.12472165e+00 -1.66409716e-01 1.18505752e+00
5.19418158e-02 -7.06982732e-01 6.14325225e-01 6.44130170e-01
-7.93167502e-02 -5.13802707e-01 -1.08376980e+00 -8.96976769e-01
-1.84798583e-01 -5.18554807e-01 -1.48087248e-01 1.09918439e+00
-5.02247632e-01 -3.60192172e-02 -8.21625173e-01 3.32608461e-01
1.25110424e+00 -1.20629519e-01 9.94793594e-01 -1.14546931e+00
-3.76380563e-01 -5.01619637e-01 -4.70705003e-01 -1.40848053e+00
1.98787600e-01 -5.50408185e-01 3.37162793e-01 -1.23607957e+00
2.78948963e-01 -1.19783908e-01 -1.26513109e-01 2.49850273e-01
-3.42465967e-01 3.33989888e-01 -3.98094617e-02 5.02438128e-01
-3.86760652e-01 5.95700979e-01 1.02830124e+00 -3.44496161e-01
-1.86607346e-01 8.43424872e-02 -4.25297976e-01 1.25831330e+00
2.51705438e-01 -2.74601430e-01 -1.87001675e-01 -4.83214378e-01
3.42253834e-01 -1.59665585e-01 2.86015064e-01 -7.82324791e-01
3.08293521e-01 -3.21852237e-01 3.28640193e-01 -3.36522073e-01
4.20700550e-01 -8.51913929e-01 5.00827171e-02 5.59154153e-03
1.57505825e-01 -6.53593838e-02 -9.93336271e-03 7.39565492e-01
-3.11462015e-01 -2.07093328e-01 1.06897795e+00 1.33279953e-02
-3.71774077e-01 4.03212339e-01 -1.74807742e-01 2.20978409e-01
8.63414168e-01 -3.17721903e-01 -1.37972623e-01 -7.08906054e-01
-5.08652747e-01 1.49679393e-01 7.29271054e-01 -1.92051679e-02
7.20123708e-01 -1.42683363e+00 -9.65834081e-01 3.36763471e-01
-2.30982170e-01 1.89767078e-01 2.07552269e-01 1.65561426e+00
-5.77784061e-01 -1.66436985e-01 -4.32007872e-02 -8.90038610e-01
-1.24233639e+00 6.52928352e-01 1.91825166e-01 -1.71389375e-02
-7.40646183e-01 9.32286620e-01 6.59652293e-01 -2.96904612e-02
4.89118546e-01 -2.95557618e-01 2.06123531e-01 -9.92124975e-02
7.48085380e-01 8.02022696e-01 -9.90819186e-02 -8.89420331e-01
-4.10680771e-01 1.02673411e+00 1.63345844e-01 -8.57704505e-02
1.17587793e+00 -1.81872413e-01 -5.75171769e-01 3.69920701e-01
1.07645416e+00 3.91998291e-01 -1.65182507e+00 -3.16236794e-01
-1.16994172e-01 -8.42198372e-01 1.03191346e-01 -1.81616209e-02
-1.26368451e+00 9.36871350e-01 4.09460157e-01 1.24555133e-01
1.34651589e+00 -3.29596311e-01 6.39632702e-01 2.59727329e-01
1.78675994e-01 -1.22930062e+00 -8.02273825e-02 5.18267930e-01
1.09730947e+00 -1.09830213e+00 2.53115684e-01 -7.08356440e-01
-4.15667713e-01 9.30735826e-01 3.38554144e-01 -1.41603872e-01
5.43980837e-01 2.88986742e-01 8.46366435e-02 2.97766060e-01
-1.76979393e-01 -1.81329101e-01 4.37190115e-01 5.75420558e-01
2.99457610e-01 -1.91187337e-01 -2.78686732e-01 2.39379182e-01
1.40647694e-01 -2.95317769e-01 5.56225240e-01 6.54725552e-01
-3.35746318e-01 -7.84196138e-01 -1.00113380e+00 3.76527101e-01
-5.18346846e-01 -6.27076402e-02 -1.61641270e-01 4.09070492e-01
-1.10323131e-01 8.66942644e-01 -2.42183089e-01 7.18685240e-02
3.49890471e-01 -4.75274846e-02 5.41678548e-01 -1.98778480e-01
1.56527206e-01 7.22427785e-01 -3.74415606e-01 -7.65295386e-01
-5.66648722e-01 -7.88086653e-01 -9.04440045e-01 -3.26627702e-01
-4.75375086e-01 -9.90434512e-02 5.78520477e-01 6.89438522e-01
2.09842712e-01 -5.46629280e-02 4.58198339e-01 -1.11048090e+00
-7.23533750e-01 -8.65337789e-01 -8.46536338e-01 7.40095198e-01
4.38441396e-01 -6.62857354e-01 -7.41471410e-01 3.95602912e-01] | [7.677067279815674, 4.30270528793335] |
412d5010-4463-400c-887c-046b71c7ce71 | thai-nested-named-entity-recognition-corpus | null | null | https://aclanthology.org/2022.findings-acl.116 | https://aclanthology.org/2022.findings-acl.116.pdf | Thai Nested Named Entity Recognition Corpus | This paper presents the first Thai Nested Named Entity Recognition (N-NER) dataset. Thai N-NER consists of 264,798 mentions, 104 classes, and a maximum depth of 8 layers obtained from 4,894 documents in the domains of news articles and restaurant reviews. Our work, to the best of our knowledge, presents the largest non-English N-NER dataset and the first non-English one with fine-grained classes. To understand the new challenges our proposed dataset brings to the field, we conduct an experimental study on (i) cutting edge N-NER models with the state-of-the-art accuracy in English and (ii) baseline methods based on well-known language model architectures. From the experimental results, we obtained two key findings. First, all models produced poor F1 scores in the tail region of the class distribution. There is little or no performance improvement provided by these models with respect to the baseline methods with our Thai dataset. These findings suggest that further investigation is required to make a multilingual N-NER solution that works well across different languages. | ['Sarana Nutanong', 'Attapol Rutherford', 'Peerat Limkonchotiwat', 'Can Udomcharoenchaikit', 'Weerayut Buaphet'] | null | null | null | null | findings-acl-2022-5 | ['nested-named-entity-recognition'] | ['natural-language-processing'] | [-5.08325696e-01 -2.97592618e-02 -2.50808865e-01 -4.45791811e-01
-8.24250400e-01 -5.68633378e-01 4.22956973e-01 1.46849483e-01
-1.04956734e+00 9.12547708e-01 6.98732078e-01 -4.64823246e-01
1.87067628e-01 -5.62350214e-01 -5.27212322e-01 -2.34672993e-01
-6.57608034e-03 4.64839131e-01 1.98934004e-01 -3.37543964e-01
2.55224168e-01 5.66475987e-01 -1.05592036e+00 2.96350688e-01
1.08008337e+00 6.04020894e-01 -1.82853639e-02 3.79092485e-01
-5.18806219e-01 8.60154271e-01 -6.37789428e-01 -8.06369126e-01
9.47011169e-03 -5.28344885e-02 -8.09472620e-01 -4.41284865e-01
3.74209821e-01 -3.49218510e-02 -1.68260500e-01 8.04017663e-01
6.94900930e-01 2.93190718e-01 5.04507840e-01 -6.26255155e-01
-1.11789715e+00 1.25790036e+00 -3.44871849e-01 3.15544754e-01
1.27727017e-01 -2.91511565e-01 1.09487343e+00 -1.20730579e+00
8.36332619e-01 1.02142036e+00 1.14019775e+00 6.29299045e-01
-7.25324214e-01 -7.52391279e-01 1.23961531e-01 -1.34230152e-01
-1.50401509e+00 -5.18595636e-01 3.28864723e-01 6.09092116e-02
1.49292195e+00 -9.19269696e-02 2.44602337e-01 1.15847778e+00
2.24532694e-01 8.90367091e-01 1.17826438e+00 -5.70463955e-01
1.51543826e-01 3.02055091e-01 2.91362703e-01 2.93100744e-01
4.43714917e-01 -2.41416007e-01 -4.86812592e-01 4.94528823e-02
3.76187384e-01 -3.42564017e-01 -3.63215618e-02 1.94641352e-01
-1.30694890e+00 6.98556721e-01 3.78004819e-01 9.67768371e-01
-5.44722736e-01 -3.88408810e-01 5.73951364e-01 8.83947089e-02
4.71501142e-01 5.48148096e-01 -1.23141968e+00 -3.72777253e-01
-9.50164258e-01 -3.04845661e-01 1.26350379e+00 1.12888825e+00
4.89511222e-01 1.48433015e-01 1.98449880e-01 8.30418408e-01
8.99109840e-02 4.99221265e-01 4.89875019e-01 -3.66822898e-01
8.57281029e-01 5.20502687e-01 3.06900404e-02 -6.69745386e-01
-5.97148478e-01 -3.96775067e-01 -6.58092082e-01 -6.56988502e-01
4.47433621e-01 -7.40183830e-01 -8.79065394e-01 1.55539107e+00
-3.36125754e-02 -2.32699558e-01 6.08111203e-01 3.70572537e-01
1.24152482e+00 8.13805461e-01 4.57099497e-01 -5.73998354e-02
1.38872206e+00 -1.22356284e+00 -9.50053215e-01 -2.39554375e-01
9.20341790e-01 -9.31805074e-01 8.12230527e-01 9.53983441e-02
-7.21238792e-01 -5.27650774e-01 -7.15446353e-01 -3.35153550e-01
-1.05819011e+00 6.57866597e-01 7.43076146e-01 9.53844607e-01
-9.97535288e-01 2.40003020e-01 -8.79224002e-01 -8.83622527e-01
-2.17090413e-01 3.36557955e-01 -6.49261534e-01 2.88703348e-02
-1.44809914e+00 1.12725341e+00 8.32045436e-01 3.52442473e-01
-4.25331920e-01 -5.84665596e-01 -9.87420022e-01 6.87671006e-02
2.81300694e-01 -1.42583232e-02 1.14295220e+00 -4.18313444e-01
-1.31513786e+00 6.33810163e-01 -2.70215839e-01 -3.11041981e-01
8.77935216e-02 -5.90575516e-01 -9.73265648e-01 -3.74518931e-01
3.32990795e-01 5.72911561e-01 -2.84066737e-01 -1.13683379e+00
-7.55878150e-01 -3.24169636e-01 -2.67033756e-01 4.98332344e-02
-5.13269782e-01 5.35405397e-01 -3.83751571e-01 -8.33300889e-01
-1.93415537e-01 -9.74171102e-01 -3.74313563e-01 -9.15682137e-01
-5.91266036e-01 -4.77751791e-01 4.31875676e-01 -8.93636048e-01
1.54119956e+00 -2.13662219e+00 -4.54331160e-01 -1.55749008e-01
-1.77458256e-01 3.13695878e-01 -2.68378079e-01 9.03458118e-01
-7.03984052e-02 6.37218297e-01 4.23790030e-02 -2.69357026e-01
1.35832503e-01 1.37173861e-01 -1.59280986e-01 1.03613056e-01
6.39044568e-02 9.73533511e-01 -7.16479421e-01 -3.74108791e-01
-2.18748018e-01 7.62162030e-01 -1.31444171e-01 1.53633356e-01
2.76747614e-01 1.74870774e-01 -4.59869862e-01 6.63736641e-01
5.92686117e-01 1.37993976e-01 2.74552971e-01 -2.59301215e-01
-7.83211589e-01 6.47334397e-01 -1.05525613e+00 1.54726136e+00
-3.88645500e-01 5.23885727e-01 -9.87638310e-02 -3.84275824e-01
1.04996407e+00 5.72047412e-01 1.60116166e-01 -6.37848854e-01
1.62307099e-01 7.61779368e-01 6.03669286e-02 -3.09269488e-01
9.81996119e-01 -2.33275853e-02 -5.17076552e-01 1.05222277e-01
3.62485439e-01 3.90636981e-01 3.72321308e-01 -2.32884049e-01
9.82778549e-01 -9.32583734e-02 5.33275425e-01 -3.16789687e-01
3.69025946e-01 1.26847059e-01 7.99787104e-01 9.00970340e-01
-2.98092157e-01 3.03525984e-01 1.44482538e-01 -4.49140012e-01
-1.14252698e+00 -8.09597790e-01 -2.75374979e-01 1.40023744e+00
-3.66983026e-01 -3.45053285e-01 -7.77329445e-01 -9.42446709e-01
-5.11638820e-01 9.16541934e-01 -4.95785445e-01 6.21293843e-01
-9.42565084e-01 -8.24828267e-01 1.19251215e+00 7.70630419e-01
7.99297333e-01 -1.49520767e+00 -1.03864394e-01 4.00655955e-01
-2.61048317e-01 -1.53246725e+00 -5.85374892e-01 6.82839870e-01
-9.56427276e-01 -6.48202837e-01 -9.39173877e-01 -1.31493247e+00
2.98838854e-01 -2.69702733e-01 1.16251862e+00 -4.49607521e-01
4.14354742e-01 1.13756247e-01 -6.63241446e-01 -4.48599666e-01
-2.79882580e-01 8.23331892e-01 6.28466234e-02 -4.47229177e-01
9.29801524e-01 -1.58386007e-01 -1.19695321e-01 1.44468233e-01
-6.26935065e-01 -4.66552764e-01 1.11529493e+00 6.59979641e-01
4.77381229e-01 2.49862909e-01 7.25107610e-01 -1.04570627e+00
6.02858663e-01 -4.92414266e-01 -2.28510931e-01 5.00701785e-01
-5.02958834e-01 6.77822456e-02 8.51780593e-01 -3.76257598e-01
-1.45133889e+00 -4.48959805e-02 -5.62683880e-01 3.90186548e-01
-6.25933111e-01 8.69296253e-01 -2.73318917e-01 2.86725610e-01
5.26428998e-01 2.08101675e-01 -9.43585813e-01 -1.00999165e+00
4.70933914e-01 9.65916216e-01 5.62413096e-01 -5.04359603e-01
2.58769602e-01 9.63062122e-02 -6.62895799e-01 -1.09862256e+00
-9.53280866e-01 -7.52501607e-01 -1.03628743e+00 3.40343207e-01
1.12980652e+00 -9.93408084e-01 -3.26409876e-01 6.86583161e-01
-1.25324047e+00 -2.43816689e-01 -1.05595231e-01 7.45673835e-01
4.74924147e-02 2.93151587e-01 -1.28976357e+00 -7.11615145e-01
-5.00361621e-01 -7.95254588e-01 8.42601776e-01 4.00113910e-01
1.35149853e-02 -1.12580180e+00 2.96194047e-01 2.49554023e-01
4.98690546e-01 -6.33293986e-02 6.94485009e-01 -1.38481164e+00
1.58935133e-02 -1.14302218e-01 -1.25958607e-01 2.39639074e-01
5.03649041e-02 -1.32494017e-01 -8.11126709e-01 -9.91216823e-02
-1.07030645e-01 -2.54133016e-01 7.21698105e-01 7.48176426e-02
3.97949964e-01 -3.09889793e-01 -2.63658345e-01 1.64350107e-01
1.45931602e+00 4.82843816e-01 7.29697883e-01 7.86981761e-01
7.83635497e-01 4.24173892e-01 4.57238257e-01 2.19838381e-01
8.41143429e-01 2.58509457e-01 -1.84099570e-01 -2.37397656e-01
-1.28237456e-01 -4.74739015e-01 4.86287594e-01 1.55841064e+00
-5.61112761e-02 -5.44348419e-01 -1.27851212e+00 9.84259605e-01
-1.50524342e+00 -7.68949807e-01 -1.30452082e-01 1.86621547e+00
9.97181833e-01 1.27123088e-01 -1.21932991e-01 -3.25791180e-01
8.60767186e-01 1.06983580e-01 -2.68671751e-01 -6.88655794e-01
-3.61701488e-01 3.52782995e-01 6.23628080e-01 6.40902966e-02
-1.44818771e+00 1.29002452e+00 6.61883640e+00 7.70307183e-01
-9.47584569e-01 1.05762415e-01 3.96159172e-01 6.37151122e-01
-8.57062116e-02 6.84384853e-02 -1.47624898e+00 1.62670031e-01
1.63566482e+00 1.32242084e-01 1.25470543e-02 8.35331619e-01
-5.85483424e-02 2.57066190e-01 -8.77447784e-01 5.17393291e-01
1.45803764e-01 -9.49373662e-01 4.40041795e-02 3.68472300e-02
9.86200631e-01 6.40986145e-01 -3.84066254e-01 7.33339846e-01
4.78118211e-01 -9.37097132e-01 5.78223348e-01 4.77100194e-01
6.13796115e-01 -1.00553668e+00 1.41678214e+00 4.10089433e-01
-1.38245964e+00 5.89053519e-02 -3.31827134e-01 2.35836491e-01
4.24229354e-01 4.37484324e-01 -7.18389809e-01 6.57361448e-01
8.48932266e-01 6.67826653e-01 -5.16023815e-01 1.00491679e+00
-2.85532653e-01 8.69389594e-01 -4.15042788e-01 -2.99985737e-01
4.60226953e-01 2.18433738e-01 2.65065074e-01 1.94595897e+00
3.31597954e-01 6.28120750e-02 1.68122560e-01 3.52657557e-01
-4.18240964e-01 6.42809927e-01 -4.48538929e-01 -3.31349820e-01
6.25987470e-01 1.12734199e+00 -9.25880730e-01 -1.94255665e-01
-8.17233920e-01 8.40343535e-01 5.88164151e-01 3.82214636e-01
-6.62455857e-01 -9.81482923e-01 1.51614994e-01 -2.58342534e-01
7.03594744e-01 -5.04975438e-01 -3.82923692e-01 -1.32610524e+00
-1.46174684e-01 -8.39191437e-01 4.81914103e-01 -4.10524577e-01
-1.77022469e+00 1.12205088e+00 -2.78870523e-01 -7.47369111e-01
-3.24281938e-02 -8.03888619e-01 -2.36005709e-01 6.02365613e-01
-1.77715576e+00 -1.30418861e+00 2.57924944e-01 3.52449417e-01
4.83658850e-01 -2.38238677e-01 1.07702219e+00 6.63454175e-01
-6.07901871e-01 9.55651224e-01 3.27972829e-01 9.53465939e-01
8.16177070e-01 -1.18551433e+00 6.81528747e-01 9.73898470e-01
2.68647641e-01 9.05569613e-01 2.50916451e-01 -8.03407013e-01
-1.10296249e+00 -8.28843951e-01 1.92660129e+00 -5.96441627e-01
8.64145815e-01 -3.51708800e-01 -8.29954743e-01 9.64768112e-01
5.81116736e-01 -1.56371221e-01 9.19853330e-01 4.37343597e-01
-3.68123680e-01 2.22184286e-01 -1.16954815e+00 5.81709564e-01
7.61453629e-01 -5.48178315e-01 -9.78184521e-01 -1.43562585e-01
8.69808733e-01 -3.58507901e-01 -1.45661104e+00 5.14761209e-01
4.67889249e-01 -7.20024943e-01 4.50352728e-01 -6.00713730e-01
9.29667428e-02 -1.52496800e-01 -2.96672970e-01 -1.04831779e+00
-2.04877138e-01 -2.66042054e-01 2.53298342e-01 1.84949863e+00
9.98622358e-01 -5.91037869e-01 5.09661317e-01 4.81724560e-01
-3.03688675e-01 -5.90200126e-01 -9.23235238e-01 -7.27425218e-01
4.10562038e-01 -5.23274839e-01 5.75808346e-01 1.21646011e+00
-2.01022863e-01 5.51561236e-01 -3.25833738e-01 1.59461841e-01
2.22330213e-01 -3.32291186e-01 3.18406790e-01 -1.00216973e+00
2.57097960e-01 -1.01762667e-01 -1.59887582e-01 -1.02927148e+00
3.82851690e-01 -7.33388722e-01 2.12718219e-01 -1.66839576e+00
3.12033385e-01 -6.07298255e-01 -3.74482810e-01 1.02686119e+00
1.99473500e-02 1.57885954e-01 1.75034136e-01 1.27025738e-01
-7.46886253e-01 5.14691651e-01 8.23920548e-01 1.72132730e-01
-3.30437690e-01 -2.52622157e-01 -7.62617469e-01 6.98698640e-01
6.92482352e-01 -7.08511889e-01 7.66767114e-02 -6.73553467e-01
2.51593500e-01 -2.04500079e-01 -4.40867007e-01 -8.87365222e-01
4.75524366e-01 7.43511394e-02 5.14184773e-01 -7.98591554e-01
-1.16182692e-01 -8.16428006e-01 -5.87604120e-02 8.20636228e-02
-3.38654190e-01 4.96075243e-01 4.10238951e-01 1.52752951e-01
-3.88596565e-01 -2.57725298e-01 3.93653959e-01 -2.22038060e-01
-1.09950280e+00 -5.84392138e-02 -6.95440173e-01 5.05267799e-01
6.16916478e-01 -6.96024150e-02 -3.85741681e-01 4.54138890e-02
-6.74653351e-01 3.88006233e-02 1.53577730e-01 7.04874933e-01
2.72984713e-01 -1.20384789e+00 -8.36076617e-01 -4.00340967e-02
1.29236400e-01 -4.13496405e-01 9.74214301e-02 5.34327269e-01
-3.76709789e-01 8.31441045e-01 -3.86196673e-02 -2.31446978e-02
-7.31460154e-01 3.71654361e-01 1.74156398e-01 -7.19266057e-01
-4.13820207e-01 5.82286000e-01 -2.78920591e-01 -1.27319694e+00
2.76600033e-01 -3.10320288e-01 -6.85665548e-01 2.10995942e-01
4.43329573e-01 5.83569586e-01 3.06995124e-01 -1.09127724e+00
-6.90869331e-01 4.41470772e-01 -2.86232442e-01 -1.08104184e-01
1.57338333e+00 -1.80430651e-01 -2.34217033e-01 6.88326538e-01
1.08260345e+00 6.56246305e-01 -4.50227737e-01 -4.78354134e-02
6.00911140e-01 2.69104302e-01 -1.45297050e-01 -1.14881539e+00
-9.44561481e-01 5.43199420e-01 4.57937390e-01 -2.89369212e-03
9.96035874e-01 7.41960621e-03 1.15852940e+00 7.59315908e-01
4.84551162e-01 -1.25684392e+00 -4.97586995e-01 1.28598273e+00
3.82056922e-01 -1.06594288e+00 -3.32379490e-01 -9.31243822e-02
-8.85537267e-01 1.14834034e+00 7.24866211e-01 1.02746986e-01
7.63184249e-01 5.21830022e-01 4.84885722e-01 -1.17097758e-01
-5.40653288e-01 -2.82569408e-01 2.91948527e-01 3.63112271e-01
1.16427302e+00 2.43699644e-02 -7.20303714e-01 9.84076619e-01
-3.67145926e-01 4.34522890e-02 4.00521576e-01 8.55905652e-01
-2.89386332e-01 -1.18201518e+00 -9.76501778e-02 1.69136778e-01
-1.10751331e+00 -5.89392126e-01 -3.44906479e-01 1.23025072e+00
1.17017545e-01 1.18521988e+00 -2.86878496e-01 -2.61986911e-01
7.45611072e-01 2.67055929e-01 -1.73730746e-01 -6.64328158e-01
-1.04774761e+00 -5.62823005e-02 5.18746197e-01 -1.78165659e-01
-5.45400500e-01 -5.35924375e-01 -1.54835987e+00 -3.11117470e-01
-5.72502434e-01 6.71552777e-01 7.46232986e-01 8.91292155e-01
3.67206722e-01 1.62333578e-01 2.32110605e-01 -5.80808461e-01
-3.03845495e-01 -1.13960457e+00 -6.70571983e-01 1.60236005e-02
-2.97221933e-02 -1.87690169e-01 -3.59869331e-01 -2.87476093e-01] | [9.89746379852295, 9.726642608642578] |
43b0eca4-a102-4b93-89ec-0d8662aeac2d | language-agnostic-bert-sentence-embedding | 2007.01852 | null | https://arxiv.org/abs/2007.01852v2 | https://arxiv.org/pdf/2007.01852v2.pdf | Language-agnostic BERT Sentence Embedding | While BERT is an effective method for learning monolingual sentence embeddings for semantic similarity and embedding based transfer learning (Reimers and Gurevych, 2019), BERT based cross-lingual sentence embeddings have yet to be explored. We systematically investigate methods for learning multilingual sentence embeddings by combining the best methods for learning monolingual and cross-lingual representations including: masked language modeling (MLM), translation language modeling (TLM) (Conneau and Lample, 2019), dual encoder translation ranking (Guo et al., 2018), and additive margin softmax (Yang et al., 2019a). We show that introducing a pre-trained multilingual language model dramatically reduces the amount of parallel training data required to achieve good performance by 80%. Composing the best of these methods produces a model that achieves 83.7% bi-text retrieval accuracy over 112 languages on Tatoeba, well above the 65.5% achieved by Artetxe and Schwenk (2019b), while still performing competitively on monolingual transfer learning benchmarks (Conneau and Kiela, 2018). Parallel data mined from CommonCrawl using our best model is shown to train competitive NMT models for en-zh and en-de. We publicly release our best multilingual sentence embedding model for 109+ languages at https://tfhub.dev/google/LaBSE. | ['Yinfei Yang', 'Daniel Cer', 'Naveen Arivazhagan', 'Fangxiaoyu Feng', 'Wei Wang'] | 2020-07-03 | null | https://aclanthology.org/2022.acl-long.62 | https://aclanthology.org/2022.acl-long.62.pdf | acl-2022-5 | ['multilingual-nlp'] | ['natural-language-processing'] | [-3.06127906e-01 -7.78246075e-02 -4.44350451e-01 -2.15984672e-01
-1.49697983e+00 -7.86576331e-01 1.01169419e+00 3.53683859e-01
-8.43872607e-01 7.49803722e-01 4.83291566e-01 -8.37619960e-01
2.28963777e-01 -4.64822650e-01 -9.86753702e-01 -1.43160731e-01
6.89267814e-02 5.69015265e-01 -2.30716914e-01 -4.66654480e-01
-1.82004403e-02 -1.63239648e-03 -1.07871473e+00 3.87792468e-01
9.96447861e-01 4.47073698e-01 4.22956049e-01 8.81158888e-01
-1.61586419e-01 6.19822502e-01 -3.37632418e-01 -6.53135419e-01
5.29818982e-02 -1.71438843e-01 -9.84326482e-01 -5.74525833e-01
8.17763388e-01 -1.13820925e-01 -3.95214051e-01 8.68529320e-01
8.54960084e-01 -2.16929838e-01 8.33649099e-01 -7.98569262e-01
-1.35131097e+00 9.80772376e-01 -4.11655217e-01 2.80373454e-01
1.99465737e-01 3.10639553e-02 1.31056380e+00 -1.37614107e+00
8.00300777e-01 1.25936258e+00 7.06905305e-01 4.55772191e-01
-1.25172448e+00 -7.19885528e-01 -1.95908055e-01 4.05660540e-01
-1.30152202e+00 -5.96483111e-01 2.68030822e-01 -2.77426779e-01
1.49293137e+00 1.59306586e-01 2.56427974e-01 1.36273491e+00
3.67697358e-01 9.93992329e-01 1.23954058e+00 -8.02155793e-01
-2.58014560e-01 4.25927252e-01 6.74958974e-02 6.58200085e-01
8.87927338e-02 -1.46940783e-01 -5.18368244e-01 -1.95925236e-01
9.62120220e-02 -2.65857071e-01 -1.46359382e-02 -1.29317865e-01
-1.27563632e+00 9.44564283e-01 4.22529161e-01 5.11129498e-01
-6.20542429e-02 1.32301599e-01 9.02361333e-01 8.12073529e-01
9.33627844e-01 5.85761786e-01 -8.43148828e-01 -1.13614097e-01
-8.52300942e-01 1.26539394e-01 6.20095491e-01 8.97720098e-01
7.10208833e-01 -2.99368761e-02 -1.07504003e-01 1.16272366e+00
1.63204461e-01 7.59810448e-01 7.46566832e-01 -5.41719854e-01
6.97475195e-01 1.09764963e-01 -1.81884900e-01 -4.47992086e-01
-1.03304528e-01 -4.20704693e-01 -4.33631063e-01 -3.50357741e-01
2.88582206e-01 -2.24667042e-01 -5.87109804e-01 1.85042453e+00
-1.15629837e-01 -1.13014705e-01 5.60904741e-01 5.16760349e-01
7.70037532e-01 8.84598196e-01 1.16275437e-01 1.95838660e-01
1.39356327e+00 -1.16128623e+00 -4.50642943e-01 -4.56065655e-01
1.30488896e+00 -1.29131067e+00 1.38677084e+00 3.11485510e-02
-1.03416276e+00 -5.02969623e-01 -1.03133118e+00 -5.41371167e-01
-6.07046366e-01 2.93087900e-01 6.38177156e-01 3.64918202e-01
-1.27260387e+00 2.74874866e-01 -8.73217762e-01 -8.42125475e-01
1.34956598e-01 1.57428533e-01 -5.30491948e-01 -4.58477914e-01
-1.60916424e+00 1.45976233e+00 1.87653482e-01 -1.77043051e-01
-7.63451099e-01 -8.18576157e-01 -1.02721882e+00 -1.13328099e-01
-2.08985418e-01 -4.66275513e-01 1.06945205e+00 -7.72276282e-01
-1.30508792e+00 1.13789153e+00 -1.22242786e-01 -7.48523772e-01
2.13093981e-01 -4.14317638e-01 -5.44880390e-01 -9.49935541e-02
2.69455492e-01 9.11348104e-01 2.46164247e-01 -7.75151789e-01
-2.57525384e-01 -2.68656015e-01 3.27033848e-02 2.89927542e-01
-8.47906351e-01 4.61351424e-01 -1.31382704e-01 -5.19240499e-01
-3.96241069e-01 -9.24495101e-01 8.40006843e-02 -3.34707916e-01
-9.95481014e-02 -5.42238832e-01 5.27934790e-01 -1.06080496e+00
9.84167755e-01 -1.82686162e+00 1.49065275e-02 -4.38730747e-01
-1.30420670e-01 4.78683054e-01 -8.00441682e-01 9.66703832e-01
-1.36964634e-01 2.53024280e-01 -1.61877885e-01 -7.76790798e-01
2.76933879e-01 8.05276036e-02 -3.85360867e-01 4.72941250e-01
2.21142024e-01 1.39784896e+00 -8.62080812e-01 -4.07797396e-01
2.04332381e-01 7.44495451e-01 -5.44561803e-01 4.85181995e-02
1.73933469e-02 1.56702623e-01 2.13672891e-01 5.75465500e-01
3.56863976e-01 -2.68899500e-02 3.13118607e-01 -1.15568027e-01
-1.69036269e-01 9.18767452e-01 -5.92327714e-01 2.04439545e+00
-1.15294278e+00 9.80885148e-01 -7.35033602e-02 -8.21523547e-01
8.06919456e-01 5.50650179e-01 2.60472745e-01 -7.88830280e-01
2.44805943e-02 7.26657212e-01 5.60659589e-03 -4.09322202e-01
6.79226279e-01 -1.19013980e-01 -3.90690058e-01 6.78059697e-01
4.63763297e-01 -1.77785113e-01 2.72142172e-01 2.99225181e-01
9.43892896e-01 1.23160347e-01 1.37513891e-01 -6.12853289e-01
3.83960992e-01 -1.00644026e-02 1.51807070e-01 5.37910938e-01
-1.57558501e-01 2.95172453e-01 1.49272025e-01 -2.70803750e-01
-1.35465658e+00 -1.26916528e+00 -4.87361133e-01 1.40123069e+00
-3.69237572e-01 -5.75100005e-01 -6.76255643e-01 -5.16620874e-01
1.00166656e-01 8.48876715e-01 -4.32313889e-01 -1.63488045e-01
-5.91695487e-01 -7.37917125e-01 8.95392299e-01 3.16962868e-01
5.59114702e-02 -9.93713796e-01 1.36367932e-01 1.69803083e-01
-2.70670742e-01 -1.05974233e+00 -7.44039893e-01 3.09196651e-01
-6.07038617e-01 -6.39391780e-01 -7.99207568e-01 -1.06674337e+00
3.08527380e-01 -1.08285099e-01 1.32731414e+00 -3.34600180e-01
-4.07387316e-01 2.52151221e-01 -2.87488550e-01 -3.04608971e-01
-6.54971778e-01 5.11810005e-01 4.16037917e-01 -4.93940264e-01
7.13249922e-01 -3.91214639e-01 -3.12737703e-01 -2.73748785e-01
-7.44461715e-01 -1.86328366e-01 5.91092825e-01 1.12294197e+00
4.54463601e-01 -8.55525017e-01 7.64078557e-01 -6.88695490e-01
9.28440690e-01 -5.75731218e-01 -4.06876653e-01 4.93809313e-01
-7.31900334e-01 1.94144621e-01 7.36313939e-01 -4.73238498e-01
-5.10043859e-01 -5.12107730e-01 -3.12330127e-01 -2.53520131e-01
6.77271411e-02 5.95637321e-01 1.94922507e-01 1.38809428e-01
6.65011287e-01 2.52882063e-01 -7.94120431e-02 -7.00803697e-01
8.03671718e-01 1.19974780e+00 2.26691306e-01 -6.85072899e-01
5.90948939e-01 -4.96085249e-02 -6.08299494e-01 -5.93434691e-01
-8.16858172e-01 -3.13610345e-01 -8.00198257e-01 2.69299090e-01
9.70562756e-01 -1.23293316e+00 -1.89396933e-01 -4.52710912e-02
-1.33786678e+00 -3.20787758e-01 -1.49904087e-01 7.89346635e-01
-4.24332500e-01 2.80883431e-01 -1.13917291e+00 -4.15180981e-01
-6.68937325e-01 -1.14962137e+00 1.07702565e+00 -4.09805298e-01
-4.11740661e-01 -1.35501468e+00 4.96869743e-01 4.36980844e-01
6.28343999e-01 -1.45505354e-01 1.06078935e+00 -8.54759455e-01
-1.45442054e-01 -1.05043657e-01 -1.19712114e-01 6.67104542e-01
7.31319562e-02 -2.52799928e-01 -9.44972098e-01 -7.72597551e-01
-2.45347843e-01 -8.94266546e-01 8.39139819e-01 7.63661638e-02
6.19453132e-01 -2.53568977e-01 -8.69734399e-03 5.52350938e-01
1.46243203e+00 -3.05941373e-01 2.21251443e-01 5.83634079e-01
5.79086363e-01 4.76199210e-01 2.77595520e-01 -2.45442301e-01
7.52026916e-01 5.65161765e-01 -5.49862534e-02 -1.58604398e-01
-3.65083963e-01 -3.99885476e-01 9.68066692e-01 1.81322181e+00
3.67713511e-01 -1.70760095e-01 -9.71067250e-01 8.08669329e-01
-1.63519788e+00 -7.80668020e-01 1.42502738e-02 2.13614225e+00
1.13140571e+00 -3.78939360e-02 -2.15747252e-01 -4.28156316e-01
4.11204338e-01 1.50659397e-01 -2.39935622e-01 -1.06233406e+00
-4.79053825e-01 6.22311592e-01 6.95155025e-01 1.00287008e+00
-1.01100969e+00 1.42153287e+00 5.49340343e+00 9.22119439e-01
-1.04760873e+00 7.06873298e-01 4.31131512e-01 -1.69140026e-01
-5.16050100e-01 3.22431587e-02 -8.57862115e-01 3.83371383e-01
1.51875699e+00 -4.25034493e-01 5.86636186e-01 4.90597606e-01
-2.04245403e-01 3.41464370e-01 -1.17761350e+00 7.70818710e-01
4.53748405e-01 -1.13499880e+00 -7.46847540e-02 -9.97200161e-02
1.07330096e+00 1.02482653e+00 2.92930603e-01 8.05328786e-01
5.01710474e-01 -1.21234012e+00 4.58158463e-01 1.39548033e-01
9.63293314e-01 -7.67926335e-01 7.40461409e-01 3.18215758e-01
-8.89163315e-01 1.86858088e-01 -5.77847600e-01 7.09322318e-02
9.59501937e-02 3.28240514e-01 -8.11200500e-01 6.53651118e-01
5.47071874e-01 8.76775265e-01 -7.12237418e-01 4.03465271e-01
-3.40414971e-01 8.08862209e-01 -2.65546232e-01 -1.76605657e-01
4.34177667e-01 -7.17991740e-02 3.94881517e-01 1.49481058e+00
2.36675233e-01 -8.28565836e-01 9.46221799e-02 5.68913579e-01
-3.82284433e-01 6.55065775e-01 -7.40819454e-01 -2.54210860e-01
4.98812437e-01 1.05142236e+00 -1.64544694e-02 -3.98261756e-01
-6.65790617e-01 1.14785397e+00 7.67867625e-01 3.88926268e-01
-5.25366008e-01 -4.63873148e-01 8.03454459e-01 -1.57764688e-01
6.38049170e-02 -3.80354375e-01 -1.30003244e-01 -1.39171147e+00
1.30870238e-01 -9.72943246e-01 2.25848258e-01 -5.61675131e-01
-1.67250729e+00 8.47351491e-01 -1.93470195e-01 -1.06634879e+00
-4.15432185e-01 -7.56157637e-01 -3.93900007e-01 1.35233343e+00
-1.64391899e+00 -1.58251905e+00 5.76763093e-01 2.00683877e-01
6.79684758e-01 -3.70538324e-01 1.26733446e+00 6.46166265e-01
-3.74581039e-01 9.23747957e-01 7.44123161e-01 2.16614217e-01
1.17894614e+00 -1.32789755e+00 7.45779395e-01 6.34845436e-01
5.00573874e-01 8.20616603e-01 4.06367779e-01 -3.61167312e-01
-1.51220417e+00 -1.10687864e+00 1.77406704e+00 -8.49388838e-01
1.20672297e+00 -8.88486266e-01 -8.83867323e-01 1.02003813e+00
9.62784648e-01 -2.04627559e-01 7.38303423e-01 4.12985504e-01
-6.42005503e-01 -9.37127471e-02 -6.76172972e-01 8.03084493e-01
6.02512538e-01 -1.19761109e+00 -6.78310573e-01 8.26250315e-01
1.00256681e+00 -1.82069875e-02 -1.18669152e+00 3.87485623e-01
4.03695256e-01 -4.84223604e-01 9.44461644e-01 -8.85470450e-01
5.59295595e-01 1.07373185e-01 -3.92940134e-01 -1.53653502e+00
8.90581869e-03 -2.53460020e-01 1.43224448e-01 1.31028366e+00
7.36577153e-01 -8.55605364e-01 5.49447909e-02 -5.89556172e-02
-2.38137856e-01 -1.07378733e+00 -1.00276113e+00 -1.13701451e+00
1.15140176e+00 -4.26367998e-01 4.02867556e-01 1.23270822e+00
2.12393075e-01 6.90253079e-01 -2.30782598e-01 -9.97021720e-02
4.77749020e-01 -1.64337426e-01 6.39690757e-01 -7.15596855e-01
-2.38661200e-01 -4.50708419e-01 -1.88712552e-01 -9.65516806e-01
7.88531899e-01 -1.94146335e+00 -2.31594265e-01 -1.80475819e+00
3.65163624e-01 -3.65491778e-01 -5.57510257e-01 4.05540913e-01
-1.97445303e-01 3.58938128e-01 8.36388171e-02 1.57210931e-01
-4.70986873e-01 6.97849095e-01 7.76407540e-01 -2.38736242e-01
3.28489065e-01 -6.52244806e-01 -5.48960447e-01 3.20280463e-01
8.27591598e-01 -5.65081358e-01 -1.52168557e-01 -1.08695889e+00
7.83144236e-02 -3.33604097e-01 8.39034468e-02 -6.34232521e-01
1.75846964e-01 3.54879588e-01 1.04337901e-01 -2.77803957e-01
3.51788878e-01 -2.98281997e-01 -3.94382149e-01 5.42152405e-01
-6.36431694e-01 7.28608668e-01 4.37938124e-01 1.12710506e-01
-3.51601481e-01 -2.40594417e-01 5.66170692e-01 9.08303037e-02
-3.92090052e-01 1.42473921e-01 -3.42186660e-01 3.09325337e-01
4.89289522e-01 3.80596489e-01 -6.01084173e-01 -2.54490644e-01
-4.42758888e-01 4.04361218e-01 4.30672884e-01 7.91870296e-01
1.65226907e-01 -1.45743012e+00 -1.17909873e+00 1.26672983e-01
2.90905744e-01 -6.22740746e-01 -2.53679417e-02 9.01117742e-01
-4.26627189e-01 7.41475999e-01 3.12485695e-02 -3.72798860e-01
-1.03935850e+00 3.56349945e-01 1.30374297e-01 -5.38506985e-01
-2.53726333e-01 9.08063293e-01 -2.27819443e-01 -1.24285829e+00
-7.75841102e-02 -1.85426809e-02 1.46391749e-01 -5.12731820e-02
2.73218066e-01 1.73703805e-01 3.08318943e-01 -6.48815155e-01
-4.97207433e-01 4.76814806e-01 -3.80226314e-01 -3.70258749e-01
1.35238874e+00 -7.45907277e-02 -3.68346900e-01 8.15872490e-01
1.88007092e+00 2.00091198e-01 -4.10286814e-01 -4.97833997e-01
2.25351050e-01 -2.04767217e-03 -4.59163673e-02 -8.79015446e-01
-3.50608557e-01 1.31102812e+00 6.08607471e-01 -1.58314273e-01
6.55837715e-01 1.74996942e-01 1.18135369e+00 5.95138550e-01
3.30771506e-01 -1.11788249e+00 -1.96039379e-01 9.23250377e-01
8.38099480e-01 -1.40843153e+00 -1.97861478e-01 1.83871806e-01
-4.80175227e-01 9.33453500e-01 4.49900895e-01 -1.61373898e-01
5.96730113e-01 3.40545386e-01 3.56868029e-01 7.26481527e-02
-1.06882870e+00 -7.76086152e-02 4.15274203e-01 2.33894035e-01
9.88118410e-01 3.52037281e-01 -6.46812737e-01 3.16587359e-01
-4.77178127e-01 -3.86795104e-01 1.85756296e-01 6.82123661e-01
-2.09705964e-01 -1.49099004e+00 -9.60514471e-02 4.63691294e-01
-5.53992987e-01 -8.83029044e-01 -2.93195754e-01 7.38422215e-01
-1.00780748e-01 7.82794178e-01 2.00532541e-01 -3.32822710e-01
1.44601032e-01 4.45874304e-01 5.20492315e-01 -8.65375698e-01
-8.23810518e-01 -1.08861685e-01 3.29396576e-01 -6.84144199e-02
-3.84837389e-02 -6.80025995e-01 -8.45478952e-01 -3.22361082e-01
-2.42931187e-01 3.57953936e-01 8.35221529e-01 8.81330729e-01
3.97460401e-01 2.22484112e-01 4.97441649e-01 -4.77024198e-01
-7.46527553e-01 -1.49967301e+00 -1.77011564e-01 2.27068290e-01
1.63786411e-01 -1.91944003e-01 -5.14502823e-01 -1.53438032e-01] | [11.200488090515137, 9.751641273498535] |
9885f038-13c9-46ac-81a3-6818ddd34c0a | assessing-the-efficacy-of-large-language | 2307.04274 | null | https://arxiv.org/abs/2307.04274v1 | https://arxiv.org/pdf/2307.04274v1.pdf | Assessing the efficacy of large language models in generating accurate teacher responses | (Tack et al., 2023) organized the shared task hosted by the 18th Workshop on Innovative Use of NLP for Building Educational Applications on generation of teacher language in educational dialogues. Following the structure of the shared task, in this study, we attempt to assess the generative abilities of large language models in providing informative and helpful insights to students, thereby simulating the role of a knowledgeable teacher. To this end, we present an extensive evaluation of several benchmarking generative models, including GPT-4 (few-shot, in-context learning), fine-tuned GPT-2, and fine-tuned DialoGPT. Additionally, to optimize for pedagogical quality, we fine-tuned the Flan-T5 model using reinforcement learning. Our experimental findings on the Teacher-Student Chatroom Corpus subset indicate the efficacy of GPT-4 over other fine-tuned models, measured using BERTScore and DialogRPT. We hypothesize that several dataset characteristics, including sampling, representativeness, and dialog completeness, pose significant challenges to fine-tuning, thus contributing to the poor generalizability of the fine-tuned models. Finally, we note the need for these generative models to be evaluated with a metric that relies not only on dialog coherence and matched language modeling distribution but also on the model's ability to showcase pedagogical skills. | ['Tushaar Gangavarapu', 'Wentao Guo', 'Abhishek Masand', 'Yann Hicke'] | 2023-07-09 | null | null | null | null | ['benchmarking', 'benchmarking'] | ['miscellaneous', 'robots'] | [-1.91565007e-01 7.02258945e-01 1.15728177e-01 -2.99910814e-01
-9.57314193e-01 -8.98198247e-01 9.79288101e-01 -2.82112323e-02
-2.29690135e-01 1.06128025e+00 6.38677299e-01 -5.02750456e-01
-3.01833481e-01 -5.97762048e-01 -3.05582374e-01 -2.60204673e-01
1.19908527e-01 8.99426341e-01 3.30001563e-01 -6.08217478e-01
1.21689439e-01 3.56953084e-01 -1.43430007e+00 4.25104588e-01
1.27549434e+00 1.49417654e-01 3.55149537e-01 9.26767707e-01
-2.52469301e-01 1.20321083e+00 -1.04080188e+00 -5.16076326e-01
-2.40190268e-01 -7.30443537e-01 -1.18607891e+00 3.27210985e-02
4.10783321e-01 -5.23710251e-01 -2.63402071e-02 5.55809617e-01
7.29523063e-01 6.29084051e-01 6.81575060e-01 -1.18328023e+00
-4.53753024e-01 1.27707696e+00 1.77216053e-01 1.29271805e-01
6.05479419e-01 5.70137322e-01 9.12340581e-01 -4.52221781e-01
6.68491900e-01 1.69494784e+00 5.20916879e-01 9.12555814e-01
-1.35011411e+00 -5.75789988e-01 6.39590621e-02 -1.74815178e-01
-8.06290269e-01 -6.06424570e-01 2.42679358e-01 -5.84497869e-01
8.94724727e-01 2.26893082e-01 7.75914371e-01 1.21933401e+00
-7.70662501e-02 7.18433738e-01 1.44638014e+00 -5.41689098e-01
1.67064920e-01 6.49756551e-01 8.63280594e-02 5.55619538e-01
-3.20564359e-01 1.61903836e-02 -5.60963988e-01 -2.99846709e-01
5.58718383e-01 -7.25018740e-01 -2.18435183e-01 -1.12343267e-01
-7.89117873e-01 1.21171641e+00 -1.59327015e-01 3.62988502e-01
6.39193575e-04 -1.33554086e-01 2.06153929e-01 5.02293885e-01
4.50378120e-01 9.53425407e-01 -4.99313354e-01 -8.63601625e-01
-7.08929539e-01 5.59400439e-01 1.43061006e+00 1.01805675e+00
4.06285346e-01 6.51811436e-02 -6.71676457e-01 9.38111842e-01
3.60455900e-01 1.06548831e-01 8.22796464e-01 -1.29494488e+00
3.82636845e-01 5.65036714e-01 -4.50578779e-02 -6.96534887e-02
-3.13469827e-01 2.01334129e-03 -1.19222865e-01 1.24525703e-01
6.42604351e-01 -6.61279261e-01 -5.89078426e-01 1.79603660e+00
3.23106170e-01 5.78828193e-02 4.53484774e-01 3.16617399e-01
1.43063879e+00 5.25786757e-01 3.73176843e-01 -2.72952437e-01
1.18857527e+00 -1.07497752e+00 -6.78740978e-01 -3.51753831e-03
7.67575502e-01 -1.14257812e+00 1.32956147e+00 2.85717398e-01
-1.58097124e+00 -5.60423136e-01 -3.56531531e-01 4.76781689e-02
-1.06158502e-01 -1.81237683e-01 3.83634448e-01 8.14612210e-01
-1.21939003e+00 4.56438214e-01 -5.71938217e-01 -4.26877856e-01
-1.77856043e-01 1.52090371e-01 2.07274370e-02 3.43510181e-01
-1.12333214e+00 1.07300353e+00 2.86194831e-01 -6.07527018e-01
-1.08940661e+00 -1.04005921e+00 -6.61278605e-01 3.82805854e-01
2.73429900e-01 -6.23627841e-01 1.89532709e+00 -3.32444549e-01
-2.29219770e+00 5.75936675e-01 3.27992320e-01 -4.90539312e-01
7.43027151e-01 -1.24310359e-01 1.90914303e-01 -1.27668502e-02
-2.52127033e-02 9.46935117e-01 3.50106321e-02 -1.14979482e+00
-4.94069487e-01 1.10735692e-01 4.38426524e-01 5.43019831e-01
-2.13852838e-01 -5.86536387e-03 -1.32211789e-01 -3.09638083e-01
-5.27924716e-01 -1.02888727e+00 -3.15812618e-01 -9.57483709e-01
-1.10406645e-01 -8.51279795e-01 5.91237247e-01 -3.83649468e-01
1.24548316e+00 -1.73263037e+00 -1.11153042e-02 -5.41801155e-02
1.65684268e-01 3.19504946e-01 -2.42991626e-01 9.64201868e-01
1.40601769e-01 2.55919993e-01 3.61370504e-01 -4.11700457e-01
1.74119040e-01 2.37040982e-01 -2.60470092e-01 -2.74051756e-01
3.04286517e-02 8.71242523e-01 -9.17115867e-01 -6.10764861e-01
3.67152572e-01 2.67362177e-01 -8.01102281e-01 7.50939608e-01
-5.96917689e-01 6.33989811e-01 -5.04373789e-01 -5.19286171e-02
-7.89299309e-02 -1.57594696e-01 1.50528848e-01 3.70571107e-01
-3.67689848e-01 7.90362716e-01 -9.96308744e-01 1.42929113e+00
-8.03309977e-01 5.11384547e-01 3.11257523e-02 -3.77407283e-01
9.02092755e-01 6.72659874e-01 2.45705754e-01 -4.19337958e-01
4.68222685e-02 -2.65538469e-02 4.91894811e-01 -6.53117180e-01
7.26981461e-01 -8.02271068e-02 4.52464260e-02 9.23945963e-01
5.27960539e-01 -7.70079792e-01 4.97330844e-01 4.67940480e-01
8.17388833e-01 1.56340554e-01 1.16982132e-01 -4.30760264e-01
3.10516655e-01 -5.32749072e-02 4.29389477e-02 9.89817679e-01
-6.20661750e-02 3.37918937e-01 5.65672576e-01 1.46647349e-01
-6.29300237e-01 -7.13751256e-01 1.87683050e-02 1.63624096e+00
-5.61402380e-01 -5.34115911e-01 -1.10251570e+00 -7.21849203e-01
-2.81361073e-01 1.29865575e+00 -3.58303785e-01 -3.92468013e-02
-5.42657614e-01 -4.24821705e-01 8.44384134e-01 2.75766462e-01
2.95027524e-01 -1.26941061e+00 -5.96444845e-01 4.24459368e-01
-2.26510629e-01 -1.00052774e+00 -4.49157149e-01 1.90185636e-01
-7.34497428e-01 -9.06966150e-01 -4.67198908e-01 -5.80464303e-01
2.34075442e-01 -1.42083587e-02 1.40881419e+00 1.32747918e-01
1.33564129e-01 1.02341461e+00 -4.87049192e-01 -5.04830420e-01
-1.28764808e+00 3.50341439e-01 -2.20732823e-01 -8.43982399e-01
1.88314617e-01 -5.42630196e-01 -2.01143906e-01 2.17515230e-01
-6.54641509e-01 1.56449899e-01 4.23917234e-01 9.70627367e-01
-7.12765679e-02 -2.16790050e-01 7.24309623e-01 -1.15410817e+00
1.49812877e+00 -2.88257182e-01 -6.05366826e-01 4.59489226e-01
-7.32829630e-01 1.51310429e-01 4.36345518e-01 -6.36211216e-01
-1.53732967e+00 -4.03554350e-01 -4.50180739e-01 -2.37921193e-01
-3.66148323e-01 4.57686692e-01 3.11621167e-02 -1.39905632e-01
7.73921251e-01 1.07208930e-01 1.33785367e-01 -4.90935206e-01
6.20275676e-01 5.92024505e-01 1.32160038e-01 -1.18607938e+00
4.12945479e-01 -4.82780725e-01 -4.26622748e-01 -8.68175387e-01
-7.53122449e-01 -1.05786771e-01 -4.52915609e-01 -2.62226671e-01
7.09324002e-01 -9.85035062e-01 -9.33637202e-01 4.27722745e-02
-1.08693790e+00 -1.10749853e+00 -5.98536611e-01 5.38982034e-01
-7.29866326e-01 1.45540640e-01 -8.61050487e-01 -9.57015932e-01
-1.98368400e-01 -1.55909991e+00 6.41710877e-01 5.58498442e-01
-6.01571500e-01 -1.53428233e+00 4.41219777e-01 9.36875820e-01
5.32461643e-01 -2.45690554e-01 1.00661957e+00 -1.46153963e+00
-4.43552494e-01 3.82071912e-01 3.46766919e-01 2.52926469e-01
-1.49036214e-01 2.43358910e-01 -1.18760228e+00 -1.99018657e-01
-5.39564453e-02 -9.73264873e-01 3.33684653e-01 2.12772235e-01
7.00746179e-01 -5.94436646e-01 1.23155333e-01 -2.31996528e-03
7.38285422e-01 2.20973954e-01 1.54521197e-01 2.03636527e-01
3.49743724e-01 9.48730528e-01 4.94287610e-01 3.83472711e-01
6.60697401e-01 6.64156675e-01 2.52694003e-02 3.26802462e-01
-4.59892541e-01 -3.96719545e-01 3.79285336e-01 1.18711591e+00
1.85499676e-02 -3.62806231e-01 -9.64770675e-01 5.45149803e-01
-1.64555955e+00 -9.71415222e-01 1.36174858e-02 1.95125401e+00
1.48935902e+00 1.46127805e-01 4.30981666e-01 -5.46330571e-01
2.30774134e-01 -7.16291666e-02 4.92510460e-02 -7.56929755e-01
2.61731982e-01 5.36784291e-01 -1.24596775e-01 1.07535183e+00
-3.69213253e-01 1.20094335e+00 6.59642792e+00 9.02219653e-01
-7.63030112e-01 8.11220482e-02 5.93231797e-01 1.12486601e-01
-5.12324452e-01 7.35758394e-02 -1.26921320e+00 3.05308253e-01
1.45685244e+00 -4.83704180e-01 4.24956411e-01 6.91732049e-01
3.28139186e-01 -3.06231733e-02 -1.20409751e+00 3.99242304e-02
-1.09620430e-01 -1.29262662e+00 -1.14509342e-02 1.30638629e-01
1.02844298e+00 -1.77084655e-01 1.84891224e-02 9.29565489e-01
1.26165390e+00 -1.08635974e+00 3.82338524e-01 1.50298566e-01
2.96163797e-01 -5.81872344e-01 5.68145335e-01 5.56875944e-01
-7.00801373e-01 9.40299034e-02 -8.39280337e-02 -3.65921333e-02
-8.24617147e-02 -6.14945292e-02 -1.82458866e+00 1.88982725e-01
3.86651993e-01 2.91293878e-02 -5.42357981e-01 7.94617057e-01
-4.42111790e-01 1.12903404e+00 -1.90326720e-01 -3.91440839e-01
4.53903347e-01 -1.11640006e-01 5.11666358e-01 1.35192299e+00
1.22297198e-01 4.42837447e-01 5.94742537e-01 9.97990489e-01
2.00041737e-02 9.87179875e-02 -4.60688353e-01 9.48394081e-05
9.15092826e-01 1.36742461e+00 -2.85627604e-01 -2.89539844e-01
-1.87794283e-01 1.87666729e-01 2.44405746e-01 2.17085570e-01
-3.37679237e-01 7.14259893e-02 4.28654224e-01 6.22036085e-02
-2.22903993e-02 1.16950750e-01 -9.18712541e-02 -6.33789182e-01
-7.44980574e-01 -1.41635537e+00 2.16244191e-01 -7.73350298e-01
-1.24816096e+00 6.39598429e-01 4.82424706e-01 -7.34188855e-01
-9.79564369e-01 -1.58724710e-01 -9.78074670e-01 1.03481400e+00
-9.62695241e-01 -1.07846522e+00 -2.36739859e-01 5.33236265e-01
9.33853686e-01 -3.90937507e-01 9.98242915e-01 -1.46839127e-01
-4.43238258e-01 7.32790112e-01 -2.24373460e-01 -1.69261009e-01
9.80015635e-01 -1.61308765e+00 3.13948393e-01 3.21141124e-01
3.11385226e-02 6.64480388e-01 1.07272220e+00 -6.65100932e-01
-9.80706990e-01 -7.87044168e-01 8.39738965e-01 -6.87198997e-01
8.24314177e-01 -6.39304146e-02 -9.93595421e-01 7.50356197e-01
9.54147398e-01 -8.35514307e-01 8.83392453e-01 2.99704999e-01
1.66994885e-01 3.96048933e-01 -1.14033008e+00 6.30576313e-01
6.46053851e-01 -2.21024066e-01 -9.75069046e-01 4.60420161e-01
9.28623915e-01 -8.62972677e-01 -1.39371896e+00 7.20834136e-02
2.72229105e-01 -1.07862461e+00 6.79962277e-01 -6.41161025e-01
5.73751688e-01 4.85321164e-01 3.50711346e-01 -1.69214416e+00
-5.97022101e-03 -1.00129902e+00 1.56330794e-01 1.71595395e+00
4.73143101e-01 -4.31098282e-01 8.27213347e-01 8.77628565e-01
-3.77420068e-01 -7.19052255e-01 -6.37410879e-01 -4.93761599e-01
6.32526696e-01 -5.24042547e-02 6.74545348e-01 9.53493476e-01
1.32925123e-01 7.52761602e-01 -7.30596110e-02 -3.05663377e-01
2.98789769e-01 1.62719358e-02 1.16408885e+00 -1.25153291e+00
-5.61500967e-01 -6.51420534e-01 3.31838727e-01 -1.04583704e+00
3.13330740e-01 -5.34879923e-01 1.81913897e-01 -1.33235371e+00
5.49780391e-02 -7.31969297e-01 3.90965700e-01 4.61539179e-01
-3.50700498e-01 -3.35448474e-01 3.41195822e-01 -5.20686880e-02
-5.95918894e-01 5.80335557e-01 1.54982424e+00 2.46123925e-01
-5.68561852e-01 3.57181579e-01 -7.11154759e-01 6.52548313e-01
6.79700971e-01 -1.61476076e-01 -7.95336068e-01 -1.16723061e-01
-2.22575754e-01 5.48579395e-01 6.49780482e-02 -6.57804549e-01
3.53212059e-01 -4.92413312e-01 -5.15597016e-02 -3.15868109e-01
3.75920683e-01 -3.99202645e-01 -2.18155280e-01 3.03025097e-01
-8.61311257e-01 3.33049223e-02 4.53760982e-01 -1.02050833e-01
-1.34265155e-01 -7.09032595e-01 5.83311200e-01 -3.26143622e-01
-2.63685673e-01 -1.97730586e-01 -6.53737247e-01 5.71742058e-01
8.59960139e-01 -2.08806600e-02 -6.42559946e-01 -7.61393845e-01
-6.60794199e-01 5.42927444e-01 1.45770580e-01 4.66179192e-01
2.97337472e-01 -8.33469689e-01 -8.75969291e-01 2.48275530e-02
-1.35451302e-01 -4.03137580e-02 2.49387458e-01 6.43893003e-01
-2.79083937e-01 7.79321671e-01 -1.39553249e-01 -4.98290569e-01
-1.61841500e+00 -6.03396632e-02 3.60367358e-01 -7.27592349e-01
-3.05456668e-01 1.05754554e+00 1.89762786e-01 -8.51761281e-01
4.18421835e-01 -3.48920166e-01 -5.04248381e-01 1.12509236e-01
3.03975195e-01 4.86604244e-01 -1.16528153e-01 -1.44860104e-01
3.24539185e-01 -9.10558850e-02 -1.58907160e-01 -4.62413281e-01
1.10488272e+00 -2.14993116e-02 2.51387209e-01 4.41635311e-01
5.13288736e-01 2.82503456e-01 -1.18251026e+00 -1.33033201e-01
-5.25645018e-02 1.06097665e-02 -2.78620899e-01 -1.23257327e+00
-5.18072963e-01 7.70207644e-01 3.17755759e-01 5.56595385e-01
5.57542443e-01 8.38486627e-02 2.75436223e-01 3.93794239e-01
6.23109117e-02 -8.17134559e-01 3.91535491e-01 8.20182085e-01
7.64645875e-01 -1.18756831e+00 -3.62388529e-02 -2.62609094e-01
-9.11153018e-01 1.08183444e+00 1.26473927e+00 3.88401300e-01
2.03345045e-01 2.73549497e-01 1.58330843e-01 -1.47743881e-01
-1.46953821e+00 -2.27674525e-02 3.16465348e-01 4.24444079e-01
1.10626185e+00 8.07964951e-02 -2.80767024e-01 4.89873767e-01
-9.22393322e-01 -1.33237183e-01 7.29761302e-01 7.24795759e-01
-6.19098306e-01 -1.38411355e+00 -2.78369248e-01 2.22078517e-01
-3.29757452e-01 -1.67632073e-01 -7.92101920e-01 1.06205010e+00
-1.68791726e-01 1.21532583e+00 -2.10074246e-01 -1.86681241e-01
4.72470075e-01 3.01087976e-01 4.77928311e-01 -1.16994417e+00
-1.57661462e+00 1.06559657e-01 5.78045905e-01 -1.29007593e-01
-2.16685265e-01 -5.37184596e-01 -8.58888865e-01 -3.35757852e-01
-5.23054302e-01 8.15559864e-01 3.29847246e-01 9.86694813e-01
8.50545317e-02 7.54444420e-01 3.19657475e-01 -4.79568094e-01
-1.17596555e+00 -1.55420780e+00 -7.82723576e-02 5.33724837e-02
-1.06986418e-01 -4.47656721e-01 -2.84024149e-01 -7.30640888e-02] | [12.559064865112305, 8.104456901550293] |
cf9d9302-3d15-433a-9c2c-39f393610dd8 | bridging-continuous-and-discrete-spaces | 2305.14599 | null | https://arxiv.org/abs/2305.14599v1 | https://arxiv.org/pdf/2305.14599v1.pdf | Bridging Continuous and Discrete Spaces: Interpretable Sentence Representation Learning via Compositional Operations | Traditional sentence embedding models encode sentences into vector representations to capture useful properties such as the semantic similarity between sentences. However, in addition to similarity, sentence semantics can also be interpreted via compositional operations such as sentence fusion or difference. It is unclear whether the compositional semantics of sentences can be directly reflected as compositional operations in the embedding space. To more effectively bridge the continuous embedding and discrete text spaces, we explore the plausibility of incorporating various compositional properties into the sentence embedding space that allows us to interpret embedding transformations as compositional sentence operations. We propose InterSent, an end-to-end framework for learning interpretable sentence embeddings that supports compositional sentence operations in the embedding space. Our method optimizes operator networks and a bottleneck encoder-decoder model to produce meaningful and interpretable sentence embeddings. Experimental results demonstrate that our method significantly improves the interpretability of sentence embeddings on four textual generation tasks over existing approaches while maintaining strong performance on traditional semantic similarity tasks. | ['Dong Yu', 'Muhao Chen', 'Hongming Zhang', 'Kaiqiang Song', 'Wenlin Yao', 'James Y. Huang'] | 2023-05-24 | null | null | null | null | ['sentence-embeddings', 'sentence-embeddings', 'semantic-textual-similarity', 'semantic-similarity'] | ['methodology', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 7.15752602e-01 5.41629374e-01 -1.07671268e-01 -9.33623791e-01
-4.64875609e-01 -6.62864387e-01 8.50754261e-01 7.18434930e-01
-3.48195821e-01 3.07752699e-01 1.21718025e+00 -6.94190025e-01
1.46438971e-01 -8.15313756e-01 -4.72398162e-01 -8.85182712e-03
1.33369744e-01 3.25409204e-01 -2.62706906e-01 -5.23431480e-01
1.85592830e-01 9.25246812e-03 -1.25107169e+00 7.20898509e-01
7.87339330e-01 5.07502913e-01 9.31460485e-02 7.82662749e-01
-5.66353977e-01 9.05033708e-01 -5.23298264e-01 -6.84842229e-01
-2.50939950e-02 -5.76037645e-01 -1.01569378e+00 -1.43668741e-01
3.98859382e-01 -2.79302388e-01 -3.01748753e-01 8.76726270e-01
8.13635811e-02 2.48873189e-01 8.17885935e-01 -1.07640743e+00
-1.64310730e+00 1.17062640e+00 -2.63671689e-02 2.18418330e-01
5.46910942e-01 -6.86041033e-03 1.91947043e+00 -1.03457427e+00
4.86426085e-01 1.66895962e+00 5.90876162e-01 5.40502369e-01
-1.59521651e+00 -5.39052114e-02 2.62033463e-01 1.94748305e-02
-6.95608556e-01 -4.24827248e-01 8.24032426e-01 -2.26965964e-01
1.35564661e+00 5.28489232e-01 3.95590365e-01 1.07475781e+00
2.60288268e-01 7.69085824e-01 4.53135699e-01 -5.11960804e-01
1.67494968e-01 4.95339148e-02 5.11856914e-01 6.63742125e-01
2.02978566e-01 -4.23994660e-01 -5.08686960e-01 -2.21034452e-01
3.90599847e-01 3.04478735e-01 -1.54716328e-01 -1.48426831e-01
-1.26301432e+00 1.37054038e+00 7.76727557e-01 2.19280645e-01
-1.51198819e-01 6.44995689e-01 9.58877325e-01 5.32222748e-01
7.93781102e-01 9.75862384e-01 -4.32902426e-01 -1.44833848e-01
-3.77015144e-01 2.26138219e-01 6.35807037e-01 7.03480601e-01
5.06254792e-01 -1.59743994e-01 -4.55238581e-01 7.01906323e-01
4.72309351e-01 2.39802986e-01 5.41905284e-01 -9.52893376e-01
6.69488728e-01 8.09068859e-01 -1.01374343e-01 -1.00388217e+00
-3.87060679e-02 -7.11172959e-03 -3.38211358e-01 -3.79226357e-01
-1.22550257e-01 1.59830078e-01 -3.88378590e-01 1.90373981e+00
-4.16487493e-02 -2.41464842e-02 3.52718920e-01 7.68859386e-01
8.43028605e-01 6.78687811e-01 1.17541879e-01 1.51484042e-01
1.67620981e+00 -1.02130806e+00 -7.09170222e-01 -5.93541920e-01
1.02972448e+00 -5.03425419e-01 1.68492579e+00 -4.94004607e-01
-1.27251375e+00 -4.70670491e-01 -1.15933669e+00 -7.97954619e-01
-2.43606076e-01 -2.60133207e-01 9.49056685e-01 3.59329373e-01
-9.91089463e-01 6.79659724e-01 -9.06314254e-01 -3.32187176e-01
2.98434883e-01 1.27952725e-01 -3.38966370e-01 1.12985902e-01
-1.43101144e+00 1.00831366e+00 4.32626069e-01 -7.39780590e-02
-1.00933097e-01 -9.34834421e-01 -1.53871071e+00 4.29245979e-01
-1.97772175e-01 -1.34634185e+00 1.51048696e+00 -8.77532899e-01
-1.17104733e+00 6.18380308e-01 -7.07602084e-01 -7.41441309e-01
-1.96275264e-01 -2.54574031e-01 -1.23110570e-01 1.91580281e-01
2.82046974e-01 6.62361324e-01 4.98520881e-01 -7.97539413e-01
-1.90028414e-01 -2.48310938e-01 4.57878441e-01 4.42157060e-01
-8.48066151e-01 2.19679892e-01 3.88749331e-01 -8.03555727e-01
2.81716157e-02 -6.98208749e-01 -1.34390563e-01 4.06051993e-01
-2.79547542e-01 -5.80991983e-01 5.13277650e-01 -5.29798567e-01
1.22676551e+00 -2.08501363e+00 3.32884878e-01 -2.94595808e-01
4.30157483e-01 -3.12237799e-01 -3.63766491e-01 8.67521465e-01
-2.30798289e-01 5.00921845e-01 -5.25234222e-01 -7.01589525e-01
5.42472124e-01 4.79752332e-01 -9.17830944e-01 9.43571031e-02
6.70257986e-01 1.42476428e+00 -1.23963988e+00 -2.33924210e-01
9.31934938e-02 3.80236536e-01 -7.86736012e-01 1.00173339e-01
-2.36318216e-01 -3.27420384e-01 -2.89955705e-01 -3.35434116e-02
2.14002460e-01 -2.04870597e-01 3.40999752e-01 -1.90237626e-01
4.17792618e-01 1.20499516e+00 -5.58325529e-01 2.01958394e+00
-8.64269853e-01 9.33595717e-01 -3.88651639e-01 -1.09664571e+00
7.50764728e-01 2.95293421e-01 -1.08095139e-01 -3.00209880e-01
-9.80551690e-02 5.74273011e-03 1.45013025e-02 -4.32861239e-01
1.09547126e+00 -6.68901086e-01 -4.52040970e-01 1.12929344e+00
-5.13933450e-02 -3.92505169e-01 2.76776291e-02 6.01726830e-01
1.12150824e+00 -2.04619139e-01 2.79540539e-01 -2.56082386e-01
3.04794848e-01 -3.03246051e-01 1.06089920e-01 3.68310004e-01
2.19499215e-01 4.22721297e-01 6.99122548e-01 -5.09735227e-01
-1.28273988e+00 -1.66270053e+00 -9.48946700e-02 1.25641465e+00
5.68784438e-02 -8.39029372e-01 -5.08202434e-01 -7.20464647e-01
2.87804127e-01 1.29329336e+00 -7.82609344e-01 -6.64593577e-01
-4.65955943e-01 -3.87468785e-01 7.57116437e-01 1.08978713e+00
-2.38134399e-01 -9.37361836e-01 -2.44257420e-01 2.18218908e-01
-2.11904526e-01 -1.00293136e+00 -9.50787783e-01 9.83307138e-02
-9.89749849e-01 -6.62085891e-01 2.09776893e-01 -8.69317055e-01
8.82948875e-01 3.31173718e-01 1.29961944e+00 1.98922120e-02
-1.81893498e-01 1.88810185e-01 -3.93658340e-01 -2.05997452e-01
-8.44980419e-01 -4.75672670e-02 3.87788303e-02 -2.59972394e-01
6.02196693e-01 -5.44588566e-01 -3.34907174e-01 -3.87034059e-01
-1.20530832e+00 1.86118290e-01 5.60152158e-02 1.14208043e+00
-6.10347055e-02 -4.54420865e-01 6.86797142e-01 -9.17390466e-01
1.38806772e+00 -3.86081278e-01 1.45819783e-01 2.81217217e-01
-4.96599793e-01 8.88462722e-01 8.60406995e-01 -1.83490932e-01
-9.15838361e-01 -4.82606769e-01 -9.65800602e-03 -1.78054303e-01
2.03936130e-01 6.52689934e-01 9.73666236e-02 6.46858692e-01
6.47444367e-01 2.00474277e-01 3.69088352e-01 -2.41937369e-01
9.61176753e-01 8.04927111e-01 3.61984938e-01 -6.84228480e-01
7.99200475e-01 4.12067503e-01 -2.37389177e-01 -6.38892591e-01
-1.01389956e+00 -1.53714165e-01 -4.92504358e-01 6.13531590e-01
1.05578470e+00 -8.09978068e-01 -4.36717331e-01 -5.31818569e-01
-1.70843506e+00 2.54512042e-01 -7.13927567e-01 2.82091707e-01
-3.63530815e-01 6.25039339e-01 -1.09871292e+00 -4.93505359e-01
-6.50377393e-01 -1.05473208e+00 1.41175604e+00 -2.09310323e-01
-1.15908456e+00 -1.60514534e+00 -8.67559239e-02 3.73088717e-01
4.52452004e-01 -2.37372075e-03 1.41380489e+00 -7.03319073e-01
-1.99452788e-01 -9.76604521e-02 -1.72169447e-01 5.15728712e-01
3.55312556e-01 -2.10559115e-01 -8.76495659e-01 -1.09548382e-01
3.37627307e-02 -3.00680935e-01 9.60434139e-01 -1.93644892e-02
1.05968404e+00 -5.90767086e-01 -5.97763760e-03 4.60272282e-01
1.05780959e+00 -3.38905752e-01 4.06586021e-01 2.16091387e-02
5.90050161e-01 5.93968928e-01 4.40831855e-02 1.88325435e-01
5.64390719e-01 3.98995250e-01 1.77205309e-01 5.57498336e-02
-2.12825518e-02 -7.54778862e-01 5.91874123e-01 1.19656670e+00
5.22861123e-01 -6.72535822e-02 -5.24594665e-01 3.93249065e-01
-1.69723201e+00 -1.17420101e+00 4.47795540e-02 1.72567534e+00
1.09140742e+00 2.11281389e-01 -4.99837220e-01 1.03116177e-01
4.42776889e-01 5.38641393e-01 -3.37106138e-01 -1.25145984e+00
4.45940346e-02 5.00166893e-01 7.40125775e-02 8.26687098e-01
-8.45997453e-01 8.20584536e-01 6.32230330e+00 2.98354238e-01
-8.19502831e-01 2.70139635e-01 3.52280229e-01 -2.55540609e-01
-1.33861768e+00 1.18355021e-01 -3.15377265e-01 3.53438437e-01
9.17935967e-01 -5.57897627e-01 3.89624327e-01 6.93284690e-01
2.47619063e-01 5.42265773e-01 -1.78927350e+00 6.12381220e-01
3.71587843e-01 -1.68204200e+00 3.65839303e-01 -2.81402797e-01
4.79330420e-01 -2.37711400e-01 -1.90094933e-02 2.28236094e-01
3.50730747e-01 -1.24149442e+00 7.48997152e-01 1.35152444e-01
7.57159054e-01 -4.99435872e-01 6.22490346e-01 -3.75994976e-04
-1.24460113e+00 -1.00472584e-01 -4.31373596e-01 -6.88835979e-01
3.88610989e-01 2.34320134e-01 -1.13854909e+00 4.07066256e-01
-3.27958688e-02 1.03744841e+00 -6.13164365e-01 -3.88711244e-02
-6.67558253e-01 3.99915606e-01 1.43582433e-01 -4.61637855e-01
2.69123733e-01 -1.69757068e-01 4.38441217e-01 1.10721850e+00
8.72472301e-02 -2.40201905e-01 -1.25914305e-01 1.43019986e+00
-3.41343343e-01 -1.84059307e-01 -9.40652728e-01 -6.20731294e-01
7.71409452e-01 9.40863192e-01 -3.09199244e-01 -4.18824911e-01
-5.09290934e-01 1.36285663e+00 4.74859267e-01 2.74703681e-01
-8.47752273e-01 -5.44006288e-01 1.31707215e+00 -1.06877036e-01
4.53548171e-02 -3.75078201e-01 -6.82536781e-01 -1.43019605e+00
3.16554725e-01 -5.42039931e-01 4.69219722e-02 -9.15581822e-01
-1.61252248e+00 4.02479321e-01 -1.29745692e-01 -7.95275927e-01
-4.39294010e-01 -7.17038929e-01 -1.07643223e+00 1.02729523e+00
-1.05578017e+00 -1.22171819e+00 2.70050704e-01 -8.17673728e-02
8.85606170e-01 6.99096173e-02 1.20154333e+00 -2.48860866e-01
-2.33052835e-01 6.94991112e-01 -7.56533220e-02 1.41377091e-01
3.85232568e-01 -1.53398585e+00 1.11878276e+00 8.18478525e-01
4.67686445e-01 1.28389525e+00 7.63418376e-01 -2.28109241e-01
-1.44764876e+00 -1.14740849e+00 1.58826435e+00 -8.98620784e-01
1.03941512e+00 -8.06401491e-01 -8.64766657e-01 9.64874268e-01
5.98512471e-01 -2.34674513e-01 1.08163297e+00 4.43410009e-01
-8.34236324e-01 9.99813005e-02 -6.72069848e-01 9.46452677e-01
1.21541357e+00 -1.34524369e+00 -1.39238429e+00 3.99061888e-01
1.76889443e+00 1.93759333e-02 -8.43800783e-01 -8.06308985e-02
2.95782894e-01 -3.38238537e-01 1.12872577e+00 -1.25229990e+00
1.20703697e+00 -2.54358858e-01 -2.79287666e-01 -1.59279227e+00
-3.22998166e-01 -3.37127060e-01 -9.95278656e-02 1.04333305e+00
7.12100506e-01 -7.34280050e-01 4.29268330e-01 8.74124885e-01
-3.93111497e-01 -8.62694442e-01 -7.66204655e-01 -6.87519133e-01
2.81363070e-01 -5.37232459e-01 9.63179410e-01 9.81085181e-01
7.20141530e-01 1.03766191e+00 2.53847212e-01 -1.31850600e-01
2.89925158e-01 2.79739529e-01 3.24986517e-01 -1.03337467e+00
-4.95325387e-01 -5.81785440e-01 -6.05496824e-01 -1.20255423e+00
7.87477672e-01 -1.68868208e+00 -8.80667716e-02 -1.90756643e+00
3.03184628e-01 9.72192362e-02 -4.61486094e-02 4.53188479e-01
-6.09943211e-01 -5.69173731e-02 3.23231846e-01 -1.49782151e-01
-4.19845909e-01 1.02644467e+00 9.44838166e-01 -3.54476154e-01
2.33321592e-01 -6.74018681e-01 -1.22525203e+00 4.74403679e-01
6.30583525e-01 -2.77766675e-01 -8.59947681e-01 -1.18998778e+00
5.26938736e-01 -2.12940499e-01 3.48139226e-01 -2.60576159e-01
6.17048182e-02 -8.39360803e-02 1.42549174e-02 -2.78323162e-02
4.03299451e-01 -4.92587626e-01 -3.99059623e-01 4.76051986e-01
-1.15870976e+00 5.20091057e-01 -6.88052624e-02 6.32389545e-01
-3.48406285e-01 -2.28043541e-01 2.41604954e-01 8.05983618e-02
-4.07956153e-01 -1.92451909e-01 -2.73655742e-01 3.72847646e-01
7.52547920e-01 -2.61660784e-01 -4.90155399e-01 -3.68420124e-01
-4.27386224e-01 2.09751502e-01 5.81906199e-01 9.02234614e-01
9.01758015e-01 -1.57692409e+00 -6.84784889e-01 2.94994563e-01
3.35629374e-01 -1.79455414e-01 -1.99744329e-01 4.65478361e-01
-5.04528701e-01 4.53076154e-01 2.32597977e-01 -2.38852933e-01
-1.26990974e+00 4.35081005e-01 1.84257939e-01 -3.09231188e-02
-6.10538185e-01 9.74083602e-01 3.80063027e-01 -8.16376686e-01
-1.24599032e-01 -9.54431593e-01 1.95804387e-01 -1.73988000e-01
5.46991408e-01 -7.12122070e-03 -1.99456275e-01 -3.66073132e-01
-3.64909381e-01 1.21213853e-01 -4.99922149e-02 -1.67531401e-01
1.32034922e+00 -1.48710459e-01 -5.82470894e-01 7.17207372e-01
1.76305604e+00 -1.69311702e-01 -7.63998151e-01 -1.60116047e-01
2.03976139e-01 -3.73674929e-01 -2.61834174e-01 -2.70348608e-01
-3.45447063e-01 1.18219841e+00 -2.79503524e-01 2.97147632e-01
6.69929445e-01 2.51805395e-01 1.23239398e+00 5.71243525e-01
-1.30985051e-01 -8.27374995e-01 3.35936069e-01 7.29207695e-01
1.05270374e+00 -1.10491562e+00 -2.34458804e-01 -4.06616807e-01
-7.48010039e-01 1.38467705e+00 3.38279188e-01 -2.09887609e-01
2.92989463e-01 2.30458155e-01 -3.06376815e-01 -1.43901154e-01
-1.17842078e+00 3.00161123e-01 3.60111445e-01 1.94584996e-01
1.04448080e+00 2.82782972e-01 -4.82768863e-01 7.27415383e-01
-6.07307971e-01 -4.37125027e-01 5.14312565e-01 7.79625773e-01
-4.05453533e-01 -1.24116433e+00 9.91344675e-02 5.78039110e-01
-8.73898044e-02 -6.97768033e-01 -5.44075966e-01 1.59820735e-01
-2.69747674e-01 8.76709282e-01 6.15017593e-01 -2.86311775e-01
1.95879892e-01 3.47172916e-01 3.42281222e-01 -1.14011145e+00
-6.19965732e-01 -7.40892112e-01 4.14110750e-01 -2.86522567e-01
5.03142737e-02 -4.68965173e-01 -1.64911389e+00 -3.51278424e-01
-1.01881169e-01 3.20006788e-01 5.27275920e-01 1.02195764e+00
5.35654128e-01 6.73715889e-01 5.80291986e-01 -3.86327296e-01
-1.21532750e+00 -9.90375340e-01 -3.20295036e-01 1.02381313e+00
3.67017061e-01 -8.23071375e-02 -5.37132800e-01 2.73324102e-01] | [10.774727821350098, 8.798331260681152] |
bda1fafb-455c-4b3a-874b-34016e3d1f7e | introducing-the-prague-discourse-treebank-10 | null | null | https://aclanthology.org/I13-1011 | https://aclanthology.org/I13-1011.pdf | Introducing the Prague Discourse Treebank 1.0 | null | ["Eva Haji{\\v{c}}ov{\\'a}", "{\\v{S}}{\\'a}rka Zik{\\'a}nov{\\'a}", "Pavl{\\'\\i}na J{\\'\\i}nov{\\'a}", "Ji{\\v{r}}{\\'\\i} M{\\'\\i}rovsk{\\'y}", "Lucie Pol{\\'a}kov{\\'a}", 'Anna Nedoluzhko'] | 2013-10-01 | introducing-the-prague-discourse-treebank-10-1 | https://aclanthology.org/I13-1011 | https://aclanthology.org/I13-1011.pdf | ijcnlp-2013-10 | ['morphological-tagging'] | ['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] | [-7.290279388427734, 3.7589826583862305] |
d33f743a-5ca1-4a8d-94b5-c42f12893b5d | source-summary-entity-aggregation-in | null | null | https://aclanthology.org/2022.coling-1.526 | https://aclanthology.org/2022.coling-1.526.pdf | Source-summary Entity Aggregation in Abstractive Summarization | In a text, entities mentioned earlier can be referred to in later discourse by a more general description. For example, Celine Dion and Justin Bieber can be referred to by Canadian singers or celebrities. In this work, we study this phenomenon in the context of summarization, where entities from a source text are generalized in the summary. We call such instances source-summary entity aggregations. We categorize these aggregations into two types and analyze them in the Cnn/Dailymail corpus, showing that they are reasonably frequent. We then examine how well three state-of-the-art summarization systems can generate such aggregations within summaries. We also develop techniques to encourage them to generate more aggregations. Our results show that there is significant room for improvement in producing semantically correct aggregations. | ['Jackie Chi Kit Cheung', 'Annie Louis', 'José Ángel González'] | null | null | null | null | coling-2022-10 | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 2.20364839e-01 7.49584496e-01 -2.22351447e-01 -3.79346699e-01
-1.11458743e+00 -9.40318286e-01 1.01186562e+00 7.59349108e-01
-2.44053185e-01 1.19231439e+00 1.19796252e+00 -7.64562637e-02
1.06834896e-01 -6.71708882e-01 -7.40507722e-01 -1.32661819e-01
1.10098921e-01 5.02750278e-01 2.33854771e-01 -4.40919816e-01
5.19785047e-01 2.53101766e-01 -1.25452805e+00 8.24834645e-01
1.18689501e+00 2.70629913e-01 -7.34574050e-02 7.81219304e-01
-4.29553211e-01 8.64587426e-01 -1.62733626e+00 -6.49761498e-01
-1.64523497e-01 -7.70037711e-01 -1.24778724e+00 8.93863216e-02
9.09268677e-01 -2.96557993e-01 -1.33039774e-02 8.13993514e-01
5.67891777e-01 3.60637993e-01 9.32801962e-01 -9.69638109e-01
-7.10888028e-01 1.48894477e+00 -3.80130559e-01 4.35450166e-01
6.08653247e-01 -3.26423943e-01 1.28524411e+00 -5.84349930e-01
1.06440222e+00 1.28163123e+00 4.82860714e-01 7.71809518e-01
-8.47584724e-01 -3.05106670e-01 2.06405800e-02 -2.71950401e-02
-9.66227591e-01 -6.79318607e-01 4.48640287e-01 -9.49795693e-02
1.13199091e+00 7.36943483e-01 5.96893132e-01 7.31430948e-01
1.01838581e-01 1.24823165e+00 3.16757768e-01 -4.21179175e-01
9.13291052e-02 -1.64969474e-01 4.60895866e-01 4.57402974e-01
7.07106054e-01 -7.16035843e-01 -6.58506691e-01 -2.08545104e-01
2.84238845e-01 -5.10936677e-01 -4.50262129e-01 5.23784339e-01
-1.38029075e+00 8.12154472e-01 4.37094271e-01 5.28826356e-01
-4.97856289e-01 2.51012266e-01 6.58611178e-01 -9.78801250e-02
6.24504983e-01 1.07184303e+00 -2.23626420e-01 -2.14656547e-01
-1.27969456e+00 7.69904733e-01 1.28625381e+00 1.44608402e+00
4.43515599e-01 -1.54813737e-01 -5.00799060e-01 6.65053844e-01
-2.43321061e-01 -7.77422339e-02 3.81705403e-01 -1.31104434e+00
7.89688945e-01 5.77299178e-01 2.45893553e-01 -9.06597614e-01
-2.46608704e-01 -3.12880039e-01 -7.79885232e-01 -5.45554042e-01
2.64127195e-01 -4.06852543e-01 -4.89233762e-01 1.43018341e+00
-6.26492649e-02 -1.40143871e-01 3.39812279e-01 3.63222063e-01
1.38636923e+00 7.42127359e-01 -7.91758299e-02 -4.49735850e-01
1.39119112e+00 -1.19743359e+00 -1.09856868e+00 -1.34221107e-01
8.48882854e-01 -7.03520596e-01 7.34649658e-01 5.33973277e-02
-1.44651532e+00 -2.78308928e-01 -9.44047332e-01 -3.43181044e-01
-3.06785733e-01 3.47547412e-01 5.36614180e-01 2.16322079e-01
-9.23777461e-01 8.65019739e-01 -7.23159492e-01 -7.41619825e-01
3.25312495e-01 6.24344386e-02 -2.08806694e-01 4.38443393e-01
-1.15162969e+00 9.05228555e-01 7.63481259e-01 -1.55945525e-01
-3.50365490e-01 -5.79656124e-01 -8.63742530e-01 1.90467775e-01
5.09850383e-01 -1.12108874e+00 1.76028991e+00 -8.39180350e-01
-1.01680875e+00 9.49944854e-01 -5.48260212e-01 -9.00219023e-01
3.60759377e-01 -4.46787119e-01 -4.34986174e-01 3.35387170e-01
3.94664913e-01 8.61111820e-01 1.72243580e-01 -1.23957276e+00
-9.95916545e-01 -3.82006690e-02 4.67226893e-01 4.84483689e-01
-8.24289545e-02 3.52410376e-01 -2.65018553e-01 -8.92774761e-01
-1.29284382e-01 -7.25018501e-01 -1.63000729e-02 -5.16427815e-01
-1.14985347e+00 -6.52433097e-01 4.82607841e-01 -7.93169856e-01
1.80563891e+00 -1.57982373e+00 1.94038600e-01 -3.18993658e-01
2.98032314e-01 2.61868507e-01 2.28140399e-01 7.83519149e-01
1.42222077e-01 7.80803084e-01 -2.78760225e-01 -3.52915466e-01
-6.41871570e-03 1.59669042e-01 -4.58544254e-01 -1.61496356e-01
1.79620922e-01 1.12562454e+00 -1.11159766e+00 -5.98686934e-01
-4.00246173e-01 8.80205855e-02 -4.18110073e-01 4.91986051e-02
-3.81203949e-01 2.66833603e-01 -6.03965938e-01 3.51688325e-01
4.57575992e-02 -2.41412565e-01 -1.05220005e-01 -4.51930761e-01
-2.12275699e-01 7.96426594e-01 -6.65589273e-01 1.67117715e+00
-1.02149576e-01 8.99238586e-01 -1.93533286e-01 -7.23164856e-01
7.19183207e-01 4.67595696e-01 -7.35864090e-03 1.19338229e-01
7.51303434e-02 3.27377230e-01 4.61207218e-02 -3.80357057e-01
1.49599075e+00 1.39631897e-01 -3.25159520e-01 4.71284330e-01
-9.37622339e-02 -5.10907173e-01 9.34048653e-01 7.18495369e-01
1.01997840e+00 -1.24880582e-01 7.38579988e-01 -3.12964827e-01
3.28967810e-01 4.74582404e-01 3.61038864e-01 1.10931098e+00
2.03298897e-01 7.45983243e-01 6.91501737e-01 2.63444595e-02
-1.15650833e+00 -7.24038005e-01 1.62424713e-01 8.45695376e-01
-8.34330022e-02 -1.01035309e+00 -8.70914519e-01 -8.22920024e-01
-1.94790393e-01 1.28642309e+00 -5.09086967e-01 4.42867838e-02
-1.02370393e+00 -3.93029302e-01 8.78936410e-01 7.58708358e-01
5.13355017e-01 -1.13497663e+00 -7.85590470e-01 2.56870419e-01
-5.47804475e-01 -9.40837145e-01 -5.92348516e-01 -5.09640984e-02
-8.37774813e-01 -8.52113307e-01 -8.69069457e-01 -5.93346775e-01
5.31912267e-01 7.53436983e-02 1.60017776e+00 3.06205124e-01
1.25178620e-01 3.11527848e-01 -5.93729794e-01 -6.94106519e-01
-9.58378255e-01 6.18221223e-01 -1.95844322e-01 -4.80029702e-01
1.53880909e-01 -3.07844907e-01 -3.82666975e-01 -2.48160392e-01
-1.11352122e+00 1.95826352e-01 2.71474391e-01 4.98129785e-01
2.95665741e-01 -3.23102146e-01 8.88911068e-01 -1.27452838e+00
1.23210382e+00 -4.60157603e-01 1.70825735e-01 3.86896253e-01
1.10025797e-02 1.15657188e-01 6.26041114e-01 -1.98711112e-01
-9.91448462e-01 -3.65693301e-01 -1.47126138e-01 1.69471964e-01
-1.89157590e-01 7.78005600e-01 1.12459287e-02 6.88751519e-01
9.50354159e-01 -4.31593060e-02 -4.57982153e-01 -2.44929552e-01
7.49756217e-01 7.34761298e-01 7.56880879e-01 -6.79017842e-01
3.85148376e-01 3.06795508e-01 -3.88020247e-01 -1.05042446e+00
-1.11340880e+00 -3.94254923e-01 -3.86015743e-01 -1.23163454e-01
7.85164535e-01 -7.80926824e-01 -3.09121519e-01 6.38302192e-02
-1.70733130e+00 4.71610855e-03 -7.83310235e-01 3.90061140e-02
-6.26812816e-01 4.92047518e-01 -7.24980116e-01 -4.22365576e-01
-6.21560037e-01 -7.02695072e-01 1.23050940e+00 5.82121134e-01
-1.02745235e+00 -1.04451489e+00 -2.27116957e-01 2.10050363e-02
2.29698732e-01 3.67414892e-01 8.24539661e-01 -1.39694476e+00
-3.79618853e-01 -3.27919014e-02 1.05188459e-01 1.56602979e-01
4.72532153e-01 3.07594031e-01 -3.24713707e-01 8.17478225e-02
-3.16375822e-01 -5.39987870e-02 1.06747127e+00 3.67657870e-01
9.52508152e-01 -9.32848096e-01 -4.10605282e-01 2.30411142e-02
8.94231498e-01 9.23991278e-02 6.13513708e-01 1.89809054e-01
5.86138546e-01 7.03918934e-01 5.46468735e-01 1.94604158e-01
4.62335706e-01 5.22064388e-01 -6.01051152e-02 2.39177585e-01
-3.77071291e-01 -2.99662173e-01 8.33147243e-02 1.19207954e+00
-2.22660154e-01 -7.73606539e-01 -6.93320632e-01 9.90834355e-01
-2.02434635e+00 -1.37002325e+00 -4.42637295e-01 1.59432375e+00
1.28140116e+00 5.19062020e-02 5.58407046e-02 -2.19661936e-01
9.85842228e-01 4.40899760e-01 -1.50348783e-01 -5.33260643e-01
-4.65542614e-01 5.63459955e-02 4.72750515e-01 4.76990849e-01
-9.65797722e-01 1.12505078e+00 6.87892675e+00 9.37478483e-01
-6.17991924e-01 -1.85470864e-01 6.02537334e-01 7.92684313e-03
-5.36871314e-01 -5.43394089e-02 -1.06462240e+00 2.03052148e-01
8.30795944e-01 -8.81169498e-01 -1.58380479e-01 5.11482120e-01
6.84647784e-02 -3.70101660e-01 -1.33844590e+00 4.07100946e-01
4.20469522e-01 -1.90557754e+00 5.34015954e-01 -3.69702429e-01
8.55010808e-01 -4.12024796e-01 -5.78839779e-01 2.05809996e-01
5.55708408e-01 -8.91183615e-01 7.95451164e-01 5.31799257e-01
4.05244708e-01 -5.83582342e-01 6.06941342e-01 3.98681283e-01
-8.54503572e-01 3.01370740e-01 -2.86832005e-01 1.36900529e-01
6.12198710e-01 5.28475225e-01 -9.49148595e-01 1.02526104e+00
2.12967053e-01 8.11957121e-01 -6.12172484e-01 1.19589496e+00
-4.99324441e-01 6.85657263e-01 -3.08612078e-01 -3.69578004e-01
2.56901383e-01 3.23273949e-02 9.72437680e-01 1.60029626e+00
4.44462508e-01 2.88877547e-01 -4.68068272e-02 8.00170243e-01
-4.59024906e-01 4.05652598e-02 -5.01329005e-01 -1.92363933e-01
6.98144138e-01 1.10422397e+00 -8.80886137e-01 -1.01642787e+00
-3.32359001e-02 9.21346664e-01 1.07528448e-01 2.89536804e-01
-7.37462699e-01 -5.61283588e-01 2.54328668e-01 8.89601111e-02
7.50192329e-02 -4.59621958e-02 -1.96037516e-01 -1.29716325e+00
-1.14651844e-01 -8.49574983e-01 2.92330772e-01 -9.67384815e-01
-1.20695806e+00 6.68541253e-01 4.57662642e-01 -9.34560597e-01
-5.67413628e-01 1.16164098e-02 -9.57793295e-01 5.92248201e-01
-9.97645020e-01 -8.62014890e-01 -1.55959025e-01 1.63836807e-01
1.08541560e+00 -2.71994919e-02 6.98775113e-01 9.61930752e-02
-2.80776739e-01 2.56163597e-01 -1.67199641e-01 3.12012583e-01
8.88661742e-01 -1.47325516e+00 7.83638120e-01 9.56202388e-01
3.13228309e-01 8.45008194e-01 1.14691138e+00 -9.64165092e-01
-6.68360174e-01 -1.17548096e+00 1.49589837e+00 -3.91141683e-01
6.52945817e-01 8.57112259e-02 -9.97444272e-01 1.07423425e+00
1.05987370e+00 -7.65146852e-01 5.74354112e-01 1.26670199e-02
-1.34396583e-01 5.18102169e-01 -7.83831656e-01 8.21344197e-01
1.12464738e+00 -2.03842998e-01 -1.41985929e+00 6.44257724e-01
1.04568744e+00 -7.28166580e-01 -6.12511873e-01 2.86944479e-01
7.81819299e-02 -8.71658325e-01 6.57451689e-01 -1.05747247e+00
1.00501859e+00 -2.79605508e-01 2.12485626e-01 -1.67164481e+00
7.05780461e-02 -9.21862662e-01 -4.93996114e-01 1.74666071e+00
6.34688199e-01 -3.32084715e-01 4.33675379e-01 5.79590201e-01
-6.52310073e-01 -5.35005212e-01 -6.92403555e-01 -5.85263312e-01
2.62259454e-01 1.86233938e-01 6.68209791e-01 9.67483699e-01
3.43933135e-01 6.37434006e-01 5.89606799e-02 -3.57797056e-01
1.80716887e-01 1.41885458e-02 7.93820858e-01 -1.03528380e+00
-8.83155596e-03 -7.10041940e-01 -5.38113806e-03 -1.38278437e+00
2.79872715e-01 -9.42834139e-01 8.92040320e-03 -2.22634339e+00
1.55663937e-01 -4.42067012e-02 5.25443137e-01 3.63948852e-01
-3.98024678e-01 7.76803493e-02 3.04529130e-01 2.92638153e-01
-8.18503857e-01 3.45974296e-01 1.15777445e+00 -2.05215126e-01
-2.64526874e-01 8.71141534e-03 -1.19272149e+00 7.92003155e-01
8.75102222e-01 -3.96271110e-01 -2.42665075e-02 -4.00614172e-01
4.15491134e-01 2.87695616e-01 1.19335316e-01 -8.73183668e-01
4.74033684e-01 1.11416737e-02 -3.41944471e-02 -9.55139339e-01
3.82165536e-02 -1.56445518e-01 1.40970066e-01 2.32103970e-02
-1.01912165e+00 1.96288586e-01 1.15754023e-01 1.90534592e-01
-4.95174587e-01 -6.80737078e-01 3.36877197e-01 -3.66798192e-01
-3.61820430e-01 -2.61979699e-01 -6.08391762e-01 5.28779387e-01
6.83680236e-01 -1.24358162e-01 -9.38985527e-01 -8.45033824e-01
-6.94610298e-01 3.53584945e-01 3.36400658e-01 3.90064687e-01
2.84067541e-01 -1.18548512e+00 -1.09393132e+00 -6.72399759e-01
-3.33455168e-02 3.16672236e-01 -6.74524680e-02 7.51505375e-01
-6.95443630e-01 3.68567884e-01 9.18925926e-02 -2.25591034e-01
-1.35679090e+00 1.90905944e-01 1.40792370e-01 -3.69205624e-01
-6.88885868e-01 7.13439584e-01 -1.91770010e-02 -1.28650859e-01
1.88055456e-01 -5.78821957e-01 -4.86827493e-01 6.13120139e-01
6.85633540e-01 4.14643794e-01 -1.77972727e-02 -7.50635684e-01
-4.50406633e-02 1.08838886e-01 -2.10825816e-01 -1.69854149e-01
1.05884361e+00 -1.73440799e-01 -3.92852187e-01 6.18586540e-01
9.05264080e-01 6.04778171e-01 -5.72760999e-01 8.02025571e-02
3.33404094e-01 -5.66143682e-03 -6.32247865e-01 -7.16821849e-01
-6.43697143e-01 4.64394897e-01 -6.57178402e-01 9.45756555e-01
9.42501545e-01 4.72278357e-01 8.95342708e-01 6.32334411e-01
2.49027722e-02 -1.03435194e+00 -2.94677794e-01 8.40954721e-01
1.17709899e+00 -5.52423179e-01 3.94953817e-01 -6.21436775e-01
-9.31188762e-01 1.20593643e+00 4.06694710e-01 -3.78810078e-01
4.73874062e-02 -1.84781477e-02 -2.78628230e-01 -3.44315588e-01
-8.13481152e-01 -1.31851539e-01 3.41027260e-01 2.30326504e-01
7.99728513e-01 -2.95413602e-02 -9.60518539e-01 7.16961145e-01
-7.93527246e-01 -2.09438369e-01 1.19486094e+00 8.77905071e-01
-6.73443377e-01 -8.43863249e-01 -1.81605846e-01 7.15012968e-01
-8.39513123e-01 -3.45507562e-01 -9.60872591e-01 7.97703385e-01
-2.76447177e-01 1.07930994e+00 2.48958588e-01 1.25288889e-01
4.45325404e-01 1.96151152e-01 4.42089587e-01 -1.05021024e+00
-1.21119225e+00 6.19711019e-02 9.40472543e-01 -1.40636906e-01
-1.02466559e+00 -6.46617115e-01 -1.46056557e+00 -2.86706895e-01
-4.55998152e-01 6.13803506e-01 2.65528798e-01 9.33875561e-01
4.62467790e-01 9.26447690e-01 -3.76363546e-02 -8.08296084e-01
-3.19735616e-01 -1.16431844e+00 -3.21143448e-01 3.67632121e-01
2.06187919e-01 -7.43632913e-02 -4.01767403e-01 4.42708075e-01] | [12.34532356262207, 9.366759300231934] |
7687646f-48c9-4710-9cc9-1f1d9df31487 | speechocean762-an-open-source-non-native | 2104.01378 | null | https://arxiv.org/abs/2104.01378v2 | https://arxiv.org/pdf/2104.01378v2.pdf | speechocean762: An Open-Source Non-native English Speech Corpus For Pronunciation Assessment | This paper introduces a new open-source speech corpus named "speechocean762" designed for pronunciation assessment use, consisting of 5000 English utterances from 250 non-native speakers, where half of the speakers are children. Five experts annotated each of the utterances at sentence-level, word-level and phoneme-level. A baseline system is released in open source to illustrate the phoneme-level pronunciation assessment workflow on this corpus. This corpus is allowed to be used freely for commercial and non-commercial purposes. It is available for free download from OpenSLR, and the corresponding baseline system is published in the Kaldi speech recognition toolkit. | ['Yujun Wang', 'Daniel Povey', 'Ke Li', 'YuKai Huang', 'Qiong Song', 'Zhiyong Yan', 'Yongqing Wang', 'Zhiwen Zhang', 'Junbo Zhang'] | 2021-04-03 | null | null | null | null | ['phone-level-pronunciation-scoring'] | ['speech'] | [-1.71152025e-01 2.90952474e-01 2.80694783e-01 -5.68894684e-01
-1.23542428e+00 -7.48752296e-01 1.11083075e-01 8.16899464e-02
-5.09600222e-01 4.94868279e-01 5.94171464e-01 -6.67443693e-01
4.98583555e-01 -3.54378283e-01 -2.82932609e-01 -2.92447567e-01
1.32207006e-01 5.81830263e-01 -2.36750040e-02 -3.31920683e-01
-1.50475711e-01 8.23857542e-03 -1.38733661e+00 4.43163335e-01
9.29895222e-01 7.54425883e-01 5.84999561e-01 1.05008781e+00
-1.57494426e-01 7.83405960e-01 -1.00490427e+00 -7.53776789e-01
-1.74111292e-01 -1.00356713e-01 -9.43736315e-01 -2.45072946e-01
4.16786999e-01 -5.08110940e-01 -1.12941995e-01 1.15348089e+00
8.42207015e-01 2.14194998e-01 2.30651692e-01 -5.77606738e-01
-8.43505204e-01 1.21371889e+00 6.87644362e-01 1.02209017e-01
6.38170719e-01 4.85583488e-03 8.76250744e-01 -1.34927464e+00
1.92113444e-01 1.33674467e+00 2.54077554e-01 8.23469400e-01
-8.94539654e-01 -8.62755954e-01 -1.09861847e-02 7.44028762e-02
-1.35816181e+00 -9.85033870e-01 3.95207047e-01 -3.97759736e-01
1.48051798e+00 3.83062333e-01 5.18757522e-01 1.58845985e+00
-4.19902831e-01 1.06367147e+00 1.10792267e+00 -6.96838856e-01
1.63031712e-01 1.15485162e-01 4.06907558e-01 1.81065470e-01
-5.26263535e-01 1.33228272e-01 -8.33233893e-01 2.22305223e-01
4.35306013e-01 -7.87081361e-01 -2.75636017e-01 4.23067302e-01
-1.19852436e+00 5.98891735e-01 -2.43411422e-01 5.01262426e-01
-2.34323978e-01 -2.76091367e-01 5.86812973e-01 8.08672249e-01
8.02072227e-01 1.69428527e-01 -8.15805197e-01 -1.02845073e+00
-5.00894487e-01 7.33164921e-02 8.87652159e-01 1.13006210e+00
2.41393641e-01 4.74955022e-01 -2.37250309e-02 1.65919745e+00
4.25615072e-01 8.11495662e-01 1.05308783e+00 -7.71574676e-01
5.79955816e-01 1.23610109e-01 -1.58642292e-01 -1.02131031e-01
2.83493221e-01 -8.76550972e-02 -3.87434155e-01 9.62875504e-03
1.93539411e-01 -1.68362021e-01 -9.82916772e-01 1.58002341e+00
1.99067637e-01 -5.36585636e-02 5.57155073e-01 4.56990868e-01
1.51957595e+00 9.39334691e-01 2.71195412e-01 -1.11223161e-01
1.30325139e+00 -1.09948206e+00 -1.16436303e+00 -4.10568379e-02
4.90999997e-01 -1.13547456e+00 1.52731669e+00 7.96674609e-01
-1.52825797e+00 -6.44888818e-01 -8.33781242e-01 -1.02181710e-01
-7.74554551e-01 2.08582148e-01 2.07386360e-01 1.27460611e+00
-1.32575619e+00 1.01904243e-01 -6.82734847e-01 -1.34864151e-01
-1.71219986e-02 1.03732720e-01 -4.41055626e-01 1.81987733e-01
-1.58980632e+00 9.28072691e-01 4.40910965e-01 -3.06896597e-01
-1.05188096e+00 -5.58347821e-01 -1.20354736e+00 -9.56359059e-02
-1.01768732e-01 4.16085094e-01 2.07489014e+00 -7.70112813e-01
-2.23553348e+00 1.10033131e+00 -1.33766055e-01 -2.44824082e-01
2.64761180e-01 -4.86561567e-01 -9.25302804e-01 -1.16263956e-01
-4.61940318e-02 4.73324746e-01 4.80919629e-01 -6.09121859e-01
-6.86169088e-01 -1.94987476e-01 -1.81618527e-01 4.74753410e-01
-4.71785367e-01 7.11164713e-01 -2.90198207e-01 -1.15651691e+00
-1.46461427e-01 -6.71786606e-01 1.22046627e-01 -8.35587859e-01
-3.15543383e-01 -6.51271164e-01 3.77808124e-01 -1.21662629e+00
1.28645468e+00 -2.34795380e+00 -2.52968073e-01 -1.82682946e-01
-4.08865571e-01 7.27916896e-01 -1.70697980e-02 6.10486507e-01
-7.22338408e-02 8.65792409e-02 -1.53563812e-01 -6.60032928e-01
9.10650343e-02 3.20380740e-03 -9.18704197e-02 3.27715576e-02
6.54729605e-02 6.55932426e-01 -1.13476324e+00 6.14023507e-02
6.95144653e-01 6.44251943e-01 -9.13460702e-02 5.72596073e-01
4.47748713e-02 5.67144036e-01 4.99810465e-02 5.66649556e-01
6.48468673e-01 8.20057988e-01 -2.62588710e-01 6.96924329e-01
-4.34183657e-01 1.32213235e+00 -1.09216332e+00 1.60228682e+00
-9.19173777e-01 5.44546306e-01 3.79576683e-01 -4.15396631e-01
9.11038935e-01 9.59572077e-01 -2.94240743e-01 -3.38518977e-01
-4.30075079e-02 5.70123136e-01 1.88195780e-01 -1.99690014e-01
6.29416227e-01 1.65455639e-01 -1.36646211e-01 1.36456981e-01
4.69524503e-01 -6.47703469e-01 1.92143992e-01 7.49633387e-02
1.05800104e+00 -1.64275184e-01 4.34661835e-01 -4.30325687e-01
7.17152953e-01 -3.94282371e-01 1.27281383e-01 4.53800768e-01
-1.44039556e-01 5.86138725e-01 -1.84863418e-01 1.54271916e-01
-6.58171833e-01 -1.62213457e+00 -4.77738559e-01 1.66827285e+00
-7.91304290e-01 -4.70113784e-01 -1.13366830e+00 -2.66394556e-01
-4.10453409e-01 1.09759176e+00 -2.78203357e-02 3.12004715e-01
-3.86922181e-01 5.01432717e-01 8.99869442e-01 5.90951025e-01
2.55169570e-01 -1.62004697e+00 1.83099285e-01 3.85018617e-01
-2.39409968e-01 -1.02945435e+00 -6.05716288e-01 5.11138104e-02
-8.31184164e-02 -3.43208134e-01 -9.09175217e-01 -1.18041611e+00
1.63220853e-01 -8.82048532e-02 9.94975090e-01 -1.66973114e-01
2.68342674e-01 4.29594666e-01 -8.09303582e-01 -7.35351980e-01
-1.09747648e+00 1.46002829e-01 3.37110430e-01 -2.76056916e-01
6.88457370e-01 -2.73760378e-01 -1.88833978e-02 5.33551127e-02
-3.42418283e-01 -1.39308736e-01 2.26926878e-01 5.10377705e-01
4.87592041e-01 -3.21002394e-01 8.08389068e-01 -6.02313042e-01
8.18610668e-01 -3.33689779e-01 -4.56930250e-01 -4.20777388e-02
-3.65097560e-02 -5.10422468e-01 7.44327128e-01 -2.56223828e-01
-1.24054897e+00 4.48069014e-02 -1.45764267e+00 -5.82173169e-02
-8.57265353e-01 4.88186806e-01 -8.65112066e-01 3.95293057e-01
1.91678375e-01 1.47701427e-01 -3.36128175e-01 -9.00762439e-01
5.32642722e-01 1.62409866e+00 1.01623189e+00 -5.22535205e-01
1.75201803e-01 -6.76089585e-01 -1.19953132e+00 -1.35292625e+00
-4.83511865e-01 -6.38892353e-01 -5.85695624e-01 -2.15865359e-01
6.98031962e-01 -1.13332677e+00 -4.32313383e-01 1.08963609e+00
-1.22751880e+00 -4.69964564e-01 -2.91575700e-01 6.22401357e-01
-4.44006771e-01 8.19433294e-03 -7.84436584e-01 -8.06104660e-01
-4.40240949e-01 -1.35370946e+00 8.40923369e-01 3.77380173e-03
-4.50053781e-01 -9.44564521e-01 2.99267918e-01 4.59387720e-01
4.90805000e-01 -7.03946531e-01 4.32612658e-01 -1.18755949e+00
1.58456996e-01 -1.63095668e-01 4.21281874e-01 1.19753873e+00
1.89178541e-01 1.02706745e-01 -1.50774825e+00 -1.68370366e-01
-6.76354766e-02 -7.81827688e-01 3.01114261e-01 2.26232454e-01
1.10360789e+00 -4.93985862e-01 4.20657367e-01 1.89986140e-01
6.46186292e-01 3.82670790e-01 5.91199636e-01 -2.13147268e-01
6.33182824e-01 8.66550028e-01 2.87050217e-01 1.99910417e-01
5.92492878e-01 6.69781446e-01 -7.37356618e-02 3.07175875e-01
-2.94825643e-01 -5.82337201e-01 8.97296607e-01 2.04060411e+00
3.29132199e-01 -3.13059092e-01 -1.17362094e+00 8.44617844e-01
-1.25288093e+00 -7.09410965e-01 -2.59552300e-01 2.37423515e+00
1.25371432e+00 -1.25488639e-01 7.82834291e-02 3.01632255e-01
8.08958709e-01 6.07849248e-02 6.44011870e-02 -1.12011766e+00
-2.32729480e-01 6.00225747e-01 -1.35892063e-01 9.48822975e-01
-9.06753957e-01 1.35305011e+00 6.45316696e+00 1.20466387e+00
-9.00974989e-01 4.92978454e-01 6.00346267e-01 2.50227600e-01
-3.17128420e-01 -2.29913086e-01 -8.93503308e-01 6.61436141e-01
1.86260986e+00 -3.01697254e-01 4.37867135e-01 1.06560147e+00
1.09717861e-01 7.86672458e-02 -9.98102903e-01 7.82836914e-01
-1.17947394e-02 -8.07004809e-01 -2.40402818e-01 -3.20827544e-01
3.56054038e-01 3.90252203e-01 1.32473409e-01 6.72446311e-01
5.86259842e-01 -9.91267443e-01 8.89383793e-01 -3.71778086e-02
1.28380311e+00 -8.92602503e-01 5.14896452e-01 2.81753868e-01
-1.18189275e+00 4.27774698e-01 -4.92636040e-02 -1.56420305e-01
2.70755857e-01 2.50267863e-01 -9.88231480e-01 1.46863028e-01
7.15588331e-01 2.32646093e-01 -4.45102066e-01 7.08483100e-01
-6.23179674e-01 1.42282486e+00 -2.36139387e-01 -2.05639079e-01
6.18764162e-02 4.93688770e-02 5.00930488e-01 1.64397013e+00
3.02084506e-01 1.69916213e-01 6.19765483e-02 1.69637427e-01
-3.99117172e-02 7.03672647e-01 -5.06293237e-01 -1.40979648e-01
9.53471541e-01 1.04018509e+00 -2.72523344e-01 -3.78544152e-01
-5.89013875e-01 1.06604075e+00 4.89072323e-01 6.06723838e-02
-9.18357447e-02 -6.35895550e-01 7.82847166e-01 -4.30384167e-02
6.15062416e-02 -1.79334581e-01 -2.45509688e-02 -8.73475850e-01
-1.51832670e-01 -1.12917864e+00 -2.10277401e-02 -6.90950096e-01
-1.33138108e+00 9.94814515e-01 2.96081183e-04 -8.48866165e-01
-8.30289245e-01 -9.66451943e-01 -7.02283025e-01 1.28440309e+00
-9.02982533e-01 -9.41127598e-01 9.76063162e-02 4.78199512e-01
1.20034611e+00 -5.37106514e-01 1.38726556e+00 4.19005096e-01
-4.44663823e-01 8.42225730e-01 1.07310563e-01 4.71839696e-01
5.20029902e-01 -1.61511838e+00 1.06774831e+00 7.97228098e-01
2.13550150e-01 6.27614141e-01 5.99313736e-01 -6.49219811e-01
-7.09625363e-01 -1.15508032e+00 1.53927016e+00 -5.30468524e-01
9.25779104e-01 -8.11986387e-01 -1.06922758e+00 7.78817117e-01
6.71694517e-01 -2.40536019e-01 9.83052015e-01 1.42077468e-02
8.86551440e-02 1.06113888e-01 -1.05946612e+00 4.44811463e-01
8.30950379e-01 -8.75765681e-01 -8.98823440e-01 3.03861409e-01
1.17418540e+00 -6.90744042e-01 -1.07899523e+00 2.29473293e-01
2.91156292e-01 -6.70796096e-01 4.68967110e-01 -1.90699011e-01
6.34277612e-02 1.44531548e-01 -4.07960862e-01 -2.03169513e+00
7.35152662e-02 -1.06118178e+00 3.90381180e-02 1.85794175e+00
6.33349001e-01 -8.22121143e-01 -5.17105358e-03 2.13810593e-01
-8.44369769e-01 -3.67813796e-01 -1.38681400e+00 -8.39812219e-01
6.41142547e-01 -1.10439587e+00 8.03542495e-01 6.99068308e-01
4.37495112e-01 3.72853160e-01 5.22743352e-02 1.20197527e-01
1.01655096e-01 -8.45944524e-01 4.22041953e-01 -7.22033262e-01
-3.35600823e-01 -5.03178835e-01 -2.81916916e-01 -1.09074533e+00
4.39639807e-01 -8.24926794e-01 3.63625139e-01 -1.15067482e+00
-5.57063282e-01 -4.02620584e-01 -1.22762933e-01 3.79698694e-01
-2.07817614e-01 1.44919232e-01 1.30613804e-01 -3.01802546e-01
-2.24749908e-01 5.85788190e-01 6.37956262e-01 6.40806481e-02
-1.16470285e-01 3.49920332e-01 -2.80102640e-01 7.60187209e-01
1.06246698e+00 -1.72431782e-01 -3.87651116e-01 -5.83199203e-01
-4.46038783e-01 7.28074014e-02 -2.63427436e-01 -1.18568552e+00
-9.87307653e-02 1.17436722e-01 -1.60356238e-01 -8.90874684e-01
5.00681162e-01 -1.71561122e-01 -1.66640729e-01 5.50503694e-02
-5.19111872e-01 2.15241447e-01 3.49400699e-01 -4.53448325e-01
-4.59233314e-01 -5.48309565e-01 7.33856976e-01 -1.48167446e-01
-7.26291478e-01 8.33291560e-02 -8.14842463e-01 2.08185256e-01
7.71722436e-01 2.18620542e-02 -3.06469202e-01 -6.77976072e-01
-8.05567086e-01 1.02712162e-01 1.37368217e-01 7.29461789e-01
6.75692797e-01 -1.40729237e+00 -1.06421673e+00 6.35563433e-01
3.69783401e-01 -2.13420153e-01 1.92805246e-01 1.74476624e-01
-5.08815169e-01 6.64479375e-01 6.16882965e-02 -1.67817578e-01
-1.22782874e+00 2.28626039e-02 1.59909084e-01 3.58697087e-01
-3.31196547e-01 1.36865842e+00 -8.06969684e-03 -1.02795732e+00
6.39640510e-01 -3.99129212e-01 -3.91468376e-01 -2.14618832e-01
1.04998410e+00 5.74861825e-01 4.52037752e-01 -1.14696920e+00
-3.27327222e-01 -1.81257010e-01 -1.54635668e-01 -7.94396937e-01
7.79001713e-01 -3.36050987e-01 8.65560621e-02 1.03704965e+00
9.99077976e-01 6.04718268e-01 -8.05067480e-01 5.91733726e-03
-2.43047744e-01 -1.89487204e-01 6.83558956e-02 -1.07386768e+00
-5.54039001e-01 1.07935202e+00 5.03695786e-01 3.17943186e-01
7.53501832e-01 2.78464556e-02 9.44669127e-01 3.13129276e-01
3.09106350e-01 -1.36327553e+00 -2.85699874e-01 1.20184302e+00
1.08600473e+00 -1.16951382e+00 -8.24899495e-01 -4.04201299e-01
-9.81917858e-01 7.89230645e-01 6.61150634e-01 1.14908487e-01
7.18987525e-01 4.99704629e-01 7.06363857e-01 4.02850151e-01
-7.29428828e-01 -2.71690428e-01 2.07614422e-01 1.08634496e+00
1.13943446e+00 5.84498048e-01 -2.41782889e-01 8.53545487e-01
-1.14670002e+00 -5.81608355e-01 5.82606375e-01 4.74299192e-01
-3.87907833e-01 -1.26983941e+00 -4.03210074e-01 6.95359111e-02
-8.79801393e-01 -7.60783851e-01 -4.95252877e-01 2.83450454e-01
2.78925989e-02 1.44342542e+00 2.35265255e-01 -2.09433168e-01
3.91747445e-01 3.39524895e-01 1.49101853e-01 -1.18254161e+00
-6.85149252e-01 1.24819882e-01 5.85363805e-01 -2.13781178e-01
1.66109234e-01 -8.12698305e-01 -1.19779325e+00 -4.92050592e-03
-1.69001788e-01 2.96510488e-01 1.00841069e+00 7.12654829e-01
-1.13209963e-01 2.81842172e-01 5.49660861e-01 -7.30916023e-01
-5.95062077e-01 -1.58325541e+00 -4.56344008e-01 6.03095442e-02
3.64552170e-01 -2.72444934e-01 -3.81400675e-01 -3.97350155e-02] | [14.246417045593262, 6.908626079559326] |
2282a592-3a8d-4936-9be6-2182509f1205 | weakly-supervised-audio-visual-sound-source | 2104.02606 | null | https://arxiv.org/abs/2104.02606v1 | https://arxiv.org/pdf/2104.02606v1.pdf | Weakly-supervised Audio-visual Sound Source Detection and Separation | Learning how to localize and separate individual object sounds in the audio channel of the video is a difficult task. Current state-of-the-art methods predict audio masks from artificially mixed spectrograms, known as Mix-and-Separate framework. We propose an audio-visual co-segmentation, where the network learns both what individual objects look and sound like, from videos labeled with only object labels. Unlike other recent visually-guided audio source separation frameworks, our architecture can be learned in an end-to-end manner and requires no additional supervision or bounding box proposals. Specifically, we introduce weakly-supervised object segmentation in the context of sound separation. We also formulate spectrogram mask prediction using a set of learned mask bases, which combine using coefficients conditioned on the output of object segmentation , a design that facilitates separation. Extensive experiments on the MUSIC dataset show that our proposed approach outperforms state-of-the-art methods on visually guided sound source separation and sound denoising. | ['Leonid Sigal', 'Tanzila Rahman'] | 2021-03-25 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [ 6.28957272e-01 -1.21351242e-01 1.30633652e-01 -2.73098856e-01
-1.23523188e+00 -8.71774793e-01 2.46993944e-01 -7.04050735e-02
-2.07758486e-01 1.94689184e-01 3.36844504e-01 2.21945256e-01
-1.06020346e-01 -6.86102435e-02 -1.02771091e+00 -7.26899266e-01
-1.00840971e-01 7.01578557e-02 2.90097058e-01 3.29504579e-01
1.91432565e-01 -3.21980007e-02 -1.78819954e+00 7.44795799e-01
6.65409923e-01 1.23432481e+00 3.22523266e-01 1.17446601e+00
4.98841610e-03 8.36816430e-01 -4.50227529e-01 -7.37270247e-03
4.15234268e-01 -7.35860705e-01 -7.69714892e-01 4.09754097e-01
1.07600093e+00 -6.98464215e-02 -8.00859779e-02 1.11446953e+00
3.59065026e-01 1.56445086e-01 7.12110519e-01 -1.21518981e+00
-4.00381148e-01 9.00527477e-01 -5.36034405e-01 1.76299691e-01
1.67090699e-01 1.90854162e-01 1.30752993e+00 -1.03552604e+00
1.37168407e-01 1.18200624e+00 6.50535166e-01 3.74640703e-01
-1.66994190e+00 -6.58137500e-01 5.57280660e-01 -1.63612962e-02
-1.17084444e+00 -9.38488662e-01 1.08460045e+00 -7.76324034e-01
5.02749562e-01 4.21759695e-01 6.26361012e-01 7.50489593e-01
-4.52873617e-01 1.06692791e+00 8.01256537e-01 -5.22884548e-01
2.96786129e-01 -3.73129025e-02 -7.47626694e-03 6.42837048e-01
-2.09233582e-01 -2.04495072e-01 -9.13823843e-01 -1.01953633e-01
8.05290937e-01 -2.46865124e-01 -6.01426363e-01 -5.34138501e-01
-1.15567994e+00 3.87558699e-01 2.45646015e-01 1.28262654e-01
-1.82976395e-01 6.34195328e-01 2.18084082e-01 1.31568789e-01
5.70874214e-01 1.78354487e-01 -3.52223665e-01 1.25979811e-01
-1.61518180e+00 1.49526268e-01 5.97519875e-01 7.18259275e-01
5.72649598e-01 3.92638415e-01 -2.59996742e-01 9.28090036e-01
5.60760677e-01 2.77522534e-01 2.34428942e-01 -1.23080432e+00
4.31096852e-01 -5.43450415e-02 8.18364769e-02 -7.03471065e-01
-1.63476288e-01 -6.53093338e-01 -6.04483426e-01 2.68194079e-01
4.79279399e-01 -1.47743091e-01 -1.25290501e+00 1.75587356e+00
2.33792394e-01 1.04350626e+00 -2.27353632e-01 1.34903145e+00
1.00050366e+00 6.49679303e-01 4.24154196e-03 -1.88945666e-01
1.10356617e+00 -1.29254889e+00 -6.43492162e-01 -5.00692606e-01
-1.63906813e-01 -8.20769608e-01 1.01166749e+00 7.70516753e-01
-1.37387240e+00 -1.08211231e+00 -9.08507288e-01 4.50759418e-02
2.61601865e-01 5.69723427e-01 2.62508720e-01 5.84818244e-01
-9.37774956e-01 4.98888701e-01 -9.80816543e-01 1.58840030e-01
5.80050111e-01 2.49832138e-01 -3.56898159e-02 4.41264093e-01
-4.76315200e-01 2.42072996e-02 8.91172886e-02 1.72769621e-01
-1.72902215e+00 -8.65884423e-01 -8.63071322e-01 3.37239385e-01
6.93828166e-01 -6.65743470e-01 1.43665373e+00 -1.38185692e+00
-1.56442392e+00 7.45078266e-01 -3.53072405e-01 -5.52148402e-01
2.47136459e-01 -6.26340032e-01 -1.32172868e-01 5.32272339e-01
1.14727721e-01 7.71380305e-01 1.72191215e+00 -1.85739708e+00
-8.29010665e-01 -1.34495229e-01 -1.69206366e-01 1.90307796e-01
-1.43590629e-01 1.10625789e-01 -7.00590014e-01 -1.06108594e+00
2.56895214e-01 -6.36882126e-01 -6.30921349e-02 -3.90689261e-03
-5.35295248e-01 2.26780683e-01 5.83521783e-01 -9.06814337e-01
1.08696163e+00 -2.43868136e+00 6.31154418e-01 1.40328854e-01
2.53718406e-01 -8.04245025e-02 -4.25554574e-01 -1.06587104e-01
-1.10412836e-01 -1.87088437e-02 -6.22077525e-01 -8.27616096e-01
1.47580042e-01 -3.92487228e-01 -5.79040468e-01 4.20496941e-01
2.35855043e-01 2.49999806e-01 -9.38494027e-01 -4.58309770e-01
9.02963430e-02 4.69371706e-01 -8.29007626e-01 3.78882825e-01
-3.09474587e-01 6.09276056e-01 -3.94516028e-02 6.66529715e-01
5.71277916e-01 3.98632921e-02 5.22844270e-02 -2.75371909e-01
1.05166957e-01 2.54768491e-01 -1.59298277e+00 2.07532358e+00
-3.80348295e-01 7.51015246e-01 9.70259547e-01 -1.03319776e+00
7.25983083e-01 6.28762364e-01 4.94957864e-01 5.11375815e-02
3.11131822e-03 1.92028239e-01 -1.43520743e-01 -5.23336411e-01
2.16542348e-01 -2.25389406e-01 1.39757410e-01 3.17250729e-01
5.94286680e-01 -3.02108377e-01 1.79422632e-01 3.27027291e-02
7.52350926e-01 3.16696763e-01 -1.76214546e-01 -5.86075708e-03
6.28826916e-01 -4.71523404e-01 5.72706461e-01 8.76872063e-01
-2.21638635e-01 1.03681862e+00 2.29114309e-01 1.84901088e-01
-5.16713321e-01 -1.30298352e+00 1.36243850e-01 1.61873496e+00
9.05821919e-02 -4.75464731e-01 -1.11849201e+00 -6.92992389e-01
-2.89084390e-02 3.04652154e-01 -2.41791040e-01 2.10228682e-01
-5.92335284e-01 -2.35545620e-01 4.84566838e-01 4.92340803e-01
4.14759256e-02 -1.06887949e+00 -4.37657624e-01 1.82726160e-01
-3.56230497e-01 -1.12583911e+00 -7.01332748e-01 4.16521609e-01
-6.19377136e-01 -9.77133214e-01 -7.41695940e-01 -1.19296169e+00
3.92440945e-01 3.68552327e-01 1.07975733e+00 -2.01035574e-01
-4.18154240e-01 7.97251582e-01 -6.22420125e-02 -3.60701203e-01
-2.98514694e-01 -2.60581702e-01 1.05067417e-01 7.55478144e-01
-1.45676419e-01 -9.13483143e-01 -7.23501623e-01 2.03140303e-01
-9.04886603e-01 -2.18524989e-02 3.31738591e-01 4.85702902e-01
8.77253592e-01 1.08027548e-01 6.60682619e-01 -4.26701814e-01
4.23399776e-01 -1.88307300e-01 -4.10450339e-01 5.33389375e-02
2.23124519e-01 -2.36704424e-01 6.39532208e-01 -5.82724571e-01
-9.83469069e-01 5.99469423e-01 1.34505227e-01 -8.34913194e-01
-6.89435124e-01 5.35229258e-02 -2.85551906e-01 4.20805998e-02
4.82729971e-01 8.93039927e-02 -3.22933435e-01 -9.93072510e-01
6.64662480e-01 7.41981566e-01 1.02403021e+00 -4.20886755e-01
5.89396656e-01 7.49333978e-01 -2.61039644e-01 -9.33207691e-01
-9.78058934e-01 -8.37111235e-01 -7.04759121e-01 -3.44151050e-01
1.06903410e+00 -1.21750331e+00 -5.81625760e-01 3.76383424e-01
-1.07368922e+00 -6.04144096e-01 -2.72955328e-01 5.26166856e-01
-8.00800622e-01 2.92867988e-01 -4.42274660e-01 -1.12827122e+00
-7.33761415e-02 -1.11906397e+00 1.34974468e+00 8.48894045e-02
-1.51086271e-01 -5.37366986e-01 1.21667914e-01 5.38942397e-01
-1.44801930e-01 -1.59376279e-01 5.25104165e-01 -3.57168853e-01
-6.77193642e-01 3.08551133e-01 -4.46839370e-02 6.46963775e-01
1.90932915e-01 -6.00352138e-02 -1.52442396e+00 -2.09151670e-01
1.25645190e-01 -4.32942420e-01 1.55222631e+00 8.37871432e-01
1.28570163e+00 -2.97566295e-01 -1.36064276e-01 8.28694761e-01
1.16819453e+00 -2.37403624e-02 8.59413743e-02 -1.52609855e-01
9.77629662e-01 6.67587459e-01 4.21467185e-01 3.11839998e-01
-9.42500494e-03 5.77134728e-01 5.83345413e-01 -1.99797601e-01
-4.65874970e-01 -2.74773449e-01 5.42760491e-01 6.01515234e-01
1.04209661e-01 -1.50277436e-01 -5.73326170e-01 8.30069602e-01
-1.87254632e+00 -7.78951406e-01 5.80031909e-02 1.84681845e+00
9.81586337e-01 1.11643136e-01 5.16134143e-01 5.79966962e-01
7.70849228e-01 4.01216835e-01 -3.05406183e-01 1.22870080e-01
5.74139170e-02 3.82253826e-01 1.85087323e-02 7.75064707e-01
-1.71002293e+00 8.96622181e-01 5.98292017e+00 9.15772200e-01
-1.05768001e+00 1.80353388e-01 4.44893748e-01 -5.10325313e-01
-3.90007868e-02 -4.33092117e-02 -4.85998720e-01 3.03245276e-01
5.92481852e-01 5.27423441e-01 6.27336383e-01 5.88387370e-01
3.23034227e-01 -1.00709401e-01 -1.33552861e+00 1.24714518e+00
3.43522996e-01 -1.16599333e+00 -1.43032953e-01 -5.13049006e-01
6.20576680e-01 -2.89747328e-01 1.70417041e-01 -1.07819401e-01
-1.43221812e-02 -8.94676626e-01 1.51425278e+00 5.52870810e-01
5.10934114e-01 -6.34312451e-01 -1.42827062e-02 2.39933565e-01
-1.55161059e+00 -3.66084307e-01 3.53952050e-02 7.84339234e-02
3.62255871e-01 3.60355109e-01 -6.23625815e-01 3.10782969e-01
1.03279030e+00 7.56544948e-01 -4.21641111e-01 1.39268315e+00
-3.36125940e-01 1.27658093e+00 -2.35589251e-01 6.03055894e-01
1.01756886e-01 -3.23934667e-02 9.61344779e-01 1.63802862e+00
6.34111837e-02 -2.69911349e-01 4.55931127e-01 9.89044666e-01
-6.26929477e-02 1.91363469e-02 -2.23912206e-02 -7.59948343e-02
2.14443311e-01 1.37146628e+00 -9.09667492e-01 -1.77279249e-01
-3.08444917e-01 8.99856806e-01 -6.75020041e-03 5.84293306e-01
-7.96849668e-01 -2.82910377e-01 6.33001089e-01 1.42426744e-01
8.12550962e-01 -1.02636985e-01 -4.14242059e-01 -8.97827327e-01
1.64256394e-01 -9.40990269e-01 2.77593046e-01 -8.10826659e-01
-1.20626330e+00 3.61719728e-01 -1.50176674e-01 -1.38523972e+00
7.11618140e-02 -5.85476339e-01 -6.57096803e-01 4.73020881e-01
-1.40663505e+00 -1.15676439e+00 -2.06692532e-01 5.35623789e-01
8.20473433e-01 -1.69920042e-01 4.45790827e-01 3.25497001e-01
-4.27169532e-01 3.09321880e-01 1.77484471e-02 3.64392102e-01
9.16885316e-01 -1.27742994e+00 1.77760363e-01 8.95776629e-01
1.02681100e+00 1.77180588e-01 7.58996129e-01 -5.28923035e-01
-9.51690376e-01 -1.12973022e+00 2.17046842e-01 -1.38962701e-01
5.19831419e-01 -7.39417851e-01 -7.97916234e-01 5.21955729e-01
1.97853744e-01 9.59354788e-02 7.48065829e-01 -1.09311759e-01
-3.16915452e-01 -3.75225276e-01 -5.77426553e-01 2.67577022e-01
9.94986475e-01 -7.29674816e-01 -6.15747035e-01 1.48914784e-01
7.39341021e-01 -1.72581390e-01 -1.78849801e-01 2.48552069e-01
4.85288054e-01 -1.08591735e+00 1.30371714e+00 -6.70748830e-01
2.36292183e-01 -6.87610328e-01 -1.12422451e-01 -1.28354239e+00
-3.60673070e-01 -1.01635659e+00 -3.67043048e-01 1.45743680e+00
2.85107821e-01 2.48731107e-01 6.19720876e-01 -1.26590490e-01
-3.07075828e-01 -2.24498868e-01 -7.86365330e-01 -6.55735910e-01
-4.50879097e-01 -7.66181290e-01 1.48168400e-01 5.57028592e-01
-1.34987980e-01 3.20060998e-01 -5.55380285e-01 4.06762898e-01
8.45854938e-01 3.00189704e-01 7.78055727e-01 -1.21551752e+00
-8.86516750e-01 -5.05647242e-01 -2.73280382e-01 -1.29985690e+00
3.71108770e-01 -8.20520937e-01 7.06747949e-01 -1.45397401e+00
1.02377646e-01 1.27333344e-03 -4.85538602e-01 4.06104416e-01
-2.15494096e-01 6.59644842e-01 4.66815352e-01 1.92637220e-01
-8.26591790e-01 5.13471603e-01 9.78012621e-01 -3.88358444e-01
-5.18803298e-01 1.10208906e-01 -5.88335156e-01 1.22081435e+00
4.23079729e-01 -6.08805656e-01 -4.35961425e-01 -5.96125364e-01
-1.20796829e-01 1.95375264e-01 8.40458810e-01 -1.24497974e+00
2.31246099e-01 -3.75976861e-02 2.61630148e-01 -5.41177869e-01
6.89604163e-01 -7.42235303e-01 -2.05082938e-01 -1.09004796e-01
-7.06869245e-01 -8.51296782e-01 2.39199340e-01 9.58700120e-01
-5.00278890e-01 -2.19690219e-01 8.15965056e-01 -6.59579709e-02
-2.34489202e-01 5.85664287e-02 -5.09609044e-01 2.81728953e-01
6.18274927e-01 -7.63455331e-02 2.00312704e-01 -5.24396360e-01
-1.26430917e+00 -1.68011170e-02 -1.23458490e-01 2.90634423e-01
6.93652153e-01 -1.17724741e+00 -7.63325870e-01 1.62643954e-01
-2.86914080e-01 1.31307453e-01 4.13892776e-01 8.42698395e-01
-1.66334406e-01 1.63533427e-02 -1.17630223e-02 -9.11092937e-01
-1.58138692e+00 5.43574631e-01 4.49033916e-01 2.45822161e-01
-5.00638962e-01 1.39513278e+00 8.39235127e-01 1.30844533e-01
9.49042201e-01 -6.40755355e-01 -1.34475172e-01 1.62206024e-01
5.06838441e-01 4.09875393e-01 -1.91764668e-01 -6.74019575e-01
-2.52104193e-01 6.68173552e-01 3.50339800e-01 -5.84842741e-01
1.43870330e+00 -2.53103942e-01 -3.30473408e-02 7.17541397e-01
9.13880169e-01 3.43715459e-01 -1.72240865e+00 -2.28859127e-01
-2.52608269e-01 -5.30149579e-01 2.82668740e-01 -7.03621805e-01
-1.18473721e+00 1.16975510e+00 6.25770152e-01 3.55668992e-01
1.33597398e+00 9.23327580e-02 5.34160495e-01 1.25857681e-01
-2.86119908e-01 -1.07905686e+00 5.15310764e-01 1.96532264e-01
9.51989532e-01 -1.06261003e+00 -3.31450671e-01 -6.67199492e-01
-5.81493199e-01 9.31568265e-01 5.15769541e-01 -3.33781868e-01
7.40740895e-01 5.18717587e-01 2.31581181e-01 1.23163737e-01
-4.79096323e-01 -5.29606938e-01 8.45056415e-01 6.05938077e-01
4.45339739e-01 -1.29520312e-01 4.72575039e-01 1.13364911e+00
-1.03706248e-01 -2.21891299e-01 9.82705355e-02 8.17702472e-01
-7.77211428e-01 -6.97953701e-01 -7.66198754e-01 6.30956665e-02
-6.03820622e-01 -2.39696935e-01 -7.32565403e-01 -1.23227388e-02
4.73623186e-01 1.26803732e+00 1.06353074e-01 -2.09495217e-01
5.68914227e-02 3.52327943e-01 6.64372683e-01 -7.53822386e-01
-7.44592488e-01 1.10855293e+00 -1.47989467e-01 -4.27554131e-01
-7.19986856e-01 -5.50576329e-01 -1.34216559e+00 8.18236530e-01
-3.29849571e-01 1.96759149e-01 3.73136848e-01 7.20401824e-01
1.64135575e-01 8.40961874e-01 3.91881675e-01 -1.53252792e+00
-2.38641962e-01 -9.29975927e-01 -7.47277021e-01 4.31498379e-01
9.93486881e-01 -4.50381488e-01 -4.99168664e-01 9.15389717e-01] | [14.863975524902344, 4.959963321685791] |
aad06732-35e5-4d6d-8868-425edbe834c4 | deep-virtual-stereo-odometry-leveraging-deep | 1807.02570 | null | http://arxiv.org/abs/1807.02570v2 | http://arxiv.org/pdf/1807.02570v2.pdf | Deep Virtual Stereo Odometry: Leveraging Deep Depth Prediction for Monocular Direct Sparse Odometry | Monocular visual odometry approaches that purely rely on geometric cues are
prone to scale drift and require sufficient motion parallax in successive
frames for motion estimation and 3D reconstruction. In this paper, we propose
to leverage deep monocular depth prediction to overcome limitations of
geometry-based monocular visual odometry. To this end, we incorporate deep
depth predictions into Direct Sparse Odometry (DSO) as direct virtual stereo
measurements. For depth prediction, we design a novel deep network that refines
predicted depth from a single image in a two-stage process. We train our
network in a semi-supervised way on photoconsistency in stereo images and on
consistency with accurate sparse depth reconstructions from Stereo DSO. Our
deep predictions excel state-of-the-art approaches for monocular depth on the
KITTI benchmark. Moreover, our Deep Virtual Stereo Odometry clearly exceeds
previous monocular and deep learning based methods in accuracy. It even
achieves comparable performance to the state-of-the-art stereo methods, while
only relying on a single camera. | ['Jörg Stückler', 'Rui Wang', 'Nan Yang', 'Daniel Cremers'] | 2018-07-06 | deep-virtual-stereo-odometry-leveraging-deep-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Nan_Yang_Deep_Virtual_Stereo_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Nan_Yang_Deep_Virtual_Stereo_ECCV_2018_paper.pdf | eccv-2018-9 | ['monocular-visual-odometry'] | ['robots'] | [-1.60826176e-01 8.40326995e-02 -1.47078931e-01 -3.98407608e-01
-4.19716358e-01 -5.00124156e-01 9.09722328e-01 -4.30899024e-01
-3.30658764e-01 7.29607463e-01 4.35592115e-01 -9.20844451e-02
4.73248959e-01 -6.77429259e-01 -9.52069759e-01 -2.89341211e-01
4.78969991e-01 8.43158722e-01 4.10965413e-01 -7.38340542e-02
4.09644455e-01 5.17767370e-01 -1.51469648e+00 1.27808228e-01
5.05174637e-01 8.85523617e-01 4.67196554e-01 8.31499994e-01
7.94146731e-02 1.21279073e+00 1.73082933e-01 8.95879269e-02
6.21295512e-01 -1.91303819e-01 -8.11144769e-01 1.14087455e-01
1.13551021e+00 -1.00515556e+00 -9.92756784e-01 9.10338640e-01
6.07522666e-01 -1.97430328e-01 4.77811843e-01 -8.53866518e-01
-3.80706340e-01 -2.20285639e-01 -4.09541935e-01 1.06683532e-02
8.90271962e-01 3.66278201e-01 7.57914305e-01 -1.02766812e+00
1.14437950e+00 1.18728161e+00 7.80245841e-01 5.87513268e-01
-1.34146655e+00 -4.18651193e-01 1.80939846e-02 1.77650347e-01
-1.23199189e+00 -6.08217716e-01 7.50866354e-01 -8.25759828e-01
1.41952097e+00 -3.39576602e-01 8.23210537e-01 9.69086826e-01
3.80123615e-01 6.88639283e-01 1.08259809e+00 -1.73132509e-01
2.67699569e-01 -2.08685353e-01 -4.17133778e-01 7.84723520e-01
2.29542017e-01 6.34392202e-01 -8.80385399e-01 1.90306515e-01
1.11081088e+00 1.56554013e-01 -4.44204181e-01 -1.13990915e+00
-1.27399385e+00 7.61351049e-01 6.77976072e-01 -2.10540622e-01
-2.70389348e-01 4.98967171e-01 1.03585407e-01 2.84524500e-01
6.04416370e-01 3.07180047e-01 -4.63309735e-01 -2.23265991e-01
-1.04512215e+00 3.05062056e-01 6.31999373e-01 8.99846673e-01
1.26898885e+00 2.48204619e-01 5.12016296e-01 4.41706806e-01
3.83139431e-01 6.35291338e-01 5.94551206e-01 -1.66674626e+00
5.09890795e-01 4.72870380e-01 1.73483893e-01 -7.02423990e-01
-5.73609173e-01 -1.65016666e-01 -6.69128895e-01 6.52128994e-01
1.80067122e-01 9.85107757e-03 -9.87228692e-01 1.41484332e+00
4.03519034e-01 3.74024898e-01 9.37592238e-02 1.23757434e+00
8.55660260e-01 3.57175708e-01 -7.26919532e-01 2.99888730e-01
4.71679538e-01 -1.06142056e+00 -2.16701239e-01 -7.41155922e-01
7.21579790e-01 -7.22813189e-01 5.57933509e-01 4.43473071e-01
-1.08002174e+00 -4.95838106e-01 -1.25338697e+00 -5.56760609e-01
7.73827881e-02 -2.14695752e-01 5.44779003e-01 2.73988366e-01
-1.63749158e+00 7.56538808e-01 -9.59008098e-01 -4.12028015e-01
3.19712460e-02 4.10297960e-01 -6.49135470e-01 -3.86518657e-01
-8.19974065e-01 9.68953669e-01 1.96876213e-01 -2.51600981e-01
-1.14220369e+00 -6.80123508e-01 -1.59110093e+00 -4.41424042e-01
-1.34216428e-01 -1.37035406e+00 1.42665315e+00 -5.96293569e-01
-1.72623646e+00 1.26715839e+00 -5.04536688e-01 -6.55065894e-01
7.76739597e-01 -4.14053708e-01 2.82628715e-01 2.78311849e-01
2.67374963e-01 1.33227718e+00 4.30746764e-01 -1.52327609e+00
-5.97893655e-01 -4.85151887e-01 2.33698845e-01 6.45608127e-01
4.85820204e-01 -7.54344344e-01 -5.23917377e-01 4.64685597e-02
8.20821166e-01 -1.09479260e+00 -2.96338409e-01 4.01140213e-01
-3.55544657e-01 5.78169823e-01 6.20167077e-01 -5.01139343e-01
4.24009681e-01 -1.67758965e+00 2.89311796e-01 -3.79851639e-01
4.39879626e-01 -9.04962234e-03 1.40117452e-01 1.90656632e-01
1.75632387e-01 -6.80563986e-01 1.05050124e-01 -1.06765485e+00
-1.19363606e-01 3.65435570e-01 -2.87607908e-01 9.05581892e-01
-1.61845088e-01 8.64626110e-01 -1.01788855e+00 -2.63887018e-01
9.10265744e-01 5.54179192e-01 -8.71546030e-01 2.31906623e-01
-1.71981737e-01 9.22925234e-01 1.46132931e-01 6.59538090e-01
8.23728025e-01 -2.73174345e-01 9.29898173e-02 -5.25711849e-02
-4.35048491e-01 7.96339810e-01 -1.00448310e+00 2.37992692e+00
-7.47783124e-01 9.90926564e-01 -7.80996978e-02 -5.19856811e-01
8.54262233e-01 6.90383837e-02 5.07160068e-01 -9.59034860e-01
5.45137599e-02 5.30867159e-01 -5.90712011e-01 -1.00432597e-02
7.25971341e-01 -1.75946280e-01 3.48919213e-01 1.43753871e-01
1.63454250e-01 -9.06965256e-01 -2.63162047e-01 1.10504314e-01
1.05062723e+00 7.59315789e-01 5.03119528e-01 -2.29100659e-02
4.90778476e-01 8.51059407e-02 5.44368386e-01 4.54742074e-01
-9.34120566e-02 1.20807767e+00 7.26129636e-02 -7.95140266e-01
-1.48892713e+00 -1.17129612e+00 -9.52991396e-02 7.30825141e-02
6.20814323e-01 -9.79948193e-02 -2.20571339e-01 -1.50573358e-01
2.43824929e-01 1.79746643e-01 -4.51811522e-01 1.01940393e-01
-5.45637310e-01 -1.34084374e-01 2.33456731e-01 6.61720157e-01
6.06730044e-01 -6.29680336e-01 -6.52900994e-01 3.01941901e-01
-6.72060773e-02 -1.58128405e+00 -1.66938379e-01 1.98575944e-01
-1.18837118e+00 -9.52426136e-01 -9.82137680e-01 -6.57954633e-01
3.33584189e-01 7.92290270e-01 1.29107237e+00 -4.09764796e-01
1.61083922e-01 2.31866553e-01 9.61946025e-02 -3.53291854e-02
-2.25523472e-01 1.00375712e-01 2.45874345e-01 -4.99081939e-01
3.20542336e-01 -8.78598034e-01 -1.02397382e+00 3.31630319e-01
-4.84169513e-01 5.16384184e-01 1.32347837e-01 8.21974456e-01
5.75629473e-01 -7.60576308e-01 -4.33580369e-01 -4.68056053e-01
-3.48934948e-01 -4.35135394e-01 -1.07174218e+00 -6.59170926e-01
-6.11902058e-01 1.80849046e-01 4.54475880e-01 1.70047909e-01
-8.57391000e-01 6.17609799e-01 -1.41896367e-01 -7.76457369e-01
-3.18316594e-02 1.63876578e-01 3.17587227e-01 -4.61735368e-01
7.64597237e-01 3.19157243e-01 1.17443703e-01 -3.18340838e-01
7.42163658e-02 3.84227216e-01 6.69092953e-01 5.83918244e-02
8.51236761e-01 1.32390678e+00 4.08752382e-01 -6.07990086e-01
-8.62316847e-01 -7.80024350e-01 -1.10488522e+00 1.09192841e-02
9.05821383e-01 -1.85016704e+00 -6.22095644e-01 7.12115586e-01
-1.44184566e+00 -7.24652886e-01 3.61422002e-02 7.00483739e-01
-9.49789464e-01 7.14581549e-01 -7.98811555e-01 -4.53153282e-01
-6.42015189e-02 -1.13588369e+00 1.68724728e+00 -4.94242311e-02
-2.64065266e-01 -1.20494652e+00 6.70312941e-01 2.69033551e-01
-5.82038276e-02 2.49435917e-01 -5.46806827e-02 4.04084623e-01
-1.28463447e+00 2.07974063e-03 -3.92655134e-01 1.16339736e-01
-1.71946529e-02 -4.19534594e-01 -1.25966442e+00 -3.42410654e-01
-4.62445430e-02 -4.92602140e-01 1.02181661e+00 6.54556215e-01
4.89707619e-01 2.24942863e-01 -2.13501960e-01 1.44801486e+00
1.77701914e+00 -4.78478745e-02 6.69708550e-01 8.83449316e-01
1.23306358e+00 4.70504105e-01 5.02551377e-01 5.95557690e-01
8.78900886e-01 9.65170860e-01 9.14961696e-01 -1.81299642e-01
-1.23889320e-01 -5.57447255e-01 3.45604151e-01 6.18565440e-01
9.84871462e-02 6.93536922e-02 -1.01687670e+00 7.43579149e-01
-1.87683785e+00 -7.92283177e-01 -3.03848624e-01 2.11991453e+00
5.11620522e-01 4.67347801e-02 -2.61877924e-01 5.56305125e-02
1.29077762e-01 5.38010359e-01 -6.01915538e-01 -2.90872425e-01
-3.16953748e-01 -8.61518458e-02 9.93143976e-01 1.06704116e+00
-9.28541958e-01 1.21863663e+00 6.04227352e+00 -1.34872064e-01
-1.42721081e+00 1.06951475e-01 1.64981917e-01 -3.32854152e-01
-4.77220088e-01 2.14041367e-01 -9.55576301e-01 1.78912804e-01
5.73470950e-01 2.96402484e-01 5.77470601e-01 9.74387646e-01
3.13093632e-01 -4.94314075e-01 -1.27559721e+00 1.56954539e+00
1.73954874e-01 -1.73056281e+00 -8.60154927e-02 3.67858529e-01
1.41177666e+00 8.96713912e-01 -1.02280259e-01 -1.64966211e-01
6.93740904e-01 -9.53606904e-01 8.49062204e-01 2.79309571e-01
8.49498451e-01 -2.65334636e-01 7.38305211e-01 5.50866604e-01
-1.21728194e+00 1.43033117e-01 -6.48738325e-01 -7.13488221e-01
6.25292182e-01 6.80061519e-01 -8.95310879e-01 6.34788454e-01
7.05219626e-01 1.36822534e+00 -3.41928840e-01 8.62820148e-01
-3.59540015e-01 -2.45580271e-01 -2.80883938e-01 6.06210589e-01
5.40043473e-01 -2.06425297e-03 4.04481858e-01 6.91464543e-01
5.32486200e-01 -2.23141715e-01 -1.30913453e-02 6.76220953e-01
1.40432417e-01 -4.20001715e-01 -1.24812460e+00 7.04338014e-01
2.94864118e-01 5.68136454e-01 5.68014830e-02 -4.54300940e-01
-6.10345662e-01 1.38070631e+00 4.35667694e-01 1.08721964e-01
-3.92324984e-01 3.63109797e-01 1.08543169e+00 1.54284582e-01
3.00458640e-01 -6.82691455e-01 -6.61420226e-01 -1.69874728e+00
2.16417499e-02 -3.75506908e-01 -2.55033430e-02 -1.31363511e+00
-8.04353535e-01 5.61539412e-01 -3.72112989e-01 -1.65718341e+00
-8.66001487e-01 -7.86846220e-01 -3.28909844e-01 1.03186738e+00
-2.06872272e+00 -8.79410028e-01 -1.03316462e+00 6.40562356e-01
4.63576317e-01 6.48917407e-02 4.30856138e-01 1.61743820e-01
2.09254012e-01 -1.13613017e-01 1.39020711e-01 -1.58221290e-01
8.42270494e-01 -1.40882862e+00 1.19157219e+00 8.68589699e-01
7.60980248e-02 2.69762516e-01 8.09958875e-01 -5.56125164e-01
-1.39844739e+00 -9.42087412e-01 1.05688465e+00 -7.75375843e-01
3.87984723e-01 -1.59053817e-01 -4.80286807e-01 9.76364732e-01
-9.99064669e-02 2.91682661e-01 -1.37127578e-01 -3.37738991e-01
-2.64189363e-01 -1.54315238e-03 -8.72778535e-01 6.92689478e-01
1.44995916e+00 -9.98340070e-01 -3.94477218e-01 1.40917584e-01
7.37773776e-01 -1.06314611e+00 -4.35573518e-01 3.95394415e-01
7.29918957e-01 -1.71601510e+00 1.20471144e+00 6.72049150e-02
7.25089431e-01 -4.58497792e-01 -3.95293683e-01 -1.31422317e+00
-2.04528227e-01 -5.62039852e-01 -1.77252263e-01 3.41677547e-01
1.23882815e-02 -6.33722365e-01 1.39285731e+00 1.41179577e-01
-3.70489478e-01 -3.30427557e-01 -1.02024615e+00 -7.89008498e-01
-7.81025216e-02 -4.84402746e-01 3.92586797e-01 8.34303439e-01
-1.88027218e-01 1.60174221e-01 -7.04380155e-01 2.86945701e-01
8.29055369e-01 8.27021375e-02 1.37899935e+00 -1.17258191e+00
-5.03988683e-01 -2.67715356e-03 -1.02922440e+00 -2.00496578e+00
2.87214547e-01 -5.93497753e-01 1.45897895e-01 -1.70348835e+00
-1.04780179e-02 5.00916205e-02 3.21159393e-01 -6.04956932e-02
2.71671444e-01 6.97651148e-01 2.57383939e-02 4.37219262e-01
-2.35638037e-01 6.90666616e-01 1.32057703e+00 4.90970649e-02
-2.79773504e-01 -4.23060894e-01 -2.25241836e-02 9.52801883e-01
5.08091629e-01 -3.39115292e-01 -4.93249238e-01 -8.83395135e-01
3.27170581e-01 4.45884347e-01 6.39386594e-01 -1.30680072e+00
3.40310335e-01 -4.24682982e-02 4.32641268e-01 -8.85391355e-01
8.67898226e-01 -6.07039154e-01 2.29172230e-01 6.60436213e-01
1.92305177e-01 5.67410290e-02 -5.04790992e-02 5.00302911e-01
-3.93465102e-01 1.04501642e-01 8.66402209e-01 -4.19965327e-01
-1.20403433e+00 5.12781739e-01 -1.99817419e-01 -6.10259138e-02
5.36235571e-01 -6.84115171e-01 -3.80494118e-01 -7.17943668e-01
-5.15273333e-01 1.81040615e-02 1.39971805e+00 2.53764957e-01
7.68918455e-01 -1.26886570e+00 -2.55301893e-01 2.80324101e-01
3.27811658e-01 4.56212401e-01 1.78955778e-01 7.98165262e-01
-1.35108221e+00 8.95841241e-01 -4.18275654e-01 -1.21042275e+00
-9.93090272e-01 2.87786394e-01 6.96101665e-01 -1.15111709e-01
-7.93398678e-01 6.87985957e-01 7.80547738e-01 -8.72296929e-01
-1.33440616e-02 -6.32817447e-01 1.82918966e-01 -5.64454377e-01
1.93149850e-01 2.36922219e-01 7.07178712e-02 -7.79010594e-01
-3.69371921e-01 1.08755064e+00 3.15386385e-01 -5.39717197e-01
1.07824337e+00 -6.09096944e-01 2.45289937e-01 4.28735018e-01
1.46973729e+00 -1.44141391e-01 -2.11164284e+00 -2.67929137e-01
-4.00104254e-01 -8.24006796e-01 3.25609803e-01 -2.48476997e-01
-8.49574685e-01 1.08671999e+00 3.44731510e-01 -6.30930364e-01
5.91110706e-01 -1.58087060e-01 9.45829690e-01 4.16477829e-01
8.26384902e-01 -7.76571989e-01 1.42897651e-01 1.05929244e+00
5.64641893e-01 -1.69206750e+00 2.46713609e-01 -4.42414880e-01
-4.34528381e-01 1.16256917e+00 7.32197821e-01 -3.49322170e-01
4.73640978e-01 7.34150633e-02 3.52091521e-01 -1.58678070e-01
-7.78922200e-01 -2.29669169e-01 2.22039700e-01 7.66403675e-01
3.36144686e-01 -3.69795889e-01 2.95134038e-01 -5.85705161e-01
-3.52629095e-01 2.09750250e-01 7.62461901e-01 1.00136423e+00
-5.14248133e-01 -8.54073465e-01 -2.36792475e-01 -3.05523485e-01
4.56567705e-02 -2.74984360e-01 -3.86447608e-01 7.32909083e-01
-1.04390576e-01 5.98774552e-01 2.07109466e-01 -4.59826767e-01
1.04832925e-01 -4.41278815e-01 7.70993769e-01 -7.14186370e-01
4.63631973e-02 -2.14916110e-01 1.99901640e-01 -1.19346166e+00
-4.52122658e-01 -6.85224473e-01 -1.04991198e+00 -6.17299855e-01
3.27063620e-01 -6.50814712e-01 7.26848066e-01 1.07039547e+00
2.93665320e-01 -2.35307999e-02 5.30139565e-01 -1.74872422e+00
-1.96002007e-01 -5.80039322e-01 -4.95316267e-01 1.97909787e-01
7.96555817e-01 -7.49192059e-01 -5.60239255e-01 -5.49744777e-02] | [8.164275169372559, -2.3116939067840576] |
f43d2c64-2f1f-45a2-95e0-314e96498d0a | long-term-photometric-consistent-novel-view | 2304.10700 | null | https://arxiv.org/abs/2304.10700v1 | https://arxiv.org/pdf/2304.10700v1.pdf | Long-Term Photometric Consistent Novel View Synthesis with Diffusion Models | Novel view synthesis from a single input image is a challenging task, where the goal is to generate a new view of a scene from a desired camera pose that may be separated by a large motion. The highly uncertain nature of this synthesis task due to unobserved elements within the scene (i.e., occlusion) and outside the field-of-view makes the use of generative models appealing to capture the variety of possible outputs. In this paper, we propose a novel generative model which is capable of producing a sequence of photorealistic images consistent with a specified camera trajectory, and a single starting image. Our approach is centred on an autoregressive conditional diffusion-based model capable of interpolating visible scene elements, and extrapolating unobserved regions in a view, in a geometrically consistent manner. Conditioning is limited to an image capturing a single camera view and the (relative) pose of the new camera view. To measure the consistency over a sequence of generated views, we introduce a new metric, the thresholded symmetric epipolar distance (TSED), to measure the number of consistent frame pairs in a sequence. While previous methods have been shown to produce high quality images and consistent semantics across pairs of views, we show empirically with our metric that they are often inconsistent with the desired camera poses. In contrast, we demonstrate that our method produces both photorealistic and view-consistent imagery. | ['Marcus A. Brubaker', 'Konstantinos G. Derpanis', 'Fereshteh Forghani', 'Jason J. Yu'] | 2023-04-21 | null | null | null | null | ['novel-view-synthesis'] | ['computer-vision'] | [ 5.27834237e-01 9.87567455e-02 3.28767091e-01 -3.26384962e-01
-6.41164184e-01 -9.56972301e-01 9.84260261e-01 -6.81550801e-01
1.58208922e-01 6.66341305e-01 1.13940306e-01 4.59490269e-02
8.61086845e-02 -5.90682089e-01 -1.00867736e+00 -6.13591373e-01
4.91148144e-01 3.46939594e-01 2.66501874e-01 1.57517985e-01
2.18507558e-01 7.11344302e-01 -1.60781300e+00 3.18621844e-01
4.88173455e-01 5.20974398e-01 5.43238163e-01 1.05684328e+00
4.30321187e-01 7.89303124e-01 -3.56151938e-01 -4.26161617e-01
5.02857208e-01 -8.74646544e-01 -6.98296726e-01 1.07116616e+00
7.54044354e-01 -5.21154404e-01 -2.19953135e-01 1.05324471e+00
-2.97163948e-02 2.14404643e-01 6.61689401e-01 -1.31324732e+00
-3.42294335e-01 -8.31627846e-02 -4.44400102e-01 -5.01007512e-02
8.36612284e-01 2.92111993e-01 6.53229117e-01 -9.74495649e-01
1.28994536e+00 1.11890757e+00 3.02075922e-01 5.42760313e-01
-1.58090699e+00 -1.34950265e-01 2.46125892e-01 -3.34698021e-01
-1.42612123e+00 -7.09072709e-01 8.25453937e-01 -6.93023086e-01
5.07491291e-01 3.13391864e-01 6.67801380e-01 1.16198552e+00
4.45906460e-01 2.42866978e-01 1.21407068e+00 -5.23208082e-01
4.37670708e-01 2.36090273e-01 -5.28373003e-01 5.67545652e-01
9.84781235e-02 3.96523029e-01 -4.82756019e-01 -8.49443227e-02
1.16063797e+00 2.21140414e-01 -5.26068449e-01 -8.97323430e-01
-1.33715904e+00 6.24783576e-01 9.53721926e-02 1.32466763e-01
-5.26675344e-01 4.37177531e-02 -3.99980098e-01 8.74420628e-02
3.19952905e-01 2.31798381e-01 -6.20891638e-02 1.08774491e-01
-9.52637136e-01 3.02390128e-01 7.54426777e-01 1.33019602e+00
7.38606274e-01 2.45277464e-01 2.96005577e-01 3.21346045e-01
1.91200703e-01 6.36572897e-01 -1.02620587e-01 -1.62284863e+00
2.62937248e-01 2.04071641e-01 2.83082873e-01 -1.00607741e+00
1.87830597e-01 -1.16252951e-01 -7.51373470e-01 7.45977581e-01
4.34499294e-01 -9.84297618e-02 -8.39154363e-01 1.84665859e+00
4.40823287e-01 1.87249944e-01 1.94174036e-01 8.29833627e-01
2.38774940e-01 7.87579536e-01 -5.34854770e-01 -6.06057167e-01
7.97931433e-01 -7.46616960e-01 -4.43488121e-01 -3.52329820e-01
-2.44731139e-02 -1.06147015e+00 6.31341517e-01 5.42652488e-01
-1.58909261e+00 -6.94173753e-01 -9.50389147e-01 -3.23382765e-02
1.65036812e-01 -2.85605997e-01 -1.91516969e-02 4.28709120e-01
-1.20761180e+00 4.14236486e-01 -7.04824567e-01 -3.52287948e-01
-2.03420907e-01 -3.81833361e-03 -6.43279612e-01 -4.62134659e-01
-4.88651752e-01 9.31215882e-01 1.60162956e-01 1.72525132e-03
-1.07153714e+00 -3.23290706e-01 -7.88587213e-01 -2.33944710e-02
4.44473118e-01 -1.25536633e+00 1.05384040e+00 -1.23776233e+00
-1.38529515e+00 8.74585211e-01 -4.56364572e-01 -1.93003818e-01
7.35694706e-01 2.25128885e-02 -2.60332733e-01 2.63425797e-01
6.20835125e-02 9.06199932e-01 1.22833455e+00 -1.96346712e+00
-6.01230204e-01 -3.38378727e-01 1.19176537e-01 5.15312910e-01
5.02059817e-01 -2.18745545e-01 -5.95462084e-01 -3.82064074e-01
4.57399666e-01 -1.24594545e+00 -3.42748225e-01 -5.89226373e-02
-3.78616512e-01 5.10069847e-01 6.72553420e-01 -4.63424742e-01
7.14505672e-01 -2.09656429e+00 5.11715770e-01 5.88882789e-02
1.13806434e-01 -3.21359605e-01 7.26512372e-02 6.62009180e-01
-1.70255065e-01 -1.89707607e-01 -3.08292031e-01 -5.90546012e-01
-4.17509764e-01 3.11249614e-01 -6.26576364e-01 6.49094105e-01
8.82417243e-03 5.48239946e-01 -9.10503387e-01 -3.20725381e-01
6.11835480e-01 5.51744282e-01 -5.94970465e-01 4.72088605e-01
-2.81115174e-01 8.14790785e-01 -9.48902071e-02 2.43040979e-01
7.24228621e-01 -2.45061100e-01 2.52907395e-01 -2.03098059e-01
-2.10526332e-01 -1.84437379e-01 -1.53913116e+00 1.82406855e+00
-4.45954204e-01 6.40304148e-01 -1.29777029e-01 -4.06547666e-01
8.55958045e-01 3.15182686e-01 3.28904510e-01 -1.17418878e-01
6.38903771e-03 4.64909011e-03 -1.80973977e-01 -3.49558383e-01
5.98874152e-01 -3.37945849e-01 1.54994443e-01 5.26803792e-01
-3.40329623e-03 -6.94972634e-01 6.18287101e-02 3.05473506e-01
8.35437894e-01 3.59845221e-01 4.55582291e-01 2.39505284e-02
3.11254472e-01 -7.50158504e-02 3.97584915e-01 6.63458765e-01
1.53740585e-01 1.42486966e+00 1.46852940e-01 -3.83785099e-01
-1.56754398e+00 -1.39566743e+00 1.00760408e-01 1.04107320e-01
3.93199027e-01 -1.74732402e-01 -7.11627305e-01 -4.09515142e-01
-4.63270783e-01 9.00776386e-01 -4.31849182e-01 3.72369550e-02
-5.06411672e-01 -4.97924238e-02 -1.71249807e-01 2.79955268e-01
3.52042615e-01 -7.33090818e-01 -8.41004312e-01 5.38645275e-02
-3.96769524e-01 -1.41034639e+00 -7.79910624e-01 -1.05928861e-01
-9.92491007e-01 -1.04665971e+00 -8.08471501e-01 -5.94646871e-01
1.03781748e+00 7.02471673e-01 1.28048730e+00 -3.70736688e-01
-2.61922572e-02 8.39180589e-01 -4.26551998e-02 1.41926318e-01
-7.40122199e-01 -6.92591727e-01 4.88875359e-02 3.37244660e-01
-2.61390299e-01 -4.97159243e-01 -6.15287304e-01 3.21183741e-01
-1.21836174e+00 6.42072499e-01 2.95126021e-01 9.01547313e-01
7.20016479e-01 7.82302693e-02 -1.98919013e-01 -7.50052929e-01
1.34636581e-01 -3.71785551e-01 -7.08392680e-01 2.65448153e-01
-4.54405278e-01 -4.44155857e-02 6.37323856e-01 -3.46066415e-01
-1.38284242e+00 5.64863920e-01 3.44093055e-01 -8.98072064e-01
-3.26894224e-01 -8.58569797e-03 -1.62901089e-01 2.43435696e-01
6.02062762e-01 3.61436307e-01 3.24330921e-03 -9.01389960e-03
7.00083017e-01 2.09504187e-01 7.85649240e-01 -3.15770835e-01
8.32938135e-01 9.03033257e-01 1.32946849e-01 -8.01764369e-01
-5.23926973e-01 -3.62939835e-01 -1.00129914e+00 -4.53952968e-01
9.67810452e-01 -1.02963996e+00 -2.70757616e-01 3.39805245e-01
-1.40720785e+00 -4.15158570e-02 -5.31705916e-01 6.64989948e-01
-1.08012462e+00 4.82472032e-01 -1.75962701e-01 -9.33522046e-01
2.42678523e-01 -1.29450798e+00 1.26449299e+00 5.65730780e-02
-3.41306388e-01 -1.04879212e+00 1.46504715e-01 1.94600090e-01
-1.88452497e-01 3.81565273e-01 4.82136846e-01 8.70227665e-02
-1.13532698e+00 -6.26764446e-02 1.10898875e-01 4.69351143e-01
1.25889823e-01 2.70180255e-01 -9.78885770e-01 -3.42431098e-01
5.70317686e-01 3.95676643e-02 4.49184537e-01 7.07729280e-01
4.79487568e-01 -1.80028185e-01 -2.30315939e-01 6.45755589e-01
1.71397233e+00 6.68441117e-01 7.88617134e-01 8.67993236e-02
5.23050964e-01 6.53880477e-01 4.46952045e-01 3.23713511e-01
8.77118483e-02 7.48572290e-01 5.04062116e-01 -7.65648335e-02
-1.91974625e-01 -5.76727450e-01 3.33236307e-01 5.93151391e-01
-1.42210305e-01 -6.98219597e-01 -5.09189546e-01 5.81909299e-01
-1.74441147e+00 -1.32139528e+00 -1.64089531e-01 2.56977749e+00
2.73163617e-01 -4.27719243e-02 -1.40717432e-01 -8.37653428e-02
8.84756386e-01 2.00451255e-01 -5.35932183e-01 -3.07075709e-01
-1.46504313e-01 -2.52306432e-01 3.55419040e-01 9.21736658e-01
-5.22140145e-01 5.89004755e-01 6.63679504e+00 2.40400523e-01
-8.60855877e-01 -1.77232206e-01 6.53516352e-01 2.02621892e-02
-5.86646140e-01 4.49752033e-01 -7.01176643e-01 4.05327797e-01
5.64144075e-01 -9.60856229e-02 4.80400801e-01 6.18130922e-01
2.70647049e-01 -4.57099319e-01 -1.36961854e+00 1.02526271e+00
6.40931904e-01 -1.48409832e+00 1.47649601e-01 2.33461499e-01
1.23973000e+00 -3.63177180e-01 1.42363831e-01 -5.49581051e-01
5.87909400e-01 -8.23311269e-01 1.00919878e+00 7.94031680e-01
8.54531467e-01 -5.45180917e-01 1.72503650e-01 5.64330697e-01
-9.07413244e-01 1.60761476e-01 -1.83542594e-01 -5.73361013e-03
6.71621203e-01 3.28869104e-01 -7.50558436e-01 5.86872101e-01
5.28664351e-01 6.68662608e-01 -2.62430817e-01 7.77404249e-01
-1.68636575e-01 -8.62874538e-02 -2.05177978e-01 6.32225752e-01
1.20401166e-01 -5.92001140e-01 7.28612840e-01 5.98904133e-01
7.93248773e-01 2.34276369e-01 8.42293501e-02 1.19073856e+00
2.21193060e-01 -5.05174518e-01 -1.29198933e+00 4.80846167e-01
3.16740602e-01 9.86958265e-01 -6.88298821e-01 -4.06181842e-01
-4.80931550e-01 1.40050375e+00 -1.09100990e-01 5.74862063e-01
-7.40106463e-01 2.62887180e-01 2.69324869e-01 2.19324946e-01
4.06223804e-01 -4.50468749e-01 -1.41980156e-01 -1.41420591e+00
1.45040169e-01 -7.22669601e-01 4.17675972e-02 -1.55559599e+00
-1.02915609e+00 8.49026918e-01 2.96210319e-01 -1.75906169e+00
-7.67287552e-01 -6.19215667e-02 -5.26353776e-01 1.03361559e+00
-1.02622950e+00 -1.05636775e+00 -3.95057291e-01 6.40000105e-01
1.04696476e+00 -1.75406821e-02 5.38532495e-01 -2.00509384e-01
2.95620449e-02 -2.14953944e-01 2.16440007e-01 -4.36106414e-01
6.51842952e-01 -1.12295556e+00 4.28008527e-01 1.26772952e+00
4.65415686e-01 5.54524839e-01 9.75042522e-01 -6.04073942e-01
-1.26893687e+00 -9.12703693e-01 9.16832387e-01 -7.23686934e-01
1.80743679e-01 -1.62266329e-01 -6.71195328e-01 1.04807627e+00
4.69909459e-01 2.01014932e-02 1.08470246e-01 -6.01682603e-01
-8.61847177e-02 9.57326964e-02 -9.65063751e-01 7.92512059e-01
1.10015690e+00 -5.47814667e-01 -5.51560283e-01 1.05831176e-01
4.20128047e-01 -5.11836767e-01 -6.55386746e-01 2.42230311e-01
5.72330773e-01 -1.54426229e+00 1.02453721e+00 -2.62314320e-01
6.36635423e-01 -5.81003010e-01 -3.04300517e-01 -1.43647051e+00
-3.14375937e-01 -6.50550246e-01 2.29596227e-01 1.04055882e+00
1.08805530e-01 -3.07627350e-01 6.72414958e-01 8.51276875e-01
-1.02444090e-01 -2.86821961e-01 -7.00193763e-01 -8.68231952e-01
-2.20603988e-01 -2.83397645e-01 2.75981337e-01 7.41703153e-01
-3.29518080e-01 4.84890938e-01 -7.07569897e-01 4.41592157e-01
7.97501624e-01 4.39503223e-01 1.13158584e+00 -9.45305467e-01
-4.94441628e-01 5.70757175e-03 -3.42327625e-01 -1.27726495e+00
-7.86844194e-02 -3.95736158e-01 2.17904359e-01 -1.39834225e+00
2.38356769e-01 -9.61342156e-02 5.62687814e-01 -3.48301739e-01
1.00786760e-01 2.13207714e-02 3.48944932e-01 4.57849681e-01
-2.12166563e-01 2.95431554e-01 1.34631968e+00 1.37788042e-01
-2.28535309e-01 1.03041083e-01 -2.97483236e-01 1.00174487e+00
3.07327062e-01 -4.14716542e-01 -8.60596359e-01 -5.54979384e-01
2.15809554e-01 7.81972110e-01 5.58088481e-01 -8.69527042e-01
3.32250595e-01 -3.67310882e-01 6.16239429e-01 -6.33291960e-01
7.45068967e-01 -1.09133387e+00 1.15754306e+00 2.51186877e-01
-2.32829526e-01 3.58206540e-01 -2.78072000e-01 8.54288697e-01
-2.47786179e-01 -3.50611269e-01 8.27278912e-01 -4.39078420e-01
-6.38223886e-01 2.74605781e-01 -3.07444751e-01 -1.07777029e-01
1.31287909e+00 -7.19402254e-01 -6.92580417e-02 -7.06467986e-01
-8.92489195e-01 -3.81632686e-01 1.20104694e+00 3.46871674e-01
9.18035090e-01 -1.29170394e+00 -5.62893867e-01 4.54486549e-01
9.80314240e-02 1.69001132e-01 3.85347933e-01 4.66240197e-01
-6.48656547e-01 2.45852575e-01 -1.19162172e-01 -9.82510984e-01
-1.38060975e+00 7.07700253e-01 2.82360107e-01 -1.19488512e-03
-7.55923271e-01 4.40906554e-01 7.63228655e-01 -4.50765435e-03
-1.83139652e-01 -2.95358092e-01 2.63541281e-01 -4.28159148e-01
2.98270941e-01 1.79534152e-01 -1.90650418e-01 -9.37202096e-01
9.46176797e-02 7.91326106e-01 1.93516955e-01 -6.20974123e-01
9.68395412e-01 -6.66959286e-01 2.85657942e-01 6.28070533e-01
1.11228716e+00 2.65227675e-01 -1.86844778e+00 2.61609931e-03
-5.81808627e-01 -1.00538015e+00 -2.66611218e-01 -4.50320691e-01
-8.24239731e-01 6.98541939e-01 4.14420813e-01 1.68351367e-01
9.67286050e-01 3.21657434e-02 4.24460948e-01 -6.80566281e-02
5.06280661e-01 -5.85619628e-01 1.06740043e-01 1.29793614e-01
1.04345655e+00 -1.18348622e+00 5.77955395e-02 -5.64329147e-01
-8.57676506e-01 1.10616553e+00 4.33411628e-01 -1.17028080e-01
3.83907974e-01 1.99027762e-01 -1.02532007e-01 -1.49806932e-01
-9.71583962e-01 -3.85369174e-02 2.02006668e-01 8.12666357e-01
1.23641007e-01 -2.06255391e-01 3.24483246e-01 -3.75609189e-01
-9.59576219e-02 -4.13181633e-03 1.02494025e+00 8.38970482e-01
-2.72439212e-01 -8.18500936e-01 -6.09189451e-01 -8.57102349e-02
-2.32548453e-02 8.03758502e-02 -3.55305940e-01 7.14063942e-01
8.03064033e-02 1.01907253e+00 2.31553227e-01 -1.01170011e-01
2.82057911e-01 -9.88297388e-02 8.15879226e-01 -7.20142066e-01
-7.83260260e-03 3.79648417e-01 -1.16803065e-01 -4.62418705e-01
-7.37187207e-01 -8.05643380e-01 -7.71669924e-01 -2.74974793e-01
-3.25902432e-01 -1.39567345e-01 4.11203355e-01 7.79798448e-01
2.00443983e-01 1.38479605e-01 1.00333011e+00 -1.02244508e+00
-2.37344310e-01 -3.94641340e-01 -7.79518604e-01 6.79138422e-01
4.36497241e-01 -4.67070878e-01 -6.34651542e-01 9.32447910e-01] | [9.319002151489258, -2.701500654220581] |
db97c229-8a97-499d-8816-13055d6682c9 | multi-step-ahead-stock-price-prediction-using | 2212.14687 | null | https://arxiv.org/abs/2212.14687v1 | https://arxiv.org/pdf/2212.14687v1.pdf | Multi-step-ahead Stock Price Prediction Using Recurrent Fuzzy Neural Network and Variational Mode Decomposition | Financial time series prediction, a growing research topic, has attracted considerable interest from scholars, and several approaches have been developed. Among them, decomposition-based methods have achieved promising results. Most decomposition-based methods approximate a single function, which is insufficient for obtaining accurate results. Moreover, most existing researches have concentrated on one-step-ahead forecasting that prevents stock market investors from arriving at the best decisions for the future. This study proposes two novel methods for multi-step-ahead stock price prediction based on the issues outlined. DCT-MFRFNN, a method based on discrete cosine transform (DCT) and multi-functional recurrent fuzzy neural network (MFRFNN), uses DCT to reduce fluctuations in the time series and simplify its structure and MFRFNN to predict the stock price. VMD-MFRFNN, an approach based on variational mode decomposition (VMD) and MFRFNN, brings together their advantages. VMD-MFRFNN consists of two phases. The input signal is decomposed to several IMFs using VMD in the decomposition phase. In the prediction and reconstruction phase, each of the IMFs is given to a separate MFRFNN for prediction, and predicted signals are summed to reconstruct the output. Three financial time series, including Hang Seng Index (HSI), Shanghai Stock Exchange (SSE), and Standard & Poor's 500 Index (SPX), are used for the evaluation of the proposed methods. Experimental results indicate that VMD-MFRFNN surpasses other state-of-the-art methods. VMD-MFRFNN, on average, shows 35.93%, 24.88%, and 34.59% decreases in RMSE from the second-best model for HSI, SSE, and SPX, respectively. Also, DCT-MFRFNN outperforms MFRFNN in all experiments. | ['Mohammad Mehdi Ebadzadeh', 'Hamid Nasiri'] | 2022-12-24 | null | null | null | null | ['time-series-prediction', 'stock-price-prediction'] | ['time-series', 'time-series'] | [-5.18959224e-01 -8.15658152e-01 -9.01133940e-02 -5.58982827e-02
-2.12546736e-01 -2.87683636e-01 3.31620365e-01 -4.06694591e-01
-8.21911320e-02 6.47550821e-01 2.39885584e-01 -3.94786537e-01
-1.55695170e-01 -9.16553855e-01 -1.05842531e-01 -9.36076403e-01
-2.43998915e-02 -6.56476244e-02 1.25063211e-01 -3.49901348e-01
3.94748747e-01 3.53407532e-01 -1.58779812e+00 1.46490455e-01
8.89143527e-01 1.54369807e+00 2.65361041e-01 6.87181205e-02
-9.46305320e-02 8.17847669e-01 -4.95499343e-01 -4.32020873e-01
4.51810718e-01 -3.60383779e-01 -1.58105642e-01 -3.13680023e-01
-5.83455026e-01 -5.20830989e-01 -2.15540037e-01 1.15321302e+00
2.31187627e-01 4.48550701e-01 5.66747069e-01 -1.00572991e+00
-8.23045373e-01 4.82066244e-01 -6.92981541e-01 4.69278574e-01
-1.65680498e-01 -1.55531332e-01 8.55270803e-01 -1.03267443e+00
2.16751024e-01 8.93442929e-01 8.03033888e-01 2.63391316e-01
-1.07119608e+00 -7.23947167e-01 -6.95590451e-02 4.43132758e-01
-1.17797709e+00 -9.72118378e-02 1.08185005e+00 -4.19092566e-01
1.14214754e+00 3.13454360e-01 8.37892056e-01 4.38832790e-01
5.89494586e-01 4.91070271e-01 1.21558666e+00 -1.84634417e-01
1.77298576e-01 -1.38188615e-01 1.29893363e-01 2.94606201e-02
1.48045376e-01 2.45134994e-01 -1.57259032e-01 -1.78112298e-01
7.82279015e-01 4.21455920e-01 -4.15670156e-01 6.90791547e-01
-9.82056737e-01 8.57302845e-01 7.08405748e-02 7.61752903e-01
-8.31889629e-01 -4.12291259e-01 3.54808271e-01 3.81078333e-01
9.41474319e-01 -8.44016895e-02 -6.34514153e-01 -1.76463842e-01
-1.43930328e+00 3.16287637e-01 6.30951405e-01 2.16991439e-01
3.32824290e-01 6.93151593e-01 1.90465659e-01 7.00314164e-01
5.18931806e-01 6.06385827e-01 1.19748664e+00 -1.02121973e+00
4.68492538e-01 2.51287401e-01 2.55932808e-01 -1.53493726e+00
-2.09743112e-01 -6.51240230e-01 -1.15108585e+00 1.92952693e-01
2.35053319e-02 -2.92049259e-01 -3.80828947e-01 1.22583222e+00
8.45550839e-03 3.97719532e-01 3.67925346e-01 8.34376037e-01
5.97409129e-01 1.38588572e+00 -3.82694781e-01 -1.01406264e+00
1.01499569e+00 -8.88757229e-01 -9.13150012e-01 4.04454738e-01
1.31190922e-02 -9.94212806e-01 6.06208026e-01 7.49126136e-01
-1.18925035e+00 -8.28812838e-01 -9.85986471e-01 5.80966055e-01
-9.50121433e-02 2.05446392e-01 1.88175082e-01 3.73763561e-01
-1.02712846e+00 1.09152162e+00 -8.99136245e-01 4.32787567e-01
-2.78905123e-01 1.34327427e-01 1.87634468e-01 3.80312890e-01
-1.47184110e+00 8.70747566e-01 3.05847824e-01 4.58514899e-01
-3.46008539e-01 -5.42782426e-01 -3.68957102e-01 2.33065829e-01
-2.15633377e-01 -3.34001333e-01 1.19864178e+00 -8.93960238e-01
-1.64228559e+00 -1.60433799e-02 -4.14608181e-01 -6.34114623e-01
2.60562181e-01 1.82781205e-01 -1.11087370e+00 2.79050797e-01
2.06504166e-02 -8.81592557e-02 9.84264374e-01 -6.49898171e-01
-6.54533684e-01 -2.20215395e-01 -5.40144145e-01 1.28789747e-03
-4.63649809e-01 -4.31580320e-02 3.61946672e-01 -1.20459831e+00
5.01986861e-01 -5.18815815e-01 2.76580509e-02 -7.94209003e-01
3.05210031e-03 -3.73454422e-01 8.28637838e-01 -1.24957359e+00
1.63255227e+00 -2.06898236e+00 -2.73436278e-01 2.67799467e-01
-3.09466839e-01 3.89416635e-01 2.72860587e-01 6.04307413e-01
-4.03788567e-01 1.26057059e-01 -8.04504827e-02 -1.17213011e-01
-5.77942468e-02 3.00433654e-02 -8.05809319e-01 3.02618861e-01
-8.90286341e-02 5.14312029e-01 -4.41243082e-01 -1.79342315e-01
2.43254200e-01 4.59164739e-01 -1.62910253e-01 -1.65099144e-01
8.87855887e-02 2.93696672e-01 -3.54076087e-01 6.20955586e-01
9.68208909e-01 -2.29520723e-02 -1.62430853e-01 -4.09156531e-01
-8.35389555e-01 8.81368443e-02 -1.25256383e+00 7.11009026e-01
-1.90631405e-01 3.61263096e-01 -1.77435234e-01 -1.16777122e+00
1.38123775e+00 8.18557441e-01 7.33741224e-01 -7.15818703e-01
4.28036191e-02 8.11142385e-01 -2.22781807e-01 -5.20666242e-01
4.73795891e-01 -5.89250267e-01 3.63510460e-01 4.27637160e-01
-3.93838316e-01 2.60340124e-01 4.14911628e-01 -3.91970187e-01
3.60926300e-01 1.74438640e-01 1.11262172e-01 -2.24460244e-01
7.84606099e-01 -1.12202279e-01 1.05526769e+00 -1.15214080e-01
-1.36436507e-01 3.85209590e-01 6.21509142e-02 -6.21396601e-01
-8.56636107e-01 -7.34546483e-01 -2.17233166e-01 4.55061167e-01
-5.91959059e-02 3.95145081e-02 -5.69519579e-01 1.20066740e-02
2.04449315e-02 1.05644798e+00 -2.57931501e-01 1.22681290e-01
-5.03558934e-01 -1.12998748e+00 1.96891665e-01 2.67538965e-01
9.51231599e-01 -9.81038988e-01 -4.73668694e-01 5.66349924e-01
-4.43462521e-01 -6.42390788e-01 -3.49684626e-01 -3.28777015e-01
-1.47776353e+00 -6.52769685e-01 -1.18222725e+00 -5.98382592e-01
3.29465985e-01 4.51486707e-01 7.55115747e-01 -1.49034858e-02
5.66590607e-01 -1.46467939e-01 -3.91992062e-01 -2.85723180e-01
-3.02079439e-01 -3.10343802e-01 2.31984705e-01 4.22081023e-01
2.92255223e-01 -9.39491391e-01 -6.21224821e-01 4.17881846e-01
-7.41932631e-01 -1.31667882e-01 3.53859484e-01 8.38668287e-01
7.09840536e-01 8.80257964e-01 9.70854342e-01 -1.64194360e-01
8.13404560e-01 -5.08542895e-01 -9.75613058e-01 9.01507139e-02
-1.15635347e+00 -3.01060021e-01 1.03976047e+00 -3.66060615e-01
-1.27877581e+00 -3.86500776e-01 -2.29799137e-01 -7.92064190e-01
3.25688034e-01 1.07780659e+00 3.55462134e-01 9.77221504e-02
1.41124815e-01 7.70344079e-01 1.73289984e-01 -7.52636135e-01
-4.32175905e-01 7.01469779e-01 5.48029840e-01 -4.88193817e-02
6.57868385e-01 2.06638634e-01 -1.00988105e-01 -6.33640289e-01
-4.40354049e-01 -3.37221295e-01 -2.76901692e-01 -3.66933227e-01
6.41957283e-01 -9.86388981e-01 -8.59649658e-01 9.16854858e-01
-1.09572875e+00 2.84910440e-01 -1.30623728e-01 9.64221478e-01
-3.31915289e-01 2.97723472e-01 -9.54591632e-01 -1.16093779e+00
-5.90050519e-01 -8.73548865e-01 3.30126315e-01 4.44018036e-01
9.20428559e-02 -1.10477006e+00 7.23819807e-02 2.34537840e-01
5.92253685e-01 3.87688696e-01 7.01696515e-01 -5.90941250e-01
-1.73372790e-01 -3.10325176e-01 1.28914759e-01 7.95588195e-01
1.80197939e-01 2.39894927e-01 -6.46577239e-01 -2.18412846e-01
9.86271918e-01 2.84659952e-01 6.47058666e-01 8.71053278e-01
6.00020289e-01 -5.91974378e-01 1.58721223e-01 4.84514117e-01
1.62653136e+00 9.50938582e-01 6.45425677e-01 6.64391994e-01
5.90506420e-02 3.25612277e-01 7.38353908e-01 6.47988200e-01
4.29298282e-01 1.28017098e-01 3.51815552e-01 1.83672771e-01
5.18934608e-01 9.27652419e-03 6.44902885e-01 1.62955081e+00
-7.66288459e-01 1.09299250e-01 -5.95325828e-01 3.68208557e-01
-1.69171643e+00 -1.48604679e+00 -3.52197558e-01 1.94322050e+00
5.40128231e-01 1.69143915e-01 2.82332152e-01 7.71688402e-01
7.99025953e-01 2.95838475e-01 -6.09687567e-01 -2.44172469e-01
-2.41921499e-01 9.58106145e-02 2.45724469e-01 9.90097821e-02
-9.28707540e-01 6.73094168e-02 5.17136240e+00 7.92870343e-01
-1.45648420e+00 8.83864835e-02 7.08748817e-01 1.93280801e-01
-3.03956896e-01 -1.35760888e-01 -6.66495442e-01 1.05832243e+00
1.18152797e+00 -5.88262916e-01 6.24797046e-01 6.40501142e-01
8.85563433e-01 5.56894280e-02 -1.70127928e-01 9.48895991e-01
-3.39323223e-01 -1.25061250e+00 -1.81654081e-01 1.51268765e-02
8.83567452e-01 -1.15354568e-01 1.76387951e-01 2.41958633e-01
-3.26850742e-01 -5.28015614e-01 8.51533353e-01 1.06872392e+00
1.44754648e-01 -1.25682056e+00 1.20334256e+00 6.97556436e-01
-1.53495634e+00 -2.22722247e-01 -5.90848565e-01 -4.28663462e-01
3.74709964e-01 1.03288531e+00 -8.68852832e-04 9.70014811e-01
8.01863730e-01 9.94456053e-01 8.61583874e-02 7.24693835e-01
6.48066625e-02 8.43967795e-01 -1.37147889e-01 -1.23244815e-01
2.35441461e-01 -7.53794491e-01 4.74939793e-01 5.87732136e-01
9.84322190e-01 3.79579991e-01 -3.63106690e-02 8.06282341e-01
4.56981301e-01 1.31657794e-01 6.98692352e-02 -6.60930276e-02
4.55302566e-01 9.40997720e-01 -6.02857113e-01 -5.66108167e-01
-5.12394905e-01 5.05136907e-01 -5.21971703e-01 4.85549152e-01
-7.96594799e-01 -5.21764100e-01 4.31990296e-01 -3.14536281e-02
6.31933808e-01 -2.20364600e-01 -3.87213945e-01 -1.41909802e+00
2.75999576e-01 -8.16675603e-01 3.03087652e-01 -8.26581120e-01
-1.26229370e+00 6.75629973e-01 -9.03406069e-02 -1.87362397e+00
-6.04862988e-01 -3.15966964e-01 -7.92910457e-01 1.36265171e+00
-1.71977854e+00 -5.26429415e-01 1.65377021e-01 6.33944273e-01
6.08100533e-01 -4.85029548e-01 7.12298512e-01 4.32749718e-01
-6.93161249e-01 1.10273715e-02 8.66449594e-01 -1.78835746e-02
-2.64778621e-02 -6.98787391e-01 2.52363712e-01 1.06935024e+00
-2.42803827e-01 5.70726693e-01 7.33362377e-01 -7.86599219e-01
-9.27631617e-01 -9.44971323e-01 1.12726951e+00 5.22907138e-01
7.05455005e-01 6.40577614e-01 -1.03042662e+00 4.14757013e-01
2.24058300e-01 -1.24061123e-01 4.05371934e-01 -6.89931691e-01
1.53113738e-01 -4.44265544e-01 -1.33338034e+00 2.38712817e-01
6.25665532e-03 -1.83326483e-01 -7.80218542e-01 -1.22934490e-01
4.99047160e-01 -8.51416215e-02 -1.34484923e+00 4.16495115e-01
6.75924599e-01 -1.48461080e+00 8.76228690e-01 8.88892487e-02
4.01382715e-01 -4.90122348e-01 -2.20709324e-01 -1.37158036e+00
-5.12384713e-01 -5.45176327e-01 -2.63037354e-01 1.33357501e+00
2.40652099e-01 -1.19782019e+00 4.18267459e-01 4.38628256e-01
-1.18945993e-01 -8.42016935e-01 -9.92213666e-01 -1.01642263e+00
-3.15780528e-02 -3.59263837e-01 9.65920329e-01 9.67481196e-01
-1.15586102e-01 -1.06604159e-01 -3.84528577e-01 1.15906142e-01
5.89567542e-01 6.18648171e-01 1.02679923e-01 -1.34026694e+00
-3.23079884e-01 -6.51634157e-01 5.96786626e-02 -8.05263460e-01
-1.38830781e-01 -5.11936426e-01 -3.67141485e-01 -1.46299887e+00
-3.01310778e-01 7.04486892e-02 -5.60089529e-01 1.62951007e-01
4.72451262e-02 1.05045840e-01 4.05201584e-01 1.01285601e+00
4.31476891e-01 5.85626245e-01 1.25099945e+00 -8.47312659e-02
-3.46037447e-01 6.35785341e-01 -1.65750265e-01 8.22813094e-01
8.50853562e-01 -2.17955291e-01 -3.51595998e-01 -5.86237647e-02
-5.75678796e-02 5.48342764e-01 -1.52548170e-02 -1.03463650e+00
1.33053496e-01 -2.28224784e-01 5.88192105e-01 -1.20246220e+00
2.24967971e-01 -7.85004616e-01 6.79198921e-01 8.50052357e-01
1.98700130e-01 5.90025425e-01 1.68048544e-03 1.23346530e-01
-8.63544106e-01 -4.82475281e-01 5.70577383e-01 -2.26289049e-01
-4.99343157e-01 8.52648243e-02 -5.73553205e-01 -3.03064823e-01
7.66299725e-01 -3.56463224e-01 -3.23073305e-02 -4.79185283e-01
-7.03922272e-01 9.96368378e-03 3.55234183e-02 -2.04271171e-02
8.25474322e-01 -1.53379405e+00 -7.66811371e-01 2.96433836e-01
-9.49687839e-01 -2.20614836e-01 4.27367121e-01 1.16069150e+00
-5.73107541e-01 5.84005058e-01 -1.37369307e-02 -2.14665771e-01
-9.15821850e-01 5.05226851e-01 2.69540459e-01 -5.77176750e-01
-2.98282474e-01 3.78857404e-01 -2.55527765e-01 -5.44872284e-02
-1.79985777e-01 -6.23163521e-01 -5.81770062e-01 5.31651795e-01
5.86824656e-01 9.25431848e-01 2.09342048e-01 -8.65563154e-01
-5.48044704e-02 8.38750839e-01 3.31952423e-01 -2.53203928e-01
1.83053982e+00 -2.08502799e-01 -4.43260550e-01 7.03297377e-01
1.15377331e+00 -9.86743718e-02 -1.00952733e+00 7.71220401e-02
5.21925017e-02 -2.03317434e-01 2.62046784e-01 -5.49783230e-01
-1.37739420e+00 6.19312644e-01 3.81501347e-01 8.65737498e-01
1.77519333e+00 -9.40055728e-01 1.40618575e+00 5.43411039e-02
4.56711471e-01 -1.17537642e+00 -4.12966788e-01 5.14490902e-01
8.24504972e-01 -6.57839179e-01 4.50133123e-02 -1.78723484e-02
-4.03975457e-01 1.60721052e+00 7.30046537e-03 -3.40873748e-01
9.99449849e-01 2.59086620e-02 3.85214798e-02 3.58035207e-01
-8.02816749e-01 3.98766488e-01 2.73896307e-01 3.36233014e-03
3.23720664e-01 -4.43327874e-02 -5.66567779e-01 1.09374464e+00
-4.70166773e-01 3.44942927e-01 3.84257317e-01 7.52652943e-01
-4.73098665e-01 -7.70200849e-01 -8.32618117e-01 5.40716708e-01
-8.50422442e-01 -9.78241786e-02 3.99857283e-01 7.32584298e-01
6.63126335e-02 9.89850402e-01 2.49664873e-01 -5.89204073e-01
3.93872470e-01 5.58431633e-02 -1.84442535e-01 6.45276606e-02
-6.44241035e-01 6.74323320e-01 -3.01708192e-01 -9.55576226e-02
-8.13850582e-01 -1.10516083e+00 -1.33830190e+00 -5.38991153e-01
-4.76968527e-01 5.35270751e-01 6.52171731e-01 9.01716948e-01
5.15303165e-02 4.11885083e-01 1.35722709e+00 -9.59671080e-01
-9.04087126e-01 -9.64255512e-01 -1.08648646e+00 -2.18855515e-02
3.64829481e-01 -4.17982399e-01 -8.26637685e-01 1.58193544e-01] | [4.735313892364502, 4.066105842590332] |
de8cbea3-d4c8-4be9-aa62-d0e3a40fcd3d | trimap-guided-feature-mining-and-fusion | 2112.00510 | null | https://arxiv.org/abs/2112.00510v3 | https://arxiv.org/pdf/2112.00510v3.pdf | Trimap-guided Feature Mining and Fusion Network for Natural Image Matting | Utilizing trimap guidance and fusing multi-level features are two important issues for trimap-based matting with pixel-level prediction. To utilize trimap guidance, most existing approaches simply concatenate trimaps and images together to feed a deep network or apply an extra network to extract more trimap guidance, which meets the conflict between efficiency and effectiveness. For emerging content-based feature fusion, most existing matting methods only focus on local features which lack the guidance of a global feature with strong semantic information related to the interesting object. In this paper, we propose a trimap-guided feature mining and fusion network consisting of our trimap-guided non-background multi-scale pooling (TMP) module and global-local context-aware fusion (GLF) modules. Considering that trimap provides strong semantic guidance, our TMP module focuses effective feature mining on interesting objects under the guidance of trimap without extra parameters. Furthermore, our GLF modules use global semantic information of interesting objects mined by our TMP module to guide an effective global-local context-aware multi-level feature fusion. In addition, we build a common interesting object matting (CIOM) dataset to advance high-quality image matting. Particularly, results on the Composition-1k and our CIOM show that our TMFNet achieves 13% and 25% relative improvement on SAD, respectively, against a strong baseline with fewer parameters and 14% fewer FLOPs. Experimental results on the Composition-1k test set, Alphamatting benchmark, and our CIOM test set demonstrate that our method outperforms state-of-the-art approaches. | ['Hongtao Lu', 'Zehuan Yuan', 'Yaoyi Li', 'Zhaozhi Xie', 'Dongdong Yu', 'Weihao Jiang'] | 2021-12-01 | null | null | null | null | ['image-matting'] | ['computer-vision'] | [ 1.66617244e-01 -2.91790366e-01 -2.74886936e-01 -5.69042802e-01
-8.53457510e-01 -1.82290599e-02 1.92938626e-01 -1.17232122e-01
-1.82037175e-01 3.75001460e-01 1.91641569e-01 1.16726279e-01
-3.32720056e-02 -9.58336771e-01 -9.58155036e-01 -7.32904136e-01
-2.84841433e-02 7.28154182e-02 5.27428746e-01 -2.13304266e-01
7.89996833e-02 2.23291889e-01 -1.63821805e+00 1.03237772e+00
1.21961784e+00 1.53788137e+00 5.82805157e-01 2.34390125e-01
-3.80930126e-01 3.94624680e-01 -2.16817290e-01 -3.02890271e-01
6.34982646e-01 -1.94544077e-01 -4.93945986e-01 4.38256711e-01
9.65628386e-01 -3.24285150e-01 -2.86204070e-01 1.10886645e+00
2.75025278e-01 -4.97763753e-02 1.86534896e-01 -1.40566301e+00
-4.72316027e-01 6.95362151e-01 -9.36811626e-01 2.00612158e-01
-1.25706717e-01 5.49519360e-01 8.58809531e-01 -1.29221785e+00
4.78449792e-01 1.38312507e+00 5.01591444e-01 -3.45692900e-03
-9.73640382e-01 -7.20207274e-01 5.60498834e-01 3.79330873e-01
-1.26282620e+00 -1.59576640e-01 6.48226619e-01 -1.34078085e-01
7.37677097e-01 3.94667208e-01 7.13903427e-01 4.94097382e-01
5.83263457e-01 9.85346854e-01 1.19945776e+00 -7.78694302e-02
9.40183550e-02 -9.82229486e-02 2.70451576e-01 9.95876133e-01
1.12396494e-01 -3.19602519e-01 -9.53136444e-01 8.69276524e-02
7.11852968e-01 4.14599091e-01 -2.27044955e-01 -7.79108778e-02
-1.43020213e+00 6.19642854e-01 8.29563320e-01 1.41536772e-01
-5.54661512e-01 3.06225747e-01 2.00122908e-01 4.27762866e-01
5.36971152e-01 1.93051800e-01 -5.55618346e-01 2.04972371e-01
-1.20801044e+00 1.75087258e-01 2.47711465e-01 9.18821216e-01
1.26776552e+00 -3.06780860e-02 -5.51282108e-01 8.52130651e-01
2.34054312e-01 6.27195537e-01 2.57650822e-01 -1.22519350e+00
7.74073958e-01 1.00156939e+00 -3.09172601e-01 -1.19094825e+00
-2.12166175e-01 -5.91749966e-01 -1.06970108e+00 1.81590721e-01
-9.42156538e-02 1.56071827e-01 -1.47434545e+00 1.45285738e+00
3.09257865e-01 3.40344787e-01 -1.86454982e-01 1.03088152e+00
9.90172207e-01 8.60665798e-01 -3.13041478e-01 -3.31322670e-01
1.49238956e+00 -1.48914981e+00 -5.55739105e-01 -4.10475224e-01
3.94632846e-01 -8.89296770e-01 1.16504776e+00 6.27462924e-01
-1.22401321e+00 -7.59397566e-01 -1.09281969e+00 -8.65263492e-02
-3.18286091e-01 1.17848702e-01 8.84261847e-01 2.40923285e-01
-1.05190694e+00 4.63023305e-01 -7.40674078e-01 -8.97046775e-02
7.62188852e-01 4.39096421e-01 -4.36544657e-01 -6.42505407e-01
-7.29240835e-01 4.74492520e-01 3.86743426e-01 2.02078134e-01
-9.54886198e-01 -8.72626185e-01 -8.93000841e-01 1.55803673e-02
6.01873279e-01 -9.18082356e-01 7.63701379e-01 -9.81043339e-01
-8.10758352e-01 5.59149981e-01 -5.51630974e-01 -3.98303866e-01
1.88124374e-01 -3.37804854e-01 -3.47610623e-01 3.39926839e-01
3.72622699e-01 1.04940617e+00 9.99268413e-01 -1.05460918e+00
-8.87760758e-01 -3.16921353e-01 6.62779585e-02 3.24103653e-01
-4.94014323e-01 -9.91020054e-02 -8.53450835e-01 -1.02537358e+00
6.24501467e-01 -5.53710103e-01 -3.42882544e-01 2.54982054e-01
-4.04355735e-01 4.23007458e-02 1.04559231e+00 -6.23741567e-01
1.26446474e+00 -2.26853466e+00 7.80880526e-02 2.23148987e-01
4.28744584e-01 3.65853049e-02 -5.14054894e-01 3.97003116e-03
-2.61030323e-03 2.01675622e-03 -3.63835692e-01 -5.18426418e-01
-2.74758399e-01 2.89306045e-01 -1.82997882e-01 9.60672945e-02
4.46899742e-01 1.17237175e+00 -7.48559952e-01 -5.99901021e-01
2.50196248e-01 2.02262715e-01 -7.06723213e-01 -7.10861906e-02
-3.00420284e-01 7.82283619e-02 -3.47338587e-01 1.12810707e+00
1.15077341e+00 -2.88550705e-01 -3.00479978e-01 -7.56838500e-01
-6.04502298e-02 1.12405553e-01 -1.26786196e+00 2.04115033e+00
-3.72035354e-01 3.29059988e-01 1.87613845e-01 -7.69489408e-01
8.29200029e-01 -2.84422636e-01 5.76075435e-01 -7.71025956e-01
-5.80230914e-02 3.04856211e-01 -8.68471041e-02 -2.22873718e-01
5.91860116e-01 1.87590227e-01 1.39350727e-01 2.26263002e-01
2.20154852e-01 1.92656636e-01 1.05338879e-01 4.69171077e-01
1.24448383e+00 -1.75840870e-01 -1.38246164e-01 -3.32432210e-01
3.43036801e-01 3.81161575e-03 9.95341778e-01 5.99429011e-01
-1.87523868e-02 9.99370754e-01 1.64198816e-01 -6.33112609e-01
-6.56422496e-01 -1.05875456e+00 -4.65803221e-02 9.40369010e-01
6.48255587e-01 -6.48868322e-01 -5.79505146e-01 -7.86327183e-01
8.00376460e-02 3.17660838e-01 -7.43383169e-01 -1.29267070e-02
-6.96426630e-01 -1.16723526e+00 -1.62258327e-01 5.78611851e-01
1.17496002e+00 -9.83967841e-01 -1.75774619e-01 1.87036663e-01
-3.65921080e-01 -1.10822141e+00 -7.99757481e-01 1.38464734e-01
-9.91948128e-01 -8.33944917e-01 -4.37128454e-01 -5.99474192e-01
8.29710364e-01 9.48511779e-01 1.10683823e+00 2.81631023e-01
-1.84710547e-01 7.85628036e-02 -4.75479543e-01 -2.18756452e-01
1.30833939e-01 -6.71842769e-02 -2.00307563e-01 2.49720410e-01
1.14650257e-01 -7.00154841e-01 -8.89478266e-01 6.85115278e-01
-9.93396401e-01 5.38887620e-01 9.88716960e-01 7.32370257e-01
8.82833123e-01 1.08469293e-01 5.74303031e-01 -4.23654616e-01
-5.21343164e-02 -4.51682270e-01 -1.65944368e-01 2.97711462e-01
-5.40762424e-01 -6.15431443e-02 1.21305600e-01 -1.91561401e-01
-1.00353420e+00 -5.89792319e-02 3.59834991e-02 -5.96937656e-01
2.05868080e-01 4.06058192e-01 -6.97796583e-01 -2.47425914e-01
3.74623001e-01 4.60383475e-01 -2.20214645e-03 -5.62093914e-01
3.11778277e-01 3.56376231e-01 5.95373809e-01 -5.08517206e-01
8.27554286e-01 6.71276331e-01 -4.91205193e-02 -1.65482953e-01
-9.33251679e-01 -4.67055500e-01 -4.73397166e-01 -1.22173525e-01
7.12902248e-01 -1.37702692e+00 -3.12810749e-01 6.88347638e-01
-9.34941232e-01 -2.14992553e-01 -3.56933236e-01 1.30146757e-01
-3.99686724e-01 3.92583609e-01 -5.63098431e-01 -2.54821360e-01
-6.56616151e-01 -1.40784562e+00 1.26268530e+00 2.47979596e-01
3.77994120e-01 -2.76225239e-01 -6.44981682e-01 7.03215003e-01
4.54618156e-01 5.01018874e-02 5.65434039e-01 -1.46244615e-01
-1.21276176e+00 2.00343058e-01 -6.84355795e-01 5.48620582e-01
2.57079571e-01 -1.97393924e-01 -8.56929898e-01 -3.07147473e-01
-1.41722232e-01 -1.45524800e-01 1.57736993e+00 5.43368042e-01
1.45645857e+00 -3.08796346e-01 -2.92822421e-01 9.68843460e-01
1.38601673e+00 -1.71391353e-01 7.56732583e-01 4.03808087e-01
9.46314394e-01 2.53900975e-01 1.18469512e+00 3.84329289e-01
6.44928932e-01 7.75871277e-01 8.33806276e-01 -3.08450162e-01
-4.05079544e-01 1.23132206e-01 7.09633470e-01 8.15576673e-01
4.49562334e-02 2.68477872e-02 -4.45455581e-01 6.33159101e-01
-2.21132374e+00 -7.38300622e-01 -2.07172453e-01 1.98065734e+00
8.51646960e-01 1.57050580e-01 8.01803637e-03 -1.96562409e-01
5.24424255e-01 2.01432675e-01 -5.99995553e-01 -6.61480576e-02
-6.45235956e-01 9.99144390e-02 3.65629256e-01 2.06240058e-01
-1.20929992e+00 7.96225488e-01 4.87753344e+00 1.42951369e+00
-1.06999373e+00 3.85643750e-01 9.59197462e-01 -3.67045522e-01
-4.62069422e-01 -7.35216141e-02 -9.80965316e-01 7.15319693e-01
3.00233930e-01 1.30922347e-01 2.20080391e-01 7.14604855e-01
2.99653560e-01 -4.95355576e-01 -7.94512868e-01 1.07845092e+00
1.49369806e-01 -1.67469263e+00 2.94161052e-01 2.22926006e-01
9.89356935e-01 2.06548691e-01 1.24810278e-01 2.47548491e-01
-6.84945509e-02 -9.62171912e-01 6.69222713e-01 6.39429450e-01
4.79485363e-01 -7.40581334e-01 9.24396634e-01 6.58147186e-02
-1.47932255e+00 -1.93224311e-01 -4.83557671e-01 1.98740155e-01
1.00301944e-01 1.22558272e+00 -4.16763246e-01 9.59501624e-01
9.28868234e-01 9.49701965e-01 -9.30524647e-01 1.24870336e+00
1.19813763e-01 5.87721169e-01 -5.22091508e-01 6.22715056e-01
4.53117728e-01 -9.05889366e-03 5.42254984e-01 1.11099374e+00
4.73690391e-01 3.82371433e-02 6.13814771e-01 8.53358686e-01
-1.31001949e-01 4.31227824e-03 -3.29245590e-02 4.46606368e-01
3.05702239e-01 1.70975447e+00 -8.49141181e-01 -6.22486532e-01
-4.60987926e-01 9.91408885e-01 3.18022743e-02 2.02301890e-02
-8.15265775e-01 5.20320470e-03 8.83729696e-01 3.87662113e-01
5.41012228e-01 -9.39738229e-02 -6.61615133e-01 -1.35437787e+00
4.36221272e-01 -1.03846860e+00 4.58450407e-01 -7.89690673e-01
-1.23080504e+00 5.45903981e-01 4.64786701e-02 -1.40013897e+00
3.63124460e-01 -3.95905703e-01 -8.40387881e-01 8.60823989e-01
-1.63191450e+00 -1.62657678e+00 -8.20554614e-01 6.49272442e-01
9.49528515e-01 -1.39778584e-01 1.62872940e-01 3.28000456e-01
-5.95714748e-01 6.13106072e-01 -2.81079054e-01 -1.85455859e-01
7.27608144e-01 -1.01644170e+00 3.89820099e-01 1.05475128e+00
2.81860754e-02 5.40846169e-01 2.17469782e-01 -8.31838191e-01
-1.41302514e+00 -1.63190711e+00 3.24802995e-01 -1.69330686e-01
2.81328231e-01 -2.60083884e-01 -1.04927671e+00 3.36566925e-01
2.41499454e-01 4.17431891e-01 2.56733447e-01 -5.77249676e-02
-2.88274646e-01 -5.45588553e-01 -1.14567685e+00 4.16694582e-01
1.31220663e+00 2.90353782e-02 -2.76433975e-01 3.36524308e-01
1.18096948e+00 -2.56453365e-01 -8.55536401e-01 8.78064692e-01
3.93027604e-01 -1.05397511e+00 9.88704979e-01 -4.01536822e-02
5.41709900e-01 -8.98010969e-01 -5.45071900e-01 -1.01715827e+00
-4.78825450e-01 -3.89734924e-01 -1.81000158e-01 1.29196572e+00
2.33336926e-01 -3.00089985e-01 8.58896315e-01 2.46032670e-01
-6.40032351e-01 -1.20050573e+00 -8.80286932e-01 -6.74278617e-01
-4.04768795e-01 -6.10577703e-01 6.44111812e-01 7.32428610e-01
-4.13545549e-01 -1.48730427e-01 -4.45683241e-01 2.31313884e-01
6.88209474e-01 5.19954860e-01 8.43757629e-01 -7.82028973e-01
-2.98955619e-01 -3.05677652e-01 -5.66071093e-01 -1.05577433e+00
-3.42718899e-01 -8.64748120e-01 1.84427444e-02 -1.47401917e+00
8.40677977e-01 -5.02103031e-01 -5.53403378e-01 8.81360829e-01
-5.24006426e-01 6.33987486e-01 4.67169642e-01 3.41115594e-02
-9.03415084e-01 6.89794242e-01 1.46259999e+00 -4.45290387e-01
-1.06195867e-01 -2.96349347e-01 -8.38416338e-01 6.93898082e-01
5.43468654e-01 -3.68844658e-01 -3.30887854e-01 -4.40403879e-01
1.34801373e-01 -4.03964311e-01 6.72977507e-01 -1.26042545e+00
2.14671254e-01 -4.17504400e-01 7.01810837e-01 -1.08140445e+00
4.83772427e-01 -5.51022232e-01 9.99014601e-02 2.42491797e-01
2.74788082e-01 -1.90844573e-02 3.53224188e-01 4.60843563e-01
-3.71678144e-01 1.49341807e-01 6.40250802e-01 -2.17848763e-01
-1.10410357e+00 6.66619718e-01 3.49373072e-02 -2.66891778e-01
9.42841947e-01 -4.37325656e-01 -4.24478412e-01 -1.69017747e-01
-6.12878144e-01 6.28618717e-01 5.54881036e-01 4.32226449e-01
1.05433774e+00 -1.44260073e+00 -7.61072576e-01 4.07215416e-01
1.18863516e-01 3.61754596e-01 5.25422871e-01 1.14873433e+00
-1.76774964e-01 4.15753108e-03 -3.57012689e-01 -8.66304219e-01
-1.33523619e+00 3.04522783e-01 8.18922520e-02 -4.13933456e-01
-6.43770337e-01 1.09994817e+00 8.42806637e-01 -1.75397918e-01
-4.42452803e-02 -5.41458905e-01 3.62192214e-01 -4.50297482e-02
6.99370921e-01 9.32924077e-02 2.81170160e-01 -4.89994526e-01
-3.96345347e-01 6.01953626e-01 -4.28918839e-01 1.85861334e-01
1.39537835e+00 -2.45502323e-01 -3.82472098e-01 8.39758217e-02
9.23887849e-01 -2.22407296e-01 -1.28772342e+00 -5.11945426e-01
-4.62469786e-01 -8.14365745e-01 2.67890990e-01 -9.53806341e-01
-1.61401939e+00 7.89713621e-01 7.18223751e-01 -4.77580935e-01
1.68927312e+00 4.55187596e-02 1.14181173e+00 1.10537559e-01
5.44080019e-01 -9.62109864e-01 4.41318154e-01 1.74699813e-01
1.22169793e+00 -1.34141004e+00 1.76845610e-01 -8.61282051e-01
-5.08837938e-01 8.55889261e-01 1.17494345e+00 -1.10695632e-02
6.08672202e-01 2.76927561e-01 -1.29907995e-01 -2.28502378e-01
-8.64108622e-01 -1.75889283e-01 6.19733691e-01 2.94989437e-01
-6.46844059e-02 5.71542718e-02 -1.15831845e-01 7.25064814e-01
5.90081327e-02 -3.32083255e-02 3.00611258e-01 7.93611348e-01
-9.27372038e-01 -1.00580347e+00 -4.65079129e-01 8.98935199e-01
-1.88613802e-01 -4.88303185e-01 2.11943835e-02 4.35989529e-01
6.91167235e-01 8.29272866e-01 1.82632774e-01 -7.11136520e-01
9.73484665e-02 -2.66731769e-01 3.38754505e-01 -5.88240027e-01
-6.87634706e-01 2.39371583e-01 -9.85234529e-02 -1.07655251e+00
-4.46084559e-01 -5.43220997e-01 -1.17104685e+00 -4.09327030e-01
-3.93657386e-01 -2.21497551e-01 3.37347984e-01 9.20013428e-01
7.49450684e-01 5.72124362e-01 6.32355452e-01 -9.15526032e-01
4.75321114e-02 -1.02654684e+00 -5.16449869e-01 2.32501954e-01
1.97796598e-01 -6.29297435e-01 3.03425454e-03 1.31485844e-02] | [10.619043350219727, -0.8748883605003357] |
d87e9177-98a8-472e-990a-4f7bc8a50679 | coreference-resolution-for-polish | null | null | https://aclanthology.org/2022.crac-mcr.2 | https://aclanthology.org/2022.crac-mcr.2.pdf | Coreference Resolution for Polish: Improvements within the CRAC 2022 Shared Task | The paper presents our system for coreference resolution in Polish. We compare the system with previous works for the Polish language as well as with the multilingual approach in the CRAC 2022 Shared Task on Multilingual Coreference Resolution thanks to a universal, multilingual data format and evaluation tool. We discuss the accuracy, computational performance, and evaluation approach of the new System which is a faster, end-to-end solution. | ['Karol Saputa'] | null | null | null | null | crac-acl-2022-10 | ['coreference-resolution'] | ['natural-language-processing'] | [-3.41956466e-01 3.34300339e-01 -9.48518589e-02 -3.01136136e-01
-1.54373348e+00 -7.85414577e-01 9.27964211e-01 1.29092112e-01
-1.01097369e+00 1.26751661e+00 9.29406524e-01 -1.45510480e-01
-5.64174473e-01 -3.75361085e-01 -1.54474273e-01 -3.56750101e-01
2.27898851e-01 1.58020329e+00 3.77907723e-01 -1.00433636e+00
2.57255025e-02 4.32062685e-01 -1.44271076e+00 5.29096842e-01
8.03023875e-01 6.82379827e-02 4.85818714e-01 6.35314703e-01
-1.67523269e-02 4.28513199e-01 -3.69768828e-01 -7.05584645e-01
-1.20332487e-01 3.62242833e-02 -1.76268589e+00 -1.12009966e+00
6.03995681e-01 4.15938020e-01 -2.68434227e-01 1.31020355e+00
9.30463433e-01 1.53301343e-01 -8.91768467e-03 -8.49814773e-01
3.04855883e-01 1.39924896e+00 -1.35505661e-01 1.83429718e-01
1.10373831e+00 -6.82814598e-01 1.03359008e+00 -7.14792728e-01
1.65545166e+00 1.48933184e+00 8.87659907e-01 3.52524638e-01
-8.91226053e-01 -8.88307333e-01 -4.11236554e-01 4.72594440e-01
-1.50431108e+00 -7.11552262e-01 1.61078811e-01 1.51755754e-03
1.27513039e+00 5.63778341e-01 2.67767489e-01 9.39294338e-01
-1.93179995e-01 4.63794708e-01 8.25303793e-01 -8.94860983e-01
-5.66008270e-01 7.59505853e-02 3.93081486e-01 2.28212297e-01
4.32733387e-01 2.62171894e-01 -5.32389939e-01 -2.85176277e-01
2.23599881e-01 -8.31700623e-01 -4.75979030e-01 -2.91172475e-01
-1.28793895e+00 5.86532176e-01 6.51516691e-02 1.04264748e+00
-1.79365098e-01 -5.90984702e-01 7.56347120e-01 5.85916042e-01
-2.23817583e-02 5.64359725e-01 -6.69464111e-01 -3.86496842e-01
-9.55853403e-01 5.52141309e-01 1.17756760e+00 1.07973886e+00
2.51765758e-01 -5.89048028e-01 2.17649326e-01 1.13153934e+00
1.36340350e-01 6.71011984e-01 2.91391671e-01 -1.19377279e+00
8.41565609e-01 1.61579788e-01 3.56318891e-01 -6.77065849e-01
-1.04682696e+00 -7.07886666e-02 -3.76641363e-01 -7.30261058e-02
6.78036213e-01 -2.53944576e-01 2.14701161e-01 1.79174590e+00
1.62287265e-01 -3.91270220e-01 8.46771061e-01 8.60174060e-01
1.20582247e+00 3.06171238e-01 3.58350761e-02 -6.15690649e-01
1.59227955e+00 -9.61013258e-01 -1.15743256e+00 4.02865976e-01
9.85968828e-01 -1.51122308e+00 2.00163782e-01 5.03107905e-01
-1.18734622e+00 -3.63407731e-01 -8.82731974e-01 -1.41873151e-01
-4.62766767e-01 -1.82925940e-01 4.94339377e-01 5.75560212e-01
-1.05652142e+00 4.60950643e-01 -3.04063588e-01 -1.03458226e+00
-6.69281483e-01 5.18136382e-01 -1.03259444e+00 -1.68520808e-02
-1.79874504e+00 1.42418289e+00 1.01287329e+00 -1.91083670e-01
1.70402944e-01 -5.46645164e-01 -9.37979519e-01 -2.78016597e-01
2.63356328e-01 -2.04170182e-01 1.42500424e+00 -5.39544344e-01
-1.41698849e+00 1.16323042e+00 -8.31770375e-02 -5.60679495e-01
4.63749200e-01 -4.76795495e-01 -1.18476903e+00 2.44847927e-02
3.89365852e-01 6.10461414e-01 -5.22701383e-01 -1.00171745e+00
-1.32102644e+00 -1.63987458e-01 -6.06591627e-02 4.44340199e-01
4.70056236e-01 6.63658559e-01 -6.07095838e-01 -3.67258370e-01
-5.44979163e-02 -8.73089433e-01 7.56240785e-02 -1.09146631e+00
9.70483124e-02 -3.62994879e-01 6.28030002e-01 -9.73087907e-01
1.27846909e+00 -1.78379345e+00 2.42137492e-01 2.30794951e-01
-4.48271871e-01 5.28048456e-01 -3.19391310e-01 8.34519506e-01
-4.87441391e-01 -4.30713117e-01 3.55425780e-03 1.66249778e-02
1.37327239e-01 2.38826230e-01 -1.38261437e-01 3.39553326e-01
-4.42700267e-01 4.35651153e-01 -1.34094512e+00 -7.85529852e-01
2.74576426e-01 5.33169448e-01 -4.40823972e-01 -3.99540439e-02
3.49069983e-01 3.37239683e-01 8.44871774e-02 2.74132282e-01
7.48214543e-01 8.52439702e-01 1.02054143e+00 -4.10229117e-01
-7.53788412e-01 4.96862739e-01 -1.57762480e+00 2.36622000e+00
-5.44932067e-01 5.32860100e-01 5.18124819e-01 -7.83374488e-01
1.06732666e+00 1.10291135e+00 2.49248117e-01 -7.91656494e-01
-8.78617838e-02 8.13238561e-01 3.57374810e-02 -4.09048021e-01
1.11096740e+00 1.19362667e-01 -5.73674858e-01 4.04699922e-01
5.43757498e-01 -4.09723781e-02 5.99255502e-01 2.28428900e-01
5.35383284e-01 3.72415304e-01 5.98193765e-01 -1.06545377e+00
1.20209253e+00 2.33603254e-01 7.88789928e-01 4.13486183e-01
-9.08614174e-02 3.83143991e-01 1.65385127e-01 -4.79032725e-01
-8.65007341e-01 -7.50604212e-01 -4.53179032e-01 9.85688746e-01
4.13226821e-02 -6.34932578e-01 -7.59756148e-01 -4.06222075e-01
-5.21312952e-01 6.71500742e-01 -1.18279666e-01 4.13008690e-01
-1.11509919e+00 -4.17869031e-01 1.06894410e+00 -1.48808658e-02
3.51311684e-01 -1.26888716e+00 -2.06689849e-01 5.59291720e-01
-1.03530145e+00 -1.49697244e+00 -4.46839333e-01 -2.10813563e-02
-4.95625019e-01 -1.69453132e+00 -2.07792148e-01 -1.18763423e+00
-1.47157818e-01 -1.49547249e-01 1.36355627e+00 -2.58439779e-01
-1.44525856e-01 2.90092766e-01 -3.48497003e-01 -5.04397266e-02
-6.15788698e-01 3.75976175e-01 3.13932538e-01 -6.62991822e-01
9.28202927e-01 -2.52234042e-01 8.79918784e-02 8.83416608e-02
-5.14690161e-01 -4.10453320e-01 1.62409514e-01 8.14209759e-01
6.96954578e-02 -4.01729316e-01 3.87506276e-01 -1.20784247e+00
8.48975718e-01 -5.49499951e-02 -7.28244722e-01 4.12717700e-01
-6.40415907e-01 3.69805127e-01 1.38359010e-01 2.55677402e-01
-1.41707706e+00 9.62342024e-02 -6.52124643e-01 3.05220157e-01
-2.29314044e-01 3.33971709e-01 -5.26248515e-01 -7.04832226e-02
4.99263734e-01 -2.79836237e-01 -3.04915547e-01 -8.77940834e-01
5.75011075e-01 8.54797542e-01 1.46238935e+00 -1.12608981e+00
2.02118233e-01 -3.34247835e-02 -2.11828038e-01 -6.59635127e-01
-3.06077898e-01 -9.40191746e-01 -1.21269608e+00 -5.94095849e-02
6.62987828e-01 -1.10535896e+00 -6.93487763e-01 2.67823309e-01
-1.67808688e+00 2.41217092e-01 -2.75786668e-02 8.77121925e-01
-6.84174180e-01 4.77947712e-01 -6.28517509e-01 -4.69114363e-01
-9.76744354e-01 -1.16233969e+00 7.22179174e-01 1.35004250e-02
-7.92645097e-01 -1.21424282e+00 9.50163364e-01 3.75429034e-01
-2.47032568e-02 2.18712822e-01 5.18650830e-01 -8.73407364e-01
4.95539159e-01 -7.25886375e-02 -1.66532993e-01 -3.96495193e-01
-2.81630725e-01 -1.61663160e-01 -5.44480622e-01 -3.86277169e-01
-6.18181348e-01 -2.07320467e-01 4.23851460e-01 -8.06940198e-02
-3.76197040e-01 9.24912691e-02 -4.56320703e-01 4.23239738e-01
1.60419297e+00 2.28781626e-01 6.90693498e-01 9.61964548e-01
1.03962906e-01 8.10996115e-01 1.03567219e+00 1.72165409e-02
5.52877009e-01 1.17264009e+00 -1.41029507e-01 2.74854362e-01
-1.19753860e-01 -5.86382225e-02 7.17002302e-02 1.34072959e+00
-5.15569389e-01 3.46226126e-01 -1.19261432e+00 7.03487217e-01
-2.18884254e+00 -1.28549302e+00 -5.21752715e-01 2.22983623e+00
9.44791436e-01 -9.80777219e-02 3.76391783e-02 1.11333124e-01
1.00579059e+00 -1.67481869e-01 7.89833009e-01 -8.50288987e-01
-6.98580623e-01 6.12250209e-01 4.04064924e-01 1.35256612e+00
-1.19774806e+00 1.33049130e+00 7.23047447e+00 5.43676376e-01
-7.18526840e-01 4.60762471e-01 -7.25615680e-01 1.29536361e-01
9.58638564e-02 6.13231212e-02 -1.01914418e+00 -1.88148499e-01
1.33552706e+00 -3.38633776e-01 3.52175713e-01 2.18880460e-01
-1.89913169e-01 2.93817241e-02 -8.31170797e-01 1.07250094e+00
2.99356449e-02 -1.31386268e+00 -6.22991920e-02 -3.16004604e-01
5.26803732e-01 3.83146614e-01 -7.64487624e-01 3.51674616e-01
7.89577603e-01 -4.43262815e-01 6.64552510e-01 4.35139537e-01
8.32130849e-01 -1.20048487e+00 1.43275583e+00 -1.01786822e-01
-1.48405337e+00 2.27159932e-01 -2.96050608e-01 2.68984467e-01
3.70276034e-01 -7.04476833e-02 -1.05838358e-01 1.50591660e+00
7.39845037e-01 4.26155001e-01 -1.14812523e-01 9.47878242e-01
-2.88154155e-01 -2.21916467e-01 -2.05741361e-01 5.09692311e-01
3.35391879e-01 -1.00595355e-01 8.98873150e-01 2.01536822e+00
2.87530512e-01 1.42415985e-01 -2.71038190e-02 -1.24413557e-01
1.23160288e-01 4.32717562e-01 -4.74957526e-01 5.91558099e-01
9.78671789e-01 1.21979511e+00 -2.52962653e-02 -2.20856979e-01
-3.66761237e-01 6.55446768e-01 5.21466792e-01 -8.12143832e-02
-1.98205158e-01 -8.23105216e-01 5.36496580e-01 -4.77201790e-01
-2.51247715e-02 2.65850037e-01 3.73535424e-01 -9.19957817e-01
-4.39744413e-01 -1.29841387e+00 1.14257395e+00 -4.33531940e-01
-6.40161753e-01 1.01546395e+00 3.56025666e-01 -1.05000854e+00
-7.77513683e-01 -5.24114788e-01 -3.36376786e-01 9.69378591e-01
-1.23433185e+00 -1.14942932e+00 3.07018280e-01 8.69631469e-01
1.26815453e-01 -5.19232273e-01 1.48460722e+00 8.02705944e-01
-1.24717161e-01 9.04281080e-01 2.38102436e-01 2.11325809e-01
1.48222303e+00 -1.07270479e+00 9.52964425e-02 7.43075907e-01
-1.06348857e-01 6.44293010e-01 1.10822761e+00 -2.01231688e-01
-1.09885764e+00 -5.87519109e-01 1.94430649e+00 -2.64842808e-01
9.61309493e-01 2.00503562e-02 -5.83799958e-01 7.91366339e-01
9.02600527e-01 -5.10191083e-01 5.34066200e-01 6.59465015e-01
-6.52064502e-01 -2.00162139e-02 -1.34629071e+00 1.17259346e-01
1.01834369e+00 -5.60726762e-01 -1.38707685e+00 2.15473361e-02
4.50505167e-01 -5.42907715e-01 -1.50870037e+00 4.91349488e-01
6.87861204e-01 -6.95830703e-01 8.73079836e-01 -3.56125027e-01
-3.10771763e-01 -7.11041391e-02 -4.34644848e-01 -1.36430073e+00
-4.18014526e-01 -8.69156241e-01 4.90055442e-01 1.37585092e+00
5.57047725e-01 -5.97200811e-01 9.83158424e-02 6.41374066e-02
-2.83018410e-01 5.41966081e-01 -1.39255404e+00 -4.60269630e-01
3.58995855e-01 -2.91973740e-01 7.22125232e-01 1.28233111e+00
9.68660295e-01 6.33771777e-01 -3.12343389e-01 2.25154042e-01
6.94486260e-01 1.14465848e-01 6.23789728e-01 -1.80278170e+00
1.28543094e-01 -6.06377125e-01 -2.28659838e-01 -1.15487561e-01
4.92407352e-01 -1.02537525e+00 2.81690172e-05 -1.15981090e+00
1.00062042e-01 -1.56887472e-01 -2.88781732e-01 1.36492476e-01
6.23359643e-02 1.35928690e-01 4.15945113e-01 3.15228939e-01
-8.48890305e-01 -2.71880805e-01 6.24583125e-01 -1.50569379e-01
-1.17039211e-01 -5.04842043e-01 -4.99773830e-01 5.43981373e-01
4.31513727e-01 -5.21766484e-01 4.11141396e-01 -4.01289195e-01
1.86589226e-01 4.27425206e-01 -5.07567167e-01 -9.54629481e-01
6.00836337e-01 3.08953166e-01 -2.97087997e-01 -8.94466996e-01
6.61242753e-02 -8.12803626e-01 6.31959438e-01 8.94753456e-01
-5.46929426e-02 3.32083821e-01 4.68315691e-01 -1.94655910e-01
-6.58989489e-01 -4.60551858e-01 8.16428006e-01 -1.32359624e-01
-1.10192943e+00 -3.15054059e-01 -1.98595375e-01 2.57505715e-01
6.44490898e-01 4.21177626e-01 -5.26302695e-01 -1.41315525e-02
-1.02566040e+00 6.00190222e-01 2.52026379e-01 8.35241139e-01
-5.09883724e-02 -1.55362678e+00 -1.30615914e+00 -4.13176902e-02
3.86544228e-01 -6.85050428e-01 -7.40863681e-02 9.18333232e-01
-8.03941727e-01 1.34286451e+00 -7.41264880e-01 -1.64225355e-01
-1.79523671e+00 4.51718390e-01 4.23328966e-01 -8.16115379e-01
-6.48219943e-01 2.34770790e-01 -6.27291739e-01 -1.22706831e+00
3.98412257e-01 5.12679338e-01 -9.68648434e-01 4.11119521e-01
8.30721438e-01 4.20453429e-01 2.06569195e-01 -1.31220233e+00
-8.18291306e-01 7.59195507e-01 -7.44219348e-02 -5.24790645e-01
1.19797003e+00 -2.98975110e-01 -5.62118888e-01 3.63782018e-01
9.08098936e-01 2.53825694e-01 1.32311910e-01 -5.99466145e-01
6.87925279e-01 9.62335616e-02 -1.16723649e-01 -8.30054104e-01
-7.67102718e-01 2.82008410e-01 7.77837574e-01 -2.72526503e-01
6.74994230e-01 -4.92199771e-02 5.02827764e-01 6.13052249e-01
8.41989040e-01 -1.46614254e+00 -1.21521115e+00 1.04554939e+00
9.34593439e-01 -1.02312040e+00 -8.25819466e-03 -3.02109152e-01
-3.77534151e-01 1.37967408e+00 2.19868138e-01 1.15016170e-01
6.40228868e-01 6.60058618e-01 4.71323043e-01 -6.15112595e-02
-6.78048134e-01 -3.81518453e-01 -2.30717827e-02 8.21066320e-01
8.64436865e-01 3.50067556e-01 -1.29433215e+00 7.27708340e-01
-5.01858771e-01 2.20654812e-02 2.00461179e-01 7.16255128e-01
-2.97227114e-01 -1.77813160e+00 -6.70075476e-01 -5.49363256e-01
-7.64076829e-01 -2.55281985e-01 -5.19106507e-01 1.42891240e+00
2.29342476e-01 1.02641416e+00 1.26674682e-01 -3.70040268e-01
8.71642888e-01 2.35532895e-01 6.97452784e-01 -2.20519185e-01
-1.03184962e+00 -2.57280115e-02 9.53144312e-01 -6.74247980e-01
-9.41025853e-01 -1.23942196e+00 -1.26985025e+00 -5.63376248e-01
-1.84137329e-01 1.07175171e+00 4.19851810e-01 8.86756301e-01
-1.19392417e-01 1.07985586e-01 2.69108385e-01 -7.42199957e-01
4.76813689e-02 -1.33992589e+00 -4.50643331e-01 4.98606622e-01
-1.89066663e-01 -3.91387463e-01 -6.38524396e-03 -1.51642546e-01] | [9.327295303344727, 9.614083290100098] |
642128a3-ad34-4907-bd1c-26109edbc47b | approximate-conditional-coverage-via-neural | 2205.14310 | null | https://arxiv.org/abs/2205.14310v3 | https://arxiv.org/pdf/2205.14310v3.pdf | Approximate Conditional Coverage & Calibration via Neural Model Approximations | A typical desideratum for quantifying the uncertainty from a classification model as a prediction set is class-conditional singleton set calibration. That is, such sets should map to the output of well-calibrated selective classifiers, matching the observed frequencies of similar instances. Recent works proposing adaptive and localized conformal p-values for deep networks do not guarantee this behavior, nor do they achieve it empirically. Instead, we use the strong signals for prediction reliability from KNN-based approximations of Transformer networks to construct data-driven partitions for Mondrian Conformal Predictors, which are treated as weak selective classifiers that are then calibrated via a new Inductive Venn Predictor, the Venn-ADMIT Predictor. The resulting selective classifiers are well-calibrated, in a conservative but practically useful sense for a given threshold. They are inherently robust to changes in the proportions of the data partitions, and straightforward conservative heuristics provide additional robustness to covariate shifts. We compare and contrast to the quantities produced by recent Conformal Predictors on several representative and challenging natural language processing classification tasks, including class-imbalanced and distribution-shifted settings. | ['Danielle Rasooly', 'Allen Schmaltz'] | 2022-05-28 | null | null | null | null | ['protein-secondary-structure-prediction', 'grammatical-error-detection'] | ['medical', 'natural-language-processing'] | [ 3.51069570e-01 5.67482412e-01 -5.30483067e-01 -1.03306031e+00
-8.07707608e-01 -6.47758543e-01 9.62536395e-01 4.66976941e-01
-3.28049272e-01 9.93948817e-01 3.09446752e-01 9.69258696e-03
-8.37193668e-01 -8.79976928e-01 -8.13067675e-01 -1.00594246e+00
-1.07699715e-01 1.06232083e+00 3.10069650e-01 1.83141068e-01
2.55221844e-01 2.80500621e-01 -1.49156356e+00 2.66478211e-01
6.31683230e-01 9.38302696e-01 -3.13028842e-01 4.26410288e-01
2.54268706e-01 5.92143774e-01 -2.76274532e-01 -7.44397461e-01
1.25378698e-01 -2.96280861e-01 -6.22427106e-01 -5.77055991e-01
6.05405569e-01 1.72035649e-01 -8.88440236e-02 1.06417108e+00
2.65550137e-01 -4.60026301e-02 1.30273426e+00 -1.58553636e+00
-5.50442100e-01 1.56391490e+00 -4.83855635e-01 3.01093102e-01
-3.68453205e-01 -2.48873889e-01 1.39757311e+00 -5.11834502e-01
5.68169296e-01 1.31103766e+00 1.15478790e+00 2.01121420e-01
-1.62329030e+00 -6.24867976e-01 1.63323015e-01 -6.44681826e-02
-1.36223006e+00 -4.37403977e-01 3.69790256e-01 -5.96241057e-01
3.65518510e-01 2.88370967e-01 2.83217490e-01 1.38214934e+00
6.30523562e-01 5.84510341e-02 9.96254146e-01 -2.44719088e-01
5.65613389e-01 5.76684475e-01 3.35797340e-01 1.90981135e-01
5.64438939e-01 3.05582017e-01 -5.90752900e-01 -4.70130742e-01
4.23114687e-01 -1.47589371e-01 -3.69873941e-01 -8.35591316e-01
-1.28461111e+00 9.53546584e-01 4.70495522e-01 1.69865802e-01
-1.82099208e-01 2.14435637e-01 5.49691200e-01 1.55240431e-01
7.91033924e-01 5.47907650e-01 -6.82077765e-01 2.38787457e-01
-1.09896827e+00 3.69181305e-01 8.25215161e-01 9.83147204e-01
4.81269658e-01 -3.33992869e-01 -6.92197025e-01 6.40851796e-01
1.88675508e-01 3.34442377e-01 2.13607728e-01 -8.68993998e-01
3.22664380e-02 1.75116539e-01 -1.28978282e-01 -9.24332559e-01
-6.56158328e-01 -1.01235688e+00 -1.07346165e+00 -7.18364492e-02
7.22672403e-01 2.57415146e-01 -7.49781847e-01 2.13453317e+00
2.74942368e-01 -5.03577106e-02 -2.58544177e-01 6.53898597e-01
3.19672614e-01 3.51339012e-01 4.43105996e-01 -1.07056960e-01
1.26236570e+00 -4.23826337e-01 -1.73523694e-01 -1.74194768e-01
4.49529737e-01 -2.11228266e-01 1.13908947e+00 4.36256737e-01
-7.42893815e-01 -2.75989413e-01 -1.19545031e+00 1.23881817e-01
-2.18349546e-01 -3.02694440e-01 5.21256566e-01 7.32949495e-01
-1.02848351e+00 8.87291849e-01 -4.69074458e-01 -1.76233336e-01
6.14685118e-01 1.96680978e-01 -1.62247553e-01 1.68667108e-01
-1.11862767e+00 9.32126403e-01 4.53204632e-01 -2.69708514e-01
-1.03915966e+00 -1.16888654e+00 -5.21373570e-01 3.36928368e-01
5.00001246e-03 -5.30106366e-01 1.12249458e+00 -1.12069786e+00
-8.65219772e-01 1.02947748e+00 1.62419662e-01 -7.57018805e-01
7.03880727e-01 3.07612747e-01 -6.83634281e-02 -2.65125573e-01
9.56923291e-02 8.71018350e-01 8.22025061e-01 -1.37930000e+00
-5.28790951e-01 -4.44871873e-01 -5.38588315e-02 -7.19815609e-04
-3.35993558e-01 -3.20707440e-01 1.84502572e-01 -6.13737285e-01
1.80805489e-01 -7.69096315e-01 -2.32647151e-01 8.68778974e-02
-6.46070063e-01 -2.44829014e-01 1.60439312e-02 -1.91149175e-01
8.80355179e-01 -2.15208507e+00 9.50474069e-02 5.34823418e-01
4.36022133e-01 -5.34878612e-01 -8.40765685e-02 -7.08427206e-02
-2.47353345e-01 7.55799785e-02 -5.72808981e-01 -2.44143397e-01
3.61700714e-01 9.56154317e-02 -5.45788586e-01 9.66837347e-01
1.77861482e-01 5.44908226e-01 -7.22789705e-01 -4.93795663e-01
-9.13586318e-02 5.04977405e-01 -5.54692745e-01 -1.81166962e-01
-2.80753404e-01 1.61101192e-01 1.09819353e-01 2.25627273e-01
6.53060138e-01 -1.86248630e-01 2.72641122e-01 -2.92135067e-02
2.92496532e-01 4.95017409e-01 -7.87513554e-01 9.05650079e-01
-2.65466213e-01 7.13813484e-01 -1.83604509e-01 -1.29889071e+00
1.11992431e+00 -1.10960573e-01 1.78438157e-01 -4.20563161e-01
1.42688408e-01 5.12326546e-02 -1.23582007e-02 2.39422068e-01
3.40292186e-01 -4.81997073e-01 -1.96584806e-01 3.56762826e-01
3.12693775e-01 -8.14334303e-02 -2.03008682e-01 5.66501431e-02
9.47826445e-01 3.77257280e-02 3.57523412e-01 -1.10164046e+00
2.59687826e-02 -1.44400299e-01 7.11294174e-01 1.07165921e+00
-6.59927502e-02 8.18777561e-01 1.14534867e+00 -2.70977587e-01
-1.53126192e+00 -1.36645389e+00 -1.17461944e+00 1.25062096e+00
-1.37477145e-01 -1.57950923e-01 -8.04627538e-01 -6.43383205e-01
2.19010621e-01 1.01536000e+00 -1.06134033e+00 -5.33502102e-01
6.98617892e-03 -1.11530447e+00 6.26226246e-01 6.32956445e-01
-2.28576437e-01 -6.91287160e-01 -4.34021115e-01 -1.03887819e-01
9.12335515e-02 -6.71272814e-01 -1.75324693e-01 8.69078755e-01
-7.64828444e-01 -1.09104502e+00 -6.09529436e-01 -5.14951527e-01
5.58282137e-01 -3.93693447e-01 1.66054642e+00 -2.27053657e-01
1.50874346e-01 7.32631832e-02 3.11610904e-02 -4.95736688e-01
-5.00190318e-01 3.30268502e-01 3.86050284e-01 -1.47499487e-01
5.55147946e-01 -7.82612562e-01 -5.26263714e-01 5.32590091e-01
-6.64671421e-01 -2.56505787e-01 4.53665942e-01 9.56575572e-01
8.33538294e-01 -1.00575732e-02 7.28430867e-01 -1.11446500e+00
4.82629865e-01 -7.65894711e-01 -6.18879080e-01 4.05647457e-01
-8.79844904e-01 3.04950804e-01 6.00243449e-01 -6.70479298e-01
-7.40046740e-01 -1.79843917e-01 3.96545947e-01 -1.02428347e-01
4.66170013e-02 3.68720382e-01 -1.16094314e-01 2.10302696e-01
1.17500508e+00 -3.17784071e-01 -3.31632257e-01 -3.20559070e-02
3.27219903e-01 5.95975995e-01 7.07480848e-01 -7.14981914e-01
5.81346452e-01 3.61264110e-01 1.12656586e-01 -3.16374779e-01
-1.10759652e+00 -7.96578359e-03 -7.49602079e-01 -3.16876993e-02
4.91821647e-01 -7.89976180e-01 -5.31115711e-01 1.36797624e-02
-8.36220682e-01 -4.28859562e-01 -6.64725065e-01 4.12502497e-01
-6.75853908e-01 -1.49430096e-01 -2.80040234e-01 -7.61049330e-01
-1.40083849e-01 -8.50327015e-01 9.13459539e-01 -4.51373309e-02
-4.82657939e-01 -1.01345778e+00 1.32073879e-01 -2.16683999e-01
3.95847261e-01 1.80985883e-01 1.37753165e+00 -1.06102681e+00
-1.75171390e-01 -1.66684255e-01 -3.33471090e-01 2.34928682e-01
-3.60800773e-01 2.54197866e-01 -1.37967300e+00 -1.91813886e-01
-1.10683732e-01 -1.70005798e-01 1.16260958e+00 8.01450789e-01
1.46367586e+00 -2.39509091e-01 -5.40349364e-01 6.46283746e-01
1.10682499e+00 -2.67681152e-01 4.03031766e-01 1.08784698e-01
1.35987908e-01 9.06392872e-01 3.01544219e-01 3.90361756e-01
1.66916385e-01 4.46063936e-01 3.40631485e-01 2.25848809e-01
2.17348695e-01 -5.63680112e-01 -4.57503460e-02 3.84915560e-01
3.08606952e-01 -2.17666894e-01 -1.05607605e+00 5.90474308e-01
-1.67976165e+00 -7.48572886e-01 -1.37181595e-01 2.52244782e+00
1.06749654e+00 5.35832584e-01 1.82668194e-02 1.20125614e-01
9.10096467e-01 -1.49489548e-02 -7.74115741e-01 -1.34375095e-01
-4.23157603e-01 7.17216507e-02 8.20308924e-01 4.52965170e-01
-1.17877936e+00 4.16835696e-01 6.96698380e+00 9.47269678e-01
-7.42195427e-01 2.61402875e-01 1.29121184e+00 -2.14470387e-01
-8.25446844e-01 5.78728653e-02 -7.20875859e-01 4.52032804e-01
1.40441871e+00 -2.38285333e-01 1.27207592e-01 1.03401268e+00
-3.03850353e-01 -1.17485963e-01 -1.59612393e+00 6.41224802e-01
-8.72218758e-02 -1.29253078e+00 -1.58895060e-01 6.32468835e-02
6.42800987e-01 5.26150465e-02 3.37832510e-01 2.04132333e-01
5.80840230e-01 -1.26057458e+00 1.02396059e+00 7.62267828e-01
9.17281747e-01 -9.63860631e-01 9.14435685e-01 1.58825368e-01
-4.94550616e-01 -2.10633725e-01 -8.74221146e-01 1.81357220e-01
-2.21735939e-01 1.21835053e+00 -7.54298925e-01 -7.50489011e-02
9.48599458e-01 4.52460080e-01 -5.37123501e-01 1.05246103e+00
8.67298022e-02 8.17329943e-01 -4.54291552e-01 -1.15829505e-01
-1.01760954e-01 4.55395281e-02 3.40100527e-01 1.10601604e+00
1.80220962e-01 -3.32525909e-01 -4.46767598e-01 1.10158849e+00
-2.09746957e-01 -1.06045693e-01 -5.03722250e-01 2.66945869e-01
6.69588029e-01 1.17833602e+00 -9.06595051e-01 -6.58601746e-02
1.02416441e-01 2.20109731e-01 5.03113270e-01 1.71239272e-01
-1.05677080e+00 -1.50315046e-01 7.69206643e-01 3.17402542e-01
1.08582765e-01 2.51305223e-01 -7.87857473e-01 -8.73190284e-01
-1.07292689e-01 -5.76046526e-01 5.54523408e-01 -6.58817947e-01
-1.93167472e+00 5.50541401e-01 2.59042323e-01 -9.92395759e-01
7.26383850e-02 -6.61260545e-01 -3.29376847e-01 5.08861482e-01
-1.03612363e+00 -7.26512313e-01 -5.72016127e-02 2.45559305e-01
-1.21309996e-01 -1.05642229e-01 7.47994721e-01 -8.81111994e-03
-2.27680892e-01 9.39151943e-01 3.58555555e-01 -1.75296590e-01
1.14441860e+00 -1.35626841e+00 2.92832464e-01 5.50328732e-01
2.88059890e-01 4.71958101e-01 9.42622840e-01 -4.09714252e-01
-4.87422138e-01 -1.10282886e+00 7.75005519e-01 -1.03330708e+00
5.96402645e-01 -6.59576654e-01 -9.36091840e-01 7.31509745e-01
-1.70387834e-01 -7.36801773e-02 5.62498450e-01 6.46891475e-01
-8.20398450e-01 -4.95713204e-01 -1.37534440e+00 1.61295742e-01
1.15487552e+00 -3.47395182e-01 -4.79280829e-01 5.27131140e-01
8.16604555e-01 -1.85316131e-01 -9.56715345e-01 5.89820862e-01
5.63169420e-01 -9.92422342e-01 8.75348449e-01 -1.00028241e+00
7.18500733e-01 -1.53175695e-02 -4.58500862e-01 -1.39093745e+00
-5.05899131e-01 -1.12122558e-01 3.51114601e-01 1.30180347e+00
8.25109124e-01 -6.44810736e-01 5.86310267e-01 6.21102691e-01
2.14477181e-02 -6.78127348e-01 -1.27713954e+00 -4.19151634e-01
6.49351895e-01 -5.44463813e-01 7.67932296e-01 1.10528910e+00
3.14688273e-02 2.04006284e-01 -1.28945366e-01 1.28352642e-01
9.87641811e-01 -1.15170531e-01 2.10571319e-01 -1.82752001e+00
-1.61083341e-01 -6.51595175e-01 -5.88573515e-01 -4.12678629e-01
3.33022356e-01 -1.03487194e+00 2.49884814e-01 -9.07853007e-01
6.55544698e-01 -9.91641879e-01 -4.51114386e-01 2.98596829e-01
1.63993806e-01 2.44549423e-01 -3.58403474e-01 1.21813186e-01
-3.33984554e-01 4.90611881e-01 6.20496690e-01 -2.17221200e-01
1.47115335e-01 1.70834899e-01 -1.14801419e+00 6.88405097e-01
6.73343182e-01 -8.07199895e-01 -5.34730256e-01 -4.33565788e-02
6.90714002e-01 -3.26334536e-01 5.97043753e-01 -8.36829543e-01
2.10448876e-01 -1.07214436e-01 7.30788887e-01 -3.03172350e-01
8.21418762e-02 -5.95576465e-01 1.80029050e-01 7.82574341e-02
-9.20214713e-01 -1.01583570e-01 -3.19307745e-02 7.51613617e-01
1.11570470e-01 -1.37066737e-01 1.19912040e+00 1.55732661e-01
-2.12990474e-02 3.00084680e-01 -2.30978671e-02 3.81687880e-01
9.21651483e-01 -5.39983995e-03 -6.75947189e-01 -4.90342110e-01
-5.39890766e-01 9.55323875e-02 3.77279550e-01 3.77726465e-01
1.63336366e-01 -1.12630594e+00 -8.65178406e-01 6.68968335e-02
2.40044951e-01 9.60074514e-02 5.56666441e-02 8.26988459e-01
-1.78676024e-01 3.59907895e-01 -2.58530855e-01 -1.02206242e+00
-6.70069396e-01 4.94593114e-01 6.88616395e-01 1.31135613e-01
-4.56431717e-01 1.07130337e+00 4.68937129e-01 -8.21289837e-01
4.45342302e-01 -5.03836036e-01 1.18040696e-01 2.60047972e-01
3.75102580e-01 4.56509553e-02 1.36278272e-01 -2.31360987e-01
-4.17953908e-01 1.63706258e-01 1.44070819e-01 -3.94187346e-02
1.47288454e+00 -3.46007980e-02 -3.44114751e-01 7.63297856e-01
7.81369686e-01 -2.11473972e-01 -1.35432696e+00 -2.55924463e-01
1.73066080e-01 -2.33033076e-01 -1.52525716e-02 -8.59085381e-01
-9.22812998e-01 8.44885945e-01 4.63703543e-01 3.41641665e-01
7.30526507e-01 3.76987964e-01 -2.53899723e-01 1.26104087e-01
3.66076231e-01 -8.76350462e-01 -5.28935611e-01 2.64587164e-01
8.95018697e-01 -9.59124446e-01 6.10793419e-02 -1.62495837e-01
-5.70467174e-01 1.00495422e+00 5.59308708e-01 8.49231426e-03
9.24065709e-01 4.79145139e-01 -1.82588369e-01 -2.17273515e-02
-9.52520728e-01 4.82863873e-01 3.28204989e-01 8.20805252e-01
3.67380679e-01 3.71765703e-01 4.95175123e-02 1.04267645e+00
-6.86972678e-01 -5.64190447e-01 4.22139764e-01 -8.34901333e-02
-3.70437771e-01 -4.47337449e-01 -1.86016306e-01 8.83389831e-01
-1.30297899e-01 -1.84857428e-01 -2.52005547e-01 8.11777651e-01
-1.12987146e-01 4.73131180e-01 6.65610492e-01 -3.50144744e-01
3.92090110e-03 2.54308134e-01 3.35563719e-01 -2.55313575e-01
-2.59245545e-01 -5.96248806e-01 9.51392427e-02 -3.44428331e-01
-3.12306255e-01 -8.59759390e-01 -7.55297303e-01 -5.43562531e-01
-2.48662755e-01 2.08081186e-01 4.03352022e-01 8.07210863e-01
1.85610175e-01 3.56670052e-01 2.92358071e-01 -7.38919437e-01
-1.07972074e+00 -1.10861337e+00 -6.17868960e-01 2.28246301e-01
1.43231377e-01 -8.87805343e-01 -7.87619352e-01 -4.04251605e-01] | [8.01925277709961, 4.102053165435791] |
35956758-8d5e-4ce5-9ef6-a21b6389e941 | domainmix-learning-generalizable-person-re | 2011.11953 | null | https://arxiv.org/abs/2011.11953v3 | https://arxiv.org/pdf/2011.11953v3.pdf | DomainMix: Learning Generalizable Person Re-Identification Without Human Annotations | Existing person re-identification models often have low generalizability, which is mostly due to limited availability of large-scale labeled data in training. However, labeling large-scale training data is very expensive and time-consuming, while large-scale synthetic dataset shows promising value in learning generalizable person re-identification models. Therefore, in this paper a novel and practical person re-identification task is proposed,i.e. how to use labeled synthetic dataset and unlabeled real-world dataset to train a universal model. In this way, human annotations are no longer required, and it is scalable to large and diverse real-world datasets. To address the task, we introduce a framework with high generalizability, namely DomainMix. Specifically, the proposed method firstly clusters the unlabeled real-world images and selects the reliable clusters. During training, to address the large domain gap between two domains, a domain-invariant feature learning method is proposed, which introduces a new loss,i.e. domain balance loss, to conduct an adversarial learning between domain-invariant feature learning and domain discrimination, and meanwhile learns a discriminative feature for person re-identification. This way, the domain gap between synthetic and real-world data is much reduced, and the learned feature is generalizable thanks to the large-scale and diverse training data. Experimental results show that the proposed annotation-free method is more or less comparable to the counterpart trained with full human annotations, which is quite promising. In addition, it achieves the current state of the art on several person re-identification datasets under direct cross-dataset evaluation. | ['Ling Shao', 'Cuicui Kang', 'Fang Zhao', 'Shengcai Liao', 'Wenhao Wang'] | 2020-11-24 | null | null | null | null | ['generalizable-person-re-identification'] | ['computer-vision'] | [ 8.86304602e-02 -3.49232674e-01 -3.11814368e-01 -5.13040602e-01
-6.16667628e-01 -5.82435787e-01 5.36109209e-01 -6.83873072e-02
-6.84709489e-01 8.91800344e-01 1.40116308e-02 2.90733755e-01
-3.13870073e-03 -6.44080460e-01 -3.86856943e-01 -6.99029624e-01
5.63611984e-01 6.89519823e-01 7.92281777e-02 -2.51217615e-02
-1.65097088e-01 2.03785554e-01 -1.38405561e+00 -2.78008491e-01
1.14754117e+00 8.47464740e-01 -1.26053244e-02 -1.23579890e-01
2.03676760e-01 5.92279360e-02 -6.37335718e-01 -6.04795635e-01
4.02736932e-01 -5.30920923e-01 -6.11219347e-01 3.08093011e-01
4.55551565e-01 -3.74971956e-01 -5.05317450e-01 1.24654245e+00
8.13216269e-01 2.47579277e-01 5.29810250e-01 -1.40960872e+00
-8.91564727e-01 2.58219928e-01 -6.20144188e-01 4.21955204e-03
4.09870207e-01 1.85365617e-01 5.88777661e-01 -6.86100900e-01
3.56377006e-01 1.20131218e+00 7.94534445e-01 1.04443574e+00
-1.29779994e+00 -1.30201840e+00 2.34219357e-01 7.73281753e-02
-1.78890717e+00 -3.33595067e-01 8.15717041e-01 -5.39521575e-01
2.89104283e-01 1.65077090e-01 4.84039932e-01 1.39766943e+00
-5.34213901e-01 6.60947502e-01 1.17345905e+00 -2.74148107e-01
4.00097221e-02 5.34605682e-01 1.61732540e-01 2.59895742e-01
5.89632690e-01 2.21498683e-01 -1.75448686e-01 -1.15386099e-01
7.80569732e-01 4.65745658e-01 -3.92791182e-01 -4.79333669e-01
-1.40963233e+00 4.49591547e-01 2.94501513e-01 2.40511715e-01
-5.62482923e-02 -4.84367549e-01 7.57791281e-01 2.62055337e-01
3.98605049e-01 2.14839920e-01 -2.36930922e-01 1.40316531e-01
-8.20446312e-01 4.06755120e-01 3.79614860e-01 9.97099578e-01
7.46102929e-01 -1.58157796e-01 -2.84770370e-01 1.24119866e+00
6.77226409e-02 7.85333395e-01 1.00585592e+00 -5.21716833e-01
6.32507682e-01 8.89820993e-01 3.75952721e-01 -9.54889655e-01
-3.47332329e-01 -4.28560257e-01 -1.26365387e+00 -1.70245543e-01
6.50927186e-01 -1.60682797e-01 -7.27367043e-01 2.01793814e+00
3.50511521e-01 4.60972905e-01 2.57743776e-01 1.07723951e+00
7.35393941e-01 2.68793851e-01 3.84063959e-01 -1.91242188e-01
1.51445818e+00 -8.34429443e-01 -5.16293824e-01 -3.38946968e-01
2.71372646e-01 -3.27816576e-01 1.08346045e+00 1.68430693e-02
-6.31273687e-01 -9.41730738e-01 -1.03796387e+00 1.98361859e-01
-3.05742711e-01 4.27665502e-01 4.06147391e-01 8.60670686e-01
-5.02123773e-01 2.09802747e-01 -3.09358448e-01 -6.28980219e-01
4.20076758e-01 4.86890107e-01 -7.84030437e-01 -3.62367570e-01
-1.39740300e+00 5.00803649e-01 8.97344053e-01 -6.14857860e-02
-6.71254039e-01 -6.06925488e-01 -9.18419719e-01 9.57648549e-03
3.72688472e-01 -7.20383883e-01 1.04062307e+00 -1.18468988e+00
-1.32250655e+00 1.22855878e+00 -1.39784683e-02 -1.98699430e-01
6.98214293e-01 -2.04979889e-02 -9.07504022e-01 -1.29758820e-01
4.00941789e-01 4.27883625e-01 8.18035066e-01 -1.32141209e+00
-7.89751947e-01 -7.22011924e-01 -6.81135207e-02 1.82825312e-01
-8.89502048e-01 -6.35801777e-02 -6.25801325e-01 -9.18100953e-01
-1.54388264e-01 -1.02472329e+00 -4.45492081e-02 -2.72555202e-01
-4.01861697e-01 -3.65615904e-01 5.14154792e-01 -7.36978292e-01
1.14004409e+00 -2.20601535e+00 -3.85123976e-02 3.00766796e-01
2.25675046e-01 6.34182513e-01 -9.44572389e-02 8.14071074e-02
-3.00390452e-01 8.32372811e-03 -2.77855694e-01 -3.23300779e-01
-5.07309102e-02 -7.06804767e-02 -1.22757182e-01 4.29878592e-01
-1.15793332e-01 8.34782720e-01 -9.70692873e-01 -4.33452636e-01
2.75482774e-01 2.10949317e-01 -3.20329994e-01 3.41302872e-01
2.70489067e-01 8.98923099e-01 -6.12135291e-01 5.95370293e-01
1.01053274e+00 -1.50995687e-01 2.05830872e-01 -1.99425608e-01
1.32640377e-01 -3.66199493e-01 -1.34680879e+00 1.68706584e+00
-3.02486181e-01 5.83456866e-02 -2.70086914e-01 -1.32692337e+00
1.16548860e+00 3.12952071e-01 4.24970806e-01 -8.20004821e-01
-7.94282649e-03 3.70448589e-01 -2.78164387e-01 -2.96040684e-01
1.40083790e-01 -2.07089588e-01 -3.53114605e-01 3.09871554e-01
6.69004628e-03 5.52748799e-01 1.16595335e-01 -3.18709165e-02
5.42778194e-01 2.60150898e-02 3.98094267e-01 -1.59685686e-01
1.08450150e+00 -9.81639028e-02 9.60140109e-01 4.80898291e-01
-5.16667783e-01 5.40979922e-01 -4.31907065e-02 -4.13528264e-01
-1.13494122e+00 -8.87780070e-01 -9.43129435e-02 9.45749700e-01
7.83339143e-01 -1.38534561e-01 -8.87010217e-01 -9.42936718e-01
1.92172304e-01 2.90788263e-01 -6.32700741e-01 -3.93669248e-01
-5.63303292e-01 -8.47088277e-01 7.49174595e-01 6.52086675e-01
1.08386791e+00 -8.45537543e-01 1.66679889e-01 7.81219602e-02
-3.90131116e-01 -1.18859923e+00 -8.54287386e-01 -5.57637453e-01
-6.50443971e-01 -1.16689777e+00 -1.39356828e+00 -1.02276492e+00
7.84292936e-01 5.63580334e-01 7.96071649e-01 -1.46156237e-01
-3.76533642e-02 3.23753059e-01 -3.14083189e-01 -8.33458751e-02
-1.43750176e-01 2.35908888e-02 5.89347899e-01 4.83107805e-01
8.32680643e-01 -4.24726963e-01 -7.71409631e-01 8.23542416e-01
-7.25713730e-01 -1.75854892e-01 5.72083831e-01 1.25542450e+00
6.42697394e-01 2.22027630e-01 9.98499751e-01 -8.62590611e-01
5.76249003e-01 -4.13256407e-01 -3.49733114e-01 5.61440229e-01
-7.54915357e-01 -2.56137967e-01 9.95682299e-01 -8.07351112e-01
-1.35229135e+00 1.01270840e-01 5.97796850e-02 -3.05473626e-01
-3.76043469e-01 1.73077434e-01 -7.22284079e-01 -9.40364301e-02
6.39451742e-01 4.60634857e-01 -9.25875902e-02 -6.22116208e-01
1.74673170e-01 9.79331017e-01 8.29594135e-01 -6.89947605e-01
1.14959788e+00 4.80146974e-01 -4.20834929e-01 -4.14999545e-01
-6.22636557e-01 -7.44210362e-01 -7.58742213e-01 1.17371753e-01
5.97203434e-01 -1.30180025e+00 -6.68541253e-01 8.49441946e-01
-8.39759648e-01 -4.25992459e-02 -3.14124018e-01 6.36553228e-01
-3.14049125e-01 6.65642440e-01 -2.65812725e-01 -5.84802151e-01
-3.40659171e-01 -8.35961223e-01 9.27165329e-01 6.38846159e-01
-4.00989987e-02 -8.89032125e-01 7.17629716e-02 5.73406518e-01
1.26122922e-01 2.08009511e-01 4.54702020e-01 -1.03806734e+00
-2.26889387e-01 -6.07983232e-01 -5.98136127e-01 5.03933012e-01
3.60088706e-01 -7.53943264e-01 -9.40285146e-01 -6.99116826e-01
-2.77933151e-01 -5.15510321e-01 6.43618107e-01 -1.28199875e-01
1.33153760e+00 -1.84111804e-01 -6.39805496e-01 6.21845841e-01
1.25430620e+00 8.00845921e-02 4.29806173e-01 4.88445133e-01
8.59789550e-01 6.75161242e-01 8.41973305e-01 5.29402196e-01
5.78292310e-01 9.22602475e-01 -1.49488002e-01 -2.42560938e-01
-4.69799601e-02 -6.18952572e-01 5.32071181e-02 4.37769771e-01
-2.78318197e-01 -1.48957595e-01 -7.68676937e-01 4.92088854e-01
-1.87979937e+00 -1.04498518e+00 3.21081072e-01 2.59885097e+00
8.19454730e-01 -2.05615789e-01 6.21582925e-01 9.12145078e-02
1.26272178e+00 -2.08166853e-01 -9.48526680e-01 4.30620641e-01
-1.35545954e-01 -1.94064349e-01 5.02938211e-01 -2.66100075e-02
-1.30994487e+00 7.93274045e-01 4.95460558e+00 1.17883205e+00
-1.03645360e+00 2.65556633e-01 4.65917498e-01 2.29019284e-01
1.04248961e-02 -3.68898749e-01 -9.18243229e-01 9.42744076e-01
6.78116977e-01 -5.19503593e-01 3.60831887e-01 9.18108881e-01
1.25714526e-01 5.10062397e-01 -1.07095706e+00 1.61508358e+00
2.15204969e-01 -7.35980690e-01 9.82564390e-02 7.48028308e-02
6.69839263e-01 -6.44501567e-01 5.10643758e-02 5.31169176e-01
1.48879513e-01 -9.01569426e-01 4.41375911e-01 2.46719748e-01
1.33682728e+00 -8.79047453e-01 9.62613165e-01 4.66990411e-01
-1.32339692e+00 -3.76325160e-01 -5.51193774e-01 3.08871865e-01
-1.23550780e-02 2.52956629e-01 -3.08851123e-01 7.28252411e-01
7.18608975e-01 8.77249718e-01 -6.82636142e-01 1.04997301e+00
7.12369531e-02 3.26226741e-01 -4.82746176e-02 3.68334472e-01
-3.06297660e-01 -1.31299376e-01 2.06348762e-01 1.08398557e+00
3.73420686e-01 9.77526605e-02 6.04695261e-01 6.73601568e-01
-1.92931935e-01 2.44504750e-01 -4.45444226e-01 1.40838102e-01
6.31566942e-01 1.13073707e+00 -3.12830120e-01 -4.14558232e-01
-5.22841394e-01 1.39603996e+00 2.96106309e-01 5.24377525e-01
-8.20999265e-01 -4.07531172e-01 5.53454876e-01 7.09446669e-02
-1.14296317e-01 2.42936000e-01 -1.05329053e-02 -1.60482955e+00
1.64272010e-01 -1.00956798e+00 7.44569421e-01 -1.66398242e-01
-1.87503576e+00 5.07803500e-01 -4.41343151e-03 -1.73695970e+00
-2.14737460e-01 -4.43648189e-01 -3.71842831e-01 1.04167509e+00
-1.45414197e+00 -1.56109619e+00 -8.02961886e-01 1.05786860e+00
2.85136133e-01 -6.60550654e-01 8.16211581e-01 8.01455200e-01
-9.40294445e-01 1.43968010e+00 3.39692116e-01 5.50842643e-01
1.24013472e+00 -1.00509691e+00 2.83361226e-01 7.70733535e-01
-3.46683562e-01 9.34527814e-01 2.73648292e-01 -6.45557582e-01
-9.48284447e-01 -1.35728347e+00 6.09347403e-01 -2.82313406e-01
2.69804150e-01 -3.79781663e-01 -9.60953355e-01 6.82327986e-01
-4.11078274e-01 1.59302905e-01 9.94985521e-01 2.65798979e-02
-6.73031509e-01 -4.66772377e-01 -1.47811842e+00 3.23134750e-01
1.29836726e+00 -5.51708579e-01 -5.38018286e-01 2.87354022e-01
5.58705866e-01 -1.80261612e-01 -1.00102234e+00 5.38771272e-01
6.60261452e-01 -6.62851393e-01 1.33202517e+00 -6.01937592e-01
-1.29419699e-01 -4.62073982e-01 7.08303154e-02 -1.27414680e+00
-5.09143114e-01 -3.57831270e-01 1.33786634e-01 1.81060433e+00
1.76307175e-03 -1.07819676e+00 9.00283694e-01 8.09099078e-01
2.99790949e-01 -1.39989242e-01 -7.77893841e-01 -1.20768404e+00
8.99183601e-02 2.44446188e-01 8.83570969e-01 1.29441822e+00
-1.98926687e-01 2.26991996e-01 -6.15444243e-01 2.01853231e-01
8.86684716e-01 1.62143379e-01 1.03574908e+00 -1.54070318e+00
-1.15763471e-01 -1.72036812e-01 -4.99877453e-01 -1.18164039e+00
4.73934501e-01 -9.60495710e-01 -2.27431372e-01 -1.05191088e+00
6.78607821e-01 -8.85977507e-01 -4.91747350e-01 3.30567181e-01
-5.56713700e-01 4.22739118e-01 8.78055319e-02 6.62032247e-01
-6.05876267e-01 6.97714210e-01 1.23309112e+00 -4.14562076e-01
-1.79909393e-01 2.42389470e-01 -9.81343687e-01 5.59882164e-01
7.51070619e-01 -1.83628187e-01 -4.41400319e-01 -1.13081023e-01
-5.18696070e-01 -2.07080707e-01 6.27196789e-01 -1.11709905e+00
2.52079546e-01 -1.16033815e-01 7.02761471e-01 -5.68727292e-02
1.53640300e-01 -8.80520165e-01 3.82253341e-02 2.77016282e-01
-2.63604492e-01 -3.56131583e-01 4.44853082e-02 7.90318310e-01
-3.32815260e-01 -3.25269923e-02 9.17462885e-01 -9.73320976e-02
-1.05721784e+00 8.15705776e-01 5.30182958e-01 3.10316563e-01
1.21861589e+00 -4.52631563e-01 -2.10168272e-01 -1.47865146e-01
-5.35245597e-01 3.72273833e-01 7.20140517e-01 5.63911140e-01
3.76175642e-01 -1.71015203e+00 -8.93954575e-01 3.63685936e-01
6.29985809e-01 -1.30382955e-01 6.32895529e-01 3.46062243e-01
-6.94975117e-03 3.69856238e-01 -4.47154284e-01 -5.12581289e-01
-1.13578212e+00 8.78658295e-01 3.14924866e-01 -1.95242345e-01
-4.89718795e-01 5.81382632e-01 4.58262920e-01 -8.16710830e-01
1.36715978e-01 3.84107053e-01 -3.65011126e-01 -8.33585113e-02
8.91796827e-01 3.59423369e-01 -3.19520861e-01 -9.48485255e-01
-4.40955698e-01 7.92000771e-01 -1.46483466e-01 1.43818006e-01
7.41032243e-01 -3.53541732e-01 1.23732686e-01 -5.78083424e-03
1.06058049e+00 -2.59486176e-02 -1.26215398e+00 -6.46358550e-01
-2.06070647e-01 -6.96132243e-01 -6.31758690e-01 -7.21176744e-01
-9.80708659e-01 7.27495730e-01 9.91123497e-01 -5.97541519e-02
1.20620430e+00 -1.87321767e-01 1.17632616e+00 2.06841379e-01
5.08390129e-01 -1.13755178e+00 1.14984781e-01 9.81476679e-02
5.61677277e-01 -1.63165879e+00 -1.19251147e-01 -4.83688831e-01
-8.14508796e-01 6.97576761e-01 1.05246186e+00 2.70568524e-02
3.03133249e-01 -4.54319388e-01 -3.99316009e-03 3.59819263e-01
2.83287764e-01 -1.96475372e-01 3.37235361e-01 1.01044118e+00
7.59398043e-02 3.53598595e-01 -3.42316210e-01 1.19390941e+00
-6.54876977e-02 1.89807057e-01 -9.21033397e-02 2.31966913e-01
-8.90600979e-02 -1.40153289e+00 -4.99738067e-01 2.17612341e-01
-3.26198816e-01 2.90165812e-01 -2.28252470e-01 8.34788084e-01
5.35343945e-01 9.37268853e-01 -1.33081317e-01 -6.12112284e-01
5.31254113e-01 -1.03899986e-01 2.79320836e-01 -4.91303831e-01
-2.94590294e-01 -4.31449771e-01 -2.08668485e-01 -3.06688160e-01
-5.05720079e-01 -5.16877174e-01 -9.76821959e-01 -3.55199307e-01
-2.41918102e-01 2.33392075e-01 1.28681093e-01 8.09293747e-01
3.83819401e-01 1.20515987e-01 8.15915465e-01 -6.72252417e-01
-8.28899920e-01 -8.80867600e-01 -6.90205336e-01 1.11108768e+00
1.37300715e-01 -8.76340270e-01 -1.15864195e-01 8.04775432e-02] | [14.779193878173828, 1.1138030290603638] |
c1a2c32b-beef-4915-856c-f38efcea2032 | investigating-the-reordering-capability-in | 2105.04840 | null | https://arxiv.org/abs/2105.04840v1 | https://arxiv.org/pdf/2105.04840v1.pdf | Investigating the Reordering Capability in CTC-based Non-Autoregressive End-to-End Speech Translation | We study the possibilities of building a non-autoregressive speech-to-text translation model using connectionist temporal classification (CTC), and use CTC-based automatic speech recognition as an auxiliary task to improve the performance. CTC's success on translation is counter-intuitive due to its monotonicity assumption, so we analyze its reordering capability. Kendall's tau distance is introduced as the quantitative metric, and gradient-based visualization provides an intuitive way to take a closer look into the model. Our analysis shows that transformer encoders have the ability to change the word order and points out the future research direction that worth being explored more on non-autoregressive speech translation. | ['Hung-Yi Lee', 'Chih-Chiang Chang', 'Yung-Sung Chuang', 'Shun-Po Chuang'] | 2021-05-11 | null | https://aclanthology.org/2021.findings-acl.92 | https://aclanthology.org/2021.findings-acl.92.pdf | findings-acl-2021-8 | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 1.73035994e-01 1.52682126e-01 -5.02422810e-01 -4.56025898e-01
-6.71053410e-01 -5.86044431e-01 9.78789032e-01 -2.92518228e-01
-1.75335094e-01 5.60921609e-01 6.78733945e-01 -1.19211078e+00
8.43361989e-02 -4.36226726e-01 -4.29919988e-01 -4.93221849e-01
-1.59939155e-01 4.76113379e-01 1.28242835e-01 -5.68361163e-01
1.62052915e-01 4.33895022e-01 -9.65998828e-01 4.47138101e-01
7.51076996e-01 4.49019164e-01 1.22213773e-01 6.34398282e-01
-2.15268955e-01 9.06737924e-01 -6.14326000e-01 -6.74695969e-01
9.00784656e-02 -8.27783942e-01 -8.29700589e-01 -2.33268738e-01
-1.42257899e-01 -9.12761465e-02 -1.44020677e-01 7.57785320e-01
2.81016469e-01 6.40528649e-02 7.00999379e-01 -1.00913656e+00
-9.98819232e-01 8.76743615e-01 -4.68788072e-02 6.47292078e-01
4.94671315e-01 -1.12493953e-03 1.13070071e+00 -8.58926535e-01
5.65331519e-01 1.35722578e+00 5.68589747e-01 4.60536301e-01
-1.09873176e+00 -5.16627371e-01 2.08939761e-01 4.71439302e-01
-1.08462524e+00 -5.09406686e-01 7.90417492e-01 -3.58106703e-01
1.51746130e+00 7.35129058e-01 8.06845903e-01 1.46357095e+00
5.24688601e-01 7.84136117e-01 1.46621704e+00 -7.48133063e-01
2.52848975e-02 3.16116571e-01 -8.67922306e-02 4.29893225e-01
-3.88734996e-01 4.25625682e-01 -6.33216739e-01 1.26357943e-01
6.55943930e-01 -1.85677484e-01 1.35226306e-02 -1.54493945e-02
-1.15102565e+00 8.83707285e-01 2.41875485e-01 7.99902737e-01
-3.12398851e-01 6.43968433e-02 3.94565433e-01 9.08806205e-01
8.96460474e-01 6.58994555e-01 -6.94123983e-01 -8.67676675e-01
-1.01123083e+00 -4.42033738e-01 5.50171137e-01 6.36682987e-01
2.19317570e-01 3.78235936e-01 -2.19209164e-01 7.31521726e-01
2.29554087e-01 3.55364859e-01 9.60854471e-01 -4.79545981e-01
4.51372623e-01 3.64050418e-01 -3.05886894e-01 -6.28392041e-01
-1.52934745e-01 -6.92321479e-01 -6.25377059e-01 -2.05128700e-01
2.42211342e-01 3.44410427e-02 -9.46296692e-01 1.37250364e+00
-3.35514516e-01 -2.46571135e-02 -2.93252561e-02 9.52245355e-01
1.85291439e-01 7.32699335e-01 -9.00394171e-02 -6.33626699e-01
1.22150135e+00 -1.13001275e+00 -1.15069008e+00 2.22018100e-02
8.72826755e-01 -7.51878381e-01 1.45741427e+00 1.89209044e-01
-8.58881474e-01 -1.61596239e-01 -8.64647508e-01 -6.72840178e-02
-5.93376756e-01 3.22958499e-01 7.95812607e-01 8.61167371e-01
-1.21407890e+00 5.63271701e-01 -1.35556376e+00 -5.94339430e-01
-2.65443958e-02 2.99183816e-01 -9.97014344e-02 3.63082975e-01
-1.52848458e+00 1.46469760e+00 -1.45703703e-01 1.15335703e-01
-6.02753878e-01 -2.70086467e-01 -5.89855850e-01 1.21321470e-01
4.54998314e-02 -6.24101520e-01 1.06929028e+00 -1.24195051e+00
-2.24748182e+00 5.81048131e-01 -5.01743376e-01 -7.03190565e-01
6.18521035e-01 1.10231169e-01 -6.10559523e-01 -7.15284944e-02
-2.37696081e-01 2.04374969e-01 7.93752909e-01 -8.46859992e-01
-3.49070162e-01 -3.55890036e-01 -1.47256717e-01 1.14245415e-01
-5.89957893e-01 4.94747102e-01 4.84479405e-02 -9.82080102e-01
2.13804185e-01 -1.08546782e+00 2.45825537e-02 -6.34725571e-01
-2.54533708e-01 -2.18374103e-01 5.89039922e-01 -9.76189792e-01
1.70139527e+00 -1.80603433e+00 3.50705594e-01 1.16586529e-01
-2.03099698e-01 5.83010167e-02 -1.82402834e-01 8.27509284e-01
-3.35998774e-01 4.55916166e-01 -8.53909850e-02 -3.06615382e-01
-2.24148914e-01 1.31728962e-01 -5.20588219e-01 2.34094530e-01
3.85959148e-01 1.13807011e+00 -7.58361459e-01 -1.62856966e-01
1.72939509e-01 5.16187191e-01 -1.14107348e-01 -2.39428803e-01
2.55434867e-02 4.20894980e-01 -1.98734030e-01 3.98957819e-01
6.65100291e-02 -5.22805527e-02 2.70405531e-01 1.75390378e-01
-2.24306867e-01 1.07659602e+00 -3.22978377e-01 1.35100877e+00
-6.11727893e-01 1.16066813e+00 -4.75125939e-01 -1.00466633e+00
1.05474377e+00 3.58360440e-01 1.32883668e-01 -1.00678742e+00
-3.61872953e-04 2.59235352e-01 3.05262446e-01 -3.25802118e-01
4.80871469e-01 -3.88770163e-01 3.33525270e-01 5.57845354e-01
-2.75843859e-01 -8.28337949e-03 -1.51578903e-01 -1.10358587e-02
1.06471276e+00 2.37842306e-01 1.52603418e-01 -1.95836037e-01
4.04981107e-01 1.86928496e-01 1.21707611e-01 3.67524654e-01
1.28185958e-01 4.03714389e-01 3.68057430e-01 -3.67558986e-01
-1.08051968e+00 -8.07543457e-01 9.78525281e-02 1.25638759e+00
-4.31501448e-01 -5.61926365e-01 -6.73565686e-01 -7.20484018e-01
-4.70600635e-01 1.10376203e+00 -6.99260592e-01 -2.47042820e-01
-5.72797239e-01 -6.17938340e-01 5.73732734e-01 4.76899087e-01
1.03272311e-01 -7.61432767e-01 -3.59638989e-01 7.63923675e-02
-4.82063979e-01 -7.45823920e-01 -4.52180624e-01 4.82681602e-01
-1.31417513e+00 -3.80958796e-01 -8.71621728e-01 -7.25945055e-01
4.59243834e-01 1.21481955e-01 7.64022052e-01 -1.89458728e-01
4.12515640e-01 2.41321400e-01 -5.55898905e-01 -2.06445307e-01
-7.80987561e-01 3.76479089e-01 1.89165726e-01 -2.49791488e-01
4.50049311e-01 -5.90839863e-01 -2.28735223e-01 4.56996232e-01
-5.93485534e-01 8.15997496e-02 5.61772227e-01 1.01218545e+00
9.87558365e-02 -1.61474451e-01 2.95545906e-01 -5.59225559e-01
1.14406490e+00 -1.55208632e-01 -2.87419379e-01 4.13854212e-01
-9.79014456e-01 2.55531520e-01 6.40555859e-01 -6.96967125e-01
-9.19321239e-01 -1.66128606e-01 5.28842136e-02 -5.48526227e-01
4.43995982e-01 8.37246597e-01 3.81333232e-01 4.16219324e-01
5.39160311e-01 5.26629508e-01 6.13667816e-02 -4.70511526e-01
3.91390830e-01 8.66147161e-01 1.11481771e-01 -1.58380389e-01
7.41929531e-01 1.68636829e-01 -3.25274229e-01 -7.94876873e-01
-1.74402595e-01 -1.67819008e-01 -6.93462193e-01 -8.49871114e-02
7.01118886e-01 -6.91889226e-01 -5.93603015e-01 -7.13374540e-02
-1.30657423e+00 -5.67281842e-01 -4.70206924e-02 7.93495834e-01
-5.99552155e-01 1.86356544e-01 -8.91645432e-01 -1.26112628e+00
-2.41942912e-01 -1.05311000e+00 9.60845053e-01 -5.00775397e-01
-5.80886185e-01 -1.26760077e+00 -3.98616567e-02 5.90265274e-01
6.10902309e-01 -5.24700403e-01 9.18088853e-01 -9.25675869e-01
-3.62349868e-01 -2.87621394e-02 1.51375070e-01 3.53896827e-01
2.62231499e-01 1.31575435e-01 -8.27979386e-01 -1.99743226e-01
2.95980513e-01 4.66307342e-01 6.75202191e-01 3.53782266e-01
6.67252660e-01 -8.06114614e-01 -2.28630424e-01 1.35454178e-01
8.50898981e-01 5.43736458e-01 9.37195301e-01 5.50624371e-01
4.15324837e-01 5.92704117e-01 6.10271871e-01 5.06611168e-02
2.91519731e-01 9.30856347e-01 1.32630631e-01 -1.34654403e-01
-1.30003244e-01 -5.54592133e-01 9.97366965e-01 1.64400458e+00
-2.00068608e-01 -3.41893911e-01 -9.93508756e-01 1.73673242e-01
-1.71261179e+00 -8.93276334e-01 -2.82399684e-01 2.09697223e+00
7.46008039e-01 3.56455952e-01 2.09885225e-01 2.29733884e-01
5.02491474e-01 -1.19713478e-01 8.14642608e-02 -1.02606177e+00
-6.33809566e-02 1.21685378e-01 4.14209366e-01 7.94195414e-01
-5.57653666e-01 1.31407940e+00 7.21178055e+00 8.80624771e-01
-1.54641044e+00 3.05810094e-01 6.71781182e-01 7.01579079e-02
-5.25184691e-01 2.45228097e-01 -1.79083154e-01 6.56234026e-01
1.49561822e+00 -2.28521749e-01 6.95973754e-01 4.71791208e-01
6.98157609e-01 6.14468567e-02 -1.11080778e+00 7.45988429e-01
7.78625011e-02 -1.25493598e+00 1.31836772e-01 9.66736004e-02
3.43658388e-01 1.74264684e-02 2.36030832e-01 3.33754867e-01
2.17854932e-01 -9.40906405e-01 6.91827714e-01 5.31222820e-01
7.59900570e-01 -4.70390826e-01 7.01913655e-01 3.39228123e-01
-8.57286811e-01 -8.95392299e-02 2.25746185e-02 -5.07896245e-01
-5.64165339e-02 2.50692070e-01 -1.23062134e+00 4.21432674e-01
4.13835853e-01 6.71646357e-01 -6.54435575e-01 4.02045339e-01
-2.16803834e-01 1.14387465e+00 -1.77150831e-01 -4.07235593e-01
2.86694854e-01 -4.82615530e-01 6.44701481e-01 1.31032062e+00
4.42964822e-01 -1.17416643e-01 -3.27758074e-01 5.50262690e-01
2.67538816e-01 3.23455900e-01 -7.09925592e-01 -4.26703990e-01
3.06888223e-01 6.44260287e-01 -9.66009915e-01 -3.96045744e-01
-2.51636952e-01 1.40490699e+00 -1.01043671e-01 4.49626356e-01
-6.93828344e-01 -4.24763858e-02 4.35649037e-01 2.51570255e-01
5.03435619e-02 -5.55357397e-01 -5.22399664e-01 -1.16516244e+00
1.20324023e-01 -1.05515647e+00 4.80396189e-02 -1.13512146e+00
-8.00586998e-01 6.80041909e-01 -2.70480901e-01 -1.30637538e+00
-5.95854282e-01 -3.81111085e-01 -4.16043818e-01 8.42939377e-01
-1.19479716e+00 -1.12804389e+00 5.11151671e-01 2.89649934e-01
8.23483944e-01 -2.74584025e-01 1.09154570e+00 1.97917685e-01
-5.12821436e-01 7.91687965e-01 1.09218091e-01 4.59183902e-02
5.16625822e-01 -9.91872013e-01 5.08834839e-01 8.06841195e-01
5.39661407e-01 7.96514630e-01 1.01668942e+00 -6.51669681e-01
-1.62772000e+00 -6.34186447e-01 1.56675792e+00 -8.10132205e-01
1.04549170e+00 -4.40040201e-01 -7.59478092e-01 8.56130242e-01
4.26348925e-01 -6.19133472e-01 4.98918980e-01 2.62292475e-01
-1.86801523e-01 -1.04175292e-01 -6.48817539e-01 9.82835650e-01
1.10858858e+00 -9.71458972e-01 -6.49915755e-01 4.34608132e-01
1.02333760e+00 -3.23448256e-02 -8.89983058e-01 1.41734794e-01
4.65894163e-01 -6.87668622e-01 5.66763580e-01 -8.29785287e-01
4.30485696e-01 9.08784568e-02 -9.15008262e-02 -1.58292103e+00
-2.27740645e-01 -1.08216393e+00 2.27333412e-01 1.12242091e+00
9.82126474e-01 -1.02382290e+00 6.07694268e-01 2.90619612e-01
-1.35566205e-01 -7.51330316e-01 -1.12101543e+00 -1.18513668e+00
1.75448507e-01 -7.13716149e-01 3.94024879e-01 1.17887568e+00
6.25874639e-01 7.61972308e-01 -4.67360914e-01 -1.71209842e-01
-3.76705408e-01 -3.07233989e-01 3.33337009e-01 -8.28605473e-01
-1.15911424e-01 -6.32397175e-01 -4.71931219e-01 -1.23896122e+00
1.17655642e-01 -9.31178272e-01 -3.44163001e-01 -1.36677837e+00
-6.41769916e-02 -8.93292353e-02 -1.13109425e-01 4.81326997e-01
3.64614785e-01 4.67994511e-02 1.62260070e-01 4.45318937e-01
-5.36162369e-02 6.71133697e-01 1.24606478e+00 -3.34164679e-01
-4.11614537e-01 2.52440095e-01 -3.61754179e-01 1.24793701e-01
9.49066639e-01 -4.78932261e-01 -5.14492035e-01 -4.39232826e-01
3.89487237e-01 3.94221127e-01 -1.12080552e-01 -3.89459729e-01
8.68382752e-02 -1.07198812e-01 -3.25639546e-03 -3.06758344e-01
5.64242661e-01 -9.70765710e-01 1.96037576e-01 6.17687941e-01
-7.13491201e-01 6.99344158e-01 5.90276867e-02 3.49242777e-01
-3.39382797e-01 2.08349779e-01 2.12934464e-01 6.71214759e-02
-3.57692949e-02 -2.31918246e-01 -9.32167530e-01 -4.77219015e-01
5.70571721e-01 -2.68854082e-01 -1.09340310e-01 -7.91823804e-01
-9.42675650e-01 -7.08915219e-02 2.43470475e-01 8.98614168e-01
4.33659643e-01 -1.28268278e+00 -5.75316906e-01 1.51290774e-01
-5.83482757e-02 -1.25246704e+00 -4.76353407e-01 1.29092228e+00
-1.98746443e-01 9.40767884e-01 8.76592323e-02 -6.41834795e-01
-1.39925241e+00 5.65711319e-01 3.21906924e-01 -3.74778181e-01
-5.10509014e-01 3.96971881e-01 -4.18292254e-01 -3.47430557e-01
8.45316127e-02 -6.39684916e-01 -1.82251871e-01 1.94966003e-01
-3.22394036e-02 4.38502908e-01 3.31829101e-01 -4.81779277e-01
-5.75191259e-01 2.37033486e-01 -1.33209042e-02 -7.47457862e-01
1.21489263e+00 -4.78063583e-01 -1.83613911e-01 9.15518820e-01
1.13780951e+00 -1.04188301e-01 -5.94801188e-01 -3.06190886e-02
4.21575040e-01 -1.89480066e-01 1.30355442e-02 -1.04214525e+00
-6.83837712e-01 1.14312029e+00 6.88682199e-01 5.70512116e-01
9.86539960e-01 -1.65824354e-01 3.54199469e-01 4.17025506e-01
2.26331323e-01 -1.19364548e+00 -4.70001772e-02 7.98585176e-01
1.01564717e+00 -1.02805996e+00 -2.00791165e-01 -3.83882195e-01
-8.79972875e-01 1.33127725e+00 3.99973057e-02 3.23396534e-01
5.48667133e-01 3.87120306e-01 3.67794365e-01 8.60769674e-02
-1.13507855e+00 -1.07640497e-01 1.55971736e-01 4.15937543e-01
8.83356452e-01 2.79502749e-01 -7.12040246e-01 2.48717532e-01
-6.88287973e-01 -1.64188832e-01 3.60665888e-01 5.67498684e-01
6.13107532e-02 -1.17598295e+00 -2.37766832e-01 3.51811796e-01
-4.01346326e-01 -4.88065273e-01 -7.85118043e-01 5.56232095e-01
-6.36486948e-01 1.31817091e+00 2.26746976e-01 -7.21495271e-01
8.64893571e-02 5.00373363e-01 3.39238733e-01 -4.38087404e-01
-7.46949792e-01 2.67807037e-01 3.14387202e-01 -4.48775828e-01
-2.25712478e-01 -6.22369647e-01 -9.72304523e-01 -3.05969626e-01
-4.41920042e-01 3.47178310e-01 1.17288530e+00 9.58363950e-01
4.36338186e-01 5.79451203e-01 8.89359295e-01 -4.23874795e-01
-5.49141407e-01 -1.32423401e+00 9.88997146e-02 -1.66783761e-02
3.73002946e-01 -3.55446339e-01 -5.51118612e-01 1.48931528e-02] | [14.421964645385742, 7.203676223754883] |
c787d2ee-5be6-4063-a463-23d3fd17c058 | re-benchmarking-pool-based-active-learning | 2306.08954 | null | https://arxiv.org/abs/2306.08954v1 | https://arxiv.org/pdf/2306.08954v1.pdf | Re-Benchmarking Pool-Based Active Learning for Binary Classification | Active learning is a paradigm that significantly enhances the performance of machine learning models when acquiring labeled data is expensive. While several benchmarks exist for evaluating active learning strategies, their findings exhibit some misalignment. This discrepancy motivates us to develop a transparent and reproducible benchmark for the community. Our efforts result in an open-sourced implementation (https://github.com/ariapoy/active-learning-benchmark) that is reliable and extensible for future research. By conducting thorough re-benchmarking experiments, we have not only rectified misconfigurations in existing benchmark but also shed light on the under-explored issue of model compatibility, which directly causes the observed discrepancy. Resolving the discrepancy reassures that the uncertainty sampling strategy of active learning remains an effective and preferred choice for most datasets. Our experience highlights the importance of dedicating research efforts towards re-benchmarking existing benchmarks to produce more credible results and gain deeper insights. | ['Hsuan-Tien Lin', 'Chun-Liang Li', 'Po-Yi Lu'] | 2023-06-15 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 3.53326276e-02 4.20990795e-01 -7.54832685e-01 -5.79504251e-01
-1.38334823e+00 -7.53372252e-01 6.49117053e-01 3.54980230e-01
-5.53922296e-01 8.61559331e-01 3.10005903e-01 -5.44944882e-01
-2.14624226e-01 -3.51984382e-01 -6.80490017e-01 -5.26021421e-01
2.15208933e-01 4.52491164e-01 9.42712128e-02 3.98220271e-01
6.09498739e-01 2.88437277e-01 -1.60813665e+00 1.32118374e-01
9.03860748e-01 5.17429888e-01 -2.77554214e-01 2.76744574e-01
-1.74900562e-01 1.29584169e+00 -3.97904098e-01 -6.06232941e-01
4.30255413e-01 -2.45414212e-01 -1.23304760e+00 -8.29294100e-02
4.25124705e-01 -1.84621453e-01 2.31409878e-01 6.37772739e-01
4.73158628e-01 1.93701372e-01 3.23539376e-01 -1.54504025e+00
-4.75583583e-01 7.42646337e-01 -3.20041180e-01 4.31988716e-01
1.35458231e-01 5.08541107e-01 1.21433997e+00 -8.07496071e-01
7.03922868e-01 7.47079015e-01 7.70225704e-01 6.31643534e-01
-1.26824737e+00 -8.88587117e-01 3.90762895e-01 4.22242105e-01
-1.25628161e+00 -1.07036054e+00 6.74990833e-01 -4.76743460e-01
9.14647877e-01 6.32210851e-01 6.81644797e-01 1.34636974e+00
-2.03416616e-01 1.08193231e+00 1.22723603e+00 -6.44077003e-01
4.21483696e-01 4.85109448e-01 3.83731544e-01 4.80619550e-01
5.55079639e-01 2.78095454e-01 -7.06387162e-01 -5.59460521e-01
2.42953405e-01 2.09428892e-02 -7.09057227e-02 -6.18523061e-01
-9.21097040e-01 8.46280396e-01 -2.81650349e-02 7.33353347e-02
-2.88938046e-01 -1.28398821e-01 2.15696931e-01 7.78859854e-02
8.87638152e-01 8.29720795e-01 -6.71581864e-01 -8.45603824e-01
-1.10391593e+00 2.32808977e-01 8.18041146e-01 7.48089433e-01
6.46457911e-01 -2.11534455e-01 1.81143820e-01 8.45531821e-01
6.52632892e-01 -2.65726715e-01 4.55004685e-02 -1.35647166e+00
2.56245971e-01 9.12757099e-01 2.74105281e-01 -6.76070988e-01
-5.17909825e-02 -5.21340847e-01 2.22109631e-02 3.86736602e-01
5.35743535e-01 -1.46087095e-01 -5.44320107e-01 1.37504494e+00
2.05542952e-01 1.75296560e-01 -2.52625525e-01 5.67806900e-01
6.95503592e-01 1.75523669e-01 4.94973958e-01 -3.20260435e-01
5.43422639e-01 -1.11520028e+00 -4.97853309e-01 -4.68838692e-01
7.58831024e-01 -8.76110077e-01 1.13112724e+00 5.14296293e-01
-1.29953372e+00 -1.93476439e-01 -1.10727561e+00 9.43711624e-02
-3.23906422e-01 -4.25360799e-01 9.90488827e-01 7.03597605e-01
-8.03637326e-01 4.56396431e-01 -1.32265592e+00 -3.66186708e-01
5.84475458e-01 1.47108138e-01 -2.48450503e-01 1.09775797e-01
-8.45068693e-01 1.05394292e+00 3.94864231e-01 4.23312709e-02
-6.98854506e-01 -1.16106570e+00 -5.69867432e-01 -2.56554127e-01
6.25651956e-01 -2.87302583e-01 1.58248162e+00 -1.09932256e+00
-1.16041291e+00 9.14147913e-01 -2.30450649e-02 -3.50915223e-01
7.95670450e-01 -3.00173998e-01 -4.10051048e-01 -4.66831505e-01
-1.89107701e-01 5.81730068e-01 1.56843051e-01 -1.35164738e+00
-6.94172025e-01 -2.42211848e-01 -5.08842804e-02 2.61205107e-01
-3.94691586e-01 3.09892476e-01 -3.21050853e-01 -3.58999133e-01
1.28124980e-02 -8.44414115e-01 -1.81317076e-01 -1.06536016e-01
-1.35468453e-01 -2.99722195e-01 6.52486622e-01 -2.76407123e-01
1.46821201e+00 -2.04252625e+00 -3.07433814e-01 -2.53254950e-01
2.78021783e-01 2.32585207e-01 8.37261900e-02 7.54920900e-01
-3.94011848e-02 5.17659843e-01 -2.22593725e-01 -5.92207491e-01
-8.14163089e-02 4.58670706e-02 -3.25251311e-01 3.74652326e-01
3.97894979e-01 9.14565563e-01 -9.31222081e-01 -7.07252324e-01
2.08525509e-01 3.28339189e-01 -4.39828783e-01 2.87683368e-01
-2.47287005e-01 3.02499980e-01 -1.91855371e-01 7.05575705e-01
5.90984285e-01 -6.47273719e-01 1.06417894e-01 2.19844609e-01
-4.85001415e-01 7.21159577e-01 -1.24895787e+00 1.57576227e+00
-1.29197374e-01 6.55489802e-01 -1.65008485e-01 -7.24936426e-01
8.39202881e-01 2.02432707e-01 7.05266178e-01 -5.07386863e-01
-1.62737235e-01 2.08890498e-01 9.80855972e-02 -4.87686604e-01
4.34453666e-01 2.37605378e-01 4.38318044e-01 7.18313396e-01
2.36209240e-02 1.68162808e-01 1.40058756e-01 1.06970116e-01
1.05065322e+00 4.08205479e-01 3.60477746e-01 -3.16107512e-01
-1.12133391e-01 4.18819457e-01 7.17037976e-01 8.51945102e-01
-6.10050857e-01 6.83407247e-01 2.91639090e-01 -3.05964470e-01
-7.06959724e-01 -9.33276534e-01 -3.26203078e-01 1.24787533e+00
-3.38618308e-01 -5.03411770e-01 -3.15159678e-01 -9.92395639e-01
-7.29650483e-02 1.12514842e+00 -6.73432946e-01 1.03304282e-01
-2.96588808e-01 -1.11361837e+00 4.67665583e-01 5.68794966e-01
2.46004656e-01 -8.31775546e-01 -6.64291918e-01 1.22112900e-01
-2.21143305e-01 -5.39449513e-01 5.06425798e-02 4.63686854e-01
-1.06944442e+00 -1.34317791e+00 -3.06865931e-01 -3.34211588e-01
4.14975733e-01 1.98397562e-01 1.43850887e+00 1.74740449e-01
-3.71295935e-03 6.05493188e-01 -4.19344395e-01 -9.55634356e-01
-4.46290791e-01 4.49721128e-01 -4.98377800e-01 -5.39847851e-01
8.83277833e-01 -1.60536185e-01 -7.45559752e-01 3.17772031e-01
-6.92846775e-01 1.67481840e-01 3.05161744e-01 5.73149264e-01
4.13717479e-01 -3.22504342e-01 7.33278453e-01 -1.40629458e+00
5.51150799e-01 -7.72837698e-01 -6.85929537e-01 3.59754115e-01
-1.49283600e+00 -1.61487684e-01 -1.46527467e-02 -1.09238222e-01
-1.30821896e+00 -1.13375343e-01 -1.19151786e-01 1.22014321e-01
-2.05433473e-01 5.19881070e-01 2.71460950e-01 9.08076093e-02
9.99827445e-01 -3.99094343e-01 -3.39340307e-02 -3.65522444e-01
-1.57614704e-02 7.47187376e-01 1.73539612e-02 -4.82979655e-01
5.35070717e-01 3.54322642e-01 -7.50602782e-01 -5.09564817e-01
-8.25314820e-01 -6.54313266e-01 -4.13796484e-01 -1.40032619e-01
2.34794796e-01 -9.35054719e-01 -2.35360041e-01 2.56032884e-01
-7.63925731e-01 -6.80085897e-01 -3.84781688e-01 6.22033834e-01
-2.17751473e-01 2.00957358e-01 -3.89439970e-01 -9.73229289e-01
-2.64100343e-01 -1.11367965e+00 6.07238054e-01 3.42459738e-01
-8.16039085e-01 -1.34602213e+00 7.32159793e-01 1.10015404e+00
5.60815454e-01 9.02093425e-02 4.89871114e-01 -1.03607011e+00
-7.98907101e-01 -8.99583697e-02 1.66068852e-01 1.83551580e-01
1.18630260e-01 7.79870093e-01 -1.47839749e+00 -2.48163462e-01
-1.91801369e-01 -6.54245436e-01 7.07760215e-01 1.33816794e-01
1.15044284e+00 -1.15256518e-01 -1.43365502e-01 1.77113444e-01
1.20935357e+00 2.78871834e-01 5.69796622e-01 6.77236319e-01
1.40560418e-01 5.63456714e-01 8.43539417e-01 2.99355984e-01
6.16502464e-01 4.33970869e-01 2.20545828e-01 -3.57683748e-01
5.44508919e-02 -4.52614427e-02 2.04636753e-01 7.08754063e-01
-4.19280417e-02 -9.78163034e-02 -1.52107096e+00 4.70926672e-01
-1.96350265e+00 -1.01135135e+00 -8.20394531e-02 2.46310019e+00
1.10705602e+00 4.33598161e-01 1.11469045e-01 1.46175250e-01
2.45238751e-01 1.88440353e-01 -6.78690434e-01 -2.60524720e-01
3.87495011e-02 2.12667435e-01 2.20256418e-01 4.90658164e-01
-9.52661693e-01 6.24213576e-01 7.03599596e+00 5.41487694e-01
-1.20427871e+00 2.92655140e-01 7.81772971e-01 -3.48106921e-01
-4.20234680e-01 3.99472028e-01 -6.67304516e-01 4.09806341e-01
1.15653110e+00 -3.98882091e-01 1.96049005e-01 9.93610740e-01
2.33678266e-01 -3.47324103e-01 -1.27639329e+00 5.08189261e-01
-6.44070357e-02 -1.59228480e+00 -6.09583318e-01 3.38145532e-02
7.09874153e-01 5.65467775e-01 -6.63669920e-03 4.26896185e-01
5.36590993e-01 -9.04091716e-01 7.30020046e-01 3.92691553e-01
3.30380261e-01 -4.55460161e-01 5.61761856e-01 4.11839813e-01
-5.77338219e-01 -9.97753218e-02 -1.50815859e-01 -3.70333314e-01
-1.54291272e-01 5.14330089e-01 -1.17639422e+00 3.54351580e-01
6.40952647e-01 7.05253184e-01 -1.02338409e+00 1.35260367e+00
1.69705540e-01 1.33395028e+00 1.30480360e-02 2.60158598e-01
-1.13101982e-01 2.52721868e-02 2.69914061e-01 1.09067571e+00
-2.44328007e-01 -1.35489523e-01 2.08956506e-02 9.38481390e-01
-1.32604502e-02 1.76346853e-01 -5.77704966e-01 -6.21738195e-01
8.41984272e-01 1.29143667e+00 -8.66525710e-01 -1.48052856e-01
-7.04887331e-01 2.67368168e-01 5.90472758e-01 3.67088765e-01
-6.83879137e-01 1.79055452e-01 4.43970054e-01 1.41185731e-01
-3.91046196e-01 -1.30986646e-01 -7.69065976e-01 -1.04107916e+00
-7.20547214e-02 -1.12468314e+00 7.36139596e-01 -6.16892278e-01
-1.26070702e+00 1.76752284e-01 2.31352031e-01 -1.02462816e+00
-3.29993486e-01 -1.70021579e-01 -5.15520871e-01 6.81659043e-01
-1.29969513e+00 -9.01405394e-01 -4.57143784e-01 9.60937515e-02
7.06512034e-01 6.48089945e-02 9.60430920e-01 2.60627210e-01
-9.59716976e-01 6.77590609e-01 -6.64104745e-02 -1.86688989e-01
1.04559660e+00 -1.16170394e+00 5.00730276e-01 6.30165577e-01
3.80256444e-01 9.77599144e-01 6.00689590e-01 -6.52578652e-01
-1.14710343e+00 -7.43377209e-01 8.05712998e-01 -1.13733029e+00
6.23659790e-01 -3.85415524e-01 -9.98958051e-01 1.13596892e+00
4.44584399e-01 -2.48477578e-01 1.29763329e+00 6.88924372e-01
-2.56190479e-01 1.04087368e-01 -9.74393070e-01 3.38002414e-01
9.50103343e-01 -4.72170681e-01 -5.55790544e-01 2.33783215e-01
4.86464202e-01 -2.89059877e-01 -8.87759924e-01 6.31223023e-01
5.76222479e-01 -1.25423801e+00 5.02595782e-01 -7.55112112e-01
4.08214986e-01 2.28580996e-01 1.92717239e-01 -9.72437918e-01
-1.64004922e-01 -5.32474697e-01 -1.48420721e-01 1.55398250e+00
9.63637352e-01 -8.57230067e-01 1.07409573e+00 1.20698166e+00
-1.11641504e-01 -1.05963922e+00 -4.35633659e-01 -5.10009348e-01
2.87811249e-01 -7.42113948e-01 3.77236933e-01 1.45521235e+00
3.21744382e-02 -4.46844362e-02 -6.85231248e-03 -1.88941792e-01
5.59142470e-01 -2.19355121e-01 7.38881528e-01 -1.47163129e+00
-6.94470406e-02 -2.61490017e-01 7.86968973e-03 -4.09419864e-01
1.07070521e-01 -6.55896783e-01 -2.57708102e-01 -1.44557548e+00
6.61341369e-01 -8.14035714e-01 -5.97274244e-01 8.11458588e-01
-2.28881165e-01 2.02368155e-01 3.69175412e-02 4.08773363e-01
-7.16959000e-01 1.75099805e-01 7.95735359e-01 7.38645419e-02
-3.25283438e-01 7.99815580e-02 -1.12999523e+00 6.39484763e-01
1.01204538e+00 -5.93958259e-01 -6.72863603e-01 -4.51577902e-01
3.95410150e-01 -5.00186265e-01 2.27276683e-01 -7.68743277e-01
3.05586606e-01 -6.30886078e-01 2.99146920e-01 -5.18677831e-01
1.83536395e-01 -7.57246375e-01 4.50446755e-01 1.50396675e-01
-7.95674801e-01 2.98524182e-02 3.34845185e-01 9.54181552e-02
-1.49967909e-01 -3.89746368e-01 6.06918752e-01 -1.30332053e-01
-7.81261206e-01 8.32078010e-02 -1.46916538e-01 4.21585083e-01
1.09902489e+00 -3.98697168e-01 -6.94023311e-01 -1.85183376e-01
-3.20947766e-01 1.70832559e-01 8.67852449e-01 5.75132728e-01
1.28675684e-01 -9.86727655e-01 -7.03016996e-01 1.72673285e-01
3.73532504e-01 -2.42508273e-03 2.86520775e-02 9.63540435e-01
-3.71679872e-01 5.93120098e-01 -7.37700686e-02 -4.48431790e-01
-1.31011641e+00 3.96779440e-02 3.07704061e-01 -5.75993396e-02
-3.05333167e-01 7.38101065e-01 -4.18471873e-01 -6.19627893e-01
6.09110653e-01 6.14613630e-02 4.87652346e-02 2.42245898e-01
2.38522962e-01 8.36404979e-01 4.98620749e-01 -6.69153631e-02
-4.07519072e-01 -3.46240520e-01 -3.68466645e-01 -1.38138175e-01
1.40601265e+00 -2.13986859e-02 1.79409370e-01 9.86654282e-01
8.70643914e-01 2.47478664e-01 -1.44504893e+00 -2.32657082e-02
4.65993732e-01 -4.63631034e-01 2.69235194e-01 -1.15740001e+00
-7.66757131e-01 5.08142412e-01 7.15967894e-01 3.41336846e-01
8.01289380e-01 -4.19737212e-02 -3.99854481e-02 3.82360697e-01
3.31372350e-01 -1.32396698e+00 -4.71250108e-03 2.73283124e-01
7.35782981e-01 -1.62165892e+00 4.00956452e-01 -2.23213404e-01
-7.49341428e-01 7.16381013e-01 8.10640514e-01 4.12937224e-01
5.46685994e-01 4.55614448e-01 6.50802135e-01 -3.42373163e-01
-1.34709024e+00 1.51478395e-01 1.46903753e-01 3.14209729e-01
1.17003906e+00 -5.04025258e-02 -4.76836354e-01 1.73688352e-01
5.54834353e-03 3.66148204e-01 3.34362388e-01 1.30877936e+00
-8.15020576e-02 -1.51123512e+00 -2.22730651e-01 5.10153472e-01
-2.70003706e-01 -2.56247725e-02 -7.99821794e-01 1.13363457e+00
8.98848101e-03 1.18686962e+00 1.62494004e-01 -1.38373777e-01
1.21412896e-01 4.33086812e-01 4.77600574e-01 -9.22797322e-01
-7.64598727e-01 -2.46251076e-01 3.76769990e-01 -5.94680727e-01
-5.90346992e-01 -8.74334931e-01 -9.02741551e-01 -2.11880013e-01
-5.03746212e-01 3.81960601e-01 6.66545808e-01 7.12011755e-01
7.44646907e-01 5.02883606e-02 3.59066665e-01 -1.54662520e-01
-6.80154443e-01 -8.94627094e-01 -6.90370575e-02 2.59297401e-01
-9.10005718e-02 -8.01527143e-01 -6.44255817e-01 -1.27581403e-01] | [9.61316204071045, 4.370011806488037] |
8ab5969d-26fb-4fdb-bcc9-c0faebda1174 | horae-an-annotated-dataset-of-books-of-hours | 2012.00351 | null | https://arxiv.org/abs/2012.00351v1 | https://arxiv.org/pdf/2012.00351v1.pdf | HORAE: an annotated dataset of books of hours | We introduce in this paper a new dataset of annotated pages from books of hours, a type of handwritten prayer books owned and used by rich lay people in the late middle ages. The dataset was created for conducting historical research on the evolution of the religious mindset in Europe at this period since the book of hours represent one of the major sources of information thanks both to their rich illustrations and the different types of religious sources they contain. We first describe how the corpus was collected and manually annotated then present the evaluation of a state-of-the-art system for text line detection and for zone detection and typing. The corpus is freely available for research. | ['Christopher Kermorvant', 'Dominique Stutzmann', 'Marie-Laurence Bonhomme', 'Mélodie Boillet'] | 2020-12-01 | null | null | null | null | ['line-detection'] | ['computer-vision'] | [-3.25430483e-01 3.07515562e-01 -4.96689528e-01 -5.26410379e-02
-2.99983293e-01 -7.13688135e-01 1.24446952e+00 3.42534065e-01
-4.05880481e-01 8.15134466e-01 7.58249938e-01 2.21041679e-01
-7.75894970e-02 -8.05166960e-01 -2.05697492e-01 -3.53735745e-01
2.87384391e-02 1.01888585e+00 5.08355677e-01 -5.92596233e-01
1.09375143e+00 4.14682835e-01 -1.36466372e+00 4.05949354e-01
6.88199282e-01 5.08587480e-01 1.08043388e-01 6.30683124e-01
-3.48730296e-01 7.33054519e-01 -8.39814603e-01 -1.04795134e+00
-6.78789150e-03 -5.98524392e-01 -1.06052685e+00 2.96325386e-01
6.43282533e-01 -4.91307139e-01 -3.43056142e-01 8.29187512e-01
2.77152210e-01 -3.06369126e-01 8.96768212e-01 -4.38894451e-01
-8.53827536e-01 1.05020881e+00 -3.56692463e-01 1.64605215e-01
6.45885408e-01 -7.60645688e-01 9.42410231e-01 -9.18146372e-01
1.37533724e+00 9.81052697e-01 5.80990314e-01 5.15726507e-01
-8.47104669e-01 -8.32480639e-02 -5.45084000e-01 2.88028002e-01
-1.29808056e+00 -5.23133218e-01 9.26368117e-01 -7.70741403e-01
4.71755594e-01 1.01872250e-01 1.00801313e+00 1.19226801e+00
1.64129004e-01 8.17296028e-01 1.29885161e+00 -9.57617700e-01
1.28877223e-01 5.19133151e-01 1.88247249e-01 1.00863802e+00
1.28785253e-01 -3.90832096e-01 -9.70458984e-01 -1.02727018e-01
1.04823935e+00 -5.28345704e-01 -2.26789311e-01 -9.80160851e-03
-1.35180211e+00 1.02588797e+00 -8.00780281e-02 6.65577590e-01
1.10983811e-01 -3.10387462e-01 5.27593613e-01 7.57851601e-02
5.40715575e-01 3.53704333e-01 -5.55073991e-02 -4.86924469e-01
-1.26031244e+00 1.68228671e-01 1.16792381e+00 9.65820074e-01
3.31770023e-03 -2.97925085e-01 2.31214046e-01 1.33832335e+00
3.36748540e-01 1.76823258e-01 6.80478930e-01 -6.71929121e-01
4.37454611e-01 4.77444887e-01 -1.01962298e-01 -9.39312577e-01
-3.65726352e-01 -1.68846771e-02 -5.66071689e-01 -9.03821737e-03
8.66503835e-01 2.27738500e-01 -5.37731528e-01 6.82466507e-01
1.97932541e-01 -8.74115884e-01 -2.16303453e-01 4.22350973e-01
1.13293850e+00 5.41484237e-01 -5.44442892e-01 -6.29380196e-02
1.74311316e+00 -9.30763423e-01 -8.89680326e-01 1.39608994e-01
3.39777678e-01 -1.16361725e+00 9.70101953e-01 8.67993116e-01
-1.20295596e+00 -1.06458083e-01 -1.23576009e+00 -3.14880341e-01
-5.78999758e-01 3.48055184e-01 5.04885793e-01 8.62429321e-01
-6.42903149e-01 7.46074021e-01 -5.19418597e-01 -8.68225873e-01
5.56341529e-01 -3.03249776e-01 -1.99500456e-01 3.75623435e-01
-8.99216175e-01 1.11401796e+00 4.29379344e-01 -1.72308356e-01
-2.97225744e-01 -2.32424200e-01 -5.02848327e-01 -4.13731396e-01
3.75660896e-01 -4.65577357e-02 1.21516907e+00 -7.04328179e-01
-1.47080004e+00 1.61792195e+00 3.42233479e-01 -4.45889741e-01
1.09166169e+00 -5.06689429e-01 -4.83786851e-01 6.25360727e-01
5.24245650e-02 1.70925364e-01 6.01638198e-01 -1.01942480e+00
-5.46092808e-01 -3.24366301e-01 -2.88565904e-01 -1.67768616e-02
-3.32858205e-01 4.53906834e-01 -5.49250066e-01 -1.17580998e+00
-3.26543152e-02 -5.31978965e-01 5.43032050e-01 -1.27030656e-01
-4.32754517e-01 -2.25795805e-01 4.04375076e-01 -1.43306696e+00
1.31332028e+00 -1.87716329e+00 2.87001505e-02 1.96721122e-01
9.90054384e-02 -2.88786680e-01 8.05133820e-01 1.07655764e+00
5.48625827e-01 2.98119307e-01 -1.38825849e-01 -8.10671449e-02
2.29876354e-01 1.71677694e-01 -4.34129715e-01 6.46913886e-01
-6.44624054e-01 3.58139098e-01 -7.36965954e-01 -9.56615686e-01
3.45198512e-02 3.46728593e-01 -3.75264324e-02 -1.83691233e-01
-1.38625100e-01 5.32682694e-04 -2.34953225e-01 7.71644890e-01
2.56310940e-01 5.26299663e-02 9.89042688e-03 5.49354628e-02
-6.18174851e-01 4.61820066e-01 -9.62011814e-01 1.73645592e+00
1.10494643e-01 1.15946829e+00 -2.82396883e-01 -9.76629630e-02
8.84886920e-01 2.89672285e-01 1.54292867e-01 -5.72195053e-01
1.24186121e-01 4.57041800e-01 -3.10064495e-01 -4.38699007e-01
1.27976429e+00 -4.15289253e-02 -1.25322536e-01 4.03276950e-01
6.34144843e-02 -2.39326507e-01 7.34425247e-01 4.13871318e-01
4.23257232e-01 4.33500201e-01 5.96307993e-01 -7.47174859e-01
5.57061613e-01 3.41101497e-01 2.29068249e-01 3.20141077e-01
1.21508636e-01 8.44155312e-01 5.60689509e-01 -5.77542782e-01
-1.61037838e+00 -9.13790226e-01 -8.83936167e-01 9.90188360e-01
-1.97031632e-01 -5.48868358e-01 -9.43733692e-01 -2.37032801e-01
-3.32669646e-01 4.99426007e-01 -9.54709053e-01 5.67037344e-01
-4.52356398e-01 -3.31306577e-01 7.08491504e-01 2.30786741e-01
8.33754361e-01 -1.11969984e+00 -9.29603159e-01 -3.74527946e-02
6.36007031e-03 -7.11832166e-01 -3.80491853e-01 -1.10382706e-01
-6.49714887e-01 -1.07649529e+00 -1.20885706e+00 -8.74884605e-01
4.54026699e-01 -6.05010867e-01 1.23221970e+00 2.23345026e-01
-4.09785092e-01 2.45087758e-01 -5.89638472e-01 -5.69595933e-01
-7.43093789e-01 3.85820955e-01 -2.09893361e-01 -3.28197867e-01
4.06991690e-02 3.57391685e-02 -2.67772704e-01 1.01311192e-01
-8.02230239e-01 2.21839741e-01 1.03771597e-01 8.38446259e-01
2.61548758e-02 -3.27296779e-02 1.50118321e-01 -1.47497380e+00
4.11950767e-01 2.68471334e-02 -6.56683564e-01 3.36114705e-01
-3.19539338e-01 2.19156947e-02 2.23245263e-01 -3.06935422e-02
-1.27813280e+00 -4.75571066e-01 -3.00462961e-01 7.80158043e-01
-4.29399498e-02 3.75970125e-01 1.84429109e-01 -6.25944957e-02
6.69777095e-01 -1.12416204e-02 -4.83831882e-01 -7.71351278e-01
1.89373404e-01 1.03239703e+00 9.63475525e-01 -5.77888787e-01
6.86785936e-01 4.86985981e-01 -1.78148463e-01 -1.46586275e+00
-7.78736413e-01 -4.65089828e-01 -1.23956847e+00 -5.09597301e-01
9.61198628e-01 -5.16056716e-01 -3.71332228e-01 8.34318995e-01
-8.83028567e-01 -5.25405407e-01 -2.46442527e-01 1.82483643e-01
-5.24480164e-01 6.26523733e-01 -1.04554462e+00 -8.69209349e-01
-3.91529053e-01 -6.28170744e-02 6.89539969e-01 2.90660977e-01
-4.01394844e-01 -1.38848400e+00 5.36609292e-01 3.79934609e-01
-7.13996440e-02 6.22188270e-01 1.00456119e+00 -4.50579524e-01
1.24924676e-02 1.68036297e-01 1.66863203e-01 -1.73162952e-01
-1.29461125e-01 4.47147161e-01 -8.48592758e-01 8.19811821e-02
-4.09421623e-02 -2.11790785e-01 7.91843116e-01 -1.13697276e-01
5.02588630e-01 -5.89616597e-01 -1.53052285e-02 2.97393173e-01
1.61423719e+00 -1.71393435e-02 9.56683338e-01 1.15875423e+00
5.53842008e-01 7.44174898e-01 3.65421385e-01 8.29892457e-01
2.97180533e-01 6.70998454e-01 -1.90264255e-01 9.48989615e-02
-2.01517329e-01 -2.67397344e-01 3.05440515e-01 1.06653428e+00
-4.61859703e-01 4.69071865e-02 -1.16655433e+00 7.59340346e-01
-1.61198258e+00 -1.12914968e+00 -4.57042396e-01 2.34275818e+00
1.19397247e+00 3.40810895e-01 7.17571557e-01 3.67032617e-01
5.18054843e-01 -2.63866559e-02 2.30368286e-01 -5.37992358e-01
-4.48758781e-01 1.26148880e-01 5.98420084e-01 3.51843864e-01
-1.08841622e+00 7.91698039e-01 7.37622452e+00 7.74430096e-01
-6.51184797e-01 -1.73840165e-01 2.96716630e-01 2.65678972e-01
7.41153881e-02 -2.44341686e-01 -9.20076013e-01 5.14971733e-01
6.68434560e-01 -3.79486680e-02 1.90830916e-01 6.04508936e-01
-2.46169671e-01 -5.99455655e-01 -9.62843359e-01 5.93062460e-01
5.41811824e-01 -1.55454290e+00 -3.75389159e-02 1.89586699e-01
6.76791072e-01 -4.62995656e-02 -2.10060984e-01 -3.41349214e-01
1.62896261e-01 -8.52966487e-01 1.52424467e+00 7.06453502e-01
6.89294755e-01 -7.47540116e-01 6.52808964e-01 2.20141366e-01
-6.33314550e-01 2.58931816e-01 -5.92047572e-01 6.44198135e-02
7.02774078e-02 4.48174745e-01 -5.77903092e-01 4.08566654e-01
7.55758047e-01 7.16913640e-01 -1.16227663e+00 1.21995580e+00
-4.58890259e-01 7.50564337e-01 -1.94288090e-01 -2.93105274e-01
5.43458834e-02 -3.44625771e-01 4.90207702e-01 1.56624103e+00
2.23092258e-01 -1.09160118e-01 -2.04430714e-01 5.94123304e-01
4.57678549e-02 4.83907610e-01 -2.94644415e-01 -1.35089636e-01
3.92358601e-01 1.31037760e+00 -1.42247295e+00 -3.25257421e-01
-3.96790743e-01 1.03968024e+00 3.66621837e-02 -2.31543168e-01
-4.58805472e-01 -7.51612067e-01 -3.88554275e-01 5.40148318e-01
2.06775025e-01 -3.86011302e-01 -5.80999374e-01 -1.22502029e+00
-1.40987858e-01 -5.74692011e-01 5.96594214e-01 -6.66742861e-01
-1.34058440e+00 4.08530831e-01 1.56276986e-01 -7.94010997e-01
-2.60593686e-02 -7.09312856e-01 -3.04237872e-01 3.41403306e-01
-9.41347897e-01 -1.27704430e+00 -1.91466719e-01 4.05207723e-01
7.25392401e-01 -4.19234425e-01 7.63807654e-01 1.72892287e-02
-3.90698433e-01 3.91160190e-01 6.27245128e-01 7.66955197e-01
9.60322499e-01 -1.72089064e+00 2.73721457e-01 5.66957474e-01
3.34227145e-01 3.67415667e-01 8.10695767e-01 -7.53178418e-01
-9.59977806e-01 -2.35223055e-01 1.16181231e+00 -6.44766212e-01
9.33774531e-01 -5.43260098e-01 -7.32015550e-01 5.69204688e-01
7.45361328e-01 -1.17302573e+00 7.48502433e-01 -8.42359383e-03
-2.55033910e-01 2.40444005e-01 -1.25771940e+00 5.47270834e-01
6.46121502e-01 -4.88414854e-01 -1.29143918e+00 5.74810326e-01
-1.58624455e-01 -5.20427883e-01 -1.21055222e+00 -4.09118354e-01
8.29900026e-01 -1.11092830e+00 5.67626953e-01 3.81731361e-01
9.83453989e-01 7.74687380e-02 2.50670224e-01 -9.15668488e-01
-2.16308594e-01 -7.72856057e-01 1.45937398e-01 1.76447153e+00
4.96333748e-01 -4.92108464e-01 7.52804756e-01 2.11805388e-01
-2.24196032e-01 -2.69607484e-01 -8.10341775e-01 -4.24282461e-01
3.20049286e-01 3.67107421e-01 1.92265883e-01 1.04868793e+00
4.72748160e-01 4.54893827e-01 -3.49733889e-01 -6.54511511e-01
7.19850004e-01 1.56655349e-02 5.16118407e-01 -1.50013709e+00
-2.76579201e-01 -8.41379464e-01 -2.97681808e-01 -5.85812926e-01
-4.81696695e-01 -8.13661814e-01 -7.94177800e-02 -1.72432888e+00
2.98591495e-01 -4.67784740e-02 8.11259210e-01 2.84207672e-01
3.25783372e-01 6.30649567e-01 3.01494002e-01 7.87661076e-01
-2.51265287e-01 4.47217375e-02 9.33473349e-01 1.17806315e-01
-1.06633186e-01 -4.37272161e-01 -3.66499752e-01 9.88292038e-01
7.10234404e-01 -2.35797808e-01 2.28974923e-01 -8.74145031e-02
7.86134720e-01 -3.71667892e-01 1.97183322e-02 -9.92475927e-01
-2.07115337e-02 7.54938796e-02 8.32287908e-01 -9.31647778e-01
8.09616670e-02 -5.71901798e-01 5.39875105e-02 5.21651745e-01
-2.39937052e-01 -2.98734065e-02 -1.01153046e-01 2.39823967e-01
-1.90046970e-02 -7.31731415e-01 8.34439158e-01 -4.21772867e-01
-7.05915868e-01 -3.88966978e-01 -5.67502260e-01 1.17485717e-01
8.67817581e-01 -4.53091323e-01 -8.76477361e-01 -1.07888013e-01
-4.42478508e-01 -1.93965703e-01 1.06207120e+00 1.21191338e-01
2.56025702e-01 -1.24993622e+00 -8.93174410e-01 -6.52084649e-02
4.67994548e-02 -4.68636841e-01 -2.68387347e-01 5.72197378e-01
-1.49999714e+00 1.48913965e-01 -8.01951766e-01 -2.54969537e-01
-1.25971425e+00 3.87319237e-01 -2.87017431e-02 -1.88599288e-01
-1.21557045e+00 4.83007401e-01 -3.13043743e-01 2.89387051e-02
1.82836175e-01 4.24349681e-02 -8.49272549e-01 5.03025711e-01
7.11769521e-01 8.80318999e-01 -9.26314369e-02 -8.82088482e-01
-9.67426747e-02 6.81643307e-01 -1.58799179e-02 -5.60181379e-01
1.41582239e+00 -2.88118154e-01 -4.09641415e-01 1.30410528e+00
5.66202044e-01 7.24763215e-01 -7.56849527e-01 -4.85590147e-03
2.31587917e-01 -5.14213443e-01 -2.51568913e-01 -1.08822143e+00
-5.60081244e-01 5.91171980e-01 1.44981563e-01 4.96697575e-01
6.07160032e-01 6.40747547e-02 6.29871905e-01 2.76518971e-01
3.59466463e-01 -1.91029143e+00 1.16518233e-02 5.14996052e-01
1.13290381e+00 -6.79856122e-01 6.91224754e-01 -4.67152655e-01
-6.41722500e-01 1.63809466e+00 -1.34105816e-01 -2.04437748e-01
4.89898682e-01 1.93913072e-01 4.86357026e-02 -3.69770318e-01
-1.78654030e-01 1.63778335e-01 4.97222513e-01 5.85554540e-01
9.07075465e-01 -9.08554047e-02 -9.84657228e-01 4.15706366e-01
-1.04626739e+00 -3.26654524e-01 1.17760658e+00 1.14951277e+00
-7.17041135e-01 -1.05909240e+00 -6.46872163e-01 2.97661692e-01
-8.53253901e-01 6.71121255e-02 -1.24240506e+00 1.16675186e+00
-1.29915560e-02 6.00460410e-01 -1.13458661e-02 9.23216641e-02
-7.13824779e-02 1.67978369e-02 1.00816381e+00 -5.15661240e-01
-9.61398542e-01 3.74120116e-01 5.58148444e-01 2.22952530e-01
-5.65829277e-01 -9.52309728e-01 -1.23583460e+00 -5.98451972e-01
-2.20927224e-01 4.02500838e-01 6.52375877e-01 7.73285449e-01
-6.32149279e-01 1.17355086e-01 1.99734315e-01 -7.09796727e-01
-1.68178037e-01 -1.16781259e+00 -1.30396307e+00 2.73689270e-01
-3.50722671e-02 -4.16803628e-01 -8.40939060e-02 6.41941726e-01] | [10.154866218566895, 10.307100296020508] |
2019234a-7722-453b-bc4b-9e71900a2446 | estimating-the-performance-of-entity | 2210.01230 | null | https://arxiv.org/abs/2210.01230v2 | https://arxiv.org/pdf/2210.01230v2.pdf | Estimating the Performance of Entity Resolution Algorithms: Lessons Learned Through PatentsView.org | This paper introduces a novel evaluation methodology for entity resolution algorithms. It is motivated by PatentsView.org, a U.S. Patents and Trademarks Office patent data exploration tool that disambiguates patent inventors using an entity resolution algorithm. We provide a data collection methodology and tailored performance estimators that account for sampling biases. Our approach is simple, practical and principled -- key characteristics that allow us to paint the first representative picture of PatentsView's disambiguation performance. This approach is used to inform PatentsView's users of the reliability of the data and to allow the comparison of competing disambiguation algorithms. | ['Christina Jones', 'Sarvo Madhavan', 'Youngsoo Baek', 'Emma Hickerson', 'Sokhna A York', 'Olivier Binette'] | 2022-10-03 | null | null | null | null | ['entity-resolution'] | ['natural-language-processing'] | [ 2.81511635e-01 2.70334244e-01 -9.15642262e-01 -3.12356874e-02
-8.41799080e-01 -1.06055188e+00 7.60477364e-01 4.17970330e-01
-2.92506933e-01 1.14516973e+00 1.47419363e-01 -1.04185760e+00
-9.47712421e-01 -5.22802889e-01 -3.85465294e-01 -1.92323774e-01
4.44074534e-02 5.78332424e-01 -8.02905336e-02 2.85758048e-01
9.87662077e-01 7.91725755e-01 -1.41554844e+00 -9.44979861e-02
1.32321560e+00 9.51896310e-01 2.86624670e-01 1.53391391e-01
-3.97001058e-01 2.80544221e-01 -7.21936941e-01 -3.69206458e-01
6.09183788e-01 2.68382758e-01 -7.39237309e-01 -6.76485837e-01
4.91317004e-01 1.32139549e-01 1.56408310e-01 1.18097317e+00
4.36272651e-01 -2.31357425e-01 7.47469366e-01 -9.93158996e-01
-6.46083176e-01 5.81703901e-01 -3.59064370e-01 5.01852989e-01
7.14331269e-01 -8.88797194e-02 1.29590094e+00 -7.41638422e-01
1.29211283e+00 1.03855121e+00 4.56617296e-01 -1.14166001e-02
-1.18378544e+00 -1.02012908e+00 3.24560463e-01 -4.06088799e-01
-1.58466530e+00 -5.43378949e-01 4.95418251e-01 -8.82011831e-01
9.91736770e-01 5.00512064e-01 4.64028299e-01 9.86024022e-01
4.38764185e-01 3.29653293e-01 1.36907041e+00 -3.09413493e-01
3.45358700e-01 3.65344286e-01 3.91569525e-01 2.12021843e-01
1.29193139e+00 4.61976141e-01 -4.28514481e-01 -6.13349080e-01
9.10493195e-01 -1.68822050e-01 -6.36605471e-02 -5.88567555e-01
-8.16992402e-01 6.09266520e-01 5.03904335e-02 5.84579945e-01
-7.56114185e-01 -2.09521532e-01 -1.69504493e-01 3.99255246e-01
2.31898621e-01 1.60494435e+00 -5.37004888e-01 -1.32369295e-01
-9.63589013e-01 5.08783877e-01 9.34356809e-01 1.19672155e+00
6.03939116e-01 -3.10733378e-01 -2.03418322e-02 2.40760773e-01
3.77713978e-01 2.80165821e-01 2.40108579e-01 -8.50330949e-01
3.71323407e-01 6.53560340e-01 6.24492824e-01 -6.37188971e-01
-1.62168756e-01 -5.87197065e-01 2.74260268e-02 3.04736793e-01
8.09066966e-02 -1.63155332e-01 -9.44862008e-01 1.10448754e+00
-2.16088414e-01 -2.91320443e-01 4.61306982e-02 3.11568350e-01
6.37000322e-01 5.84145030e-03 5.42197526e-01 -4.94935304e-01
1.34008431e+00 -6.60674199e-02 -7.72262692e-01 2.68582374e-01
2.92780995e-01 -9.47118163e-01 6.07702374e-01 6.67980731e-01
-1.09186876e+00 -4.27334845e-01 -1.07134557e+00 3.71749789e-01
-9.27164793e-01 -1.97913423e-01 1.01227903e+00 8.72434378e-01
-5.77775121e-01 9.06394005e-01 -2.40603045e-01 -2.65361160e-01
8.26420367e-01 6.12775385e-01 1.01048306e-01 3.53475034e-01
-1.13853288e+00 1.15657127e+00 6.28744185e-01 -6.73727214e-01
-2.88441598e-01 -9.56256926e-01 -2.21771210e-01 -4.74690925e-03
5.73792040e-01 -6.10160649e-01 1.07073450e+00 -4.46110934e-01
-1.02781129e+00 4.84008819e-01 -6.60280809e-02 -2.98672557e-01
3.26006502e-01 -2.00531453e-01 -1.03765428e+00 -2.70632803e-01
6.75542951e-01 3.18175197e-01 5.10032296e-01 -1.03961813e+00
-1.04372430e+00 -4.08547044e-01 -2.23604500e-01 -1.05040573e-01
2.38166779e-01 -5.76772466e-02 -3.09065133e-01 -8.30173433e-01
6.91804141e-02 -6.51776612e-01 -4.10747111e-01 -1.02738190e+00
-6.32698298e-01 -4.51567173e-01 3.33068669e-01 -2.50947565e-01
1.68602455e+00 -1.73606539e+00 -3.86263222e-01 7.82565534e-01
1.97827131e-01 2.67863154e-01 7.44335279e-02 3.95123273e-01
-5.30334473e-01 8.75499547e-01 2.11612940e-01 3.92605484e-01
8.98390561e-02 -3.90432805e-01 -5.56232095e-01 3.26354094e-02
3.35064232e-01 9.78003263e-01 -7.97449529e-01 -1.55708805e-01
1.41141370e-01 -5.16433045e-02 -1.48062959e-01 -3.10953826e-01
-1.15273654e-01 2.11868435e-01 -1.17688274e+00 1.24969947e+00
5.99335492e-01 -4.85930964e-02 3.28612268e-01 -1.84110552e-01
-8.06523919e-01 8.48106146e-01 -1.28046083e+00 1.59936929e+00
1.24721758e-01 3.95232588e-01 -3.14987957e-01 -3.07828158e-01
1.07074094e+00 6.12180121e-02 6.45008385e-01 -8.90788138e-01
-6.36475459e-02 6.37043297e-01 -2.44639710e-01 -3.28987777e-01
8.15755010e-01 1.33464217e-01 -1.80666298e-01 1.87665895e-01
-1.64231434e-01 -8.11392888e-02 2.58831620e-01 -2.00278889e-02
9.73943055e-01 2.24522889e-01 7.87103534e-01 -6.91939950e-01
5.27384058e-02 3.86160523e-01 6.02265358e-01 1.00059438e+00
1.69758573e-01 -1.24446437e-01 3.80657792e-01 -3.69493872e-01
-7.81578600e-01 -9.55427527e-01 -6.90290868e-01 4.18166220e-01
3.76575813e-02 -4.92290884e-01 -2.82568425e-01 -7.63485193e-01
5.65251052e-01 7.86210060e-01 -4.44838554e-01 4.63263005e-01
-6.20515971e-03 -4.15246457e-01 1.68737173e-01 1.40822530e-01
-7.34759942e-02 -6.15781248e-01 -5.69625497e-01 2.52941310e-01
4.50373977e-01 -5.96840858e-01 -3.22399922e-02 3.93955946e-01
-9.85346615e-01 -1.38278723e+00 -4.99844134e-01 -4.64689523e-01
2.87503064e-01 -3.51147018e-02 1.06153750e+00 -6.42766654e-01
-3.73500049e-01 3.24382067e-01 -2.50817034e-02 -9.58065808e-01
-3.27634275e-01 4.23929065e-01 4.61707413e-02 -7.34880209e-01
8.78785849e-01 -6.05417550e-01 -4.04725194e-01 2.37311587e-01
-6.32647574e-01 -8.91996741e-01 1.09501398e+00 1.37259141e-01
8.14573884e-01 2.45651856e-01 9.30553496e-01 -1.31572735e+00
1.24113417e+00 -5.46201825e-01 -1.18916345e+00 6.37548387e-01
-1.80454981e+00 3.26310813e-01 -2.34875247e-01 -2.83941120e-01
-1.03898084e+00 1.57739908e-01 5.63311994e-01 -2.03705356e-01
-2.99621761e-01 7.80795753e-01 -3.18919897e-01 -2.68611699e-01
7.31066406e-01 -4.71807897e-01 -3.59320700e-01 -7.84691215e-01
4.02750552e-01 6.96872413e-01 5.16478479e-01 -5.59287548e-01
5.79276025e-01 -9.31336656e-02 7.27744177e-02 -4.54527110e-01
-1.93587556e-01 -7.25026309e-01 -5.24471343e-01 2.29789495e-01
4.23088372e-01 -8.40851188e-01 -8.48563433e-01 -2.08297029e-01
-9.98586059e-01 4.23665315e-01 -3.82187366e-01 6.18552685e-01
-1.96925089e-01 1.33574352e-01 3.23589444e-01 -8.16968381e-01
-1.94202259e-01 -1.06448483e+00 5.90088725e-01 2.41440848e-01
-7.94633150e-01 -7.49247968e-01 3.79383713e-01 2.13025689e-01
3.05142283e-01 4.08988357e-01 6.27117038e-01 -1.18250751e+00
-9.53733981e-01 -2.97756076e-01 -1.41693428e-01 -2.30128527e-01
4.30684716e-01 1.76771551e-01 -9.18604553e-01 -5.34613393e-02
-2.74440765e-01 4.72435355e-01 7.56408930e-01 6.71613514e-01
8.54174674e-01 -2.66496867e-01 -1.06449091e+00 9.06000137e-02
1.42679608e+00 9.37666774e-01 7.22902238e-01 7.79042959e-01
3.65972638e-01 7.50455022e-01 8.64211798e-01 3.37019533e-01
-1.57857358e-01 5.94626427e-01 -6.39408901e-02 6.96055442e-02
4.31596130e-01 -3.30082774e-01 -3.81102532e-01 -1.01915393e-02
-3.71520907e-01 -1.20308757e-01 -8.66486907e-01 4.45450544e-01
-1.67339146e+00 -9.32227671e-01 -6.62704930e-02 2.44299030e+00
5.44610143e-01 5.26620269e-01 4.06310618e-01 -3.12536061e-01
6.85997725e-01 -9.84931514e-02 -7.01499939e-01 -3.84990424e-01
-1.29729316e-01 7.17310846e-01 1.25089037e+00 4.08410400e-01
-9.76167798e-01 8.57542157e-01 7.82996273e+00 6.73279643e-01
-5.34202933e-01 -4.94865417e-01 1.90023482e-01 5.30714728e-02
-8.18983972e-01 3.29568624e-01 -8.54086280e-01 3.59308034e-01
9.19726610e-01 -8.57729018e-01 -2.27052066e-03 7.71076262e-01
-3.95323820e-02 -1.79822519e-01 -1.06101918e+00 7.58858800e-01
-5.28555632e-01 -1.92703986e+00 3.58664542e-01 7.81866550e-01
5.25659740e-01 -2.92031646e-01 2.66526014e-01 -1.38589507e-03
6.08458996e-01 -1.13864303e+00 4.57338899e-01 7.58227587e-01
7.66170323e-01 -8.48956525e-01 4.20111716e-01 -4.14490610e-01
-9.56923485e-01 -3.39693308e-01 -2.11013719e-01 1.30768359e-01
3.94057631e-02 7.00332880e-01 -1.20104694e+00 1.02634537e+00
5.07578313e-01 7.87476063e-01 -3.92414123e-01 1.02741349e+00
1.84317713e-03 1.09896779e-01 6.68314695e-02 -1.18166059e-02
5.68539873e-02 -3.29866886e-01 7.29841709e-01 1.03574836e+00
4.91175711e-01 2.48722918e-02 -2.64640898e-01 1.22500885e+00
1.81212239e-02 3.79032120e-02 -9.75155056e-01 -6.53607726e-01
1.35411167e+00 8.56993258e-01 -4.79597032e-01 -1.82856694e-01
-2.05743909e-01 2.10021824e-01 -3.68866801e-01 4.14680213e-01
-1.36506364e-01 -6.69467986e-01 7.37955689e-01 2.57441491e-01
5.23572028e-01 4.41520773e-02 -6.62671149e-01 -5.87366700e-01
-2.36558869e-01 -9.29493070e-01 5.89416862e-01 -4.96112108e-01
-9.74833548e-01 2.42349327e-01 1.79514050e-01 -1.27094340e+00
-1.70649663e-01 -8.18655610e-01 -4.98468816e-01 1.14925420e+00
-1.38937712e+00 -4.94930595e-01 2.71479934e-01 1.19517520e-01
4.22741193e-03 -7.23746657e-01 7.62522101e-01 4.04460393e-02
-5.06986439e-01 4.51062769e-01 2.22194359e-01 -4.05728519e-01
7.07110465e-01 -1.41402495e+00 6.39930725e-01 4.67967927e-01
5.30234501e-02 1.42047822e+00 7.08662212e-01 -1.22902083e+00
-1.16400087e+00 -6.55969501e-01 8.63363385e-01 -9.89088297e-01
1.03242731e+00 2.70400494e-01 -6.33140206e-01 5.22672832e-01
2.19357118e-01 -9.41339612e-01 9.80746806e-01 5.83665967e-01
-2.24803954e-01 -2.07724124e-01 -1.12563431e+00 3.59183609e-01
8.73100519e-01 -6.02932990e-01 -1.08129275e+00 5.03232144e-02
5.07490754e-01 -9.21241790e-02 -1.50770307e+00 2.87812084e-01
8.96535337e-01 -4.53602344e-01 1.02397180e+00 -7.50021219e-01
-9.92243364e-02 -2.20049784e-01 -4.57135737e-02 -9.03174877e-01
-5.64279854e-01 -1.14738750e+00 2.44097933e-01 1.17163813e+00
1.05750847e+00 -1.14293754e+00 4.52961504e-01 1.05783761e+00
1.58922225e-02 -4.05523807e-01 -8.88809741e-01 -1.07570004e+00
-2.92694345e-02 -1.59380570e-01 1.02023757e+00 1.28535795e+00
1.88074902e-01 3.49898040e-01 1.66209728e-01 6.29629046e-02
6.56934977e-01 4.38572615e-01 5.68391383e-01 -1.90309823e+00
1.04440682e-01 -8.37467492e-01 -2.25561157e-01 -5.92278361e-01
-2.31768176e-01 -4.44053590e-01 -7.02691317e-01 -1.23865330e+00
-1.11197270e-01 -6.35424256e-01 -8.63296032e-01 1.80188373e-01
8.78686681e-02 -5.32255530e-01 -1.59963846e-01 7.29556680e-01
-3.21403086e-01 -3.59818786e-01 9.75647807e-01 -7.19300807e-02
-7.34389663e-01 1.57322526e-01 -1.48290849e+00 2.75778949e-01
9.87153947e-01 -5.29178083e-01 -5.37553966e-01 1.85479030e-01
4.82795775e-01 -1.96663082e-01 9.01967660e-02 -5.78598201e-01
2.90922821e-01 -4.65053439e-01 4.82046664e-01 -8.95647645e-01
-2.82344908e-01 -7.73322642e-01 8.25657308e-01 5.73311150e-01
-3.65275502e-01 5.90970106e-02 4.29268867e-01 7.04518974e-01
-1.68368164e-02 -2.42153183e-01 9.51205119e-02 -1.60120562e-01
-7.71175146e-01 2.35862494e-01 -1.79386437e-01 -2.96053320e-01
9.53276575e-01 -5.52358031e-01 -7.97126174e-01 1.95683613e-01
-6.47543252e-01 1.11509897e-01 6.86293960e-01 5.36196113e-01
1.75738052e-01 -1.19964206e+00 -4.15036261e-01 -6.53329417e-02
3.56602430e-01 -3.13259661e-01 -4.01628435e-01 4.74227071e-01
-3.21401544e-02 1.09999371e+00 -2.97247201e-01 2.89368741e-02
-9.13000643e-01 7.90612221e-01 1.28451111e-02 -1.78448841e-01
-1.93182714e-02 5.09438753e-01 -5.33276759e-02 1.99163765e-01
1.11306131e-01 -4.87045467e-01 -4.58277285e-01 4.77515608e-01
2.56984413e-01 6.03547156e-01 -6.85494542e-02 -1.25548378e-01
-7.79970527e-01 5.09392858e-01 -4.41429168e-01 -8.98699164e-02
1.26551926e+00 1.54431850e-01 1.61229983e-01 2.88183361e-01
6.04406536e-01 4.98392522e-01 -6.31070316e-01 -1.46894082e-01
7.70124972e-01 -5.83227515e-01 8.96025151e-02 -1.40367436e+00
-4.15704519e-01 5.91232069e-02 7.89140999e-01 6.21362329e-01
7.80788064e-01 -1.73011478e-02 -3.73019464e-02 3.59979302e-01
4.45627481e-01 -1.43961418e+00 -4.79031473e-01 -1.58958837e-01
8.10118079e-01 -9.37173009e-01 7.07532942e-01 -3.95425618e-01
-2.82599866e-01 9.42699850e-01 5.97963855e-02 4.36601371e-01
3.44401687e-01 -3.03201955e-02 -2.69573212e-01 -6.35562599e-01
-3.44324917e-01 -1.23049088e-01 6.59354985e-01 7.28656173e-01
4.91654605e-01 3.11575562e-01 -7.07308710e-01 7.08754838e-01
3.66497040e-02 1.91911146e-01 1.09512329e-01 1.14602351e+00
-6.89828336e-01 -1.65181839e+00 -3.29069346e-01 6.64793074e-01
-5.41575015e-01 -2.69112974e-01 -1.31812429e+00 1.30056298e+00
2.52713948e-01 8.74764204e-01 -2.88785309e-01 -3.38849097e-01
7.92758107e-01 1.06980167e-01 3.48212898e-01 -4.97183293e-01
-3.55284452e-01 3.23905349e-01 4.87070233e-01 -5.47676086e-01
-6.14900649e-01 -8.88022661e-01 -7.34375596e-01 -1.45229042e-01
-6.32539451e-01 5.96240699e-01 8.33232224e-01 4.65341598e-01
1.03620720e+00 5.89210331e-01 6.49691403e-01 -9.82483327e-02
-2.28727221e-01 -7.44364440e-01 -6.74434185e-01 -3.39286998e-02
1.47522494e-01 -8.88565660e-01 -1.22598179e-01 -4.23768014e-01] | [9.722464561462402, 8.252643585205078] |
174d00b3-76d0-4944-b32a-e566f5ddd2d1 | bag-of-tricks-and-a-strong-baseline-for-fgvc | null | null | http://star.informatik.rwth-aachen.de/Publications/CEUR-WS/Vol-3180/paper-182.pdf | http://star.informatik.rwth-aachen.de/Publications/CEUR-WS/Vol-3180/paper-182.pdf | Bag of Tricks and a Strong Baseline for FGVC | Fine-grained visual classification (FGVC), as a subclass classification task under the superclass, brings
more challenges. However, in addition to fine-grained features, the FungiCLEF 2022 dataset is also
characterized by imbalance and rich meta-information. This motivates us to explore the impact of different
methods and components in fine-grained classification on FungiCLEF 2022. We explore the impact of
different data augmentations, backbones, loss functions, and attention mechanisms on classification
performance. Additionally, we explore different metadata usage scenarios. In the end, we win second
place in the CVPR2022 FGVC Workshop FungiCLEF 2022 challenge. Our code is available at https:
//github.com/wujiekd/Bag-of-Tricks-and-a-Strong-Baseline-for-Fungi-Fine-Grained-Classification. | ['Feng Shuang', 'Fang Gao', 'Zepeng Liu', 'Zhihong Wei', 'Shenshen Du', 'Zhongpeng Cai', 'Liwen Zhang', 'Guochen Xie', 'Keda Lu', 'Hao Chang', 'Jun Yu'] | 2022-09-05 | null | null | null | conference-and-labs-of-the-evaluation-forum | ['fine-grained-image-classification'] | ['computer-vision'] | [-3.36287677e-01 -5.04077971e-01 -3.25201899e-01 -4.98222023e-01
-9.64502394e-01 -1.01286304e+00 9.32964683e-01 2.46139407e-01
-2.84695387e-01 9.84416366e-01 4.25757855e-01 -1.21556140e-01
-4.70767729e-02 -5.42727590e-01 -6.70921683e-01 -6.42357945e-01
2.01591447e-01 3.75383079e-01 3.97886932e-02 3.26207072e-01
1.14750594e-01 2.78264433e-01 -1.56784725e+00 8.59156728e-01
7.75738597e-01 8.13921630e-01 1.75624579e-01 7.61764884e-01
-1.55383527e-01 5.62261879e-01 -7.08349049e-01 -8.79154205e-01
2.65546530e-01 6.15993105e-02 -8.72813702e-01 -1.50385901e-01
1.19201922e+00 -8.11524913e-02 -2.55191088e-01 8.74586701e-01
5.34956396e-01 -2.21958116e-01 8.86817455e-01 -1.40892398e+00
-7.54612207e-01 3.18468928e-01 -6.70751572e-01 5.89658499e-01
1.42198075e-02 4.13045466e-01 1.03081512e+00 -8.76363993e-01
8.89714301e-01 1.28777432e+00 7.80679464e-01 5.95678687e-01
-1.20059550e+00 -7.63251781e-01 3.08913112e-01 5.32771707e-01
-1.34689188e+00 -3.59977841e-01 5.59339896e-02 -7.38848507e-01
9.87948358e-01 4.45883930e-01 2.34413564e-01 1.50512314e+00
7.71584511e-02 6.08477056e-01 1.54696357e+00 -2.87899166e-01
-2.24131066e-02 4.81891781e-02 5.62271714e-01 5.51307023e-01
4.22448337e-01 2.25517765e-01 -5.10370851e-01 -1.22403599e-01
3.82312864e-01 -7.22740665e-02 -4.77681279e-01 -1.18188813e-01
-1.25816393e+00 7.49206722e-01 4.63211238e-01 3.57846797e-01
1.69746354e-02 2.95246869e-01 6.87255442e-01 2.85475820e-01
7.80775070e-01 4.73598242e-01 -8.46528590e-01 -2.17134967e-01
-8.92106593e-01 4.84723538e-01 6.17753804e-01 6.45719767e-01
3.79625499e-01 -3.32102060e-01 -6.36146486e-01 9.10669148e-01
-3.41598019e-02 2.09128484e-01 4.14463371e-01 -7.92701364e-01
6.09776139e-01 1.08631074e-01 -2.18819827e-01 -6.01140857e-01
-2.13002264e-01 -7.78632939e-01 -8.86485577e-01 4.25049007e-01
7.00703740e-01 2.56266803e-01 -1.07657802e+00 1.70308459e+00
3.53523999e-01 1.61854159e-02 -7.27862418e-01 9.66308534e-01
9.25617874e-01 1.71122804e-01 2.65962809e-01 1.96067005e-01
1.60686195e+00 -1.39963710e+00 -4.62510884e-01 -4.10087407e-02
5.47218978e-01 -6.44405425e-01 1.47728550e+00 5.70315182e-01
-5.79814613e-01 -6.44269586e-01 -8.99283051e-01 -4.03339624e-01
-7.93013692e-01 6.63594157e-02 6.38432443e-01 6.20073557e-01
-8.71079624e-01 6.72044337e-01 -6.57647491e-01 -4.85329866e-01
7.37173021e-01 -1.12790816e-01 -6.78516150e-01 -2.36481920e-01
-7.80563712e-01 7.22744823e-01 1.86788648e-01 -4.19673622e-01
-9.37421739e-01 -1.20396793e+00 -2.49935150e-01 5.76819852e-02
3.97464111e-02 -8.50863338e-01 1.07949984e+00 -6.12165868e-01
-6.40093088e-01 1.17559147e+00 -1.01160549e-01 -3.48269671e-01
6.67517364e-01 -1.74062133e-01 -1.69891417e-01 -1.25824213e-01
1.08796284e-01 7.77383566e-01 5.59014261e-01 -9.80251789e-01
-6.70355737e-01 -7.02198505e-01 2.56928932e-02 -1.68438643e-01
-1.26938060e-01 1.23230182e-02 -5.73297203e-01 -9.78365719e-01
-5.92774212e-01 -7.53064275e-01 6.76684454e-02 -7.57616311e-02
-5.56950510e-01 -3.34078521e-01 3.66090328e-01 -7.83666372e-01
1.24632657e+00 -2.29029083e+00 6.63797930e-02 -2.59626925e-01
4.52675700e-01 1.38721853e-01 -6.54822528e-01 1.99966893e-01
-3.58306974e-01 6.53877199e-01 1.86356589e-01 -5.94076157e-01
2.30133697e-01 -3.63589495e-01 -1.97366878e-01 2.69737273e-01
1.62110552e-01 9.62865889e-01 -4.81819540e-01 -3.25111508e-01
2.04823881e-01 7.04061747e-01 -5.62961996e-01 1.00955695e-01
-3.83801997e-01 3.94985527e-01 -1.26751691e-01 9.86430109e-01
8.26636255e-01 -5.54577947e-01 -1.12711951e-01 -5.13583064e-01
1.62027571e-02 3.28495771e-01 -7.07567573e-01 1.73721862e+00
-3.13303143e-01 4.94954199e-01 -1.11622782e-02 -4.07754034e-01
4.52913523e-01 2.93767303e-02 -1.20264292e-01 -8.20564747e-01
-7.46802539e-02 1.95002764e-01 -2.55219460e-01 -1.42800897e-01
1.97806716e-01 1.15012310e-01 -6.22073971e-02 1.66118786e-01
2.68719614e-01 8.69593546e-02 4.99347776e-01 1.57148153e-01
1.35644734e+00 3.46712232e-01 2.50054687e-01 -3.01425010e-01
9.98604298e-02 -3.97508629e-02 5.79895198e-01 8.00602615e-01
-5.27345598e-01 8.65312517e-01 3.79003346e-01 -5.39302349e-01
-8.95803690e-01 -1.09887707e+00 -3.49017888e-01 1.22731602e+00
-3.14277619e-01 -8.76032114e-01 -6.45673811e-01 -1.06290662e+00
3.58742267e-01 6.02413297e-01 -8.94380808e-01 1.22302078e-01
-4.99810092e-02 -9.72061157e-01 6.65546060e-01 2.59652585e-01
2.29961306e-01 -9.49104309e-01 -1.60146415e-01 -3.31341505e-01
-3.04472893e-01 -1.02278364e+00 -5.55818379e-01 5.65713763e-01
-6.40224218e-01 -1.35450518e+00 -5.46873510e-01 -4.36202675e-01
1.08264729e-01 2.40641177e-01 1.65382385e+00 -3.45857628e-02
-6.35258257e-01 1.55460492e-01 -5.64942837e-01 -2.68832326e-01
-7.20944479e-02 3.70421201e-01 -2.42656454e-01 -4.25250173e-01
4.62204158e-01 -5.57520390e-01 -5.84602594e-01 7.69256800e-02
-6.12237453e-01 -3.10402713e-03 3.42392206e-01 8.38129938e-01
9.13973272e-01 -3.17148119e-02 3.15442652e-01 -1.11054385e+00
5.62807560e-01 -4.93396372e-01 -5.12098134e-01 5.44747055e-01
-4.67636287e-01 -8.94652978e-02 6.77701473e-01 -2.18492970e-01
-8.61843884e-01 -3.50551367e-01 -9.64755565e-02 -6.46503031e-01
-5.98153591e-01 1.36154711e-01 -3.71387810e-01 3.60808596e-02
8.08011830e-01 -1.72945768e-01 -5.30223668e-01 -9.99342859e-01
5.12635708e-01 6.70739889e-01 5.30198157e-01 -5.81193984e-01
7.06093311e-01 2.40857214e-01 -3.42291981e-01 -3.56436342e-01
-1.01427424e+00 -3.82334590e-01 -4.78838712e-01 3.24656576e-01
9.48224008e-01 -1.18594134e+00 -6.12258196e-01 6.95365071e-01
-1.00327039e+00 -4.79496330e-01 -3.41447592e-01 1.85956582e-01
-4.45388287e-01 2.75916755e-01 -9.44252849e-01 -2.11051524e-01
-5.02916276e-01 -1.05507994e+00 1.06812334e+00 1.60452034e-02
-2.25430474e-01 -8.16762507e-01 2.23267928e-01 9.89226699e-01
2.90106148e-01 2.56849647e-01 9.22296584e-01 -5.67897856e-01
-5.28449953e-01 1.12447843e-01 -6.06999993e-01 2.45565474e-01
1.70685854e-02 1.53791845e-01 -1.32349372e+00 -4.11363721e-01
-4.57789600e-01 -7.46924877e-01 1.40281463e+00 3.22402745e-01
1.79702079e+00 -1.39319852e-01 -1.36763617e-01 1.10876000e+00
1.44496965e+00 -3.41047138e-01 5.78376293e-01 6.45206392e-01
8.29618216e-01 5.04039705e-01 6.61221325e-01 4.22012001e-01
4.39214885e-01 8.03641260e-01 5.57311118e-01 1.41758248e-01
-7.81519413e-01 3.37823071e-02 7.44573325e-02 4.03260201e-01
-1.24823656e-02 -5.37368715e-01 -6.63766980e-01 3.73766571e-01
-1.46747518e+00 -1.00086761e+00 -2.43230343e-01 2.10523629e+00
9.92598176e-01 -1.62829965e-01 1.17942043e-01 -7.95080736e-02
8.09041083e-01 2.13213235e-01 -4.96720016e-01 -2.94176966e-01
-4.39711958e-01 3.24072659e-01 3.04034889e-01 3.15157712e-01
-1.37954402e+00 1.03400815e+00 5.08824110e+00 1.55674863e+00
-1.12274885e+00 5.36725581e-01 1.01513159e+00 -4.55627531e-01
-3.10943037e-01 -1.66397244e-01 -1.12525332e+00 8.22969258e-01
8.86960924e-01 4.22548711e-01 6.13252282e-01 5.84071219e-01
-2.32575312e-01 1.50452733e-01 -9.14465368e-01 9.87776279e-01
-1.24407396e-01 -1.53380847e+00 3.68315913e-02 2.68015683e-01
5.35969138e-01 5.74579060e-01 5.15656844e-02 3.55713308e-01
2.38967165e-01 -1.06752002e+00 7.66440511e-01 2.31820777e-01
1.19353640e+00 -5.01113892e-01 5.57593584e-01 -1.25504673e-01
-1.00154698e+00 2.14146059e-02 -3.03014070e-01 2.78052837e-01
-2.08721682e-01 1.10553479e+00 -5.56148291e-01 4.12646174e-01
1.21231282e+00 5.20682216e-01 -1.04155493e+00 1.23976302e+00
-2.10390463e-01 6.55869007e-01 -9.60756391e-02 2.24743485e-01
-1.42725199e-01 2.44644210e-01 1.53518736e-01 1.51105237e+00
1.97228789e-01 -4.39514816e-01 -3.87257114e-02 6.24628246e-01
-2.06277370e-01 3.23629752e-02 -1.34680256e-01 -3.88495028e-01
6.39493465e-01 1.59733915e+00 -6.55960679e-01 -1.47828728e-01
-3.25592756e-01 1.01815760e+00 8.54462802e-01 3.00509483e-01
-9.42325413e-01 -1.00628421e-01 1.01064742e+00 2.86964983e-01
4.18618768e-01 1.21864378e-01 -2.70983279e-01 -1.47628856e+00
-9.19234157e-02 -1.32468760e+00 6.97818398e-01 -7.90216029e-01
-1.87954712e+00 5.62384129e-01 -4.50955719e-01 -7.76675522e-01
1.45937681e-01 -7.36839116e-01 -4.14650530e-01 8.14863205e-01
-1.59009588e+00 -1.57185996e+00 -6.26484573e-01 6.25591516e-01
5.71373820e-01 -7.41538852e-02 7.94586957e-01 5.45517385e-01
-5.95115364e-01 1.01439869e+00 2.41705999e-01 -1.08390197e-01
1.37743676e+00 -1.40659666e+00 7.32597530e-01 6.27955317e-01
2.77372748e-01 4.21719551e-01 5.66507995e-01 -7.29985118e-01
-9.34662342e-01 -1.37033081e+00 9.36774433e-01 -7.34194815e-01
6.70492351e-01 -6.49692893e-01 -7.56722510e-01 5.64737499e-01
1.33600384e-01 3.12697321e-01 8.56706858e-01 5.17969728e-01
-1.29675198e+00 3.12034972e-02 -1.34441662e+00 2.81278580e-01
1.17650247e+00 -6.49305999e-01 -1.97641805e-01 6.36594892e-01
9.93082523e-01 -6.21737912e-02 -1.04751301e+00 4.81113613e-01
7.83775985e-01 -1.09316790e+00 1.17405689e+00 -7.65632629e-01
5.88314414e-01 -2.01442882e-01 -6.07858896e-01 -1.39368868e+00
-9.11309421e-01 2.68211104e-02 -9.51132700e-02 1.43425250e+00
2.22209439e-01 -5.27431846e-01 8.77993405e-01 -3.85224372e-02
-2.18875986e-02 -6.25933707e-01 -8.77650201e-01 -8.15825760e-01
5.00207067e-01 -1.73817992e-01 5.31819642e-01 1.02402210e+00
-4.29822952e-01 2.69941598e-01 -2.45483458e-01 -1.28086343e-01
7.47403204e-01 2.69605964e-01 7.92913079e-01 -1.21087253e+00
-7.32879221e-01 -4.38119173e-01 -3.26181859e-01 -2.91067362e-01
1.13019459e-01 -1.11990750e+00 -3.64644915e-01 -1.34890115e+00
6.87101007e-01 -5.20306051e-01 -4.91879910e-01 6.73019648e-01
-4.43819463e-01 7.06980884e-01 6.28439844e-01 2.02289417e-01
-7.87967086e-01 1.75971359e-01 1.04414594e+00 -2.59384573e-01
3.18451583e-01 -2.95179278e-01 -7.02404678e-01 3.80596250e-01
1.05462599e+00 -3.15698385e-01 -3.72410640e-02 -5.28906524e-01
-3.50507982e-02 -4.32101876e-01 5.46203732e-01 -9.48974371e-01
-1.41854450e-01 -4.41211984e-02 6.46102309e-01 -6.71868980e-01
3.72316808e-01 -3.69949877e-01 4.15687978e-01 3.52075279e-01
-1.37639061e-01 -7.28074536e-02 2.66369313e-01 3.49148631e-01
-1.51555687e-01 -1.98538303e-02 7.86042154e-01 -2.71337688e-01
-7.62262166e-01 4.92734402e-01 3.56020629e-02 5.03779888e-01
7.16420352e-01 1.91973001e-02 -1.17596424e+00 1.39383718e-01
-7.27484822e-01 -2.40274100e-03 7.67652333e-01 5.54532409e-01
3.78326848e-02 -1.25868607e+00 -9.39968109e-01 9.18762162e-02
4.55070108e-01 -5.48376501e-01 5.14908850e-01 7.73482919e-01
-4.50959086e-01 7.45558858e-01 -4.99158740e-01 -2.35505193e-01
-1.56147265e+00 6.23013973e-01 1.71533927e-01 -7.49941111e-01
-9.81831476e-02 1.27099383e+00 7.02661932e-01 -8.37531626e-01
3.81754100e-01 -4.90864627e-02 1.78493038e-01 3.24450046e-01
5.68684697e-01 4.94503528e-01 4.34902281e-01 -1.55587450e-01
-5.96755087e-01 4.26744044e-01 -2.92972475e-01 2.35750780e-01
1.24144673e+00 5.62793538e-02 -8.26307312e-02 1.36527747e-01
1.26720881e+00 2.02273205e-01 -1.03334296e+00 1.86212152e-01
-8.67957622e-02 -5.67669928e-01 -9.32197273e-03 -1.62940311e+00
-1.11477828e+00 1.06447029e+00 7.89612949e-01 -1.90843776e-01
9.94821310e-01 5.99927232e-02 4.21748191e-01 -5.00023477e-02
4.51643199e-01 -6.56636357e-01 -1.42651305e-01 4.17160332e-01
9.88788426e-01 -1.22120798e+00 -2.76355855e-02 -5.99573076e-01
-2.76546180e-01 8.15950811e-01 8.21594894e-01 1.15564391e-01
5.28835893e-01 5.95530391e-01 1.42405838e-01 -3.45002227e-02
-1.18749034e+00 -2.40157083e-01 3.04019392e-01 7.56653965e-01
7.70046711e-01 3.48652810e-01 -3.55410159e-01 6.21556759e-01
-3.03411782e-01 2.27565020e-01 2.08469346e-01 4.24722642e-01
-4.69301380e-02 -1.20474899e+00 -4.57878530e-01 7.64336646e-01
-7.44369447e-01 -5.35556078e-01 -6.79802299e-01 6.44012392e-01
6.37888312e-01 8.84050965e-01 1.04240425e-01 -4.45757002e-01
9.66812000e-02 1.77887287e-02 7.80836999e-01 -4.66833532e-01
-1.02772748e+00 -2.86042020e-02 2.96108544e-01 -9.20608044e-01
1.88117512e-02 -5.74863613e-01 -6.60862029e-01 -5.38622856e-01
-1.64523169e-01 8.36527273e-02 7.41818368e-01 4.30127770e-01
6.87482238e-01 4.98744130e-01 1.67373553e-01 -5.37811518e-01
-5.26502311e-01 -1.05959404e+00 -4.96256053e-01 4.90071744e-01
1.43538728e-01 -5.94697356e-01 -7.26681232e-01 -1.00249499e-01] | [9.587434768676758, 2.2905683517456055] |
a5829321-4fde-4fc7-b607-7c72b65f8e76 | the-open-catalyst-2022-oc22-dataset-and | 2206.08917 | null | https://arxiv.org/abs/2206.08917v3 | https://arxiv.org/pdf/2206.08917v3.pdf | The Open Catalyst 2022 (OC22) Dataset and Challenges for Oxide Electrocatalysts | The development of machine learning models for electrocatalysts requires a broad set of training data to enable their use across a wide variety of materials. One class of materials that currently lacks sufficient training data is oxides, which are critical for the development of OER catalysts. To address this, we developed the OC22 dataset, consisting of 62,331 DFT relaxations (~9,854,504 single point calculations) across a range of oxide materials, coverages, and adsorbates. We define generalized total energy tasks that enable property prediction beyond adsorption energies; we test baseline performance of several graph neural networks; and we provide pre-defined dataset splits to establish clear benchmarks for future efforts. In the most general task, GemNet-OC sees a ~36% improvement in energy predictions when combining the chemically dissimilar OC20 and OC22 datasets via fine-tuning. Similarly, we achieved a ~19% improvement in total energy predictions on OC20 and a ~9% improvement in force predictions in OC22 when using joint training. We demonstrate the practical utility of a top performing model by capturing literature adsorption energies and important OER scaling relationships. We expect OC22 to provide an important benchmark for models seeking to incorporate intricate long-range electrostatic and magnetic interactions in oxide surfaces. Dataset and baseline models are open sourced, and a public leaderboard is available to encourage continued community developments on the total energy tasks and data. | ['C. Lawrence Zitnick', 'Edward H. Sargent', 'Oleksandr Voznyy', 'Jehad Abed', 'Felix Therrien', 'Brandon M. Wood', 'Zachary Ulissi', 'Anuroop Sriram', 'Nima Shoghi', 'Ammar Rizvi', 'Adeesh Kolluru', 'Javier Heras-Domingo', 'Abhishek Das', 'Siddharth Goyal', 'Muhammed Shuaibi', 'Janice Lan', 'Richard Tran'] | 2022-06-17 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [ 3.75514865e-01 2.50914302e-02 -3.80161613e-01 -1.09590411e-01
-8.72810364e-01 -2.97107965e-01 4.86952454e-01 3.81299824e-01
-3.20312321e-01 9.89012480e-01 1.38903260e-01 -5.85381925e-01
-2.77106851e-01 -1.00334954e+00 -9.65631247e-01 -8.25342178e-01
-1.65334389e-01 5.07331967e-01 2.33566225e-01 -6.95833623e-01
2.75616676e-01 6.44190669e-01 -1.31791437e+00 4.95603889e-01
1.05319452e+00 1.00029135e+00 3.74022931e-01 4.18501377e-01
4.10314739e-01 4.23103064e-01 -9.25627798e-02 -4.43422273e-02
4.47637103e-02 1.21514529e-01 -8.75652254e-01 -7.85486460e-01
2.51022398e-01 2.96442121e-01 -4.92577225e-01 7.81457305e-01
5.71498513e-01 3.19343776e-01 1.16174364e+00 -6.41743302e-01
-8.34331930e-01 6.26173556e-01 -2.91307092e-01 5.99302389e-02
9.85301286e-02 4.37540442e-01 1.43992639e+00 -8.97318721e-01
6.62157297e-01 7.02104151e-01 9.17084157e-01 7.16571510e-01
-1.43701684e+00 -6.81517839e-01 -2.11982384e-01 2.35933363e-01
-1.16546524e+00 -5.30955017e-01 6.56795204e-01 -2.19065174e-01
1.99840784e+00 4.62038100e-01 6.68097615e-01 1.10770059e+00
6.42500997e-01 7.96924904e-02 9.09187555e-01 -3.42299193e-01
1.34447441e-01 -1.89037174e-02 2.02197790e-01 4.22918409e-01
9.74814296e-01 6.03606850e-02 -5.37449121e-01 -3.96261424e-01
8.53202716e-02 -1.30998850e-01 -1.23080425e-01 -2.17934579e-01
-5.88168383e-01 6.34791613e-01 5.06707549e-01 2.04505712e-01
-4.08669114e-01 3.89345318e-01 2.71894395e-01 2.02425361e-01
4.85659480e-01 1.05269682e+00 -6.86057627e-01 -1.81369036e-02
-4.68781888e-01 4.69302386e-01 1.18468821e+00 7.34441042e-01
8.26109886e-01 2.29607224e-02 1.70694605e-01 9.84587789e-01
2.60380626e-01 5.04497826e-01 8.29334557e-02 -8.38496208e-01
6.68538034e-01 4.84195709e-01 1.89150557e-01 -5.65478444e-01
-4.38863575e-01 -7.41079897e-02 -5.15140295e-01 6.46593049e-02
-1.58055454e-01 -3.68452668e-01 -7.23629057e-01 1.47358382e+00
-6.21252209e-02 -6.00918829e-01 2.84036905e-01 2.76545584e-01
1.00650883e+00 7.53682137e-01 2.25237951e-01 -9.86491144e-03
8.87850523e-01 -8.05739105e-01 -2.57134020e-01 -2.37597197e-01
9.39463317e-01 -3.68740380e-01 8.71305466e-01 3.88413280e-01
-1.36589646e+00 -1.20063253e-01 -1.37785673e+00 -3.77407908e-01
-7.73528039e-01 -2.49681398e-01 1.02776480e+00 5.32097101e-01
-1.07603979e+00 1.29214358e+00 -9.32168007e-01 -3.48146886e-01
3.15548956e-01 8.50619733e-01 -2.75371760e-01 -2.57389218e-01
-1.36972046e+00 1.09264326e+00 3.57725471e-01 -3.30146998e-01
-7.36116648e-01 -9.54454005e-01 -5.66215277e-01 2.38529310e-01
2.91144907e-01 -4.78080124e-01 9.41436350e-01 -1.67629659e-01
-1.31345844e+00 4.72384661e-01 -1.21522442e-01 -4.29391205e-01
1.89506814e-01 4.89138626e-02 -7.11688817e-01 -1.62502348e-01
-5.31449206e-02 7.37009406e-01 1.37759820e-01 -8.32561791e-01
-6.60875291e-02 -1.70463100e-02 -1.97054639e-01 2.16115534e-01
-6.77872658e-01 -4.09986138e-01 -7.04599097e-02 -1.05584063e-01
-2.63830096e-01 -1.03468275e+00 -3.35540503e-01 -6.68764114e-01
-4.70245361e-01 -4.89880294e-01 4.02979583e-01 -3.57020468e-01
9.96546865e-01 -1.55856061e+00 7.27036688e-03 4.16737497e-01
3.79340440e-01 8.72118305e-03 -8.52483884e-02 7.82471538e-01
-2.91666716e-01 4.88929063e-01 -1.79134101e-01 3.97458598e-02
7.14057311e-02 -2.23065108e-01 1.19063720e-01 6.55673444e-02
3.63006532e-01 1.07018566e+00 -5.98187804e-01 6.55865576e-03
4.98344451e-02 5.95422745e-01 -9.48498607e-01 -2.69589663e-01
-4.85429913e-01 -1.35778204e-01 -4.46651816e-01 7.39652514e-01
5.07982194e-01 -7.61513770e-01 3.87432963e-01 -4.17978078e-01
-3.73492777e-01 8.65955591e-01 -4.80273455e-01 1.38318968e+00
-1.44153178e-01 5.34958184e-01 -8.50696266e-02 -8.90038610e-01
8.63281906e-01 1.50765717e-01 1.06599188e+00 -8.39757562e-01
-7.71655887e-02 5.69496512e-01 4.87125248e-01 -4.52246249e-01
7.82752275e-01 -2.64189571e-01 1.12306923e-01 3.33741009e-02
-1.00954689e-01 -3.37232649e-01 3.27238619e-01 1.43233091e-01
1.56737161e+00 -1.80803195e-01 -9.66554657e-02 -1.05965173e+00
9.89862680e-02 4.07897413e-01 1.47259444e-01 7.74138570e-01
1.76238671e-01 3.27410758e-01 3.42033982e-01 -4.61800903e-01
-1.45703936e+00 -5.42669952e-01 -5.92604220e-01 1.05832803e+00
-5.51508330e-02 -6.81374311e-01 -5.89167893e-01 -5.90878874e-02
3.93827498e-01 7.48766899e-01 -5.91227055e-01 -4.73819733e-01
-2.44009390e-01 -1.41372490e+00 5.22083580e-01 6.53363943e-01
2.66805470e-01 -1.01219022e+00 1.69261530e-01 4.96348053e-01
2.55054951e-01 -7.94274688e-01 -1.57124832e-01 6.88137710e-01
-8.18142593e-01 -1.18839967e+00 -3.17609727e-01 -5.08258343e-01
2.14166805e-01 2.64106542e-01 1.10188377e+00 1.77228153e-01
-3.57288152e-01 2.07307369e-01 1.32691562e-01 -5.64857483e-01
-7.04669833e-01 8.20951641e-01 2.83469796e-01 -8.33417475e-01
5.74937880e-01 -5.99840879e-01 -8.37184370e-01 2.11315915e-01
-6.03674650e-01 8.36707205e-02 6.15942776e-01 5.08604348e-01
6.37638390e-01 -2.94475377e-01 6.74947619e-01 -9.54277039e-01
8.30623150e-01 -7.66771972e-01 -3.85843158e-01 3.88011783e-01
-1.07725549e+00 3.57581705e-01 7.02917635e-01 -1.96089700e-01
-9.99681473e-01 -3.64511192e-01 -4.59009022e-01 3.66801023e-02
1.91382989e-01 6.96760714e-01 -3.40117514e-03 -4.05895174e-01
1.16384161e+00 -1.61730185e-01 -1.58537298e-01 -3.53107333e-01
-6.26440197e-02 6.20440066e-01 1.62284106e-01 -1.22404718e+00
3.54278237e-01 1.82129920e-01 4.06252921e-01 -1.18904662e+00
-5.79648316e-01 -2.08560735e-01 8.89932830e-03 3.76189768e-01
8.91000211e-01 -9.70814466e-01 -1.19941199e+00 7.42109492e-02
-8.27513456e-01 -7.59651065e-01 -7.27047026e-02 3.35219294e-01
-5.34218661e-02 1.68643415e-01 -6.55153155e-01 -4.97259140e-01
-8.80874157e-01 -1.31313181e+00 8.20535660e-01 5.59988283e-02
-1.11779660e-01 -1.39107907e+00 1.84947431e-01 3.08834046e-01
7.34235585e-01 3.36844712e-01 1.24288702e+00 -7.33277977e-01
-4.80792522e-01 -1.67850599e-01 -1.72801942e-01 3.24802756e-01
3.05856496e-01 6.95720911e-02 -8.20705712e-01 -5.41716576e-01
-3.38183701e-01 -6.20006442e-01 1.42166948e+00 4.56108093e-01
1.45658052e+00 7.02297688e-02 -7.53813028e-01 4.73538131e-01
1.39045632e+00 2.46463120e-01 7.44957089e-01 7.31349513e-02
1.14079499e+00 1.34563550e-01 -1.82463247e-02 -8.87961015e-02
2.99431682e-01 3.23712975e-01 2.98076212e-01 -2.82807052e-02
2.15021268e-01 -1.12198986e-01 3.89241636e-01 1.00503612e+00
-7.43778944e-01 -4.26240206e-01 -1.12767303e+00 2.87703499e-02
-1.33403838e+00 -8.50610733e-01 -3.17500979e-01 2.09261799e+00
8.10467482e-01 4.29266006e-01 -1.60297617e-01 -3.19770217e-01
4.45334435e-01 1.82956532e-01 -1.31815016e+00 -3.31393898e-01
-1.20619722e-01 6.66215181e-01 1.00860465e+00 5.26175559e-01
-6.55000448e-01 7.98463106e-01 7.44224977e+00 8.80207062e-01
-1.35761213e+00 -9.54323709e-02 8.75685632e-01 -2.81457126e-01
-7.54685819e-01 5.06410263e-02 -1.04684865e+00 2.71323085e-01
1.28892219e+00 -6.66762963e-02 7.52590179e-01 6.88230872e-01
-2.43047848e-02 2.27923796e-01 -1.29738295e+00 5.51830649e-01
-2.76267499e-01 -1.92836630e+00 -1.29012629e-01 5.65455556e-01
1.18762064e+00 9.76231694e-01 -1.76766634e-01 4.08347726e-01
4.55309540e-01 -1.31173003e+00 2.10534856e-01 5.01619518e-01
1.23038125e+00 -4.91190284e-01 1.46142527e-01 -1.71825960e-01
-1.17343378e+00 1.92668676e-01 -8.10128629e-01 -1.78999215e-01
-3.99213552e-01 4.50028092e-01 -6.93677783e-01 7.01248646e-01
6.62313700e-01 9.54163432e-01 -3.45243037e-01 5.09564340e-01
3.95088732e-01 5.86710215e-01 -6.36171460e-01 -5.76065600e-01
7.93563277e-02 -3.02706301e-01 2.86333710e-01 1.09518588e+00
3.21489424e-01 1.33483365e-01 -2.05935612e-01 1.06276524e+00
-7.98096001e-01 -5.30173033e-02 -7.03432620e-01 -3.66697162e-01
7.28554487e-01 1.39304209e+00 -3.73974651e-01 -1.62351355e-01
-4.40987885e-01 2.81104416e-01 5.00012755e-01 4.72479135e-01
-8.22080076e-01 -4.37981784e-01 7.01320827e-01 3.56118709e-01
-4.61066775e-02 -9.15352404e-02 -2.08458126e-01 -1.14799023e+00
-1.41419113e-01 -5.29951215e-01 8.18856880e-02 -7.73959935e-01
-1.50718248e+00 1.04420662e-01 3.11146639e-02 -2.15841860e-01
2.11708918e-01 -1.43646574e+00 -8.82911921e-01 8.90121996e-01
-1.52270734e+00 -6.49970472e-01 -2.99136877e-01 2.71747038e-02
1.83232218e-01 -2.13387031e-02 6.89103782e-01 2.52474517e-01
-5.65349579e-01 3.50175560e-01 6.46683812e-01 -4.87914920e-01
6.93021059e-01 -1.31311464e+00 6.89210355e-01 -4.51628119e-03
-3.48211139e-01 7.48713553e-01 6.64431751e-01 -9.06637430e-01
-1.87956369e+00 -1.13554811e+00 3.71768355e-01 -7.31307328e-01
7.22226799e-01 -4.80779767e-01 -1.12524247e+00 6.24010921e-01
2.25259379e-01 -2.58272707e-01 5.86477935e-01 4.12491471e-01
6.73587173e-02 -1.22567229e-01 -8.93993676e-01 6.11029744e-01
1.45918703e+00 -5.72004855e-01 -9.11041126e-02 8.17719221e-01
8.92317533e-01 -4.37161326e-01 -1.62343764e+00 6.94524527e-01
5.25347948e-01 -7.53775299e-01 1.06541991e+00 -7.59688377e-01
7.20298231e-01 4.42687482e-01 -1.56535879e-01 -1.03583968e+00
-4.85537946e-01 -6.12874687e-01 -1.51717380e-01 7.43091345e-01
1.08684206e+00 -1.01153994e+00 9.12708402e-01 1.16588318e+00
-7.04531789e-01 -1.25810099e+00 -5.39722919e-01 -5.72402835e-01
7.66610444e-01 -3.47775608e-01 5.53013086e-01 9.07396972e-01
1.88003093e-01 5.00740588e-01 -4.12721559e-02 -1.64882213e-01
2.66829878e-01 -2.54041612e-01 2.89765269e-01 -1.47765613e+00
-1.62077114e-01 -4.09085542e-01 3.08941334e-01 -7.67700672e-01
6.39876053e-02 -1.38881457e+00 -3.08434784e-01 -1.60769224e+00
4.44941312e-01 -7.67791629e-01 -6.79649055e-01 5.61673284e-01
-1.07769296e-02 7.02978820e-02 -4.11713481e-01 2.87816316e-01
-6.98667288e-01 7.46291459e-01 1.09791887e+00 -2.05773234e-01
-2.90187091e-01 -4.62041020e-01 -1.08370674e+00 3.31265420e-01
8.70420039e-01 -6.16101384e-01 -1.38739631e-01 -3.00721169e-01
7.20690429e-01 -2.45332822e-01 1.14284717e-01 -1.27203071e+00
1.25758097e-01 -2.82339990e-01 5.85744023e-01 -2.09576130e-01
4.99708027e-01 -3.68351310e-01 3.73313278e-01 4.82497483e-01
-3.31844866e-01 -4.02516127e-02 3.76237273e-01 3.87588769e-01
4.53898638e-01 -2.16379330e-01 5.08908033e-01 -2.98225135e-01
-1.99153483e-01 5.92211246e-01 -1.03271514e-01 1.87823907e-01
5.69007158e-01 -2.17344254e-01 -5.13364136e-01 7.09387958e-02
-4.66414094e-01 2.91435599e-01 8.10931325e-01 2.03169748e-01
5.72836921e-02 -1.06418467e+00 -2.38626316e-01 -2.09285051e-01
6.43775472e-03 8.67806301e-02 2.39734590e-01 5.60718298e-01
-5.93167961e-01 6.45619452e-01 -3.28703448e-02 -3.07288855e-01
-7.74468899e-01 2.45117083e-01 6.68966591e-01 -4.06794548e-01
-2.75595337e-01 4.56110537e-01 -1.00869620e-02 -3.68565768e-01
-5.09089530e-01 -5.16173840e-01 1.67487249e-01 -4.68807846e-01
-1.07486695e-01 6.83860362e-01 5.19507110e-01 -1.21465653e-01
-2.57865280e-01 1.91193745e-01 -2.73863375e-01 3.28512788e-01
1.88632834e+00 5.31488657e-01 -5.60449772e-02 1.96391344e-01
1.34131455e+00 8.48170146e-02 -1.40623534e+00 2.50154793e-01
-9.50892940e-02 5.07701337e-01 -9.00662765e-02 -6.24094725e-01
-6.47506058e-01 7.33932734e-01 3.57788891e-01 4.83969957e-01
5.05865097e-01 7.77651072e-02 5.50375104e-01 1.21977758e+00
1.09279647e-01 -1.30498588e+00 1.71072781e-01 5.90674400e-01
7.11529851e-01 -1.25485373e+00 3.16981733e-01 -2.84732342e-01
-3.70313525e-01 1.03638947e+00 8.11914325e-01 -7.38400370e-02
5.16638160e-01 1.89341843e-01 -7.69187987e-01 -8.02783847e-01
-1.10618341e+00 4.42366987e-01 2.93184221e-01 2.61213660e-01
6.80152059e-01 1.61475409e-02 -2.41350278e-01 5.50062239e-01
-1.60370827e-01 -3.59332919e-01 2.29992807e-01 8.82683039e-01
-7.08425045e-01 -1.31208694e+00 2.19706625e-01 9.72327888e-01
-5.87097347e-01 -4.91132677e-01 -4.82648432e-01 9.52421606e-01
-2.65448004e-01 7.51792610e-01 -1.19554810e-01 -3.22435319e-01
3.57021809e-01 2.29455158e-01 4.88893241e-01 -5.60410380e-01
-2.79341578e-01 -3.65157068e-01 5.31324923e-01 -1.66026711e-01
-2.61509329e-01 -4.77048665e-01 -1.40143037e+00 -6.04495585e-01
-7.35960186e-01 3.63898396e-01 6.90523922e-01 8.80299211e-01
7.57063985e-01 8.03264201e-01 2.13490024e-01 -1.26774371e+00
-5.65848291e-01 -1.14490616e+00 -3.62357467e-01 4.68742490e-01
-9.56650302e-02 -7.09035456e-01 -3.02814454e-01 -5.37070930e-01] | [5.234614372253418, 5.498034954071045] |
782734ab-dddd-43d7-bab9-0ac19fd78213 | perceptual-based-deep-learning-denoiser-as-a | 2107.05222 | null | https://arxiv.org/abs/2107.05222v1 | https://arxiv.org/pdf/2107.05222v1.pdf | Perceptual-based deep-learning denoiser as a defense against adversarial attacks on ASR systems | In this paper we investigate speech denoising as a defense against adversarial attacks on automatic speech recognition (ASR) systems. Adversarial attacks attempt to force misclassification by adding small perturbations to the original speech signal. We propose to counteract this by employing a neural-network based denoiser as a pre-processor in the ASR pipeline. The denoiser is independent of the downstream ASR model, and thus can be rapidly deployed in existing systems. We found that training the denoisier using a perceptually motivated loss function resulted in increased adversarial robustness without compromising ASR performance on benign samples. Our defense was evaluated (as a part of the DARPA GARD program) on the 'Kenansville' attack strategy across a range of attack strengths and speech samples. An average improvement in Word Error Rate (WER) of about 7.7% was observed over the undefended model at 20 dB signal-to-noise-ratio (SNR) attack strength. | ['Shrikanth Narayanan', 'Dillon Knox', 'Raghuveer Peri', 'Nicholas Mehlman', 'Anirudh Sreeram'] | 2021-07-12 | null | null | null | null | ['speech-denoising'] | ['speech'] | [ 6.34589553e-01 3.69912177e-01 6.29997730e-01 -3.09882909e-01
-1.20330203e+00 -8.75498891e-01 5.87530315e-01 -1.40220791e-01
-4.18557703e-01 2.52338946e-01 4.29215401e-01 -9.43866193e-01
2.73862869e-01 -3.54880393e-01 -5.49293399e-01 -6.43091559e-01
2.28417162e-02 -1.08940750e-01 9.69192609e-02 -6.30762160e-01
-1.58620849e-01 8.99702966e-01 -8.95242333e-01 4.12749946e-01
4.13467646e-01 6.03124321e-01 -2.93016583e-01 1.48225307e+00
5.41273057e-01 6.09131575e-01 -1.54256845e+00 -4.76956904e-01
6.72321558e-01 -3.97176445e-01 -3.88548821e-01 -6.04076572e-02
3.67518753e-01 -3.07654977e-01 -1.08066666e+00 1.47985935e+00
8.79093111e-01 1.04099177e-01 1.07530974e-01 -8.07111442e-01
-4.58416104e-01 6.33757174e-01 -2.92839087e-03 5.18951535e-01
3.52019072e-01 4.99052972e-01 6.23110712e-01 -4.36434329e-01
6.13417923e-02 1.56097531e+00 5.90941668e-01 8.48534644e-01
-1.33021224e+00 -8.76166344e-01 -8.26343447e-02 -1.38598204e-01
-1.35482013e+00 -1.25770330e+00 7.18986690e-01 2.40536883e-01
1.04406023e+00 7.80457377e-01 -5.74151007e-03 1.29011393e+00
1.32891476e-01 4.06066120e-01 9.25084889e-01 -4.28213924e-01
2.84985781e-01 -2.44532645e-01 -2.97254950e-01 1.74985945e-01
5.91056347e-02 6.30753994e-01 -3.70815754e-01 -4.50798750e-01
3.93221796e-01 -7.68805981e-01 -4.07517642e-01 5.43989241e-01
-6.26346350e-01 5.89112639e-01 4.16763842e-01 1.92448243e-01
-3.21788311e-01 5.11228800e-01 3.59180689e-01 7.71349669e-01
5.48010767e-01 5.16848505e-01 -3.45344037e-01 -6.21675625e-02
-8.40442479e-01 6.29521757e-02 7.30345428e-01 4.71942782e-01
-1.29177794e-01 1.07108915e+00 1.15110530e-02 9.74096596e-01
5.77636600e-01 8.46434772e-01 4.17739898e-01 -7.72441566e-01
5.59626222e-01 -4.85682130e-01 -5.51884435e-02 -8.24875951e-01
1.71664685e-01 -6.25071526e-01 -5.61185539e-01 6.90080285e-01
3.07505131e-01 -4.79574412e-01 -1.52511907e+00 1.67649651e+00
-3.46210226e-02 1.65036321e-01 5.41809320e-01 8.10994625e-01
2.02964529e-01 7.31562018e-01 4.63882647e-02 -1.47243440e-01
1.00593424e+00 -4.07679260e-01 -9.21892047e-01 -5.59393585e-01
2.09114030e-01 -1.27076423e+00 7.11578131e-01 7.10370362e-01
-1.21055460e+00 -3.33643943e-01 -1.53787541e+00 4.30676997e-01
-2.32216582e-01 -3.97588432e-01 -3.02440654e-02 1.64416265e+00
-1.22924864e+00 4.24330950e-01 -7.42834210e-01 1.20421626e-01
1.05556048e-01 3.72086436e-01 -3.43664825e-01 -1.91450100e-02
-1.61742914e+00 9.17412281e-01 -1.14327259e-02 2.93650270e-01
-1.30351412e+00 -5.43095112e-01 -7.22591102e-01 4.62531857e-02
2.27872208e-02 -2.08416358e-01 1.61912274e+00 -1.06304181e+00
-1.63270283e+00 6.39396727e-01 -4.59566712e-02 -9.59482789e-01
7.70504415e-01 -1.82489470e-01 -1.40561283e+00 2.18492195e-01
-5.02183616e-01 9.51982439e-02 1.58447552e+00 -1.17418814e+00
-2.82704830e-01 -1.94585219e-01 -7.63395280e-02 5.87703176e-02
-1.39110833e-01 4.25367594e-01 5.80432303e-02 -1.34827900e+00
2.45806262e-01 -9.20658827e-01 -3.04031760e-01 -5.20505071e-01
-3.82516056e-01 5.27685463e-01 1.09760368e+00 -1.09436524e+00
1.20179021e+00 -2.39483976e+00 -3.56272638e-01 5.69462836e-01
-1.07928559e-01 1.08342934e+00 -1.47979543e-01 2.45579392e-01
-6.67502284e-01 3.06250721e-01 -1.87904119e-01 -2.13929474e-01
5.03292978e-02 1.32955045e-01 -7.04769313e-01 7.46159196e-01
2.64738183e-02 3.50314558e-01 -7.43165612e-01 4.65958565e-01
2.46571407e-01 6.86567366e-01 -3.61515909e-01 2.40446717e-01
9.52756926e-02 -9.50246379e-02 4.58716713e-02 5.38782418e-01
6.72601998e-01 8.43139172e-01 7.90789053e-02 -1.19437980e-04
4.11772579e-01 6.26155913e-01 -1.06555462e+00 9.62746084e-01
-5.61377525e-01 9.62631464e-01 9.20762897e-01 -6.38362467e-01
1.00807226e+00 7.52021790e-01 -3.52240980e-01 -4.78406549e-01
2.33574554e-01 2.84165323e-01 5.09133160e-01 9.73375663e-02
2.68589169e-01 -3.64311934e-01 -9.24989730e-02 9.25200135e-02
-1.77118108e-01 -2.48373970e-01 -4.62925196e-01 2.89312959e-01
1.65086055e+00 -8.21489632e-01 -7.70278368e-03 -1.14833631e-01
5.86686671e-01 -4.07645941e-01 3.55948359e-01 9.63800371e-01
-5.62602162e-01 5.37177980e-01 6.14480786e-02 -4.36003841e-02
-1.26487923e+00 -1.47323895e+00 2.77631008e-03 7.23912537e-01
-3.61938626e-01 -1.26516581e-01 -9.79424298e-01 -5.11717021e-01
-2.32499260e-02 1.18018234e+00 -4.23075147e-02 -5.35508931e-01
-7.39746869e-01 -5.32931685e-01 1.62931073e+00 3.38903904e-01
4.46318179e-01 -6.98953748e-01 3.05642039e-01 2.45393798e-01
2.83716410e-01 -1.15789437e+00 -7.14781642e-01 4.67736274e-01
-4.25247848e-01 -5.34406304e-01 -3.87011886e-01 -4.58098739e-01
5.72938204e-01 2.04055339e-01 5.39432943e-01 -3.37499790e-02
-4.55726609e-02 3.18660975e-01 -3.82655174e-01 -5.20728588e-01
-1.57758415e+00 -3.67988586e-01 5.17143786e-01 -4.21344116e-02
1.58463046e-01 -4.37590182e-01 -2.25448534e-01 4.61332083e-01
-1.01554728e+00 -9.09637749e-01 4.08598989e-01 7.43618071e-01
-8.10518190e-02 4.97171611e-01 8.46017420e-01 -6.73705459e-01
1.10030472e+00 -1.41789079e-01 -5.58631003e-01 -3.12610343e-02
-3.73804212e-01 -1.20276483e-02 8.88181806e-01 -5.35645723e-01
-1.15102899e+00 -3.67070474e-02 -9.97011006e-01 -4.88244891e-01
-2.99111694e-01 1.17025316e-01 -7.26292074e-01 -3.72021347e-01
1.00353205e+00 1.60942376e-01 7.17234388e-02 -2.98085004e-01
4.67171133e-01 1.18000281e+00 8.85146618e-01 -2.87353665e-01
1.46121216e+00 6.45853505e-02 -3.12918305e-01 -1.21349478e+00
-3.38216245e-01 -2.51485556e-01 2.39385054e-01 -8.60121623e-02
3.14360797e-01 -9.25811589e-01 -4.54590142e-01 6.74799025e-01
-1.04676831e+00 -1.71874240e-01 5.01907617e-02 4.06891018e-01
-1.78314522e-01 5.30557811e-01 -6.24276161e-01 -1.03276980e+00
-4.15088505e-01 -1.01565540e+00 6.05895519e-01 -2.17271864e-01
-3.57612371e-01 -8.06293547e-01 -3.23451906e-01 6.01108313e-01
7.70316541e-01 -4.61191460e-02 4.94898498e-01 -1.11460626e+00
-1.03075087e-01 -7.57361293e-01 2.79636025e-01 1.18306470e+00
2.30590686e-01 -3.77924629e-02 -1.44414818e+00 -6.41066074e-01
5.63717306e-01 1.06619380e-01 5.69662571e-01 1.06862999e-01
7.55112767e-01 -6.05086446e-01 2.49723226e-01 4.61086780e-01
9.30650830e-01 6.54070497e-01 1.13036263e+00 5.92010766e-02
4.11496520e-01 1.25345916e-01 1.76124841e-01 -1.90282110e-02
-7.43691564e-01 5.02339303e-01 3.68222386e-01 2.78218724e-02
-4.03413355e-01 -2.23696113e-01 8.89622569e-01 6.60097837e-01
4.29434121e-01 -6.76404774e-01 -9.01415169e-01 4.59978461e-01
-9.76133645e-01 -1.12593031e+00 2.37698182e-02 2.20675302e+00
7.58519173e-01 7.80958235e-01 -8.72355476e-02 5.94175160e-01
1.00755310e+00 5.82568228e-01 -3.78100425e-01 -1.01750982e+00
-2.29330957e-01 5.21177888e-01 1.07025790e+00 1.19344079e+00
-1.01843905e+00 1.10988247e+00 6.83719063e+00 1.05017388e+00
-1.09431303e+00 7.87715241e-02 4.98802811e-01 -8.17776918e-02
-1.68018386e-01 -2.68704861e-01 -4.94951785e-01 3.38757098e-01
1.57600713e+00 -2.41139665e-01 6.93637848e-01 6.79577589e-01
4.55898821e-01 5.09851515e-01 -6.50752544e-01 5.90864837e-01
1.06534429e-01 -9.49308455e-01 -6.75101802e-02 2.95385588e-02
1.27715066e-01 4.76065353e-02 3.75969082e-01 1.31376818e-01
7.47941792e-01 -1.28357816e+00 6.49905980e-01 -2.77441610e-02
8.34948301e-01 -1.09351850e+00 5.43032587e-01 7.86767602e-02
-8.12495649e-01 -4.92680408e-02 -1.08012401e-01 5.76343648e-02
1.71211138e-01 4.32981521e-01 -1.29160964e+00 1.06801674e-01
3.33324194e-01 -3.27457964e-01 -8.34411979e-02 5.66777587e-01
-5.97366154e-01 1.24865818e+00 -4.54027474e-01 2.31552705e-01
2.39781946e-01 5.68583250e-01 1.18980289e+00 1.39469707e+00
-8.91803671e-03 2.24700853e-01 -3.15472811e-01 2.34268695e-01
-3.03287238e-01 -5.62341809e-01 -5.03728747e-01 -1.60174325e-01
8.25214088e-01 6.56220734e-01 -3.12626541e-01 -2.66313367e-02
2.55097747e-01 1.03723490e+00 -4.54733551e-01 5.19802570e-01
-8.02748561e-01 -8.38067830e-01 1.18258095e+00 5.20185828e-02
3.06889176e-01 -3.53634536e-01 -1.73209876e-01 -7.38851070e-01
-9.75385234e-02 -1.56993389e+00 -8.12343955e-02 -6.40107572e-01
-1.19110322e+00 7.77489483e-01 -3.35423142e-01 -8.93266380e-01
-2.46499628e-01 -4.75493670e-01 -6.40804291e-01 1.33847916e+00
-8.88588369e-01 -7.76826680e-01 4.03123438e-01 5.41449785e-01
6.26467109e-01 -6.54520929e-01 8.03165436e-01 2.80479491e-01
-3.57324570e-01 1.18065357e+00 1.08433180e-01 2.72613525e-01
4.33960587e-01 -1.17245471e+00 1.34616554e+00 1.70146704e+00
1.58597261e-01 8.18988860e-01 1.30932403e+00 -7.30170965e-01
-1.24828255e+00 -1.15865588e+00 5.82587779e-01 -2.85909563e-01
8.69358599e-01 -5.60016036e-01 -1.01042593e+00 5.71800411e-01
4.07322019e-01 -3.86120714e-02 4.39456284e-01 -3.10041159e-01
-7.17203259e-01 -3.84606868e-02 -1.58059251e+00 7.62120008e-01
7.39109874e-01 -1.02138150e+00 -8.02431047e-01 1.34145081e-01
9.88812447e-01 -5.01488328e-01 -5.55513978e-01 1.71263441e-01
1.62938520e-01 -5.16475081e-01 1.25178492e+00 -7.41448998e-01
-2.76085198e-01 -6.28922820e-01 -6.11924291e-01 -1.67186034e+00
-1.10306114e-01 -1.46170163e+00 -4.59196046e-02 1.16384494e+00
5.32158971e-01 -7.87927151e-01 7.05997229e-01 4.54870105e-01
-4.51109052e-01 -3.69585827e-02 -1.15924776e+00 -1.09127629e+00
-2.11240184e-02 -9.74580288e-01 2.61932760e-01 6.11501873e-01
-3.78701001e-01 8.18797350e-02 -4.68492627e-01 1.13288641e+00
6.14237547e-01 -1.26357269e+00 5.81702828e-01 -2.56370902e-01
-5.22934437e-01 -1.96788967e-01 -7.36903131e-01 -9.26953554e-01
7.20665678e-02 -6.57703996e-01 2.22944960e-01 -8.32484961e-01
-9.78173137e-01 -6.03679158e-02 -1.96433634e-01 3.18917930e-01
-2.13866368e-01 7.76777640e-02 3.23950589e-01 -2.59480327e-01
1.81108862e-01 1.82581589e-01 4.10258085e-01 -3.41000110e-01
1.78400859e-01 4.30800974e-01 -7.28518128e-01 7.66749799e-01
1.04572940e+00 -8.20272565e-01 -4.63702172e-01 -3.69073898e-01
-3.94128323e-01 3.22842151e-02 2.77579546e-01 -1.19314766e+00
1.97051451e-01 2.38403440e-01 1.72973186e-01 -1.23723581e-01
5.43787956e-01 -8.38661432e-01 2.15209469e-01 6.23854876e-01
-5.59431851e-01 -5.28919958e-02 4.74358708e-01 6.78840458e-01
-2.67766297e-01 -2.47105345e-01 1.21584165e+00 2.09114984e-01
-2.11017534e-01 -3.85625094e-01 -7.69890606e-01 -5.42813912e-02
7.61461556e-01 -8.78861845e-02 -5.21252751e-01 -7.64954627e-01
-8.36023152e-01 -3.46688718e-01 2.07319751e-01 3.62388223e-01
6.86751544e-01 -8.56597364e-01 -8.65220904e-01 5.17807186e-01
-4.40153927e-01 -6.91340089e-01 1.12646222e-01 1.39934069e-03
-7.64723301e-01 1.87596619e-01 2.96087325e-01 -8.76500383e-02
-1.86551583e+00 4.35840130e-01 7.18464673e-01 1.44117206e-01
-3.09952140e-01 1.11056507e+00 -3.83128107e-01 -8.08814093e-02
4.91617233e-01 4.58283275e-02 6.30509019e-01 -5.72105527e-01
7.57482350e-01 3.54316920e-01 5.70292950e-01 -6.81873262e-01
-3.85199964e-01 -1.86353177e-01 -2.56955713e-01 -5.67897081e-01
9.22729790e-01 6.10502884e-02 3.23110312e-01 -1.88549981e-01
1.30377460e+00 5.33587158e-01 -9.38217580e-01 6.57539722e-03
-1.57800809e-01 -6.11540020e-01 4.77972776e-01 -1.07064724e+00
-8.56855273e-01 6.71289504e-01 8.40948224e-01 7.11696208e-01
1.15964162e+00 -5.41327894e-01 9.31551635e-01 4.11725998e-01
-1.11163566e-02 -9.51309979e-01 -1.54123932e-01 5.21836698e-01
1.02316594e+00 -7.87204027e-01 -1.75126016e-01 -1.53879747e-01
-5.24644732e-01 1.02903104e+00 -1.78745370e-02 -2.25749537e-01
6.70848906e-01 7.05625713e-01 6.14243805e-01 2.99663067e-01
-5.35396576e-01 1.85650751e-01 6.83097839e-02 8.41093183e-01
8.47908333e-02 1.47730306e-01 1.04793765e-01 1.01204403e-01
-5.29385865e-01 -7.01462626e-01 6.73911810e-01 9.12717342e-01
-6.58275843e-01 -1.17287934e+00 -1.03645313e+00 1.27617076e-01
-9.61156428e-01 -3.73674124e-01 -6.23152375e-01 4.11151260e-01
-3.29641968e-01 1.58731782e+00 -3.23990583e-01 -6.60616338e-01
6.37139440e-01 -5.77036999e-02 5.33299288e-03 -4.73359585e-01
-8.93392384e-01 2.18516588e-01 6.50415480e-01 -3.66320699e-01
2.52679646e-01 -4.96116549e-01 -9.93179619e-01 -6.70942426e-01
-1.07855231e-01 7.89579153e-02 1.04956055e+00 7.16966808e-01
2.18624860e-01 9.15768862e-01 9.81974363e-01 -3.83207917e-01
-1.16431618e+00 -8.84183526e-01 -3.81166041e-01 2.22902745e-01
7.57015944e-01 5.09859398e-02 -9.50187743e-01 -2.87608784e-02] | [14.115492820739746, 5.849719047546387] |
f3ecfb51-3246-42e3-8bde-7cf3fae1d645 | self-supervised-amodal-video-object | 2210.12733 | null | https://arxiv.org/abs/2210.12733v1 | https://arxiv.org/pdf/2210.12733v1.pdf | Self-supervised Amodal Video Object Segmentation | Amodal perception requires inferring the full shape of an object that is partially occluded. This task is particularly challenging on two levels: (1) it requires more information than what is contained in the instant retina or imaging sensor, (2) it is difficult to obtain enough well-annotated amodal labels for supervision. To this end, this paper develops a new framework of Self-supervised amodal Video object segmentation (SaVos). Our method efficiently leverages the visual information of video temporal sequences to infer the amodal mask of objects. The key intuition is that the occluded part of an object can be explained away if that part is visible in other frames, possibly deformed as long as the deformation can be reasonably learned. Accordingly, we derive a novel self-supervised learning paradigm that efficiently utilizes the visible object parts as the supervision to guide the training on videos. In addition to learning type prior to complete masks for known types, SaVos also learns the spatiotemporal prior, which is also useful for the amodal task and could generalize to unseen types. The proposed framework achieves the state-of-the-art performance on the synthetic amodal segmentation benchmark FISHBOWL and the real world benchmark KINS-Video-Car. Further, it lends itself well to being transferred to novel distributions using test-time adaptation, outperforming existing models even after the transfer to a new distribution. | ['Zheng Zhang', 'Yanwei Fu', 'David Wipf', 'Francesco Locatello', 'Tong He', 'Tianjun Xiao', 'Chiyu Wang', 'Yuxin Hong', 'Jian Yao'] | 2022-10-23 | null | null | null | null | ['video-object-segmentation', 'temporal-sequences'] | ['computer-vision', 'reasoning'] | [ 3.12584460e-01 3.37457448e-01 -1.62753090e-01 -3.40341508e-01
-4.73507673e-01 -7.67814934e-01 3.68568003e-01 -2.00320646e-01
-3.04168254e-01 7.45931149e-01 -2.34497339e-01 4.47717980e-02
2.82003403e-01 -5.10227025e-01 -1.46752405e+00 -1.01274943e+00
3.51739042e-02 5.29176891e-01 6.70565009e-01 -2.22749323e-01
-6.79134065e-03 3.15202743e-01 -1.78992188e+00 3.57753485e-01
9.69296038e-01 1.29252839e+00 5.64269662e-01 4.84345466e-01
-9.50184464e-03 6.36319339e-01 -2.30999514e-01 -2.27702796e-01
4.15009707e-01 -2.94797450e-01 -9.62713897e-01 5.00111282e-01
8.65048051e-01 -5.59732914e-01 -3.74264449e-01 1.00571585e+00
-5.55863604e-02 1.65219203e-01 8.36285055e-01 -1.28623199e+00
-5.40993273e-01 3.88435036e-01 -5.49416006e-01 1.97979525e-01
2.57227063e-01 4.75351006e-01 8.32478106e-01 -7.29280889e-01
6.57144904e-01 1.15750921e+00 2.29754582e-01 7.25570560e-01
-1.18037951e+00 -3.70250314e-01 5.96150458e-01 3.29760432e-01
-8.92466486e-01 -3.02865386e-01 7.96111166e-01 -7.15306878e-01
4.19786870e-01 -5.27499951e-02 6.82687283e-01 1.33709824e+00
-6.46950826e-02 1.21690547e+00 1.18628395e+00 -8.53446499e-02
4.70930785e-01 -1.91329084e-02 3.19059789e-01 5.61893344e-01
1.83804870e-01 1.88543230e-01 -4.94606614e-01 2.96472728e-01
7.40498841e-01 4.05824520e-02 -4.29844111e-01 -4.86430645e-01
-1.23181844e+00 3.67053807e-01 4.30248827e-01 -4.90632765e-02
-3.91870737e-01 2.08754450e-01 1.51108861e-01 1.71084046e-01
2.50442058e-01 6.69552088e-02 -5.93483925e-01 4.95552197e-02
-7.96064734e-01 3.49593326e-03 5.86716235e-01 1.02950239e+00
1.21134341e+00 2.01578364e-01 -1.85155913e-01 4.29460049e-01
3.02464634e-01 6.63867891e-01 3.00788462e-01 -1.12150800e+00
2.82761961e-01 5.03686726e-01 3.25275838e-01 -4.73226875e-01
-8.34073946e-02 -1.51732266e-01 -5.87746739e-01 4.61291909e-01
8.96060288e-01 -5.31739369e-02 -1.50860572e+00 1.91391206e+00
3.81268591e-01 5.99130392e-01 1.26529839e-02 1.15595937e+00
7.62846470e-01 6.87508106e-01 -1.11702345e-01 -4.20050584e-02
1.28429413e+00 -1.01780355e+00 -4.86488163e-01 -6.93941712e-01
8.02668184e-02 -3.99906546e-01 1.07310951e+00 4.79432523e-01
-1.07832170e+00 -8.28389645e-01 -8.89883935e-01 2.68536471e-02
-2.70434350e-01 -1.01183638e-01 7.14521348e-01 4.21187848e-01
-9.99817312e-01 2.98115939e-01 -1.14817429e+00 -3.89000326e-01
6.60115600e-01 3.16274226e-01 -4.53766525e-01 -2.77215272e-01
-9.16435301e-01 5.65349638e-01 5.71036041e-01 8.33054259e-02
-1.46774638e+00 -5.85838377e-01 -1.02762628e+00 -5.24221659e-02
5.67406416e-01 -7.96870053e-01 9.61713493e-01 -1.54081166e+00
-1.40445065e+00 7.98873365e-01 -2.11317316e-01 -5.84306896e-01
5.00445426e-01 -2.33328998e-01 -1.61389227e-03 5.90469480e-01
4.09167036e-02 1.18974972e+00 1.39143407e+00 -1.70291710e+00
-6.43603384e-01 -4.45126265e-01 2.58531570e-01 1.40826687e-01
-2.24443585e-01 -5.77150524e-01 -6.34453893e-01 -7.08263397e-01
-3.76714990e-02 -1.06700826e+00 1.60756454e-01 2.38985240e-01
-4.09448147e-01 -1.51392326e-01 1.01452720e+00 -6.55644774e-01
6.68038905e-01 -2.10757375e+00 4.49153543e-01 -7.85609186e-02
-4.00254913e-02 1.61124215e-01 -2.88757712e-01 9.05128196e-02
2.42052041e-02 -2.99123526e-01 -4.90334332e-01 -3.26638222e-01
-1.72124833e-01 5.82864583e-01 -5.91725111e-01 5.17875016e-01
3.66545886e-01 9.77432013e-01 -8.21512997e-01 -4.36143816e-01
2.29361415e-01 3.10953468e-01 -5.75087726e-01 4.19580579e-01
-6.40786231e-01 9.31759417e-01 -2.72561610e-01 8.94259989e-01
8.76578569e-01 -3.38055819e-01 -3.76075745e-01 -2.52654642e-01
5.68021350e-02 -2.71147341e-01 -1.04353940e+00 1.95659077e+00
1.59607716e-02 5.46748757e-01 2.12030292e-01 -1.44169474e+00
5.07047176e-01 1.33467317e-01 5.51527917e-01 -3.36579055e-01
5.35426997e-02 5.37697636e-02 -5.61616421e-02 -8.62154484e-01
9.33274999e-02 7.69715337e-03 1.90654844e-01 1.96616516e-01
4.42347169e-01 -1.19312242e-01 3.67313415e-01 1.07639104e-01
7.99092770e-01 5.33513546e-01 -1.51622519e-01 -1.36518762e-01
4.90853071e-01 1.05778137e-02 6.50607526e-01 7.28317082e-01
-1.93445459e-01 6.81131601e-01 3.23207557e-01 -1.88654214e-01
-8.33118081e-01 -1.31151068e+00 -3.10645878e-01 1.07704937e+00
6.86911702e-01 3.63891274e-01 -7.90074348e-01 -7.51997709e-01
1.14118326e-02 5.50602317e-01 -8.86244357e-01 -1.63842253e-02
-3.92372638e-01 -3.00219983e-01 2.41251037e-01 7.17718303e-01
4.26496834e-01 -1.07753742e+00 -7.21898198e-01 -2.15929404e-01
-3.23208123e-01 -1.44119740e+00 -4.64230061e-01 1.64847285e-01
-8.55118155e-01 -1.13902128e+00 -9.39947546e-01 -9.35688794e-01
8.33360195e-01 3.43235821e-01 8.06187451e-01 -2.24045426e-01
-1.77853152e-01 8.35611224e-01 -3.67675036e-01 -2.24523172e-01
-1.96260765e-01 -3.53960037e-01 9.56841465e-03 5.84182501e-01
3.19908336e-02 -5.42474091e-01 -8.39492261e-01 4.36409920e-01
-1.08812213e+00 -5.06107649e-03 6.68383181e-01 9.02508616e-01
7.13620186e-01 5.00128604e-03 5.70878804e-01 -6.81584239e-01
-3.16189557e-01 -5.90111852e-01 -4.73930955e-01 3.17469925e-01
-1.51769131e-01 4.50087674e-02 6.06237173e-01 -7.61666656e-01
-1.26385260e+00 3.88477772e-01 1.09255947e-01 -7.55925477e-01
-5.51667333e-01 1.79811671e-01 -3.06106538e-01 -3.08505539e-02
4.90566403e-01 3.11712861e-01 2.24161059e-01 -3.03911269e-01
4.40135539e-01 3.45290154e-01 8.25340569e-01 -8.06508243e-01
8.01181257e-01 1.08415937e+00 -1.61139727e-01 -9.13238108e-01
-8.82783890e-01 -6.01207376e-01 -7.58227587e-01 -3.35250616e-01
1.22836924e+00 -1.00784099e+00 -7.32400596e-01 6.37949109e-01
-1.11650705e+00 -6.81376934e-01 -2.87293941e-01 4.18123633e-01
-7.92130411e-01 6.56206191e-01 -4.79214579e-01 -7.54899263e-01
1.20274961e-01 -1.27419555e+00 1.23406529e+00 3.56247008e-01
2.31480107e-01 -9.75419164e-01 -3.31430674e-01 5.68363249e-01
-1.80722144e-03 2.05848336e-01 7.22231448e-01 -5.12777448e-01
-1.08814991e+00 2.87377201e-02 -1.28683150e-01 6.00253344e-01
-2.28698365e-02 -1.34916723e-01 -1.22169852e+00 -3.94746125e-01
2.79213786e-02 -7.07776725e-01 1.26013124e+00 6.19583786e-01
1.30062616e+00 -1.07819609e-01 -3.12117755e-01 7.10916042e-01
1.10965538e+00 1.06442146e-01 5.89091778e-01 3.39466743e-02
7.54092515e-01 7.70801127e-01 7.48935997e-01 2.22379774e-01
3.05157989e-01 5.20389736e-01 8.84941816e-01 -1.07611358e-01
-1.98754862e-01 -1.16088346e-01 5.79160154e-01 3.69044960e-01
-5.91736175e-02 -3.30882251e-01 -7.00159311e-01 5.80084264e-01
-1.96023667e+00 -8.51798892e-01 9.26518217e-02 2.14945364e+00
7.30812013e-01 6.21274784e-02 2.92653143e-01 9.85878520e-03
6.86786592e-01 -5.64777851e-02 -1.04481936e+00 1.69568852e-01
-2.96058476e-01 -3.83105688e-02 3.78174037e-01 3.72895300e-01
-1.13304484e+00 8.98775578e-01 5.61108732e+00 5.85421860e-01
-1.08357787e+00 -3.60571332e-02 6.72101140e-01 3.90473455e-01
-2.54799098e-01 1.66112691e-01 -7.36156762e-01 6.52735114e-01
3.47501487e-01 1.81704000e-01 4.86826718e-01 7.87622154e-01
1.69612225e-02 -3.76492798e-01 -1.60253739e+00 9.43089306e-01
3.06315243e-01 -9.73735094e-01 2.03240395e-01 -7.19061270e-02
7.10782290e-01 -9.92061719e-02 3.05821776e-01 1.99501470e-01
-5.56872366e-03 -8.29847336e-01 8.23772192e-01 6.97847784e-01
6.92806542e-01 -1.71168923e-01 5.34354806e-01 5.15182137e-01
-9.85103011e-01 -2.42600963e-01 -3.44856977e-01 9.64121670e-02
2.53169835e-01 2.53197521e-01 -5.74734032e-01 4.46556211e-01
7.32805014e-01 9.46003675e-01 -6.42838299e-01 1.14145935e+00
-2.66903788e-01 7.48363733e-01 -3.35838586e-01 5.01338661e-01
1.83239818e-01 -2.47579172e-01 7.35863209e-01 9.61049378e-01
2.72575676e-01 -1.72586064e-03 3.09126765e-01 9.84295726e-01
7.18474910e-02 -3.47463876e-01 -4.34928060e-01 6.35293918e-03
2.04699293e-01 1.03534198e+00 -7.77175903e-01 -3.43274087e-01
-5.82503200e-01 1.18398392e+00 1.59251317e-01 8.77442837e-01
-9.13739324e-01 7.77490959e-02 2.66836852e-01 1.35670200e-01
7.42864609e-01 -4.98478338e-02 -1.40392959e-01 -1.32521534e+00
-3.56932171e-02 -6.50229275e-01 5.22439182e-01 -1.03204322e+00
-1.53454471e+00 4.86501843e-01 1.70668647e-01 -1.40246296e+00
-1.91832036e-01 -9.34153378e-01 -4.97032225e-01 5.35805464e-01
-1.68770730e+00 -1.38996756e+00 -6.23685539e-01 7.87860453e-01
6.73211277e-01 -1.00433782e-01 4.45558190e-01 1.41767904e-01
-4.10534501e-01 3.99673969e-01 -2.39487574e-01 1.37550667e-01
8.08377624e-01 -1.28936923e+00 -1.71165988e-01 9.59703624e-01
1.82280600e-01 3.12709063e-01 6.97764099e-01 -5.39568305e-01
-1.46437287e+00 -9.47070897e-01 -1.16477542e-01 -5.90993285e-01
6.80557907e-01 -3.18343252e-01 -1.13399887e+00 7.49600112e-01
3.05739850e-01 4.01463658e-01 2.18349710e-01 -3.28177750e-01
-3.76018494e-01 -2.86979496e-01 -8.43698323e-01 3.86953443e-01
9.51387823e-01 -3.95439088e-01 -7.24308610e-01 2.82975644e-01
7.69405782e-01 -5.40659726e-01 -5.67955732e-01 5.32914937e-01
4.01649356e-01 -9.92967129e-01 1.06471145e+00 -5.61230540e-01
5.05489230e-01 -4.83651698e-01 -1.51079744e-01 -1.17504740e+00
3.16589177e-01 -5.79817593e-01 -4.52977091e-01 1.09520543e+00
9.30793062e-02 -5.11562049e-01 7.86114812e-01 6.30289435e-01
-3.52772236e-01 -6.22353733e-01 -9.27295685e-01 -8.39596212e-01
7.12712556e-02 -4.98629063e-01 2.07287416e-01 8.21568131e-01
-3.21480185e-01 1.11510217e-01 -3.55228931e-01 4.33745891e-01
8.35402012e-01 2.71524400e-01 8.10906172e-01 -1.12234282e+00
-4.86473680e-01 -1.32179320e-01 -6.37921751e-01 -1.67219353e+00
5.07792652e-01 -7.63460040e-01 2.10339293e-01 -1.38125050e+00
2.24628672e-01 -2.81036258e-01 -2.75896937e-01 5.34024477e-01
-1.87705562e-01 1.94576576e-01 1.18716180e-01 2.02792168e-01
-6.74418449e-01 6.06560707e-01 1.57276952e+00 -3.53779018e-01
-1.66311145e-01 1.97041988e-01 -3.58653396e-01 7.64955044e-01
3.66563022e-01 -1.81093752e-01 -6.09058619e-01 -4.56788570e-01
-6.06002957e-02 1.94946781e-01 7.70309091e-01 -7.69037545e-01
2.33483464e-01 -2.44313300e-01 2.58508593e-01 -7.86101878e-01
5.58640122e-01 -1.01905441e+00 -1.66236520e-01 2.23534733e-01
-1.78451374e-01 -4.01341856e-01 2.85611540e-01 1.09958756e+00
-3.49978745e-01 -1.79912657e-01 7.82004952e-01 -1.13194108e-01
-1.12903488e+00 5.21747053e-01 -2.21857697e-01 3.09181213e-01
1.15606582e+00 -5.17453730e-01 -3.80613416e-01 -3.15841556e-01
-8.67074847e-01 4.15512145e-01 7.40476787e-01 4.24340516e-01
6.83206797e-01 -9.27045882e-01 -4.12830353e-01 3.39073420e-01
2.80617356e-01 4.16692913e-01 4.83062744e-01 9.82176661e-01
-2.38724768e-01 1.58570960e-01 -3.37334663e-01 -1.24393213e+00
-8.43468070e-01 8.53325248e-01 2.21592978e-01 6.10160887e-01
-7.72550762e-01 8.20427895e-01 9.07941103e-01 9.55434740e-02
5.28463006e-01 -5.48586607e-01 -1.10736974e-01 -1.29177729e-02
3.49590391e-01 8.91230330e-02 -3.14495444e-01 -7.09770679e-01
-1.52076975e-01 7.20324576e-01 1.43403583e-03 -2.01173685e-02
1.13572943e+00 -3.50071311e-01 -3.16753201e-02 7.67748773e-01
9.30266440e-01 -2.96469182e-01 -2.03264976e+00 -2.43150145e-01
-2.67401785e-01 -4.58737880e-01 -2.30689734e-01 -6.31113708e-01
-1.11887777e+00 1.11070716e+00 5.08369148e-01 5.80861345e-02
1.12671292e+00 2.74364769e-01 8.25841367e-01 4.06061262e-01
2.89287001e-01 -1.06855690e+00 6.26079023e-01 3.10209155e-01
6.88286543e-01 -1.51521802e+00 -3.09706211e-01 -4.64197278e-01
-9.38333511e-01 1.04485869e+00 8.00347686e-01 -1.21626876e-01
6.54071391e-01 1.26579791e-01 -3.60430591e-02 -3.63880135e-02
-5.78843236e-01 -4.04235244e-01 5.69541216e-01 7.25424767e-01
-8.11709687e-02 -2.74815798e-01 2.63247937e-01 5.18382311e-01
2.99409837e-01 -1.31863981e-01 4.74483460e-01 6.73343956e-01
-5.06677747e-01 -6.28180087e-01 -2.67514706e-01 2.86172450e-01
-1.97246522e-01 6.55293092e-02 -2.11373225e-01 7.08268523e-01
3.36482495e-01 7.23751485e-01 1.15312643e-01 -5.88162914e-02
-5.98829612e-02 -5.56199951e-03 6.77815855e-01 -6.84333861e-01
1.44885778e-01 2.62335300e-01 -2.82845378e-01 -5.10294259e-01
-8.38029921e-01 -7.06549346e-01 -1.48713088e+00 3.46804053e-01
-1.64752483e-01 -1.25333533e-01 3.71807843e-01 1.10453081e+00
1.47897020e-01 3.65423560e-01 4.39700067e-01 -1.15334976e+00
-3.14729869e-01 -7.41650999e-01 -4.86078709e-01 8.14664900e-01
7.12164402e-01 -1.05335832e+00 -5.61232984e-01 6.31108761e-01] | [9.299084663391113, 0.05589378625154495] |
fd6f4cf6-764b-4327-8db6-45a64dde197e | a-structural-causal-model-for-mr-images-of | 2103.03158 | null | https://arxiv.org/abs/2103.03158v3 | https://arxiv.org/pdf/2103.03158v3.pdf | A Structural Causal Model for MR Images of Multiple Sclerosis | Precision medicine involves answering counterfactual questions such as "Would this patient respond better to treatment A or treatment B?" These types of questions are causal in nature and require the tools of causal inference to be answered, e.g., with a structural causal model (SCM). In this work, we develop an SCM that models the interaction between demographic information, disease covariates, and magnetic resonance (MR) images of the brain for people with multiple sclerosis. Inference in the SCM generates counterfactual images that show what an MR image of the brain would look like if demographic or disease covariates are changed. These images can be used for modeling disease progression or used for image processing tasks where controlling for confounders is necessary. | ['Jerry L. Prince', 'Aaron Carass', 'Jacob C. Reinhold'] | 2021-03-04 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [ 5.39235175e-01 4.71660703e-01 -4.88639534e-01 -6.22117162e-01
-1.85634762e-01 -8.78092200e-02 7.90600538e-01 1.07171677e-01
-5.52172065e-01 1.16252196e+00 8.63843858e-01 -8.53057265e-01
-3.56915712e-01 -8.49467754e-01 -9.18746948e-01 -5.69705188e-01
-4.67676103e-01 6.07229829e-01 -2.99719155e-01 1.93272904e-01
3.16588432e-01 2.97715008e-01 -9.34297740e-01 3.01163375e-01
8.22590947e-01 2.31557921e-01 2.65589923e-01 4.00133342e-01
-2.78757848e-02 6.30482256e-01 -1.70377925e-01 -3.24402153e-01
-2.47345403e-01 -5.41526973e-01 -1.04984081e+00 -2.67723978e-01
1.77091300e-01 -4.46780533e-01 -1.94799706e-01 9.40222859e-01
4.23144221e-01 8.33806954e-03 9.69602406e-01 -1.17266834e+00
-6.52604878e-01 9.39410567e-01 -7.20444143e-01 4.17840868e-01
3.75151187e-01 2.64586449e-01 4.81878906e-01 -3.43710423e-01
7.64868855e-01 1.88361216e+00 2.76670516e-01 6.79660499e-01
-1.90095520e+00 -6.92698777e-01 1.59486204e-01 3.28219414e-01
-5.25357842e-01 -2.43516460e-01 3.98533314e-01 -7.11247325e-01
4.37224418e-01 5.02603054e-01 6.08833313e-01 1.31333911e+00
1.02465987e+00 2.38301903e-01 1.53209186e+00 -3.77251089e-01
3.43125552e-01 -5.07217228e-01 7.66316354e-02 3.81579697e-01
5.08473158e-01 6.86204553e-01 -3.85626256e-01 -5.16239703e-01
1.09568596e+00 -4.67062136e-03 -3.80424082e-01 -2.84233004e-01
-1.67714572e+00 1.11884558e+00 5.24682403e-01 3.38298321e-01
-9.44095075e-01 4.75272477e-01 1.29588187e-01 1.21437900e-01
2.43226185e-01 5.11470616e-01 -3.65539849e-01 5.60932636e-01
-9.13649201e-01 4.95440185e-01 1.80153280e-01 -4.42901291e-02
1.14521258e-01 -2.56553799e-01 -4.45168972e-01 3.52264434e-01
4.10778880e-01 7.81925023e-01 4.40955251e-01 -1.17066169e+00
3.56588978e-03 2.36206234e-01 2.15505183e-01 -6.68392897e-01
-6.98439837e-01 -1.64089963e-01 -8.46777558e-01 2.86954820e-01
4.66703326e-01 -3.72074634e-01 -1.02567315e+00 1.98748159e+00
3.58494043e-01 2.92149067e-01 -2.54828781e-01 9.85320628e-01
4.86879051e-01 1.82717383e-01 6.92278862e-01 -6.47750974e-01
1.57819557e+00 1.27001509e-01 -1.00928938e+00 -2.95372516e-01
5.80889344e-01 -4.39694732e-01 9.46907580e-01 -6.08204752e-02
-1.10688281e+00 4.17636801e-03 -4.45225477e-01 2.45318055e-01
7.22481236e-02 -4.46407318e-01 8.03664982e-01 5.46673954e-01
-8.59277904e-01 6.88484013e-01 -7.88932621e-01 -3.11061352e-01
6.24736190e-01 2.58501977e-01 -2.82230586e-01 -1.65352359e-01
-1.41746318e+00 1.12688792e+00 1.16174825e-01 -3.11908364e-01
-9.68602180e-01 -1.36693966e+00 -6.50174141e-01 2.24767216e-02
8.63406733e-02 -1.44694459e+00 1.20212483e+00 -7.98290670e-01
-6.93376899e-01 9.29754317e-01 -4.05876458e-01 -5.97465396e-01
5.68652987e-01 1.97900146e-01 -3.36161017e-01 1.64391875e-01
4.57549959e-01 7.60267913e-01 7.65276551e-01 -1.08643937e+00
-4.03201193e-01 -1.13277519e+00 -5.73229864e-02 1.33386537e-01
5.74195683e-01 3.41669858e-01 5.25144637e-01 -8.30404341e-01
9.46552455e-02 -7.73748994e-01 -7.97647238e-01 -9.27328244e-02
-6.06796086e-01 2.59392709e-01 5.30791700e-01 -8.73146355e-01
7.91063607e-01 -1.67209756e+00 -2.72618502e-01 1.25433400e-01
2.37595588e-01 -4.53969300e-01 -1.49630398e-01 -1.02526158e-01
-8.01978111e-01 5.48802972e-01 -2.37204581e-01 4.94486004e-01
-2.82643020e-01 -1.38272181e-01 -1.58773527e-01 4.99703288e-01
5.05085364e-02 1.05917299e+00 -8.33633959e-01 -6.40789211e-01
2.25739151e-01 3.74143809e-01 -6.04824126e-01 -1.32417649e-01
-6.62871301e-02 9.02189314e-01 -4.94030774e-01 -4.58633807e-03
5.91936588e-01 -1.25881910e-01 5.15446723e-01 -1.16501421e-01
-1.02793753e-01 3.28235030e-01 -6.54861271e-01 1.18448532e+00
-2.43375137e-01 1.53303534e-01 2.09088903e-03 -1.11871111e+00
4.01770353e-01 6.09679997e-01 4.86262560e-01 -9.40906346e-01
1.85532719e-01 -1.34254694e-01 5.81994474e-01 -6.60050154e-01
-4.39207792e-01 -1.05280042e+00 2.03185320e-01 7.70733535e-01
-4.42354292e-01 2.09808871e-02 -3.84493470e-02 2.21637294e-01
1.24625194e+00 -3.42121959e-01 4.73524809e-01 -4.19008315e-01
1.24165252e-01 2.49119967e-01 6.85700417e-01 8.45274448e-01
-9.18477103e-02 2.40115270e-01 6.68708682e-01 -4.35605168e-01
-1.05994499e+00 -1.32794845e+00 -4.81932670e-01 2.81425327e-01
-2.94483453e-01 6.31659508e-01 -6.51172638e-01 -3.89402032e-01
1.87884599e-01 1.43604422e+00 -9.13355649e-01 -3.62533718e-01
-5.76944888e-01 -1.47501767e+00 -3.66862901e-02 2.82169759e-01
2.43941173e-01 -1.06275141e+00 -9.07425523e-01 2.96870708e-01
-5.03119171e-01 -3.07904541e-01 -2.04063103e-01 -3.38207424e-01
-1.52705717e+00 -1.36166644e+00 -9.21501577e-01 -1.38492689e-01
5.20381868e-01 4.33278754e-02 1.19914663e+00 -1.35315165e-01
-4.58506972e-01 1.78687528e-01 6.72265366e-02 -5.50055146e-01
-3.86482805e-01 -7.41600215e-01 1.98724538e-01 -3.60218473e-02
-5.63166775e-02 -9.05049324e-01 -1.18844211e+00 6.90596253e-02
-8.70862603e-01 3.03194851e-01 7.13833451e-01 7.19807208e-01
5.17412901e-01 -1.46305189e-01 7.87105858e-01 -1.19993830e+00
7.66366303e-01 -7.33965814e-01 -1.59183472e-01 2.56948739e-01
-6.45821214e-01 8.60076174e-02 1.29577637e-01 -4.92316902e-01
-1.58409631e+00 -3.11675519e-01 2.20497280e-01 1.01554751e-01
-2.82064468e-01 6.89387023e-01 -2.38733575e-01 5.59739113e-01
8.01171303e-01 -3.20105731e-01 2.18069926e-02 -4.46600735e-01
3.60501915e-01 2.98438877e-01 5.40008426e-01 -4.66391951e-01
1.86188772e-01 8.78060102e-01 2.71001786e-01 -3.36707264e-01
-5.10085225e-01 2.20566466e-01 -3.20907593e-01 -1.55686662e-01
1.01099896e+00 -4.95112866e-01 -1.00278616e+00 -5.42111583e-02
-9.67394650e-01 -4.22595471e-01 -1.78142235e-01 1.06384850e+00
-7.66189992e-01 -1.74081117e-01 -6.08745441e-02 -7.02073812e-01
-1.76194832e-01 -1.04047084e+00 5.70307016e-01 1.03690088e-01
-6.38107896e-01 -1.15881205e+00 7.02456534e-02 4.68834847e-01
4.70459551e-01 5.54427087e-01 1.73601675e+00 -1.81000605e-01
-3.61730516e-01 -2.44897939e-02 -2.28285864e-01 -5.85852623e-01
2.07194656e-01 -2.89687932e-01 -3.51678908e-01 7.83713087e-02
3.27796429e-01 1.28554538e-01 5.67400157e-01 1.56810784e+00
1.15019226e+00 -5.77383935e-01 -7.03031301e-01 6.62514418e-02
1.27603972e+00 5.23747802e-01 7.91768432e-01 9.15031955e-02
2.23285034e-01 1.08878672e+00 5.94507277e-01 2.58953243e-01
4.23680007e-01 6.66732669e-01 3.50307077e-01 -7.81843066e-02
4.75051906e-03 -1.01201937e-01 -2.90876590e-02 -3.62366438e-01
3.72420661e-02 1.30239338e-01 -1.12572765e+00 7.52435625e-01
-1.63090873e+00 -1.13786483e+00 -8.84480774e-01 2.12087369e+00
8.15063179e-01 -2.22229630e-01 5.72931953e-04 -2.62263685e-01
9.18702900e-01 -7.87979364e-02 -7.24753797e-01 -4.06932443e-01
-1.19410027e-02 1.49674654e-01 4.36399192e-01 5.60968220e-01
-6.59632683e-01 1.98769346e-01 7.42380524e+00 2.59163678e-01
-8.15668941e-01 2.26876587e-01 1.17269862e+00 -1.75951943e-01
-7.61644304e-01 4.94638294e-01 -4.52018389e-03 5.19205987e-01
1.23833382e+00 -4.92385238e-01 1.96603075e-01 4.58543748e-02
1.32260668e+00 -7.76194394e-01 -1.15244710e+00 5.49025178e-01
-5.42468488e-01 -1.51912963e+00 2.45521337e-01 1.83866724e-01
5.62123477e-01 -2.97373295e-01 1.49929039e-02 -3.04010034e-01
9.78422403e-01 -1.23729539e+00 4.24869061e-01 1.08802378e+00
9.10754144e-01 -3.98691446e-01 6.83337927e-01 3.77093792e-01
-4.98824753e-02 8.58008340e-02 -6.95807561e-02 3.51084559e-03
6.42639756e-01 1.13549542e+00 -8.88703108e-01 2.30422348e-01
6.86150014e-01 -5.86160794e-02 -9.68602300e-02 9.13997948e-01
-2.99126446e-01 7.73675263e-01 9.69732851e-02 5.18137038e-01
-1.86958954e-01 -1.24462150e-01 6.11374676e-01 6.62454665e-01
2.54283041e-01 4.12355363e-01 -4.97485816e-01 1.27252853e+00
1.67988956e-01 1.00152353e-02 -6.46352530e-01 8.84000678e-03
2.75885403e-01 7.33016491e-01 -6.34009719e-01 -2.99048096e-01
-1.77389592e-01 5.15172362e-01 -2.46732756e-01 3.77959013e-01
-3.71071011e-01 4.65773284e-01 4.39925224e-01 6.67467296e-01
-4.73198503e-01 3.73979896e-01 -7.76079893e-01 -9.78316426e-01
-3.53190720e-01 -8.55777144e-01 6.57955348e-01 -1.04761565e+00
-1.22716558e+00 -1.61601603e-01 4.35794890e-01 -2.98962891e-01
-2.94181705e-01 -1.11182161e-01 -9.35532033e-01 1.15531242e+00
-1.01258743e+00 -8.48086894e-01 3.14337254e-01 4.99114901e-01
-5.60976937e-02 2.93398470e-01 7.80760169e-01 -9.93016735e-02
-3.98894966e-01 -1.39676332e-01 -2.62841791e-01 -1.49997249e-01
5.78857005e-01 -1.24473000e+00 1.03364348e-01 5.72947025e-01
-4.06958967e-01 7.08600819e-01 1.07245016e+00 -1.33414209e+00
-7.72574425e-01 -8.35265279e-01 1.13914514e+00 -2.46600434e-01
5.73122621e-01 2.60165542e-01 -6.99638546e-01 6.65197015e-01
6.69370219e-02 -4.30359304e-01 6.44336998e-01 2.15103105e-01
2.79781781e-02 2.38581896e-01 -1.53101718e+00 9.46818769e-01
8.75935316e-01 4.18949872e-02 -9.19302225e-01 4.03196633e-01
6.44154966e-01 2.80939221e-01 -8.68461609e-01 4.57798928e-01
5.35841465e-01 -6.66408777e-01 1.19275939e+00 -1.30112398e+00
7.66727209e-01 2.97007021e-02 5.58302402e-02 -1.61871374e+00
-2.85656124e-01 -1.35936782e-01 3.67110729e-01 6.51486695e-01
3.29704314e-01 -6.08335555e-01 6.12785339e-01 1.12382376e+00
4.49236691e-01 -4.00015771e-01 -1.08857906e+00 -4.40431446e-01
3.83029848e-01 -1.96344808e-01 6.71038747e-01 1.25631464e+00
8.35693628e-02 3.99797350e-01 -1.68639898e-01 9.52062830e-02
8.96867514e-01 2.19355166e-01 2.14272127e-01 -1.39168286e+00
-1.34770563e-02 -1.45726457e-01 4.04013175e-04 1.16112322e-01
1.41328290e-01 -7.83079743e-01 -6.20700717e-01 -1.80142581e+00
8.44130397e-01 -3.99829000e-01 1.72803834e-01 2.91013509e-01
-4.64486301e-01 -3.55509162e-01 3.81705235e-03 9.19249356e-02
4.09555852e-01 4.78703409e-01 1.43444085e+00 -1.10135287e-01
-2.71515585e-02 3.51814553e-02 -9.33605492e-01 8.26393068e-01
9.35902357e-01 -8.69325697e-01 -2.72255152e-01 -5.14601655e-02
1.21100552e-01 9.62121606e-01 9.78599131e-01 -1.72847018e-01
-3.16326648e-01 -7.33624458e-01 5.75002015e-01 -3.09269309e-01
-6.02930672e-02 -6.45761013e-01 6.53223634e-01 1.10632873e+00
-4.60302711e-01 1.30809203e-01 -9.13031027e-02 4.69776869e-01
-1.02110412e-02 -2.90186703e-01 8.11685622e-01 -3.59186023e-01
-1.47136793e-01 8.79132450e-02 -5.86612523e-01 -4.95571382e-02
8.29683661e-01 1.43910334e-01 -4.02999729e-01 -5.87469220e-01
-1.27685916e+00 2.94557780e-01 1.96623325e-01 4.29109856e-02
5.90099514e-01 -1.29349804e+00 -1.00156045e+00 -6.05896592e-01
-3.29551280e-01 -4.83716607e-01 6.61671162e-01 1.29634500e+00
-2.74740875e-01 6.06029332e-01 -4.06730860e-01 -1.16802335e-01
-1.10965538e+00 7.64192104e-01 4.99074668e-01 -5.82188666e-01
-5.43666661e-01 3.60722005e-01 8.52226198e-01 -2.84780890e-01
-4.52430934e-01 9.62174758e-02 -2.48392209e-01 1.15135320e-01
7.29070246e-01 3.90433758e-01 -2.24736243e-01 -2.34928027e-01
-3.48351449e-01 -2.80015022e-01 -1.23462556e-02 -4.98036355e-01
1.60822177e+00 -4.04952228e-01 -4.01922017e-01 3.11154127e-01
6.29978836e-01 -4.69915539e-01 -9.22416508e-01 -1.45132631e-01
-4.53633815e-02 -5.46731353e-01 3.29207003e-01 -1.25561011e+00
-9.56262410e-01 7.59368777e-01 9.60299492e-01 -1.73834100e-01
1.01951456e+00 1.85862392e-01 1.00288078e-01 -3.04672658e-01
1.65107027e-01 -6.13229752e-01 -3.63731295e-01 -2.74870753e-01
1.26129854e+00 -9.67532098e-01 -3.05936728e-02 -2.99269080e-01
-5.34131408e-01 5.78106105e-01 2.76101381e-02 1.06615648e-02
7.55349159e-01 1.00033805e-01 -1.13775522e-01 -6.75894976e-01
-8.96553218e-01 3.83073203e-02 2.36102208e-01 7.28809059e-01
6.68235600e-01 6.63404703e-01 -1.10657728e+00 5.73945642e-01
-2.37902150e-01 2.67266721e-01 5.31090319e-01 5.00432372e-01
-2.05938079e-04 -1.08541322e+00 -9.14299846e-01 1.11993122e+00
-6.22484505e-01 -2.50405729e-01 -3.45613182e-01 6.55711293e-01
6.65068207e-03 8.44996870e-01 2.43549496e-01 3.83230031e-01
3.08078676e-01 2.69516319e-01 5.54780364e-01 -5.30943871e-01
-1.37844399e-01 4.21603490e-03 1.71145305e-01 -5.73221147e-01
-8.12697589e-01 -1.31792724e+00 -1.30047560e+00 -4.55290407e-01
-6.91999355e-03 -9.47504714e-02 5.72344661e-01 1.14750123e+00
7.50137419e-02 6.79604590e-01 2.61809051e-01 -3.19363594e-01
-1.71170905e-01 -9.98255730e-01 -5.34247339e-01 3.35218579e-01
1.88410982e-01 -7.46832252e-01 -3.67618203e-02 2.52792090e-01] | [8.030915260314941, 5.460659980773926] |
09ebd734-4acf-4289-9c4e-24a958ca9f02 | exploring-the-value-of-multi-view-learning | null | null | https://openreview.net/forum?id=05JLuxoSbX5 | https://openreview.net/pdf?id=05JLuxoSbX5 | Exploring the Value of Multi-View Learning for Session-Aware Query Representation | Recent years have witnessed a growing interest towards learning distributed query representations that are able to capture search intent semantics. Most existing approaches learn query embeddings using relevance supervision making them suited only to document ranking tasks. Besides, they generally consider either user’s query reformulations or system’s rankings whereas previous findings show that user’s query behavior and knowledge change depending on the system’s results, intertwine and affect each other during the completion of a search task. In this paper, we explore the value of multi-view learning for generic and unsupervised session-aware query representation learning. First, single-view query embeddings are obtained in separate spaces from query reformulations and document ranking representations using transformers. Then, we investigate the use of linear (CCA) and non linear (UMAP) multi-view learning methods, to align those spaces with the aim of revealing similarity traits in the multi-view shared space. Experimental evaluation is carried out in a query classification and session-based retrieval downstream tasks using respectively the KDD and TREC session datasets. The results show that multi-view learning is an effective and controllable approach for unsupervised learning of generic query representations and can reflect search behavior patterns. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['multi-view-learning', 'document-ranking'] | ['computer-vision', 'natural-language-processing'] | [-2.93080751e-02 -5.20337880e-01 -7.10079670e-01 -4.76598710e-01
-9.44779038e-01 -8.89817357e-01 1.23057067e+00 5.63487947e-01
-5.14784157e-01 -6.75216243e-02 7.34977186e-01 -6.69702590e-02
-7.69691110e-01 -6.31116331e-01 -2.80870140e-01 -4.11660880e-01
-1.37059778e-01 7.09067106e-01 2.13215783e-01 -5.62697768e-01
4.54295069e-01 3.60483348e-01 -1.70467412e+00 5.72698474e-01
3.48940134e-01 7.07919121e-01 2.65383325e-03 7.61904180e-01
-4.04561982e-02 6.86246991e-01 -5.06313264e-01 -1.06457449e-01
6.54635951e-02 1.00681797e-01 -1.02065337e+00 -2.85653591e-01
3.84978443e-01 -2.28796482e-01 -4.15420353e-01 4.70946640e-01
7.44262457e-01 1.34962916e-01 9.09896553e-01 -1.07315969e+00
-8.61988962e-01 2.09525108e-01 -1.75668210e-01 3.90794098e-01
6.89081311e-01 -5.12959599e-01 1.56048286e+00 -8.78815949e-01
7.36326396e-01 1.11924672e+00 6.65271953e-02 1.94467932e-01
-1.47834861e+00 -9.78000984e-02 1.44538298e-01 4.12976533e-01
-1.19807148e+00 -2.20203817e-01 9.26061153e-01 -5.05831599e-01
9.03768420e-01 5.36917508e-01 3.00345004e-01 1.24177754e+00
1.44155934e-01 9.06416118e-01 1.03219056e+00 -4.09775555e-01
2.17540979e-01 6.15938723e-01 3.62619311e-01 1.09147280e-02
1.87691953e-02 2.19020575e-01 -6.98794186e-01 -5.38452029e-01
1.81803361e-01 4.16788489e-01 -2.72755831e-01 -1.23882651e+00
-1.16968179e+00 9.12768245e-01 5.01491249e-01 6.64578855e-01
-4.96495008e-01 -1.16380258e-02 6.31647050e-01 7.45121717e-01
5.26896536e-01 6.70099735e-01 -6.20120227e-01 -1.58456206e-01
-6.95615530e-01 1.87810153e-01 8.86668086e-01 7.55874217e-01
7.71902680e-01 -4.63342130e-01 -5.72215736e-01 9.66982543e-01
1.85220838e-01 3.73100579e-01 8.51857781e-01 -3.51670593e-01
3.94455284e-01 9.76589978e-01 1.98547453e-01 -1.14707077e+00
-1.76583141e-01 -4.12625492e-01 -2.24935785e-01 -3.93707365e-01
-8.13454986e-02 5.09247601e-01 -8.88060629e-01 1.37524891e+00
-1.45705007e-02 -3.04455161e-01 2.82511503e-01 7.52518892e-01
5.83313107e-01 2.91538179e-01 1.17322437e-01 -7.26748481e-02
1.41603041e+00 -7.74970651e-01 -4.65383291e-01 8.51460323e-02
1.09188724e+00 -6.24961436e-01 1.18312263e+00 1.34257272e-01
-7.22800255e-01 -6.15298510e-01 -7.81367004e-01 -8.10455903e-02
-9.39889073e-01 2.15374425e-01 6.08593524e-01 5.58242381e-01
-1.03008199e+00 5.10649420e-02 -6.83400929e-01 -6.81106508e-01
-1.28820270e-01 4.55892593e-01 -3.31489027e-01 -2.91421086e-01
-1.29061282e+00 8.45366180e-01 2.34776139e-01 -1.75620839e-01
-9.00858998e-01 -7.02947378e-01 -5.33701181e-01 3.13646019e-01
3.59082252e-01 -4.44354326e-01 1.11377668e+00 -3.88720334e-01
-1.08716023e+00 8.95434558e-01 -2.33871713e-01 -1.56138822e-01
1.14394873e-01 -3.07226330e-01 -6.49249971e-01 4.85825315e-02
1.38419867e-01 1.23176411e-01 7.16946006e-01 -1.31090999e+00
-2.91246712e-01 -8.83273065e-01 3.00569087e-01 5.68952382e-01
-6.76236331e-01 -1.43878087e-01 -7.93266475e-01 -1.92946747e-01
8.90701637e-02 -8.02183032e-01 6.14844039e-02 -5.55781066e-01
-1.13507755e-01 -6.85001075e-01 1.01106381e+00 -2.49035999e-01
1.54894304e+00 -2.04380417e+00 4.71762300e-01 4.02888149e-01
-3.20765488e-02 1.68335319e-01 -1.58054978e-01 1.23427033e+00
-6.14890568e-02 1.73005372e-01 4.66010392e-01 -8.14989954e-02
6.21971637e-02 1.86538309e-01 -5.50881863e-01 3.32923889e-01
-1.81981251e-01 1.09155452e+00 -8.39827836e-01 -2.96576738e-01
1.11112610e-01 3.16153228e-01 -6.11364782e-01 3.87990087e-01
-1.42301649e-01 1.37787759e-01 -8.15060139e-01 5.58446884e-01
1.01243787e-01 -4.00548011e-01 2.84085512e-01 -3.76674712e-01
-4.32963334e-02 2.42431805e-01 -7.44140565e-01 1.89513254e+00
-9.83801365e-01 3.46061558e-01 -3.13327938e-01 -1.13170743e+00
8.35001647e-01 4.96669173e-01 6.69540584e-01 -1.05010355e+00
-2.85935313e-01 1.70485929e-01 -3.83510083e-01 -6.87956512e-01
7.94938922e-01 3.18876028e-01 -3.76789421e-01 8.28428924e-01
2.27884054e-01 1.88715570e-02 -3.08748353e-02 3.95259023e-01
9.80105340e-01 -1.33821890e-01 2.18720689e-01 -1.56290364e-02
7.55014062e-01 -1.90386429e-01 -2.23344907e-01 7.66051173e-01
1.11575730e-01 2.94197232e-01 5.08767128e-01 -2.57265031e-01
-6.82842433e-01 -9.89459157e-01 -7.50735477e-02 1.85773528e+00
2.56153405e-01 -5.19656122e-01 5.72927818e-02 -8.01080525e-01
1.69679165e-01 7.04794466e-01 -8.01937819e-01 -6.33067131e-01
-1.95669681e-01 -3.52643639e-01 1.52397290e-01 4.24396634e-01
8.48985314e-02 -8.46311867e-01 -4.88031089e-01 -4.72043790e-02
1.52839674e-02 -7.52482295e-01 -5.25302887e-01 2.12456271e-01
-9.40629005e-01 -1.00498223e+00 -7.60638654e-01 -5.92188418e-01
3.80513281e-01 6.86600566e-01 1.18886662e+00 -1.16874963e-01
-2.70910710e-01 1.59712338e+00 -5.48176289e-01 -1.18836433e-01
1.59807149e-02 4.45246011e-01 3.01301610e-02 7.09355325e-02
8.23320687e-01 -5.29995739e-01 -7.66687751e-01 4.39167351e-01
-1.27949381e+00 -7.20746160e-01 5.77774882e-01 9.64986861e-01
5.38917959e-01 -3.85890454e-01 3.44561458e-01 -9.63083208e-01
1.18022001e+00 -6.91059947e-01 -3.19211751e-01 8.45250189e-01
-1.25587797e+00 3.63899231e-01 6.61158860e-02 -3.66211265e-01
-9.31347370e-01 -3.72241557e-01 6.15993619e-01 -5.27576029e-01
-6.70626834e-02 7.62564659e-01 1.57540977e-01 2.66353607e-01
7.95716882e-01 5.54534912e-01 -2.28425220e-01 -5.25315642e-01
7.38317132e-01 8.25039148e-01 -6.95388438e-03 -3.67753029e-01
7.79636025e-01 5.18107891e-01 -4.55995709e-01 -7.82376289e-01
-5.83834708e-01 -1.41417372e+00 -6.59605622e-01 -1.06653176e-01
8.08085024e-01 -9.90023375e-01 -5.39804220e-01 -4.74145226e-02
-7.11973667e-01 2.83678532e-01 -2.13545546e-01 5.22479475e-01
-5.51574886e-01 1.05630450e-01 -7.93440118e-02 -7.06691742e-01
-1.63716689e-01 -9.59327161e-01 1.45337570e+00 7.03141466e-02
-6.27604872e-02 -1.34233856e+00 7.12989151e-01 4.97362107e-01
5.60147822e-01 -4.63504493e-01 1.20504785e+00 -1.36344898e+00
-5.58860242e-01 -4.97180998e-01 -1.13991633e-01 1.24019049e-01
3.95995945e-01 -5.43384969e-01 -1.16744578e+00 -6.42039418e-01
-5.49601167e-02 -4.38019812e-01 7.43461549e-01 -1.01082318e-01
8.98512542e-01 -2.25643367e-02 -5.17357826e-01 2.21098304e-01
1.47893548e+00 4.25545201e-02 3.53354007e-01 3.99084389e-01
3.65995020e-01 6.42556906e-01 5.01305759e-01 3.23110431e-01
3.33450705e-01 1.03992879e+00 3.24036747e-01 1.90580487e-01
2.92982697e-01 -6.07269943e-01 2.61381149e-01 6.88947618e-01
8.44746679e-02 -3.33171755e-01 -7.94086397e-01 7.12205231e-01
-1.61543715e+00 -1.06835103e+00 2.92863339e-01 2.40507269e+00
3.67348135e-01 -2.30593145e-01 1.03324801e-02 -1.78243309e-01
2.15317354e-01 4.92993087e-01 -5.23307979e-01 -5.10698259e-01
2.68838674e-01 5.75701222e-02 3.19335490e-01 3.48494232e-01
-9.54481244e-01 8.10013831e-01 5.64850378e+00 2.85765380e-01
-1.10305572e+00 -5.60187269e-03 1.59154996e-01 -5.64782619e-02
-8.78330529e-01 5.63708581e-02 -5.50768495e-01 -3.45118679e-02
1.01010346e+00 -3.67983192e-01 2.30863154e-01 8.37277591e-01
-6.04892150e-02 8.90090987e-02 -1.44714212e+00 9.85740781e-01
4.24277157e-01 -9.10710692e-01 3.80043626e-01 2.68294513e-01
4.80283141e-01 1.15542606e-01 4.47307736e-01 7.40393996e-01
-2.81027127e-02 -8.18645000e-01 -1.53881699e-01 6.88868463e-01
3.59738708e-01 -4.64531481e-01 5.93688250e-01 2.93994993e-01
-1.17053568e+00 -2.07837105e-01 -1.09008998e-01 4.17953372e-01
-1.38430580e-01 -9.37337801e-03 -1.03804862e+00 8.34152043e-01
5.20883799e-01 7.46590972e-01 -1.00984693e+00 7.78889596e-01
2.57789254e-01 3.09013575e-01 6.53732195e-02 -1.60272956e-01
1.94971159e-01 -2.62676358e-01 5.00109315e-01 8.44381392e-01
1.14399083e-01 -2.99167216e-01 1.31130308e-01 5.47704875e-01
1.01172104e-01 3.66160572e-01 -9.82638299e-01 -4.38970983e-01
3.34856778e-01 9.80851352e-01 -3.34056884e-01 -1.50302678e-01
-5.44314802e-01 1.14669645e+00 2.06092566e-01 8.08769166e-01
-2.97557056e-01 -6.46843165e-02 6.33777440e-01 -9.71780047e-02
4.53959107e-01 -1.53858230e-01 3.78815919e-01 -1.30078948e+00
1.70939267e-01 -6.90861523e-01 5.75886190e-01 -4.60987628e-01
-1.24326980e+00 4.44158763e-01 3.41222823e-01 -1.39927411e+00
-7.75855362e-01 -3.12099874e-01 -3.42635334e-01 7.63768017e-01
-1.63977098e+00 -1.12466872e+00 -4.46579158e-02 7.82029748e-01
5.29650748e-01 -4.84078199e-01 1.19969285e+00 3.19257259e-01
-2.54182182e-02 5.40192962e-01 5.38841844e-01 -2.02839941e-01
8.84926975e-01 -1.25037563e+00 -2.03130454e-01 -5.14005832e-02
6.56730771e-01 9.90379035e-01 3.47362816e-01 -2.48820066e-01
-1.89524603e+00 -5.94170928e-01 1.05470371e+00 -8.79811525e-01
6.88050985e-01 -4.64723647e-01 -8.28907609e-01 5.44542909e-01
8.58342797e-02 6.89320965e-04 1.08798170e+00 7.15457380e-01
-7.29824424e-01 -2.50245184e-01 -5.73064506e-01 3.50650847e-01
6.20349169e-01 -1.31070590e+00 -7.09516823e-01 4.38320011e-01
6.52887106e-01 1.31247550e-01 -1.04035985e+00 3.45877558e-01
5.44879735e-01 -1.05313051e+00 1.33958721e+00 -1.09744644e+00
-8.25545564e-02 2.72335354e-02 -5.48630834e-01 -1.07246995e+00
-1.25794560e-01 -7.93567449e-02 -2.46745288e-01 1.01830626e+00
3.38603824e-01 -4.58524466e-01 7.61076033e-01 4.38941807e-01
3.65706116e-01 -7.13238597e-01 -8.25083673e-01 -5.55501938e-01
-6.16687238e-02 -1.91106066e-01 3.32605004e-01 7.76411474e-01
-4.78774607e-02 5.81636667e-01 -1.69491112e-01 2.64763147e-01
2.04051465e-01 5.34505129e-01 8.46696079e-01 -1.42728806e+00
-1.54325664e-01 -2.74287075e-01 -5.67002654e-01 -1.08811307e+00
6.37236163e-02 -9.87867951e-01 -4.77690876e-01 -1.41496575e+00
1.99253321e-01 -1.95508033e-01 -7.54248321e-01 -1.53447747e-01
6.11252822e-02 -4.06553090e-01 -1.31830305e-01 6.11420512e-01
-8.14223826e-01 6.86041236e-01 8.13148260e-01 -2.93547690e-01
-5.50803363e-01 4.39916074e-01 -6.24567330e-01 1.57484226e-02
4.27569836e-01 -2.12682143e-01 -1.04875243e+00 -4.89789903e-01
4.40807074e-01 1.01960011e-01 3.63074332e-01 -5.91011822e-01
4.21703964e-01 1.34384573e-01 8.80627111e-02 -5.97941458e-01
5.37449479e-01 -1.00757003e+00 -4.36747521e-01 3.65181565e-01
-8.61171782e-01 2.49141544e-01 -1.90193206e-01 1.21974552e+00
-6.98965490e-01 -1.66394021e-02 -7.45650604e-02 -3.41021270e-02
-7.87062407e-01 3.14139575e-01 -1.55765101e-01 7.80426636e-02
7.81375885e-01 -7.17779025e-02 -6.28455058e-02 -6.28256023e-01
-7.62623727e-01 4.86048371e-01 1.19552583e-01 1.05659580e+00
6.18710756e-01 -1.29005826e+00 -3.06458145e-01 1.16382532e-01
8.89519572e-01 -5.47400713e-01 8.16240683e-02 5.90177417e-01
1.44620731e-01 1.13732302e+00 3.22470248e-01 -7.85517931e-01
-1.15907991e+00 7.47820139e-01 9.62132029e-03 -6.30893886e-01
-1.48123786e-01 6.45044625e-01 5.03471643e-02 -8.00815523e-01
4.50822979e-01 -2.37731710e-02 -6.98837101e-01 6.04882300e-01
1.40593365e-01 2.17300594e-01 2.32161358e-01 -4.95243549e-01
-1.86829254e-01 5.03019035e-01 -6.15092039e-01 -2.86665976e-01
1.35265970e+00 -1.93954870e-01 -4.27505523e-02 8.75805855e-01
1.80046892e+00 -1.60401970e-01 -2.83121079e-01 -6.52736366e-01
6.44698501e-01 -3.36161166e-01 1.31456017e-01 -6.73692107e-01
-5.81031442e-01 7.66948342e-01 1.09212458e+00 4.85726476e-01
1.04464269e+00 5.50559878e-01 2.48255357e-01 8.15226674e-01
3.73268694e-01 -1.06043971e+00 4.69075024e-01 4.34734225e-01
9.18609619e-01 -1.27395475e+00 3.52400094e-02 1.86433226e-01
-5.24163544e-01 9.92665827e-01 4.42017496e-01 1.77145556e-01
8.14973414e-01 -7.55057454e-01 9.50727686e-02 -7.74583519e-01
-9.96853471e-01 -1.73370764e-01 6.95469320e-01 3.86306196e-01
6.02625370e-01 -1.29050836e-01 -1.73862204e-01 1.42853707e-01
4.33109403e-02 -2.43775785e-01 4.90148067e-02 1.17695510e+00
-1.65017903e-01 -1.46703434e+00 6.92188963e-02 5.62094152e-01
-9.29494873e-02 -2.32362058e-02 -5.18782139e-01 7.57556617e-01
-4.65336055e-01 8.74916494e-01 -4.42815498e-02 -4.95353609e-01
6.75759792e-01 5.35561800e-01 1.24057576e-01 -9.68856633e-01
-5.39695919e-01 -5.91099337e-02 -1.92465767e-01 -7.80657887e-01
-6.28458679e-01 -6.32647038e-01 -4.73071247e-01 4.38854486e-01
-4.07530725e-01 5.48471987e-01 7.13028550e-01 7.06004083e-01
6.38569474e-01 2.41261452e-01 9.72494960e-01 -3.94594133e-01
-1.08646357e+00 -9.27700162e-01 -5.22374570e-01 6.21496737e-01
2.91145504e-01 -6.00479245e-01 -4.32814270e-01 -2.46853143e-01] | [11.550670623779297, 7.608453750610352] |
9c19629c-cf12-42a0-b384-05ebdb6d543d | identity-driven-three-player-generative | 2305.00358 | null | https://arxiv.org/abs/2305.00358v1 | https://arxiv.org/pdf/2305.00358v1.pdf | Identity-driven Three-Player Generative Adversarial Network for Synthetic-based Face Recognition | Many of the commonly used datasets for face recognition development are collected from the internet without proper user consent. Due to the increasing focus on privacy in the social and legal frameworks, the use and distribution of these datasets are being restricted and strongly questioned. These databases, which have a realistically high variability of data per identity, have enabled the success of face recognition models. To build on this success and to align with privacy concerns, synthetic databases, consisting purely of synthetic persons, are increasingly being created and used in the development of face recognition solutions. In this work, we present a three-player generative adversarial network (GAN) framework, namely IDnet, that enables the integration of identity information into the generation process. The third player in our IDnet aims at forcing the generator to learn to generate identity-separable face images. We empirically proved that our IDnet synthetic images are of higher identity discrimination in comparison to the conventional two-player GAN, while maintaining a realistic intra-identity variation. We further studied the identity link between the authentic identities used to train the generator and the generated synthetic identities, showing very low similarities between these identities. We demonstrated the applicability of our IDnet data in training face recognition models by evaluating these models on a wide set of face recognition benchmarks. In comparison to the state-of-the-art works in synthetic-based face recognition, our solution achieved comparable results to a recent rendering-based approach and outperformed all existing GAN-based approaches. The training code and the synthetic face image dataset are publicly available ( https://github.com/fdbtrs/Synthetic-Face-Recognition ). | ['Naser Damer', 'Arjan Kuijper', 'Fadi Boutros', 'Jurek Elliesen', 'Tim Rieber', 'Jan Niklas Kolf'] | 2023-04-30 | null | null | null | null | ['face-recognition'] | ['computer-vision'] | [ 4.17747229e-01 2.66973555e-01 2.49350339e-01 -4.39436674e-01
-6.08149290e-01 -6.62090659e-01 1.01177871e+00 -7.99134016e-01
-1.91278830e-01 8.51824701e-01 -1.61486566e-01 -4.06163484e-02
8.82873386e-02 -1.00881529e+00 -6.15738928e-01 -7.33104408e-01
2.29578048e-01 6.55096769e-01 -3.62205058e-01 -2.41795853e-01
-2.19157487e-02 6.62301600e-01 -1.89861596e+00 3.19710523e-01
6.49513125e-01 1.08279848e+00 -4.12633449e-01 4.56969738e-01
7.26137832e-02 5.46043098e-01 -8.08823824e-01 -9.75370646e-01
6.43811703e-01 -9.05929863e-01 -6.00065112e-01 2.92243380e-02
5.39595008e-01 -4.02591765e-01 -2.94561952e-01 9.26716566e-01
8.20450306e-01 -2.02882573e-01 4.96518552e-01 -1.65846062e+00
-1.02395689e+00 3.00686419e-01 -2.14412272e-01 -2.91806072e-01
5.83751023e-01 2.69146174e-01 4.60327417e-01 -9.41713095e-01
8.24400902e-01 1.22203219e+00 6.20806694e-01 1.21623099e+00
-1.39357924e+00 -1.02457976e+00 -5.50079644e-01 -1.09580960e-02
-1.60222125e+00 -8.88910651e-01 5.78814566e-01 -3.72293979e-01
2.97355592e-01 4.30429906e-01 3.81253093e-01 1.81489742e+00
-3.85063678e-01 3.29212546e-01 1.49779844e+00 -5.83926201e-01
9.91424099e-02 4.96130228e-01 -5.41498661e-01 3.52627099e-01
1.62212253e-01 5.35068095e-01 -5.16307771e-01 -2.04210460e-01
8.43495846e-01 -3.25298369e-01 -3.51311147e-01 -3.06776971e-01
-8.05936515e-01 7.95151949e-01 9.70515758e-02 3.33867937e-01
-3.60346824e-01 -4.19375394e-03 2.07335591e-01 4.39262718e-01
4.49130982e-01 3.21843177e-01 3.08651086e-02 -1.59502327e-01
-9.41483200e-01 3.15108359e-01 1.03490949e+00 8.46699119e-01
5.71257353e-01 2.55162269e-01 -3.39156836e-01 8.98419440e-01
2.14677751e-01 5.27177811e-01 4.29681063e-01 -8.81750464e-01
1.18314467e-01 4.62402493e-01 -1.12084210e-01 -8.81724238e-01
3.60721529e-01 -2.52335936e-01 -9.18611646e-01 5.82897604e-01
6.48851335e-01 -1.39816418e-01 -1.04470336e+00 1.93825948e+00
2.76596487e-01 3.74294937e-01 2.74929672e-01 5.80105841e-01
7.92866766e-01 2.44800657e-01 9.24653467e-03 6.28848299e-02
1.18219960e+00 -3.81919086e-01 -5.53586900e-01 1.66405052e-01
6.77042976e-02 -9.21695292e-01 9.49414968e-01 1.56560734e-01
-1.02779746e+00 -5.99663198e-01 -9.00319993e-01 3.19369704e-01
-5.92724860e-01 4.18034457e-02 4.90441531e-01 1.60182142e+00
-1.49937987e+00 5.82873881e-01 -3.38992715e-01 -3.19472075e-01
1.03332782e+00 6.02408469e-01 -8.23339403e-01 -2.86504656e-01
-1.17858040e+00 7.08903253e-01 1.80903241e-01 1.73063815e-01
-1.09898865e+00 -7.98296094e-01 -6.92751825e-01 -1.69756755e-01
1.50636524e-01 -6.19988441e-01 9.25022840e-01 -1.49608874e+00
-1.66634154e+00 1.43131387e+00 1.62257016e-01 -3.67839247e-01
1.06211829e+00 1.55313134e-01 -6.19685113e-01 -1.14237130e-01
-8.21232796e-02 8.37288022e-01 1.02198565e+00 -1.56832540e+00
-1.06467776e-01 -3.65254641e-01 1.88683886e-02 -2.76179075e-01
-1.86345398e-01 2.65864760e-01 -3.17716867e-01 -6.62762105e-01
-5.30384421e-01 -9.80915427e-01 1.93106681e-01 -5.90730682e-02
-4.32211220e-01 1.47199109e-01 9.42277551e-01 -5.59219360e-01
4.96215492e-01 -2.10440183e+00 -1.48966402e-01 3.40281188e-01
-4.37117042e-03 7.31876433e-01 -3.34666222e-01 4.20565695e-01
-3.40816766e-01 4.70820010e-01 -2.04636455e-01 -6.78894818e-01
6.79943860e-02 2.01116383e-01 -3.18373144e-01 3.37238103e-01
4.99317378e-01 8.89234602e-01 -6.05682075e-01 -2.21028894e-01
-2.39370298e-02 1.00568700e+00 -5.26073456e-01 4.83442456e-01
2.23586541e-02 8.56061280e-01 -1.60941869e-01 6.81159675e-01
9.68459070e-01 2.61450052e-01 5.97065799e-02 5.87316789e-02
2.96642661e-01 -1.98524311e-01 -1.04802799e+00 1.31190205e+00
-3.71190995e-01 5.13748825e-01 7.01559708e-02 -6.65483773e-01
1.12046301e+00 6.22757971e-01 3.92861277e-01 -7.41244435e-01
2.94564158e-01 3.20828050e-01 1.32920399e-01 -1.58989191e-01
1.54637292e-01 -1.67003557e-01 2.42030054e-01 5.19030213e-01
4.05413181e-01 -1.81704566e-01 1.74307495e-01 -1.67799428e-01
8.23114216e-01 3.05202842e-01 -2.57228278e-02 -1.54723883e-01
8.55545044e-01 -4.00258660e-01 4.17241335e-01 5.28173447e-01
-6.74535111e-02 1.07624257e+00 5.04253566e-01 -3.30105960e-01
-1.07066000e+00 -1.05784738e+00 -1.88153327e-01 5.50463915e-01
-3.51808041e-01 -1.46061584e-01 -1.30025470e+00 -7.26469159e-01
-1.13558635e-01 6.29400730e-01 -8.60709250e-01 -1.64308965e-01
-4.21358615e-01 -6.52840614e-01 1.23204362e+00 1.41843274e-01
7.37012446e-01 -1.29856265e+00 -1.38238937e-01 -9.70862880e-02
1.27547204e-01 -1.10399628e+00 -2.75271177e-01 -4.11915094e-01
-2.84124315e-01 -1.23156428e+00 -8.23224247e-01 -5.26361048e-01
8.88306022e-01 -4.57684338e-01 1.27550256e+00 2.15698004e-01
-4.38311398e-01 5.75658560e-01 -3.69467854e-01 -6.22781157e-01
-1.01948261e+00 -6.06550612e-02 5.82532696e-02 5.21272480e-01
3.47462088e-01 -7.76969314e-01 -6.43308938e-01 5.25440216e-01
-1.14910579e+00 -1.26547113e-01 3.39139163e-01 9.39553261e-01
2.46956572e-01 -1.86453655e-01 8.50347519e-01 -1.22329485e+00
7.25736797e-01 -4.27675366e-01 -5.56909144e-01 1.75903291e-01
-5.42289495e-01 -1.71898678e-01 6.95331097e-01 -4.60196763e-01
-1.09251070e+00 3.76277114e-03 -5.60900509e-01 -4.32020903e-01
-4.73533273e-01 -1.08977832e-01 -5.33593833e-01 -5.04312575e-01
6.78789914e-01 2.15600640e-01 5.31969249e-01 -2.07751736e-01
2.89018333e-01 7.34777749e-01 5.12518942e-01 -7.29075968e-01
1.02538121e+00 3.24826360e-01 -3.95830870e-02 -6.04912400e-01
-8.83090720e-02 2.40226701e-01 -5.21242917e-01 -2.42950052e-01
5.33327758e-01 -7.30638444e-01 -8.42956245e-01 9.17195618e-01
-9.88186002e-01 -3.26542050e-01 -5.95155180e-01 1.32532090e-01
-6.22047246e-01 2.72408463e-02 -3.42543066e-01 -8.66196513e-01
-3.62161636e-01 -1.21893346e+00 9.53886628e-01 2.75320619e-01
-8.83535072e-02 -8.67985010e-01 3.93448174e-02 5.16835630e-01
8.37867320e-01 7.27496982e-01 5.19680619e-01 -7.18923926e-01
-4.54957455e-01 -3.22396964e-01 -6.87895417e-02 7.16260254e-01
3.94771039e-01 6.32402077e-02 -1.41295743e+00 -3.36922646e-01
1.00638501e-01 -4.53294963e-01 3.78711641e-01 -2.60890752e-01
1.01215398e+00 -3.79412234e-01 7.05564320e-02 7.29585469e-01
1.36249650e+00 2.36023366e-01 1.24370229e+00 6.60792831e-03
5.45815945e-01 8.68665576e-01 2.26808727e-01 3.43759567e-01
1.14134466e-02 8.12212229e-01 4.40749556e-01 -2.48259783e-01
-3.55973780e-01 -2.91874796e-01 2.31646284e-01 1.21422298e-01
-4.15137172e-01 -3.24742675e-01 -8.32678676e-01 2.63958573e-01
-1.36705947e+00 -1.12572038e+00 2.16365993e-01 2.39341664e+00
9.42551196e-01 -3.16741675e-01 2.30059490e-01 6.97235540e-02
7.67605007e-01 -1.35432780e-01 -3.74363661e-01 -3.49161595e-01
-2.15004355e-01 8.68547499e-01 2.87376374e-01 2.01452181e-01
-8.83777201e-01 8.72234523e-01 6.24078608e+00 7.76965618e-01
-1.30127430e+00 2.07009748e-01 1.08682287e+00 4.42840122e-02
-2.37861082e-01 -3.03918868e-01 -7.76644051e-01 5.75591922e-01
1.04680896e+00 -2.74898082e-01 6.60714984e-01 7.48622894e-01
-2.90701166e-02 4.07567710e-01 -1.22809052e+00 1.04801178e+00
2.87021071e-01 -1.32683730e+00 1.78712040e-01 3.94362420e-01
8.25462878e-01 -3.24734300e-01 4.45303530e-01 1.34687990e-01
3.83073121e-01 -1.64528370e+00 7.47405887e-01 4.50163215e-01
1.38579476e+00 -8.36427510e-01 9.06422257e-01 8.11551977e-03
-7.69106328e-01 1.79670721e-01 -1.44454464e-01 2.56864429e-01
-7.16038644e-02 2.82936126e-01 -9.12806869e-01 6.82612419e-01
5.58857322e-01 1.74746618e-01 -6.54258490e-01 5.92624247e-01
-1.78458214e-01 5.11447012e-01 -2.66626447e-01 4.10637975e-01
-1.67633191e-01 -3.40988547e-01 2.17242897e-01 8.03486764e-01
5.01604795e-01 -9.55920592e-02 -3.09717625e-01 1.41694498e+00
-4.78351116e-01 9.72432196e-02 -8.59253645e-01 -1.56848356e-01
4.31432307e-01 1.21542788e+00 -4.54631656e-01 3.48890610e-02
-2.41608843e-01 1.01069856e+00 4.84433770e-02 2.88715631e-01
-8.33876550e-01 -4.81167994e-02 9.35634077e-01 2.95915753e-01
1.79351732e-01 1.29110962e-01 -6.89469650e-02 -8.58341455e-01
1.28980964e-01 -1.54824984e+00 1.11694038e-01 -3.65377635e-01
-1.45849848e+00 1.18109262e+00 -8.59653503e-02 -1.05505490e+00
-6.41648412e-01 -5.86218059e-01 -6.43921793e-01 1.28788638e+00
-1.19750369e+00 -1.64816260e+00 -4.62071747e-01 7.95945406e-01
2.16695648e-02 -6.62862301e-01 1.32221735e+00 5.10747135e-01
-4.64535415e-01 1.11990738e+00 -9.86383632e-02 3.40442598e-01
6.43523753e-01 -8.29216778e-01 7.31828153e-01 7.64568031e-01
2.60645807e-01 6.87785268e-01 4.19570506e-01 -3.86877865e-01
-1.33624375e+00 -1.03692186e+00 4.20963079e-01 -5.61159551e-01
1.32888660e-01 -7.46699214e-01 -7.92229295e-01 5.21383762e-01
3.26041877e-01 1.10473953e-01 9.43526030e-01 -4.09144312e-01
-4.10381973e-01 -1.49439380e-01 -1.77411699e+00 5.25550723e-01
1.17493224e+00 -5.92029154e-01 -7.29444474e-02 1.07521608e-01
1.84004396e-01 -1.89221278e-01 -9.27901864e-01 5.30617654e-01
6.54899955e-01 -1.27834988e+00 1.05785787e+00 -3.90494317e-01
4.13200706e-01 -1.68248594e-01 -2.69064941e-02 -1.21667588e+00
1.27766445e-01 -8.15671444e-01 2.12948427e-01 1.88278222e+00
3.50640208e-01 -1.05664229e+00 1.15837479e+00 7.95974433e-01
3.93526703e-01 -5.38253725e-01 -9.92068589e-01 -8.28583956e-01
1.43454373e-01 -1.02900833e-01 9.55010414e-01 9.17774618e-01
-7.59044588e-01 -1.73631772e-01 -4.83588487e-01 -1.31052762e-01
6.33141100e-01 -2.81320930e-01 1.07592154e+00 -1.07607305e+00
-2.22293317e-01 -3.66997600e-01 -7.55399644e-01 -1.42749205e-01
4.18156296e-01 -9.48823690e-01 -2.81097412e-01 -9.10106778e-01
-1.48875296e-01 -6.95730984e-01 1.87401101e-02 5.16041636e-01
3.22126448e-02 9.52523887e-01 2.29560986e-01 3.89677323e-02
3.11886102e-01 5.11127889e-01 1.12732983e+00 8.44357535e-03
1.24268517e-01 6.66497648e-02 -7.88316131e-01 3.05442154e-01
9.30365086e-01 -4.45123464e-01 -5.30888498e-01 -1.55526074e-02
-2.41831928e-01 -3.57306004e-01 3.96079153e-01 -1.20781636e+00
3.35239917e-02 1.87786490e-01 5.24177432e-01 1.58078983e-01
5.03674984e-01 -8.56567502e-01 8.90631020e-01 3.50383401e-01
-7.45946467e-02 -1.83818430e-01 2.78484464e-01 -2.37947721e-02
-3.48608643e-01 -4.72074337e-02 8.98739517e-01 -1.14821985e-01
-4.38919544e-01 5.24581015e-01 1.11536860e-01 3.62189189e-02
1.27798522e+00 -5.08351505e-01 -2.70917505e-01 -5.46463192e-01
-5.11332214e-01 -3.12129915e-01 7.61541307e-01 6.04177475e-01
5.70459366e-01 -1.40597224e+00 -1.11059308e+00 7.37575531e-01
3.69516388e-02 -3.50635588e-01 1.16784230e-01 4.13809448e-01
-5.96619368e-01 -1.45645428e-03 -7.61019409e-01 -3.10661316e-01
-1.36364973e+00 4.55595821e-01 5.74061930e-01 -1.07229359e-01
-2.08099693e-01 7.45416164e-01 2.22446516e-01 -6.21261835e-01
4.55401056e-02 5.18820584e-01 2.19634585e-02 -1.16753355e-01
5.89982986e-01 2.02923164e-01 4.95801158e-02 -9.83108044e-01
-2.53271013e-01 3.62441540e-01 9.09693465e-02 -2.42382929e-01
1.29871273e+00 3.55848998e-01 -2.33885005e-01 -1.24603823e-01
1.09766829e+00 1.37530550e-01 -1.17898858e+00 1.65675849e-01
-2.05301225e-01 -8.91507983e-01 -5.73420346e-01 -8.50823224e-01
-1.40817797e+00 5.54753423e-01 8.40022624e-01 2.02961773e-01
1.16250062e+00 -2.97957301e-01 3.88160229e-01 -2.01781809e-01
7.18185425e-01 -5.66114366e-01 -1.00711107e-01 9.78222564e-02
1.12563467e+00 -1.28201556e+00 -3.11209768e-01 -7.80799448e-01
-4.67681289e-01 9.21184957e-01 5.83580911e-01 -2.32828036e-03
4.34453934e-01 4.12261218e-01 3.86325061e-01 2.79212724e-02
-2.93737799e-01 2.61748638e-02 2.20605463e-01 1.19316614e+00
4.60925758e-01 -1.30147696e-01 -9.19930413e-02 4.57039863e-01
-5.03763199e-01 1.31946236e-01 3.10846120e-01 6.83152676e-01
5.36523640e-01 -1.93546343e+00 -5.27872384e-01 2.00400352e-01
-7.84918606e-01 1.27267260e-02 -7.97589302e-01 9.34821844e-01
4.17802662e-01 9.60326612e-01 -9.60596129e-02 -3.96482527e-01
3.35406363e-01 2.61061043e-01 5.92014730e-01 -4.74607885e-01
-9.22334969e-01 -6.87796950e-01 1.48874551e-01 -5.60372353e-01
-3.65302473e-01 -5.63197076e-01 -5.72153389e-01 -6.16356969e-01
-9.73420590e-02 6.67106956e-02 8.60058844e-01 5.17096937e-01
5.88293850e-01 3.34391415e-01 7.80879378e-01 -8.78664970e-01
-5.13849139e-01 -8.74960423e-01 -5.13002038e-01 8.57407808e-01
-1.66036785e-01 -5.09441197e-01 -3.19188982e-01 2.84171224e-01] | [12.84754467010498, 0.5794481039047241] |
fa088547-71e5-4d6b-ad45-fa38e931a81e | hyspecnet-11k-a-large-scale-hyperspectral | 2306.00385 | null | https://arxiv.org/abs/2306.00385v2 | https://arxiv.org/pdf/2306.00385v2.pdf | HySpecNet-11k: A Large-Scale Hyperspectral Dataset for Benchmarking Learning-Based Hyperspectral Image Compression Methods | The development of learning-based hyperspectral image compression methods has recently attracted great attention in remote sensing. Such methods require a high number of hyperspectral images to be used during training to optimize all parameters and reach a high compression performance. However, existing hyperspectral datasets are not sufficient to train and evaluate learning-based compression methods, which hinders the research in this field. To address this problem, in this paper we present HySpecNet-11k that is a large-scale hyperspectral benchmark dataset made up of 11,483 nonoverlapping image patches. Each patch is a portion of 128 $\times$ 128 pixels with 224 spectral bands and a ground sample distance of 30 m. We exploit HySpecNet-11k to benchmark the current state of the art in learning-based hyperspectral image compression by focussing our attention on various 1D, 2D and 3D convolutional autoencoder architectures. Nevertheless, HySpecNet-11k can be used for any unsupervised learning task in the framework of hyperspectral image analysis. The dataset, our code and the pre-trained weights are publicly available at https://hyspecnet.rsim.berlin . | ['Begüm Demir', 'Martin Hermann Paul Fuchs'] | 2023-06-01 | null | null | null | null | ['image-compression'] | ['computer-vision'] | [ 8.01259398e-01 -4.59303886e-01 1.53255656e-01 -2.68296063e-01
-5.67520022e-01 -3.34674031e-01 2.91811764e-01 1.04492478e-01
-2.53628612e-01 6.14327788e-01 -1.37253478e-01 -2.78389901e-01
-6.09982014e-01 -1.23233593e+00 -6.35592341e-01 -1.06665146e+00
-3.31647158e-01 1.94035426e-01 -4.78537291e-01 -7.12608546e-02
-2.32528914e-02 5.85577786e-01 -1.72339928e+00 3.36320668e-01
8.40471268e-01 1.29177332e+00 5.77039599e-01 5.95078647e-01
3.96613747e-01 3.77663344e-01 -2.76849031e-01 -1.08109498e-02
7.64106989e-01 -5.00246584e-01 -7.68164217e-01 2.26724133e-01
6.70612991e-01 -5.56552768e-01 -6.11981034e-01 1.42237544e+00
7.78875947e-01 2.57454693e-01 5.82470119e-01 -8.22503328e-01
-4.88904625e-01 7.55485177e-01 -4.93606865e-01 1.77610070e-01
-5.64416707e-01 1.25224784e-01 9.69200909e-01 -7.37617671e-01
2.51411468e-01 5.60819566e-01 6.95913732e-01 4.75325175e-02
-1.10952365e+00 -5.72672248e-01 -4.15497094e-01 3.57460618e-01
-1.63065112e+00 -2.60988384e-01 7.03134000e-01 -4.09688264e-01
1.01428485e+00 4.74528462e-01 9.87463713e-01 5.43630481e-01
-1.22516394e-01 1.58837378e-01 1.24046183e+00 -3.17612350e-01
3.74494731e-01 -4.02227730e-01 1.33165410e-02 3.88872027e-01
5.29846072e-01 5.07731318e-01 -2.84942299e-01 -4.97935191e-02
5.79916716e-01 2.35644296e-01 -6.93860352e-01 5.24209701e-02
-7.72741258e-01 1.02906787e+00 8.29001188e-01 3.29880774e-01
-4.70147461e-01 -9.27439108e-02 1.74861580e-01 3.69200915e-01
5.89772582e-01 5.58342338e-01 -5.64278543e-01 3.28533590e-01
-1.08017504e+00 1.24158375e-01 4.93628889e-01 6.27529383e-01
1.11320806e+00 2.38525331e-01 2.87158430e-01 1.25097680e+00
8.29716176e-02 7.39051461e-01 6.60867274e-01 -8.16538393e-01
2.65754402e-01 4.86320674e-01 -3.36590111e-01 -9.81762648e-01
-3.28918695e-01 -6.46744907e-01 -1.40912652e+00 3.44264746e-01
-4.24091339e-01 -1.54545903e-01 -7.96237528e-01 1.07717514e+00
-2.19283719e-02 1.90836579e-01 3.40580583e-01 1.03273320e+00
8.22514296e-01 1.10088909e+00 -2.52547562e-01 -1.90911993e-01
9.60422695e-01 -9.30785000e-01 -2.50257164e-01 -2.37183392e-01
4.22518700e-01 -7.09277213e-01 8.27992022e-01 4.39868540e-01
-7.77311921e-01 -4.37748343e-01 -1.26522017e+00 2.40222052e-01
-6.90729618e-01 3.96690786e-01 7.32080400e-01 4.57346678e-01
-8.02607417e-01 9.09767628e-01 -6.25155747e-01 -1.23891093e-01
7.51721442e-01 2.00616121e-01 -3.97124797e-01 -2.82838374e-01
-1.01126611e+00 5.98606706e-01 1.00645757e+00 1.38129577e-01
-9.70172584e-01 -7.53154993e-01 -7.48238206e-01 2.89045066e-01
4.22588326e-02 -2.60773480e-01 1.10004151e+00 -8.77542138e-01
-1.13419807e+00 7.40034878e-01 3.70464802e-01 -5.80531061e-01
-1.26952052e-01 -2.35629687e-03 -6.16953552e-01 4.88030493e-01
-2.87783712e-01 6.03247881e-01 1.01698506e+00 -1.09481764e+00
-5.41421354e-01 -4.46735412e-01 -2.24612489e-01 1.91623658e-01
-8.00088108e-01 -5.69819748e-01 1.52574882e-01 -7.13700116e-01
2.06139728e-01 -1.00719762e+00 -1.62536353e-01 8.02288949e-03
-3.22136313e-01 4.55474854e-01 1.07683325e+00 -5.67421615e-01
9.14292157e-01 -2.09419966e+00 7.93962553e-03 2.07175836e-01
3.30164246e-02 7.31777310e-01 -5.56678832e-01 4.85274106e-01
-4.98728722e-01 1.33802667e-01 -9.23864484e-01 -9.11333412e-02
-1.30696893e-01 1.62950754e-01 -3.57718140e-01 6.82916820e-01
7.85529763e-02 6.04660332e-01 -4.96228069e-01 -4.74926345e-02
4.59995657e-01 6.00472093e-01 -4.44447219e-01 3.99244368e-01
-3.76939178e-01 8.97681490e-02 -2.22561523e-01 9.56907094e-01
1.18772900e+00 -3.59558374e-01 -4.12003808e-02 -3.53895098e-01
-2.41868213e-01 -6.33805767e-02 -7.05260515e-01 1.74000287e+00
-1.42677948e-01 7.38935769e-01 1.02883197e-01 -1.40927947e+00
8.52835417e-01 2.16346607e-01 8.49860191e-01 -5.87323725e-01
1.89795289e-02 3.08705509e-01 -6.00088015e-02 -4.91722256e-01
5.00220597e-01 -2.43668109e-01 7.08747804e-01 4.70990032e-01
5.73800094e-02 -7.30396807e-01 2.98569620e-01 -4.30124581e-01
7.18832374e-01 -3.36171418e-01 1.42069772e-01 -2.55584925e-01
3.91272604e-01 3.64025980e-01 1.00041673e-01 4.86604244e-01
1.13282338e-01 5.35532773e-01 -1.42078355e-01 -4.83467758e-01
-1.40048230e+00 -3.93957347e-01 -7.65254915e-01 7.84942865e-01
-9.60035101e-02 -2.64385074e-01 -5.62584043e-01 -1.37399733e-01
9.46292430e-02 4.41478640e-01 -5.12319565e-01 -1.33864477e-01
6.27283156e-02 -1.36579359e+00 5.66876709e-01 1.74758248e-02
9.32600796e-01 -1.14199007e+00 -7.30034947e-01 1.87428355e-01
-2.72871815e-02 -8.50615561e-01 1.15484223e-01 4.06308919e-01
-1.27529836e+00 -1.30521405e+00 -5.37529647e-01 -6.15607023e-01
4.96796250e-01 7.67755210e-01 9.30789590e-01 -2.67373055e-01
-5.30180573e-01 8.25296864e-02 -6.26576424e-01 -6.89248860e-01
-2.96842396e-01 2.95886189e-01 -3.50238830e-01 -1.89448386e-01
3.02067101e-01 -9.73550737e-01 -6.50587201e-01 6.15568571e-02
-1.64402020e+00 1.26634672e-01 8.21581900e-01 9.13155138e-01
1.03793597e+00 8.58536541e-01 7.62267876e-03 -5.85948706e-01
2.58141607e-01 -5.13767838e-01 -8.62040579e-01 1.42820656e-01
-7.63630390e-01 -3.38839501e-01 6.87155068e-01 -5.25794141e-02
-6.59060001e-01 9.98024195e-02 -7.41801485e-02 -4.68529135e-01
-3.14340800e-01 1.36573982e+00 1.77223295e-01 -3.80960733e-01
9.14024293e-01 7.02169597e-01 8.57003778e-02 -4.85117078e-01
-2.83639412e-02 9.13427532e-01 3.76389325e-01 -1.36364460e-01
8.92796218e-01 4.97810692e-01 1.70006648e-01 -1.26990545e+00
-8.62858891e-01 -4.32237774e-01 -4.22423661e-01 -9.95620936e-02
5.05009830e-01 -1.26995385e+00 -3.27650994e-01 6.15368187e-01
-6.46466196e-01 -6.46784663e-01 -3.12347889e-01 6.28464460e-01
-3.33360493e-01 3.40819091e-01 -3.73169988e-01 -4.49378937e-01
-8.95483434e-01 -9.46015298e-01 8.32833171e-01 -1.48873059e-02
7.42550910e-01 -6.91380739e-01 1.71731874e-01 2.32953236e-01
7.24789023e-01 3.75726014e-01 7.53151715e-01 -2.25698873e-01
-4.98649120e-01 -3.08875620e-01 -2.99063414e-01 8.64503086e-01
1.75276622e-01 -4.54835296e-02 -1.10794628e+00 -6.35870516e-01
1.39585450e-01 -8.71776104e-01 1.33086097e+00 5.18370807e-01
1.94081795e+00 -4.01106387e-01 1.40657246e-01 1.17311573e+00
1.98912024e+00 1.70111805e-01 7.97780991e-01 4.02312964e-01
3.98158342e-01 4.86790568e-01 1.86061651e-01 7.89381564e-01
-2.67999321e-01 7.24432394e-02 1.11364424e+00 -1.86628774e-01
2.21496522e-01 9.04672444e-02 -8.34487751e-02 8.09597969e-01
-3.51874471e-01 -2.82165468e-01 -1.00291884e+00 1.92892343e-01
-1.33222747e+00 -1.26043391e+00 -3.72597501e-02 1.99429214e+00
8.68407369e-01 -5.81032693e-01 -5.01210272e-01 5.54305494e-01
5.33976674e-01 6.17040813e-01 -7.33333647e-01 3.39425921e-01
-5.13772011e-01 5.85461080e-01 8.11362147e-01 2.28057578e-01
-1.57792628e+00 6.03441715e-01 5.56187296e+00 7.52220511e-01
-1.57930529e+00 -7.67312422e-02 7.18063951e-01 -5.27417175e-02
-1.29846394e-01 -1.62552148e-01 -3.77044231e-01 2.19369665e-01
1.01448166e+00 1.10095106e-02 9.16900396e-01 8.71473789e-01
1.31615147e-01 -5.33042476e-02 -5.51116288e-01 1.34218383e+00
8.62791315e-02 -1.44867718e+00 1.89113334e-01 1.94562182e-01
1.09119809e+00 6.26144946e-01 3.04130554e-01 -5.00874966e-02
-1.43652841e-01 -1.48556042e+00 3.12881768e-01 3.61679912e-01
1.15978765e+00 -7.59004354e-01 6.70085847e-01 2.57868469e-01
-1.03446448e+00 -2.26516932e-01 -1.16503882e+00 1.12478964e-01
-4.57401395e-01 1.13383603e+00 -4.68165606e-01 8.50175858e-01
7.71412909e-01 1.19477189e+00 -4.72347379e-01 1.16599560e+00
1.50958344e-01 9.02432501e-01 -3.88552874e-01 3.91487300e-01
4.92225319e-01 -5.72095573e-01 3.02111119e-01 1.06130254e+00
8.58237922e-01 5.54075241e-01 -4.45930287e-03 6.83245718e-01
-2.17298552e-01 7.84358829e-02 -7.16343641e-01 -6.15212858e-01
2.01401398e-01 1.29463685e+00 -2.06679106e-01 -1.95452839e-01
-2.31172711e-01 6.69317424e-01 -4.64706533e-02 3.08968037e-01
-6.58151984e-01 -3.52829278e-01 7.80942500e-01 -1.19199812e-01
4.02323872e-01 -2.95380175e-01 -4.64407504e-02 -1.24490309e+00
-3.00649434e-01 -1.15612507e+00 4.37056273e-01 -1.07654500e+00
-1.20701635e+00 8.21070552e-01 -6.51397333e-02 -1.46701825e+00
2.13108867e-01 -1.00221181e+00 -2.93100446e-01 8.92977118e-01
-2.25841975e+00 -9.27789390e-01 -1.17661572e+00 6.87779903e-01
2.39934519e-01 -4.90763873e-01 1.35439599e+00 2.77509600e-01
-6.29700840e-01 -2.05027983e-02 5.77216268e-01 8.83869734e-03
2.85308570e-01 -8.23566854e-01 -1.78516343e-01 6.95056438e-01
4.35194373e-02 1.88004114e-02 3.41541797e-01 -2.66177773e-01
-1.56234896e+00 -1.75856650e+00 4.19957191e-01 4.67611223e-01
5.38667262e-01 3.47842187e-01 -9.89324629e-01 4.78591412e-01
3.44800502e-01 2.97302306e-01 1.11593807e+00 -5.31266332e-01
-3.01269501e-01 -5.28820217e-01 -1.00950408e+00 1.15771703e-01
8.53563249e-01 -5.66734791e-01 -1.58337310e-01 8.25624347e-01
3.61246198e-01 -2.56213486e-01 -1.13015783e+00 6.33037925e-01
2.17145294e-01 -1.13974023e+00 9.78506565e-01 -1.60000071e-01
1.11259234e+00 -1.97574481e-01 -5.25246620e-01 -1.45118189e+00
-4.47466403e-01 -9.73460898e-02 6.98522180e-02 3.91215026e-01
3.75953823e-01 -6.50981307e-01 6.66450560e-01 -5.24591953e-02
-4.86613899e-01 -5.07515073e-01 -6.38271034e-01 -6.69430912e-01
5.75631894e-02 -2.56717414e-01 9.04730678e-01 1.11878645e+00
-3.02940935e-01 -1.39807433e-01 -2.90934652e-01 4.90331024e-01
6.55940950e-01 5.40581822e-01 4.10111189e-01 -1.31492043e+00
-3.59099001e-01 -6.27597153e-01 -1.33318171e-01 -6.77924216e-01
8.93816426e-02 -1.14723444e+00 -8.65751281e-02 -1.30662441e+00
1.56571046e-01 -3.78118396e-01 -1.94049552e-01 8.41488361e-01
2.60834813e-01 4.75104153e-01 -8.34450349e-02 6.18674397e-01
1.72059968e-01 8.88985038e-01 1.22692311e+00 -7.14196205e-01
-1.76009710e-03 -3.22618663e-01 -5.08620560e-01 2.63276905e-01
1.50600898e+00 -3.76942605e-01 -4.56492692e-01 -7.81314373e-01
6.64941296e-02 -7.76218921e-02 4.94502038e-01 -1.39278018e+00
1.56538248e-01 -3.50827456e-01 5.91580212e-01 -6.54411674e-01
3.54922533e-01 -1.05180097e+00 5.69124937e-01 6.31737292e-01
-3.03736210e-01 -2.95962602e-01 3.21790636e-01 3.39956939e-01
-6.81340575e-01 -5.64363897e-01 9.71871495e-01 -3.30927402e-01
-8.27790439e-01 9.60896730e-01 -4.39873226e-02 -5.19992590e-01
8.77403438e-01 4.01860401e-02 -4.29930925e-01 -7.67686963e-02
-4.05395836e-01 -3.31336528e-01 3.82818848e-01 -2.67146349e-01
8.42045605e-01 -1.02597237e+00 -1.06946921e+00 5.59471667e-01
4.42502469e-01 1.77420661e-01 4.81440336e-01 3.80846292e-01
-1.19147527e+00 4.78784233e-01 -3.98582727e-01 -3.71214122e-01
-1.00910282e+00 2.62521029e-01 6.43826663e-01 -3.71571295e-02
-5.49749553e-01 7.41034746e-01 -3.29911679e-01 -3.99754941e-01
-8.84084925e-02 -1.22503921e-01 -2.50017196e-01 -3.02507281e-02
6.54346228e-01 2.57690340e-01 3.52544963e-01 -4.78618771e-01
1.02616161e-01 3.91027749e-01 5.34702957e-01 1.30680338e-01
2.07955289e+00 2.84172624e-01 -6.29212320e-01 -7.98628703e-02
1.38567340e+00 -5.10344565e-01 -1.24236393e+00 -2.15339288e-01
-5.90991080e-01 -7.07970977e-01 6.07218564e-01 -7.88087606e-01
-1.51396382e+00 8.09511483e-01 8.58262539e-01 3.96293640e-01
1.74946737e+00 -4.98869598e-01 5.36834478e-01 7.35573351e-01
1.82331368e-01 -9.23301220e-01 -1.60809204e-01 6.22706056e-01
1.19517398e+00 -1.41714883e+00 2.68700331e-01 -1.50944471e-01
-2.67940611e-01 1.29825139e+00 2.49949008e-01 1.85248200e-02
9.27023351e-01 5.67701347e-02 -5.73696978e-02 -2.86077797e-01
-2.35865578e-01 -2.36581817e-01 1.80458471e-01 5.44353247e-01
3.77738833e-01 3.24655563e-01 -9.43041816e-02 6.32188618e-02
-4.77217615e-01 -6.16232678e-02 2.86051840e-01 8.93361330e-01
-7.23914027e-01 -8.75197828e-01 -3.63743097e-01 8.20654452e-01
-2.36337319e-01 -4.23437685e-01 -4.50962596e-02 3.85201931e-01
1.82895526e-01 7.63531506e-01 -6.15902953e-02 -4.15927410e-01
7.63074085e-02 -1.09866597e-01 1.20962813e-01 -4.29080218e-01
-1.77459583e-01 -4.68936004e-02 -3.38639408e-01 -1.81485280e-01
-9.02442038e-01 -4.54385042e-01 -6.42713547e-01 -5.81303418e-01
-2.42593482e-01 1.55425861e-01 9.99559462e-01 3.81170988e-01
2.97902912e-01 2.45217592e-01 7.58409560e-01 -1.05892980e+00
-7.92030513e-01 -1.34610653e+00 -1.20736372e+00 1.74166292e-01
3.49191934e-01 -3.65496017e-02 -4.67511475e-01 2.41932169e-01] | [10.072078704833984, -1.9165961742401123] |
f91cf2e6-e606-45e3-bf45-f237171763fc | few-shot-in-context-learning-for-knowledge | 2305.01750 | null | https://arxiv.org/abs/2305.01750v2 | https://arxiv.org/pdf/2305.01750v2.pdf | Few-shot In-context Learning for Knowledge Base Question Answering | Question answering over knowledge bases is considered a difficult problem due to the challenge of generalizing to a wide variety of possible natural language questions. Additionally, the heterogeneity of knowledge base schema items between different knowledge bases often necessitates specialized training for different knowledge base question-answering (KBQA) datasets. To handle questions over diverse KBQA datasets with a unified training-free framework, we propose KB-BINDER, which for the first time enables few-shot in-context learning over KBQA tasks. Firstly, KB-BINDER leverages large language models like Codex to generate logical forms as the draft for a specific question by imitating a few demonstrations. Secondly, KB-BINDER grounds on the knowledge base to bind the generated draft to an executable one with BM25 score matching. The experimental results on four public heterogeneous KBQA datasets show that KB-BINDER can achieve a strong performance with only a few in-context demonstrations. Especially on GraphQA and 3-hop MetaQA, KB-BINDER can even outperform the state-of-the-art trained models. On GrailQA and WebQSP, our model is also on par with other fully-trained models. We believe KB-BINDER can serve as an important baseline for future research. Our code is available at https://github.com/ltl3A87/KB-BINDER. | ['Wenhu Chen', 'Yu Su', 'Yu Gu', 'Alex Zhuang', 'Xueguang Ma', 'Tianle Li'] | 2023-05-02 | null | null | null | null | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-2.72631437e-01 3.05510670e-01 -1.09331027e-01 -1.94595799e-01
-1.55095196e+00 -8.70944262e-01 2.88814545e-01 2.74773203e-02
-8.42337906e-02 8.50008428e-01 1.90763801e-01 -7.26924181e-01
-1.18587010e-01 -1.09687364e+00 -1.15774751e+00 1.92322750e-02
2.84571558e-01 8.53170872e-01 8.03542495e-01 -9.35416877e-01
-1.39087588e-01 -1.11217350e-01 -1.26441646e+00 7.70216346e-01
1.15269184e+00 7.14378059e-01 2.16533512e-01 9.64829862e-01
-4.90901589e-01 1.43605089e+00 -7.37176955e-01 -1.10912848e+00
8.01258013e-02 -3.66636038e-01 -1.53518140e+00 -4.74197924e-01
8.53020668e-01 -5.06284058e-01 -3.91884536e-01 6.92078948e-01
3.31932813e-01 1.10214561e-01 3.91754806e-01 -1.40788412e+00
-1.05937576e+00 7.67746150e-01 3.61272953e-02 3.11557472e-01
8.76609981e-01 5.55742264e-01 1.64765656e+00 -6.12512529e-01
7.20990658e-01 1.30253434e+00 3.74975920e-01 8.51242125e-01
-8.94404411e-01 -3.60810757e-01 2.62547024e-02 8.98616314e-01
-1.19288242e+00 -2.27533802e-01 3.83228958e-01 -3.06870908e-01
1.34592140e+00 2.86823243e-01 5.24495721e-01 1.04607642e+00
-1.96546867e-01 1.10944414e+00 7.16298103e-01 -4.22889680e-01
3.77885215e-02 -1.22012943e-01 5.24864316e-01 1.20747018e+00
1.29846647e-01 -4.23927277e-01 -4.52436358e-01 -3.74072075e-01
2.70892411e-01 -4.12970245e-01 -4.24398869e-01 -3.05142999e-01
-1.06796372e+00 9.34039533e-01 3.37413162e-01 -6.11388013e-02
-1.04720615e-01 4.27741379e-01 3.93475920e-01 6.71333432e-01
-1.79775745e-01 9.53378856e-01 -7.79463112e-01 -4.00110453e-01
-3.16573203e-01 7.35819638e-01 1.34974444e+00 1.41931677e+00
9.98444378e-01 -3.21725488e-01 -4.11060542e-01 5.71071744e-01
2.08105206e-01 6.10008061e-01 3.19218934e-01 -1.04446137e+00
1.02564776e+00 8.25654984e-01 2.28938267e-01 -3.22092175e-01
-5.28233312e-02 4.21056151e-03 1.57396891e-03 -4.06785280e-01
5.65665722e-01 -1.96779415e-01 -7.08596587e-01 1.51389182e+00
5.22937715e-01 1.81933403e-01 3.59292030e-01 7.57268488e-01
1.30087399e+00 9.66937006e-01 7.27871880e-02 3.44420075e-01
1.69903672e+00 -1.50687695e+00 -4.05022800e-01 -2.81186998e-01
9.59961295e-01 -3.50466698e-01 1.77941155e+00 2.86325932e-01
-1.12836874e+00 -3.22349250e-01 -8.47790241e-01 -6.10037386e-01
-3.78378361e-01 -1.89319894e-01 5.30366957e-01 5.00148356e-01
-1.12268877e+00 -7.09241405e-02 -5.67421138e-01 -4.05600965e-01
1.93437517e-01 -1.82235390e-01 -5.77547587e-02 -5.15520453e-01
-1.47609031e+00 9.33555603e-01 4.51979488e-01 -1.46032467e-01
-1.38896382e+00 -1.05084658e+00 -1.00330436e+00 1.18951090e-01
9.03171659e-01 -1.25687933e+00 1.89759672e+00 -3.23020250e-01
-1.66040897e+00 5.04777789e-01 -1.56195194e-01 -5.03431499e-01
2.43415967e-01 -4.22465473e-01 -4.32593942e-01 3.14681470e-01
2.74850875e-01 6.65355384e-01 5.13637304e-01 -8.86141181e-01
-5.24366856e-01 -6.12720251e-02 1.02106833e+00 2.40905464e-01
8.51703733e-02 -2.38040581e-01 -6.68141365e-01 -1.40172811e-02
-4.73609388e-01 -7.30368078e-01 7.89032280e-02 -3.66960227e-01
-2.37403259e-01 -7.28608787e-01 3.78154516e-01 -8.08333576e-01
1.18682027e+00 -1.79059982e+00 2.22719371e-01 -3.91188145e-01
7.57999197e-02 2.73788780e-01 -5.55170596e-01 7.49618053e-01
5.05936146e-01 1.50632367e-01 -1.20297730e-01 2.39565462e-01
3.55976909e-01 3.64677966e-01 -7.85037398e-01 -1.94438368e-01
6.10778391e-01 1.47186267e+00 -1.21955276e+00 -5.48825622e-01
-2.02664852e-01 -8.13669488e-02 -7.95396149e-01 6.07432246e-01
-1.17480183e+00 -3.81291844e-02 -6.94873929e-01 7.69260108e-01
1.66345358e-01 -6.83990717e-01 4.82840203e-02 -5.07913306e-02
5.47001779e-01 8.26884568e-01 -7.32832730e-01 2.10452271e+00
-6.69112086e-01 3.54528606e-01 -2.08081782e-01 -7.63635755e-01
6.50662005e-01 4.41932917e-01 -4.11486089e-01 -7.66935170e-01
-4.31351274e-01 2.54913956e-01 1.27091452e-01 -1.01928270e+00
6.77385747e-01 -1.89713687e-02 -2.55803019e-01 3.48251015e-01
4.59034681e-01 -6.94612384e-01 6.05760574e-01 6.57581925e-01
1.53227401e+00 1.91850677e-01 1.79401994e-01 1.14201270e-01
4.99204338e-01 4.32450354e-01 1.96946263e-01 1.05116868e+00
-1.42768756e-01 1.85910851e-01 4.59081501e-01 -2.56458133e-01
-6.64041996e-01 -1.31306279e+00 3.67864728e-01 1.54036272e+00
1.70665860e-01 -6.17105782e-01 -5.20228326e-01 -1.08163249e+00
7.46973306e-02 1.13753128e+00 -3.01728427e-01 -1.43555000e-01
-6.45552218e-01 -6.80869892e-02 1.00869536e+00 4.90990818e-01
4.43661958e-01 -1.24851465e+00 -3.52655143e-01 1.93417162e-01
-5.44159770e-01 -1.18652689e+00 -2.12899134e-01 -1.47489861e-01
-7.40745962e-01 -1.33051264e+00 -3.47376108e-01 -6.89009309e-01
2.04950735e-01 3.24050486e-01 1.83844018e+00 3.27993393e-01
-1.86655194e-01 9.93806362e-01 -7.42592692e-01 -3.93131495e-01
-6.19096875e-01 4.81802046e-01 -6.52672827e-01 -7.71337211e-01
3.30577374e-01 -2.62375981e-01 -5.28725147e-01 1.46901652e-01
-8.71638119e-01 -6.81601698e-03 5.30139089e-01 6.59215569e-01
3.18303525e-01 -4.42907393e-01 8.47432137e-01 -9.49520051e-01
8.66655469e-01 -8.12649012e-01 -6.76379681e-01 8.14405620e-01
-2.22165778e-01 1.09435342e-01 6.77930951e-01 -2.23183200e-01
-1.13909280e+00 -6.72825694e-01 -2.63586581e-01 -5.97193763e-02
-9.39368904e-02 7.71659553e-01 -1.65381998e-01 3.27922165e-01
1.00573218e+00 3.35050225e-01 -3.84245217e-01 -2.50077665e-01
9.87737119e-01 3.89232278e-01 3.73583287e-01 -1.23369265e+00
8.11978579e-01 1.26620963e-01 -4.97277975e-01 -5.53707063e-01
-1.15657055e+00 -6.70405984e-01 -1.12140365e-01 1.57543272e-01
7.61570394e-01 -1.09169841e+00 -7.59301901e-01 -2.63559222e-02
-1.23665988e+00 -9.80953813e-01 -3.61593813e-01 2.21940060e-03
-6.87814355e-01 3.43700379e-01 -7.54192352e-01 -4.82674271e-01
-6.06597900e-01 -1.13568270e+00 1.05103397e+00 1.56760201e-01
-2.02755064e-01 -9.92408931e-01 3.84424746e-01 1.18205619e+00
2.63095766e-01 -4.20139790e-01 1.39637196e+00 -8.66105378e-01
-1.06287563e+00 5.68890348e-02 -1.10862628e-01 3.73485208e-01
-8.77493769e-02 -1.11759037e-01 -6.40426815e-01 -9.81776640e-02
-4.46192890e-01 -1.26336181e+00 6.77308798e-01 -3.31736267e-01
9.51439500e-01 -4.84494328e-01 7.92281926e-02 2.25739151e-01
1.18864727e+00 -1.21376410e-01 7.80234277e-01 5.10302246e-01
6.09878480e-01 2.12354198e-01 7.30457187e-01 -1.06209479e-01
1.01567161e+00 4.46742833e-01 4.30560797e-01 4.72821355e-01
-2.81252563e-01 -4.39700782e-01 5.54016113e-01 9.99758959e-01
2.83075213e-01 -3.72291595e-01 -1.15063739e+00 1.04760170e+00
-1.77501452e+00 -9.07505214e-01 -7.03060478e-02 1.61832452e+00
1.32363093e+00 -1.43130302e-01 7.22656548e-02 -5.49967825e-01
1.72524899e-01 8.46343562e-02 -5.69210231e-01 -3.69083673e-01
1.64193556e-01 5.01264572e-01 -1.28725052e-01 5.97055674e-01
-5.75033128e-01 1.15117478e+00 5.58805943e+00 9.46916699e-01
-6.29472852e-01 2.24888712e-01 -2.71154847e-02 6.01638407e-02
-5.38976789e-01 2.49418437e-01 -1.02987790e+00 1.95083231e-01
1.05329359e+00 -2.86608756e-01 4.76832747e-01 1.08176219e+00
-4.94319886e-01 -4.27778848e-02 -1.21805000e+00 5.32658458e-01
3.92823946e-03 -1.49174237e+00 6.06277764e-01 -4.91495728e-01
6.83126748e-01 2.99763471e-01 -8.30825195e-02 1.29371464e+00
8.53991210e-01 -9.63424087e-01 6.12600982e-01 3.46395940e-01
5.53102493e-01 -3.03139836e-01 6.08603477e-01 5.25190473e-01
-9.80570018e-01 -3.97634832e-03 -6.23438537e-01 -2.01219440e-01
9.83195677e-02 7.14326128e-02 -1.53495181e+00 8.56468678e-01
5.71810186e-01 1.79261819e-01 -6.38093114e-01 8.93751025e-01
-8.63169193e-01 1.05501473e+00 -5.84061965e-02 -4.94303226e-01
2.73663580e-01 2.97046211e-02 2.68764287e-01 1.15841949e+00
2.37105116e-01 3.94872069e-01 3.89093459e-02 1.02800608e+00
-4.21869665e-01 2.47234553e-01 -4.26014096e-01 -2.40571886e-01
5.85247695e-01 1.18886971e+00 1.74562305e-01 -7.19381452e-01
-7.55126297e-01 8.95163834e-01 9.11649883e-01 4.96752143e-01
-9.11695659e-01 -4.35956061e-01 4.80449975e-01 -1.19902641e-01
4.57926899e-01 -4.87832651e-02 4.30011034e-01 -1.38569748e+00
3.02844167e-01 -1.40497661e+00 8.39275599e-01 -1.22622061e+00
-1.53284156e+00 4.46312368e-01 1.21081382e-01 -7.25153208e-01
-3.64853173e-01 -6.59528792e-01 -5.44960022e-01 7.69088566e-01
-1.86536610e+00 -1.32076538e+00 -5.28105378e-01 8.11990798e-01
7.95859575e-01 -2.52643079e-02 9.61449385e-01 3.19835059e-02
-3.02727997e-01 5.22852898e-01 -2.60473609e-01 2.58321404e-01
8.02119017e-01 -1.55926001e+00 6.34041250e-01 5.98978817e-01
4.32153702e-01 7.97392905e-01 4.99522954e-01 -5.77112079e-01
-1.90189207e+00 -1.24074745e+00 5.67398727e-01 -1.12335515e+00
1.13099325e+00 -1.58829898e-01 -1.31524384e+00 1.09602571e+00
3.85386795e-01 -9.78703275e-02 8.38725150e-01 3.59709471e-01
-9.82291758e-01 -1.10514902e-01 -6.70843005e-01 7.94223785e-01
9.67501998e-01 -8.41431141e-01 -1.31580293e+00 5.66028178e-01
1.32794261e+00 -6.64549470e-01 -1.02819610e+00 2.51792878e-01
2.62849838e-01 -7.75144517e-01 8.60990465e-01 -1.10160983e+00
7.33562529e-01 -3.73356670e-01 -1.81083784e-01 -1.32687902e+00
-9.65772793e-02 -4.95634884e-01 -5.73676169e-01 1.07464516e+00
8.15456688e-01 -4.82274920e-01 6.36750996e-01 6.46729410e-01
-3.26121926e-01 -8.23608816e-01 -8.75266850e-01 -9.44682777e-01
2.87749201e-01 -5.22388577e-01 8.31968129e-01 7.92573750e-01
2.79739857e-01 8.61807108e-01 1.37628958e-01 3.01600933e-01
1.14194192e-01 2.62155712e-01 1.35472047e+00 -6.41058981e-01
-8.10367882e-01 -1.50376916e-01 -5.23288548e-02 -1.47509241e+00
3.36856335e-01 -1.08565676e+00 6.08000159e-02 -1.88318622e+00
2.01275021e-01 -3.42372358e-01 1.32594006e-02 6.47747040e-01
-6.06748819e-01 -1.85050964e-01 8.21232721e-02 -6.80648237e-02
-1.22006238e+00 6.07702613e-01 1.46786964e+00 -4.80912417e-01
2.93551292e-02 -1.97377428e-01 -8.38626325e-01 3.31986785e-01
6.67970598e-01 -2.19393015e-01 -8.07996809e-01 -7.71656811e-01
7.40311205e-01 3.70029330e-01 3.69095981e-01 -7.82904148e-01
3.53179157e-01 -3.04880083e-01 -4.54459101e-01 -2.87576765e-01
3.63263875e-01 -4.66528207e-01 -2.36854583e-01 2.21629232e-01
-3.58768046e-01 3.99616584e-02 4.34665322e-01 5.95469892e-01
-3.21244955e-01 -6.30123317e-01 9.60034877e-02 -4.72110629e-01
-1.04632580e+00 1.44704401e-01 -1.45916432e-01 8.45012426e-01
6.63016081e-01 2.70585567e-01 -1.15897644e+00 -6.01876557e-01
-2.05618382e-01 7.38850176e-01 2.58607596e-01 4.04327393e-01
5.18759608e-01 -9.12486434e-01 -7.97502875e-01 -4.19019222e-01
7.14930236e-01 4.44866568e-01 3.98372352e-01 4.72501725e-01
-7.05374777e-01 6.94059193e-01 1.06938474e-01 -4.18787718e-01
-1.00230551e+00 5.12346029e-01 3.29816312e-01 -6.23244762e-01
-3.14584941e-01 1.12039220e+00 1.92197472e-01 -8.80726278e-01
-4.74157929e-02 -6.11372769e-01 1.10389464e-01 -3.40918243e-01
5.72439373e-01 2.86027223e-01 2.71966755e-01 3.23231630e-02
-1.50717840e-01 1.74944445e-01 -3.76625448e-01 -8.37318227e-02
9.97421086e-01 2.20807523e-01 -2.09084392e-01 4.02305067e-01
8.75096321e-01 2.91010439e-02 -8.60338092e-01 -3.10062855e-01
9.28089321e-02 -1.18883111e-01 -5.99565625e-01 -1.20944715e+00
-4.06372696e-01 9.10938442e-01 -2.00645506e-01 1.32379532e-01
7.75203347e-01 4.37530339e-01 8.58069956e-01 1.38336158e+00
6.52684033e-01 -5.91438055e-01 5.86170554e-01 8.48323882e-01
1.20533466e+00 -1.18652368e+00 -1.38605520e-01 -2.83028185e-01
-6.66287243e-01 8.68454993e-01 1.23862660e+00 1.24980390e-01
-3.56103294e-02 -2.55039096e-01 2.16040209e-01 -4.43484694e-01
-1.24850678e+00 -2.79384375e-01 2.59399891e-01 5.63481271e-01
3.78472447e-01 9.34217311e-03 7.03336522e-02 8.16828787e-01
-5.06508827e-01 6.72472194e-02 6.47121787e-01 9.61051643e-01
-5.19743860e-01 -1.02808428e+00 -1.28727958e-01 3.88186485e-01
-2.58003324e-01 -3.79211366e-01 -4.23578560e-01 1.08692932e+00
-4.36997831e-01 1.05275333e+00 -2.59447992e-01 -6.24575689e-02
6.35811985e-01 5.90229571e-01 8.25312436e-01 -1.10328841e+00
-6.77600384e-01 -7.88275301e-01 5.79468191e-01 -5.05911648e-01
-1.11670494e-02 -1.04178771e-01 -1.32773077e+00 -4.88205366e-02
-4.57736820e-01 4.47001994e-01 1.80544004e-01 9.25097525e-01
5.41657925e-01 4.44611311e-01 -1.36128366e-01 1.91039741e-01
-9.53942716e-01 -1.08650446e+00 -2.07719043e-01 5.51981091e-01
2.06968561e-01 -3.32411945e-01 -1.78702369e-01 9.97358859e-02] | [10.795254707336426, 7.960903167724609] |
c1b5aa77-0330-4aba-9625-64f04882c2cb | prophnet-efficient-agent-centric-motion | 2303.12071 | null | https://arxiv.org/abs/2303.12071v3 | https://arxiv.org/pdf/2303.12071v3.pdf | ProphNet: Efficient Agent-Centric Motion Forecasting with Anchor-Informed Proposals | Motion forecasting is a key module in an autonomous driving system. Due to the heterogeneous nature of multi-sourced input, multimodality in agent behavior, and low latency required by onboard deployment, this task is notoriously challenging. To cope with these difficulties, this paper proposes a novel agent-centric model with anchor-informed proposals for efficient multimodal motion prediction. We design a modality-agnostic strategy to concisely encode the complex input in a unified manner. We generate diverse proposals, fused with anchors bearing goal-oriented scene context, to induce multimodal prediction that covers a wide range of future trajectories. Our network architecture is highly uniform and succinct, leading to an efficient model amenable for real-world driving deployment. Experiments reveal that our agent-centric network compares favorably with the state-of-the-art methods in prediction accuracy, while achieving scene-centric level inference latency. | ['Xiaodong Yang', 'Fang Da', 'Tong Su', 'Xishun Wang'] | 2023-03-21 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_ProphNet_Efficient_Agent-Centric_Motion_Forecasting_With_Anchor-Informed_Proposals_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_ProphNet_Efficient_Agent-Centric_Motion_Forecasting_With_Anchor-Informed_Proposals_CVPR_2023_paper.pdf | cvpr-2023-1 | ['motion-prediction', 'motion-forecasting'] | ['computer-vision', 'computer-vision'] | [-2.79003456e-02 -9.72717777e-02 -6.20918214e-01 -4.28435415e-01
-8.38575602e-01 -6.49809301e-01 9.86101091e-01 -6.28879517e-02
-3.40701073e-01 7.16894984e-01 4.68871564e-01 -2.69159645e-01
-8.12858343e-03 -6.95248127e-01 -7.09479690e-01 -4.05487567e-01
-2.39507362e-01 6.49056911e-01 4.02447581e-01 -5.94923079e-01
1.40073240e-01 4.17580187e-01 -1.92028606e+00 3.18941891e-01
6.77982271e-01 9.21549737e-01 5.55926800e-01 9.60974634e-01
8.65701512e-02 1.20011997e+00 -1.67317703e-01 -3.62411618e-01
1.67238370e-01 -2.02551372e-02 -6.55907094e-01 -2.84342449e-02
3.27659011e-01 -4.05098945e-01 -7.42287040e-01 6.13449633e-01
2.74483293e-01 2.76134402e-01 5.68176270e-01 -1.87270176e+00
-6.56678304e-02 3.63065422e-01 -1.53154731e-01 5.57121495e-03
5.04679143e-01 6.53748810e-01 9.50915158e-01 -8.88018548e-01
1.01791584e+00 1.11952221e+00 2.97405928e-01 7.16224849e-01
-9.28857386e-01 -3.36379409e-01 6.92388594e-01 6.63811624e-01
-1.19175029e+00 -6.19506896e-01 8.69897783e-01 -4.47048157e-01
1.06280911e+00 4.10416096e-01 6.91854835e-01 1.31667876e+00
5.21563292e-01 9.74026561e-01 5.49289644e-01 2.00681955e-01
4.99156415e-01 9.30427946e-03 -1.54840887e-01 8.47422063e-01
5.48326150e-02 3.07402294e-02 -8.92003536e-01 -1.21169522e-01
2.33837709e-01 2.72352714e-02 4.77442592e-02 -6.74125612e-01
-1.59591889e+00 6.67800009e-01 3.26964945e-01 -3.70717376e-01
-5.11274695e-01 5.69388449e-01 5.28708935e-01 4.34702672e-02
1.89424232e-01 2.43774518e-01 -2.21218154e-01 -5.55904269e-01
-7.95965433e-01 7.46271908e-01 7.73704469e-01 1.25969064e+00
8.97004306e-01 3.36994350e-01 -9.16367918e-02 4.43424404e-01
4.08545524e-01 6.65060759e-01 1.44786447e-01 -1.47344279e+00
7.01466262e-01 5.44731557e-01 2.15386659e-01 -1.03278899e+00
-7.22872019e-01 -1.82388797e-01 -7.70396173e-01 3.40545535e-01
1.63272306e-01 -3.58043849e-01 -6.99506700e-01 1.94328701e+00
4.98971909e-01 3.56982410e-01 3.65250498e-01 1.19047320e+00
8.09339464e-01 9.23534751e-01 3.28648359e-01 8.77547711e-02
1.28190696e+00 -1.36840618e+00 -3.76322120e-01 -6.04050338e-01
5.70924580e-01 -3.85902613e-01 8.82681966e-01 1.15686849e-01
-9.25605357e-01 -5.56486905e-01 -1.07575381e+00 -1.74615867e-02
-3.49011123e-01 -2.95969099e-02 8.57481897e-01 2.99830943e-01
-1.14157796e+00 1.44711733e-02 -8.59117091e-01 -3.67387593e-01
1.68386176e-01 3.27218831e-01 -4.82920110e-01 -1.41988054e-01
-8.43055367e-01 1.17866254e+00 4.13355529e-01 -4.83684381e-03
-1.12917483e+00 -5.74495912e-01 -1.09936321e+00 -1.93634868e-01
3.39521617e-01 -1.11578417e+00 1.51313686e+00 -5.14104426e-01
-1.60333264e+00 1.43434986e-01 -5.16231596e-01 -6.00978673e-01
5.11260986e-01 2.49605179e-01 -4.61424619e-01 7.47435093e-02
1.92404956e-01 1.30202389e+00 7.47233272e-01 -1.39187825e+00
-1.06485069e+00 7.13759146e-05 3.50080073e-01 6.62167311e-01
-1.68213829e-01 -3.78459275e-01 -5.68181098e-01 -1.83099046e-01
-1.81574538e-01 -1.14279068e+00 -7.80179203e-01 -1.39748588e-01
-1.61595136e-01 -2.27028415e-01 9.84420180e-01 -2.19161451e-01
1.03232658e+00 -1.91930616e+00 2.93902695e-01 2.48750914e-02
3.61737788e-01 -2.60058790e-01 -4.02793050e-01 8.50513458e-01
6.16047084e-01 -3.89395267e-01 -1.74580157e-01 -7.97034562e-01
5.57659686e-01 4.69042122e-01 -5.78849435e-01 3.30491275e-01
3.44690621e-01 9.95051324e-01 -1.17949235e+00 -5.91892838e-01
3.43607962e-01 2.36849949e-01 -5.57137787e-01 3.23104948e-01
-6.53817117e-01 5.24630189e-01 -5.85032403e-01 9.70428944e-01
4.62201983e-01 -9.67596695e-02 1.19052596e-01 -1.14799581e-01
-2.65305400e-01 1.52168468e-01 -9.76208985e-01 2.05251002e+00
-5.57303727e-01 9.25822377e-01 3.09597459e-02 -5.37413776e-01
7.05361307e-01 1.46174759e-01 4.18971449e-01 -6.45524800e-01
-2.17592329e-01 9.50661749e-02 -1.93985209e-01 -6.11248076e-01
1.25138032e+00 3.67931813e-01 -6.01833642e-01 2.67433733e-01
-1.94938868e-01 -1.55266941e-01 3.85297358e-01 3.16329271e-01
1.29104710e+00 4.80878443e-01 -1.68225896e-02 9.97917429e-02
2.21778765e-01 7.13642597e-01 5.83305418e-01 7.39567161e-01
-4.48435992e-01 4.02771294e-01 2.16542810e-01 -8.03702652e-01
-1.11351752e+00 -8.54780614e-01 4.24819708e-01 1.17334461e+00
7.00287879e-01 -5.84256351e-01 -4.67881978e-01 -5.29574931e-01
-8.52041095e-02 7.42030084e-01 -2.89977372e-01 1.00630391e-02
-8.19601834e-01 -5.48229277e-01 4.84217346e-01 6.18594646e-01
4.42449868e-01 -8.60196471e-01 -1.04181492e+00 4.24258143e-01
-3.97666693e-01 -1.40605509e+00 -3.18784088e-01 -2.25108042e-01
-4.84872609e-01 -7.62866139e-01 -3.09097677e-01 -5.41577697e-01
4.46253151e-01 5.05973339e-01 1.21132922e+00 -3.87868375e-01
1.63371369e-01 6.17080927e-01 -9.85099003e-02 -2.37267599e-01
-4.40712899e-01 2.15837091e-01 5.80995493e-02 -1.45283472e-02
1.87410042e-01 -3.64828110e-01 -7.66152680e-01 2.05788136e-01
-6.57550752e-01 4.33659256e-01 6.69544637e-01 6.02006257e-01
4.13976341e-01 -2.45763093e-01 6.48980439e-01 -4.08527464e-01
5.21904051e-01 -9.66247618e-01 -5.91885328e-01 -1.01266243e-02
-3.94895136e-01 -9.44830850e-02 7.23951399e-01 -3.74414951e-01
-1.05085421e+00 5.71336448e-01 -5.63980639e-02 -3.56983274e-01
-5.40932059e-01 5.51246643e-01 1.06568754e-01 -1.65370271e-01
4.20055062e-01 3.50214720e-01 1.74530655e-01 1.54523998e-01
7.40081608e-01 3.72514337e-01 7.56379247e-01 -4.38843518e-01
6.82382107e-01 6.74218357e-01 3.48215342e-01 -7.53722310e-01
-1.13046169e-01 -4.05529529e-01 -3.26808184e-01 -7.36197948e-01
7.77624309e-01 -1.28963327e+00 -1.07095969e+00 1.31771982e-01
-1.38463223e+00 -4.76045072e-01 -3.14572267e-02 5.94992161e-01
-9.13935483e-01 1.77872658e-01 -4.53173190e-01 -8.74676585e-01
-1.48211360e-01 -1.56329250e+00 1.24308443e+00 -2.27425620e-02
-3.74692589e-01 -8.88247430e-01 3.40443373e-01 3.42105538e-01
4.19975698e-01 2.88808316e-01 5.54739237e-01 -2.66257316e-01
-1.30767059e+00 -2.75946319e-01 -1.19529374e-01 -5.94213307e-01
-4.81867403e-01 -1.02334313e-01 -8.94742191e-01 -8.60164836e-02
-5.83181560e-01 -2.33626425e-01 8.15910816e-01 1.71379536e-01
7.68076003e-01 -2.95290917e-01 -5.99859595e-01 4.34734166e-01
1.28569329e+00 -3.57211679e-02 3.80151957e-01 5.22199094e-01
7.41754472e-01 5.94524205e-01 9.32670593e-01 5.79072356e-01
1.41292906e+00 8.19460392e-01 9.86130834e-01 1.03026211e-01
4.25865762e-02 -2.41803557e-01 5.50216854e-01 6.82907879e-01
-1.70614719e-02 -6.33293509e-01 -8.44505906e-01 7.45090485e-01
-2.49593711e+00 -1.31471252e+00 -7.83331171e-02 1.68874955e+00
1.64028659e-01 4.73512784e-02 5.02163827e-01 -3.70342642e-01
1.99998647e-01 5.48129082e-01 -5.62430918e-01 -4.20927137e-01
-5.18583842e-02 -8.05374980e-01 6.04876339e-01 5.32385290e-01
-1.05314076e+00 9.84260380e-01 6.76496029e+00 6.91356599e-01
-1.09908879e+00 1.40242964e-01 3.71757001e-01 -3.15638453e-01
-5.88348866e-01 -1.22279875e-01 -8.67440581e-01 4.59915727e-01
1.14779186e+00 -3.34492363e-02 3.32033813e-01 9.95635569e-01
4.35272634e-01 -5.75006939e-02 -1.15657449e+00 9.24091756e-01
5.53949401e-02 -1.98258770e+00 2.25555245e-02 -6.64478168e-02
6.74425244e-01 5.32895803e-01 1.27930179e-01 3.78374726e-01
4.62886482e-01 -8.23979855e-01 1.19921005e+00 6.13133967e-01
4.58005458e-01 -7.13922739e-01 5.41736186e-01 5.71162343e-01
-1.51906979e+00 -2.68145800e-01 -2.12617964e-01 -2.99869776e-01
6.18176818e-01 -8.40761960e-02 -9.81984735e-01 6.26628339e-01
3.49041075e-01 7.81971455e-01 -2.80899286e-01 1.01561499e+00
2.97826380e-01 3.90018851e-01 -3.11471283e-01 -4.00722682e-01
7.07880795e-01 1.01810716e-01 8.31136227e-01 1.38847995e+00
3.44592601e-01 -1.11607842e-01 5.59835076e-01 5.40200412e-01
1.61834389e-01 -2.01659903e-01 -1.15043139e+00 3.64186585e-01
6.91305757e-01 1.38614714e+00 -3.86431903e-01 -4.21777159e-01
-5.84533274e-01 8.36290777e-01 4.54081655e-01 5.26807308e-01
-9.96708930e-01 1.13283262e-01 1.03971696e+00 -2.69504577e-01
1.94718152e-01 -5.93083680e-01 -3.09957653e-01 -9.25219655e-01
9.11269486e-02 -5.96563518e-01 1.86099753e-01 -6.58343792e-01
-9.80183005e-01 8.98041427e-01 1.14550799e-01 -1.70054078e+00
-8.02847326e-01 -4.34836209e-01 -7.03526378e-01 5.35416186e-01
-1.74127805e+00 -1.57143390e+00 -4.49559093e-01 5.22566736e-01
8.99711430e-01 -4.87367809e-01 8.31564188e-01 2.67011821e-01
-5.67775667e-01 3.62396538e-01 -2.05325767e-01 -4.97635543e-01
3.49185526e-01 -9.82079685e-01 5.85127771e-01 9.42930818e-01
-1.87561929e-01 1.16723061e-01 9.16915894e-01 -5.04242957e-01
-2.27899957e+00 -1.47036541e+00 7.24918306e-01 -7.31646836e-01
7.03910649e-01 -2.53793657e-01 -3.59346300e-01 6.49156988e-01
4.37659800e-01 -7.63289407e-02 4.40145701e-01 -2.97836244e-01
-2.05874160e-01 -1.91767946e-01 -7.90482402e-01 1.18819225e+00
1.08055878e+00 -2.98371553e-01 -1.44152641e-01 3.68644953e-01
7.82006860e-01 -6.78661585e-01 -5.98286688e-01 4.24941838e-01
6.54955089e-01 -7.76009619e-01 9.69776988e-01 -5.25360763e-01
5.24019480e-01 -6.82247102e-01 -4.65000361e-01 -1.12006497e+00
-4.62044269e-01 -7.40412772e-01 -4.57515687e-01 6.90503001e-01
6.39141798e-01 -4.80070561e-01 9.92118180e-01 7.62424648e-01
-6.95539415e-01 -8.10575485e-01 -1.02221131e+00 -4.35187966e-01
-4.36097801e-01 -8.08962464e-01 8.21173608e-01 5.07912040e-01
2.45636940e-01 2.26438642e-01 -7.53644168e-01 5.55002570e-01
5.97944856e-01 2.84848660e-01 1.30850863e+00 -7.70541489e-01
3.48927081e-02 -5.34470379e-01 -7.15238154e-01 -1.31529760e+00
3.82080495e-01 -6.64350033e-01 3.59746188e-01 -1.67919540e+00
-5.85401691e-02 -4.61365402e-01 4.38343771e-02 3.09766471e-01
1.53638586e-01 3.89109045e-01 3.46886575e-01 1.46736160e-01
-1.14064467e+00 7.36027539e-01 1.02053452e+00 -3.36248487e-01
-1.84557125e-01 -1.36810347e-01 -3.41793239e-01 6.78322077e-01
8.59383762e-01 -6.68493956e-02 -7.03765452e-01 -6.56214178e-01
1.88730195e-01 4.05463308e-01 6.04664862e-01 -1.07225120e+00
7.86591589e-01 -7.05351114e-01 6.64959773e-02 -1.04941177e+00
1.19085503e+00 -9.13668334e-01 3.66039723e-01 2.98572510e-01
-2.74396122e-01 4.61005986e-01 2.17681676e-01 7.78314412e-01
-2.26356223e-01 2.28068918e-01 3.48552093e-02 2.16172053e-03
-1.57710433e+00 5.66419244e-01 -9.04203355e-01 -2.93088675e-01
1.15175676e+00 -3.13025564e-01 -7.32385695e-01 -7.31406093e-01
-2.83963621e-01 6.01337373e-01 5.41004479e-01 6.53917968e-01
7.65832067e-01 -1.46849859e+00 -7.00543463e-01 1.86667952e-03
5.20692348e-01 -1.03833325e-01 4.81315047e-01 7.38032043e-01
-4.95651901e-01 4.12973821e-01 -2.36247584e-01 -8.35611105e-01
-8.27171504e-01 4.03380781e-01 2.04562489e-02 5.61304390e-02
-6.30495667e-01 5.09054422e-01 7.97247961e-02 -5.25452971e-01
1.65331483e-01 -2.17830747e-01 -1.37794942e-01 -3.52095395e-01
4.80115771e-01 4.20369357e-01 -2.57831395e-01 -9.23010528e-01
-3.53387743e-01 3.26656193e-01 2.37758473e-01 -2.41675064e-01
9.54818547e-01 -4.82629597e-01 1.47533312e-01 4.37683523e-01
7.27527976e-01 -1.79875791e-01 -1.62665975e+00 6.39061108e-02
-2.13758662e-01 -2.91035831e-01 -7.55353644e-02 -5.36601365e-01
-6.83660328e-01 4.60669816e-01 2.73739368e-01 2.39523262e-01
7.09378421e-01 -8.09218362e-02 9.05728519e-01 6.72645807e-01
6.60025120e-01 -1.12096965e+00 -1.91128090e-01 5.91862857e-01
6.78708494e-01 -1.46861053e+00 -3.00959051e-01 -3.74516517e-01
-9.39128399e-01 9.48087931e-01 8.97635341e-01 1.43939927e-01
1.40816629e-01 2.30334297e-01 1.39242321e-01 -9.87259373e-02
-1.45874786e+00 -1.90580085e-01 2.28982493e-01 7.57776201e-01
-2.06785187e-01 2.31275395e-01 1.89963460e-01 2.89454162e-01
-1.04634121e-01 -3.18720102e-01 5.41457891e-01 9.93588567e-01
-5.84267855e-01 -7.44855046e-01 -2.50496238e-01 2.03033403e-01
3.17329615e-01 2.58101165e-01 -6.97915966e-04 6.55315757e-01
-5.47658168e-02 1.14419889e+00 1.32390037e-01 -7.01302707e-01
1.52703881e-01 -1.52184099e-01 1.52942866e-01 -1.27725512e-01
-1.59612447e-01 -2.90915400e-01 5.99966466e-01 -9.62964118e-01
-3.91667902e-01 -6.01430058e-01 -1.30695653e+00 -5.62992811e-01
2.67553508e-01 -2.06069216e-01 8.26920569e-01 1.00790143e+00
9.84810531e-01 4.17916059e-01 5.92758834e-01 -1.47963047e+00
-1.58708796e-01 -4.93458122e-01 8.11925530e-02 1.93731293e-01
5.97727180e-01 -8.03239465e-01 1.46057338e-01 9.63503495e-03] | [5.870617866516113, 0.7825969457626343] |
e7d17bfe-1299-487e-8593-0571ca161e03 | extracting-motion-and-appearance-via-inter | 2303.00440 | null | https://arxiv.org/abs/2303.00440v2 | https://arxiv.org/pdf/2303.00440v2.pdf | Extracting Motion and Appearance via Inter-Frame Attention for Efficient Video Frame Interpolation | Effectively extracting inter-frame motion and appearance information is important for video frame interpolation (VFI). Previous works either extract both types of information in a mixed way or elaborate separate modules for each type of information, which lead to representation ambiguity and low efficiency. In this paper, we propose a novel module to explicitly extract motion and appearance information via a unifying operation. Specifically, we rethink the information process in inter-frame attention and reuse its attention map for both appearance feature enhancement and motion information extraction. Furthermore, for efficient VFI, our proposed module could be seamlessly integrated into a hybrid CNN and Transformer architecture. This hybrid pipeline can alleviate the computational complexity of inter-frame attention as well as preserve detailed low-level structure information. Experimental results demonstrate that, for both fixed- and arbitrary-timestep interpolation, our method achieves state-of-the-art performance on various datasets. Meanwhile, our approach enjoys a lighter computation overhead over models with close performance. The source code and models are available at https://github.com/MCG-NJU/EMA-VFI. | ['LiMin Wang', 'Gangshan Wu', 'Youxin Chen', 'Haonan Wang', 'Yuhan Zhu', 'Guozhen Zhang'] | 2023-03-01 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhang_Extracting_Motion_and_Appearance_via_Inter-Frame_Attention_for_Efficient_Video_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhang_Extracting_Motion_and_Appearance_via_Inter-Frame_Attention_for_Efficient_Video_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-frame-interpolation'] | ['computer-vision'] | [-9.77332518e-03 -2.91804612e-01 -1.83958590e-01 -3.96428287e-01
-6.49791241e-01 -3.02648574e-01 3.87323081e-01 -2.07821161e-01
-3.08338135e-01 5.21436393e-01 2.30629638e-01 -2.02238604e-01
2.25334316e-01 -7.50327051e-01 -7.31005847e-01 -6.14454865e-01
2.95828849e-01 -3.35480541e-01 3.79153341e-01 -7.80606791e-02
8.30419809e-02 3.01059276e-01 -1.36587012e+00 3.46752703e-01
9.49053049e-01 1.11437631e+00 3.23180139e-01 4.44201559e-01
-1.34968266e-01 9.93545532e-01 -2.49762252e-01 -2.95353532e-01
1.89290464e-01 -3.65185767e-01 -7.96206474e-01 2.53422502e-02
5.23586810e-01 -7.07534373e-01 -6.92190766e-01 9.81659293e-01
4.50014323e-01 4.45709340e-02 2.06726193e-01 -1.13963723e+00
-6.48071527e-01 2.57616073e-01 -9.44889426e-01 2.56772667e-01
2.20980108e-01 3.92778218e-01 6.74232960e-01 -1.10542381e+00
4.02072370e-01 1.15577161e+00 6.64241850e-01 3.87786090e-01
-1.12454486e+00 -8.30839813e-01 3.81645828e-01 4.97256666e-01
-1.38436162e+00 -7.28771091e-01 9.95377302e-01 -3.82835358e-01
6.24337912e-01 1.54397979e-01 7.30161548e-01 7.78124273e-01
4.78520319e-02 1.13984692e+00 6.66292071e-01 -1.30836919e-01
-1.97894648e-01 -3.07610989e-01 1.19130528e-02 7.90784240e-01
7.42696449e-02 -2.64861509e-02 -4.03777868e-01 2.96440750e-01
1.12036872e+00 1.75959289e-01 -6.03601694e-01 -9.46030542e-02
-1.36144328e+00 4.89987284e-01 4.55220968e-01 2.85709679e-01
-2.69204378e-01 2.92505920e-01 3.77702802e-01 5.89483827e-02
5.35000741e-01 -2.09425867e-01 -3.86227846e-01 -4.71852794e-02
-9.58617628e-01 2.27200091e-01 2.07388118e-01 1.12115860e+00
9.73331928e-01 1.34923667e-01 -5.01715958e-01 6.82400882e-01
3.21816653e-01 2.65047252e-01 3.22969526e-01 -1.05735981e+00
4.49988246e-01 4.47929323e-01 3.35823327e-01 -1.04809797e+00
-2.86569476e-01 -3.65004420e-01 -1.12575090e+00 1.37890622e-01
4.19975787e-01 1.86635125e-02 -9.08786952e-01 1.72011614e+00
5.10728002e-01 6.03424788e-01 -3.88295799e-01 1.14488590e+00
1.18568635e+00 6.20799780e-01 1.76523581e-01 -7.34921098e-02
1.35464430e+00 -1.46834683e+00 -8.04596484e-01 -1.58748940e-01
3.30891907e-01 -8.83438587e-01 1.06650639e+00 1.17621999e-02
-1.45905161e+00 -9.83129203e-01 -9.60739374e-01 -6.04836702e-01
1.71081498e-02 4.46529418e-01 5.90014815e-01 1.58880562e-01
-1.08303869e+00 5.63063145e-01 -9.22668636e-01 1.46455333e-01
6.47484958e-01 4.53593642e-01 -2.42731288e-01 -2.18869388e-01
-1.09386528e+00 4.00143743e-01 2.18291685e-01 4.13044214e-01
-5.78229547e-01 -8.22277427e-01 -1.01689565e+00 6.88274428e-02
1.71929181e-01 -9.58980739e-01 1.22100687e+00 -1.04024827e+00
-1.46389341e+00 4.65149790e-01 -6.35200083e-01 -8.74807760e-02
5.00512362e-01 -3.12940061e-01 -2.50033110e-01 3.96039188e-02
5.04220696e-03 8.35302413e-01 9.52964604e-01 -1.14807248e+00
-7.37311542e-01 -1.95664793e-01 1.85035810e-01 9.04887095e-02
-3.63126487e-01 4.80180755e-02 -9.74788070e-01 -9.80011165e-01
-4.50101867e-03 -6.59668684e-01 -6.85479715e-02 3.11596781e-01
-3.51693593e-02 -1.09056190e-01 9.38480258e-01 -9.17021394e-01
1.50872183e+00 -2.23368716e+00 1.06098741e-01 -2.86313683e-01
4.65531558e-01 5.53245485e-01 -1.58696666e-01 3.74676809e-02
-2.35523712e-02 -3.16165835e-02 -2.81889468e-01 -6.19820654e-01
-1.93757206e-01 9.09731537e-02 -3.60178798e-02 4.87677157e-01
4.58467662e-01 1.23779309e+00 -1.00484157e+00 -5.87365150e-01
5.37830710e-01 8.80721152e-01 -6.94982052e-01 2.01226339e-01
1.58901766e-01 8.45577300e-01 -3.89423192e-01 6.93355978e-01
1.05366206e+00 -4.63841081e-01 -9.31254402e-02 -7.33065605e-01
-1.88567609e-01 1.88614458e-01 -1.06813836e+00 2.00048614e+00
-6.20386243e-01 7.15152383e-01 2.53093839e-01 -8.50105524e-01
6.28310740e-01 2.22680464e-01 5.62053502e-01 -5.82021475e-01
1.80984691e-01 5.49836233e-02 -9.40757096e-02 -4.28424180e-01
5.38896203e-01 2.78492600e-01 3.71905625e-01 5.09444550e-02
2.79359315e-02 5.97215295e-01 5.64116836e-02 -7.10630938e-02
7.18445837e-01 5.84839046e-01 2.52757609e-01 -1.62394837e-01
8.74452889e-01 -4.19931442e-01 9.72111106e-01 3.59652340e-01
-3.32332402e-01 8.87086987e-01 -2.01625694e-02 -6.61330760e-01
-9.99853909e-01 -8.14631522e-01 -2.06338577e-02 8.59901428e-01
6.05183780e-01 -5.57021260e-01 -5.55355728e-01 -4.52874839e-01
-2.87803441e-01 1.75979733e-01 -6.21474147e-01 5.38119450e-02
-8.78189504e-01 -6.07416987e-01 1.67813540e-01 7.83473730e-01
8.02779973e-01 -9.57241118e-01 -4.83637691e-01 3.59799445e-01
-6.27259076e-01 -1.13366222e+00 -8.84810090e-01 -4.41308737e-01
-7.71220386e-01 -8.28550637e-01 -9.22028005e-01 -9.14938629e-01
7.27338314e-01 7.20807791e-01 1.09188950e+00 4.35287803e-01
-9.77631733e-02 -7.97768608e-02 -3.98462325e-01 1.75078809e-02
6.97733015e-02 8.96557122e-02 -5.09519100e-01 3.26224148e-01
1.84025496e-01 -7.22077370e-01 -1.27587247e+00 3.37322712e-01
-8.45764875e-01 5.95012844e-01 4.69322771e-01 9.35073614e-01
5.00182629e-01 -2.84510285e-01 3.15971375e-01 -4.13422525e-01
1.37836248e-01 -4.62625951e-01 -5.28590202e-01 1.23632371e-01
-2.82921672e-01 -1.27312556e-01 6.22946382e-01 -3.60538095e-01
-1.28017461e+00 2.18677342e-01 -3.26975703e-01 -7.18809426e-01
-1.66806623e-01 1.84062719e-01 -3.84757608e-01 -2.20057771e-01
-5.48070893e-02 4.13204193e-01 -1.89290002e-01 -6.32955909e-01
2.06687391e-01 4.97544616e-01 6.28546476e-01 -5.29560804e-01
7.80059755e-01 7.02160478e-01 -2.38498956e-01 -5.95480680e-01
-6.99506819e-01 -4.29547280e-01 -6.61333025e-01 -2.55355328e-01
8.68922055e-01 -1.21278894e+00 -7.03455865e-01 7.74585247e-01
-1.46284962e+00 -3.68547320e-01 4.80978517e-03 3.86035442e-01
-2.87181318e-01 6.78884268e-01 -8.70001793e-01 -5.01666903e-01
-5.89232981e-01 -1.49817455e+00 1.34186375e+00 3.76738071e-01
1.23608120e-01 -8.03300798e-01 -2.05956742e-01 3.68127823e-01
5.19308805e-01 9.56091285e-02 4.90035027e-01 7.82796890e-02
-9.46548879e-01 1.15503393e-01 -6.63203180e-01 1.43729737e-02
2.63793617e-01 7.71083981e-02 -1.01232839e+00 -4.76918459e-01
-1.38034165e-01 3.49353701e-02 1.00801420e+00 4.56401289e-01
1.51599932e+00 -3.54838848e-01 -1.57258809e-01 1.17653286e+00
1.42192006e+00 1.96549356e-01 9.50056672e-01 4.07198757e-01
1.10711813e+00 3.30475688e-01 6.62585199e-01 5.64698398e-01
6.57011867e-01 9.35752451e-01 3.63561809e-01 -4.55885142e-01
-3.11938643e-01 -9.54165608e-02 2.98262745e-01 9.65542495e-01
-2.91064739e-01 -5.29404283e-02 -5.72164297e-01 6.54160857e-01
-2.04719520e+00 -1.03912640e+00 -1.56919628e-01 2.06323099e+00
8.48136127e-01 -2.87192553e-01 7.95350671e-02 1.03863336e-01
7.63566673e-01 3.68427664e-01 -4.36655015e-01 8.33266005e-02
-6.97893556e-04 9.75945368e-02 4.40211356e-01 6.52706623e-01
-1.28000069e+00 8.35659266e-01 5.00471640e+00 9.99760926e-01
-1.16503227e+00 3.00912410e-01 9.23028111e-01 -6.44885600e-02
-3.50132853e-01 -6.25573844e-03 -7.48869061e-01 8.46458316e-01
5.09006321e-01 -1.35770617e-02 3.96850646e-01 4.53264832e-01
3.84010971e-01 2.22284555e-01 -8.02601933e-01 1.14041507e+00
-1.82922333e-01 -1.59489429e+00 -1.85070504e-02 -1.11818060e-01
6.94968164e-01 -4.70730029e-02 -1.30759969e-01 9.91708040e-02
-2.00779393e-01 -8.26928139e-01 9.07930851e-01 4.88531619e-01
8.72336686e-01 -7.35247314e-01 7.05508292e-01 4.02886756e-02
-1.84457290e+00 1.00311711e-01 -2.39409417e-01 -7.58999884e-02
2.64973879e-01 4.86856818e-01 1.42908832e-02 9.39644933e-01
8.04932833e-01 1.10227489e+00 -5.69375157e-01 1.21186852e+00
-2.58526430e-02 4.47074503e-01 -1.39224589e-01 4.93220896e-01
1.57476932e-01 -2.96497732e-01 3.12484950e-01 1.23532796e+00
3.78419697e-01 1.63536072e-01 1.88476980e-01 8.11681271e-01
6.47904947e-02 5.84304705e-02 -2.88141906e-01 4.55809772e-01
5.62526524e-01 1.32376397e+00 -5.87851763e-01 -5.17790139e-01
-8.60451102e-01 1.12358952e+00 3.41588855e-01 4.82992768e-01
-1.20234728e+00 -4.21649724e-01 9.06964779e-01 1.83482692e-01
5.82553685e-01 -2.46640325e-01 -1.73127830e-01 -1.56340206e+00
3.15973103e-01 -6.05242610e-01 1.00755662e-01 -7.49904454e-01
-9.07526255e-01 6.50462508e-01 -1.90815866e-01 -1.68279111e+00
3.15468796e-02 -3.33693922e-01 -7.50130117e-01 8.82120490e-01
-1.84477818e+00 -1.43805182e+00 -8.00434768e-01 7.13150084e-01
7.62283385e-01 2.45205626e-01 3.37808341e-01 8.42144012e-01
-8.46551061e-01 8.13359380e-01 6.16184436e-02 4.33751017e-01
7.53132343e-01 -7.93126702e-01 6.23307228e-01 1.07117462e+00
-1.67361543e-01 5.32535672e-01 1.85529694e-01 -4.45935607e-01
-1.40043879e+00 -1.51498222e+00 6.38082147e-01 -5.73093593e-02
3.42858106e-01 -8.87237042e-02 -1.10640717e+00 6.16096914e-01
3.41429740e-01 3.53331983e-01 3.32208127e-01 -4.06136423e-01
-1.47624806e-01 -1.92665532e-01 -8.35248888e-01 6.89789593e-01
1.18058825e+00 -3.24044704e-01 -1.62158325e-01 -9.71219018e-02
8.91559839e-01 -6.05601847e-01 -9.98735845e-01 5.30346632e-01
6.57666445e-01 -1.10162449e+00 1.26251543e+00 -1.94633722e-01
7.19801247e-01 -8.33875597e-01 5.70560880e-02 -7.81273425e-01
-6.85645342e-01 -7.41808951e-01 -5.25398016e-01 1.36366320e+00
-2.78309938e-02 -5.08555830e-01 5.14081061e-01 4.76933569e-01
-2.21074089e-01 -1.04898715e+00 -6.35584772e-01 -4.58975196e-01
-2.08667353e-01 -3.05223376e-01 8.08621526e-01 9.91180003e-01
-3.91920179e-01 2.25043654e-01 -7.45863080e-01 1.51060358e-01
7.00673044e-01 3.74630094e-01 8.13258529e-01 -8.20583820e-01
-3.31061423e-01 -5.31401336e-01 -3.80509049e-01 -1.51369345e+00
7.81882703e-02 -7.20323145e-01 -6.76551163e-02 -1.57210767e+00
2.32394248e-01 -2.94924289e-01 -4.14318353e-01 5.42653561e-01
-5.35523772e-01 5.44379830e-01 3.23838800e-01 2.98682958e-01
-5.54484129e-01 7.83026397e-01 1.55152857e+00 -1.75358504e-01
-1.60031974e-01 -3.10552776e-01 -6.38110220e-01 8.34133208e-01
7.82583296e-01 -1.12844676e-01 -4.59816366e-01 -8.60755265e-01
-2.61776179e-01 1.70347154e-01 6.57719433e-01 -1.01214099e+00
1.49831623e-01 -1.47988886e-01 6.33678079e-01 -6.92371190e-01
3.21616858e-01 -6.79602087e-01 2.44826064e-01 2.69972354e-01
-3.98599729e-02 1.35663152e-01 3.13862115e-01 5.34132898e-01
-4.97925818e-01 5.37939891e-02 8.41729701e-01 2.35363226e-02
-9.27658856e-01 7.13436246e-01 -1.24959037e-01 -5.30344546e-02
8.13702703e-01 -3.38120043e-01 -2.04176456e-01 -4.22648638e-01
-4.94280100e-01 2.44707838e-01 6.98488712e-01 4.45472717e-01
6.86728835e-01 -1.55888557e+00 -7.11722076e-01 3.29339474e-01
-6.99991584e-02 3.14859748e-01 5.95847130e-01 1.16822338e+00
-4.70986843e-01 1.42798856e-01 -3.05290878e-01 -6.17016196e-01
-1.14933348e+00 5.44642389e-01 2.13569745e-01 -9.75407958e-02
-7.95506418e-01 7.60235369e-01 7.77137101e-01 1.34494394e-01
4.65954579e-02 -4.11064684e-01 -3.55505645e-02 -1.75805241e-01
7.80967236e-01 2.44284987e-01 -1.93194509e-01 -8.69564772e-01
-2.85772413e-01 8.43213141e-01 -1.56557381e-01 3.58048618e-01
1.22834802e+00 -4.61576909e-01 -6.78797662e-02 -6.06762245e-03
1.39610612e+00 -8.24136809e-02 -1.62604403e+00 -2.08498701e-01
-5.09659767e-01 -9.44441080e-01 1.52316779e-01 -3.09570044e-01
-1.53420496e+00 9.85977411e-01 4.61360544e-01 -3.13567519e-01
1.55773556e+00 -2.87644237e-01 1.20142627e+00 -2.00984925e-01
2.16198862e-01 -5.73270679e-01 -2.33332127e-01 2.44575337e-01
7.93990672e-01 -1.32079542e+00 3.51923145e-02 -6.14950001e-01
-3.58619750e-01 9.59215581e-01 8.53797793e-01 -2.17053816e-02
5.87905467e-01 2.57074177e-01 -2.16663450e-01 2.00771004e-01
-6.32696748e-01 -1.91783458e-01 4.74959671e-01 4.04434115e-01
8.20545316e-01 -2.53486961e-01 -4.07322824e-01 5.55493057e-01
3.59263718e-01 2.66000628e-01 2.22552791e-01 8.76581848e-01
-1.06905617e-01 -1.11509836e+00 -1.24323681e-01 1.84472933e-01
-4.60635245e-01 -3.03650647e-01 2.02612460e-01 6.36880517e-01
2.36224294e-01 8.11868787e-01 7.15787113e-02 -4.51008767e-01
7.08585382e-02 -2.66684771e-01 4.28891629e-01 -1.05628129e-02
-4.49025780e-01 3.43657255e-01 -3.91118862e-02 -8.22779417e-01
-6.84231162e-01 -3.37436914e-01 -1.10677946e+00 -4.12693024e-01
-2.23353341e-01 -1.86359704e-01 1.19222611e-01 7.33553946e-01
6.77439928e-01 7.09119201e-01 5.99788189e-01 -1.36059141e+00
6.18074834e-02 -7.90679812e-01 -6.42715544e-02 4.49689746e-01
4.66658682e-01 -6.83780611e-01 2.03034957e-03 3.75540555e-01] | [10.787790298461914, -1.425649881362915] |
40f4a435-7a9c-4d8d-9efc-aa693bdacfae | teamuncc-lt-edi-eacl2021-hope-speech | null | null | https://aclanthology.org/2021.ltedi-1.20 | https://aclanthology.org/2021.ltedi-1.20.pdf | TeamUNCC@LT-EDI-EACL2021: Hope Speech Detection using Transfer Learning with Transformers | In this paper, we describe our approach towards utilizing pre-trained models for the task of hope speech detection. We participated in Task 2: Hope Speech Detection for Equality, Diversity and Inclusion at LT-EDI-2021 @ EACL2021. The goal of this task is to predict the presence of hope speech, along with the presence of samples that do not belong to the same language in the dataset. We describe our approach to fine-tuning RoBERTa for Hope Speech detection in English and our approach to fine-tuning XLM-RoBERTa for Hope Speech detection in Tamil and Malayalam, two low resource Indic languages. We demonstrate the performance of our approach on classifying text into hope-speech, non-hope and not-language. Our approach ranked 1st in English (F1 = 0.93), 1st in Tamil (F1 = 0.61) and 3rd in Malayalam (F1 = 0.83). | ['Samira Shaikh', 'Erfan Al-Hossami', 'Khyati Mahajan'] | 2021-04-19 | null | null | null | null | ['hope-speech-detection-for-tamil', 'hope-speech-detection-for-malayalam', 'hope-speech-detection', 'hope-speech-detection-for-english'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-8.81831795e-02 1.08727597e-01 -2.49689117e-01 -1.23702332e-01
-1.63036108e+00 -5.70475221e-01 1.00456643e+00 2.39050135e-01
-4.19818282e-01 6.36857629e-01 9.65620279e-01 -7.60228992e-01
-5.83728217e-02 -5.02360404e-01 -2.32513085e-01 -2.86233902e-01
2.17728689e-01 5.24427414e-01 -3.63713689e-02 -4.81245786e-01
3.60065758e-01 4.26673383e-01 -9.72825885e-01 6.63404465e-01
6.45854175e-01 3.38075787e-01 8.63714293e-02 1.43160582e+00
3.11237387e-02 1.02765501e+00 -8.63762498e-01 -4.42694992e-01
-8.72782394e-02 -6.03798091e-01 -1.20674944e+00 -1.69596687e-01
6.15570307e-01 -1.43657669e-01 -5.43457031e-01 9.13639069e-01
7.57493973e-01 6.38526538e-03 8.52897584e-01 -7.73638844e-01
-4.75080162e-01 1.23041594e+00 -3.93536091e-01 4.81618553e-01
4.40717459e-01 1.68153197e-01 1.09669304e+00 -1.10200477e+00
6.45998240e-01 1.51461565e+00 7.47136652e-01 5.77453017e-01
-9.44446683e-01 -4.71164256e-01 -5.80059111e-01 -1.15380026e-01
-1.55040693e+00 -1.34900331e+00 4.64118809e-01 -3.84172261e-01
1.11842847e+00 7.46922016e-01 4.06713635e-01 1.27010894e+00
-3.64988923e-01 1.29675543e+00 1.01607811e+00 -5.49655795e-01
4.42935564e-02 3.85373294e-01 -6.81457520e-02 7.01207936e-01
-5.48067808e-01 -2.38282904e-01 -8.15803230e-01 -2.90645510e-01
5.73587641e-02 -7.05062151e-01 1.34416465e-02 6.66027129e-01
-1.10463989e+00 1.21773160e+00 -1.06532268e-01 5.54364443e-01
-1.17021896e-01 -5.29205561e-01 5.54278791e-01 3.03249329e-01
5.95220208e-01 4.96207148e-01 -2.49104232e-01 -6.72801971e-01
-1.42502475e+00 1.21100038e-01 7.19958127e-01 8.25176835e-01
-1.89857796e-01 6.14587925e-02 -5.08839667e-01 1.43472135e+00
2.89812475e-01 5.77898324e-01 7.62390971e-01 -5.09141624e-01
7.04062939e-01 2.87739754e-01 -1.69456974e-01 -4.63926882e-01
-3.97355974e-01 -3.51293594e-01 -5.70889831e-01 -2.43258268e-01
5.21289229e-01 -1.34301364e-01 -7.47581959e-01 1.57634997e+00
-1.14276677e-01 -4.83499050e-01 7.19050884e-01 4.37241167e-01
1.10501647e+00 1.02415490e+00 -1.88671514e-01 -4.32756126e-01
1.26322150e+00 -1.20081878e+00 -4.47935164e-01 -1.80158108e-01
9.51741576e-01 -1.43050885e+00 1.27581608e+00 2.02639073e-01
-1.25229216e+00 -1.75887138e-01 -7.39524961e-01 -1.84253901e-01
-3.61605436e-01 5.32877088e-01 9.65450518e-03 1.10252500e+00
-1.12182605e+00 8.52000639e-02 -2.33262926e-01 -8.45825016e-01
2.08920121e-01 -1.53484687e-01 -3.55776995e-01 1.70341417e-01
-1.00468314e+00 6.80713773e-01 2.32263818e-01 -4.24688488e-01
-1.03938627e+00 -4.11437362e-01 -6.64902091e-01 -7.69085288e-02
-6.05524927e-02 5.28948717e-02 1.21739602e+00 -4.38344985e-01
-1.02713883e+00 1.28786170e+00 -3.04336369e-01 -5.61306298e-01
6.88355386e-01 8.06977749e-02 -7.93011665e-01 -1.19046234e-01
2.73338497e-01 4.95903581e-01 6.20850623e-01 -7.61898994e-01
-8.60669792e-01 -2.17141956e-01 -2.32047766e-01 1.72315449e-01
-5.09207487e-01 5.39546907e-01 1.96163990e-02 -6.37325585e-01
-1.34347260e-01 -6.64185643e-01 3.68067175e-01 -6.26735449e-01
-8.89100254e-01 -4.89776254e-01 8.19879293e-01 -1.16900134e+00
1.55612934e+00 -2.25195289e+00 -4.14817393e-01 9.89262089e-02
1.87318981e-01 3.25206041e-01 -1.93368614e-01 7.65489280e-01
1.08929522e-01 5.02845943e-01 5.82890399e-02 -5.03283441e-01
1.39431790e-01 -3.53200622e-02 -4.74822998e-01 4.20733273e-01
1.91968754e-01 5.51722527e-01 -9.22318399e-01 -5.74832141e-01
-9.73370597e-02 4.65251744e-01 -2.41643026e-01 7.20837563e-02
3.34068924e-01 -1.26171950e-03 7.63029680e-02 8.57706785e-01
4.19693410e-01 2.70081937e-01 -6.17269017e-02 1.87772572e-01
-5.76247633e-01 9.87005949e-01 -8.08234870e-01 7.57838726e-01
-5.18667340e-01 1.01098883e+00 2.67406911e-01 -5.58106005e-01
1.01983488e+00 4.41503018e-01 6.66896766e-03 -6.28328204e-01
-1.52811250e-02 5.41251242e-01 2.41137773e-01 -4.98256445e-01
9.17114437e-01 -7.87703134e-03 -1.43474653e-01 1.34178013e-01
-3.78217958e-02 -2.47300535e-01 4.25498992e-01 4.04236227e-01
1.07175434e+00 -8.44538569e-01 4.90480483e-01 -5.25133133e-01
8.14979732e-01 -2.21616700e-01 1.25983849e-01 1.21601939e+00
-5.88002980e-01 6.73929930e-01 5.31391025e-01 5.25324829e-02
-1.19394028e+00 -9.83854413e-01 -3.77004355e-01 1.32334340e+00
-6.44504130e-01 -7.77665198e-01 -5.24586558e-01 -7.41156876e-01
-3.86376619e-01 1.18523169e+00 -3.25997174e-01 -2.27941826e-01
-4.65592504e-01 -5.17683268e-01 1.19514525e+00 4.60173748e-03
3.90926123e-01 -1.01552367e+00 -1.60295576e-01 -1.42641857e-01
-4.65823144e-01 -1.00007057e+00 -7.65501380e-01 1.71995714e-01
-3.18787873e-01 -5.34421861e-01 -9.61213708e-01 -7.41884291e-01
3.81656289e-02 -1.28202692e-01 8.76559138e-01 -5.39337285e-02
-2.68980563e-01 2.12677523e-01 -2.96909511e-01 -1.53736323e-01
-1.18201923e+00 3.57106835e-01 -7.76044205e-02 -2.30312943e-01
1.50865406e-01 1.07570782e-01 6.57071695e-02 4.94941417e-03
-3.90101641e-01 3.43841799e-02 3.08625907e-01 9.29095984e-01
1.01864427e-01 6.31993040e-02 4.49779093e-01 -7.07989872e-01
1.02129710e+00 -4.04763341e-01 -2.29970291e-01 2.32797980e-01
-2.64536798e-01 -2.42932782e-01 4.57570791e-01 -3.29871565e-01
-7.64411628e-01 2.03550294e-01 -9.11992371e-01 7.57200867e-02
-1.70886397e-01 3.87352884e-01 -2.26046428e-01 2.07553133e-01
7.51206815e-01 6.01516545e-01 -3.95930558e-01 -4.85760093e-01
3.38965863e-01 1.42052078e+00 6.03800595e-01 -8.49465877e-02
4.78048235e-01 -7.67671093e-02 -5.26396036e-01 -1.70366561e+00
-6.69832468e-01 -8.07828844e-01 -3.60018194e-01 -2.30019689e-01
6.94981158e-01 -9.13473785e-01 -3.72980475e-01 5.71462810e-01
-1.06312346e+00 -2.64297545e-01 -2.27440864e-01 3.03003311e-01
-3.60770464e-01 3.89390260e-01 -7.06143916e-01 -1.53315282e+00
-6.35144830e-01 -1.16528893e+00 1.28415298e+00 -4.64150235e-02
-7.16216207e-01 -9.19310212e-01 1.01009063e-01 8.56726706e-01
3.22421044e-01 -3.95373553e-01 9.79969263e-01 -1.22294772e+00
3.77682716e-01 -1.35264486e-01 -1.27401561e-01 9.03824195e-02
1.11018997e-02 4.62420583e-02 -1.12709069e+00 -2.17699870e-01
-1.69497043e-01 -5.50634623e-01 8.90248835e-01 1.57748833e-01
5.20017028e-01 -6.60644054e-01 7.29357777e-03 1.08324727e-02
7.00052381e-01 3.17132771e-02 4.98884499e-01 2.02487409e-01
3.18636209e-01 6.82598472e-01 4.08440977e-01 5.33431411e-01
2.81050414e-01 7.02841163e-01 6.01376733e-03 4.54555601e-02
-3.54858935e-01 -7.04509556e-01 9.04923558e-01 9.05925870e-01
5.48668265e-01 -5.56880474e-01 -1.27486396e+00 1.05308771e+00
-1.34443176e+00 -1.13764882e+00 -5.93883455e-01 2.18092251e+00
9.53043699e-01 2.81965584e-01 9.84080732e-01 2.46924207e-01
5.57749510e-01 2.32635692e-01 1.45897776e-01 -8.16511214e-01
-3.66747111e-01 -4.51258458e-02 2.56166935e-01 1.14104211e+00
-1.40763712e+00 1.15588343e+00 6.39698505e+00 1.21676278e+00
-9.98365104e-01 1.32624835e-01 8.40255022e-01 -9.95819047e-02
-1.72405154e-01 -1.29666656e-01 -1.14827931e+00 4.39141899e-01
1.51196074e+00 -2.91415066e-01 2.78085768e-01 6.84672773e-01
2.89929688e-01 2.30394229e-02 -6.83275461e-01 9.51768577e-01
2.91226774e-01 -1.08075237e+00 -7.20964968e-02 -1.22004561e-02
5.51132441e-01 2.20005751e-01 1.96316317e-01 5.76071799e-01
2.64261156e-01 -1.16376102e+00 8.61136556e-01 1.57912776e-01
6.05518222e-01 -9.45032001e-01 7.18225598e-01 8.85285020e-01
-8.45028698e-01 -8.56959149e-02 -2.65457034e-01 1.49459630e-01
-1.32594514e-03 4.71321523e-01 -1.44285917e+00 -2.90248133e-02
4.18436021e-01 3.44067127e-01 -7.59706557e-01 7.52065837e-01
-8.53660628e-02 1.01304877e+00 -3.10449183e-01 -4.68420833e-01
4.16307300e-01 5.04866481e-01 9.04424012e-01 1.74318600e+00
3.55444103e-01 -4.03434128e-01 3.56216878e-01 5.05656898e-01
-2.97150671e-01 6.66343212e-01 -5.24652302e-01 -5.96260667e-01
4.63618994e-01 1.07957518e+00 -4.65884268e-01 -5.15583932e-01
-1.31312206e-01 9.18128252e-01 1.25462919e-01 -2.77943492e-01
-3.35283905e-01 -4.42227006e-01 3.18670750e-01 2.70531267e-01
-9.43966210e-02 5.01581207e-02 -1.67678788e-01 -8.80103648e-01
-2.89262921e-01 -1.06913936e+00 6.32121384e-01 -6.89303935e-01
-1.35857332e+00 7.33612359e-01 -3.04346561e-01 -6.12723589e-01
-5.25256872e-01 -3.72935385e-01 -5.88491619e-01 1.13049841e+00
-9.13103282e-01 -1.01189423e+00 4.02979940e-01 4.85844254e-01
1.13106060e+00 -8.26041222e-01 8.48924518e-01 3.28564793e-01
-3.70318919e-01 1.04274797e+00 3.43279779e-01 3.91125560e-01
5.70131540e-01 -1.29374397e+00 5.60988784e-01 9.69174922e-01
4.05355781e-01 3.78783941e-01 8.07463169e-01 -6.16326034e-01
-1.05049336e+00 -8.56679022e-01 1.92875469e+00 -4.22851861e-01
1.05766153e+00 -7.23266900e-01 -5.92607558e-01 2.29281157e-01
2.58665681e-01 -6.18338585e-01 5.49866796e-01 1.86734930e-01
-3.42495471e-01 3.53040844e-01 -1.35520470e+00 7.32060552e-01
7.91550756e-01 -1.13843608e+00 -4.59338456e-01 8.74441147e-01
4.84206229e-01 -2.98254490e-01 -7.24236548e-01 -1.77121297e-01
4.21693861e-01 -8.41894150e-01 8.03041697e-01 -6.52681589e-01
6.03000939e-01 2.46650502e-01 -4.86820936e-01 -1.21344531e+00
-1.19937837e-01 -8.13767791e-01 1.89712733e-01 1.58604825e+00
8.25931251e-01 -4.06226128e-01 6.72831953e-01 -2.35008355e-02
-8.44137520e-02 -3.80756855e-01 -1.29045284e+00 -8.33460927e-01
3.77867818e-01 -3.55300277e-01 -3.71502861e-02 9.90171134e-01
2.53743738e-01 8.94940853e-01 -6.23544872e-01 7.29424953e-02
2.05435276e-01 -3.62284660e-01 6.47356749e-01 -7.90940046e-01
-3.94810587e-01 -8.08404207e-01 -3.53795946e-01 -8.46565187e-01
5.20555601e-02 -1.15806711e+00 1.21759154e-01 -1.16091657e+00
4.80616957e-01 -1.18246280e-01 1.82271868e-01 7.20160067e-01
-1.05108768e-01 2.57340014e-01 3.28529537e-01 2.82273382e-01
-2.85658807e-01 3.46431524e-01 4.85541373e-01 -3.33656132e-01
-1.82488903e-01 3.71415794e-01 -7.60999978e-01 4.12805974e-01
9.98746932e-01 -5.06510913e-01 -2.64701545e-01 1.48740844e-04
-1.04125768e-01 1.34099886e-01 1.23270981e-01 -8.71118128e-01
3.44498344e-02 3.26659158e-02 9.29426104e-02 -8.78719628e-01
4.43521321e-01 -2.01987594e-01 -2.37736955e-01 6.08432233e-01
-7.75812209e-01 2.25076959e-01 2.10445955e-01 -2.25587755e-01
-7.42525980e-02 -6.57842755e-01 7.86361098e-01 -2.18255550e-01
-2.88290143e-01 -3.15532446e-01 -8.97468269e-01 5.67944765e-01
4.82411295e-01 2.31493145e-01 -5.89581132e-01 -8.00751328e-01
-6.31885946e-01 1.68373939e-02 8.84485468e-02 6.88798308e-01
6.45035923e-01 -9.33738530e-01 -1.31754386e+00 3.67111504e-01
3.22990447e-01 -1.14228189e+00 -3.13756615e-02 8.60937297e-01
-3.37568969e-01 4.90862101e-01 2.30268642e-01 -3.74490172e-01
-2.05377603e+00 7.78420269e-02 2.63650566e-01 -3.13984275e-01
-2.55731642e-01 1.08818185e+00 -1.74599811e-01 -9.16564941e-01
3.40322435e-01 4.59274173e-01 -2.36253858e-01 2.67985016e-02
8.09416354e-01 4.31868464e-01 7.45910555e-02 -1.24024761e+00
-4.80937779e-01 -1.33833051e-01 -2.81805009e-01 -5.35570800e-01
1.09200788e+00 3.23158763e-02 -1.30247340e-01 6.72534287e-01
1.43165600e+00 7.87926614e-01 -2.32021466e-01 2.49947205e-01
1.71238840e-01 -3.40773314e-01 3.53216171e-01 -1.11100781e+00
-4.31772649e-01 9.94718552e-01 8.44872475e-01 4.66167212e-01
6.72886848e-01 2.75771946e-01 6.49762630e-01 2.80583203e-01
-3.25492293e-01 -1.21727240e+00 -1.42824456e-01 9.78142381e-01
9.75084782e-01 -1.22761393e+00 -2.55724609e-01 -9.55747291e-02
-7.02069223e-01 1.21083057e+00 4.99268733e-02 6.23873591e-01
1.82265610e-01 1.38140127e-01 1.16692506e-01 2.04223935e-02
-5.99254191e-01 -1.82840556e-01 2.89080113e-01 5.23487866e-01
9.15085435e-01 3.63878787e-01 -3.97814721e-01 1.74296126e-01
-8.76819849e-01 -5.39750576e-01 5.71670532e-01 5.51002681e-01
-6.96178555e-01 -7.64666677e-01 -3.87591720e-01 5.69581926e-01
-7.22703338e-01 -5.56589186e-01 -1.12054455e+00 7.28580117e-01
-2.66419768e-01 1.20315611e+00 -9.48350057e-02 -5.58443367e-01
2.95297831e-01 2.83177286e-01 -1.24234073e-01 -4.14428741e-01
-7.39873290e-01 4.51568156e-01 8.83588791e-01 1.86320201e-01
2.51727581e-01 -9.35623407e-01 -8.81776512e-01 -5.86560667e-01
-1.47339061e-01 3.27629775e-01 5.24055541e-01 6.94201052e-01
1.95809659e-02 2.42162988e-01 5.75307012e-01 -9.73458588e-02
-7.09344149e-01 -1.04842401e+00 -5.23101926e-01 1.21815726e-01
4.49490875e-01 6.19486123e-02 -3.85783643e-01 -2.48768836e-01] | [9.367219924926758, 10.718904495239258] |
e17f119c-97ec-462b-af96-e40618f8d093 | deep-region-and-multi-label-learning-for | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Zhao_Deep_Region_and_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Zhao_Deep_Region_and_CVPR_2016_paper.pdf | Deep Region and Multi-Label Learning for Facial Action Unit Detection | Region learning (RL) and multi-label learning (ML) have recently attracted increasing attentions in the field of facial Action Unit (AU) detection. Knowing that AUs are active on sparse facial regions, RL aims to identify these regions for a better specificity. On the other hand, a strong statistical evidence of AU correlations suggests that ML is a natural way to model the detection task. In this paper, we propose Deep Region and Multi-label Learning (DRML), a unified deep network that simultaneously addresses these two problems. One crucial aspect in DRML is a novel region layer that uses feed-forward functions to induce important facial regions, forcing the learned weights to capture structural information of the face. Our region layer serves as an alternative design between locally connected layers (i.e., confined kernels to individual pixels) and conventional convolution layers (i.e., shared kernels across an entire image). Unlike previous studies that solve RL and ML alternately, DRML by construction addresses both problems, allowing the two seemingly irrelevant problems to interact more directly. The complete network is end-to-end trainable, and automatically learns representations robust to variations inherent within a local region. Experiments on BP4D and DISFA benchmarks show that DRML performs the highest average F1-score and AUC within and across datasets in comparison with alternative methods. | ['Wen-Sheng Chu', 'Kaili Zhao', 'Honggang Zhang'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [ 1.88127175e-01 2.01583520e-01 -5.06931365e-01 -3.86946917e-01
-6.76453948e-01 -3.63360256e-01 5.90989411e-01 -3.27358872e-01
-3.28554362e-01 4.58029509e-01 1.76651463e-01 2.03433231e-01
1.79559484e-01 -6.59469724e-01 -8.21126521e-01 -8.73347223e-01
3.90672982e-02 8.53503309e-03 1.59496903e-01 -4.99341711e-02
-4.31254059e-02 7.57190764e-01 -1.48600221e+00 5.57889938e-01
5.79040051e-01 1.22799325e+00 -1.52250409e-01 2.02189878e-01
1.12806626e-01 8.92808974e-01 -3.17176014e-01 -2.71117210e-01
4.09269810e-01 -4.89633203e-01 -6.57017171e-01 9.26204994e-02
7.58141458e-01 -3.75709116e-01 -1.57237574e-01 8.82168949e-01
6.08815312e-01 5.83513267e-02 8.74126494e-01 -1.08961666e+00
-6.29560053e-01 2.69058734e-01 -1.25730443e+00 2.44494155e-03
9.95010287e-02 1.12868980e-01 1.06402659e+00 -1.21710634e+00
5.20454526e-01 1.46680331e+00 4.90689665e-01 8.57956767e-01
-1.36111379e+00 -8.38517666e-01 3.17202330e-01 -1.99168310e-01
-1.66761804e+00 -4.08283949e-01 6.84505939e-01 -4.70511973e-01
5.28302789e-01 2.26200093e-02 2.59258896e-01 1.31064475e+00
2.30879888e-01 1.04037130e+00 1.17929554e+00 -2.69035697e-01
7.13345222e-03 2.36643746e-01 -1.77067608e-01 1.12639880e+00
-3.48982513e-02 9.65112373e-02 -4.53936815e-01 -1.72534749e-01
1.19313240e+00 1.04346991e-01 -1.27535522e-01 -4.68333304e-01
-7.90437877e-01 9.82224226e-01 5.83155334e-01 2.89573103e-01
-2.16222495e-01 3.30255747e-01 3.78216296e-01 -2.46466994e-02
7.85815418e-01 2.49125689e-01 -4.00443286e-01 4.91225868e-01
-9.72516060e-01 1.22546434e-01 2.07411736e-01 6.31956339e-01
1.00308371e+00 -1.34877577e-01 -6.77157700e-01 1.11094379e+00
3.50175261e-01 9.21029672e-02 4.25914466e-01 -6.59856915e-01
1.41811103e-01 7.77558804e-01 -2.05778554e-01 -8.85820925e-01
-5.82677960e-01 -5.01449525e-01 -8.54610682e-01 5.59792519e-01
5.04293323e-01 -1.96781307e-01 -9.24568653e-01 2.05923796e+00
3.69711339e-01 4.88588452e-01 -2.57171154e-01 8.52440000e-01
9.59375262e-01 5.66260993e-01 1.21981487e-01 4.60014492e-02
1.38175631e+00 -1.18787539e+00 -4.19104129e-01 -3.67379397e-01
6.54416144e-01 -3.89922708e-01 9.22019660e-01 2.02332497e-01
-9.55999494e-01 -6.46363974e-01 -7.07189322e-01 -2.01369449e-02
-4.48591113e-01 5.20851791e-01 8.05518389e-01 3.51369560e-01
-1.21030211e+00 2.91795254e-01 -3.56808901e-01 -2.81579256e-01
9.22035098e-01 5.44105828e-01 -6.21352196e-01 -8.73659104e-02
-1.03023374e+00 7.43295610e-01 6.98501989e-02 1.30137876e-01
-1.14874220e+00 -6.89485908e-01 -9.15808260e-01 2.34019626e-02
4.46856111e-01 -3.21024448e-01 9.48188305e-01 -1.37249911e+00
-1.48938727e+00 1.13900483e+00 -1.19712301e-01 -2.16979817e-01
4.26940650e-01 -1.23928919e-01 -1.87606722e-01 2.55591214e-01
6.31071255e-02 1.16455483e+00 1.08260727e+00 -1.18837225e+00
-5.15514135e-01 -3.39874983e-01 1.54833132e-02 -1.97141226e-02
-3.23135704e-01 4.04189557e-01 -4.55355823e-01 -6.77598059e-01
-1.22707911e-01 -7.64242411e-01 -7.93755725e-02 4.21779573e-01
-1.35345250e-01 -7.06270754e-01 7.58933365e-01 -3.56491774e-01
1.07488310e+00 -2.25076866e+00 1.65793315e-01 2.14905471e-01
4.48081166e-01 2.31626272e-01 -5.15209615e-01 -5.62840048e-03
-4.29689616e-01 2.20222119e-02 -1.57966688e-01 -2.92143494e-01
-2.50475943e-01 -2.00619921e-02 -3.67977321e-02 8.65928650e-01
6.86682642e-01 1.19058549e+00 -7.63195395e-01 -5.87462008e-01
-4.02239710e-02 5.54589033e-01 -5.53117931e-01 2.41721883e-01
-2.03714699e-01 5.63322663e-01 -3.71853590e-01 9.65000570e-01
6.49409056e-01 -3.32800508e-01 -2.33497232e-01 -1.46132529e-01
1.89049244e-02 -2.38212496e-01 -9.41954911e-01 1.66752100e+00
-5.20670235e-01 4.12778109e-01 1.52816460e-01 -9.43474293e-01
9.84190106e-01 4.30931866e-01 7.38257468e-01 -7.58017182e-01
2.55181789e-01 2.78814719e-03 -1.96114004e-01 -3.19981784e-01
-8.11532699e-03 -1.62039623e-01 1.27356648e-02 5.21432936e-01
4.34986591e-01 6.23288333e-01 -1.40498236e-01 -6.19928688e-02
9.82268631e-01 2.75345176e-01 4.39431816e-01 -2.21451223e-01
5.84135115e-01 -6.44185603e-01 7.84723938e-01 4.43845540e-01
-4.25877839e-01 7.51905143e-01 6.68176532e-01 -5.17872810e-01
-5.61517358e-01 -1.01132500e+00 -1.89744726e-01 1.48038888e+00
1.28959771e-02 -2.21589923e-01 -6.75668120e-01 -1.11159039e+00
1.96660981e-01 8.60387832e-02 -1.21297562e+00 -2.60437131e-01
-4.70552713e-01 -6.51462793e-01 6.83037698e-01 5.31228304e-01
4.78280306e-01 -1.11990619e+00 -3.58389646e-01 1.87808275e-02
7.04324618e-03 -1.04509497e+00 -5.57800233e-01 1.80289134e-01
-4.98333037e-01 -1.01511931e+00 -1.01824856e+00 -7.75692582e-01
7.79270649e-01 2.21187413e-01 9.94144857e-01 -1.22213416e-01
-5.06097376e-01 2.17499539e-01 -2.16876417e-01 -2.47417942e-01
-4.41203341e-02 -3.66125628e-02 -6.38644248e-02 7.32586443e-01
4.01130646e-01 -4.72719520e-01 -6.61903143e-01 4.28299189e-01
-8.67932558e-01 -1.10323235e-01 1.01861286e+00 8.49063516e-01
6.34299934e-01 -4.00840521e-01 7.60953009e-01 -1.00133634e+00
3.83084357e-01 -6.22172713e-01 -4.75225091e-01 4.37254578e-01
-2.03385651e-01 -9.34603214e-02 3.30014616e-01 -5.31082392e-01
-9.32217419e-01 4.19644058e-01 -6.00268468e-02 -7.01332510e-01
-3.89808536e-01 -5.54777645e-02 -2.70615250e-01 -4.53346252e-01
6.55524313e-01 6.99350014e-02 2.59514093e-01 -3.98608983e-01
5.24532020e-01 4.97338474e-01 2.47142136e-01 -3.94540191e-01
5.35949349e-01 6.85467958e-01 -3.25554647e-02 -6.64806962e-01
-1.12973678e+00 -6.38280153e-01 -8.30321670e-01 -4.84428108e-01
1.05224407e+00 -1.16576290e+00 -6.19269729e-01 3.46682817e-01
-1.07351184e+00 -3.47760141e-01 -9.86862779e-02 1.86140120e-01
-4.88463521e-01 1.54493272e-01 -5.71688235e-01 -7.14198887e-01
-1.96888715e-01 -1.09516406e+00 1.48241198e+00 3.62414122e-01
-1.84275076e-01 -9.29345429e-01 -8.08546320e-02 1.85085490e-01
1.58352897e-01 6.82144165e-01 4.96732056e-01 -3.80765647e-01
-3.22754681e-01 -7.80769289e-02 -6.55716121e-01 4.46406603e-01
2.51146585e-01 -2.44722962e-02 -1.42258739e+00 -3.78062218e-01
-1.96775690e-01 -7.21723080e-01 1.18186080e+00 3.70131552e-01
1.39644468e+00 -3.21898669e-01 -4.65300381e-01 6.03535175e-01
1.35058618e+00 -1.32529154e-01 5.79650700e-01 8.53664428e-03
7.32083619e-01 6.93669081e-01 4.59368646e-01 3.94196272e-01
7.94058070e-02 8.39956760e-01 5.76767862e-01 -6.74367607e-01
-2.93207258e-01 -1.04356132e-01 6.11015677e-01 -1.02324806e-01
2.66624168e-02 1.02162473e-01 -5.27353168e-01 3.32068980e-01
-1.77896106e+00 -9.40917611e-01 2.17495963e-01 1.97759068e+00
8.94491196e-01 -9.65659842e-02 5.34754038e-01 -4.15318310e-01
5.98702371e-01 3.50150853e-01 -5.38695931e-01 -4.32337075e-01
-2.45653316e-01 3.96916211e-01 5.20611048e-01 2.09838301e-01
-1.37368429e+00 1.14746058e+00 6.16516018e+00 1.03852856e+00
-1.24945056e+00 2.70874023e-01 1.02895331e+00 -2.42642000e-01
8.30315799e-03 -4.26279634e-01 -1.16412199e+00 3.18297327e-01
6.85595036e-01 4.85771418e-01 2.89433867e-01 8.66002381e-01
2.19355881e-01 -4.58797626e-02 -1.11510742e+00 1.03963184e+00
2.61953086e-01 -1.13865089e+00 5.46069853e-02 2.62924820e-01
8.80973518e-01 7.54259303e-02 1.67395040e-01 2.61775076e-01
3.11820596e-01 -1.34335828e+00 6.10126019e-01 6.58053339e-01
1.08007741e+00 -7.69412756e-01 5.58156312e-01 2.01595902e-01
-1.22343111e+00 -2.24490434e-01 -4.56386149e-01 1.52731106e-01
-2.34639436e-01 4.52887744e-01 -7.04999089e-01 1.98049352e-01
5.14200091e-01 8.50945413e-01 -7.51270771e-01 8.50494266e-01
-3.40080887e-01 5.24507642e-01 -1.72668591e-01 2.85025924e-01
6.23973906e-01 -1.86738715e-01 7.49185979e-02 1.34442925e+00
-1.14352405e-01 -7.52770677e-02 1.73945367e-01 1.09192777e+00
-2.10446790e-01 2.34465465e-01 -8.06356788e-01 2.37363487e-01
-7.10561872e-02 1.58655870e+00 -6.95574760e-01 2.52833422e-02
-7.94666350e-01 1.13836348e+00 8.45451713e-01 3.92469138e-01
-9.71006036e-01 -3.34312283e-02 8.98317337e-01 6.90372959e-02
3.04137319e-01 5.51046431e-02 -1.12734653e-01 -9.01759744e-01
-1.31316215e-01 -6.35018349e-01 4.45049644e-01 -2.81611949e-01
-1.29790139e+00 4.45367336e-01 -3.18172812e-01 -1.20570576e+00
-1.43421339e-02 -7.43668795e-01 -5.93593717e-01 9.29921389e-01
-1.86996710e+00 -1.46847653e+00 -2.36759067e-01 6.90494835e-01
4.64678913e-01 -7.29826093e-02 7.77562380e-01 3.20755899e-01
-8.28648031e-01 1.00964463e+00 -7.48239690e-03 4.20426607e-01
8.69860470e-01 -9.98244524e-01 -3.79197709e-02 7.34829843e-01
3.46359968e-01 4.36817914e-01 -2.19392329e-02 -3.51146758e-01
-8.90425742e-01 -1.42339814e+00 5.72960258e-01 -3.94027382e-01
3.61678123e-01 -7.39459693e-01 -8.53263795e-01 5.65631986e-01
5.29547334e-02 5.94221652e-01 8.43618095e-01 -1.75820086e-02
-7.52639711e-01 -2.62422889e-01 -1.13178205e+00 4.36796010e-01
9.84240651e-01 -5.44941962e-01 -4.46788594e-02 4.20967340e-01
3.94692570e-01 -1.15801819e-01 -8.22369933e-01 4.07977670e-01
6.02504134e-01 -1.04453325e+00 9.36652541e-01 -5.12921691e-01
4.50960696e-01 -2.54630685e-01 2.86711659e-02 -1.05130708e+00
-5.39312541e-01 -3.71150196e-01 -1.94455922e-01 1.24956059e+00
4.85457927e-01 -6.02993667e-01 8.26646447e-01 1.36268809e-01
4.41793278e-02 -1.33490705e+00 -8.85136664e-01 -6.41544580e-01
1.48200393e-01 -1.60835534e-02 4.04671967e-01 9.40062702e-01
-2.24438980e-01 3.23333710e-01 -4.12926704e-01 9.96984690e-02
5.22176087e-01 9.87284780e-02 5.50117373e-01 -1.26933491e+00
-2.20793396e-01 -7.69376338e-01 -2.60490090e-01 -1.04513252e+00
5.20622671e-01 -1.01228988e+00 -3.22032929e-03 -1.29672611e+00
2.41696447e-01 -5.03974259e-01 -4.69400823e-01 9.04977858e-01
-2.15268344e-01 5.96199453e-01 -1.11037612e-01 4.13677208e-02
-5.49512923e-01 6.07720315e-01 1.40245843e+00 -1.42894536e-01
-1.10714465e-01 9.99464691e-02 -9.02654707e-01 6.74852014e-01
7.21154749e-01 -2.38287151e-01 -2.52368420e-01 -2.19810396e-01
-8.70145187e-02 -3.27443689e-01 4.48719233e-01 -7.87083447e-01
1.75496552e-03 -5.47412969e-02 6.33310735e-01 -3.13923717e-01
3.53149116e-01 -7.34786153e-01 -4.68063086e-01 1.18088834e-01
-3.53725463e-01 -4.21712577e-01 2.43537173e-01 5.09577453e-01
-2.64081389e-01 1.49275079e-01 1.26313722e+00 -2.91931719e-01
-8.96651864e-01 6.60511374e-01 -2.07353354e-01 -1.90293670e-01
1.41579521e+00 -3.53657812e-01 -1.66661814e-02 -1.61126256e-01
-5.80036998e-01 2.08475918e-01 1.96390986e-01 5.82578301e-01
6.46486700e-01 -1.38904142e+00 -8.57947409e-01 3.92105877e-01
4.01025146e-01 3.38096060e-02 3.05004157e-02 8.46779048e-01
-1.75824672e-01 3.21105719e-01 -2.30128706e-01 -6.07160926e-01
-1.28442991e+00 4.96936172e-01 6.50265813e-01 -1.90322593e-01
-4.47914541e-01 1.34952092e+00 7.84778714e-01 -4.30748105e-01
3.62673700e-01 -9.72036049e-02 -4.65023786e-01 4.07172918e-01
7.05026567e-01 2.71382891e-02 -1.02383144e-01 -9.11499798e-01
-3.47000748e-01 6.63983047e-01 -2.37058312e-01 3.29054207e-01
1.33437872e+00 1.31001592e-01 -2.30148405e-01 2.84082979e-01
1.48979962e+00 -1.53267696e-01 -1.60251904e+00 -4.02779818e-01
-2.44379491e-02 -2.81549931e-01 1.14925571e-01 -7.14691520e-01
-1.42517138e+00 8.24055851e-01 6.39256477e-01 -2.91696906e-01
1.13406098e+00 3.66999686e-01 4.51725125e-01 -2.11801037e-01
2.27479681e-01 -8.39103281e-01 4.36645836e-01 2.04239801e-01
8.90385807e-01 -1.38034880e+00 -1.24275684e-01 -4.91031587e-01
-5.42960107e-01 1.06361568e+00 1.02816117e+00 -1.27686918e-01
5.60049117e-01 7.49318525e-02 2.09818128e-03 -3.44862372e-01
-5.32358825e-01 -4.25981849e-01 4.79972541e-01 3.81302059e-01
5.36256671e-01 -1.31924137e-01 -1.10712200e-01 3.64701629e-01
4.82821852e-01 -1.38948679e-01 -4.49510897e-03 6.94725394e-01
-4.39555764e-01 -1.03262782e+00 -3.23618114e-01 5.34047246e-01
-4.93238062e-01 1.19374409e-01 -6.61472142e-01 7.82670259e-01
5.26836038e-01 6.36653662e-01 1.23909399e-01 -3.27422470e-01
1.16953306e-01 2.72151586e-02 4.17388767e-01 -7.75034666e-01
-6.38574064e-01 1.44788668e-01 -4.31403965e-01 -9.07462180e-01
-3.61719728e-01 -6.63148403e-01 -1.16029263e+00 2.04119742e-01
-4.22103375e-01 -1.13828115e-01 3.29379082e-01 8.31511676e-01
3.54189485e-01 4.31234568e-01 9.04698551e-01 -7.44080067e-01
-2.46232226e-01 -1.06854987e+00 -8.02403748e-01 2.82985210e-01
3.78711849e-01 -9.31261182e-01 -1.63190067e-01 -2.93224871e-01] | [13.599496841430664, 1.526954174041748] |
08d96658-db7d-4e2c-9f6a-d857bcf6ae42 | hybrid-precoder-and-combiner-designs-for | 2306.14301 | null | https://arxiv.org/abs/2306.14301v1 | https://arxiv.org/pdf/2306.14301v1.pdf | Hybrid Precoder and Combiner Designs for Decentralized Parameter Estimation in mmWave MIMO Wireless Sensor Networks | Hybrid precoder and combiner designs are conceived for decentralized parameter estimation in millimeter wave (mmWave) multiple-input multiple-output (MIMO) wireless sensor networks (WSNs). More explicitly, efficient pre- and post-processing of the sensor observations and received signal are proposed for the minimum mean square error (MMSE) estimation of a parameter vector. The proposed techniques exploit the limited scattering nature of the mmWave MIMO channel for formulating the hybrid transceiver design framework as a multiple measurement vectors (MMV)-based sparse signal recovery problem. This is then solved using the iterative appealingly low-complexity simultaneous orthogonal matching pursuit (SOMP). Tailor-made designs are presented for WSNs operating under both total and per-sensor power constraints, while considering ideal noiseless as well as realistic noisy sensors. Furthermore, both the Bayesian Cramer-Rao lower bound and the centralized MMSE bound are derived for benchmarking the proposed decentralized estimation schemes. Our simulation results demonstrate the efficiency of the designs advocated and verify the analytical findings. | ['Lajos Hanzo', 'Aditya K. Jagannatham', 'Naveen K. D. Venkategowda', 'Kunwar Pritiraj Rajput', 'Suraj Srivastava', 'Priyanka Maity'] | 2023-06-25 | null | null | null | null | ['benchmarking', 'benchmarking'] | ['miscellaneous', 'robots'] | [ 5.91650605e-01 4.93172944e-01 1.68240860e-01 -2.72102118e-01
-8.60378861e-01 -3.27940613e-01 2.73057997e-01 -1.68434739e-01
-4.72783923e-01 5.78525484e-01 1.71431735e-01 -3.64319772e-01
-6.85485184e-01 -5.17998993e-01 -4.25117612e-01 -1.36481714e+00
-4.53968793e-01 -1.19457453e-01 -6.14443123e-01 1.32134512e-01
-1.90967008e-01 3.96016508e-01 -7.64780760e-01 -6.47898078e-01
3.95138294e-01 1.27582705e+00 3.88587505e-01 9.42337930e-01
5.51434815e-01 2.10987329e-01 -5.78713357e-01 -1.86723724e-01
5.13780475e-01 -1.44191220e-01 5.85465014e-01 5.76256327e-02
-1.66197643e-01 -3.19790572e-01 -7.74059594e-01 1.08766031e+00
8.34460199e-01 -7.35757411e-01 6.15625203e-01 -1.17438817e+00
2.06409797e-01 7.65488923e-01 -6.46362305e-01 -3.00364703e-01
1.53687134e-01 -2.41673350e-01 5.40011585e-01 -7.74652302e-01
2.56325364e-01 8.22464645e-01 7.85650253e-01 -8.67638215e-02
-7.90582478e-01 -9.07321513e-01 -4.91048992e-01 -2.26180851e-01
-1.48800540e+00 -9.03115690e-01 6.92273080e-01 5.00357375e-02
4.43459541e-01 4.64318156e-01 5.10556042e-01 7.71671474e-01
8.08753490e-01 3.70015800e-01 8.56890678e-01 -6.03489935e-01
5.62286139e-01 -1.57303795e-01 -1.97683632e-01 6.47426367e-01
1.28897285e+00 4.23808962e-01 -5.80677569e-01 -4.22814190e-01
6.24921441e-01 4.36200649e-02 -2.00383380e-01 -2.58389592e-01
-1.33623278e+00 4.34975654e-01 2.52146095e-01 4.29735959e-01
-1.09615636e+00 8.14975023e-01 -5.31995773e-01 6.42389834e-01
-4.12077457e-02 1.09927826e-01 -2.41223469e-01 5.18060803e-01
-1.22524321e+00 -2.27210477e-01 1.20883954e+00 1.58896494e+00
6.83690488e-01 3.23783040e-01 -2.26392940e-01 2.10282221e-01
1.29912508e+00 1.68109131e+00 -2.08033323e-01 -7.25468278e-01
7.53968894e-01 -1.88387811e-01 2.74739981e-01 -1.05846477e+00
-9.45605278e-01 -1.56436968e+00 -1.41723692e+00 -3.13402086e-01
1.32454887e-01 -9.73242104e-01 -7.89486885e-01 1.52273929e+00
2.79957235e-01 2.37247840e-01 7.23789096e-01 4.57287878e-01
6.54320657e-01 5.22041202e-01 -1.65857479e-01 -8.65342975e-01
1.17939806e+00 6.27496019e-02 -9.09490466e-01 -6.27405882e-01
3.64561439e-01 -7.30164826e-01 -5.75262368e-01 4.10378575e-01
-1.12766743e+00 3.83279562e-01 -1.60976386e+00 7.93602228e-01
4.98557627e-01 2.82205939e-01 4.82087851e-01 1.29543841e+00
-8.12867105e-01 -1.28119230e-01 -9.65679824e-01 -7.78895468e-02
2.76661307e-01 5.87044835e-01 -8.24214667e-02 -3.53051692e-01
-6.04619443e-01 3.31509233e-01 -1.22575328e-01 4.80310857e-01
-6.70365930e-01 -7.34564304e-01 -6.85032964e-01 4.08331193e-02
-7.94832706e-02 -1.10868800e+00 1.05576599e+00 -1.33337537e-02
-1.39890337e+00 3.21308412e-02 -3.27586591e-01 -6.01438642e-01
1.12914369e-01 1.38496071e-01 -3.32035333e-01 3.50104868e-01
-2.55896717e-01 -4.24056262e-01 7.79573739e-01 -1.37345195e+00
-4.48004454e-01 -6.49974704e-01 -5.40286899e-01 -1.12758286e-01
-1.47781566e-01 -3.84792298e-01 4.07431014e-02 -6.68137610e-01
1.34570432e+00 -7.73874640e-01 -9.92404342e-01 -3.15362453e-01
-6.61068678e-01 7.33998418e-01 2.64166176e-01 -4.59577173e-01
8.92495275e-01 -2.12759304e+00 5.18753305e-02 1.03114545e+00
2.42939726e-01 -3.36463183e-01 -2.48181939e-01 1.07317436e+00
4.33501303e-01 -6.31361783e-01 -1.00435838e-01 -5.04355729e-01
-2.15472002e-02 3.76328677e-02 1.00527093e-01 1.28911400e+00
-6.58378243e-01 6.25797272e-01 -6.00335598e-01 -3.03852987e-02
1.14785917e-01 3.50493133e-01 -1.83352590e-01 1.29060879e-01
4.27973807e-01 4.56131756e-01 -9.61842835e-01 8.37221503e-01
1.07983017e+00 6.74634725e-02 4.95799154e-01 -6.38723493e-01
-1.42088160e-01 -2.58974552e-01 -1.57264507e+00 1.63873065e+00
-7.73063421e-01 2.52238482e-01 1.10781074e+00 -8.43600452e-01
9.19158041e-01 6.43251419e-01 8.75084639e-01 -7.90327847e-01
4.55541223e-01 5.32958388e-01 -1.87625401e-02 -5.64886212e-01
-8.79622027e-02 -3.80034506e-01 -2.38154948e-01 4.66565758e-01
1.55952036e-01 -2.47446641e-01 -2.39807233e-01 2.59582072e-01
1.78340304e+00 -4.75477576e-01 8.35253060e-01 -5.25085211e-01
2.31247514e-01 -4.67446655e-01 5.92103839e-01 1.08188534e+00
3.00713271e-01 -1.06818229e-01 -4.00312901e-01 3.96435678e-01
-5.49595654e-01 -9.69309926e-01 -3.54963206e-02 1.75768167e-01
4.41119909e-01 -1.68746144e-01 -2.12800577e-01 3.06064993e-01
8.89026821e-02 6.01173460e-01 -1.03218168e-01 2.22270489e-01
-2.00731337e-01 -1.12020087e+00 4.73015726e-01 -1.02994524e-01
4.00525779e-01 1.14877068e-01 -9.12689090e-01 6.50037110e-01
2.29956552e-01 -1.24015963e+00 4.22509521e-01 7.31385827e-01
-6.94876850e-01 -8.14900875e-01 -5.61970890e-01 -5.21376491e-01
9.21803176e-01 6.19908750e-01 4.88633007e-01 -2.50820011e-01
-7.91172162e-02 1.10992980e+00 -4.00197655e-01 -7.50780761e-01
-1.13110974e-01 -3.67812306e-01 2.94402242e-01 1.47307679e-01
6.40307087e-04 -1.14386058e+00 -7.56839335e-01 1.95790976e-01
-6.56178296e-01 -9.28299427e-02 1.54352188e+00 3.75282317e-01
1.73923701e-01 8.09614174e-03 5.70965111e-01 -1.99432284e-01
3.14299971e-01 -7.32457101e-01 -7.83722639e-01 3.66728336e-01
-4.13481772e-01 3.01540583e-01 2.40127906e-01 1.89679507e-02
-7.66355038e-01 4.42649364e-01 -5.41905165e-02 4.50230926e-01
2.05162600e-01 6.98709548e-01 -5.08275509e-01 -9.12097454e-01
3.94366801e-01 3.93010765e-01 -3.12371224e-01 -1.66746050e-01
3.92187446e-01 9.09104586e-01 5.40974438e-01 -1.96011707e-01
1.77812016e+00 6.85066998e-01 8.00948441e-01 -1.48803520e+00
-3.58135343e-01 -6.86419070e-01 -2.44461507e-01 -2.30630383e-01
9.13814306e-02 -1.36572278e+00 -6.91957653e-01 1.07167989e-01
-1.08652103e+00 9.52356160e-02 1.90105110e-01 1.14602494e+00
-3.73629808e-01 3.34362388e-01 -8.25681910e-02 -1.52693629e+00
-5.60663879e-01 -8.01214516e-01 8.49285007e-01 6.05055951e-02
1.29360668e-02 -6.97180271e-01 -7.72109553e-02 -1.20581545e-01
8.89470637e-01 4.88378376e-01 2.51398414e-01 -9.53069106e-02
-8.34627032e-01 -6.43449485e-01 -2.32394729e-02 -4.89815265e-01
-6.59051910e-02 -1.00603139e+00 -6.98602080e-01 -6.76027358e-01
3.68439019e-01 3.88499379e-01 5.02174258e-01 8.89633119e-01
1.42949343e-01 -6.03146493e-01 -7.17534304e-01 1.00739515e+00
1.93232083e+00 9.35549587e-02 4.80685622e-01 3.79376020e-03
6.14116043e-02 1.43911228e-01 3.90349835e-01 1.13691163e+00
2.98375994e-01 2.17131272e-01 7.59787977e-01 1.81587562e-01
1.57140523e-01 3.72787625e-01 8.83644894e-02 9.33886707e-01
4.33387130e-01 -5.21728456e-01 -2.39470094e-01 4.17562723e-01
-1.54670501e+00 -6.39456928e-01 -4.89832997e-01 2.13012838e+00
1.83942780e-01 -4.29399788e-01 -6.71343148e-01 2.67009526e-01
1.78697482e-01 1.03339009e-01 -5.07162511e-01 5.74857891e-01
-3.32872689e-01 2.16200978e-01 1.67212355e+00 5.24894059e-01
-8.40995371e-01 -1.58878461e-01 5.41765547e+00 4.48820561e-01
-5.47720790e-01 1.92513883e-01 -2.31402621e-01 8.13960209e-02
-6.45700872e-01 -6.61625713e-03 -5.22378147e-01 2.20809877e-02
9.74091589e-01 -3.06951910e-01 9.64514390e-02 3.53979290e-01
4.07136440e-01 -4.06835258e-01 -7.29832470e-01 1.50872540e+00
-1.41925544e-01 -1.22147214e+00 -7.67739832e-01 3.82393360e-01
7.17710316e-01 1.57585621e-01 -6.59336448e-02 -3.35644931e-01
2.74769723e-01 -4.02736723e-01 6.24123633e-01 6.74044967e-01
5.21506369e-01 -3.59057069e-01 8.00232530e-01 4.94317174e-01
-1.18407261e+00 -5.33442795e-01 -3.03027719e-01 -4.35054153e-02
4.02397633e-01 1.60823941e+00 -5.86274803e-01 1.08666015e+00
1.44417509e-01 3.04577917e-01 1.03455983e-01 1.51733255e+00
-3.07154059e-01 6.75963163e-01 -1.12533844e+00 -4.92477924e-01
7.20065385e-02 -1.48003981e-01 1.07704425e+00 1.02320600e+00
1.02714562e+00 3.94974679e-01 -2.63855726e-01 2.07557991e-01
4.21074778e-02 -7.91058391e-02 -4.41287011e-01 3.53480965e-01
1.05711567e+00 1.23208678e+00 -2.91581333e-01 2.81891197e-01
-3.52623284e-01 3.59704345e-01 -7.66062737e-01 8.31557274e-01
-1.19008586e-01 -5.16261697e-01 5.74971735e-01 -1.54774353e-01
4.54813927e-01 -8.33895862e-01 -8.51818264e-01 -7.62033045e-01
6.70186654e-02 -3.78623188e-01 -1.50197566e-01 -7.80840218e-02
-7.60842025e-01 -1.42778018e-02 -1.86588198e-01 -1.41264582e+00
-2.55073518e-01 -1.59111127e-01 -5.12331843e-01 5.68893969e-01
-1.70159149e+00 -1.15504909e+00 -3.46376657e-01 2.91372120e-01
-2.94206142e-01 -2.88645148e-01 7.68751204e-01 3.15796673e-01
-3.02914143e-01 5.29254735e-01 9.14713860e-01 -1.83084682e-01
-5.53291626e-02 -6.01608574e-01 -2.70665675e-01 1.25459325e+00
-1.33831158e-01 7.07677126e-01 1.39527249e+00 -5.65420151e-01
-2.54819155e+00 -1.05141842e+00 5.93405008e-01 4.75981206e-01
6.32819712e-01 -3.45120430e-01 3.93131942e-01 4.45450753e-01
2.77696878e-01 -2.82826692e-01 1.02863300e+00 -3.76939118e-01
8.33945870e-02 -2.10518479e-01 -1.39358652e+00 4.09237355e-01
8.00412536e-01 3.34657758e-01 9.53494385e-02 5.04281640e-01
8.23699757e-02 -2.19427168e-01 -9.81560469e-01 5.52393854e-01
8.07795048e-01 -5.04161894e-01 1.10714388e+00 3.93478245e-01
-3.73671979e-01 -3.41796368e-01 -1.05684686e+00 -1.18436062e+00
-3.73627514e-01 -1.32561433e+00 -2.32675895e-01 9.58297789e-01
4.93574440e-01 -6.33087456e-01 1.03237402e+00 1.05219074e-01
5.02725132e-02 -4.94438022e-01 -1.38322854e+00 -7.39489853e-01
-6.59969747e-01 -7.95020044e-01 2.46413097e-01 4.85432483e-02
-1.86383687e-02 1.33382142e-01 -6.20579302e-01 1.24343765e+00
1.70329154e+00 -3.13897252e-01 1.02064121e+00 -1.20894325e+00
-4.76065964e-01 2.08419129e-01 -6.02485836e-01 -1.51499367e+00
-4.23936844e-01 -6.01701617e-01 2.84003317e-01 -1.83306098e+00
-2.80964345e-01 -5.89665234e-01 -8.37704390e-02 -3.59983034e-02
4.43032295e-01 1.81492999e-01 -1.52233029e-02 -2.67463201e-03
-4.56268609e-01 6.39784575e-01 5.74101985e-01 -1.81981534e-01
-1.20711196e-02 7.80989408e-01 -6.70907855e-01 4.01613027e-01
5.58788657e-01 -7.41302729e-01 -2.59826362e-01 -5.82305491e-01
4.91450429e-01 7.20196128e-01 2.50834316e-01 -1.58566892e+00
8.15496266e-01 2.49652471e-02 2.54303038e-01 -7.32085347e-01
5.88099957e-01 -1.50296450e+00 5.37496746e-01 1.00794089e+00
1.32713348e-01 -5.69296360e-01 -2.95975477e-01 1.14730048e+00
3.89946669e-01 -1.16236068e-01 6.46641970e-01 2.25026920e-01
-7.00927228e-02 2.12208703e-01 -6.03342175e-01 -6.77933633e-01
1.06146252e+00 -2.84874588e-01 -2.57678945e-02 -1.01485169e+00
-4.97283667e-01 3.12097520e-01 -3.51677626e-01 -4.34262991e-01
7.73364604e-01 -9.73147690e-01 -1.02941465e+00 2.89560169e-01
1.96275227e-02 -4.58421648e-01 2.82708883e-01 1.19548523e+00
-1.74498469e-01 5.62256455e-01 3.11353296e-01 -4.56573844e-01
-1.14824021e+00 -3.12103689e-01 1.16782628e-01 9.35837347e-03
-4.30818617e-01 7.89750814e-01 -3.35182756e-01 -7.11383149e-02
4.27695245e-01 -1.52213559e-01 2.32420027e-01 -4.85712647e-01
5.50376832e-01 2.79522508e-01 2.19283149e-01 -3.09529573e-01
-4.06119734e-01 8.05526078e-01 6.41556025e-01 -4.39056069e-01
1.67879522e+00 -6.41848922e-01 -1.53116763e-01 -2.06681669e-01
1.03996539e+00 3.77582103e-01 -7.26759017e-01 -6.47995353e-01
3.05736233e-02 -2.66676515e-01 2.65150100e-01 -5.63337564e-01
-9.25320327e-01 1.40597314e-01 7.19481766e-01 -3.51832621e-02
1.07758498e+00 -1.62493616e-01 5.67174494e-01 1.09657371e+00
1.14159048e+00 -5.59435487e-01 -4.77989763e-01 4.70866002e-02
5.01203001e-01 -6.41120136e-01 6.38914466e-01 -5.68475604e-01
4.61560369e-01 1.08087289e+00 -4.84707296e-01 -9.32651013e-02
1.28682804e+00 9.29173470e-01 1.15967587e-01 -2.78162539e-01
-4.34728652e-01 -4.47203070e-02 -2.49714732e-01 8.67294669e-01
6.97173700e-02 1.12518884e-01 -5.19368589e-01 6.80097878e-01
-3.73716414e-01 -4.34171617e-01 6.71933353e-01 1.24727297e+00
-8.84708703e-01 -9.04240847e-01 -8.38100910e-01 6.84299648e-01
-3.45813960e-01 8.22210964e-03 3.11802864e-01 3.37937742e-01
-3.55293959e-01 1.75415206e+00 -5.89286089e-01 -3.59306663e-01
2.80967712e-01 -7.62275755e-01 6.26002789e-01 -2.25600645e-01
-3.41986679e-02 3.39300036e-01 1.72290318e-02 -4.30105031e-01
-4.41256493e-01 -5.76932669e-01 -8.79740596e-01 3.08261663e-01
-3.74314576e-01 4.16819334e-01 1.50069439e+00 1.22630799e+00
2.94924468e-01 5.76110125e-01 1.01977324e+00 -7.39640296e-01
-8.42185497e-01 -8.46999466e-01 -8.91819835e-01 -5.31111360e-01
4.94572580e-01 -2.88438946e-01 -5.30110300e-01 -4.88428235e-01] | [6.283343315124512, 1.2648673057556152] |
376ed628-6192-47e5-8b83-587b834d794e | automatic-prosody-prediction-for-chinese | 1511.00360 | null | http://arxiv.org/abs/1511.00360v1 | http://arxiv.org/pdf/1511.00360v1.pdf | Automatic Prosody Prediction for Chinese Speech Synthesis using BLSTM-RNN and Embedding Features | Prosody affects the naturalness and intelligibility of speech. However,
automatic prosody prediction from text for Chinese speech synthesis is still a
great challenge and the traditional conditional random fields (CRF) based
method always heavily relies on feature engineering. In this paper, we propose
to use neural networks to predict prosodic boundary labels directly from
Chinese characters without any feature engineering. Experimental results show
that stacking feed-forward and bidirectional long short-term memory (BLSTM)
recurrent network layers achieves superior performance over the CRF-based
method. The embedding features learned from raw text further enhance the
performance. | ['Chuang Ding', 'Yang Liu', 'Weini Zhang', 'Lei Xie', 'Jie Yan'] | 2015-11-02 | null | null | null | null | ['prosody-prediction'] | ['natural-language-processing'] | [-9.31326523e-02 -3.96179333e-02 -2.89139569e-01 -5.54273248e-01
-6.12904489e-01 -8.20138082e-02 8.41489285e-02 -3.58035356e-01
-3.79070759e-01 8.38174820e-01 7.50632644e-01 -4.95736003e-01
6.30140245e-01 -6.35972440e-01 -3.01751763e-01 -5.00164032e-01
4.82496560e-01 -9.92765278e-02 1.06994890e-01 -2.04599306e-01
-6.73916936e-02 1.69355512e-01 -8.12486768e-01 4.22463357e-01
7.81237006e-01 9.89425898e-01 6.61762416e-01 6.45877302e-01
-5.64826012e-01 1.01442099e+00 -6.95628762e-01 -1.14351764e-01
-4.03583735e-01 -5.79372823e-01 -6.57239616e-01 -1.74557000e-01
-5.10934114e-01 -2.20733508e-01 -2.82740980e-01 1.10675156e+00
5.37271142e-01 2.53631026e-01 3.46610308e-01 -3.66621494e-01
-9.42848682e-01 1.35013354e+00 -3.75965163e-02 6.93153962e-02
1.30245373e-01 -9.22211930e-02 1.06980753e+00 -9.80645061e-01
3.31642419e-01 1.12861073e+00 6.48094893e-01 9.13032472e-01
-1.05360138e+00 -7.57487476e-01 3.32651466e-01 2.21627593e-01
-1.27039659e+00 -6.36147916e-01 1.02925158e+00 -2.01004669e-01
1.13503146e+00 1.84779074e-02 2.49647319e-01 1.22158527e+00
-4.15744670e-02 7.97568679e-01 8.74073625e-01 -5.80093086e-01
-2.28985921e-02 1.91896901e-01 2.35837698e-01 3.57543111e-01
-7.89795101e-01 1.83869094e-01 -6.04102612e-01 3.24088395e-01
5.09736717e-01 -2.33143166e-01 -3.08845907e-01 6.27590239e-01
-1.01225126e+00 8.07890475e-01 2.40715206e-01 5.89125872e-01
-3.66829574e-01 7.05776289e-02 5.62486887e-01 6.73998520e-02
4.30316299e-01 1.99721023e-01 -5.64393520e-01 -3.93004984e-01
-8.63604546e-01 -2.61437833e-01 6.33945048e-01 9.75222647e-01
3.59026343e-01 8.42748344e-01 -9.25677866e-02 1.15869689e+00
3.74503165e-01 5.07081568e-01 8.01728725e-01 -5.76273263e-01
4.94294345e-01 8.79639536e-02 -1.20401613e-01 -7.78538644e-01
-1.47405311e-01 -6.80728629e-02 -7.25316644e-01 -2.70365179e-01
9.22001898e-02 -8.84008825e-01 -8.68433058e-01 1.62438285e+00
-1.65456414e-01 -7.77604878e-02 5.14028072e-01 9.13359284e-01
1.04894447e+00 1.43980443e+00 2.49737367e-01 -5.79802394e-01
9.15238202e-01 -1.19712055e+00 -1.48573256e+00 -1.90388769e-01
2.37867504e-01 -1.10885954e+00 1.23379230e+00 3.52125406e-01
-1.08089375e+00 -7.83209622e-01 -9.84251261e-01 -1.44916639e-01
-2.09137261e-01 4.89527136e-01 2.05521777e-01 8.05194378e-01
-7.04096496e-01 4.46551502e-01 -8.54876637e-01 3.64033699e-01
-1.21916845e-01 3.86974663e-01 -4.62410040e-02 5.51244855e-01
-1.48944032e+00 6.35493219e-01 4.82387513e-01 4.18136179e-01
-2.13179171e-01 -3.21234077e-01 -8.58323753e-01 2.74019867e-01
1.62868887e-01 1.48978144e-01 1.43475425e+00 -8.26477587e-01
-2.48125172e+00 1.98820923e-02 -4.94576961e-01 -5.31648517e-01
-1.41815111e-01 -4.54594344e-01 -6.76109612e-01 -6.24554381e-02
-4.84578580e-01 7.17996538e-01 6.74105465e-01 -8.14128995e-01
-4.22172874e-01 1.28872216e-01 -7.47986019e-01 6.10712767e-02
-5.19971430e-01 6.59353018e-01 1.35016799e-01 -9.27270472e-01
9.58564058e-02 -7.11737394e-01 -2.98464328e-01 -8.13626826e-01
-5.58346927e-01 -6.08746529e-01 8.10196698e-01 -1.01553166e+00
1.42655957e+00 -2.18025398e+00 -2.23091871e-01 -2.79137194e-01
-3.27953905e-01 6.40827119e-01 1.46357566e-01 1.77872583e-01
2.18708366e-01 2.88009852e-01 5.00198230e-02 -3.37705523e-01
-1.54248383e-02 1.52234197e-01 -7.65969694e-01 -1.55322745e-01
5.13871610e-01 8.18864167e-01 -4.61355686e-01 -5.05365610e-01
2.85896748e-01 6.24394715e-01 -3.22531998e-01 5.15967667e-01
-4.64413166e-01 6.88534915e-01 -2.68524945e-01 9.57836732e-02
3.55400562e-01 2.23463193e-01 3.08550507e-01 1.85625583e-01
-5.36105335e-01 1.15546513e+00 -5.79142451e-01 1.51911449e+00
-8.04824948e-01 5.62104821e-01 -4.96987924e-02 -6.40243530e-01
1.59814870e+00 9.62430120e-01 -8.98345858e-02 -5.69400191e-01
2.15826333e-01 2.21051574e-01 7.26686269e-02 -4.11854506e-01
5.74053526e-01 -6.22918129e-01 -1.25296041e-01 9.88208354e-02
1.19174942e-01 -5.19211330e-02 -5.28597295e-01 -4.37194854e-01
5.63667893e-01 1.00609429e-01 2.05705658e-01 3.81838083e-02
5.68573952e-01 -2.65659571e-01 1.18155074e+00 8.95754769e-02
-1.93473101e-01 7.43575633e-01 3.21906775e-01 -1.83970079e-01
-9.78926897e-01 -1.01504993e+00 -2.47404352e-02 1.08769727e+00
-2.97003120e-01 -4.77721512e-01 -7.48074055e-01 -3.91002685e-01
-6.17348373e-01 1.01782775e+00 -4.63079102e-03 9.40631256e-02
-1.12189102e+00 -2.06297100e-01 7.02412426e-01 8.01450014e-01
3.47334743e-01 -1.69425976e+00 4.69834618e-02 6.97214007e-01
-3.67815584e-01 -1.42291141e+00 -8.43356729e-01 2.20468193e-01
-5.77095091e-01 3.03058960e-02 -7.57043302e-01 -1.28736424e+00
2.75719818e-02 -2.61333078e-01 4.76561844e-01 -2.81380445e-01
4.16678727e-01 -5.69017708e-01 -6.33412957e-01 -2.95944661e-01
-8.20334017e-01 2.35363677e-01 1.06107198e-01 -2.66740676e-02
6.08297110e-01 -5.25939822e-01 -8.38818774e-02 2.04799756e-01
-3.35297585e-01 1.85279459e-01 4.68802691e-01 9.23670888e-01
6.32492125e-01 -1.58146068e-01 1.29965246e+00 -6.40237927e-01
6.75300419e-01 -1.12696595e-01 -6.07336581e-01 1.90804511e-01
-4.82939363e-01 1.16700657e-01 1.14369702e+00 -6.52625740e-01
-1.68644464e+00 2.49628112e-01 -8.63860965e-01 -2.44388118e-01
-4.07146811e-01 7.29157805e-01 -5.55711150e-01 5.61293244e-01
1.53183684e-01 2.73788899e-01 -1.17786720e-01 -9.28999007e-01
3.50773901e-01 1.30657434e+00 5.82105696e-01 -2.54678488e-01
1.92042887e-01 -2.75681227e-01 -6.52459323e-01 -1.14207935e+00
-7.57193685e-01 -1.29778221e-01 -5.89812994e-01 -4.25448781e-03
1.33699071e+00 -7.97821581e-01 -7.49291301e-01 4.59176302e-01
-1.66947222e+00 -4.00280356e-01 5.39460965e-03 1.06088698e+00
-4.79328275e-01 3.19290489e-01 -1.30115449e+00 -9.10652995e-01
-5.28815031e-01 -1.08950996e+00 3.66274148e-01 3.81528676e-01
-8.62532035e-02 -8.87977660e-01 -4.36088853e-02 2.78652459e-01
5.02456784e-01 -3.10880035e-01 8.62739086e-01 -6.55309498e-01
-2.78219104e-01 9.09567475e-02 -1.48671925e-01 8.84399533e-01
3.55286181e-01 -6.07499480e-02 -1.20495689e+00 4.26799923e-01
3.10198933e-01 -2.99229443e-01 7.35516012e-01 4.99416590e-01
9.38368678e-01 -5.62828600e-01 1.26910642e-01 3.77873451e-01
9.73103404e-01 5.20416498e-01 5.83058953e-01 -3.19429517e-01
8.77626896e-01 7.02154040e-01 5.71149588e-01 1.95640832e-01
2.65420169e-01 4.15149063e-01 -2.42248580e-01 3.72340754e-02
-2.80591339e-01 -5.65156400e-01 8.65819931e-01 1.89490104e+00
2.82113045e-01 -1.40209943e-01 -9.01401818e-01 4.98269707e-01
-1.51422429e+00 -7.18804717e-01 -1.28576249e-01 1.82588291e+00
1.40455949e+00 4.24076498e-01 -1.45247132e-01 8.31388831e-02
1.07351875e+00 2.16707572e-01 -8.84895995e-02 -9.53822553e-01
-9.68116298e-02 2.67256409e-01 1.30881906e-01 7.01713860e-01
-8.70516896e-01 1.49484062e+00 6.00202036e+00 1.05783284e+00
-1.49957848e+00 1.96502849e-01 5.74242711e-01 3.46390873e-01
-3.20257396e-01 -2.27262620e-02 -1.26299024e+00 4.27164048e-01
1.48479939e+00 -7.25182295e-02 3.29175174e-01 7.74565220e-01
5.65330982e-01 2.63108224e-01 -6.57420456e-01 6.55605018e-01
-2.11993381e-01 -1.29570198e+00 -1.56111434e-01 -2.31416106e-01
5.39963365e-01 3.57950851e-02 3.66267264e-02 4.45495725e-01
2.63542503e-01 -1.17156863e+00 6.78206801e-01 5.82057297e-01
8.78265798e-01 -9.29675639e-01 7.42935717e-01 5.07678032e-01
-1.25307584e+00 2.20348418e-01 -4.59889472e-01 -2.91473776e-01
7.16624856e-01 5.36074519e-01 -1.29698122e+00 1.57625228e-01
2.99293399e-01 5.14009535e-01 -1.03187431e-02 5.40080965e-01
-4.83080149e-01 1.49844205e+00 -1.27982542e-01 -5.96596479e-01
2.79196709e-01 -4.24266309e-02 4.32113439e-01 1.55805278e+00
2.03605089e-02 1.85829267e-01 2.69678295e-01 1.01679599e+00
-9.75952670e-02 3.67474794e-01 -3.55028600e-01 -4.36996490e-01
4.19727892e-01 9.16893661e-01 -4.85359728e-01 -4.87601571e-02
-5.13982177e-01 5.95100582e-01 3.11920792e-01 2.55267978e-01
-7.31743574e-01 -6.21608913e-01 4.95997131e-01 -2.20327586e-01
6.54857934e-01 -6.25416398e-01 -6.08368337e-01 -9.44755733e-01
-1.10984445e-01 -4.73078310e-01 -1.07455961e-01 -6.58306897e-01
-1.36264324e+00 1.02626252e+00 -5.65263927e-01 -1.03379405e+00
-5.06127715e-01 -4.22001302e-01 -8.04085672e-01 1.34653068e+00
-1.53143036e+00 -1.04231346e+00 3.69143277e-01 2.01562002e-01
9.72889006e-01 -2.52269983e-01 1.12324107e+00 1.70106515e-02
-7.06968784e-01 5.52508891e-01 3.43140215e-02 6.58836246e-01
4.23090756e-01 -1.20013952e+00 6.46674514e-01 8.44155729e-01
1.18031427e-01 5.09965897e-01 5.10549188e-01 -7.36482680e-01
-6.20047152e-01 -1.06301820e+00 1.29612565e+00 2.51766235e-01
7.52904177e-01 -4.67402041e-01 -1.12360704e+00 4.82167810e-01
5.76481640e-01 -1.93304032e-01 8.86187315e-01 1.29057601e-01
7.60502592e-02 -9.55450460e-02 -4.80285645e-01 6.33907557e-01
4.37714577e-01 -7.07757115e-01 -8.04852784e-01 -2.01571822e-01
1.54288554e+00 -3.21987346e-02 -8.92514050e-01 3.73187482e-01
2.98494846e-01 -5.99323690e-01 4.52817708e-01 -5.05211174e-01
2.49174431e-01 -1.46380678e-01 -2.59182692e-01 -1.14450049e+00
7.01986440e-03 -8.35700274e-01 2.21032634e-01 1.66429090e+00
9.29040134e-01 -4.01868194e-01 7.57148743e-01 4.55515087e-01
-4.81977016e-01 -5.29423535e-01 -7.79835224e-01 -7.94581592e-01
4.32880551e-01 -5.18912315e-01 5.27231634e-01 6.49751306e-01
4.82377082e-01 8.66347075e-01 -6.08980358e-01 -2.03473475e-02
-6.60259947e-02 2.00163163e-02 1.92383364e-01 -7.62918472e-01
-2.98569888e-01 -2.26303667e-01 -2.98862923e-02 -1.45809960e+00
5.41515589e-01 -5.44779658e-01 6.69963062e-01 -1.24839282e+00
-4.11128581e-01 -5.63243985e-01 -3.07836384e-01 4.12646711e-01
-2.11611643e-01 -2.22536847e-01 2.62203127e-01 -1.12233102e-01
-2.06334352e-01 1.09951329e+00 1.01273084e+00 1.43872470e-01
-6.95916414e-01 3.73142630e-01 -3.76308709e-02 7.32726514e-01
1.37976372e+00 -5.41283071e-01 -3.10127605e-02 -3.27596158e-01
-1.20584778e-01 7.80223250e-01 -2.76969075e-01 -9.42506552e-01
3.96827757e-01 -3.34820509e-01 1.99262768e-01 -8.00989389e-01
4.96994913e-01 -3.96801800e-01 -3.97024632e-01 1.62543908e-01
-6.47725940e-01 -2.32471768e-02 1.36912033e-01 2.46980637e-01
-6.73677087e-01 -4.65202868e-01 6.75995469e-01 -1.05657382e-02
-5.37453592e-01 -1.31205112e-01 -9.03702021e-01 -1.97485089e-01
3.83950055e-01 8.83268341e-02 -3.42531167e-02 -4.63124931e-01
-9.99013126e-01 -1.29955143e-01 -2.40505114e-01 5.30217409e-01
8.34074855e-01 -1.15795648e+00 -5.80378234e-01 2.89605349e-01
-5.40625334e-01 -5.11256494e-02 1.00428313e-01 5.42970538e-01
-2.14383751e-01 6.31506443e-01 -2.60724165e-02 -2.52867728e-01
-1.15939116e+00 4.55706865e-01 1.35427326e-01 -1.14285432e-01
-6.41303360e-01 8.49931300e-01 -2.67194565e-02 -5.07233441e-01
3.27584624e-01 -2.99584478e-01 -6.85390770e-01 -3.34384859e-01
4.83281851e-01 1.19182773e-01 -1.19832695e-01 -7.57416427e-01
-1.14803806e-01 9.89662409e-02 -1.97563112e-01 -5.31110823e-01
1.20983183e+00 -3.08458537e-01 1.68537498e-01 7.96559691e-01
1.11065936e+00 5.30288279e-01 -1.27359891e+00 -2.58034319e-01
3.92157048e-01 1.60323411e-01 1.91576734e-01 -7.77336597e-01
-9.04168010e-01 1.43420112e+00 6.81139603e-02 1.96421117e-01
9.01676953e-01 -2.55071193e-01 1.47538376e+00 4.83618528e-01
-5.05795516e-02 -1.39986920e+00 -5.26707098e-02 1.01337969e+00
6.52008176e-01 -9.93444085e-01 -6.73983753e-01 -4.91094977e-01
-1.02361763e+00 1.46516061e+00 6.21410072e-01 -1.90450907e-01
1.00242448e+00 5.24529636e-01 4.77700651e-01 5.50284743e-01
-8.94406021e-01 -9.82376188e-02 1.07811153e-01 4.80931342e-01
7.63246179e-01 2.96798438e-01 -3.55038047e-01 1.23934352e+00
-6.77376926e-01 -1.83879957e-01 6.37735248e-01 5.13052046e-01
-6.53206170e-01 -1.40921772e+00 -1.97073132e-01 1.74219936e-01
-7.40159631e-01 -4.52455282e-01 -3.40367079e-01 1.59528684e-02
-1.04915880e-01 1.29552698e+00 9.14255455e-02 -6.48375750e-01
-6.46135509e-02 4.06972975e-01 -2.88732760e-02 -8.68087649e-01
-5.79423189e-01 5.55908084e-01 3.07712257e-01 -5.11027500e-02
-2.54089564e-01 -6.07637882e-01 -1.84908736e+00 1.09563112e-01
-7.88618565e-01 3.99237782e-01 9.34388638e-01 9.46567476e-01
1.23543061e-01 4.64276075e-01 1.02175796e+00 -4.03942525e-01
-5.33187807e-01 -1.31764698e+00 -4.79890615e-01 -3.22994649e-01
4.14393067e-01 -2.41658628e-01 -2.71395266e-01 3.88233662e-01] | [14.728447914123535, 6.77061128616333] |
32de6110-7818-4a13-8253-7ccd70554abb | e2gan-end-to-end-generative-adversarial | null | null | https://doi.org/10.24963/ijcai.2019/429 | https://www.ijcai.org/proceedings/2019/0429.pdf | E2GAN: End-to-End Generative Adversarial Network or Multivariate Time Series Imputation | The missing values, appear in most of multivariate time series, prevent advanced analysis of multivariate time series data. Existing imputation approaches try to deal with missing values by deletion, statistical imputation, machine learning based imputation and generative imputation. However, these methods are either incapable of dealing with temporal information or multi-stage. This paper proposes an end-to-end generative model E2GAN to impute missing values in multivariate time series. With the help of the discriminative loss and the squared error loss, E2GAN can imputethe incomplete time series by the nearest generated complete time series at one stage. Experiments on multiple real-world datasets show that our model outperforms the baselines on the imputation accuracy and achieves state-of-the-art classification/regression results on the downstream applications. Additionally, our method also gains better time efficiency than multi-stage method on the training of neural networks. | ['Xiangrui Cai', 'Yonghong Luo', 'Xiaojie Yuan', 'Ying Zhang'] | 2019-08-10 | null | null | null | proceedings-of-the-twenty-eighth | ['multivariate-time-series-imputation'] | ['time-series'] | [-1.37936631e-02 -1.25798047e-01 -4.36038196e-01 -4.96771365e-01
-1.13364875e+00 -3.22143376e-01 4.10508454e-01 -3.19835424e-01
-1.61538050e-01 1.37150466e+00 3.13064545e-01 -2.65054107e-01
-2.51343518e-01 -6.77448571e-01 -9.71349895e-01 -7.84487069e-01
-2.20276907e-01 4.42706615e-01 -7.71427751e-01 4.50012051e-02
-1.29869476e-01 1.70948114e-02 -1.14331591e+00 3.10745597e-01
9.78943884e-01 9.64999557e-01 -4.62765187e-01 2.84257889e-01
-5.88775016e-02 9.54458177e-01 -5.25753915e-01 -5.35316825e-01
2.61097789e-01 -5.97132325e-01 -6.78624511e-02 -5.60841143e-01
-2.47134045e-01 -3.77640188e-01 -2.33087003e-01 3.93029869e-01
5.96134603e-01 5.39631620e-02 6.90860450e-01 -1.72914064e+00
-4.01391566e-01 8.49604130e-01 -7.67262995e-01 -3.53875337e-03
6.25589490e-02 -7.41275325e-02 1.32932514e-01 -8.43207061e-01
3.04011077e-01 8.38438332e-01 1.38233113e+00 4.56852525e-01
-1.48198926e+00 -1.03445554e+00 -2.13856190e-01 -6.59308434e-02
-1.20959699e+00 -5.85601747e-01 8.99380088e-01 -4.04618829e-01
9.81915295e-01 2.43202627e-01 1.41659990e-01 1.58562553e+00
4.87213314e-01 5.54154396e-01 1.05947399e+00 3.95938642e-02
3.00696641e-01 -5.78096330e-01 -1.09524608e-01 -1.20057061e-01
2.97616962e-02 5.05838871e-01 -6.79664135e-01 -3.66397411e-01
7.11145699e-01 5.42040467e-01 1.91536725e-01 1.39931902e-01
-1.45092523e+00 5.36420226e-01 -1.16407871e-02 -1.05997724e-02
-7.64736056e-01 3.54586601e-01 5.66004574e-01 5.95405817e-01
9.43119109e-01 -1.89300850e-01 -7.98456848e-01 -4.30286676e-01
-1.29628611e+00 1.18563332e-01 7.55395353e-01 1.01263738e+00
3.70583802e-01 4.72376198e-01 -3.85893524e-01 6.33160233e-01
-7.52522498e-02 5.96464574e-01 5.08325756e-01 -8.94852102e-01
9.58614886e-01 1.73430815e-01 1.05297677e-01 -5.38598478e-01
-2.96110541e-01 -5.41693151e-01 -1.84795845e+00 -5.35967797e-02
5.00100255e-01 -8.57385635e-01 -9.14927006e-01 2.06332040e+00
1.48894385e-01 8.05567682e-01 1.99284464e-01 5.04485130e-01
4.99505401e-01 5.53807437e-01 3.13938260e-01 -6.59169853e-01
6.64232135e-01 -5.25091588e-01 -8.93848896e-01 1.80191752e-02
4.67047542e-01 -6.02369130e-01 1.90140039e-01 3.33188713e-01
-1.05689692e+00 -5.02971649e-01 -3.55405003e-01 2.55790204e-02
-1.48118198e-01 1.21835463e-01 8.55103314e-01 4.39649642e-01
-5.87077379e-01 7.75893748e-01 -1.20768809e+00 2.95939833e-01
5.07446945e-01 4.15419102e-01 -4.84606415e-01 1.94872066e-01
-1.09332407e+00 6.00461900e-01 1.59958571e-01 4.01487321e-01
-8.82339060e-01 -1.40440655e+00 -6.35576248e-01 5.57640102e-03
-2.52044648e-01 -1.02677703e+00 9.43883240e-01 -9.55170035e-01
-1.06904519e+00 3.72705519e-01 -4.49629515e-01 -5.80068171e-01
9.53516781e-01 -1.55244783e-01 -5.79729617e-01 -7.44483888e-01
1.01880655e-01 3.15970570e-01 7.97675729e-01 -7.22323596e-01
-2.89742321e-01 -5.40209413e-01 -8.26068878e-01 -3.00362021e-01
2.56502956e-01 -2.41778359e-01 7.75023848e-02 -1.20080376e+00
2.08675727e-01 -6.16699994e-01 -3.57960582e-01 -5.36253095e-01
-5.20487249e-01 1.14501482e-02 6.19845033e-01 -1.26015937e+00
1.07334292e+00 -2.04927802e+00 1.36470184e-01 6.09224439e-02
-9.32639614e-02 -3.15914273e-01 -1.18353583e-01 6.55395150e-01
-5.51431119e-01 4.04985808e-02 -4.05141771e-01 -9.29834187e-01
7.02487677e-02 1.44074947e-01 -6.65651023e-01 4.57094818e-01
-4.40996587e-02 1.14362061e+00 -5.81466198e-01 -2.17184335e-01
1.56442910e-01 7.82640934e-01 -2.30227605e-01 3.04220676e-01
1.41856195e-02 1.00207782e+00 -1.91128850e-01 8.92590702e-01
9.05002713e-01 7.20121935e-02 9.97310653e-02 8.86208564e-02
-3.35028744e-03 2.51315050e-02 -7.73528874e-01 1.98354006e+00
-3.61700624e-01 3.26633066e-01 -2.08492264e-01 -9.84547377e-01
9.84597206e-01 6.05724216e-01 7.48569131e-01 -5.82052708e-01
4.38762903e-02 1.30803138e-01 -3.76662493e-01 -4.18757856e-01
9.08129141e-02 -2.13652760e-01 -3.74671906e-01 2.88081884e-01
-1.11069232e-01 4.98969913e-01 -1.38098970e-01 -3.55099827e-01
1.00427032e+00 7.84582794e-01 -8.85973976e-04 3.10497761e-01
-8.69271066e-03 -1.46008804e-01 1.13552129e+00 9.20424581e-01
2.03970745e-01 8.34300697e-01 5.34979284e-01 -5.52515268e-01
-1.32299531e+00 -1.16005909e+00 -8.56528953e-02 8.04961383e-01
-6.75479174e-01 7.27409571e-02 -5.23341358e-01 -4.04655039e-01
2.61620790e-01 7.82447517e-01 -6.38661921e-01 1.95065923e-02
-6.24886036e-01 -1.38230872e+00 7.60648787e-01 7.39517629e-01
3.33050668e-01 -1.04786980e+00 -2.01574877e-01 5.60768723e-01
-7.64046013e-01 -7.18997002e-01 -1.95203841e-01 2.96277344e-01
-1.47199273e+00 -6.04864538e-01 -5.93300521e-01 -4.25112993e-01
6.19434595e-01 -3.47508997e-01 1.16269279e+00 -2.83133775e-01
5.75726703e-02 -2.25810260e-01 -2.52496213e-01 -6.21985972e-01
-6.23553246e-02 2.02574581e-01 -7.57966042e-02 1.15174845e-01
6.01705313e-01 -1.28879762e+00 -5.20007968e-01 -6.26770854e-02
-6.80125356e-01 2.12802023e-01 2.99634397e-01 9.87342596e-01
7.22550452e-01 -1.17348269e-01 1.14147043e+00 -6.65622413e-01
3.51315558e-01 -1.18407512e+00 -5.39895833e-01 -1.03849010e-03
-6.57520533e-01 -2.32121386e-02 7.52035081e-01 -5.01150191e-01
-1.04137301e+00 1.67093948e-01 -1.23580322e-01 -6.28550589e-01
-2.94654161e-01 8.47069860e-01 -7.82116726e-02 5.38571715e-01
2.56915420e-01 3.77749741e-01 4.68156599e-02 -8.59182477e-01
5.99051118e-02 5.61802924e-01 8.27150941e-01 -3.53803515e-01
6.55104280e-01 3.24719340e-01 3.11012447e-01 -1.27068415e-01
-5.24843991e-01 -5.56156673e-02 -6.65895998e-01 -9.04376339e-03
4.31723714e-01 -1.47240305e+00 -6.85792685e-01 8.48204553e-01
-1.12941730e+00 -2.96882838e-01 -2.17159465e-01 7.32478023e-01
-7.70015001e-01 4.47993726e-02 -6.39658570e-01 -1.10511851e+00
-6.47341549e-01 -4.31492001e-01 9.34689760e-01 -1.80792496e-01
-2.99454063e-01 -9.24414337e-01 1.29652530e-01 2.48729691e-01
5.25877357e-01 1.04869747e+00 9.16505396e-01 -4.18165624e-01
-5.05142152e-01 -4.57185775e-01 2.47186065e-01 7.09303617e-02
9.99479592e-02 -1.97121084e-01 -9.50130403e-01 -1.33953765e-01
3.55990790e-02 -1.47415791e-02 9.25541162e-01 7.96111703e-01
1.64782071e+00 -4.61674213e-01 -2.15028077e-01 1.09151602e+00
1.42559409e+00 1.38555646e-01 1.07077813e+00 1.05069622e-01
6.02931619e-01 3.85375291e-01 4.34945166e-01 9.65886831e-01
5.30952096e-01 2.37380818e-01 4.46641356e-01 -1.88871056e-01
3.07758093e-01 -4.80272412e-01 4.20707315e-01 8.27408493e-01
-2.91431159e-01 -2.99893707e-01 -7.72875726e-01 7.78875887e-01
-2.30519867e+00 -1.50417852e+00 -5.70267975e-01 2.42397857e+00
9.82392192e-01 -2.33748212e-01 2.25385904e-01 1.13822095e-01
4.88941938e-01 -1.66611314e-01 -8.21617424e-01 -1.64193347e-01
-4.60155100e-01 7.97811076e-02 6.14027560e-01 1.37443645e-02
-1.09580910e+00 4.15830165e-01 6.99085236e+00 5.95956385e-01
-7.53412724e-01 3.34701300e-01 9.88438070e-01 -3.17204148e-01
-1.59630179e-01 -8.26179162e-02 -5.65959871e-01 1.06124139e+00
1.34041655e+00 -3.23824473e-02 6.69502139e-01 4.70775634e-01
6.90641761e-01 2.07006678e-01 -1.15316951e+00 9.97685313e-01
-2.87543625e-01 -1.18814480e+00 -4.02339965e-01 -2.84642596e-02
8.43503177e-01 2.92624384e-01 5.98367527e-02 4.90063608e-01
5.31771243e-01 -1.22547317e+00 4.86785650e-01 1.23159802e+00
8.75566840e-01 -9.56863046e-01 9.95215654e-01 7.13029385e-01
-6.45993292e-01 -4.11690623e-02 -1.49822578e-01 -3.77978176e-01
4.62226361e-01 9.65959907e-01 -3.40462208e-01 7.89532065e-01
7.19479322e-01 8.84266913e-01 -1.99445918e-01 1.08741665e+00
-7.17014372e-02 8.82534325e-01 -3.48518819e-01 5.61416984e-01
-3.27575892e-01 -4.58913743e-01 2.47869968e-01 8.60045612e-01
9.15176749e-01 5.38108051e-02 -1.40824348e-01 8.50081325e-01
-6.40891865e-02 -3.97493392e-01 -5.70168197e-01 5.04984930e-02
4.86204356e-01 8.49556744e-01 2.58880220e-02 -2.15929061e-01
-3.39771211e-01 7.55631447e-01 4.20336574e-02 6.61319017e-01
-1.11353624e+00 -2.11799741e-01 5.81087708e-01 -1.48338377e-01
7.65312910e-02 -2.33358771e-01 -7.78922915e-01 -1.40421987e+00
1.30235448e-01 -8.38291049e-01 5.84811985e-01 -1.03667891e+00
-1.78162122e+00 2.70719022e-01 2.67741922e-02 -1.43610930e+00
-9.50754464e-01 1.26958102e-01 -6.46806180e-01 1.24395657e+00
-1.28324795e+00 -1.45876288e+00 -2.13257656e-01 6.71680510e-01
2.73537278e-01 -2.42088467e-01 8.71314704e-01 5.28787196e-01
-6.66603446e-01 6.01720393e-01 8.06962907e-01 3.32121462e-01
6.16575897e-01 -9.29243743e-01 5.34315348e-01 8.55075002e-01
-4.49456453e-01 4.51588929e-01 6.62915587e-01 -9.24342453e-01
-1.53652751e+00 -1.41643381e+00 1.35224724e+00 -6.81753337e-01
3.22699308e-01 -4.04051632e-01 -8.90025496e-01 1.03728712e+00
2.72726506e-01 -1.69240147e-01 8.91498864e-01 3.03898960e-01
-3.78518879e-01 -3.35445762e-01 -1.33929610e+00 2.56253421e-01
9.04219210e-01 -2.82282710e-01 -5.58117807e-01 2.74353325e-01
4.86119390e-01 -3.64943534e-01 -1.22239470e+00 4.83666450e-01
7.07229555e-01 -6.09024942e-01 1.12680781e+00 -9.00449812e-01
8.11254323e-01 -1.45465761e-01 -4.46471646e-02 -1.26899028e+00
-3.09938252e-01 -7.79180229e-01 -5.04527807e-01 1.62438154e+00
4.59189057e-01 -6.51172459e-01 7.42637694e-01 8.63801062e-01
-4.98215333e-02 -1.96908921e-01 -1.28139353e+00 -8.45930040e-01
3.05172801e-01 -5.37121415e-01 1.26440394e+00 1.21115935e+00
-3.37820202e-01 -1.41730353e-01 -1.17464125e+00 -1.44293839e-02
1.18285441e+00 2.51262993e-01 8.92592013e-01 -1.23546433e+00
-1.06520437e-01 6.23373054e-02 -1.07062787e-01 -3.50481898e-01
3.68579030e-01 -6.97382033e-01 -1.43631473e-01 -1.24509025e+00
1.94652855e-01 -4.45224375e-01 -7.18947291e-01 1.00503063e+00
1.11834250e-01 3.00955206e-01 -2.69677311e-01 2.39897400e-01
-1.25473186e-01 6.04032815e-01 7.51700699e-01 -3.83733101e-02
-3.08289737e-01 1.36290014e-01 -4.32685763e-01 2.74474770e-01
1.04288697e+00 -9.33517039e-01 -1.81805283e-01 -6.43459141e-01
1.81241319e-01 9.31099415e-01 7.19961762e-01 -6.56673312e-01
1.69217095e-01 -4.44245875e-01 9.70608056e-01 -1.16087544e+00
2.15549409e-01 -8.39369953e-01 1.26850605e+00 1.71647042e-01
-1.47193745e-01 5.01618266e-01 1.46191806e-01 5.98755717e-01
-2.44867906e-01 2.28712112e-01 2.35492095e-01 5.75234108e-02
-3.35856676e-01 4.25598741e-01 -2.04228401e-01 -9.35578123e-02
7.23787785e-01 -1.78182915e-01 -2.43808448e-01 -4.58084911e-01
-9.13523138e-01 3.25606376e-01 8.76868218e-02 5.36474228e-01
4.31823403e-01 -1.59476566e+00 -1.16334152e+00 4.14202958e-01
-2.76815504e-01 -1.68448061e-01 5.03717601e-01 1.25738084e+00
-5.71643794e-03 1.28400892e-01 -2.69406676e-01 -3.78262103e-01
-8.24192941e-01 5.58777690e-01 1.91550344e-01 -2.38048360e-01
-4.68513817e-01 2.02574611e-01 -2.37365559e-01 -5.95087767e-01
1.25279039e-01 -1.07582018e-01 1.98813722e-01 8.55419412e-02
3.21746618e-01 6.11597359e-01 8.77405554e-02 -3.40613753e-01
-2.22637087e-01 2.74160326e-01 3.80603224e-01 1.86211374e-02
1.85876739e+00 -1.82153627e-01 -4.16909009e-01 6.08080506e-01
1.07547736e+00 -4.26329464e-01 -1.34821546e+00 5.07424697e-02
-1.79912716e-01 -4.25429463e-01 -7.10566640e-02 -1.14760303e+00
-1.05497384e+00 5.99998355e-01 4.47029412e-01 -1.18146762e-01
1.59059155e+00 -7.20939696e-01 7.53954649e-01 -1.04038194e-01
4.55214322e-01 -7.21992135e-01 -8.18729997e-01 3.94382447e-01
7.75955439e-01 -1.33228552e+00 -1.99701339e-01 1.79286338e-02
-3.86508316e-01 7.17373013e-01 2.97061950e-01 -2.81928107e-02
6.32760644e-01 5.19656181e-01 -3.01705617e-02 5.29198587e-01
-1.13682878e+00 3.88156772e-01 3.65758426e-02 7.85005331e-01
5.39465368e-01 2.01788336e-01 -3.93491387e-01 1.07093501e+00
-2.38500267e-01 5.84489167e-01 2.86841452e-01 5.83298087e-01
5.33096135e-01 -1.18016958e+00 -3.82155359e-01 6.71215355e-01
-7.90327013e-01 -2.26160318e-01 5.71227744e-02 7.02148855e-01
1.27532423e-01 1.07685339e+00 1.09847821e-01 -8.59420598e-02
2.25570723e-01 3.16302478e-01 1.57860816e-01 2.86093533e-01
-8.72070312e-01 3.04983348e-01 -7.22207129e-02 -3.75674009e-01
-2.69254684e-01 -1.00244331e+00 -9.07907665e-01 -8.17984462e-01
5.77697754e-02 5.03384434e-02 6.55684054e-01 7.52120554e-01
7.43893027e-01 4.31202084e-01 6.62959218e-01 -6.57626390e-01
-6.21522903e-01 -1.06145990e+00 -4.66026604e-01 5.78104198e-01
6.26622736e-01 -1.93795279e-01 -2.98443019e-01 3.31508428e-01] | [7.098151206970215, 3.3685669898986816] |
d52a508d-658b-4d30-8936-8734090512b7 | multilingual-denoising-pre-training-for | 2001.08210 | null | https://arxiv.org/abs/2001.08210v2 | https://arxiv.org/pdf/2001.08210v2.pdf | Multilingual Denoising Pre-training for Neural Machine Translation | This paper demonstrates that multilingual denoising pre-training produces significant performance gains across a wide variety of machine translation (MT) tasks. We present mBART -- a sequence-to-sequence denoising auto-encoder pre-trained on large-scale monolingual corpora in many languages using the BART objective. mBART is one of the first methods for pre-training a complete sequence-to-sequence model by denoising full texts in multiple languages, while previous approaches have focused only on the encoder, decoder, or reconstructing parts of the text. Pre-training a complete model allows it to be directly fine tuned for supervised (both sentence-level and document-level) and unsupervised machine translation, with no task-specific modifications. We demonstrate that adding mBART initialization produces performance gains in all but the highest-resource settings, including up to 12 BLEU points for low resource MT and over 5 BLEU points for many document-level and unsupervised models. We also show it also enables new types of transfer to language pairs with no bi-text or that were not in the pre-training corpus, and present extensive analysis of which factors contribute the most to effective pre-training. | ['Xi-An Li', 'Mike Lewis', 'Luke Zettlemoyer', 'Jiatao Gu', 'Sergey Edunov', 'Yinhan Liu', 'Naman Goyal', 'Marjan Ghazvininejad'] | 2020-01-22 | null | null | null | null | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 4.11276311e-01 3.75767238e-02 -1.12906896e-01 -3.91959876e-01
-1.79340577e+00 -8.43918562e-01 7.89438307e-01 -1.99524805e-01
-5.56313813e-01 9.31115210e-01 4.53742892e-01 -7.67826617e-01
5.90006530e-01 -9.36196819e-02 -1.07457972e+00 -4.07611132e-01
3.62245411e-01 9.34440255e-01 -2.53220528e-01 -6.08119905e-01
-2.06974521e-01 -7.04923198e-02 -8.57759655e-01 6.13831460e-01
1.05509591e+00 1.30941436e-01 6.36715770e-01 8.67885947e-01
-3.76426205e-02 5.73825896e-01 -6.81185305e-01 -9.39649343e-01
4.17751998e-01 -9.28035438e-01 -7.00630426e-01 -1.91289410e-02
6.37669802e-01 -3.69285494e-01 -2.45601982e-01 9.63991821e-01
8.02991927e-01 -2.67950684e-01 5.90928793e-01 -5.66146255e-01
-6.28426433e-01 9.42817509e-01 -4.02539164e-01 2.06915021e-01
2.38242939e-01 2.48641312e-01 7.58833051e-01 -1.07092547e+00
9.29301977e-01 1.46947980e+00 8.09666455e-01 5.85976899e-01
-1.39954150e+00 -4.39211011e-01 -4.97603625e-01 2.16649231e-02
-1.01167572e+00 -1.00054193e+00 1.98590621e-01 -3.12994570e-01
1.51766324e+00 1.16064906e-01 1.38347089e-01 1.45156753e+00
4.65945542e-01 7.40450978e-01 1.15976059e+00 -8.25470328e-01
-3.27283204e-01 -3.71415988e-02 -3.14269990e-01 5.45035779e-01
-9.32865962e-02 5.50079644e-02 -5.76216578e-01 1.48449708e-02
5.80242634e-01 -6.08291447e-01 -1.07844509e-02 2.11061195e-01
-1.59391022e+00 7.48827219e-01 -2.62559265e-01 4.28946644e-01
-3.31054688e-01 7.20871761e-02 8.83063197e-01 9.33914900e-01
8.02942276e-01 4.71110731e-01 -7.92560279e-01 -6.00915432e-01
-1.27685416e+00 -2.52989888e-01 8.32249641e-01 1.32487106e+00
8.59103918e-01 2.68072546e-01 -3.01210195e-01 9.97895777e-01
-2.99324751e-01 7.82125533e-01 6.55787170e-01 -8.86834145e-01
1.12677574e+00 -2.13975549e-01 -1.81731790e-01 -1.77135453e-01
2.01983169e-01 -6.01559520e-01 -9.06886935e-01 -2.98596889e-01
2.33218551e-01 -4.46780562e-01 -9.56853926e-01 1.84738255e+00
-3.19441631e-02 -2.57118165e-01 2.85511255e-01 5.24875760e-01
3.69852513e-01 9.33667421e-01 -2.20307291e-01 -5.92312992e-01
1.15764225e+00 -1.40685487e+00 -9.38907564e-01 -4.75600004e-01
1.00956404e+00 -1.44718480e+00 1.30788732e+00 3.56723845e-01
-1.38363290e+00 -6.84505522e-01 -8.91816974e-01 -4.85314339e-01
-2.30695307e-02 2.53698260e-01 2.93651619e-03 4.59739655e-01
-1.37505686e+00 7.06279516e-01 -9.85916853e-01 -5.27149618e-01
-1.02128685e-01 2.39229649e-01 -5.89195430e-01 -3.81544977e-01
-1.06885517e+00 1.40901017e+00 2.20118180e-01 -1.68238536e-01
-1.04244018e+00 -4.61434871e-01 -8.48821044e-01 -9.61503536e-02
-9.54996198e-02 -8.73329282e-01 1.59444535e+00 -1.38954711e+00
-1.65107930e+00 7.91596353e-01 -6.78335667e-01 -5.11079431e-01
7.70112991e-01 -4.04180199e-01 -3.74015868e-01 4.61861119e-03
3.08647215e-01 8.12256515e-01 9.43698347e-01 -9.63616252e-01
-4.26669151e-01 -2.11261269e-02 -5.78216851e-01 5.62621951e-01
-2.80037463e-01 5.53856611e-01 -5.91009378e-01 -9.90325749e-01
-2.48855010e-01 -7.62951314e-01 -1.42362043e-01 -8.18682969e-01
-2.11376265e-01 1.51220530e-01 6.61222816e-01 -1.50468993e+00
9.46187198e-01 -1.85630310e+00 5.14751732e-01 -3.79891694e-01
-3.49908233e-01 3.40259522e-01 -6.86876357e-01 8.78791928e-01
-1.41936652e-02 2.24835232e-01 -3.79607201e-01 -8.09673429e-01
-1.28293410e-01 4.28460687e-01 -2.69192189e-01 3.37719977e-01
3.02261233e-01 1.06334341e+00 -9.17842031e-01 -4.99761969e-01
2.77294163e-02 5.76526105e-01 -3.57365519e-01 2.85767168e-01
-2.59372652e-01 8.27548623e-01 2.13785902e-01 2.69338965e-01
4.35379565e-01 2.73016006e-01 3.31031889e-01 1.53972194e-01
1.14135809e-01 8.48124385e-01 -4.80101645e-01 2.09152842e+00
-7.43656337e-01 1.06892085e+00 2.69817710e-01 -7.83000708e-01
6.60639584e-01 6.65123641e-01 1.86487094e-01 -8.90500844e-01
-1.83836266e-01 7.47637510e-01 5.64430021e-02 -3.71069789e-01
7.03407884e-01 -3.63803238e-01 -7.44948313e-02 6.51992738e-01
5.42481601e-01 -2.77123034e-01 4.84448642e-01 1.95069820e-01
1.22605371e+00 3.27447563e-01 1.20864592e-01 -2.60035366e-01
3.69690001e-01 2.39934504e-01 4.81317461e-01 6.87361062e-01
3.37783433e-02 7.50164330e-01 1.97249651e-01 -5.94748296e-02
-1.68733978e+00 -8.29443693e-01 9.16209668e-02 1.37083709e+00
-4.77499187e-01 -4.63776797e-01 -1.17158461e+00 -6.04297101e-01
-4.59860891e-01 8.97457123e-01 -1.08417012e-01 7.44046941e-02
-9.58445609e-01 -5.93191206e-01 9.83881652e-01 1.54817760e-01
2.90668339e-01 -8.61792445e-01 3.48305553e-01 4.12582546e-01
-8.93579602e-01 -1.21725726e+00 -1.06387138e+00 4.48275566e-01
-1.24787951e+00 -3.71869683e-01 -8.42115164e-01 -1.20069098e+00
5.89661658e-01 3.76732051e-02 1.52196395e+00 -1.52028203e-01
3.12446475e-01 -6.79921955e-02 -3.74021113e-01 -5.42803667e-02
-1.36842561e+00 5.57112753e-01 2.59581178e-01 -5.04282057e-01
2.91942298e-01 -4.95070249e-01 -1.71286147e-02 1.97408304e-01
-7.13686168e-01 2.03020766e-01 8.54463279e-01 1.09823775e+00
5.27087569e-01 -4.39982086e-01 4.94265407e-01 -7.99839973e-01
7.16486216e-01 -1.40008554e-01 -2.58858174e-01 4.06178415e-01
-4.32699859e-01 2.84546643e-01 8.06832612e-01 -4.94076431e-01
-9.72213268e-01 -4.55617122e-02 -4.67050016e-01 -2.84398884e-01
3.29130553e-02 5.03145993e-01 -2.77356863e-01 3.16923618e-01
8.70258987e-01 5.38858950e-01 1.99071709e-02 -7.02250063e-01
6.01011276e-01 1.00967038e+00 6.89788818e-01 -6.37334943e-01
7.66438901e-01 -2.54582822e-01 -5.12991548e-01 -7.04165697e-01
-6.10485077e-01 -4.23576444e-01 -8.98361564e-01 2.37336293e-01
6.30656898e-01 -1.32749414e+00 2.24782363e-01 2.69830704e-01
-1.49408185e+00 -6.98180318e-01 -1.62141487e-01 5.11437178e-01
-7.56825686e-01 5.17324686e-01 -1.23695707e+00 -2.87866116e-01
-7.59598732e-01 -1.40023303e+00 1.16806483e+00 -5.78354001e-01
-2.99678743e-01 -9.93232667e-01 3.18627775e-01 5.28496325e-01
3.87067527e-01 -3.76309812e-01 9.13765967e-01 -5.53853631e-01
-2.61297971e-01 6.34218678e-02 4.10115123e-02 7.78170705e-01
1.11812651e-01 -2.50318050e-01 -7.22687304e-01 -6.72390640e-01
1.60297155e-01 -3.99294913e-01 7.27892101e-01 2.12732792e-01
3.59056890e-01 -5.65750062e-01 6.75326288e-02 7.64802575e-01
1.20367205e+00 -1.96749598e-01 7.95531154e-01 2.67216295e-01
6.59660757e-01 2.79905230e-01 4.18098360e-01 -1.71973839e-01
2.67646611e-01 6.52192593e-01 -1.68084130e-01 -2.61957765e-01
-3.54490131e-01 -4.41311985e-01 1.19548142e+00 1.91685426e+00
1.21102341e-01 -3.88176262e-01 -7.57271290e-01 6.07666850e-01
-1.60918033e+00 -9.18158948e-01 -1.89750880e-01 2.09893489e+00
1.49916959e+00 -1.49742365e-01 -2.94954982e-02 -3.75301510e-01
8.53682101e-01 -1.33320883e-01 -9.29499418e-02 -8.55267107e-01
-3.99785638e-01 4.30571198e-01 6.86198473e-01 1.03071010e+00
-9.20003653e-01 1.51502740e+00 6.95731115e+00 1.19570291e+00
-9.89625692e-01 7.45237887e-01 6.26320839e-01 -5.10408729e-02
-3.13805133e-01 1.03996791e-01 -6.85056388e-01 3.55980963e-01
1.49261534e+00 -4.33425903e-02 7.86216557e-01 5.07957041e-01
4.80848789e-01 2.81049997e-01 -1.26683152e+00 7.47442484e-01
1.04371823e-01 -1.12546861e+00 1.90192178e-01 7.66259879e-02
1.10651886e+00 6.44158840e-01 -2.31876329e-01 5.36817789e-01
5.30126333e-01 -9.35908437e-01 7.94914663e-01 6.63619488e-02
1.39425778e+00 -7.01226234e-01 7.61365950e-01 6.65614009e-01
-8.08517694e-01 4.94264424e-01 -4.90967900e-01 1.21682391e-01
4.17835146e-01 6.18306339e-01 -9.75961089e-01 7.12376595e-01
2.94232488e-01 7.18002915e-01 -3.56367886e-01 4.55324739e-01
-4.45190430e-01 1.12171435e+00 -2.90358007e-01 3.52101386e-01
2.26844981e-01 -3.01884472e-01 5.87731600e-01 1.87167346e+00
5.89987814e-01 -4.14185226e-01 3.64026334e-03 2.81299174e-01
-4.25845891e-01 2.72510409e-01 -5.60079813e-01 -2.76282638e-01
1.82773069e-01 1.04495811e+00 -2.20692724e-01 -6.29425585e-01
-3.30899924e-01 1.66227543e+00 4.48049814e-01 4.74272847e-01
-6.03121400e-01 -3.21439683e-01 5.73639393e-01 1.25189293e-02
2.35290110e-01 -6.21810615e-01 -4.57119644e-01 -1.51019883e+00
-2.30590943e-02 -1.63376868e+00 2.97321938e-02 -6.83610737e-01
-1.14559281e+00 8.31527054e-01 -4.27247852e-01 -1.03365791e+00
-8.69506299e-01 -3.73723000e-01 -2.73752302e-01 1.38151371e+00
-1.25209212e+00 -1.35051203e+00 4.13071215e-01 4.50148225e-01
9.59415495e-01 -3.22751462e-01 8.67213547e-01 4.81494427e-01
-2.66044199e-01 8.13753307e-01 8.00159216e-01 1.83118761e-01
1.25206304e+00 -1.18813634e+00 9.75741982e-01 1.28886032e+00
3.70882303e-01 5.62202394e-01 7.43301988e-01 -9.33291852e-01
-1.48027539e+00 -1.10537267e+00 1.56134140e+00 -7.98117161e-01
6.33280635e-01 -7.92041183e-01 -7.06147313e-01 1.00293839e+00
9.34830606e-01 -5.98202884e-01 3.54548395e-01 5.88426776e-02
-1.61869004e-01 5.96469454e-02 -8.53701711e-01 6.81314886e-01
1.07953691e+00 -8.82050633e-01 -6.57834828e-01 5.69400311e-01
9.63473618e-01 -5.67973793e-01 -9.14578438e-01 1.74561501e-01
2.29905397e-01 -5.35571635e-01 7.10589170e-01 -5.94637990e-01
7.64715075e-01 6.66880608e-02 -2.42759645e-01 -1.73091376e+00
-3.43660772e-01 -1.15590370e+00 1.65259186e-02 1.30668366e+00
8.05311859e-01 -3.99762243e-01 3.06856811e-01 -2.32670978e-01
-5.78493893e-01 -1.15575217e-01 -9.51762021e-01 -9.17050540e-01
5.44256568e-01 -4.04896051e-01 2.98289806e-01 9.57928538e-01
-5.96155785e-02 8.33944082e-01 -6.60453081e-01 -9.12084207e-02
5.20915508e-01 -4.33321953e-01 9.42545950e-01 -4.48013783e-01
-6.07921720e-01 -3.02791744e-01 1.30719110e-01 -1.27971876e+00
2.46852025e-01 -1.22974074e+00 4.76283401e-01 -1.58955097e+00
3.91986996e-01 2.19677631e-02 1.77865013e-01 4.55233276e-01
-3.87932450e-01 2.84576327e-01 6.10657893e-02 5.41506588e-01
-3.04291129e-01 5.34224570e-01 1.34844792e+00 -4.71225865e-02
-2.87311710e-03 -3.76098603e-01 -4.58207786e-01 1.70416117e-01
6.24891698e-01 -6.90824747e-01 -2.51953006e-01 -1.28402519e+00
5.03892684e-03 1.11636929e-01 -2.33077317e-01 -6.32507622e-01
-6.33361191e-02 3.32846977e-02 1.72082677e-01 -5.89358091e-01
1.38071686e-01 -4.67632413e-01 1.19782008e-01 4.22233164e-01
-3.11922282e-01 5.50731182e-01 3.92687440e-01 1.48068875e-01
-3.79533350e-01 -2.59534299e-01 7.50093877e-01 -2.60119826e-01
-3.06956083e-01 -5.24430685e-02 -6.95697486e-01 3.51318091e-01
3.70244771e-01 2.72096824e-02 -2.83316672e-01 -6.57539308e-01
-4.57721233e-01 -2.58254353e-02 6.69909000e-01 3.84285599e-01
-3.98012474e-02 -1.20618069e+00 -1.30444598e+00 1.84205279e-01
-1.94406122e-01 -2.96292901e-01 -2.87691325e-01 8.51408958e-01
-7.02128232e-01 5.42801082e-01 1.13889133e-03 -7.94948220e-01
-1.20127654e+00 2.69565135e-01 2.87051529e-01 -6.97912157e-01
-4.26021725e-01 6.68128312e-01 -7.57417083e-02 -9.57876086e-01
-4.58204299e-02 -3.62481959e-02 6.51573360e-01 -1.10902689e-01
2.54919231e-01 1.50769055e-01 4.57392514e-01 -9.10587609e-01
-7.54337385e-02 2.49139175e-01 -2.16153145e-01 -8.93034577e-01
1.36285222e+00 -4.03259009e-01 -2.66582906e-01 2.98922062e-01
1.29739809e+00 8.86221826e-02 -1.10080099e+00 -3.53008568e-01
-2.97843665e-03 -9.30457786e-02 -2.17465702e-02 -1.04218829e+00
-6.34121478e-01 1.25764036e+00 2.29747027e-01 -5.07132292e-01
1.03130984e+00 -2.80275166e-01 1.12123299e+00 6.03295803e-01
3.97939116e-01 -1.30248082e+00 -3.93210500e-02 1.18471432e+00
8.31832826e-01 -1.20833755e+00 -3.35487753e-01 -1.51846513e-01
-6.70093715e-01 9.74133372e-01 1.75791368e-01 2.86363125e-01
-5.06218895e-02 5.55532157e-01 4.50355262e-01 5.48916221e-01
-8.57313991e-01 -4.63752337e-02 2.84103155e-01 6.37014270e-01
8.50175023e-01 4.27055173e-02 -3.58896106e-01 2.53855109e-01
-4.70467657e-01 -1.19086616e-01 4.07188535e-01 8.32190216e-01
-3.15027863e-01 -1.64987719e+00 -3.78304005e-01 2.16844589e-01
-6.51603043e-01 -8.14633667e-01 -4.69114512e-01 4.73328918e-01
-9.30024609e-02 1.02265310e+00 -3.33779827e-02 -4.23319936e-01
3.31629328e-02 4.82668847e-01 6.23041749e-01 -7.74687588e-01
-1.02009666e+00 6.23346090e-01 5.64073324e-01 -2.29238972e-01
5.56078320e-03 -7.26082921e-01 -7.67395675e-01 -6.41524374e-01
-3.30534399e-01 2.02779949e-01 8.93479824e-01 1.07073832e+00
4.25611466e-01 2.02511594e-01 5.37753403e-01 -9.02150810e-01
-7.72299707e-01 -1.55171263e+00 6.66372329e-02 3.09770733e-01
3.61935288e-01 1.95427522e-01 -2.30024502e-01 6.24287784e-01] | [11.633651733398438, 10.271897315979004] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.